In-implanted ZnO: Controlled degenerate surface layer

Wright State University CORE Scholar Physics Faculty Publications Physics 5-1-2009 In-Implanted ZnO: Controlled Degenerate Surface Layer David C. Look Wright State University - Main Campus, [email protected] Gary C. Farlow Wright State University - Main Campus, [email protected] F. Yaqoob L. H. Vanamurthy M. Huang Follow this and additional works at: https://corescholar.libraries.wright.edu/physics Part of the Physics Commons Repository Citation Look, D. C., Farlow, G. C., Yaqoob, F., Vanamurthy, L. H., & Huang, M. (2009). In-Implanted ZnO: Controlled Degenerate Surface Layer. Journal of Vacuum Science & Technology B, 27 (3), 1593-1596. https://corescholar.libraries.wright.edu/physics/4 This Article is brought to you for free and open access by the Physics at CORE Scholar. It has been accepted for inclusion in Physics Faculty Publications by an authorized administrator of CORE Scholar. For more information, please contact [email protected]. In-implanted ZnO: Controlled degenerate surface layer D. C. Looka兲 Semiconductor Research Center, Wright State University, Dayton, Ohio 45435 and Materials and Manufacturing Directorate, Air Force Research Laboratory, Wright-Patterson Air Force Base, Ohio 45433 G. C. Farlow Department of Physics, Wright State University, Dayton, Ohio 45435 F. Yaqoob Department of Physics, University at Albany, Albany, New York 12222 L. H. Vanamurthy and M. Huang College of Nanoscience and Engineering, University at Albany, Albany, New York 12203 共Received 21 November 2008; accepted 2 February 2009; published 28 May 2009兲 In was implanted into bulk ZnO creating a square profile with a thickness of about 100 nm and an In concentration of about 1 ⫻ 1020 cm−3. The layer was analyzed with Rutherford backscattering, temperature-dependent Hall effect, and low-temperature photoluminescence measurements. The implantation created a nearly degenerate carrier concentration n of about 2 ⫻ 1019 cm−3, but with a very low mobility ␮, increasing from about 0.06 cm2 / V s at 20 K to about 2 cm2 / V s at 300 K. However, after annealing at 600 ° C for 30 min, n increased to about 5 ⫻ 1019 cm−3, independent of temperature, and ␮ increased to about 38 cm2 / V s, almost independent of temperature. Also, before the anneal, no excitons bound to neutral In donors, called I9 in literature, were observed in the photoluminescence spectrum; however, after the anneal, the I9 line at 3.3568 eV was by far the dominant feature. Analysis of the Hall-effect data with a parametrized, two-layer model showed that the conduction before the anneal was mainly due to very high concentrations of native donors and acceptors, produced by the implantation, whereas the conduction after the anneal was due to In ions that were nearly 100% activated. These results show that strongly degenerate conductive layers with designed profiles can be created in ZnO with implantation and relatively low-temperature anneals. © 2009 American Vacuum Society. 关DOI: 10.1116/1.3089375兴 I. INTRODUCTION Ion implantation 共II兲 is an important dopant technology in many semiconductor systems.1 It can be applied directly to bulk substrates and allows precise control of donor and acceptor concentrations and profiles. However, the main disadvantage of II is that high annealing temperatures are usually necessary to reduce lattice damage caused by the implantation, and these high temperatures sometimes alter the dopant profile and lead to other deleterious effects.2,3 The rule of thumb is that the required annealing temperature TA,crit should be about 2 / 3 of the melting temperature, and this would give a TA,crit of about 1275 ° C in ZnO. Unfortunately, some researchers