Aerosol-assisted CVD of thioether-functionalised indium aminoalkoxides
- First Online:
- 176 Downloads
Thioether-functionalised indium aminoalcoholates have been used as single-source precursors in aerosol-assisted CVD processes. The obtained In2O3−xSx oxysulphide deposits show either a single indium sulphide phase for deposits with high sulphide content [>75% (S/(S + O)) for the t-butyl derivatives] or pronounced phase separation in indium oxide and indium sulphide for lower sulphide content [<62% (S/(S + O)) for the n-butyl derivatives]. In addition to thin films, polycrystalline 1D structures are obtained at slightly modified synthesis conditions. The materials are analysed by EDX, XRD, XPS, SEM, and TEM.
KeywordsAACVD Oxysulphides Indium aminoalcoholates Thin films Nanowires
Indium oxide In2O3 is a wide band gap material with a direct band gap of 2.9–3.0 eV and transparency in the visible range of the spectrum . Therefore, In2O3 has been widely used in optoelectronics, sensing and as a transparent conducting oxide [2, 3]. Similarly, indium sulphide In2S3 with a band gap of approximately 2.2 eV has been largely exploited for optoelectronics and as buffer layer in photovoltaic devices [4, 5].
In contrast, reports on indium oxysulphides are rare despite the fact that they could be a favourable replacement for CdS in buffer layers of Cu(In,Ga)Se solar cells . The term oxysulphide represents in this paper a material composition with oxide and sulphide ions In2O3−xSx, which can be related to a sulphide phase with substitutional oxygen in In2S3 or phase-separated In2O3/In2S3. Some reports indicate a linear dependence of the band gap with increasing oxygen content in a In2S3 matrix which should allow direct tuning of the electronic properties of the material, but no local oxygen distribution has been reported . There are different methods for the growth of In2O3−xSx described in the literature, but also the unintentional incorporation in indium sulphide during processing should be considered. Processes leading to indium-based materials with mixed sulphur and oxide anions include stepwise formation of superstructures via ALD , sputtering techniques combined with post-growth oxidation , electrodeposition , co-precipitation sol-gel synthesis , or annealing of In2S3 prepared by solid-state reaction .
Metal alkoxides are often used as precursors in molecule-to-material processes for oxide synthesis due to preformed metal–oxygen bonds and the tunability of physical and chemical properties . For instance, the formation of monomeric alkoxides of indium aminoalcoholates by the combination of sterical crowding and donor bonds from aminoligands has been described and the same strategy has proven successful for thioether-functionalised aminoalcoholates [13, 14]. The thermally labile thioether moiety in the organic backbone of thioether-functionalised aminoalcoholates allows the simultaneous incorporation of oxide and sulphide ions in the material . The thioether-functionalised indium alkoxides showed promising results in hot-injection pyrolysis; however, these indium alkoxides are not volatile enough for low-pressure chemical vapour deposition.
This study describes the application of these precursors in an aerosol-assisted CVD (AACVD) process and the corresponding thin film formation and nanostructure evolution with a general composition of In2O3−xSx. The advantage of the AACVD approach is the independency of a precursor’s volatility because this technique relies only on the solubility of precursors and the aerosol stream is used for the transport of the precursor species to the substrate .
Results and discussion
The t-butyl thioether derivatives In2 and In4 show significantly higher sulphur contents compared to n-butyl species In1 and In3. Similar behaviour has been previously observed in LPCVD of gallium homologues indicating the formation of more effective secondary decomposition products of the thioether moiety. The n-butyl derivatives In1 and In3 do not show a comparable trend. The sulphur content in the In2O3−xSx deposits obtained using In1 remains in the range of x = 1.28–1.56 and In3 have slightly higher sulphur contents x = 1.68–1.85 (< 62% (S/(S + O))) while In2 and In4 deposits contain more sulphur [x > 2.25; >75% (S/(S + O))]. All films exhibited carbon contamination, but values are not provided due to the inaccuracy of the EDX results for elements with such a small atomic number. In general, tertiary alcoholate derivative based material shows higher carbon contamination of their secondary alcoholate counterparts. This could be caused by kinetic limitations of the heterogeneous surface reactions at the moderate decomposition temperatures in combination with rather high precursor delivery rates . X-ray photoelectron spectroscopy (XPS) will give more insight into the carbon contamination, which is actually pronounced at the surface and highly diminished in the bulk of the films.
The peak positions for In2 can be well referenced to In2S3, while signals for In1 are shifted 0.4–0.6 eV to lower binding energies [23, 24]. This can be most likely attributed to the higher oxygen content in In1-derived thin films as In2O3 was reported to give signals in this range of the In3d region . However, the use of a laboratory X-ray source in the XPS setup without monochromatisation should also be considered as it leads to increased signal widths and, therefore, the signal might not be sufficiently resolved to show the two different species expected for In1-derived coating, which showed phase separation in the XRD pattern. Sulphide contents after sputtering were found at x = 1.1 [37% (S/(S + O))] and 2.0 [67% (S/(S + O))], which is in good agreement with sulphide contents obtained from EDX analysis (x = 1.28 and 2.3 for In1 and In2, respectively). However, the exact relative atomic proportion of each element in the films cannot be determined accurately by XPS because of the different etching rates of indium, sulphur, and oxygen. Therefore, we generally relate to the sulphur content determined by EDX.
