J O U R N A L O F M AT E R I A L S S C I E N C E : M AT E R I A L S I N E L E C T RO N I C S 1 2 ( 2 0 0 1 ) 4 7 3 ± 4 7 6
Characterization of p-In2Se3 thin ®lms A. F. QASRAWI Faculty of Engineering, Atõlõm University, 06836, Ankara, Turkey È NAL M. PARLAK, CË . ERCË ELEBI, I. GU Department of Physics, Middle East Technical University, 06531, Ankara, Turkey E-mail:
[email protected] Indium selenide thin ®lms were deposited onto glass substrates kept at 150 C by thermal evaporation of a-In2 Se3 . Some of the ®lms were annealed at 150 C and 200 C and they all were found to exhibit p-type conductivity without intentional doping. Scanning electron microscopy (SEM) established that the ®lms have an atomic content of In51 Se49 . X-ray diffraction (XRD) indicated that the as-grown ®lms were amorphous in nature and became polycrystalline b-In2 Se3 ®lms after annealing. The analysis of conductivity temperaturedependence measurements in the range 320±100 K revealed that thermal excitation and thermionic emission of the carriers are the predominant conduction mechanisms above 200 K in the amorphous and polycrystalline samples, respectively. The carrier transport below 200 K is due to variable range hopping in all the samples. Hall measurements revealed that the mobility of the polycrystalline ®lms is limited by the scattering of the charged carriers through the grain boundaries above 200 K. # 2001 Kluwer Academic Publishers 1. Introduction
VI In2 Se3 is a semiconducting compound of the AIII 2 B3 family, the structures of which are defective with respect to their metal atoms: only two thirds of the sites in the cation sublattice are occupied. In2 Se3 has been shown to be a material with attractive properties for applications in electrochemical and photovoltaic devices [1, 2]. The physical properties of n-InSe thin ®lm have been investigated by several researchers [3±7]. However, no signi®cant attention has yet been devoted to the study of the properties of p-In2 Se3 thin ®lms. The conduction type and the properties of InSe thin ®lms are very sensitive to the deposition technique, evaporation source material and temperature. These parameters yield different phases of InSe, such as In2 Se3 , and In6 Se7 [8]. Although the polycrystalline InSe thin ®lms usually show n-type conduction and amorphous ®lms are p-type, there is a discrepancy in the literature on the majority carrier type of the InSe ®lms even of the same composition [3]. In this study, the structural and electrical properties of undoped p-In2 Se3 thin ®lms produced by a thermal evaporation technique from a-In2 Se3 lumps, have been investigated.
2. Experimental Procedure
InSe thin ®lms have been deposited onto glass substrates by the thermal evaporation of a-In2 Se3 kept at 800 C. The evaporation cycle was done in a Leybold vacuum system at a pressure of about 1:3610 3 Pa. The ®lms were evaporated in Hall bar (six-arm bridge), Maltese cross and square shapes at a substrate temperature
Ts of 150 C. 0957±4522
# 2001 Kluwer Academic Publishers
After deposition, some of the ®lms were annealed at Ta 150 C and 200 C for 1 h under nitrogen ¯ow to examine the in¯uence of annealing on the properties of the ®lms. The details of the structural and electrical characterization techniques of the ®lms were described elsewhere [4±6].
3. Results and discussion
The X-ray diffraction patterns have indicated that the evaporation source consists of a-In2 Se3 and InSe phases. The unannealed ®lms (B1) are amorphous in structure. The annealed samples B2 and B3 show intense peaks of (0 0 6)-oriented polycrystalline b-In2 Se3 as illustrated in Fig. 1. The other regular peaks of b-In2 Se3 observed at (1 1 8), (2 2 5), and (3 1 5) are weak when compared to the (0 0 6) re¯ection peak intensity. The XRD patterns improve as the annealing temperature is increased to 200 C, indicating better crystallinity with annealing. SEM and EDXA analysis indicated that the ®lms are composed of In and Se only and no other impurities are detected. To examine the homogeneity of the ®lms, different areas for each ®lm were considered and the composition was found to be almost the same. The atomic percentages were found to be * 51% indium and * 49% selenium regardless of annealing temperature. All the deposited ®lms showed p-type conduction. In our previous investigations [4±6], we have reported that when the source material was e-modi®ed InSe
Tsource 650 C, the ®lms grown at the same substrate temperatures were polycrystalline InSe, composed of 65% In and 35% Se exhibiting n-type conductivity. In this study, we have found that the ®lms deposited by 473
Figure 2 Variation of conductivity with temperature for all samples. p The inset shows the ln
s T T 1 plots for the polycrystalline samples in the high-temperature region.
