Passivating Contacts for Silicon Solar Cells: A Zinc Oxide Breakthrough?

In the field of silicon solar cells, the search for more efficient and cost-effective technologies never ceases. One of the most exciting developments of the past decade has been the rise of so-called passivating contacts, with the current industry leader being TOPCon (Tunnel Oxide Passivated Contact) cells. I remember I was a PhD student when I visited the SiliconPV conference in 2014, where all of a sudden this totally new TOPCon approach was presented with an (at the time) highly impressive efficiency of 23%[1]. By now, TOPCon seems to have overtaken the previous PERC cell (passivated emitter rear contact) in industry, and risen to 26.7% efficiency in lab-scale and 25.9% efficiency in mass-production by Trina solar.

In this blog post, I want to highlight that maybe (and I stress maybe!) a follow-up or alternative to TOPCon is on the horizon, i.e. passivating contacts based on zinc oxide (ZnO) prepared by atomic layer deposition (ALD). This concept, pioneered by our research group, received a major boost when recently Gao et al., alongside LONGi, published a 24.3% efficient solar cell, setting a new record for metal oxide passivating contacts.[2] This breakthrough signals that ZnO passivating contacts—what they nicely coined “ZOPCon”, or Zinc Oxide Passivating Contact —could be the next big innovation in solar cell manufacturing. In this blog, I will introduce this novel “ZOPCon” technology, compare it to the currently dominant TOPCon structure, and provide an overview of the key findings from our research that laid the groundwork for this innovation. I’ll also briefly discuss the recent work by LONGi and offer my thoughts on what this might mean for the future.

For those of you not fully familiar with “passivating contact” technologies and how they differ from traditional solar cells, I refer to two review articles from our group (Melskens et al. [3], and more recently Michel et al. [4]), as well as from others (Allen et al.[5], Richter et al.[6]). Also, we have three earlier blogs on AtomicLimits on passivating contacts, describing the presentations at the passivating contacts workshop we hosted in Eindhoven in 2018 (blogs one, two, three).

PERC vs. TOPCon vs. ZOPCon: A Structural Comparison

In order to appreciate the potential benefits of ZOPCon, it is instructive to first explain why TOPCon became the superior structure over conventional PERC cells. To aid this discussion, I have placed these three solar cell structures below. For this blog, we only need to consider the rear side contact of the solar cells.

PERC: In a PERC cell, most of the rear side is covered by Al₂O₃ and SiN_x dielectrics, which act as surface passivation layers to minimize recombination of electrons and holes (See also our recent blog on surface passivation [7]). At the Al₂O₃-Si interface, a SiO₂ interlayer forms automatically during deposition (see inset TEM image), which helps in achieving a low-defect interface. Negative fixed charges at the Si-Al₂O₃ interface help repel electrons from the silicon surface, resulting in excellent passivation especially on p-type silicon. Through the dielectric, local p-type contacts are made, where the Si/Al interface forms a highly p⁺-doped eutectic. While this contact has been thoroughly optimized and only represents a small fraction of the rear side, ultimately, the presence of this Si(p⁺)/Al contact region limits the efficiency by inducing Auger recombination in the highly-doped Si and surface recombination by defects at the Si-metal interface.

TOPCon: The key strength of TOPCon (and passivating contacts in general) is to avoid direct contact between the silicon wafer and the metal contact, since this inevitably leads to recombination losses. Instead, the wafer surface is fully covered by a (stack of) passivating layer(s), which selectively lets only electrons (n-type) or holes (p-type) pass to the metal contact. The latter functionality can be achieved through careful engineering of the band diagram, typically relying on doping, inducing band bending in the wafer and/or band offsets.

In the specific case of TOPCon, a passivating contact at the rear side is achieved by intentionally growing an ultrathin (1-2 nm) SiO₂ layer, which is a thin enough dielectric to let charge carriers tunnel through (see also inset TEM image in the figure). Selective conductivity for electrons (n-type contact) is achieved by heavy (degenerate) n-type doping of the polycrystalline Si. Importantly, in order to reach really excellent surface passivation, the SiO₂ has to be hydrogenated, which is typically achieved by thermal annealing, during which hydrogen from the SiN_x or Al₂O₃ dielectric diffuses to the SiO₂ and binds to any remaining defects.[8]

ZOPCon: Let’s now compare the well-established TOPCon structure to the new ZOPCon structure, which replaces the poly-Si contact in TOPCon with a ZnO passivating contact. While the polycrystalline silicon layer in TOPCon solar cells yields excellent surface passivation and a low-resistance contact, our research has shown that a ZnO layer can offer similar benefits, with potential advantages in terms of processing, scalability, and transparency.

