EP4626828A2 - Methods of forming bonded diamond membrane heterostructures - Google Patents

Abstract

A method of forming a bonded diamond membrane heterostructure may comprise: (a) subjecting a surface of a target substrate to plasma ashing to provide a plasma treated target substrate having a plasma treated surface; and (b) contacting the plasma treated surface of the plasma treated target substrate with a surface of a diamond membrane to form a bonded diamond membrane heterostructure comprising the target substrate bound via covalent bonds to the diamond membrane at a bonding interface formed between the plasma treated surface of the plasma treated target substrate and the surface of the diamond membrane. The bonded diamond membrane heterostructures formed using the method are also provided.

Description

METHODS OF FORMING BONDED DIAMOND MEMBRANE HETEROSTRUCTURES

CROSS-REFERENCE TO REEATED APPEICATIONS

[0001] The present application claims priority to U.S. provisional patent application number 63/428,236 that was filed November 28, 2022, the entire contents of which are incorporated herein by reference.

REFERENCE TO GOVERNMENT RIGHTS

[0002] This invention was made with government support under DE FOA-0002253 awarded by the Department of Energy. The government has certain rights in the invention.

BACKGROUND

[0003] Diamond is a leading material platform in quantum information science with several landmark demonstrations in quantum sensing and quantum communication. These demonstrations rely on devices carved directly into bulk, monolithic diamond at high cost, low throughput, and low yield, limiting scalability and device functionality. For instance, millimeter scale electronic-grade diamond has limited availability and costs thousands of dollars. Additionally, device functionality is highly limited by the diamond material, which lacks any significant nonlinear optical response and is challenging to fabricate at the nanoscale without degradation of the optical and spin qubit properties. While color centers in diamond can be used to probe biological and chemical phenomena, it is challenging to integrate the diamond substrates with non-diamond materials used in the quantum applications.

SUMMARY

[0004] Provided are methods of forming bonded diamond membrane heterostructures.

The methods make use of diamond membranes (e.g., ultrathin, single-crystal (100) diamond) and non-diamond target substrates (e.g., lithium niobate) and achieve strongly bound (e.g., via covalent bonds) disparate materials without using an intervening material to join the diamond membrane and the target substrate. This is by contrast to existing methods involving non-covalent interactions (e.g., Van der Waals forces) or intermediate bonding layers such as epoxy and hydrogen silsesquioxane. In addition, the bonding interfaces of the present bonded diamond membrane heterostructures are highly crystalline and extremely thm, by contrast to the amorphous and/or thick bonding interfaces produced using existing methods. The present methods allow for integration of diamond membranes into a variety of devices comprising non-diamond materials such as those used in quantum sensing and quantum communication applications.

[0005] An embodiment 1 is a method of forming a bonded diamond membrane heterostructure, the method comprising: (a) subjecting a surface of a target substrate to plasma ashing to provide a plasma treated target substrate having a plasma treated surface; and (b) contacting the plasma treated surface of the plasma treated target substrate with a surface of a diamond membrane to form a bonded diamond membrane heterostructure comprising the target substrate bound via covalent bonds to the diamond membrane at a bonding interface formed between the plasma treated surface of the plasma treated target substrate and the surface of the diamond membrane.

[0006] An embodiment 2 is the method of embodiment 1, further comprising subjecting the surface of the diamond membrane to plasma ashing to provide a plasma treated surface of the diamond membrane prior to step (b).

[0007] An embodiment 3 is the method of any of embodiments 1-2, wherein the plasma ashing is carried out using an O2 plasma.

[0008] An embodiment 4 is the method of any of embodiments 1-3, wherein the plasma ashing provides the plasma treated surface of the plasma treated target substrate with oxygen termination.

[0009] An embodiment 5 is the method of embodiment 1, wherein the surface of the diamond membrane is an untreated surface.

[0010] An embodiment 6 is the method of any of embodiments 1-5, wherein the diamond membrane is single-crystalline and the surface of the diamond membrane is (100).

[0011] An embodiment 7 is the method of any of embodiments 1-6, wherein the diamond membrane has a thickness of no more than 500 nm.

[0012] An embodiment 8 is the method of any of embodiments 1-7, wherein the diamond membrane contacted with the target substrate in step (b) is provided on an intermediate substrate, wherein the diamond membrane is adhered to the intermediate substrate via a layer of a photoresist directly between and in contact with the diamond membrane and the intermediate substrate. An embodiment 9 is the method of embodiment 8, wherein the photoresist is a positive photoresist. An embodiment 10 is the method of any of embodiments 8-9, wherein the photoresist is characterized by a T_g of no greater than 200 °C and over a range of no more than 20 °C. An embodiment 11 is the method of any of embodiments 8-10, wherein the positive photoresist comprises a cresol novolak resin or polymethylmethacrylate.

[0013] An embodiment 12 is the method of any of embodiments 1-11, wherein the target substrate is fused silica, thermal oxide silicon, sapphire, lithium niobate, silicon, or yttrium iron garnet.

[0014] An embodiment 13 is any of embodiments 1-12, wherein step (b) comprises heating via a first heating stage and a second heating stage. An embodiment 14 is the method of embodiment 13, wherein the first heating stage comprises heating to an intermediate temperature selected to soften a layer of a photoresist in contact with the diamond membrane; and further wherein the second heating stage comprises heating to a final temperature greater than the intermediate temperature and under a non-oxidizing atmosphere. An embodiment 15 is the method of embodiment 14, wherein the final temperature is at least 500 °C.

[0015] An embodiment 16 is any of embodiments 1-15, wherein the bonding interface is crystalline across its thickness as measured using high resolution transmission electron microscopy (HRTEM). An embodiment 17 is the method of embodiment 16, wherein the bonding interface has a thickness of no more than 0.5 nm as measured using HRTEM.

[0016] An embodiment 18 is any of embodiments 1-17, wherein the diamond membrane is provided on an intermediate substrate, wherein the diamond membrane is adhered to the intermediate substrate via a layer of a photoresist directly between and in contact with the diamond membrane and the intermediate substrate, and further wherein the method comprises subjecting the surface of the diamond membrane to plasma ashing to provide a plasma treated surface of the diamond membrane prior to step (b).

[0017] An embodiment 19 is a bonded diamond membrane heterostructure comprising a plasma treated target substrate having a plasma treated surface and a diamond membrane having a surface, the plasma treated target substrate bound via covalent bonds to the diamond membrane at a bonding interface formed between the plasma treated surface of the plasma treated target substrate and the surface of the diamond membrane, wherein the bonding interface is crystalline across its thickness as measured using HRTEM. [0018] An embodiment 20 is the bonded diamond membrane heterostructure of embodiment 18. wherein the bonding interface has a thickness of no more than 0.5 nm as measured using HRTEM.

[0019] An embodiment 21 is the bonded diamond heterostructure of any of embodiments 19-20, wherein the surface of the diamond membrane is a plasma treated surface.

[0020] An embodiment 22 is the bonded diamond membrane heterostructure of any of embodiments 19-21, wherein the diamond membrane is single-crystalline and the surface of the diamond membrane is (100).

[0021] An embodiment 23 is the bonded diamond membrane heterostructure of any of embodiments 19-22, wherein the diamond membrane has a thickness of no more than 500 nm.

[0022] An embodiment 24 is the bonded diamond membrane heterostructure of any of embodiments 19-23, wherein the plasma treated target substrate is fused silica, thermal oxide silicon, sapphire, lithium niobate, silicon, or yttrium iron garnet

[0023] Other principal features and advantages of the disclosure will become apparent to those skilled in the art upon review of the following drawings, the detailed description, and the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Illustrative embodiments of the disclosure will hereafter be described with reference to the accompanying drawings.

[0025] FIG. 1 A shows a schematic depiction of steps in the diamond membrane transfer onto a photoresist coated intermediate wafer via two patterned stamps, PDMS1 and PDMS2. FIGS. 1B-1D show images corresponding to the three steps in FIG. 1A. Specifically, FIG. IB is an image of alignment and pick-up of the diamond membrane with PDMS 1 (four little squares in contact with the membrane’s comers). FIG. 1C is an image of membrane flipping by transferring the membrane from PDMS1 to PDMS2 (the larger, outer square). FIG. ID is an image of membrane placement on the photoresist coated intermediate wafer via PDMS2.

Here, the intermediate wafer is a fused silica substrate with a stool elevated by 5 pm from the surrounding substrate. [0026] FIG. 2A shows a schematic depiction of a steps in a method of forming a bonded diamond membrane heterostructure according to an illustrative embodiment. FIGS. 2B-2E show images corresponding to the four steps in FIG. 2A. Specifically, FIG. 2B is an image of membrane alignment to the target wafer. A rainbow color was observed due to optical interference from using a non-zero approaching angle. FIG. 2C shows the membrane after contact with the target wafer while heating. The image shows the heated photoresist flowing over the membrane to surround it on all sides. FIG. 2D shows the membrane on the target wafer after dragging and lifting off the intermediate wafer. Residual photoresist remains. FIG. 2E shows the final bonded diamond membrane heterostructure after annealing and photoresist removal.

[0027] FIGS. 3A-3B illustrate vertical strain in transferred diamond membranes. FIG. 3A shows a Raman image of the original membrane and overgrowth layers. FIG. 3B shows an image of a curved diamond membrane. The arrow indicates the presence of a rainbow color indicative of the curved membrane on the PDMS stamp.

[0028] FIGS. 4A-4H show small (200 nm range) and large (10 pm range) area AFM images of the He⁺ damaged side of diamond membranes under various conditions. The rootmean-square roughness values (Rq) are indicated in the images. FIGS. 4A-4B show images obtained after Ar/Ch etching described in “Multi-cycle ICP etching” of the Example. FIGS. 4C-4D show images obtained after O2 etching, no plasma added. FIGS. 4E-4F show images obtained after O2 plasma treatment according to a first set of conditions (“O2 descum”) as described in “Plasma treatment” in the Example. FIGS. 4G-4H show images obtained after O2 plasma treatment according to a second set of conditions (“high power”) as described in “Plasma treatment” in the Example. The defect-free area has surface roughness of 0.29 nm (0.35 nm) in FIG. 4G (FIG. 4H).

[0029] FIG. 5A is an image of a bonded diamond membrane heterostructure formed according to an illustrative embodiment of the present methods. The bonded diamond membrane heterostructure is composed of a diamond membrane (smaller, lighter square) bound to an underlying thermal oxide silicon wafer (larger, darker square) without any intervening layer therebetween. FIG. 5B shows microscope images of 155 nm-thick diamond membrane bonded to a thermal oxide silicon substrate with markers left (left) and a fused silica substrate with a 5 pm-deep trench etched prior to bonding (right). [0030] FIG. 6 is a schematic depiction of a method of forming a bonded diamond membrane heterostructure according to an illustrative embodiment.

