FIELD OF THE INVENTION

The invention is generally related to the tech nology of quantum computing hardware. In particular the invention is related to an advantageous structural so lution of a quantum computing circuit.

BACKGROUND OF THE INVENTION

The hardware used for quantum computing is based on superconducting chips. This term is generally used to mean a device in which a number of microscopic- scale circuit elements, at least some of which are made of superconductive material, have been produced on a substrate using photolithography, micromachining, and/or other suitable methods. A quantum processor is a superconducting chip that comprises a selection of cir cuit elements, including one or more qubits, as well as their interconnections in an arrangement that enables using the one or more qubits for quantum computing op erations.

Examples of circuit elements that may be needed in a quantum processor include but are not limited to qubits, resonators, couplers, qubit reset circuitry, semiconducting quantum dots, single-electron transis tors, amplifiers, and others. Of these, the qubit reset circuitry may comprise for example quantum circuit re frigerators, known as QCRs for short. Whatever the exact composition of the quantum processor, it has been found that problems may arise from the various materials and processing steps that are required during its manufac turing process, as well as from unwanted interactions between the various circuit elements in the completed quantum processor. As an example, a manufacturer may have a pro cess that has been optimized to produce qubits of ex tremely high quality. It may turn out, however, that the process is not suitable as such for manufacturing more complicated superconducting chips such as quantum pro cessors, because some of the materials and/or process steps needed for the other circuit elements are incom patible with those needed for the qubits. The result is often a compromise in which the materials and/or process steps are reasonably suitable for all circuit elements, even if they may not be exactly optimal for any indi vidual circuit element.

As another example, while the qubits and other circuit elements of a quantum processor may have a va riety of desired interactions that are essential for quantum computing, they may also interact in undesired ways that cause dissipation and shorten the coherence time of the qubit states. Such a phenomenon is a source of quantum information loss.

There exists a clear need for solutions in cir cuit design and manufacturing methods for quantum com puting circuits that would enable better optimization of the materials and/or process steps and/or circuit operation.

SUMMARY

It is an objective to present a quantum compu ting circuit and a method for its manufacturing that enable optimization of the materials and/or process steps involved. Another objective is to enable reaching optimal performance of all or at least a majority of circuit elements in a quantum computing circuit.

The objectives of the invention are achieved by using a flip-chip approach in which circuit elements, the manufacturing and/or operation of which together could involve incompatible aspects, are manufactured on separate chips, which are subsequently attached together in a sandwiched configuration.

According to a first aspect there is provided a quantum computing circuit, which comprises a first chip, with at least one qubit thereon, and a second chip, with at least other quantum circuit element than qubit thereon. Said first chip and said second chip are stacked together in a flip-chip configuration and at tached to each other with bump bonding that comprises bonding bumps.

According to an embodiment, said first chip is made of a first set of constituent materials and said second chip is made of a second set of constituent ma terials. In such a case said first and second sets con sist of at least partly different constituent materials. This involves the advantage that in fabricating the qubits the use of such materials can be avoided that might cause disadvantageous contamination of the qubit (s).

According to an embodiment, said second set of constituent materials comprises at least one material that is not present in said first set of constituent materials and is one of: aluminum oxide, copper, palla dium, other non-superconductive metal. This involves the advantage that in particular, contamination by this kind of materials can be avoided.

According to an embodiment, said first chip is a chip manufactured in a first manufacturing process that consists of a first sequence of manufacturing steps, and said second chip is a chip manufactured in a second manufacturing process that consists of a second sequence of manufacturing steps. Said first and second sequences may be at least partly different sequences of manufacturing steps. This involves the advantage that one may avoid subjecting the qubit(s) to manufacturing steps that are not needed for manufacturing the qubit(s) and that could cause harmful effects to the qubit(s). According to an embodiment, at least some of said bonding bumps are galvanically conductive and con stitute galvanically conductive contacts between said first and second chips. This involves the advantage that signal lines can be routed to and from, and/or ground planes and other conductive constructions may be con nected together, between the first and second chips.

