Characterization of Solid Particulates to Be Used as Storage as Well as Heat Transfer Medium in Concentrated Solar Power Systems

Abstract

Using solid particulates as a heat transfer medium for concentrated solar power (CSP) systems has many advantages, positioning them as a superior option compared with conventional heat transfer media such as steam, oil, air, and molten salt. However, a critical imperative lies in the comprehensive evaluation of the properties of potential solid particulates intended for utilization under such extreme thermal conditions. This paper undertakes an exhaustive examination of both ambient and high-temperature thermophysical properties of four naturally occurring particulate materials, Riyadh white sand, Riyadh red sand, Saudi olivine sand, and US olivine sand, and one well-known engineered particulate material. The parameters under scrutiny encompass loose bulk density, tapped bulk density, real density, sintering temperature, and thermal conductivity. The results reveal that the theoretical density decreases with the increase in temperature. The bulk density of solid particulates depends strongly on the particulate size distribution, as well as on the compaction. The tapped bulk density was found to be larger than the loose density for all particulates, as expected. The sintering test proved that Riyadh white sand is sintered at the highest temperature and pressure, 1300 °C and 50 MPa, respectively. US olivine sand was solidified at 800 °C and melted at higher temperatures. This proves that US olivine sand is not suitable to be used as a thermal energy storage and heat transfer medium in high-temperature particle-based CSP systems. The experimental results of thermal diffusivity/conductivity reveal that, for all particulates, both properties decrease with the increase in temperature, and results up to 475.5 °C are reported.

1. Introduction

The particle heating receiver (PHR), which uses solid particles as thermal energy storage (TES) and heat transfer media (HTM), has recently attracted many researchers’ attention for its use in concentrated solar power (CSP) systems. Many of the drawbacks of the common HTMs that are now in use, such as air, steam, molten salt, and synthetic oil, are avoided with the use of such receivers. Using air as an HTM requires excessively large receivers and heat exchangers [1]. Furthermore, the pressure drop through the receiver and heat exchanger piping is an additional issue; so, pumping becomes problematic and therefore increases the investment cost. Despite the high thermal capacity, low cost, and availability of water, and its capability to be used as a direct steam source to feed the power unit, its drawbacks are oxidization, freezing, and difficulty in storing steam at high pressure [2]. The advantages of oil are high operating temperatures, low specific volume, and low pressure [1]. The main issue with heavy oils is that typical hydrocarbons decompose at around 400 °C, which, in turn, limits their operating temperature and therefore restricts their application in low-efficiency CSP cycles. Other disadvantages of oils are deterioration over time, expense, and flammability.

In current molten salt commercial solar tower power plants, the molten salt used is a 40/60% mixture of KNO₃/NaNO₃ [3]. The operational temperature range, as stated previously, is between 290 °C and 565 °C. With the highest temperature of 565 °C, only a steam cycle can be used, and more efficient power cycles cannot be integrated since their operational temperature is higher than 565 °C. Other commonly considered molten salts include carbonate salts, which have a higher stability limit (>650 °C) and a higher melting point (<390 °C), but a relatively higher price. Fluoride salts have an even higher stability limit (>700 °C), but they are toxic, relatively expensive, and have a higher melting point (>450 °C) [4]. The U.S. Department of Energy (DOE) is supporting the development of a chloride salt, which has high heat capacity but a high melting point (>750 °C) [5,6,7]. However, increasing the operational temperature range for molten salts brings additional challenges. In particular, it increases the corrosiveness of the material. This limits the lifetime of the components and increases the levelized cost of electricity (LCOE) of the plant.

Particulate materials are found to be a good alternative to molten salts both as the HTM and the TES in higher-temperature systems for multiple reasons, including their thermal conductivity not limiting their capacity to store heat [8]. Most importantly, silica sand and similar particulates have demonstrated exceptional stability well beyond 1000 °C [9], which is far above the decomposition thresholds of commonplace nitrate salts [10,11] and organic fluids [12]. In addition, in comparison to conventional HTMs, solid particulates (1) simplify energy storage without the need for separate tanks or heat exchangers; (2) when natural, are abundantly available at a low cost; (3) undergo no toxic decomposition and carry no environmental risk; and (4) do not require freeze protection. Recent endeavors, backed by the US DOE, have showcased the viability of natural particulates (sand) as both an HTM and TES in CSP systems, particularly through the direct heating of sand in PHR systems [13,14]. However, the particulates’ thermophysical properties, heat resistance, crack resistance, hazards, and cost should be taken into account since the particulates will be stored in a larger amount in TES to be used during off-sun times.

