Influence of boat material on the structure, stoichiometry and optical properties of gallium sulphide films prepared by thermal evaporation

Influence of boat material on the structure, stoichiometry and optical properties of gallium sulphide films prepared by thermal evaporation

Materials Chemistry and Physics 149-150 (2015) 164e171 Contents lists available at ScienceDirect Materials Chemistry and Physics journal homepage: w...

2MB Sizes 0 Downloads 3 Views

Materials Chemistry and Physics 149-150 (2015) 164e171

Contents lists available at ScienceDirect

Materials Chemistry and Physics journal homepage: www.elsevier.com/locate/matchemphys

Influence of boat material on the structure, stoichiometry and optical properties of gallium sulphide films prepared by thermal evaporation Pritty Rao a, Sanjiv Kumar a, *, N.K. Sahoo b a b

National Centre for Compositional Characterization of Materials, Bhabha Atomic Research Centre, ECIL Post, Hyderabad 500062, India Atomic and Molecular Physics Division, Bhabha Atomic Research Centre, Mumbai 400085, India

h i g h l i g h t s  Gallium sulphide films are prepared by thermal evaporation from a Mo or Ta boat.  Mo-boat prepared pristine film has Ga and S in 1:1 atomic ratio and is transparent.  Ta-boat prepared pristine film is Ga rich and absorbing.  Mo/Ta-boat prepared films crystallise into Ga2S3/GaS on vacuum annealing.  Diffusion of gallium in glass on vacuum annealing improves transmission of films.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 August 2013 Received in revised form 22 August 2014 Accepted 3 October 2014 Available online 11 October 2014

The paper describes the deposition of thin films of gallium sulphide on soda-lime glass substrates by thermal evaporation of chemically synthesized powders consisting of gallium sulphide and gallium oxyhydroxide from a Mo or Ta boat and the evolution of their compositional, structural and optical properties on vacuum annealing. The films deposited from Mo or Ta boats possessed distinctly different properties. The Mo-boat evaporated pristine films were amorphous, transparent (a ~ 103 cm1) in visible region and had a direct band gap of about 3.2 eV. Vacuum annealing at 723 K brought about their crystallization predominantly into cubic g-Ga2S3 and a blue shift by about 0.2 eV. The Ta-boat evaporated pristine films were also amorphous but were absorbing (a ~ 104 cm1) and had a direct band gap of about 2.1 eV. These crystallized into hexagonal GaS and experienced a blue shift by more than 1.0 eV on vacuum annealing at 723 K. The dissimilar properties of the two kinds of films arose mainly from their different atomic compositions. The Mo-boat evaporated pristine films contained Ga and S in ~1:1 atomic proportions while those prepared using Ta-boat were Ga rich which impaired their transmission characteristics. The former composition favoured the stabilization of S rich gallium sulphide (Ga2S3) phase while the latter stabilised S deficient species, GaS. Besides inducing crystallization, vacuum annealing at 723 K also caused the diffusion of Ga in excess of atomic composition of the phase formed, into soda-lime glass which improved the optical transmission of the films. Gallium oxyhydroxide, an inevitable coproduct of the chemical synthetic process, in the evaporant introduced oxygen and hydrogen impurities in the films which do not seem to significantly influence their optical properties. © 2014 Elsevier B.V. All rights reserved.

Keywords: Thin films Evaporation Rutherford backscattering spectrometry Optical properties

1. Introduction Group III oxide and chalcogenide semiconductors are promising materials for photovoltaic and optoelectronic applications due to their interesting electrical and optical properties [1,2]. Gallium oxide (Ga2O3) and gallium sulphide are the typical representatives

* Corresponding author. E-mail address: [email protected] (S. Kumar). http://dx.doi.org/10.1016/j.matchemphys.2014.10.002 0254-0584/© 2014 Elsevier B.V. All rights reserved.

of the Group III oxide and chalcogenide families respectively. Ga2O3 has five polymorphs. It occurs in monoclinic b phase under ambient conditions. The other polymorphs are formed at higher temperatures. b-Ga2O3 is a wide band gap (5.0 eV) intrinsic insulator but acquires n-type conductivity on doping. As a result, it can serve as an ultraviolet transparent conducting oxide for flat panel displays and solar cells. Intrinsic b-Ga2O3, on the other hand, has applications in semiconducting lasers and field effect devices and is also used as an anti-reflecting coating.

P. Rao et al. / Materials Chemistry and Physics 149-150 (2015) 164e171

Gallium sulphide exists in two stoichiometric formulations: gallium monosulphide (GaS) and gallium sesquisulphide (Ga2S3). Both types of gallium sulphide exhibit polymorphism with b-GaS (hexagonal) and a-Ga2S3 (monoclinic) being the most stable polymorphs under normal conditions. GaS has, similar to graphite, a layered structure. A strong covalent interaction exists within the layers which, in turn, are bonded by weak van der Waals forces [2]. The layered structure imparts anisotropicity to the functional properties of the material. Gallium sulphide is probably one of the least studied compounds in the IIIeVI chalcogenide family, however, the unique structure of GaS has evoked considerable interest among the researchers in recent years. In terms of optical characteristics, GaS has a direct band gap (Eg) of about 3.0 eV and an indirect Eg of about 2.6 eV [3]. It exhibits, depending on composition, n-type or ptype conductivity. Ga-rich GaS is an n-type and S-rich GaS is a p-type semiconductor [4]. On the other hand, Ga2S3 has a direct Eg of about 3.5 eV and exhibits p-type conductivity [5]. In view of rather large Eg, gallium sulphide is a candidate material for buffer layers in a photovoltaic cell. GaS provides effective surface passivation to GaAs and has been shown to enhance its photoluminescence yield by two orders of magnitude [6]. It is also being investigated for the fabrication of near-blue-light emitting devices [7]. Thin films of gallium sulphide have been prepared by several methods that include metal-organic chemical vapour deposition (MOCVD) [5,8], modulated flux deposition [9,10], microwave glow discharge [11], reactive RF sputtering [3] and molecular beam epitaxy [12]. MOCVD has, in fact, been used to prepare GaS or Ga2S3 films using suitable precursors. However, in this technique the crystallinity of the films is strongly affected by the nature of the substrate and deposition temperatures [5,8]. The films prepared by modulated flux deposition, on the other hand, were amorphous and possessed Ga rich composition. Moreover these contained up to as high as 34 at.% oxygen, suggesting the formation of essentially ternary GaSxOy films [9]. The GaS films prepared by microwave glow discharge also contained up to 5 at.% oxygen as an impurity [11]. Interestingly, recently Chowdury and Ichimura have reported the deposition of GaSxOy films by photochemical deposition (PCD) and electrochemical deposition techniques [13,14]. GaSxOy films, like gallium sulphide films can be used as buffer layers in photovoltaic devices. The films reported in these studies contained only 5e18 at.% S signifying difficulties in its incorporation in the films by photochemical or electrochemical reactions. In the present paper we report our investigations on the preparation of gallium sulphide films by the thermal evaporation of chemically co-precipitated gallium sulphide and gallium oxyhydroxide (GaOOH) powders from a Mo or Ta-boat. The effects of vacuum annealing on the compositional, structural and optical properties of the films have also been probed. These investigations form a part of our current studies on the development of metal sulphide films for photovoltaic applications by the resistive evaporation of chemically synthesized sulphide powders. The synthesis of pure gallium sulphide powders, unlike those of copper or indium sulphide, by precipitation reaction in aqueous medium is difficult due to the propensity of gallium salts to undergo hydrolysis. Meanwhile the choice of thermal evaporation based on that the fact that it is a simple yet an effective technique for the deposition of elemental as well as compound films. It has been previously employed for the deposition of GaS films using GaS crystals as evaporants but the preparative method and the chemical composition of the films were not described in detail which prompted us to undertake the present investigations [15]. It is to be noted that these are important considerations in view of the fact that the optical properties of the films, as observed in current study, depend on the preparative conditions, particularly on the type of boat material used for evaporation. Furthermore, an assessment of

165

oxygen content of the films is desirable since gallium sulphide films often contain, for reasons not explained, oxygen as an impurity in fairly large concentrations [9,11]. Presently, the composition of the films has been comprehensively examined by ion beam analysis techniques with particular emphasis on the analysis of hydrogen and oxygen, and an attempt has been made to correlate it with the structural and optical properties of the films. 2. Experimental details 2.1. Chemical synthesis of evaporant Gallium sulphide was synthesized by co-precipitation method using sodium sulphide as the precipitating agent. The procedure involved the addition of 50 mL of 1500 mM sulphide solution to 250 mL of 190 mM gallium chloride solution in ambient conditions. The precipitation was slow and therefore the solution was aged overnight for complete precipitation. The precipitate was filtered and washed copiously initially with deionised water and subsequently with isopropanol. It was later dried under the flow of argon at 383 K for 2 h. The dried powder was used as an evaporant for the deposition of films. 2.2. Deposition of films The films were deposited on soda lime glass substrates at a rate of 1e2 Å/s by resistively heating the evaporant in a Mo-boat or Taboat at ~3  104 Pa vacuum in a thermal evaporation unit. Both Mo and Ta boats had a rating of 200 A and weighed ~1.63 g and 2.34 g respectively. Vacuum in the deposition chamber was created and maintained by an oil diffusion pump backed by a rotary pump. The substrates were initially cleaned chemically and were subsequently sputter cleaned in situ in Ar plasma. The distance between the boat and the platen onto which the substrates were fixed was about 70 mm. The platen rotated at 25 rpm to ensure uniform deposition. Though the substrates were not intentionally heated, their temperature rose to 323 K at the end of the depositions. The temperature of the substrates was measured by a thermocouple held close to the platen. Films of three different thicknesses namely ~130 nm, 280 nm and 390 nm were deposited by evaporation from Mo boats. These are referred to as Mo1, Mo2 and Mo3 films respectively. Similarly 150 nm, 230 nm and 280 nm films, referred to as Ta1 Ta2, Ta3 films respectively, were deposited by evaporation from Ta boats. The thicknesses of the films were determined by Rutherford backscattering spectrometry (RBS) with a precision of ~5%. A number of films of were deposited in a run. The thickness of these films varied within ±10%. Subsequent to their deposition the films were annealed in vacuum in a quartz tubular furnace in 523 Ke723 K temperature range for 4 h. The vacuum during annealing, created by a turbomolecular pump was better than 7  103 Pa. 2.3. Characterization 2.3.1. Phase analysis The phase evolution in the films was examined by glancingincidence X-ray diffraction (GI-XRD) (incidence angle ¼ 2 ; step size ¼ 0.1 ; scan speed ¼ 1 per minute) by a Rigaku (Ultima IV) diffractometer using Cu Ka radiation (l ¼ 1.5402 Å). The same instrument was used to analyse residues in Mo or Ta boat in powder mode. 2.3.2. Compositional analysis 2.3.2.1. Determination of atomic ratio of Ga, S and O by backscattering spectrometry. Films, pristine as well as those vacuum

