EP0836232A1 | 1998-04-15 |
SIMON U: "CHARGE TRANSPORT IN NANOPARTICLE ARRANGEMENTS", ADVANCED MATERIALS, vol. 10, no. 17, 1 December 1998 (1998-12-01), pages 1487 - 1492, XP000790419, ISSN: 0935-9648
MATSUMOTO K ET AL: "ROOM TEMPERATURE OPERATION OF A SINGLE ELECTRON TRANSISTOR MADE BY THE SCANNING TUNNELING MICROSCOPE NANOOXIDATION PROCESS FOR THE TIOX/TI SYSTEM", APPLIED PHYSICS LETTERS, vol. 68, no. 1, 1 January 1996 (1996-01-01), pages 34 - 36, XP000548791, ISSN: 0003-6951
PATENT ABSTRACTS OF JAPAN vol. 095, no. 011 26 December 1995 (1995-12-26)
1. | A quantumsize electronic device comprising electrodes, at least one cluster and a tunnelly transparent gap, characterised in that the cluster has at least one distinguished size found from the formula r = a. rO, wherein r. is the circular radius of an electron wave under the formula rO = h/(mea2c), wherein h is the Planck's constant; meis the electron mass; a = 1/137036 is the fine structure constant; c is the light speed; a is a coefficient determined in the range of 1 < a < 4, the thickness of the tunnelly transparent gap being not more than ro, the spacing between the electrodes being more than ro. |
2. | A device according to claim 1, characterised in that the said cluster is made of metal. |
3. | A device according to claim 1, characterised in that the cluster is made of semiconductor. |
4. | A device according to claim 1, characterised in that the cluster is made of superconductor. |
5. | A device according to claim 1, characterised in that the cluster is made of a high molecular organic substance. |
6. | A device according to claim 1, characterised in that the cluster is made in the form of a cavity having a sheath of a tunnelly transparent gap. |
7. | A device according to claim 1, characterised in that the cluster has a centrally symmetric form. |
8. | A device according to claim 1, characterised in that cluster has an axisymmetric form. |
9. | A device according to claim 8, characterised in that the cluster is made extended and has a distinguished crosssectional size determined by the formula d=bro, 2<b<4. |
10. | A device according to claim 9, characterised in that the cluster is made extended along the axis and has a regular structure with a period determined by the formula T = b ro, 1 <b<4, 1. |
11. | A device according to claim 1, characterised in that a plurality of clusters is regular located at least in one layer, and the spacings between clusters are tunnelly transparent not exceeding ro. |
12. | A device according to ciaim 1, characterised in that the clusters are connected to at least two electrodes, one of which is a control electrode. |
13. | A device according to claim 1, characterised in that the clusters are connected to at least three electrodes, at least one of which is a control electrode. |
14. | A device according to claim 1, characterised in that the electrodes are made of metal or/and semiconductor, or/and superconductor or/and conductive organic materials. |
15. | A device according to claim 1, characterised in that the clusters are integrated into groups, forming thereby onedimensional or/and twodimensional or/and three dimensional structures. |
16. | A device according to claim 15, characterised in that the clusters are integrated into groups by means of reciprocal location of discrete electrodes. |
17. | A device according to claim 16, characterised in that the clusters are integrated into groups by means of reciprocal location of discrete electrodes and the form thereof. |
18. | A device according to claim 1, characterised in that the clusters are integrated into insulated spacial groups that are connected to corresponding electrodes. |
19. | A device according to claim 1, characterised in that the electrodes are made of superconductor and have a crosssectional dimension d 2 2 ru. |
20. | A device according to claim 1, characterised in that the electrodes are made of material having a metalsemiconductor phase transition, and have a crosssectional dimension d 2 2 ru. |
21. | A device according to claim 1, characterised in that the clusters are connected to at least two control electrodes, a set of such clusters forming a storage cell matrix. |
22. | A device according to claim 1, characterised in that two or more series clusters are connected to at least two control electrodes, a set of such clusters forming a storage cell matrix. |
23. | A device according to claim 1, characterised in that two or more clusters are connected to at least two supply electrodes at least through one resistive layer. |
24. | A device according to claim 23, characterised in that two or more clusters are connected to supply electrodes and integrated in a group in the form of one layer of directly mutually contacting clusters, one or more clusters being connected to control output electrodes, and the other cluster or clusters being connected to output electrodes, forming thereby the output of the gate"OR". |
25. | A device according to claim 23, characterised in that two or more clusters are integrated in a group in the form of a onedimensional series circuit, the even elements of the said circuit are connected through resistive layers to the first supply electrode and the odd elements are connected through resistive layers to the second supply electrode, the said odd elements forming a logical shift register. |
26. | A device according to claim 1, characterised in that two or more clusters are connected to supply electrodes and integrated in a group in the form of one layer, one or more clusters are connected to control input electrodes, the other cluster or clusters are connected to output electrodes, the input and output clusters are joined together through additional electrodes having the same thickness and width, said electrodes can be connected to one or more clusters of the next group. |
27. | A device according to claim 1, characterised in that two or more clusters, are connected to supply electrodes and integrated in a group in the form of one layer, one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to the output electrode, input and output clusters are joined together through additional electrodes tapered at one side in the signal direction, the said electrodes can be connected to one or more clusters of the next group. |
28. | A device according to claim 1, characterised in that the input voltage is supplied directly through one or more control electrodes connected to clusters through the tunnelly transparent gap, the cluster is connected through a resistive element to the supply voltage, the junction point is connected to the output electrode, forming thereby the output of the gate"NOT'. |
29. | A device according to claim 1, characterised in that two clusters are connected to the supply voltage through resistive elements, the first input voltage is supplied directly through the first control electrode connected to one cluster through tunneliy transparent gap, the second input voltage is supplied to the second electrode connected to the other cluster through the tunnelly transparent gap, some of junction points of the resistive elements of each cluster are joined together, the other junction points of the resistive elements are connected to output electrodes, forming thereby the outputs of the analogue comparator of two signals. |
30. | A device according to claim 1, characterised in that two clusters are connected through resistive elements to the supply voltage, the junction points thereof are connected to the output electrodes, the first input voltage is supplied directty through the first control electrode that is connected to the first cluster through the tunnelly transparent gap, the second input voltage is supplied directly through the second control electrode that is connected to the second cluster through the tunnelly transparent gap, the first output electrode is connected to the second cluster through the tunnelly transparent gap, the second output electrode is connected to the first cluster through the tunnelly transparent gap forming thereby a bistable trigger. |
31. | A device according to claim 1, characterised in that one and more clusters are connected through the resistive layer to the supply voltage and form isolated groups combined by one common output electrode, each isolated group of clusters is connected to one or more control input electrodes, the number of clusters in each group determines the weight function according to the input signal and forms a neuronetype logical componenta weight summator. |
32. | A device according to claim 1, characterised in that one or more clusters are connected to the supply electrodes at least through one additional cluster layer. |
33. | A device according to claim 32, characterized in that two or more clusters are connected to the supply electrodes and joined together in a group in the form of a single layer of directly together contacting clusters, one or more clusters are connected to control input electrodes, the other cluster or cluster are connected to output electrodes, forming thereby the output of the gate"OR"with memory. |
34. | A device according to claim 32, characterised in that two or more clusters are connected to supply electrodes and integrated in a group of in the form of a single layer, one or more clusters are connected to the control input electrodes, the other cluster or clusters are connected to the output electrodes, the input and output clusters are joined together through additional electrodes of same thickness and width, the said electrodes can be connected to one or more clusters of the next group. |
35. | A device according to claim 32, characterised in that two or more clusters are connected to supply electrodes and integrated in a group in the form of a single layer, one or more clusters are connected to control input electrodes, the other cluster or Clusters are connected to output electrodes, the input and output clusters are joined together through additional electrodes tapered at one side in the signal direction, the said electrodes can be connected to one or more clusters of the next group. |
36. | A device according to claim 32, characterised in that one and more clusters are connected through additional clusters to the supply voltage and form isolated groups combined by a single common output electrode, and each isolated group of clusters is connected to one or more control input electrodes, the number of clusters in each group determines the weight function according to the input signal forming thereby a neuronetype logical componenta weight summator with memory. |
37. | A device according to claim 1, characterised in that the cluster is connected through additional cluster to the supply voltage, a junction point is connected to the output electrode, the input voltage is supplied directly through one or more control electrodes that are connected to clusters through the tunnelly transparent gap, forming thereby an inversion gate with memory. |
38. | A device according to claim 1, characterised in that two clusters are connected to the supply voltage through additional clusters, the first input voltage is supplied directly through the first control electrode that is connected to one cluster through the tunnelly transparent gap, the second input voltage is supplied to the second electrode that is connected to the other cluster through the tunnelly transparent gap, some of junction points of resistive elements of each cluster are joined together and connected to the supply electrode through the resistive element, other junction points of additional clusters are connected to output electrodes that are outputs of the two signal analogue comparator with memory. |
39. | A device according to claim 1, characterised in that two clusters are connected through additional clusters to the supply voltage, junction points thereof are connected to output etectrodes, the first input voltage is supplied directly through the first control electrode, that is connected to the first cluster through the tunnelly transparent gap, and the second input voltage is supplied directly through the second control electrode, that is connected to the second cluster through the tunnelly transparent gap, the first output electrode is connected to the second cluster through the tunneily transparent gap, and the second output electrode is connected to the first cluster through tunnelly transparent gap, forming thereby a bistable trigger circuit. |
40. | A device according to claim 1, characterised in that two or more clusters are connected at least to two control electrodes, at least one of which is a light transparent electrode, spacings between clusters are filled with a photosensitive semiconductor and a set of such clusters forms a photosensitive matrix. |
41. | A device according to claim 1, characterised in that one or more layers of clusters are connected to at least two distributed electrodes, at least one of which is a light transparent electrode, the spacings between clusters are filled with photosensitive semiconductor and form a lightsensitive storage medium. |
42. | A device according to claim 1, characterised in that one or more layers of clusters are connected to at least two distributed electrodes made in the form of a resonator, forming thereby a highfrequency generator with a maximal boundary frequency determined by the formula f: f:. |
43. | A device according to claim 1, characterised in that one or more clusters are joined together by direct contacting or through electrodes and connected to the current source, at least one of the contacts is connected to the output electrode, allowing thereby to create a standard voltage source with levels U= na3c2me/2e, wherein n is a number of clusters connected in series. |
44. | A process of operating the devices according to claims143, comprising applying the electric field in the working range of strengths, characterised in that the magnitude of the field control strength at one cluster is determined in the range wherein E= ma/2eh, E= E47ra ; h is the Planck's constant; me is the electron mass; e is the electron charge; a = 1/137036 is the fine structure constant; c is the light speed. |
45. | A process of operating the device according to claim 44, characterised in feeding a continuos or/and pulse supply. |
46. | A quantumsized electronic device comprising electrodes and located between them a layer of the material having a metalsemiconductor phase transition, characterised in that the layer of the material having a metalsemiconductor phase transition is made in the form of clusters with crosssectional dimensions determined by the formula: r = a ro, wherein rO is a circular radius of a of an electron wave under the formula: rO= h/(anea2c), wherein h is Planck's constant, me is the electron mass, a = 1/137,036 is the fine structure constant; cis the light speed; a is a coefficient determined in the range 2 < a < 4, the distance between the electrodes being more than ro. |
47. | A device according to claim 46, characterised in that the cluster is connected to the supply electrodes and at least to one load, one or more control electrodes are connected through tunnelly transparent gap, the width of the tunnelly transparent gap is not more than ro and the distance between the electrodes is not less than ro. |
48. | A device according to claim 46, characterised in that the electrodes are made of superconductor and a crosssectional dimension d > 2 ro. |
49. | A device according to claim 46, characterised in that the said electrodes are made of material having a metalsemiconductor phase transition, and have a cross section dimension d > 2 ro. |
50. | A device according to claim 46 characterised in that one or more clusters are connected to supply electrodes at least through one resistive layer. |
51. | A device according to claim 50, characterised in that two or more clusters are connected to supply electrodes and joined together in a group in the form of one layer of directly together contacting clusters. one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to the output electrodes forming thereby the output of the gate"OR". |
52. | A device according to claim 50, characterised in that two or more clusters are connected to supply electrodes and integrated in a group in the form of one layer, one or more clusters are connected to the control input electrodes, and the other cluster or clusters are connected to the output electrodes, the input and output clusters are joined together through additional electrodes of the same thickness and width, the said electrodes can be connected to one or more clusters of the next group. |
53. | A device according to claim 50, characterised in that two or more clusters are connected to supply electrodes and integrated in a group in the form of one layer, one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to output electrodes, the input and output ciusters are joined together through additional electrodes tapered at one side in the signal direction, the said electrodes can be connected to one or more clusters of the next group. |
54. | A device according to claim 46, characterised in that the input voltage is supplied directly through one or more control electrodes connected to clusters through tunnelly transparent gap, the cluster is connected through the resistive element to the supply voltage, and the junction point is connected to the output electrode, that is the output of the gate"NOT". |
55. | A device according to claim 46, characterised in that two clusters are connected to the supply voltage through resistive elements, the first input voltage is supplied directly through the first control electrode connected to one cluster through the tunnelly transparent gap, the second input voltage is supplied to the second electrode connected to the other cluster through the tunnelly transparent gap, some of the junction points with the resistive elements of each cluster are joined, and the other junction points with the resistive elements are connected to output electrodes forming thereby the outputs of the two signal analogue comparator. |
56. | A device according to claim 46, characterised in that two clusters are connected through resistive elements to the supply voltage, the junction points thereof are connected to output electrodes, the first input voltage is supplied directly through the first control electrode connected with the first cluster through the tunnelly transparent gap, and the second input voltage is supplied directly through the second control electrode connected to the second cluster through the tunnelly transparent gap, first output electrode is connected to the second cluster through the tunnelly transparent gap, and the second output electrode is connected to the first cluster through the tunnelly transparent gap, forming thereby a bistable trigger. |
57. | A device according to claim 46, characterised in that one or more clusters are connected through the resistive layer to the supply voltage and form isolated groups joined together by one common output electrode, each isolated group of clusters is connected to one or more control input electrodes, the number of clusters in each group determines the weight function according to the input signal, forming thereby the neuronetype logical componenta weight summator. |