Various researchers have suggested many potential solid particulates as an HTM/TES, like quartz, sand, cristobalite, bauxite, alumina, ceria, graphite, olivine, silicon carbide, calcined flint clay, etc. [15,16]. A comparison of the thermophysical properties and cost of some solid particulates is shown in

Table 1. Comparison of some thermophysical properties and the cost of solid particulates. Adapted from [15,16,17,18,19,20,21,22,23].

Material	Specific Heat (kJ/kgK)	Melting Point (°C)	Sintering Temperature (°C)	Energy Density (kJ/m3)	Density (kg/m3)	Thermal Conductivity (W/mK)	Cost ($/ton)
Quartz; SiO2	0.95	1670	1670–1714	2480	2600	0.4	150
Sand; SiO2	0.95	1700	1760	2490	2550	1.2	75
Cristobalite; SiO2	1.05	1713	1714	2447	2330	0.48	200
Sintered Bauxite	1.05	2000	1550	4095	3600	0.65	400
Silicon carbide; SiC	1.05	2210	1800	3380	3200	100	1900
Alumina; Al2O3	1.05	2054	2072	4200	3970	23	300
Olivine; (Mg,Fe)2SiO4	1.25	1870	1450	4290	3300	0.56	175
Ceria; CeO2 [17]	0.39	2340	1000–1400	-	7600	12	1610
Graphite; C	0.79	3652	900–1500	1131	2200	45	362
Calcined Flint Clay; Al2O3/SiO2 [18,19]	0.74–0.87	1500	1200	2730	3300	0.85	180

.

Numerous investigations have been carried out to understand how particles affect the receiver’s performance. The first to employ calcined coke gemstone solid particles as the HTM in solar receivers was the National Renewable Energy Laboratory [5]. The study conducted by our collaborators at the Georgia Institute of Technology, and it investigated how temperature and pressure affected the particles’ sintering characteristics [24]. However, the sintering study was limited to temperatures up to 800 °C.

Numerous experts have studied the specific heat of solid particles utilized in next-generation CSP reactors [25,26,27,28,29]. They either measured or calculated the specific heat at high temperatures up to a maximum of 800 °C. Kang et al. [21] calculated the thermal conductivity of various particulates up to 1000 °C. Nonetheless, their calculation was based on a predictive model. Lv et al. [29] measured the thermal conductivity of only desert sand. Thermal conductivity was measured at temperatures close to ambient conditions only.

An obstructed flow particle heating receiver (PHR) using sand was designed and developed by King Saud University (KSU) and Georgia Institute of Technology (GIT) [30,31]. Several on-sun tests were conducted in the 300 kW_th particle-based CSP plant at KSU [32,33]. Riyadh red sand (RRS) and Carbobead (CB) were used as HTMs. The system demonstrated that the particle temperature at the outlet of the PHR reached 720 °C. Our team is conducting ongoing de-risking activities to ensure the reliability, efficiency, and economic viability of particle-based CSP systems before their widespread commercial deployment [34,35]. Based on this initiative, a 1.3 MWe pre-commercial plant will be built by our team in Waad-Al-Shamal, Saudi Arabia [36].

Many studies have been carried out by researchers worldwide, as seen in the literature reviewed herein. Nonetheless, those studies are somewhat limited. In particular, no study has been found that experimentally investigated thermal conductivity at high temperatures and sintering at both high temperatures and pressures for many particulate materials of interest. Since the authors are the developers of an innovative CSP system [37], investigation of these properties is needed. Consequently, this paper provides the experimental measurements of the sintering behavior at high pressure and temperature, thermal conductivity at high temperature, and other pertinent properties of the candidate particulate materials. Four naturally occurring particulate materials, namely, Riyadh white sand (RWS), Riyadh red sand (RRS), Saudi olivine sand (SOS), and US olivine sand (USOS), and one engineered particulate material called Carbobead (CB) have been investigated. RWS, RRS, and SOS are naturally occurring sands readily available in Saudi Arabia, making them cost-effective and practical for large-scale CSP systems. CB is a commercially engineered ceramic particle, widely used in energy applications due to its high solar absorptance, thermal stability, and durability. CB allows for benchmarking the performance of natural particulates against a well-established reference material. The inclusion of olivine sands from different sources (SOS vs. USOS) adds insight into regional variations in natural materials and their potential impact on CSP performance. Loose bulk density, tapped bulk density, real density, sintering temperature, and thermal diffusivity have been measured.

