Journal of Cleaner Production 65 (2014) 389e396

Contents lists available at ScienceDirect

Journal of Cleaner Production

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

Durability performance of concrete incorporating coarse aggregates

from marble industry waste
Antnio Andr, Jorge de Brito*, Alexandra Rosa, Diogo Pedro
Department of Civil Engineering, Architecture, and Geo-resources, Instituto Superior Tcnico, Technical University of Lisbon, Av. Rovisco Pais, 1049-001
Lisbon, Portugal

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 18 March 2013
Received in revised form
21 August 2013
Accepted 21 September 2013
Available online 1 October 2013

The quarrying of marble, a well-known ornamental stone, has a substantial positive impact on Portugals
economy, but it also generates large environmental impacts. The amount of waste produced during
quarrying can be as much as 80% of all stone/soil extracted. That waste is then dumped near the quarry,
where it accumulates indiscriminately because a viable alternative for its disposal has not yet been
found. In this context, solutions must be found that can transform this waste into a by-product and
restore some of its economic value.
The main goal of this study was to evaluate the inuence of the replacement of primary aggregates
(PA) with marble aggregates. No additions or admixtures were used, as those could change the fresh or
hardened properties of the resulting concrete and disguise the inuence of the replacement under study.
This evaluation required the production of three concrete families. The conventional primary aggregates
(PA; basalt, granite and limestone) were replaced in the three families by coarse marble aggregates
(CMA) at ratios of 20%, 50% and 100% of the total volume of aggregates. These mixes were tested in the
concretes fresh state for workability and density and in its hardened state for compressive strength,
water absorption by capillarity and immersion, carbonation and chloride penetration. The results indicate that there are no signicant differences between a concrete produced using CMA and one made with
PA in terms of durability, making the use of this waste as concrete aggregate perfectly feasible.
The mechanical properties of the concrete made with CMA were assessed in the scope of concurrent
work also performed at the Instituto Superior Tcnico (Lisbon, Portugal).
2013 Elsevier Ltd. All rights reserved.

Keywords:
Coarse marble aggregates
Recycled aggregate concrete
Durability
Quarrying industry waste

1. Introduction
The stone industry is of major importance to the Portuguese
economy and to that of the world. However, by its very nature, this
industry frequently has signicant environmental impacts. Advances in the exploitation of quarries have led to the extraction of
blocks containing more impurities and of worse quality, which
lowers the productivity of ornamental stone. The waste produced
during the extraction process can amount to as much as 80% of the
total volume of stone extracted (Hebhoub et al., 2011). Bulkier
waste is disposed of at sites located far from the excavation front,
Abbreviations: PA, Primary aggregates; CMA, Coarse marble aggregates; RC,
Reference concrete; BRC, Basalt reference concrete; GRC, Granite reference concrete; LRC, Limestone reference concrete; MRC, Marble reference concrete; BCA,
Basalt coarse aggregates; GCA, Granite coarse aggregates; LCA, Limestone coarse
aggregates.
* Corresponding author. Tel.: 351 218443659; fax: 351 218443071.
E-mail addresses: [email protected], [email protected] (J. de Brito).
0959-6526/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jclepro.2013.09.037

contributing to the degradation of the environment and the natural

landscape by occupying the land indiscriminately.
The construction industry can use marble waste for purposes
other than as ornamental stone. Marble dust (also called sludge, an
agglomeration of water and very ne particles produced by cutting
the stone) is the most widely studied alternative, primarily because
it is quite versatile. According to Saboya et al. (2007), marble dust
can enhance the properties of ceramic bricks. Aruntas et al. (2010)
studied the effects of adding marble dust to cement, whereas Ergn
(2011) evaluated the effect of replacing cement with diatomite and
marble dust on the mechanical properties of concrete. Marble dust
can also be used as ller for road construction and asphalt mixes
(Karasahin and Terz1, 2005). Binici et al. (2007) showed that using
up to 15% marble dust as an additive may result in more durable
concrete. It is also possible to improve all of the properties of a selfconsolidating concrete by adding 200 kg/m3 of marble dust, as
shown by Topu et al. (2009). Finally, Corinaldesi et al. (2010) reported that when 10% of sand is replaced with marble dust and

390

A. Andr et al. / Journal of Cleaner Production 65 (2014) 389e396

Table 1
Compositions of reference mixes (per m3).
Sieve (mm)

Coarse
aggregates

Fine aggregates

16e22.4
11.2e16
8e11.2
5.6e8
4e5.6
Coarse sand
Fine sand

CEM II A-L 42.5

R (kg/m3)
Water (kg/m3)
w/c

Fig. 1. Experimental concrete mixes.

