J. Inst. Eng. India Ser.

A (June 2020) 101(2):247–257

https://doi.org/10.1007/s40030-020-00434-8

ORIGINAL CONTRIBUTION

Carbonation-Induced Corrosion: A Brief Review on Prediction

Models
Ramesh Gopal1 • Bhaskar Sangoju1

Received: 25 May 2019 / Accepted: 24 January 2020 / Published online: 3 February 2020
 The Institution of Engineers (India) 2020

Abstract This paper presents a brief review on carbona- Introduction

tion depth prediction models. The prediction models are
classified into different ways, namely empirical and phys- Deterioration due to corrosion is one of the major prob-
ical models; and probabilistic and deterministic models. In lems faced by reinforced concrete structures as it affects
this paper, empirical and physical models for natural car- the durability [1]. Industrial buildings, tunnels and struc-
bonation depth prediction have been reviewed. As a para- tures away from coastal regions are more prone to car-
metric study, carbonation depths are predicted using bonation-induced corrosion than chloride-induced
empirical and physical models for buildings within CSIR- corrosion, and it affects the long-term durability of the
Structural Engineering Research Centre, Chennai campus. structure. Carbonation causes deterioration of concrete
From the parametric study, Fib model code, Ekolu’s model, structures and induces corrosion by decreasing the pH in
Kokubu’s and Hakkinen’s models showed good correlation reinforced concrete structures affecting the service life of
(R2 = 0.94) with the measured carbonation depths. Car- the structures [2–4]. Carbonation is given importance as
bonation depths are predicted using Ekolu’s model for there is a gradual increase in the CO2 concentration in the
ordinary, standard and high-strength concretes for exposure atmosphere due to vehicular emissions, industrial effluents
periods of 25, 50 and 100 years at various key locations and use of machineries [5]. Concrete is deemed to be
such as rural areas, sea coast regions, urban settings, durable if minimum water-to-cement ratio and minimum
industrial areas and tunnels. Carbonation depth is higher in grade of concrete as specified in standards are used in
tunnels than coastal region due to increased CO2 concen- field for different exposure conditions. But, there is no
tration and low relative humidity. Therefore, the cover actual method in field to check these parameters. Further,
depth required to be provided for corrosion against car- durability of concrete is affected by environmental factors
bonation is more for tunnels, followed by industrial sites such as relative humidity, CO2 concentrations and shelter
and urban areas. conditions which are not considered in the durability
design. The relative humidity and atmospheric CO2 react
Keywords Reinforced concrete  Durability  with the hydration products, particularly Ca(OH)2, caus-
Carbonation  Corrosion  Prediction models ing a reduction in pH levels and in turn leads to cover
concrete neutralization and the active corrosion is initi-
ated. The hydration process and the mechanism of car-
bonation are depicted in Eqs. 1.1–1.4. Tricalcium and
dicalcium silicates present in cement react with water to
& Ramesh Gopal produce calcium silicate hydrate (C–S–H) and calcium
grm@serc.res.in
hydroxide as their products. During carbonation reaction,
Bhaskar Sangoju the calcium hydroxide in the aqueous state reacts with
bhaskar@serc.res.in
carbon dioxide and calcium carbonate precipitate is
1
CSIR- Structural Engineering Research Centre, CSIR formed [5–8].
Campus, Taramani, Chennai, India

