Cement & Concrete Composites 56 (2015) 59–72

Contents lists available at ScienceDirect

Cement & Concrete Composites

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

Mechanical properties, durability, and life-cycle assessment

of self-consolidating concrete mixtures made with blended
portland cements containing fly ash and limestone powder
Kemal Celik a, Cagla Meral a,b, A. Petek Gursel a, P. Kumar Mehta a, Arpad Horvath a, Paulo J.M. Monteiro a,⇑
a
Department of Civil and Environmental Engineering, University of California, Berkeley, CA 94720, USA
b
Department of Civil Engineering, Middle East Technical University, 06800 Ankara, Turkey

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

Article history: This paper reports the composition and properties of highly flowable self-consolidating concrete (SCC)
Received 28 January 2014 mixtures made of high proportions of cement replacement materials such as fly ash and pulverized
Received in revised form 5 October 2014 limestone instead of high dosage of a plasticizing agent or viscosity-modifying chemical admixtures.
Accepted 6 November 2014
Self-consolidating concrete mixtures are being increasingly used for the construction of highly reinforced
Available online 15 November 2014
complex concrete elements and for massive concrete structures such as dams and thick foundation. In
this study, by varying the proportion of portland cement (OPC), Class F-fly ash (F), and limestone powder
Keywords:
(L), SCC mixtures with different strength values were produced, and the properties of both fresh and
Self-consolidating concrete (SCC)
Fly ash
hardened concrete were determined. For a comprehensive analysis and quantification of emissions and
Emissions global warming potential (GWP) from concrete production, life-cycle assessment (LCA) was employed.
Life-cycle assessment (LCA) We find that high volume, up to 55% by weight replacement of OPC with F, or F and L produces highly
Limestone powder workable concrete that has high 28-day and 365-day strength, and extremely high to very high resistance
Global warming potential (GWP) to chloride penetration along with low GWP for concrete production.
Sustainability Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Roughly 60% of these emissions come from the calcination of lime-
stone, which is the main raw material for making portland-cement
According to data compiled by the U.S. Geological Survey clinker. For every tonne of calcium carbonate calcined in the kiln to
(USGS) in 2011 [1], the yearly global production of portland form calcium oxide, 440 kg of CO2 are released into the atmosphere
cement was about 3.3 billion metric tons (mt). Considering typical as the chemical reaction progresses. The combustion of fuel
concrete mixture proportions for ordinary concrete [2], this required to generate the heat necessary for the reactions forming
amount of cement is incorporated into approximately 27 billion the clinker minerals accounts for the remaining CO2 emissions.
mt of concrete, which requires 22 billion mt of aggregates and As a result, considering an average clinker factor (kg of clinker
2.2 billion mt of fresh water, leading to an annual global average per kg of cement) of 0.78 [4], annual worldwide CO2 emissions
consumption rate of about 4 mt of concrete per person. The mas- from cement manufacturing add up to almost 2.3 billion mt, which
sive production and consumption cycle of concrete has substantial is nearly 7% of the global emissions from fossil fuel combustion [2].
environmental impact. For an average of 918 kg of CO2 per mt of cement [5], the U.S.
Carbon dioxide (CO2) emissions from portland cement cement industry generated about 56 million mt of CO2 based on
manufacturing is one of the major sustainability issues facing the the 2010 portland cement production rate of 61 million mt [1].
concrete industry. Although considerable gains in energy efficiency These numbers correspond to direct emissions only, i.e., those gen-
during cement production manufacturing have been realized over erated in the cement factory. Based on economic input–output
the last two decades, according to industry data [3], about 866 kg analysis-based life-cycle assessment (EIO-LCA) using U.S. data
of CO2 are being generated for every 1000 kg of clinker made. [6], supply-chain inclusive, life-cycle greenhouse gas emissions
associated with cement manufacturing are expected to be 13%
higher than direct emissions.
⇑ Corresponding author at: Department of Civil and Environmental Engineering,
While the current environmental impact of the concrete indus-
725 Davis Hall, University of California, Berkeley, CA 94720, USA. Tel.: +1 (510) 643
8251; fax: +1 (510) 643 8928. try is indeed considerable, the increased use of supplementary
E-mail address: [email protected] (P.J.M. Monteiro). cementitious materials (SCMs) offers a possible reduction in global

http://dx.doi.org/10.1016/j.cemconcomp.2014.11.003
0958-9465/Ó 2014 Elsevier Ltd. All rights reserved.
60 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72

CO2 emissions. A study involving business leaders and academics Emissions from the concrete mixtures were compared using the
[7] singled out construction materials as one of the seven most ‘‘GreenConcrete LCA’’ tool developed by some of the co-authors of
promising technologies for investment (together with wind, this paper [24]. This cradle-to-gate life-cycle assessment (LCA) tool
biofuels, photovoltaics, and concentrating solar power, nuclear, estimates direct and supply-chain global warming potential (GWP)
and building efficiency). The report concluded that within the con- in units of CO2-equivalent (CO2-eq) emissions and some criteria air
struction materials sector, ‘‘the biggest single opportunity for CO2 pollutants (CO, NOx, PM10, and SO2) associated with the use of elec-
reduction is a low-carbon cement,’’ and that annual savings of tricity, fuel, transportation, and production processes taking place
1 billion mt of CO2 could be reached through the concrete sector within the boundary of concrete production system. In LCA appli-
if 50% of portland cement were replaced by a low-carbon cations, drawing the system boundaries, i.e., decisions on inclusion
alternative. or exclusion of processes in an analysis, is an essential step [25].
In order to achieve such a level of CO2 reductions, the industry This study incorporates the following parameters in the system
must embrace a comprehensive, integrated approach that boundary: extraction of cement raw materials, manufacturing of
necessarily involves the use of less concrete for new structures, cement, extraction and processing of aggregates, manufacturing
consumption of less cement in concrete mixtures, and use of less of superplasticizers, preparation and treatment of fly ash prior to
clinker for making cements [8]. mixing into concrete, extraction and processing of limestone, and
Replacing half of portland cement would require about 1.7 bil- concrete batching, and transportation of raw materials and prod-
lion mt of alternative materials, according to USGS data [1]. ucts within the system.
High-volume fly ash (HVFA) concrete has been used successfully
for many years in numerous applications with technical and envi-
ronmental advantages as compared to conventional portland 2. Materials
cement concrete, and its use is expected to keep increasing over
time [8–10]. Yet already the global availability of fly ash is roughly For all the HVFA-L SCC trial mixtures, the common goal was to
800 million mt [11], which is less than half of the overall amount of reduce the cement content in order to lower the environmental
materials needed. Thus, other materials, such as limestone powder, footprint while maintaining the required flowability specifications.
must be increasingly brought into the mixture. The powder materials used in the mixes are ASTM Type I/II
Limestone powder (L) as calcite (or crystalline CaCO3) is a portland cement (C), Class F-fly Ash (F), and ground limestone
widely available resource that has been added to cement and con- powder (L). The Class F-fly ash is obtained from the Jim Bridger
crete in small volumes for many years, particularly in Europe. Power Plant, Wyoming, United States. Fig. 1 shows volume-based
Recent research has shown that larger amounts can be successfully particle size distributions of the powder materials obtained by
used in low water-cementitious materials ratio (w/cm) mixtures to laser light scattering. Table 1 summarizes the particle size distribu-
conserve portland cement [12,13]. The added limestone has two tion data. The measured mean particle diameters were 10.4 lm,
main functions. It acts as a limited participant in the hydration pro- 22.2 lm, and 48.1 lm, respectively, for C, F and L. The D10, D50,
cess at early ages and/or as a relatively inert calcareous filler and D90 values correspond to diameters at which the cumulative
depending on levels of calcite and replacement ratio [14]. As the sample was under 10%, 50%, and 90%, respectively. In general, C
portland cement hydrates, the ground CaCO3 reacts with various and F had much finer particle sizes when compared to L; however,
calciumaluminate hydrates to form high and low forms of carboa- F had some coarser particles of 25 lm and larger. The chemical
luminates [15]. Calcium hemicarboaluminate forms an early compositions of powder materials are presented in Table 2.
hydration product in calcite containing ordinary portland cement X-ray powder diffraction (XRD) measurements on L and F were
(OPC) blends. After about 28 days, it converts nearly completely performed using a PANalytical X’Pert PRO Materials Research Dif-
to calcium monocarboaluminate, a stable AFm phase [16]. Thermo- fractometer. Finely ground samples were loaded into metal sample
dynamic calculations and experimental observations showed that holders and placed into the diffractometer. Data were collected
monocarboaluminate formation is favored instead of monosulfoa- using a cobalt target that produces X-rays with a wavelength of
luminate [17]. The available sulfate reacts with water and calcium 1.789 Å. XRD patterns taken at ambient conditions are presented
hydroxide, crystalizing as ettringite [17,18]. Due to additional in Fig. 2 together with schematic diagrams for relevant phases.
ettringite formation, the total volume of the hydrated phase The F measurement and the L measurement are plotted with the
increases, and the overall porosity decreases [18]. same intensity scales. Although the F is mainly composed highly
This paper presents a study on the development of lower-cost,
environmental less burdensome, self-consolidating concrete mix- 100
Portland cement
tures with high-volume fly ash (HVFA) and limestone powder (L). 90 Limestone
Self-consolidating concrete (SCC) has been used increasingly in Fly Ash
the field due to several advantages when compared to conventional 80
Cumulave passing (%)

