Improved Nickel Silicide Ohmic Contacts to N-Type 4H and 6H-SiC Using Nichrome
- PDF / 947,702 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 92 Downloads / 213 Views
0.012
0.008
0.004 E eS
0.000
-0.004-
900 -C,
/
1000 OC/
-0.008oos
-i
-0.012
, -10
/
OC
/1100
IO°, , -8 -6 -4 -2 0 2 4 V [Volts]
6
- ,I 8
I 10
FIG. 1. Current as a function of voltage between two NiCr ohmic contacts on n-type 4H-SiC (4.8E17 cm"3) following three minute anneals at temperatures between 800 and 1100 C. Anneals were performed using a graphite strip heater, and the anneal temperature was measured using an optical pyrometer positioned to view a 5 nun x 5 mm area of the strip heater immedialely adjacent to the 5 mm x 5 mm SiC samples. Each sample was fabricated with two TLM patterns, each pattern having ten 200 x 200 gtm 2 ohmic contacts separated by gaps of constant width and variable lengths which increase from 4 grm to 20 gim (see Figure 2).
(A)1 gtm
_
5
_(B)5
tm
FIG. 2. SEM micrographs of the SiC surface between adjacent ohmic contacts in TLM patterns for (A) sample 187 and (B) sample 155 following 3 min anneals at 1IOOC. Four contacts in each TLM pattern on sample 155 had the structure shown in Figure 2. None of
the other NiCr samples had these features; however, we have observed similar structure in TLM patterns with ohmic contacts formed by high temperature annealing following sputter deposition
of Ni2Si and AMTi on n- and p-type SiC, respectively. We suspect that the white areas in the
120
ohmic contact near the edge of the gap are associated with carbon segregation during the anneal cycle; however, our attempts to confirm this supposition using Auger spectroscopy have not been successful. We have not been able to isolate the white areas from the rest of the contact with our AES electron beam. The results of specific contact resistance measurements for several 4H and 6H-SiC samples are shown in Table I. Different doping concertrations refer to different wafers, and repeated doping concentrations indicate different 5 mm x 5 mm samples from a given wafer. The sheet resistance values listed in Table I were measured using Van der Pauw patterns fabricated on the samples with the TLM patterns. In all cases, sheet resistances determined using the Van der Pauw patterns agreed within ±10% with those determined from the TLM calculations. Table I. Summary of Specfic Contact Resistance Meaurements
Pc (Q-cm2 )
Sample No.
ND (cm 3 )
RSHEET (f0/sq)
155 (6H)
3.2E17
3708
209 (6H)
3.2E17
3478
2.5E-6 4.2E-6
211 (6H)
1.4E18
1975
9.1E-5 6.5E-5
187 (4H)
4.8E17
1923
188 (4H)
4.8E17 I _1.5E-4 1.3E19
1870
1.6E-5 1.4E-5 1.OE-4
97
1.2E-5
185 (4H)
1.2E-6 (4 gaps) 1.3E-5 (4 gaps)
8.9E-6 (8 gaps) 9.5E-6 (8 gaps)
1.6E-5
I_
Variations in the measured contact resistances for sample 155 result from the number of pairs of contacts used to determine Pc. For the remaining samples, the sheet resistances are generally in good agreement and decrease with increasing doping concentration as expected. However, the specific contact resistances do not depend strongly on the doping concentrations. For research grade epi layers using Hg probe profiling, we have measured doping concent
Data Loading...
