A R T I C L E

Journ

al of

Materials

Ch

emistry

www.rsc.o

rg/m

aterials

Sol-gel formation of ordered nanostructured doped ZnO films

Nicola R. S. Farley,*a,b Christopher R. Staddon,a Lixia Zhao,a Kevin W. Edmonds,a

Bryan L. Gallaghera and Duncan H. Gregoryb

aSchool of Physics and Astronomy, University of Nottingham, Nottingham NG7 2RD,

UK. E-mail: Nicola.Farley@nottingham.ac.uk; Fax: 144-115-951-5180;Tel: 144-115-951-5189bSchool of Chemistry, University of Nottingham, Nottingham NG7 2RD, UK

Received 21st October 2003, Accepted 21st January 2004

First published as an Advance Article on the web 23rd February 2004

A novel sol-gel route to c-axis oriented undoped and Co, Fe and Mn doped ZnO films is reported. Sols were

prepared from a hydrated zinc acetate precursor and dimethylformamide (DMF) solvent. Films were spin-

coated on to hydrophilic sapphire substrates then dried, annealed and post-annealed, producing almost purely

uniaxial ZnO nanocrystallites and a high degree of long-range structural order. Specific orientation of

hexagonal crystallites is demonstrated both perpendicular and parallel to the substrate surface. Cobalt doping

resulted in the formation of disconnected oriented ZnO nanocrystals. Vanadium doped films formed the spinel

oxide ZnAl2O4. Structural, optical and morphological characterisation demonstrated the high quality of the

films and shows the suitability of the method for cost-effective fabrication studies in areas of oxide research

that traditionally employ epitaxial growth techniques.

Introduction

ZnO has a long history of usage for pigments and protectivecoatings on metals. The electrical, optoelectronic and photo-chemical properties of undoped ZnO has resulted in usefor solar cells,1,2 transparent electrodes24 and blue/UV lightemitting devices.5 Ordered c-axis orientation of ZnO crystal-lites perpendicular to the substrate surface is desirable forapplications where crystallographic anisotropy is a prerequisitee.g. piezoelectric surface acoustic wave or acousto-opticdevices.6 Furthermore, the introduction of large numbers ofsubstitutional transition metal (TM) dopants into widebandgap oxides7 and IIIV semiconductors8 has been shownto induce ferromagnetic ordering. The ferromagnetism oper-ates via a mechanism by which itinerant carriers are spinpolarised and mediate ferromagnetic ordering between thewidely spaced dopant ions. Such long-range ferromagneticordering demands materials with high structural quality, asdemonstrated for IIIV ferromagnetic semiconductors such asGa1 2 xMnxAs.

810 Films with such a high degree of structuralorder are usually deposited epitaxially by molecular beamepitaxy (MBE),5 pulsed laser deposition,11 chemical vapourdeposition6 or sputtering.12

Sol-gel methods are being employed increasingly for the low-cost fabrication of ordered high-specification materials sincestructural and morphological characteristics may be tuned inorder to tailor the optical, electrical or magnetic propertiesof the material. In particular, sol-gel methods are now beingrecognised as a potential route to the preparation of dilutemagnetic semiconductors.13 Sol-gel methods may also befavoured since it is possible to produce a large number ofsamples rapidly and at a fraction of the cost of MBE. Sol-gelpreparations are therefore ideal for exploratory studies forwhich large numbers of candidate materials, compositionsor preparative conditions require screening, and can thereforebe utilised for areas of oxide research that traditionallyemploy MBE growth techniques. The fabrication of dilutemagnetic semiconductors with Curie temperature above roomtemperature is such a challenge. Such materials have thepotential to revolutionise the electronics industries, which at

present are actively pursuing the next generation successors toboth silicon-based volatile data storage and processing, andnon-volatile ferromagnetic recording media. TM doped ZnOhas been shown theoretically to be one of the most promisingmaterials for room temperature ferromagnetism.7 We havedeveloped a new method for producing ZnO films doped withtransition metals, suitable for applications in spin electronics(spintronics).In this study we describe a method that has several advant-

ages over existing sol-gel procedures for depositing undopedand doped ZnO. Sol preparation is simplified compared tothose reported in the literature, and greater control of dopantincorporation is facilitated.Alcoholic solvents are often used to prepare sols; for

instance, most sol-gel syntheses of undoped and Al, Li, B,Cd, or Co, doped c-axis oriented ZnO use solvents such asmethoxyethanol1416 or ethanol/propanol.17,18 Alcohols arecommonly used because the solvent also acts as a reagent.However, the solvent does not participate in the reactionforming ZnO from zinc acetate. Traditional methods requireheating of the sol to 60100 uC to dissolve the precursor, anduse additives such as mono/di/triethylamine4,1417,19,20 or lacticacid18 to improve the stability and homogeneity of the sol bypreventing uncontrolled hydrolysis and precipitation of ZnO.The procedure reported here is simpler than established

methods since the sols consist only of two or three components:zinc acetate precursor, a dopant salt and the solvent. Zincacetate is more soluble in highly polar solvents such asdimethylformamide (DMF) than in alcohols, so it is notnecessary to heat the starting mixture. The relatively lowvapour pressure of DMF prevents premature and unevendrying, which may cause cracking and disrupt uniaxial crystal-lisation of the films. Also, an additive is not required toimprove sol stability and homogeneity.The oxidation state (and spin state), homogeneity and

substitutional incorporation of the dopants must be consideredwhen TM doping for producing dilute magnetic semicon-ductors. Heating the mixture (which is necessary whendissolving in alcohols) increases the possibility of inadvertentlyaltering the oxidation state of the dopants or forming dopantD

OI:10.1039/b313271d

J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 1 0 8 7 1 0 9 2 1 0 8 7T h i s j o u r n a l i s T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

oxide secondary phases prior to ZnO film formation. Tocircumvent this problem, sols were prepared at room tem-perature. Heating was performed only after amorphous dopedZnO films were formed. This could only be achieved whenusing a solvent that stabilised the sol and did not requireheating to dissolve the precursors, which is why the new methodwas developed.We present structural, morphological and optical character-

isation of ZnO/ZnAl2O4 films, both undoped and dopednominally with 6 mol% Co21, Fe31, Mn21, Mn31 or V31 withrespect to Zn.

Experimental

Each sol was synthesised from a zinc acetate Zn(CH3CO2)2.2H2O precursor (Fluka, 99.51%) and DMF solvent (Fisher,w99%). Dopant ions were provided by transition metalchlorides or acetates: CoCl2?56H2O (99.999%, Aldrich),FeCl3?6H2O (971%, East Anglia Chemicals), Mn(CH3CO2)2?4H2O (991%, Aldrich), Mn(CH3CO2)3?2H2O (97%, Aldrich)and VCl3 (Aldrich). Zinc acetate and dopant salts weredissolved completely in DMF by stirring at room temperature.Six 20 ml, 0.6 M sols were prepared: undoped (transparent),Fe31 doped (red), Co21 doped (pink), Mn21 doped (trans-parent), Mn31 doped (dark brown) and V31 doped (darkgreen). Henceforth, the initial oxidation state of the dopantswill be used to distinguish the samples. The sols were stableand homogeneous; no particulates or precipitates were visibleto the eye and their appearance was unchanged for severalmonths. Sols were usually stirred overnight at room tempera-ture before use.Sapphire (0001) substrates (Union Carbide, USA) were

cleaned and hydrophilically terminated by immersion in 1 : 1by volume 0f concentrated H2SO4 : 30% vol. H2O2 (piranhasolution) for 15 min, followed by rinsing with deionised waterand drying in air. Films were deposited by dropping sol ontothe substrate using a 1ml syringe fitted with a polypropylenefilter with 0.2 mm pores (Whatman, BDH). Filtration was anecessary precaution to ensure that dust particulates were notincorporated into the films, since they are thought to disruptuniaxial crystallisation.15 As an additional precaution, themagnetic PTFE-covered stirrer bead also served to withholdany solid ferromagnetic impurities that may have been presentin the sol. Each film consisted of 5 layers, each spun at 4000 rpmfor 30 s. After deposition, each layer was left to hydrolyse inhumid air at room temperature for 1 h then dried at 300 uC for10 min in an air furnace before spinning the next layer. Eachspun layer must be thin enough to ensure the simultaneousevaporation of all solvent, thus preventing the cracking anddisruption to crystallisation caused by gradual evaporation.14