58. | A device according to claim 46, characterised in that two or more clusters are connected at least to two control electrodes, the gaps between clusters are filled with the photosensitive semiconductor, a set of such clusters forms a photosensitive matrix. |
59. | A device according to claim 46, characterized in that one or more layers of clusters are connected to at least two distributed electrodes made in the form of resonator, forming thereby a highfrequency generator with a maximal boundary frequency determined from the formula f s mea4c2/h. |
60. | A process for operating the devices according to claims 4658 that comprises transmitting an electric current at least through one cluster, characterised in that the current density through the cluster is limited by the value j < 4weme a c/h wherein h is the Planck's constant; me is the electron mass; e is the electron charge : a = 1/137,036 is the fine structure constant; c is the light speed. |
61. | A process for operating the devices according to claims 4657,59, that comprises transmitting an electric current at least through one cluster, characterised in that in case the cluster is made of materials having a temperature of the metal semiconductor phase transition higher than the operation temperature of the devices, t the the electric field strength should correspond to E mcV (2eh), wherein h is the Planck's constant; me is the electron mass; e is the electron charge; (x = 1/137,036 is the fine structure constant; c is the light speed. |
62. | A process for operating the devices according to claim 58 that comprises transmitting an electric current at least through one cluster, characterised in that in the cluster are used materials having a temperature of the metalsemiconductor phase transition higher than the operating temperature of devices. |
63. | A quantumsize electronic device comprising electrodes and located between them at least one cluster, characterised in that, the cluster is made of the material of the superconductor and the cluster has a crosssectional dimension found from the formula: r = a. ro, wherein ro is determined as the circular radius of a of an electron wave under the formula: pro = tr Imea2c, wherein h is the Planck's constant; me is the electron mass; a = 1/137,036 is the fine structure constant; c is the light speed; a is a coefficient determined in the range of 2 < a < 4, the distance between the electrodes being more than ro. |
64. | A device according to claim 63, characterised in that the cluster is connected to the supply electrodes and at least one load, and to one or more control electrodes through tunnelly transparent gaps, the thickness of tunnelly transparent gaps not exceeding rO. |
65. | A device according to claim 63, characterised in that the said electrodes are made of superconductor and have a crosssectional size d 2 2 ru. |
66. | A device according to claim 63, characterised in that the electrodes are made of the material having a metalsemiconductor phase transition, and have a crosssectional dimension d 2 2 ru. |
67. | A device according to claim 63, characterised in that one or more clusters are connected to the supply electrodes at least through one resistive layer. |
68. | A device according to claim 63, characterised in that two or more clusters, are connected to the supply electrodes and integrated in a group in the form of one layer of directly together contacting clusters, one or more clusters are connected to the control input electrodes, the other cluster or clusters are connected to the output electrodes, forming thereby the output of the gate"OR". |
69. | A device according to claim 63, characterised in that two or more clusters, are connected to supply electrodes and integrated in a group in the form of one layer, one or more clusters are connected to the control input electrodes, the other cluster or clusters are connected to the output electrodes, the input and output clusters are joined together through additional electrodes of the same thickness and width, the said electrodes can be connected to one or more clusters of the next group. |
70. | A device according to claim 63, characterised in that two or more clusters are connected to supply electrodes and joined together in a group in the form of one layer, one or more clusters are connected to the control input electrodes, and the other cluster or clusters are connected to output electrodes, the input and output clusters are connected together through additional electrodes tapered at one side in the signal passing direction, the said electrodes can be connected to one or more clusters of the next group. |
71. | A device according to claim 63, characterised in that the input voltage is supplied directly through one or more control electrodes connected to clusters through the tunnelly transparent gap, the cluster is connected through the resistive element to the supply voltage, and the junction point is connected to the output electrode, forming thereby the output of the gate"NOT'. |
72. | A device according to claim 63, characterised in that two clusters are connected to the supply voltage through resistive elements, the first input voltage is supplied directly through the first control electrode, that is connected to one cluster through the tunnelly transparent gap, the second input voltage is supplied to the second electrode, that is connected to the other cluster through the tunnelly transparent gap, some of junction points with the resistive elements of each cluster are joined together, and the junction points of connection to resistive elements are connected to the output electrodes, forming thereby the outputs of the analogue comparator with two signals. |
73. | A device according to claim 63, characterised in that two clusters are connected through resistive elements to the supply voltage, the junction points thereof are connected to the output electrodes, the first input voltage is supplied directly through the first control electrode that is connected to the first cluster through the tunnelly transparent gap, the second input voltage is supplied directly through the second control electrode, that is connected to the second cluster through the tunnelly transparent gap, the first output electrode is connected to the second cluster through the tunnelly transparent gap, and the second output electrode is connected to the first cluster through the tunnelly transparent gap, forming thereby a bistable trigger. |
74. | A device according to claim 63, characterised in that two or more clusters are connected through the resistive layer to the supply voltage and form isolated groups joined together by one common output electrode, each isolated group of clusters is connected to one or more control input electrodes, the number of clusters in each group determines the weight function according to the input signal, forming thereby a neuronetype logical componenta weight sumator. |
75. | A process for operating the devices according to claim 6374, characterised in that the operating mode of the devices is limited by the critical temperature of transition into the superconducting state of the used materials, that is determined from the formula T, < mea3c2/(2kw), wherein k is the Boltzmann's constant; me is the electron mass; a = 1/137036 is the fine structure constant; c is the light speed. |
76. | A process for operating the devices according to claim 75, characterised in that the transition from the superconducting state into the normal sate under the effect of the control voltages takes place at the cluster in presence of the strength E > me2a5c3/ (2he), wherein h is Planck's constant. |
77. | A quantumsize electronic device, comprising electrodes at least one of which is made of superconductor or of material with the metalsemiconductor phase transition, characterised in that the electrodes have a crosssectional dimension determined from the formula: d = nd r0 wherein ro is determined as a circular radius of a of an electron wave under the formula: rO = hI (mea2c), wherein ri is the Planck's constant: me is the electron mass; a= 1/137036 is the fine structure constant; c is the light speed; nd is a coefficient determined in the range 1 < nd < 2. |
78. | A device according to claim 77, characterised in that a group of electrodes has at least one region of approach to the distance not exceeding ro, forming thereby a transformer of direct or alternating current. |
79. | A device according to claim 77, characterised in that, two electrodes have at least two regions of approach to the distance not more than ro, a direct current flows through one of the electrodes, and the second electrode is connected to the load, alternating current flows through the load, forming thereby a converter of direct current to alternating current. |
80. | A device according to claim 77, characterised in that two electrodes have at least two regions of approach to the distance not more than ro,, the direct current flows through one of the electrodes, and the second electrode is connected to the load, the direct current flows through the load, forming thereby a converter of direct current to alternating current. |
Such elements working on basis of quantum size resonance effects are used for constructing two-dimensional (planar) and three-dimensional electronic devices, designed for processing and converting of analog and digital information.
2. Background of the invention IC elements tend to scale down. However at downsizing of IC elements to less than 100 nm, charge carriers start revealing the discreet nature and the quantum mechanical characteristics thereof, what makes influence on constructive features of active devices, i. e. transistors.
At the same time at the dimension of less than 100 nm separate transistor elements actually are small particles, i. e. clusters [1]. Downsizing a cluster may create a condition allowing to design devices that are able to control groups of electrons, and even one electron.
Prior art describes a large class of electronic devices basing on single-electron tunnelling through a small size cluster [2]. The simples variant of such a device is a kind of analogue of a field-effect semiconductor transistor comprising between the drain and the source thereof an isolator with a built-in small cluster in the centre. Such a transistor is generally referred as SET- (Single Electron Transistor).
A cluster built in the isolator of a SET device has its own capacity in relation to the substrate Cc. The core of the effect disclosed in [2] is that during tunnel passage through the cluster of an electron with e-charge, the electron changes the potential of the cluster by the magnitude AU=e/Cc and blocks by its field the passage of other electrons for a while it is present at the cluster. In the process it is necessary that the potential at the cluster exceeded the potential of the thermal noises of the cluster capacitance: AU22kT/e (1) wherein k is the Boltzmann's constant, T is an absolute temperature.
For example a spherical silicon cluster with a radius rc = 5 nm having a <BR> <BR> <BR> <BR> dielectric permeability s = 11.7, will have the capacitance CC=47tcoErc and, hence on basis of (1) will have the maximum operating temperature of the device <BR> <BR> <BR> <BR> T = e 2/ (8z£o£rck) = 143K (-130°C) (2)<BR> <BR> <BR> <BR> <BR> <BR> <BR> wherein £o is the vacuum dielectric constant.
This condition shows that use of materials with s< 5.6 or clusters of the smaller size, generally provides a possibility of designing a single-electronic quantum device, operable at the normal temperature-290-300K (17-27°C). However there is no physical sense in considering a separate cluster as a microcircuit component without taking into consideration the capacitance of transistor electrodes. All stray capacitances of a SET device are considered further.
As it is disclosed in [3] a field semiconductor transistor with the isolated gate may register a single electron. In this case the structure of the proper transistor canal does not influence the analysis. Therefore for any devices of this kind, including nanometer devices, it is necessary to consider the input capacitance Ci as well as the output capacitance Ca. Thus, the multiplier Ca/Cj, should be added to the formula (1) in accordance with [3, formula (7.36)] (Ca/Cj) (e/Cc) 2 2kT/e (3) From this expression follows that if the entrance control signal is present on the gate or cluster and the conductor has an admissible size, e. g. the conductor length is about-1000 nm and the conductor width is about 10 nm, the conductor capacitance for a silicon substrate will be C ;. Accordingly, at the acceptable speed the operating temperature of the device is in all T=1.43°K (-271.72°C). Right this temperature is the iimit for the most of known SET-devices. [4-7]. The said researches, which describe approaches to realising high temperature single electronic tunnelling, in fact made use of one and the same method. For example, metal clusters of a size less than 50 nm were placed between two electrodes applied to a dielectric [4] or similarly, fullerene clusters of a size at all 0.634 nm were regularly spaced in the dielectric layer [5]. Various logical devices for designing the digital memory with logical elements of a size from 0.2 nm to 100 nm are investigated in [6,7]. In a number of other researches [8] the more traditional methods make use of building-in a cluster in a gate insulator of a field transistor. Charging and discharging the said cluster yet
by a group of electrons tunnelling through the dielectric (insulator) provides a possibility to change the characteristics of the field transistor so that to create an analog or digital memory. However the time of charge storage is insufficient in this case.
It is obvious from the descriptions of the aforementioned patents that space capacitances of conductors that connect transistors were not considered there. And naturally, operating temperatures exceeding the temperature of fluid helium were not obtained there.
The research [9] corresponds to a certain progress in the field of increasing operating temperatures of SET devices to normal conditions. The authors placed a 30-nm titan cluster between titanium electrodes of the 3-nm thickness spaced at a distance of 50 nm. The gap between the cluster and electrodes was filled with tunnelly transparent dielectric of TiOX. Supplying at normal temperature a small voltage of 0.1- 0.7 V produces four N-shape regions in voltage-current characteristics. This extraordinary effect was explained by a single electron tunnelling. Meanwhile, being aware that the titanium oxide has s = 24 and taking additionally into consideration capacitance of the cluster and electrodes respective the substrate it is obvious that the operating temperature should be well below normal. It is clear that the researchers faced the effect that might be caused by extraordinary characteristics of the proper dielectric TiOX film.
In fact, all dielectrics to a more or less extent have non-liner regions, the specific resistivity of which depends on the electric-field intensity. At the starting section of this response the specific resistivity does not change up to the electric-field <BR> <BR> <BR> <BR> intensity of 104 V/cm. The specific resistivity further decreases on account of creation of additional carriers released from traps-donators [10, c. 264]. If a dielectric is a high- molecular compound, the current flows through it along spherolites, i. e. certain <BR> <BR> <BR> <BR> channels formed by long moleculars. Excess of the value 105-106 V/cm generally causes irreversible breakdown of the dielectric i. e., mass transfer and destruction of molecules start directly in breakdown paths. The volume content of donators of thin- film dielectrics is not sufficient for formation of breakdown paths, therefore film breakdown takes place at the electric-field intensity of a greater order. For example, f <BR> <BR> <BR> <BR> a Si203 film of a thickness of about 15 nm has the electric-field intensity not more than 8 MV/cm. If the number of traps-donators is sufficient the dielectric has an ability to store the charge that had passed through the said path. This ability to store a charge
is widely used in electronics for designing a reprogrammed memory. However, the said memory operates with a large number of electrons, accumulated in numerous traps of different energy characteristics. This causes a constant charge leakage from the traps, and consequently, changes device characteristics. Therefore they can not be used in nanosized devices operating with single traps.
The further significant characteristic of dielectrics is the avalanche discharge of a dielectric. The function of the said discharge is to limit the output signal. In this case the discharge proceeds without any destruction of a material, for example in wide-gap semiconductors, designed in the form of ZnO multilayer polycrystalline films. The size of crystallites-clusters in these films is 0.2-15.0 , m. They are divided by Bi203 tunnelly transparent gaps of the thickness 2.0-10.0 nm [11]. Meanwhile the researchers do not disclose the nature of changes of film characteristics upon minimising crystallites to nanosizes, i. e. to less than 0.1 pm. Moreover, such output signal stoppers lack any amplifying characteristics, what limits the field of their use.
It is known that the classes of elements having N-and S-shaped characteristics allow amplifying and non-liner transforming a signal. It is common knowledge that N-shaped characteristics are found in the devices in which electron drops-domains have been formed. Generally S-shaped characteristics appear due to generation of current paths [12,13]. However, the provided characteristics of non- linear elements are drawn as a rule for samples of a micron and a larger size, what prevents from mechanical applying such characteristics to nanosized samples.
Moreover, these non-liner characteristics are specific only for two electrode devices, what limits the field of their use in nanoelectronics.
Important kind of non-liner characteristics are hysteresis loops based on the josephson effect in superconductors as well as similar hysteresis characteristics [14].
However, devices based on the josephson effect and other tunnel effects between superconductor and semiconductor, superconductor and metal and etc. are controlled by applied current or applied magnetic field. Designing current sources for control of josephson devices require a rather high voltage, resulting in overall energy loss.
Moreover inducing of an applied magnetic field requires coils or loops produced by means of lithography what renders this approach rather bulky. Available superconductors, that may be used in josephson devices have critical temperatures not more than-182°C, what requires use of cryostats and enlarges overall dimensions of devices. All this renders problematic the use of these devices in nanoelectronics.
Similar non-linear characteristics have a number of amorphous and polycrystalline films of semiconductors [15] or materials, having metal-semiconductor phase transition (MSPT) [16], including MSPT basing on high-molecular organic semiconductors (BEDT-TTF) mXn [17] or basing on Lengmurr-Blodgett films of stearic acid in the form of a molecular single electronic transistor [18].