found that even lower annealing temperatures of 1000– 1200 ° C can actually cause a heavily damaged implanted layer to partially or totally evaporate,3 which would seem to limit applications of II in ZnO. Still, about 300 papers on this subject have been published so far, according to the ISI Web of Science. The reason for this continued strong interest is that successful II in ZnO would reap large dividends, as is the case in almost every important semiconductor material. However, only a very few of these former investigations have involved In,4,5 the subject of this study. a兲 Electronic mail: [email protected] 1593 J. Vac. Sci. Technol. B 27„3…, May/Jun 2009 An important issue in any implantation study is potential diffusion of the dopant during the high-temperature anneal necessary to eliminate the lattice damage. With regard to In diffusion in ZnO, it is known that implanted In diffuses in hydrothermally grown bulk ZnO with a diffusion coefficient of about 1.1 exp共−2.7 eV/ kT兲 cm2 / s,4 which translates to diffusion lengths of about 8, 52, 230, 780, and 2200 nm in 30 min at annealing temperatures of 600, 700, 800, 900, and 1000 ° C, respectively. Our results discussed below are in qualitative agreement, showing relatively weak diffusion at 700 ° C, and very strong diffusion at 1000 ° C. Fortunately, however, we find that a 600 ° C anneal is sufficient to create a highly degenerate n-type layer with a relatively high mobility. II. EXPERIMENTAL CONSIDERATIONS The substrate chosen for implantation, Q8-14d, was a 5 ⫻ 5 ⫻ 0.5 mm3 c-axis-perpendicular plate of hydrothermally grown ZnO purchased from Tokyo Denpa.6 Substrates of this type normally have good structural quality, as evidenced by typical x-ray rocking curve full width at half maximum of about 20 arc sec.7 The as-grown room-temperature resistivity, mobility, and carrier concentration were 72.8 ⍀ cm, 204 cm2 / V s, and 4.21⫻ 1014 cm−3, respectively. The sample was implanted at 350 ° C on the Zn face 关0001兴 with 1071-1023/2009/27„3…/1593/4/$25.00 ©2009 American Vacuum Society Downloaded 24 Sep 2012 to 130.108.121.217. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions 1593 1594 Look et al.: In-implanted ZnO: Controlled degenerate surface layer FIG. 1. 共Color online兲 O-interstitial profiles. TRIM 1594 calculations of In, Zn-interstitial, and four different energy/dose combinations: 8 ⫻ 1013 ions/ cm2 at 40 keV, 1.2⫻ 1014 ions/ cm2 at 100 keV, 1.6 ⫻ 1014 ions/ cm2 at 200 keV, and 6.4⫻ 1014 ions/ cm2 at 350 keV, in an attempt to produce a square In profile of about 1020 cm−3 in a 100 nm layer. The actual In profile was calculated by using the TRIM program,8 which employs Monte Carlo simulations of quantum-mechanical-based ion/ atom collisions in an amorphous material. The calculated profile was close to the desired profile, except for a tail that extended from about 100 to 150 nm 共cf. Fig. 1兲. Rutherford backscattering 共RBS兲 measurements were carried out with a 2 MeV He+ beam, in both random and aligned 共c-axis兲 directions. Temperature-dependent Hall-effect 共T-Hall兲 measurements were performed over the range of 15– 320 K with a LakeShore 7507 system. Finally, 4 K photoluminescence 共PL兲 was excited with a 30 mW 共⬃30 W / cm2 at sample position兲, 325 nm HeCd laser and dispersed with a 1.25 m spectrometer having a resolution of about 0.01 meV in the near-band-edge region. No strong color changes in the implanted layer were noted by visual inspection immediately after implantation, although it was perhaps somewhat more opaque. However, the unimplanted sides of some of the samples sometimes showed a yellowish tinge, evidently from effects of the 2 MeV He+ beam used for RBS measurements. Lowtemperature PL measurements