Understanding the behavior of recycled aggregate concrete by using thermogravimetric analysis

Abstract

The physio-chemical changes in concrete mixes due to different coarse aggregate (natural coarse aggregate and recycled coarse aggregate (RCA)) and mix design methods (conventional method and Particle Packing Method (PPM)) are studied using thermogravimetric analysis of the hydrated cement paste. A method is proposed to estimate the degree of hydration (\alpha) from chemically bound water (W_{\mathrm{B}}). The PPM mix designed concrete mixes exhibit lower \alpha. Recycled aggregate concrete (RAC) mixes exhibit higher and \alpha after 7 d of curing, contrary to that after 28 and 90 d. The chemically bound water at infinite time (W_{\mathrm{B}\infty }) of RAC mixes are lower than the respective conventional concrete mixes. The lower W_{\mathrm{B}\infty }, Ca(OH)₂ bound water, free Ca(OH)₂ content and FT-IR analysis substantiate the use of pozzolanic cement in the parent concrete of RCA. The compressive strength of concrete and \alpha cannot be correlated for concrete mixes with different aggregate type and mix design method as the present study confirms that the degree of hydration is not the only parameter which governs the macro-mechanical properties of concrete. In this regard, further study on the influence of interfacial transition zone, voids content and aggregate quality on macro-mechanical properties of concrete is needed.

Keywords

recycled aggregate concrete / Particle Packing Method / thermogravimetric analysis / chemically bound water / degree of hydration / Fourier transform infrared spectroscopy

Introduction

Concrete is the second most-consumed material in the world and aggregates occupy about 70%–80% of its total volume. This causes the rapid exhaustion of non-renewable natural resources. The construction activities also produce construction and demolition (C&D) waste in abundance. The bulk amount of C&D waste is waste concrete. The waste concrete is considered as a source to recycle and extract qualified aggregates, which can be used as a new raw material for concrete production. However, the earlier investigations [¹–⁵] substantiated the influence of inferior quality of recycled coarse aggregate (RCA) on the macro-mechanical properties of recycled aggregate concrete (RAC). Hence, it is required to study the influence of micro-level characteristics on the macro-level performance. The present paper discusses the effect of RCA on the degree of hydration of cement in RAC and its relationship with the mechanical properties of the prepared concrete.

The four major compounds present in ordinary Portland cement (OPC) are tricalcium silicate (C₃S), dicalcium silicate (C₂S), tricalcium aluminate (C₃A), and tetracalcium aluminoferrite (C₄AF). The hydration products of these compounds are the bridging units between the aggregates, as well as the strength contributors. The chemical compounds formed as a result of hydration reactions in concrete at the microstructure level can be examined by thermogravimetric analysis (TGA). The TGA technique measures the mass loss due to the decomposition of hydration products, such as ettringite, calcium silicate hydrate (CSH), calcium hydroxide (CH) and carbonated calcium hydroxide. The measured mass loss is due to the decomposition of chemically bound water (W_{\mathrm{B}}) present in the hydration products. The extent of hydration can be estimated from the mass loss recorded in the TGA test.

The heterogeneity of cement and mineral admixture makes the stoichiometry of cement hydration very complicated. For this, the TGA technique is used by researchers to understand the development of the hydration process of OPC as well as blended cement [⁶–¹⁴]. Furthermore, the degree of hydration (\alpha) of the cementitious materials is correlated with the mechanical properties of the concrete and observed a direct relationship with the compressive strength parameter [⁹,¹⁰,¹⁴]. In the field of RAC, Tam et al. [¹⁵] used the differential scanning calorimetry (DSC) technique to understand the physiochemical changes in RAC and reported that the hydration process of RAC was hindered by the presence of microcracks in recycled coarse aggregate (RCA). Moreover, in two-stage mixing approach (TSMA) higher summation of hydration products was observed in comparison to the conventional mixing approach.

Research significance

The TGA technique was employed by the researchers to estimate the degree of hydration of the cementitious materials and subsequently, the correlation between \alpha and macro-mechanical properties was examined. As per the authors’ knowledge, the TGA technique has not been considered yet to understand the behavior of RAC. The physio-chemical behavior of concrete is influenced by the use of different types of aggregate, aggregate contents, cement contents, mix design method and mixing approaches. In the present study, the natural aggregate concrete (NAC) and RAC are proportioned using conventional mix design (IS: 10262-2009) [¹⁶] and Particle Packing Method (PPM) of mix design [¹⁷] approaches. The normal mixing approach and TSMA [¹⁸] are used for conventional mix design and PPM mix design, respectively. Hence, the role of mix design methods and mixing approaches on the degree of hydration of NAC as well as RAC needs to be studied. The TGA technique is applied to measure the mass loss due to the decomposition of hydration products and subsequently, \alpha in different types of concrete is estimated. In this context, a method is proposed to estimate the degree of hydration from the measured W_{\mathrm{B}} by combining the methods proposed by Pane and Hansen [⁸] and Monteagudo et al. [⁹], which ameliorates the limitations of the existing methods. The mortar adhered to RCA disintegrates during the mixing process and the unhydrated cement and cementitious material present in it influences the hydration process. In this regard, Fourier Transform Infrared Spectroscopy (FT-IR) technique is used to characterize the cement used in the parent concrete of RCA. Subsequently, the hydration products in the four types of concrete are quantified and the correlation between and macro-mechanical properties is analyzed.

