Influence of Zn/O flux ratio and Mn-doped ZnO buffer on the plasma-assisted molecular beam epitaxy of ZnO on c-plane sapphire

(1)

Inﬂuence of Zn/O ﬂux ratio and Mn-doped ZnO buffer on the plasma-assisted

molecular beam epitaxy of ZnO on c-plane sapphire

J.S. Wang

a,

, C.S. Yang

b

, M.J. Liou

a

, C.X. Wu

a

, K.C. Chiu

a

, W.C. Chou

c

a

Department of Physics and Center for Nano-Technology, Chung Yuan Christian University, Chung-Li 32023, Taiwan b_{Graduate Program in Electro-optical Engineering, Taipei 10452, Taiwan}

c

Department of Electrophysics, National Chiao Tung University, Hsin-Chu 30010, Taiwan

a r t i c l e

i n f o

Article history: Received 2 June 2008 Received in revised form 28 July 2008

Accepted 30 July 2008 Communicated by D.P. Norton Available online 6 August 2008 PACS: 81.15.Hi 81.10.Aj 81.40.Tv 78.55.m 68.60.p Keywords: A1. Characterization

A1. Reﬂection high-energy electron diffraction

A3. Molecular beam epitaxy B1. Oxides

B1. Zinc compounds

B2. Semiconducting II–VI materials

a b s t r a c t

This work investigated the influence of Zn/O flux ratio and Mn-doped ZnO buffer layer on the epitaxial growth of ZnO grown by plasma-assisted molecular beam epitaxy on c-plane sapphire substrates. Atomic force microscopy (AFM), photoluminescence (PL) and X-ray diffraction (XRD) measurements indicated that a small amount residual strain of ZnO epilayers was further relaxed under stoichiometric growth conditions due to the better surface migration of the adatoms. Moreover, we observed that a small amount of Mn doping led to obtain a flatter surface with stronger lattice relaxation maybe due to the greatly enhanced surface migration of the adatoms. By adding a Mn-doped ZnO buffer layer the optical and electrical properties of the ZnO epilayers had significant improvement.

1. Introduction

Wurtzite-structure ZnO with a wide direct band gap of 3.37 eV at room temperature (RT) has a large exciton binding energy of 60 meV. The fundamental properties of ZnO make it an attractive candidate for UV light-emitters or laser diodes. Optical pumped

excitonic lasing at RT from ZnO has been reported [1,2], and

excitonic-stimulated emissions have been observed at

tempera-tures up to 550 1C[3]. Another possible application of ZnO is in the

ﬁeld of spintronics [4,5]. After Dietl et al. [6] predicted RT

ferromagnetism for Mn-doped ZnO, a number of studies have been carried out experimentally and theoretically to obtain

reliable Mn-doped ZnO diluted magnetic semiconductors[7–9].

Due to the hexagonal symmetry and low cost, ZnO epitaxial layers grown by plasma-assisted molecular beam epitaxy (PA-MBE)

have mostly been studied using c-plane sapphire as a substrate

[10–13]. However, there is a large in-plane lattice mismatch (18%) and two types of in-plane rotation domains between c-oriented ZnO and the oxygen sublattice of c-plane sapphire

(ZnO½1 ¯1 0 0||Al2O3½1 1 ¯2 0 and ZnO½1 1 ¯2 0||Al2O3½1 1 ¯2 0).

