US 3585462 A
Abstract available in
Claims available in
Description (OCR text may contain errors)
United States Patent Kurt Lehovec Williamstown, Mass. 775,240
Nov. 13, 1968 June 15, 1971 Sprague Electric Company North Adams, Mass.
Inventor Appl. No. Filed Patented Assignee SEMICONDUCTIVE MAGNETIC TRANSDUCER 11 Claims, 9 Drawing Figs.
Int. Cl 1101111/00 Field of Search 317/235,
 References Cited UNITED STATES PATENTS 2,778,956 1/1957 Dacey et a1. 317/235 2,943,269 6/1960 Huang 317/235 3,389,230 6/1968 Hudson 317/235 3,401,319 9/1968 Watkins 317/235 Primary Examiner-Jerry D. Craig Attorneys-Connolly and Hutz, Vincent H. Sweeney, James P.
OSullivan and David R. Thornton PATENTED JUN! 51971 SEMICONDUCTIVE MAGNETIC TRANSDUCER BACKGROUND OF THE INVENTION This invention relates to magnetic transducers and more particularlyto a semiconductive magnetic transducer having an elongated semiconductive channel.
Magnetic transducers which are perhaps best known for their application in tape recorders, find many applications in modern electronics, including switching applications and the like. In the prior art, semiconductive magnetic transducers, which utilize magnetic deflection of minority carriers of a baselike semiconducting region to point contact collectors or to an underlying pair of collector regions, is available. However, devices of this type are generally low in sensitivity and limited in their application.
It is an object of this invention to provide a highly sensitive semiconductive magnetic transducer.
It is another object of this invention to provide a magnetic transducer having an elongated channel contiguous along one side with a collector region.
It is still another object of this invention to provide a magnetic transducer having an electric field disposed in an elongated semiconductive channel so as to enhance the flow of minority carriers along the channel.
It is a further object of this invention to provide a semiconductive magnetic transducer in combination with semiconductive regions which provide amplifying characteristics.
It is a still further object of this invention to provide a semiconductive magnetic transducer which also provides unipolar and bipolar characteristics.
SUMMARY OF THE INVENTION Broadly, a semiconductive magnetic transducer providedin accordance with the invention comprises an elongated semiconductive channel bounded on one side by a collector region of minority carriers, means disposed at one end of said channel for injection of minority carriers therein, and said channel having an electric field therein of such polarity as to direct the minority carrier flow along the axis of said channel such that a magnetic field applied across the channel will deflect said drifting minority carriers and control their collection in said collector region.
Broadly, a magnetic transducer-amplifier providedin accordance with the invention comprises a body of semiconductor material, a semiconductive channel in said body, at least one collector region of said body adjoining said channel, means for injecting minority carriers in said channel, an amplifying unit disposed in said body with its input coupled to said collector such that the output of said amplifyingdevice is proportional to the minority carrier flow to said collector.
In a more limited sense, the transducer-amplifier includesa feedback within said body from said amplifying device such that minority carrier flow is increased. This is accomplished in one embodiment by making said collector also operate as source of majority carriers of said channel which thereby increases the minority carrier injection and flow of said channel.
BRIEF DESCRIPTION OF THE DRAWING FIG. I is a view in section of a semiconductive magnetic transducer provided in accordance with the invention;
FIG. 2 is a view in section of another embodiment of the invention wherein an electric field of the channel is provided by externally charged contacts at opposing ends of the channel;
FIG. 3 is a plan view of the semiconducting structure .of FIG. 2;
FIG. 4 is a view in section of another embodiment which provides a semiconductive channel parallel to the wafer surface;
FIG. 5 is a graph of the lateral impurity distribution in the channel of the structure of FIG. 4;
FIG. 6 is a view in section of still another embodiment of the invention wherein a collector of the device is provided by a Schottky barrier contact;
FIG. 7 shows a further embodiment of the invention wherein a collector of the device is provided by an inversion layer region;
FIG. 8 is a view in section of a transducer-amplifier provided in accordance with the invention; and
FIG. 9 is a view in section of a further embodiment wherein a transducer and an MOS amplifier are integrally combined.