Experimental program

Concrete preparation

The concrete mixes were prepared by using the natural coarse aggregate (NCA), RCA collected from IL&FS Environmental Infrastructure and Services Ltd. Plant (New Delhi), river sand of Zone II (IS: 383-1970 [¹⁹]) and OPC of 53 grade (IS: 12269-2013 [²⁰]). The NCA and RCA used in the present study for the preparation of NAC and RAC, respectively are shown in Fig. 1. Both NAC and RAC were prepared by the conventional mix design method specified in IS: 10262-2009 [¹⁶] and PPM mix design approach [¹⁷]. During the concrete mixing the normal mixing approach and established TSMA [¹⁸] were used for concrete proportioned using IS code method and PPM mix design approach, respectively. An effective ratio of 0.45 was used for all the mixes. The mix proportions for different types of concrete are shown in Table 1. The detail description about the mix design methods used in the present study and mix proportions can be found elsewhere [¹⁷].

Thermogravimetric analysis

The samples for the TGA test were collected after 7, 28, and 90 d of curing period. For each curing age, at least three samples were tested. The cubes and cylinders prepared for the compressive strength and tensile strength test were used to collect the sample for TGA test. In this context, the core portion of the tested cubes and cylinders was preferred. The hydrated cement paste was collected from the vicinity of aggregate as suggested by Tam et al. [¹⁵]. The collected powder sample was then sieved through 75 μm sieve [¹³]. The hydration process of the collected sample was arrested by using isopropanol solvent. Zhang and Scherer [²¹] compared the different available physical and chemical techniques to seize the hydration process of cement paste and finally recommended the use of isopropanol solvent, as the microstructure of cement paste is less disturbed by this technique. Subsequently, the sample was collected in a closed microcentrifuge tube and placed in a vacuum desiccator to avoid the influence of atmospheric moisture on the hydration process.

The TGA test was performed using Perkin Elmer Pyris Diamond TG-DTA with a balance of accuracy of 0.1 μg. About 10 to 15 mg of sample was placed in an alumina crucible. This is heated in the dynamic heating ramp between 30°C and 1000°C at a heating rate of 10°C/min. The test was conducted under Ar atmosphere.

Method of investigation

TGA is a technique which monitors the change in mass of the sample material as a function of time at a specified temperature (isothermal mode) or over a temperature range. In this test, both the sample and reference materials are subjected to the same environment at a preferred rate of heating or cooling. The hydrated cement paste, when subjected to TGA test, the decomposition of cement hydrates can be witnessed from the descending TGA curve (Fig. 2).

The mass loss of the hydrated cement paste subjected to 105°C to 1000°C in the TGA test is used to assess the degree of hydration [⁶,¹⁰,²²]. The endothermic effects can be separated into three major phases based on the decomposition of cement hydrates. The first phase is characterized by the dehydration (Ldh) of ettringite and CSH, the second phase corresponds to the dehydroxylation (Ldx) of CH and the third phase is due to the decarbonation (Ldc) of CaCO₃ [⁶,⁷,⁹,¹⁰]. The temperature range of Ldh, Ldx, and Ldc phases reported in the earlier studies [⁶,⁸–¹⁰] are represented in Table 2. Apart from Pane and Hansen [⁸], the initial temperature for the dehydration phase is considered as 105°C [⁶,⁹,¹⁰,²²]. Monteagudo et al. [⁹] observed that the mass loss associated between temperature 105°C and 140°C due to the physically bound water contributes in improving the degree of saturation of dissolutions and precipitation of the hydration products. Furthermore, Monteagudo et al. [⁹] determined the temperature range for the dehydroxylation from the second derivative of the DTA curve for each sample.

In the present study, the initial temperature for the dehydration phase is considered as 105°C. The temperature range for Ldx phase is determined for the individual sample from the first derivative of the DTA curve (dDTA). A typical dDTA curve is shown in Fig. 3, where between the temperature 400°C and 600°C the temperature corresponds to the minimum and maximum values of dDTA, provide the lower limit and upper limit of Ldx phase. From the dDTA curves, the temperature range for Ldx phase is observed to be between 420°C and 500°C for all the sample. Hence, in the present study, the temperature range for Ldh, Ldx, and Ldc are specified between 105°C to 420°C, 420°C to 500°C, and 500°C to 1000°C.

Degree of hydration according to Bhatty [⁶]

Bhatty [⁶] proposed a method to estimate \alpha from the mass loss obtained from the TGA test of the hydrated cement paste. In this context, W_{\mathrm{B}} is calculated from the mass loss measured in Ldh, Ldx, and Ldc phases within the temperature range of 105°C to 1100°C. This method does not account the change in mass for the carbonated compound of anhydrous material. The conversion factor of value 0.41 (ratio of the molecular mass of H₂O and CO₂) is multiplied with Ldc to satisfy the assumption that, the derived chemically bound water owing to the decarbonation is from carbonated CH. Equation (1) represents the expression to calculate W_{\mathrm{B}} and subsequently, \alpha is estimated from Eq. (2) as suggested by Bhatty [⁶]. At the infinite time the maximum chemically bound water (W_{\mathrm{B}\infty }) estimated according to the stoichiometry of cement is 0.23 to 0.25 g per gram of cement paste [⁶,²³,²⁴]. Hence, Bhatty [⁶] assumed the value 0.24 to estimate \alpha of the hydrated cement paste from Eq. (2).

