ZnO Doping and Co-doping Paradigm and Properties

Journal of The Electrochemical Society, 157 共9兲 H891-H895 共2010兲

H891

0013-4651/2010/157共9兲/H891/5/$28.00 © The Electrochemical Society

ZnO Doping and Co-doping Paradigm and Properties K. Shtereva,a,b,z I. Novotny,b,* V. Tvarozek,b P. Sutta,c A Vincze,b,d and A. Pullmannovab a

Department of Electronics, University of Rousse, BG-7017 Rousse, Bulgaria Department of Microelectronics, Slovak University of Technology, 81219 Bratislava, Slovakia c New Technologies - Research Center, West Bohemia University, 30614 Plzen, Czech Republic d International Laser Centre, 84104 Bratislava, Slovak Republic b

Here, we report on experimental studies of the role of doping and co-doping on the properties of ZnO thin films deposited by radio-frequency diode sputtering at varying nitrogen content 共0 ÷ 100%兲 in the sputtering Ar/N2 gas. Co-doping improved the crystalline structure, and ZnO:Al:N films maintain a c-axis texture in the direction of the surface normal. Depending on the N2 content, the estimated crystallite size varies from 8 to 37 nm for ZnO:N and 21–33 nm for ZnO:Al:N. Nitrogen doping results in an increased absorption around the bandedge and the bandgap narrowing 共Eg ⬍ 3.2 eV兲. © 2010 The Electrochemical Society. 关DOI: 10.1149/1.3459900兴 All rights reserved. Manuscript submitted January 11, 2010; revised manuscript received May 14, 2010. Published July 23, 2010.

Zinc oxide 共ZnO兲 is a wide bandgap 共⬃3.37 eV兲, group II–VI compound semiconductor with a hexagonal wurtzite structure and diverse properties that depend on doping, including a type of conductivity 共n-type or p-type兲, which can range from metallic to insulating, high transparency, piezoelectric, ferromagnetic, and sensing qualities. Therefore, this versatile material has been intensively studied for a wide range of applications from optoelectronic and transparent electronic devices,1 surface and bulk acoustic wave devices and piezoelectric transducers,2 and spintronics3 to chemical and gas sensors4 and solar cells.5 Moreover, ZnO has a widely abundant and inexpensive source material and is an environment friendly material. Low cost, high resistance to hydrogen plasma, and ease of patterning make ZnO an alternative to indium tin oxide in transparent conducting oxide applications such as transparent electrodes for liquid crystal displays, in organic light emitting diodes, and in photovoltaic solar cells.6 ZnO has been studied as a semiconductor for light emitting diodes and UV semiconductor lasers due to its advantages over gallium nitride 共GaN兲, namely, a larger exciton binding energy 共⬃60 meV兲 and availability of high quality ZnO single crystals. The fabrication of ZnO-based devices and the progress of transparent electronics are predetermined from the advance in quality and processing of both n-type and p-type ZnO materials. Among numerous methods for ZnO thin-film deposition, sputtering is often used due to its important industrial advantages, namely, simplicity, low processing temperature, and low cost. High quality n-type ZnO with electron concentrations of ⬃1021 cm−3 has been achieved using aluminum 共Al兲, gallium 共Ga兲, or indium 共In兲 as donor dopants.7,8 Much effort has been put in the development of transparent p-type ZnO materials with high hole concentrations needed for transparent electronics, organic optoelectronics applications, and photovoltaics. Among the possible acceptors, nitrogen has been one of the most studied potential p-type dopant, and p-type ZnO via nitrogen doping has been reported.9,10 Nitrogen dopant source is easy to get, economically lucrative, and nontoxic. The novelty of our approach is in the use of radio-frequency 共rf兲 diode reactive sputtering toward the goal of obtaining a p-type ZnO material via nitrogen doping. In general, sputter deposition is characterized by complex processes occurring 共i兲 on the target bombarded by energetic ions, 共ii兲 in low temperature plasma, and 共iii兲 on the substrate surface and the growing film. Thin-film growth is influenced by the kinetic energy of coating species on the substrate. In addition to substrate temperature, a total energy flux that depends mainly on the amount and the energy of 共i兲 sputtered coating species 共1–20 eV兲, 共ii兲 energetic neutral working gas atoms neutralized and reflected at the target 共ⱕ100 eV兲, 共iii兲 energetic secondary electrons emitted from the target 共⬃100 eV兲, and 共iv兲 negative ions coming from the working gas

