Chemical vapor deposition of TiSi2 using an industrial integrated cluster tool

Maury, Delphine; Rostoll, M. L.; Gayet, Philippe; Régolini, J. L.

doi:10.1116/1.589238

Cited by 6 publications

(1 citation statement)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…21 Most studies on CVD TiSi 2 have concentrated mainly on understanding the first principles of the process on undoped or moderately doped substrates, or on heavily doped silicon with an undoped silicon epitaxial interlayer. 22,23 Only a few studies exist concerning TiSi 2 deposition on heavily doped substrates. [24][25][26][27] It appears, however, that arsenic does effect deposition and consumption.…”

mentioning

confidence: 99%

Effects of Arsenic Doping on Chemical Vapor Deposition of Titanium Silicide

Fang¹,

Öztürk²,

Seebauer

et al. 1999

J. Electrochem. Soc.

View full text Add to dashboard Cite

Titanium silicide (TiSi 2 ) has long been used in complementary metal oxide silicon (CMOS) integrated circuits because of its low resistivity and compatibility with standard processes. 1 With the scaling of device dimensions into the deep submicron regime, two problems related to the conventional self-aligned silicide (SALICIDE) process have emerged. The first problem is the so-called narrow-line effect which describes the difficulty in conversion of TiSi 2 from the high resistivity C49 phase into the low resistivity C54 phase on narrow polysilicon lines. 2 Several new growth methods have been suggested to solve this problem, including the use of titanium alloys 3 and molybdenum implants. 4 The second problem is the silicon substrate consumption during silicide formation, which makes the SALICIDE process incompatible with ultrashallow junctions. Metal oxide semiconductor field effects transistor (MOSFETs) with elevated source/ drain structures have been proposed to address this problem. 5,6 Selective chemical vapor deposition (CVD) of TiSi 2 has been proposed as an alternative process to form contacts to ultrashallow junctions without silicon consumption. It has also been suggested that C49-C54 conversion occurs more readily for this process although this area still lacks an in-depth investigation. The CVD process was first introduced by Tedrow et al. 7 in 1985, and it has attracted considerable interest from researchers in U.S., Japan, and France. [8][9][10][11][12][13][14][15][16] Silicides are formed on heavily doped source/drain regions and polycrystalline silicon gates. For the conventional SALICIDE process, the effects of dopants on the formation and characteristics of silicides have been extensively studied. For TiSi 2 , several general remarks can be made: (i) TiSi 2 growth rate depends on dopant type and concentration. The growth rate is highest on undoped and borondoped substrates, lower on heavily phosphorous-doped substrates, and lowest on heavily arsenic-doped substrates 17,18 ; (ii) Ti reacts with boron and arsenic to form stable compounds such as TiAs and TiB 2 , which makes TiSi 2 unsuitable as a dopant out-diffusion source to form junctions [18][19][20] ; (iii) TiSi 2 films formed on arsenic-doped substrates are thermally less stable. 21 Most studies on CVD TiSi 2 have concentrated mainly on understanding the first principles of the process on undoped or moderately doped substrates, or on heavily doped silicon with an undoped silicon epitaxial interlayer. 22,23 Only a few studies exist concerning TiSi 2 deposition on heavily doped substrates. [24][25][26][27] It appears, however, that arsenic does effect deposition and consumption. 25 In a recent report, 28 this group presented experimental results on TiSi 2 deposition on heavily boron-doped substrates. It was conclud-ed that implanted boron does not have a strong influence on the deposition process. Shallow p ϩ -n junctions (X j ϭ 1000 Å) with 800 Å thick TiSi 2 contacts were successfully fabricated. It was also shown by a preliminary experiment...

show abstract

mentioning

confidence: 99%