Atomic layer deposition of aluminum sulfide thin films

0 downloads 0 Views 2MB Size Report
Dec 11, 2014 - Atomic layer deposition of aluminum sulfide thin films using trimethylaluminum and hydrogen sulfide. Soumyadeep Sinha, Neha Mahuli, and ...
Atomic layer deposition of aluminum sulfide thin films using trimethylaluminum and hydrogen sulfide Soumyadeep Sinha, Neha Mahuli, and Shaibal K. Sarkar Citation: Journal of Vacuum Science & Technology A 33, 01A139 (2015); doi: 10.1116/1.4903951 View online: http://dx.doi.org/10.1116/1.4903951 View Table of Contents: http://scitation.aip.org/content/avs/journal/jvsta/33/1?ver=pdfcov Published by the AVS: Science & Technology of Materials, Interfaces, and Processing Articles you may be interested in Molecular layer deposition of alucone films using trimethylaluminum and hydroquinone J. Vac. Sci. Technol. A 33, 01A115 (2015); 10.1116/1.4900934 Waterless TiO2 atomic layer deposition using titanium tetrachloride and titanium tetraisopropoxide J. Vac. Sci. Technol. A 32, 01A114 (2014); 10.1116/1.4839015 Atomic layer deposition of cobalt oxide thin films using cyclopentadienylcobalt dicarbonyl and ozone at low temperatures J. Vac. Sci. Technol. A 31, 01A145 (2013); 10.1116/1.4772461 Atomic layer deposition of zinc indium sulfide films: Mechanistic studies and evidence of surface exchange reactions and diffusion processes J. Vac. Sci. Technol. A 31, 01A131 (2013); 10.1116/1.4768919 Infrared spectroscopic study of atomic layer deposition mechanism for hafnium silicate thin films using HfCl 2 [ N ( SiMe 3 ) 2 ] 2 and H 2 O J. Vac. Sci. Technol. A 22, 2392 (2004); 10.1116/1.1806442

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40

Atomic layer deposition of aluminum sulfide thin films using trimethylaluminum and hydrogen sulfide Soumyadeep Sinha Department of Energy Science and Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

Neha Mahuli Centre for Research in Nanotechnology and Science, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

Shaibal K. Sarkara) Department of Energy Science and Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

(Received 28 August 2014; accepted 1 December 2014; published 11 December 2014) Sequential exposures of trimethylaluminum and hydrogen sulfide are used to deposit aluminum sulfide thin films by atomic layer deposition (ALD) in the temperature ranging from 100 to 200  C. ˚ per ALD cycle is achieved by in-situ quartz crystal microbalance Growth rate of 1.3 A measurements. It is found that the growth rate per ALD cycle is highly dependent on the purging time between the two precursors. Increased purge time results in higher growth rate. Surface limited chemistry during each ALD half cycle is studied by in-situ Fourier transformed infrared vibration spectroscopy. Time of flight secondary ion-mass spectroscopy measurement is used to C 2014 American Vacuum Society. confirm elemental composition of the deposited films. V [http://dx.doi.org/10.1116/1.4903951]

I. INTRODUCTION

II. EXPERIMENT

The device application of pure aluminum sulfide (Al2S3) is highly precluded due to its instability under ambient.1 However, the major use of Al2S3 is realized in Al doped chalcogenide materials that show wide range of optoelectronic properties.2–5 Al3þ doping in metal sulfides is successfully materialized either by cophysical vapor deposition, sputtering, or sometimes by forming nanolaminate or multilayer structures with Al2S3. The deposition chemistry of pure Al2S3 is yet to be widely explored, barring a few scattered reports. Wehmschulte and Power6 reported ligand stabilized Al2S3 nanoparticles synthesis through solution growth process. In addition, multilayer structures consisting of thin films of Al2S3 were consistently deposited by e-beam evaporation by some groups.1,7,8 Most recently, Rachmady et al.9 have successfully grown Al2S3 using trimethylaluminum (TMA) and hydrogen sulfide (H2S) for thin film transistor application; however, no detail chemistry was revealed. Our interest lies on the Al3þ doping in atomic layer deposition (ALD) grown II–VI materials, particularly for transparent conductor application. Thus, understanding the deposition mechanism is the major goal here. In this paper, we show the deposition of Al2S3 by atomic layer deposition using TMA and H2S in the temperature range 100–200  C. Film growth is studied by in-situ quartz crystal microbalance (QCM). The deposition chemistry during individual ALD half cycle is monitored by in-situ Fourier transform infrared (FTIR) spectroscopy. Notably, higher mass gain per ALD cycle is found with increase in purge time.