According to an embodiment, one of said first and second chips is a larger chip and the other of said first and second chips is a smaller chip that covers only a part of said larger chip in said flip-chip con figuration. This involves the advantage that the exposed area of the larger chip can be used to make connections to and from the quantum computing circuit.

According to an embodiment, the larger chip comprises at least a first contact pad on that part of its surface facing the smaller chip that is not covered by said smaller chip. The larger chip may then comprise a first connection connecting said first contact pad and a first galvanically conductive bonding bump. The smaller chip may comprise a second connection connecting said first galvanically conductive bonding bump and a first quantum circuit element on the smaller chip. Said first contact pad may constitute a signal connection to said first quantum circuit element. This involves the advantage that signal connections to components on the smaller chips may be made through contact pads on the larger chip.

According to an embodiment, the smaller chip comprises a second contact pad on its surface facing away from the larger chip, and a third connection through a first conductive via, connecting said second contact pad to a second quantum circuit element on the surface of the smaller chip facing the larger chip. This involves the advantage that signal connections to com ponents on the smaller chip may be made through contact pads on the exposed surface of the smaller chip. According to an embodiment, the larger chip comprises a second conductive via, connecting a third quantum circuit element on that part of the surface of the larger chip facing the smaller chip that is covered by said smaller chip to a fourth connection that is at least partly located on the surface of the larger chip facing away from the smaller chip. This involves the advantage that signal connections can be made in a very effective way also to components on any one of the chips located within the area where the chips overlap.

According to an embodiment, the quantum compu ting circuit comprises a non-galvanic connection for conveying signals between said first and second chips, said non-galvanic connection comprising matching non- galvanic connector structures on the surfaces of said first and second chips facing each other. This involves the advantage offering a large degree of control over the ways in which the components on the two chips are coupled to each other.

According to an embodiment, said matching non- galvanic connector structures comprise mutually aligned conductive areas on the surfaces of said first and sec ond chips facing each other for making a capacitive connection. This involves the advantage that the prop erties of the corresponding connection, including e.g. inherent filtering capability, can be tuned by selecting the capacitance of the capacitive connection appropri ately.

According to an embodiment, said matching non- galvanic connector structures comprise mutually aligned inductive elements for making a magnetic connection. This involves the advantage that the properties of the corresponding connection, including e.g. inherent fil tering capability, can be tuned by selecting the in ductance of the magnetic connection appropriately.

According to an embodiment, said second chip comprises a quantum circuit refrigerator. Said quantum computing circuit may comprise a controllable connection between said quantum circuit refrigerator and at least one qubit on said first chip for allowing said quantum circuit refrigerator to be controllably used to reset the state of said at least one qubit. This involves the advantage that the manufacturing of the qubit(s) and the quantum circuit refrigerator(s) can be kept separate from each other, so that both can be optimized without causing disadvantageous effects to the other.

According to an embodiment said second chip comprises at least one filter that comprises at least one of: a non-superconductive metal, a lossy dielectric. This involves the advantage that the method steps and materials that are needed to fabricate filters can be kept from compromising the quality of the qubit(s).

According to an embodiment, the separating dis tance between said first and second chips is between 1 and 100 micrometers. This involves the advantage that the distance can be made to play a desired role in any non-galvanic connections that may be made between the two chips.

According to a second aspect there is provided a method for producing a quantum computing circuit. The method comprises manufacturing a first chip, and pro ducing at least one qubit on said first chip; manufac turing a second chip, and producing at least one quantum circuit element other than qubit on said second chip; and bump bonding said first and second chips together into a stacked configuration where bonding bumps attach the first and second chips to each other.

According to an embodiment, the method com prises using a first set of materials in manufacturing said first chip and using a second set of materials in manufacturing said second chip, so that said first and second sets consist of at least partly different mate rials. This involves the advantage that in fabricating the qubits the use of such materials can be avoided that might cause disadvantageous contamination of the qubit (s).

