Journal of Cleaner Production 182 (2018) 427e436

Contents lists available at ScienceDirect

Journal of Cleaner Production

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

Development of a new supplementary cementitious material from the

activation of copper tailings: Mechanical performance and analysis of
factors
Felipe Vargas a, **, Mauricio Lopez a, b, *
a
Department of Construction Engineering and Management, School of Engineering, Pontificia Universidad Catolica de Chile, Vicuna Mackenna 4860, Chile
b
Center for Sustainable Urban Development (CEDEUS), Pontificia Universidad Catolica de Chile, Vicuna Mackenna 4860, Casilla 306, Correo 22, Santiago,
Chile

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

Article history: The use of tailings as aggregate and supplementary cementitious materials has been studied previously.
Available online 3 February 2018 Nevertheless, tailings are generally used as collected, without treatment, showing low cementitious
capacity allowing low replacement levels of cement (below 15%). This research studies eight copper
Keywords: tailings to determine which tailings are likely to improve their cementitious capacity as supplementary
Tailing cementitious material by using thermal and mechanical treatments. In the first stage, using TGA and PSD,
Heat treatments
the capacity of the tailings to undergo chemical transformations through a thermal treatment of up to
Mechanical treatments
1000
C and physical transformation through low-energy milling for up to 180 min, were explored.
Relevant chemical modifications were observed over the range of 500
Ce900
C, with peaks between
600
C and 800
C. It was also observed that due to the nature of rock processing prior to mineral
extraction, the particle size was able to be reduced by 50% after 60 min of milling or less. With these
results, a central composite design was proposed using thermal treatment temperatures between 600
C
and 800
C (central point at 700
C) and milling with a central point at 30 min. The results show that the
mechanical performance of the mixtures can be improved by up to 40% at 90 days compared to untreated
tailings at a 40% replacement level.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction 2009), phosphates (Zheng et al., 2015), iron (Zhao et al., 2014),
and copper (Onuaguluchi and Eren, 2012) both as SCM and as a fine
The impact of cement production on the environment is one of aggregate.
the main factors in the search for new materials and construction The copper mining industry is one of the most important
methods. The production of 1 metric ton of Portland cement in- extractive industries worldwide, with production levels of
volves the emission of nearly 1 t of CO2 to the atmosphere: approximately 19.4 million t in 2016 (U.S. Geological Survey, 2017).
approximately 50% comes from the decomposition of carbonates Each ton of refined copper generates 196.5 tons of tailing
(Schneider et al., 2011). Globally, clinker production accounts for 5% (Onuaguluchi and Eren, 2012) andddepending on the technique
of total CO2 emissions (Ludwig and Zhang, 2015). Supplementary applied, which is associated with the type of rockdbetween 2.2
cementitious materials (SCM) have been used to replace part of the and 3 t of copper slag (Shi et al., 2008). The tailings are a mixture of
cement (Schneider et al., 2011) to reduce the impact of the pro- by-products generated from the extraction process from the ores
duction of cement. In recent developments, the use of tailings as using mineral and hydrometallurgical processes (Wang et al., 2014),
SCM has been studied: the use of tungsten tailings (Choi et al., specifically when the crushed rock enters the separation process
performed by floating of the “copper concentrate”, in which the
fraction of high copper concentrations (30% by mass) is further
* Corresponding author. Department of Construction Engineering and Manage- processed, and the rest is discarded as tailing (Davenport et al.,
ment, School of Engineering, Pontificia Universidad Catolica de Chile, Vicuna 2002).
Mackenna 4860, Chile.
The use of copper tailings has been researched both as SCM with
** Corresponding author.
E-mail addresses: [email protected] (F. Vargas), [email protected] (M. Lopez).
Portland cement (Deng et al., 2014; Onuaguluchi and Eren, 2015,

https://doi.org/10.1016/j.jclepro.2018.01.223
0959-6526/© 2018 Elsevier Ltd. All rights reserved.
428 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436

2013, 2012; Thomas et al., 2013a) and as a fine aggregate replace- milling. The combined effect of different treatments has also been
ment (Zhang et al., 2014). Copper tailings have also been investi- documented for some pozzolanic and non-pozzolanic materials
gated as a principal element in the production of bricks through a (Jain, 2012).
geopolymerization process (Ahmari and Zhang, 2012). In blends The aim of this research was to scientifically define promising
with Portland cement and tailings, one study (Onuaguluchi and treatments to maximize the cementitious activity of copper tailings
Eren, 2012) indicated that at replacement levels of 10% by mass, to enable their feasible use as SCM in concrete. Therefore, a first
concrete strength exceeds the normal performance by 16.2% over stage was developed on tailings to observe their performance as
28 days. However, it has been demonstrated that with higher SCM as found without any treatment. The optimal temperatures
replacement levels (until 20%), there is no sizable impact or even and the appropriate grinding time needed to obtain changes in
loss in compressive strength over 28 days up to 15% (Zhang et al., particle size were then determined. Finally, a central composite
2014). factorial design was used to observe the effect of each treatment
The effect of the use of any SCM in concrete can be divided into separately and in combination on selected tailings.
two contributions occurring simultaneously (Juenger and Siddique,
2015; Thomas, 2013): 2. Materials and methods

