1 Introduction

The most exceedingly used man-made building material in the world is concrete [1]. One of the major drawbacks of the concrete at the early ages is its lesser strength which slows down the construction process [2]. High early strength concrete (HESC) is one of the types in high-performance concretes. It attains high early strength than the conventional concrete [3]. Mainly this type of concrete is used for fast-track paving, precast concrete elements and concrete works in colder regions where it facilitates rapid pace of construction [4].

Generally, fast-track concrete structures need sufficient strength at early ages. In some cases, it is obligatory to accelerate the concrete hardening for removal of formwork such as in precast construction and exigent repair works [5]. There are three ways to improve the early strength of concrete; the first of them is the use of special cement, this cement can reach the predetermined strength in a short time. However, due to its low production and high prices its use in restricted not to be widely used. The second way is improving the construction and maintenance of concrete, such as hot-mix concrete, steam curing treatment methods. However, this method also causes inconvenience in actual industrial production, therefore not been widely adopted. The third way is to use the early strength admixture; practice has proved that this method is the easiest and least costly method, so it has been widely adopted in the construction projects [6].

Recently, chemical admixtures have affected huge changes in cement concrete construction. The mechanism of function of accelerating admixture (non-chloride accelerator) is that it increases the rate of hydration of tricalcium silicate (C3S) and tricalcium aluminate (C3A) phases of cement, thereby providing heat evolution and early strength development [7]. It acts as a catalyst in the hydration of tricalcium silicate (C3S) and tricalcium aluminate (C3A) [8].

The past researchers have studied the properties of concrete by using different accelerators. Up to now the most effective accelerator has been Calcium chloride, but it promotes the reinforcement corrosion by depletion of passive oxide layer [9]. Due to this, researchers have developed non-chloride accelerators which include thiosulfates, nitrates, nitrites, thiocyanate and formates [10].

A review of non-corrosive and non-chloride accelerating admixtures has been conducted by Dodson. Calcium formate of formula Ca(HC00)2, is a by-product in the manufacture of polyhydric alcohol, pentaerthritol. It is a powder and has a low solubility of about 15% in water at room temperature. It is a popular non-chloride chemical that is advocated in practice. Many non-chloride accelerating admixture formulations contain formates. Calcium formate is an accelerator for the hydration of C3S; at equal concentration however, CaC12, is more effective in accelerating the hydration of C3S [11]. Angtadt found that (patented in 1969) Calcium nitrite was the second admixture to meet the requirements of non-chloride accelerator and found that it is a very efficient corrosion inhibitor for concrete-based metals. Justnes has investigated the effect of calcium nitrate on the setting properties of cement and steel corrosion. Calcium nitrate is often used in industry because of its high solubility and cost efficiency. Some studies have shown that the effect of nitrate salts depends on the quantity of belite and aluminium in cement. Triisopropanolamine and triethanolamine are the general grinding chemicals in the cement production process. They are also formulated and used as chemical additives that can improve the mechanical properties of concrete by accelerating the hydration of specific mineral cement-based compounds [12].

The ordinary portland cement which is a major constituent of conventional concrete plays a significantly important role in contributing to the strength properties. But now- a-days cement has become a major source for pollution which compels the researchers to replace cement by some alternative materials which can provide the desirable mechanical and durability properties to concrete as well as address the pollution menace such as dumping of industrial waste.

Alccofine 1203 is proprietary low calcium silicate based mineral additives. Controlled granulation process results in unique particle size distribution. Its latent hydraulic property and pozzolanic reactivity results in enhanced hydration process. Addition of alccofine 1203 improves the packing density of paste component. The production of alccofine is more environmentally friendly compared to the production of OPC, thus producing a more environmentally friendly concrete than the OPC concrete. It has been well documented that alccofine is a very good mineral admixture to be used in improving the properties of the concrete due to its positive effects on its sustainable development and the environment. In the present study, an attempt has been made to study the effects of alccofine as a constant replacement of cement on durability and mechanical properties of concrete and its scope to address environmental pollution caused by industrial by-products. A study was carried out on the effect of non-chloride hardening accelerators on the properties of alccofine concrete in hardened state. It highlights the outcomes of the mechanical properties and durability of different concrete mixes. Hence this paper presents studies of the effect of non-chloride hardening accelerator with addition to alccofine on the properties of concrete and also checks its mechanical properties and durability when high early strength is developed.

