- Email: [email protected]
The role of particle size on the performance of pumice as a supplementary cementitious material
Cement and Concrete Composites 80 (2017) 135e142
Contents lists available at ScienceDirect
Cement and Concrete Composites journal homepage: www.elsevier.com/locate/cemconcomp
The role of particle size on the performance of pumice as a supplementary cementitious material Saamiya Seraj, Rachel Cano, Raissa D. Ferron, Maria C.G. Juenger* The University of Texas at Austin, 301 E. Dean Keeton St., Austin, TX 78712, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 September 2016 Received in revised form 9 February 2017 Accepted 16 March 2017 Available online 20 March 2017
A critical area overlooked in previous research on pumice is understanding how its physical characteristics influence its behavior as a supplementary cementitious material (SCM). This study investigated three pumices with different particle size distributions to observe whether these porous materials exhibit enhanced nucleation and growth of hydration products, in the same way as non-porous materials, and whether the rate of pozzolanic reaction can be changed through particle size. The effect of particle size on compressive strength, rheology and resistance to alkali silica reaction (ASR) was also evaluated. Results showed that reducing particle size increased the rates of cement hydration, pozzolanic reaction, and compressive strength gain, while also increasing mixture viscosity. Interestingly, particle size did not impact the yield stress of the mixture or the resistance to ASR. These new findings give insight about how the particle size of pumice can be used to overcome drawbacks reported in previous literature. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Pozzolan Particle size distribution Compressive strength Durability Rheology
1. Introduction The push towards more economical, sustainable and durable concrete mixtures has made supplementary cementitious materials (SCMs) very important in the concrete industry. However, at a time of increasing SCM demand, global changes in the power generation industry have led to regional supply shortages of fly ash, one of the most widely used SCMs. Consequently, there is increased interest in the cement and concrete community to evaluate other materials that can serve as suitable SCMs to overcome this scarcity. Pumice, a natural pozzolan, is one such promising material that has seen renewed interest in being used as an SCM [1e4]. Pumice is formed from highly silicic volcanic lava that tends to hold more dissolved gases than other types of lava due to its viscosity. As the lava rapidly cools and hardens into a glassy structure, the dissolved gases form pores or vesicles that result in low density volcanic rock [5]. The high silica content and amorphous structure are what give pumice its pozzolanic properties and make it a candidate for use as an SCM in concrete. In fact, pumice was used in concrete throughout history dating back to the ancient Greek and Roman civilizations [3]. Even
* Corresponding author. E-mail addresses: [email protected] (S. Seraj), [email protected] (R. Cano), [email protected] (R.D. Ferron), [email protected] (M.C.G. Juenger). http://dx.doi.org/10.1016/j.cemconcomp.2017.03.009 0958-9465/© 2017 Elsevier Ltd. All rights reserved.
as recent as the 20th century, finely ground pumice was used as an SCM in mass concrete structures, before the price and availability of fly ash made it more economical than pumice. Some notable examples of US structures with pumice are the Los Angeles aqueduct in 1912, the Friant Dam in 1942, the Altus Dam in 1945 and the Glen Canyon Dam in 1964 [3,6]. With the rising demand for SCMs, pumice is experiencing a renaissance in concrete research. Recent studies have shown finely ground pumice to meet most of the requirements listed in ASTM C 618 for natural pozzolans [4,7] and to lower calcium hydroxide content in cementitious mixtures, which is an indication of pozzolanic reaction [4,8,9]. Researchers have also evaluated the performance of pumice mixtures in terms of compressive strength, durability, and mixture workability [1e4,8,9]. However, a critical area of research that has been overlooked in previous literature is understanding how the characteristics of the pumice itself, like particle size and composition, influence its behavior in cementitious mixtures. For example, previous literature has indicated that replacing cement with finely ground pumice decreased the compressive strength of mortar and concrete specimens [1,4,7,9], but there has not been any investigation on how these effects on strength relate back to characteristics of the pumice itself, and whether those characteristics can be modified to improve the compressive strength of pumice blended mixtures. Similarly, for durability, although studies have found pumice to be effective at
136
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
controlling alkali silica reaction (ASR), results vary on the minimum percentage needed to mitigate expansions [2,7]. Since pumice is a natural pozzolan whose properties can vary depending upon its sourcing, these differences in reported results are not unusual. However, it is important to understand how these properties of pumice change its effectiveness in mitigating ASR so that minimum replacement dosages can be predicted for new sources. Evaluating the effects of pumice characteristics on mixture workability is also important, as fresh state properties ultimately affect hardened concrete properties. While there are some studies that have qualitatively measured workability of pumice concrete mixtures using the slump test [4,8,9], a more comprehensive analysis is needed to evaluate how the viscosity and yield stress of pumice blended mixtures change with characteristics of the pumice itself. The main objective of the study presented here was to address the shortcomings in understanding how physical properties of pumices influence their behavior in cementitious systems. We used three pumices with similar compositions but differing particle size distributions to evaluate the role of particle size in the performance of pumice as an SCM. In this way we could hold all material variables constant other than pumice particle size, so as to eliminate the overwhelming effects of parameters such as composition on pozzolanicity, particularly amorphous content and soluble silica and alumina contents [10]. The influence of pumice particle size was examined for its impact on the following: cement hydration kinetics, pozzolanic reactivity, compressive strength, resistance to ASR, and rheology. Based on previous research on other SCMs and filler materials [11,12], we hypothesized that in the early stages of hydration, when the process is dominated by the availability of nucleation and growth sites, pumice with a fine particle size distribution can increase the rate of cement hydration by providing an increased surface area for the nucleation and growth of hydration products. While this is simply referred to as “filler effect” in previous literature [11,12], in this paper it will be referred to as “size filler effect” to differentiate it from “space filler effect,” which is the growth of additional hydration products due to increased space per grain of cement when SCM or filler materials are used [11,12]. Much of the work examining size filler effects has been done with nonporous materials, and it is of interest to determine if the same principles apply for porous SCMs like pumice. Isothermal calorimetry was used to observe how pumice particle size can change cement hydration kinetics through size filler effects. Thermal gravimetric analysis (TGA) was used to measure calcium hydroxide consumption in hydrated pastes to see whether pumice particle size affected the rate of pozzolanic reaction. Compressive strengths of pumice-containing mortars were measured from 1 day to 1 year, to observe how the changes to early cement hydration kinetics and pozzolanic reaction rate correlate to the rate of strength gain. Based on previous research on fly ash and limestone [13,14], we hypothesized that the low early strengths of the pumice mixtures reported in literature [1,4,7,9] were most likely due to the dilutionary effect of replacing hydraulic cement with a slower reacting pozzolanic material like pumice. However, we expected that this dilutionary effect could be overcome, at least partially, to achieve higher early age strengths by using a finer pumice pozzolan that can provide increased surface area to enhance nucleation and growth of hydration products during the early stages of hydration. Further, other than a paper by Mielenz et al. [15], most studies report compressive strength data for pumice-containing mixtures only until 28 days. Therefore, we measured the strength of pumice-containing mortars for one year to provide insight into the role of particle size on the rate of strength gain at later ages and to evaluate the amount of time it takes to offset the dilutionary effect of replacing cement with pumice. ASR resistance of pumice-containing mortar bars was evaluated to see if particle size changed the effectiveness of the
pumice pozzolans at mitigating expansions due to ASR. Finally, the effect of pumice particle size on mixture workability was measured in terms of yield stress and plastic viscosity of pastes. Pumicecontaining mixtures were compared against mixtures with an inert quartz filler and control samples without SCMs or filler materials. 2. Materials and methods The three pumice powders used in this paper were sourced from USA and are referred to as Pumice-N, Pumice-D and Pumice-S. Pumice-N and Pumice-D are commercially available SCMs, while Pumice-S was originally a fine aggregate that was ground in the laboratory using a Bico Inc. UA V-Belt Drive Pulverizer and sieved through a No. 200 sieve (75 mm opening) for use as an SCM. All three pumices were tested to determine if they could be classified as Class N pozzolans according to the guidelines stated in ASTM C 618 [16]. The results, presented in Table 1, show that the physical and chemical characteristics of the pumices fulfilled all of the requirements for Class N pozzolans. X-ray diffraction (XRD) tests, conducted using a Siemens D-500 X-ray Diffractometer with Cu-Ka radiation and a dwell time of 4 s, showed the pumices to be mostly amorphous (Fig. 1). X-ray florescence (XRF) analyses of the pumices were conducted using a Bruker S4 Explorer according to the procedures of ASTM D 4326 [17]. The results, presented in Table 2, confirmed the oxide compositions of the pumices to be almost identical to each other. It was assumed, therefore, that the amorphous content and soluble silica and alumina contents of the pumices were nearly identical, so that the only difference between them was the gradation. An inert quartz filler, purchased from Old Hickory Clay Company, was also used in this study. The cement used in all mixtures was an ASTM C 150 [18] Type I portland cement sourced from USA. The XRF oxide composition of the cement, found using the same procedure as the pumices, is also presented in Table 2. The particle size distributions of the pumices and quartz filler were analyzed using a Horiba Partica LA 950-V2 Laser Scattering Particle Size Distribution Analyzer and are presented in Fig. 2. Table 3 lists the d10, d50, and d90 values of the particle size distributions, which respectively refer to the particle size below which 10%, 50% and 90% of the sample lies. From the particle size distribution results, Pumice-N, with a median particle size (d50) of approximately 3 mm, was observed to have the finest particle size distribution among all the SCMs/filler materials tested. Pumice-D was slightly coarser than the quartz, with a d50 value of approximately 13 mm. Finally, Pumice-S was observed to have an even coarser distribution than Pumice-D, with its d50 value being approximately 20 mm. Since Pumice-S was originally a fine aggregate that was ground in the laboratory, it was expected that it would have a coarser distribution than both Pumice-N and PumiceD, which are commercially marketed as SCMs. The surface areas of the pumices and quartz filler were measured using an ASAP 2020 Micromeritics Surface Analyzer with nitrogen gas and are presented in Table 3. Due to the porous nature of the pumice pozzolan [5] that gives rise to internal surface area, Pumice-D had a much higher surface area than the quartz, despite having a similar particle size distribution. The control sample was a mixture that had 100% portland cement (OPC) as its binding material. The other mixtures had 20% of the cement content by weight replaced with either pumice or the quartz filler. The isothermal calorimetry and thermal gravimetric tests used pastes with a water to cementitious (and/or filler) material ratio (w/cm) of 0.45 and 50 g of cementitious (and/or filler) material. Prior to each test, the pastes were mixed by hand for 2 min. For the calorimetry tests, approximately 10 g of paste was
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
137
Table 1 ASTM C 618 results for pumices. ASTM C 618 Criteria
Pumice-N
Pumice- D
Pumice-S
ASTM Requirements
SiO2þAl2O3þFe2O3 (wt %) SO3 (wt %) Moisture Content (wt %) LOI (wt %) Fineness (wt% retained#325 sieve) Strength Activity Index - 7 day Strength Activity Index - 28 day Water Requirement (% Flow) Autoclave Expansion (%)
83 0.05 2.2 4.0 0 99 119 107 0.01
83 0.04 1.5 4.4 2 82 93 104 0.01
82 0.05 0.7 4.4 19 79 83 103 0.02
70% min 4.0% max 3.0% max 10.0% max 34% max 75% min 75% min 115% max ±0.8% max
Fig. 2. Particle size distributions of pumices and quartz. Fig. 1. X-ray diffractograms of pumices. Table 3 Particle sizes and surface areas of pumices and quartz.
