Direct gas heating in linear concentrating solar collectors for power and industrial process heat production: Applications and challenges

Parabolic trough collectors and linear Fresnel collectors are mature technologies for power production, and they are being recently applied to provide solar heat for industrial needs. Conventionally, they use a liquid as heat transfer fluid, either thermal oil or water, to carry heat from the receivers up to the point of conversion or delivery. Although liquids offer excellent thermal properties, they show technical limitations, besides environmental concerns, which have encouraged the research on alternative solutions. This work reviews the main research works on the use of gases as heat transfer fluid in linear concentrating collectors, including solar power and heat production, highlighting the potential applications and technical challenges. The review indicates: first, gases offer potential to replace liquids, and second, there is a need for more research and development to define the best technical compromises to reach practical application in every sector.


| INTRODUCTION
In the past decades, Linear concentrating technologies such as parabolic trough collectors (PTCs) and linear Fresnel collector (LFC) have been developed for large-scale solar thermal plants aiming at electricity generation (Fuqiang et al., 2017) known as solar thermal electricity (STE).
The PTC is the most mature thermal concentrating solar power (CSP) technology (Bilal Awan et al., 2020;Fuqiang et al., 2017;Gharat et al., 2021, among others) including STE. As a basic common feature, in a PTC a cylindrical parabolic-shaped reflector focuses the incoming direct component of the solar irradiance. The tubular receiver is located on the focal line as in Figure 1a. The receiver is an evacuated tube with high thermal performance, usually made of a stainless-steel tube coated with a spectrally selective coating, embedded into a coaxial high solar transparency glass According to International Energy Agency (2018), heat takes a large share of energy consumption in the industrial sector. Solar energy has a great potential for replacing fossil fuel consumption for heat production in industry, aiming at greenhouse gases and pollutant emission reduction, nonforgetting environmental sustainability (Farjana et al., 2018; F I G U R E 2 Linear concentrating technologies. Example of thermal concentrating technologies at Plataforma Solar de Almeria (PSA, n.d.): (a) PTC (Eck et al., 2003); (b) LFC (Bernhard et al., 2008). Example for SHIP: PTC (c) by Inventive Power (n.d.) and LFC (d) by Solatom (n.d.). LFC, linear Fresnel collector; PTC, parabolic trough collector; SHIP, solar heat for industrial processes.
F I G U R E 3 Simplified basic diagram of conventional STE plant using thermal oil as HTF. The solar field is simplified to a single loop of PTCs or LFCs. Molten salt-based thermal energy storage (TES) can be present (dashed lines). HTF, heat transfer fluids; LFCs, linear Fresnel collectors; PTCs, parabolic trough collectors; STE, solar thermal electricity. Mekhilef et al., 2011;A. K. Sharma et al., 2017). Among solar thermal technologies, linear concentrating collectors, either PTC or LFC, are the suitable solution in the low-to-medium temperature range, below 400 C, corresponding to around half of the total heat consumption in the industry (International Renewable Energy Agency, n.d.). Due to their modularity, they can be adapted to small-and medium-scale facilities suitable for installation in industries. Both LFCs and PTCs optimized for SHIP application are commercially available (Figure 2c,d). The SHIP database (SHIP Database) 2022 reports 60 projects using PTCs with solar field areas ranging from 34 to 5060 m 2 . From the same source, 21 existing plants use LFCs for SHIP from 123 m 2 of aperture surface up to 2000 m 2 .
Regarding the mismatch of the heat demand in front of the solar production time profile, there are three possibilities: (1) implementing heat storage (TES), (2) using a backup source while solar production cover only a portion (solar fraction) of the overall consumption, and (3) a combination of (1) and (2) (Figure 4). This figure does not show any backup source as they are diverse and can be implemented in series or in parallel either in the supply circuit or in the consumption apparatus of the thermal process.
Current practice for HTF is oriented to liquid substances with suitable thermal properties, mainly thermal oil for STE applications and pressurized water predominating in SHIP projects, where internal evaporation is avoided. The use of gases as an HTF seems to offer some potential for application both in STE and SHIP due to their lower environmental impact, cost, and leakage risks, among other technical advantages, but some technical challenges are still unsolved, and experience is short.
Throughout the following sections, this work reviews the main research and development carried out and highlights the main advantages and limitations of several gaseous HTFs in comparison with the widespread liquid alternatives. It focuses on emerging SHIP applications and challenges.

| APPLICATIONS
Although sharing the same principles, both PTC and LFC technologies differ when applied to STE or SHIP, in several terms. Such are a smaller scale of the solar field, collector sizes and concentration ratio, differences in components, and operating conditions (temperature, flow rate). The HTF requirements also change according to the specific applications. The present section reviews the main trends and research work on alternative gaseous HTFs either for STE or SHIP.

