JPS598070B2 - semiconductor equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a charge transfer device that sequentially transfers majority carriers in a semiconductor as an information source, and particularly to a charge transfer device having a novel electrode structure and a method for manufacturing the same.

In the explanation of the conventional example, the principle of the present invention, and the embodiments of the present invention described below, a semiconductor is used as a substrate as a charge transfer device, and many of the semiconductor layers formed on this substrate have a conductivity type different from that of the substrate. A charge transfer device that uses carriers as an information source is used, but this is, for example, a charge transfer device that uses a substrate other than a semiconductor such as an insulator and uses majority carriers in a semiconductor layer formed on the substrate as an information source. It can be applied in the same way.

In addition, in the following explanation, the charge transfer device uses electrons as the information source, but by reversing the polarity of the charge, the conductivity type of the semiconductor, and the potential relationship, the charge transfer device uses holes as the information source. can also be applied in the same way.

Figure 1 shows a conventional charge transfer device using majority carriers as an information source (BulkChar).
FIG. 2 is a diagram illustrating the relationship between the cross-sectional structure of an e-transfer device (hereinafter abbreviated as BCD) and the channel potential that optimizes the charge transfer characteristics of the BCD.

In the figure, 1 is a p-type silicon (Si) crystal that serves as a substrate, and 2 is an n-type S crystal formed on the substrate 1 to form a channel.
i-layer 3 is a SiO2 film formed on n-type Si layer 2;
4 is a polycrystalline Si electrode which is a first electrode embedded in the SiO2 film 3 at a distance, and 5 is made of aluminum (Al) and covers the SiO2 film in the gap between the first type electrodes 4. This is the second type of electrode.

In the structure shown in FIG. 1, an n-type Si layer 2 serving as a channel
Since the minimum required voltage VO (considered in terms of absolute value) to completely deplete is almost the same for each electrode, we now apply a driving pulse with the same DC level and pulse amplitude to each electrode 6 to 11. When applied, the channel potential has poor directionality, and the majority carriers (hereinafter referred to as signal charges (indicated by diagonal lines in the figure)), which serve as information sources, partially flow backwards and do not operate normally, as shown in Figure b.

In order to prevent this phenomenon, it is necessary to apply a driving pulse P1 to the first type electrodes 7, 9, 11, which normally accumulate signal charges, and to apply a driving pulse P1 to the second type electrodes 6, 8, 10, which are usually used as switches.
, and the drive pulse P2 applied to the drive pulse P2, for example, it is sufficient to drive BCD by 1 based on the potential relationship when considering a drive pulse with an amplitude of 10V.

However, this method requires three power supplies other than the ground potential (0V), which is not practical. On the other hand, FIG. 1c shows the case where too much DC bias is applied between the drive pulses.

As can be seen from the figure, a potential barrier is now formed between electrodes 9 and 11, and a portion of the signal charge is not transferred, resulting in deterioration of the BCD characteristics. Figure 1 d is the same figure b
(It is in the middle of 5C, and the BCD characteristics are good.

Channel potential under the first type electrode 11 and second type electrode 1
If the difference from the channel potential below 0 is VO6, then from the above discussion, the difference between 1 drive pulse amplitude V)
It turns out that it is sufficient if the relationship CP exists.

Here, when driving the BCD with four-phase driving pulses, the voltage is V as shown in FIG. 1d.

. Even if you keep it small, the rise time of the drive pulse (
Since most of the charge is transferred within a period of about 5 to 20 nsec, there is no need to worry about the charge flowing backwards, and the advantage is that the amount of signal charge that can be handled increases.
When driven with two-phase drive pulses (same drive pulse for electrodes 8 and 9, different drive pulses for electrodes 6, 7, 10, and 11, two types in total), ! :The best way to increase the amount of signal charge is to increase the amount of signal charge.

This is because if VOO deviates from this point, either the transfer time or the amount of signal charge becomes insufficient. However, this is accurate depending on the specifications for the frequency to be driven and the amount of signal charge to be transferred.

Of course, it does not have to be an o-ball. Now, applying a DC bias between drive pulses in order to create conditions such as the above equations (1) or (2) within the channel requires an increase in the number of power supplies, the need for DC bias adjustment, etc., and the BCD system In terms of implementation, handling is extremely complicated, and this is a major obstacle in the application of BCD.

An object of the present invention is to provide a device that eliminates the above-mentioned problems.

