Patent Application
20220149216
Aspects of the present disclosure are related to a processing element for an advanced processing system, and particularly, but not exclusively, to a quantum processing system and a processing element for a quantum processing system.
The developments described in this section are known to the inventors. However, unless otherwise indicated, it should not be assumed that any of the developments described in this section qualify as prior art merely by virtue of their inclusion in this section, or that those developments are known to a person of ordinary skill in the art.
Large-scale quantum processing systems hold the promise of a technological revolution, with the prospect of solving problems which are out of reach with classical machines. To date, a number of different structures, materials, and architectures have been proposed to implement quantum processing systems and fabricate their basic information units (or quantum bits).
One way of fabricating quantum bits, for example, is to use silicon quantum dots. In this case, quantum dots are formed using an interface between a 28Si substrate and a dielectric material. A confining arrangement is utilized for confining one or more electrons in the silicon substrate to form the quantum dot and a control arrangement (e.g., a gate) is formed on the dielectric material to control the confined electron (e.g., by applying a voltage to tune the electronic spin resonance frequency of the confined electron). One technique for fabricating such quantum dots and processing systems utilizing these quantum dots is described in International Patent Applications PCT/AU2014/000596 and PCT/AU2016/050713, which are incorporated herein in their entirety.
When a number of such quantum dots are fabricated at a suitable distance from each other, quantum information can be moved through the quantum dot array via spin shuttling or exchange mediated coupling of the electrons.
Electron spin-based quantum bits show high control fidelity and can utilise the fabrication technologies that already exist for the manufacture of metal-oxide-semiconductor-field-effect-transistors (MOSFETs), making them a popular choice for semiconductor based quantum bits. However, they may suffer from short coherence times.
Accordingly, an improvement is desirable.
According to an aspect of the present disclosure, there is provided a processing element for a quantum processing apparatus, the processing element comprising: a silicon substrate; a dielectric material, wherein the silicon substrate and the dielectric material form an interface; an electrode formed on the dielectric material for isolating one or more electrons in the silicon substrate to form a quantum dot; a group IV atom having a nuclear spin located in the wavefunction of the one or more electrons, where the nuclear spin of the group IV atom is entangled with the one or more electrons; and a control arrangement for controlling a quantum property of the quantum dot and/or nuclear spin of the group IV atom to operate as a qubit.
In some embodiments, the atom may be a 29Si atom. In other embodiments, the atom may be an atom of a group IV isotope that has a nuclear spin. For example, the atom may be a germanium-73 (73Ge) atom having a spin of 9/2. In other examples, the atom may be a carbon-13 (13C) atom, a tin-115 (115Sn), tin-117 (117Sn) or tin-119 (119Sn) atom.
In some embodiments, where the atom is a silicon-29 atom, the silicon substrate is an isotopically enriched 28Si substrate. In certain embodiments, the isotopically enriched 28Si substrate contains less than or equal to 800 ppm of 29Si atoms. It will be appreciated that the 29Si atoms can naturally exist in the isotopically enriched 28Si substrate or they can be engineered into a pure 28Si substrate—i.e., located within the wavefunction of a quantum dot electron.
Further, the quantum dot electron wavefunction diameter of the processing element can be less than or about 50 nm. And in some preferred embodiments, this diameter can be less than or about 15 nm.
It will be appreciated that the nuclear spin of the group IV atom can be entangled with the electron of the quantum dot via a several mechanisms. One such mechanism is hyperfine interaction between the electron and the nuclear spin. The strength of the hyperfine interaction can vary depending on the size of waveform diameter of the electron and the location of the group IV atom within the waveform. In some embodiments, the hyperfine interaction between the electron and the group IV atom is between about 100 KHz-1 MHz.
According to another aspect of the present disclosure, there is provided a method of operating a plurality of quantum processing elements, each quantum processing element comprising a silicon substrate, a dielectric material, wherein the silicon substrate and the dielectric material form an interface, an electrode formed on the dielectric material for isolating one or more electrons in the silicon substrate to form a quantum dot, a group IV atom having a nuclear spin located in the wavefunction of the one or more electrons, the nuclear spin of the group IV atom entangled with the one or more electrons, and a control arrangement for controlling a quantum property of the quantum dot and/or the nuclear spin to operate as a qubit. The method includes the step of applying a signal via the control arrangement to control the state of the qubit in a quantum processing element.