| Solar thermal electricity
With STE, the large-scale applications dominate, which goes together with large apertures and concentration ratios. Thermal oils, such as Dowtherm™ A and Therminol™ VP-1, are the most common choices as HTF due to their good F I G U R E 4 Linear concentrating collector layout for industrial process heat (SHIP). The solar field is simplified as a single loop with PTCs or LFCs. Thermal Energy Storage TES can be present (dashed lines). LFCs, linear Fresnel collectors; PTCs, parabolic trough collectors; SHIP, solar heat for industrial processes. thermal properties (Zarza, 2016). Their main drawbacks are related to labor safety and environmental concerns in case of leaks from the solar plant circuit, besides the associated fire hazard. From the thermodynamics point of view, the HTF's limited working temperature (≲400 C) curtails the maximum power Rankine cycle temperature and consequently the conversion efficiency of the plant (Figure 3). Oil also degrades during the operation. For around 4% of the total oil must be replaced and disposed of every year (Giglio et al., 2017). Crystallization at low ambient temperatures represents an issue affecting operating strategies and increasing O&M (OPEX) concerns and costs.
Among the alternative HTFs, nitrate molten salts have received considerable attention in replacing thermal oil in CSP applications especially with PTC, although commercial applications are still a few. Their main advantage is the higher operating temperature, above 500 C, enabling a higher conversion efficiency, but the high solidification temperature, $220 C, requires a dedicated heating system to avoid plant damage, increasing the complexity, costs, and maintenance operations (Zarza, 2016).
In general terms, gases can operate at much higher temperatures without any degradation, and most of them do not present risks because of leakages. Under this framework, research on gaseous HTFs has been encouraged (Benoit et al., 2016). According to Zarza (2016), the larger temperature difference across the solar field reduces the volume of the heat storage needed (Figure 3). Air, nitrogen (N 2 ), and carbon dioxide (CO 2 ) have received most of the attention. Steam is considered an HTF in the framework of direct steam generation (DSG), where the water is preheated and evaporated in a once-through process taking place inside the receiver tubes, to obtain either saturated or even superheated steam. The direct use of single-phase steam as a gaseous HTF has not received attention.
One of the most relevant research projects on gaseous HTF with PTCs has been carried out at Plataforma Solar de Almeria (PSA; Muñoz-Anton et al., 2014;PSA, n.d.). CO 2 was preferred among other gases as HTF in an experimental test facility made of two 50-m length ET-II PTC collectors with an aperture of 5.76 m, operating in a closed circuit, pressurized up to 100 bar and at 525 C of maximum temperature. Biencinto et al. (2014) presented a new concept for solar thermal power plants through a theoretical investigation on a PTC-based STE using pressurized N 2 as HTF. Therefore, the solar field is configured by parallel subfields using ET-II collectors. Molten salt is used as secondary HTF carrying heat from the subfields to the power block, as represented in Figure 5. Biencinto et al. (2019) presents a similar concept applied to supercritical CO 2 as HTF inside large aperture PTCs, equipped with an evacuated receiver with a larger inner diameter than the commercial standard of 70 mm (Rioglass, n.d.) so that pressure loss is diminished and capturing concentrated rays is easier. Bellos et al. (2016) compared various liquid and gaseous HTFs for their usage inside a PTC. Air, N 2 , CO 2 , He, and argon (Ar) were analyzed to determine the best working conditions for each gas. In a further study, the PTC technology operating with gas using longitudinal internal fins was studied by (Bellos, Tzivanidis, Daniil, et al., 2017). Bellos, Tzivanidis, and Antonopoulos (2017a) evaluated the energy and exergy performances of pressurized air, CO 2 , and He, compared with liquid HTFs in PTCs. Supercritical CO 2 was theoretically explored by .
Air in large PTCs was investigated using nonconventional concentrating technology, based on very large PTCs with original cavity receivers. Good et al. (2015) proposed an original cavity receiver for operating with air temperatures F I G U R E 5 Simplified diagram of a STE plant using pressurized gases as HTF, following Biencinto et al. (2014) and Biencinto et al. (2019). The solar field is simplified as a single loop of PTCs. HTF, HTF, heat transfer fluid; PTCs, parabolic trough collectors; STE, solar thermal electricity. above 600 C. Bader et al. (2015) also developed an innovative cavity receiver for usage with a large-span PTC using a 9 m of aperture heating air up to 500 C.
Pressurized air heated inside PTCs has been considered for integration with Brayton cycles for power production. To reach competitive efficiency, this cycle needs high air temperature, in the range of 600-1000 C and up. Linear concentrating collectors, with the current technology, do not allow these temperatures, but they can be still used as air preheaters for a combustion chamber (cc) to power a Brayton cycle as a solar-assisted gas turbine, either as a simple cycle or as a Rankine combined cycle (Ahmadi et al., 2018; Figure 6). Air can either be directly preheated by flowing inside the solar collectors or indirectly through a heat exchanger (HX; Behar, 2018). Amelio et al. (2014) studied an open solar-assisted combined Brayton cycle using PTCs, heating compressed air into a PTC solar field up to 580 C before entering the combustion chamber. One similar example of a solar-assisted gas turbine is studied by Bellos, Tzivanidis, and Antonopoulos (2017b). The pressure ratio is limited by inlet air temperature at the solar field (outlet compressor air temperature) unless intercooled compression is implemented (Dabwan et al., 2022). The use of a linear concentrating collector as the unique heat source in the power cycle has been studied to less extent, aiming at lower maximum temperature and modest efficiency, considering the Brayton cycle (Ferraro & Marinelli, 2012), or a discrete steps Ericsson cycle (Cinocca et al., 2018).