superplasticizers are used, the compressive strength of mortar is

maximized but the level of workability is unchanged.
Despite the successful use of marble dust, there is little information on the potential use of larger marble waste as a coarse
aggregate, especially with respect to its effect on concrete durability. However, Akbulut and Gurer (2007) explored this potential
and concluded that CMA can be used in binder layers of asphalt
pavements bearing low to medium trafc volume. Moreover,
Gencel et al. (2012) determined that concrete pavement blocks
made with CMA are of adequate quality. Concerning the use of CMA
in concrete, Hebhoub et al. (2011) studied the effect of the
replacement of conventional aggregates with several ratios of ne,
coarse and ne and coarse marble waste on concrete. The results
obtained revealed that for incorporation ratios up to 75%, the use of
CMA is benecial for concrete strength. Furthermore, Binici et al.
(2008) showed that CMA signicantly increases resistance to
chloride penetration, and Pereira et al. (2009) found that the
durability of class C30/37 concrete is not affected by the aggregates
mineralogy. Overall, marble aggregates can be used to enhance the
mechanical properties and chemical resistance of concrete mixes.

BRC

LRC

GRC

MRC

Mass
(kg/m3)

Mass
(kg/m3)

Mass
(kg/m3)

Mass
(kg/m3)

366.4
362.4
140.6
139.0
122.4
650.7
183.5
350.0

324.9
321.3
124.6
123.3
108.5
650.7
183.5

337.5
333.8
129.5
128.0
112.7
650.7
183.5

331.5
327.9
127.2
125.8
110.7
619.1
174.6

189.0
0.55

NP EN 933-1 (2000), Grading size analysis;

NP EN 933-4 (2002), Shape index;
NP EN 1097-3 (2002), Loose bulk density and voids;
NP EN 1097-6 (2003), Particle density and water absorption;
LNEC E 237 (1971), Los Angeles abrasion test;
NP EN 12350-2 (2009), Slump test with Abrams cone;
NP EN 12350-6 (2009), Density;
NP EN 12350-3 (2009), Compressive strength, 4 specimens/test;
LNEC E 391 (1993), Carbonation resistance, 3 specimens/test (7,
28, 56 days);
- LNEC E 393 (1993), Absorption by capillarity, 4 specimens/test;
- LNEC E 394 (1993), Absorption by immersion, 4 specimens/test;
- LNEC E 463 (2004), Chloride penetration resistance, 3 specimens/test (7, 28, 56 days).

3. Experimental results and discussion

3.1. Aggregate properties
Table 3 summarizes the experimental results of the tests performed to characterize the aggregates.

2.2. Testing

3.1.1. Particle density and bulk density

Particle density depends on the source rock, and it obviously
varies according to the mineralogical composition of each material.
Basalt has the highest particle density because of its rich content of
iron-magnesium minerals, such as pyroxenes and olivine, according to Rutley (1916). Although its origin is similar to that of basalt,
granite is composed of feldspar minerals that have a lower percentage of heavy elements than of iron-magnesium minerals.
Therefore, the particle density of granite is lower than that of basalt.
The limestone formation process produces rocks with the worst
characteristics in terms of particle density. Limestone results when
a deposition of inorganic sediments is combined with some organic
material and solidied by compression, creating less dense and
more porous rocks. It is thus expected that limestone will have the
lowest particle density. Marbles particle density values are within
the range obtained by the authors who have studied this material
(Hebhoub et al. (2011), Pereira et al. (2009) and Binici et al. (2008))
and are similar to those of the PA.
Basalt has the highest bulk density, followed by limestone,
marble and granite, contrary to the results obtained for particle
density. Taking the void content of limestone, given in Table 3, this
result was expected, meaning that limestone aggregates have a
better spatial arrangement of particles than the others.

The characterization tests of the aggregates and the fresh and

hardened concrete were conducted according to the following
standards and specications:

3.1.2. Water absorption

CMA has the lowest water absorption of all aggregates tested,
followed by the basalt (BCA) and granite (GCA) aggregates due to

2. Experimental program
2.1. Concrete mix design
The 10 concrete mixes used are depicted in Fig. 1: three reference concrete (RC) mixes with 100% basalt (BRC), granite (GRC) or
limestone (LRC) coarse aggregates; six concrete mixes obtained by
replacing each of the primary aggregates (PA) used in the reference
concretes with marble coarse aggregates (CMA) at ratios of 20% or
50% of the total volume; and nally, a concrete with 100% CMA
(MRC). To create a set of concrete families that were usable for a
signicant number of current structural applications, it was guaranteed that the concrete produced met the requirements specied
in NP EN 206-1 (2005) for a given class of environmental exposure.
Therefore, the concrete produced complied with the characteristics
described in Tables 1 and 2.
All of the mixes had a slump range of 115  10 mm, which is
lower than the standard set. The three reference concretes were
designed using Faury reference curves.

