You are on page 1of 7

Cement & Concrete Composites 46 (2014) 19–25

Contents lists available at ScienceDirect

Cement & Concrete Composites

journal homepage:

Effect of waste latex paint on concrete

Nasser Almesfer ⇑, Jason Ingham
Department of Civil and Environmental Engineering, The University of Auckland, Auckland 1142, New Zealand

a r t i c l e i n f o a b s t r a c t

Article history: Approximately 20% of paint sold in western countries does not get used for its intended purpose, and in
Received 10 December 2012 time much of this left-over paint ends up in land-fill as part of the household waste collection, at a sig-
Received in revised form 20 October 2013 nificant economic and environmental cost. In New Zealand, a comprehensive product stewardship cam-
Accepted 31 October 2013
paign has been initiated to recover waste paint before it enters the waste stream, while the collaboration
Available online 9 November 2013
amongst a cohort of companies has contributed to utilising waste latex paint in concrete. The objective of
this study was to investigate the effects of waste latex paint on concrete, with special focus on a low
strength standard 20 MPa concrete mix. It is demonstrated that waste latex paint can improve the work-
ability and durability of concrete, whilst achieving sufficient compressive strength. Phase analysis indi-
Polymers cated that the addition of waste latex paint does not affect the concrete hydrates. It was established
Characterisation that concrete with 12 L/m3 of waste latex paint is suitable for use in low strength non-structural concrete.
Mechanical properties Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction However, excesss air entrainment can be controlled with the prop-
er selection and amount of latex and antifoaming agent used [10].
Latex dispersions, which consist of small organic polymer parti- Despite its improved flowability, PMC has improved resistance to
cles dispersed in water, have been widely used to modify Portland segregation and bleeding due to the hydrophilic properties of the
cement-based concrete and produce what is commonly known as polymers and the air-entraining and water-reducing effects of
polymer-modified concrete (PMC) [1–3], with latex dispersions the surfactants, and hence reductions in strength and waterproof-
improving some undesired fresh and hardened cement properties ness caused by bleeding and segregation are reduced in PMC
such as cohesion, adhesion, brittleness, low flexural strength, and [1,11].
poor durability. Additionally, PMC is suitable for various structural Generally the compressive strength of PMC is lower than that of
and non-structural products and is ideal for overlay of pavements, ordinary concrete after 28 days, but equalises or exceeds the com-
bridges, and industrial applications [4,5], due to PMC having a pressive strength of ordinary concrete after 90 days due to the im-
superior monolithic cement matrix in which the organic polymer proved water retention of PMC [1,8]. Ohama [1] states that it is
phase and cement phase are interpenetrated and homogenised evident that optimum strength in PMC is obtained by achieving a
[6,7]. The formation of the microstructure in PMC in relation to reasonable degree of cement hydration under wet conditions at
the polymer film formation being coincident with cement hydra- early stages, followed by polymer film formation under dry condi-
tion is described by various authors and qualitative models [1,4,8]. tions. This theory is also supported by Beeldens et al. [8], stating
Previous studies have indicated that the workability of PMC is that at high relative humidity, the influence of polymer modifica-
enhanced due to surfactants from the latex dispersion being ad- tion on short-term strength is limited, but as soon as a dry curing
sorbed onto the polymer and cement particles, generating a dis- period is introduced, a polymer film starts to build up through the
persing effect and an increase in entrained air, and also due to binder phase and an increase in strength is noticed with an in-
the ‘‘ball bearing’’ action of the polymer particles [1], while Lewis crease in polymer/cement ratio. Polymers have been reported to
and Lewis [9] reported that surfactants have a lubricating effect improve the cohesion of the bulk cement paste, with polymer
on the wet concrete mix, reducing the viscosity, and hence high bridges being detected between the portlandite crystal layers, act-
workability of the cement paste is achieved at a much lower water ing as an additional bond and gluing the layers together [12]. These
to cement ratio. The action of the surfactants contained as emulsi- bond properties are why PMC is used extensiveley for overlays on
fiers and stabilizers in the polymer latex means that a large quan- concrete bridges and parking decks [10]. The resistance of PMC
tity of air is entrained in comparison to ordinary cement concrete. against freezing and thawing, and wearability is superior, due to
PMC’s low permeability to water, while 30 years of field experience
⇑ Corresponding author. Tel.: +64 21 022 55239. has shown no freeze–thaw deterioration [10]. The pores in PMC
E-mail address: (N. Almesfer).
can be filled with polymers or sealed with continuous polymer

