Thermally Insulating, Thermal Shock Resistant Calcium Aluminate Phosphate Cement Composites for Reservoir Thermal Energy Storage

This paper presents the use of hydrophobic silica aerogel (HSA) and hydrophilic fly ash cenosphere (FCS) aggregates for improvements in the thermal insulating and mechanical properties of 100- and 250 °C-autoclaved calcium aluminate phosphate (CaP) cement composites reinforced with micro-glass (MGF) and micro-carbon (MCF) fibers for deployment in medium- (100 °C) and high-temperature (250 °C) reservoir thermal energy storage systems. The following six factors were assessed: (1) Hydrothermal stability of HSA; (2) Pozzolanic activity of the two aggregates and MGF in an alkali cement environment; (3) CaP cement slurry heat release during hydration and chemical reactions; (4) Composite phase compositions and phase transitions; (5) Mechanical behavior; (6) Thermal shock (TS) resistance at temperature gradients of 150 and 225 °C. The results showed that hydrophobic trimethylsilyl groups in trimethylsiloxy-linked silica aerogel structure were susceptible to hydrothermal degradation at 250 °C. This degradation was followed by pozzolanic reactions (PR) of HSA, its dissolution, and the formation of a porous microstructure that caused a major loss in the compressive strength of the composites at 250 °C. The pozzolanic activities of FCS and MGF were moderate, and they offered improved interfacial bonding at cement-FCS and cement-MGF joints through a bridging effect by PR products. Despite the PR of MGF, both MGF and MCF played an essential role in minimizing the considerable losses in compressive strength, particularly in toughness, engendered by incorporating weak HSA. As a result, a FCS/HSA ratio of 90/10 in the CaP composite system was identified as the most effective hybrid insulating aggregate composition, with a persistent compressive strength of more than 7 MPa after three TS tests at a 150 °C temperature gradient. This composite displayed thermal conductivity of 0.28 and 0.35 W/mK after TS with 225 and 150 °C thermal gradients, respectively. These values, below the TC of water (TC water = 0.6 W/mK), were measured under water-saturated conditions for applications in underground reservoirs. However, considering the hydrothermal disintegration of HSA at 250 °C, these CaP composites have potential applications for use in thermally insulating, thermal shock-resistant well cement in a mid-temperature range (100 to 175 °C) reservoir thermal energy storage system.


Introduction
This study is a continuation of efforts in designing cement composites with very low thermal conductivity (TC) under water-saturated conditions for use in reservoir energy storage and heat recovery wells [1]. For deep wells, significant energy savings are made possible if insulating cement is used for heat storage and recovery wells. The economic benefits of using insulating cement can amount to more than 75% heat loss reduction with a 50% decrease in cement thermal conductivity (TC) for a 3 km deep well [2]. Insulation can also decrease heat losses in borehole underground storage systems [3]. Research
To improve compressive toughness, which suppresses crack development and propagation, non-crystalline MGF were used. The fibers were 16 µm in diameter and 120 µm in length and had a bulk density of 0.93 ± 0.08 g/cc. MCF derived from a polyacrylonitrile (PAN) precursor was 7-9 µm in diameter and 100-200 µm in length. The visual appearance of both MGF and MCF was powder-like products.

Cement Formulas and Samples Preparation
In our study, the FCS/HSA aggregate wt. ratios were 100/0, 94/6, 90/10, 88/12, and 83/17, whereas the CAC/total aggregate wt. ratio was 70/30 for all formulations. The content of MGF and MCF was 10 and 5 wt.%, respectively, by total weight of CAC, FCS, and HSA. SHMP was used at 6 wt.% by total weight of CAC in all aggregates. All solid ingredients described above were prepared as a dry blend prior to mixing with water.
The composite and reference samples were prepared in the following sequences. A certain amount of water was added to a dry blended composite, followed by hand mixing for 1 min. The hand-mixed slurry was poured into molds (the mold shape depended on the test) and was left for 24 h at ambient temperature, allowing it to initiate acid-base reactions. Thereafter, the hardened composite was removed from the molds and placed in a 99 ± 1% relative humidity environment for 24 h at 85 • C. Finally, the pre-cured composite was autoclaved in a non-stirred Parr Reactor 4622 (Hillsboro, OR, USA) for 24 h at 100, 175, and 250 • C.

Measurements
Assuming that CaP cement is used in storage wells at the medium and high temperatures of 100 and 250 • C, the thermal shock (TS) test was carried out by heating the 100-or 250 • C-autoclaved CaP composite samples for 24 h in an oven at 175-or 250 • C, respectively, followed by immersion of hot cement in 25 • C water. This heat-quenching process was repeated three times. Therefore, in the 175 and 250 • C storage environment, the thermal shock-related temperature gradients were set at 150 and 225 • C, respectively. TS resistance was evaluated based on the changes in compressive strength and TC after the TS.
To evaluate the hydrothermal stability of HSA, we autoclaved HSA for 24 h at 100, 175, and 250 • C, and visually checked its miscibility with water. We also determined the water droplet contact angle on autoclaved HSA samples after 100 • C drying and conducted ATR-FTIR analyses. Water droplet contact angle measurements at ambient temperatures were assessed using Model CAA 3, Imass Inc. (Marshfield, MA, USA).
Pellet samples, 13.0 mm diameter by 1.7 mm height, were prepared for contact angle measurements according to the following sequence: (1) An amount of 0.6 mg HSA was blended with 15.0 g deionized water in a 20 mL sealed glass ampoule at ambient temperature; (2) The ampoule was autoclaved for 24 h at 100, 175, and 250 • C; (3) The autoclaved HSA was dried for 24 h in an oven at 100 • C; (4) An amount of 0.2 mg of dried sample was placed in the CrushIR™ 15 Ton Hydraulic Press (Fitchburg, WI, USA) mold; (5) A 10 ton pressure was loaded over the powder to make pellet-shaped samples. The dried autoclaved samples were also used in ATR-FTIR and TGA/DTA analyses. The former analysis was aimed at investigating the hydrothermal stability of the hydrophobic trimethylsiloxy group linked to silica aerogel. The latter was used to collect the thermal decomposition parameters, including the overall decomposition pattern, onset decomposition temperature, and weight losses. TGA/DTA (model Q50, TA Instruments, New Castle, DE, USA) analyses were run at a heating rate of 10 • C/min in a N 2 flow.
The slump flow of the cement slurries was measured by using a polyethylene flow cone with a top hole with a diameter of 20 mm, bottom hole with a diameter of 45 mm, and a height of 40 mm. The cement slurry was filled in the cone and placed on a flat carbon steel plate. Thereafter, the cone was slowly lifted, allowing the slurry to flow freely. The slurry slump, in mm, was determined 20 s after flow onset.
For compressive strength and compressive toughness, the composite samples were prepared in cylindrical molds (20 mm in diameter and 40 mm in height). Electromechanical Instron System Model 5967 (Norwood, MA, USA) was used to obtain these mechanical properties. Compressive strength is the capacity of a material or structure to withstand under compression. Cement with high compressive strength may be brittle and lack stress energy absorption, resulting in the rapid propagation of pre-existing and newly created cracks in cement bodies under stress conditions. Thus, cement ideally possesses toughness that delays or prevents crack propagation. To obtain quantitative data on compressive toughness, we determined the total energy consumed during the completion of cement compressive failure; this was computed from the enclosed area of the compressive stressstrain curve with the baseline extending from the beginning to the end of the curve. Toughness depended primarily on the balance between compressive strength and ductility.
To investigate the pozzolanic activity of each insulating aggregate and micro-fiber reinforcement, the 100, 175, and 250 • C-autoclaved CaP cements, containing one of FCS, HSA, MGF, and MCF, were prepared and studied using XRD, ATR-FTIR, and SEM-EDX probes. Changes in mechanical behavior due to pozzolanic activity were assessed by compressive-strength and compressive-toughness measurements.
To obtain information on the hydration of CaP composite slurries with different FCS/HSA ratios, a calorimetric study was done at the isothermal temperatures of 25, 50, and 85 • C, using TAM air isothermal microcalorimetry (TA Instruments, New Castle, DE, USA). We determined the initial and final setting times and heat energy evolved during the acid-base and hydration reactions.
TC was measured with the Quick Thermal Conductivity Meter, QTM-500, Kyoto Electronic (Kyoto, Japan) on rectangular prism samples (60 mm wide, 120 mm long, and 20 mm thick). Measuring TC in hydrous conductive material such as cement requires the use of a water-proofed probe consisting of a single heater and thermocouple. We placed the probe on the cement surface after removing the excess water on the water-saturated cement using a dry paper towel.
XRD (40 kV, 40 mA copper anode X-ray tube) and ATR-FTIR were used to identify amorphous and crystalline phase compositions and phase transitions of tested samples. The PDF-4/Minerals 2021 database of the International Center for Diffraction Data (ICDD) was used to analyze the XRD patterns.
JEOL 7600F SEM (Pleasanton, CA, USA) image analysis coupled with EDX elemental composition survey were used in this work.
The experiment included seven steps: (1) Assessment of hydrothermal stability of HSA at temperatures of up to 250 • C; (2) Investigation of workability and pumpability of the CaP composite slurry by slump and slurry density measurements; (3) Understanding pozzolanic and mechanical behaviors of individual insulating aggregates and micro-fibrous reinforcements in CaP composite system after autoclaving at 100, 175, and 250 • C; (4) Identification of cement-forming pathways by following exothermic chemical and hydration reactions of CaP composite slurries as a function of the FCS/HSA ratio at 85 • C; (5) Evaluation of compressive-strength and compressive-toughness of 100 and 250 • C-autoclaved CaP cement composites, before and after thermal shock (TS) tests; (6) Determination of TC and water-saturated bulk density as a function of the FCS/HSA ratio for 100 and 250 • Cautoclaved pre-and post-TS samples; (7) Study of phase composition and transitions along with microstructures and their alterations for an optimized composite formula. Figure 3 shows the phase separation for HSA and water, water droplets on the HSA surfaces, and ATR-FTIR analysis of HSA samples autoclaved at 100, 175, and 250 • C. Both 100 and 175 • C samples show water-HSA phase separation and water repulsion by the hydrophobic surfaces of HSA exposed to these temperatures for 24 h. In contrast, at 250 • C, water and HSA form a uniform mixture and water spreads over the HSA surface, suggesting that HSA has become hydrophilic at this elevated temperature. In fact, the contact angle, θ, of the water droplet on compressed dry HSA surfaces after the autoclaving for 24 h was 135 ± 1.5 (not shown), 136 ± 0.9, and 132 ± 1.8 degrees ( • ) for "as received" HSA, 100, and 175 • C-autoclaved samples, respectively. The water repellent behavior of solid surfaces based on the contact angle, in degrees ( • ), is categorized as follows [52]: superhydrophobic > 150 • , over-hydrophobic between 150 and 120 • , hydrophobic 120-90 • , and hydrophilic < 90 • . Thus, the surfaces of all samples except for the 250 • C-autoclaved sample could be categorized as over-hydrophobic. Therefore, these three degradation stages may involve the degradation of silicon-like polymer resulting in a total WL of 1.29%.   As expected, it was very difficult to measure the contact angle for 250 • C-autoclaved HSA; in fact, the compressed sample allowed the water to permeate through it, thereby no longer showing hydrophobicity. ATR-FTIR results supported this observation of transformation from a hydrophobic to hydrophilic sample. Namely, all hydrocarbon (CH 3 )-related bands at 2962, 1421, 1256, 845, and 757 cm −1 in the hydrophobic trimethylsilyl group vanished after 250 • C-autoclaving. Furthermore, hydrophilic silanol-related bands at 3457 and 947 cm −1 were also eliminated, whereas the peak height (∆A) of two Si-O-Si linkage-related bands at 1069 and 800 cm −1 noticeably rose, compared with that of "as-received", 100, and 175 • C samples. For instance, ∆A at 1069 cm −1 for 250 • C-treated HSA was 0.57, corresponding to 54, 62, and 36% higher values than "as-received", 100, and 175 • C-autoclaved samples, respectively. Although there is no experimental evidence, it is possible to rationalize that at 250 • C, the scission of -Si≡(CH 3 ) 3  If so, such self-condensation explains why the peak height of Si-O-Si bands remarkably increased, whereas silanol bands disappeared.

