D.O. Dumcencoa, H.P. Hsua, Y.S. Huanga,∗, C.H. Liangb, K.K. Tiongb, C.H. Duc
aDepartment of Electronic Engineering, National Taiwan University of Science and Technology, Taipei 106, Taiwan
bDepartment of Electrical Engineering, National Taiwan Ocean University, Keelung 202, Taiwan
cDepartment of Physics, Tamkang University, Tamsui 251, Taiwan
a r t i c l e i n f o
Article history:
Received 7 December 2007 Received in revised form 1 April 2008 Accepted 30 April 2008
Single crystals of WS2doped with gold have been grown by the chemical vapour transport method using iodine as a transporting agent. X-ray diffraction (XRD) pattern analysis revealed presence of mixed three-layer rhombohedral (3R) and two-three-layer hexagonal (2H) polytypes for the doped crystals while the undoped one shows only 2H form. Hall measurements indicate that the samples are p-type in nature. The dop-ing effects of the materials are characterized by surface photovoltage (SPV), photoconductivity (PC) and piezoreflectance (PzR) measurements. Room temperature SPV and PC spectra reveal a feature located at∼60 meV below the A exciton and has been tentatively assigned to be an impurity level caused by Au dopant. Excitonic transition energies of the A, B, d and C excitons detected in PzR spectra show red shift due to the presence of a small amount of Au and the broadening parameters of the excitonic transi-tion features increase due to impurity scattering. The values of the parameters that describe the electron (exciton)–phonon interaction of excitonic transitions of A–B are about two times larger than that of d–C excitonic pairs. The possible assignments of the different origins of A–B and d–C excitonic pairs have been discussed.
© 2008 Elsevier B.V. All rights reserved.
1. Introduction
The tungsten disulphide WS2semiconductor belongs to transi-tion metal dichalcogenides MX2(M = Mo, W; X = Se, S) that exhibit many physical properties with a pronounced two-dimensional character[1–5]. The peculiar properties of these materials result from their layered structure, consisting of mostly covalently bonded X–M–X sheets linked by weak Van der Waals (VdW) forces. The lay-ered transition metal dichalcogenides MX2 semiconductors have been extensively investigated due to the potential application of the materials as solid lubricants, catalysts and photovoltaic solar cell materials[6–12]. Particularly WS2 was the subject of great interest because its band gap is well matched to the solar spec-trum. Like the other transition metal dichalcogenide compounds, WS2is an indirect-gap semiconductor with the top of the valence band situated at point of the Brillouin zone and the bottom of the conduction band located at halfway between and K points while the direct gap is placed at K point[13–16]. The main advan-tage of WS2is the photocorrosion resistance of the nonbonding d–d orbitals of W atoms involved in the phototransitions[17], WS2can act as an efficient photoconductive layer in photovoltaic devices and photoelectrochemical solar cells[18–20]. Owing to these potential
∗ Corresponding author. Tel.: +886 2 27376385; fax: +886 2 27376424.
E-mail address:[email protected](Y.S. Huang).
benefits, the main investigations of MX2compounds deal with the strong anisotropy of physical properties[5,21], the role of the d orbital of the transition metal atom in the electronic band struc-ture[22]and the sharp excitonic structures in the visible and IR region[23–25].
The anisotropic mechanical and electrical properties of the material are a result of the sandwich interlayer structure, loosely bonded by the weak VdW force. The intralayer bonding is thought to be partly ionic and partly covalent, with the latter being domi-nant. There are two known polytypes of WS2: two-layer hexagonal and three-layer rhombohedral, termed 2H and 3R, respectively [1,26,27]. The 2H polytype has two layers per unit cell stacked in hexagonal symmetry and belongs to the space group D46h, while the 3R form has three layers in rhombohedral symmetry and belongs to the space group C3v5. Earlier studies[28]suggested that natural rhombohedral MX2is consistently rich in certain minor elements for example Re, Nb, Ti, Zr and Fe, and that the incorporation of such impurity elements has predetermined the adoption of the lower 3R symmetry for MX2. However, only few works concerning the effect of the dopants on the physical properties of WS2have been reported[26,29,30].
