optical properties of alon aluminum oxynitride

Ž .Infrared Physics & Technology 39 1998 203–211

ž /Optical properties of ALON aluminum oxynitride

T.M. Hartnett ), S.D. Bernstein, E.A. Maguire, R.W. TustisonRaytheon Electronic Systems, Lexington Laboratory, 131 Spring St., Lexington, MA 02173, USA

Abstract

Ž .The optical properties of ALON aluminum oxynitride are presented. Optical scatter and index of refraction, andabsorption of several different compositions of ALON are compared. The temperature dependence of emissivity of ALONwas measured in the temperature range 468C to 12008C. q 1998 Published by Elsevier Science B.V. All rights reserved.

Keywords: ALON, aluminum oxynitride; Optical properties; Scatter; Spectral emittance

1. Introduction

Aluminum oxynitride or ALON is a polycrys-talline ceramic material of high strength and hard-ness. This material is made by conventional sinteringand hot pressing of a powder compact. Powdercompacts of near net shape have been made byisostatic pressing, by slip casting, and by injectionmolding methods of forming. Full density and intrin-sic transparency extending from ultraviolet wave-

Ž . Ž .lengths UV to mid-infrared wavelengths MID-IRhave been achieved.

The performance of ALON windows in an opticalsystem will depend on the scatter which can cause areduction in signal-to-noise and degradation in image

w xresolution. ALON has the cubic spinel crystal struc-ture. The optical properties of ALON are thereforeisotropic and a polycrystalline ALON ceramic willnot scatter light due to birefringence. Scattering due

) Corresponding author.E-mail: thomas_m_hartnett@ raytheon.com.

to other sources is present in ALON and recentdevelopments have led to lower scatter levels thanhave previously been achieved.

In applications where the material will becomehot due to aerodynamic heating the emittance ofALON becomes important. IN this paper, recentmeasurements of scatter in ALON at several wave-lengths from the UV to the MID-IR and spectralemittance measurements made from room tempera-ture to 12008C will be presented.

Measurements of the variation of the refractiveindex with composition and wavelength are alsopresented. A comparison of the values of the refrac-tive index, dnrd t, scatter and emissivity from theliterature is made.

2. Composition range of cubic ALON

The stability range for the spinel phase, ALON, inthe AlN–Al O system is outside the ideal 9AL O2 3 2 3

Ž .P5AlN composition 35.7 mol % AlN . The constantw x w xanion model of McCauley 1 and Lejus 2 for the

Ž .spinel crystal structure given in Eq. 1 describes the

1350-4495r98r$19.00 q 1998 Published by Elsevier Science B.V. All rights reserved.Ž .PII S1350-4495 98 00007-3

( )T.M. Hartnett et al.r Infrared Physics & Technology 39 1998 203–211204

defect structure of ALON. In this equation Al isaluminum, V is cation lattice vacancy, O is oxygen,and N is nitrogen. The composition range of stabilityfor single-phase ALON was determined to extend

Ž .from 5.52 at.% nitrogen 23.7 mol % AlN to 8.20Ž .percent nitrogen 32.9 mole % AlN .

Al V O N 1Ž .Ž64qx .r3 Ž8yx .r3 32yx x

Fig. 1 graphs the lattice constant of cubic ALONvs. nitrogen content. The relationship between latticeconstant and nitrogen content was determined using

Ž .the constant anion model of Eq. 1 in conjunctionwith X-ray diffraction measurements of the latticeconstant and density measurements. Chemical analy-sis of the nitrogen content of ALON was performed

Ž .by inert gas fusion accurate to "0.1% nitrogen toverify the composition determined by X-ray diffrac-

w xtion and density measurements. Corbin 3 providesan excellent review of previous work on the compo-sition range of stability of ALON and other phasesthat can form in the AlN–AI O system.2 3

The useful transmission range of ALON extendsfrom 0.22 mm to 5.2 mm. The long wavelengthcutoff as a function of composition was determinedby measuring the transmission of two sample ofthickness and calculating the extinction coefficient.The extinction coefficient in the infrared cutoff re-gion increases with increasing nitrogen content asshown in Fig. 2. Higher temperatures will shift the

Fig. 1. Variation of the lattice constant of single phase ALONŽwith nitrogen content. Lattice constants7.9032q0.00625= at.%

.nitrogen .