(1)W_{\mathrm{B}}=Ldh+Ldx+\mathrm{0.41}Ldc,

(2)\alpha ={\displaystyle \frac{W_{\mathrm{B}}}{\mathrm{0.24}}}\times \mathrm{100}.

Degree of hydration according to Pane and Hansen [⁸]

Later Pane and Hansen [⁸] measured the mass loss within the temperature range of 140°C to 1100°C of TGA test. In this method, the mass loss due to the decarbonation of carbonated anhydrous material is accounted. Hence, the correction due to the carbonated anhydrous material ({{\displaystyle Ldc}}_{\mathrm{a}}) is subtracted from Ldc and the final expression for W_{\mathrm{B}} is represented in Eq. (3). However, the conversion factor of 0.41 is not considered by Pane and Hansen [⁸] to account the chemically bound water of the carbonated CH. Moreover, unlike Bhatty [⁶] a fixed value is not assumed for W_{\mathrm{B}\infty }. In this context, a three-parameter equation (Eq. (4)) is suggested, which fits the experimental data of W_{\mathrm{B}} obtained at different curing age (t). The parameters τ and \alpha of three-parameter equation represented in Eq. (4) control the intercept and curvature of the graph, respectively on a logarithmic scale. Finally, \alpha can be estimated from Eq. (5) by using W_{\mathrm{B}} and W_{\mathrm{B}\infty }.

(3)W_{\mathrm{B}}=Ldh+Ldx+\left(Ldc-Ld{c}_{\mathrm{a}}\right),

(4)W_{\mathrm{B}}=\hspace{0.17em}{W}_{\mathrm{B}\infty }\times \exp \left[-{\left({\displaystyle \frac{\tau }{t}}\right)}^a\right],

(5)\alpha ={\displaystyle \frac{W_{\mathrm{B}}}{W_{\mathrm{B}\infty }}}\times \mathrm{100}.

Degree of hydration according to Monteagudo et al. [⁹]

The method suggested by Monteagudo et al. [⁹] to calculate W_{\mathrm{B}} is based on the contributions of Bhatty [⁶] and Pane and Hansen [⁸]. Similar to Pane and Hansen [⁸], there is no fixed value is set for W_{\mathrm{B}\infty }. An equation similar to Michaelis-Menten equation is suggested to determine W_{\mathrm{B}\infty }. The experimental data of W_{\mathrm{B}} at different curing age (t) is fitted using Eq. (7) and subsequently, W_{\mathrm{B}\infty } is obtained. Consequently, using the calculated values of W_{\mathrm{B}} and W_{\mathrm{B}\infty }, the \alpha is estimated from Eq. (8).

(6)W_{\mathrm{B}}=Ldh+Ldx+\mathrm{0.41}\left(Ldc-Ld{c}_{\mathrm{a}}\right),

(7)W_{\mathrm{B}}={\displaystyle \frac{W_{\mathrm{B}\infty }\times t}{t+K}},

(8)\alpha ={\displaystyle \frac{W_{\mathrm{B}}}{W_{\mathrm{B}\infty }}}\times \mathrm{100}.

Degree of hydration according to present study

The aforementioned methods to estimate the degree of hydration have some advantages as well as disadvantages. The method suggested by Bhatty [⁶] is simple and easy to employ. However, the calculation of W_{\mathrm{B}} does not include the carbonated anhydrous material content. In addition to this, W_{\mathrm{B}\infty } is taken as 0.24 as observed from the stoichiometric analysis of cement. Practically it is difficult to obtain the value 0.24 for W_{\mathrm{B}\infty }. Pane and Hansen [⁸] and Monteagudo et al. [⁹] also observed that, W_{\mathrm{B}\infty } value is less than 0.24 for OPC cement and for blended cement the W_{\mathrm{B}\infty } value is even lower than that of OPC cement. Hence, a fixed value (0.24) cannot be considered for W_{\mathrm{B}\infty }. In this context, Pane and Hansen [⁸] accounted the carbonated anhydrous material content and used the three-parameter equation to calculate W_{\mathrm{B}\infty } from the available W_{\mathrm{B}} data at different curing period. However, while calculating W_{\mathrm{B}} the conversion factor 0.41 is not applied to account only the chemically bound water correspond to the CH equivalent in carbonated portlandite. The method suggested by Monteagudo et al. [⁹] included the conversion factor (0.41) while calculating W_{\mathrm{B}}. In addition to this, a Michaelis-Menten type equation is proposed to obtain the value of W_{\mathrm{B}\infty } from the calculated W_{\mathrm{B}} values at different curing period. Although Michaelis-Menten type equation yields a single solution of W_{\mathrm{B}\infty }, the value of W_{\mathrm{B}\infty } is significantly low. This yields the estimated \alpha even more than 100% at advanced curing age of the cement paste [⁹], which is scientifically inappropriate.

The method used in the present study to calculate W_{\mathrm{B}} and W_{\mathrm{B}\infty } is the synergy of the methods proposed by Pane and Hansen [⁸] and Monteagudo et al. [⁹]. In this context, the expression suggested by Monteagudo et al. [⁹] to estimate W_{\mathrm{B}} is adopted, as it accounts only the chemically bound water correspond to the CH equivalent in carbonated portlandite by incorporating the conversion factor of 0.41. Furthermore, instead of the Michaelis-Menten type equation, the three-parameter equation suggested by Pane and Hansen [⁸] to predict W_{\mathrm{B}\infty } value is adopted, which avoids the underestimation of W_{\mathrm{B}\infty }. Consequently, \alpha is estimated from Eq. (11), where W_{\mathrm{B}} and W_{\mathrm{B}\infty } are calculated by conjugating the contributions of Bhatty [⁶], Pane and Hansen [⁸] and Monteagudo et al. [⁹].

