Progress in Organic Coatings 48 (2003) 194200

Design of coating resins by changing the macromolecular

architecture: solid and liquid coating systems
M. Johansson , T. Glauser, A. Jansson, A. Hult, E. Malmstrm, H. Claesson
Department of Fibre and Polymer Technology, Royal Institute of Technology, SE-100 44 Stockholm, Sweden

Abstract
An increased demand for new and improved coating systems, both due to environmental as well as performance reasons, have appeared
during the last decades. Techniques such low temperature curing powder coatings, radiation curable systems, and high solids have gained
an increased interest and obtained significant market shares. Although improved in many aspects, these systems still have limited use in
certain applications due to technical reasons.
One way to change the properties of thermoset resins that has obtained significant interest during the last decade is by changing the
molecular architecture of the resin. An example of polymers which exhibit different properties compared to conventional linear structures are
highly branched, dendritic, polymers [13]. These polymers, for example, exhibit a higher solubility and lower melt viscosity compared to
their linear counterparts. They can also be tailored with respect to functionality and polarity to adjust the properties for certain applications.
Coating resins based on hyperbranched polymers have been described both for liquid UV-curable systems as well as powder coatings [46].
This presentation will focus on how properties of resins based on dendritic polymers can be tailored and how this can applied to coating
systems such as powder coatings, radiation curable resins and no-solvent liquid systems. The correlation between resin structure and its
properties both before and after cure will be discussed.
2003 Elsevier B.V. All rights reserved.
Keywords: Resin chemistry; Macromolecular architecture; Hyperbranched polymers; Dendritic; Coatings

1. Introduction
The development of new coating systems have during the
last decades been pushed forward by new demands on either improved technical performance and/or environmental
concerns. Several new technologies such as radiation curable systems, waterborne systems and powder coatings have
obtained an increased market share since they address some
of these issues. The development of new resins, adopted to
these systems have been presented and significant improvement have been obtained. The development of a new systems needs numerous aspects to be considered to make them
function all the way in the process. Both the production of
the coating formulation, storage of the coating, application
and film formation must work with the techniques intended.
The final film must also exhibit properties at least as good as
the ones it is intended to replace. Finally, must the economical aspects be considered, i.e. it should not cost too much.
New monomers and polymers have traditionally been developed by the introduction of new reactive groups and/or
new backbone structures. For example, will the incorpora Tel.: +46-8-790-92-87; fax: +46-8-790-89-34.
E-mail address: matskg@polymer.kth.se (M. Johansson).

0300-9440/$ see front matter 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0300-9440(03)00105-X

tion of a rigid moiety such as an aromatic group into the

backbone increase the Tg of the systems. Another way that
during the last 10 years have obtained an increased interest
is to modify the properties of polymers by changing the
macromolecular architecture. The macromolecular architecture can be divided into several subgroups ranging from
linear to highly branched or side chain grafted structures.
One of these sub-groups is dendritic polymers comprising of hyperbranched polymers and dendrimers (Fig. 1).
Both dendrimers and hyperbranched polymers are built
from monomers containing three or more reactive groups
introducing a potential branching point in every repeating
unit, so called ABx -monomers. The difference between
dendrimers and hyperbranched polymers is that dendrimers
are perfectly built, monodisperse polymers built onto a core
molecule while hyperbranched polymers are polydisperse
and not fully reacted at every repeating unit.
Dendrimers can be synthesized according to several different synthetic principles such as divergent or convergent
synthesis. The main feature in common for most of these
synthesis routes is that they are time consuming and often
produce substances in low yield. This makes dendrimers
rather costly to produce and less attractive for large volume
applications. Hyperbranched polymers on the other hand can

M. Johansson et al. / Progress in Organic Coatings 48 (2003) 194200

195

B
A
B
*

*
*

*
*

*
*

*
*

*
*

*
*

*
*

*
*

*
*

Hyperbranched polymer

Dendrimer

Fig. 1. Schematic description of dendrimers and hyperbranched polymers.