[0031] FIG. 7 is a schematic depiction of a device incorporating a bonded diamond membrane heterostructure according to an illustrative embodiment.

[0032] FIGS. 8A-8E show the characterization of a bonded diamond membrane heterostructure. FIG. 8A is an AFM image of the diamond bonding interface (the etched side) post ICP etching. Atomically flat surfaces with Rq < 0.3 nm were observed in both small (200 nm by 100 nm, the upper figure) and large (10 pm by 5 pm, the lower figure) scanning areas. FIG. 8B shows a plot of the contact angle and XPS of diamond and sapphire pre- and post- high power plasma treatments. An increase of hydrophilicity is observed via the decrease of the contact angle, and the effect of oxygen termination is observed through the reduction of the carbon sp² as obtained from C KLL extrapolation of the sp²/sp³ ratio and the enhancement of the sapphire-0 signals as obtained from the O Is peak quantification. A PL map of GeV centers in a membrane bonded to a DBR mirror was obtained but not shown. The signal-to- background ratio around zero phonon line (ZPL) was as high as 65, with the signal surpassing 65 kc s '. FIG. 8C shows the profilometry of a membrane-silicon heterostructure. The membrane region is highlighted by two dashed lines. The thickness of the membrane is 493.7 nm with a standard deviation of 1. 1 nm. FIG. 8D shows a HRTEM image of a 10 nm-thick membrane bonded to a c-plane sapphire substrate. The 2 nm layer on top of diamond comes from the lack of surface control before gold deposition. FIG. 8E shows (top): the zoomed-in HRTEM image of the diamond-sapphire bonding interface, the dashed rectangle region in FIG. 8D, showing a sub-0.5 nm thickness of the bonding interface and (bottom): EDS elemental analysis across the bonding interface.

[0033] FIG. 9A show schematics of a TiOi-based (top) and diamond-based (bottom) nanophotonic devices on bonded diamond membrane heterostructures. Fused silica (thermal oxide silicon) wafers were used as carrier wafers for the TiCh (diamond)-based demonstrations. The grating couplers for excitation (collection) are labeled. FIG. 9B is a schematic illustration of a flow channel device fabricated using a diamond membrane bonded heterostructure (diamond membrane bound to a fused silica coverslip). Also shown is use of the device to image a cell illuminated by total internal reflection through the diamond membrane. DETAILED DESCRIPTION

[0034] In one aspect, methods of forming bonded diamond membrane heterostructures are provided. The methods comprise generating a plasma comprising plasma activated species; exposing a surface of a target substrate to the plasma activated species to provide a plasma treated target substrate having a plasma treated surface; and contacting the plasma treated surface of the plasma treated target substrate with a surface of a diamond membrane under conditions to form a bonded diamond membrane heterostructure. As further described below, the plasma treatment may be “plasma ashing.” The bonded diamond membrane heterostructure comprises the target substrate which is bound, e.g., via covalent bonds, to the diamond membrane at a bonding interface formed between the plasma treated surface and the surface of the diamond membrane.

[0035] The composition of the diamond membrane is that of a solid carbon matrix in which the carbon atoms are substantially sp³ hybridized. The diamond membrane may be characterized by its degree of crystallinity, which is generally high. In embodiments, the diamond of the diamond membrane is single-crystalline, i.e., the diamond membrane is a single-crystal diamond membrane. The diamond membrane may be characterized by its lattice structure at the surface of the diamond being contacted with the plasma treated surface of the plasma treated target substrate. In embodiments, this surface is (100) diamond. In addition to carbon, the diamond membrane may be doped such that the diamond membrane may comprise other elements (including ions or isotopes thereof), e.g., N, Ge, Si, Sn, etc. In embodiments, the diamond membrane is a ¹²C isotopic purified diamond membrane.

[0036] The diamond membrane is characterized by having a thickness that is substantially less than that of the other two dimensions of the diamond membrane. The thickness is generally nanoscale, i.e., no more than 1 pm. This includes having a thickness of no more than 750 nm, no more than 500 nm, no more than 250 nm, or in a range of from 3 nm to 250 nm, from 100 nm to 200 nm, or from 5 nm to 50 nm. Extremely thin diamond membranes may be used, including those having a thickness in a range of from 10 nm to 15 nm. The thickness of the diamond membrane may be measured from atomic force microscopy (AFM) images and a profilometer. The thickness values may refer to an average value as determined from such AFM images/profilometry. The other two dimensions of the diamond membrane are not particularly limited, although they are greater than its thickness, e.g., in a range of from 10 pm to 10 mm. Thus, the diamond membranes may be characterized as having a planar, two-dimensional morphology. The shape of the diamond membranes as defined by the two dimensions perpendicular to the thickness is not particularly limited.

[0037] The diamond membrane to be used in the present methods is further characterized by having low surface roughness and low surface curvature. Surface roughness may be quantified by reference to root-mean-square roughness values (Rq) as determined, e.g., by using atomic force microscopy (AFM). In embodiments, the diamond membrane has an Rq value in a range of from 0.2 nm to 0.9 nm as measured over an AFM area of from 0.04 to 100 pm²). This includes from 0.2 nm to 0.6 nm and from 0.2 nm to 0.4 nm.

[0038] The present diamond membranes are distinguished from bulk diamond which refers to diamond having substantially greater thicknesses than those described above, including thicknesses of greater than about 50 pm. Bulk diamond is also generally characterized by having surface curvatures greater than that of the diamond membranes, due to surface polishing that is required for bulk diamond. By way of illustration, the diamond used in the following references was bulk diamond: Matsumae, T., et al., Scripta Materialia 175 (2020) 24-28; Matsumae, T., et al., Scientific reports 11.1 (2021): 11109; Liang, J. et al.. Applied Physics Express 12.1 (2018): 016501 ; and Liang, J. et al., Applied Physics Letters 110.11 (2017).

[0039] Techniques for synthesizing the diamond membranes to be used in the present methods include those described in described in X. Guo, et al., Nano Letters 21, 10392 (2021), which is hereby incorporated by reference in its entirety. Briefly, such a method involves carrying out He⁺ implantation and annealing on a single-crystal, optical grade diamond mother substrate, diamond overgrowth via plasma enhanced chemical vapor deposition (PE-CVD), in situ doping (if desired), and diamond membrane undercutting via electrochemical etching (EC). Diamond membranes synthesized using such methods may be characterized by having a region of He-damaged lattice therein due to the He⁺ implantation. Other methods may be used to synthesize the diamond membranes.

[0040] The diamond membrane to be used in the present methods may be provided on an intermediate substrate (in this phrase the term “substrate” and “wafer” may be used interchangeably). This is illustrated in box 104 of FIG. 1A showing a diamond membrane 116 adhered to an intermediate substrate 118 via a layer of a photoresist 120. The diamond membrane 116 is adhered at its surface opposite that of the surface to be contacted with the target substrate. Boxes 100 and 102 of FIG. 1A illustrate the origin the of the diamond membrane 116 as synthesized from a diamond mother substrate 122 using the diamond membrane synthesis technique described above. Box 102 illustrates the transfer of the diamond membrane 116 from its diamond mother substrate 122 to the photoresist coated intermediate substrate 118 using poly dimethylsiloxane (PDMS) stamps PDMS1 and PDMS2. Any visibly transparent material may be used as for the intermediate substrate.

[0041] Positive photoresists may be used to provide the layer of the photoresist 118. Suitable such positive photoresists include those comprising cresol novolak resins such as AZ 1505 photoresist (available from EMD Performance Materials Corp.), Microposit™ SI 805™ photoresist (available from The Dow Chemical Company), AZ MiR 703 photoresist (available from EMD Performance Materials). Other suitable such positive photoresists include those comprising polymethylmethacrylate (PMMA) such as PMMA A4 photoresist (e.g., 950 PMMA A4, 495 PMMA A4). It has been found that these positive photoresists, particularly AZ 1505, are useful in the present methods as they have a relatively low, well- defined glass transition temperature T_g range and exhibit and relatively low viscosities over this temperature range. For example, AZ 1505 exhibits a T_g range of from about 100 °C to about 110 °C; AZ MiR 703 exhibits a T_g range of from about 130 °C to about 135 °C; and PMMA A4 photoresists exhibit a T_g range of from about 95 °C to about 106 °C. In embodiments, the photoresist exhibits a T_g of below about 200 °C (e.g., less than about 150 °C or less than about 140 °C) and over a range of no more than about 20 °C (e.g., about 15 °C or about 10 °C). In embodiments, the photoresist is not a negative photoresist. In embodiments, the photoresist does not comprise hydrogen silsesquioxane and hydrogen silsesquioxane is not used in the present methods. A variety of thin-film coating techniques may be used to coat the intermediate substrate with the layer of the photoresist. Generally, the layer of the photoresist is quite thin, e.g., from 80 nm to 500 nm. This includes from 100 nm to 400 nm and from 150 nm to 300 nm.

[0042] The present methods may be used to bond the diamond membrane to a variety of target substrates (in this phrase the term “substrate” and “wafer” may be used interchangeably). However, the target substrate has a composition different from that of the diamond membrane. The composition of the target substrate generally depends upon the application for the bonded diamond membrane heterostructure. However, illustrative materials include optically non-linear materials, piezo-electric materials, superconducting materials, materials that benefit from thermal management, magnetic materials, biocompatible materials (glasses, oxides), metals that are amenable to oxygen terminated interfaces and resilient to stamping processes. Specific such illustrative materials include fused silica, thermal oxide silicon, sapphire, lithium niobate, silicon, and yttrium iron garnet (YIG). The shape and dimensions of the target substrate are not particularly limited, but rather, depend upon the application.

[0043] As noted above, the present methods involve plasma treating the surface of the target substrate to be bonded to the diamond membrane. Plasma treatment is a dry functionalization process as distinguished from wet chemical functionalization processes, e.g., using wet chemicals such as H2SO4, H2O2, NH3, etc. The plasma treatment is carried out by generating the plasma in a gas (which may be a gas mixture), which creates the plasma activated species comprising ions, free radicals, etc. derived from the gas(es). The plasma, and thus, the plasma activated species, may be generated at a location remote from the location of the target substrate. Such a configuration involves subsequently transporting the plasma activated species to the target substrate at its remote location. For example, the plasma may be generated in a first chamber and the plasma activated species transported to the target substrate positioned in a different chamber downstream from the first chamber.