Processes of electron passing in lengmur films having built-in nanosized clusters were controlled by the orthogonal needle CTM [18]. Naturally, this geometrical configuration makes the capacitance of the control electrode-needle substantially lower than in the formula (3) what allows to observe quantum effects at normal temperatures and low speed. However the capacitance grows when a control electrode is applied to the substrate, and at acceptable speed the device may work only at low temperatures.
Elements having characteristics with hysteresis loops allow to store information, i. e. to design memory cells. Information in the cells is recorded by means of electric current. Moreover, there are hysteresis characteristics of magnetic materials, which may be similarly used for designing memory cells. Information in these cells may be recorded by means of additional external fields. For example [19] reports on a variant of writing information into a single cluster by means of turning the electronic spin under the action of photons in the magnetic semiconductor films of a nanometer thickness in the system PbTe-EuTe-PbTe.
The superior magnetic material such as SmCo has the magnetic field stored energy of not more than 5 J/mole. Thermodynamic analysis for this material show that the minimum size of the magnetic cluster meeting reasonable requirements for information storage for more than a year at a normal temperature should be above 100 nm. Accordingly, the device of [19] is actually limited by these dimensions.
Therefore magnetic materials are not yet perspective for nanosize devices.
Further approach to designing of active nanosize quantum devices [12] is based on producing a kind of atom-like device-a superatom of a spherical form made of semiconductors according to the superlattice technology. Here spherical layers of superlattices surround the nucleus of a size 3-10 nm. The overall diameter of such cluster is 71 nm. The electrons in the device move atong surfaces of superlattices around the charged nucleus. However, the electron bound energy in such"envelopes"is about what requires accordingly helium temperatures.
Thus, such an approach to designing electronic nanosize devices for normal temperatures is not the future-technology.
Meanwhile among the aforementioned examples at least only a metal- semiconductor phase transition (MSPT) is supported by the axiomatic theory that approximately describes non-liner S-shaped volt-amps diagrams. This theory is based on thermodynamic instability and presence of hysteresis in the metal- semiconductor transformation point on the account of change of the crystalline structure. Persistence of all thermal processes in such S-devices at MSPT renders them unpromising for use in microelectronics and all the more so in nanoelectronics.
Non-liner characteristics advantageous for use in active nanoelectronic elements often appear with decreasing device dimensions. For example a tunnel current develops between electrodes when the thickness of the dielectric located between electrodes decreases to less than 8 nm [12, p. 93]. This current is described as developing due to probable tunnelling of electrons through the energy barrier of a predetermined form. However substantial abnormalities developing on the barrier at small voltages of 1-300 mV [13, p. 371] can not be described by probabilistic approaches. Moreover the critical size of the barrier of-8 nm is not evident from any theory.
N-shaped characteristic develops in semiconductor diodes with a high doping level, namely tunnel diodes. This characteristic is described by tunnelling of carriers in a semiconductor through p-n junction. A normal width of p-n junction in such a diode subject to voltage is 10-15 nm, and the electron wavelength is not more than 3 nm.
Hereupon the tunnel effect should not be observed under the classic theory [20, p. 349]. In fact volt-amps diagrams of tunnel semiconductor diodes have a valley that is claimed to the tunnel current appearing upon interaction of carriers with phonons and photons of the semiconductor of the p-n junction lattice. Nevertheless neither models available in this case are able to describe such abnormalities at volt-amps diagrams as residual stored current in the volt-amps valley and rising a bulge at volt- amps diagrams at the additional doping of the junction and some other abnormalities [13].
As it is made apparent in this chapter the state of art of available models of SET devices and other non-liner devices considering all additional factors do not allow to reckon and, consequently, to design a high temperature logical circuit. Use of thermodynamic models for reckoning causes great problems since they do allow
computing the concrete speed of devices. The known models that describe tunnelling of electrons with N-shaped characteristics do not express numerous features of volt- amps diagrams and do not allow determining design requirements for nanosized devices.
Designing integrated circuits basing on quantum size electronic devices still has a current interest to the problem of galvanic uncoupling of single microcircuit parts, i. e. creation of certain quantum size transformers.
Thus, nanoelectronic has faced a problem of what physical principles are to be applied to designing active elements and what the physical limitations are to a size, speed and operating temperature of a device.
Meanwhile these problems may be solved only by creating more precise quantum-mechanical models and designing on the basis thereof very large scale ICs for processing information at normal temperature.
3. Summary of the Invention.
The aim of the invention is increasing operating temperatures of quantum size devices, that control one two or more electrons passing through such a quantum size device and that have the extreme accessible speed at minimum permissible dimensions. Such devices may be fabricated by means of two-dimensional and three -dimensional technology.
Theoretical investigations and analysis of experimental evidences resulted in designing a model of interaction of electrons in condensed mediums. The model provides a rather accurate correspondence to the experimental evidences. According to the said model, in condensed mediums electrons may occupy certain stable states with minimum energy and a low profile of interaction with atoms of the medium. The way it may be implemented is shown below.
At present the only way to increase the operating temperature of a device is to reduce the cross section of interaction of an electron with short-wave phonons and infrared photons of the substrate and the proper material of the device.
At the same time it is a common knowledge that the cross section of interaction of a free electron is close to the classical radius thereof r1 = ah/mec. The cross section of Compton scattering of a gamma quantum at an electron in condensed mediums gives an electron radius r2 = h/mec, and a cross section of scattering of a
hydrogen atom is equal to Bohr radius r3= hImeac. Here h=h/27t is the Planck's<BR> <BR> <BR> <BR> <BR> <BR> constant, a=1/137.036 is the fine structure constant, me is the free electron mass and c is the light speed.
One of the basic theorems of differential geometry teaches that any space may be discomposed only to embedded tori (anchor rings), or in a particular torus case- to embedded spheres [21]. Let us select the step of space decomposition at torus 1/a . Basing on this theorem we postulate that in condensed medium the maximum size of an electron wave is: rO=/(mea2c)(mea2c) = 7.2517 nm (4) At the said size and speed of the wave motion a2C, the electron will have the minimum possible energy in condensed medium.
Basically this assumption meets the original idea of de Broglie (1924). He assumed that a particle is similar to a torsion pendulum, i. e. it has internal vibrations in a phase with its space wave motion. Later together with co-authors, de Broglie developed this model by presenting an electron as a vortex ring [22]. Other authors who presented an electron as a torus [23,24] used similar models. However their models have the size of the major radius of torus not more than the Compton radius thereof, r2 = a2ro, and the minor radius of torus tend to zero.
Let us extend this model. For a spaced layer with the radius ro we shall introduce the major torus radius equal to ro, and the minor torus radius will be limited by a classical electron radius r1 = a3ro. Such formation is the only closed oriented two- dimensional surface that allows a vector field non-singular in every point. As it is evident from the differential geometry, no other topology allows equilibrium of an isolated system with a self-action in the form of the like charged medium.
Considering the aforementioned the electron ring wave with equally distributed charge e will get the termin"ring electron". Extremely important is the fact that due to axial symmetry of the distributed rotating charge, such a ring emits neither electromagnetic nor gravitation waves, i. e. it has an absolute stability.
While presenting an electron as a point charge orbiting around a nucleus in the Bohr model or a certain charge-distributed rate in the Schroedinger model we must postulate the stability of the charge by introducing discrete energetic levels, at which
the electron does not emit electromagnetic waves. In our case the stability of the electron is automatically conditioned by the geometry thereof.
Experimental evidences of rightfulness of such handling of size and form of an electron are provided below.
Quantum size effects occur in various condensed mediums, e. g. the quantum Hall effect in thin semiconductor layers at low temperatures [25]. Here the density of the allowed state at Landau levels is equal to the quantum density of the magnetic flow n4 =l/2zr42, where r4 z 7 nm is a so-called magnetic length discretely relating to the radius of the electron orbit for the lowest Landau level, i. e. here electrons are presented as thing ring-like waves with intervals between rings z-2-r4. In their turn the rings are located in one plane.
At normal temperatures there was registered a characteristic formation, the size of which was determined by the size of an electron ring. The formation emerges at mechanical interaction of two planes in the 0.1 M HCI electrolyte. This formation has a size of an order 7.5 nm [26, p. 170]. Moreover it is exceptionally rigid. In the course of a number of experiments it was observed that formations of a similar size and rigidity usually emerged at the initial phase of the transition of a substance from a liquid-phase to a solid-phase [27].
The suggested electron model with a radius ro allows explaining the anormal effects occurring in metal-dielectric junctions in a rather simple way without referring to probabilistic models. If we conceive an electron as a certain ring with a radius ro, then such a ring may easily cross a potential well of a smaller size, e. g. less than 8 nm. Such an utterly simplified mechanic explanation bears a fundamental meaning, which is not connected with a particle regarded as a certain density of probability distributed in the space. And in this case it is not necessary to conceive the particle tunnelling through any potential barrier.
Using a ring electron model one may describe all main features of current- voltage curves of tunnel semiconductor diodes. It is possible to conceive that formation of clusters with the radius ro is possible in specific oversaturated solid solutions of heavily doped semiconductors. This cluster acts as a nucleus and it is surrounded by a solvating sphere of a less doped semiconductor, i. e. here a kind of pseudoatom with a tunnel transparent single-layer or multilayer sphere of a thickness not more than ro.. This results in creating a bulk formation with a total diameter of dx
4ru~29 nm. With such a structure there is a probability of formation of an electron ring that is moving between the sphere and the surface of a nucleus totally environing the nucleus. Once being formed such an electron may be presented in a form of a current ring, characteristics of which may be calculated.
It is a common knowledge that a thin ring current of a radius ro with a charge e generates on the x-axis, an electric and a magnetic field, accordingly [28]: E = (e x/4Z£0) (ro 2 + (5) H = (I ro2/2) (ro2 + x(6) wherein I = a-ec/2zrO is a ring current. It follows from this expressions that on the axis x E-field has the maximum potential at the distance from the centre of the ring ru/-, and the H-field has no maximum characteristics. Accordingly, at the distance <BR> <BR> <BR> <BR> from the ring centre ro/V2 the other free electron will be electrostatically attracted to the ring. And besides, in the centre of the cross section plane of such a ring the electric filed density is equal to null. Because of this in the centre of the ring there is formed a potential well having at its bottom point the energy of interaction with the point charge equal to null. In the process of passage of an electron in the crystal, some ions of crystal lattice core occur into the electron potential well. It results in reducing the energy of ion interaction with the ring electron at least by < a/2X. In this case interaction of the ring electron with all other surrounding charges will be mainly determined by the part of the electrostatic field tying beyond the square surrounded by this ring. It is possible to show that the value of this filed will be of an order ae.
If a ring is placed into the external magnetic filed B, then the precession <BR> <BR> <BR> <BR> frequency thereof will be (Og = Bae/me. It follows from this equation that the effective<BR> <BR> <BR> <BR> <BR> <BR> <BR> electron mass is m* = 137.036 me. Therefore reduction of the cross section of electron interaction with fluctuations of lattice ions (with phonons or infra-red photons) may be regarded as an increase of the electron effective mass, and accordingly, the decrease of the space shift, imparted to it by a phonon (or IR photon). The reduction of interaction cross section may also be interpreted as the reduction of the Coulumb interaction between the lattice charge and the proper electron by factor of ae. In consequence of the aforesaid the electron looks like being"senseless"towards characteristics of the medium, which it is passing.
Vivid proofs of existence of heavy electrons are superconductors with f- electronic systems. For example for systems of the type UBe13UPt3 M*-= 137Me [29]. It will be noted that for semiconductors m* < me, for metals m* _ me. At normal temperatures it is possible to find heavy-electron systems. For example in materials having a VO2 type metal-semiconductor phase junction the effective electron mass is m*-60 me [16, p. 33].
Thus, the validity of the proposed theoretical model of an electron ring in a condensed phase is fully founded and supported by independent experiments.
However occurrence of such electron is possible only under specific external impacts, for example temperature, high forced external field, and other transient processes.
Owing to this the phenomenon can not be registered by standard measurements in stationary conditions, such as registering an electron mass in semiconductors.
One of such non-stationary states in semiconductors occurs at pulse lightening thereof. With this are formed bounded states electron hole-exitons. They are usually described by a Bohr model with the radius rS=£r3/m* where m5*-is the equivalent exiton mass. However, the multiplier 8/mu* may be presented be means of a model of an electron ring. As the interaction cross seduction of the electron ring <BR> <BR> <BR> <BR> with the lattice goes down as ae, then £ ~ 1, and m3* ~ me/a. Accordingly, the exiton radius r5 should not be more than ro. With this the exiton energy will not exceed <BR> <BR> <BR> <BR> (axis W3 = 15.8 meV, where W3 = me (ac) 2/2 is the energy of the main level of the Bohr atom. In this case the spaced three-dimensional condensation of electrons into <BR> <BR> <BR> <BR> exciton droplets should have the concentration N5 < (ro/)-3=7. 42101g cm-3. The densest droplets appear in Si. They have Ns= (3.0. 3 37) 1018 cm~3 and the energy of bounding excitons into droplets W5 = 8.2 meV, what fully meets the aforementioned limits and agrees with experimental data described in [30].
Such a big exciton with the radius of an order ro is generally named as the Wannier-Mott exciton. Experimental data show that when the exciton size diminishes to 0.1-1.0 nm it is transformed into the Fenckel exciton [30]. In this case electron ring with the radius ro will simply roll up to the size of the period of lattice of the atom frame, and the ring speed will increase in the order of the speed on the Fermi surface.
Electrons of Be will have the maximum speed value on the Fermi surface. The said value does not exceed the magnitude of ac.
Thus, the proposed theoretical model of an electron ring allows, without using any probability models, a new approach in describing most of time-varying and non- linear processes occurring in condensed medium.
It follows from the foregoing analysis that in certain materials it is possible to induce a condition of formation an electron ring by means of an external action and/or by nanostructuring of a medium. By that are provided resonance conditions for operating nanoelectronic devices, which conditions allow their functioning at normal and higher temperatures. This model has become a basis for designing a number of new devices with new operation modes in according with the further going specification and the attached claims.
The essence of the invention is as follows.
In accordance with one embodiment of the invention a quantum size electronic device comprising electrodes, at least one cluster and a tunnelly transparent layer is characterised in that the cluster has at least one distinctive size, determined from the formula: r = a rO, wherein ro is a (ring) radius of an electron wave under the formula: wherein h is the Planck's constant, me is the electron mass, a is the fine structure constant = 1/137,036, c is the light speed, a is a coefficient, determined in the range of 1 < a < 4.
