Structural modifications of GaN after cerium implantation

Research Article

Received: 22 March 2012 Revised: 6 June 2012 Accepted: 6 June 2012 Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI 10.1002/jrs.4143

Structural modifications of GaN after ceriumimplantationAbdul Majid,a* Jianjun Zhu,b Najam al Hassanc and Abdul Shakoord

Among the family of rare earth (RE) dopants, the doping of first member Ce into GaN is the least studied system. This article reportsstructure properties of Ce-doped GaN realized by technique of ion implantation. Ce ions were implanted intometal organic chemical
vapor deposition grown n- and p-GaN/sapphire thin films at doses 3� 1014 and 2� 1015 cm�2. X-ray diffraction scans and Ramanscatteringmeasurements exhibited expansion of lattice in the implanted portion of the samples. First order Raman scattering spectrashow appearance of several disorder-activated Raman scattering modes in addition to typical GaN features. A dose-dependent de-crease in intensity of E2 mode was observed in Raman the spectra of the implanted samples. Ultraviolet Raman spectra of implantedsamples show complete quenching of photoluminescence emission and appearance of multiple A1(LO) phonon scattering modesup to fifth order. Moreover, a decrease in intensity and an increase in line width of LO modes as a function of wavenumber wereobserved for implanted samples. Copyright © 2012 John Wiley & Sons, Ltd.
Supporting information may be found in the online version of this article.

Keywords: GaN; ion implantation; rare earth; lattice expansion; Raman scattering

* Correspondence to: Abdul Majid, Physics Department, University of Gujrat,Gujrat 50700, PakistanE-mail: [email protected]

a Physics Department, University of Gujrat, Gujrat, Pakistan

b Nano Fabrication Facility, Suzhou Institute of Nano-Tech and Nano-Bionics,Chinese Academy of Sciences, Suzhou, China

c Physics Department, Hazara University, Mansehra, Pakistan

d Physics Department, Bahauddin Zakariya University, Multan, Pakistan

Introduction

Rare earth (RE) metals offer an interesting class of dopants forsemiconductors due to having unfilled 4f subshell shielded fromenvironment by 5s and 5p filled subshells. These 4fn electrons areresponsible for 4f-4f and 4f-5d transitions to produce sharp emis-sion lines of wavelength, ranging from infrared to visible regionof spectrum, independent of host and temperature.[1] RE-dopedmaterials, due to having such properties, present variety ofapplications such as solid state laser, electroluminescent devices,phosphors, fiber optic telecommunication, data storage andmagnetic materials.[1–5] The possibility of using RE3+-dopedmaterials for use in light emitters due to intra-4f shell transitionsmotivated the researchers to use different hosts for RE doping. Inthe beginning, only narrow band gap semiconductors weretested for RE doping.[6] Therefore, the infrared light emittingdevices, based on the intra-4f electron transition between theground state and the first excited state of Er3+ were common atthat time.[7] A large number of Er3+-doped semiconductors canbe found in literature, e.g. Si, Ge, SiGe, SiC, GaAs etc. However,commercial implementation of early narrow gap semiconductorsdoped by RE was unsatisfactory because of thermal quenching ofdopant-related luminescence. This provoked the community tofind larger band gap semiconductors consequent upon fabrica-tion of many such materials in early 1990s. Gallium nitride is sucha versatile wide band gap semiconductor of III-nitride family,which has gone through a long journey starting from a researchgrade material up to the time when it is now an industrial/commercial grade material. Researchers introduced plenty ofdopants into GaN for its modifications to realize interestingproperties for use in devices. In addition to transition metalsand other dopants, RE-doped GaN can be extensively found inthe literature such as GaN : Pr3+,[8] GaN : Nd,[9] GaN : Sm3+,[10] GaN :Eu3+,[11] GaN : Tb3+,[12] GaN : Dy3+,[13] GaN : Ho3+,[14] GaN : Er3+,[11]

GaN : Tm3+[13,14] etc. The main advantage of GaN over small bandgap semiconductors was that it is resistive to thermal quenching

J. Raman Spectrosc. (2012)

of intra shell transitions.[15,16] Literature survey indicatesavailability of RE–GaN systems with nearly all members of REfamily except for Ce, which is having one 4f electron withthe largest radius. The Ce–GaN system is expected to presentinteresting properties as seen from other RE–GaN materials.[17]

This article reports on structural properties of Ce-implantedGaN thin films.

