Electrical properties of chlorine doped ZnO thin films ...

3 downloads 0 Views 121KB Size Report
Jan 17, 2008 - ... D.a,c, J. Moralesa, W. Estrada L.a, E. Andradeb, M. Miki-Yoshid, Thin Solid Films ... [22] S. F. J. Cox, E. A. Davis, S. P. Cottrell, et al, Phys. Rev ...
Original Paper

phys. stat. sol. (a) 203, No. z, zzz – zzz (2006) / DOI 10.1002/pssa.200600000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

Electrical properties of chlorine doped ZnO thin films grown by MOCVD E.Chikoidze1, M.Modreanu2, V.Sallet1, O.Gorochov1 and P.Galtier1 Groupe d’Etude de la Matière Condensée, CNRS-Université de Versailles-Saint-Quentin, 1 Place Aristide Briand. 92195 Meudon Cedex, France. 2 Tyndall National Institute-University College Cork, Lee Maltings, Prospect Row, Cork, Ireland 1

Received zzz, Published online zzz

revised

zzz,

accepted

zzz

PACS 81.05.Dz, 81.15.Gh, 73.50.-h Chlorine doped ZnO thin films were grown by MOCVD on sapphire and fused silica. Chlorine is incorporated in substitution to oxygen and acts as a donor, leading to an increase of electron concentration. Transport properties were studied for samples with different content of chlorine. Hall Effect measurements show the increase of electron carrier concentration and decreases of electron mobility while increasing the amount of chlorine incorporated in ZnO. Carrier concentration as high as 6.51020cm-3 has been achieved with resistivity of ρ=1.4x10-3 Ohm cm for layers deposited on sapphire substrate. copyright line will be provided by the publisher

1

Introduction

Nowadays significant issues remain concerning p-type doping of ZnO and related alloys. But ZnO, which has been recognized as an intrinsic, n-type semiconductors more than 70 years ago [1,2] has a broad range of applications in optoelectronic, acoustic devices and lasers light-emitting diodes. As a semiconductor oxide material with low resistivity, high transmittance up to UV spectral range, and with a good chemical stability under strong reducing environments, ZnO is thus a promising Transparent Conductive Oxide (TCO) and a possible alternative to tin oxide and indium oxide to be used as transparent electrode for photovoltaic solar cells or electrodes on flat panel displays [3]. The only way to obtain good transparent conductors is to create electron degeneracy in a wide band gap (greater than 3eV) oxide by controllably introducing non-stoichiometry and /or appropriate dopants. In 1980’s the group III impurities have been reported as effective dopants for ZnO. The resistivity of the order of 10-4 ohm cm was obtained with 2-3% Al, Ga, In and with 10% B doping [4]. Later on much study has been done toward ZnO doping, using different metal elements like Al [5,6], Ga [7,8], Ba [9], Sc [9], Cu [10], Fe [10], Sn [10], In [10,11], group IV elements as Si, Ge, Ti, Zr, Hf [12] substituting to Zn. While apart from a few works [13-15] where fluorine doping of ZnO was studied, little is known about the doping of ZnO with anion impurity in substitution to oxygen. Lately, the use of non-metal dopants in substitution to oxygen was suggested as a better way to achieve high carrier concentration and mobility while keeping good transparency, thanks to the weaker perturbation of the ZnO conduction band expected in this configuration [16]. Until now the best result regarding n type doping of ZnO with low resistivity, typically ρ=4x10-4ohm cm, was achieved with Fluorine [13]. Following the above mentioned idea and the fact

copyright line will be provided by the publisher

000000 physica (a) 203/z

WordXP

Art.: W0000/Autor

2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

E.Chikoidze, et al: Electrical and optical properties of chlorine doped ZnO grown by MOCVD

that, to our knowledge, with the exception of one brief report [17], there are no data in literature about chlorine doping of ZnO, we decided to study this issue. Though, is clear that one of the most important role could play the ionic radius of the dopant atom, being larger or smaller than substituted host anion radius, it can cause the various distortion of the lattice.

