Effect of annealing atmosphere on physical characteristics and photoluminescence properties of nitrogen-implanted ZnO thin films

(1)

Journal of Crystal Growth 285 (2005) 96–102

Effect of annealing atmosphere on physical characteristics

and photoluminescence properties of nitrogen-implanted

ZnO thin ﬁlms

Chih-Cheng Yang, Chin-Ching Lin, Cheng-Hsiung Peng, San-Yuan Chen

Department of Materials Science and Engineering, National Chiao Tung University, 1001 Ta-Hsueh Road, 300, Hsinchu, Taiwan, ROC Received 18 June 2005; received in revised form 30 July 2005; accepted 30 July 2005

Available online 3 October 2005 Communicated by D.P. Norton

Abstract

ZnO films were pre-treated with nitrogen implantation in the range from 5 1012to 5 1015cm2. Effect of nitrogen concentration on structural crystallinity and photoluminescence properties of nitrogen-implanted ZnO films under different atmospheres and annealing treatments was investigated. It was found that there exists a solubility issue of nitrogen in ZnO films at 850 1C via secondary ion mass spectrometry (SIMS). It was found that the peak intensity of near band-edge (NBE) emission remarkably decreases with the increase of concentration of implanted nitrogen if annealed in nitrogen atmosphere due to the increased oxygen vacancies. However, when the ZnO was implanted with 5 1012cm2and annealed in oxygen atmosphere, the optical properties are improved probably because the effective incorporation of O atom diminishes those donor levels (oxygen vacancies) and the crystallinity is also improved due to implanted nitrogen. However, excess nitrogen would reduce the crystallinity and promote the formation of the deep-level emission due to high amount of intrinsic and structure defects.

PACS: 64.75.+g; 82.65.+r; 68.35.Dv

Keywords: A1. Photoluminescence; B1. ZnO; A1. Nitrogen implantation; A1. Annealing atmosphere

1. Introduction

In wide band gapoptoelectronics, zinc oxide (ZnO) is attracting more attention because of its potential applications in various ﬁelds, such as ultraviolet (UV) resistive coating, gas sensors, solar cells and optical devices [1,2]. One of these

www.elsevier.com/locate/jcrysgro

Corresponding author. Tel.: +886 3 5731818; fax: +886 3 5725490.

(2)

unique properties is that ZnO thin films present strong spontaneous and simulated emissions by excitons even at room temperature [3,4]. It has been recognized that to grow high quality both p and n types ZnO thin films are necessary for the development of ZnO-based optoelectronic devices. The n-type ZnO is easily available even without any doping because ZnO is a natural n-type semiconductor, while it has been recognized that p-type ZnO is very difficult to develop although high densities of holes could be achieved with nitrogen as the dopant. The presence of intrinsic defects such as interstitial zinc and oxygen vacancies has been considered to cause a deviation from stoichiometry [5]. Ogata et al. [6] reported that those intrinsic defects can be reduced by thermal annealing of ZnO layer in O2atmosphere

but the electron carrier density was also decreased. A number of groups have been trying to fabricate p-type ZnO films using N or As [7,8]. Nitrogen (N) doping has been considered as an effective method to realize p-type ZnO films because N has the smallest ionization energy in groupV-dopants [9]. Recently, several investiga-tions focused on nitrogen and donor co-implanted ZnO to study the solubility and diffusion behavior of nitrogen in ZnO crystal[10,11]. Georgobiani et al. [12] studied the electrical properties of ZnO films with ion implantation of nitrogen from 3 1014 to 3 1015cm2 in oxygen atmosphere. It was shown that the nitrogen implantation could result in the formation of the hole type of conductivity but new peaks had appeared in the ultraviolet and visible ranges of photolumines-cence spectra.

Even though it was reported that p-type ZnO may be developed by using nitrogen as an acceptor dopant, no detailed studies were made to investi-gate the effect of implanted-nitrogen concentra-tion on the structural change and optical properties of sputtered ZnO films. The situation makes it difficult to study the optimal properties dependence of the doping parameters, which is essential for the full understanding of the physical and optoelectronic characterization. Furthermore, it is well known that both physical characteriza-tion and optoelectronic properties are strongly influenced by the defect concentration in ZnO

films and this can be modified and controlled via nitrogen treatment under different atmosphere and annealing conditions. Therefore, in this work, ZnO films were pre-treated with nitrogen implan-tation in the range from 5 1012to 5 1015cm2. The photoluminescence behavior of the N-im-planted ZnO films annealed in different atmo-spheres as a function of nitrogen concentration will be primarily focused. This study can provide more valuable optical information for the devel-opment of p-type ZnO doped with nitrogen in the applications of optoelectronic and other optical devices.

