DE2341855A1 - BUCKET CHAIN CARGO TRANSPORT COMPONENT - Google Patents

Description

BLUMBACH ■ WESER ■ BERGEN & KRAMER

PATENT LAWYERS IN WIESBADEN AND MUNICH

DIPI.-ING. C 6. BlUMBACH · DiPL-PHYS. Dr. W. WESER DIPL-ING. DR. JUR. P. BERGEN DIPL-ING. R. KRAMER

WIESBADEN · SONNENBERGER STRASSE 43 ■ TEL (06121) 542943, 561798 MUNICH

WESTERN ELECTRIC COMPANY, INCORPORATED BERGLUND NEWYORK, N.Y., USA

BUCKET CHAIN CARGO TRANSPORT COMPONENT

The invention relates to a bucket chain charge transport device with a storage medium having a main surface, a plurality of spaced apart along a path on the main " localized, zone-immobile charge lying on the surface, an insulating layer arranged over the surface and the zones, and a plurality of localized electrodes disposed over the dielectric layer and so aligned with the localized zones are that separate electrodes extend across the space between a separate pair of consecutive zones "as well extend further over one zone of the pair than the other.

As is well known in the art by now, work the bucket chain type charge transport devices by selectively transferring packets or the absence of

409821/0730

Parcels of load carriers move one after the other from zone to Zone along the surface of a semiconductor body comprising a body part of one semiconductor type and a plurality of spaced apart includes mutually lying localized zones of the other semiconductor type. In these components, each zone is considered to be one Potential well operated, the limit of which is determined by the pn junction defining the zone.

Conceptually simple, you can get the load transport components from the bucket chain! Imagine yp as a cascade of IGFETS (Insulated Gate Field Effect Transistor), in which each of the surface zones serves as the drain of a particular IGFET and as the source of the next IGFET. At any given time of operation, a pair of successive zones can then be regarded as the source and drain of an IGFET, respectively, and the overlying transport electrode can be regarded as the gate electrode of the IGFET.

As is well known, a significant factor in signal degradation in such devices is what has been termed "dynamic drain conductance", a feedback effect, the 409821/0730

to an increase in the effective distance between source and Drain during a transport process and leads to the reduction of the drain voltage during the transport of charge carriers is based on this. This effect has been described by C. N. Berglund and R. J. Strain in the article "Fabrication and Performance Considerations of Charge Transfer Dynamic Shift Registers, "published in Bell System Technical Journal, Volume 51, No. 3, March 1972, page 655.

One method of significantly reducing the effect of the dynamic drain conductance was proposed by F. L. J. Sangster in the Article "Integrated Bucket-Brigade Delay Line Using MOS Tedrodes" described in Philips Technical Review, Volume 31, 1970, Page 266 has appeared. Sangstefj's method involves doubling the number of IGFETS per bit of stored information and indicates the use of a trace to bias the added IGFETS in addition to the pair of conductor tracks, which are normally used to couple clock signals to the bucket brigade electrodes is used. In other words, Sangster ^ Methoae involves the use of twice the number of surface cells per Bits of stored information and a 50 percent increase

409821/0730

the number of conductors required to bring about a scheduled transport of information is used by the component. This method is considered overly complex for many applications, and brings with it in addition, a larger dimension per bit of stored information with it. Both disadvantages are intended to be reduced by the invention.

This problem is addressed in accordance with the invention by a bucket chain charge transport device solved, which is characterized by separate, each space between a pair of successive zones associated devices for generating a plurality of different Threshold voltages in the intermediate space such that the threshold voltages from the rear to the front part of the intermediate space decrease and there is a practically abrupt transition between successive threshold voltages.

In the drawing show:

Fig. 1 is a cross-sectional view of part of a basic known one Bucket chain component along the information channel,

Figure 2 is a cross-sectional view along the information channel of a To describe bucket chain component according to a first

409821/0730

the embodiment according to the invention,

3 shows a diagram in which typical surface potentials

are shown in the structure of Figure 2 when typical working voltages are applied,

4A is an enlarged view of a particularly important part

the structure of Figure 2,

4B is a diagram showing typical surface potentials

are shown, the typical operation of the structure of Figures 2 and 4A occur with a working voltage applied, and

Figure 5 is a cross-sectional view along the information channel

a part of a bucket chain component according to another embodiment of the invention.

