pss Structural differences in Mg-doped InN indication of polytypism solidi status physica

Transcription

1 Phys. Status Solidi C 7, No. 7 8, (2010) / DOI /pss Strutural differenes in Mg-doped InN indiation of polytypism physia pss urrent topis in solid state physis Z. Liliental-Weber *,1, M. E. Hawkridge 1, X. Wang 2, and A. Yoshikawa 3 1 Materials Sienes Division, Lawrene Berkeley National Laboratory, 1 Cylotron Road, Berkeley, CA , USA 2 State Key Laboratory of Artifiial Mirostruture and Mesosopi Physis, Shool of Physis, Peking University, Beijing , China 3 Chiba University, Chiba , Japan Reeived 8 Otober 2009, revised 10 November 2009, aepted 10 November 2009 Published online 30 April 2010 Keywords InN, MBE, doping, disloations, struture, TEM * Corresponding author: Z_Liliental-Weber@lbl.gov Transmission Eletron Mirosopy shows that the InN samples doped with either inreasing or onstant Mg onentration follow a ation or anion substrate polarity. In-polar samples hange growth polarity when the Mg onentration is >10 19 m -3. N-polar samples have muh higher density of planar defets than In-polar samples and their presene leads to a derease in disloation density. In the N-polar samples equally spaed planar defets are observed for Mg onentration >10 19 m -3. Three different polytypes (2H, 3C and 4H) were observed in this type of samples. A band of planar defets with thik layers of a ubi material (3C) is observed for Mg onentration >10 20 m -3. At this Mg onentration only n-type ondutivity was reported earlier. 1 Introdution Among III-Nitrides InN was the least studied ompound beause obtaining high strutural quality was diffiult mainly due to low dissoiation temperature and high equilibrium vapor pressure of Nitrogen and weak bonding of In and N [1]. However, alloying with GaN and AlN allows light emission from ultraviolet to red due to its narrow band gap [2], overing in this way the whole solar spetrum. This ompound has a potential appliation in miroeletronis, optoeletronis and solar ells [1,2]. In order to apply InN to devies one needs to be sure that p doping in this material is possible. While undoped InN films are always n-type, p-type doping is still diffiult to ahieve due to the unusual band struture of InN [3,4] and the presene of an eletron aumulation layer [5]. Only reently was this eletron aumulation layer overome by using eletrolyte-based apaitane-voltage (ECV) measurements [6,7] and a ombination of Hall effet, photoluminesene and ECV analysis [8]. Mg doped InN thin films were onfirmed to be p-type. However, there are only few reports on strutural properties of this ompound [9-11]. Detailed knowledge on the strutural properties might lead to better understanding of eletrial measurements and shead light on an earlier dispute of band gap of this ompound [12]. In this paper the results of Transmission Eletron Mirosopy (TEM) studies will be presented on the struture of Mg doped InN layers grown by MBE. 2 Experimental Four samples doped with Mg will be desribed. All samples were grown by Moleular Beam Epitaxy (MBE) either on Ga-or N- GaN/-Al 2 O 3 substrates. The detailed growth proedure was desribed earlier [9,13,14]. The different growth polarities GaN substrates were used in order to obtain InN either with an anion or a ation growth polarity. Two samples were grown in opposite polarities with multilayers of InN:Mg, where eah layer had an equal thikness but inreased Mg onentration (1x10 18 m -3, 5.6x10 18 m -3, 2.9x10 19 m -3 and 1.8x10 20 m -3 ) interseted by thin undoped layers leading to a total layer thikness of 2.2 μm. Two other samples were grown on idential substrates in different polarities with a onstant Mg onentration of 6x10 18 m -3 and 3x10 19 m -3 and a thikness of 3.5 μm and 7.3 μm, respetively. As mentioned in earlier papers [9,13,14] different growth temperatures were used for growth on the substrates with different polarities. Different methods of transmission eletron mirosopy have been used to haraterize these layers: lassial methods for defet identifiation, lattie image for more in depth haraterization of partiular defet and Convergent Beam Eletron Diffra-