* Electrochemical Society Active Member. z

E-mail: [email protected]

plasma or target 共ⱕ1 keV兲 is an important parameter for film microstructure evolution. The effects of high energy particle bombardment on ZnO films 共O2, Ar, O, and O−兲 during diode and magnetron sputtering have been known for a long time.11 Rf diode sputtering in Ar/N2 discharge provides a very reactive environment and supports the formation of charged and neutral species 共ZnO, ZnO+, Zn, Zn+, Ar, Ar+ , O, O−, e−, O+2 , N, N+, and N+2 兲 that can more or less affect the growing film. We found previously12 that the crystalline structure of sputtered ZnO films and, hence, film properties are determined by both the substrate temperature and the total energy flux density. When rf diode sputtering in a low pressure region 共p ⱕ 1.3 Pa兲 is used for film growth, the mean free path of sputtered particles 共⬃10−2 m兲 is comparable with the distance between the target and the substrate holder 共usually 4.10−2 m兲, i.e., we can assume a “collision-less” regime, particularly for energetic particles that passed through the rf discharge. In rf diode sputtering, three important effects with regard to p-type doping of ZnO can be presented. The first is an increased molecular dissociation that can lead to the formation of nitrogen atoms 共N2 ↔ 2N兲 and NO, whose incorporation in the growing film forms desirably for p-type doping nitrogen acceptors 共NO兲. Moreover, N atoms can form AlN/GaN molecules when co-dopants 共Al and Ga兲 are added to the growth ambient. Thus, the concentration of N that can reconvert to N2 is reduced and more N atoms reach the growing film and substitute for O leading to p-type ZnO. Second, in rf diode sputtering, the substrate/growing film are subjected to energetic particle bombardment 共sputtered particles, reflected neutrals, and negative ions兲 that can affect the formation of acceptor or donor complexes in the growing film. Third, there is a significant contribution of secondary electron bombardment to the atomic scale heating of the film when it is prepared by the rf diode sputtering. Our aim was to exploit the special technological conditions of rf diode sputtering 共in comparison, e.g., with magnetron sputtering兲 for developing a p-type doping paradigm and for investigating the role of doping and co-doping on the properties of rf-sputtered ZnO thin films deposited at varying nitrogen contents 共0 ÷ 100%兲 in the sputtering Ar/N2 gas. Experimental The ZnO:N and ZnO:Al:N thin films were deposited on Corning glass 7059 substrates by rf diode sputtering in an Ar/N2 working gas. The percentage of nitrogen in the sputtering gas varied from 0 to 100%. Other deposition parameters, such as base pressure 共2 ⫻ 10−5 Pa兲, working gas pressure 共1.3 Pa兲, and sputtering power 关500 W 共ZnO:N兲/400 W 共ZnO:Al:N兲兴, were maintained constant during the deposition process. The ZnO:N films were sputtered from a ZnO target 共purity 99.99%兲. The deposition rate decreased linearly with the increasing content of nitrogen 共from 25 to 75%兲 in the sputtering gas 共Fig. 1兲. A sintered ceramic target, a mixture of ZnO

Downloaded 06 Oct 2010 to 91.139.245.14. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

H892

Journal of The Electrochemical Society, 157 共9兲 H891-H895 共2010兲

Figure 1. Deposition rate as a function of nitrogen content.