This and other embodiments of the plasma treatment may involve filtering the plasma activated species and exposing the surface of the target substrate to filtered plasma activated species. Both such embodiments prevent more energetic/reactive plasma activated species from impacting the surface of the target substrate in favor of less energetic/reactive (including electrically neutral) plasma activated species. Each of these features characterize the plasma treatment as “plasma ashing” which is distinguished from “plasma etching” or “reactive ion etching.” Thus, in embodiments, the present methods comprise subjecting the surface of the target substrate to plasma ashing using a plasma comprising plasma activated species to provide a plasma treated target substrate having a plasma treated surface; and contacting the plasma treated surface of the plasma treated target substrate with a surface of a diamond membrane under conditions to form a bonded diamond membrane heterostructure.

[0044] Regarding plasma etching and reactive ion etching, these techniques involve configurations in which the substrate to be treated is positioned in the same chamber in which the plasma is generated, the plasma activated species are unfiltered, and/or, the more energetic/reactive (including electrically charged) plasma activated species are allowed to impact the exposed surface. By way of illustration, surface treatments used in the following references involved one or more of plasma etching, reactive ion etching, and wet functionalization, all as distinguished from plasma ashing: Matsumae, T., et al., Scripta Materialia 175 (2020) 24-28; Matsumae, T., et al., Scientific reports 11.1 (2021): 11109; Wang, F., et al.. Applied Sciences 12.7 (2022): 3261. Surface treatments used in the following references made use of fast Ar beams rather than any plasma: Liang, J. et al. , Applied Physics Express 12.1 (2018): 016501 ; and Liang, J. et al., Applied Physics Letters 1 10.11 (2017).

[0045] A variety of gases may be used to generate the plasma being used, including in plasma ashing. However, in embodiments, the plasma is an O2 plasma. In such embodiments, the plasma activated species impacting the surface of the target substrate may comprise or consist of monatomic oxygen. Without wishing to be bound to any particular theory, it is believed that the O2 plasma treatment (including O2 plasma ashing) carried out as described herein generates an oxygen terminated plasma treated surface. Oxygen termination refers to termination with oxygen atoms and is distinguished from hydroxyl termination. Oxygen termination may be confirmed using X-ray photoelectron spectroscopy (XI’S) as described in the Example below. (See also FIG. 8B.)

[0046] The plasma treatment (including plasma ashing) may be characterized by the conditions used to generate the plasma, including the gas flow rate, the power (which may be a radio-frequency (RF) power), the treatment temperature, and the treatment time (i.e., length of time the target substrate is exposed to the plasma activated species). These conditions may be adjusted to facilitate the bonding between the plasma treated surface of the target substrate and the diamond membrane. This may include facilitating the oxygen termination noted above. Illustrative values of these parameters include gas flow rates of from 5 seem to 250 seem; RF powers of from 100 W to 650 W; treatment temperatures of room temperature (20 °C to 25 °C) to 150 °C; and treatment times of a few seconds to minutes. Gas flow rates of from 75 seem to 150 seem and 175 seem to 225 seem are encompassed. RF powers of from 150 W to 250 W and from 575 W to 625 W are encompassed. In embodiments, the treatment temperature is room temperature. As discussed in the Example, below, room temperature was found to improve the bonding process. Treatment times of from 10 s to 60 s and 95 s to 175 s are encompassed. The plasma treatment may be carried out a single time (i.e., once) or multiple times (e.g., 2, 3, etc. times).

[0047] In embodiments, the surface of the diamond membrane to be bound to the plasma treated surface of the plasma treated target substrate is untreated. By "untreated" it is meant that the surface of the diamond membrane is not exposed to the plasma treatments described herein. It is further meant that the surface is not exposed to wet chemical functionalization, e.g., using sulfuric acid, ammonia, peroxide. However, the term “untreated” does not preclude the processing of the diamond membrane that accompanies formation of the diamond membrane itself, e.g., using the synthesis techniques described above. The term “untreated” further does not preclude processing that may occur after formation of the bonded diamond membrane heterostructure.

[0048] In other embodiments, the surface of the diamond membrane to be bound to the plasma treated surface of the plasma treated target substrate is also plasma treated, i.e., is also exposed to plasma activated species from a generated plasma. The plasma gas(es) and plasma conditions used may be the same or different as compared to those used to treat the target substrate. However, in embodiments, the plasma is an O2 plasma. In embodiments, the surface of the diamond membrane is subjected to plasma ashing, including using an O2 plasma.

[0049] If the diamond membrane comprises a region of He-damaged lattice as described above, this region may be removed prior to contacting the diamond membrane to the target substrate to induce bonding. If the diamond membrane is to be plasma treated, the He- damaged lattice may be removed prior to the plasma treatment. The removal may be carried out using an inductively coupled plasma (ICP) etching process as described in the Example, below.

[0050] After plasma treatment (including plasma ashing) of the target substrate (and optionally, plasma treatment (including plasma ashing) of the diamond membrane), the plasma treated surface of the target substrate and the (plasma treated) surface of the diamond membrane are brought together until they contact one another across their respective surfaces, thereby forming a bonding interface. The bonding interface is formed by direct contact of the target substrate and the diamond membrane with one another without any intervening material therebetween. Without wishing to be bound to any particular theory, it is believed that covalent bonds may form between the individual atoms of the target substrate and the individual carbon atoms of the diamond membrane. Again, without wishing to be bound to any particular theory, the covalent bonds may compnse those represented by the formula (-O-), where each represents a covalent bond to an atom of the target substrate and a carbon atom of the diamond membrane, respectively (“0” represents oxygen). The bonding interface is discussed further below.

[0051] The contacting step is earned out under conditions to facilitate the bonding between the plasma treated surface of the target substrate and the (plasma treated) diamond membrane. This may include formation of the covalent bonds noted above. For example, the contacting step generally comprises heating. The heating may be carried out in more than one stage, which is useful for embodiments in which the diamond membrane is provided on a photoresist coated intermediate substrate as shown in FIG. 2 A. For example, as shown in box 200, the target substrate 224 and the diamond membrane 216 are brought together until they contact one another across their respective surfaces 224a, 216a. In this embodiment, both respective surfaces 224a and 216a have been plasma treated as indicated by the bold dashed line. As shown in box 202, in a first heating stage, heat is applied to raise the temperature from an initial temperature (e.g., room temperature) to an intermediate temperature. The intermediate temperature is selected to soften the layer of photoresist 220. Thus, the specific intermediate temperature depends upon the type of photoresist being used. The heating rate, heating time (i.e., length of time of heating), and use of one or more isothermal holds during the heating may be adjusted as desired, e.g., to ensure a uniform reflow of the softened photoresist 220 over the entire diamond membrane 216 and to facilitate subsequent removal of the intermediate substrate as shown in box 204. By way of illustration, as described in the Example below, when AZ 1505 was used as the layer of photoresist, heat was applied to raise the temperature from room temperature to 125 °C and included holds at 75 °C and 95 °C. The first heating stage may be carried out without applying any mechanical force (other than that from the overlying photoresist coated intermediated membrane). The first heating stage may be carried out under atmospheric pressure.

[0052] As shown in box 206, after removing the intermediate substrate, additional heat may be applied in a second heating stage (which may be referred to as an annealing stage), to a final temperature that is generally greater than the intermediate temperature of the first heating stage. Prior to the second heating stage, the diamond membrane heterostructure may be cooled to room temperature and thus, the second heating stage may be initiated at room temperature. The second heating stage may be carried out under a non-oxidizing atmosphere (e.g., an Ar/H? gas mixture) selected to prevent oxidation of the diamond membrane during the annealing and at the final temperature. The final temperature, heating rate, heating time, use of one or more isothermal holds, and atmosphere may be adjusted as desired, e.g., to facilitate bonding without inducing oxidation. However, in embodiments, the final temperature is at least 450 °C, at least 500 °C, at least 525 °C, at least 550 °C, or in a range of from 500 °C to 550 °C. As described in the Example, below, lower temperatures may not achieve bonding or the bonding may fail after a final cleaning step. The second heating stage may be carried out without applying any mechanical force.

[0053] The result of contacting the target substrate and the diamond membrane together under heat as described above is a bonded diamond membrane heterostructure 226 as illustrated in box 206 of FIG. 2A. Any residual photoresist present on the heterostructure 226 may be removed, e.g., by applying a cleaning composition. In embodiments, the cleaning composition is a di-acid cleaning composition comprising HiSOi HNOi. Use of a cleaning composition may not be necessary for all types of photoresists, e.g., PMMA.

[0054] An illustrative embodiment of the present methods is further illustrated in FIG. 6. Box 600 corresponds to the steps depicted in boxes 100-104 of FIG. 1A. However, in this embodiment, the He-damaged region of the diamond membrane is explicitly labeled as 628. Also shown in this embodiment is the removal of the He-damaged region 628 in box 602. Box 604 corresponds to the steps depicted in boxes 200-206 of FIG. 2A.

[0055] Plasma treated (including plasma ashed) surfaces as provided by the present methods may be characterized by a variety of properties, including Rq value, water contact angle, and oxygen termination. Regarding Rq, this value may be less than 0.35, less than 0.33, less than 0.30, or less than 0.28. (See also FIGS. 4A-4H and 8A.) The Rq for nondiamond plasma treated surfaces may be higher, e.g., less than 0.53, less than 0.40, less than 0.30. Rq values may be determined via AFM as described in the Example below and may refer to both a small area (e.g., 200 nm by 100 nm) and a large area (e.g., 10 pm by 5 pm). Regarding water contact angle, it may be less 40°, less than 35°, less than 30°, less than 28°, less than 25°, less than 22°, or less than 20°. (See FIG. 8B.) Water contact angles may be determined using the technique described in the Example, below. As noted above and described in the Example, below, oxygen termination may be confirmed using XPS. For plasma treated diamond, oxygen termination may be confirmed through a reduction of the amount of carbon sp² in atomic (at.) % (as obtained from C KLL extrapolation of the sp²/sp³ ratio), which may be a reduction of at least 2 or 3. For plasma treated non-diamond surfaces, oxygen termination may be confirmed through an enhancement in the amount of non- diamond-0 signal (as obtained from O ls peak quantification), which may be an enhancement of at least 2 or 3.

[0056] The present disclosure further encompasses the bonded diamond membrane heterostructures formed using the present methods. As noted above, bonded diamond membrane heterostructures are provided which comprise the target substrate covalently bound (e.g., via -0- bonds) to the diamond membrane at a bonding interface formed between the plasma treated surface and the surface of the diamond membrane.