166

P. Rao et al. / Materials Chemistry and Physics 149-150 (2015) 164e171

annealed in 523 Ke623 K temperature range were examined by 2.4 MeV a-Rutherford backscattering spectrometry (RBS) and 3.035 MeV 16O(a, a)16O resonant scattering to determine their atomic composition and thickness. The resonant scattering is a variant of backscattering spectrometry and has enhanced sensitivity for oxygen (~1 at.%) in comparison to conventional RBS [16]. The backscattering spectrometry experiments involved the bombardment of films at normal incidence with ion beams of requisite energies and the detection of the particles scattered at 170 angle by a Si-surface barrier detector. The quantitative analysis was performed by simulating the experimental spectra by SIMNRA. The overall uncertainty in the determination of atomic ratio of Ga, S and O was about ±5%. 2.3.2.2. Depth profiling of hydrogen by nuclear resonance reaction analysis. The depth profile of hydrogen in films was determined by nuclear resonance reaction analysis (NRRA) using the resonance at 6.44 MeV in 1H(19F,ag)16O nuclear reaction [17]. The measurements were accomplished by bombarding the films at normal incidence with a well collimated 19Fþ3 beam in energy steps of 15 keV beyond the resonance energy and measuring 6.1, 6.9 and 7.1 MeV grays, characteristic of the reaction, by a bismuth germanate detector at each step. Representing the film by the formula GaSpOqHr, the content of hydrogen (r) at a depth x was calculated using the relationship.

r ¼R

f ðstdÞ½εðGaÞ þ pεðSÞ þ qεðOÞ εðstdÞ  ½R  f ðstdÞ  εðHÞ

(1)

where f(std) is the atomic fraction of hydrogen in the standard, which in the present case is mylar (aluminium coated), R is ratio of the charge normalized yields of 6e7 MeV g-rays for the films and the standard, ε is the stopping cross section of the element indicated in bracket for 6.44 MeV 19F beam, and p and q are the relative contents of S and O determined by backscattering measurements. The depth x was be calculated by the equation



ðE  ER Þ εðfilmÞ

(2)

where E is the incident beam energy, ER is the resonance energy and ε (film) is the stopping cross-section of the beam in the film. The stopping cross sections of the elements were calculated using ZieglereBiersack formulations while the stopping cross sections of the films were calculated by the Bragg's rule of linear addition. The uncertainty in the determination of hydrogen content of the films was less than ±10%. The backscattering spectrometry measurements and nuclear reaction analysis were carried out using a 3 MV Tandetron (HVE, Europa). The experiments were conducted in a typical scattering chamber maintained at ~7  104 Pa vacuum.

density of films to be 3.65  103 kg/m3, the density of bulk gallium sulphide. The true density of a film can vary with its thickness but is invariably less than that of the corresponding bulk material. Therefore the conversion provides an underestimation, which to a rough estimate can be ~10%, of the thickness of the films in linear dimensions. The electrical resistivity of the films was measured by a four point probe resistivity measurement unit (PRO 4) using Keithley 2610A source metre. No specific metallic contacts were made for the measurements. 3. Results The precipitate was white and had an odour typical of sulphides. The powder obtained after heat treating the precipitate at 383 K in argon atmosphere was also white but had diminished odour. As evidenced by the X-ray diffractogram in Fig. 1(a), the heat treated powder was amorphous. An analysis of chemical composition by 1.5 MeV proton RBS showed that the powder contained Ga (31 at. %), O (58 at. %) and S (11 at. %), and was devoid of any metallic impurity. The proton backscattered spectrum of the powder overlapped with the simulated curve is shown in Fig. 1(b) for illustration. In addition to Ga, O and S the powder contained hydrogen as well which existed as hydroxyl (OH) species. The presence of OH group was established by a broad band around 3200 cm1 in the infrared spectrum (not shown) recorded for the powder. It can be ascribed to GaOOH, produced by the hydrolysis of gallium chloride in aqueous medium. Since gallium sulphide has a high solubility product, unlike group I metal sulphides its precipitation by sulphur bearing precipitating agents is difficult. However the hydrolysis of gallium chloride causes the co-precipitation of gallium sulphide and is present in the evaporant along with GaOOH. 3.1. Phase evolution The films prepared by evaporation from either a Mo-boat or a Ta-boat were amorphous in their pristine state. It is exemplified by the XRD patterns of as-deposited Mo3 and Ta2 films in Figs. 2(a) and 3(a) respectively. The Mo-boat evaporated films remained amorphous on vacuum annealing up to 623 K but underwent crystallization at 723 K. Such an amorphous to crystalline transition is typically represented by the diffraction patterns of the Mo3 films vacuum annealed at 623 K and 723 K in Fig. 2(b) and (c)

2.3.3. Optical and electrical measurements The optical absorbance of the films was measured in 300e900 nm wavelength region by a UVevisible spectrophotometer. The optical band gaps (Eg) were calculated using the relation:

ahy ¼ C hy  Eg

m

(3)

where a is the absorption coefficient, C is an energy independent constant and m is a constant that is ½ for a direct allowed transition. The a values were derived from the measured absorbance and thickness of films determined by RBS experiments. It is to be noted that RBS provides thickness in terms of areal density (at./cm2) which was converted into linear dimensions by assuming the

Fig. 1. (a) X-ray diffraction pattern (b) and 1.5 MeV proton backscattered spectrum of the precipitate after heat treatment at 383 K in argon atmosphere.