123
248 J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257

Hydration reactions predicted to new and existing buildings provided the data
2ð3CaO  SiO2 Þ þ 6H2 O ! ð3CaO  2SiO2  3H2 OÞ on compressive strength and curing period are known.
þ 3CaðOHÞ2 From the compressive strength of the concrete, the slump,
cement content, fly ash replacement ratios are directly
ð1:1Þ accounted into the model [9]. The environmental factors
ð2CaO  SiO2 Þ þ 4H2 O ! ð3CaO  2SiO2  3H2 OÞ such as CO2 concentrations, relative humidity and tem-
þ CaðOHÞ2 ð1:2Þ perature influences are not considered in this model.
 pffiffiffiffiffiffipffi
Carbonation reactions D ¼ 10  1:6 fck t
ð1:5Þ
CaðOHÞ2 ðsÞ $ Ca2þ ðaqÞ þ 2OH ðaqÞ ð1:3Þ ðOutdoor exposure)
 pffiffiffiffiffiffipffi
Ca2þ ðaqÞ þ 2OH ðaqÞ þ CO2 ðgÞ ! CaCO3 ðsÞ þ H2 O D ¼ 17  2:2 fck t
ð1:4Þ ð1:6Þ
ðIndoor exposure)
The aim of this research is to find a suitable model that has
the highest correlation with the observed carbonation where D = carbonation depth (mm); fck = characteristic
depths. The carbonation depth at different regions for dif- compressive strength of concrete at 28 days (MPa); t =
ferent grades of concrete, exposure conditions and age of time of exposure (years).
exposure is predicted using the chosen model, and thus the
knowledge of minimum cover depth to be adopted can be Hakkinen’s Model [10]
gained for different environmental conditions.
This is also a strength-based empirical model, and this can
Carbonation Depth Prediction Models also be used for new and existing structures with the
available data on air entrainment. The model proposed by
To predict the carbonation depth, various types of models Hakkinen’s used to predict the depth of carbonation using
are available and it is divided into different categories such Eqs. 1.7 and 1.8. The essential input parameter in the
as empirical model, statistical model, numerical and sim- model is air content coefficient from the concrete, and the
ulation models. Empirical models are proposed based on variables a and b assume different values for various types
the exposure of structures to natural environment. Empir- of cement are given in Table 1. The air content of concrete
ical model also grouped based on permeability, diffusion, can be determined as per ASTM C 173 [11]. This model is
strength and composition. Statistical model is purely based limited to specified replacement level of fly ash, silica fume
on data generation through regression of multiple variables and slag.
pffi
and standard mathematical functions. Numerical models D¼K t ð1:7Þ
work based on computational approach through software
program using hybrid physico-chemical equations and K ¼ Cenv Cair aðfck þ 8Þb ð1:8Þ
transport process happening in concrete. Simulation mod- where K = carbonation rate factor (mm/year0.5); Cenv-
els are based on artificial neural network algorithm with an = environmental coefficient; Cair = air content coefficient;
input and output parameters, to generate the data set by a and b are constants depending on the type of cement;
training to get the reliable output. The depth of carbonation fck = characteristic compressive strength of concrete
in reinforced concrete structures can be found out without (MPa).
actual measurements with prediction models. Out of many
models available in the literature, five models are chosen Parrott’s Model [12]
based on the availability of the input parameters. These
models are briefly described in the following sections. Parrott’s model predicts the carbonation depth on new and
existing structure based on air permeability coefficient by
Kokubu and Nagataki’s Model [9] using the expression in Eq. 1.9. The air permeability test
mimics the actual transport mechanism in concrete, and the
This is purely strength-based empirical model obtained model assumes that the CO2 diffusion in concrete is rep-
from regression analysis to predict the carbonation depth resented by the air permeability coefficient [12]. Air per-
using minimum available parameters and defined by the meability coefficient values are taken as 1 9 10-16 for all
expression in Eqs. 1.5 and 1.6. This model can be used for structures according to Torrent air permeability coefficient
ordinary portland cement as well as for cement containing values for normal quality concrete. The classification of
fly ash. Using this model, the carbonation depth can be concrete quality according to Torrent air permeability

123
J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257 249

Table 1 Values of a and b for different binders in Hakkinen’s model and execution transfer parameter obtained by using number
[10] of curing days. As per Indian practice, the concrete struc-
Binder a b tures are subjected to a minimum curing period of 7 days.
For this reason, the number of curing days is taken as
Portland cement 1800 - 1.7
7 days for all the buildings. The CO2 concentration, taken
Portland cement ? fly ash 28% 360 - 1.2 as a fixed quantity in Fib 34, is 0.00082 kg/m3 which is
Portland cement ? silica fume 9% 400 - 1.2 equal to 360–380 ppm of CO2 in the atmosphere. But, CO2
Portland cement ? blast furnace slag 70% 360 - 1.2 concentration in the present study is about 500 ppm which
is equal to 0.000982 kg/m3 which is used in the calcula-
tions. 50% of the rainy days are assumed to have wind
Table 2 Concrete quality classification according to Torrent air
permeability coefficient [13] direction that can wet the surface. This assumption is valid
as wind blows more in the eastern direction, Chennai is in
Torrent air permeability (kT 9 10-16 m2) Quality of concrete
the east coast line of the country, and all the cores are
\ 0.01 Very good extracted in the eastern direction only. Time of wetting
0.01–0.1 Good (ToW) is 0.143 as given in Table 5. The predicted car-
0.1–1.0 Normal bonation depth is listed in Table 4.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 
1.0–10 Poor pffiffiffiffiffiffi
[ 10 Very poor Xc;d ðtSL Þ ¼ 2ke;d  kc;d  R1 NAC;0  Cs;d 
 tSL  W ðtÞ

