1. Introduction

High-concentration photovoltaic (HCPV) cells have been widely prepared using group III-V compound semiconductors, such as GaInP/GaInAs/Ge tandem solar cells, to achieve very high efficiency [1]. In PV concentration systems, sunlight is focused by lens or mirrors to increase incident light flux density to deliver more power compared with nonconcentrated systems. When the open-circuit voltage (V_OC) of PV cells is logarithmically increased with concentrated sunlight, power conversion efficiency (PCE) is enhanced, and cell cost is reduced by saving the cell area. To achieve even higher efficiency, a single material system of the In_xGa_1-XN alloys simultaneously covers a wide solar-spectrum range by the combination of InN with GaN compound semiconductors, especially those larger than 2 eV and lower than 1 eV, which could be obtained for specific optimum multi-bandgaps for future tandem cells with ultra-high power conversion efficiency [2–4]. This field has drawn much attention [5–7]. However, there are still few reports on HCPV characteristics of InGaN-based materials. Although the first Si p-n junction solar cell was developed in 1954 [8], GaAs in 1956 [9], and concentrator tandem solar cells and organic PV cells in the 1990’s, the group III-nitride semiconductors for PV applications are still in its infancy. To date, the GaN-based materials, mostly grown on sapphire (Al₂O₃) substrates, have been widely applied to laser diodes, light-emitting diodes, high power convertors, UV-photodetectors, and spintronics recently [10,11]. This is because GaN-based materials have excellent properties, such as tunable bandgap engineering, high mobility, excellent chemical tolerance, superior radiation resistance, self-organized InGaN quantum dots and very high absorption coefficients (> 10⁵ cm⁻¹) near the material bandgap [10]. At present, however, there still lacks for suitable substrate to grow III-nitride materials with low structure defect density. Lattice constant and thermal coefficient mismatches induce threading dislocations and high background carrier concentration that limit the performance of InGaN-based PV devices [10]. Dahal et al. reported their efforts on 30-sun concentration levels by a multiple-quantum-well(MQW)-type cell with back metal reflectors. Although the photocurrent was enhanced 15% by the aluminum back reflectors, the fill factor (FF) value significantly decreased by 11% mainly due to serious recombination [12]. In this study, we proposed a blue-band MQW-type PV cell grown on patterned sapphire substrates (PSSs), which could efficiently improve internal quantum efficiency by decreasing the dislocation density [13]. In addition, the surface of PV cells was textured to allow less light reflection thereon. The solar response under concentrated levels up to 150-sun conditions shall be discussed.

2. Device fabrication and experiment methods

InGaN epitaxial layers were deposited on c-face sapphire substrates by a metalorganic vapor-phase epitaxy reactor. The substrate used in this study was PSS through the conventional photolithography and dry etching process. PSS featured truncated cone patterns in the closely triangular package with 3 μm spacing. The top and bottom diameters of the truncated cone were 2 and 3 μm, respectively, and the height was 1.5 μm. Figure 1 shows the schematic structure of the blue-band MQW-type InGaN-based PV cell. The deposition process involved a 30 nm thick low-temperature GaN nucleation layer, followed by a 1.6 μm thick undoped GaN (u-GaN) with chamber pressure at 500 torr and then a 3 μm thick u-GaN at 200 torr. Afterward, a p-i-n heterostructure consisting of 3 μm thick Si-doped n⁺-GaN (n~5 × 10¹⁸ cm⁻³), an undoped InGaN/GaN (2.5/14.5 nm for 10 pairs) multiple quantum well structure, and 200 nm thick Mg-doped p-GaN (p~5 × 10¹⁷ cm⁻³) were sequentially deposited. The peak value of the electroluminescence (EL) spectrum of the MQW-type PV cell was characterized at 448 nm. The fabrication process was similar to conventional nitride-based light-emitting diodes (LEDs) with lateral electrode pads. First, a 200 nm thick indium tin oxide (ITO) film prepared by RF sputtering was deposited on p-GaN to form the p-type ohmic contact. The cell area was defined into 1 × 1 mm² by photolithography. A self-aligned etching process was performed to expose the underlying n⁺-GaN layer by ITO etching solutions and inductively coupled plasma in sequence. On the ITO surface, circle patterns in the closely triangular package (both diameter and spacing were 5 μm) were formed by selectively etching to reduce light reflection at the cell surface. Afterward, the Cr/Au (50/1000 nm) bilayer metal was deposited on the ITO and the exposed n⁺-GaN to form the anode and cathode electrode, respectively [14]. The lateral electrode pads were in digitated configurations for better current collection. Finally, a 200 nm thick SiO₂ layer was coated using plasma-enhanced chemical-phase deposition to function as the antireflection and passivation layer.

