JPS5817671A - Charge transfer device - Google Patents

Abstract

PURPOSE:To nullify the level of a harmful boundary surface between a semiconductor and an insulating layer formed onto the semiconductor, to increase the quantity of charges treated and to reduce the loss of the transfer charges by disposing metallic layers isolated and shaped at positions corresponding to transfer electrodes and on the boundary surface between the semiconductor and the insulating layer. CONSTITUTION:A metallic film made of a metal such as aluminum is formed onto the whole surface of a P type silicon substrate 1 in a surface channel type CCD through evaporation in thickness such as approximately 2,000Angstrom , sections corresponding to the positions of the transfer electrodes are left, other excessive sections are removed through etching, etc. and a plurality of the metallic layers 5 mutually isolated are formed. The insulating layer 6 is molded onto the metallic layers through the method of evaporation or CVD or the like, and the transfer electrodes 3 are arranged at positions corresponding to the metallic layers 5 on the insulating layer 6 by a substance such as silicon.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to charge transfer devices such as charge coupled devices (CODs). Charge-coupled charge transfer devices are
Technical Journal, April 1-70 issue (Vol. 4, No. 4), page ss, KW@8 Boyle and G-D-
Smith announced it as a ``charge-coupled semiconductor device.'' This is the potential well F formed in Table 11 of the semiconductor.
``The electric charge is accumulated in a potential well, and the accumulated electric charge is moved along the semiconductor surface using a pulse voltage.The element configuration is an MO8 capacitor, and the electric charge of a large number of semiconductors is Usually, these are minority carriers accumulated at the interface between the silicon semiconductor and the oxide layer (8tOt) of an MO8 capacitor, and these are the
Therefore, it is transported along the boundary [K] with the oxide layer.

Such a 5-1 charge transfer device is called a surface channel charge coupled device (1!IccD). This 8CCD has 1
．． There are five problems in that the loss of transferred charges is extremely large due to the trapping of carriers by the surface states of the interface, and the transfer efficiency is poor.

A conventional example of such a surface channel type COD is shown in FIG. This ICCD has a structure in which an oxide layer (eg, 1io, etc.) is formed on a de-shaped silicon substrate 1, and transfer electrodes 1 are individually arranged on this oxide layer 1. The structure of each cell is an MO1e structure, which is also generally referred to as a metal-insulated semiconductor MIII structure.

For the transfer electrode 1, polysilicon or the like is used in addition to a metal film. A band model of the device when a positive voltage is applied to the transfer electrode 1 is shown in FIG.

FIG. 3(1) shows the state immediately after the voltage v□ is applied to the transfer electrode IK, and is a case where there is no signal charge or accumulated charge. A depletion layer, a so-called potential well, is formed on the surface of the silicon semiconductor substrate l. When signal charges begin to accumulate in this potential well F', these charges are collected at the lower potential boundary 1[K, as shown in FIG. 19(b). A portion of the signal charge is captured by the level at the interface.

The method of transmitting the signal charge under one transfer electrode to the next transfer electrode is based on a known method (for example, multiphase drive pulse), but it is also possible to form a deeper potential well under the next transfer electrode and pour the charge one after another. Use it for. During this charge transfer, the charges trapped in the deep units of the interface are transferred to the transfer electrode 1.
Transfer efficiency deteriorates due to charge trapping caused by This transfer efficiency is generally *S, *S%111
It is L. As a method to avoid such drawbacks, there is a method called the zero-zero method, in which charges corresponding to the horizontal level of the boundary surface are injected in advance as bias charges. Although this method improves the transfer efficiency, it has the disadvantage that the bias charge typically requires 16-40% 11Ilt of the saturation level, which reduces the maximum amount of transfer charge and therefore reduces the dynamic range. arise anew.

On the other hand, a buried channel charge transfer device (BCOD), which transfers charge in the bulk of a semiconductor, is described in the Bell System Technical Journal by 8. H. Waldey et al. Published in September issue of 1-Tego (Vol. 51, No. 1), page 16, $5. In this BCOD, the surface of the semiconductor is formed to have a conductivity type opposite to that of the substrate, and charges are carried in a channel remote from the interface with the oxide layer. However, although this BCOD solves the problem of the interlayer of the 8CCD described above, it has a drawback that the transfer charge is generally less than t618ccD. When the amount of charge is increased, the channel type transitions from the filled channel type to the surface channel type, and trapping at the interface becomes a problem again.