According to an embodiment, the method com prises using a first manufacturing process to manufac ture said first chip, said first manufacturing process consisting of a first sequence of manufacturing steps; and using a second manufacturing process to manufacture said second chip, said second manufacturing process con sisting of a second sequence of manufacturing steps; so that said first and second sequences are at least partly different sequences of manufacturing steps. This in volves the advantage that one may avoid subjecting the qubit (s) to manufacturing steps that are not needed for manufacturing the qubit(s) and that could cause harmful effects to the qubit(s).

According to an embodiment, the method com prises making said bump bonding attach the first and second chips to each other at a separating distance selected for optimized non-galvanic signal coupling be tween circuit elements on said first and second chips. This involves the advantage that the distance can be made to play a desired role in any non-galvanic connec tions that may be made between the two chips.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are included to provide a further understanding of the invention and constitute a part of this specification, illustrate em bodiments of the invention and together with the de scription help to explain the principles of the inven tion. In the drawings:

Figure 1 illustrates a part of a known quantum processor,

Figure 2 illustrates an exploded view of the part shown in fig. 1, Figure 3 illustrates a principle of a quantum computing circuit that includes a sandwiched configura tion of two chips,

Figure 4 illustrates an example of applying the principle of fig. 3,

Figure 5 illustrates another example of apply ing the principle of fig. 3,

Figure 6 illustrates another example of apply ing the principle of fig. 3,

Figure 7 illustrates couplings between circuit elements in a quantum computing circuit,

Figure 8 illustrates an exploded view of an example of applying the principle of fig. 3, and

Figure 9 illustrates a method according to an embodiment of the invention.

DETAILED DESCRIPTION

Fig. 1 illustrates schematically a part of a quantum computing circuit seen from above. The quantum computing circuit in question may be for example a quan tum processor. Fig. 2 is an exploded view that shows how the patterns seen in fig. 1 may be constructed on the surface of a dielectric substrate 201. Both fig. 1 and fig. 2 are simplified for the purpose of graphical clar ity, but the following explanation is applicable to also more complete corresponding structures. Cross-hatched portions in fig. 1 illustrate areas where the surface of the substrate is visible between the patterns made of substances of desired degree of conductivity and/or superconductivity .

The X-formed or plus-sign-formed part 101 is a qubit of the transmon type. The simplification made here involves e.g. not showing a nonlinear inductance compo nent of the qubit that could appear e.g. as a Josephson junction and/or SQUID (Superconductive Quantum Inter ference Device) at the end of one of the branches shown here. The fork-like structure 102 around one of the branches is a capacitive coupling element, from which a transmission line 103 may lead to some other part of the quantum computing circuit. The intertwined fingers 104 constitute another capacitive coupling element, from which there is a connection to a qubit reset circuit 105, which here is a quantum circuit refrigerator, also known by the acronym QCR. As such, the QCR could also be coupled to the capacitive coupling element 102 in stead of being coupled directly to the qubit 101.

Fig. 2 shows how most of the surface of the substrate is covered by the superconductive ground plane 202, in which the small, square openings serve to pre vent the occurrence of unwanted eddy currents. The pat terns shown as 203, 204, and 205 are also made of a superconductive material, which as a characterization means a material that becomes superconductive in the low temperatures at which the quantum computing circuit is to be used.

The part shown as 206 comprises the SINIS junc tion needed in the QCR, as well as the contact pad through which the QCR makes a connection to the ground plane. Part 207 is a part of a microwave filter, and part 208 forms a transmission line through which a con trol signal may be brought to the QCR. A dielectric layer 209 is needed between the ground plane 202 and the conductive part 207 to form a capacitor as a part of the microwave filter.