Table 2 XRF oxide compositions of the cementitious materials. Oxide Composition
Pumice-N
Pumice-D
Pumice-S
Cement
SiO2 (%) Al2O3 (%) Fe2O3 (%) CaO (%) MgO (%) SO3 (%) Na2O (%) K2O (%)
69 13 1.0 0.96 0.45 0.05 4.0 4.9
69 12 1.1 0.94 0.44 0.04 3.8 5.2
69 12 1.2 1.4 0.33 0.05 5.1 4.4
19 5.2 2.5 63 1.1 3.2 0.12 0.91
added to vials after mixing, and the heat from hydration was measured at 23 C for 72 h using a Thermometric TAM Air Isothermal Calorimeter, using reference ampoules containing water with the same specific heat of the paste samples. Tests were replicated at least twice and the data shown are from one representative test. For the TGA tests, pastes were cured at room temperature and 100% relative humidity after mixing until they reached the desired test age, which was 7, 28 or 90 days. After curing, the samples were weighed and then broken into small chunks (approximately 10 mm) that were stored under vacuum for 14 ± 1 days. The samples were reweighed, and the change in weight was recorded as the amount of water lost on drying. The samples were then crushed and sieved through a No. 325 sieve (45 mm opening) to ensure uniformity during testing. The sieved samples were stored under vacuum in a desiccator until they were tested on a Mettler Thermogravimetric Analyzer, Model TGA/DSC 1. During the test, the chamber gas used was nitrogen (at a flow rate of 50 ml/ s) and the samples were contained in alumina crucibles with lids.
Material Properties
Pumice-N
Pumice-D
Pumice-S
Quartz
d10 (mm) d50 (mm) d90 (mm) BET Surface Area (m2/g)
2.03 3.82 8.12 10.9
3.97 13.2 69.3 7.58
4.57 20.4 118 1.57
2.21 12.3 57.9 1.26
The weight losses of the samples were recorded as they were heated from 40 C to 1000 C, at a rate of 20 C per minute, and were used to plot the TGA curves. The heat flow during this interval was recorded as well and was used to plot the differential scanning calorimetry (DSC) curve. The DSC curve was used to pinpoint the exact temperatures between which the calcium hydroxide in the paste decomposed. Using those temperatures, the weight loss due to calcium hydroxide decomposition was calculated from the TGA curve for 2e3 specimens per sample type. Finally, using molecular weights and the recorded weight change from water loss in the desiccator, the weight loss from calcium hydroxide decomposition was converted to the calcium hydroxide content per gram of cement in the initial paste. For testing compressive strength, mortar cubes were prepared according to ASTM C 109 [19]. However, instead of using a w/cm that would result in a flow of 110 ± 5, as specified by the standard, the w/cm of the mortar cubes used in this study was fixed at 0.5. Standard graded sand from USA, which met the requirements of ASTM C 778 [20], was used for the compressive strength mortar mixtures. The average compressive strength was calculated using three mortar cube specimens from two separate batches that were made consecutively, on the same day. In certain cases, if the range
138
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
of the compressive strength data from the three mortar cubes was greater than 8.7% of the average, the result that deviated most from the average was discarded, as instructed by ASTM C 109 [19]. The average was then calculated from two mortar specimens instead of three, as required by ASTM C 109 [19]. Resistance to ASR was measured according to the guidelines stated in ASTM C 1260 [21] and ASTM C 1567 [22]. A mixed quartz and chert sand from USA, which has been shown in previous literature to cause deleterious expansions in concrete [23], was used in the mortar mixes that tested resistance to ASR. This reactive fine aggregate was re-graded in the laboratory to meet the requirements of ASTM C 1260 [21] and ASTM C 1567 [22]. Mixture workability was evaluated by measuring the plastic viscosity and the yield stress of cementitious pastes that had an identical mixture design to those used for the calorimetry and TGA tests. The rheology measurements on the pastes were performed using an MCR 301 Anton Paar Rotational Rheometer that was fitted with a cup and bob measurement system with a 1.0 mm gap. Before the test, the pastes were mixed mechanically for 2 min using a Caframo Compact Digital BDC 2002 overhead stirrer at 1000 rpm. The test was performed on approximately 19 mL of the mixed sample, which was pre-sheared for 4 min at a shear rate of 50 s1. Pre-shearing was done to reduce the effects of shear history and ensure a similar starting point across all tests. The pre-shearing was followed by 30 s of rest, after which the shear rate was gradually increased from 10 s1 to 50 s1 and then decreased from 50 s1 to 10 s1, in increments of 10 s1. After each increment, the shear rate was held constant for 3 min to ensure that equilibrium had been reached. Data from the equilibrium range of each shear rate were used to plot a rheological flow curve, from which the plastic viscosity and yield stress of the mixture were found by fitting the Bingham model [24,25]. The tests were performed at least twice per mixture type. 3. Results and discussion 3.1. Cement hydration kinetics Fig. 3 shows representative rate of heat curves of the paste mixtures during hydration, normalized to the amount of cement in the paste. The curves were plotted from the results of isothermal calorimetry, which was conducted to understand how pumice particle size influenced cement hydration kinetics. The first hour of
Fig. 3. Rate of heat released from control, pumice, and quartz-containing pastes during hydration.
data was omitted to allow for equilibration of the sample to the calorimeter temperature. Since Table 2 shows the oxide compositions of all three pumices to be almost identical, any observed differences in the rate of heat curves of the pumice-containing pastes were attributed to the differences in pumice particle size. Fig. 3 shows the rate of heat curve for the paste with Pumice-N to display characteristics that indicate size filler effects [11,12,26], including a significant leftward shift of the curve relative to the control, a higher maximum rate of heat evolution, and a higher aluminate peak than the silicate peak. This supports our hypothesis that, similar to non-porous materials, porous materials like pumice can also enhance the nucleation and growth of hydration products and accelerate cement hydration kinetics when a fine particle size distribution is used. While the curve for the paste containing Pumice-D, a coarser material than Pumice-N, also showed a leftward shift and a maximum rate of heat evolution higher than the control, both the shift and the maximum rate were less than that of the paste with Pumice-N. Furthermore, unlike the Pumice-N-containing paste, the aluminate and silicate peaks for the Pumice-D-containing paste had similar heights. This indicates that, similar to a non-porous material, the enhanced nucleation and growth of hydration products decreases as pumice particle size increases. The curve for the paste containing Pumice-S, the coarsest pumice, did not differ from the control significantly in terms of either position or shape, which suggests that a certain critical level of fineness is needed for size filler effects to occur. The curve for the inert quartz-containing paste also did not differ significantly from the control, meaning that size filler effects from the quartz were minimal. This is interesting because the paste with Pumice-D, a material which has a similar particle size distribution to quartz but a higher surface area due to porosity, clearly showed size filler effects. It is possible that, given a similar particle size distribution, a porous material can enhance nucleation and growth of hydration products more than a non-porous material, due to the increased availability of surface area from the porous material. More research is needed to understand how much of the pumice's internal surface area from porosity is available as nucleation and growth sites for hydration products and whether differences other than porosity, such as having a favorable surface for calcium silicate hydrate (CeSeH) precipitation, or having higher shearing between particles due to interparticle distance [27], can account for the variation in size filler effects between the Pumice-D and quartz paste. Space filler effects, which refer to the growth of additional hydration products from increased space per grain of cement, can also be discerned from the rate of heat curves. However, unlike size filler effects that are observed early, space filler effects are more apparent during the later stages of hydration, when the process is dominated by the impingement of hydration products [28]. Previous literature [12] has characterized space filler effects by an elongation of the rate of heat curve with respect to time during the deceleratory period. From Fig. 3, it can be observed that the rate of heat curves for the pumice and quartz pastes are more elongated than the control, which suggests that these pastes exhibit space filler effects. This is not surprising, as additional space per grain of cement often arise when SCMs or filler materials are used, since these materials typically do not produce hydrates during the first few days of hydration [12]. As expected, all of the pumices and the quartz reduced the cumulative heat of hydration of the pastes because they dilute the cement content. The cumulative heat was calculated over the 24 h period measured, starting from the minimum from each curve shown in Fig. 3. The control sample produced the most heat, 194 J/g cementitious material. The Pumice-N containing paste had a
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
cumulative heat of 177 J/g cementitious material, while the other pumices and quartz had cumulative heats of 164e168 J/g. The higher value for the Pumice-N containing paste compared to the other pumices and quartz is consistent with the increased size filler effects shown by Pumice-N, as discussed earlier. 3.2. Pozzolanic reactivity
139
even further to about 45% lower than that of the control. On the other hand, the paste with Pumice-D did not show a significant reduction in calcium hydroxide content until 90 days, when the amount of calcium hydroxide was seen to be 18% lower than that of the control. Even at 90 days, the Pumice-S-containing paste did not show any significant reduction of calcium hydroxide content, relative to that of the control. Since the pumices are almost identical in composition, the differences in calcium hydroxide content can be attributed to particle size. This is interesting because it shows that effects of particle size play a role even during the later stages of hydration, and that the rate of pozzolanic reactivity can be improved through a decrease in particle size. Similar to observations in previous literature [26], Fig. 4 shows the calcium hydroxide content of the quartz-containing paste to be similar to that of the control at 7 days, and to gradually increase over time due to space filler effects. Since the pumice-containing pastes also exhibit space filler effects, comparing the calcium hydroxide content of the pastes with pumice to that of the paste with quartz provides a better idea of the total calcium hydroxide reduction than when the pumice-containing pastes are compared to the control. The 90-day calcium hydroxide contents of the pastes with Pumice-N, Pumice-D and Pumice-S were lower than that of the paste with quartz by 50%, 26% and 12%, respectively, consistent with their particle sizes.
