| Budget Amount *help |
¥6,900,000 (Direct Cost: ¥6,900,000)
Fiscal Year 1996: ¥2,700,000 (Direct Cost: ¥2,700,000)
Fiscal Year 1995: ¥4,200,000 (Direct Cost: ¥4,200,000)
|
|---|
| Research Abstract |
There is much interest in intersubband-transition (ISB-T) in semiconductor quantum wells (QWs). The ISB-T has a large transition probability and a very fast energy relaxation time (about lps). The optical phenomena in the QW structure thus far have been mainly concentrated on either this ISB-T or a conventional interband-transition (IB-T). To investigate the phenomena by the simultaneous utilizetion of the ISB-T and IB-T is physically very interesting and may lead to new device application fields such as ultrafast all-optical devices.
Based on this concept, the IB-resonant light modulation by the ISB-resonant light has been recently proposed and demonstrated, where modulation schemes are two-fold : (i) utilization of n-doped QWs and (ii) utilization of undoped QWs. The former makes use of the actual carrier transition from the first to second subbands in the conduction band, while the latter utilizes the quasi-virtual transition between them. Ultrafast modulation from picosecond to femt
… More
second range can be expected in these schemes.
In this work, we have succeeded in ultrafast IB-resonant light modulation by the ISB-resonant light in the case of former n-doped QW.The sample utilized for the experiment has a multiple quantum well structure with 30 periods of GaAs wells (7.6nm) and Al_<0.3>Ga_<0.7>As barrier layrs (14.1nm) grown by MBE on a GaAs substrate, where the barrier layrs were selectively n-doped with Si (n-3.0x10^<18>cm^<-3>). The theoretical calculation using a density matrix theory predicted that the IB-absorption change of about 3000cm^<-1> can be expected by using an ISB-resonant light intensity of 1MW/cm^2. This value of absorption change is large enough for actual device performance. The QWs were sandwitched by AlGaAs cladding layrs and processed to the ridge waveguide structure to let the IB-resonant light propagate through it.
The IB-resonant light is modulated by ISB-resonant light when it propagates through the waveguide. A free-electron laser (FEL) was utilized for the ISB-resonant light which delivered macropulses with a width of about 15mum at a 10Hz repetition rate. Each macropulse contained a train of 300-400 ultrashort micropulses with a 45ns duration. A pump and probe measurement of the absorption saturation of the ISB-T of the wafer utilized for this study revealed that the micropulse width was as short as -5ps. The FEL pulse was incident to the sample from the substrate side with a Brewster's angle with TM polarization. On the other hand, a cw semiconductor laser with a wavelength of 830nm was used for the IB-resonant light, which was incident to the ridge waveguide of the sample through an objective lens.
The modulated IB-resonant light was detected by a high-speed Si-pin photodetector whose rise and fall times were 400ps, respectively, and was analyzed by a digitizing oscilloscope triggered by the FEL micropulse signal. It was found that the modulation signals were clearly observed as negative spikes, corresponding to each FEL micropulses. These negative spikes indicate the increase of absorption coefficient for the IB-resonant light induced by the ISB excitation in accordance with the theoretical prediction. The fall and rise times of the observed IB signal were about 400ps, respectively, which were limited by the detection system as described above. The observed modulation depth m was found to be 0.8-0.9%, where m was defined as m= (I_1-I_2) /I_1 : I_1 and I_2 are the intensities of the IB-resonant light without and with ISB-resonant light, respectively.
It should be noted that the actual modulation depth should be more than 100 times larger than the observed one, because the response time of the detection system was about 800ps, while the actual time of the modulation would be as fast as -5ps (FEL micropulse width). The rise time observed is much faster than the band-to-band relaxation time (typically, -a few nanosecond), which indicates that the modulation speed is not limited by the hole accumulation in the valence band, and we could not find the speed limitation factors except for our proposed modulation principle. To confirm this, we have also measured the dependence of the modulation depth on the FEL pulse energy. The modoulation depth saturated to be 0.8-0.9% when the effective pulse energy exceeds -5pJ (effective peak power supplied to the QWs was about 300kW/cm^2). The result also supports that the almost full switching was achieved by this experiment. We have also succeeded in shortening of the intersubband trasnsition down to 1.9mm by using InGaAs./AlAs QW system. Less
|
