JPS633467B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a charge transfer element that transfers charge carriers in the vicinity of a semiconductor surface by an electric field associated with an ultrasonic wave propagating through a piezoelectric material, and more specifically, the present invention relates to a charge transfer element that transfers charge carriers near a semiconductor surface by an electric field accompanying an ultrasonic wave propagating through a piezoelectric material. The present invention relates to a monolithic charge transfer device having a channel stop that does not displace.

Traditionally, elements (devices) for transmitting information
There are various types. For example, devices that utilize magnetic properties include devices called magnetic bubbles that move parts of different magnetic polarities on a flat plate, and devices that sequentially change the polarity of magnetism stored in a perforated magnetic core. There is.

In addition, there are devices called BBDs that integrate circuits using MOS field effect transistors and capacitors that utilize electrical characteristics.
There are devices called CCDs that sequentially change the potential near the semiconductor surface by applying voltage one after another to a series of electrodes provided on an insulating oxide film on the semiconductor surface, thereby moving minority carriers one after another. . Among these, CCD is a device that is currently attracting a lot of attention.

However, all of these devices have the following drawbacks. For example, magnetic bubbles have a long access time, approximately 1 msec, and BBD
The loss of charge carriers during transfer (hereinafter referred to as transfer loss) is large. In addition, in CCDs, the potential under the transfer electrode is generally flat and has no slope, so some of the charge carriers move due to the concentration diffusion of the signal charge itself, and therefore the transfer rate cannot be made higher than this diffusion rate. However, it is difficult to manufacture because it requires a series of electrodes for transfer and lead wires to sequentially apply voltage to them, and the structure is complicated.

Also, unlike the above devices, a device (hereinafter referred to as
(referred to as SAW-CTD) are also known. However, all of them have many drawbacks and are not practical.

For example, as disclosed in U.S. Pat. No. 3,858,232, a device in which a series of metal electrodes is provided between a piezoelectric material and a semiconductor has a potential that is generated by surface waves under the electrodes due to the provision of the metal electrodes. The gradient is eliminated, and as a result, the transfer rate is limited by the concentration diffusion of the charge carriers, making it suffer from the same drawbacks as conventional CCDs. Furthermore, the same specification discloses a device in which a piezoelectric material and a semiconductor are in direct contact, but in this device, charge carriers are also injected and released between the semiconductor and the piezoelectric material, and the injected charge carriers become piezoelectric. It has the disadvantage that the S/N ratio decreases and hysteresis occurs in the electrical characteristics of the device because it enters into traps in the material. A further disadvantage of this device is that there is no blocking region (hereinafter referred to as channel stop) for preventing charge carriers from being removed from the path to which they are to be transferred (hereinafter referred to as transfer channel). Therefore, during transfer, some of the charge carriers are removed from the transfer channel, causing transfer loss. Furthermore, since there is no channel stop around the transfer channel, the transfer channel can only be designed in a straight line, so when this device is used as a delay line, it has the disadvantage that it becomes long. Furthermore, as disclosed in Japanese Patent Publication No. 52-36677, a device in which an insulating layer is provided between the piezoelectric material and the semiconductor has been devised, but this device also has no channel stop and therefore suffers from large transfer losses. It is not suitable for practical use. Also published in Ultrasonics Symposium Proceedings, IEEE Cat.
#76 As described on pages 201 to 204, devices have been devised in which an insulating layer is provided on a semiconductor, a piezoelectric material layer is provided on top of the insulating layer, and an electrode is further provided on top of the insulating layer. Since this method also does not have a channel stop, the transfer loss is large and it is not suitable for practical use. Also, Applied Physics Letters, Volume 29,
As described in No. 2, page 82, a device in which a gap is provided between a piezoelectric material and a semiconductor is also known, but this device has unstable electrical characteristics due to the gap, and It has the disadvantage of being easily deformed and difficult to handle.

As mentioned above, with conventional CCDs, it is practically impossible to increase the transfer frequency while keeping transfer losses small, and with SAW-CTDs, transfer losses are large and the characteristics are unstable, making it practically impossible to increase the transfer frequency. It was not suitable for Furthermore, in the SAW-CTD, when charge carriers are injected, although there is an injection diode, there is no injection gate, so it is difficult to control the injection time and amount.