2. Experimental procedure

The ZnO thin films (150 nm thick) were deposited on 4 in diameter Si substrates by RF magnetron sputtering, using 99.99% ZnO as the target. The growth chamber was evacuated by a turbo pump and mechanical pump. Before the deposition, a target was pre-sputtering by Ar ions for 10 min to remove contamination on the target. Argon and oxygen mixtures with oxygen molar ratio (OMR) of 5% were used as sputtering gases. Silicon substrates were cleaned by usual semicon-ductor technology methods before loading into the chamber. Sputtering conditions for ZnO films were performed at a substrate temperature of 50 1C, RF power of 50 W, sputtering pressure of 10 m Torr and sputtering time of 40 min. The as-grown ZnO film was subsequently subjected to N implantation at room temperature. The nitrogen ions with energy of 80 KeV were injected into the as-grown ZnO films. The fluences range studied was from 5 1012 to 5 1015cm2. Based on Rutherford back scattering measurement followed by (the transport of ions in matter) TRIM simulations, the nitrogen ion distribution in ZnO films forms a nearly perfect Gaussian shape with the peak position at (0.08+0.01) mm below the surface. After ion implantation, the specimens were an-nealed at 850 1C for 20 min under pure oxygen and nitrogen atmospheres. The thickness of ZnO films was measured using a surface profilometer (Tencor Alpha-Step 200) and the crystal structure was determined by X-ray diffractometer using CuKa

(3)

radiation. Photoluminescence measurement was performed by the excitation from 325 nm He–Cd laser at room temperature. The depth proﬁle of the implanted ions was measured by secondary ion mass spectrometry (CAMECA IMS-SF). A 10 KeV Cs+primary beam was used to determine the atomic distribution of the implanted N.

3. Results and discussion

3.1. Physical characteristics

Fig. 1 shows the depth profile of nitrogen-implanted ZnO films with various fluences. As can be seen, a normal shape with almost Gaussian distribution was observed for the ZnO film after annealed at 850 1C in nitrogen atmosphere. An abrupt change as marked in Fig. 1appears in the slope of the concentration profile. The concentra-tion at this point corresponds to the solubility of N in ZnO film that is determined as 9.1 1017, 4.2 1018, and 6.3 1018ions/cm3with the fluence of 5 1012, 1 1014, and 5 1015ions/cm2, respec-tively[13]. The value of solubility is equal to about 4–5 atom ppm and this implies that the N is a very insoluble element in ZnO.

Fig. 2 shows the XRD patterns of the N-implanted (fluence of 5 1012, 1 1014, and 5 1015cm2) and non-implanted ZnO thin films annealed at 850 1C in nitrogen atmosphere. A strong and sharp(0 0 2) peak is observed in non-implanted sample (Fig. 1a), which indicates that the ZnO film exhibits a preferred (0 0 2) orientation with c-axis perpendicular to the substrate. With an increase of nitrogen concentration, it was found that diffraction peak was shifted towards the smaller 2y direction and the shift increases with an increase of fluence from 5 1012 to 1 1014cm2. According to the Bragg Law, the shift implies that the lattice constant increases. This was considered due to the incorporation of the nitrogen at interstitial sites in ZnO, leading to an increase in the lattice constant. However, as the fluence exceeds 1 1014cm2, the position of diffraction peak was almost unchanged, but a weak diffraction peak was observed at the fluence of 1 1014cm2 for the N-implanted ZnO films annealed at 850 1C in nitrogen atmosphere. It was believed that the phenomenon is strongly depen-dent on the solubility limit of N ion in the ZnO matrix. It could be postulated that part of nitrogen ions could have occupied the sites of O atom. Therefore, the (0 0 2) peak was shifted to lower diffraction angle side and the crystal structure was deteriorated. 1021 1020 1019 1018 1017 0.00 0.02 0.04 006 0.08 0.10 0.12 0.14 0.16 0.18 Depth (µm) N concentration (cm -3) a b c

Fig. 1. SIMS depth profile of various fluences of nitrogen implanted into ZnO films after annealing at 850 1C in nitrogen atmospheres. (002) 5x1015 1x1014 5x1012 non-implanted 30 32 34 36 38 40 2θ (deg) Intensity (a.u.)