A. Material synthesis

a)

Electronic mail: [email protected]

01A139-1 J. Vac. Sci. Technol. A 33(1), Jan/Feb 2015

A hot wall viscous flow ALD reactor was used to deposit aluminum sulfide (Al2S3) on microscopic glass and Si (111) substrates using (TMA; 97% from Aldrich) and H2S (99.99% from Asia Advanced Gas, Hong Kong). Ultra high purity N2 gas with the flow rate of 160 SCCM was used in continuous flow to keep the reactor base pressure constant at 1 Torr. All the reactants were kept at room temperature and dosed into the reactor with N2 carrier gas. The dose times were controlled through the computer controlled pneumatic valves and carrier gas flow was controlled through mass flow controllers from MKS Instruments. The two precursors were dosed into the reactor at a constant partial pressure of 0.1 and 1.5 6 0.1 Torr, respectively, for 1 s each. During the course of film deposition, reactant exposure and purging time were maintained by time sequence m*t1  t2  n*t3  t4, where t1 and t3 are the dose times for TMA and H2S, respectively, while t2 and t4 are the purging times and m and n, the respective number of variable exposures. All the pressure changes in the reactor were monitored by the capacitance manometer (Baratron, MKS Instruments). B. Material characterizations

Film growth was monitored by in-situ QCM. Measurements were recorded using Inficon SQM-160 thickness monitor. Gold-coated AT-cut quartz crystals having 6 MHz resonance frequency were used for in-situ QCM studies. The crystals were mounted in a crystal drawer and retainer assembly and sealed with nonconducting silver epoxy to prevent gas flow into the crystal chamber. An

0734-2101/2015/33(1)/01A139/8/$30.00

C 2014 American Vacuum Society V

01A139-1

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40

01A139-2 Sinha, Mahuli, and Sarkar: Atomic layer deposition of aluminum sulfide thin films

additional positive pressure was applied inside the crystal drawer assembly by keeping constant N2 flow. The ZnS capped Al2S3 films on Si substrate were used for ex-situ thickness measurement by x-ray reflectivity (XRR) measurement using Rigaku Smartlab x-ray diffractometer equipped with Cu-Ka source. Experimentally, obtained data were fitted using GLOBALFIT software to obtain the film properties. In-situ FTIR spectroscopy studies were performed to understand the surface-limiting chemical reactions in a different reactor similarly equipped with programmable gas feed and capacitance manometer as the one described earlier. FTIR spectra were recorded with Bruker Vertex 70 spectrophotometer. IR beam was passed through the reactor equipped with ZnSe window and recorded with liquid nitrogen cooled MCT detector. Each scan presented here is an average of 100 scans taken within 370–4000 cm1 with a resolution of 4 cm1. KBr pellets were used as substrate for all FTIR measurements. Secondary ion-mass spectroscopy (SIMS) measurements were performed to examine relative concentration of the different elements along the depth of the film. The time of flight SIMS (TOF-SIMS) profile was acquired using a PHI TRIFT V nanoTOFTM instrument from UULVAC-Physical Electronics, MN, USA. Pulsed primary beam of 30 keV, 40 lm Gaþ ion source was used as primary gun for the elemental analysis. Csþ ions of 3 kV were bombarded in a raster size of 600  600 lm in a dual beam mode for depth sputtering. Alternately, the Gaþ and Csþ ions were used to acquire data and sputter the layers, respectively.

