Construction and Building Materials 36 (2012) 205215

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Durability of mortar using waste glass powder as cement replacement

Ana Mafalda Matos a,, Joana Sousa-Coutinho a,b
a
b

LABEST FEUP, Faculty of Engineering, University of Porto, Portugal

FEUP, Department of Civil Engineering, Faculty of Engineering, University of Porto, Portugal

h i g h l i g h t s
" Glass powder is thoroughly analysed as a partial cement replacement material in mortar.
" Testing is carried out far beyond the usual mechanical testing and ASR by most authors.
" Chloride ingress, carbonation, sulphate attack and sorptivity are also considered for durability assessment.
" SEM demonstrated effect of ne glass particles well encapsulated into a dense matrix.
" This holistic study conrms that glass powder contributes to sustainability in construction.

a r t i c l e

i n f o

Article history:
Received 21 August 2011
Received in revised form 28 March 2012
Accepted 25 April 2012
Available online 23 June 2012
Keywords:
Waste
Glass powder
Mortar
Addition
Durability
Pozzolanic activity

a b s t r a c t
It is well known that Portland cement production is an energy-intensive industry, being responsible for
about 5% of the global anthropogenic carbon dioxide emissions worldwide. An important contribution to
sustainability of concrete and cement industries consists of using pozzolanic additions, especially if
obtained from waste such as waste glass.
Crushed waste glass was ground (WGP) and used in mortar as a partial cement replacement (0%, 10%
and 20%) material to ascertain applicability in concrete.
An extensive experimental program was carried out including pozzolanic activity, setting time, soundness, specic gravity, chemical analyses, laser particle size distribution, X-ray diffraction and scanning
electron microscopy (SEM) on WGP and resistance to alkali silica reaction (ASR), chloride ion penetration
resistance, absorption by capillarity, accelerated carbonation and external sulphate resistance on mortar
containing WGP.
Glass particles well encapsulated into dense and mature gel observed by SEM, may help explaining
enhanced durability results and thus conrming that waste glass powder can further contribute to sustainability in construction.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
Concrete is the most commonly used construction material in
the world and it is the second most consumed product on the
planet after water [1].
Todays annual global cement production has reached 2.8 billion tons, and is expected to increase to some 4 billion tons per
year. At the same time, the cement industry is facing challenges
such as cost increases in energy supply, requirements to reduce
CO2 emissions and the supply of raw materials in sufcient qualities and amounts [2]. It is estimated that about 0.91.0 tons of CO2
are produced for a ton of clinker depending on the type of fuels
used [3].
Corresponding author. Address: Rua do Dr. Roberto Frias, 4200-465 Porto,
Portugal. Tel.: +351 225081936; fax: +351 225081441.
E-mail address: [email protected] (A.M. Matos).
0950-0618/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2012.04.027

While cement production in its beginnings only focused on ordinary Portland cement, later cements with several main constituents were produced by replacing parts of the clinker content by
supplementary cementitious materials [3].
Ecological or environmental benets of alternative supplementary materials include (1) the diversion of non-recycled waste from
landlls for useful applications, (2) the reduction in the negative
effects of producing cement powder, namely the consumption of
non-renewable natural resources, (3) the reduction in the use of
energy for cement production and (4) the corresponding emission
of greenhouse gasses [4].
Solid industrial by-products, such as siliceous and aluminous
materials, as well as some natural pozzolanic materials are increasingly being used in the cement and concrete industry. The incorporation of these materials in concrete has been giving encouraging
results regarding the mechanical and durability properties of concrete [5].

206

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

Despite the existence of large amounts of industrial waste such

as blast furnace slag, y ash, silica fume, slag and agricultural residues such as rice husk ash, which have been used for many years
in large amount as raw materials and components in the cement
industry, there are still many other industrial wastes not used
yet [7], such as waste glass. In fact a few studies have successfully
investigated the potential use of nely ground waste glass as a pozzolanic material focusing on strength development and alkali-silica
reactivity [918,2023].
Glass is the result of the merger of several inorganic mineral
raw materials, which after undergoing a process of controlled cooling becomes a hard, homogeneous, stable, inert, amorphous and
isotropic material [8]. Based on the major composition, glass can
be classied into several categories but soda-lime glass is the most
widely used to manufacture containers, oat and sheet glass and
therefore composing over 80% by weight of waste glass [9].
With the exception of Al2O3 and CaO, the percentages of the
main constituents of different types of glass are similar [4,6]. For
soda lime glass the typical glass composition is approximately
70% of silica, 1317% Na2O and 10% CaO [9].
In fact, glass can be considered a pozzolanic-cementitious material according to the chemical requirements in ASTM C 618 [33] if
the alkali content is disregarded.
Alkalis can cause alkali-aggregate reaction and expansion if
aggregates are alkali-reactive. Therefore high alkali content of glass
is a typical concern for its use in concrete but studies have shown
that nely ground glass does not contribute to ASR [15,20,21]. The
pozzolanic properties of glass are rst notable at particle sizes below approximately 300 lm. Below 100 lm, glass can have a pozzolanic reactivity which is greater than that of y ash at low percent
cement replacement levels and after 90 days of curing [5]. In fact,
according to Shi and Zheng [9], ground glass powders exhibit very
good pozzolanic reactivity and can be used as cement replacement.
As expected its pozzolanic activity increases as neness increases
[9]. The combined use of other supplementary cementing materials
such as coal y ash, ground blast furnace slag and metakaolin can
also decrease the expansion from ASR. Lithium salt can be also very
effective to prevent ASR expansion of concrete containing glass
powders [10]. In any case compressive strength is often lower
when waste glass powder is used as cement replacement, especially at early ages [15,18,20].

