Electrical-double-layer-force-enabled transfer of van der waals materials

Abstract

In a method for transferring van der Waals materials, a van der Waals material is grown on or transferred onto a source substrate. A support layer is adhered to the van der Waals material. The source substrate with the van der Waals material is then immersed in a weak base solution, producing an electrical double layer at and between surfaces of (a) the van der Waals material and (b) the source substrate, wherein the electrical double layers generate a repulsive force that detaches the van der Waals material from the source substrate. The support layer with the adhered van der Waals material is separated from the source substrate. The support layer with the adhered van der Waals material is then applied to a target substrate. The support layer is then removed from the van der Waals material, leaving the van der Waals material adhered to the target substrate.

Description

[0001] Attorney Docket No. mit-25634pct

[0002] ELECTRICAL-DOUBLE-LAYER-FORCE-ENABLED TRANSFER OF VAN DER WAALS MATERIALS

[0003] GOVERNMENT SUPPORT

[0004] This invention was made with government support W911NF-23-2-0057 and W911NF-22-1-0023 awarded by the US Army Research Office. The US Government has certain rights in the invention.

[0005] BACKGROUND

[0006] The discussion of the background state of the art, below, may reflect hindsight gained from the disclosed invention(s); and these characterizations are not necessarily admitted to be prior art.

[0007] With the scaling of silicon-based complementary metal-oxide-semiconductor (CMOS) devices approaching its fundamental limit, tremendous efforts have been devoted to searching for breakthroughs in new materials and integration strategies. A complementary metal-oxide-semiconductor (CMOS) is a type of semiconductor technology that uses both N-type and P-type metal-oxide-semiconductor field-effect transistors (MOSFETs), which are voltage-controlled devices, to implement logic gates and other digital circuits. CMOS circuits use complementary pairs of N-type and P-Type transistors, where one type is used to represent a logic 'o' and the other type represents a logic '1'. In a CMOS logic gate, when the N-type transistor is on (conducting), the P- type transistor is off (non-conducting), and vice versa. This complementary operation ensures that at any given time, only one type of transistor is conducting, minimizing power consumption. CMOS technology offers several advantages over other semiconductor technologies, including low power consumption (CMOS circuits consume very little power when in a static state, as there is no direct path between the power supply and ground); high noise immunity (CMOS circuits have high noise immunity due to the complementary nature of the transistors, making them less susceptible to noise and interference); scalability (CMOS technology can be scaled down to smaller feature sizes, enabling higher integration densities and faster switching speeds); and versatility (CMOS can be used to implement a wide range of digital and analog circuits, as well as mixed-signal systems). CMOS technology is widely used in various electronic devices, including microprocessors, memory chips, digital signal processors, and system-on-chip (SoC) designs.

[0008] It has become the dominant semiconductor technology due to its low power consumption, high integration density, and scalability, enabling the development of increasingly complex and power-efficient electronic systems.

[0009] Van der Waals (vdW) materials (i.e., materials, such as graphene and metal dichalcogenides) composed of “two-dimensional” layers, which strongly bonded within Attorney Docket No. mit-25634pct each layer but held together weakly between layers by van der Waals forces, have aroused immense interest for next-generation CMOS technology with their excellent electronic and optoelectronic properties at an atomically thin limit. Their potentials lie in both front-end-of-line (FEOL) applications as ultra-scaled channel materials and back-end-of-line (BEOL) applications as building blocks for 3D monolithic integration with Si CMOS platform. However, high-quality vdW materials are often not grown on the target substrates directly. On the one hand, the harsh growth conditions, such as high temperatures (>700°C), reactive vapor, CMOS-incompatible promotors (such as NaCl), may create damage to the underlying substrates and circuits. On the other hand, lattice-matched substrates are advantageous for epitaxial growth. All of these involve post-growth layer transfer, i.e., removing the vdW materials from their source substrates and transferring them onto a target substrate.

[0010] During layer transfer, the vdW materials have been separated from the source substrate by chemical etching of the source substrate, electrochemical bubbling, or metal-assisted transfer, etc.. Despite all of the progress that has been made, from a practical application point of view, the transferred results and processes are still far from satisfying. Most of the processes are not CMOS-compatible (e.g., there could be K+or Na+contaminations in the chemical etching method), and there are still noticeable damages, especially for wafer scale transfers. Transfer via electrochemical bubbling uses conducting substrates and introduces strain and damage. Metal-assisted dry transfer also suffers from low yield, high cost, and CMOS-incompatible contamination issues. Wedging transfer and surface-energy-assisted transfer only utilize water to peel off vdW materials; but they were found unreliable, and the range of applications is narrow. These limitations have made the transfer-free, low-temperature synthesis processes much more appealing, but high-temperature growth processes are still beneficial for higher-quality materials. Until now, a reliable CMOS-compatible transfer technique remained to be developed.

[0011] Another objective of vdW materials transfer is cleanliness. An atomically clean interface not only boosts the device performances of transistors and light-emitting diodes but also enables the construction of vdW heterostructures with designed properties (e.g., twistronics and interlayer exciton electronics). The clean transfer of large-area vdW materials, however, is still challenging.

[0012] SUMMARY

[0013] Methods for transferring van der Waals materials and devices produced via these methods are described herein, where various embodiments of the methods and devices may include some or all of the elements, features, and steps described below.

[0014] A method for transferring van der Waals materials includes growing or transferring a van der Waals material on or onto a source substrate, and a support layer Attorney Docket No. mit-25634pct is adhered to the van der Waals material. The source substrate is then immersed with the van der Waals material in a weak base solution, producing an electrical double layer at and between surfaces of (a) the van der Waals material and (b) the source substrate, wherein the electrical double layers generate a repulsive force that detaches the van der Waals material from the source substrate. The support layer with the adhered van der Waals material is separated from the source substrate. Then, the support layer is applied with the adhered van der Waals material to a target substrate, and then the support layer is removed from the van der Waals material, leaving the van der Waals material adhered to the target substrate.