ð2:0Þ
coefficient is given in Table 2. The effect of curing time rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  ffi
pffiffiffiffiffiffi
and type of cement used are considered as least significant Xc;d ðtSL Þ ¼ 2ke;d  kc;d  R1 ACC;0;k þ a_ t;d  Cs;d  tSL
in the model.  W ðt Þ
0:4 tn
D ¼ ak ð1:9Þ ð2:1Þ
c0:5
where Xc,d = design value of the carbonation depth at the
where D = carbonation depth (mm); k = air permeability of
time tSL (mm). tSL = design service life (years). ke,d = de-
cover concrete and it depends on the relative humidity (in
sign value of the environmental function based on relative
units of 10-16 m2); n = power exponent based on the rel-
humidity. kt,d = execution transfer parameter for different
ative humidity; c = calcium oxide content in the hydrated
curing periods. kt,d = design value of the regression
cement matrix of the cover concrete (kg/m3); a = 64,
parameter = 1.25. R1
ACC;0;k = inverse effective carbonation
coefficient on the basis of the available published data.
resistance of concrete [(mm2/years)/(kg/m3)]. a€ = partial
safety factor = 1.5. a_ t;d = design value of the error
Fib Model Code
term = 315.5. Cs,d = design value of the CO2; concentra-
tion = 0.00082 (kg/m3) [for 360–380 ppm]. W(t) =
Fib model uses inverse natural carbonation relationship
weather function.
using characteristic compressive strength of concrete [14].
To predict the carbonation depth, most of the influencing
Ekolu’s Model [16]
factors are considered in this model [15]. Fib model can be
used for carbonation depth prediction for natural and
This model is based on time-dependent strength growth of
accelerated carbonation by using Eqs. 2.0 and 2.1. Envi-
concrete characteristic compressive strength, defined by the
ronmental factor is calculated using the relative humidity

Table 3 Input parameters for different models

Prediction model fck RH CO2 Shelter conditions/ k Curing Age of
rainfall period exposure

Kokubu and Nagataki’s model 4 4

Hakkinen’s model 4 4 4
Parrott’s model 4 4 4
Fib model code for service life design 4 4 4 4 4 4
Ekolu’s model 4 4 4 4 4
fck = characteristic compressive strength of concrete, RH = relative humidity, [CO2] = CO2, concentration, k = permeability of concrete