 

Fig. 1 Schematic structure of fabricated InGaN-based PV devices.

Download Full Size | PDF

3. Results and discussion

Figure 2 shows the typical characteristics of current density and power density versus voltage (J-V and P-V, respectively) of MQW-type PV cells illuminated by the solar simulator, which was calibrated by the calibration cell of NREL with global air-mass 1.5 (AM1.5G) testing conditions. The measured V_OC, short-circuit current density (J_SC) and FF of the PV cell were 2.28 V, 1.27 mA/cm², and 0.74, respectively, corresponding to the PCE of 2.13%. Under AM1.5G conditions, the experiment results indicated that the photocurrent of our InGaN PV cells reached 63% of the theoretical value (approximately 2.0 mA/cm²). Although this was significantly higher than reported results with similar blue-band InGaN materials [15,16], the PCE was still far from the theoretical value (~4.5%) by single-junction materials with optical bandgap at 2.77 eV (448 nm) [17]. It was mainly limited by material qualities due to the large lattice mismatch between sapphire substrates and GaN-based semiconductors [10]. In p-i-n type PV cells with single InGaN absorption layers, the GaN/sapphire hetero-epitaxy limits the critical thickness of InGaN, especially when the indium contents are raised to convert more sunlight. That is, the epitaxial structure could be relaxed to induce leakage to lower the shunt resistance of PV devices. Noteworthly, the V_OC value in this study was competatively high enough compared with other similar PV cells with blue-band absorption layers. This could be attributed to the fact that material quality of GaN-based epitaxial layers was improved by the use of PSS [13].

 

Fig. 2 Typical J-V and P-V characteristics of the MQW-type PV devices illuminated by the AM1.5G conditions.

Download Full Size | PDF

With material properties of high thermal conductivity and low saturation current, PV cells made from InGaN alloys are expected to have excellent potential to operate under high solar concentration. Since GaN-based epitaxial layers are generally grown on sapphire or SiC substrates, mismatch-induced structure defects (e.g. threading dislocations) that play a key role degrade the cell performance. On the other hand, the sapphire substrates with low thermal conductivity may cause further degradation of performance of InGaN-based PV cells operating under high solar concentration. To study the concentrated solar response of MQW-type PV cells made from InGaN/sapphire-based materials, Fresnel lens were used to concentrated light up to 150-sun conditions by the AM1.5G spectrum, and the cell temperature was maintained at 25 °C. Figure 3 shows the typical J-V characteristics under concentrated conditions. In an equivalent circuit of a solar cell modeled by an ideal current source (J_SC) in parallel with a diode, the J-V characteristic is given by Eq. (1):

 

Fig. 3 Typical J-V characteristics under concentrated levels up to 150-sun conditions.

Download Full Size | PDF

From Eq. (2), V_OC increases logarithmically with J_SC (i.e., radiation intensity). Generally, at the high concentration ratio, the measured value of V_OC may be smaller than the 1-sun value. This is because cell temperature rises by the light with high energy at the high concentration ratio, and V_OC decreases. The decrease in V_OC with increasing temperature arises mainly from changes in intrinsic carrier concentration (n_i); the value of J₀ increases exponentially with decreasing T⁻¹, causing V_OC to decrease almost linearly with increasing T. The conversion efficiency is thus decreased by the decrease of V_OC at the high concentration ratio. In this study, the cell temperature was well maintained at 25 °C by using a cooling stage. Therefore, V_OC increased logarithmically with radiation intensity although sun concentrations increased to 150. The V_OC values increased from 2.28 to 2.50 V, and FF values decreased from 0.74 to 0.70 when the concentrated level was increased from 1- to 150-sun conditions, as shown in Fig. 4(a) . Compared with 1-sun illumination, the measured V_OC increased by 9.6% and measured FF decreased by 5.4% when the PV cells were tested under 100-sun conditions. This corresponded to an enhancement of PCE by approximately 4.7%. Peak PCE (2.23%) occurred at 100-sun conditions rather than 150-sun conditions, as shown in Fig. 4(b). The PCE rolling off as the sun concentration over 100 could be due to the reduced FF. The lower FF at high concentration could be attributed to the relatively significant power loss compared with low-concentration illumination due to high current. In other words, power consumption in the series resistance (R_S) increased with an increase in cell current (i.e., under high illumination) [18]. Therefore, it is important to reduce R_S further to allow for GaN-based PV cells operating under high concentration.

