Journal of Crystal Growth 312 (2010) 2896–2903

Contents lists available at ScienceDirect

Journal of Crystal Growth

0022-02

doi:10.1

n Corr

E-m

journal homepage: www.elsevier.com/locate/jcrysgro

Sublimation crystal growth of yttrium nitride

Li Du a, J.H. Edgar a,n, Roberta A. Peascoe-Meisner b, Yinyan Gong c, Silvia Bakalova c, Martin Kuball c

a Kansas State University, Department of Chemical Engineering, Durland Hall, Manhattan, KS 66506-5102, USAb Department of Materials Science and Engineering, University of Tennessee/High Temperature Materials Laboratory, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USAc H.H. Wills Physics Laboratory, University of Bristol, Bristol BS8 1TL, United Kingdom

a r t i c l e i n f o

Article history:

Received 4 March 2010

Received in revised form

2 June 2010

Accepted 7 June 2010

Communicated by Dr. M. Skowronskiwith a nitrogen pressure from 125 to 960 Torr. The highest rate was 9.64�10�5 mol/h (9.92 mg/h). The

Available online 11 June 2010

Keywords:

A1. Crystal morphology

A1. Crystal structure

A1. X-ray diffraction

A2. Growth from vapor

B1. Yttrium compounds

B1. Nitrides

48/$ - see front matter & 2010 Elsevier B.V. A

016/j.jcrysgro.2010.06.011

esponding author.

ail addresses: lidu@ksu.edu (L. Du), edgarjh@

a b s t r a c t

The sublimation–recombination crystal growth of bulk yttrium nitride crystals is reported. The YN

source material was prepared by reacting yttrium metal with nitrogen at 1200 1C and 800 Torr total

pressure. Crystals were produced by subliming this YN from the source zone, and recondensing it from

the vapor as crystals at a lower temperature (by 50 1C). Crystals were grown from 2000 to 2100 1C and

YN sublimation rate activation energy was 467.1721.7 kJ/mol. Individual crystals up to 200 mm in

dimension were prepared. X-ray diffraction confirmed that the crystals were rock salt YN, with a lattice

constant of 4.88 A. The YN crystals were unstable in air; they spontaneously converted to yttria (Y2O3)

in 2–4 h. A small fraction of cubic yttria was detected in the XRD of a sample exposed to air for a limited

time, while non-cubic yttria was detected in the Raman spectra for a sample exposed to air for more

than 1 h.

& 2010 Elsevier B.V. All rights reserved.

1. Introduction

The transition metal nitrides exhibit a wide range of physical(electrical, magnetic, and optical) and chemical properties that areof technological interest and have commercial applications.Examples include TiN [1] and HfN [2] diffusion barriers forintegrated circuits [3]; CrN for hard, wear resistant coatings; ScNfor high temperature Ohmic contacts to IIIA nitride semiconduc-tors [4]; and VN which is being investigated as a catalyst [5]. Thetransition metal nitrides also form alloys, which can be exploitedto control their lattice constants and electrical properties, as hasbeen demonstrated with Ti1�xScxN [6] and Y1�xScxN [7].

Many researchers are investigating the possibility of combin-ing transition metal nitrides with the IIIA nitride semiconductors(aluminum nitride, gallium nitride, and indium nitride) either aslayered structures or as alloys, to realize new functional proper-ties. The similar lattice constants and the shared common element(N) have inspired efforts to combine layers as epitaxial films.Scandium nitride [8] and zirconium nitride [9] have beenemployed as buffer layers between silicon substrates and GaNepitaxial films, to block the initiation and propagation of defects.Additions of chromium, magnesium, and iron to AlN and GaN

ll rights reserved.

ksu.edu (J.H. Edgar).

have all been studied in attempts to create a ferromagneticsemiconductor [10,11].