In order to achieve the above object, the present invention utilizes two types of electrodes of a charge transfer device, namely, a first type electrode (hereinafter referred to as carrier storage electrode) that stores carriers as an information source and a switch role. Among the second type of electrodes (hereinafter referred to as carrier transfer electrodes), the length of the carrier transfer electrode is shorter than the length of the carrier storage electrode, and
By forming the interface between the carrier transfer electrode and the gate oxide film to be closer to the semiconductor substrate or the back surface of the semiconductor layer, which serves as a charge transfer path, than the interface between the carrier storage electrode and the gate oxide film, the gap between the drive pulses is reduced. It is possible to obtain an electrode that can obtain the relationship of formula (1) or (2) above in the channel of a BCD and easily control the potential of the channel without applying a DC bias to the BCD channel. As a result, a charge transfer device that uses a small number of power supplies, is easy to handle, has a high degree of integration, and can be driven at high speed is realized.

The present invention will be explained below with reference to Examples.

This figure shows an embodiment of the present invention shown in FIG.
22 is a p-type Si layer forming a channel, and 23 is a SiO layer formed on the n-type Si layer 22.
2 film, 24 is a polycrystalline Si electrode which becomes a carrier storage electrode and is embedded in the SiO2 film 23 at a distance, and 25 is a first polycrystalline Si electrode.
This is an Al electrode that serves as a carrier transport electrode and is formed closer to the back surface of the n-type Si layer 22 than the seed electrode 24 .

As shown in the figure, when the carrier transfer electrode 25 is formed closer to the back surface of the n-type Si layer 22 than the carrier storage electrode 24, the carrier transfer electrode 25 is formed closer to the back surface of the n-type Si layer 22 than the carrier storage electrode 24. Since the voltage VO2 to be applied to the carrier storage electrode can be made smaller than the voltage VOl to be applied to the carrier storage electrode, there is no directivity in the channel potential even when driven by drive pulses of the same amplitude at the same DC level. This allows normal operation to occur. Here, the minimum voltage V is required to completely deplete the n-type Si layer. The difference between l and V62 (collectively referred to as VO) is approximately the difference in channel potential when the same voltage is applied to completely deplete the n-type Sl layer under the carrier storage electrode and the carrier transport electrode. equal, so (1) or (2
) In order to satisfy the condition of the expression (or), the carrier transfer electrode 25 should be brought close to the back surface of the n-type Si.

Here, as mentioned above, equation (3) is expressed by BC with four-phase drive pulses.
This is the case when driving D, and equation (4) is the optimum condition especially in the case of two-phase drive pulses. However, even in the case of two-phase drive, the voltage is exactly V.

As mentioned above, it is not necessary that l-VO2=0P. V above.

It is expressed as follows. However, VFB: n-type Si
Flat band voltage on the layer surface φ8: Surface potential when the n-type Si layer is completely depleted (potential difference from the point with the lowest potential relative to the signal charge in the channel to the surface) V
The voltage applied to OX:SiO2, φS, and Ox are expressed as follows.

φs214−Fdx″n(x)Dx(6) where q: elementary charge (1.6×10−19 coulombs) ε, i: permittivity of Si (1.0×10−12 farads/C77 !)ε
0X: Dielectric constant of SlO2 (3.4X10-13 farad/CTIL) TOx: Thickness of SiO2 film under the electrode (c!
n) d: Distance from the n-type Si layer surface to the point with the lowest potential for signal charges in the channel ((177!) x: n
Distance measurement (CTIL) n(x
): Impurity concentration in the n-type Si layer (C!!L-3) Therefore, using equations (3) to (7) above, the difference Δl in distance from each electrode to the back surface of the n-type Si layer is calculated. You can also ask for conditions.

Here, when the drive pulse is four-phase, VO, -
.

Experiments have shown that the optimum condition for 2 is about 1 to about 4.

Therefore, for example, if TOx~10-5 the n-type Si layer has a Gaussian distribution and the n(0)-
2 to 5 x 1016CfL-3, the total depth of the n-type Si layer is approximately 10-4 (: Tn. When the distribution is just deleted by Δl from the side,
It can be seen that the optimum value of Δl is 200λ to 2000 people. FIG. 3 shows another embodiment of the present invention, in which 41 is a p-type Si substrate, 42 is an n-type Si layer forming a channel, 43 is a SiO2 film formed on the n-type Si layer 42, and 44 is a p-type Si substrate. S
A polycrystalline Si electrode serving as a first type electrode is embedded in the iO2 film 43 at a distance, and a carrier transport electrode 45 is formed closer to the back surface of the n-type Si layer than the carrier storage electrode 44. This is an Al electrode.