In certain embodiments, the method further includes applying a signal via the control arrangement to store information in the qubit. This can include applying the signal to store information in the electron spin of the quantum dot and swapping this information from the electron spin to the nuclear spin of the 29Si atom.
In certain embodiments, the method further includes transferring information from a first processing element of the plurality of processing elements to a second processing element of the plurality of processing elements. This includes swapping the information from the nuclear spin of the group IV atom of the first processing element to the electron spin of the first processing element, transporting the electron spin from the first processing element to the quantum dot of the second processing element, causing the transported electron spin to entangle with the nuclear spin of the group IV atom of the second processing element, and swapping the quantum information from the transported electron spin to the nuclear spin of the group IV atom of the second processing element.
According to yet another embodiment, a method for manufacturing an advanced processing apparatus is disclosed. The method comprises the steps of: manufacturing a plurality of processing elements by: providing a silicon substrate; comprising a 28Si layer; forming a dielectric layer in a manner such that the dielectric layer and the 28Si layer form an interface; forming a plurality of electrodes suitable to isolate one or more electrons about the interface to define a plurality of quantum dots; locating one or more group IV atoms having nuclear spin in the wavefunction of the one or more electrons such that the nuclear spins of the one or more group IV atoms entangles with the electrons of the quantum dots such that the pair of quantum dots and nuclear spins operate as qubits; forming a plurality of control members comprising switches arranged to interact with the plurality of electrodes; and forming a plurality of control lines; each control line being connected to one or more control members to enable simultaneous operation of the plurality of processing elements; wherein the plurality of electrodes, control members and control lines are formed by using a silicon metal-oxide-semiconductor manufacturing process.
Also disclosed is a quantum processing apparatus, comprising: a plurality of quantum processing elements arranged in a matrix, each processing element comprising: a silicon substrate and a dielectric material forming an interface, an electrode arrangement suitable to confine one or more electrons in the silicon to form a quantum dot and a nuclear spin entangled to the one or more electrons; a plurality of control members disposed about the processing elements; each control member comprising one or more switches arranged to interact with the electrode arrangement to perform quantum operations with the processing elements; and a plurality of control lines; each control line being connected to a plurality of control members to enable simultaneous operation of a plurality of processing elements.
While the invention is amenable to various modifications and alternative forms, specific embodiments are shown by way of example in the drawings and are described in detail. It should be understood, however, that the drawings and detailed description are not intended to limit the invention to the particular form disclosed. The intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the present invention as defined by the appended claims.
Typically, the surface silicon layer in which the gate-based quantum dots are formed is isotopically enriched to predominately contain 28Si atoms, which are known to have no nuclear spin of their own and therefore allow for long coherence times in the quantum dots thus formed. Even isotopically enriched silicon typically includes a small amount of 29Si atoms. These 29Si atoms have a nuclear spin and can sometimes affect the coherence times of quantum dots.
The inventors of the present application have however found that the nuclear spin of a 29Si atom can intimately interact with a nearby quantum dot and specifically with the electron spin of the nearby quantum dot (e.g., via hyperfine coupling) such that the pair of nuclear spin and quantum dot can form a quantum bit (or qubit).
Electron spins typically have poor information storage capabilities (e.g., an electron spin can generally store information for 5-10 microseconds) but they are very mobile and can be used to move information about in a quantum processing system. On the other hand, nuclear spins have good storage capabilities (e.g., a nuclear spin can typically store information for 5-10 milliseconds) but nuclear spins are fixed in the lattice and immobile and are generally not suitable for transferring information.
In the present disclosure, by combining a quantum dot with a 29Si atom, the properties of electron and nuclear spins can be utilized together to form a processing element such that the resulting processing element has high information storage capabilities and high mobility. In one example, the quantum information may be stored in the nuclear spin until it is ready for manipulation, at which stage the quantum information is swapped back to the electron of the quantum dot.
In order to form such a processing element, the quantum dot needs to be small enough and the density of 29Si atoms needs to be sufficient enough such that there is a strong hyperfine coupling (relative to the electron spin resonance linewidth) between the electron spin of the quantum dot and a nearby 29Si atom. In some embodiments, the wavefunction diameter of the electron in a quantum dot is less than 50 nm and in some preferred embodiments the diameter is about 15 nm or less (e.g., between 8-15 nm). In such cases, the density of 29Si atoms in the silicon substrate can be between 50-800 ppm.
