Materials Today: Proceedings 61 (2022) 466–476

Contents lists available at ScienceDirect

Materials Today: Proceedings

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

Statistical analysis of fly ash and slag blended geopolymer concrete

Kailash Kumar Singaram, Mohd Ataullah Khan ⇑, V. Talakokula
Department of Civil Engineering, Mahindra University, Ecole Centrale School of Engineering, Hyderabad 500043, India

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

Article history: Geopolymer concrete (GPC) allows a complete replacement of cement clinker with industrial wastes,
Available online 24 December 2021 contributing to a reduction in global carbon footprint. The recent evolution of GPC cured at ambient tem-
peratures, widens its suitability and applicability to concrete-based structures. To improve the mix
Keywords: design method of such a material, in this study, a predictive model for the compressive strength of fly
Nonlinear regression ash and slag blended GPC was developed using non-linear regression analysis, as a machine learning tool,
Sustainable on a database of 156 samples, aggregated from the literature. Statistical investigations were carried out
Machine learning
on the influence of parameters, identified from existing literature while developing the model. 75% of the
Blended slag
Blended fly ash
data were randomly chosen to train the model while the remaining 25% was used to test the efficacy of
GPC the proposed model. Root Mean Squared Error (RMSE) of the predictions on the training and testing data-
sets and R-squared value of the model based on training data were used as criteria to select the model.
The proposed model shows RMSE of 6.3 MPa, and 6.7 MPa, respectively, for the training and test datasets.
The statistical investigation reveals that the compressive strength is significantly affected by the com-
bined effect of the proportion of Na2SiO3-to-NaOH ratio, activator-to-binder ratio, Silica-to-Alumina
(Si-to-Al) ratio, Lime-to-Silica (Ca-to-Si) ratio, and the curing temperature. From the prediction of the
proposed model, it was found that the compressive strength increases with the increase in Ca-to-Si ratio
and curing temperature while it decreases with the increase in the activator-to-binder ratio. Further, the
model prediction indicates that the Fly ash and Slag blended GPC can reach a compressive strength of
about 70–80 MPa, for the following range of mix proportions and conditions: the ratio of Na2SiO3-to-
NaOH (1.5–2.5), activator-to-binder ratio (0.2–0.4), Si-to-Al ratio (1.4–1.8), Ca-to-Si ratio (0.4–0.6), and
curing temperature (30 °C to 60 °C).
Copyright Ó 2022 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the F-EIR Conference 2021
on Environment Concerns and its Remediation: Materials Science

1. Introduction
The production of cement is utilized in the construction of
infrastructure which is essential to the development of society,
however, it increases the CO2 emission into the atmosphere [1],
causing environmental problems such as the greenhouse effect
along with the reduction in fossil fuels [2–5]. Hence, there is a need
for an environment-friendly alternate binding material that can
replace cement. The natural pozzolanic materials or the industrial
by-products which are rich in reactive silica and alumina can be
utilized as cementitious materials to replace the cement. In this
line of view, aluminosilicate materials such as Fly ash and Ground
Abbreviations: GPC, Geopolymer Concrete; Ca-to-Si, Calcium oxide to Silica
ratio; Si-to-Al, Silica-to-Alumina.
⇑ Corresponding author.
E-mail address: mohdataullah.khan@mahindrauniversity.edu.in (M.A. Khan).
Granulated Blast Furnace Slag (GGBS) with the help of alkaline
solutions such as NaOH and Na2SiO3 can produce three-
dimensional polymeric chains and ring structures consisting of
Polysialates-siloxy geopolymers [6]. The amorphous materials pro-
duced during geo-polymerization contribute to the hardened
geopolymer concrete (GPC) with a smooth surface and high rigid-
ity. Fly-ash is a by-product of burned coal originated from thermal
power plants, is used as a binding material which under high tem-
perature and alkaline environment gets polymerized, leading to
the formation of amorphous and crystalline compounds, required,
to attain desirable mechanical properties [7], however, for in-situ
application, providing a high amount of heat energy is not advis-
able [8]. One barrier to the widespread use of fly ash-based GPC
is the necessity for high-temperature curing [9]. To mitigate the-
heat demand, blending of slag which is rich in silica, alumina,
and calcium is adopted. In geo-polymerization, blending of slag

https://doi.org/10.1016/j.matpr.2021.12.026
2214-7853/Copyright Ó 2022 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the F-EIR Conference 2021 on Environment Concerns and its Remediation: Materials Science
K.K. Singaram, M.A. Khan and V. Talakokula
and fly ash brings out the compounds such as calcium
aluminosilicate hydrate (C-A-S-H) in addition to sodium alumi-
nosilicate hydrate compound, which is produced by using Fly ash
alone [10]. The compounds which are obtained by blending slag
and fly ash helps in early age strength gain and forms a dense
structure which could lead to enhanced mechanical properties at
ambient curing conditions [8]. Yazdi et al. [11] reported that the
compressive and flexural strengths increased up to 100 MPa and
10 MPa, respectively, when the mixture contains 50% GGBS. In
comparison to ordinary Portland Cement Concrete (OPCC), the
amorphous structure of GPC provides several remarkable features,
including resistance to acid/ sulfate attack [12,13], chloride pene-
tration [14], and fire [15], and an ability to reach compressive
strength greater than 100 MPa [16]. GPC is an attractive choice
for 3D printing applications owing to its viscous nature in the fresh
state, contributing to enhanced bonding between the printed lay-
ers [17]. The utilization of fly ash and slag blended GPC in compar-
ison to conventional concrete, results in a reduction of carbon
dioxide emissions in the atmosphere, consumption of energy, dis-
posal of waste, and cost of the construction.
The compressive strength is a parameter that indicates the stan-
dard of concrete that could be used as a structural element. Some
statistical methods, algorithms, and codes will give the empirical
equations to evaluate the relationship between mechanical proper-
ties, but the models for the relationship between the compressive
strength and the material mix parameters of GPC are scarce [18].
The proposed study utilizes a non-linear regression analysis as a
statistical technique using MATLAB R2017a [19] to obtain a predic-
tive model to estimate the compressive strength. The developed
model could help in deriving the standard mix design to facilitate
wide usage of GPC in construction.
Previous studies identified that the chemical compositions and
the physical properties of the mix, were the most influential factors
affecting the compressive strength of GPC. These factors are alka-
line liquid to binder ratio, Na2SiO3-to-NaOH ratio, additional
water-to-binder ratio, NaOH to fly ash ratio, Na2SiO3 to fly ash
ratio, sand to fly ash ratio, and workability [20-22]. On the other
hand, Ukritnukun et al. [23] reported that the curing temperature,
blast furnace slag content, and silicate modulus also have a positive
impact. Moreover, Asghar et al. [24] have given the optimal values
for molar ratios Na + K/Si ((sodium + potassium)/silica), Ca/Si (cal-
cium oxide/silica), and volume ratio (H2O/solid) vol to increase the
compressive strength of GPC. Songpiriyakij et al. [25] reported that,
at a Si-to-Al ratio of 15.9, GPC with a comparatively high compres-
sive strength value of 73 MPa was obtained. Puertas et al. [26]
investigated the variation in strength and development of fly
ash/slag paste hydration products. They reported that the com-
pressive strength of the mix with a high fly ash/slag ratio of 1.0,
activated with 10 M NaOH solution, and cured at 25 °C reached
more than 50 MPa after 28 days. A statistical analysis by Rai
et al. [27] reported that the combined effect of curing temperature,
the molarity of NaOH, and activator-to-binder ratio notably affects
early compressive strength, while the NaOH/Na2SiO3 ratio is not
statistically significant, and the target strength can be achieved fas-
ter at elevated temperatures than at ambient temperatures. Lokuge
et al. [28] proposed the mix design for low calcium fly ash-based
GPC to obtain an optimal percentage of mix proportion to get the
target compressive strength using the Multi-variate Adaptive
Regression Splines (MARS) model. Ahmed et al. [18] proposed the
mix design using linear, nonlinear, and multi-logistic regression
analysis to develop predictive models for the compressive strength
of the fly ash based-GPC and reported that the nonlinear regression
model has better performance than the other two models. Previous
studies [30–32] have proposed models for the prediction of uncon-
fined compressive strength of soft grounds, stabilized clayey soil,
and granulated blast furnace slag using techniques such as artifi-
467
Materials Today: Proceedings 61 (2022) 466–476
cial neural network, multilayer perceptron, radial basis function,
and genetic programming.
Most of the statistical analyses reported in the literature to
obtain optimal mix design were focused on low calcium fly ash-
based GPC cured at elevated temperature [27–29], and alkali-
activated slag GPC cured at ambient conditions [23]. In the pro-
posed study, a non-linear regression analysis technique was used
to develop a model to predict the compressive strength of fly ash
and slag blended GPC cured at elevated and ambient temperatures.
In the process, an experimental database was developed and the
influence of material mix proportions on the compressive strength
was presented. To the best of the authors’ knowledge, limited work
has been carried out on the statistical analysis of fly ash and slag
blended GPC to predict the compressive strength using material
mix parameters. The developed model can be very useful to opti-
mize the ratios of different mix proportions.
2. Methodology
To develop a model, in order to predict the performance of fly
ash and slag blended GPC, the following three steps were carried
out viz. 1. Development of database from Scopus and web of
science, 2. Identification of parameters affecting the performance
of GPC from literature and 3. Statistical analysis to develop a
non-linear regression model along with validation criteria. These
steps are discussed in detail, in the following sub-sections.
2.1. Geopolymer concrete database
A systematic literature review was carried out on slag and fly
ash blended GPC from the Scopus and Web of Science (WoS) and
a database was developed using the published literature [8,33–
42] and is presented in Appendix. The output parameter is the
compressive strength of GPC (y) at 28 days in MPa. The motiva-
tion to do so is that the compressive strength value of the con-
crete also provides the overall measure of the performance of
concrete in terms of durability. From the previous studies, as dis-
cussed before, a total of eight input parameters were identified to
influence the compressive strength of GPC. These input parame-
ters are listed as
x1 ¼ Na2SiO3-to-NaOH ratio :
x2 ¼ Activator-to-binder ratio :
x3 ¼ Si-to-Al ratio.
x4 ¼ Aggregate-to-binder ratio :
x5 ¼ Ca-to-Si ratio :
x6 ¼ Superplasticizer dosage :
x7 ¼ Molarity of NaOH :
x8 ¼ Curing temperature.
where, Activator-to-binder ratio (x2 ) is the ratio of the sum of
Na2SiO3 and NaOH solutions in kg/m3 to the sum of fly ash and slag
content in kg/m3. The Si-to-Al ratio (x3 ) is the ratio of total silica in
kg/m3 to the total alumina in kg/m3, in the mixture of slag and fly
ash i.e., the binder. Ca-to-Si ratio (x5 ) is the ratio of the total lime in
kg/m3 to the total silica in kg/m3, in the mixture of slag and fly ash
i.e., the binder. The ranges of input parameters such as Na2SiO3-to-
NaOH ratio (0.5–6.3), activator-to-binder ratio (0.16–6.39), Si-to-Al
ratio (1.36–2.61), Ca-to-Si ratio (0.035–1.14), curing temperature
(20–60 °C) are presented in Table 1. The overall influence of the
type of the fly ash and slag on the performance is considered in this
study by including the respective oxide composition as an input
parameter. However, the influence of the particle size distribution
and other physical properties of the binder constituents which
were not reported in the published literature were not accounted
K.K. Singaram, M.A. Khan and V. Talakokula Materials Today: Proceedings 61 (2022) 466–476