A. Andr et al. / Journal of Cleaner Production 65 (2014) 389e396

391

Table 2
Mixes with CMA (per m3).
BC
% of marble incorporation
Sieves (mm)
Aggregates

Coarse

Fine

16e22.4
11.2e16
8e11.2
5.6e8
4e5.6
Coarse sand
Fine sand

CEM II A-L 42.5 R (kg/m3)

Water (kg/m3)
w/c
a

LC

20%

50%
a

GC

20%

50%

20%

50%

PA

MA

PA

MA

LA

MA

LA

MA

GA

MA

GA

MA

293.1
289.9
112.5
111.2
97.9
619.1
174.6
350.0
189.0
0.55

66.3
65.6
25.4
25.2
22.1
e
e

183.2
181.2
70.3
69.5
61.2
619.1
174.6

165.8
163.9
63.9
62.9
55.4
e
e

259.9
257.0
99.7
98.6
86.8
619.1
174.6

66.3
65.6
25.4
25.2
22.1
e
e

162.4
160.7
62.33
61.66
54.2
619.1
174.6

165.8
163.9
63.9
62.9
55.44
e
e

270.0
267.0
103.6
102.4
90.2
619.1
174.6

66.3
65.6
25.4
25.2
22.1
e
e

168.7
166.9
64.7
64.0
56.3
619.1
174.6

165.8
163.9
63.6
62.9
55.4
e
e

PA e Primary aggregates; MA e Marble aggregates.

Table 3
Aggregate properties (average value and standard deviation).
Particle dry
density (kg/m3)

Fine sand
Coarse sand
BCA
LCA
GCA
CMA

Loose bulk
density (kg/m3)

Particle
saturated
surface-dried
density (kg/m3)

Void content
(%)

Water
absorption (%)

Los Angeles
coefcient (%)

Shape index
(%)

2576
2621
2953
2641
2705
2687

e
e
22.2
9.6
48.3
22.5

2584
2625
2976
2671
2734
2705

e
e
20.5
13.2
33.9
17.5

1500
1543
1475
1430
1350
1352

e
e
48.5
7.4
30.3
33.8

41.8
41.1
50.0
45.9
50.1
49.7

e
e
1.4
0.2
2.0
1.0

0.091
0.048
0.782
1.149
1.077
0.662

e
e
0.2
0.1
0.6
0.2

e
e
11.8
32.3
24.7
38.8

e
e
2.8
1.77
3.9
0.4

e
e
23.4
16.2
37.7
30.1

e
e
5.3
3.0
11.8
1.4

the dense matrix of their source rocks, and nally the limestone
(LCA) aggregates. Most of the results are below the means reported
by the authors cited, as shown in Table 4. Because these gures are
relatively low, this property was considered unnecessary for the
determination of the effective w/c ratio.
3.1.3. Los Angeles coefcient
The Los Angeles coefcient of the CMA is high, 38.8%, but still lower
than 50%, which is the maximum recommended for incorporation in
structural concrete. BCA is the least susceptible to wear with a coefcient of 11.8%, followed by GCA with 24.7%, indicating that the
mineralogical composition is relevant to this property. As such, it is
expected that the mass loss of LCA (32.3%) will be considerably higher
than that of BCA and GCA as well as that of the CMA because of LCAs
richer content of fragile minerals that are sensitive to mechanical
actions. These results are summarized in Table 3.
3.1.4. Shape index
CMAs shape index demonstrates that most of its particles are
both elongated and angular. These two characteristics may lead to a

Table 4
Water absorption (comparison with previous studies).
Authors

Binici et al. (2008)

Pereira et al. (2009)
Hebhoub et al. (2011)
Mean value
Andr et al.

Water absorption (%)