(9)W_{\mathrm{B}}=Ldh+Ldx+\mathrm{0.41}\left(Ldc-Ld{c}_{\mathrm{a}}\right),

(10)W_{\mathrm{B}}=\hspace{0.17em}{W}_{\mathrm{B}\infty }\times \exp \left[-{\left({\displaystyle \frac{\tau }{t}}\right)}^a\right],

(11)\alpha ={\displaystyle \frac{W_{\mathrm{B}}}{W_{\mathrm{B}\infty }}}\times \mathrm{100}.

Results and discussion

Calculation of mass loss at different phases and W_{\mathrm{B}}

The mass of the sample at different stages of the decomposition of hydrated cement paste during the TGA test is specified in Table 3. In addition to the mass of the sample at 105°C, the initial and final mass of Ldx phase (420°C and 500°C) and 1000°C, the mass of the sample correspond to 140°C is also reported in Table 3 to estimate \alpha using Pane and Hansen [⁸] method.

The recorded mass at a specified temperature is used to calculate the mass loss in Ldh, Ldx, and Ldc phases and reported in Table 4. The correction in the decarbonation phase due to {{\displaystyle Ldc}}_{\mathrm{a}} is used to calculate W_{\mathrm{B}} in Pane and Hansen [⁸] and Monteagudo et al. [⁹] methods. The {{\displaystyle Ldc}}_{\mathrm{a}} value is recorded as 1.055% for the anhydrous cement. Subsequently, the W_{\mathrm{B}} is calculated for each sample using the expressions suggested by Bhatty [⁶], Pane and Hansen [⁸], and Monteagudo et al. [⁹] and represented in Table 4. The W_{\mathrm{B}} content calculated by Pane and Hansen [⁸] method exhibit larger variation in comparison to Bhatty [⁶] and Monteagudo et al. [⁹] methods for different samples of a particular type of concrete after a certain curing period. This variation in W_{\mathrm{B}} content of different replications in Pane and Hansen [⁸] method is due to the absence of conversion factor 0.41 for the decarbonation part. The W_{\mathrm{B}} content increases as the curing age of the sample increases. For a particular mix design approach the W_{\mathrm{B}} content of RAC mixes are observed to be higher than NAC mixes after 7 d of curing and after 28 d of curing the W_{\mathrm{B}} content of NAC mixes is marginally higher than RAC mixes. However, after 90 d of curing higher W_{\mathrm{B}} content is observed for NAC samples in comparison to the RAC samples. The higher W_{\mathrm{B}} content at the early stage (7 d) signifies the formation of a higher amount of hydration products. During the mixing of RAC, the attached mortar present in RCA gets disintegrated. The unhydrated cement and cementitious material present in the disintegrated mortar increase the total cementitious material content, which may be increasing the hydration products content of RAC at the early stage. On the contrary, the increased surface area of RAC mixes due to the disintegrated unhydrated cementitious material and fine aggregate the effective w/c of RAC mixes become lower than 0.45. A comparative study between w/c 0.35–0.45 showed that, the heat of hydration of cement paste increases as the w/c increases [⁸,²⁵]. The availability of less capillary space retards the growth of hydration products at lower w/c [⁸]. This may be the possible cause which hinders the growth of hydration products in RAC and subsequently, lowers the W_{\mathrm{B}} content at the later stage (28 and 90 d of curing).

Estimation of W_{\mathrm{B}\infty }

The calculated W_{\mathrm{B}} content is employed to estimate the W_{\mathrm{B}\infty } for each type of concrete using Pane and Hansen [⁸] and Monteagudo et al. [⁹] methods. For a particular type of concrete mix the W_{\mathrm{B}\infty } values of each sample at different curing period are operated. The W_{\mathrm{B}\infty } values of a particular type of concrete mix are plotted and the data points are fitted by employing the expressions specified by Pane and Hansen [⁸] and Monteagudo et al. [⁹]. Consequently, the estimated W_{\mathrm{B}\infty } values in Pane and Hansen [⁸] method, Monteagudo et al. [⁹] method and present method are represented in Table 5. The W_{\mathrm{B}\infty } values obtained in Monteagudo et al. [⁹] method are significantly low with respect to that of Pane and Hansen [⁸] method and present method.

The estimated W_{\mathrm{B}\infty } values of RAC mixes are lower than that of NAC mixes in each of the discussed method (Table 5). The earlier studies [⁸–¹¹] observed lower W_{\mathrm{B}\infty } values for the blended cement. This indicates the blending of pozzolanic material with the OPC to prepare the parent concrete of the RCA. During the mixing, the attached mortar of RCA containing unhydrated cement and pozzolanic material gets disintegrated. The unhydrated cementitious material of the RCA takes part in the hydration process and the presence of pozzolanic material yields lower W_{\mathrm{B}\infty } values for RAC mixes. Furthermore, the W_{\mathrm{B}\infty } of PPM mixes are lower than conventional mixes for both NAC and RAC. The lower W_{\mathrm{B}\infty } values of PPM mixes is possibly related to the higher fine aggregate content. The higher fine aggregate content in the PPM mixes increases the surface area, which resulted in lowering the interface w/c ratio. Pane and Hansen [⁸] observed a lower W_{\mathrm{B}\infty } for w/c ratio 0.35 in comparison to w/c ratio of 0.45 owing to the limited availability of the capillary space. Consequently, the increase in surface area in PPM mixes decreases the effective w/c ratio in the interface and causes the yielding of lower W_{\mathrm{B}\infty }.