2. Methodology

All natural particulate materials were sifted to the targeted particle size for the 1.3 MWe pre-commercial plant that will be built in Waad-Al-Shamal, Saudi Arabia [36]. Figure 1 shows the candidate particulate materials investigated in this study. Two different particle sizes of RWS were tested, 210 to 425 µm and 500 to 1000 µm. CB, SOS, and USOS were used as received. CB is an engineered material that has uniform sphericity and diameters [38,39,40,41]. After sifting, all particulate materials were washed with water and then dried to remove dirt and any other contaminants. Chemical compositions of the particulates were either obtained from the literature or measured experimentally and are provided in

Table 2. Particulate materials’ properties.

Sample	Average Particle Diameter [µm]	Shape	Chemical Composition
RWS	500 to 1000	Irregular	98% SiO2, 1.56% Al2O3, 0.44% others
RWS	210 to 425	Irregular	98% SiO2, 1.56% Al2O3, 0.44% others
RRS	210 to 425	Irregular	90% SiO2, 2.93% Al2O3, 3.15% Fe2O3, 2.23% CaSiO3, and 1.69% others
SOS *	200 to 500	Irregular	59.32% O, 22.91% Mg, 12.91% Si, 4.87% Fe
USOS	250 to 400	Irregular	50.2% SiO2, 31.4% MgO, 15.9% Fe2O3, 1.6% Al2O3, 0.7% CaO, <0.1% Ni & NiO, 0.1% Cr2O3 and <0.1% Trace Elements & Compounds
CB	300	Regular	75% Al2O3, 11% SiO2, 9% Fe2O3, 3% TiO2, and 2% others [42]

* Energy-dispersive X-ray analysis (Jeol Limited, Akishima, Tokyo, Japan, EDX), together with a scanning electron microscope (Jeol Limited, Akishima, Tokyo, Japan, SEM) both by Jeol Limited (Akishima, Tokyo, Japan), were used to identify the elemental composition of the SOS materials. Qualitative and quantitative analyses were conducted to identify the types of elements that were present, as well as the percentage of each element’s concentration within the sample.

. It should be noted that all natural particulate materials were sifted to match the particle size mentioned in Table 2. The effect of the particle size was investigated on RWS by conducting the sintering test on two different particle sizes, 210 to 425 µm, and 500 to 1000 µm. In contrast, CB is an engineered material that has a uniform circular size and diameter; so, sifting is unnecessary and would be redundant.

2.1. Density Measurement

Two types of bulk density for all particulate materials were measured in this study: loose particle density, and tapped bulk density. The volume of the solid particles was measured using a high-precision graduated Pyrex beaker, while the mass of the particles was determined using a high-accuracy lab scale (PLJ 4000-2M), manufactured by KERN & Sohn GmbH, Baden-Württemberg, Germany [43]. Tapped density was obtained by knocking on the beaker until the volume occupied by the particles became fixed and there was no further change with knocking. In the case of loose density, the particles were placed in the beaker; then, the volume was measured as it was, without knocking, and then the same procedures with the calculation of tapped density were applied. The measurements were repeated seven times to increase the reliability and limit the uncertainty.

By filling the voids between the solid particles with distilled water, Archimedes’ method [44] was utilized to measure the theoretical density of particles (also called real density). The density of the distilled water was measured at various temperatures ranging from room temperature to 70 °C. Subsequently, the particle sample was gently poured into distilled water, ensuring that the particles were fully submerged and the formation of air bubbles between the distilled water and the particles was avoided. The sample was weighed once again. By deducting the measured water mass from the overall mass, the particle mass was computed. Additionally, by deducting the measured water volume from the total volume, the particle volume was computed. In the end, the particle mass divided by the particle volume yielded the particle theoretical density.