The environmental and economic benets from the reuse of

recycled waste glass in cement and concrete production can also
be very signicant depending on the end uses and production scale
[9].
Therefore the present work besides the usual strength and ASR
testing encompasses a wider range of analysis of effect of waste
glass powder, including pozzolanic activity, setting time, soundness, specic gravity, chemical analyses, laser particle size distribution, X ray diffraction and scanning electron microscopy (SEM),
chloride ion penetration resistance, absorption by capillarity,
accelerated carbonation and external sulphate attack. This holistic
study corroborates that glass powder as a partial replacement
material is a waste material that contributes to sustainability in
construction.

2. Experimental program
2.1. Materials
Typical commercial Type I 42.5R Portland cement was used and commercial silica fume (SF) was employed as a reference pozzolanic material. Waste glass was obtained from a recycling glass industry in Portugal where waste glass such as car
windscreens is crushed and sold to the bottle industry. Fine waste glass from this
industry was ground during 48 h in a ball mill in the laboratory (WGP). Fig. 1 shows
particle size distribution of cement and WGP. X-ray diffraction on WGP showed that
no substantial crystalline phases were detected. Soundness and setting time tests
(NP EN 196-3 [36]) were carried out on pastes with 100% CEM I 42.5R (CTL) and
on paste with 10% (WGP10) and with 20% (WGP20) cement replacement with
WGP. Pozzolanicity test was carried out following NP EN 196-5 [37] on a sample
with 90% of cement and 10% of WGP. Chemical composition and physical characteristics of WGP and cement are given in Table 1. Scanning Electron Microscopy (SEM)
was carried out on CEM I 42.5R, silica fume and WGP samples shown in Fig 2.

2.2. Mortar production and workability

Four mortar types were prepared following the procedure described in NP EN
196-1 [35] a control mix with 100% cement (CTL), mixes with 10% (WGP10) and
20% (WGP20) of waste glass powder as cement replacement and a mix with 10% cement replacement with silica fume (SF). CEN Reference sand for mortar production
and super plasticizer (SP) SikaVisocrete 3000 complying with NP EN 2 [41] were
used.
Mortar workability (measured according to ASTM C 109/90 [31] and ASTM C230
[32]) increased with cement replacement with WGP, but SF mortar showed a significant workability decrease so super plasticizer was added. Mix proportions and
workability results are shown in Table 2.

Fig. 1. Particle size distribution of cement and WGP.

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

Table 1
Chemical and physical characteristics of cement and glass.
CEM I 45.2R
Chemical composition (% mass)
LOI
2.61
Insoluble residue
1.33
SiO2
20.36
5.1
Al2O3
Fe2O3
3.12
CaO
62.72
MgO
1.81
Na2O
K2O
Na2O eq
SO3
3.44
Cl
0012
Free lime
1.62
Pozzolanic activity
Physical properties
Specic gravity (g/cm3)
Soundness (mm)
Initial setting time
WGP10/WGP20
CTL
Final setting time
WGP10/WGP20
CTL

3.16
0

WGP

Silica fume

0.92
70
1.2
0.65
8.7
3.7
16
0.35
17
<0.05
<0005

>99

Positive

Positive

2.68
0

2.20
0

207

calculation of the chloride diffusion coefcient (Dns) from an accelerated test. An

external electrical potential is applied axially across the specimen and forces the
chloride ions outside to migrate into the specimen. After a certain test duration,
the specimen is axially split and a silver nitrate solution is sprayed on to one of
the freshly split sections. The chloride penetration depth can then be measured
from the visible white silver chloride precipitation, after which the chloride migration coefcient can be calculated from this penetration depth as follows:

p
RTL xd  a xd
:
ZFU
t

r
RTL
e
ZFU

Dns

a2

e erf 1 1 

2C d
C0


5

where Dns is the apparent diffusion coefcient obtained in a non-steady state migration test (cm2/s), R the gas constant R = 8.314 J/(mol K), T the absolute temperature
(K), L the thickness of specimen (cm), Z the ion valence, F the Faraday constant,
F = 9.648  104 J (V mol.), U the effective voltage applied (V), xd the depth of chloride
penetration measured by using a colorimetric method (cm), t the time of test duration (s), a the laboratory constant, e = 0.764 if external chloride concentration of
0.5 M. Cd the concentration of free chloride at which the colour changes when using
the colorimetric method to measure the chloride penetration depth (kg Cl/m3 solution) and C0 is the concentration of free chloride in the external solution.
Results can be seen in Fig. 7.