Basalt

Limestone

Granite

Marble

e
1.50
e
1.50
0.78

2.30
2.80
e
2.55
1.15

1.50
0.30
e
0.90
1.08

1.40
0.05
0.39
0.92
0.66

concrete with a poorer performance in terms of its mechanical and

durability properties.
Table 3 shows that, except for LCA, the aggregates shape indexes are higher than 20%. This may cause workability issues for
fresh concrete because angular particles lead to a larger internal
friction angle and to a worse spatial arrangement. Additionally, the
hardened concrete may be more porous and consequently less
durable. The 20% gure is the maximum recommended limit for
recycled aggregates incorporated in concrete with a class strength
of C20/25 or higher, according to the Technical guideline for recycled
aggregate concrete in Hungary.
It is important to note that the crushing process may inuence
an aggregates shape. Crushing may lead to rounder shapes, which
will help improve the mechanical and durability performance of
the concrete.
Further discussion of the aggregates properties is available in
another paper by the authors Martins et al. (2013).
3.2. Fresh concrete properties
3.2.1. Slump
Fig. 2 shows the results of the Abrams cone test for all of the
mixes. Note that to ensure the reliability of the comparison between the various properties of the concrete mixes produced, it was
established that all of the mixes should have similar workability.
Therefore, all results are within the range of 115  10 mm.
Fig. 2 also proves that the workability exhibits no clear trend as
PA is replaced by CMA, contrary to what was found by Binici et al.
(2008) and Hebhoub et al. (2011) in their studies. The workability of
mixes made with GCA is not signicantly affected by its replacement with CMA, possibly due to the high shape index of GCA. For
mixes made with BCA and GCA, there is an increase in workability

392

A. Andr et al. / Journal of Cleaner Production 65 (2014) 389e396

4.0

Slump (mm)

125
120
BCA
115

LCA
GCA

110

2.0
0.0
-2.0
-4.0

BCA

-6.0

LCA

-8.0

GCA

-10.0
-12.0
-14.0

105
0 (RC)

20
50
Incorporation ratio of CMA (%)

0 (RC)

100

20
50
Incorporation ratio of CMA (%)

100

Fig. 4. Compressive strength relative to that of the reference concrete mixes.

Fig. 2. Slump.

Table 6
Water absorption by immersion (average, standard deviation and difference from
the reference mix).

2500
Fresh concrete density (kg/m3)

fcm, 28days/fcmRC, 28days (%)

130

2480

x (%) s (%) D (%)

2460
2440
BCA
2420

LCA

2400

BRC
BC20
BC50
MRC

13.6
14.4
14.4
14.0

0.3
5.8
5.8
3.0

0
5.8
5.8
3.0

x (%) s (%) D (%)

LRC
LC20
LC50
MRC

14.1
13.8
13.3
14.0

0.3
0.5
0.3
0.4

0
2.3
5.4
0.8

x (%) s (%) D (%)

GRC
GC20
GC50
MRC

13.8
13.6
14.0
14.0

0.44
0.2
0.2
0.4

0
1.6
1.9
1.5

GCA

2380
2360
0 (RC)

20
50
100
Incorporation ratio of CMA (%)
Fig. 3. Fresh concrete density.

for the 20% incorporation ratio, most likely because of the CMAs
plain surface and low absorption. However, the results show a loss
of workability for the 50% replacement ratio, most likely because
the shape index has more inuence than the ow increase
observed for the 20% replacement ratio.
3.2.2. Density
Fig. 3 presents the density results for the mixes. The results
show that the density of the concrete made with BCA decreases
with higher incorporation ratios of CMA, but it is still higher than
those of the LCA and GCA mixes. This result was expected because
the aggregates have similar densities, the concretes were produced
with the same workability, and no extra water was needed. The
slight increase observed in the 20% replacement ratio mix in the
GCA and LCA families may be due to an improved spatial
arrangement of the particles provided by the compacted mixes, as
observed in the slump test results.
3.3. Hardened concrete properties
3.3.1. Compressive strength
Although concrete durability cannot be directly characterized
by compressive strength, that parameter is still frequently used to

assess concretes quality and thus indirectly reveals its durability.

As such, the results obtained by Martins et al. (2013) in their
concurrent research work were analyzed, and the mean
compressive strength at 28 days is presented in Table 5. Although
Fig. 4 indicates a general downward trend of the mean compressive strength at 28 days with increasing incorporation ratio, this
decrease may be considered almost insignicant with variations
up to 10.3% (for the granite concrete mix). Hebhoub et al. (2011)
obtained identical results for the difference between the reference concrete and the one produced with 100% marble aggregates.
However, the concrete strength improved for the 25%, 50% and
70% incorporation ratios. Binici et al. (2008) concluded that the
use of marble and granite improves the exural and splitting
tensile strengths of concrete.
3.3.2. Water absorption by immersion
The results from the water absorption by immersion test are
presented in Table 6. This property characterizes the open porosity
of the cement matrix at 28 days. Open porosity is due to either
excess water in the mix that is neither consumed in the cement
hydration reactions nor absorbed by the aggregates or to air that is
still trapped after the vibration process.
From Fig. 5, it can be concluded that although there is an upward
linear trend (R2 0.93), the introduction of CMA does not signicantly inuence water absorption. All mixes exhibited very similar
absorption values with variations of between 5.4% and 5.8%. This
can be explained both by the weak absorption of the aggregates in
the mix and by the (expected) similarity of the concretes microstructure. These results are in agreement with those obtained by
Pereira et al. (2009).