共purity 99.99%兲 and Al2O3 共2 wt %, purity 99.99%兲, was utilized for the deposition of ZnO:Al:N thin films. The film thickness was measured by a Talystep instrument. Structural properties were assessed using X-ray diffraction 共XRD兲. The crystallographic structure of the ZnO:N thin films was analyzed on an AXS Bruker D8 powder diffractometer equipped with a two-dimensional detector 共Co K␣, ␭ = 0.179 nm兲. The XRD patterns of ZnO:Al:N thin films were recorded on a thin-film attachment of an X’pert Pro powder diffractometer 共Co K␣, ␭ = 0.154 nm兲. In both cases, a powdered alumina from the National Institute of Standards and Technology was used as an instrumental standard. Size-strain analysis 共microstrains and crystallite sizes兲 was performed using a method proposed by Langford based on the XRD line profile analysis. The depth profile of the films was measured with secondary-ion mass spectrometry 共SIMS兲 using a time of flight SIMS IV analyzer from ION TOF GmbH, Muenster. Optical spectrophotometry measurements were carried out on an Ava Spec-2048 fiber optic spectrometer from the UV region to the near-IR region. Results and Discussion XRD structure characterization.— XRD measurements of ZnO thin films deposited on Corning glass substrates show the growth of crystallites with different orientations as a result of nitrogen incorporation. Three diffraction lines typical for ZnO共100兲, 共002兲, and 共101兲 appear in the XRD patterns 共30° ⬍ 2␪ ⬍ 80°兲 of all ZnO:N films 共Fig. 2兲. The fourth diffraction line, 共110兲, is available only in the patterns of the films deposited at 25, 50, and 75% N2 in the sputtering gas. The diffraction lines’ position, full width at halfmaximum 共fwhm兲, and relative intensity vary with N2 content. A strong 共002兲 line and weak 共100兲 and 共101兲 features for undoped 共0% N2兲 films and those grown at 100% N2 provide evidence for their polycrystalline hexagonal structure with a preferential c-axis orientation perpendicular to the substrate. For nitrogen contents ranging from 25 to 75%, the 共002兲 line intensity decreases and its fwhm increases. The growth of the diffraction lines corresponding to the 共100兲, 共101兲, and 共110兲 crystal planes and the dominant 共100兲 line show that the ZnO:N thin films become more randomly orientated. XRD measurements reveal that changes in Ar/N2 ratio and nitrogen incorporation in ZnO, as proved by SIMS, have considerable effects on the crystallographic texture of the deposited films. The impurity species, most likely, interact differently with different crystallographic planes, promoting or hindering their growth, which causes changes in the microstructure. Moreover, nitrogen incorpora-

Figure 2. XRD patterns of ZnO:N thin films deposited on a Corning glass substrate at 0, 25, 50, 75, and 100% N2 in the growth ambient.

tion is accompanied with the formation of defect complexes composed of nitrogen and intrinsic defects 共NO–ZnO, NO–Zni, and NO–VO complexes兲, or composed purely of nitrogen 共NO–N2 O兲, which influence the structural and, hence, the electrical and optical properties of ZnO. According to first principles calculations,13 these complexes are donors in ZnO and compensate NO acceptors, making it difficult to obtain p-type conductivity. This conclusion is supported by our Hall measurement results, which were published elsewhere,14 and showed that changes in hole concentration and other electrical parameters with increasing nitrogen content corresponded to observed microstructure changes. The comparison of the XRD reference data for hexagonal ZnO, calculated using an APX 63 STRUC software,15 with those of ZnO:N shows that the 2␪ diffraction angles of the N2-doped films are slightly smaller than the standard values 共37.09, 40.22, 42.39, and 66.85°兲 for the 共100兲, 共002兲, 共101兲, and 共110兲 lines of undoped ZnO. Depending on the N2 content, the diffraction line positions shift lower for the ZnO:N films. According to the Bragg law, the decrease in the diffraction angle with increasing N2 content increases the interplanar spacing d, thus introducing stress into the film. Therefore, this shift can be explained by compressive lattice strains 共stresses兲 created during a sputtering process and can be quantitatively evaluated from the equation for biaxial lattice stress16 ␴1 + ␴2 = −

E d − d0 · ␮ d0

关1兴

where E is Young’s modulus, ␮ is Poisson’s ratio, d0 is the reference strain-free interplanar spacing, and d is the interplanar spacing obtained from the experiment. Crystallite size is estimated using Scherrer’s formula 具D典 =

␤Cf

K␭ cos ⌰

关2兴

where K = 2冑ln 2/␲ ⬇ 0.94 is the Scherrer’s constant, ␭ is the wavelength of the X-rays used, ␤fC is the pure 共physical兲 Cauchy component of the integral breadth of the line taken in radians, and ⌰ is the Bragg’s angle.17 For the ZnO:N films, the estimated crystallite size varies from 8 to 37 nm, depending on the N2 content in working gas 共Table I兲. The average microstrains are in the order of ⬃10−2 and were determined using the equation