2. Experimental details ZnO and ZnO:Cl layers have been grown by Metal-Organic Chemical vapour Deposition (MOCVD) in a vertical geometry reactor. Diethyl-zinc (DEZ), n-Butil Chloride (n-BuCl), (CH3(CH2)3Cl ) and tertiarybutanol (t-Bu) have been used respectively as zinc, chlorine and oxygen sources. The chlorine precursor is liquid (-123°C melting point) and was used at different temperatures from 0°C to 20°C. The carrier gas was H2 with a 2 L/min flow. The growth was performed on two types of substrates: (001) oriented sapphire pre-treated in oxygen atmosphere at 1100 °C for 8 hours and synthetic fused silica. For both of them, growth has been done at T=425°C. When hydrogen is used as a vector gas, the growth temperature is limited T 1018 cm-3 in our samples. The role of hydrogen in the electrical conductivity of ZnO started to be discussed 50 years ago by D.C.Thomas and J.J.Lander [20]. Hydrogen has been found to diffuse easily in ZnO and thus increase the carrier concentration. Later from first -principles calculations, it became evident that hydrogen incorporating in high concentration behaves as a shallow donor. [21] Number of experiments has been performed to verify this idea. In all studies, it was found that there is non intrinsic nature shallow donor state in ZnO and that carrier concentration enhances with hydrogen incorporation [22-24]. Not surprisingly, also in our pure ZnO and slightly doped ZnO:Cl samples, hydrogen might play the important role in conductivity. During the deposition, keeping the same pressures for DEZ and t-Bu (VI/II =5.3), and increasing the chlorine vapour pressure from 3.8 to 104Pa, for the layers grown on sapphire, lead to an increase of carrier concentrations, reaching a maximal value of 1.77x1020 cm-3 (Fig.2 (a)). The electron concentration increases with chlorine vapour pressure also in the case of silica substrate, although this increase is more gradual. This might be connected to some differences in the incorporation of chlorine in ZnO on textured (on silica) or epitaxially grown (on sapphire) films. Interestingly the same effect was observed for the incorporation of Mn in ZnO with MOCVD [19]. Another explanation can be that total amount of incorporated chlorine is the same on both substrates, but the accumulation of dopant in grain boundaries in polycristalline layers grown on silica can cause the fewer amounts of electrically active chlorines.

www.pss-a.com

000000 physica (a) 203/1

copyright line will be provided by the publisher

WordXP

Art.: W0000/Autor

E.Chikoidze, et al: Electrical and optical properties of chlorine doped ZnO grown by MOCVD

on sapphire on fused silica

35 20

2

30

-1 -1

Electron mobility (V s cm )

10

25

19

10

on sapphire on fused silica 18

10

20 15 10 5 0

0

50

100

150

200

250

0

50

100

150

200

250

n-BuCl pressure (Pa)

n-BuCl pressure, (Pa)

a) b) Fig.2 a) Electron concentration and b) Electron mobility versus chlorine vapor pressure; All samples are grown with DEZ= 30Pa and t-Bu= 161Pa.

sapphire fused silica

2

30

-1 -1

35

25

Electron mobility (V s cm )

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

Carrier concentration (cm-3)

4

20 15 10 5 0 18 1,0x10

19

5,0x10

1,0x10

20

1,5x10

20

2,0x10

20

-3

Carrier concentration (cm )

Fig. 2 c) Electron mobilities versus electron concentrations for ZnO:Cl thin films. All samples are grown with DEZ= 30Pa and t-Bu= 161Pa. At 0.02% Cl doping level ( P(n-BuCl)=31 Pa), when the amount of incorporated chlorine measured from SIMS is N(Cl)=1.2 x1019cm-3 , the free carrier concentrations measured from Hall effect is higher (n=7.6 x1019cm-3). Thus at this doping range, the conductivity has still an intrinsic character. With the further increase of doping level (up to 0.5%Cl for (P(n-Bu)=84 Pa), when N(Cl)=4x1020cm-3, the measured electron concentration is n=1.6 x1020cm-3; That meance that most of the incorporated chlorines are electrically active and impurity conduction is realized. When chlorine pressure exceeds 104 Pa, the concentration of free carriers starts to decrease. It seems that, for this growth conditions, we reach the solubility limit of Cl in ZnO matrix. This result must be compared to the solubility limit of other dopants in ZnO. For example, the solubility limit of group III elements (Al, In, Ga) is very high and can reach 6-7% [8, 25, 26]. The Fluorine solubility limit was also reported to be around 5% [13]. An important point concerning solubility is the ratio between ionic radii of dopant and host cation/anion. This reflects the possibility to incorporate a dopant in substitution without inducing much perturbation in the host lattice. The ionic radii of Al (53pm), Ga (62pm), In (80pm) are smaller or very close to one of Zn (74pm). Group VII elements are supposed to www.pss-a.com

copyright line will be provided by the publisher

000000 physica (a) 203/z

WordXP 17.01.2008

Art.: W0000/Autor Bearb.:

10; insgesamt 6 Seiten

Original Paper

phys. stat. sol. (a) 203, No. z (2006)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

5

substitute oxygen which exhibit a 140 pm ionic radius. This explain why fluorine, due to its small (133pm) ionic radius, is highly soluble whereas the low solubility of chlorine is caused by its larger (181pm) ionic radius. If we compare two ZnO:Cl samples grown in the same conditions P(DEZ)=30 Pa, P(t-Bu) =161 Pa, P(nBuCl) = 50Pa with different substrates (see Fig 2(a)), we obtain the same value of electron concentration, n=1.6x1020cm-3, confirming that the carrier concentration we measured for ZnO:Cl layers is not influence by the impurities from substrate (as aluminum from sapphire). When the amount of electrically active impurity is a small (n-1/3>> aH , where aH is a donor Bohr radius in material), we speak about doping with non interacting donors discrete levels [27]. Upon increasing the concentration of donor impurities, the overlap of the donor wave functions leads the creation of impurity band which is still separated from conduction band. This is so called Mott metal-insulator transition and is realized at nc critical concentration, when nc1/3 aH ≈0.27 [28]. Assuming that donor Bohr radius in ZnO is about 0.9-1.5nm [29], we get nc>1.2 x1019 cm-3. Following increase of impurity concentration caused the enlargement of impurity band which merges the conduction band and we obtain so called heavily doped or degenerated semiconductor, with Fermi level inside the conduction band. The condition for heavy doping is n-1/3= aH, for ZnO it will be when n >5.8x1020 cm-3 [27]. As we can see in Fig.2 (b), it is possible to distinguish two regimes for the dependence of the mobility on chlorine vapour pressure: the first regime, for P(n-BuCl)100Pa, with comparatively small (9 - 12 V-1s-1cm2) variation of mobility. At room temperature, mobility in the films with low carrier concentrations is influenced mostly by the scattering of free carriers on phonons, while with high doping levels, scattering by the ionized dopant atoms becomes an important mechanism that limits mobility and can explain its sharp decrease during the first regime. However was shown that undoped and impurity doped ZnO films can be affected by grain boundary scattering as well [12]. This is demonstrated in Fig.2(c) which reflects the decrease of the mobility upon increasing the n–carrier concentration. In the case of polycrystalline samples mobility can also decrease due to grain-boundary scattering. This can explain the lower mobility observed on ZnO:Cl layers grown on fused silica substrate. The already published values of mobility on heavily doped zinc oxide films show a large scattering and are in the range of 5 to 60 cm2V-1s-1 [26]. In case of ZnO:F (0.1%-1% fluorine) reported mobilities are ranging between 10 and 40cm2V-1s-1 [13]. For our ZnO:Cl films, it change from 27 to 7 cm2V-1s-1 for layers grown sapphire and from 15 to 0.62 cm2V-1s-1 in case of fused silica one. The lower values of mobilities observed in our case can also be related to the large ionic radius of Cl with respect to O which lead to a strong distortion of the lattice and directly affects the electron mobility through an increase of the interaction of free carrier with the lattice. In addition, it is well known that the mobility strongly depends on growth temperature (i.e. on the quality of crystallinity of the layers) and can be improved using higher deposition temperatures [26]. In our case we were limited to a deposition temperature of 425°C, due to the use of hydrogen as a vector gas [18]. For fixed DEZ and chlorine pressures, a decrease of t-Bu, (VI/II =1.07) pressure leads to a decrease of the resistivity. Probably in oxygen poor conditions the formation of oxygen vacancies is favorable, leading the increase of chlorine solubility in zinc oxide. The lowest resistivity obtained was ρ=1.38x10-3ohm cm, (n=6.13x1020 cm-3and µ=7.4 V-1s-1cm2) for ZnO:Cl grown on sapphire. Layers grown on silica could have the carrier concentration as high as in the case of sapphire substrate, but since the films on fused silica are polycrystalline with a lower mobility, a resistivity below ~10-2 Ohm cm has not been obtained.