Table 5
Compressive strength at 28 days (average, standard deviation and difference from the reference mix).
fcm,
BRC
BC20
BC50
MRC

28

45.9
44.0
44.1
41.8

(MPa)

s (MPa)

D (%)

0.62
0.91
0.57
1.28

0
4.1
3.9
8.9

fcm,
LRC
LC20
LC50
MRC

28

43.4
43.2
44.2
41.8

(MPa)

s (MPa)

D (%)

1.89
1.48
1.10
1.28

0
4.1
3.9
8.9

fcm,
GRC
GC20
GC50
MRC

28

46.6
43.7
41.3
41.8

(MPa)

s (MPa)

D (%)

0.34
1.24
3.81
1.28

0
4.1
3.9
8.9

A. Andr et al. / Journal of Cleaner Production 65 (2014) 389e396

393

Fig. 5. Water absorption by immersion.

Water absorption by capillarity coefcient at Water height by

72 h
capillarity at 72 h

BRC
BC20
BC50
MRC
LRC
LC20
LC50
MRC
GRC
GC20
GC50
MRC

x (g/mm2 x 104) s (g/mm2 x 104) D (%)

x (mm) s (mm) D (%)

7.2
10.0
12.1
8.0
14.2
11.7
13.3
8.0
13.6
9.4
11.6
8.0

17.4
15.8
20.5
15.4
21.3
12.6
19.3
15.4
31.2
16.3
16.0
15.4

0.9
0.8
0.6
0.6
0.1
0.3
0.8
0.6
0.7
0.1
0.7
0.6

0.0
38.9
68.1
11.1
0.0
17.6
6.3
43.7
0.0
30.9
14.7
41.2

11.1
13.4
14.7
9.8
7.0
9.2
13.2
9.8
16.4
11.7
13.8
9.8

0.0
9.2
17.8
11.5
0.0
40.8
9.4
27.7
0.0
47.8
48.7
50.6

Absorption by capillarity
relative to the reference
concrete (%)

80
60
40
20

BCA
LCA

0
0

20

50

100

GCA

-20
-40
-60

Incorporation ratio of CMA (%)

Fig. 6. Absorption by capillarity compared with the reference concrete (%).

Absorption by capillarity
relative to the reference
concrete (%)

35
BRC

30

BC20
25

BC50

20

LRC

15

LC20
LC50

10

GRC
5

GC20

GC50
0

3
6
24
Incorporation ratio of CMA (%)

72

Fig. 7. Water height by capillary action for every mix versus Time.

MRC

3.3.3. Water absorption by capillarity

The results for water absorption by capillarity at 72 h are
shown in Table 7, and the comparison with the reference concretes
is shown in Fig. 6. They reveal that this property slightly decreases
with the addition of CMA for the mixes made with LCA and GCA,
with variations of 17.4% and 31% for absorption by capillarity and
40.6% and 48.7% for capillary height, respectively. However, in the
case of BCA concrete, such water absorption increases by up to
67.7% (for the 50% replacement ratio) as a result of the poorer
adherence between the CMA and the cement paste compared to
the BCA, thereby increasing the interface zone pores in which
absorption is more severe. According to Pereira et al. (2009), the
LCA mix registered the highest capillarity coefcient, followed by
the GCA mix. For all mixes, the absorption is more intense in the
rst hours when the specimen is in contact with water, as shown
in Fig. 7.
Ferreira (2000) and Bravo and de Brito (2012) concluded that the
capillary height at 72 h varies exponentially with absorption by
immersion at 28 days. Fig. 8 demonstrates that, for this experimental
campaign, this parameter is insensitive to increased water absorption by immersion. Therefore, there is no statistically signicant
relationship between them, which contradicts the previous studies.
3.3.4. Carbonation resistance
Table 8 and Fig. 9 show the carbonation depth for all of the
concrete mixes.
As expected, carbonation depth is quite similar for the various
mixes at any given age. The analysis of water absorption, which is
inuenced by the same factors, led to a similar conclusion. The
variations are quite small; the most signicant, 25.1%, occurred for
GC50 at 7 days in comparison to its reference concrete. However, at
91 days, the variation was only 11.1%. It is also noted that the
carbonation depth increased suddenly in the rst days and exhibited
a slower rate of increase at 56 and 91 days. Overall, it is inferred that

Water height by capillary height

at 72 h (mm)

Table 7
Water absorption by capillarity (average, standard deviation and difference from the
reference mix).