[0015] Van der Waals (vdW) materials are a class of materials composed of strongly bonded two-dimensional (2D) layers (e.g., single- or few-atomic-layer structures) that are held together in the third dimension by relatively weak van der Waals forces. Van der Waals materials have a layered structure; they consist of individual atomic planes or layers that are strongly bonded within each layer by covalent or ionic bonds. The weak vdW forces holding the VdW layers together are much weaker than the intralayer bonds. Due to the anisotropic nature of the bonding, the properties of these materials can be highly anisotropic, exhibiting different characteristics along the in-plane and out-of-plane directions. VdW materials can have a wide range of chemical compositions, including elements (e.g., graphite), compounds (e.g., hexagonal boron nitride, transition metal dichalcogenides), and various naturally occurring minerals. VdW materials are of significant interest due to their unique properties and potential applications in electronics, optoelectronics, energy storage, and other fields.

[0016] The weak base is a type of chemical base that, when dissolved in water, does not completely dissociate into ions, resulting in a solution with few hydroxide ions and many undissociated base molecules. When a weak base dissolves in water, it establishes an equilibrium between the undissociated base molecules, the hydroxide ions, and the conjugate acid: B + H20 BH++ OH , where B represents the weak base, and BH+is its conjugate acid. Examples of weak bases that may be used include ammonia (NH3), pyridine (C5H5N), and methylamine (CH3NH2). The dissolution of ammonia in water can be represented as NH3+ H20 NH4++ OH . The OH’ ions in the alkali environment can charge the substrates via chemical reactions or adsorptions, such as deprotonation reactions in the case of metal oxide substrates.

[0017] The integration and stacking of van der Waals (vdW) materials to target substrates or circuits is advantageous to their applications in high-end electronics, optics, moire electronics, etc. Because high-quality vdW materials are often grown under harsh conditions, the transfer of grown vdW materials to target substrates is one of the key steps in the integration. However, the step of detaching nanomaterials from the source substrate typically requires the use of either chemical etchants, Attorney Docket No. mit-25634pct electrochemical bubbling, or metal-assisted mechanical strain, which often leaves residue or introduces contamination, making the process CMOS-incompatible and low- yield; and the substrate becomes nonreusable.

[0018] Here, we present an electrical-double-layer-force-enabled transfer method (termed as EFT or EDL force transfer) that is CMOS-compatible, with significantly reduced damage, and wide applicability, and which is cost-effective. The unique strategy is to leverage the negative zeta potential of both the substrate and the vdW material in concentrated ammonia solution. With the formation of the EDL, the vdW material is immediately repelled from the substrate by the strong EDL repulsion force. This method is widely applicable to various vdW materials (e.g., carbon nanotube, M0S2, h-BN, etc.) and substrates (e.g., oxide, nitride, etc.) with negative zeta potential because EDL formation is universal. The as-transferred vdW materials show ultra-high nanoscale cleanliness despite the use of poly(methyl methacrylate) (PMMA), along with reduced wrinkles, cracks, transfer-induced defects, and metal / polymer contaminations. The M0S2 transistors fabricated with this transfer method show higher on current and reduced threshold-voltage variation than devices made via chemical etching transfer. This EDL-force-enabled layer transfer offers a facile and manufacturing-viable solution for vdW material integration, which will significantly advance the future development of atomically thin electronics.

[0019] The methods and devices described herein can offer a variety of advantages, including (a) being widely applicable to different types of materials and substrates, (b) producing damage-free devices with these van der Waals materials (post-transfer), (c) generating atomically clean interfaces between the target substrate and the vdW material, (d) providing a fast transfer of the vdW material, (e) transferring vdW materials with a high yield (of successful transfers without damage), (f) compatibility with the silicon industry, and (g) offering a unique transfer mechanism based on the electrical-double-layer force.

[0020] BRIEF DESCRIPTION OF THE DRAWINGS

[0021] FIG. 1 is a schematic illustration showing the electrical double layer (EDL) of an isolated plate 12 in a liquid medium 14 with negative zeta potential with an exemplary illustration of ionic charges. Only one side of surface charge is plotted for simplicity. The potential profile is shown schematically in the lower part of FIG. 1.

[0022] FIG. 2 is a schematic illustration showing formation of an EDL and the resulting EDL repulsion force when a vdW material joined with a support layer 16 is immersed in concentrated ammonia 14.

[0023] FIG. 3 is a schematic illustration showing the EDLs of the plate 12 and the vdW material joined with the support layer 16 facing each other, both with negative zeta Attorney Docket No. mit-25634pct potential. Only one side of surface charge is plotted for simplicity. The potential profile is shown schematically in the lower part of FIG. 3.

[0024] FIG. 4 shows the dependence of zeta potential on pH value for different substrates and vdW materials. The central band for each composition shows the isoelectric point. The dotted line shows the pH value of concentrated ammonia (wt% ~32%).

[0025] FIG. 5 is a schematic illustration showing the scope of the application explored in this work, wherein vdW materials, such as carbon nano tubes (CNTs) 18; single or fewlayer structures 20, such as graphene and hexagonal boron nitride; or transition metal dichalcogenides 22, such as M0S2, Wse2, NbSe2, etc., are transferred from a source substrate 12, formed of, for example, an oxide 24, such as Si02, A12O3, CuO, Hf02, NbO5, SrTiO3, LaA103, etc.; a metal 26, such as Cu, Ti, Au, etc.; or a nitride 28, such as GaN, or SiNx. Note that the detachment from insulating substrates is more effective than on metal substrates, as graphene or hexagonal boron nitride (hBN) grown on Cu typically have strong interactions. An oxidation step is utilized, as described in the Methods section, 80% graphene can be transferred by water oxidation and EFT.

