A Brief Review of High Efficiency III-V Solar Cells for Space Application

The demands for space solar cells are continuously increasing with the rapid development of space technologies and complex space missions. The space solar cells are facing more critical challenges than before: higher conversion efficiency and better radiation resistance. Being the main power supply in spacecrafts, III-V multijunction solar cells are the main focus for space application nowadays due to their high efficiency and super radiation resistance. In multijunction solar cell structure, the key to obtaining high crystal quality and increase cell efficiency is satisfying the lattice matching and bandgap matching conditions. New materials and new structures of high efficiency multijunction solar cell structures are continuously coming out with low-cost, lightweight, flexible, and high power-to-mass ratio features in recent years. In addition to the efficiency and other properties, radiation resistance is another sole criterion for space solar cells, therefore the radiation effects of solar cells and the radiation damage mechanism have both been widely studied fields for space solar cells over the last few decades. This review briefly summarized the research progress of III-V multijunction solar cells in recent years. Different types of cell structures, research results and radiation effects of these solar cell structures under different irradiation conditions are presented. Two main solar cell radiation damage evaluation models—the equivalent fluence method and displacement damage dose method—are introduced.


INTRODUCTION
Space solar cells, being the most important energy supply unit, have been employed in spacecrafts and satellites for over sixty years since the first satellite was launched in 1958 [1]. It has been developed from the initial single junction low efficiency silicon solar cells [2] to the now high efficiency multi-junction III-V compound multi-junction solar cells [3]. The main objectives of space solar cell development are directed toward to improving the conversion efficiency and reducing the mass power ratio and increase the radiation hardness [4][5][6][7]. At present, the highest conversion efficiency of solar cells is 47.1% achieved by six-junction inverted metamorphic (6 J IMM) solar cells under 143 suns [8]. The high-efficiency III-V triple-junction cells are also becoming the mainstream of space solar cells. The best research-grade multi-junction space solar cell efficiency so far is 35.8% for five-junction direct bonded solar cell and 33.7% for the monolithically grown 6 J IMM multijunction solar cell [9,10]. Despite the high fabrication cost, they offer excellent performance and reliable stability for space missions [11][12][13]. GaInP/GaAs/Ge (1.82/1.42/0.67 eV) lattice-matched triple-junction cells are well established with efficiencies of over 30% and fulfilled many space applications in the past two decades. However, the current mismatch between its subcells makes it difficult to improve the conversion efficiency further [14]. New structures of current matched or lattice mismatched solar cell structures and different fabrication methods are proposed to overcome this problem, such as the metamorphic (MM) growth method [15], mechanical stack [16], wafer bonding technology [17], etc.
While improving the efficiency of space solar cells, the radiation resistance should also be considered. In-orbit solar cells suffer from irradiation damages due to high energy protons and electrons in the earth's radiation belt and cosmic rays [18,19], and consequently, the photoelectric performance of solar cells will be degraded. The main reason of the degradation of solar cell performance is due to the radiation-induced displacement damage in the solar cell lattice, resulting in a decrease in the lifetime of the photo-generated carriers [20][21][22]. Therefore, the degradation mechanism and performance of solar cells under an irradiation environment must be explored, and it is necessary to apply radiation hardening methods before the space mission starts. The degradation of electrical performance in solar cells directly affects the life-time of space missions. The researchers aimed at improving the radiation resistance of solar cells by adding a certain thickness of protective cover to the solar cell to shield the damage of certain particles [23], using back-surface (BSF) [24] or distributed Bragg reflector (DBR) [25], and thinning the base layer thickness of the current-limiting subcell [26], or using the p-i-n structure and different doping methods for multi-junction solar cells [27]. The experimental observations show that annealing of the multi-junction solar cell can restore certain electrical properties after being radiated by high-energy particles [28].
In recent years, various new types of multi-junction solar cells with different combinations of materials have been developed by different research groups, and the expectations for future development are different. Solar cell conversion efficiencies are rapidly being updated, and scientists are still struggling to come up with solar cells which have high conversion efficiency and possess good radiation resistance. Although there are several reviews available which cover the manufacturing, efficiency, and application prospects of photovoltaic modules [29,30], the new types of high efficiency space solar cells based on III-V compound materials have not been summarized yet. This review attempts to give a brief review on different types of space solar cells and emphasize the high energy particle irradiation effects of solar cells and recent results on the most promising types of solar cells, including dilute nitride, metamorphic, mechanical stack, and wafer bonding multijunction solar cells.

