ECS Meeting Abstracts, 31(MA2018-02), p. 1065-1065, 2018
DOI: 10.1149/ma2018-02/31/1065
The Electrochemical Society, ECS Transactions, 7(86), p. 357-372
Full text: Download
We investigated the effects of Sn, Si and C7 implantation into both P-well and N-well doped regions of Ge-epi on Si wafers after RTA and laser annealing, on crystal quality using XRD analysis and mobility using ALPro profiler of Active Layer Parametrics , which provides mobility and carrier concentration profiles at sub-nm resolution. SIMS analysis was also used for chemical depth analysis. XRD analysis shows improved Ge-epi crystallinity after >1.5J/cm2 laser melt annealing (liquid phase epitaxial regrowth). Since µ=1/ρqN where µ=mobility (cm2/volt-sec), ρ=resistivity (ohm-cm), q=charge of an electron and N=dopant concentration (cm3), if we assume qN=K a constant for these low dopant well regions then we can estimate change in mobility based on change in P-well or N-well resistance then compare to mobility measured by ALPro profiler. For the P-well cases doped with B=5keV at 1E13/cm2 dose for a retrograde P-well ~1E18/cm3 the Ge-epi Rs decreased to 1,600Ω/□ compared to Si at 3,400Ω/□ suggesting a hole mobility increase of 2.1x from 150cm2/V-s to 315cm2/V-s based on literature with actual measured Hall mobility of 465.2cm2/V-s, an increase of 3.1x. The 1.8J/cm2 laser melt anneal boosted P-well Hall mobility by 22% to 567cm2/V-s to a depth of 80nm and adding Sn implant improved mobility 91% to 889cm2/V-s and at a depth beyond 80nm mobility increases to 3,000cm2/V-sec. Si implant improved mobility by 43% to 667cm2/V-s to a depth of >100nm. For the N-well cases doped with P=40keV at 1E13/cm2 dose for a retrograde N-well ~1E18/cm3 the Ge-epi Rs increased to 2,000Ω/□ compared to Si at 1,600Ω/□ suggesting an electron mobility decrease of 0.8x to 216cm2/V-s rather than the expected increase of 3.3x from 270cm2/V-s for Si to 900cm2/V-s for Ge reported in the literature and expected N-well Rs of 485Ω/□. The actual Ge-epi N-well Hall mobility measured was 449cm2/V-s so only a 1.7x increase and nearly identical to the P-well mobility of 465cm2/V-s but the mobility drops off beyond a depth of 40nm. The 1.8J/cm2 laser melt anneal reduces N-well mobility to 191cm2/V-s but is uniform to a depth of 70nm. Adding Si implant had a slight decrease in mobility to 183cm2/V-s while Sn implant increases mobility to 280cm2/V-s uniformly to a depth >100nm. C7 (cluster-C) implant reduced N-well surface mobility to 50.3cm2/V-s to a depth of 30nm but then mobility increases 10x up to ~500cm2/V-s to a depth of 80nm and then drops to ~100cm2/V-s in the non-carbon N-well region as revealed by C-SIMS analysis to a depth of 110nm, therefore measuring mobility versus depth can be very critical in engineering surface and bulk mobility improvements based on chemical compositional changes in the Ge-epilayer. B-SIMS analysis shows retrograde P-well profile with 2E18/cm3 Rp at 50nm and Xj at 150nm also the Ge/Si epilayer interface. The laser melt flattens the B P-well profile to ~1E18/cm3 with a B pile-up at ~2E18/cm3 at the Ge/Si interface. C-SIMS analysis revealed only slight to no C diffusion even in the liquid Ge-melt phase while we observe 2-5% Si intermixing in the Ge-melt but no B nor P diffusion when C was present. Additional results for RTA anneals will be presented.