An object of the present invention is to provide a SAW-CTD which has extremely low transfer loss and can therefore be put to practical use as a charge transfer device for transferring information for the first time. Transfer loss has not been sufficiently considered in conventional SAW-CTD. In other words, in conventional SAW-CTDs, it was only considered that the charge carriers are forced to move by the electric field because they fall into the potential valleys, but essentially no consideration was given to the connection of the charge carriers. However, as a result of research conducted by the present inventor, it was actually discovered that even in SAW-CTDs prior to the present invention, there is a transfer loss due to straining of the fibers, and the present invention has been made based on this knowledge. The present inventor believes that the causes of transfer loss are, firstly, deviation of charge carriers from the transfer channel in a direction perpendicular to the direction in which charge carriers are transferred (lateral direction); and secondly, charge carriers exceeding the transfer channel. Taking into account the fact that the potential is not completely constrained by the potential troughs associated with sound waves, and thirdly the capture of charge carriers by empty traps on the semiconductor surface, the transfer loss caused by them has been drastically reduced. In other words, for the first cause, channel
A stop was set up. This prevented charge carriers from escaping laterally from the transfer channel. As for the second cause, we correctly analyzed for the first time the conditions that constrain the charge carriers so that they do not fall out, and thereby clarified the transfer conditions that do not cause them to fall out. Note that restraining charge carriers in the valley of the potential that progresses with the ultrasonic wave is called bunching. Further, as a structure that can perform bunching perfectly, a monolithic type in which a piezoelectric body and a semiconductor are integrated is preferable because a strong potential necessary for bunching can be applied to the semiconductor surface. Further, the smaller the wavelength of the ultrasonic wave, the smaller the potential of the ultrasonic wave required for bunching. When surface acoustic waves are used as ultrasonic waves, the potential associated with the surface acoustic waves exists within about one wavelength from the surface, so the thickness of the piezoelectric body needs to be shorter than the wavelength of the surface acoustic waves. As for the third cause, we estimated the statistical time it takes for charge carriers to fall into a trap on the semiconductor surface, and clarified that it is possible to transfer charge carriers in a shorter time than that time.

Furthermore, the purpose of the present invention is to provide a SAW with stable operation.
- To provide CTD. A further object of the present invention is to provide a SAW-CTD with high transfer density. Furthermore, it is an object of the present invention to provide a charge transfer device with extremely low transfer loss, high transfer density, high transfer speed, no transfer electrodes, low power consumption, stable operation, and simple structure. The purpose is to provide a product that is easy to manufacture and inexpensive. Another object of the present invention is to provide a SAW-CTD for use in high-density memories, high-speed memories, signal processing such as delay lines, facsimile machines, and imaging plates.

The present invention relates to an ultrasonic generation unit for propagating ultrasonic waves to a piezoelectric layer, and a transfer unit for bunching charge carriers into potential valleys associated with the ultrasonic waves and transferring them through a depletion layer near the semiconductor surface. In a monolithic charge transfer device having a channel, a detection section for the transferred charge carriers, and a channel stop for preventing the charge carriers from being removed from the transfer channel, the bunching is a space created by the transferred charge carriers themselves. A charge transfer element characterized in that it provides a transfer channel with a potential greater than the sum of a potential due to electric charges, a potential due to a diffusion effect due to transferred charge carriers, and a potential for bunching one charge carrier (test charge). be.

A further preferred embodiment of the invention is a charge transfer device, characterized in that the time for the ultrasonic waves to travel over the distance of one wavelength is shorter than the statistical time for the charge carriers to fall into the surface traps of the semiconductor. .

Examples of semiconductors used in the present invention include
Si, Ge, ZnO, ZnS, CdS, Te, Se, ZnSe,
InSb, InAs, InP, GaP, GaAs, GaSb, Ga _1-x
There are Al _x As, GaAs _1-x P _x and InGaP.

Note that the material is not limited to the above-mentioned materials, and any semiconductor can be applied to the present invention in principle. Among these, preferred semiconductors are Si, Ge, GaAs, InAs, InSb, and CdS.
More preferred is Si. Furthermore, two or more types of semiconductors may be used in combination.

