Investigation of Passivation Effects in InP HEMT Layers
- PDF / 430,513 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 22 Downloads / 190 Views
271 Mat. Res. Soc. Symp. Proc. Vol. 406 01996 Materials Research Society
EXPERIMENT Two different layer structures were grown by molecular beam epitaxy (MBE) on semiinsulating InP substrates. Structure A consists of the following sequence: a 250 nm undoped In.52A1.48As buffer layer, a 20 nm undoped In.53Ga.47As quantum well (QW), a 6 nm In.52A1.48As spacer layer, a Si sheet doping layer (5 x 1012 cm- 2 ), a 20 nm undoped In .52A1.48As Schottky contact layer and a 7 nm highly doped In.53Ga.47As ohmic contact layer (6 x 1018 cm- 3 ). Structure B consists of the 250 nm InAlAs buffer layer, immediately followed by the 7 nm highly doped InGaAs ohmic contact layer. For the electrical measurements a Hall bar geometry was defined by optical lithography and wet etching. Standard AuGe/Ni ohmic contacts were used. The samples were standard cleaned in aceton and isopropyl alcohol and dried in a N2 ambient at 150 °C, before loading them in the plasma chamber. Passivating PECVD SiN layers were deposited on the patterned samples at a 2
temperature of 250 'C with a deposition rate of 8 nm/min and a rf power density of 45 mW/cm .
The gas mixture consisted of NH3, N2 and SiH4. The nominal SiN layer thickness varied between 1 and 200 nm. Some samples have been loaded in the heated and pumped-down chamber without plasma exposure (indicated further as "heated") and on other samples the nitride was removed by wet etching in buffered HF (indicated as "passivated + BHF etch") before the measurements. The Hall and magnetoresistance measurements were done in an Oxford SM4-7 4 He-cryostat that covers the temperature range from 1.2 to 300 K and magnetic fields up to 7 T. The Raman spectra were taken using the 488 nm line of an Ar+ laser . The scattered light was analysed by a spex 1403 monochromator and detected by an optical multichannel analyser (EGG 1462). The spectral resolution was I cm- 1 . To decrease the thermal broadening of the peaks, the spectra were recorded at 90 K. To avoid the contribution of photoinduced carriers we used unpatterned passivated samples and a low power density (60 W/cm 2 ). ELECTRICAL MEASUREMENTS The as-grown structure A has a room temperature (RT) sheet-resistance Rsh of 215 D/sq, a free electron density ns of 2.4 x 1012 cm- 2 and a mobility gXof 12000 cm 2 N/- s. The as-grown structure B is very highly resistive (Rsh > 106 Q/sq), due to the depletion of the surface states carriers a high density Fermi level at that the pin the by position of the of
growth time (min) 0
1 SI
3 I
2 I
4 I
5
220
,.,200
180-
160I
0
I
I
I
I
40 30 20 10 nitride thickness (nm)
I
50
Fig. 1: Sheet resistance versus nitride thickness at RTfor the HEMT layer (structureA).
surface [6]. After passivation both the HEMT layer (structure A) and the InGaAs ohmic cap layer (structure B) show a remarkable gradual decrease of Rsh with increasing SiN thicknesses. Fig. I shows Rsh versus the nitride layer thickness for structure A at RT. Saturation is observed for nitride thicknesses larger than 50 nm. The behaviour of struct
Data Loading...