2 Research significance

The large flow of material driven by concrete global production and consumption and the lack of durability of concrete structures have created large economic, social, and environmental impacts. A new class of repair material that achieves high-early-age strength and durability against typical cracking and delamination failures is urgently needed. This research departs from the traditionally narrow emphasis on only the high or high-early-age compressive strength of concrete repair materials but establishes a detailed understanding and database of the balanced material properties of HESC. The resulting compressive strength and tensile strength characteristics of the new HESC are expected to contribute to fast and durable concrete repairs with a reduced environmental burden with the combination of alccofine and non-chloride accelerator.

3 Materials

3.1 Cement

Khyber (Brand name) 43 grade of OPC was used and it was tested as per BIS: 8112 specifications [13]. Tables 1 and 2 show the physical and chemical properties of cement.

Table 1 Physical properties of cement
Full size table
Table 2 Chemical properties of cement
Full size table

3.2 Aggregates

Natural river sand was used as a fine aggregate having specific gravity 2.68 and conforming to Zone II. Crushed rocks were used as coarse aggregate and having specific gravity 2.78, the maximum size of the aggregate was limited to 20 mm as per BIS:383-1970 specification [14].

3.3 Alccofine

Alccofine is an ultra-fine material which is significantly superior to other mineral admixtures which are available in market [15,16,17,18,19,20,21,22,23,24,25,26,27]. In this research, the replacement of cement with alccofine was 25% for all mixes. SEM image of alccofine is shown in Fig. 1. Tables 3 and 4 Show the chemical properties and physical properties of alccofine respectively.

Fig. 1
figure 1

SEM image of alccofine

Full size image
Table 3 Chemical composition of Alccofine
Full size table
Table 4 Physical properties of alccofine
Full size table

3.4 Non-chloride hardening accelerator

A number of inorganic and organic components are available in the present market as non-chloride accelerators. Non-chloride accelerating admixtures include thiosulfates, nitrates, nitrites, thiocyanate and formates [28,29,30,31,32,33,34,35,36,37,38,39,40,41,42]. In the present research, Master set AC 100 composed primarily of calcium nitrite was used as non-chloride hardening accelerator which conforms to ASTM C494: Type C. The properties of Master set AC 100 are shown in Table 5.

Table 5 Properties of non-chloride hardening accelerator
Full size table

4 Fabrication of concrete mixes

The fabrication of concrete mixes was done as per IS: 10262-2009 specifications to produce M30 grade of concrete [43]. The binder (Cement + Alccofine) to aggregate ratio was obtained as 1:1.40:2.65 for M30 grade concrete (water to binder ratio = 0.45). Alccofine was added as a constant replacement (i.e. 25%) to all concrete mixes and four mixes in which percentage of chemical admixture was altered were prepared in this research work. Concrete mixes prepared had 328.59 kg/m3 of cement, 109.53 kg/m3 of alccofine, 614.59 kg/m3 of fine aggregate and 1162.53 kg/m3 of coarse aggregate with water-binder ratio of 0.45. Concrete mix (M1) did not contain non-chloride hardening accelerator. From concrete mixes M2 to M4 non-chloride accelerator was added in different proportions as shown in Table 6 and list of experiments for evaluating the mechanical properties and durability are shown in Table 7.

Table 6 Mix details of different grades of concrete
Full size table
Table 7 Summary of tests
Full size table

5 Methodology

5.1 Compressive strength

Compressive strength was evaluated by using cube specimens of size 150 mm as per code IS: 516-1959 [44]. Total four concrete mixes were casted with varying dosage of non-chloride hardening accelerator and tested under dry conditions. For each mix, three cubes were tested at 3, 7 and 28 days using CTM of 2000KN capacity at a uniform rate of 5.2 KN/s by placing the specimen tested at the center of the testing machine. The compressive strength of concrete specimen was evaluated.