Figure 1 X-ray diffraction patterns of as-grown and annealed In2 Se3 thin ®lms.
using a-In2 Se3 lumps kept at Tsource 800 C are amorphous in structure, the annealed ones are polycrystalline and are all composed of 51% In and 49% Se exhibiting p-type conduction. Marsillac et al. [7] have reported that the Inx Se100 x thin ®lms obtained by the thermal evaporation technique exhibited p-type conduction for compositions of 45±60% Se and n- type conduction for composition below 40% Se. The values of atomic percentages reported in this study are in the indicated range for p-type conduction. The room-temperature electrical conductivity values sRT listed in Table I were found to increase with increasing annealing temperature. The temperaturedependent conductivity illustrated in Fig. 2 was measured in the range of 100±320 K. Below 100 K, the sample resistance exceeds 1010 O and electrical measurements could not be carried out due to instrumental limitations. In general, the conductivity decreases exponentially with decreasing temperature at higher rates above 200 K, below which the conductivity variation with temperature becomes less dominant. The different characteristics of the s-T variations in different
temperature regions imply the existence of different conduction mechanisms. To deduce the dominant conduction mechanism, s-T data for the as-grown (annealed) samples are analyzed using thermal excitation (thermionic emission), thermally assisted tunneling and hopping models. The details of the analysis are given below; The conductivity temperature dependence of the asgrown amorphous ®lms (B1) are ®rst analyzed according to the conductivity expression due to the excitation of carriers into the extended states beyond the mobility shoulders [9±11] Es s s0 exp
1 kT where s0 is the pre-exponential factor and Es is the conductivity activation energy. As shown in Fig. 2, the ln
s T 1 variation is linear above 200 K with an activation energy value of 224 meV. Below 200 K ln
s T 1 plots are not linear, indicating the temperature dependence of the activation energy. In the temperature range 320±200 K conductivity measurements ®t to (1), with an activation energy less than 1/2 Eg , which implies that not the intrinsic but tail states conduction is the dominant conduction mechanism above 200 K [9]. Below 200 K the non-linear behavior
T A B L E I Room temperature conductivity sRT , carrier concentration pRT , Hall mobility mRT , conductivity activation energy Es, carrier concentration activation energy Ep grain boundary barrier height Fb and Mott's parameters T0 , g, R, W and N
EF Sample sRT
O 1 cm 1 610 pRT
cm 3 61014 mRT
cm2 V 1 s 1 Es (meV) (T 4 200 K) Ep (meV) F0 (meV) T 4 200 K T 5 200 K T0 6106 (K) g6106
cm 1 R
130 K610 7
cm W (130 K) (meV) gR N
EF
cm 3 eV 1
474
4
B1 (as-grown)
B2 (Ta 150 C)
B3 (Ta 200 C)
1.7 Ð Ð 224 Ð Ð Ð 9.8 36.3 1.7 47 5.6 1:061021
3.2 1.2 17 259 159 89 14 2.6 1.35 33.0 33 4.4 2:061017
15.0 16.0 6 263 163 94 10 1.9 0.75 54.8 30 4.1 4:761016
of ln
s T 1 variation, with temperature-dependent activation energy that decreases with decreasing temperature, rules out the validity of the same kind of transport mechanism described by (1). For the polycrystalline samples (B2, B3), if the transport mechanism is dominated by the thermionic emission, the conductivity is expressed [12] by p Es s T s1 exp
2 kT where ps1 is the pre-exponential factor. The plots of ln
s T T 1 for the polycrystalline samples showed linear variations with two different slopes above and below 200 K. Values of Es calculated from the slopes of linear ®ts (see the inset of Fig. 2) above 200 K are tabulated in Table I. The activation energies are found to be insensitive to annealing temperature within an error of + 5 meV. Above 200 K the conductivity ®ts to the relation given by (2) with activation energies larger than kT, implying that the thermionic emission over the grain boundary potential proposed by Seto [12] is the dominant conduction. Below 200 K the weak temperature dependence of the conductivity with very low activation energy values indicates the possibility of a transport mechanism other than thermally activated conduction. To deduce the dominant conduction mechanism below 200 K, the conductivity±temperature data were also analyzed according to other possible models, namely, thermally assisted tunneling and hopping. If the conduction is due to the thermally assisted tunneling of carriers, the conductivity is expressed by F2 2 s s2 1 T
3 6 where s2 and F are constants [13]. For the samples in this study, the tunneling transport mechanism is not identi®ed to be adequate due to the non-linearity of s versus T 2 plots. The conduction mechanisms below 200 K for both amorphous and polycrystalline In2 Se3 thin ®lms were analyzed according to Mott's hopping mechanism [11]. The conductivity with the correction for the temperature dependence of the effective density of states [14] is p T0 14 s T s3 exp
4 T
the variations exhibited linear behavior in the temperature range 100±190 K, as illustrated in Fig. 3. Mott parameters; the average hopping distance
R and energy
W, respectively, were calculated using the related expressions [11, 14, 15] and tabulated in Table I together with T0 and N
EF values. The degree of disorder T0 and the density of the localized states N
EF decrease as the samples were annealed, predicting a better crystallinity of the ®lms, which is consistent with X-ray results. The values of gR and W satisfy Mott's requirements
gR 4 1; W 4 kT for variable-range hopping. Thus, below 200 K, variablerange hopping conduction is probably the transport mechanism in both amorphous and polycrystalline In2 Se3 thin ®lms. The temperature dependence of the Hall mobility is measured by the standard d.c. Hall measurement technique for the annealed samples. The room-temperature carrier concentration, pRT , calculated from the relation p 1=eRH (RH is the measured Hall coef®cient), was found to increase with increasing annealing temperature (see Table I). The increase in the carrier concentration with increasing annealing temperature is accompanied by a decrease in the Hall mobility, calculated from the relation m sRH . This similar behavior of Hall parameters was also observed for In2 Se3 thin ®lms grown by the elemental evaporation technique and it was attributed to the reduction in the disorder reducing the loss of carriers at the grain boundaries [8, 16]. The temperature dependence of the Hall mobility and carrier concentration was found to be similar to that of the conductivity, both decreasing exponentially with decreasing temperature. Thus, the m T variations have been analyzed using the model proposed by Petritz [17] and later extended by Seto [12] in which the effective mobility is given by Fb m m0 exp
7 kT Here, Fb is the potential barrier height at the grain boundaries of the polycrystalline ®lm and m0 is given by qd m0 p 2pm* kT
8
in which d is the average grain size and m* is the
where T0 is the degree of disorder and s3 is the preexponential factor given by T0
lg3 kN
EF
5
and s3 e2 R2 nph N
EF
6
respectively where, g is the inverse of the localization length, N
EF is the density of localized states at the Fermi level, l is a dimensionless constant and nph is the phonon frequency (1012 Hz, typically). In the low-temperature region,
sT 1=2 was plotted as a function of T 1=4 on a semi-logarithmic scale and
p Figure 3 ln
s T temperature region.
