Contactless Electroreflectance Study of In x Ga 1-x As/InP Multiple Quantum Well Structures Including the Observation of
- PDF / 427,019 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 43 Downloads / 199 Views
consisted of two excitonic transitions (le-lhh and le-llh), corresponding to the fundamental conduction to heavy (h)- and light (1)- hole transitions, respectively, in the MQW portion and a complicated FKO pattern originating in the InP regions. Comparison between the experimental energies of le-lhh/le-llh and a theoretical envelope function calculation (including the effect of strain) [10,11] made it possible to evaluate the conduction band offset parameters Q, = 0.34+0.03 and 0.57+0.03 for x = 0.53 and 0.75, respectively. The InP related FKO beat patterns were analyzed by a fast Fourier transform (FFT) method [12,13]. It was found that the FKO spectra were due to the simultaneous contribution of at least three different fields (106 kV/cm, 36 kV/cm, and 23 kV/cm), which originate in the various interfaces, i.e., substrate/buffer, cap layer/surface, and buffer/QW structure. Identification of the different fields has been accomplished by comparison of the Fourier-transformed spectra before and after sulfur passivation of the structure surface [14]. EXPERIMENTAL DETAILS Two InxGalxAs/InP QW samples were studied in this paper: sample A contained three latticematched QWs (x = 0.53) while in sample B the three QWs were compressively strained (x = 0.75). The two materials were grown by gas-source molecular beam epitaxy (GSMBE) on (001) semi-insulating InP:Fe (001) substrates. Each sample was composed of three identical InxGalixAs QWs alternating between undoped InP barriers. The nominal well (Lw) and barrier widths are 30A and 300A, respectively, while the thicknesses of nominally undoped InP cap and buffer layers were 1500 A and 2000 A, respectively. The background doping level in the InP layers was n = 2x10 5 cm-'. The CER measurements were performed using a condenser-like system at room temperature. An ac modulating voltage 2 (= 1 kV peak-to-peak) at a frequency of 200 Hz was employed. The value of the modulating .. Expt. voltage was sufficiently small so that all the (a) measured line shapes were independent of the modulating voltage. The experimental details have been described previously [6,7]. The sulfur passivation procedure consisted of an ammonium 0 sulfide ((NH 4 ) 2 Sx) solution treatment of a previously etched sample, followed by drying in N 2 atmosphere in a glove box [14]. EXPERIMENTAL RESULTS (b)(b
Displayed by the solid lines in Fig. l(a) and 1(b) are the experimental CER spectra originating from the QW portion of samples A and B, respectively. The strong, lowest lying features are the fundamental intersubband electron to
heavy-hole excitonic transition (le-lhh), while
o
-5 0.7
0.8
0.9
1.0
Photon energy (WV) the weaker high lying features are fundamental intersubband electron to light-hole excitonic transition (le-llh). The dashed lines in these Fig. 1 Experimental CER spectra (solid figures are least-square fits to the first-derivative lines) from the MQWs of samples (a) A and of a Lorentzian profile, which is appropriate for (b) B. The dotted lines are fits to a lineshape function. 482
excito
Data Loading...