2 h, 20 m/2 h, 15 m
2 h, 30 m
3 h, 25 m/3 h, 25 m
3 h, 25 m

2.6. Absorption by capillarity

Several test specimens were produced. After demoulding the following day, test
specimens were cured in water at 20 C in a fog room until testing.
2.3. Strength
Flexural and compressive strength testing was undertaken at 7, 28, 90 and
180 days (and 562 days for CTL and WGP10) following the standard procedure in
NP EN 196-1 [35]. Results are presented in Fig. 3. The pozzolanic effect of glass powder may be assessed with the Activity Index (AI) (Fig. 4) which is the ratio between
strength of WGP containing mortar and strength of equivalent control mortar at the
same age.

Absorption by capillarity testing was undertaken on two months old cast test
specimens, approximately 50 mm diameter by 100 mm length generally following
standard procedure RILEM TC 116-PCD [45]. The samples were put to dry in a ventilated heater at 40 C until constant mass. For the test itself, test specimens were
placed moulded face downwards, in a shallow water bath. Water level was adjusted
automatically so that the formwork face was dipped to a depth of approximately
3 mm. During the test, water was drawn into the core by capillary forces and
weighed at time intervals up to 4.5 h from the start of the test [45].
The absorption of water into concrete under capillary action is dependent on
the square-root of time and may be modelled [24] by the following equation:

A A0 St 0;5

6
2

2.4. Alkali silica reaction (ASR)

The potential risk of alkali-silica reaction in concrete was monitored in accordance with the accelerated mortar bar test following ASTM 1567 [34], where two
mortar prisms with 25  25  250 mm were used for each mortar type.
In this test method, higher temperature (80 C) and increased alkalinity (1 N
NaOH) accelerates the reaction. ASTM C 1567 method is not intended to capture
the effect of increased alkalis from glass powder; rather it is used in this study to
understand the expansion characteristics of plain and modied mortar. In fact this
test method can be used to evaluate the effectiveness of supplementary cementing
materials in reducing the expansion due to ASR [17].
Two slight modications were performed when following the procedure in
ASTM. Instead of 1:2.25:0.47 cement:aggregate:water ratio required in ASTM, the
ratio used was 1:3:0.5 which is considered in NP EN 196-1 [35]. ASTM C 1567 also
requires particle size distribution of (reactive) aggregates shown in Table 3, but the
particle size distribution used was the one considered in NP EN 196-1 CEN sand,
which was found to be reactive. ASTM C 1567 was followed for the remaining procedure, so immediately after casting, the moulds were covered and demoulded 24 h
later. Then they were preconditioned for a further 24 h in water maintained at
80 C. The lengths of these mortar bars after immersion in hot water, measured
along the 4 faces of each specimen, were then taken as the initial readings (L0).
The mortar bars were subsequently transferred to 1 N NaOH solution maintained
at 80 C and periodically measured for 14 days.
Therefore the actual length (Lx) of each specimen, on day x is taken as

P4

Lx

i1 Li

The expansion of each specimen at day x is given by

Lx  L0
 100%
250

Expansion with time is shown in Fig. 5 as well as nal expansion in Fig. 6.

where A (mg/mm ) is the water absorption by unit area of concrete surface since the
moment the core was dipped in water, S is the sorptivity of the material, t is the
elapsed time and A0 (mg/mm2) is the water absorbed initially by pores in contact
with water. The above equation was found to provide a very good t to the data with
correlation coefcients of over 0.96 (Fig. 8). The average sorptivity value of each
mortar type is shown in Fig. 9.
2.7. Carbonation
Resistance to carbonation was assessed in accordance with the procedure described RILEM CPC-18 [44] where test specimens were exposed to 5 0.1% carbon
dioxide, relative humidity (RH) of 60 5% and temperature of 23 3 C, in an accelerated carbonation chamber.
Carbonation depth was evaluated on three test specimens for each mortar type,
water cured at 20 C followed by 2 and 4 months in the chamber. After splitting the
specimens, the surface was cleaned and sprayed with a phenolphthalein pH indicator. In the noncarbonated part of the specimen, where the mortar was still highly
alkaline, a purple-red colour was obtained. In the carbonated part of the specimen
where the alkalinity of mortar is reduced, no coloration occurred. Each value corresponds to the mean of twelve measurements taken around the four sides of the
freshly split and sprayed surface of each test specimen. Results can be seen in Fig. 10.
2.8. External sulphate attack
Resistance to external sulphate attack was evaluated according to the Portuguese standard E-462 [42], on six (1, 2, 3, 4, 5 and 6) mortar prisms of
16  16  160 mm. Test specimens were immersed in calcium hydroxide solution
during 28 days, measured along the 4 side faces of each specimen and taken as initial readings (L0). Test specimens 2, 4 and 6 were transferred to a sodium sulphate
solution and the others were maintained in calcium hydroxide saturated solution.
Readings were taken throughout 26 weeks and sulphate solution was renewed
every 2 weeks. The actual length increase of each specimen on day x is taken as:

