c ~ 10-7 
úcm2
for W and c of 4x 10-7
úcm2 for WSix. InN metallized with W
produced ohmic contacts with c ~ 10-7
úcm2 and c~ 10-6
úcm2.for WSix at room temperature.The temperature dependence of the specific contact resistance of W and WSi0.44 contacts on n+ In0.65Ga0.35N and InN was measured in the range -50 °C to 125 °C. The results were compared to theoretical values for different conduction mechanisms, to further elucidate the conduction mechanism in these contact structures. The data indicates the conduction mechanism is field emission for these contact schemes for all but as-deposited metal to InN where thermionic emission appears to be the dominant mechanism. The contacts were found to produce low specific resistance ohmic contacts to InGaN at room temperature,
It has proven difficult to produce low resistance ohmic contacts to the
III-nitride materials because of their wide bandgaps.1-11 To date
little work has been done regarding the conduction mechanism in ohmic contacts
to the nitrides. We would like to establish a high temperature contact
technology for the nitrides, for applications such as electronics capable of
operation at >= 500 °C, or for power switching. Cole et. al.7
reported that W produced contact resistivities of ~10-4
úcm2 on n+ GaN, and was stable for
annealing temperatures up to ~ 1000 °C. In particular the use of lower
bandgap In-containing nitrides should be able to reduce the contact resistance
on GaN, in analogy to the situation for InxGa1-xAs on
GaAs.
In this paper we report the results of W and WSi0.44 contacts
deposited on n+ In0.65Ga0.35N and
n+ InN. Temperature dependent transmission line measurements (TLM)
in the range -50 °C to 125 °C were used to obtain information
about the conduction mechanism in these contact structures. Room temperature
TLM measurements were also measured as a function of annealing temperature, in
order to establish the stability of the contacts. A key feature of using
In-based nitrides is the trade-off between contact resistance and thermal
stability.
The 2000 thick InN and InGaN samples were grown using Metal Organic Molecular Beam Epitaxy (MO-MBE) on semi-insulating, (100) GaAs substrates in an Intevac Gen II system as described previously.12,13 The InN and In0.65Ga0.35N were highly autodoped n-type (~1020 cm-3, and ~ 1019 cm-3 respectively) due to the presence of the native defects endemic to these materials. The samples were rinsed in H2O:NH4OH (20:1) for 1 min just prior to deposition of the metal to remove native oxides. The metal contacts were sputter deposited to a thickness of 1000 and then etched in SF6/Ar in a Plasma Therm reactive ion etch (RIE) system to create TLM patterns.14,15 The nitride samples were subsequently etched in Cl2/CH4/H2/Ar in an Electron Cyclotron Resonance (ECR) etcher to produce the mesas for the TLM patterns.16 The samples were annealed at temperatures from 300 to 900 °C for 15 sec under a nitrogen ambient in a RTA system (AG-410). Temperature dependent TLM measurements were made over the range -50 °C to 125 °C on the as-deposited and 900 °C (InGaN) and 500 °C (InN) annealed samples. These measurements make it possible to determine the dominant conduction mechanism over the barrier, and the results were compared to theoretical values. The error in these measurements was estimated to be ±10 % due mainly to geometrical contact size effects. The widths of the TLM pattern spacings varied slightly due to processing, (maximum of ± 5 %) as determined by SEM measurements, which were taken into account when calculating the contact resistances.
Figure 1 shows the theoretical curves for contacts to InGaN of this doping
level exhibiting thermionic, thermionic field, or field emission as their
dominant conduction mechanisms. The curves are shown only to give the expected
temperature dependence of c and the magnitude of the specific
contact resistance is arbitrary. The theoretical values are calculated
from17
c  exp( b/E00) for
field emission (1)
c
exp[b/E00coth(qE00/kT)] for thermionic
field emission | (2) |
| c exp(qb/kT) for thermionic
emission | (3) |
where
E00=
h/4 [Nd/m* s]1/2 | (4) |
with b being the barrier height, Nd the donor
concentration in the semiconductor, m* the effective mass of electrons in the
material and s the permittivity of the semiconductor. For
field emission qE00/kT >> 1, for thermionic field emission
qE00/kT ~ 1, and for thermionic emission qE00/kT <<
1, with q/kT 
0.026 eV at 300 K. A fixed barrier height (1 eV) was
assumed for calculations of the three conduction mechanisms. As values have not
been definitively established for m* and s for all the
nitride compounds, the best available values for InN were used, (m*=
0.1me and s=
8o).18
Over the temperature range we studied there was little difference between the
slope expected for the theoretical field emission and thermionic field emission
plots (Figure 1). The thermionic field emission does have a slight upward slope
with increasing reciprocal temperature, but it is less than the error found in
the experimental measurements on the samples. By contrast, the thermionic
emission case shows an obvious trend over the temperature range. Temperature
dependent contact resistance values for InGaN contacted with W and
WSixare shown in Fig. 2. The specific contact resistance is very low
(< 10-5 úcm2) for both
metals. There is no clear pattern to the data over this temperature range.
There is however no upward trend that would indicate thermionic emission. For
this material, the value of E00 was estimated to be 0.63 eV based on
doping levels. This gives a value of qE00/kT ~ 77 indicating field
emission conduction is expected to be dominant.