DSG represents an important field of research that up to now has involved both the PTC (Eck et al., 2003;Zarza et al., 2006) and LFC (Bernhard et al., 2008;Montes et al., 2014, among others). The HTF in the solar field is replaced F I G U R E 6 Simplified diagram of direct air heating in linear concentrating collectors with a Brayton cycle configuration. The solar field is simplified to a single loop of PTCs or LFCs. A combustion chamber (cc) can be present (dashed lines). LFCs, linear Fresnel collectors; PTCs, parabolic trough collectors.
F I G U R E 7 Simplified diagram of direct steam generation DSG for STE. The solar field is simplified to a single loop of PTCs or LFCs. A recirculation circuit and separator (dashed line) can be used. DSG, direct steam generation; LFCs, linear Fresnel collectors; PTCs, parabolic trough collectors; STE, solar thermal electricity. by water which is preheated, evaporated, and converted into saturated or superheated steam inside the receivers (Figure 7). A power block steam generator and an HTF/steam HX are not required anymore, achieving layout simplification and potentially lower installation costs. The solar field is directly coupled with the power block in an STE layout. Zarza (2016) summarizes the research effort carried out on DSG development. DSG has reached commercial application for both technologies, although LFC counts with a greater number of plants than PTC. Minor attention has been given to direct vapor generation in organic Rankine cycles, systems, worth mentioning (Loni et al., 2021;Xu et al., 2015).
Despite the published relevant numerical works as well as important experimental research projects on gases heated inside PTC, there is a lack of commercial examples of STE. Concerning LFC, it has been considered a technology for DSG (Bellos, 2019). Their use with pressurized gases as HTF has been rarely studied, as in one numerical study by Montes et al. (2016), where air at 90 bar is considered as HTF, besides water/steam, for comparison with PTC under similar conditions.

| Solar heat for industrial processes
Regarding SHIP, the trends differ from STE. The typical scale is much lower than in conventional STE, so shorter rows and loops are commonly used for smaller fields. Concentrating technologies go toward lower concentration ratios, aiming at lower procurement and maintenance costs, and less skyline invasion because of a smaller aperture. This is especially noticeable for PTCs, which even exhibit so small apertures as 1 m and nonevacuated small diameter receivers in most of the existing applications. LFCs, thanks to their lower cost, low suspended weight, and small frontal area, are viable with larger apertures. The higher concentration ratios overcome their lower optical efficiency. Lower temperatures than STE (<250 C) are common practice in SHIP applications so increasing the operating temperature has limited relevance.
Water is used as HTF in most of the SHIP installations using PTCs. It offers excellent thermophysical properties but presents the main issue of low boiling point, not forgetting the need for pure water. For applications above 100 C, pressurization is needed to maintain the liquid state. The higher the operating temperature, the higher the pressure required and hence the technical requirements and leakage risk, with associated costs and safety issues. Easily affordable 15-20 bar is enough for an operating temperature of 200 C, but very high pressure is required for increased operating temperatures (85 bar @300 C, 165 bar @350 C). Indeed, SHIP applications using water do not exceed 200-250 C. Technical issues concerning water usage are the corrosion problem affecting the carbon steel components, in addition to the freezing problem, arising during nonoperating cold hours with ambient temperature below 0 C. Water/glycol mixtures have been used in a few SHIP plants to lower the freezing point, incorporating the need of replacing the anti-freeze agent for its potential corrosion action. Alternatively, a drain-back operation is required to empty the outdoor water circuit during cold hours. This is complex for the largest SHIP plants. In addition to that, pretreating equipment is required to demineralize fresh water before injecting it into the circuit, for start-up and makeup water.
The large use of steam in industry and the associated technological maturity motivated the efforts toward DSG. In existing applications, saturated steam seems a more attractive option than superheated steam due to its easy integration into the user steam grid. A review of existing installations (SHIP Database, n.d.) showed that the use of PTC as direct solar thermal steam boiler systems is less common than LFC. A few additional works cover the topic. Holler et al.
(2021) exposed a feasibility study of superheated steam production in the beverage industry. Superheated steam was provided at 160-180 C and a pressure of 8 bar in the solar field, considering two aperture areas of 3360 and 4480 m 2 Cundapí et al. (2017) performed a study on the absorber tube for the direct steam generation with small aperture PTCs for industrial applications. They considered a module aperture of 1 m, length of 2 m, an absorber diameter external/ internal 18/15 mm, with 76 m of loop length. LFCs seem more suitable for steam generation than PTCs, according to Sepúlveda et al. (2019). Haagen et al. (2015) reported an applied example where a saturated steam generation plant for industry using LFC is described. Eighteen LFC modules are installed on the rooftop of the factory in two rows with a total aperture area of 396 m 2 . Water is partially evaporated in the receivers. Steam at 6 bar, 166 C is separated from water in the external steam drum, while water is recirculated back to the solar field. To maintain a constant mass inside the system, softened water is pumped into the bottom of the steam drum while steam is withdrawn. Hongn and Flores Larsen (2018) offered a study of the hydrothermal behavior of DSG in a small LFC facility for industry, using saturated and superheated steam for a multitube trapezoidal receiver. The model was validated against experimental data of a prototype installed in San Carlos, Argentina.