DIFFERENT TYPES OF HIGH-EFFICIENCY SOLAR CELLS
With the improvement of the manufacturing process and deposition technology of materials, the solar cells industry has developed tremendously. Solar cell materials are developed from a single material (single crystal Si, single-junction GaAs, CdTe, CuInGaSe, and amorphous Si:H) to compound materials, such as III-V multi-junction solar cells, perovskite cells, dye-sensitized cells, organic cells, inorganic cells, and quantum dot cells [31][32][33]. The structure of solar cells also forms homogeneous junction cell to heterogeneous junction solar cell, Schottky junction solar cell, compound junction solar cell, and liquid junction solar cell. In the purpose of its usage, it has also been developed from flat cells to concentrator cells and flexible cells [34,35].
The silicon solar cells were used as the first choice in the spacecraft since the first solar-powered satellite was launched in 1958. The Soviet Space station, MIR, was launched in 1986, was equipped with 10 kW GaAs solar cells, and the power per unit area reached 180 W/m 2 [36]. Then, the fabrication technique of GaAs-based cells experienced changes from Liquid Phase Epitaxy (LPE) to metal organic vapor phase epitaxy (MOVPE), from homogeneous epitaxy to heterogeneous epitaxy, from single junction to multi-junction structure [37][38][39]. Notably, their efficiency was continuously improved from the initial 16-25%, and over 100 kW industrial-scale power output per year has been reached [40]. Higher efficiency reduces the size and weight of the array, increases the payload of the spacecraft and results in lower costs for the entire satellite power system. Therefore, GaAs-based solar cells are widely used in space systems and continue to be used today [41][42][43]. Comparing with silicon solar cells, GaAs solar cells have the following advantages [42]: (1) Higher photoelectric conversion efficiency.
(3) The band-gap tailoring by controlling the composition and doping of material. (4) Superior radiation resistance.
However, the processes involved in GaAs solar cell fabrication are complicated, and its cost is much higher than that of silicon solar cells owing to the expensive equipment and material preparation. Therefore, GaAs solar cells cannot be widely utilized in the civil market. Nevertheless, GaAs solar cells have gradually replaced silicon solar cells in the aerospace field, where higher cell efficiency and better radiation resistance are needed.
The loss in the efficiency of solar cells can be divided into two parts: the unabsorbed loss and excessive energy loss. When the photon interacts with the semiconductor materials, where the photon energy is smaller than the bandgap width, the valence band electrons are not excited, and they do not generate an electron-hole pair to form an electrical current. However, when the photon energy is greater than the gap width, the excess energy is lost in the form of phonons or heat [44]. Fortunately, multijunction solar cells successfully solved this problem. Semiconductor materials with different band-gaps are composed from top to bottom from large to small band-gaps, and the higher energy photons are absorbed by the top large band-gap material. The lower energy photons go through the upper large band-gap material and reach the appropriate band-gap width to generate power. Therefore, for multi-junction solar cells, finding a current matching and lattice matching cell material is the critical and general focus [14,45]. The following sections present a brief introduction of different types of multijunction solar cells in terms of their performance.  [46]. Since then diluted nitride GaInNAs materials have been widely used in heterojunction bipolar transistors (HBTs) [47] and lasers [48], where GaInNAs HBTs base layer can reduce the open voltage and run under low working voltage. These features of diluted nitride GaInNAs materials are also useful in wireless communications and power amplifier applications. GaInAsN is a direct band-gap semiconductor material, which can change its band-gap by adjusting the component content of nitrogen and indium while keeping its lattice constant matching to conventional substrate materials such as GaAs and Ge. These advantages bring great potential for using 1.0 eV subcell in a high efficiency multijunction solar cell [49].
The Ga 1-x In x N y As 1-y is used as a sub-cell material for GaInP/ GaAs/GaInNAs/Ge four-junction solar cell by NREL [50]. When y 0.3x, the lattice constant of Ga 1-x In x N y As 1-y matches GaAs and Ge, which is an ideal material to construct a GaInP/GaAs/ GaInNAs/Ge (1.88/1.42/1.05/0.67 eV) four-junction solar cell with band-gap matching. Figure 1A shows the representative GaInNAs devices structure grown by MOVPE. The device is grown with dimethylhydrazine (DMHy) as the nitrogen source. At the same time, the experimental results showed that the remaining factor of GalnAsN cell efficiency is 0.93 and 0.89 after 5 × 10 14 and 1 × 10 15 e/cm 2 electron fluence of 1 MeV electrons irradiation, respectively [51]. The specific degradation of the device parameters is summarized in Table 1. The results showed that this type of cell structure possesses superior radiation resistance comparing to the traditional lattice matched multijunction solar cell.
However, difficulties in epitaxial growth of diluted nitride materials have prevented its further development. It is also found that the diffusion length of minority carriers is short, the internal quantum efficiency is low for the GaInNAs subcell, and these poor minority carrier transmission characteristics resulted in lower current and voltage. Since the current of a multijunction solar cell is determined by the smallest sub-cell, the low current in GaInNAs subcell has a significant impact on the overall cell performance [51]. The quality of GaInNAs solar cells was improved by fabrication of p-i-n structure cells [52], annealing [53], doping Sb in GaInNAs material [54], and changing substrate epitaxial orientation [55]. Miyashita et al. demonstrated a solar cell with p-i-n (p-GaAs/i-GaInAsN (Sb)/ n-GaAs) structure, and tested the cell under the AM1.5 spectrum.
The short-circuit current density (J sc ) reached 21.5 mA/cm 2 , the open-circuit voltage (V oc ) reached 0.42 V, and the filling factor (FF) reached 0.71 [56]. They further optimized the content of Sb and found that the content of Sb was less than 1%, which is more beneficial to improve GaInAsN crystal quality and device performance [57]. Han et al. found that GaAsN material grown by (311) B substrate epitaxy could not only improve the incorporation efficiency of nitrogen, but also effectively enhance the carrier lifetime of the material [58]. Although the theoretical conversion efficiency of GaInNAs multijunction solar cell could reach 41%, the actual growth problem still needs to be overcome, which needs to be focused on as a development goal in the future.