2.2. Sintering Test

Sintering of particles into clumps or larger masses would be detrimental in a CSP system since small clumps of particles can clog the PHR, heat exchanger, or other processes, and any large agglomeration of mass in the high-temperature TES would quickly disrupt and deactivate the entire system. The most likely situations for sintering are in the highest-temperature parts of the PHR and in the highest-pressure lower level of the high-temperature TES. In commercial-scale TES bins, pressures up to 1 MPa or more should be expected. Since sintering is promoted by pressure, the TES conditions are likely the most critical. At present, the actual TES pressures cannot yet be reliably specified, nor can the expected residence time for particles in this TES. In any case, long-duration testing of numerous candidate particles at moderate pressures is not currently feasible. Consequently, the development teams have conducted rapid high-temperature and high-pressure testing to quickly screen candidate HTM materials.

To investigate the sintering of particulates, the samples were tested in a High-Frequency Induction Heat Sintering Furnace (HFIHS), as shown in Figure 2. In this device, the samples are heated to a desired temperature and then rapidly compressed to a specified pressure. Since this type of testing is relatively unfamiliar, the apparatus and procedure are described below in some detail.

In this apparatus, the test sample is placed in a robust high-temperature- and high-pressure-resistant cylindrical graphite die. The inner diameter is 12.7 mm in this case. Two plungers, one from each end, are utilized to consolidate the tested sample, as shown in Figure 2a). To prepare for testing, a small, 1.0 g sample of the solid particulate is loaded into the die, and the plungers are inserted until they touch the sample. Then, the loaded die is placed in the sintering chamber of the HFIHS, as shown in Figure 2c, and the chamber is evacuated. At this point, the apparatus is ready for testing.

In a typical test, the electrically conductive graphite die, along with the sample, is rapidly heated (100–150 °C/min) by induction from the ambient temperature to the test temperature. To prevent actual melting, the temperature in the chamber is typically kept at less than 75% of the melting point of the material being processed. Subsequently, uniaxial pressure is applied by the piston seen in Figure 2 in 10 kPa steps until the ultimate test pressure, 20–50 MPa, is reached or until sintering is automatically detected. Sintering is detected by sudden piston movement. At this point, the test is terminated, and pressure is relieved. The typical test time is 10 to 15 min.

The sintering test was conducted at different high temperatures (800, 1000, 1100, 1200, and 1300 °C) and pressures (20, 30, 40, and 50 MPa). For each particulate, testing was carried out at each selected temperature for a range of pressures until a temperature and pressure were found to be causing sintering. Partially consolidated pellets from the sintering process were also mechanically finished and used in the thermal diffusivity testing described next since that apparatus is not intended for loose particles. We use the term “Partially consolidated” because, normally, the HF sinter system (Figure 2) produces completely solidified pellets for most of the samples for other applications. However, in the case of solid particulates, the produced pellets were porous masses. It is important to mention that, while making the pellets, we reached the apparatus’s maximum limits of pressure and temperature.

2.3. Thermal Conductivity Measurement & Calculation

The bulk thermal conductivity of the candidate particulate materials at ambient temperatures was measured using a KD2 Pro thermal conductivity analyzer supplied by Decagon Devices Inc., Pullman, WA, USA [45]. This instrument implements the transient line source measurement. This instrument was calibrated using distilled water at ambient temperature. This test was conducted at ambient temperature. The particulate sample was first filled in a glass cylinder and then a KD2 Pro needle was inserted into the sample. The reading was recorded once the analyzer completed the transient measurement and related calculations. For the accuracy of the results, three measurements were carried out at different times.

In the case of determining the bulk thermal conductivity at high temperatures, an indirect method was used. The thermal diffusivity of the samples was measured first by the transient laser flash method [46,47]. This apparatus is not suited for loose particles; so, sintered pellets or discs were used. The pellets, which were agglomerated during the sintering process, were in a state intermediate to the solid material and the loose particles, but these were the best available upper-bound approximation to the conductivity of particle beds in a CSP system. To this effect, particulate samples were slightly sintered at high temperature and pressure in the HFIHS, as described above, to obtain a disc-shaped pellet of bonded particles. The sintered samples were then sized to the pan dimensions of the machine by cutting and polishing the samples using a Dremel tool 4000-4/34 by KVM Tools Inc., Cypress, TX, USA [48] and fine-grit SiC abrasive paper, respectively. Bulk density at ambient temperature was measured using an Electronic Densimeter SD 200L by Alfa Mirage Co. Ltd., Osaka, Japan [49] with a density resolution of up to 0.0001 units. Thermal diffusivity was measured using a laser flash apparatus (LFA 457) by Netzsch, Selb, Germany [46,47]. It was measured at different temperatures starting from ambient temperature up to 475 °C.