The use of pressurized gases such as CO 2 or N 2 for HTF in SHIP applications, although they received some interest for STE, has not been studied, according to the open literature. On the other hand, air heating in PTCs and LFCs has recently received moderate interest. Together with steam, air is one of the most used fluids in industrial processes (Farjana et al., 2018;A. K. Sharma et al., 2017), for thermal processing of materials, manufactured products, biomass, and water treatments, drying, curing, and cooking. Depending on the application, the air temperature requirements vary from the low-temperature range (<150 C) up to medium temperatures and high temperatures (>400 C). Solar air heating up to medium temperature can be achieved by using linear concentrating collectors either indirectly, using a primary liquid HTF in the solar field, or even directly, using air as HTF inside the receivers. A few examples of indirect heating up to 160-180 C are reported in the technical literature using LFCs (Pietruschka et al., 2016;Rehman et al., 2019;SHIP Database, n.d.) or using PTCs (SHIP Database, n.d.). In these works, the layout is using water or thermal oil as HTF and an HTF/air HX upstream of the industrial thermal processes.
Nowadays, direct solar air heating is widely experimented with by using flat plate air collectors below 150 C (Saxena et al., 2015). Recent works focus on the use of ambient air as HTF in concentrating collectors for SHIP applications near ambient pressure in an open to atmosphere circuit (OAC) layout up to medium temperatures (300-400 C) primarily aiming at layout simplification and lower costs. As for a nonconcentrating solar air heater, the OAC could be implemented for industrial solar hot air generation, avoiding the need for a primary HTF as well as an HTF/air HX, with the advantage of a higher delivery temperature, achievable thanks to the corresponding higher thermal efficiency of linear concentrating collectors. Xuening et al. (2015) offer an original numerical study using PTCs delivering air at 280 C at 2 bar. Hot pressurized air is used to evaporate brine from a desalination process in a spray evaporation tower. Kasperski and Nem s (2013) proposed a concentrating solar air heater, which uses PTCs for direct air heating. They analyzed different types of internal longitudinal fins for enhancing the receiver-to-flow heat transfer and evaluated the thermohydraulic collector efficiency, which includes pumping power consumption. Famiglietti et al. (2020) offered a comprehensive theoretical analysis of a concentrating solar air heater (SAH). The high pumping power arises as a practical constraint to the design of the solar field, limiting to short rows and moderate apertures. In the main conclusion, they stated that direct atmospheric air heating with PTC/LFC is energetically feasible within a restricted range of design parameters. In the same study, the authors introduced the innovative concept of Turbo-assisted concentrating solar air heater (T-SAH; Lecuona-Neumann, 2016; Figure 8). The original layout uses an automotive turbocharger coupled with the solar field operating with air. This minimizes the pressure drops and avoids the external auxiliary pumping power consumption, while an axillary electrical compressor only is needed for starting.
A further study (Famiglietti et al., 2021) presented a feasibility study of T-SAH on an industrial scale. Commercial LFCs and large-size automotive turbochargers were considered for the scope. A first small-scale experimental prototype of the T-SAH was designed and installed in Madrid, Spain (Famiglietti & Lecuona, 2021a). The results corroborate the practical viability of the concept and indicated relevant features and critical aspects for scaling up to industrial size. Further work (Famiglietti & Lecuona, 2021b) details the experimental investigation performed on the T-SAH prototypes F I G U R E 8 Open to atmosphere circuit concentrating solar air heater for industrial processes according to Famiglietti et al. (2020). The solar field is simplified as a single loop of PTCs or LFCs. Turbine e, compressor c, auxiliary electrical compressor ac. A turbocharger can be used for pumping (turbo-assisted SAH), otherwise only the electrical compressor ac is active.
focusing on the air heating and receiver behavior. High temperatures are reached under continuous operation, up to 500 C for the air at the receiver outlet, and close to 600 C for the tube wall without damage. One exceptional quality of this innovation is the self-adjusting outlet temperature when solar input varies.

| CHALLENGES
Gases of interest as HTFs such as CO 2 , air, N 2 , steam, and He, offer thermophysical properties which differ from the common liquid HTFs (Benoit et al., 2016). Compared to a liquid substance, they offer lower specific heat capacity c p , lower density ρ and thermal conductivity k, and higher viscosity μ. Besides, they vary among them Table 1. At the same pressure and temperature, CO 2 has the highest density, which is favorable. He has higher thermal conductivity and specific heat, favorable too. Air, very similar to nitrogen, shares the highest viscosity with He, and it is the only one that can be used in an OAC, carrying a substantial layout simplicity.