0958-9465/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
20 N. Almesfer, J. Ingham / Cement & Concrete Composites 46 (2014) 19–25

films, reducing water absorption and permeability as the polymer– water-reducing and air-entraining chemical admixtures, with the
cement ratio increases. Such good water impermeability provides evaluated properties including slump-loss, compressive strength,
high resistance to chloride ion penetration and reduced transmis- flexural strength and splitting tensile strength, while fundamental
sion of such gases as carbon dioxide (CO2) and oxygen (O2), which components of durability, microstructure, and hydration were also
are all important factors contributing to the corrosion of steel rein- assessed as reported below.
forcement in reinforced concrete structures [1,5]. Yang et al. [7] re-
ported that the incorporation of styrene–butadiene rubber (SBR)
2. Materials
latex led to a more dense and more refined microstructure of ce-
ment mortar, reducing water permeability and chloride diffusion.
Test specimens were made using the same materials as used in a
Paint, which is a latex dispersion, is considered a hazardous
standard 20 MPa concrete mix produced by a New Zealand ready-
product and leftover paint is a hazardous waste which must be
mix concrete plant, with a maximum nominal aggregate size of
managed appropriately. Similar to the program that is currently
19 mm, and general purpose (ASTM C150 Type I) cement being used.
used in North America [13], a comprehensive product stewardship
Details of the standard 20 MPa concrete mix are presented in Table 1,
campaign has been initiated in New Zealand to recover waste latex
and the chemical composition and physical properties of the cement
paint (WLP) before it enters the waste stream so that it can be recy-
are presented in Tables 2 and 3 respectively. A commercial air-
cled in applications which add value for the key participants in the
entraining admixture was used along with a commercial water-
recovery programme, while the collaboration amongst a cohort of
reducing admixture composed of modified polycarboxylates. All
companies has contributed to using WLP into potentially more
materials were supplied by a local concrete plant, while various
economical and environmentally-friendly concrete.
WLP samples were sourced from the New Zealand paint collection
WLP has been used in urban concrete sidewalks in Ontario, Can-
program. WLP collected through the paint collection program was
ada [3], where WLP contributed in a similar form to virgin latex in
separated from its packaging either by the use of a decanter (dec-
concrete, such as increasing the flexural strength and decreasing
anted paint) or by manually pouring the paint out of its packaging
chloride ion diffusion. A study was also conducted at the University
(pailed paint), and was stored in 200 litre drums until re-use. Six rep-
of Bath [14] where the rheological and hardened properties of PMC
resentative samples of collected WLP were obtained and analysed,
mixes using WLP as the enhancing admixture were investigated.
and their properties are listed in Table 4. There was a clear difference
The addition of WLP was found to either enhance or deteriorate
in properties between the samples of paint that were sourced from
the compressive and tensile strength depending on the water con-
drums of pailed paint (P) and the samples of paint that were sourced
tent, with air entrainment believed to be the main cause of deteri-
from drums of decanted paint (D), especially in pH and viscosity.
oration. The study demonstrated that polymers contribute to the
Recognising that this WLP was a non-controlled waste product,
enhanced strength, and that the effects of polymers on the work-
the variation shown in Table 4 was considered acceptable. To control
ability of the mix are dependent on the polymer type, with recom-
air-entrainment, a proprietary anti-foaming agent was mixed into
mendations that further study be conducted into the effects of
the paint prior to making the WLP concrete specimens.
different WLP dosages, air entrainment, and mixing procedures.
Mohammed et al. [13] investigated the effects of WLP on a range
of fresh and hardened properties of concrete, incorporating 5%, 3. Experimental methods
10%, 15%, 20%, and 25% of WLP as partial replacement by mass
for mixing water (0.3%, 0.6%, 0.9%, 1.2%, and 1.5% polymer/cement The effect of WLP as a replacement for conventional water-
ratio). WLP enhanced the workability of concrete, and did not neg- reducing and air entraining admixtures and partial replacement
atively affect the setting time or slump loss. In addition, the partial for mixing water in concrete was investigated. The mixing se-
replacement of mixing water by WLP led to a considerable increase quence initially involved mixing the WLP and 50% of the mix water
in the flexural strength of concrete, but a tendency to decrease the for 30 s, then adding the saturated coarse and fine aggregates, ce-
compressive strength. It was suggested that WLP forms a polymer ment, and the remaining water and chemical admixtures (only
film in concrete, while its pigments and extenders fill additional for control mix) to the premixed water (containing WLP) and
pores, resulting in a cementitious matrix with a relatively denser resuming mixing for a further 3 min.
microstructure that enhances the long-term durability of concrete The initial tests involved finding the optimum WLP dosage to
to rapid freezing-and-thawing cycles, surface scaling by de-icing use in concrete, using indications from preliminary results [11]
salts, and exposure to sulfuric acid. Moreover, the resistance of and previous testing using WLP in concrete [13]. To find this opti-
concrete to rapid chloride ion penetrability increased with WLP mum WLP dosage, compressive strength and slump loss tests were
partial replacement for mixing water. Almesfer et al. [11] investi- conducted. Compressive strength testing consisted of a control
gated the use of WLP as a partial replacement for mixing water concrete mixture with no WLP; and six WLP concrete mixtures
in masonry blockfill, and established that WLP dosages of 8–16% (incorporating the six representative WLP samples) with WLP at
replacement of water were optimum for maintaining strength 8, 12, and 16 L/m3 of concrete (approximately 8, 12, and 16% partial
and improving workability. It was concluded that WLP improves replacement by mass for mixing water). For each mixture, nine
the workability and rheology of masonry blockfill, resulting in an 100  200 mm cylindrical samples were cast. After demoulding
increase in spread and a significant decrease in viscosity and sep- at 24 h, all specimens were cured by immersion in water at
aration, and hence WLP was found to be viable as a replacement approximately 23oC until testing. Compressive strength testing
for conventional chemical admixtures. was carried out for all mixtures at time periods of 7, 28, and
The objective of this study was to investigate the use of WLP in 56 days in accordance to NZS 3112 [15], with three specimens
low strength standard 20 MPa concrete, recognising the potential being tested at each time period for each mixture. To investigate
for use of WLP as a viable admixture in non-structural concrete the effect of curing on the compressive strength of concrete pro-
applications such as pavements and sidewalks [13,3]. While preli- duced with WLP, instead of using the prescriptive wet curing con-
minary results were previously published [11] that indicated opti- ditions where concrete is immersed in water until testing, a wet–
mum WLP dosages in masonry blockfill, the aim of the present dry curing condition was utilised. This wet–dry procedure entailed
study was to investigate the effects of WLP on a standard 20 MPa three additional cylindrical specimens being made for mixtures
concrete mix produced by a New Zealand ready-mix concrete incorporating WLP, that were immersed in water for 16 days and
supplier. WLP was used as a replacement for the conventional then dry cured at approximately 23 °C until being tested at
N. Almesfer, J. Ingham / Cement & Concrete Composites 46 (2014) 19–25 21