in these areas revealed new donor-bound-exciton lines, evidently related to point defects. Analysis of these effects will be presented separately. III. RUTHERFORD BACKSCATTERING MEASUREMENTS AND TRIM CALCULATIONS Figure 2 displays channeling 共c-axis-aligned兲 RBS measurements for the sample before implantation, and both random and aligned RBS measurements after implantation. Analysis of these results showed that about 88% of the In ions were aligned along the c axis 共but not necessarily substitutional兲 immediately after the implantation. Further RBS measurements, with the He+ beam at an angle of 32° with respect to the c axis, showed that up to about 47% of the ions were substitutional on Zn sites, before any postimplantation annealing. The spatial profile of the In ions, calculated from FIG. 2. 共Color online兲 Rutherford backscattering measurements on sample Q8-14d, before and after implantation. the TRIM program, is shown in Fig. 1. The profiles of displaced Zn and O atoms can also be determined from this program, if the displacement energies Ed共Zn兲 and Ed共O兲 are known. Values of 34 and 44 eV have been recently calculated for Ed共Zn兲 and Ed共O兲, respectively, using molecular dynamics simulations.9,10 Although TRIM does not allow for vacancy/interstitial recombinations, we made an ad hoc assumption that any pairs less than two nearest-neighbor lattice distances apart will recombine, and the resulting profiles of ZnI and OI are also shown in Fig. 1. 共Note that the profiles of VZn and VO, not shown, are approximately the same as those of Znl and Ol, respectively, in this calculation.兲 Even though this model is very rough, still it is clear that the defect concentrations after implantation are much larger than the In concentration. However, the picture is even more complicated because it is known from theoretical studies that Znl and Ol should be mobile at room temperature,11 and thus it is likely that they will form stable complexes, with each other and with impurities of opposite charge. More accurate theoretical modeling of defect profiles is underway. IV. PHOTOLUMINESCENCE Near-band-edge, 4 K PL spectra of sample Q8-14d after implantation, and after a 600 ° C anneal in flowing N2 for 30 min, are presented in Fig. 3. Also shown is the spectrum of an adjacent sample, Q8-14a, before any implantation or anneal. As is typical in high-quality ZnO the dominant features are excitons bound to neutral donors, mostly in the 3.357– 3.363 eV range. Meyer et al.12 have summarized results for a large number of donor-bound exciton 共D0X兲 lines observed in vapor-phase-grown ZnO; some of the most common lines are I4共H兲 at 3.3628 eV, I6共Al兲 at 3.3608 eV, I8共Ga兲 at 3.3598 eV, and I9共In兲 at 3.3567 eV. In Q8-14a, our unimplanted, unannealed control sample, there are three clearly observed D0X lines, at 3.3625, 3.3612, and 3.3602 eV. We attribute the latter two lines to Al and Ga, respectively, recognizing that the 0.4 meV difference with respect to Meyer’s results can arise from several factors: 共1兲 sample strain, 共2兲 inaccurate spectrometer calibration, and 共3兲 the use of different wavelength/energy conversion fac- J. Vac. Sci. Technol. B, Vol. 27, No. 3, May/Jun 2009 Downloaded 24 Sep 2012 to 130.108.121.217. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions 1595 Look et al.: In-implanted ZnO: Controlled degenerate surface layer 1595 FIG. 3. 共Color online兲 Photoluminescence spectra of unimplanted ZnO sample Q8-14a, In-implanted sample Q8-14d, and also Q8-14d after a 600 ° C anneal in flowing N2 for 30 min. FIG. 4. 共Color online兲 Mobilities of unimplanted ZnO sample Q8-14a, Inimplanted sample Q8-14d, and also Q8-14d after a 600 ° C anneal in flowing N2 for 30 min. The solid lines are only a guide to the eye. tors. 