Processes and designs for manufacturing high- quality qubits are known, and they have typically been optimized so that the produced qubits have a long co herence time and other advantageous characteristics. Manufacturing a structure like that of figs. 1 and 2, in particular the parts of the QCR, may require a dif ferent kind of process. The difference may appear in the form of different process steps; different order of pro cess steps; different materials; different process pa- rameters such as temperatures, pressures and/or dura tions; and the like. For example, manufacturing the SINIS junction in the QCR involves using one or more materials, i.e. normal metal(s) and insulator(s), that would not be used if only the qubit(s) would be manu factured. Similarly the insulating layer 209 may com prise a material that would not appear in the manufac turing process of just one or more qubits. All these differences in materials and/or process steps may lead to the disadvantageous consequences discussed above in the background section.

Fig. 3 illustrates schematically a quantum com puting circuit in which said disadvantageous conse quences can be at least partly avoided. The quantum computing circuit of fig. 3 comprises a first chip 301 and a second chip 302. Both chips have one or more quantum circuit elements built on them. The quantum cir cuit element(s) on the first chip 301 are schematically shown as 303, and the quantum circuit elements on the second chip 302 are schematically shown as 304. Based on their appearance in fig. 3 the first chip 301 may also be called the bottom chip and the second chip 302 may be called the top chip.

One of the first and second chips is the so- called qubit chip. Without losing generality we may as sume here that the first chip 301 has at least one qubit included in its quantum circuit elements 303. The second chip 302 has at least one quantum circuit element other than qubit included in its quantum circuit elements 304. In an advantageous embodiment the second chip 302 has no qubits included in its quantum circuit elements 304, so that all qubits of the quantum computing entity con sisting of the first and second chips 301 and 302 are included in the quantum circuit elements 303 of the first chip 301. With reference to the description of figs. 1 and 2 we may assume that the manufacturing pro cess used to manufacture the first chip 301 differs from that used to manufacture the second chip 302, and this difference is a direct consequence of the fact that there is at least one quantum circuit element other than qubit included in the quantum circuit elements 304 of the second chip 302.

In this approach, the qubit chip will not ex perience the manufacturing process of the QCR and/or other non-qubit circuit elements located on the other chip, and will not have direct contact to the materials involved in the fabrication of the QCR and/or other non qubit circuit elements. In that sense the qubit is 'pristine' and can be manufactured with the standard process which is known to produce high quality qubits and long coherence time. In addition, this can minimize the contact of dissipative components to the qubit chip.

An example of a manufacturing step that may be needed to fabricate another quantum circuit element but that is disadvantageous to any qubit located on the same chip as such another quantum circuit element is baking, which means the application of high temperature in order to e.g. cure a layer of resist. Resists are used in photolithographic manufacturing methods to define, how the various patterns will be formed on the surface of the chip. For example, the junctions of qubits and QCRs both require baking during their fabrication, but since the materials involved are different, both require at least one baking step of their own. On the other hand, any subjecting to high temperature is known to acceler ate the disadvantageous aging of any such junction. Thus, if a qubit and a QCR are fabricated on the same chip, whichever of their respective junctions is made first, it will experience disadvantageous accelerated aging during the baking step(s) needed to subsequently fabricate the other junction.

Another example of a manufacturing step that may be needed to fabricate another quantum circuit el ement but that is disadvantageous to any qubit located on the same chip is etching. For example, when a micro- wave filter is fabricated there is needed a layer of dielectric material, like aluminum oxide for example. After the deposition of an aluminum oxide layer, etching is used to remove the aluminum oxide from those parts of the chip surface where it is not needed. Etching may create surface roughness on the qubit region, which may impede the optimal performance that the qubit might oth erwise achieve.

Additionally, any step of a manufacturing method that involves using a material that would not be needed to manufacture just the qubit(s) may be disad vantageous to any qubit located on the same chip. Any such step may cause contamination, which in this frame work means unwanted appearance of material residues at or close to the qubit region in the completed chip. Contaminating materials may include materials that are actually needed at other parts of the quantum computing circuit, such as non-superconductive metals and dielec tric substances, and/or materials that are only needed during manufacturing such as resists.