Slurry Properties
Additionally, a TGA survey in the temperature range of 25-800 • C supported the above information. Figure 4 shows TGA and DTG curves of "as-received" HSA as control, and 100 • 175, and 250 • C-autoclaved HSA samples. All the TGA curves except for the 250 • C sample showed a one stage-decomposition pattern with a degradation onset point at 414 • C. The decomposition onset and end points were determined from the DTG curves. The total mass loss rate (TMLR, %/mm) was computed from the area under the TGA curve, between the onset and end points. The weight loss (WL, %) was estimated from the TGA curve. For "as-received" HSA, it is possible to assume that the WL of~9.52% was due to the decomposition of all trimethylsilyl groups. Similar curves, except for the smaller WL of~7.87%, was observed for 100 • C-autoclaved sample, suggesting that~1.65% of hydrocarbons may be decomposed during autoclaving at this temperature. At 175 • C, even though the WL value was almost the same as that at 100 • C, the TMLR declined from 0.73 (at 100 • C) to 0.42%/min. It is likely that the carbonaceous degradation products of CH 3 in the trimethylsilyl groups contributed to the WL. This suggests that some trimethylsilyl groups might be degraded at 175 • C, but HSA remains hydrophobic. The sample autoclaved at 250 • C displayed a three-stage decomposition (I, II, III in the Figure 4) with onset points at 110, 314, and 568 • C. The WL at each stage was 0.28, 0.48, and 0.53%. As aforementioned in the ATR-FTIR study, the hydrophobic trimethylsilyl groups at 250 • C underwent three steps of molecular alterations: (1) Scission of the -Si≡(CH 3 ) bond; (2) Substitution of the -CH 3 group for a hydroxyl (-OH) group, to form silanol -Si≡(OH) 3 ; (3) Self-condensation between silanols, making it possible to assemble a silicon-like polymer. Therefore, these three degradation stages may involve the degradation of silicon-like polymer resulting in a total WL of 1.29%.    Table 2 gives the properties of CaP slurries as a function of the FCS/HSA ratio, including the water/blend (W/B) weight ratio (blend weight includes that of CAC, FCS, and HSA), density, slump size, and pH of pore solution extracted from the slurry by centrifuging. The W/B ratio was decided based on the addition of water up to the point where some water was pushed to the surface of the slurry, a phenomenon called "water bleeding". As the water uptake of the reference slurry without HSA was the same as the slurry with the highest HSA content, it can be concluded that HSA did not take up any additional water, due to the hydrophobic treatment of its surface. The cement slurries were made with W/B ratios in the range of 0.48-0.55. As expected, a downward trend for slurry density was observed; the density declined with increasing HSA content, from 1.29 g/cm 3 at FCS/HSA 100/0 to 1.07 g/cm 3 at 83/17. Similar to the W/B ratio, no specific trend was observed for slump size; it was in the range between 78 and 83 mm for all samples. Such slurry properties indicate that all the aggregates and reinforcements were uniformly distributed in a slurry of good fluidity, which is important for the pumpability of the blend in the field.

Pozzolanic and Mechanical Behaviors of Aggregates and Reinforcements
To understand pozzolanic and mechanical behaviors of individual insulating aggregates and micro-fibrous reinforcements in the CaP composite system, after autoclaving at 100, 175, and 250 • C, four CaP cement composites, each with one of FCS, HSA, MGF, and MCF, were prepared. These blends had 24.9 g FCS, 5.1 g HSA, 10 g MGF, and 5 g MCF, each mixed with 70 g CAC containing 6 g SHMP. For comparison purposes, neat CaP cement without any aggregates or reinforcements was prepared as a control. Figure 5 presents XRD patterns of neat CaP cement (pore solution of pH 9.77) after autoclaving for 24 h at 100, 175, and 250 • C. Three major crystalline phases were identified in the patterns of 100 • C-autoclaved samples: Si-free katoite (C 3 AH 6 ), gibbsite (AH 3 ), and monocalcium aluminate (CA), krotite. Additionally, two minor phases were calcium dealuminate (CA 2 ), grossite, and hydroxyapatite [(HOAp, Ca 10 (PO 4 ) 6 (OH) 2 ]. Hydration of CA and CA 2 present in CAC resulted in the formation of Si-free katoite and gibbsite through the following reactions: 3CA + 12H 2 O → C 3 AH 6 + 2AH 3 and 3CA 2 + 21H 2 O → C 3 AH 6 + 5AH 3 [59,60]. Thus, some non-hydrated clinkers remained in the 100 • C-24 h-autoclaved neat CaP cement. It appears that cement slurry made from the dry blend of CAC and SHMP at 100 • C provides two different cementitious binding phases: (1) CAC hydration products, such as katoite and gibbsite; (2) HOAp, the reaction product of SHMP with CAC. At 175 • C, the features of the XRD patten were strikingly different from those at 100 • C. There were four major differences: (1) Non-hydrated CA and CA 2 clinkers at this temperature were not found, suggesting that more alkalis such as Ca 2+ and OH − were dissociated from CA and CA 2 ; (2) Disappearance of all gibbsite-related d-spacings, whereas the line intensity of katoite remained unchanged; (3) Emergence of boehmite (γ-AlOOH) phase high-intensity peaks; (4) Increased intensity of HOAp-related peaks. Increased content calcium and hydroxide ions from CAC hydration resulted in higher HOAp concentrations and higher intensities of its peaks. As for (2) and (3), the alkali dissolution and recrystallization of gibbsite produces boehmite, Al(OH) 3 [61]. Hence, three major crystalline phases, HOAp, katoite, and boehmite, were responsible for strengthening neat CaP cement at 175 • C. A similar pattern, except for the appearance of intense line signals of calcium carbonate (CaCO 3 ), was observed for the 250 • C-autoclaved sample.