Examples of piezoelectric materials include ZnO, AlN,
GaAs, GaP, GaSb, InAs, InP, InSb, Se,
Te, Cds, ZnS, PZT ceramics [Pb(Sn _0.5
Sb _0.5 ) O ₃ −PbZrO ₃ −PbTiO ₃ etc.] γ-Bi ₂ O ₃ group compounds (Bi ₁₂ GeO ₂₀ , Bi ₁₂ PbO ₁₉ , Bi ₁₂ SiO ₂₀ )
There are LiNbO ₃ , LiTaO ₃ , and crystal. In principle, materials other than those mentioned above can be applied to the present invention as long as they are piezoelectric materials. Among these, preferred are ZnO, CdS, and AlN. Further, two or more types of piezoelectric bodies may be used in combination.

Since the charge transfer device of the present invention is a monolithic type, when a semiconductor and a piezoelectric material are laminated, if the piezoelectric material is highly insulating, the piezoelectric material layer is directly deposited on the semiconductor by vacuum evaporation or sputtering. Although they may be laminated by various thin film forming methods such as ion plating or ion plating, it is generally preferable to form an insulator layer on a semiconductor and then laminate a piezoelectric layer thereon.

Examples of substances forming such an insulating layer include SiO, SiO ₂ , Al ₂ O ₃ , Si ₃ N ₄ , TaN,
These include TiO ₂ , Ta ₂ O ₅ , AlN, boron glass, phosphorus glass, and oxygen-excess polycrystalline silicon. Note that it is not necessary to limit the materials to the above materials, and in principle, any insulator can be used in the present invention. Preferred are Si ₃ N ₄ ,
These include Al ₂ O ₃ and SiO ₂ , with SiO ₂ being more preferred. Furthermore, two or more types of insulators may be used in combination. When creating a piezoelectric material on an insulator, it is desirable to create an insulator layer in which neither ions nor surface states are introduced into the insulator.

Furthermore, if the semiconductor also serves as a piezoelectric material, there is no need to perform the above-described lamination. Examples of such semiconductors include CdS, ZnO, ZnS, GaAs,
Semiconductors such as GaP, GaSb, InAs, InP, InSb, Se, Te can be used.

A depletion layer near the semiconductor surface is, for example, in the case of a p-type semiconductor, when a positive voltage is applied, holes near the surface of the p-type semiconductor are repelled by this positive voltage.
There is a region where there are almost no holes that are pushed from the surface toward the bulk, and this region is referred to as this region. Also p
Even when an n-type semiconductor is embedded in the surface of a p-type semiconductor, if the vicinity of the surface of the p-type semiconductor, which is the substrate, is in the above state, there will be a region from the n-type semiconductor to the p-type semiconductor where there are almost no holes. up to. In the case of an n-type semiconductor, this also refers to a region where there are almost no electrons.

To form such a depletion layer, for example, p
In the case of a type semiconductor, positive ions, and in the case of an n-type semiconductor, electrons or negative ions can be formed by causing them to exist in an insulating layer provided on the semiconductor by a method such as ion implantation. Furthermore, a depletion layer can be formed by applying a voltage through an insulating layer on the semiconductor, so in this case, a conductor serving as an electrode may be provided on the insulating layer. Examples of such conductors include Al, Au, Ag, Cr,
Metals such as Ni, Pt, and Mo are representative, but other metals and alloys can also be applied to the present invention. Particularly preferred are conductive materials such as In ₂ O ₃ --SnO ₂ and polysilicon. Generally, this conductive layer may be laminated on a piezoelectric layer laminated directly or indirectly on a semiconductor, but if the piezoelectric layer has low insulation properties, it is preferable to interpose an insulating layer between them. . The reason for this is that when electrons or holes are injected from the conductor layer into the low-insulating piezoelectric layer, the electrical characteristics near the semiconductor surface change.

For example, the ultrasonic wave generator may be placed on a piezoelectric layer.
A comb-shaped electrode made of metal such as Al or Cr may be used, but other configurations based on known ultrasonic generation methods may also be used. In any case, any device that allows ultrasonic waves to propagate through the piezoelectric layer may be used, and the number of ultrasonic wave generators may be two or more, and their positions can be freely determined depending on the purpose. In addition, when using a comb-shaped electrode, ultrasonic waves can be generated by applying a high-frequency voltage to the comb-shaped electrode, but as shown in Figure 16a, a pad made of metal such as Al or Cr is provided. This is preferable because the efficiency of generating ultrasonic waves becomes high. In addition, ultrasonic waves are not limited to surface acoustic waves; pseudo surface waves,
Boundary waves, bulk waves, etc. may also be used.