Mechanically Stacked Solar Cell
According to the theoretical calculation, the optimum bandgap energy for top and bottom subcell for a tandem multijunction solar cell is 1.65-1.8 eV and 1.0-1.5 eV, respectively, and the conversion efficiency of this structure reaches 32.5% under 1-sun AM0 spectrum [59]. Typically, the III-V compound material based multijunction solar cells are fabricated by MOVPE or molecular beam epitaxy (MBE) techniques, where the lattice matching and energy matching between subcells is a critical  problem. The presence of mechanical stacks makes this feasible and allows to use of lattice-and current-mismatched combinations of semiconductor materials. By using the mechanical stacking method, III-V compound materials can be stacked together regardless of their bandgap energy and lattice constants. This technique can reduce the production cost significantly, and the substrate removal process greatly reduces the weight of solar cells, which is useful if used in space solar cell applications [60][61][62]. Still, the utilization of multiple substrates requires the removal of substrates, thereby, it makes the solar cell fabrication process complicated and affects the interface quality [63]. At the same time, a complicated design requirement becomes an obstacle for large-scale applications of mechanical stacked solar cells in space [64]. Since each subcell in a multi-junction solar cell consists of a p-n junction, if individual subcells directly stacked together in series, it will form a reverse p-n junction between subcells which block the current flow. Therefore, interconnecting individually processed solar cells keeping both electrical and optical properties still is a technical challenge need to consider with high conductivity and low transmission loss [65]. This problem can be solved by adding a tunnel junction between the subcells [66]. The 30% conversion efficiency of III-V//Si multi-junction solar cells using smart stack technology have been reported [67]. A schematic diagram of InGaP/AlGaAs//Si triple-junction solar cell is shown in Figure 1B. A "smart stack," "areal current matching," and "solar concentration" two-terminal GaAs/Si tandem solar cell, having potential to achieve an efficiency of approximately 30%, are proposed and the indoor experiment proves its feasibility, although the experiment result is lower than the simulation [64]. These types of mechanical stacking processes are quite flexible and allow different cells to be combined, and consequently reduces the fabrication cost and increases the cell efficiency.