5.2 Split tensile strength

Split tensile strength tests were carried out on cylinder specimens of size 150 mm X 300 mm as per IS 5816-1999 at 7 and 28 days [45]. The cylinder specimen was placed longitudinal to the CTM surfaces and the load was applied until the failure occurred along the vertical diameter of the cylinder. The tensile strength of the cylindrical specimens was calculated.

5.3 Modulus of rupture

Modulus of rupture was evaluated on a beam specimen at 7 and 28 days on a CTM as per code IS 516-1959 specification [44]. The modulus of rupture was determined using Eq. 1 below.

$$\upsigma \, = \frac{FL}{{bd^{2} }}$$
(1)

where: σ = modulus of rupture of the concrete specimen in MPa, F = failure load (N), L = Length of the cylinder (mm), b = width of the beam specimen (mm), d = depth of the beam specimen (mm).

5.4 Water absorption test

Water absorption test was carried out on cube specimen of size 150 mm with varying dosage of non-chloride hardening accelerator. The specimens were taken out from the curing tank at the age of 7 and 28 days. These specimens were dried in an oven at 105 °C for 24 h followed by cooling to ambient temperature after which the dry weight was measured. These dried specimens were placed in water at 23 °C for 48 h [46]. Specimens were taken out from water and the surface was wiped with a dry cloth before weighing. The water absorption value of the cube specimen was determined as per Eq. 2 below.

$${\text{Water}}\;{\text{absorption }}\left( \% \right) \, = \frac{{{\text{Saturated}}\;{\text{weight}}\;{\text{of}}\;{\text{the}}\;{\text{sample}} - {\text{Dry}}\;{\text{weight}}\;{\text{of}}\;{\text{the}}\;{\text{sample }}}}{{{\text{Dry }}\;{\text{weight}}\;{\text{of}}\;{\text{the}}\;{\text{sample}}}}$$
(2)

5.5 Resistance against sulphate attack

The sulphate resistance of concrete was studied by placing the specimens in a sulphate solution. The test was conducted on cubes with and without non-chloride hardening accelerator. The cube specimens of 150 mm were casted and demoulded after 24 h followed by curing in water for 28 days [47]. The cubes were taken out from curing tank after 28 days and initial weights were taken. For sulphate attack these cubes were immeresed in sodium sulphate solution having a concentration of 5% by weight of water. The concentration of sodium sulphate solution was maintained and specimens were taken out from solution at the end of 28 days. The cube surfaces were cleaned with water and kept in normal atmosphere for getting constant weight. The difference between initial and final dry weights indicated the degree of sulphate attack and percentage of Mass loss were calculated by using Eqs. 3 and 4 below.

$${\text{Strength}}\;{\text{deterioration}}\;{\text{factor}}\left( \% \right) = \frac{{\upsigma_{1} -\upsigma_{2} }}{{\upsigma_{1} }} \times 100$$
(3)
$${\text{Mass}}\;{\text{loss}} = \frac{{W_{1} - W_{2} }}{{W_{1} }} \times 100$$
(4)

σ1 = cube compressive strength before immersion in sulphate solution (MPa), σ2 = cube compressive strength after immersion in sulphate solution (MPa), W1 = weight of cube specimen before immersion in sulphate solution (g), W2 = weight of cube specimen after immersion in sulphate solution (g).

6 Results and discussion

6.1 Compressive strength

In this study, the strength of concrete was evaluated by compression test. Generally in the construction industry, compressive strength of concrete is not only important at later stages but it is equally important at early ages. In this research, a constant replacement of cement with alccofine volume (i.e. 25%) was taken as a reference mix and the non-chloride accelerator was added with different proportions. The compressive strength of cubes tested at the age of 3, 7 and 28 days for the water-binder ratio of 0.45 is shown in Fig. 2. It has been recognized that non-chloride hardening accelerator increases the compressive strength of concrete at early ages compared to reference mix. The compressive strength of concrete at early age (3 days) for mixes M2, M3 and M4 were enhanced by 22.34%, 35.05% and 29.02% compared to reference mix M1. Similar results were also reported by Meagher et al. [5] wherein it was found that chloride free accelerators increased the hydration rate of tricalcium aluminate and tricalcium silicate phases of cement, consequently giving not only heat evolution but also strength enhancement at early ages [48]. At 7 days, ratio of strength improvement was less as compared to 3 days i.e. 5.76%, 11.89% and 6.95% compared to the reference mix at 7 days. The long term compressive strengths for all concrete mixes were slightly higher than those of reference concrete mix (i.e. Accelerators don’t affect the long term properties).