Table 1
Statistics for input and predictive parameters.

Variable Minimum Maximum Mean Median St. deviation

Na2SiO3-to-NaOH ratio 0.5 6.3 2.48 2.5 1.29
Activator-to-binder 0.16 6.39 0.45 0.45 0.10
Si-to-Al ratio 1.36 2.61 2.03 2.0259 0.23
Aggregate-to-binder ratio 3.85 5.18 4.54 4.63 0.4095
Ca-to-Si ratio 0.035 1.14 0.36 0.3107 0.21
Superplasticizer dosage [Kg/m3] 0 180 87.05 144 80.83
Molarity of NaOH [M] 4 16 9.4 8 2.94
Curing temp. oC 20 60 35.9 30 16.05
Comp. strength (MPa) 10.5 89.6 42.71 42.78 15.32

for, therefore, the coefficient of determination was expected to be
less than unity in the modelling stage. The output parameter i.e.,
the compressive strength ranges between 10.5 and 89.6 MPa
which corresponds to low to high strength concrete. In Table 1,
the mean value for all variables, median, which shows the variation
of a specific variable, and standard deviation, used to determine
the spread of datasets compiled from the literature [8,33–42], are
likewise provided.
2.2. Statistical assessment
As discussed before, the dataset comprises eight input parame-
ters with output parameter compressive strength while the param-
eters include a Na2SiO3-to-NaOH ratio, activator-to-binder, Si-to-Al
ratio, aggregate-to-binder ratio, Ca-to-Si ratio, superplasticizer
dosage, the molarity of NaOH, curing temperature in oC. Statistical
analysis was performed to investigate whether any strong correla-
tions exist between the input parameters and the compressive
strength of fly ash and slag blended GPC with the help of the cor-
rplot function in MATLAB R2017a [19]. The function corrplot gives
a measure of covariance between the individual variables such that
it ranges between 1 and 1. The corrplot presented in Fig. 1, graph-
ically showcases the frequency distribution of each of the datasets
and the Pearson linear coefficient between all the parameters
which will help in visualizing the subsets of data and the relation
between them. The diagonal entries show the histogram of each
dataset while the off-diagonal entries depict the fit between
individual parameters. A detailed explanation of correlations
between the input parameters and the compressive strength is
provided in Appendix-A.
2.3. Non-linear regression modelling
The non-linear regression model developed with the help of a
machine learning technique wherein the database was grouped
into training and testing datasets by randomly selecting the train-
ing subset while the remaining subset was allocated for testing.
75% of datasets were used to train the non-linear regression model
and determine the best weights and biases. The remaining 25% of
datasets were used for testing to confirm the predictability of the
proposed model. The training dataset was employed to build the
model in two stages using MATLAB R2017a [19]. In the first stage,
a linear regression model was developed with the help of the fitlm
function in MATLAB. The function fitlm develops a linear regression
model of the dependent variable, fit to the input predictors, of the
form given by
y ¼ a0 þ a1 Z 1 þ a2 Z 2 þ    an Z n þ  ð1Þ
where, y is the dependent variable, ai are the model coefficients,
Z 1 ; Z 2 ;    Z n are the predictors, and  is the residual error, not
explained by the model. In the second stage of the modelling, a
polynomial regression model of order two was developed. A generic
second-order polynomial regression model with multiple predictor
variables will have the following form

Fig. 1. Correlation matrix of input parameters and strength.