35
BRC

30

BC20
25

BC50

20

LRC

15

LC20
LC50

10

GRC
5
0
13.2

GC20
GC50
13.4

13.6
13.8
14.0
14.2
Water absorption by immersion (%)

14.4

14.6

MRC

Fig. 8. Water height by capillary action versus water absorption by immersion at 72 h.

394

A. Andr et al. / Journal of Cleaner Production 65 (2014) 389e396

Table 8
Carbonation depth (average, standard deviation and difference from the reference mix).
Carbonation depth (mm)
Concrete

7 days

s (mm)

D (%)

28 days

s (mm)

D (%)

56 days

s (mm)

D (%)

91 days

s (mm)

D (%)

BRC
BC20
BC50
LRC
LC20
LC50
GRC
GC20
GC50
MRC

3.47
3.56
3.81
3.99
3.65
3.76
3.75
3.96
4.70
3.69

0.26
0.21
0.70
0.70
0.44
0.21
0.15
0.31
0.77
0.60

0.0
2.4
9.6
0.0
8.6
5.7
0.0
5.6
25.1
1.6

8.49
7.74
8.28
7.83
7.36
7.38
8.29
8.03
9.08
8.51

0.36
1.10
0.27
1.14
0.91
0.66
0.75
0.57
0.36
0.57

0.0
8.9
2.5
0.0
6.0
5.8
0.0
3.1
9.5
2.7

10.68
11.79
10.06
10.32
10.81
11.45
10.51
10.22
12.15
10.83

0.81
0.54
0.57
1.53
0.80
0.49
0.81
1.15
0.50
0.93

0.0
10.4
5.8
0.0
4.8
11.0
0.0
2.8
15.6
3.0

13.54
13.36
11.12
13.06
13.24
14.47
12.61
14.04
14.01
13.53

0.59
0.52
0.71
0.86
1.29
0.40
0.64
0.75
0.37
1.33

0.0
1.3
17.9
0.0
1.4
10.7
0.0
11.3
11.1
7.3

16.00
Carbonation depth (mm)

14.00
12.00
10.00
7 days
8.00

28 days

6.00

56 days

4.00

91 days

2.00
0.00
BRC

BC20

BC50

LRC

Basalt

LC20

LC50

GRC

Limestone

GC20

GC50

Granite

MRC
Marble

Fig. 9. Carbonation depth.

the incorporation of CMA contributes to the concrete matrixs

carbonation-related characteristics exactly as the other PAs do.
De Brito and Alves (2010), de Brito and Robles (2010) and
Amorim et al. (2012) also evaluated the relationship between
carbonation resistance and the weighted bulk density of the recycled aggregates. Their results and those of our study are summarized in Fig. 10. Contrary to the other campaigns, the carbonation
depth in this study decreases as the bulk density increases. This can
be explained by the CMAs considerable inuence on the basalt
family, for which small carbonation depth decreases were recorded
with the introduction of CMA.

CarbonationCMAC/CarbonationPAC

3.3.5. Chloride penetration resistance

The results for chloride penetration at 28 and 91 days are presented in Table 9 and in Fig. 11 (where it is noticeable that some
results are missing because of experimental problems that could

not be solved). They show that the chloride migration coefcient

increases with the replacement ratio of PA with CMA. This increase
may be due to the low percentage of alumina (Al2O3), approximately 0.64%, present in the CMA. Alumina favors the formation of
tricalcium aluminate (C3A), which xes the chloride ions. With a
low alumina content, the free chloride percentage within the
concrete matrix increases and therefore promotes chloride migration and worsens the concretes performance (Uysal et al., 2012).
However, Binici et al. (2008) obtained contrary results. Concrete
produced with CMA exhibited a clearly superior chloride penetration resistance compared to concrete with LCA. The CMA chemical
analysis performed in their work showed that the CMA alumina
content was 21.9%, whereas the LCA alumina content was 2.1%.
Apart from the contradictory results, the others corroborate the
previous hypothesis that the alumina content undoubtedly inuences the chloride penetration depth.

Table 9
Chloride migration coefcient (average, standard deviation and difference from the
reference mix).

1.80
1.65

Chloride migration coefcient (1012 m2/s)

1.50
De Brito and Alves

1.35

Amorim et al.
1.20

De Brito and Robles

1.05

Andr et al.
Linear (Andr et al.)

0.90
0.75
-0.01

0.01

y = -2.1287x + 1.0375
R = 0.2006
0.03
0.05

1-(dCMA/dPA)
Fig. 10. Concrete carbonation coefcient versus weighted bulk density of the
aggregates.