a Chemical contribution: also called the pozzolanic effect, related Eight samples of tailings were collected from different sources
to the reaction and hydration products of portlandite and in Chile and used in this study in a first exploratory stage. Six of the
aluminosilicate phases present in the SCM's (Lothenbach et al., tailings came from long-standing tailings (LS) in Copiapo, Chile and
2011). two from ongoing production (OP) mines in Salamanca and Til-Til,
b Physical contribution: also called the filler effect, related to the Chile. Among the eight original tailings, three were selected to
space between particles and the surface in the hydration assess the effect of the treatments in their performance as
products (Lothenbach et al., 2011) which can be separated from supplementary cementitious materials. The tailings were collected
the chemical effect (Zunino and Lopez, 2016). by means of sampling in at least three locations of the deposit and
then homogenized by mixing them in one representative sample of
Considering both effects and the poor performance shown by 40 kg. For this research, only the fraction of tailing passing through
copper tailings as SCM at higher replacement levels (Thomas et al., a No. 30 sieve (600 mm) was considered; i.e., the larger particles
2013b; Zhang et al., 2014), activation stands as an alternative to (approximately 15%) were discarded. For the experimental program
transformation into a more attractive cementing material (Peng conducted, type I OPC was used.
et al., 2015). The improvement of performance in supplementary
cementitious materials has been researched since the mid-20th 2.1. Chemical characterization
century, with the use of thermal treatments. Optimal tempera-
tures for chemical and mineralogical changes with the potential to The chemical composition of the tailings was determined by
improve the chemical effect of SCM are 800
for andesite (Hamidi X-ray fluorescence (XRF), using a Bruker S8 Tiger spectrometer, and
et al., 2013) and 650e700
for kaolinite and kaolinite-bentonite the results are shown on Table 1. The specific gravity (SG) of the raw
(Souri et al., 2015; Taylor-Lange et al., 2015). However, other SCM materials is included in the table.
show lower improvements with calcination: zeolites have been The mineralogy composition of the tailings was determined by
researched with temperatures above 300
without showing great qualitative X-ray diffraction (XRD), using a Miniflex diffractometer
pozzolanic capacity after the treatment, and changes are more of 30 kV, 10 Maq; Rigaku with a Cu Ka source. The diffractograms
closely related due to morphology and particle size than chemical were recorded between angles 2q ¼ 5
and 40
with a scan rate of
contribution (Kuçukyıldırım and Uzal, 2014; Seraj et al., 2016). 2
/min at room temperature.
Changes in particle size also have been researched for SCM as an
effective method to improve performance, generally with high- 2.1.1. Thermogravimetric analyses (TGA)
energy methods (Burgos et al., 2014; Fuentes et al., 2014). The Simultaneous thermal gravimetric analysis (TGA) and differen-
combined effect of thermal and milling treatments has increased tial scanning calorimetry (DSC) were performed on the eight
effectiveness, with particle size above 45 mm (Hamidi et al., 2013; untreated tailings to understand how thermal treatment affects
Tironi et al., 2013). Treatments without high-temperature pro- phase decomposition and to determine the best treatment tem-
cesses also show improvements in the reactivity of zeolites (Burris perature. The TGA/DSC tests were performed on crushed samples,
and Juenger, 2016) and kaolin (Ili
c et al., 2016) when combined with passed through a No. 200 sieve (75 mm opening), using a TA

Table 1
Chemical composition of OPC and tailings. Chemical composition data are presented as oxide percentage by weight.

Percentage of Oxides (%) OPC OP1 OP2 LS1 LS2 LS3 LS4 LS5 LS6

SiO2 20.3% 62.90% 59.90% 49.40% 60.20% 51.90% 73.70% 42.60% 31.8%
Al2O3 4.9% 9.50% 16.90% 10.30% 12.20% 16.70% 5.40% 8.80% 6.9%
Fe2O 3.1% 12.60% 2.80% 18.70% 7.00% 10.60% 12.50% 11.00% 37.9%
CaO 63.0% 2.80% 2.70% 5.10% 5.90% 4.40% 1.80% 12.70% 8.6%
MgO 0.3% 1.10% 1.90% 2.70% 1.70% 3.80% 0.50% 2.50% 2.4%
SO3 2.5% 0.70% 1.30% 2.60% 1.40% 0.40% 0.40% 1.80% 0.1%
Na2O 0.2% 1.30% 3.20% 1.30% 2.30% 3.80% 0.50% 1.60% 2.7%
K2O 0.8% 3.30% 4.10% 2.60% 2.60% 2.90% 1.30% 1.80% 1.8%
TiO2 0.4% 0.20% 0.70% 0.50% 0.70% 1.10% 0.40% 0.40% 0.5%
P2O5 0.1% 0.10% 0.30% 0.20% 0.30% 0.40% 0.10% 0.10% 0.1%
MnO 0.1% 0.10% <LOD 0.20% 0.10% 0.30% 0.10% 0.70% 0.20%
CuO <LOD 0.70% 1.00% 0.30% 0.80% 0.70% 0.40% 0.20% 0.30%
LOI 2.3% 4.40% 5.30% 5.80% 4.20% 2.80% 3.00% 14.50% 6.4%
Specific Gravity 3.124 3.01 2.992 2.925 2.79 2.778 2.887 2.864 3.211
 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436 429