Thermal conductivity (k) in W/mK was then calculated as follows:

$$k=\alpha \rho C_p$$

(1)

where ρ is the density in kg/m³, α is the thermal diffusivity in m²/s, and C_p is the specific heat in J/kgK.

Since silica (SiO₂) is by far the dominant ingredient of both RRS and RWS, it is assumed that the specific heat for both of them is the same as that of SiO₂. To calculate the specific heat of SiO₂, Chase [50] suggested the following relation:

$$C_{p_{\mathrm{RWS},\mathrm{RRS}}}=\mathrm{A}+\mathrm{B}T+\mathrm{C}{T}^2+\mathrm{D}{T}^3+\mathrm{E}/T^2$$

(2)

where A = −101.134, B = 4188.707, C = −5405.677, D = 2805.398, E = 0.042407, and T is the particulate temperature in Kelvin/1000. $C_{p_{\mathrm{RWS},\mathrm{RRS}}}$ represents the specific heat of RWS and RRS in J/kgK.

At Sandia National Laboratories, the specific heat of CB was measured with the application of differential scanning calorimetry. Ho et al. [51] developed the following relation for the calculation of the specific heat of CB:

$${\mathrm{C}}_{p_{\mathrm{CB}}}=0.365 T^{0.18}$$

(3)

where ${\mathrm{C}}_{p_{\mathrm{CB}}}$ is in kJ/kgK, and T is the CB temperature in K, which falls between 50 °C and 1100 °C.

The specific heat of SOS was measured using a STA 409C/CD apparatus by Netzsch, Selb, Germany [52], which is based on differential scanning calorimetry. The measurements were carried out from room temperature to 1200 °C.

These measurements and research sources resulted in the needed density, specific heat, and thermal diffusivity data.

3. Results and Discussion

3.1. Bulk Density

The results of loose density (ρ_l), tapped density (ρ_t), and theoretical density are summarized in

Table 3. Measurement results of loose and tapped density of particulate materials at ambient temperature.

Material Type	Theoretical Density ${\mathit{\rho }}_{\mathit{th}}$ [kg/m3]	Loose Density ${\mathit{\rho }}_{\mathit{l}}$ [kg/m3]	Tapped Density ${\mathit{\rho }}_{\mathit{t}}$ [kg/m3]	Porosity [38,42] $\left(1-\frac{{\mathit{\rho }}_{\mathit{b}}}{{\mathit{\rho }}_{\mathit{th}}}\right)$ × 100
RWS	2756 ± 28.7	1571 ± 8.8	1653 ± 37	43	Loose
				40	Tapped
RRS	2771 ± 29	1574 ± 8.8	1679 ± 38	43	Loose
				39	Tapped
SOS	3260 ± 29.7	1873 ± 9.3	2155 ± 37	43	Loose
				34	Tapped
USOS	1913 ± 29.2	1445 ± 8.6	1572 ± 36	24	Loose
				18	Tapped
CB–0.3 mm	3321 ± 36.6	1796 ± 10.2	1944 ± 43	46	Loose
				41	Tapped
CB–0.5 mm	3321 ± 42.3	1813 ± 10.3	1828 ± 44	45	Loose
				45	Tapped

. The results show that the bulk densities of solid particles depend somewhat on the particulate size distribution, as well as on the compaction. The loose bulk density of CB with the small particle size (0.3 mm) is slightly lower (around 0.9% lower) than the tapped bulk density of the 0.5 mm beads, indicating that the larger and nearly spherical 0.5 mm beads pack slightly better under gravity alone. In contrast, the tapped bulk density of the 0.3 mm CBs is noticeably greater (by around 6.0%), indicating that the smaller CBs may be slightly less uniform in size and shape and that the surface friction between the smaller beads may deter close packing under gravity alone. In all cases, the porosity of the tapped CB beads (41% and 46%) is reasonably close to but not less than the expected porosity of randomly packed spheres (around 36%). This behavior is expected for the nearly spherical shape of these artificial particles. Generally, for all particulate materials, the results showed that the tapped bulk density is at least somewhat larger than the loose density for solid particulates of a small size. This is attributed to the rearrangement of particles after disturbance thereby decreasing the inter-particle voids.