The first chip 301 and the second chip 302 are stacked together in a flip-chip configuration and at tached to each other with bump bonding 305 that com prises bonding bumps. A flip-chip configuration of two chips is sometimes referred to as (one form of) 3D in tegration of circuits. Producing the flip-chip config uration involves using special machine known as a flip- chip bonder to place bonding bumps at selected locations on the surface of at least one of the chips and to then press the chips together, applying a predetermined tem perature and pressing force. This causes partial defor mation of the bonding bumps and attaches the chips to each other. If some or all of the bonding bumps are made of an electrically conductive (or superconductive) ma terial, and if they were placed at locations where mu tually aligned conductive (or superconductive) patterns exist on the surfaces of both chips, they can be used to make electric connections of desired kind between the chips.

The difference between the manufacturing pro cesses of the two chips 301 and 302 may be for example such that the first chip 301 is made of a first set of constituent materials, the second chip 302 is made of a second set of constituent materials, and the first and second sets consist of at least partly different con stituent materials. Assuming that the first chip 301 comprises the qubit(s) and the second chip 302 comprises at least some other quantum circuit elements than qubits, the second set of constituent materials may com prise at least one material that is not present in said first set of constituent materials, like aluminum oxide, copper, palladium, and/or other non-superconductive metal. In general, said at least one material that is not present in said first set of constituent materials is a material that is fundamentally not compatible with the aim of optimizing the manufacturing process and later operative use of qubits.

Additionally or alternatively, the difference between the manufacturing processes of the two chips 301 and 302 may be in the steps of their manufacturing pro cesses. The first chip 301 may be a chip manufactured in a first manufacturing process that consists of a first sequence of manufacturing steps, and the second chip 302 may be a chip manufactured in a second manu facturing process that consists of a second sequence of manufacturing steps. These first and second sequences are then at least partly different sequences of manu facturing steps. In particular, the second sequence may involve one or more manufacturing steps that by their nature would be disadvantageous in producing qubits of highest possible quality. Additionally or alterna tively, the second sequence may lack one or more manu facturing steps that are essential in producing qubits of highest possible quality. Additionally or alterna tively, the second sequence may involve one or more manufacturing steps in which the selected value of a process parameter, such as a temperature, pressure, or duration for example, would be disadvantageous in pro ducing qubits of highest possible quality.

Fig. 4 illustrates a quantum computing circuit according to one embodiment that follows the principle explained above with reference to fig. 3. In figs. 4, 5, and 6, one of the chips (here the bottom chip 301) is a larger chip and the other (here the top chip 302) is a smaller chip that covers only a part of the larger chip in the flip-chip configuration. In fig. 4 the larger chip comprises at least one electrically conduc tive (or superconductive) contact pad 401 on that part of its surface facing the smaller chip that is not cov ered by the smaller chip. Additionally the larger chip comprises a connection, called here the first connection for unambiguous reference, that connects the contact pad 401 and a (first) galvanically conductive bonding bump 403. Said first connection may comprise one or more conductive (or superconductive) patterns 402 on the sur face of the larger chip.

Further in fig. 4, the smaller chip comprises a connection, called here the second connection, that connects the first galvanically conductive bonding bump 403 and a first quantum circuit element 405 on the smaller chip. Said second connection may comprise one or more conductive (or superconductive) patterns 404 on the surface of the smaller chip. This way at least one conductive (or superconductive) contact pad 401 on the larger chip constitutes a signal connection to the first quantum circuit element 405. The use of the preposition "to" does not limit the direction in which signals flow in said signal connection; it could also be called a connection "from" the first quantum circuit element 405. The quantum computing circuit of fig. 4 may involve a large variety of other connections and circuit elements. As examples, a further contact pad 406, fur ther conductive (or superconductive) patterns 407, and a further quantum circuit element 408 are shown on the surface of the larger chip.