Materials and methods

Sample preparation

The wurtzite GaN/sapphire thin films in form of two types ofsamples, undoped GaN and Mg-doped p-type GaN, were usedfor this study. Because undoped GaN shows n-type conductivity,so these samples will be regarded as n-GaN (unintentionallyn-doped) in this article to keep them distinctive from p-GaN.Both types of the samples were implanted at room temperaturewith Ce ions at 300 keV with doses of 3� 1014 and 2� 1015 cm�2.The first set of samples was n-GaN for which as-grown andsamples implanted with doses 3� 1014 and 2� 1015 cm�2 werenamed as nCe, Ce1 and Ce2, respectively. Whereas, the secondset of samples consisted of p-GaN for which as-grown andsamples implanted with doses 3� 1014 and 2� 1015 cm�2 were

Copyright © 2012 John Wiley & Sons, Ltd.

A. Majid et al.

named as pCe, Ce3 and Ce4, respectively. The implanted ionspenetrated to a depth of 100 nm from the sample surface asestimated by Transport of Ions in Matter (TRIM) simulation. Foractivation of dopants, the samples were annealed using rapidthermal annealing at 900 �C for 30 s in nitrogen ambient whileplacing faced down on another GaN wafer in order to minimizethe possible decomposition.

X-ray diffraction measurements

High resolution X-ray diffraction (XRD) measurements were per-formed using Rigaku SLX-1A diffractometer to study the struc-tural properties of the materials.

Raman scattering measurements

Raman scattering (RS) measurements were carried out underbackscattering geometry at room temperature. First order Ramanspectra were recorded by using 514.5 nm (2.41 eV) line of argonlaser at spectral resolution of ~ 0.5 cm�1. The laser beam wasfocused up to a spot size of 1mm and scattered light was col-lected with a 100� objective lens. The scattered light wasdetected by Jobin Yvon T64000 spectrometer with a LN2-cooledcharge-coupled device. Ultraviolet (UV) Raman spectra weremeasured by using UV Renishaw micro-Raman system with a325 nm (3.841 eV) line of He–Cd laser as excitation source.

Results

X-ray diffraction

Figure S1 shows the XRD scans taken for as-grown andCe-implanted n-GaN and p-GaN samples in 2θ range of 20–100�.The scans show usual peaks of GaN and sapphire substrate andno new diffraction peak is observed in whole measurement range.X-ray diffraction scans showing (0002) GaN peaks for as-grown

and Ce-implanted n-GaN and p-GaN samples are shown inFig. S2. The panels given on right side of this figure show thermalannealing effects on (0002) peaks of as-grown and Ce-implantedn-GaN (with logarithmic intensity scale) indicating annealing-induced lattice recovery.[18]

The implanted samples exhibited additional peak structure onthe lower angle side of the main GaN peak. Samples Ce1 and Ce3exhibit a clear peak along with a shoulder peak on lower angleside of main GaN peak that represents a damaged layer. XRDspectra of the samples Ce2 and Ce4 exhibited a similar shoulderpeak but wider and less intense in comparison with that of thenew peak for previous sample.To find the Bragg position of these subsidiary peaks, we fitted

the Gaussian multi peaks as shown in four panels given in the leftside of Fig. S2. Subsidiary peaks are named as P1, P2 and P1*, P2*

for samples Ce1 and Ce2, respectively. Similarly, these peaksare named as P3, P4 and P3*, P4* for samples Ce3 and Ce4,

Table 1. Out-of-plane lattice constants corresponding to the additional X

Samples Ce1 Ce2

Peaks P1 P2 P1*

Lattice constants (Å) 5.195 5.189 5.191

wileyonlinelibrary.com/journal/jrs Copyright © 201

respectively. The values of out-of-plane lattice constant ‘c’were calculated corresponding to all these peaks using Bragg’slaw and are given in Table 1. This table indicates that latticeconstants corresponding to additional peaks are higher thanthat of unimplanted GaN (5.185 Å).