A transfer channel is a path for moving charge carriers, and there may be a plurality of them depending on the purpose, or they may be curved, and the width thereof does not need to be always constant. Furthermore, the transfer channels may be split or merged.

As a method of providing a channel stop, for example, in the case of a p-type semiconductor, a p ⁺ -type semiconductor may be diffused near the surface of the p-type semiconductor, or a channel stop may be provided on the p-type semiconductor. A conductive layer may be provided on top of the insulating layer and a negative voltage may be applied thereto. A channel stop can also be provided in the case of an n-type semiconductor based on a similar idea. Also, the thickness of the insulator layer on the semiconductor may be thinner on the transfer channel and thicker at the periphery of the transfer channel.

The detection section may be any device that can detect charge carriers near the semiconductor surface. For example, when p-Si is used as the semiconductor, p-Si and n ⁺
- Any diode made of Si may be used. However, a detection section having a diode and a gate made of a metal such as Al or Au is preferable. This detection section is for detecting the transferred charge carrier, and uses the so-called current output method.
A configuration based on a known detection method such as the FDA method, FGA method, or DFGA method can be used. Although the element of the present invention is not necessarily required, if an injection section for injecting charge carriers is provided in the transfer channel, the element of the present invention can be used as a delay line element, for example. Such an injection section may have essentially the same function as the detection section. Alternatively, a configuration based on a known injection method such as the so-called DI method, DC method, PE method, etc. may be used. These detection sections or injection sections only need to be located in contact with the transfer channel. For example, a device located next to the transfer channel has advantages such as not attenuating the ultrasonic wave, making it easy to create a structure that guides the propagation of the ultrasonic wave, and preventing noise from entering the detected signal. An arbitrary number of detection sections or injection sections can be provided at arbitrary positions as necessary. Further, when the transfer channel is irradiated with light, pairs of electrons and holes are generated in the semiconductor, and a signal is detected by the minority carriers being transferred and detected. In this case, the injection part may not be provided.

In the present invention, the perfect bunching condition refers to the potential due to the space charge created by the transferred charge carrier itself, the potential due to the diffusion effect of the transferred charge carrier, and the potential due to the space charge created by the transferred charge carrier itself, and
This is a condition obtained by correctly elucidating the propagation characteristics of ultrasonic waves near the semiconductor surface for the first time. The value of the potential that satisfies this condition is larger than the value known as the potential that satisfies charge transfer in conventional SAW-CTDs. (Reference IBM Technical Disclosure Bulletin Vol. 18, No. 5, pp. 1613-1614) In other words, conventional charge transfer potentials could only carry a portion of the charge carrier; By providing a potential that satisfies the above conditions, charge transfer without transfer loss due to unloading became possible. For example, in a system in which a semiconductor, an insulator, a piezoelectric material, and a metal are stacked, this complete bunching condition is as follows: |Φ||(S+D/qμ)(∂q/∂θ)+νλ/2π
μ|(1) Here, Φ is the potential required for complete bunching.

Also, S = γ / C _i + 2 δ / ε _p + ε _s (reciprocal of practical capacitance felt by ultrasonic waves) γ = 0.5 to 1.2 (parameter) Ci = combined capacitance of piezoelectric layer and insulator layer, δ = Thickness of piezoelectric layer, ε _p = permittivity of piezoelectric material,
ε _s = dielectric constant of semiconductor, D = diffusion constant of charge carrier, q = charge amount of charge carrier, μ = mobility of charge carrier, θ = phase angle of ultrasonic wave at 2π(x/λ-ft), ν=ultrasonic velocity, λ=ultrasonic wavelength, x
= coordinates in the direction in which the ultrasound waves travel, f = frequency of the ultrasound waves, and t = time. That is, a preferred embodiment of the charge transfer device of the present invention is one that generates ultrasonic waves having a potential that satisfies equation (1).

Further, if the statistical time for a charge carrier to fall into a surface trap of a semiconductor is τ _f , and the time for an ultrasonic wave to travel a distance of one wavelength is τ _s , then a more preferable embodiment of the charge transfer device of the present invention is τ _f It generates ultrasonic waves with a relationship of >τ _s . Here τ _f
is determined as being inversely proportional to the product of the capture cross section σ _o of the semiconductor surface trap for charge carriers, the number of times the charge carrier encounters the surface trap 1/τ, and the charge carrier density n _s . That is, τ _f
=1/(σ _o ×1/τ×n _s ).

For example, σ _o for the charge carrier of the surface trap at the Si-SiO ₂ interface is 10 ^-15 to 10 ^-18 cm ^-2
In the case of 1/τ is (1.9×10 ^-13 ) ^-1 sec ^-1 at room temperature, when n _s = 7×10 ¹⁰ cm ^-2 τ _f = 2.8×10 ^-8
× ~ 2.8 × 10 ^-6 sec (the frequency at which the ultrasonic wave travels for one wavelength is τ _f is 360 MHz to 360 KHz).
When comb-shaped electrodes are used in the ultrasonic generation section, τ _s is determined mainly by the piezoelectric material and the semiconductor, which affect the distance between the electrodes and the speed of the ultrasonic waves.
For example, stacked Si (200μ)/SiO ₂ (1000Å)/
For ZnO (6000 Å), if the sound velocity is 4.3×10 ⁵ cm/sec, the wavelength must be within 12 μ in order to make the ultrasonic frequency 360 MHz or higher. In this way, the shape of the comb-shaped electrode that satisfies the above-mentioned τ _s can be determined. The ultrasonic wavelength and electrode shape can be determined using similar calculations for electrodes other than comb-shaped electrodes.

According to the present invention, SAW with extremely low transfer loss
- A CTD can be created. Further, according to the present invention, a SAW-CTD with extremely low transfer loss can be freely designed. Therefore, according to the present invention, there is almost no transmission loss, and the wavelength of the ultrasonic wave can be changed.
Since it can be made as small as about 1μ, it is possible to achieve a spatial resolution of about 1/20 of the size of conventional CCD electrodes, and it is possible to obtain a large number of bits.

In the following embodiments, a case will be explained in which charge carriers are generated from the injection part and transferred, but optical information can be detected by irradiating light or the like onto the transfer channel and photoelectrically converting it to generate charge carriers. Of course, this is possible.

The present invention will be specifically described below using Examples. First, a method for manufacturing an example of the charge transfer device of the present invention will be explained, and then examples of various aspects of the charge transfer device of the present invention will be explained. It goes without saying that the charge transfer device of the present invention is not limited only to these examples. Although most of the description below will be made in the case of a p-type semiconductor and the charge carriers are electrons, the present invention is not limited thereto. The present invention can also be applied to, for example, an n-type semiconductor in which the charge carrier is a hole. In that case, in this specification, p should be read as n, p ⁺ as n ⁺ , electrons as holes, and the polarity of the potential reversed.

Example 1 A method for manufacturing an example of the charge transfer device of the present invention will be described with reference to FIGS. 1 to 15 (all cross-sectional views). A p-Si substrate 10 with a thickness of approximately 200μ is heated in an oxygen atmosphere at approximately 1200°C for 7.5 hours in an electric furnace to thermally oxidize its surface.
A SiO ₂ layer 11 was formed (see FIG. 1). Then _SiO2
A photoresist layer 12 was applied to a thickness of about 5000 Å on layer 11, and then prebaked by heating at 80° C. for 25 minutes (see FIG. 2). Next, ultraviolet light 14 is applied to the photoresist layer 1 through a mask 13 having a pattern of light-transmitting parts corresponding to an injection part for injecting charge carriers and a detection part for detecting the transferred charge carriers.
2 (see Figure 3). Next, the photoresist layer 12 was developed to remove the irradiated area, and after washing with pure water, the remaining photoresist layer 12 was post-baked by heating at 80° C. for 30 minutes or more to form a resist layer 12'. (See Figure 4). Next, the above laminate was immersed in 20 g of ammonium fluoride, 5 c.c. of hydrogen fluoride, and 20 c.c. of pure water for 8 minutes, and the SiO ₂ layer exposed through the area where the resist layer 12' was removed was removed. 11
was etched (see Figure 5). Next, this laminate was placed in an acetone solution and subjected to ultrasonic cleaning to remove the remaining resist layer 12' (6th
(see figure). Next, SiO ₂ 15 doped with phosphorus (P) is applied to a thickness of about 2000 Å on top of this laminate.
Pre-baking was performed by heating at approximately 150°C for 40 minutes in an electric furnace (see Figure 7). Next, this laminate is heated to about 1200℃.
It is placed in an N ₂ gas atmosphere for about 2 hours to selectively diffuse phosphorus into the Si substrate 10 in the region where the phosphorus-doped SiO ₂ 15 is in contact with the Si substrate 10, ^and
Then, an injection part 16 and a detection part 17 were formed (see FIG. 8). Next, the SiO ₂ layer 11 was removed by etching to expose the surface of the Si substrate 10 (see FIG. 9). Next, this surface was heated at about 1050° C. for about 2.5 hours to form an SiO ₂ layer 18 with a thickness of about 1000 Å as an insulating layer (see FIG. 10).

Next, a photoresist layer is formed on this surface, pre-baked, the photoresist layer is developed by irradiating ultraviolet rays through a mask having a transparent area in a narrower area above the injection part 16 and the detection part 17, and then post-baked. , etching the _SiO2 layer 18,
By repeating the removal of the resist layer in the same manner as described above, the implantation part 1 having a cross-sectional view as shown in FIG.
6 and the insulator SiO ₂ layer 18 on top of the detection part 17
A perforated laminate was formed.

Next, Al19 was deposited on the surface of this laminate, and a photoresist layer 20 was applied thereon and prebaked. Next, the above-mentioned injection section 16 and detection section 17
The ultraviolet rays 22 are applied to the photoresist through a mask 21 which has a wiring pattern of a channel stop and its pattern, an injection gate and its wiring pattern, a detection gate and its wiring pattern, and a pad pattern below a comb-shaped electrode (to be described later) as non-transparent parts. Layer 20 was irradiated (see FIG. 12). Next, after developing and washing with water to form a resist layer in the above pattern, the exposed Al layer 19 was etched with a mixed solution of phosphoric acid and nitric acid, and then the resist layer was removed.

In FIG. 13, 23 is a pad, 24 is a channel stop, 25 is an injection gate, and 26 is a detection gate. Note that Al for wiring is omitted from the drawing. Next, ZnO 27 was formed as a piezoelectric material on this laminate by sputtering to a thickness of about 0.6 μm (see FIG. 14). Next, ZnO layer 2
A comb-shaped electrode 28 and an electrode 29 for forming a depletion layer were formed of Al by a lift-off method on 7, each having a line width and a spacing of 3 μm. Furthermore, Au 30 was deposited on the back side of the Si substrate 10 of this laminate (see FIG. 15).

FIG. 16a is an enlarged cross-sectional view along the transfer channel of an example of the charge transfer device according to the present invention manufactured as described above, and FIG. 16b is a plan view thereof. Various non-visible components are also shown to illustrate their relative positions. In the drawing, terminals for applying voltage to injection diodes and other devices are shown outside the transfer path, but the arrangement and size of these terminals can be selected as appropriate, and this drawing is used for ease of understanding for the sake of explanation. It has been deformed and enlarged.
Further, the device for applying voltage to these terminals is omitted from the drawing.

In this charge transfer element, a high frequency voltage is applied to the comb-shaped electrode 28 to generate surface acoustic waves on the piezoelectric layer 27. Here, surface acoustic waves 32 traveling in the direction of arrow 31 are used for charge transfer. 3 in the drawing
3 is a potential wave associated with a surface acoustic wave 32, which travels in the direction of an arrow 31. Now, in order to transfer charge carriers, for example, +10V is applied to the conductor layer 29 made of Al to deplete the surface of the semiconductor 10, and 0.1V is applied to the injection part (injection diode) 16, and 0.1V is applied to the injection gate 25.
1.0V to the detection part (detection diode) 17, 0.5V to the detection part (detection diode) 17,
1.5V is applied to the detection gate 26 and -10V is applied to the channel stop. Injection part 16
Before charge carriers are injected into the surface acoustic wave, a surface acoustic wave is generated, and the charge carriers existing near the semiconductor surface are transferred and removed by being confined to a potential valley by the electric field associated with the surface acoustic wave. Next, when the injection gate 25 is opened from the injection part 16 and charge carriers are injected into the depletion layer near the surface of the semiconductor, the charge carriers 3
4 is restrained by the potential wave 33 and transferred within a transfer channel consisting of a depletion layer surrounded by channel stops, and detected by the detection section 17 by opening the detection gate 26.