Yttrium nitride is particularly intriguing because it is one ofthe few transition metal nitrides that is also a semiconductor (asis scandium nitride). Several groups reported the rocksalt crystalstructure for YN with lattice constants between 4.8 and 4.9 A[12–15]. No other crystal structure has been experimentallyreported for YN, but a recent first principle calculation comparedthe wurtzite and bcc structures to the rocksalt structure (thelatter was the most stable) [16]. Although no measurement hasbeen reported, studies predicted an indirect bandgap for YN of0.8 eV [16], 0.85 eV [17], and 0.544 eV [18]. Yttrium nitride is alsopredicted to exhibit a high Mn solubility, which could impart itwill good magnetic properties while retaining its semiconductorproperties [19].

In the past, only a few studies have reported the synthesis of YN. Inthe 1950s, a group produced YN powder by first converting yttriummetal to YH2 by reacting with hydrogen at 550 1C in a quartz tube,then heating this gas to 900 1C in the presence of nitrogen [12]. Laterin the 1960s, YN powders were obtained by reacting yttrium metalwith nitrogen at 1400 1C [13] and arc-melting under 0.3 MPa nitrogen[14]. Recently, YN thin films were grown on both silicon and sapphiresubstrates by laser ablation deposition [20] and reactive magnetronsputtering [21], respectively. Although the lattice constants reportedfrom these different material preparation methods are very close,there are still variations.

L. Du et al. / Journal of Crystal Growth 312 (2010) 2896–2903 2897

In the present study, the sublimation–recondensation growthmethod was employed to produce YN bulk crystals. Thistechnique is attractive for its ability to produce bulk crystalswith much lower dislocation densities than in thin films onforeign substrates. In addition, its growth rate can be orders ofmagnitude higher than thin film techniques, i.e. greater than10 mm/h. Previously, our group showed ScN [22] and TiN [23]crystals produced by this technique have defect selective etch-pitdensities on the order of 106 cm�2. The YN growth process wasanalyzed and the materials produced were thoroughly character-ized. The YN crystal morphology was studied by optical andscanning electron microscopy, while its crystal structure andlattice constants were evaluated by X-ray diffraction. Thedependence of the YN growth rate on temperature and pressurewas established and compared with ScN and TiN sublimationgrowth under similar conditions. Lastly, the stability of the YNcrystals in air was examined, and the resulting oxidation productsreported.

2. Experimental

The experiments started with YN source synthesis, which wasused for the YN crystals growth. Both YN synthesis and crystalgrowth were conducted in a resistively heated tungsten furnacewith tungsten wire mesh heating elements that provided an axialtemperature drop between the source and crystal growth zones. Acovered tungsten crucible within a covered tungsten retort wasused as the reactor/crystal growth chamber, and the growthtemperature was measured by an optical pyrometer focused onthe top of the retort.

The YN source was produced by nitridizing the yttrium metal(99.9%) in ultrapure nitrogen atmosphere of 800 Torr at 1200 1C(Fig. 1). The conversion of the yttrium metal to YN was monitoredby measuring the mass change of Y/YN solid mixture in thecrucible to determine its Y/N ratio. The Y/YN mixture was groundinto small pieces after the weight measurement, to ensure good

Fig. 1. The sketch of Y metal nitridizing setup. The metal chunks turned into

nearly black lump 1 h later and were ground before continuing nitridizing. The

color of Y/YN changed to blue-green as the nitridizing time was increased.

contact of nitrogen gas with the mixture in the subsequentnitridizing step.

In the YN crystal growth process, tungsten foil (25 mmdiameter) served as the substrate for the deposited YN crystals(Fig. 2). The YN source was annealed before each growth to reducethe surface oxide as YN crystal easily reacts with oxygen andmoisture in ambient air. First, the YN source was baked in a 5%hydrogen and 95% argon gas mixture at 900 Torr and 1200 1C forabout 4 h. Then it was heated in ultra pure nitrogen gas at250 Torr and 1900 1C for another 8 h. A similar process has proveneffective in reducing the surface oxide in our AlN [24] and TiN [23]crystal growths. The YN growth rate was investigated as afunction of growth temperature (2000–2100 1C), pressure(125–960 Torr), and time (16–48 h). A temperature difference ofapproximately 50 1C was maintained between the source andcrystal growth zones. The overall growth rate was determined bydividing the YN crystal mass increase by the growth time. Afterthe YN crystals were synthesized, some were sealed into glasstubes at high vacuum (2�10�5 Torr) to prevent oxidation duringstorage.