In this work, we report the characterization of Au-doped WS2 single crystals grown by chemical vapour transport method with I2 as transport agent. The crystal structure was analyzed using XRD patterns. The doping effects of the materials were charac-terized by surface photovoltage (SPV), photoconductivity (PC) and
0254-0584/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.matchemphys.2008.04.052
Au-doped WS2 to detect the possible distinction in the spectral features of the excitonic transitions, namely A, B, d, and C. The parameters which describe the temperature behaviour of excitonic transitions indicate that A–B and d–C excitonic pairs, caused by interlayer interaction and spin–orbit splitting, correspond to exci-tonic transitions with different origins. The effects of dopant on the optical properties are analyzed and discussed.
2. Experimental
Single crystals of WS2:Au were grown by the chemical vapour transport method using I2as a transport agent. The weight of doping material was determined stoi-chiometrically to obtain the small concentration of dopant, i.e. 50 mg of Au on about 10 g of the total charge weight. Single crystalline platelets up to 10 mm× 5 mm sur-face area and 0.5 mm thickness were obtained. Hall-effect measurements indicate that the Au-doped sample is p-type in nature. At room temperature, the carrier concentration is 3× 1018cm−3and the Hall mobility corresponds to 18 cm2V−1s−1. The crystal structure of undoped and Au-doped WS2single crystals was analyzed using Rigaku RTP300RC X-ray diffractometer with Cu K␣ radiation ( = 1.5418 ˚A).
For X-ray studies, several small crystals were finely ground with a mixture of glass powder, and the X-ray powder patterns were recorded by means of a slow-moving radiation. Cu K␣ radiation was employed and a silicon standard was used to calibrate the diffractometer. The lattice parameters were calculated with the aid of a computer using a least squares refinement program.
In SPV measurements, the photovoltage was measured between the sample and a reference metal grid electrode in a capacitive manner as a function of the photon energy of the probe beam. The illumination system consisted of a 150 W tungsten–halogen lamp chopped at 200 Hz and a 0.25 m grating monochromator.
A beam splitter was placed in the path of the incident light. The intensity of the incident beam was monitored by a pyroelectric detector and was maintained at a constant level of∼10−5W cm−2by a stepping motor connected to a variable neutral density filter, which was also placed in the path of the incident beam. The photo-voltage spectrum on the metal grid was measured with a copper plate as the ground electrode, using a buffer circuit and a lock-in amplifier. For the PC measurements, selected samples were cut into rectangular shape. The PC spectra were recorded on a standard experimental set-up with chopped light beam incident onto the basal plane of the crystals. Good ohmic contacts to the crystal were facilitated by means of two gold wires laid across the VdW surface of the crystal and attached to the crystal surface by highly conducting silver epoxy. Phase sensitive detection technique was used to measure the PC signals at room temperature.
The PzR measurements were achieved by gluing the thin single-crystal speci-mens on a 0.15 cm thick lead zirconate titanate piezoelectric transducer driven by a 200 Vrmssinusoidal wave at 200 Hz. A 150 W xenon arc lamp filtered by a 0.25 m grating monochromator provided the monochromatic light. The reflected light was detected by an UV-enhanced silicon photodiode, and the signal was recorded from a lock-in amplifier. A closed-circle cryogenic refrigerator equipped with a digital thermometer controller was used for low-temperature measurements. The mea-surements were made between 15 and 300 K with a temperature stability of 0.5 K or better.