Before we delve into these findings, you may have noticed one thing in the image above: The recent ZOPCon cell has a planar rear side with full area metallization, as opposed to the other cells shown. Ideally, you want also the rear side to be pyramid-textured to enhance light-trapping in the solar cell, and having local rear metallization means that you can also collect sunlight from the back (so-called bifaciality). Initial PERC and TOPCon cells also did not have this, but the recent industrial editions, often coined PERC⁺ and i-TOPCon (i stands for industrial, by Trina Solar[9]), have evolved to have these features. Later in this blogpost, I’ll reflect on how we might achieve this for ZOPCon as well.

Key Findings from Our Research

Our work was instrumental in showing that ZnO could serve as a highly effective passivating contact for silicon. Figure 2 shows the main pathfinding we had to do, and below we summarize our key findings and insights:

1. Surface Passivation by ALD ZnO (van de Loo et al., DOI: 10.1063/1.5054166)[10]:
   Initially, we faced challenges in passivating silicon with ZnO, but two breakthroughs helped us succeed. First, adding an interfacial SiO₂ layer was vital for achieving good passivation. In this sense, ALD ZnO is very different from ALD Al₂O₃, for which you get a good interfacial SiO₂ “for free”. Second, we found that hydrogen present within the ALD ZnO – crucial for passivation – could easily effuse from this polycrystalline material during thermal annealing. By capping the ZnO with Al₂O₃, we could retain hydrogen in the structure during annealing, which was critical for improving passivation. Finally, we also found that doping the ZnO enhanced field-effect passivation.
2. ZnO as an Electron Contact (Macco et al., DOI: 10.1016/j.solmat.2021.111386)[11]:
   In this study, we demonstrated that ZnO could not only act as a passivating layer, but also as an effective electron contact. The main hurdle was that the Al₂O₃ capping layer, that was required to prevent hydrogen effusion during annealing, is a dielectric, preventing electrically contacting the ZnO by a metal. Fortunately, we found a pH-based alkaline etch by which we could selectively remove the Al₂O₃ cap after annealing without impairing the passivation. This crucial step allowed us to finally measure directly the contact resistivity between the ZnO:Al layer and the silicon wafer (over the SiO₂ tunnel oxide). We found that the key to achieving low contact resistivity was strong doping of the ZnO and silicon, a crucial insight for understanding future developments. Finally, we showed another great benefit that arises automatically from the Al₂O₃ capping layer and annealing approach: The optoelectronic quality of the ZnO:Al also strongly improves, specifically increasing the carrier mobility and thereby electrical conductivity, and improving transparency in the infrared through reduced Drude (free-carrier) absorption. Likely, hydrogen also plays a key role in bulk passivation of the ZnO:Al (grain boundaries). Through our approach, optical transparency and electrical conductivity close to levels typically only achieved with indium-based transparent conductive oxides (TCOs) were obtained.
3. Industrial Feasibility of ZnO Contacts (Macco et al., DOI: 10.1016/j.solmat.2022.111869)[12]:
   We explored how to make these layers more practical for industrial use, reducing the thickness of ZnO and Al₂O₃ layers to 5 nm and 8 nm, respectively. We also demonstrated the viability of spatial ALD for large-scale applications on a tool of the company SALD and showed that excellent passivation can be achieved on pyramid-textured silicon surfaces. Furthermore, we showed selective growth on SiO₂ surfaces, opening the door to self-aligned processing.

The Recent Breakthrough: LONGi’s 24.3% Efficient ZOPCon Cell

A few months ago, we were thrilled to see Cai et al. publish a paper on a 23.2% efficient solar cell contacts, exactly based on our approach of adding a tunnel SiO₂ layer, capping the ALD ZnO:Al with Al₂O₃ and annealing, followed by selective Al₂O₃ removal.[13] Our excitement grew, as this was quickly followed by LONGi’s announcement of a 24.3% efficient ZnO-based solar cell. What surprised us most was their ability to achieve low contact resistivity on lowly-doped n-type Si (1 Ωcm). In our earlier research, we found that highly doped n-Si was necessary for a good contact, but these new studies introduced a clever trick: the addition of a low work function layer, i.e. ~1 nm of LiF. This low work function induces more downward band bending at the silicon surface, improving the contact quality without the need for excessive doping of the silicon.

