Patent Application
20220292384
The invention relates to the field of devices with electrostatic control gates. The invention particularly applies to the field of quantum devices, quantum information processing, and quantum computing.
Quantum computing is based on the use of a quantum state with two measurable levels as an information vector, referred to as qubit or quantum bit, and the laws of quantum mechanics (superposition, entanglement, measurement) in order to execute algorithms. A quantum device with one or more spin qubits can be used to manipulate the quantum state of this or these qubits.
Spin qubits can be formed in semiconductor, advantageously, silicon. Semiconductor technologies are studied for producing qubits due to the high integration potential thereof, like conventional electronics. Electrons or holes are individually confined in quantum wells located in a cryostat at cryogenic temperatures (less than 4 K, or less than 1 K) produced in confinement structures of nanometric sizes defined electrostatically and, in the case of silicon, of similar architecture to that of MOSFETs. These confinement structures correspond to quantum dots. A quantum dot behaves as a potential well confining one or more elementary charges (electrons or holes) in a semiconductor region.
Applying a static magnetic field (e.g., between 100 mT and 1 T) makes it possible to remove the spin degeneracy of quasi-particles which is a quantum system with two measurable levels, known as “spin down” and “spin up”. The quantum state thereof is a superposition of these two states represented by the direction of an arrow in a sphere, known as the Bloch sphere. The manipulation (rotation) of these spins is performed using an alternating component of the magnetic field (of frequency for example between 1 GHz and several dozen GHz) or using an alternating electric field at the same frequencies in the case of sufficient spin-orbit interaction (case of holes in silicon) or in the presence of static magnetic field gradient (e.g., micro-magnet integration). Reading the spins (up or down) makes it possible to extract a portion of the quantum information according to the probability of it being in each state, determined by the quantum superposition.
To optimally control several quantum dots formed side by side in the same semiconductor portion, it is necessary to be able to adjust, in addition to the depth of each of the quantum dots in the semiconductor, i.e., the minimum potential of each of the quantum dots, the height of the tunnel barriers, or tunnel-effect barriers, which separate the quantum dots, independently.
The control of the height of the tunnel barriers and the minimum potential of each quantum dot is performed by means of voltages applied to independent electrostatic control gates. The solutions proposed to date consist of defining control gates positioned directly above each quantum dot and each tunnel barrier.
The document by C. H. Yang et al., “Dynamically controlled charge sensing of a few-electron silicon quantum dot”, AIP Advances, 1, 042111, 2011, describes for example the production of such gates by implementing two or three successive lithography cycles, with partial overlap of the different levels to manage misalignment problems. The document by R. Pillarisetti et al., IEDM 18, proposes the self-aligned production of such gates.
These solutions, however, have the drawback of requiring the production of electrostatic control gates with a density two times more than with which the quantum dots are produced. The gate pitch to be achieved with these solutions is between 40 nm and 60 nm, which renders gate production complex.
Similar problems to those described above for quantum devices also arise for other types of electrostatic control gate devices.
Therefore, there is a need to provide a device with electrostatic control gates wherein the structure enables easier production of the electrostatic control gates.
For this, a device is proposed including at least:
wherein, in a first configuration, the first and second electrostatic control gates are disposed above the first regions of the semiconductor portion, or wherein, in a second configuration, the first and second electrostatic control gates are disposed above the second regions of the semiconductor portion.
In this device, it is proposed to produce gate structures with two stages, or with two different superposed gates, the potential whereof applied to each of the two gates can be adjusted independently in relation to one another.
This superposition of the first and second gates therefore makes it possible to produce gate stacks with a pitch that is two times greater than in the solutions according to the prior art, thus reducing for the most part the problems associated with producing electrostatic control gates.
The second electrostatic control gates may be solely located in the vertical extension of the first electrostatic control gates.
The second electrostatic control gates may be entirely or completely located above a plane including the top faces of the first electrostatic control gates.