DESCRIPTION OF THE PREFERRED EMBODIMENTS In order to provide a clear and concise description of the invention, a brief analysis of the magnetic deflection of minority carriers travelling in a semiconductive channel is first presented.
In the preferred embodiments, the novel transducer utilizes an extended semiconductive channel having a pair of collectors of minority carriers which border on the sides of the channel, and an emitter, or other means for injection of minority carriers is disposed at one end of the channel such that minority carriers injected into the channel may be preferentially deflected towards or away from the collectors by a magnetic field.
In such a case, the force acting on the charge q moving at velocity v in a magnetic flux density B which is orthogonal to v F H =vB 10 where the force F H expressed in terms of an equivalent electric-field as V/cm., v is in cm./sec. and Bis in gauss.
The lateral (i.e., orthogonal to v and B) displacement in the time t is:
5=p,F t =p.L'B 10" where L'=vt is the distance travelled along the channel in the time t, and if a channel having a length L L' and a width W between collectors is assumed, we have a transducer efficiency of olw =p.B 10 L'/W. For instance, if p.=l0 cmF/volt sec., B =10 gauss, L'=l0 cm and W 10 cm, then 6/W= 10- and consequently, Al 10- I assuming that the entire emitter current I is injected into the channel, and AI is the change in the collector current caused by the magnetic flux.
Of course L, the distance travelled along the channel bounded by two collectors, can be limited by the bulk lifetime 1, but for narrow channels bounded by two collectors it is more likely that it will be limited by lateral Brownian motion whereby minority carriers are sucked up by the collector space charge regions. Now since the transit time t from the center of the channel to one of its boundaries by Brownian motion is:
1 IZVI where D is the diffusion constant of the minority carriers which is related to their mobility by Einstein's law D =p.kT/q when k is the Boltzmann constant and Tthe absolute temperature, if there is no electric field along the channel the effective length L in the region between the collectors is:
L'-=-'(2Dr) or (ZDI 1/2 whichever is smaller. Thus, L cannot be larger than about W/2. However, by providing an electric field F along the channel, the distance travelled becomes L' fLFT0l' L' pFt =F( W/2) /2(kT/q) whichever is smaller.
Thus, using an electric field along the channel, the ratio L/W/2 can be made larger than unity with corresponding increase in the sensitivity of the device. For instance, consider the case of an exponential acceptor distribution in a p-type channel:
C=C,, exp(x/x,,, where C, is the concentration at the end of the channel where minority carriers (-electrons) are injected, and x is the distance along the channel from this injection point.
Then the built-in field is: I
If the value of L is limited by lateral diffusion of electrons to one of the channel boundaries, rather than by bulk recombination, one obtains by substitution of F into one of the previous equations:
Therefore, the ratio L'lW, which enters the transducer efficiency as outlined previously, becomes:
In order to maximize this expression, it would appear desirable to choose x, as small as possible, i.e., to select a steep impurity gradient, However, the impurity gradient has to extend along the entire channel length L, and the concentration, C at x=L has to be still sufficiently large to provide a significant doping of the semiconductor, and since:
x,,=L/Ln(C,,/C, one has, L,(lI'/2) Ln (C,,,'C )/2 For C,,=10 CHI-3, C =1O cn1. Ln(C /C )/2 becomes equal to 2.4, i.e., the channel length should be chosen at least 2.4 times the half-width of the channel.
Using the above equations, we may also express the optimized x, in terms of the channel width Was follows:
The denominator is 4.8 for the numerical example chosen. Thus for a channel width of W=l cm. and a dopant concentration at the ends of the channel as indicated above, one should select x cm. and a channel length L of at least l.2 10cm., and for silicon at room temperatures the electron lifetime r in the p-type channel should be at least (W/2) 2DES see.