01A139-2

time was set to 1 s each separated by 110 s purging. Almost, a linear growth was observed apart from the first 1–2 cycles. During the first couple of cycles of Al2S3 growth on ZnS, comparatively low deposition rate was observed. This could be well attributed to the nucleation regime of the film growth. Sauerbrey equation11,12 was used for the calculation of mass gain from the changes in resonance frequency measurements by the QCM. The total mass gain during the first 20 cycles of the deposition was found to be ca. 960 ng on the QCM surface. Thus, the average mass gain was calculated to be 48 ng/cm2 per ALD cycle. Considering the density of the Al2S3 as 2.3 g/cm3, as found from the XRR measurements, ˚ per ALD cycle was the average growth rate of ca. 1.3 A found. Figure 1(b) shows a couple of representative ALD cycles from the linear growth regime. The pressure transient due to the reactant dosage measured by the Baratron capacitance manometer was also shown alongside. As expected from the proposed chemical reaction [Eq. (1)], considerable mass increase was clearly observed upon TMA exposure. The total mass change measured by QCM upon 1  105 L (1 L ¼ 106 Torr in 1 s) of TMA exposure was found to be ca. 40–45 ng/cm2 (Dm1). During the second half of the ALD reaction [Eq. (2)], upon H2S exposure, the surface became reterminated by the thiol (–SH) groups thus replicating the initial surface. The net increase in the mass of ca. 9–10 ng/

III. EXPERIMENTAL RESULTS A. Study of growth kinetics-in-situ quartz crystal microbalance

Atomic layer deposition of Al2S3 was performed in a custom-built laminar flow reactor using TMA and H2S within the temperature range 100–200  C. Reactant dosages and purge times were maintained by the computer controlled pneumatic valves and mass flow controllers. Following the zinc sulfide (ZnS) ALD reaction chemistry, where diethylzinc (DEZ) and H2S were used as the ALD precursors,10 the proposed deposition chemistry of aluminum sulfide could be written similarly in two separate surface limited half-reactions, as follows: SH þ AlR3 ! S  Al  R þ RH; 



Al  R þ H2 S ! Al  SH þ RH;

(1) (2)

where “*” denotes the surface species and R denotes the methyl group. Repeated AB cycles felicitated aluminum sulfide deposition. The film growth was monitored by in-situ QCM. Hundred cycles of ZnS were deposited on the QCM crystal inside the same ALD reactor using DEZ and H2S prior to Al2S3 deposition. This allowed similar starting surface for any consequent QCM experiments carried out. Unless specified otherwise, all QCM experiments were performed at 150  C. Figure 1(a) shows the mass gain versus time during first 20 cycles of Al2S3 deposition. Here, the reactant dose

FIG. 1. (Color online) (a) QCM growth characteristics of Al2S3 ALD on ZnS coated gold crystal at 1*TMA-110 s-1*H2S-110 s sequence of the precursors at 150  C. (b) QCM mass gain during two complete ALD cycles along with the precursor exposure at 150  C in the linear growth regime.

J. Vac. Sci. Technol. A, Vol. 33, No. 1, Jan/Feb 2015

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40

01A139-3 Sinha, Mahuli, and Sarkar: Atomic layer deposition of aluminum sulfide thin films

cm2 (Dm2) was measured after H2S exposure. Thus, the overall mass gain of a representative ALD cycle was found to be ca. 50–55 ng/cm2, which contributed to the net growth of the material per cycle. Considering the mass density of aluminum sulfide (2.3 g/cm3), the growth rate was found to ˚ per cycle at 150  C, which is comparable to be ca. 1.2–1.5 A the average growth rate as obtained and discussed previously. Figure 2 shows the experimentally determined x-ray reflectivity measurement along with the fitting using commercially available software, GLOBALFIT. The software independently calculates material density and thickness of individual layers. Al2S3 was being moisture sensitive, 300 cycles of the same grown on silicon substrate was encapsulated with 200 cycles of ALD grown ZnS. The resultant thickness of Al2S3 obtained by fitting the experimental data was 36 nm, corresponding to ˚ per cycle. This is in well agreement a growth rate of ca. 1.2 A with the QCM measurements within certain experimental error. The as deposited Al2S3 capped with ZnS showed no signature x-ray diffraction (XRD) peak of the material (apart from ZnS). Also due to its instability under ambient condition, it was difficult to perform any selected area diffraction in transmission electron microscope. Hence, on the basis of the XRD, we therefore conclude that as deposited Al2S3 films were amorphous in nature. Furthermore, in-situ QCM was employed to investigate the average growth rate per ALD cycle with the variation of the purge times at 150  C. As shown in Fig. 3, the average ˚ per cycle with the growth rate increased from 0.4 to 1.3 A increase in purge time from 40 to 90 s. Beyond 90 s, there was no change in the growth rate observed upon further increase in the purge time. Increase in the growth rate with increasing purge time is certainly nontrivial as per the mechanism described in ALD literature.13–16 It is also obvious that very low purge time can result in CVD type of reaction due to inefficient pumping. This may result in higher growth rate than ALD. While increase in the purge time can lead to some desorption, but ultimately a constant growth rate can be achieved. However,

FIG. 2. (Color online) X-ray reflectivity measurement data and fitting of the ALD grown ZnS capped Al2S3 film grown on Si substrate at 150  C.