In a parallel plane with the top faces of the first electrostatic control gates, the cross-section of each second gate dielectric may have a similar shape and dimensions to those of the cross-section of at least one of the first and second portions of electrically conductive material disposed against said second gate dielectric.
The second gate dielectric may not be in contact with the semiconductor portion.
The second gate dielectric may rest solely on the first portions of electrically conductive material, or solely on the first portions of electrically conductive material and a part of lateral spacers disposed against the first portions of electrically conductive material.
In the first configuration, at least a part of each of the second regions of the semiconductor portion may not be covered by the first and second electrostatic control gates, and in the second configuration, at least a part of each of the first regions of the semiconductor portion may not be covered by the first and second electrostatic control gates.
The device may be such that:
The semiconductor of the first and second regions may be unintentionally doped.
Advantageously:
In the first configuration, because the two-stage gate structures are located above the quantum dots, the electric potentials applied to the first electrostatic control gates will make it possible to control essentially the minimum potentials of the quantum dots formed in the first regions of the semiconductor portion located below the first gates. Furthermore, the electric potentials applied to the second electrostatic control gates will make it possible to control essentially the heights of the tunnel barriers formed in the second regions of the semiconductor portion, because the first gates form screens against the electric fields created by the second gates.
In the second configuration, because the two-stage gate structures are located above the quantum dots, the electric potentials applied to the first electrostatic control gates will make it possible to control essentially the heights of the tunnel barriers formed in the second regions of the semiconductor portion located below the first gates, and the electric potentials applied to the second electrostatic control gates will make it possible to control essentially the minimum potentials of the quantum wells formed in the first regions of the semiconductor portion.
The quantum device may be used with hole spins and/or electron spins.
In this quantum device, the gate dielectrics do not correspond to tunnel oxides, i.e., do not allow the passage of the charges by the tunnel effect therethrough.
The device may further include electrical charge carrier reservoirs between which the first and second regions of the semiconductor portion are disposed. These electrical charge carrier reservoirs may correspond to doped parts of the semiconductor portion, and the first and second regions may include unintentionally doped (uid) semiconductor.
The device may be such that:
When the first electrostatic control gates each include a gate dielectric, the first gate dielectrics may each have a thickness between 2 nm and 10 nm.
The device may further include means for applying at least a first electric potential to each of the first portions of electrically conductive material and at least a second electric potential, of a separate value from that of the first electric potential, to each of the second portions of electrically conductive material. These application means may particularly include electrical contacts which are connected to the first and second portions of electrically conductive material, as well as the elements for transporting the desired electric potentials to these electrical contacts.
The device may be such that the first and second regions of the semiconductor portion are aligned along a first direction, or parallel with the first direction, and:
These alternative embodiments of the electrostatic control gates make it possible to adjust the effects of one of the gates in relation to the other by producing one of the first and second gates with a different width from that of the other gate. In the first case, by reducing the width of the first gate, the part of the semiconductor portion controlled by the second gate therefore has greater dimensions. In the second case, by increasing the width of the first gate in relation to that of the second gate, the part of the semiconductor portion controlled by the first gate has greater dimensions.
The device may be such that:
This advantageous configuration enables easy contacting of the first gates which are disposed below the second gates.
In this case, the device may be such that:
This advantageous configuration makes it possible to distribute the first electrically conductive contacts of the first gates on the first and second ends of the first portions of electrically conductive material, in alternation, which makes it possible to reduce the stress associated with the dimensions of the first electrically conductive contacts which can be greater than when the first electrically conductive contacts are produced next to one another on ends of the first portions all located on the same side.
The semiconductor portion may correspond to a semiconductor nanowire. A semiconductor nanowire corresponds to a semiconductor portion in which the diameter or thickness is nanometric and in which the geometry enables one-dimensional confinement of charges thanks to the electrostatic control of the gates of the device.
In a second embodiment, the device may be such that:
In this configuration, the first semiconductor regions may be formed below the top edges of the semiconductor portion.