[0057] An illustrative bonded diamond membrane heterostructure formed according to the present methods is shown in FIG. 5A. In this embodiment, the diamond membrane (smaller, lighter square) is covalently bound to a thermal silicon dioxide target substrate (larger, darker square) without any intervening material therebetween.

[0058] The bonded diamond membrane heterostructures fabricated using the present methods are characterized by high quality as evidenced by high resolution transmission electron microscope (HRTEM) images, as further discussed in the Example, below. Briefly, HRTEM images reveal that the diamond membrane retains uniform crystal 1 i ni ty and morphology throughout its thickness. Specifically, HRTEM images such as those shown in FIGS. 8D-8E, show that the diamond membrane remains single-crystalline post bonding. In this embodiment, the non-diamond target substrate is sapphire, which the HRTEM images also reveal retains its crystallinity post bonding, HRTEM images such as that shown in FIG. 8E also show that the bonding interface, which refers to the region formed between the surface of the diamond membrane in contact with the surface of the non-diamond target substrate, is also crystalline (as opposed to amorphous). This is evidenced by the bonding interface producing a lattice image via HRTEM throughout its thickness. This is by contrast to a bonding interface producing a dark image via HRTEM, which is indicative of an amorphous, rather than crystalline, atomic structure. Finally, HRTEM images such as that shown in FIG. 8E, also show that the bonding interface is extremely thin, in this embodiment, no more than 0.5 nm. The thickness of the bonding interface may be measured from such HRTEM images and corresponds to the thickness of the transition region in the lattice image of the bonded diamond membrane heterostructure (labeled by the arrow in FIG. 8E.)

[0059] Thus, the present bonded diamond membrane heterostructures may be characterized by having a crystalline bonding interface. The bonding interface may be further characterized by having a thickness of no more than 0.5 nm, no more than 0.4 nm, or no more than 0.3 nm.

[0060] By contrast, existing techniques either result in amorphous bonding interfaces, much thicker bonding interfaces, or both. See Matsumae, T., et al., Scrlpta Materialla 175 (2020) 24-28; Matsumae, T., etal., Scientific reports 11.1 (2021): 1 1 109; Wang, E. etal., Applied Sciences 12.7 (2022): 3261; Liang, J. et al. , Applied Physics Express 12.1 (2018): 016501; and Liang, J. et al., Applied Physics Letters 110.11 (2017).

[0061] Finally, the high quality of the present bonded diamond membrane heterostructures fabricated using the present methods is further evidenced by color centers with in the diamond membrane (e.g., GeV, NV) exhibiting high signal-to-noise via photoluminescence spectroscopy. These results are further discussed in the Example, below.

[0062] Devices incorporating the bonded diamond membrane heterostructures fabricated using the present methods are also provided. Any device in which high qualify diamond is generally used is encompassed by the present disclosure. However, an illustrative device is an electrically-reconfigurable multiplexed quantum photonic device 700 shown in FIG. 7. This device comprises a lithium niobate substrate 724 (target substrate) and patterned diamond membranes 716 which are directly bonded together using the present methods. The device 700 further comprises electrodes 730 (two of which are labeled) in electrical communication with the lithium niobate for phase shifting. Other illustrative photonic devices are shown in FIG. 9A and an illustrative flow channel device is shown in FIG. 9B, each of which is further described in the Example below.

EXAMPLE

[0063] Introduction

[0064] Diamond has superlative material properties for a broad range of quantum and electronic technologies. However, heteroepitaxial growth of single crystal diamond remains limited, impeding integration and evolution of diamond-based technologies. In this Example, single-crystal diamond membranes are directly bound to a wide variety of materials including silicon, fused silica, sapphire, thermal oxide silicon, and lithium niobate. The bonding process combines customized membrane synthesis, transfer, and dry surface functionalization based on certain plasma treatments, allowing for minimal contamination while providing pathways for near unify yield and scalability. As further described below, bonded crystalline membranes with thickness as low as 10 nm, sub-nm interfacial regions, and nanometer-scale thickness variability over 200 by 200 pm² areas were generated. In addition, the resulting bonded diamond membrane heterostructures were integrated with high quality factor nanophotonic cavities, highlighting the platform versatility in quantum photonic applications. Furthermore, it has been shown that the bonded diamond membrane heterostructures are compatible with total internal reflection fluorescence (TIRF) microscopy, enabling interfacing coherent diamond quantum sensors with living cells while rejecting unwanted background luminescence. The processes demonstrated below provide a full toolkit to synthesize heterogeneous diamond-based hybrid systems for quantum and electronic technologies.

[0065] Fabrication Process

[0066] Diamond membrane synthesis and patterning

[0067] Diamond membranes were synthesized according to the method described in the paper by X. Guo, et al. , Nano Letters 21, 10392 (2021). Briefly, single crystal, optical grade diamond substrates were subjected to He⁺ implantation (dose 5 x 10¹⁶ cm ². energy 150 keV), followed by an annealing process in an argon forming gas environment (4 % H2, 96% Ar). The annealing included three isothermal holds, 400 °C for 8 h, 800 °C for 8 h, and 1200 °C for 2 h. The diamond overgrowth was performed in a microwave plasma chemical vapor deposition (MPCVD) chamber at Argonne National Laboratory. Four membranes were used in this Example using diamond overgrowth layers grown to thicknesses of 185 nm, 260 nm, 400 nm, and 660 nm, respectively. Post overgrowth, samples received either ion implantation (Si⁺, Ge⁺, Sn⁺ or N⁺) or ^-doping ¹⁵N. Next, samples were patterned to form individual 200 pm by 200 pm membranes via lithography, inductively coupled plasma (ICP)-etching, and undercutting via electrochemically (EC) etching in deionized (DI) water.

[0068] Patterned PDMS stamps for membrane transfer to intermediate wafers

[0069] Patterned PDMS stamps were used for membrane transfer to intermediate substrates. Two different PDMS patterns were transferred from inverse SU-8 (3050, with thickness 55 pm) structures lithographically defined on a 4-inch silicon wafer. The first pattern consisted of four squares (see FIG. 1A, box 100, and FIG. IB) to pick up the diamond membrane from its diamond mother substrate, while the second contained a single large square to realize membrane flipping by utilizing a larger adhesion area (see FIG. 2A, box 102, and FIG. 2C). The membrane pick-up, flipping, and placement were carried out using a probe station (Signatone SI 160).

[0070] The diamond membrane transfer process shown in FIGS. 1A-1D had several advantages. The first PDMS stamp (PDMS1) only broke the tether to the diamond mother substrate and picked up the membrane from its comers, which effectively reduced the contact area and thus brought less transfer-induced contamination to the growth side. In addition, neighboring partially -etched membranes on the diamond mother substrate were protected, which effectively improved the overall transfer yield to 100%. This process also allowed for EC etching and transferring of multiple membranes in a single cycle, with a current record of 6. The second PDMS stamp (PDMS2) was 300 pm by 300 pm, which preserved the existing structures on the intermediate wafer outside of the transfer area. In summary, the patterned PDMS method achieved a unity yield, protected both mother substrates and intermediate wafers, and greatly improved the scalability.

[0071] Intermediate wafer preparation

[0072] The intermediate wafers used in this Example were 13 mm by 13 mm substrates diced from a 4-inch fused silica wafer. However, other transparent substrates could be used as an intermediate wafer. Prior to the dicing step, the 4-inch wafer was patterned and ICP-etched to generate a 400 pm by 400 pm square at the center of each chip with 5 pm height to provide stools. Fabricating stools was useful to compensate for the residual tilt angle between intermediate and final (target) wafers. In addition, the stool fabrication also limited the contact region and protected existing structures on the final (target) wafer. Alternatively, wafer bonders may be used.

[0073] Prior to membrane transfer to the intermediate wafer, the patterned intermediate wafers were spin-coated with a thin layer of positive photoresist (AZ 1505, ~ 500 nm) or of electron beam resist (950 K PMMA A4, Mi croChem «250 nm). AZ 1505 is particularly suitable as it softens over a well-defined temperature range (100 °C to 110 °C), has relatively low viscosity over this temperature range, and results in minimal contamination. As shown in box 104 of FIG. 1A, after the membrane was flipped and placed on the stool of the photoresist coated intermediate wafer, because the membrane’s adhesion to the photoresist exceeded that of the PDMS2 stamp, the membrane was left attached to the intermediate wafer, ready for multi-cycle ICP etching.

[0074] Multi-cycle ICP etching [0075] Post transfer to the intermediate wafer, the He-damaged side of the diamond membrane was ICP-etched, thinning the overgrown diamond layer to a membrane target thickness. The membrane target thickness (and as such, the total etching depth) is variable and tailored to the intended application (e.g., to match a desired photonics wavelength). The minimum etching depth is set only by the thickness of the He-damaged diamond layer. This process also released the out-of-plane strain induced by the lattice mismatch between damaged and overgrown layers. Since the membrane thinning step took place on the intermediate wafer, pre-existing structures on final (target) wafers could be effectively protected. To prevent the photoresist from crosslinking at elevated temperatures, a multicycle Ar/Ch-O2/C12-O2 etching sequence was used with only 15 s etching time per cycle. Multi-cycle ICP etching also maintained a higher quality chamber environment as compared to continuous etching by utilizing more pump-purge cycles in between etching cycles. Multicycle ICP etching also provided greater control over the etching rate and thus, the membrane thickness.

[0076] Plasma treatment on membranes and target wafers

[0077] A downstream plasma asher (YES-CV200 RFS Plasma Strip/Descum System from Yield Engineering Systems Inc.) was used to activate surfaces prior to bonding. The plasma was generated using O2, but other gases may be used (N2, Ar, etc.). Two different sets of plasma conditions were used to treat ICP-etched diamond membranes and target wafers. The first set of plasma conditions was as follows (which may be referred to herein via the phrase “O2 descum”): O2 gas flow at 100 seem; RF power at 200 W; temperature at room temperature; and an exposure time of 25 s. The second set of plasma conditions was as follows (which may be referred to herein via the phrase “high-power”): O2 gas flow at 200 seem; RF power at 600 W; temperature at room temperature; and an exposure time of 150 s. The room temperature condition was used since a degradation of hydrophilicity was observed after a brief baking (90 °C on a hot plate for 1 min) post plasma treatments (discussed further below). Some target wafers, e.g., LiNbCh wafers, were subjected to the plasma treatment more than once, e g., 3 times.