2 Report Doumentation Page Form Approved OMB No Publi reporting burden for the olletion of information is estimated to average 1 hour per response, inluding the time for reviewing instrutions, searhing existing data soures, gathering and maintaining the data needed, and ompleting and reviewing the olletion of information. Send omments regarding this burden estimate or any other aspet of this olletion of information, inluding suggestions for reduing this burden, to Washington Headquarters Servies, Diretorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington VA Respondents should be aware that notwithstanding any other provision of law, no person shall be subjet to a penalty for failing to omply with a olletion of information if it does not display a urrently valid OMB ontrol number. 1. REPORT DATE OCT REPORT TYPE 3. DATES COVERED to TITLE AND SUBTITLE Strutural differenes in Mg-doped InN - indiation of polytypism 5a. CONTRACT NUMBER 5b. GRANT NUMBER 5. PROGRAM ELEMENT NUMBER 6. AUTHOR(S) 5d. PROJECT NUMBER 5e. TASK NUMBER 5f. WORK UNIT NUMBER 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) Lawrene Berkeley National Laboratory,Materials Sienes Division,1 Cylotron Road,Berkeley,CA, PERFORMING ORGANIZATION REPORT NUMBER 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR S ACRONYM(S) 12. DISTRIBUTION/AVAILABILITY STATEMENT Approved for publi release; distribution unlimited 11. SPONSOR/MONITOR S REPORT NUMBER(S) 13. SUPPLEMENTARY NOTES See also ADM Presented at the International Conferene on Nitride Semiondutors (8th) (ICNS8) Held in Jeju, Korea on Otober 18-23, Sponsored by AOARD. 14. ABSTRACT 15. SUBJECT TERMS 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT a. REPORT unlassified b. ABSTRACT unlassified. THIS PAGE unlassified Same as Report (SAR) 18. NUMBER OF PAGES 4 19a. NAME OF RESPONSIBLE PERSON Standard Form 298 (Rev. 8-98) Presribed by ANSI Std Z39-18

3 physia p s s 2026 Z. Liliental-Weber et al.: Strutural differenes in Mg-doped InN tion (CBED) for determination of growth polarity of several areas of eah sample. 3 TEM results CBED patterns were taken on four samples in different areas. These patterns show that the InN layers start to grow following the substrate polarity. For the sample with multilayers of InN:Mg CBED patterns indiate that InN layer started to grow with In-growth polarity following the Ga-polarity of a GaN substrate. (a) (b) (b) 200 nm N In 200 nm Figure 1 (a) An area of the InN sample with a zig-zag shape inversion boundary where the growth polarity hanged from In polarity in the lower part to N-growth polarity in the upper part. (b) The entire layer of an InN sample grown with a N-polarity grown on N-polar GaN is presented. Eah double-arrow indiates where Mg onentration is hanging from 1x10 18 m -3 to 5.6x10 18 m -3 (the lowset one), then to 2.9x10 19 m -3 (entral one), and to 1.8x10 20 m -3 (the upper one). Note the formation of twins in the top part of the layer and many planar defets arranged in a spaer layers when Mg onentration is rapidly hanging. Two darker areas marked by I might be either imbedded inversion domains or areas with a different omposition. A drasti differene in strutural perfetion of the InN samples grown with different polarities was observed. The density of disloations in this sample is as high as 3.6x10 10 m -2. When the thikness of the sample is about 1.7 μm growth polarity hanged to N-polarity (Fig. 1a). This would suggest that Mg is floating to the sample surfae and most probably reahes the onentration of about 5x10 19 m -3 at whih point the growth polarity hanges. At these onditions an inversion boundary with a ontinuous zig-zag (or V) shape was formed (Fig. 1a). This type of boundary was reported earlier for InN:Mg [9] and GaN:Mg [15] approximately for the same Mg onentration. Density of disloations above this boundary was signifiantly redued. With a further inrease of Mg onentration a layer approximately a 100 nm thik of densely distributed planar defets was formed leading to a very low disloation density in the 0.2 μm of the subsurfae area (Fig. 2a). Earlier studies on similar samples [8] show that an inrease of Mg onentration above 1.8x10 20 m -3 using ECV measurements leads to a hange from p-ondutivity to n-type. I GaN N I This onentration would be expeted at the sample thikness where subsurfae planar defets are formed, assuming Mg floatation to the sample surfae. The multilayer InN:Mg sample grown on N-polar GaN has N-growth polarity (Fig. 1b). This sample has muh lower disloation density already lose to the interfae with GaN. At the top part of the layer disloation density did not exeed 3.7x10 7 m -2. One an notie that the majority of disloations in the InN layer propagated from the GaN buffer layer that was grown on Al 2 O 3 with a large lattie mismath. However, there are areas in the N-polar sample where vertial voids were formed. Some voids started from the interfae with GaN and others were formed within the layer (not shown here for a lak of spae). (a) (b) 100 nm 2nm Figure 2 (a) A band of planar defets formed in the subsurfae area in the sample that started to grow with In polarity and transferred to N polarity above a zig-zag inversion boundary; (b) high resolution image from this band of faults. Note a large thikness of a ubi (3C) material. On the same figure one an see some imbedded domains with a different diffration ontrast (marked by I in Fig.1b), most probably inversion domains or areas with a different omposition. In addition to disloations a high density of planar defets in a form of disloation loops were formed. Many of them were lining up in the areas of the undoped spaer layers (Fig. 1b). These disloation loops often interated with threading disloations leading to their lower density above the loops (see right hand side of the mirograph, Fig. 1b). Formation of these loops within the spaer layers onfirms the assumption of Mg flotation to the sample surfae and non-uniform distribution. In the upper part of the third layer where Mg onentration reahed 2.9x10 19 m -3 (where the hange of growth polarity took plae in the previous sample) many planar defets almost equally spaed (Fig. 3) are formed. With a further inrease of Mg onentration to 1.8x10 20 m -3 a density of planar defets beomes more irregular, but at a distane of about 400nm from the surfae a band of densely distributed planar defets appear again similar to that observed in the previous sample (as presented in Fig. 2). A lattie image from this area with planar defets appearing lose to the sample surfae show formation of dif-

4 Contributed Artile Phys. Status Solidi C 7, No. 7 8 (2010) 2027 ferent thiknesses of ubi material, meaning a different types of staking fault. However, in many plaes the thikness of a ubi layer is muh larger than required for the formation of any type of staking fault (I 1, I 2 or E), typial for 2H material [16]. Some top areas of the sample are ompletely onverted to ubi material and twin defets typial for a ubi material are observed (see top enter part of Fig. 1b). In the majority of areas the subsurfae areas still show 2H struture. 50 nm Figure 3 Staking faults almost equally distributed in the sample grown with N-polarity. As shown earlier [8] for this high Mg onentration (>1.8x10 20 m -3 ), a hange from p-ondutivity to n- ondutivity is taking plae in both samples that started to grow with a different polarities. For the same Mg onentration a band of planar defets is formed. It is worth noting that these defets appear always in N-polar material sine a hange of polarity took plae in the sample that has started to grow with In-growth polarity. Figure 4 Disloation distribution in a sample grown with onstant In polarity (a) and a onstant N polarity (b). The arrows indiate areas where planar defets are formed leading to disloation interation and derease their density in the upper part of the layer. For the 3μm thik sample grown with In-growth polarity with onstant Mg onentration of 6x10 18 m -3, the density of disloations is rather high and reahes 1.6 x10 10 m -2 (Fig. 4a). They are of edge type and their density does not derease with sample thikness. This is in agreement with our earlier studies of the thik In-polar samples grown in a different laboratory with Mg onentration below m -3 [10]. The 7.3 μm thik sample grown with N-growth polarity with onstant Mg onentration of 3x10 19 m -3 is the most struturally poor sample among four samples studied. The density of disloations at the area lose to the interfae with GaN is about 1 x10 11 m -2 and slightly dereases toward the sample thikness (Fig. 4b). These derease is aused by the formation of planar defets (indiated by arrows) at a disreate distanes from the interfae suh as 0.4 μm, 1.2 μm interfae and ~ 3 μm suggesting again Mg floating to the growth surfae. Due to the formation of the planar defets disloation density dereases to 2 x10 9 m -2 in the subsurfae area. (a) nm (b) Figure 5 (a) A seletive area diffration pattern (SAD) indiating division of the 0002 distane into four indiating some ordering onsistent with 4H polytype struture; (b) High resolution lattie image from the same area. In few areas of this partiular sample seletive area diffration (SAD) patterns showed additional spots dividing 0002 distane into four (Fig. 5a). High resolution images show that eah fourth 0002 lattie fringe has higher intensity and a speifi lattie arrangement suggesting some ordering onsistent with the formation of 4H polytype (Fig. 5b). The reason for the formation of suh a polytype is not lear sine in some areas of the sample this poltype starts from a non-uniform interfae, but with an ubrupt and uniform interfae this polytype starts from a band of planar defets, mentioned above, loated at some distane from the interfae. This 4H polytype might extend to another band of planar defets, formed for different sample thikness, onverting again to 2H InN at the surfae. To our knowledge this is the first experimental observation of the 4H type of polytype in this type of samples. These studies show that type of defets formed in the InN:Mg depends strongly on growth polarity. Samples that grow with N polarity are more suseptable to the formation of planar defets than these grown with In polarity. Two other strutures besides 2H have been observed: 3C and 4H struture found in this study. 4 Disussion and onlusion The addition of Mg to InN leads to the formation of different types of strutural defets. The type of defet depends on sample growth polarity. In the samples grown with N-polarity for Mg onentration >10 19 m -3, equally spaed staking faults are formed in 2H struture. For the higher onentration >10 20 m -3, a band of planar defets with a large thikness of a ubi material an be observed in the sub-surfae area. For the thik samples with a onstant Mg onentration of

5 physia p s s 2028 Z. Liliental-Weber et al.: Strutural differenes in Mg-doped InN 3x10 19 m -3, the 4H polytype ould be found in some areas of the sample. The question remains what leads to this behavior. Is it a Mg a leading fator or the fat that we are dealing at least with three elements with different radii? There are many Mg rih ternaries like Zn 1-x Mg x Se [17] or Cd 1-x Mg x Te [18] with these type of defets and polytypes but there are also other materials like Zn 1-x Cd x S and ZnS 1-x Se x [19], or SiC [20] with different dopants where polytypes are also often observed. The observation of densely distributed staking faults in InGaN and large areas of ubi materials with either high In ontent or large sample thikness [21] and InGaN:Mg [22] would suggest that different atomi radii might be the reason for the formation of different polytypes. This is an important issue sine hange of polytypes from 3C 8H 4H 2H in ZnMgSe [17] leads to a hange of a bandgap from 2.94 ev to 3.08 ev, an important issue for III-Nitrides. This would shed a light that a previous dispute of a different value of a bandgap in InN most probably was releated to strutural variety of the samples grown in different laboratories. In onlusion, one an see that In-polar samples hange growth polarity to N when Mg onentration is exeeding a speifi onentration of few times m -3. In ontrast, N polar samples do not hange growth polarity for he same Mg onentration. However, they are more vunerable to the formation of different types of planar defets (staking faults and disloation loops) and also to the formation of different polytypes. In these N-polar samples there are several areas where planar defets interat with disloations dereasing in this way their density. For both starting growth polarities a band of planar defets with extended layers of ubi material (defetive polytypes) is formed when the Mg onentration exeeds 1.8x10 20 m -3. The presene of this band of defets with a large thikness of ubi material might be responsible for n-type ondutivity reported earlier [8] for heavily Mg-doped films, sine layers with lower Mg onentration, where these defets were not observed, show p-ondutivity. Aknowledgements This work was supported by the Diretor, Offie of Siene, Offie of Basi Energy Sienes, Materials Sienes and Engineering Division, of the U.S. Department of Energy under Contrat No. DE-AC02-05CH All studied samples were grown at Chiba Univerity, Chiba, Japan. The use of the faility of the National Center for Eletron Mirosopy in LBNL, Berkeley is appreiated. [3] A.V. Blant, T.S. Cheng, N.J. Jeffs, L.B. Flannery, I. Harrison, J.F.W. Mosselmans, A.D. Smith, and C.T. Foxon, Mater. Si. Eng. B 59, (1999). [4] V.V. Mamutin, V.A. Vekshin, V.Y. Davydov, V.V. Ratnikov, Y.A. Kudriavtsev, B.Y. Ber, V.V. Emtsev, and S.V. Ivanov, Phys. Status Solidi A 176, (1999). [5] W. Walukiewiz, Physia B 302, (2001). [6] H. Lu, W.J. Shaff, L.F. Eastman, and C.E. Stutz, Appl. Phys. Lett. 82, (2003). [7] S.X. Li, K.M. Yu, J.Wu, R.E. Jones, W. Walukiewiz, J.W. Ager III, W. Shan, E.E. Haller, H.Lu, and W.J. Shaff, Phys. Rev. B 71, (2005). [8] X. Wang, S. B. Che, Y. Ishitani, and A. Yoshikawa, Appl. Phys. Lett. 91, (2007). [9] X. Wang, S. B. Che, Y. Ishitani, and A. Yoshikawa, Appl. Phys. Lett. 91, (2007). [10] Z. Liliental-Weber, in: Indium Nitride and Related Alloys, edited by T.D. Veal, C.F. MConville, and W.J. Shaff (CRS Press Taylor & Franis Group, Boa Raton, London, New York, 2008), pp [11] Y. Nanishi, T. Araki, and T. Yamaguhi, in: Indium Nitride and Related Alloys, edited by T.D. Veal, C.F. MConville, and W.J. Shaff (CRS Press Taylor & Franis Group, Boa Raton, London, New York, 2008), pp [12] T.L. Tansley and C.P. Foley, J. Appl. Phys. 59, (1986). [13] X. Wang, S.B. Che, Y. Ishitani, and A. Yoshikawa, J. Appl. Phys. Part 2 45, L730 (2006). [14] X. Wang, S,B. Che, Y. Ishitani, and A. Yoshikawa, J. Appl. Phys. 99, (2006). [15] L. Romano, J.E. Northrup, A.J. Ptak, and T.H.Myers, Appl. Phys. Lett. 77, 2479 (2000). [16] D. Hull and D.J. Baon, in: Introdution to Disloations, 3rd ed., Internat. Series on Materials Siene and Tehnology, Vol. 37 (1985), pp [17] W. Paszkowiz, P. Dluzewski, Z.M. Spolnik, F. Firszt, and H. Mezynska, J. Alloys Compd. 286, 224 (1999). [18] E. Mihalski, M. Demianiuk, S. Kazmarek, and J, Zmija, Ata Phys. Polon. A 56, 333 (1979). [19] M. J. Kozielski, J. Cryst. Growth 30, (1975) 86. [20] J.F. Kelly, G.R. Fisher, and P.Barnes, Mater. Res. Bull. 40, 249 (2005). [21] Z. Liliental-Weber, K.M. Yu, M. Hawkridge, S. Bedair, A.E. Berman, A. Emara, J. Domagala, and J. Bak-Misiuk, Phys. Status Solidi C 6(S2), S433 (2009). [22] M. E. Hawkridge, Z. Liliental-Weber, K. M. Yu, L. A. Reihertz, W.J. Shaff, J. W. Ager, and W. Walukiewiz, Phys. Status Solidi C 6(S2), S421(2009). Referenes [1] V.Y. Davydow, A.A. Klohikhin, R.P. Seisyan, V.V. Emtsev, S.V. Ivanov, F.Behstedt, J. Furthmüller, H.Harima, A.V. Mudryi, J. Aderhold, O. Semhnova, and J. Graul, Phys. Status Solidi B 229, R1-R3 (2002). [2] J. Wu, W. Walukiewiz, K.M. Yu, J.W. Ager III, E.E. Haller, H. Lu, W.J. Shaff, Y. Saito, and Y. Nanishi, Appl. Phys. Lett. 80, (2002).