The mobility of a relatively narrow bandgap semiconductor material can be significantly enhanced by incorporating it into a multilayered structure (10) comprising a first plurality of relatively narrow bandgap layers (12) of the material and a second plurality of wider bandgap semiconductor layers (14) interleaved with and contiguous with the first plurality. The wide bandgap and narrow bandgap layers are substantially lattice-matched to one another, and the wide bandgap layers are doped such that the impurity concentration-thickness product therein is greater than the same product in the narrow bandgap layers. The fabrication of the structure by MBE to enhance the mobility of GaAs is specifically described. In this case, the narrow bandgap layers (12) comprise GaAs and are unintentionally doped to about 1014 /cm3, whereas the wide bandgap layers (14) comprise AlGaAs doped n-type to about 1016 to 1018 /cm3. The incorporation of this structure in an FET is also described.
|
1. A method of enhancing the mobility of a narrow bandgap semiconductor material comprising the steps of:
(a) forming said narrow bandgap material is a first plurality of spaced apart, narrow bandgap semiconductor layers, (b) forming a second plurality of wide bandgap semiconductor layers interleaved with and contiguous with said first plurality, and (c) forming said wide bandgap layers from a material which (i) is substantially lattice-matched to that of said narrow bandgap layers (ii) forms a conductor or valence band step at the interfaces with said narrow bandgap layers of sufficient magnitude to confine carriers, and (iii) is doped such that the impurity-concentration-thickness product thereof exceeds that of said narrow bandgap layers.
7. A method of fabricating a field effect transistor comprising the steps of:
(a) forming a narrow bandgap semiconductor first layer which includes the channel of said transistor, (b) forming a wide bandgap semiconductor second layer on one side of said first layer, (c) forming a wide bandgap semiconductor third layer on the other side of said first layer, (d) forming said layers from materials which (i) form a conduction or valence band step at the interfaces with said first layer of sufficient magnitude to confine carriers, and (ii) are doped such that the impurity-concentration thickness product of at least one of said second and third layers exceeds that of said first layer, and (e) forming electrode means on said transistor including source and drain electrode means which are electrically coupled to said channel and gate electrode means for controlling the flow of carriers in said channel.
6. A method of enhancing the mobility of a narrow bandgap semiconductor material comprising the steps of:
(a) forming said narrow bandgap material as a narrow bandgap first layer, (b) forming an essentially undoped wide bandgap semiconductor second layer contiguous with one side of said first layer, (c) forming a wide bandgap semiconductor third layer contiguous with said second layer, (d) forming an essentially undoped wide bandgap semiconductor fourth layer contiguous with the other side of said first layer, (e) forming a wide bandgap semiconductor fifth layer contiguous with said fourth layer, (f) forming said layers from materials which (i) form a conduction or valence band step at the interfaces with said first layer of sufficient magnitude to confine carriers, and (ii) are doped such that the impurity-concentration thickness product at least one of of said third layer and said fifth layer exceeds that of said first layer, and (g) forming electrode means electrically coupled to said layers and capable of causing said carriers to flow in said first layer in a direction essentially parallel to said interfaces.
2. The method of
3. The method of
4. The method of
5. The method of
8. The method of
|
22 26 are depleted from the wide bandgap layers 14 and accumulate in the potential wells formed by the narrow bandgap layers 12. The flow of electrons in the conduction band, which is depicted by the arrows 30 in FIG. 2, results in band bending as shown in FIG. 3. That is, the depletion of electrons from the wide bandgap layers 14 causes a downward bending in the conduction band 32 and valence band 34 of these layers. In contrast, the accumulation of electrons in the narrow bandgap layers 12 produces a corresponding upward bending in the conduction band 36 and the valence band 38. The accumulation of electrons in narrow bandgap layer 12 fills the energy states (e.g., states at 40) below the Fermi level EF.
In essence, therefore, the modulated conduction band of the structure 10 results in a depletion of electrons from the wide bandgap layers and accumulation of those electrons in the narrow bandgap layers. The modulated doping of the structure 10, on the other hand, insures that the electrons which accumulate in the narrow bandgap layers exhibit greatly reduced ionized impurity scattering. As a result, the structure 10 as a whole exhibits significantly enhanced mobility. As mentioned previously, however, the heterojunction barriers are not infinitely high (in energy) so that there is a finite quantum-mechanical probability for electrons to penetrate a few Angstroms into the wide bandgap layers. Such electrons could therefore experience ionized impurity scattering by donors near the heterojunctions. To reduce the likelihood of such scattering, and further enhance mobility, it is a feature of another embodiment of our invention that, in each wide bandgap layer 14, doping is terminated short of the heterojunctions so as to leave thin buffer zones 14.1 substantially free of ionized impurities. Thus, only the central portion 14.2 of each wide bandgap layer is doped. This feature also reduces the likelihood that impurities in the wide bandgap layers 14 will diffuse into the narrow bandgap layers 12 where they increase scattering. Thus, it is also preferable to dope the wide bandgap layers 14 with slow diffusing impurities (e.g., Si for AlGaAs), and to grow the layers by MBE since it enables abrupt changes in doping and utilizes low growth temperatures.