Comparing the measured theoretical density results in Table 3 with those in Table 1, which are adapted from [16,17,18,19,20,21,22,23], it is clear that for RWS and RRS, the theoretical densities are slightly higher than that of pure SiO₂. This is likely due to the small quantities of heavier impurities (like iron oxides) in the natural particulate samples. The density of SOS is almost identical to that of pure olivine [17], indicating that it has minimal impurities. However, the density of USOS is much smaller. That is due to a large proportion of impurities in the sample. The loose density measured by [53] for typical sand is 1450 kg/m³, whereas that for RWS is 1571 kg/m³. The higher loose density for RWS is due to its 98% SiO₂ content, whereas that in [53] is 91 to 97%. The other reason for the deviation is that [53] used relatively bigger sizes of particles up to 465 μm. The porosity measured in [38] for CB is almost the same as that measured in this study.

3.2. Sintering of Candidate Materials

Figure 3 shows the pellets after sintering and machining, demonstrating that some particulate nature has been retained. The thickness of the pellets was averaged for each specimen by measuring their thickness at three different places along the circumference. These measurements are shown in

Table 4. Dimensions and density of the solid pellets produced by the sintering test.

Material Type	Average Diameter [mm]	Thickness [mm]	Pellet Density [kg/m3]
RRS	12.60	2.93	2454 ± 13.3
RWS	12.60	3.52	2581 ± 16.7
SOS	12.65	2.96	3169 ± 28.7
CB	12.65	3.36	3210 ± 18.2

. All pellets were then coated with highly absorptive graphite, as required for this measurement.

Table 5. Sintering results of the candidate particulate materials at different temperatures and pressures.

Test No.	Sintering Temperature [°C]	RWS (500 to 1000 µm)			RRS			SOS			USOS			CB
		Sintering Pressure [MPa]	Dwell Time [min]	Sintering	Sintering Pressure [MPa]	Dwell Time [min]	Sintering	Sintering Pressure [MPa]	Dwell Time [min]	Sintering	Sintering Pressure [MPa]	Dwell Time [min]	Sintering	Sintering Pressure [MPa]	Dwell Time [min]	Sintering
1	800	-	-	-	-	-	-	-	-	-	20	5	Yes **	-	-	-
2	1000	20	10	No	20	10	No	20	10	No	20	5	Yes **	20	5	No
3	1100	-	-	-	-	-	-	-	-	-	20	5	Fused ***	20	7	Yes
4	1200	20	10	No	20	10	No	20	10	No	-	-	-	-	-	-
5	1300	30	5	No	-	-	-	-	-	-	-	-	-	-	-	-
6	1300	40	5	No	-	-	-	-	-	-	-	-	-	-	-	-
		RWS (210 to 425µm)			-	-	-	-	-	-	-	-	-	-	-	-
7	1300	20	4	No	20	5	No	20	3	No	-	-	-	-	-	-
8	1300	30	4	No	30	1	No	30	3	No	-	-	-	-	-	-
9	1300	40	4	No	40	1	Yes	40	3	No	-	-	-	-	-	-
10	1300	50	4	Yes *	-	-	-	50	3	Yes	-	-	-	-	-	-

* A weak bond was determined between the particles of RWS; ** incompletely sintered with significant cracks; *** SOS melted and formed lumps; so, it was excluded from further tests due to its undesired sintering behavior, which makes it unsuitable for particle-based CSP.

summarizes the sintering behavior of the selected particulate materials—RWS, RRS, and SOS—under varying temperatures and pressures. The table also details the experimental protocols employed for the sintering process. The investigation highlights the critical role of choosing particulate materials with minimal sintering potential to enhance the thermal performance and operational reliability of particle-based CSP systems.

These screening results for RWS, RRS, and SOS were all very favorable, as expected. RWS and SOS both resisted sintering at 40 MPa and 1300 °C, much higher pressures than expected in practice, while 1300 °C is somewhat higher than the maximum expected temperature, possibly 1000 °C to 1100 °C, in operational hot TES bins. All of these materials should be considered viable candidates for CSP and similar applications. SOS is particularly interesting.