2.5. Chloride ion diffusion

ExpCaOH2 x

Lx  L0
1600

The resistance to chloride permeability was evaluated by the CTH Rapid Method
according NT Build 492 [44] which is a non-steady state migration method based on
a theoretical relationship between diffusion and migration which enables the

ExpSO4 Na2 x

Lx  L0
1600

208

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

CEM I 42.5 R 200 times enlarged

CEM I 42.5 R 1000 times enlarged

CEM I 42.5 R 5000 times enlarged

GP 500 times enlarged

GP 2000 times enlarged

GP 5000 times enlarged

SF 10000 times enlarged

SF 20000 times enlarged

SF 40000 times enlarged

Fig. 2. Scanning electron microscopy on powder samples.

2.9. Scanning electron microscopy

Table 2
Mixture proportions and workability of mortar.
Materials/mortar

CTL

SF

WGP10

WGP20

CEM I 42.5R (g)

SF (g)
WGP (g)
Sand (g)
Water (ml)
SP/C ratio (%)
W/binder ratio
Workability (mm)

450
0
0
1350
225
0
0.5
200

405
45
0
1350
225
0.34
0.5
198

405
0
45
1350
225
0
0.5
204

360
0
90
1350
225
0
0.5
206

3. Discussion
3.1. Chemical and physical properties of WGP

The expansion for each mortar is taken as:

ExpCaOH2

Exp1 Exp3 Exp5

3

ExpSO4 Na2

Exp2 Exp4 Exp6

3

10

The expansion due to sulphate is:

Expansionx ExpSO4 Na2 x  ExpCaOHk2 x

Expansion during 26 weeks is shown in Fig. 11.

SEM was also carried out on samples CTL and WGP10 mortar samples cured in
water at 20 C for 562 days, as shown in Fig. 12, for 2500, 5000, 10,000 and 20,000
magnication in order to investigate changes in the cement hydrated micro
structure.

11

Liu compared composition of soda lime glass in general to y

ash, stating that the SiO2 and (Na2O + K2O) contents are much
higher in glass and (SiO3 + Al2O3 + Fe2O3) is similar [6]. Chemical
composition of WGP is in accordance with those values and with
results by other authors using soda lime glass waste in mortar
and concrete [1418].
Considering that no specic European standard covers material
such as WGP and that it is mainly composed of silica, WGP properties were compared in Table 4, to requirements for silica fume (NP
EN 13263-1 [38]), also for y ash (NP EN 450-1 [39]) and slag (NP
EN 15617-1 [40]). As can be observed glass powder is basically in
accordance with these requirements, except for Na2O content

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

209

Fig. 3. Compressive and exural strength.

Fig. 4. Activity index at 28 and 90 days.

Table 3
Required grading requirements in ASTM C 1567 and grading for CEN sand used.
Grading requirements ASTM C 1567

CEN sand

Retained on sieve size

Mass (%)

Retained on sieve size

Mass (%)

2.36 mm
1.18 mm
0.60 mm
0.30 mm
0.15 mm

10
25
25
25
15

2,00 mm
1.60 mm
1.00 mm
0.50 mm
0.16 mm

0
7
26
34
20

(No.
(No.
(No.
(No.
(No.

8)
16)
30)
50)
100)

with WGP, respectively, compared to control paste. Final setting

time was similar for all types of mortar.
Total CaO was determined but, as free CaO was unknown,
soundness testing was carried out with results well under standard
limits, as can be seen in Table 1.
SEM observations (Fig. 2) indicate that WGP particles seem
more angular, denser and more prismatically shaped compared
to cement and naturally much larger than spherical silica fume
particles. Fineness obtained through laser particle size distribution,
is similar for WGP and cement (Fig. 2).
3.2. Strength and activity index

(17%), far above the limit (5%) imposed in the standard for y ash.
This corroborates Shao et al [12] who observed general accordance
with ASTM C 618 [33] but not meeting the optional requirement
for the alkali content, causing concern due to potential ASR.
Initial setting time (Table 1) was marginally shorter for WGP10
(10 and 20 min) for paste with 10% and 20% cement replacement

Mechanical and durability related properties for WGP were

compared to those of CTL and SF mortar for the present study as
well as to results by other authors.
Strength for WGP mortar was lower than control at 7 and 28
days, decreasing with replacement dosage but at 90 days generally

210

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

Fig. 5. Expansion vs. time due to ASR.