[0026] FIG. 6 is a schematic illustration showing the general process flow of an EFT process, wherein vdW materials 32 are detached from source substrates 12 using EDL repulsion force by simply immersing the vdW materials 32 and source substrates 12 into a 32 wt% concentrated (in water) ammonia solution 14. From left-to-right, (a) a source substrate 12 coated with a vdW material 32 is provided; (b) a support layer 30 is coated onto the vdW material; (c) the support layer 30 with the vdW material attached is detached from the source substrate 12 in a 32 wt% ammonia solution by EDL repulsion; (d) the support layer 30 with the vdW material attached to a target substrate 34; (e) this laminate structure is vacuum baked with the layers in intimate contact; and (f) the support layer is removed, leaving the vdW material 32 adhered to the target substrate 34.

[0027] FIG. 7 is a schematic illustration of the key steps of EFT. The detachment of vdW materials and support layers 16 from source substrates 12 is carried out in concentrated ammonia 14 enabled by the strong EDL repulsion force. When using a thermal release tape (TRT) and poly(methyl methacrylate) (PMMA) as the support layer, the release of vdW materials onto a target substrate is a dry step that enables deterministic and scalable transfer.

[0028] FIG. 8 plots the Raman spectrum of graphene / Cu before (as-grown) 36 and after 38 a decoupling treatment, including water immersion for 46 hours. The 2D / G ratio increases from 0.52 (line 36) to 2.03 (line 38) after the decoupling.

[0029] FIG. 9 plots the Raman spectrum of graphene transferred onto a Si02 / Si substrate. Attorney Docket No. mit-25634pct

[0030] FIGS, io and 11 include Raman mappings of graphene after it is transferred onto a Si02 / Si substrate. The plots show the intensity of the 2D peak and G peak, respectively.

[0031] FIG. 12 is a schematic illustration that shows the possible detaching mechanism of a vdW material from the substrate. The gap between the vdW material and the substrate is first enlarged by steric hydration force and further enlarged by EDL repulsion force, which eventually enables the detachment.

[0032] FIG. 13 is a plot that shows the calculated forces between the vdW material and the substrate in concentrated ammonia as a function of the interface gap distance, r. Plotted therein are the absolute value of the attraction force 46 and the total repulsion force 48, which is larger than the attraction force in the range under consideration. The repulsion forces include the contribution from the EDL force 50 and the steric hydration force 52. The parameters used for the plot here are: f0=101mJ / m2, r0= 0.165 nm, C = 30 mJ / m2,H= 0.236 nm, LD = 1.37 nm, i / q = 60 mV, ip2= 40 mV.

[0033] FIG. 14 plots the expansion length of the Si02gel pieces as a function of time. The o-second point is taken at the point when the Si02gel pieces get in touch with the solution. Measurements with the ammonia solution are plotted with circles, while measurements with the NH4C1 and KOH solutions (in water) are plotted respectively with upward-facing and downward-facing triangles

[0034] FIG. 15 is a plot evidencing the dependence of the detaching rate on ionic strength (experimentally measured). The ionic strength is tuned by adding NaCl into the concentrated ammonia solution.

[0035] FIG. 16 is a plot of the detaching rate of MoS2from Si02as a function of the ammonia solution concentration. Only a PMMA layer is used in the transfer without TRT to quantify the detaching rate. The top plots 54 represent detachment in pure ammonia. The middle dots and bottom dots represent the addition of NaCl or NH4C1, respectively, into the ammonia solution.

[0036] FIG. 17 is a plot evidencing the dependence of EDL force 50 and Debye length 60 on the ionic strength (calculated). The parameters used for the plot here are as follows: LD = 1.37 nm, i / q = 60 mV, ip2= 40 mV, and r = 3 nm.

[0037] FIG. 18 plots the dependence of the EFT detaching rate on the solution pH value (experimentally measured). The pH of the solution is tuned by adding HC1 into the concentrated ammonia.

[0038] FIG. 19 plots the dependence of the EDL repulsion force on pH (calculated). The parameters used for the plot here are as follows: r = 3 nm and I = 0.02 mol / L;. i / q and >2at different pHs are taken from literature.

[0039] FIG. 20 includes bar charts showing the surface Si02thickness loss from ellipsometry measurement and roughness gain from AFM characterization of the Attorney Docket No. mit-25634pct

[0040] Si02 / Si substrates after immersion in ammonia and KOH solutions (for 30 minutes each). The error bars come from multiple ellipsometry measurements of the same piece of substrate.

[0041] FIG. 21, in the top panels, shows M0S2 grown (with a first growth at left and a second growth at right) on a pristine Si02 / Si substrate and on the same reused substrate that has gone through chemical-vapor-deposition (CVD) growth and EFT transfer once; the scale bar represents 10 pm. The bottom panels are SEM images of carbon nano tubes (CNTs) grown (again with a first growth at left and a second growth at right) on a pristine single-crystalline quartz substrate and on the same reused substrate that has gone through CVD growth and EFT once; the scale bar represents 20 pm. The aligned CNT ratio for the first growth was 99.3%, and the aligned CNT ratio of 99.1% for the second growth.

[0042] FIGS. 22-24 are typical AFM images of bare (pristine) Si02 / Si (FIG. 22), wherein Ra= 124.9 Pm; after immersion in an ammonia solution (in water) for 30 minutes (FIG. 23), wherein Ra= 125.7 Pm; and after immersion in a KOH solution for 30 minutes (FIG. 24), and wherein Ra= 123.9. The roughness for each FIGURE is 124.9 pm, 125.7 pm, and 233.9 Pm, respectively. A much rougher surface is observed in the KOH case.

[0043] FIGS. 25-27 are plots of typical ellipsometry measurements (reflectance as a function of wavelength) of Si02 / Si after immersions in the ammonia solution 54 and in the KOH solution 58. Both calculated values 62 and measured values 64 are plotted. The Si02thickness is greatly reduced in the KOH case. Specifically, the Si02thickness is 306.51 nm for the values in FIG. 25, 306.22 nm for the values in FIG. 26, and 209.41 nm for the values in FIG. 27.