It should be noted that for convenience and clarity the illustration and explanation, the figures are not necessarily to scale.

The present invention is based on the recognition that Songster's additional conduction path can be avoided its added transistors are constructed in such a way that they are a

409821 / G730

Have threshold voltage different from that of the original charge transport transistors is different and is based on the further knowledge that by using the different threshold voltages Sangster ^ s added surface zones can also be eliminated.

In an embodiment according to the invention, which can be seen as an example a bucket string assembly comprises first and second different from each other Threshold voltages in each transport area between Each pair of consecutive storage locations, the transport area is also characterized by a practical abrupt transition between the first and second threshold voltages and by the fact that the threshold voltage the rear part of each transport area with respect to the information transport direction is greater than the threshold value of the front part of the transport area.

This difference in threshold voltage, coupled with a practically abrupt transition therebetween, reduces feedback the voltage from the receiving zone (drain) to the transferring zone (source) is just as effective as Sangster * s

409821/0730

Two transistor-per-bit structure without the one described above excessive complexity.

Specific reference will now be made to the drawing. In Fig. 1 Figure 3 is a cross-sectional view along the information channel of a part Π of a fundamentally known bucket chain component, as disclosed, for example, in US Pat. No. 3,660,697. As shown is, the part 11 comprises a storage medium, the body 12, which is, for example, n-semiconducting and over which an insulating layer 13, typically silicon oxide, is formed. Above the insulator 13 is a A plurality of field plate electrodes 14M, 15M, 14N and 15N are arranged, each of which in E? ns-to-one match with a plurality localized zones 18M, 19M, 18N and 19N adjacent to the surface of the body part 12 is aligned. As is further indicated in the drawing, where the body part 12 is n-semiconducting, the upper surface areas are 18 and 19 of heavily doped p-type semiconductivity.

From Figure 1 it can be seen that the electrodes and the localized Zones are arranged in relation to one another in such a way that each electrode is considerably larger than the one below it on the right Zone than the zone below on the left. More specifically, for example, the electrode 15M covers one

409821/0730

considerably larger part of Zone 19M (below on the right) than that Zone 18M (below on the left).

A pair of conductor tracks 16 and 17 are each connected to every second electrode / ie conductor track 16 is connected to electrode 14 and conductor track 17 to electrode 15. During operation, two phase clock voltages V. and V- are applied to electrodes 14 and 15 via conductor tracks 16 and 17, respectively. It is known that the voltages V ₁ and V ~ are advantageously large enough to keep the semiconductor surface of the part 11 always depleted in order to minimize the effects of the surface conditions on the charge to be transported.

In a clock phase, for example when the amplitude of V. is greater than that of V ", the zones 18 under the electrodes 14 become a much larger surface potential than the zones

19 under the electrodes 15. As a result of this intended built-in Difference in the surface potential, which arises from the asymmetrical arrangement of the electrodes with respect to the zones, becomes movable Charge carriers that were in the zones 19 before the application of this momentarily described clock phase, from their respective Zone 19 transported into the zone immediately to the right in FIG. 1.

409821/0730

From the above description of the structure and the mode of operation is to see that each pair of adjacent zones, e.g. 18M and 19M, 19Mund 18N, and 18N and 19N can be regarded as the source and drain of an IGFET (Insulated Gate Field Effect Transistor). In such a context, the n-surface part between any part of adjacent zones can be understood as the channel of an IGFET.

It will be seen, however, that there are no drain and source electrodes which maintain the voltages at the source and drain regions when the transport of the charge carriers precedes. When charge carriers are transported into a zone, the magnitude of the surface potential in this zone and accordingly the voltage of the zone itself decrease as a result as each subsequent mobile charge carrier enters it. In a structure of the type shown in Figure 1, this decreasing potential creates an effective increase in length of the IGFET channel and thus makes it more difficult for further charges to be transported. This is the effect referred to above as "dynamic drain conductance".

