DTIC ADA550039: Cathodoluminescence Study of the Properties of Stacking Faults in 4H-SiC Homoepitaxial Layers

0.08 MB·English

by  Defense Technical Information Center

APPLIEDPHYSICSLETTERS94,092101(cid:1)2009(cid:2)
Cathodoluminescence study of the properties of stacking faults
in 4H-SiC homoepitaxial layers
Serguei I. Maximenko,1,a(cid:1) Jaime A. Freitas, Jr.,1 Paul B. Klein,1 Amitesh Shrivastava,2
and Tangali S. Sudarshan2
1NavalResearchLaboratory,Washington,D.C.20375,USA
2DepartmentofElectricalEngineering,UniversityofSouthCarolina,Columbia,
SouthCarolina29208,USA
(cid:1)Received 26 November 2008; accepted 9 January 2009; published online 3 March 2009(cid:2)
In-grown stacking faults in n-type 4H-SiC epitaxial layers have been investigated by real-color
cathodoluminescenceimagingandspectroscopycarriedoutatroomandliquidheliumtemperatures.
Stacking faults with 8H stacking order were observed, as well as double layer and multilayer
3C-SiC structures and a defect with an excitonic band gap at 2.635 eV. It was found that 8H
stacking faults and triangular surface defects can be generated from similar nucleation sources.
Time-resolved measurements reveal that compared to defect-free regions, the carrier lifetimes are
severely reduced by the presence of stacking faults corresponding to triangular surface defects and
three-dimensional 3C-SiC inclusions. © 2009 American Institute of Physics.
(cid:3)DOI:10.1063/1.3089231(cid:4)
Despite significant improvements in the quality of dividual IGSFs. In addition, time-resolved SEM-CL mea-
4H-SiC bulk substrates and epitaxial films during the past surements were carried out to determine the effect of IGSFs
decade, there are still a number of crystallographic defects on the carrier lifetime.
that limit the full development of SiC-based electronic de- The 4H-SiC epitaxial films used in these experiments
vices. Since high micropipe densities are now commonly weretypically18–20 (cid:1)mthickwithn-typedopinglevelsof
avoided and are no longer credited as the most harmful de- about 7(cid:2)1015 cm−3. The homoepitaxial layers were depos-
fects for device performance,1 the elimination of other crys- ited in a vertical hot-wall inverted chimney chemical vapor
tallographic defects, particularly various types of stacking deposition (cid:1)CVD(cid:2) reactor at a temperature of (cid:5)1500 °C us-
faults (cid:1)SFs(cid:2), remains an important issue. This problem be- ing SiH Cl –C H –H precursors. Commercially purchased
2 2 3 8 2
comes more severe as off-cut angles of substrates used for off-cut (cid:1)0001(cid:2) 4H-SiC wafers oriented 8° toward the (cid:6)11¯20(cid:7)
epitaxialgrowtharedecreasedfrom8°to4°.Theseextended direction were used as substrates.ALEO 435VPSEM fitted
defects can result in significant deterioration in the perfor- witharetractableparaboliclightcollectingmirrorandliquid
mance and yields of unipolar and bipolar devices.2,3 While He cold stage was employed to obtain CLluminescence im-
SFs often form as cubic SiC (cid:1)3C-SiC(cid:2) lamella of variable agingandCLspectrainthetemperaturerangebetween5and
thicknesses embedded in the 4H-SiC host material, 8H-like 300K.TheSEMisalsoequippedwithanelectrostaticbeam
structures have also been reported.2 SFs can be introduced blanker, which allows fast time-resolved CLdecay measure-
during film growth (cid:1)in-grown SFs or IGSFs(cid:2) or by post- ments. The light emitted by the sample is dispersed by a
growth processing such as annealing, oxidation, and by me- Triax550 spectrometer fitted with a liquid nitrogen cooled
chanicalstress.Inaddition,ifShockley-typebasalplanedis- charge-coupled device (cid:1)CCD(cid:2) or with a Bialkali photomulti-
locationsarepresent,SFscandevelopundercarrierinjection plier for the luminescence decay studies.
in the active regions of bipolar devices, or they can be in- Figure 1 shows the room temperature (cid:1)RT(cid:2) CL image
duced by high power laser or e-beam excitation. In spite of from a highly defective region, which includes a variety of
their great technological importance and the relatively large defects found across the sample. The image was captured
amount of work reported in the literature, neither the defect with15keVacceleratingvoltage(cid:1)estimatedmaximumprob-
formation mechanisms nor the effect of specific SFs on the
electrical properties of epitaxial films is completely under-
stood.
