Method for Generation of Tunable Far Infrared Radiation From Two-Dimensional Plasmons

The invention relates to generation of far infrared (FIR) radiation from two-dimensional (2D) plasmons, and more particularly to a method and apparatus for optically tuning the FIR wavelength.

Plasmons in a two-dimensional system were first reported by C. C. Grimes and G. Adams, in Phys, Rev. Lett. 36, 145 (1976) on the surface of liquid helium, and later by S. J. Allen, Jr., D. C. Tsui, and R. A. Logan in Phys. Rev. Lett. 38, 980 (1977) in the inversion layer of electrons at the interface of a metal-oxide-semiconductor field-effect transistor (MOSFET) shown in FIG. 1, and more recently by N. Okisu and T. Kobayashi, in Electron. Lett. 22, 877 (1986) at the heteroinferface of a A1GaAs/GaAs heterostructure. The plasmons are coupled to the radiation field by a conductive metal grating structure superimposed on a second semitransparent metal film, with the grating having a periodicity which is typically several micrometers as shown in a typical example illustrated in FIG. 1.

The grating appears as a short circuit for radiation polarized parallel to the plane of the grating, thereby reflecting nearly all of the radiation. For radiation perpendicular to the plane of the grating, there is modulation of the far-infrared field with the periodicity of the grating, resulting in the coupling of the plasma oscillations of the inversion layer electrons with wave vectors q=n2.pi./.LAMBDA., where n=1,2 . . . , to the radiation field. By passing current, I, through the device, far infrared radiation is produced at the wave vector q=2.pi./.LAMBDA.. Coupling of the plasmons and the radiation field at the resonance peak take place at large inversion-layer electron densities. Consequently, tuning the infrared radiation is possible only by varying the periodicity of the grating, which is not possible in the prior art as that is fixed at the time of fabrication, and by varying the two-dimensional electron concentration, which requires deposition of a gate electrode below the grating. Due to pinch-off effects, this gate limits the power density level that can be generated to low levels (.ltorsim.10.sup.-7 W/cm.sup.2).

In accordance with the present invention, tunable far-infrared radiation from two-dimensional plasmons is produced from a semiconductor device which provides large inversion-layer electron densities without the need for a metallic grating on a planar surface parallel to the inversion layer. Instead, two coherent beams are directed to the planar surface of the device at an angle to each other, and at a wavelength selected for the light to be strongly absorbed, generating a photoconductivity grating through the inversion layer. Maximum and minimum photogenerated carrier locations correspond to the respective maxima and minima of the interfering light beams. This photogenerated conductivity grating thus produced by the interfering beams couples the two-dimensional plasmons to the far infrared radiation beam. The wavelength of the far infrared radiation can then be readily tuned by varying the angle between the beams, or varying the wavelength of the two interfering coherent light beams.

The novel features that are considered characteristic of this invention are set forth with particularity in the appended claims. The invention will best be understood from the following description when read in connection with the accompanying drawings.

FIG. 1 illustrates schematically a prior-art device for infrared radiation from two-dimensional plasmons using a metal grating on a semitransparent metal film over p-doped silicon.

FIG. 2 illustrates a prior-art device for infrared radiation from two-dimensional plasmons using a metal grating over an A1GaAs heterostructure.

FIG. 3 is an energy flow diagram of the physical process in generating FIR radiation from 2D plasmons.

FIG. 4 illustrates an A1GaAs heterostructure for infrared radiation from two-dimensional plasmons without using a metal grating, and instead using a photoconductivity grating in accordance withe the present invention.

As noted above, two-dimensional (2D) plasmons have been utilized for generating far-infrared (FIR) radiation. Both silicon MOSFET-type inversion layers and A1GaAs heterostructure devices have been utilized as shown in FIGS. 1 and 2, respectively, to establish the necessary hot two-dimensional electron population. FIG. 3 illustrates the physical processes involved, starting with an electric field to generate the hot 2D electrons. Those electrons in turn produce the plasmons that are then coupled out as FIR radiation. Some energy is lost as heat from the hot 2D electrons. In the diagram, the arrows indicate the direction of energy flow, from which it should be appreciated that the physical process is reversible. Starting with input radiation, the output is then a detectable electric field output. In that case, the device becomes an infrared radiation detector, and in either case wavelength tunability would be desirable.

The dispersion relation of the 2D plasmon is: ##EQU1## where: k and f are the plasmon wave vector and frequency, respectively; q and m* are the electron charge and effective mass, respectively; n.sub.s is the electron surface density; and d is the distance between the 2D electrons and the metal grating structure that couples the plasmons to the FIR radiation; g=coth(kd) in silicon MOSFET type devices and g=tanh(kd) in A1GaAs heterostructures. .epsilon..sub.1 and .epsilon..sub.2 and the dielectric constants of the A1GaAs and vacuum (in A1GaAs heterostructure devices), and oxide and metal (in MOSFET devices). The grating periodicity .LAMBDA. determines the plasmon wave vector ##EQU2## and, thus, for a given device, the FIR frequency, f, is determined by the electron surface density, n.sub.s. In silicon MOSFETS, a gate (thin film metallization) is required below the grating as shown in FIG. 1 in order to create the inversion layer (FET channel), which is where the 2D electron population is generated. By varying the gate voltage, the electron surface density n.sub.s is varied, and tunability is easily achieved. However, the phenomenon of channel pinch-off (common to all FET devices), limits the electric field that can be usefully applied between the drain and source, which limits the power density levels that can be generated by such devices to .ltorsim.10.sup.-7 W/cm.sup.2.