Fig. 4 presents the calcium hydroxide content of the pastes over time, calculated using the mass loss results from TGA. The calcium hydroxide contents in Fig. 4 have been normalized to the amount of cement in the mixture, with the error bars representing the range of the data. Calcium hydroxide contents of the paste were measured to see whether decreasing the pumice particle size could cause an increase in the rate of the pozzolanic reaction. Initially, the calcium hydroxide content per gram of cement of the pumicecontaining pastes was expected to be the same as the control or higher, to account for the additional growth of hydration products from size and/or space filler effects [11,12]. Over time, the calcium hydroxide content of the pumice-containing pastes was expected to decrease as the pozzolans react with the calcium hydroxide to form CeSeH. Since quartz is inert, the calcium hydroxide content of the quartz-containing paste was not expected to decrease over time. In fact, previous literature has indicated that, due to space filler effects, the calcium hydroxide content of quartz-containing pastes can actually increase over time, relative to the control [26]. It is somewhat surprising that, within error, the calcium hydroxide content of the control pastes does not increase after 7 days. This has been seen by other researchers for portland cement control pastes [26] and may be related to the conditions of the test method. From Fig. 4, it can be seen that as early as 7 days, the paste with Pumice-N showed a 10% decrease in calcium hydroxide content relative to that of the control. On the other hand, the paste with Pumice-D, which has similar chemical composition to Pumice-N but a coarser particle size distribution, did not show a decrease in calcium hydroxide content at 7 days. This illustrates that, although pozzolanic reaction is considered to be negligible in the first few days of hydration [12], it is possible to increase the rate of pozzolanic reaction by grinding pozzolans to a smaller particle size. Even after 7 days, the rate of calcium hydroxide depletion in the PumiceN-containing paste continued to be faster than the two other pumice-containing pastes. At 28 days, the calcium hydroxide content in the paste with Pumice-N was about 24% lower than that of the control. By 90 days, the calcium hydroxide content dropped
The compressive strength results were used to evaluate how the changes to cement hydration kinetics and rate of pozzolanic reaction from the use of a smaller pumice particle size influenced the rate of strength gain for the pumice-containing mortars. Since the three pumices are almost identical in composition, any differences in the compressive strength of their mortars can be attributed to the differences in pumice particle size. Fig. 5 shows the average compressive strength of mortars, with the error bars representing the range of the data. From Fig. 5, it can be observed that all three pumice-containing mortars have a lower compressive strength than the control at 1 day. Since the rate of heat curves in Fig. 3 did not show any retardation of cement hydration kinetics from the addition of pumice, the lower 1-day strength of the mortars with pumice must be a result of diluting the cement content with a slower reacting, pozzolanic material. Interestingly, the compressive strength results indicate that the dilutionary effect from replacing cement can be partially compensated by using a pumice with a
Fig. 4. Calcium hydroxide content of the control, pumice, and quartz-containing pastes.
Fig. 5. Average compressive strengths of the control, pumice, and quartz-containing mortars.
3.3. Compressive strength
140
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
smaller particle size, since the difference in strength between the control and the pumice-containing mortars becomes progressively smaller with the use of a finer pumice. The mortar with Pumice-N, the finest pumice, had a 1-day strength that was about 90% of the control, while the mortar strength with Pumice-D, which was coarser than Pumice-N, was approximately 75% of the control. Finally, the mortar with Pumice-S, which was the coarsest pumice, had a 1-day strength that was 65% of the control. Since the pozzolanic reactions are considered to be negligible during the first day [12], the higher strength of the Pumice-N-containing mortar compared to those with Pumice-D and Pumice-S must be due to the fine particle size distribution of Pumice-N that provides more area for the nucleation and growth of hydration products compared to the area provided by the two coarser pumices. Since Pumice-D is finer than Pumice-S, the higher 1-day strength of the Pumice-Dcontaining mortar relative to the Pumice-S-containing mortar can also be similarly explained. This is corroborated by the rate of heat curves from isothermal calorimetry (Fig. 3) that show the enhanced nucleation and growth of hydration products to increase as pumice particle size decreases, with the paste with Pumice-S, the coarsest pumice, displaying no size filler effects and the paste with PumiceN, the finest pumice, to exhibit the most pronounced size filler effects. The rapid strength gain of the Pumice-N-containing mortar continued even after 1 day, with its strength reaching that of the control by 7 days. By 28 days, it surpassed the strength of the control by 40%. In fact, from 28 days up to one year, the Pumice-Ncontaining mortar had the highest strength out of all the mortars tested. The rates of strength gain for mortars with Pumice-D and Pumice-S were comparatively slower, with the Pumice-Dcontaining mortar reaching a strength similar to the control at 28 days, and the mortar with Pumice-S reaching an equivalent strength to the control at 90 days. In other words, the compressive strength results showed a clear trend of increasing rate of strength gain with decreasing particle size. However, it must be noted that the benefits of a smaller particle size seem to diminish with time. For example, from Fig. 5, it can be seen that the difference in strength between the Pumice-N-containing mortar and the Pumice-D-containing mortar is lower at 1 year than their difference in strength at 90 days. This suggests that although a smaller particle size enhances the rate of strength development initially, eventually the long term strength is dependent on the composition or the total amount of reactive material in an SCM. While the higher strength of the Pumice-N-containing mortar during the initial stages of hydration was a result of size filler effects, during the later stages, when hydration is no longer dominated by the availability of nucleation and growth sites, the faster rate of strength gain was most likely due to the increase in the rate of pozzolanic reaction from the use of a smaller pumice particle size. This is corroborated by the TGA results of the paste with Pumice-N, which showed a 10% decrease in calcium hydroxide content relative to the control at an age as early as 7 days, while it took up to 90 days for the calcium hydroxide content of the paste with Pumice-D, a compositionally identical material with a coarser particle size distribution, to show a significant decrease compared to the control. 3.4. Resistance to alkali silica reaction (ASR) ASR resistance of pumice-containing mortar bars was evaluated to see if particle size could change the effectiveness of the pumice pozzolans at mitigating expansions due to ASR. Fig. 6 shows the amount of expansion experienced by the mortar bars when testing them for resistance to ASR using ASTM C 1567 [22]. The error bars in Fig. 6 represent the range of the expansion data. Please note, in
Fig. 6. Expansion of control and pumice-containing mortar bars in ASTM C 1567 ASR test.
some cases the error bars cannot be seen because they are smaller than the point marker. From Fig. 6, we see that at a 20% replacement level by weight, the expansion of all three pumice-containing mortars are significantly below the limit of ASTM C 1567 [22], which states that specimens having less than 0.10% expansion, after 14 days of being submerged in 1N sodium hydroxide solution at 80 C, will be unlikely to undergo deleterious expansions from ASR in the field. While the expansion results show that all three pumices are effective at mitigating expansions from ASR, they do not show any clear trend with respect to particle size. Although the mortar containing the coarsest pumice, Pumice-S, had the highest expansion out of all the pumice-containing mortars, the particle size difference between the Pumice-D and Pumice-N did not result in any difference in performance, with both of their mortar expansions being close to 0%. It could be the case that, for ASR mitigation, after a certain level of fineness, a smaller pumice particle size does not yield any additional benefits at the same SCM replacement dosage.
3.5. Rheological properties The effect of pumice particle size on mixture workability was evaluated by measuring the rheological properties of pastes containing pumice. The level of workability needed from a cementitious mixture is often dependent on the application. Since the focus of the research presented here was on the effect of particle size on general performance, rather than on performance in a particular application, the desired workability of the pumice-containing pastes was for them to have a similar or lower plastic viscosity as that of the control paste. If an increased viscosity is observed compared to the control, this suggests that additional effort would be needed for the mixture to achieve a similar flow condition as the control paste (i.e., increasing water content, adding a waterreducer, optimizing aggregate gradation, etc.). Fig. 7 presents representative rheological flow curves, which show how the shear stress in the pastes changed as the shear rate was decreased from 50 s1 to 10 s1. The flow curves in Fig. 7 show strong linear behavior with coefficients of determination (R2) above 0.98. As such, the Bingham model, which defines the slope of the linear trend line as plastic viscosity and the y-axis intercept as yield stress, was used to evaluate the rheological properties of all pastes [24,25]. The average Bingham plastic viscosities and yield stress from duplicate rheology tests are shown in Table 4, with the error
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
141
pastes had similar results, where the plastic viscosity of fly ash pastes increased with the fineness of the fly ash particles used [30]. 4. Conclusions
Fig. 7. Representative rheological flow curves for the control and pumice-containing pastes.