3. Results and discussion
Fig. 1shows the X-ray powder diffraction patterns for undoped and Au-doped WS2samples, respectively. The XRD pattern of Au-doped WS2 crystals (Fig. 1(b)) differs from the undoped WS2 (Fig. 1(a)). The peak positions and intensities of the XRD pattern of undoped WS2crystals correspond to the hexagonal phase while for WS2:Au the XRD patterns show a mixture of hexagonal (marked with an H upon the peaks associated with 2H polytype inFig. 1(b)) and rhombohedral phases. It should be noted that for 3R-polytype a lattice parameter (3.1581 ˚A) is closed to the 2H-polytypic one (a = 3.154 ˚A) while the c lattice parameter of magnitude 18.491 ˚A is about 1.5 larger than the hexagonal c value (12.323 ˚A)[4]. The change in relative line intensities is the result of a different crys-tal quality of 2H-WS2to that of the mixed 3R- and 2H-polytypes WS2:Au single crystals. The rocking curves for the features of 3R-(0 1 5) and 3R-(0 1 8) illustrated inFig. 1(c) show a full width at half maximum (FWHM) of 0.85◦and 0.33◦, respectively. Large FWHM values may also be related to the formation of mixed 3R- and 2H-polytypes of the gold-doped WS2crystals.
Fig. 1. X-ray diffraction patterns of (a) undoped and (b) Au-doped WS2. The XRD pat-tern of undoped WS2crystals correspond to the hexagonal phase while for WS2:Au the XRD patterns show a mixture of hexagonal (marked with an H upon the peaks associated with 2H polytype) and rhombohedral phases. (c) The full width at half maximum (FWHM) of the 3R- (0 1 5) and (0 1 8) features determined from rocking curves.
Fig. 2shows the PzR, SPV and PC spectra in the vicinity of band edge at room temperature. The first-derivative PzR spectrum indi-cates the presence of a well known excitonic feature A at∼1.96 eV near the direct band gap Edg of WS2:Au as well as interference induced oscillations on the lower energy side (below 1.9 eV). A band-gap-related feature A at∼1.96 eV has also been observed in SPV and PC spectra. In the sub-band-gap energy range for both SPV and PC spectra, an additional feature X located at ∼1.9 eV
Fig. 2. Piezoreflectance, surface photovoltage and photoconductivity spectra of Au-doped WS2at room temperature. The arrows show the position of band-gap-related future A and additional feature X.
Fig. 3. Piezoreflectance spectra of (a) undoped and (b) Au-doped WS2at several temperatures between 15 and 300 K. The dashed curves are the experimental results and the solid curves are least squares fits of Eq.(1).
is observed. This feature is not detected in the PzR spectrum.
From the previous study of single crystals using PC measure-ments[31–33], the photoresponse of undoped WS2sample sharply increases starting at about 1.8 eV without the appearance any addi-tional pronounce feature below the A excitonic peak. Moreover, it is also known that the SPV signal is proportional to the absorption coefficient ˛ and previous absorption measurements[1,31–33]on pure WS2indicated absence of such feature. Thus the extra sub-band-gap feature at∼60 meV below exciton A is most likely related to the impurity state as a result of Au doping of WS2.
The experimental PzR spectra showing the prominent excitonic features over the range 1.8–3.0 eV for the undoped and Au-doped WS2single crystals at several representative temperatures between 15 and 300 K are displayed inFig. 3. The spectra are characterized by four excitonic transitions marked by A, B, d, and C. As the
temper-ature increases, the PzR fetemper-atures shift towards lower energies and broaden. In the case of undoped WS2(Fig. 3(a)), the second feature of the A exciton sequence and an adjacent resonance feature above the A series denoted as A2and AR, respectively, are also detected.
In order to determine the position of the transitions accurately, we performed a theoretical line shape fitting for the first-derivative functional Lorentzian line shape form appropriated for the bounded states, such as excitons or impurity transitions[34].
R R = Re
n
i=1
Aexi exp(jϕexi )(E − Eexi + jiex)−2
, (1)
where Aexi and ϕexi are the amplitude and phase of the line shape, Eexi and iexare the energy and broadening parameters of the interband excitonic transitions, respectively. The fits shown as solid curves in
Fig. 4. Temperature variations of the energies of (a) the A–B and (b) d–C excitonic pairs for undoped and Au-doped WS2. Representative error bars are shown. The dashed and solid curves are least squares fits to Eqs.(2) and (3), respectively.