Downloaded 06 Oct 2010 to 91.139.245.14. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 157 共9兲 H891-H895 共2010兲

H893

Table I. XRD data “size-strain” taken from (002) line. Nitrogen content 共%兲 0 25 50 75 100

Average crystallite size 共nm兲 ZnO:N

ZnO:Al:N

37 20 12 8 24

23 32 21 33 —

具␧典 =

f ␤G 4tg⌰

Average microstrains ZnO:N 1.0 1.6 1.3 1.2 1.2

⫻ ⫻ ⫻ ⫻ ⫻

10−2 10−2 10−2 10−2 10−2

ZnO:Al:N 8.3 4.5 9.6 9.0

⫻ 10−3 ⫻ 10−3 ⫻ 10−3 ⫻ 10−3 —

关3兴

f is the pure 共physical兲 Gaussian component of the integral where ␤G breadth of the line taken in radians.17 The XRD analysis indicates a c-axis preferential orientation in all ZnO:Al:N films, which is more or less expressed stronger depending on the N2 content in the sputtering gas 共Fig. 3兲. The preferential orientation is slightly more random for the films deposited at 25% N2, where 共100兲 and 共101兲 diffraction lines can be observed except for the 共002兲 diffraction line. In contrast to them, the ZnO:Al:N films deposited at 50 and 75% N2 show a strong preferential orientation of the crystallites. An improvement of texture was also caused by lower deposition rates 共Fig. 1兲, i.e., a better condition for the film growth. The diffraction lines are asymmetric, more likely due to the incorporation of Al and the formation of different phases 共AlO兲 at the interface, which causes broadening of the line toward the higher diffraction angles. The XRD “size-strain” data for ZnO:Al:N films are listed in Table I. The estimated grain size changes from 21 to 33 nm, and the microstrains vary from 4.5 ⫻ 10−3 to 9.6 ⫻ 10−3 with the N2 content. The improved crystalline structure of Al–N co-doped films results in better electrical properties of p-type ZnO:Al:N 共hole concentration of 7.8 ⫻ 1017 cm−3, mobility of 1 cm2 /Vs, and resistivity of 21 ⍀ cm兲 compared to p-type ZnO:N 共hole concentration of 3.6 ⫻ 1014 cm−3, mobility of 22 cm2 /Vs, and resistivity of 7.9 ⫻ 102 ⍀ cm兲.

SIMS depth profile.— The p-type conductivity in ZnO:N and ZnO:Al:N results from the nitrogen incorporation and formation of NO acceptors. Nitrogen incorporation was confirmed by SIMS measurements that were carried out using 共Au+兲 primary analysis ions. A Cs+ ion beam was used as a primary ion source for the sputtering

Figure 3. XRD patterns of ZnO:N:Al thin films deposited on a Corning glass substrate at 25, 50, and 75% N2 in the growth ambient.

Figure 4. Depth profile of ZnO:N prepared at 75% N2.

of various negative secondary ions 共O−, ZnO−, ZnN−, N−, NO−, and NO−2 兲. Figure 4 shows the depth profile of the main ionic species for nitrogen-doped ZnO prepared at 75% N2. In the figure, negative ions of 共NO兲− and 共NO2兲−, along with 共ZnO兲− and their uniform distribution from the surface to the interface can be observed. The negative nitrogen ions N− create a constant noise level intensity due to their low ionization. Therefore, N− cannot be used to evaluate N introduction in the film. Figure 5 shows nitrogen depth profile in co-doped ZnOAl:N films prepared at 75% N2. The SIMS depth profiles provide an evidence for the nitrogen introduction in the film, and the observed p-type features are due to the formation of nitrogen acceptors NO. Optical properties.— To investigate the optical properties of ZnO:N and ZnO:Al:N thin films prepared at different N2 contents, their transmittance was measured in the wavelength range of 200 ⬍ ␭ ⬍ 1000 nm. The absorption coefficient of direct bandgap semiconductors can be calculated using the formula ␣=

1 1 ln t T

关4兴

where t is the thickness of the film and T is the transmittance. The optical bandgap for allowed direct transitions can be estimated by extrapolating the linear portion of the square of the absorption coefficient against photon energy from the most commonly used equation for ZnO18-20

Figure 5. Depth profile of ZnO:Al:N prepared at 75% N2.