be produced in large-scale to a reasonable cost making them

more attractive in several areas. The drawbacks with hyperbranched polymers are that they are polydisperse, are less
well-defined and that they contain linear segments in the
structure.
The properties of dendritic polymers have been shown
to exhibit significant differences compared to their linear
analogs. Dendritic polymers exhibit lower viscosities, are
non-entangled globular structures, and often have increased
solubility in various solvents. Dendritic polymers contain a
large number of end-groups in the structure and it has been
shown that the structure of these end-groups have a large
influence on the properties. The possibility to modify the
structure of the end-groups can be used to tool to adjust the
properties of the polymer for specific applications such as
toughening additives [7] or rheology modifiers [8]. The use
of hyperbranched polymers in several coating applications
[9] such as alkyds [10], UV-curable systems [11], and low
VOC PU-coatings [12] have been presented recently.
The present paper aims to describe the use of hyperbranched polymers as scaffolds for different thermoset coating resin systems. The effect of structural variations on
important properties, both before and after curing, are addressed. It is shown that the same hyperbranched backbone
can be used for such different resins as air-drying alkyds to
semi-crystalline powder coatings. The use of hyperbranched
polyesters as toughening additives is also shown.

are based on 2,2-bis(methylol)propionic acid (bis-MPA)

as AB2 -monomer and ethoxylated pentaerytritol as core
molecule. The hyperbranched polyesters are hydroxyfunctional aliphatic polyesters with a size and OH-functionality
depending on the core:monomer ratio (Table 1). Dendrimers
corresponding to some of the hyperbranched polyester have
been used for some test series.
-Caprolactone (-CL) was dried over CaCl2 and distilled
twice before use. All other chemicals were purchased from
Aldrich, Lancaster, Ciba-Geigy and used as-received.

2. Experimental

Base
structure
no.

Average
molecular
weight
(g/mol)

Average
functionality
(number of
end-groups)

End-group
structure

Polymer type

H20
H30
H40
D30

1750
3600
7300
2745

16
32
64
24

OH
OH
OH
OH

Hyperbranched
Hyperbranched
Hyperbranched
Dendrimer

2.2. Equipment
1H

and 13 C NMR were recorded on a Bruker AM 400,

at 400 and 100 MHz, respectively, using CDCl3 as solvent.
The solvent signal was used as internal standard. Size exclusion chromatography (SEC) was performed on a Waters
SEC-system using a solvent delivery system (Model 510),
automatic injector (WISP 710B) and a differential refractometer (Waters 410) as a detector and three PLgel 10 m
mixed-B columns from Polymer Labs. THF, 1.0 ml min1 ,
was used as the mobile phase. Calibration was carried
out using linear polystyrenes of known molecular weight
and dispersity. Calorimetric studies were carried out on a
Mettler Toledo DSC 820 equipped with a Mettler Toledo
Table 1
Basic data on the dendritic polyesters used in the studies

2.1. Materials
The hyperbranched polyesters used as base for most
resins were kindly supplied by Perstorp AB, Perstorp, Sweden and used as-received. The hyperbranched polyesters