01A139-3

FIG. 3. Growth rate per cycle of the Al2S3 ALD as function of purge time for the single dosing of each precursor at 150  C.

in this current observation for Al2S3 ALD, certainly the expected trends are not followed. Figures 4(a) and 4(b) show the self-limiting behavior of the reaction mechanism. Self-limiting reaction mechanism is a common phenomenon for any ALD process16–18 where the

FIG. 4. (Color online) Mass gains as a function of number (a) TMA and (b) H2S exposure (self-limiting characteristics) for the Al2S3 ALD at 150  C for different purge time.

JVST A - Vacuum, Surfaces, and Films

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40

01A139-4 Sinha, Mahuli, and Sarkar: Atomic layer deposition of aluminum sulfide thin films

growth rate saturates upon full conversion of the surface limited species. Beyond this, no further mass increase is found with increasing reactant dosage. Here, the reactant dosages were increased stepwise while the subsequent growth rate was measured. Figure 4(a) shows the growth rate with increase in the TMA dosage at two different purge times; 40 and 110 s. It is clearly depicted that the average growth rate saturates beyond a single dose of TMA; however, the value depends on the purge time. For 110 s purge time, the average mass gain per ALD cycle was ca. 46.5 ng/cm2, correspond˚ . In comparison shorter dose time ing to the growth rate 1.3 A of 40 s results in a mass gain of ca. 12 ng/cm2 per ALD ˚. cycle, corresponding to a growth of 0.4 A Figure 4(b) shows similar experiments where the average growth rate is plotted with varying H2S pulse. It is found that for 110 s purge time, growth rate almost saturates beyond the single dosage of H2S, equivalent to the exposure of ca. 1.5  106 L (1 L ¼ 106 Torr in 1 s). However, for low purge time of 40 s, the growth rate increases initially and saturates beyond 3–4 pulse of H2S dosage corresponding to the total exposure of 4.5  1066  106 L. Thus, it can be well concluded that increase in the purge time resulted in faster approach to the self-limiting regime. The reproducibility of the purge time dependent growth rate was further investigated. In-situ QCM was used to monitor the Al2S3 growth. Figure 5(a) shows a section of a train of repeated sequences as described below Sequence A: Ten ALD cycles with 40 s purge time were deposited that terminates with a 110 s purge. Sequence B: Three ALD cycles were deposited with 110 s purging that ends up with a 40 s purge. The above two sequences (A and B) were repeated for ten consecutive times while a fraction of it is shown in Fig. 5(a). Figures 5(b) and 5(c) depict the changeover from sequences

01A139-4

A to B and from B to A, respectively. It is very clear from Fig. 5(b) that the certain increase in the mass gain is only triggered by the 110 s purge in between. Similar but slightly different behavior can also be noticed from Fig. 5(c). Here, the reduced growth rate was triggered when the purge time decreased from 110 to 40 s. Such increased or decreased growth rate that was triggered by the preceding purge time was found highly reproducible. This observation probably underlines the fact that the growth rate, which is a combination of sequential chemical reactions, is kinetically limited. Longer time enables more surface sites that can be availed for the next half cycle, leading to an increased mass gain. The dependency on the surface sites was critically studied for first 6–7 cycles with in-situ QCM. Figures 6(a) and 6(b) show the evolutions of mass gain during Al2S3 deposition with purge time 40 and 110 s, respectively, on –SH terminated ALD grown ZnS surfaces. Here, ZnS ALD was performed in the same reactor with sequential exposure of diethylzinc and H2S. Figure 6 clearly portrays nonlinear growth during first few cycles in both the cases that had same starting surface. Hence, it is noteworthy here that the total mass gain for the first cycle remains constant at ca. 18 ng/cm2, irrespective of the purge time. When the purge time was set to 40 s, beyond couple of cycles the average growth rate decreased rapidly and attained linear growth per cycle ˚ . On the other hand, when the purge time was 110 s, as 0.4 A the growth rate increased and reached to a constant value of ˚ per 50–52 ng/cm2, corresponding to a growth rate of 1.3 A cycle. Furthermore, we investigated the temperature dependency of the said reaction with 40 and 110 s of purging while keeping 1 s dosage constant for both TMA and H2S. Figure 7 shows the growth rate obtained at three different deposition temperatures; 100  C, 150  C, and 200  C. We refrained

