A little while ago I wrote a blog about my PhD project. In this blog, I explained that most of my PhD project revolved around an atomic layer deposition (ALD) process that I developed to make hydrogen-doped indium oxide (In2O3:H): This material combines extreme transparency with good electrical conductivity, making it highly interesting in for example solar cell applications. In that blog, which was a mini-review, I gave an overview on why this material is so interesting, and what the key new insights were that we obtained. Meanwhile, two other papers have appeared from our group on In2O3:H. One on the use of In2O3:H in solar cells, one on area-selective ALD of In2O3:H. There is actually quite a nice story and link between these two seemingly separate papers, which I would like to tell in this blog.
If you have read my previous blog, you probably got the impression that my PhD research went like a breeze. It probably did, and I probably should be the last guy to complain. Nonetheless, not everything went exactly as planned, and I did strike some adversity when trying to apply this In2O3:H in solar cells. As said, the In2O3:H material that I developed was highly interesting for solar cells. Having realized this, I tried to convince my advisor, Erwin Kessels, that we should immediately apply this material in a so-called silicon heterojunction (SHJ) solar cell. I remember him saying something along the lines of “Well Bart, you can try it, but once you go to solar cells things usually turn tough and you encounter unexpected things.” Of course I was stubborn and didn’t believe him. I simply had to deposit a layer onto a solar cell. Surely, nothing could go wrong. Of course, Erwin was right.
The problem I encountered when greedily applying my In2O3:H on a SHJ solar cell for the first time is probably best explained by a photo, see below. As most readers are probably aware of, ALD is known to give highly uniform films. Looking at the photo, you can probably agree with me that something clearly went wrong: The cell was full of “smudges”: some brown patches where In2O3:H had grown, other areas where there was no growth…a big mess.
Photo of a SHJ cell on which it was attempted to grow In2O3:H by ALD. A very non-uniform deposition is observed, where the brown smudges have ~40 nm of In2O3:H. Ideally, there would have been a uniform deposition of 75 nm, giving the solar cell a nice, dark blue appearance.
How to achieve uniform growth?
So what was causing this non-uniformity, and how did we solve it? Well, as we are becoming increasingly aware of, a lot of subtleties in ALD lie in the nucleation process of the film. Also in this case, it was the film nucleation that was playing a major role. Let me try to explain this using the figure below, which schematically shows different nucleation types. In ideal ALD growth, film growth starts directly from the first cycle with the steady-state growth-per-cycle (GPC). However, since ALD inherently relies on surface chemistry, the chemical nature of the starting surface can strongly influence the initial growth. This is also something I observed for my ALD In2O3:H process, which I developed on both SiO2 or Al2O3 substrates. On these substrates, it takes a few tens of cycles before growth starts. In such a case of delayed growth, as depicted in blue in the figure below, the precursor and/or reactant is not very reactive towards the surface. Therefore, it takes more cycles (i.e. more exposure) for growth to start. In addition, growth often preferentially nucleates at surface defects, and so-called island growth is often observed, as has for instance been shown for the case of ZnO.
Examples of different types of ALD growth: Ideal growth, delayed growth, and no growth.
So for the case of ALD In2O3:H on SiO2 and Al2O3 substrates, a growth delay is observed. However, for the SHJ solar cell as shown in the photo, we need to deposit the In2O3:H on a different surface: hydrogenated amorphous silicon (a-Si:H). As it turned out, the InCp precursor molecule that we use in our ALD process does not react with the Si-H terminated surface of a-Si:H. In other words, the film growth does not start, because our indium precursor does not “stick” to such a surface. This is what we call a “non-growth surface”, and corresponds to the case presented in black in the figure above. Therefore, we probably only observed patches of growth in places where there were some defects or some contaminants on the surface.
So the objective was actually quite simple: We had to alter the chemical nature of the a-Si:H surface in order to go from a non-growth surface to delayed growth or preferably ideal ALD growth. At that time, a new post-doc joined our group, Yinghuan Kuang. Together, and later mostly by him, we pursued this goal. Together with the help of some computational power and density functional theory (courtesy Chaitanya Ande and Bora Korasalu) we confirmed that the InCp precursor molecule that we use in our ALD process does not react with the Si-H terminated surface of a-Si:H. So how could we circumvent this problem? What we found is that if we pretreat the surface with a mild plasma, we can modify the surface of the a-Si:H such that the InCp precursor becomes reactive to it! This way, uniform growth could be achieved on the solar cell, and this breakthrough enabled us to finally apply In2O3:H to a SHJ solar cell. As the photo below shows, this enabled us to make beautiful solar cells! Details on the insights behind the plasma surface activation and the solar cell results are described in a recent Solar Energy Materials & Solar Cells paper by Yinghuan. 