Instruments TGA/DSC Q600. During the test, the chamber gas used plastic containers until the date of mixture.
was nitrogen at a 50 mL/min flow. The samples were stabilized in
alumina crucibles at 30
C for 10 min and then heated from 30
C to 2.5. Isothermal calorimetry
1000
C at a rate of 20
C/min.
Isothermal calorimetry tests were conducted to assess the effect
2.2. Physical characterization of the treatments at the central point and compare it with the
untreated tailings, analysing only the pozzolanic capacity of the
Particle size distributions (PSD) of the eight tailings and OPC tailings. The measurements were taken using a TAM Air isothermal
were measured using a Malvern Mastersizer 2000 laser diffrac- calorimeter using the procedure proposed previously (Avet et al.,
tometer. Isopropanol (refractive index 1.378) was used as a 2016; Snellings and Scrivener, 2015) at 20
C, in mixes with por-
dispersant, and the PSD was measured for 12 s five times while tlandite and calcium sulphate (LC3-50 system), with single
stirring at 2000 rpm to obtain an average measurement. The PSD replication.
results for the untreated tailings are shown on Fig. 1. For the SEM
images, samples were dried and gold-coated. Magnifications be- 3. Exploration of factors: results and discussion
tween 300  and 5000  with an accelerating voltage of 15 kV were
used. In this stage, an exploratory procedure was developed to obtain
a relevant experimental region of interest for the activation stage
2.3. Tailing strength development and the experimental design.
The first step was a review of the compressive strength of paste
Mixture proportioning was designed by mass to avoid the effect mixes of untreated tailing at 7, 28 and 90 days, at a replacement
of volume changes due to thermal and mechanical treatments. The levels of 20%, 40% and 50% by mass.
water-to-cementitious (w/c) materials ratio of the pastes was fixed The second analysis was related to the use of TGA/DSC curves to
at 0.5 by mass. Three 100% OPC mixtures were selected as refer- obtain a range of temperatures in which chemical changes on
ences, with w/c ratios of 0.5, 0.55 and 0.6. The samples were tested tailings could be observed. With these results, and the analysis of
at 7, 28 and 90 days. the XRD patterns, the potential of the tailings as a supplementary
cementing material could be assessed. With the result of this step,
2.4. Treatment on tailings three tailings were selected considering the potential capacity of
tailings as supplementary cementitious material and their origin, so
The treatments in tailings were performed according to the recent and long-standing tailings were evaluated. The final three
following procedure: first, the tailings in batches of 200 g were tailings were used in a third analysis, which was related to the
introduced in the mill. For the milling procedure, a rotational mill particle size changes due to different milling times. The mill was
was chosen, similar to the conditions present in mills currently operated at 40 rpm, with samples milled for 15, 30, 60, 90, 120, 150
used to grind other SCMs, since this type of milling would be more and 180 min.
feasible to implement in an industrial environment. The thermal
treatment on the tailings was then performed in the same batch of 3.1. Compressive strength
200 g using a programmable Nabertherm HT 40/16 electric muffle
oven. The heating cycle started at room temperature and proceeded The results of the untreated tailings pastes are shown in Fig. 2. It
at 10
C/min to the maximum temperature indicated by the can be seen that depending on the tailing, the results vary widely.
experimental program. The maximum temperature was stabilized All tailings exhibit lower strength ratios than the OPC paste at the
for 20 min, and then the oven was left to cool to room temperature same w/c ratio, the expected result considering the replacement
at a free rate. The resulting calcined tailings were stored in sealed (Onuaguluchi and Eren, 2013, 2012; Thomas et al., 2013b).

Fig. 1. Particles size distributions (PSD) of OPC and 8 tailings.

430 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436

0,9
20% 40% 50%
0,8 W/C 0.55

0,7 W/C 0.6

Rtailing/RPC 7 days 0,6 W/C 0.625

0,5

0,4

0,3

0,2 W/C 0.833

W/C 1
0,1

0
OP1 OP2 LS1 LS2 LS3 LS4 LS5 LS6
0,9
20% 40% 50%
0,8
W/C 0.55
0,7
Rtailing/RPC 28 days

W/C 0.6
0,6
W/C 0.625

0,5

0,4

0,3

0,2 W/C 0.833

0,1 W/C 1

0
OP1 OP2 LS1 LS2 LS3 LS4 LS5 LS6
0,9
20% 40% 50%
W/C 0.55
0,8

0,7 W/C 0.6

Rtailing/RPC 90 days

W/C 0.625
0,6

0,5

0,4

0,3
W/C 0.833
0,2
W/C 1
0,1

0
OP1 OP2 LS1 LS2 LS3 LS4 LS5 LS6
Fig. 2. Compressive strength ratios (referenced to the 100% OPC mixture) versus tailing mixtures at 7, 28 and 90 days for 20%, 40% and 50% replacement levels.

Specifically, reductions in compressive strength between 20% and between OPC paste at w/c of 0.4 and the pastes with tailings tend to
80% were obtained depending on the tailing and replacement level. decrease, which is attributable to the presence of phases capable of
This shows that the base cementitious capacity of the tailings generating pozzolanic reactions in the tailings (Aïtcin, 2016;
can vary importantly. At later ages, the gap in compressive strength Lothenbach et al., 2011).
 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436 431