Concrete 28 days s (1012 m2/s) D (%) 91 days s (1012 m2/s) D (%)

MRC
BRC
BC20
BC50
LRC
LC20
LC50
GRC
GC20
GC50

e
13.51
e
15.24
17.92
21.49
e
e
18.73
20.51

2.18
2.76
0.83
2.34
2.54
0.85
2.69
1.29
1.37
1.72

e
0.0
e
12.8
0.0
19.9
e
e
0.0
9.5

17.91
e
e
e
14.34
14.86
16.27
15.21
17.3
19.86

2.00
1.57
1.61
1.94
2.45
1.67
1.09
2.07
1.39
1.42

e
e
e
e
0.0
3.6
13.5
0.0
14.3
30.5

A. Andr et al. / Journal of Cleaner Production 65 (2014) 389e396

395

Chloride migration coefficient

(x10-2 m2/s)

25
20
15
28 days

10

91 days

5
0
BRC

BC20
Basalt

BC50

LRC

LC20

LC50

GRC

Limestone

GC20
Granite

GC50

MRC
Marble

Fig. 11. Chloride migration coefcient at 28 and 91 days.

4. Conclusions

References

The experimental results allow the following conclusions to be

drawn:

Akbulut, H., Grer, C., 2007. Use of aggregates produced from marble quarry waste
in asphalt pavements. Build. Environ. 42 (5), 1921e1930.
Amorim, P., de Brito, J., Evangelista, L., 2012. Concrete made with coarse concrete
aggregate: inuence of curing on durability. ACI Mater. J. 109 (2), 195e204.
Aruntas, H.Y., Gr, M., Dayi, M., Tekin, I., 2010. Utilization of waste marble dust as
an additive in cement production. Mater. Des. 31 (8), 4039e4042.
Binici, H., Kaplan, H., Yilmaz, S., 2007. Inuence of marble and limestone dusts as
additives on some mechanical properties of concrete. Sci. Res. Essay 2 (9), 372e
379.
Binici, H., Shah, T., Aksogan, O., Kaplan, H., 2008. Durability of concrete made with
granite and marble as recycle aggregates. J. Mater. Process. Technol. 208 (1e3),
299e308.
Bravo, M., de Brito, J., 2012. Concrete made with used tyre aggregates: durabilityrelated performance. J. Clean. Prod. 29 (1), 42e50.
Corinaldesi, V., Moriconi, G., Naik, T.R., 2010. Characterization of marble powder for
its use in mortar and concrete. Constr. Build. Mater. 24 (1), 113e117.
De Brito, J., Alves, F., 2010. Concrete with recycled aggregates: the Portuguese
experimental research. Mater. Struct. 43 (S1), 35e51.
De Brito, J., Robles, R., 2010. Recycled aggregate concrete (RAC) e methodology for
estimating its long-term properties. Indian J. Eng. Mater. Sci. 17 (6), 449e462.
Ergn, A., 2011. Effects of the usage of diatomite and waste marble powder as
partial replacement of cement on the mechanical properties of concrete. Constr.
Build. Mater. 25 (2), 806e812.
Ferreira, R., 2000. Evaluation of Concrete Durability Tests (in Portuguese). MSc
Thesis in Civil Engineering. Minho University, Guimares, Portugal.
Gencel, O., Ozel, C., Koksal, F., Erdogmus, E., Martnez-Barrera, G., Brostow, W., 2012.
Properties of concrete paving blocks made with waste marble. J. Clean. Prod. 21
(1), 62e70.
Hebhoub, H., Aoun, H., Belachia, M., Houari, H., Ghorbel, E., 2011. Use of waste
marble aggregates in concrete. Constr. Build. Mater. 25 (3), 1167e1171.
Karasahin, M., Terzi, S., 2005. Evaluation of marble waste dust in the mixture of
asphaltic concrete. Constr. Build. Mater. 21 (3), 616e620.
Martins, P., de Brito, J., Rosa, A., Pedro, D., 2013. Mechanical performance of concrete
with incorporation of coarse waste from the marble industry. Magazine of
Concrete Research (submitted for publication).
Pereira, C.G., Castro-Gomes, J., Oliveira, L., 2009. Inuence of natural coarse
aggregate size, mineralogy and water content on the permeability of structural
concrete. Constr. Build. Mater. 23 (2), 602e608.
Rutley, F., 1916. Elements of Mineralogy, 19th ed., vol. 2. University of Michigan,
Herbert Harold Read, p. 394.
Saboya, F., Xavier, G.C., Alexandre, J., 2007. The use of the powder marble byproduct to enhance the properties of brick ceramic. Constr. Build. Mater. 21
(10), 1950e1960.
_
lu, T., 2009. Effect of waste marble dust content as ller
Topu, I.B.,
Bilir, T., Uygunog
on properties of self-compacting concrete. Constr. Build. Mater. 23 (5), 1947e
1953.
Uysal, M., Yilmaz, K., Ipek, M., 2012. The effect of mineral admixtures on mechanical
properties, chloride ion permeability and impermeability of self-compacting
concrete. Constr. Build. Mater. 27 (1), 263e270.