4. Conclusion In this work, we have studied the possibility to incorporate Chlorine as n-dopant in ZnO thin films with MOCVD on sapphire and fused silica as substrates. We studied doping range with chlorine incorporation bellow 1%. The n-carrier concentration increases with the incorporation of chlorine reaching a www.pss-a.com

000000 physica (a) 203/1

copyright line will be provided by the publisher

WordXP

Art.: W0000/Autor

6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

E.Chikoidze, et al: Electrical and optical properties of chlorine doped ZnO grown by MOCVD

maximal value of n=6.13x1020 cm-3 when grown on sapphire substrate. The electron mobility decreases when increasing the doping level, even though the results obtained on films deposited on sapphire exhibit a low resistivity (ρ~10-3 ohm cm). This study shows that ZnO thin films can be n-type doped effectively with chlorine using MOCVD technique, making it a possible candidate for TCO applications. The following work will be by changing deposition conditions to increase the solubility of chlorine in ZnO and study optical and electrical properties of ZnO:Cl in heavy doping regime. Acknowledgements Work was supported by NATCO-FP6-511925 project and INTAS N 05-114-4526 grant. References [1] H.H.Baumbach and C.Wagner, Zeits.F.physik.Chemie B22, 199 (1933). [2] O.Fritsch, Ann.D.Physik 22, 375 (1935). [3] A.Guillen-Santiago, M.Olivera, A.Maldonado, et al. Phys.Stat.Sol.(a) 201, 952, (2004). [4]T. Minami, H. Sato, H. Nanto and S. Takata, Jap. J. Appl. Phys. 24, L781 (1985). [5] B-Y Oh, M-C Jeong, Woong Lee, Jae-Min Myoung, J.Cryst.Growth 274, 453 (2005). [6] A.V.Singh, M.Kumar, R.M.Mehra, et al, J.Indian Inst.Sci., 81 527, (2001). [7] H. Kato, M. Sano, K. Miyamoto, T.Yao, J.Cryst.Growth, 273, 538 (2002). [8] J.D.Ye, S.L.Gu, S.M.Zhu et al, J.App.Phys.86, 192111 (2005). [9] Y.Morianga, K. Sakuragi, N. Fujilura, Tachiro Ito, J.Cryst.Growth, 174, 691 (1997). [10] F. Paraguay D.a,c, J. Moralesa, W. Estrada L.a, E. Andradeb, M. Miki-Yoshid, Thin Solid Films 366, 16 (2000). [11] L. Castaneda , A. Garcı´a-Valenzuel , E.P. Zironi ,T, et al, Thin Solid Films 503 212, (2006). [12] T. Minami, MRS Bull. 25, 38 (2000) [13] J.Hu and R.G.Gordon , Sol.Cells 30, 437 (1991). [14]. H.Y.Xu, Y.C.Liu, R.Mu, Appl.Phys.lett. 86, 123107 (2005) [15] A.Guillen-Santiago, M.Olivera, A.Maldonado, et al. Phys.Stat.Sol.(a) 201, 952, (2004) [16] R.Gordon, MRS Bulettin,/August, 52 (2000) [17] B.Hahn, G.Heindel, E.Pschorr-Schoberer et al, Semic.Sci.Techol.13, 788 (199 [18]V. Sallet, C. Thiandoume, J.F. Rommeluere, A. Lusson, et al. Mater. Lett. 53 (2002) 126. [19] E.Chikoidze, Y.Dumont, F.Jomlard, et al, Mat. Res Bull, 41, 1038, (2006) [20] D.C.Thomas and J.J.Lander [20,]Phys.Rev. 25, 1136, 1957 [21]C. G. van de Walle, Phys. Rev. Lett. 85, 1012 (2000). [22] S. F. J. Cox, E. A. Davis, S. P. Cottrell, et al, Phys. Rev. Lett. 86, 2601 (2001). [23]F. Ruske, V. Sittinger, W. Werner,et al, Surf. Coat. Techn. 200, 236 (2005). [24]Bertrand Theys, Vincent Sallet, Francois Jomard, J.Appl.Phys.91, 3922, 2001 [25]Y.Liu, J.Lian Appl.Surf.Sci. 253, 3727, (2007) [26] Ellmer, J. Phys. D: Appl. Phys.34, 3097 (2001) [27]B.L.Bonch-Bruevich, S.G.Klashnikov, “The physics of semiconductors”, Moscow, Nauka 1990, p.677 [28] D. M. Hofmann, A.Hofstaetter, F.Leiter et al, Phy.Rev.Lett, 88, 045504-1 (2002) [29]. N.F.Mott “Metal Insulator Transitions”, Taylor and Francis Ltd, London, 1990, p.219

www.pss-a.com

copyright line will be provided by the publisher

000000 physica (a) 203/z

WordXP 17.01.2008

Art.: W0000/Autor Bearb.:

10; insgesamt 6 Seiten