Fig. 2
figure 2

Compressive strength of alccofine concrete for different proportions of NCHA

Full size image

6.2 Split tensile strength

Early age cracking may occur because of change in volume due to thermal contraction and shrinkage. Thermal and shrinkage effects induce tensile stresses in concrete. When these stresses are greater than the tensile strength of concrete, which is somewhat low at early ages, it can develop cracks in concrete. Hence it is required to evaluate the tensile strength of concrete at which concrete may crack. The cylindrical specimens were tested to evaluate the split tensile strength of concrete at the age of 7 and 28 days for water binder ratio 0.45 as shown in Fig. 3. From the results, it is clearly observed that the concrete mixes made with non-chloride hardening accelerator showed higher tensile strength values than the reference concrete. The early age split tensile strength (7 days) for concrete mixes M2, M3 and M4 were improved by 9.54%, 23.32% and 16.21% respectively. The degree of enhancement in split tensile strength for alccofine concrete is slightly higher strengths with non-chloride accelerator at a later age [10].

Fig. 3
figure 3

Split tensile strength of alccofine concrete with different proportion of NCHA

Full size image

6.3 Modulus of rupture

The modulus of rupture results with and without non-chloride hardening accelerator for concrete mixes with water-binder ratio 0.45 at the age of 7 and 28 days are shown in Fig. 4. In this study, it was observed that reference concrete exhibited less modulus of rupture at given water-binder ratio (i.e. 0.45). The modulus of rupture of concrete at an early age for mixes M2, M3 and M4 were increased by 7.24%, 29.03% and 26.66% compared to reference alccofine concrete M1. The modulus of rupture of concrete at a later stage for mixes M2, M3 and M4 were almost same compared to reference mix M1. Brook et al. also reported that concrete with non-chloride hardening accelerator showed higher modulus of rupture compared to reference concrete at early age but the modulus of rupture of concrete with non-chloride hardening accelerator at 28th day was the same as that of reference concrete.

Fig. 4
figure 4

Modulus of rupture of alccofine concrete with different proportion of NCHA

Full size image

6.4 Water absorption test

Water absorption test is one of the most important parameter for finding the durability of concrete. The perforation of ions, water and gases depend on the porosity and microstructure of concrete. Figure 5 shows the results of water absorption test with and without non-chloride hardening accelerator. From the results, it can be seen that reference concrete has more water absorption than that of concrete with non-chloride hardening accelerator. Concrete mixes with the accelerator (i.e. M2, M3 and M4) showed a decrease in the percentage of water absorption values with respect to reference concrete (M1) due to the formation of ettringite at early ages along with alccofine fills the voids between the particles and thus, decreases the permeable voids in the cubes as shown in Fig. 5.

Fig. 5
figure 5

Percentage of water absorption of alccofine concrete with different proportion of NCHA

Full size image

6.5 Resistance to sulphate attack

Sulphate attack on the concrete is a chemical disintegration mechanism where sulphate ions attack components of cement paste. The compounds responsible for sulphate attack on concrete are water soluble salts such as alkali (potassium and sodium) and alkali earth (magnesium and calcium) sulphates that are capable of chemically reacting with concrete components. In the present study, the effect of sulphate attack was analyzed on concrete with and without non-chloride hardening accelerator [49]. The concrete samples were tested for four mixes to find out strength deterioration and mass loss factors. These are shown in Figs. 6 and 7. It was seen that concrete with non-chloride accelerator specimens were less affected than that of reference concrete. From Fig. 7 it is clearly seen that the mass loss decreased with the increment of non-chloride accelerator to concrete. Figure 8 shows the non-chloride accelerator increases the resistance of concrete against sulphate attack.