concrete, including shortened placement time, labor savings, 70

improved compaction, and better encapsulation of rebar. However,
60
typical SCC mixtures usually have an excessively high cement
content, high heat of hydration, and utilize a high dosage of high- 50
performance superplasticizers and viscosity-modifying agents 40
[19–23]. Developed herein are self-consolidating concrete mixtures
30
with low cement content (less than 250 kg/m3), and a limited
amount of a low range superplasticizer, and without any viscosity 20
modifying agents, with ternary blends of cementing material con- 10
taining portland cement, limestone powder and fly ash. The
mechanical properties and durability performance of these 0
0.1 1 10 100 1000
‘‘greener concrete mixes’’ are presented. The properties determined
Parcle diameter (μm)
include the slump flow, normal consistency, and setting time. Those
of the hardened concrete include the compressive strength, chlo- Fig. 1. Particle size distributions of portland cement (C), limestone powder (L) and
ride-ion penetration, water absorption, and gas permeability. Class F-fly ash (F).
 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72 61

Table 1 0.65–1.14% by weight of cementitious materials as a high-range

Particle size distribution of portland cement (C), limestone powder (L) and Class F-fly water reducer and 0.36–0.65% as a medium-range water reducer.
ash (F).
The maximum suggested dosage was 1.43%. The dosage used for
Mean Median Mode Standard D10 D50 D90 the trial mixes prepared in this study ranges from 1.00% to
(lm) (lm) (lm) deviation (lm) (lm) (lm) (lm) 1.43%. No viscosity-modifying admixture was used.
C 10.4 11.7 18.5 10.2 0.3 10.4 25.7
F 22.2 12.5 24.4 30.5 0.4 12.5 55.1
L 48.1 38.7 72.0 41.9 7.0 15.6 38.7 3. Experimental and analytical methods

3.1. Concrete mixture proportions

Table 2
The concrete mixture proportions are shown in Table 3. The dif-
Chemical composition of powder materials (oxides, % by weight).
ferences considered for mixture selection were aggregate fractions,
C F L proportion of powder materials, and workability of concrete. The
SiO2 20.44 62.0 0.70 w/cm ratio was held constant at 0.35. For SCC mixtures, typical
Al2O3 3.97 18.90 0.50 aggregate-to-cementing materials ratio varies between 2:1 and
Fe2O3 4.07 4.90 0.12 4:1 by weight [30]. To reduce overall cement content compared
CaO 62.90 5.98 47.40
MgO 2.42 1.99 6.80
to typical SCCs, the total aggregate-to-cementing materials ratio
Na2O 0.37 2.41 – was experimentally set at 4:1. The ratio of portland cement
K2O 0.43 1.14 – replacement (CR) to F and L was varied between 45% and 75% by
P2O5 0.16 0.26 – weight. For the ternary blends (C–F–L), L content was fixed either
TiO2 0.23 1.09 –
at 15% by weight or 25% by weight. The amount of fly ash varied
MnO 0.32 0.04 –
L.O.I. 4.69 1.30 44.48 between 20% and 60% by weight in order to match the desired total
replacement ratio for each mixture. The ratio between coarse and
fine aggregates was fixed at 50:50% by weight, and the coarse
aggregate consisted of pea gravel 30% by weight and basalt 70%
amorphous aluminosilicate glass [26], the XRD analysis reveals by weight.
that it also has several crystalline phases, including mainly quartz The workability factor that satisfied SCC requirements was tar-
(SiO2), but also hematite (Fe2O3), magnetite (Fe3O4), mullite geted and water-reducing admixture was added until the slump
(3Al2O3SiO2), and anhydrite. The XRD analysis of L showed that flow diameter was between 56 and 69 cm, and the flow time to
it was composed mainly of calcite (CaCO3) and dolomite achieve a diameter of 50 cm (T50) was between 2 and 5 s.
[CaMg(CO3)2], but also quartz, and muscovite. Similar phases were
observed in fly ash [27] and limestone [27–29] samples by other 3.2. Sample preparation
researchers.
The fine aggregate used was a quartzitic sand (with fineness A total volume of 22 liters of concrete was prepared in a pan
modulus 3.1); the coarse aggregate consisted of pea gravel (maxi- planetary-type mixer for each mixture. First, coarse aggregates
mum size 12.7 mm) and basalt (maximum size 19.0 mm). and a small amount of water were mixed for 30 s. Then cement,
The only chemical admixture used was a carboxylated poly- fly ash, and more water were added and mixed for 1 min. Lime-
ether-based high-range water reducer (ADVA 140) with a specific stone powder and the rest of the water were added and mixed
gravity of 1.010–1.120 in order to maintain the required workabil- for 1 min before the water reducer was added and mixed again
ity. The chosen superplasticizer meets the requirements of ASTM for 1 min. Fine aggregate was added and mixed for 3 min. During
C494 as Type A and Type F, and ASTM C1017 as Type I. Typical that time, the mixer was stopped if necessary, and the bottom of
addition proportions were provided by the manufacturer as the mixer was scraped to remove fine particles.

Fig. 2. X-ray diffraction (XRD) patterns of the limestone powder (L) and the Class F-fly ash (F).
62 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72

Table 3
Concrete mixture proportionsa.

(w/cm = 0.35) Proportions (by weight) kg/m3 of concrete

C–F–L C F L Aggregates SP (%) CM C CR
Fine Coarse
Control mixtures 100–0–0 1.00 – – 2.00 2.00 1.43 461 461 0
85–0–15 0.85 – 0.15 2.00 2.00 1.43 458 389 69
75–0–25 0.75 – 0.25 2.00 2.00 1.32 456 342 114
Binary HVFA blends 70–30–0 0.70 0.30 0.00 2.00 2.00 1.39 453 317 136
50–50–0 0.50 0.50 0.00 2.00 2.00 1.14 449 224 225
Ternary HVFA-L blends 55–30–15 0.55 0.30 0.15 2.00 2.00 1.14 451 248 203
45–40–15 0.45 0.40 0.15 2.00 2.00 1.03 448 202 246
35–50–15 0.35 0.50 0.15 2.00 2.00 1.00 446 156 290
25–60–15 0.25 0.60 0.15 2.00 2.00 1.00 444 111 333
55–20–25 0.55 0.2 0.25 2.00 2.00 1.34 451 248 203
45–30–25 0.45 0.3 0.25 2.00 2.00 1.14 449 202 247
35–40–25 0.35 0.4 0.25 2.00 2.00 1.14 447 156 291
25–50–25 0.25 0.5 0.25 2.00 2.00 1.14 445 111 334
a
Keys: w/cm = water/cementitious materials ratio; C = ordinary portland cement; F = Class F-fly ash; L = limestone powder; SP = superplasticizer; CM = cementitious
materials; CR = cement replacement.