These 301-rotation domains are surrounded by highly defective domain boundaries with threading dislocations leading to

deterioration of the crystal quality[14,15]. Therefore, the lattice

strain relaxation and accompanying structural evolution in the initial growth stage directly affect the surface morphology and the

crystalline quality [16,17]. Low-temperature ZnO homo-buffer

layers have been tried to reduce the defects and dislocation

density caused by the large lattice mismatch[18,19]. To eliminate

the hexagon-on-hexagon growth (ZnO½1 1 ¯2 0||Al2O3½1 1 ¯2 0),

some researchers have suggested that depositing a thin MgO hetero-buffer layer, which has lattice mismatch of 8.4% and 9.1% between ZnO and sapphire, respectively, at the interface between

ZnO and c-plane sapphire [20,21]. Furthermore, a dopant can

inﬂuence both the kinetic and thermodynamic factors in the

ARTICLE IN PRESS

Contents lists available atScienceDirect

journal homepage:www.elsevier.com/locate/jcrysgro

Journal of Crystal Growth

doi:10.1016/j.jcrysgro.2008.07.108

Corresponding author. Tel.: +886 3 2653210; fax: +886 3 2653299. E-mail address:jswang@cycu.edu.tw (J.S. Wang).

(2)

formation and growth of grains during ﬁlm deposition. Kim et al.

[22]demonstrated that Mn doping (Mn ¼ 5%) greatly enhanced

the atomic alignment and led to a singly oriented ﬁlm by totally

suppressing the hexagon-on-hexagon growth. Han et al. [23]

reported that ZnO doped with Mn increased the surface diffusion.

Nakayama et al. [24] showed that a Zn0.95Mn0.05O thin ﬁlm is

preferentially oriented along the [0 0 0 1] axis and that the interface and surface are very ﬂat.

In this work, we present our study on the growth of ZnO epilayers varying with Zn/O ﬂux ratio and the improvements in crystal structures, optical and electrical properties by using a Mn-doped ZnO buffer layer at the interface between ZnO and c-plane sapphire substrates grown by PA-MBE.

2. Experiment

The growth was carried out in a Veeco EPI 620 MBE system. An Addon RF-plasma source with independently separated pumping design was used to provide reactive oxygen radicals. The ﬂux of oxygen gas was controlled by a mass ﬂow controller system. Knudsen-cells were used for evaporating Zn and Mn metal sources. Before deposition, the c-plane sapphire substrates were cleaned by acetone and chemical etched by the heated acid

solution of H2SO4and H3PO4for 15 min. Following, the substrate

was desorbed at 920 1C. During growth, the substrate temperature was fixed at 650 1C and the oxygen RF–plasma source was kept at 300 W with an oxygen flow rate of 1.5 SCCM. The whole growth process was monitored by reflection high-energy electron diffrac-tion (RHEED). Film thickness was measured by employing the cross-sectional scanning electron microscope (SEM). After growth, the samples were investigated by atomic force microscopy (AFM), X-ray diffraction (XRD), Hall measurements, and photolumines-cence (PL) spectroscopy. A 325 nm of He–Cd laser was employed as the exciting source for PL measurements.

3. Results and discussion 3.1. Effect of Zn/O ﬂux ratio

Recent reports have demonstrated that the ZnO layers grown under the stoichiometric and/or the slight Zn-rich ﬂux conditions showed better crystalline qualities due to the better surface

migration of the adatoms[25,26].Fig. 1shows the growth rates of

ZnO, which was grown directly on sapphire substrates without buffer layers, estimated from the cross-sectional SEM images as a function of reciprocal Zn temperature. On the whole, the ﬂux of solid source is an exponential function of the reciprocal

temperature in a suitable range. As shown in Fig. 1, the ZnO

growth rate linearly increases up to near 310 1C and then saturates at higher Zn temperature. This behavior indicates that the growth mode is limited by the minority atom on the surface. When the growth rate is governed under oxygen-rich conditions, the growth rate linearly increases when Zn beam flux increases. When the Zn/O flux ratio is close to unity (i.e. the stoichiometric flux condition) the growth rate displays a gradual increase. Finally, the growth condition gets into Zn-rich, as the growth rate becomes saturated with further increases in Zn beam flux.

Fig. 2shows the root-mean-square surface roughness of the

ZnO ﬁlms over the scanned size of 5

mm 5mm with different Zn

temperature. The smallest roughness occurred at near the stoichiometric conditions due to the better surface migration of

the adatoms[25,26].