Moreover, in this case, the device may be such that:
and further including a third electrostatic control gate comprising at least a third portion of electrically conductive material disposed facing at least a fourth region of the semiconductor portion not covered by the first and second electrostatic control gates.
In the case of a quantum device, the fourth region of the semiconductor portion may form second tunnel barriers each separating two of the quantum dots formed at the two top edges.
In this configuration, an arrangement of several quantum dots disposed pairwise facing one another along the nanowire may be obtained. Thus, a two-fold greater number of quantum dots can be formed in the semiconductor nanowire.
In a third embodiment, the first regions of the semiconductor portion may be arranged in the form of a two-dimensional array, several of the first and second electrostatic control gates being arranged in the form of rows and the other first and second control gates being arranged in the form of columns.
In this third embodiment, the semiconductor portion may correspond to a semiconductor layer of nanometric thickness or to a grating of nanowires intersecting by forming an array.
The semiconductor portion may be disposed on a buried dielectric layer of a semiconductor-on-insulator type substrate.
The device may further include dielectric spacers covering lateral sides of the first and second electrostatic control gates, and disposed, in the first configuration, above the second regions of the semiconductor portion, or, in the second configuration, above the first regions of the semiconductor portion.
The dielectric spacers may include at least a first dielectric material covering the lateral sides of the first portions of electrically conductive material and at least a second dielectric material covering the lateral sides of the second portions of electrically conductive material, the permittivity of the first dielectric material being less than that of the second dielectric material.
The present invention will be understood more clearly on reading the description of embodiment examples given merely by way of indication and not limitation with reference to the appended drawings wherein:
Identical, similar, or equivalent parts of the various figures described hereinafter bear the same reference numbers so as to facilitate the transition from one figure to another.
The various parts in the figures are not necessarily shown on a uniform scale, to render the figures more readable.
The various possibilities (alternatives and embodiments) should be understood as not being mutually exclusive and may be combined with one another.
A device 100 with superposed electrostatic control gates according to a first embodiment is described hereinafter with reference to
The device 100 includes a semiconductor portion 102. The semiconductor of the portion 102 corresponds for example to silicon, or alternatively to GaAs.
The portion 102 is for example disposed on a dielectric layer 104 which may include SiO2.
Advantageously, the portion 102 corresponds to a part of a superficial layer of a silicon-on-insulator type substrate, e.g., SOI (Silicon-On-Insulator). The dielectric layer 104 corresponds in this case to the buried dielectric, or BOX (Buried-Oxide), layer of this substrate.
In the embodiment example described here, the portion 102 corresponds to a nanowire comprising the following features:
The device 100 includes several qubits produced side by side in the portion 102. Each of these qubits comprises a quantum dot formed in a first region 108 of the portion 102 and disposed between two tunnel barriers formed by second regions 110 of the portion 102. In
The device 100 includes first electrostatic control gates each comprising at least a first portion of electrically conductive material 112 and at least a first gate dielectric 114 disposed between the portion 102 and the first portion of electrically conductive material 112. In the embodiment example described here, the first gate dielectrics 114 each have a thickness between 2 nm and 10 nm, for example equal to 5 nm, and include SiO2 and/or HfO2 and/or Al2O3. The first portions 112 each have a thickness between 5 nm and 30 nm, for example equal to 10 nm, and include degenerate doped polysilicon and/or TiN and/or W.
The device 100 also includes second electrostatic control gates disposed on the first gates. These second gates each include at least a second portion 116 of electrically conductive material and at least a second gate dielectric 118 disposed between the second portion 116 and the first portion 112 whereon the second gate is disposed. Thus, each of the first gates is disposed between the semiconductor portion 102 and one of the second gates. In the example described here, the second gate dielectrics 118 each have a thickness between 2 nm and 15 nm, for example equal to 10 nm, and include SiO2 and/or HfO2 and/or Al2O3, and the second portions 116 each have a thickness between 5 nm and 40 nm, for example equal to 20 nm, and include degenerate doped polysilicon and/or TiN and/or W.
