Thermomechanical Lift-Off and Recontacting of CdTe Solar Cells

Abstract

Controlled delamination of thin-film photovoltaics (PV) post-growth can reveal interfaces that are critical to device performance yet are poorly understood because of their inaccessibility within the device stack. In this work, we demonstrate a technique to lift off thin-film solar cells from their glass substrates in a clean, reproducible manner by first laminating a polymeric backsheet to the device and then thermally shocking the system at low temperatures (T ≤ −30 °C). To enable clean delamination of diverse thin-film architectures, a theoretical framework is developed and key process control parameters are identified. Focusing on cadmium telluride (CdTe) devices, we show that the lamination temperature and device architecture control the quality of lift-off, while the rate at which the film stack is removed is controlled by the delamination temperature. Crack-free CdTe devices are removed and successfully recontacted, recovering up to 80% of the original device efficiency. The areal density of these devices is ∼0.4 kg m–2, a reduction of over an order of magnitude relative to their initial weight on glass. The framework developed here provides a pathway toward both the development of inexpensive, flexible PV with high specific power and the study of previously buried interfaces in thin-film architectures.

Exceeding 200 ns Lifetimes in Polycrystalline CdTe Solar Cells

“…To characterize the interfaces contributing to the very long lifetimes observed here, we applied a thermomechanical cleaving method to cleanly separate a number of devices at the MgZnO/CdSeTe interface. [ 43 ] This method exposes the critical p–n junction interface that changes during deposition, CdCl 2 annealing, and back‐contact processing for examination by XPS and other surface‐sensitive probes. Cleaving occurs in an argon‐filled glovebox that is connected to a surface analysis cluster tool.…”

Mitigating Measurement Artifacts in TOF-SIMS Analysis of Perovskite Solar Cells

“…The sample is then placed in an argon atmosphere glovebox (<2 ppm water typical) and cured (typically overnight). Once cured, the sample is submerged in liquid nitrogen, where the sample adhered to the epoxy cleaves at the back CdTe/CdS interface due to the large difference in thermal expansion coefficient between the film stack on glass and the epoxy + metal. − When applying this to a PSC stack consisting of glass/TCO/PSC/2 nm ALD Al 2 O 3 /50 nm indium tin oxide (ITO), we did not have success because the epoxy reacted with the PSC absorber material, most likely due to prolonged contact before curing. By using a much thicker (∼200 nm) capping TCO layer, as well as a thicker (40 nm) ALD alumina layer, the epoxy was able to cure to the metal and perovskite overnight in the glovebox without apparent degradation; it successfully cleaved when immersed in liquid nitrogen.…”

“…Alumina were deposited by atomic layer deposition (ALD) by trimethylaluminum and water at 85 °C on some films under identical conditions as our aluminum-doped zinc oxide process previously reported . The thermomechanical cleaving procedure used was similar to those used at NREL recently for other PV applications and have been discussed in several publications. − The cleaving procedure as applied to PSC materials will be discussed in further detail here. TOF-SIMS measurements using a gas-cluster ion source for profiling were performed on the ION-TOF TOF-SIMS V instrument at the Colorado School of Mines (CSM) in this work.…”

Revealing the Importance of Front Interface Quality in Highly Doped CdSexTe1–x Solar Cells

“…While the increase in efficiency via addition of the ARC is large, it is consistent with corrections made to a significantly more reflective air/ZnO:Al interface (the refractive index of ZnO is ∼2) compared to the standard air/glass interface in the superstrate configuration (the refractive index of glass is ∼1.5); this is applicable to both Cu- and As-doped devices. 15.1% is the highest efficiency for substrate CdTe reported in the literature to date (largely owing to the low-temperature absorber growth that is typically required in the substrate configuration); the device can also readily be made lightweight and flexible, as described in earlier work . Applying the same technique to As-doped devices, however, resulted in substantial reductions in all performance parameters and <8% device efficiency (a > 50% decrease from the original efficiency).…”

“…15.1% is the highest efficiency for substrate CdTe reported in the literature to date (largely owing to the low-temperature absorber growth that is typically required in the substrate configuration); 27 the device can also readily be made lightweight and flexible, as described in earlier work. 28 Applying the same technique to As-doped devices, however, resulted in substantial reductions in all performance parameters and <8% device efficiency (a > 50% decrease from the original efficiency). The remainder of this work explores a few key differences between reconstructed Cu-and As-doped devices that may account for this discrepancy.…”