If the comparison is made against OPC pastes considering only carbonate minerals such as calcite and dolomite (Cao et al., 2016;
the cement reductiondi.e., an increase to a new w/c ratio (dashed Hamidi et al., 2013). In the case of tailings LS1, LS2 and LS5, this
lines)dthe improvement in compressive strength is important. mass loss is between 2% and 6%. With these results, the chosen
This means that the use of raw tailings in addition to the cement, tailings were OP1, LS1 (from the previous step) and LS5. As the
instead of replacing cement, improves compressive strength results show major mass loss at temperatures between 500
C and
significantly. For example, tailings OP1, OP2, LS2 and LS3 show 800
C, the experimental space of interest was chosen between
important increases in compressive strengths in the analysis 600
C and 800
C, centred at 700
C. At these temperatures, major
compared to their respective dash lines. improvement is expected in the performance of the tailings as SCM.
For the next stage, mixtures with a 40% replacement level will The XRD patterns (Figs. 4e6) show similar results to TGA, with
be used to assess the improvements due to the treatments more the previous selected temperatures of thermal treatment. First, the
precisely. In addition, tailings OP1 and LS1 were selected due to high crystallinity of the minerals of the tailings was reported pre-
their performance and their difference in precedence (one is from viously, with the presence of kaolin, calcite and dolomites, due to
an ongoing production mine, and the other is from a long-standing the nature of the rock from which copper is extracted in Chile
tailing). (Smuda et al., 2014, 2008). Second, it can be seen that at 600
C, the
kaolin phase peaks disappear due to the change into amorphous
phases (Hamidi et al., 2013; Taylor-Lange et al., 2015). At 800
C,
3.2. Temperature: TGA-DSC and XRD pattern finally, the calcite and dolomite peaks disappear due to dehy-
droxylation of these phases (Hamidi et al., 2013). The presence of
TGA-DSC results obtained for all tailing samples are shown in
Fig. 3. The results show the differences between the tailings
regardless of the similarities in chemical composition. Neverthe-
less, there are some common peaks in mass loss and endothermic
peaks. After the mass loss due to water content below 150
C, the
first mass loss and endothermic peak, between 400
C and 640
C,
is related to the transformations of kaolinic phases present in the
tailings, due to dehydroxylation and the loss of hydroxyl groups
(Cao et al., 2016; Hamidi et al., 2013; Taylor-Lange et al., 2015). Due
to the low amount on the mass loss (below 3%), it is expected that
the pozzolanic activity of the tailings will be low (Teklay et al.,
2014). The second endothermic peak and mass loss is centred at
750
C. This temperature is related to the decarbonation of

Fig. 4. XRD Pattern of OP1 tailing, untreated and calcined at 600, 700 and 800
C (Q:
quartz; Cr: cristobalite; Z: zinc oxide (i.s.); F: feldspar).

Fig. 3. Thermogravimetric curve (TGA) and differential scanning calorimetry (DSC) Fig. 5. XRD pattern of LS1 tailing, untreated and calcined at 600, 700 and 800
C (Q:
curves of 8 tailings. quartz; Cr: cristobalite; A: anorthite; Z: zinc oxide (i.s.); F: feldspar).
432 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436

Fig. 6. XRD pattern of LS5 tailing, untreated and calcined at 600, 700 and 800
C (Q:
quartz; a: albite; Z: zinc oxide (i.s); F: feldspar).

another clay mineral, montmorillonite, is also noted, whose struc-

ture also changes between 600
C and 800
C (Taylor-Lange et al.,
2015). Even with this change, the total amount of this phase by
weight (measured in TGA) is not expected to exceed 8%, which is
consistent with previous tailing measurements in Chile (Smuda
et al., 2014, 2008).
Fig. 8. SEM microphotographs of tailings (1: OP1; 2: LS1; 3: LS5). Particle micro-
fractures are observed in the LS5 image, some of them denoted by arrows.
3.3. Grinding time: particle size distribution (PSD) and SEM images

In Fig. 1, it is observed that the PSD of untreated tailings has milling treatments such as those subjected to the mineralogical
values comparable to those of other supplementary cementitious process. In image number 3 (LS5 tailing), the presence of micro-
materials (Thomas, 2013). The results of the exploratory milling fractures that facilitate the milling process of the tailings has been
process of 3 tailings can be seen in Fig. 7. There is a significant emphasized.
decrease in particle size in the first 60 min of milling for the old
tailings. In the case of mine tailings in production, the reduction is 3.4. Experimental design for treated tailing pastes
smaller because the particle size is already small when the process
is started. Due to the differences in the rock process and the For the next stage, the results of the exploration of factors were
development of the mine production, the particle size of the tail- considered to obtain an experimental region of interest. The chosen
ings has been reduced to optimize the extraction of copper from the replacement level was 40% by mass. The range of temperature
ores for newer productions (Davenport et al., 2002). The results considered for the thermal treatment was centred at 700
± 100
.
show that with 60 min max. Milling, the particle size decrease is The base range of time for the milling treatment was centred at
important. The times chosen for the next step were centred at 30 min ± 21.22 min. A second-order model of response surface
30 min. called central composite design for two factors is proposed, which
In Fig. 8, the SEM images shows the morphology of the tailing allows high levels of rotatability of factors. The model considers a
particles, which follow the characteristic pattern of particles from significance level of 95%. Another advantage of the chosen

Fig. 7. Particle size fraction d(0.9) and d(0.5) of 3 tailings after grinding.
 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436 433