The contact resistance for the as-deposited contact, however, rises with
temperature, characteristic of thermionic emission. This may be a result of
changing doping levels in the InN because of the sputter deposition of the
contact, as is the case for GaAs. In comparing the data in Fig. 2 and 3 it is
seen that contacts to InN are more sensitive to temperature than InGaN. The
specific contact resistance of InN contacted with W as deposited and after a
500 °C anneal was also measured (Fig. 3, bottom). Again the annealed
contact shows a relatively constant contact resistance over the range while the
as-deposited contact shows an upward trend.
The contact resistance for W and WSix on InGaN as a function of
subsequent annealing temperature is shown in Figure 4 (top). Both contacts had
similar contact resistance as-deposited, ~ 2-4x10-7
úcm2. Above 600 °C the WSix showed
signs of degradation, with c ~ 10-5
úcm2 at 900 °C. c for the W
contact sample dropped to ~ 6x10-8 úcm2 at
600 °C and then increased slightly above that temperature. The contact
resistances for ohmic contacts of W and WSix to InN as a function of
annealing temperature are shown in Fig. 4 (bottom). As-deposited samples had
similar contact resistances to InGaN, indicating a similar conduction
mechanism. WSix contacts showed the most degradation at low
annealing temperatures, with the resistance rising a factor of 5 after 300
°C annealing and then remaining constant. The W contacts began to
degrade at 500 °C.
In summary, theoretical calculations based on the doping levels of InGaN and
InN indicate that the dominant conduction mechanism in W-based ohmic contacts
to these materials should be field emission. The experimental data fit curves
for field emission or thermionic field emission for InGaN contacted with
WSix and W. InN samples contacted with both W and WSix
showed similar behavior after annealing at 500 °C, while for
as-deposited the curves fit better to the thermionic emission case. This may
indicate that the deposition of the contact metal lowered the doping levels in
the InN, while annealing returned them to a higher level. W and WSix
were found to produce low resistance ohmic contacts on n+ InGaN and
InN. W contacts proved to be the most stable, and also gave the lowest
resistance to InGaN and InN, c < 10-7
úcm2 after 600 °C anneal, and 1x10-7
úcm2 after 300 °C anneal, respectively.
The work at Sandia is supported by DOE contract DE-AC04-94AL85000. The technical help of J. Escobedo, M.A. Cavaliere, D. Tibbets, G.M. Lopez, A.T. Ongstad, J. Eng and P.G. Glarborg at SNL is appreciated. The work at the UF is supported by DARPA (monitored by AFOSR, G.L. Witt), an AASERT grant through ARO (Dr. J. M. Zavada), and a University Research Initiative grant #N00014-92-J-1895 administered by AFOSR.
2. M.A. Khan, T.N. Kuznia, A.R. Bhattaraia and D.T. Olson, Appl. Phys. Lett. 62 1786 (1993).
3. S. Nakamura, T. Mukai and M. Senoh, Jpn. J. Appl. Phys. 30 L1998 (1991).
4. S.C. Binari, L.B. Rowland, W. Kruppa, G. Kelner, K. Doverspike and D.K Gaskill, Electron. Lett. 30 1248 (1994).
5. S. Nakamura, M. Senoh and T. Mukai, Appl. Phys. Lett. 62 2390 (1993).
6. M.E. Lin, Z. Ma, F.Y. Huang, Z.F. Fan, L.A. Allen and H. Morkoc, Appl. Phys. Lett. 64 1003 (1994).
7. M.W. Cole, D.W. Eckart, T. Monahan, R.L. Pfeffer, W.Y. Han, F. Ren, C.Yuan, R.A. Stall, S.J. Pearton, Y. Li and Y. Lu, J. Appl. Phys. 80 278 (1996).
8. M.E. Lin, F.Y. Huang and H. Morkoc, Appl. Phys. Lett. 64 2557 (1994).
9. F. Ren, C.R. Abernathy, S.N.G. Chu, J.R. Lothian and S.J. Pearton, Appl. Phys. Lett. 66 1503 (1995).
10. F. Ren, C.R. Abernathy, S.J. Pearton and P.W. Wisk, Appl. Phys. Lett. 64 1508 (1994).
11. F. Ren, in GaN and Related Materials, ed. S.J. Pearton (Gordon and Breach, NY 1996).
12. C.R. Abernathy, J. Vac. Sci. Technol. A 11 869 (1993).
13 C.R. Abernathy, Mat. Sci. Eng. Rep. 14, 203 (1995).
14. R.J. Shul, D.J. Rieger, A.G. Baca, C. Constantine and C. Barratt, Electron. Lett. 30 85 (1994).
15. R.J. Shul, M.E. Sherwin, A.G. Baca and D.J. Rieger, Electron. Lett. 32 70 (1996).
16. R.J. Shul, S.P. Kilcoyne, M. Hagerott-Crawford, J.E. Parmeter, C.B. Vartuli, C.R. Abernathy and S.J. Pearton, Appl. Phys. Lett. 66 1761 (1995).
17. A.Y.C. Yu, Solid State Electron, 13, 239 (1970).
18. L.L. Smith and R.F. Davis, in Properties of Group III Nitrides, ed. J.H. Edgar, EMIS Datareview (INSPEC, London 1994).
© 1997 The Materials Research Society