It will be appreciated that in some cases, e.g., when the density of 29Si atoms is low or naturally non-existent in the silicon substrate (e.g., if it is pure 28Si substrate), the 29Si atoms can be precisely placed within the wavefunction of the electron(s) of a quantum dot (e.g., via ion implantation) such that they entangle with the electron(s).
The strength of the hyperfine interaction between the electron and nuclear spins is set by the size of the wavefunction of the electron in the quantum dot and the placement of the 29Si atoms within that wavefunction. The stronger the quantum dot is confined (i.e., the smaller the size of the quantum dot), the smaller the wavefunction of its electron and larger the hyperfine coupling between the electron and a nuclear spin located in the wavefunction of the electron.
Accordingly, it will be appreciated that if the quantum dots are too large, the wavefunction of the quantum dots would be large, resulting in a smaller hyperfine interaction between the electron and nuclear spins located in the wavefunction of the electron. Similarly, if the density of 29Si is high, a large number of 29Si atoms may be located in the wavefunction of the electron which can act as background noise and reduce the coherence time of the quantum dot. Alternatively, if the density of 29Si is low and the quantum dot is very small, a 29Si atom may not be located in the small wavefunction of the electron and therefore may not be able to entangle with the electron spin of the quantum dot. Accordingly, a balance between the quantum dot size and 29Si atom density is important.
In order to allow entanglement between the 29Si nuclear spin and quantum dot electron spin states and operate the pair as a qubit, specific operational frequencies should be used. In addition, for the qubit to be able to transfer information along a quantum processing system, the hyperfine coupling between the electron and nuclear spins of the qubit should be such that the movement of quantum information through the quantum dot array via spin shuttling or exchange mediated coupling is allowed. For example, the hyperfine coupling between the electron and nuclear spins could be weaker than the quantum dot inter-dot coupling.
It will be appreciated that the qubit thus formed can be used to form nuclear spin-based quantum processing systems or electron spin-based quantum processing systems. Further, the resulting device can function as a memory device for storing quantum information or as an information processing device, or both.
Referring to
In this embodiment, the processing element 100 comprises a silicon substrate 102 and a dielectric 104. The silicon substrate 102 in this example is isotopically enriched silicon having 29Si atoms density of less than or equal to 800 ppm and the dielectric 104 is silicon dioxide. The isotopically enriched silicon may be an epitaxial layer grown on a conventional silicon substrate. It will be appreciated that in other examples, the density of 29Si atoms in the silicon substrate may be different and outside this range. For example, in some cases the silicon substrate may have a lower density than 50 ppm and 29Si atoms may be engineered or fabricated into the silicon substrate such that they are located in and around the wavefunction of the electron(s) in the quantum dot.
A gate electrode 106 is provided and is operable to form a quantum dot proximate the Si/SiO2 interface. The gate electrode 106 is also arranged to modify a quantum property of the quantum dot. In one example, the quantum dot includes one or more electrons 108 and the electrode gate 106 is configured to control the effective g-factor of the electrons 108. Throughout this specification, the term “effective g-factor”, is used broadly to indicate a ratio between the spin resonance frequency for the spin system and the DC magnetic field.
In addition to controlling a quantum property of the quantum dot, the gate electrode 106 may also be used to directly control the spin of the electrons 108 using an AC electric or magnetic field.
In order to form electrons 108 in the quantum dot, sufficiently positive voltages are applied to the gate electrode 106. This causes electrons in the area below the gate 106 (see
It will be appreciated that the gate electrode 106 is also used to control a quantum property of the 29Si atom and/or control the spin of the nucleus. In addition, the gate electrode 106 can be used to control the spin of the 29Si nucleus (e.g., using similar techniques to those used for electron spin resonance), swap quantum information between the nucleus and the electron and/or between two such processing elements.
The quantum device 200 includes an isotopically enriched 28Si layer 202 topped by a dielectric layer 204. Gate electrodes 206A-206P are formed on the dielectric layer 204. When a sufficiently large positive voltage is applied via the gate electrodes 206, one or more electrons 208 are isolated in the area under each gate electrode 206. These isolated electrodes 208 then entangle with one or more 29Si atoms 210 located in the wavefunction of each of the electrons 208 to form entangled electron and nuclear spins, where each combination of quantum dot and 29Si atom can be used to encode a qubit.
In the illustrated example, an array of 4×4 qubits can be encoded in the quantum device 200. In this case, an electromagnetic field may be applied to the gate electrodes 206 to control inter quantum dot coupling and the hyperfine coupling between the electron and nuclear spins in a particular processing element.
In certain embodiments, the gate electrodes 206 and hence the quantum dot location can be lithographically defined anywhere on the quantum device 200 providing flexibility in the device design. Typically, however, the requirements for forming such processing element (i.e., optimal dot size and 29Si atom density, and trade-off between the hyperfine coupling and inter-dot coupling) would be satisfied with a quantum device design having gate widths of less than 30 nm fabricated on top of a 800 ppm isotopically enriched silicon substrate, separated by a high-quality thermally-grown oxide.
The electron spin resonance (ESR) frequency (fESR) of an entangled nuclear and electron spin is—
fESR=|γeB|+A/2, when the nuclear spin is up
fESR=|γeB|−A/2, when the nuclear spin is down
Where γe is the electron gyromagnetic ratio, A is the electron-nuclear hyperfine interaction and B is the DC magnetic field. Further, spin up can be defined as a spin being parallel to, or aligned with, the DC magnetic field (B), whereas spin down can be defined as anti-parallel, or aligned opposite to, the DC magnetic field.
Performing a spin inversion with a pi-pulse on one of these frequencies implements a 2-qubit operation known as the controlled-not (CNOT) gate, which can be used to generate entanglement.
As noted previously, the gate electrode can be used to control the nuclear spin of the 29Si atom. In particular, the nuclear spin can be controlled via an AC magnetic field, in the same way electron spin resonance is performed on the electron of the quantum dot.
Depending on whether the quantum dot is loaded with an electron or not, the resonance frequency will be either
|γ_N B|, when no electron is loaded
|γ_N B|+A/2 when the electron is loaded with spin down
|γ_N B|−A/2 when the electron is loaded with spin up
Typical rotation Rabi frequencies that can be obtained are between 1-10 kHz.
As noted previously, nuclear spins have longer quantum memory storage capabilities than electron spins. To take advantage of this, the processing elements of
To this end, once quantum information is encoded on the electron spin of a quantum dot using the mechanisms described in International Patent Applications PCT/AU2014/000596 and PCT/AU2016/050713, a memory operation is performed to map the qubit information encoded in the electron spin state onto the nuclear spin state.
This operation, known as a SWAP operation, can be implemented with three controlled rotations and typically includes three stages—initialisation, transfer and recovery. In one embodiment, nuclear initialisation in the spin up state is achieved with a sequence of microwave and/or RF pulses. Starting from the spin down electron state and the nuclear spin in an unknown state, the microwave pulse flips the electron to the spin up state only if the nucleus is in the spin up state as well. In that case, the subsequent RF pulse is off-resonance with the nucleus and leaves the nucleus in the target spin up state. Conversely, if the nucleus is initially spin down, the microwave pulse is off-resonance with the electron, and the subsequent RF pulse flips the nucleus from spin down to spin up state. With one more electron readout/initialization step, the system is unconditionally prepared such that the electron is in the spin down state and the nucleus is in the spin up state.
In the transfer stage, to transfer the quantum information from the electron to the nucleus, a first RF pulse conditionally shifts the spin down component of the electron spin state onto the spin down component of the nucleus, creating a double quantum coherence between the electron and nuclear spin ups or spin downs if the electron possessed a spin up component. The subsequent microwave pulse conditional on the nuclear spin up subspace, translates the spin up electron and nuclear component onto the electron spin down nucleus spin up state, leaving the nuclear spin with all the quantum information and the electron spin in the spin down eigenstate. The ‘transfer’ operation of the memory protocol is now complete and the quantum information may be left in the nuclear state for the desired wait time before reversing the order of the pulses (‘recovery’ sequence) to bring the quantum information back to the electron state.
