Integration of power electronics into crystalline silicon solar cells

Abstract

Power electronics (PE) plays a crucial role in optimizing the performance of photovoltaic (PV) systems. At present, module and sub-module level PE is increasingly being adopted in global PV installations. Typically, PE for sub-module purposes is installed in the PV junction box. However, in this dissertation another approach is investigated: integrating power electronics into or onto crystalline silicon (c-Si) PV cells. This approach has the potential to contribute to the development of shade-tolerant PV modules, increase the reliability of PV module-integrated PE, reduce the volume and weight of the PV system, and support the development of autonomous devices powered by low-power and low-current PV cells. Specifically, the aim of this dissertation is to study the integration of different PE components into c-Si solar cells, and identify the most promising approaches. The dissertation is structured in two parts.

In Part 1, the objective is to gain a comprehensive understanding of the impedance of c-Si solar cells. To this end, Chapter 2 reports a characterization of the impedance of eight single-cell laminates. Each laminate contains a different commercially available c-Si solar cell, featuring various cell architectures, namely Aluminium Back Surface Field (Al-BSF), Passivated Emitter and Rear Contact (PERC), Tunnel Oxide Passivated Contact (TOPCon), Interdigitated Back Contact (IBC), and Silicon Heterojunction (SHJ). It is found that the two main factors contributing to a high PN junction capacitance (C_j) at maximum power point (MPP) are (1) a low wafer dopant concentration and (2) a high MPP voltage. Furthermore, the studied cell laminates exhibit inductances between 63 and 130 nH. Following this, the impedance of PN junctions is further investigated in Chapter 3. Specifically, PN homojunction devices are investigated through Technology Computer-Aided Design (TCAD) simulations. This methodology allows to study the junction impedance in a detailed way, which may be difficult to do experimentally due to noise and reactance of metal contacts. Through analysis of the impedance data it is revealed that the PN junction exhibits the behaviour of a parallel resistor-capacitor circuit (RC-loop) at low frequencies, but undergoes relaxation in both PN junction resistance R_j and capacitance C_j as frequency increases. While various publications on solar-cell impedance model the low-high (LH) junction using an RC-loop, the findings presented in this chapter indicate that such a model does not accurately represent the underlying physics. Instead, this approach is likely compensating for the frequency-dependent behavior of R_j and C_j. Finally, in Chapter 4, the PN junction impedance of modern c-Si solar cells is studied across varying temperature and illumination conditions. In the tested conditions, the range in which the area-specific MPP R_j varies is similar for different cell architectures, despite their different properties. Conversely, the range in which the areal MPP C_j varies is significantly affected by the substrate dopant concentration and MPP voltage of the cell.

In Part 2, the aim is to assess the feasibility of leveraging solar-cell impedance at the input of a power converter, and to explore various methods for integrating additional PE components into solar cells. In Chapter 5, the feasibility of integrating different PE components into c-Si solar cells is explored. First of all, diodes exhibit high ease of integration into PV cells and successfully integrated designs have already been demonstrated in prior work. Alternatively, the integration of transistors is more complex. Since transistor fabrication processes require lithographic steps, it is necessary for cost-effective integration to combine as many processing steps as possible with PV fabrication. Regarding passive component integration, it is found that the self-capacitance of modern c-Si solar cells is sufficiently large to replace the input capacitor of an exemplary boost converter. However, for thin-film capacitor integration, it is challenging to achieve a sufficiently high areal capacitance. Moreover, the self-inductance of a solar-cell string could potentially be leveraged to replace the inductor at the input of a power converter. By analyzing this approach for an exemplary boost converter, it was found that high switching frequencies in the MHz range are required. Alternatively, the required switching frequency may be reduced through the integration of planar inductors. It was found that the area of PV cells is sufficiently large to facilitate the integration of planar coils exhibiting inductance values that are useful for power conversion. Finally, general challenges that should be considered for successful PE-PV integration are appropriate thermal management, opto-electric behaviour under illumination, and repairability. The inductor integration is further studied in Chapter 6. Specifically, it is explored whether large-area planar air-core inductors can yield the required inductor properties to support sub-module power conversion in PV modules. First, it is shown how the interplay between the different design parameters, such as track spacing, track width, number of turns, and middle gap size, play an important role in the inductor properties. This analysis includes changes due to high-frequency effects, which significantly impact the results. The coil geometries that are simulated yield inductance values between 0.3 and 3.2 μH. Considering the power losses, the applicability of such inductors in sub-module power converters is discussed. Finally, in Chapter 7, the concept of COSMOS (COmbined Solar cell and MOSFET) devices is introduced and a process flow is proposed in which back-contact TOPCon solar cells and lateral power MOSFETs are simultaneously fabricated on a single substrate. This process is successfully employed to manufacture both n-type solar cells with integrated p-channel MOSFETs (PMOS) and p-type solar cells with integrated n-channel MOSFETs (NMOS). Notably, efficiencies exceeding 20% are achieved for both n-type and p-type solar cells, highlighting the potential of COSMOS solar cells. Furthermore, two main integration challenges are identified. Firstly, the off-state leakage currents of the MOSFETs increase due to illumination. Secondly, specific topologies of monolithic integration lead to increased off-state leakage currents.