Each complete five layer film was annealed at 550 uC for 5 h,warming at 1 uCmin21 in air, then post-annealed at 15 uCmin21

to 850 uC for 5 h, cooling naturally. This annealing regime issimilar to that reported by Wessler et al.15 for undoped ZnOZnAl2O4 with methoxyethanol solvent. All undoped and dopedfilms appeared smooth, uniform and transparent. By contrast,films prepared using methoxyethanol and monoethanolaminewere the colour of dopant oxides, suggesting either phaseinhomogeneity or non-substitutional doping, even though noother phase was detected by X-ray diffraction (XRD).Structural characterisation was performed using a Philips

Xpert materials research X-ray diffractometer with a Cuanode and a b-filtering parafocusing mirror, generating Karadiation at 1.542 A and operating at 45 kV and 40 mA. Allmeasurements were performed using a parallel plate detectorand the beam size was 20 mm2. 2h/v scans (analogous to h/2hscans) were recorded to determine the crystallinity andcrystalline orientation of the films. The grazing incidencegeometry was used for 2h and w scans in order to access specific

out-of-plane, high symmetry reflections. The Scherrer formulawas used to estimate the ZnO particle size from the 0002 peakwidth. Surface morphology was investigated using a DigitalInstruments multimode scanning probe microscope for tappingmode atomic force microscopy (AFM), utilising Al-coveredsingle Si cantilevers (Ultrasharp, MikroMasch) and operatingat about 150 kHz. Images were processed using the programWSxM (Nanotec Electronica, Spain). Optical qualities wereassessed by photoluminescence (PL) measurements, utilising aCd-He laser with an excitation wavelength of 325 nm and abeam diameter of about 1mm. A photomultiplier tube was usedto detect emission normal to the sample surface. All structural,morphological and optical measurements were performed atroom temperature.

Results and discussion

Crystalline orientation and nanostructuring

The diffraction patterns in Fig. 1 show the crystallographicphases present in the samples. Each film shows the Ka1 andKa2 sapphire 0006 peaks around 42u and other sapphire peaksat 20.7u, 64.7u and 64.9u.

In-plane film structure. The undoped and Co21 doped filmsare single phase ZnO with a strong 0002 peak at 34.5u. A highdegree of preferred orientation is evident, giving rise to spectraresembling single crystal diffraction patterns, although thefilms are polycrystalline. The c-axis of most ZnO crystallitesaligns with the c-axis of the substrate in the [0001] direction.The crystals therefore stand with c-axis perpendicular to thesubstrate surface, and the c-plane of both film and substrate areparallel to the surface. The logarithmic scale reveals a secondpeak at 36.3u for all ZnO samples; this corresponds to the 1011reflection, which is the dominant peak in randomly orientedpolycrystalline ZnO (Fig. 1a21).A mixture of hexagonal ZnO and cubic (spinel) ZnAl2O4

phases is evident in Fe31, Mn21 and Mn31 doped films. Thespinel is the product of the reaction between ZnO and sapphire:ZnO 1 Al2O3A ZnAl2O4.15 The height of the ZnO 0002 peakrelative to other reflections signifies the extent of c-axisorientation and hence film structural quality. The relativeZnO peak intensities given in Table 1 indicate that the qualityof Fe31, Mn21 and Mn31 films is low compared to undopedand Co21 doped samples, with TM31 doped films having thepoorest structure. The sapphire peaks are almost absent in theMn21 scan (Fig. 1g) because the c-axis of sapphire and ZnO aremisaligned, making sapphire (0006) and ZnO (0002) planesnon-parallel. Only spinel ZnAl2O4 peaks are visible above thenoise level for the V31 doped film (scan 1i), suggesting extensiveintermixing of the initial film with the substrate. The ZnAl2O4

Fig. 1 XRD patterns: a) zincite ZnO reference 21, b) gahniteZnAl2O4 reference 21, c) sapphire (0001) substrate, d) undoped,e) Co21, f) Fe31, g) Mn21, h) Mn31 and i) V31 doped film. ZnO andZnAl2O4 peaks are marked with dotted lines and arrows, respectively.Scans are offset for clarity.

1 0 8 8 J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 1 0 8 7 1 0 9 2

phase is polycrystalline, with a slightly preferred orientationin the [111] direction since the main peak is at 59.5u (333reflection), as opposed to 36.9u (311 reflection) in a randomlyoriented powder (Fig. 1b21 and Table 1). The spinel phasedeveloped more readily in V31, Mn31 and Fe31 doped samplesthan in undoped and Co21 and Mn21 doped films preparedunder the same conditions, suggesting that spinel formation ispromoted by TM31 ions. Since TM doped ZnO has theoreticalpotential for exhibiting room temperature ferromagnetism, thisobservation has relevance to those seeking to deposit single-phase doped films on sapphire using any method.

Out-of-plane film structure. A grazing incidence configura-tion, with the incident angle v fixed typically at 0.35u, providedaccess to specific out-of-plane reflections in the ZnO andZnAl2O4 crystallites. Contributions from the substrate are alsoexcluded at such low angles of incidence.Fig. 2 shows all of the polycrystalline ZnO reflections since

the samples are tilted, giving access to more internal planes. Aprominent ZnO 1013 peak is present at 63.2u in the undopedand Co21 doped samples. The intensity of the 1013 peakrelative to the other reflections diminishes passing from theundoped to the Mn31 doped samples (scans 2ae). TheZnAl2O4 phase is not apparent in the Fe

31 doped scan(Fig. 2c), but two oxide phases are visible in both Mn dopedfilms (scans 2de), suggesting that the reaction had notextended beyond the interfacial region in the Fe31 dopedsample. The V31 doped film again shows only the spinel phase,with a dominant 440 reflection at 65.5u (scan 2f).The c and a parameters in Fig. 3 were calculated using ZnO

0002 (34.5u) and 1010 (31.9u) peak positions in Fig. 1 and 2,respectively. The ZnO a parameter varies significantly with

varying TM dopant and corresponds with the ionic radii of theions,22 indicating that Co21, Fe31 and Mn21 dopants are likelyto be substituting for Zn in the ZnO host lattice. The c para-meter is independent of dopant. High spin Zn21 is tetrahedrallycoordinated to O in ZnO, but high spin d4 Mn31 and d2 V31

are unstable in this environment. Conversely, all of the otherdopants readily occupy tetrahedral sites. Zn21 is also tetra-hedrally coordinated to O in ZnAl2O4, and Al

31 occupiesoctahedral sites. Spinel formation is likely to be energeticallypreferable due to the charge compensating defects necessary tosubstitute TM31 for Zn21 in ZnO, and the availability ofoctahedral coordination sites in Al31-substituted ZnAl2O4.Whilst at grazing incidence, 2h was fixed either at the ZnO

1013 peak or the ZnAl2O4 440 peak, then the sample wasrotated 360u to perform a w scan in the azimuthal circle in orderto explore the hexagonal 6-fold and cubic 3-fold symmetry inZnO (Fig. 4a) and ZnAl2O4 (Fig. 4b), respectively. Fig. 1demonstrates the high degree of ZnO phase c-axis orientation,and Fig. 5 shows that the crystallites are also highly directionalin the (a,b) plane. This is the first report of out-of-planeordering in sol-gel-derived ZnO films. None of the films areas ordered as the CVD-grown GaN reference single crystal

Table 1 Peak intensity ratios derived from 2h/v XRD scans andreference spectra displayed in (Fig. 1)

Sample

ZnO Reflection

0002 (35.5u) 1011 (36.3u)

Zincite ref. 0.4 1.0Undoped 281.1 1.0Co21 doped 37.9 1.0Fe31 doped 5.8 1.0Mn21 doped 9.3 1.0Mn31 doped 1.9 1.0

ZnAl2O4 Reflection

333 (59.5u) 311 (36.9u)

Gahnite ref. 0.4 1.0V31 doped 1.1 1.0

Fig. 2 Grazing incidence 2h XRD patterns at v# 0.35u: a) undoped,b) Co21, c) Fe31, d) Mn21, e) Mn31 and f) V31 doped films. ZnAl2O4peaks are marked with dotted lines; all other peaks originate fromZnO. Scans are offset for clarity.

Fig. 3 Correlation between dopant tetrahedral ionic radius22 andZnO a and c parameters. Error bars indicate 95% confidence level.

Fig. 4 Schematic diagrams of hexagonal 6-fold symmetry about[0001], and cubic 3-fold symmetry about [111], respectively ina) zincite unit cell, (1013) and commensurate planes are shaded, andb) gahnite unit cell, (440) and commensurate planes are shaded. Thegrazing incidence beam geometry and axis for w rotation are illustrated.

Fig. 5 Grazing incidence XRD w scans about: hexagonal wurtzite(1013) peak of a) GaN (0002) reference film, b) undoped, c) Co21,d) Fe31 and e) Mn21 doped films, and cubic gahnite 440 peak off) Mn21 doped and g) V31 doped film. Peaks marked with arrows arereferred to in the text. Scans are arbitrarily offset for clarity.

J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 1 0 8 7 1 0 9 2 1 0 8 9

(Aixtron AG, Germany), which has a very similar wurtzitestructure to ZnO, but it is clear that the ZnO crystallites are notrandomly arranged in the (a,b) plane. The 60u separated peaksin GaN (scan 5a) are reflections from six commensurate planesin the hexagonal lattice shown in Fig. 4a.The primary peaks in the undoped, Co21 and Fe31 doped

samples have a 60u separation, showing that most of thehexagonal crystallites have the same orientation in the (a,b)plane. However, some of the undoped and Co21 dopedcrystallites are rotated 30u relative to most of the crystallites,giving rise to a set of peaks separated by 60u and offset from theprimary orientation by 30u. Two more orientations are evidentin undoped and Co21 doped samples, one offset by 20u and theother by 40u relative to the primary direction. A degree ofrandom orientation is evident in all of the films, as shown bythe background signal. The broad featureless curve from theMn21 doped film (scan 5e) shows that no out-of-planepreferential orientation was found in the hexagonal phase ofthe film. However, a slight preference for specific orientations isvisible in the cubic phase (scan 5f), showing a primary set ofthree small features separated by 120u.The single spinel ZnAl2O4 phase in the V

31 doped film has amore complex and strongly developed structure than in theMn21 doped sample. Within the 120u separation of the threeprimary peaks, sets of 120u-separated peaks are offset from theprimary peaks by 30, 60 and 90u. Additionally, two satellitepeaks are visible, shouldering the peaks offset from the primaryset by 60u (arrows on scan 5g). These are offset from the 60upeaks by approximately 1 and 26u, and are therefore offsetfrom the main peaks by about 54 and 66u. So in total sixdistinct orientations are observed in the V31 doped sample.This complex arrangement of directions is repeatable over each120u cycle separating the primary peaks. The Mn31 dopedsample was not crystalline enough to display preferentialorientation in either the hexagonal or cubic phases.AFM yielded additional insight into the orientation of

crystallites in the (a,b) plane. Fig. 6a shows that the undoped

film consists mainly of closely packed crystallites with a few smallvoids. The close proximity of the grains makes the shape andorientation of the crystallites difficult to distinguish and the sizedifficult to measure from the image. However, an approximateparticle size of 185nmwas calculated using the Scherrer formula.The Co21 doped film has a more open structure with large voidsthat are irregularly shaped and randomly distributed. Thediscontinuity of the Co21 doped film made the size andarrangement of crystallites easier to determine.

Crystallite dimensions. The diameter of the crystallites wasmeasured from profiles of image 6b as the trough-to-troughhorizontal distance across each grain. Fig. 7 shows that the Co21

doped ZnO crystallites are approximately 100nm across at thebase. This value is close to the particle size estimated using theScherrer formula (dashed line onFig. 7). Fig. 6bc show thatCo21

doped grains have hexagonal bases, and the flat edges of manyhexagonal crystallites appear to be aligned in the (a,b) plane.Since the cantilever tip was 1520 mm high and v20 nm

across at the point, it was able to follow the profile of the film,

Fig. 6 2D (ab) and 3D (cd) AFM images of: a) undoped, bc) Co21 doped and d) V31 doped films.

Fig. 7 Approximate base diameter of Co21 doped ZnO crystallites,measured from AFM image 6b. Dashed line indicates average Co21

doped ZnO particle size derived using the Scherrer formula.

1 0 9 0 J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 1 0 8 7 1 0 9 2

through the voids to the substrate, resulting in a thicknessmeasurement of about 40 10 nm for all films. Undoped andCo21 doped grains have different height distributions. In theundoped film, all the grains are approximately the same height.However, some Co21 doped crystallites are low flattenedhexagons about 3040 nm high, while many others are tallerand taper to a point at around 4060 nm. These shapes arecharacteristic of oxygen and zinc termination of planar andtapered crystals, respectively.23

The V31 doped film is very different to the others. The film isdensely packed and has large regions consisting either ofslanted blocks or small, rough, irregular grains. The largeblocks are approximately 400 nm across and appear to besimilarly oriented. The Mn21 doped film displays a mixture ofhexagonal and cubic ordering (image not shown).

Optical properties and crystal quality

In addition to structural evaluation, photoluminescence spectro-scopy provides additional insight into semiconductor crystalquality through emission from discrete states bordering andwithin the band gap, or states arising from structural defects.These measurements can also be used to reveal quantum effectsin very small crystallites. Due to the low coverage of the Co21

doped sample, photoluminescence hardly exceeded the back-ground level (scan 8b). The characteristic green mid-band(2.4 eV) and UV band-edge (3.3 eV) emissions of ZnO5,24 aredisplayed by undoped, Fe31, Mn21 and Mn31 films (scans8gd, respectively). The broad deep-level emission bandoriginates from structural defects, which form electron-holerecombination centres. The band-edge luminescence is due tofree exciton emission or exciton transitions with shallow donorsor acceptors; this feature is usually associated with highstructural quality. Contrary to this interpretation, the trendobserved in Fig. 8 demonstrates that the band-edge to mid-band intensity ratio increases as ZnO structural qualitydiminishes, culminating with zero ZnO mid-band emissionfrom the V31 doped sample (scan 8c). Fig. 1, 2 and 5, andTable 1 show that the reduction in ZnO structural quality isaccompanied by an increase in the proportion of ZnAl2O4present, suggesting that this phase may be responsible for theobserved trend. Since ZnAl2O4 has a direct band gap of 4.04.5 eV,25 the excitation energy of 3.8 eV is insufficient to initiateabsorption across the energy gap, thereby diminishing theluminescent intensity in the range measured. Emission from theV31 doped sample coincides with the ZnO band edge, so couldoriginate from residual ZnO below the detectable limit forXRD. However, this is unlikely since PL measurements werenot achieved on Co21 doped ZnO, but XRD was successful.Note that the band-edge to mid-band intensity ratio is the samefor Mn21 and Mn31 doped samples.

Conclusions

A novel sol-gel synthesis of c-axis oriented undoped and TMdoped ZnO films was described. The structure and morphologyof the films is strongly dependent on the dopant species. X-raydiffraction revealed single-phase ZnO for undoped and Codoped films. TM31 doped films formed ZnAl2O4, whichresulted from the reaction between ZnO and the sapphiresubstrate. This reaction was probably promoted by TM31 ions.Mixed hexagonal zincite ZnO and cubic spinel ZnAl2O4 phaseswere detected in the Fe and Mn doped samples. The stronglydirectional nature of the ZnO crystallites was demonstrated byXRD. A high degree of long-range structural order was evidentboth parallel and perpendicular to the substrate surface forundoped and Co doped films. V doped ZnAl2O4 crystalliteswere also preferentially oriented. The ordered arrangement ofZnO and ZnAl2O4 films was confirmed by AFM imaging,which showed that undoped and Co doped films had differentdistributions of crystallites, and the V doped ZnAl2O4 filmcontained regions of oriented crystallites. Optical characterisa-tion suggested the high structural quality of the films by thepresence of the ZnO band edge feature. Mid-band emissionappeared to be suppressed by the presence of ZnAl2O4.Our results illustrate the utility of sol-gel methods for

producing high quality oxide semiconductors suitable for adiverse range of technological applications.

Acknowledgements

The authors thank the University of Nottingham Institute forMaterials Technology (UNIMAT) for funding. NRSF thanksProfessor Tom Foxon for helpful discussions, Dr PhilipMoriarty for use of AFM equipment and Dr I. Harrison foruse of PL instruments.

References

1 D. G. Baik and S. M. Cho, Thin Solid Films, 1999, 354, 227.2 Y. Ohya, M. Ueda and Y. Takahashi, Jpn. J. Appl. Phys, 1996, 35,

4738.3 Y. Natsume, Thin Solid Films, 2000, 372, 30.4 G. K. Paul, S. Bandyapadhyay and S. K. Sen, Phys. Status Solidi

A, 2002, 191(2), 509.5 D.C. Look,D.C.Reynolds, C.W. Litton, R. L. Jones, D. B. Eason

and G. Cantwell, Appl. Phys. Lett., 2002, 81(10), 1830.6 E. Y. M. Lee, N. Tran, J. Russell and R. N. Lamb, J. Appl. Phys.,

2002, 92(6), 2996.7 S. J. Pearton, C. R. Abernathy, M. E. Overberg, G. T. Thaler,

D. P. Norton, N. Theodoropoulou, A. F. Hebard, Y. D. Park,F. Ren, J. Kim and L. A. Boatner, J. Appl. Phys., 2003, 93(1), 1.

8 T. Dietl, H. Ohno, F. Matsukura, J. Cibert and D. Ferrand,Science, 2000, 287, 1019.

9 K. W. Edmonds, K. Y. Wang, R. P. Campion, A. C. Neumann,N. R. S. Farley, B. L. Gallagher and C. T. Foxon, Appl. Phys.Lett., 2002, 81(26), 4991.

10 K. W. Edmonds, P. Bogusawski, K. Y. Wang, R. P. Campion,S. N. Novikov, N. R. S. Farley, B. L. Gallagher, C. T. Foxon,M. Sawicki, T. Dietl, M. B. Nardelli and J. Bernholc, Phys. Rev.Lett., 2003, 92(3), , 037201.

11 K. Ueda, H. Tabata and T. Kawai, Appl. Phys. Lett., 2001, 79(7),988.

12 S. G. Yang, A. B. Pakhomov, S. T. Hung and C. Y. Wong, IEEETrans. Magn., 2002, 38(5), 2877.

13 H-J. Lee, S-Y. Jeong, C. R. Cho and C. H. Park, Appl. Phys. Lett.,2002, 81(21), 4020.

14 M. Ohyama, H. Kozuka and T. Yoko, Thin Solid Films, 1997, 306,78.

15 B.Wessler, F. F. Lange andW.Mader, J. Mater. Res., 2002, 17(7),1644.

16 S. Fujihara, C. Sasaki and T. Kimura, Appl. Surf. Sci., 2001, 180,341.

17 K. Nishio, S. Miyake, T. Sei, Y. Watanabe and T. Tsuchiya,J. Mater. Sci., 1996, 31, 3651.

18 D. Bao, H. Gu and A. Kuang, Thin Solid Films, 1998, 312, 37.19 Z. Jiwei, Z. Liangying and Y. Xi, Ceram. Int., 2000, 26, 883.

Fig. 8 Room temperature photoluminescence spectra of: a) sapphiresubstrate, b) Co21, c) V31, d) Mn31, e) Mn21, f) Fe31 and g) undopedfilms. Excitation wavelength 325 nm. Scans are offset for clarity.

J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 1 0 8 7 1 0 9 2 1 0 9 1

20 S. Bandyapadhyay, G. K. Paul, R. Roy, S. K. Sen and S. Sen,Mater. Chem. Phys., 2002, 74, 83.

21 Inorganic Crystal Structure Database (ICSD), CCLRCDaresburyLaboratory, UK.

22 R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr.,Theor. Gen. Cryst., 1976, 32, 751; CRC Handbook of Chemistryand Physics, 83rd edition, ed. D. R. Lide, CRC Press, London,2002.

23 A. N. Mariano and R. E. Hanneman, J. Appl. Phys., 1963, 34(2),384.

24 B. D. Yao, Y. F. Chan and N. Wang, Appl. Phys. Lett., 2002,81(4), 757.

25 S. K. Sampath, D. G. Kanhere and R. Pandey, J. Phys.: Condens.Matter, 1999, 11, 3635.

1 0 9 2 J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 1 0 8 7 1 0 9 2