Raman scattering spectroscopy

The first order Raman spectra, carried out using visible line ofargon laser, for as-grown and Ce-implanted GaN samples aregiven in Fig. 1. Raman measurements were also performed onsapphire to find the contribution of the substrate in the Ramanspectra of GaN samples. Raman modes appearing at 378, 417,428, 447, 575 and 750 cm�1 are of sapphire.[18,19] As per grouptheory, wurtzite GaN has eight modes 2A1 + 2B+ 2E1 + 2E2 nearK = 0, out of which the six modes 1A1(TO) + 1A1(LO) + 1E1(TO) + 1E1(LO) + 2E2 are Raman active.[20] Out of these Raman activemodes, only A1(LO) and E2 are allowed in backscatteringconfiguration.[20,21]

Raman spectrum of as-grown n-GaN exhibits two dominantpeaks at 570 and 735 cm�1, whereas that of as-grown p-GaNshows two peaks at 566 and 734 cm�1. The respective peaksin both the samples are identified as E2 and A1(LO) modes ofGaN.[22] Several Raman modes were observed to appear at223, 289, 363, 378, 417, 428, 447, 570, 663, 735 and 750 cm�1

in the spectra of implanted n-GaN samples and at 199,289, 359, 379, 417, 428, 447, 566, 660, 732 and 750 cm�1

for implanted p-GaN samples.[18] The modes at 570 and566 cm�1 are identified as E2 modes for implanted n-GaNand p-GaN, respectively.[22] The broad bands observed in theRaman spectra of implanted samples at around 199, 223 and289 cm�1 are expected to be due to the disorder-activatedRaman scattering (DARS).[23,24] An excellent discussion on theorigin of disorder modes below 400 cm�1 in wurtzite struc-tures on the basis of Raman scattering analysis can be foundin Chi et al., Gouadec and Colomban, and Havel et al.[25–27]

The modes that appeared at 736.8, 736.4 and 736.1 cm�1 forsamples nCe, Ce1 and Ce2, and at 734.1, 732.2, 731.7 cm�1

for pCe, Ce3 and Ce4, respectively, are identified as A1(LO)modes.

The data shown in Fig. 2 are Gaussian fitted, indicating a redshift in wavenumber of A1(LO) mode for implanted GaN samples.Furthermore, values of full width at half maximum (FWHM) for A1

(LO) mode of nCe, Ce1 and Ce2 are found to be 7.94, 10.22 and10.93 cm�1 and for pCe, Ce3 and Ce4 are 7.77, 7.97 and9.93 cm�1, respectively, which point to the implantation-inducedbroadening of the peaks.

Raman scattering measurements were also performed usingHe–Cd laser UV excitation source, line 325 nm. This excitationsource gives several advantages during Raman scatteringmeasurements of GaN over the visible laser sources.[28,29]

The penetration depth of this laser into GaN is calculated as40 nm[28], which is much smaller than the implantation depth

RD peaks of Ce-implanted n-GaN and p-GaN samples

Ce3 Ce4

P2* P3 P4 P3* P4*

5.188 5.195 5.188 5.189 5.187

2 John Wiley & Sons, Ltd. J. Raman Spectrosc. (2012)

Figure 1. First order Raman spectra for (a) as-grown and Ce-implanted n-GaN, and (b) as-grown and Ce-implanted p-GaN samples. Spectrum ofsapphire is also given for comparison.

Figure 2. First order Raman spectra showing the A1(LO) Raman mode for as-grown and Ce-implanted (a) n-GaN and (b) p-GaN samples.

Figure 3. UV Raman spectra for as-grown and Ce-implanted n-GaN samples.

Structural modifications of GaN after cerium implantation

of 100 nm for implanted samples used in this study. Thisallows true determination of Raman scattering modes fromthe implanted region rather to obtain the signal that is super-imposed from the implanted, unimplanted GaN and substrateregions of the samples. Figure 3 shows the UV Raman spectrafor as-grown and implanted n-GaN samples showing multipleA1(LO) phonon scattering up to fifth order. Raman spectrumfor as-grown samples exhibits strong photoluminescence (PL)emission at 3144 cm�1 owing to a clear outgoing multiphononresonance along with some other peaks found at 3423, 3572,3750, 4007, 4280 and 3423 cm�1. Raman spectrum for samplesCe1 exhibited dominant Raman lines with consistently decreas-ing intensity at 731, 1469, 2208, 2944 and 3687 cm�1. Thesemodes have nearly the same energy interval of ~92meV thatis the LO phonon energy for GaN[30] and thus assigned to1A1(LO), 2A1(LO), 3A1(LO), 4A1(LO) and 5A1(LO), respectively.The samples Ce2 exhibited dominant peaks at 727, 1460and 2200 cm�1 with consistently decreasing intensity, whichare identified as 1A1(LO), 2A1(LO) and 3A1(LO), respectively.In addition to mA1(LO) modes, implanted samples also exhibittwo series of peaks denoted by Pm and Pm

0(m = 1–5).

The spectra of samples Ce1 have Pm peaks located at 578,1300, 2033, 2783 and 3525 cm�1and Pm

0peaks at 878, 1634,

2351, 3147 and 3898 cm�1. On the other hand, sample Ce2exhibited Pm peaks at 578 and 1300 cm�1 and Pm

0peaks at

J. Raman Spectrosc. (2012) Copyright © 2012 John Wiley

1230 and 1678 cm�1. This sample could exhibit mA1(LO) up tom=3 and Pm peaks up to m=2 only. Some other peaks werefound at 635 and 681 cm�1 for the samples Ce1 and at 624 and687 cm�1 for sample Ce2. These values are given in Table S1.

& Sons, Ltd. wileyonlinelibrary.com/journal/jrs

Figure 4. Variation in intensity and FWHM of mA1(LO) modes forCe-implanted n-GaN samples. Only three data points could be observedfor samples implanted with the highest dose. Data points were extractedfrom UV Raman spectra after Lorentz fittings.

A. Majid et al.

Figure 4 gives the variation of intensity and line width ofLO modes for cerium-implanted samples as a function ofwavenumber. An exponential fall in intensity is observed for boththe implanted samples with increase in wavenumber. Similarly,a linear increase in line width of LO modes and an exponentialincrease in the width for sample Ce2 is observed.

Discussion

X-ray diffraction curves of the implanted samples carry infor-mation from both the implanted and unimplanted parts of thelattice, and therefore GaN peak appearing at its usual position(i.e. 2θ=34.56�) is a characteristic peak of GaN material and addi-tional peak structure can be assumed because of the implantedpart of the lattice.[31–33] XRD scans of implanted samples areresults of superposition of the diffracted signals from the originalhexagonal GaN lattice and a similar hexagonal structure buthaving different lattice constants.[33] It is reported that expansionof GaN lattice along c-axis is the reason for appearance of suchadditional peaks at smaller Bragg angles in XRD spectrum.[31–33]

In our opinion, the additional peak structure may consist of singleor multiple peaks depending upon the distribution of the latticeconstants in the implanted part of the lattice. Both the implantedsamples exhibit two peaks on lower angle side of the main GaNpeak (Fig. S2). Appearance of two well-defined peaks indicatesthe presence of two different lattices in the implanted partof the samples. Out-of-plane lattice constants correspondingto these peaks (Table 1) are larger than that of unimplantedGaN (5.185 Å), which points to the expansion of lattice in theimplanted region of samples. It is observed that the values oflattice constants corresponding to subsidiary peaks for sampleshaving dose 2� 1015 cm�2 are smaller than those of peakscorresponding to samples having dose 3� 1014 cm�2. It can beassumed that subsidiary peaks P1 and P2 (for n-GaN samplesimplanted with dose 3� 1014 cm�2) are the same as P1* andP2*, respectively (for n-GaN samples having dose 2� 1015 cm�2).Similar assumption can be made for Ce-implanted p-GaN sam-ples. Validity of this assumption leads us to conclude that thelattice constant shrinks with increase in implantation dose. It isknown that implantation and substitution of large ionic sized rare