Fig. 2. XRD patterns of ZnO ﬁlms with different ﬂuences annealed in nitrogen atmospheres at 850 1C.

(4)

On the other hand, as both N-implanted and non-implanted ZnO ﬁlms were annealed in oxygen atmosphere at 850 1C, the XRD patterns inFig. 3

illustrate that the diffraction peak (0 0 2) of N-implanted ZnO samples is also shifted with the implanted N concentration up to 1 1014cm2. Similar phenomenon is also observed in the samples annealed at nitrogen atmosphere and this indicates that there exists a solubility limit for N-implanted ZnO ﬁlm annealed at 850 1C but seems independent of the annealing atmosphere. How-ever, as comparingFig. 2withFig. 3, it was found that the (0 0 2) peak is stronger for the samples annealed at O2 than N2 atmosphere. This may

suggest that the nitrogen-implanted ZnO ﬁlms present better crystallinity when annealed in oxygen atmosphere compared to that in nitrogen atmosphere.

3.2. Photoluminescence properties

In our previous studies, it was found that the predominant lattice defects in the as-grown ZnO films are the oxygen vacancies because the as-grown ZnO films were sputtered at the OMR 5%. As the samples were annealed below 500 1C, a very low intensity emission with the same feature as the as-grown films was observed. However, with the increase of annealing temperature, some lattice

and surface defects could be removed and the ZnO film would be re-structured into more perfect structure, indicating that the role of predominant defect may be different. At 850 1C, a strong NBE emission with a weak deep-level emission of 528 nm was obtained. However, when the films were annealed at a higher annealing temperature such as 1000 1C, a wide emission band appears around 560 nm and the deep-level emission peak shifts from 2.35 to 2.26 eV in comparison with that at 850 1C [14]. Therefore, 850 1C was used in this work to study the influence of implanted nitrogen concentration on the ZnO films in different atmo-spheres. As shown in Fig. 4(a), a very stronger NBE peak (378 nm) and a relatively low deep-level emission (528 nm) were obtained for the non-implanted sample. However, as the ZnO films were implanted with different fluences of nitrogen ion and annealed in this condition ( N2atmosphere),

the NBE peak becomes weaker and presents slightly red shift as compared to the non-implanted one. The decrease in the NBE emission must be related to the variation of the intrinsic defects such as zinc vacancy (VZn), oxygen vacancy

(VO) and interstitial zinc (Zni). As the samples

were annealed at 850 1C in nitrogen atmosphere, either VO or Zni should be apparently increased

and this could promote the possibility of Ni to

occupy the oxygen vacancies to form NOdefects.

Therefore, both Zniand NOwill be increased with

increasing implanted nitrogen concentration. Furthermore, it was found that with an increase of implanted nitrogen concentration, the NBE emission of ZnO film becomes weak. This PL result along with the XRD analysis implies that there exists a critical implanted nitrogen concen-tration corresponding to the defect transition that in turn influences the related PL properties. Generally, the defect reaction for non-implanted films can be expressed as follows:

ZnZnþOO)ZniþVOþ1₂O2. (1)

When the ZnO ﬁlm was annealed at a higher temperature, both zinc interstitials and oxygen vacancies would be induced from the ZnO ﬁlms as illustrated in Eq. (1). However, as the ZnO was implanted with nitrogen and then annealed at high temperature, the induced defect concentration and

(002) 5x1015 1x1014 5x1012 non-implanted 30 32 34 36 38 40 2θ (deg) Intensity (a.u.)

Fig. 3. XRD patterns of ZnO ﬁlms with different ﬂuences annealed in oxygen atmospheres at 850 1C.

(5)

defect type would be changed with different fluences for the nitrogen-implanted ZnO films. In order to clarify the role of N-implanted concen-tration in the ZnO films, it can be further elucidated as follows. First, if the fluence was below 1 1014cm2, the nitrogen ions may possi-bly substitute for the oxygen atoms in ZnO films depending on annealing conditions especially on atmosphere. Because the ZnO films in our experi-ment grew in lower oxygen partial pressure (OMR ¼ 5%), VO and Zni can be easily formed

especially as the ZnO ﬁlms were further annealed in nitrogen atmosphere at high temperature.