Wafer Bonded Multijunction Solar Cell
In multi-junction solar cells, the lattice dislocations induced by the lattice mismatch during the epitaxial growth reduce the material quality and deteriorates the device performance. The wafer bonding technique, which refers to the physical integration of the two different materials, overcomes this issue very well and allows the formation of a monolithic multijunction solar cell structure without both electrical and optical losses, regardless of the different lattice constant of subcells [68]. As early as 1986, Lasky et al. demonstrated surface treated silicon wafer bonding at room temperature and obtained very good bonding strength through high-temperature annealing [69]. The freedom of material selection in the process of semiconductor device design is greatly improved because the bonding technology can realize the lamination structure of materials with different thermal expansion coefficient and lattice constant, and can limit the dislocation and defect to the area near the bonding interface. Figure 2A shows the structure of GaInP/GaAs//InGaAsP/ InGaAs wafer-bonded four junction solar consists of GaInP/ GaAs and InGaAsP/InGaAs dual junction cells, whereas Figure 2B shows the degradation of InGaAsP and InGaAs subcell electrical properties under 1 MeV electron irradiation. The result shows that the bonding interface has no effects on overall cell performance and radiation resistance, the main degradation happened in the third and fourth subcell, and InGaAsP subcell has a superior radiation resistance than InGaAs cell owing to In-P bonds [70].
At present, the wafer bonding technology has been studied extensively and widely used in the structural design and integration field of microelectronics and optoelectronics. A variety of bonding methods have been developed, however, they can be mainly divided into two categories: direct bonding and intermediate-layer bonding [71]. The direct wafer bonding process includes cleaning and activating two polished wafers and sticking them together at room temperature. Heterogeneous integration of multi-junction solar cells has to meet three stringent conditions: good mechanical strength at the bonding interface, high optical transmittance, and low resistivity. It does not work without each condition being met. Intermediate-layer bonding technology introduces a layer with good ductility and adhesion to alleviate stress and improve the bonding interface [72]. GaInP/GaAs//GaInAsP/GaInAs four-junction waferbonded concentrator solar cells with an efficiency of 46% at 508 suns have been reported [73]. It demonstrates that the wafer bonding is a feasible method to combine lattice mismatched

Inverted Metamorphic and Upright Metamorphic Solar Cells
Another approach of solving current matching issue in multijunction solar cell is the metamorphic growth method, which involves applying a compositionally graded buffer (CGB) layer between the lattice mismatched subcells [74]. The main purpose of using a CGB layer is to distribute the strain relaxation layers into a lattice constants gradually changed thick buffer layer instead of growing a highly lattice mismatched layer onto its under layer [75]. According to the growth method, the metamorphic growth technique can be divided into two types: IMM [76] and upright metamorphic (UMM) [77]. In the IMM method, for example GaInP/GaAs/InGaAs triple junction solar cell, a CGB layer is inserted between the GaAs and InGaAs subcells and the growth direction of solar cell epilayers is in the inverse direction, in order to delay the strain relaxation in CBG layer at the later stages of growth process. However, the IMM method requires an additional step of substrate lift-off before solar cell device fabrication processes. On the other hand, in the UMM method, the growth direction is from bottom subcell to top subcell, so it requires an even higher quality CBG layer to control the strain relaxation. Figures 3A,B show the general structure of IMM and UMM cells, respectively. Comparing to traditional lattice matched (LM) solar cells, metamorphic multijunction solar cells are aiming to achieve current matching between its subcells and expected to have higher conversion efficiency. Experimentally, NREL has shown that IMM solar cells have a conversion efficiency of 40.8% for triple-junction solar cells at high concentration [15] and 47.1% efficiency for 6 J solar cells at 143 suns concentration [8]. Furthermore, IMM cell can reduce the cell weight by removing the substrate and increase the mass power ratio. IMM cells also can be utilized to develop flexible solar cells for any non-flat surfaces. These high efficiency, flexible, and lightweight properties of IMM multi-junction solar cells make them the most promising for space applications [ [14,78,79]]. For UMM multijunction solar cell, the conversion efficiency exceeded 31% under one Sun AM0 spectrum compared to traditional lattice matched solar cell conversion efficiency is limited to 30% [80]. The critical point of increasing UMM solar cells is improving the quality of CGB layer to suppress lattice dislocation and threading dislocations induced by strain relaxation. However, the fabrication process of UMM solar cells is similar to the matured fabrication technology of LM cells, therefore, this advantage makes the UMM multi-junction solar cell utilization possible for potential applications [77].
The reported experimental efficiency, theoretical limits, and their advantages of all above mentioned solar cells in Lattice Matched GaInNAs Multi-Junction Solar Cells, Mechanically Stacked Solar Cells, and Wafer Bonded Multijunction Solar Cells are presented in Table 2. According to the Shockley-Queisser balance model, the theoretical efficiency limit of single-junction, triple-junction, and four-junction solar cells are 33.5%, 56% and 62%, respectively [85]. Along with the rapid development of new materials and high quality fabrication technologies, all these solar cells could achieve higher efficiency and better performance. Furthermore, UMM and IMM solar cell efficiencies tend to surpass LM cells even more as their manufacturing technology continues to be innovated and developed, and they are expected to become the next generation space solar cells.