Characterization of the YN crystals included Raman spectro-scopy and X-ray diffraction. Raman measurements were per-formed with a Renishaw InVia micro-Raman system using a488 nm Ar ion (60 mW) laser as excitation source. To identify thephases presented by x-ray diffraction, both the source materialsand grown crystals were ground into fine powders and stored inan air sensitive sample holder (to prevent oxidation). The XRDdata were collected on a PANalytical X’Pert Pro diffractometerwith Cu Ka radiation (45 kV/40 mA) using parabolic mirrors withparallel plate collimator (0.091) and miniprop point detector.Continuous scans were taken between 201 and 1001 (2y).Diffraction patterns were analyzed using the ICDD powderdiffraction files and JADE or High Score software packages toidentify crystalline materials. The grain sizes (the projected areaof individual crystals) were measured using scanning electronmicroscopy (SEM, S-3500N, Hitachi Science Systems, Ltd., Japan)and averaged from a number of condensed YN crystals

Fig. 2. Schematic sketch of YN crystal growth. Tungsten foil was added and served

as substrate.

0

0.2

0.4

0.6

0.8

1

0Time (hour)

N/Y

Ato

m R

atio

0.79

0.89

1 2 3 4 5

0.90 0.91

Fig. 3. The calculated nitrogen to yttrium atom ratio versus nitridizing time in

ultrapure nitrogen atmosphere of 800 Torr at 1200 1C.

Fig. 4. Optical microscope image of YN crystals on tungsten. The As-produced YN

is gray-green and cube or cubiod is the regular shape of the crystals.

L. Du et al. / Journal of Crystal Growth 312 (2010) 2896–29032898

3. Results and discussion

Yttrium metal has a melting point of 1509 1C and a boilingpoint of 3030 1C [25], and it is easily nitridized at 1200 1C in apure nitrogen atmosphere with the possible reaction of:

Y ðsÞþ0:5N2 ðgÞ2YN ðsÞ ð1Þ

In analyzing yttrium nitridation, the loss of yttrium by vaportransport to the furnace chamber was negligible; it is preventedby employing double containers (crucible and retort). Thus, themass increase is due primarily to the addition of nitrogen to thesolid. From the change in mass, the nitrogen to yttrium atom ratiocan be calculated. Fig. 3 showed the calculated nitrogen toyttrium atom ratio versus nitridizing time. After only 1 h ofnitridizing, the nitrogen to yttrium atomic ratio increased rapidlyfrom 0 to 0.79. Subsequently, the rate slowed, increasing to 0.89and 0.90 after the second and third hours, respectively. This ratiostayed nearly constant around 0.91 with longer nitridation times.Although the Y/YN solid mixture was ground into powder everytime before the next nitridizing step, heating still caused it tosinter together. This sintering and reduction of exposed surfacearea may slow down nitridation, preventing its completion. Afternitridizing for more than 90 h, the source materials were metallicgray blue-green, with the cubic crystal grains formed on thesource surface. X-ray diffraction confirmed that there was no un-reacted yttrium metal left in the source material.

Thus the reactions involved in the crystal growth process wereyttrium nitride sublimation in the source zone at a relatively hightemperature and recondensation in the crystal growth zone at arelatively low temperature

2YN ðsÞ22Y ðgÞþN2 ðgÞ: ð2Þ

3.1. Morphology

The as-received yttrium was hard silvery metal chunks. After1 h nitridizing, the chunks turned black and were easily fractured.As the nitridizing time was increased, the color of the Y/YNmixture changed towards the blue-green. In the end, the YNproduced by sublimation crystal growth is a gray blue-greencrystal as shown in Fig. 4.