Neat CaP Cement
These XRD results were supported by the ATR-FTIR results. Figure 6 depicts ATR-FTIR spectra in the frequency region of 4000 to 650 cm −1 for "as-received" Calgon SHMP salt, hydrolyzed SHMP solution, and 100-, 175-, and 250 • C-autoclaved neat CaP cements. For "as-received" SHMP, the two bands located at 774 and 720 cm −1 represent the P-O-P linkage in the Calgon cyclic chain structure [62]. The other bands at 1256, 1090, 976, and 866 cm −1 , respectively, are due to asymmetric stretching (ν as P=O ) of the P=O bond, asymmetric (ν as P-O ) and symmetric (ν s P-O ) vibration modes of the P-O bond, as well as symmetric (ν s P-O-P ) of the P-O-P linkage in Calgon [62][63][64][65]. Compared with these results, the spectral features of hydrolyzed SHMP exhibited three differences: (1) Disappearance of all cyclic P-O-P linkage-associated bands at 774 and 720 cm −1 ; (2) Shift of the 866 cm −1 band of ν s P-O-P towards a higher frequency site at 876 cm −1 ; (3) Presence of water-related bands at 3356, 3245, and 1637 cm −1 , whereas no shift was observed from ν as P=O , ν as P-O , and ν s P-O bands. For the first two differences, it is possible to rationalize that the hydrolysis of SHMP led to cycle-opening and cycle-scission, thereby initially transforming the Calgon into ionic liner polyphosphate, which was further hydrolyzed to orthophosphate [62,[66][67][68][69][70]. The pH of SHMP solution was~6. 5 tent calcium and hydroxide ions from CAC hydration resulted in higher HOAp concentrations and higher intensities of its peaks. As for (2) and (3), the alkali dissolution and recrystallization of gibbsite produces boehmite, Al(OH)3 + OH − → Al(OH)4 − → AlH(OH)4 → -AlOOH + 2H2O [61]. Hence, three major crystalline phases, HOAp, katoite, and boehmite, were responsible for strengthening neat CaP cement at 175 °C. A similar pattern, except for the appearance of intense line signals of calcium carbonate (CaCO3), was observed for the 250 °C-autoclaved sample. These XRD results were supported by the ATR-FTIR results. Figure 6 depicts ATR-FTIR spectra in the frequency region of 4000 to 650 cm −1 for "as-received" Calgon SHMP salt, hydrolyzed SHMP solution, and 100-, 175-, and 250 °C-autoclaved neat CaP cements. For "as-received" SHMP, the two bands located at 774 and 720 cm −1 represent the P-O-P linkage in the Calgon cyclic chain structure [62]. The other bands at 1256, 1090, 976, and 866 cm −1 , respectively, are due to asymmetric stretching (νas P=O) of the P=O bond, asymmetric (νas P-O) and symmetric (νs P-O) vibration modes of the P-O bond, as well as symmetric (νs P-O-P) of the P-O-P linkage in Calgon [62][63][64][65]. Compared with these results, the spectral features of hydrolyzed SHMP exhibited three differences: (1) Disappearance of all cyclic P-O-P linkage-associated bands at 774 and 720 cm −1 ; (2) Shift of the 866 cm −1 band of νs P-O-P towards a higher frequency site at 876 cm −1 ; (3) Presence of water-related bands at 3356, 3245, and 1637 cm −1 , whereas no shift was observed from νas P=O, νas P-O, and νs P-O bands. For the first two differences, it is possible to rationalize that the hydrolysis of SHMP led to cycle-opening and cycle-scission, thereby initially transforming the Calgon into ionic liner polyphosphate, which was further hydrolyzed to orthophosphate [62,[66][67][68][69][70]. The pH of SHMP solution was ~6.5. As the predominant orthophosphate species at pH~6.5 are H2PO4 − and HPO4 2− [70], these anions likely play a major role in promoting acid-base reactions with Ca 2+ alkali hydrolysate from CAC.
As described in the XRD study, the 100 °C-autoclaved neat CaP cement had two major crystalline phases: Si-free katoite and gibbsite, derived from the hydration of CAC. Correspondingly, the ATR-FTIR spectrum indicated the presence of these hydration products. The spectrum involved three groups: (1) OH in the ranges of 3700-3300 cm −1 and 1100-900 cm −1 ; (2) Condensed tetrahedral AlO4 in 800-700 cm −1 ; (3) Condensed octahedral AlO6 in 680-650 cm −1 . Regarding gibbsite [71,72], the O-H bond stretching (νO-H) and O-H bending (δO-H) vibration modes involved six frequencies, with 3623, 3529, 3453, and 3384 As described in the XRD study, the 100 • C-autoclaved neat CaP cement had two major crystalline phases: Si-free katoite and gibbsite, derived from the hydration of CAC. Correspondingly, the ATR-FTIR spectrum indicated the presence of these hydration products. The spectrum involved three groups: (1) OH in the ranges of 3700-3300 cm −1 and 1100-900 cm −1 ; (2) Condensed tetrahedral AlO 4 in 800-700 cm −1 ; (3) Condensed octahedral AlO 6 in 680-650 cm −1 . Regarding gibbsite [71,72] 3 2− group in carbonate compounds; namely, the first two bands are assignable to ν as C-O in CO 3 2− and the last one [73,74]. As crystalline carbonate was not found by XRD, the detected carbonate is likely amorphous. Additionally, the presence of HOAp as a minor phase in XRD can be identified from a shoulder band at 1072 cm −1 attributed to the ν as P-O of the PO 4 3− group [62,75,76]. On the other hand, Si-free katoite-related bands [77][78][79] involved the stretching (ν O-H ) mode at 3667 cm −1 . The contributors to the other four bands at 1025, 963, 736, and 663 cm −1 were the same as those of gibbsite. At 175 • C, there were three major changes in the spectrum compared with 100 • C. First, there was a disappearance of all gibbsite-related bands, whereas Si-free katoite-related bands remained unchanged. Secondly, there was an emergence of three new principal bands at 3294, 3090, and 1159 cm −1 . Thirdly, the shoulder band at became the principal one. The three new bands were associated with the ν as O-H , ν s O-H , and δ O-H of the OH group in boehmite [80][81][82]. The shoulder band at 1072 cm −1 was relevant to the ν as P-O of the PO 4 3− group in HOAp, as the major phase. All the results agreed with the XRD findings. Spectral features of the 250 • C-autoclaved sample closely resembled the spectrum at 175 • C.
cm −1 belonging to νO-H, and 1025 and 963 cm −1 as δO-H, whereas the Al-O bond stretching (νAl-O) in tetrahedral AlO4 and octahedral AlO6 groups can be recognized by the presence of 736 and 663 cm −1 bands, respectively. The bands at 1466, 1430, and 853 cm −1 are related to CO3 2− group in carbonate compounds; namely, the first two bands are assignable to νas C-O in CO3 2− and the last one is O-C-O out-of-plane bending (δO-C-O) [73,74]. As crystalline carbonate was not found by XRD, the detected carbonate is likely amorphous. Additionally, the presence of HOAp as a minor phase in XRD can be identified from a shoulder band at 1072 cm −1 attributed to the νas P-O of the PO4 3− group [62,75,76]. On the other hand, Si-free katoite-related bands [77][78][79] involved the stretching (νO-H) mode at 3667 cm −1 . The contributors to the other four bands at 1025, 963, 736, and 663 cm −1 were the same as those of gibbsite. At 175 °C, there were three major changes in the spectrum compared with 100 °C. First, there was a disappearance of all gibbsite-related bands, whereas Si-free katoite-related bands remained unchanged. Secondly, there was an emergence of three new principal bands at 3294, 3090, and 1159 cm −1 . Thirdly, the shoulder band at became the principal one. The three new bands were associated with the νas O-H, νs O-H, and δO-H of the OH group in boehmite [80][81][82]. The shoulder band at 1072 cm −1 was relevant to the νas P-O of the PO4 3− group in HOAp, as the major phase. All the results agreed with the XRD findings. Spectral features of the 250 °C-autoclaved sample closely resembled the spectrum at 175 °C.
Thus, the three major phases, HOAp, boehmite, and Si-free katoite, were formed in neat CaP cement in the range from 175 to 250 °C. At 100 °C, HOAp, as a minor phase, was induced by acid-base reactions between the orthophosphoric acid liberated from hydrolyzed SHMP and Ca 2+ and OH − alkali hydrolysates of CA and CA2. Higher temperatures promoted more acid formation from SHMP and alkalis from CA and CA2, resulting in the HOAp becoming the major phase. Aluminum from CA and CA2 formed gibbsite at 100 °C, Thus, the three major phases, HOAp, boehmite, and Si-free katoite, were formed in neat CaP cement in the range from 175 to 250 • C. At 100 • C, HOAp, as a minor phase, was induced by acid-base reactions between the orthophosphoric acid liberated from hydrolyzed SHMP and Ca 2+ and OH − alkali hydrolysates of CA and CA 2 . Higher temperatures promoted more acid formation from SHMP and alkalis from CA and CA 2 , resulting in the HOAp becoming the major phase. Aluminum from CA and CA 2 formed gibbsite at 100 • C, which transferred to boehmite at 175 • C. Based on the above information, Figure 7 illustrates two different cement-forming pathways by acid-base and hydration reactions in neat CaP cement.  As seen in Figure 7, the formation of HOAp as the major phase engenders release of Na+ ions due to the consumption of orthophosphate. Simultaneously, hydrolysis of CA and CA2 creates an alkaline environment through the release of Ca 2+ and OH − . This is the reason why the reference pH of neat CaP is 9.77. The high pH raises concerns about activating pozzolanic decomposition of lightweight aggregates and glass fibers. In fact, as seen in SEM-EDX data ( Figure 8) on a freshly fractured surface of 100 °C-autoclaved CaP, the predominant element in areas marked as No.1 and 2 is Na, representing more than As seen in Figure 7, the formation of HOAp as the major phase engenders release of Na + ions due to the consumption of orthophosphate. Simultaneously, hydrolysis of CA and CA 2 creates an alkaline environment through the release of Ca 2+ and OH − . This is the reason why the reference pH of neat CaP is 9.77. The high pH raises concerns about activating pozzolanic decomposition of lightweight aggregates and glass fibers. In fact, as seen in SEM-EDX data ( Figure 8) on a freshly fractured surface of 100 • C-autoclaved CaP, the predominant element in areas marked as No. 1 and 2 is Na, representing more than half of the total mass weight (wt%) of all elements. As seen in Figure 7, the formation of HOAp as the major phase engenders release of Na+ ions due to the consumption of orthophosphate. Simultaneously, hydrolysis of CA and CA2 creates an alkaline environment through the release of Ca 2+ and OH − . This is the reason why the reference pH of neat CaP is 9.77. The high pH raises concerns about activating pozzolanic decomposition of lightweight aggregates and glass fibers. In fact, as seen in SEM-EDX data ( Figure 8) on a freshly fractured surface of 100 °C-autoclaved CaP, the predominant element in areas marked as No.1 and 2 is Na, representing more than half of the total mass weight (wt%) of all elements.  Pozzolanic reactions of FCS, HSA, and MGF in alkaline environments of CaP slurries after curing at 100, 175, and 250 • C were investigated in the next step. A formulation with a non-pozzolan MCF reinforcement was used as a control. Figure 9 shows XRD patterns for 100-, −175-, and −250 • C-autoclave CaP/FCS cements (pore solution pH 10.06). The features of 100 • C-autoclaved CaP/FCS and neat CaP cement were alike, except for the two FCS-associated crystalline phases, mullite (ICDD#04-016-1586) and silica (#00-014-0654). At 175 • C, similar to neat CaP, the two non-hydrated phases, CA and CA 2 , disappeared and HOAp became the major phase. The gibbsite phase was transferred to the boehmite phase, whereas katoite remained a major phase. The only difference from the neat cement was the appearance of lowintensity albite (NaAlSi 3 O 8 ) peaks in the presence of FCS. The reaction of albite formation is Na + + Al(OH) 4 − +3H 4 SiO 4 (aq) ↔ NaAlSi 3 O 8 + 8H 2 O [83]. Orthosilicic acid (H 4 SiO 4 ) was possibly derived from the alkaline dissolution of silica and aluminosilicate mullite in the FCS shell, whereas Al(OH) 4 − came from the hydrolysis of CAC. Thus, FCS appears to undergo PR; namely, alkali dissolution of the FCS shell leading to the precipitation of albite, a PR product. At 250 • C, another PR product was identified in addition to albite: zeolite hydrosodalite [Na 6 (AlSiO 4 ) 6 ·8H 2 O]. However, strong peaks of mullite and silica were still present in the 250 • C XRD pattern, suggesting that a certain amount of these pozzolans remained unreacted.

CaP/FCS System
(aq) ↔ NaAlSi3O8 + 8H2O [83]. Orthosilicic acid (H4SiO4) was possibly derived from the alkaline dissolution of silica and aluminosilicate mullite in the FCS shell, whereas Al(OH)4 − came from the hydrolysis of CAC. Thus, FCS appears to undergo PR; namely, alkali dissolution of the FCS shell leading to the precipitation of albite, a PR product. At 250 °C, another PR product was identified in addition to albite: zeolite hydrosodalite [Na6(Al-SiO4)6·8H2O]. However, strong peaks of mullite and silica were still present in the 250 °C XRD pattern, suggesting that a certain amount of these pozzolans remained unreacted.     In summary, in agreement with the XRD results, FCS was susceptible to PR at hydrothermal temperatures of ≥175 • C. As a result, the albite (anhydrous N-A-S system) and hydrosodalite zeolite (N-A-S-H system) were present as minor phases in the cement matrix along with the major phases of HOAp, boehmite, and katoite. Thus, the 663 cm FCS strongly suggests that the FCS surface was susceptible to PR, not only with Na + , but also with Ca 2+ ions. Relating this to ATR-FTIR results, it appears that amorphous C-A-H and C-A-S-H also contribute to the bands in the region of 1000 to 870 cm −1 . Some typical hydrosodalite-related porous crystalline balls were seen in the cement/FCS interfacial boundary region and at locations adjacent to FCS. Hence, the pozzolanic activity of FCS offered improved interfacial bonding of FCS with cement through the formation of crystalline albite, hydrosodalite, and amorphous C-A-H and C-A-S-H PR products.      Figure 14 shows the XRD patterns of 100-, 175-, and 250 • C-autoclaved CaP/HSA cements (pore solution pH 10.74). At 100 and 175 • C, the major crystalline products of this system were very similar to those in neat CaP. At 100 • C, they included katoite and gibbsite, and at 175 • C, HOAp, boehmite, and katoite. Unlike the CaP/FCS system, N-A-S type minerals, albite, and N-A-S-H phases were not identified at 175 • C. The TGS/DTG results suggested some degradation of trimethylsilyl groups in HSA at 175 • C. If the loss of trimethylsilyl groups was significant, the PR of HSA with sodium could proceed with the formation of N-A-S or N-A-S-H type reaction products. As such products were not detected, it is reasonable to assume that even with some degradation of trimethylsilyl groups, the PR of HSA were negligible at 175 • C. In contrast, because all hydrophobic trimethylsilyl groups were decomposed at 250 • C, the trimethylsilyl-free hydrophilic silica aerogels became susceptible to alkali dissolution, similar to FCS. The XRD patterns of the 250 • C-autoclaved sample included peaks of albite and hydrosodalite brought about by the hydrothermal interactions of the following three reactants: (1) Na + ; (2) Al(OH) 4 − from the hydrolysis of CAC; (3) (H 4 SiO 4 ) (aq) , derived from the alkali dissolution of silica gel.
ATR-FTIR results ( Figure 15) at 100 • C detected all HSA-related bands coexisting with gibbsite and katoite as major phases, clearly demonstrating that HSA was not susceptible to PR at that temperature. However, at 175 • C, all CH 3 -related bands in trimethylsilyl disappeared. As mentioned in previous hydrothermal stability studies of HSA, HSA exposed to 175 • C in deionized water retained CH 3 bands. However, trimethylsilyl was vulnerable to a 175 • C alkaline cement pore water solution with a pH of 10.74. Despite the degradation of hydrophobic trimethylsilyl groups, no alkali dissolution of silica aerogel was detected, as there were no noticeable bands related to amorphous or crystalline N-A-S and N-A-S-H phases in the region between 1000 to 870 cm −1 , in agreement with the XRD results. In contrast, at 250 • C, anhydrous and hydrated N-A-S-(H) bands attributed to albite and hydrosodalite, respectively, were detected at 958 and 914 cm −1 , indicating the alkali dissolution of silica aerogel at 250 • C. The band at 842 cm −1 was ascribed to CO 3 2− in conjunction with the CO 3 2− band at 853 cm −1 .  Figure 14 shows the XRD patterns of 100-, 175-, and 250 °C-autoclaved CaP/HSA cements (pore solution pH 10.74). At 100 and 175 °C, the major crystalline products of this system were very similar to those in neat CaP. At 100 °C, they included katoite and gibbsite, and at 175 °C, HOAp, boehmite, and katoite. Unlike the CaP/FCS system, N-A-S type minerals, albite, and N-A-S-H phases were not identified at 175 °C. The TGS/DTG results suggested some degradation of trimethylsilyl groups in HSA at 175 °C. If the loss of trimethylsilyl groups was significant, the PR of HSA with sodium could proceed with the formation of N-A-S or N-A-S-H type reaction products. As such products were not detected, it is reasonable to assume that even with some degradation of trimethylsilyl groups, the PR of HSA were negligible at 175 °C. In contrast, because all hydrophobic trimethylsilyl groups were decomposed at 250 °C, the trimethylsilyl-free hydrophilic silica aerogels became susceptible to alkali dissolution, similar to FCS. The XRD patterns of the 250 °Cautoclaved sample included peaks of albite and hydrosodalite brought about by the hydrothermal interactions of the following three reactants: (1) Na + ; (2) Al(OH)4 − from the hydrolysis of CAC; (3) (H4SiO4)(aq), derived from the alkali dissolution of silica gel. ATR-FTIR results ( Figure 15) at 100 °C detected all HSA-related bands coexisting with gibbsite and katoite as major phases, clearly demonstrating that HSA was not susceptible to PR at that temperature. However, at 175 °C, all CH3-related bands in trimethylsilyl disappeared. As mentioned in previous hydrothermal stability studies of HSA, HSA exposed to 175 °C in deionized water retained CH3 bands. However, trimethylsilyl was vulnerable to a 175 °C alkaline cement pore water solution with a pH of 10.74. Despite the degradation of hydrophobic trimethylsilyl groups, no alkali dissolution of silica aerogel was detected, as there were no noticeable bands related to amorphous or crystalline N-A-S and N-A-S-H phases in the region between 1000 to 870 cm −1 , in agreement with the XRD results. In contrast, at 250 °C, anhydrous and hydrated N-A-S-(H) bands attributed to albite and hydrosodalite, respectively, were detected at 958 and 914 cm −1 , indicating the alkali dissolution of silica aerogel at 250 °C. The band at 842 cm −1 was ascribed to CO3 2− in conjunction with the CO3 2− band at 853 cm −1 . Furthermore, the information described above was supported by the SEM-EDX exploration of the microstructure in the 250 • C-autoclaved sample. The image of the fractured sample ( Figure 16) showed a disappearance of all HSA aggregates due to their alkali dissolution and the formation of numerous large (~250 µm) cavities in their place. PR products of silica gel were observed in the cavities (Figure 17). EDX elemental analysis of PR products showed the presence of Na, Al, Si, and small amounts of Ca (less than 4%). The Al/Na, Al/Na, and Si/Al atomic ratios at locations No. 1 and 2 were attributable to hydrosodalite, possibly coexisting with some C,N-Al-S-H phases. However, the morphological features of these hydrosodalite crystals differed from the hydrosodalite formed in the CaP/FCS system, suggesting a poorly crystallized hydrosodalite. Furthermore, the information described above was supported by the SEM-EDX exploration of the microstructure in the 250 °C-autoclaved sample. The image of the fractured sample ( Figure 16) showed a disappearance of all HSA aggregates due to their alkali dissolution and the formation of numerous large (~250 µ m) cavities in their place. PR products of silica gel were observed in the cavities ( Figure 17). EDX elemental analysis of PR products showed the presence of Na, Al, Si, and small amounts of Ca (less than 4%). The Al/Na, Al/Na, and Si/Al atomic ratios at locations No. 1 and 2 were attributable to hydrosodalite, possibly coexisting with some C,N-Al-S-H phases. However, the morphological features of these hydrosodalite crystals differed from the hydrosodalite formed in the CaP/FCS system, suggesting a poorly crystallized hydrosodalite.   Furthermore, the information described above was supported by the SEM-EDX exploration of the microstructure in the 250 °C-autoclaved sample. The image of the fractured sample ( Figure 16) showed a disappearance of all HSA aggregates due to their alkali dissolution and the formation of numerous large (~250 µ m) cavities in their place. PR products of silica gel were observed in the cavities (Figure 17). EDX elemental analysis of PR products showed the presence of Na, Al, Si, and small amounts of Ca (less than 4%). The Al/Na, Al/Na, and Si/Al atomic ratios at locations No. 1 and 2 were attributable to hydrosodalite, possibly coexisting with some C,N-Al-S-H phases. However, the morphological features of these hydrosodalite crystals differed from the hydrosodalite formed in the CaP/FCS system, suggesting a poorly crystallized hydrosodalite.

CaP/MGF System
As seen in Figure 18, the XRD patterns of CaP/MGF cement (pore solution pH 11.05) at 100 and 175° were like those of the CaP/HSA system, namely, no crystalline anhydrous and hydrate N-A-S phases were detected. At 250 °C, unlike in the CaP/HSA system, only hydrosodalite formed as a minor phase, and no albite.
Corresponding to XRD results, the N-A-S-H phases associated with hydrosodalite could be recognized from the bands at 958 and 914 cm −1 at 250 °C ( Figure 19). As MGF is a non-crystalline calcium aluminosilicate (C-A-S), two bands at 910 and 702 cm −1 in the spectrum of "as-received" MGF were implicated in M-O-Si (M: Ca and Al) bonds. Therefore, the bands at 958 and 914 cm −1 may also belong to amorphous C,N-A-S-H phases formed by the PR of MGF with Na. Similar to HSA at 175 °C, no representative bands of N-A-S-H and C,N-A-S-H phases were found in the region of 1000-870 cm −1 , verifying little susceptibility of MGF to PR at that temperature.

CaP/MGF System
As seen in Figure 18, the XRD patterns of CaP/MGF cement (pore solution pH 11.05) at 100 and 175 • were like those of the CaP/HSA system, namely, no crystalline anhydrous and hydrate N-A-S phases were detected. At 250 • C, unlike in the CaP/HSA system, only hydrosodalite formed as a minor phase, and no albite.

CaP/MGF System
As seen in Figure 18, the XRD patterns of CaP/MGF cement (pore solution pH 11.05) at 100 and 175° were like those of the CaP/HSA system, namely, no crystalline anhydrous and hydrate N-A-S phases were detected. At 250 °C, unlike in the CaP/HSA system, only hydrosodalite formed as a minor phase, and no albite.
Corresponding to XRD results, the N-A-S-H phases associated with hydrosodalite could be recognized from the bands at 958 and 914 cm −1 at 250 °C ( Figure 19). As MGF is a non-crystalline calcium aluminosilicate (C-A-S), two bands at 910 and 702 cm −1 in the spectrum of "as-received" MGF were implicated in M-O-Si (M: Ca and Al) bonds. Therefore, the bands at 958 and 914 cm −1 may also belong to amorphous C,N-A-S-H phases formed by the PR of MGF with Na. Similar to HSA at 175 °C, no representative bands of N-A-S-H and C,N-A-S-H phases were found in the region of 1000-870 cm −1 , verifying little susceptibility of MGF to PR at that temperature. Corresponding to XRD results, the N-A-S-H phases associated with hydrosodalite could be recognized from the bands at 958 and 914 cm −1 at 250 • C ( Figure 19). As MGF is a non-crystalline calcium aluminosilicate (C-A-S), two bands at 910 and 702 cm −1 in the spectrum of "as-received" MGF were implicated in M-O-Si (M: Ca and Al) bonds. Therefore, the bands at 958 and 914 cm −1 may also belong to amorphous C,N-A-S-H phases formed by the PR of MGF with Na. Similar to HSA at 175 • C, no representative bands of N-A-S-H and C,N-A-S-H phases were found in the region of 1000-870 cm −1 , verifying little susceptibility of MGF to PR at that temperature.    Figure 20 shows the lack of PR of MGF, whereas Figure 21 demonstrates the degradation of MGF in PR. A SEM image of the non-reacted area of MGF presents a smooth fiber texture. The EDX of this smooth surface was consistent with the elemental composition of MGF, except for a small presence of Na. The EDX for location No. 2 of the cement adhering to MGF indicated the presence of three major elements (O, Al, and Ca) with atomic fractions ≥25% and three minor elements (Mg, Si, and P) with atomic fractions <4% (except C). Therefore, similar to the FCS surface, the cement was made by amorphous C-A-H and C-A-S-H with Ca/Al, Si/Al, and Ca/Si atomic ratios of 1.04, 0.15 and 6.77, respectively, verifying that the MGF surface was also susceptible to PR with alkaline calcium species. In contrast, the SEM image of the reacted area of MGF ( Figure 21) highlighted the development of an excessive interfacial bonding between the cement and MGF, leading to a shear-induced fracture of MGF during sample preparation. The EDX determined that atomic fractions were similar in locations No. 1 and 2, with the only difference being a higher Si content in the PR-affected area, with the cement strongly adhering to the fiber. Thus, the atomic ratios of Si-rich amorphous C-A-S-H were as follows: Ca/Al=0.82, Si/Al=0.47, and Ca/Si=1.75. On the other hand, it was very difficult to identify hydrosodalite phases on MGF surfaces. This phase probably formed in the interfacial cement-MGF region beneath the C-A-S-H layer.  Figure 20 shows the lack of PR of MGF, whereas Figure 21 demonstrates the degradation of MGF in PR. A SEM image of the non-reacted area of MGF presents a smooth fiber texture. The EDX of this smooth surface was consistent with the elemental composition of MGF, except for a small presence of Na. The EDX for location No. 2 of the cement adhering to MGF indicated the presence of three major elements (O, Al, and Ca) with atomic fractions ≥25% and three minor elements (Mg, Si, and P) with atomic fractions <4% (except C). Therefore, similar to the FCS surface, the cement was made by amorphous C-A-H and C-A-S-H with Ca/Al, Si/Al, and Ca/Si atomic ratios of 1.04, 0.15 and 6.77, respectively, verifying that the MGF surface was also susceptible to PR with alkaline calcium species. In contrast, the SEM image of the reacted area of MGF ( Figure 21) highlighted the development of an excessive interfacial bonding between the cement and MGF, leading to a shear-induced fracture of MGF during sample preparation. The EDX determined that atomic fractions were similar in locations No. 1 and 2, with the only difference being a higher Si content in the PR-affected area, with the cement strongly adhering to the fiber. Thus, the atomic ratios of Si-rich amorphous C-A-S-H were as follows: Ca/Al = 0.82, Si/Al = 0.47, and Ca/Si = 1.75. On the other hand, it was very difficult to identify hydrosodalite phases on MGF surfaces. This phase probably formed in the interfacial cement-MGF region beneath the C-A-S-H layer.      Na + , Ca 2+ , and OH − at 250 °C. In fact, the spectral features at 100 and 250 °C were the same as for neat CaP. In contrast, upward trends could be seen for the CaP/FCS system with temperature increase. The compressive strength developed at 100 °C (9.1 MPa) increased at 175 °C (20.4 MPa) and increased further at 250 °C (23.3 MPa). This is a result of FCS shell pozzolanic activity at temperatures ≥175 °C. Pozzolanic reactions form N-A-S phases (crystallizing as albite) at 175 °C, and hydrosodalite and additional N-A-S-H phases at 250 °C. These PR products on the surface of FCS improved the adhesion of hard-shells to the CaP matrix, thereby resulting in increased compressive strength with the enhancement of pozzolanic activity at elevated temperatures. On the other hand, the strength of the CaP/HSA system at 100 °C dropped by nearly 43% to 5.1 MPa. It is evident that HSA itself did not contribute to compressive strength in CaP cement, and neither did the PR products of HSA. As described in the SEM-EDX study, the alkali dissolution of silica gel by the hydrothermal degradation of HSA at 250 °C engendered the creation of numerous cavities, including cavities as large as ~250 µ m in size. This was the major reason why strength was drastically decreased. Furthermore, the contribution of HSA itself to the compressive strength of CaP was almost zero compared with that of FCS. Incorporation of MGF into CaP noticeably improved cement strength. The compressive strengths of 47.7 and 45.7 MPa developed at 100 and 175 °C was tantamount to nearly 1.6-and more than 5-times greater than the strengths of neat CaP and CaP/HSA, respectively, strongly demonstrating that MGF reinforcements could help recover the considerable strength loss brought about by adding HSA. However, at 250 °C, this strength markedly decreased, decreasing by 17% compared with its strength at 175 °C, implying that hydrosodalite formation from PR of MGF may have a negative effect in sustaining adequate compressive strength. This may be due to the conversion of ductile fibers into brittle ones by PR products. In contrast     In contrast, upward trends could be seen for the CaP/FCS system with temperature increase. The compressive strength developed at 100 • C (9.1 MPa) increased at 175 • C (20.4 MPa) and increased further at 250 • C (23.3 MPa). This is a result of FCS shell pozzolanic activity at temperatures ≥175 • C. Pozzolanic reactions form N-A-S phases (crystallizing as albite) at 175 • C, and hydrosodalite and additional N-A-S-H phases at 250 • C. These PR products on the surface of FCS improved the adhesion of hard-shells to the CaP matrix, thereby resulting in increased compressive strength with the enhancement of pozzolanic activity at elevated temperatures. On the other hand, the strength of the CaP/HSA system at 100 • C dropped by nearly 43% to 5.1 MPa. It is evident that HSA itself did not contribute to compressive strength in CaP cement, and neither did the PR products of HSA. As described in the SEM-EDX study, the alkali dissolution of silica gel by the hydrothermal degradation of HSA at 250 • C engendered the creation of numerous cavities, including cavities as large as~250 µm in size. This was the major reason why strength was drastically decreased. Furthermore, the contribution of HSA itself to the compressive strength of CaP was almost zero compared with that of FCS. Incorporation of MGF into CaP noticeably improved cement strength. The compressive strengths of 47.7 and 45.7 MPa developed at 100 and 175 • C was tantamount to nearly 1.6-and more than 5-times greater than the strengths of neat CaP and CaP/HSA, respectively, strongly demonstrating that MGF reinforcements could help recover the considerable strength loss brought about by adding HSA. However, at 250 • C, this strength markedly decreased, decreasing by 17% compared with its strength at 175 • C, implying that hydrosodalite formation from PR of MGF may have a negative effect in sustaining adequate compressive strength. This may be due to the conversion of ductile fibers into brittle ones by PR products. In contrast MCF, which is not susceptible to PR, showed a greater ability to recover strength loss due to the addition of HSA at temperatures up to 250 • C than MGF.

Calorimetric Study of CaP Composites in Hydration
Compressive toughness depends on a balance between compressive strength and ductility. The data of compressive toughness mirrored the trends of compressive strength. MCF showed better performance than MGF. Importantly, MCF enhanced toughness, rather than compressive strength, for neat CaP. For example, compared with the~1.6-fold improvement in compressive strength, toughness improved by~2.2-fold. Nonetheless, both MGF and MCF offered improved toughness of the CaP cement matrix. The reinforcements can help considerably alleviate the losses in strength and toughness caused by HSA in CaP cement. Figure 25 illustrates the isothermal calorimetry curves for neat CaP cement and CaP composite slurries with 100/0, 90/10, and 83/17 FCS/HSA ratios at 25 • C. There were two distinct heat release stages (I and II) during the hydration of the slurries. First stage (I) reactions occurred within the first~80 min from the start of measurement; second stage (II) reactions took place between~3 and~15 h from the start of measurement. The total normalized heat released during the first stage, computed by integrating the area under the curve, was 56.06 J/g for neat CaP and 32.56, 35.17, and 34.01 J/g for 100/0, 90/10, and 83/17 CaP/HSA ratios, respectively. During stage II, the average total normalized heat value was 43.20 ± 2.75 J/g for all samples. The conspicuously high normalized heat value of neat CaP in stage I, compared with an average of 33.91 ± 1.07 J/g for the other CaP composites, can be explained by the acid-base reactions between orthophosphoric acid from SHMP and Ca + and OH − alkalis from CAC, as well as generally faster chemical reaction rates in neat CaP than in reactions of cement hydrate formations. stage II values were attributable to the hydration reactions of CAC, leading to the formation of katoite and gibbsite as end products. The average of total normalized heat release for all slurries was 82.15 ± 8.93 J/g. Although the contribution of CAC hydration to stage I heat release cannot be excluded completely, for the most part, CaP cement had two different cement-forming pathways induced by acid-base and hydration reactions. FCS pozzolanic reactions could have contributed to stage II heat release.   At an isothermal temperature of 50 • C (Figure 26), the calorimetry curve of neat CaP was characterized by a striking increase in normalized heat release during the first stage that extended to~2 h from the start of measurement, and extremely low heat release during the second stage. This suggests that temperature increase accelerated CAC hydration reactions so that acid-base and hydration reactions happened during stage I, whereas cement hydration during stage II was minor. Increasing the temperature to 50 • C shortened the onset time of the neat cement hydration reactions from approximately 4 h 30 min at 25 • C to within 2 h. In contrast, heat release increased for all CaP/HSA composites during both stages, compared with those at 25 • C. Interestingly, the acid-base reaction heat release rate depended on the FCS/HSA ratios, namely, a decrease in the proportion of FCS to HSA enhanced heat release values. In fact, 61.76 J/g of an 87/13 ratio had 23.9 and 18.3% higher values than that of 100/0 and 90/10 ratios, respectively. As described in the CaP composite slurry properties, pH value increased with a decreasing FCS/HSA ratio. Thus, an 87/13 ratio with the highest pH of 10.58 increased the acid-base rection rate. An explanation for this could be that alkalinity was partially consumed by FCS during PR. This could have led to lower pH values in slurries with higher FCS content. The average value of total normalized heat release for all slurries was 98.91 ± 5.60 J/g.   At 85 • C (Figure 27), there was a major shift in reaction onset time for stage II to an earlier time. As a result, both stages I and II started within the first 3 h after starting measurement. For neat CaP cement, the rate of the acid-base reaction during stage I further increased by 17.4% to 109 J/g, compared with that at 50 • C, whereas the heat release rate during stage II was similar to that at 50 • C. In contrast, for all composites, heat release during stage I markedly declined. Furthermore, this value showed a downward trend with decreasing FCS/HSA ratios. This phenomenon was reversed in comparison with results at 50 • C, whereas the heat release values during stage II considerably increased. For instance, the average value for all the composites, 82.45 ± 6.12 J/g, was 1.73 times greater than that at 50 • C. Moreover, stage I duration was~3 h at 50 • C, but at 85 • C it was only~45 min. The average of total heat release for all slurries was the highest in this series of experiments: 102.45 ± 12.06 J/g.