 Workability is neither signicantly affected (for a constant w/c

ratio) nor does it exhibit a dened trend when CMA is incorporated. Despite this, there is still an increase in workability for
the 20% incorporation ratio and a decrease for the 50% ratio for
mixes made with BCA and LCA. The workability of mixes made
with GCA is unchanged;
 Concrete density reects the density of each of its components
and their packing. Concrete density therefore decreases with the
addition of CMA for mixes made with BCA and is approximately
constant for concretes made with LCA and GCA;
 Overall, there is a slight loss of compressive strength at 28 days
with an increase in the replacement ratio of PA with CMA;
 As for water absorption by immersion, the behavior of concrete
made with CMA is similar to that of the reference concretes. This
may be due to the low absorption of the aggregates used and the
microstructural similarity between the mixes;
 The incorporation of CMA in the mixes made with LCA and GCA
results in lower water absorption by capillary action. The
opposite trend is observed in the mixes made with BCA. This can
be explained by the worse adhesion of the CMA and consequent
increase of the interface zone pores, which is where this phenomenon is more severe.
 Concrete made with CMA has a similar carbonation depth to
that of the reference concretes, thus supporting the microstructural similarity hypothesis between all the families;
 The incorporation of CMA in the mixes resulted in a signicant
increase of the chloride migration coefcient. The low alumina
percentage of these aggregates may be the primary cause, as this
compound promotes the formation of tricalcium aluminate,
which favors the xation of the chloride ions.
 In terms of durability, it was shown that CMA produces concrete
with similar characteristics to those made with BCA, LCA and
GCA, which are typically used in the construction industry.
Therefore, the incorporation of CMA in concrete is perfectly
feasible. However, great care must be taken in chloridecontaminated environments, where CMA with low alumina
content exhibited its worst performance.
Acknowledgments
The authors gratefully acknowledge the support of SECIL, MotaEngil and Lena Agregados for the materials required for the
experimental campaign and also thank the ICIST Research Institute,
IST, Technical University of Lisbon, Portugal and the Foundation for
Science and Technology (FCT).

Standards used
LNEC E 237, 1971. Aggregates: Los Angeles Abrasion Test (in Portuguese). LNEC,
Lisbon, Portugal.
LNEC E 391, 1993. Concrete: Determination of the Carbonation Resistance (in Portuguese), Lisbon, Portugal.
LNEC E 393, 1993. Concrete: Determination of the Water Absorption by Capillarity
(in Portuguese), Lisbon, Portugal.
LNEC E 394, 1993. Concrete: Determination of the Water Absorption by Immersion
(in Portuguese), Lisbon, Portugal.
LNEC E 463, 2004. Concrete: Determination of the Chloride Diffusion Coefcient by
Migration Test in a Non-stationary Regime (in Portuguese), Lisbon, Portugal.

396

A. Andr et al. / Journal of Cleaner Production 65 (2014) 389e396

NP EN 206e1, 2005. Concrete e Part 1: Specication, Performance, Production and

Conformity. IPQ, Lisbon, Portugal.
NP EN 933e1, 2000. Tests for Geometrical Properties of Aggregates e Part 1:
Determination of Particle Size Distribution e Sieving Method. IPQ, Lisbon,
Portugal.
NP EN 933e4, 2002. Tests for Geometrical Properties of Aggregates e Part 4:
Determination of Particle Shape e Shape Index. IPQ, Lisbon, Portugal.
NP EN 1097e3, 2002. Tests for Mechanical and Physical Properties of Aggregates e
Part 3: Determination of Loose Bulk Density and Voids. IPQ, Lisbon, Portugal.

NP EN 1097e6, 2003. Tests for Mechanical and Physical Properties of Aggregates e

Part 6: Determination of Particle Density and Water Absorption. IPQ, Lisbon,
Portugal.
NP EN 12350e2, 2009. Testing Fresh Concrete. Part 2: Slump-test. IPQ, Lisbon,
Portugal.
NP EN 12350e6, 2009. Testing Fresh Concrete. Part 6: Density. IPQ, Lisbon, Portugal.
NP EN 12390e3, 2009. Testing Hardened Concrete. Part 3: Compressive Strength of
Test Specimens. IPQ, Lisbon, Portugal.