Table 1
Details of standard 20 MPa concrete mix used.

Cement (kg) Water (L) Aggregate (kg) Water-reducing agent (ml) Air-entraining agent (ml)
19 mm 13 mm Sand
210 145 505 584 865 530 100

C 1585 [16] and ASTM C 1556 [17] guidelines respectively, whilst

Table 2
Chemical composition of cement. using wet curing for the concrete mix with no WLP, and wet–dry
curing for the concrete mix with WLP. Sorptivity testing involved
Compounds % (By weight)
immersing an exposed face of three unsaturated concrete speci-
SiO2 22.8 mens per mixture in water after being cured, while the rate of
Al2O3 4.2
absorption (dominated by the capillary suction mechanism) was
Fe2O3 2.3
CaO 64.8 determined by measuring the increase in mass of the specimens
MgO 1.0 as a function of time. The bulk chloride-diffusion coefficient was
Na2O 0.2 calculated for the average chloride content profile from three spec-
K2O 0.5 imens per mix.
Free lime (CaO) 1.3
Hypothetical compounds
Lastly, to analyse the effect of WLP on the hydration products
Lime saturation factor 95.0 and microstructure of OPC concrete, the concrete mixture with
C3S 68.0 no WLP; and six WLP concrete mixtures (incorporating the six rep-
C2S 14.0 resentative WLP samples at 12 L/m3) were analysed for the follow-
C3A 7.0
ing tests:
C4AF 7.0