P. Rao et al. / Materials Chemistry and Physics 149-150 (2015) 164e171

167

and have been indexed accordingly. A comparison of JCPDS PDF No. 43-0916 and JCPDS PDF No. 30-0576 shows that bothg-Ga2S3 and GaS exhibit reflections in the vicinity of 2q ¼ 29 which makes an unambiguous distinction between the two phases, particularly when the crystallization is not extensive, difficult. In this condition, the reflection at 2q ¼ 49.4 (g-Ga2S3) and those at 2q ¼ 11.4 , 50.6 and or 50.9 (GaS) serve to provide better discrimination. The prevalence of reflections at latter 2q values in the diffraction patterns of Ta-boat evaporated films, together with the sharpness of the line at 2q ¼ 29 suggest that the specimens are phase-singular in GaS. These considerations further suggest that the broadness of the reflections of Mo-boat evaporated films (Fig. 2 (c)) can also arise due to the presence of GaS as a minor phase in association with the dominant g-Ga2S3 phase. It is to be noted that the patterns of pristine or vacuum annealed films do not contain any discernible reflection either from any impurity or GaO species. 3.2. Compositional analysis

Fig. 2. X-ray diffraction patterns of 390 nm films prepared by evaporation from Mo boat: (a) pristine, (b) vacuum annealed at 623 K for 4 h and (c) vacuum annealed at 723 K for 4 h.

respectively. The latter pattern consists of reflections at 29.6 and 49.8 which can be attributed to, in accordance with JCPDS PDF No. 43-0916, (111) and (220) planes respectively of g-Ga2S3 phase. It can be seen that the reflections are weak and broad, suggesting a rather poor crystallization of the films. The films prepared by evaporation from a Ta-boat, on the other hand, crystallized into hexagonal GaS above 623 K exhibiting re flections at 11.3 , 28.9 and 50.9 in the diffraction patterns. These reflections, as shown in the diffractograms of Ta2 film vacuum annealed at 623 K and Ta1 film at 723 K in Fig. 3(b) and (c) respectively, are consistent with the JCPDS PDF No. 30-0576 of GaS

Fig. 3. X-ray diffraction patterns of films prepared by evaporation from Ta boat: (a) 230 nm pristine film, (b) 230 nm film vacuum annealed at 623 K for 4 h and (c) 150 nm film vacuum annealed at 723 K for 4 h.

3.2.1. Atomic ratio of Ga, S and O Fig. 4 shows the spectra of a-particles backscattered from pristine Mo1, Mo2 and Mo3 films at an incidence beam energy Ea  3.038 MeV. The spectra are superimposed with their respective simulated curves. These are marked by the strong peaks of Ga, S and O from the films, and steps arising from Ca, Si, Na and O that constitute the substrate. The Mo2 and Mo3 films also contain In and Mo as impurity elements. Though the source of In is presently not known, Mo might have its origin in the Mo boat used for evaporation. The signals of Ga and S are due to Rutherford backscattering and their widths represent the thickness of the films. The signal of O in the films, on the other hand, results from 16O(a, a)16O resonant scattering at 3.035 MeV beam energy and is a measure of its content at a depth of about 15 nm at Ea ¼ 3.038 MeV and about 90 nm at Ea ¼ 3.055 MeV. The films were also examined by 2.4 MeV a-RBS wherein the scattering for O, similar to Ga and S, is Rutherford in

Fig. 4. 16O(a, a)16O resonant scattering spectra of (a) 130 nm, (b) 280 nm and (c) 390 nm pristine films prepared by evaporation from Mo boat. The spectra (a) and (c) are acquired at 3.038 MeV incident beam energy while the spectrum (b) is acquired at 3.055 MeV. The curve (d) is the 2.4 MeV a-RBS spectrum of the 390 nm pristine film. The spectra are superimposed with their respective simulated curves.

168

P. Rao et al. / Materials Chemistry and Physics 149-150 (2015) 164e171

Table 1 Areal densities, thickness, absorption coefficient (a) and direct band gaps of pristine and differently annealed films prepared by evaporation from a Mo boat. The asterisk (*) represents vacuum annealing. The areal densities and the thickness of film vacuum annealed at 723 K correspond to the undiffused layer. S. no.