Fig. 5 shows the SEM image for YN crystals grown for 16 h at960 Torr, 2000 1C. After 16 h growth, although most grains wereregularly shaped cubes or cuboids, wedge, tetrahedron, pyramid,

frustum, truncated octahedrons, and truncated tetrahedrons werealso observed. As the grains merged together, they displayed lessregular shapes, and the surfaces became rough. Straight lines onthe perimeters of the crystals and microhollows (upper-left inFig. 5) with size about 2–3 mm on the grain boundary wereobserved. The straight lines were seen on all the samples whilemicrohollows were found mostly on samples grown at highertemperature or longer time. The line features were found on themerged grains (Fig. 5b1), on the grains grown on substrate grainboundary (Fig. 5b2), and also on single grain grown on singlesubstrate grain (Fig. 5b3). If the lines occurred at regular shapedgrains (cubes or cuboids), they usually were at an angle of 451 tothe square edges. For the same growth time and pressure, theaverage crystal size for higher growth temperature was larger,increasing from 1.5�10�3 at 2000 1C to 2.5�10�3 mm2 at2050 1C (Fig. 6a) and 5.0�10�3 mm2 at 2100 1C(Fig. 6b–d). Thisresult was the same for other transition metal nitride ScN [22]and TiN [23] crystal growth in our previous studies. However, thelines and microhollows became more evident at highertemperature as more crystal grains merged together. Fig. 6bshows the nearby microhollows on the crystal boundary 16 h at960 Torr, 2100 1C. Whereas single crystal grains on the samplestill maintained cube, pyramid (Fig. 6c), or frustum (Fig. 6d)shapes.

Reducing the pressure also has a significant effect on growthrate. Growing crystals for the same amount of time andtemperature but at a lower pressure produced larger crystals.For a YN sample grown for 16 h at 740 Torr, 2000 1C, theestimated average grain size of crystals was 10�10�3

mm2—much larger than one grown for 16 h at 960 Torr,2100 1C. This result is also confirmed for ScN [22] and TiN [23]crystal growth in our previous studies.

3.2. Overall growth rate

The temperature and pressure effects on the sublimation growthof YN is the same as the other transition metal nitrides, ScN and TiN;

Fig. 5. SEM top view of YN crystals grown for 16 h at 960 Torr, 2000 1C. Most individual grains were regular shaped, while merged grains were less regular, and had rough

surfaces. Lines (b) on the surface and microhollows (a) with size about 2–3 mm on the grain boundaries were observed.

Fig. 6. SEM images of YN crystals grown for 16 h at 2050 1C, 960 Torr (a) and 2100 1C, 960 Torr (b–d). Most grains were cubes and cuboids (a), other shapes like wedge (c),

and fustum (d) were also observed; micro hollows became more evident at higher temperature as more crystal grains merged together (b).

L. Du et al. / Journal of Crystal Growth 312 (2010) 2896–2903 2899

increasing the temperature or decreasing the pressure enhances itscrystal growth rate. However, the growth rate of YN is lower thaneither ScN or TiN. The YN sublimation growth rate was 3.24�10�5

mol/h (3.33 mg/h) at 2000 1C, 125 Torr, while the ScN sublimationgrowth rate from our unpublished work was 9.65�10�5 mol/h(5.69 mg/h) at same temperature but higher pressure of 150 Torr.The highest growth rate for YN in this study was 9.64�10�5 mol/h

(9.92 mg/h) at 2100 1C and 125 Torr, while the TiN had a growthrate of 3.22�10�4 mol/h (19.92 mg/h) at same temperature buthigher pressure of 150 Torr from our previous study [26]. Decreasingthe pressure has a more significant effect on overall growth ratethan increasing the temperature Figs. 7 and 8 show the dependenceof the growth rate on temperature and pressure: the logarithmicgrowth rate is first order dependent on reciprocal temperature,

200020502100

2

5

10

1 Gro

wth

Rat

e m

g/hr

-4.5

-5.5

-6.5

-7.5

Ln (r

)

0.42 0.425 0.43 0.435 0.44

125 torr 960 torr

R2 = 0.9999

R2 = 0.9992

100/T (K)

T C

Fig. 7. The variation of the logarithmic growth rate on reciprocal growth

temperature at constant growth pressure.