409821/0730

ι ο

This channel lengthening effect can be understood if one considers that the voltages induced in zones 18 and 19 prove typical are of a polarity which is sufficient to generate a reverse bias at the pn junctions associated with these zones. As a result a depletion region of this extends from the zone in all directions and, importantly, to the left in Figure 1 into the channel region. The effect of this expansion of the depletion area in the channel is stone effect! ve Reduce the channel length. Like it for IGFETS is known, a decrease in the channel length produces a higher transmission conductance, which for a given applied voltage enables the charge carriers to be transported more easily from source to drain. If the amount of the surface potential in the drain zones, the depletion area is reduced in width and effectively extends the channel, which successively reduces the transmission conductance and at the same time the transport of the rest Makes charges more and more difficult.

It is easy to see that this effect increases as the load to be transported increases. For example, a relatively large charge carrier packet that may represent a "one" is more attenuated than a relatively small charge carrier ^{packet that represents a 1} NuII ". Since the effect is not uniform, can

409821/0730

11

significant signal degradation result.

Since, as stated above, the mode of operation of the structure of FIG is not directly analogous to the IGFET mode of operation, since there is no direct electrical connection to the zones, and there is also is much simpler for the following description, the use of IGFET terminology is discontinued at this point. Instead, the localized zones are called "charge storage sites" and the channel areas are along between zones on the surface as "transport areas". Additionally, the use of the terms "front" and "back" is useful what is related to the direction of movement of the signal information representing movable charge carriers. Because through the built-in Asymmetry in the structure of Figure 1 is defined that this direction that of transport to the right, the rightmost parts of some particular feature become front parts and the leftmost parts are referred to as the rear parts.

Taking into account this terminology and the above description of the fundamentally known bucket chain component and its Problems will be considered next in the structure shown in Figure 2.

409821/0730

tet, a cross-sectional view along the information channel of a part a bucket chain component according to a first embodiment of the invention.

As shown, the portion 21 comprises a semiconductor body portion 22, the is for example n-semiconducting and, adjacent to the surface thereof, comprises a plurality of p-conducting localized regions 28 and 29 which correspond to zones 18 and 19 in FIG. Above the surface of the body 22 and the zones 28 and 29, an insulating layer 23 is more uneven Thickness, and over the insulator 23 are a plurality of electrodes 24 and 25. As can be seen Each electrode comprises two different parts, the front part being marked with the suffix B and the rear part with the suffix A. is.

More specifically, there are a plurality of electrodes 24 and 25 in one-to-one correspondence with the plurality of located zones 28 and 29 arranged. As can be seen, the front portion 24MB of the electrode 24 extends over a large percentage of the zone 28M also approximately to the middle above the space between the zone 28M and the immediately preceding p-type zone; 0 »r rear

409821/0730

Part of the electrode 24M labeled 24MA is essentially only sufficient across the rear half of the space, i.e. the transport area, between zone 28M and the immediately preceding zone.

Likewise, each of the electrodes 25M, 24N and 25N comprises parts A and B analogous to the described parts A and B of the electrode 24M.

For operation, when two-phase clock voltages V, and V- are supplied via conductor tracks 26 and 27 to electrodes 24 and 25, a typical surface potential distribution (if there is no signal information-representing movable conductor carrier) is shown chemically in the diagram in FIG. In FIG. 3, the size of the surface potential in the structure of FIG. 2 is shown increasing in the downward direction, an arbitrarily selected zero reference level at the base of the arrow 31 being assumed. As can be seen, the horizontal dimension of the graph of Figure 3 is in accordance with the horizontal dimension of the structure shown in Figure 2. As can be seen further, it is assumed in the diagram of Figure 3 that larger the magnitude of the voltage V "than that of the Voltage V. is.

In operation , a voltage applied to the electrode 24M generates a first surface potential SI under the part 24 MA, its

409821/0730

Amount smaller than that of a second surface potential S2 below the rear part of that part of the electrode 24MB which is not above the Zone 28M lies, the smaller amount being due to the greater distance of the part 24MA from the semiconductor surface. In the case of the FIG The type of semiconductor shown are the voltages V, and V_ typically negative to cause depletion mode operation, in this case the negatively ionized acceptors in zone 28M tend to have the greatest surface potential in the negative direction to increase.