[0044] FIG. 28 plots the etching rate of Si02as a function of ammonia concentration 54 and KOH concentration 58. The etching rate for KOH is found from literature, while the detachment rate with ammonia was measured by us.

[0045] FIG. 29 plots the detachment rate of vdW materials from substrates as a function of ammonia concentration 54 and KOH concentration 58.

[0046] FIG. 30 plots the change of Si02thickness after being immersed in concentrated ammonia for an elongated time. The thickness is measured using ellipsometry. No observable thickness change was found even after 100 hours of immersion, indicating a minimized etching effect in EFT.

[0047] FIG. 31 plots the detachment rate of vdW materials from substrates as a function of KOH concentration and temperature. Both the z axis and shading bar show the value of the detachment rate.

[0048] FIG. 32 plots the Raman spectrum of MoS2grown on a pristine Si02 / Si substrate 66 and on reused substrates 68 after ammonia transfer.

[0049] FIG. 33 illustrates the desired transfer of a vdW material 32 via the EFT method. Attorney Docket No. mit-25634pct

[0050] FIG. 34, for comparison with FIG. 33, shows the many issues present in the previous chemical etching transfer (CET) methods, including wrinkles and cracks, polymer residues, metal-ion contaminants, and defects.

[0051] FIGS. 35 and 36 include bright-field optical images of M0S2 transferred by EFT and CET, respectively. The M0S2 transferred by CET shown in FIG. 36 shows cracks and macroscopic wrinkles. The scale bar is 20 pm for the optical images.

[0052] FIGS. 37 and 38 include atomic-force-microscopy (AFM) images of M0S2 transferred by EFT and CET, respectively. The scale bar is 500 nm for the AFM images. The micro-wrinkle area for the M0S2 transferred by EFT (FIG. 37) is 0.06%, and the micro-wrinkle area for the M0S2 transferred by CET (FIG. 38) is 2.33%. Microscopic wrinkles 70 can be clearly seen in FIG. 38.

[0053] FIGS. 39 and 40 include dark field optical images of PMMA residues on a Si02 / Si substrate after PMMA removal. Before being fished onto Si02 / Si and removed, the PMMA film is floated on ammonia solution (FIG. 39) and hot KOH solution (FIG. 40), respectively, for an elongated time of 30 minutes. The scale bar in each represents 50 pm. The “dirt” area for the PMMA film floated on ammonia solution (FIG. 39) was 0.04%, and the “dirt” area for the PMMA film floated on KOH (FIG. 40) was 6.41%. Polymer residues 72 can be clearly seen in FIG. 40.

[0054] FIG. 41 is an SEM image of M0S2 transferred by EFT. Wrinkles, cracks, or polymer contaminations are barely observed.

[0055] FIG. 42 is an SEM image of M0S2 transferred by CET. The wrinkles exist everywhere 70, and polymer contaminations 72 can be observed.

[0056] FIG. 43 is an SEM image of M0S2 transferred by CET. Different from the exemplification of FIG. 42, the PMMA / M0S2 are not taken out of KOH for rinse right after detaching. Instead, the PMMA / M0S2 are kept floating on KOH for 90 minutes to see the chemical reaction between PMMA and KOH, which has resulted in the severe cracks 74, wrinkles 70, and polymer contaminations 72 shown in the FIG. 43.

[0057] FIG. 44 includes an SEM image and a segmentation and particle analysis image (produced using the particle analysis tools of ImageJ) of CET-M0S2, providing a substantial wrinkle area ratio measurement.

[0058] FIG. 45 includes an SEM image and a similarly produced segmentation and particle analysis image of EFT-M0S2, providing a substantially reduced wrinkle area ratio measurement.

[0059] FIGS. 46-49 include STEM images of EFT-M0S2 under different scales. The scale bar represents 2 nm in FIG. 46, which shows a monolayer 76 of EFT-M0S2; 5 nm in FIG. 47, which shows a monolayer 76 and a bilayer 78 of EFP-MoS2; 10 nm in FIG. 48, which shows a monolayer 76, a bilayer 78, a trilayer 80; and a tetralayer 82 of EFT- Attorney Docket No. mit-25634pct

[0060] MOS2; and 50 nm in FIG. 49, which shows a monolayer 76 and a bilayer 78 of EFT- M0S2.

[0061] FIG. 50 includes plots of the Raman spectrum of as-grown M0S2 on Si02 / Si 84, the Raman spectrum of as-transferred M0S2 on another identical Si02 / Si substrate using ammonia solution (EFT) 86, and the Raman spectrum of as-transferred M0S2 on another identical Si02 / Si 88 using KOH solution (CET), respectively.

[0062] FIG. 51 includes intensity and peak position mapping of E^g and AigRaman mode of EFT-M0S2. The scale bar is 2 pm. A triangular M0S2 flake is characterized to show the contrast between the substrate and the sample. The high integrity of both the position mapping and the intensity mapping of Aigand EAg shows the well-preserved crystallinity after transfer by EFT.

[0063] FIG. 52 is a plot showing the deconvolution of strain and doping in M0S2 through the correlation analysis of Aigand E^g vibration modes for as-grown M0S2 84, EFT-transferred M0S2 86, and CET-transf erred M0S2 88.

[0064] FIG. 53 is a plot of Raman spectra of CNTs transferred by EFT (left) and CET (right). The defect peaks in CET-CNTs are magnified 5 times for clarity.

[0065] FIG. 54 is a schematic illustration and false-color SEM image showing the device structure of a back-gated field-effect transistor (FET) with monolayer M0S276 as the channel. Ni is used as the contact 90; a 7.2-nm-thick layer of ALD-grown Hf0292 is used as the dielectric; and Si is used as the back gate 94. The scale bar for the SEM image is 1 pm.