As shown in FIGS. 2 and 3, an electrode with a structure in which the carrier transport electrode is closer to the back surface of the n-type Si layer 22 to 42 than the carrier storage electrode can reduce the DC bias between drive pulses as described above. In addition to the removal effect, it also prevents the potential in the n-type Si layer under the carrier transport electrode from fluctuating due to the influence of the potential in the n-type Si layer under the carrier storage electrodes on both sides. The carrier transport electrode allows the underlying n-type Si layer to be sufficiently controlled even at a low voltage, and has the effect of easily operating as a switch.

In particular, in the structures shown in Figures 2 and 3, carriers as an information source are stored at carrier storage electrodes.
The carrier transfer electrode is generally used simply as a switch, and in order to shorten the carrier transfer time, the shorter the carrier transfer electrode, the better the characteristics.
Due to the characteristics of the device, it is desirable to shorten the length of the carrier transport electrode. In this case, it is necessary to sufficiently eliminate the influence of the potential of the N8Si layer under the carrier storage electrode as described above to ensure stable operation of the carrier transport electrode. This effect becomes greater as the carrier transfer electrode is brought closer to the back surface of the n-type Si layer, but on the other hand (3)
~The limit as shown in equation (7) is the distance difference Δ! =
2, -12, and in some cases, it may not be possible to completely eliminate the influence from below the carrier storage electrode. The embodiment shown in FIG. 3 solves this difficulty by adding the above-mentioned restrictions to the distance difference Δl from the carrier storage electrode and the carrier transport electrode to the back surface of the n-type Si layer. In addition, the back surface of the n-type Si layer itself is also made uneven according to the unevenness of the upper electrode, and potential fluctuations in the n-type Si layer under the carrier transport electrode due to the influence from the n-type Si layer of the carrier storage electrode are created. As a result, it is possible to obtain a BCD that uses less power, has a high degree of integration, and can be driven at a low voltage.

However, in the case of the embodiment shown in Fig. 3, it is necessary to ensure that there is enough n-type Si layer to allow the signal charge to pass through the intermediate portion between the carrier storage electrode and the carrier transfer electrode, and to avoid cutting this layer. You need to be careful.

In the embodiments of the present invention described above, a SiO2 film was used as the Si insulating film as the semiconductor, and polycrystalline Sil or Al was used as the electrode. It is not limited to what is given.

For example, polycrystalline Si and Al that form the electrodes may be the same material, and each may be made of other conductive materials, such as metals such as tantalum (Ta), tungsten (W), and molybdenum (MO), and these materials. Combinations may also be used.

In addition to SiO2, alumina (Al2
O5), silicon nitride (Si3N4), etc., and composite films thereof can be used. Also, although we have described the case where Si is used as a semiconductor, this is not limited to Si, and any semiconductor may be used. For example, , germanium (G
The same objective can be achieved by using composite semiconductors such as e) or gallium arsenide (GaAs), or by using a combination of these. As explained above,
According to the present invention, by bringing the carrier transfer electrode closer to the back surface of the semiconductor layer serving as the channel than the carrier storage electrode, the DC bias between the driving pulses of the BCD is removed.
The number of power supplies can be reduced, and the carrier transfer electrode can sufficiently control the channel immediately below it even with short electrode lengths and low drive voltages, and accordingly,
It can be seen that the present invention has a remarkable effect on improving the performance of semiconductor devices, such as by being able to obtain a multi-carrier type charge transfer device that can be driven at high density and at low voltage and has excellent high frequency characteristics.

[Brief explanation of drawings]

FIG. 1 is a diagram illustrating the structure of a conventional multi-carrier charge transfer device and the relationship between channel potentials that optimize its charge transfer characteristics, and FIGS. 2 and 3 respectively illustrate the structure of an embodiment of the present invention. FIG.

Claims (1)

[Claims] 1 The substrate formed on the substrate includes a semiconductor layer of one conductivity type that is electrically insulated from the substrate, and a plurality of electrodes that are sequentially arranged on the semiconductor layer with an insulating film interposed therebetween, and serves as an information source. In a charge transfer device that sequentially stores and transfers charge carriers in the semiconductor layer under the electrode, the electrode includes a charge storage electrode and a charge transfer electrode adjacent to the upstream side of the charge storage electrode, and the charge transfer electrode is shorter than the electrode length of the charge storage electrode, and the interface between the charge transfer electrode and the insulating film is closer to the back surface of the semiconductor layer than the interface between the charge storage electrode and the insulating film. The insulation value of the minimum voltage that should be applied to each electrode to completely deplete the semiconductor layer is V_c, and V_c at the charge storage electrode is V_
c_1, V_c at the charge transfer electrode is V_c_2
, the amplitude of the driving pulse for driving the charge transfer device is V_
When c_p, 0V<V_c_1−V_c_2<
A semiconductor device characterized in that V_c_p.