In the other half bit, that is, in the transport sector and in which the electrode 25M unallocated space 29M marked the surface potentials in Figure 3, assuming that V "is sufficiently greater than V _1, so that the minimum size of the through the rear edge 25MA Electrode 25M caused surface potential S4 more attractive, i.e. .H. is of a greater amount than the surface potential S2 assigned to the front part of the transport zone directly preceding the half-bit. More specifically, the surface potential applied to electrode 25M caused by voltage V_ can be viewed as having three wedges like that of the surface potential under electrode 24M, except that it is shifted in the direction of increased attraction for movable supports.

409821/0730

As shown, the surface potential S3 of the zone is 28M practically equal to the surface potential S4 of the rear part of the subsequent transport zone, which is due to the copious amount of mobile charge carriers in the heavily doped zone 28M. As you can see that the potential diagram repeats itself with two-electrode periodicity, for example, from the rear edge of electrode 24M to the rear edge of 24N.

At this point it should be noted that the relative potential differences remain constant within a given half-bit and are determined by the physical structure. More precisely, the difference between the surface potential S2 and the surface potential Sl becomes, assuming complete emptying in the region of S1, caused only by the difference in oxide thickness under electrode 24M. The abruptness of the transition from S1 to S2 is due to the abruptness in the transition from the thicker ones to the thinner oxide parts. The difference in potential between S2 and S3 is due to the relative concentrations of the Donor atoms in the transport area and the acceptor atoms in storage space 28M. Of course, the same physical structure is below each electrode repeated or can be repeated, and so on are the relative levels of the wedges under electrode 25M and

409821/0730

other electrodes are the same as those under electrode 24M. Such a relationship is shown in FIG.

Referring to Figure 4A, there is an enlarged view of a particularly important wedge of the structure according to the invention of FIG. 2, with a typical transport area forming the center. Figure 4B is similar to Figure 4A aligned diagram and shows typical surface potentials S, which occur during typical operation of the structure of FIGS. 2 and 4A.

As can be seen, the structure of FIG. 4A is just an enlarged view of that part of FIG most of the right-hand part of zone 28M, the transport area between zones 28M and 29M, the left-most part Part of zone 29M and includes the insulator and the overlying electrode structure. As can be seen, the transport area closes between the zones 28M and 29M two parts, which are designated with T1 and T2, respectively. Tl is that part / advantageously about half of the transport zone closest to zone 2ΘΜ and below the leftmost part of electrode 25M, this part being labeled 25MA. The part 25MA is arranged at a greater distance, i.e. above one larger insulator thickness than the 25MB part / the above the other

409821/0730

Half T2 of the transport zone lies.

As shown in Figure 4B, applying a voltage to the Electrode 25M has a surface potential distribution in the transport area, which has two different levels S4 and S5 in the areas TI and T2, and there is a relatively abrupt transition between the two levels of surface potential that is on the relatively abrupt one transition between the different insulator thicknesses is based. As mentioned above are both the presence of the two levels and the existence of the relatively abrupt transition between them important for this invention for reasons to be explained.

First of all, however, it should be noted that the presence of the two different levels of surface potential in the transport area of a structure, as shown in Figure 4A, is based solely on the fact that the electrode part 25MA is arranged over a thicker insulator area and therefore a greater distance from one Has semiconductor surface than the electrode part 25MB. As is known in the relevant prior art, this described difference in the surface potential, which is caused by the application of a uniform voltage to the electrode 25M, results from the different spacing of the parts of 25M from the half

409821/0730

leather surface and creates correspondingly different ones underneath Threshold voltages. It can thus be said that the structure according to FIG. 4A is assigned different parts Tl and T2 Has threshold voltages and further has a relatively abrupt transition therebetween.

With particular reference to Figure 4B, the importance of the two different threshold voltages in conjunction with a relatively abrupt transition between them will now be explained. First of all, it should be noted that the current, i.e. the flow rate of conduction carriers, through the transport area for any given voltage applied to the overlying electrode is limited by the part of the transport area which has the greater threshold voltage, since the part of the underlying surface is less attractive to movable ones Load carrier is. Thus, in FIGS. 4A and 4B, the current flow rate through the transport area is limited by the rear part T1 of the transport area. Since the current flow is limited in this way, length variations of the half-channel assigned to the area T2 due to variations in the depletion width, denoted by XD2 in FIG , 409821/0730

very little effect on the ease with which the current passes through the entire transport area can flow. This is of course based on the fact that the flow of current is limited in principle by the half-channel T1 which is only influenced in the second order by voltage variations at S6.