[0066] FIG. 55 is a plot of the transfer characteristic curves (drain current, Id, versus gate voltage, Vgs) with a drain-source voltage (Fds) of 1 V for EFT-transferred M0S2 86 and for CET-transf erred M0S2 88.

[0067] FIG. 56 is a plot of the output characteristic curve of an EFT-Ni-MoS2-HfO2fieldeffect transistor, as shown in FIGS. 54, where Lch = 1 pm and where VGS = o to 3 V. Noted, for both EFT-M0S2 and CET-M0S2, while some of the devices exhibit an ideal output curve similar to the graph shown here, some other devices show non-linear behaviors. The variation in contact resistance can be caused by multiple reasons, such as the conditions for metal deposition.

[0068] FIG. 57 is a boxplot comparison of the on-current (Ion) for the field-effect transistor, where Vds = 1 V.

[0069] FIG. 58 is a boxplot comparison of the threshold voltages ( r) for the field-effect transistor, where Vds = 1 V.

[0070] FIG. 59 is a schematic illustration of a device with CVD M0S276 (batch #1) and an SEM image of a M0S2 batch #1 device. The M0S2 is grown by liquid precursors, and the device is fabricated on a 300-nm-thick Si02layer 92 atop a Si back gate 94, with a Attorney Docket No. mit-25634pct channel length of 500 nm and a contact metal width of 200 nm. The scale bar for the SEM image is 2 |im.

[0071] FIG. 60 is a plot showing transfer characteristics (drain current, Id, versus gate- to-source voltage, Vgs) with a drain-source voltage ( Vds) of 1 V for the device of FIG. 59 with CVD M0S2 batch #1.

[0072] FIG. 61 provides a boxplot comparison of the on-current (Ion) for the CVD M0S2 batch #1 devices, where Vds = 1 V.

[0073] FIG. 62 provides a boxplot comparison of the threshold voltages (VT) for the CVD M0S2 batch #1 devices, where Vds = 1 V.

[0074] FIG. 63 is a schematic illustration of a device with CVD M0S276 (batch #2) and an optical image of a M0S2 batch #2 device. The M0S2 is grown by solid precursors, and the device is fabricated on a 300-nm-thick Si02layer 92 atop a Si back gate 94, with a channel length of 1 pm and a contact metal width of 2 pm. The scale bar for the optical image represents 5 pm.

[0075] FIG. 64 is a plot showing transfer characteristics (drain current, Id, versus gate- to-source voltage, Vgs) for the CVD M0S2 batch #2 devices with a drain-source voltage Vds) of 1 V for EFT-M0S2 86 and CET-M0S2 88, where Vds = 1 V.

[0076] FIG. 65 provides a boxplot comparison of the on-current (Ion) for the batch #2 devices, where Vds = 1 V.

[0077] FIG. 66 provides a boxplot comparison of the threshold voltages ( VT) for batch #2 devices.

[0078] FIG. 67 is a schematic and optical image of M0S2 devices with a transfer-length- method (TLM) structure. The M0S2 is grown by MOCVD on a sapphire substrate and transferred by EFT (EDL repulsion) onto a pre-patterned substrate with local bottom gates. The bottom gate is Au / Ti (50 nm / 10 nm), and the gate dielectric is 10-nm-thick atomic-layer-deposition (ALD)-grown Hf02. The channel length is 200 nm. The width of the contact metal isipm. The scale bar is 3 pm.

[0079] FIG. 68 plots the output characteristic curves of a Bi-MoS2-HfO2FET with Left = 200 nm and with a channel length of 200 nm for a range of gate-to-source voltage, VGS, values.

[0080] FIG. 69 plots the transfer characteristics of M0S2 with bismuth as contact metals. The top solid plot 98 is for Vds = 1 V, while the bottom solid plot is for Vds = 0.05 V. The dashed line is the gate leakage, which indicates that the off-current is limited by the gate leakage instead of the channel material. Mobility was about 36.8 cm^S’1.

[0081] FIG. 70 extracts the contact resistance using the TLM method from the device in FIGS. 67-69. The total resistance (Rtot) is plotted against channel length (Left). The extracted contact resistance is Rc= ~ns D. Attorney Docket No. mit-25634pct

[0082] FIG. 71 is a plot of a transfer characteristic of one device set that includes seven large-scale FETs fabricated with EFT-M0S2. The inset shows the device set. Ni is used as the contact, and a Si substrate with a 300-nm-thick Si02coating (where measurements to coatings or layers herein generally refer to thickness unless otherwise indicated or implied) is used as a back gate and dielectric. The channel length is 1 pm, and the channel width is 5 pm. Vds = 1V.

[0083] FIG. 72 is an on-current (Ion) map of 24 large-scale EFT-MoS2-fabricated device sets over 5 mm x 6 mm. The x and y axes in the map show the location of the device sets.

[0084] FIG. 73 is a plot of the benchmark of transfer yield of transferring MoS2from Si02, sapphire, and gold tape via the EFT method 102 described herein, via chemical etching 104, via hydrophilicity 106, via metal or adhesive-polymer assist 108, and via ice assist no.

[0085] FIG. 74 provides a comparison of various transfer approaches, including EFT 102, electrobubbling 112, chemical etching 104, and metal assist 108 in terms of industrially relevant factors, including scalability, cost-effectiveness, yield, time efficiency, and versatility. The time efficiency considers the time cost of processing steps (for example, the time cost of vacuum pumping, metal deposition, and metal etching in metal-assisted transfer 108).

[0086] FIG. 75 shows the process flow of an EFT method with only poly(methyl methacrylate (PMMA) as a support layer 30 and a comparison with the CET method.

[0087] FIG. 76 shows the process flow of an EFT method with a thermal release tape (TRT) 120 and PMMA as a support layer.

[0088] FIG. 77 shows the process flow of an EFT method with PDMS as a support layer.

[0089] FIG. 78 is a schematic illustration showing detachment of TRT / PMMA in the case that the TRT is not in intimate contact with the PMMA. At those places that are not closely adhered, the attraction force between the PMMA and the substrate is larger than that between the PMMA and the TRT; therefore, the detachment happens between the PMMA and TRT while the PMMA remains on the original substrate. This is similar to the observation in graphene transfer by EFT. Due to the fact that graphene is only partially decoupled, some places are still strongly adhered with the growth copper substrates and will remain on the original substrate.

[0090] FIG. 79 is a plot of adhesion energy (Fo) of graphene (denoted as Gr) and MoS2on different substrates.

[0091] FIG. 80 shows the qualitative relation between vdW energy, Uvdw, and separation distance, r.

[0092] FIG. 81 plots the magnitude of the vdW attraction force as a function of adhesion energy at separation distance, r = 3 nm. Attorney Docket No. mit-25634pct

[0093] FIGS. 82 and 83 are SEM images of as-grown M0S2 on Si02 / Si (FIG. 82) and M0S2 after being transferred by EFT onto a new Si02 / Si substrate (FIG. 83). The M0S2 is grown by MOCVD using NaCl as the seeding promoter. During growth, the NaCl is also deposited on the substrate surface and stays at the interface between the M0S2 and the substrate. The most serious place will have a film of NaCl on the substrate, as shown in the upper SEM image. Scale bars for both upper and lower SEM images are 1 pm.

[0094] FIG. 84 is a plot from energy dispersive x-ray spectroscopy (EDS) on the M0S2 flake before (as-grown) 126 and after transfer 128. The as-grown M0S2 126 shows an obvious Na peak, which is attributed to the NaCl at the interface of the M0S2 and the substrate.

[0095] FIGS. 85 and 86, respectively, plot energy and total force 130 between M0S2 and sapphire in concentrated ammonia as a function of the separation distance, r. The repulsion force 132 is defined to be positive, and the attraction force 134 is defined to be negative.

[0096] FIGS. 87 and 88, respectively, plot energy and total force as a function of the separation distance, r, when the solution pH is changed. Plots are provided for pH=i2 136, pH=io 138, pH=8 140, pH=6 142, and pH=4 144.

[0097] FIGS. 89 and 90, respectively, plot energy and total force as a function of the separation distance, r, when the ionic strength is changed. Plots are provided for 0.02 mol / L 146, 0.1 mol / L 148, 0.2 mol / L 150, 0.5 mol / L 152, and 1 mol / L 154.

[0098] FIGS. 91 and 92, respectively, plot energy and total force as a function of the separation distance, r, for vdW materials / substrates with different adhesion energies. Plots are provided for 100 mJ / m2156, 200 mJ / m2158, 300 mJ / m2160, 500 mJ / m2162, 1,000 mJ / m2164, 3,000 mJ / m2166, and 5,000 mJ / m2168.

[0099] FIGS. 93 and 94 show two different paths for the surface (support layer and vdW material 16) to separate from a source substrate 12, involving the same net change in energy but involving very different repulsion forces. In general, since APE = Fdx, the force can be large or small depending on the path taken, which is prescribed by x.

[0100] FIG. 95 is an SEM image of the as-transferred hBN from sapphire onto Si02 / Si substrate by EFT.

[0101] FIG. 96 plots the Raman spectrum of hBN before and after transfer. The bottom line 170 is as-grown hexagonal boron nitride (h-BN) on sapphire. The top line 172 is as- transferred h-BN on Si02 / Si substrate by EFT. The h-BN characteristic peak is marked with the dashed line.

[0102] FIG. 97 is a schematic illustration of a device structure of a back-gated fieldeffect transistor (FET) with monolayer M0S276 as the channel and an SEM image of a typical short-channel device with Lch of 25 nm. The FET was transferred by the EFT Attorney Docket No. mit-25634pct method. Nickel (Ni) is used as the contact 90, and ALD-grown Hf02is used as the dielectric 92. The scale bar for the SEM image is 50 nm.

[0103] FIG. 98 plots the transfer characteristic curves (drain current, Id, versus gate-to- source voltage, Vgs) for Vds = 1 V 174 and for Vds = 0.05 V 176 of a long channel device with a channel length, Lch, of 1 pm. The channel is monolayer M0S2 transferred by the EFT method. The inset plot includes zoomed-in curves (Vds = 0.05V) near Vth to show the voltage hysteresis, wherein AVH= 11.1 mV, and wherein the subthreshold swing (SS) is 68.4 mV / dec.

[0104] FIG. 99 shows the transfer characteristics (drain current, Id, versus gate-to- source voltage, Vgs) of a short channel device with a channel length, Lch, of 25 nm, where s = 1 V 174, and where Vds = 0.05 V 176. The channel is monolayer M0S2 transferred by the EFT method. The inset plot includes zoomed-in curves near Vth to show the drain-induced-barrier lowering (DIBL).

[0105] FIG. 100 plots the output characteristics (drain current, Id, versus drain-to- source voltage, Vds) of a short channel device with a channel length, Lch, of 25 nm, where Vg= -1.5 V to 3 V, and where the voltage step was 0.5 V.

[0106] FIG. 101 represents large-scale FETs fabricated with EDL-M0S2, wherein an on- current (Ion) map of 300 devices (with a transfer yield of 100%) from two device batches (with an area of 10 x 10 mm2for each). The devices boxed by the dashed line are short channel devices with Lch = 25 nm. The others are long channel devices with Lch = 1 pm. The x and n axis in the map shows the locations of the devices.