With this the thickness of the tunnelly transparent layer does not exceed ro, and the distance between electrodes is not more than ro.
According to the invention a cluster may be made of metal, semiconductor, superconductor, high-molecular organic material. Besides, it may be also made as a cave with an enclosure in the form of a tunnelly transparent layer.
In a number of embodiments the cluster has a central-symmetric form.
Under a still further embodiment a cluster may have an axisymmetric form, and also may be made extended and have a distinctive cross section size determined by the formula
d=bro, 2b4, In a further embodiment an extended cluster may have along its axis a regular structure with a period determined by the formula T = b rO, 1 < b < 4, In accordance with further development of the invention plurality of clusters may be regular arranged at least in one layer, and besides the gaps between the clusters should be tunnelly transparent and not exceed ro.
At least two electrodes should be connected to clusters, one of the electrodes being a control one.
Clusters may be also connected to at least three electrodes, one of which is a control electrode.
According to further development of the invention electrodes may be made of metal and/or semiconductor, and/or superconductor, and/or conducting organic materials.
Clusters may be also united into groups and form one-dimensional and/or two- dimensional and/or three-dimensional structures.
Arrangement of clusters into groups may be performed by means of mutual location of discreet electrodes, and also by the form of discreet electrodes.
In a further embodiment of the invention clusters may be arranged into isolated space groups, which are connected to the corresponding electrodes.
In the event of using electrodes of superconductor, the cross section size of electrodes should be limited by the size d 2 2 ru According to another embodiment electrodes are made of the material having a MSPT and a cross section size d 2 2 ru Each cluster may be also is connected to at least two control electrodes, an ensemble of such clusters forming a memory cell matrix.
Two control electrodes may be also connected to at least two or more clusters and an ensemble of such clusters forms a memory cell matrix, capable of storing information even at de-energising.
Clusters may be also connected to supply electrodes through a resistive layer.
According to one embodiment two or more clusters are connected to supply electrodes and are arranged in a group in the form of a single layer of clusters directly contacting one another, and one or more clusters are connected to control input
electrodes, other cluster or clusters are connected to output electrodes, forming thereby the output of the logical element « OR ».
In another embodiment two or more clusters are arranged into a group in the form of a serial one-dimensional chain, the even elements of which are connected through resistive layers to the first supply electrode, and the odd elements are connected through resistive layers to the second supply electrode, forming thereby a logical shift register.
Clusters and groups of clusters may be arranged by way of direct contacting and joined together by means of electrodes as well.
In one embodiment of the invention two or more clusters are connected to supply electrodes and joined together in a group in the form of a single layer, and besides one or more clusters are connected to control input electrodes, the other cluster or clusters are connected to output electrodes, the input and output clusters are connected to one another through additional electrodes of the similar thickness and width, and besides the said electrodes can be connected to one or more clusters of the next group.
In the other variant of ensembling two or more clusters connected to supply electrodes and joined together in a group in the form of one layer, one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to output electrodes, and besides the input and output clusters are connected to one another through additional electrodes, tapered at one side in the signal direction, the said electrodes can be connected to one or more clusters of the next group.
The logical inversion operation may be carried out if the cluster is connected to supply voltage through resistive layer, and the supply point is connected to the output electrode, and besides the input voltage is supplied directly through one or more control electrodes, connected to clusters through the tunnelly transparent gap.
Analogue comparing of two signals may be performed when two clusters are connected to the supply voltage through resistive elements and the first input voltage is supplied directly through the first control electrode, connected to one cluster through the tunnelly transparent gap, and the second input voltage is supplied to the second electrode connected to the other cluster through tunnelly transparent gap, while some of the junction points to resistive elements of each cluster are joined
together and the other junction points of resistive elements are connected to output electrodes, forming thereby the outputs of a two signal analogue comparator.
Two clusters are connected to the supply voltage through resistive elements, and the junction points thereof are connected to output electrodes, and besides the first input voltage is supplied directly through the first control electrode connected to the first cluster through the tunnelly transparent gap, and the second input voltage is supplied directly through the second control electrode connected to the second cluster through the tunnelly transparent gap, the first output electrode is connected to the second cluster through the tunnelly transparent gap, the second output electrode is connected to the first cluster through the tunnelly transparent gap forming thereby a bistable trigger circuit.
Two or more clusters may be connected through the resistive layer to the supply voltage and form isolated groups, united by one common output electrode, and each isolated group of clusters is connected to one or more control input electrodes, and besides the number of clusters in each group determines the weight function according to the input signal and forming thereby the neurone-type logical component -a weight summator.
Still one more improvement is in one or more clusters being connected to the supply electrodes at least through one additional cluster layer.
Two or more clusters may be connected to the supply electrodes and joined together in a group in the form of a single layer of directly together contacting clusters, one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to the output electrodes, forming thereby the output of a logical memory component"OR"with memory.
Two or more clusters may be connected to the supply electrodes and joined together in a group in the form of a single layer, one or more clusters are connected to the control input electrodes, and the other cluster or clusters are connected to output electrodes, and besides the input and output clusters are joined together through additional electrodes of the same thickness and width, and these electrodes can be connected to one or more clusters of the next group.
Two or more clusters may be connected to supply electrodes and joined together in a group in the form of a single layer, one or more clusters are connected to control input electrodes, and the other cluster or other clusters are connected to output electrodes, and besides input and output clusters are joined together through
additional electrodes tapered at one side in the in the signal direction, and the said electrodes can be connected to one or more clusters of the next group.
One and more clusters may be connected through additional clusters to supply voltage and form isolated groups, which are combined by a single common output electrode, and each isolated group of clusters is connected to one or more control input electrodes, the number of clusters in the each group determines the weight function according to the input signal, forming thereby a neurone-type logical component-a weight summator with memory.
An important logic operation may be performed if a cluster is connected through additional cluster to the supply voltage, a junction point is connected to the output electrode, the input voltage is supplied directly through one or more control electrodes, which is connected to the clusters through the tunnelly transparent gap, forming thereby an inversion logical component with memory.
In the further embodiment two clusters are connected to the supply voltage through additional clusters, the first input voltage is supplied directly through the first control electrode connected to one cluster through the tunnelly transparent gap, and the second input voltage is supplied to the second electrode connected to the other cluster through the tunnelly transparent gap, some of junction points of the resistive elements of each cluster are joined together and connected to the supply electrode through the resistive element, and the other junction points of additional clusters are connected to the output electrodes, forming thereby the outputs of the two signal analogue comparator with memory.
Still further improvement is in two clusters being connected through additional clusters to the supply voltage, the junction points thereof are connected to the output electrodes, the first input voltage is supplied directly through the first control electrode connected to the first cluster through the tunnelly transparent gap, and the second input voltage is supplied directly through the second control electrode connected to the second cluster through the tunnelly transparent gap, the first output electrode is connected to the second cluster through the tunnelly transparent gap, and the second output electrode is connected to the first cluster through tunnelly transparent gap forming thereby a bistable trigger with memory.
In the further improvement used for image conversion into digital form two or more clusters are connected to two control electrodes at least one of which is light
transparent, the gaps between clusters are filled with the photosensitive semiconductor, a set of such clusters forms a photosensitive matrix.
If one or more cluster layers are connected to at least two spaced electrodes, at least one of which is light transparent, a gaps between clusters are filled with photosensitive semiconductor, forming thereby a light-control memory medium, which may be used, for example in laser discs.
According to one more improvement one or more cluster layers are connected to at least two spaced electrodes made in the form of resonator, forming thereby a high- frequency generator with a maximal boundary frequency determined by the formula f < mea4c2/2z.
In a still further improvement one or more clusters are combined by direct contacting or joined together through electrodes and are connected to the voltage supply, at least one of the contacts is connected to the output electrode allowing thereby to form a standard voltage supply with levels U= na3c2me/2e, wherein n-a number of serially connected clusters.
The operation process of devices is also characterised in that the field control strength per one cluster is determined in the range, Emin: E < EmaX X wherein E-maV/2eh, E-Ea.
The process of operation of the above-described devices is characterised in using a continuos and/or pulse supply.
Other class of electronic devices in accordance with the further invention comprises the following improvements.
A quantum-size electronic device comprising electrodes and located between them a layer of material having MSPT in which the layer of material having MSPT is made in the form of clusters, which have cross section size determined from the formula: r = a rO, wherein a is a coefficient determined in the range 2 < a < 4, the distance between the electrodes being more than ro.
The further improvement consists in connecting the cluster to the supply electrodes and at least one load and through tunnelly transparent gaps-to one or more control electrodes, the thickness of tunnelly transparent gaps does not exceed ro and the distance between electrodes is not less than ro.
Electrodes made of superconductor or material having MSPT may have a cross section size d 2 2 ru Such clusters may be connected to supply electrodes at least through one resistive layer.
Such a layer may be connected to two or more clusters that are connected to supply electrodes and combined into a group in the form of a single layer of directly contacting clusters, and besides one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to output electrodes, forming thereby the output of the logical element"OR".
Further improvement consists in that two or more clusters are connected to supply electrodes and joined together in a group in the form of a single layer, and besides one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to output electrodes, input and output clusters are connected in-between through additional electrodes of the same thickness and width, and besides the said electrodes can be connected to one or more clusters of the next group.
In case of one-directional signal passing two or more clusters are connected to supply electrodes and joined together in a group in the form of a single layer, an besides one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to output electrodes, input and output clusters are connected in-between through additional electrodes, tapered at one side in the signal direction, and besides the said electrodes can be connected to one or more clusters of the next group.
If the input voltage is supplied directly through one or more control electrodes connected to the cluster through the tunnelly transparent gap, and besides the cluster is connected through the resistive element to the supply voltage, and the connection point is connected to the output electrode, then it is the output of the logical element "NOT".
For carrying out the operation of analog comparison of two signals, two clusters are connected to the supply voltage through resistive elements, the first input voltage is supplied directly through the first control electrode connected to one cluster through the tunnelly transparent layer, and second input voltage is supplied to the second electrode connected to the other cluster through the tunnelly transparent gap, and besides some of the connection points to the resistive elements of each clusters are joined together, the other connection points of resistive elements are connected to the output electrodes, forming thereby the outputs of the analog comparator of two signals.
One more device according to the invention is made so that two clusters are t. connected to the supply voltage through resistive elements, and the points of connection thereof are connected to the output electrodes, the first input voltage is supplied directly through the first control electrode connected to the first cluster through the tunnelly transparent gap, and the second input voltage is supplied directly through the second control electrode connected with the second cluster through the tunnelly transparent gap, the first output electrode is connected to the second cluster through the tunnelly transparent layer, and the second output electrode is connected to the first cluster through the tunnelly transparent layer, forming thereby a bistable trigger.
One more variant of development of the invention consists in that one and more clusters are connected to the supply voltage through a resistive layer and form isolated groups combined by one common output electrode, and each isolated group of clusters is connected to one or more control input electrodes, and the number of clusters in each group determines the weight function according to the input signal, forming thereby a neurone type logic element-a weight combiner.
The further improvement consists in that two or more clusters are connected to at least two control electrodes, the gaps between clusters are filled with photosensitive semiconductor, a set of such clusters forming a photosensitive matrix.
If one or more layers of clusters are connected to at least two spaced electrodes made in the form of resonator, they form a high frequency resonator with maximum cut-off frequency determined from the formula f < mea4c2/h
The operating procedure of devices with clusters made of material having MSPT consists in transmitting electric current through at least one cluster, and is characterised in that the current density through cluster is limited by the value j < 4weme a c/h If in the cluster are used materials having the temperature of the metal- semiconductor phase transition higher than the operation temperature of devices, then under the invention it is necessary to provide the electric field density E > me'aSc3/2eh, This condition is optional in case of use in the cluster of materials having the temperature of the metal-semiconductor phase transition lower than the operating temperature of devices.
For a photosensitive matrix the process of operation that comprises transmitting of the electric current at least through one cluster, is characterised in using in the cluster of materials having the temperature of the metal-semiconductor phase transition higher than the operating temperature of devices.
One more variant of the invention consists in that the quantum size electronic device, comprising electrodes and at least one cluster located in-between, is characterised in that the cluster is made of the material of the superconductor and has the cross section size determined by the formula: r = a. rO, wherein a-is a coefficient determined within the range 2 < a < 4, and besides the distance between the electrode exceeds ro.
The further improvement consists in that the cluster is connected to the supply electrodes and at least one load, and through tunnelly transparent gaps they are connected to one or more control electrodes, and besides the thickness of the tunnelly transparent gaps does not exceed ro.
The device may be additionally characterised in that the electrodes are made of superconductor or of material having MSPT and having the cross section size d > 2 ro.
One or more clusters according to the improvement may be connected to the supply electrodes at least through one resistive layer.
Through such a resistive layer two or more clusters are connected to the supply electrodes and joined together in a group in the form of a single layer of directly contacting in-between clusters, and besides one or more clusters are connected to
the control input electrodes, and the other cluster or clusters are connected to the output electrodes, forming thereby the output of the logical element"OR".
The further improvement consists in that two or more clusters are connected to the supply electrodes and joined together in a group in the form of one layer, and besides one or more clusters are connected to control input electrodes, and the other cluster or clusters are connected to the output electrodes, the input and output clusters are connected in-between through additional electrodes of the same thickness and widths, and besides electrodes can be connected to one or more clusters of the next group.
To ensure a directed signal passing two or more clusters are connected to the supply electrodes and joined together in a group in the form of a single layer, one or more clusters are connected to the control input electrodes, and the other cluster or clusters are connected to the output electrodes, the input and output clusters are connected in-between through additional electrodes tapered at one side in the signal direction, and besides these electrodes can be connected to one or more clusters of the next group.
Such an important logic element as an input signal inversion may be created if the input voltage is supplied directly through one or more control electrodes connected to the cluster through the tunnelly transparent layer, the cluster is connected to the supply voltage through the resistive element, and besides the connection point is connected a to the output electrodes, forming thereby the output of the logic element"NOT".
According to the next improvement two clusters are connected to the supply voltage through resistive elements, and with this the first input voltage is supplied directly through the first control electrode that is connected to one cluster through the tunnelly transparent gap, and the second input voltage is supplied to the second electrode that is connected to the other cluster through the tunnelly transparent gap, and besides some of the connection points that are connected to the resistive elements of each cluster are joined together, and the other connection points of resistive elements are connected to the output electrodes that are outputs of the comparator circuit of two signals.