Among the gases considered for HTFs, CO 2 represents a superior behavior due to its higher density, as indicated by Muñoz-Anton et al. (2014). As the main drawback, it is incompatible with graphite sealing above 400 C (Zarza, 2016). In case of water presence, CO 2 could react to form carbonic acid (H 2 CO 3 ), which is corrosive to carbon steel. CO 2 shows even better properties when working under supercritical conditions, above its critical point (30.98 C, 73.77 bar). For its good thermal properties, it has been considered the coolant in nuclear power plants, and also proposed in Rankine and Brayton cycles as a working fluid (W. L. Cheng et al., 2017;Crespi et al., 2020;Ma & Turchi, 2011;Sarkar, 2015, among others). N 2 poses a less corrosion issue than air or CO 2 , and it can be obtained on-site from the air. He presents the risk of leakage due to its small atoms.
Thermophysical properties concur to determine the thermo-hydraulic behavior of the gases, which is the most challenging aspect of their design and operation. This issue is considered by most of the studies reviewed and mainly supported by theoretical and numerical analysis, but just in a few cases supported by experiments. The inclusion of O 2 , H 2 , CO, CH 4 , and other gases for the innovative techniques, thermal solar fuel synthesis at high temperatures (Z. D. Cheng et al., 2019;Warren & Weimer, 2022) or photosolar water splitting for green H 2 (Salcedo-Abraira et al., 2021), seems of high interest. Whether linear concentrating solar collectors would play a role in those technologies has still to be determined.

| Basic equations
To clarify the relationship between the different design and operating parameters on the performance of the technology and their dependence on HTF properties, simple 1D steady-state models, considering constant fluid-thermal properties, are much highlighting and accurate, owing to the slenderness of the receiving tubes. Linear concentrating collectors, either of the PTC or LFC types, reflect the direct component of solar irradiance G bn impacting on the primary reflector T A B L E 1 Main thermal properties of gaseous and liquid heat transfer fluids (HTFs). area A a ¼ W a L toward the receiver tube with a certain optical efficiency η op . The heat flux _ q s on a tubular receiver external surface A r ¼ LP ex can be expressed as in Equation (1).

Temp. ( C)
P ex ¼ D ex π is the external perimeter of the receiver tube with an external diameter D ex . The concentration ratio C ≝ W a =P ex is a relevant parameter for the collector characterization. The collectors' row length L can be adjusted to any desired value, arranging in series several collector modules in a row. In STE installations, very long rows L $ 100 to 1000 m are the standard practice, owing to the high thermal capacity of liquids. Smaller L $ 10 to 100 m are typically considered for SHIP applications. The primary reflector aperture W a varies according to the collector model, from $ 1 m for small size PTC (SHIP), up to $ 8 m for the largest examples in CSP. For Linear Fresnel collectors, W a is the sum of the aperture w m of n m mirrors constituting the discretized primary reflector, typically W a ¼ w m n m $ 5 m in SHIP, up to W a $ 20 m for CSP. Concentration ratios are above C $ 25 for PTC in CSP, while limited to C ≲ 20 for SHIP. LFCs for CSP exhibit up to C $ 50 or C $ 20 for SHIP. The maximum optical efficiency range between η op, max $ 0:5 to 0:8 from low-end LFCs to high-end PTCs. The off-design figures are lower due to the cosine effect, and optical end losses, among other effects, according to the solar incidence angle, summing up both lateral and axial components through an incidence angle modifier IAM.
HTF temperature T and wall temperature T w increase downstream of the row, making the last receiver tubes critical. Wall temperature is higher in the direction of the impacting rays, but turbulent inner convection helps in homogenization besides wall heat conduction. Assuming a radially averaged tube wall temperature and _ q s , the thermal balance equation on the receiver cross-section, Equation (2) shows that part of the incoming heat flux _ q s is lost to the ambient as losses _ q L . The resulting heat flux to the HTF _ q u is controlled by the heat loss transfer coefficient h across the internal circular perimeter P ¼ Dπφ. This is indicated in Equation (2) as the metallic wall of the tube is a much lower thermal resistance from the surface absorption to the HTF. The factor φ ≥ 1 accounts for extended surfaces (internally finned tube). It is φ ¼ 1 for smooth tubes, which is the common case.
The thermal losses coefficient U L is a wall temperature-dependent coefficient resulting from the heat transfer mechanisms involved toward ambient, according to the receiver configuration and material properties. Under nominal operation, an indicative value can be 3 -5 W m À2 K À1 (Duffie et al., 1985) for evacuated tubes. They offer relatively low U L since the vacuum in the annulus minimizes the convection losses, the selective coating reduces the reradiation losses and self-radiation (Burkholder & Kutscher, 2009;Forristall, 2003) while the external glass cover tube can be IR opaque and include antireflective coatings. Other configurations, such as nonevacuated tubes, noncovered tubes, and glazed cavity receivers, offer higher U L .
The Dittus-Boelter correlation (Incropera & DeWitt, 1996) gives a first approximation of the turbulent wall-to-flow internal heat transfer h (Equation 3), which depends on the flow Reynolds and Prandtl numbers (Equation 4), being D h ¼ Dφ the circular hydraulic diameter, and _ m is the mass flow rate.
Equation (3) shows that low thermal conductivity, typical of gaseous HTF excepting He, Table 1, limits h, although thermal oil shows a value in between. As a consequence, the wall over-temperature ΔT w normally almost negligible with liquid HTFs becomes relevant and can turn into a design or operating constraint for the material of the most downstream tubes. As shown in Equation (5), ΔT w is higher as higher the heat flux _ q s is; hence its effect grows with collector aperture, ceteris paribus. This equation also indicates the effect of density and velocity.