Fig. 2. Extinction coefficient of ALON vs. composition in theinfrared cutoff.

infrared cutoff to shorter wavelengths. Joseph et al.w x4 have developed a multiphonon absorption modelfor ALON and several other oxide window materialsŽ .e.g., sapphire, yttria, and spinel . In their paper theypresented the absorption coefficient of these materi-als as a function of temperature and wavelength inthe infrared cutoff region. Later in this paper wepresent high temperature spectral emittance measure-ments of ALON.

3. Optical scatter in ALON

The reduction in optical scatter in transparentpolycrystalline materials like ALON is very impor-tant in for improving the performance of an opticalsystem. The most detrimental effects of scatter are areduced signal-to-noise ratio and reduced image res-olution. Unlike glass materials such as calcium alu-minate or single crystal materials such as sapphireALON contains grain boundaries. Since ALON hascubic symmetry and is optically isotropic, scatteringdue to the refractive index difference along differentcrystallographic directions is not a contributor tooptical scatter. Grain boundaries as well as the indi-vidual grains themselves are, however, potential scat-

Ž .tering sites for the following reasons: 1 refractiveindex variations due to composition variations at

Ž .grain boundaries; 2 refractive index variations dueto composition variations between adjacent grains;Ž . Ž3 structural defects dislocations, twins, stacking

. Ž .faults located at grain boundaries; 4 particulateinclusions of noncubic ALON phase or aluminum

( )T.M. Hartnett et al.r Infrared Physics & Technology 39 1998 203–211 205

Table 1Comparison of scatter in ALON

Ž .Sample Grain size mm CTIS from 2.58 to 908 at wavelength indicated

0.6328 mm 1.06 mm 3.39 mm

Ž .ALON Sample C1 151 0.15% 0.1% 0.078%Ž .ALON Sample C 84 2.72% 3.55% 5.5%

Ž .nitride; 5 refractive index variations due to compo-sition variations due to strain.

In addition to these bulk scattering sources, thereis also a contribution to the total scatter due tosurface scattering. Differences in polishing rates atgrain boundaries and at twin boundaries can result ina grain relief structure at the surface which willcause scatter. A comparison of surface scattering andbulk scattering has not yet been made for ALON.

The optical scatter and imaging performance ofALON and several other polycrystalline oxides andsingle crystal sapphire has been described previously

w xby Duncan et al. 5 . Reduction of the optical scatterin ALON has been made since their study and theserecent measurements will be examined here. Scattermeasurements on ALON test samples was performed

Ž .at TMA Technologies Bozeman, MT . Details ofw xthe measurement technique are given by Stover 6 .

The bidirectional transmissive scatter distributionŽ .function BTDF was measured at several laser

wavelengths. Using this method, the angular distribu-tion of the optical scatter is measured and calculated

Ž .total integrated scatter CTIS can be obtained. Inthis paper the effect of grain size on scatter will be

Fig. 3. Comparison of the measured BTDF for samples C and C1Ž .of ALON measured at a wavelength of 3.39 mm .

Ž .examined at wavelengths from the UV 325 nm toŽ .IR 3.39 mm .ALON samples of two grain sizes were prepared

for scatter measurements. Both large- and small-grained samples are compared in Table 1. Samples CŽ . Ž .small size and C1 large grain size are ALONsamples having grain sizes of 84 mm and 150 mm,respectively.