 

Fig. 4 The summarized solar parameters under concentrated illumination by the AM1.5G spectrum: (a) PCE in semi-logarithmical scales and (b) V_OC and FF in linear scales.

Download Full Size | PDF

To extract R_S from PV cells, J-V characteristics taken from solar cells under illumination with different concentrations were performed. The R_S value of the PV cells was determined to be 23.5 Ωcm² on average. This R_S value was larger than that of conventional PV cells (e.g., InGaP/GaAs/Ge tandem cells). In conventional tandem PV cells, tunneling junctions between subcells dominate R_S. In single-junction GaN-based PV cells, InGaN/GaN-based p-i-n heterojunctions have relatively larger band discontinuity and low-conductivity p-GaN contact layer, which result in high R_S. In addition, GaN-based tunneling diodes are important to future InGaN/GaN-based tandem cells. However, this research topic is still in its infancy. In addition to the R_S issue, the material quality is also important. In conventional PV cells, different semiconductor layers with tiny lattice mismatches between them are stacked by heteroepitaxy to form subcells on a substrate in sequence. Generally, the density of structure defects in GaAs-based cells is low enough; thus, performance is reasonably good. InGaN-based PV cells grown on sapphire substrates suffer from large lattice mismatches between III-nitride semiconductors and sapphire. Therefore, dense structure defects (e.g., threading dislocation) in InGaN-based PV cells significantly degrade the device performance. In a fabricated PV diode, extracting the ideality factor from J-V characteristics can be indirectly applied to evaluate the effect of material quality on device performance. The calculated V_OC values of PV cells based on different diode ideality factors (n = 2) under different concentrated levels are also shown in Fig. 4(a). In conventional GaAs-based tandem cells, the ideality factor is larger than unity and typically greater than 3. This could be attributed to the epitaxial interface profile and mismatched current in each subcell [19,20]. At low current levels, photogenerated carriers could be possibly recombined by defect centers. At high concentrated levels, the ideality factor increases rapidly with the increase of diode saturation current. In this study, the ideality factor was estimated to be 1.7 at high current levels (i.e., 100-sun concentration conditions) but 2.0 at low current levels (i.e., 10-sun concentration conditions). This indicated that the component of recombination current cannot be ignored in PV cells. In a heterojunction, many factors could result in an ideality factor much greater than 2, including tunneling through barriers, interface defects and charge-related defects with a pined Fermi level [1,12]. This ideality factor was markedly lower than those of our previous GaN-based PV cells grown on flat sapphire substrates [16]. In other words, PSS-grown GaN-based PV cells exhibited positive effect on material qualities and resulted in PCE enhancement.

4. Conclusions

We have demonstrated the InGaN-based solar cells operating under high concentrated conditions over 100-sun intensities of the AM1.5G spectrum. This result indicates that the InGaN/sapphire-based PV cells have great potential for HCPV applications. The blue-band MQW-type PV cells were deposited on PSSs and fabricated with a textured surface through selective etching of the ITO top contact layer. V_OC values increased from 2.28 to 2.50 V when the concentrated level was increased from 1- to 150-sun conditions, and FF values remained as high as 0.70. PCE was enhanced to 2.23% when PV cells were tested under 100-sun conditions. The PCE rolled off as the sun concentration over 100 could be due to the reduction in FF. The declined FF at high concentration could be attributed to the high series resistance.