196

M. Johansson et al. / Progress in Organic Coatings 48 (2003) 194200

Sample Robot TSO801RO calibrated using standard procedures. Photo-DSC experiments were performed on a similar
Mettler Toledo DSC equipped with a Hamamatsu UV Spot
Light Source L5662. FTIR measurements were performed
on a Perkin-Elmer Spectrum 2000 either in transmission
mode or as ATR-spectra utilizing a Golden Gate (diamond
crystal) accessory from Grasseby Specac. A Leitz Ortholux
POL BK II optical polarizing microscope equipped with
a Mettler Hot Stage FP 82 central processor was used for
studying crystallinity. UV curing was performed on a high
power microwave irradiator from Fusion UV Curing system, using an electrodeless H-bulb regulated with a P300
Power Supply. The spectral range used for curing was in
between 320 and 390 nm. The dose was measured with an
UVICURE Plus device from EIT Inc.
Rheology measurements were carried out on a Rheometrics Dynamical Analyzer RDAII, using parallel plates with
a radius of 3.95 mm. Dynamic mechanical properties were
analyzed using a Polymer Laboratories dynamic mechanic
thermal analyzer (DMTA) in tensile mode.
2.3. Synthesis
The base polyesters have been modified in several different ways to obtain thermoset resin structures. The different
syntheses have been base catalyzed esterification of the hydroxyl end-groups with suitable anhydrides or acid chlorides
or ring-opening polymerization of -CL onto the hydroxyl
groups. Direct esterification with suitable acids, such as tall

oil fatty acids, have also been performed. Detailed description of the synthesis routes can be found elsewhere [13].

3. Results and discussion

3.1. Synthesis of dendritic polymers
Modification of a hyperbranched hydroxyfunctional
polyester to a thermoset resin structure is readily achieved
via standard reactions such as esterification with acid chlorides or anhydrides. The hydroxy end-groups can either
be functionalized with reactive end-groups suitable for
cross-linking or non-reactive end-groups, which only affect
the polarity of the system.
Different fractions of non-reactive/reactive end groups
can be achieved to adjust the cross-link density of the final network. Very different resins ranging from alkyds to
UV-curable resins and powder coatings can be obtained depending on the structure and type of the end-groups (Fig. 2).
The large number of hydroxyl end-groups does, however,
present one small obstacle that may occur. The ester backbone together with the large number of hydroxyl end-groups
allows hydrogen bonding to form a thermoreversible physical network with time. This makes the polyester difficult
to dissolve unless it is first heat-treated. The secondary network of hydrogen bonding can be cleaved by heating the
polymer to a temperature above 130 C. This is seen as
an endotherm in the first temperature sweep in DSC-scans

Hyperbranched
polymer

+ Reactive
end-groups

+ Non-reactive
end-groups

Hyperbranched
polymer

Hyperbranched
polymer

Hyperbranched
polymer

Curing

End-groups Resin type

Fatty acids
Acrylate
Acrylate
Maleate

Alkyd
UV-coating
UV-powder
UP-resin

Hyperbranched
polymer

Thermoset
Fig. 2. Principle of synthesis routes for different resin structures based on hyperbranched polymer scaffolds.

M. Johansson et al. / Progress in Organic Coatings 48 (2003) 194200

Fig. 3. DSC thermogram of an hydroxyfunctional hyperbranched polyester

heating and cooling rate of 10 C min1 after the sample has been annealed
at 70 C for 16 h.

on annealed samples of hydroxyfunctional hyperbranched

polyesters (Fig. 3). The hydrogen bond formation process,
have been shown to be fully thermoreversible and not due to
any chemical reaction or moisture absorption. The hydrogen
bonding will, after heat treatment, slowly reform with time
but this is not a problem as long as the polyester is used
within reasonable time frame. A more extensive description
of the annealing process has previously been presented elsewhere [14].
The strong interaction between the numerous end-groups
also presents a problem when analyzing the polymers with
respect to molar mass. Aggregation and interaction with
column materials often makes the determination of the
molecular weight difficult to perform when the end-groups
are highly polar. Studies on modified polyesters where the
end-groups have been altered to non-polar moieties have,
however, revealed that the data presented in Table 1 is
reasonably correct [15].