earth atoms on cationic lattice sites produce strain. Many groupsreported decrease of strain at certain implantation dose whilestudying the dose dependence of strain.[34,35] We assign strainrelaxation to increase in substitutional efficiency of dopants onGa sites in the implanted regions of the samples.

The positions of E2 and A1(LO) Raman modes for GaN are notestablished in literature and are found to depend upon the sam-ple and experimental setup.[20,36,37] It is well known that built-instrain can cause the shift of E2 mode in GaN.[38] Therefore, theaforementioned factors account for the different values of peakpositions for E2 and A1(LO) Raman modes for n-GaN and p-GaN.A dose-dependent decrease in intensity of E2 was observed inRaman spectra of the implanted samples as shown in the insetof Fig. 1a. This is due to the expansion of lattice after implanta-tion, which was also seen in XRD measurements. The E2 phononline width, which is a measure of crystalline quality, was found tobe 5.63, 5.93 and 6.45 cm�1 for nCe, Ce1 and Ce2 samples, re-spectively. Similarly, E2 phonon line widths found for Ce-implanted p-GaN are found to be 3.8, 3.97 and 4.07 cm�1 forpCe, Ce3 and Ce4, respectively. The increase in Raman line widthmeans the shortening of phonon life time, which indicates thatprobability of phonon scattering has been increased in GaN afterimplantation. It is clearly seen that E2 phonon line is broadenedpreferentially from lower wavenumber side with increase in dose.It is possibly due to the appearance of forbidden modes A1(TO)and E1(TO) at 532 and 558 cm�1, respectively. These modes arenot allowed in our experimental backscattering configurationaccording to Raman selection rules[20,21,39] but possibly implanta-tion-induced disorder in lattice has activated them.

The red shift and broadening of A1(LO) mode in GaN is usuallyattributed to the extended defects and lattice disordering.[40]

Therefore, we assign the dose-dependent red shift and broaden-ing of A1(LO) in case of our implanted samples to the straininduced by the point defects produced by implantation.

The implanted samples exhibited some new Raman modes at360, 363, 660, 663 and 703 cm�1, as shown in Fig. 1. None ofthese modes matches the cerium-related local vibrational mode

(LVM) calculated using the relation oCe-NoGa-N ¼

ffiffiffiffiffiffiffiffimGa-NmCe-N

q.[41,42] The mode

at 363 cm�1 for n-GaN and 359 cm�1 for p-GaN cannot be takenas DARS mode because the phonon dispersion curve for GaN inthis range has no large density of states.[43] Limmer et al.[23]

and Wang et al.[43] have observed a Raman mode at 363 cm�1

in Si-implanted GaN and assigned it to the LVM of vacancyrelated defects. Therefore, we assign the 363 cm�1 mode forimplanted n-GaN and 360 cm�1 mode for implanted p-GaN tothe vacancy-related defects. The peaks at 663 cm�1 for implantedn-GaN and 660 cm�1 for implanted p-GaN are close to 670 cm�1

reported for variety of ions implanted into GaN.[19,23] Thesemodes are assigned to the nitrogen vacancy-related defects,as density of states spectrum in this region is dominated bynitrogen-related stretching modes.[19]

The appearance of a strong PL emission resulting from theband-to-band recombination has masked the usual Ramanmodes in UV Raman spectrum of as-grown samples. However,such PL emission is quenched and could not be detected inRaman spectra of both the implanted samples because ofimplantation-induced damages and incomplete recovery of lat-tice after post-implantation annealing. Complete quenching ofPL emission was reported under similar experimental conditionsfor Be-implanted GaN at dose of 5� 1015 cm�2,[28] whereas thesame happened in present samples when implanted by Ce at