We have fabricated several structures of the type shown in FIG. 1 utilizing molecular beam epitaxy to grow n--Alx Ga1-x As (x=0.2 to 0.35) wide bandgap layers 14 and unintentionally doped GaAs layers 12. Due to background contamination, the GaAs layers would tend to have an impurity concentration of about 1014 /cm3. The AlGaAs layers, on the other hand, were doped with Si in different structures to levels ranging between about 1016 to 1018 /cm3. Doping in some structures was terminated 10-60 Angstroms short of the heterojunctions. Also in different structures, the thickness of the layers ranged from 100 to 400 Angstroms, but in each the thicknesses of wide and narrow bandgap layers were equal and the impurity concentration-thickness ratios ranged from 102 :1 to 104 :1. The results which follow, however, were found to be substantially independent of layer thickness of this range.
Some of the results are plotted in FIG. 4 and compared with GaAs grown by other techniques. The dashed line, which results from the Brooks-Herring theory, predicts the maximum mobility for n-type GaAs at room temperature over the doping range from 1014 to 1019 /cm3. In the prior art, A. Y. Cho has grown n-type, Si-doped GaAs epitaxial layers by MBE on n-type GaAs substrates and has measured the mobility which is shown by the black dots. W. Wiegmann has grown similar layers depicted by the cross data point. Similar results have been obtained for n-type, Te-doped GaAs grown by liquid phase epitaxy. The significance of this prior art data is that in all cases the mobilities fell below the theoretical maximum predicted by the Brooks-Herring theory.
For comparison purposes, we fabricated a structure analogous to that shown in FIG. 1 using GaAs and Alx Ga1-x As (x=0.27) with uniform doping throughout; that is, the GaAs and AlGaAs layers were all doped with Si to substantially the same level (about 1018 /cm3). The mobility of this structure also fell below the theoretical maximum as shown by the square data point on FIG. 4. In contrast, the mobilities for GaAs-AlGaAs multilayered structures with modulated band-gap and doping as prescribed by our invention all exhibited mobilities above the theoretical maximum as shown by the open circle data points of FIG. 4. It should be noted that although these data points appear to be close to those of the prior art, they are considerably higher because the ordinate scale is logarithmic.
While FIG. 4 shows how mobility varies with doping level at a given temperature, it is also important to know how mobility varies with temperature at a given doping level. Temperature was controlled by cooling means 15 (e.g., cryogenic apparatus) shown schematically in FIG. 1 as surrounding device 10. Thus, FIG. 5, curve I, shows the mobility-temperature variation for a multi-layered structure of GaAs-Alx Ga1-x As (x=0.30) doped with Si uniformly throughout to a level of about 1017 /cm3. At room temperature, the mobility of the uniformly doped structure was about, 2,500 cm2 V-1 sec-1 and decreased drastically with temperature to about 100 cm2 V-1 sec-1 at liquid helium temperatures. Another uniformly doped (about 6×1017 /cm3) multilayered structure had somewhat higher mobilities at 77 K and 300 K. As with bulk GaAs samples, in the cryogenic temperature range the mobility of the 1017 /cm3 uniformly doped sample decreased as T3/2 which is characteristic of ionized impurity scattering. However, a similar structure (x=0.26) fabricated in accordance with our invention with modulated doping and bandgap, exhibited much higher mobilities as shown by curve II of FIG. 5. At room temperature, the mobility was about 6,000 cm2 V-1 sec-1, nearly 2.5 times greater than that of the uniformly doped multilayered sample. Moreover, as the temperature was decreased, the mobility dramatically increased to about 10,000 cm2 V-1 sec-1 at about 150 K and to 16,000 cm2 V-1 sec-1 at 50 K and below, more than 200 times greater than that of the uniformly doped sample. The dramatic increase of mobility with decreasing temperature was evidence of the efficacy of our structure to substantially reduce the adverse effects of ionized impurity scattering on mobility.
The high mobility of our invention can be exploited in a number of devices such as the FET shown in FIG. 6.
In general, a MESFET includes separated source and drain regions coupled by a channel in which depletion is created by voltage applied to an overlying gate electrode. Typically, when no voltage is applied to the gate, current flows between the source and drain, but when a voltage of suitable magnitude and polarity is applied, depletion (i.e., pinch-off) occurs in the channel and current flow between the source and drain is interrupted.
In the MESFET devices shown in FIG. 6, a multilayered semiconductor structure 100 having modulated doping and bandgap in accordance with the previous description is epitaxially grown on a semi-insulating substrate 102. Since the wide bandgap layers of the structure 100 are preferably n-type in order to exploit the higher mobility of electrons compared to holes, the source zone 104 and drain zone 106 are typically formed by diffusing, implanting or otherwise placing donors in localized zones 104 and 106 which extend at least through the structure 100 to the substrate 102. Source and drain electrodes 108 and 110 are then deposited by conventional techniques over the zones 104 and 106, respectively. The portion 112 of structure 100 which is located between source and drain zones 104 and 106 forms the channel of the FET. Gate electrode 114 is formed as a Schottky barrier contact directly on the channel. When a negative voltage is applied to the gate electrode 114, the channel is depleted and no conduction occurs between source 104 and drain 106. Conversely, when no voltage is applied to the gate, conduction occurs between the source and drain, thus exploiting the enhanced mobility of the multilayered channel 112.
It is to be understood that the above-described arrangements are merely illustrative of the many possible specific embodiments which can be devised to represent application of the principles of the invention. Numerous and varied other arrangements can be devised in accordance with these principles by those skilled in the art without departing from the spirit and scope of the invention. In particular, while our invention has been described with specific reference to a GaAs-AlGaAs example, it will be obvious that other lattice-matched materials which exhibit sufficiently large conduction or valence bans steps are also suitable; for example, Aly Ga1-y As--Alx Ga1-x As (0±y, x-y≧0.02 to yield a sufficiently large ΔEc), GaAs-AlGaAsP: InP-InGaAsP; InP-InGaAs; or InAs-GaAsSb.
It will also be appreciated that our invention, viewed from another aspect, is a method for enhancing the mobility of a given narrow bandgap semiconductor material by fabricating the narrow bandgap material in the form of a first plurality of spaced layers which are undoped or unintentionally doped and fabricating a second plurality of wide bandgap semiconductor layers contiguous with and interleaved with the narrow bandgap layers from a material which is substantially lattice-matched to the narrow bandgap layers, and is doped such that the impurity concentration-thickness product of the wide bandgap layers exceeds that of the narrow bandgap layers. When utilizing molecular beam epitaxy to perform this process, the Knudsen cells or ovens containing dopant source material for the wide bandgap layers would be shuttered closed during the growth of the narrow bandgap layers so that any impurities incorporated into the narrow bandgap layers would result primarily from background contamination in the ultra high vacuum system.
Stormer, Horst L., Dingle, Raymond, Gossard, Charles
Patent | Priority | Assignee | Title |
5142341, | Apr 08 1991 | Motorola, Inc. | Enhanced conductivity quantum well structure having resonant interface phonon induced charge coupling |
5296395, | May 17 1991 | International Rectifier Corporation | Method of making a high electron mobility transistor |
5521735, | Aug 09 1990 | Canon Kabushiki Kaisha | Electron wave combining/branching devices and quantum interference devices |
6888170, | Mar 15 2002 | Cornell Research Foundation, Inc | Highly doped III-nitride semiconductors |
6953740, | Mar 15 2002 | Cornell Research Foundation, Inc | Highly doped III-nitride semiconductors |
7482191, | Mar 15 2002 | Cornell Research Foundation, Inc. | Highly doped III-nitride semiconductors |
7485901, | Mar 15 2002 | Cornell Research Foundation Inc. | Highly doped III-nitride semiconductors |
7993938, | Mar 15 2002 | Cornell Research Foundation, Inc. | Highly doped III-nitride semiconductors |
Patent | Priority | Assignee | Title |
3626257, | |||
3626328, | |||
3721583, | |||
3737737, | |||
3758875, | |||
3838359, | |||
3882533, | |||
3911376, | |||
3915765, | |||
4016505, | Mar 20 1973 | Matsushita Electronics Corporation | Double heterostructure semiconductor laser |
4088515, | Apr 15 1973 | International Business Machines Corporation | Method of making semiconductor superlattices free of misfit dislocations |
4103312, | Jun 09 1977 | International Business Machines Corporation | Semiconductor memory devices |
4137542, | Apr 20 1977 | International Business Machines Corporation | Semiconductor structure |
4141025, | Mar 24 1977 | Gosudarstvenny Nauchno-Issle-Dovatelsky I Proektny Institut | Semiconductor structure sensitive to pressure |
4169997, | May 06 1977 | Bell Telephone Laboratories, Incorporated | Lateral current confinement in junction lasers |
DE2607940, |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
May 26 1987 | AT&T Bell Laboratories | (assignment on the face of the patent) | / |
Date | Maintenance Fee Events |
Nov 05 1998 | ASPN: Payor Number Assigned. |
Date | Maintenance Schedule |
Aug 20 1994 | 4 years fee payment window open |
Feb 20 1995 | 6 months grace period start (w surcharge) |
Aug 20 1995 | patent expiry (for year 4) |
Aug 20 1997 | 2 years to revive unintentionally abandoned end. (for year 4) |
Aug 20 1998 | 8 years fee payment window open |
Feb 20 1999 | 6 months grace period start (w surcharge) |
Aug 20 1999 | patent expiry (for year 8) |
Aug 20 2001 | 2 years to revive unintentionally abandoned end. (for year 8) |
Aug 20 2002 | 12 years fee payment window open |
Feb 20 2003 | 6 months grace period start (w surcharge) |
Aug 20 2003 | patent expiry (for year 12) |
Aug 20 2005 | 2 years to revive unintentionally abandoned end. (for year 12) |