468
K.K. Singaram, M.A. Khan and V. Talakokula
y ¼ a0 þ a1 Z 1 þ a2 Z 2 þ    an Z n þ a11 Z 21 þ a22 Z 22 þ    ann Z 2n
þ a12 Z 1 Z 2 þ    aij Z i Z j þ 
where, Z i ; Z 2i and Z i Z j are the linear, the quadratic, and the interac-
tion terms, respectively, ai andaij are the model coefficients. In this
stage, stepwise non-linear regression is carried out using the step
function in MATLAB on the linear model of the form presented in
Eq. (1). The step function will modify the model by removing the
parameters from Eq. (1) and/or by adding quadratic or interaction
terms from Eq. (2), in a stepwise manner. The basis for selecting
or rejecting a term in a model is its significance in predicting the
response. The function step utilizes the p-value of an F-statistic to
test models with and without a potential term at each step. The null
hypothesis employed is that the term to be removed/added has a
zero coefficient. In case, sufficient evidence is gathered against the
null hypothesis, the term is added to the model. At the end of the
last step in regression analyses of the step function, a non-linear
regression model is obtained along with the statistical significance
of model coefficients in terms of standard error, t-statistics, and
p-value. To ensure statistical significance, in case the p-value of
the model coefficients is found to be greater than 0.05, the respec-
tive term is removed from the model.
The following assessment criteria were used to choose the
model: the prediction of the model should have an error of smaller
percentage between the true and predicted values, and have a
higher R-squared value closer to one. The Root Mean Squared Error
(RMSE) of the model predictions are minimized for both the train-
ing and testing dataset while choosing the model. The RMSE is
computed as
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 X 2
RMSE ¼ pffiffiffiffi i
ytrue
i  ypred i
N
where, ytrue and ypred are the true and predicted values, respectively,
i i
of the compressive strength for sample data i and N is the number
of samples observed. The coefficient of determination or R-squared
value is the measure of the amount of variability in the output
parameter explained by the input parameters which is computed
using MATLAB R2017a [19]. Note that the adjusted R-squared value
accounts for the number of terms in the regression model. The R-
squared value for the model predictions of training dataset could
be obtained from MATLAB, however, the R-squared value for the
model predictions of training and/or test dataset could be computed
as below
P  true 2
i yi  ypred
i
R2 ¼ 1  P   2
true
i yi  mean ytrue
i to-Si ratio, Si-to-Al ratio with remaining variables such as Na2SiO3-
3. Compressive strength model for fly ash and slag blended GPC
Minimizing the RMSE of the model predictions for the training
and testing datasets, and maximizing the R-squared value close to
unity, the proposed nonlinear model for the compressive strength
prediction of fly ash and slag blended GPC is obtained as:
f c ¼ b0 þ b1 x1 þ b2 x2 þ b3 x3 þ b4 x5 þ b6 x8 þ b7 x1 x3
The coefficients bi ’s and their statistical significance values are
given in Table 2. The standard error, t-statistics, and the p-value
for the respective model co-efficient are also presented. Note that
the p-value for the estimated model coefficients of all the predictor
variables is less than 0.05. The adjusted R-squared value of the
model, accounting for the number of predictors is 0.815. The RMSE
of the model predictions for the training and testing datasets are
6.3 MPa and 6.7 MPa, respectively, indicating that the model is
Materials Today: Proceedings 61 (2022) 466–476
generalized. Note that the training dataset was not used for the
development of the model. The R-squared value obtain based on
ð2Þ
the predictions of the training dataset is 0.682. From the total eight
parameters considered based on the preliminary investigations,
the significant parameters influencing the compressive strength
found after the regression are the Na2SiO3-to-NaOH ratio (x1 ), the
activator-to-binder ratio (x2 ), Si-to-Al ratio (x3 ), Ca-to-Si ratio
(x5 ), and curing temperature (x8 ). In most of the predictive mod-
elling studies [16,20,23] the molarity of NaOH as an input param-
eter has a significant effect on the compressive strength. In this
study, the dataset of NaOH molarity has not affected the compres-
sive strength due to the usage of an identified optimal value of
molarity of NaOH (about 10 M) by most of the recent studies, used
to develop database, as observed from the histogram of respective
data, presented in Fig. 1. It is a convention to present a validation
plot akin to the present study to understand the performance of
the different models based on training and test datasets (see for
instance [43–45]. The relationship between the predicted com-
pressive strength and experimentally measured compressive
strength of fly ash and slag blended GPC for training and testing
datasets are displayed in Fig. 2a and 2b, respectively. In these fig-
ures, the values being closer to the diagonal line depicts the accu-
racy of the model. The high accuracy reported in terms of RMSE
values could be graphically observed in Fig. 2a and 2b. Note that
a higher error in the predictions of the testing dataset indicates
the overfitting characteristic of a model. However, the present
model shows efficacy in the predictions of both training and test-
ing datasets (see Fig. 2a and 2b).
4. Results and discussions
ð3Þ
To understand the influence of mix proportions on the compres-
sive strength of fly ash and slag blended GPC, the proposed model,
given by Eqn. (3) is employed and contour maps are presented. The
proposed model contains five input parameters as predictors. The
contour plots depicted in Figs. 3 to 5 illustrate the dependency of
the compressive strength on the two predictors simultaneously
for chosen values of the remaining predictors. The constant values
of the predictors chosen here are the respective median values of
the dataset obtained as {x1 , x2 , x3 , x5 , x8 } = {2.48, 0.45, 2.03, 0.36,
35.9 °C}. The contour plots should have to be interpreted carefully,
as some material mix proportions are dependent on each other,
however, the influences of the material mix proportions on con-
crete strength are apparent. The influence of each parameter on
the compressive strength is discussed below.
ð4Þ In the first series, the influence of oxide constituents such as Ca-
to-NaOH ratio, activator-to-binder ratio, and curing temperature
on compressive strength was discussed through contour plots.
The contour plot, depicted in Fig. 3a, illustrates the variation in
compressive strength with the activator-to-binder ratio and the
Ca-to-Si ratio in the binder. It was observed that the compressive
strength increases from about 20 MPa to 100 MPa as the Ca-to-Si
ratio increases from 0.15 to 1. This could be attributed to the
increase in the formation of C-A-S-H gel associated with a higher
Ca-to-Si ratio that helps in developing the compressive strength
ð5Þ
[10]. Fig. 3a also showcases that for a constant Ca-to-Si ratio, the
compressive strength marginally decreases with an increase in
the activator-to-binder ratio. Fig. 3b showcases a contour plot,
illustrating the variation in compressive strength in relation to
the Na2SiO3-to-NaOH ratio and the Ca-to-Si ratio. It can be seen
that the compressive strength decreases as the Na2SiO3-to-NaOH
ratio increases. For instance, at a Ca-to-Si ratio of 0.8, the compres-
sive strength decreases from about 100 MPa to 80 MPa as the Na2-
SiO3-to-NaOH ratio increases from 0.5 to 6. This could be due to the
469
K.K. Singaram, M.A. Khan and V. Talakokula Materials Today: Proceedings 61 (2022) 466–476

Table 2
Results of nonlinear regression model for fly ash and slag blended GPC using MATLAB: coefficient values, standard error, t-statistics, and p-value.

Estimated Coefficients:
Estimate Standard error t-statistics p-value
(Intercept) 85.8 12.672 6.7708 1.63E-09
x1 76.472 9.9704 7.6699 2.77E-11
x2 47.858 9.7495 4.9088 4.45E-06
x3 64.296 5.8663 10.96 7.15E-18
x5 95.046 6.4415 14.755 4.83E-25
x8 0.273 0.052594 5.1908 1.43E-06
x1 x3 40.219 4.6784 8.5968 3.85E-13

Fig. 2. Validation of the proposed model for compressive strength prediction of fly ash and slag blended GPC for (a) training dataset and (b) testing dataset.

Fig. 3. Influence of input parameters on the compressive strength, Input parameters: (a) activator-to-binder ratio and Ca-to-Si (b) Na2 SiO3-to-NaOH ratio and Ca-to-Si ratio;
(c) Ca-to-Si ratio and curing temperature. (d) activator-to-binder ratio and Si-to-Al ratio; (e) Na2SiO3-to-NaOH ratio and Si-to-Al ratio; (f) Si-to-Al ratio and curing
temperature.