[0078] The two recipes were compared according to three metrics measured on the resulting products, namely, surface morphology, surface hydrophilicity, and the optical performance of GeV- and NV“ centers. The effects on both carrier (i.e., target) wafer and diamond membrane surfaces were assessed. Results are discussed below. For the carrier wafer choices, fused silica and thermal oxide silicon were tested. Both recipes led to enhanced surface hydrophilicity while maintaining good surface morphology, which are favorable for the subsequent bonding process. The high-power recipe was used for all carrier substrates due to the better hydrophilicity achieved. For diamond membranes, both recipes improved the signal to background ratio for optical characterizations of color centers. However, an increased number of particle-like contaminates were observed under atomic force microscopy (AFM) on the treated surfaces post high-power recipes. Therefore, the O2 descum recipe was used for diamond membranes (except for NV sensing applications which required a more adequate charge state preparation). However, contamination could be eliminated by using specialized tooling.

[0079] It is noted that plasma treatment of target substrates alone (no diamond membrane plasma treatment) is sufficient to achieve high quality bonding. However, diamond membrane plasma treatment is desirable in other embodiments.

[0080] Membrane bonding to the target wafer

[0081] Membrane bonding to the target wafer was carried out using the same probe station described above. FIG. 2A schematically illustrates the steps in the process while FIGS. 2B-2E show images corresponding to these steps. The bold dashed line in FIG. 2A is used to indicate a plasma treated surface. To begin, the intermediate wafer was placed on a glass slide with a large PDMS stamp, while the final (target) wafer was placed on a temperature-controlled stage and held by vacuum. Due to the lack of full angle control on the micropositioner (Signatone CAP - 946), the approaching angle was set to 0° along the y direction, leaving the angle along the x to be a small but not well-defined value. The diamond membranes were aligned by looking through the transparent PDMS stamp and intermediate wafer via microscope. The alignment precision was limited to 30 pm and 0.1° due to the weakly defined approaching angle. Alignment precision may be improved by using a wafer bonder. As shown in box 200 of FIG. 2A, once the target location was found, the intermediate wafer was slowly lowered until the membrane made contact with the final (target) wafer. Next, as shown in box 202 of FIG. 2A, heating was applied to initiate bonding and soften the photoresist (PR). The temperature was increased from room temperature step- wise (75 °C, 95 °C, and 125 °C for AZ 1505; 90 °C, 130 °C, and 170 °C for PMMA), allowing the resist to reach thermal equilibrium at each stage. No mechanical force was applied during heating (other than that from the overlying photoresist coated intermediated membrane). The membrane was fully covered and surrounded by the photoresist when the flow pattern entered an equilibrium state and the intermediate wafer had a slight shift with respect to the membrane due to the non-zero contact angle and the softening of the photoresist. Next, as shown in box 204 of FIG. 2A, the intermediate wafer was slowly- dragged off of the membrane region and lifted up, leaving the membrane on the target wafer along with some remaining photoresist. Finally, as shown in box 206 of FIG. 2A and further described below, after cooling to room temperature, the diamond membrane heterostructure annealed to complete the bonding.

[0082] Annealing and photoresist removal

[0083] Annealing was used to facilitate bonding between the diamond membrane and the target wafer. Argon forming gas (96% Ar, 4% H2, ~ 1 atm) was used during annealing to eliminate oxidization of the diamond surface at elevated temperatures. Post annealing, di-acid cleaning (1: 1 FhSC HNCh, at a temperature of 225 °C for 2 h) was used to remove the residual crosslinked photoresist (if necessary).

[0084] Two annealing temperatures were used, 450 °C and 550 °C. Membranes annealed at 450 °C ultimately detached from their underlying target wafer during acid cleaning. By contrast, membranes annealed at 550 °C remained bonded after acid cleaning.

[0085] When PMMA was used as the resist, residual resist was fully baked out post annealing and only a brief O2 descum cleaning was used to clean the bonded diamond membrane heterostructure. However, when AZ 1505 was used, residual photoresist crosslinked, rendering it challenging to remove. However, a boiling di-acid cleaning step was found to remove crosslinked AZ 1505, involving 1: 1 FbSOrHNCh for 2 hours at the nitric boiling point. Due to the reduced viscosity of AZ 1505 compared with PMMA, PMMA- based bonding was used for structured surfaces and acid-sensitive substrates, while AZ 1505 was used on other substrates. The bonded diamond membrane heterostructure was found to be compatible with isopropyl alcohol, acetone, potassium or tetramethylammonium hydroxide (TMAH) based developers (such as AZ 300 MIF or AZ 400K), heated (80 °C) N- Methyl-2-pyrrolidone (NMP), and room temperature NanoStrip. However, the tri-acid cleaning (1: 1 : 1 H2SO4: HNO3 HCIO4 at refluxing temperature), hot (>80 °C) Piranha (3 : 1 H2SO4:H2O2), and hot NanoStrip may damage the bonds and loosen the diamond membranes from the target wafer.

[0086] Material characterizations [0087] Out-of-plane strain of smart-cut membranes

[0088] Unlike isotropically -etched diamond frames or ICP-etched diamond slabs, smartcut diamond membranes naturally contain out-of-plane strain originating from the lattice mismatch between the damaged layer generated from He⁺ implantation and the subsequent overgrowth layer. This strain brings a curvature to freestanding membranes due to their high geometry aspect ratio (usually beyond 500). To roughly estimate the strain magnitude, Raman spectroscopy was performed on a transferred diamond membrane with a 100 nm overgrowth layer prior to ICP etching. The experimental data is shown in FIG. 3A as individual points, which can be fit by two Lorentzian curves. The original membrane (He⁺ damaged layer) is indicated as the dashed curve with a center wave number of 1326 cm '. while the overgrowth layer exhibited a center wave number of 1332 cm as indicated by the other dashed curve. Differences in wave number indicate a ~ 0.5% lattice mismatch.

[0089] In FIG. 3B, a test membrane partially attached to a PDMS2-stamp is shown, wi th the upper and lower parts floated, as indicated by the arrow. From the interference pattern, the extension of the original layer and the compression to the overgrowth layer can be observed, causing the membrane to be curved up. The strain elimination via ICP etching has been discussed above.

[0090] Surface morphology of diamond and final wafers

[0091] The success of plasma-enhanced bonding is dependent upon the surface morphology of the diamond membrane and target substrates. AFM was performed to characterize the surface roughness during the fabrication process. Both small (200 nm by 200 nm) and large (10 pm by 10 pm) scale scans were applied to capture features of various sizes.

[0092] For diamond membrane surfaces, only the etched side is discussed in this Example since the growth side preserved the excellent surface morphology. Prior to the ICP etching, the etched side had an R_q of 1.17 nm due to the He⁺ implantation and EC etching. By applying 15 cycles of Ar/Ch etching discussed above, a polishing effect was shown w ith R_q reduced to 0.54 nm (0.44 nm) in small (large) area scans, as shown in FIGS. 4A-4B. It was noted that small area scans usually revealed greater roughness compared to large area scans, which might indicate the Cl-based contamination on the diamond surface. Such contamination can be removed by O2/CI2-O2 ICP cycles, shown as an R_q of 0.25 nm (0.34 nm) in small (large) areas (FIGS. 4C-4D). A change of R_q post O2 descum treatment (0.28 nm and 0.34 nm in small and large area scans) was not observed, as depicted in FIGS. 4E-4F. In contrast, the high-power recipe was found to have a negative impact on the surface morphology by elevating the R_q to 0.84 nm (1.09 nm) in small (large) areas. This can be interpreted as an appearance of particle-like dust since the R_q of the contamination-free area remains < 0.35 nm. Such contamination can be reduced by using process specific tooling.

[0093] The impact of the two downstream plasma asher recipes was also analyzed on two carrier substrates, namely fused silica and thermal oxide silicon wafer. Values of R_q are shown in Table 1, below. Both wafers exhibited sub-nm R_q out of the box, and their surface morphology was maintained post ashing with no correlation to power or duration settings. Therefore, it was concluded that the plasma recipe had no significant effect on the surface morphology of these carrier wafers.

[0094] Table 1: Rq values of target wafers under various plasma treatment conditions.

[0095] Height variation across diamond membranes

[0096] One-dimensional (ID) height detection via profilometrv

[0097] In this Example, two methods were applied for global flatness characterizations. For ID characterization presented, a profilometer (Dektak XT) was used to scan across the membrane. The scan range was 350 pm with a 6.5 pm height detection limit. The membrane showed a height variation o of ~ 1 nm, which was 10 times smaller than the HSQ-bonded membranes’ value measured on the same equipment. This o was also below the minimum detectable height of the tool (10 nm) and the instrument resolution (1.5 nm) for large scale scanning.

[0098] Two-dimensional (2D) height mapping via confocal laser scanning microscopy (CLSM)

[0099] The 2D membrane height map and surface topology were measured via an Olympus LEXT OLS4100405 nm laser confocal microscope. The microscope image of the measured membrane-thermal oxide silicon heterostructure was obtained along with its height map. The bonded membrane profile reveals a uniform height of 309 ± 8 nm across the membrane, with the standard deviation o below the height resolution of the CLSM (~ 10 nm). Currently, the dominant sources of height inhomogeneity are assigned to diamond membrane crystallographic growth defects and transfer process contamination, which can be minimized by performing the totality of the processing in a clean environment (e.g., cleanroom).

[00100] Hydrophilicity characterization of bonding interfaces

[00101] Contact angle measurement setup

[00102] Water contact angle was measured to charactenze the surface hydrophilicity of diamond and target substrates. Measurements were performed using a Kruss DSA100A dropped shape analyzer. DI water was dispensed from a sterile syringe (14-817-25, Fisher Scientific) through a thin needle (75165 A761, McMaster-Carr). The dispense rate was set to 2.67 pL s ', resulting in atypical droplet size between 4 pL to 5 pL. The diamond used for contact angle measurements were 3 mm by 3 mm single crystal fine-polished diamond substrates (R_q < 0.3 nm). Contact angle analysis of diamond and thermal oxide silicon substrates before and after high power plasma ashing was conducted, indicating an improvement of the surface hydrophilicity. It was noted that the weaker O2 descum recipe showed minimal influence on diamond hydrophilicity, with the contact angle above 40°.

[00103] Aging and temperature dependence of the surface hydrophilicity

[00104] Surface hydrophilicity is positively correlated to the plasma activated bonding quality. Here, the decay of the hydrophilicity was characterized by recurring measurements of contact angle on various substrates, including diamond, fused silica, thermal oxide silicon, sapphire, and lithium niobate on insulator. The aging trend of the hydrophilicity indicated reduced hydrophilicity over a timescale of from 3 to 24 hours, demonstrating the advantages of timely bonding. The temperature dependence of the hydrophilicity was also tested by baking the plasma-treated diamond sample on a 90 °C hotplate for 30 s pnor to the contact angle measurements. A decay of hydrophilicity was observed, possibly due to the loss of surface-absorbed water molecules.