123
250 J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257

expression in Eq. (2.2). The input parameters considered in the concrete cores are extracted in unsheltered outdoor
this model are relative humidity, shelter conditions, varied exposure conditions.
CO2 levels, time-dependent strength growth of compres-
sive strength. This model can be used for different types of Carbonation Depth Based on Prediction Models
binders as well which is effected into the conductance
factor, g. But, a shortcoming of this model is that it cannot Environmental Conditions of the Selected Buildings
be applied to concretes having compressive strength less
than 20 MPa [16]. In order to predict the carbonation depth using the selected
 g pffi models, the environmental and other parameters are
dðf ;tÞ ¼ eh  es  eco  cem FcðtÞ  t ð2:2Þ
required. The parameters are collected through the avail-
where dðf ;tÞ = carbonation depth (mm); eh = environmental able information and published literature. To determine the
factor for relative humidity; es = environmental factor for carbonation depth in buildings, concrete cores are to be
shelter conditions. eco = environmental factor for varied extracted. The buildings considered are about four kilo-
CO2 concentration. FcðtÞ = time-dependent strength growth metres away from the seashore and are within the city at
function; cem and g are cement factors for carbonation CSIR-Structural Engineering Research Centre (at coordi-
conductance. nates 12.9862 N, 80.2489 E). In 2015, CO2 concentration
The summary of the input parameters to be used for the in the atmosphere surpassed 400 ppm, at the rate of
selected models is presented in Table 3. increase of 2 ppm per year (http://www.climatecentral.org).
In general, the level of CO2 in cities varies from
455–511 ppm [17]. Thus, the average CO2 concentration
Experimental Investigation within the CSIR campus is considered as 500 ppm. The
optimum relative humidity required for carbonation to take
Carbonation Depth in Concrete from Core Samples place is 50–70%. Data regarding relative humidity obtained
from Chennai International Airport for the past few years
The buildings chosen for the present investigation are (2014–2018) indicate that the average relative humidity has
having varying concrete strength and age of the structure not fallen below 50% except very few days [18]. Gener-
and with SCMs such as fly ash and slag. AML, SSL and ally, carbonation progression is more when relative
WEL building constructed 38, 31 and 27 years back, and humidity is between 50 to 70%, therefore, the average
all the structural members were concreted with M20 grade annual relative humidity is taken as 60% [19]. Rainfall data
concrete. And the samples were extracted from an for Chennai region in 2017 show that the area experienced
unsheltered location. ACTEL building is one the oldest rainfall for 76 days, out of which 54 days had rainfall
building in CSIR-SERC campus and the whole building is above 2.5 mm [18]. Rainfall recorded above 2.5 mm is
constructed with the movable formwork portal frame considered as significant in this study according to the
technique. The grade of concrete during that time is M15, models chosen. The relative humidity, CO2 concentration,
and the age of the building is 52 years. SHML building is precipitation and time of wetness data used in the present
completely a prefab building with waffle-type slab, and the are listed in Table 5. In addition to the above, wherever
age of the building is 44 years. The concrete grade for required, the grade of concrete mentioned in Table 4 is
beams, columns and slabs is M25 with 30% fly ash used in the prediction of carbonation depth.
replacement. The concrete road (CR) laid very recently in
CSIR campus is also M20 grade with 30% fly ash Prediction of Carbonation Depth Based on Different
replacement. Models
In this investigation, core samples of 25 mm diameter
and about 50 mm long were extracted from different The carbonation depths are predicted based on the avail-
buildings and the carbonation depth is evaluated by able input parameters and are presented in Table 6. The
spraying phenolphthalein indicator. Concrete in the car- input parameters required for estimation of carbonation
bonated region is colourless when the indicator is sprayed, depth are already presented in Table 3. From the measured
and the concrete turns violet red in the un-carbonated carbonation depth readings, it can be seen that the car-
region. Carbonation depth is measured at three locations bonation depth is higher for ACTEL building as it is the
along the circumference of the core extracted with a ver- oldest building in the campus. CR has the least carbonation
nier scale, and the average is reported. The average car- depth as it is exposed to carbonation only for 2 years. Thus,
bonation depth along with the grade of concrete and type of it is evident that carbonation depth increases with the
binder used in the buildings are presented in Table 4. All increase in the age of the building. ASTAR has low car-
bonation depth due to the use of higher grade of concrete,

123
J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257 251

Table 4 Measured carbonation depth in different buildings within CSIR-SERC

Sl. Name and age of the Grade of Type of cement Measured carbonation depth Average Remarks cores sprayed with
nos. building in years concrete (mm) carbonation depth phenolphthalein indicator
(mm)
Location Location Location
1 2 3

1. AML (38) M20 OPC 23.04 20.16 20.02 21.07

2. SSL (31) M20 OPC 18.78 16.28 16.34 17.14

3. WEL (27) M20 OPC 15.50 11.94 12.72 13.39

4. ACTEL (52) M15 OPC 34.66 33.54 31.82 33.34

5. SHML (44) M25 OPC with 30% fly 18.62 19.54 14.56 17.57
ash replacement

123
252 J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257

Table 4 continued
Sl. Name and age of the Grade of Type of cement Measured carbonation depth Average Remarks cores sprayed with
nos. building in years concrete (mm) carbonation depth phenolphthalein indicator
(mm)
Location Location Location
1 2 3

6. CL (46) M15 OPC 24.12 24.80 24.72 24.55

7. ASTAR (12) M30 OPC with 50% 6.58 4.80 10.00 5.56
slag replacement

8. CR (2) M20 OPC with 30% fly 4.80 4.46 4.48 4.58
ash replacement

AML, Advanced Materials Laboratory; SSL, Steel Structures Research Facility; WEL, Wind Engineering Laboratory; ACTEL, Advanced
Concrete Testing and Evaluation Laboratory; SHML, Structural Health Monitoring Laboratory; CL, Chemistry Laboratory; ASTAR, Advanced
Seismic Testing and Research Laboratory; CR, Concrete Road

as a buffer. But, when SCM added concrete is exposed to

Table 5 Environmental data used in the study carbonation for longer duration, they are more advanta-
Parameter used Value geous as they completely seal the pores due to pozzolanic
activity which decreases the rate of carbonation [20]. When
Relative humidity 60%
comparing the measured average carbonation depths of
CO2 concentration 500 ppm
SHML and CL, the CO2 penetration rate is lower (0.399/
Number of rainy days (2017) 76 days out of which 54 days
year) in the case of buildings with SCMs (fly ash) when
have significant rainfall
compared to that of the buildings with OPC alone (0.53/
Time of wetness (annual frequency of 0.143
days with significant rainfall) year).