At this time, the specific resistance of Si is 130Ωcm, the thickness of ZnO is 0.6μ, the wavelength of surface acoustic waves is 12μ, and when the charge carrier density to be transferred is 10 ¹⁰ /cm ² , there is a sufficient amount to satisfy the bunching condition. We generated ultrasonic waves with a potential of 30 mW and a power that could be applied to the end of the transfer channel. In addition, the statistical time τ _f for charge carriers to fall into the semiconductor surface trap in this charge transfer element is the numerical value mentioned above, and the distance that the ultrasonic wave travels in that time is approximately 12μ, so the charge transfer in this example There is almost no transfer loss due to traps in the device.

Hereinafter, various structures will be explained, but the manufacturing method and detailed structure will be omitted from the drawings, and only the main parts will be illustrated and explained.

Embodiment 2 A channel stop 24 is provided in a semiconductor 10, as shown in a cross-sectional view and a plan view in FIGS. 17a and 17b. The channel stop 24 in this case is, for example, a p-Si semiconductor 10 in which p ⁺ -Si is diffused surrounding the transfer channel. Such channel stops may be created by ion implantation. In the case of this embodiment, the structure is not only simpler than that of the first embodiment, but also the effect is great because the escape of charge carriers is prevented without using the insulating layer 18.

Embodiment 3 A channel stop structure having a transverse cross-sectional shape of a transfer channel portion as shown in FIG. 18 may be formed. The thickness of the insulator layer 18 outside the transfer channel is increased. Even if the conductor layer 29 is wider than the transfer channel width, the effect of the electric field is smaller outside the transfer channel, so that a shallower potential than the potential in the transfer channel prevents charge carriers from deviating from the transfer channel. A method for forming such an insulator layer is, for example, to first form an insulator layer uniformly by thermal oxidation, then etch the transfer channel portion, and then form an insulator layer again by thermally oxidizing the entire surface. Bye.

Hereinafter, descriptions of channels, stops, etc. may be omitted to avoid complexity.

Example 4 When the piezoelectric material is highly insulating, for example, AlN
When this is used, the cross-sectional structure of the transfer channel portion of the charge transfer device of the present invention can be obtained by removing the insulating layer as shown in FIG. At this time, the channel stop may be provided in the piezoelectric layer 27 as in the first embodiment, or may be provided in the semiconductor 10 as in the second embodiment. The feature of this structure is that the structure and manufacturing are simple because an insulating layer is not required.
For example, such a piezoelectric material has a specific resistance.
10 ¹⁰ Ωcm or more is preferable.

Example 5 When the piezoelectric material has low insulation, for example, 10 ¹⁰ Ω
cm or less, for example, when CdS is used, the cross-sectional structure of the transfer section of the charge transfer element of the present invention is as shown in FIG.
8' is preferably provided. With this configuration, it is possible to create a charge transfer element using a piezoelectric material with low insulation properties.

Example 6 As shown in FIG. 21, n ⁺ − is formed on the surface of p-Si10
It is embedded with Si10'. This configuration has the advantage that the ultrasonic bunching potential is small because the mobility of the charge carriers is high, and although the charge carriers move more inside than on the surface, the trap density is higher inside than on the surface. This has the advantage that there is less transfer loss due to trapping of charge carriers. Note that the configurations of the piezoelectric layer and the insulating layer may be other than those shown in FIG. 21.

Example 7 As shown in FIGS. 22a and 22b, the piezoelectric layer 2
7 does not need to be entirely on the insulator layer 18, but may be on the charge carrier transfer channel. Further, if the piezoelectric layer 27 is used only on the transfer channel as shown in FIG. 22c (the side view thereof is shown in FIG. 22a, for example), the ultrasonic waves can be confined therein, so that a stronger potential can be obtained. It can occur in transfer channels in semiconductors.

Embodiment 8 The transfer channel may be curved as shown in FIG. In this case, the piezoelectric layer is preferably on the transfer channel, with a channel stop around it.

Note that the width of the transfer channel does not need to be constant in the direction of propagation of the ultrasonic waves, and may be widened or narrowed. In that case, the transferred charge carriers are dispersed or concentrated.