Fig. 3displays the 10 K normalized PL spectra measured at the near band edges from ZnO epilayers. The spectra show exciton

emissions at 3.357, 3.361, and 3.376 eV. According to their energy, these emission lines can be ascribed to excitons bound to neutral

donors (D1X) and free excitons (FXA), respectively. However, the

spectra of samples, which were grown under near-stoichiometric growth conditions (Zn ¼ 310 and 315 1C), revealed an energy red-shift of about 2 meV. This indicates that a small amount residual strain of ZnO epilayer was further relaxed under stoichiometric growth conditions due to the better surface migration of the adatoms. The (0 0 0 2) XRD measurements of ZnO epilayers, as

shown inFig. 4, agreed with the results of PL. The larger angle

peaks in each spectrum are due to the Ka2 line of the X-ray source, and the diffraction angles of all spectra have been shifted to get the same angle from the sapphire substrates in different samples. Because of the compressive residual strain in ZnO epilayer, the lattice constant of c-axis became smaller as the strain was further relaxed under stoichiometric growth conditions.

ARTICLE IN PRESS

(III) (II) Growth rate (nm/s) 280 290 300 310 320 330 (I) (I): Zn-rich (II): Stoichiometry (III): O-rich 1000/TZn (1/K) 1.65 1.70 1.75 1.80 TZn (°C) 10-3 10-2

Fig. 1. Growth rates of ZnO films, grown under various Zn temperatures with fixed O-plasma conditions, estimated from the cross-sectional SEM images as a function of reciprocal Zn temperature. The solid line is the linear fitting of the lower-temperature region. 0 5 10 15 20 25 30 35 40 Root-mean-square roughtness (nm) Temperature of Zn (_°C) 295 300 305 310 315 320 325 330 335

Fig. 2. Root-mean-square surface roughness of the ZnO ﬁlms, grown under various Zn temperatures with ﬁxed O-plasma conditions, over the scanned size of 5mm 5mm as a function of Zn temperature.

J.S. Wang et al. / Journal of Crystal Growth 310 (2008) 4503–4506 4504

(3)

3.2. Effect of Mn-doped ZnO buffer

Fig. 5b and c shows the RHEED patterns of 6 nm-thick ZnO and Mn-doped ZnO (ZnO:Mn) at substrate temperature of 650 1C,

respectively. The sapphire substrate displays a sharp streaky

pattern, as shown in Fig. 5a, directly after oxygen plasma

treatment, indicating a clean and ﬂat surface. A 301 rotation in the basal plane of the ZnO layer was observed by RHEED, and the ½1 ¯1 0 0 direction of ZnO was found to align with the ½1 1 ¯2 0

direction of Al2O3. This indicates the ZnO lattice aligns itself with

the oxygen sublattice in Al2O3, and the lattice mismatch is

reduced from 32% to around 18%[13]. The RHEED patterns show

that the crystal structure of ZnMnO buffer layers is the same as that of the ZnO. Besides, the in-plane lattice mismatch between the sapphire substrate and the 6 nm-thick ZnO and ZnO:Mn layers was estimated from the rod spacing of RHEED patterns to be around 12.1% and 15.4%, respectively. Moreover, AFM measure-ments revealed the root-mean-square surface roughness of the ZnO and ZnO:Mn layers was about 0.4 and 0.1 nm, respectively. We determined the concentration of Mn in this 6 nm-thick ZnO:Mn epilayer is less than 0.6% through the X-ray photoelectron spectroscopy (XPS) measurements. Our results suggest that a

ARTICLE IN PRESS

3.34

Normalized PL intensity (a.u.)

Photon energy (eV) 10 K FXA D 0 X Zn=320 °C Zn=315 °C Zn=310 °C Zn=300 °C 3.35 3.36 3.37 3.38

Fig. 3. Normalized 10 K PL spectra measured at the near band edge from ZnO epilayers grown under various Zn temperatures with ﬁxed O-plasma conditions.