Comparison Between Measured and Predicted

as carbonation depth progresses at a slower rate with the Carbonation Depths
increase in the strength of the concrete. Although SHML
and CL have almost the same age of exposure, the car- A comparison between the measured and predicted car-
bonation depth is more in CL. The lower carbonation depth bonation depths is made using the correlation coefficients.
in SHML could be due to the higher grade and addition of Correlation coefficient R2 for each model is depicted in
supplementary cementitious materials (SCMs). Addition of Fig. 1. It can be seen that all the selected models except
SCMs accelerates carbonation only in the early period due Parrott’s model show good correlation (R2 = 0.94) with the
to lower quantity of CaO content in them which cannot act measured values.

123
J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257 253

Table 6 Predicted carbonation depths using different models

Sl. nos. Name and age of the Predicted carbonation depth (mm)
building in years
Kokubu and Hakkinen’s Parrott’s Fib model code for Ekolu’s
Nagataki’s model model model service life design model

1. AML (38) 21.33 19.23 18.60 15.33 17.32

2. SSL (31) 19.26 17.37 16.80 13.84 16.17
3. WEL (27) 17.98 16.21 15.67 12.91 15.40
4. ACTEL (52) 31.20 31.43 21.75 24.25 –
5. SHML (44) 17.90 – 28.30 13.04 17.29
6. CL (46) 29.32 29.56 20.46 22.80 –
7. ASTAR (12) 6.97 – 12.15 5.63 7.57
8. CR (2) 4.90 4.42 4.27 3.52 6.70
AML, Advanced Materials Laboratory; SSL, Steel Structures Research Facility; WEL, Wind Engineering Laboratory; ACTEL, Advanced
Concrete Testing and Evaluation Laboratory; SHML, Structural Health Monitoring Laboratory; CL, Chemistry Laboratory; ASTAR, Advanced
Seismic Testing and Research Laboratory; CR, Concrete Road

Fib model code and Ekolu’s model are the suit- These locations are chosen based on varying CO2 con-
able models as they incorporate most of the factors influ- centrations and relative humidity values. Ekolu’s model is
encing carbonation in reinforced concrete structures and used as the ‘weather function’ in Fib model code cannot be
show a good correlation with the measured carbonation assumed easily for unsheltered exposure conditions. Car-
depth. However, Ekolu’s model cannot be used to predict bonation depth is more in indoor conditions (sheltered)
carbonation depth for concrete grade less than 20 MPa than in outdoor conditions (unsheltered) due to non-dis-
[16]. Kokubu and Nagataki’s and Hakkinen’s models also persive effect of air inside the structures. In unsheltered
showed the correlation coefficient R2 = 0.94; however, conditions, the pores in concrete are filled with water
they are strength-based models only. In situations, where during rainy season and the supply of CO2 is intermittent
adequate data are not available, these models can be used whereas; this is not the case in sheltered conditions where
to the predict carbonation depth. Parrott’s model has the there is a constant supply of ambient CO2. Due to this
least correlation obtained in all the buildings, and this reason, the carbonation depth for indoor exposure condi-
model cannot be used since air permeability coefficient is tion (sheltered) is more. Relative humidity and CO2 con-
generally not available. Air permeability varies greatly centrations reported in these regions are tabulated in
with concrete grade and addition of SCMs. These param- Table 7 [16].
eters are not considered in the model due to which the It is reported that the atmospheric CO2 concentration has
correlation coefficient is very less. Although Fib model increased by 40% from a preindustrial level of 278 ppm in
code and Ekolu’s model have the same correlation, Fib 1750 to 390.5 ppm in 2011 and further increase in atmo-
model code is used to predict the carbonation depth as it spheric CO2 concentrations may result in significant cli-
can be applied to all grades of concrete unlike in Ekolu’s mate changes in future. According to Abha and Ankit [21],
model. the average atmospheric CO2 level in India is around
From the values of corrosion initiation time for 40 mm 399 ppm during 2015. In coastal cities like Goa, the CO2
cover depth, it is concluded that ACTEL, CL and AML levels were shot up to 408 ppm, and in Madhya Pradesh,
show the corrosion initiation trends. It is observed that Uttar Pradesh and Himachal Pradesh the CO2 concentration
during the core extraction from the outside columns, sign is much higher in the range of 405–410 ppm. In the
of corrosion in ACTEL building is shown. Thus, these southern part of India and west coast, the CO2 concentra-
buildings have started to corrode and in need of attention in tion lies between 395 to 400 ppm and central and northern
order to ensure the safety of the buildings. regions registered between 400 to 405 ppm. CO2 concen-
tration is taken as 200 ppm, 300 ppm, 500 ppm and
Carbonation Depth Prediction at Different Key 1000 ppm for coastal areas, urban outdoor settings,
Locations industrial sites and tunnels, respectively. These values are
chosen within the range of CO2 specified so as to incor-
Carbonation depth is predicted using Ekolu’s model for porate the eco factor in Ekolu’s model.
different key locations such as rural areas, urban outdoor The optimum relative humidity for carbonation to take
settings, sea coast regions, industrial areas and tunnels. place is 50–70%. Relative humidity is taken as 70% in the

123
254 J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257

Fig. 1 Correlation between measured and predicted carbonation depth using various models

123
J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257 255

Table 7 CO2 concentration and relative humidity at key locations These grades are chosen as they are classified as normal,
[16] standard and high-strength concrete as per IS 456:2000
Key location Range of CO2 Average relative [23]. According to Fib model code, design life for tem-
concentration (ppm) humidity (%) porary structures, buildings structures and monumental
structures is 25, 50 and 100 years. Due to this reason, these
Rural areas 150–200 60
three different design life is chosen. Figures 2 and 3 show
Sea coast 150–250 70
the carbonation depth predicted for sheltered and unshel-
Urban outdoor settings 300–400 50
tered conditions.
Industrial areas 450–600 50
It is seen from the graphs that carbonation depth
Tunnels 1000–2000 60
increases with the increase in the concentration of CO2. In
all the cases, carbonation depth is the maximum for tunnels
sea coast region as large sea body evaporates and the and minimum for coastal region. It is least for sea coast
atmosphere has more water content and thus relative region as it has lower CO2 concentration and high relative
humidity is taken to be the highest in the coastal region, humidity. This can be further clearly noted between rural
whereas relative humidity is taken to be least in industrial areas and sea coast regions. Although both are assumed to
areas and urban areas due to no transpiration from vege- have the same CO2 concentration, carbonation depth is less
tation and the rainwater is unable to be absorbed into the in sea coast region than in rural areas because of low rel-
ground to evaporate into the air. Thus, relative humidity is ative humidity (50–60%) in rural areas. Thus, cover depth
the least for industrial and urban areas. In rural areas, rel- required in rural areas is more than the cover depth in sea
ative humidity is slightly higher than in urban areas due to coast to prevent the initiation of corrosion by carbonation.
the presence of vegetation. Thus, 60% is chosen as the From Figs. 2 and 3, it is inferred that the carbonation
average relative humidity in rural areas and tunnels depth decreases with the increase in the grade of concrete.
[16, 22]. This is witnessed in all the plots for sheltered as well as for
With the known CO2 concentration, temperature and unsheltered conditions. Also, it is noted that carbonation
relative humidity, minimum cover depth is predicted for front increases with the increase in age of the concrete. For
sheltered and unsheltered exposure conditions for 25, 50 the same concrete grade and key location (Table 7), car-
and 100 years for M20, M40 and M60 grades of concrete. bonation depth is more when the age of exposure is

Fig. 2 Predicted carbonation depths for sheltered conditions

123
256 J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257

Fig. 3 Predicted carbonation depths for unsheltered conditions

100 years followed by 50 years and 25 years. Outdoor • The measured carbonation depths are ranging from 13.4
urban settings and industrial areas are assumed to have the to 33.3 mm for the selected buildings of the exposure
same relative humidity in the study, but they have different ranging from 27 to 52 years.
carbonation depth because of more CO2 in the atmosphere. • Parametric study indicated that carbonation depth is
Thus, carbonation depth is more in industrial areas than in more in the case of tunnels, bridge abutments, deck
urban settings and more cover depth is required in the slabs, etc., and it is low in marine environments. Thus,
industrial areas. The cover depth required in these key more cover depth is required in tunnels to prevent
locations for a given service life should be more than the corrosion against carbonation.
predicted carbonation depth to prevent corrosion against • When comparing the measured average carbonation
carbonation. Thus, these values of carbonation depth depths of SHML and CL, the CO2 penetration rate is
indicate the minimum cover depth to be adopted in these lower in the case of buildings with SCMs (fly ash)
locations for a given grade of concrete, exposure conditions when compared to that of the buildings with OPC
and design service life of the building. alone.
• Parametric study indicates that higher cover depth is
Conclusions required in sheltered outdoor conditions than in the
unsheltered exposure conditions. Also, carbonation
The following conclusions are drawn from the present depth is more in industrial areas than in urban settings,
study: and hence more cover depth is required in the industrial
areas.
• Ekolu’s model, Fib model code, Kokubu and Nagata-
ki’s and Hakkinen’s models showed a good correlation
(R2 = 0.94) with the measured carbonation depths.
However, Kokubu and Nagataki’s and Hakkinen’s
models are strength based only and cannot take into References
account the parameters such as relative humidity,
1. E. Roziere, A. Loukili, F. Cussigh, Constr. Build. Mater. 23(1),
shelter conditions, etc. 190–199 (2009)
2. L. Jiang, B. Lin, Y. Cai, Cem. Concr. Res. 30(5), 699–702 (2000)

123
J. Inst. Eng. India Ser. A (June 2020) 101(2):247–257 257

3. K. Simsomphon, L. Franke, Cem. Concr. Res. 37, 1047–1058 14. Fib Model Code 2010 Final Draft, vols. 1 and 2 (Fib Bulletin N.
(2007) 65 and 66, Lausanne, 2012)
4. P. Sulapha, S.F. Wong, T.H. Wee, S. Swaddiwudhipong, J. 15. Fib Bulletin 34, (Fib bulletin, Lausanne, 2006), pp. 126. ISBN:
Mater. Civ. Eng. 15, 2 (2003) 978-2-88394-074-1
5. V.G. Papadakis, C.G. Vayenas, M.N. Fardis, J. AIChE 35, 10 16. S. Ekolu, Cem. Concr. Comput. 36, 688–696 (2018)
(1989) 17. L. Czarnecki, P. Woyciechowski, Bull. Pol. Acad. Sci. 61(1),
6. V.G. Papadakis, C.G. Vayenas, M.N. Fardis, ACI Mater. 88, 173–181 (2013)
363–373 (1991) 18. M. Thiery, G. Villain, P. Dangla, G. Platret, Cem. Concr. Com-
7. V.G. Papadakis, C.G. Vayenas, M.N. Fardis, ACI Mater. 8, 2 put. 37, 1047–1058 (2007)
(1991) 19. H. Cui, W. Tang, W. Liu, Z. Dong, F. Xiang, Constn. Build.
8. M.G. Alexander, M. Santhanam, Y. Ballim, Int. J. Adv. Eng. Sci. Mater. 93, 258–265 (2015)
Appl. Math. 2(3), 95–105 (2010) 20. V.G. Papadakis, Cem. Concr. Res. 30, 1031–1036 (2000)
9. M. Kokubu, S. Nagataki, ACI-SP 114, 315–329 (1989) 21. C. Abha, G. Ankit, Curr. Sci. 112, 12 (2017)
10. T. Hakkinen, Pub. No. 141, Technical Research Centre, Espoo, 22. S. Ekolu, Constr. Build. Mater. 127, 306–320 (2016)
Finland (1993) 23. IS 456:2000, Bureau of Indian Standards (BIS), New Delhi
11. ASTM C 173, ASTM International, West Conshohocken, PA, (2005)
USA (1993)
12. L. Parrott, in Proceedings of Third CANMET-ACI International
Conference (ACI SP-145), ed. by M. Malhotra, pp. 283–298 Publisher’s Note Springer Nature remains neutral with regard to
(1994) jurisdictional claims in published maps and institutional affiliations.
13. R.Torrent, G. Frenzer, Federal Office for Roads (VSS Report No.
516, Bern, Switzerland, 1995), p. 106

123