Structural modifications of GaN after cerium implantation

dose of 3� 1014 cm�2. This is expected because of comparativelymore lattice damage caused by heavy ion of Ce in comparisonwith that of Be. The strong suppression of PL emission forimplanted samples is caused by the presence of defects andproduction of non-radiative recombination centers emergedduring high energy ion implantation.[29] The intensity reductionprovided an indirect advantage and many Raman features areobserved in spectra of implanted samples. The observation ofmultiphonon mA1(LO) peaks indicates high crystalline quality ofthe samples.[44] The appearance of A1(LO) peaks up to third orderin samples implanted with dose of 2� 1015 cm�2 tells about theimplantation-induced degradation of crystalline quality in com-parison with that of 3� 1014 cm�2. The wavenumber of 1A1(LO)phonon for as-grown and implanted samples nCe, Ce1, Ce2 wasobserved (Fig. 2a) at 736, 731 and 728 cm�1, which shows a cleardose-dependent red shift in the wavenumber. This red shift canbe attributed to the strain induced by implantation and Ce sub-stitution on Ga sites. It is expected because of the larger size ofcerium ion occupying the Ga sites. The dose-dependent increasein broadening and decrease in intensity of mA1(LO) modes (Fig. 4)point to the level of implantation-induced damages. The increasein Raman line width also means the shortening of phonon lifetime, which indicates that probability of phonon scattering hasbeen increased in GaN after implantation.[45]

It is believed that majority of rare earth ions substitute Ga intrivalent state upon implantation into GaN lattice.[46–48] Alves et al.have already reported the Ce substitution on Ga sites in GaN;[49]

therefore, it is assumed that majority of implanted Ce ions havesubstituted the Ga sites to produce a diluted magnetic semicon-ductor alloy GaCeN.[18,50]

Conclusions

X-ray diffraction and Raman measurements were carried out onCe-implanted GaN samples to study the modifications in structuralproperties of the material. XRD measurements including reciprocalspace maps indicated expansion of lattice along c-axis of GaN,which is interpreted as incorporation of heavy Ce atom on Ga sites.First order Raman scatteringmeasurements performed using argonlaser show appearance of several disorder-activated Ramanscattering modes in addition to typical GaN features. None of thesemodes matches the cerium-related LVM; however, we assigned the363 cm�1 mode for implanted n-GaN and 360 cm�1 mode forimplanted p-GaN to the vacancy-related defects. The observeddose-dependent decrease in intensity of E2 is due to the expansionof lattice after implantation, which was also seen in XRD and RSMmeasurements. The wavenumber of 1A1(LO) phonon for as-grownand implanted samples nCe, Ce1 and Ce2 was observed at 736,731 and 728 cm�1, which shows a clear dose-dependent red shiftin the wavenumber. This red shift can be attributed to the straininduced by implantation and Ce substitution on Ga sites. UV Ramanscatting measurements show that strong PL emission resultingfrom the band to band recombination is found quenched andcould not be detected in Raman spectra of both the implantedsamples. The strong suppression of PL emission for implantedsamples is caused by the presence of defects and productionof non-radiative recombination centers emerged during highenergy ion implantation. The observation of multiphonon mA1(LO) peaks up to fifth order indicates high crystalline quality ofthe implanted samples.

J. Raman Spectrosc. (2012) Copyright © 2012 John Wiley

Supporting information

Supporting information may be found in the online versionof this article.

References[1] M. M. Mezdrogina, E. Yu Danilovskii, R. V. Kuz’min, Inorg. Mater.