If two clusters are connected to the supply voltage through resistive elements, the connection points thereof are connected to output electrodes, and besides the first input voltage is supplied directly through the first control electrode that is connected to
the first cluster through the tunnelly transparent gap, and the second input voltage is supplied directly through the second control electrode that is connected to the second cluster through the tunnelly transparent gap, the first output electrode is connected to the second cluster through the tunnelly transparent gap, the second output electrode is connected to the first cluster through the tunnelly transparent gap, forming thereby a bistable trigger.
Further one and more clusters are connected the supply voltage through the resistive layer and form isolated groups that are joined together by one common output electrode, and one or more control input electrodes are connected to each isolated group of clusters, with this the number of clusters in each group determines the weight function according to the input signal, forming thereby a neurone-type logic element-a weight comparator.
The operating procedure of the device is determined by that the operating range of devices is limited by critical temperature of the junction in the superconducting state of the used materials, which is determined from the formula Tc < mea3c2/2kn, wherein k-is the Boltzmann's constant, The operating procedure of the devices listed in this variant is characterised in that transition from the superconducting state to the normal sate under the action of the control voltages takes place with the following strength at the cluster E > me-a5c3/2he where h = 2jrh-Ptanck's constant.
In one more variant the quantum size electronic device comprising electrodes, at least one of which is made of superconductor or material with MSPT is characterised by that the electrodes have a cross section size determined from the formula: d = nd r09 wherein nd-a coefficient determined in the range 1 < nd< 2, The device is further characterised by that a group of electrodes have at least one region of approach up to the distance not exceeding ro, forming thereby a transformer a direct or altemate current.
According to the next improvement two electrodes have at least two regions of approach up to the distance not exceeding ro, and besides a direct current flows through one of the electrodes, the second electrode is connected to the load, with this
an alternate current flows through the load, and forming thereby a dc-to-ac transducer.
Moreover the two electrodes have at least two regions of approach to the distance not exceeding ro, and besides alternate current flows through one of the electrodes, and the second electrode is connected to the load through which direct current flows, forming an ac-to-dc transducer.
All the itemised devices are illustrated by the following below examples that are depicted on the drawings.
Brief description of the drawings Fig. 1. Spherical logic cell.
Fig. 2. Spin configurations of paired ring electrons in the spherical logic cell.
(The arrow indicates the direction of the ring current).
Fig. 3. Cylinder logic cell.
Fig. 4. Spin configurations of the paired ring electrons in the cylinder logic cell.
Fig. 5. Cylinder logic cell basing on the material with MSPT.
Fig. 6. Spin configurations of paired ring electrons in a cylinder logic cell on basis of material with MSPT.
Fig. 7. Graph of current against strength of the electric field for a cylinder logic cell on basis of material with MSPT.
Fig. 8. Cylinder logic cell on basis of superconductor.
Fig. 9. Spin configurations of paired ring electrons in the cylinder logic cell on basis of superconductor.
Fig. 10. Current-voltage diagram of cylinder logic cell on basis of superconductor.
Fig. 11. Experimental current-voltage diagram of the spherical logic cell.
Fig. 12. Experimental current-voltage diagram of a group of logical cells formed in semiconductor.
Fig. 13. Current-voltage diagram of a thin-film logic cell on basis of TiOX.
Fig. 14. Design of a multilayer logic cell on basis of spherical clusters.
Fig. 15. Design of a logic cell on basis of a group of cylinder clusters.
Fig. 16. Volatile memory matrix..
Fig. 17. Non-volatile memory matrix.
Fig. 18. A variant of connecting control resistive electrodes to a group of contacting clusters.
Fig. 19. A variant of connecting control electrodes to a group of contacting clusters through the additional cluster layer.
Fig. 20. Group logic cell"OR".
Fig. 21. Group logic cell"OR"with branching at the output.
Fig. 22. Group logic cell"OR"with a high-power output.
Fig. 23. A variant of integrating logic cells with a predetermined signal direction.
Fig. 24. A variant of integrating logical cells with a predetermined direction of signal passing and branching.
Fig. 25. Shift register.
Fig. 26. Volatile logic element"NOT".
Fig. 27. Non-volatile logic element"NOT".
Fig. 28. Non-volatile trigger.
Fig. 29. Analog signal comparator.
Fig. 30. Standard voltage source.
Fig. 31. Neurone logic element.
Fig. 32. Non-volatile weight summator of a neurone logic element.
Fig. 33. Volatile weight summator of a neuron logic element on basis of materials with metal-semiconductor phase transition.
Fig. 34. Current transformer.
Fig. 35. Dc-to-ac transformer The aforesaid devices may be classified and described as follows: Resonant electronic devices.
As a matter of principle it is possible to design any artificial flaw that may be a certain resonator for a ring wave with a radius ro and an effective quality factor 1/a.
Fig. 1 presents a spherical logic cell. 1 is the cluster-nucleus; 2 is the tunnelly transparent enclosure; 3,4 are supply electrodes; 5,6 are control electrodes. Actually, a nucleus may be made of any material-metal, semiconductor, superconductor, material with MSPT, high-molecular organic compound or be a mere empty cave-a bubble. The main thing is that the nucleus radius had dimensions aliquot ro. The cluster enclosure may be made of any dielectric or semiconductor and have a single- or multilayer structure. The thickness of the envelope generally should not exceed ro.
The said cluster can be connected to the supply electrodes 3,4 and control
electrodes 5,6. With this one or two electrons in various spin configurations may be present at the nucleus. Fig. 2 shows spin configurations of paired ring electrons in a spherical logic cell. The arrow direction shows the direction of the ring current.
Taking as an example the cell of Fig. 1 it is possible to determine optimal design characteristics, which may be applied also to the other electronic quantum size devices considered below.
Thermodynamic stability of a ring electron is found from the kinetic equation of movement of a spaced ring mass speeding at oc2c. In this case we take the effective electron mass for m* = me/a, then the critical temperature up to which the existence of a ring electron is possible may be expressed as follows: Te = me (a2c) 2/ (2ka) = 1 151.86K (878.71 °C) (7) It is characteristic that temperatures of phase transitions of the first and the second order in some materials correspond to the critical temperature, e. g. the upper limit of semiconductor-metal phase transitions [16, p. 4]: MoTe2, T _ 1053K (780°C); NbO2, T _ 1070K (797°C); FeSi2, T _ 1123K (850°C) etc.. An exception is rather only ZrO2, T = 1273. 1443K, (1000. 1170°C), probably due to the presence of two modifications of the crystal latitude. The upper limits of phase transitions of the second kind lie close to this temperature, e. g. ferromagnetics in the paramagnetic state and anti-ferromagnetics in the state of magnetic disordering.
Let us determine energetic characteristics of a cell. The equations (5) and (6) show that when two rings approach at the distance < rO/there may be created a pair with a maximum energy: W, = (8/27) e'/ (4r,) (8) At a transverse orthogonal overlap of two rings having a common centre the energy maximum is: W2 = (l/4) e2/ (4Z£0r0) (9) The bounding energy W1 corresponds to the phase transition temperature T, = W,/k = 8/27Te = 341.3K (68.3°C). This magnitude perfectly corresponds to the experimental temperature of a lightly smeared centre of the phase transition in VO2, Tn =340K (67°C). Bounding energy gives a good characteristic to the V203 system. The temperature of the beginning of the phase transition will be T2 = 1/8 Te =144K (-129°C) (during the experiment T=145K (-128°C) [16]). Energy state of the bound electron
pairs may undergo changes only with multiplicity of a rational number- <BR> <BR> <BR> <BR> n= n1/n2, where n,, n2-are natural integer numbers and n2 W 0. By changing multiplicity of the distance between electron rings with provision for vanadium oxides forming Magneli series VnO2n 1 = V203 + (n-2) VO2, it is possible to determine to high precision all temperatures of metal-semiconductor phase transitions of the series.
Similariy the temperature of metal-semiconductor phase junctions for any material may be theoretically determined.
As it follows from theoretical and experimental data for metal-semiconductor phase transitions, ring electrons may be condensed into drops at various space configurations of their magnetic moments - spin. Unlike bound electron pairs in atoms, <BR> <BR> having only 4 spin states: ring electrons may have additional states: 1"- and etc. Beside these states a ring allows discrete compression into a ring with a smaller radius (Fig. 2a) or directly into an ellipse. (Fig. 2b). It is important that at maximum compression of two ring-like electrons, their spins will be directed*, +, ¢, A, and rings must transform into ellipses with a maximum size of semiaxis Hence, the maximum size of a cluster nucleus will be as follows: dmaX = rO (4-X) = 18.75 nm. With this the configuration of electrons will have a form of orthogonal inter-crossing ellipses, and besides one of the ellipses passes through focuses of the second ellipse (Fig. 2b). There is a probabiiity of joining together electrons without compressing them into an ellipse, obtaining thereby a nucleus d = 2rO (Fig. 2c).
Naturally, the maximum size of a pair of rings at To will be at their coaxial inter-crossing d=3ro (Fig. 2d), and at a bounding through the surface of rings- d = 4ro (Fig. 2. By adding an enclosure with a thickness ro we shall get the maximum cell size Sro = 36.26 nm.
Now let us determine the conditions to the minimum cell size. We allow the condition of the cell located on a substrate being heated to at least a normal temperature.
It is shown in [1, p. 20], that even at sizes of an order 2-10 nm, phase and structural transformations of clusters start to differ from the bulk material. At smaller sizes the melting temperature of particles reduces and the crystal structure thereof changes as well. At the same time may increase the mobility of cluster on the
substrate surface and a coefficient of diffusion of the cluster material into the bulk of the substrate material. In order to reduce the thermodynamic unsuitability, i. e. cluster fusing at the elevated operating temperatures, the atom binding energy in the cluster or the energy of the phase transition of the cluster material should be not less than 5 kJ/mol.
The other criterion limiting the minimum size of cluster may be found from the condition of compression of the electronic ring by factor of 2,3,4 and more. Namely, de Broglie wave function will have the second and higher modes, i. e. harmonics. Then ro* = ro/n, and at n=2 the ring diameter will decrease to 7.25 nm. But as it was shown above upon decreasing the distance between electrodes < 8 nm there appears the short-circuit « tunnel » current. Therefore there is no physical sense in reducing the cluster size to less than 7.25 nm, and to diminishing the distance between control and supply electrodes to less than 7.25 nm.
Hence the size of 7.25 nm is the fundamental limit of nanoelectronics.
Taking into consideration the aforesaid let us select the operation dimensions for an elementary logic cell in the range: 7.25 nm <_ d <_ 36.26 nm Beside the centrally symmetric configuration there may be designed axisymmetric configurations of clusters that will also have resonance features and form ring electrons.
Fig. 3 shows a planar modification of a logic cell-a cylinder logic cell, wherein 7 is cluster-nucleus; 8 is a tunnelly transparent enclosure; 9,10 are supply electrodes; 11 is a control electrode.
Here the function of the nucleus-resonator plays a cylinder-like channel 7 directly built in a dielectric or a large-gap semiconductor, and functioning as an enclosure 8. Moreover the channel may be filled with any metal, semiconductor, superconductor, high molecular compound, material with MSPT or partially filled with residual gases. Supply electrodes 9,10 and control electrodes 11 are connected to the channel. It is important that the materials were thermally and chemically resistant, i. e., processable. Besides, they should not form additional defects scattering one or two ring electrons during their movement around the nucleus. Dimensional requirement towards axisymmetric cells somewhat differ from that to centrally symmetric cells. The minimum diameter of the cell nucleus is 2ro, and the maximum should not exceed 4ro. Such cells may contact one another through tunnelly
transparent gaps having a maximum size not more than ro and form periodic structures extending both parallel to the axis and along the axis.
Axisymmetric configurations may have both orthogonal-spin (OS) pairing of electrons of the types as shown at Fig. 2 and parallel pairing and forming chains. Fig.
4 shows spin configurations of paired ring electrons in a cylinder logic cell. The arrow indicates the direction of the ring current.
General requirements to constructive materials are defined below.
It is a common knowledge that compounds of transient metals have a variable valence. Higher oxides of such metals as Ti, V, Nb, Mo, W at deviation from stechiometry, for example in oxygen deficiency, rather then point defects form new homological Magneli series: MnO2n"MnO3n1, MnO3n2, where n = 1,2,3... etc. [16, p.
48]. Besides, the said metals form compounds, having MSPTs at temperatures higher <BR> <BR> <BR> <BR> than normal. For example, NbO2, Tk =1070K (797°C); V305, Tk = 450K (177°C); Ti203, Tk=600K (327°C); Ti305, Tk = 448K (175°C); ZrO2, Tk = 1273K (1000°C). Hence, such materials are extremely advantageous in designing enclosures. Moreover for designing a nucleus it is possible to use other materials, having a high temperature of metal-semiconductor junction: MoTe2, Tk = 1053K (780°C); FeSi2, Tk = 1123K (850°C).
Parallel-spin electronic devices.
In a number of embodiments it is possible to make a direct use of the characteristics of the proper material avoiding creation of an artificial cluster formation that initiates formation of a ring electron. For example, a ring electron automatically appears in materials with MSPT. An elementary logic cell on basis of such material is shown at Fig. 5. In this cell a channel of material 12 with MSPT functions as a nucleus, 13 is a tunnelly transparent gap; 14,15 are supply electrodes; 16 is a control electrode. The radius of such a nucleus-channel is within a range ro to 4ro. Similar to the previous cases the distance between the electrodes should not be less than ro.
The distinctive feature of this kind of the cell is lack of tunnelly transparent gaps between supply electrodes. Such a medium is not able to store information. But the size of the channel along the axis is not limited what is useful in many applications, for example in integrated circuits with a complex topology.
In case of an extended axisymmetric cell, ring electrons may both form pairs and chains of the type TTTTTT... Fig. 6 presents spin configurations of paired ring
electrons in a cylinder logic cell on basis of material with MSPT. The arrow indicates the direction of the ring current.
This type of condensation of ring electrons will be named'parallel-spin condensation' (PS). In this chain distances between separate ring electrons will not exceed rO/The chain performs charge transport between supply electrodes 14,15 (Fig. 5). Current is controlled by a control electrode 16, which is connected to the nucleus-channel 12 through the tunnelly transparent dielectric enclosure 13. Actually, such a structure is a quantum-size analogue of a field transistor with coherent electrons.
During experiments such chains of ring electrons are observed in the form of current filamentary nanopaths at field breakthrough phenomenon in thin films. The breakthrough leads to formation of S-shaped volt-amps diagrams and has a number of technical applications [16,15]. However the prior art described hitherto the formation mechanism of S-shaped characteristics basing on thermodynamic instability appearing in the hysteresis loop due to lattice strain at the temperature metal- semiconductor and semiconductor-metal transition. But this explanation does not allow describing certain experimentally low temperature of the proper filament path, which is lower than the temperature of phase transition material. In case of our model the temperature of the filament path is determined from the condition Te/n, where n- is a number of ring electrons forming the said path. Thus, the more ring electrons participate in forming the filament path, the lower is the effective temperature thereof.