4. Resin properties
4.1. Liquid resins
One of the major issues for coating resins is the rheological behavior before curing, i.e. during the application and
film formation process. The optimum for a liquid coating
is to have low enough viscosity to allow application with
the desired techniques and subsequent wetting and leveling
without any sagging effects. This is, however, difficult to
achieve just by decreasing the molecular weight of the resin
since this would lead to too much cure shrinkage and internal
stresses in the final film. One of the advantages with highly
branched structures is that the viscosity is lower than corresponding linear polymer. This is seen for hyperbranched
polymers where the viscosity is significantly reduced compared to their linear analogues. The glass transition, Tg , has
been shown to largely depend on the structure and polarity

197

of the end-groups. The Tg have been shown to differ up to

100 C as reported by several authors [16]. The Tg of the hyperbranched polyesters in the present study has been shown
to differ from, 10 C for alkyl (propionate), to 35 C for
hydroxyl end-groups. This corresponds to a large viscosity difference at ambient temperature. The possibility to affect the viscosity is more limited for reactive resins since
a fraction of the end-groups needs to be a specific reactive
group and only the remaining fraction can be modified. Rheological studies of the hyperbranched polyesters revealed a
Newtonian behavior in all cases with small end-groups. The
aforementioned physical network structure due to hydrogen
bonding gave a shear thinning behavior for annealed hydroxyfunctional polyesters. The thermally treated hydroxyfunctional polyester had a Newtonian behavior.
4.2. Solid resins
Powder coating is one group of coatings that have obtained significant interest during the last decades. These
coatings are solid powders at ambient conditions and applied as powders to a substrate. The applied powders are then
melted at elevated temperature to form a homogeneous film,
and in most cases thermally cross-linked to a thermoset. One
limit for conventional powder coatings is that the powder
must be stable upon storage at ambient conditions, i.e. they
must have a Tg above 60 C. The lowest temperature for the
film formation of normal amorphous powder coating is determined by this Tg . The temperature need to be raised at
least 50 C above Tg to obtain good enough film properties.
This means that the lowest curing temperature for conventional is around 110 C which omits several substrates due
to thermal stability reasons.
Several ways to reduce this lowest curing temperature for
coatings such as the use of UV-curing, crystalline monomers
or combination of these have been proposed [17]. Early studies on hyperbranched polymers showed that it was possible to produce semi-crystalline structures by attachment of
end-groups that could crystallize. It has also been shown that
it is possible to graft crystalline end-groups by ring opening
polymerization of -CL to hyperbranched polyester cores
to produce a semi-crystalline starbranched hybrid structure
with suitable melting temperatures [18]. The grafting of
-CL also allows the resin to be further functionalized to an
UV-curable structure (Fig. 4).
The resins had a melting point of approximately 60 C
above which the viscosity rapidly dropped to allow a film
formation around 80 C. The resins solidified readily upon
cooling and powders could be produced by grinding and
sieving procedures. The grinding process was necessary to
cool since the friction heat otherwise increased the temperature to the melting temperature region.
A series of resins based on dendritic polyesters with
different core structure and length of the crystalline grafts
have been produced and studied with respect to the rheological behavior in comparison with linear PCL-polymers.

198

M. Johansson et al. / Progress in Organic Coatings 48 (2003) 194200

OH

O
O

O
n

OH

O
O

Dendritic
core

Dendritic
core

SnOct2
O

O
n

OH
Et3N, DMAP

O
O
Dendritic
core

OH

O
O

Fig. 4. Schematic description of synthesis of semi-crystalline resins for powder coating applications.

The results reveal that the main factor affecting the melt
viscosity is the length rather then number of grafts on the
dendritic polyester (Fig. 5). The viscosities presented in
Fig. 5 will overlap for the different structures if it is plotted
against the molecular weight of the grafts rather than the
total molecular weight. No significant differences were seen
when comparing the viscosity between a dendrimer and a
hyperbranched polymer as core structure.
The conclusion from this is that low melt viscosity in
combination with high molecular weight can be obtained for
resins based on highly functional core structures. A rheological behavior suitable for low temperature curing powder

coatings can be obtained by introduction of semi-crystalline

segments as end-groups on the hyperbranched polyester
backbone.
4.3. Curing performance and final properties
A series of acrylate functional hyperbranched polyesters
with 30% of the end-groups acrylate functionalized and the
rest either hydroxyl, propionate or benzoate groups have
been UV-cured as thin films. Measurements of the residual
unsaturation with Raman spectroscopy reveal very low levels
remaining unreacted acrylate groups. This means that the

1000

Viscosity (Pa s)

100

10

0,1

0,01
1000

10000

100000

1000000

Total molecular weight (g/mol)

Fig. 5. Viscosity vs. total molecular weight for () linear, () starbranched (hyperbranched core) and () starbranched (dendrimer core) polyesters.