2011, 47, 1450.[2] C. W. Thiel, T. Böttger, R. L. Cone, J. Lumin. 2011, 131, 353.[3] D. S. Lee, A. J. Steckl, Appl. Phys. Lett. 2001, 79, 1962.[4] A. J. Steckl, J. C. Heikenfeld, D. S. Lee, M. J. Garter, C. C. Baker,

Y. Wang, R. Jones, IEEE J. Sel. Top. Quantum Electron. 2002, 8, 749.[5] S. Dhar, L. Pérez, O. Brandt, A. Trampert, K. H. Ploog, J. Keller,

B. Beschoten, Phys. Rev. B 2005, 72, 245203.[6] V. K. Dugaev, V. I. Litvinov, A. Lusakowski, Phys. Rev. B 1999, 59, 15190.[7] W. X. Ni, C. X. Du, D. Duteil, A. Elfving, G. V. Hansson, Opt. Mater.

2001, 17, 65.[8] R. H. Birkhahn, M. J. Garter, A. J. Steckl, Appl. Phys. Lett. 1999, 74, 2161.[9] E. D. Readinger, G. D. Metcalfe, P. H. Shen, M. Wraback, N. Jha,

N. Woodward, P. Capek, V. Dierolf, MRS Proceedings. 2008, 1111,1111-D02-04.

[10] J. B. Gruber, B. Zandi, H. J. Lozykowski, W. M. Jadwisienczak, J. Appl.Phys. 2001, 91, 2929.

[11] A. J. Steckl, R. H. Birkhahn, Appl. Phys. Lett. 1998, 73, 1700.[12] H. Bang, S. Morishima, Z. Li, K. Akimoto, M. Nomura, E. Yagi, J. Cryst.

Growth 2002, 237–239, 1027.[13] H. J. Lozykowski, W. M. Jadwisienczak, I. Brown, Appl. Phys. Lett.

1999, 74, 1129.[14] H. J. Lozykowski, W. M. Jadwisienczak, I. Brown, Solid State Commun.

1999, 110, 253.[15] Ei Ei Nyein, U. Hömmerich, J. Heikenfeld, D. S. Lee, A. J. Steckel,

J. M. Zavada, Appl. Phys. Lett. 2003, 82, 1655.[16] K. Lorenz, U. Wahl, E. Alves, E. Nogales, S. Dalmasso, R.W. Martin,

K. P. O’Donnell, M.Wojdak, A. Braud, T.Monteiro,Opt.Mater. 2006, 28, 750[17] A. Majid, A. Ali, J. Phys. D: Appl. Phys. 2009, 42, 045412.[18] A. Majid, J. Iqbal, A. Ali, J. Supercond. Nov. Magn. 2011, 24, 585.[19] M. Katsikini, K. Papagelis, E. C. Paloura, S. Ves, J. Appl. Phys. 2003, 94, 4389.[20] V. Yu. Davydov, Yu. E. Kitaev, I. N. Goncharuk, A. N. Smirnov, J. Graul,

O. Semchinova, D. Uffmann, M. B. Smirnov, A. P. Mirgorodsky,R. A. Evarestov, Phys. Rev. B 1998, 58, 12899.

[21] U. Haboeck, H. Siegle, A. Hoffmann, C. Thomsen, Phys. Status. Solidi C2003, 6, 1710.

[22] H. Siegle, G. Kaczmarezyk, L. Filippidis, A. P. Litvinchuk, A. Hoffmann,C. Thomsen, Phys. Rev. B 1997, 55, 7000.

[23] W. Limmer, W. Ritter, R. Sauer, B. Mensching, C. Liu, B. Rauschenbach,Appl. Phys. Lett. 1998, 72, 2589.

[24] R. S. Berg, P. Y. Yu, E. R. Weber, Appl. Phys. Lett. 1985, 47, 515.[25] T. T. K. Chi, G. Gouadec, Ph. Colomban, G. Wang, L. Mazerolles,

N. Q. Liem, J. Raman Spectrosc. 2011, 42, 1007.[26] G. Gouadec, Ph. Colomban, Prog. Cryst. Growth. Charact. Mater.2007, 53, 1.[27] M. Havel, D. Baron, L. Mazerolles, Ph. Colomban, Appl. Spectrosc.