FIG. 5. (Color online) QCM study of the Al2S3 ALD for single sequential dose of TMA and H2S for 40 and 110 s purging time, where (a) mass gain for the alternative purge time of 40 and 110 s, respectively, (b) mass gain for 40 s purge time which terminating by 110 s purge time, and (c) mass gain for 110 s purge time with the termination of 40 s purge time. J. Vac. Sci. Technol. A, Vol. 33, No. 1, Jan/Feb 2015

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40

01A139-5 Sinha, Mahuli, and Sarkar: Atomic layer deposition of aluminum sulfide thin films

01A139-5

FIG. 6. (Color online) QCM studies of the first few cycles of the Al2S3 ALD on the ALD grown ZnS for (a) 40 s and (b) 110 s purge time, respectively, for the single dose of both the precursors at 150  C. QCM mass gain as a function of number of ALD cycles at (c) 40 s and (d) 110 s purge time.

from either increase or decrease in the temperature of the reaction due to two inherent reasons. Decrease in temperature lower than 100  C often results in difficulties of pumping the stray chemical out, particularly H2S, which might initiate some CVD type deposition. Above 200  C, the reliability of the QCM decreases where statistical analysis usually plays considerable role. As can be seen from Fig. 7, when the films were deposited with higher purge time the growth rate remains constant till 150  C but decreased with increase in temperature. This can be well attributed to desorption of surface species. However, a relatively opposite behavior can be noted when the material was grown with a

FIG. 7. (Color online) Growth rate per Al2S3 ALD cycle with the variation of substrate temperature at 40 and 110 s purge time.

40 s purge time. Here, increase in the temperature, beyond 150  C, results in increase of the growth rate. It should be mentioned here that the growth rate is not only determined by the available surface sites, but the reaction kinetics also play a major role.19,20 At higher temperature, thermal energy helps to increase the reaction rate that in turn results escalation in the deposition rate. B. In-situ FTIR spectroscopy studies of surface chemistry

In-situ FTIR spectroscopy was recorded during each half reaction of aluminum sulfide ALD at 150  C, under saturated precursor dosage condition, to study the changes in surface chemistry during each monolayer formation. The infrared difference spectra after every half cycle during aluminum sulfide growth are represented in Fig. 8. All measurements were carried out after 50 cycles of aluminum sulfide growth on bare KBr substrate. This was done to limit the reaction unperturbed from the surface related effect. The difference spectra were obtained taking the spectrum of the previous half reaction as reference. Spectra were recorded after each AB cycle for five consecutive AB cycles of aluminum sulfide ALD with saturating exposures of TMA and H2S, respectively. In these difference absorbance spectra, the removal of surface species was indicated by the negative absorbance features, and the positive absorbance features indicated the presence of surface species. The FTIR spectra in Fig. 8(a) show the absorption peak corresponding to C–H stretching21,22 at around 3016 cm1 that characterizes the appearance of that particular surface species. The positive nature of the absorption

JVST A - Vacuum, Surfaces, and Films

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40

01A139-6 Sinha, Mahuli, and Sarkar: Atomic layer deposition of aluminum sulfide thin films

FIG. 8. (Color online) In-situ FTIR difference spectra during four alternate dosing of TMA and H2S showing the (a) presence and (b) removal of C–H surface species.

peak corresponding to C–H stretch was observed after every dose of TMA. The clear presence of C–H group after the TMA dose indicated the presence of the absorbed surface species during the first half reaction. Alternately, the appearance of a symmetrical negative spectrum after every saturated dose of H2S shown in Fig. 8(b) indicated the absence of the C–H group. Thus, the flip–flop appearances of the absorption peak corresponding to C–H stretch indeed proved the two surface half reactions as mentioned earlier. It can be mentioned here that no particular changes for the