In the embodiment example described here, the first and second gates are disposed above the first regions 108 of the portion 102.
As seen in
The second electrostatic control gates are entirely or completely located above a plane including the top faces of the first control gates.
In a parallel plane with the top faces of the first control gates, the cross-section of each second gate dielectric 118 has a similar shape and dimensions to those of the cross-section of at least one of the first and second portions of electrically conductive material 112, 116 disposed against said second gate dielectric 118.
The second gate dielectric 118 is not in contact with the portion 102. In the configuration seen in
The device 100 also includes electrical charge carrier reservoirs 120 between which the first and second regions 108, 110 are disposed. In the embodiment example described here, the reservoirs 120 correspond to doped semiconductor portions formed by epitaxy on ends 121 of the portion 102. Alternatively, these reservoirs 120 may correspond to the ends 121 of the portion 102 wherein dopants are directly implanted.
The first and second regions 108, 110 are here aligned along a first direction which is parallel with the length of the nanowire formed by the portion 102 (parallel with the axis X in
Though not visible in
In order to be able to access the first gates which are located below the second gates electrically, at least one of the ends of each of the first portions 112 is not covered by the second gate dielectrics 118 and the second portions 116. In the embodiment example shown in
Advantageously, and as is the case in the configuration shown in
The device 100 also includes third electrically conductive contacts 130 produced on the reservoirs 120.
The contacts 126, 128, 130 include for example at least one metal such as for example Ti, TiN or W.
The device 100 also includes dielectric spacers 132 covering lateral sides of the first and second gates. The dielectric spacers 132 include for example SiO2. In the configuration in
All the elements of the device 100 described above other than the contacts 126, 128 and 130 are covered by an etch stop layer 134, including for example SiN. The whole is covered by a pre-metal dielectric material 136, including for example SiO2 and wherein the top surface is planarized and through which the contacts 126, 128 and 130 are produced. In
Thus, the device 100 includes, at each qubit, a gate structure with two stages, or with two different superposed gates, where the electric potential whereof applied to each of the two gates can be adjusted independently in relation to one another. Because, in the embodiment example described here, the two-stage gate structures are located above the quantum dots, the electric potentials applied to the first electrostatic control gates will make it possible to control the minimum potentials of the quantum dots formed in the first regions 108 located below the first gates, and the electric potentials applied to the second gates will make it possible to control the levels of the tunnel barriers formed in the second regions 110. Because the first gates form screens against the electric fields created by the second gates, these fields will therefore serve to control not the minimum potentials of the quantum dots formed in the first regions 108, but to control the levels of the tunnel barriers in the second regions 110 not covered by the first gates.
For the tunnel barriers which are located between two quantum wells, the level of each of these tunnel barriers is controlled by means of the two second gates located above the two quantum wells adjacent to the tunnel barrier.
The field generated by each of the first gates will, notwithstanding, have an influence on the levels of the tunnel barriers adjacent to the quantum dot controlled by each of the first gates. Furthermore, the electric field generated by each of the second gates also has an influence on the minimum potential of the quantum well present below each of the second gates. Thus, the values of the electric potentials to be applied to each of the first and second gates are chosen accounting for the influence of the other gates present in the vicinity thereof.
These two alternative embodiments make it possible to adjust the effects of the first and second gates in relation to one another, according to the values of the minimum potentials sought for the quantum dots and the levels of the tunnel barriers formed in the portion 102.
In the devices 100 described above with reference to
In the devices 100 described above, the first and second gates are disposed above the first regions 108 wherein the quantum dots are formed.
In a second configuration, the first and second gates may be disposed above the second regions 110 which form the tunnel barriers. In this case, the electric potentials applied to the first gates make it possible to control the level of the tunnel barriers formed in the second regions 110, and the electric potentials applied to the second gates make it possible to control the minimum potentials of the quantum wells formed in the first regions 108. The level of each quantum well is controlled by means of the two second gates located above the two tunnel barriers adjacent to the quantum well.