In FIG. 1, a semiconductor body 10 of P-type silicon or the like is provided with a pair of spaced apart N-type collector regions l2 and 14 which form a channel region 16 of the P-type body interposed between them. An injector region 18, in this case an N-type emitter region, is provided at one end of channel 16 where it extends to the wafer surface 20 as shown. Conductive contacts 22 and 24 with appropriate leads 26, 28 provide low resistance or ohmic contact to collectors l2 and 14 respectively. Contact 30 and lead 32 connect to the injection end of channel 16, and contact 34 and lead 36 connect to injector region 18.
For an understanding of the invention, consider that channel 16 has a width W as measured between opposed collector faces 40 and 42 and a length L, as measured from surface 20 to a point 44 near the bottom of regions 12 and 14. Then, minority carriers injected from region 18 fiow along channel 16 to regions 12, 14, or recombine with majority carriers either in channel 16 or in the semiconductive body beyond the channel.
If a magnetic flux B (not shown) is directed into the paper, that is, a magnetic field is applied substantially perpendicular to the longitudinal axis of the channel and parallel to the interface 40 and 42, the minority carriers will be deflected away from one collector region and towards the other depending upon the polarity of the applied magnetic field.
In the preferred embodiments, the channel is elongated to increase its length, and the drift velocity v is increased by an electric field provided in the channel by either a suitable impurity profile or an externally applied electric bias. The elongated channel is provided in the structure of FIG. 1 between deep diffused pockets [2 and 14 which extend to the wafer surface, and a longitudinal channel field is built in to this structure by providing an impurity distribution in channel 16 having large impurity concentration at wafer surface 20 which gradually diminishes to a low concentration at the channel end 44. Consequently, electrons injected at the end of high impurity concentration (adjacent to region 18) are driven by this field towards the low concentration at end 44 such that drift velocity v is increased.
The structure of FIG. 1 may be made by conventional means, as for example, by first doping a silicon wafer 10 with acceptor impurities such as boron or the like. For instance, boron may be diffused into surface 20 from an ambient vapor phase to provide P-type conduction having a high concentration at this surface and a gradually diminishing concentration in the direction of the base of the wafer. Thereafter, spaced apart N-type regions 12, 14 and an N-type injector region 18 are formed in wafer 10 by conventional techniques, for example, by diffusion of donor impurities such as phosphorus or the like through mask openings into spaced apart portions of surface 20. Low resistance contact films of aluminum or the like, may then be deposited by conventional means on appropriate surface portions, and finally, leads are then provided by any of the bonding techniques commonly employed in the semiconductor arts.
Since regions 12 and 14 are deep regions while region 18 is comparatively shallow, the former would generally be diffused into wafer 10 before the formation of region 18. Of course, the diffusant chosen for region 18 could also be slow moving as compared to that of regions 12 and 14.
A preferred channel length of at least 2.4 times the halfwidth of the channel is provided in this embodiment by spacing of regions 12 and 14in accordance with the diffusion constant of the impurities used. In addition, the values of X and the dopant concentrations at the ends of the channel are made to conform to those indicated in the preceding theoretical discussion.
As previously indicated, there are inherent restrictions in providing a large built-in longitudinal field over a long channel, however, these restrictions can be overcome by generating an electric field from an external power source in connection to contacts located at opposite ends of the channel.
In the case of an externally applied channel field, the length L' is restricted by the saturation drift velocity of minority carriers which is approached with increasing field F. In the case of silicon, this saturation velocity is v,=l0 cm./sec. Thus In this estimate, we have used the diffusion constant for holes at low field conditions which strictly does not apply to the hot" electrons flowing at the saturation drift velocity. Nevertheless, the error thereby introduced is minor and does not detract from the fact that large channel length/width ratios can be employed with externally applied fields.
An externally applied field is provided in the structure of FIGS. 2 and 3 by means of an additional contact 50 and lead 52 which connect to wafer 10 to permit application of an electric bias to channel 16 from an external source (not shown). In these FIGS. parts identical to those of FIG. 1 are identified by corresponding reference numerals.
Preferably, the unit of FIGS. 2 and 3 are constructed in a manner similar to that of FIG. 1 to provide a long narrow channel, however, since a field is externally applied a graded impurity concentration of channel 16 is unnecessary. Assuming, a P-type channel (P-type wafer) in the structure of FIG. 2 contact 50, if positively biased as compared to contact 30, will act as a source of majority carriers (holes) while contact 30 acts as a drain of these carriers, and majority carrier flow from contact 50 to 30 proceeds along the dotted line 54.
The majority drain contact 30, preferably surrounds injector region 18 and is substantially concentric to it at surface 20. Consequently, the deepest point of the injector region faces the most positive value of channel 16 comparing points along the injector-channel interface, and the injected electrons are concentrated at the center of the channel.
In contrast to devices having a built-in field, when the field is externally applied between electrodes located at the channel ends, steps must be taken to confine the majority carrier flow to the channel. Hence, it is necessary to completely surround the channel by the collectors or external boundaries of the semiconductive wafer. This may be accomplished by making regions 12 and 14 extend around the channel. This can be seen from FIG. 3 wherein regions 12 and 14 substantially conform to contacts 22, 24 which they underlie. A small gap 62 and 64 are provided at the ends of the collector regions, and these are biased positive with respect to the surrounding ptype layer such that the space charge layer generated at the collector junctions extends from regions 12, 14 to the dotted lines 56 and 58. This space charge layer leaves a p-type opening for the channel 16 (within line 56), but covers completely gaps 62 and 64 between the horseshoe-shaped collector regions 12, 14. In this manner, channel 16 is laterally pinched off from the surrounding p-layer 66, yet regions l2, 14 remain insulated from each other by gaps 62, 64.
The structures illustrated in FIGS. 1-3 provide a vertical channel perpendicular to the wafer surface; however, there are several advantages of structures having a horizontal channel parallel to the wafer surface since this provides: (1 access of the channel to a highly localized magnetic field, for instance, the field ofthe domains on magnetic'tape; (2) an elongated channel without the necessity of deep diffusion of the collector regions; (3) other means of collecting injected minority carriers than the use of reverse biased pn junctions; and (4) case of providing a lateral impurity profile in the channel.
A structure having a horizontal channel is shown in FIG. 4 which depicts a structure having a p-type channel 70 parallel to the wafer surface 72. In this embodiment, the bulk of wafer 74 is an N-type substrate which provides one collector region. Substrate 74 is overlaid by a p-type epitaxial layer 78 which has an elongated N-type region 76 disposed in upper surface 72. Region 76 provides the other collector, and together with substrate 74 forms a ribbonlike channel 70 from the interposed portion of layer 78. An N-type injector region 80 is laterally spaced from region 76 at one end of channel 70, and is positioned between region 76 and a majority drain contact 82 which connects to layer 78. A majority source contact 84 is provided at the other end of channel 70, and contacts 86 and 88 are provided to regions 76 and 80, respectively. It should be noted that drain contact 82 is, in this case, made negative with respect to source contact 84 so that majority carrier flow is along line 90 from contact 84 to 82, and enhances minority carrier flow in an opposite direction through channel 70 from region 80. The minority carrier flow in channel 70 is then magnetically deflected to either collector 74 or 76 by a magnetic field applied parallel to wafer surface 72 and orthogonal to channel flow 90; that is, orthogonal to the plane of the figure.
A built-in electric field can also be utilized in this embodiment to increase sensitivity. That is, the lateral transit time may be increased by providing a lateral drift field in the channel which opposes the flow of injected minority carriers toward the channel-collector interface.
The lateral field may be accomplished by providing a concentration valley, that is a saddlelike impurity profile, perpendicular to the longitudinal axis of the channel. For example, the impurity concentration (e.g., boron concentration) of the gaseous ambient may be decreased during epitaxial growth of layer 78 and then again increased to the original level so as to provide the distribution plotted in FIG. 5.