Degree of hydration (\alpha)

The calculated W_{\mathrm{B}\infty } values in different methods are used to estimate \alpha of different mixes at different curing period and shown in Figs. 4 to 7. The \alpha values obtained in Bhatty [⁶] method are lower in comparison to the other discussed methods due to the assumption of higher W_{\mathrm{B}\infty } value (0.24). The lower W_{\mathrm{B}\infty } values obtained in Monteagudo et al. [⁹] method yield the degree of hydration more than 100% for the 90 d cured samples. Interestingly, the \alpha values of RAC mixes are observed to be higher with respect to the NAC mixes in Monteagudo et al. [⁹] method. The \alpha values of PPM mix designed concrete are lower than the conventional mix designed concrete for both NAC and RAC. This may be due to the same reason explained for lower W_{\mathrm{B}\infty } values observed in PPM mix designed concrete.

Investigation on the presence of pozzolanic material in RCA

Estimation of CH bound water and free CH content

The presence of pozzolanic material in RCA and its participation in the hydration process of RAC is discussed in this section. For both NAC and RAC mixes only OPC is used. However, the lower W_{\mathrm{B}\infty } content for RAC mixes indicates the presence of pozzolanic material in RCA, which governs the secondary hydration reaction in RAC. In this context, the CH bound water and free CH content are calculated by using the established expressions. The CH bound water is calculated by using Eq. (12) and represented in Table 6. The CH bound water content increases as the curing period increases for each type of concrete. Apart from the silica fume blended OPC mix, Monteagudo et al. [⁹] witnessed an increasing trend of CH bound water for only OPC as well as blended cement (OPC+ GGBS and OPC+ fly ash). This indicates the absence of silica fume in RCA. However, the presence of other pozzolanic materials; such as GGBS and fly ash cannot be confirmed from the estimated CH bound water content. For this, the free CH content is estimated by using Eq. (13) [²⁶] as well as the modified expression (Eq. (14)) suggested by Monteagudo et al. [⁹].

(12)\mathrm{CH}\hspace{0.17em}\mathrm{b}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{d}\mathrm{w}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}=Ldx+\mathrm{0.41}\left(Ldc-Ld{c}_{\mathrm{a}}\right),

(13)\mathrm{F}\mathrm{r}\mathrm{e}\mathrm{e}\hspace{0.17em}\mathrm{CH}=\mathrm{4.11}Ldx+\mathrm{1.68}Ldc,

(14)\mathrm{F}\mathrm{r}\mathrm{e}\mathrm{e}\hspace{0.17em}\mathrm{C}{\mathrm{H}}_{\mathrm{M}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{a}\mathrm{g}\mathrm{u}\mathrm{d}\mathrm{o}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{l}.}=\mathrm{4.11}Ldx+\mathrm{1.68}\left(Ldc-Ld{c}_{\mathrm{a}}\right).

The free CH content estimated from Eqs. (13) and (14) are shown in Table 6. It is observed that for a particular type of mix design approach the estimated free CH content of NAC mixes are higher than RAC mixes. Monteagudo et al. [⁹] reported a lower free CH content for GGBS and fly ash blended cement with respect to OPC. Hence, the similar relationship between free CH content and RAC mixes with respect to the respective NAC mixes provides the evidence for the existence of pozzolanic material (GGBS or fly ash or both) in RCA. Moreover, the increase in free CH content with age is less for the RAC mixes. This indicates the presence of pozzolanic material in very less quantity, which is from the disintegrated mortar of RCA.

Fourier transform infrared spectroscopy analysis

The presence of unhydrated cement is analyzed using FT-IR. In this context, the cement paste present in the attached mortar of RCA and hydrated cement paste present in NAC (prepared using OPC) were analyzed. The collected samples were mixed with KBr (acts as beam splitter) at a weight ratio of 1/100 to prepare pellets. The prepared pellets were scanned using the spectrometer (Nicolet 6700 FT-IR spectrometer) in the mid-infrared spectra (wavelength ranges from 400 to 4000 cm⁻¹).

The FT-IR spectra for the aforementioned samples are represented in Fig. 8 and following major peaks are discussed in the present study. The peak corresponds to the wavelength 3640 m⁻¹ is related to the stretching of O-H bond of {{\displaystyle \mathrm{C}\mathrm{a}\left(\mathrm{O}\mathrm{H}\right)}}_{\mathrm{2}} [²⁷–³⁴]. For the sample collected from RCA, the peak at 3640 cm⁻¹ is absent (Fig. 8). This is possibly due to the transformation of {{\displaystyle \mathrm{C}\mathrm{a}\left(\mathrm{O}\mathrm{H}\right)}}_{\mathrm{2}} to {{\displaystyle \mathrm{C}\mathrm{a}\left(\mathrm{C}\mathrm{O}\right)}}_{\mathrm{3}} because of the exposure of RCA to atmospheric {{\displaystyle \mathrm{C}\mathrm{O}}}_{\mathrm{2}}. The bands correspond to the peaks approximately at 2983, 2875, 2510, and 1795 cm⁻¹ are due to the stretching of C-O bond of {{\displaystyle \mathrm{C}\mathrm{O}}}_{\mathrm{3}}^{-\mathrm{2}} [²⁷,³²,³⁴–³⁶]. The peaks observed at wavelength 1425, 874, and 712 cm⁻¹ are respectively due to the asymmetric stretching, out of plane bending and doubly degenerate planar bending of {{\displaystyle \mathrm{C}\mathrm{O}}}_{\mathrm{3}}^{-\mathrm{2}} [²⁷–²⁹,³¹–³³,³⁵–³⁸]. The broad peak between 2800 and 3700 cm⁻¹ is due to the stretching vibration of O-H bond of {{\displaystyle \mathrm{H}}}_{\mathrm{2}}\mathrm{O}, which is strongly associated with cementitious material by hydrogen bond [²⁷–²⁹,³²–³⁴,³⁷,³⁹]. The peak corresponds to 1640 cm⁻¹ of NAC sample is due to the bending vibration of H-O-H bond of {{\displaystyle \mathrm{H}}}_{\mathrm{2}}\mathrm{O} molecule [²⁷–³⁰,³²–³⁵,³⁷,³⁸]. In addition to the band associated with water between 1500 and 1700 cm⁻¹ the presence of a hump between 970 and 1100 cm⁻¹ is the indicative of the presence of CSH [³¹,³²]. The bands or peaks of FT-IR spectra discussed above are associated with the compounds of hydrated cement paste.

For both the samples the peak observed at wavelength 525 cm⁻¹ is associated with the presence of unhydrated cement. The peak corresponds to 525 cm⁻¹ is resulted due to the out of plane bending vibration of Si-O bond of silicate [²⁹,³⁰,³⁵,⁴⁰]. This confirms the presence of unhydrated cement in RCA. Interestingly, a significant peak is observed at wavelength 1630 cm⁻¹ along with the peaks at 1035 and 3450 cm⁻¹ for the sample collected from RCA. It is worth mentioning that, in the FT-IR spectra the peaks correspond to 1035, 1630, and 3450 cm⁻¹ are observed in fly ash [⁴¹–⁴⁴]. The peak at 1035 cm⁻¹ is due to the asymmetric stretching vibration of Al-O or Si-O bonds, whereas the peaks at 1630 and 3450 cm⁻¹ are because of the stretching and deformation vibration of OH and H-O-H groups. Hence, the presence of fly ash in the adhered mortar of RCA is substantiated, which is possibly due to the use of fly ash blended cement in the parent concrete.

Relationship between degree of hydration and compressive strength

The compressive strength of different types of concrete recorded in the present study is represented in Fig. 9. Irrespective of the method, the estimated \alpha value increases with the increase in curing age. This is similar to the usual trend of compressive strength of concrete with age. Apart from the Monteagudo et al. [⁹] method, the trend of compressive strength of concrete (Fig. 9) is very similar to the relationship of \alpha with curing age. The \alpha values estimated by Monteagudo et al. [⁹] method is observed to be higher for RAC mixes with respect to the NAC mixes (Fig. 6), which is contrary to the recorded compressive strength results. The \alpha of RAC IS mix is observed to be significantly higher than the other concrete mixes after 7 d of curing, which is also reflected in the compressive strength behavior. For 28 d cured mixes the concrete mix designed by IS code method exhibited higher compressive strength than PPM mix designed concrete (Fig. 9). The \alpha values obtained in Bhatty [⁶] method (Fig. 4) and Pane and Hansen [⁸] method (Fig. 5) and present method (Fig. 7) also exhibit similar relationship for 28 d cured specimens. Furthermore, higher compressive strength values are witnessed for NAC mixes in comparison to the RAC mixes after 90 d of curing. This observation is correlated with the \alpha values estimated by using the Bhatty [⁶] method (Fig. 4) and Pane and Hansen [⁸] method (Fig. 5). In the present method, the \alpha of 90 d cured NAC IS mix is observed to be significantly higher, whereas for RAC IS mix the \alpha value is marginally lower than the NAC PPM mix, which is contrary to compressive strength results after 90 d of curing.

Similar to the earlier studies [⁸–¹⁰] the present investigation also observed a direct relationship between \alpha and compressive strength at different curing age. However, this is limited only for the same concrete mix. While comparing the \alpha and compressive strength of different concrete mixes a consistent relationship is not observed. Consequently, \alpha of cement is not the only parameter which influences the mechanical parameters of concrete. In addition to \alpha, ITZ characteristics, interface voids and its distribution and aggregate quality also influence the macro-mechanical properties of concrete.

Conclusions

The present study investigated the TGA test results of the hydrated cement paste collected from the four different types of concrete prepared using NCA and RCA and two different mix design approaches. The \alpha of the prepared concrete mixes were estimated after 7, 28, and 90 d of curing period. In this context, a method is proposed by combining two earlier methods to estimate \alpha from the recorded mass loss in the TGA test, which accounts the carbonated anhydrous material content and determines W_{\mathrm{B}\infty } using the three-parameter equation. The W_{\mathrm{B}\infty } of RAC mixes were estimated to be lower than the respective NAC mixes, which indicates the presence of pozzolanic material in RCA used in the present study. In this context, the CH bound water, free CH and FT-IR analysis supported the use of pozzolanic material blended cement in the parent concrete of RCA.

It was observed that the \alpha values of RAC after 7 d and 28 d of curing period were higher than NAC irrespective of mix design method. However, after 90 d of curing the \alpha values of RAC were lower than NAC for respective mix design methods. Moreover, irrespective of the aggregate type the \alpha values of PPM mix designed concrete were estimated to be lower than the concrete proportioned by IS code method after 90 d of curing. Because of the higher degree of hydration of RAC after 7 and 28 d of curing the material inferiority was successfully counter balanced, which was also reflected in the compressive strength results of RAC. In this regard, the contribution of pozzolanic material present in adhered mortar of RCA with secondary hydration products to improve the ITZ as well as interface characteristics is justified. However, the lower degree of hydration after 90 d of curing indicates that at the later stage the availability of pozzolanic material for secondary hydration reaction was not significant. This analysis suggests that, by using pozzolanic materials with appropriate quantity the RAC can be prepared with satisfactory short-term as well as long-term properties.

For a particular type of concrete, the compressive strength parameter was observed to be directly proportional to \alpha at different curing age, whereas this relationship was not substantiated for concrete with different aggregate type and mix design approach. Hence, in addition to \alpha, the parameters, such as, ITZ characteristics, interface voids content and its distribution, and aggregate quality needs to be integrated while correlating the macro-level performance of concrete with micro-level characteristics.

References

[1] Chakradhara Rao M, Bhattacharyya S K, Barai S V. Influence of field recycled coarse aggregate on properties of concrete. Materials and Structures, 2011, 44(1): 205–220

[2] Mukharjee B B, Barai S V. Influence of Nano-Silica on the properties of recycled aggregate concrete. Construction & Building Materials, 2014, 55: 29–37

[3] Fathifazl G, Abbas A, Razaqpur A G, Isgor O B, Fournier B, Foo S. New mixture proportioning method for concrete made with coarse recycled concrete aggregate. Journal of Materials in Civil Engineering, 2009, 21(10): 601–611

[4] Knaack A M, Kurama Y C. Design of concrete mixtures with recycled concrete aggregates. ACI Materials Journal, 2013, 110: 483–492

[5] Pepe M, Toledo Filho R D, Koenders E A B, Martinelli E. A novel mix design methodology for Recycled Aggregate Concrete. Construction & Building Materials, 2016, 122: 362–372

[6] Bhatty J I. Hydration versus strength in a Portland cement developed from domestic mineral wastes—A comparative study. Thermochimica Acta, 1986, 106: 93–103

[7] Bhatty J I. A review of the application of thermal analysis to cement-admixture systems. Thermochimica Acta, 1991, 189(2): 313–350

[8] Pane I, Hansen W. Investigation of blended cement hydration by isothermal calorimetry and thermal analysis. Cement and Concrete Research, 2005, 35(6): 1155–1164

[9] Monteagudo S M, Moragues A, Gαlvez J C, Casati M J, Reyes E. The degree of hydration assessment of blended cement pastes by differential thermal and thermogravimetric analysis. Morphological evolution of the solid phases. Thermochimica Acta, 2014, 592: 37–51

[10] Deboucha W, Leklou N, Khelidj A, Oudjit M N. Hydration development of mineral additives blended cement using thermogravimetric analysis (TGA): Methodology of calculating the degree of hydration. Construction & Building Materials, 2017, 146: 687–701

[11] Zeng Q, Li K, Fen-Chong T, Dangla P. Determination of cement hydration and pozzolanic reaction extents for fly-ash cement pastes. Construction & Building Materials, 2012, 27(1): 560–569

[12] Ye G, Liu X, De Schutter G, Poppe A M, Taerwe L. Influence of limestone powder used as filler in SCC on hydration and microstructure of cement pastes. Cement and Concrete Composites, 2007, 29(2): 94–102

[13] Vedalakshmi R, Sundara Raj A, Srinivasan S, Ganesh Babu K. Quantification of hydrated cement products of blended cements in low and medium strength concrete using TG and DTA technique. Thermochimica Acta, 2003, 407(1–2): 49–60

[14] Hemalatha T, Mapa M, George N, Sasmal S. Physico-chemical and mechanical characterization of high volume fly ash incorporated and engineered cement system towards developing greener cement. Journal of Cleaner Production, 2016, 125: 268–281

[15] Tam V W Y, Gao X F, Tam C M, Ng K M. Physio-chemical reactions in recycle aggregate concrete. Journal of Hazardous Materials, 2009, 163(2–3): 823–828

[16] IS 10262-2009. Concrete Mix Proportioning—Guidelines. New Delhi: Bureau of Indian Standards, 2009

[17] Pradhan S, Kumar S, Barai S V. Recycled aggregate concrete: particle packing method (PPM) of mix design approach. Construction & Building Materials, 2017, 152: 269–284

[18] Tam V W Y, Gao X F, Tam C M. Microstructural analysis of recycled aggregate concrete produced from two-stage mixing approach. Cement and Concrete Research, 2005, 35(6): 1195–1203

[19] IS 383-1970. Specification for Coarse and Fine Aggregate from Natural Sources. New Delhi: Bureau of Indian Standards, 1970

[20] IS 12269-2013. Ordanary Portland Cement 53 Grade-Specification. New Delhi: Bureau of Indian Standards, 2013

