InGaN alloys have great potential in the field of photovoltaics due to their excellent light absorption and tunable bandgap (0.7-3.4eV) which spans the visible spectrum. However, the growth of this material remains a challenge due to thermal decomposition, indium surface segregation, and phase separation. To combat these issues, low substrate temperatures and high growth rates must be implemented in combination with in situ surface monitoring via RHEED. Metal modulated epitaxy (MME) has been shown to achieve single-phase InGaN alloys throughout the miscibility gap with sub-nm RMS roughness.

 

In this study, MME is applied to the growth of n-GaN/i-InGaN/p-GaN solar cells. Unintentionally doped InGaN layers 50/200nm thick with 16% In composition are grown on a 3μm n-type GaN template. The topmost layer consists of p-type GaN with varying hole concentrations. Transient RHEED signals are monitored closely during InGaN growth to prevent indium surface segregation and ensure a smooth film.

 

The absorption characteristics are determined as shown in Figure 1, and the material is characterized via AFM, XRD and PL prior to being fabricated into device structures. The InGaN layers are found to be approximately 80% strain relaxed as determined by XRD analysis of the (20-25) reflection (Figure 2). During device testing, the solar cells exhibit low turn-on voltages below the expected value, as well as low external quantum efficiency. These inconsistencies indicate carrier loss due to electrical shorts and defects. The scanning TEM images in Figure 3 show striations in the InGaN layer similar to those in a superlattice structure but with a period that does not correspond to the shutter cycles. Even though the x-ray diffraction indicates predominantly relaxed epitaxy, TEM images are unable to resolve individual threading dislocations. It is not clear at present if this is indicative of extremely large dislocation density or a new method of relaxation related to the above mentioned striations. Finally, post-growth annealing processes are explored as possible ways to improve device performance.