The 35th conventional example of such a salted channel type COD
IIIC shown. In the case of transferring electrons, the BCOD uses an r-type silicon substrate 1 and forms the N-type silicon layer 4 in the same way as the above-mentioned 8CCD. After forming the oxide layer 2 on top of the N-type silicon layer 4, the transfer electrode 1 is arranged separately.

FIG. 4 is a band model diagram when the transfer electrode JVC voltage vO of the BCCD is applied, and FIG. 4 (1) shows a state where no signal charge is accumulated immediately after the voltage is applied. If there is a signal charge, it is collected at the minimum potential portion in FIG. 4 (crow), so the signal charge is accumulated away from the boundary surface. When the amount of signal charge accumulated is large (as shown in Fig. 4(b) K, the flat part of the potential spreads, and there is no barrier to electrons at the boundary surface 11. From this point on, the BCCD becomes the same as the 8CCD described above. This results in a loss of signal charge depending on the unit of the boundary surface.For this reason, the normal B
In COD, charge is handled within a range where there is no influence from the unit of the boundary surface, and therefore, in the charge transfer operation, there is no loss due to the level of the boundary surface, so the transfer efficiency is 9919% 1 degree. Under such conditions, the amount of charge accumulation is
While ccB uses up to oxide layer I, 11CCD
"C is only the capacitance of the internal channel, and the difference is the difference in the amount of charge handled between the two, and the general K BCO
D has less professional charge than IIccD.

As mentioned above! 5K 8COD, ICCD #iC
It has problems due to harmful levels in the lump boundary [K].

The present invention was made in view of the above-mentioned circumstances, and the metal layer formed at the position corresponding to the transfer electrode and separated from the semiconductor and the insulating layer formed on the semiconductor by ξ Therefore, it is an object of the present invention to provide a charge transfer element that can unify harmful interface levels, increase the amount of charge to be handled, and reduce the loss of transferred charges.

Hereinafter, one embodiment of the present invention will be described with reference to the drawings. FIG. 2 shows the charge transfer device of the present invention, which is applied to a surface channel type CCD. In Jin's charge transfer device, first, a metal film such as aluminum is deposited on the entire surface of a P-type silicon substrate l to a thickness of about 000A, and then a portion is left to be placed at the transfer electrode position. Then, other excess portions are removed by etching capes to form a plurality of metal layers I separated from each other. An insulating layer C is formed on this by a method such as vapor deposition or CVN (chemical vapor deposition), and furthermore, on this insulating layer C, a transfer electrode Jv is provided at a position corresponding to the hooked metal layer 5 using polysilicon, etc. do. ζHere, the above-mentioned insulating layer 6 is used not only to oxidize silicon but also to generally refer to an insulating film formed by vapor deposition, CVD, or the like.

When a voltage is applied to the transfer electrode 3 of the charge transfer element, the potential distribution of the band model is as shown in Fig. 6.
become. Since the metal layer 5 is not applied with an external voltage, it is in a floating state, and the band sag exhibits the same tendency as the first intention. Since the signal charges are collected at the minimum point of the potential, they are collected at the surface of the silicon substrate l, as shown in Figure 2, but a metal layer S is formed between this part and the insulating layer. , charges are incorporated into this metal layer I. In the present device, the deep unit at the boundary surface, which has conventionally had a harmful interaction, is electronically buried by the metal layer #, so that it loses the function of capturing and releasing signal charges, and is rendered harmless. Therefore, during charge transfer, the charges move from the metal layer I in which charges are accumulated to the next metal layer S separated from it through a short interface between the silicon substrate 1 and the insulating layer, in the same way as in a normal surface channel type COD.

Here, 7 Hermi levels of metal layer I and silicon substrate 1
If there is no large energy gap between the conductive bands of is carried out, and the metal layer IK is collected again in the transferred cell I. Therefore, in the above device, the adverse effect due to the unit of the boundary surface can be almost ignored within each cell.

Therefore, the loss of transferred charges disappears in proportion to the area occupied by the separated metal layer l, and the transfer efficiency becomes extremely high.This device improves the transfer efficiency of the conventional 8CCD to the high value of BCOD. . In addition, since the amount of bias charge per unit of the transfer portion between each cell is sufficient, the amount of signal charge that can be handled increases accordingly. It shows. In this device, when the metal layer I is disposed in the form of being embedded in the silicon substrate I, there is a difference of 41%. This can be done by forming thin irregularities on the surface of the silicon substrate in advance, or by applying metal to the silicon substrate through a heat process.
5 to form a separated shape in the silicon substrate l as shown in the figure. This element also operates in the same way as the first embodiment and has the same effects.

FIG. 8 shows a charge transfer device according to a third embodiment of the present invention. This case is an example in which an embedded channel type CCDK is applied. That is, first, an N-type silicon layer 4 is formed on the surface of an R-type silicon substrate l. After applying a metal film over the entire surface of the N-type silicon layer 4 by vapor deposition or the like in the same manner as in the first embodiment described above, the metal layer is separated by leaving a part corresponding to the position of the transfer electrode 1 and removing the other part. Provide l. Furthermore, after forming an insulating layer C on top of this by a method such as vapor deposition or CVD, a transfer electrode 1 is provided for 5K.

sg is the band state t* when a voltage is applied to the element in FIG.
q, and No. 9 II (a) shows the situation immediately after applying the positive voltage vB to the transfer electrode IK.

Since the metal layer 2 is electrically floating, the shape of the band is similar to that shown in FIG. 4(1) described above. When signal charges are accumulated from this state, they are initially collected at the minimum potential within the semiconductor bulk (inside the silicon layer), and the potential of that portion becomes flat. Further, as charges are accumulated, a state is reached where the buried channel type transitions to the surface channel type as shown in Figure 3). When the signal charge is accumulated, the charge flows into the metal layer i because there is no barrier (I) in the bulk (inside the semiconductor).As mentioned above, the boundary between the insulating layer and the silicon layer 4 Since the deep unit has already been placed and rendered harmless by the metal layer 5, signal charge is accumulated in the same manner as a surface channel type COD without loss of charge. In addition, the metal layer SK also shows an accumulated state.

The conventional BCOD transfer method can be applied as is to transfer between the cells of the element. In this case, for one electrode, Figure 9 (
From the state of e), the applied voltage is applied in the negative direction, and the next transfer electrode 1 to which the charge is transferred will be in the state shown in the same figure (Hatatake), so a part of the signal charge is transferred from the metal layer I to the next metal layer. The charge is transferred through the boundary between the insulating layer 6 and the silicon layer 4 between the cells in the layer 1, and other charges are propagated through the bulk in the same manner as in a normal BCOD. When the signal charge is small, charge transfer is performed only through the bulk.

By arranging a separate metal layer S in a part of the BCOD as in the third embodiment, it is possible to effectively utilize the region from the salt channel operation to the surface channel operation.
It becomes possible to improve the amount of charge handled compared to the conventional method, and expand the dynamic range.

Note that the metal layer 1 shown in FIG. 1 may be formed to have a structure embedded in a silicon layer as shown in FIG. 1 described above. Also,
In FIG. 9, the contact state between the separated metal layer i and the N-type silicon layer 4 is shown as neutral, where no barrier can be formed, but this contact state may of course be ohmic. . Further, even if a shut-off barrier is formed, only the barrier portion does not contribute to the accumulation of signal charges, and the other operations are the same as those described in FIG. 9.

The effects of the present invention are achieved because the metal layer 5 is separated and interposed between the insulating layer 6 and the semiconductor substrate 1 or the N-type silicon layer 4, and the metal layer 1 is electrically floating. Therefore, the adverse effect of the interface unit, which has been a problem in conventional CCDs, can be eliminated by the element having the structure of the present invention. Further, another advantage of placing such an interface unit by the function of the metal layer I is that the dark current caused by the interface state can be reduced.

In this way, since the charge transfer loss can be almost ignored and the dark current can be considered, it is possible to form a long-distance cell 1 as shown in FIG. 10 as a modification of the device shown in FIG. 5, for example. I can do it.