In FIG. 5, the impurity concentration is plotted along the ordinate while the distance from the channel center is plotted along the abscissa. Herein, points a and 17 represent the channel edges or channel-collector interfaces 71 and 73 of FIG. 4. This graphically illustrates the desired saddlelike impurity distribution having maxima at points c and d, adjacent the channel edges, and a minimum at e which corresponds to the channel center.
Hence, between c and d, there is a built-in field which tends to drive majority carriers away from and minority carriers toward the channel axis e. Consequently, the applied magnetic force has to overcome this lateral drift field in order to propel minority carriers toward either collector.
It should also be understood of course, that a built-in longitudinal drift field could also be utilized in the parallel channel structure of FIG. 4, and conversely, the lateral field of FIG. 5 could also be employed by appropriate construction in the vertical channel structure of FIG. 1.
In another embodiment, shown in FIG. 6, the upper collector is provided by a Schottky barrier contact 92. In this case, region 76 and junction 73 of .FIG. 4 is replaced by metallic layer 92 which is deposited on surface 72 in accordance with well known methods of preparation. For example, a thin layer of chromium may be deposited on surface 72 of the p-type layer 78 to provide the Schottky barrier contact. This then provides a channel portion 70 of epitaxial layer 78 interposed between collector substrate 74 and collector 92.
In still another embodiment as shown in FIG. 7, collector region 76 of FIG. 4 is replaced by an inversion layer 94, induced in the upper surface of layer 78 by a positively biased metal film 96 which overlies an insulating film 98. For example, a layer of aluminum or the like deposited over a silicon dioxide film by conventional MOS techniques would be suitable.
In this case, contact to inversion region 94 is provided by a connection to its lateral edge (not shown) or by a lead (not shown) insulatively extended through film 96 and 98.
In he embodiments shown in FIGS. 4, 6, and 7, the upper collector regions (76, 92 or 94) and their appropriate contacts may be laterally extended (that is in a direction into the paper) to provide a thin ribbonlike channel, and the extreme edges of the channel can be laterally terminated in each case by the wafer edge or by a pair of grooves or the like, cut or etched, in surface 72 parallel-to channel 70 and penetrating to collector 74.
The described transducers can also be combined in an integrated arrangement with an amplifying unit, for example by utilizing one collector as a base region of a bipolar PNP transistor such as that shown in FIG. 8, or may be combined with an MOS-type unipolar transistor, for example as shown in FIG. 9.
In FIG. 8, the transducer of FIG. 4 is modified by providing a p-type emitter region 100, formed by diffusion or the like, in upper collector 76. In this case, contact 86 is affixed to region 100 so that current flowing from channel 70 into region 76 must pass through the p-n junction 101 between region 76 and 100. Thus, channel collector 76 can be made to also operate as the base region of a bipolar PNP transistor 106 whose collector is a portion of channel 70.
In operation, contact 86 is connected at 102 to the positive side of a voltage source 104 such that PNP transistor 106 operates in a manner similar to contact 84 of FIG. 4, as a source of holes in channel 70. Consequently, the input to transistor 106 is partly controlled by the collector current of the transducer. Drain contact 82 is also connected in common to injector contact 88 at 108 and then over a load resistance 110, such as a 5 kilo ohm resistor, to the negative terminal of source 104 as shown.
Then, the voltage drop caused by the fiow of holes in channel 70 between transistor 106 and drain contact 82 provides a slightly negative bias for the injecting p-n junction 112 since contact 88 of region is connected to 82. Consequently, minority carriers are injected over junction 112 into channel 70, and some of them are collected by the n-layer 76, thus providing the base current of the transistor 106. Hence, the device has internal feedback, since the minority carrier current from channel 70 to collector 76 governs the majority carrier current of channel 70 which, in turn, determines the voltage drop in the channel between junction 112 and contact 82, V
Switching from the on"-state to the ofi-state is provided by magnetic deflection of minority carriers away from the collector region 76. However, for some applications, it is also desireable to provide an additional majority carrier source contact 114 connected over a resistance 116, for example 50 kilo ohms, to the positive terminal 102 of the battery. In this case, there will be a finite channel current and some limited injection overjunction 112, although the transistor 106 might still be off due to magnetic deflection of most of the electrons away from the collector 76. Reversal of the polarity of the magnetic flux can trigger such a device into the on-stage of transistor 106 thereby providing greatly enhanced power through load resistor 110.