STUDIES ON RADIATION EFFECTS OF SOLAR CELLS
The solar cell arrays in a spacecraft are exposed to a harsh space environment during its mission. Therefore, radiation resistance is also a critical index to evaluate the quality of space solar cells. The main reason for the performance deterioration of space solar cells is the irradiation of high-energy particles, that include electrons (energy up to 10 MeV) and protons (energy up to several hundred MeV) from the earth's radiation belt, as well as solar cosmic rays (energy up to GeV) [86]. When these high-energy particles collided with the cell materials, the transmitted energies cause the lattice atoms to shift their original positions and form displacement damages, and, consequently, the diffusion length of minority carriers decreases and the performance of solar cells is degraded [87]. Therefore, only solar cells with high conversion efficiency and good radiation resistance could be employed as space solar cells.

Radiation Effects of Double-Junction GaAs/ Ge Solar Cell
Messenger et al. investigated the electron and proton irradiation effects of double-junction GaAs/Ge solar cell with different energies [88]. The cell performance is degraded with an increase of the particle fluence under both electron and proton irradiation, as shown in Figure 4. Additionally, electron irradiation with greater incident energy resulted in more severe deterioration of cell performance for the same irradiation fluence. For proton irradiation, on the other hand, low energy protons produced a bigger reduction in cell performance when the proton energy is in the range of 0.2-9.5 MeV. Proton irradiation experiments on GaInP/GaAs/ Ge, GaAs/Ge solar cells with proton energy of less than 200 KeV are reported and it was found that the low energy particle irradiation caused defects in different subcells and different regions in the tandem multi-junction solar cell [23,89].
The radiation-induced defects in the base and the emitter layers form non-radiative recombination centers and capture photo-generated electron-hole pairs before they are collected by the junction region, and eventually reduce the short circuit current. On the other hand, the radiation-induced damages in the junction region mainly cause the degradation of open circuit voltage by introducing deep energy levels into the band gap and accelerating the recombination of valence band holes and conduction band electrons [89].
The effects of 40, 100, 170 keV energy proton irradiation on GaAs/Ge cells with different fluences were discussed, the degradation of spectral response results indicated that the largest damages are caused by 170 keV protons, and the lowest by 40 keV protons, in the long wavelength range (720-900 nm). The degradation effect on the normalized maximum power (P max ) is the largest for the 170 keV protons and lowest for 100 keV protons because irradiation with 170 keV protons produces the most severe damage in the junction region of the cells. The short circuit current decreases with increasing proton energy under the energies of 40-170 keV proton irradiation and the degradation extent of V oc is the largest for 170 keV proton because they produce defects with deep energy levels in the space charge region, accelerating the recombination of electrons and holes, which is also the reason for the significant decline in P max for solar cell [23].

Radiation Effects of Lattice Matched GaInP/ GaAs/Ge Triple-Junction Solar Cell
Sharps et al. investigated the electron and proton radiation effects of GaInP/GaAs/Ge solar cells with different energies [91]. Same as that of GaAs/Ge double-junction solar cell, the P max of solar cells declined with the increase of the particle fluence for the same energy. For 1-12 MeV electron irradiation, the degradation of cell performance is severer for higher energy electrons for same irradiation fluence, as shown in Figure 5A. The degradation of P max is 9 and 13% for 1 MeV electrons at fluences of 5 × 10 14 and 1 × 10 15 e/cm 2 . However, for proton irradiation, the degradation of cell performance is bigger for lower energy protons in the energy range of 50-200 keV, as shown in Figure 5B. The main reason for this is that less relative damage occurs for energies below 200 keV, and lower energies would have more of an effect on the emitter as compared with the base region. At the same time, GaInP top cell and GaAs middle cell current matching point was studied in more detail by quantum efficiency measurements. It was found that the crossover from GaInP current-limited subcell to GaAs currentlimited subcell occurs at 2 × 10 15 e/cm 2 for the 1 MeV electrons [90]. This result indicated that advanced triple junction solar cells with current matching at end of life (EOL) can be achieved by reducing the amount of beginning of life (BOL) current mismatch, which is helpful for designing high efficiency radiation hardened space solar cells.
Wang et al. investigated the electron irradiation effects of GaInP/GaAs/Ge solar cell with 1.0-11.5 MeV energy electron beams [91]. The degradation of normalized P max of GaInP/GaAs/ Ge solar cell under different fluence of electron irradiation is shown in Figure 6A, whereas the degradation of external quantum efficiency (EQE) changes irradiated with 1.8 MeV electron beam with different fluences are shown in Figure 6B.
The degradation effects of electron irradiation of the P max and the EQE of LM solar cells increase with the increase of irradiation fluences and electron energy, as observed in GaAs/Ge doublejunction cell. Figure 6C exhibits that the EQE of GaAs middle cell degrades more than GaInP top cell under same irradiation fluence which is indicating the radiation resistance of GaInP/ GaAs/Ge triple-junction cells is dominated by the GaAs middle cell.   [92]. In this study, the electrical parameters V oc , I sc (shortcircuit current) and η decayed o 89.3%, 96.3%, 84.1% and 88.6%, 98.6%, and 79.5% of its original values for one group of cells have higher efficiency at BOL and another group of cells have higher efficiency at EOL, respectively, when the electron irradiation fluence reaches 1 × 10 15 cm −2 . These details of degradation of electrical parameters of these two types of cells are summarized in Table 3.
Imaizumi et al. studied the radiation response of In 0.5 Ga 0.5 P, GaAs, In 0.2 Ga 0.8 As, and In 0.3 Ga 0.7 As single-junction solar cells, whose materials are also used as component subcells of inverted metamorphic triple-junction solar cells, and results show that the photo-generation current in the InGaAs bottom subcell of InGaP/GaAs/InGaAs IMM3J cells was severely damaged under the electron and proton radiation, which can be attributed to the stronger decrease of minority-carrier diffusion length in InGaAs compared with that in InGaP and GaAs subcells after irradiation [93]. By comparing the irradiation resistance of two InGaAs cells (In 0.2 Ga 0.8 As and In 0.3 Ga 0.7 As cells), GaAs and InGaP cells, it was found that radiation resistance of these two InGaAs cells is approximately equivalent to InGaP and GaAs cells from the initial material qualities. However, the InGaAs cells show lower radiation resistance especially for the I sc comparing to InGaP and GaAs cells due to the bigger decrease of minoritycarrier diffusion length in InGaAs materials. And the InGaP and two InGaAs cells exhibited equivalent radiation resistance of V oc , but with different degradation mechanisms.
Zhang et al. investigated the 1 MeV electron radiation effects of IMM GaInP/GaAs/InGaAs solar cells by electrical properties, spectral response, and photoluminescence (PL) signal amplitude analysis [94]. The results show that the electrical parameters of IMM solar cell decrease continuously with an increase in the electron fluence same as traditional LM GaInP/GaAs/Ge solar cells. As shown in Figure 7, P max illustrates the maximum degradation compared to V oc and I sc , and, V oc is degraded more compared to I sc . This phenomenon is explained as V oc is the sum of three series sub-cell voltages, where the I sc is the smallest current produced in three series sub-cells. The In 0.3 Ga 0.7 As bottom subcell exhibited most severe damage in the irradiated IMM triple junction cells due to the drastic degradation of the effective minority carrier lifetime (τ eff ) of In 0.3 Ga 0.7 As subcell than that of GaAs subcell. Therefore, the radiation hardness of IMM GaInP/GaAs/InGaAs solar cell is mainly determined by the InGaAs subcell.

RADIATION DAMAGE EVALUATION OF SPACE SOLAR CELLS
To obtain better radiation hardened performance, it is essential to explore the radiation damage mechanism of the solar cells. The interaction of high energy charged particles in irradiation environment with the solar cell materials includes ionization and non-ionization (displacement damage) processes [95]. The displacement damage effect is the main reason for the degradation of solar cell performance. It is an effective method to explain the formation, distribution, and evolution of displacement defects after irradiation by different high-energy particles. To examine the At present, for space solar cells, two assessment methods for evaluating radiation damages of solar cells are available, which are through ground irradiation simulation experiments: equivalent fluence method and equivalent displacement damage method [88]. Both methods are used to study the radiation damage effect and to reveal the degradation mechanism of solar cells. They also provide theoretical guidance and an experimental basis for scientifically predicting the on-orbit behavior of solar cells.