With the above in mind, some other outstanding properties will now be discussed the structure and the surface potential distribution as shown in FIGS. 4A and 4B. as The first is the surface potential transition between the areas T1 and T2 have been described as being relatively abrupt and shown in Figure 4B. The degree of abruptness for use in this invention is advantageous, does not in itself lead to a precise quantification. Due to a difference between the potentials S4 and S5 occurs however, in the transition area between these two potentials, an electric field occurs, and results from this electric field a depletion zone, the width of which is indicated in Figure 4B with XDl. All that can be said about the degree of abruptness in the transition between the surface potentials S4 and S5 is desirable, can say, is that such a transition should be sufficiently abrupt, to generate an electric field of sufficient strength to produce a depletion zone XDl, so that the concentration moves

409821/0730

Licher carrier in the channel area at the left edge of the depletion area XD1, based on the concentration of such carriers at the left edge of the area P1 ₇ ie adjacent to the right edge of the zone 28M, effectively goes to zero. Only when this condition is met is the second order effect of the variations in the surface potential S6 on the above-mentioned transmission conductance achieved.

The size of the difference between the potentials S4 and S5 is according to the invention also of importance, but also does not lead to a precise quantification. However, it is believed that this difference

_3 nigstens should kT be, which is about 26 X 10 volts at 300 degrees Celsius, and DA13 it should not be larger, than the magnitude of the difference between the surface potentials S3 and S6, as caused by the applied voltages V ₁ Uncl V " , the latter difference typically being in the range of 10 volts. In practice, a difference of about 1 or 2 volts between the surface potentials S 4 and S 5 is considered to be practicable. This difference can easily be achieved with simple manufacturing differences in the oxide thicknesses among the electrode parts.

The total length of the duct area, i.e. the distance between the

Zone 28M and Zone 29M should preferably be so large that 409821/0730

the depletion area XD2 from zone 29M never fully extends extends to the right edge of the depletion region XDl, even below the highest expected operating voltages supplied, since otherwise the second order effect, as is desired according to the invention, is not fully reached. The width of the half-channel Tl should also be so be large that the left edge of the depletion region XDT never extends to zone 28M.

As an example of the relative dimensions and clearances as shown in the Structure of the type described so far are applicable, the following may be considered to be typical at present, although as will be seen, further variations in all of them according to the teaching herein Parameters are possible. The body part 22 can be doped to a concentration of 10 donors per cubic centimeter and the p-conducting localized zones to be as expedient as possible

19 attainable concentration, typically at least 10 acceptors per cubic centimeter. The zones 28 and 29 can have lateral dimensions along the channel of 40 microns and the spacing between the zones can be about 10 microns. The isolator 23 can 1000 angstrom silicon oxide in the thinner parts and 3000 angstrom silicon oxide in the thicker parts, i.e. under the rear edges like 25MA of the electrodes. With these dimensions you can

409821/0730

the electrodes have a lateral dimension along the channel of 40 micrometers have, with 10 micron intervals in between, to make an easy Enable manufacture. Such a structure can be operated with clock voltages V. and V_ of 10 and 20 volts, respectively.

After now a particular embodiment of the invention in detail reference is made to Figure 5 which shows another embodiment in which the threshold voltage differences generated along the channel by differences in the .Kanaldotierung instead of the differences in the insulator thicknesses will. More specifically, the structure of Figure 5 is a cross-sectional view along the information channel of a portion 31 of a double-threshold bucket string component according to another Embodiment of the invention. As shown, the part 31 comprises a semiconductor body part 32, for example n-semiconducting, and has a plurality of p-type conductors adjoining its surfaces localized zones 38 and 39 similar to zones 18, 19, 28 and in the previous figures. Above the surface of the body 32 and the zones, an insulating layer 33 of practically uniform thickness is arranged, and there are a plurality of electrodes 34 and 35. As can be seen, Figure 5 differs from Figure 1 only in that in Figure 5 the rear half of each

409821/0730

Transport area includes a η-conductive zone 40 or 41, which is stronger is doped as the body part 32.