Therefore, the oxygen could be evaporated from the as-grown ZnO ﬁlms and more oxygen vacan-cies will be produced in nitrogen atmosphere. Furthermore, it is possible for a certain amount of nitrogen ions to occupy oxygen sites in the N-implanted ZnO ﬁlms after nitrogen annealing at high temperature. The defect chemistry (reaction) could be depicted as follows:

NiþZnO ! N2

ZnZnþNOþ1₂O2. (2)

Therefore, the peak intensity of NBE emission decreases with the increase of implanted nitrogen concentration as illustrated inFig. 4(a). Therefore, it is likely that high nitrogen concentration is responsible for producing high amount of struc-ture defects (NO) in ZnO ﬁlms, which are the

centers of non-radiative recombination.

In addition, according to XPS analysis for oxygen (O1s) peak in the ZnO films, the intensity of the peak (531 eV) shown in Fig. 5(a) for the non-implanted sample annealed at 850 1C in nitrogen atmosphere represents the measured amount of oxygen atoms in a fully oxidized stoichiometric surrounding. The binding energy side of the O1s spectrum at 531.2570.20 eV is associated with O2 ions in the oxygen-deficient regions within the matrix of ZnO[15]. Therefore, the change in this peak intensity (O1s at 531.2570.20 eV) may be partly correlated with the variation in the concentration of oxygen vacancies. Dependence of the relative intensity ratio of this component on O1s at 531.2570.20 eV can be further depicted inFig. 5(b). It was found that the concentration of oxygen vacancies de-creases with an increase of the N-implanted fluence in ZnO film but this phenomenon is more obvious for the samples annealed in oxygen atmosphere. This indicates that part of nitrogen ions have occupied at the oxygen vacancies (sites). Based on above-mentioned results, it can be inferred that the amount of both zinc interstitials and concentration of antisite NOwould increase in proportion to the

ﬂuence of nitrogen ions for the N-implanted ZnO ﬁlms annealed at 850 1C. The amount of both zinc interstitials and concentration of antisite NO

would increase in proportion to the ﬂuence of nitrogen ions. non-implant 5E 12 1E 14 5E 15 non-implant 5E 12 1E 14 5E 15 Intensity (Arb.) Intensity (Arb.) 350 400 450 500 550 600 650 Wavelength (nm) 350 400 450 500 550 600 650 Wavelength (nm) (a) (b)

Fig. 4. Dependence of ﬂuences conditions on the room temperature PL spectra of the annealed ZnO ﬁlms treated at 850 1C in (a) N2and (b) O2atmospheres.

(6)

However, if the annealing atmosphere was exchanged and the oxygen atmosphere was used for annealing the N-implanted ZnO, the number of oxygen vacancies existing in the as-deposited ZnO ﬁlm will be further decreases probably because the effective incorporation of O atom diminishes those donor levels. Therefore, higher (0 0 2) peak intensity and better crystallinity were obtained for the ZnO ﬁlms annealed in oxygen

atmosphere compared to those annealed in nitro-gen atmosphere (comparingFig. 2withFig. 3). In this condition, part of the implanted nitrogen ions are probably located as the interstitial defects (Ni)

in the ZnO films because of limited available oxygen vacancies. Therefore, when the N-im-planted ZnO films were imN-im-planted with a lower dose of 5 1012cm2nitrogen ions which is below the solubility limit, some of nitrogen ions could be incorporated into the oxygen sites to reduce the intrinsic defect and restore the radiation damage. Therefore, the peak intensity of NBE emission was slightly enhanced by the implantation of nitrogen ions. Furthermore, as comparing Fig. 4(a) with Fig. 4(b), it is clear that for the same lower N-implanted concentration, i.e. 5 1012fluence, a stronger NBE peak intensity was detected in the N-implanted ZnO film annealed in oxygen atmo-sphere compared to that annealed in nitrogen atmosphere. However, with an increase of im-planted nitrogen concentration, more imim-planted nitrogen ions were incorporated into the ZnO matrix and the structure defects were further increased. As the implanted fluence exceeds 1 1014cm2 i.e., 5 1015cm2, a broader deep level peak around 532 nm due to high levels of nitrogen incorporation was observed in the sample annealed in oxygen atmosphere compared to that annealed in nitrogen atmosphere because a large amount of nitrogen ions could be dissolved in the ZnO films as the interstitial defects and the produced unstable defects would deteriorate the crystallinity and become some nonradiative center to absorb the light emission during photolumines-cence process.