In an A1GaAs heterostructure-based device, as shown in FIG. 2, the 2D electrons can be generated at the heterointerface without the need for a metal gate and, thus, much higher power density levels (.gtorsim.10.sup.-5 W/cm.sup.2)have been demonstrated, and even higher levels (.about.10.sup.-2 W/cm.sup.2) are expected. See Okisu cited above. However, elimination of the metal gate makes it impossible to modulate the electron surface density n.sub.s and, thus, to achieve wavelength tunability.

The present invention involves a modified A1GaAs heterostructure-based device as shown in FIG. 4. It does not have a metal grating on top. Instead, two tunable coherent laser beams derived from a laser 10 and beam splitter 12, and adjustable mirrors 14 and 16, interfere on top of the device. The wavelength .lambda..sub.laser of the interfering laser beams is sufficiently short so the laser light is strongly absorbed, generating electron-hole pairs. Maximum and minimum photogenerated carrier density locations correspond to maxima and minima, respectively, of the interfering light beams which produce a photogenerated conductivity grating between etched isolation channels 20 and 21 between the drain 22 and source 23. This photogenerated conductivity grating will couple the 2D plasmons to the FIR radiation. (Coupling requires conductivity modulation provided by the grating.) Since the photogenerated grating periodicity .LAMBDA. depends on the angle .theta. between the optical beams and the laser wavelength .lambda..sub.laser, ##EQU3## the plasmon wave vector Eq. (2) and, thus, the FIR frequency Eq. (1) can be tuned by varying the angle .theta. between the beams, or varying the wavelength .lambda..sub.laser. Techniques for effecting each of the alternate variations for FIR tuning are well known.

Assuming that each period of the grating is composed of a region of width t with low-sheet conductance Y.sub.1 and an adjacent region of width (.LAMBDA.-t) with high-sheet conductance Y.sub.2 (for gratings formed optically, t.about..LAMBDA./2), the coupling efficiency of the grating is proportional to ##EQU4## From Eq. (4) it is clear that to increase the coupling efficiency, it is necessary for Y.sub.2 >Y.sub.1.

Assuming the photogenerated conductivity grating is produced with a 10 mW He-Ne laser (.lambda..sub.laser =6238 .ANG.) capable of generating a peak power density of 200 mW/cm.sup.2 (laser light) on a 2.times.2 mm.sup.2 device (which will result in absorption of .about.6.multidot.10.sup.17 ph/cm.sup.2 /sec), the absorption coefficient is 2.multidot.10.sup.4 cm.sup.-1 at this wavelength, corresponding to an absorption length of 0.5 .mu.m. Since the decay length of the FIR radiation in the semiconductor material is much larger than the absorption length of the light, the FIR field will interact with the entire photogenerated conductivity grating. Absorption of .about.6.multidot.10.sup.17 ph/cm.sup.2 /sec over a length of 0.5 .mu.m corresponds to a generation rate G of .about.1.2.multidot.10.sup.22 carriers/cm.sup.3 /sec. The carrier concentration N is

where .tau. is the carrier lifetime. For .tau.=5.multidot.10.sup.-9 sec, the carrier concentration N is 6.multidot.10.sup.13 cm.sup.-3. This concentration is comparable to the background carrier concentration in the undoped A1GaAs region, so the coupling coefficient factor in Eq. (4) can be expected to exceed 0.1. Higher coupling coefficients can be achieved with higher power lasers.

Some degradation is expected due to diffusion and drift effects. If diffusion effects are taken into account, the grating concentration modulation, N, will be reduced from N=G.tau. Eq. (5) to: ##EQU5## where L.sub.D is the diffusion length of the carriers (typically 2 or 3 .mu.m) and .LAMBDA. is the grating periodicity which is typically of the order of several micrometers for these devices. Another measure that may be taken to reduce diffusion effects is to reduce the carrier mobility constant (and hence the diffusion constant) by growing the top layer with aluminum concentration of x.apprxeq.0.43. Since the grating serves only to couple the plasmons out, and does not seem to play a major role in their actual excitation, such an aluminum concentration may be the optimum choice.

Heat dissipation from the light absorption is negligible. For a 100 .mu.m thick GaAs substrate with a thermal conductivity of 0.5 W.cm-K (and even higher at reduced temperatures), absorption of 10 mW" will cause heating of only 2K.

Although particular embodiments of the invention have been described and illustrated herein, it is recognized that modifications and variations may readily occur to those skilled in the art, such as other hetero- or homostructures used with a photoconductive grating. Consequently, it is intended that the claims be interpreted to cover such modifications and variations.