The Equivalent Fluence Method
The equivalent fluence method was proposed by Tada et al. from Jet Propulsion Laboratory, California Institute of Technology [96,97]. The key point of this approach is the corresponding relative damage coefficient, which is related to the radiation damage effects caused by different types of charged particles with a different energy to relative damage coefficients.
In the first step, the critical fluence (ϕ) of electron or proton irradiation, which is corresponding to the electrical properties of the solar cell degraded to a specified level of its original value (such as 75% of I sc0 , V oc0 , P max0 ), has to be determined according to the experimental results [98]. Then, the relative damage coefficient (RDC) of different energy electron and proton regarding 1 MeV electron and 10 MeV proton is calculated. The ratio of the critical fluence for 1 MeV electrons to the critical fluence for other electron energies is taken as a measure of the RDC of electrons, and similarly, the relative damage coefficients of different proton energies are normalized regarding the 10 MeV proton critical fluence for RDC of protons, as shown in the following equations [88]: where ϕ e and ϕ p are the critical fluence for electrons and protons. The next step is using the orbital environment parameters to calculate the corresponding relative damage coefficients for omnidirectional particles on bare cells from the measured values for normally incident particles. By substituting the electron and proton environment parameters under consideration into following integration equations, one can obtain the equivalent 1 MeV electron fluence for the mission in question.
where ϕ e (E) and ϕ p (E) is the particle fluence of electron and proton, respectively at energy E. D e (E) and D p (E) are the relative damage coefficient of electron and proton, respectively. D pe is the proton to electron damage equivalency ratio, which converts 10 MeV proton fluence to an equivalent 1 MeV electron fluence. The calculation result of Eqs. 3, 4 are the 1 MeV electron fluence normally incident on solar cells that will cause same damage as the selected omnidirectional spectrum. The result of calculated relative damage coefficient for omnidirectional proton irradiation and electron irradiation of GaAs/Ge solar cells are shown in Figures 8A,B.
In the equivalent fluence method, it requires a sufficient measurement of data to calculate the RDC and generate a detailed degradation curve, as shown Figure 4, which shows eight proton energies (0.05, 0.1, 0.2, 0.3, 0.5, 1.3, and 9.5 MeV) and four electron energies (0.6, 2.4, and 12 MeV). Another relatively simple way of predicting the degradation of a solar cell performance regarding a given electron or proton fluence is using a semi-empirical equation [99]: where P 0 and P ϕ are the output power (also can be replaced with I sc and V oc ) of the solar cell before and after irradiation with different irradiation fluences ϕ, respectively. C and ϕ x are the fitting parameters for the same structure solar cell using a large amount of experimental data upon specific irradiation particles.

The Displacement Damage Dose Method
The displacement damage dose method was initiated by Naval Research Laboratory (NRL) [100]. The key point of this method is finding the non-ionizing energy loss (NIEL) values for different materials. By using the NIEL method, the radiation fluence of particles is converted into displacement damage dose (DDD), and the degradation curve of the electrical parameters of solar cells with the change of DDD can be obtained.
The NIEL values of different materials upon different particles with different energies can be calculated by using MULASSIS software [101] or the following equation [102]: where n is the atomic density of the target material, T d is the threshold energy to displace atom, Q max is the maximum energy that can be given to a recoil nuclei by an incident particle of a given energy E, G(Q) is the energy partition function and (dσ/ dQ) E the differential interaction cross section. Figure 9A shows the calculated results of electron and proton NIEL for GaAs with the energy range from zero up to 200 MeV. The DDD induced by irradiation particles can be calculated by the following equation using particle fluence: where D d (E) is the DDD, ϕ(E) is the particle fluence, and NIEL is the non-ionizing energy loss value of the target material. When Frontiers in Physics | www.frontiersin.org February 2021 | Volume 8 | Article 631925 9 calculating D d of proton irradiation, n value is 1, and the n value of electron irradiation is in the range of 1 to 2. Using NIEL of the same particle with different energies, the relative damage coefficient for different energies of a charged particle, i.e. proton, the irradiation can be obtained from using the following equation [99]: where D x→10 is the X MeV proton to 10 MeV proton relative damage coefficient, and NIEL is the corresponding nonionization damage energy for different energies. Electron to proton irradiation damage equivalency factor Rep can be calculated by following equation [99]: where Rep is electron to proton irradiation equivalent damage coefficient, D j is the actual DDD values by using electron irradiation and D i is the corresponding fitting values of DDD. By considering Rep degradation curves of electrical parameters of solar cells against the irradiation fluence can be described into a single curve against the D d . For example, Figure 9B shows the replot of Figure 4 using NIEL method. The DDD of solar cells in complex electron and proton environment can be calculated by Eq. 10, and applied to evaluate the degradation curve of the electrical performance of space solar cells with the DDD.
Besides, predicting the degradation of a solar cell performance regarding a given electron or proton fluence can also be conducted by using the following semi-empirical equation: where P max and P 0 are the output power (also can be I sc or V oc ) of solar cell before and after irradiation, D d is the given DDD regarding to the given irradiation fluence, respectively. C and D x are the fitting parameters obtained by on-ground experimental data. Eq. 4 is more versatile than Eq. 5 for different kinds of particles by using D d as variable for given irradiation condition.
Comparing to the equivalent fluence method, the NIEL approach requires fewer experimental measurements to successfully predict the radiation damage of different particles   [88]. The displacement damage dose method was successfully applied for predicting on-board solar cell measurements for GaAs/Ge and CIS solar cells at 500 × 67,300 km near equatorial orbit, shown in Figure 10.