FIG. 9 shows a further arrangement whereby the transducer illustrated in H6. 4 is modified such that the current change to collector 76 is amplified by an MOS-transistor 120. Herein, transistor 120 consists of an n-type region 122 whose surface is partially covered with an insulating film 124 and an overlying gate contact 126.
A p-type region 128 is disposed within region 122 and spaced away from collector 76. Region 128 provides a conventional MOS source region of transistor 120 and is connected through a resistor 123 to gate 126 and region 76. Resistor 123 and region 128 are connected, in turn, to the positive side ofa voltage source 132.
Injector region 80 and contact 82 are connected in common through load resistor 110 to a ground tenninal 138 and the negative side of voltage source 132. As in the structure of FIG. 8, source terminal 114 is connected to voltage source 132 over resistor 116, and finally an output terminal 140 is provided in connection to load resistor 110 as shown. This provides a voltage output between terminal 140 and ground 138.
A sufficiently large negative bias voltage is applied against region 122 by means of the positive voltage of gate electrode 126 which causes a p-inversion layer to form at the interface between region 122 and insulator 124, thereby providing an inversion conductance between the p-pocket 128 and the terminal end 142 of the p-channel 70. The negative bias to region 122 results from electron current flow from channel 70 into collector 76 which, in turn, flows over register 123 to the positive terminal of battery 132.
Again, as in the structure of FIG. 8, we have an on"-stagc in which electron current over junction 112, magnetically directed to collector 76, causes a gate bias which turns on MOS-transistor 120. This, in turn, provides the hole channel current along dotted line 144 causing increased electron injection from region 80. In the off-stage, with no electron collection by region 76 (due to the magnetic field) transistor 120 is turned off, and little or no electron injection from region 80 is present.
The described devices can of course be utilized in many different applications. For example, the structures shown in FIGS. 8 and 9 are suitable for use in switching applications such as for a magnetically activated light switch, or the like. This may be accomplished by any suitable arrangement which provides reversal of the magnetic polarity across the channel. That is, a conventional U-shaped magnet (not shown) may be rotatably mounted over the device so that its field may first pass perpendicular to the channel in one direction and then made to pass in the reverse direction by rotation of the magnet through 180 to provide off-on states of the switch. Concentration of the magnetic field can also be enhanced by the use of soft iron pole pieces or the like extended laterally from the sides of the transducer to the magnet faces. Then, rotation of the magnet to bring its opposite poles to bear on the pole pieces provides field reversal and circuit switching.
Many different embodiments are possible, of course. For example, N-type channels with appropriate collector and injector regions may be employed. In this case, regions of opposite type to those given in the description would also be required for the amplifier portions ofthe transducer-amplifier structures.
In addition, substrate 74 may he an insulative member such as sapphire or the like which provides a supporting base upon which layer 78 may be epitaxially deposited, for example. In this embodiment the channel is sandwiched between one collector and an insulating means.
Semiconductive materials other than silicon may also be employed. Both the built-in and externally applied channel field may be utilized, and many different embodiments of the transducer-amplifier structure may be possible.
Consequently, many different embodiments are possible without departing from the spirit and scope of the invention, and it should be understood that the invention is not to be limited except as in the appended claims.