Fig. 6. Final expansion of mortar after 14 days in 1 N NaOH.

attaining CTL although under SF levels. WGP20 gained signicant

strength between 28 and 90 days showing pozzolanic reaction taking place in this period. Idir et al [26] for ne glass with particle size
range under 41 lm and mean particle size of 7.8 lm obtained similar results concluding that the general trend is for replacement of
cement by glass to lead to a decrease in compressive strength [26].
Strength activity indexes obtained were higher than 90% at 90
days of curing for 10% and 20% replacement by ne glass powder,
conrming results obtained by Jin et al [25]. Shao et al [12], used
higher replacement (30%) by ne glass powder (<38 lm) in concrete and obtained concrete strength activity indexes higher than
90% at 90 days. Shi et al, for 25% replacement and coarser material
(<100 lm) obtained strength activity values around 74% at 28 days,
with which the results of the present study for 20% replacement
seem to be in accordance.
3.3. Alkali silica reaction
Results obtained for ASR expansion respecting control mortar
indicated potential deleterious expansion according to ASTM
1567 (higher than 0.10% after 14 days in NaOH). Results for both

SF and WGP were effective in reducing ASR expansion. Increasing

glass powder content led to lower expansion despite the high alkali
content. Frederico and Chidiac [4] reported results on ASR expansion relating to particle size of glass and cement replacement dosage including work by Jin et al [25] and Shao et al [12] with which
present results are in accordance.
ASR results in the present study also conrm ndings by Schwarz et al [20] where expansion for 10% and 20% replacement with
glass powder was slightly higher, presumably due to use of coarser
glass particles. Sacanni and Bizigoni [21], where ne aggregate was
also replaced by glass particles, obtained ASR expansion similar to
values in the present study.
Therefore ASR testing conrmed that glass powder assisted in
hindering expansion compared to control specimens, conrming
conclusions by other authors [2022].
The possible explanations for this phenomenon according to
Taha and Nounu [22] is that available reactive silica in glass powder as an amorphous material will dissolve very quickly during the
pozzolanic reaction and react with other chemicals to form the
mineral phases of concrete. Therefore, the dissolved reactive silica
of glass powder will be accommodated in the crystals of concrete

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

211

Fig. 7. Chloride diffusion coefcient for each mortar type.

Fig. 8. Linear regression for absorption by capillarity during 4 h 30 min on test cores.

minerals and not available for the ASR which occurs in later stages.
With the presence of other chemicals the dissolved reactive silica
of glass powder will not be able to form ASR gel during the pozzolanic reaction, or even a gel having a swelling property likewise the
ASR gel. Otherwise, if there was any chance to form ASR gel during
the pozzolanic reaction from the reactive silica of glass powder one
can envisage that any existing ASR expansion would be noticeable.
If any ASR could occur during the pozzolanic reaction stage, the
produced expansion would vanish and accommodate within the
plasticity of fresh concrete. On the other hand, in the rst 4 weeks
of the pozzolanic reaction and hydration process, most of the alkalis would have been consumed in order to catalyse the pozzolanic
reaction and act as chemical activators to compensate for the lower
calcium content in the glass powder. Therefore, in later age, there
would not be enough free alkali to interact with the reactive silica
in the pore solution of the concrete to induce ASR, as illustrated in
Fig. 13.

Additions with low CaO content act as ASR inhibitors, whereas

those with more than 10% CaO content act as ASR promoters [27]
and therefore WGP proved once again to inhibit ASR.

3.4. Chloride diffusion, absorption by capillarity, carbonation and

sulphate attack
CTH rapid method results show that using WGP as a partial cement replacement drastically enhances resistance to chloride penetration compared to control mortar specially for 20% replacement
dosage (on average, CTL Dns is about double compared to WGP20)
and even compared to 10% SF mortar. Samples of silica fume and
WGP10 have very similar diffusion coefcients. Results obtained
in the present study, conrm values by other authors [16,19,20],
where performance of glass powder mortar is better than control
mortar regarding chloride penetration. Only Jain and Neithalath

212

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

Fig. 9. Sorptivity of each mortar type.

Fig. 10. Carbonation depth results after 2 months curing followed by 4 months in the chamber.