prevention of water evaporation delaying setting of GPC structure
in presence of excess Na2SiO3 leading to diminishing of compres-
sive strength, although, Na2SiO3 works as a charge balancing ion
and is essential for the development of GPC [43,46]. The contour
plot in Fig. 3c shows the variation in compressive strength, as a
function of the Ca-to-Si ratio and the curing temperature. It was
observed that, irrespective of the curing temperature, the compres-
sive strength increases from about 30 MPa to 100 MPa as the Ca-
to-Si ratio increases from about 0.15 to 1.0. The contour plot in
Fig. 3d demonstrates the variation in compressive strength in
response to the change in the activator-to-binder ratio and the
Si-to-Al ratio. It can be observed that the compressive strength
marginally increases as the activator-to-binder ratio decreases.
470
Fig. 3d also shows that compressive strength decreases from about
85 to 40 MPa as the Si-to-Al ratio increases from 1.4 to 2.6 at an
activator-to-binder ratio of about 0.2. This could be attributed to
the formation of monomers in the GPC which majorly depends
upon the ratio of Si-to-Al present in the source material. Previous
studies have indicated that the total Si-to-Al ratio should be in
the range of 1.6–2.3, for the formation of stable geopolymer gel
[47,48]. The contour plot in Fig. 3e depicts the variation in com-
pressive strength in relation to the Na2SiO3-to-NaOH ratio and
the Si-to-Al ratio. It was observed that at a lower Si-to-Al ratio of
about 1.5, the compressive strength increases with an increase in
the Na2SiO3-to-NaOH ratio. Conversely, it was observed that at a
relatively higher Si-to-Al ratio of about 2.5, the compressive
K.K. Singaram, M.A. Khan and V. Talakokula
Fig. 4. Influence of Si-to-Al ratio and Ca-to-Si ratio on the compressive strength.
strength decreases with the increase in the Na2SiO3-to-NaOH ratio.
Fig. 3e also shows that at a relatively low Na2SiO3-to-NaOH ratio of
about 1.2, the compressive strength increases continuously with an
increase in the Si-to-Al ratio. Conversely, at a relatively higher Na2-
SiO3-to-NaOH ratio greater than about 4, the compressive strength
decreases continuously with an increase in the Si-to-Al ratio. Fig. 3f
showcases a contour plot of compressive strength as a function of
the Si-to-Al ratio and the curing temperature. It can be noted that
the compressive strength increases with an increase in the curing
temperature while it decreases with an increase in the Si-to-Al
ratio. The overall observations from Fig. 3 show that the compres-
sive strength increases with the increase in the Ca-to-Si ratio and
the curing temperature while more complex non-linear variations
in compressive strength, as discussed before, were noted in rela-
tion to the Na2SiO3-to-NaOH ratio and Si-to-Al ratio. A significant
improvement in compressive strength from about 30 to 120 MPa
was noted with an increase in Ca-to-Si ratio from about 0.2 to 1
irrespective of the other parameters.
The influence of the oxide composition of the mixture of fly ash
and slag is presented in the next result. The contour plot in Fig. 4,
illustrates the dependency of the compressive strength on the two
predictors i.e., the Ca-to-Si ratio and Si-to-Al ratio simultaneously,
for median values of the remaining predictors. It can be seen that
the compressive strength increases from about 60 MPa to
140 MPa with an increase in Ca-to-Si ratio from 0.2 to 1 at a lower
Si-to-Al ratio. At the higher value of the Si-to-Al ratio, the compres-
sive strength is seen to increase from about 20 MPa to 100 MPa
with an increase in Ca-to-Si ratio from 0.2 to 1.
In the next series of results, the influence of parameters other
than the oxide constituents is studied. The contour plot in Fig. 5a
showcases the changes in compressive strength as a function of
the activator-to-binder ratio and the Na2SiO3-to-NaOH ratio. It
can be seen that the compressive strength decreases from about
70 MPa to 50 MPa as the activator-to-binder ratio increases about
the range of 0.2–0.6. It was also observed that at the activator-to-
binder ratio of 0.2, the compressive strength decreases from about
Fig. 5. Influence of input parameters on the compressive strength, Input parameters: (a) Na2SiO3-to-NaOH ratio and activator-to-binder ratio, (b) activator-to-binder ratio and
curing temperature, (c) Na2SiO3-to-NaOH ratio and curing temperature.
471
Materials Today: Proceedings 61 (2022) 466–476
70 MPa to 45 MPa as the Na2SiO3-to-NaOH ratio increases from
about 1 to 6. Fig. 5b showcases a contour plot, illustrating the vari-
ation in compressive strength in relation to the activator-to-binder
ratio and the curing temperature. It can be seen that at a relatively
low curing temperature of 23 °C, the compressive strength
decreases from about 60 MPa to 40 MPa as the activator-to-
binder ratio increases from 0.1 to 0.6. It was also noted that at
an activator-to-binder ratio of 0.2, the compressive strength
increases from about 60 MPa to 70 MPa as the curing temperature
increases from about 20 °C to 60 °C. The improvement in compres-
sive strength could occur due to an acceleration in geo-
polymerization and polycondensation processes associated with
high temperatures [49]. The contour plot in Fig. 5c illustrates the
variations in compressive strength as a function of the Na2SiO3-
to-NaOH ratio and the curing temperature. It was observed that
at a curing temperature of about 23 °C, the compressive strength
decreases from about 55 MPa to 30 MPa as the Na2SiO3-to-NaOH
ratio increases from 0.5 to 5.5. It was also noted that at the Na2-
SiO3-to-NaOH ratio of about 1, the compressive strength increases
from about 55 MPa to 65 MPa as the curing temperature increases
from 20 °C to 60 °C. The overall observations from Fig. 5 show that
the compressive strength increases with the increase in the curing
temperature and decreases with the increase in the activator-to-
binder ratio and Na2SiO3-to-NaOH ratio. From the Figs. 3, 4, and
5, it was inferred that the Fly ash and Slag blended GPC can be fab-
ricated without the need of high thermal curing with the compres-
sive strength in the range of (70 MPa to 80 MPa), provided that the
following range for material mix parameters is adopted: the ratio
of Na2SiO3-to-NaOH (1.5–2.5), activator-to-binder ratio (0.2–0.4),
Si-to-Al ratio (1.4–1.8), Ca-to-Si ratio (0.4–0.6), and curing temper-
ature (30 °C to 60 °C). The optimum values of input parameters for
the above range of compressive strength can be clearly observed,
Na2SiO3-to-NaOH ratio in Fig. 3e, the activator-to-binder ratio in
Fig. 3d and Fig. 5a, Si-to-Al ratio in Fig. 3d, Ca-to-Si ratio in
Fig. 3a and Fig. 3b, and curing temperature in Fig. 5b, and Fig. 5c.
5. Conclusions
A database has been collated for the compressive strength of fly
ash and slag blended GPC, detailing various material mix propor-
tions adopted in the literature. A statistical model was developed
for the compressive strength prediction of fly ash and slag blended
GPC based on non-linear regression analysis as a machine learning
technique, minimizing Root Mean Square Error (RMSE) for the pre-
dictions of both the training and testing datasets while ensuring
the R-squared value close to unity. The RMSE values for the predic-
tions of training and test datasets were 6.3 MPa and 6.7 MPa,
respectively. The R-squared value of the model for the predictions
of training and testing datasets are 0.827 and 0.682, respectively.
The following conclusion can be derived based on the predictions
of the proposed model:
K.K. Singaram, M.A. Khan and V. Talakokula
 The chemical composition of the binder in terms of the Lime-to-
Silica (Ca-to-Si) ratio and the Silica-to-Alumina (Si-to-Al) ratio
has a significant influence on the performance of fly ash and
slag blended GPC.
 The compressive strength was found to increase with the
increase in Ca-to-Si ratio and curing temperature and it
decreases with the increase in activator-to-binder ratio while
a non-linear dependence was seen with the Si-to-Al ratio and
Na2SiO3-to-NaOH ratio.
 A significant improvement in compressive strength greater than
about 200% was noted with an increase in the Ca-to-Si ratio
from about 0.2 to 1, irrespective of the other parameters.
 The compressive strength in the range of about 80 MPa could be
obtained by providing oxide compositions such as Ca-to-Si ratio
of 0.5 and Si-to-Al ratio of 1.4 while maintaining the median
values of other parameters, i.e., the activator-to-binder ratio of
0.45, the Na2SiO3-to-NaOH ratio of 2.5, and the curing temper-
ature of 30 °C.
The proposed model could be used to develop a mix design for
fly ash and slag blended GPC which can be cured at ambient
temperatures.
CRediT authorship contribution statement
Kailash Kumar Singaram: Validation, Formal analysis, Investi-
gation, Resources, Data curation, Writing – original draft. Mohd
Ataullah Khan: Conceptualization, Methodology, Software, Visual-
ization, Writing – review & editing. V. Talakokula: Project admin-
istration, Funding acquisition.
Declaration of Competing Interest
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
Acknowledgments
We acknowledge the research grant received for the project
titled ‘‘High-Performance Multi-functional Geopolymer Compos-
ites” from Mahindra University, Ecole Centrale School of
Engineering.
Appendix A. Correlation of input parameters to strength
As discussed before, the correlation between the input parame-
ters and strength is performed with the help of the corrplot func-
tion in MATLAB. The diagonal entries in Fig. 1, present the
histograms of the input and the output variables based on the col-
lated dataset. A histogram typically helps in visualizing the distri-
bution of data in terms of spread, skewness, and symmetry. The
correlations between the input parameters and the compressive
strength, as depicted in Fig. 1, are discussed below:
(a). Na2SiO3-to-NaOH ratio:
The ratio of Na2SiO3-to-NaOH was in the range of 0.5–6.3, with
a mean of 2.48. The median and standard deviation were 2.5 and
1.29 respectively as reported in Table 1. The relationship between
compressive strength and Na2SiO3-to-NaOH with respective his-
tograms for slag and fly ash blended GPC mixtures is shown in
Fig. 1. As can be seen from this figure that the compressive strength
is increasing with an increase in the Na2SiO3-to-NaOH ratio.
472
Materials Today: Proceedings 61 (2022) 466–476
(b). Activator-to-binder:
The Activator-to-binder ratio of the fly ash and slag blended
GPC was in the range of 0.16–6.39 with a mean value of 0.45, the
median of 0.45, the standard deviation of 0.10 as presented in
Table 1. The variation between compressive strength and
Activator-to-binder, as well as the histogram analysis, is shown
in Fig. 1. As can be seen from the figure a very poor relationship
existed between compressive strength and the Activator-to-
binder. However, confounding effects of other variables are
expected to show the significance of the activator-to-binder ratio
in the prediction of compressive strength. These confounding
effects could be observed from the significance of the activator-
to-binder ratio in the model.
(c). Si-to-Al ratio:
The Si-to-Al ratio of the fly ash and slag blended GPC was in the
range of 1.36–2.61 with a mean value of 2.03, the median of 2.02,
the standard deviation of 0.23, as presented in Table 1. The rela-
tionship between compressive strength and Si-to-Al ratio with His-
togram of fly ash and slag blended GPC mixtures is presented in
Fig. 1. As it can be seen from the figure that the compressive
strength is increasing with an increase in the Si-to-Al ratio.
(d). Aggregate-to-binder ratio:
The aggregate-to-binder ratio was in the range of 3.85–5.18.
The statistical analysis of the dataset shows that the mean value
of the aggregate-to-binder ratio content was 4.54, the median
value of 4.63, and the standard deviation was 0.4095 were shown
in Table 1. From Fig. 1, a poor relationship is observed between
compressive strength and the Aggregate-to-binder ratio.
(e). Ca-to-Si ratio:
The ratio of Ca-to-Si ratio was in the range of 0.035–1.14, with a
mean value of 0.36. The median and standard deviation were
0.3107 and 0.21, respectively, as shown in Table 1. The relationship
between compressive strength and Ca-to-Si ratio along with
respective histograms for the fly ash and slag blended GPC mixture
is shown in Fig. 1. The figure illustrates that the compressive
strength increases with the increase in the Ca-to-Si ratio.
(f). Superplasticizer dosage:
The superplasticizer dosage was varying in the range of 0–
180 kg/m3, with a mean value of 87.05, the median value of 144,
and the standard deviation of 80.83 as depicted in Table 1. The
relationship between compressive strength and superplasticizer
dosage along with their respective histograms is shown in Fig. 1.
It can be seen that the compressive strength increases with the
increase in superplasticizer dosage.
(g). Molarity of NaOH:
The Molarity of NaOH was varying in the range of 4–16 M, with
a mean value of 9.4 M, the median value of 8 M, and the standard
deviation of 2.94 M as depicted in Table 1. Variation between com-
pressive strength and molarity of NaOH along with their respective
histograms is illustrated in Fig. 1. As it can be seen that a very poor
relationship existed between compressive strength and the molar-
ity of NaOH. Although the molarity of NaOH is an important
parameter influencing the performance of GPC, however, most of
the recent studies have used an optimal value of 10 M which was
depicted in the histogram of the dataset.
K.K. Singaram, M.A. Khan and V. Talakokula Materials Today: Proceedings 61 (2022) 466–476

(h). Curing temperature: (i). Compressive strength:

The range of the curing temperature was varying from 20 to 60
°C, with a mean value of 35.9 °C and the median value of 30 °C, and
the standard deviation of 16.05 °C as depicted in Table 1. Fig. 1
illustrates that the compressive strength increases with the
increase in curing temperature.
The compressive strength of the fly ash and slag blended GPC
was in the range of 10.5–89.6 MPa, with a mean value of
42.71 MPa and the median value of 42.78 MPa, and the standard
deviation of 15.32 MPa as depicted in Table 1.

Appendix B. Mix design database for fly ash and slag blended GPC

Fly Ash Fly Ash GGBS GGBS Composition Aggregates Alkaline SP NaOH Temp. Comp. Ref.
(kg/m3) Composition (kg/m3) (% weight) (kg/m3) Activators dosage Molarity (°C) strength
(% weight) (kg/m3) (kg/m3) (MPa)
SiO2 Al2O3 CaO SiO2 Al2O3 CaO Fine Coarse NaOH Na2SiO3
252 60.11 26.53 4 108 34.06 20 32.6 774 1090.8 46.3 115.7 144 8 30 33.83 [33]
252 60.11 26.53 4 108 34.06 20 32.6 774 1090.8 51.4 128.6 144 8 30 36.19
252 60.11 26.53 4 108 34.06 20 32.6 774 1090.8 56.6 141.4 144 8 30 31.11
252 60.11 26.53 4 108 34.06 20 32.6 774 1090.8 61.7 154.3 144 8 30 25.71
294 60.11 26.53 4 126 34.06 20 32.6 810.6 966 54 135.0 168 8 30 36.69
294 60.11 26.53 4 126 34.06 20 32.6 810.6 966 60 150.0 168 8 30 38.16
294 60.11 26.53 4 126 34.06 20 32.6 810.6 966 66 165.0 168 8 30 26.17
294 60.11 26.53 4 126 34.06 20 32.6 810.6 966 72 180.0 168 8 30 25.11
315 60.11 26.53 4 135 34.06 20 32.6 760.5 972 57.9 144.6 180 8 30 33.81
315 60.11 26.53 4 135 34.06 20 32.6 760.5 972 64.3 160.7 180 8 30 38.96
315 60.11 26.53 4 135 34.06 20 32.6 760.5 972 70.7 176.8 180 8 30 25.71
315 60.11 26.53 4 135 34.06 20 32.6 760.5 972 77.1 192.9 180 8 30 20.76
216 60.11 26.53 4 144 34.06 20 32.6 774 1090.8 46.3 115.7 144 8 30 42.32
216 60.11 26.53 4 144 34.06 20 32.6 774 1090.8 51.4 128.6 144 8 30 47.92
216 60.11 26.53 4 144 34.06 20 32.6 774 1090.8 56.6 141.4 144 8 30 46.87
216 60.11 26.53 4 144 34.06 20 32.6 774 1090.8 61.7 154.3 144 8 30 43.38
252 60.11 26.53 4 168 34.06 20 32.6 810.6 966 54 135.0 168 8 30 46.08
252 60.11 26.53 4 168 34.06 20 32.6 810.6 966 60 150.0 168 8 30 51.22
252 60.11 26.53 4 168 34.06 20 32.6 810.6 966 66 165.0 168 8 30 44.67
252 60.11 26.53 4 168 34.06 20 32.6 810.6 966 72 180.0 168 8 30 46.57
270 60.11 26.53 4 180 34.06 20 32.6 760.5 972 57.9 144.6 180 8 30 45.88
270 60.11 26.53 4 180 34.06 20 32.6 760.5 972 64.3 160.7 180 8 30 50.69
270 60.11 26.53 4 180 34.06 20 32.6 760.5 972 70.7 176.8 180 8 30 41.85
270 60.11 26.53 4 180 34.06 20 32.6 760.5 972 77.1 192.9 180 8 30 45.68 [33]
180 60.11 26.53 4 180 34.06 20 32.6 774 1090.8 46.3 115.7 144 8 30 55.37
180 60.11 26.53 4 180 34.06 20 32.6 774 1090.8 51.4 128.6 144 8 30 59.79
180 60.11 26.53 4 180 34.06 20 32.6 774 1090.8 56.6 141.4 144 8 30 51.61
180 60.11 26.53 4 180 34.06 20 32.6 774 1090.8 61.7 154.3 144 8 30 46.68
210 60.11 26.53 4 210 34.06 20 32.6 810.6 966 54 135.0 168 8 30 56.86
210 60.11 26.53 4 210 34.06 20 32.6 810.6 966 60 150.0 168 8 30 60.38
210 60.11 26.53 4 210 34.06 20 32.6 810.6 966 66 165.0 168 8 30 46.48
210 60.11 26.53 4 210 34.06 20 32.6 810.6 966 72 180.0 168 8 30 47.57
225 60.11 26.53 4 225 34.06 20 32.6 760.5 972 57.9 144.6 180 8 30 48.91
225 60.11 26.53 4 225 34.06 20 32.6 760.5 972 64.3 160.7 180 8 30 58.53
225 60.11 26.53 4 225 34.06 20 32.6 760.5 972 70.7 176.8 180 8 30 48.45
225 60.11 26.53 4 225 34.06 20 32.6 760.5 972 77.1 192.9 180 8 30 47.64
252 60.11 26.53 4 108 34.06 20 32.6 774 1090.8 46.3 115.7 144 8 60 41.53
252 60.11 26.53 4 108 34.06 20 32.6 774 1090.8 51.4 128.6 144 8 60 42.56
252 60.11 26.53 4 108 34.06 20 32.6 774 1090.8 56.6 141.4 144 8 60 33.94
252 60.11 26.53 4 108 34.06 20 32.6 774 1090.8 61.7 154.3 144 8 60 28.58
294 60.11 26.53 4 126 34.06 20 32.6 810.6 966 54 135.0 168 8 60 40.41
294 60.11 26.53 4 126 34.06 20 32.6 810.6 966 60 150.0 168 8 60 43.5
294 60.11 26.53 4 126 34.06 20 32.6 810.6 966 66 165.0 168 8 60 31.24
294 60.11 26.53 4 126 34.06 20 32.6 810.6 966 72 180.0 168 8 60 26.89
315 60.11 26.53 4 135 34.06 20 32.6 760.5 972 57.9 144.6 180 8 60 38.63
315 60.11 26.53 4 135 34.06 20 32.6 760.5 972 64.3 160.7 180 8 60 44.49
315 60.11 26.53 4 135 34.06 20 32.6 760.5 972 70.7 176.8 180 8 60 35.2

(continued on next page)

473
K.K. Singaram, M.A. Khan and V. Talakokula Materials Today: Proceedings 61 (2022) 466–476

Appendix B (continued)

Fly Ash Fly Ash GGBS GGBS Composition Aggregates Alkaline SP NaOH Temp. Comp. Ref.
(kg/m3) Composition (kg/m3) (% weight) (kg/m3) Activators dosage Molarity (°C) strength
(% weight) (kg/m3) (kg/m3) (MPa)
SiO2 Al2O3 CaO SiO2 Al2O3 CaO Fine Coarse NaOH Na2SiO3
315 60.11 26.53 4 135 34.06 20 32.6 760.5 972 77.1 192.9 180 8 60 29.46
216 60.11 26.53 4 144 34.06 20 32.6 774 1090.8 46.3 115.7 144 8 60 55.57 [33]
216 60.11 26.53 4 144 34.06 20 32.6 774 1090.8 51.4 128.6 144 8 60 57.37
216 60.11 26.53 4 144 34.06 20 32.6 774 1090.8 56.6 141.4 144 8 60 53.39
216 60.11 26.53 4 144 34.06 20 32.6 774 1090.8 61.7 154.3 144 8 60 51.22
252 60.11 26.53 4 168 34.06 20 32.6 810.6 966 54 135.0 168 8 60 58.8
252 60.11 26.53 4 168 34.06 20 32.6 810.6 966 60 150.0 168 8 60 62.29
252 60.11 26.53 4 168 34.06 20 32.6 810.6 966 66 165.0 168 8 60 58.93
252 60.11 26.53 4 168 34.06 20 32.6 810.6 966 72 180.0 168 8 60 56.64
270 60.11 26.53 4 180 34.06 20 32.6 760.5 972 57.9 144.6 180 8 60 56.56
270 60.11 26.53 4 180 34.06 20 32.6 760.5 972 64.3 160.7 180 8 60 61.3
270 60.11 26.53 4 180 34.06 20 32.6 760.5 972 70.7 176.8 180 8 60 54.38
270 60.11 26.53 4 180 34.06 20 32.6 760.5 972 77.1 192.9 180 8 60 49.83
180 60.11 26.53 4 180 34.06 20 32.6 774 1090.8 46.3 115.7 144 8 60 61.96
180 60.11 26.53 4 180 34.06 20 32.6 774 1090.8 51.4 128.6 144 8 60 65.26
180 60.11 26.53 4 180 34.06 20 32.6 774 1090.8 56.6 141.4 144 8 60 53.39
180 60.11 26.53 4 180 34.06 20 32.6 774 1090.8 61.7 154.3 144 8 60 52.4
210 60.11 26.53 4 210 34.06 20 32.6 810.6 966 54 135.0 168 8 60 57.54
210 60.11 26.53 4 210 34.06 20 32.6 810.6 966 60 150.0 168 8 60 64.28
210 60.11 26.53 4 210 34.06 20 32.6 810.6 966 66 165.0 168 8 60 50.64
210 60.11 26.53 4 210 34.06 20 32.6 810.6 966 72 180.0 168 8 60 49.44
225 60.11 26.53 4 225 34.06 20 32.6 760.5 972 57.9 144.6 180 8 60 54.97
225 60.11 26.53 4 225 34.06 20 32.6 760.5 972 64.3 160.7 180 8 60 61.9
225 60.11 26.53 4 225 34.06 20 32.6 760.5 972 70.7 176.8 180 8 60 54.38
225 60.11 26.53 4 225 34.06 20 32.6 760.5 972 77.1 192.9 180 8 60 53.79
293 46 33 2.6 88 21 17 56.1 760 1005 143.3 71.7 6 20 37.40 [34]
293 46 33 2.6 88 21 17 56.1 760 1005 107.5 107.5 6 20 26.60
293 46 33 2.6 88 21 17 56.1 760 1005 143.3 71.7 4 20 29.90
293 46 33 2.6 88 21 17 56.1 760 1005 107.5 107.5 4 20 30.70
344 46 33 2.6 34 21 17 56.1 760 1005 143.3 71.7 4 20 15.50
316 46 33 2.6 63 21 17 56.1 760 1005 143.3 71.7 4 20 23.00
293 46 33 2.6 88 21 17 56.1 760 1005 143.3 71.7 4 20 30.60
253 46 33 2.6 108 21 17 56.1 760 1005 143.3 71.7 4 20 39.01
272 46 33 2.6 108 21 17 56.1 760 1005 143.3 71.7 4 20 27.80
253 46 33 2.6 126 21 17 56.1 760 1005 143.3 71.7 4 20 28.00
237 63.53 27.4 1.26 158 34.26 11.32 38.34 547 1277 52 129 7.9 8 32 28.36 [35]
237 63.53 27.4 1.26 158 34.26 11.32 38.34 547 1277 52 129 7.9 4 32 34.84
237 63.53 27.4 1.26 158 34.26 11.32 38.34 547 1277 52 129 7.9 4 32 37.24
237 63.53 27.4 1.26 158 34.26 11.32 38.34 547 1277 52 129 7.9 4 32 33.2
400 53.71 27.2 1.9 0 29.96 12.25 45.45 651 1209 45.7 114.3 14 23 25.6
360 53.71 27.2 1.9 40 29.96 12.25 45.45 651 1209 45.7 114.3 14 23 38.3
340 53.71 27.2 1.9 60 29.96 12.25 45.45 651 1209 45.7 114.3 14 23 46.6
400 53.71 27.2 1.9 0 29.96 12.25 45.45 655.9 1218.1 40 100 6 14 23 32.5
360 53.71 27.2 1.9 40 29.96 12.25 45.45 655.9 1218.1 40 100 6 14 23 33.3
200 52.6 34.1 1.32 200 33.8 14.8 38.8 716 1074 9 56 31.7 [37]
200 52.6 34.1 1.32 200 33.8 14.8 38.8 712 1068 12 74 49.3
200 52.6 34.1 1.32 200 33.8 14.8 38.8 708 1062 15 93 58.2
200 52.6 34.1 1.32 200 33.8 14.8 38.8 712 1068 19 62 44.4
200 52.6 34.1 1.32 200 33.8 14.8 38.8 708 1062 15 93 58.2
200 52.6 34.1 1.32 200 33.8 14.8 38.8 712 1068 12 74 49.3 [37]
100 52.6 34.1 1.32 300 33.8 14.8 38.8 712 1068 12 74 52.1
0 52.6 34.1 1.32 400 33.8 14.8 38.8 712 1068 12 74 63.1
200 52.6 34.1 1.32 200 33.8 14.8 38.8 716 1074 15 93 65.7
200 52.6 34.1 1.32 200 33.8 14.8 38.8 708 1062 15 93 58.2
200 52.6 34.1 1.32 200 33.8 14.8 38.8 700 1050 15 93 42.3
360 53.71 27.2 1.9 40 29.96 12.25 45.45 651 1209 45.7 114.3 14 22 40 [38]
320 53.71 27.2 1.9 80 29.96 12.25 45.45 651 1209 45.7 114.3 14 22 47
360 53.71 27.2 1.9 40 29.96 12.25 45.45 651 1209 64 96 14 22 43

474
K.K. Singaram, M.A. Khan and V. Talakokula Materials Today: Proceedings 61 (2022) 466–476

Appendix B (continued)

Fly Ash Fly Ash GGBS GGBS Composition Aggregates Alkaline SP NaOH Temp. Comp. Ref.
(kg/m3) Composition (kg/m3) (% weight) (kg/m3) Activators dosage Molarity (°C) strength
(% weight) (kg/m3) (kg/m3) (MPa)
SiO2 Al2O3 CaO SiO2 Al2O3 CaO Fine Coarse NaOH Na2SiO3
320 53.71 27.2 1.9 80 29.96 12.25 45.45 651 1209 64 96 14 22 54
400 53.71 27.2 1.9 0 29.96 12.25 45.45 658 1222 40 100 6 14 22 25
360 53.71 27.2 1.9 40 29.96 12.25 45.45 655 1216 40 100 6 14 22 27
320 53.71 27.2 1.9 80 29.96 12.25 45.45 655 1216 40 100 6 14 22 35
400 53.71 27.2 1.9 0 29.96 12.25 45.45 658 1222 56 84 6 14 22 27
360 53.71 27.2 1.9 40 29.96 12.25 45.45 655 1216 56 84 6 14 22 27
320 53.71 27.2 1.9 80 29.96 12.25 45.45 655 1216 56 84 6 14 22 45
360 53.24 26.42 3.65 40 36.77 13.56 37.6 644 1197 53 107 4 10 22 21.9 [8]
340 53.24 26.42 3.65 60 36.77 13.56 37.6 646 1200 53 107 4 10 22 28
320 53.24 26.42 3.65 80 36.77 13.56 37.6 648 1203 53 107 4 10 22 41.5
300 53.24 26.42 3.65 100 36.77 13.56 37.6 650 1207 53 107 4 10 22 46
280 53.24 26.42 3.65 120 36.77 13.56 37.6 652 1210 53 107 4 10 22 56.5
340 53.24 26.42 3.65 60 36.77 13.56 37.6 646 1200 53 107 4 12 22 35
320 53.24 26.42 3.65 80 36.77 13.56 37.6 648 1203 53 107 4 12 22 45
300 53.24 26.42 3.65 100 36.77 13.56 37.6 658 1222 53 107 4 12 22 57 [8]
340 53.24 26.42 3.65 60 36.77 13.56 37.6 651 1209 47 93 4 10 22 30
320 53.24 26.42 3.65 80 36.77 13.56 37.6 661 1227 47 93 4 10 22 40
300 53.24 26.42 3.65 100 36.77 13.56 37.6 671 1246 47 93 4 10 22 48
340 53.24 26.42 3.65 60 36.77 13.56 37.6 637 1184 64 96 4 10 22 28.5
300 53.24 26.42 3.65 100 36.77 13.56 37.6 659 1223 46 114 4 10 22 45
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 8 10.5 [39]
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 10 13
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 12 19
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 14 22
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 16 24
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 108 54 4.05 14 22.4
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 64.8 97.2 4.05 14 27
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 54 108 4.05 14 33
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 46.28 115.72 4.05 14 30
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 40.5 121.5 4.05 14 18.2
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 8 22.2
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 10 23.2
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 12 24
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 14 24.8
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 81 81 4.05 16 21.8
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 108 54 4.05 14 23.8
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 64.8 97.2 4.05 14 33.8
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 54 108 4.05 14 34.5
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 46.28 115.72 4.05 14 35
303.75 45.8 21.4 13.7 101.25 34.52 20.66 32.43 683 1269 40.5 121.5 4.05 14 31.7
400 37.73 14.42 37.34 810 990 57.1 143 8 12 25 89.6 [40]
400 37.73 14.42 37.34 810 990 57.1 143 12 12 25 89.2
400 37.73 14.42 37.34 810 990 57.1 143 16 12 25 88.7
400 37.73 14.42 37.34 810 990 57.1 143 20 12 25 86.11
400 37.73 14.42 37.34 810 990 57.1 143 24 12 25 84.1
204.5 65.6 28 1 204.5 30.61 16.24 34.48 554 1293 41 102 10 53.5 [41]
102 65.6 28 1 307 30.61 16.24 34.48 554 1293 41 102 10 55.5
0 65.6 28 1 409 30.61 16.24 34.48 554 1293 41 102 10 58.6
400 31.63 13.42 36.35 740 1110 12.9 82.5 22 34.6 [42]
400 31.63 13.42 36.35 785 1085 12.9 82.5 22 53.6
400 31.63 13.42 36.35 790 1065 12.9 82.5 22 66.7