The slump flow test was performed on the prepared mixture were checked qualitatively by observing aggregate density and
(Section 3.3.2). When the concrete mixture performed satisfacto- excess liquid near the slump flow perimeter.
rily with regard to slump flow diameter and T50, it was returned
to the mixer and mixed for an additional minute before casting. 3.3.3. Compressive strength test
If the slump flow time was too low or too high, the concrete was Compressive strength of concrete was determined at 7, 28, and
returned to the mixer, mixed for an additional minute and the 91 days, and one year on 75  150 mm cylinders according to
water reducing admixture was adjusted until workability looked ASTM C39 [38]. The results reported present the mean values of
to be sufficient. Then the slump flow test was performed once three samples for each mixture and curing period. Rubber pads
more. If the concrete was then satisfactory in terms of consistency were used to cap all 7-day samples and 14-day samples (ASTM
and workability, it was remixed for an additional minute before C1231 [39]), while all others were capped with sulfur mortar
casting. Otherwise, it was discarded and a new trial mixture was (ASTM C617 [40]). The compressive strength of cylinders was
prepared with a proper amount of water-reducer admixture. measured under a stress rate control machine until significant soft-
Once the concrete mixture was ready, it was cast into eighteen ening was observed. The maximum load value was taken as the
75  150 mm cylinders and three 100  200 mm cylinders in two compressive strength. To keep the coefficient of variation at less
lifts without mechanical vibration. Light shaking was allowed as than 10% for each mixture and curing period, the outliers were
the only method of consolidation. Cylinders were immediately determined and removed from the data. The cylinder size was
covered with plastic wrap and left undisturbed for 24 h at the preferred for convenience and economy, yet testing 75  150 mm
ambient laboratory conditions. The cylinders were then demolded concrete cylinders with 19.0 mm maximum size aggregate under-
and placed in an environmental chamber at 100% relative humidity value compressive strength by 2.94%, compared to the standard
and room temperature (23 ± 2 °C) until testing in accordance with 100  150 mm concrete cylinders due to the ‘‘wall effect’’
ASTM C 192 [31]. [41,42]. For that reason, the correction factor of 102.94% was used
in compressive strength calculations.
3.3. Experimental procedures
3.3.4. Non-steady state chloride migration test
Each mixture was evaluated for normal consistency and setting To determine the resistance of one-year-old concrete samples to
time by testing the cement pastes, and slump flow, compressive chloride penetration, a non-steady state chloride migration test was
strength, chloride penetration coefficient, water absorption, and performed in accordance with Nordtest Method, NT BUILD [43]. The
gas permeability by testing concretes. The test results were used 100  200 mm cylinder samples were sawed into 50 ± 2 mm-thick
as basic indicators of workability, mechanical strength, and dura- sections and cured in the fog room until the testing date.
bility properties. Three specimens were tested for each mixture. They were pre-
conditioned by vacuuming in the saturated Ca(OH)2 solution and
3.3.1. Normal consistency and setting time then settled between a catholyte solution (10% NaCl) and an ano-
The binary and ternary cement blends and control mixtures lyte solution (0.3 N NaOH). The specimens were tested under a
were tested for normal consistency in accordance with ASTM 30 V electrical potential for 24 h. Each specimen was then split axi-
C187 [32]. The cement pastes were proportioned and mixed to nor- ally into two pieces, and a 0.1 M AgNO3 solution was sprayed on
mal consistency, and thereafter the Vicat needle penetration test the freshly split surfaces where the areas containing chloride ions
was conducted to obtain the initial and the final time of setting colored white. The chloride penetration depth was measured on
according to ASTM C191 [33]. photographic images of the specimens enlarged with image-pro-
cessing software at seven points over 70 mm distance from the
3.3.2. Slump flow test white silver chloride precipitation. From the mean penetration
To assess the properties of the fresh concrete, the slump flow depth, the non-steady state chloride migration coefficient Dnssm
test was carried out according to ASTM C1611 [34]; and the flow (2) was calculated, as described in NT BUILD 492 [43], using:
diameter and T50 time was recorded. To evaluate for SCC character- pffiffiffiffiffi
RT xd  a xd
istics, the flow diameter and T50 were checked at 550–700 mm and Dnnsm ¼  ð1Þ
zFE t
2–5 s, respectively [35–37]. In addition, segregation and bleeding
 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72 63

where the volume flow rate is determined, assumed in this test to be

atmospheric pressure (N m2).
U2
E¼
L 3.4. Analytical procedures
and
rffiffiffiffiffiffiffiffi A life-cycle approach systematically studies the environmental
 
RT 1 2cd impact of concrete, quantifies resource inputs and environmental
a¼2  erf 1
outputs (life-cycle inventory), estimates the impact of these inputs
zFE c0
and burdens on humans and nature (impact analysis), and reveals
where Dnssm is the non-steady-state migration coefficient, m2/s; z is areas with improvement potential [47,48]. LCA is particularly cru-
the absolute value of ion valence for chloride, z = 1; F is the Faraday cial for a methodical analysis of the cradle-to-gate environmental
constant, F = 9.648  104 J/(Vmol); U is the absolute value of the impact of concrete production given the high volumes of concrete
applied voltage, V; R is the gas constant, R = 8.314 J/(K mol); T is use, the growing importance of environmentally sustainable infra-
the average value of the initial and final temperatures in the anolyte structure decisions, and the fact that once concrete is put in a
solution, K; L is the thickness of the specimen; xd is the average structure certain environmental factors are locked in for many
value of the penetration depths, m; t is the test duration in seconds; years [49].
erf1 is the inverse of the error function; cd is the chloride concen-
tration at which the color of the concrete changes, and cd generally 3.4.1. LCA methodology
equals 0.07 N (N is the molar concentration divided by an equiva- The LCA approach assesses the environmental impact and
lence factor) for OPC concrete; and c0 is the chloride concentration resources (including raw materials, energy, and water) used
in the catholyte solution, about 2 N. Three specimens were tested throughout a product’s life-cycle from raw material acquisition
for each mixture, and the average result was calculated. through production, use, maintenance, recycling, and ultimate
disposal. The ISO 14040 series provide a consensus framework, ter-
3.3.5. Water absorption test minology, and some methodological choices for performing a LCA
A water absorption test was carried out on the one-year-old [47,48]. The LCA consists of four major stages, as illustrated in
concrete specimens, 100 mm in diameter and 50 mm in thickness, Fig. 3.
in accordance with the ASTM C948 [44]. The specimens were The goal and scope describe the product system, the system
immersed in 21 °C water and weighed every 24 h until obtaining boundary, and the functional unit. The functional unit, measured
SSD weight stabilization, which requires less than 0.5% weight dif- in cubic meter of ready-mixed concrete, is the basis for comparison
ference between two successive measurements. The last weight throughout the study [47,48,50]. In the GreenConcrete LCA tool,
obtained of specimens was designated as B. Then, the specimens the system boundary includes the extraction of raw materials for
were dried at 100–110 °C in oven, and weighed every 24 h. Once cement and concrete, major material production and preparation
a weight loss of less than 0.5% of previously measured weight processes (e.g., cement, aggregates, admixtures, fly ash, and lime-
was obtained, the specimen was cooled in a vacuum desiccator stone), transportation of materials, mixing of concrete materials
to room temperature. The weight of the specimen in room temper- at the batching plant, and ends at the gate of the concrete plant
ature was assigned as C. The water absorption is determined as with the final product being the concrete mixture ready to be used
follows: at the construction site.
In the life-cycle inventory (LCI), we estimate GWP (in CO2-eq) as
BC well as some criteria air pollutants for which we had data, CO, NOX,
Water absorption ð%Þ ¼  100 ð2Þ
C PM10, and SO2.
The life-cycle impact assessment (LCIA) stage estimates the
potential human and ecological impact. This stage involves classi-
3.3.6. Gas permeability test
fication, characterization, and normalization of impact categories.
The gas permeability was performed on the one-year-old oven-
We only assess concrete production’s GWP, relative to 1 kg of the
dried specimens according to the CEMBUREAU method [45] to
reference gas CO2 (in CO2-eq), calculated based on IPPC guidelines
determine the gas permeability coefficient. Nitrogen gas was
[51,52]:
applied to the specimens at five particular gas pressures in the
range from 0.5 bar (0.05 MPa) to 2.5 bar (0.25 MPa), in 0.5 bar GWP100YEAR ¼ 1  CO2 þ 21  CH4 þ 310  N2 O
(0.05 MPa) increments. Flow times of nitrogen gas were recorded
until a steady-state flow was obtained. The steady-state condition
was reached once the difference between two successive flow
times was less than 3% within 5 min. For each gas pressure stage, Life Cycle Assessment
the gas permeability coefficient, Kg, was calculated using the Framework Direct Applications
Hagen-Poiseuille relationship for laminar flow of a compressible Goal and
• Product Development and
fluid through a porous medium under steady-state conditions Scope
Improvement
Definition
[46], so that:
Interpretation

• Strategic Planning
2P0 QLl
Kg ¼ ð3Þ
AðP2  P2a Þ Inventory • Public Policy Making
Analysis

where Kg is the gas permeability coefficient (m2); Q is the volume • Marketing

flow rate of the fluid (m3 s1); A is the cross-sectional area of the
Impact • Other
specimen (m2); L is the thickness of the specimen in the direction Assessment
of flow (m); l is the dynamic viscosity of the nitrogen at test tem-
perature (N s m2); P is the inlet (applied) pressure (absolute)
(N m2); Pa is the outlet pressure assumed in this test to be equal
to atmospheric pressure (N m2); and P0 is the pressure at which Fig. 3. Stages and applications of an LCA, adapted from [47,50].
64 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72