Zn=320 °C

Zn=315 °C

Zn=310 °C

Zn=300 °C

Normalized intensity (a. u.)

2θ (degree)

34.3 34.4 34.5 34.6 34.7 34.8

Fig. 4. Normalized (0 0 0 2) XRD spectra of ZnO epilayers grown under various Zn temperatures with ﬁxed O-plasma conditions. The larger angle peaks in each spectrum are due to the Ka2 line of the X-ray source.

Fig. 5. RHEED patterns of (a) the sapphire substrate after oxygen plasma treatment, Al2O3½1 1 ¯2 0, (b) 6 thick ZnO epilayer, ZnO ½1 1 ¯2 0, and (c) 6 nm-thick ZnO:Mn epilayer, ZnO:Mn ½1 1 ¯2 0. The substrate temperatures were 650 1C. J.S. Wang et al. / Journal of Crystal Growth 310 (2008) 4503–4506 4505

(4)

small amount of Mn doping greatly enhanced surface migration of the adatoms, and led to obtain a ﬂatter surface with stronger lattice relaxation. Therefore, the ZnO:Mn layer could be a very promising buffer for the growth of high-quality ZnO.

Fig. 6shows the 10 K normalized PL spectra measured at the

near band edges from 0.25

mm-thick ZnO epilayers grown under

near-stoichiometric growth conditions with and without 6 nm-thick ZnO:Mn buffer layers, respectively. By comparison the

sample with ZnO:Mn buffer layer shows the much stronger FXA

emission and the smaller FWHM of D1X peak. On the other hand, decreased defect and/or dislocation density might result in a lower residual carrier concentration, and a higher mobility due to the decreased concentration of scattering centers. Hall

measure-ments showed the mobility was enhanced from 24 to 50 cm2_{/V s,}

and residual carrier concentration decreased from 1.4 1019 _to

4.3 1018_cm3 _{by adding a 6 nm-thick ZnO:Mn buffer layer,}

indicating that this initial ZnO:Mn buffer layer could improve the crystalline quality of ZnO ﬁlm.

4. Conclusion

The effects of the Zn/O ﬂux ratio and the ZnO:Mn buffer layers on the material properties of ZnO epilayers grown by PA-MBE were investigated. A small amount of residual strain of ZnO epilayer was further relaxed under stoichiometric growth condi-tions, and caused a red-shift of near band edge emissions. Moreover, we proposed a ZnO:Mn buffer layer with simple growth

process for the ZnO PA-MBE growth on c-plane sapphire substrates. A small amount of Mn doping could greatly enhance surface migration of the adatoms, and led to obtain a ﬂatter surface with stronger lattice relaxation. Our results demonstrated that by using a ZnO:Mn buffer layer the optical and electrical properties of the ZnO epilayers had signiﬁcant improvement.

Acknowledgments

This work was supported by the National Science Council of the Republic of China, Taiwan, under grant numbers NSC 95-2112-M-033-008-MY3. The authors are grateful for the support from the National Nano-Device Lab.

References

[1] D.M. Bagnall, Y.F. Chen, Z. Zhu, T. Yao, S. Koyama, M.Y. Shen, T. Goto, Appl. Phys. Lett. 70 (1997) 2230.

[2] Z.K. Tang, G.K.L. Wong, P. Yu, M. Kawasaki, A. Ohmoto, H. Koinuma, Y. Segawa, Appl. Phys. Lett. 72 (1998) 3270.

[3] D.M. Bagnall, Y.F. Chen, Z. Zhu, T. Yao, M.Y. Shen, T. Goto, Appl. Phys. Lett. 73 (1998) 1038.

[4] H. Ohno, Science 281 (1998) 951.

[5] S.J. Pearton, et al., J. Appl. Phys. 93 (2003) 1.

[6] T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science 287 (2000) 1019. [7] T. Fukumura, Z. Jin, A. Ohtomo, H. Koinuma, M. Kawasaki, Appl. Phys. Lett. 75

(1999) 3366.