M. Johansson et al. / Progress in Organic Coatings 48 (2003) 194200

reactive end-groups are available for polymerization and not

trapped inside the resin structure. The Tg of the final films are
between 67 and 77 C (tan peak) depending on the structure
of the non-reactive end-groups. Changing the end-groups
from hydroxyl groups to propionate only shifts the Tg 10 C
indicating that the major factor affecting the Tg is type and
amount of cross-linking.
Changing the reactive end-groups to methacrylates shifts
the Tg upwards. The effect of functionality was studied on
three different methacrylate functional resins with 30, 45,
and 60% of the end-groups functionalized. The Tg in the
final films shifted from 100 to 160 C with increasing functionality. The amount of residual unsaturation also increased
with increased functionality going from 5 to 30%. The increase in residual unsaturation after UV-curing is mainly
due to vitrification effects since the Tg increases well above
ambient temperature.
Films of the PCL-grafted semi-crystalline resins were
readily UV-cured in the molten state to form cross-linked
films with little or no residual unsaturation. The properties of the films differed depending on the resin structure.
Resins with short grafts, up to DP 10, were amorphous networks with a Tg of 56 C while longer grafts resulted in a
semi-crystalline film with a melting point around 50 C. The
degree of crystallinity was strongly reduced compared to the
uncross-linked resin. Shorter grafts result in an increased
cross-link density, which affected the modulus above Tg /Tm .
4.4. Resin additives
Another material of interest in many thermoset applications is controlled phase separated systems to obtain high
modulus materials with improved toughness. Many high
modulus thermoset materials suffer from brittle fracture behavior and must be modified to improve this property. One
way to retain the high modulus at the same time as the
toughness is increased is by the introduction of small micron size phase separated particles into the thermoset. This
can, for example, be achieved by reaction induced phase
separation in thermally cured two component systems, e.g.
epoxyamine-structures. Reaction induced phase separation
is more difficult to obtain for radiation curable system since
the gel point occur at early stages of conversion, i.e. the
system must be phase separated from the beginning. The
possibility to tailor the polarity and functionality of hyperbranched polymers by end-group functionalization, make
them good candidates for this approach.
A series of resin additives with different polarity was synthesized by attachment of alkyl chains of different length to
previously partly methacrylated hyperbranched polyester.
The longer the alkyl chain length, the lower the polarity
of the additive. These resins were mixed in small amounts
with a conventional radiation curable methacrylate thermoset resin (EB610 from UCB). The mixtures were more
and more phase separated with decreasing polarity of the
additive. The additive with the lowest polarity formed a sta-

199

Table 2
Data on EB-cured resin with respect to impact toughness and Tg

K1c (MPa m0.5 )

Tg ( C)