 X-ray diffraction (XRD), carried out on a Bruker D2 Phaser dif-

fractometer, to record diffractograms in the range 15–55o.
Table 3  Thermogravimetry (TGA/DTG), using a thermogravimetric anal-
Physical properties of cement. ysis instrument with the temperature set to increase at a rate of
10 oC/min from room temperature to 600 oC. The rate of weight
loss as a function of temperature indicated by a thermogravi-
Specific surface area (m2/kg) 329
metric curve was recorded and used to provide quantitative
Initial setting time (min) 145
Final setting time (h:min) 4:18 and qualitative information on concrete hydration products
Soundness (mm) 1 such as portlandite (calcium hydroxide).
SO3 (%) 1.9  Fracture specimens obtained from OPC concrete and WLP con-
Specific gravity (g/cm3) 3.09 crete samples that fractured during mechanical testing were
attached to aluminium plates using carbon tape, after first being
28 days. Slump loss testing was completed on concrete mixtures sputter coated with platinum for analysis with a Quanta 200
incorporating WLP at 0, 8, 12, and 16 L/m3 of concrete, and was FEG scanning electron microscope (FE-SEM).
carried out at 20 min intervals over 120 min, with the concrete
being left static over the 20 min interval between tests. The con- A summary of the aforementioned tests conducted is presented
crete was mixed for 25 s prior to each slump test, to ensure that in Table 5.
there were no inconsistencies within the concrete due to bleeding
and/or segregation. 4. Results and discussion
After finding an optimum paint dosage based on compressive
strength and slump loss, additional strength and durability testing 4.1. Optimum paint dosage
was conducted on the control concrete mix and a concrete mix
with WLP at a dosage of 12 L/m3 based on the optimum proportion 4.1.1. Compressive strength
suggested by initial testing. The flexural strength and splitting ten- The 7-day compressive strength results (Fig. 1a) indicated that
sile strength were obtained at 28 days using the wet and wet–dry the control mix had a larger compressive strength (16.6 MPa) than
curing conditions on three 100  100  350 mm beam specimens all the mixes with WLP, suggesting that WLP either delayed the
(satisfying ASTM C 78 requirements) and three 100  200 mm strength development of concrete, or simply reduced the strength
cylindrical specimens respectively, according to NZS 3112. In addi- of the cement matrix. It can also be observed in Fig. 1a that too lit-
tion, the rate of water absorption (sorptivity) and bulk chloride- tle WLP (8 L/m3) can be as disadvantageous to strength gain as too
diffusion coefficients were obtained for each mix as per the ASTM much WLP (16 L/m3), although there was no significant difference

Table 4
Properties of waste latex paint.

Parameters Pailed paint Average of pailed paint Decanted paint Average of decanted paint Average
P1 P2 P3 D1 D2 D3
pH 7.82 7.95 7.67 7.81 8.29 8.17 8.24 8.23 8.02
Viscosity (KU) 77 81 73 77 97 94 106 99 88
Density (kg/L) 1.292 1.283 1.260 1.278 1.258 1.288 1.303 1.283 1.281
Pigments (%) 28.1 27.5 26.5 27.4 24.6 27.6 30.2 27.5 27.4
Polymers (%) 24.2 22.5 20.9 22.5 26.9 25.4 24.1 25.5 24.0
Water (%) 47.8 50.0 52.6 50.1 48.5 47.0 45.7 47.1 48.6
22 N. Almesfer, J. Ingham / Cement & Concrete Composites 46 (2014) 19–25

Table 5
Summary of tests conducted.