On the other hand, the Reynolds number of the flow plays an important role in determining h and subsequently ΔT w . For a given fluid, Re D h is related to the mass flow rate flowing into the receiver _ m in Equation (4), which is imposed by the thermal balance of the collector row as in Equation (6). Once the extreme temperatures are defined by the application, _ m is determined by the collector aperture W a according to the row length L, determining the total solar heat power delivered to the flow _ Q u .
For a given fluid, a low _ m translates into a low Re D h , thus a low h (Equations 3 and 4). The thermal balance equation on a generic cross-section of the receiver shows as the lower h the higher the wall overtemperature ΔT w is (Equation 5), which leads to higher T w for a given T. Since the thermal losses increase with T w , a low h leads to higher thermal losses, reducing the thermal efficiency of the collector η t (Equation 7).
Besides, when ΔT w is not negligible, the maximum thermal limit of the receiver tube can be reached even if the fluid temperature is not critical by itself, limiting the maximum fluid temperature achievable to T max ¼ T w,max À ΔT w : The maximum allowable temperature T w, max of the receiver tube is indicated by the manufacturer, with the main goal of preventing selective coating degradation. T w,max varies from 450 to 600 C with current technology. When T max is defined by the application and T w,max by the receiver technology, the allowable ΔT w turns into a constraint on other design parameters such as W a or L. (Famiglietti et al., 2020) found a minimum L allowable by this thermal limit for direct air heating with PTC and LFC, being short rows associated with both low _ m and h. (Famiglietti & Lecuona, 2021b) experimentally observed a decrease in thermal efficiency due to high ΔT w (up to $ 200 C) in a small scale LFC with air as HTF.

| Hydraulic issues
As indicated above, increasing _ m is beneficial for the internal heat rate removal from the wall but affects the stagnation t pressure drop along the receiver tube and external tubing Δp t ¼ p in,t À p ou,t . Equation (8) shows both its kinetic and frictional components, respectively Δp k,t and Δp f,t while K n considers localized pressure losses (bends, connections, and cross-section variations). The average m Darcy friction factor f and ρ characterize the distributed losses on the internal tube, considered as the smallest D h of the loop. They can be estimated using the Blasius correlation for smooth tubes (Equation 9) for a low Mach number but dilatating flow.
It is worth mentioning that the overall circuit length can be larger than just the collector row length L, as connection tubes are needed, L c in Equation (8), connecting more rows at their ends to configure the loop. Besides when scaling up to a large solar field made of several subfields, a nonnegligible pipe length is needed to connect the loops with the fluid delivery point.
The pumping power required to move the HTF across the collector row can be evaluated by considering a pumping compressor isentropic efficiency, total to total, η ptt (Equation 10) with average properties p. For the compressible flow, the pumping power is expressed as in Equation (11), which considers a perfect gas.
For a given receiver internal diameter D, the higher _ m, the higher the flow mean velocity v is, which mainly increases Δp t . The lower ρ, the higher v, as well as Δp t (Equation 8). It follows that Δp t is always a constraint when dealing with gaseous HTF, having much lower densities than liquids. Moreover, there is a relation between _ m and L, which is controlled by imposing the desired temperature increment across the row and C (Equation 1). A long row with a large collector aperture requires a high _ m (as higher as lower the c p is) to absorb the incoming thermal power and limit the row wall outlet temperature (Equation 6). The pumping power (Equations 10 and 11) can be the limiting constraint on the collector row length and aperture, much more limiting using gases than liquid HTFs.
In CSP with a gaseous HTF, as large-scale plants are the standard, the pumping power will be a limiting constraint due to the long rows as well as the large aperture of the typical collectors (PTC and LFC).
High pressure is unavoidable for reducing the pressure drops down to reasonable values. The consulted literature is also clear about gas pressurization as an efficient way to reduce the pressure drops and the pumping power in a collector row. The pressure drop reduction is related to pressure by a quadratic relation (Equation 8; 10 times pressure increase means 100 times Δp t ). As a drawback, high pressurization poses problems in the joints and flanges, especially for moving elements such as ball joints in PTC plants so that robust designs must be used, and safety issues considered.