The screening results for CB were less reassuring. CB passed at 20 MPa and 1000 °C, which is likely the minimal but acceptable performance for the highest-temperature CSP applications. Since CB has a desirable dark color, it is highly preferred for directly irradiated PHR-based CSP. Since its resistance to sintering is questionable, further investigation and longer-term testing are recommended before deployment in even pre-commercial applications.

The sintering results for USOS were very disappointing. As seen in Figure 4, it fused at rather a modest temperature and pressure, disqualifying it from any anticipated, even moderate-temperature, particle-based CSP applications. Figure 4a shows the sample after sintering at a temperature of 800–1000 °C. The pellet was formed, but it was broken due to the initiation of melting. Figure 4c shows the fused USOS lumps after sintering at 1100 °C. Figure 4b shows the leakage of molten USOS from the fine gap between the die and plunger. This result is especially disappointing since, as it is described as a magnesium-orthosilicate olivine material, it would be expected to have a very high melting point, maybe 1800 °C, and a correspondingly high sintering temperature. In fact, USOS is not a freshly mined material but a spent ore from a discontinued nickel production facility. It is, in fact, crushed from quenched molten ore. It is very likely that either the crystal structure was disrupted or impurities were introduced during processing. While this olivine-like material is not suitable for the intended application, many other olivine minerals exist, olivine being a major constituent of the Earth’s mantle and being rather common in the crust. The good performance of the recently mined SOS attests to the potential of other olivines generally. This USOS might be used in a fixed-bed or brickwork type of TES since it can be rather easily sintered into solid blocks for such applications. Overall, this USOS is not viable for applications with permanent stability as free-flowing particles at high temperatures, such as in particle-based CSP systems.

3.3. Thermal Conductivity

The results of thermal conductivity for all samples at ambient temperature are summarized in

Table 6. Measured bulk thermal conductivity of particulate materials at ambient temperature using KD2-Pro.

Sample	Thermal Conductivity [W/mK]
RWS	0.267 ± 0.02
RRS	0.260 ± 0.02
CB	0.202 ± 0.02
SOS	0.217 ± 0.02
USOS	0.155 ± 0.02

. It should be noted that the pellets were held at 500 °C for 5 h in an electrically heated clean-air muffle-type furnace [53] in order to remove any impurities left after processing.

The temperature-dependent thermal diffusivity (α) in [mm²/s] of pelletized samples of RWS, RRS, SOS, and CB is presented in Figure 5 over a temperature range of 25 °C to about 475 °C. The thermal conductivity (k) [W/mK] is also presented, which is calculated based on the thermal diffusivity, density, and specific heat, as mentioned in Equations (1)–(3). Note that while measurements of thermal diffusivity and data for the specific heats are available over the entire temperature range, only ambient-temperature density is available. No suitable apparatus for measuring the density change was available; however, the density change is expected to be very small and is estimated to cause less than 1.0% variation in the calculated thermal conductivity. This variation is entirely negligible compared with the approximately ±60% variation in conductivity seen in Figure 5.

Figure 5 illustrates the temperature dependence of the measured thermal diffusivity and the calculated thermal conductivities of the four pelletized samples. The thermal diffusivity values at low temperatures are all grouped from 1.2 to 1.6 mm²/s and are mostly grouped from 0.2 to 0.8 at the highest temperature. The diffusivity trends lower with temperature for all samples at similar gradients over most of the range. This trend is most likely caused by the expected decrease in thermal conductivity for solid particles in all these somewhat solidified pellets. The calculated thermal conductivities inferred from the measured diffusivities of all materials do show the expected decrease. Based on a comparison of CB with polycrystalline alumina, the conductivities appear to be about one order of magnitude lower than the conductivity of a fully dense crystalline solid, confirming that the pellets are not overly densified. The uncertainties associated with all experimentally measured quantities were carefully evaluated. For parameters with only two available measurements, such as thermal diffusivity, the uncertainty was estimated using the half-range method, calculated as half the difference between the maximum and minimum recorded values. This approach provides a practical estimate of uncertainty when limited data points are available and aligns with the Type B evaluation method described in the Guide to the Expression of Uncertainty in Measurement [54]. For thermal conductivity, the uncertainty was determined using uncertainty propagation techniques, taking into account the contributions from the uncertainties in thermal diffusivity, specific heat, and density.