The scatter from ALON exhibits a maximum inscattered intensity close to the specular beam. In-creased grain size shifts this maximum in scatteredintensity to smaller angles. In Fig. 3 the angulardistribution of scatter is compared for samples C andC1 measured at 3.39 mm. The maximum in scatteredintensity can be seen for sample C. A similar com-parison can be made for other wavelengths with themaximum shifting to smaller angles with decreasingwavelength. In Table 2 the location of the maximumin scattered intensity is shown for sample as a func-tion of wavelength. Fig. 4 plots the angle at whichthe maximum occurs as a function of wavelength.Assuming the scattered light intensity maximum re-sults from a diffraction effect with an Airy distribu-tion, the particle size of the scattering sites can be

Ž . w xcalculated from Eq. 2 5 .

angle radians s1.635lrD 2Ž . Ž .where: l is the wavelength, D is the size of thescattering site.

For sample C the calculated size of the scatteringsites is 65 mm and the measured grain size 84 mm.

Table 2Angular location of the maximum in scattered intensity

Ž .Wavelength Angle sample C

3.3900 4.868

1.0600 1.588

0.63280 1.058


Fig. 4. Angle of maximum scattered light intensity vs. wavelengthfor C.

The calculated size of the scattering sites is on theorder of the grain sizes. The location of the maxi-mum in the scattered intensity distribution vs. anglemoves to smaller angles and the magnitude of themaximum decreases substantially. The measuredgrain size of sample C1 is 151 mm. Scattering sitesof this size would move the maximum in scatteredintensity to about 2.1% for sample C1.

Scatter was measured on a hemispherical ALONdome at five wavelengths in the UV to MID-IRspectral region. This dome was anti-reflection coatedon one side after polishing. On this sample moreprecise measurements were made close to the specu-lar direction than for the previous flat samples C andC1. Table 3 compares the angular and wavelengthdependence of scatter from this ‘low scatter’ ALON

Ž .dome sample dome D1, 0.050 in. thick . As indi-cated in this table a much higher degree of scatter is

Fig. 5. Angle of maximum scattered light intensity vs. wavelengthfor dome sample D1.

measured within 0.58 of the specular direction anddecreases dramatically beyond 0.58. The angular lo-cation of the maximum in scattered intensity vs.wavelength for ALON dome D1 is shown in Fig. 5.

Ž .Fitting this data to Eq. 1 the size of the scatteringsite is found to be 80 mm. The CTIS calculated from0.18 to 908 has a wavelength dependence that isproportional to the inverse 2.5 power of the wave-

Ž 2.5.length 1rl as shown in Fig. 6. This indicatesthat the scattering sites are larger than the wave-length of the light used which is consistent with thescattering site being on the order of the grain size.

4. Emittance measurements

Emittance measurements were made on a polishedALON sample having a 1-in. diameter and a 0.0775-

Table 3Comparison of the angular and wavelength dependence of scatter in ALON dome sample D1a

Ž .Wavelength CTIS % integrated scatter from angle indicated to 708 % In-line transmission

0.18 0.58 1.08 2.58

325 nm 9.0000 3.6800 0.64800 0.29600 86.400488 nm 3.0900 2.3800 0.70600 0.19200 89.500633 nm 1.9000 1.3800 0.66500 0.10200 86.5001.08 mm 0.63100 0.56800 0.54000 0.20000 90.0003.39 mm 0.07800 0.06200 0.06200 0.05700 91.800

a This sample was anti-reflection coated on one side and has a higher transmission at selected wavelengths than would uncoated ALON.


Fig. 6. Wavelength dependence of scatter in ALON.

in. thickness. The measurements were performed byŽ .Advanced Fuel Research East Hartford, CT . Spec-

tra were measured at 32 cmy1 resolution using aBomem M151 FTIR spectrometer. In the methodused, the directional–hemispherical transmittanceŽ . Ž .tl , the directional–hemispherical reflectance rl

Ž .and the sample radiance R are measured. The1Ž .spectral emittance ´l is obtained from the relation-

ship:

´ls1ytlyrl 3Ž .In the technique used, the measured radiance and

the spectral emittance are used to determine thesample surface temperature. A complete descriptionof the measurement technique is described in detail

w xin Refs. 7,8 .The spectral emittance of ALON in the 2 mm to

20 mm wavelength region at temperatures of 468Cand 5258C is shown in Fig. 7. In this figure acomparison of the measured emittance at 5258C withand without subtraction of the broad CO peak2

centered at 4.34 mm is shown. The CO peak at 4.342

mm is due to differences in the background concen-tration of CO in the air in an unpurged section of2

the emissometer setup. The apparent emittance ofgreater than one in the high temperature emission

Ž .measurement 5258C at long wavelengths is due tointerference from the flame used to heat the sample.