Neat
resin

Medium polarity
additive

Low polarity
additive

154
38

135
44

142
47

ble emulsion with the resin with phase separated domains

in the micron size. EB-curing of the resin mixtures revealed
the desired effectretained high modulus and increased
toughness. Some of the results are summarized in Table 2
with values of Tg and K1c .
5. Conclusions
The use of hyperbranched polyesters as base for thermoset coating resins has been demonstrated. Very different
resins ranging from liquid UV-systems to low temperature
curing powder coatings and resin additives based on the
same hyperbranched polyester have been described. The
use of highly branched resins present advantages such as
reduced need for solvents, and low melt viscosities. The
numerous end-groups can be modified to produce very
different cross-link densities and polarities of the films.
Semi-crystalline resins can be produced by attachment of
crystalline end-groups yielding resins with rheological properties suitable for low temperature cured powder coatings.
References
[1] A. Hult, M. Johansson, E. Malmstrm, Adv. Polym. Sci. 143 (1999)
1.
[2] F. Vgtle, S. Gestermann, R. Hesse, H. Schwierz, B. Windisch, Prog.
Polym. Sci. 25 (2000) 987.
[3] J.M.J. Frchet, D.A. Tomalia (Eds.), Dendrimers and Other Dendritic
Polymers, Wiley Series in Polymer Science, Wiley, Chichester, UK,
2002, pp. 1647.
[4] M. Johansson, E. Malmstrm, A. Jansson, A. Hult, J. Coat. Technol.
72 (2000) 49.
[5] D. Schmaljohann, B.I. Voit, J.F.G.A. Jansen, P. Hendriks, J.A.
Loontjens, Macromol. Chem. Eng. 275 (2000) 31.
[6] M. Johansson, A. Hult, J. Coat. Technol. 67 (1995) 35.
[7] (a) R. Mezzenga, L. Boogh, J.-A.E. Manson, Compos. Sci. Technol.
61 (2001) 787;
(b) G. Jannerfeldt, L. Boogh, J.-A.E. Manson, Polymer 41 (2000)
7634.
[8] Y. Hong, S.J. Coombs, J.J. Cooper-White, M.E. Mackay, C.J. Hawker,
E. Malmstrm, N. Rehnberg, Polymer 41 (2000) 7705.
[9] (a) R.A.T.M. van Benthem, Prog. Org. Coat. 40 (2000) 203;
(b) D. Schmaljohann, B.I. Voit, J.F.G.A. Jansen, P. Hendriks, J.A.
Loontjens, Macromol. Mater. Eng. 275 (2000) 31.
[10] G. Lin, J.A. Baghdachi, J.L. Massingill Jr., Polym. Mater. Sci. Eng.
85 (2001) 135.
[11] A. Hult, E. Malmstroem, M. Johansson, Polym. Mater. Sci. Eng. 72
(1995) 528.
[12] J. Huybrechts, Proceedings of the FATIPEC 24th Congress, vol. C,
1998, pp. C/43C/56.
[13] (a) M. Johansson, T. Glauser, G. Rospo, A. Hult, J. Appl. Polym.
Sci. 75 (2000) 612;

200

M. Johansson et al. / Progress in Organic Coatings 48 (2003) 194200

(b) M. Trollsas, C.J. Hawker, J.F. Remenar, J.L. Hedrick, M.

Johansson, H. Ihre, A. Hult, J. Polym. Sci. 36 (1998) 2793;
(c) J. Lange, E. Stenroos, M. Johansson, E. Malmstrom, Polymer 42
(2001) 7403;
(d) E. Malmstrm, M. Johansson, A. Hult, Macromolecules 28 (1995)
1698.
[14] M. Johansson, A. Hult, J.-A.E. Mnson, Proceedings of the
Macroakron94, 35th IUPAC International Symposium on Macromolecules, Akron, OH, USA, 1994.

[15] E. Malmstrm, KTH, Personal communication.

[16] (a) B.I. Voit, Acta Polym. 46 (1995) 87;
(b) B.I. Voit, S.R. Angew, Macromol. Chem. 223 (1994) 13.
[17] (a) L. Moens, J.-M. Loutz, D. Maetens, P. Loosen, M. van Kerkhove,
US Patent 5 639 560, UCB, SA, 1997;
(b) M. Johansson, H. Falkn, A. Irestedt, A. Hult, J. Coat. Technol.
70 (1998) 57.
[18] H. Claesson, M. Doyle, E. Malmstrm, M. Johansson, J.-A.E.
Mnson, A. Hult, Prog. Org. Coat. 44 (2001) 63.