关Our conversion factor was E共eV兲 = 12 395/ ␭共Å兲. Note that the use of 12 394 instead of 12 395 would result in a 0.3 meV difference!兴 The 3.3625 eV line may well be due to some form of H since it disappears after the 600 ° C anneal, as expected for H in ZnO.13 Note that no I9共In兲 is observed in our control sample on the linear scale of Fig. 3. After implantation, sample Q8-14d displays what appears to be a broad D0X line centered at 3.361 eV, although it actually is better modeled as a superposition of several sharp lines, to be discussed in detail elsewhere. The most surprising features, however, are several new lines between 3.33 and 3.35 eV, which we tentatively attribute to defect complexes. After a 400 ° C anneal 共data not shown兲, these defectrelated lines disappear, and after a 600 ° C anneal, the I9共In兲 line at 3.3568 eV finally becomes dominant, much larger than the I6共Al兲 and I8共Ga兲 lines, which return with the same relative strengths that had existed before the implantation. To summarize, the PL spectrum after the 600 ° C anneal is indicative of a very large In donor concentration, and relatively much smaller Al and Ga donor concentrations. about 6.7⫻ 1020 and 6.5⫻ 1020 cm−3, respectively, much higher than the implanted In concentration of about 1 ⫻ 1020 cm−3. This means that the electrical characteristics of the implanted sample, especially the mobility, are controlled by point defects, not by the In itself; details of this subject will be discussed elsewhere. Our main interest here is the electrical properties of the implanted sample after an anneal at 600 ° C, i.e., sample I/A. As seen in Fig. 5, the carrier concentration of sample I/A is nearly degenerate, varying by only a few percent over the range of 15– 320 K. The mobility is not quite as degenerate, varying from about 29 cm2 / V s at 14 K, to about 37 cm2 / V s at room temperature. However, if we treat both n and ␮ as degenerate, using the degenerate form of BrooksHerring theory17 to describe ␮, we get a donor concentration NDs = 1.4⫻ 1020 cm−3 and acceptor concentration NAs = 1.0 ⫻ 1020 cm−3, at 14 K. If we instead apply the same model at V. HALL-EFFECT MEASUREMENTS The temperature-dependent mobilities and carrier concentrations for the unimplanted 共UI兲, implanted 共I兲, and implanted/annealed 共I/A兲 samples are shown in Figs. 4 and 5, respectively. The resulting curves are fitted with a two-layer model,14–16 with one layer representing the bulk region, comprising most of the 0.5 mm thick sample, and the other layer, the implanted region, known to be about 100 nm thick in this case 共cf. Fig. 1兲. If both layers are contributing to the conductivity, then the apparent carrier concentration n will go through a minimum,16 as appears to be the case for sample UI, although the surface layer in this case is present in the as-grown crystal and is not due to implantation. However, for samples I and I/A, there is no apparent minimum in n, and in fact the electrical properties are totally dominated by a single layer, resulting from the implantation. A preliminary analysis of this layer, treating n as degenerate, and the mobility ␮ as nondegenerate, gives donor and acceptor concentrations of FIG. 5. 共Color online兲 Carrier concentrations of unimplanted ZnO sample Q8-14a, In-implanted sample Q8-14d, and also Q8-14d after a 600 ° C anneal in flowing N2 for 30 min. The solid lines are only a guide to the eye. The factor of 5000 on the top two curves is necessary because all three curves have been normalized to the total sample thickness, 0.05 cm, whereas the top two curves should be normalized to the implanted-layer thickness, 100 nm. Thus, e.g., the top curve has a nearly constant carrier concentration of 9 ⫻ 1015 cm−3共5000兲 = 4.5⫻ 1019 cm−3. JVST B - Microelectronics and Nanometer Structures Downloaded 24 Sep 2012 to 130.108.121.217. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions 1596 Look et al.: In-implanted ZnO: Controlled degenerate surface layer room temperature, the numbers are 1.2⫻ 1020 and 0.7 ⫻ 1020 cm−3, respectively. It is reasonable to consider these two fits as constituting a range of potential values for NDs and NAs, i.e., NDs = 共1.3⫾ 0.1兲 ⫻ 1020 cm−3, and NAs = 共0.8⫾ 0.2兲 ⫻ 1020 cm−3. VI. DISCUSSION In this work, we have attempted to create a flat profile of 1 ⫻ 1020 cm−3 In donors from the surface to a depth of 100 nm, and have shown that the actual donor concentration after annealing is NDs = 共1.3⫾ 0.1兲 ⫻ 1020 cm−3. This success has been obtained with a very simple, low-cost, and robust process: implantation of In at 350 ° C and annealing at 600 ° C in flowing N2 for 30 min. Note that these conditions represent a first attempt, and are not optimized. For example, it should be possible to vary the implantation and annealing temperatures to reduce the acceptor concentration NAs and thus increase the mobility. Nevertheless, a mobility of about 40 cm2 / V s has already been achieved at room temperature, and this mobility is quite respectable for a carrier concentration of about 5 ⫻ 1019 cm−3. Thus, we have demonstrated that ion implantation is a viable technique for the controlled creation of highly conducting layers in ZnO. ACKNOWLEDGMENTS The authors wish to thank T. A. Cooper for the Hall-effect measurements, W. Rice for the photoluminescence measure- 1596 ments, and B. Claflin and Z.-Q. Fang for helpful discussions. Support is gratefully acknowledged from the following sources: AFOSR Grant No. FA9550-07-1-0013 共K. Reinhardt兲, NSF Grant No. DMR0513968 共L. Hess兲, DOE Grant No. DE-FG02-07ER46389 共R. Kortan兲, ARO Grant No. W911NF-07-D-0001/Task07275 共M. Gerhold兲, and AFRL Contract No. FA8650-06-D-5401 共D. Silversmith兲. 1 S. J. Pearton, D. P. Norton, K. Ip, Y. W. Heo, and T. Steiner, Prog. Mater. Sci. 50, 293 共2005兲. 2 S. O. Kucheyev, J. S. Williams, C. Jagadish, J. Zou, C. Evans, A. J. Nelson, and A. V. Hamza, Phys. Rev. B 67, 094115 共2003兲. 3 V. A. Coleman, H. H. Tan, C. Jagadish, S. O. Kusheyev, and J. Zou, Appl. Phys. Lett. 87, 231912 共2005兲. 4 I. Sakaguchi, D. Park, Y. Takata, S. Hishita, N. Ohashi, H. Haneda, and T. Mitsuhashi, Nucl. Instrum. Methods Phys. Res. B 206, 153 共2003兲. 5 D. Park, I. Sakaguchi, N. Ohashi, S. Hishita, and H. Haneda, Appl. Surf. Sci. 203–204, 359 共2003兲. 6 Tokyo Denpa Co., Ltd., 5-6-11 Chuo, Ohta-ku, Tokyo 143-0024, Japan. 7 K. Maeda, M. Sato, I. Niikura, and T. Fukuda, Semicond. Sci. Technol. 20, S49 共2005兲. 8 J. P. Biersack and L. G. Haggmark, Nucl. Instrum. Methods 174, 257 共1980兲. 9 D. C. Look, G. C. Farlow, P. Reunchan, S. Limpijumnong, S. B. Zhang, and K. Nordlund, Phys. Rev. Lett. 95, 225502 共2005兲. 10 J. Nord, K. Nordlund, and J. Keinonen, Phys. Rev. B 68, 184104 共2003兲. 11 P. Erhart and K. Albe, Appl. Phys. Lett. 88, 201918 共2006兲. 12 B. K. Meyer et al., Phys. Status Solidi B 241, 231 共2004兲. 13 N. H. Nickel and K. Fleischer, Phys. Rev. Lett. 90, 197402 共2003兲. 14 D. C. Look, H. L. Mosbacker, Y. M. Strzhemechny, and L. J. Brillson, Superlattices Microstruct. 38, 406 共2005兲. 15 D. C. Look, Surf. Sci. 601, 5315 共2007兲. 16 D. C. Look, J. Appl. Phys. 104, 063718 共2008兲. 17 D. C. Look, Electrical Characterization of GaAs Materials and Devices 共Wiley, New York, 1989兲, p. 76. J. Vac. Sci. Technol. B, Vol. 27, No. 3, May/Jun 2009 Downloaded 24 Sep 2012 to 130.108.121.217. Redistribution subject to AVS license or copyright; see http://avspublications.org/jvstb/about/rights_and_permissions