475
K.K. Singaram, M.A. Khan and V. Talakokula Materials Today: Proceedings 61 (2022) 466–476

References ´ nez-Ramı́rez, S. Alonso, T. Vázquez, Alkali-activated fly ash/

[26] F. Puertas, S. Martı
slag cements: strength behavior and hydration products, Cem. Concr. Res. 30
(10) (2000) 1625–1632.
[1] R. Quadrelli, S. Peterson, The energy-climate challenge: recent trends in CO2
[27] B. Rai, L.B. Roy, M. Rajjak, A statistical investigation of different parameters
emissions from fuel combustion, Energy Policy 35 (11) (2007) 5938–5952.
influencing compressive strength of fly ash-induced geopolymer concrete,
[2] A.A. Aliabdo, A.E.M. Abd Elmoaty, M.A. Emam, Factors affecting the mechanical
Struct. Concr. 19 (5) (2018) 1268–1279.
properties of alkali-activated ground granulated blast furnace slag concrete,
[28] W. Lokuge, A. Wilson, C. Gunasekara, D.W. Law, S. Setunge, Design of fly ash
Constr. Build. Mater. 197 (2019) 339–355.
geopolymer concrete mix proportions using Multivariate Adaptive Regression
[3] A.M. Rashad, A comprehensive overview about the influence of different
Spline model, Constr. Build. Mater. 166 (2018) 472–481.
additives on the properties of alkali-activated slag–A guide for Civil engineers,
[29] Y. Cui, K. Gao, P. Zhang, Experimental and statistical study on mechanical
Constr. Build. Mater. 47 (2013) 29–55.
characteristics of geopolymer concrete, Materials 13 (7) (2020) 1651.
[4] V. Bhikshma, R.M. Koti, R.T. Srinivas, An experimental investigation on
[30] B.S. Narendra, P.V. Sivapullaiah, S. Suresh, S.N. Omkar, Prediction of unconfined
properties of geopolymer concrete (no cement concrete), Asian J. Civ. Eng.
compressive strength of soft grounds using computational intelligence
(Build. Hous.) (2012) 841–853.
techniques: a comparative study, Comput. Geotech. 33 (3) (2006) 196–208.
[5] N. Mahasenan, S. Smith, K. Humphreys, The cement industry and global
[31] R.A. Mozumder, A.I. Laskar, Prediction of unconfined compressive strength of
climate change: current and potential future cement industry CO2 emissions,
geopolymer stabilized clayey soil using artificial neural network, Comput.
in: Greenhouse Gas Control Technologies-6th International Conference,
Geotech. 69 (2015) 291–300.
Pergamon, 2003, January, pp. 995-1000.
[32] N.T. Sithole, T. Mashifana, Geosynthesis of building and construction materials
[6] J. Davidovits, Geopolymers and geopolymeric materials, J. Therm. Anal. 35 (2)
through alkaline activation of granulated blast furnace slag, Constr. Build.
(1989) 429–441.
Mater. 264 (2020) 120712, https://doi.org/10.1016/
[7] G. Nagalia, Y. Park, A. Abolmaali, P. Aswath, Compressive strength and
j.conbuildmat.2020.120712.
microstructural properties of fly ash-based geopolymer concrete, J. Mater.
[33] G. Mallikarjuna Rao, T.D. Gunneswara Rao, A quantitative method of approach
Civ. Eng. 28 (12) (2016) 04016144, https://doi.org/10.1061/(ASCE)MT.1943-
in designing the mix proportions of fly ash and GGBS-based geopolymer
5533.0001656.
concrete, Aust. J. Civil Eng. 16 (1) (2018) 53–63.
[8] G. Fang, W.K. Ho, W. Tu, M. Zhang, Workability and mechanical properties of
[34] N.K. Lee, H.K. Lee, Setting and mechanical properties of alkali-activated fly ash/
alkali-activated fly ash-slag concrete cured at ambient temperature, Constr.
slag concrete manufactured at room temperature, Constr. Build. Mater. 47
Build. Mater. 172 (2018) 476–487.
(2013) 1201–1209.
[9] D. Adak, S. Mandal, Strength and durability performance of fly ash–based
[35] A. Karthik, K. Sudalaimani, C.T. Vijaya Kumar, Investigation on mechanical
process-modified geopolymer concrete, J. Mater. Civ. Eng. 31 (9) (2019)
properties of fly ash-ground granulated blast furnace slag-based self-curing
04019174, https://doi.org/10.1061/(ASCE)MT.1943-5533.0002793.
bio-geopolymer concrete, Constr. Build. Mater. 149 (2017) 338–349.
[10] P. Nath, P.K. Sarker, Effect of GGBFS on setting, workability, and early strength
[36] P. Nath, P.K. Sarker, Fracture properties of GGBFS-blended fly ash geopolymer
properties of fly ash geopolymer concrete cured in ambient condition, Constr.
concrete cured in ambient temperature, Mater. Struct. 50 (1) (2017) 1–12.
Build. Mater. 66 (2014) 163–171.
[37] Y. Ding, C.-J. Shi, N. Li, Fracture properties of slag/fly ash-based geopolymer
[11] M.A. Yazdi, M. Liebscher, S. Hempel, J. Yang, V. Mechtcherine, Correlation of
concrete cured in ambient temperature, Constr. Build. Mater. 190 (2018) 787–
microstructural and mechanical properties of geopolymers produced from fly
795.
ash and slag at room temperature, Constr. Build. Mater. 191 (2018) 330–341.
[38] P.S. Deb, P. Nath, P.K. Sarker, The effects of ground granulated blast-furnace
[12] T. Bakharev, Resistance of geopolymer materials to acid attack, Cem. Concr.
slag blending with fly ash and activator content on the workability and
Res. 35 (4) (2005) 658–670.
strength properties of geopolymer concrete cured at ambient temperature,
[13] T. Bakharev, Durability of geopolymer materials in sodium and magnesium
Mater. Des. 1980–2015 (62) (2014) 32–39.
sulfate solutions, Cem. Concr. Res. 35 (6) (2005) 1233–1246.
[39] M. Verma, N. Dev, Sodium hydroxide effect on the mechanical properties of fly
[14] K. Kupwade-Patil, E.N. Allouche, Examination of chloride-induced corrosion in
ash-slag based geopolymer concrete, Struct. Concr. 22 (2021) E368–E379.
reinforced geopolymer concretes, J. Mater. Civ. Eng. 25 (10) (2013) 1465–1476.
[40] C. Jithendra, S. Elavenil, Role of superplasticizer on GGBS based geopolymer
[15] D.L.Y. Kong, J.G. Sanjayan, Effect of elevated temperatures on geopolymer
concrete under ambient curing, Mater. Today:. Proc. 18 (2019) 148–154.
paste, mortar, and concrete, Cem. Concr. Res. 40 (2) (2010) 334–339.
_ [41] J.G. Jawahar, G. Mounika, Strength properties of fly ash and GGBS based
[16] C.D. Atisß, E.B. Görür, O.K.A.N. Karahan, C. Bilim, S.E.R.H.A.N. Ilkentapar, E. Luga,
geopolymer concrete, Asian J. Civ. Eng. (BHRC) 17 (1) (2016) 127–135.
Very high strength (120 MPa) class F fly ash geopolymer mortar activated at
[42] K. Yang, C. Yang, B. Magee, S. Nanukuttan, J. Ye, Establishment of a
different NaOH amounts, heat curing temperature, and heat curing duration,
preconditioning regime for air permeability and sorptivity of alkali-activated
Constr. Build. Mater. 96 (2015) 673–678.
slag concrete, Cem. Concr. Compos. 73 (2016) 19–28.
[17] M.A. Khan, Mix suitable for concrete 3D printing: a review, Mater. Today:.
[43] F. Farooq, W. Ahmed, A. Akbar, F. Aslam, R. Alyousef, Predictive modeling for
Proc. 32 (2020) 831–837.
sustainable high-performance concrete from industrial wastes: a comparison
[18] H.U. Ahmed, A.S. Mohammed, A.A. Mohammed, R.H. Faraj, T. Xie, Systematic
and optimization of models using ensemble learners, J. Cleaner Prod. 292
multiscale models to predict the compressive strength of fly ash-based
(2021) 126032, https://doi.org/10.1016/j.jclepro.2021.126032.
geopolymer concrete at various mixture proportions and curing regimes, PLoS
[44] C. Gunasekara, P. Atzarakis, W. Lokuge, D.W. Law, S. Setunge, Novel analytical
ONE 16 (6) (2021) e0253006, https://doi.org/10.1371/journal.pone.0253006.
method for mix design and performance prediction of high calcium fly ash
[19] MathWorks Inc, MATLAB R2017a Documentation, 2017, https://
geopolymer concrete, Polymers 13 (6) (2021) 900.
www.mathworks.com/help (last accessed July 2021).
[45] S. Aneja, A. Sharma, R. Gupta, D.Y. Yoo, Bayesian regularized artificial neural
[20] A.R. Kotwal, Y.J. Kim, J. Hu, V. Sriraman, Characterization and early age physical
network model to predict strength characteristics of fly-ash and bottom-ash
properties of ambient cured geopolymer mortar based on class C fly ash, Int. J.
based geopolymer concrete, Materials 14 (7) (2021) 1729.
Concr. Struct.Mater. 9 (1) (2015) 35–43.
[46] M.S. Morsy, S.H. Alsayed, Y. Al-Salloum, T. Almusallam, Effect of sodium
[21] K. Pimraksa, P. Chindaprasirt, A. Rungchet, K. Sagoe-Crentsil, T. Sato,
silicate to sodium hydroxide ratios on strength and microstructure of fly ash
Lightweight geopolymer made of highly porous siliceous materials with
geopolymer binder, Arabian J. Sc. Eng. 39 (6) (2014) 4333–4339.
various Na2O/Al2O3 and SiO2/Al2O3 ratios, Mater. Sci. Eng., A 528 (21) (2011)
[47] D.E. Ortega-Zavala, O. Burciaga-Díaz, J.I. Escalante-García, Chemically bonded
6616–6623.
ceramic/cementitious materials of alkali activated metakaolin processed by
[22] M.N.S. Hadi, H. Zhang, S. Parkinson, Optimum mix design of geopolymer pastes
cold pressing, Constr. Build. Mater. 279 (2021) 121365, https://doi.org/
and concretes cured in ambient condition based on compressive strength,
10.1016/j.conbuildmat.2020.121365.
setting time, and workability, J. Build. Eng. 23 (2019) 301–313.
[48] J.G.S. Van Jaarsveld, J.S.J. Van Deventer, L. Lorenzen, The potential use of
[23] S. Ukritnukun, P. Koshy, A. Rawal, A. Castel, C.C. Sorrell, Predictive model of
geopolymeric materials to immobilise toxic metals: Part I. Theory and
setting times and compressive strengths for low-alkali, ambient-cured, fly ash/
applications, Minerals Eng. 10 (7) (1997) 659–669.
slag-based geopolymers, Minerals 10 (10) (2020) 920.
[49] S. Inti, M. Sharma, V. Tandon, Influence of curing on the properties of
[24] A. Gholizadeh-Vayghan, M.H. Nofallah, A. Khaloo, Techno-economic study of
geopolymer mortar made with low molarity sodium hydroxide, Transport.
alkali-activated slag concrete with a focus on strength, CO 2 emission, and
Develop. Econom. 3 (2) (2017) 11.
material cost, J. Mater. Civ. Eng. 33 (7) (2021) 04021136.
[25] S. Songpiriyakij, T. Kubprasit, C. Jaturapitakkul, P. Chindaprasirt, Compressive
strength and degree of reaction of biomass-and fly ash-based geopolymer,
Constr. Build. Mater. 24 (3) (2010) 236–240.

476