125960 P torr

0 0.002 0.004 0.006 0.008 0.011/P (1/torr)

2050 C 2100 C

R2 = 1

R2 = 0.9998

10

8

6

4

2

0

Gro

wth

rate

mg/

hr

740

Fig. 8. The growth rate variation with reciprocal pressure at constant growth

temperature.

Fig. 9. X-ray diffraction for YN crystals g

L. Du et al. / Journal of Crystal Growth 312 (2010) 2896–29032900

while growth rate itself varies approximately linearly withreciprocal pressure.

The activation energy of the growth was determined by com-bining the Arrhenius equation with sublimation–recondensationkinetics:

r¼ da=dt¼ kf ðaÞ

lnr¼ �Ea

RT

� �þ lnA

9=;) ln

dadt¼ �

Ea

RT

� �þ lnAþ ln f a½ � ð3Þ

where Ea is the activation energy, A is the pre-exponential factoror frequency factor, R is the ideal gas constant, r is the growthrate, a is the degree of conversion, and f ðaÞ is a mathematicalfunction whose form depends on the reaction type.

In Eq. (3), although the relationship between the rate constantk and the growth rate (ðda=dtÞ ¼ kf ðaÞ) has not been determinedfor Y+N system, we previously showed that if the equilibrium canbe assumed at the surfaces of the source and seed, f(a) isapproximately constant [25]. This is demonstrated as follows.

Since the transport of growth species from source to growingsurface is often the rate limiting step in vapor crystal growth [20],equilibrium can be assumed at the surfaces of the source andseed. For reaction (4):

Keq ¼½Y�2½N2�

½YN�2¼ðf eq

Y Þ2ðf eq

N2Þ

ðaeqYNÞ

2ð4Þ

The growth pressures used in this study were not high(o1.25 atm), so solid species activity is close to unity and gasfugacity can be represented by gas species partial pressure, that is:

Keq ¼ ðPeqY Þ

2ðPeq

N2Þ ð5Þ

For a fixed growth temperature and pressure, the local gasphase species concentration at the growth surface remains thesame and f(a) is approximately constant. Therefore, the sublima-tion growth rate is proportional to the rate constant and yields afirst order dependence of logarithmic growth rate on reciprocaltemperature at same pressure (confirmed in Fig. 7)

dlnr

dT¼

Ea

RT2r¼

dadt

� �ð6Þ

rown for 48 h at 2000 1C, 740 Torr.

L. Du et al. / Journal of Crystal Growth 312 (2010) 2896–2903 2901

The deduced YN sublimation activation energy is467.1721.7 kJ/mol (from Fig. 7), which is similar to the activationenergy of ScN (456.0 kJ/mol [22]), and lower than that of TiN(775.8729.8 kJ/mol [26]).

Table 1X-ray powder diffraction data for YN.

Peak no. (h k l) 2y d-spacing (A) Intensity (%)

1 (1 1 1) 31.7285 2.81790 100.00

2 (2 0 0) 36.7998 2.44038 89.54

3 (2 2 0) 53.0250 1.72561 51.05

4 (3 1 1) 63.1269 1.47160 34.11

5 (2 2 2) 66.2843 1.40895 15.23

6 (4 0 0) 78.2915 1.22019 6.58

7 (3 3 1) 86.9343 1.11972 13.24

8 (4 2 0) 89.7900 1.09137 18.49

Table 2literature information for YN structure reference from [12] to [15]; calculated from

LPF using POWD-12++.