[8] T. Fukumura, Z. Jin, M. Kawasaki, T. Shono, T. Hasegawa, S. Koshihara, H. Koinuma, Appl. Phys. Lett. 78 (2001) 958.

[9] K. Sato, H. Katayama-Yoshida, Jpn. J. Appl. Phys. Part 2 40 (2001) L334. [10] Y. Chen, D.M. Bagnall, Z. Zhu, T. Sekiuchi, K. Park, K. Hiraga, T. Yao, S. Koyama,

M.Y. Shen, T. Goto, J. Crystal Growth 181 (1997) 165.

[11] P. Fons, K. Iwata, S. Niki, A. Yamada, K. Mastubara, M. Watanabe, J. Crystal Growth 201/202 (1999) 627.

[12] K. Sakurai, M. Kanehiro, K. Nakahara, T.T. Tanabe, Sz. Fujita, Sg. Fujita, J. Crystal Growth 214/215 (2000) 92.

[13] Y. Chen, D.M. Bagnall, H.J. Koh, K.T. Park, K. Hiraga, Z. Zhu, T. Yao, J. Appl. Phys. 84 (1998) 3912.

[14] H. Kato, M. Sano, K. Miyamoto, T. Yao, Jpn. J. Appl. Phys. Part 1 42 (2003) 2241. [15] C. Liu, S.H. Chang, T.W. Noh, M. Abouzaid, P. Ruterana, H.H. Lee, D.W. Kim,

J.S. Chung, Appl. Phys. Lett. 90 (2007) 011906.

[16] A.B.M. Almamun Ashraﬁ, N.T. Binh, B. Zhang, Y. Segawa, Appl. Phys. Lett. 84 (2004) 2814.

[17] H.C. Ong, A.X.E. Zhu, G.T. Du, Appl. Phys. Lett. 80 (2002) 941.

[18] K. Nakahara, H. Takasu, P. Fons, K. Iwata, A. Yamada, K. Matsubara, R. Hunger, S. Niki, J. Crystal Growth 227/228 (2001) 923.

[19] Y.S. Jung, O. Kononenko, J.S. Kim, W.K. Choi, J. Crystal Growth 274 (2005) 418. [20] Y. Chen, H.J. Ko, S.K. Hong, T. Yao, Appl. Phys. Lett. 76 (2000) 559. [21] K. Miyamoto, M. Sano, H. Kato, T. Yao, Jpn. J. Appl. Phys. 41 (2002) L1203. [22] S.S. Kim, J.H. Moon, B.T. Lee, O.S. Song, J.H. Je, J. Appl. Phys. 95 (2004) 454. [23] J. Han, P.Q. Mantas, A.M.R. Senos, J. Am. Ceram. Soc. 88 (2005) 1773. [24] M. Nakayama, H. Tanaka, K. Masuko, T. Fukushima, A. Ashida, N. Fujimura,

Appl. Phys. Lett. 88 (2006) 241908.

[25] H.J. Ko, T. Yao, Y. Chen, S.K. Hong, J. Appl. Phys. 92 (2002) 4354.

[26] A. Setiawan, Z. Vashaei, M.W. Cho, T. Yao, H. Kato, M. Sano, K. Miyamoto, I. Yonenaga, H.J. Ko, J. Appl. Phys. 96 (2004) 3763.

ARTICLE IN PRESS

0.01 0.1

1 _{with ZnO:Mn buffer}

without buffer

Normalized intensity (a.u.)

Photon energy (eV) FXA 10 K

3.34 3.35 3.36 3.37 3.38 3.39

Fig. 6. Normalized 10 K PL spectra measured at the near band edge from 0.25 mm-thick ZnO epilayers grown under near stoichiometric growth conditions with and without 6 nm-thick ZnO:Mn buffer layers, respectively.

J.S. Wang et al. / Journal of Crystal Growth 310 (2008) 4503–4506 4506