[00105] Partially due to the strong association between elevated temperature and reduced hydrophilicity, resist AZ 1505 was chosen as a suitable resist material for the bonding process due to its much-reduced viscosity at a fairly low glass transition temperature (softening temperature).

[00106] High resolution transmission electron microscopy (HRTEM) and Energy’ dispersive X-ray spectroscopy (EDS)

[00107] To obtain an atomic level understanding of the bonding interface, HRTEM was performed on a cross-sectional sample from a diamond-sapphire heterostructure. The sapphire substrate was C-axis (0001) from University Wafer. To start, a 200 nm-thick gold mask was deposited on the surface to protect the diamond membrane from being damaged by the Ga ion beam. Using a Zeiss NVision 40 system, a cross-sectional TEM specimen with thickness of a few tens of nanometers was prepared by a standard FIB lift-out procedure. The HRTEM image was obtained by a FEI Titan operated at 200 kV, which was equipped with an aberration corrector and a chromatic corrector. The scanning transmission electron microscope (STEM) image was acquired by using high angle annular dark field (HAADF) detector. A FEI Talos S/TEM equipped with a Super X energy-dispersive spectrometer (EDS) was employed for STEM-EDS elemental mapping. The results confirmed the presence of carbon at the position of the diamond membrane and the presence of oxygen and aluminum at the position of the sapphire target substrate.

[00108] X-ray photoelectron spectroscopy (XPS)

[00109] The experimental samples were identical to those employed in the rest of the demonstrations: diamond, fused silica, and sapphire. A set of these was left unprocessed to be used as a reference, whereas the other set received surface activation via oxygen plasma ashing ~ 90 min before loading into the XPS chamber. Two different incidence angles were taken for the XPS analysis to confirm that the perceived effect was related to near-surface species. One set of characterization was taken at 0° incidence, with another set taken at 35° incidence. No significant differences were found for both datasets; as such, only the 0° incidence dataset is shown. In agreement with the contact angle measurements, in that timeframe, the surface was known to have degraded somewhat; however, it remained ‘bondready’. As such, all the quantitative XPS analysis provides a lower bound on the surface activation related species.

[00110] An Al-Ka source in a Thermo Scientific ESCALAB 250Xi was used to perform the XPS characterization at both normal and 35° incidence angles. Elemental peaks were taken with a pass energy of 50 eV, 50 ms dwell time, and a step size of 0. 1 eV, whereas the C KLL peak had 100 eV pass energy and 0.5 eV step size. For all scans, a charge compensating electron flood gun was used, whereas the X-ray spot size was maintained at ~ 200 pm in spread. C Is, C KLL, A12p, Si 2p, O Is, N Is peaks were all collected in high-resolution mode and the presence of other unplanned contaminants was verified via survey scans. All scans were references to the C sp³ peak at 284.8 eV binding energy (BE). Elemental analysis was taken by quantitative comparison of the high-resolution component fits (with appropriate sensitivity factors accounted for). The C Is KLL signature was used to extrapolate and corroborate sp²-to-sp³ ratio obtained from the C Is line-fit by calculating and fitting the peak- to-peak separation of the first derivative of the KLL signal. All peak fitting parameters were taken from commonly agreed on values for equivalent materials and systems from the NIST XPS database. XPS high-resolution spectra of relevant lines with fitting parameters were obtained for Unprocessed-Diamond, Ashed-Di amend, Unprocessed-Sapphire, Ashed- Sapphire, Unprocessed-Silica, and Ashed-Silica, respectively. The analysis is consistent with the conclusions of this Example. However, an unresolved abnormality was observed, specifically the presence of an unidentified third component of the O Is peak in the fused silica samples. This peak is tentatively identified as organics. Looking at the appearance of an intermediate SiCh-Silicate peak in the fused silica Si 2p quantification, it is plausible that these are the oxygen species from low coordination quartz. Regardless, these are in the low at. % range (= 2 at. %) and do not impact any of the conclusions of this Example. Table 2 presents the full quantification summary of the XPS experiments.

[00111] Table 2. High-resolution XPS quantification for all fitted components and peaks. Values in at. %.

[00112] Optical characterizations of GeV and NV centers

[00113] Optical setup

[00114] In this Example, membrane samples were mounted inside a closed-loop cryostation (Montana S200) and cooled down to 4K for low temperature measurements. The position of the membrane was controlled by three closed-loop piezo micropositioners (Attocube ANC 350). Light beams were navigated by a fast-steering mirror (Newport FSM- 300). For photoluminescence (PL) measurements, either a 519 nm green diode (Thorlabs LP520-SF15) or a 532 nm continuous wave (CW) laser (Lighthouse Photonics Sprout-G) was used as the excitation source. For photoluminescence excitation (PLE) measurements of GeV-, the excitation laser was generated by a wave mixing module (AdvR Inc.) combining a tunable CW Ti: Sapphire laser (M Squared Solstis) and a monochromatic CW laser (Thorlabs, SFL 1550P). A single photon counting module (SPCM) (Excelitas Technologies) was applied to plot PL maps, while a spectrometer (Princeton Instruments, SpectraPro HRS) was used to measure the spectra of the color centers. Two bandpass filters (Semrock FF01-615/24-25, Semrock FFO 1 -600/14-25) were combined for GeV PL measurements, and bandpass filters (2x Semrock FF01-647/57-25) were used for PLE measurements. For NV measurements, a single long pass filter (Semrock LP02-561RE-25) was used to capture both NV° and NV“ signals.

[00115] Effect of plasma treatments on the optical coherence of GeV centers

[00116] As mentioned above, plasma treatment of diamond membranes is not necessarily required to achieve bonding. In this portion of the Example, the effect of the plasma on the PL and PLE properties of GeV centers was investigated. Two membranes, each having ~ 200 nm thickness and containing 40 nm-deep implanted GeV centers from the top surface, were transferred onto a single thermal oxide silicon wafer. Membrane 1 received no plasma treatment prior to the bonding, while membrane 2 received a strong plasma ashing treatment as described above. The 4K PL map of the GeV centers on membranes 1 and 2 clearly indicated a signal-to-background improvement via plasma treatments. The average background dropped from ~ 7000 to ~ 1900, leading to an improvement of signal -to- background ratio from ~ 4.5 to ~ 11. The slightly lower signal for the plasma treated diamond membrane indicates a slight oxygen termination which shifts the Fermi level away from the optimal value for GeV centers. A plot of the single (2.5 min average) ZPL linewidths with resonant excitation was obtained. No statistical difference of the linewidth distribution was observed, with mean single scan linewidth of 97 MHz (85 MHz) and mean average scan linewidth of 212 MHz (196 MHz) for membrane 1 (2). The measured line widths may be broader than the real value due to the resolution limit of the wavelength meter (High Finesse WS6-600, 20 MHz measurement resolution, 500 MHz wavelength accuracy).

[00117] Strain characterization via GeV centers

[00118] Group IV centers in diamond are good sensors for local strain environment due to their relatively large strain susceptibilities. The strain magnitude can be estimated via the relative shift of the wavelength and the increased ground state splitting

2 ^E_(JX + ^eE_gy + ( ~ ) ■ Spectra were recorded from 69 (52) GeV centers in diamond membranes direct-bonded to fused silica (thermal oxide silicon) earner wafers. The ZPL wavelength and ground state splitting distribution were obtained plotted. Since implantation- induced strain has a large span which greatly affects GeV“’s ZPL and ground state splitting statistics, the color centers with ZPL from 601.5 nm to 603.2 nm and ground state splitting < 800 GHz were used to estimate the strain from bonded membrane crystals. This region covered ~ 80% of the data points.

[00119] For membranes bonded to fused silica (thermal oxide silicon) wafers, the average ZPL wavelength of GeV centers was 602.68(20) nm (602.53(8) nm), with the average ground state splitting to be 307(158) GHz (224(75) GHz). These ZPL wavelength distributions are comparable with those obtained in bulk diamonds. A slight positive strain was observed with diamond-fused silica heterostructures, which could be explained by the lower thermal expansion ratio of fused silica. Thermally induced negative strain was barely visible for diamond membrane-thermal oxide silicon substrates, which may be due to the fact that membranes wi th such a high aspect ratio (> 1000) could deform instead of generating negative strain under compressive stress. The average strain level of diamond membranes was estimated to be ~ 2.9 x 10 ⁴ (~ -1.7 x 1 () ⁴) on fused silica (thermal oxide silicon) carrier wafers.

[00120] NV centers at 4K

[00121] Thanks to the low optical background of the direct bonding method, individual NV centers were resolvable in diamond membrane heterostructures. Typical NV PL map taken at 4K showed a signal-to-background ratio of over 1.4. This enables NV sensing applications, as discussed further below. In addition, the charge stability of NV centers is a good indicator of a membrane’s surface termination with respect to various plasma treatments on the diamond bonding interface. Here, the NV spectra was characterized in three bonded membranes. They were picked up from a single mother substrate doped in-situ with ¹⁵N, thus contain the same NV densities. They were bonded to SiCh surfaces with (1) no plasma treatment; (2) O2 descum; and (3) high power plasma ashing on the diamond bonding side. NV spectra and statistical data were obtained. It was observed that with no plasma treatment, the NV center stayed at the neutral charge state due to the hydrogen termination effect of the Ar/H2 annealing process. However, membranes treated with plasma had considerably higher NV7NV⁰ ratio. This ratio was positively correlated with the strength of the O2 ashing process, indicating a better oxygen termination performance which helped maintain the NV center in its negatively charged state. This supports use of the dry O-termination method to desirably engineer the diamond near-surface Fermi level for NV based applications.

[00122] Diamond-based nanophotonic devices were fabricated using bonded diamond membrane heterostructures prepared as described above. The devices are schematically illustrated in FIG. 9A and include a TiCh-based device formed on a diamond membrane bonded to fused silica (top) and a diamond-based device in which the diamond membrane was bonded to thermal oxide silicon (bottom). The results are summarized below.

[00123] NV centers in diamond membranes bonded to fused silica covershps (fabricated as described above) were also characterized, including after chemical functionalization of the diamond membranes with labeled biomolecules. The results are summarized below.

[00124] A flow channel device was fabricated using a bonded diamond membrane heterostructure prepared as described above (diamond membrane bonded to a fused silica coverslip). The device is schematically illustrated in FIG. 9B, which also shows use of the device to image a cell illuminated by total internal reflection through the diamond membrane. This device and its use are further described below.