As noted previously, in addition to isolating electrons, electrical signals can be applied to the electrodes 106/206 to perform selected operations on the processing elements and/or readout a quantum state of the processing element.
In certain embodiments, the nuclear spin state readout may be performed by performing electron spin resonance (ESR) on the electron of the quantum dot using, e.g., microwave pulses. These microwave pulses in one example may be delivered by an on-chip broadband planar transmission line (not shown).
In response to the microwave pulses, the system will exhibit two possible ESR frequencies depending on the state of the nuclear spin—i.e., for a nuclear spin up, the processing element will exhibit an ESR frequency of ESR plus half the hyperfine interaction (γeB+A/2) and for a nuclear spin down, the processing element will exhibit an ESR frequency of ESR minus half the hyperfine interaction (γeB−A/2). Depending on the detected ESR, the state of the nuclear spin can therefore be determined.
Further,
As noted previously, nuclei, therefore nuclear spins, are generally immobile whereas electrons, and therefore electron spins are very mobile and can be shuttled along the quantum device 200 to transport information to other circuit nodes.
Typically, in donor-electron spin systems, when the electron is unloaded from the donor, the nuclear coherence of the donor is destroyed. This is because the ionisation rate of unloading an electron is slow relative to the typical hyperfine coupling between the donor and electron, which results in a random phase shift of the nuclear spin state depending on the random time the electron was ionized.
However, because in the presently disclosed system it is possible for the electron to tunnel between two neighbouring quantum dots via spin shuttling or exchange mediated coupling, the tunnel rate for unloading an electron from the quantum dot is much higher than the ionisation rate of electron-donor systems. In the presently disclosed system, it is therefore possible to retain the nuclear spin coherence when unloading the electron.
This is important when the electron is used as a mediator of entanglement between nuclear spins in different quantum dots. To do this, the electron 302 can first be entangled with a nuclear spin 304 in a first quantum dot 306. The electron 302 is then transferred to a second quantum dot 308, where it can be entangled with a second nuclear spin 310. If the state of the electron spin of the second quantum dot is measured at this time, it will be determined that the two nuclear spins 304 and 310 are projected onto an entangled state, even though the two nuclear spins did not interact directly. In this manner, information can be shuttled across the quantum device 200 from one quantum dot to the next.
It will be appreciated that the architectures described in the embodiments above utilize an electron in the quantum dot that is entangled with the nucleus of a 29Si atom. In other embodiments, the quantum dot can be configured to confine holes instead of electrons and the holes can be entangled with the 29Si atoms.
In some embodiments, instead of using 29Si atoms and entangling their nuclear spins with nearby quantum dots, any other (stable) group IV atoms (e.g., Carbon, Germanium, Tin or Lead) that have a nuclear spin may be utilized. In one example, the carbon-13 (13C) isotope may be utilized. In another example, a Germanium-73 (73Ge) isotope may be utilized. Tin has a number of stable isotopes with nuclear spins, such as Tin-115 (115Sn), tin-117 (117Sn), Tin-119 (119Sn) and any of these isotopes of tin may be utilized instead of 29Si atoms.
The atoms of one or more of these elements can be implanted with low energy to place them at a shallow depth and therefore within the wavefunction of a nearby quantum dot (similar to the way described for the 29Si atoms above). The nuclear spin of the atom could then be entangled with the electron spin of the nearby quantum dot(s) via hyperfine coupling.
The carbon and tin isotopes have a spin of ½ (similar to 29Si) and therefore function in the same manner as the 29Si atoms. The Germanium isotope has a spin of 9/2. This means that the germanium nuclear spin has 10 states that can be used for quantum computation as opposed to the 2 states available for silicon, carbon, and tin atoms.
One example method for artificially introducing one or more group IV atoms in a silicon substrate is described in U.S. Pat. No. 7,002,166, titled “method and system for single ion implantation”, which is incorporated herein in its entirety. Although implanting techniques are described in that patent with respect to Phosphorus atoms, those techniques can be utilized (with minor modifications) to precisely implant carbon, silicon, germanium or tin atoms in a silicon substrate.
The term “comprising” (and its grammatical variations) as used herein are used in the inclusive sense of “having” or “including” and not in the sense of “consisting only of”.
It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2019900297 | Jan 2019 | AU | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/AU2020/050066 | 1/31/2020 | WO | 00 |