The life-cycle interpretation is important to ‘‘identify, quantify, Tables 4 and 5 summarize the LCA inputs used in the GreenCon-
check, and evaluate information from the results of the LCI and the crete LCA tool regarding production technologies, geographic loca-
LCIA, and communicate them effectively’’ [48,53,54]. Results tions, electricity grid mix percentages, transportation distances
obtained from the GreenConcrete LCA tool are discussed in the and modes, as well as type of material choices.
following sections.
4. Results and discussion
3.4.2. Environmental assessment of concrete mixtures
This study used a concrete production LCA tool, GreenConcrete 4.1. Normal consistency and setting time
LCA [24], to assess the environmental profiles of concrete mixtures
through both direct and supply-chain inputs. For example, when a Water-to-cementitious material ratios desired for normal con-
process involves the use of electricity, the tool estimates not only sistency and initial and final times of setting of cement blends
direct electricity generation impact but also supply-chain impact are shown in Table 6 and Fig. 4. Water demand of C70–F30 and
that encompasses the construction and operation of a power plant, C50–F50 was 8% and 12% lower than that of control C100, respec-
as well as the life-cycle impact of the major resources used in the tively. The ternary C25–F60–L15 blends showed 19% less water
construction of the plant, the operation of the plant, and so on. demand than the control C85–L15. This indicates that addition of
Moreover, integration of regional variations and technological
alternatives in the material production processes within the tool
offers a wide range of applicability and flexibility for cement and Table 5
Electricity grid mix percentage by source of energy adapted from EIA [69,70] for
concrete manufacturers in the United States and worldwide.
United States and from CEA [71] for Canadaa.
The Class F-fly ash was obtained from Jim Bridger Plant located
in Wyoming, United States, and it was transported by rail. The User-Input Data California (%) Wyoming (%) Canada (%)
coarse aggregate was sourced from a quarry located in British Coal 1 91 19
Colombia, Canada. The demand for aggregates in California is quite Natural gas 55 1 6
high, and importing coarse aggregates from British Columbia by Fuel oil 0.1 0.1 3
Petroleum coke 1
barges and ships is a common approach. The coarse aggregate used Nuclear 16 13
in this study was transported to a nearby port by barge, and then Hydropower 14 2 58
carried by trucks to the casting site. The limestone and cement Biomass 3
used were purchased from local producers located 130 km and Geothermal 6
Solar 0.3 1
60 km away from the casting site respectively. Hence, in the LCA
Wind 3 5
calculations, the electricity grid mixes of Wyoming and Canada
a
were used for the fly ash and the coarse aggregate production Percentages may not add up to 100% due to rounding.
phases. The California grid mix was used for the remaining phases.

Table 4
Assumptions for the concrete mixture production LCA calculations.a

User-Input Data Assumption

Type of cement Type I
Type of SCMs Fly ash, limestone
Type of admixture Superplasticizer
Electricity grid mix for Location
Cement supplier California, US
Fine aggregates supplier California, US
Coarse aggregates supplier Canada, average
Gypsum supplier California, US
Limestone supplier California, US
Fly ash supplier Wyoming, US
Transportation details for Mode Distance (km)
Cement raw materials to cement plant Truck_Class 8b (Model 2005) 1
Gypsum to cement plant Truck_Class 8b (Model 2005) 200
Cement to concrete plant Truck_Class 8b (Model 2005) 60
Fine aggregates to concrete plant Truck_Class 8b (Model 2005) 50
Coarse aggregates to concrete plant Water_Barge and Truck_Class 8b (Model 2005) 1000 km by barge and 10 km by truck
Admixture to concrete plant Truck_Class 8b (Model 2005) 1000
Fly ash to concrete plant Rail 1000
Limestone to concrete plant Truck_Class 8b (Model 2005) 130
Technology options for Type of technology selected Distance (m)
Cement raw materials prehomogenization Dry, raw storing, preblending
Cement raw materials grinding Dry, raw grinding, ball mill
Cement raw meal blending/homogenization Dry, raw meal blending, storage
Clinker pyroprocessing Preheater/Precalciner kiln with US average kiln fuel mix [75]
Clinker cooling Reciprocating grate cooler (modern)
Cement finish milling/grinding/blending Roller press
Cement PM control technology ESP
Conveying within the cement plant Screw pump 20 m between process stations
Concrete batching plant loading/mixing Mixer loading (central mix)
Concrete batching plant PM control Fabric filter
a
For further details see GreenConcrete LCA tool (Website: http://greenconcrete.berkeley.edu/).
 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72 65

Table 6 The visual stability index (VSI) values of the mixtures were
Setting time of the mixtures. evaluated between zero and one: zero showed no evidence of
C–F–L w/cm for normal Setting time, segregation or bleeding, and one showed no evidence of segrega-
consistency minutes tion and slight bleeding as a sheen on the concrete mass, in accor-
Initial Final dance with ASTM C1611 [32]. Due the constant water content and
Control mixtures 100–0–0 0.25 136 185
variable use of water reducing agent, the impact of F and L replace-
85–0–15 0.25 134 185 ment on flowability is not clearly discernible. However, it was
75–0–25 0.25 128 185 noted that generally the necessary water-reducer content, and
Binary HVFA blends 70–30–0 0.23 189 260 T50 decreased, or ds increased with the addition of fly ash to the
50–50–0 0.22 353 420 mixtures (Table 7). Notably, the 100C control concrete had a lower
Ternary HVFA-L blends 55–30–15 0.24 267 315 flow diameter and longer flow time compared to the more work-
45–40–15 0.22 282 345 able 50C–50F concrete, even though the amount of water reducer
35–50–15 0.22 340 360 used in the 100C concrete was 25% greater.
25–60–15 0.21 278 390
55–20–25 0.24 214 270
45–30–25 0.23 245 300
4.3. Compressive strength
35–40–25 0.22 285 345
25–50–25 0.22 359 405
Table 9 presents the average compressive strength for all con-
crete mixtures, while Fig. 5 shows the results in graphic form. A
wide range of strengths are attainable depending on the specific
ratios of cement replacement and curing time specified for a given
Final Inial w/cm for normal consistency
project. Similarly, a given strength can be obtained using a variety
450 0.300
of mixtures and cement replacement ratios. For example, a
400 strength level of 30 MPa can be obtained either at 28 days with a
0.250
total cement replacement level of up to 55% (either using 15%
Seng me (min.)

350 L + 40% F or 25% L + 30% F), or at 91 days with a cement replace-

0.200
ment level as high as 75%. Higher strengths can also be easily
300
w/cm

achieved. The 45% replacement mixture with 30% F and 15% L

0.150
250 reached approximately 38 MPa at 28 days, increasing to about
0.100
47 MPa at 91 days.
200 The early age data are very similar for the 15% L and 25% L ser-
0.050
ies. At 7 days, F–L mixtures ranged between 10 MPa for 75%
150
replacement (either using 15% L + 60% F or 25% L + 50% F) and
100 0.000
about 20 MPa for 45% replacement (either using 15% L + 30% F or

Table 7
Slump flow diameter and T50.
C-F-L (% )
C–F–L ds (mm) T50 (sec.) SP (%)
Fig. 4. w/cm for normal consistency, and initial and final times of setting of the 100–0–0 584 4.5 1.43
cement blends. 85–0–15 622 4.4 1.43
75–0–25 559 4.6 1.32
70–30–0 610 4.5 1.43
F in the cement system greatly decreased the water demand in the 50–50–0 660 3.8 1.14
cement paste. In general, more L in the cement blends requires 55–30–15 622 3.6 1.14
45–40–15 572 3.2 1.03
slightly more w/cm for normal consistency. The water demand of 35–50–15 635 2.6 1.00
C45–F30–L25, for instance, was approximately 5% higher than that 25–60–15 667 2.8 1.00
of C45–F40–L15. 55–20–25 635 4.0 1.34
Fig. 4 shows that changing the mixture proportions of F and L 45–30–25 622 3.8 1.14
35–40–25 653 3.0 1.14
had a strong influence on the initial and final times of setting. Bin-
25–50–25 692 3.5 1.14
ary and ternary blends showed longer setting times compared to
the control specimens. The larger amounts of cement replacement
with F elevate the initial and final setting times in both of the bin-
Table 8
ary and ternary cement systems. Significantly, the final setting Specifications and recommended values for SCC [55].
times of C70–F30 and C50–F50 were about 1.4 and 2.3 times longer
Workability characteristic Test Recommended
than the control C100. When the F30 mixtures are compared, the
methods values
C70–F30 had much lower initial and final setting times than the
Deformability and flow rate (filling ability, Slump Hwang et al. .620–
C55–F30–L15 and C45–F30–L25. However, when compared to
unrestricted flow) flow 720 mm
C50–F50, the cement replacement with L in ternary mixtures low- EFNARC: 650–
ered the setting times. For example, the final setting time of C45– 800 mm
F40–L15 and C45–F30–L25 was almost 18% and 29% shorter than JSCE: 600–700 mm
that of C50–F50, respectively. PCI: P660 mm
Swedish concrete
association:
4.2. Fresh concrete flowability 650 mm to 750 mm
T50 EFNARC: 2–5 s
The slump flow diameter (ds) and T50 times of the concrete mix- PCI: 3–5 s
Swedish concrete
tures are presented in Table 7. According to the slump flow results,
association: 3–7 s
all mixtures met the specified SCC requirements; see Table 8 [52].
66 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72