Table 4 Rheological properties of the control and pumice-containing pastes. Type of Paste
Average Plastic Viscosity (Pa$s)
Control 20% Pumice-N 20% Pumice-D 20% Pumice-S
0.57 1.31 0.73 0.53
± ± ± ±
0.01 0.03 0.03 0.02
Average Yield Stress (Pa) 7.8 ± 0.6 12.5 ± 0.5 10.7 ± 0.4 13.2 ± 0.1
representing the range of the data. The yield stress of a material is defined to be the shear stress that must be exceeded before the material can start to flow [24]. From Table 4, it can be observed that while all the pumicecontaining pastes had higher yield stresses than the control paste, there was no clear trend with respect to particle size. While prior studies have not looked at the rheological properties of blended pastes with pumice, previous literature on cement pastes suggest that a material exhibits yield stress because of attractive forces between particles that allow the material to support a certain amount of stress without flowing [29]. As such, yield stress is perhaps more dependent on the composition of the material rather than the particle size. Past rheological research has linked low yield stress in fly ash pastes to particularly weak attractive forces between the fly ash particles, and between the fly ash and cement particles, compared to stronger forces between the cement particles [30]. Although, more research needs to be done in this area, perhaps the higher yield stress of the pumice-containing pastes is because the attractive forces between the pumice particles, or between the pumice and cement particles, are stronger than the forces between the cement particles themselves. The plastic viscosity of a material can be defined as a material's internal resistance to deformation [25]. The results in Table 4 show that the plastic viscosity of the pumice mixtures increased with decreasing particle size. While the plastic viscosity of the paste with Pumice-S, the coarsest pumice, was similar to the control, the plastic viscosity of the paste with Pumice-N, the finest pumice, was found to be significantly higher than that of the control. This is not surprising, as previous rheology literature on cement pastes suggests that plastic viscosity is dependent upon interparticle friction that can be changed by either particle spacing or surface contact [31], both of which are dependent on material fineness or particle size. While previous studies have not measured the viscosity of blended pastes with pumice, previous rheology research on fly ash
The main motivation for this study was to understand whether the particle size of pumice, which is a porous material, influenced its performance as an SCM in terms of effects on the following: cement hydration kinetics, pozzolanic reactivity, compressive strength in mortars, resistance to ASR, and rheology. Results from isothermal calorimetry showed that, similar to non-porous materials, a finer pumice particle size also led to enhanced nucleation and growth of hydration products, which partially compensated for the dilutionary effect of replacing cement with a slower reacting, pozzolanic material. Consequently, this led to increased early age compressive strength of the pumice-containing mortars. The TGA results demonstrated that a smaller pumice particle size could also increase the rate of pozzolanic reaction as measured through calcium hydroxide consumption in pastes, which decreased the amount of time needed for the pumice-containing mortars to reach an equivalent strength to the control. It must be noted that, despite the initial strength improvement, the benefits to compressive strength from using a smaller pumice particle size seemed to diminish with time, so that the differences in strength between materials decrease with time. Similar to long term compressive strength, the ability of pumice to mitigate expansions from ASR seemed to depend little on particle size. Finally, although the higher yield stress of pumice-containing pastes seemed independent of particle size, a clear trend of increasing plastic viscosity with decreasing pumice particle size was observed. This indicates that while the use of smaller pumice particle size is beneficial in terms of enhancing cement hydration kinetics, pozzolanic reaction rate and rate of strength gain, it will come at the cost of increasing mixture viscosity and potentially lowering mixture workability. Acknowledgements The authors wish to thank the Texas Department of Transportation for funding this project through project 0-6717. The authors also wish to acknowledge the help of Cliff Coward, Victoria Valdez and Juan Pablo Gevaudan for their help with the data collection. References [1] M. Tapan, T. Depci, A. Ozvan, T. Efe, V. Oyan, Effect of physical, chemical and electro-kinetic properties of pumice on strength development of pumice blended cements, Mater. Struct. 46 (2013) 1695e1706. [2] H.S. Gokce, O. Simsek, S. Korkmaz, Reduction of alkali-silica reaction expansion of mortars by utilization of pozzolans, Mag. Concr. Res. 65 (2013) 441e447. [3] ACI Committee 232.1R, Report on the Use of Raw or Processed Natural Pozzolans in Concrete, American Concrete Institute, Farmington Hills, MI, 2012. [4] K.M.A. Hossain, S. Ahmed, M. Lachemi, Lightweight concrete incorporating pumice based blended cement and aggregate: mechanical and durability characteristics, Constr. Build. Mater. 25 (2011) 1186e1195. [5] G.C. Presley, Pumice, pumicite, and volcanic cinder, in: J.E. Kogel, N.C. Trivedi, J.M. Barker, S.T. Krukowski (Eds.), Industrial Minerals and Rocks, seventh ed., Society for Mining, Metallurgy, and Exploration, Inc, Littleton, CO, 2006, pp. 743e754. [6] R.J. Elfert, Bureau of Reclamation Experiences with Fly Ash and Other Pozzolans in Concrete, Information Circular No. 8640, U.S. Bureau of Mines, Washington D.C, 1974, pp. 80e93. [7] K.M.A. Hossain, Potential use of volcanic pumice as a construction material, ASCE J. Mater. Civ. Eng. 16 (2004) 573e577. [8] K.M.A. Hossain, Chloride induced corrosion of reinforcement in volcanic ash and pumice based blended concrete, Cem. Concr. Compos. 27 (2005) 381e390. [9] K.M.A. Hossain, M. Lachemi, Performance of volcanic ash and pumice based blended cement concrete in mixed sulfate environment, Cem. Concr. Res. 36 (2006) 1123e1133.
142
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
[10] M.C.G. Juenger, R. Siddique, Recent advances in understanding the role of supplementary cementitious materials in concrete, Cem. Concr. Res. 78 (2015) 71e80. [11] T. Oey, A. Kumar, J.W. Bullard, N. Neithalath, G. Sant, The filler effect: the influence of filler content and surface area on cementitious reaction rates, J. Am. Ceram. Soc. 96 (2013) 1978e1990. [12] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials, Cem. Concr. Res. 41 (2011) 1244e1256. [13] J. Camiletti, A.M. Soliman, M.L. Nehdi, Effects of nano- and micro-limestone addition on early-age properties of ultra-high-performance concrete, Mater. Struct. 46 (2013) 881e898. [14] K.D. Weerdt, M.B. Haha, G.L. Saout, K.O. Kjellsen, H. Justnes, B. Lothenbach, The effect of temperature on the hydration of composite cements containing limestone powder and fly ash, Mater. Struct. 45 (2012) 1101e1114. [15] R.C. Mielenz, L.P. Witte, O.J. Glantz, Effect of Calcination on Natural Pozzolana, Symposium on Use of Pozzolanic Materials in Mortars and Concretes, STP-99, American Society for Testing and Materials, 1950, pp. 43e91. [16] ASTM C 618, Standard Specification for Coal Fly Ash and Raw or Calcined Natural Pozzolan for Use in Concrete, ASTM International, West Conshohocken, PA, 2012. [17] ASTM D 4326, Standard Test Method for Major and Minor Elements in Coal and Coke Ash by X-Ray Fluorescence, ASTM International, West Conshohocken, PA, 2011. [18] ASTM C 150, Standard Specification for Portland Cement, ASTM International, West Conshohocken, PA, 2011. [19] ASTM C 109, Standard Test Method for Compressive Strength of Hydraulic Cement Mortars (Using 2-in. Or [50-mm] Cube Specimens), ASTM International, West Conshohocken, PA, 2012. [20] ASTM C 778, Standard Specification for Standard Sand, ASTM International, West Conshohocken, PA, 2012. [21] ASTM C 1260, Standard Test Method for Potential Alkali Reactivity of
[22]
[23]
[24]
[25] [26]
[27] [28]
[29]
[30]
[31]
Aggregates (Mortar-bar Method), ASTM International, West Conshohocken, PA, 2007. ASTM C 1567, Standard Test Method for Determining the Potential Alkalisilica Reactivity of Combinations of Cementitious Materials and Aggregate (Accelerated Mortar-bar Method), ASTM International, West Conshohocken, PA, 2013. J.H. Ideker, A.F. Bentivegna, K.J. Folliard, M.C.G. Juenger, Do current laboratory test methods accurately predict alkali-silica reactivity, ACI Mater. J. 109 (2012) 395e402. V. Mechtcherine, A. Gram, K. Krenzer, J.H. Schwabe, S. Shyshko, N. Roussel, Simulation of fresh concrete flow using Discrete Element Method (DEM): theory and applications, Mater. Struct. 47 (2014) 615e630. G.H. Tatersall, P.F.G. Banfill, The Rheology of Fresh Concrete, first ed., Pitman Books Limited, London, 1983. M. Antoni, J. Rossen, F. Martirena, K. Scrivener, Cement substitution by a combination of metakaolin and limestone, Cem. Concr. Res. 42 (2012) 1579e1589. E. Berodier, K. Scrivener, Understanding the filler effect on the nucleation and growth of C-S-H, J. Am. Ceram. Soc. 97 (2014) 3764e4773. J.W. Bullard, H.M. Jennings, R.A. Livingston, A. Nonat, G.W. Scherer, J.S. Schweitzer, K.L. Scrivener, K.L, J.J. Thomas, Mechanisms of cement hydration, Cem. Concr. Res. 41 (2011) 1208e1223. N. Roussel, A. Lemaitre, R.J. Flatt, P. Coussot, Steady state flow of cement suspensions: a micromechanical state of the art, Cem. Concr. Res. 40 (2010) 77e84. D.P. Bentz, C.F. Ferraris, M.A. Galler, A.S. Hansen, J.M. Guynn, Influence of particle size distributions on yield stress and viscosity of cementefly ash pastes, Cem. Concr. Res. 42 (2012) 404e409. K. Vance, G. Sant, N. Neithalath, The rheology of cementitious suspensions: a closer look at experimental parameters and property determination using common rheological models, Cem. Concr. Compos. 59 (2015) 38e48.