SFs in SiC behave as quantum-well (cid:1)QW(cid:2) structures,4
exhibiting highly efficient radiative recombination due to
strong carrier localization in the wells. Consequently, the
presence of SFs is easily observed by various luminescence
techniques. Due to its unique characteristics, cathodolumi-
nescence (cid:1)CL(cid:2) based scanning electron microscopy (cid:1)SEM(cid:2)
provides the necessary lateral/depth resolution necessary to
investigate the optical and electronic properties of SFs.5 In
this work we employ real-color SEM-CL imaging and
SEM-CL spectroscopy techniques to detect and identify in-
FIG.1. (cid:1)Color(cid:2)(cid:1)a(cid:2)RTreal-colorCLSEMand(cid:1)b(cid:2)opticalNomarskiimages
a(cid:2)Electronicmail:[email protected]. capturedfromthesamepartofthesample.
0003-6951/2009/94(cid:2)9(cid:1)/092101/3/$25.00 94,092101-1 ©2009AmericanInstituteofPhysics
Downloaded 03 Mar 2009 to 132.250.134.32. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp
Report Documentation Page Form Approved
OMB No. 0704-0188
Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and
maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,
including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington
VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it

does not display a currently valid OMB control number.
1. REPORT DATE 3. DATES COVERED
NOV 2008 2. REPORT TYPE 00-00-2008 to 00-00-2008
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
Cathodoluminescence study of the properties of stacking faults in 4H-SiC
5b. GRANT NUMBER
homoepitaxial layers
5c. PROGRAM ELEMENT NUMBER
6. AUTHOR(S) 5d. PROJECT NUMBER
5e. TASK NUMBER
5f. WORK UNIT NUMBER
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION
Naval Research Laboratory,4555 Overlook Avenue REPORT NUMBER
SW,Washington,DC,20375
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)
11. SPONSOR/MONITOR’S REPORT
NUMBER(S)
12. DISTRIBUTION/AVAILABILITY STATEMENT
Approved for public release; distribution unlimited
13. SUPPLEMENTARY NOTES
14. ABSTRACT

15. SUBJECT TERMS
16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF
ABSTRACT OF PAGES RESPONSIBLE PERSON
a. REPORT b. ABSTRACT c. THIS PAGE Same as 3
unclassified unclassified unclassified Report (SAR)
Standard Form 298 (Rev. 8-98)
Prescribed by ANSI Std Z39-18
092101-2 Maximenkoetal. Appl.Phys.Lett.94,092101(cid:2)2009(cid:1)
ing depth of (cid:5)2 (cid:1)m(cid:2) (cid:1)Ref. 6(cid:2) and 100 nA electron beam
current. High excitation level conditions were used to allow
the simultaneous observation of defects with high and low
radiative efficiencies. Six representative types of defects ob-
servedinthesamplearemarked1through6inFig.1(cid:1)a(cid:2)and
are distinguished by their characteristic CL spectra. The tri-
angular features of different colors (cid:1)defects 2–6(cid:2), with bases
alignedparallelto(cid:6)1¯100(cid:7)andapexespointingtoward(cid:6)11¯20(cid:7)
direction, are IGSFs commonly observed in SiC epitaxial
films,7 while the defect “1” is a SF induced by the e-beam
during the experiments. An optical image (cid:3)Fig. 1(cid:1)b(cid:2)(cid:4) of the
same region, acquired with a Nomarski microscope, reveals
that many of SFs observed by CLcorrespond to well known FIG.2. Lowtemperature(cid:1)5K(cid:2)CLspectraofsomeofthedefectsindicated
triangular surface defects (cid:1)TSDs(cid:2).8 The defects marked inFig.1:(cid:1)a(cid:2)2,(cid:1)b(cid:2)4,(cid:1)c(cid:2)5,and(cid:1)d(cid:2)6.