4. Conclusion

The ZnO films implanted with nitrogen ions from 5 1012to 5 1015cm2show high preferred orientation (c-axis) films with strong NBE emis-sion property. It was found that there exists a solubility issue of nitrogen in ZnO films. Further-more, the annealing atmosphere plays an impor-tant role in crytallinity and photoluminescence of nitrogen-implanted ZnO films. When annealed in nitrogen atmosphere, the peak intensity of near

Relative intensity (a.u.)

Relative intensity

528 530 532 534

Binding energy (eV)

0.40 0.35 0.30 0.25 0.20 0.15 non-implanted N+5E12 cm -2 N+1E14cm -2 N+5E15cm -2 N2 atmosphere O2 atmosphere Fluence (cm-2₎ (a) (b)

Fig. 5. (a) Displays the O1s fitted spectra of non-implanted sample. (b) Dependence of relative intensity of the fitted components centered at O1s 531.25_{70.2 eV, on N-implanted} fluence in different atmospheres.

(7)

band-edge (NBE) emission remarkably decreases with the increase of concentration of implanted nitrogen. However, when the ZnO films were implanted with 5 1012cm2 and annealed in oxygen atmosphere, the optical properties are improved probably because the effective incor-poration of O atom diminishes those donor levels (oxygen vacancies) and the crystallinity is also improved due to implanted nitrogen. However, excess nitrogen would reduce the crystallinity and promote the formation of the deep-level emission due to high amount of intrinsic and structure defects. This study reveals that it is possible to control the crystallinity and the NBE emission of the ZnO films by using different annealing atmo-spheres and changing the nitrogen-implanted fluence.

Acknowledgments

The authors gratefully acknowledge the ﬁnan-cial support of the National Science Council of Taiwan in the Republic of China through Contract NSC-94-2216-E-009-016.

References

[1] T. Soki, Y. Hatanaka, D.C. Look, Appl. Phys. Lett. 76 (2000) 3257.

[2] K. Hummer, Phys. Stat. Sol. 56 (1973) 249.

[3] P. Zu, Z.K. Tang, G.K.L. Wong, M. Kawasaki, A. Ohtomo, H. Koinuma, Y. Segawa, Solid State Commun. 103 (1997) 459.

[4] D.M. Bagnall, Y.F. Chen, Z. Zhu, T. Yao, Appl. Phys. Lett. 70 (1997) 2230.

[5] T. Yamamoto, H. Yoshida, Physica B 302, 303 (2001) 155. [6] K. Ogata, K. Sakurai, Sz. Fujita, Sg. Fujita, K.

Matsushige, J. Crystal Growth 214/215 (2000) 312. [7] K. Minegishi, Y. Koiwai, Y. Kikuchi, K. Yano, M.

Kasuga, A. Shimizu, Jpn J. Appl. Phys. 36 (1997) L1454. [8] Y.R. Ryu, S. Zhu, D.C. Look, J.M. Wrobel, H.M. Jeong,

H.W. White, J. Crystal Growth 216 (2000) 330.

[9] C.H. Park, S.B. Zhang, S.H. Wei, Phys. Rev. B 66 (2002) 3202.

[10] M. Komatsu, N. Ohashi, I. Sakaguchi, S. Hishita, H. Haneda, Appl. Surf. Sci 189 (2002) 349.

[11] D.C. Park, I. Sakaguchi, N. Ohashi, S. Hishita, H. Haneda, Appl. Surf. Sci 203, 204 (2003) 359. [12] A.N. Georgobiani, A.N. Gruzintsev, V.T. Volkov,

M.O. Vorobiev, V.I. Demin, V.A. Dravin, Nucl. Instrum. Methods Phys. Res. A 514 (2003) 117.

[13] S.M. Myers, J. Vac. Sci. Technol. 15 (1978) 1650. [14] C.C. Lin, C.S. Hsiao, S.Y. Chen, S.Y. Cheng, J.

Electro-chem. Soc. 151 (2004) G285.

[15] J.C.C. Fan, J.B. Goodenough, J. Appl. Phys. 48 (1977) 3524.