[13] carried out the NT Build 492 [43] test as in the present study
and obtained similar improvement to WGP 20 at 56 days.
Absorption by capillarity and sorptivity proved to be similar for
WGP and CTL mortar and a little lower for SF mortar. This situation
may be explained by similar neness of both WGP and cement (see
Fig. 9). Much ner SF particles physically ll capillary pores reducing capillary absorption. Other authors, such as Taha and Nounu
[23] also showed similar results for concrete with cement replacement materials, but following BS 1881.
Carbonation depth for all blended cement mixtures was greater
than for the Portland cement mixture. It was found that carbonation increased along with the increase in WGP content, which is
consistent with the trend observed in concrete for various pozzolanic materials and probably due to CH reduction. WGP10 carbonated almost half of SF10 mortar, albeit pore renement due to

silica fume. This may be explained by higher SF reactivity and thus

further CH reduction, compared to WGP. When WGP replacement
is doubled, carbonation depth is similar to SF. Although a less reactive material, higher WGP replacement and less cement in the mixture, both contribute to CH reduction levelling out intense CH
reduction in SF mortar.
Sulphate expansion at 26 weeks should be under 0.10% according to the Portuguese standard E-462 and therefore Portland cement used is not sulphate resistant. Blended Portland cement
with 10% replacement with WGP showed an impressive resistance
to sulphate attack, far higher than SF marginally within the limit of
0.10%. The pozzolanic activity of WGP and SF binds portlandite
(CH) released in the hydration of calcium silicates (C3S and C2S),
so CH is no longer available for reaction with sulphates, in
accordance with [28,29]. This prevents the formation of gypsum.

213

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

Fig. 11. Expansion during 26 weeks due to sulphate.

CTL 562 days 2500 times enlarged

CTL 562 days 5000 times enlarged

CTL 562 days 10000 times enlarged

CT L 562 days 20000 times enlarged

WGP10 562 days 2500 times enlarged

WGP10 562 days 5000 times enlarged

WGP10 562 days 10000 times enlarged

WGP 562 days 20000 times enlarged

Fig. 12. SEM on 562 days old mortar.

Pozzolanic reaction produces a secondary CSH that also decreases the capillary porosity of mortar and enhances signicantly
the pasteaggregate interface [28].
3.5. Scanning electron microscopy
Regarding SEM observations, glass particles seem to be thoroughly encapsulated and dispersed into to the hydrated products
of a dense and mature gel with needle shaped ettringite crystals
(see WGP10 10,000 and 20,000 times enlarged, Fig. 12). CH layered
crystals were observed in the CTL mortar (see CTL 5000 times enlarged, Fig. 12) and CSH gel lling the hydrated structure. CSH gel

in WGP10 was found to contain more calcium as well as more alkalis compared to CTL mortar. These ndings are in accordance with
Sobolev et al [30] who suggested that the main difference between
glass containing mortar and reference Portland cement is related to
the decrease in size and the amount of CH, caused by consumption
of CH as a result of the pozzolanic reaction involving glass grains.
More alkalis found in the gel of WGP10 compared to CTL also conrm ndings by Shayan and Xu [17] who proposed that the pozzolanic reaction of glass powder with cement causes the binding of
alkalis in the paste, making it unavailable for ASR [17] and therefore
in accordance with the mechanism proposed by Taha and Nounu
[22] explaining lower expansion due ASR when glass is used.

214

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

Table 4
Requirements in standards.
European Standards

Chemical requirements

Physical requirements

Fly ash (NP EN 450-1)

Silica fume
(NP EN
13263-1)

Slag (NP EN 15167-1)

GP

Remarks

Cl
SO4
SO3
Free CaO

60.10%
63.0%

62.5%

60.3%
62.0%

61.0%

60.10%
62.5%
62.0%

<0005

<0.05
8.70%

OK
OK

Reactive CaO
SiO2
SiO2 + Al2O3 + Fe2O3
CaO + MgO + SiO2
NaO2eq
LOI

610%
P25%
P70%

P85%
P67.0%

Pozzolanicity

65.0%
Class A 6 5%
Class B 27%
Class C 49%

Activity index

P75% at 28 days

Start setting time

Soundness

70%
71.82%

P85% at 90 days
6120 min plus start time of
the paste-setting reference
610 mm

64.0%

63.0%

17%
0.92%

Only the total content of CaO

is known but soundness is OK
OK
OK
OK
KO, but OK in ASR testing
OK

P100% at
28 days old

Positive

OK

P45% at 7 days and P75%

at days

KO

Almost OK for 28 days, and ok

for 90 days

62 times the start time of

the paste-setting reference

OK

610 mm

Fig. 13. Model illustrates the consumption of alkali in the pozzolanic reaction in the
rst 4 weeks [22].