Boat

State

NGa

NS

NO

Thickness (1015 at./cm2)

a (cm1)

248 585 742 782 744 727

84 150 312 202 284 264

591 1323 1815 1770 1748 1583

2.5 3.2 2.6 2.5 3.4 1.8

Band gap (eV)

(1015 at./cm2) 1. 2. 3. 4. 5. 6. a b c d

Mo1 Mo2 Mo3 Mo3 Mo3 Mo3 N(In N(In N(In N(In

þ Mo) þ Mo) þ Mo) þ Mo)

¼ ¼ ¼ ¼

18 12 18 17

   

(130 (280 (390 (390 (390 (390

1015 1015 1015 1015

nm) nm) nm)a nm)b nm)c nm)d

Pristine Pristine Pristine 523 K* 623 K* 723 K*

259 588 761 786 720 592

     

103 103 103 103 103 103

3.4 3.0 3.2 3.2 3.2 3.4

at./cm2. at./cm2. at./cm2. at./cm2.

nature. A typical 2.4 MeV a-RBS spectrum of the films (e.g. Mo3) is shown in Fig. 4(d) for comparison. It can be seen that the signal of oxygen is not as well pronounced as in the case of resonant scattering due to the low sensitivity of the technique for oxygen. The areal densities of Ga(NGa), S(NS) and O(NO) in the films determined by the two backscattering spectrometry techniques were in good agreement. The values listed in Table 1 are obtained by the resonance scattering method. A comparison of the data shows that the atomic ratio of Ga, S and O (NGa:NS:NO) in the films is ~43:42:15 and is not influenced significantly by the film thickness. However the content of impurity elements in the films was found to vary with thickness as simulations showed that while Mo3 and Mo2 films contained ~1.0 at. % and 0.1 at. % impurities respectively, Mo1 film did not contain them in a detectable level. Table 1 also lists the areal densities of films annealed in vacuum at different temperatures. It shows that vacuum annealing up to 623 K does not induce any significant compositional changes in the films. Spectrally, such films were analogous to their respective pristine films. However the spectra of films annealed at 723 K were marked by a tailing in the rear edge of Ga, for example as in Fig. 5(a) for a Mo3 film, indicating an interfacial diffusion of Ga into the glass substrate. Notably, there is no perceptible diffusion of sulphur. The depth profile of Ga in the vacuum annealed specimen is presented in the inset. A detailed spectral analysis by simulating the experimental data showed that while NGa:NS (inclusive of diffused Ga) in

the annealed specimen remains similar to that in the pristine Mo3 film within the experimental error of measurement, NGa:NS:NO in the undiffused layer that measures about 1600  1015 at./cm2 (~340 nm) in thickness, is ~37:46:16. Though the layer does not strictly bear the stoichiometry of Ga2S3, its S rich composition (NGa:NS ¼ 37:46) tentatively explains the formation of Ga2S3 phase in films vacuum annealed at 723 K. The composition, in fact, lends credence to the postulation made on the basis of XRD measurements about the coexistence of Ga2S3 and GaS as major and minor phases respectively in the films. At this point it is instructive to mention that the present results also explain the anomaly reported in reference [18] concerning the chemical composition (GaS) and the phase (Ga2S3) of gallium sulphide films. It is to be noted that the films were analysed by energy dispersive X-ray analysis (EDS) which measures the overall composition of the films and is oblivious to the occurrence of any diffusion process. The backscattered spectra of pristine Ta1, Ta2 and Ta3 films recorded at Ea ¼ 3.038 MeV along with the 2.4 MeV a-RBS spectrum of a Ta3 film are shown in Fig. 6. The spectral features of the films are, in general, similar to those prepared by Mo-boat evaporation. However the films are devoid of Mo or any other metal impurity. The most distinguishing feature of these films, as can be observed from Table 2 that lists their areal densities determined by resonant scattering, is that, irrespective of thickness they are invariably rich in Ga in comparison to Mo-boat evaporated films. For instance,

Fig. 5. 16O(a, a)16O resonant scattering spectra of films vacuum annealed at 723 K for 4 h: (a) 390 nm Mo-boat evaporated and (b) 280 nm Ta-boat evaporated films. The spectra are acquired at 3.050 MeV incident beam energy and are superimposed with the respective simulated spectra. The insets show the depth profiles of Ga in the respective specimens.

Fig. 6. 16O(a, a)16O resonant scattering spectra of (a) 150 nm, (b) 230 nm and (c) 280 nm pristine films prepared by evaporation from Ta boat. The spectra are acquired at 3.038 MeV incident beam energy. The curve (d) is 2.4 MeV a-RBS spectrum of the 280 nm pristine film. The spectra are superimposed with their respective simulated curves.

P. Rao et al. / Materials Chemistry and Physics 149-150 (2015) 164e171

169

Table 2 Areal densities, thickness, absorption coefficient (a) and direct band gaps of pristine and differently annealed films prepared by evaporation from a Ta boat. The asterisk (*) represents vacuum annealing. The areal densities and the thickness of films vacuum annealed at 723 K correspond to the respective undiffused layers. S. no.

Film

State

NGa

NS

NO

Thickness (1015 at./cm2)

a (cm1)

190 376 343 374 239 454 407

122 142 114 108 115 161 96

714 1078 981 1030 601 1288 926

6.7 2.3 1.5 2.9 3.5 2.8 2.6

Band gap (eV)

(1015 at./cm2) 1. 2. 3. 4. 5. 6. 7.