Table 9
Average compressive strength (MPa).
(a) 80

70
C–F–L Compressive strength (MPa)

Compressive Strength (MPa)

7 days 28 days 91 days 365 days 60

100–0–0 43.9 53.9 66.3 77.1 50

85–0–15 28.0 43.2 49.7 53.0
75–0–25 22.2 33.6 38.7 42.2 40
70–30–0 35.6 51.0 64.6 71.8
50–50–0 27.9 41.6 54.8 64.8 30
100-0-0
55–30–15 21.5 37.5 47.3 56.8
20 C-F-L
45–40–15 17.1 33.9 46.2 56.5 70-30-0
(L=0 wt%)
35–50–15 12.9 26.0 39.3 49.0
25–60–15 10.3 20.6 29.4 39.5 10 50-50-0
55–20–25 19.6 36.3 42.9 54.2
45–30–25 14.9 31.5 43.3 55.3
0
0 50 100 150 200 250 300 350
35–40–25 12.4 28.9 41.8 50.4
25–50–25 10.1 20.6 32.8 38.0
Curing me (days)

(b) 80

70

Compressive Strength (MPa)

25% L + 20% F). At 28 days, these mixtures reached 21 MPa for 75% 60
replacement and approximately 37 MPa for 45% replacement.
As expected, at a fixed w/cm ratio as used herein, increased F 50
content led to relatively lower early strengths but steady strength 40
gain over time, as shown in Fig. 5, which plots r’c against time.
Note the remarkable synergistic effect between limestone pow- 30 85-0-15
der and fly ash shown in Fig. 5(b) and (c) compared to the control 20 C-F-L
55-30-15
mixtures, shown in Fig. 5(d). The addition of fly ash led to reduced (L=15 wt%) 45-40-15
35-50-15
compressive strength in many cases. But in several cases, the com- 10
25-60-15
pressive strength actually increased for the limestone series mix-
0
tures when portland cement was replaced by fly ash up to a 0 50 100 150 200 250 300 350
certain amount. For the 25% limestone series with 20% fly ash Curing me (days)
(45% total replacement ratio), at 28 days the strength was 8%
higher than the zero fly ash mixture (with only 25% replacement); (c) 80

the strength was kept increasing over time. With 30% fly ash and 70
Compressive Strength (MPa)

25% limestone powder, the strength was only slightly lower than
60
for the zero fly ash mixture (75C–25L) at 28 days, however, it
exceeded 75C–25L at 91 days and 365 days. At one year the 50
strength of 55C–30F–15L was somewhat higher than the control
40
specimen of 85C–15L. This indicates that a less environmentally
burdensome product can be obtained with more cement replace- 30 75-0-25
ment by fly ash. Investigating the influence of fine additives on 55-20-25
20 C-F-L
the viscosity of cement paste, Diamantonis et al. [56] also observed (L=25 wt%) 45-30-25
a synergistic effect in a mixture with 20% limestone powder and 10 35-40-25
20% fly ash. The authors concluded that the synergy between the 25-50-25
materials can lead to a higher packing density that in turn results 0
0 50 100 150 200 250 300 350
in a denser microstructure, a hypothesis also supported by Liu and
Curing me (days)
Yan [57].
(d) 80

70
4.4. Durability assessment
Compressive Strength (MPa)

60
4.4.1. Coefficient of chloride migration
50
Fig. 6(a) demonstrates that the mixing ratio of F and F–L as a
cement replacement had a strong effect on the chloride migration 40
coefficient of the experimental concretes. All the concretes with
30
blended cement mixtures demonstrated a higher resistance to 100-0-0
chloride migration relative to the 100C, 85C–15L and 75C–25L. 20 C-F-L
(F=0 wt%) 85-0-15
Based on standard guidelines [58], the chloride penetration resis-
tance of the C–F binary mixes and C–F–L ternary mixtures ranged 10 75-0-25
from extremely high to very high (Table 10). This result suggests 0
that hydration of F and L impedes voids and pores, leading to 0 50 100 150 200 250 300 350
pore-size reduction and smaller effective chloride diffusivity. Curing me (days)
The 50C–50F showed higher resistance to chloride penetration Fig. 5. Compressive strength development over time. (a) binary mixtures (without
than the 70C–30F. This suggests that the more the F replacement, limestone powder) (b) 15% L series, (c) 25% L series, (d) control mixtures. Mixtures
the higher chloride penetration resistance for the binary mixtures. with higher F content gain strength steadily over time.
 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72 67

Fig. 6. Results of durability tests of the experimental C–F–L concretes: (a) non-steady state chloride migration coefficient as a function of cement replacement of the concrete
mix. Zones 1, 2, 3 and 4 indicate extremely high to moderate resistance to chloride penetration (Table 10); (b) relationship between water absorption and cement
replacement; and (c) gas permeability coefficient in function of cement replacement of the concrete mixture. All ratios listed as wt% C–F–L.
68 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72

Table 10 F replacement in ternary blends, which is the more straightforward

Resistance against chloride penetration based on non-steady state migration testing measurement. Long-term drying in the gas permeability tests may
[58].
have introduced artifacts into the specimens that obscure the
Chloride diffusion D (1012 m2/s) Concrete resistance meaning of the experimental results. Thus, microcracks induced
>15 Low in the concrete during longer drying periods may result in higher
10–15 Moderate gas permeability in specimens with less OPC replacement
5–10 High [37,59–63].
2.5–5 Very high
<2.5 Extremely high
4.5. Environmental assessment

However, the 35C–50F–15L showed lower resistance to chloride Estimated GWP (in CO2-eq) and criteria air emissions associated
penetration than the 45C–40F–15L, whereas the 45C–40F–15L pre- with each concrete mixture are shown in Table 11. Calculations are
sented higher resistance to chloride penetration than 55C–30F–15L. obtained from the GreenConcrete LCA tool based on assumptions
This indicates that more than 55% replacement for the ternary mix- presented in Table 4 and Table 5.
tures lowers the resistance to chloride migration. For comparison purposes, typical ordinary portland cement
Significantly, for the same ratio of cement replacement with F (OPC) concrete having a 28-day strength of 18, 30, and 60 MPa
and L, the ternary 35C–40F–25L mixture results in lower chloride (low-strength, medium-strength, and high-strength, respectively)
migration than the 35C–50F–15L mixture. This indicates that the causes emissions of about 220, 300 and 440 kg CO2/m3,
25% L replacement in ternary blends had the added advantage of respectively.
improving chloride resistance. These estimates can be compared to typical SCC mixtures
obtained from the literature [19,20,64,65]. An example SCC, which
4.4.2. Water absorption used 30% silica fume (SF) and 10% F, would result in higher CO2
Fig. 6(b) shows that alterations in the cement replacement emissions than typical concrete mixtures. Because of the high
mixtures have a strong influence on the water absorption of the volumes of SCMs in the mixture, it has similar estimated CO2-eq
experimental concretes. The 70C–30F and 50C–50F binary mixtures emissions (299 kg CO2/m3) compared to medium strength OPC
displayed lower water absorption compared to the control mix- concrete. This is still higher than the F-L SCCs studied herein.
tures and ternary mixtures due to possibly lower porosity. For Fig. 7 shows strength versus GWP trends for the mixtures and
the control specimens and ternary blends, higher L replacement compares average compressive strength (MPa) of the concrete
resulted in more water absorption, with the exception of the mixtures over time (days). The red line shows the calculated total
55C–20F–25L which had lower absorption than the 55C–30F–15L. GWP for concrete production (kg CO2-eq/m3 of concrete) and the
In general, increasing the ratio of F replacement resulted in the blue line shows the contribution to the GWP of the portland
higher water absorption in the binary and ternary mixtures. cement used in the mixtures. GWP increases with the quantity of
portland cement used in concrete mix. Fig. 8 shows the calculated
4.4.3. Gas permeability intensity of CO2-eq emissions per unit volume of concrete per 365-
Fig. 6(c) indicates that F and F–L mixtures had a strong day compressive strength versus 365-day compressive strength for
influence on the coefficient of gas permeability in the experimental the concrete mixtures used in the study. The CO2-eq intensity is
concretes. Greater amounts of cement replacement with F lower considered a good measure of the impact of concrete use [66]
the gas permeability in both the binary and ternary cement mix- because this indicator allows for the consideration of both perfor-
tures, with the exception of the 45C–40F–15L and the 35C–40F– mance (e.g., compressive strength) and contribution of concrete
25L mixtures. Adding L acts to increase the gas permeability with mixtures to GWP per unit volume and strength. For a given
the exception of the 45C–30F–25L that has 10% lower gas perme- strength, lower CO2-eq intensities are achieved by portland cement
ability than the 45C–40F–15L; note that gas permeability is being replaced with SCMs as observed in Fig. 8. For example, at
affected by drying procedure [59–62]. Although there is a correla- about 53 MPa strength, a 85% OPC – 15% L concrete mixture pro-
tion between resistance to chloride penetration and gas permeabil- duces 9.2 kg CO2-eqm3/MPa, while a 55% OPC – 20% F – 25% L
ity of concrete subjected to short-term air or oven drying, this mixture has a lower intensity, 6.1 kg CO2-eqm3/MPa. Therefore,
correlation weakens with longer drying periods [63]. This is dem- for a given performance measure, with improvements in mixture
onstrated for the specimens tested here with the CEMBUREAU design as well as selection of materials, it is possible to reduce
method: even though increasing F replacement of C in the binary the CO2-eq intensity of concrete mixtures.
and ternary mixtures generally decreased gas permeability, it As shown in Table 11, similarly to GWP, criteria air emissions
increased the chloride penetration coefficient for the 50% and more appear to increase with increase in portland cement use, mostly

Table 11
Material GWP, and criteria air pollutants.

C–F–L CO2-eq (kg/m3) CO (kg/m3) NOX (kg/m3) PMtotal (kg/m3) SO2 (kg/m3)
100–0–0 5.69E+02 1.35E01 2.99E+00 3.64E01 1.35E+00
85–0–15 4.87E+02 1.26E01 2.61E+00 3.09E01 1.15E+00
75–0–25 4.34E+02 1.20E01 2.36E+00 2.73E01 1.03E+00
70–30–0 4.12E+02 1.49E01 2.27E+00 2.62E01 9.74E01
50–50–0 3.11E+02 1.57E01 1.82E+00 1.97E01 7.31E01
55–30–15 3.33E+02 1.39E01 1.91E+00 2.09E01 7.84E01
45–40–15 2.82E+02 1.43E01 1.68E+00 1.77E01 6.63E01
35–50–15 2.32E+02 1.47E01 1.45E+00 1.44E01 5.44E01
25–60–15 1.83E+02 1.52E01 1.23E+00 1.12E01 4.27E01
55–20–25 3.32E+02 1.29E01 1.90E+00 2.07E01 7.83E01
45–30–25 2.81E+02 1.33E01 1.67E+00 1.74E01 6.61E01
35–40–25 2.31E+02 1.38E01 1.44E+00 1.42E01 5.42E01
25–50–25 1.82E+02 1.43E01 1.22E+00 1.10E01 4.26E01
 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72 69

90 600
356 days 91 days
80 28 days 7 days
Total GWP OPC GWP 500
70

GWP (kg CO2-eq / m3 of concrete)

Compressive strength (MPa)
60 400

50
300
40

30 200

20
100
10

0 0
L 0 0 0 15 25 15 25 15 25 15 25 15 25
F 0 30 50 60 40 50 50 40 30 30 20 0 0
C 100 70 50 25 35 35 25 45 45 55 55 85 75
Concrete mix

Fig. 7. Comparison of average compressive strength (MPa) of the concrete mixtures over time (days). Red line shows the calculated total GWP for concrete production (kg
CO2-eq/m3 of concrete). Blue line shows the contribution of the portland cement used in the mixtures to the GWP. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)

90 12.0
50-50-0 306
365-d Compressive Strength (MPa) Concrete
70-30-0 409
80 CO2-eq Intensity (kg CO2-eq.m-3/MPa)
10.0 75-0-25 433
365-day Compressive Strength (MPa)

70 85-0-15 487
Concrete Mixture Label

100-0-0 569
CO 2-eq Intensity (kg.m-3 /MPA)

60 8.0 25-50-25 177

(C-F-L)

35-40-25 226
50
45-30-25 277
6.0
55-20-25 329
40
25-60-15 177

30 4.0 35-50-15 227

45-40-15 278
20 55-30-15 330
2.0 0 100 200 300 400 500 600
10
Total GWP for concrete producon (kg CO2-eq/m 3 of concrete)

0 0.0 Fig. 9. Comparison of total GWP for the studied concrete mixtures (kg CO2-eq/m3 of
55-30-15

45-40-15

35-50-15

25-60-15

55-20-25

45-30-25

35-40-25

25-50-25
85-0-15

75-0-25
100-0-0

70-30-0

50-50-0

L
F concrete).
C

Concrete Mixture Label

50-50-0 258
Cement
Fig. 8. Comparison of 365-day compressive strength (MPa) of the studied concrete
70-30-0 363
mixtures and CO2-eq intensity values.
75-0-25 392

85-0-15 446

100-0-0 528
because of fuel combustion during pyroprocessing. The CO emis-
Concrete Mixture Label

sions are the only exception, which appear higher for concrete 25-50-25 127

mixtures with higher fly ash content. This is attributable to the nat- 35-40-25 179
(C-F-L)

ural gas, with a comparably higher CO emission factor, used for 45-30-25 231

drying the fly ash. 55-20-25 284

GWP by major concrete ingredients are investigated in further
25-60-15 127
details through Figs. 9–12. In Fig. 9, total GWP for each of the con-
35-50-15 179
crete mixtures involve direct and supply-chain emissions from all
the quarrying, production, and transportation processes taking 45-40-15 231

place within the system boundary. With a total of about 570 kg 55-30-15 284

of CO2-eq, the concrete mixture with 100 wt% portland cement 0 100 200 300 400 500
(which is responsible for about 93% of the total) has the largest Total GWP during concrete producon due to cement
producon (kg/m 3 of concrete)
of GWP (see Figs. 9 and 10). With the decreasing amount of port-
land cement and increasing amount of SCMs, e.g. for the 60F–15L Fig. 10. Comparison of GWP associated with cement production only (kg CO2-eq/
mixture, GWP from portland cement can be as low as 69% of the m3 of concrete).
70 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72

50-50-0 4.3 2.5 3.9 1.5 31.2 aggregate to cementing powders, and a low dose of
70-30-0 4.3 2.5 5.0 1.5 29.7 superplasticisers.
75-0-25 4.3 2.6 4.6 1.5 28.0 2. A wide range of early and long-term strengths were attainable
85-0-15 depending on the selected mixture proportions. In some cases,
Concrete Mixture Label

4.4 2.6 5.0 1.5 27.8

100-0-0 4.4 2.6 5.1 1.5 27.3 strength gain can be achieved by increasing fly ash content
25-50-25 4.2 2.5 3.9 1.5 31.9 when limestone powder is also present. For example, it was
(C-F-L)

35-40-25 4.3 2.5 3.9 1.5 31.1 observed that by 28 days the mixture with 20% fly ash exceeded
45-30-25 4.3 2.5 3.9 1.5 30.3 the strength of the control mixture containing only 25% lime-
55-20-25 4.3 2.5 4.6 1.5 29.6 stone powder. The mixtures with 30% or 40% fly ash and 25%
25-60-15 4.2 2.5 3.4 1.5 32.3 limestone powder also gained strength relative to the control
35-50-15 4.3 2.5 3.4 1.5 31.5 mixture containing only 25% limestone powder, but by 91 days.
45-40-15 4.3 2.5 3.5 1.5 30.7 The 20%, 30%, and 40% fly ash and 25% limestone mixtures had
55-30-15 4.3 2.5 3.9 1.5 30.0
higher 365-day strength than the mixture containing only 25%
0 5 10 15 20 25 30 35 40 45 limestone. 365-day strength gain was also observed for 30%
Total GWP during concrete producon excluding cement,
fly ash and limestone producon (kg/m3 of concrete)
and 40% fly ash mixtures, that contain 15% limestone powder,
relative to the mixture containing only 15% limestone. Zero
Fine Aggregate Coarse Aggregate
Superplascizer Concrete Mixing and Batching limestone powder and 30% or 50% fly ash mixtures had higher
Transportaon to Concrete Plant
strength than any mixtures containing fly ash and limestone.
Fig. 11. Total GWP associated with concrete production, excluding cement, fly ash
3. Concrete mixtures with 50% or less SCMs attained 30 MPa com-
and limestone production (kg CO2-eq/m3 of concrete). pressive strength by 28 days. All but one concrete mixtures
attained 30 MPa by 91 days.
4. The binary and ternary mixtures with fly ash or fly ash and
limestone powder resulted in higher resistance to chloride
50-50-0 5.0
Fly Ash
penetration compared to 100C, 85C–15L, and 75C–25L control
70-30-0 3.0
mixtures. The binary mixtures and ternary mixtures with
75-0-25 0.5 Limestone 25% L showed lower water absorption capacity than the ref-
85-0-15 0.3 erence mixtures of 100C and 75C–25L, respectively. However,
Concrete Mixture Label