The Development Trends of Space Solar Cells
There is no doubt that space solar cells should move toward higher efficiency, low cost and better radiation resistance. In this direction, many types of new technologies are trying to solve these problems. Currently, LM triple-junction solar cells are the main stream in space applications. In theory, the solar cell has more junction number has higher efficiency, but it is difficult to increase the number of cell junctions in real cell fabrication. The theoretical limit of N-junction (N for the infinite) solar cells conversion efficiency can reach 68.2% [3]. But the cell fabrication becomes more difficult when the number of junctions is increased, especially for more than 5-6 junctions. The efficiency of multi-junction solar cells from single-junction to six-junction is presented in Table 4. Besides, apart from the high cost of III-V materials, the price of GaAs is ten times than that of Si, the growth of III-V materials requires expensive equipment, and hence, the production cost of multijunction solar cells is very high and mostly used in space applications now. Therefore, in the future, the development trends of III-V multijunction solar cells are low-cost, high efficiency, high radiation resistance and simple fabrication method.

CONCLUSION
III-V multijunction solar cells are the primary power supply for space application due to its super high photoelectric conversion efficiency and better radiation resistance. Despite the high fabrication cost, it is widely used in different space applications. New types of space solar cells with new materials and new structures are continuously coming out and the performance characteristics of various space solar cells have been extensively studied. Owing to the development of the balance between the bandgap matching and lattice matching, the structure of solar cells is continuously optimized, more suitable new materials are discovered, and more mature manufacturing processes are utilized. The highest conversion efficiency of solar cells is constantly being refreshed. Comparing to conventional silicon solar cells, the conversion efficiency of III-V multijunction solar cells is significantly improved. Till now, the highest efficiency of crystalline silicon heterojunction solar cell reached 25.6%, and the world record of 47.1% conversion efficiency is achieved by six-junction inverted metamorphic solar cells under 143 suns. Furthermore, latticematched GaInP/GaAs/Ge triple junction solar cell fabrication technology is becoming more and more mature with large scale of mass production while keeping the conversion efficiency over 30%. The space radiation environment is the main threat for inorbit solar cell performance and lifetime. The high energy particles, such as electrons and protons, induce displacement damages in different regions of solar cell structure and lead to electrical and spectral performance degradation of solar cells. The main reason for the deterioration of space solar cells is that the radiation induced displacement damages forms nonradiative recombination centers and reduces the minority carrier lifetime, and, consequently, results in a reduction of solar cell electrical and spectral parameters. The equivalent fluence method and the displacement damage dose method FIGURE 10 | The displacement damage does method measurements for GaAs/Ge and CIS solar cells [88]. are two main approaches for evaluating solar cell radiation effects and degradation of cell parameters. Especially, the displacement damage dose method is a more convenient and effective tool. Although the radiation effects of solar cells are widely studied, the radiation damage mechanism of multijunction solar cells with different materials and different structures are not fully explored yet. More experimental and theoretical studies are still needed for further investigating feasible radiation hardening methods.

DATA AVAILABILITY STATEMENT
The data that supports the findings of this study are properly cited with literature and also available from the corresponding author upon reasonable request.

AUTHOR CONTRIBUTIONS
JL, AA, and YL designed the research, conducted the literature review and wrote this manuscript. All authors contributed to the literature review, discussion of the results and edited the manuscript.