Ta1 Ta2 Ta2 Ta2 Ta2 Ta3 Ta3

(150 (230 (230 (230 (230 (280 (280

nm) nm) nm) nm) nm) nm) nm)

Pristine Pristine 523 K* 623 K* 723 K* Pristine 723 K*

402 560 524 548 247 673 423

NGa:NS:NO of pristine Ta3 films which have thickness similar to Mo2 films, is ~52:35:12 against the atomic composition of ~44:44:12 of the latter. Notwithstanding this difference, the influence of vacuum annealing up to 623 K is nearly the same as witnessed by Moevaporated films, however, the interfacial diffusion at 723 K is comparatively more pronounced. The phenomenon typically occurring under such conditions is displayed in Fig. 5(b) by the resonant backscattered spectrum of a Ta3 film vacuum annealed at 723 K and the depth profile of Ga in the specimen presented in the inset. The simulation of the spectrum showed that the undiffused layer measuring about 926  1015 at./cm2 (~200 nm) in thickness bears NGa:NS:NO ~ 46:44:10. As can be seen from Table 2, similar results were obtained for other films as well. The NGa:NS values are consistent with the prevalence of GaS phase in the films revealed by XRD. Another interesting observation pertains to the fact that the overall NGa:NS (inclusive of Ga diffused in glass) in the annealed films is about 15% less than that for the corresponding pristine films which is indicative of a preferential loss of Ga during vacuum annealing at 723 K. It must be mentioned that the films produced either by Mo or Ta boat evaporation were stable during beam irradiation; no loss of any element, S in particular, was observed during the measurements. It was ascertained by analysing the composition of the films as a function of beam fluence. This observation suggests that elements are chemically bonded even in pristine films since the loss of elemental sulphur under beam irradiation has been observed on several occasions. 3.2.2. Depth profile of hydrogen Since the evaporant contains hydrogen in significant proportions, the element is likely to get incorporated into the films during deposition. To probe this pristine and vacuum annealed Mo3 and Ta2 films were depth profiled for hydrogen by 1H(19F, ag)16O resonance reaction. The results of these measurements (calculated using Equation (1)) are displayed in Fig. 7. It shows that the element has an interesting distribution in the films which can be broadly divided into two regions namely R1 that spans from surface to a depth corresponding to about 75% of film thickness and R2 that covers the rest of the film. As shown in Fig. 7(a), the pristine Mo3 film contained ~5 at. % hydrogen in R1 region and ~12 at. % hydrogen in R2 region. In pristine Ta2 film (Fig. 7(d)), on the other hand, the hydrogen contents in the two regions were ~10 at.% and ~14 at.% respectively. It indicates that the incorporation of the element occurs mainly in the early stages of deposition. An examination of the profiles in Fig. 7 further shows that vacuum annealing brings about a desorption of hydrogen from both regions which increases with the temperature of annealing. As a result, the contents of hydrogen in R1 and R2 regions (Fig. 7(c)) declined to about 1 at.% and 3 at.% respectively on annealing the Mo3 film at 723 K. The Ta2 film also experienced reduction in its hydrogen content by nearly similar extent on vacuum annealing. The film annealed at 723 K contained about 3 at.% and 6 at.% hydrogen in R1 and R2

      

104 104 104 104 103 104 103

e 2.2 2.3 2.1 3.3 2.2 3.6

regions respectively (Fig. 7(e)). It is important to note that there are several instances of beam induced desorption of hydrogen from materials during the course of analysis by 1H(19F, ag)16O resonance reaction [17]. The films under investigation also experienced such a desorption which was ascertained by directly probing the R2 regions. The desorption, however, was not extensive and hydrogen concentrations in Fig. 7 have been duly corrected for the loss. 3.3. Optical and electrical characterization The pristine films deposited by Mo boat evaporation had a greenish tinge but were visibly transparent. There was no noticeable change in their physical appearance on vacuum annealing. The pristine films deposited by Ta boat evaporation, on the other hand, were brownish and became increasingly transparent with annealing temperature. The optical absorption characteristics of the films can be compared in detail from the absorption spectra of pristine and differently annealed Mo3 and Ta2 films shown in Fig. 8. The films prepared by Mo-boat evaporation have, in general, very low absorption above 450 nm with the average a in 450e750 nm region typically being ~3.0  103 cm1. An increase in annealing temperature causes marginal but a discernible blue shift. The direct band gaps of the pristine and differently vacuum annealed films calculated using Tauc's relationship (Equation (3)) are given in Table 1. It can be observed that Eg increases from ~3.2 eV in the pristine state to ~3.4 eV on annealing at 723 K. So far as the films produced by evaporation from Ta-boat are concerned, the pristine specimens exhibit high absorption in visible

Fig. 7. Depth profile of hydrogen in 390 nm films prepared by evaporation from Mo boat: (a) pristine and, (b) and (c) vacuum annealed at 523 K and 723 K respectively for 4 h, and the depth profile of hydrogen in 230 nm films prepared by evaporation from Ta boat: (d) pristine and (e) vacuum annealed at 723 K for 4 h.

170

P. Rao et al. / Materials Chemistry and Physics 149-150 (2015) 164e171

mixed homogeneously in a mortar before recording the XRD patterns (not shown). The analysis revealed that the residues contained Ga2O3, GaS and Ga2S3. The sulphides (GaS and Ga2S3) present in the evaporant in amorphous state apparently underwent crystallization during resistive heating. The occurrence of Ga2O3 in the residue can be attributed to the decomposition of the evaporant as per the following schemes:

Fig. 8. Absorbance spectra of 390 nm films prepared by evaporation from Mo boat: (a) pristine and (b), (c) and (d) vacuum annealed at 523 K, 623 K and 723 K respectively for 4 h, and 230 nm films prepared by evaporation from Ta boat: (e) pristine and (f), (g) and (h) vacuum annealed at 523 K, 623 K and 723 K respectively for 4 h.

region with the average a being >1.0  104 cm1. The high a values are manifestations of their large Ga content. As can be seen from Fig. 8 and Table 2, vacuum annealing at 723 K had a dramatic influence on the optical properties of the films. The films originally brownish in colour became transparent and acquired an Eg ~3.6 eV with average a in the visible region typically being ~103 cm1. It is to be noted that the pristine as well as vacuum annealed films prepared by Mo or Ta boat evaporation were insulating and exhibited electrical resistivity in 104e105 U cm range.