As for comparison, a typical STE plant loop of PTCs (W a ¼ 5:7 m) is of L $ 570 m, operating with _ m $ 5 kg s À1 of thermal oil (Figure 3). An optimized equivalent DSG LFC loop is typically longer, L $ 860 to 1,140 m, operating with _ m $ 2 kg s À1 of water (Figure 7). Due to the high-pressure drop of the steam section, working pressures below 50 bar are not recommended. Pressure of 75 bar seems a good compromise considering the leakage issue, which intensifies with pressurization (Zarza, 2016). DSG for SHIP at lower than 50 bar pressure is the current practice, due to the smaller size of the facilities and lower temperature required. An additional problem arises in DSG due to the coexistence of liquid and vapor phases in the same horizontal row, with spontaneous large variations of pressure drop owing to the twophase flow. They can cause not only flow instabilities by themselves but also when interacting with neighbor ones. These issues can result in complicated control and/or the need for pressure reduction valves between rows. Numerical and experimental studies proved the benefit of pressurization (Muñoz-Anton et al., 2014). They estimated, for a Euro-Trough type 100 m long PTC with CO 2 , a pumping power going from 5.24 kW at 25 bar down to 0.53 kW at 75 bar, processing _ m ¼ 1:1 kg s À1 . Similar values are reported for N 2 under similar conditions. In spite of the pressurization effect, the same ET 100 m PTC would need a lower pumping power of 0.24 kW working with thermal oil (Zarza, 2016). Although pressurization is beneficial, the pressure drops are still a problem when scaling up due to the long piping connecting the loops between them and with the power block. One of the studied solutions is the usage of a molten salt as a secondary fluid carrying heat from the subfield up to the power block, this way keeping the loops at a moderated length of 100 m , or up to 200 m using a larger diameter receiver and larger aperture PTCs, for example, the Iber-Trough type (Biencinto et al., 2019; Figure 5). Bader et al. (2015) estimated the pumping power for an original cavity receiver with a diameter of 450 mm using atmospheric air as HTF. They found a large variation with _ m, going from 2.5 kW with _ m ffi 2 kg s À1 up to 47 kW with _ m ffi 5 kg s À1 on a PTC with a 9 m aperture and 200 m of length.
In SHIP studies, the hydraulic problem is rarely considered explicitly. The original work on an open circuit concentrating solar air heater by Famiglietti et al. (2020) shows that working with air at atmospheric pressure, $10% of the net solar power gain is needed for pumping using a collector LFC with a 5 m aperture and length of 30 m. It grows to $20% for a length of 40 m, more than $50% for a length of 50 m. To overcome the problem, they propose implementing an OAC solar Brayton cycle with null mechanical (power) efficiency aiming to provide heat instead of mechanical power (and/or electricity; Figure 8). It is based on compressing ambient air before solar heating, this way increasing its density for reducing the mean velocity and the stagnation pressure drop. A turbine placed at the receiver outlet expands the airflow down to approximately ambient pressure, releasing a hot air stream at its outlet and providing both the compressing and pumping power. Since no power is extracted at the turbine and compressor common shaft, a standard turbocharger can be used for the duty. As shown in Famiglietti et al. (2021), in a well-designed system, using a compressing ratio from 1.5 to 3.5, typical of turbochargers, and considering the commercial evacuated tube thermal limit, hot airflow at 300-400 C can be provided at the turbine outlet, available for the industrial usage, with no external power consumption for pumping.

| Thermomechanical issues
The other issue consequent of the low h imposed by a gaseous HTFs is both the increased mechanical stress of the receiver tube and bending, as a consequence of the circumferential temperature gradient.
The concentrated irradiance distribution on the receiver perimeter can be non-homogeneous, resulting from the collector geometry and its optical properties. This is a well-known issue for both PTCs and LFCs , even using liquid HTFs. In the case of PTCs, most of the concentrated irradiance heats the receiver tube on its parabola-facing peripheral arc, while the opposite portion of the tube perimeter receives a negligible heat flux. In a typical large PTC for a STE plant, the impacting side of the receiver pipe receives more than 40 kW m À2 while the opposite side only receives around 1 kW m À2 of nonconcentrated radiation (Figure 9a). The peripheral distribution of heat flux varies with the rim angle, focal length, and collector aperture. Most of the studies use raytracing techniques to simulate the heat flux distribution , but also measurements have been also provided (Riffelmann et al., 2006;Schiricke et al., 2009, among others).
LFCs implement more complex optics than PTCs. In principle, the heat flux distribution can be even peakier on the bottom half of the receiver, but the second reflection taking place on the secondary optics increases the sun rays' impact on the upper part of the tube (Figure 9b).
It is worth mentioning that the heat flux is not homogenous along the row length due to non-irradiated segments, which appear due to spacing between consecutive modules or at the extremities of the row due to "optical end losses" caused by the longitudinal incidence angle. These increase with de focal length, and its importance is higher in SHIP applications as a result of the smaller L.
The non-uniform heat flux distribution induces a circumferential thermal gradient within the tube wall. The temperature distribution is affected by the internal heat transfer, hence with the flow conditions and the wall heat conduction. Since circumferential temperature gradient drives an axial bending moment originated by differential thermal dilatation of the material, low h for a constant _ q s leads to increased deformation and mechanical stress, up to plastic deformation risk and loss of focus. Moreover, it can lead to the critical lateral displacement of contact between inner and outer tubes, with the corresponding risk of glass tube rupture and/or coating degradation. This phenomenon is recognized as one of the main causes of failure in PTC-based STEs when a critical flow condition is reached such as low _ m, transient operation, laminar flow, maximum _ q s (Fuqiang et al., 2017). Operating with gases, that problem is always an issue, even under normal operation, unless adequate secondary optics jointly with a focusing strategy are devised for homogenizing the heat flux.