showing a chemical transformation of phases between 600
C and

800
C. The endothermic peak observed was at 550
C, indicating
that in the experimental space selected, the thermal effect may be
hidden in the average value obtained by the regression. Although
this tailing showed a smaller particle size decrease when analysing
the grinding times (Fig. 7), the statistical analysis suggests that it
may have been acted on by more reactive particles, which allow the
increase and the positive effect on the result by increasing the
milling times.
In the case of LS5 tailing, the estimates of the regression show
that at early ages, the grinding time and temperature are statisti-
cally significant, which agrees with the previous observation in the
TGA and XRD test (Figs. 3 and 5). Those figures show important
Fig. 9. Central composite design and experimental region covered. mass losses in the experimental region of interest, related to
changes in crystalline structures present in the tailings and changes
in particle size due to the milling procedure. At 90 days, the
experimental design is that it allows better mapping of the chosen
grinding time loss relevance becomes statistically insignificant.
experimental area and determines a direction of maximum
This agrees with what has been reported on the chemical and
response (Fig. 9) (Montgomery, 2012). In the case of the grinding
physical effects of supplementary cementitious materials
time, the factorial points were chosen to allow the use of the star
(Lothenbach et al., 2011; Zunino and Lopez, 2016), where it has
points (axial points) at 0 and 60 min. This is due to the mathe-
been observed that the physical effect has a greater impact at early
matical formulation of the model, which requires a values of 1.414
ages, while the chemical effect has a greater impact at later ages.
times the distance to the factor points of the design. The design
The estimates for temperature also show a maximum in the
factor levels and runs for the studied experimental region of in-
experimental region of interest, indicating that thermal treatment
terest are given in Table 2. Three replications were made for each
above 750
C decreases compressive strength. At that temperature,
point of the design.
the decarbonation of dolomite and calcite contributes to the for-
mation of the calcium and magnesium oxides, and the effect of
4Treatment on tailing: results and discussion these oxides at later ages is less relevant for the pozzolanic activity.
Finally, the LS1 tailing shows a similar behaviour to the LS5
4.1. Experimental design for treated tailing pastes tailings, explained partially by their similarities in TGA and particle
size reduction. However, the differences between statistically sig-
The individual compressive strength obtained for each tailing at nificant factors in each case and the differences between the esti-
7 and 90 days can be seen in Table 3. The results are expressed as mates are related to the different chemical compositions between
ratios between the compressive strength of a treated tailing and LS1 and LS5. In the case of the LS1 tailing, the untreated tailing had
that obtained with the same tailing, but untreated; thus, the im- a higher compressive strength than the LS5 tailing, so it is possible
provements due to treatments can be easily observed. As expected, to assume that with this tailing, there is no severe impact on the
the results vary widely among tailings. Table 4 presents the resistance because we can already be at the limit of its capacity as
regression-coded estimates for each tailing and the statically sig- supplementary cementitious material for the chosen experimental
nificant estimates. region. In this case, also, the tailing does not have a maximum in the
No correlation was found between the thermal treatment and experimental region of interest, which indicates that for this case,
the grinding time for the three tailings due to the small amount of there are temperatures higher than those of the conducted test that
material that effectively undergoes transformations due to the high allow its increased capacity as SCM.
crystallinity and presence of crystalline phases of silicon oxide in By analysing the obtained resistances (Fig. 10), it can be seen
the tailings (Figs. 4e6). In addition, the changes in particle size that the final resistance can increase by up to 40% at 7 days and up
before and after the thermal treatment were measured, but no to 35% related to the resistance of the untreated tailings. Mixtures
statistically significant differences in particle size were found. such as these with high replacement levels and low relative
For the OP1 tailing, there is a general decrease in the ratios strengths show a significant improvement when using treated
between 7 and 90 days, and the significant estimates are in both tailings.
ages and grinding times. With the TGA analysis (Fig. 3), the OP1 test
shows a loss mass less than 2%, without any endothermic peak 4.2. Isothermal calorimetry

The increase in the pozzolanic activity related to the activation

Table 2
Design factor levels and runs at 40% replacement level.
of the tailings can be seen in Fig. 11a. Although the tailings show a
low pozzolanic capacity, mainly due to the absence of reactive
Run Experimental Factors
phases and the high crystallinity of the compounds present in
Coded Factors
them, an improvement in the pozzolanic capacity of the tailings can
A B Temperature C
Grinding Time min be observed due the proposed treatments. When comparing the
1 e e 600 8.78 results for the 3 tailings, those with greater increases in their
2 þ e 800 8.78 pozzolanic capacity (LS1 and LS5) are the same as those with
3 þ 600 51.21
e
greater losses of mass measurable by TGA (Fig. 3) until the analysed
4 þ þ 800 51.21
5 a 0 558.6 30 temperature of 700
C (central point of the experimental region of
6 a 0 841.2 30 interest).
7 0 a 700 0 Chemical transformations are considered to occur at tempera-
8 0 a 700 60 tures up to 650
C for the kaolinitic phases, which indicates a
9e12 0 0 700 30
greater contribution to the pozzolanic activity (Taylor-Lange et al.,
434 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436

Table 3
Compressive strength ratios (referenced to the same tailing without treatment) versus mixtures at 7 and 90 days of age, 40% replacement level.

Tailing Factor Levels 40% Replacement Level

OP1 LS1 LS5

Run Temperature C Grinding Time min 7 days 90 days 7 days 90 days 7 days 90 days

1 600 8.78 1.32 0.99 0.67 1.04 1.27 1.04

2 800 8.78 0.57 1.19 0.73 1.20 1.44 1.25
3 600 51.21 1.45 1.61 1.03 1.04 1.47 1.14
4 800 51.21 1.41 1.33 1.02 1.16 1.53 1.56
5 558.6 30 1.44 1.07 0.64 1.08 1.17 1.20
6 841.4 30 1.31 1.12 0.27 1.17 1.46 1.36
7 700 0 1.09 1.04 0.82 1.06 1.35 1.38
8 700 60 1.23 1.30 1.13 1.01 1.58 1.36
9 700 30 1.33 1.26 0.91 0.99 1.66 1.43
10 700 30 1.33 1.23 0.85 1.08 1.57 1.47
11 700 30 1.32 1.22 0.65 1.08 1.62 1.41
12 700 30 1.31 1.23 0.69 1.08 1.71 1.49

Table 4
Estimates and significance of coded estimates on regression models for treatments on tailings, at 95% significance level, for 7 and 90 days. Boldface indicates
estimates that are statistically significant.