100-0-0 the greater than 45% fly ash and limestone powder addition
25-50-25 5.0 0.4 in the ternary mixtures containing 15% limestone powder
(C-F-L)

35-40-25 4.0 0.4 demonstrated higher water absorption than the 85C–15L
45-30-25 3.0 0.4 control mixture. It is worth noting that an additional benefit
55-20-25 2.0 0.4 of using large amounts of SCMs in the reduction of heat of
25-60-15 5.9 0.3 hydration [68] was not tested in this study. The reduction
35-50-15 5.0 0.3 of heat of hydration could result in higher strength and
45-40-15 4.0 0.3
durability performance for concrete with high cement con-
55-30-15 3.0 0.3
tent and concrete members with large size (i.e., foundations,
0 1 2 3 4 5 6 piers, and dams).
Total GWP during concrete producon due to supplementary 5. GHG emissions and the analyzed criteria air pollutants were
cemenous materials producon (kg/m3 of concrete) in all cases similar to or lower than emissions from typical
ordinary portland cement concrete. The only exception to this
Fig. 12. Total GWP associated with fly ash preparation and limestone production
processes (kg CO2-eq/m3 of concrete). was CO emissions, which were estimated to be slightly higher
for mixtures with higher fly ash content due to the use of
total concrete production. Transportation of materials to the con- natural gas in drying the fly ash prior to use in concrete.
crete batching plant is the second highest source of emissions, Regarding GHG emissions from other non-cementitious ingre-
changing between 4% and 18% of the total depending on the dients, their mass contribution remains almost constant for
amount of materials conveyed and transportation distance and all mixtures. The GWP from fly ash is found to be 5–10 times
mode (Fig. 11). When we further examine the sources of the major greater for the same weight of limestone powder used in the
GHG emissions from other non-cementitious ingredients, their mixture. This difference can be explained by the higher
mass contribution remains almost constant for all mixtures, about amount of fuel utilized per unit mass of fly ash handled dur-
4 kg for fine aggregates (1–2%), 4–5 kg for superplasticizers (1–2%), ing the drying process as part of treatment prior to mixing in
3 kg for coarse aggregates (1%), and 1.5 kg for concrete mixing and the concrete.
batching activities (0.3–1%), as shown in Fig. 11. Finally, Fig. 12
summarizes the GWP from limestone production and fly ash prep-
aration, which can be as high as 3% of the total for the mixture with
highest cement replacement content, e.g., 25C–60F–15L. GWP from Acknowledgement
fly ash is found to be larger by a factor of 5–10 for the same weight
of limestone powder used in the mixture. This difference can be This research is funded by the Republic of Singapore’s National
explained by the higher amount of fuel utilized per unit mass of Research Foundation through a grant to the Berkeley Education
fly ash during the drying process as part of treatment prior to mix- Alliance for Research in Singapore (BEARS) for the Singapore-
ing in the concrete [67]. Berkeley Building Efficiency and Sustainability in the Tropics (Sin-
BerBEST) Program. BEARS has been established by the University of
5. Conclusions California, Berkeley as a center for intellectual excellence in
research and education in Singapore. The Advanced Light Source
1. Without the use of any viscosity-modifying admixtures, self- is supported by the Director, Office of Science, Office of Basic
consolidating concrete mixtures can be produced with cement, Energy Sciences, of the U.S. Department of Energy under Contract
high-volume of fly ash, limestone powder, high ratio of the No. DE-AC02-05CH11231.
 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72 71