[00125] Results and Discussion

[00126] This Example demonstrates surface plasma activation-based synthesis of diamond heterostructures in which diamond membranes are directly bonded to technologically relevant materials, including silicon, fused silica, thermal oxide silicon, sapphire, and lithium niobate (LiNbCh), with the capability of pre-existing on-chip structures. The fabrication process begins with membrane synthesis via smart-cut, followed by homoepitaxial diamond overgrowth and ex situ or in situ color center formation. Substrates are then patterned to define individual membrane shapes via either photo- or electron beam lithography. Target membranes are undercut by selectively removing sp² carbon via electrochemical (EC) etching, leaving a small tether attached to the diamond substrate for deterministic manipulation. Membrane sizes were limited to 200 pm by 200 pm squares. However, larger and more intricate membrane shapes can be generated by extending the EC etching time and slightly modifying the patterning step of the process flow.

[00127] Following EC etching, templated area-controlled polydimethylsiloxane (PDMS) stamps were utilized to transfer and manipulate membranes with improved process yield and scalability. The PDMS stamps had two different patterns, allowing for smaller (PDMS1- stamp) and larger (PDMS2-stamp) contact areas, and by extension, adhesion strength. The PDMS 1 -stamp was used to break the diamond tether and pick up the membrane, whereas the PDMS2-stamp was used for flipping the diamond membrane from the PDMS 1 -stamp and subsequent placement. In both cases, the prominence of the adhesion region, which was 50 pm taller than the rest of the stamp, ensured only the targeted membrane was contacted. This method enabled multiple membrane transfers following EC etching, which can ultimately be automated into a single step for the entire diamond substrate.

[00128] Next, the underlying diamond layer that was damaged by He⁺ implantation was removed. This improved the overall crystallographic quality and fully decoupled the final membranes, which were isotopically purified with controlled doping, from the low-cost type- Ila diamond substrate. This thinning was performed via inductively coupled plasma (ICP) reactive ion etching (RIE). To protect the final bonded substrate from being etched, the membrane was thinned by placing it on an intermediate fused silica carrier wafer. Intermediate wafers were coated with photo- (AZ 1505) or electron beam resist (PMMA), which soften in the temperature range from 100 °C to 130 °C with reduced viscosity at subsequent stages. This additional step flipped the membrane again so the growth side was facing up (exposed) on the target substrate, which eliminated growth side morphology constraints for bonding and enabled precise depth control for near-surface and 5-doped color centers. To prevent the resist from overheating and crosslinking, a multi-cycle etching recipe with short plasma duration of < 15 s per cycle was developed. Using this methodology, precise thickness control from 10 nm to 500 nm was realized. The maximum thickness was determined by the homoepitaxial overgrowth step and can be modified to meet application needs.

[00129] Downstream O2 plasma ashing was used for surface activation on both the diamond membrane and target substrate to enable subsequent bonding. The target substrates were subjected to a high-power ashing recipe (gas flow 200 seem, RF power 600 W for 150 s) with extended process duration for inert substrates such as sapphire and LiNbOs. The diamond membranes received either this high-power recipe or an O2 descum clean (gas flow 100 seem, RF power 200 W for 25 s), which did not etch or roughen the diamond surface. The downstream O2 plasma cleaned and oxygen-terminated the membrane and carrier material surfaces without using or requiring any wet processing. To prevent functionalization degradation at elevated temperatures, all ashing recipes were performed at room temperature.

[00130] Next, the membrane was bonded to the target substrate. The patterned intermediate wafer was mounted onto a micropositioner-controlled glass slide via a flat, chip size PDMS stamp. The target substrate was vacuum secured on a temperature-controlled stage. Leveraging optical access through the transparent intermediate wafer for alignment, the membrane was moved to the target location and it was brought into contact with the target substrate, which coincided with the appearance of membrane-scale interference fringes/pattems. Using this method, an alignment precision of 30 pm and 0.1° was achieved. The heterostructure was subsequently heated by elevating the temperature of the stage through multiple steps. After reaching the resist softening point, the intermediate wafer was slid away, leaving the bonded structure behind. Future utilization of dedicated wafer-bonding equipment will significantly improve the precision and tolerance of all transfer steps.

[00131] Finally, to ensure a robust, covalently bonded interface between the membrane and the target wafer, the heterostructure was annealed at 550 °C under argon forming gas atmosphere to minimize undesired oxidation. This annealing also removed the polymethyl methacrylate (PMMA) residue and left a clean direct-bonded membrane as the final product (for PMMA-based transfer). In some cases, the diamond membrane was bonded to fused silica having a trench patterned therein, emphasizing the capability of bonding membranes to structured materials. The overall process yield stands above 95%, limited only if the plasma ashing chamber conditions are unstable and the approach angle of the transfer station is under limited control. However, both can be readily improved by transitioning to process specific tooling.

[00132] In-depth material characterization revealed the preserved diamond quality throughout the bonding process. Atomic force microscopy (AFM) was utilized to characterize the membrane surface morphology, which is critical for successful plasma-activated bonding, as well as for the coherence and stability of near-surface color centers. As shown in FIG. 8A, both small and large area scanning results returned atomically flat surface profiles with R_q < 0.3 nm. Additionally, target substrates were characterized via AFM to ensure sub-nm roughness post plasma treatments as detailed in Table 1, above.

[00133] Beyond local height variation, the bonded membrane showed a general flatness of ~ 1 nm, as characterized via profilometry. The thickness profile of the membrane, shown in FIG. 8C, revealed a uniform height of 493.7 ± 1.1 nm, with the standard deviation less than the instrument resolution (1.5 nm) for large scale scanning. Beyond line scans, a two- dimensional flatness map of the membrane was studied via confocal laser scanning microscopy as described above. The effectiveness of the plasma surface activation was also characterized by tracking the change in surface hydrophilicity of the bonding interfaces via contact angle measurements, as shown in FIG. 8B. For the diamond surface, the contact angle reduced from 52.4° ± 0.7° to 6.1° post high-power plasma treatment, indicating a considerable increase in hydrophilicity. This was confirmed via quantitative X-ray photoelectron spectroscopy (XPS) characterization of surface species also shown in FIG. 8B. It was noted that surface hydrophilicity was directly correlated with a reduction of surface amorphous carbon sp² bonds as quantified by the more reliable D-parameter extrapolation of the C KLL line. Furthermore, the decrease from the raw C is quantification (likely an increase of ether-like terminations) and an increase of surface available sapphire-0 bonds indicated an effective surface preparation and oxygen termination to both surfaces. Similar behavior was confirmed on all target bonding materials with observed contact angles below 20° post treatments.

[00134] The quality of the bonded diamond heterostructure was directly studied via high resolution transmission electron microscopy (HRTEM). FIGS. 8D-8E show an ICP -thinned (from ~ 309 nm to 10 ± 0.3 nm) diamond membrane bonded to a sapphire substrate. The thinness and uniformity reflected the high level of process control and allowed single field of view characterization of both diamond membrane interfaces. The HRTEM image revealed several critical features. Firstly, the membrane exhibited uniform crystallinity and morphology throughout its thickness. Secondly, a sharp, sub-0.5 nm interface was observed between the crystalline diamond and sapphire. Thirdly, there was a repeating atomic arrangement throughout the interface profile, evidence of a covalently cross-linked interface. Energy Dispersive X-ray Spectroscopy (EDS) analysis (see bottom of FIG. 8E) of the various elements associated with the intersection (C, Al, O) placed an upper limit on the bonding interface to be less than 2 nm. It was noted that the EDS analysis artificially broadened the interface as a result of the slight angular mismatch between the electron beam depth projection and the actual physical interface.

[00135] Furthermore, the optical properties of color centers in the bonded membranes were characterized. Confocal imaging revealed that germanium vacancies (GeV ) within the bonded membranes had high signal-to-background and sufficient optical coherence for applications in quantum technologies. Typical photoluminescence (PL) maps of individual GeV centers hosted in a membrane bonded to a distributed Bragg reflector (DBR) mirror showed that the signal-to-background ratio of GeV can be as large as ~ 65 with an average value of ~ 40, a significant improvement from any suspended HSQ-based membranes (5 to 30) and bulk diamond (—25). The bonding process also preserved GeV centers’ optical coherence and introduced minimal strain. [00136] This Example further demonstrated the suitability of the fabricated bonded diamond membrane heterostructures as a platform for quantum technologies. First, nanophotonic integration was explored which improves qubit addressability and is broadly utilized in quantum photonics. Photonic integration is typically achieved by patterning diamond into undercut, suspended structures, creating geometrical constraints that complicate further multiplexing and integration with on-chip single-photon detectors, electronics, or other devices that could otherwise enhance quantum network functionality. This Example demonstrated that the bonded diamond membrane heterostructures enabled multiple approaches to photonic integration.

[00137] Briefly, templated atomic layer deposition (ALD) of TiCh was used to create nanophotonic devices on the surface of a 50 nm-thick diamond membrane bonded to a fused silica substrate. The schematic is shown as the upper image of FIG. 9A, with excitation and collection ports (grating couplers) labeled. In these devices, the optical mode effectively hybridizes between the TiCh and diamond. Microscope images of TiCh fishbone cavities and ring resonators were taken at the same location but in different fabrication rounds, highlighting the robustness of the bonded diamond membranes to cleanroom processing and the recyclability of photonic integration.

[00138] Transmission measurements of the cavities also revealed significant improvement of quality factors. Transmission spectra of a typical fishbone cavity with target wavelength 737nm (the wavelength of SiV emission) were obtained. The highest measured quality factor Q was 10,640, with a three-device average of 10,150±350. These values are 2.5 times higher than existing demonstrations, resulting from elimination of intermediate bonding layers and improved diamond crystal quality. Based on updated Q factors, a maximum Purcell enhancement factor of 270 in the diamond is estimated. These metrics are suitable for state- of-the-art experiments in cavity quantum electrodynamics. Similarly, ring resonators measured through the drop port exhibited quality factors as QTM = 16319 (QTE = 12620) for transverse magnetic, TM (transverse electric, TE) modes.

[00139] To explore the robustness of the bonded diamond membrane heterostructure platform, nanophotonic ring resonators were etched directly into diamond using a lithographically defined hard mask and RIE. The schematic is shown in the lower image of FIG. 9A. Bright and dark field images of the devices demonstrated the high quality and uniformity of the fabrication process. These ring resonators also exhibited quality factors Q of 21,883, with large field confinement within the diamond. The compatibility' and versatility of these high aspect ratio structures makes them ideal for integrating with other materials and thus paves the path for hybrid quantum technologies.