Given that the first and second regions 108, 110 include a similar material, the fact that the device 100 is found in one or the other of these two configurations is dependent on the values of the electric potentials applied to the first and second portions 112, 116.
In the case where electrons are confined in quantum dots, and relative to the work functions of the different materials present, more positive (or negative) voltages on the first (or second) gates will produce the first configuration, whereas more negative (or positive) voltages on the first (or second) gates will produce the second configuration. In the case of quantum dots confining holes and not electrons, the voltage polarizations described above are inverted.
A device 100 according to a second embodiment is described hereinafter with reference to
In this second embodiment, each of the first and second gates does not cover the portion 102 across the entire width of the portion 102, but solely on a part of this width so as to cover one of the top edges 146, 148 of the portion 102, the other top edge of the portion 102 being covered by other first and second electrostatic control gates. With such control gates, quantum dots are therefore formed below each of the top edges 146, 148 of the portion 102, which makes it possible, in relation to the first embodiment, to double the number of quantum dots that can be obtained in the portion 102. The quantum dots located facing one another below the two top edges 146, 148 are separated by a tunnel barrier in which the potential is controlled by a third electrostatic control gate comprising at least a third electrostatic control gate comprising at least a first portion of electrically conductive material 150 and at least a third gate dielectric disposed between the portion 102 and the third portion 150. This third gate is disposed facing a fourth region 152 of the portion 102 which is not covered by the first and second gates and wherein the tunnel barriers separating each of the two quantum dots formed below the two top edges 146, 148 are formed. In the embodiment example shown in
As in the first embodiment, the values of the electric potentials to be applied to each of the first and second gates and to the third gate will be chosen accounting for the influence of the other gates present in the vicinity thereof.
Detailed features of an embodiment example of the device 100 according to this second embodiment are given below:
The various alternative embodiments and other features described above for the first embodiment may be applied to the device 100 according to the second embodiment.
A device 100 according to a third embodiment is described hereinafter with reference to
In this third embodiment, the qubits are arranged in the form of a two-dimensional array of quantum dots. Unlike the previous embodiments, the quantum dots are not produced in a semiconductor nanowire, but in a semiconductor portion corresponding to a semiconductor layer 101 or to a grating of nanowires intersecting by forming an array.
Several of the first and second electrostatic control gates are arranged in the form of rows and bear the reference 154, and other first and second gates are arranged in the form of columns which bear the reference 156 and which are superposed on the gate rows 154 at the intersections between the gate rows and columns.
In this third embodiment, the quantum dots are formed in the regions of the portion 102 which are not covered by the first and second gates. These quantum wells are symbolically represented, in
The various alternative embodiments and other features described above for the first embodiment may be applied to the device 100 according to the third embodiment.
In all the embodiments and alternative embodiments described above, the semiconductor portion 102 may be part of a stack of several different semiconductors forming a heterostructure. Such a heterostructure corresponds for example to an alternating stack of layers of Si and SiGe, or of layers of GaAs and AlGaAs, or of Ge and SiGe. In this case, it is possible that the first electrostatic control gates may not include the first gate dielectric(s) 114, and the portions of electrically conductive material 112 may be disposed directly against the stack of semiconductors. The interface(s) between the stack of semiconductors and the portion(s) of electrically conductive material 112 of the first electrostatic control gates form in this case Schottky barriers, and the semiconductor portion 102 comprising the first and second regions 108, 110 corresponds to the layer of the heterostructure wherein the two-dimensional electron gas (2DEG) is formed. Examples of choices of metallic and semiconductor materials for forming Schottky barriers are for example described in the documents “The role of surface-gate technology for AlGaAs/GaAs nanostructures”, R P Taylor 1994 Nanotechnology 5 183, and “High-mobility Si and Ge structures”, Friedrich Schaffler 1997 Semicond. Sci. Technol. 12 1515.
