Characterization of Nonradiative Recombination Centers in GaPN Alloys by Two-Wavelength Excited Photoluminescence

概要

Among renewable energies, power generation using solar cells is drawing attention. However, the problem with solar power generation lies in its generation cost. It is essential to reduce the power generation cost of renewable energy, when making it the main power source. Higher efficiency of solar cells has been desired with the aim of reducing the cost of photovoltaic power generation. One of the attempts to achieve high efficiency is an intermediate band solar cell (IBSC). Although there is a rapid technological progress of growth technique, the performance of IBSCs shows less than the expectation partly due to the presence of deep level defects which originate from point defects, dislocations and so on. The deep level defects which are located below the conduction band minima or intermediate band act as nonradiative recombination (NRR) centers. It is crucial to reduce the NRR centers to improve photoconversion efficiency of IBSCs.
GaPN alloys can be a material candidate for efficient IBSCs due to the cascade excitation of photo-generated carriers via the IB in addition to the direct excitation of carriers from the valence band (VB) to the conduction band (CB). Doping with nitrogen into GaP can lead to an extremely large bandgap bowing and demonstrate quasi-direct bandgap behavior. Isolated N atoms at a low concentration induces highly localized isoelectronic traps, and the formation of the impurity band occurs sequentially by spatial and energetic overlap among different isoelectronic trap centers with increasing N concentration.
In this research work, a purely optical and non-destructive technique of two-wavelength excited photoluminescence (TWEPL) method was used for the detection and characterization of NRR centers. We have constructed a schematic energy diagram of NRR processes among CB, IB and VB considering the impact of the irradiation of an intermittent below-gap excitation (BGE) light on the IB PL under two kinds of above-gap excitation (AGE) light (CB or IB excitation). By combining experimental data of time-resolved photoluminescence (TRPL), solving the rate equations based on the postulated energy diagram and fitting the results with experimental data of TWEPL, the relative contributions of NRR parameters are evaluated.
The key contributions of this work are as follows:
The GaPN sample was grown by metalorganic chemical vapor deposition (MOCVD). A 300 nm-thick GaP buffer layer and a 500 nm-thick GaP₁-xNx (x = 0.105%) were grown in sequence on a 400 μm-thick sulfur doped GaP substrate with an orientation of (100). Trimethylgallium, phosphine, and dimethylhydrazine were used as Ga, P, and N sources, respectively. The sample was excited by two types of AGE light sources: 2.33 eV (532 nm) for IB and 3.49 eV (355 nm) for CB excitation. Another intermittent BGE light was superposed over AGE at the same point of the sample. The NRR level was detected and evaluated by measuring the change in the PL intensity with the irradiation of BGE. Five types of lasers (0.81, 0.93, 1.17, 1.27, 1.46 eV) were used as the BGE light sources. The temperature effect due to BGE irradiation on the sample was eliminated by immersing it in liquid N2. Depending on the excitation energies of BGE, different contribution of one-level model and two-level model took place for CB and IB excitation. The effect of BGE energies depends also on the AGE excitation power density through a shift of Fermi level in the forbidden energy gap. These results are successfully interpreted by the distribution of NRR centers and NRR processes among CB, IB and VB. Finally, the NRR parameters were evaluated by rate equation analysis. The rate equation analysis has been performed to justify the phenomenological recombination model combining the result of TRPL measurements.
Similarly, the carrier recombination model of GaP₁-xNx (x = 0.56 and 0.75%) samples are interpreted from the experimental results of TWEPL and the NRR parameters are evaluated by TRPL measurements and rate equation analysis based on the recombination model.
It was observed from the experimental results that with increasing N concentration, the PL spectrum of GaP₁-xNx transforms from superposition of discrete peaks to a broader PL band corresponding to the formation of IB. The increase in the N concentration of GaPN alloy causes a shift of the IB emission toward lower energies. When BGE light was irradiated on GaP₁-xNx (x = 0.56 and 0.75%) samples, the PL intensity increased for all five BGE, whereas in the sample with a N concentration of 0.105%, such a clear tendency was not shown for all BGE lights, and different BGE energy dependence was observed. This is because the energy distribution of NRR centers is different in each sample, and different carrier recombination processes can be considered for each sample. The evaluated NRR parameters by TRPL measurements and rate equation analysis of three samples shows that the defect density of GaP₁-xNx (x = 0.56%) sample is the lowest among three samples. The experimental results shows that GaP₁-xNx (x = 0.56%) sample has the highest emission efficiency with a low NRR density.
It was found from the evaluation that the energy distribution of the carrier recombination process via IB was different due to a slight change in N concentration. For actual application of GaPN alloy for IBSC, the Fermi level should be very near or inside the IB. The reduction of the defect density inside the forbidden gap is also important to raise the Fermi level up to IB. This work also implies that the combination of TRPL results with TWEPL is important to improve the accuracy toward quantitative determination of NRR parameters. However, more samples with higher N concentrations should be investigated to find the best condition of GaPN alloys for the application of IBSC. Considering the detection sensitivity, merit of non-contacting and non-destructive measurement, the way of utilizing the IB luminescence as a probe of characterizing defect levels is beneficial not only for GaPN but also InAs/GaAs, N δ-doped GaAs superlattice and other IBSC materials.

この論文で使われている画像