[0107] FIG. 102 plots the on current from a control experiment comparing the on- current of Ni-MoS2FETs transferred by EFT, H20, and KOH. The average Ion and their standard deviations are 48.4 ± 13.5 pA / pm (EDL-M0S2), 38.2 ± 8.2 pA / pm (H20- M0S2), and 4.6 ± 2.5 pA / pm (K0H-M0S2), with relatively standard deviation being 27.9%, 21.5%, and 54.3%, respectively. Lch = 1 pm, and Vds = 1 VThe K0H-M0S2 has fewer data points because KOH transfer has much lower yield.

[0108] FIG. 103 plots the benchmarking transfer yield versus mobility with a few representative reports with similar geometries (back-gated FETs with monolayer M0S2 as the channel). Plots are provided for the EFT method 102 described herein, for a water-based method 178, for an ice-assisted method no, for chemical etching 104, for a dry-exfoliation method 180. The shading corresponds to the grain size. The lightest shading corresponds to single-crystalline M0S2. The unfilled points with grey frame indicate the grain sizes are not mentioned in the literature.

[0109] FIG. 104 plots the benchmarking hysteresis versus channel length in the representative reports with similar device structures (monolayer M0S2 FETs in a back- gated geometry). Plots are provided for the EFT method 102 described herein, for a water-based method 178, for chemical etching 104, and for no transfer 182. Attorney Docket No. mit-25634pct

[0110] FIG. 105 plots the benchmarking subthreshold swing versus channel length in the representative reports with similar device structures (monolayer M0S2 FETs in a back-gated geometry). Plots are provided for the EFT method 102 described herein, for dry exfoliation 180, for a water-based method 178, for chemical etching 104, and where unmentioned 184.

[0111] FIG. 106 plots the output characteristics (drain current, Id, versus drain-to- source voltage, Vds) of a bismuth contact short channel M0S2-FET with a channel length, Lch, of 22 nm, where Vg= -1 V to 2.5 V, and where the voltage step is 0.5 V.

[0112] FIG. 107 plots the benchmarking on-current versus channel length at the same drain-to-source 433 voltage (Vds) of 1 V in monolayer M0S2 FETs in a back-gated geometry. The data point from A. Mondal, et al., “Low Ohmic Contact Resistance and High On / Off Ratio in Transition Metal Dichalcogenides Field- Effect Transistors via Residue-Free Transfer,” Nat. Nanotechnol. 19, 34-43 (2024), was measured at a low temperature of 15 K. All the other data points in the benchmark are room temperature measurements.

[0113] In the accompanying drawings, like reference characters refer to the same or similar parts throughout the different views; and apostrophes are used to differentiate multiple instances of the same item or different embodiments of items sharing the same reference numeral. The drawings are not necessarily to scale; instead, an emphasis is placed on illustrating particular principles in the exemplifications discussed below. For any drawings that include text (words, reference characters, and / or numbers), alternative versions of the drawings without the text are to be understood as being part of this disclosure; and formal replacement drawings without such text maybe substituted therefor.

[0114] DETAILED DESCRIPTION

[0115] The foregoing and other features and advantages of various aspects of the invention(s) will be apparent from the following more-particular description of various concepts and specific embodiments within the broader bounds of the invention(s). Various aspects of the subject matter introduced above and discussed in greater detail below may be implemented in any of numerous ways, as the subject matter is not limited to any particular manner of implementation. Examples of specific implementations and applications are provided primarily for illustrative purposes.

[0116] Unless otherwise herein defined, used, or characterized, terms that are used herein (including technical and scientific terms) are to be interpreted as having a meaning that is consistent with their accepted meaning in the context of the relevant art and are not to be interpreted in an idealized or overly formal sense unless expressly so defined herein. For example, if a particular composition is referenced, the composition Attorney Docket No. mit-25634pct may be substantially (though not perfectly) pure, as practical and imperfect realities may apply; e.g., the potential presence of at least trace impurities (e.g., at less than 1 or 2%) can be understood as being within the scope of the description. Likewise, if a particular shape is referenced, the shape is intended to include imperfect variations from ideal shapes, e.g., due to manufacturing tolerances. Percentages or concentrations expressed herein can be in terms of weight or volume. Processes, procedures, and phenomena described below can occur at ambient pressure (e.g., about 50-120 kPa— for example, about 90-110 kPa) and temperature e.g., -20 to 50°C— for example, about 10- 35°C) unless otherwise specified. Solutions are in water unless otherwise specified.

[0117] Although the terms, first, second, third, etc., maybe used herein to describe various elements, these elements are not to be limited by these terms. These terms are simply used to distinguish one element from another. Thus, a first element, discussed below, could be termed a second element without departing from the teachings of the exemplary embodiments.

[0118] Spatially relative terms, such as “above,” “below,” “left,” “right,” “in front,” “behind,” and the like, maybe used herein for ease of description to describe the relationship of one element to another element, as illustrated in the figures. It will be understood that the spatially relative terms, as well as the illustrated configurations, are intended to encompass different orientations of the apparatus in use or operation in addition to the orientations described herein and depicted in the figures. For example, if the apparatus in the figures is turned over, elements described as “below” or “beneath” other elements or features would then be oriented “above” the other elements or features. Thus, the exemplary term, “above,” may encompass both an orientation of above and below. The apparatus maybe otherwise oriented (e.g., rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly. The term, “about,” can mean within ±5% or ±10% of the value recited. In addition, where a range of values is provided, each subrange and each individual value between the upper and lower ends of the range is contemplated and, therefore, disclosed.

[0119] Further still, in this disclosure, when an element is referred to as being “on,” “connected to,” “coupled to,” “in contact with,” etc., another element, it maybe directly on, connected to, coupled to, or in contact with the other element or intervening elements may be present unless otherwise specified.

[0120] The terminology used herein to describe particular embodiments is not intended to limit the represented concepts to the particulars of the exemplary embodiments. As used herein, singular forms, such as those introduced with the articles, “a” and “an,” are intended to include the plural forms as well, unless the context indicates otherwise. Additionally, the terms, “includes,” “including,” “comprises,” and “comprising,” specify Attorney Docket No. mit-25634pct the presence of the stated elements or steps but do not preclude the presence or addition of one or more other elements or steps.

