PHD Thesis PSantos PDF

Assessment of the Shear Strength
between Concrete Layers

Thesis submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy in Civil Engineering, specialty of Mechanics of
Structures and Materials
Author
Pedro Miguel Duarte dos Santos
Scientific Supervisor
Eduardo Nuno Brito Santos Júlio
Coimbra, July, 2009

Avaliação da Resistência ao Corte
entre Camadas de Betão
Tese submetida para obtenção do grau de Doutor em Engenharia Civil,
na especialidade de Mecânica das Estruturas e dos Materiais
Autor
Orientador Científico
Eduardo Nuno Brito Santos Júlio
Coimbra, Julho, 2009

In memory of my father
Ernesto Santos Lucas
Assessment of the Shear Strength between Concrete Layers
ABSTRACT
The bond strength at the interface between concrete layers cast at different ages is
important to ensure the monolithic behaviour of reinforced concrete composite
members. Precast beams with cast-in-place slabs, bridge decks strengthened by adding a
new concrete layer, repair and strengthening of existing concrete structural members by
adding a new concrete layer are typical examples of reinforced concrete composite
members.
Current design codes of reinforced concrete structures, such as the Eurocode 2 (2004)
and the ACI 318 (2008), present design expressions for the assessment of the
longitudinal shear strength of concrete-to-concrete interfaces. These expressions are
based on the shear-friction theory and the longitudinal shear strength is evaluated
considering basically the following four parameters: a) compressive strength of the
weakest concrete; b) normal stress at the interface; c) amount of shear reinforcement
crossing the interface; and d) roughness of the substrate surface.
A qualitative evaluation of the surface roughness is currently adopted by all design

codes. It is common to classify the surface as very smooth, smooth, rough or very rough
by visual inspection. This classification is clearly inaccurate, since it is subjected to
human error. Typical finishing treatments of concrete surfaces are usually linked to this
classification and the values of two coefficients, friction and cohesion, are given to be
adopted in the design expressions.
These design codes do not take into account the curing conditions of the substrate
concrete (old concrete) and of the added concrete layer (new concrete) either and,
therefore, the differential shrinkage is neglected. The differential stiffness, due to the
difference between Young modulus of both concrete layers, is not addressed either. This
has a significant influence because it can create additional stresses at the interface.
For all these reasons, current design expressions need improvements to increase their
accuracy in the prediction of the longitudinal shear strength of concrete-to-concrete
interfaces. This research study aims to add a contribution to the development of such
–I–
design expressions.
Aiming this, three main goals were defined: to develop an in situ non-destructive
method for the characterization of concrete surfaces; to quantify the influence of the
differential shrinkage and of the differential stiffness on the longitudinal shear strength
at the interface between concrete layers cast at different times; and to present design
guidelines for composite reinforced concrete structures.
A new optical measuring device was developed to characterize the texture of concrete
surfaces and a full in situ and non-destructive methodology presented for the assessment
of the bond strength of concrete-to-concrete interfaces.
An experimental study was conducted to investigate the influence of several parameters

on the bond strength of concrete-to-concrete interfaces, including: a) the preparation of
the substrate surface; b) the differential shrinkage and stiffness between the substrate
concrete and the added concrete layer; and c) the failure mode.
Changes to the current shear-friction provisions of Eurocode 2 (2004) (Section 6.2.5)

are proposed.
– II –
RESUMO
A aderência na interface entre camadas de betão de diferentes idades é importante para

assegurar o comportamento monolítico de elementos compósitos de betão armado.
Vigas pré-fabricadas com lajes betonadas in situ, tabuleiros de pontes reforçados por
adição de uma nova camada de betão, reparação e reforço de elementos estruturais de
betão por adição de uma nova camada são exemplos típicos de elementos compósitos de
betão armado.
Os actuais códigos de dimensionamento de estruturas de betão armado, tais como o

Eurocódigo 2 (2004) e o ACI 318 (2008), apresentam expressões de dimensionamento
para a avaliação da resistência ao corte longitudinal de interfaces betão-betão. Estas
expressões são baseadas na teoria do corte-atrito e a resistência ao corte longitudinal é
determinada considerando essencialmente os seguintes quatro parâmetros: a) resistência
à compressão do betão mais fraco; b) tensão normal à interface; c) quantidade de
armadura transversal cruzando a interface; e d) rugosidade da superfície do substrato.
A avaliação qualitativa da rugosidade da superfície é actualmente adoptada por todos os

códigos de dimensionamento. É habitual classificar a superfície como muito lisa, lisa,
rugosa ou muito rugosa recorrendo a uma inspecção visual. Esta classificação é
obviamente pouco precisa, uma vez que se encontra sujeita a erro humano. Tratamentos
correntes de preparação de superfícies de betão são habitualmente interligados com esta
classificação e os valores de dois coeficientes, atrito e coesão, são definidos para serem
adoptados nas expressões de dimensionamento.
Estes códigos de dimensionamento não têm em consideração as condições de cura do

betão do substrato (betão existente) e do betão adicionado (betão novo) e, portanto,
desprezam a retracção diferencial. A rigidez diferencial, devido à diferença entre
módulos de Young de ambas as camadas de betão, é igualmente desprezada. Isto tem
uma influência significativa porque pode criar tensões adicionais na interface.
Por todas estas razões, as actuais expressões de dimensionamento necessitam de

melhoramentos para aumentar a sua precisão na previsão da resistência ao corte
– III –
longitudinal de interfaces betão-betão. Este trabalho de investigação pretende contribuir

para o desenvolvimento das referidas expressões.
Assim sendo, definiram-se três objectivos principais: desenvolver um método in situ

não-destrutivo para a caracterização de superfícies de betão; quantificar a influência da
retracção diferencial e da rigidez diferencial na resistência ao corte longitudinal da
interface entre camadas de betão de diferentes idades; e apresentar especificações de
dimensionamento para estruturas compósitas de betão armado.
Foi desenvolvido um novo equipamento óptico de medição para caracterizar a textura

de superfícies de betão e apresentada uma metodologia não-destrutiva e completamente
in situ para avaliação da resistência à aderência de interfaces betão-betão.
Foi desenvolvido um estudo experimental para investigar a influência de diversos

parâmetros na resistência à aderência de interfaces betão-betão, incluindo: a) a
preparação da superfície do substrato; b) a retracção e rigidez diferencial entre o betão
do substrato e o betão adicionado; e c) o modo de rotura.
São propostas modificações às actuais especificações do corte-atrito do Eurocódigo 2

(2004) (Secção 6.2.5).
– IV –
KEYWORDS
Concrete, Shear Strength, Interface, Texture, Roughness, Differential Stiffness,

Differential Shrinkage.
–V–
This page was intentionally left blank
– VI –
ACKNOWLEDGMENTS
To Prof. Eduardo Júlio, my thesis supervisor, for his assistance, guidance and friendship
in the development of this research work.
To MAPREL - Empresa de Pavimentos e Materiais Pré-Esforçados Lda, SIKA Portugal,

WEBER Cimenfix, AFAssociados - Projectos de Engenharia SA, CIMPOR - Cimentos
de Portugal, BETÃO - LIZ Adémia and Euro-Planning - Engenharia e Gestão Lda,
TrueGage and SYCODE for their financial and material support in the developing of this
work.
To the laboratory technicians Luís Gaspar, Miguel Clara and Olegário João; and to the
students of the Strengthening and Rehabilitation of Structures course, Diogo Carvalho,
Hugo Machado and José Santos, for the collaboration and help in the developing of the
experimental tasks.
To Prof. Jaime Santos and Prof. Helder Araújo, from the Department of Electrical and
Computers Engineering of the University of Coimbra, for the collaboration in the
development of the 2D laser roughness analyser and for making available the 3D laser
scanner, respectively.
To my colleagues and friends, Paulo Fernandes, from the Polytechnic Institute of Leiria,
for making available the formwork for the slant shear specimens; to Jónatas Valença for
helping in the use of the 3D laser scanner; and to Ricardo Costa for helping in the
literature review about shear-friction design expressions.
To my family, to Ana.

pedromdsantos@gmail.com
– VII –
– VIII –
To the Portuguese Science and Technology Foundation (FCT) for their financial
support, PhD Grant number SFRH/BD/25510/2005.
– IX –
–X–
TABLE OF CONTENTS
CHAPTER I - INTRODUCTION
I.1 Overview................................................................................................................I.1
I.2 Objectives..............................................................................................................I.3
I.3 Research Significance............................................................................................I.4
I.4 Thesis Organization...............................................................................................I.5
CHAPTER II - LONGITUDINAL SHEAR STRENGTH
II.1 Introduction..........................................................................................................II.1
II.2 Load Transfer Mechanisms..................................................................................II.2
II.3 Quantification Proposals......................................................................................II.5
II.3.1 Introduction............................................................................................II.5
II.3.2 Design Expressions................................................................................II.7
II.3.2.1 Anderson (1960).........................................................................II.8
II.3.2.2 Hanson (1960).............................................................................II.9
II.3.2.3 Mattock and Kaar (1961)............................................................II.9
II.3.2.4 Saemann and Washa (1964)........................................................II.9
II.3.2.5 Gaston and Kriz (1964).............................................................II.10
II.3.2.6 Birkeland and Birkeland (1966)................................................II.11
II.3.2.7 Badoux and Hulsbos (1967)......................................................II.12
II.3.2.8 Birkeland (1968).......................................................................II.13
II.3.2.9 Mast (1968)...............................................................................II.14
II.3.2.10 Hofbeck, Ibrahim and Mattock (1969).....................................II.14
II.3.2.11 Mattock and Hawkins (1972)....................................................II.15
– XI –
II.3.2.12 Mattock (1974)..........................................................................II.16
II.3.2.13 Hermansen and Cowan (1974)..................................................II.17
II.3.2.14 Mattock, Johal and Chow (1975)..............................................II.17
II.3.2.15 Mattock, Li and Wang (1976)...................................................II.18
II.3.2.16 Raths (1977)..............................................................................II.18
II.3.2.17 Shaikh (1978)............................................................................II.19
II.3.2.18 Loov (1978)..............................................................................II.20
II.3.2.19 Mattock (1981)..........................................................................II.21
II.3.2.20 Vecchio and Collins (1986).......................................................II.21
II.3.2.21 Walraven, Frénay and Pruijssers (1987)...................................II.22
II.3.2.22 Mattock (1988)..........................................................................II.24
II.3.2.23 Mau and Hsu (1988).................................................................II.25
II.3.2.24 Lin and Chen (1989).................................................................II.25
II.3.2.25 Tsoukantas and Tassios (1989)..................................................II.26
II.3.2.26 Patnaik (1992)...........................................................................II.27
II.3.2.27 Loov and Patnaik (1994)...........................................................II.27
II.3.2.28 Mattock (1994)..........................................................................II.28
II.3.2.29 Randl (1997).............................................................................II.29
II.3.2.30 Ali and White (1999)................................................................II.31
II.3.2.31 Valluvan, Kreger and Jirsa (1999).............................................II.31
II.3.2.32 Patnaik (2000)...........................................................................II.32
II.3.2.33 Mattock (2001)..........................................................................II.33
II.3.2.34 Patnaik (2001)...........................................................................II.34
II.3.2.35 Kahn and Mitchell (2002).........................................................II.35
II.3.2.36 Papanicolaou and Triantafillou (2002)......................................II.35
II.3.2.37 Gohnert (2003)..........................................................................II.37
– XII –
II.3.2.38 Mansur, Vinayagam and Tan (2008).........................................II.38
II.3.3 Design Codes........................................................................................II.39
II.3.3.1 CEB-FIP Model Code (1990)...................................................II.39
II.3.3.2 Eurocode 2 (2004).....................................................................II.43
II.3.3.3 BS 8110-1 (1997)......................................................................II.44
II.3.3.4 ACI 318 (2008).........................................................................II.46
II.3.3.5 CAN/CSA A23.3 (2004)...........................................................II.48
II.3.3.6 AASHTO LRFD Bridge Design Specifications (2007)............II.49
II.3.3.7 AASHTO Standard Specifications for Highway Bridges (1996)

...................................................................................................II.50
II.3.3.8 PCI Design Handbook (2004)...................................................II.52
II.3.4 Comparative Analysis Between Quantification Proposals...................II.53
II.3.4.1 Design Expressions...................................................................II.53
II.3.4.2 Design Codes............................................................................II.61
II.4 Conclusions........................................................................................................II.67
CHAPTER III - CHARACTERIZATION OF CONCRETE SURFACES
III.1 Introduction.........................................................................................................III.1
III.2 Methods of Surface Preparation..........................................................................III.2
III.2.1 Mechanical Removal.............................................................................III.2
III.2.2 Particle Impact Removal.......................................................................III.4
III.2.3 Hydrodemolition...................................................................................III.4
III.2.4 Chemical Removal................................................................................III.5
III.3 Measuring Methods of Surface Texture..............................................................III.6
III.3.1 Concrete Surface Profiles......................................................................III.7
III.3.2 Sand Patch Test.....................................................................................III.8
III.3.3 Outflow Meter.......................................................................................III.9
– XIII –
III.3.4 Mechanical Stylus...............................................................................III.10
III.3.5 Circular Track Meter...........................................................................III.12
III.3.6 Digital Surface Roughness Meter.......................................................III.13
III.3.7 Microscopy..........................................................................................III.14
III.3.8 Ultrasounds.........................................................................................III.16
III.3.9 Slit-Island Method...............................................................................III.17
III.3.10 Roughness Gradient Method...............................................................III.18
III.3.11 Close-Range Digital Photogrammetry................................................III.19
III.3.12 Shadow Profilometry..........................................................................III.19
III.3.13 TDI Method.........................................................................................III.20
III.3.14 Patented Devices.................................................................................III.21
III.4 Methods of Characterization.............................................................................III.22
III.4.1 Filters..................................................................................................III.24
III.4.2 Texture Parameters..............................................................................III.28
III.5 Conclusions.......................................................................................................III.34
CHAPTER IV - LASER ROUGHNESS ANALYSER
IV.1 Introduction.........................................................................................................IV.1
IV.2 Development of a 2D Laser Roughness Analyser...............................................IV.3
IV.2.1 Software Application SurfTEX.............................................................IV.5
IV.2.2 Experimental Study...............................................................................IV.7
IV.2.3 Discussion of Results..........................................................................IV.12
IV.2.4 Conclusions.........................................................................................IV.14
IV.3 Comparison of Texture Quantification Methods...............................................IV.16
IV.3.1 Work Description................................................................................IV.16
– XIV –
IV.4 Influence of Filtering on Texture Parameters....................................................IV.26
IV.5 Influence of Surface Irregularities on Texture Parameters................................IV.34
IV.6 Conclusions.......................................................................................................IV.51
CHAPTER V - MATERIALS AND METHODS
V.1 Introduction...........................................................................................................V.1
V.2 Literature Review on Bond Tests..........................................................................V.2
V.2.1 Bending Test...........................................................................................V.4
V.2.2 Bi-Surface Shear Test.............................................................................V.7
V.2.3 Butterfly Test..........................................................................................V.8
V.2.4 Direct Shear Test.....................................................................................V.9
V.2.5 Direct Tension Test...............................................................................V.10
V.2.6 Friction-Transfer Test...........................................................................V.10
V.2.7 Guillotine Test.......................................................................................V.11
V.2.8 Patch Test..............................................................................................V.12
V.2.9 Pull-Off Test.........................................................................................V.13
V.2.10 Push-Off Test........................................................................................V.14
V.2.11 Slant Shear Test....................................................................................V.15
V.2.12 Splitting Test.........................................................................................V.16
V.2.13 Wedge Splitting Test.............................................................................V.17
V.3 Numerical Modelling of the Slant Shear Test.....................................................V.18
– XV –
V.3.1 Influence of the Specimen Geometry...................................................V.19
V.3.2 Influence of the Confinement of the Support Conditions.....................V.22
V.3.3 Influence of the Differential Stiffness between Concrete Parts............V.24
V.3.4 Influence of the Differential Shrinkage between Concrete Parts.........V.28
V.4 Materials and Methods........................................................................................V.30
V.4.1 Concrete Mixture..................................................................................V.31
V.4.2 Curing Conditions.................................................................................V.32
V.4.3 Difference of Ages between Concrete Layers......................................V.35
V.4.4 Concrete Compressive Strength...........................................................V.35
V.4.5 Shrinkage..............................................................................................V.37
V.4.6 Bond Tests.............................................................................................V.40
V.4.7 Surface Preparation and Characterization............................................V.41
V.5 Conclusions.........................................................................................................V.43
CHAPTER VI - ASSESSMENT OF THE BOND STRENGTH
VI.1 Introduction.........................................................................................................VI.1
VI.2 Tasks Description................................................................................................VI.2
VI.3 Experimental Results..........................................................................................VI.5
VI.4 Analytical Approach.........................................................................................VI.11
VI.5 Numerical Modelling........................................................................................VI.16
VI.5.1 Differential Shrinkage.........................................................................VI.16
VI.5.2 Influence of the Differential Stiffness.................................................VI.19
VI.5.3 Combined Effect of Differential Shrinkage and Stiffness with

Compressive Loading.........................................................................VI.20
VI.6 Discussion of Results........................................................................................VI.22
VI.7 Conclusions.......................................................................................................VI.26
– XVI –
CHAPTER VII - CONCLUSIONS, DESIGN GUIDELINES AND FUTURE

DEVELOPMENTS
VII.1 Conclusions.......................................................................................................VII.1
VII.2 Design Guidelines..............................................................................................VII.8
VII.2.1 Introduction.........................................................................................VII.8
VII.2.2 Background Information.....................................................................VII.9
VII.2.2.1 Surface Preparation and Characterization................................VII.9
VII.2.2.2 Coefficients of Cohesion and Friction...................................VII.10
VII.2.2.3 Differential Shrinkage............................................................VII.12
VII.2.2.4 Differential Stiffness..............................................................VII.12
VII.2.2.5 Modified Design Equation.....................................................VII.13
VII.2.3 Proposed Draft for Section 6.2.5 of Eurocode 2 (2004)....................VII.15
VII.3 Future Developments.......................................................................................VII.19
REFERENCES
APPENDIX
– XVII –
– XVIII –
LIST OF FIGURES
CHAPTER I - INTRODUCTION
Figure I.1 - Thesis organization.............................................................................I.6
Figure II.1 - Shear-friction model..........................................................................II.2
Figure II.2 - Possible applications of the shear-friction theory..............................II.3
Figure II.3 - Contribution of adhesion, shear-friction and shear reinforcement

according to Zilch and Reinecke (2001)............................................II.4
Figure II.4 - Simply supported beam: a) with mobilization of shear stresses at

interface; and b) without mobilization of shear stresses at interface.II.5
Figure II.5 - Shear-friction model........................................................................II.12
Figure II.6 - Aggregate size and crack width.......................................................II.22
Figure II.7 - Influence of concrete strength.........................................................II.23
Figure II.8 - Cement matrix and aggregates........................................................II.24
Figure II.9 - Sphere model...................................................................................II.24
Figure II.10 - Indented surface according to Eurocode 2 (2004)...........................II.43
Figure II.11 - Experimental tests: a) Anderson (1); b) Hanson (2); c) Saemann and
Washa (3); d) Hofbeck et al. (4); e) Mattock (5); f) Walraven et al. (6);
and g) Patnaik (7)............................................................................II.55
Figure II.12 - Comparison of design equations for shear-friction.........................II.61
CHAPTER III - CHARACTERIZATION OF CONCRETE SURFACES
Figure III.1 - Scarification.....................................................................................III.3
Figure III.2 - Sand-blasting....................................................................................III.4
Figure III.3 - Hydrodemolition..............................................................................III.5
– XIX –
Figure III.4 - Concrete surface profiles: a) CSP 1; b) CSP 2; c) CSP 3; d) CSP 4;

e) CSP 5; f) CSP 6; g) CSP 7; h) CSP 8; and I) CSP 9.....................III.8
Figure III.5 - Sand Patch Test................................................................................III.9
Figure III.6 - Outflow Meter................................................................................III.10
Figure III.7 - Mechanical stylus...........................................................................III.11
Figure III.8 - Incapacity of the probe tip to measure the surface.........................III.12
Figure III.9 - Circular Track Meter......................................................................III.13
Figure III.10 - Digital Surface Roughness Meter...................................................III.14
Figure III.11 - Scanning Electron Microscope.......................................................III.15
Figure III.12 - Slit-Island Method..........................................................................III.17
Figure III.13 - Roughness Gradient Method..........................................................III.18
Figure III.14 - Shadow Profilometry......................................................................III.20
Figure III.15 - TDI method: a) preparation of the specimen; b) specimen after filled
with additional material; c) sliced specimen; d) digital image of the
interface; e) acquisition of profile coordinates; and f) texture profile.. . .
........................................................................................................III.20
Figure III.16 - Apparatus for measuring surface roughness..................................III.22
Figure III.17 - Sampling, evaluation and traverse length......................................III.23
Figure III.18 - Analysis phases of a surface texture..............................................III.23
Figure III.19 - Primary, waviness and roughness profile.......................................III.25
Figure III.20 - Cut-off length.................................................................................III.25
Figure III.21 - Plateaued surface............................................................................III.27
Figure III.22 - Average Roughness........................................................................III.29
Figure III.23 - Different surfaces with same average roughness...........................III.30
Figure III.24 - Mean Peak Height and Mean Valley Depth...................................III.31
Figure III.25 - Mean Peak-to-Valley Height..........................................................III.32
Figure III.26 - Ten Points Height...........................................................................III.32
Figure III.27 - Maximum Peak Height and Maximum Valley Depth......................III.33
– XX –
Figure III.28 - Total Roughness Height..................................................................III.34
Figure IV.1 - Laser roughness analyser: a) internal view; b) aluminium cover box;
and c) equipment, connected to a laptop, being used.......................IV.4
Figure IV.2 - Graphical user interface....................................................................IV.5
Figure IV.3 - SurfTEX application.........................................................................IV.6
Figure IV.4 - Inspected surfaces: a) left as-cast; b) wire-brushed; and c) sand-

blasted...............................................................................................IV.7
Figure IV.5 - Roughness profile of the surface left as-cast....................................IV.8
Figure IV.6 - Roughness profile of the wire-brushed surface................................IV.8
Figure IV.7 - Roughness profile of the sand-blasted surface.................................IV.8
Figure IV.8 - Roughness parameters assessed with the 2D laser roughness analyser..
..........................................................................................................IV.9
Figure IV.9 - Roughness parameters obtained with both laser sensors for the left as-
cast surface.....................................................................................IV.10
Figure IV.10 - Coefficient of variation of the adopted roughness parameters.......IV.10
Figure IV.11 - Bond strength in shear....................................................................IV.11
Figure IV.12 - Bond strength in tension.................................................................IV.11
Figure IV.13 - Roughness parameters assessed for the wire-brushed surface.......IV.12
Figure IV.14 - Roughness parameters assessed for the sand-blasted surface........IV.12
Figure IV.15 - Roughness parameters assessed for the left as-cast surface...........IV.13
Figure IV.16 - Konica Minolta VIVID 910............................................................IV.17
Figure IV.17 - 3D surface obtained from Konica Minolta VIVID 910..................IV.18
Figure IV.18 - Coefficient of variation of the primary parameters........................IV.24
Figure IV.19 - Coefficient of variation of the roughness parameters.....................IV.24
Figure IV.20 - Coefficient of variation of the waviness parameters......................IV.24
Figure IV.21 - Sand Patch Test: a) left as-cast; b) wire-brushed; and c) sand-blasted....
........................................................................................................IV.25
– XXI –
Figure IV.22 - Primary parameters with Gaussian filter (0.8mm)..........................IV.29
Figure IV.23 - Roughness parameters with Gaussian filter (0.8mm).....................IV.29
Figure IV.24 - Waviness parameters with Gaussian filter (0.8mm).......................IV.29
Figure IV.25 - Primary parameters for the Gaussian filter.....................................IV.30
Figure IV.26 - Roughness parameters for the Gaussian filter................................IV.30
Figure IV.27 - Waviness parameters for the Gaussian filter...................................IV.30
Figure IV.28 - Primary parameters for the Spline filter.........................................IV.31
Figure IV.29 - Roughness parameters for the Spline filter.....................................IV.31
Figure IV.30 - Waviness parameters for the Spline filter.......................................IV.31
Figure IV.31 - Primary parameters for the Rk filter................................................IV.32
Figure IV.32 - Roughness parameters for the Rk filter............................................IV.32
Figure IV.33 - Waviness parameters for the Rk filter..............................................IV.32
Figure IV.34 - Surface preparation: a) left as-cast; b) free after vibration; c) wire-
brushed; d) sand-blasted; e) shot-blasted; and f) hand-scrubbed.. .IV.36
Figure IV.35 - Roughness parameters....................................................................IV.41
Figure IV.36 - Average and deviation of the Average Roughness..........................IV.42
Figure IV.37 - Average and deviation of the Mean Peak-to-Valley Height............IV.42
Figure IV.38 - Average and deviation of the Maximum Peak-to-Valley Height.....IV.42
Figure IV.39 - Average and deviation of the Mean Third Highest Peak-to-Valley
Height.............................................................................................IV.43
Figure IV.40 - Average and deviation of the Maximum Third Highest Peak-to-Valley
Height.............................................................................................IV.43
Figure IV.41 - Average and deviation of the Ten Points Height............................IV.43
Figure IV.42 - Average and deviation of the Total Roughness Height...................IV.44
Figure IV.43 - Average and deviation of the Root-Mean-Square Profile Height...IV.44
Figure IV.44 - Average and deviation of the Mean Peak Height...........................IV.44
Figure IV.45 - Average and deviation of the Maximum Peak Height....................IV.45
Figure IV.46 - Average and deviation of the Mean Valley Depth..........................IV.45
– XXII –
Figure IV.47 - Average and deviation of the Maximum Valley Depth....................IV.45
Figure IV.48 - Coefficient of variation...................................................................IV.47
Figure IV.49 - Influence of the Chauvenet's criterion on the decrease of the

coefficient of variation....................................................................IV.48
Figure V.1 - Types of failures: a) in the bonding agent; b) in the old concrete; c) in
the new concrete; d) between the bonding agent and the old concrete;
e) between the bonding agent and the new concrete; and f) mixed,
crossing the new concrete, the bonding agent and the old concrete.. V.2
Figure V.2 - Bending test proposed by Ohama et al. (1986)..................................V.5
Figure V.3 - Bending test proposed by Wall et al. (1986)......................................V.5
Figure V.4 - Bending test proposed by Abu-Tair et al. (1996)...............................V.6
Figure V.5 - Bending test proposed by Kunieda et al. (2000)................................V.6
Figure V.6 - Bending test proposed by Kamada and Li (2000).............................V.7
Figure V.7 - Bi-surface shear test...........................................................................V.8
Figure V.8 - Butterfly test: a) perspective; and b) side view..................................V.9
Figure V.9 - Direct shear test: a) single shear plane; b) double shear plane........V.10
Figure V.10 - Direct tension test............................................................................V.10
Figure V.11 - Friction-transfer test.........................................................................V.11
Figure V.12 - Guillotine test: a) front view; and b) side view................................V.12
Figure V.13 - Patch test: a) in compression; and b) in tension...............................V.12
Figure V.14 - Patch test: a) with positive bending; and b) with negative bending........
..........................................................................................................V.13
Figure V.15 - Pull-off test.......................................................................................V.13
Figure V.16 - Push-off test.....................................................................................V.14
Figure V.17 - Slant shear test.................................................................................V.15
Figure V.18 - Splitting test: a) cubic specimen; b) cylindrical specimen...............V.16
Figure V.19 - Wedge splitting test..........................................................................V.17
– XXIII –
Figure V.20 - Finite elements mesh........................................................................V.19
Figure V.21 - Mean line of the interface................................................................V.19
Figure V.22 - Specimen´s geometry.......................................................................V.21
Figure V.23 - Variation of the shear stress with the height of the smaller side of half
specimen..........................................................................................V.21
Figure V.24 - Variation of the normal stress with the height of the smaller side of
half specimen...................................................................................V.22
Figure V.25 - Variation of the shear stress with the restraining of the plates.........V.23
Figure V.26 - Variation of the normal stress with the restraining of the plates......V.24
Figure V.27 - Concrete strength class of the substrate and added concrete layer.. V.25
Figure V.28 - Variation of the shear stress with different Young modulus for the
added layer, with restrained supports...............................................V.25
Figure V.29 - Variation of the normal stress with different Young modulus for the
added layer, with restrained supports...............................................V.26
Figure V.30 - Variation of the shear stress with different Young modulus for the
added layer, with unrestrained supports...........................................V.26
Figure V.31 - Variation of the normal stress with different Young modulus for the
added layer, with unrestrained supports...........................................V.27
Figure V.32 - Normal versus shear stresses at the interface...................................V.28
Figure V.33 - Variation of the shear stress with the differential shrinkage strain.. V.29
Figure V.34 - Variation of the normal stress with the differential shrinkage strain.......
..........................................................................................................V.29
Figure V.35 - Variation of the equivalent stress (von Mises stress) with the
differential shrinkage strain.............................................................V.30
Figure V.36 - Granulometric, reference and mixture curves of the adopted concrete
mixture.............................................................................................V.32
Figure V.37 - Curing conditions: a) laboratory; and b) exterior.............................V.32
Figure V.38 - Hygro-thermograph..........................................................................V.33
Figure V.39 - Variation with time of the temperature in the laboratory.................V.34
Figure V.40 - Variation with time of the relative humidity in the laboratory.........V.34
Figure V.41 - Variation with time of the temperature in the exterior.....................V.34
– XXIV –
Figure V.42 - Variation with time of the relative humidity in the exterior.............V.34
Figure V.43 - Concrete compressive strength........................................................V.36
Figure V.44 - Concrete specimen for measuring deformations due to shrinkage.. V.37
Figure V.45 - Variation with time of the concrete shrinkage strain.......................V.38
Figure V.46 - Adopted bond tests: a) slant shear test; and b) splitting test............V.41
Figure V.47 - Surface preparation and a typical texture profile: a) left as-cast;
b) wire-brushed; c) sand-blasted; d) shot-blasted; and e) hand-
scrubbed...........................................................................................V.42
Figure VI.1 - Placement of the substrate concrete for the slant shear test (top view)...
..........................................................................................................VI.3
Figure VI.2 - Placement of the added concrete for the slant shear test (top view).......
..........................................................................................................VI.3
Figure VI.3 - Concrete placement for the slant shear test with the hand-scrubbed
surface (side view)...........................................................................VI.4
Figure VI.4 - Concrete placement for the splitting test (side view): a) cast of the
concrete substrate; and b) placement of the added concrete............VI.4
Figure VI.5 - Concrete placement for the splitting test with the hand-scrubbed
surface (side view): a) cast of the concrete substrate; and b) placement
of the added concrete.......................................................................VI.5
Figure VI.6 - Slant shear test.................................................................................VI.6
Figure VI.7 - Bond strength in shear for L series...................................................VI.7
Figure VI.8 - Bond strength in shear for E series..................................................VI.7
Figure VI.9 - Splitting test.....................................................................................VI.8
Figure VI.10 - Splitting test procedure: a) proposed by EN 12390-6 (2004); and

b) adopted.........................................................................................VI.9
Figure VI.11 - Bond strength in tension for L series...............................................VI.9
Figure VI.12 - Bond strength in tension for E series.............................................VI.10
Figure VI.13 - Failure mode: a) adhesive; and b) cohesive...................................VI.11
Figure VI.14 - Number of cohesive failures in shear.............................................VI.11
– XXV –
Figure VI.15 - Failure envelope of both substrate concrete and added concrete...VI.12
Figure VI.16 - Pure shear stress for slant shear specimens presenting adhesive failure.
........................................................................................................VI.13
Figure VI.17 - Mohr-Coulomb criterion................................................................VI.13
Figure VI.18 - Pure shear strength, for L series, for a difference of 28, 56 and 84 days
between substrate and added concrete layer..................................VI.14
Figure VI.19 - Pure shear strength, for L series, for different surface preparation.........
........................................................................................................VI.15
Figure VI.20 - Pure shear strength, for E series, for a difference of 28, 56 and 84 days
between substrate and added concrete layer..................................VI.15
Figure VI.21 - Pure shear strength, for E series, for different surface preparation.........
........................................................................................................VI.15
Figure VI.22 - Shear stress distribution at the interface........................................VI.18
Figure VI.23 - Normal stress distribution at the interface.....................................VI.18
Figure VI.24 - Equivalent stress (von Mises stress) distribution at the interface.. VI.18
Figure VI.25 - Shear stress distribution at the interface........................................VI.21
Figure VI.26 - Normal stress distribution at the interface.....................................VI.21
Figure VI.27 - Equivalent stress (von Mises stress) distribution at the interface.. VI.21
Figure VI.28 - Comparison of the pure shear strength for both curing conditions.........
........................................................................................................VI.24
Figure VI.29 - Differential stiffness versus cohesive failures................................VI.25
Figure VI.30 - Stress concentrations in the slant shear specimen..........................VI.25
Figure VI.31 - Correlation between the Maximum Valley Depth and the number of
cohesive failures.............................................................................VI.26

DEVELOPMENTS
Figure VII.1 - Correlation between the Mean Valley Depth (Rvm) and the coefficients
of cohesion and friction................................................................VII.12
– XXVI –
LIST OF TABLES
Table II.1 - Shear-friction time line......................................................................II.7
Table II.2 - Values for the constants of the design expression...........................II.30
Table II.3 - Coefficients of friction and cohesion...............................................II.36
Table II.4 - Coefficients of cohesion and friction according to CEB-FIP Model

Code 1990 (1990)............................................................................II.41
Table II.5 - Coefficient of cohesion and friction according to Eurocode 2 (2004)....

.........................................................................................................II.44
Table II.6 - Angle of internal friction according to BS 8110-1 (1997)...............II.45
Table II.7 - Design ultimate longitudinal shear stress according to BS 8110-1

(1997)...............................................................................................II.45
Table II.8 - Coefficient of friction proposed by ACI 318 (2008).......................II.47
Table II.9 - Cohesion and coefficient of friction proposed by CAN/CSA A23.3

(2004)...............................................................................................II.49
Table II.10 - Cohesion, coefficient of friction and factors K1 and K2 according to

AASHTO LFRD Bridge Design Specification (2007)....................II.50
Table II.11 - Coefficient of friction proposed by AASHTO Standard Specifications

for Highway Bridges (1996)............................................................II.51
Table II.12 - Coefficient of friction proposed by PCI Design Handbook (2004).II.53
Table II.13 - Comparison between design expressions........................................II.56
Table II.14 - Comparison between design codes..................................................II.62
Table II.15 - Comparison between CAN/CSA A23.3 (2004), ACI 318 (2008), PCI
Design Handbook (2004) and AASHTO Standard Specifications for
Highway Bridges (1996) and AASHTO LFRD Bridge Design
Specification (2007)........................................................................II.65
– XXVII –
Table IV.1 - Comparison of texture quantification methods................................IV.2
Table IV.2 - Roughness parameters......................................................................IV.9
Table IV.3 - Correlation between roughness parameters and bond strength......IV.13
Table IV.4 - Comparison of the notation of the texture parameters...................IV.19
Table IV.5 - Primary parameters.........................................................................IV.20
Table IV.6 - Roughness parameters....................................................................IV.20
Table IV.7 - Waviness parameters......................................................................IV.21
Table IV.8 - Analysis of primary parameters......................................................IV.22
Table IV.9 - Analysis of roughness parameters..................................................IV.23
Table IV.10 - Analysis of waviness parameters....................................................IV.23
Table IV.11 - Roughness parameters for the left as-cast surface..........................IV.37
Table IV.12 - Roughness parameters for the free after vibration surface.............IV.38
Table IV.13 - Roughness parameters for the wire-brushed surface......................IV.38
Table IV.14 - Roughness parameters for the sand-blasted surface.......................IV.39
Table IV.15 - Roughness parameters for the shot-blasted surface........................IV.39
Table IV.16 - Roughness parameters for the hand-scrubbed surface...................IV.40
Table IV.17 - Decrease of the coefficient of variation with the application of the
Chauvenet's criterion......................................................................IV.40
Table V.1 - Temperature and relative humidity..................................................V.33
Table V.2 - Concrete compressive strength........................................................V.36
Table VI.1 - Material and time-dependent properties........................................VI.17
Table VI.2 - Difference between Young modulus of the slant shear specimens..........
........................................................................................................VI.19
– XXVIII –

DEVELOPMENTS
Table VII.1 - Coefficient of cohesion.................................................................VII.10
Table VII.2 - Coefficient of friction...................................................................VII.11
APPENDIX
Table A.1 - Concrete compressive strength (L series)..........................................A.2
Table A.2 - Concrete compressive strength (E series).........................................A.3
Table A.3 - Experimental shrinkage strain...........................................................A.4
Table A.4 - Shrinkage strain for the specimens cured in the laboratory...............A.5
Table A.5 - Shrinkage strain for the specimens cured in the exterior...................A.6
Table A.6 - Failure load of the slant shear specimens (L28 Series).....................A.7
Table A.9 - Failure load of the slant shear specimens (E28 Series)...................A.10
Table A.12 - Failure load of the splitting specimens (L28 Series).......................A.13
Table A.15 - Failure load of the splitting specimens (E28 Series).......................A.16
– XXIX –
– XXX –
NOTATIONS
Acronyms
AA – Arithmetic Average
ACI – American Concrete Institute
ANSI – American National Standards Institute
ASTM – American Society for Testing and Materials
AVG – Average
BS – British Standard
CCD – Charge Coupled Device
CEB – Euro-International Concrete Committee
CLA – Center Line Average
COV – Coefficient of Variation
CSA – Canadian Standards Association
CSP – Concrete Surface Profile
CTM – Circular Track Meter
DIN – German Institute for Standardization
DSRM – Digital Surface Roughness Meter
FAV – Free After Vibration
FIP – International Federation for Prestressing
FHWA – United States Federal Highway Administration
– XXXI –
HS – Hand-Scrubbing
ICRI – International Concrete Repair Institute
INPI – National Institute of Industrial Property
ISO – International Organization for Standardization
JIS – Japanese Standards Association
LAC – Left as-Cast
LNEC – National Laboratory of Civil Engineering
LVDT – Linear Variable Differential Transducer
MPD – Mean Profile Depth
MTD – Mean Texture Depth
PCMCIA – Personal Computer Memory Card International Association
RH – Relative Humidity
RMS – Root Mean Square
SAB – Sand-Blasting
SC – Strength Control
SEM – Scanning Electron Microscopy
SHB – Shot-Blasting
SHR – Shrinkage
SIM – Slit Island Method
SLS – Serviceability Limit State
SPT – Sand Patch Test
– XXXII –
SRA – Shrinkage Reducing Admixture
STD – Standard Deviation
TDI – Treatment of Digital Image
TEM – Transmission Electron Microscopy
ULS – Ultimate Limit State
USPTO – United States Patent and Trademark Office
WB – Wire-Brushing
Latin upper case letters
A – Area
Ac – Area of the cross section of a concrete element
As – Area of the steel reinforcement
C – Generalized cohesion term (Papanicolaou and Triantafillou,

2002)
C1 – Equation parameter (Walraven et al., 1987)
C2 – Equation parameter (Walraven et al., 1987)
Cs – Equation parameter (Raths, 1977)
D – Diameter, Fractal dimension
EL – Evaluation length
F – Force
I – Moment of inertia (2nd moment of area)
– XXXIII –
K1 – Equation parameter (Mattock, 1988)
L – Length, Length of the contact line of the splitting test
LO – Sampling length
P – Perimeter
Pa – Primary average
Pmax – Maximum primary depth
Pp – Maximum profile peak height
Pq – Root-mean-square primary
Ppm – Mean peak height
Pt – Maximum height of the profile
Ptm – Mean total height
Pv – Maximum profile valley depth
Pvm – Mean valley depth
R – Roughness parameter
R3z – Mean third highest peak-to-valley height
R3z,max – Maximum third highest peak-to-valley height
Ra – Average roughness
Rmax – Maximum peak-to-valley height, Maximum roughness depth
Rp – Maximum peak height
– XXXIV –
Rpm – Mean peak height
Rq – Root-mean-square profile height
Rt – Maximum height of the profile
Rtm – Mean total height
Rv – Maximum valley depth
Rvm – Mean valley depth
Ry – Total roughness height
Rz – Mean peak-to-valley height, Ten-points height
Rz(DIN) – Mean peak-to-valley height
Rz(ISO) – Ten-points height
Rz(JIS) – Ten-point height of irregularities
RH0 – Parameter equal to 100% (Eurocode 2, 2004)
S – Static moment (1st moment of area)
SL – Sampling length
T – Temperature
TL – Traverse length
V – Volume, Shear force
VR – Coefficient of variation
W – Section modulus
Wa – Waviness average
Wmax – Maximum waviness depth
– XXXV –
Wp – Maximum profile peak height
Wpm – Mean peak height
Wq – Root-mean-square waviness
Wt – Waviness height
Wtm – Mean total height
Wv – Maximum profile valley depth
Wvm – Mean valley depth
X – Ration between the shear span and the effective depth of the
section (Saemann and Washa, 1964)
Latin lower case letters
a – Equation parameter (Ali and White, 1999)
ad – Absolute volume of the admixture for unit of apparent volume

of concrete
ag – Absolute volume of the aggregates for unit of apparent volume

of concrete
b – Equation parameter (Ali and White, 1999; Papanicolaou and

Triantafillou, 2002), Width of the cross section
c – Cohesion, Coefficient of cohesion, Coefficient that quantifies

the interface size effect (Papanicolaou and Triantafillou, 2002),
Absolute volume of cement for unit of apparent volume of
concrete
cd – Design value of the coefficient of cohesion
– XXXVI –
d – Equation parameter (Papanicolaou and Triantafillou, 2002),

Effective depth (Mattock and Kaar, 1961), Dimension of the
cross section of the splitting specimen
di – Initial distance between two points
df – Final distance between two points
fc – Compressive strength of concrete
fcd – Design value of the compressive strength of concrete
fc,prism – Compressive strength of concrete obtained with prismatic

specimens
fc,cube – Compressive strength of concrete obtained with cubic specimens
fc,cyl – Compressive strength of concrete obtained with cylindrical

specimens
fci – Positive compressive stress due to internal and external loads

(Vecchio and Collins, 1986)
fcm – Mean compressive strength of concrete
fcmo – Parameter equal to 10MPa (Eurocode 2, 2004)
fs – Equation parameter (Mattock, 1961)
ft – Tensile strength of concrete
fct – Tensile strength of concrete
fctd – Design value of the tensile strength of concrete
fctm – Mean tensile strength of concrete
fy – Yield strength of reinforcement
h – Height of the cross section
– XXXVII –
ho – Notional size of the cross section
k – Equation parameter (Loov, 1978; Mau and Hsu, 1988; Randl,

1997), Factor that defines the target fractile of the Normal
Distribution
kh – Coefficient that depends on the notional size of the cross section
ko – Equation parameter (Anderson, 1960)
n – Number of points
p – Peak height
s – Displacement, Absolute volume of the addition for unit of

apparent volume of concrete
t – Age of the concrete at the moment considered
ts – Age of the concrete at the beginning of the drying shrinkage
u – Perimeter of the cross section exposed to drying
v – Valley depth, Distance from a fiber to the neutral axis
vv – Void volume for unit of apparent volume of concrete
w – Water volume for unit of apparent volume of concrete,

Displacement, Crack width, Dilatancy
x – Shear span (Mattock and Kaar, 1961)
z – Peak-to-valley height
Greek lower case letters
α – Equation parameter (Mau and Hsu, 1988), Coefficient for the

flexural resistance of reinforcement (Randl, 1997), Parameter
– XXXVIII –
that depends on cement class type, Angle of the shear plane of

the slant shear specimen, Angle between the shear reinforcement
and the shear plane, Weighting factor
β – Coefficient allowing for angle of concrete diagonal strut

(Randl ,1997), Reliability factor
βRH – Parameter to relative humidity (Eurocode 2, 2004)
β(τ, τs) – Function that depends on notional size of the cross section and
on the age of the concrete at the moment considered and at the
beginning of the drying shrinkage
γcoh – Partial safety factor for the coefficient of cohesion, Partial safety
factor for the cohesion (Randl, 1997)
γs – Partial safety factor for the shear reinforcement (Randl, 1997)
γc – Partial safety factor for the concrete (Randl, 1997)
γfr – Partial safety factor for the coefficient of friction
εcs – Shrinkage strain
εcd – Drying shrinkage strain
εca – Autogenous shrinkage strain
εcd,0 – Basic drying shrinkage strain
θ – Orientation of the shear reinforcement crossing the interface
λ – Correction factor related with the concrete density
µ – Coefficient of friction, Average
µe – Effective coefficient of friction
– XXXIX –
µd – Design value of the coefficient of friction
ν – Reduction factor for strength of concrete diagonal strut (Randl,

1997), Strength reduction factor
νo – Equation parameter (Anderson, 1960)
νRdi – Design value of the shear resistance at the concrete-to-concrete

interface
νRdi,coh – Design value of the shear resistance at the concrete-to-concrete

interface due to cohesion
νRdi,fr – Design value of the shear resistance at the concrete-to-concrete

interface due to friction
νu – Ultimate horizontal shear stress of the interface
νcimax – Maximum shear stress that a crack can resist (Vecchio and
Collins, 1986)
ρ – Reinforcement ratio (As/Ac)
σ – Normal stress, Standard deviation
σn – Normal stress to the interface
σs – Normal stress of the shear reinforcement
τ – Tangential stress
τ(s) – Tangential stress function at the interface
τo – Pure shear stress
τadh – Tangential stress due to adhesion
τcoh – Tangential stress due to adhesion (Randl, 1997)
– XL –
τsf – Tangential stress due to shear-friction
τsr – Tangential stress due to shear reinforcement
φ – Internal friction angle, Capacity reduction factor
– XLI –
– XLII –
Chapter I
Introduction
I.1 Overview
Concrete structures are usually repaired and/or strengthened by adding a new concrete
layer. Concrete jacketing is a wide spread technique where beams and columns are,
partially or totally, involved by a new concrete layer. Bridge decks and building slabs
strengthened by increasing their thickness is another example of this type of
intervention. Precast concrete members are used in almost every type of structure, from
buildings to bridges and tunnels. Very often, concrete members combine precast with
cast-in-place parts. Precast beams with cast-in-place slabs are typical examples of this.
For the given examples, the bond strength at the interface between concrete layers cast
at different times is important to ensure a monolithic behaviour for SLS and safety for
ULS.
Current design codes of reinforced concrete structures, such as: CEB-FIP Model Code
1990 (1990), Eurocode 2 (2004), BS 8110-1 (1997), ACI 318 (2008), CAN/CSA A23.3
(2004), AASHTO LFRD Bridge Design Specifications (2007), AASHTO Standard
Specifications for Highway Bridges (1996) and PCI Design Handbook (2004), present
design expressions for the assessment of the longitudinal shear strength at the interface
between concretes cast at different times. These design expressions are based on the
– I.1 –
shear-friction theory, as proposed by Birkeland and Birkeland (1966), and the following
four fundamental parameters are considered: a) compressive strength of the weakest
concrete; b) normal stress at the interface; c) shear reinforcement crossing the interface;
and d) roughness of the substrate surface.
The surface preparation plays an important role on the interface shear strength. Even
when the substrate concrete is not damaged, it is usual to increase the surface roughness
with the purpose of improving the bond between both concrete layers. Methods like
wire-brushing, sand-blasting, shot-blasting, hydrodemolition and chipping are usually
adopted to prepare the substrate surface. Common to all these design codes is the
qualitative evaluation of the surface roughness of the concrete substrate.
Eurocode 2 (2004) classifies the roughness of the substrate surface as very smooth,
smooth and rough. This classification is clearly inaccurate since it depends on a
subjective assessment of the technician. To help in the classification of the roughness,
typical finishing treatments for the concrete surface are defined, which include: surface
left as-cast against steel, plastic or wooden formwork (very smooth), surface left without
further treatment after vibration (smooth), and surface intentionally roughened, e.g. by
raking, exposing aggregates (rough). Surfaces can also be classified as indented.
Depending on the finishing method of the concrete surface, the values of two factors,
friction and cohesion, are given to be used in the design expressions proposed by
Eurocode 2 (2004). The coefficient of cohesion and the coefficient of friction can vary
100% and 80%, respectively, in relation to the very smooth surface. This shows the
relevance of the surface preparation in concrete-to-concrete interfaces.
To overcome the subjectiveness associated to the classification of the roughness, some

attempts have been made. The International Concrete Repair Institute (ICRI) (1997) has
defined nine standard profiles of increasing roughness in order to classify, by
comparison with these, the roughness of a concrete surface. An alternative to this
method is to use the Sand Patch Test, defined by the standard ASTM E 965 (2001), in
which the surface roughness is assessed by spreading calibrated sand on the surface,
making a circle, measuring its diameter and then correlating this with the mean texture
depth of the surface.
– I.2 –
Equally important is the understanding of the early age and of the long term behaviour
of the reinforced concrete composite member. Cracking and debonding of the interface,
mainly due to the use of materials with different mechanical properties; inadequate
surface preparation; and to thermal and stiffness incompatibility between the substrate
concrete and the added concrete layer, are some of the main problems associated with
reinforced concrete composite members.
These design codes also do not take into account the curing conditions of the substrate
concrete (old concrete) and of the added concrete layer (new concrete). This can have a
significant influence because additional stresses can appear at the interface between
both concrete layers due to differential shrinkage. Differential stiffness due to the
difference between Young modulus is not addressed either.
For these reasons – roughness evaluation, differential shrinkage and differential

stiffness – current design expressions may not be safe and therefore improvements are
needed to increase the accuracy in the prediction of the longitudinal shear strength of
concrete-to-concrete interfaces.
Some studies have been conducted to better understand the behaviour of reinforced
concrete composite structures and construction guidelines already exist for repairing and
strengthening operations, as well as for precast concrete members with cast-in-place
parts. In order to achieve a better bond, the roughness of the substrate surface should be
increased; a quantification method should be adopted to assess the effectiveness of the
surface preparation; abrasive methods should be adopted instead of heavy impact tools;
the added material should present a Young modulus equal or higher relative to the
substrate material; the substrate should be saturated but presenting a dry surface, when
placing the new concrete layer; bonding agents do not lead to an increase in bond when
the surface is adequately prepared. Nevertheless, several issues remained unanswered,
thus justifying the aim of this thesis.
I.2 Objectives
Three major objectives have been defined for this research study:
– I.3 –
First, to develop an in situ non-destructive method for the characterization of concrete

surfaces. To reach this goal, it was decided to improve the method previously developed
by designing a measuring device to obtain 2D texture profiles of the surface of the
concrete substrate. This device should be portable; without requiring contact with the
surface; non-destructive; easy to use; fast to perform; and adequate for both in situ and
laboratory use. The major goal was to overcome the disadvantages of other roughness
quantification methods.
Second, to quantify the influence of the differential shrinkage and of the differential
stiffness on the shear strength at the interface between concrete layers cast at different
times.
Third, from the conclusions reached with the study undertaken, to present design
guidelines for composite reinforced concrete structures, namely to propose a modified
design expression, based on the current shear-friction provisions of Eurocode 2 (2004),
including a quantitative characterization of the surface of the concrete substrate and
taking into account the influence of differential shrinkage and stiffness between
concrete layers.
I.3 Research Significance
A measuring device was specifically developed to characterize the texture of concrete

surfaces in the scope of this PhD thesis: the 2D laser roughness analyser. The principal
aim was to overcome the disadvantages of the TDI method previously developed by the
team, Santos et al. (2007), which was based on the processing of digital image and
presented some disadvantages, namely, the fact of being partially destructive, work
intensive and of including a laboratorial component.
With this device, the method can be performed totally on site and it is really non-
destructive. Moreover, since a fast on site quantitative characterization of the surface
roughness is obtained, it can easily replace the qualitative approach of current design
codes.
It was demonstrated that the proposed new method presents all the advantages, with
– I.4 –
even more accurate results, and overcomes all the disadvantages of the TDI method and
other methods.
I.4 Thesis Organization
This thesis is structured into seven chapters. In this first chapter, it is presented: a
general overview on the research topic; the motivation for the development of this
study; the objectives; the research significance; and the thesis organization.
Chapter 2 includes the state of the art on design expressions to assess the longitudinal
shear strength at the interface between concrete layers cast at different times. The shear-
friction theory is presented and the load transfer mechanisms between concrete parts are
described. Several quantification proposals, some of these adopted in design codes of
reinforced concrete structures, are presented and compared.
In chapter 3, a bibliographic review on texture of concrete surfaces is presented. First,

surface preparation methods are presented. Then, measuring devices and roughness
quantification methods, using both mechanical and optical devices, are described and
compared. And finally, the procedures needed to perform a quantitative characterization
of a concrete surface – texture parameters and filters – are described and compared.
In chapter 4, it is described the 2D laser roughness analyser that was specifically

developed to measure and quantify the texture of concrete surfaces. A comparison
between the method developed, based on this device, and other methods is presented.
The influence of filters and the need of filtering in the entire process of characterization
is discussed. The influence of localized surface irregularities, such as deep air holes and
exposed aggregates, is also described.
Chapter 5 presents the materials and methods adopted in the development of the
experimental study. It starts with a bibliographic review on the experimental assessment
of bond strength at concrete-to-concrete interfaces, followed by a numerical simulation
to define the geometry of the adopted slant shear specimen. The concrete mixture;
curing conditions; difference of ages between concrete layers; shrinkage quantification;
bond tests; and surface preparation methods are presented.
– I.5 –
Chapter 6 presents the experimental study conducted to assess the influence of

differential shrinkage and stiffness on the bond strength of the interface between
concretes cast at different times. Tasks are described, results presented and discussed. A
numerical simulation, conducted to better explain the experimental results, is presented.
In chapter 7, are presented the conclusions obtained with the development of this
research study. Design guidelines for structural concrete-to-concrete interfaces are
proposed. Research topics for possible future investigations are enumerated.
In Figure I.1 is presented a graphical illustration about the thesis organization.
Overview
Conclusions
Assessment
Motivation Characterization Development
State of the art Definition of of the bond Design
of the texture of a new
on design materials strength guidelines
Objectives of concrete measuring
expressions and methods between
surfaces device
concrete layers Future
Research
developments
significance
Figure I.1 - Thesis organization.
In the scope of this PhD thesis, following the research studies conducted during the Msc
dissertation (Santos, 2005), several papers were published in international and national
journals and several communications were presented at international and national
conferences. In relation to papers published in international journals (cited in ISI Web of
Knowledge), two have already been published in Construction and Building Materials
(Santos et al., 2007) and in Magazine of Concrete Research (Santos and Júlio, 2008);
two more are accepted for publication in ACI Materials Journal (Santos and Júlio,
2009a) and in Materials and Structures (Santos and Júlio, 2009b); and a fifth paper is
submitted to ACI Materials Journal (Santos and Júlio, 2009c).
– I.6 –
Chapter II
Longitudinal Shear Strength
II.1 Introduction
The present chapter is dedicated to the assessment of the longitudinal shear strength of
concrete-to-concrete interfaces. Herein the load transfer mechanisms of shear forces
between two concrete layers according to the shear-friction theory is described. The
contribution of several parameters, such as: a) cohesion between particles, or adhesion
as mentioned by some researchers; b) friction between concrete parts; and c) shear
reinforcement crossing the interface; is considered.
An extensive bibliographic review was conducted on this subject covering what has
been published between 1960 and 2008. Several milestones are identified. A comparison
between some of the most relevant design expressions proposed is presented.
The shear-friction provisions of several design codes, including the: a) CEB-FIP Model
Code 1990 (1990); b) Eurocode 2 (2004); c) British Standards BS 8110-1 (1997);
d) American Building Code Requirements for Structural Concrete ACI 318 (2008);
e) Canadian Code CAN/CSA A23.3 (2004); f) AASHTO LFRD Bridge Design
Specifications (2007); g) AASHTO Standard Specifications for Highway Bridges
(1996); and h) PCI Design Handbook (2004), are also presented and compared.
– II.1 –
II.2 Load Transfer Mechanisms
The behaviour of concrete-to-concrete interfaces, subjected to longitudinal shear

stresses, can be predicted using the shear-friction theory. This theory was developed
around 1960 and was adopted by several design codes of reinforced concrete structures,
including the CEB-FIP Model Code 1990 (1990); the Eurocode 2 (2004); the BS 8110-1
(1997); the ACI 318 (2008); the CAN/CSA A23.3 (2004); the AASHTO LFRD Bridge
Design Specifications (2007); the AASHTO Standard Specifications for Highway
Bridges (1996); and the PCI Design Handbook (2004).
The shear-friction theory assumes that the transfer mechanism of shear forces at a
concrete-to-concrete interface subjected simultaneously to shear and compression forces
is ensured by friction only. A simple saw-tooth model is usually adopted to exemplify
the basic principles of this theory, Figure II.1. The influence of both shear reinforcement
crossing the interface and normal stresses to the shear plane are considered.
σs
σ
τ
w
τ s
σ
σs
Figure II.1 - Shear-friction model.
The shear-friction theory can be used to predict the shear strength by friction in several
situations. The Argentinian design code CIRSOC 201 (2005) presents a detailed
illustration with several zones of a typical structure where the shear-friction theory can
be applied, Figure II.2. This includes: a) the interface between precast elements with
cast-in-place parts; b) corbels; c) metallic supports subjected mainly to shear forces;
d) regions near supports; e) the connection between precast elements and existing
– II.2 –
concrete; and f) the connection between columns and foundations.
Figure II.2 - Possible applications of the shear-friction theory.