“1–5” were characterized by a substantial luminescence in-
tensity variation across their lengths, while the defect “6” The 5 K spectrum of the SF associated with the broad
exhibited a quite uniform intensity distribution. Because the RT CLpeak observed at (cid:5)500 nm (cid:1)(cid:5)2.48 eV(cid:2) also shows
CLintensity is proportional to the concentration of e-h pairs sharp lines assigned to the 3C-SiC phonon replicas (cid:1)not de-
generated and recombining in the probed region, variations picted in Fig. 2(cid:2). The TA phonon energy yields an E of
gx
in the CL intensity are expected to result from variations in 2.506 eV.According to Ref. 13, this excitonic gap is associ-
the probed defect depth or thickness.9Therefore we can rec- ated with double layer Shockley type SFs (cid:3)(cid:1)6,2(cid:2)SSF(cid:4). The
ognize defects 1–5 as thin planar defects embedded within observationofthisdefectinlowN-dopedsamplesisunusual
thebasalplane(cid:1)0001(cid:2)4H-SiCmatrix.ThefairlyuniformCL since 2SSFs are typically observed in heavily N-doped
imagingcontrastofdefect6indicatesthepresenceofarather SiC materials subjected to a high temperature annealing
three-dimensional (cid:1)3D(cid:2) 3C inclusion inclined 8° in the basal process.14The5KCLspectrumofdefect“4”(cid:1)broadRTCL
plane.10 band at (cid:5)481 nm, (cid:5)2.58 eV(cid:2), shown in Fig. 2(cid:1)b(cid:2), also ex-
The CL spectra, acquired at 5 K and RT, were excited hibitssharp3C-SiCphononreplicasatlowtemperature.The
with15keVacceleratingvoltageand1nAbeamcurrent.The energypositionoftheTAphononreplicaofthisdefectyields
electron beam was aimed at the central part of each selected anEgx at2.635eV.Sincehigherenergyshiftoftheemission
SF to maximize the density of carriers generated at the SF lines of the SFs is expected from lamella with reduced
thickness,15thisdefectisexpectedtohaveathicknessofless
and prevent luminescence peak shifts resulting from screen-
ing introduced by boundary charge accumulation.5 In addi- than that of two double layers.
TheRTCLbandoftheIGSFsassociatedwithTSDsand
tion, the position of the electron beam along the SFs was
3D-3C inclusions, marked 5 and 6 in Fig. 1(cid:1)a(cid:2), respectively,
adjustedsothatthemaximuminthedistributionofgenerated
havepeaksat(cid:5)540 nm(cid:1)2.29eV(cid:2).ThelowtemperatureCL
carriers occurred at the depth of SF, which is approximately
41% of the electron penetration depth.9 This approach en- spectraofbothdefects,showninFigs.2(cid:1)c(cid:2)and2(cid:1)d(cid:2),exhibit
broad and poorly resolved phonon replicas. The estimated
sures similar injection conditions for all studied defects.The
excitonic gap of these defects is about 2.356 eV, which is
RT CLspectra of the selected SFs exhibited broad emission
below the reported value for bulk 3C-SiC (cid:1)E =2.39 eV(cid:2).
bands with peaks at 422, 461, 481, 500, and 550 nm, which gx
The redshift of the phonon replicas for the three 3D-3C in-
correspond to the different colors appearing in the real-color
clusions, as well as its broadening and intensity quenching,
CLimage depicted in Fig. 1(cid:1)a(cid:2).These spectra also include a
could be related to a high level of incorporated nitrogen im-
peak at (cid:5)390 nm (cid:1)3.18 eV(cid:2), which reflects the near band
puritiesand/ortothepresenceofresidualstrain.16,17Thefact
edge (cid:1)NBE(cid:2) emission from the 4H-SiC host material. De-
that IGSFs associated with surface triangular defects (cid:1)defect
fects with a blue emission peak at (cid:5)422 nm (cid:1)(cid:5)2.94 eV, 5 in Fig. 1(cid:2) have excitonic gaps similar to bulk 3C-SiC can
SF1 in Fig. 1(cid:2) are the common single layer Shockley-type be understood from recent computational studies,15,18 which
SFs (cid:1)1SSF(cid:2) (cid:1)1,3(cid:2) generated from basal plane dislocations
show that with increasing number of stacked 3C layers in a
present in the epitaxial film.11 heteropolytype structure 4H/QW/4H, the SF gap ap-
The 5 K CLspectrum of the IGSF “2” (cid:1)(cid:5)461 nm; 2.61
proaches that of the bulk 3C-SiC polytype.
eVatRT(cid:2),representedinFig.2(cid:1)a(cid:2),showsfourwell-resolved
The effect of SFs on the carrier lifetime of 4H-SiC was
and relatively sharp peaks at 2.678, 2.647, 2.629, and 2.619 probed by the time decay of the RT NBE (cid:1)(cid:5)390 nm(cid:2)
eV. Their energy separations correspond to the TA emission.19 The sample was excited with an SEM e-beam
(cid:1)(cid:5)46 meV(cid:2), LA (cid:1)(cid:5)76.6 meV(cid:2), TO (cid:1)94.5 meV(cid:2), and LO
pulse width of 50 ns and a repetition rate of 0.1 MHz. An
(cid:1)(cid:5)104.5 meV(cid:2) phonon replicas associated with the 3C-SiC accelerating voltage of 20 keVwas employed and current of
polytype and are considered typical optical signatures of 40 nA was necessary to produce sufficient signal strength.
SFs.12 Using theTAphonon spectral position as a reference, Under these conditions, the estimated injection level was
the excitonic band gap (cid:1)E (cid:2) of this defect was estimated at about (cid:5)1.5(cid:2)1018 cm−3.20 The effective carrier lifetime
gx
2.724 eV. Previous studies correlated this spectrum with a (cid:1)(cid:3) (cid:2) was obtained from the exponential decay of the lumi-
eff
Shockley-type SF of (cid:1)4, 4(cid:2) stacking structure, in terms of nescence intensity and reflects contributions from both bulk
Zhdanov notation.2 It was observed that many of the (cid:1)4, 4(cid:2) and surface recombination, (cid:3)−1=(cid:3)−1 +(cid:3)−1 .
eff bulk surf
SFspairwithTSDs(cid:3)defecttype“5”inFig.1(cid:1)a(cid:2)(cid:4),suggesting The carrier lifetime in the region near several defects of
that they have common nucleation sites. eachtypewasprobedinordertoprovidereliableresults.The
Downloaded 03 Mar 2009 to 132.250.134.32. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp
092101-3 Maximenkoetal. Appl.Phys.Lett.94,092101(cid:2)2009(cid:1)
In summary, it was demonstrated that real-color CLim-
aging provides a quick and nondestructive identification of
several different types of SFs with high spatial resolution.
CLspectra of several distinct SF structures have been deter-
mined. Time-resolved CL measurements show shorter life-
times near IGSFs and that SFs corresponding to TSDs and
3D-3C inclusions strongly reduce the carrier lifetime in the
regionneartheseSFs.Multiple-layerSFstructuresappearto
also have a strong influence on the carrier lifetime.
The research at NRLwas partially supported by the Of-
fice of Naval Research. S.I.M. thanks the NRC program at
NRLfor support.
FIG.3. (cid:1)Coloronline(cid:2)TypicalCLdecaycurvesdetectedfroma4SSF(cid:1)4,4(cid:2)
defectandfromanearbyDFR. 1R.T.Leonard,Y.Khlebnikov,A.R.Powell,C.Basceri,M.F.Brady,I.
Khlebnikov, J. R. Jenny, D. P. Malta, M. J. Paisley, V. F. Tsvetkov, R.
Zilli, E. Deyneka, H. Hobgood, V. Balakrishna, and C. H. Carter, Jr.,
majority of the SFs investigated (cid:3)(cid:1)4,4(cid:2), (cid:1)6,2(cid:2), and (cid:1)E Mater.Sci.Forum 600,7(cid:1)2009(cid:2).
gx
=2.356(cid:2) IGSFs(cid:4)reflectedaverysimilarinfluenceonthecar- 2H.Fujiwara,T.Kimoto,T.Tojo,andH.Matsunami,Appl.Phys.Lett. 87,
051912(cid:1)2005(cid:2).
rier lifetime of the surrounding 4H-SiC epitaxial layer. Fig-
3H. Lendenmann, F. Dahlquist, N. Johansson, R. Söderholm, J. Per Åke
ure 3 shows a CL decay from a region near these defects Nilsson,P.Bergman,andP.Skytt,Mater.Sci.Forum 353,727(cid:1)2001(cid:2).
compared to a decay from a neighboring defect-free region 4H. Iwata, U. Lindefelt, S. Öberg, and P. R. Briddon, Phys. Rev. B 65,
(cid:1)DFR(cid:2). As indicated in Table I, summarizing the lifetime 033203(cid:1)2001(cid:2).
measurements for the various SFs, the lifetime in the DFR 5S. Juillaguet, M. Albrecht, J. Camassel, and T. Chassagne, Phys. Status
SolidiA 204,2222(cid:1)2007(cid:2).
was found to be fairly uniform across the sample, in the 6K.KanayaandS.Okayama,J.Phys.D 5,43(cid:1)1972(cid:2).
range of (cid:3)540 ns (cid:1)but lower at the sample edges, (cid:3)277 ns(cid:2), 7G.Feng,J.Suda,andT.Kimoto,Appl.Phys.Lett. 92,221906(cid:1)2008(cid:2).
while near IGSFs the lifetime was reduced by about 40% to 8T.Okada,T.Kimoto,K.Yamai,H.Matsunami,andF.Inoko,Mater.Sci.
roughly (cid:3)330 ns. Significantly, larger lifetime reductions Eng.,A 361,67(cid:1)2003(cid:2).
9B.G.YacobiandD.B.Holt,CathodoluminescenceMicroscopyofInor-
were observed in regions containing surface triangular de-
ganicSolids,1sted.(cid:1)Springer,NewYork,1990(cid:2),p.65.
fects and 3D-3C inclusions. The decay time for one of the 10A.Shrivastava,P.Muzykov,J.D.Caldwell,andT.S.Sudarshan,J.Cryst.
doublelayerIGSFsdecreaseddramatically,similartosurface Growth 310,4443(cid:1)2008(cid:2).
triangular defects, which could be explained by the presence 11A.Galeckas,A.Hallen,S.Majdi,J.Linnros,andP.Pirouz,Phys.Rev.B
of multiple heteropolytype structures aligned in the basal 74,233203(cid:1)2006(cid:2).
plane.21 12S. Bai, R. P. Devaty, W. J. Choyke, U. Kaiser, G. Wagner, and M. F.
MacMillan,Appl.Phys.Lett. 83,3171(cid:1)2003(cid:2).
13K. Irmscher, J. Doerschel, H.-J. Rost, D. Schulz, D. Siche, M. Nerding,
TABLE I. Summary of time-resolved measurements for different defects. andH.P.Strunk,Eur.Phys.J.:Appl.Phys. 27,243(cid:1)2004(cid:2).
Indicatedlifetimeerrorsare(cid:4)2standarddeviations. 14J.Q.Liu,H.J.Chung,T.Kuhr,Q.Li,andM.Skowronski,Appl.Phys.
Lett. 80,2111(cid:1)2002(cid:2).
Label (cid:5) E (cid:3) (cid:3) 15J.CamasselandS.Juillaguet,J.Phys.D 40,6264(cid:1)2007(cid:2).
gx eff eff 16A.Henry,U.Forsberg,M.K.Linnarsson,andE.Janzén,Phys.Scr. 72,
Defect (cid:1)Fig.1(cid:2) (cid:1)nm(cid:2)atRT (cid:1)eV(cid:2)at5K (cid:1)ns(cid:2)DFR (cid:1)ns(cid:2)atSF 254(cid:1)2005(cid:2).
4SSF(cid:1)4,4(cid:2) 2 461 2.724 544(cid:4)28 326(cid:4)24 17W. J. Choyke, Z. C. Feng, and J. A. Powell, J. Appl. Phys. 64, 3163
(cid:1)1988(cid:2).
2SSF(cid:1)6,2(cid:2) 3 500 2.506 548(cid:4)24 330(cid:4)30 18M.S.MiaoandW.R.L.Lambrecht,J.Appl.Phys. 101,103711(cid:1)2007(cid:2).
? 4 481 2.635 537(cid:4)26 339(cid:4)26 19J.P.Bergman,O.Kordina,andE.Janzén,Phys.StatusSolidiA 162,65
TSDb 5 540 2.356 277(cid:4)24 15(cid:4)5a (cid:1)1997(cid:2).
3D-3Cincl.b,c 6 540 2.356 277(cid:4)25 15(cid:4)5a 20D. K. Schroder, Semiconductor Material and Device Characterization,
2nded.(cid:1)Wiley,NewYork,1998(cid:2).
aValueclosetotheexperimentalresolutionlimit. 21J.D.Caldwell,P.B.Klein,M.E.Twigg,R.E.Stahlbush,O.J.Glemb-
bMeasuredclosetothesampleedge. ocki,K.X.Liu,K.D.Hobart,andF.Kub,Appl.Phys.Lett. 89,103519
cFoundonlyaboutsampleedges. (cid:1)2006(cid:2).
Downloaded 03 Mar 2009 to 132.250.134.32. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp