DE2240249C3 - Charge coupled semiconductor device - Google Patents

Description

Provide electrode constructions that are asymmetrical Generate potential wells in the substrate, thus ensuring unidirectional charge propagation will.

The in F i g. 1 shown electrode structure ergjbi such asymmetrical potential wells over the substrate 6. As already mentioned, the electrodes are arranged in pairs. Namely, the electrode is 12 of each pair via a switch (schematically at

has, which only takes a few seconds or less, the voltage source - V can be switched off and the charge remains stored.

After the charge has accumulated in the insulation layer in the manner described, are the switch 14 switched to its second position. As a result, all of the electrodes 12 each directly with the corresponding electrode IC

Embodiments of the invention are shown in the drawing. It shows

F i g. 1 shows a cross section of a charge coupled semiconductor element according to an embodiment the invention,

F i g. 2 shows the diagrammatic representation of the course of the operation of the semiconductor component according to F i g. 1 two-phase voltage used,

Fig. 3 shows a representation of occurring in the operation of the semiconductor component according to Fig. 1 poten- io 14-1, 14-2 and 14-3 indicated), with a source tialwannen, negative voltage - V, which is about 40 volts (for a

F ic ·. 4, 5 and 6 cross-sectional views of an- layer of 40 Å) may be connected. Be ^; their embodiments of the invention, application of this negative voltage together

F i g. 7 shows an illustration of the operation of the semi-detached under each electrode 12 on the silicon conductor component: es According to Fig. 6 occurring potential 15 dioxide-silicon nitride interface a positive latial wells. * application. After this charge has accumulated

The in Fig. 1 consists of a substrate 6 made of p-silicon with an insulation layer on one surface in the form of a very thin (20 to 50 Å) layer 8 made of SiIi- 20 ciumdioxyd (SiO.,) And a relatively thicker (about 400 to 80 "0 A) Layer 9 made of silicon nitride (Si ₃ N ₁ ). In transistor technology, such multilayer arrangements are known as MNOS arrangements

(MNOS = metal nitride oxide semiconductor) known. _i5 connected. Now also be; Absence of a pair of multiphase voltages under each pair of electrodes arranged on the silicon nitride surface electrodes 10-1, 12-1; 10-2, 12-2 creates an asymmetrical potential well. Because of etc. the accumulated under the 12 electrode of each pair

Before the mode of operation of the semiconductor component is charged, the potential well is in the substrate area element according to Fig. 1 is received, should briefly 30 below this electrode deeper than in the substrate area the general theory of the mechanism of action under the associated electrode 10. If a charge coupled circuits are considered. Electrode pair 12, 10 with a positive voltage If a positive chip is applied to an electrode such as 12-1, the potential well below the voltage pulse applied, then forms in the immediate electrode 12 as well as the potential well below the part of the p-type silicon 35 of the electrode 10 located below this electrode is lower; however, remains because of the zucium substrate a so-called deep depletion electric field generated by the permanent area. That is, the applied positive voltage is generated which stored positive charge pulse pushes majority carriers, in the case of a p-channel asymmetry.

holes in the substrate, from the substrate surface

just below the electrode 12-1 away. As a result, 40 I- i g. 1 is illustrated graphically in FIGS. 2 and 3, there is a potential on the surface of the p-type silicon. The signals Φ, and Φ The depth of the potential well is the square point t ₀ generates the voltage Φ ,, which is relatively positively proportional to the depth of the depletion area.

A material well on the surface of the silicon substrate in the semiconductor substrate under the electrical element The potential well formed strives to trap minority carrier pairs 10-1, 12-1, as in the upper part of FIG. 3 (in this case electrons) to accumulate or represented. The charge signal, in this case accumulate; n. If they are not formed by electrons from another, it will run along the surface Source are available, the minority carriers of the silicon substrate come under the "channel oxide" (dei from the substrate itself. As with those in the initially 50 thin silicon dioxide layer). To the aforementioned Patent application described half-refinement purposes is, however, the charge signal through Ladder components can, however, symbolize the minority cross, which symbolizes the lowering of the charges initially from a source electrode have a surface potential due to the presence of the which are embedded in the p-type silicon and display the charge signal. Since the potential tub be arranged at one end of the silicon substrate 55 under the electrode 12-1 (because of the in the oxide subtei can, or the minority charges can be raised by 12-1 stored fixed positive charge) light or other funds are brought in.

The stored in the substrate under an elecrode
Charge can be made with the help of multiphase voltages
into the substrate area under the next electrode 60
be transmitted. As described in the cited patent application, when using
three or more phases a unidirectional (in only
one direction) charge propagation. However, it is τη
Interest in simplifying construction wiin- 65 couple 10-1, 12-1. The shallower potential well at 10-2 is worth seeing that a charge-coupled half generates a potential threshold or barrier for the ladder-orientated with a two-phase drive voltage forward flow (to the right,! From iiung stored at 1.8. To do this, however, you have to operate a special charge and for the reflux (to the left) of the under

Borrowed is deeper than under the electrode 10-1, this negative charge collects essentially entirely in the substrate area under the electrode 12-1. At time t ₀ , OVoH is applied to the next adjacent pair of clectrodes 10-2, 12-2. The asymmetrical potential well formed under this pair of electrodes is therefore relatively shallower than the potential well under the electrode

12-2 stored charge. (Although 0 volts are applied to this electrode, the potential can is shown with a certain finite depth. This can also be interpreted to mean that there is a fixed but relatively low reverse or reverse voltage on the substrate, as explained in the patent application mentioned.) Likewise the asymmetrical potential well under the pair of electrodes 10-0, 12-0 (not shown in FIG. 1) has the same shape as the potential well under 10-2, 12-2. The potential well under the electrode 12-0 is shallower than the potential well under the electrode 10-1. This requirement is necessary to ensure a unidirectional flow of the charge signal (ie the electrons) from left to right. ¹ p

At the time f, the voltage Φ _{2 is} already positive by +10 volts. The voltage Φ switches from its positive value of + 10 volts to 0 volts. As a result of these voltage relationships, a transfer of negative charge takes place, as in the middle part of FIG. 3 indicated. The asymmetrical potential well under the pair of electrodes 10-2, 12-2 is now at its greatest depth. The asymmetrical potential well under the pair of electrodes 10-1, 12-1 becomes shallower. When the potential well under electrode 12-1 becomes shallower than the potential well under electrode 10-2, the charge stored under the first-mentioned electrode "falls" into the potential well under the last-mentioned electrode. At this point, however, the potential well is under electrode 10-2, so the charge continues to "run down" until it comes to rest in the potential well under electrode 12-2. Under the influence of the self-induced drift field, the thermal diffusion and the stray field, the charge is essentially completely transferred within nanoseconds from the potential well under the electrode 12-1 to the potential well under the electrode 12-2 .

Shortly after time r, all of the charge previously stored on electrode 12-1 is stored on electrode 12-2 . Considerably later, at time f (g last part of F i. 3) _2, the charge remains in the potential well under the electrode 12-2, as shown. Shortly before time I ₃ , this charge would begin to be transferred to the next pair of electrodes 10-3, 12-3 .

F i g. 4 shows a preferred electrode design according to an exemplary embodiment of the invention. A method for producing this general type of construction is described in detail in the cited patent application, although there the insulating layer consists of only a single material. The positive charge shown can therefore initially be introduced without the need for switches such as 14 in FIG. 1. All that is necessary is to connect all, - and lines together to a relative voltage source - V and then switch off the voltage source and put the semiconductor component into operation in the manner already explained. (The same applies to the components according to FIGS. 5 and 6.) As in the exemplary embodiment P «i F i g. 1 Consign the positive charges once accumulated at the silicon nitride-silicon dioxide interface there, provided that the positive phase voltages are always kept considerably smaller than the initial negative injection pulse (by a factor of e.g. 3 or 4).

Instead of, as shown, with a p-substrate, the component according to FIG. 4 also with a substrate be made of n-silicon. In this case, fixed negative charges may initially be below the Aluminum electrodes by briefly applying a positive voltage to all electrodes get saved. The polysilicon electrodes, which in the present case from η-Tyρ are either p-type or η-type using either the n-conductive or p-conductive Silicon substrate.

The mode of operation of the two-phase described here Charge-coupled semiconductor component is based on the permanent storage of different Quantities or different polarities of charge in the channel oxide under two adjacent electrodes, which are subjected to the same phase voltage. One embodiment is preferred with overlapping or overlapping electrodes. These can be polysilicon electrodes act with overlapping aluminum electrodes as shown in FIG. 4, 5 and 6 shown. Instead of this However, both electrodes of the individual pairs made of polysilicon or another can also be difficult consist of fusible material.

While in the embodiment of F i g.4 the polarity of the charge is chosen so that the generated The potential well under one electrode is deeper than under the other electrode of the pairs, one can also choose the polarity so that a potential threshold or barrier under one electrode of the couple arises. A corresponding exemplary embodiment is illustrated with reference to FIGS. 6 explained.

F i g. 5 illustrates another embodiment of the invention. The substrate here is of the η-type instead of the p-type, and the insulation layer contains a relatively thick (1000 to 2000A) layer of aluminum oxide (Al ₂ O ₁₁ ) between the aluminum electrodes and the substrate and between the aluminum and polysilicon electrodes. The charge stored in the aluminum oxide under the aluminum electrodes is negative and creates deeper potential wells (more negative surface potential) in the substrate areas under the stored charges than in the substrate areas under the polysilicon electrodes. The silicon dioxide layer (approximately 40 Å thick) deliberately used in the manufacture of MNOS devices, as shown in FIG. 4 is not really needed when using alumina. In this case, the permanent storage of a negative charge in the z. B. in the article by S. Nakanuma et al: "A Read-Only Memory Using MAS-Transistors" in Digist of Technical Papers, 1970 International Solid State Circuit Conference, Philadelphia, Pa., Pp. 68 and 69, described manner be achieved.

The aluminum oxide of the component according to Fig.! but can instead also be carried out by pyrohydrolysis of AlCL (S. K. Tung and R. E. Caffney, Trans Met. Soc. AIME, 233, 572, 1968) or by Redu adorn with Al isopropoxide (M. T. Duffy um A.Revesz, J. Electrochemical Soc. 117,372,1969 be generated. In both cases, the Al.O. can be produced in such a way that a fixed negative

Charge already stored in it in a "natural" way that is enriched among the permanently negatively charged. The charge-always finite potential thresholds produced in this way are coupled two-phase semiconductor ribbon elements which can thus operate the complete charge transfer without preventing them between two potential wells. One of the electrodes, under which the negative charge must be applied to a voltage of the signal charge, in which the charge is to be stored in each potential, an initial positive feed-in trough. However, it can also have a comparable value, is desirable in these cases, when the amount of negative miniumoxyds initially built into the production of the aluminum effect of trapping or holding charge at the silicon-aluminum oxide interface leads to minimal charge for the not to malise given purposes. This can be at the Ausführungsbeireicht, the '.rforderliche additional negative charge to play F i g. 6 or an embodiment with the means already explained above (on (not shown) with p-substrate, silicon dioxide under a positive voltage for a given the aluminum electrodes and aluminum oxide under time interval), as in other things the polysilicon electrodes and with storage context also in the above-mentioned work 15 of relatively large negative charges in the by S. N aka η uma et al. Aluminum oxide under the polysilicon electrodes

During operation of the semiconductor component according to FIG. Reach 5.

multiphase voltages are used, the case of which is the case in the exemplary embodiments explained above

Values instead of 0 to + 10VoIt from OVoIt up to for each pair of electrodes in the insulation

a negative voltage value of, for example, 20 layers under one electrode charge permanently

- 10 full enough. The phase relationship between stored and under the other electrode is none

the voltages Φ, and Φ, is stored analogously to that after charging. Of course, other options are also

Fig. 2 and as in Fig. 13 of the several examples of management possible. For example, can also

Patent application. The one in the component after each pair of electrodes in the insulation layer

F i g. 5 each accumulating minority carrier 25 under one electrode negative and under the one

are of course holes instead of electrons. positive charge must be stored on the other electrode,

Since it is usually easy to only use the negative or those stored under the two electrodes Charge to be stored in the pyrolytic alumina- charges can be of different amounts, rather, determines the type of cable used, one of which may or may not be a zero. Subtrats, whether deep potential wells or potential- 30 What is important in each case is that the in the swells are generated. In the case of the component after the insulation layer under the two electrodes each F i g. 5, the substrate is η-conductive, so that deep pot pairs of stored charges are different tialwannen are generated, as already explained. At (either the amount or the sign), In the component according to FIG. 6, the substrate is p-line, thus the required asymmetry in the substrate tend, so that the negative charges in the aluminum 35 potential well is achieved. For the isooxide A potentialation layer can be swelled instead of, for example, under the aluminum electrodes generate against the flow of charge. With the given materials also other suitable matesers Arrangement you can use an extremely good material.

Achieve performance if you take into account the spacing The design of the electrodes (in plan)

between the polysilicon electrodes can relatively 40 the components described above

makes tight, for example 3 μπι, as in F i g. 6 shown. be similar to the patent application mentioned

The operation of charge coupled two is described. For example, a two-dimensional component operating as a shift phase semiconductor component, for example in a register, can be placed in shift registers using the aluminum components shown in FIG. 17 to 19 of said patent application oxide stored charge for the formation of energy 45 be designed in the manner shown. The invention has thresholds for the signal charge can also correspond to other possible uses in FIG - Signal channel from a relatively thin Schichi th energy thresholds is placed. This function of silicon dioxide or some other insulating material is illustrated in FIGS. 7a and 7b. Fig. 7a rial between thicker insulating layer areas. It shows the potential wells generated in the absence of S ₁ and <& ₂ voltage detection is also suitable for charge-coupled halves. 7b illustrates the case in which the channel regions are shown in which Φ ₂ becomes negative, while Φ, how relatively weakly doped substrate regions remain between in FIG. 7a .7c illustrates. Here the th are formed as recently described by Bell Telephone Laboratories maximum value of the phase voltages.