(picture extracted from CIRSOC 201 (2005))
Two different situations are usually considered: a) shear transfer across an initially
cracked plane; and b) shear transfer across an initially uncracked plane.
The shear-friction theory applies to the first case, where cohesion, friction and dowel
action control the behaviour, being design expressions empirically calibrated using
experimental results. This is consistent with main published research studies, where
push-off tests are usually adopted and pre-cracked shear planes are imposed.
The shear transfer across an initially uncracked plane was investigated by Hsu et al.
(1987). In opposition to the shear-friction theory, these authors proposed a shear transfer
theory, based on a truss model, where failure is caused by crushing of concrete struts.
Hwang et al. (2000) proposed a similar theory but applicable to both initially cracked
and uncracked shear planes. Gohnert (2000) also proposed a theory for shear at the
interface between precast concrete and cast-in-place parts. This theory was developed
for both cracked and uncracked sections and is based on an alternative method proposed
– II.3 –
by ACI 318 (1995), considering the tension force change at an elemental segment.
As stated by Zilch and Reinecke (2001), Figure II.3, the shear strength at a concrete-to-
concrete interface can be described by a combination of three different load carrying
mechanisms: a) adhesion; b) shear-friction; and c) shear reinforcement.
τ(s) τsr(s)
τadh(s)
τsf(s)
Figure II.3 - Contribution of adhesion, shear-friction and shear reinforcement according to Zilch
and Reinecke (2001).
The adhesion component is originated by chemical bond connections between the

particles of old and new concrete. When its maximum load capacity is reached,
debonding occurs at the concrete-to-concrete interface and the shear stresses will be
transferred by mechanical interlocking. If the interface is subjected to compression, the
shear stresses will be transferred by shear-friction. With the increase of the relative
displacement between concrete parts, the reinforcement that crosses the interface will be
tensioned and yielding can occur. Therefore, the shear reinforcement will induce
compression at the interface and the shear load will be transferred by friction. Due to
slippage, the shear reinforcement will also be subjected to shear, usually named as
dowel action.
The shear stress at a concrete-to-concrete interface, τ (s), for a given crack with a
relative longitudinal displacement between concrete parts equal to s, is then given by:
 s =adh  s sf  s sr s  (II.1)
where τadh (s) is the contribution of the adhesion, τsf (s) is the contribution of the shear-
– II.4 –
friction and τsr (s) is the contribution of the shear reinforcement for the shear stresses.
II.3 Quantification Proposals
II.3.1 Introduction
Structural design must respect several conditions. Safety is essential but economy can
not be left out. In the case of composite concrete members, with parts cast at different
times, it is fundamental to ensure that stresses are adequately transferred between both
parts. Otherwise, either safety or economy is not achieved.
In Figure II.4, a typical example of a simply supported concrete beam is represented:

a) with mobilization; and b) without mobilization, of the shear stresses at the interface.
This type of example is usually used to introduce this subject in the Mechanics and
Strength of Materials unit course. Consider the simply supported concrete beam, cast at
two different phases with a rectangular cross section with a height/width ratio equal to
2.
a)
b)
Figure II.4 - Simply supported beam: a) with mobilization of shear stresses at interface; and
b) without mobilization of shear stresses at interface.
The moment of inertia of the whole cross section in relation to the horizontal axis is
given by:
– II.5 –
3
b h3 b⋅ 2 b  2 b4
I= = = (II.2)
12 12 3
Therefore, the section modulus is equal to:
I
W= =
2 b4
3   
=
2b 4
3
=
2b
3
(II.3)
v h
2 b 3
Considering that there is not any shear transfer across the interface, the corresponding
moment of inertia is given by:
b h 3 b⋅b 3 b 4
I= = = (II.4)
12 12 12
and the corresponding section modulus is equal to:
I
W= =
b4
12
=
  
b4
12
=
b3 (II.5)
v h
4  
b
2
6
With this simple example, it can be shown that the moment of inertia of the monolithic
cross section is eight times higher and the section modulus four times higher than the
non-monolithic cross section, assuming that there is not shear transfer at the interface.
Considering the same example, the shear stresses at the interface can be computed using
the elastic beam theory, given by:
VS
= (II.6)
Ib
where τ is the shear stress; V is the shear force; S is the first moment of area (static
moment) above or below the interface; I is the second moment of area (moment of
inertia); and b is the section width at the interface level.
The field of application of the elastic theory can be extended to cracked sections if the
geometrical properties, the first and second moment of area, are evaluated for the
– II.6 –
cracked section instead of the uncracked one. This solution is valid for the serviceability
limit states (SLS) but not for the ultimate limit states (ULS). It was also proposed by
Loov and Patnaik (1994) and, according to these researchers, this was the simplest and
most practical method to calculate shear stresses at the interface.
The influence of several parameters, such as: a) material constitutive law; b) existence
of cracking; c) material time-dependent properties (creep, shrinkage and relaxation);
and d) existence of different materials; makes the previous expression unusable, for the
assessment of the shear stresses at the concrete-to-concrete interface, leading to the
development of more accurate design expressions.
II.3.2 Design Expressions
In Table II.1 are presented the publication year and the researcher(s) name(s) that made
contributions to this subject since the shear-friction theory was first proposed in 1960.
Mattock has a particular relevance, since he is the only one present in all decades.
Table II.1 - Shear-friction time line.

1960-1969 1970-1979 1980-1989 1990-1999 2000-2009
1960 1972 1981 1992 2000
Anderson Mattock and Hawkins Mattock Patnaik Patnaik
Hanson
1974 1986 1994 2001
1961 Mattock Vecchio and Collins Loov and Patnaik Mattock
Mattock and Kaar Hermansen and Cowan Mattock Patnaik
1987
1964 1975 Walraven, Frénay and 1997 2002
Saemann and Washa Mattock, Johal and Pruijssers Randl Kahn and Mitchell
Gaston and Kriz Chow Papanicolaou and
1988 1999 Triantafillou
1966 1976 Mattock Ali and White
Birkeland and Birkeland Mattock, Li and Wang Mau and Hsu Valluvan, Kreger and 2003
Jirsa Gohnert
1967 1977 1989
Badoux and Hulsbos Raths Lin and Chen 2008
Tsoukantas and Tassios Mansur, Vinayagam and
1978 Tan
1968
Birkeland Shaikh
Mast Loov
1969
Hofbeck, Ibrahim and
Mattock
– II.7 –
Next, the corresponding design expressions are presented. Since some of these were
originally proposed in imperial units, while others were expressed in SI units, a dual
unit format is adopted.*
Also the notation originally used in design expressions was changed in order to use an
uniform notation in all of them. The same procedure was adopted to the design
expressions proposed by several design codes and later presented.
II.3.2.1 Anderson (1960)
Anderson (1960) was one of the first to propose a design expression to predict the
longitudinal shear strength of concrete interfaces. The proposed expression was of the
following type:
v u =v ok  (II.7)
where νu is the ultimate longitudinal shear stress at the interface; vo and k are two
parameters that are experimentally obtained from push-off tests; and ρ is the
reinforcement ratio.
This expression was calibrated for two different concretes, with a compressive strength
of 20.68MPa (3000psi) and 51.71MPa (7500psi). For the weakest concrete, the design
expression is as follows:
v u =4.41229  (MPa) (II.8)
v u =64033180  (psi) (II.9)
For the strongest concrete, the design expression is as follows:
v u =5.52276  (MPa) (II.10)
v u =80040000  (psi) (II.11)
* Conversion factors from SI Units to Imperial Units:

1mm = 0.039in.
1mm2 = 0.001550in.2
1MPa = 145psi
– II.8 –
II.3.2.2 Hanson (1960)
Hanson (1960) presented a design expression of the same type of the one proposed by
Anderson (1960), also calibrated with experimental results obtained from push-off tests.
The proposed expression was developed for rough interfaces and, probably for this
reason, presented different coefficients than those proposed by Anderson (1960).
v u =3.45121  (MPa) (II.12)
v u =50017500  (psi) (II.13)
where νu is the ultimate longitudinal shear stress at the interface; and ρ is the
reinforcement ratio.
II.3.2.3 Mattock and Kaar (1961)
Mattock and Kaar (1961) proposed a design expression based on the shear
span/effective depth ratio to determine the ultimate longitudinal shear stress at the
interface of composite reinforced concrete beams. The proposed expression is as
follows:
18.6
vu = 121 
 
x
d
5
(MPa) (II.14)
2700
vu = 17500 
 
x
d
5
(psi) (II.15)
where νu is the ultimate longitudinal shear stress at the interface; x is the shear span; d is
the effective depth; and ρ is the reinforcement ratio. The minimum reinforcement ratio
should be equal or higher than 0.15%.
II.3.2.4 Saemann and Washa (1964)
Saemann and Washa (1964) developed an experimental study to determine a design
– II.9 –
expression for the prediction of the longitudinal shear strength of reinforced composite
concrete elements. The proposed expression is as follows:
18.6 33− X
vu = 207  2 (MPa) (II.16)
X 5 X 6 X 5
2700 33− X
vu = 30000  2 (psi) (II.17)
X 5 X 6 X 5
where νu is the ultimate longitudinal shear stress at the interface; ρ is the reinforcement
ratio; and X is the ratio between the shear span and the effective depth of the section.
The surface condition was not considered in the proposed expression, because the
authors concluded that its contribution for the shear strength is variable and decreases
with the increase of the shear reinforcement ratio.
The first term of the proposed expression represents the ultimate longitudinal shear
stress with no reinforcement crossing the interface and the second term represents the
contribution of the clamping stresses when steel reinforcement is added.
II.3.2.5 Gaston and Kriz (1964)
Gaston and Kriz (1964) suggested the following design expressions to estimate the
ultimate longitudinal shear stress in scarf joints of precast concrete.
For smooth unbonded interfaces the design expression is as follows:
v u =0.300.78 n (MPa) (II.18)
v u =430.78 n (psi) (II.19)
For smooth bonded interfaces the design expression is as follows:
v u =0.760.70  n (MPa) (II.20)
v u =1100.70 n (psi) (II.21)
where νu is the ultimate longitudinal shear stress at the interface and σn is the normal
stress at the interface.
– II.10 –
II.3.2.6 Birkeland and Birkeland (1966)
Birkeland and Birkeland (1966) were the first to propose a linear expression to evaluate
the ultimate longitudinal shear stress of concrete interfaces. The proposed expression is
as follows:
v u = f y tan = f y  (II.22)
ratio; fy is the yield strength of the reinforcement; and φ is the internal friction angle.
The tangent of the internal friction angle is also designated as coefficient of friction,
being represented by the Greek letter µ, and the term ρ fy is designated as clamping
stresses.
This expression was proposed for smooth concrete surfaces, artificially roughened
concrete surfaces and concrete-to-steel interfaces. The coefficient of friction was
empirically determined, varying with the surface preparation, and it was defined for
several situations, namely: a) µ = 1.7, for monolithic concrete (59.5º); b) µ = 1.4, for
artificially roughened construction joints (54.5º); and c) µ = 0.8 to 1.0, for ordinary
construction joints and for concrete to steel interfaces (38.7º to 45.0º). This expression
was limited to the following conditions:
1.5% (II.23)
v u5.52
(MPa) (II.24)
f c 27.58
v u 800
(psi) (II.25)
f c 4000
The adopted design philosophy stated that: the tensile strength of concrete should be
neglected; all tensile forces are absorbed by the steel reinforcement; and shear forces are
transmitted by friction, Figure II.5.
As the unbonded concrete parts slide one over another, the crack will open and the
– II.11 –
reinforcement steel will be tensioned. Therefore, it is assumed that the longitudinal

shear reinforcement will compress the interface, resulting in frictional resistance along
the interface.
This expression, usually known as the shear-friction expression, presents several

advantages: a) the shear-friction model, in which it is based, is clearly understandable;
b) the expression is simple and easy to use; and c) it gives accurate results.
Figure II.5 - Shear-friction model.

(picture extracted from Birkeland and Birkeland (1966))
II.3.2.7 Badoux and Hulsbos (1967)
Badoux and Hulsbos (1967) proposed a design expression to predict the ultimate
longitudinal shear stress between precast concrete beams and cast-in-place slabs, when
the composite beam is subjected to repeated loading. The proposed expressions, defined
as conservative by the researchers, takes into account: the reinforcement ratio; the
surface preparation; and the shear span/effective depth ratio of the cross section of the
beam.
Two different types of finishing surfaces were considered. Intermediate surfaces, as

named by the researchers, obtained by applying a retarding agent on the fresh concrete
and by steel brushing in the day after to cast. Rough surfaces were obtained by two
different methods: a) with a board having a protruding nail; and b) by a metal plate with
teeths.
– II.12 –
For construction joints with an intermediate finish, the following expression was
proposed:
13.79
vu = 137.9 
11  
a
d
(MPa) (II.26)
2000
vu = 20000 
11 
a
d
(psi) (II.27)
For rough construction joints the following expression should be adopted:
24.14
vu = 137.9 
11  
a
d
(MPa) (II.28)
3500
vu = 20000 
11 
a
d
(psi) (II.29)
ratio; a is the effective depth of the cross section; and d is the shear span.
According to Badoux and Hulsbos (1967), the proposed design expressions are the sum
of two terms. The first term represents the contribution of the natural bond, as
designated by the researchers, while the second represents the contribution of the
reinforcement crossing the interface.
II.3.2.8 Birkeland (1968)
Birkeland (1968), cited by Patnaik (1992), was the first researcher to introduce a non-
linear expression to predict the ultimate longitudinal shear stress at the interface
between concrete parts. The proposed expression was derived from a previous research
work by the author, Birkeland and Birkeland (1966), but it was only published in lecture
notes. A parabolic function was fitted to the available experimental data, being the
ultimate longitudinal shear stress predicted by:
– II.13 –
v u =2.78   f y (MPa) (II.30)
v u =33.5   f y (psi) (II.31)
ratio; and fy is the yield strength of the reinforcement.
II.3.2.9 Mast (1968)
Mast (1968) adopted the expression proposed by Birkeland and Birkeland (1966) but
suggested an upper limit of (0.15 fc tanφ) for the ultimate longitudinal shear stress.
Different values for the coefficient of friction were also proposed: a) µ = 1.4, for
concrete-to-concrete rough interfaces (54.5º); b) µ = 1.0, for concrete-to-steel composite
beams (45.0º); c) µ = 0.7, for concrete-to-steel field-welded inserts (35.0º); and
d) µ = 0.7, for concrete-to-concrete smooth interfaces (35.0º).
II.3.2.10 Hofbeck, Ibrahim and Mattock (1969)
Hofbeck et al. (1969) performed an experimental study to quantify the influence of the
following parameters on the shear strength of concrete-to-concrete interfaces: a) pre-
cracked shear plane; b) strength, size and arrangement of shear reinforcement;
c) concrete strength; and d) dowel action.
A design expression was not presented but the proposals of others researchers, such as
Gaston and Kriz (1964), Birkeland and Birkeland (1966) and Mast (1968), were
analysed and discussed.
Hofbeck et al. (1969) stated that pre-existing cracks along the shear plane results in a
decrease of the shear strength and in a increase of the relative slip between both
concrete parts. Changes in the strength, geometry and arrangement of shear
reinforcement crossing the interface also affects the clamping stresses ρfy.
The concrete strength showed to have influence in the shear strength of the interface.
For values of the clamping stresses below 4.14MPa (600psi), the concrete strength does
– II.14 –
not affect the shear strength. Above this limit, the shear transfer is affected and the shear
strength increases with the increase of the concrete strength.
The dowel action due to the shear reinforcement was significant only for concrete
specimens with a pre-existing crack along the shear plane. For initially uncracked
specimens, the influence of the dowel action was insignificant since the relative slip
between concrete parts was too small.
The shear strength of initially cracked specimens was computed using the design
expression proposed by Birkeland and Birkeland (1966). A coefficient of friction of 1.4
was adopted. According to Hofbeck et al. (1969), the shear-friction theory gives a
conservative estimate of the shear strength of the concrete-to-concrete interface for
specimens with a pre-existing crack along the shear plane.
II.3.2.11 Mattock and Hawkins (1972)
Mattock and Hawkins (1972) presented a design expression to predict the ultimate
longitudinal shear stress. The lower bound of the experimental tests was represented by:
v u =1.380.8   f y  n  (MPa) (II.32)
v u =2000.8   f y  n  (psi) (II.33)
ratio; fy is the yield strength of the reinforcement; and σn is the normal stress at the
interface.
The ultimate longitudinal shear stress is limited by the minimum value between 0.3fc
and 10.34MPa (1500psi). The clamping stresses have to be higher than 1.38MPa
(200psi). The first term is due to an apparent cohesion of the interface and to dowel
action of the reinforcement and the second term is due to clamping stresses. The
coefficient of friction was equal to 0.8.
– II.15 –
II.3.2.12 Mattock (1974)
The design expression proposed by Mattock and Hawkins (1972) was developed for the
lower bound of the experimental tests used for calibration. In a subsequent publication,
Mattock (1974) presented a modified expression calibrated with average values from
experimental results, given by:
v u =2.760.8   f y  n  (MPa) (II.34)
v u =4000.8   f y  n  (psi) (II.35)
ratio; fy is the yield strength of the reinforcement; and σn is the normal stress at the
interface.
The ultimate longitudinal shear stress is limited to a minimum between 0.3fc and
10.34MPa (1500psi) and the clamping stresses should be higher than 1.38MPa (200psi).
It was also presented a modified design expression to include the orientation of the
reinforcement steel crossing the interface. The proposed expression is as follows:
v u =2.76 sin  f s  0.8sin −0.5sin 2 

2 2
(MPa) (II.36)
v u =400 sin 2  f s  0.8sin 2 −0.5sin 2  (psi) (II.37)
where νu is the ultimate longitudinal shear stress at the interface; θ is the angle between
the reinforcement and the shear plane; and ρ is the reinforcement ratio. The ultimate
longitudinal shear stress is limited to the maximum of 0.3fc.
The term fs was defined from experimental tests, for a coefficient of friction equal to
0.8, and is given by:
f s =0 : 051.3º (II.38)
f s =−1.6 f y cos  38.7º  : 51.3º90º (II.39)
– II.16 –
f s= f y : 90º180º (II.40)
where fy is the yield strength of the reinforcement.
II.3.2.13 Hermansen and Cowan (1974)
Hermansen and Cowan (1974) presented a modified shear-friction theory for the design
of concrete brackets. This new theory was proposed because the adoption of the basic
shear-friction theory led to conservative values of the brackets shear strength.
Using two simple modifications, limiting the use of the shear-friction theory to brackets
where the failure mode by shear is assured and considering an apparent coefficient of
cohesion, these researchers have successfully predicted the shear strength of forty
concrete specimens. The other failure modes, due to bending and secondary causes
(high bearing stresses and anchorage failures), were also considered. For failure in
bending, a design expression was presented and for failure due to secondary causes it
has been considered that this can be avoided by adequately detailing.
Hermansen and Cowan (1974) also concluded that the ultimate longitudinal shear stress
of an uncracked concrete interface, which fails in shear, can be predicted by the
following expression:
v u =4.00.8  f y (MPa) (II.41)
v u =5800.8  f y (psi) (II.42)
II.3.2.14 Mattock, Johal and Chow (1975)
Mattock et al. (1975) investigated the suitability of using two design expressions, one
proposed by himself, Mattock (1974), and another suggested by Birkeland (1968), for
the design of concrete connections, such as corbels and columns foundations, where the
bending moment and normal forces must be transferred between concrete parts.
– II.17 –
Mattock et al. (1975) concluded that both design expressions can be used for this
purpose, being the expression proposed by Birkeland (1968) more conservative than his
proposal. A slight modification of the design expression was proposed, adopting a
capacity reduction factor equal to 0.85.
II.3.2.15 Mattock, Li and Wang (1976)
Mattock et al. (1976) presented a design expression for lightweight reinforced concrete
with an initially cracked condition. For all (aggregates and sand) lightweight concrete,
the design expression is as follows:
v u =1.380.8  f y (MPa) (II.43)
v u =2000.8  f y (psi) (II.44)
For sanded lightweight concrete, the design expression is as follows:
v u =1.720.8  f y (MPa) (II.45)
v u =2500.8  f y (psi) (II.46)
The maximum value for the ultimate longitudinal shear stress is limited to the minimum
value between 0.2fc and 5.52MPa (800psi) for all lightweight concrete and 0.2fc and
6.90MPa (1000psi) for sanded lightweight concrete. These design expressions were
obtained for clamping stresses with a minimum value of 1.38MPa (200psi).
II.3.2.16 Raths (1977)
Raths (1977) suggested a design expression very similar to the one proposed by
Birkeland (1968). The main difference was the inclusion of the concrete density, being
this expression applicable to both normal and lightweight concrete. For monolithic
specimens, the expression is as follows:
– II.18 –
v u =C s 3.11   f y (MPa) (II.47)
v u =C s 37.42   f y (psi) (II.48)
For a smooth interface in shear, with a coefficient of friction equal to 0.6, the design
v u =C s 2.03   f y (MPa) (II.49)
v u =C s 24.49   f y (psi) (II.50)
where νu is the ultimate longitudinal shear stress at the interface; Cs is a constant related
with the concrete density; ρ is the reinforcement ratio; and fy is the yield strength of the
reinforcement.
The following values were proposed for the constant Cs : a) Cs = 1.00, for normal
weight concrete; b) Cs = 0.85, for sand-lightweight concrete; and c) Cs = 0.75, for all
lightweight concrete. An effective coefficient of friction was also defined as follows:
C s2 
e =6.90 (MPa) (II.51)
vu
C s2 
e =1000 (psi) (II.52)
vu
II.3.2.17 Shaikh (1978)
Shaikh (1978) presented a revision of the shear-friction design provisions and suggested
the following expression to evaluate the ultimate longitudinal shear stress of concrete-
to-concrete or concrete-to-steel interfaces:
v u = f y e (II.53)
where νu is the ultimate longitudinal shear stress at the interface; φ is a capacity

reduction factor, equal to 0.85 for shear; ρ is the reinforcement ratio; fy is the yield
strength of the reinforcement; and µe is an effective coefficient of friction.
– II.19 –
The effective coefficient of friction is defined by:
C s2 
e =6.90 (MPa) (II.54)
vu
C s2 
e =1000 (psi) (II.55)
vu
where Cs is a constant related with the concrete density; and µ is the coefficient of
friction.
The concrete density is taken equal to: a) Cs = 1.00, for normal weight concrete;
b) Cs = 0.85, for sand-lightweight concrete; and c) Cs = 0.75, for all-lightweight
concrete. The coefficient of friction is taken equal to: a) µ = 1.4, for concrete-to-
concrete cast monolithically (54.5º); b) µ = 1.0, for a concrete to hardened concrete
interface with a roughness amplitude of 6.4mm (0.25in.) (45.0º); c) µ = 0.6, for a
concrete-to-steel interface (31.0º); and d) µ = 0.4, for a concrete-to-concrete smooth
interface (21.8º).
II.3.2.18 Loov (1978)
Loov (1978), cited by Patnaik (1992), was the first researcher to explicitly include the
concrete strength, by proposing the following non-dimensional expression:
vu
fc
=k

 f y  n
fc
(II.56)
where νu is the ultimate longitudinal shear stress at the interface; fc is the concrete
compressive strength; k is a constant; ρ is the reinforcement ratio; fy is the yield strength
of the reinforcement; and σn is the normal stress at the interface. For initially uncracked
interfaces, Loov (1978) suggested the value of 0.5 for the constant k.
For a concrete with a compressive strength equal to 30.89MPa (4480psi), this

expression is equal to the one proposed by Birkeland (1968). The proposed design
expression can also be used with any consistent system of units (SI or imperial).
– II.20 –
II.3.2.19 Mattock (1981)
Mattock (1981) investigated the behaviour of concrete-to-concrete interfaces under

cyclic loading. The main goal of this researcher was to determine how the design
expressions, developed for monotonic loading, should be modified in order to be used
also for cyclic loading.
Two design expressions developed by this researcher and his co-workers, Mattock and
Hawkins (1972) and Mattock et al. (1976), for both normal and lightweight concrete,
were adopted to assess the ultimate longitudinal shear stress at the interface of
composite specimens under cyclic loading.
Mattock (1981) suggested that the shear strength of the concrete-to-concrete interface
under cyclic loading, should be taken equal to 0.8 of the shear strength under monotonic
loading, for monolithic specimens made of normal and lightweight concrete and rough
interfaces between concrete parts cast a different ages.
If the bond between concrete parts is destroyed, the shear strength under cyclic loading
should be taken as 0.6 of the shear strength under monotonic loading. It was observed
that the shear transfer mechanism of composite specimens after cracking, for both
monotonic and cyclic loading, is identical to that of monolithic specimens.
II.3.2.20 Vecchio and Collins (1986)
Vecchio and Collins (1986) proposed an analytical model to predict the behaviour of
reinforced concrete members subjected to in-plane shear and normal stresses. Based on
previous investigations by Walraven (1981), these researchers proposed a design
expression for the assessment of the ultimate longitudinal shear stress of a concrete
crack. This expression has the particularity of requiring the estimation of the crack
width, Figure II.6, and is as follows:
f ci2
v u =0.18 v cimax 1.64 f ci−0.82 (II.57)
v cimax
– II.21 –
v cimax =
fc
 0.31
24 w
a16  (MPa; mm) (II.58)
12  f c
v cimax =
 0.31
24 w
a0.63  (psi; in.) (II.59)
where νu is the ultimate longitudinal shear stress of the crack; vcimax is the maximum
shear stress that a crack can resist; fci is the positive compressive stress due to internal
and external loads; fc is the concrete compressive strength; w is the average crack width;
and a is the maximum aggregate size.
Figure II.6 - Aggregate size and crack width.

(picture extracted from Vecchio and Collins (1986))
II.3.2.21 Walraven, Frénay and Pruijssers (1987)
In order to consider the concrete strength, Walraven et al. (1987) developed a large
experimental study with 88 push-off specimens and proposed a non-linear function to
predict the shear strength of initially cracked interfaces.
This design expression, including the reinforcement ratio; the yield strength of the
reinforcement; and the concrete compressive strength, is as follows:
C2
v u =C 1   f y  (MPa) (II.60)
C 1=0.822 f c 0.406 (MPa) (II.61)
– II.22 –
0.303
C 2=0.159 f c (MPa) (II.62)
ratio; fy is the yield strength of the reinforcement; and fc is the concrete compressive
strength of 150mm cubic specimens.
In imperial units, the design expression appears as:
C2
v u =C 1  0.007  f y  (psi) (II.63)
0.406
C 1=15.686 f c (psi) (II.64)
C 2=0.0353 f c 0.303 (psi) (II.65)
The influence of the concrete strength on the longitudinal shear strength of the interface
is clearly visible in Figure II.7.
16
14
Shear strength (MPa)
12
fc = 20MPa
10 fc = 25MPa
fc = 30MPa
8 fc = 35MPa
6 fc = 40MPa
fc = 45MPa
4 fc = 50MPa
fc = 55MPa
2
fc = 60MPa
0
0 1 2 3 4 5 6 7 8 9 10
Clamping stresses (MPa)
Figure II.7 - Influence of concrete strength.
This design expression is based on a model proposed by Walraven (1981), where the
concrete is represented by the binding paste and the aggregates (assumed as spheres)
and where the interface between both is considered as the weakest zone, Figure II.8, and
therefore cracks will develop along this border Figure II.9.
– II.23 –
Figure II.8 - Cement matrix and aggregates.

(picture extracted from Walraven (1981))
Figure II.9 - Sphere model.

(picture extracted from Walraven (1981))
II.3.2.22 Mattock (1988)
Mattock (1988), in the discussion of the paper by Walraven et al. (1987), presented a
modified design expression incorporating the normal stress at the interface. In this new
expression, the first term represents the shear strength due to cohesion and the second
one represents the shear strength due to friction between aggregates. The modified
0.545
v u =0.467 f c 0.8   f y  n  0.3 f c (MPa) (II.66)
v u =4.5 f 0.545
c 0.8   f y  n 0.3 f c (psi) (II.67)
compressive strength; ρ is the reinforcement ratio; fy is the yield strength of the
reinforcement; and σn is the normal stress at the interface. The ultimate longitudinal
shear stress is limited to the maximum value of 0.3fc.
– II.24 –
II.3.2.23 Mau and Hsu (1988)
Mau and Hsu (1988), also in the discussion of the paper by Walraven et al. (1987),
presented a non dimensional expression. The general form of the proposed design
expression is given by:

vu
fc
=k
 
fy
fc
(II.68)
reinforcement; and the coefficients k and α are two parameters to be experimentally
calibrated.
This expression is identical to the one previously proposed by Loov (1978) for initially
uncracked interfaces. The main difference is that Mau and Hsu (1988) assumed a
constant value of 0.66 for the parameter k, for both initially uncracked and cracked
interfaces. The proposed expression is as follows:
vu
fc
=0.66

fy
fc
0.3 (II.69)
compressive strength; ρ is the reinforcement ratio; and fy is the yield strength of the
reinforcement.
II.3.2.24 Lin and Chen (1989)
According to Lin and Chen (1989), the shear-friction provisions presented by ACI 318
(1983) are too conservative while the proposal of the PCI Design Handbook (1985) is
non-conservative for large values of the clamping stresses. Moreover, Lin and Chen
(1989) stated that the design expressions proposed by Mattock (1988) and Walraven et
al. (1987) are only adequate for concretes with lower compressive strengths, whereas
for concrete with high compressive strengths these could be non-conservative.
– II.25 –
Therefore, Lin and Chen (1989) proposed a design expression, calibrated from sixty
eight push-off tests using concrete with a compressive strength between 20.59MPa and
68.65MPa, given by:
v u =e   f y  n  (II.70)
where νu is the ultimate longitudinal shear stress at the interface; µe is an equivalent

coefficient of friction; ρ is the reinforcement ratio; fy is the yield strength of the
reinforcement; and σn is the normal stress at the interface. The ultimate longitudinal
shear stress is upper limited to the minimum value between 0.3fc and 12.46MPa.
The equivalent coefficient of friction is given by:
0.5
e =1.75  f c
 f y  n   0.8 f c 0.25 (MPa) (II.71)
0.5
e = 
21.12  f c
 f y  n   0.8 f c 0.25 (psi) (II.72)
II.3.2.25 Tsoukantas and Tassios (1989)
Tsoukantas and Tassios (1989) proposed two design expressions to estimate the ultimate
longitudinal shear stress at the interface between precast members. The first expression
was proposed for smooth interfaces and is given by:
v u =0.40  n (MPa) (II.73)
The second expression, for rough interfaces, is as follows:
3
v u =0.5  f 2c  n (MPa) (II.74)
where νu is the ultimate longitudinal shear stress at the interface; σn is the normal stress
at the interface due to internal and external loads; and fc is the concrete compressive
strength.
– II.26 –
II.3.2.26 Patnaik (1992)
Patnaik (1992) proposed a design expression to evaluate the ultimate longitudinal shear
stress of composite concrete beams, based on the results of sixteen simply supported
specimens, given by:
v u =0.6   0.1 f y  fc (MPa) (II.75)
v u =0.6   15 f y  f c (psi) (II.76)
strength. The maximum value of the ultimate longitudinal shear stress is limited to
0.25fc.
II.3.2.27 Loov and Patnaik (1994)
Loov and Patnaik (1994) proposed a design expression based in previous research
studies of Loov (1978). The proposed expression, only applicable to rough interfaces,
combines the effects of the concrete strength, clamping stresses and concrete density.
For composite beams without stirrups, the ultimate longitudinal shear stress at the
interface can be predicted by the following expression:
v u =0.6  0.1 f c (MPa) (II.77)
v u =0.6  15 f c (psi) (II.78)
where νu is the ultimate longitudinal shear stress at the interface; and fc is the concrete
compressive strength.
Combining this design expression with the one proposed by Loov (1978), the ultimate
longitudinal shear stress at the interface of composite beams, with reinforcement
crossing the interface, can be predicted by:
– II.27 –
v u =k    0.1 f y  fc (MPa) (II.79)
v u =k    15 f y  f c (psi) (II.80)
where νu is the ultimate longitudinal shear stress at the interface; k is a constant; λ is a

correction factor related with the concrete density; ρ is the reinforcement ratio; fy is the
yield strength of the reinforcement; and fc is the concrete compressive strength. The
maximum value of the longitudinal shear stress is limited to 0.25fc.
The value of the constant k was suggested to be equal to 0.5 and 0.6, for composite and
monolithic reinforced concrete members, respectively. The coefficient λ is taken equal
to 1.00 for normal weight concrete; 0.85 for sand-lightweight concrete; and 0.75 for all
lightweight concrete.
II.3.2.28 Mattock (1994)
Mattock (1994), commenting the paper by Loov and Patnaik (1994), suggested that the
ultimate longitudinal shear stress at the interface should not be proportional to the
term fc0.5. A modified design expression, defined with the experimental tests data
conducted by Loov and Patnaik (1994), was proposed and is given by:
vu =
  f y f c 0.73 (MPa) (II.81)
4.536
vu =
  f y f c 0.73 (psi) (II.82)
14.25
strength. The maximum value of the longitudinal shear stress is limited to 0.3fc.
This expression was calibrated for the lower bound of the experimental tests data and
can be used to predict the ultimate longitudinal shear stress across a crack in monolithic
normal weight concrete. For initially cracked specimens with a rough interface, Mattock
(1994) suggested the following expression:
– II.28 –
vu =
  f y f c 0.73 −0.02 f (MPa) (II.83)
c
4.536
vu =
  f y f c 0.73 −0.02 f (psi) (II.84)
c
14.25
II.3.2.29 Randl (1997)
Randl (1997) has one of the most significant contributions to the improvement of the
accuracy of the design expressions for the assessment of the ultimate longitudinal shear
stress at concrete-to-concrete interfaces. After Birkeland and Birkeland (1966); Mattock
and Hawkins (1972); Loov (1978); and Walraven et al. (1987), Randl (1997) presented
a design expression that explicitly includes the contribution of cohesion, friction and
dowel action.
The first term – cohesion – is related to the contribution of the interlocking between
aggregates; the second term – friction – is related to the contribution due to the
longitudinal relative slip between concrete parts and is influenced by the surface
roughness and the normal stress at the shear interface; and the third term – dowel
action – is related to the contribution of the flexural resistance of the shear
reinforcement crossing the interface.
The first two terms are clearly related to the Coulomb shear friction hypothesis while
the third term represents the contribution of the deformation of the shear reinforcement
due to the relative slip between concrete parts. The proposed design expression is given
as follows:
v u =coh  n   f c f y (II.85)
where νu is the ultimate longitudinal shear stress at the interface; τcoh is the concrete
cohesion due to aggregate interlock; µ is the coefficient of friction; σn is the normal
stress at the interface due to external loading and tension in the shear reinforcement; α
is a coefficient to take into account for the flexural resistance of reinforcement (dowel
action); ρ is the reinforcement ratio; fc is the concrete compressive strength; and fy is the
– II.29 –
yield strength of the reinforcement.
With the inclusion of partial safety factors, the design expression is as follows:
vu=
c f ck 1/ 3
coh  f
  k yk  n  
s  
f yk f ck
 s c
f
  ck
c
(II.86)
where νu is the ultimate longitudinal shear stress at the interface; c is the coefficient of
cohesion; fck is the characteristic value of concrete compressive strength; γcoh is the
partial safety factor for the cohesion; µ is the coefficient of friction; ρ is the
reinforcement ratio; k is a coefficient of efficiency for the tensile force that can be
transmitted to the shear reinforcement; fyk is the characteristic value of yield strength of
the reinforcement; γs is the partial safety factor for the shear reinforcement; σn is the
normal stress at the interface due to external loading; α is a coefficient for the flexural
resistance of reinforcement (dowel action); γc is the partial safety factor for concrete; β
is a coefficient allowing for angle of concrete diagonal strut; and ν is a reduction factor
for strength of concrete diagonal strut.
The values of these constants were calibrated by Randl (1997) and the values presented
in Table II.2 were proposed.
Table II.2 - Values for the constants of the design expression.

Surface Coefficient
Coefficient of friction
roughness of cohesion
Surface preparation k α β
R c µ
(mm) - (fck ≥ 20MPa) (fck ≥ 35MPa) - - -
High-pressure water-blasting ≥ 3.0 0.4 0.8 1.0 0.5 0.9 0.4
Sand-blasting ≥ 0.5 0.0 0.7 0.7 0.5 1.1 0.3
Smooth - 0.0 0.5 0.5 0.0 1.5 0.2
The surface roughness is quantitatively evaluated using the Sand Patch Test proposed by
the ASTM E 065 (2001). The adopted partial safety factors were of 1.15 and 1.50 for
steel and concrete, respectively. Randl (1997) proposed the value of 2.00 for the partial
safety factor of concrete cohesion since this is strongly influenced by the surface
preparation.
– II.30 –
II.3.2.30 Ali and White (1999)
Ali and White (1999) presented a methodology to predict the shear strength capacity of
concrete-to-concrete interfaces made of both normal and high strength concretes. The
four fundamental parameters were considered: a) surface roughness; b) concrete
strength; c) amount of shear reinforcement crossing the interface; and d) normal stress
at the interface.
The similarity between the proposed design expression and the one proposed by Mau
and Hsu (1988) is remarkable, as stated by Ali and White (1999). Moreover, while this
expression was developed on a pure analytical basis and calibrated with experimental
data, the design expression proposed by Mau and Hsu (1988) is based on semi-
empirical experiments.
The proposed design expression is as follows:
vu
fc
=1.47 a

 f y  n
fc
1.2 b (II.87)
reinforcement; σn is the normal stress at the interface due to external loads; and a and b
are two parameters to be experimentally calibrated.
II.3.2.31 Valluvan, Kreger and Jirsa (1999)
Valluvan et al. (1999) presented a revision of the shear-friction provisions of ACI 318
(1995). Based on the results of an experimental study, these researchers concluded that
the actual specifications of this design code are too conservative and proposed the
following design expression to estimate the ultimate longitudinal shear stress:
v u =   f y  n  (II.88)
where νu is the ultimate longitudinal shear stress at the interface; µ is the coefficient of
friction; ρ is the reinforcement ratio; fy is the yield strength of the reinforcement; and σn
– II.31 –
is the normal stress at the interface due to external loads. This design expression is valid
when the normal stress, due to external loads, is lower or equal to 5.52MPa (800psi).
The ultimate longitudinal shear stress at the interface is also limited by the minimum
value between 0.25fc and 5.52MPa (800psi).
When the normal stress, due to external loads, is higher than 5.52MPa (800psi), the
design expression takes the following form:
v u = n  (II.89)
In this case, the ultimate longitudinal shear stress is limited by the minimum value
between 0.6fc and 14.49MPa (2100psi).
II.3.2.32 Patnaik (2000)
Patnaik (2000), in the discussion of the paper by Valluvan et al. (1999), proposed the
following expression for the prediction of the ultimate longitudinal shear stress of
intentionally roughened surfaces:
v u =0.55   0.25 f y  f c (MPa) (II.90)
v u =0.55   36 f y  fc (psi) (II.91)
strength. According to Patnaik (2000), this expression also represents a lower bound of
the ultimate longitudinal shear stresses in monolithic concrete.
For surfaces not intentionally roughened, a slight modification to the expression was
proposed:
v u =0.5   0.25 f y  f c (MPa) (II.92)
v u =0.5   36 f y  fc (psi) (II.93)
The upper limit of the ultimate longitudinal shear stress should be limited to the
– II.32 –
minimum value given by: a) 0.2fc and 5.52MPa (800psi), for surfaces not intentionally
roughened; b) 0.25fc and 7.93MPa (1150psi), for intentionally roughened surfaces; and
c) 0.2fc and 8.96MPa (1300psi), for monolithic concrete.
II.3.2.33 Mattock (2001)
Mattock (2001) presented design expressions applicable to all strengths of concrete,

from normal to high-strength concrete. To estimate the ultimate longitudinal shear stress
across monolithic concrete and across the interface between concrete cast against
hardened concrete, with the substrate surface intentionally roughened, two expressions
were proposed by this researcher.
When the normal stress at the interface, due to internal and external loads, is greater or
equal to K1 /1.45, or when the ultimate longitudinal shear stress is greater or equal to
1.55K1, the ultimate longitudinal shear stress at the interface is given by:
v u =K 1 0.8   f y  n  (II.94)
where νu is the ultimate longitudinal shear stress at the interface; K1 is a coefficient that
depends of the concrete density; ρ is the reinforcement ratio; fy is the yield strength of
the reinforcement; and σn is the normal stress at the interface due to external loads. The
ultimate longitudinal shear stress is limited by the minimum value between K2 fc and K3.
When the normal stress at the interface, due to internal and external loads, is less than
K1 /1.45 or when the ultimate longitudinal shear stress is less than 1.55K1, the ultimate
longitudinal shear stress at the interface is given by:
v u =2.25   f y  n  (II.95)
For normal weight and monolithic concrete, the coefficient K1 is given by the minimum
between 0.1fc and 5.52MPa (800psi); K2 is equal to 0.3; and K3 is equal to 16.55MPa
(2400psi).
For normal weight concrete placed against hardened concrete with the substrate surface
intentionally roughened, the coefficient K1 is equal to 2.76MPa (400psi); K2 is equal to
– II.33 –
0.3; and K3 is equal to 16.55MPa (2400psi).
For sand-lightweight concrete, the coefficient K1 is equal to 1.72MPa (250psi); K2 is

equal to 0.2; and K3 is equal to 8.27MPa (1200psi). For all lightweight concrete, the
coefficient K1 is equal to 1.38MPa (200psi); K2 is equal to 0.2; and K3 is equal to
8.27MPa (1200psi).
For concrete placed against hardened concrete with the substrate surface not
intentionally roughened, the ultimate longitudinal shear stress at the interface can be
predicted by:
v u =0.6   f y (II.96)
and for concrete anchored to clean, unpainted, as-rolled steel by headed studs or by
reinforcing bars, the ultimate longitudinal shear stress at the interface is predicted by:
v u =0.7   f y (II.97)
with the ultimate longitudinal shear stress limited to the minimum value between 0.2fc
and 5.52MPa (800psi) for both cases. The coefficient λ is taken equal to 1.00 for normal
weight concrete; 0.85 for sand-lightweight concrete; and 0.75 for all lightweight
concrete.
II.3.2.34 Patnaik (2001)
Patnaik (2001) concluded that the shear-friction provisions for smooth surfaces,
proposed by the ACI 318 (1999), are too conservative. Based on experimental test
results, this researcher proposed the following expression to estimate the ultimate
longitudinal shear stress at smooth concrete interfaces:
v u =0.6 f y (MPa) (II.98)
v u =87 f y (psi) (II.99)
ratio; and fy is the yield strength of the reinforcement. The ultimate longitudinal shear
– II.34 –
stress is limited by the minimum value between 0.2fc and 5.52MPa (800psi). This
researcher also suggests that shear strength should not be considered for clamping
stresses lower than 0.35MPa (50psi).
II.3.2.35 Kahn and Mitchell (2002)
Kahn and Mitchell (2002) developed a research study to extend the current shear-
friction provisions of the ACI 318 (1999) to high strength concretes. The suggested
design expression, to be used with normal and high strength concretes, is as follows:
v u =0.05 f c 1.4  f y (II.100)
reinforcement. The adopted coefficient of friction is equal to 1.4, corresponding to a
monolithic concrete connection, and the maximum longitudinal shear stress is limited to
0.2fc.
II.3.2.36 Papanicolaou and Triantafillou (2002)
Papanicolaou and Triantafillou (2002) developed an experimental study to investigate

the shear transfer capacity of interfaces between pumice aggregate concrete and high-
performance concrete. Pumice is a natural and porous glass aggregate, very common in
volcanic regions such as Greece.
The push-off test was adopted and 126 specimens were tested. Two interface lengths
were considered: a) small, specimens with 170mm of length; and b) large, specimens
with 240mm of length. The following parameters were considered:
a) compressive/tensile strength of the pumice aggregate concrete; b) density of the
pumice aggregate concrete; c) ratio of shear reinforcement crossing the interface;
d) interface length; e) surface preparation; f) lateral confinement; and g) loading rate.
Rough surfaces were prepared with a special hammer, as designated by these
researchers, while smooth surfaces were prepared using an abrasive disk.
– II.35 –
The general form of the proposed design expression is given by:
b
v u =   f y  n  C (II.101)
where νu is the ultimate longitudinal shear stress at the interface; µ is the coefficient of
friction; ρ is the reinforcement ratio; fy is the yield strength of the reinforcement; σn is
the normal stress at the interface due to external loads; and C is a generalized cohesion
term. The latter depends on the tensile strength of the pumice aggregate concrete, and is
given by:
d
C=c f ctm (II.102)
where c is a coefficient to take into account the interface size effect; and fctm is the mean
tensile strength of the pumice aggregate concrete.
The values proposed for the coefficients b and d, in equations II.101 and II.102,
respectively, and for the coefficients of friction and interface size are presented in Table
II.3.
Table II.3 - Coefficients of friction and cohesion.

Size – Surface preparation Coefficient of friction Coefficient of cohesion
(b ≈ 1; d ≈ 0.5) µ c
Small – Smooth 0.33 3.63
Small – Rough 0.45 2.97
Large – Smooth 0.33 2.33
Large – Rough 0.45 1.90
Equations II.103 and II.104 were proposed for smooth and rough interfaces,
respectively, being given by:
v u =0.30   f y  n  1.7  f ct (smooth interfaces) (II.103)
v u =0.45   f y  n  1.4  f ct (rough interfaces) (II.104)
ratio; fy is the yield strength of the reinforcement; σn is the normal stress at the interface
– II.36 –
due to external loads; and fct is the tensile strength of the pumice aggregate concrete and
is given by:
f ct =0.2 f ck
2/ 3
[ 0.40.6 2000

] (II.105)
where ρ is the density of pumice aggregate concrete (kg/m3).
These expressions were defined for a coefficient c that quantifies the interface size
effect, assumed as three-quarters of the minimum value presented in Table II.3 for
smooth and rough surfaces.
II.3.2.37 Gohnert (2003)
Gohnert (2003) proposed a design expression to evaluate the ultimate longitudinal shear
stress at the interface between precast ribs and in situ concrete, based on the results of
ninety composite specimens, given by:
v u =0.2090 R z0.7719 (MPa; mm) (II.106)
v u =777 R z112 (psi; in.) (II.107)
where νu is the ultimate longitudinal shear stress at the interface and Rz is a texture
parameter calculated as the difference between the average height of the peaks and the
average height of the valleys given from an arbitrary baseline of the surface profile.
The tested specimens were composed by commercial precast ribs from five different
manufacturers and, therefore, presented different geometry, surface roughness and
concrete compressive strength. The cross section of the specimens, including the precast
rib and the in situ concrete, was rectangular. The width and thickness of the rib varied
between 100 and 250mm (3.94 and 9.84in.) and 60 and 70mm (2.36 and 2.76in.),
respectively. The width of the cross section was the same of the rib and the total depth
of the cross section was of 210mm (8.27in.). The adopted length for the specimens was
of 750mm (29.53in.). The substrate surface was prepared by the means of a stiff wire
brush or a rake. The parameter Rz was determined and presented values between 0.89
– II.37 –
and 4.22mm (0.035 and 0.1661in.). The concrete compressive strength varied between
22.8MPa (3306psi) and 56.2MPa (8149psi).
Gohnert (2003) concluded that a better correlation was obtained between the ultimate
longitudinal shear stress at the interface and the surface roughness, herein represented
by the texture parameter Rz. A poor correlation was observed when the surface
roughness was replaced by the concrete compressive strength. This researcher also
concluded that a actual measurement of the surface roughness should be specified
instead of simply describing the finishing procedure or the equipment used to prepare
the surface.
II.3.2.38 Mansur, Vinayagam and Tan (2008)
Mansur et al. (2008) investigated the shear transfer across a crack, both analytically and
experimentally. A comparison between several design expressions, including the ones
proposed by the ACI 318 (2005) and PCI Design Handbook (1992) and those suggested
by Mattock and his co-workers (1972, 1976, 2001), Walraven et al. (1987), Mau and
Hsu (1988), Lin and Chen (1989) and Loov and Patnaik (1994), was made.
It was concluded that the design expressions proposed by Walraven et al. (1987) and by
Mau and Hsu (1988) give unsafe predictions of the interface shear strength. Moreover,
the design expression proposed by Loov and Patnaik (1994), similar to the one proposed
by Mau and Hsu (1988) although with different values for the coefficients, presented
identical unsafe predictions.
A single curve formulation was proposed by Mansur et al. (2008), based on the design
expression by Mau and Hsu (1988) and calibrated with a set of 154 test results. The
compressive strength of the concrete adopted in the experimental study was between
18MPa (2611psi) and 100MPa (14504psi), while the normalized clamping forces (ρfy/fc)
were between 0.02 and 0.39. The proposed expression is given by:
vu
fc
=0.566

fy
fc
0.3 (II.108)
– II.38 –
reinforcement.
Comparing the proposed expression with the experimental data, Mansur et al. (2008)
concluded that this could be unsafe for low values of the normalized clamping forces.
Therefore, a trilinear formulation was proposed.
For normalized clamping forces lower or equal to 0.075, corresponding to the first
branch of the trilinear formulation, the normalized ultimate longitudinal shear stress at
the interface is given by:
vu
fc
=2.5
 
fy
fc
(II.109)
For normalized clamping forces between 0.075 and 0.270, corresponding to the middle
branch of the trilinear formulation, the normalized ultimate longitudinal shear stress at
the interface is given by:
v u 0.56
= 0.385 0.55
fc fc
fy
fc  (II.110)
For normalized clamping forces equal or higher than 0.270, corresponding to the third
and last branch of the trilinear formulation, the normalized ultimate longitudinal shear
stress at the interface is given by:
vu
=0.3 (II.111)
fc
II.3.3 Design Codes
II.3.3.1 CEB-FIP Model Code (1990)
According to the CEB-FIP Model Code 1990 (1990), the design shear stress at the
concrete-to-concrete interface νu is given by:
– II.39 –
 u=c f ctd    n f y   0.25 f cd (II.112)
where c and µ are factors that depend on the roughness of the interface; fctd is the design
tensile strength of the weakest concrete; σn is the external normal stress acting on the
interface; ρ is the reinforcement ratio (> 0.1%); fy is the yield strength of the
reinforcement; and fcd is the design compressive strength of concrete.
The coefficients of cohesion and friction proposed by this design code, for each type of
surface preparation, are presented in Table II.4.
For situations where the shear stress at the interface is considered low and, therefore, no
shear reinforcement is necessary, the CEB-FIP Model Code 1990 (1990) proposes that
the design shear stress at the concrete-to-concrete interface νu can be predicted only by:
 u=c f ctd (II.113)
where c is a factor that depend on the roughness of the interface, given in Table II.4; and
fctd is the design tensile strength of the weakest concrete.
The orientation of the shear reinforcement that crosses the interface is also considered.
For plain interfaces, the design shear stress at the interface νu is given by:
 u= [  n f y  sin cos   ]  0.3 f cd (II.114)
where µ is the coefficient of friction, assuming the value of 0.5 for smooth surfaces and
0.9 for rough surfaces; σn is the external normal stress acting on the interface; ρ is the
reinforcement ratio; fy is the yield strength of the reinforcement; α is the angle between
the shear reinforcement and the shear plane; and fcd is the design compressive strength
of concrete.
As mentioned before, the shear strength at a concrete-to-concrete interface can be

described by a combination of three different load carrying mechanisms: a) cohesion;
b) shear-friction; and c) shear reinforcement. CEB-FIP Model Code 1990 (1990) also
states that for a given slip value the shear resistance of a concrete joint can be calculated
as the sum of each resisting mechanism.
– II.40 –
Table II.4 - Coefficients of cohesion and friction according to CEB-FIP Model

Code 1990 (1990).
Type of surface Coefficient of Coefficient of
cohesion friction
Category Description c µ
I : a smooth surface, as obtained by casting a steel or
timber shutter
II : a surface which lies between trowelled or floated to
a degree, which is effectively as smooth as (I)
III : a surface which as been trowelled or tamped in
Type 1 such a way that small ridges, indentations or
undulations have been left 0.2 * 0.6
(smooth)
IV : a surface achieved by slip forming or vibro-beam
screeding
V : a surface achieved by extrusion
VI : a surface, which has been deliberated textured by
lightly brushing the concrete when wet
VII : as for (IV), but with more pronounced texturing,
as obtained by brushing, by a transverse screeder, by
combining with a steel rake or with an expanded metal
VIII : a surface which has been thoroughly compacted,
but no attempt has been made to smooth, tamp or
Type 2 texture the surface in any way, having a rough surface
with coarse aggregate protruding, but firmly fixed in 0.4 0.9
(rough)
the matrix
IX : where the concrete has been sprayed when wet, to
expose the coarse aggregate without disturbing it
X : a surface which has been provided with
mechanical shear keys
* For very smooth surfaces (type I and II) it is recommended to the use of c = 0.1.
Concrete-to-concrete friction is of particular interest because there are situations where

it is not necessary to provide shear reinforcement. For these cases, where the resisting
mechanism is a combination of cohesion and friction, CEB-FIP Model Code 1990
(1990) considers two types of surfaces: a) smooth; and b) rough. A smooth surface can
be obtained from concrete cast against metal or wooden formwork, usually named as
left as-cast surface, or by smoothing after cast with a trowel. A rough surface is obtained
by washing fresh concrete, scrabbling or scraping.
For smooth interfaces, the shear resistance due to concrete-to-concrete friction is given
by:
– II.41 –
 fu ,d =0.40  cd (II.115)
where σcd is the average normal compressive stress to the interface. The shear slip su
needed for the mobilization of τfu,d is given by:
s u=0.15   cd (II.116)
with su in mm and σcd in MPa.
For rough interfaces, the shear resistance due to concrete-to-concrete friction is given
by:
2 1
3
 fu ,d =0.40 f cd   cd  f yd  3 (II.117)
where fcd is the design value of the compressive strength of concrete; σcd is the average
normal compressive stress to the interface; ρ is the reinforcement ratio; and fyd is the
design value of the yield stress of the reinforcement crossing the interface. The shear
slip su needed for the mobilization of τfu,d is approximately equal to 2mm.
For a shear slip lower than su, the mobilized shear stress is given by:
 fd =5  fu ,d : if s10 mm (II.118)
4 3
   
 fd
 fu ,d
−0.5
 fd
 fu ,d
=0.3 s−0.03 : if s10 mm (II.119)
with s in mm.
The shear slip along a rough interface is accompanied by a crack opening, usually
designated as dilatancy, which is given by:
2
3 (II.120)
w=0.6 s
with both w and s in mm.
– II.42 –
II.3.3.2 Eurocode 2 (2004)
Eurocode 2 (2004) indicates that, in the absence of more detailed information, surfaces
may be classified as very smooth, smooth, rough or indented. The first three types of
surfaces are obtained by preparing the concrete surface in order to achieve the desired
roughness. The last one, indented surface, is a surface with indentations complying with
Figure II.10.
Figure II.10 - Indented surface according to Eurocode 2 (2004).

