Field testing of a spectrum-splitting transmissive concentrator photovoltaic module

Highlights

• A spectrum-splitting photovoltaic module is prototyped for use in solar cogeneration.

• Module photovoltaic regions show 34.7% in-band electrical conversion efficiency.

• Module photovoltaic regions transmit 58.8% of infrared light to a thermal receiver.

• Module is field tested at 160 suns, showing 75% of incident solar power is collected.

• Design enables low temperature PV operation with high temperature process heat capture.

Abstract

Hybrid photovoltaic-thermal systems can decouple IR light from visible light, allowing it to be collected separately by spectrum-optimized mechanisms for increased total efficiency. To demonstrate this, we have designed and prototyped a transmissive spectrum-splitting concentrator photovoltaic module that maximizes solar energy conversion by utilizing the entire solar spectrum. The system first collects visible light using IR-transmissive triple-junction photovoltaic cells to achieve an in-band module efficiency of ${\eta }_m^{IB}$  = 34.7% for light of wavelengths λ < 870 nm. Simultaneously, 58.8% of light with λ > 870 nm is transmitted through the cells for collection by a thermal receiver. By combining electrical and thermal power collection, 75% of incident solar power is collected, far surpassing the collection capability of only photovoltaics. The module was tested on a dual-axis tracked parabolic concentrator dish at up to 160 suns for 60 cumulative on-sun hours while maintaining photovoltaic cell temperatures at an average of 50 °C via active cooling. The system performed as expected based on modeled values, and represents a cost-effective path forward for dual-generation of electricity and high-temperature heat with increased total efficiency. The capability is valuable in a wide range of commercial and industrial cogeneration applications.

Introduction

To fill the growing demand for more secure, reliable, and cost effective thermal and electric power, public and business interests are shifting towards renewable energy systems [1]. This energy demand includes both electrical power [2] and process heat [3]. As ∼33% of solar power incident on photovoltaics (PV) is in the form of infrared (IR) light and wasted as a heat byproduct, there has been a movement over the past 15 years towards hybrid photovoltaic-thermal (PV/T) systems that output both electricity and heat while more efficiently converting the entire solar spectrum [4,5]. Significant research has been done on topping mode designs in which PV components absorb all wavelengths and waste heat is captured through a cell cooling system [[6], [7], [8], [9]]. However, this architecture generally cannot yield heat outputs at temperatures above 80 °C, and therefore cannot satisfy the heat requirements for most commercial and industrial applications [3]. Furthermore, topping PV/T suffers from reduced PV cell efficiency due to elevated cell temperatures. An alternative is to use spectrum-splitting PV cells that selectively absorb ultraviolet and visible (UV–vis) light for energy conversion, while either reflecting or transmitting IR wavelengths [[10], [11], [12], [13], [14], [15], [16]]. Transmissive photovoltaics have been developed in the past for applications such as building and window-integrated PV, with 1%–13% [17]; However, by utilizing concentrated sunlight and high-efficiency III–V based photovoltaics and focusing transmitted light into a dedicated thermal receiver, the total efficiency of the system can be drastically improved [18]. This configuration offers several advantages, such as the thermal decoupling of the PV and thermal systems, allowing the PV components to operate at low temperatures and thermal components at high temperatures [19]. Parabolic troughs can be used to concentrate sunlight onto such a hybrid transmissive receiver, but troughs generally achieve lower concentration than two-axis parabolic mirrors, which can achieve concentration in excess of 1000 suns [20,21]. Such concentrations are critical for high-efficiency concentrator photovoltaic (CPV) designs in which the cells operate at >500 suns with thermal system stagnation temperatures of up to 450 °C [22]. Operation at high concentrations allows for both increased PV cell cost effectiveness ($/W) and higher process heat temperatures. Furthermore, high temperature output heat enables use in industrial processes, the largest consumers of heat [23]. The shared PV and thermal collector infrastructure in concentrator PV/T (CPV/T) designs also enable cost reductions, facilitating commercialization in previously unavailable markets [24]. In light of this, we here present a transmissive CPV (tCPV) module designed for high solar concentrations with applications in industrial process heating and electricity production. For the first time, we demonstrate a working two-axis transmissive CPV solar collection system and detail its performance on-sun.

The system herein, similar to PV/T systems, seeks to optimize performance within the trade-space of electrical performance, thermal performance, and PV thermal management. Most importantly, unlike topping modes, specular transmissivity must be achieved in the tCPV/T module. Previous reports have demonstrated transmissive GaAs-substrate III–V triple junction PV cells that absorb light of wavelength λ < 870 nm for electricity conversion, while transmitting IR light for thermal collection [25]. These bare cells achieved up to 24.9% full-spectrum efficiency under 1-sun, and 29.5% efficiency at 500x concentration [25]. We here demonstrate a module that encapsulates, protects, cools, and electrically connects these cells for outdoor use on a concentrator dish and 2-axis tracker. Along with the cells, all module subsystems are designed to be transparent, including substrates, superstrates, encapsulant, and active cooling systems [26]. Non-transparent module components, such as copper wire electrodes, were utilized sparingly to reduce shadowing while maintaining acceptable series resistance. The work described here constitutes the experimental results of our fifth and sixth PV module prototype iterations, which are referred to as “Module 5” and “Module 6”.

We introduce three metrics for describing the PV efficiency of the tCPV module, as shown in Table 1; the cell efficiency, ${\eta }_c$, describes the power conversion efficiency (PCE) of a bare PV cell in air. The cell area is defined as 0.29 cm² and excludes the regions covered by busbar. Module efficiency, ${\eta }_m$, describes the collective PCE of all working cells in the module, and accounts only for sunlight incident on the cell area, given that light spilled outside of the cell area is directed to the thermal receiver by design. Receiver efficiency, ${\eta }_r$, describes the total PCE of the module aperture and is dependent on the number of cells in the module and their spacing. Because the cells are only designed to convert light of wavelength λ < 870 nm, we define solar spectrum light of wavelength λ < 870 nm as “in-band” light, and similarly define solar spectrum light of wavelength λ > 870 nm as “out-band” light. Therefore, each efficiency metric can be further specified as full-spectrum (AM1.5D) or in-band (λ < 870 nm). All electrical output power numbers represent two-terminal, max-power-point measurements. One sun measurements are taken at 25 °C and use an AM1.5D calibrated multi-zone solar simulator to define input power. High flux measurements use real outdoor testbed conditions for temperature and input flux, as described further in the text. The key figure of merit for our tCPV system is in-band module efficiency, ${\eta }_m^{IB}$, which most directly speaks to the electrical quality of the module in the context of a tCPV system. We also highlight the out-band transmission (λ > 870 nm) through the cell area as a key metric.