Test WLP dosages tested (L/m3) Number of WLP samples tested Curing condition(s) Concrete age(s) when tested (days)
Compressive strength 0, 8, 12, 16 Six (P1, P2, P3, D1, D2, D3) Wet 7, 28, 56
Compressive strength (wet–dry 8, 12, 16 Six (P1, P2, P3, D1, D2, D3) Wet–dry 28
Slump loss 0, 8, 12, 16 One (average of WLP samples) – –
Flexural strength 0, 12 One (average of WLP samples) Wet, wet–dry 28
Splitting tensile strength 0, 12 One (average of WLP samples) Wet, wet–dry 28
Sorptivity 0, 12 One (average of WLP samples) Wet, wet–dry 28
Chloride diffusion 0, 12 One (average of WLP samples) Wet, wet–dry 28
XRD, TGA/DTG, ESEM 0, 12 Six (P1, P2, P3, D1, D2, D3) Wet, wet–dry 28

(less than 5%) in the compressive strength of mixes with different or decrease the overall tensile strength of concrete, although other
WLP dosages. The variation between the compressive strength of researchers have shown that the addition of WLP to concrete gen-
mixes with the six different samples of paint was not substantial, erally does result in an increase in flexural strength due to
while the compressive strength at 28 days (Fig. 1b) and at 56 days improvements in the bond at the paste–aggregate interfacial tran-
(Fig. 1c) followed a similar trend to that shown at 7 days. The over- sition zone [3,11,13,14] depending on the quantities of WLP and
all finding was that the 12 L/m3 WLP dosage was most optimal cement, and the curing method used.
when considering the compressive strength of concrete with
WLP, exceeding the 24.5 MPa average compressive strength 4.3. Durability
requirement at 28 days as specified by NZS 3104 [18] for the pro-
duction of concrete having a lower characteristic strength of Fig. 4 shows the water absorption plotted against the square
20 MPa, although none of the WLP concrete samples reached the root of time, and from the developed trend line the rate of water
strength of the control samples. absorption was determined, with the rate of water absorption
The ratio of compressive strength at 28 days for WLP concrete being the slope of the developed trend line. For the control con-
specimens cured according to the wet–dry (WD) procedure, com- crete the rate of water absorption was found to be 0.196 mm/
p p
pared to WLP concrete specimens cured under wet (W) conditions min, while for WLP concrete it was 0.175 mm/ min. The slightly
is presented in Fig. 2, indicating that the strength of WLP concrete lower rate of water absorption for WLP concrete suggests that WLP
can increase by up to 8% when first cured under wet conditions and only slightly improves the resistance of concrete to water and
then dry-cured. This finding is likely due to water evaporation and chemical absorption, potentially due to the polymer membrane
polymerisation of WLP filling pores in the internal concrete struc- forming, although WLP has been shown to reduce moisture pene-
ture [1,8,13], with the WD/W compressive strength ratio increas- trability in concrete as implied by mercury intrusion porosimetry
ing as the WLP dosage increased. results, showing a pore size refinement and lower porosity than
that of conventional concrete [13].
4.1.2. Slump loss The results of the chloride ion penetration test are shown in
The slump loss results of concrete incorporating various dos- Fig. 5, where the chloride content profiles for the control mix and
ages of WLP are presented in Fig. 3. The highest initial slump of WLP concrete mix are presented. A bulk chloride-diffusion coeffi-
200 mm was achieved at 8 L/m3, followed closely by the other cient was calculated for each mix using non-linear regression anal-
three mixes with an initial slump of 190 mm. Generally, the WLP ysis, and was determined to be 9.74  1012 m2/s for WLP
concrete mixes maintained workability more effectively than the concrete, being approximately 10% less than the
control mix, whilst after 120 min the largest slump of 70 mm 10.91  1012 m2/s coefficient for the control mix. It is necessary
was obtained for the 12 L/m3 concrete mix. to note that the Fig. 5 clearly shows the total chloride ion concen-
tration at different depths in the concrete, and that the slightly
4.2. Tensile strength lower chloride ion concentration in the WLP mix is indicative that
porosity may be lower in WLP concrete. These results indicate that
The flexural strength of the control concrete mix after 16 and WLP concrete is potentially more resistant to chloride ion penetra-
28 days wet curing was 3.3 MPa and 3.9 MPa respectively, and tion, especially if reinforcing steel is present. This increased resis-
for the WLP concrete mix the flexural strength was 3.7 MPa and tance to chloride ion penetration may be due to a pore blocking
3.9 MPa respectively. The splitting tensile strength of the control effect of the polymers, and also due to the fine pigments which in-
concrete mix after 16 and 28 days wet curing was 2.9 MPa and crease the chloride ion resistance through mechanisms similar to
3.0 MPa respectively, and for the WLP concrete mix the splitting those of silica fume [3,13,19].
tensile strength was 2.8 MPa and 2.9 MPa respectively. Both the
flexural tensile strength results and the splitting tensile strength 4.4. Xrd, tga/dtg, esem
results indicated that the tensile strength of the control concrete
mix and the WLP concrete mix both increased as the period of Fig. 6 shows the XRD patterns of the control concrete and differ-
wet-curing increased (prior to being tested at 28 days). This finding ent WLP concrete mixes, with no difference in the type of hydra-
was expected for the control mix, but for the WLP concrete mix it tion products identified. All mixes consisted of general hydration
was expected that a period of moist curing for 16 days followed by products usually found, such as ettringite (Ca6Al2(SO4)3(OH)12-
dry curing would result in greater tensile strength, which would 26H2O), portlandite, calcium silicate (Ca2SiO5), and silica (SiO2)
also follow the trend of the compressive strength results. It has which is predominantly attributed to the quartz sand contained
been reported that when flexural strength specimens are left to within the concrete. Because of its semi-crystalline nature, the cal-
dry prior to testing, a tensile stress is induced within the outer skin, cium silica hydrate (C–S–H) gel that formed was hard to identify
weakening the specimens [10]. Based on both tensile tests con- using XRD. However the existence of a peak shown overlapping
ducted, results indicated that WLP does not considerably increase with calcite (CaCO3) at approximately 29o was associated with
N. Almesfer, J. Ingham / Cement & Concrete Composites 46 (2014) 19–25 23

