JPS61242082A - Semiconductor element - Google Patents

Abstract

PURPOSE:To improve the efficiency of gate control while being conformed to the variation of the conductivity of a channel by directly forming source-drain electrodes consisting of a superconductor to a semiconductor channel section and changing effective semiconductor channel length of width thereof by a super conductive proximity effect. CONSTITUTION:Source-drain electrodes 21, 31 composed of a superconductive material are shaped directly (direct contact) to a semiconductor channel 4, and effective semiconductor channel length is shortened by using a superconducting region in a semiconductor formed by a superconducting proximity effect from the source and drain electrodes 21, 31. Accordingly, effective semiconductor channel length is changed by altering the width of the superconducting regions in source and drain sections by the superconductive proximity effect, and the variation (decrease and increase) of semiconductor channel length by the control of a gate 5 and the change (increase and decrease) of the conductivity of a semiconductor channel section are summed up particularly, thus largely improving a current control factor.

Description

[Detailed Description of the Invention] [Summary] Source and drain electrodes made of a superconductor are provided directly in the semiconductor channel portion, and the effective semiconductor channel is shortened by superconducting regions of the source and drain due to the superconducting proximity effect. Alternatively, by changing the width of the region, the semiconductor channel length can be changed (or, by combining the increase or decrease in the conductivity of the channel with the change (decrease or increase) in the semiconductor channel length, the gate control efficiency can be improved. do.

[Industrial application field]

The present invention relates to a semiconductor device, and particularly to a semiconductor FET device that has good gate control efficiency, high speed, and low power consumption.

[Conventional technology]

Figures 5 and 6 show the element structure of a conventional semiconductor FET (Figure 5).
Fig. 6 shows the MES type; Fig. 6 shows the MIS type.

In the MES type shown in FIG. 5, the current between the source/drain (2b15b) is controlled by applying the gate voltage to the gate electrode 5 (by controlling the thickness of the depletion layer 4α in the channel semiconductor layer 4). 1α is a semiconductor substrate, 2α and 6α are source and drain electrodes. In the MIS type shown in Fig. 6, the current between the source and drain (2b75 b ) is controlled by increasing or decreasing the carrier concentration in the channel part by applying a gate voltage. .In FIG. 6 (2), parts corresponding to those in FIG. (usually high concentration) or a normal conduction electrode, and the length of the semiconductor in the channel portion was 4r: There was no change.

[Problem that the invention seeks to solve]

Conventionally, the efficiency of current control using C was still insufficient, even though the efficiency was 5:g. An object of the present invention is to improve the efficiency of current control by a gate.

[Means for solving problems]

The present invention uses a superconducting region in a semiconductor that is formed from the source and drain electrodes by the superconducting proximity effect (the source and drain electrodes made of a superconducting material are formed directly in the semiconductor channel and in direct contact with the semiconductor channel). The most important features are shortening the effective semiconductor channel length and controlling the FET current to increase or decrease the proximity effect region.

[For production]

According to the present invention, by changing the width of the superconducting region in the source and drain parts due to the superconducting proximity effect,
The effective semiconductor channel length changes. In particular, by combining changes (decrease, increase) in the semiconductor channel length due to gate control (2) and changes (increase, decrease) in the conductivity of the semiconductor channel portion, it is possible to significantly improve the current control rate.

〔Example〕

1 to 3 are cross-sectional views of semiconductor devices in Examples 1 to 3:2 of the present invention, in which source and drain electrodes 21 and 51 made of a superconducting material are formed directly in the semiconductor channel portion. When a superconductor is in contact with a normal conductor, mutual diffusion of superconducting electrons and normal conducting electrons occurs at the interface.

This is known as the superconducting proximity effect.

FIG. 4 shows the superconducting pair potential distribution (corresponding to the number of superconducting electrons) near the channel portion due to this effect. In other words, the portion of the original semiconductor region near the superconducting electrode becomes a superconductor due to the proximity effect. The size of this proximity effect region is determined by the superconducting diffusion length ξ in the semiconductor, and is free In the electronic model, the cube root of carrier concentration and the square root of carrier mobility C are proportional to each other, and inversely proportional to the square root of carrier effective mass.