T
1=4
variations for all the samples in the low-
475
higher temperatures, current transport is mainly due to the thermal excitation in amorphous ®lms and thermionic emission in the polycrystalline ®lms. As-grown ®lms exhibited no Hall mobility due to disordered structure. In polycrystalline samples, the Hall mobility was found to be limited by the scattering related to the intergrain barriers of about 90 meV.
References 1.
2. Figure 4 ln
mT 1=2
T
1
variations for the polycrystalline samples.
effective mass. The plots of ln
mT 1=2 T 1 illustrated in Fig. 4 are linear above 200 K, giving a barrier height of 89 meV for sample B2 and 94 meV for sample B3. The carrier concentration activation energies
Ep calculated from the slopes of ln
p T 1 are given in Table I. The values of Fb and Ep satisfy the relation Es &Ep Fb indicating that the mobility is limited by the scattering of charge carriers at the grain boundaries. Similar mobility behavior was also reported for n-InSe thin ®lms obtained by various growth techniques [4, 6, 18].
4. Conclusion
The structure, composition and electrical properties of InSe thin ®lms, thermally deposited at a substrate temperature of 150 C and the annealing effect on these properties were investigated. As-grown ®lms have amorphous structure and the heat treatment resulted in a transformation into a polycrystalline state and also increased the conductivity. Over the temperature range 100±200 K, variable-range hopping was found to be the dominant conduction mechanism in all samples. At
476
3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
C . J U L I E N , M . E D D R I E F, K . K A M B A S and M . B A L K A N S K I , Thin Solid Films 137 (1986) 27. A . A R U C H A M Y, ``Photoelectrochemistry and Photovoltaics of Layered Semiconductors'', (Kluwer Academic Publisher, Netherlands, 1992). J . C . B E R N E D E , S . M A R S I L L AC , A . C O N A N and A . G O D O Y, J. Phys. Condens Matter 8 (1996) 3439. È N A L , Z . S A L A E VA and K . M . P A R L A K , CË . E R CË E L E B I , I . G U A L L A K H V E R D I E V, Thin Solid Films 258 (1995) 860. M . PA R L A K and C Ë . E R CË E L E B I , ibid. 322 (1998) 334. M . P A R L A K and CË . E R CË E L E B I , J. Mater. Sci.: Mater. Electron. 10 (1999) 313. S . M A R S I L L AC , J . C . B E R N E D E and A . C O N A N , J. Mater. Sci. 31 (1996) 581. B . T H O M A S , Appl. Phys. A 54 (1992) 293. D . V. K . S A S T R Y and P. J . R E D DY, Thin Solid Films 105 (1983) 139. I . G U N A L and A . F. Q A S R AW I , J. Mater. Sci. 34 (1999) 1. N . F. M O T T and E . A . D AV I S , ``Electronic Process in Non Crystalline Materials'', 2nd Edn. (Clarendon Oxford, 1979). J . Y. S E T O , J. Appl. Phys. 46 (1975) 5247. M . V. G A R C I A - C U E N CA , J . L . M O R E N Z A and J . E S T E V E , ibid. 56 (1984) 1738. R . M . H I L L , Phil. Mag. 24 (1971) 1307. D . K . P A U L and S . S . M I T R A , Phys. Rev. Lett. 31 (1973) 1000. M . P E R S I N , A . P E R S I N , B . C E L U S T K A and B . E T L I N G E R , Thin Solid Films 11 (1972) 153. R . L . P E T R I T Z , Phys. Rev. 104 (1956) 1508. G . M I C O C C I , A . T E P O R E , R . R E L L A and P. S I C I L I A N O , Phys. Status Solidi A 148 (1995) 431.
Received 5 January and accepted 6 May 2001