Several studies deal with the mentioned challenge, although mainly referring to liquid HTFs, using combined numerical simulation techniques (Khanna et al., 2013(Khanna et al., , 2014(Khanna et al., , 2015Marug an-Cruz et al., 2016;Valdés et al., 2014). Still, research is needed to address and quantify the impact of the phenomenon on the large variety of possible applications as well as on the mitigation techniques. Dealing with pressurized gases (Muñoz-Anton et al., 2014) recommended h $ 1000 W m À2 K À1 to keep maximum circumferential temperature difference in the order of $50 C. Under similar conditions bending was not excessive and cover glass rupture was avoided. The problem is considered in detail for a DSG plant by Rold an et al. (2013), supported by experimental measurements of the thermal gradient on a PTC wall obtained at the DISS facility at PSA (n.d.). Temperature differences up to 45 C were reported in the superheated steam section. A typical value of wall-to-gas h in DSG plants for CSP is 800 W m À2 K À1 (Zarza, 2016). Famiglietti and Lecuona (2021b) observed tube bending in a small-scale LFC with air as HTF up to 500 C, with lower h $ 100 W m À2 K À1 . No damage to the evacuated receiver was reported, underlining the beneficial optical diffusion effect of the secondary optics. Besides, a lack of more experimental studies is noticed.
One method of thermal deformation reduction of the tubes through reduction of ΔT w includes internal heat transfer enhancement. Some studies have been carried out aiming at increasing wall-to-HTF h by installing longitudinal inserts inside the tubular receiver to augment flow turbulence and/or heat transfer surface (M. Shahzad et al., 2021;Sharma & Jilte, 2021). The investigations focus on optimizing the insert features to keep a reasonable excess of friction introduced, either considering wire coils (Yılmaz et al., 2020), a helicoidal screw (Sami, 2021;Song et al., 2014), or a twisted tape (Jaramillo et al., 2016;Mwesigye et al., 2016, among others).
Optical design optimization for more uniform heat flux peripheral distribution is a way for mitigating the problem. A detailed study performed by (Chaitanya Prasad et al., 2017) achieved the homogenization on an LFC for DSG, modifying the shape of the secondary reflector, the location of the focus line, or the curvature of the primary mirrors ( Figure 9b). The possibility of adding a secondary reflector in PTC, analogous to the case of LFC, has been investigated to achieve uniform heat flux using different shapes (Figure 9a; Canavarro et al., 2013;He et al., 2019;Rodriguez-Sanchez & Rosengarten, 2015;Wang et al., 2014).
Structural improvement of the receiver tube has been also considered, with the adoption of advanced materials, besides increasing the wall thickness. A compound copper-steel tube has been studied for improving the temperature distribution within the framework of a DSG plant (Almanza et al., 1998;Flores & Almanza, 2004;Khanna et al., 2018). Alternative materials with superior mechanical properties (nickel superalloy or copper) have also been considered (Fuqiang et al., 2017).

| CONCLUSIONS
The review of the technical literature, summarized in the previous sections, indicates that relevant research efforts have been devoted to direct gas heating in linear concentrating collectors. The motivations have been explained showing a high potential for improvements. The main issues are presented and analyzed, and the basic steady-state equations are presented so that the interleaving requirements can be fulfilled with a low computing effort. Transient operation is an additional issue that deserves additional research, including the study of failures on some components. At the same time, a critical analysis suggests that still more work is needed to deeply explore the large variety of technical aspects involved, considering several points of view: theoretical, experimental, economic, environmental, and industrial, for the different applications considered as potential ones.
It is noticed that efforts been made to analyze deeply certain issues, but a systematic and comprehensive assessment has been found still missing. This work is an effort to fill this gap. To complicate the assessment of the solar direct gas heating technology, the applications, as well as the technologies involved, are diverse and some of them still emerging.
The analysis of the literature indicates that the thermal constraints imposed by heat transfer, and the related mechanical aspects such as bending deflection and stresses must be considered together with the hydraulic issues of pressure drop and pumping power. This is of paramount importance for evaluating the viability of a certain fluid as HTF in the herewith case considered of linear concentrating technologies and for determining the optimal design and operating conditions. Although quite a variety of theoretical investigations are available, experimental studies are limited to a few examples.
It is surprising that beyond the authors' publications, there have not been found complementary studies on using pressurized gases for SHIP. A closed loop circuit in SHIP using pressurized air, nitrogen, or carbon dioxide has not been investigated yet but seems promising from the above discussion. Open-to-air circuits concentrating solar energy for air heating is an interesting solution for SHIP for widespread usage of medium-temperature air in industrial processes.

FUNDING INFORMATION
The partial funding of the project TURBOSOL, Ref. IND2017/AMB7769 "Producci on directa de aire a alta temperatura y a presi on turboalimentada en colectores solares de concentraci on" from the Madrid autonomous government through the research program on doctoral industry is highly appreciated. No additional funding excepting from the Universidad Carlos III de Madrid, has been received.

CONFLICT OF INTEREST STATEMENT
The authors have declared no conflicts of interest for this article.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study ORCID Antonio Lecuona-Neumann https://orcid.org/0000-0002-5763-4982 Antonio Famiglietti https://orcid.org/0000-0002-0631-345X RELATED WIREs ARTICLES Thermodynamic cycles for solar thermal power plants: A review Solar thermal CSP technology