Exfoliation and Blistering of Cd^sub 0.96^Zn^sub 0.04^Te Substrates by Ion Implantation

Journal of Electronic Materials, Jun 2005 by Miclaus, C, Malouf, G, Johnson, S M, Goorsky, M S

As part of a series of wafer bonding experiments, the exfoliation/blistering of ion-implanted Cd^sub 0.96^Zn^sub 0.04^Te substrates was investigated as a function of postimplantation annealing conditions. (211) Cd^sub 0.96^Zn^sub 0.04^Te samples were implanted either with hydrogen (5 × 10^sup 16^ cm^sup -2^; 40-200 keV) or co-implanted with boron (1 × 10^sup 15^ cm^sup -2^; 147 keV) and hydrogen (1-5 × 10^sup 16^ cm^sup -2^; 40 keV) at intended implant temperatures of 253 K or 77 K. Silicon reference samples were simultaneously co-implanted. The change in the implant profile after annealing at low temperatures (

Key words: CdZnTe, ion implantation, blistering

INTRODUCTION

HgCdTe epitaxial structures make use of either lattice-mismatched (211) CdTe/ZnTe/Si composite substrates1,2 or bulk (211) Cd(Zn)Te substrates. Growth on a silicon substrate is beneficial for overall thermal expansion matching to the silicon read-out circuitry that is part of the final device,3-5 but the lower defect density associated with growth on the CdZnTe substrates leads to better performance at longer wavelengths.6 Wafer bonding and layer exfoliation offer a means to produce the combination of high-quality CdZnTe layers on a silicon substrate for subsequent HgCdTe deposition. We have already demonstrated successful bonding of (211) CdZnTe wafers to (001) silicon substrates using SiN dielectric interfacial layers.7 In that case, a thin CdZnTe layer could be produced by etching, grinding, and polishing the CdZnTe substrate. The inclusion of a layer transfer step by hydrogen-induced exfoliation, however, allows for the reuse of the original CdZnTe substrate. Layer transfer by wafer bonding and hydrogen ion implantation8-10 has been extensively addressed for silicon,8,11 SiC,12,13 and SiGe graded alloys14 but only recently with III-V materials,15 in which it was demonstrated that a lower temperature (150°C) postimplant defect nucleation step followed by a higher temperature (~300°C) defect growth exfoliation step promoted layer exfoliation. To our knowledge, there are no published reports16 of CdZnTe blistering/exfoliation. To integrate layer exfoliation into the wafer bonding process, the hydrogen implantation of CdZnTe substrates and subsequent thermal annealing to promote surface blistering is addressed.

It has been demonstrated7,17 that annealing of hydrogen-implanted silicon at sufficiently high temperatures induces blistering of the implanted layer with the exfoliated layer thickness controlled by the implant energy. A minimum implant dose is required18 to facilitate layer blistering and compensate for ion loss through diffusion during annealing. At elevated temperatures, additional defects may form from the agglomeration of initial point defects. Aside from the process annealing steps, these elevated temperatures may also be reached unintentionally due to self-heating that occurs during implantation. The exfoliation initiates at these defects and expands laterally by increasing internal gas pressure in the defect. When the critical stress for crack propagation is exceeded, the defects/cracks will expand laterally. In the case of an unconstrained surface, these will eventually lead to blistering of the surface and the exfoliation of small (micron scale) fragments. However, constraining the top surface of the wafer by bonding it to a handle wafer forces the defect growth to continue in a lateral direction, leading to complete layer transfer.

In this paper, we monitor the changes in the implant profile and blister formation during the exfoliation anneal of ion-implanted CdZnTe bulk substrates. Results from atomic force microscopy (AFM) and x-ray diffraction were used to determine the implant stability at low (

EXPERIMENTAL

(211) A Cd^sub 0.96^Zn^sub 0.04^Te wafers20 were implanted under various conditions, as shown in Table I. During the implant, the CZT samples were attached to a 100-mm-diameter Si wafer using silver paste to allow for improved thermal contact. After implantation, the wafers were cleaned using a series of organic soaks (10 min) in acetone, isopropanol, and methanol followed by annealing at temperatures up to 300°C.

The state of the implant after implantation and during annealing was investigated using double-and triple-axis x-ray diffraction. The measurements were conducted using a Bede D3 diffractometer (Bede Scientific Inc., Englewood, CO).21 Double-axis ω/2θ scans were employed to monitor the anneal-induced strain changes in the implanted layer. Triple-axis rocking curves were also employed at each anneal step to measure the change in the mosaic structure on the lattice, based on the concept that nucleation and growth of defects leads to localized tilting of the lattice planes.10 The surface morphology of the sample surface was monitored during annealing by Nomarski microscopy and AFM.

RESULTS AND DISCUSSION

The (422) x-ray diffraction scattering measurement22 for the hydrogen-implanted CdZnTe (CZT) sample at 253 K is presented in Fig. 1a together with the data following the anneal steps at temperatures up to 500°C. No differences were observed in the diffraction profiles and no optically detectable blisters were formed during annealing. The surface roughness of the sample was preserved (1.7 nm) during the entire annealing sequence. To further investigate the inhibition of blister formation for this sample, we measured the Si carrier using (004) x-ray diffraction measurements and compared the results with a similarly implanted Si wafer from a previous run that was successfully exfoliated (Fig. 1b). There are clear differences in the diffraction profiles. The lack of strain fringes in our current implant data shows a similar behavior as we generally observed in an annealed sample10 that may inhibit or prevent23 blistering during annealing. Given the difference in thermal conductivity between Si and CdZnTe, we speculated that greater heating occurred in the CZT structure and that reduced implant temperatures might be necessary.