In a narrow bandgap high electron mobility compound semiconductor such as ILSB, it is relatively large, on the order of 0.1 to 0.3 microns. On the other hand, since the proximity effect ≦ 2, the region can be regarded as superconducting, ld is the condition at the semiconductor/superconductor interface, that is, the ease of transition of superconducting electrons into the semiconductor (the size of Δ in Fig. 4: transition It is also related to probability=Δ0/Δ,). In other words, even with the same ξU, Δ. ld is long from the larger one. For this reason, it is effective to clean the semiconductor/superconductor interface and, in some cases, highly dope the semiconductor surface to increase the transition of superconducting electrons. Usually 1. ．． 1d can be considered to be on the order of ξ7 or several times longer. In the present invention, by converting a part of the semiconductor into a superconductor (2) due to the proximity effect, the effective semiconductor channel length can be shortened, and the increase or decrease in the semiconductor effective length due to this change in length can be interpreted as a change in the conductivity of the channel part. Tomo (; used.

First, an embodiment of the junction type or MES type of the first factor will be described in detail. 1α is a semi-insulating or P-type semiconductor substrate, and 21.31 is a superconducting electrode.

4 is an R-type (for channel) semiconductor layer, 41 is a P-type (for gate Cp-rb junction) semiconductor layer (this layer is not present in the case of MES type), 4a is a depletion layer region, and 5 is a date electrode.

The length of the original semiconductor layer is either L or 1.0 due to the superconducting proximity effect as shown in FIG. ．． 1d becomes superconducting, and the actual semiconductor channel length is 1. becomes. For example, as 4,
If an n-type 1nAz (Dlo"~10"cm-') is used, the mobility is 4.2 and the mobility is 20000~5000a.
n>”/V, and the import is 0.1 pm −0,
It is 2 μm. In the case of L type 1rLAz, the semiconductor/metal interface barrier has a negative value (no barrier), so a large transition probability can be obtained by cleaning the semiconductor surface, so ld is about 0.5 to 1 μm, and L is 1 μm. Even in the above case, a submicron short channel FET can be easily formed.

FIG. 2 shows an element structure similar to that of the previous embodiment, except that an ohmic electrode is used as the gate electrode. For example, as the channel semiconductor layer 4, a square type JnSAk is used. In this case, the semiconductor/metal interface barrier has a positive finite value of ~0.17#V (see Figure 2 (5)), so the transition of superconducting electrons into the semiconductor is suppressed. When the carrier concentration is not very high, Δ0 (Figure 4) is small and 1.
．． 1. is close to zero.

If a positive bias is applied to the semiconductor 4 with respect to the source and drain superconducting electrodes, or if electrons are injected from the gate electrode (see the opening in FIG. 2 (G)), the transition of superconducting electrons increases.1. , 1 cL can be increased. This changes the semiconductor channel length.

Next, an example of the MIS type element shown in FIG. 3 (two will be explained). 1h is a semiconductor substrate, 6 is an insulating film for a gate (for Mis). The length of the original semiconductor layer is L.

When no voltage is applied between the gate and source or between the substrate, or when the carrier number in the channel part is set to zero or very small by applying a voltage (reverse voltage), l,
, l, are almost zero (Off state). Next, when applying a voltage in the operating direction between the gate and source (0 state)
, a surface inversion layer is formed (carrier concentration increases), and channel conductance increases. However, in this case, the carrier concentration further increases (ξ due to 2). Therefore, due to the superconducting proximity effect, l, , l, 1 has a finite value, and the actual semiconductor channel width is from L! In other words, the l, l part becomes superconducting (resistance is zero), and the conductivity of the channel part is large.These two effects make it possible to increase the FET current, so
Control efficiency by gates can be greatly increased. As a specific example, a semiconductor substrate 11J using P-type 1nAs will be explained. In p-type IrLAz, a naturally inverted channel layer is formed on the surface. When formed over the entire surface of the gate insulating film, this channel is maintained or disappears due to fixed charges. In the former case, the channel can be eliminated by applying a negative bias to the gate. Both correspond to the off state, there are no carriers, l(t is zero, and the channel conductance is also zero.If the gate bias is a positive electron gas, C; increases as the carrier a degree increases, and □ An increase in electron mobility occurs. These effects cause an increase in the axis, an increase in 1., 1ct, and a decrease in !. In this example, the current Since gate control efficiency can be controlled, gate control efficiency can be greatly increased.