Reference

code

Lattice

constantSample preparation

This study 4.88075 AGray-green color, prepared by sublimation at

2000 1C, 960 Torr

[12] 4.877 AObtained by converting yttrium metal to YH2 then

to YN (powder diffraction)

[13] 4.8699 ABlue-violet color, prepared by reaction of metal at

1673 K (powder diffraction)

[14] 4.8920 AArc-melted under 0.3 MPa nitrogen (powder

diffraction)

[15] 4.8935 AAnnealed at 1673 K under argon or nitrogen

(powder diffraction)

[21] 0.491(3) nm Reactive magnetron sputtering

Fig. 10. Raman spectrum of YN crystals grown for 48 h at 2000 1C, 760 Torr. Photolumi

subtracted) from three different crystals, showing consistent Raman spectra over diffe

3.3. Composition and structure

The YN crystals grown for 48 h at 2000 1C and 740 Torr wereground into a fine powder and examined by X-ray diffraction(Fig. 9). Table 1 lists the peak position, peak intensity, d-spacing,and corresponding crystal planes of the eight strongest peaks. Thespace group of the sample was identified as Fm-3 m and thecrystal structure was confirmed as rocksalt (NaCl), with 4YN perunit cell. The calculated lattice parameters were 4.88 A (a¼b¼c)and 901 (a¼b¼g) with a cell volume of 116.27�105 pm3.The calculated density for produced YN crystals was 5.87 g/cm3

at 25 1C. YN structure information from literatures is listed inTable 2 and their calculated lattice constants bracketed ourrefined value. Since these YN crystals were produced at differentconditions, it is possible that variations in the stoichiometry andresidual impurity concentrations may have caused the smallfluctuation lattice constant values.

Fig. 10 shows the Raman spectrum of YN crystals grown for48 h at 2000 1C and 760 Torr. First order Raman scattering issymmetry forbidden for the rock salt crystal structure. Theapparent features in the Raman spectrum are disorder-inducedfirst-order Raman scattering and correspond to the densityof phonon states. Three crystals were selected for analysis(inset) and they all have similar spectra. Note that thephotoluminescence was superimposed on the Raman spectra inthe spectral range of the Raman spectra, i.e., as using 488 nm laserexcitation, around this wavelength. There is no literature data onRaman spectra of YN, thus, we compared our results with that ofother transition metal nitrides, e.g. ZrN (same rock salt structurewithout greatly changing the mass of the metallic ion; due todifferent electronic structure an effect of the free carriers onthe phonon modes however may be expected) [27] and ScN (samerock salt structure, close electronic structure, but the atomic massratio is MSc/MYE0.5, implying that the Raman peaks are shifted infirst approximation as oYN/oScNE(MSc/MY)1/2

¼0.711) [28].Hence, the broad Raman peaks at �318 and 360–400 cm�1

nescence background is subtracted. Inset shows Raman spectrum (background not

rent crystals measured.

Fig. 11. The apparent color of YN crystals changed with time upon exposure time to ambient air (optical micrographs). The time interval between each image is 5 min.

L. Du et al. / Journal of Crystal Growth 312 (2010) 2896–29032902

could be attributed to acoustic phonons in cubic YN. The broadpeak with lower intensity at �500 cm�1 most probablyoriginated from optical phonons.

3.4. Oxidation

The yttrium nitride crystals readily oxidized in air at roomtemperature. After exposure to air for 20 min, the gray-greencolor turned into blue-violet, and 2 h later it changed to black. Thecolors observed under microscope with strong reflection light(Fig. 11) included red, orange, yellow, green, blue, indigo, violet,and mauve. Photos were taken every 20 min. Sample crystalsshowed a yellow color 5 min after initial air exposure (Fig. 11a),and turned to orange 20 min later (Fig. 11b). The sample thenlooked blue-violet 40 min after the first photo was taken (Fig. 11c)and changed to green-yellow 20 min later (Fig. 11d). These colorswere repeated as the thickness of the oxidation layer increased,and the surface roughness increased.

Even for YN samples sealed under a high vacuum in a glasstube and ground to powder in an air-free glove box before X-raydiffraction, oxidation products were still observed. The tiny peak(3.86% intensity) appearing at 29.1665 (2y) in Fig. 9 did not

originate from YN. Rietveld refinement for the XRD pattern of YNin an air sensitive sample holder indicated that roughly 96 wt% ofthe examined sample was yttrium nitride; the remaining 4 wt%was crystalline yttrium oxide. The space group of this yttriumoxide is Ia-3, the calculated O and Y ratio in its unit cell is 3:2, andthe lattice parameters are 10.598 A (a¼b¼c) and 90o (a¼b¼g),which suggested the formation of c-type yttrium oxide afterlimited air exposure.