Re-doped WS2 following notations of Wilson and Yoffe[1]and Beal and Knights [2]. For undoped 2H-WS2, the Rydberg series of A exciton can be described by three-dimensional Mott–Wannier exciton model[35]
which predicts that the intensity of the discrete excitonic transi-tion falls off as n−3. The direct band gap Egand the binding energy of the A exciton series for pure WS2 at 15 K are estimated to be 2.119± 0.002 eV and 52 ± 2 meV, respectively[36]. For Au-doped WS2, only the ground state (n = 1) of the A exciton is observed.
The excitonic series can be more appropriately described by the two-dimensional Mott–Wannier excitons model with the intensity fallen as (2n− 1)−3[35]. The much broader linewidth of the ground state of the A exciton and a more severe intensity degradation pre-vent the observation of the higher series. It is also noted that the red-shift rate of A and B excitons is higher than that of d and C excitons as temperature increased.
The temperature variations of the energies of A, B, d and C exci-tons for undoped and Au-doped WS2are shown inFig. 4. The dashed curves are the least squares fits to the Varshni empirical relation-ship[37].
Ei(T) = Ei(0)− ˛iT2
ˇi+ T, (2)
where i = A, B, d or C, Ei(0) is the excitonic transition energy at 0 K and, ˛iand ˇiare constants referred to as the Varshni coefficients.
The constant ˛iis related to the electron (exciton)–phonon interac-tion and ˇiis closely related to the Debye temperature. The obtained values of Ei(0), ˛iand ˇifor the excitonic transition of undoped and Au-doped WS2are listed inTable 1. The values of the A and B exci-tonic transitions for Re-doped WS2are also listed for comparison [36]. The Debye temperature Dof WS2has been estimated to be 210 K using the Lindemann formula[38]. The fitted values of ˇiare in reasonable agreement with the estimated D.
The temperature variations of the energies of A, B, d and C exci-tons for undoped and Au-doped WS2are also fitted (solid curves) by an expression containing the Bose–Einstein occupation factor for the phonon inFig. 4 [39,40].
Ei(T) = EiB(0)− aiB
where i = A, B, d or C, aiB represents the strength of the electron (exciton)–phonon interaction and iBcorresponds to the average phonon temperature. The fitted values for EiB(0), aiBand iB are given inTable 1. For comparison purposes, we have also included corresponding values for WS2:Re[36]. High-temperature limits of Eqs.(2) and (3)yield the relation ˛i≈ 2aiB/iB which is satisfied
within error bars by the fitted numbers for ˛i, aiBand iBas listed inTable 1. The values of ˛iand aiBwhich are related to the electron (exciton)–phonon interaction for A–B are about two times larger than that of d–C excitonic pairs. These results indicate that A–B and d–C excitonic pairs correspond to excitonic transitions with differ-ent origins. From the recdiffer-ent theoretical and experimdiffer-ental studies [14–16], the A and B peaks near the absorption edge correspond to the smallest direct transitions at the K point while the d and C excitons are attributed to the direct transitions of the point of the Brillouin zone. The A and B excitons belong respectively to K4→ K5
and K1→ K5optical transitions originating from interlayer interac-tions and spin–orbit splitting of the valance band at the K point. The K states have been shown to be predominantly metal d states with a small contribution from the nonmetal p states. According to the band structure calculations, in the absence of spin–orbit interaction
4−→ 5+ is the only allowed transition from top of the valance band to conduction band[15,16]. The 5+→ 9++ 8+spin–orbit splitting of the conduction band level due to the metal d and non-metal p contributions to the wave functions have different signs, and the metal d contribution to these states is rather small. As a result, the d and C exciton peaks can be prescribed, respectively, as
4−→ 8+and 4−→ 9+optical transitions.
The experimental values of i(T) of the A and B excitons at several temperatures between 15 and 300 K as obtained from the line shape fit with Eq.(1)for Au-doped and undoped WS2are dis-played inFig. 5. The temperature dependence of the broadening parameters of semiconductors can be expressed as[39,40]:
i(T) = i0+ iLO
exp(iLO/T) − 1, (4)
where i = A or B. The first term iLOof Eq.(4)represent the broaden-ing contribution of temperature independent mechanisms, such as electron–electron interaction, impurity, dislocation, electron inter-action and alloy scattering, whereas the second term is caused by the electron (exciton)–LO phonon (Fr ¨ohlich) interaction. The quantity iLOrepresents the strength of the electron (exciton)–LO phonon coupling while iLOis the LO phonon temperature[39,40].