1. A semiconductive magnetic transducer comprising an elongated semiconductive channel, means disposed at one end of said channel for generation and injection ofa constant flow of minority carriers therein, at least one region for collection of minority carriers contiguous with a side ofsaid channel that is substantially parallel to the flow of minority carriers and of opposite conductivity type from said channel, wherein a magnetic field will deflect said drifting minority carriers thereby affecting their collection in said collector region, and a nonvariable electric field provided within said channel by a gradually decreasing concentration of impurities from the injection end of said channel to the opposite end thereofimparting a drift to said minority carriers along the axis ofsaid channel and away from the point of injection such that the natural drift velocity of the injected minority carriers within the channel is substantially increased thereby permitting a corresponding increase in sensitivity for the transducer.
2. The transducer of claim 1 including a lateral variation in impurity concentration having a saddle-shaped concentration provided by an impurity distribution of gradually decreasing concentration from the channel sides towards its axis.
3. The transducer of claim 1 including at least one other region for collection of minority carriers, said other collector region also being contiguous with a side of said channel and spaced from said one collector region.
4. The transducer-amplifier of claim 3 wherein said collector regions extend around said channel and are spaced apart a short distance at their ends, and said collector regions are biased electrically against said channel to an extent that the space charge layers between said collectors and channel extend across the spacing between said collectors so that the channel is fully surrounded by said collectors and their space charge regions.
5. A semiconductive magnetic transducer-amplifier comprising in combination a body of semiconductive material; a semiconductive channel disposed within said body; at least one collector region ofsaid body adjoining said channel and of opposite conductivity type from said channel; means to inject minority carriers into said channel from an emitter located at said surface at one end of said channel in order that minority carriers moving along said channel can be deflected to said collector by a magnetic field applied to said channel; and an enhancement MOS transistor disposed adjacent said channel with its inversion layer in series connection therewith, and said collector region being coupled to the gate of said transistor and biased in accordance with the flow of minority carriers into said collector such that said MOS transistor operates as a source of majority carriers to said channel.
6. The transducer of claim 5 wherein said MOS transistor includes a first region of the same conductivity type as said collector region, said first region spaced apart from said collector and disposed in said channel at the opposite end from said injection region, said first region having a majority source region disposed therein, said majority source region being spaced away from said collector region, said transistor having a gate electrode overlying and insulatively spaced from said first region, said gate overlying a portion of said first region which is disposed between its source region and said collector region, said collector region coupled to said gate electrode, and said minority flow into said collector providing a bias of said gate electrode which inverts the underlying region such prising in combination a body of semiconductive material; a
semiconductive channel of opposite conductivity type from said body and disposed within said body so as to be contiguous with said body along a substantially planar surface; a collector region of opposite conductivity type from said channel and contiguous with the outside planar surface of said channel,
means to inject minority carriers into said channel from an emitter located at the surface at one end of said channel; a source of majority carriers located at the surface at the other end of said channel; circuit means to provide a flow of majority carriers along said channel from said source of majority carriers toward said emitter of minority carriers, thereby generating a drift field along said channel so that minority carriers drifting along said channel are deflected to said collector by a magnetic field applied to said channel; and an emitter region within said collector and of opposite conductivity type from said collector, which together with said collector and said channel forms a bipolar transistor such that the output of said transistor amplifier operates as a source of majority carriers for said channel and is proportional to minority carrier flow deflected to said channel.
9. The transducer-amplifier of claim 8 including means for controlling the injection of minority carriers into said channel in accordance with said majority carrier flow thereby providing a bistable device having one state characterized by a comparatively large minority carrier flow and another state of little minority carrier flow in said channel.
10. The transducer of claim 9 including a magnetic field in said channel, said magnetic field directed across said channel so as to control minority flow into said collector, and means for varying said magnetic field to control said minority flow into said collector and thereby switch said bistable device from one state to the other.
11. The transducer-amplifier of claim 9 wherein said channel includes a source and a drain of majority carriers, said majority carrier drain being located at the injection end of said channel, said injection means being disposed between said drain and said collector, and said injection means being conductively connected to said drain so as to control minority carrier injection in accordance with majority carrier flow.