Calorimetric Study of CaP Composites in Hydration
In this experiment, all the samples were prepared at 25 • C for 24 h prior to a steam pretreatment at 85 • C, following by an autoclave final treatment. Thus, the calorimetric results at 25 • C provided us with information on the two-stage chemical reaction pathway, with acid-base (first) and hydration (second) reactions. The first reaction occurred within the elapsed time of~80 min from the start of measurement, whereas the second reaction happened between~3 and~15 h.

Compressive Strength and Toughness of 100 and 250 • C-Autoclaved CaP Cement Composites before and after Thermal Shock (TS) Tests
Next, we evaluated the thermal shock (TS) resistance of the MGF-and MCF-reinforced composites of CaP cement with various FCS/HSA ratios after two curing temperatures of 100 and 250 • C. The samples were subjected to three cycles of TS, with one cycle being 175 • C heat → 25 • C water quenching for 100 • C-autoclaved samples, and 250 • C → 25 • C water quenching for 250 • C-autoclaved samples. Figures 28 and 29 present the changes in compressive strength and toughness of 100 • C-autoclaved composites as a function of the FCS/HSA ratio for pre-and post-TS test samples. The pre-TS test sample with a 100/0 FCS/HSA ratio had a compressive strength of 12.6 MPa, compared with~29 MPa in the neat CaP; this low strength was due to a much lower volumetric fraction of CAC in the composite systems. As evident from the figure, compressive strength tends to decline as the mass of HSA substituting FCS increases. The strength of the 87/13 ratio FCS/HSA was as low as 2.7 MPa, corresponding to ã 79% reduction compared with the 100/0 ratio composite. After TS tests, the 100/0 ratio composite exhibited a~10% reduction in strength. This strength reduction was similar for composites with 94/6 and 90/10 ratios. In contrast, the 87/13 ratio composite showed the lowest reduction rate of~4%. The average reduction for all samples including 100/0, 94/6, 90/10, 88/12, and 83/17 ratios was 13.30% ± 3.35.
For pre-TS samples, the toughness of 100/0 ratio was 0.51 N-mm/mm 3 , which was 33% higher than that of neat CaP, emphasizing a large improvement in toughness by MGF and MCF. Even the low-strength 87/13 ratio sample had a toughness of 0.24 N-mm/mm 3 , which was similar to that of the CaP/FCS system. After TS, the average toughness reduction in all samples was 28.61% ± 4.90, corresponding to~2.2 times greater value than that of average compressive strength reduction. Despite a large reduction in toughness, the toughness values of 100/0, 96/4, and 90/10 ratios were greater than those for the CaP/FCS system. Furthermore, the toughness of 0.17 N-mm/mm 3 for the 83/17 ratio sample was equivalent to a toughness improvement of~55% compared with the CaP/HSA system, due to the reinforcements that minimized the considerable losses in toughness caused by the incorporation of HSA.
MCF. Even the low-strength 87/13 ratio sample had a toughness of 0.24 N-mm/mm which was similar to that of the CaP/FCS system. After TS, the average toughness reduction in all samples was 28.61% ± 4.90, corresponding to ~2.2 times greater value than that of average compressive strength reduction. Despite a large reduction in toughness, the toughness values of 100/0, 96/4, and 90/10 ratios were greater than those for the CaP/FCS system. Furthermore, the toughness of 0.17 N-mm/mm 3 for the 83/17 ratio sample was equivalent to a toughness improvement of ~55% compared with the CaP/HSA system due to the reinforcements that minimized the considerable losses in toughness caused by the incorporation of HSA.  The extent of strength reduction depended on the FCS/HSA ratios. As mentioned earlier, this strength decrease may be caused by two factors. One factor could be the phase transitions occurring in the cement, namely, HOAp becomes the major phase at 250 °C The other is the gibbsite to boehmite phase transition. After a TS with the temperature gradient of ~225 °C, a small reduction (average 8.85% ± 2.9) in strength was observed strongly suggesting that all composites possessed an adequate compressive strength for       The extent of strength reduction depended on the FCS/HSA ratios. As mentioned earlier, this strength decrease may be caused by two factors. One factor could be the phase transitions occurring in the cement, namely, HOAp becomes the major phase at 250 • C. The other is the gibbsite to boehmite phase transition. After a TS with the temperature gradient of~225 • C, a small reduction (average 8.85% ± 2.9) in strength was observed, strongly suggesting that all composites possessed an adequate compressive strength for TS resistance. As for toughness, the pre-TS samples had a toughness similar to that of the 100 • C-autoclaved samples. This suggests that, regardless of the precipitation of N-A-S-H as a PR product of MGF at 250 • C, the MGF/MCF dual reinforcement aided in providing adequate toughness. However, TS for all samples engendered a striking reduction in toughness, with an average value of 56.14% ± 3.23, which was more than six times higher than the average reduction in compressive strength. As adequate toughness comes from a proper balance in compressive strength and ductility, this result suggests that the ductile nature of the lightweight composites was likely transformed to a brittle one at this TS temperature gradient. Thus, the development of a densified microstructure brought about by excessive PRs of FCS, HSA, and MGF in composites may cause brittleness at this TS heat temperature. For FCS, the excessive PR may lead to development of a strong, dense bond structure in cement/FCS shell interfacial boundary regions. For HSA, the PR may induce morphological alterations, such as the transformation of original nanoporous structure to a micro-porous one, whereas for MGF, the ductile fibers may be converted by PR into brittle ones. Although these composites demonstrated good strength resistance to TS, they did not sustain an adequate toughness following the TS tests. Figures 32 and 33 indicate the changes in thermal conductivity (TC) and watersaturated bulk density (WBD) as a function of the FCS/HSA ratio for 100-and 250 • Cautoclaved pre-and post-TS samples. For 100 • C-autoclaved pre-TS samples, the TC value decreased with the reduction of WBD. The 100/0 ratio sample had a TC of 0.56 W/mK with a WBD of 1.53 g/cm 3 . In contrast, the 83/17 ratio sample with the lowest WBD of 1.34 g/cm 3 displayed a desirable decrease in the TC value by~43% to 0.32 W/mK, due to the increased HSA content. Interestingly, all post-TS samples revealed decreases in both TC and WBD values. The 250 • C-autoclaved samples had two noticeable characteristics: (1) Further reduced TC and WBD values for 100/0, 94/6, and 90/10 ratios; (2) An upward trend of TC for 88/12 and 83/17 ratios. The 90/10 ratio sample had the lowest TC of 0.29 W/mK while exhibiting good TS resistance. The upward TC trend, on the other hand, resulted in 0.41 and 0.43 W/mK TC of the 88/12 and 83/17 ratios, respectively, which was tantamount to a~21 and~27% increase in TC. This implied that the incorporation of excessive HSA could lead to such undesirable results. The degradation of sample insulation performance was caused by the disintegration of HSA at 250 • C. As explained above, the high temperature degradation of HSA resulted in the development of microstructures with large pores. The loss of insulating air and water ingress through the porous structure led to an increased WBD. To support this, Figure 34 shows photomicrographs of 250 • Cautoclaved 90/10 and 83/17 ratio composites. The abundance of large cavities that have developed due to the decomposition of HSA is clearly visible in the 83/17 sample images, whereas the 90/10 sample shows few of these cavities. Small craters of~100 µm size were formed by the debonding of FCS particles from the cement matrix. Such a hydrophilic porous structure allows water to permeate readily through the cement matrix, resulting in the increase of TC and WBD. in TC. This implied that the incorporation of excessive HSA could lead to such undesirable results. The degradation of sample insulation performance was caused by the disintegration of HSA at 250 °C. As explained above, the high temperature degradation of HSA resulted in the development of microstructures with large pores. The loss of insulating air and water ingress through the porous structure led to an increased WBD. To support this, Figure 34 shows photomicrographs of 250 °C-autoclaved 90/10 and 83/17 ratio composites. The abundance of large cavities that have developed due to the decomposition of HSA is clearly visible in the 83/17 sample images, whereas the 90/10 sample shows few of these cavities. Small craters of ~100 µ m size were formed by the debonding of FCS particles from the cement matrix. Such a hydrophilic porous structure allows water to permeate readily through the cement matrix, resulting in the increase of TC and WBD.      (Figure 35), the locations of all absorption bands, except for a weak band at 914 cm −1 ascribed to MGF, were similar to those of the neat CaP cement. In contrast, the spectra of 250 °C-autoclaved samples ( Figure 36) showed emergence of two new bands at 1256 and 917 cm −1 . The former band may be representing amorphous silica gel [91] from HSA. The peak intensity (height, ∆A) of this band for the 87/13 ratio sample was much greater than for the 100/0 and 90/10 ratio samples: 0.0216 ∆A of the 87/13 ratio was equivalent to a 54.3% increase from the averaged ∆A of the 100/0 and 90/10 ratios, indicating that a large free silica gel was liber-  (Figure 35), the locations of all absorption bands, except for a weak band at 914 cm −1 ascribed to MGF, were similar to those of the neat CaP cement. In contrast, the spectra of 250 • C-autoclaved samples ( Figure 36) showed emergence of two new bands at 1256 and 917 cm −1 . The former band may be representing amorphous silica gel [91] from HSA. The peak intensity (height, ∆A) of this band for the 87/13 ratio sample was much greater than for the 100/0 and 90/10 ratio samples: 0.0216 ∆A of the 87/13 ratio was equivalent to a 54.3% increase from the averaged ∆A of the 100/0 and 90/10 ratios, indicating that a large free silica gel was liberated from the disintegrated HSA at 250 • C. As the 917 cm −1 band comes from amorphous and crystalline hydrated and anhydrous N-A-S(-H), C,N-A-S(-H), and C-A-S(-H), the striking increase of this band's intensity for the 87/13 ratio sample clearly manifested from the precipitation of large amounts of these phases; the increase in this peak by 0.1182 ∆A was 47.8% bigger than the averaged ∆A of the 100/0 and 90/10 ratios. This correlated with another intense peak at 663 cm −1 attributed to these phases, as aforementioned in "CaP/FCS System". Hence, PR of silica gels promoted the precipitation of these phases in the 87/13 ratio sample, but barely in the 100/0 and 90/10 samples. In other words, the pozzolanic activity of HSA in the 90/10 ratio sample was minor, similar to the case of FCS.  To identify crystalline phases, XRD analysis was conducted. Figure 37 shows the XRD pattern of the 250 °C-autoclaved 87/13 ratio composite. The three PR-related crystalline phases included two anhydrous phases (albite (N-A-S), a Na-plagioclase feldspar and dmisteinbergite (CaAl2Si2O8, C-A-S), which is a high-temperature hexagonal polymorph of anorthite, a Ca-plagioclase feldspar) and one hydrous phase (hydrosodalite zeolite (N-  To identify crystalline phases, XRD analysis was conducted. Figure 37 shows the XRD pattern of the 250 °C-autoclaved 87/13 ratio composite. The three PR-related crystalline phases included two anhydrous phases (albite (N-A-S), a Na-plagioclase feldspar and dmisteinbergite (CaAl2Si2O8, C-A-S), which is a high-temperature hexagonal polymorph of anorthite, a Ca-plagioclase feldspar) and one hydrous phase (hydrosodalite zeolite (N- To identify crystalline phases, XRD analysis was conducted. Figure 37 shows the XRD pattern of the 250 • C-autoclaved 87/13 ratio composite. The three PR-related crystalline phases included two anhydrous phases (albite (N-A-S), a Na-plagioclase feldspar and dmisteinbergite (CaAl 2 Si 2 O 8 , C-A-S), which is a high-temperature hexagonal polymorph of anorthite, a Ca-plagioclase feldspar) and one hydrous phase (hydrosodalite zeolite (N-A-S-H)). Among them, dmisteinbergite was the major PR product. Three other crystalline phases formed in CaP cement( katoite, HOAp, and boehmite) were also present as major phases. To support the above information, Figure 38 shows the photomicrograph and elemental composition of phases formed in the printing of disintegrated HSA on the freshly fractured surface of the 250 °C-autoclaved 87/13 ratio composite. The No. 1 location had two principal elements, O and Al, corresponding to an atomic O/A ratio of 1.51. This was likely a boehmite-rich region. The second location (No. 2) encompassed four major elements, O, Al, Si, and Ca, and the Ca/Al, Si/Al, and O/Al atomic ratios were 0.68, 1.19, and 2.28, respectively. These atomic ratios closely resemble those of dmisteinbergite: Ca/Al = 0.5, Si/Al = 1.0, and O/Al = 4.0. Interestingly, similar atomic ratios of Ca/Al = 0.70, Si/Al = 0.88, and O/Al = 2.48 were observed for the PR products formed on the surfaces of FCS shells where Ti and Fe signals came from the FCS itself ( Figure 39). Nevertheless, the presence of mullite in the XRD patterns suggests limited pozzolanic activity in FCS. Thus, most of Si likely came from the disintegrated has, resulting in the precipitation of dmisteinbergite. To support the above information, Figure 38 shows the photomicrograph and elemental composition of phases formed in the printing of disintegrated HSA on the freshly fractured surface of the 250 • C-autoclaved 87/13 ratio composite. The No. 1 location had two principal elements, O and Al, corresponding to an atomic O/A ratio of 1.51. This was likely a boehmite-rich region. The second location (No. 2) encompassed four major elements, O, Al, Si, and Ca, and the Ca/Al, Si/Al, and O/Al atomic ratios were 0.68, 1.19, and 2.28, respectively. These atomic ratios closely resemble those of dmisteinbergite: Ca/Al = 0.5, Si/Al = 1.0, and O/Al = 4.0. Interestingly, similar atomic ratios of Ca/Al = 0.70, Si/Al = 0.88, and O/Al = 2.48 were observed for the PR products formed on the surfaces of FCS shells where Ti and Fe signals came from the FCS itself ( Figure 39). Nevertheless, the presence of mullite in the XRD patterns suggests limited pozzolanic activity in FCS. Thus, most of Si likely came from the disintegrated has, resulting in the precipitation of dmisteinbergite.
The various crystalline phases formed in these composites may increase their TC (λ). For example, the λ of moist silica gel was reported to be in the range of~0.19 W/mK at 0.1 water/gel wt. ratio to~0.23 W/mK at 0.3 water/gel wt. ratio [92]. These λ values were considerably higher than the 0.012 W/mK of "as-received" HSA, corresponding to more than a fifteen-fold increase of the HSA. The thermal conductivity of~3.3,~2.0, and 1.7 W/mK for N-A-S-H hydrosodalite [93], albite, and dmisteinbergite [94], respectively, was~3.1-,~1.9-, and~1.6-fold higher than the~1.1 W/mK thermal conductivity of hydroxyapatite [95] and~1.05 W/mK of neat CAC hydrate containing C 3 AH 6 and AH 3 at a W/C ratio of 0.3 [96]. Nonetheless, these PR products may have not significantly affected composite TC. The boehmite that formed as a major phase of CaP cement at ≥175 • C has a very high λ of 30 W/mK. It is known as a high thermal conductor [97]. Thus, boehmite is an undesirable phase for thermally insulating cements. Nevertheless, in the case of the tested formulations, the products with high TC did not compromise the insulating nature of the composites In summary, the incorporation of an excessive amount of HSA into the FCS/HSA hybrid insulating system engendered a negative effect on the thermal insulating performance of the 88/12 and 83/17 ratio composites, raising λ and WBD values compared with the 90/10 ratio samples at 250 • C.  As in the case of albite, the orthosilicic acid (H4SiO4) derived from alkali (OH) − dissolution of silica gels can serve as a source of Si for dmisteinbergite formation [83]:  As in the case of albite, the orthosilicic acid (H4SiO4) derived from alkali (OH) − dissolution of silica gels can serve as a source of Si for dmisteinbergite formation [83]: Ca 2+ + 2Al(OH)4 − + 2H4SiO4 (aq) ↔ CaAl2Si2O8 + 8H2O.