Downloaded 06 Oct 2010 to 91.139.245.14. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

H894

Journal of The Electrochemical Society, 157 共9兲 H891-H895 共2010兲

Figure 6. Optical transmittance spectra of ZnO:N films deposited at 0, 25, 50, 75, and 100% N2 in the growth ambient.

␣h␯ = A共h␯ − Eg兲1/2

关5兴

Figure 6 shows the transmission of undoped and N2-doped ZnO films 共including glass substrate兲 in the visible range of 300–900 nm. The transmittance of ZnO:N and ZnO:Al:N are compared in Fig. 7 The optical transmittance depends on the nitrogen content in the working gas. The undoped ZnO films and monodoped ZnO:Al films 共0% N2兲 are colorless, and their average transmittance including glass substrate is ⬎80%. Nitrogen doping results in an increased absorption around the bandedge of the ZnO:N and ZnO:Al:N films. Their color changes from yellow to brown with increasing N2 content, and their transmittance is reduced. The absorption in the visible region and coloration of the film, associated with nitrogen doping, have been assigned to deep N-related states in the bandgap of ZnO.21,22 The shift in the bandedge for nitrogen-doped films to higher wavelengths is due to the bandgap narrowing 共Fig. 8兲. Conclusion

Figure 8. Plot of the absorption squared 共␣hv兲2 vs phonon energy for undoped 共0% N2兲 and nitrogen-doped 共75% N2兲 ZnO thin films.

25–75% N2 become randomly oriented. Co-doping improves the crystalline structure. The dominant 共002兲 diffraction line, which appears in the XRD patterns of ZnO:Al:N thin films, reveals a strong preferential c-axis orientation in the direction of the surface normal. Aluminum incorporation and the formation of different phases 共AlO兲 at the interface cause asymmetry and broadening of this line toward the higher diffraction angles. Depending on the N2 content, the crystallite size varies from 8 to 37 nm for the ZnO:N films and 21–33 nm for ZnO:Al:N films. SIMS depth profiles confirm the nitrogen incorporation. Nitrogen doping results in an increased absorption around the bandedge, yellow to brown coloration of the ZnO:N and ZnO:Al:N films, and narrowing of the bandgap 共Eg ⬍ 3.2 eV兲. We observed p-type ZnO:N with hole concentration of 3.6 ⫻ 1014 cm−3, mobility of 22 cm2 /Vs, and resistivity of 7.9 ⫻ 102 ⍀ cm and p-type ZnO:Al:N with hole concentration of 7.8 ⫻ 1017 cm−3, mobility of 1 cm2 /Vs, and resistivity of 21 ⍀ cm. Acknowledgments

Nitrogen-doped and aluminum–nitrogen co-doped ZnO films were deposited by rf diode sputtering at different N2 contents 共ranging from 0 to 100%兲 in an Ar/N2 working gas. The XRD analysis indicates that undoped ZnO films 共0% N2兲 have a preferential c-axis orientation perpendicular to the substrate. Nitrogen doping causes the change in the growth direction, and the ZnO:N films deposited at

Slovak University of Technology assisted in meeting the publication costs of this article.

Figure 7. Comparison between the transmittance spectra of ZnO:Al films 共0% N2兲 and the transmittance spectra of ZnO:Al:N and ZnO:N films deposited at 75% N2 in the growth ambient.