Contents lists available at ScienceDirect
Cement and Concrete Composites journal homepage: www.elsevier.com/locate/cemconcomp
The role of particle size on the performance of pumice as a supplementary cementitious material Saamiya Seraj, Rachel Cano, Raissa D. Ferron, Maria C.G. Juenger* The University of Texas at Austin, 301 E. Dean Keeton St., Austin, TX 78712, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 September 2016 Received in revised form 9 February 2017 Accepted 16 March 2017 Available online 20 March 2017
A critical area overlooked in previous research on pumice is understanding how its physical characteristics influence its behavior as a supplementary cementitious material (SCM). This study investigated three pumices with different particle size distributions to observe whether these porous materials exhibit enhanced nucleation and growth of hydration products, in the same way as non-porous materials, and whether the rate of pozzolanic reaction can be changed through particle size. The effect of particle size on compressive strength, rheology and resistance to alkali silica reaction (ASR) was also evaluated. Results showed that reducing particle size increased the rates of cement hydration, pozzolanic reaction, and compressive strength gain, while also increasing mixture viscosity. Interestingly, particle size did not impact the yield stress of the mixture or the resistance to ASR. These new findings give insight about how the particle size of pumice can be used to overcome drawbacks reported in previous literature. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Pozzolan Particle size distribution Compressive strength Durability Rheology
1. Introduction The push towards more economical, sustainable and durable concrete mixtures has made supplementary cementitious materials (SCMs) very important in the concrete industry. However, at a time of increasing SCM demand, global changes in the power generation industry have led to regional supply shortages of fly ash, one of the most widely used SCMs. Consequently, there is increased interest in the cement and concrete community to evaluate other materials that can serve as suitable SCMs to overcome this scarcity. Pumice, a natural pozzolan, is one such promising material that has seen renewed interest in being used as an SCM [1e4]. Pumice is formed from highly silicic volcanic lava that tends to hold more dissolved gases than other types of lava due to its viscosity. As the lava rapidly cools and hardens into a glassy structure, the dissolved gases form pores or vesicles that result in low density volcanic rock [5]. The high silica content and amorphous structure are what give pumice its pozzolanic properties and make it a candidate for use as an SCM in concrete. In fact, pumice was used in concrete throughout history dating back to the ancient Greek and Roman civilizations [3]. Even
* Corresponding author. E-mail addresses: [email protected] (S. Seraj), [email protected] (R. Cano), [email protected] (R.D. Ferron), [email protected] (M.C.G. Juenger). http://dx.doi.org/10.1016/j.cemconcomp.2017.03.009 0958-9465/© 2017 Elsevier Ltd. All rights reserved.
as recent as the 20th century, finely ground pumice was used as an SCM in mass concrete structures, before the price and availability of fly ash made it more economical than pumice. Some notable examples of US structures with pumice are the Los Angeles aqueduct in 1912, the Friant Dam in 1942, the Altus Dam in 1945 and the Glen Canyon Dam in 1964 [3,6]. With the rising demand for SCMs, pumice is experiencing a renaissance in concrete research. Recent studies have shown finely ground pumice to meet most of the requirements listed in ASTM C 618 for natural pozzolans [4,7] and to lower calcium hydroxide content in cementitious mixtures, which is an indication of pozzolanic reaction [4,8,9]. Researchers have also evaluated the performance of pumice mixtures in terms of compressive strength, durability, and mixture workability [1e4,8,9]. However, a critical area of research that has been overlooked in previous literature is understanding how the characteristics of the pumice itself, like particle size and composition, influence its behavior in cementitious mixtures. For example, previous literature has indicated that replacing cement with finely ground pumice decreased the compressive strength of mortar and concrete specimens [1,4,7,9], but there has not been any investigation on how these effects on strength relate back to characteristics of the pumice itself, and whether those characteristics can be modified to improve the compressive strength of pumice blended mixtures. Similarly, for durability, although studies have found pumice to be effective at
136
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
controlling alkali silica reaction (ASR), results vary on the minimum percentage needed to mitigate expansions [2,7]. Since pumice is a natural pozzolan whose properties can vary depending upon its sourcing, these differences in reported results are not unusual. However, it is important to understand how these properties of pumice change its effectiveness in mitigating ASR so that minimum replacement dosages can be predicted for new sources. Evaluating the effects of pumice characteristics on mixture workability is also important, as fresh state properties ultimately affect hardened concrete properties. While there are some studies that have qualitatively measured workability of pumice concrete mixtures using the slump test [4,8,9], a more comprehensive analysis is needed to evaluate how the viscosity and yield stress of pumice blended mixtures change with characteristics of the pumice itself. The main objective of the study presented here was to address the shortcomings in understanding how physical properties of pumices influence their behavior in cementitious systems. We used three pumices with similar compositions but differing particle size distributions to evaluate the role of particle size in the performance of pumice as an SCM. In this way we could hold all material variables constant other than pumice particle size, so as to eliminate the overwhelming effects of parameters such as composition on pozzolanicity, particularly amorphous content and soluble silica and alumina contents [10]. The influence of pumice particle size was examined for its impact on the following: cement hydration kinetics, pozzolanic reactivity, compressive strength, resistance to ASR, and rheology. Based on previous research on other SCMs and filler materials [11,12], we hypothesized that in the early stages of hydration, when the process is dominated by the availability of nucleation and growth sites, pumice with a fine particle size distribution can increase the rate of cement hydration by providing an increased surface area for the nucleation and growth of hydration products. While this is simply referred to as “filler effect” in previous literature [11,12], in this paper it will be referred to as “size filler effect” to differentiate it from “space filler effect,” which is the growth of additional hydration products due to increased space per grain of cement when SCM or filler materials are used [11,12]. Much of the work examining size filler effects has been done with nonporous materials, and it is of interest to determine if the same principles apply for porous SCMs like pumice. Isothermal calorimetry was used to observe how pumice particle size can change cement hydration kinetics through size filler effects. Thermal gravimetric analysis (TGA) was used to measure calcium hydroxide consumption in hydrated pastes to see whether pumice particle size affected the rate of pozzolanic reaction. Compressive strengths of pumice-containing mortars were measured from 1 day to 1 year, to observe how the changes to early cement hydration kinetics and pozzolanic reaction rate correlate to the rate of strength gain. Based on previous research on fly ash and limestone [13,14], we hypothesized that the low early strengths of the pumice mixtures reported in literature [1,4,7,9] were most likely due to the dilutionary effect of replacing hydraulic cement with a slower reacting pozzolanic material like pumice. However, we expected that this dilutionary effect could be overcome, at least partially, to achieve higher early age strengths by using a finer pumice pozzolan that can provide increased surface area to enhance nucleation and growth of hydration products during the early stages of hydration. Further, other than a paper by Mielenz et al. [15], most studies report compressive strength data for pumice-containing mixtures only until 28 days. Therefore, we measured the strength of pumice-containing mortars for one year to provide insight into the role of particle size on the rate of strength gain at later ages and to evaluate the amount of time it takes to offset the dilutionary effect of replacing cement with pumice. ASR resistance of pumice-containing mortar bars was evaluated to see if particle size changed the effectiveness of the
pumice pozzolans at mitigating expansions due to ASR. Finally, the effect of pumice particle size on mixture workability was measured in terms of yield stress and plastic viscosity of pastes. Pumicecontaining mixtures were compared against mixtures with an inert quartz filler and control samples without SCMs or filler materials. 2. Materials and methods The three pumice powders used in this paper were sourced from USA and are referred to as Pumice-N, Pumice-D and Pumice-S. Pumice-N and Pumice-D are commercially available SCMs, while Pumice-S was originally a fine aggregate that was ground in the laboratory using a Bico Inc. UA V-Belt Drive Pulverizer and sieved through a No. 200 sieve (75 mm opening) for use as an SCM. All three pumices were tested to determine if they could be classified as Class N pozzolans according to the guidelines stated in ASTM C 618 [16]. The results, presented in Table 1, show that the physical and chemical characteristics of the pumices fulfilled all of the requirements for Class N pozzolans. X-ray diffraction (XRD) tests, conducted using a Siemens D-500 X-ray Diffractometer with Cu-Ka radiation and a dwell time of 4 s, showed the pumices to be mostly amorphous (Fig. 1). X-ray florescence (XRF) analyses of the pumices were conducted using a Bruker S4 Explorer according to the procedures of ASTM D 4326 [17]. The results, presented in Table 2, confirmed the oxide compositions of the pumices to be almost identical to each other. It was assumed, therefore, that the amorphous content and soluble silica and alumina contents of the pumices were nearly identical, so that the only difference between them was the gradation. An inert quartz filler, purchased from Old Hickory Clay Company, was also used in this study. The cement used in all mixtures was an ASTM C 150 [18] Type I portland cement sourced from USA. The XRF oxide composition of the cement, found using the same procedure as the pumices, is also presented in Table 2. The particle size distributions of the pumices and quartz filler were analyzed using a Horiba Partica LA 950-V2 Laser Scattering Particle Size Distribution Analyzer and are presented in Fig. 2. Table 3 lists the d10, d50, and d90 values of the particle size distributions, which respectively refer to the particle size below which 10%, 50% and 90% of the sample lies. From the particle size distribution results, Pumice-N, with a median particle size (d50) of approximately 3 mm, was observed to have the finest particle size distribution among all the SCMs/filler materials tested. Pumice-D was slightly coarser than the quartz, with a d50 value of approximately 13 mm. Finally, Pumice-S was observed to have an even coarser distribution than Pumice-D, with its d50 value being approximately 20 mm. Since Pumice-S was originally a fine aggregate that was ground in the laboratory, it was expected that it would have a coarser distribution than both Pumice-N and PumiceD, which are commercially marketed as SCMs. The surface areas of the pumices and quartz filler were measured using an ASAP 2020 Micromeritics Surface Analyzer with nitrogen gas and are presented in Table 3. Due to the porous nature of the pumice pozzolan [5] that gives rise to internal surface area, Pumice-D had a much higher surface area than the quartz, despite having a similar particle size distribution. The control sample was a mixture that had 100% portland cement (OPC) as its binding material. The other mixtures had 20% of the cement content by weight replaced with either pumice or the quartz filler. The isothermal calorimetry and thermal gravimetric tests used pastes with a water to cementitious (and/or filler) material ratio (w/cm) of 0.45 and 50 g of cementitious (and/or filler) material. Prior to each test, the pastes were mixed by hand for 2 min. For the calorimetry tests, approximately 10 g of paste was
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
137
Table 1 ASTM C 618 results for pumices. ASTM C 618 Criteria
Pumice-N
Pumice- D
Pumice-S
ASTM Requirements
SiO2þAl2O3þFe2O3 (wt %) SO3 (wt %) Moisture Content (wt %) LOI (wt %) Fineness (wt% retained#325 sieve) Strength Activity Index - 7 day Strength Activity Index - 28 day Water Requirement (% Flow) Autoclave Expansion (%)
83 0.05 2.2 4.0 0 99 119 107 0.01
83 0.04 1.5 4.4 2 82 93 104 0.01
82 0.05 0.7 4.4 19 79 83 103 0.02
70% min 4.0% max 3.0% max 10.0% max 34% max 75% min 75% min 115% max ±0.8% max
Fig. 2. Particle size distributions of pumices and quartz. Fig. 1. X-ray diffractograms of pumices. Table 3 Particle sizes and surface areas of pumices and quartz.