4. Discussion These studies clearly show the formation of GaS/Ga2S3 films notwithstanding GaOOH being one of the constituents of the evaporant. It is important to note that the content of oxygen in the present films is considerably lower, in comparison to the GaeS films prepared by modulated flux deposition technique [9]. These films were significantly rich in Ga (NGa:Ns ¼ 1.5:1), remained amorphous even at a substrate temperature of 723 K and contained, though not intentionally introduced, as high as 35 at.% oxygen. Even as the authors did not put forward any explanation, the occurrence of oxygen to such an extent is rather unusual given that the depositions were carried out at pressures better than 1  103 Pa. Assuming that the deposition system had certain residual oxygen, its inclusion in the films might have taken place as a result of the following reaction: 4GaS þ 3O2 ¼ 2Ga2O3 þ 4S

(4)

It is to be noted that GaS exhibits high reactivity towards oxygen, particularly at elevated temperatures. The recent report on the formation of Ga2O3 on annealing of GaS single crystals in 4N pure Ar gas (O2 content ~ 20 ppm) is a case in the point [19]. Presently, the films in general, bear compositions that are significantly different from that of the evaporant. Sulphur, present in the evaporant in relatively low proportions, is preferentially incorporated in the films. It is indicative of a substantial loss of oxygen during the process of deposition. To get an insight into the development of the films, the residues in Mo and Ta boat were examined by XRD. It is worth mentioning that the residues contained white, grey and yellowish powders. Since the apportioning of the residue in terms of colours was difficult, it was ground and

4GaOOH ¼ 2Ga2O3 þ 2H2 þ O2

(5)

2GaOOH ¼ Ga2O3 þ H2O

(6)

These reactions also explain the incorporation of hydrogen in the early stages of deposition process. Ga2O3 thus formed undergoes decomposition into substoichiometric gallium oxide (Ga2Ox) and oxygen with the former species being mainly responsible for the presence of oxygen bearing species in the films. Lieth et al. have reported that an equilibrium between solid GaS and GaS, Ga and S in the vapour phase is established when the compound is heated below its melting point (1235 K). The equilibria can be represented by the equations [20]: GaS(s) 4 GaS(g)

(7)

2GaS(s) 4 2Ga (g) þ S2(g)

(8)

The vapour pressure (p) of GaS (mm Hg) at a temperature T (K) is given by the relation:

23190 þ 19:39 log p ¼  T

(9)

The application of this equation shows that the vapour pressure of GaS at 1023 K, the approximate temperature of deposition, is substantially high (~7  102 Pa). Assuming that Ga2S3 (m.p. ~ 1528 K) also exhibits similar equilibria, it can be surmised that vapours of GaS, Ga2S3, Ga and S undergo condensation and or solid state reaction to produce gallium sulphide films. Even as these chemical reactions underline the basic processes involved in the development of the films, the significant difference in the NGa:NS values of the films prepared by Mo-boat and Ta-boat evaporation is intriguing and an explicit explanation for such an occurrence is lacking at the moment. However different constitution of the vapour phase, particularly with respect to sulphur, in the two approaches is presumably responsible for the dissimilar composition of the films. It was observed that the deposition from a Ta-boat sets in at ~1000 K while it commences at ~1123 K from a Mo-boat. It is well known that sulphur exists in polyatomic form (Sx (2e8)) with its precise composition varying with temperature. It exits predominantly as S6eS8 below 1073 K and as S2 at 1123 K. Therefore it can be assumed that the vapours emanating from a Ta boat are predominantly rich in S6eS8 and while those emerging from a Mo boat, in S2. The transportation and subsequently sticking (condensation) of S8 molecules may be sluggish and inefficient (due to higher mass) in comparison to S2 molecules. As a result, the films deposited from the Ta-boat are deficient in sulphur. These arguments are well supported by the colour of the films described in the previous (optical and electrical characterization) Section 3.3. The atomic ratio of Ga to S appears to have a direct bearing on the structural and optical properties of the films. Taking the cognizance of the RBS and XRD results it can be deduced that sulphur rich composition (NGa:NS ~ 1:1) favours the stabilization of g-Ga2S3 phase whereas sulphur deficient composition (NGa:NS > 1:1) leads to the formation of GaS phase with the bulk of Ga in excess of the composition of the phase formed diffusing into