The success of this approach lies in keeping the ZnO layer thin (around 5 nm), so the work function effect from the LiF isn’t shielded by free carriers in the ZnO. This is ideal from an industrial perspective, as thinner layers are easier to integrate into large-scale production.

The Potential of ZnO Passivating Contacts

So, what does this mean for the future of solar cell technology? Of course, it’s still really early to tell, but some signs are promising. While TOPCon started at an efficiency of 22.3% and progressed through its learning curve up to 26.7% now, the ZOPCon cell is at a much higher efficiency of 24.3% at the start of its learning curve. Hence, ZOPCon cells could potentially surpass TOPCon cells through iterative improvements. Regarding this, it is important to note that for current-day solar cells the last few percent of efficiency are “squeezed out” by processing millions of wafers, storing all their process settings (and variations) and cell parameters, and use machine learning to further optimize the production process – often with limited connection to physical understanding. So in this sense, we might not be surprised if ZOPCon’s learning curve could match or outperform that of TOPCon.

Besides efficiency, ALD ZnO also offers several potential industrial advantages over doped poly-Si contacts. This includes lower-temperature processing and the elimination of toxic gases (phosphine) used to dope the poly-Si during deposition. The ZnO:Al layer can also be only 5 nm thick, compared to the tens or hundreds of nanometers of poly-Si typically used. And importantly, the aforementioned work of Cai et al., shows that ALD ZnO:Al should be highly compatible with industrial processing, especially when only 5 nm thin. This is also not surprising, as ALD aficionados will know that ALD ZnO is a very ideal ALD process, much like ALD Al₂O₃ which is being used to coat billions of solar cell wafers annually. 

Of course, on the other hand, one can still raise questions on the industrial feasibility of placing evaporated LiF on cells, the type of metallization used, … and many more questions. Will ZnO passivating contacts be scalable to the gigawatt scale? Will these contacts be stable? Can the manufacturing process be streamlined further? Will the efficiency be on par with poly-Si based contacts? While it’s hard to predict, it’s still exciting to think that research from our group could play a pivotal role in this evolution.

I wouldn’t bet on it just yet—but I wouldn’t be surprised either!

Outlook: future directions for ZnO-based passivating contacts

Next, I would like to share some thoughts on future directions for ZnO-based passivating contacts in silicon solar cells. Let’s first look back at the current ZOPCon design. As mentioned earlier, the rear side is currently planar, and due to the full-area metallization at the rear, no sunlight can enter from there (no bifaciality). Moving to a pyramid-textured rear side would probably be relatively easy, as our research showed good passivation also on such texture. Moving to a bifacial cell with local rear metal contacts might be trickier though. The figure below shows what such a future bifacial ZOPCon cell might look like.

If you want to make local metal rear contacts, you need to ensure that the n-type contact at the rear is capable of laterally conducting electrons very well (i.e., it should have a low sheet resistance). In principle, ZnO:Al is perfectly suited for this! It is a transparent conductive oxide, like ITO, and for thicker ZnO:Al layers (~80 nm) you get good lateral conduction and as an added bonus a proper antireflection coating. Yet, herein lies the crux: The ZOPCon cell uses low-work function LiF to enhance the contact, but if you make the ZnO:Al thicker than 5 nm (or: thicker than the Debye length therein), the silicon will not “feel” the low work function of the LiF anymore. If we would not have to rely on LiF, that would solve the problem. One way to do this would be to reduce the work function of the ZnO:Al itself. This can be achieved by stronger degenerate doping, as this raises the Fermi level further into the conduction band. Interestingly, the two papers with ZnO:Al/LiF solar cell contacts used trimethylaluminum (TMA, Al(CH₃)₃) for doping the ZnO with Al during ALD. TMA is actually not a good dopant for ZnO for many reasons, and we have shown that other dopant precursors such as dimethylaluminumisopropoxide (DMAI) can yield more effective doping and thereby higher degenerate carrier densities (and hence lower work functions).[14] Perhaps this could help alleviate the reliance of LiF.

Beyond ZOPCon, the ZnO:Al contact might actually also prove useful in other cell designs, which I also presented at the SiliconPV 2022 conference and have summarized in the figure above. A potential device application is to utilize the ZnO-based stack as a full-area passivating and contacting layer on the front side of a PERC-type solar cell, replacing the SiN_x layer. The ZnO’s full-area passivation would make the shielding of minority carriers by the n⁺-diffused region less critical, while also providing an additional lateral current pathway for electron extraction. These effects could potentially lower the required n⁺ doping level at the wafer front surface and, together with the absence of direct Si-metal contact, enable a higher V_oc.

Integrating such a front TCO on a PERC solar cell could also be a key step toward enabling its use as a bottom cell in silicon-perovskite tandem cells. As an n⁺ layer, the front TCO can readily form a recombination junction with the p-layer from the top perovskite solar cell. So-called silicon heterojunction cells based on passivating and doped silicon layers (not discussed in this blog) naturally have TCOs on both the front and rear, and because of this – together with their high efficiency – silicon heterojunction bottom cells have been at the forefront of silicon-perovskite tandem efficiency records at the research level. Yet, from an industry perspective, using mass-produced PERC bottom cells would be much more appealing. The recent announcement of Qcells in the European PEPPERONI project (where TU/e is also part of) has shown great promise in this respect, showcasing a 28.6% tandem efficiency using their full size industrial Q.antum product as a bottom cell.

While I have always found the ZnO:Al contact a possible fascinating innovation for PERC bottom cells, one aspect that always made me a bit skeptical is the fact that we found that high doping is needed to make a good contact, as mentioned before. This high doping will increase free-carrier absorption in the infrared. Yet, in tandem cells in particular, the infrared transparency of all layers above the silicon bottom cell should be high, since this is the main stream of photons towards the bottom cell. But perhaps also here the introduction of LiF can help alleviate the need for high doping. Remains to be seen!

As a final structure, I would like to mention a more straightforward possible use of the ZnO:Al/Al₂O₃ stack. While I did not mention it before, in one of my listed papers [11] we showed that this stack can also strongly enhance poly-Si contacts like those found in TOPCon. Here, adding a ZnO:Al/Al₂O₃ stack on the poly-Si contact and annealing serves a double purpose: enhancing the ZnO:Al optoelectronic properties and hydrogenating the SiO₂ interlayer, resulting in better passivation.[15] As the structure above shows, the ZnO:Al would then be positioned at the rear, serving as additional lateral conductivity channel (perhaps reducing the required poly-Si thickness) and as antireflection coating, both of which would enhance bifaciality.

A personal note to wrap up this blog. I truly find this ZnO:Al contact extremely fascinating, exactly because it brings together so many aspects in physics, materials and applications. It contains surface passivation, tunneling contacts, degenerate doping and Drude absorption, hydrogen migration, ALD chemistry (for efficient doping), and much more … The fact that we can spend almost a whole blog on possible applications of this contact also attests to its versatility. And even while there are many steps to be taken before we might see the ZnO:Al contact in mass production, I really feel like there are many opportunities to be explored and questions to be answered!

Acknowledgements

This work would not have been possible without the contributions of many talented individuals. A special thanks to Bas, who was the first to discover the excellent passivation properties of ZnO, laying the foundation for this research. I am also grateful to Marc, Dennis, and Bart, the master’s students I had the pleasure of guiding, whose dedicated work made the electrical contact possible. This was truly a team effort, and I am proud of what we have accomplished together.

I also acknowledge the financial support that enabled this research. In particular, my NWO VENI grant, which provided the opportunity to explore and develop ZnO passivating contacts, and various TKI programs, which further supported this work.

References

[1]       F. Feldmann, M. Bivour, C. Reichel, M. Hermle, S.W. Glunz, Passivated rear contacts for high-efficiency n-type Si solar cells providing high interface passivation quality and excellent transport characteristics, Sol. Energy Mater. Sol. Cells. 120 (2014) 270–274. https://doi.org/10.1016/j.solmat.2013.09.017.

[2]       K. Gao, D. Xu, J. Wang, Q. Bi, Z. Wu, H. Lin, S. Wang, W. Shi, C. Yu, F. Cao, Y. Diao, J. Xie, X. Wang, K. Li, X. Lou, W. Li, C. Xing, Y. Wang, T. Yan, D. Zhang, S. de Wolf, X. Zhang, X. Yang, Efficient Silicon Solar Cells with Aluminum-Doped Zinc Oxide-Based Passivating Contact, Adv. Funct. Mater. 2415039 (2024) 1–10. https://doi.org/10.1002/adfm.202415039.

[3]       J. Melskens, B.W.H. van de Loo, B. Macco, L.E. Black, S. Smit, W.M.M. Kessels, Passivating Contacts for Crystalline Silicon Solar Cells: From Concepts and Materials to Prospects, IEEE J. Photovoltaics. 8 (2018) 373–388. https://doi.org/10.1109/JPHOTOV.2018.2797106.