Table 2 XRF oxide compositions of the cementitious materials. Oxide Composition
Pumice-N
Pumice-D
Pumice-S
Cement
SiO2 (%) Al2O3 (%) Fe2O3 (%) CaO (%) MgO (%) SO3 (%) Na2O (%) K2O (%)
69 13 1.0 0.96 0.45 0.05 4.0 4.9
69 12 1.1 0.94 0.44 0.04 3.8 5.2
69 12 1.2 1.4 0.33 0.05 5.1 4.4
19 5.2 2.5 63 1.1 3.2 0.12 0.91
added to vials after mixing, and the heat from hydration was measured at 23 C for 72 h using a Thermometric TAM Air Isothermal Calorimeter, using reference ampoules containing water with the same specific heat of the paste samples. Tests were replicated at least twice and the data shown are from one representative test. For the TGA tests, pastes were cured at room temperature and 100% relative humidity after mixing until they reached the desired test age, which was 7, 28 or 90 days. After curing, the samples were weighed and then broken into small chunks (approximately 10 mm) that were stored under vacuum for 14 ± 1 days. The samples were reweighed, and the change in weight was recorded as the amount of water lost on drying. The samples were then crushed and sieved through a No. 325 sieve (45 mm opening) to ensure uniformity during testing. The sieved samples were stored under vacuum in a desiccator until they were tested on a Mettler Thermogravimetric Analyzer, Model TGA/DSC 1. During the test, the chamber gas used was nitrogen (at a flow rate of 50 ml/ s) and the samples were contained in alumina crucibles with lids.
Material Properties
Pumice-N
Pumice-D
Pumice-S
Quartz
d10 (mm) d50 (mm) d90 (mm) BET Surface Area (m2/g)
2.03 3.82 8.12 10.9
3.97 13.2 69.3 7.58
4.57 20.4 118 1.57
2.21 12.3 57.9 1.26
The weight losses of the samples were recorded as they were heated from 40 C to 1000 C, at a rate of 20 C per minute, and were used to plot the TGA curves. The heat flow during this interval was recorded as well and was used to plot the differential scanning calorimetry (DSC) curve. The DSC curve was used to pinpoint the exact temperatures between which the calcium hydroxide in the paste decomposed. Using those temperatures, the weight loss due to calcium hydroxide decomposition was calculated from the TGA curve for 2e3 specimens per sample type. Finally, using molecular weights and the recorded weight change from water loss in the desiccator, the weight loss from calcium hydroxide decomposition was converted to the calcium hydroxide content per gram of cement in the initial paste. For testing compressive strength, mortar cubes were prepared according to ASTM C 109 [19]. However, instead of using a w/cm that would result in a flow of 110 ± 5, as specified by the standard, the w/cm of the mortar cubes used in this study was fixed at 0.5. Standard graded sand from USA, which met the requirements of ASTM C 778 [20], was used for the compressive strength mortar mixtures. The average compressive strength was calculated using three mortar cube specimens from two separate batches that were made consecutively, on the same day. In certain cases, if the range
138
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
of the compressive strength data from the three mortar cubes was greater than 8.7% of the average, the result that deviated most from the average was discarded, as instructed by ASTM C 109 [19]. The average was then calculated from two mortar specimens instead of three, as required by ASTM C 109 [19]. Resistance to ASR was measured according to the guidelines stated in ASTM C 1260 [21] and ASTM C 1567 [22]. A mixed quartz and chert sand from USA, which has been shown in previous literature to cause deleterious expansions in concrete [23], was used in the mortar mixes that tested resistance to ASR. This reactive fine aggregate was re-graded in the laboratory to meet the requirements of ASTM C 1260 [21] and ASTM C 1567 [22]. Mixture workability was evaluated by measuring the plastic viscosity and the yield stress of cementitious pastes that had an identical mixture design to those used for the calorimetry and TGA tests. The rheology measurements on the pastes were performed using an MCR 301 Anton Paar Rotational Rheometer that was fitted with a cup and bob measurement system with a 1.0 mm gap. Before the test, the pastes were mixed mechanically for 2 min using a Caframo Compact Digital BDC 2002 overhead stirrer at 1000 rpm. The test was performed on approximately 19 mL of the mixed sample, which was pre-sheared for 4 min at a shear rate of 50 s1. Pre-shearing was done to reduce the effects of shear history and ensure a similar starting point across all tests. The pre-shearing was followed by 30 s of rest, after which the shear rate was gradually increased from 10 s1 to 50 s1 and then decreased from 50 s1 to 10 s1, in increments of 10 s1. After each increment, the shear rate was held constant for 3 min to ensure that equilibrium had been reached. Data from the equilibrium range of each shear rate were used to plot a rheological flow curve, from which the plastic viscosity and yield stress of the mixture were found by fitting the Bingham model [24,25]. The tests were performed at least twice per mixture type. 3. Results and discussion 3.1. Cement hydration kinetics Fig. 3 shows representative rate of heat curves of the paste mixtures during hydration, normalized to the amount of cement in the paste. The curves were plotted from the results of isothermal calorimetry, which was conducted to understand how pumice particle size influenced cement hydration kinetics. The first hour of
Fig. 3. Rate of heat released from control, pumice, and quartz-containing pastes during hydration.
data was omitted to allow for equilibration of the sample to the calorimeter temperature. Since Table 2 shows the oxide compositions of all three pumices to be almost identical, any observed differences in the rate of heat curves of the pumice-containing pastes were attributed to the differences in pumice particle size. Fig. 3 shows the rate of heat curve for the paste with Pumice-N to display characteristics that indicate size filler effects [11,12,26], including a significant leftward shift of the curve relative to the control, a higher maximum rate of heat evolution, and a higher aluminate peak than the silicate peak. This supports our hypothesis that, similar to non-porous materials, porous materials like pumice can also enhance the nucleation and growth of hydration products and accelerate cement hydration kinetics when a fine particle size distribution is used. While the curve for the paste containing Pumice-D, a coarser material than Pumice-N, also showed a leftward shift and a maximum rate of heat evolution higher than the control, both the shift and the maximum rate were less than that of the paste with Pumice-N. Furthermore, unlike the Pumice-N-containing paste, the aluminate and silicate peaks for the Pumice-D-containing paste had similar heights. This indicates that, similar to a non-porous material, the enhanced nucleation and growth of hydration products decreases as pumice particle size increases. The curve for the paste containing Pumice-S, the coarsest pumice, did not differ from the control significantly in terms of either position or shape, which suggests that a certain critical level of fineness is needed for size filler effects to occur. The curve for the inert quartz-containing paste also did not differ significantly from the control, meaning that size filler effects from the quartz were minimal. This is interesting because the paste with Pumice-D, a material which has a similar particle size distribution to quartz but a higher surface area due to porosity, clearly showed size filler effects. It is possible that, given a similar particle size distribution, a porous material can enhance nucleation and growth of hydration products more than a non-porous material, due to the increased availability of surface area from the porous material. More research is needed to understand how much of the pumice's internal surface area from porosity is available as nucleation and growth sites for hydration products and whether differences other than porosity, such as having a favorable surface for calcium silicate hydrate (CeSeH) precipitation, or having higher shearing between particles due to interparticle distance [27], can account for the variation in size filler effects between the Pumice-D and quartz paste. Space filler effects, which refer to the growth of additional hydration products from increased space per grain of cement, can also be discerned from the rate of heat curves. However, unlike size filler effects that are observed early, space filler effects are more apparent during the later stages of hydration, when the process is dominated by the impingement of hydration products [28]. Previous literature [12] has characterized space filler effects by an elongation of the rate of heat curve with respect to time during the deceleratory period. From Fig. 3, it can be observed that the rate of heat curves for the pumice and quartz pastes are more elongated than the control, which suggests that these pastes exhibit space filler effects. This is not surprising, as additional space per grain of cement often arise when SCMs or filler materials are used, since these materials typically do not produce hydrates during the first few days of hydration [12]. As expected, all of the pumices and the quartz reduced the cumulative heat of hydration of the pastes because they dilute the cement content. The cumulative heat was calculated over the 24 h period measured, starting from the minimum from each curve shown in Fig. 3. The control sample produced the most heat, 194 J/g cementitious material. The Pumice-N containing paste had a
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
cumulative heat of 177 J/g cementitious material, while the other pumices and quartz had cumulative heats of 164e168 J/g. The higher value for the Pumice-N containing paste compared to the other pumices and quartz is consistent with the increased size filler effects shown by Pumice-N, as discussed earlier. 3.2. Pozzolanic reactivity
139
even further to about 45% lower than that of the control. On the other hand, the paste with Pumice-D did not show a significant reduction in calcium hydroxide content until 90 days, when the amount of calcium hydroxide was seen to be 18% lower than that of the control. Even at 90 days, the Pumice-S-containing paste did not show any significant reduction of calcium hydroxide content, relative to that of the control. Since the pumices are almost identical in composition, the differences in calcium hydroxide content can be attributed to particle size. This is interesting because it shows that effects of particle size play a role even during the later stages of hydration, and that the rate of pozzolanic reactivity can be improved through a decrease in particle size. Similar to observations in previous literature [26], Fig. 4 shows the calcium hydroxide content of the quartz-containing paste to be similar to that of the control at 7 days, and to gradually increase over time due to space filler effects. Since the pumice-containing pastes also exhibit space filler effects, comparing the calcium hydroxide content of the pastes with pumice to that of the paste with quartz provides a better idea of the total calcium hydroxide reduction than when the pumice-containing pastes are compared to the control. The 90-day calcium hydroxide contents of the pastes with Pumice-N, Pumice-D and Pumice-S were lower than that of the paste with quartz by 50%, 26% and 12%, respectively, consistent with their particle sizes.
