Abstract

Tunnel layer passivated contacts have been successfully demonstrated for next-generation silicon solar cell concepts, achieving improved device performance stemming from the significantly reduced contact recombination of the solar cell contacts. However, these carrier-selective passivated contacts are currently deployed only at the rear side of the silicon solar cell, while the front side adopts a conventional diffused junction and contacting scheme. In this work, we report on the novelty and feasibility of deploying tunnel layer passivated contacts on both sides of a silicon wafer-based solar cell, featuring a textured front surface and a planar rear surface. In particular, we demonstrate that silicon solar cells incorporating our in-house developed electron-selective thermal-SiOx/poly-Si(n<sup>+</sup> ) and hole-selective thermal-SiOx/poly-Si(p<sup>+</sup> ) passivated contacts have a solar cell efficiency potential approaching 24%. Deploying contact passivation only at the rear side of the solar cell, we have reached a solar cell efficiency of 21.7%, using conventional screen printing for metallization. We present a systematic approach of evaluating our in-house developed electron-selective and hole-selective passivated contacts on both textured and planar lifetime test structures, as well as dark I–V test structures, to extract the recombination current density j<sup>0</sup> and the contact resistance R<sup>c</sup> of the passivated contact, which is used for process optimization as well as for subsequent efficiency potential prediction. The two key challenges aiming at a double-sided integration of passivated contacts are (1) parasitic absorption within the front-side highly doped poly-Si capping layer, requiring the processing of ultrathin (≤10-nm) contact passivation layers. This has been quantified by numerical simulation (using SunSolve™) and also solved experimentally, i.e., processing ultrathin 3-/10-nm hole/electron extracting SiOx/poly-Si(p<sup>+</sup> /n<sup>+</sup> ) passivated contact layers, reaching an implied open-circuit voltage of 690/703 mV on a planar/textured silicon surface, which will even further enhance after SiNx capping. (2) Ensuring process compatibility with conventional screen printing: Screen printing on electron extracting poly-Si(n<sup>+</sup> ) seems feasible; however, screen printing on holeextracting poly-Si(p<sup>+</sup> ) is still a challenge. Solar cell precursors have been processed, showing excellent pre-metallization results (implied-VOC 713 mV). According to

our efficiency potential prediction (using the measured j<sup>0</sup> and R<sup>c</sup> values of our developed contact passivation layers), bifacial double-sided passivated contact solar cells (efficiency potential of 23.2%, using our layers) can clearly outperform rearside-only passivated contact solar cells (efficiency potential of 22.5%).

being contacted in an all-back-contact configuration) can be attributed to the highly effective and simplified full-area rear-side passivating contact scheme, which inserts an electron-selective tunnel layer passivated rear-side contact between the wafer and the full-area rear-side contact of the solar cell, comprising a wetchemically formed silicon oxide tunnel layer (wet-SiOx) and a highly n-doped polysilicon capping layer. This achieves both excellent interface passivation toward the silicon wafer and a highly selective collection of excess electron charge carriers.

Double-Sided Passivated Contacts for Solar Cell Applications: An Industrially Viable Approach…

silicon wafer, adopting a conventional front-side selective emitter, photolithography processes, and evaporated contacts, it has set the stage for immense research interests such as those reported in Refs. [12–26]. Contact passivation presents a clear advantage over the popular passivated emitter rear contact (PERC) solar cell concept by UNSW [27], which is currently a large scale adopted by the industry (as of Jan. 2019), as an even higher solar cell efficiency can be reached (i.e., by directly passivating the metal solar cell contacts instead of "only" reducing the

An ideal tunnel layer, suited for contact passivation, (i) exhibits a tunneling relevant thickness (i.e., <2 nm) [14], (ii) exhibits excellent interface passivation toward the crystalline silicon wafer [28, 29], and (iii) contributes only minimally to the total contact resistance of the solar cell (in the order of maximal 1 Ω cm<sup>2</sup>

Furthermore, an ideal capping layer, suited for contact passivation, should be either (i) highly doped or (ii) exhibit a high/low work function [31] in order to ensure

The already proven success on electron-selective passivated contacts is also generating huge interest and research activities on hole-selective passivated contacts now. Pertaining to the feasibility studies of different tunnel layer candidates for hole-extracting passivated contacts, most previous reports had focused on using silicon-based oxides formed via either wet-chemical approaches (wet-SiOx) or UV/ozone photo-oxidation (ozone-SiOx) approaches. In our published works [28, 29, 32–34], a comprehensive evaluation of passivation quality and interface properties of silicon-based oxides (SiOx) and atomic layer-deposited aluminum oxides (ALD-AlOx) had revealed a larger potential for ALD-AlOx to be integrated in hole-selective passivated contacts as compared to the commonly used wet-SiOx or ozone-SiOx. This stems from a significantly higher negative fixed interface charge

relevant thickness (just a few ALD cycles) while maintaining a relatively low

Dit of SiOx-based tunnel layers. The high negative fixed interface charges of the ALD-AlOx tunnel layer will accumulate holes at the c-Si interface, which will simultaneously enhance hole extraction probability and reduce surface recombination rates due to an efficient field-effect passivation in addition to the chemical passivation at the interface, as evident from the higher measured effective carrier lifetime (two orders of magnitude higher) than the passivation by either wet-SiOx or ozone-SiOx alone on symmetrically tunnel layer passivated n-type Cz wafers in our previous work [28]. These findings were consistent with literature for much thicker AlOx layers [35–39]. For hole-extracting capping layer materials, various candidates had been suggested, which includes highly p-doped polysilicon, transition metal oxide films with high work function such as molybdenum oxide (MoOx) [40–45], tungsten oxide (WOx), vanadium oxide (V2O5), cuprous oxide (Cu2O) [46], or alternatively organic polymers, such as poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS) [47–49], among others. It is worthy to highlight that the transition metal oxide films exhibit a tunable work function between 4.7 and 7 eV [50, 51] by an appropriate combination of materials, while organic

) float-zone n-type

) even at a tunneling

, which is comparable to the

) [30].

Although this work was established on a small-sized (4 cm2

metal contact area fraction).

91

selective excess charge carrier extraction.

DOI: http://dx.doi.org/10.5772/intechopen.85039

density (1 order of magnitude higher at <sup>6</sup> 1012 cm<sup>2</sup>

interface defect density (Dit) of <sup>2</sup> <sup>10</sup><sup>12</sup> cm<sup>2</sup> eV<sup>1</sup>

Keywords: passivated contacts, contact passivation, silicon solar cells, double-sided passivated contacts