17 1.08

WD / W compressive strength ratio

16 1.06
Compressive strength (MPa)

15 1.04

14 1.02

13 1.00

8L 12L 16L Control 8L 12L 16L

12 0.98
P1 P2 P3 D1 D2 D3 P1 P2 P3 D1 D2 D3

Paint sample Paint sample

(a) 7-day compressive strength results. Fig. 2. Ratio of wet–dry (WD) to wet (W) 28-day compressive strength.


27 200
Compressive strength (MPa)

26 8L
25 12L
Slump (mm)

24 140

8L 12L 16L Control
P1 P2 P3 D1 D2 D3
Paint sample 0 20 40 60 80 100 120
(b) 28-day compressive strength results. Time (min)
Fig. 3. Slump loss of concrete mixtures.

Compressive strength (MPa)

y = 0.1959x + 0.0548
28 R² = 0.9978
Water absorbed (mm 3/mm 2 )


26 2

24 1 y = 0.1751x -0.0532
R² = 0.9975
8L 12L 16L Control
22 1 Control
P1 P2 P3 D1 D2 D3
Paint sample WLP
(c) 56-day compressive strength results. 0
0 5 10 15
Fig. 1. Compressive strength results. Elapsed time (min1/2 )

Fig. 4. Water absorption of concrete mixtures.

the presence of C–S–H gel. The XRD patterns showed little differ-
ence in the quantity of hydration products between the mixes. temperature and the weight loss of the specimen is similar to a
When a concrete specimen is heated to 1000 °C using a thermo- thermogravimetric (TG) curve showing the percentage change in
gravimetric analysis instrument the relationship between the weight as the temperature increases. The first derivative of the
24 N. Almesfer, J. Ingham / Cement & Concrete Composites 46 (2014) 19–25

0.7 Temperature ( oC)

150 250 350 450 550
0.6 0.000
Chloride (% ion conc)



DTG (1/ oC)


Control 0.004
0.0 0.005
0 5 10 15 20 25 30 35 WLP
Penetration Depth (mm)
Fig. 5. Chloride ion concentration in concrete mixtures.
Fig. 7. Differential thermograms (mass loss downwards) of concrete mixtures.