CVD growth of indium oxide nanowires usually requires a metal growth seed for an effective nanowire formation . Similarly, the literature reports on gas phase processes mainly deal with gold-supported vapour–liquid–solid (VLS) growth . No metal growth promoter is used in the here-described study and the morphology resembles the morphology of In2S3 nanostructures prepared by AACVD using a dithiocarbamate precursor without the presence of a metal growth seed . However, the study on the pure elongated In2S3 nanostructures does not provide any data, which allows to assign a growth mechanism that could also be present in the here-described samples.
The inhomogeneity is most likely caused by spinoidal decomposition, surface diffusion of the growing material and site-specific adsorption and diffusion of decomposition products preferentially forming phase-separated indium oxide and sulphide crystallites.
We have demonstrated that thioether-functionalised indium aminoalcoholates are well suited as single-source precursors for the direct synthesis of crystalline indium oxysulphide In2O3−xSx thin films via AACVD processes. The resulting thin film coatings exhibit varying degrees of sulphur incorporation ranging from x = 1.28 to 2.80 [42–93% (S/(S + O))] depending on the selected precursor species as well as deposition temperature. Furthermore, depending on the process parameters nanowires and rod-type 1D structures are formed showing nanophase separation concomitant with high number of defects and strong twinning of the crystallites. High sulphide contents lead to materials with phase-pure sulphide crystal structure while lower sulphide content leads to phase separation and the formation of coexisting indium oxide and sulphide phases.
All chemicals were purchased from Sigma-Aldrich, ABCR, and TCI Europe and used as received. Manipulations of metal alkoxides were conducted taking stringent precautions against atmospheric moisture using Schlenk or modified Stock-type apparatuses. Solvents were desiccated using standard techniques and stored over Na wire, if applicable. Synthesis procedures and characterisation of the metal alkoxide precursors have been described in the literature . Silicon (911) substrates were purchased from CrysTec.
Aerosol-assisted chemical vapour deposition
AACVD was conducted in a home-built hot-wall reactor consisting of a laboratory-grade NS 29 glass tube in a tube furnace equipped with a gas inlet nozzle attached to a 50 cm3 flask. For a deposition, 50–75 mg of precursor were dissolved in 10 cm3 dry n-hexane and placed in the aerosol generator. The aerosol was generated ultrasonically and transported into the reactor by a stream of dry, deoxygenised nitrogen, which was controlled with a mass flow controller (MKS PR4000B) and set to 40–70 sccm. The apparatus was desiccated by heating to >400 °C for 1 h in a carrier gas stream prior to use. The Si (911) substrate temperature was initially monitored ex situ via a thermocouple under reaction conditions and found to be approx. 100 °C lower than the oven temperature. Therefore, all temperatures mentioned in the manuscript refer to the substrate temperature and not the settings of the tube furnace. In addition to thin film growth, gas phase nucleation was generally observed forming a particulate deposit at the end of the heated zone.
Diffraction patterns were recorded on a PANanalytical XPERT Pro PW 3050/60 diffractometer with Cu Kα radiation (λ = 1.5406 Å) and an X’Celerator detector. Data analysis was performed using HighScore Plus software. SEM analysis was carried out with a FEI Quanta 200 FEG SEM equipped with an EDX detector for elemental analysis. TEM micrographs were acquired using a FEI TECNAI F20 TEM operated at 200 kV, which is equipped with high-angle annular dark-field (HAADF) STEM and EDX detector. XPS measurements were performed on a commercially available SPECS STM/XPS UHV setup equipped with a non-monochromatised dual-anode X-ray tube with Al and Mg Kα anodes (XRC 50, SPECS) and a hemispherical analyser PHOIBOS 100 (SPECS) with multi-channeltron detector. The base pressure was in the low 10−9 mbar regime. The samples were introduced into the UHV system via a load lock after mounting them on a steel plate. Samples were measured at room temperature with Al Kα radiation in the as is state and after sputtering for 45 min at 5 × 10−6 mbar Ar and an acceleration voltage of 1 kV. The obtained data were analysed using the commercial software CasaXPS and peak positions were corrected via the signal for graphitic carbon and double checked by measuring the Fermi Edge. Peak positions were identified using the NIST XPS database (http://srdata.nist.gov/xps/). Peaks were fitted after Shirley background subtraction using a Gaussian–Lorentian mixed peak shape (30% L) without further constraints. For comparing indium, carbon and sulphur signals the intensity was corrected by the cross sections obtained from the Elletra Trieste database (https://vuo.elettra.eu/services/elements/WebElements.html).
Open access funding provided by TU Wien (TUW). We thank the X-ray centre for access to the powder XRD machine and the University Service Centre for TEM (USTEM) for access to the electron microscopes at TU Wien. The TU Wien is acknowledged for financial support. We thank Thomas Haunold for acquiring some XPS spectra.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.