Conclusions
Low TC and lightweight hydrophobic TS-resistant cement composites were formulated using a combination of FCS and HSA as insulating particles and CaP cement as a binder. The composites were cured at 100 or 250 • C and tested in three cycles of TS tests with temperature gradients of 150 or 225 • C (1 cycle: 175 • C heat for 24 h → 25 • C water quenching for 100 • C-autoclaved samples and 250 • C heat for 24 h → 25 • C water quenching for 250 • C autoclaved samples). Combining low-strength HSA with FCS and high-strength CaP cement binder resulted in an improvement in the mechanical properties of composites with very weak HSA particles and low TC values (below that of water) when measured under water saturated conditions. Further strength improvements were achieved with the use of reinforcement micro-glass and carbon fibers. As CaP cement has a lower pH (pH~10 in tested blends) than OPC-based blends (pH~13), this minimized pozzolanic reactions that lead to the degradation of solid FCS shells, release of insulating gasses, and the loss of cement integrity. The hydrophobic nature of HSA treated with HMDS provided cements with water-repellant properties, preventing water ingress and TC increase.
As a result, the optimized 90/10 FCS/HSA ratio samples after TS tests possessed a TC of 0.35 and 0.28 W/mK for 100 and 250 • C, respectively. However, at 250 • C, trimethylsilyl groups of HMDS underwent hydrothermal degradation, followed by trimethylsilyl-free silica aerogel alkali dissolution and development of a porous microstructure with large cavities in place of the dissolved HSA. This resulted in a 43% loss in strength compared with the 100-and 175 • C-made samples. Considering the hydrothermal degradation of the hydrophobic trimethylsilyl groups present on silica aerogel surfaces at 250 • C, the 90/10 FCS/HSA CaP composite has potential to be used as a thermally insulating, thermally shock-resistant well cement in mid-temperature range (~100 to 175 • C) geothermal energy storage systems. The compressive strength of more than 7 MPa for this lightweight composite persisted after three TS tests at 175 • C.
Author Contributions: All authors (T.S. and T.P.) contributed to conceptualization, methodology, validation, formal analysis, investigation, writing, project administration, and funding acquisition. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.