P. Rao et al. / Materials Chemistry and Physics 149-150 (2015) 164e171

glass during vacuum annealing. It is well known that gallium oxide films deposited by thermal evaporation of Ga2O3 powder are significantly deficient in oxygen and have therefore very low transmittance. The high transmission in even pristine films deposited by evaporation from Mo-boat results from the introduction of sulphur and the extensive formation of GaS bonds. As pointed out by Morii et al. the short-range atomic arrangements in an amorphous phase is similar to those in a crystalline phase [18]. Similar situation prevails in Ta-boat evaporated films but the relatively higher content of Ga which may lead to the formation of mid gap states is probably responsible for their low transmittance. The substantial improvement in the transmission characteristics and increase in the band gap of Ta-boat evaporated films on vacuum annealing at 723 K are possibly due to the reason that the undiffused layers do not contain Ga in excess of that permitted by the stoichiometry of GaS phase. (Excess Ga is removed partly by diffusion and partly through evaporation). In fact, the presence of excess Ga is responsible for comparatively higher a values of films annealed at 623 K in spite of the prevalence of GaS. It is to be noted that the structural disorder that may affect the optical properties can also arise due to dangling bonds within the energy gap. Hydrogen is known to saturate the dangling bonds and reduce the associated absorption. Since these films, particularly in their pristine states, contain hydrogen in significant proportions only in the interiors, it is expected to have a limited influence on the optical properties of the films. Before concluding the discussion it is worthwhile mentioning that the presence of impurities does not affect the optical properties of the optical properties of Mo-boat evaporated films. This inference is based on the fact that Mo1, Mo2 and Mo3 films exhibit nearly identical optical characteristics. It is also supported by the results in reference [10] wherein the presence of as high as 10 at.% In does not seem to have a conclusive influence on the properties of GaeIneS films. Lastly, though the optical properties of even pristine Mo boat evaporated films make them amenable for buffer layers in photovoltaic devices, their conductivity needs to be improved for such applications. 5. Conclusions Thermal evaporation by resistively heating chemically synthesized powders consisting of gallium oxyhydroxide and gallium sulphide provides a simple method for preparing gallium sulphide films with traceable oxygen and hydrogen contaminations. The structural, compositional and optical properties of the films vary

171

with the boat material (Mo or Ta) used for evaporation. There exists a strong correlation between the optical properties and composition of the films. Amorphous in their pristine states, Mo-boat evaporated films are transparent while Ta-boat evaporated films are absorbing which are manifestations of their gallium to sulphur atomic ratio with the latter films being comparatively rich in gallium. Vacuum annealing at 723 K brings about significant modification in the optical properties of the films, in particular, Taevaporated films. It promotes the crystallization of Ga2S3 or GaS phases in the films and induces the diffusion of gallium into glass that makes the films bereft of excess gallium, imparting them improved transmission characteristics. The properties of as deposited or vacuum annealed films deposited from a particular boat remain largely invariant with thickness. Acknowledgements The authors thank Dr. B. N. Jagatap, Director, Chemistry Group, BARC and Dr. Sunil Jai Kumar, Head, NCCCM, Hyderabad for their comments and suggestions in improving the quality of the manuscript. References [1] H. He, R. Orlando, M.A. Blanco, R. Pandey, E. Amzallay, I. Baraille, M. Rerat, Phys. Rev. B 74 (2006) 195123e195128. [2] G. Shen, D. Chen, P.-C. Chen, C. Zhou, ACS Nano 3 (2009) 1115e1120. [3] M. Ohyama, H. Ito, M. Takeuchi, Jpn. J. Appl. Phys. 44 (2005) 4780e4783. [4] R.M.A. Lieth, F. Van der Maesen, Phys. Status. Solidi. A 10 (1972) 73e81. [5] S. Suh, D.M. Hoffman, Chem. Mater. 12 (2000) 2794e2797. [6] Q.S. Xin, S. Conrad, X.Y. Zhu, Appl. Phys. Lett. 69 (1996) 1244e1246. [7] T. Aono, K. Kase, A. Kinoshita, J. Appl. Phys. 74 (1993) 2818e2820. [8] P.J. Pernot, A.R. Barron, Chem. Vap. Depos. 1 (1999) 75e78. [9] C. Sanz, C. Guillen, M.T. Gutierrez, J. Phys. D Appl. Phys. 42 (2009) 085108e085113. [10] C. Sanz, C. Guillen, M.T. Gutierrez, Phys. Stat. Sol. 204 (2007) 3367e3372. [11] X. Chen, X. Hou, X. Cao, X. Ding, L. Chen, G. Zhao, Wang X. Wang, J. Cryst. Growth 173 (1997) 51e56. [12] H. Yamada, K. Ueno, A. Koma, Jpn. J. Appl. Phys. 25 (1996) L568eL570. [13] S. Chowdhury, M. Ichimura, Jpn. J. Appl. Phys. 49 (2010) 062302e062304. [14] S. Chowdhury, M. Ichimura, Jpn. J. Appl. Phys. 48 (2009) 061101e061104. [15] G. Micocci, R. Rella, A. Tepore, Thin Solid Films 172 (1989) 179e183. [16] J.A. Leavitt Jr., L.C. McIntyre, M.D. Ashbaugh, J.G. Oder, Z. Lin, B. DezfoulyArjomandy, Nucl. Instr. Methods B 44 (1990) 260e265. [17] W.A. Lanford, Nucl. Instr. Methods B 66 (1992) 65e82. [18] K. Morii, H. Ikeda, Y. Nakayama, Mat. Lett. 17 (1993) 274e281. [19] E. Filoppo, T. Siciliano, A. Genga, G. Micocci, M. Siciliano, A. Tepore, Appl. Sur. Sci. 261 (2012) 454e457. [20] R.M.A. Lieth, H.J.M. Heijligers, C.W.M. Heijden, Mater. Sci. Eng. 2 (1967) 193e200.