(picture extracted from Eurocode 2 (2004))
The very smooth surface, is considered as a surface cast against steel, plastic or specially
prepared wooden moulds. The smooth surface is a spliformed or extruded surface, or a
free surface left without further treatment after vibration. The rough surface is a surface
that has at least 3mm roughness at about 40mm spacing, achieved by raking, exposing of
aggregate or other methods giving an equivalent behaviour.
For each type of surface, two coefficients are defined and used in the prediction of the
shear at the interface between concretes cast as different times. These are the cohesion
and friction coefficients, being adopted the values presented in Table II.5. The
corrigendum to Eurocode 2 (2004), effective since 16 January 2008, presents significant
modifications to the coefficient of cohesion, in particular, for the very smooth surface.
The design shear stress at the concrete-to-concrete interface νu is given by:
 u=c f ctd  n f y  sin cos    0.5  f cd (II.121)
where c and µ are factors that depend on the roughness of the interface; fctd is the design
tensile strength of concrete; σn is the external normal stress acting on the interface; ρ is
– II.43 –
the reinforcement ratio; fy is the yield strength of the reinforcement; α is the angle
between the shear reinforcement and the shear plane; ν is a strength reduction factor;
and fcd is the design compressive strength of concrete.
Table II.5 - Coefficient of cohesion and friction according to Eurocode 2 (2004).

Coefficient of cohesion Coefficient of friction
c µ
Type of surface
Corrigendum Corrigendum
Eurocode 2 (2004) Eurocode 2 (2004)
(16 January 2008) (16 January 2008)
Very smooth 0.25 0.025 - 0.10 0.50 0.50
Smooth 0.35 0.20 0.60 0.60
Rough 0.45 0.40 0.70 0.70
Indented 0.50 0.50 0.90 0.90
II.3.3.3 BS 8110-1 (1997)
The British standard BS 8110-1 (1997) includes specifications for the design and
detailing of precast and composite construction. Several situations are considered for the
stresses at the interface and design provisions are presented for each case.
When shear forces are transmitted in plane and the design ultimate shear stress does not
exceed 0.23MPa, there is not any need to provide shear reinforcement crossing the
interface. Nevertheless, very smooth surfaces should be roughened.
Concrete-to-concrete interfaces under compression in all design conditions do not need

shear reinforcement crossing the interface. Nevertheless, the interface surface should
present a rough finishing and the design ultimate shear stress should not exceed
0.45MPa.
When the shear stresses due to the ultimate loads do not exceed 1.3MPa, the crack
opening at the interface should be prevented using steel ties or by the normal
compression stresses at the interface under all loading conditions.
When the shear reinforcement crossing the interface has to resist to the entire acting
shear force, the ultimate longitudinal shear stress at the interface vu is given by:
– II.44 –
 u=0.6 F b  tan  (II.122)
where Fb is the minimum value between 0.95 fy As and the anchorage value of the
reinforcement; fy is the yield strength of the reinforcement; As is the area of the shear
reinforcement crossing the interface; ρ is the reinforcement ratio; and α is the angle of
internal friction between the faces of the joint. The proposed expression is very similar
to one suggested by Birkeland and Birkeland (1966).
The values of α are already proposed for three different situations, Table II.6. This
standard also suggests that the angle of internal friction should vary between 0.7 and
1.7, being best determined with experimental tests.
Table II.6 - Angle of internal friction according to BS 8110-1 (1997).

Type of surface tan α
Smooth interface, as in untreated concrete 0.7
Roughened or castellated joint without continuous in-situ strips across the ends of joints 1.4
Roughened or castellated joint with continuous in-situ strips across the ends of joints 1.7
In Table II.7 is presented, for the ultimate limit state, the design longitudinal shear stress
at the interface, which is depends of: a) the preparation of the substrate surface; b) the
strength of the in-situ concrete; and c) the existence of shear reinforcement crossing the
interface.
Table II.7 - Design ultimate longitudinal shear stress according to BS 8110-1

(1997).
In-situ concrete strength
Precast unit Type of surface
25MPa 30MPa > 40MPa
As-cast or as-extruded 0.40 0.55 0.65
Without shear
Brushed, screeded or rough-tamped 0.60 0.65 0.75
reinforcement crossing the
interface Washed to remove laitance or treated
0.70 0.75 0.80
with retarder and cleaned
As-cast or as-extruded 1.20 1.80 2.00
With shear reinforcement Brushed, screeded or rough-tamped 1.80 2.00 2.20
crossing the interface
Washed to remove laitance or treated
2.10 2.20 2.50
with retarder and cleaned
– II.45 –
The surface designated as-cast includes cases where the concrete is placed and vibrated
leaving a rough finishing. The surface designated as-extruded includes cases where an
open-textured surface is produced directly from an extruding machine. The surface
designated brushed, screeded or rough-tamped includes roughened surfaces without
exposing aggregates.
II.3.3.4 ACI 318 (2008)
The Building Code Requirements for Structural Concrete ACI 318 (2008), from the
American Concrete Institute, includes design specifications for the shear-friction and
also particular specifications applicable to composite reinforced concrete elements.
The shear-friction provisions indicates that the ultimate longitudinal shear stress at the
interface vu is given by:
 u= f y  sin cos   (II.123)
where ρ is the reinforcement ratio; fy is the yield strength of the reinforcement; µ is the
coefficient of friction; and α is the angle between the shear reinforcement and the shear
plane.
The proposed design expression is identical to the one suggested by Birkeland and
Birkeland (1966) but includes the influence of the orientation of the shear reinforcement
crossing the interface.
The coefficient of friction µ, Table II.8, is defined for four situations: a) concrete placed
monolithically; b) concrete placed against hardened concrete with surface intentionally
roughened; c) concrete placed against hardened concrete not intentionally roughened;
and d) concrete anchored to as-rolled structural steel by headed studs or by reinforcing
bars.
The parameter λ is a modification factor related to concrete density and shall be taken
equal to: a) 1.00, for normal weight concrete; b) 0.85, for sand-lightweight concrete;
and c) 0.75, for all lightweight concrete.
– II.46 –
For normal weight concrete, placed monolithically or cast against hardened concrete
with surface intentionally roughened, the ultimate longitudinal shear stress vu is upper
limited by the minimum value between 0.2fc, (480 + 0.08 fc ) ((3.3 + 0.08 fc ) psi) and
11MPa (1600psi), where fc is the compressive strength of concrete. The yield strength of
the reinforcement fy shall not be taken greater than 414MPa (60000psi).
For all other cases, the ultimate longitudinal shear stress vu is upper limited by the
minimum value between 0.2fc and 5.52MPa (800psi), where fc is the compressive
strength of concrete. The yield strength of the reinforcement fy shall not be taken greater
than 414MPa (60000psi).
Table II.8 - Coefficient of friction proposed by ACI 318 (2008).

Type of surface
µ
Concrete placed against hardened concrete not intentionally
0.60λ
roughened
Concrete placed against hardened concrete with surface intentionally
1.00λ
roughened
Concrete placed monolithically 1.40λ
Concrete anchored to as-rolled structural steel by headed studs or by
0.70λ
reinforcing bars
When concrete is placed against hardened concrete, the interface shall be clean and free
of laitance. If the interface is intentionally roughened, it shall be to a full amplitude of
one-quarter of inch (6.35mm).
Additional provisions are specified for composite concrete members in bending, such as
precast concrete with cast-in-place parts.
When contact surfaces are clean, free of laitance and intentionally roughened, or not
intentionally roughened but provided with minimum ties, the longitudinal shear strength
of the interface Vu should not be taken greater than 552bd kN (80bd lb). When contact
surfaces are clean, free of laitance, intentionally roughened to a full amplitude of one-
quarter of inch (6.35mm) and ties are provided to transfer longitudinal shear forces, the
longitudinal shear strength of the interface Vu should not be taken greater than
2415bd kN (350bd lb). The terms b and d are the width of cross section at contact
– II.47 –
surface and the distance from extreme compression fibre to centroid of longitudinal
tension reinforcement, respectively. Both dimensions of the cross section should be used
in meters or inches, depending on the adopted unit system.
II.3.3.5 CAN/CSA A23.3 (2004)
The Canadian design code CAN/CSA A23.3 (2004) considers that cracking shall be
assumed to occur along a shear plane and that the relative displacement at the interface
is resisted by cohesion and friction.
The ultimate longitudinal shear stress at the interface νu is given by:
 u= c  c n  s  f y cos  (II.124)
where λ is a factor related to the concrete density; φc is the resistance factor for
concrete; c is the cohesion; µ is the coefficient of friction; σn is the external normal
stress acting on the interface due to external loading and tensioned shear reinforcement;
φs is the resistance factor for reinforcing bars; ρ is the reinforcement ratio; fy is the yield
strength of the reinforcement; and α is the angle between the shear reinforcement and
the shear plane.
The term λφc(c+µσn) shall not exceed 0.25φcfc, where fc is the specified compressive
strength of concrete.
For the parameters c and µ, four situations are considered: a) concrete placed against
hardened concrete with the surface clean but not intentionally roughened; b) concrete
placed against hardened concrete with the surface clean and intentionally roughened to a
full amplitude of at least 5mm; c) concrete placed monolithically; and d) concrete
anchored to as-rolled structural steel by headed studs or by reinforcing bars.
In Table II.9 are presented the proposed values for the cohesion and for the coefficient
of friction.
– II.48 –
Table II.9 - Cohesion and coefficient of friction proposed by CAN/CSA A23.3

(2004).
Cohesion Coefficient of friction
Type of interface c µ
(MPa) -
Not intentionally roughened 0.25 0.60
Intentionally roughened 0.50 1.00
Concrete placed monolithically 1.00 1.40
Concrete anchored to as-rolled structural steel 0.00 0.60
An alternative expression is proposed by CAN/CSA A23.3 (2004) for concrete placed

monolithically or placed against hardened concrete with the surface clean and
intentionally roughened to a full amplitude of at least 5mm. In this case, the ultimate
longitudinal shear stress at the interface is given by:
 u= c k   n f c  s  f y cos  (II.125)
where k is equal to 0.5, for concrete placed against hardened concrete, and 0.6, for
concrete placed monolithically. The term  c k   n f c shall not exceed 0.25φcfc.
II.3.3.6 AASHTO LRFD Bridge Design Specifications (2007)
The AASHTO LFRD Bridge Design Specifications (2007) considers that a crack across
a given plane can occur due to an existing or potential crack; an interface between
different materials; an interface between two concretes cast at different times; or at the
interface between different materials of the cross section.
Adopting a similar design approach as the one proposed by Mattock and Hawkins
(1972), the ultimate longitudinal shear stress at the interface νu is given by:
 u=c   f y  n  (II.126)
where c is the cohesion; µ is the coefficient of friction; ρ is the reinforcement ratio; fy is

the yield strength of the reinforcement; and σn is the external normal stress acting on the
interface.
– II.49 –
The ultimate longitudinal shear stress at the interface shall not exceed the minimum
value between K1 fc and K2, where fc is the specified compressive strength of concrete.
The cohesion, coefficient of friction and factors K1 and K2 depend on the surface
roughness and are presented in Table II.10.
Table II.10 - Cohesion, coefficient of friction and factors K1 and K2 according to

AASHTO LFRD Bridge Design Specification (2007).
Coefficient
Cohesion
of friction K1 K2
Type of interface c
µ
(MPa) - - (MPa)
12.5
Cast-in-place concrete slab on clean concrete
(normal weight concrete)
girder surfaces, free of laitance with surface 1.90 1.00 0.30
9.0
roughened to an amplitude of 6mm
(light weight concrete)
Normal density concrete placed monolithically 2.80 1.40 0.25 10.3
Lightweight concrete placed monolithically, or
non-monolithically, against a clean concrete
surface, free of laitance, with surface 1.70 1.00 0.25 6.9
intentionally roughened to an amplitude of
6mm
Normal weight concrete placed against a clean
concrete surface, free of laitance, with surface
1.70 1.00 0.25 10.3
intentionally roughened to an amplitude of
6mm
Concrete placed against a clean concrete
surface, free of laitance, but not intentionally 0.52 0.60 0.20 5.5
roughened
Concrete anchored to as-rolled structural steel
by headed studs or by reinforcing bars where
0.17 0.70 0.20 5.5
all steel in contact with concrete is clean and
free of paint
II.3.3.7 AASHTO Standard Specifications for Highway Bridges (1996)
The AASHTO Standard Specifications for Highway Bridges (1996) also adopts the
shear-friction theory to predict the longitudinal shear strength of concrete-to-concrete
interfaces. With a design philosophy very similar to the ACI 318 (2008), based in the
design expression suggested by Birkeland and Birkeland (1966), this design code
proposes that the ultimate longitudinal shear stress at the interface νu should be given
by:
– II.50 –
 u= f y  sin cos   (II.127)
where ρ is the reinforcement ratio; fy is the yield strength of the reinforcement; µ is the
coefficient of friction; and α is the angle between the shear reinforcement and the shear
plane.
The coefficient of friction µ is defined for four different situations: a) concrete placed
monolithically; b) concrete placed against hardened concrete with surface intentionally
roughened; c) concrete placed against hardened concrete not intentionally roughened;
and d) concrete anchored to as-rolled structural steel by headed studs or by reinforcing
bars. The suggested values for this coefficient are presented in Table II.11.
Table II.11 - Coefficient of friction proposed by AASHTO Standard Specifications

for Highway Bridges (1996).
Type of surface
µ
Concrete placed against hardened concrete not intentionally
0.60λ
roughened
Concrete placed against hardened concrete with surface intentionally
1.00λ
roughened
Concrete placed monolithically 1.40λ
Concrete anchored to as-rolled structural steel by headed studs or by
0.70λ
reinforcing bars
The parameter λ is a modification factor related to concrete density and shall be taken
equal to: a) 1.00, for normal weight concrete; b) 0.85, for sand-lightweight concrete;
and c) 0.75, for all lightweight concrete.
The ultimate longitudinal shear stress vu is upper limited by the minimum value between
0.2fc and 5.52MPa (800psi), where fc is the compressive strength of concrete.
When concrete is placed against hardened concrete, the interface shall be clean and free
of laitance. If the interface is intentionally roughened, it shall be to a full amplitude of
one-quarter of inch (6.35mm).
– II.51 –
II.3.3.8 PCI Design Handbook (2004)
The PCI Design Handbook (2004) proposes the use of the shear-friction theory to
predict the longitudinal shear strength of concrete-to-concrete interfaces, namely,
between precast concrete members and cast-in-place parts.
The proposed design expression to determine the ultimate longitudinal shear stress at
the interface νu, identical to the one proposed by Shaikh (1978), is given as follows:
 u= f y e (II.128)
where φ is a factor equal to 0.75; ρ is the reinforcement ratio; fy is the yield strength of
the reinforcement; and µe is the effective coefficient of friction.
The effective coefficient of friction is given by:
6.90  
e = (MPa) (II.129)
u
1000  
e = (psi) (II.130)
u
where λ is a factor related to the concrete density; µ is the coefficient of friction; and νu
is the ultimate longitudinal shear stress at the interface.
The parameter λ is a modification factor related to concrete density and the following
values can be adopted: a) 1.00, for normal weight concrete; b) 0.85, for sand-
lightweight concrete; and c) 0.75, for all lightweight concrete.
The yield strength of the reinforcement fy shall not be taken greater than 414MPa
(60000psi).
The coefficient of friction µ is defined for four different situations: a) concrete to

concrete, cast monolithically; b) concrete to hardened concrete, with roughened surface;
c) concrete to concrete; and d) concrete to steel. The recommend values for the
coefficient of friction and the maximum admissible values for the effective coefficient
of friction and for the ultimate longitudinal shear stress at the interface are given in
– II.52 –
Table II.12.
Table II.12 - Coefficient of friction proposed by PCI Design Handbook (2004).

Effective coefficient of
Coefficient of friction Maximum
Type of surface friction
µ νu
µe
Concrete to concrete, 0.30fc
1.4λ 3.4
cast monolithically 6.90MPa (1000psi)
Concrete to hardened
0.25fc
concrete, with 1.0λ 2.9
6.90MPa (1000psi)
roughened interface
0.20fc
Concrete to concrete 0.6λ 2.2
5.52MPa (800psi)
0.20fc
Concrete to steel 0.7λ 2.4
5.52MPa (800psi)
II.3.4 Comparative Analysis Between Quantification Proposals
II.3.4.1 Design Expressions
The performed bibliographic review revealed that several design expressions for the
assessment of the longitudinal shear strength of concrete-to-concrete interfaces have
been proposed since 1960. The scope of application of all these expressions is very
large, since several situations are assumed: monolithic concrete; composite concrete
with intentionally roughened surfaces (rough surfaces) and composite concrete with
surfaces not intentionally roughened (smooth surfaces). The consideration of the
material density, normal weight concrete and lightweight concrete, is also present in
some design expressions.
These expressions were developed and calibrated using results of experimental tests
and, therefore, they are representative of those results, giving an adequate prediction of
the longitudinal shear strength for those particular cases. Different researchers, such as
Anderson (1960); Hanson (1960); Saemann and Washa (1964); Hofbeck et al. (1969);
Mattock (1974); Walraven et al. (1987) and Patnaik (1992), adopted different tests and
used specimens with different geometries, Figure II.11. The push-off test is the most
common bond test but composite beams were also adopted. The mechanical properties
– II.53 –
of the adopted materials, concrete and steel reinforcement, are also very different
between studies. The concrete strength is taken into consideration by some design
expressions but not by all of them.
The main differences between design expressions are better identified in Table II.13,
where all expressions and their limits of application are presented. Some conclusions
are presented after a brief analysis of the contents of Table II.13.
From 1960 until the present date, the shear-friction theory proved to be effective in the
assessment of the shear stresses at concrete-to-concrete interfaces. Assumptions are
simple to understand and the design expression is easy to apply. Improvements, in order
to adequate it to different situations and to make it more accurate have easily been
introduced. Also the addition of other influencing parameters, such as external normal
stresses; dowel action; and concrete density, have been easily incorporated in existing
terms or as additional terms.
Several milestones can be identified from 1960 to the present date. The first milestone is
of course the publication of the shear-friction theory by Birkeland and Birkeland (1966).
The second milestone is the expression proposed by Mattock and Hawkins (1972), also
known as the modified shear-friction theory, since it explicitly includes the contribution
of cohesion.
The third milestone is the explicit inclusion of concrete strength in the design
expression, proposed by Loov (1978).
The fourth milestone is the expression by Walraven et al. (1987), since a totally
different approach to quantify the shear stresses at the interface is adopted.
The fifth and last milestone, is the design expression proposed by Randl (1997), by
explicitly considering the three mechanisms of shear transfer between concrete parts:
cohesion; friction; and dowel action.
– II.54 –
a) b) c)
d) e)
f) g)
Figure II.11 - Experimental tests: a) Anderson (1); b) Hanson (2); c) Saemann and Washa (3);
d) Hofbeck et al. (4); e) Mattock (5); f) Walraven et al. (6); and g) Patnaik (7).
( pictures extracted from: (1) Anderson (1960); (2) Hanson (1960); (3) Saemann and Washa (1964); (4) Hofbeck et al. (1969); (5) Mattock
(1974); (6) Walraven et al. (1987); and (7) Patnaik (1992))
– II.55 –
Table II.13 - Comparison between design expressions.

Design expression
Researcher(s) Year Limitations / Notes
(SI units)
For a concrete with 20.7MPa:
v u= 4.41 229
Anderson 1960 -
For a concrete with 51.7MPa:
v u=5.52276 
For rough bonded interfaces:
Hanson 1960 v u=3.45121  -
18.6
Mattock and v u= 121 
Kaar
1961
 
x
d
5
0.15 %
Saemann and 18.6 33− X

1964 v u= 207  2 -
Washa X 5 X 6 X 5
For smooth unbonded interfaces:
v u=0.300.78  n
Gaston and
1964 -
Kriz For smooth bonded interfaces:
v u=0.760.70  n
µ = 1.7 for monolithic concrete;
µ = 1.4 for artificially roughened
joints;
µ = 0.8-1.0 for ordinary construction
Birkeland and v u=  f y  joints.
1966
Birkeland
1.5 %
v u5.52 MPa
f c 27.58 MPa
For construction joints with an intermediate finish:
13.79
v u= 137.9 
Badoux and
1967
11  
a
d
-
Hulsbos For construction joints with a rough finish:
24.14
v u= 137.9 
11
a
d 
Birkeland 1968 v u=2.78   f y -
Adopted the design expression proposed by Birkeland and µ = 1.4 for rough interfaces;
Mast 1968 Birkeland (1966) µ = 0.7 for smooth interfaces;
v u=  f y  v u0.15 f c 
Hofbeck, Adopted the design expression proposed by Birkeland and
Ibrahim and 1969 Birkeland (1966)  fy 4.14 MPa
Mattock v u=  f y 
Mattock and Developed for the lower bound of the experimental tests  fy n 1.38 MPa
1972 v u=1.380.8   f y  n 
Hawkins v u{ 0.3 f c ; 10.34 MPa }
– II.56 –
Design expression
(SI units)
Developed for the mean values of the experimental tests
v u= 2.760.8   f y n 
v u=2.76 sin  f s  0.8sin −0.5sin 2 
2 2
 fy n 1.38 MPa
Mattock 1974
f s =0 : 051.3º v u{ 0.3 f c ; 10.34 MPa }
f s =−1.6 f y cos  38.7º  : 51.3º90º
f s = f y : 90º180º
Hermansen and v u= 4.00.8 f
1974 y -
Cowan
 fy1.38 MPa
For all lightweight concrete:
v u=1.380.8  f For all lightweight concrete:
y
Mattock, Li and v u { 0.2 f c ; 5.52 MPa }
1976
Wang For sanded lightweight concrete:
v u=1.720.8 f y For sanded lightweight concrete:
v u{ 0.2 f c ; 6.90 MPa }
Cs = 1.00 for normal weight concrete;
For monolithic specimens of normal and lightweight
Cs = 0.85 for sand-lightweight
concrete:
concrete;
v u=C s 3.11   f y Cs = 0.75 for all lightweight concrete.
Raths 1977
For smooth interfaces in normal and lightweight concrete: 2
Cs 
v u=C s 2.03   f e =6.90
y vu
Cs = 1.00 for normal weight concrete;
Cs = 0.85 for sand-lightweight
concrete;
Cs = 0.75 for all-lightweight concrete.
µ = 1.4 for monolithic concrete

Shaikh 1978 v u=  f y e µ = 1.0 for rough interfaces
µ = 0.4 for smooth interfaces
2
Cs 
e =6.90
vu
Loov 1978
vu
fc
=k

 f y  n
fc
2
For initially uncracked interfaces:
k = 0.50
f ci
v u=0.18 v cimax 1.64 f ci−0.82
vcimax
Vecchio and
1986 -
Collins
v cimax =
 fc
 0.31
24 w
a16 
C2
Walraven, v u=C 1   f y 
Frénay and 1987 -
0.406 0.303
Pruijssers C 1=0.822 f c C 2=0.159 f c
0.545
Mattock 1988 v u=0.467 f c 0.8   f y  n  v u0.3 f c
Design expression equal to the one proposed by Loov

(1978) For initially cracked and uncracked
Mau and Hsu 1988 interfaces: k = 0.66
v u= k    f y n  f c
– II.57 –
Design expression
(SI units)
v u= e   f y n 
 fy n 1.38 MPa
Lin and Chen 1989 0.5
e =
 1.75  f c
 f y  n   0.8 f c
0.25 v u{ 0.3 f c ; 12.46 MPa }
For smooth interfaces:

v u=0.40  n
Tsoukantas and
1989 -
Tassios For rough interfaces:
3
v u=0.5  f c  n
2
Patnaik 1992 v u=0.6   0.1  f y  fc v u0.25 f c

k = 0.5 for composite concrete
elements;
k = 0.6 for monolithic concrete.
Loov and λ = 1.00 for normal weight concrete;

1994 v u= k    0.1 f y  fc λ = 0.85 for sand-lightweight
Patnaik
concrete;
λ = 0.75 for all lightweight concrete.
v u0.25 f c
For initially cracked, monolithic, normal weight concrete:
v u=
  f y f c0.73
3.820
Mattock 1994 For composite reinforced concrete elements cast at different
v u0.3 f c
times, with a rough interface:
v u=
  f y f c0.73 −0.02 f
c
3.820
v u  f c
For water-blasted surfaces

(R ≥ 3.0mm):
c=0.4
=0.8 to 1.0
1 /3
Randl 1997 v u= c f c    k f y  n     f y f c For sand-blasted surfaces
(R ≥ 0.5mm):
c=0
=0.7
For smooth surfaces:
c=0
=0.5
Ali and White 1999

vu
fc
=1.47 a
fc
 f y  n
1.2 b -
Valluvan, n 5.52 MPa

Kreger and 1999 v u=   f y  n 
v u{ 0.25 f c ; 5.52 MPa }
Jirsa
– II.58 –
Design expression
(SI units)
v u{ 0.25 f c ; 7.93 MPa }
For intentionally roughened surfaces: For monolithic concrete (lower
v u=0.55   0.25 f y  fc bound):
Patnaik 2000 v u { 0.2 f c ; 8.96 MPa }
For surface not intentionally roughened:
v u=0.5   0.25 f v u { 0.2 f c ; 5.52 MPa }
y fc
For monolithic concrete and intentionally roughened
surfaces:
v u= K 1 0.8   f y  n  K1
 f y  n
1.45
K 1 =min { 0.1 f c ; 5.52 MPa } v u1.55 K 1
K 2 =0.3 v u{ K 2 f c ; K 3 }
K 3 =16.55MPa
For monolithic concrete and intentionally roughened
surfaces:
K1
v u=2.25   f y n   f y  n
1.45
K 1 =2.76 MPa v u1.55 K 1
K 2 =0.3 v u{ K 2 f c ; K 3 }
K 3 =16.55MPa
Mattock 2001 For monolithic sand-lightweight concrete and intentionally roughened surfaces:
K 1 =1.72 MPa
K 2 =0.2
K 3 =8.27 MPa
For all lightweight monolithic concrete and intentionally roughened surfaces:

K 1 =1.38 MPa
K 2 =0.2
K 3 =8.27 MPa
For concrete placed against hardened concrete not
intentionally roughened: v u { 0.2 f c ; 5.52 MPa }
v u=0.6   f y
λ = 1.00 for normal weight concrete;
For concrete anchored to clean, unpainted, as-rolled λ = 0.85 for sand-lightweight
structural steel by headed studs or by reinforcing bars: concrete;
v u=0.7   f λ = 0.75 for all lightweight concrete.
y
For smooth concrete interfaces:

Patnaik 2001 v u=0.6 f v u { 0.2 f c ; 5.52 MPa }
y
Kahn and v u=0.05 f c 1.4  f v u0.2 f c

2002 y
Mitchell
f ct=0.2 f ck
2/3
[ 0.40.6 2000

]
Papanicolaou For shear transfer capacity of
For smooth interfaces: interfaces between pumice aggregate
and 2002 v u=0.30   f y  n  1.7  f concrete and high-performance
ct
Triantafillou concrete
For rough interfaces:
v u=0.45   f y  n  1.4  f ct
Gohnert 2003 v u=0.2090 R z 0.7719 -
– II.59 –
Design expression
(SI units)
For normalized clamping forces lower or equal to 0.075:
vu
fc
=2.5
 
fy
fc
Mansur, For normalized clamping forces between 0.075 and 0.270: Normalized clamping forces:
fy
Vinayagam and 2008
Tan
v u 0.56
= 0.385 0.55
fc fc
fy
fc   fc
For normalized clamping forces equal or higher than 0.270:

vu
=0.3
fc
In Figure II.12, a comparison of the five design expressions, mentioned as the five
milestones of the shear-friction theory is presented. The design expression proposed by
Birkeland and Birkeland (1966) is plotted for both smooth and rough surfaces. The
design expression proposed by Randl (1997) is considered for three types of surfaces:
smooth; sand-blasted; and water-blasted. The remaining expressions plotted were
suggested by Mattock and Hawkins (1972), Loov (1978) and Walraven et al. (1987).
The adopted materials are: concrete with a compressive strength of 25MPa; and shear
reinforcement steel with a yield stress of 400MPa. All values are assumed as
characteristics values. No safety factors are adopted.
The design expressions proposed by Loov (1978) and Walraven et al. (1987) are very
similar and present, in general, the highest value for the shear strength of the interface.
The expressions proposed by Mattock and Hawkins (1972) and Birkeland and Birkeland
(1966) appear as a lower bound of the design expressions proposed by Loov (1978) and
Walraven et al. (1987), being similar to a bilinear approximation of the non-linear
expressions proposed by the latter researchers.
The influence of the surface preparation is already visible in the design expression
suggested by Birkeland and Birkeland (1966) but it is amplified by the design proposal
of Randl (1997). In fact, comparing the shear strength predicted using the design
expression proposed by Randl (1997) for surfaces with different preparations – smooth,
sand-blasted and water-blasted – the difference between curves is significant. These
differences reveal the significance of the surface preparation in the bond strength of the
– II.60 –
interface.
Significant differences are also observed in the predicted shear strength with the
increase of the steel reinforcement ratio.
10
8
Birkeland and Birkeland
7 (1966)
Smooth
Shear strength (MPa)
Birkeland and Birkeland

6 (1966)
Rough
5 Mattock and Hawkins
(1972)
4 Loov (1978)
Walraven et al. (1987)
Randl (1997)
3
Smooth
Randl (1997)
2 Sand-blasting
Randl (1997)
1 Water-blasting
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
As/Ac (%)
Figure II.12 - Comparison of design equations for shear-friction.
II.3.4.2 Design Codes
Some of the most important design codes of concrete structures present design
expressions based on the shear-friction theory. However, these expressions are quite
different since some parameters are not considered in some cases or since different
values for the same parameters are assumed.
In Table II.14, it is presented a comparison between the CEB-FIP Model Code 1990
(1990); the Eurocode 2 (2004); the BS 8110-1 (1997); the ACI 318 (2008); the
CAN/CSA A23.3 (2004); the AASHTO LFRD Bridge Design Specifications (2007); the
AASHTO Standard Specifications for Highway Bridges (1996); and the PCI Design
Handbook (2004).
– II.61 –
Table II.14 - Comparison between design codes.

AASHTO
AASHTO
CEB-FIP Standard PCI
BS ACI CAN/CSA LFRD Bridge
Model Eurocode 2 Specifications Design
Parameter 8110-1 318 A23.3 Design
Code 1990 (2004) for Highway Handbook
(1997) (2008) (2004) Specifications
(1990) Bridges (2004)
(2007)
(1996)
Compressive strength
no no no no no no no no
of concrete
Tensile strength of
yes yes no no no no no no
concrete
Reinforcement
yes yes yes yes yes yes yes yes
crossing the interface
Reinforcement
yes yes no yes yes no yes no
orientation
Coefficient of
yes yes no no yes yes no no
cohesion (or cohesion)
Coefficient of friction yes yes yes yes yes yes yes yes
Normal stress at the
yes yes no no yes yes no no
interface
Qualitative evaluation
yes yes yes yes yes yes yes yes
of roughness
Quantitative
evaluation of no no no no no no no no
roughness
The following parameters are considered: a) compressive and tensile strength of

concrete; b) reinforcement crossing the interface; c) cohesion; d) coefficient of friction;
e) normal stress at the interface due to external loading; and f) roughness quantification.
It must be highlighted that only the parameters that are explicitly present in the design
expression are considered.
Concrete Strength
Concrete strength is a common parameter to all design codes. The approach of CEB-FIP
Model Code 1990 (1990) and Eurocode 2 (2004) is similar. Both design codes makes
use of the concrete strength, adopting the tensile strength directly in the design
expression to evaluate the contribution of cohesion and the compressive strength to
define an upper limit for the shear strength.
ACI 318 (2008), BS 8110-1 (1997), CAN/CSA A23.3 (2004), AASHTO LFRD Bridge
Design Specifications (2007), AASHTO Standard Specifications for Highway Bridges
(1996) and PCI Design Handbook (2004) approaches are very similar and all these
design codes use the compressive strength of concrete indirectly in the design
– II.62 –
expression to define an upper bound for the longitudinal shear strength. This way, the
shear strength does not vary with the concrete strength but it is limited by it.
None of these design codes consider the influence of both substrate concrete and added
concrete layer. This way, the specified concrete strength in the design expressions
corresponds to the weakest concrete, usually the substrate concrete, and not to both
concrete strengths or even a combination of these.
Reinforcement Crossing the Interface
The consideration of the reinforcement crossing the interface, namely by considering its
area, orientation and yield strength, is common to all design codes.
Nevertheless, the BS 8110-1 (1997), the AASHTO LFRD Bridge Design Specifications
(2007) and the PCI Design Handbook (2004) considers that the shear reinforcement
crossing the interface is always normal to the shear plane, which is in fact the most
common situation. The remaining codes consider the angle between the shear
reinforcement and the shear plane as a design parameter.
Cohesion and Friction
The coefficient of friction is considered in all these design codes. On the contrary,
cohesion is only considered by four design codes: the CEB-FIP Model Code 1990
(1990); the Eurocode 2 (2004); the CAN/CSA A23.3 (2004); and the AASHTO LFRD
Bridge Design Specifications (2007).
The coefficient of friction is always linked to the surface finishing treatment of the
interface. Since each design code present its definition of smooth and rough surface, the
surface preparation methods do not have a direct correspondence between design codes.
Therefore, for the same surface finishing treatment, different values are proposed for the
coefficient of friction. Common to all these design codes is the definition of the left as-
cast surface or a similar type of surface, i.e., surface resulting from concrete cast against
steel or wooden formwork, not intentionally roughened and without further treatment.
BS 8110-1 (1997) has the simplest approach of all design codes. Only three surface
finishing treatments are considered, with the coefficient of friction varying between 0.7
– II.63 –
and 1.7.
The design expressions proposed by the Eurocode 2 (2004) and by the CEB-FIP Model
Code 1990 (1990) are very similar. Nevertheless, the classification of the surface
finishing treatment is quite different between both codes. Eurocode 2 (2004) considers
four surface conditions (very smooth, smooth, rough or indented), varying the
coefficient of friction between 0.5 and 0.9. The CEB-FIP Model Code 1990 (1990) only
considers two situations for the interface surface, smooth or rough, with a coefficient of
friction of 0.6 and 0.9, respectively.
The surface finishing terminology proposed by ACI 318 (2008), Table II.15, was also
adopted by the CAN/CSA A23.3 (2004), both AASHTO standards, LFRD Bridge
Design Specifications (2007) and Standard Specifications for Highway Bridges (1996),
and by the PCI Design Handbook (2004). The latter code presents four categories for
the surface finishing treatment but a more simplified definition for each one (Table
II.12), while the AASHTO LFRD Bridge Design Specifications (2007) considers two
additional categories for the case of cast-in-place slabs and lightweight concrete (Table
II.10).
The values for the coefficient of friction are common to these five design codes.
Nevertheless, a slight difference exist in the coefficient of friction for concrete interface
anchored to as-rolled structural steel by headed studs or by reinforcing bars, with the
CAN/CSA A23.3 (2004) presenting a slightly lower value.
The main difference between the CAN/CSA A23.3 (2004) and the ACI 318 (2008) is
the consideration of cohesion by the first code, while the latter code simply ignores this
parameter, Table II.15. AASHTO Standard Specifications for Highway Bridges (1996)
and PCI Design Handbook (2004) also neglects the contribution of cohesion.
CAN/CSA A23.3 (2004) and AASHTO LFRD Bridge Design Specifications (2007)
propose very different values for the cohesion, varying this parameter between 0 and
1.0MPa, for the first design code, and between 0.17 and 2.80MPa, for the latter design
code.
– II.64 –
Table II.15 - Comparison between CAN/CSA A23.3 (2004), ACI 318 (2008), PCI
Design Handbook (2004) and AASHTO Standard Specifications for Highway
Bridges (1996) and AASHTO LFRD Bridge Design Specification (2007).
AASHTO Standard AASHTO LFRD
CAN/CSA PCI Design
ACI 318 Specifications for Bridge Design
A23.3 Handbook
(2008) Highway Bridges Specification
(2004) (2004)
(1996) (2007)
Type of surface Coeff. of Coeff. of Coeff. of Coeff. of Coeff. of
Cohesion Cohesion Cohesion Cohesion Cohesion
friction friction friction friction friction
c µ c µ c µ c µ c µ
(MPa) - - - - - - - (MPa) -
Concrete placed
against hardened
concrete not 0.25 0.60 - 0.60 - 0.60 - 0.60 0.52 0.60
intentionally
roughened
Concrete placed
against hardened
concrete with
0.50 1.00 - 1.00 - 1.00 - 1.00 1.70 1.00
surface
intentionally
roughened
Concrete placed
1.00 1.40 - 1.40 - 1.40 - 1.40 2.80 1.40
monolithically
Concrete
anchored to as-
rolled structural
0.00 0.60 - 0.70 - 0.70 - 0.70 0.17 0.70
steel by headed
studs or by
reinforcing bars
A different approach is adopted by the CEB-FIP Model Code 1990 (1990) and the
Eurocode 2 (2004). Both codes consider that the contribution of cohesion is given by
multiplying a coefficient of cohesion, that depend on the roughness of the interface, by
the tensile strength of concrete. According to these design codes, the contribution of
cohesion is proportional to the concrete strength.
Normal Stress at the Interface
The normal stress at the interface, caused by external loads, is a design parameter in the
design expressions proposed by the CEB-FIP Model Code 1990 (1990); the Eurocode 2
(2004); the CAN/CSA A23.3 (2004); and the AASHTO LFRD Bridge Design
Specifications (2007). The BS 8110-1 (1987) ignore the contribution of this parameter
but has some considerations about the existence or not of compression stresses at the
interface. The ACI 318 (2008), the AASHTO Standard Specifications for Highway
Bridges (1996) and the PCI Design Handbook (2004) neglects the normal stress at the
– II.65 –
interface due to external loading.
Surface Roughness
Common to all these design codes is the consideration of the roughness of the interface
surface. It must be highlighted that in all design codes a qualitative approach is adopted,
easy and fast to perform but subjective since it depends on the technician opinion and
therefore subject to human error.
The preparation of a concrete surface, although being a simple process, is dependent of

several parameters such as: a) worker skills; b) adopted equipment; c) application time;
d) concrete physical and mechanical properties; among others. It is clear that, just
varying one of the mentioned parameters, very different textures can be obtained.
ACI 318 (2008), CAN/CSA A23.3 (2004), AASHTO LFRD Bridge Design
Specifications (2007), AASHTO Standard Specifications for Highway Bridges (1996)
and PCI Design Handbook (2004) provisions are very similar. Procedures or
equipments are not described, just a brief reference to the intention of having or not a
roughened surface is made and a minimum value for the roughness amplitude is
presented.
BS 8110-1 (1997) considers several finishing surfaces, from left as-cast surface to
brushed and washed surfaces. Preparation methods are not presented.
CEB-FIP Model Code 1990 (1990) and Eurocode 2 (2004) distinguish the surface
roughness as a function of the finishing process. CEB-FIP Model Code 1990 (1990)
considers a smooth surface as the one resulting from the cast against steel or wooden
moulds, from smoothening a surface after cast with a trowel or a surface without any
finishing. A rough surface can be obtained by washing fresh concrete, scrabbling or
scraping. Eurocode 2 (2004) considers the surface cast against steel, plastic or wooden
moulds as a very smooth surface; a slipformed, extruded or a surface without finishing,
as a smooth surface; and a surface resulting from raking or exposing of aggregates, as a
rough surface.
Comparing the CEB-FIP Model Code 1990 (1990) with the Eurocode 2 (2004), it is
– II.66 –
possible to conclude that the adopted terminology is very different regarding the surface
resulting from cast against steel moulds. The same surface (left as-cast) is classified by
the first code as smooth, presenting coefficients of cohesion and friction of 0.2 and 0.6,
respectively, while the second code will consider it as very smooth, presenting
coefficients of cohesion and friction of 0.25 and 0.5, respectively. Moreover, the
coefficients of cohesion proposed by the Eurocode 2 (2004) were recently modified
with the corrigendum of 2008 and changed from 0.25 to 0.025 - 0.10 for the very
smooth surface. This modification increases even more the difference between design
codes.
II.4 Conclusions
The behaviour of concrete-to-concrete interfaces, subjected to longitudinal shear

stresses, can be predicted using the shear-friction theory. This theory is accepted by
researchers worldwide. Moreover, two of the most important design codes of concrete
structures – Eurocode 2 (2004) and ACI 318 (2008) – present design expressions based
on the shear-friction theory.
The load transfer mechanism of shear forces between two concrete layers, according to
the shear-friction theory, is composed by: a) cohesion, due to mechanical interlocking
between particles; b) friction, due to the existence of compression stresses at the
interface and to the relative displacement between concrete parts; and c) dowel action,
due to the deformation of the reinforcement bars crossing the interface.
The elastic beam theory can be adopted to assess the shear strength of the concrete-to-
concrete interface for SLS. Nevertheless, this theory can not be used for ULS because of
the material constitutive law; existence of cracking; material time-dependent properties
(creep, shrinkage and relaxation); and existence of different materials. Therefore, more
accurate design expressions were developed and empirically calibrated using, mainly,
push-off tests.
The literature review identified several design expressions that were proposed since
1960 to the present date. The scope of application of all design expressions is very large,
– II.67 –
from monolithic concrete members to composite concrete members with intentionally

(rough surfaces) and not intentionally roughened surfaces (smooth surfaces). The
consideration of the material density, normal weight concrete and lightweight concrete,
is also present in some design expressions.
Five milestones of the shear-friction theory can be highlighted: 1) the publication of the
shear-friction theory by Birkeland and Birkeland (1966); 2) the modified shear-friction
theory, which explicitly includes the contribution of cohesion, proposed by Mattock and
Hawkins (1972); 3) the explicit inclusion of the concrete strength in the design
expression by Loov (1978); 4) the design expression developed using a sphere model,
representing the binding paste and the aggregates, by Walraven et al. (1987); and 5) the
first design expression that explicitly considers the three mechanisms of shear transfer
between concrete parts, proposed by Randl (1997).
The design expressions proposed by several design codes, such as the CEB-FIP Model
Code 1990 (1990); the Eurocode 2 (2004); the BS 8110-1 (1997); the ACI 318 (2008);
the CAN/CSA A23.3 (2004); the AASHTO LFRD Bridge Design Specifications (2007);
the AASHTO Standard Specifications for Highway Bridges (1996); and the PCI Design
Handbook (2004), are significantly different between themselves since some parameters
are not considered in some cases or different values for the same parameters are
assumed.
There are some common parameters to all design expressions, such as the coefficient of
friction and the reinforcement crossing the interface (percentage and yield strength).
Other parameters, such as the cohesion; the normal stress at the interface; the orientation
of the reinforcement crossing the interface; and the concrete strength, are only
considered by some design codes.
The roughness of the interface surface plays an important role since it is indirectly
considered in the design expressions of all design codes. The coefficients of cohesion
and friction are linked to the surface finishing treatment. A qualitative approach, based
on a visual inspection, is adopted to classify the surface roughness. This methodology is
easy and fast to perform but subjective since it depends on the technician opinion. It is
subject to human error and insufficient for an accurate prediction of both coefficients of
– II.68 –
cohesion and friction and, therefore, the shear strength of the interface.
The preparation of a concrete surface, although being a simple process, is significantly

influenced by: a) worker skills; b) adopted equipment; c) application time; d) concrete
physical and mechanical properties; among others. Varying one of the mentioned
parameters, very different textures can be obtained. No equipment is defined by design
codes to prepare the substrate surface. Moreover, no method is specified to control and
assess the effectiveness of the surface finishing treatment.
Even the comparison between design codes is difficult to perform. Besides different
design expressions, each design code has its own terminology and surface roughness
classification. Different coefficients of cohesion and friction are, therefore, obtained for
the same surface finishing treatment.
Improvements are necessary to increase the accuracy of the design expressions of

current design codes and to decrease the difference between them. The inclusion of a
quantitative methodology to characterize the surface roughness can help achieving this
goal.
– II.69 –
– II.70 –
Chapter III
Characterization of Concrete Surfaces
III.1 Introduction
The quality of the bond between concrete layers is usually linked to the roughness
degree of the interface between both parts. When preparing the substrate surface of a
reinforced concrete member, before casting a new concrete layer, the main question is to
know if it is rough enough or not. The meaning of rough differs from person to person
and, therefore, different people will make different classifications for a given surface.
The most common type of evaluation, purely qualitative and based in a visual
inspection, can be useful as a first approach but it is inadequate since it is subjective. To
overcome the discrepancy of opinions, in the definition of what is smooth or rough, a
quantitative characterization, using numerical parameters, is required.
The correct roughness preparation and characterization of a concrete surface must meet
three essential steps: a) an adequate preparation method, to remove the superficial layer
of concrete and/or increase the surface roughness; b) a method capable of obtaining a
2D profile or 3D sample of the surface; and c) a method to calculate numerical
parameters from these 2D profiles or 3D samples and to assess the effectiveness of the
preparation method.
– III.1 –
III.2 Methods of Surface Preparation
Several methods are available to remove the superficial layer of concrete, either for the
removal of damaged concrete or only to increase the surface roughness. Taking into
account the type of equipment used and the technical specifications of each method,
these can be classified, as proposed by ACI 555R (2001), in: a) mechanical removal;
b) particle impact removal; c) hydrodemolition; and d) chemical removal. The standard
ASTM D 4259 (1999) also defines the practice of surface preparation.
Special attention must be given to mechanical removal techniques, since heavy

machinery must not be used in surface preparation. Several researchers, such as
Silfwerbrand (1990), Talbot et al. (1994) and Bissonnette et al. (2006), concluded that
some removal techniques, such as hand-held hammers, can potentially create micro-
cracking in the concrete substrate and, therefore, can lead to lower values of the
interface bond strength.
If possible, these removal techniques should be avoided. Otherwise, the use of an

abrasive preparation method, such as sand-blasting, after the use of a mechanical
removal technique can help in the reduction of micro-cracking.
III.2.1 Mechanical Removal
Mechanical removal involves the use of manual or mechanized systems to remove the
superficial layer of concrete. Techniques like chipping, wire-brushing and scarification
are included in this category.
Chipping is the term used to designate the use of hand, electric or pneumatic hammers
with a chisel or a bushing tool. The equipment used should be lightweight, with high
manoeuvrability, and adequate to be used both in vertical and in horizontal areas. The
choice of the chisel or bushing tool is very important, being the pyramidal points and V-
shaped grooves the most appropriate, since they define the finishing surface due to the
type of face and consequently the imposed roughness.
Wire-brushing involves the removal of concrete using a wire brush attached to an
– III.2 –
electrical rotary hammer. This technique is easy to use and can be applied to almost all
surfaces, but it is more adequate for small surface areas and for early age concrete
because it is very work intensive and a light removal technique, respectively. The
surface roughness can be easily increased by increasing the application time, the
rotation speed and/or the pressure applied to the surface.
Scarification, also known as milling, is a technique that removes concrete by applying a

rotating cutting wheel to the surface, Figure III.1. The field of application of this
technique, due to the size of the equipment, is mainly horizontal concrete surfaces such
as bridge decks or pavements.
Figure III.1 - Scarification.

(picture extracted from Rampart Hydro Services website http://www.rampart-hydro.com)
Special attention should be given to the concrete cover of reinforcement bars. In areas
of reduced concrete cover, the reinforcement bars can be damaged with the concrete
removal. This can be avoided by selecting the appropriate cutting head, with the ideal
spacing and shape of the teeth.
The associated cost of these techniques is, in general, low, with the exception of
scarification because of the cost of the cutting wheels. Common disadvantages of all
these mechanical removal methods are the reduced speed and the noise and dust
– III.3 –
produced.
III.2.2 Particle Impact Removal
Particle impact removal is the term for the process of roughening a surface by forcing
solid particles across it at high speeds. These methods also have a wide spread use in
concrete removal. The most used techniques are shot-blasting and sand-blasting. When
metal particles are used this technique is called shot-blasting and when sand particles
are used it is called sand-blasting, Figure III.2.
The main advantages are their capability to be used both in horizontal and vertical
surfaces in a fast and effective way. They are also suitable to be used in places where
the use of water is prohibited and, therefore, hydrodemolition can not be adopted.
Figure III.2 - Sand-blasting.

(picture extracted from Tennessee Metallizing website htpp://www.tennesseemetallizing.com)
III.2.3 Hydrodemolition
The term hydrodemolition, also known as water-jetting or water-jet blasting, defines the
technique that uses high pressure water jets to remove the superficial layer of concrete,
Figure III.3.
This technique is highly recommended to apply in situations where the reinforcement

bars will be reused and where the use of water is admissible, since it does not introduce
– III.4 –
any damage to the concrete members. The main advantage of this technique is that
higher bond strengths of the interface are obtained in comparison with other methods, as
referred to by Silfwerbrand (1990), Talbot et al. (1994) and Bissonnette et al. (2006).
The finishing surface of concrete members is influenced by several parameters, such as:
a) nozzle type; b) distance to surface; c) water pressure; and d) operating time. All these
parameters should be considered and defined prior to the preparation of concrete
surfaces by this technique.
Figure III.3 - Hydrodemolition.

(picture extracted from Rampart Hydro Services website http://www.rampart-hydro.com)
III.2.4 Chemical Removal
Chemical removal is the term used for the application of an acid solution to a concrete
surface, allowing the acid to react with it and etch the concrete. The most common
designation of this type of technique is called acid etching, being this less popular than
other methods for the preparation of concrete surfaces, such as wire-brushing, sand-
blasting, shot-blasting or hydrodemolition.
The standard ASTM D 4260 (1999) includes the use of several acid solutions such as
Hydrochloric (Muriatic), Sulfamic, Phosphoric or Citric Acid.
The major disadvantage of this method is the fact that some of the products used are
– III.5 –
highly dangerous acid solutions. When improperly used, these solutions can lead to
severe injury. Protective clothing and equipment for eyes, face, hands and feet are
required. Disposal of all residual materials, including acid solutions, is also a major
problem and it must be made according to specific regulations for chemical residuals.
III.3 Measuring Methods of Surface Texture
There is not a widely accepted method for the texture characterization of concrete
surfaces. Current methods can be classified as contact methods or non-contact methods,
depending on if they require or not direct contact with the inspected surface,
respectively. These can also be classified as qualitative methods or quantitative
methods, depending on if the surface characterization is made by human assessment
only or if it is done using a numerical value, respectively. Finally, these can also be
classified as non-destructive methods, partially destructive methods or destructive
methods, respectively when they do not introduce any damage, when a small damage is
unavoidable or when a significant damage is provoked. The best method must not
require contact with the inspected surface, should be quantitative and non-destructive.
A qualitative approach, based on a visual inspection, is currently proposed by several

design codes such as CEB-FIP Model Code 1990 (1990), Eurocode 2 (2004), BS 8110-1
(1997), ACI 318 (2008) and CAN/CSA A23.3 (2004). These design codes already
present several typical finishing methods for the concrete surface which makes the
classification easier. However, this type of approach, focused on the surface macro-
texture, is always limited because the surface is simply classified as very smooth,
smooth, rough, very rough or indented and, for each case, different shear strengths can
be obtained.
An alternative to this qualitative approach is the adoption of a quantitative one, capable

of describing the surface texture with a numerical value, a texture parameter. This is
enough to describe the macro-texture of the surface and, if more advanced methods,
such as microscopy, are used, it can even characterize the surface micro-texture.
Moreover, this quantitative approach also presents the advantage of promoting the
standardization of measurement methods of the surface texture and the definition of the
– III.6 –
most adequate quantification method for a specific surface.
A quantitative method requires the definition of texture parameters, usually obtained

from 2D profiles or 3D samples of the surface. These texture parameters can be divided
into roughness and waviness parameters and are based on geometrical features of the
surface such as spacing, height and depth between peaks and valleys. Some quantitative
methods, although presenting a few relevant drawbacks, do not even require the
assessment of these profiles or surfaces, such as the Sand Patch Test proposed by
ASTM E 965 (2001).
Previous research studies, dealing with the characterization of the surface texture in
Civil Engineering works, were oriented to concrete and asphalt road pavements
surfaces. In this case, this subject is very important for safety concerns, to ensure that an
adequate friction between the pavement and the vehicle is reached and to facilitate
water drainage from the surface.
An extensive list of methods were tested and compared by Wambold and Henry (2002)
at the National Aeronautics for Space Administration Wallops facility, in a research
program sponsored by the Transportation Development Centre of Quebec. All these
methods were applied in the analysis of the surface texture of road pavements.
Nevertheless, some of these methods can be adopted to characterize the roughness of
concrete surfaces for repair, strengthening or precast purposes.
Next, a bibliographic review on quantification methods for the assessment of the surface
texture, which includes non-contact and contact methods, qualitative and quantitative
methods and non-destructive, partially destructive and destructive methods, is
presented. Some of these are described in more detail by Whitehouse (2002).
III.3.1 Concrete Surface Profiles
The most well known qualitative method for the quantification of the texture of a
concrete surface was proposed by the International Concrete Repair Institute (ICRI)
(1997). In this method, the concrete surface to be characterized is visually compared
with nine concrete surface profiles (CSP) of increasing roughness, Figure III.4. The
– III.7 –
advantages and disadvantages are clear: fast and easy to perform but results are
subjected to the opinion and sensitiveness of the technician.
a) b) c)
d) e) f)
g) h) i)
Figure III.4 - Concrete surface profiles: a) CSP 1; b) CSP 2; c) CSP 3; d) CSP 4; e) CSP 5; f) CSP 6;
g) CSP 7; h) CSP 8; and I) CSP 9.
(pictures extracted from Maerz and Myers (2001))
III.3.2 Sand Patch Test
The Sand Patch Test, or simply Sand Method, is proposed by ASTM E 965 (2001) and it
is the simplest quantitative method to assess the texture of a surface. This test involves
the spreading of a known volume of material, usually calibrated sand, on the surface and
the subsequent measurement of the total area covered, Figure III.5. More specifically,
the diameter of the circle formed is measured at quarter points and the resulting average
value is determined. The Mean Texture Depth (MTD) is a texture parameter that
– III.8 –
quantifies the surface texture and is given by:
4V
MTD= (III.1)
 D2
where V is the volume of sand dropped over the surface (mm3) and D is the average
value of the diameter of the area covered by the sand (mm).
Figure III.5 - Sand Patch Test.

