Relaxed In x .Ga 1-x as Graded Buffers Grown With Organometallic Vapor Phase Epitaxy on GaAs
- PDF / 2,108,082 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 86 Downloads / 188 Views
ter each change in TMI flow setting to ensure the expected composition, during which time the sample was kept at the growth temperature. All samples, except the structure with the x,.=0.33 graded buffer grown at 700 °C, had an undoped 1 gm uniform cap. The sample with an x,=0.33 graded buffer at 700 0C had a 2 gm cap which incorporated a PIN diode structure. RESULTS A visual inspection of the surface morphology reveals a strong dependence on growth temperature, a surprising result for a such a low lattice mismatch. AFM surface topology data taken on 10 pm x 10 pm areas of each sample, including the x•,=0.15 and x1,=0.33 structures is depicted in figure 1. The data show that the rms roughness for the nominal x1.=0.06 sample grown at 550 *C has a significantly greater rms roughness value (52 nm) than the other structures which have an rms roughness value of about 10 rum. In fact, despite the low mismatch, the sample grown at 550 0C is not specular. In addition, the surface roughness for the xl,=0.33 structure grown at 700 *C was less than the surface roughness of x1.=0.06 structures grown at all lower temperatures. Most notably, the x,=0.33 structure grown at 550 *C has a surface roughness 20X that of the structure grown at 700 0C, showing the greater disparity in surface roughness with increasing indium content. A key criterion for applications is that there must be a great amount of strain relief and a low threading dislocation density. To determine the degree of strain relaxation and the indium composition, glancing exit (224) reciprocal space maps were conducted with triple axis XRD. Since the xj,=0.06 structures were of low mismatch and relatively thin (2 pM) the effect of epilayer tilt was expected to be negligible, and thus no glancing incidence (224) or (004) reciprocal space maps were acquired to extract this effect. X-TEM was used to measure the thickness and in combination with the final composition, the grading rate. Table I shows the growth temperature, composition, and grading rate for the nominal x,.=0.06 structures. The table shows that the indium composition steadily increased with decreasing temperature (with the exception of the structure grown at 600 0C), which is due to the lower cracking temperature for TMI. In addition, the growth rate decreased with decreasing temperature, which in turn provided 100
33% In
80
180 nm 8.2 %In
.E 500C C
6039.8%
04o
20
0
550 OCg1%fl I7.%I
9.aI 50" 1 0 1500
%I
6.160( %In 7.8%1In L1l15%ln 33%1n 700 • 700
Figure 1- RMS roughness for InAGalxAs structures grown at different temperatures 632
for a higher grading rate at lower temperature. It should be noted that there is a small error (n 300 A) in the measurement of the graded buffer thickness due to the calibration of the TEM and the tilting of the TEM specimens. Table I shows the residual strain in each of the nominal x1,=0.06 structures as a function temperature. The structure grown at 500 'C had a noticeably greater residual strain left in the structure, and there is no general trend among the other sam
Data Loading...