Generally in technical application high overworks are created on films with MSPT, thus naturally raising thereby the temperature of filament path above the temperature of phase transition material. This in turn leads to overheating of the junction and increasing the time of switching the device due to temperature relaxations.
Thus, taking into consideration specialities of forming the current nanopath it is possible to design a speed device without any temperature relaxation, having a high reliability due to absence of re-crystallisation and temperature breakdown of material.
Fig. 7 shows a graph of standard current/electric field strength ratio for a cylinder logic cell on basis of material with MSPT. The scale of current values and field strength is shown in conditional units. El is a threshold strength, U1<U2<U3 is the voltage at the control electrode 16 (Fig. 5). The passage of current through the
channel is possible only after crossing the field threshold E,. After the channel « breakdown » the channel current is controlled by the electrode 16, what is reflected at the diagram as a family of characteristics at different values of U. If a device operates at temperatures higher than MSPT, than the channel material is in the metal phase and El is near null. But with this metal phase is not a classical metal and it does not screen the electric field of the control electrode 16 due to the specific configuration of ring electrons. Hence, this case makes possible gaining an effective control of the current flowing through the channel.
Antiparallel-spin electronic devices.
Under determinate conditions and in certain materials ring electrons may pair <BR> <BR> <BR> <BR> into chains with opposite spins of the type 1\t/N/N/L. This state will be further refereed to as antiparallel-spin state (APS). In this phase are met all main and sufficient conditions for formation of superfluid quantum liquid: pulses of electrons (of current and field) are directed oppositely and the whole chain forms a space-coherent periodic function. It should be specially noted that in this case paired electrons interact with one another due to electromagnetic fields at the near-light speed. Logic cells with hysteresis characteristic may be formed on such materials. Fig. 8 shows a cylinder logic cell on basis of superconductor, where 17 is a cluster on basis of superconductor; 18 is a tunnelly transparent layer; 19,20 are supply electrodes; 21 is the control electrode.
Fig. 9 shows spin configurations of paired ring electrons in a cylinder logic cell on basis of superconductor. The arrow indicates the direction of the ring current.
Fig. 10 shows the voltage-current characteristic of the cylinder logic cell on basis of superconductor. The current scale is denoted in conventional units. I,-is a critical current of superconductor; U,-is a threshold voltage.
In essence this cell repeats the structure of the cell made of material with MSPT. Here there are no tunnelly transparent layers between supply electrodes, and the cell is controlled through the tunnelly transparent layer 18 (Fig. 8) of thickness not more than ro. Though the material of the channel 17 is in the superconducting phase, the electric field with control electrodes 21 penetrates therethrough due to the specific configuration of ring paired electrons. In this case using the field it is possible to control the critical current of superconductor, i. e. to get a certain quantum-size analogue of a superconducting field transistor. The cell should be supplied by a current source. The voltage drop at the cell will be equal to 0. At reaching the critical
current 11 (point A) the cell state changes from superconducting state to the resistive state (point B). The characteristic remains resistive at further voltage gain. At further loss of voltage at the hysteresis loop it returns to the original location. As shows the provided hysteresis loop it fully corresponds to volt-amps diagram of superconducting devices widely used in cryotechnics, for example on basis of thin superconductor filaments [14], though having the additional function of the critical current control by means of the control electrode.
A great number of logical elements on basis of superconductor devices are investigated in [14]. They are non-topical now because of their low-temperature operating regime. Under our assumptions there is a possibility to give the second birth to this class of devices by bringing them to normal temperatures.
The modern superconductivity theories, e. g. the Bardeen-Cooper-Schrieffer theory, are based on the phonon action of at large distance paring. However these theories assume that interaction of electrons by means of phonons goes with a perpetual speed. [13, p. 287.]., i. e. there is no retarded multiplier in the Hamiltonian function. Hence it is impossible to determine the electron speed in the superconducting phase. An unlimited speed of interaction results in violation of energy conservation law.
In our case the speed of electron movement is limited by the value of a2c, corresponding thereby to the energy conservation law.
A turn of electron spin will change the energy of the system. The bonding energy in APS-phase will be lower than the binding energy of ring electrons in PS- <BR> <BR> <BR> phase by the magnitude l/27rn. Taking into consideration that the charge of the pair is 2e, the critical temperature is found from (7) according to the formula: T, =2T27rn.
Eventually we get the final formula for finding the critical temperature high- temperature superconductors: Tc= me (a'c) 2/ (2kan) (10) The maximum critical temperature accessible in high-temperature superconductors will be at n=1 and will form Tcl= 366.65K (93.5°C). This temperature conform, e. g. powdered superconductors on basis of xCuBr. CuBr2-the experiment gives a temperature of the order 365K (92°C) [31]. Micron powders on basis of Y-Ba- Cu-O composition have a critical temperature of the order 300 K (27°C) [32]. Similar effects were found in many powder materials of the said class. [33]. Micron regions of
the surface of oxide polypropylene films have the critical temperature of the order 300K (27°C) [34].
Temperature n=2, Toc2=183.2K (-89.95°C) corresponds to powders on basis of system C60/Cu in respect of 7/1 (the experiment gives the temperature of the centre of the smearing transition of the order 185K (-88°C) [35].
Temperature n=4, Te4=91.66K (-181.49°C) corresponds to the majority of discovered at present high-temperature superconductors of the YBa2Cu307 type. For monocrystalline samples of such ceramics many authors report the experimental critical temperature equal to 91.6K (-181.55°C) [36]).
Temperature n=16, Toc16=22 92K (-250.23°C) conforms the critical temperature 200 nm of Nb2Ge films T,= 23.2 0.2K (-250.15°C) [37, p. 267] Temperature n=32, Tu32=1 1.46K (-261.69°C) conforms a large number of low- temperature superconductors of the second order of the Y2C3, NbC, Nb3Au type, etc.
[37]. In essence all the aforementioned superconductors have one common feature, and namely, they are superconductors of the second order on ring electrons.
Accordingly, the cited data are the best proof for existence of ring electrons.
In should be noted that in low-temperature superconductors of the first order on basis of monomaterials electron bounding with the lattice differs from that of the aforementioned. However they are of no practical importance for quantum size devices. The distinctive features of such interaction will be described in further publications.
Only materials that are superconductive at the temperature Tc, have practical importance for designing nanoelectronic devices. Creating such materials is of especial importance also for use in power engineering. Therefore the present specification comprises no know-how disclosure of art of manufacture of such materials. However it is really possible to create nanosized devices as well as supply and control electrodes therefore on basis of the aforementioned powder materials with high-temperature superconductivity.
Selection of a logic cell performance A model of ring electron describes OS-, PS-and APS-electronic devices.
Accordingly performance analysis is similar for all of such cells. The energy of forming a ring electron from bound states, i. e. the energy that is necessary for an electron transition from the electrode and current into the nucleus region will be determined by the expression:
αeUe = me(α²c)²/2 = 1,16 *10-22 J, (11) what corresponds to Ue= 0.09928 V. At pairing of electrons in OS-phase the energy will be identical to 2Ue= 0.19856 V. As electrons in the APS-phase are turned versus one another by n, the threshold voltage at the superconducting cell is even to U1= Ue/w = 0.0316V. This voltage complies with the threshold energy of electron bounding in any superconductors < 31. meV. This This supported supported experimental data, obtained, e. g. regarding the energy of electron bounding in YBa2Cu307 superconductors about 30 meV [38]. As the space separates, the bounding energy of a pair of electrons in APS-phase will correspond to the definite discrete energy levels n, i.e. U1/n. For example the best NbN/MgO/NbN supercondiucting device gives a threshold voltage (value of the energy gap band) 5.3 mV [39]. In our case this corresponds to n=6, i. e., 31.6 mV/6 = 5.266 mV.
The rotation frequency of an electronic ring: fe= a2c/2ZrO= me (a2c) 2/h= 3.5037 10"Hz. (12) To this frequency will correspond the maximum possible frequency of OS-and PS-phase generators The time for departure by an electron of the nucleus cell at supply of the external field will be equal to one turn thereof, i. e. Te =l/fe, and, hence, the maximum current through one element: Ie=efe= eme (a'c) 2/h = 5.6 10-8 A. (13) Considering that an average radius of a cluster is z 2ro, we find the current density at 1cm2 of a surface of densely packed cells: j = I/ (W r 2) = 47rem 3a8C4lh3 = 3.4 104 A/cm'-. (14) The maximum required at the cell field strength, capable of switching it into another state is: Ee = Ue/ro = me'a5c3/2eh = 1.37 105 V/cm. (15) The resistivity of a cell is: Re = Ue/Ie = h/2e2a = 1. 768 l06 Q (16) that is of 137.036 higher than the Josephson resistance. The magnetic field of such ring current is: Be = me#e/e = (me/e)[me(α2c)2/h] = 12.5 T. (17) The magnetic flow is equal to:
(De = TrrB, = h/2e = (Do 067810'"Wb,(18) that is accurate to the quantum of the magnetic flow.
The cited expressions describe phase changes of the first and second order in critical points (singular behaviour) in condensed substances. The aforecited expressions show that they are comprised only of common constants and determine the maximum permissible characteristics of any high-temperature electronic quantum size devices produced of any solid-state materials.
As it is shown above the maximum speed of a ring electron in a condensed medium does not exceed the speed a2c. Right this speed is the limit of existence of a ring electron of the maximum size. Absolutely, it is possible to design electronic devices having speed of carriers higher than oc or higher than the maximum sound <BR> <BR> <BR> <BR> speed in condensed materials umaX=3a2c/ (X). However in this case ring electrons wrap up to atomic sizes. Their energy increases up to the energy of electrons at the Fermi level what leads to the increase of the cross section of <BR> <BR> <BR> <BR> interaction of electrons with the lattice by factor of 1/a. This naturally leads to the excess heating of devices thus limiting the use thereof in superlarge integrated circuits.
In order to support the theoretical data there were designed two kinds of OS- electronic cells.
First variant ZrO2. Hollow spheres of an average hollow radius of the order ro and the thickness of the enclosure ro/2 were plasma deposited on a metal substrate as a single layer with minor addition elements. As a result we get an analogue of the cell on Fig. 1. Spheres having diameters being maximal close to 3ro were selected by a scanning tunnel microscope with the needlepoint of a radius of the order ro, The <BR> <BR> <BR> <BR> maximum Q-factor of the hollow-resonator equal to 1/a was determined according to the maximum change of the tunnel current. This Q-factor complied spheres having the nucleus of the diameter 2ro. The volt-amps diagram of such a cell was taken by changing the needle voltage. Fig. 11 shows an experimental volt-amps characteristic of a spherical logic cell, where 22-is a resistive load. The starting segment of volt- amps diagram from 0.0V to 0.1V (point A) corresponds to parasitic loss currents of the cell enclosure dielectric. This is due to the fact that the microscope was operated at the room atmosphere with the end humidity and at normal temperature. At gaining the voltage value of 0.1 V the current increased abruptly (up to point B) what
corresponded to formation a ring electron around the nucleus. At further voltage increase the current was growing due to the passage of ring electron through the cell (line BC). Thus, electrons unwrapping to their maximum sizes are constantly coming from the metal of the substrate through the tunnelly transparent enclosure into the region of hollow-nucleus. Further ring electrons flow through the tunnelly transparent enclosure into the microscope needle, where they unwrap anew to the sizes- determined by the characteristics of needle material. The current increase is in direct ratio to the voltage increase at the needle. In this case the control electrode is absent.
At the gaining at the needle of the voltage o. 2V (point C) current falls (point D) up to the value of the loss current, determined by the dielectric, forming the cell enclosure. Actually orthogonal pairing of two ring electrons takes place at the voltage 0.2 V as the energy 0.2 eV is the threshold energy of forming the shared potential hole thereof. At this parameter the cell may operate as a logic memory at the external maintaining voltage 0.1 V< U < 0.2V (segment AD). At these parameters the cell practically drains no energy except the energy determined by the parasite leakage current.
At the voltage increase of more than 0.2 V, the current passing through the cell will be determined yet by paired electrons-electronic drops and will continue increasing at a lower slope in a direct proportion to the voltage.
At decrease of the voltage to null the cell discharges and is ready to work again.
Second variant In order to test possibilities of logic cell operation in photosensitive matrixes there was designed a cell in a photosensitive semiconductor.
Atoms of transition metals were particularly injected into an n-Si substrate. At concentrations of 1011_1020 cm~3 and special anneal parameters clusters with nuclear size close to the radius ro were formed in the said supersaturated solution. Then the substrate surface was coated by semitransparent electrodes and thereafter volt-amps diagrams were taken. The area of a single electrode was 100 pm2; i. e. such an electrode overlapped together a group of clusters. The measurements were tested at normal temperatures in absence of illumination.
Fig. 12 shows an experimental volt-amps diagram of a group of logic cells, formed in a semiconductor. Here at the section of 0.0V to 0.1 V through this composite material flows the current, determined by a dark current of proper substrate semiconductor-Si. At gaining the voltage 0.1 V (point A) there appear conditions for
forming ring electrons in the cluster zone. The effective mass thereof increases causing thereby the decrease of the slope of volt-amps diagram (segment AB). At gaining the voltage of 0.2 V (point B) there appears the condition for electron pairing <BR> <BR> <BR> <BR> that result in an abrupt current slope. (line BC). At further voltage decrease to 0.11V (line CD) the current stays approximately constant. However if the voltage is increased again, the current will grow but with a smaller slope (line DF). As a result there is formed a closed contour at the volt-amps diagram, in other words-a hysteresis loop. The nature of these loops differs from the hysteresis loop at Fig. 11 by virtue of a high concentration of free carriers at normal temperature in the proper Si substrate.
At a proper control voltage such a cell may function as a current or charge switch for reading the stored photocarriers from the substrate material.
At other concentrations of doping agents the volt-amps diagram may be transformed into a N-shape form, that is characteristic for tunnel diodes (dashed lines 23 and 24, Fig. 12) or fully degenerate into a line.
Comparison with the prior art analogues.
The validity of the aforementioned analysis and experimental data may be additionally proved by independent known experimental data for commercial tunnel diodes.
Let us assume that for semiconductor tunnel diodes on p-n junction the process of forming ring electrons on clusters of doping agents starts in the point of peak voltage. In this case there is a probability of their condensation into drops or filaments along 2e, 3e, 4e... It is known that threshold rated voltages for commercial diodes of various kinds made in Russia are as follows: for germanium diodes- 1181104E-Un < 100mV, for gallium-arsenide diodes-3M2OlA-U, < Un < Theses Theses correspond correspond Un Un = nUe = 99.3 mV; 198.6 nV; 298.0 mV accordingly at n = 1,2,3.
The formed drops of ring electron will increase their masses mn*=nme/a, and, hence the current will go down under the geometric progression law at b<1: The decay of drops starts at gaining the drop size 4e for germanium diodes and 6e for gallium-arsenide diodes. The evaporation of drops occurs due to the
constriction of p-n junction to the size less than the length of the electronic filament < 2ro. Hence at b>1 the falling branch transfers into the ascending one.
Through statistic spread of clusters along the space (bulk) of the p-n junction the quantum steps of the current at normal temperatures are straighten. They may be described by experimental functions commonly used for tunnel-wave description of such effects. If we assume that a part of electrons has not condensed into drops, then the part of current caused by them will determine the excess current at the trough of the volt-amps diagram (shaded area, Fig. 12) [13].
The validity of the equation (12) supports the fact that the limiting rated generation frequency for any tunnel diodes is not more than f 40GHz (1104E).
This corresponds to data computed underthe formula (12), giving the generation <BR> <BR> <BR> <BR> frequency fe < 350 Such operation frequencies frequencies may retained retained to to Te <<BR> <BR> <BR> <BR> <BR> <BR> <BR> 1150K (877°C), what fully agrees with the maximum possible operation temperatures of gallium-arsenide diodes where the operation temperatures in the pulsed mode may go up to 870K (597°C) [20]. Further the material of proper diode degrades under the temperature, but with this the physics of the process itself is not violated. The general current density through p-n junction for such diodes is j ~ 1K/Vcm2, what agrees with (14).
The limiting field strength Ee (15) may cause evaporation of electronic drops both in tunnel p-n junctions and in metal-semiconductor junctions. For example for V203 the increase of the field 34 KV/cm leads to the fall of phase transition temperature from 145K (-128°C) to 100K (-173°C) and further at E=3Ee= 95 KV/cm to Tc~ OK (-273°C) [16, p. 16].
One more experimental proof may serve the work [9]. The author investigated the nanosized cell, filled with TiOX. As TiOX has a high-temperature MSPT (500K-600K [16]). In the film are created conditions for condensation and decay of electronic pairs.
This result in formation of four N-shaped segments on the volt-amps diagrams as shown at Fig. 13, line 25. This figure also shows dependencies computed according to the formula (19) for b<1 (line 26) and b>1 (line 27). As it is seen from the drawing threshold energies of experimental data fully correspond to our analysis. Minor variations are related to the fact that the thickness of titanium electrodes is in all of the order of 2 nm, what is well over ro, i. e., ring electrons are inclined in relation to the motion vector.
A memory cell is an important structure of any logic digital circuit. It may be designed keeping in mind the specialities of interaction of ring electrons (Fig. 2). In order to create a stable logic state we would need minimum two ring electrons, as in this case their overall energy will be minimal. Then using (11-13,16) we find the switching energy We = 2Ie2ReTe = 2me (crc)/2a = 3.18 10-'°J,(20) that is equivalent to 0.2 eV.
It is possible to show that the condition We 2 2kT/e. witl be met for the whole range of operating temperatures T<Te, i. e. the signal-to-noise ratio will be above two.
Hence the system will be always noise-immune at performing logic operations.
Embodiment of the invention The modern commercial lithographic techniques in 1998 do not allow designing elements smaller than 180-250 nm. Therefore at present we need a step-by-step conversion to nanosized electronic devices created on basis of submicron techniques.
Active logical films Making use of the suggested resonance OS-electronic cells in the form of fused single-and multilayer films-composites it is possible to design various digital and analog-digital devices of a micron and submicron size. It is apparent that multilayer active films may be designed of sphere-like clusters: Example 1. Fig. 14. shows a structure of a multiyear logic cell based on spherical clusters, where 28 are spherical clusters-cells; 29,30 are control electrodes ; 31 is the substrate. Bulk element, consisting of No clusters 28 that are laid in N layers, intervene orthogonal electrodes 29,30 and are located on the common substrate 31.
Single layer active films may be made of cylinder-like clusters: Example 2. Fig. 15 shows the structure of logic cell on basis of a group cylinder clusters, where 32 are cylinder clusters-cells; 33,34 are control electrodes; 35 are substrate. Cyllinder-like clusters are laid in a single layer between two orthogonal electrodes 33,34. The whole device is located on a common substrate 35.
In the both examples gaps between clusters may be filled with gas, dielectric or semiconductor.
Example 3. Fig. 16 shows a volatile memory matrix, where 36 are single spherical or cylinder clusters; 37,38,39,40 are orthogonal control electrodes. The
record and storage of information in such a cell are based on non-liner hysteresis characteristics (Fig. 11). At disconnection the cell discharges through the voltage distribution bus.
Example. 4. Fig. 17 shows a non-volatile memory matrix, where 41 are spherical clusters; are orthogonal control electrodes. This memory cell may comprise serially connected two and more clusters what permits to store information at disconnection. Overcharging of such a cell is performed by back voltage.
For example a submicron memory cell of a size 150x150x60 nm with an average number of clusters No = 50 and the number of layers N=2 may be created for such a cell of spherical clusters of a size 30 nm. There may be written from 2 to 100 elementary charges in these cells. In this case he time of storing information in such a cell is unlimited. In the storage mode the cell consumes no energy. Writing pulses U>2UeNe or reading pulses U<2UeNe may be fed to buses 42,44. A charge proportional to 0 or 1 of the logic signal may be taken off the control bus in the reading mode.
Active ana) og photo fitms.
Beside a digital level a fused multilayer film may store an analog level that is proportional to the total charge of elementary clusters.
Example 5. Using films, shown at Fig. 14, the free gaps of which are filled with light-sensitive semiconductor it is easily possible to create matrixes for TV or photo cameras. In this case cluster layer should absorb effectively the photons, i. e. the thickness of the film should be commensurable with light wavelength. This corresponds to 15-17 layers what constitutes of an order 500 nm. Naturally, in this case one electrode 29 should be made of a transparent conductor.
At designing megapickcell light-sensitive matrixes with orthogonal structure (Fig. 17,) a photosensitive cell will occupy an area of an order 100 lim2. In order to increase of the spectral photosensitivity the gaps between clusters should be completely filled with the semiconductor material and besides, it is desirable that the enclosures around the cluster nucleus were made of a light-sensitive semiconductor as well. This is connected with the fact that ring electrons as it was mentioned above have a small cross section of interaction with photons and the whole process of photo conversion takes place mainly in the semiconductor.
The work of the light-sensitive matrix is based on the known in the art principles -storing photo carriers during exposition of the image in the semiconductor within a picture. After the end of the exposition of the picture, the pulse signal initiating "breakthrough"of clusters is fed to the buses 29 and 30 (Fig. 14) and the stored photoelectrons are read at the buses in a standard way in analog form similar to digital memory (see. Fig. 17). Then the analog signal is amplified by the bus amplifier and thereafter digitised.
Example 6. A light-sensitive matrix may be designed on basis of materials with MSPT. To do this the gaps between cylinder cells (Fig. 15) are filled with light- sensitive semiconductor, and moreover the cell itself is made without tunnelly transparent gaps (Fig. 5). In this case the material with MSPT should work below the point of its phase transition into the metal phase. Photo information is stored similar manner to that described in the Example 5. Reading is performed due to the transition of the channel material into the metal phase under the electrostatic field of the electrodes 33,34 (Fig. 15) according to the formula (15).
Active analog distributed memory Each single cluster having a hysteresis characteristic may store up to 2 electrons. Therefore, by joining them in groups it is possible to make a discrete analog memory with an accuracy of storing analog information rising in proportion to the size of the cluster group.
Example 7. If for the base of the analog memory we take a film of the kind shown at Fig. 14, and supply to the buses 29,30 a voltage difference within the time of analog signal retrieval, then a part of clusters will « breakdown » and electrify. The quantity of the charge will determine the level of the « memorised » signal. The information may be read by reducing voltage at the cell < UeNC. Then a standard analog-digital converter may digitise the charge. The number of charged clusters is counted by differentiating the output signal and feeding it to the impulse conter. In the last case due to the large number of elementary clusters each cell has a kind of its own serial analog-digital converter with a capacity determined by log2NO.
Example 8. If a layer of clusters is placed between electrodes functioning as a distributed super high frequency resonator, then due to presence of the section of dropping characteristic at the volt-amps diagram (Fig. 11, line CD or Fig. 12, line BC or Fig. 10, line BO), it is possible to create a distributed super high frequency generator with the frequency determined by the resonator characteristics. The
maximum boundary frequency of such a generator is determined by the formula (12).
An important characteristic of such a generator is a low noise level.
Three-dimensional logic It is significant that spherical and sphere-like clusters allow three-dimensional connection with direct contact between cells. Such a contact functions as a control or a supply electrode. As a result such a structure allows to change the modern planar approach for the three-dimensional approach increasing thereby the circuit density and hence the operating speed and performance of integrated circuits. Three- dimensional logic is a future-technology in designing parallel matrix computing structures optimised for specific application as well as structures with a large branching level, e. g., in designing neurocomputers.
A large number of clusters connected in series will afford to compensate defective elements by a simple a statistic averaging, increasing thereby in general the yield ratio. Moreover, upon radiation of a layer tracks of the most dangerous heavy alpha particles will be localised by the bulk of clusters situated long the particle track.
Hence, radiation tolerance as well as temperature tolerance will sharply rise in comparison with classical semiconductor devices.
Spaced group logical elements.
An important positive characteristic of cluster films is a possibility to design in principle a homogeneous active computing medium allowing both two-and three- dimensional connection architecture. As it follows of the formula (3), the main problem in transmitting signals between nanoelements is presence of parasitic capacitances of connecting elements and output contacts of logical circuits. The problem may be solved by giving the amplifying functions to connecting elements, i. e., a signal should be additionally amplified in the process of passing along buses. The problem of adapting nanologic dimension to the dimensions of the IC contacts requires designing buffer power amplifiers having the operating speed close to the speed of a single cluster, i. e. it is necessary to create a certain scale transformer.
Some variants of designing distributed active logical devices are considered further.
Example 9. Fig. 18 shows a variant of connecting control resistive electrodes to a group of contacting clusters, where 46 is a spherical or cylinder cluster; 47,48 are supply electrodes, 49 is a spaced resistive layer; 50 is a control electrode; 51 is an isolating layer, 52 is the output electrode. Clusters 46 are located on metal substrate
48 and connected to the control electrode 47 through semiconductor medium 49. This medium is a spaced load. Such a load is used in neuristor lines on S-diodes [15], or in spaced tunnel junctions or Josephson spaced junctions [14]. It is important that the specific dielectric permeability of the semiconductor was £ >> 1, and the specific resistivity thereof could form the restive load, which crosses the dropping section of the volt-amps diagram of the cluster (Fig. 11, line 22). In this case a wave of switches having a form of a front edge or a soliton may spread in the cluster medium upon pulse supply to electrodes 48,47. Here the process of signal amplification is conditioned by a negative differential resistance (Fig. 11, section CD of the volt-amps diagram). Initiation of the process may be controlled by the electrode 50, and the signal pick-up-by the electrode 52.
Example 10. Fig. 19. Shows a variant of connecting control electrodes to the group of contacting clusters through an additional cluster layer, wherein 53 are spherical clusters; 54,55 are supply electrodes; 56 is the spaced additional cluster layer; 57 is the control electrode; 58 is the isolated layer; 59 is the output electrode.
Due to use of the additional cluster layer both is possible here: the process of transmission of a maintained wave, i. e. soliton, and also storage at one time of the element state at de-energising. Upon activation the process will go on from the pre- power-down state. The whole system resets at change of the supply polarity.
Example 11. Fig. 20 shows a group logic cell"OR", wherein 60 are spherical or cylinder clusters; 61,62,63 are input control electrodes; 64 is the output electrode. A start-up signal initiating the charge wave through clusters 60 comes through any of input electrodes 61,62,63. Electrodes 64 collect the amplified charge in clusters. All clusters are supplied by electrodes as shown at Fig. 18 (electrodes 47,48).
Electrodes 61,62,63 correspond to the electrode 50 and electrode 64 correspond to the electrode 52.
Example 1, 2. Fig. 21 shows a group logic cell"OR"with output branching, wherein 65 are spherical or cylinder clusters; 66,67,68 are input control electrodes; 69,70,71 are output electrodes. Here in the similar manner as in the example 11 the initial start-up of the charge wave is performed by any of the electrodes 66,67,68.
However the outputs are independent, branching the input signal to three outputs.
Clusters 65 are supplied in the same way as in example 11.
Example 13. Fig. 22 shows a group logic cell"OR"with a power output, wherein 72 are spherical or cylinder clusters; 73,74,75,76,77 are input control
electrodes; 78 is the output electrode. Here the input signal is supplied to any cluster group by means of any one of electrodes 73-77. The charge wave is amplified by a group of clusters and collecte by the output electrode 78. This electrode may function as an input of a standard logical element based on field or bipolar transistors of micron size or may directly be an output contact of a logic nanoelectronic microcircuit.
Directionality of a signal As it is shown on Fig. 1,3, cluster is an electrically symmetric element. Some applications yet require a signal with one directional propagation. In order to perform this function it is necessary to assign a correct direction of the electric field strength gradient. To do this it is required to form geometrically in the proper direction the electrode tapers connecting clusters or produce electrodes of materials with various output works.
Example 14. Fig. 23 shows a variant of integrating logical cells with targeted signal direction, wherein 79 are spherical or cylinder clusters; 80 are information electrodes; 81,82 are supply electrodes. Here the input signal propagates from left to right along the electrodes 80 through clusters 79. Both the continuous and the pulse supply is connected directly to electrodes 81,82.
Example 15. Fig. 24 shows a variant of integrating logical cells with targeted direction of signal passing and branching, wherein 83 are spherical or cylinder clusters; 84 are information electrodes; 85,86 are supply electrodes, 87 is the branched information electrode. Here the input signal is fed through the electrode 84, amplified in the cluster 83, and further branched into two or more outputs by the electrode 87 and fed to the similar device through electrodes 88.
Local logical elements Local logical elements may be created by using a method of targeted propagation of a signal.