[4]       J. Ibarra Michel, J. Dréon, M. Boccard, J. Bullock, B. Macco, Carrier‐selective contacts using metal compounds for crystalline silicon solar cells, Prog. Photovoltaics Res. Appl. 31 (2023) 380–413. https://doi.org/10.1002/pip.3552.

[5]       T.G. Allen, J. Bullock, X. Yang, A. Javey, S. De Wolf, Passivating contacts for crystalline silicon solar cells, Nat. Energy. (2019) 2–3. https://doi.org/10.1038/s41560-019-0463-6.

[6]       A. Richter, R. Müller, J. Benick, F. Feldmann, B. Steinhauser, C. Reichel, A. Fell, M. Bivour, M. Hermle, S.W. Glunz, Design rules for high-efficiency both-sides-contacted silicon solar cells with balanced charge carrier transport and recombination losses, Nat. Energy. 6 (2021) 429–438. https://doi.org/10.1038/s41560-021-00805-w.

[7]       R.J. Theeuwes, W.M.M. Kessels, B. Macco, Surface passivation approaches for silicon, germanium, and III–V semiconductors, (2024). https://doi.org/10.1116/6.0004030.

[8]       B.W.H. van de Loo, B. Macco, M. Schnabel, M.K. Stodolny, A.A. Mewe, D.L. Young, W. Nemeth, P. Stradins, W.M.M. Kessels, On the hydrogenation of poly-Si passivating contacts by Al₂O₃ and SiN_x thin films, Sol. Energy Mater. Sol. Cells. 215 (2020) 110592. https://doi.org/10.1016/j.solmat.2020.110592.

[9]       Y. Chen, D. Chen, C. Liu, Z. Wang, Y. Zou, Y. He, Y. Wang, L. Yuan, J. Gong, W. Lin, X. Zhang, Y. Yang, H. Shen, Z. Feng, P.P. Altermatt, P.J. Verlinden, Mass production of industrial tunnel oxide passivated contacts (i‐TOPCon) silicon solar cells with average efficiency over 23% and modules over 345 W, Prog. Photovoltaics Res. Appl. (2019) pip.3180. https://doi.org/10.1002/pip.3180.

[10]     B.W.H.W.H. Van De Loo, B. Macco, J. Melskens, W. Beyer, W.M.M.M.M. Kessels, Silicon surface passivation by transparent conductive zinc oxide, J. Appl. Phys. 125 (2019) 105305. https://doi.org/10.1063/1.5054166.

[11]     B. Macco, B.W.H. van de Loo, M. Dielen, D.G.J.A. Loeffen, B.B. van Pelt, N. Phung, J. Melskens, M.A. Verheijen, W.M.M. Kessels, Atomic-layer-deposited Al-doped zinc oxide as a passivating conductive contacting layer for n⁺-doped surfaces in silicon solar cells, Sol. Energy Mater. Sol. Cells. 233 (2021) 111386. https://doi.org/10.1016/j.solmat.2021.111386.

[12]     B. Macco, M.L. van de Poll, B.W.H. van de Loo, T.M.P. Broekema, S.B. Basuvalingam, C.A.A. van Helvoirt, W.J.H. Berghuis, R.J. Theeuwes, N. Phung, W.M.M. Kessels, Temporal and spatial atomic layer deposition of Al-doped zinc oxide as a passivating conductive contact for silicon solar cells, Sol. Energy Mater. Sol. Cells. 245 (2022) 111869. https://doi.org/10.1016/j.solmat.2022.111869.

[13]     H. Cai, Z. Zhong, Q. Nong, P. Gao, J. He, Design and Optimization of SiO_x/AZO Transparent Passivating Contacts for High‐Efficiency Crystalline Silicon Solar Cells, Adv. Funct. Mater. 2411207 (2024) 1–9. https://doi.org/10.1002/adfm.202411207.

[14]     B. Macco, W.M.M. Kessels, Atomic layer deposition of conductive and semiconductive oxides, Appl. Phys. Rev. 9 (2022) 041313. https://doi.org/10.1063/5.0116732.

[15]     L. Tutsch, F. Feldmann, B. Macco, M. Bivour, E. Kessels, M. Hermle, Improved Passivation of n-Type Poly-Si Based Passivating Contacts by the Application of Hydrogen-Rich Transparent Conductive Oxides, IEEE J. Photovoltaics. 10 (2020) 986–991. https://doi.org/10.1109/JPHOTOV.2020.2992348.