A critical role of hydrogen sulfide evolution during MOCVD of single phase thin film tin sulfide using ditertiarybutylsulfide as a less toxic precursor

This item is brought to you by Swansea University. Any person downloading material is agreeing to abide by the terms of the repository licence. Copies of full text items may be used or reproduced in any format or medium, without prior permission for personal research or study, educational or non-commercial purposes only. The copyright for any work remains with the original author unless otherwise specified. The full-text must not be sold in any format or medium without the formal permission of the copyright holder.

A critical role of hydrogen sulfide evolution during MOCVD of single phase thin film tin sulfide using ditertiarybutylsulfide as a less toxic precursor

INTRODUCTION
Thin film tin sulfide (SnS) is an attractive material for solar absorption in photovoltaics due to its high absorption coefficient at visible wavelengths [1]. Its earth abundant non-toxic constituents give potential for high throughput module production, which is an important consideration as energy demands increase and materials availability of existing technologies, such as cadmium telluride (CdTe) and copper indium gallium diselenide (CIGS) become uncertain [2]. To date, progress in conversion efficiencies for SnS solar cells has been slow, with a certified world record of 4.4 % [3]. This is very short of the theoretical conversion efficiency limit of 32 % [4]. The challenges with using SnS as a thin film solar cell absorber have mainly been associated to its small grains and poor band alignment with p-n junction counter-parts [3]. This leads to heavy recombination within the bulk material due to grain boundaries acting as recombination centres and poor energy matches to n-type buffers creating losses at the junction.
There are many reports focusing on the appropriate n-type buffer material to create small conduction band offsets with SnS, which has a low electron affinity [5]. Post-growth annealing steps are generally carried out to enlarge the SnS grains and reduce grain boundary defect concentration [3]. This is typically carried out in a sulfur atmosphere to reduce sulfur loss from the annealed film. Hydrogen sulphide (H 2 S) has been shown [3] to be a successful sulfur precursor, also assisting in grain enlargement during post-growth annealing. Unfortunately, H 2 S is a highly toxic gas so special measures are required for its use [6]. Ditertiarybutylsulfide (t-Bu 2 S) is an alternative sulfur source, with less toxicity than H 2 S and has been used by the authors employing the metal organic chemical vapor deposition (MOCVD) process [7]. Tetraethyltin (Et 4 Sn) was used in this study as the tin precursor with successful growth of single phase SnS on to molybdenum (Mo) substrates. Post-growth annealing was not necessary with high growth temperatures of 470 o C facilitating large grain growth. In situ mass spectroscopy was used to give insight into the reaction process.

EXPERIMENT
The MOCVD process used either nitrogen or hydrogen as carry gas to transport the Et 4 Sn and t-Bu 2 S vapor to the reaction chamber. The Mo substrates were 25 × 75 mm 2 in area and placed on to a heated graphite block with injector position 3 mm above over the upstream region of the substrate. The vapor flow followed a horizontal configuration towards the exhaust. The chamber was kept at atmospheric pressure and the concentration of the precursor vapor was determined by the partial pressure, which was 5.9 × 10 -4 and 3.7 × 10 -3 atmospheres for Et 4 Sn and t-Bu 2 S, respectively. It was necessary to use a 6-fold greater concentration of t-Bu 2 S to Et 4 Sn due to the higher vapor pressure of sulfur compensating for its reevaporation from the growing SnS film. The authors have reported [7] these optimised parameters in another article.
A Cirrus portable mass spectrometer was used to sample exhaust gas species after the growth chamber to determine by-products produced from the chemical reaction process after SnS film formation. Scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX) were carried out with a Hitachi TM3000 bench top instrument and used to assess grain size/shape and approximate stoichiometry of the Sn-S deposits. High resolution (HR)-SEM was carried out with a Hitachi S-4800 instrument equipped with field emission gun (FEG) using beam settings at 2 -10 keV and 10 µA. X-ray diffraction (XRD) was used with a Phillips X'PERT MPD instrument to determine the thin film crystal structure.

DISCUSSION
It was not possible to obtain thin film SnS when using nitrogen as the carrier gas. This was even the case at chamber temperatures up to 600 o C. Thin film SnS only deposited on to the substrates when using hydrogen as carrier gas and at temperatures greater than 410 o C.

Thermal decomposition of Et 4 Sn
The decomposition products of Et 4 Sn, from Et 3 Sn to EtSn, with loss of ethyl groups from the precursor, were all detected as soon as Et 4 Sn was injected into the chamber using hydrogen as carrier (Figure 1). On heat up, detected concentrations of these by-products increased, correlating with outgassing of sulfur from parasitic wall deposits within the reactor chamber. At a critical temperature between 410 o C and 420 o C, H 2 S was detected. The temperature reached a maximum of 470 o C with the hydrogen flow over the substrate. The carrier gas was then switched from hydrogen to nitrogen, resulting in an immediate drop in all detected exhaust gas by-products, even though the chamber temperature increased towards 600 o C. This gave two observations, the first being that hydrogen was cooling the substrate surface to a certain degree and secondly that the Et 4 Sn thermal decomposition was reliant on hydrogen being used as the carrier gas.

Evolution of H 2 S
The initial formation of H 2 S during temperature increase with hydrogen flow is attributed by the sulfur release from parasitic wall deposits. Addition of t-Bu 2 S results in approximately an order of magnitude increase in the partial pressure of H 2 S detected ( Figure 2). After t-Bu 2 S injection, the sulfur follows a similar trend to H 2 S, but at lower partial pressures. The t-Bu 2 S decomposition products of t-Bu-S and butane were also detected upon t-Bu 2 S injection. When Et 4 Sn was then added to the reaction chamber, the detection of both t-Bu-S and butane increased and the butyl group released from t-Bu 2 S became detectable, suggesting the Et 4 Sn precursor enhanced t-Bu 2 S decomposition. Also at the point of Et 4 Sn injection, some of the H 2 S and sulfur was observed to be consumed. It was determined that this was due to SnS being formed at the substrate surface.  The strong presence of Mo in the XRD pattern suggested the SnS film was thin, or that there was incomplete coverage of the SnS film on the Mo substrate. SEM was carried out on the film to get an indication of grain size ( Figure  4). The grains were large, spanning 2 µm across and 4 µm in length in many cases. However, there were areas of significant contrast which were identified as Mo using EDX. This correlated with the strong peak associated with Mo observed in Figure 3. This signified a large vertical growth rate with incomplete coalescence of the grains. Further growth of the SnS should result in coalescence of the films.

CONCLUSIONS
XRD determined that single phase thin film SnS was produced using t-Bu 2 S as a less toxic sulfur precursor alternative to H 2 S. However, high growth temperatures excessive of 410 o C were required to produce SnS with use of hydrogen as carrier gas being essential. This resulted in the formation of H 2 S within the growth chamber and was determined to be a necessity for Et 4 Sn and its decomposition products to react and form SnS. Large grains were produced using the as-grown process, but vertical growth rate was high leading to incomplete coalescence of the SnS grains, exposing areas of the Mo substrate. Further growth of the SnS would be expected to lead to coalescence of the grains.