[00140] Diamond heterostructures also have distinct advantages in quantum metrology such as nanoscale magnetic field, electric field, and temperature sensing. In these applications, a diamond sensor is ty pically brought in close proximity of a sensing target. The most sensitive diamond sensors rely on high-purity single-crystals with the sensing target bound on the diamond’s top surface. The large thickness and refractive index of conventional bulk diamond requires optical initialization and readout of the sensing qubits from the top surface. As a consequence of this geometry, the target systems need to possess optical transparency, low auto-fluorescence, and high photo-stability, which are significant restrictions for the study of biological systems. The present bonded diamond membrane heterostructures overcome these challenges by enabling optical addressability through the back of the diamond membrane, without the need for passing through the top surface and the sensing target.

[00141] First, the stability and addressability of individual nitrogen vacancy (NV ) centers (i.e., a qubit sensor) in the bonded diamond membranes was investigated. Widefield and confocal images show individually resolvable photostable NV centers in a diamond membrane bonded to a fused silica coverslip. These emitters were confirmed to be NV centers through the presence of their characteristic 2.87 GHz zero-field splitting by optically detected magnetic resonance (ODMR) spectroscopy. NV centers in membranes are known to exhibit excellent spin coherence. Next, the diamond surface of the bonded diamond membrane heterostructure was chemically functionalized. Using back illumination, it was shown that the fluorescence of individual Alexa 488-labeled streptavidin molecules and streptavidin-conjugated Qdot-525 quantum dots could be detected. This enables imaging the positions of not only NV centers but also target proteins bound to the diamond membrane’s surface. The ability to fluorescently detect the position of individual NV centers and proteins is important for NV-based single-molecule nuclear magnetic resonance and electron paramagnetic resonance spectroscopy, as this will allow for the efficient identification of NV centers that have a desired molecular target within the sensing range.

[00142] The bonded diamond membrane heterostructures were also used with TIRF- microscopy to demonstrate imaging at a reduced level of background luminescence. A schematic representation of a bonded diamond membrane integrated in a flow channel with macrophage-like RAW cells grown on the diamond surface is shown in FIG. 9B. Optical excitation above the critical angle ensured that only a small section above the diamond membrane was excited by the optical field. Staining the toll-like receptor 2 (TLR2) with Alexa488-labeled anti-TLR2 antibody revealed in TIRF imaging the location of individual proteins distributed across cell surface. This is in stark contrast with the epiluminescence mode where background luminescence prevents the imaging of individual molecules. It is noted that the larger index of refraction (n = 2.4) of diamond results in an evanescent field that falls 1.6-times faster off compared with a conventional glass microscope coverslip (n = 1.5). Likewise, the sedimentation of living Escherichia coli bacteria on the diamond membrane which had been introduced via the flow channel was also observed in real time. Experiments enabled by the flow channel demonstrate remarkable flexibility in interfacing target samples with quantum diamond sensors, a significant challenge to fulfill using existing techniques.

[00143] Conclusions

[00144] This Example has demonstrated a complete process flow to create diamond-based heterogenous materials and technologies. The bonded diamond membrane heterostructures combine isotopic engineering, in-situ doping, and precise thickness control, while maintaining the surface morphology, flatness, and crystal qualify necessary for quantum technologies. Bonded, continuous crystalline diamond films as thin as 10 nm were fabricated, which is well below previous demonstrations and comparable to material geometries in state- of-the-art microelectronics. HRTEM results revealed ordered, sub-nanometer bond interfaces while PL measurements demonstrated high signal-to-background ratio for all hosted color centers. The process is compatible with nanostructured substrates, has a compact footprint and requires no post-bond etching, ensuring the integrity of pre-existing target substrate structures. Bonded diamond membranes are robust to multiple subsequent nanofabrication steps, and the process is compatible with standard semiconductor manufacturing techniques including wafer-bonding. Crucially, by avoiding intermediary adhesion materials, optimal material heterostructures are generated for applications in quantum photonics and quantum biosensing. Technological suitability for quantum photonics is demonstrated via the integration of high-quality factor nanophotonics by either TiCh deposition or direct diamond patterning and etching. These diamond-based heterostructures, with minimal optical loss, are ideal candidates for on-chip nanophotonic integration and spin-photon coupling devices. Furthermore, it was demonstrated that diamond membrane bonding unlocks novel experimental possibilities for quantum biosensmg and imaging by integrating flow channels with diamond membranes. The simultaneous resolution of fluorescent molecules and NV centers enables accurate identification of proximal NV sensors for desired sensing targets. The ultrathin diamond membranes also allow for TIRF illumination which strongly improves the signal contrast of local sensing targets while minimizing undesired laser excitation.

[00145] The disclosed manufacturing process opens up a broad range of heterogeneous diamond-based platforms for quantum technologies. The integration of diamond with electro- optical and piezoelectric materials such as LiNbCh enables on-chip, electrically- reconfigurable nonlinear quantum photonics and allows for studies of quantum spin-phonon interactions. The diamond bonding disclosed herein unlocks additional coupling possibilities with other solid-state qubits, magnonic hybrid systems, and superconducting platforms. Furthermore, combining the bonded diamond membranes with established techniques for the creation of highly coherent near-surface NV centers will achieve ultra-sensitive diamondprobes optimized for the study of molecular binding assays, two dimensional dichalcogenides (TMD), and thin-film magnetic materials. Lastly, with high thermal conductivity, large bandgap and high critical electric field, bonded diamond membranes have a myriad of applications in high power electronics.

[00146] The word "illustrative" is used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as "illustrative" is not necessarily to be construed as preferred or advantageous over other aspects or designs. Further, for the purposes of this disclosure and unless otherwise specified, "a" or "an" means "one or more.”

[00147] The foregoing description of illustrative embodiments of the disclosure has been presented for purposes of illustration and of description. It is not intended to be exhaustive or to limit the disclosure to the precise form disclosed, and modifications and variations are possible in light of the above teachings or may be acquired from practice of the disclosure. The embodiments were chosen and described in order to explain the principles of the disclosure and as practical applications of the disclosure to enable one skilled in the art to utilize the disclosure in various embodiments and with various modifications as suited to the particular use contemplated. It is intended that the scope of the disclosure be defined by the claims appended hereto and their equivalents. [00148] If not already included, all numeric values of parameters in the present disclosure are proceeded by the term “about” which means approximately. This encompasses those variations inherent to the measurement of the relevant parameter as understood by those of ordinary skill in the art. This also encompasses the exact value of the disclosed numeric value and values that round to the disclosed numeric value.

Claims

WHAT IS CLAIMED IS:

1. A method of forming a bonded diamond membrane heterostructure, the method comprising:

(a) subjecting a surface of a target substrate to plasma ashing to provide a plasma treated target substrate having a plasma treated surface; and

(b) contacting the plasma treated surface of the plasma treated target substrate with a surface of a diamond membrane to form a bonded diamond membrane heterostructure comprising the target substrate bound via covalent bonds to the diamond membrane at a bonding interface formed between the plasma treated surface of the plasma treated target substrate and the surface of the diamond membrane.

2. The method of claim 1, further comprising subjecting the surface of the diamond membrane to plasma ashing to provide a plasma treated surface of the diamond membrane prior to step (b).

3. The method of claim 1, wherein the plasma ashing is earned out using an O2 plasma.

4. The method of claim 1, wherein the plasma ashing provides the plasma treated surface of the plasma treated target substrate with oxygen termination.

5. The method of claim 1, wherein the surface of the diamond membrane is an untreated surface.

6. The method of claim 1, wherein the diamond membrane is single-crystalline and the surface of the diamond membrane is (100).

7. The method of claim 1, wherein the diamond membrane has a thickness of no more than 500 nm.

8. The method of claim 1, wherein the diamond membrane contacted with the target substrate in step (b) is provided on an intermediate substrate, wherein the diamond membrane is adhered to the intermediate substrate via a layer of a photoresist directly between and in contact with the diamond membrane and the intermediate substrate.

9. The method of claim 8, wherein the photoresist is a positive photoresist.

10. The method of claim 8, wherein the photoresist is characterized by a Tg of no greater than 200 °C and over a range of no more than 20 °C.

11. The method of claim 8, wherein the positive photoresist comprises a cresol novolak resin or polymethylmethacrylate.

12. The method of claim 1, wherein the target substrate is fused silica, thermal oxide silicon, sapphire, lithium niobate, silicon, or yttrium iron garnet.

13. The method of claim 1, wherein step (b) comprises heating via a first heating stage and a second heating stage.

14. The method of claim 13, wherein the first heating stage comprises heating to an intermediate temperature selected to soften a layer of a photoresist in contact with the diamond membrane; and further wherein the second heating stage comprises heating to a final temperature greater than the intermediate temperature and under a non-oxidizing atmosphere.

15. The method of claim 14, wherein the final temperature is at least 500 °C.

16. The method of claim 1, wherein the bonding interface is crystalline across its thickness as measured using high resolution transmission electron microscopy (HRTEM).

17. The method of claim 16, wherein the bonding interface has a thickness of no more than 0.5 nm as measured using HRTEM.

18. The method of claim 1, wherein the diamond membrane is provided on an intermediate substrate, wherein the diamond membrane is adhered to the intermediate substrate via a layer of a photoresist directly between and in contact with the diamond membrane and the intermediate substrate, and further wherein the method comprises subjecting the surface of the diamond membrane to plasma ashing to provide a plasma treated surface of the diamond membrane prior to step (b).

19. A bonded diamond membrane heterostructure comprising a plasma treated target substrate having a plasma treated surface and a diamond membrane having a surface, the plasma treated target substrate bound via covalent bonds to the diamond membrane at a bonding interface formed between the plasma treated surface of the plasma treated target substrate and the surface of the diamond membrane, wherein the bonding interface is crystalline across its thickness as measured using HRTEM.

20. The bonded diamond membrane heterostructure of claim 19, wherein the bonding interface has a thickness of no more than 0.5 nm as measured using HRTEM.

21. The bonded diamond membrane heterostructure of claim 19, wherein the surface of the diamond membrane is a plasma treated surface.

22. The bonded diamond membrane heterostructure of claim 19, wherein the diamond membrane is single-crystalline and the surface of the diamond membrane is (100).

23. The bonded diamond membrane heterostructure of claim 19, wherein the diamond membrane has a thickness of no more than 500 nm.

24. The bonded diamond membrane heterostructure of claim 19, wherein the plasma treated target substrate is fused silica, thermal oxide silicon, sapphire, lithium niobate, silicon, or yttrium iron garnet.