[21] Zhang J, Scherer G W. Comparison of methods for arresting hydration of cement. Cement and Concrete Research, 2011, 41(10): 1024–1036

[22] Scrivener K L, Lothenbach B, De Belie N, Gruyaert E, Skibsted J, Snellings R, Vollpracht A. TC 238-SCM: Hydration and microstructure of concrete with SCMs. Materials and Structures, 2015, 48(4): 835–862

[23] Copeland L E, Kantro D L, Verbeck G. Chemistry of hydration of Portland cement. In: The 4th International Symposium on the Chemistry of Cement. Washington, D.C., 1960, I: 429–465

[24] Young J F, Hansen W. Volume Relationships for C-S-H Formation Based on Hydration Stoichiometries. In: MRS Online Proceedings Library Archive. 1986, 85: 313

[25] Lura P, Winnefeld F, Fang X. A simple method for determining the total amount of physically and chemically bound water of different cements. Journal of Thermal Analysis and Calorimetry, 2017, 130(2): 653–660

[26] El-Jazairi B, Illston J M. The hydration of cement paste using the semi-isothermal method of derivative thermogravimetry. Cement and Concrete Research, 1980, 10(3): 361–366

[27] Mendes A, Gates W P, Sanjayan J G, Collins F. NMR, XRD, IR and synchrotron NEXAFS spectroscopic studies of OPC and OPC/slag cement paste hydrates. Materials and Structures, 2011, 44(10): 1773–1791

[28] Pan Z Y, Li G, Hong C Y, Kuang H L, Yu Y, Feng F X, Liu D M, Du H. Modified recycled concrete aggregates for asphalt mixture using microbial calcite precipitation. Royal Society of Chemistry Advances, 2015, 5(44): 34854–34863

[29] Bhat P A, Debnath N C. Theoretical and experimental study of structures and properties of cement paste: The nanostructural aspects of C-S-H. Journal of Physics and Chemistry of Solids, 2011, 72(8): 920–933

[30] Mollah M Y A, Yu W, Schennach R, Cocke D L. A Fourier transform infrared spectroscopic investigation of the early hydration of Portland cement and the influence of sodium lignosulfonate. Cement and Concrete Research, 2000, 30(2): 267–273

[31] Peyvandi A, Holmes D, Soroushian P, Balachandra A M. Monitoring of sulfate attack in concrete by Al 27 and Si 29 MAS NMR spectroscopy. Journal of Materials in Civil Engineering, 2015, 27(8): 04014226

[32] Ylmén R, Jäglid U, Steenari B, Panas I. Early hydration and setting of Portland cement monitored by IR, SEM and Vicat techniques. Cement and Concrete Research, 2009, 39(5): 433–439

[33] Trezza M A, Lavat A E. Analysis of the system 3CaO·Al₂O₃–CaSO₄·2H₂O–CaCO₃–H₂O by FT-IR spectroscopy. Cement and Concrete Research, 2001, 31(6): 869–872

[34] Delgado A H, Paroli R M, Beaudoin J J. Comparison of IR techniques for the characterization of construction cement minerals and hydrated products. Applied Spectroscopy, 1996, 50(8): 970–976

[35] Nasrazadani S, Eghtesad R, Sudoi E, Vupputuri S, Ramsey J D, Ley M T. Application of Fourier transform infrared spectroscopy to study concrete degradation induced by biogenic sulfuric acid. Materials and Structures, 2016, 49(5): 2025–2034

[36] Hughes T L, Methven C M, Jones T G J, Pelham S E, Fletcher P, Hall C. Determining cement composition by Fourier transform infrared spectroscopy. Advanced Cement Based Materials, 1995, 2(3): 91–104

[37] Yu P, Kirkpatrick R J, Poe B, McMillan P F, Cong X. Structure of calcium silicate hydrate (C-S-H): near-, mid-, and far-infrared spectroscopy. Journal of the American Ceramic Society, 1999, 82(3): 742–748

[38] Hidalgo López A, García Calvo J L, García Olmo J, Petit S, Alonso M C. Microstructural evolution of calcium aluminate cements hydration with silica fume and fly ash additions by scanning electron microscopy, and mid and near-infrared spectroscopy. Journal of the American Ceramic Society, 2008, 91(4): 1258–1265

[39] Choudhary H K, Anupama A V, Kumar R, Panzi M E, Matteppanavar S, Sherikar B N, Sahoo B. Observation of phase transformations in cement during hydration. Construction & Building Materials, 2015, 101: 122–129

[40] Govin A, Peschard A, Guyonnet R. Modification of cement hydration at early ages by natural and heated wood. Cement and Concrete Composites, 2006, 28(1): 12–20

[41] Zhang Z, Wang H, Provis J L. Quantitative study of the reactivity of fly ash in geopolymerization by FTIR. Journal of Sustainable Cement-Based Materials, 2012, 1(4): 154–166

[42] Guo X, Shi H, Dick W A. Compressive strength and microstructural characteristics of class C fly ash geopolymer. Cement and Concrete Composites, 2010, 32(2): 142–147

[43] Chindaprasirt P, Jaturapitakkul C, Chalee W, Rattanasak U. Comparative study on the characteristics of fly ash and bottom ash geopolymers. Waste Management, 2009, 29(2): 539–543

[44] Rożek P, Król M, Mozgawa W. Spectroscopic studies of fly ash-based geopolymers. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 2018, 198: 283–289

RIGHTS & PERMISSIONS

Higher Education Press