Implementations such as that in fig. 4 allow routing signal lines, such as control and bias lines for example, of the top chip 302 via the bump bonds to the bottom chip 301, and making them accessible through con tact pads so that connections to and from them can be made by other components. Also the signal lines of the bottom chip 301 may be routed further out so that they are not covered by the smaller chip but become similarly accessible. This allows a straightforward way of provid ing access for electrical and microwave signals to and from both chips after they have been bonded together.

Fig. 5 illustrates a quantum computing circuit according to another embodiment that follows the prin ciple explained above with reference to fig. 3. Here the smaller chip (the top chip 302) comprises a second con tact pad 501 on its surface facing away from the larger chip (the bottom chip 301). The smaller chip comprises a third connection through a first conductive via 502, connecting the second contact pad 501 to a quantum cir cuit element 503 on the surface of the smaller chip facing the larger chip. Using one or more conductive vias (also known as TSVs or Through Silicon Vias) in the smaller chips allows using the upper surface (the sur face facing away from the larger chip) of the smaller chip for useful purposes, such as for contact pads, connections, and even quantum circuit elements. Another exemplary feature shown in fig. 5 is the possibility to use also non-conductive bonding bumps 504, at least at such locations where there is no need to make conductive connections between the two chips in the flip-chip con- figuration. It should be noted that despite the refer ence to silicon in the term "through silicon via", the substrate of the chips may be other than silicon, such as sapphire for example.

The embodiment of fig. 5 may allow routing even more signal lines on any or both chips to the central region, thus allowing more components to be integrated at the central region. It should be noted that while fig. 5 does not show any quantum circuit elements or connections on the surface of the bottom chip 301 within the area covered by the top chip 302, such quantum cir cuit elements and connections could very well exist there. The approaches shown in figs. 4 and 5 may also be combined, for example so that signal lines to some components on the top chip 302 go through contact pads on the bottom chip 301 and through conductive pads, while some other signal lines take routes through con ductive vias in the top chip 302. It is also possible to route some signal lines to quantum circuit elements on the bottom chip 301 through contact pads on the top chip 302, through conductive vias in the top chip 302, and through conductive bonding bumps.

Fig. 6 illustrates a quantum computing circuit according to another embodiment that follows the prin ciple explained above with reference to fig. 3. Here the larger chip (the bottom chip 301) comprises a second conductive via 601, connecting a third quantum circuit element 602 on that part of the surface of the larger chip facing the smaller chip that is covered by said smaller chip to a fourth connection that is at least partly located on the surface of the larger chip facing away from the smaller chip. In the embodiment of fig. 6 said fourth connection continues through a conductive (or superconductive) pattern 603 on the lower side of the bottom chip 301 to a further conductive via 604, and therethrough to a further contact pad 605 on the top side of the bottom chip 301. As there are even more versatile ways of rout ing the connections and placing the circuit element in fig. 6 than in the two preceding drawings, the embodi ment of fig. 6 may allow routing even more signal lines to the central region and/or integrating even more com ponents to the central region.

In fig. 6 the top chip 302 is of the kind shown earlier in fig. 5, i.e. one with conductive vias therethrough and with contact pads on its top surface offering access for signals to and from quantum circuit elements on its bottom side. The top chip 302 could also be of the kind shown earlier in fig. 4. The approaches shown in figs. 4, 5, and 6 can be combined also in many other ways. Additionally there could be contact pads on the lower side of the bottom chip 301.

Fig. 6 shows also a further advantageous fea ture, according to which there may be one or more non- galvanic connections for conveying signals between the first and second chips. Such one or more non-galvanic connections may comprise matching non-galvanic con nector structures on the surfaces of the first and sec ond chips facing each other. As an example, the quantum circuit element 602 shown in fig. 6 on the bottom chip 301 may have a capacitive or inductive coupling to an opposite quantum circuit element 606 on the top chip 302. In general, said matching non-galvanic connector structures may comprise mutually aligned conductive ar eas on the surfaces of said first and second chips facing each other for making a capacitive connection, and/or mutually aligned inductive elements for making a mag netic connection.