01A139-6

S–H vibration stretch could be discernible here. This is however not uncommon due to the difficulty in observation of this particular vibration frequency owing to its weak absorbance intensity as was seen earlier too.23 To understand the underlying chemistry that results in the purge-time dependent growth rate as obtained from QCM, we performed the in-situ FTIR spectroscopy study during aluminum sulfide ALD at 150  C. The infrared difference spectra were recorded after each half-cycle reaction. Similar to Figs. 8(a) and 8(b), in Fig. 9(a), we also show the appearance of C–H stretching21,22 at ca. 3016 cm1 after each TMA dose, which indicates the TMA incorporation while the negative C–H stretching represents removal of –CH3 thus incorporation of –SH groups. The spectra represented by the dotted line was obtained after 40 s of purging while the one represented by the solid line was obtained after 110 s of purge between the H2S and TMA. Thus, from Fig. 9(a), it can be easily inferred that longer purge time between the H2S and TMA results in increase in next TMA dose absorbance intensity, which can be related to the increase in the available –SH surface sites per unit area from the previous half-cycle. This is in agreement with the QCM growth as seen earlier in Fig. 5. To investigate the mechanism further, we performed the time dependent FTIR measurement during the purging after the H2S dose. Here, we probe negative C–H stretching as an indicative of positive –SH. It is found in Fig. 9(b) that the peak intensity initially decreases with time and comes to a saturating level beyond 110 s. Such observation clearly indicates an effective increase in the active surface sites available for the

FIG. 9. (Color online) In situ FTIR difference spectra during aluminum sulfide ALD showing the effect of purge time on the (a) presence and (b) removal of C–H surface species after TMA and H2S dose, respectively. (c) Effect of purge time on C–H surface species after TMA dose. J. Vac. Sci. Technol. A, Vol. 33, No. 1, Jan/Feb 2015

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40

01A139-7 Sinha, Mahuli, and Sarkar: Atomic layer deposition of aluminum sulfide thin films

next half-cycle. A similar experiment was performed with the TMA, where positive C–H stretch was monitored with respect to time during purging after the TMA dose, as shown in Fig. 9(c). There is no noticeable change in the peak intensity observed. This indeed is in agreement with QCM studies that show that negligible change occurs in the mass change after H2S dose half cycle irrespective of the purge time after TMA dose. Due to the lack of evidences, it is difficult to comment on such slow reaction kinetics but our experimental finding definitely suggest that purge time plays a crucial role to improve the effective number of surface sites after H2S dose only. C. Secondary ion mass spectrometry measurement

Figure 10 shows the depth profile elemental analysis of the aluminum sulfide film deposited on Si substrate, sandwitched between ALD grown ZnS layers due to the moisture sensitive nature of aluminum sulfide.1 The depth profile of the deposited film was carried out by the TOF-SIMS in a negative mode. It is clear from Fig. 10 that the top surface layer consists of high ZnS counts contributing to the ZnS top capping layer. The ZnS counts go down sharply with an increase in AlS intensity, indicating the interface where formation of Al2S3 starts. Throughout the layer of aluminum sulfide, AlS intensity is constant indicating uniform distribution of the elements. Further crater sputtering confirms the presence of bottom ZnS layer with expected decrease and increase in AlS and ZnS profiles, respectively, before reaching the Si substrate. The gradual decrease in AlS intensity can be attributed to minimal diffusion of Al2S3 film in bottom ZnS layer. Here, we also show the profile of AlO as the combination of aluminum and oxygen. This profile clearly shows that the counts of AlO were lower and below that of the Si substrate, which prove the absence of oxygen in the deposited film.

FIG. 10. (Color online) Elemental analysis of Al2S3 film by using TOFSIMS depth profile which was deposited on ZnS coated Si substrate and capped with ZnS.

01A139-7

IV. SUMMARY AND CONCLUSIONS Atomic layer deposition technique was used to deposit Al2S3 films using TMA as the metal source and H2S as the sulfur source, respectively. The growth rate was found self-limiting for 1 s exposure of both TMA and H2S precursors with 110 s of N2 purging at 150  C. The average ˚ per ALD cycle was obtained, measgrowth rate of 1.3 A ured by in-situ QCM and subsequently verified by XRR. Reaction chemistry during each ALD half cycle was examined by in-situ FTIR measurements. TOF-SIMS measurement proved the elemental presence in the films. Film growth was found to be considerably dependent on the purging time after H2S dose. Slow reaction kinetics was hypothesized where the dwelling time of H2S and/or reaction temperature plays a substantial role in net film growth. ACKNOWLEDGMENTS This paper was based upon work supported in part under the US–India Partnership to Advance Clean EnergyResearch (PACE-R) for the Solar Energy Research Institute for India and the United States (SERIIUS), funded jointly by the U.S. Department of Energy (Office of Science, Office of Basic Energy Sciences, and Energy Efficiency and Renewable Energy, Solar Energy Technology Program, under Subcontract No. DE-AC36-08GO28308 to the National Renewable Energy Laboratory, Golden, Colorado) and the Government of India, through the Department of Science and Technology under Subcontract No. IUSSTF/ JCERDC-SERIIUS/2012 dated November 22, 2012. S.S. thanks National Center for Photovoltaics Research and Education (NCPRE) for fellowship. 1