Fig. 4 presents the calcium hydroxide content of the pastes over time, calculated using the mass loss results from TGA. The calcium hydroxide contents in Fig. 4 have been normalized to the amount of cement in the mixture, with the error bars representing the range of the data. Calcium hydroxide contents of the paste were measured to see whether decreasing the pumice particle size could cause an increase in the rate of the pozzolanic reaction. Initially, the calcium hydroxide content per gram of cement of the pumicecontaining pastes was expected to be the same as the control or higher, to account for the additional growth of hydration products from size and/or space filler effects [11,12]. Over time, the calcium hydroxide content of the pumice-containing pastes was expected to decrease as the pozzolans react with the calcium hydroxide to form CeSeH. Since quartz is inert, the calcium hydroxide content of the quartz-containing paste was not expected to decrease over time. In fact, previous literature has indicated that, due to space filler effects, the calcium hydroxide content of quartz-containing pastes can actually increase over time, relative to the control [26]. It is somewhat surprising that, within error, the calcium hydroxide content of the control pastes does not increase after 7 days. This has been seen by other researchers for portland cement control pastes [26] and may be related to the conditions of the test method. From Fig. 4, it can be seen that as early as 7 days, the paste with Pumice-N showed a 10% decrease in calcium hydroxide content relative to that of the control. On the other hand, the paste with Pumice-D, which has similar chemical composition to Pumice-N but a coarser particle size distribution, did not show a decrease in calcium hydroxide content at 7 days. This illustrates that, although pozzolanic reaction is considered to be negligible in the first few days of hydration [12], it is possible to increase the rate of pozzolanic reaction by grinding pozzolans to a smaller particle size. Even after 7 days, the rate of calcium hydroxide depletion in the PumiceN-containing paste continued to be faster than the two other pumice-containing pastes. At 28 days, the calcium hydroxide content in the paste with Pumice-N was about 24% lower than that of the control. By 90 days, the calcium hydroxide content dropped
The compressive strength results were used to evaluate how the changes to cement hydration kinetics and rate of pozzolanic reaction from the use of a smaller pumice particle size influenced the rate of strength gain for the pumice-containing mortars. Since the three pumices are almost identical in composition, any differences in the compressive strength of their mortars can be attributed to the differences in pumice particle size. Fig. 5 shows the average compressive strength of mortars, with the error bars representing the range of the data. From Fig. 5, it can be observed that all three pumice-containing mortars have a lower compressive strength than the control at 1 day. Since the rate of heat curves in Fig. 3 did not show any retardation of cement hydration kinetics from the addition of pumice, the lower 1-day strength of the mortars with pumice must be a result of diluting the cement content with a slower reacting, pozzolanic material. Interestingly, the compressive strength results indicate that the dilutionary effect from replacing cement can be partially compensated by using a pumice with a
Fig. 4. Calcium hydroxide content of the control, pumice, and quartz-containing pastes.
Fig. 5. Average compressive strengths of the control, pumice, and quartz-containing mortars.
3.3. Compressive strength
140
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
smaller particle size, since the difference in strength between the control and the pumice-containing mortars becomes progressively smaller with the use of a finer pumice. The mortar with Pumice-N, the finest pumice, had a 1-day strength that was about 90% of the control, while the mortar strength with Pumice-D, which was coarser than Pumice-N, was approximately 75% of the control. Finally, the mortar with Pumice-S, which was the coarsest pumice, had a 1-day strength that was 65% of the control. Since the pozzolanic reactions are considered to be negligible during the first day [12], the higher strength of the Pumice-N-containing mortar compared to those with Pumice-D and Pumice-S must be due to the fine particle size distribution of Pumice-N that provides more area for the nucleation and growth of hydration products compared to the area provided by the two coarser pumices. Since Pumice-D is finer than Pumice-S, the higher 1-day strength of the Pumice-Dcontaining mortar relative to the Pumice-S-containing mortar can also be similarly explained. This is corroborated by the rate of heat curves from isothermal calorimetry (Fig. 3) that show the enhanced nucleation and growth of hydration products to increase as pumice particle size decreases, with the paste with Pumice-S, the coarsest pumice, displaying no size filler effects and the paste with PumiceN, the finest pumice, to exhibit the most pronounced size filler effects. The rapid strength gain of the Pumice-N-containing mortar continued even after 1 day, with its strength reaching that of the control by 7 days. By 28 days, it surpassed the strength of the control by 40%. In fact, from 28 days up to one year, the Pumice-Ncontaining mortar had the highest strength out of all the mortars tested. The rates of strength gain for mortars with Pumice-D and Pumice-S were comparatively slower, with the Pumice-Dcontaining mortar reaching a strength similar to the control at 28 days, and the mortar with Pumice-S reaching an equivalent strength to the control at 90 days. In other words, the compressive strength results showed a clear trend of increasing rate of strength gain with decreasing particle size. However, it must be noted that the benefits of a smaller particle size seem to diminish with time. For example, from Fig. 5, it can be seen that the difference in strength between the Pumice-N-containing mortar and the Pumice-D-containing mortar is lower at 1 year than their difference in strength at 90 days. This suggests that although a smaller particle size enhances the rate of strength development initially, eventually the long term strength is dependent on the composition or the total amount of reactive material in an SCM. While the higher strength of the Pumice-N-containing mortar during the initial stages of hydration was a result of size filler effects, during the later stages, when hydration is no longer dominated by the availability of nucleation and growth sites, the faster rate of strength gain was most likely due to the increase in the rate of pozzolanic reaction from the use of a smaller pumice particle size. This is corroborated by the TGA results of the paste with Pumice-N, which showed a 10% decrease in calcium hydroxide content relative to the control at an age as early as 7 days, while it took up to 90 days for the calcium hydroxide content of the paste with Pumice-D, a compositionally identical material with a coarser particle size distribution, to show a significant decrease compared to the control. 3.4. Resistance to alkali silica reaction (ASR) ASR resistance of pumice-containing mortar bars was evaluated to see if particle size could change the effectiveness of the pumice pozzolans at mitigating expansions due to ASR. Fig. 6 shows the amount of expansion experienced by the mortar bars when testing them for resistance to ASR using ASTM C 1567 [22]. The error bars in Fig. 6 represent the range of the expansion data. Please note, in
Fig. 6. Expansion of control and pumice-containing mortar bars in ASTM C 1567 ASR test.
some cases the error bars cannot be seen because they are smaller than the point marker. From Fig. 6, we see that at a 20% replacement level by weight, the expansion of all three pumice-containing mortars are significantly below the limit of ASTM C 1567 [22], which states that specimens having less than 0.10% expansion, after 14 days of being submerged in 1N sodium hydroxide solution at 80 C, will be unlikely to undergo deleterious expansions from ASR in the field. While the expansion results show that all three pumices are effective at mitigating expansions from ASR, they do not show any clear trend with respect to particle size. Although the mortar containing the coarsest pumice, Pumice-S, had the highest expansion out of all the pumice-containing mortars, the particle size difference between the Pumice-D and Pumice-N did not result in any difference in performance, with both of their mortar expansions being close to 0%. It could be the case that, for ASR mitigation, after a certain level of fineness, a smaller pumice particle size does not yield any additional benefits at the same SCM replacement dosage.
3.5. Rheological properties The effect of pumice particle size on mixture workability was evaluated by measuring the rheological properties of pastes containing pumice. The level of workability needed from a cementitious mixture is often dependent on the application. Since the focus of the research presented here was on the effect of particle size on general performance, rather than on performance in a particular application, the desired workability of the pumice-containing pastes was for them to have a similar or lower plastic viscosity as that of the control paste. If an increased viscosity is observed compared to the control, this suggests that additional effort would be needed for the mixture to achieve a similar flow condition as the control paste (i.e., increasing water content, adding a waterreducer, optimizing aggregate gradation, etc.). Fig. 7 presents representative rheological flow curves, which show how the shear stress in the pastes changed as the shear rate was decreased from 50 s1 to 10 s1. The flow curves in Fig. 7 show strong linear behavior with coefficients of determination (R2) above 0.98. As such, the Bingham model, which defines the slope of the linear trend line as plastic viscosity and the y-axis intercept as yield stress, was used to evaluate the rheological properties of all pastes [24,25]. The average Bingham plastic viscosities and yield stress from duplicate rheology tests are shown in Table 4, with the error
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
141
pastes had similar results, where the plastic viscosity of fly ash pastes increased with the fineness of the fly ash particles used [30]. 4. Conclusions
Fig. 7. Representative rheological flow curves for the control and pumice-containing pastes.