Fig. 7 is a simplified circuit diagram of an example of a quantum computing circuit in which the top chip 302 comprises two qubits 701 and 702 and the bottom chip 301 comprises two QCRs 703 and 704 that can be used to reset the two qubits 701 and 702 respectively. Con cerning the routing of signal lines the embodiment shown in fig. 7 follows the general approach taken in fig. 4 above, in which also those signal lines that eventually go to quantum circuit elements in the top chip 302 come in through the bottom chip 301. One such signal line is shown, namely the signal line 705 that is used to drive the qubits 701 and 702. The two other signal lines 706 and 707 are used to control the operation of the QCRs 703 and 704 respectively. Additional there is the ground connection 708. Any suitable method can be used to make connections from the signal lines 705, 706, and 707, and ground connection 708 to other parts of the quantum computing device, including but not being limited to wire bonding.

The point 709 in the circuit diagram that rep resents connecting the ground connection 708 between the two chips 301 and 302 may in practice take the form of a plurality of conductive (or superconductive) bonding bumps at a number of locations distributed around the ground planes on the surfaces of the chips facing each other. In general, it is advantageous to ensure that the grounding connection between the two chips is as effec tive as possible, for which purpose the use of a plu rality of conductive (or superconductive) bonding bumps that interconnect a large number of ground plane points on the two chips is often recommendable.

In the embodiment of fig. 7 the couplings of signal lines between the two chips are capacitive, as is shown by the capacitor symbols at those locations where a signal line passes from one chip to another. These couplings may go through e.g. mutually aligned conductive areas on the surfaces of the two chips 301 and 302 facing each other. The capacitance of such ca pacitive couplings can be controlled by dimensioning the mutually aligned conductive areas. Additionally or al ternatively, all non-galvanic signal couplings between circuit elements on the surfaces of the two chips 301 and 302 may be deliberately affected by controlling the flip-chip bonder so that during the bonding process it makes the final separating distance between the first and second chips assume a particular value selected for optimized non-galvanic signal coupling.

Many kinds of connectivity can be achieved be tween the two chips by using some or all of the ap proaches described above. Such connectivity may include but is not limited to

- drive control lines and coupling elements (capacitors) for qubit driving, i.e. the RF or microwave lines for driving qubit transitions,

- readout control lines for passing the readout pulses across the qubit system, readout elements (res onators), and coupling elements for coupling the readout resonator to the readout control lines,

- connectivity to couple the qubits with each other with dedicated coupling elements,

- control lines and elements to control the qubit coupling,

- connectivity and coupling elements from reset circuitry such as QCRs to the qubits,

- reset control, such as QCR control lines for biasing and providing the reset pulses, and

- connectivity ensuring proper grounding within the system.

In preferred embodiments the parts of the cir cuit including quantum coherent operation such as the qubits themselves, and the couplers including any con nectivity between the qubit and couplers, is included on the qubit chip. As fabrication layers, the non-qubit chip can include normal metals, and possibly lossy di electrics used for multi-layer structures for routing or potentially as filters that are on the non-qubit chip. In preferred embodiments the non-qubit chip in cludes at least the reset circuitry, such as QCR devices themselves and the QCR control lines. In an embodiment all connections to the qubit chip are performed by ca pacitive coupling as in fig. 7 (which shows just some connectivity and components as examples) apart from the grounding connection which is advantageously performed galvanically (by single or preferably multiple bump bonds forming the connection) for proper RF design. In principle the capacitive connections can be at any part replaced by galvanic connections. A capacitive connec tion can also be implemented by placing the coupling capacitor on one of the chips entirely, followed by a galvanic connection from one of the capacitor plates to the other chip.

In some embodiments, in addition to or in place of galvanic and capacitive couplings one can use mag netic coupling by mutual inductance.