If negative voltages are applied to the electrodes via the conductor tracks 36 and 37, the presence of more heavily doped ones creates As is known, η-conductive zones 40 and 41 in the transport areas include a threshold voltage which is greater than that is in the less heavily doped area, i.e. in the front part of the transport areas. This difference in threshold voltage results on the effect of the ionized donor dopants in the more heavily doped zones 40 and 41. As defined with this area As is clear, the presence of the zones 40 and 41 has the consequence that the structure according to FIG step-shaped insulator structure of Figures 2 and 4A is the same. Therefore, a further detailed description of the structure is given not considered necessary according to Figure 5.

It is now evident that the important properties of a structure according to the invention, the presence of more than one different threshold voltage in the transport area of the bucket chain component with a practically abrupt transom in between is important. It goes without saying, however, that the

409821/0730

the various arrangements listed are only descriptions of the general principles of the invention; and numerous modifications to those skilled in the art within the scope of this invention. For example, it is obvious that the semiconductor species can be exchanged / grounded, provided that appropriate voltage polarity changes are made.

It is also apparent that a wide variety of other methods can be used to provide the plurality of threshold voltages in each transport area according to the invention. For example, the electrodes can be formed from adjacent segments of different metals that have different behavioral functions or the insulator may have parts of suitably different dielectric constants, or the electrodes can be formed from overlapping, insulated metals (or other η conductive materials such as silicon) with a bias between assigned parts. There are of course other possibilities within the scope of this invention.

Finally, within the scope of this invention, transport areas with more than two threshold value areas can of course be used. If three are used, the feedback effect will be

409821/0730

to which this invention is directed is a third order effect ! reduced. If four are used, a reduction to a fourth order effect is achieved, etc. If more than two threshold voltages is used, an arrangement would be preferable with successively decreasing threshold voltages in the information transport direction and of course there should be an abrupt transition between successive threshold ranges of the plurality to be available. However, increasing the number of different threshold ranges increases the device size and involves considerable manufacturing problems. Accordingly, the use of only two such areas is not the current one Technology to view the most advantageous compromise.

409821/0730

Claims (1)

PATENT CLAIMS ( 1. Bucket chain charge transport device comprising a storage medium having a main surface, a plurality of localized zones of immobile charge lying at a distance from one another along a path on the main surface, an insulating layer arranged over the surface and the zones, and a plurality of localized electrodes, the disposed over the dielectric layer and aligned with the localized zones such that separate electrodes extend farther than the other across the space between a separate pair of successive zones and over one zone of the pair, characterized by separate, every space between a pair of consecutive ones Devices (25MA, 25MB, 40, 41) assigned to zones for generating a plurality of different threshold voltages in the space such that the threshold voltages from the rear to the front of the space decrease, and there is a practically abrupt transition between successive threshold voltages. 409821/0730 ■. ■ η - 2. The component according to claim 1, characterized in that the. -. Threshold voltage generation line across the rear Part of the space has a thicker Isollerschlchttell than over. the front part of the gap. 3. Component according to claim 1,. · -. ■ ^: r characterized in that ",,""-. that the threshold voltage generation facility for bringing about different threshold voltages in the rear part. -The interspace has a greater concentration of oxygen than Im '. having front part of the gap. ■ · "- 4. Component according to claim 1 / ". ' ^: S .... ■; _; marked by Circuit center (V. / V.) connected to the plurality of localized electrodes for supplying two-phase voltages, whose ^: Greetings and polarity are sufficient to convey information, but not To do this, a depletion layer extends from a localized zone to -In the middle of the gap between this zone and the directly. - to extend the preceding zone. ■ 5 component according to claim A, characterized by 40 021/0730 ORIGINAL l ^ fSPECTEb if that the voltage supply switching means comprise a pair of conductor tracks (26 _r 27, 36/37), each second electrode being connected to one conductor track and the remaining electrodes to the other conductor track each being commonly connected. 6. The component according to claim 5, characterized in that the two-phase voltages are applied to the conductor tracks. 7. Component according to claim 1, characterized in that that the storage medium is silicon semiconductor material and the insulating layer comprises silicon oxide. 8. The component according to claim 1, ■ characterized in that the plurality is two with the rear portion of the gap has the larger threshold voltage. 409821/0730