Fig. 6
figure 6

Percentage of mass loss of alccofine concrete with different proportion of NCHA

Full size image
Fig. 7
figure 7

Strength deterioration factor of alccofine concrete with different proportion of NCHA

Full size image
Fig. 8
figure 8

Response surface plot for “compressive strength” versus “%NCHA”, “curing age”

Full size image

7 Design of experiments

7.1 Response surface method

In this research Response Surface Method (RSM) was used for predicting the compressive strength of the concrete. Response surface method is a statistical and mathematical technique which can be used for refining, developing and optimizing processes in the research and industrial field [25]. It can evaluate the effects of one or more variables and their interactions on one or more response variables. It can be represented by Eq. 5, which relates the variables £ and responses y for a specified system [50].

$${\text{y }} = {\text{ f }}\left( {\pounds_{1} , \, \pounds_{2} , \, \pounds_{3} \ldots \ldots ..\pounds_{\text{k}} } \right) +\upvarepsilon$$
(5)

where \({\pounds_{1} , \, \pounds_{2} , \, \pounds_{3} \ldots \pounds_{\text{k}} }\) are the input variables, y is the response, f represents approximate response function and ɛ denotes the statistical error.In this research, the second order polynomial was used instead of the linear polynomial to evaluate response from input variables as shown in Eq. 6.

$${\text{y}} =\upbeta_{\text{o}} + \mathop \sum \limits_{i = 1}^{\text{k}}\upbeta_{\text{i}} \pounds_{\text{i}} + \mathop \sum \limits_{i = k}^{\text{k}} \beta_{\text{ii}} \;\pounds^{2}_{\text{ii}} + \mathop \sum \limits_{i = 1}^{\text{k}} \mathop \sum \limits_{j = 1,j \ne 1}^{\text{k}} \beta_{\text{ij}} \pounds_{\text{i}} \pounds_{\text{j}} +\upvarepsilon$$
(6)

where y = response, β = coefficient,  \(\pounds\) = factor and ɛ = error. βo, βi, βii and βij are the regression coefficients for intercept, linear, quadratic and interaction terms respectively, and \(\pounds\)I and \(\pounds\)j are the input variables

In the present research, two independent variables were “dosage of Non-chloride hardening accelerator” and “Curing age” and the dependent response variable was compressive strength of concrete. By using this method which employs the regression analysis, response surface and residuals plots of the variables were evaluated in addition to main effect.

In this study, MINITAB was used to obtain the values of the coefficients by virtue of regression analysis for 95% confidence levels (α = 0.05) wherein the general Eq. 8 takes the following form Eq. 7.

$${\text{Z }} = {\text{ A }} + {\text{ BX }} + {\text{ CY }} + {\text{ DX}}^{2} + {\text{ EY}}^{2} + {\text{ FXY}}$$
(7)

where Z = compressive strength of concrete, X = percentage of non-chloride accelerator, Y = curing age, The coefficient values evaluated were A = 9.878, B = 6.57, C = 1.421, D = − 1.22, E = − 0.01792 and F = − 0.1563.

Table 8 shows the predicted and experimental values of compressive strength along with residual error. The regression model obtained from response surface analysis consisted compressive strength (Z) versus the curing age (Y) and the %NCHA (X) as shown in Fig. 8. The percentage of error less than 5% was observed for the predicted response surface plot thereby showing 95% confidence level [51].

Table 8 Predicted and experimental values by using regression analysis
Full size table

7.2 Residual plots

The graph between fitted and residual values shows the closeness of predicted compressive strength to experimentally observed data. The points near to the reference line in Fig. 9 designate less error and the points far from the reference line show more error. The error observed in Fig. 9 was found to be within acceptable range of tolerance with two residual values very close to the reference line. In the range of 0–0.4 and 0.4–0.8 the observed residual values were 2 and 3 respectively. Similarly, under the reference line 2 and 3 residual values were found to be in the range 0 to − 0.4 and − 0.4 to − 0.8 respectively.