A method for producing the device 100 according to the first embodiment is described hereinafter with reference to
The device 100 is produced from a semiconductor layer 101 which, in the embodiment example described here, corresponds to the superficial layer of a semiconductor-on-insulator type, and more particularly SOI type, substrate (
The layer 101 is then etched so that at least one remaining portion 102 forms a nanowire wherein the quantum dots of the device 100 will be produced. For this, a first etching mask 158, including for example a resin, is deposited on the semiconductor layer 101 (
Alternatively, it is possible not to implement the steps described above with reference to
A stack of material layers intended to produce the first and second electrostatic control gates is then deposited by covering the portion 102 and the dielectric material of the layer 104. This stack includes a first dielectric layer 160 intended to produce the first gate dielectrics 114, a first electrically conductive layer 162 intended to produce the first portions of electrically conductive material 112, a second dielectric layer 164 intended to produce the second gate dielectrics 118, and a second electrically conductive layer 166 intended to produce the second portions of electrically conductive material 116. In order to be able to etch this stack of layers 160, 162, 164, 166 and form the first and second electrostatic control gates, a hard mask layer 168, corresponding for example to a layer of SiN of thickness equal to 30 nm, is deposited on this stack (
A second etching mask 170 is then deposited on the hard mask layer 168, the pattern whereof corresponds to that of the control gates to be produced (
The dielectric spacers 132 are then produced by covering the lateral sides of the first and second gates, these lateral sides corresponding to the faces of the first and second gates which are perpendicular to the surface of the portion 102 whereon the gates are produced. In the example shown in
As shown in
The remaining hard mask portions 172 are then removed (
A third mask 174 is then produced so as to cover the second portions 116, except for the regions covering the first and second ends 122, 124 of the first portions 122 at which the first electrically conductive contacts 126 are intended to be produced (
The parts of the second portions 116 and the second gate dielectric 118 not covered by the third mask 174 are then etched, thus revealing the first and second ends 122, 124 of the first portions 112. The third mask 174 is then removed (
An etch stop layer 134 and a pre-metal dielectric material 136 are then deposited and planarized (see
A fourth mask 176 is then produced on the structure produced, wherein the openings define the positions of the electrical contacts of the device 100: first openings 178 define the positions of the first contacts 126, second openings 180 define the positions of the second contacts 128, and third openings 182 define the positions of the third contacts 130 (see
The device 100 is then completed by etching the pre-metal dielectric material 136 and the etch stop layer 134 plumb with the openings 178, 180 and 182 of the fourth mask 176, thus forming the positions of the first, second and third contacts 126, 128, 130, then by depositing at least one electrically conductive material in these positions and planarizing the electrically conductive material, thus forming the first, second and third contacts 126, 128, 130. The fourth mask 176 is then removed. The device obtained after this method corresponds to that shown in
In the method described above, the dielectric spacers 132 are produced by depositing a single dielectric material. To produce dielectric spacers 132 as described above with reference to
As shown in view a) in
The layer 184 is then etched so that the remaining portions form the first parts 138 which cover the lateral sides of the first portions 112 and the first gate dielectrics 114 (view b) of
A layer 186 including the second dielectric material is then deposited in a conforming manner (view c) in
In the embodiments and alternative embodiments described above, the device 100 includes several spin qubits coded by an electron stored in a quantum dot. Alternatively, each qubit may be coded by several electrons stored in several neighboring quantum dots, e.g., double or triple quantum dots.
In the different embodiments and alternative embodiments described above, the device 100 corresponds to a spin qubit quantum device. These different embodiments and alternative embodiments may, however, be applied for a device 100 not corresponding to a quantum device and used in a similar operating mode to that of field-effect transistors coupled in series. Such a device 100 may function in a cryogenic environment, i.e., in a cryostat at cryogenic temperatures (less than 1 K, or even less than 4 K). In such an operating mode, the electrostatic control gates of the device 100 are used as conventional transistor gates, the polarizations whereof control the conductance between the regions 121 of the structure.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2102536 | Mar 2021 | FR | national |