Portlandite Calcium silicate hydrate chemically bound water from C–S–H and ettringite, while peak 2
Calcium silicate Ettringite corresponds to the dehydroxilation/decomposition of portlandite
Silicon oxide (silica) Calcite
[20,21]. The amounts of both chemically bound water and dehy-
droxilation are a measure of the degree of hydration of different
phases in concrete [20,21].
P3 From these DTG results it is clear that concrete with WLP pre-
sents different results than for conventional concrete, especially
P2 regarding the portlandite content. Regarding WLP concrete, the
weight loss related to the dehydration of chemically bound water
P1 (C–S–H) is only slightly less than for the control concrete, while
the weight loss related to dehydroxilation (portlandite) is consid-
D3 erably less. This decrease in portlandite content and hence the
overall calcium/silica ratio of the concrete binder due to the addi-
D2 tion of polymer latex has been confirmed by other studies using TG
and microstructural analysis [12,22], and is caused by the polymer
D1 film partially or fully encapsulating the cement grains and reduc-
Control ing their hydration rate, while the cement hydration process is also
influenced by the presence of the surfactants at the surface of the
polymer particles being adsorbed onto portlandite nucleation sites,
15 25 35 45 55
poisoning the growth of portlandite crystals [23]. It has been re-
ported that the portlandite content can be underestimated by TG
Fig. 6. XRD diffraction patterns of control concrete and WLP concrete mixtures. techniques due to the interaction of polymers with unhydrated ce-
ment particles, the hydrated compounds, or the released hydroxide
ions, resulting in the formation of carbonate-like phases which
TG curve with respect to temperature is a differential thermogravi- interfere with the decomposition and mass loss of portlandite
metric (DTG) curve [20]. DTG results for the control concrete and upon heating [12].
WLP concrete mixes are presented in Fig. 7. DTG results indicate Micrographs of control concrete and WLP concrete are shown in
two peaks of mass loss, at 150–180 °C (peak 1) and at 480– Fig. 8. The micrograph of WLP concrete indicates changes in the
520 °C (peak 2). Peak 1 belongs to the dehydration of free and/or microstructure, especially with regards to the density of the

Fig. 8. FESEM images of control OPC concrete (left) and WLP concrete (right).
N. Almesfer, J. Ingham / Cement & Concrete Composites 46 (2014) 19–25 25