OK

Higher resistance to chloride penetration was obtained for WGP

containing mortar, increasing with dosage replacement;
Sorptivity of glass mixes was equivalent to control due to the
effect of similar particle size distribution;
Carbonation depth for all blended cement mixtures was greater
than for the Portland cement mixture, increasing with replacement dosage in the trend for pozzolanic materials;
WGP10 mortar showed a remarkable resistance to sulphate
attack, far higher than SF;
SEM observations may help explaining these ndings as glass
particles seem to be completely encapsulated and dispersed
into dense and matured gel.
Although soda lime glass presents a high alkali content, use of
ground waste glass as cement replacement in mortar, improved
resistance to ASR and chloride penetration with replacement
dosage and greatly improved sulfate resistance without compromising strength. Therefore this waste material can successfully enhance durability and further contribute to sustainability
in construction

4. Conclusions

5. Future developments

This study aimed to evaluate use of glass powder in cement

based materials. An experimental study on the mechanical and
durability properties of mortar containing waste glass powder as
cement replacement provided the following results:

Future research comprising use of glass powder is underway,

including use in self-compacting concrete, taking advantage of increased workability provided by this waste material.

In terms of physical and chemical characteristics, glass can be

considered a pozzolanic-cementitious material according to
requirements in ASTM C 618 [33] as well as in European standards [3840], if the alkali content is disregarded;
Attractive strength activity indexes were obtained at 90 days of
curing in WGP containing mortar, with signicant increase
between 28 and 90 days which means pozzolanic activity
occurred;
In spite of high alkali content in WGP, ASR expansion was drastically reduced. The explanation of this phenomenon can be
referred to the consumption of the alkali in the structure of
the CSH gel during the pozzolanic reaction which occurs earlier than ASR, therefore not leaving enough alkali to induce ASR;

Acknowledgement
The authors would like to thank Mr. Oliveira from the glass
industry, LEMC, LABEST and to FCT under Project PTDC/ECM/
098117/2008.
References
[1] Mehta PK. Greening of the concrete industry for sustainable development.
Concr Int 2002;24:238.
[2] Schneider M, Romer M, Tschudin M, Bolio H. Sustainable cement production
present and future. Cem Concr Res 2011;43:64250.
[3] Ali MB, Saidur R, Hossain MS. A review on emission analysis in cement
industries. Renew Sustain Energy Rev 2011;15:225261.

A.M. Matos, J. Sousa-Coutinho / Construction and Building Materials 36 (2012) 205215

[4] Frederico LM, Chidiac SE. Waste glass as a supplementary cementitious
material in concrete critical review of treatment methods. Cem Concr
Compos 2009;31:60610.
[5] Papadakis VG, Tsimas S. Supplementary cementing materials in concrete Part
I: efciency and design. Cem Concr Res 2002;32:152532.
[6] Miao Liu. Incorporing ground glass in self-compacting concrete. Constr Build
Mater 2011;25:91925.
[7] Uchikawa Horishi, Hanehara Shunsuke. Recycling of waste as an alternative
raw material and fuel in cement manufacturuing. In: Waste materials used in
concrete manufacturing. Noyes publications: United States of America; 1997.
p. 430551.
[8] Ortiz SA, O Vidro e as Embalagens de Vidro. Campinas, Brasil: CETEA/ITAL;
1996.
[9] Shi Caijun, Zheng Keren. A review on the use of waste glasses in the production
of cement and concrete. Resour Conserv Recycl 2007;52:23447.
[10] Byars EA, Morales-Hernandez B, Zhu HY. Waste glass as concrete aggregate
and pozzolan. Waste glass as concrete aggregate and pozzolan: laboratory and
industrial projects. Concr Soc 2004;38(1):414.
[11] Archibald JF, DeGagne DO, Lausch P, De Souza EM. Ground waste glass as a
pozzolanic consolidation agent for mine backll. CIM Bull 1995;88(807):20.
[12] Shao Y, Lefort T, Moras S, Rodriguez D. Studies on concrete containing ground
waste glasses. Cem Concr Res 2000;30:91100.
[13] Dyer TD, Dhir RK. Chemical reactions of glass cullet used as cement
component. J Mater Civ Eng 2001;13(6):4127.
[14] Shayan A, Xu A. Value-added utilisation of waste glass in concrete. Cem Concr
Res 2004;34:819.
[15] Shi C, Wu Y, Shao Y, Rieer C. Alkali-aggregate reaction of concrete containing
ground glass powder. In: Proceedings of the 12th international conference on
AAR in, concrete; 2004. p. 78995.
[16] Shi C, Wu Y, Rieer C, Wang H. Characteristics and pozzolanic reactivity of
glass powders. Cem Concr Res 2005;35:98793.
[17] Shayan A, Xu A. Performance of powder as pozzolanic materials in concrete: a
led trail on concrete slabs. Cem Concr Res 2006;36:45768.
[18] Nathan Schwarz, Narayanan Neithalat. Inuence of a ne glass powder on
cement hydration: comparison to y ash and modeling the degree of
hydration. Cem Concr Compos 2008;30:48696.
[19] Jain Jitendra A, Neithalath Narayanan. Chloride transport in y ash and glass
powder modied concretes inuence of test methods on microstructure.
Cement Concr Compos 2010;32:14856.
[20] Nathan Schwarz, Hieu Cam, Narayanan Neithalath. Inuence of a ne glass
powder on the durability characteristics of concrete and its comparison to y
ash. Cem Concr Compos 2008;30:48696.
[21] Saccani Andrea, Bignozzi Maria Chiara. ASR expansion behavior of recycled
glass ne aggregates in concrete. Cem Concr Res 2010;40:5316.
[22] Taha Bashar, Nounu Ghassan. Using lithium nitrate and pozzolanic glass
powder in concrete as ASR suppressors. Cem Concr Compos 2008;30:497505.
[23] Taha Bashar, Nounu Ghassan. Properties of concrete contain mixed colour
waste recycled glass as sand and cement replacement. Constr Build Mater
2008;22:71320.
[24] Hall C. Water sorptivity of mortar and concretes a review. Mag Concr Res
1989;41(147):5161.