References [35] S.E.P. Group. The European guidelines for self compacting concrete; 2005.
[36] Koehler EP, Fowler DW. Inspection manual for self-consolidating concrete in
precast members. Center for Transportation Research The University of Texas
[1] USGS, Cement – Mineral Commodity Summaries. In: Cement statistics and
at Austin; 2007.
information - annual publications. U.S. Department of the Interior, U.S.
[37] Celik K, Jackson MD, Mancio M, Meral C, Emwas AH, Mehta PK, et al. High-
Geological Survey; 2011.
volume natural volcanic pozzolan and limestone powder as partial
[2] Mehta PK, Monteiro PJM. Concrete: microstructure, properties, and materials.
replacements for portland cement in self-compacting and sustainable
4th ed. USA: McGraw-Hill; 2014.
concrete. Cement Concr Compos 2014;45:136–47.
[3] World Business Council for Sustainable Development (WBCSD-CSI). Cement
[38] ASTM, C39/C39M – 10. Standard test method for compressive strength of
industry energy and CO2 performance ‘‘Getting the Numbers Right’’,
cylindrical concrete specimens. ASTM; 2010. p. 1–7.
Washington, DC; 2009.
[39] ASTM, C1231/C1231M-12. Standard practice for use of unbonded caps in
[4] International Energy Agency & World Business Council for Sustainable
determination of compressive strength of hardened concrete cylinders. ASTM;
Development (IEA-WBCSD). Cement Technology Roadmap 2009 - Carbon
2012. p. 1–5.
emissions reductions up to 2050; 2009.
[40] ASTM, C617/C617M-12. Standard practice for capping cylindrical concrete
[5] Marceau ML, Nisbet MA, VanGeem MG. Life cycle inventory of portland cement
specimens; 2012. p. 1–6.
manufacture. Portland Cement Association [PCA], Skokie, IL; 2006.
[41] Issa SA, Islam MS, Issa MA, Yousif AA, Issa MA. Specimen and aggregate size
[6] Carnegie Mellon University Green Design Institute (CMU GDI). Economic
effect on concrete compressive strength. Cem Concr Aggr 2000;22:103–15.
input–output life cycle assessment (EIO-LCA), US 2002 industry benchmark
[42] Tokyay M, Ozdemir M. Specimen shape and size effect on the compressive
model; 2014 (accessed November 8, 2014).
strength of higher strength concrete. Cem Concr Res 1997;27:1281–9.
[7] Tomkins CD. Redefining what’s possible for clean energy by 2020. Gigaton
[43] NORDTEST, NT BUILD 492: Concrete, mortar and cement-based repair
Throwdown, San Francisco; 2009.
materials: chloride migration coefficient from non-steady-state migration
[8] Mehta PK. Global concrete industry sustainability: tools for moving forward to
experiments. Nordtest, Finland; 1999. p. 1–8.
cut carbon emissions. Concr Int 2009:45–8.
[44] ASTM, C948 – 81. Standard test method for dry and wet bulk density, water
[9] Malhotra VM. High-performance high-volume fly ash concrete. Concr Int
absorption, and apparent porosity of thin sections of glass-fiber reinforced
2002;24:30–4.
concrete. ASTM; 2009. p. 1–2.
[10] Mehta PK, Manmohan D. Sustainable high-performance concrete structures.
[45] RILEM TC 116-PCD. Recommendations of TC 116-PCD: Tests for gas
Concr Int 2006;28:37–42.
permeability of concrete. A. Preconditioning of concrete test specimens for
[11] Mehta PK. Sustainable cements and concrete for the climate change era – a
the measurement of gas permeability and capillary absorption of water, B.
review. In: Zachar PC, Naik TR, Ganjian E, editors. Second international
Measurement of the gas permeability of concrete by the RILEM—CEMBUREAU
conference on sustainable construction materials and technologies. Coventry
method, C. Determination of the capillary absorption of water of hardened
University and The University of Wisconsin Milwaukee Centre for By-products
concrete. Mater Struct 1999;32:174–9.
Utilization, Italy; 2010.
[46] Kollek JJ. The determination of the permeability of concrete to oxygen by the
[12] Bentz DP, Irassar EF, Bucher BE, Weiss J. Limestone fillers conserve cement Part
Cembureau method-a recommendation. Mater Struct 1989;22:225–30.
1: An analysis based on Powers’ model. Concr Int 2009.
[47] ISO, Environmental management - Life cycle assessment: principles and
[13] Bentz DP, Irassar EF, Bucher BE, Weiss WJ. Limestone fillers conserve cement
framework. ISO 14040:2006, International Organization for Standardization,
Part 2: Durability issues and the effects of limestone fineness on mixtures.
Geneva, Switzerland; 2006.
Concr Int 2009:35–9. December.
[48] ISO, Environmental management. Life cycle assessment: requirements and
[14] Matschei T, Lothenbach B, Glasser FP. The role of calcium carbonate in cement
guidelines. ISO 14044:2006, International Organization for Standardization,
hydration. Cem Concr Res 2007;37:551–8.
Geneva, Switzerland; 2006.
[15] Poppe A-M, De Schutter G. Cement hydration in the presence of high filler
[49] Gursel AP, Masanet E, Horvath A, Stadel A. Life-cycle inventory analysis of
contents. Cem Concr Res 2005;35:2290–9.
concrete production: a critical review. Cement Concr Compos 2014;51:38–48.
[16] Ipavec A, Gabrovšek R, Vuk T, Kaučič V, Maček J, Meden A. Carboaluminate
[50] EPA, Life cycle assessment: principles and practice. U.S. Environmental
phases formation during the hydration of calcite-containing portland cement. J
Protection Agency – National Risk Management Research Laboratory, Office
Am Ceram Soc 2011;94:1238–42.
of Research and Development, Cincinnati, Ohio; 2006.
[17] Lothenbach B, Le Saout G, Gallucci E, Scrivener K. Influence of limestone on the
[51] IPCC, IPCC guidelines for national greenhouse gas inventories. In:
hydration of portland cements. Cem Concr Res 2008;38:848–60.
Intergovernmental Panel on Climate Change – National Greenhouse Gas
[18] Bonavetti V, Donza H, Rahhal V, Irassar E. Influence of initial curing on the
Inventories Programme Technical Support Unit Geneva, Switzerland; 2006.
properties of concrete containing limestone blended cement. Cem Concr Res
[52] IPCC, Climate Change 2007: The Physical Science Basis. In: Press CU, editor.
2000;30:703–8.
Contribution of Working Group I to the Fourth Assessment Report of the
[19] Ouchi M, Nakamura S, Osterson T, Hallberg S, Lwin M. Applications of self-
Intergovernmental Panel on Climate Change, World Meteorological
compacting concrete in Japan, Europe and the United States. In: International
Organization (WMO) and the United Nations Environment Programme
symposium on high performance concrete, Orlando, FL; 2003. p. 20.
(UNEP), Cambridge, United Kingdom; 2007. p. 996.
[20] Saak AW, Jennings HM, Shah SP. New methodology for designing self-
[53] Hans-Henning S. Fundamentals of corrosion. Corrosion mechanisms in theory
compacting concrete. ACI Mater J 2001;98:429–39.
and practice, 3rd ed. CRC Press; 2011. p. 1–104.
[21] Saak AW, Jennings HM, Shah SP. A generalized approach for the determination
[54] Vieira PS, Horvath A. Assessing the end-of-life impacts of buildings. Environ Sci
of yield stress by slump and slump flow. Cem Concr Res 2004;34:363–71.
Technol 2008;42:4663–9.
[22] Le Saoût G, Kocaba V, Scrivener K. Application of the Rietveld method to the
[55] Hwang SD, Khayat KH, Bonneau O. Performance-based specifications of self-
analysis of anhydrous cement. Cem Concr Res 2011;41:133–48.
consolidating concrete used in structural applications. ACI Mater J
[23] Maggenti R. From passive to active thermal control. Concr Int
2006;103:121–9.
2007;2007:24–30.
[56] Diamantonis N, Marinos I, Katsiotis MS, Sakellariou A, Papathanasiou A,
[24] Gursel AP, Horvath A. GreenConcrete LCA Tool. University of California,
Kaloidas V, et al. Investigations about the influence of fine additives on the
Berkeley, CA; 2012.
viscosity of cement paste for self-compacting concrete. Constr Build Mater
[25] Cicas G, Hendrickson C, Horvath A, Matthews HS. A regional version of a US
2010;24:1518–22.
economic input-output life-cycle assessment model. Int J Life Cycle Ass
[57] Liu S, Yan P. Effect of limestone powder on microstructure of concrete. J
2007;12:365–72.
Wuhan Univ Technol-Mat Sci Edit 2010;25:328–31.
[26] Meral C, Benmore CJ, Monteiro PJM. The study of disorder and
[58] Gjørv OE. Durability design of concrete structures in severe
nanocrystallinity in C–S–H, supplementary cementitious materials and
environments. New York: Taylor & Francis; 2009.
geopolymers using pair distribution function analysis. Cem Concr Res
[59] Hassan KE, Cabrera JG, Maliehe RS. The effect of mineral admixtures on the
2011;41:696–710.
properties of high-performance concrete. Cem Concr Compos
[27] Bilgrami SA. Serpentinite-limestone contact at Taleri Mohammad Jan, Zhob
2000;22:267–71.
Valley, West Pakistan. The American Mineralogist 1960;45:1008–19.
[60] Abbas A, Carcasses M, Ollivier JP. Gas permeability of concrete in relation to its
[28] Lumsden DN. Discrepancy between thin-section and X-ray estimates of
degree of saturation. Mater Struct 1999;32:3–8.
dolomite in limestone. J Sediment Res 1979;49:429–35.
[61] Sugiyama T, Bremner TW, Holm TA. Effect of stress on gas permeability in
[29] Durn G, Ottner F, Slovenec D. Mineralogical and geochemical indicators of the
concrete. ACI Mater J 1996;93:443–50.
polygenetic nature of terra rossa in Istria, Croatia. Geoderma 1999;91:125–50.
[62] Abbas A, Carcasses M, Ollivier JP. The importance of gas permeability in
[30] Bonen D, Shah SP. Fresh and hardened properties of self-consolidating
addition to the compressive strength of concrete. Mag Concrete Res
concrete. Prog Struct Eng Mat 2005;7:14–26.
2000;52:1–6.
[31] ASTM, C192/C192M – 07. Standard practice for making and curing concrete
[63] Sugiyama T, Bremner TW, Tsuji Y. Determination of chloride diffusion
test specimens in the laboratory, safety precautions, manual of aggregate and
coefficient and gas permeability of concrete and their relationship. Cem
concrete testing. ASTM; 2009, p. 1–8.
Concr Res 1996;26:781–90.
[32] ASTM, C187 – 11. Standard test method for amount of water required for
[64] DOE. Transportation data book. Oak Ridge National Laboratory-Vehicle
normal consistency of hydraulic cement paste. ASTM; 2011. p. 1–3.
Technologies Program Office of Energy Efficiency and Renewable Energy U.S.
[33] ASTM, C191 – 08. Standard test methods for time of setting of hydraulic
Department of Energy (USDOE), Oak Ridge, Tennessee; 2011.
cement by Vicat needle. ASTM; 2008. p. 1–8.
[65] DOE. Buildings energy data book. Chapter:1 Buildings sector. 1.1 Buildings
[34] ASTM, C1611/C1611M – 09b. Standard test method for slump flow of self-
sector energy consumption. 1.1.1 U.S. Residential and commercial buildings
consolidating concrete. ASTM; 2009. p. 1–6.
72 K. Celik et al. / Cement & Concrete Composites 56 (2015) 59–72

total primary energy consumption, U.S. Department of Energy [DOE] - Energy [69] EIA, Electricity Net Generation by State by Type of Producer by Energy Source,
Efficiency and Renewable Energy; 2011. Annual Back to 1990 (EIA-906, EIA-920, and EIA-923). U.S. Energy Information
[66] Damineli BL, Kemeid FM, Aguiar PS, John VM. Measuring the eco-efficiency of Administration Washington, DC; 2011.
cement use. Cement Concr Compos 2010;32:555–62. [70] Shi H-S, Xu B-W, Zhou X-C. Influence of mineral admixtures on compressive
[67] Chen C, Habert G, Bouzidi Y, Jullien A, Ventura A. LCA allocation procedure strength, gas permeability and carbonation of high performance concrete.
used as an incitative method for waste recycling: an application to mineral Constr Build Mater 2009;23:1980–5.
additions in concrete. Resour Conserv Recycl 2010;54:1231–40. [71] CEA, Canadian Electricity Association – Industry Data; 2012. http://
[68] Lothenbach B, Scrivener K, Hooton RD. Supplementary cementitious materials. www.electricity.ca/resources/industry-data.php, (accessed November 8,
Cem Concr Res 2011;41:1244–56. 2014).