The conductivity of the HTM will impact the size and cost of the particle-to-fluid heat exchanger (PFHX) in a CSP system. Since Figure 5 indicates the decrease in thermal conductivity at high temperatures for all the samples, that means a bigger PFHX will be needed to operate at higher temperatures. However, particles in the PFHX will be in a moving bed or other type of flow, and the thermal conductivity in such a bed or flow cannot yet be reliably predicted. In any case, the conductivities do not differ greatly. With respect to the design of TES bins, the conductivities of all these materials are similar and are considerably lower, by an order of magnitude, than the conductivity of typical insulating materials expected to be used in TES bins in CSP systems. Consequently, the heat loss from the TES bins will be nearly insensitive to the choice of HTM, and the HTM can be selected on the basis of economics and other performance criteria with relatively little effect from thermal conductivity.

4. Conclusions

There are many benefits to using solid particulates as an HTM in CSP systems, making them a better choice than traditional heat transfer media like steam, oil, air, and molten salt. These advantages encompass the capacity to operate at exceptionally high temperatures (>1000 °C), widespread availability, cost effectiveness, high thermal energy storage capabilities, absence of freezing constraints, and direct compatibility with solar irradiation. These attributes collectively contribute to significantly enhanced operational efficiencies and reduced levelized cost of energy expected from particle-based CSP plants. On the other hand, a crucial requirement is the thorough assessment of the characteristics of possible solid particulates meant for application in such severe-heat environments.

In this article, four naturally occurring particulate materials—Riyadh white sand, Riyadh red sand, Saudi olivine sand, and US olivine sand—and one synthetic particulate material called Carbobead were investigated. All natural particulate materials were sifted to the targeted particle size for the 1.3 MWe pre-commercial plant that will be built in Waad-Al-Shamal, Saudi Arabia [36]. In particular, the relevant ambient- and high-temperature thermophysical properties were thoroughly examined. The parameters that were examined include thermal conductivity, sintering temperature, actual density, tapped bulk density, and loose bulk density. This meticulous analysis aimed to ascertain the suitability of these materials for deployment in high-temperature CSP applications, thereby advancing the understanding and subsequent design and optimization of particle-based CSP and similar heat transfer systems. The results revealed or confirmed the following:

• Particulate size and size distribution have a significant effect on bulk density. For most of the particles, their porosity ranges from 35% to 45%, consistent with the commonly expected value of 40% for monodisperse particles. In contrast, the porosity of the USOS material is much lower, 24% loose and 18% tapped, which is indicative of the presence of fines in the mix filling the spaces that are voids between more uniform particles.
• The tapped bulk density for all particle materials is always greater than the loose density for all particles. In the case of 0.5 mm CB, the beads are so uniform and smooth that the effect of tapping is minimal.
• During the sintering test, when the temperature and pressure were raised to 1300 °C and 40 MPa, respectively, for RWS (0.5–1 mm), the specimen did not exhibit any signs of sintering. For RWS (0.21–0.425 mm), only weak binding formed at the high pressure of 50 MPa. Sintered RRS was produced at 40 MPa and 1300 °C, and CB at 20 MPa and 1100 °C. Similar behavior was displayed by SOS and RRS. At 800 °C, USOS solidified, while at higher temperatures, it melted. This demonstrates why USOS cannot be used in high-temperature particle-based CSP systems as TES or an HTM.
• Particulate materials’ thermal diffusivity and thermal conductivity decrease as the temperature rises, reaching their lowest values at the highest temperatures. This indicates that a bigger-size heat exchanger is needed to run at elevated temperatures. When comparing various particulates, RWS has the lowest thermal conductivity and minimum thermal diffusivity of 0.15 mm²/s and 0.46 W/mK, respectively, at the maximum temperature of 475.5 °C. Conversely, at the same temperature, the maximum values of the RRS are 1.25 W/mK and 0.432 mm²/s, respectively.

For future studies, it is planned to upgrade the transient laser flash apparatus at KSU for the measurement of the thermal diffusivity of loose particles. A furnace is also planned to be installed with the apparatus to extend the measurement temperatures up to 1200 °C. An on-sun testing of the solid particulates is also planned in the future at the KSU CSP demonstration facility to understand the thermal behaviors of the candidate materials with cyclic thermal loadings.