Table 4 Rheological properties of the control and pumice-containing pastes. Type of Paste
Average Plastic Viscosity (Pa$s)
Control 20% Pumice-N 20% Pumice-D 20% Pumice-S
0.57 1.31 0.73 0.53
± ± ± ±
0.01 0.03 0.03 0.02
Average Yield Stress (Pa) 7.8 ± 0.6 12.5 ± 0.5 10.7 ± 0.4 13.2 ± 0.1
representing the range of the data. The yield stress of a material is defined to be the shear stress that must be exceeded before the material can start to flow [24]. From Table 4, it can be observed that while all the pumicecontaining pastes had higher yield stresses than the control paste, there was no clear trend with respect to particle size. While prior studies have not looked at the rheological properties of blended pastes with pumice, previous literature on cement pastes suggest that a material exhibits yield stress because of attractive forces between particles that allow the material to support a certain amount of stress without flowing [29]. As such, yield stress is perhaps more dependent on the composition of the material rather than the particle size. Past rheological research has linked low yield stress in fly ash pastes to particularly weak attractive forces between the fly ash particles, and between the fly ash and cement particles, compared to stronger forces between the cement particles [30]. Although, more research needs to be done in this area, perhaps the higher yield stress of the pumice-containing pastes is because the attractive forces between the pumice particles, or between the pumice and cement particles, are stronger than the forces between the cement particles themselves. The plastic viscosity of a material can be defined as a material's internal resistance to deformation [25]. The results in Table 4 show that the plastic viscosity of the pumice mixtures increased with decreasing particle size. While the plastic viscosity of the paste with Pumice-S, the coarsest pumice, was similar to the control, the plastic viscosity of the paste with Pumice-N, the finest pumice, was found to be significantly higher than that of the control. This is not surprising, as previous rheology literature on cement pastes suggests that plastic viscosity is dependent upon interparticle friction that can be changed by either particle spacing or surface contact [31], both of which are dependent on material fineness or particle size. While previous studies have not measured the viscosity of blended pastes with pumice, previous rheology research on fly ash
The main motivation for this study was to understand whether the particle size of pumice, which is a porous material, influenced its performance as an SCM in terms of effects on the following: cement hydration kinetics, pozzolanic reactivity, compressive strength in mortars, resistance to ASR, and rheology. Results from isothermal calorimetry showed that, similar to non-porous materials, a finer pumice particle size also led to enhanced nucleation and growth of hydration products, which partially compensated for the dilutionary effect of replacing cement with a slower reacting, pozzolanic material. Consequently, this led to increased early age compressive strength of the pumice-containing mortars. The TGA results demonstrated that a smaller pumice particle size could also increase the rate of pozzolanic reaction as measured through calcium hydroxide consumption in pastes, which decreased the amount of time needed for the pumice-containing mortars to reach an equivalent strength to the control. It must be noted that, despite the initial strength improvement, the benefits to compressive strength from using a smaller pumice particle size seemed to diminish with time, so that the differences in strength between materials decrease with time. Similar to long term compressive strength, the ability of pumice to mitigate expansions from ASR seemed to depend little on particle size. Finally, although the higher yield stress of pumice-containing pastes seemed independent of particle size, a clear trend of increasing plastic viscosity with decreasing pumice particle size was observed. This indicates that while the use of smaller pumice particle size is beneficial in terms of enhancing cement hydration kinetics, pozzolanic reaction rate and rate of strength gain, it will come at the cost of increasing mixture viscosity and potentially lowering mixture workability. Acknowledgements The authors wish to thank the Texas Department of Transportation for funding this project through project 0-6717. The authors also wish to acknowledge the help of Cliff Coward, Victoria Valdez and Juan Pablo Gevaudan for their help with the data collection. References [1] M. Tapan, T. Depci, A. Ozvan, T. Efe, V. Oyan, Effect of physical, chemical and electro-kinetic properties of pumice on strength development of pumice blended cements, Mater. Struct. 46 (2013) 1695e1706. [2] H.S. Gokce, O. Simsek, S. Korkmaz, Reduction of alkali-silica reaction expansion of mortars by utilization of pozzolans, Mag. Concr. Res. 65 (2013) 441e447. [3] ACI Committee 232.1R, Report on the Use of Raw or Processed Natural Pozzolans in Concrete, American Concrete Institute, Farmington Hills, MI, 2012. [4] K.M.A. Hossain, S. Ahmed, M. Lachemi, Lightweight concrete incorporating pumice based blended cement and aggregate: mechanical and durability characteristics, Constr. Build. Mater. 25 (2011) 1186e1195. [5] G.C. Presley, Pumice, pumicite, and volcanic cinder, in: J.E. Kogel, N.C. Trivedi, J.M. Barker, S.T. Krukowski (Eds.), Industrial Minerals and Rocks, seventh ed., Society for Mining, Metallurgy, and Exploration, Inc, Littleton, CO, 2006, pp. 743e754. [6] R.J. Elfert, Bureau of Reclamation Experiences with Fly Ash and Other Pozzolans in Concrete, Information Circular No. 8640, U.S. Bureau of Mines, Washington D.C, 1974, pp. 80e93. [7] K.M.A. Hossain, Potential use of volcanic pumice as a construction material, ASCE J. Mater. Civ. Eng. 16 (2004) 573e577. [8] K.M.A. Hossain, Chloride induced corrosion of reinforcement in volcanic ash and pumice based blended concrete, Cem. Concr. Compos. 27 (2005) 381e390. [9] K.M.A. Hossain, M. Lachemi, Performance of volcanic ash and pumice based blended cement concrete in mixed sulfate environment, Cem. Concr. Res. 36 (2006) 1123e1133.
142
S. Seraj et al. / Cement and Concrete Composites 80 (2017) 135e142
[10] M.C.G. Juenger, R. Siddique, Recent advances in understanding the role of supplementary cementitious materials in concrete, Cem. Concr. Res. 78 (2015) 71e80. [11] T. Oey, A. Kumar, J.W. Bullard, N. Neithalath, G. Sant, The filler effect: the influence of filler content and surface area on cementitious reaction rates, J. Am. Ceram. Soc. 96 (2013) 1978e1990. [12] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials, Cem. Concr. Res. 41 (2011) 1244e1256. [13] J. Camiletti, A.M. Soliman, M.L. Nehdi, Effects of nano- and micro-limestone addition on early-age properties of ultra-high-performance concrete, Mater. Struct. 46 (2013) 881e898. [14] K.D. Weerdt, M.B. Haha, G.L. Saout, K.O. Kjellsen, H. Justnes, B. Lothenbach, The effect of temperature on the hydration of composite cements containing limestone powder and fly ash, Mater. Struct. 45 (2012) 1101e1114. [15] R.C. Mielenz, L.P. Witte, O.J. Glantz, Effect of Calcination on Natural Pozzolana, Symposium on Use of Pozzolanic Materials in Mortars and Concretes, STP-99, American Society for Testing and Materials, 1950, pp. 43e91. [16] ASTM C 618, Standard Specification for Coal Fly Ash and Raw or Calcined Natural Pozzolan for Use in Concrete, ASTM International, West Conshohocken, PA, 2012. [17] ASTM D 4326, Standard Test Method for Major and Minor Elements in Coal and Coke Ash by X-Ray Fluorescence, ASTM International, West Conshohocken, PA, 2011. [18] ASTM C 150, Standard Specification for Portland Cement, ASTM International, West Conshohocken, PA, 2011. [19] ASTM C 109, Standard Test Method for Compressive Strength of Hydraulic Cement Mortars (Using 2-in. Or [50-mm] Cube Specimens), ASTM International, West Conshohocken, PA, 2012. [20] ASTM C 778, Standard Specification for Standard Sand, ASTM International, West Conshohocken, PA, 2012. [21] ASTM C 1260, Standard Test Method for Potential Alkali Reactivity of
[22]
[23]
[24]
[25] [26]
[27] [28]
[29]
[30]
[31]
Aggregates (Mortar-bar Method), ASTM International, West Conshohocken, PA, 2007. ASTM C 1567, Standard Test Method for Determining the Potential Alkalisilica Reactivity of Combinations of Cementitious Materials and Aggregate (Accelerated Mortar-bar Method), ASTM International, West Conshohocken, PA, 2013. J.H. Ideker, A.F. Bentivegna, K.J. Folliard, M.C.G. Juenger, Do current laboratory test methods accurately predict alkali-silica reactivity, ACI Mater. J. 109 (2012) 395e402. V. Mechtcherine, A. Gram, K. Krenzer, J.H. Schwabe, S. Shyshko, N. Roussel, Simulation of fresh concrete flow using Discrete Element Method (DEM): theory and applications, Mater. Struct. 47 (2014) 615e630. G.H. Tatersall, P.F.G. Banfill, The Rheology of Fresh Concrete, first ed., Pitman Books Limited, London, 1983. M. Antoni, J. Rossen, F. Martirena, K. Scrivener, Cement substitution by a combination of metakaolin and limestone, Cem. Concr. Res. 42 (2012) 1579e1589. E. Berodier, K. Scrivener, Understanding the filler effect on the nucleation and growth of C-S-H, J. Am. Ceram. Soc. 97 (2014) 3764e4773. J.W. Bullard, H.M. Jennings, R.A. Livingston, A. Nonat, G.W. Scherer, J.S. Schweitzer, K.L. Scrivener, K.L, J.J. Thomas, Mechanisms of cement hydration, Cem. Concr. Res. 41 (2011) 1208e1223. N. Roussel, A. Lemaitre, R.J. Flatt, P. Coussot, Steady state flow of cement suspensions: a micromechanical state of the art, Cem. Concr. Res. 40 (2010) 77e84. D.P. Bentz, C.F. Ferraris, M.A. Galler, A.S. Hansen, J.M. Guynn, Influence of particle size distributions on yield stress and viscosity of cementefly ash pastes, Cem. Concr. Res. 42 (2012) 404e409. K. Vance, G. Sant, N. Neithalath, The rheology of cementitious suspensions: a closer look at experimental parameters and property determination using common rheological models, Cem. Concr. Compos. 59 (2015) 38e48.