Fig. 12 is the Raman spectra from an YN crystal sample afterexposed to air for different time durations. A sharp peak at�587 cm�1 was observed after the sample was exposed to air for1 h; it became more pronounced with time. This peak position isclose to that reported for yttrium oxide [27] but not the cubic(c-type) structure. This result implies the formation of non-cubicyttrium oxide, monoclinic (b-type) and rhombohedral (a-type)after exposure to air for more than 1 h.

4. Conclusions

With fully nitridized yttrium as the source, the sublimation–recondensation technique proved viable for producing YN bulk

Fig. 12. Raman spectra for YN crystal after exposed to air for different duration of

time.

L. Du et al. / Journal of Crystal Growth 312 (2010) 2896–2903 2903

crystals on tungsten substrates. The growth temperature andpressure are the major factors that impact the morphology andgrowth rate of YN crystals. Experiments proved that the growthrate increases exponentially with temperature, inversely with thetotal pressure, and this trend is the same with ScN and TiNcrystal growth under sublimation recondensation technique. Thehighest growth rate was 9.64�10�5 mol/h after 24 h of growthat 2100 1C and 125 Torr. Combining the Arrhenius equationwith growth kinetics, the deducted activation energy was467.1721.7 kJ/mol. Crystal structure of YN was confirmed asrocksalt (NaCl) with 4YN in the unit cell by X-ray diffraction. Thecalculated lattice constant was 4.88 A, which is bracketed byvalues reported in the literature. YN has a poor stability in the air.XRD patterns indicated presence of cubic yttrium oxide in YNsample after exposure to air for limited time, while the Ramanspectra implied the present possibility of non-cubic yttrium oxidein samples that exposure to air for more than 1 h.

Acknowledgement

The support of II–VI Foundation is greatly appreciated. X-raydiffraction analysis at the Oak Ridge National Laboratory’s HighTemperature Materials Laboratory was sponsored by the U.S.Department of Energy, Office of Energy Efficiency and RenewableEnergy, Vehicle Technologies Program. We would like to thankMs. Yi Zhang and Mr. Clinton Whiteley for sample preparation.

References

[1] M.Y. Kwak, D.H. Shin, T.W. Kang, K.N. Kim, Characteristics of TiN barrier layeragainst Cu diffusion, Thin Solid Films 339 (1999) 290–293.

[2] R.A. Araujo, X. Zhang, H. Wang, Cubic HfN thin films with low resistivity on Si(001) and MgO (001) substrates, J. Electron. Mater. 37 (2008) 1828.

[3] B. Navinsek, P. Panjan, I. Milosev, Industrial applications of CrN (PVD)coatings, deposited at high and low temperatures, Surf. Coat. Technol. 97(1997) 182–191.

[4] X.W. Bai, M.E. Kordesch, Structure and optical properties of ScN thin films,Appl. Surf. Sci. 175-176 (2001) 499–504.

[5] H. Kwon, S. Choi, L.T. Tompson, Vanadium nitride catalysts: synthesis andevaluation for n-butane dehydrogenation, J. Catal. 184 (1999) 236.

[6] J.M. Gregoire, S.D. Kirby, M.E. Turk, R.B. Van Dover, Structural, electronicand optical properties of (Sc,Y)N solid solutions, Thin Solid Films 517 (2009) 1607.

[7] D. Gall, M. Stoehr, and J.E. Greene, Vibrational modes in epitaxialTi1-xScxN(001) layers: an ab initio calculation and Raman spectroscopy study,Phys. Rev. B, 64, 174302.

[8] M.H. Oliver, J.L. Schroeder, D.A. Ewoldt, I.H. Wildeson, V. Rawar, R. ColbyP.R. Cantwell, E.A. Stach, S.D. Sands, Organometallic vapor phase epitaxialgrowth of GaN on ZrN/AlN/Si substrates, Appl. Phys. Lett. 93 (2008) 023109.

[9] C.F. Johnston, M.A. Moram, M.J. Kappers, C.J. Humphreys, Defect reduction in(1122) semipolar GaN on m-plane sapphire using ScN interlayers, Appl. Phys.Lett. 94 (2009) 161109.

[10] S.G. Yang, A.B. Pakhomov, S.T. Hung, C.Y. Wong, Room-temperaturemagnetism in Cr-doped AlN semiconductor films, Appl. Phys. Lett. 81(2002) 2418.

[11] M. Zajac, J. Gosk, E. Grzanka, M. Kaminska, A. Twardowski, B. StrojekT. Szyzko, S. Podsiadlo, Possible origin of ferromagnetism in (Ga, Mn)N,J. Appl. Phys. 93 (2003) 4715.

[12] Charles P. Kempter, N.H. Krikorian, Joseph C. McGuire, The crystal structure ofyttium nitride, J. Phys. Chem. 61 (1957) 1237.

[13] F. Anselin, C.r. Hebd Seances Acad. Sci. 256 (1963) 2616.[14] H. Holleck, Smailose E. Thummler, J. Nucl. Mater. 32 (1969) 281.[15] H.P. Klesnal, P. Rogl, High Temp. High Pres. 22 (1990) 453.[16] S. Zerroug, F. Ali Sahraoui, N. Bouarissa, Ab initio calculations of yttrium

nitride: structural and electronic properties, Appl. Phys. A 97 (2009)345–350.

[17] C. Stampfl, W. Mannstadt, R. Asahi, A.J. Freeman, Electronic structure andphysical properties of early transition metal mononitrides: density-func-tional theory LDA, GGA, and screened-exchange LDA FLAPW calculations,Phys. Rev. B 63 (2001) 155106.

[18] T. Lu, M. Huang, Electronic structure of ScN and YN: density-functionaltheory LDA and GW approximation calculations, Chin. Phys. 16 (2007) 62–66.

[19] X. Jia, W. Yang, M. Qin, Magnetism in Mn doped yttrium nitride: first-principles calculations, Appl. Phys. Lett. 93 (2008) 222501 - 222501-3.

[20] W. De La Cruz, J.A. Diaz, L. Mancera, N. Takeuchi, G. Soto, Yttrium nitride thinfilms grown by reactive laser ablation, J. Phys. Chem. Solids 64 (2003)2273–2279.

[21] John M. Gregoire, Steven D. Kirby, George E. Scopelianos, Felix H. Lee, R. Brucevan Dover, High mobility single crystalline ScN and single-orientationepitaxial YN on sapphire via magnetron sputtering, J. Appl. Phys. 104(2008) 074913.

[22] Z. Gu, J.H. Edgar, J. Pomeroy, M. Kuball, D.W. Coffey, Crystal growth andproperties of scandium nitride, J. Mater. Sci.: Mater. Electron 15 (2004) 555.

[23] Li Du, J.H. Edgar, Edward A. Kenik, Harry Meyer III, Sublimation growth oftitanium nitride crystals, J. Mater. Sci.: Mater. Electron 21 (2010) 78–87.

[24] J.H. Edgar, L. Du, R.G. Lee, L. Nyakiti, J. Chaudhuri, Native oxide andhydroxides and their implications for bulk AlN crystal growth, J. Cryst.Growth 310 (2008) 4002–4006.

[25] H.E. Boyer, T.L. Gall (Eds.), Metals Handbook—Desk Edition, ASM, MetalsPark, OH, 1985.

[26] D.W. Greenwell, B.L. Markham, F. Rosenberger, J. Cryst. Growth 51 (1981) 413.[27] W. Spengler, R. Kaiser, Solid State Commun. 18 (1976) 881.[28] G. Travaglini, F. Marabelli, R. Monnier, E. Kaldis, P. Wachter, Phys. Rev. B 34

(1986) 3876.