Peter Deák, professor of physics Department of Atomic Physics Surface , Physics Laboratory Budapest University of Technology and Economics Budafoki út 8. , Budapest H-1111, Hungary Tel. 463-4207 Fax 463-4357 E-mail: [email protected] Proposal for an FP6 STREP FET based on nanodevice ideas with SiC Information Society Technologies 2003-2004 Workprogramme 2/3 of the budget will be devoted to IPs and NoEs. HOWEVER On average two to three IPs and NoEs are expected to be supported for each Strategic Objectiv. The IST thematic priority will also support Specific Targeted Research Projects (STREPs) several STREPs are also foreseen in most objectives. IST will support research into future visions and emerging technologies (FET). All instruments should have adequate industrial participation. The selection criteria and weights and thresholds for the FET open scheme are different. Scientific and technological excellence, Potential impact,Quality of the consortium The Community support for IST in FP6 will help mobilise the industrial and research community around high-risk long term goals. 184.108.40.206 Micro and nano systems Additional STREPs will be restricted to explore highly promising alternative approaches to prepare new technological fields to explore the application potential of micro-nano technology Nano-devices with SiC Ideas of a theorist: 1. EXPERIMENT Heat treatment of a SiO2/Si system of electronic quality in a CO containing gas: CO diffusion dissociation SiC formation at SiO2/Si interface. SUBSTRATE: Si (100); P- or N-doped (4-6 cm). OXIDE: 100 nm, thermally grown at 1050 °C in dry oxygen, Quality: density of electrically active states in the oxide and at the interface < 5·1010 cm-2 (CV) TREATMENT: furnace anneal in Ar gas flow of 100 cm3/min, containing 5 % CO (gas purity: 99.995%) - for 3, 8 and 20 hours - between 900-1190C Characterization of samples by: SIMS, XPS, TEM, AFM, ESR, CV SiC is present in form of cubic crystallites at the interface: 3 hours 20 hours 45 45 20 nm 90 90 35 nm ratio of average grain volumes: 7 (reaction limited growth) nucleation density, 2.5109 cm-2, independent of time About 90 % of the crystallites are epitaxially oriented: Orientation of the grains: Top view (001)Si || (001)SiC Si || SiC Side view There are no voids at the Si/SiO2 interface! SiC 4 : 5 fit between lattice constants: ESR: dangling bond density ~ 41012 cm-2 (courtesy of M. Brandt, TU-Munich) LIKELY GROWTH MECHANISM CO 1 O C 2 SiO2 SiC O C 3 O C Si 1. 2. 3. (CO) + 2 <SiO2> <SiO2:Ci,Oi> loss of ~ 7 eV/CO (K et al. PRB 2001) 2 (CO) + 2 <SiO2> 2 <SiC> + 3(O2) loss of ~ 8 eV/CO ; V=-63 Å3 4(CO) + 6 <Si> 4 <SiC> + 2<SiO2> gain of ~ 6 eV/CO ; V=+48 Å3 AFM after etching off the SiO2 layer shows etch pits at the Si/SiC interface: Together with lateral growth direction and shape of the cross section, proof for CO dissociation at the Si/SiC interface! GRAINS COALESCENCE WITHOUT GRAIN BOUNDARY IDEA 1.a SiO2 Si (SOI) SiO2 Si Band structure: C V SiC Problems to be solved (before thinking of devices) 1. Lateral growth rate: depends only on [CO] and T - growth volume is linear with time growth rate T =1100 C; pCO 0.10 atm - T (> 1000 C) is critical 0.12 growth rate T =1190 C; pCO 0.05 atm - Rate with 1.00 atm CO2 at 1190 C is the same as with 0.05 atm CO but quality somewhat worse 2. Control of nucleation: presently random - density independent of process parameters: ~ 2.5 ·109 cm-2 - suspicion: depends on defects at interface (irradiation in pattern?) 3. During long anneals the oxide degrades 4. Passivation of dangling bonds: - dangling bond density corresponds to 4:5 lattice parameter ratio - “wet” treatments had no effect IDEA 1.b SiO2 Si Band structure: C V SiC IDEA 1.+ (on the side) Nucleation process for 3C-SiC heteroepitaxy 2. EXPERIMENT ALE of SiC: - monolayer growth of 3C-SiC on Si in ALE proven [Hara et al. Thin Solid Films 225, 240 (1993)] - monocrystalline growth was possible even on Si in a commercial planetary reactor [Sumakeris et al., ibid. p. 219; Nagasawa & Yamaguchi, ibid. p.230] - layer by layer 3C- SiC homoepitaxy in ALE demonstrated [Fuyuki et al., ibid. p. 225] 2. THEORY Calculation of the stability and electronic structure of various extended defects in SiC: alternating layer sequences in SiC. C Si C C Si Si Si C face substrate Si face substrate ALE layers with polarity change ALE layers with polarity change Si C Si Si C C C C Si 3.6 eV 1.8 eV 1.4 eV 0.3 eV ~ 3Å ~ 2Å Based on bulk band off-sets, strong QW effect expected. No stress in the system! Methods: ab initio DFT electronic structure calculations; DFTB-MD - DFTB-MD: if the “homo-double-layers” are formed during ALE, H H they are stable up to temperatures where surface H is released Si C Si C Si C C Si Si C Studies regarding stability during growth (attack of single CH3 on Si-C-Si-Si-H and of SiH3 on C-Si-C-C-H sequence) are under way. - ab initio DFT: to keep periodicity, both type of “homo-doublelayers” have to be built in into the supercell. C C Si C Si Si C Si Si Si C C C Si C Si > 4 eV 2.9 eV ?? ~ 5A Wave function at the VB edge: (LDA problem with CB) 1.0 eV Further possibilities…to be examined by calculations Admittedly: it is a long shot before thinking of devices but IST will support research into future visions (FET) and FP6 will help mobilise around high-risk long term goals. So, while theory is working: - why not study ways to use such systems - why not study problems of contacting, integrating, etc. - why no try to grow it with ALE!?!? Device ideas????? • HEMT ? • Idea on the side: blocking of stacking fault motion in SiC! Possible distribution of tasks Managment FhG-IIS-B nano SiC in Si/SiO2 synthesis MFA + FhG-IIS-B QW within SiC synthesis UE(Ley) Theory BUTE + UPB structural characterization MFA + BUTE+ UE(Ley) electrical/optical.EPR characterization Ue(Pensl) + LiU Device considerations FhG-IIS-B, MFA, LiU/Laussane "interested industry" ???