When the transfer gate electrode is lengthened in the conventional configuration, there is an effect on the interface unit in proportion to the increase in area, but in the configuration of the present invention, the cell can be expanded to any length and stack. Although FIG. 10 shows an example in which the wire is elongated in one direction, the structure of the present invention allows it to take any shape, which has the advantage of allowing JIK to change the transfer direction, which was difficult in the past. Furthermore, connections to other transfer lines in the transfer am, transfer in the opposite direction, etc. can be freely designed without requiring complicated transfer electrodes.

Note that the material of the metal layer I, which is arranged separately at the interface between the semiconductor substrate l or the N-type silicon layer and the insulating layer C, is not only aluminum (M1) but also gold (M1), #I! (
Mug), platinum (Pi), tungsten (W), siliconium (Zr), hafnium (Hf), molybdenum (
Most metal materials such as Mo), nickel (N1), and palladium (Pd) can be used. This metal layer 5 may be a composite layer made of multiple materials, or it may be made of materials such as not only silicon (8N) but also lium arsenide (GaAB) and indium antimony (In8b) as a semiconductor. Any material can be used as long as it generates charge and can perform a charge transfer operation. Sarak,
The conductivity types of semiconductors include P-type surface channel, P-
It can also be applied to N-structured buried channel channels. Moreover, as an insulating layer, 810. Others 11c8i0 and AJ
The insulating layer may be formed from a composite layer of a plurality of materials, or may be formed by step oxidation. Further, the metal layer 5 described above may be provided under all the transfer electrodes of the charge transfer element, or may be provided under a necessary part of the transfer electrodes.

As explained above, according to the present invention, a separate metal layer is provided at a position corresponding to the transfer electrode and at the interface between the semiconductor and the insulating layer formed on the semiconductor, thereby preventing harmful It is possible to provide a charge transfer element that can reduce the adverse effects of boundary surface units, increase the amount of charge that can be handled, reduce loss of transferred charges, and improve transfer efficiency.

[Brief explanation of the drawing]

Fig. 1 is a cross-sectional view of a conventional surface channel type COD, and Fig. ** is a band model diagram for explaining the operation of Fig. 1.
Fig. 3 is a cross-sectional view of a conventional salted channel type COD, Fig. 4E is a self-made band model diagram for explanation of Fig. 3,
FIG. S shows a surface channel shape C according to the first embodiment of the present invention.
A structural diagram of a charge transfer element used in conjunction with OD, #1I11 is a band model diagram for explaining the operation of Fig. 5, and Fig. 7 is a structural diagram of a charge transfer element according to a second embodiment of the present invention.
No. ** is a structural diagram of a charge transfer device applied to a built-in channel type COD according to the third embodiment of the present invention, FIG. 9 is a band model diagram for explaining the operation of the device in FIG. FIG. 10 is a structural diagram of a charge transfer element according to a modification of the present invention. l... silicon substrate%j"' oxide layer, S... transfer electrode, 4... N-type silicon layer, 5... metal layer, 6
...Insulating layer, 1...Cell. Applicant's agent Patent attorney Takehiko Suzue If! Figure 5 Figure 7 Figure 8

Claims (1)

[Claims] +1) A semiconductor of a first conductivity type, a plurality of metal layers formed separately on the surface of this semiconductor, an insulating layer formed on these metal layers and the semiconductor, and this insulating layer. and a transfer electrode disposed at a position corresponding to each of the metal layers and to which a pulse voltage for charge transfer is applied, and a unit harmful to charge transfer generated at the interface between the semiconductor and the insulating layer. A charge transfer element characterized by being made ineffective. If a semiconductor layer of a second conductivity type is provided on the semiconductor layer of the first conductivity type, and the metal layer is provided on the semiconductor layer of the second conductivity type, it becomes v41I mold.
The charge transfer device described in Section 1. @ According to any one of claims 1 and 2, wherein the metal layer is formed in the semiconductor layer of the first conductivity type or in the semiconductor layer of the third conductivity type. The charge transfer device described above. (41) The unit cell composed of the first conductivity type semiconductor, the second conductivity type conductor layer, the insulating layer, the metal layer, and the transfer electrode is set to have an arbitrary direction and length. A charge transfer device according to any one of claims 1 to 3.