Example 16. Using targeted electrodes it is only suffice to make a unidirectional shift register. Fig. 25 presents such a shift register, wherein 89 are spherical or cylinder clusters; arrows indicate information electrodes; 90,91 are <BR> <BR> <BR> <BR> <BR> supply electrodes; Ut, U2 are antiphase pulse supply voltages; U3 iS the input voltage; U4 is the output voltage. The out-of-phase pulse supply is fed to buses U1, U2. Logical <BR> <BR> <BR> <BR> signal from the input U3 is successfully shifted along the circuit at cells 89 to the output<BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> U4.
Example 17. Fig. 26 shows volatile logical element"NOT", wherein 92 is the spherical or cylinder cluster; Ui, is the pulse or direct supply voltage; U2 is the input voltage; U3 is the output voltage; R is the resistive load. The pulse or direct supply voltage is fed to the supply bus U1, the control electrode U2 functions as the control electrode similar to the gate function of the field transistor, i. e., performs the function of inversion-"NOT".
Example 18. Fig. 27 shows a non-volatile logical element"NOT", wherein 93 is a spherical cluster; 94 is the additional cluster; U1, is the pulse supply voltage; U2 is the input voltage; U3 is the output voltage. The device here works in the similar way as in the example 17, however the load is the additional cluster 94. Such a circuit serves for information storage at the off mode. The supply of the device unlike in the example 17 may be only the pulse supply. Cluster 94 needs overcharging. At de-energising two paired electrons may remain in one of the two serially connected clusters. In this case the pair of clusters is, in fact, a logical trigger.
Example. 19. A two level trigger for four states may be made by transverse joining of two logical elements"NOT"with memory (Fig. 27) and etc. Logical elements having the base >2 may be produced by connecting a large number of clusters in various configurations. Fig. 28 shows a non-volatile trigger, wherein 95 are spherical <BR> <BR> <BR> <BR> <BR> clusters; 96 are additional clusters; U, U2 are pulse supply voltages; Us, U4 are input voltages; U5, U6 are output voltages. The said trigger operates under the standard circuit using transverse joint of logical elements"OR"and"NOT".
Example 20. Joining of two devices depicted at Fig. 26 makes it possible to design a differential amplifier of analog signals on the basis of which various known analog devices may be produced. For example Fig. 29 shows a comparator of analog signals, wherein 97 is the spherical cluster or cylinder cluster; U, is the supply voltage; U2 is the input voltage; U3 is the reference voltage; U4, U5 is the output voltage; R.,, R are resistive loads; R3 is the common resistor.
As the aforementioned examples cover the most typical variants of connecting elements used both in analog and in digital devices for processing information, on the basis thereof those skilled in the art may easily combine any required IC elements.
This is made by a mere combination of a resistive or cluster load. The more so as clusters themselves, depending upon requirements, may be made on basis of OS-, PS-or APS-electronic effects. Supply electrodes may be made of metals, superconductors or materials with metal-semiconductor phase transition.
Example 21. As it follows of (16), the quantum resistance of a cell will exceed the Josephson resistance equal to h/2e, by 1/a. Thus, there may be designed a sample low noise resistance at normal temperature with an accuracy determined by <BR> <BR> common constants h, e, a, c, i. e. with an error of an order 10-'. It may be connected to the current source, e. g. as shown at Fig. 30 forming thereby a standard voltage source. Fig. 30 shows: 98 is the spherical cluster; I is the current source; U, U2 are standard output voltages.
Neurone logics.
Neurocumputers are the most advanced approach to the future development of parallel computing structures. The base element thereof should perform functions of threshold comparison and summing of several inputs with different weight [40].
Example 22. Fig. 31 shows a neurone logical element, wherein U1, U2, U3 are input voltages; Ai, A2, A3 are weight multipliers; 99 are analog summators; 100 is the threshold element; U4 is the summator output voltage; U5 is the output voltage.
Active films connected as shown on Fig. 20 and having only one input are useful for weight summing operations. A group of such elements joined together by a common electrode forms a non-volatile weight summator of a neurone logical element, as shown at Fig. 32, wherein 101 are weight groups of spherical clusters with weights A,, A2, A3; Ul U2, U3 are input voltages; U4 is the summator output voltage.
Due to discreteness of charge carriers the operation of weight multiplying is <BR> <BR> accomplished in a discreet manner. In this case the input signal U1, U2, U3 will multiply in clusters-cells in proportion to the number of these cells. U, will correspond to the weight A, =1, U2-to the weight A2 = 3 and U3 to the weight A3= 2. Summing of the signal is accomplished at the electrode 102.
The function of weight summing may be also accomplished by films having MSPT. Fig. 33 shows volatile weight summator of a neurone logical element on basis of materials with MSPT, wherein 103,104,105 are weight groups of clusters with <BR> <BR> weights Ai, A2, A3; U1, U2, U3 are input voltages; U4 is the summator output voltage; 106 is the output electrode. In this case the film area will determine the ratio of signal multiplication at the charge. The charge level will determine the weight according to A,, A2, A3 s in the similar way to the aforementioned case.
For devices shown at Fig. 32,33 the output signal U4 is suppiied to the comparison element made, e. g. according to the circuit (Fig. 29), wherein the threshold comparison is accomplished.
If the weight summator is made under Fig. 32, and the supply is fed through additional cluster layer under Fig. 19, then it is possible to perform weight summing with memory. This function with memory is the closest description of the real neurone.
Voltaic cross-coupling On frequent occasions combined operation of analog and analog-digital parts of an integrated circuit requires elimination of cross-voltaic effect along control and supply buses. For this may be used the cross effect of ring electrons in two electrodes located at the distance less than ro. When electrodes are present at such a distance there occurs current induction without any galvanic coupling. In this case altemating voltage need not be used, as the circuital field is the ring electron itself.
Example 23. Fig. 34 presents current transformer, wherein 107,108 are input electrodes made of superconductor or material with metal-semiconductor phase transition; 109 is an output electrode; 11,12are input currents; 13 is the output current.
The input currents 11 or 12 may be both alternating and direct. The output current 13 will be connected with input currents through transformation ratio.
Example 24. Fig. 35 presents a dc-to-ac transducer, wherein 110 is an input electrode made of superconductor or material with metai-semiconductor phase transition; 111 is the output electrode; 11, is the direct input current; 12 is the altemating output current. When one of the transformer electrodes is made in the square waveform having regions of approach with the output electrode < ro and distant regions with output electrode > ro, then currents in the output electrode 111 will'be induced only at the regions of approach. This induces altemating current in the output circuit, i. e., the said device allows transforming direct current into altemating current. If the electrode 111 is made of superconductor or material with MSPT connected to the alternating current, then the direct current will pass through the electrode 110. With this the function of rectification of altemating current is accomplished only at frequencies proportional to the time of carrier movement along the approach region.
At the same time the device performs galvanic de-coupling provided the voltage between the electrodes is not above 0.1V.
Technologica/implementation The possibility of producing the claimed logical cell on the modem technological base is illustrated below.
There are three methods of forming spherical and sphere-like particles [1]. The first method-metal or semiconductor clusters of a diameter up to 37 nm are formed of a gas phase with their further oxidation in the oxygen flow or similar chemicals.
Formation of such particle is similar to formation of hail in the Earth atmosphere. The second method is the colloidal method. It is based on cluster precipitation from metal salt solutions following by the chemical coating with corresponding enclosures. The third method is based on collective effects in solid supersaturated solutions similar to pyroceram.
Hollow clusters may be made of xerogels [41] or by gas blowing of the liquid drops at adding of volatile components-additives with further cooling in upward gas current [42]. Nanosized hollow spheres of zirconium dioxide are automatically obtained during the process of high-frequency plasma-chemical denitrification; therefore they may be applied to the substrate directly from plasma. [43].
Designing planar vertical nanochannels is based on collective formation methods, e. g. according to electrochemical oxidation Al, Ta, Nb, Hf, etc. The formed channel may be filled with metal or semiconductor by the galvanic technique [44].
The aforementioned examples show that the modem techniques allow producing nanometer logical cell and integrated circuits on the basis thereof. For example clusters may be quickly produced in colloidal solutions and precipitated on substrates by spinning within 60c. Clusters may be also produced by evaporation of a metal by means of a plasmatrone at a speed of the order 1 cm3/min followed by condensation in a gas flow and precipitation on a substrate. This results in producing clusters having d-35 nm. Both colloidal and vapour-phase techniques allow to obtain from 1 cm3 material about 2.3 1016 clusters that are full-value quantum devices, i. e. transistors. It should be noted that digital microcircuit personal computing machines manufactured in the world during a year comprise approximately the aforementioned number of transistors.
It is important that the cluster layer, which is not located directly under electrodes, forms a passive area that not involved into signal transforming. Due to collective methods of producing, the cost of clusters is so small that it practically makes no effect on the overall cost of a product.
Thus, nanotechnology allows just now provides possibilities for leaving the frameworks of silicon and gallium-arsenic techniques and starting to review the conceptions of designing solid-state electronic high-temperature microcircuits and to produce devices operable up to 600K (327°C).
It will be apparent to those skilled in the art that modifications of this invention may be practised without departing from the essential scope of this invention.
Information Sources.
1. Petrov U. I. Cluster and minor particles. M. Nauka. 1986,368 pp.
2. Likharev K. K. On possibilities of designing analog and digital integrated circuits on basis of the effect of discrete single electronic tunnelling. Microelectronics. Vol. 16, ed. 3.1987, p. 195-209.
3. Rouse A. Human sight and electronic vision. M. Mir. 1977 4. US 5731598 5. US 5420746 6. US 5677637 7. US 5694059 8. US 5389567 9. Matsumoto K., Sepawa K., Oka Y., Vartanian B. J., Haris J. S. Room temperature operation of a single electron transistor made by the scanning tunnelling microscope nanooxidation process for the TiOX/Ti system. Appl. Phys. Lett. 1996.
V. 68. N1.
10. High voltage technique: Theoretical and practical principles of use. M. Beier et al.
M. Energatomizdat 1989.
11. PCT WO 98/21754 of 22.05.1998 12. Buzaneva E. V. Microstructures of integral electronics. M. Radio. 1990.
13. Tunnel effects in solid bodies. Edited by E. Burstein, S. Lundquist. Translation from English. M. Mir. 1973.
14. Likharev K. K. Introduction to dynamics of josephson junction. M. Hauka. 1985.
320 pp.
15. Stafeev V. M., et al. Neuruistor and other functional circuits with three-dimensional coupling. M. Radio and communication. 1981.111 pp.
16. Metal-semiconductor-phase transition and use thereof. Bugaev A. A. et. al. Nauka.
1979.183 pp.
17. Demishev S. V., et al. ThermoEMF quasi two-dimensional organic conductors of the assemblage (BEDT-TTF) mXn. Journal of Theoretical and Experimental Physics.
1998. V113. Ed. 1. p. 323-338.
18. WO 97/36333.
19. US 5530263 20. Fistul V. I. Heavy doped semiconductors. M. Nauka. 1967.
21. Dubrovin B. A., et. al., Modern geometry. Methods of homology theory. M. Nauka.
1984.
22. L. De Broglie, D. Bohn, P. Hillion, F. Halbwachs, T. Takabayasi, G. P. Vigier. Rotator Model of Elementary Particles Considered as Relativistic Extended Structures in Mincowski. Phys. Rev. 1963. V. 129. Na 1. p. 438-440.
23. Buneman O. Proc. Cambr. Phil. Soc. 1954. V. 50. p. 77.
24. Motora I. M. Preprint P4-81-81. Assoiate Institute of Nuclear Researches, Dubna.
1981 25. Rabsha E. I., Timofev V. B. Hall quantum effects. (Review). Physics and techniques of semiconductors. 1986. V. 20. Ns6, p. 977.
26. Deriagin B. V et al. Surface forces. M. Nauka. 1985.
27. Quantum and exchange forces in condensed mediums. Kulakov A. V. et al. M.
Nauka. 1990.120pp.
28. Druszkin L. A. Objectives of the field theory. M. Energia. 1964 29. Maple M. B. New types of superconductivity in f-electronic systems. Physics abroad. Articles. Series A. M. Mir, 1987. pp.
30. Electronic-hole drops in semiconductors. Keldysh et al., Edited by Jefrice K. D. et al., M., Nuaka, 1988.477 pp.
31. Tennakone K., Lokunetti C. S. et. al. The Possibility of an Above-Room- Temperature Superconducting Phase in xCuBr. CuBr2. J. Phys. C Solid State Phys.
21.1988. P. L643-L647.
32. Riley J. F., Sampath W. S., at al. Meissner Effect up to 300 K in Microscopic Regions of Y-Ba-Cu-O. Phys. Rev. B. 1988. V. 37. N1. p. 559-561/ 33. Superconductivity Researchers Tease Out Facts From Artifacts. Science. 1994.
V. 265. p. 2014-2015.
34. Enicolopian N. S. et al., Possible superconductivity of oxidised polypropylene in the area 300 K. Letters to Journal of Theoretical and Experimental Physics. 1989.
V. 49. Ed.. 6. p. 326-330.
35. Masterov V. F. et al., High-temperature superconductivity in the system carbon- cuprum. Letters to Journal of Theoretical and Experimental Physics. 1994. Vol. 20.
Ed. 15. p. 17-21.
36. Obolensky M. A. et al., Anizothropy of critical current at vortexes pinning at twins in YBa2Cu307x monocrystalls. Superconductivity: Physics, Chemistry, and Techniques. 1994. Vol. 7. Ns1. p. 43-47.
37. Superconducting materials. M. Metallurgy. 1976.
38. Reznitskich O. G. et al., Study of normal state of electronic Y-123 spectrum by methods of microcontact and tunnel spectroscopy. Superconductivity: Physics, Chemistry, Techniques, 1994. Vol. 7. Na2. C. 322-326.
39. Hayakawa H. Technology of josephson computers. Physics abroad. Articles Series A., M. Mir. 1987.272 pp.
40. Neurocomputers and intellectual robots/Amosov N. M., et al., Naukova dumka, Kiew 1991.
41. JP 07021716 of 09.02.1995. Publication No. 08218163 of 27.08.1996.
42. Budov V. V. Physical-chemical processes in the technology of hollow glass microspheres. Glass and Ceramics. 1990. Ns3, p. 9-10.
43. Dedov N. V. et al., Structural studies of powders on basis of zirconium dioxide produced by HF-plasmachemical denitration method. Glass and Ceramics. 1991.
Ne10, p. 17-19.
44. Averianov E. E. Anodization manual, M., Mashinostroenie, 1988
Next Patent: SEMICONDUCTOR SYSTEM WITH TRENCHES FOR SEPARATING DOPED AREAS