2007, 61, 855.[28] D. Pastor, J. Ibanez, R. Cusco, L. Artus, G. Gonzalez-Dıaz, E. Calleja,

Semicond. Sci. Tech. 2003, 22, 70.[29] B. Boudart, Y. Guhel, J. C. Pesant, P. Dhamelincourt, M. A. Poisson,

J. Phys. Condens. Matter 2004, S49,16.[30] W. H. Sun, S. J. Chua, L. S. Wang, X. H. Zhang, J. Appl. Phys. 2002, 91, 4917.[31] C. Liu, B. Mensching, M. Zeitler, K. Volz, B. Rauschenbach, Phys. Rev. B

1998, 57, 2530.[32] S. B. Qadri, B. Molnar, M. Yousuf, C. A. Carosella. Nucl. Instrum. Meth. B

2002, 190, 878.[33] C. Liu, B. Mensching, K. Volz, B. Rauschenbach. Appl. Phys. Lett. 1997,

71, 2313.[34] P. Partyka, R. S. Averback, D. V. Forbes, J. J. Coleman, P. Ehrhart,

J. Appl. Phys. 1998, 83, 1265.[35] A. Majid, A. Ali, J. J. Zhu, Y. T. Wang, J. Phys. D: Appl. Phys. 2008, 41, 115404.[36] J. Gleize, F. Demangeot, J. Frandon, M. A. Renucci, F. Widmann,

B. Daudin, Appl. Phys. Lett. 1999, 74, 703.[37] T. Azuhata, T. Sota, K. Suzuki, S. Nakamura, J. Phys. Condens. Matter

1995, L129, 7[38] C. Kisielowski, J. Kruger, S. Ruvimov, T. Suski, J. W. Ager, E. Jones,

Z. Liliental-Weber, M. Rubin, E. R. Weber, M. D. Bremser, R. F. Davis,Phys. Rev. B 1996, 54, 17745.

& Sons, Ltd. wileyonlinelibrary.com/journal/jrs

A. Majid et al.

[39] Z. C. Feng, W. Wang, S. J. Chua, P. X. Zhang, K. P. J. Williams, G. D. Pitt.J. Raman Spectrosc. 2001, 32, 840.

[40] R. Fornari, M. Bosi, D. Bersani, G. Attolini, P. P. Lottici, C. Pelosi, Semicond.Sci. Tech. 2001, 16, 776.

[41] A. Hoffmann, A. Kaschner, C. Thomsen, Phys. Status Solidi C,2003, 6, 1783.

[42] M. D. McCluskey. J. Appl. Phys. 2000, 87,3593.[43] H. T. Wang, L. S. Tan, E. F. Chor, Semicond. Sci. Tech. 2004, 19, 142.[44] S. Dhara, S. Chandra, G. Mangamma, S. Kalavathi, P. Shankar, K. G. M. Nair,

A. K. Tyagi, C. W. Hsu, C. C. Kuo, L. C. Chen, K. H. Chen, K. K. Sriram,Appl. Phys. Lett. 2007, 90, 213104.


[45] S. Tripathy, S. J. Chua, M. S. Hao, E. K. Sia, A. Ramam, J. Zhang,W. H. Sun, L. S. Wang, J. Appl. Phys. 2002, 91, 5840.

[46] B. Pipeleers, S. M. Hogg, A. Vantomme, J. Appl. Phys. 2005, 98, 123504[47] U. Wahl, E. Alves, K. Lorenz, J. Correia, T. Monteiro, B. De Vries,

A. Vantomme, R. Vianden, Mater. Sci. Eng. B 2003, 105, 132[48] A. Svane, N. E. Christensen, L. Petit, Z. Szotek, W. M. Temmerman,

Phys. Rev. B 2006, 74, 165204[49] E. Alves, K. Lorenz, R. Vianden, C. Boemare, M.J. Soares, T. Monteiro,

Mod. Phys. Lett. B 2001, 15, 1281.[50] A. Majid, Ion Implantation into GaN and AlInN, LAP Lambert

Academic Publishing, Germany, 2012.


Structural modifications of GaN after cerium implantation

Documents