In the above, an example of gate control by providing a gate electrode has been shown in the present invention, but by controlling the semiconductor channel portion by irradiating the semiconductor with light, electron beam, etc.,
Control of elements similar to those described above is possible. It can also be used as a detector.

〔Effect of the invention〕

As explained above, in the semiconductor element of the present invention,
By shortening the channel length using the superconducting proximity effect and improving the gate control efficiency, it is possible to significantly improve the gate control efficiency by simultaneously increasing the conductivity of the semiconductor channel and reducing the channel length due to the proximity effect. I can do it. In addition, compared to so-called semiconductor-based superconducting devices, the device of the present invention allows for a longer distance between superconducting electrodes.

This facilitates device fabrication (1) and is advantageous in improving the yield of device characteristics (2), which is important in integration.

[Brief explanation of the drawing]

Figure 's1 is a sectional view of a junction type or MES type half-shaped semiconductor device according to Embodiment 1 of the present invention, and Figure 2 (A) is a sectional view of Embodiment 2 of the present invention.
2(5) is a band structure diagram, FIG. 3 is a sectional view of the Mis-type semiconductor device of Example 3 of the present invention, and FIG. 4 is a superconducting proximity effect: FIGS. 5 and 6 are T diagrams showing element structures of conventional and other conventional semiconductor FETs. 1α, 16... Semiconductor substrate, 2α, 3α... Source and drain electrodes, 2b, 5b... Source and drain (semiconductor layer), 4... Semiconductor layer for channel, 21
, 31... Superconducting electrode, 41... Semiconductor layer for pn junction, 5... Gate electrode, 6... Insulating film for MIS... Superconducting electrode spacing (semiconductor region spacing for channel) ,
lct...superconducting region due to proximity effect, l,...effective channel length (l, = r, -t, -t,) Patent applicant Nippon Telegraph and Telephone Corporation agent Patent attorney Gobe Tamamushi (external) 2 people) Cross-sectional view of Example 3 Figure 3 Superconductor Semiconductor Superconductor Bare Botentia due to the effects of superconductivity Figure 5. Cross-sectional view of another conventional FET Figure 6

Claims (1)

[Claims] 1. Source and drain electrodes made of a superconductor are provided directly in the semiconductor channel portion, and superconductivity in the semiconductor channel portion is formed from the source and drain electrodes by the superconducting proximity effect. A semiconductor device characterized in that a semiconductor channel length is shortened by a region. 2. Source and drain electrodes made of a superconductor are provided directly in the semiconductor channel portion, and the semiconductor channel length is increased by the superconducting region in the semiconductor channel portion formed from the source and drain electrodes by the superconducting proximity effect. 1. A semiconductor device characterized in that a semiconductor channel length is changed by changing the width of the superconducting region due to a superconducting proximity effect. 3. Source and drain electrodes made of a superconductor are provided directly in the semiconductor channel region, and a gate electrode is provided between the two electrodes. and a semiconductor device characterized in that a semiconductor channel length is changed by contracting and stretching a superconducting region in a semiconductor channel portion formed from a drain electrode by a superconducting proximity effect. 4. Source and drain electrodes made of a superconductor are provided directly on the semiconductor channel portion, and a gate electrode is provided between the two, so that the conductivity of the semiconductor channel portion can be increased or decreased by gate control, and the conductivity can be increased or decreased by controlling the gate. As the conductivity of the semiconductor channel increases, the superconducting region in the semiconductor channel formed from the source and drain electrodes due to the superconducting proximity effect expands and contracts, and the semiconductor channel length decreases as the conductivity of the semiconductor channel increases. 1. A semiconductor device characterized in that the number of semiconductor devices increases.