(picture extracted from Hanson and Prowell (2004))
This test presents the advantage of being inexpensive and easy to perform both in situ
and in laboratory, but it is limited to horizontal floor surfaces. Furthermore, it provides
an average depth value of the surface macro-texture, being insensitive to micro-texture
characteristics.
III.3.3 Outflow Meter
The Outflow Meter is a volumetric method, originally develop by the US Federal

Highway Administration (FHWA) and proposed by ASTM E 2380 (2005), to assess the
surface texture of road pavements, Figure III.6, but it can also be used in other surfaces.
The Outflow Meter gives an average value for the surface texture but no other
additional information.
In this method, a cylinder is placed on the surface with a rubber ring between both. This
– III.9 –
ring is intended to simulate the contact of a vehicle tire with the road pavement. Then,
the cylinder is filled with water and the time needed for this to flow by the voids
between the surface and the rubber ring is measured.
The main concept of this method is that the necessary time for a know volume of water
to flow by the voids between a surface and a rubber ring can be correlated with the
surface texture. For smooth surfaces the water will flow slower and, therefore, the
measured time is higher in comparison with rough surfaces.
The major disadvantage of this method is that the permeability of the measured surface
also has a significant influence on the time needed to drain the water. Therefore, it is
essential to ensure that the water flow occurs between the surface and the rubber ring
and not by the substrate surface.
The results obtained with the Outflow Meter are related with the Mean Texture Depth
(MTD), as with the Sand Patch Test defined by ASTM E 965 (2001).
Figure III.6 - Outflow Meter.

(picture extracted from Wambold and Henry (2002))
III.3.4 Mechanical Stylus
The mechanical stylus, described by Whitehouse (2002), is a common mechanical
– III.10 –
device to assess the texture of a surface, Figure III.7. It is essentially composed by a

stylus, a conditioner/amplifier, a mechanical unit for advancement and a computer unit
for data acquisition.
The stylus probe, usually with a diamond tip, is dragged over the surface, along a
straight line, and the surface texture is measured. The precision of the measurement is
influenced by several parameters such as: a) size and hardness of the probe tip; b) signal
acquisition technology; c) movement speed; and d) surface irregularities.
This method has several disadvantages. First, it requires contact between the probe tip
and the measured surface. Hard materials are usually adopted to make the probe tip, in
order to avoid wear as possible, but they can also damage the surface, being
inappropriate for several applications.
Figure III.7 - Mechanical stylus.

(picture extracted from Taylor Hobson website http://www.taylor-hobson.com)
Another problem is the impossibility to perform measurements between peaks less

spaced than the probe tip size, Figure III.8. This difficulty is extensible to the lateral
resolution and it results in a loss of information on the surface texture. To avoid this
problem, the probe tip size must be carefully chosen.
Irregularities on the surface are identified using a Linear Variable Differential

Transformer (LVDT). As the probe tip moves over the surface, the LVDT produces an
– III.11 –
analog signal that is converted to a digital format through an analog-to-digital converter.

Other devices, which makes uses of a laser sensor, are also commonly used to detect the
stylus vertical movements. The digital signal is then stored in the computer unit for
future analysis.
The speed of the movement is also a critical parameter since, if it is too fast, the stylus
can not follow the entire contour of the surface and loss of information will occur.
Figure III.8 - Incapacity of the probe tip to measure the surface.
The mechanical stylus also measures the surface texture along a line. To obtain the full
surface texture, several lines must be measured. Very irregular surfaces, with curved or
steeped slopes are also very difficult to measure.
Some advantages can be pointed to this method. A scan line is relatively fast to perform,
the sample preparation is not necessary and, although not being a portable device, it can
be easily moved.
III.3.5 Circular Track Meter
A quantitative method with applications to road pavements is the Circular Track Meter
(CTMeter) proposed by ASTM E 2157 (2001), Figure III.9. This device is composed by
a charge coupled device (CCD) laser displacement sensor which is mounted on an arm.
This arm rotates so that the displacement sensor follows a circular track having a
diameter of 11.2in. (284mm). A portable computer controls the measurement operations.
While the Sand Patch Test, defined by ASTM E 965 (2001), gives an indication of the
– III.12 –
surface texture, with the MTD parameter, the Circular Track Meter gives the profile of
the surface. Additional calculations are therefore necessary to assess the Mean Profile
Depth (MPD) from the surface profile. The control software directly evaluates the Mean
Profile Depth (MPD) and the Root Mean Square (RMS) from the texture profile.
Figure III.9 - Circular Track Meter.

(picture extracted from Hanson and Prowell (2004))
Results of the Sand Patch Test, defined by ASTM E 965 (2001), and of the Circular
Track Meter can be converted in each other using an equation proposed by
ASTM E 2157 (2001).
MTD=0.947 MPD0.069 (mm) (III.2)
MTD=0.947 MPD0.0027 (in.) (III.3)
III.3.6 Digital Surface Roughness Meter
The Digital Surface Roughness Meter (DSRM) from MaGaNa Instruments, Figure
III.10, results from the development of the prototype proposed by Maerz and Myers
(2001).
– III.13 –
Figure III.10 - Digital Surface Roughness Meter.

(picture extracted from MaGaNa Instruments website http://www.magana-instruments.com)
This device was initially developed to measure the roughness of the concrete substrate,
to study the bond between concrete and epoxy bonded FRP laminates, used to
strengthen reinforced concrete members. It can be used as a quality control tool of the
surface roughness in order to identify when an adequate surface preparation has been
achieved.
The device consists of a portable lightweight aluminium box, which holds a camera and
a striping laser, and it can be operated against a horizontal or an upwards surface. The
device uses a 670nm/20mW striping laser mounted at 45 degrees to the concrete surface
to generate profile lines and a miniature video camera, mounted at 90 degrees to the
surface, to image these profiles. A laser band pass filter is mounted on the lens to admit
the laser-illuminated profile, and to reject other ambient light. The video signal is
transmitted to a laptop computer, with control software developed to run under
Microsoft Windows, and digitized by a PCMCIA card. Several roughness parameters
are automatically assessed with the surface profiles.
III.3.7 Microscopy
Improved methods for surface characterization are based in microscopy, Brundle et al.
(1992), Goodhew et al. (2001) and Whitehouse (2002), which are an expensive, but
– III.14 –
accurate and reliable, alternative to the previously presented non-contact profilometers,

Figure III.11. Their main advantage is the precision achieved, in the scale of
nanometers.
Figure III.11 - Scanning Electron Microscope.

(picture extracted from FEI Company website http://www.fei.com)
Optical and electron microscopy are techniques that can be used to inspect the surface
texture of a body using a light or electron beam, respectively, being the latter much
more powerful than the first. While optical microscopy uses visible light and a system
of lenses to magnify and focus the inspected body, electron microscopy uses an electron
beam and a set of electrostatic and electromagnetic lenses to imaging the body.
Two types of electron microscopy can be used. Scanning Electron Microscopy (SEM)
scans the body surface with the electron beam and measures how many electrons are
scattered back, giving information about its external morphology such as surface
texture. Transmission Electron Microscopy (TEM) shoots electrons through the body
and measures how the electron beam changes when it is scattered in the body, giving
information about its internal micro-structure such as micro-cracking.
Scanning Electron Microscope has some disadvantages, when compared with contact
profilometers, like the mechanical stylus, such as: a) high acquisition and maintenance
costs; b) surface sample must be inserted into a vacuum system; c) it requires stable and
high-voltage supplies; and d) it is a large, heavy and non-portable device.
Light microscopy does not require the inspected samples to be in vacuum and high-
– III.15 –
voltage power supplies but it is sensitive to light diffraction phenomena.
Improved techniques were developed to optimize the obtained texture data by

microscopy. Stereo imaging takes two images at angles of incidence differing just a few
degrees and by image processing it provides 3D models of the surface. Confocal
microscopy improves the resolution and contrast of images by eliminating the scattered
and reflected light from out-of-focus planes.
III.3.8 Ultrasounds
Ultrasounds is an alternative to contact methods. Although no specific application to

concrete surfaces was found in the conducted literature review, several researchers
adopted this technique to characterize the surface texture of other materials, such as
Flannery and von Kiedrowski (2002) and Mitri et al. (2009).
Sukmana and Ihara (2006) proposed this novel non-contact method for characterizing
the texture of surfaces by air-coupled ultrasound scattering. The researchers used nine
samples of sandpaper sheets with different grit sizes and a polished stainless steel plate
as reference.
III.3.9 Slit-Island Method
Issa et al. (2003) adopted a modified version of the Slit-Island Method to characterize
the fractured surface of concrete specimens. In Figure III.12 is present the overall view
of the experimental setup.
The original version of this method is based in covering the surface with a layer of a
different material, usually an epoxy resin, and then polishing it parallel to its plane.
With the removal of successive layers of the added material, several islands appear. For
each level the area and perimeter of each island are evaluated and plotted in a log-log
graph. A linear least squares regression is defined and the fractal dimension D is given
by:
– III.16 –
D=D '1 (III.4)
where D' is a power function obtained with the area and perimeter.
 2 / D ´
A= P (III.5)
Issa et al. (2003) used dyed water instead of an epoxy resin. From a top view, photos
were taken and the fractal dimension was evaluated by image processing. The
correlation between the fractal dimension, evaluated with the Slit-Island Method, and
the roughness of the surface was considered very good.
Figure III.12 - Slit-Island Method.

(picture extracted from Issa et al. (2003))
As stated by Bigerelle and Iost (2004), the assessment of the fractal dimension is
influenced by several parameters. The introduction of noise in the measurements of the
area and perimeter is an important factor that can lead to a false correlation between the
fractal dimension and the surface roughness. The evaluation method is another
– III.17 –
important factor. Alternatively to the Slit-Island Method, it is possible to use the Vertical
Section Method or the Box Counting Method leading to different values of the fractal
dimension and, therefore, to different correlations between this parameter and the
surface roughness.
III.3.10 Roughness Gradient Method
Abu-Tair et al. (2000) developed the Roughness Gradient Method to describe the
surface roughness. In this method, the surface of the concrete substrate is defined as a
series of irregular waves being the roughness parameter dependent of the peak-to-trough
depth and respective wavelength.
The roughness profile of the surface is evaluated using a profile texture-meter,

composed by 500 needles of 0.8mm in diameter and spaced 1.0mm apart and freely
dropped over the concrete surface, Figure III.13. The profile formed by the needles is
then photographed. By enlarging the photography, more accurate readings can be made
and the roughness profile is achieved.
Figure III.13 - Roughness Gradient Method.

(picture extracted from Abu-Tair et al. (2000))
The method considers that the profile is sequentially composed by a series of waves,
having each one its wavelength and maximum and minimum amplitude. For each wave,
it is possible to define the average double amplitude using the maximum and minimum
– III.18 –
amplitude of the respective wave. The averaging of all the double amplitudes gives the
roughness gradient.
III.3.11 Close-Range Digital Photogrammetry
Photogrammetry, in particular Close-Range Digital Photogrammetry, is a non-contact

method that can be used to characterize the surface texture. With this technique,
geometric properties of a body are determined using several photographic images taken
from different positions and angles. From common points, identified in each image, a
3D model of the body is generated using mathematical procedures.
This technique can be used to solve a wide range of measurement problems, with
applications in different fields, such as topography, architecture, engineering, among
others.
This technique has already been used to assess the roughness of rock surfaces, as
described by Lee and Ahn (2004), Unal and Unver (2004) and Unal et al. (2004). Seker
and Tavil (1996) also adopted photogrammetry to evaluate the surface roughness of
building facades.
III.3.12 Shadow Profilometry
A technique based on photography, called Shadow Profilometry, was proposed by

Maerz et al. (1990). A source of light is projected against a ruler revealing the roughness
profile across a joint surface, Figure III.14. For smooth surfaces the edge of the shadow
is a straight line, while for rough surfaces it will be irregular.
The projection of the light source on the surface can be recorded by photography and
then processed using commercial software for image analysis in order to isolate the
shadow edge. Once obtained the surface profile, texture parameters can be evaluated for
surface characterization.
– III.19 –
Figure III.14 - Shadow Profilometry.

(picture extracted from Maerz et al. (1990))
III.3.13 TDI Method
Santos et al. (2007) developed an in situ method to quantify the texture of concrete
surfaces and correlate this with the shear strength of concrete-to-concrete interfaces.
This was named TDI method because it is based on the processing of digital image,
Figure III.15.
c)
a) b) d)
e) f)
Figure III.15 - TDI method: a) preparation of the specimen; b) specimen after filled with additional
material; c) sliced specimen; d) digital image of the interface; e) acquisition of profile coordinates;
and f) texture profile.
– III.20 –
The TDI method is partially destructive since it involves the extraction of samples from
the concrete member. Extracted samples are subject to a preparation procedure
consisting on the following steps: a) cleaning of specimen surface with acetone to
remove oils, grease, etc.; b) application of an epoxy resin over the specimen surface;
c) sawing on a plane perpendicular to the surface and polishing of the specimen;
d) creation of a digital image, by photographic or scanning methods, using an auxiliary
metric scale; e) processing of the obtained digital image to identify the parts
corresponding to resin and concrete; f) determination of the coordinates of the
roughness profile; and g) transformation of coordinates to remove the effect of the
profile orientation.
To help in the identification of the surface profile, the magnification and special
selection tools of a common software package for image processing can be used.
III.3.14 Patented Devices
A bibliographic research at the United States Patent and Trademark Office (USPTO)
and at the National Institute of Industrial Property (INPI) revealed several techniques,
including apparatus and methods, that can be used for the characterization of concrete
surfaces. The scope of application of these techniques is the evaluation of the surface
profile of road pavements, such as runways, highways and raceways, etc.
One of the most significant devices was invented by Ozkul (2006). It uses air leakage to
predict the roughness of the concrete surface, Figure III.16. This device is composed by:
a) an air pump; b) a pressure sensor; c) an electronic control device; and d) a rigid skirt.
The device is placed on the concrete surface, involved by the rigid skirt which is in
contact with the substrate surface, and then the air is pumped inside. Since the contact
between the rigid skirt and the concrete surface is not completely perfect, due to surface
irregularities, the air leaks and the pressure inside the reservoir will change. The
variation in the pressure is compared with the same parameter measured using the
method on the nine concrete surface profiles from the International Concrete Repair
Institute (ICRI) (1997).
– III.21 –
Indirectly, the roughness of the surface is correlated with the variation of pressure inside
the reservoir and with the nine standard roughness profiles from ICRI (1997).
1) air pump
2) inside cavity
3) pressure sensor
4) rigid skirt
5) electronic control device
6) measured surface
Figure III.16 - Apparatus for measuring surface roughness.

(picture extracted from international patent WO2006111795A1)
III.4 Methods of Characterization
According to Smith (2003), surface texture is a complex condition resulting from a

combination of: a) roughness; b) waviness; and c) form. Roughness results from
irregularities that can occur due to the process of material removal, tool geometry, and
other factors. Waviness results from machine deflections, vibrations, etc., being a large
component of the surface texture upon which roughness is superimposed. Form is the
overall shape of the surface and can be caused by errors in machine tool.
Surface texture parameters can be grouped in the following four categories:

a) roughness or amplitude parameters; b) waviness or spacing parameters; c) hybrid
parameters; and d) statistical parameters. Roughness and waviness parameters are
related with, respectively, vertical and horizontal characteristics of the texture profile,
while hybrid parameters depend on both. Statistical parameters are surface texture
parameters treated as statistical data sets.
According to EN ISO 4287 (1998), to compute texture parameters it is necessary to

define the following three characteristic lengths, as defined in Figure III.17: a) sampling
length; b) evaluation length; and c) traverse length. The sampling length (SL) is the
– III.22 –
minimum length of the sample's surface in which all patterns of roughness and waviness
are present. The evaluation length (EL) is the length of the surface considered for
evaluation purposes which, generally, includes five sampling lengths. The traverse
length (TL) is the total length of the surface traversed by the measuring device, including
the evaluation length plus two sampling lengths (LO), one at the beginning and one at
the end of the evaluation length, for filtering purposes if needed. These additional
sampling lengths are necessary when a weighting filter is used to separate the waviness
profile from the roughness profile.
L0 SL SL SL SL SL L0
EL
TL
Figure III.17 - Sampling, evaluation and traverse length.
The analysis of a surface texture, Figure III.18, starts with the assessment of the surface
micro-topography. In two dimensions, this is represented by an irregular curve (2D
profile), while in three dimensions, this is given by an irregular surface (3D surface).
Sometimes, the irregular curve or surface is defined from a set of points, using
interpolation or other numerical methods, since data acquisition is performed in a
discrete rather than a continuous way.
Evaluation Evaluation
Form Definition
Assessment Evaluation of of other
removal of
of the of the Filtering roughness parameters
of the roughness
primary primary and (hybrid,
primary and waviness
profile parameters waviness statistical,
profile profiles
parameters etc.)
Figure III.18 - Analysis phases of a surface texture.
The measured profile/surface is called the primary profile/surface. It contains

information about the surface, as measured, i.e., without filtering and with the
limitations of the measuring device.
Next, the form of the surface profile is removed to ensure that the mean line of the
– III.23 –
profile, or the mean plane of the surface, is “horizontal”. The mean line/plane is a
straight line/plane generated by weighted average of each point resulting in equal areas/
volumes above and below the line/plane. In general, an auxiliary line from a linear
regression can be used to remove the form. This step is necessary because the surface is
not always perfectly parallel to the measuring device. Texture parameters can be
computed directly from this profile/plane and, therefore, these are usually simply
designated as primary parameters or roughness and waviness primary parameters.
The next step is filtering, which is a very important concept in surface metrology. The
filter is an electronic, mechanical, optical or mathematical transformation that can be
applied to the primary profile/plane to extract profiles/planes with different
wavelengths, namely roughness and waviness.
The final step is the computation of parameters, such as roughness and waviness
parameters from both roughness and waviness profiles/planes, respectively, to perform a
quantitative characterization of the surface texture.
III.4.1 Filters
In general, a profile is composed by a wide range of frequency components. The highest

frequency components, or the shortest waves, correspond to roughness, and the lowest
frequency components, or the longest waves, correspond to waviness.
Using a filter, Figure III.19, the waviness profile is extracted from the primary profile.
The roughness profile is then given, depending on the adopted filter, by the difference
between the primary profile and the waviness profile. Waviness and roughness
parameters are then assessed from these profiles.
Filtering is one of the most critical phases of the analysis of a surface texture. Basically,
it consists on smoothing and weighting, using an average function, the filter function,
the entire primary profile. The level of smoothing is defined by the cut-off length,
Figure III.20.
There are four standard cut-off lengths, defined by EN ISO 11562 (1997): a) 0.08mm;
– III.24 –
b) 0.25mm; c) 2.50mm; and d) 8.00mm.
Primary profile as measured
Primary profile after form removal
Waviness profile
Roughness profile
Figure III.19 - Primary, waviness and roughness profile.
The selection of the cut-off length can be made adopting a value corresponding to
approximately 2.5 times the spacing between two consecutive peaks or valleys. This
way, each sampling length should present at least two peaks and valleys. This is the
minimum required value for each sampling length.
Cut-off length
Figure III.20 - Cut-off length.
The evaluation length is, in general, equal to five sampling lengths and, therefore, the
traverse length is given by the evaluation length, added by two more sampling lengths at
both ends for stabilization purposes when filters are used. Therefore, the minimum
recommended value to be adopted for the traverse length should be of: 0.56mm,
1.75mm, 17.5mm and 56mm, for the cut-off length of 0.08mm, 0.25mm, 2.50mm and
8.00mm, respectively.
– III.25 –
Changes in the cut-off length will have a significant influence in the obtained waviness
and roughness profiles. A short cut-off length will result in an irregular waviness profile
and a smooth roughness profile. A long cut-off length will give a smooth waviness
profile and higher/irregular roughness profiles. This behaviour is independent of the
smoothness or roughness of the primary profile.
Defining the right cut-off length is a difficult task, since this will define waviness and
roughness. While the primary profile represents the geometry of a surface, as it is, given
the limitations of the measuring device, the waviness and roughness profiles are defined
by the technician. Therefore, the waviness and roughness profiles should always be
presented with the considered filter and cut-off length.
A large number of filters is available to extract the waviness and the roughness profile
from the primary profile, such as: a) the 2RC filter; b) the Gaussian filter; c) the Spline
filter; d) the Morphological filters; and e) the Wavelets filters.
The first three mentioned filters are the most common. Raja et al. (2002) makes a
comparison between these and some of the most advanced methods, the other two
filters. The latter were proposed by several researchers, including Chen et al. (1999),
Lingadurai and Shunmugam (2006) and Volk and Ville (2008), to overcome some
disadvantages of the former filters.
The analog 2RC was the first filter to be developed and it is fundamentally
characterized by its non-linear behaviour, resulting in a phase distortion effect, making
it impossible to obtain the roughness profile by a simple subtraction.
The Gaussian filter was developed to overcome the disadvantages of the 2RC filter and
is, today, the most used filter. It is defined by the EN ISO 11562 (1997). It has a linear
phase, allowing the roughness profile to be obtained directly by subtracting the
waviness profile to the primary profile.
The main disadvantage of this filter is the computational effort required, which has been
significantly reduced with the implementation of several numerical methods by
researchers such as Krystek (1996c), Hara et al. (1998) and Yuan et al. (2000).
– III.26 –
In some cases, specially when the surface has a strong curvature, the Gaussian filter
places the mean line outside the surface profile, which is unrealistic. Moreover, it
requires an additional length for stabilization purposes at the beginning and at the end of
the evaluation length. Without this additional length, the mean line will exhibit some
distortion due to the weighting function. In the presence of deep valleys, this filter also
exhibits some deficiencies and do not realistically represent the mean line of the surface
profile, as stated by Li et al. (2004).
The Rk filter is a modified Gaussian filter, defined by EN ISO 13565-1 (1997). It is used
in surfaces with very deep valleys but where the mean line should not be influenced by
these valleys, usually called plateaued surfaces, Figure III.21. It has the same
disadvantages of the Gaussian filter, excepting for the anomalies associated to deep
valleys.
Figure III.21 - Plateaued surface.
The Rk filter is a two-step Gaussian filter, being therefore also known as the double-
Gaussian filter. In the first step, the filter is applied and all points of the primary profile
below the mean line are replaced by the mean line itself, resulting in the removal of all
deep valleys. In the second step, the filter is again applied and the obtained mean line
will be considered as the real mean line. The roughness profile is then obtained by
subtracting the mean line to the primary profile.
More advanced methods, named Spline filters, were proposed by Krystek (1996a,
1996b], Goto et al. (2005) and Numada et al. (2007). Spline filters are also defined by
standards such as ISO/TS 16610-22 (2006). The use of high-order polynomials splines,
continuous functions also presenting continuous first and second derivatives, ensures a
smooth transition between multiple splines and a consequent smooth mean line. The
Spline filter is also a phase corrective filter, as the Gaussian filter. It does not causes
– III.27 –
edge distortion and it is most adequate for surfaces with strong curvatures. It usually
requires much less computational effort in comparison to the conventional Gaussian
filter.
III.4.2 Texture Parameters
Texture parameters are based on the geometry of 2D profiles or 3D surfaces, such as the
peaks height, valleys depth and spacing between peaks and valleys. These parameters
are defined by standards associations, such as the International Organization for
Standardization (ISO), the German Institute for Standardization (DIN) and the Japanese
Standards Association (JIS).
Some texture parameters have the same notation in different standards but present a
different terminology, usually related with the number of sampling lengths and with the
location of the considered peaks and valleys. For the sake of clearness, the same
notation is adopted and the standard designation is added in subscript, such as Rz(ISO),
Rz(DIN) or Rz(JIS).
Thomas (1982) and Mummery (1992) state that different standards, such as the
DIN 4768 (1990), DIN 4762 (1989), ISO 4287-1 (1984) and BS 1134-1 (1988), define
different parameters but use the same notation. Taking as an example the roughness
parameter Rz, the first mentioned standard, DIN 4768 (1990), uses the notation Rz and
the terminology Mean Peak-to-Valley Height, while the other mentioned three standards
use the same notation but a different terminology: Ten Points Height. Both parameters
are described ahead.
The main difference between these standards is the location of the peaks and valleys
used in the evaluation of the roughness parameter. The Mean Peak-to-Valley Height is
computed adopting the highest peak-to-valley height in each sampling length, while the
Ten Points Height is computed using the average of the five highest peaks and the five
deepest valleys from the entire evaluation length. Obviously, different results are
obtained from the same surface profile.
Next, it will be presented the definition of several roughness parameters, for the case of
– III.28 –
a 2D profile, where pi is the peak height and vi is the valley depth, relatively to the
profile mean line, and zi is the peak-to-valley height. Mummery (1992) presents a more
extensive and detailed list of texture parameters. These parameters can be divided into
two different categories: a) those that are based on average values; and b) those that are
based on extreme values.
The most used texture parameter is the Average Roughness (Ra), Figure III.22, also
known as the Center Line Average (CLA) or Arithmetic Average (AA). It is usually
adopted by its simplicity, being defined as the average deviation of the profile in
relation to its mean line. It is given by the following expression:
1 L
L ∫0
Ra = ∣z  x∣dx (III.6)
where L is the evaluation length and z(x) is the profile height at the coordinate x.
Ra
EL
Figure III.22 - Average Roughness.
In the case of discrete measurements, the Average Roughness (Ra) can be approximately
computed using the expression:
n
1
Ra ≈ ∑ ∣z i∣ (III.7)
n i=1
where n is the number of measurements and z(i) is the coordinate of each measurement.
The Root-Mean-Square Roughness (Rq) is a parameter more sensitive to peaks and

valleys and it is given by the expression:
– III.29 –
Rq =
 1 L
L
∫0
z  x 2 dx (III.8)
where L is the evaluation length and z(x) is the profile height at the coordinate x. For
discrete measurements, the Root-Mean-Square Roughness (Rq) is evaluated by:

n
1
Rq≈ ∑
n i=1
z i2 (III.9)
where n in the number of measurements and z(i) is the coordinate of each measurement.
The Average Roughness (Ra) does not provide any type of information on the local
variability of the surface profile. Therefore, quite different profiles can present the same
average roughness, Figure III.23. A texture parameter can only be used if it is sensitive
to changes in the surface texture, otherwise another one should be selected.
To overcome the limitations of the Average Roughness (Ra), other texture parameters
were defined, taking into consideration the location and spacing between peaks and
valleys, such as the Mean Peak Height (Rpm); the Mean Valley Depth (Rvm); the Mean
Peak-to-Valley Height (Rz(DIN)); and the Ten Points Height (Rz(ISO)). These parameters are,
in general, evaluated at five different sampling lengths, each one corresponding to a
fifth of the evaluation length, and then taken as the average values of these.
Figure III.23 - Different surfaces with same average roughness.
The Mean Peak Height (Rpm), Figure III.24, is defined as the average of the maximum
peak height from each sampling length and is given by:
– III.30 –
5
1
R pm= ∑p
5 i=1 i
(III.10)
where pi is the maximum peak height at each sampling length.
p2
p3 p4
p1
p5
v2 v5
v1 v4
v3
SL SL SL SL SL
Figure III.24 - Mean Peak Height and Mean Valley Depth.

The Mean Valley Depth (Rvm), Figure III.24, is defined as the average of the maximum
valley depth from each sampling length and is given by:
5
1
Rvm = ∑v
5 i=1 i
(III.11)
where vi is the maximum valley deep at each sampling length.
The Mean Peak-to-Valley Height (Rz(DIN)), Figure III.25, is defined as the average of the
maximum peak-to-valley height from each sampling length and is given by:
5
1
R z DIN = ∑z
5 i =1 i
(III.12)
where zi is the height between the maximum peak and the deepest valley in each
sampling length.
The Ten Points Height (Rz(ISO)), Figure III.26, is defined as the average of the sum of the
five highest peaks with the five lowest valleys within the entire evaluation length, and is
given by:
– III.31 –
∑ 
5 5
1
R z ISO = pi ∑ v i (III.13)
5 i =1 i=1
where pi is the peak height and vi is the valley deep.
z2 z5
z1 z4
z3
SL SL SL SL SL
Figure III.25 - Mean Peak-to-Valley Height.
p2
p3 p4
p1
p5
v2 v5
v1 v4
v3
EL
Figure III.26 - Ten Points Height.

The parameter Ten Points Height (Rz(ISO)) presents some limitations. Since the five
highest peaks and deepest valleys can be located anywhere along the entire evaluation
length, or even concentrated all together, this texture parameter is more adequate to be
used in very short evaluation lengths, where could not exist one peak and one valley in
each sampling length. Otherwise, the Mean Peak-to-Valley Height (Rz(DIN)) is a better
choice.
Several other parameters are based on extreme values of the profile, such as the
maximum peak height and the maximum valley depth. The Maximum Peak Height (Rp),
Figure III.27, is one of these and it is defined as the maximum peak height of the
– III.32 –
evaluation length, being given by:
R p=max { p i } (III.14)
where pi is the peak height.
The Maximum Valley Depth (Rv), Figure III.27, is another example and it is defined as
the maximum valley depth of the evaluation length, being given by:
Rv =max { v i } (III.15)
where vi is the valley deep.
Rp
Rv
SL SL SL SL SL
Figure III.27 - Maximum Peak Height and Maximum Valley Depth.

The Maximum Peak-to-Valley Height (Rmax), Figure III.25, is defined as the maximum
peak-to-valley height, in any of the sampling lengths within the evaluation length, and is
given by:
Rmax =max { z i } (III.16)
where zi is the height between the maximum peak and the deepest valley in each
sampling length.
The Total Roughness Height (Ry), Figure III.28, is defined as the height between the
maximum peak and the deepest valley within the evaluation length, and is given by:
R y = p maxv max (III.17)
where pmax is the highest peak and vmax is the deepest valley.
– III.33 –
The Total Roughness Height (Ry) represents the true total height of the profile, between
the highest peak and the deepest valley. Since the highest peak and the deepest valley
could be far away between them, as an example, one at the beginning and the other at
the end of the profile, these could be unrelated to one another and the texture parameter
do not represent a characteristic of the surface profile.
pmax
Ry
vmax
EL
Figure III.28 - Total Roughness Height.

For these reason, the parameters Mean Peak-to-Valley Height (Rz(DIN)) and Maximum
Peak-to-Valley Height (Rmax) are usually adopted. If two different surfaces present
similar values of both parameters, this indicates that they are identical, presenting a
consistent and uniform surface finishing. Otherwise, they have to present a local or
global significant discrepancy, since the average distance between the highest peak and
the deepest valley does not match with the extreme values.
III.5 Conclusions
The condition of the substrate surface of reinforced concrete composite members plays
an important role on the development of the bond strength at the concrete-to-concrete
interface. The unsound concrete must be removed and a certain degree of roughness is
required for the substrate surface in order to achieve an adequate bond between concrete
layers.
The substrate surface can be prepared using different methods. Chipping, wire-brushing,
scarification, sand-blasting, shot-blasting and hydrodemolition are among the most
common methods. Factors such as the available conditions on site; the existence of steel
– III.34 –
reinforcement near the surface; the possibility of using water; and the associated cost;
usually controls the selection of the preparation method.
Heavy concrete removal techniques should not be used. Abrasive and light removal
techniques, such as hydrodemolition, sand-blasting and shot-blasting, are most adequate
since micro-cracking can be avoided and, therefore, an higher bond strength at the
interface can be achieved. Chemical methods are prohibited for safety and
environmental reasons.
After the preparation of the substrate surface it is essential to control the quality of the
finishing surface. A visual inspection is usually adopted to check if the roughness
degree is adequate. This type of approach is easy and fast to perform but it is influenced
by the technician opinion and subjected to human error. It should only be adopted as a
first approach.
To overcome the disadvantages of a visual inspection it is possible to use several

methods that are based on a quantitative approach. The Sand Patch Test, defined by
ASTM E 965 (2001), is one of the most widely used standard tests. It is fast and easy to
use but limited to horizontal floor surfaces.
Other methods that are equally easy to use, non-destructive and do not require contact
with the measured surface, such as the Circular Track Meter (ASTM E 2157, 2001) and
the Digital Surface Roughness Meter (Maerz and Myers, 2001), can be used to obtain
2D profiles of the substrate surface. These two methods are laser based and are,
therefore, extremely accurate in the determination of the profile coordinates.
From these 2D profiles, or even from 3D surfaces, it is possible to compute texture

parameters that describes the geometrical characteristics of the measured surface. Filters
can also be used to extract profiles with different wavelengths from the measured
profile.
The Average Roughness (Ra) is one of the most used texture parameters but different
surfaces can present the same value for this parameter. To overcome this inconvenient,
parameters based on the average of the five highest peaks, five deepest valleys or five
highest peak-to-valley height, taken from the entire evaluation length or from the five
– III.35 –
sampling lengths, can be adopted. The Mean Peak-to-Valley Height (Rz(DIN)) and the Ten
Points Height (Rz(ISO)) are two of these parameters.
– III.36 –
Chapter IV
Laser Roughness Analyser
IV.1 Introduction
From the literature review previously presented on texture quantification methods, it is

possible to state that the characterization of a concrete surface should be made using a
quantitative evaluation; a non-destructive method; with a low/medium cost of
acquisition and maintenance; based on a portable device in order to be possible to be
used both in situ and laboratory; easy to perform without being work intensive; and
without requiring contact with the inspected surface.
The main characteristics of all texture quantification methods previously described are
summarized in Table IV.1. It is possible to conclude that the Circular Track Meter,
adopted by the ASTM E 2157 (2001), and the Digital Surface Roughness Meter,
proposed by Maerz and Myers (2001), are two methods satisfying the referred
conditions.
A first study by Júlio et al. (2004) proved that the bond strength between concrete layers
cast at different times is highly influenced by the surface texture of the concrete
substrate. In a subsequent study, Santos et al. (2007), a technique to assess the texture
profile of the substrate surface, named of TDI method and based on the processing of
digital image, was developed.
– IV.1 –
Table IV.1 - Comparison of texture quantification methods.

Contact
Quantitative Non Associated Work
Texture quantification method Portability with
evaluation destructive cost intensive
surface
Concrete Surface Profiles No Yes Low Yes No No
Sand Patch Test Yes Yes Low Yes No Yes
Outflow Meter Yes Yes Low Yes No Yes
Mechanical stylus Yes No Medium No Yes Yes
Circular Track Meter Yes Yes Medium Yes No No
Digital Surface Roughness Meter Yes Yes Medium Yes No No
Microscopy Yes No High No Yes No
Ultrasounds Yes Yes Medium Yes No Yes
Slit-Island Method Yes No Low No Yes Yes
Roughness Gradient Method Yes No Low No Yes Yes
Close-Range Digital Photogrammetry Yes Yes Medium Yes Yes No
Shadow Profilometry Yes Yes Low Yes Yes Yes
TDI Method Yes No Low No Yes Yes
Air Leakage No Yes Low Yes No Yes
The TDI method overcomes the disadvantages of other methods, in particular, the use of
the concrete surface profiles proposed by the ICRI (1997) and the Sand Patch Test
defined by ASTM E 965 (2001), allowing a quantitative assessment of the texture of
concrete surfaces but present some disadvantages: a) needs to be processed at the
laboratory; b) it is work intensive; and c) it is partially destructive.
Eleven roughness parameters were determined from profiles obtained with the TDI
method and correlated with the bond strength of the concrete-to-concrete interface both
in shear and in tension. High correlation coefficients were obtained, mainly with three
of the roughness parameters considered.
Due to the good results obtained with the TDI method and pointing to the development
and proposal of an innovative in situ methodology, for the assessment of the texture of
concrete surfaces and prediction of the shear strength between concrete layers, it was
decided to develop a new measuring device. This present the advantages of the existing
methods, described in Table IV.1, overcoming all the disadvantages.
The first part of the present chapter describes the development and specifications of the
newly measuring device, the 2D laser roughness analyser.
– IV.2 –
A comparison of results obtained with four texture quantification methods, including:

a) the newly developed 2D laser roughness analyser; b) a 3D laser scanner; c) the TDI
method, developed in a previous research study by Santos et al. (2007); and d) the Sand
Patch Test, a standard method proposed by the standard ASTM E 965 (2001), is
presented in the second part of the present chapter. The performance of each texture
quantification method is assessed and the advantages and disadvantages are discussed.
In the third part is discussed the influence of filtering. Texture profiles with different
wavelength can be extracted from the measured profile and used to define roughness
and waviness parameters. These parameters can improve the analysis of the surface
texture but require the specification of a filter and a cut-off length. The problems
associated to this selection are discussed.
The fourth part describes a study conducted to assess the influence of surface
irregularities in texture parameters. These parameters can be significantly affected by
strong irregularities, such as deep air holes and exposed aggregates. Therefore, special
attention must be given to the measuring path used to obtain the texture profiles.
IV.2 Development of a 2D Laser Roughness Analyser
The new measuring device, a portable laser roughness analyser, was entirely developed
during this research study, in collaboration with the Department of Electrical
Engineering and Computers of the University of Coimbra. It is fully customizable,
presenting all the advantages of the existing roughness quantification methods and
simultaneously overcoming all the disadvantages. This new technique was named as 2D
Laser Roughness Analyser Method (2D-LRA method), Santos and Júlio (2008).
The device uses two laser sensors with a range of 30-50mm and 30-130mm with a
resolution of 10µm and 60µm, respectively. A linear displacement table allows the laser
sensors to perform a maximum evaluation length of 220mm. In Figure IV.1a), these and
the remaining components can be observed in the interior of the device. The hardware is
protected by an aluminium box, Figure IV.1b). This equipment is connected to a laptop
that controls its use, Figure IV.1c).
– IV.3 –
a)
b)
c)
Figure IV.1 - Laser roughness analyser: a) internal view; b) aluminium cover box; and
c) equipment, connected to a laptop, being used.
The control software was developed in National Instruments LabView 7.1 and runs on a
NT based Microsoft Corporation operating system, Windows 2000 or XP, on a laptop
connected to the 2D laser roughness analyser using a USB interface. The software was
developed to: a) control the equipment; b) to assess the data; and c) to generate an
– IV.4 –
output text file, containing the coordinates of the texture profile. Control operations
include the following possibilities, Figure IV.2: a) definition of the evaluation length,
useful when readings are made on small concrete specimens; b) selection of the laser
sensor, allowing to increase or decrease the accuracy of the reading; c) positioning of
the laser sensor at the start position, necessary at the beginning of the reading
operations; and d) aborting of the reading process, in case of emergency. The texture
profile is recorded and graphically represented, in real time, during the readings of the
concrete surface. The graphic shows the distance between the laser sensor and the
concrete surface for each position along the evaluation path.
Figure IV.2 - Graphical user interface.
The obtained results are stored in a text file, with tab separated values, compatible with
any spreadsheet application, for future editing and processing.
IV.2.1 Software Application SurfTEX
The output data files generated by the developed software, can be imported after each
measurement to almost any commercial application of surface metrology. Since these
are text files, they can easily be treated to become compatible with the format of input
data files of other applications. Instead of using a commercial application for this
purpose, a full customizable software package, named SurfTEX, was developed.
SurfTEX was developed, using the Delphi RAD tool from Borland Corporation, Figure
– IV.5 –
IV.3. This application allows the user to: a) remove invalid profile coordinates; b) to
calculate several texture parameters; c) to determine the mean line of the texture profile;
and d) to draw the surface profile with and without form removal.
Twelve roughness parameters (Mummery, 1992) were implemented in SurfTEX:

a) Average Roughness (Ra); b) Mean Peak-to-Valley Height (Rz(DIN)); c) Maximum Peak-
to-Valley Height (Rmax); d) Mean Third Highest Peak-to-Valley Height (R3z);
e) Maximum Third Highest Peak-to-Valley Height (R3z,max); f) Ten Points Height (Rz(ISO));
g) Total Roughness Height (Ry); h) Root-Mean-Square (RMS) Profile Height (Rq);
i) Mean Peak Height (Rpm); j) Maximum Peak Height (Rp); k) Mean Valley Depth (Rvm);
and l) Maximum Valley Depth (Rv). These parameters were described in Section III.4.2.
Figure IV.3 - SurfTEX application.
Filters were not implemented in SurfTEX. Preliminary tests conducted with the 2D laser
roughness analyser suggested that roughness parameters, evaluated directly from the
primary profile, are sufficient to distinguish the texture of surfaces prepared using
different techniques.
For this reason, and to avoid the implementation of several routines in SurfTEX that
could potentially not be used in the future, it was adopted a commercial application to
study the influence of filters in the values of several texture parameters.
– IV.6 –
IV.2.2 Experimental Study
To test the 2D laser roughness analyser and compare this new method with the TDI
method, Santos et al. (2007), the same three different types of surface roughness were
adopted: a) smooth - surface left as-cast (LAC) against steel formwork, considered to
serve as reference, Figure IV.4a); b) rough without exposing the aggregates - surface
prepared with wire-brushing (WB), Figure IV.4b); and c) rough with aggregate
exposure - surface prepared with sand-blasting (SAB), Figure IV.4c).
a) b) c)
Figure IV.4 - Inspected surfaces: a) left as-cast; b) wire-brushed; and c) sand-blasted.
Ten records were produced for each surface type and the corresponding texture profiles
were generated. In Figures IV.5 to IV.7 is represented one record for each situation, left
as-cast surface; wire-brushed surface and sand-blasted surface, respectively.
Subsequently, based on these profiles, the surface texture was quantitatively evaluated
by determining the twelve roughness parameters implemented in SurfTEX.
– IV.7 –
0.080
Amplitude (mm) 0.060
0.040
0.020
0.000
-0.020
-0.040
-0.060
-0.080
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Position along the evaluation length (mm)
Figure IV.5 - Roughness profile of the surface left as-cast.
0.500
0.400
0.300
Amplitude (mm)
0.200
0.100
0.000
-0.100
-0.200
-0.300
-0.400
-0.500
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Figure IV.6 - Roughness profile of the wire-brushed surface.
0.800
0.600
Amplitude (mm)
0.400
0.200
0.000
-0.200
-0.400
-0.600
-0.800
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
Figure IV.7 - Roughness profile of the sand-blasted surface.
In Figure IV.8 the values assumed by the twelve roughness parameters, assessed with
the 2D laser roughness analyser, are plotted for each surface condition. These and the
corresponding values determined with the TDI method are presented in Table IV.2.
– IV.8 –
1.800
1.600
1.400
1.200
Amplitude (mm)
1.000
0.800
LAC (10E-6m)
0.600 WB (60E-6m)
0.400 SAB (60E-6m)
0.200
0.000 z
)
Ra
)
ax
ax
m
Ry
Rq
Rp
Rv
IN
O
R3
Rp
Rv
Rm
IS
D
(
(
R3
Rz
Rz
Roughness Parameter
Figure IV.8 - Roughness parameters assessed with the 2D laser roughness analyser.
Table IV.2 - Roughness parameters.

TDI method 2D-LRA method
Roughness Parameter
(mm) LAC LAC WB SAB
LAC WB SAB
(10µm) (60µm) (60µm) (60µm)
Average Roughness, Ra 0.032 0.095 0.202 0.034 0.019 0.129 0.227
Mean Peak-to-Valley Height, Rz(DIN) 0.214 0.413 0.818 0.121 0.106 0.391 1.015
Maximum Peak-to-Valley Height, Rmax 0.371 0.749 0.999 0.199 0.146 0.555 1.553
Mean Third Highest Peak-to-Valley Height, R3z 0.206 0.403 0.781 0.110 0.097 0.376 0.998
Maximum Third Highest Peak-to-Valley Height, R3z,max 0.370 0.707 0.994 0.185 0.136 0.541 1.537
Ten Points Height, Rz(ISO) 0.356 0.853 1.192 0.224 0.146 0.846 1.697
Total Roughness Height, Ry 0.371 0.887 1.235 0.240 0.156 0.864 1.710
Root-Mean-Square (RMS) Profile Height, Rq 0.049 0.119 0.254 0.041 0.024 0.160 0.309
Mean Peak Height, Rpm 0.084 0.162 0.411 0.065 0.054 0.209 0.623
Maximum Peak Height, Rp 0.133 0.282 0.529 0.143 0.079 0.413 1.130
Mean Valley Depth, Rvm 0.131 0.252 0.406 0.056 0.052 0.182 0.392
Maximum Valley Depth, Rv 0.239 0.605 0.706 0.097 0.078 0.451 0.580
It should be noted that the definition of roughness parameters like the Mean Third Highest
Peak-to-Valley Height (R3z) and Maximum Third Highest Peak-to-Valley Height (R3z,max)
implies the manual selection of the third highest peak-to-valley height. In Santos et al.
(2007), the authors performed this manual selection but have concluded that this
procedure is not completely objective and that further work is needed to automate this
operation. The values presented in Table IV.2 were obtained without manual selection.
The roughness profile of the smooth surface was assessed using both laser sensors, with
– IV.9 –
a resolution of 10µm and 60µm, to check if the latter is sufficiently accurate to analyse
smooth surfaces, such as those obtained with concrete cast against steel formwork.
Results are presented in Figure IV.9. A comparison between the coefficients of variation
of the roughness parameters obtained with each sensor is shown in Figure IV.10.
0.275
0.250
0.225
0.200
0.175
Amplitude (mm)
0.150
0.125
0.100 LAC (10E-6m)
0.075 LAC (60E-6m)
0.050
0.025
0.000
z
)
Ra
)
ax
ax
m
Ry
Rq
Rp
Rv
IN
O
R3
Rp
Rv
Rm
IS
D
(
(
R3
Rz
Rz
Roughness Parameter
Figure IV.9 - Roughness parameters obtained with both laser sensors for the left as-cast surface.
70%
60%
Coefficient of Variation (%)
50%
40%
30% LAC (10E-6m)

LAC (60E-6m)
20% WB (60E-6m)
10% SAB (60E-6m)
0%
)
Ra
)
ax
ax
m
Ry
Rq
Rp
Rv
m
IN
O
R3
Rp
Rv
Rm
IS
D
(
(
R3
Rz
Rz
Roughness Parameter
Figure IV.10 - Coefficient of variation of the adopted roughness parameters.
The slant shear test and the pull-off test were adopted to assess the bond strength, of the
interface between the old and the new concrete layers, in shear and in tension,
respectively.
In the slant shear test, a 200×200×400mm3 prismatic specimen, with the interface at 30
degrees to the vertical, is tested under compression. A standard testing machine
– IV.10 –
AMSLER 500 TNF and a load rate of 10kN/s were adopted to test the specimens. In the
pull-off test, a 200×200×200mm3 cubic specimen is used, with the interface at middle
height, and a core of 75mm diameter, drilled into the added concrete and extending
15mm beyond the interface into the concrete substrate, is tested in tension. The LOK
TEST equipment and a load rate of 0.05MPa/s were adopted to test the specimens.
Five slant shear specimens and five pull-off specimens were considered for each type of
surface roughness: smooth; rough without exposing the aggregates; and rough with
aggregate exposure. After casting the concrete substrate, the surface of each specimen
was prepared with the corresponding method: left as-cast; wire-brushing; and sand-
blasting, respectively. A period of 84 days was adopted between casting the substrate
and placing the added concrete. All specimens were stored in laboratory conditions,
without control of temperature and relative humidity. Tests were performed 28 days
after placing the added concrete.
The average compressive strength of the substrate concrete and of the added concrete,
assessed with cubic specimens at 28 days of age, was of 50.4MPa and 46.2MPa,
respectively. All slant shear specimens were tested, presenting an adhesive failure. For
the pull-off specimens, debonding occurred when drilling the core in all five specimens
with the substrate surface left as-cast. All the remaining pull-off specimens were tested,
also presenting an adhesive failure. Test results are shown in Figure IV.11 and Figure
IV.12.
16 3.0
Bond strength in tension (MPa)
14
Bond strength in shear (MPa)
2.5
12
2.0
10
8 1.5
6
1.0
4
0.5
2
0 0.0
LAC WB SAB LAC WB SAB
Surface Preparation Surface Preparation
Figure IV.11 - Bond strength in shear. Figure IV.12 - Bond strength in tension.
– IV.11 –
IV.2.3 Discussion of Results
In Table IV.2 are presented the results obtained with the proposed new method, using
the developed 2D laser roughness analyser, and those obtained with the TDI method
adopted in the previous study of Santos et al. (2007), based on digital image processing.
Comparing both methods, it can be observed that, for the rough without exposing the
aggregates surface type, the roughness parameters present, in general, similar values
when calculated using each of the methods, Figure IV.13. For the rough with aggregates
exposure surface type, most of the roughness parameters considered present higher
values when assessed with the 2D-LRA method than when determined by the TDI
method, Figure IV.14. For the smooth surface type, most of the roughness parameters
considered present smaller values when assessed with the 2D-LRA method than when
determined by the TDI method, Figure IV.15. This analysis confirms that with the
proposed new method, more precise roughness profiles are obtained.
1.0
0.8
Amplitude (mm)
0.6
0.4
TDI method
0.2 2D-LRA method
0.0
)
)
z
Ra
ax
ax
m
Ry
Rq
Rp
Rv
IN
O
R3
Rp
Rv
m
Rm
IS
D
(
(
R3
Rz
Rz
Roughness Parameter
Figure IV.13 - Roughness parameters assessed for the wire-brushed surface.
2.0
1.5
Amplitude (mm)
1.0
0.5 TDI method

2D-LRA method
0.0
)
)
z
Ra
ax
ax
m
Ry
Rq
Rp
Rv
IN
O
R3
Rp
Rv
m
Rm
IS
D
(
(
R3
Rz
Rz
Roughness Parameter
Figure IV.14 - Roughness parameters assessed for the sand-blasted surface.
– IV.12 –
0.5
Amplitude (mm) 0.4
0.3
0.2
TDI method
0.1
2D-LRA method
0.0
z
)
Ra
)
ax
ax
m
Ry
Rq
Rp
Rv
IN
O
R3
Rp
Rv
Rm
IS
D
(
(
R3
Rz
Rz
Roughness Parameter
Figure IV.15 - Roughness parameters assessed for the left as-cast surface.
The values of the coefficient of correlation are presented in Table IV.3. Linear
correlations are assumed between the bond strength of the interface, both in shear and in
tension, and each of the roughness parameters considered, determined with the proposed
new method and the TDI method. It can be observed that values are always higher than
0.85 and, for some roughness parameters, higher than 0.95.
Table IV.3 - Correlation between roughness parameters and bond strength.

Coefficient of Correlation (R2)
Roughness Parameter TDI method 2D-LRA method
Shear Tension Shear Tension
Average Roughness, Ra 0.9178 0.9203 0.9748 0.9762
Mean Peak-to-Valley Height, Rz(DIN) 0.8985 0.9013 0.8908 0.8936
Maximum Peak-to-Valley Height, Rmax 0.9898 0.9907 0.8788 0.8817
Mean Third Highest Peak-to-Valley Height, R3z 0.9049 0.9075 0.8891 0.8919
Maximum Third Highest Peak-to-Valley Height, R3z,max 0.9773 0.9786 0.8780 0.8810
Ten Points Height, Rz(ISO) 0.9886 0.9895 0.9504 0.9523
Total Roughness Height, Ry 0.9890 0.9900 0.9520 0.9539
Root-Mean-Square (RMS) Profile Height, Rq 0.9043 0.9070 0.9596 0.9613
Mean Peak Height, Rpm 0.8509 0.8541 0.8691 0.8721
Maximum Peak Height, Rp 0.9203 0.9228 0.8927 0.8955
Mean Valley Depth, Rvm 0.9463 0.9483 0.9240 0.9264
Maximum Valley Depth, Rv 0.9984 0.9980 0.9999 0.9998
The best result was obtained with the Maximum Valley Depth (Rv), presenting an almost
perfect linear correlation with the bond strength in shear and in tension, using either the
2D-LRA method or the TDI method, with values always over 0.99. In relation to the
– IV.13 –
Total Roughness Height (Ry), high values of coefficient of correlation are registered
using the TDI method, 0.9890 in shear and 0.9900 in tension, and using the 2D-LRA
method, 0.9520 in shear and 0.9539 in tension. The Maximum Peak-to-Valley Height
(Rmax), exhibits high values of coefficient of correlation using the TDI method, 0.9898 in
shear and 0.9907 in tension, but significantly lower values, using the 2D-LRA method,
0.8788 in shear and 0.8817 in tension. With the Average Roughness (Ra), the opposite is
observed, with high values of coefficient of correlation using the 2D-LRA method,
0.9748 in shear and 0.9762 in tension, and lower values, using the TDI method, 0.9178
in shear and 0.9203 in tension.
IV.2.4 Conclusions
This study corroborates the conclusions drawn in previous studies of Júlio et al. (2004)
and Santos et al. (2007), namely: a) the texture of the substrate surface influences
significantly the bond strength of concrete-to-concrete interfaces; b) it is possible to
obtain a texture profile of the substrate surface and, from this, to determine roughness
parameters; and c) it is possible to define linear correlations between some of these and
the bond strength of the interface, both in shear and in tension, with very high
coefficients of correlation.
Moreover, it was demonstrated that the proposed new method, using the developed 2D
laser roughness analyser, presents all the advantages, with even more accurate results,
and overcomes all the disadvantages of the TDI method, adopted in the previous study
of Santos et al. (2007). In fact, the proposed new method is really non-destructive, easy
to use and allows a fast in situ prediction of the bond strength of concrete-to-concrete
interfaces.
The use of the 10µm laser sensor for the left as-cast surface seems to be more adequate,
since all roughness parameters are slightly higher when compared with the same
parameters measured with the 60µm laser sensor. This indicates that for very smooth
surfaces, the resolution of the laser sensor has to be carefully selected.
The coefficient of variation of the roughness parameters presents significantly smaller
– IV.14 –
values for the wire-brushed surface, almost always lower than 20%, than for each of the
remaining two surface types: left as-cast and sand-blasted. This can be justified with the
uniformly rough surface, obtained with this treatment. For the left as-cast surface, this
coefficient varies between 20% and 40%. This higher value can be justified with the
existence of air holes in the concrete surface, which have a very significant influence in
the assessed roughness parameters when these are included in the path of the evaluation
length. For the sand-blasted surface, this coefficient presents the highest values. This
can be justified taking into account that this treatment implies the exposure of the
aggregates which produces an irregularly rough surface.
Further work is needed to create a large database, statistically representative, before

proposing a design expression that replaces the current qualitative approach of the
design codes. Nevertheless, results indicate that it is expectable this method to be
adequate to quantify the texture of the substrate surface and the Maximum Valley Depth
(Rv) roughness parameter to be adequate to incorporate in a design expression for the
longitudinal shear strength of the interface between concretes cast at different times.
In this first study on the influence of the surface texture on the bond strength of
concrete-to-concrete interfaces, filtering was not used. In fact, the texture profile was
considered equal to the primary profile and the roughness and waviness profiles were
neglected.
Next, a comparison between the Sand Patch Test, the TDI method, the 2D-LRA method
and a 3D method is presented. The latter method presents the same working principles
of the 2D-LRA method, with the obvious advantage of obtaining a 3D sample of the
concrete surface: a) it is laser based; b) do not require contact with the measured
surface; and c) it is non-destructive.
Two powerful applications for surface metrology, TrueMap and TrueSurf from
TrueGage, are used for data processing and analysis of the surface texture. These
applications allow the fast evaluation of several texture parameters, based on the
primary, roughness and waviness profiles.
– IV.15 –
IV.3 Comparison of Texture Quantification Methods
The Sand Patch Test (SPT), defined by ASTM E 965 (2001), is the simplest quantitative
method to assess the texture of a surface. It is fast and easy to perform and it is a non-
destructive test, but it presents the disadvantage of being limited to horizontal floor
surfaces. This method also provides an average depth value of the surface macro-
texture, but it is insensitive to micro-texture characteristics.
The TDI method, developed by Santos et al. (2007), overcomes the disadvantages of the
Sand Patch Test, as explained. However, the TDI method also presents disadvantages in
comparison to the Sand Patch Test: a) it is partially destructive; b) it is not completely
an in situ method, since it needs to be completed at the laboratory; and c) it is work
intensive and time consuming.
The 2D-LRA method, developed by Santos and Julio (2008), really overcomes all
disadvantages of both SPT and TDI method because it is a fully non-destructive and in
situ method, and it is easy and fast to perform.
Next, it is presented a comparison between four texture quantification methods: a) the

Sand Patch Test; b) the TDI method; c) the 2D-LRA method; and d) a 3D method,
named of 3D Laser Scanner Method (3D-LS method).
In comparison with the 2D-LRA method, the 3D-LS method is not portable; needs to be
conducted in laboratory; and presents a high cost of acquisition and maintenance. Its
major advantage is the possibility of obtaining a 3D sample of the entire substrate
surface instead of only a single 2D texture profile.
IV.3.1 Work Description
The selected laser scanner was the 3D high resolution laser digitizer Konica Minolta
VIVID 910, Figure IV.16, operated with TELE lens, leading to a maximum precision of
39µm. The 2D-LRA method was operated first with the laser sensor of 60µm and then
with the laser sensor of 10µm.
– IV.16 –
Figure IV.16 - Konica Minolta VIVID 910.
The selected testing samples were the same three concrete specimens, used to test the
2D-LRA method and showed in Figure IV.4, with a surface of 200×200mm2 and
prepared with different techniques, corresponding to the following situations: a) smooth,
surface left as-cast against steel formwork (reference situation); b) rough without
exposing the aggregates, surface prepared with wire-brushing; and c) rough with
aggregate exposure, surface prepared with sand-blasting.
For each situation several profiles were considered to evaluate the texture properties: ten
2D profiles, for the 2D-LRA method; four 2D profiles, for the TDI method, which is
particularly work intensive; and ten 2D profiles for the 3D-LS method, obtained by
slicing the generated 3D virtual model, Figure IV.17, in order to compare 2D results
with the remaining methods.
The data obtained from the 3D laser scanner was processed, using a commercial CAD
package and SYCODE plug-ins, to convert the original mesh of 3D face entities into a
mesh of points. Both 3D and 2D data were processed and numerical parameters were
computed using commercial software for visualization, processing and analysis of 2D
profiles and 3D surfaces, TrueSurf and TrueMap from TrueGage, respectively.
The form of the profile was removed using a “best-fit” line, computed by the Least
Squares Method, to correct misalignments between the straight surface and the
– IV.17 –
measuring device.
Figure IV.17 - 3D surface obtained from Konica Minolta VIVID 910.
Then, three types of numerical parameters were computed from all profiles. From each
primary profile, as measured and without filtering, were evaluated the primary
parameters. Filtering operations were made using a single filter, the Gaussian filter, with
a cut-off length of 8.0mm. After filtering, the primary profile was divided into roughness
and waviness profiles and the corresponding parameters were directly evaluated from
these.
The notation herein presented for the texture parameters (primary, roughness and
waviness) is the adopted by the software package TrueSurf and, therefore, slightly
different from the one previously used and adopted by Mummery (1992). A comparison
between both is presented in Table IV.4.
The following primary parameters were computed: a) Primary Average (Pa);

b) Maximum Profile Peak Height (Pp); c) Maximum Profile Valley Depth (Pv);
d) Maximum Height of the Profile (Pt); e) Mean Total Height (Ptm); f) Mean Peak
Height (Ppm); g) Mean Valley Depth (Pvm); h) Root Mean Square Primary (Pq); and
– IV.18 –
i) Maximum Primary Depth (Pmax).
The adopted roughness parameters were the following: a) Roughness Average (Ra);
b) Maximum Profile Peak Height (Rp); c) Maximum Profile Valley Depth (Rv);
d) Maximum Height of the Profile (Rt); e) Mean Total Height (Rtm); f) Mean Peak
Height (Rpm); g) Mean Valley Depth (Rvm); h) Root Mean Square Roughness (Rq);
i) Maximum Roughness Depth (Rmax); j) Valley Depth (Rz); and k) Ten-Point Height of
Irregularities (RzJIS).
The selected waviness parameters were the following: a) Waviness Average (Wa);
b) Maximum Profile Peak Height (Wp); c) Maximum Profile Valley Depth (Wv);
d) Waviness Height (Wt); e) Mean Total Height (Wtm); f) Mean Peak Height (Wpm);
g) Mean Valley Depth (Wvm); h) Root Mean Square Waviness (Wq); and i) Maximum
Waviness Depth (Wmax).
Table IV.4 - Comparison of the notation of the texture parameters.

Roughness Parameter * Primary Parameter
(Mummery, 1992) (from TrueSurf and TrueMap)
Average Roughness, Ra Primary Average, Pa
Mean Peak-to-Valley Height, Rz(DIN) Not implemented as primary parameter
Maximum Peak-to-Valley Height, Rmax Maximum Primary Depth, Pmax
Mean Third Highest Peak-to-Valley Height, R3z Not implemented as primary parameter
Maximum Third Highest Peak-to-Valley Height, R3z,max Not implemented as primary parameter
Ten Points Height, Rz(ISO) Not implemented as primary parameter
Total Roughness Height, Ry Maximum Height of the Profile, Pt
Root-Mean-Square (RMS) Profile Height, Rq Root Mean Square Primary, Pq
Mean Peak Height, Rpm Mean Peak Height, Ppm
Maximum Peak Height, Rp Maximum Profile Peak Height, Pp
Mean Valley Depth, Rvm Mean Valley Depth, Pvm
Maximum Valley Depth, Rv Maximum Profile Valley Depth, Pv
* Computed in Section IV.2 without filtering
In Tables IV.5 to IV.7 are presented the mean values of the primary, roughness and
waviness parameters, respectively, obtained from the measured profiles.
– IV.19 –
Table IV.5 - Primary parameters.

Left as-cast Wire-brushing Sand-blasting
Primary
Parameter TDI
2D-LRA 2D-LRA
3D-LS TDI
2D-LRA
3D-LS TDI
2D-LRA
3D-LS
method method method method
(mm) method method method method method method
(10µm) (60µm) (60µm) (60µm)
Pa 0.031 0.035 0.035 0.074 0.088 0.114 0.169 0.209 0.209 0.298
Pp 0.129 0.336 0.147 0.148 0.197 0.426 0.467 0.491 0.819 0.638
Pv 0.240 0.113 0.156 0.783 0.558 0.368 1.119 0.665 0.579 1.622
Pt 0.369 0.449 0.304 0.930 0.755 0.794 1.586 1.156 1.398 2.260
Ptm 0.190 0.064 0.078 0.131 0.408 0.193 0.255 0.647 0.493 0.492
Ppm 0.076 0.034 0.035 0.061 0.150 0.110 0.113 0.287 0.261 0.243
Pvm 0.113 0.030 0.043 0.070 0.258 0.083 0.142 0.361 0.233 0.249
Pq 0.049 0.049 0.045 0.135 0.110 0.144 0.253 0.257 0.267 0.433
Pmax 0.301 0.327 0.220 0.833 0.713 0.413 1.167 0.837 1.066 1.714
Table IV.6 - Roughness parameters.

Roughness
Parameter TDI
2D-LRA 2D-LRA
3D-LS TDI
2D-LRA
3D-LS TDI
2D-LRA
3D-LS
(mm) method method method method
method method method method method method
(10µm) (60µm) (60µm) (60µm)
Ra 0.024 0.008 0.011 0.032 0.054 0.028 0.058 0.100 0.061 0.105
Rp 0.130 0.239 0.098 0.230 0.174 0.186 0.285 0.315 0.322 0.417
Rv 0.201 0.132 0.111 0.437 0.448 0.138 0.531 0.419 0.244 0.590
Rt 0.331 0.371 0.209 0.668 0.623 0.324 0.817 0.735 0.566 1.007
Rtm 0.189 0.058 0.073 0.104 0.338 0.146 0.183 0.507 0.269 0.309
Rpm 0.084 0.032 0.037 0.049 0.123 0.074 0.084 0.218 0.141 0.155
Rvm 0.105 0.026 0.036 0.054 0.215 0.072 0.100 0.289 0.128 0.154
Rq 0.042 0.019 0.017 0.069 0.076 0.038 0.100 0.126 0.080 0.151
Rmax 0.306 0.303 0.199 0.662 0.565 0.311 0.767 0.735 0.525 0.914
Rz 0.189 0.058 0.073 0.104 0.338 0.146 0.183 0.507 0.269 0.309
RzJIS 0.238 0.160 0.158 0.252 0.421 0.266 0.397 0.645 0.485 0.604
– IV.20 –
Table IV.7 - Waviness parameters.

Waviness
Parameters TDI
2D-LRA 2D-LRA
3D-LS TDI
2D-LRA
3D-LS TDI
2D-LRA
3D-LS
method method method method
(mm) method method method method method method
(10µm) (60µm) (60µm) (60µm)
Wa 0.018 0.033 0.033 0.065 0.056 0.109 0.153 0.147 0.182 0.251
Wp 0.037 0.127 0.107 0.110 0.136 0.266 0.419 0.244 0.600 0.467
Wv 0.043 0.047 0.064 0.392 0.118 0.320 0.617 0.489 0.463 1.176
Wt 0.080 0.174 0.171 0.501 0.254 0.586 1.036 0.733 1.063 1.643
Wtm 0.038 0.017 0.018 0.059 0.139 0.085 0.133 0.262 0.298 0.278
Wpm 0.020 0.006 0.004 0.029 0.068 0.054 0.064 0.108 0.155 0.138
Wvm 0.019 0.011 0.014 0.030 0.072 0.031 0.069 0.154 0.143 0.140
Wq 0.022 0.042 0.041 0.100 0.064 0.136 0.205 0.194 0.230 0.353
Wmax 0.047 0.073 0.047 0.309 0.202 0.200 0.564 0.383 0.731 1.038
The analysis of results was presented according to the following criteria: a) for each
optical method, analysis of the variation of parameters with the increase/decrease of the
surface texture; b) in the case of the 2D-LRA method, and essentially for the left as-cast
surface, analysis of the influence of the sensor resolution; c) for each parameter,
analysis of the corresponding variation when assessed by each of the adopted optical
methods; and d) analysis of results obtained with the SPT and comparison with results
from the remaining methods.
Variation of parameters with surface texture
Since the considered surfaces have an increasing roughness, from smooth to rough with
aggregates exposure, the numerical parameters should express this increase. Therefore,
for each method, the value of each parameter should increase from left as-cast to sand-
blasting, being wire-brushing the intermediate value.
With the exception of three, all roughness parameters presented the expected variation.
Maximum Profile Peak Height (Rp) and Maximum Height of the Profile (Rt) decrease
from left as-cast to the wire-brushed surface and then increase from this to the sand-
blasted surface. Maximum Profile Valley Depth (Rv) decreases from the wire-brushed
– IV.21 –
surface to the sand-blasted surface. These exceptions are not significant and can be
explained. Actually, all these parameters are based on extremes values, maximum peaks
and valleys, and, therefore, the existence of deep air holes or exposed aggregates can
generate these abnormal values.
Variation of parameters with texture quantification methods
The analysis of the numerical parameters, assessed with all methods, except for the SPT,
was made using a statistical approach. For each numerical parameter, the following
values were computed: average (AVG); standard deviation (STD); and coefficient of
variation (COV), using the results obtained from the three methods, which are presented
in Tables IV.8 to IV.10 for the primary, roughness and waviness profiles, respectively.
Table IV.8 - Analysis of primary parameters.

Primary
Parameters AVG STD COV AVG STD COV AVG STD COV
(mm) (mm) (%) (mm) (mm) (%) (mm) (mm) (%)
Pa 0.046 0.024 51.1 0.124 0.041 33.4 0.239 0.051 21.4
Pp 0.204 0.114 56.1 0.363 0.145 40.0 0.650 0.164 25.3
Pv 0.379 0.356 94.0 0.682 0.391 57.3 0.955 0.579 60.6
Pt 0.583 0.303 52.1 1.045 0.469 44.9 1.605 0.580 36.2
Ptm 0.128 0.063 49.4 0.285 0.111 38.9 0.544 0.089 16.4
Ppm 0.057 0.021 37.8 0.124 0.022 18.0 0.263 0.022 8.5
Pvm 0.071 0.041 58.3 0.161 0.089 55.5 0.281 0.070 24.9
Pq 0.077 0.050 64.1 0.169 0.075 44.2 0.319 0.099 31.0
Pmax 0.487 0.300 61.6 0.764 0.379 49.6 1.205 0.455 37.7
– IV.22 –
Table IV.9 - Analysis of roughness parameters.

Roughness
(mm) (mm) (%) (mm) (mm) (%) (mm) (mm) (%)
Ra 0.021 0.012 56.1 0.047 0.017 35.6 0.089 0.024 27.1
Rp 0.200 0.060 30.3 0.215 0.061 28.4 0.351 0.057 16.2
Rv 0.257 0.160 62.3 0.372 0.207 55.7 0.418 0.173 41.4
Rt 0.456 0.184 40.3 0.588 0.248 42.3 0.769 0.222 28.9
Rtm 0.117 0.067 57.2 0.222 0.102 45.8 0.362 0.127 35.2
Rpm 0.055 0.027 48.3 0.094 0.026 27.7 0.171 0.041 23.9
Rvm 0.062 0.040 65.2 0.129 0.076 58.9 0.190 0.087 45.5
Rq 0.043 0.025 57.6 0.071 0.031 43.9 0.119 0.036 30.6
Rmax 0.424 0.207 48.8 0.548 0.228 41.7 0.725 0.195 26.9
Rz 0.117 0.067 57.2 0.222 0.102 45.8 0.362 0.127 35.2
RzJIS 0.217 0.050 22.9 0.361 0.083 23.0 0.578 0.083 14.4
Table IV.10 - Analysis of waviness parameters.

Waviness
(mm) (mm) (%) (mm) (mm) (%) (mm) (mm) (%)
Wa 0.039 0.024 61.8 0.106 0.049 45.8 0.194 0.053 27.4
Wp 0.091 0.048 52.3 0.274 0.142 51.8 0.437 0.180 41.2
Wv 0.161 0.200 124.5 0.352 0.251 71.4 0.709 0.404 57.0
Wt 0.252 0.221 87.7 0.625 0.393 62.8 1.146 0.460 40.2
Wtm 0.038 0.021 55.5 0.119 0.029 24.7 0.279 0.018 6.4
Wpm 0.018 0.012 63.8 0.062 0.007 11.4 0.134 0.024 17.7
Wvm 0.020 0.009 47.7 0.057 0.023 40.0 0.145 0.008 5.2
Wq 0.055 0.040 73.8 0.135 0.070 52.3 0.259 0.083 32.1
Wmax 0.143 0.145 101.1 0.322 0.210 65.1 0.717 0.328 45.7
In Figures IV.18 to IV.20 are represented the coefficients of variation determined for
each parameter and surface type. With the exception of three parameters, it can be stated
that the variation of the numerical parameters reduces with the increase of the surface
roughness. This indicates that the resemblance between readings increases from smooth
to rough surfaces, for the three methods. Consequently, it is advisable to use more
accurate measuring devices for very smooth surfaces in order to reduce the variability
and increase the accuracy of results.
– IV.23 –
The high values observed for the coefficient of variation can be related with the reduced
number of measured profiles that, due to the existence of deep air holes or exposed
aggregates, can have a significant influence in the obtained results. These regions
should always be avoided in the measurements by selecting an adequate path for the
evaluation length or by editing the measured profile removing these areas.
100%
90%
80%
70%
60%
50% LAC
40% WB
30% SAB
20%
10%
0%
Pa Pp Pv Pt Ptm Ppm Pvm Pq Pmax
Primary Parameter
Figure IV.18 - Coefficient of variation of the primary parameters.

70%
60%
50%
40%
30% LAC
WB
20% SAB
10%
0%
Ra Rp Rv Rt Rtm Rpm Rvm Rq Rmax Rz RzJIS
Roughness Parameter
Figure IV.19 - Coefficient of variation of the roughness parameters.

140%
120%
100%
80%
60% LAC
WB
40% SAB
20%
0%
Wa Wp Wv Wt Wtm Wpm Wvm Wq Wmax
Waviness Parameter
Figure IV.20 - Coefficient of variation of the waviness parameters.
It is expectable that with a large increase of the number of measured profiles, the
– IV.24 –
variability of the results will decrease significantly.
Discussion of results from the Sand Patch Test
The Sand Patch Test, defined by ASTM E 965 (2001), was performed on the three
surfaces considered, Figure IV.21. It were adopted four measurements for each surface
and a sand volume of 100ml. The same average diameter was measured for the left as-
cast and for the wire-brushed surfaces, 125mm, and a smaller value for the sand-blasted
surface, 110mm. Therefore, the corresponding values of the Mean Texture Depth for the
three situations are 8.15mm, 8.15mm and 10.52mm, respectively. The results achieved
with this method do not reproduce different texture surfaces and thus they are not in
accordance with results obtained with the optical methods.
a) b) c)
Figure IV.21 - Sand Patch Test: a) left as-cast; b) wire-brushed; and c) sand-blasted.
IV.3.3 Conclusions
The three tested optical methods – TDI method; 2D-LRA method; and 3D-LS
method – proved to be capable of being used for the texture characterization of concrete
surfaces.
Making a direct comparison between the three methods, it is possible to state that the
2D-LRA method presents all the advantages and overcomes the disadvantages of the
remaining two optical methods. The only advantage that can be pointed to the 3D-LS
– IV.25 –
method if the fact of making possible to obtain a 3D sample of the concrete surface.
Contrary to the tested optical methods, the SPT cannot be used with the purpose of
texture characterization, since it does not reproduce the difference between surfaces
with different textures, mainly if these are smooth or little rough. It must be highlighted
that this slight difference in roughness can correspond to a significant difference in the
shear strength of the resulting concrete-to-concrete interface. Moreover, the SPT also
presents the major disadvantage of only being applicable to horizontal floor surfaces.
The considered surfaces presented an increasing roughness, from smooth to rough with
aggregates exposure, and the corresponding texture parameters express this increase.
The only exceptions were observed for texture parameters based on extreme values and
are probably related with the existence of deep air holes or exposed aggregates in the
path of the evaluation length.
The variation of the texture parameters reduced with the increase of the surface
roughness, indicating that the resemblance between readings increases from smooth to
rough surfaces. Therefore, it is advisable to use more accurate measuring devices for
very smooth surfaces in order to reduce the variability and increase the accuracy of
results.
The texture parameters, evaluated considering only the primary profile, showed to be
capable of distinguish concrete surfaces with very different textures. Therefore, filtering
was not adopted until now because it was not necessary. Nevertheless, filtering presents
some advantages, separating texture into roughness and waviness, but also modifies
results depending on the choice of the filter and of the cut-off length.
Further research is necessary to study the influence of the adopted filter and cut-off
length and to verify if the primary profile is enough to characterize the surface texture,
allowing filtering to be avoid. This study is presented next.
IV.4 Influence of Filtering on Texture Parameters
Previous studies showed that roughness parameters are adequate to characterize the
– IV.26 –
texture of a concrete surface, Santos et al. (2007) and Santos and Júlio (2008).
Moreover, these parameters were computed from the primary profile with and without
filtering.
Filtering presents some advantages, separating texture into roughness and waviness, but
also modifies results depending on the choice of the filter and of the cut-off length. The
adoption of a single standard filter is not a feasible alternative, since several types of
surfaces cannot be analysed with the same filter, due to the existence of strong
curvatures, deep valleys, etc. In a similar way, the adoption of a standard cut-off length
is also not feasible, since surfaces with different roughness degrees requires different
values for this parameter.
A study to assess the influence of the adopted filter and cut-off length on the achieved
roughness and waviness parameters is now presented, Santos and Júlio (2009a). The
aim of the experimental study herein described is to investigate if the surface texture of
a concrete substrate can be effectively characterized using only the texture parameters
obtained from the primary profile, thus avoiding the use of filters. The main advantage
would be the reduction of two variable parameters, filter and cut-off length, also
allowing the comparison of results that, using different parameters, is not possible.
The selected testing samples were the same three concrete specimens, previously used,
Figure IV.4. The 2D-LRA method was adopted for this experimental study, Figure IV.1.
A laser sensor with a resolution of 10µm was used to measure the left as-cast surface,
while a similar laser sensor with a resolution of 60µm was used on the remaining two
surfaces.
Several texture parameters were considered, being each one computed as the average of
the values obtained from ten profiles. The software TrueSurf from TrueGage was
adopted to compute nine primary parameters; eleven roughness parameters; and nine
waviness parameters. The adopted parameters were the same adopted in the previous
study (Section IV.3).
– IV.27 –
First, a comparison of the texture parameters was made for the three considered
surfaces, using the same filter and cut-off length, to verify if distinctive texture
parameters correspond to different textures. It was adopted a Gaussian filter and a cut-
off length of 0.8mm.
In Figures IV.22 to IV.24 are presented the primary, the roughness and the waviness
parameters, obtained for these three surfaces. Analysing the results it can be concluded
that all primary, roughness or waviness parameters can be used to characterize the
surface texture of these surfaces, since distinctive values are assumed for each of the
three different surface textures considered.
Following this statement, different aspects were further studied. It was investigated if
varying the filter and the cut-off length, surely conducting to different values of the
adopted parameters, would keep the capability of distinguishing the different surface
textures adopted. Furthermore, it was analysed if the surface texture can be
characterized only by the means of primary parameters, i.e., if filtering is a necessary
step for the texture analysis of concrete surfaces.
To study the influence of the filter and cut-off length, only one type of surface was
considered, the sand-blasted surface. Three filters were adopted, the Gaussian filter, the
Spline filter and the Rk filter; and four cut-off lengths were considered, 0.25mm,
0.80mm, 2.5mm and 8.0mm. The selection of these filters and cut-off lengths, and not
others, is related with the fact that all of them are implemented in the software TrueSurf,
also corresponding to the most common.
In Figures IV.25 to IV.27 are presented the primary, the roughness and the waviness
parameters obtained with the Gaussian filter for all cut-off lengths; in Figures IV.28 to
IV.30 are presented the same parameters for the Spline filter; and in Figures IV.31 to
IV.33 are presented the same parameters obtained with the Rk filter.
– IV.28 –
1.600
1.200
1.000
0.800 LAC
0.600 WB
0.400 SAB
0.200
0.000
Pa Pp Pq Pt Ptm Ppm Pvm Pv Pmax
Primary Parameter
Figure IV.22 - Primary parameters with Gaussian filter (0.8mm).
0.090
0.080
0.070
Amplitude (mm)
0.060
0.050
0.040 LAC
0.030 WB
0.020 SAB
0.010
0.000
Ra Rp Rq Rt Rtm Rpm Rvm Rv Rz RzJIS Rmax
Roughness Parameter
Figure IV.23 - Roughness parameters with Gaussian filter (0.8mm).
1.600
1.400
Amplitude (mm)
1.200
1.000
0.800 LAC
0.600 WB
0.400 SAB
0.200
0.000
Wa Wp Wq Wt Wtm Wpm Wvm Wv Wmax
Waviness Parameter
Figure IV.24 - Waviness parameters with Gaussian filter (0.8mm).
– IV.29 –
1.60
Amplitude (mm) 1.40
1.20
1.00
0.80 0.25mm
0.80mm
0.60
2.5mm
0.40
8.0mm
0.20
0.00
Primary Parameter
Figure IV.25 - Primary parameters for the Gaussian filter.
0.60
0.50
Amplitude (mm)
0.40
0.30 0.25mm
0.80mm
0.20 2.5mm
0.10 8.0mm
0.00
Roughness Parameter
Figure IV.26 - Roughness parameters for the Gaussian filter.
1.60
1.40
Amplitude (mm)
1.20
1.00
0.80 0.25mm
0.80mm
0.60
2.5mm
0.40
8.0mm
0.20
0.00
Waviness Parameter
Figure IV.27 - Waviness parameters for the Gaussian filter.
– IV.30 –
2.80
Amplitude (mm) 2.40
2.00
1.60
0.25mm
1.20 0.80mm
0.80 2.5mm
8.0mm
0.40
0.00
Primary Parameter
Figure IV.28 - Primary parameters for the Spline filter.
0.70
0.60
Amplitude (mm)
0.50
0.40
0.25mm
0.30 0.80mm
0.20 2.5mm
8.0mm
0.10
0.00
Roughness Parameter
Figure IV.29 - Roughness parameters for the Spline filter.
2.80
2.40
Amplitude (mm)
2.00
1.60
0.25mm
1.20 0.80mm
0.80 2.5mm
8.0mm
0.40
0.00
Waviness Parameter
Figure IV.30 - Waviness parameters for the Spline filter.
– IV.31 –
2.80
Amplitude (mm) 2.40
2.00
1.60
0.25mm
1.20 0.80mm
0.80 2.5mm
8.0mm
0.40
0.00
Primary Parameter
Figure IV.31 - Primary parameters for the Rk filter.
0.80
0.70
Amplitude (mm)
0.60
0.50
0.40 0.25mm
0.80mm
0.30
2.5mm
0.20
8.0mm
0.10
0.00
Roughness Parameter
Figure IV.32 - Roughness parameters for the Rk filter.
2.80
2.40
Amplitude (mm)
2.00
1.60
0.25mm
1.20 0.80mm
0.80 2.5mm
8.0mm
0.40
0.00
Waviness Parameter
Figure IV.33 - Waviness parameters for the Rk filter.
Five primary parameters, including the Primary Average (Pa); the Maximum Profile
Peak Height (Pp); the Root Mean Square Primary (Pq); the Maximum Height of the
Profile (Pt); and the Maximum Profile Valley Depth (Pv), presented the same behaviour
with the variation of the filter and cut-off length. Minor differences between the four
– IV.32 –
cut-off lengths; similar values for the Spline and Rk filters; and lower amplitude values
for the Gaussian filter were observed for these parameters, when compared with the
remaining filters.
For the remaining primary parameters, including the Mean Total Height (Ptm); the Mean
Peak Height (Ppm); the Mean Valley Depth (Pvm); and the Maximum Primary Depth
(Pmax), the amplitude increased with the increase of the cut-off length and presented
similar values for the Spline and Rk filters and lower values for the Gaussian filter.
A similar behaviour was observed for the waviness parameters but with minor
differences. A small reduction of the amplitude of the waviness parameters was
registered, for the cut-off length of 2.5mm and 8.0mm.
Analysing the roughness parameters, a similar behaviour can be observed for all the
considered filters and cut-off lengths. The amplitude increases with the increase of the
cut-off length.
The adopted filter and cut-off length, used to compute the roughness and waviness
parameters, have a significant effect on the assumed values of several parameters. Some
parameters are clearly influenced by the filter and cut-off length, like the Maximum
Primary Depth (Pmax), while others only experience changes due to the variation of the
filter, like the Primary Average (Pa). The roughness parameters are more sensible to
changes in the filter and cut-off length, presenting major variations.
IV.4.3 Conclusions
It is concluded that primary, roughness and waviness parameters can be used to

characterize the texture of a concrete surface, since different values are observed for
different textures, filters and cut-off lengths. When used, the adopted filter and cut-off
length have a significant influence on texture parameters and, therefore, must be
carefully selected based on the surface preparation method and predicted spacing
between peaks and valleys.
Moreover, taken into account both conclusions previously referred to, it can be
– IV.33 –
concluded that texture characterization based on primary parameters can be used on

concrete surfaces with the following advantages: a) it is the fastest approach; b) it
eliminates the consideration of variables with influence on results, filter and cut-off
length; and c) it avoids the evaluation of the roughness and waviness parameters.
The study described herein adds an important contribution to the development of an

accurate non-destructive method to assess in situ the texture of concrete surfaces,
clarifying the need of filtering.
IV.5 Influence of Surface Irregularities on Texture Parameters
Texture parameters are based on the geometrical characteristics of the surface profile.
Peaks height, valleys depth and spacing between both are usually used to evaluate
numerical parameters that describes the surface texture.
The existence of strong surface irregularities, such as deep air holes and exposed
aggregates, can have a significant influence on the values assumed by several texture
parameters, in particular, those that are based on extreme values such as the highest
peak and the deepest valley.
These regions should be avoided from the path of the measurements or, if not possible,
should be taken into account in the determination of the texture parameters. The first
approach is the easiest but not always possible, because sometimes it is not possible to
define measuring paths without the referred irregularities. The second approach can be
conducted according to different principles. An upper limit can be defined for the value
of the texture parameter; the referred regions can be eliminated from the texture profile
by slicing it; among others.
The previous investigations were conducted with a reduced number of measured

profiles, in general, only ten profiles. The coefficient of variation of the texture
parameters, computed with this reduced number of profiles, presented high values for
almost all of them.
To investigate the influence of the referred deep air holes and exposed aggregates in the
– IV.34 –
values assumed by several texture parameters, and to verify the significance of the
number of measurements on the coefficient of variation of the computed texture
parameters, it was developed an experimental study that is described next. A large
number of measurements were considered and a statistical criterion adopted to detect
measurements that can lead to spurious values of the texture parameters.
Six different types of finishing surfaces were considered. The first one, cast against steel
or plastic formwork, is considered as the reference situation and is named left as-cast
(LAC), Figure IV.34a). The second one, free surface left as-cast after vibration, is
named free after vibration (FAV), Figure IV.34b). Three types of finishing treatments,
that can be applied to hardened concrete, are also considered: a) wire-brushing (WB),
Figure IV.34c); b) sand-blasting (SAB), Figure IV.34d); and c) shot-blasting (SHB),
Figure IV.34e). A sixth situation, hand-scrubbed surface (HS), Figure IV.34f), common
practice in precast industry, is produced while concrete is still fresh. This technique is
also known as raking since a rake is usually used as a tool to prepare the surface.
The 2D-LRA method was adopted to obtain the texture profiles of all surfaces. To
measure the left as-cast, free after vibration, wire-brushed, sand-blasted and shot-blasted
surfaces, it was adopted the laser sensor with the resolution of 10µm because it ensures
a better accuracy of results. For the hand-scrubbed surface it was adopted the laser
sensor with the resolution of 60µm. This change of laser sensor was made because the
texture irregularities of the hand-scrubbed surface are too high for the measuring range
of the laser sensor with the resolution of 10µm.
In order to obtain randomly located profiles, and not profiles with a preselected path for
the evaluation length, no special attention was given to the measured path. Preselected
paths are those defined by the user of the 2D-LRA method, or another texture
quantification method, in order to avoid specific areas of the surface, such as deep air
holes or exposed aggregates. Therefore, these zones are considered in the analysis.
Twelve roughness parameters were computed, based on the primary profile and without
– IV.35 –
filtering. The adopted parameters were the same used in a previous study and were
computed using the software application SurfTEX (Section IV.2). The form of the
profile was removed using a “best-fit” line, computed by the Least Squares Method.
a) b) c)
d) e) f)
Figure IV.34 - Surface preparation: a) left as-cast; b) free after vibration; c) wire-brushed; d) sand-
blasted; e) shot-blasted; and f) hand-scrubbed.
The Chauvenet's criterion (Taylor, 1997) was applied to the computed roughness
parameters in order to detect spurious values and to eliminate the corresponding
– IV.36 –
measurements.
According to this criterion, to detect spurious values in a sample it is necessary to

compute the mean and the standard deviation. Using a normal distribution function or
table, it is then possible to determine the probability of a value of being spurious. If the
product of this probability by the number of values of the sample is lower than 0.5, then
the value should be excluded from the sample, otherwise should be considered as valid.
A value of a set can be excluded if the probability of obtaining the particular deviation
from the mean is lower than 1/(2n), being n the number of values. This criterion can be
applied only once.
In order to eliminate the measurements corresponding to spurious values of the

roughness parameters, more than one hundred profiles were initially measured
(approximately one hundred and ten). From the initial measured profiles, and after the
application of the Chauvenet's criterion, were only selected the best one hundred
profiles. The average (AVG), the standard deviation (STD) and the coefficient of
variation (COV) of these roughness parameters, for each type of surface, are presented
in Tables IV.11 to IV.16 for the left as-cast, free after vibration, wire-brushed, sand-
blasted, shot-blasted and hand-scrubbed surfaces, respectively.
Table IV.11 - Roughness parameters for the left as-cast surface.

AVG STD COV
Roughness Parameter
(mm) (mm) (%)
Average Roughness, Ra 0.437 0.062 14.1
Mean Peak-to-Valley Height, Rz(DIN) 1.492 0.718 48.1
Maximum Peak-to-Valley Height, Rmax 3.409 2.132 62.5
Mean Third Highest Peak-to-Valley Height, R3z 1.420 0.673 47.4
Maximum Third Highest Peak-to-Valley Height, R3z,max 3.248 1.992 61.3
Ten Points Height, Rz(ISO) 3.747 1.867 49.8
Total Roughness Height, Ry 3.903 1.989 51.0
Root-Mean-Square (RMS) Profile Height, Rq 0.561 0.139 24.8
Mean Peak Height, Rpm 0.914 0.527 57.7
Maximum Peak Height, Rp 2.343 1.443 61.6
Mean Valley Depth, Rvm 0.578 0.249 43.1
Maximum Valley Depth, Rv 1.559 0.773 49.5
– IV.37 –
The coefficients of variation of the adopted roughness parameters, for each type of
surface, were also computed without the application of the Chauvenet's criterion
(Taylor, 1997). In Table IV.17 is presented the decrease/increase of the coefficient of
variation of the adopted roughness parameters.
Table IV.12 - Roughness parameters for the free after vibration surface.
AVG STD COV
Roughness Parameter
(mm) (mm) (%)
Table IV.13 - Roughness parameters for the wire-brushed surface.

AVG STD COV
Roughness Parameter
(mm) (mm) (%)
– IV.38 –
Table IV.14 - Roughness parameters for the sand-blasted surface.

AVG STD COV
Roughness Parameter
(mm) (mm) (%)
Table IV.15 - Roughness parameters for the shot-blasted surface.

AVG STD COV
Roughness Parameter
(mm) (mm) (%)
– IV.39 –
Table IV.16 - Roughness parameters for the hand-scrubbed surface.

AVG STD COV
Roughness Parameter
(mm) (mm) (%)
Table IV.17 - Decrease of the coefficient of variation with the application of the
Chauvenet's criterion.
Surface preparation
Left Free after Wire Sand Shot Hand
Roughness Parameter
as-cast vibration brushing blasting blasting scrubbing
(LAC) (FAV) (WB) (SAB) (SHB) (HS)
Average Roughness, Ra 1.8% 1.8% 18.1% 6.7% 1.8% -0.2%
Mean Peak-to-Valley Height,
4.5% 0.0% 26.1% 15.4% 1.6% -0.1%
Rz(DIN)
Maximum Peak-to-Valley
1.1% 0.5% 17.7% 42.6% 7.0% 2.4%
Height, Rmax
Mean Third Highest Peak-to-
4.8% 0.1% 26.0% 8.8% 1.5% -0.2%
Valley Height, R3z
Maximum Third Highest Peak-
1.2% 0.5% 16.6% 22.6% 5.0% 0.3%
to-Valley Height, R3z,max
Ten Points Height, Rz(ISO) 1.4% 1.3% 14.7% 24.5% 5.2% 1.1%
Total Roughness Height, Ry 1.5% 1.2% 15.7% 40.1% 6.8% 3.1%
Root-Mean-Square (RMS)
5.2% 1.8% 21.1% 20.5% 3.8% -0.2%
Profile Height, Rq
Mean Peak Height, Rpm 2.9% 0.6% 27.6% 9.5% 1.5% 1.0%
Maximum Peak Height, Rp -1.0% 2.9% 12.3% 17.5% 5.0% 10.3%
Mean Valley Depth, Rvm 7.3% -0.1% 20.6% 34.6% 2.1% -0.3%
Maximum Valley Depth, Rv 8.8% 0.0% 19.7% 110.2% 17.5% 0.3%
– IV.40 –
Analysis of the roughness parameters
The average values of the roughness parameters, for all surface situations, are presented
in Figure IV.35. Analysing the computed values, it is possible to conclude that smooth
or very smooth surfaces, such as the left as-cast and free after vibration surfaces,
presented higher values when compared with some of the remaining surfaces, such as
the wire-brushed and the sand-blasted surfaces. This result is not in accordance with the
results achieved in previous investigations made in this research study. The left as-cast
surface should present the lowest roughness parameters because it is clearly the
smoothest surface of all the six considered.
9
8
7
6
Amplitude (mm)
5 LAC
4 FAV
3 WB
2 SAB
SHB
1
HS
0
z
)
Ra
)
ax
ax
m
Ry
Rq
Rp
Rv
m
IN
O
R3
Rp
Rv
Rm
IS
D
(
(
R3
Rz
Rz
Roughness Parameter
Figure IV.35 - Roughness parameters.
In Figures IV.36 to IV.47 are presented the average values of each roughness parameter
but also the deviation to the minimum and to the maximum value. It is possible to
observe the range of the values corresponding to each roughness parameter and the
incongruence, previously described, related with the roughness parameters of the left as-
cast and of the free after vibration surface.
– IV.41 –
LAC
Surface preparation
FAV
WB
SAB Above the average

Below the average
SHB
HS
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
Average Roughness, Ra (mm)
Figure IV.36 - Average and deviation of the Average Roughness.
LAC
Surface preparation
FAV
WB

Below the average
SHB
HS
0 1 2 3 4 5 6 7 8
Mean Peak-to-Valley Height, Rz(DIN) (mm)
Figure IV.37 - Average and deviation of the Mean Peak-to-Valley Height.
LAC
Surface preparation
FAV
WB

Below the average
SHB
HS
0 1 2 3 4 5 6 7 8 9 10 11
Maximum Peak-to-Valley Height, Rmax (mm)
Figure IV.38 - Average and deviation of the Maximum Peak-to-Valley Height.
– IV.42 –
Surface preparation LAC
FAV
WB

Below the average
SHB
HS
0 1 2 3 4 5 6 7 8
Mean Third Highest Peak-to-Valley Height, R3z (mm)
Figure IV.39 - Average and deviation of the Mean Third Highest Peak-to-Valley Height.
LAC
Surface preparation
FAV
WB

Below the average
SHB
HS
0 1 2 3 4 5 6 7 8 9 10 11
Maximum Third Highest Peak-to-Valley Height, R3z,max (mm)
Figure IV.40 - Average and deviation of the Maximum Third Highest Peak-to-Valley Height.
LAC
Surface preparation
FAV
WB

Below the average
SHB
HS
0 1 2 3 4 5 6 7 8 9 10 11
Ten Points Height, Rz(ISO) (mm)
Figure IV.41 - Average and deviation of the Ten Points Height.
– IV.43 –
FAV
WB

Below the average
SHB
HS
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Total Roughness Height, Ry (mm)
Figure IV.42 - Average and deviation of the Total Roughness Height.
LAC
Surface preparation
FAV
WB

Below the average
SHB
HS
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8
Root-Mean-Square (RMS) Profile Height, Rq (mm)
Figure IV.43 - Average and deviation of the Root-Mean-Square Profile Height.
LAC
Surface preparation
FAV
WB

Below the average
SHB
HS
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
Mean Peak Height, Rpm (mm)
Figure IV.44 - Average and deviation of the Mean Peak Height.
– IV.44 –
FAV
WB

Below the average
SHB
HS
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Maximum Peak Height, Rp (mm)
Figure IV.45 - Average and deviation of the Maximum Peak Height.
LAC
Surface preparation
FAV
WB

Below the average
SHB
HS
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Mean Valley Depth, Rvm (mm)
Figure IV.46 - Average and deviation of the Mean Valley Depth.
LAC
Surface preparation
FAV
WB

Below the average
SHB
HS
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5
Maximum Valley Depth, Rv (mm)
Figure IV.47 - Average and deviation of the Maximum Valley Depth.
– IV.45 –
In previous studies of this research project, special attention was given to the path
adopted to perform the texture measurements. Deep air holes and exposed aggregates
were always avoided in order to reduce their influence in the computed texture
parameters.
The adoption of this type of procedure led to the identification of a linear relationship,
with high coefficients of correlation, between roughness parameters and the bond
strength of the concrete-to-concrete interface, both in shear and in tension. The adoption
of the roughness parameters presented in Figure IV.35, for the left as-cast and wire-
brushed surfaces, resulted in the lost of the linear relationship previously identified. This
type of problem does not exist, or at least is less pronounced, for the remaining surface
preparation methods.
Three different approaches can be adopted to overcome this problem. The first
approach, easy to implement, is the definition of an upper limit for each roughness
parameter. This limit can be defined as the assumed value by each roughness parameter
obtained with the preselected measuring paths, Table IV.2. This approach is very simple,
logical and understandable, since the bond strength of the concrete-to-concrete interface
is strongly influenced by the roughness of the substrate surface but it is not unlimited.
Therefore, an upper limit is defined for the roughness level and, indirectly, for the bond
strength of the interface.
The second approach is based in editing the measured profile, slicing the zones
containing the referred deep air holes and exposed aggregates and then computing the
corresponding texture parameters for the “shortened” profile. This approach requires
additional data processing.
The third and last approach is the selection of other texture parameters, less influenced
by the existence of deep air holes and exposed aggregates, or even the use of filters
although these make the process more time consuming.
Analysis of the coefficient of variation
The coefficient of variation of the roughness parameters, for all surface situations, is
– IV.46 –
presented in Figure IV.48.
In general, the free after vibration and the hand-scrubbed surfaces are the ones
presenting the lowest coefficients of variation. In the first case, it can be justified with
the uniform finishing surface obtained after the vibration of concrete, since deep air
holes and exposed aggregates do not exist or are rare. For the hand-scrubbed surface,
the roughness amplitude obtained with this preparation method is much higher than the
amplitude of both deep air holes and exposed aggregates and, therefore, these are less
significant.
80%
70%
60%
50%
40% LAC
FAV
30% WB
20% SAB
10% SHB
HS
0%
)
)
Ra
ax
ax
m
Ry
Rq
Rp
Rv
m
IN
O
R3
Rp
Rv
m
Rm
IS
D
(
(
R3
Rz
Rz
Roughness Parameter
Figure IV.48 - Coefficient of variation.
The remaining surfaces, with the exception of the left as-cast surface, presented higher
values for this parameter because the corresponding preparation methods are abrasive
and, therefore, removes the superficial concrete layer leaving an irregular surface.
The coefficient of variation for the left as-cast surface presented an higher value in
comparison with the free after vibration surface. This was observed for all roughness
parameters, with the exception of the Average Roughness (Ra), and is related with the
finishing process. The left as-cast surface presents, in general, several air holes after
demolding, while the free after vibration surface does not. Therefore, it is expectable an
higher variation of the roughness parameters for the left as-cast surface in comparison
with the free after vibration surface.
– IV.47 –
Influence of the application of the Chauvenet's criterion
The surfaces with partial or total exposure of aggregates, corresponding to the use of
abrasive methods, presented significant decreases in the coefficient of variation after the
application of the Chauvenet's criterion, Figure IV.49. The wire-brushed and sand-
blasted surfaces presented the most significant decreases but the shot-blasted surface is
also significantly affected.
120%
100%
80%
Decrease of the
60%
LAC
40% FAV
WB
20% SAB
SHB
0% HS
-20%
)
R3 z
Ra
ax
ax
m
Ry
Rq
Rp
Rv
m
SO
IN
R3
Rp
Rv
zm
Rm
D
(I
(
Rz
Rz
Roughness Parameter
Figure IV.49 - Influence of the Chauvenet's criterion on the decrease of the coefficient of variation.
The use of sand-blasting and shot-blasting produced very similar surfaces. The latter
appears to lead to a more uniform texture and, therefore, to lower coefficients of
variation of the roughness parameters.
The surfaces prepared with the concrete substrate still fresh, such as the free after
vibration and hand-scrubbed surfaces, presented a lower decrease in the coefficient of
variation, when compared with the surfaces prepared with the concrete substrate already
hardened, such as the wire-brushed, sand-blasted and shot-blasted surfaces.
The observed high variability of the measurements appears to be related with the
preparation method itself, as the case of sand-blasting versus shot-blasting, but also with
the application of the preparation method to fresh or hardened concrete.
The Chauvenet's criterion appears to be useful in the elimination of measurements that
– IV.48 –
can lead to spurious values of the texture parameters, in particular, for the case of
concrete surfaces prepared by abrasive methods.
Considerations about the laitance concrete layer
The free after vibration surface was only considered in this last investigation with the
purpose of enlarging the number of preparation methods for concrete surfaces. This type
of surface present, for all roughness parameters, highest values than the wire-brushed
surface and, in many cases, similar values to the roughness parameters of the sand-
blasted surface.
Since the bond strength of the concrete-to-concrete interface is highly influenced by the
roughness amplitude of the substrate surface, it is expectable that the free after vibration
surface, without any type of roughness improvement, can lead to a higher bond strength
of the interface when compared with the wire-brushed and sand-blasted surfaces. No
bond tests were conducted, during the remaining research project, to verify this and,
therefore, the previous statement is merely a suspicion.
When abrasive methods, such as wire-brushing, sand-blasting or shot-blasting are

adopted, the superficial concrete layer, usually named as laitance, is removed. This
concrete layer, present in the free after vibration surface, is mainly composed by fine
particles of cement and should be removed prior to the cast of the added concrete layer,
as proposed by the standard ASTM D 4259 (1999). ACI 318 (2008) also refers that the
substrate surface should be free of laitance.
Further investigations are necessary to clarify the influence of the laitance concrete
layer on the values of the texture parameters and on the bond strength of concrete-to-
concrete interfaces. Even presenting higher values for the roughness parameters, the
free after vibration surface can in fact present a lower bond strength of the interface,
since it includes the existence of laitance.
IV.5.3 Conclusions
Deep air holes and exposed aggregates have a strong influence in the values of some
– IV.49 –
roughness parameters, in particular, those that are based on extreme values of the profile
such as the highest peak and the deepest valley within the entire texture profile. When
performing the measurements of the surface texture it is extremely difficult to avoid all
these regions and, therefore, the computed roughness parameters will be affected.
It is recommended to adopt an upper limit for the roughness parameters used to

characterize, at least, the left as-cast surface. The adopted values for this upper limit can
be obtained from the roughness parameters computed with the preselected measuring
paths, avoiding the referred regions. Without this, the roughness parameters for the left
as-cast surface will present higher values than the ones computed for the free after
vibration surface and for the wire-brushed surface, which is unrealistic.
The roughness parameters of the free after vibration surface presented, in some cases,
higher values than the ones computed for the wire-brushed and for the sand-blasted
surface. This suggests that the bond strength of concrete interfaces, with the substrate
surface prepared with wire-brushing or sand-blasting, could be lower than the bond
strength of the same interface with the substrate surface left free after vibration.
This type of supposition needs to be investigated due to the existence of laitance. It is

advisable to provide a minimum degree of roughness, in order to improve the bond
between concrete layers, which could be achieved by removing the laitance.
The coefficients of variation of the free after vibration surface and of the hand-scrubbed
surface are the lowest for all the six considered situations. The coefficient of variation is
highly influenced by the existence of deep air holes, in the case of the left as-cast
surface, and also by exposed aggregates in the case of wire-brushed, sand-blasted and
shot-blasted surfaces.
The coefficient of variation of surfaces prepared with the concrete substrate still fresh,
such as free after vibration and hand-scrubbed surfaces, is lower than the ones obtained
with surfaces prepared with the concrete already hardened and, therefore, using more
abrasive methods.
– IV.50 –
IV.6 Conclusions
The Sand Patch Test showed to be unable to reproduce the difference between surfaces
with different textures, mainly if these are smooth or little rough. This quantification
method also presents the major disadvantage of only being applicable to horizontal floor
surfaces.
The TDI method, the 2D-LRA method and the 3D-LS method proved to be capable of
being used to characterize the surface texture of a concrete substrate, overcoming the
disadvantages of the Sand Patch Test. Moreover, the 2D-LRA method showed to be the
most adequate method for being non-destructive, contrary to the TDI method, and for
representing a significant lower cost of acquisition and maintenance, in relation to the
3D-LS method. The 2D-LRA method presents all the advantages and overcomes all the
disadvantages of the TDI method, adopted in a previous study.
Results corroborated the conclusions of previous studies, namely: a) the texture of the
substrate surface influences significantly the bond strength of concrete-to-concrete
interfaces; b) it is possible to obtain a texture profile of the substrate surface and, from
this, to determine texture parameters; and c) it is possible to define linear correlations
between some of these parameters and the bond strength of the interface, both in shear
and in tension, with very high coefficients of correlation.
The obtained results suggest that this fast, non-destructive and in situ method, based in
the use of the 2D laser roughness analyser, is expectable to be adequate to quantify the
texture of the substrate surface and that the Maximum Valley Depth (Rv) roughness
parameter is expectable to be adequate to incorporate in a design expression of the
longitudinal shear strength of the interface of concrete members with parts cast at
different times.
Texture parameters, evaluated from the primary, roughness and waviness profiles, can
be used to characterize the texture of a concrete surface. These parameters assume
different values for surfaces with different texture. When filtering is adopted, the
selected filter and the value assumed for the cut-off length have a significant effect on
the values of the texture parameters. Primary parameters proved to be sufficient to
– IV.51 –
characterize the concrete surface with the following advantages: a) it is the fastest
approach; b) it eliminates the need of filtering; and c) it avoids the evaluation of the
roughness and waviness parameters.
The existence of deep air holes and exposed aggregates in the concrete surface has a
strong influence on the values of several texture parameters, in particular, those that are
based in extreme values, originating a high variability in results. These surface
irregularities should be avoided in the path of the evaluation length. Alternatively, an
upper limit can be adopted for the values of the roughness parameters and, in particular,
for the left as-cast surface. This limit should correct the abnormal values, obtained due
to the existence of the mentioned surface irregularities.
The resolution of the laser sensor has influence on the values of the texture parameters.
The use of the 10µm laser sensor is more adequate for the left as-cast surface, instead of
the 60µm laser sensor. The latter should be used in rough surfaces, such as those
obtained by wire-brushing, sand-blasting, shot-blasting and hand-scrubbing.
The free after vibration surface and the hand-scrubbed surface, both corresponding to
finishing surfaces while concrete is still fresh, presented the lowest coefficients of
variation. The coefficient of variation of the remaining surfaces presented higher values.
The left as-cast surface and the wire-brushed surface are highly influenced by deep air
holes, while the exposed aggregates affect mainly the sand-blasted and shot-blasted
surfaces. Further investigations are necessary to reduce this parameter in order to
increase the reliability of the use of texture parameters.
A minimum degree of roughness should be provided to the substrate surface in order to

obtain a better bond; and to remove contaminants from the surface, such as oils and
grease, as well as concrete laitance.
– IV.52 –
Chapter V
Materials and Methods
V.1 Introduction
The present chapter presents the materials and methods adopted in the development of
the experimental study.
Previous studies, developed by Júlio (2001) and Santos (2005), showed that the slant
shear test is adequate to predict the strength of concrete-to-concrete interfaces in shear.
Given the good results previously achieved, the slant shear test was adopted in the
experimental study developed in the scope of this PhD thesis.
The pull-off test, commonly used to assess the bond strength in tension, was also
adopted in the studies referred to. However, debonding of the interface was frequently
observed when drilling the core hole, due to vibration. For this reason, and based on a
literature review on experimental assessment of the bond strength of concrete-to-
concrete interfaces, it was decided to adopt the splitting test instead.
The literature review also revealed that slant shear specimens with different geometries,
namely in size and shear plane angle, are used. For this reason, a numerical modelling
was conducted to study the influence of the specimen geometry on the stress
distribution at the interface between concrete parts. This modelling also allowed to
– V.1 –
define the geometry of the specimens in order to optimize the use of the available
formwork.
After presenting the results of the literature review and of the numerical modelling, the
materials and methods adopted in this study are revealed. Also, adopted concrete
mixture; curing conditions; difference of ages between concrete layers; shrinkage
quantification; bond tests; and surface preparation and characterization are described.
V.2 Literature Review on Bond Tests
The development of the concrete-to-concrete bond strength is an important factor on

precast, rehabilitation and strengthening of reinforced concrete structures. A wide range
of bond tests are available to evaluate the bond strength of concrete-to-concrete
interfaces. Some of these bond tests can be used both in situ and in laboratory, while
others can only be executed in laboratory.
An effective bond test should be capable of simulating the real stress state of a structure,
from the most simple as compression, tension or shear, to the most complex such as the
combination of compression with shear. It should also be capable of exhibiting different
failure modes – adhesive, cohesive and mixed – that can occur at: a) the old concrete
(substrate concrete); b) the new concrete (added concrete); or c) the bonding agent, if
used, Figure V.1.
new new new new new new

concrete concrete concrete concrete concrete concrete
old old old old old old

concrete concrete concrete concrete concrete concrete
a) b) c) d) e) f)
Figure V.1 - Types of failures: a) in the bonding agent; b) in the old concrete; c) in the new concrete;
d) between the bonding agent and the old concrete; e) between the bonding agent and the new
concrete; and f) mixed, crossing the new concrete, the bonding agent and the old concrete.
It is common to prepare the surface of the concrete substrate in order to achieve an
– V.2 –
higher bond. The use of concrete mixtures and repair materials with strengths different
from those of the concrete substrate is also common. Therefore, the selected bond test
should be sensitive to these parameters, variations of the surface roughness and of the
differential stiffness between parts.
Besides those already mentioned, surface roughness and strength of materials, each
bond test is influenced by a set of parameters such as: a) the specimen size; b) the load
application; and c) the orientation of the interface. Therefore, a bond test should not be
used alone but together with others, since it only provides limited data on the bond
characteristics.
Furthermore, the use of the results of each bond test could take the researcher to wrong
interpretations about the interface behaviour. The obtained results should be analysed,
interpreted and used very carefully, because in practical situations the interface is
usually subjected to a multi-stress state composed by compression, tension and shear,
instead of a single stress state as those simulated by several bond tests.
Chmielewska (2005) presents a large review about bond tests, grouping these in four
categories: a) direct tension; b) bending; c) splitting; and d) shearing. The first category
includes the direct tension test and the pull-off test and measures the bond strength
under tension stress. The second category includes several bond tests where the
specimens are tested under bending, such as the modulus of rupture test and the patch
test. The third category includes the splitting test and the wedge splitting test and
measures the bond strength of the tested specimen under tension stress combined with
compression. The fourth category includes the direct shear test; the slant shear test; the
guillotine test; the butterfly test; the bi-surface shear test; the patch test; the push-off
test (L-shaped test); and the friction-transfer test (twist-off test) and evaluates the bond
strength under shear stress combined with others as compression and torsion.
The specimens of the patch test have the particularity of being tested under different
configurations, as showed ahead in the present chapter. Therefore, this bond test can be
grouped in different categories.
Momayez et al. (2005) suggested that bond tests can be divided into the following three
– V.3 –
categories, depending on the stress state subjected to specimens: a) tension stresses;

b) shear stresses; and c) combination of shear and compression stresses. The first
category includes the direct tension test; the pull-off test; and the splitting test. The
second category includes the direct shear test; the bi-surface shear test; the push-off test
(L-shaped test); and the friction-transfer test (twist-off test). The third (and last)
category includes the slant shear test.
There is not a standard bond test, accepted by the majority of the researchers for testing
the bond of cementitious-based repair materials. Although there is not a single test that
can replicate all stress states, the slant shear test, defined by EN 12615 (1999) and
ASTM C 882 (1999), seems to be the most widely accepted by researchers for the
assessment of the bond strength in shear. The pull-off test, defined by BS EN 1542
(1999) and ASTM C 1583 (2004), is also commonly used for the assessment of the
bond strength in tension.
As an alternative to bond tests, Delatte et al. (2000a) suggested the adoption of the
maturity method to predict the concrete-to-concrete bond strength at early ages.
According to these researchers, the development of the strength in compression,
splitting tension, direct tension and shear is related with the mixture proportioning and
curing temperature and can be correlated with the concrete maturity. If adequate curing
conditions are provided and the effects of other variables controlled, the maturity
method can be used to estimate the bond strength in all the mentioned situations.
Next, it is presented a literature review on bond tests. Any type of classification or

grouping was not adopted. Since some of these tests were originally proposed in
imperial units, while others were expressed in SI units, a dual unit format is adopted.*
V.2.1 Bending Test
Bending tests are very common to assess the bond strength of concrete-to-concrete and
concrete-to-repair material interfaces. The stress distribution in this type of bond test is
* Conversion factors from SI Units to Imperial Units:

1mm = 0.039in.
1mm2 = 0.001550in.2
1mm3 = 0.000061023in.3
– V.4 –
complex, since it depends on the localization and orientation of the interface plane.
Only tension and compression stresses are present at the interface, for pure bending
without shear, or a combination of tension, compression and shear stresses, for bending
with shear.
Ohama et al. (1986), cited by Chmielewska (2005), proposed a composite beam test to
assess the bond strength of a concrete repair, Figure V.2. The composite beam specimen
presents a 40×40mm2 (1.57×1.57in.2) cross section and a 160mm (6.30in.) length. Half
of the specimen is produced with the substrate concrete while the other half is made
with the repair material. The specimen is tested in bending by applying a concentrated
load at middle span.
Wall et al. (1986), cited by Chmielewska (2005), suggested a beam test very similar to
the one proposed by Ohama et al. (1986), Figure V.3. The composite beam specimen
presents a rectangular cross section of 76×76mm2 (3×3in.2) and a span of 305mm
(12in.). Also, half of the specimen is produced with the substrate concrete while the
other half is made with the repair material. The major difference is that, in this case, the
interface is oriented between 30 to 45 degrees with the longitudinal axis of the beam.
Figure V.2 - Bending test proposed by Ohama et al. (1986).
Figure V.3 - Bending test proposed by Wall et al. (1986).
Abu-Tair et al. (1996) proposed a bending test slightly different from the previous ones,
– V.5 –
Figure V.4. First, a 100×100×400mm3 (3.94×3.94×15.75in.3) concrete specimen is cast

and, then, it is cut at middle into two equal halves. Afterwards, the repair material, with
20 to 25mm (0.79 to 0.98in.) of thickness, is placed between both halves. The composite
specimen is then tested in bending with two concentrated loads applied at one-fourth of
the span.
According to Abu-Tair et al. (1996), this bond test is suitable to assess the tensile bond
strength between a repair material and a concrete substrate.
Figure V.4 - Bending test proposed by Abu-Tair et al. (1996).
Kunieda et al. (2000) suggested the use of a different bending test. Here, a composite
concrete specimen is tested in bending, under the action of two concentrated loads
applied at one-fifth of the span, Figure V.5.
Figure V.5 - Bending test proposed by Kunieda et al. (2000).
The specimen is symmetrical, with a notch located at middle span, corresponding to

one-third of the specimen height. Half of the composite specimen is made with the
substrate concrete and the other half with the repair material or new concrete.
Specimens with a cross section of 100×200mm2 (3.94×7.87in.2) and a length of 1200mm

(47.24in.), corresponding to a span of 1000mm (39.37in.), were adopted. Other
geometries were also adopted, namely cross sections of 100×100mm2 (3.94×3.94in.2)
and 100×400mm2 (3.94×15.75in.2) and lengths of 400mm (15.75in.) and 600mm
– V.6 –
(23.62in.), corresponding to spans of 250mm (9.84in.) and 500mm (19.69in.),

respectively.
Kamada and Li (2000) used a bending test with a configuration quite different from
previous bending tests, such as those proposed by Abu-Tair et al. (1996) and Kunieda et
al. (2000), Figure V.6.
In this test, the substrate concrete is cast and, after hardening, it is cut into two equal
halves. The substrate surface is then prepared and the repair material placed over the
two concrete parts. The major difference is that a T-shaped notch, with a length of
51mm (2in.), is left between the two concrete blocks of the substrate.
The final composite specimen presents a rectangular cross section with 76mm (3in.) of
width and a total height of 100mm (4in.). The total length of the specimen is 355mm
(14in.). The specimen is tested in bending, with two concentrated loads applied at one-
fourth of the span, corresponding to a free span of 305mm (12in.).
According to the conclusions of Kamada and Li (2000), this bond test seems to be
insensitive to changes in the surface preparation. Smooth surfaces, resulting from the
cut of concrete specimens with a diamond saw, and rough surfaces, obtained with a
portable scarifier, present a similar behaviour.
Figure V.6 - Bending test proposed by Kamada and Li (2000).
V.2.2 Bi-Surface Shear Test
Momayez et al. (2004) proposed a bond test, the bi-surface shear test, to evaluate the
bond strength between old and new concrete, Figure V.7. This test is very similar to the
direct shear test, described ahead in this chapter, with only one shear plane and with a
– V.7 –
different load configuration.
The specimens for the bi-surface shear test can be cast with the same formwork used for
the standard cubic concrete specimens, since the proposed dimensions for the composite
concrete specimen are 150×150×150mm3 (5.91×5.91×5.91in.3).
The existing concrete occupies two-thirds of the mold and the new concrete the
remaining one-third. When casting the substrate concrete, a Styrofoam, wood block or
steel plate can be placed inside the formwork to reduce the mould size to two-thirds.
The production of specimens is very easy to carry out, avoiding special forms or saw
cutting, and can easily incorporate dowels. A standard testing machine can be used to
test the specimen.
Figure V.7 - Bi-surface shear test.
The main advantage of this test is that the loads are applied symmetrically, being
transmitted to the composite concrete specimen by three steel plates. The authors stated
that this type of test, with the combination of specimen size and loading case, causes a
state of shear stress that more closely represents the state of stress encountered in many
structures.
This bond test also allows the direct comparison between the compressive and shear
strength, since the specimen dimensions are the same. The proposed bond test can also
be used for testing repair mortars, instead of concrete.
V.2.3 Butterfly Test
Ray et al. (2005) proposed a direct shear bond test, usually known as the “butterfly” test
– V.8 –
due to the configuration of the resulting composite concrete specimen, Figure V.8. The
specimen is composed by two symmetrical wedges, with the shape of a “butterfly”, and
notches around the perimeter of the interface.
The height of the specimen is of 102mm (4in.), being the larger and smaller sides of the
cross section equals to 127mm (5in.) and 76mm (3in.), respectively. Therefore, the
interface shear plane is 76×102mm2 (3×4in.2). The load is applied through a 6mm
(0.24in.) wide strip. The thickness of each halve of the composite specimen is of 38mm
(1.5in.).
a) b)
Figure V.8 - Butterfly test: a) perspective; and b) side view.
A testing tool, designated as a block-shear apparatus, was specifically developed to

apply the compression load. The testing tool provides lateral confinement to the
composite concrete specimen in order to avoid the development of tensile stresses at the
interface. Therefore, the specimen is tested in shear.
V.2.4 Direct Shear Test
The direct shear test is the simplest bond test to assess the shear strength between two
materials. This test can be performed with a single (Li et al., 1997) or a double shear
plane (Chen et al., 1995), Figure V.9a) and Figure V.9b), respectively.
In this test, a composite specimen is subjected to two opposing compression forces,

each one acting on a different part of the specimen, Figure V.9. The average shear stress
at the interface is obtained dividing the failure load in compression by the bond area.
– V.9 –
a) b)
Figure V.9 - Direct shear test: a) single shear plane; b) double shear plane.
V.2.5 Direct Tension Test
The direct tension test, defined by standard ASTM C 1404 (2003), measures the tensile
bond strength of the interface between two materials, such as old-to-new concretes or
concrete-to-repair material, Figure V.10.
Figure V.10 - Direct tension test.
A composite specimen is subjected to two opposing tensile forces, usually transmitted

by two steel plates glued at both ends or through special grips. The alignment of the
composite specimen with the tensile force is very important since eccentricities
introduce bending, leading to false results.
V.2.6 Friction-Transfer Test
Naderi (2005) proposed the friction-transfer test, also known as the twist-off test, for the
assessment of the compressive strength of concrete, Figure V.11. This test is a partially
destructive method, since it involves drilling, and can be used both in situ and in
– V.10 –
laboratory.
Figure V.11 - Friction-transfer test.
The test involves drilling a core, with 50mm (1.97in.) of diameter and 25mm (0.98in.) of
depth, into the concrete specimen. A specially designed gripping device is glued to the
top of the core. An increasing torque is gradually applied by hand to the core, and
measured by an ordinary torque-meter, until failure occurs. The torsional shear strength
can be calculated and correlated to the compressive strength of concrete using
calibration charts previously determined.
Naderi (2005) indicates that linear relationships between the concrete tensile strength
and its torsional shear strength, obtained from the pull-off test and the friction-transfer
test, respectively, can be defined for each type of concrete. Therefore, this test method
can be used to assess the bond strength of concrete repairs. In this case, the core depth
should be extended into the substrate concrete in order to test the interface bond
strength.
V.2.7 Guillotine Test
Delatte et al. (2000b) adopted a shear test to evaluate the bond strength of concrete-to-
concrete interfaces, Figure V.12. This bond test is usually known as the guillotine test.
In this test, first a composite concrete specimen, with a cylindrical cross section of
100mm (3.94in.) of diameter, is placed inside the testing apparatus, which is similar to a
“shear box”. Then, a compression load is applied on top of this “shear box”, inducing a
shear force at the bond interface of the composite specimen.
According to Delatte et al. (2000b), the guillotine test is a bond test most adequate to be
– V.11 –
conducted in laboratory, with limited application in situ since it is very difficult to

extract an undamaged core.
Interface
a) b)
Figure V.12 - Guillotine test: a) front view; and b) side view.
V.2.8 Patch Test
The patch test was proposed by Austin and Robins (1993) and was specifically
developed to be used in the assessment of the bond strength of shallow concrete patch
repairs. The proposed bond test, as mentioned by the authors, is capable of modelling
realistically a patch repair, including several load and repair conditions such as axial
compression/tension and bending, Figures V.13 and V.14 respectively.
a) b)
Figure V.13 - Patch test: a) in compression; and b) in tension.
The test specimen is a 100×100×400mm3 (3.94×3.94×15.75in.3) prism with a circular

patch repair in one of the longitudinal faces. With this arrangement, it is possible to test
– V.12 –
the patch repair in compression, tension and bending. The specimen can also be loaded
before the repair is applied, simulating a pre-stressing condition.
a) b)
Figure V.14 - Patch test: a) with positive bending; and b) with negative bending.
V.2.9 Pull-Off Test
The pull-off test, Figure V.15, is one of the most widely used tensile bond tests. It was
initially developed by NTH Consultants, Ltd. Later, it was proposed as a standard test
for the evaluation of the bond strength of concrete repairs, being defined in BS EN 1542
(1999) and ASTM C 1583 (2004).
Figure V.15 - Pull-off test.
The main advantage of the pull-off test, is its capability of being used both in laboratory
and in situ.
The preparation and execution procedure of the pull-off test consists in, first, drilling a
circular core into the repair material, until it reaches the substrate concrete and
extending it bellow the interface surface. Then, a circular steel disk is glued to the upper
surface of the core using an epoxy resin. A device is then attached to the steel disk and a
tension force is applied until failure occurs.
The failure can occur at: a) the repair material; b) the substrate concrete; c) the interface
between the repair material and the substrate concrete; or d) both the repair material and
– V.13 –
the substrate, situation in which the failure surface crosses the interface. This last failure
mode is not very common but it can occur if the tensile force is applied eccentrically,
which naturally should always be avoided.
Previous studies conducted by Santos et al. (2007) and Júlio et al. (2004) showed that
debonding frequently occurs when drilling the core hole due to vibration.
V.2.10 Push-Off Test
The push-off test is a shear test normally adopted to assess the bond strength of
concrete-to-concrete interfaces when steel reinforcement crossing the interface is
considered, Figure V.16. It is also known as the L-shaped test.
Figure V.16 - Push-off test.
The oldest reference that was found to this bond test pointed to Anderson (1960). Later,
it was adopted and modified by several researchers, such as Hofbeck et al. (1969),
Mattock and Hawkins (1972), Mattock (1974) and Walraven et al. (1987).
It is composed by two concrete parts, each one with the shape of an L, connected
through the interface and presenting a gap at both ends which allows the test to carry on
after the failure of the interface in shear.
– V.14 –
V.2.11 Slant Shear Test
The slant shear test is one of the most wide spread bond tests, Figure V.17. It was
proposed by Kriegh (1976) and it is also known as the Arizona slant shear test, because
it was developed at the University of Arizona. It is currently defined by several
standards, such as EN 12615 (1999) and ASTM C 882 (1999).
Figure V.17 - Slant shear test.
In this bond test, the interface is subjected to a stress state of compression combined
with shear. According to Kriegh (1976), the slant shear test is a reliable and consistent
method, because it simulates a stress state close to the actual use and failure mode of
concrete.
The slant shear specimen, proposed by Kriegh (1976), is a composite concrete cylinder
with 150mm (6in.) of diameter and 300mm (12in.) of height with a bond interface at 30
degrees with the longitudinal axis of the specimen. This composite cylinder is tested in
compression as a standard concrete specimen.
In alternative, prismatic slant shear specimens can be used. One disadvantage of this
bond test is the difference between standards regarding specimen dimensions and joint
angles, since these differences make difficult the comparison between obtained results,
as stated by Clímaco and Regan (2001). Zilch and Reinecke (2000) proposed an
alternative prismatic slant shear specimen, with a notch located at both ends of the shear
– V.15 –
plane, in order to better induce adhesive failure modes.
The composite cylinder is produced by placing an elliptical plate in the standard

cylinder mold, creating a division between the old and the new concrete at 30 degrees
with the longitudinal axis of the specimen. First, the half corresponding to the old
concrete is cast. Then, the plate is removed and the specimen cured. If necessary, the
surface of the bond interface is prepared using wire-brushing, sand-blasting or any other
technique and, eventually, a bonding agent applied. Afterwards, the new concrete is
added and, finally, the composite cylinder is tested.
V.2.12 Splitting Test
The splitting test, Figure V.18, is used to measure the tensile bond strength between two
materials, usually two different concretes. It was independently developed and proposed
almost simultaneously by Carneiro and Barcellos (1953) and Akazawa (1953),
according to Thaulow (1957). This test is also known as the Brazilian test and it is
currently defined in standards, such as EN 12390-6 (2004) and ASTM C 496 (2004).
a) b)
Figure V.18 - Splitting test: a) cubic specimen; b) cylindrical specimen.
In this test a composite specimen is tested by applying two opposing compressive

forces, in a direction parallel to the interface surface. Failure is reached in tension
normal to the interface surface, due to the Poisson effect.
The splitting test is very simple and easy to perform. The formwork used to cast the
concrete specimens used for standard compressive tests, for quality control purposes,
can be used to cast the specimens of the splitting test. A standard machine can also be
– V.16 –
used to test the specimens.
V.2.13 Wedge Splitting Test
Tschegg et al. (2000) proposed the wedge splitting test to study the bond strength
between old and new concrete, with and without anchors, Figure V.19.
The wedge splitting specimen is a composite prism made of two concrete parts,
corresponding one of these to old concrete and the other one to new concrete.
Tschegg et al. (2000) used specimens with the dimensions of 150×250×270mm3
(5.91×9.84×10.63in.3) for each part of the composite specimen, based on average
dimensions of real members used in repair and strengthening operations of concrete
bridges and buildings.
Figure V.19 - Wedge splitting test.
First, the concrete substrate is cast. Then, the anchor, if considered, is positioned with
the substrate already hardened. Subsequently, the surface is prepared to improve the
roughness and the added concrete is cast. A notch is always left at the top of the
interface to apply the wedge.
Finally, a compressive load is applied to the wedge, inducing two opposing horizontal
forces on top, until splitting occurs. The opening of the specimen is controlled and
monitored by the testing apparatus.
– V.17 –
V.3 Numerical Modelling of the Slant Shear Test
Due to the good results obtained in previous research studies, conducted by Júlio (2001)
and Santos (2005), the slant shear test was adopted in the experimental study developed
in the scope of this PhD thesis to assess the bond strength of the concrete-to-concrete
interface in shear.
A specially conceived formwork was built to produce the slant shear specimens. In
order to optimize the adopted formwork and due to physical restrictions of the
equipment, the following dimensions for the slant shear specimen were adopted: a cross
section of 150×150mm2 and a maximum height of 600mm.
Many other researchers also adopted this bond test, because of its reliability in the
prediction of the bond strength between old and new concrete, such as Al-Gahtani et al.
(1995), Abu-Tair et al. (1996) and Clímaco and Regan (2001).
Nevertheless, these researchers used slant shear specimens with different geometries,
namely in size and in angle of the shear plane with the specimen's axis. Since the
geometry of the tested specimen has influence in the achieved bond strength, some
questions can be asked: 1) Is the bond strength obtained from specimens with different
geometries comparable?; 2) Which one corresponds to the most accurate value of the
bond strength?; and 3) What are the most adequate dimensions for the slant shear
specimen?
As stated by Clímaco and Regan (2001), the slant shear specimen defined in standards
from different countries, such as France, Italy, Great Britain and United States of
America, presents different sizes and different shear plane angles. It is possible to find
specimens with 100×100×300mm3 or 70×70×200mm3 and shear plane angles between
17 and 30 degrees with the vertical.
For all the reasons previously referred to, it was decided to conduct a numerical
modelling to evaluate the influence of several parameters on the stress distribution along
the concrete-to-concrete interface of the slant shear specimen.
– V.18 –
A commercial finite element package, LUSAS 14.1 (2008), was adopted. The assessed
parameters were: a) the specimen geometry; b) the confinement provided by the plates
of the testing machine, due to friction; c) the difference between the Young modulus of
the substrate concrete and of the added concrete layer; and d) the differential shrinkage
between the substrate concrete and the added concrete layer.
A 3D model of the slant shear specimen was built using a mesh of 1000 finite elements
with twenty nodes hexahedrons, Figure V.20, and a linear behaviour of the material was
assumed. In Figure V.21 is presented a sliced slant shear specimen with the location of
the considered mean line of the interface (segment AB).
Figure V.20 - Finite elements mesh. Figure V.21 - Mean line of the interface.
Several strength classes were considered depending on the modelling objectives:

a) C20/25; b) C25/30; c) C40/50; and d) C90/105. The Young modulus was determined
according to Eurocode 2 (2004): 30GPa, 31GPa, 35GPa and 44GPa, respectively. The
coefficient of Poisson was considered equal to 0.2.
V.3.1 Influence of the Specimen Geometry
The slant shear specimen is made of two prismatic parts with a trapezoidal section. This
– V.19 –
geometry can induce stress concentrations at the interface edges. In order to avoid the
influence of having two different concretes at the specimen ends, according to the Saint-
Venant's principle (2006), the height of the smaller side of half specimen should be at
least equal to the major dimension of the specimen's cross section.
Adopting a cross section of 150×150mm2, the height of the smaller side of the specimen,
for the reasons mentioned, should be at least 150mm. Considering the shear plane at 30
degrees with the vertical, which is the most common value, the total height of the slant
shear specimen corresponds to approximately 560mm, which is in agreement with the
restrictions of the equipment previously referred to. For a shear plane at 20 degrees with
the vertical, the total height of the specimen would be 712mm, bigger than the
maximum admissible value.
Taken into account the situation described before, a numerical study was developed to
assess the influence of the specimen geometry in the stress distribution along the
interface. The main goal was to define the minimum height of the smaller side of half
specimen in order to: a) obtain an acceptable stress distribution along the interface;
b) adopt a single geometry for all slant shear specimens, only varying the shear plane
angle and the corresponding total height; and c) optimize the use of the available
formwork.
It was decided to perform the numerical study using a prismatic slant shear specimen
with a cross section of 150×150mm2 with the shear plane angle at 30 degrees with the
vertical. The following values for the height of the smaller side of each half specimen
were considered in the analysis: 50mm, 75mm, 100mm, 125mm and 150mm, Figure
V.22.
A C20/25 concrete was considered. The nodal supports at the bottom and top faces of
the specimen were assumed restrained, in order to simulate the effect of friction
between the testing machine plates and the concrete specimens.
An imposed displacement of 1mm was applied to the bottom face of the model to
simulate the action of the plates of the testing machine. The specimen presenting an
height of the smaller side equal to 150mm, corresponding to a total height of 560mm,
– V.20 –
was taken as reference. The longitudinal strain corresponding to the imposed

displacement was of 1.786×10-3.
100mm 125mm 150mm

50mm 75mm
Figure V.22 - Specimen´s geometry.
Since the remaining specimens presents a smaller total height, different imposed
displacements were considered, namely 0.643mm, 0.732mm, 0.821mm and 0.911mm for
the specimens with the smaller side presenting 50mm, 75mm, 100mm and 125mm,
respectively. Therefore, the same longitudinal strain was considered acting on all
specimens and the direct comparison of stresses at the interface is possible.
The distribution of shear and normal stresses, along the mean line of the interface, is
presented in Figures V.23 and V.24, respectively. It can be observed that, for a smaller
side with an height of 50mm and 75mm, the variation of both shear and normal stresses
along the mean line of the interface is slightly curve. With a smaller side of at least
100mm, the stress distribution is almost uniform, as expected.
-23.0
-23.5
Shear stress (MPa)
50mm
-24.0 75mm
100mm
125mm
-24.5 150mm
-25.0
0 30 60 90 120 150 180 210 240 270 300
Position along the mean line of the interface (mm)
Figure V.23 - Variation of the shear stress with the height of the smaller side of half specimen.
– V.21 –
-13.0
-13.5
Normal stress (MPa)
50mm
-14.0 75mm
100mm
125mm
-14.5 150mm
-15.0
0 30 60 90 120 150 180 210 240 270 300
Figure V.24 - Variation of the normal stress with the height of the smaller side of half specimen.
From this numerical simulation, it can be concluded that the slant shear specimen
should have approximately 100mm at the smaller side to ensure a uniform stress
distribution along the mean line of the interface, also taken into account the physical
limitations of the equipment and to optimize the use of the formwork.
In the scope of another research study, dedicated to investigate the influence of the shear
plane angle of the slant shear specimen on the bond strength of the concrete-to-concrete
interface, three shear plane angles were considered: a) 20 degrees; b) 30 degrees; and
c) 40 degrees. For a smaller side of 100mm and a shear plane oriented at 20 and 40
degrees with the vertical, the total height of the resulting composite specimens is
612mm and 379mm, respectively. Therefore, with a smaller side of approximately
100mm it is possible to produce identical slant shear specimens varying only the shear
plane angle. This allows the comparison of results between both studies, this one and
the referred to.
V.3.2 Influence of the Confinement of the Support Conditions
Depending on the boundary conditions, by performing the slant shear test, the bottom
and top plates of the testing machine can restrain the deformation of the composite
concrete specimen's edges. This restraint can have significant influence in the stress
distribution at the interface. Therefore, another numerical study was conducted to
– V.22 –
analyse this influence.
The adopted geometry for the slant shear specimen was a prism with a cross section of
150×150mm2 with the shear plane angle at 30 degrees with the vertical. The height of
the smaller side of half specimen was considered equal to 100mm, defined according to
the study described before. A C20/25 concrete was assumed for both parts of the
specimen.
Two conditions were considered for the nodal supports at the bottom and top faces of
the specimen: a) totally restrained; and b) unrestrained. An imposed displacement of
1mm was applied to the bottom face of the model to simulate the action of the plates of
the testing machine.
The distribution of both shear and normal stresses, along the mean line of the interface,
is presented in Figures V.25 and V.26, respectively. It is clear that the distribution of
both shear and normal stresses along the interface is influenced by the restraining effect
of the plates of the testing machine. Ideally, this restraining effect should be avoided
using, e.g., a steel roller-bearing between the slant shear specimen and the plates.
-28.6
-28.8
Shear stress (MPa)
-29.0
Restrained
Unrestrained
-29.2
-29.4
0 30 60 90 120 150 180 210 240 270 300
Figure V.25 - Variation of the shear stress with the restraining of the plates.
– V.23 –
-16.4
Normal stress (MPa)
-16.6
-16.8
Restrained
Unrestrained
-17.0
-17.2
0 30 60 90 120 150 180 210 240 270 300
Figure V.26 - Variation of the normal stress with the restraining of the plates.
V.3.3 Influence of the Differential Stiffness between Concrete Parts
Usually, each half of the slant shear specimen is made of a different material, therefore
with a different stiffness. Even when the same material is used in both halves,
differential stiffness can exist. As an example, considering a slant shear specimen where
both halves are made of the same concrete mixture, although cast at different ages, these
will exhibit different Young modulus and, therefore, different stiffness. For this reason,
it is important to evaluate the influence of the differential stiffness in the stress
distribution along the interface.
Again, the adopted geometry for the slant shear specimen was a prism with a cross
section of 150×150mm2 with the shear plane angle at 30 degrees with the vertical and a
height of the smaller side of half specimen equal to 100mm.
The same two support conditions were considered: a) totally restrained; and
b) unrestrained, for the reasons already presented. The same loading was considered: an
imposed displacement of 1mm, applied to the bottom face of the model.
For the substrate concrete, a C20/25 class was adopted and, for the added concrete, the
following strength classes were considered: C25/30; C40/50; and C90/105, Figure V.27.
– V.24 –
C20/25 C25/30 C40/50 C90/105
C20/25 C20/25 C20/25 C20/25
Figure V.27 - Concrete strength class of the substrate and added concrete layer.
The distribution of both shear and normal stresses, along the mean line of the interface,
is presented in Figures V.28 and V.29, respectively, considering the restraining effect of
the plates of the testing machine, and in Figures V.30 and V.31, without the restraining
effect.
-20
-25
Shear stress (MPa)
-30
-35 C20/25 – C20/25

C20/25 – C25/30
C20/25 – C40/50
-40
C20/25 – C90/105
-45
-50
0 30 60 90 120 150 180 210 240 270 300
Figure V.28 - Variation of the shear stress with different Young modulus for the added layer, with
restrained supports.
– V.25 –
-12
-14
Normal stress (MPa)
-16
-18
-20 C20/25 – C20/25

C20/25 – C25/30
-22 C20/25 – C40/50
C20/25 – C90/105
-24
-26
-28
0 30 60 90 120 150 180 210 240 270 300
Figure V.29 - Variation of the normal stress with different Young modulus for the added layer, with
restrained supports.
-20
-25
Shear stress (MPa)
-30
-35 C20/25 – C20/25

C20/25 – C25/30
C20/25 – C40/50
-40
C20/25 – C90/105
-45
-50
0 30 60 90 120 150 180 210 240 270 300
Figure V.30 - Variation of the shear stress with different Young modulus for the added layer, with
unrestrained supports.
– V.26 –
-12
-14
Normal stress (MPa)
-16
-18
-20 C20/25 – C20/25

C20/25 – C25/30
-22 C20/25 – C40/50
C20/25 – C90/105
-24
-26
-28
0 30 60 90 120 150 180 210 240 270 300
Figure V.31 - Variation of the normal stress with different Young modulus for the added layer, with
unrestrained supports.
The influence of the differential stiffness on the stress distribution along the mean line
of the interface, for both shear and normal stresses, is obvious. In fact, it can be
observed an increase of stress concentrations at both ends of the specimen with the
increase of the differential stiffness, assuming the stress distribution an S-shaped form.
Moreover, it can be seen that the influence of the support conditions, namely assuming
totally restrained or unrestrained supports, is neglectable when compared to the effect of
the differential stiffness.
Analysing the relation between shear and normal stresses, it is possible to conclude that,
for a given level of shear, the corresponding normal stress increases with the increase of
the differential stiffness, Figure V.32. This corroborates the conclusions from a previous
study conducted by Júlio et al. (2006) and explains why, from a given value of
differential stiffness, a change in the failure mode type is observed, namely from
adhesive (interface debonding) to cohesive (at the weaker concrete, normally the
substrate concrete).
– V.27 –
-20
-25
Shear stress (MPa)
-30
-35 C20/25 – C20/25

C20/25 – C25/30
-40 C20/25 – C40/50
C20/25 – C90/105
-45
-50
-28 -26 -24 -22 -20 -18 -16 -14 -12
Normal stress (MPa)
Figure V.32 - Normal versus shear stresses at the interface.
V.3.4 Influence of the Differential Shrinkage between Concrete Parts
The difference of ages between both parts of the slant shear specimen implies the
existence of a differential shrinkage, even when the same material is used in both
halves. Its existence can lead to additional stresses at the interface that, in the case of
composite concrete specimens, can result in cracking. For this reason, it is important to
evaluate the influence of the differential shrinkage in the stress distribution along the
interface.
The adopted geometry for the slant shear specimen was a prism with a cross section of
150×150mm2 with the shear plane angle at 30 degrees with the vertical and a height of
the smaller side of half specimen equal to 100mm.
The nodal supports at the bottom and top faces of the specimen were assumed
unrestrained, since the concrete shrinkage develops slowly during the curing process
and it is not affected by the restraining induced by the plates of the testing machine. A
C20/25 class was adopted for both substrate and added concrete.
To simulate the effect of shrinkage, an imposed strain of 50×10-6, 100×10-6, 200×10-6

and 400×10-6 was applied to the half specimen corresponding to the substrate concrete.
The other half was not subjected to any indirect loading.
In Figures V.33 to V.35 are presented the stress distributions at the mean line of the
– V.28 –
interface, respectively for shear, normal and equivalent stresses (von Mises stress).
6.0
4.0
Shear stress (MPa)
2.0
0.0 50E-6
100E-6
200E-6
-2.0
400E-6
-4.0
-6.0
0 30 60 90 120 150 180 210 240 270 300
Figure V.33 - Variation of the shear stress with the differential shrinkage strain.
3.0
2.0
Normal stress (MPa)
1.0
0.0 50E-6
100E-6
200E-6
-1.0
400E-6
-2.0
-3.0
0 30 60 90 120 150 180 210 240 270 300
Figure V.34 - Variation of the normal stress with the differential shrinkage strain.
– V.29 –
10
9
8
Equivalent stress (MPa)
7
6
5 50E-6
100E-6
4
200E-6
3 400E-6
2
1
0
0 30 60 90 120 150 180 210 240 270 300
Figure V.35 - Variation of the equivalent stress (von Mises stress) with the differential shrinkage
strain.
The influence of the imposed differential strain on the stress distribution along the mean
line of the interface, for both shear and normal stresses, is significant and increases with
the increase of the differential strain between parts, as expected.
When comparing the stress distributions at the interface, originated by the imposed
differential strain (Figures V.33 and V.34) and differential stiffness (Figures V.30 and
V.31), a similar behaviour can be observed at both ends of the specimen, assuming the
stress distribution an S-shaped form.
Depending on the concrete strength class, the differential shrinkage between both parts
of the slant shear specimen can lead to tensile cracking, in particular, at both ends.
V.4 Materials and Methods
Herein, the materials and methods adopted in the experimental part of the thesis are
described, including: the adopted concrete mixture; the curing conditions; the
considered differences of ages between concrete layers; the instrumentation used to
measure the concrete shrinkage; the selected bond tests; and the techniques adopted to
prepare the interface surface in order to increase its roughness.
– V.30 –
V.4.1 Concrete Mixture
The concrete mixture used in the experimental study was designed based on the
absolute volume expression, described by Lourenço et al. (2006):
csad ∑ ag wvv=1 (V.1)
In this expression, c is the absolute volume of cement, s is the absolute volume of the
adopted addition, ad is the volume of the admixture, Σag is the absolute volume of
aggregates, w is the water volume of the mixture and vv is the void volume, for unit of
apparent volume of concrete.
A single concrete composition was designed. A Portland cement type I 52.5R was
adopted to produce a C50/60 class concrete. Fine sand, coarse sand, fine limestone
crushed aggregates and coarse limestone crushed aggregates, were selected. Sika
ViscoCrete 3002 HE was used as admixture to increase the concrete strength and to
reduce the water need, although keeping the workability of fresh concrete.
The maximum dimension of the aggregates was 19.1mm, the predicted void volume of
20l and the predicted compacity 0.822. The reference granulometric curve of Faury was
used to determine the proportion of each constituent, by adjusting the mixture curve to
the reference curve using the minimum square method, Lourenço et al. (2006). The
granulometric curve, the reference curve and the mixture curve of the C50/60 class
concrete is presented in Figure V.36.
An experimental mixture was executed to verify if the defined characteristics were

achieved and minor adjustments were made. The constituents adopted for the mixture of
the C50/60 class concrete were: 295kg of fine sand; 640kg of coarse sand; 375kg of fine
limestone; 545kg of coarse limestone; 350kg of Portland cement type I 52.5R; 3.675kg
of ViscoCrete 3002 HE (admixture); and 150l of water.
– V.31 –
100
90
80
70
Total passing (%)
60
Fine sand
50 Coarse sand
40 Fine limestone
30 Coarse limestone
20 Reference curve
Mixture curve
10
0
0.297
0.074
0.149
0.59
2.38
4.76
9.52
12.7
25.4
38.1
50.8
1.19
19.1
76.2
Mesh (mm)
Figure V.36 - Granulometric, reference and mixture curves of the adopted concrete mixture.
V.4.2 Curing Conditions
Two curing conditions were considered. One set of concrete specimens was stored in the
laboratory, Figure V.37a), while another was stored outside, directly exposed to the
environmental conditions such as solar radiation, rain and wind, Figure V.37b).
a) b)
Figure V.37 - Curing conditions: a) laboratory; and b) exterior.
The temperature and the relative humidity were recorded, for both situations, using a
hygro-thermograph (Dr. Friedrichs Model 9700), capable of measuring temperatures
between -20 and +40ºC, with 1% of accuracy, and relative humidity between 0 and
100%, with an accuracy of 1% for values between 20 and 80%, Figure V.38.
– V.32 –
Figure V.38 - Hygro-thermograph.
In Table V.1 the average (AVG) value; the maximum (MAX) value; the minimum
(MIN) value; the standard deviation (STD); and the coefficient of variation (COV) of
both temperature and relative humidity, for both curing conditions, are presented. The
corresponding records and respective trend lines are presented in Figures V.39 to V.42.
Table V.1 - Temperature and relative humidity.

Temperature Relative Humidity
Curing
Period AVG MAX MIN STD COV AVG MAX MIN STD COV
condition
(ºC) (ºC) (ºC) (ºC) (%) (%) (%) (%) (%) (%)
Laboratory Oct/06 - Feb/07 18.17 24.0 13.5 2.88 15.88 70.28 98 33 12.56 17.87
Exterior Feb/07 - May/07 17.50 33.0 7.0 5.15 29.42 70.75 100 15 22.28 31.49
– V.33 –
30
25
Temperature (ºC)
20
15
10
0
0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105 112
Days
Figure V.39 - Variation with time of the temperature in the laboratory.

100
90
Relative humidity (%)
80
70
60
50
40
30
20
10
0
0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105 112
Days
Figure V.40 - Variation with time of the relative humidity in the laboratory.
35
30
Temperature (ºC)
25
20
15
10
5
0
0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105 112
Days
Figure V.41 - Variation with time of the temperature in the exterior.

100
90
Relative humidity (%)
80
70
60
50
40
30
20
10
0
0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105 112
Days
Figure V.42 - Variation with time of the relative humidity in the exterior.
– V.34 –
V.4.3 Difference of Ages between Concrete Layers
Three different situations were considered for the time gap between casting the substrate
and the added concrete layer: 28 days; 56 days; and 84 days, with the aim of studying
the effect of differential shrinkage.
Six sets of specimens were produced: L28, L56 and L84 (L series); and E28, E56 and
E84 (E series), where “L” reports to the specimens cured in the laboratory and “E” to
those cured in the exterior, followed by the difference of age, in days, between casting
the substrate and the added concrete layer.
V.4.4 Concrete Compressive Strength
A large number of concrete specimens – 150 slant shear specimens; 150 splitting
specimens; 36 cubic specimens; and 4 shrinkage specimens – was cast to take into
account: two curing conditions; three differences of ages between the substrate concrete
and the added concrete layer; and two bond tests. For this reason, twelve concrete
mixtures were made during the experimental study, six to produce the substrate concrete
and more six to produce the added concrete layer.
The concrete compressive strength was assessed in accordance to EN 12390-3 (2003).

An average value, corresponding to three concrete specimens, was considered. The
obtained results are presented in Tables A.1 and A.2 (in Appendix) for both sets of
specimens cured in the laboratory and in the exterior, respectively, and summarized in
Table V.2.
The cubic specimens used were designated ”SC”, for strength control, followed by the
specimen number, from 1 to 3; the concrete layer, “S”, for substrate and “A”, for added;
the curing condition, “L”, for cured in the laboratory and “E”, for cured in the exterior;
and the difference of age between both halves of the specimens, 28, 56 and 84 days. As
an example, the first specimen used for strength control, obtained from a concrete
mixture adopted for the substrate concrete and cured in the laboratory with 28 days of
difference of age between concrete parts was labelled SC-1-S-L28.
– V.35 –
Table V.2 - Concrete compressive strength.

Compressive strength
Testing age
Series Concrete layer AVG STD COV
(days)
(MPa) (MPa) (%)
Substrate 56 79.30 4.78 6.03
L28
Added 28 66.38 2.11 3.18
Substrate 84 85.96 3.01 3.50
L56
Added 28 80.49 0.90 1.12
Substrate 112 86.44 1.98 2.29
L84
Added 28 72.63 3.83 5.27
Substrate 56 78.90 5.43 6.88
E28
Added 28 68.26 0.72 1.05
Substrate 84 77.62 1.53 1.98
E56
Added 28 71.14 3.77 5.30
Substrate 112 81.87 0.24 0.30
E84
Added 28 69.92 4.32 6.17
The compressive strength of both substrate and added concretes is summarized in

Figure V.43. The specimens were tested at the same time of the slant shear and splitting
specimens adopted to assess the bond strength of the concrete-to-concrete interface.
Therefore, the compressive strength was evaluated at different ages, namely 28, 56, 84
and 112 days.
100
90
Compressive strength (MPa)
80
70
60
50
Substrate layer
40
Added layer
30
20
10
0
L28 L56 L84 E28 E56 E84
Curing condition - Difference of age
Figure V.43 - Concrete compressive strength.
– V.36 –
V.4.5 Shrinkage
Shrinkage was recorded for both curing conditions, using in each case two prismatic
specimens of 150×150×600mm3, according to the standard LNEC E 398 (1993).
Deformations were measured using, in each specimen, two Mitutoyo 2118SB-10
transducers attached to two calibrated steel bars positioned at a distance of 300mm,
Figure V.44.
Figure V.44 - Concrete specimen for measuring deformations due to shrinkage.
For each concrete specimen, the experimental shrinkage strain εcs was computed using
the following expression:
d f −d i
 cs= (V.2)
di
where di is the initial distance between the two steel bars attached to the specimen and
df is the same distance at a certain age. The distance df is evaluated using di and the
average of the four values obtained from the transducers. An average value,
corresponding to two concrete specimens, was considered.
– V.37 –
The experimental shrinkage strain is presented in Table A.3 (in Appendix). The
shrinkage specimens were designated “SHR”, for shrinkage, followed by the specimen
number, from 1 to 2, and by the curing condition, “L” for cured in the laboratory and
“E” for cured in the exterior. As an example, the two specimens cured in the laboratory
were labelled SHR-1-L and SHR-2-L.
The variation with time of the concrete deformation due to shrinkage is presented in
Figure V.45. The experimental value of the shrinkage strain was compared with the
theoretical value evaluated according to the Eurocode 2 (2004), expressed as the sum of
two components: the drying shrinkage (εcd) and the autogenous shrinkage (εca).
450
400
Shrinkage strain (x1E-6)
350
300
250
Laboratory (exp.)
200 Laboratory (EC2)
150 Exterior (exp.)
Exterior (EC2)
100
50
0
0 14 28 42 56 70 84 98 112
Days
Figure V.45 - Variation with time of the concrete shrinkage strain.
The drying shrinkage strain develops slowly, being a function of the migration of the
water through the hardened concrete. The autogenous shrinkage strain develops during
concrete hardening, in the early days, and it is a linear function of the concrete strength.
The total shrinkage strain (εcs) is thus given by:
 cs= cd  ca (V.3)
According to Eurocode 2 (2004), the basic drying shrinkage strain (εcd,0) is given by:
[  ]⋅10
f cm
− ds2
f cmo (V.4)
 cd ,0 =0.85  220110⋅ds1 ⋅e −6
⋅RH
– V.38 –
[  ]
3
RH
 RH =1.55 1− (V.5)
RHo
where fcm is the mean compressive strength (MPa), fcmo is equal to 10MPa, αds1 and αds2
are two coefficients that depend on the type of cement, RH is the ambient relative
humidity (%) and RHo is equal to 100%. For cement CEM 52.5 R, from class R, αds1 = 6
and αds2 = 0.11.
The variation with time of the drying shrinkage strain (εcd) is given by:
 cd t=t , t s⋅k h⋅ cd ,0 (V.6)
where kh is a coefficient depending on the notional size ho of the cross section that is
equal to 2Ac/u, being Ac the area of the cross section and u the perimeter of the cross
section exposed to drying and εcd,0 is the basic drying shrinkage strain defined
previously. The coefficient β (t, ts) is given by:
t−t s
t ,t s = (V.7)
t−t s0.04  h3o
where t is the concrete age at the age considered (days) and ts is the concrete age at the
beginning of the drying shrinkage (days).
The evolution variation with time (in days) of the autogenous shrinkage strain (εca) is
given by:
 ca t =as t  ca ∞ (V.8)
 ca ∞=2.5  f ck −10 ⋅10−6 (V.9)
as t =1−e −0.2 t

0.5

(V.10)
In Tables A.4 and A.5 (in Appendix) are presented the theoretical shrinkage strain,
assessed with Eurocode 2 (2004), for the specimens cured in the laboratory and in the
exterior, respectively.
For the curing in the laboratory, the average compressive strength of the concrete is
– V.39 –
fcm = 67.32MPa; the notional size ho = 75mm; αds1 = 6; αds2 = 0.11; the average relative
humidity is 70.28%; βRH = 1.012; εcd,0 = 356.9×10-6 and εca (∞) = 123.3×10-6. For the
curing in the exterior, the average compressive strength of the concrete is
fcm = 75.76MPa; the notional size ho = 75mm; αds1 = 6; αds2 = 0.11; the average relative
humidity is 70.75%; βRH = 1.001; εcd,0 = 321.7×10-6 and εca (∞) = 144.4×10-6.
The prediction according to the Eurocode 2 (2004) proved to be very conservative. The
resemblance between the theoretical predictions, for both curing conditions, is justified
taking into account that the average relative humidity is also approximately equal in
both situations.
The difference between the experimental measurements, for both curing conditions, is
explained with the different daily fluctuations of the relative humidity, Table V.1.
V.4.6 Bond Tests
Two bond tests were adopted to assess the bond strength of the concrete-to-concrete
interface, both in shear and in tension.
For the reasons previously mentioned, it was decided to adopt the slant shear test,
defined in EN 12615 (1999) and ASTM C 882 (1999), to assess the bond strength of the
interface in shear. The splitting test, defined in EN 12390-6 (2004) and ASTM C 496
(2004), was selected to assess the bond strength of the interface in tension.
It was decided not to adopt the pull-off test, defined in ASTM C 1583 (2004) and
BS EN 1542 (1999), to assess the bond strength of the interface in tension, selecting the
splitting test, for the reasons previously presented and that are related with debonding
when driling the core hole.
The adopted geometry for the slant shear specimens was a 150×150×450mm3 prism
with the shear plane at 30 degrees with the vertical, Figure V.46a). The geometry
defined for the splitting specimens was a 150mm cube with the interface at middle
height, Figure V.46b).
– V.40 –
Added
concrete
30º
450mm
Substrate Added
150mm concrete concrete
Substrate
concrete
150mm
150mm
150mm 75mm 75mm

a) b)
Figure V.46 - Adopted bond tests: a) slant shear test; and b) splitting test.
V.4.7 Surface Preparation and Characterization
Five different situations were considered for the interface surface between the substrate
and the added concrete layer. Left as-cast (LAC) against steel formwork was considered
as the reference situation. To increase the roughness of hardened concrete, the interface
surface was subjected to the following treatments: a) wire-brushing (WB); b) sand-
blasting (SAB); and c) shot-blasting (SHB). Hand-scrubbing (HS), a technique
commonly used in the precast industry to increase the surface roughness while concrete
is still fresh, was also considered. This technique is also known as “raking” since a rake
is usually adopted as a tool to prepare the surface.
In Figure V.47 is presented each type of surface and a typical texture profile obtained
with the 2D-LRA method, Santos and Júlio (2008). This method was also adopted to
perform a quantitative characterization of the obtained surface texture. The roughness
parameter Maximum Valley Depth (Rv), previously identified (Santos and Júlio, 2008) as
the one that best correlates with the bond strength of the concrete-to-concrete interface,
both in shear and in tension, was adopted in this study.
The average values of the Maximum Valley Depth (Rv) determined in the previous
chapter, presented in Tables IV.11 to IV.16, were adopted. For the left as-cast and wire-
brushed surfaces were adopted an upper limit taken as the average values presented in
Table IV.2, as justified in Section IV.5.2. Therefore, the adopted roughness parameter
– V.41 –
assumes the values of: 0.097mm, for the left as-cast surface; 0.451mm, for the wire-
brushed surface; 1.055mm, for the sand-blasted surface; 1.122mm, for the shot-blasted
surface; and 5.270mm, for the hand-scrubbed surface.
0.080
0.060
0.040
Amplitude (mm)
0.020
0.000
-0.020
-0.040
-0.060
-0.080
0 10 20 30 40 50 60 70 80 90 100
a)
0.500
0.400
0.300

0.100
0.000
-0.100
-0.200
-0.300
-0.400
-0.500
0 10 20 30 40 50 60 70 80 90 100
b)
0.800
0.600
0.400
Amplitude (mm)
0.200
0.000
-0.200
-0.400
-0.600
-0.800
0 10 20 30 40 50 60 70 80 90 100
c)
1.500
1.000
Amplitude (mm)
0.500
0.000
-0.500
-1.000
-1.500
0 10 20 30 40 50 60 70 80 90 100
d)
3.000
2.000
1.000
Amplitude (mm)
0.000
-1.000
-2.000
-3.000
-4.000
-5.000
-6.000
0 10 20 30 40 50 60 70 80 90 100
e)
Figure V.47 - Surface preparation and a typical texture profile: a) left as-cast; b) wire-brushed;
c) sand-blasted; d) shot-blasted; and e) hand-scrubbed.
– V.42 –
V.5 Conclusions
The bond strength of concrete-to-concrete interfaces can be experimentally evaluated

using bond tests. Some of these can be used both in situ and in laboratory, while others
can only be performed in laboratory. A single bond test can not reproduce all stress
states. Therefore, two or more bond tests should be adopted to assess the bond strength
of concrete-to-concrete interfaces.
The slant shear test is a wide spread bond test that simulates a combined stress state of
compression and shear at the interface. It is most sensitive to surface roughness and to
differential stiffness between both parts of the specimen. Changes in the failure mode
can be observed mainly due to the difference between the Young modulus of each part.
The numerical modelling of slant shear tests proved that geometry, support conditions,
differential stiffness and shrinkage of specimens have a significant influence on the
stress distribution along the interface. This study allowed the optimization of the
specimens geometry, by reducing stress concentrations at the ends, by allowing to
produce specimens with different angles of the interface and simultaneously keeping a
common geometry, and by making possible to reuse the formwork fabricated
specifically for this end.
Considering a height of the smaller side of each half of the slant shear specimen equal to
150mm, an uniform stress distribution is obtained along the mean line of the interface,
both for shear and normal stresses. Considering a height of 100mm, an almost uniform
stress distribution along the mean line of the interface is achieved being compatible with
the physical limitations of the equipment and allowing an optimized use of the
formwork to produce slant shear specimens with different angles of the shear planes
and, therefore, with different heights.
From the conclusions of the numerical part of this study, it was decided to consider in
the experimental part slant shear specimens with a cross section of 150×150mm2 and a
total height of 450mm, corresponding to a shear plane oriented at 30 degrees with the
vertical, to assess the bond strength of concrete-to-concrete interfaces.
– V.43 –
Although the restraining effect induced by the plates of the testing machine on the slant
shear specimen's edges has a significant influence on the distribution of both shear and
normal stresses along the interface, this is neglectable when compared to the influence
of differential stiffness and therefore it was decided not to take any measures to avoid
this effect.
The differential stiffness and shrinkage provoke a considerable change in the stress
distribution along the interface, both for shear and normal stresses, assuming in both
cases an increasingly pronounced S-shaped form. The developed numerical model
showed that, for the same level of shear stress at the interface, the corresponding normal
stress increases with the increase of the differential stiffness between both concrete
parts. Therefore, it can be stated that the differential stiffness has influence in the
concrete-to-concrete bond strength. Moreover, changes in the failure mode of the slant
shear specimen, from adhesive to cohesive, can be observed also due to the differential
stiffness increase. Also cracking can appear at both ends of the slant shear specimen due
to differential shrinkage.
The considered storage conditions of the concrete specimens, cured in the laboratory
and in the exterior, showed to be significantly different. The average values of both
temperature and relative humidity were very similar in both situations, although the
coefficient of variation for the second situation (exterior) was approximately the double
of the one obtained for the first situation (laboratory), due to higher daily variations of
both parameters.
Several concrete mixtures were made to produce all the specimens: cubic, slant shear
and splitting specimens. Variations of the concrete compressive strength at the day of
the test were observed and, therefore, differences in specimens differential stiffness
exist. Thus this parameter was taken into account in the discussion of results.
The prediction of shrinkage by the Eurocode 2 (2004) showed to be very conservative,

presenting significantly higher values than those experimentally observed.
The quantitative characterization of the interface surface, with the 2D-LRA method,
proved that the methods selected to increase the surface roughness led to five
– V.44 –
completely different types of finishing surfaces, as planned.
The Maximum Valley Depth (Rv) was adopted to characterize the surface texture and to
be correlated with the bond strength of the concrete-to-concrete interface obtained from
the experimental study described in the next chapter. The considered surfaces presented,
for this texture parameter, the following values: 0.097mm, for the left as-cast surface;
0.451mm, for the wire-brushed surface; 1.055mm, for the sand-blasted surface;
1.122mm, for the shot-blasted surface; and 5.270mm, for the hand-scrubbed surface.
The concrete specimens prepared by hand-scrubbing need special attention. Concrete

hardening must be well controlled to achieve a roughened surface. If the surface is
scrubbed too soon, the wanted effect will not be obtained because the concrete is still
too fresh. On the contrary, if it is scrubbed too late, the wanted effect will not be
obtained either because the concrete is already hardened.
– V.45 –
– V.46 –
Chapter VI
Assessment of the Bond Strength
VI.1 Introduction
In the previous chapter the materials and methods adopted in the experimental study,
herein presented, were described. This study was conducted to assess the influence of
several parameters on the bond strength of concrete-to-concrete interfaces, namely,
differential shrinkage and differential stiffness between concrete parts. Other
parameters, such as the surface roughness and the failure mode of the composite
concrete specimens, were also analysed.
This chapter describes all tasks conducted and presents the results of the experimental
study developed in the scope of this PhD thesis. An analytical approach, based on the
Mohr-Coulomb criterion, was adopted to analyse the results obtained from slant shear
specimens presenting different failure modes, namely adhesive (interface debonding)
and cohesive failures (monolithic behaviour). A numerical modelling was also
developed to explain the influence of the combined effect of both differential shrinkage
and stiffness, with the compressive loading at testing, on the bond strength of the
interface.
The experimental results and the influence of the considered parameters – differential
shrinkage and stiffness; surface roughness; and failure mode – are discussed.
Conclusions are drawn.
– VI.1 –
VI.2 Tasks Description
For each considered situation – curing condition; surface preparation; and difference of
age between the substrate and the added concrete – five slant shear specimens and five
splitting specimens were cast.
The slant shear and splitting specimens were named, according to the surface
preparation of the interface, “LAC”, “WB”, “SAB”, “SHB” and “HS” for the left as-
cast, wire-brushed, sand-blasted, shot-blasted and hand-scrubbed surfaces, respectively;
followed by: the specimen number, varying from 1 to 5; the curing condition, “L”, if
cured in the laboratory and “E”, if cured in the exterior; and the difference of age
between concretes parts, 28, 56 or 84 days. As an example, the first specimen (out of
five) with the surface prepared by sand-blasting, cured in the laboratory and with 28
days of difference of age between the substrate and the added concrete layer was
labelled SAB-1-L28.
The hand-scrubbed surfaces were prepared while the concrete was still fresh,
approximately two hours after casting the substrate. All the remaining surface
treatments, wire-brushing; sand-blasting and shot-blasting, were executed with hardened
concrete. Wire-brushed surfaces were prepared one week later, after casting each set of
specimens. Mainly for management reasons, sand-blasted and shot-blasted surfaces
were prepared one month later after casting the first set of specimens, when all
specimens of the corresponding curing condition were already produced.
Since the hand-scrubbed treatment has to be applied on fresh concrete surfaces, both the
slant shear specimens and the splitting specimens were cast using different methods for
this surface treatment.
For all surface treatments, except hand-scrubbing, the formwork for the slant shear
specimens was divided using a steel plate to allow the simultaneous production of two
halves. Since the formwork was intended to be used in a different research study, where
different shear plane angles were wanted, a hardened concrete cube was placed at the
end. The concrete was placed vertically, being the specimens cast side by side, Figure
VI.1.
– VI.2 –
Steel formwork
Steel separator
Hardened
Fresh Fresh
150mm
cube for
substrate substrate
height
concrete concrete
adjustment
350mm 100mm 150mm
Figure VI.1 - Placement of the substrate concrete for the slant shear test (top view).
Each hardened specimen was later placed again inside the formwork and the added
concrete was cast in the same direction, Figure VI.2. For the hand-scrubbed specimens,
the substrate concrete was cast with the formwork at 30 degrees to the horizontal and
having only one specimen at the bottom. This way, the interface surface was placed
horizontally and could be treated while fresh, Figure VI.3.
Steel formwork
Treated surface
Hardened
Fresh Hardened
150mm
cube for
added substrate
height
concrete concrete
adjustment
350mm 100mm 150mm
Figure VI.2 - Placement of the added concrete for the slant shear test (top view).
For the splitting specimens and for all surface treatments, with the exception of the
hand-scrubbed, a hardened concrete half cube of 150mm, protected with a plastic sheet,
was placed vertically inside the steel formwork. The substrate concrete was then cast in
the remaining space, Figure VI.4.
– VI.3 –
Steel
formwork
Wood
plate Cast direction
Fresh
added
30º
concrete Support
structure for
formwork
Figure VI.3 - Concrete placement for the slant shear test with the hand-scrubbed surface (side
view).
After demolding, the surface of each half splitting specimen was subjected to a
roughness treatment. Later, it was again placed inside the steel formwork and the new
concrete layer was cast. For the specimens treated with the hand-scrubbed method, the
half hardened cube was placed horizontally at the bottom of the formwork allowing the
surface of the added half to be treated while the concrete was still fresh, Figure VI.5.
The added concrete was cast afterwards, adopting the same procedure previously
described.
Cast direction Cast direction
Hardened
Fresh Hardened Fresh
150mm
150mm
half cube
substrate substrate added
for width
concrete concrete concrete
adjustment
75mm 75mm 75mm 75mm
a) b)
Figure VI.4 - Concrete placement for the splitting test (side view): a) cast of the concrete substrate;
and b) placement of the added concrete.
– VI.4 –
Cast direction Cast direction
Fresh
75mm
substrate
concrete
Hardened Fresh
150mm
substrate added
concrete concrete
75mm
Hardened half cube
for height adjustment
150mm 75mm 75mm
a) b)
Figure VI.5 - Concrete placement for the splitting test with the hand-scrubbed surface (side view):
a) cast of the concrete substrate; and b) placement of the added concrete.
After surface preparation, the specimens were cleaned with compressed air to remove
dust and stored again under the specified curing conditions. Before placing back the
specimens inside the formwork to cast the added concrete, these were cleaned with
compressed air once more. All specimens were dry, substrate and surface, when the
added concrete was placed. After 28, 56 and 84 days, for each curing condition, the
added concrete was cast on the substrate concrete. When the added concrete reached 28
days of age, the specimens were tested in compression using a standard testing machine
AMSLER 500 TNF. The adopted load rate was 1kN/s for the splitting specimens; 5kN/s
for the slant shear specimens; and 10kN/s for the cubic specimens for compressive
strength control. These values were defined in accordance to the standards EN 12390-6
(2004), EN 12615 (1999) and EN 12390-3 (2003), respectively.
VI.3 Experimental Results
The bond strength in shear was assessed with the slant shear test in accordance to the
standard EN 12615 (1999), Figure VI.6. It was evaluated dividing the tangential
component of the compressive failure load of the slant shear specimen by the area of the
interface surface, being given by:
– VI.5 –
F sin cos 
= 2 (VI.1)
b
where τ is the bond strength in shear; F is the compressive failure load of the specimen;
α is the angle between the shear plane and the longitudinal axis of the specimen; and b
is the width of the square cross section of the specimen.
In Tables A.6 to A.8 (in Appendix) the compressive failure load of the slant shear
specimens, cured in the laboratory, for respectively a difference of ages of 28, 56 and 84
days between the substrate concrete and the added concrete layer is presented. In Tables
A.9 to A.11 (in Appendix) the compressive failure load of the slant shear specimens,
cured in the exterior, for respectively a difference of ages of 28, 56 and 84 days between
the substrate concrete and the added concrete layer is presented.
Figure VI.6 - Slant shear test.
An average value (AVG), corresponding to five tests, was considered. The standard
deviation (STD) and coefficient of variation (COV) were also computed for all series. In
the case of the slant shear specimens presenting a cohesive failure (monolithic
behaviour), the computed shear stress at the interface corresponds to a lower bound of
– VI.6 –
the true shear strength.
In Figures VI.7 and VI.8 the bond strength in shear of the interface is presented for the
L and E series, respectively.
30
25
20
15 28 days
56 days
10 84 days
0
LAC WB SAB SHB HS
Surface preparation
Figure VI.7 - Bond strength in shear for L series.
25
20
15
28 days
10 56 days
84 days
5
0
LAC WB SAB SHB HS
Surface preparation
Figure VI.8 - Bond strength in shear for E series.
The bond strength in tension was assessed with the splitting test in accordance to the
standard EN 12390-6 (2004), Figure VI.9. The bond strength in tension, obtained by
compression of the splitting specimen, is given by:
2F
= (VI.2)
 Ld
where σ is the bond strength in tension given in megapascal (MPa); F is the
– VI.7 –
compressive failure load given in newton (N); L is the length of the contact line of the
specimen given in millimetre (mm); and d is the dimension of the cross section of the
specimen given in millimetre (mm).
Figure VI.9 - Splitting test.
A slight modification was made to the test procedure. The half circular steel bar, used to
distribute the compression load along the interface, was used with the flat surface in
contact with the concrete specimen, Figure VI.10. This modification allowed a better
load distribution.
The testing procedure proposed by the standard EN 12390-6 (2004) is adequate for
monolithic specimens but not adequate for composite specimens since it is quite
impossible to assure that the load is equally applied to both parts and not only to a
single part. Therefore, it was decided to make this modification to the testing procedure,
changing the position of the half circular bar and assuring, this way, that the load is
applied to both parts of the composite specimens. To ensure a smooth contact between
concrete and steel was used a rubber, with approximately 3mm thickness, placed
between the specimen and the half circular bar.
In Tables A.12 to A.14 (in Appendix) the compressive failure load of the splitting
specimens, cured in the laboratory, respectively for a difference of ages of 28, 56 and 84
days between the substrate concrete and the added concrete layer is presented. In Tables
– VI.8 –
A.15 to A.17 (in Appendix) the compressive failure load of the splitting specimens,
cured in the exterior, respectively for a difference of ages of 28, 56 and 84 days between
the substrate concrete and the added concrete layer is presented.
a) b)
Figure VI.10 - Splitting test procedure: a) proposed by EN 12390-6 (2004); and b) adopted.
An average value (AVG), corresponding to five tests, was considered. In Figures VI.11
and VI.12 the bond strength in tension of the interface is presented respectively for the
L and E series. The standard deviation (STD) and coefficient of variation (COV) were
also computed for all series.
4.0
3.5
3.0
2.5
2.0 28 days
1.5 56 days
84 days
1.0
0.5
0.0
LAC WB SAB SHB HS
Surface preparation
Figure VI.11 - Bond strength in tension for L series.
The bond strength in tension, assessed with the splitting test, should be considered equal
to 90% of the achieved experimental values as defined by Eurocode 2 (2004).
In general, the bond strength of the interface, mainly in shear, increased with the
– VI.9 –
increase of the surface roughness, as expected. Nevertheless, it was also observed an

increase of the bond strength with the increase of the difference of ages between the
substrate concrete and the added concrete layer, contrary to what was expected.
3.5
3.0
2.5
2.0
28 days
1.5
56 days
84 days
1.0
0.5
0.0
LAC WB SAB SHB HS
Surface preparation
Figure VI.12 - Bond strength in tension for E series.
In fact, the increase of the difference of ages implies an increase of the differential
shrinkage between concrete layers and, therefore, an increase of stresses at the interface.
For this reason, it was expected a decrease of the bond strength of the interface, which
did not occur. For this reason, a numerical study was performed aiming to explain the
increase of the bond strength of the interface with the increase of the difference of ages
between concrete layers and it is presented next.
The assessment of the bond strength in tension, with the splitting test, proved to be
inconclusive. The variation of the bond strength, with the increase of the surface
roughness and with the increase of the difference of ages between concrete layers, is not
clear. Therefore, when comparing the results obtained with this bond test, with the
results achieved with the slant shear test, it appears that this test is unsuitable for this
purpose.
The slant shear specimens presented both adhesive (interface debonding) and cohesive
failures (monolithic behaviour), Figures VI.13a) and VI.13b), respectively. The number
of failures is presented in Figure VI.14.
– VI.10 –
a) b)
Figure VI.13 - Failure mode: a) adhesive; and b) cohesive.
The results obtained with the tested concrete specimens can not be directly compared
since, besides the mentioned different types of failures, also differences in concrete
compressive strength were registered. For this reason, the Mohr-Coulomb criterion was
adopted to determine the pure shear strength for all specimens and the results analysis
was performed based on this parameter.
5
5 5 5 5
4
4 4 4
Cohesive failures
3
3 3 3 LAC
WB
2 SAB
2 2 2
SHB
HS
1
1 1 1 1 1 1
0
L28 L56 L84 E28 E56 E84
Figure VI.14 - Number of cohesive failures in shear.
VI.4 Analytical Approach
According to the Mohr-Coulomb criterion, the failure envelope for both the substrate
concrete and the added concrete, Figure VI.15, was defined using the experimental
value of the compressive strength (fc), obtained from three cubic specimens tested on the
– VI.11 –
same date of the slant shear and splitting tests. The corresponding analytical tensile
strength (ft) was predicted according to the Eurocode 2 (2004).
For the slant shear specimens presenting an adhesive failure, the failure envelope of the
interface was defined using the experimental values of the bond strength both in shear
(σ ; τ) and in tension (ft,i), assessed from the slant shear test and the splitting test,
respectively. Then, the pure shear strength of the interface ( τo) was computed, Figure
VI.16.
For the slant shear specimens presenting a cohesive failure, the failure envelope of the
interface was defined using the experimental values of the failure stress both in
compression and in tension, obtained from the slant shear specimens presenting a
cohesive failure and from the splitting specimens, respectively.
τ
Concrete failure
envelope
fc ft σ
Figure VI.15 - Failure envelope of both substrate concrete and added concrete.
– VI.12 –
τ
Concrete failure
envelope
τ
τo
σ ft,i σ
Interface failure
envelope
Figure VI.16 - Pure shear stress for slant shear specimens presenting adhesive failure.
Since these specimens presented a failure mode similar to the one presented by the
cubic specimens, adopted to measure the concrete compressive strength, a correlation
between the failure load of both specimens was made based on the geometry. The Mohr
circle corresponding to the specimens presenting cohesive failures was defined using a
single point over the normal stress axis, Figure VI.17.
τ
Concrete failure
envelope
τo
fc,i ft,i σ
Interface failure
envelope
Figure VI.17 - Mohr-Coulomb criterion.
– VI.13 –
The correlation between cubic specimens (quality control) and prismatic specimens
(slant shear) was made adopting the coefficients suggested by Montoya et al. (2000),
being given by:
f c ,i = f c , cyl ≈0.80 f c , cube≈1.05 f c , prism ⇔ f c , cube≈1.31 f c , prism (VI.3)
where fc,i is the concrete compressive strength assumed equal to fc,cyl which is the
compressive strength obtained from cylindrical specimens with 150mm of diameter and
300mm of height; fc,prism is the compressive strength of prismatic concrete specimens
(150×150×450mm3) and fc,cube is the compressive strength of cubic concrete specimens
(150×150×150mm3). Then, the pure shear strength of the interface (τo) was computed.
An average value, corresponding to both adhesive and cohesive failures, was

considered. In Figures VI.18 and VI.19 is presented the pure shear strength of the
interface for the L series and in Figures VI.20 and VI.21 the pure shear strength for the
E series.
9
8
Pure shear strength (MPa)
7
6
5 LAC
WB
4
SAB
3 SHB
2 HS
1
0
L28 L56 L84
Figure VI.18 - Pure shear strength, for L series, for a difference of 28, 56 and 84 days between
substrate and added concrete layer.
– VI.14 –
9
Pure shear strength (MPa) 8
7
6
5
L28
4
L56
3 L84
2
1
0
LAC WB SAB SHB HS
Surface preparation
Figure VI.19 - Pure shear strength, for L series, for different surface preparation.
6
4 LAC
WB
3
SAB
SHB
2
HS
1
0
E28 E56 E84
Figure VI.20 - Pure shear strength, for E series, for a difference of 28, 56 and 84 days between
substrate and added concrete layer.
6
4
E28
3
E56
E84
2
0
LAC WB SAB SHB HS
Surface preparation
Figure VI.21 - Pure shear strength, for E series, for different surface preparation.
– VI.15 –
VI.5 Numerical Modelling
Aiming to understand the influence of differential shrinkage and differential stiffness on

the bond strength of concrete-to-concrete interfaces, a numerical study was conducted
using the commercial finite elements software LUSAS 14.1 (2008). First, both
parameters were independently considered and, afterwards, their effects were
superimposed and combined with the compressive loading corresponding to the testing.
The 3D model of the slant shear specimen, previously used to analyse the influence of
the geometry, support conditions and differential stiffness and shrinkage was adopted,
Figure V.20. A prism with a square cross section of 150mm width and a height of
450mm with the shear plane at 30 degrees with the vertical was modelled with a finite
elements mesh with 1000 twenty nodes hexahedrons. Linear material behaviour was
assumed. Nodal displacements of all nodes at the top and bottom faces were restrained.
A prescribed vertical displacement was then applied to the bottom face nodes to
simulate the displacement of the plate of the testing machine.
VI.5.1 Differential Shrinkage
To study the effect of differential shrinkage on the stress distribution at the interface, for
the considered differences of ages, the experimentally measured shrinkage strain (Table
A.3 in Appendix) was applied to each half of the concrete specimens. For the substrate
concrete cured under laboratory conditions, the values of 65.0×10-6, 68.3×10-6 and
69.2×10-6 for a difference of age between both concrete layers of 28, 56 and 84 days,
were respectively adopted. For the substrate concrete cured in the exterior, the
corresponding values of 149.2×10-6, 192.5×10-6 and 230.8×10-6 were respectively
adopted. The shrinkage strain of the added concrete was considered equal to 47.1×10-6
and 112.5×10-6, at the age of 28 days, for curing under laboratory conditions and curing
in the exterior, respectively.
These and other relevant material properties, that were not experimentally assessed,
were computed using experimental results, Table VI.1. The Young modulus of both the
– VI.16 –
substrate concrete and the added concrete was determined according to the Eurocode 2
(2004), using the values obtained for the compressive concrete strength at the test date
(Tables A.1 and A.2 in Appendix). The coefficient of Poisson was considered equal to
0.2.
Table VI.1 - Material and time-dependent properties.

Shrinkage strain Young modulus
Series Concrete layer
(×10-6) (GPa)
Substrate 65.0 40.9
L28
Added 47.1 38.8
Substrate 68.3 41.9
L56
Added 47.1 41.1
Substrate 69.2 42.0
L84
Added 47.1 39.9
Substrate 149.2 40.9
E28
Added 112.5 39.1
E56
Added 112.5 39.6
E84
Added 112.5 39.4
In Figures VI.22 to VI.24 are presented the stress distributions at the mean line of the
interface, respectively for shear, normal stress and an equivalent stress (von Mises
stress). As expected, stresses increase with the increase of the differential shrinkage
between concrete layers and are most significant at the edges.
It can also be observed that the acting tensile stress, corresponding to the shrinkage
strain, is smaller than the maximum allowed tensile stress of the existing concrete
predicted by the Eurocode 2 (2004), 5.3MPa for a C50/60 strength class. Therefore,
according to the numerical study, the concrete at the test date is not cracked due to
shrinkage, corroborating test results.
– VI.17 –
2.5
2.0
1.5
Shear stress (MPa)
1.0
0.5 L28
0.0 L56
L84
-0.5
E28
-1.0 E56
-1.5 E84
-2.0
-2.5
0 30 60 90 120 150 180 210 240 270 300
Figure VI.22 - Shear stress distribution at the interface.
1.5
1.0
Normal stress (MPa)
0.5
L28
0.0 L56
L84
-0.5 E28
E56
E84
-1.0
-1.5
0 30 60 90 120 150 180 210 240 270 300
Figure VI.23 - Normal stress distribution at the interface.
4.0
3.5
3.0
2.5
L28
2.0 L56
L84
1.5 E28
E56
1.0
E84
0.5
0.0
0 30 60 90 120 150 180 210 240 270 300
Figure VI.24 - Equivalent stress (von Mises stress) distribution at the interface.
– VI.18 –
VI.5.2 Influence of the Differential Stiffness
A single concrete composition was adopted in the experimental study herein presented.
Nevertheless, different compressive strengths were obtained since different mixtures
were made to cast all the specimens. Using the contents of Table VI.1 and Figure VI.14,
it is possible to compute the difference of Young modulus between concrete parts of the
slant shear specimens and the number of cohesive failures, respectively, for each one of
the six series, Table VI.2.
The study related with the effect of the differential stiffness on the stress distribution
along the interface was presented in Section V.3.3 of the previous chapter. Four strength
classes were considered. The strength class of the substrate concrete was kept constant
(C20/25) while the added concrete layer was modelled with four strength classes,
namely: C20/25; C25/30; C40/50; and C90/105. The difference between Young
modulus of each part, evaluated according to Eurocode 2 (2004), was respectively of
0GPa, 1GPa, 5GPa and 14GPa.
Table VI.2 - Difference between Young modulus of the slant shear specimens.
Difference between
Series Young modulus Number of cohesive failures
(GPa)
L28 2.1 7
L56 0.8 2
L84 2.1 12
E28 1.8 16
E56 1.1 9
E84 1.9 7
Analysing the computed values of the differential stiffness, Table VI.2, it is possible to
state that the conclusions obtained with the numerical modelling, previously referred to,
are valid since the obtained differential stiffness present intermediate values of the ones
considered.
Stress concentrations were observed at both ends of the slant shear specimen, for both
shear and normal stresses, with the increase of the differential stiffness. Moreover, for a
– VI.19 –
given level of shear stresses, the corresponding normal stress increases with the increase
of the differential stiffness. Therefore, for a given value of differential stiffness, a
change in the failure mode type can occur, namely from adhesive (interface debonding)
to cohesive (monolithic behaviour).
VI.5.3 Combined Effect of Differential Shrinkage and Stiffness with

Compressive Loading
The analysis of the combined effect of differential shrinkage and differential stiffness,
with the compressive loading at the test date, was performed adopting the same 3D
model previously used. The adopted material properties, including the time-dependent
properties, were defined in accordance with the experimental results, Table VI.1.
The vertical displacement imposed to simulate the compression of the testing machine
was taken equal to the displacement corresponding to 40% of the average compressive
strength of a C50/60 strength class. This value was adopted since it is used, as suggested
by Eurocode 2 (2004), for the determination of the secant modulus of elasticity of
concrete. Adopting the material properties defined in this design code, considering an
average compressive strength of 58MPa and a Young modulus of 37GPa, it is obtained
a displacement of 0.282mm for the slant shear specimen with an height of 450mm.
The distribution of shear, normal and equivalent stresses (von Mises stress), along the
mean line of the interface, is presented in Figures VI.25 to VI.27, respectively.
From Figures VI.22 to VI.24, due to the differential shrinkage between concrete layers,
it can be observed that the interface is subjected both to compression and tension
stresses. Stresses at the interface increase with the increased difference of ages between
concrete layers, as expected, since differential shrinkage also increases.
At the test date, the concrete specimens are already subjected to a stress state due to
differential shrinkage. With compressive loading, tension stresses disappear from the
interface. Since stresses due to differential shrinkage are higher for the specimens with
higher differences of age between concrete layers, the failure load also increases with
differential shrinkage, contrary to what was expected.
– VI.20 –
-4
-5
-6
Shear stress (MPa)
-7
L28
-8 L56
L84
-9 E28
E56
-10
E84
-11
-12
0 30 60 90 120 150 180 210 240 270 300
Figure VI.25 - Shear stress distribution at the interface.
-2.5
-3.0
-3.5
Normal stress (MPa)
-4.0
L28
-4.5
L56
-5.0 L84
-5.5 E28
E56
-6.0 E84
-6.5
-7.0
0 30 60 90 120 150 180 210 240 270 300
Figure VI.26 - Normal stress distribution at the interface.
26
24
22
20 L28
L56
18 L84
E28
16 E56
E84
14
12
0 30 60 90 120 150 180 210 240 270 300
Figure VI.27 - Equivalent stress (von Mises stress) distribution at the interface.
– VI.21 –
This can be observed in Figures VI.25 to VI.27, respectively for shear, normal and
equivalent stresses at the interface. The only exception is the L28 series. No justification
was found to explain this abnormal behaviour.
VI.6 Discussion of Results
Surface Roughness
From Figures VI.18 and VI.20 it can be observed that, in general, the pure shear
strength of the interface increases, for both curing conditions, with the increase of the
surface roughness. Three exceptions, in thirty, were identified: the wire-brushed surface
for the L84 series; the sand-blasted surface for the E28 series; and the hand-scrubbed
surface for the E56 series. In these cases, the pure shear strength decreased instead of
increasing with the surface roughness.
The influence of the surface preparation is more significant for the specimens cured
under laboratory conditions instead of exterior conditions. For the first case, curing
under laboratory conditions, the specimens presented an increase of the pure shear
strength that is higher than 100%, when comparing the minimum and the maximum
values obtained respectively for the left as-cast and hand-scrubbed surfaces, Figure
VI.18.
Under exterior conditions, the increase of the pure shear strength is always lower than
100%, when comparing once more the minimum and the maximum values obtained for
the pure shear strength, Figure VI.20. This decrease is related with the observed daily
variations of the relative humidity, which has a significant influence on the curing
process and, therefore, on the bond strength of the interface.
It can be stated that the surface preparation has a significant influence on the bond
strength of concrete-to-concrete interfaces. Moreover, to achieve a higher bond it is
necessary to take into consideration the curing conditions.
A linear relationship was assumed between the pure shear strength and the roughness
parameter Maximum Valley Depth (Rv). The coefficients of correlation between both
– VI.22 –
parameters were of 0.94, 0.88 and 0.94 respectively for the L28, L56 and L84 series,
and of 0.82, 0.60 and 0.92 for respectively the E28, E56 and E84 series. The E28 and
E56 series, both presenting the lowest coefficient of correlation, are two of the three
exceptions, previously identified, where the pure shear strength decreased with the
increase of the surface roughness. Therefore, the observed lower coefficient of
correlation is justified.
This proved the existence of a very good linear correlation between the bond strength of
the concrete-to-concrete interface, considered as the pure shear strength, and the surface
roughness, assumed as the Maximum Valley Depth (Rv), for different types of surface
preparation.
Difference of Ages
From Figures VI.19 and VI.21 it can be observed that the pure shear strength increases,
for both curing conditions, with the increase of the difference of ages between the
substrate concrete and the added concrete layer. Three exceptions were identified for the
L series, one for the shot-blasted surface and two for the hand-scrubbed surface. In these
cases, the pure shear strength decreased instead of increasing with the difference of ages
between concrete parts.
As previously mentioned, and as it became clear from the numerical study also
described before, this result has to do with the slant shear test, namely, with the different
interface stress states caused by differential shrinkage, differential stiffness and
compressive loading at the test date, and therefore cannot be directly extrapolated for
other structural conditions.
In Figure VI.28 is presented the comparison between the pure shear strength obtained
for both curing conditions. The curing in the exterior led, as expected, to lower values of
the pure shear strength, with an average decrease of 1.12MPa (19%). This can be
justified, as previously mentioned, with the different daily variations of the relative
humidity, which has a significant influence on the curing of the concrete at early ages.
– VI.23 –
9
Pure shear strength (MPa) 8
7
6
5
4
3 Laboratory
Exterior
2
1
0
8
4
SH 6
LA 8
LA 6
4
SH 8
6
SA 4
6
SA 8
SA 6
SH 4
4
C5
C8
B2
S5
S8
B8
B2
B5
C2
B5
B8
S2
B2
B5
B8
H
H
W
W
LA
Surface preparation - Difference of age
Figure VI.28 - Comparison of the pure shear strength for both curing conditions.
Differential Stiffness and Failure Mode
The slant shear specimens presented both adhesive and cohesive failures, Figure VI.14,
while the splitting specimens only presented adhesive failures.
The developed numerical model, described in Section V.3.3, showed that for a given
level of differential stiffness, a change in the failure mode type of the slant shear
specimen can occur, namely from adhesive (interface debonding) to cohesive
(monolithic behaviour). Moreover, the experimental results corroborated the
observations of the numerical model.
It was observed an increase of the number of cohesive failures with the increase of the
difference between the Young modulus of the substrate concrete and the added concrete,
Figure VI.29. Nevertheless, it is recommended to develop future investigations to obtain
more data since this conclusion was reached with only six series of specimens and a
single concrete composition. The adoption of concrete mixtures with different strengths
can help to clarify the influence of this parameter.
The existence of a correlation between these two parameters, cohesive failure and
differential stiffness, is extremely important since it means that it is possible to change
the failure mode of a composite concrete member by designing the differential stiffness
between both concrete layers. It must be highlighted that the obtained experimental
results allows to state this for the particular stress state present in the slant shear test. It
– VI.24 –
must be investigated for other stress states.
2.5
2.0
Young modulus (GPa)
Difference between
1.5
Laboratory
1.0
Exterior
0.5
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Cohesive failures
Figure VI.29 - Differential stiffness versus cohesive failures.
Corroborating the findings of previous studies, conducted by Júlio et al. (2006) and
Austin et al. (1999), the developed numerical model showed that the increase of the
differential stiffness between concrete layers changes the stress distribution at the
interface and stress concentrations are observed at both ends. In fact, several slant shear
specimens exhibited broken corners, Figure VI.30, that are located precisely at both
ends of the interface.
Figure VI.30 - Stress concentrations in the slant shear specimen.
– VI.25 –
It was also observed an increase of the number of cohesive failures with the increase of
the surface roughness of the interface, presenting 3, 4, 9, 18 and 19 failures for the left
as-cast, wire-brushed, sand-blasted, shot-blasted and hand-scrubbed surfaces,
respectively. This behaviour suggests that the failure mode of the slant shear specimens
is not only controlled by the differential stiffness but also affected by the degree of the
surface roughness. Moreover, the relationship between the roughness parameter
Maximum Valley Depth (Rv) and the number of cohesive failures is presented in Figure
VI.31. Assuming a linear correlation between these two parameters it is obtained a
coefficient of correlation of 0.68, indicating a “poor” correlation between both
parameters.
1.4
Maximum Valley Depth, Rv (mm)
1.2
SAB
SHB
1.0
0.8
0.6
WB
0.4
0.2
LAC
0.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Cohesive failures
Figure VI.31 - Correlation between the Maximum Valley Depth and the number of cohesive
failures.
VI.7 Conclusions
The slant shear test proved to be suitable to predict the bond strength of concrete-to-
concrete interfaces, being sensitive to the surface preparation and to the differential
shrinkage and stiffness. The modified splitting test showed to be less sensitive to these
parameters. Moreover, the modified splitting test only presented adhesive failures while
adhesive and cohesive failures were observed for the slant shear specimens.
The surface roughness proved to have a significant influence on the bond strength of
concrete-to-concrete interfaces. A good linear correlation was identified between the
– VI.26 –
roughness parameter Maximum Valley Depth (Rv) and the pure shear strength of the
interface, presenting high coefficients of correlation (R2 > 0.88).
It was observed an increase of the pure shear strength, beyond 100%, by preparing the
interface surface with a method like hand-scrubbing, instead of leaving the surface left
as-cast. This increase was observed for the specimens cured under laboratory
conditions, presenting the specimens cured under exterior conditions lower increases of
the pure shear strength.
The curing conditions have a significant influence on the bond strength of the interface.
For the same surface preparation and difference of age between concrete layers, the
specimens cured under exterior conditions led to lower values of the pure shear strength,
with an average decrease of 1.12MPa (19%). This decrease is caused by the different
daily variations of the relative humidity, which has a significant influence on the curing
of the concrete at early ages.
Therefore, it can be stated that the surface preparation has a significant influence on the
bond strength of the interface but to achieve a higher bond it is also necessary to take
into consideration the curing conditions.
The experimental study showed that the pure shear strength increases, for both curing
conditions, with the increase of the difference of ages between the substrate concrete
and the added concrete layer and, therefore, with the increase of the differential
shrinkage.
The developed numerical study corroborated the results of the experimental study. The
unexpected variation of the pure shear strength with the difference of ages is related
with the stress state at the interface of the slant shear test, namely, by the combination of
the differential shrinkage, differential stiffness and compressive loading during testing.
For the particular case of the slant shear test, the differential shrinkage appears to have a
beneficial effect. This conclusion can not be directly extrapolated for other structural
conditions.
The numerical model also showed that for a given level of differential stiffness, a
change in the failure mode type of the slant shear specimen can occur, namely from
– VI.27 –
adhesive (interface debonding) to cohesive (monolithic behaviour). Moreover, the

experimental results corroborated the observations of the numerical model and it was
observed an increase of the number of cohesive failures with the increase of the
difference between the Young modulus of the substrate concrete and the added concrete.
The existence of a correlation between the failure mode type and differential stiffness is
extremely important since it means that, in some cases, it can be possible to change the
failure mode of a composite concrete member by designing the differential stiffness
between both concrete layers.
It must be highlighted that the conclusions obtained with the numerical and
experimental studies are valid for the stress state present in the slant shear test. It must
be investigated for other stress states.
– VI.28 –
Chapter VII
Conclusions, Design Guidelines
and Future Developments
VII.1 Conclusions
Literature Review on Shear-Friction, Design Expressions and Design Codes
The shear-friction theory is worldwide accepted to predict the behaviour of concrete-to-

concrete interfaces, subjected to longitudinal shear stresses. Moreover, it is adopted in
design codes of concrete structures, such as the CEB-FIP Model Code 1990 (1990); the
Eurocode 2 (2004); the BS 8110-1 (1997); the ACI 318 (2008); the CAN/CSA A23.3
(2004); the AASHTO LFRD Bridge Design Specifications (2007); the AASHTO
Standard Specifications for Highway Bridges (1996) and the PCI Design Handbook
(2004).
This theory was first proposed by Birkeland and Birkeland (1966) and since then
several researchers proposed modifications to increase its accuracy and to extend its
application to other situations. The most relevant contributions were made by: Mattock
and Hawkins (1972), with the inclusion of the cohesion; Loov (1978), with the inclusion
of the concrete strength; Walraven et al. (1987), presenting an innovative sphere model;
– VII.1 –
and Randl (1997), explicitly considering the contribution of cohesion, friction and
dowel action.
It is recognized today that the load transfer mechanism of shear forces between two
concrete layers is as considered by Randl (1997), although current design codes do not
consider all those mechanisms. Furthermore, different design expressions are proposed,
different parameters are considered and different values are assumed for the same
parameter. The coefficient of friction and the reinforcement crossing the interface are
parameters common to all design codes. Other parameters, such as the cohesion; the
normal stress at the interface; and the concrete strength, are only considered by some of
these.
Also common to all design codes is the classification of the roughness of the interface
surface. A visual inspection is implicitly proposed being the roughness classified in
accordance to predefined categories of surface finishing treatments. The coefficients of
cohesion and friction, present in design expressions, are linked to these surface finishing
treatments. No equipment or method is defined to prepare the surface. Moreover, no
method is specified to control and assess the effectiveness of the surface finishing
treatment. Although this methodology is easy and fast to perform, it is also subjective
since it depends on the technician opinion, being therefore subjected to human error.
Literature Review on Preparation and Characterization of Concrete Surfaces
The condition of the substrate surface of reinforced concrete composite members plays
an important role on the development of the bond strength at the concrete-to-concrete
interface. The unsound concrete and the laitance must be removed before applying a
concrete overlay and a certain degree of roughness is required for the substrate surface
in order to achieve an adequate bond.
The substrate surface can be prepared using different methods. Abrasive and light
removal techniques, such as hydrodemolition, sand-blasting and shot-blasting, are the
most adequate. Heavy removal techniques should not be used to avoid micro-cracking
of the concrete substrate. Chemical methods are prohibited for safety and environmental
reasons.
– VII.2 –
The visual inspection is useful as a first approach to verify is the interface surface
presents the desired roughness. To overcome the mentioned disadvantages of the visual
assessment of the roughness, the Sand Patch Test, defined by the standard ASTM E 965
(2001), can be used. Its main inconvenient is the limitation to horizontal floor surfaces.
To overcome the disadvantages of the Sand Patch Test, a technique based in the
processing of digital image, named of TDI method, was developed to assess the texture
of concrete surface, Santos et al. (2007). With this technique it was possible to obtain
texture profiles of the substrate surface and to compute texture parameters from these,
correlating the latter with the bond strength of the interface. Nevertheless, this technique
also presents some disadvantages since it is work intensive; partially destructive; and
needs to be processed at the laboratory.
Other methods can overcome the disadvantages of the Sand Patch Test and of the TDI
method, since these are easy to use, non-destructive and do not require contact with the
measured surface, such as the Circular Track Meter, defined by ASTM E 2157 (2001),
and the Digital Surface Roughness Meter, proposed by Maerz and Myers (2001). With
these methods it is possible to obtain 2D texture profiles of the substrate surface and to
compute texture parameters that describe the geometrical characteristics of the
measured surface, as with the TDI method.
Development of the 2D Laser Roughness Analyser
To overcome the disadvantages referred to, it was decided to develop a new measuring
device, the 2D laser roughness analyser (2D-LRA method). This presents all the
advantages and overcomes all the disadvantages of other roughness quantification
methods. It is fully customizable relatively to both hardware and software, which is an
important advantage in relation to the commercial devices previously referred to.
With the development of the 2D laser roughness analyser, it was possible to conduct
tests that corroborated the conclusions of previous studies, Júlio et al. (2004) and Santos
et al. (2007), namely: a) the texture of the substrate surface influences significantly the
bond strength of concrete-to-concrete interfaces; b) it is possible to obtain a texture
profile of the substrate surface and, from this, to determine texture parameters; and c) it
– VII.3 –
is possible to define linear correlations between some of these parameters and the bond
strength of the interface, both in shear and in tension, with very high coefficients of
correlation.
A comparison between this new method and the TDI method; the Sand Patch Test; and a
3D method based in the use of a laser scanner (3D-LS method) showed that the 2D-
LRA method is in fact the most adequate method since it combines accuracy with
portability. Moreover, it is non-destructive, contrary to the TDI method, and represents a
significant lower cost of acquisition and maintenance, in relation to the 3D-LS method.
The Sand Patch Test also showed not to be sensitive to different surface textures, mainly
if these are smooth or little rough.
The resolution of the laser sensor has influence on the values of the texture parameters.
The 10µm laser sensor is more adequate for smooth surfaces, such as the left as-cast,
while the 60µm laser sensor is sufficient for rough surfaces, such as those obtained by
wire-brushing, sand-blasting, shot-blasting and hand-scrubbing.
Primary parameters proved to be sufficient to characterize the concrete surface.

Therefore, filtering can be avoided and it is not necessary to compute the roughness and
waviness profiles and corresponding parameters.
Texture parameters, in particular those that are based in extreme values, such as the
maximum peak height and the maximum valley depth, are significantly influenced by
the existence of deep air holes and exposed aggregates in the measuring path. This
originates a high variability in results and investigations to increase the reliability of the
use of these parameters should be a priority. An upper bound must be defined to limit
the values assumed by the affected parameters, in particular, for the left as-cast surface.
The results obtained with the developed methodology for the characterization of
concrete surfaces suggest that it is expectable that this can be compulsory to use, in a
near future, due to the inclusion of roughness parameters in the codes design
expressions.
– VII.4 –
Numerical Modelling of Slant Shear Tests
The literature review on bond tests showed that the slant shear test is worldwide
adopted to assess the shear strength of concrete-to-concrete interfaces. Producing a
combined stress state of compression and shear at the interface, it is most sensitive to
surface roughness. Nevertheless, other factors also have influence, and have therefore to
be taken into account in results analysis.
The numerical modelling of slant shear tests proved that geometry, support conditions,
differential stiffness and differential shrinkage of specimens have a significant influence
on the stress distribution along the interface. This study allowed the optimization of the
specimens geometry, by reducing stress concentrations at the ends, by allowing to
produce specimens with different angles of the interface and simultaneously keeping a
common geometry, and by making possible to reuse the formwork fabricated
specifically for this end. A prism with a cross section of 150×150mm2 and a total height
of 450mm, corresponding to a shear plane oriented at 30 degrees with the vertical, was
adopted.
The numerical modelling also showed that both the differential stiffness and differential
shrinkage provoke a considerable change in the stress distribution along the interface,
both for shear and normal stresses, assuming in each case an increasingly pronounced
S-shaped form. It was observed that, for the same level of shear stress at the interface,
the corresponding normal stress increases with the increase of the differential stiffness
between both concrete parts. Therefore, it can be stated that the differential stiffness has
influence in the concrete-to-concrete bond strength. Moreover, changes in the failure
mode of the slant shear specimen, from adhesive to cohesive, can be observed also due
to the differential stiffness increase. Cracking can also appear at both ends of the slant
shear specimen due to differential shrinkage.
It can be stated that both the differential stiffness and differential shrinkage must be
taken into account in practice when designing concrete-to-concrete interfaces.
Assessment of the Bond Strength
The storage conditions of the concrete specimens, curing in the laboratory and curing in
– VII.5 –
the exterior of the laboratory, showed to be significantly different. The average values of
both temperature (≈ 18ºC) and relative humidity (≈ 70%) were very similar in both
situations, although the coefficient of variation for the second situation was
approximately the double (≈ 30%) of the one obtained for the first situation (≈ 16%),
due to higher daily variations of both parameters.
The prediction of the concrete shrinkage by the Eurocode 2 (2004) showed to be too
conservative, presenting significantly higher values than those experimentally observed.
The curing conditions showed to have a significant influence on the bond strength of the
interface. For the same surface preparation and difference of age between concrete
layers, the specimens cured under exterior conditions led to lower values of the pure
shear strength. It was observed an average decrease of 1.12MPa (19%). This variation is
justified with the different daily variations of the relative humidity, which has a
significant influence on the curing of the concrete at early ages.
Contrary to what was expected, the experimental study showed that the pure shear
strength increases, for both curing conditions, with the increase of the difference of ages
between the substrate concrete and the added concrete layer and, therefore, with the
increase of the differential shrinkage.
The developed numerical study corroborated and explained the results of the
experimental study. The unexpected variation of the pure shear strength with the
difference of ages is related with the stress state at the interface of the slant shear test,
namely, by the combination of the differential shrinkage, differential stiffness and
compressive loading during testing. For the particular case of the slant shear, the
differential shrinkage appears to have a beneficial effect. This conclusion can not be
directly extrapolated for other structural conditions.
The highest values of the bond strength were obtained in concrete specimens with the
interface prepared by hand-scrubbing. The interfaces prepared with this technique
showed that concrete hardening must be well controlled to achieve a roughened surface,
being this the major inconvenient of the technique. If the surface is scrubbed too soon,
the wanted effect will not be obtained because the concrete is still too fresh. On the
– VII.6 –
contrary, if it is scrubbed too late, the wanted effect will not be obtained either because
the concrete is already hardened.
A significant increase of the pure shear strength can be achieved, slightly beyond 100%,
by preparing the interface surface with a method like hand-scrubbing, instead of leaving
the surface left as-cast. This increase was observed for the specimens cured under
laboratory conditions. The specimens cured under exterior conditions presented a lower
increase of the pure shear strength.
The roughness parameter Maximum Valley Depth (Rv) was adopted to characterize the
surface texture and to be correlated with the bond strength of the concrete-to-concrete
interface. This parameter assumes the values of: 0.097mm, for the left as-cast surface;
0.451mm, for the wire-brushed surface; 1.055mm, for the sand-blasted surface;
1.122mm, for the shot-blasted surface; and 5.270mm, for the hand-scrubbed surface.
The surface roughness proved to have a significant influence on the bond strength of
concrete-to-concrete interfaces. A linear correlation was identified between the
roughness parameter Maximum Valley Depth (Rv) and the pure shear strength of the
interface, presenting high coefficients of correlation (R2 > 0.88). To avoid problems
related with the existence of strong surface irregularities, the Mean Valley Depth (Rvm)
can be selected instead of the Maximum Valley Depth (Rv).
Based on the influence of the storage conditions and surface roughness of the interface,
it can be stated that the surface preparation has a significant influence on the bond
strength of the interface but to achieve a higher bond it is also necessary to take into
consideration the curing conditions.
The slant shear test proved to be suitable to predict the bond strength of concrete-to-
concrete interfaces and, as mentioned before, showed to be sensitive to the surface
preparation and to the differential shrinkage and stiffness. The modified splitting test
showed to be less sensitive to these parameters. Moreover, the modified splitting test
only presented adhesive failures while adhesive and cohesive failures were observed for
the slant shear specimens.
It was also observed an increase of the number of cohesive failures with the increase of
– VII.7 –
the difference between the Young modulus of the substrate concrete and of the added
concrete, corroborating the conclusions of the numerical modelling previously referred
to. The existence of such correlation is extremely important since it means that it is
possible to change the failure mode of a composite concrete member by designing the
differential stiffness between both concrete layers. This conclusion is valid in the scope
of this research study, namely for the stress state present at the interface of the slant
shear specimen. Extrapolations to other stress states must be investigated.
VII.2 Design Guidelines
VII.2.1 Introduction
As stated by Virlogeux and Walraven (1999) and Walraven (2004), a good design code
must contain several attributes, namely, it should be a) coherent; b) open-minded;
c) transparent; d) simple; e) oriented to practice; and f) flexible.
According to those researchers, to be coherent a design code should be based in

scientific theories capable of describing the structural behaviour and the physical
phenomena involved. Therefore, empiric rules should be avoided, if possible, and
carefully used, if not. To be open-minded, it should accept different theories and design
principles without excluding others. It should be transparent and the final user should
understand the basic principles. It should be simple, covering almost all practical cases
but not all. Specific cases can and should be treated apart. It should be oriented to
practice, helping designers to better design. Finally, a design code must be flexible,
being capable of adapting to the evolution of materials and to the improvement of
technical procedures. New developments must be recognized and included as much as
possible.
A design code should also include two different approaches. The first approach should
be based in the definition of simple and practical design rules, naturally leading to fast,
conservative and robust structural design. The second approach, being an alternative to
the first, should be based in a more detailed analysis, leading to more economic
– VII.8 –
although more time consuming design.
Following these principles and adopting the theoretical basis of the current design
expression proposed by the Eurocode 2 (2004), in the following paragraphs a modified
design expression for the assessment of the longitudinal shear strength of the interface
between concretes cast at different times is presented. The main proposed modifications
include the prediction of the coefficients of cohesion and friction. The influence of
differential shrinkage and stiffness must be evaluated for each case and taken into
account in the prediction of the longitudinal shear strength.
VII.2.2 Background Information
VII.2.2.1 Surface Preparation and Characterization
The Maximum Valley Depth (Rv) was identified as the texture parameter that best
correlates with the bond strength of the concrete-to-concrete interface, both in shear and
in tension. An inconvenient of this parameter is that it is defined using extreme values
of the texture profile, namely the deepest valley and, therefore, it is highly influenced by
strong irregularities, such as deep air holes and very exposed aggregates.
Instead of using that parameter, it is preferable to adopt a texture parameter based on

average values, such as the Mean Valley Depth (Rvm). Following the procedure adopted
in section V.4.7, the Mean Valley Depth (Rvm) assumes the values of: 0.056mm, for the
left as-cast surface; 0.182mm, for the wire-brushed surface; 0.771mm, for the sand-
blasted surface; 0.809mm, for the shot-blasted surface; and 3.792mm, for the hand-
scrubbed surface. The coefficients of correlation between the pure shear strength and the
Mean Valley Depth (Rvm) assumed values between 0.82 and 0.96 for the considered
situations.
Previous investigations conducted by Zilch and Reinecke (2000) showed that the bond
strength of the concrete-to-concrete interface does not increase indefinitely with the
increase of the surface roughness. According to these authors, an upper bound of 1.0mm
was suggested for the mean depth of roughness since, beyond this value, the bond
– VII.9 –
strength of the interface is less influenced by the surface roughness.
The experimental study conducted during this research corroborates the observations
made by Zilch and Reinecke (2000). This influence can be observed by comparing the
Maximum Valley Depth (Rv), for the left as-cast, wire-brushed, sand-blasted, shot-
blasted and hand-scrubbed surfaces, to the pure shear strength of the concrete-to-
concrete interface. It is notorious that the bond strength, obtained with the specimens
prepared by hand-scrubbing, does not correspond to the value of the texture parameter.
Therefore, this type of surface must be treated apart since it is identical to an indented
surface and presents an non-uniform roughness.
VII.2.2.2 Coefficients of Cohesion and Friction
Next, the texture parameter Mean Valley Depth (Rvm) of each surface condition is
correlated with the coefficients of cohesion and friction, evaluated from the
experimental results.
Using the pure shear strength, the cohesion corresponding to each specimens series and
surface preparation method was calculated, Table VII.1. Then, adopting the average
values for the computed coefficient of cohesion and considering the shear and the
normal stresses at the interface of the slant shear specimens, the coefficient of friction
for each specimens series and surface preparation method was calculated, Table VII.2.
Table VII.1 - Coefficient of cohesion.

Surface preparation method
Series
LAC WB SAB SHB HS
L28 0.71 0.74 0.97 1.27 1.79
L56 0.68 0.80 0.89 1.15 1.37
L84 0.98 0.92 1.19 1.23 1.61
E28 0.59 0.70 0.68 0.85 0.93
E56 0.68 0.71 0.90 1.02 0.98
E84 0.79 0.83 0.97 1.08 1.30
AVG 0.74 0.78 0.93 1.10 1.33
STD 0.14 0.08 0.17 0.15 0.34
COV 18.5% 10.8% 17.7% 13.8% 25.4%
– VII.10 –
Table VII.2 - Coefficient of friction.

Surface preparation method
Series
LAC WB SAB SHB HS
L28 1.18 1.25 1.42 1.48 1.50
L56 1.09 1.30 1.39 1.46 1.45
L84 1.41 1.43 1.46 1.48 1.50
E28 1.10 1.15 1.14 1.24 1.26
E56 1.14 1.22 1.29 1.32 1.39
E84 1.30 1.31 1.34 1.38 1.44
AVG 1.21 1.28 1.34 1.39 1.42
STD 0.13 0.10 0.12 0.10 0.09
COV 10.5% 7.5% 8.7% 7.1% 6.5%
Correlating the Mean Valley Depth (Rvm) with the average values of the coefficients of
cohesion and friction for each surface preparation method, it is possible to predict both
coefficients by the means of the following expressions:
0.145
c=1.062 R vm (VII.1)
0.041
=1.366 Rvm (VII.2)
where c is the coefficient of cohesion; μ is the coefficient of friction; and Rvm is the
Mean Valley Depth of the primary profile given in millimetre.
A power function was adopted since it best adjusts to the experimental values, also
presenting the highest coefficient of correlation between the Mean Valley Depth and the
coefficients of cohesion and friction, 0.92 and 0.94, respectively. In Figure VII.1, the
correlation between the parameters referred to is presented.
– VII.11 –
1.6
SHB HS
1.4
WB HS
SAB
Coefficient of cohesion
1.2 LAC
SHB
1.0 Coeff. of cohesion
SAB
WB Power Regression
0.8
LAC for Coeff. of cohe-
0.6 sion
Coeff. of friction
0.4 Power Regression
for Coeff. of friction
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Mean Valley Depth, Rvm (mm)
Figure VII.1 - Correlation between the Mean Valley Depth (Rvm) and the coefficients of cohesion
and friction.
VII.2.2.3 Differential Shrinkage
The differential shrinkage showed to have a beneficial effect on the bond strength of the
concrete-to-concrete interface. This is true for the adopted bond tests (slant shear tests);
considered curing conditions; adopted materials; among other factors, but may not be
true in practice. Thus, the most important conclusion to draw from this study is that the
curing conditions, and therefore the differential shrinkage, have influence on the
concrete-to-concrete bond strength and must therefore be taken into account and
carefully analysed in each situation.
Cracking and interface debonding are two of the major problems associated to
differential shrinkage and can be successfully reduced by adopting an adequate design
procedure, i.e., by quantifying the influence of temperature and relative humidity and/or
by considering a concrete mixture design that includes an SRA (shrinkage reduction
admixture).
VII.2.2.4 Differential Stiffness
As in the case of differential shrinkage, differential stiffness between concrete parts also
showed to have influence on the behaviour of the concrete-to-concrete interface. It is
possible that changes can occur in the failure mode of the composite reinforced concrete
– VII.12 –
member when increasing the differential stiffness. This is of major significance in

composite structures and, therefore, it must considered in practice.
Based on the results obtained from the experimental study and numerical modelling, it
is recommended to adopt an added concrete layer with a higher Young modulus in
relation to the substrate concrete.
VII.2.2.5 Modified Design Equation
The design expression of Eurocode 2 (2004), in its initial form (Eq. VII.3), is given by:
 Rdi =c f ctd  n f yd  sin cos    0.5  f cd (VII.3)
where νRdi is the shear resistance at the concrete-to-concrete interface; c and µ are
factors that depend on the roughness of the interface; fctd is the design tensile strength of
concrete; σn is the external normal stress acting on the interface; ρ is the reinforcement
ratio; fyd is the design yield strength of the reinforcement; α is the angle between the
shear reinforcement and the shear plane; ν is a strength reduction factor; and fcd is the
design compressive strength of concrete.
The quantification methodology herein presented suggests that Eq. VII.3 should be
divided into two design expressions. The first design expression (Eq. VII.4) considers
cohesion only, i.e., chemical adhesion and aggregate interlock.
If the acting shear stress is lower or equal than the shear strength provided only by
cohesion, it is not necessary to provide shear reinforcement crossing the interface. In
this case, the shear strength of the concrete-to-concrete interface, νRdi,c, is given by:
 Rdi , c =c d f ctd  0.5 f cd (VII.4)
where cd is the design coefficient of cohesion, that depends on the roughness of the
interface; and fctd is the design tensile strength of concrete.
If the acting shear stress is higher than the shear strength provided only by cohesion, it
is necessary to provide shear reinforcement crossing the interface and a different design
– VII.13 –
expression (Eq. VII.5) must be used, considering the shear strength given by friction. In
this situation the shear strength of the concrete-to-concrete interface, νRdi,f, is given by:
 Rdi , f =d  n f yd   d sin cos    0.5  f cd (VII.5)
where µd is the design coefficient of friction, that depends on the roughness of the
interface; σn is the external normal stress acting on the interface; ρ is the reinforcement
ratio; fyd is the design yield strength of the reinforcement; α is the angle between the
shear reinforcement and the shear plane; ν is a strength reduction factor; and fcd is the
design compressive strength of concrete.
The design values of the coefficient of cohesion and friction are determined as proposed
in equations VII.1 and VII.2, by measuring the surface roughness and evaluating the
Mean Valley Depth (Rvm).
1.062 Rvm0.145
cd = (VII.6)
coh
1.366 R vm0.041
d = (VII.7)
 fr
where cd is the design coefficient of cohesion; γcoh is the partial safety factor for the
coefficient of cohesion, taken as 2.6; μd is the design coefficient of friction; γfr is the
partial safety factor for the coefficient of friction, taken as 1.2; and Rvm is the Mean
Valley Depth of the primary profile given in millimetre.
The partial safety factors γ were determined, according to Sedlacek (2007), adopting the
values of the coefficient of variation calculated for the coefficient of cohesion and
friction (Tables VII.1 and VII.2) and are given by:
−k⋅ 1−k⋅V R
= = (VII.8)
−⋅⋅ 1−⋅⋅V R
where µ is the average; k is a factor to obtain a target fractile of the Normal

Distribution, taken as 1.65 for the 5% fractile; σ is the standard deviation; α is a
weighting factor, taken as 0.8; β is the reliability factor; and VR is the coefficient of
– VII.14 –
variation.
The partial safety factors were determined for ULS, adopting a reliability index β of 3.8,
to which corresponds a probability of failure of 7.5×10-5. The maximum value of the
partial safety factor was adopted, namely 2.6 and 1.2 for the coefficient of cohesion and
coefficient of friction, respectively.
The contribution of the shear reinforcement crossing the interface does not explicitly
include the dowel action. Future investigations are needed to define its contribution.
VII.2.3 Proposed Draft for Section 6.2.5 of Eurocode 2 (2004)
An alternative draft to the current shear-friction provisions of Eurocode 2 (2004) is now

presented, taken into account the conclusions taken from the study described in the
present thesis. Parts presented with a double strike through are suggested to be removed
and new parts are included in red.
(start of transcription) “...
6.2.5 Shear at the interface between concrete cast at different times

(1) In addition to the requirements of 6.2.1 – 6.2.4 the shear stress at the interface
between concrete cast at different times should also satisfy the following:
v Edi v Rdi (6.23)
vEdi is the design value of the shear stress in the interface and is given by:
v Edi = V Ed /  z bi  (6.24)
where:
β is the ratio of the longitudinal force in the new concrete area and the
total longitudinal force either in the compression or tension zone, both
calculated for the section considered;
VEd is the transverse shear force;
z is the lever arm of composite section;
bi is the width of the interface (see Figure 6.8);
vRdi is the design shear resistance at the interface and is given by:
– VII.15 –
 Rdi , c =c d f ctd  0.5 f cd (6.25a)

when no reinforcement crossing the interface is provided;
 Rdi , f =d  n f yd  d sin cos    0.5  f cd (6.25b)
when reinforcement crossing the interface is provided.
where:
cd and μd are factors which depend on the roughness of the interface (see
(2));
fctd is as defined in 3.1.6 (2)P;
σn stress per unit area caused by the minimum external normal force
across the interface that can act simultaneously with the shear force,
positive for compression, such that σn < 0.6 fcd, and negative for
tension; When σn is tensile c fctd should be taken as 0.
ρ = As / Ai;
Figure 6.8: Examples of interfaces.