Fig. 9
figure 9

Residual plot for compressive strength

Full size image

7.3 Effect of %NCHA and curing age on compressive strength of concrete (Main effects)

The effects of “%NCHA” and “Curing age” when plotted with reference to mean compressive strength were indicated by the magnitude of slope wherein larger values of slope indicated higher effect on the parameter considered (i.e. compressive strength) and vice versa. It was observed from the slope of “%NCHA – Mean compressive strength plot” that the strength enhanced up to 0.85% NCHA followed by decrease the same when computed at 1.05%. The plot of Compressive strength—curing age was seen with a steeper slope compared to “%NCHA—Mean compressive strength plot” which clearly indicated that curing age had a greater role in the compressive strength development that increased further of the 7 days as shown in Fig. 10.

Fig. 10
figure 10

Main effects plot for compressive strength

Full size image

8 Cross-validation or performance evalution

Cross-validation is a technique to determine the perfection of the performance of the predictive model (i.e. in the residual sum of squares and the coefficient of determination) and it can be used when the amount of data is limited and the response Y is a quadratic or linear regression model with a function P TX (X) = [X1X2……X 21 X 22 ….] [52].

8.1 Coefficient of determination (COD) (R2)

The determination coefficient is a statistic that indicates the rate of change in the quantity of the response variable that is predictable from the explanatory variable as explained in the regression model [53] using Eqs. 8 and 9 below.

$${\text{R}}^{2} = \, 1 - \frac{{\varSigma_{{{\text{k}} = 1}}^{n} \left( {{\text{Y}}_{\text{k}} - \hat{Y}_{\text{k}} } \right)^{2} }}{{\mathop \sum \nolimits_{{{\text{k}} = 1}}^{n} \left( {{\text{Y}}_{\text{k}} - \hat{Y}} \right)^{2} }}$$
(8)
$$\hat{Y} = \frac{1}{n}\;\mathop \sum \limits_{{{\text{k}} = 1}}^{n} \left( {Y_{\text{k}} } \right)$$
(9)

where \(\hat{Y}_{\text{k}}\) = predicted values, \(Y_{\text{k}}\) = experimental values.

The coefficient of determination value obtained in this study was with 0.9968 an acceptable range i.e. 0.0032%. The predicted and actual compressive strength values obtained by the regression equation are shown in Table 9. From this cross-validation we concluded that there is no deviation between the predicted and actual values.

Table 9 Response surface model for cross validation
Full size table

8.2 Residual sum of squares (RSS)

Residual sum of squares also called as a summed square of residuals. It is a well-known parameter to calculate the discrepancy between the estimation model and data set using below Eq. 10. The smaller residual sum of square values shows a goodness of fit of the model to the data [54].

$${\text{Residual}}\;{\text{sum}}\;{\text{of}}\;{\text{squares}} = \mathop \sum \limits_{{{\text{k}} = 1}}^{n} \left( {{\text{Y}}_{\text{k}} - \hat{Y}_{\text{k}} } \right)^{2}$$
(10)

The residual sum of square value obtained was 2.634 as shown in Table 9, thus indicating a good fit.

9 Conclusion

In this research, the effect of non-chloride hardening accelerator on mechanical properties and durability of alccofine concrete was studied. From the results, it is concluded that non-chloride accelerator enhanced compressive strength of the concrete at an early age due to increase in the rate of hydration of C3A and C3S phases of cement. The use of accelerator didn’t influence the mechanical properties of concrete at a later age thereby resulting in same values for all mixes. Water absorption values were less with alccofine in the combination of non-chloride accelerator compared to reference concrete due to micro particle size of alccofine which made the concrete more denser, more compacted and also improved the pore structure of the concrete which helped to improve strength as well as reduces the water absorption percentage. Concrete with accelerator was also found to have increased the resistance against sulphate attack. The experimental data was modelled by using the response surface method in Minitab software which employed regression analysis for predicting the values where in the error were formed to be tolerance acceptable range (95% confidence level) thus implying reliable prediction. The cross validation of data by evaluation of coefficient of determination and residual sum of squares conformed the same.