concrete binder phase, C–S–H, possibly attributed to the delay of References

cement hydration due to the paint and hence the C–S–H not being
well defined and becoming granular [24]. Large crystals of portlan- [1] Ohama Y. Polymer-based admixtures. Cem Concr Compos 1998;20(2–
dite, slender ettringite needles, and different types of C–S–H were [2] Van Gemert D, Czarnecki L, Maultzsch M, Schorn H, Beeldens A, Lukowski P,
easily identified in the control concrete and all WLP concrete et al. Cement concrete and concrete-polymer composites: two merging
mixes. The polymer film in the WLP concrete samples was not dis- worlds: a report from 11th ICPIC Congress in Berlin, 2004. Cem Concr
Compos 2005;27(9–10):926–33.
tinguishable, indicating that it is almost entirely engulfed by, or [3] Nehdi M, Sumner J. Recycling waste latex paint in concrete. Cem Concr Res
incorporated into the C–S–H phase and/or other products. 2003;33(6):857–63.
[4] Gretz M, Plank J. An ESEM investigation of latex film formation in cement pore
solution. Cem Concr Res 2011;41(2):184–90.
5. Conclusions [5] Aggarwal LK, Thapliyal PC, Karade SR. Properties of polymer-modified mortars
using epoxy and acrylic emulsions. Constr Build Mater 2007;21(2):379–83.
[6] Ohama Y. Concrete-polymer composites – the past, present, and future. Key
Fresh, mechanical, and durability properties of concrete incor- Eng Mater 2011;466(1):1–14.
porating WLP at various proportions were investigated in this [7] Yang Z, Shi X, Creighton AT, Peterson MM. Effect of styrene–butadiene rubber
latex on the chloride permeability and microstructure of portland cement
study, with this work being motivated by the fact that there is an mortar. Contr Build Mater 2009;23(6):2283–90.
available resource of WLP in New Zealand, and that there is poten- [8] Beeldens A, Van Gemert D, Schorn H, Ohama Y, Czarnecki L. From
tial for WLP to be used in concrete. The use of WLP in concrete has microstructure to macrostructure: an integrated model of structure
formation in polymer-modified concrete. Mater Struct 2005;38(6):601–7.
been found to improve workability based on slump loss, but to de- [9] Lewis WG, Lewis G. The influence of polymer latex modifiers on the properties
crease the compressive strength. The compressive strength of con- of concrete. Composites 1990;21(6):487–94.
crete utilizing WLP at a dosage of 12 L/m3 was found to satisfy [10] Lamond JF, Pielert JH. Significance of tests and properties of concrete and
concrete-making materials. New Jersey: ASTM International; 2006. p. 605–15.
compressive strength requirements, whilst a performance-based [11] Almesfer N, Haigh CJ, Ingham J. Waste paint as an admixture in concrete. Cem
curing regime that allowed the WLP to dry led to a slight increase Concr Compos 2012;34(5):627–33.
in the compressive strength of concrete utilizing WLP. WLP did not [12] Knapen E, Van Gemert D. Cement hydration and microstructure formation in
the presence of water-soluble polymers. Cem Concr Res 2009;39(1):6–13.
significantly affect the tensile strength of concrete, and slightly en-
[13] Mohammed A, Nehdi M, Adawi A. Recycling waste latex paint in concrete with
hanced the durability characteristics of concrete. added value. ACI Mater J 2008;105(4):367–74.
The concrete phase analysis indicated that the addition of WLP [14] Godfrey C. The reuse of latex paint as a polymeric admixture in concrete
does not cause the formation of any new products, but rather alters (MEng Dissertation). Bath: University of Bath; 2008.
[15] SNZ. NZS 3112 – Methods of test for concrete: Standards Association of New
the quantity of hydrates, especially portlandite. The microstructure Zealand, 1986.
of WLP concrete was not as dense as the microstructure of conven- [16] Astm, C, 1585–04 – Measurement of rate of absorption of water by hydraulic-
tional concrete, possibly due to the cement hydration process cement concretes: ASTM, International, 2004.
[17] Astm, C, 1556–04 - Determining the apparent chloride diffusion coefficient of
being retarded and not coagulating as much. Further work is re- cementitious mixtures by bulk diffusion: ASTM, International, 2004.
quired on the exact position of the polymers and pigments within [18] SNZ. NZS 3104 – Specification for concrete production: Standards Association
the cement matrix, and their effects on the porosity of concrete. of New Zealand, 2003.
[19] Ismail M, Muhammad B, Yatim JM, Noruzman AH, Soon YW. Behaviour of
Based on results from this study, it appears that concrete with concrete with polymer additive at fresh and hardened states. Procedia Eng
12 L/m3 of WLP is appropriate for use in non-structural 20 MPa 2011;14:2230–7.
applications such as sidewalks and driveways. [20] Chang C, Chen J. The experimental investigation of concrete carbonation
depth. Cem Concr Res 2006;36(9):1760–7.
[21] Ramachandran VS. Thermal analysis of cement components hydrated in the
presence of calcium carbonate. Thermochimica Acta 1988;127:385–94.
Acknowledgements [22] Almeida AE, Sichieri EP. Thermogravimetric analysis and mineralogical study
of polymer modified mortar with silica fume. Mater Res 2006;9(3):321–6.
This research study was made possible by 3R Group Ltd, espe- [23] Young J. A review of the mechanisms of set-retardation in portland cement
pastes containing organic admixtures. Cem Concr Res 1972;2(4):415–33.
cially Graeme Norton and Gareth Mentzer, the New Zealand Min-
[24] Su Z, Sujata K, Bijen J, Jennings H, Fraaij A. The evolution of the microstructure
istry of Science and Innovation, Resene Paints Ltd, and Firth styrene acrylate polymer-modified cement pastes at the early stage of cement
Industries Ltd. hydration. Adv Cem Based Mater 1996;3(3–4):87–93.