P. F. Smet, D. Poelman, and R. L. Van Meirhaeghe, J. Appl. Phys. 95, 184 (2004). 2 K. Hani, C. Guangyu, L. Oleg, C. Lee, S. Park, and S. Alfons, J. Phys. D: Appl. Phys. 41, 185304 (2008). 3 P. Prathap, N. Revathi, Y. P. V. Subbaiah, K. T. Ramakrishna Reddy, and R. W. Miles, Solid State Sci. 11, 224 (2009). 4 T. Yasuda, K. Hara, and H. Kukimoto, J. Cryst. Growth 77, 485 (1986). 5 S. Yamaga, A. Yoshikawa, and H. Kasai, J. Cryst. Growth 86, 252 (1988). 6 R. J. Wehmschulte and P. P. Power, J. Am. Chem. Soc. 119, 9566 (1997). 7 P. F. Smet, J. E. Van Haecke, R. L. Van Meirhaeghe, and D. Poelman, J. Appl. Phys. 98, 043512 (2005). 8 P. F. Smet, J. E. Van Haecke, R. L. Van Meirhaeghe, and D. Poelman, J. Electron. Spectrosc. Relat. Phenom. 148, 91 (2005). 9 W. Rachmady, M. Radosavljevic, G. Dewey, and R. S. Chau, U.S. patent 8,344,418 B2 (1 January 2013). 10 Y. S. Kim and S. J. Yun, Appl. Surf. Sci. 229, 105 (2004). 11 G. Sauerbrey, Z. Phys. 155, 206 (1959). 12 B. B. Burton, F. H. Fabreguette, and S. M. George, Thin Solid Films 517, 5658 (2009). 13 M. D. Groner, F. H. Fabreguette, J. W. Elam, and S. M. George, Chem. Mater. 16, 639 (2004). 14 J. W. Elam, M. D. Groner, and S. M. George, Rev. Sci. Instrum. 73, 2981 (2002). 15 D. M. Hausmann, E. Kim, J. Becker, and R. G. Gordon, Chem. Mater. 14, 4350 (2002). 16 C. Ozgit, I. Donmez, M. Alevli, and N. Biyikli, Thin Solid Films 520, 2750 (2012). 17 S. M. George, Chem. Rev. 110, 111 (2010).

JVST A - Vacuum, Surfaces, and Films

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40

01A139-8 Sinha, Mahuli, and Sarkar: Atomic layer deposition of aluminum sulfide thin films 18

D. J. Lee, J. Y. Kwon, J. I. Lee, and K. B. Kim, J. Phys. Chem. C 115, 15384 (2011). 19 A. W. Ott, J. W. Klaus, J. M. Johnson, and S. M. George, Thin Solid Films 292, 135 (1997). 20 A. C. Dillon, A. W. Ott, J. D. Way, and S. M. George, Surf. Sci. 322, 230 (1995).

01A139-8

21

J. D. Ferguson, A. W. Weimer, and S. M. George, J. Vac. Sci. Technol., A 23, 118 (2005). D. N. Goldstein, J. A. McCormick, and S. M. George, J. Phys. Chem. C 112, 19530 (2008). 23 S. K. Sarkar, J. Y. Kim, D. N. Goldstein, N. R. Neale, K. Zhu, C. M. Elliott, A. J. Frank, and S. M. George, J. Phys. Chem. C 114, 8032 (2010). 22

J. Vac. Sci. Technol. A, Vol. 33, No. 1, Jan/Feb 2015

Redistribution subject to AVS license or copyright; see http://scitation.aip.org/termsconditions. Download to IP: 103.21.127.80 On: Sun, 14 Dec 2014 10:57:40