215

[25] Jin W, Meyer C, Baxter S. Glascrete concrete with glass aggregate. ACI Mater J
2000;97(2):20813.
[26] Idir Rachida, Cyr Martin, Tagnit-Hamou Arezki. Pozzolanic properties of ne
and coarse color-mixed glass cullet. Cem Concr Compos 2011;33:1929.
[27] Ichikawa Tsuneki. Alkalisilica reaction, pessimum effects and pozzolanic
effect. Cem Concr Res 2009;39:71626.
[28] Sahmaran M, Kasap O, Duru K, Yaman IO. Effects of mix composition and
watercement ratio on the sulfate resistance of blended cements. Cem Concr
Compos 2007;29:15967.
[29] Ayea Thidar, Oguchi Chiaki T. Resistance of plain and blended cement mortar
exposed to severe sulfate attacks. Constr Build Mater 2011;25:298896.
[30] Sobolev Konstantin, Trker Pelin, Soboleva Svetlana, Iscioglu Gunsel.
Utilization of waste glass in ECO-cement: strength properties and
microstructural observations. Waste Manage (Oxford) 2007;27:9716.
[31] ASTM C109/C109M-11a standard test method for compressive strength of
hydraulic cement mortars (using 2 in. or [50 mm] cube specimens). West
Conshohocken; 2011.
[32] ASTM. ASTM C230/C230M-08 standard specication for ow table for use in
tests of hydraulic cement. Philadelphia; 2008.
[33] ASTM. ASTM C618-12 standard specication for coal y ash and raw or
calcined natural pozzolan for use in concrete. Philadelphia; 2012.
[34] ASTM. ASTM C1567-11 standard test method for determining the potential
alkali silica reactivity of combinations of cementitious materials and aggregate
(accelerated mortar bar method). United States: ASTM International; 2011.
[35] IPQ. EN 196-1. Methods of testing cement Part 1: determination of strength.
Lisbon: IPQ; 2006.
[36] IPQ. EN 196-3. Methods of testing cement Part 3: Determination of setting
time and soundness. Lisbon; 2005.
[37] IPQ. EN 196-5. Methods of testing cement Part 5: Pozzolanicity test for
pozzolanic cements. Lisbon; 2011.
[38] IPQ. EN 13263-1. Silica fume for concrete Part 1: Denitions, requirements
and conformity criteria. Lisbon; 2009.
[39] IPQ. EN 450-1. Fly ash for concrete Part1: Denition, specications and
conformity criteria Lisbon; 2006.
[40] IPQ. EN 15167-1. Ground granulated blast furnace slag for use in concrete,
mortar and grout Part 1: Denitions, specications and conformity criteria.
Lisbon; 2008.
[41] IPQ. EN 934-2: Admixtures for concrete, mortar and grout. Concrete
admixtures Part 2: Denitions, requirements, conformity, marking and
labelling. Lisbon; 2009.
[42] LNEC. E-462: Resistncia dos cimentos ao ataque por sulfatos. Resistance of
cements to sulphate attack. Lisbon; 2004.
[43] Nordtest method NT build 492 concrete, mortar and cement-based repair
materials: chloride migration coefcient from non-steady-state migration
experiments. Finland; 1999.
[44] RILEM. RILEM CPC-18 measurement of hardened concrete carbonation depth.
Mater Struct No. 102; 1984.
[45] RILEM. RILEM TC 116-PCD Permeability of concrete as a criterion of its
durability. Recommendations. Mater Struct/Materiaux et Constr; April de
2009. p. 174179.