p s sreprints

statu

s

soli

di

www.interscience.wiley.comph

ysi

ca

REPRINT

www.pss-rrl.com

www.pss-c.com

www.pss-b.com

www.pss-a.com

© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

p s scurrent topics in solid state physics

c

sta

tus

so

lid

i

www.pss-c.comph

ysic

aPhys. Status Solidi C 7, No. 1, 44–47 (2010) / DOI 10.1002/pssc.200982630

Study of implantation-induced blistering/exfoliation in wide bandgap semiconductors for layer transfer applications

R. Singh*, 1, U. Dadwal1, R. Scholz2, O. Moutanabbir2, S. Christiansen2, and U. Gösele2

1 Department of Physics, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India 2 Max Planck Institute of Microstructure Physics, Weinberg 2, Halle 06120, Germany

Received 21 June 2009, revised 3 October 2009, accepted 4 October 2009 Published online 9 December 2009

PACS 61.80.Jh, 68.37.Lp, 68.37.Ps, 81.05.Ea, 81.40.Ef * Corresponding author: e-mail rsingh@physics.iitd.ac.in, Phone: +91 11 2659 6495, Fax: +91 11 2658 1114

We present an overview of the implantation-induced blister-ing/exfoliation process in wide bandgap semiconductors such as GaN, AlN and ZnO. These semiconductors were implanted with 50 keV hydrogen ions at various fluences and subse-quently annealed at higher temperatures up to 800 °C in order to trigger surface blistering. In the case of GaN and AlN, a de-tailed blistering kinetics investigation was also performed. Various techniques such as optical microscopy, atomic force microscopy, cross-sectional transmission electron microscopy

and stylus profilometry were used for the characterization of the implanted samples. In the case of room temperature im-planted samples, it was observed that the damage band formed inside the samples was decorated with hydrogen filled nanovoids. These nanovoids served as precursors for the for-mation of two-dimensional extended defects called nanoc-racks and microcracks upon annealing and eventually led to surface blistering or exfoliation.

© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1 Introduction Wide bandgap semiconductors such as GaN, AlN and ZnO are of great importance due to their applications in electronic and optoelectronic devices [1-4]. The growth of epitaxial layers for device fabrication is mostly carried out on foreign substrates such as sapphire, SiC and Si due to the fact that high structural quality single crystal substrates of these wide bandgap semiconductors are very expensive and mostly available in small sizes [5-7]. Hydrogen implantation-induced layer splitting of semi-conductors in combination with direct wafer bonding can potentially be used to provide inexpensive templates of high structural quality [8-10]. Using this method the free-standing wafers of these semiconductors can be utilized to transfer multiple layers on to other foreign substrates, which in turn serve as template layer for the growth of epi-taxial device layers [8-10]. For this process to occur a nar-row parameter window of implantation dose, implantation temperature, annealing temperature and time has to be de-fined since all these parameters depend strongly on the type of semiconductor material used for implantation [9]. The physical mechanisms leading to the process of layer splitting can be conveniently investigated by studying the development of surface blisters/exfoliation in hydrogen

implanted and annealed but unbonded wafers [9]. Here an overview of the hydrogen implantation-induced blister-ing/exfoliation in GaN, AlN and ZnO has been presented. These semiconductors were implanted with 50 keV H+ ions at various fluences and after that annealed at higher tem-peratures up to 800 oC in order to observe the formation of surface blisters/exfoliation. The microscopic process lead-ing to the formation of surface blisters/exfoliation has also been discussed.

2 Experiment The samples used in this work were GaN and AlN epitaxial layers with a thickness of 4 and 2 µm, respectively, grown on 2-inch (0001) c-plane sapphire substrates. In case of ZnO, c-plane oriented bulk crystals having thickness of 500 µm were used. All these samples were implanted with 50 keV H+ ions with various fluences in the range of 5×1016 to 6×1017 cm-2. It is to be mentioned here that in some cases we have used 50 keV H+ ions [15] while in other cases we used 100 keV H2

+ ions [11,12,16-18] for implantation. But 100 keV H2

+ ions are equivalent to 50 keV H+ ions with a dose that is two times of the latter case. Here, for the sake of consistency, we have presented all the results taking ion energy as 50 keV H+ in all the cases and have doubled the dose in those cases where we

Phys. Status Solidi C 7, No. 1 (2010) 45

www.pss-c.com © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Contributed

Article

have used 100 keV H2+ ions. After implantation the sam-

ples were cut into different small pieces and annealed at higher temperatures up to 800 °C for 1 hour in air ambient. In order to check the optically detectable surface blis-ters/exfoliation in the samples, we used Nomarski optical microscopy. The surface morphology of the samples was studied using an AFM (digital Nanoscope III from Veeco in tapping mode). The microstructural characterization of the implantation-induced damage in these samples was performed using cross-sectional transmission electron mi-croscopy (XTEM). The XTEM measurements were carried out by using a Philips CM20T machine operated at 200 kV.

3 Results and discussion In case of GaN, the mini-mum hydrogen fluence required for the occurrence blister-ing to after post-implantation annealing was found to be 2.6×1017 cm−2 [11]. Figure 1 shows the Nomarski optical image of the GaN surface when the GaN layer was im-planted with hydrogen and then annealed at 700 oC for 10 min. The size of the blisters varied between 2-6 µm. The blistering times were determined at various temperatures between 350-700 oC. The Arrhenius plot of the blistering time versus reciprocal temperature revealed two activation energies: 1.79 eV in the temperature range of 350-400 oC and 0.48 eV in the temperature range of 400-700 oC. The lower activation energy is associated with the free atomic diffusion of hydrogen in GaN while the higher activation energy is associated with the free atomic diffusion limited by trapping-detrapping of hydrogen in the implantation-induced defects [11]. This feature of blistering kinetics is similar to what has been previously obtained in case of hy-drogen implanted silicon [9].

Figure 1 Nomarski optical image of GaN surface after post-implantation annealing at 700 oC/10min [11]. The implantation was done with 50 keV H+ ions at a fluence of 2.6×1017 cm−2.

The cross-sectional TEM image of the as-implanted GaN is shown in Fig. 2 wherein a damage band between 200-500 nm from the surface can be seen. Higher magnifi-cation TEM image of the damage band shows that this re-gion is filled with nanovoids having dimensions of a few nanometers. These nanovoids upon high temperature an-nealing lead to the formation of nanocracks and micro-

cracks inside the damage band (Fig. 3) that eventually lead to the formation of surface blisters [12-15].

Figure 2 XTEM image of GaN implanted with 50 keV hydrogen ions at a fluence of 2.6×1017 cm−2. Inset shows high magnifica-tion image of a selected region in the damage band [15].

Figure 3 XTEM image of microcracks formed inside the damage band in GaN implanted with 50 keV H+ ions with a fluence of 2.6×1017 cm−2 and then annealed at 700 oC/1hr [12].

In case of hydrogen implanted AlN layers the mini-mum hydrogen ion fluence to observe blistering was found out to be 2×1017 cm−2 [16]. Figure 4 shows the blistering in hydrogen implanted and annealed AlN layers. The size of the blisters varied between 10-20 µm. From the blistering kinetics investigation of hydrogen implanted AlN, two ac-tivation energies were found: 1.16 eV in the temperature range of 450−550oC and 0.44 eV in the temperature range of 550−750oC. The XTEM image of the as-implanted AlN (Fig. 5) shows a damage band extending between 330−550 nm from the surface [16]. When a high magnification XTEM image (Fig. 6) of the damage was taken, it was found to be decorated with hydrogen filled nanovoids, as in the case of GaN. The number density of these nanovoids increases towards the end region of the damage band. Moreover, these nanovoids showed a tendency to agglom-erate and line-up in the c-plane [17]. After annealing at higher temperatures, these nanovoids merged together

46 R. Singh et al.: Implantation-induced blistering/exfoliation in wide bandgap semiconductors

© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-c.com

ph

ysic

ap s sstat

us

solid

i c

leading to the formation of microcracks and ultimately to the optically observable surface blisters.

Figure 4 Nomarski optical image of 50 keV H+ implanted and annealed (700 oC/5min) AlN layer. The implantation fluence was 2.0×1017 cm−2 [16].

Figure 5 XTEM image of 50 keV H+ implanted AlN layer in the as-implanted state. The implantation fluence was 2.0×1017 cm−2 [16].

Figure 6 High magnification XTEM image of the damage band in H+ implanted AlN layer in the as-implanted state [17].

For ZnO(0001) bulk samples, the minimum hydrogen flu-ence required to observe surface blistering/exfoliation after

post-implantation annealing was 5.0×1017 cm−2 [18]. It was observed that after post-implantation annealing of ZnO, large area exfoliation of the surface was observed. The en-tire top layer was removed and only a few island regions remained intact (Fig. 7). The XTEM image of ZnO in the as-implanted state showed the formation of a damage band extending between 200−600 nm from the surface (Fig. 8). This damage band was filled with nanovoids having size between 3-10 nm. Moreover, the nanovoid could also be observed in the top 200 nm layer of ZnO that was rela-tively much less damaged. Specifically, the nanovoids in the damage band showed a tendency of aligning along the c-plane. In addition, large area microcracks were observed in the damage band in the as-implanted state. The forma-tion of microcracks in the as-implanted state was probably responsible for the large area exfoliation observed after high temperature annealing.

Figure 7 Nomarski optical image of 50 keV H+ implanted and annealed ZnO bulk crystal at 500 oC for (a) 8 min and (b) 15 min. The hydrogen fluence was 5.0×1017 cm−2 [18]

Figure 8 XTEM image of 50 keV H+ implanted ZnO bulk crystal in the as-implanted state with a fluence of 5.0×1017 cm−2 [18].

It is to be noted that the minimum implantation fluence required to observe blistering/exfoliation in GaN, AlN and ZnO after post-implantation annealing is quite higher than that for other semiconductors such as Si, SiC, InP and GaAs (for these semiconductors the minimum fluence is usually about 1.0×1017 cm−2) [10, 19-21]. Specifically for ZnO the minimum fluence is about 5 times higher than that for Si or SiC. One potential reason may be due to the high radiation resistance of GaN, AlN and ZnO. It may also be related to other factors (that are yet to be explored with fur-ther detailed experimentation) such as higher formation energy of H2 molecules [15], diffusivity of hydrogen (spe-

Phys. Status Solidi C 7, No. 1 (2010) 47

www.pss-c.com © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Contributed

Article

cifically for ZnO) [22] and details of the ion-defect interac-tion processes in these semiconductors. Nonetheless, for the technological application of thin film layer splitting of these semiconductors, the aforementioned implantation pa-rameters are very useful.

4 Conclusions A brief overview of the hydrogen im-plantation-induced blistering/exfoliation in wide bandgap semiconductors, viz. GaN, AlN and ZnO has been pre-sented. The implantation was performed with 50 keV H+ ions with fluences in the range of 5×1016 to 6×1017 cm-2

. The minimum fluences required to observe surface blister-ing/exfoliation in GaN, AlN and ZnO were 2.6×1017, 2×1017 and 5×1017 cm-2, respectively. The TEM investiga-tions revealed the formation of a damage band in all the samples that was filled with nanovoids containing H2 gas. These nanovoids served as precursors for the formation of microcracks inside the damage band after high temperature annealing eventually causing blistering/exfoliation. In combination with direct wafer bonding, the implantation parameters used in the present work for these semiconduc-tors could potentially be utilised in the future for thin film layer transfer applications for fabricating inexpensive high structural quality templates.

Acknowledgements We are thankful to Mrs. S. Hopfe from MPI Microstructure Physics, halle, german for XTEM sam-ple preparation. This work has been supported financially by the Max Planck Society (MPS), Germany under the Indo-German cooperation scheme of Max Planck India Partner Group and is co-funded by the Department of Science and Technology (DST), India. The authors also thank the German Federal Ministry of Education and Technology (BMBF) for partial support in the framework of the CrysGaN project. One of the co-authors (U. Dadwal) is thankful to the MPS-DST cooperation scheme for a research fellowship.

References [1] S. Nakamura and G. Fasol, The Blue Laser Diode: GaN

Based Light Emitters and Lasers (Springer, Berlin, 1997). [2] Y. Taniyasu, M. Kasu, and T. Makimoto, Nature 441, 325

(2006). [3] N. Mizuhara, M. Miyananga, S. Fujiwara, H. Nakahata, and

T. Kawase, Phys. Status Solidi C 4, 2244 (2007). [4] Ü. Özgür, Y. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov,

S. Dogan, V. Avrutin, S. J. Cho, and H. Morkoc, J. Appl. Phys. 98, 041301 (2005).

[5] P. Gibbart, Rep. Prog. Phys. 67, 667 (2004). [6] J. Nause and B. Nemeth, Semicond. Sci. Technol. 20, S25

(2005). [7] R. T. Bondokov, K. E. Morgan, G. A. Slack, and L. J.

Schowalter, Mater. Res. Soc. Proc. 0955-I03-08 (2007). [8] U. Gösele and Q.-Y. Tong, Ann. Rev. Mater. Sci. 28, 215

(1998). [9] S. H. Christiansen, R. Singh, and U. Gösele, Proc. IEEE 94,

2060 (2006). [10] M. Bruel, Electron Lett. 31, 1201 (1995). [11] R. Singh, I. Radu, U. Gösele, and S. H. Christiansen, Phys.

Status Solidi C 3, 1754 (2006).

[12] I. Radu, R. Singh, R. Scholz, U. Gösele, S. Christiansen, G. Bruederl, C. Eichler, and V. Haerle, Appl. Phys. Lett. 89, 031912 (2006).

[13] R. Singh, I. Radu, G. Bruederl, C. Eicheler, V. Haerle, U. Gösele, and S. H. Christiansen, Semicond. Sci. Technol. 22, 418 (2007).

[14] R. Singh, R. Scholz, S. H. Christiansen, and U. Goesele, Mater. Res. Soc. Proc. 1068-C01-08 (2008).

[15] O. Moutanabbir, Y. J. Chabal, M. Chicoine, S. Christiansen, R. Krause-Rehberg, F. Schiettekatte, R. Scholz, O. Seitz, S. Senz, F. Süβaut, and U. Gösele, Nucl. Instrum. Methods B 267, 1264 (2009).

[16] R. Singh, R. Scholz, S. H. Christiansen, and U. Gösele, Semicond. Sci. Technol. 23, 045007 (2008).

[17] R. Singh, R. Scholz, S. H. Christiansen, and U. Gösele, Phys. Status Solidi A 205, 2683 (2008).

[18] R. Singh, R. Scholz, U. Gösele, and S. H. Christiansen, Semicond. Sci. Technol. 22, 1200 (2007).

[19] R. Singh, I. Radu, M. Reiche, R. Scholz, D. Webb, U. Gösele, and S. H. Christiansen, Mater. Sci. Eng. B 125, 162 (2005).

[20] I. Radu, I. Szafraniak, R. Scholz, M. Alexe, and U. Gösele, Appl. Phys. Lett. 70, 2413 (2003).

[21] R. Singh, I. Radu, R. Scholz, C. Himcinschi, U. Gösele, and S.H. Christiansen, J. Lumin. 121, 379 (2006).

[22] K. Ip, M. E. Overberg, Y. W. Yeo, D. P. Norton, S. J. Pear-ton, C. E. Stutz, S. O. Kucheyev, C. Jagadish, J. S. Williams, C. Luo, F. Ren, D. C. Look, and J. M. Zavada, Solid State Electron. 47, 2255 (2003).