The region of interest for ALON is the 3 mm to 5mm wavelength region. Fig. 8 compares the emit-

Fig. 7. Spectral emission of ALON at 468C and at 5258C. Acomparison of the spectra with and without subtraction of the CO2

peak centered at 4.34 mm. Note A: The apparent emittance ofgreater than one is due to interference from the flame used to heatthe sample.

Žtance of ALON at three temperatures the CO peak2.at 4.34 mm has been subtracted from these spectra .

In Fig. 9 the integrated spectral emittance over thewavelength band 3 mm to 5 mm are plotted vs.temperature. In Fig. 10 the dependence of spectralemmittance on temperature is shown for severalwavelengths in the 3 mm to 5 mm wavelengthregion. The rapid increase in emittance at increasingwavelengths greater than 4 mm is apparent fromthese graphs. The useful range of ALON as anoptical window in applications where the windowwill be heated are limited to wavelengths shorterthan about 4 mm.

Fig. 8. Spectral emittance of ALON at several temperatures in the2 mm to 10 mm wavelength region.


Fig. 9. Integrated emittance of ALON vs. temperature in the 3 mmto 5 mm wavelength region.

w xSova et al. 9 measured the emissivity of severaloxide dome materials. ALON was measured at 6878Cand the measured results compared with a theoreticalmodel. The measured data of Sova et. al. is plottedon Fig. 10 in comparison with the current measuredvalues. The difference between the two may be dueto the interference of the CO peak in this wave-2

length of interest. Sova et. al. used a completelypurged system so the influence of atmospheric CO2

was not a factor. Also, they used laser heating of thesample rather than a flame. The flame method ofheating can also introduce CO as well as water2

vapor from the burning of fuel and oxygen.

Fig. 10. Spectral emittance of ALON vs. temperature for severalwavelengths in the 3 mm to 5 mm spectral region.

5. Refractive index of ALON

The variation in refractive index with compositionand wavelength was determined by measuring thetransmission of thin polished samples of ALON oftwo different thicknesses. Using the measured exter-

Ž . Žnal transmission t of the thinnest samples 125.mm to 150 mm thick and calculated extinction

Ž X.coefficient b , the refractive index was calculatedŽ . Ž .numerically from Eqs. 4 – 8 . Transmission mea-

surements were made using a dual beam spectro-Ž .photometer Perkin-Elmer 580B . This method is

considered to be more accurate than separate mea-surements of transmission and reflection because ofthe greater accuracy of the instrument in measuringtransmission compared to reflection.

21yR TŽ .ts external transmission 4Ž .2 21yR T

Tsexp ybX t internal transmission 5Ž . Ž .2 2ny1 qkŽ .

Rs single surface reflectance2 2nq1 qkŽ .6Ž .

bXl

ks index of absorption 7Ž .4p

t 2ln ž /t 1X

b sy extinction coefficient 8Ž .t y tŽ .2 1

where: t is the sample thickness, n is the refractiveindex, l is the wavelength.

The variation in refractive index with compositionand wavelength is shown in Tables 4–7. A compari-

Žson of the high and low nitrogen contents a s7.948o˚ ˚ .A and a s7.940 A indicates an increase in theo

refractive index with increasing nitrogen content overthe entire wavelength range. The intermediate com-

˚Ž .position a s7.944 A shows a smaller index ofo˚refraction than the a s7.940 A composition in the0

UV visible and near IR spectral region to 1 mm andthen follows the trend established by the high andthe low compositions at longer IR wavelengths. Thismay be due to lower scatter in this sample than in


Table 4˚Ž .Refractive index and extinction coefficient for ALON lattice constant a s7.940 Ao

XŽ . Ž . Ž . Ž .Wavelength mm Index of refraction n Extinction coefficient b External transmittance Single surface reflectance R

0.24 1.914 3.121 0.7901 0.09840.34 1.825 0.194 0.8408 0.08530.4 1.807 0.151 0.8458 0.08270.6 1.779 0.115 0.8530 0.07861 1.749 0.081 0.8610 0.07422.8 1.722 0.047 0.8681 0.07033.4 1.706 0.035 0.8721 0.06814 1.685 0.161 0.8761 0.06514.6 1.660 0.731 0.8761 0.06165 1.653 1.576 0.8690 0.06065.4 1.607 7.084 0.8230 0.05465.6 1.586 8.271 0.8159 0.0518

Table 5˚Ž .Refractive index and extinction coefficient for ALON lattice constanta s7.944 Ao

Wavelength Index of refraction Extinction External SingleŽ . Ž .mm n coefficient transmittance surface

XŽ . Ž .b t reflectanceŽ .R

0.24 1.907 2.862 0.7900 0.09730.34 1.820 0.208 0.8416 0.08460.4 1.804 0.162 0.8462 0.08220.6 1.770 0.128 0.8550 0.07721 1.748 0.084 0.8611 0.07412.8 1.722 0.108 0.8671 0.07043.4 1.709 0.044 0.8711 0.06864 1.692 0.159 0.8742 0.06604.6 1.671 0.729 0.8722 0.06325 1.656 1.598 0.8652 0.06125.4 1.618 7.300 0.8070 0.05625.6 1.604 8.383 0.7979 0.0543

Table 6˚Ž .Refractive index and extinction coefficient for ALON lattice constanta s7.948 Ao

Wavelength Index of refraction Extinction External SingleŽ . Ž .mm n coefficient transmittance surface

XŽ . Ž .b t reflectanceŽ .R

0.24 1.929 6.353 0.7411 0.10060.34 1.838 0.194 0.8370 0.08730.4 1.821 0.151 0.8420 0.08460.6 1.782 0.115 0.8520 0.07901 1.760 0.081 0.8580 0.07582.8 1.747 0.047 0.8617 0.07403.4 1.738 0.035 0.8640 0.07274 1.712 0.161 0.8688 0.06904.6 1.694 0.731 0.8657 0.06645 1.671 1.589 0.8600 0.06325.4 1.628 7.684 0.7929 0.05765.6 1.616 8.735 0.7830 0.0559


Table 7Comparison of the variation of the refractive index of ALON with composition and wavelength

Wavelength Index of Index of Index of SellmeierŽ .mm refraction refraction refraction formula index

Ž . Ž . Ž .n n n calculation˚a s7.940 a s7.944 a s7.948 a s7.945 Ao o o o

5.88 at.% N 6.53 at.% N 7.17 at.% N 6.69 at.% N

0.24 1.914 1.907 1.929 1.9010.34 1.825 1.820 1.838 1.8300.4 1.807 1.804 1.821 1.8130.6 1.779 1.770 1.782 1.7881 1.749 1.748 1.760 1.7742.8 1.722 1.722 1.747 1.7433.4 1.706 1.709 1.738 1.7284 1.685 1.692 1.712 1.7104.6 1.660 1.671 1.694 1.6885 1.653 1.656 1.671 1.6715.4 1.607 1.618 1.628 1.6525.6 1.586 1.604 1.616 1.642

the other two compositions. In the refractive indexcalculations the separate effects of scatter and ab-sorption are grouped together in the extinction coef-

Ž X.ficient b . This may cause a small error in theindex calculation for the more highly scattering sam-ples at shorter wavelengths.

The refractive index of ALON has been measuredby R.J. Scheller of Schott Glass Technologies, usingan ALON prism by Raytheon. The Sellmeier formuladetermined from these measurements is given in Eq.Ž .9 . The composition of this sample based on a

˚lattice constant of 7.945 A is 6.69 at.% nitrogen.w xTropf and Thomas 10 compiled a table of refractive

Ž .index values n and k for ALON based on themeasurements of Scheller and theoretical models. Acomparison of the values measured by transmissionmeasurements and calculated values using the Sell-

Ž .meier formula given in Eq. 9 is given in Table 7.Measurements of the variation in the refractive

index with temperature were made by several re-w xsearchers. Lange and Duncan 11 determined the

Ž .temperature coefficient of refractive index dnrdTusing measurements of optical path length differenceas a function of temperature in a Michelson interfer-ometer at a wavelength of 0.633 mm from roomtemperature to 5008C. Included in this reference is acomparison of dnrdT for ALON with several otheroxides. A value of 1.46"0.1=10y5r8C was ob-

tained. Tropf and Thomas measured a value of 1.5=

10y5r8C at 0.6328 mm and calculated a value ofy6 w x2.8=10 at 4.0 mm 10 .

The Sellmeier formula determined for ALON is:

20.32=109 1.287=1062n y1s q 9Ž .2 22 297,500 yÕ 530 yÕŽ . Ž .

6. Summary

Polycrystalline aluminum oxynitride with its cu-bic Spinel crystal structure is a very durable opticalmaterial with a high degree of transparency from theUV to the mid-IR wavelengths. Minimization ofgrain boundary defects reduces optical scatter signif-icantly. Levels of scatter of about 0.1% can beachieved.

Emittance of the material is an important factor insituations where elevated temperatures are encoun-tered. ALON experiences a rapid emittance increaseat wavelengths greater than about 4 mm. there is alinear relationship to temperature in the 3 mm to 5mm region.

Refractive index varies with wavelength and withcomposition of the ALON. A Sellmeier formula hasbeen derived. The temperature coefficient of refrac-tive index has been measured as 1.5=10y5r8C.


References

w x Ž . Ž .1 J.W. McCauley, J. Am. Ceram. Soc. 61 7–8 1978 372.w x2 A.M. Lejus, Sur la formation a haute temperature de spinelles

non stoetechiometriques et de phase derivees, Rev. HautesŽ .Temp. Refract. 1 1964 80–95.

w x3 N.D. Corbin, Aluminum oxynitride spinel, Int. J. High Tech-Ž . Ž . Ž .nol. Ceram. 5 3 1989 143–154, a review .

w x4 M.E. Thomas, R.I. Joseph, W.J. Tropf, Infrared transmissionproperties of sapphire, spinel, yttria, and ALON as a function

Ž . Ž .of temperature and frequency, Appl. Optics 27 2 1988239–245.

w x5 D.D. Duncan, C.H. Lange, D.G. Fischer, Proc. SPIE 1326Ž .1990 95.

w x Ž .6 J.C. Stover, Lasers and Optronics, August 1988 .

w x7 J.R. Markham, P.R. Solomon, P.E. Best, Rev. Sci. Instrum.Ž .61 1990 3700.

w x8 J.R. Markham, K. Kinsella, R.M. Carangelo, C.A. Brouil-lette, M.D. Carangelo, P.E. Best, P.R. solomon, Rev. Sci.

Ž .Instrum. 64 1993 2515.w x9 R.M. Sova, M.J. Linevsky, M.E. Thomas, F.F. Mark, High

temperature optical properties of oxide dome materials, Proc.SPIE 1760 Windows and Dome Technologies and MaterialsIII, 1992, p. 27.

w x Ž .10 W.J. Tropf, M.E. Thomas, in: E.D. Palik Ed. , Handbook ofOptical Constants of Solids II, Academic Press, Boston,Ž .1991 , p. 77.

w x11 C.H. Lange, D.D. Duncan, Temperature coefficient of refrac-tive index for candidate optical windows, Proc. SPIE 1326Windows and Dome Technologies and Materials II, 1990, p.71.

optical properties of alon aluminum oxynitride

Documents