Source Estimate (Compressive Strength Ratio)

OP1-7d OP1-90 LS1-7d LS1-90 LS5-7d LS5-90

Average 1.3213 1.2354 0.7777 1.0544 1.6435 1.4503

A: Temperature 0.2417 0.0040 ¡0.1186 0.0999 0.1614 0.2150
B: Grinding Time 0.2896 0.2813 0.2738 0.0240 0.1561 0.0969
A*A 0.0124 0.0632 ¡0.2490 0.0853 0.3114 ¡0.2090
A*B 0.3546 0.2394 0.0333 0.0224 0.0517 0.1058
B*B 0.1995 0.0092 0.2742 0.0034 0.1593 0.1198

400
C and 650
C (Souri et al., 2015; Taylor-Lange et al., 2015;

Teklay et al., 2014). Even though the mass loss in the 150e650
C
range may contain the decomposition of phases that promote hy-
draulic reactions, it is still a promising way to quantify the pozzo-
lanic potential of a treated tailing.

5. Conclusions

Sustainability in the construction and cement industry may

have points in common with waste disposal of other industries at
Fig. 10. Compressive strength ratios (referenced to the 100% OPC mixture) versus important levels, such as the copper mining industry. The use of
tailing mixtures at 7 and 90 days for 40% replacement level for untreated and treated
tailings, maximum value for each regression model.
tailings as a cement replacement is one alternative to reduce the
impact of the production of cement and management of the
disposal of tailings into the environment.
The present study conducted a chemical and physical charac-
terization of eight tailing samples. The samples were used as
collected, and the treatments and procedures could be scaled to
industrial levels.
Based on the findings presented in this study, the following
conclusions can be stated:

1. Each tailing has a different behaviour under compressive

strength, according to its chemical composition and physical
characteristics. With the tailings and tests carried out herein, it
is impossible to conclude that there are quantifiable factors in
Fig. 11. 1. Cumulative heat (J/g of SCM) of untreated (dashed line) and treated (700
C,
30 min grinding) tailings. 2. Relation between mass loss under 650
C and increase in common to explain this behaviour, being necessary a case-by-
pozzolanic capacity, measured as cumulative heat release at 7 days. case analysis to determine potential use as supplementary
cementing material and the levels, experimental space and ef-
fect of treatments on each tailing.
2015). A strong relation between the mass loss and the heat release 2. At replacement levels over 20%, the decrease in compressive
results obtained for these 3 tailings can be seen. The mass loss, strength is important. Nevertheless, the decrease is less than the
considered between 150
C and 650
C, coincides with the tem- lower bound case, which is to extract the portion of cement
peratures where the most important chemical transformations are without replacing it with tailings. This suggests an important
observed from the pozzolanic point of view, especially between potential use of the untreated tailings in addition to
 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436 435

cementdfor example, as fine aggregatedand of the treated References

tailings as fine aggregate with low contribution as SCM.
3. With TGA analysis, it is possible to estimate the potential ben- Ahmari, S., Zhang, L., 2012. Production of eco-friendly bricks from copper mine
tailings through geopolymerization. Construct. Build. Mater. 29, 323e331.
efits in cementitious activity of applying the thermal treat- https://doi.org/10.1016/j.conbuildmat.2011.10.048.
ments. Moreover, it was possible to link these benefits with the Aïtcin, P.-C., 2016. Supplementary cementitious materials and blended cements. Sci.
analysis of the DSC curves and with XRD patterns. In the case of Technol. Concr. Admixtures. https://doi.org/10.1016/B978-0-08-100693-
1.00004-7.
the tailings analysed in this study, the mass losses observed at Avet, F., Snellings, R., Alujas Diaz, A., Ben Haha, M., Scrivener, K., 2016. Development
550
C and 740
C were related with chemical transformations of a new rapid, relevant and reliable (R3) test method to evaluate the pozzo-
of calcium and kaolinitic phases, which improve the behaviour lanic reactivity of calcined kaolinitic clays. Cement Concr. Res. 85, 1e11. https://
doi.org/10.1016/j.cemconres.2016.02.015.
of the tailings as SCM. Burgos, D.M., Cardona, L.M., Delvasto, S., 2014. Estudio de dos materiales volca
nicos
4. Due to the nature of the mechanical processing of the rock in the y efecto del tipo de molienda en su reactividad. Rev. Ing. Construccio
n 29,
extraction of the copper minerals, the grinding of tailings de- 159e174. https://doi.org/10.4067/S0718-50732014000200003.
Burris, L.E., Juenger, M.C.G., 2016. Milling as a pretreatment method for increasing
mands relatively low energy to reduce the size of the particles in
the reactivity of natural zeolites for use as supplementary cementitious mate-
the long-standing tailings. Although the relevance of the rials. Cement Concr. Compos. 65, 163e170. https://doi.org/10.1016/
grinding process is less effective in new tailings, the time of j.cemconcomp.2015.09.008.
grinding is statistically significant to increase the reactivity of Cao, Z., Cao, Y., Dong, H., Zhang, J., Sun, C., 2016. Effect of calcination condition on
the microstructure and pozzolanic activity of calcined coal gangue. Int. J. Miner.
those tailings. Process. 146, 23e28. https://doi.org/10.1016/j.minpro.2015.11.008.
5. No statistically significant correlation was found between ther- Choi, Y.W., Kim, Y.J., Choi, O., Lee, K.M., Lachemi, M., 2009. Utilization of tailings
mal and mechanical treatments. This might be explained by the from tungsten mine waste as a substitution material for cement. Construct.
Build. Mater. 23, 2481e2486. https://doi.org/10.1016/
low percentage of material that undergoes chemical trans- j.conbuildmat.2009.02.006.
formations due to thermal treatments, so the correlation loses Davenport, W.G., King, M., Schlesinger, M., Biswas, A.K., 2002. Overview. Extr.
relevance. Metall. Copp. 1e16. https://doi.org/10.1016/B978-008044029-3/50004-6.
Deng, Y., Zhang, C., Wei, X., 2014. Influence of lithium sulfate addition on the
6. The impact of the thermal treatments is statistically significant properties of Portland cement paste. Construct. Build. Mater. 50, 457e462.
at later ages when there are changes in the chemical phases https://doi.org/10.1016/j.conbuildmat.2013.09.058.
present in the tailings due to thermal treatments. On the other Fuentes, M., Zún ~ iga, A., Díaz, M., Rocha, E., Díaz, S., 2014. Molienda mec
anica por
alta energía de minerales mexicanos para producir concreto de alto desempen ~o
hand, the impact of the mechanical treatment is statistically (CAD). Rev. Ing. Construccio
n 29, 256e269. https://doi.org/10.4067/S0718-
significant at early ages. 50732014000300003.
7. Proper use of different analytical techniques such as TGA, XRD, Hamidi, M., Kacimi, L., Cyr, M., Clastres, P., 2013. Evaluation and improvement of
pozzolanic activity of andesite for its use in eco-efficient cement. Construct.
XRF, PSD and calorimetry can predict the SCM performance of
Build. Mater. 47, 1268e1277. https://doi.org/10.1016/j.conbuildmat.2013.06.013.
the tailings and improve the cementing capacity as SCM by Ili