The solid curves inFig. 5are least squares fits to Eq.(4), which made it possible to evaluate i0, iLO and iLO for the A and B excitonic transitions of Au-doped and undoped WS2. The obtained values are presented inTable 2together with the number for Re-doped WS2[36]. The larger value of i0for the doped (Au or Re) WS2compounds is mainly due to impurity scattering. According to the previous report of Raman measurements[41], the fitted values of iLOare quite close to the longitudinal optical phonon tempera-tures for WS2(515 K). This close match between the values of iLO
and the LO phonon temperatures obtained from Raman
scatter-Fig. 5. Temperature variations of the broadening parameters of the A and B excitonic transitions for undoped and A-doped WS2. Representative error bars are shown. The solid curves are least squares fits to Eq.(4).
Table 2
Values of the parameters which describe the temperature dependence of the broad-ening function of the excitonic transitions for undoped, Au-doped WS2and Re-doped WS2
ing indicates that the temperature variation of iis indeed mainly due to the interaction of electron with the optical phonons. The
iLOvalue is significantly larger than value of iBthat is in a good agreement with existing theory[39,40].
4. Summary
Single crystals of Au-doped WS2 with surface area up to 10 mm× 5 mm surface area and 0.5 mm thickness have been grown by the chemical vapour transport method using iodine as a transport agent. The X-ray investigation indicates that WS2:Au crys-tallized in mixture of 2H- and 3R-polytypes structure instead of the pure 2H-polytype crystallization of undoped WS2. The presence of Au impurity has been determined to play a major role in influ-encing the optical properties of the doped sample. In SPV and PC
caused by Au dopant. The calculation of temperature-dependent experimental PzR data confirms theoretical results that A–B and d–C excitonic pairs have different origin of transitions. The direct band-edge excitonic transition energies show a slight red shift due to the presence of small amount of Au, and the broadening param-eter of the excitonic transition features increases due to impurity scattering.
Acknowledgement
The authors would like to acknowledge the financial supports by the National Science Council of Taiwan under Grant Nos. NSC96-2112-M-011-001, NSC96-2221-E-011-030 and NSC95-2112-M-019-001.
References
[1] J.A. Wilson, A.D. Yoffe, Adv. Phys. 18 (1969) 193.
[2] A.R. Beal, J.C. Knights, J. Phys. C: Solid State Phys. 5 (1972) 3540.
[3] W.Y. Liang, J. Phys. C: Solid State Phys. 6 (1973) 551.
[4] W.J. Schutte, J.L. De Boer, F. Jellinek, J. Solid State Chem. 70 (1987) 207.
[5] S.Y. Hu, M.C. Cheng, K.K. Tiong, Y.S. Huang, J. Phys.: Condens. Matter 17 (2005) 3575.
[6] G. Alonso, M. Del Valle, J. Cruzb, A. Licea-Claverie, V. Petranovskii, S. Fuentes, Catal. Lett. 52 (1998) 55.
[7] P.D. Fleischauer, Thin Solid Films 154 (1987) 309.
[8] T. Spalvins, Thin Solid Films 6 (1982) 17.
[9] J.M. Martin, C. Donnet, J. Le Mogne, T. Epicier, Phys. Rev. B 48 (1993) 10583.
[10] S.J. Li, J.C. Bern `ede, J. Pouzet, M. Jamali, J. Phys.: Condens. Matter 8 (1996) 2291.
[11] K.K. Kam, B.A. Parkinson, J. Phys. Chem. 86 (1982) 463.
[12] A. Jager-Waldau, M. Lux-Steiner, R. Jager-Waldau, R. Burkhardt, E. Bucher, Thin Solid Films 189 (1990) 339.
[13] K. Albe, A. Klein, Phys. Rev. B 66 (2002) 073413.