Different embodiments of methods according to the invention may comprise using different methodologies in fabricating the bump bonds such as evaporation or electroplating the bump material. The bump height de fining the inter-chip distance can be tuned based on the specifications such as the implementation of the reac tive (capacitive or mutual inductive) inter-chip cou pling strengths, and spurious (unintentional) couplings between different elements. Furthermore, the distance can be tuned based on the requirement of electric fields coupling from the qubit chip to the non-qubit chip to avoid excessive losses due to fringing fields from the qubit chip into the potentially lossy materials of the non-qubit chip. Typical inter-chip distances can be in the range of 1 - 100 ym.

Fig. 8 illustrates a part of a quantum compu ting circuit according to an embodiment of the invention in an exploded view. There is a substrate 801 of the chip that here appears as the larger chip or bottom chip, and which is also the qubit chip because the su perconductive patterns that constitute a qubit are seen as 802. A ground plane 803 covers a majority of the surface of the substrate 801, and there may be other conductive or superconductive patterns such as trans mission lines 804. The substrate 805 of the other chip, which here may be called the top chip, the smaller chip, or the non-qubit chip, is seen at the top of the exploded view, with its respective ground plane 806 produced as a layer on that surface of the substrate 805 that faces the bottom chip.

Examples of patterns that are produced on top of the ground-plane-covered surface of the top chip sub strate 805 are the pair of capacitive coupling elements 807, a short stub of a transmission line 808, a dielec tric patch 809, a conductive (or superconductive) part 810 of a microwave filter, and the tunneling junction part 811 of a QCR with its associated contacts. In a completed configuration the QCR would look essentially like the one seen in figs. 1 and 2 earlier, with the microwave filter 810 connected between the transmission line 808 and the tunneling junction part 811 and iso lated from the ground plane 806 of the top chip by the dielectric patch 809.

Fig. 8 shows some examples of places where bump bonding may be used between the two chips. In the drawing the locations of the bonding bumps are schematically shown with solid black circles on both sides for graph ical clarity. At locations 812, 813, 814, and 815, bond ing bumps are used to make galvanic connections to and from the QCR located on the top chip. At the other locations, bonding bumps are used to make galvanic con nections between the ground planes of the two chips.

Fig. 9 illustrates schematically a method for producing a quantum computing circuit. Steps 901 com prise manufacturing a first chip and include, in one or more of the steps thereof, producing at least one qubit on said first chip. Steps 902 comprise manufacturing a second chip and producing at least one quantum circuit element other than qubit on said second chip. Preferably steps 902 comprise manufacturing said second chip with out producing any qubits on said second chip. Step 903 comprises bump bonding said first and second chips to gether into a stacked configuration where bonding bumps attach the first and second chips to each other.

There is a difference between steps 901 and 902 in the method. Steps 901 may comprise e.g. using a first set of constituent materials in manufacturing said first chip, and steps 901 may comprise using a second set of constituent materials in manufacturing said second chip, so that said first and second sets consist of at least partly different constituent materials. In such an em bodiment, at least one of the materials used in at least one of the steps 902 is one that is incompatible with an optimized method of manufacturing high-quality qubits with long coherence time. Additionally or alternatively, the steps 901 may consist of a first sequence of manu facturing steps, and the steps 902 may consist of a second sequence of manufacturing steps, so that said first and second sequences are at least partly different sequences of manufacturing steps. In such an embodiment, the second sequence may comprise a manufacturing step that is incompatible with an optimized method of manu facturing high-quality qubits with long coherence time, or it may lack one or more steps that are essential in an optimized method of manufacturing high-quality qubits with long coherence time.

As shown by step 904, the method may comprise making said bump bonding attach the first and second chips to each other at a separating distance selected for optimized non-galvanic signal coupling between cir cuit elements on said first and second chips.

It is obvious to a person skilled in the art that with the advancement of technology, the basic idea of the invention may be implemented in various ways. The invention and its embodiments are thus not limited to the examples described above, instead they may vary within the scope of the claims.