c, B., Radonjanin, V., Malesev, M., Zduji
c, M., Mitrovi
c, A., 2016. Effects of me-
changing the heating and grinding conditions of the chanical and thermal activation on pozzolanic activity of kaolin containing
treatments. mica. Appl. Clay Sci. 123, 173e181. https://doi.org/10.1016/j.clay.2016.01.029.
Jain, N., 2012. Effect of nonpozzolanic and pozzolanic mineral admixtures on the
8. Different treatments may impact the behaviour of tailings as hydration behavior of ordinary Portland cement. Construct. Build. Mater. 27,
SCM differently. However, it is feasible to improve the use of 39e44. https://doi.org/10.1016/j.conbuildmat.2011.08.006.
tailings as SCM by searching for optimal levels for both thermal Juenger, M.C.G., Siddique, R., 2015. Recent advances in understanding the role of
supplementary cementitious materials in concrete. Cement Concr. Res. 78,
and mechanical treatments, and the impact of these treatments 71e80. https://doi.org/10.1016/j.cemconres.2015.03.018.
can improve the mechanical performance of tailings up to 40% at Kuçukyıldırım, E., Uzal, B., 2014. Characteristics of calcined natural zeolites for use
90 days. in high-performance pozzolan blended cements. Construct. Build. Mater. 73,
229e234. https://doi.org/10.1016/j.conbuildmat.2014.09.081.
Lothenbach, B., Scrivener, K., Hooton, R.D., 2011. Supplementary cementitious ma-
Future research needs to explore alternatives to (i) further terials. Cement Concr. Res. 41, 1244e1256. https://doi.org/10.1016/
improve the cementitious and pozzolanic such as chemical com- j.cemconres.2010.12.001.
Ludwig, H.-M., Zhang, W., 2015. Research review of cement clinker chemistry. Cem.
pounds, alkaline activators, different heating and cooling rates, Concr. Res. https://doi.org/10.1016/j.cemconres.2015.05.018.
total time of treatments, (ii) advanced characterization of the Montgomery, D.C., 2012. Design and analysis of experiments. Design. https://
tailing for identifying chemical or physical markers that could doi.org/10.1198/tech.2006.s372.
Onuaguluchi, O., Eren, O., 2013. Rheology, strength and durability properties of
predict the potential capacity of a tailing as a SCM; and (iii) study of
mortars containing copper tailings as a cement replacement material. Eur. J.
other relevant properties at the mortar and concrete scale, such as, Environ. Civ. Eng 17, 19e31. https://doi.org/10.1080/19648189.2012.699708.
rheology, workability, compressive strength, effect of free CaO and Onuaguluchi, O., Eren, O., € 2015. Reusing copper tailings in concrete: corrosion