[14] A. Klein, S. Tiefenbacher, V. Eyert, C. Pettenkofer, W. Jaegermann, Phys. Rev. B 64 (2001) 205416.
[15] R. Coehoorn, C. Hass, J. Dijkstra, C.J. Flipse, R.A. de Groot, A. Wold, Phys. Rev. B 35 (1987) 6195.
[16] R. Coehoorn, C. Hass, R.A. de Groot, Phys. Rev. B 35 (1987) 6203.
[17] H. Tributsch, Sol. Energy Mater. 1 (1979) 257.
[18] J.A. Baglio, G.S. Calabrese, E. Kamieniecki, R. Kershaw, C.P. Kubiak, A.J. Ricco, A.
Wold, M.S. Wrighton, G.D. Zoski, J. Electrochem. Soc. 129 (1982) 1461.
[19] J.A. Baglio, G.S. Calabrese, D.J. Harrison, E. Kamieniecki, A.J. Ricco, M.S. Wrighton, G.D. Zoski, J. Am. Chem. Soc. 105 (1983) 2246.
[20] Y. Santiago-Ortiz, G.I. Torres, A. D´ıaz, C.R. Cabrera, J. Electrochem. Soc. 142 (1995) 2770.
[21] J.W. Chung, Z.R. Dai, K. Adib, F.S. Ohuchi, Thin Solid Films 335 (1998) 106.
[22] L.F. Mattheiss, Phys. Rev. B 8 (1973) 3719.
[23] E. Fortin, F. Raga, Phys. Rev. B 11 (1975) 905.
[24] L. Kulyuk, D. Dumcehnko, E. Bucher, K. Friemelt, O. Schenker, L. Charron, E.
Fortin, T. Dumouchel, Phys. Rev. B 72 (2005) 075336.
[25] L. Charron, D. Dumchenko, E. Fortin, C. Gherman, L. Kulyuk, J. Lumin. 112 (2005) 45.
[26] R.H. Friend, A.D. Yoffe, Adv. Phys. 36 (1987) 2.
[27] W. Jaegermann, H. Tributsch, Prog. Surf. Sci. 29 (1988) 1.
[28] R.J. Traill, Can. Mineral. 7 (1963) 524.
[29] W.Y. Liang, Intercalation in Layered Materials, Plenum Press, New York, 1986.
[30] P.C. Yen, Y.S. Huang, K.K. Tiong, J. Phys.: Condens. Matter 16 (2004) 2171.
[31] C. Ballif, M. Regula, P.E. Schmid, M. Remskar, R. Sanjines, F. Levy, Appl. Phys. A 62 (1996) 543.
[32] C. Ballif, M. Regula, F. Levy, Sol. Energy Mater. Sol. Cells 57 (1999) 189.
[33] E. Gourmelon, O. Lignier, H. Hadouda, G. Couturier, J.C. Bernede, J. Tedd, J. Pouzet, J. Salardenne, Sol. Energy Mater. Sol. Cells 46 (1997) 11.
[34] F.H. Pollak, H. Shen, Mater. Sci. Eng. R 10 (1993) 275.
[35] A.R. Beal, W.Y. Liang, H.P. Hughes, J. Phys. C: Solid State Phys. 9 (1976) 2449.
[36] P.C. Yen, H.P. Hsu, Y.T. Liu, Y.S. Huang, K.K. Tiong, J. Phys.: Condens. Matter 16 (2004) 6995.
[37] Y.P. Varshni, Physica 34 (1967) 149.
[38] C.H. Ho, P.C. Liao, Y.S. Huang, K.K. Tiong, Phys. Rev. B 55 (1997) 15608.
[39] P. Lantenschlager, M. Garriga, S. Logothetidis, M. Cardona, Phys. Rev. B 35 (1987) 917.
[40] P. Lantenschlager, M. Garriga, L. Vi ˇna, M. Cardona, Phys. Rev. B 36 (1987) 4821.
[41] S.I. Ushida, S. Tanaka, J. Phys. Soc. Jpn. 45 (1978) 153.