1. H. Ohta and H. Hosono, Mater. Today, 7, 42 共2004兲. 2. Q. J. Wang, Ch. Pflügl, W. F. Andress, D. Ham, F. Capasso, and M. Yamanishi, J. Vac. Sci. Technol. B, 26, 1848 共2008兲. 3. G. Ji, Z. Zhang, Y. Chen, Sh. Yan, Y. Liu, and L. Mei, Acta Metall. Sin. (Engl. Lett.), 22, 153 共2009兲. 4. M. C. Carotta, A. Cervi, V. Natale, S. Gherardi, A. Giberti, V. Guidi, D. Puzzovio, B. Vendemiati, G. Martinelli, M. Sacerdoti, et al., Sens. Actuators B, 137, 164 共2009兲. 5. G. Ganguly, D. E. Carlson, S. S. Hegedus, D. Ryan, R. G. Gordon, D. Pang, and R. C. Reedy, Appl. Phys. Lett., 85, 479 共2004兲. 6. C. Guillén and J. Herrero, Thin Solid Films, 515, 640 共2006兲. 7. E. Fortunato, V. Assunção, A. Gonçalves, A. Marques, H. Aguas, L. Pereira, I. Ferreira, P. Vilarinho, and R. Martins, Thin Solid Films, 451–452, 443 共2004兲. 8. K. Shtereva, S. Flickyngerova, P. Sutta, M. Netrvalova, I. Novotny, and V. Tvarozek, J. Phys.: Conf. Ser., 100, 082021 共2008兲. 9. J. Lu, Y. Zhang, Z. Ye, L. Wang, B. Zhao, and J. Huang, Mater. Lett., 57, 3311 共2003兲. 10. X. Li, S. E. Ashera, S. Limpijumnong, B. M. Keyes, C. L. Perkins, T. M. Barnes, H. R. Moutinho, J. M. Luthera, S. B. Zhang, S.-H. Wei, et al., J. Cryst. Growth, 287, 94 共2006兲. 11. K. Tominaga, S. Iwamura, Y. Shintani, and O. Tada, Jpn. J. Appl. Phys. Part 1, 21, 688 共1982兲. 12. V. Tvarozek, I. Novotny, P. Sutta, S. Flickyngerova, K. Schtereva, and E. Vavrinsky, Thin Solid Films, 515, 8756 共2007兲. 13. E. Ch. Lee, Y. S. Kim, Y. G. Jin, and K. J. Chang, Physica B, 912, 308 共2001兲. 14. V. Tvarozek, K. Shtereva, I. Novotny, J. Kovac, P. Sutta, R. Srnanek, and A.

This work was supported by the MSMT Czech Republic project 1M06031 and in the frame of the Centre of Excellence CENAMOST 共VVCE 0049-07兲 with support of project VEGA 1/0220/09.

References

Downloaded 06 Oct 2010 to 91.139.245.14. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp

Journal of The Electrochemical Society, 157 共9兲 H891-H895 共2010兲 Vincze, Vacuum, 82, 166 共2007兲. 15. I. Kraus, in Struktura a Vlastnosti Krystalu, p. 221, Academia, Prague 共1993兲. 16. P. Šutta and Q. Jackuliak, in ASDAM’98 Proceedings of Second International Conference on Advanced Semiconductor Devices and Microsystems, J. Breza, Editor, Slovac University of Technology, Smolenice, p. 227 共1998兲. 17. R. Delhez, Th. H. de Keijser, E. J. Mittemeijer, and Z. Fresenius, Anal. Chem., 312, 1 共1982兲. 18. S. M. Sze, Physics of Semiconductor Devices, 2nd ed., pp. 40–41, John Wiley & Sons, New York 共1981兲.

H895

19. S. T. Tan, B. J. Chen, X. W. Sun, W. J. Fan, H. S. Kwok, X. H. Zhang, and S. J. Chua, J. Appl. Phys., 98, 013505 共2005兲. 20. C. Gümüş, O. M. Ozkendir, H. Kavak, and Y. Ufuktepe, J. Optoelectron. Adv. Mater., 8, 299 共2006兲. 21. F. Zhu-Ge, Z.-Z. Ye, L.-P. Zhu, J.-G. Lü, B.-H. Zhao, J.-Y. Huang, Z.-H. Zhang, L. Wang, and Z.-G. Ji, J. Cryst. Growth, 268, 163 共2004兲. 22. K. Kobayashi, Y. Kondo, Y. Tomita, Y. Maeda, and S. Matsushima, Appl. Surf. Sci., 253, 5035 共2007兲.

Downloaded 06 Oct 2010 to 91.139.245.14. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp