US 3767983 A
The transfer efficiency of a bucket-brigade charge transfer device is improved by adapting the transfer regions between successive storage sites such that in each transfer region at least two distinct threshold voltages occur, with the greater threshold voltage being in the trailing portion of the transfer zone with respect to the direction of advance of information. Advantageously, the differences in threshold voltage in each transfer region are realized through differences in oxide thickness or differences in the doping of the transfer region.
Claims available in
Description (OCR text may contain errors)
United States Patent [1 1 Berglund Oct. 23, 1973  CHARGE TRANSFER DEVICE WITH 3,697,786 10/1972 Smith 317/235 IMPROVED TRANSFER EFFICIENCY  lnventor: Carl Neil Berglund, Ottawa, r mary Examiner-Jerry D. Craig Ontario, Canada Attorney-W. L. Keefauver  Assignee: Bell Telephone Laboratories,
Incorporated, Murray Hill, NJ.  ABSTRACT  Flled: 1972 The transfer efficiency of a bucket-brigade charge  Appl. No.1 282,962 'transfer device is improved by adapting the transfer regions between successive storage sites such that in each transfer region at least two distinct threshold  317/235 317/235 33 25 voltages occur, with the greater threshold voltage  Int Cl "on 11/14 being in the trailing portion of the transfer zone with  Field 221 respect to the direction of advance of information. "5 235 Advantageously, the differences in threshold voltage in each transfer region are realized through differ-  Reierences Cited ences in oxide thickness or differences in the doping of the transfer region.
PAIENTEDuuI 2a 1915 SHEET 2 0F 3 PATENIEMcvza ma 3.767.983
SHEET 3 [IF 3 FIG. 4A
CHARGE TRANSFER DEVICEWITH IMPROVED TRANSFER EFFICIENCY BACKGROUND OF THE INVENTION each zone is operated as a potential well, the boundary of which is defined by the PN junction which defines the zone. v
A conceptually convenient way to think of the bucket-brigade type of charge transfer devices is as a cascade of Insulated Gate Field Effect Transistors (IG- FETS) in which each of the surface zones serves both as the drain of a particular IGFET and as the source of the IGFET next succeeding. At any given time in operation, then, a pair of successive zones may be thought of as the source and drain, respectively, of an IGFET; and the transfer electrode there-overlying may be thought of as the gate electrode of the IGFET.
It is known that a significant factor in signal degradation in such devices is what has been termed dynamic drain conductance, a feedback effect resulting in an increase in effective distance between source and drain during a transfer process due to decreasein drain voltage as charge carriers are transferred thereinto. This effect is described by C. N. Berglund and R. 1. Strain in an article entitled, Fabrication and Performance Considerations of Charge Transfer Dynamic Shift Registers which appeared in the Bell System Technical Jamal. sl- .1. .-.31.M* 1972, P
A technique for significantly rbiuah'gtfieerrct of dynamic drain conductance has been proposed by F I... J'. Sangster in an article entitled, Integrated Bucket- Brigade Delay Line Using MOS Tetrodes which appeared in the Philips Technical Review, Vol. 31, 1970, page 266. Sangsters approach" involves doubling the number of IGFETS per bit of stored information and also involves use of a conduction path for biasing the added IGFETS in addition to the pair of conduction paths normally used for coupling clock signals to the bucket-brigadeelectrodes. Stated anotehr way, Sangsters approach involves the use of twice as many surface zones per bit of stored information and a 50 per cent increase in the number of conduction paths employed for causing orderly advance of information through the device. This approach is considered excessively complex for many applications and, in addition, involves a larger physical size per bit of stored information, both of which problems it is the object of this invention to alleviate.
SUMMARY OF THE INVENTION The instant invention is based ,on a realization first that Sangsters additional conduction path can be obviated, provided his added transistors are designed to have a threshold voltage different from the threshold voltage of the original charge transfer transistors, and is based on the further realization that, through use of the differing threshold voltages, Sangsters added surface zones also can be eliminated.
In accordance with this invention, then, a bucketbrigade structure includes first and second different threshold voltages in each transfer region between each pair of successive storage sites. The transfer region additionally is characterized by a substantially abrupt transition between the first and second threshold voltages and by the fact that the threshold voltage of the trailing portion of each transfer region with respect to the direction of information advance is greater than the threshold voltage in the leading portion of the transfer region.
This difference in threshold voltage with a substantially abrupt transition therebetween reduces the feedback of voltage from the transferee zone (drain) to the transferor zone (source) as effectively as Sangsters two-transistor-per-bit structure without the abovedescribed undue complexity thereof.
BRIEF DESCRIPTION OF THE DRAWING It is believed the invention will be better understood from the following more detailed description taken in conjunction with the accompanying drawing in which:
FIG. 1 is a cross-sectional view taken along the information channel of a portion of a basic prior art bucketbrigade device;
FIG. 2 is a cross-sectional view taken along the information channel of a bucket-brigade device in accordance with a first-to-be-described embodiment of this invention; A
FIG. 3 is a diagram depicting typical surface potentials in the structure of FIG. 2 with typical operating voltages applied; I
FIG. 4A is an expanded view of ,a particularly relevant portion of the structure of FIG. 2 7
FIG. 4B is a diagram depicting typical surface potentials occurring in typical operation of the structure of FIGS. 2 and 4A with an operating voltage applied; and
FIG. 5 is a cross-sectional view taken along the information channel of a portion of a bucket-brigade device in accordance with an alternate embodiment of this invention. .i:
It will be appreciated that, for simplicity and clarity of illustration and explanation, the figures have no necessarily been drawn to scale. I
DETAILED DESCRIPTION I With more specific reference now to the drawing, in FIG. 1 there is shown a cross-sectional view taken along the information channel of a portion 1 l of a basic prior art bucket-brigade device such as disclosed, for example, in U.S..-Pat. No. 3,660,697, issued May 2, 1972, to C. N. Berglund and H. J. Boll. As shown, portion 11 includes a storage medium, the bulk, 12, of which illustratively is of N-type semiconductivity, and over which there has been formed an insulating layer 13, typically silicon oxide. Over insulator 13 are disposed. a plurality of field plate electrodes 14M, 15M, 14N, and 15N, each being registered in one-to-one correspondence with a plurality of localized zones 18M, 19M, 18N, and 19N adjacent the surface of bulk portion 12. As further indicated in the drawing where the bulk portion I2 is of N-type semiconductivity, surface zones 18 and 19 are of heavily dpoed p-type semiconductivity. I
It will be noted in FIG. 1 that the electrodes and the localized zones are disposed in relation to each other such that each electrode extends over significantly more of the zone lying thereunder to the right than the zone lying thereunder to the left. More specifically, for example, electrode M overlies a significantly greater portion of zone 19M (thereunderlying to the right) than zone 18M (thereunderlying to the left).
A pair of conduction paths 16 and 17 are each connected to every second electrode, i.e., conduction path 16 is connected to electrodes 14, and conduction path 17 is connected to electrodes 15. In operation, twophase clock voltages V and V: are applied to electrodes 14 and 15 through conduction paths 16 and 17, respectively. As is known, voltages V, and V advantageously are of sufficient magnitude to maintain the semiconductive surface of portion 11 always in depletion so as to minimize the effects of surface states on the charge being transferred.
In one clock phase, for example when the magnitude of V, is greater than the magnitude of V zones 18, under electrodes 14, are driven to a much greater magnitude of surface potential than are zones 19, under electrodes 15. As a result of this deliberately built-in difference in surface potential resulting from the asymmetric disposition of electrodes with respect to the zones, mobile charge carriers which were resident in zones 19 prior to application of this instant-described clock phase are transferred from their respective zone 19 into the zone 18 immediately to the right in FIG. 1.
It will be appreciated from the foregoing description of the structure and operation that each pair of closest zones, e.g., 18M and 19M, 19M and 18N, and 18N and 19N, may be thought of as the source and drain of insulated gate field effect transistor (IGFET). In such context, then, the N-type surface portion between any pair of closest zones may be thought of as the channel of an IGFET.
However, it will be seen tha there are no drain and source electrodes maintaining voltages on the source and drain zones as the transfer of charge carriers proceeds. As a result, as charge carriers transfer into a zone, the surface potential in that zone, and, accordingly, the voltage of the zone itself, decreases in magnitudeas each successive mobile charge carrier enters thereinto. In a structure of the type depicted in FIG. 1, this decreasing potential produces an effective increase in length of the channel of the IGFET and, as such, makes it more difficult for further charges to be transferred. This effect is what has been referred to hereinabove as dynamic drain conductances.
This channel lengthening effect may be understood by considering that typically the voltages induced on zones 18 and 19 are of polarity sufficient to produce a reverse bias on the PN junctions associated with those zones. As a result, a depletion region extends from the zone in all directions, and, significantly, to the left in FIG. 1 into the channel region. The effect of this extension of the depletion region into the channel is an effective decrease in channel length. As is known for IG- FETS, a decrease in channel length produces a higher transconductance which, for a given applied voltage, enables easier transfer of charge carriers from the source to the drain. As the magnitude of the surface potential in the drains decreases, the depletion region decreases in width and effectively lengthens the channel, thus successively decreasing the transconductance and concomitantly making it successively more difficult for the remaining charges to transfer.
It will be readily appreciated that this effect increases with increasing charge to be transferred. Thus, for example, a relatively large packet of charge carriers, possibly representing a one, will be attenuated more than a relatively small packet of charge carriers representing a zero. Since the effect is not uniform, significant signal deterioration can result.
Because, as discussed hereinabove, the operation of the structure of FIG. 1 is not directly analogous to IGFET operation inasmuch as there is no direct electrical connection to the zones, and further because it is simply more convenient for the following description, use of the IGFET terminology will at this point be discontinued and, instead, the localized zones will be termed charge storage sites and the channel regions between zones along the surface will be termed transfer regions. Additional terminology which will be useful are the terms leading and trailing with respect to the direction of advance of mobile charge carriers representing signal information. Inasmuch as such direction is established by the built-in asymmetry in the structure of FIG. 1 to be that of transferring to the right, the rightmost portions of any particular feature will be termed the leading portions and the leftmost portions will be termed the trailing portions.
With this terminology in mind and with the foregoing description of the basic prior art bucket-brigade device and its problems in mind, attention is directed next to the structure illustrated in FIG. 2, a cross-sectional view taken along the information channel of a portion 21 of a bucket-brigade device in accordance with a first embodiment of this invention.-
As shown, portion 21 includes a semiconductive bulk portion 22 illustratively of N-type semiconductivity and including adjacent the surface thereof a plurality of P- type localized zones 28 and 29, like zones 18 and 19 in FIG. 1. Over the surface of bulk 22 and zones 28 and 29 is disposed an insulating layer 23 of nonuniform thickness; and a plurality of electrodes 24 and 25 are disposed over insulator 23. As can be seen, each electrode includes two distinct parts, the leading part being labeled with the suffix B and the trailingpart being labeled with the suffix A.
More specifically, a plurality of electrodes 24 and 25 are disposed in one-to-one correspondence with the plurality of localized zones 28 and 29. As can be seen, the leading portion 24MB of electrode 24 extends over a great percentage of zone 28M and also extends approximately halfway over the space between zone 28M and the P-type zone next preceding. The trailing portion, labeled 24MA, of electrode 24M extends essentially only over the trailing half of the space, i.e., the transfer region, between zone 28M and the zone next preceding.
Likewise, electrodes 25M, 24N, and 25N each include parts A and B analogous to the described parts A and B of electrode 24M.
In operation, with two-phase clock voltages V, and V, applied to electrodes 24 and 25 through conduction paths 26 and 27, a typical surface potential configuration (in the absence of mobile charge carriers representing signal information) is depicted schematically in the diagram of FIG. 3. In FIG. 3 the magnitude of surface potential in the structure of FIG. 2 is plotted as increasing in the downward direction with an arbitrary zero reference level assumed to be at the base of arrow 31. As will be appreciated, the horizontal dimension of the diagram of FIG. 3 has been made to align with the horizontal dimension of the structure of FIG. 2. As will be further appreciated, the diagram of FIGS assumes that the magnitude of voltage V; is greater than the magnitude of voltage V In operation, a voltage V, applied to electrode 24M produces under portion 24MA a first surface potential S1 which is of lesser magnitude than a second surface potential S2 under the trailing portion of that part of electrode 24MB which does not overlie zone 28M, the lesser magnitude being due to the greater spacing of portion 24MA from the semiconductor surface. With the semiconductivity type-as shown in FIG. 2, voltages V, and V typically will be negative soas to operate in the depletion mode; and, in this case, the negatively ionized acceptors in zone 28M tend to enhance the magnitude of the surface potential there in the negative direction.
In the other half-bit, that is, the transfer region and storage site 29M associated with electrode 25M, the surface potentials in FIG. 3 are drawn assuming that V, is sufficiently greater than V, such that the least magnitude of surface potential S4 caused by the trailing edge 25MA of electrode 25M is more attractive, i.e., of greater magnitude, than the surface potentialS2 associated with the leading portion of the transfer zone next preceding half-bit. More specifically, the surface potential caused by voltage V, applied to electrode 25M is seen to have three parts like that of the surface potential under electrode 24M, except that it is translated in a direction of increasing attractiveness for mobile carriers.
As illustrated, the surface potential, S3, of zone 28M is essentially equal to the surface potential, S4, of the trailing portion of the succeeding transfer zone, due to the copious quantity of mobile charge carriers in heavily doped zone 28M. As seen, the potential diagram repeats with two electrode periodicity, e.g., from the trailing edge of electrode 24M to the trailing edge of 24N. i I
At this point, it should be appreciated that within a given half-bit therelative potential differences will remain constant and are fixed by physical structure. More specifically, the difference between surface potential S2 and S1, assuming complete depletion in the region of S1, is caused solely by the difference in oxide thicknesses under electrode 24M. The abruptness of the transition from S1 to S2 is due to the abruptness in transition from the thicker to the thinner oxide portions. The difference in potential between S2 and S3 is due to the relative concentrations of donor atoms in the transfer region and acceptor atoms in storage site 28M. Of course, the same physical structure is or can be made to repeat under each electrode; and so the relative levels of the portions under electrode 25M and other electrodes will be the same as those under electrode 24M. Such relationship is depicted in FIG. 3.
With reference now to FIGS. 4A and 48, there is depicted in FIG. 4A an expanded view of a particularly relevant portion of the structure of FIG. 2, centered around a typical transfer region, in accordance with this invention. FIG. 4B is a diagram aligned with FIG. 4A and depicts typical surface potentials, S, occurring in typical operation of the structure of FIGS. 2 and 4A.
As can be seen, the structure of FIG. 4A is but an expanded view of that portion of FIG. 2 which includes the rightmost portion of zone 28M, the transfer region between zones 28M and 29M, the leftmost portion of zone 29M, and the insulator and electrode structure thereoverlying. As can be seen, the transfer region between zones 28M and 29M includes two parts, labeled respectively Tl and T2. T1 is that part, advantageously approximately half, of the transfer zone closest zone 28M and underlying the leftmost portion of electrode 25M that portion labeled 25MA. As seen, portion 25MA is disposed at a greater distance, i.e., over a greater insulator thickness, than is portion 25MB which overlies the other half, T2, of the transfer zone.
As illustrated in FIG. 4B, application of a voltage to electrode 25M produces thereunder a surface potential configuration in the transfer region which includes two distinct levels S4 and S5 in regions T1 and T2, respectively; and there is a relatively abrupt transition between the two levels of surface potential, due to the relatively abrupt transition between the two distinct insulator thicknesses. As has been mentioned hereinabove, both the existence of the two levels and the existence of the relatively abrupt transition therebetween are important to this invention for reasons which will be explained.
But first, it will be appreciated that the existence of the two distinct levels of surface potential in the transfer region of a structure such as depicted in FIG. 4A is due solely to the fact that electrode portion 25A is disposed over a thicker insulator region and therefore is spaced at a greater distance from a semiconductive surface than is electrode portion 25MB. As is known in the relevant art, this described difference in surface potential caused by a uniform voltage applied to electrode 25M results because the different spacing of the portions of 25M from the semiconductive surface produces thereunder a corresponding different threshold voltage. Thus, the structure of FIG. 4A may be described as having distinctly different threshold voltages associated with portions Tl and T2 and may be further described as having a relatively abrupt transition therebetween. Y 4 v With more particular reference now to FIG. 4B, the importance of the two distinct threshold levels with a relatively abrupt transition therebetween will be explained. First, it will be appreciated that the current, i.e., the rate of flow of charge carriers, through the transfer region for any given voltage applied to the overlying electrode will be limited by that portion of the transfer region having the greater threshold voltage because the portion of the surface thereunder is less attractive to mobile charge carriers. Thus, in FIGS. 4A and 4B the rate of current flow through the transfer region will be limited by the trailing portion T1 of the transfer region. Because the current flow will be so limited, variations in the length of the half-channel associated with region T2 due to variations in the depletion width, denoted XD2 in FIG. 4B, due, in turn, to variations in surface potential S6 caused by mobile charge carriers transferring into zone 29M, will have very little effect on the ease with which current can flow through the total transfer region. This, of course, is because the current flow is principally limited by half-channel T1 which is affected only in a second order way by voltage variations at S6.'
With the foregoing in mind, a few other salient characteristics of the structure and surface potential configuration depicted in FIGS. 4A and 413 will now be mentioned. First, the transition in surface potential between regions T1 and T2 has been described and is depicted in FIG. 4B as being relatively abrupt. The degree of abruptness advantageous for use in this invention does not lend itself to precise quantification. However, because of a difference between potentials S4 and $5, an electric field occurs in the transition region between those two potentials; and, because of this electric field, a depletion region, the width of which is denoted as XDl in FIG. 48, results. About all that can be said about the degree of abruptness desired in the transition between surface potentials S4 and S5 is that such transition be sufficiently abrupt to produce an electric field of sufficient strength to produce a depletion region XDl so that the concentration of mobile carriers in the channel region effectively goes to zero at the left edge of depletion region XDl with respect to the concentration of such carriers at the left edge of region T1, i.e., adjacent the right edge of zone 28M. It is only if this condition is met that the second order effect of variations in surface potential S6 on transconductance mentioned hereinabove is achieved.
The magnitude of the difference between potentials S4 and S5 in accordance with this invention also is of importance, but also is not amenable to precise quantification. However, it is believed this difference should be at least kT, which is about 26 X 10 volts at 300 Centigrade, and should be no greater than the magnitude of the difference between surface potentials S3 and S6 caused by applied voltages V, and V which latter difference typically will be of the order of 10 volts. In practice, it is believed a practical difference between surface potentials S4 and S5 is about 1 or 2 volts, which difference can be readily achieved with conveniently fabricatable differences in oxide thickness under the portions of the electrodes.
The total length of the channel region, i.e., the distance between zone 28M and 29M, preferably should be sufficiently great that the depletion region XB2 from zone 29M never extends completely to the right edge of depletion region XDl, even under the greatest expected applied operating voltages, for otherwise the second order effect desired from this invention will not be fully achieved. Also, the width of half-channel T1 should be sufficiently great that the left edge of depletion region XBl never extends to the zone 28M.
As an example of the relative sizes and spacings usable in the structure of the type thus far described, the following may be taken as presently considered typical, although, as will be appreciated, wide variations in all parameters are possible in accordance with the teachings herein. Bulk portion 22 may be doped to a concentration of 10" donors per cubic centimeter and the P- type localized zones doped to a concentration as great as conveniently achievable, typically at least 10 acceptors per cubicc centimeter. Zones 28 and 29 may be 40 microns in lateral dimension along the channel, and the distance between the zones may be about 10 microns. Insulator 23 may be 1,000 Angstroms of silicon oxide in the thinner portions and 3,000 Angstroms of silicon oxide in the thicker portions, i.e., under the trailing edges such as 25MA of the electrodes. With these described dimensions, the electrodes may be about 40 microns in lateral dimension along the chan nel, with 10 micron spaces therebetween for enabling facile fabrication. Such a structure may be operated with clock voltages V and V, of 10 and 20 volts, respectively.
Having now described a particular embodiment of the invention in detail, attention is directed to FIG. 5 which depicts an alternate embodiment in which the differences in threshold voltage along the channel are produced by differences in channel doping rather than differences in insulator thickness. More specifically, the structure of FIG. 5 is a cross-sectional view taken along the information channel of a portion 31 of a dual threshold bucket-brigade device in accordance with an alternate embodiment of this invention. As shown, portion 31 includes a semiconductive bulk portion 32 illustratively of N-type semiconductivity and including adjacent the surface thereof a plurality of P-type localized zones 38 and 39, like zones 18, 19, 28, and 29 in the preceding figures. Over the surface of bulk 32 and the zones is disposed an insulating layer 33 of substantially uniform thickness; and a plurality of electrodes 34 and 35 are disposed thereover. As can be seen, FIG. 5 differes from FIG. 1 only in that in FIG. 5 the trailing half of each transfer region includes an N-type zone 40 or 41, more heavily doped than bulk portion 32.
As is known, when negative voltages are applied to conduction paths 36 and 37 to the electrodes, the inclusion of more heavily doped N-types zones 40 and 41 in the transfer regions produces thereunder a threshold voltage greater than the threshold voltage in the less heavily doped portion, i.e., in the leading portion of the transfer regions. This difference in threshold voltage results from the exposure of ionized donor impurities in the more heavily doped zones 40 and 41. As will be appreciated by those in the art, inclusion of zones 40 and 41 causes the structure of FIG. 5 to produce potential configurations just like the stepped insulator structure of FIGS. 2 and 4A. Accordingly, no further detailed description of the structure of FIG. 5 is believed necessary.
At this point it should be evident that the important characteristics of a structure in accordance with this invention are the existence of more than one distinct threshold voltage in thetransfer region of the. bucketbrigade device with asubstantially abrupt transition therebetween. It is to be understood, however, that the various arrangements described are merely descriptive of the general principles of the invention and that various modifications which will be apparent to those skilled in the art may be made without departing from the spirit and scope of the invention. For example, it will be readily apparent that the semiconductivity types may be interchanged, provided corresponding changes in voltage polarities also are made.
It will also be apparent that a great variety of other techniques may be employed for achieving the plurality of threshold voltages in each transfer region in accordance with this invention. For example, the electrodes may be formed of contiguous segments of different metal having different work functions; or the insulator may have portions of suitably differing dielectric constants; or the electrodes may be formed of overlapping, isolated metals (or other conductive materials such as silicon) with a bias between related portions. Still other ways also are of course possible and all are considered within the scope of this invention.
Finally, it should be understood also that transfer regions having greater than two threshold regions also may be employed within the scope of this invention. If three are used, the feedback effect to which this invention is directed will be reduced to a third order effect. If four are used, then reduction will be to a fourth order effect; and so on. If more than two threshold voltages are used, the preferred arrangement would be with successively decreasing threshold voltages in the direction of advance of information; and, of course, there should be an abrupt transition between successive ones of the plurality. However, increasing the number of distinct threshold regions adds to the size of the device and introduces serious fabrication problems. Accordingly, the use of only two such regions is believed the most advantageous compromise with present technology.
What is claimed is:
1. In a bucket-brigade charge transfer apparatus of the type including: a storage medium having a major surface; a plurality of spaced localized zones of immobile charge disposed in a path along the surface; an insulating layer disposed over the surface and the zones; and a plurality of localized electrodes disposed over the dielectric and registered with the localized zones such that separate ones of said electrodes extend over the space between a separate pair of successive zones and over one zone of said pair more than the other zone of said pair,
the improvement which comprises: separate means associated with each space between a pair of successive zones for causing in said space a plurality of different distinct threshold voltages, with said threshold voltages decreasing from the trailing portion to the leading portion of said space, said means including means for causing a substantially abrupt transition between successive ones of said threshold voltages.
2. Apparatus as recited in claim 1 wherein said plurality is two, with the greater threshold voltage being in a trailing portion of the space.
3. In bucket-brigade charge transfer apparatus of the type including: a storage medium having a major surface; a plurality of spaced, localized zones of immobile charge disposed in a path along the surface; an insulating layer disposed over the surface and the zones; and a plurality of localized electrodes disposed over the dielectric and registered with the localized zones such that separate ones of said electrodes extend over the space between a separate pairof successive zones and over one zone of said pair more than the other zone of said pair,
the improvement which comprises: means associated with each space between a pair of successive zones for causing in the trailing portion of said space a first uniform threshold voltage greater than a second uniform threshold voltage in the leading portion of said space and for further causing a substantially abrupt transition between the two regions of differing threshold voltages.
4. Apparatus as recited in claim 3 wherein the means for causing the differing threshold voltages includes in the trailing portion of said space a thicker insulating portion than is disposed over the leading portion of said space.
5. Apparatus as recited in claim 3 wherein the means for causing the differing threshold voltages includes in the trailing portion of said space a greater concentration of dopant impurities in the storage medium than are disposed in the leading portion of said space.
6. Apparatus as recited in claim 3 including means for applying to the plurality of localized electrodes twophase voltages of magnitude and polarity sufficient for causing advance of information but insufficient for causing a depletion region to extend from a localized zone halfway across the space between said zone and the zone next preceding.
7. Apparatus as recited in claim 6 wherein the means for applying said voltages includes a pair of conduction paths, every second electrode being connected to a common one of said conduction paths and the remaining electrodes being connected in common to the other of said pair of conduction paths.
8. Apparatus as recited in claim 7 wherein the twophase voltages are applied to the conduction paths.
9. Apparatus as recited in claim 3 wherein the storage medium is a silicon semiconductive material and