As is the area of reinforcement crossing the interface, including ordinary
shear reinforcement (if any), with adequate anchorage at both sides of
the interface;
Ai is the area of the joint;
α is defined in Figure 6.9, and should be limited by 45º ≤ α ≤ 90º;
ν is a strength reduction factor (see 6.2.2 (6)).
– VII.16 –
Figure 6.9: Indented construction joint.

(2) In the absence of more detailed information surfaces may be classified as very
smooth, smooth, rough or indented, with the following examples:
– Very smooth: a surface cast against steel, plastic or specially prepared wooden
moulds: c = 0.25 and μ = 0.5
– Smooth: a slipformed or extruded surface, or a free surface left without further
treatment after vibration: c = 0.35 and μ = 0.6
– Rough: a surface with at least 3mm roughness at about 40mm spacing, achieved
by raking, exposing of aggregate or other methods giving an equivalent
behaviour: c = 0.45 and μ = 0.7
– Indented: a surface with indentations complying with Figure 6.9: c = 0.50 and
μ = 0.9
(2) For uniform interface surfaces, the roughness must be measured with a minimum
accuracy of 10 micrometer; then, the Mean Valley Depth (Rvm) has to be
determined, taken as the average value obtained considering at least ten 2D
primary profiles.
Uniform interface surfaces can be considered as those obtained by cast against
steel, plastic or specially prepared wooden moulds; a slipformed or extruded
surface; a free surface left without further treatment after vibration; or surfaces
prepared by wire-brushing, sand-blasting, shot-blasting, water-blasting or other
equivalent methods.
The design value of the coefficients of cohesion and friction are predicted from
the Mean Valley Depth (Rvm) using the expressions:
0.145
1.062 Rvm
cd = (6.26)
coh
0.041
1.366 Rvm
d = (6.27)
 fr
where:
cd is the design coefficient of cohesion;
γcoh is the partial safety factor for cohesion and taken as 2.6;
μd is the design coefficient of friction;
– VII.17 –
γfr is the partial safety factor for friction and taken as 1.2;
Rvm is the Mean Valley Depth of the primary profile given in millimeter.
In the absence of more detailed information and for non-uniform interface
surfaces, such as those obtained by raking or presenting indentations complying
with Figure 6.9, the following values can be adopted: cd = 0.50 and μd = 0.90.
(3) The differential shrinkage between both concrete parts should be taken into
consideration in the design and evaluated on site for each specific case. The
influence of temperature should also be considered in design.
(4) The differential stiffness between both concrete parts should be taken into
consideration in the design. The Young modulus of the added concrete layer
should never be taken smaller than the correspondent value of the substrate
concrete.
(5) A stepped distribution of the transverse reinforcement may be used, as indicated
in Figure 6.10. Where the connection between the two different concretes is
ensured by reinforcement (beams with lattice girders), the steel contribution to vRdi
may be taken as the resultant of the forces taken from each of the diagonals
provided that 45º ≤ α ≤ 135º.
(6) The longitudinal shear resistance of grouted joints between slab or wall elements
may be calculated according to 6.2.5 (1). However in cases where the joint can be
significantly cracked, c should be taken as 0 for smooth and rough all joints and
0.5 for indented joints (see also 10.9.3 (12)).
(7) Under fatigue or dynamic loads, the values for c in 6.2.5 (1) should be halved. the
design shear resistance at the interface due to cohesion should not be considered.
Figure 6.10: Shear diagram representing the required interface reinforcement.
...” (end of transcription)
– VII.18 –
VII.3 Future Developments
Additional research studies are required to answer to several questions that came up
during the development of the present study.
Further investigations are needed to improve the reliability of results, in particular,

related with the high coefficients of variation observed for the texture parameters. It is
necessary to investigate the influence of other surface preparation methods on the bond
strength of concrete-to-concrete interfaces and to identify the variability of the
roughness degree obtained using the same preparation method, namely, due to the
exposition time; applied pressures; among other factors.
It must be highlighted that this research study was developed without steel
reinforcement. The influence of steel reinforcement bars or connectors crossing the
interface should be investigated. The dowel action must be isolated and assessed.
The use of a concrete overlay with a different density (LWAC) and strength (HPC) of
the substrate concrete presents several advantages and must be investigated in future
research studies. Being this a common solution in the strengthening of bridge decks it is
also of major interest to assess the behaviour of the concrete-to-concrete interface for
dynamic loads and fatigue, as well for fire situations.
– VII.19 –
– VII.20 –
References
AASHTO LFRD Bridge Design Specifications, American Association of State Highway

and Transportation Officials, 4th edition, SI units edition, 2007, ISBN 1-56051-354-3,
1526 p.
AASHTO Standard Specifications for Highway Bridges, American Association of State

Highway and Transportation Officials, 16th edition, 1996, ISBN 1-56051-040-6, 760 p.
Abu-Tair, A.I., Rigden, S.R. and Burley, E. - Testing the bond between repair materials
and concrete substrate, American Concrete Institute, ACI Materials Journal, Vol. 93, No.
6, pp. 53-58, November-December 1996.
Abu-Tair, A.I., Lavery, D., Nadjai, A., Rigden, S.R. and Ahmed, T.M.A. - A new method
for evaluating the surface roughness of concrete cut for repair or strengthening,
Elsevier, Construction and Building Materials, Vol. 14, No. 3, pp. 171-176, April 2000,
doi:10.1016/S0950-0618(00)00016-7.
ACI 318 - Building code requirements for structural concrete (ACI 318-08) and
commentary, Committee 318, American Concrete Institute, PO Box 9094, Farmington
Hills, MI 48333-9094, USA, 2008, 471 p.
ACI 555R - Removal and reuse of hardened concrete, Committee 555, American
Concrete Institute, PO Box 9094, Farmington Hills, MI 48333-9094, USA, 2001, 26 p.
Akazawa, T. - Tension test method for concretes, Bulletin No. 16, International
Association of Testing and Research Laboratories, pp. 11-23, Paris, November 1953.
– R.1 –
Al-Gahtani, A.S., Rasheeduzzafar and Al-Mussallam, A.A. - Performance of repair

materials exposed to fluctuation of temperature, American Society of Civil Engineers,
Journal of Materials in Civil Engineering, Vol. 7, No. 1, pp. 9-18, February 1995,
doi:10.1061/(ASCE)0899-1561(1995)7:1(9).
Ali, M.A. and White, R.N. - Enhanced contact model for shear friction of normal and
high-strength concrete, American Concrete Institute, ACI Structural Journal, Vol. 96,
No. 3, pp. 348-362, May-June 1999.
Anderson, A.R. - Composite designs in precast and cast-in-place concrete, Progressive

Architecture, Vol. 41, No. 9, pp. 172-179, September 1960.
ASTM C 496 - Standard test method for splitting tensile strength of cylindrical concrete
specimens, American Society for Testing Materials, 100 Barr Harbor Dr., West
Conshohocken, PA 19428, USA, 2004, doi:10.1520/C0496_C0496M-04E01.
ASTM C 882 - Standard test method for bond strength of epoxy-resin systems used
with concrete by slant shear, American Society for Testing Materials, 100 Barr Harbor
Dr., West Conshohocken, PA 19428, USA, 1999, doi:10.1520/C0882_C0882M-05E01.
ASTM C 1404 - Standard test method for bond strength of adhesive systems used with
concrete as measured by direct tension, American Society for Testing Materials, 100
Barr Harbor Dr., West Conshohocken, PA 19428, USA, 2003,
doi:10.1520/C1404_C1404M-98R03.
ASTM C 1583 - Standard test method for tensile strength of concrete surfaces and the
bond strength or tensile strength of concrete repair and overlay materials by direct
tension (Pull-off Method), American Society for Testing Materials, 100 Barr Harbor Dr.,
West Conshohocken, PA 19428, USA, 2004, doi:10.1520/C1583_C1583M-04E01.
ASTM D 4259 - Standard practice for abrading concrete, American Society for Testing
Materials, 100 Barr Harbor Dr., West Conshohocken, PA 19428, USA, 1999,
doi:10.1520/D4259-88R06.
ASTM D 4260 - Standard practice for acid etching concrete, American Society for
Testing Materials, 100 Barr Harbor Dr., West Conshohocken, PA 19428, USA, 1999,
doi:10.1520/D4260-05.
ASTM E 965 - Standard test method for measuring pavement macrotexture depth using
a volumetric technique, American Society for Testing Materials, 100 Barr Harbor Dr.,
West Conshohocken, PA 19428, USA, 2001, doi:10.1520/E0965-96R06.
ASTM E 2157 - Standard test method for measuring pavement macrotexture properties
using the circular track meter, American Society for Testing Materials, 100 Barr Harbor
Dr., West Conshohocken, PA 19428, USA, 2001, doi:10.1520/E2157-01R05.
– R.2 –
ASTM E 2380 - Standard test method for measuring pavement texture drainage using
an outflow meter, American Society for Testing Materials, 100 Barr Harbor Dr., West
Conshohocken, PA 19428, USA, 2005, doi:10.1520/E2380-05.
Austin, S.A. and Robins, P.J. - Development of patch test to study behaviour of shallow
concrete patch repairs, Thomas Telford, Magazine of Concrete Research, Vol. 45, No.
164, pp. 221-229, September 1993.
Austin, S., Robins, P. and Pan, Y. - Shear bond testing of concrete repairs, Elsevier,
Cement and Concrete Research, Vol. 29, No. 7, pp. 1067-1076, July 1999, doi:10.1016/
S0008-8846(99)00088-5.
Badoux, J.C. and Hulsbos, C.L. - Horizontal shear connection in composite concrete
beams under repeated loads, American Concrete Institute, Journal of the American
Concrete Institute, Vol. 64, No. 12, pp. 811-819, December 1967.
Bigerelle, M. and Iost, A. - Statistical artefacts in the determination of the fractal

dimension by the slit island method, Elsevier, Engineering Fracture Mechanics, Vol. 71,
No. 7, pp. 1081-1105, May 2004, doi:10.1016/S0013-7944(03)00136-X.
Birkeland, P.W. and Birkeland, H.W. - Connections in precast concrete construction,

American Concrete Institute, Journal of the American Concrete Institute, Vol. 63, No. 3,
pp. 345-368, March 1966.
Birkeland, H.W. - Precast and prestressed concrete, Class notes for course, University of
British Columbia, Spring 1968.
Bissonnette, B., Courard, L., Vaysburd, A.M. and Bélair, N. - Concrete removal
techniques. Influence on residual cracking and bond strength, American Concrete
Institute, Concrete International, Vol. 28, No. 12, pp. 49-55, December 2006.
Brundle, C.R., Evans Jr., C.A. and Wilson, S. - Encyclopedia of materials

characterization: surfaces, interfaces, thin films, Butterworth-Heinemann, Boston, 1992,
ISBN 0-7506-9168-9, 782 p.
BS 1134 - Assessment of surface texture. Part 1: Methods and instrumentation, British

Standard Institute, London, 1988.
BS 8110-1 - Structural use of concrete. Part 1: Code of practice for design and
construction, British Standard Institute, London, 1997.
BS EN 1542 - Products and systems for the protection and repair of concrete structures -
Test methods - Measurement of bond strength by pull-off, British Standard Institute,
London, 1999.
– R.3 –
Carneiro, F.L.L.B. and Barcellos, A. - Concrete tensile strength, Bulletin No. 13,
International Association of Testing and Research Laboratories for Materials and
Structures, pp. 97-123, Paris, March 1953.
CEB-FIP Model Code 1990 for concrete structures, Comité Euro-International du

Béton, Secretariat Permanent, Case Postale 88, CH-1015 Lausanne, Switzerland, 1990,
437 p.
Chen, P.W., Fu, X. and Chung, D.D.L. - Improving the bonding between old and new
concrete by adding carbon fibers to the new concrete, Pergamon, Cement and Concrete
Research, Vol. 25, No. 3, pp. 491-496, April 1995, doi:10.1016/0008-8846(95)00037-D.
Chen, Q., Yang, S. and Li, Z. - Surface roughness evaluation by using wavelets analysis,
Elsevier, Precision Engineering, Vol. 23, No. 3, pp. 209-212, July 1999,
doi:10.1016/S0141-6359(99)00013-6.
Chmielewska, B. - On the methods of bond strength measurements in concrete repair

systems, European Materials Research Society, Fall Meeting, Warsaw University of
Technology, Warsaw, September 2005, 11 p.
CIRSOC 201 - Argentinian design code for concrete structures, National Institute of
Industrial Technology, Buenos Aires, Argentinian Republic, July 2005, 642 p. (in
Spanish)
Clímaco, J.C.T.S. and Regan, P.E. - Evaluation of bond strength between old and new
concrete in structural repairs, Thomas Telford, Magazine of Concrete Research, Vol. 53,
No. 6, pp. 377-390, December 2001, doi:10.1680/macr.53.6.377.40800.
CAN/CSA A23.3 - Design of concrete structures - Structures design, Canadian

Standards Association, 178 Rexdale Boulevard, Rexdale, Ontario, M9W 1R3, 2004,
258 p.
Delatte, N.J., Williamson, M.S. and Fowler, D.W. - Bond strength development with
maturity of high-early-strength bonded concrete overlays, American Concrete Institute,
ACI Materials Journal, Vol. 97, No. 2, pp. 201-207, March-April 2000a.
Delatte Jr., N.J., Wade, D.M. and Fowler, D.W. - Laboratory and field testing of
concrete bond development for expedited bonded concrete overlays, American Concrete
Institute, ACI Materials Journal, Vol. 97, No. 3, pp. 272-280, May-June 2000b.
DIN 4762 - Surface roughness. Terminology, German Institute for Standardization,

1989.
DIN 4768 – Determination of roughness parameters Ra, Rz, Rmax by means of stylus
instruments. Terms, measuring conditions, German Institute for Standardization, 1990.
– R.4 –
EN 1992-1-1 - Eurocode 2 - Design of concrete structures - Part 1-1: General rules and
rules for buildings, European Committee for Standardization, 2004, 225 p. (with
corrigendum dated of 16 January 2008)
EN 12390-3 - Testing hardened concrete. Part 3: Compressive strength of test

specimens, Portuguese Institute for Quality, December 2003, 21 p. (in Portuguese).
EN 12390-6 - Testing hardened concrete. Part 6: Tensile splitting strength of test

specimens, Portuguese Institute for Quality, January 2004, 14 p. (in Portuguese).
EN 12615 - Products and systems for the protection and repair of concrete structures.
Test methods. Determination of slant shear strength, European Committee for
Standardization, 1999.
EN ISO 4287 - Geometrical Product Specifications (GPS) - Surface texture. Profile

method. Terms, definitions and surface texture parameters, International Organization
for Standardization, 1998, 25 p.
EN ISO 11562 - Geometrical Product Specifications (GPS) - Surface texture. Profile

method. Metrological characteristics of phase correct filters, Portuguese Institute for
Quality, 1997 (in Portuguese).
EN ISO 13565-1 - Geometrical Product Specifications (GPS) - Surface texture. Profile

method. Surface having stratified functional properties. Part 1: Filterin and general
measurement conditions, Portuguese Institute for Quality, 1997 (in Portuguese).
Flannery, C.M. and von Kiedrowski, H. - Effects of surface roughness on surface

acoustic wave propagation in semiconductor materials, Elsevier, Ultrasonics, Vol. 40,
No. 1-8, pp. 83-87, May 2002, doi:10.1016/S0041-624X(02)00095-1.
Gaston, J.R., Kriz, L.B. - Connections in precast concrete structures – scarf joints,
Precast/Prestressed Concrete Institute, PCI Journal, Vol. 9, No. 3, pp. 37-59, June 1964.
Gohnert, M. - Proposed theory to determine the horizontal shear between composite

precast and in situ concrete, Elsevier, Cement and Concrete Composites, Vol. 22, No. 6,
pp. 469-476, December 2000, doi:10.1016/S0958-9465(00)00050-0.
Gohnert, M. - Horizontal shear transfer across a roughened surface, Elsevier, Cement

and Concrete Composites, Vol. 25, No. 3, pp. 379-385, April 2003, doi:10.1016/S0958-
9465(02)00050-1.
Goodhew, P.J., Humphreys, J. and Beanland, R. - Electron microscopy and analysis, 3rd
edition, Taylor & Francis, 2001, ISBN 0-7484-0968-8, 264 p.
Goto, T., Miyakura, J., Umeda, K., Kadowaki, S. and Yanagi, K. - A robust spline filter
on the basis of L2-norm, Elsevier, Precision Engineering, Vol. 29, No. 2, pp. 157-161,
April 2005, doi:10.1016/j.precisioneng.2004.06.004.
– R.5 –
Hanson, N.W. - Precast-prestressed concrete bridges 2. Horizontal shear connections,

Development Department Bulletin D35, Portland Cement Association, Vol. 2, No. 2, pp.
38-58, May 1960.
Hanson, D.I. and Prowell, B.D. - Evaluation of circular texture meter for measuring
surface texture of pavements, NCAT Report 04-05, National Center for Asphalt
Technology, Auburn University, Auburn, Alabama, USA, September 2004, 26 p.
Hara, S., Tsukada, T. and Sasajima, K. - An in-line digital filtering algorithm for surface
roughness profiles, Elsevier, Precision Engineering, Vol. 22, No. 4, pp. 190-195,
October 1998, doi:10.1016/S0141-6359(98)00013-0.
Hermansen, B.R. and Cowan, J. - Modified shear-friction theory for bracket design,
pp. 55-60, February 1974.
Hofbeck, J.A., Ibrahim, I.O. and Mattock, A.H. - Shear transfer in reinforced concrete,
pp. 119-128, February 1969.
Hsu, T.T.C., Mau, S.T. and Chen, B. - Theory of shear transfer strength of reinforced
concrete, American Concrete Institute, ACI Structural Journal, Vol. 84, No. 2, pp. 149-
160, March-April 1987.
Hwang, S.J., Yu, H.W. and Lee, H.J. - Theory of interface shear capacity of reinforced
concrete, American Society of Civil Engineers, Journal of Structural Engineering, Vol.
126, No. 6, pp. 700-707, June 2000, doi:10.1061/(ASCE)0733-9445(2000)126:6(700).
ICRI (International Concrete Repair Institute) - Selecting and specifying concrete

surface preparation for sealers, coatings, and polymer overlays. Technical Guideline No.
03732, Des Plaines, IL, 1997, 41 p.
ISO 4287-1 - Surface roughness. Terminology. Part 1: Surface and its parameters,
International Organization for Standardization, 1984.
ISO/TS 16610-22 - Geometrical Product Specifications (GPS) - Filtration. Part 22:

Linear profile filters: Spline filters, International Organization for Standardization,
2006, 15 p.
Issa, M.A., Islam, M.S. and Chudovsky, A. - Fractal dimension - a measure of fracture
roughness and toughness of concrete, Pergamon, Engineering Fracture Mechanics, Vol.
70, No. 1, pp. 125-137, January 2003, doi:10.1016/S0013-7944(02)00019-X.
Júlio, E.N.B.S. - Influence of the interface on the behaviour of columns strengthened by

reinforced concrete jacketing, PhD thesis, University of Coimbra, Coimbra, 2001, 274
p. (in Portuguese).
– R.6 –
Júlio, E.N.B.S., Branco, F.A.B. and Silva, V.D. - Concrete-to-concrete bond strength.
Influence of the roughness of the substrate surface, Elsevier, Construction and Building
Materials, Vol. 18, No. 9, pp. 675-681, November 2004,
doi:10.1016/j.conbuildmat.2004.04.023.
Júlio, E.N.B.S., Branco, F.A., Silva, V.D. and Lourenço, J.F. - Influence of added
concrete compressive strength on adhesion to an existing concrete substrate, Elsevier,
Building and Environment, Vol. 41, No. 12, pp. 1934-1939, December 2006,
doi:10.1016/j.buildenv.2005.06.023.
Kahn, L.F. and Mitchell, A.D. - Shear friction tests with high-strength concrete,
American Concrete Institute, ACI Structural Journal, Vol. 99, No. 1, pp. 98-103,
January-February 2002.
Kamada, T. and Li, V.C. - The effects of surface preparation on the fracture behavior of
ECC/concrete repair system, Elsevier, Cement & Concrete Composites, Vol. 22, No. 6,
pp. 423-431, December 2000, doi:10.1016/S0958-9465(00)00042-1.
Kriegh, J.D. - Arizona slant shear test: a method to determine epoxy bond strength,
pp. 372-373, 1976.
Krystek, M. - Form filtering by splines, Elsevier, Measurement, Vol. 18, No. 1, pp. 9-15,
May 1996a, doi:10.1016/0263-2241(96)00039-5.
Krystek, M. - Discrete L-spline filtering in roudness measurements, Elsevier,

Measurement, Vol. 18, No. 2, pp. 129-138, June 1996b, doi:10.1016/S0263-
2241(96)00051-6.
Krystek, M. - A fast Gauss filtering algorithm for roughness measurements, Elsevier,

Precision Engineering, Vol. 19, No. 2-3, pp. 198-200, October-November 1996c,
doi:10.1016/S0141-6359(96)00025-6.
Kunieda, M., Kurihara, N., Uchida, Y. and Rokugo, K. - Application of tension

softening diagrams to evaluation of bond properties at concrete interfaces, Pergamon,
Engineering Fracture Mechanics, Vol. 65, No. 2-3, pp. 299-315, January 2000,
doi:10.1016/S0013-7944(99)00125-3.
Lee, H. and Ahn, K. - A prototype of digital photogrammetric algorithm for estimating

roughness of rock surface, Springer, Geosciences Journal, Vol. 8, No. 3, pp. 333-341,
September 2004, doi:10.1007/BF02910253.
Li, S., Geissert, D.G. and Frantz, G.C. - Durability and bond of high performance
concrete and repaired Portland cement concrete, Joint Highway Research Advisory
Council, Project JHR 97-257, University of Connecticut, 232 p., USA, June 1997.
– R.7 –
Li, H., Jiang, X. and Li, Z. - Robust estimation in Gaussian filtering for engineering
surface characterization, Elsevier, Precision Engineering, Vol. 28, No. 2, pp. 186-193,
April 2004, doi:10.1016/j.precisioneng.2003.10.004.
Lin, I.J. and Chen, Y.L. - Shear transfer across a crack in reinforced high strength
concrete, Proceedings 2nd East Asia-Pacific Conference on Structural Engineering &
Construction, Chiang Mai, Thailand, pp. 505-510, January 1989.
Lingadurai, K. and Shunmugam, M.S. - Metrological characteristics of wavelet filter

used for engineering surfaces, Elsevier, Measurement, Vol. 39, No. 7, pp. 575-584,
August 2006, doi:10.1016/j.measurement.2006.02.003.
LNEC E 398, Concretes - Determination of shrinkage and expansion, National

Laboratory of Civil Engineering, 1993 (in Portuguese).
Loov, R.E. - Design of precast connections, Paper presented at a seminar organized by

Compa International Pte, Ltd, 8 p., Singapore, September 1978.
Loov, R.E. and Patnaik, A.K. - Horizontal shear strength of composite concrete beams
with a rough interface, Precast/Prestressed Concrete Institute, PCI Journal, Vol. 39, No.
1, pp. 48-69, January-February 1994.
Lourenço, J., Júlio, E. and Maranha, P. - Lighweight aggregates concretes, APEB, 2006,
ISBN 972-9071-30-6, 196 p. (in Portuguese).
LUSAS 14.1, Finite Element Analysis Ltd., Forge House, 66 High Street, Kingston
upon Thames, Surrey, KT1 1HN, United Kingdom, 2008, http://www.lusas.com.
Maerz, N.H., Franklin, J.A. and Bennett, C.P. - Joint roughness measurement using
shadow profilometry, Pergamon, International Journal of Rock Mechanics and Mining
Sciences & Geomechanics Abstracts, Vol. 27, No. 5, pp. 329-343, 1990,
doi:10.1016/0148-9062(90)92708-M.
Maerz, N. and Myers, J. - Development of a prototype imaging concrete roughness

measurement device, Center for Infrastructure Engineering Studies, University of
Missouri-Rolla, Report Number 00-20, USA, April 2001, 27 p.
Mansur, M.A., Vinayagam, T. and Tan, K.H. - Shear transfer across a crack in
reinforced high-strength concrete, American Society of Civil Engineers, Journal of
Materials in Civil Engineering, Vol. 20, No. 4, pp. 294-302, April 2008, doi:10.1061/
(ASCE)0899-1561(2008)20:4(294).
Mast, R.F. - Auxiliary reinforcement in precast concrete connections, American Society

of Civil Engineers, Journal of the Structural Division, Vol. 94, No. ST6, pp. 1485-1504,
June 1968.
– R.8 –
Mattock, A.H. and Kaar, P.H. - Precast-prestressed concrete bridges, 4. Shear tests of
continuous girders, Journal of the PCA R&D Laboratories, Vol. 3, No. 1, pp. 19-46,
January 1961.
Mattock, A.H. and Hawkins, N.M. - Shear transfer in reinforced concrete – recent
research, Precast/Prestressed Concrete Institute, PCI Journal, Vol. 17, No. 2, pp. 55-75,
March-April 1972.
Mattock, A.H. - Shear transfer in concrete having reinforcement at an angle to the shear
plane, American Concrete Institute, Special Publication 42-2, pp. 17-42, January 1974.
Mattock, A.H., Johal, L. and Chow, H.C. - Shear transfer in reinforced concrete with
moment or tension acting across the shear plane, Precast/Prestressed Concrete Institute,
PCI Journal, Vol. 20, No. 4, pp. 76-93, July-August 1975.
Mattock, A.H., Li, W.K. and Wang, T.C. - Shear transfer in lightweight reinforced
concrete, Precast/Prestressed Concrete Institute, PCI Journal, Vol. 21, No. 1, pp. 20-39,
Mattock, A.H. - Cyclic shear transfer and type of interface, American Society of Civil
Engineers, Journal of the Structural Division, Vol. 107, No. ST10, pp. 1945-1964,
October 1981.
Mattock, A.H. - Reader comments of paper Influence of concrete strength and load
history on the shear friction capacity of concrete members, published in PCI Journal,
Vol. 32, No. 1, pp. 66-84, January-February 1987, by Walraven et al.,
Precast/Prestressed Concrete Institute, PCI Journal, Vol. 33, No. 1, pp. 165-166,
Mattock, A.H. - Reader comments of paper Horizontal shear strength of composite

concrete beams with a rough interface, published in PCI Journal, Vol. 39, No. 1, pp. 48-
69, January-February 1994, by Loov, R.E. et al., Precast/Prestressed Concrete Institute,
PCI Journal, Vol. 39, No. 5, pp. 106-108, September-October 1994.
Mattock, A.H. - Shear friction and high-strength concrete, American Concrete Institute,
ACI Structural Journal, Vol. 98, No. 1, pp. 50-59, January-February 2001.
Mau, S. and Hsu, T. - Reader comments of paper Influence of concrete strength and load
history on the shear friction capacity of concrete members, published in PCI Journal,
Vol. 32, No. 1, pp. 66-84, January-February 1987, by Walraven et al.,
Precast/Prestressed Concrete Institute, PCI Journal, Vol. 33, No. 1, pp. 166-168,
Mitri, F.G., Kinnick, R.R., Greenleaf, J.F. and Fatemi, M. - Continuous-wave ultrasound
reflectometry for surface roughness imaging applications, Elsevier, Ultrasonics, Vol. 49,
No. 1, pp. 10-14, January 2009, doi:10.1016/j.ultras.2008.06.011.
– R.9 –
Momayez, A., Ramezanianpour, A.A., Rajaie, H. and Ehsani, M.R. - Bi-surface shear
test for evaluating bond between existing and new concrete, American Concrete
Institute, ACI Materials Journal, Vol. 101, No. 2, pp. 99-106, March-April 2004.
Momayez, A., Ehsani, M.R., Ramezanianpour, A.A. and Rajaie, H. - Comparison of

methods for evaluating bond strength between concrete substrate and repair materials,
Elsevier, Cement and Concrete Research, Vol. 35, No. 4, pp. 748-757, April 2005,
doi:10.1016/j.cemconres.2004.05.027.
Montoya, P.J., Meseguer, A.G. and Cabré, F.M. - Hormigón armado, 14th edition,
Editorial Gustavo Gili, Barcelona, 2000, ISBN 84-252-1825-X, 443 p.
Mummery, L. - Surface texture analysis - the handbook, Hommelwerke GmbH,

Thyssen, 1992, 106 p.
Naderi, M. - Friction-transfer test for the assessment of in situ strength and adhesion of
cementitious materials, Elsevier, Construction and Building Materials, Vol. 19, No. 6,
pp. 454-459, July 2005, doi:10.1016/j.conbuildmat.2004.07.018.
Numada, M., Nomura, T., Yanagi, K., Kamiya, K. and Tashiro, H. - High-order spline
filter and ideal low-pass filter at the limit of its order, Elsevier, Precision Engineering,
Vol. 31, No. 3, pp. 234-242, July 2007, doi:10.1016/j.precisioneng.2006.09.002.
Ohama, Y., Demura, K., Nagao, H. and Ogi, T. - Adhesion of polymer-modified mortars
to ordinary cement mortar by different test methods, International Symposium on
Adhesion between polymers and concrete: bonding, protection, repair, RILEM
Technical Committee 52 - Resin Adherence to Concrete and Laboratoire Central des
Ponts et Chausées, Paris, Palais des Congrés, Aix-en-Provence, France, 16-19
September, 1986.
Ozkul, T. - Apparatus for measuring surface roughness, Patent number WO2006111795,

26 October 2006.
Papanicolaou, C.G. and Triantafillou, T.C. - Shear transfer capacity along pumice
aggregate concrete and high-performance concrete interfaces, RILEM, Materials and
Structures, Vol. 35, No. 4, pp. 237-245, May 2002, doi:10.1007/BF02533085.
Patnaik, A.H. - Horizontal shear strength of composite concrete beams with a rough
interface, PhD thesis, Department of Civil Engineering, University of Calgary, Calgary,
Canada, 1992, 263 p.
Patnaik, A.H. - Reader comments of paper Evaluation of ACI 318-95 shear-friction

provisions, published in ACI Structural Journal, Vol. 96, No. 4, pp. 473-483, July-
August 1999, by Valluvan et al., American Concrete Institute, ACI Structural Journal,
Vol. 97, No. 3, pp. 525-526, May-June 2000.
– R.10 –
Patnaik, A.H. - Behavior of composite concrete beams with smooth interface, American
Society of Civil Engineers, Journal of Structural Engineering, Vol. 127, No. 4, pp. 359-
366, April 2001, doi:10.1061/(ASCE)0733-9445(2001)127:4(359).
PCI Design Handbook, Precast/Prestressed Concrete Institute, 6th edition, 2004, ISBN
0-937040-71-1, 736 p.
Raja, J., Muralikrishnan, B. and Fu, S. - Recent advances in separation of roughness,

waviness and form, Elsevier, Precision Engineering, Vol. 26, No. 2, pp. 222-235, April
2002, doi:10.1016/S0141-6359(02)00103-4.
Randl, N. - Investigations on transfer of forces between old and new concrete at

different joint roughness, PhD thesis, University of Innsbruck, Austria, 1997, 379 p. (in
German)
Raths, C.H. - Reader comments of paper Design proposals for reinforced concrete
corbels, published in PCI Journal, Vol. 21, No. 3, pp. 18-42, May-June 1976, by
Mattock, A., Precast/Prestressed Concrete Institute, PCI Journal, Vol. 22, No. 2, pp. 93-
98, March-April 1977.
Ray, I., Davalos, J.F. and Luo, S. - Interface evaluations of overlay-concrete bi-layer
composites by a direct shear test method, Elsevier, Cement & Concrete Composites,
Vol. 27, No. 3, pp. 339-347, March 2005, doi:10.1016/j.cemconcomp.2004.02.048.
Saemann, J.C. and Washa, G.W. - Horizontal shear connections between precast beams
and cast-in-place slabs, American Concrete Institute, Journal of the American Concrete
Institute, Vol. 61, No. 11, pp. 1383-1409, November 1964.
Santos, P.M.D. - Influence of the surface roughness of the interface concrete/concrete in

its longitudinal shear resistance, MSc Thesis, University of Coimbra, 2005, 202 p. (in
Portuguese).
Santos, P., Júlio, E. and Silva, V.D. - Correlation between concrete-to-concrete bond
strength and the roughness of the substrate surface, Elsevier, Construction and Building
Materials, Vol. 21, No. 8, pp. 1688-1695, August 2007,
doi:10.1016/j.conbuildmat.2006.05.044.
Santos, P. and Júlio, E. - Development of a laser roughness analyser to predict in situ the
bond strength of concrete-to-concrete interfaces, Thomas Telford, Magazine of Concrete
Research, Vol. 60, No. 5, pp. 329-337, June 2008, doi:10.1680/macr.2007.00024.
Santos, P. and Júlio, E. - The effect of filtering on the texture assessment of concrete
surfaces, American Concrete Institute, ACI Materials Journal, June 2009a (accepted).
Santos, P. and Júlio, E. - Influence of differential shrinkage and stiffness on concrete-to-

concrete bond strength, RILEM, Materials and Structures, June 2009b (accepted).
– R.11 –
Santos, P. and Júlio, E. - Comparison of methods for the texture assessment of concrete
surfaces, American Concrete Institute, ACI Materials Journal, March 2009c (submitted).
Sedlacek, G. and Kraus, O. - Use of safety factors for the design of steel structures
according to the Eurocodes, Elsevier, Engineering Failure Analysis, Vol. 14, No. 3, pp.
434-441, April 2007, doi: 10.1016/j.engfailanal.2005.08.002.
Seker, D.Z. and Tavil, A.Ü. - Evaluation of exterior building surface roughness degrees
by photogrammetric methods, Elsevier, Building and Environment, Vol. 31, No. 4, pp.
393-398, July 1996, doi:10.1016/0360-1323(95)00051-8.
Shaikh, A.F. - Proposed revisions to shear-friction provisions, Precast/Prestressed

Concrete Institute, PCI Journal, Vol. 23, No. 2, pp. 12-21, March-April 1978.
Silfwerbrand, J. - Improving concrete bond in repaired bridge decks, American

Concrete Institute, Concrete International, Vol. 12, No. 9, pp. 61-66, September 1990.
Silva, V.D. - Mechanics and strength of materials, Springer, 2006, ISBN 978-3-540-
25131-6, 529 p.
Smith, G.T. - How rough are precision surfaces & how round is round?, Inaugural
Professorial Lecture, Southampton Institute, June 2003, 29 p.
Sukmana, D.D. and Ihara, I. - Novel noncontact method for characterizing surface
profile by air-coupled ultrasound scattering, AZoM, Advances in Technology of
Materials and Materials Processing Journal, Vol. 8, No. 2, pp. 248-255, November 2006,
doi:10.2240/azojomo0229.
Talbot, C., Pigeon, M., Beaupré, D. and Morgan D.R. - Influence of surface preparation
on long-term bonding of shotcrete, American Concrete Institute, ACI Materials Journal,
Vol. 91, No. 6, pp. 560-566, November-December 1994.
Taylor, J.R. - An introduction to error analysis - the study of uncertainties in physical

measurements, 2nd edition, University Science Books, Sausalito, California, USA,
1997, ISBN 0-935702-42-3, 349 p.
Thaulow, S. - Tensile splitting test and high strength concrete test cylinders, American
Concrete Institute, Journal of the American Concrete Institute, Vol. 53, No. 1, pp. 699-
706, January 1957.
Thomas, T.R. - Rough surfaces, Longman Higher Education, 1982, ISBN 0-582-46816-
7, 288 p.
Tschegg, E.K., Ingruber, M., Surberg, C.H. and Münger, F. - Factors influencing
fracture behavior of old-new concrete bonds, American Concrete Institute, ACI
Materials Journal, Vol. 97, No. 4, pp. 447-453, July-August 2000.
– R.12 –
Tsoukantas, S.G. and Tassios, T.P. - Shear resistance of connections between reinforced
concrete linear precast elements, American Concrete Institute, ACI Structural Journal,
Vol. 86, No. 3, pp. 242-249, May-June 1989.
Unal, M. and Unver, B. - Characterization of rock joint surface degradation under shear
loads, Elsevier, International Journal of Rock Mechanics and Mining Sciences, Vol. 43,
No. 3, pp. 145-150, May 2004, doi:10.1016/j.ijrmms.2004.03.033.
Unal, M., Yakar, M. and Yildiz, F. - Discontinuity surface roughness measurement

techniques and the evaluation of digital photogrammetric method, XXth Congress of the
International Society for Photogrammetry and Remote Sensing on Geo-Imagery
Bridging Continents, Proceedings of Commission 3, Istanbul, Turkey, 12-23 July 2004.
Valluvan, R., Kreger, M.E. and Jirsa, J.O. - Evaluation of ACI 318-95 shear-friction
provisions, American Concrete Institute, ACI Structural Journal, Vol. 96, No. 4, pp. 473-
483, July-August 1999.
Vecchio, F.J. and Collins, M.P. - The modified compression-field theory for reinforced
concrete elements subjected to shear, American Concrete Institute, Journal of the
American Concrete Institute, Vol. 83, No. 2, pp. 219-231, March-April 1986.
Virlogeux, M. and Walraven, J. - The development of an international codification for

structural concrete with the CEB-FIP model codes, Concrete Model Code for Asia,
IABSE Colloquium, Phuket, 1999.
Volk, R. and Ville, J.F. - Filters for contour measurement, Elsevier, Wear, Vol. 264, No.
5-6, pp. 469-473, March 2008.
Wall, J.S., Shrive, N.G. and Gamble, B.R. - Testing of bond between fresh and hardened
concrete, International Symposium on Adhesion between polymers and concrete:
bonding, protection, repair, RILEM Technical Committee 52 - Resin Adherence to
Concrete and Laboratoire Central des Ponts et Chausées, Paris, Palais des Congrés, Aix-
en-Provence, France, 16-19 September, 1986.
Walraven, J.C. - Fundamental analysis of aggregate interlock, American Society of Civil

Engineers, Journal of the Structural Division, Vol. 107, No. ST11, pp. 2245-2270,
November 1981.
Walraven, J., Frénay, J. and Pruijssers, A. - Influence of concrete strength and load
history on the shear friction capacity of concrete members, Precast/Prestressed Concrete
Institute, PCI Journal, Vol. 32, No. 1, pp. 66-84, January-February 1987.
Walraven, J.C. - Thinking about codes, Thomas Telford, Structural Concrete, Vol. 5, No.
3, pp. 93-100, 2004.
– R.13 –
Wambold, J.C. and Henry, J.J. - NASA Wallops Tire/Runway friction workshops: 1993-
2002, CDRM Inc., 1911 East College Avenue, PO Box 1277, State College, PA, USA
16804, September 2002, 155 p.
Whitehouse, D. - Surfaces and their measurement, Hermes Penton Science, 2002, ISBN
1-9039-9601-5, 425 p.
Yuan, Y.B., Qiang, X.F., Song, J.F. and Vorburger. T.V. - A fast algorithm for
determining the Gaussian filtered mean line in surface metrology, Elsevier, Precision
Engineering, Vol. 24, No. 1, pp. 62-69, January 2000, doi:10.1016/S0141-
6359(99)00031-8.
Zilch, K. and Reinecke, R. - Capacity of shear joints between high-strength precast

elements and normal-strength cast-in-place decks, FIB International Symposium on
High Performance Concrete, Orlando, USA, 25-27 September 2000.
– R.14 –
Appendix
Table A.1 - Concrete compressive strength (L series).

Failure load Compressive strength
Concrete Specimen in
Series compression AVG STD COV
layer designation
(kN) (MPa) (MPa) (%)
SC-1-S-L28 1712.70
Substrate SC-2-S-L28 1908.00 79.30 4.78 6.03
SC-3-S-L28 1732.05
L28
SC-1-A-L28 1548.23
Added SC-2-A-L28 1464.75 66.38 2.11 3.18
SC-3-A-L28 1467.45
SC-1-S-L56 1956.52
Substrate SC-2-S-L56 1858.07 85.96 3.01 3.50
SC-3-S-L56 1987.71
L56
SC-1-A-L56 1790.60
Added SC-2-A-L56 1831.20 80.49 0.90 1.12
SC-3-A-L56 1810.99
SC-1-S-L84 1927.10
Substrate SC-2-S-L84 1995.65 86.44 1.98 2.29
SC-3-S-L84 1912.00
L84
SC-1-A-L84 1537.98
Added SC-2-A-L84 1660.27 72.63 3.83 5.27
SC-3-A-L84 1704.10
– A.2 –
Table A.2 - Concrete compressive strength (E series).

Failure load Compressive strength
Concrete Specimen in
Series compression AVG STD COV
layer designation
SC-1-S-E28 1915.73
Substrate SC-2-S-E28 1714.50 78.90 5.43 6.88
SC-3-S-E28 1695.28
E28
SC-1-A-E28 1554.06
Added SC-2-A-E28 1529.64 68.26 0.72 1.05
SC-3-A-E28 1523.66
SC-1-S-E56 1722.64
Substrate SC-2-S-E56 1730.48 77.62 1.53 1.98
SC-3-S-E56 1785.99
E56
SC-1-A-E56 1649.67
Added SC-2-A-E56 1649.48 71.14 3.77 5.30
SC-3-A-E56 1502.57
SC-1-S-E84 1846.10
Substrate SC-2-S-E84 1844.24 81.87 0.24 0.30
SC-3-S-E84 1835.80
E84
SC-1-A-E84 1472.37
Added SC-2-A-E84 1580.83 69.92 4.32 6.17
SC-3-A-E84 1666.15
– A.3 –
Table A.3 - Experimental shrinkage strain.

Shrinkage strain (×10-6)
Age Curing in the laboratory Curing in the exterior
(days) SHR-1-L SHR-2-L SHR-1-E SHR-2-E
AVG AVG
Left Right Left Right Left Right Left Right
0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
7 13.3 6.7 13.3 23.3 14.2 16.7 56.7 40.0 26.7 35.0
14 36.7 16.7 36.7 43.3 33.3 30.0 110.0 83.3 53.3 69.2
21 50.0 23.3 36.7 56.7 41.7 40.0 141.7 113.3 68.3 90.8
28 50.0 23.3 36.7 56.7 41.7 50.0 173.3 143.3 83.3 112.5
35 58.3 28.3 43.3 61.7 47.9 56.7 186.7 156.7 93.3 123.3
42 66.7 33.3 50.0 66.7 54.2 80.0 206.7 176.7 116.7 145.0
49 66.7 33.3 56.7 83.3 60.0 66.7 200.0 180.0 103.3 137.5
56 86.7 33.3 56.7 83.3 65.0 73.3 216.7 190.0 116.7 149.2
63 86.7 36.7 56.7 83.3 65.8 83.3 230.0 203.3 126.7 160.8
70 70.0 40.0 56.7 73.3 60.0 93.3 243.3 216.7 136.7 172.5
77 80.0 43.3 66.7 83.3 68.3 106.7 256.7 226.7 153.3 185.8
84 80.0 43.3 66.7 83.3 68.3 113.3 263.3 236.7 156.7 192.5
91 83.3 43.3 66.7 83.3 69.2 120.0 273.3 243.3 156.7 198.3
98 83.3 43.3 66.7 83.3 69.2 126.7 280.0 253.3 160.0 205.0
105 83.3 43.3 66.7 83.3 69.2 130.0 296.7 276.7 173.3 219.2
112 83.3 43.3 66.7 83.3 69.2 133.3 316.7 283.3 190.0 230.8
– A.4 –
Table A.4 - Shrinkage strain for the specimens cured in the laboratory.
Shrinkage strain
Age
Drying Autogenous Total
(days)
β (t, ts) εcd (t)×10-6 βas (t) εca (t)×10-6 εcs (t)×10-6
0 0.000 0.0 0.000 0.0 0.0
7 0.212 75.7 0.411 50.7 126.4
14 0.350 125.0 0.527 65.0 189.9
21 0.447 159.5 0.600 74.0 233.5
28 0.519 185.1 0.653 80.5 265.6
35 0.574 204.8 0.694 85.5 290.4
42 0.618 220.5 0.726 89.6 310.0
49 0.654 233.2 0.753 92.9 326.1
56 0.683 243.8 0.776 95.7 339.5
63 0.708 252.7 0.796 98.1 350.8
70 0.729 260.3 0.812 100.2 360.4
77 0.748 266.8 0.827 102.0 368.8
84 0.764 272.6 0.840 103.6 376.1
91 0.778 277.6 0.852 105.0 382.6
98 0.790 282.1 0.862 106.3 388.3
105 0.802 286.1 0.871 107.4 393.5
112 0.812 289.7 0.880 108.4 398.1
– A.5 –
Table A.5 - Shrinkage strain for the specimens cured in the exterior.
Shrinkage strain
Age
Drying Autogenous Total
(days)
β (t, ts) εcd (t)×10-6 βas (t) εca (t)×10-6 εcs (t)×10-6
0 0.000 0.0 0.000 0.0 0.0
7 0.212 68.3 0.411 59.3 127.6
14 0.350 112.7 0.527 76.1 188.7
21 0.447 143.8 0.600 86.7 230.5
28 0.519 166.9 0.653 94.3 261.2
35 0.574 184.7 0.694 100.2 284.8
42 0.618 198.8 0.726 104.9 303.7
49 0.654 210.2 0.753 108.8 319.0
56 0.683 219.8 0.776 112.1 331.8
63 0.708 227.8 0.796 114.9 342.7
70 0.729 234.6 0.812 117.3 351.9
77 0.748 240.6 0.827 119.4 360.0
84 0.764 245.7 0.840 121.3 367.0
91 0.778 250.3 0.852 123.0 373.2
98 0.790 254.3 0.862 124.5 378.8
105 0.802 257.9 0.871 125.8 383.7
112 0.812 261.1 0.880 127.0 388.2
– A.6 –
Table A.6 - Failure load of the slant shear specimens (L28 Series).
Failure load Shear stress at the interface
Surface Specimen in Failure
compression AVG STD COV
preparation designation type
LAC-1-L28 884.93 Adhesive
Left as-cast LAC-3-L28 411.75 10.47 4.01 38.32 Adhesive
WB-1-L28 691.20 Adhesive
Wire-brushing WB-3-L28 715.28 11.96 1.88 15.71 Adhesive
SAB-1-L28 1133.78 Adhesive
Sand-blasting SAB-3-L28 604.58 18.52 4.65 25.13 Adhesive
SHB-1-L28 1032.08 Adhesive
SHB-2-L28 1382.63 Cohesive
Shot-blasting SHB-3-L28 1082.03 22.77 3.33 14.62 Adhesive
HS-1-L28 1390.95 Cohesive
Hand-scrubbing HS-3-L28 1331.55 25.39 3.73 14.68 Cohesive
– A.7 –
HS-1-L56 1208.38 Adhesive
Hand-scrubbing HS-3-L56 1218.97 23.39 0.48 2.07 Adhesive
– A.8 –
LAC-5-L84 1272.22 Cohesive
SAB-1-L84 1234.95 Cohesive
SAB-2-L84 1301.73 Cohesive
Shot-blasting SHB-3-L84 1303.50 25.00 1.07 4.28 Cohesive
Hand-scrubbing HS-3-L84 1384.80 26.86 0.96 3.56 Cohesive
– A.9 –
Table A.9 - Failure load of the slant shear specimens (E28 Series).
LAC-1-E28 531.38 Adhesive
Left as-cast LAC-3-E28 438.84 8.74 1.35 15.39 Adhesive
WB-1-E28 590.38 Adhesive
WB-2-E28 326.07 Cohesive
Wire-brushing WB-3-E28 643.30 10.09 2.87 28.43 Cohesive
SAB-1-E28 497.96 Cohesive
SAB-2-E28 549.68 Adhesive
Sand-blasting SAB-3-E28 361.90 10.57 1.63 15.44 Cohesive
SHB-1-E28 669.43 Cohesive
Shot-blasting SHB-3-E28 764.01 11.99 2.76 23.01 Cohesive
HS-1-E28 681.47 Cohesive
Hand-scrubbing HS-3-E28 594.65 12.45 0.84 6.73 Cohesive
– A.10 –
LAC-2-E56 508.67 Cohesive
Left as-cast LAC-3-E56 453.17 10.27 1.16 11.32 Cohesive
Wire-brushing WB-3-E56 629.59 11.91 2.41 20.23 Adhesive
Shot-blasting SHB-3-E56 797.28 14.79 1.28 8.65 Cohesive
HS-1-E56 1019.70 Adhesive
Hand-scrubbing HS-3-E56 925.85 17.73 4.03 22.72 Adhesive
– A.11 –
LAC-4-E84 - Broken
SHB-2-E84 784.83 Adhesive
Shot-blasting SHB-3-E84 1140.42 17.00 3.14 18.44 Adhesive
Hand-scrubbing HS-3-E84 1045.59 20.79 1.52 7.29 Cohesive
– A.12 –
Table A.12 - Failure load of the splitting specimens (L28 Series).

Failure load Tension stress at the interface
preparation designation compression AVG STD COV
type
– A.13 –

type
– A.14 –

type
– A.15 –
Table A.15 - Failure load of the splitting specimens (E28 Series).

type
Sand-blasting SAB-3-E28 64.88 1.67 0.13 7.57 Adhesive
– A.16 –

type
– A.17 –

type
– A.18 –

PHD Thesis PSantos PDF

Uploaded by

Copyright:

Available Formats

PHD Thesis PSantos PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PHD Thesis PSantos PDF

Uploaded by

Copyright:

Available Formats

Assessment of the Shear Strength

between Concrete Layers

Coimbra, July, 2009

Coimbra, Julho, 2009

A qualitative evaluation of the surface roughness is currently adopted by all design

An experimental study was conducted to investigate the influence of several parameters

Changes to the current shear-friction provisions of Eurocode 2 (2004) (Section 6.2.5)

A aderência na interface entre camadas de betão de diferentes idades é importante para

Os actuais códigos de dimensionamento de estruturas de betão armado, tais como o

A avaliação qualitativa da rugosidade da superfície é actualmente adoptada por todos os

Estes códigos de dimensionamento não têm em consideração as condições de cura do

Por todas estas razões, as actuais expressões de dimensionamento necessitam de

longitudinal de interfaces betão-betão. Este trabalho de investigação pretende contribuir

Assim sendo, definiram-se três objectivos principais: desenvolver um método in situ

Foi desenvolvido um novo equipamento óptico de medição para caracterizar a textura

Foi desenvolvido um estudo experimental para investigar a influência de diversos

São propostas modificações às actuais especificações do corte-atrito do Eurocódigo 2

Concrete, Shear Strength, Interface, Texture, Roughness, Differential Stiffness,

This page was intentionally left blank

To MAPREL - Empresa de Pavimentos e Materiais Pré-Esforçados Lda, SIKA Portugal,

Pedro Miguel Duarte dos Santos

This page was intentionally left blank

This page was intentionally left blank

I.3 Research Significance............................................................................................I.4

I.4 Thesis Organization...............................................................................................I.5

CHAPTER II - LONGITUDINAL SHEAR STRENGTH

II.2 Load Transfer Mechanisms..................................................................................II.2

II.3 Quantification Proposals......................................................................................II.5

II.3.2 Design Expressions................................................................................II.7

II.3.2.1 Anderson (1960).........................................................................II.8

II.3.2.2 Hanson (1960).............................................................................II.9

II.3.2.3 Mattock and Kaar (1961)............................................................II.9

II.3.2.4 Saemann and Washa (1964)........................................................II.9

II.3.2.5 Gaston and Kriz (1964).............................................................II.10

II.3.2.6 Birkeland and Birkeland (1966)................................................II.11

II.3.2.7 Badoux and Hulsbos (1967)......................................................II.12

II.3.2.8 Birkeland (1968).......................................................................II.13

II.3.2.9 Mast (1968)...............................................................................II.14

II.3.2.10 Hofbeck, Ibrahim and Mattock (1969).....................................II.14

II.3.2.11 Mattock and Hawkins (1972)....................................................II.15

II.3.2.12 Mattock (1974)..........................................................................II.16

II.3.2.13 Hermansen and Cowan (1974)..................................................II.17

II.3.2.14 Mattock, Johal and Chow (1975)..............................................II.17

II.3.2.15 Mattock, Li and Wang (1976)...................................................II.18

II.3.2.16 Raths (1977)..............................................................................II.18

II.3.2.17 Shaikh (1978)............................................................................II.19

II.3.2.18 Loov (1978)..............................................................................II.20

II.3.2.19 Mattock (1981)..........................................................................II.21

II.3.2.20 Vecchio and Collins (1986).......................................................II.21

II.3.2.21 Walraven, Frénay and Pruijssers (1987)...................................II.22

II.3.2.22 Mattock (1988)..........................................................................II.24

II.3.2.23 Mau and Hsu (1988).................................................................II.25

II.3.2.24 Lin and Chen (1989).................................................................II.25

II.3.2.25 Tsoukantas and Tassios (1989)..................................................II.26

II.3.2.26 Patnaik (1992)...........................................................................II.27

II.3.2.27 Loov and Patnaik (1994)...........................................................II.27

II.3.2.28 Mattock (1994)..........................................................................II.28

II.3.2.29 Randl (1997).............................................................................II.29

II.3.2.30 Ali and White (1999)................................................................II.31