MgO and other possible phases generated by the treatments, performance and socioeconomic implications for the Lefke-Xeros area of
Cyprus. J. Clean. Prod. https://doi.org/10.1016/j.jclepro.2015.09.036.
durability among other. Onuaguluchi, O., Eren, O., € 2012. Cement mixtures containing copper tailings as an
additive: durability properties. Mater. Res. 15, 1029e1036. https://doi.org/
10.1590/S1516-14392012005000129.
Peng, K., Yang, H., Ouyang, J., 2015. Tungsten tailing powders activated for use as
Acknowledgements cementitious material. Powder Technol. 286, 678e683. https://doi.org/10.1016/
j.powtec.2015.09.012.
This project was supported by project CONICYT/FONDECYT Schneider, M., Romer, M., Tschudin, M., Bolio, H., 2011. Sustainable cement pro-
ductiondpresent and future. Cement Concr. Res. 41, 642e650. https://doi.org/
1150251 and the doctoral scholarships CONICYT-PCHA/Doctorado 10.1016/j.cemconres.2011.03.019.
Nacional/2015e21150946 as well as by CEDEUS, CONICYT/FON- Seraj, S., Ferron, R.D., Juenger, M.C.G., 2016. Calcining natural zeolites to improve
DAP 15110020. The testing performed in this study was mainly their effect on cementitious mixture workability. Cement Concr. Res. 85,
102e110. https://doi.org/10.1016/j.cemconres.2016.04.002.
conducted at the Construction Materials Laboratories at Pontificia Shi, C., Meyer, C., Behnood, A., 2008. Utilization of copper slag in cement and
Universidad Catolica de Chile, Dictuc Laboratories in Chile, and in concrete. Resour. Conserv. Recycl. 52, 1115e1120. https://doi.org/10.1016/
the Construction Materials Research Laboratories at the University j.resconrec.2008.06.008.
Smuda, J., Dold, B., Spangenberg, J.E., Friese, K., Kobek, M.R., Bustos, C.A., Pfeifer, H.-
of Texas at Austin. Authors acknowledge the assistance provided
R., 2014. Element cycling during the transition from alkaline to acidic envi-
by the personnel working at these facilities and Prof. Maria ronment in an active porphyry copper tailings impoundment, Chuquicamata,
Juenger at the University of Texas at Austin in conducting this Chile. J. Geochem. Explor. 140, 23e40. https://doi.org/10.1016/
j.gexplo.2014.01.013.
study and the seed fund project supported by Corfo Engineering
Smuda, J., Dold, B., Spangenberg, J.E., Pfeifer, H.-R., 2008. Geochemistry and stable
2030 initiative.
436 F. Vargas, M. Lopez / Journal of Cleaner Production 182 (2018) 427e436

isotope composition of fresh alkaline porphyry copper tailings: implications on Raton.

sources and mobility of elements during transport and early stages of deposi- Tironi, A., Trezza, M.A., Scian, A.N., Irassar, E.F., 2013. Assessment of pozzolanic
tion. Chem. Geol. 256, 62e76. https://doi.org/10.1016/j.chemgeo.2008.08.001. activity of different calcined clays. Cement Concr. Compos. 37, 319e327. https://
Snellings, R., Scrivener, K.L., 2015. Rapid screening tests for supplementary doi.org/10.1016/j.cemconcomp.2013.01.002.
cementitious materials: past and future. Mater. Struct. 49, 1e15. https://doi.org/ U.S. Geological Survey, 2017. Mineral Commodity Summaries 2017. U.S. Geological
10.1617/s11527-015-0718-z. Survey. https://doi.org/10.3133/70140094.
Souri, A., Kazemi-Kamyab, H., Snellings, R., Naghizadeh, R., Golestani-Fard, F., Wang, C., Harbottle, D., Liu, Q., Xu, Z., 2014. Current state of fine mineral tailings
Scrivener, K., 2015. Pozzolanic activity of mechanochemically and thermally treatment: a critical review on theory and practice. Miner. Eng. 58, 113e131.
activated kaolins in cement. Cement Concr. Res. 77, 47e59. https://doi.org/ https://doi.org/10.1016/j.mineng.2014.01.018.
10.1016/j.cemconres.2015.04.017. Zhang, C., Zhou, T., Wu, Q., Zhu, H., Xu, P., 2014. Mechanical performances and
Taylor-Lange, S.C., Lamon, E.L., Riding, K.A., Juenger, M.C.G., 2015. Calcined microstructures of cement containing copper tailings. Asian J. Chem. 26,
kaoliniteebentonite clay blends as supplementary cementitious materials. 1371e1375.
Appl. Clay Sci. 108, 84e93. https://doi.org/10.1016/j.clay.2015.01.025. Zhao, S., Fan, J., Sun, W., 2014. Utilization of iron ore tailings as fine aggregate in
Teklay, A., Yin, C., Rosendahl, L., Bøjer, M., 2014. Calcination of kaolinite clay par- ultra-high performance concrete. Construct. Build. Mater. 50, 540e548. https://
ticles for cement production: a modeling study. Cement Concr. Res. 61e62, doi.org/10.1016/j.conbuildmat.2013.10.019.
11e19. https://doi.org/10.1016/j.cemconres.2014.04.002. Zheng, K., Zhou, J., Gbozee, M., 2015. Influences of phosphate tailings on hydration
Thomas, B.S., Damare, A., Gupta, R.C., 2013. Strength and durability characteristics of and properties of Portland cement. Construct. Build. Mater. 98, 593e601.
copper tailing concrete. Construct. Build. Mater. 48, 894e900. https://doi.org/ https://doi.org/10.1016/j.conbuildmat.2015.08.115.
10.1016/j.conbuildmat.2013.07.075. Zunino, F., Lopez, M., 2016. Decoupling the physical and chemical effects of sup-
Thomas, B.S., Damare, A., Gupta, R.C., 2013. Strength and durability characteristics of plementary cementitious materials on strength and permeability: a multi-level
copper tailing concrete. Construct. Build. Mater. 48, 894e900. https://doi.org/ approach. Cement Concr. Compos. 65, 19e28. https://doi.org/10.1016/
10.1016/j.conbuildmat.2013.07.075. j.cemconcomp.2015.10.003.
Thomas, M., 2013. Supplementary Cementing Materials in Concrete. CRC Press, Boca