Further, the transfer channel can be divided in the traveling direction of the surface acoustic waves, or can be merged with another channel.

Embodiment 9 When the transfer channel 29 is curved as shown in FIG.
Not only for putting it in 9 and transferring it,
In order to effectively transfer the charge carriers after being bent and transferred, a further comb-shaped electrode 28' may be provided.

Embodiment 10 Figure 25a shows the injection section 1 in the lateral direction of the transfer channel.
6 and a detection section 17 are provided. Any number of these injection units or detection units can be provided at any location. Of course, the injection part or the detection part can be placed before or after the transfer channel.

An advantage of having an injection part or the like in the lateral direction of this transfer channel is that the ultrasonic wave is less attenuated by the injection part or the like. It also has the effect of reducing noise due to piezoelectric potential during detection.

Furthermore, since the charge carriers are effectively detected by the detection section provided in the lateral direction, it is also effective to make the transfer channel thinner in the region of the transfer channel that is in contact with the detection section. For example, as shown in FIG. 25b, the transfer channel is tapered in the area that is in contact with the detection section, so the transferred charge carriers proceed in the direction of the detection section. The shape of such a transfer channel can be freely designed depending on the purpose.

Embodiment 11 As shown in FIG. 26, a plurality of (for example, five in the figure) transfer channels may be provided. In this case, each transfer channel is separated from one another by a channel stop.

Example 12 As shown in FIG. 27, when the semiconductor 10 is a piezoelectric material 27, the semiconductor (piezoelectric material) 1
An insulating layer 18 is provided on top of the conductive layer 29
All you have to do is form.

Embodiment 13 By implanting ions into the insulator layer 18 or the piezoelectric layer 27 as shown in FIG. 28, the vicinity of the surface of the semiconductor 10 relative to the ion implantation portion can be depleted. For example, when using a p-type semiconductor, this can be achieved by implanting positive ions into the surface of the insulator layer 18 or the piezoelectric layer 27a.

Example 14 As shown in FIG. 29, the semiconductor 10 is a piezoelectric material, and the insulating layer 18 of the semiconductor 10 is ion-implanted as in Example 13.
It is also possible to deplete a region close to . In this case, the layer structure is extremely simple.

[Brief explanation of the drawing]

1 to 15 are cross-sectional views for explaining a method of manufacturing a charge transfer device according to an embodiment of the present invention.
FIGS. 16a and 16b are a cross-sectional view and a plan view of an example of a charge transfer element. FIGS. 17a and 17b are a sectional view and a plan view of another example of a charge transfer element. FIG. 18 is a lateral cross-sectional view of another example of a charge transfer element. Figure 19~
FIG. 21 is a cross-sectional view of another example of the charge transfer device,
Figures 2a and 2b are a cross-sectional view and a plan view of another example of a charge transfer element. FIGS. 23 to 26 are plan views of other examples of charge transfer elements. FIGS. 27 to 29 are cross-sectional views of other examples of charge transfer devices. In the drawings, 10 is a semiconductor, 16 is an injection part,
17 is a detection part, 18 is an insulator layer, 24 is a channel stop, 25 is an injection gate, 26 is a detection gate, 27 is a piezoelectric material, 28 is a comb-shaped electrode, 29 is a conductor layer, 33 is associated with ultrasonic waves A potential wave, 34 is a charge carrier.

Claims (1)

[Claims] 1. An ultrasonic wave generator for propagating ultrasonic waves to a piezoelectric layer, and bunching charge carriers into a potential valley along with the ultrasonic waves to transfer them in a depletion layer near the semiconductor surface. In a monolithic charge transfer element having a transfer channel for detecting the transferred charge carriers, a detection section for the transferred charge carriers, and a channel stop for preventing the charge carriers from being removed from the transfer channel, the bunching is configured to detect the transferred charge carriers. It is characterized by giving the transfer channel a potential greater than the sum of the potential due to the space charge it creates, the potential due to the diffusion effect of transferred charge carriers, and the potential for bunching one charge carrier (test charge). Charge transfer device. 2. The charge transfer device according to claim 1, wherein the time for the ultrasonic wave to travel the distance of one wavelength is shorter than the statistical time for the charge carrier to fall into the surface trap of the semiconductor. .