For this purpose 3 sheets of drawings

Claims (7)

Patent claims: 1. Charge-coupled semiconductor component with a substrate made of semiconductor material, on one surface of which an insulation layer is applied, on which there are several pairs of adjacent electrodes electrically coupled to the substrate, in which an asymmetrical potential well is created under the two electrodes of each pair, thereby characterized in that the asymmetrical potential wells are generated in that a fixed charge is stored in the insulation layer (8, 9) under one electrode (10-1 ...) of each pair, which is derived from the charge under the other electrode (12- 1 ...) each pair is different. 2. Charge-coupled semiconductor component according to claim 1, characterized in that ao the insulation layer of a relatively thin silicon dioxide layer (8) and a relatively thick silicon nitride layer (9) under at least ¹ ; one electrode of each pair and that the silicon oxide layer is located on the substrate (6) between the silicon nitride layer and the electrode or the two electrodes of each pair. 3. Charge-coupled semiconductor component according to claim 1, characterized in that the insulating layer of an aluminum oxide layer under at least one electrode each Consists of a pair of electrodes. 4. Charge-coupled semiconductor component according to claim 1, characterized in that between the one electrode (12-0 ...) of each pair and the substrate this electrode (12-0 ...) isolating from the substrate first insulation layer, in which a solid Charge is storable. is arranged and that between the other electrode (10-1 . ..) and the substrate, this electrode (10-1 ...) Is arranged from the substrate insulating second insulation layer of a different type (Figs. 4 to 6). 5. Charge-coupled semiconductor component according to claim 4, characterized in that the first insulation layer made of a relatively thin layer of silicon dioxide under a relatively thick silicon nitride layer consists of the silicon dioxide layer on the surface of the substrate and the silicon nitride layer between the silicon dioxide layer and the one electrode is arranged and that the second insulation layer consists of a relatively thick silicon dioxide layer (Fig. 4). fi. Charge-coupled semiconductor component according to Claim 4, characterized in that the first insulation layer made of an aluminum oxide layer and the second insulation layer a silicon dioxide layer (Figs. 5 and 6). 7. Charge-coupled semiconductor component according to claim 1, characterized in that one electrode of each pair is made of polysilicon. The invention relates to a charge-coupled semiconductor component with a substrate made of semiconductor material, on one surface of which there is an insulating layer on which several pairs are located of adjacent electrodes electrically coupled to the substrate, in the case of which An asymmetrical potential well is created between the two electrodes of each pair. In the patent application P 22 01 150.3 (German Offenlegungsschrift 2 201150), various Electrode arrangements for creating asymmetrical potential wells in the semiconductor substrate of a charge-coupled semiconductor component proposed, including arrangements with electrode daars, in which one electrode is closer to the substrate than the other, as well as arrangements with electrode pairs in which one electrode is permanently set to a different DC voltage is biased than the other. Important advantages of these arrangements are that a rapid charge propagation in one direction can be achieved with a two-phase drive voltage can, with the help of a simple and compact cell geometry. A semiconductor component of the type mentioned that works in two-phase operation and with that under each pair of electrodes an asymmetrical potential well either through different Distance between the two electrodes of the pair in front of the substrate or by different permanent ones Generate DC voltages of these two electrodes! is. is from "Electronics", Volume 44, June 21, 1971, p. 50 known to 59th. Because of the two-phase operation, in which the two electrodes of a pair aul different potential, their connecting lines must each lead out of the component will. In addition, especially when using a substrate with a high specific Resistance which separates one electrode from the substrate by an insulating layer that is so much thickci than that of the other electrode, so that constructional difficulties arise. From the German Offenlegungsschrift 1 803 03f it is also known in the case of a field effect transistor, ar the boundary between a silicon dioxide layer arranged on the substrate and one thereon ordered silicon nitride layer. on which the control electrode is located, a fixed charge zi to save. The object of the invention is to provide a semiconductor element for charge propagation in a given Specify direction, which can be constructed more simply than known components of this type with a; asymmetrical potential well. The invention solves this problem with a semiconductor component of the type mentioned above by that the asymmetrical potential wells are generated by that in the insulation layer below a fixed charge is stored on one electrode of each pair, and that of the charge below the other Electrode of each pair is different. The invention has the advantage that the two electrodes of an electrode pair directly with can be connected to one another and that the required asymmetry even with substrates mi high specific resistance with insulating layer ^ smaller differences in thickness than previously achievable is. so that the component is structurally simpler