How to achieve area-selective ALD of In2O3:H?
Interestingly, our struggle to reduce the growth delay is actually opposite to the goal of most people working on area-selective ALD (AS-ALD), who want to identify non-growth surfaces with ideally an infinite growth delay. If you can structure a surface such that part of it consists of a non-growth surface and another part (where you want to deposit a structure) has a surface which gives ideal ALD growth or has a minor nucleation delay, you can in principle deposit the material only on those surfaces! Such AS-ALD processes are actually enablers for the semiconductor industry, and recently we had a very well-attended workshop on this in Eindhoven. More reading material can be found here.
Looking at our ALD In2O3:H process, we actually have the ideal ingredients for AS-ALD: We have a non-growth surface, which is H-terminated a-Si:H. When this surface is OH-terminated, e.g. by plasma exposure, a growth surface is obtained. This was recognized by another PhD student in our group, Alfredo Mameli, who is working on area-selective ALD.
Photo of a wafer on which the logo of the Eindhoven University of Technology (TU/e) has been grown by AS-ALD of In2O3:H. The pattern was made using a micro-plasma printer.
As described in Alfredo’s recent Chemistry of Materials paper, the approach for AS-ALD of In2O3:H is actually simple, but very effective: By taking an a-Si:H film and locally treating it with a plasma (e.g. using a mask or a micro-plasma printer), a growth surface is locally obtained. This way, Alfredo was able to selectively deposit ~35 nm of In2O3:H on plasma-treated areas without depositing on the non-treated surfaces. Note that such a level of selectivity is HUGE: other methods for achieving AS-ALD such as the use of self-assembled monolayers (SAMs) can typically yield selectivity up to a few nanometers of film growth. Also, Alfredo demonstrated that his approach not only works on the millimeter-scale as defined by his micro-plasma printer, but can also be employed at the nanoscale by using a focused electron beam to locally activate the non-growth surface. Bottom line: We now have an area-selective ALD process of highly conductive and transparent In2O3:H for a wide range of width scales, which can turn out to be enabling in nanopatterning as well as in various large-area electronics applications such as displays and solar cells!
 B. Macco, H.C.M. Knoops, W.M.M. Kessels, Electron Scattering and Doping Mechanisms in Solid-Phase-Crystallized In2O3:H Prepared by Atomic Layer Deposition, ACS Appl. Mater. Interfaces. 7 (2015) 16723–16729. doi:10.1021/acsami.5b04420.
 B. Macco, Y. Wu, D. Vanhemel, W.M.M. Kessels, High mobility In2O3:H transparent conductive oxides prepared by atomic layer deposition and solid phase crystallization, Phys. Status Solidi – Rapid Res. Lett. 8 (2014) 987–990. doi:10.1002/pssr.201409426.
 B. Macco, H.C.M. Knoops, M.A. Verheijen, W. Beyer, M. Creatore, W.M.M. Kessels, Atomic layer deposition of high-mobility hydrogen-doped zinc oxide, Sol. Energy Mater. Sol. Cells. (2017). doi:10.1016/j.solmat.2017.05.040.
 Y. Kuang, B. Macco, B. Karasulu, C.K. Ande, P.C.P. Bronsveld, M.A. Verheijen, Y. Wu, W.M.M. Kessels, R.E.I. Schropp, Towards the implementation of atomic layer deposited In2O3:H in silicon heterojunction solar cells, Sol. Energy Mater. Sol. Cells. 163 (2017) 43–50. doi:10.1016/j.solmat.2017.01.011.
 A. Mameli, Y. Kuang, M. Aghaee, C.K. Ande, B. Karasulu, M. Creatore, A.J.M. Mackus, W. (Erwin) M.. Kessels, F. Roozeboom, Area-Selective Atomic Layer Deposition of In2O3:H Using a µ-Plasma Printer for Local Area Activation, Chem. Mater. 29 (2017) 921-925. doi:10.1021/acs.chemmater.6b04469.