[0121] In the methods described herein, an electrical-double-layer-(EDL)-repulsion- assisted process, which can be used to effectively separate van der Waals (vdW) materials from their source substrates, enables large-scale, CMOS-compatible, facile, etching-free (for both substrates and materials), and widely applicable transfer of vdW materials. The EDL effect has been widely studied in colloidal science, and the EDL effect has also been applied to the semiconductor industry (standard clean i) for nanoparticle removal. When an object is immersed in a solution, its surface is often charged and surrounded by counterions for charge neutrality. This results in the formation of an EDL. The type of charge in the EDL and its strength are characterized by the zeta potential, which is defined as the potential at the slipping plane of the EDL (as shown in FIG. 1 and discussed in Supplementary Note 1, infra). A negative zeta potential indicates that the surface is negatively charged and surrounded by positive counterions. For two similarly charged surfaces facing each other, the counterion concentration in the gap is much higher than in its concentration in the solution, leading to entropy-driven osmotic pressure that repels the objects, as illustrated in FIG. 2 and discussed in Supplementary Note 2, infra. The EDL force is highly sensitive to the solution pH, which alters the zeta potential, and is usually inversely correlated with ion concentration in the solution.

[0122] In a concentrated ammonia solution (32 wt%, pH =~12), we found that the zeta potentials of substrates and vdW materials are usually both negative (FIGS. 3 and 4), and the ion concentration in the solution is as low as 0.02 mol / L. As a result, the vdW material 32 is easily detached from its substrate 12, as shown in FIG. 5, by the strong EDL repulsion. Due to the universal EDL formation, the electrical-double-layer-force- enabled transfer (EFT) method can be widely applicable to the transfer of various vdW materials 32, such as carbon nanotubes (CNTs) 18; single- or few-layer structures 20, such as graphene or hexagonal boron nitride (h-BN); or transitional metal dichalcogenides 22, such as M0S2, WSe2, NbSe2, etc., from various source substrates 12, including dielectrics and metals selected from oxides 24, such as Si02, A12O3, CuO, Hf02, Nb2O5, SrTiO3, LaA103, sapphire, etc.; metals 26, such as copper (Cu), titanium (Ti), gold (Au), etc.; or nitrides 28, such as GaN, SiNx, etc. (FIG. 5). Particularly, for aligned carbon nanotubes grown on quartz or sapphire substrates, where the transfer process has been more challenging than other vdW materials, we find that the EFT method is very effective, further showing its wide applicability and advantage. In the later part of this disclosure, CVD-grown MoS2on Si02 / Si will be used for detailed investigation to further compare with other existing transfer methods. Such an EDL repulsion mechanism has never been explored as a transfer method before; here, it Attorney Docket No. mit-25634pct demonstrates fast, damage-free, efficient, and large-scale transfer of vdW materials with CMOS compatibility and could greatly facilitate the development of nextgeneration electronics via vdW materials integration.

[0123] As the EFT provides a facile solution for detaching the vdW materials from substrates, the choice of the support layer in EFT is versatile (see the Methods section, infra, and FIGS. 6 and 7 for detailed descriptions and illustration), wherein the vdW materials 32 can be, e.g., carbon nanotubes, graphene, hexagonal boron nitride (h-BN), or transition metal dichalcogenides, such as M0S2, WSe2, NbSe2, etc. Here, we illustrate the process using a thermal release tape (TRT) and a polymethyl methacrylate (PMMA) layer as the support layer 30 so that after the wet detachment, the process can enable dry attachment of the vdW material 32 and the support layer 30 to a target substrate 34. An example of an EFT process is illustrated in FIG. 7.

[0124] With the simultaneous formation of an EDL repulsion force in an ammonia bath 14, the vdW materials 32 with the support layer 30 can be immediately peeled off while maintaining wafer-scale integrity. For instance, less than 5 seconds are needed to fully detach a ixi-inch MoS2layer from a Si02substrate. After delamination, the TRT / PMMA / vdW is taken out of the solution for rinsing, blow drying, and attaching it to the target substrate 34. The dry attachment allows precise positioning and multilayer stacking. We illustrated the potential of large-scale stacking of MoS2with designed angles via multiple EFT steps to transfer and stack a plurality of metal-organic chemical-vapor-deposited (MOCVD) MoS2layers using the EFT process. Furthermore, the EFT process is scalable. FIGS. 8-11 shows the transferred 2-inch-scale metal-organic chemical-vapor-deposition (MOCVD) grown MoS2using EFT with a well-intact film, which should be easily extended to even larger scales. The process flow of using TRT and PMMA is similar to the process flow using polydimethylsiloxane (PDMS) as the support layer.

[0125] To understand the mechanism of EFT quantitively, we calculated the forces in the transfer process with reference to the classical model of the Derjaguin, Landau, Verwey, and Overbeek (DLVO) theory and the additional steric hydration effect. The forces include an attractive vdW force (Fvdlv), a short-range repulsive steric hydration force and a long-range repulsive EDL force (FFDL) (see the Methods section, infra, for the calculation parameters). We define a repulsive force to be positive and an attractive force to be negative. In the case of an atomically thin monolayer and a thick 3D substrate, the Fvdwcan be expressed as: where r is the distance between the monolayer and the substrate surface; r0is the equilibrium separation distance; and f0is the interfacial adhesion energy per unit area. Attorney Docket No. mit-25634pct

[0126] This is derived from the Lennard- Jones potential (see Supplementary Note 3, infra, for the derivations). The FHydoriginated from the non-negligible sizes of molecules and ions in the confined vicinity of the aqueous solution, usually treated as an exponentially decaying force given by: where C is a constant usually determined by experiment; andHis a characteristic decaying length. With reference to the classical model of DLVO theory, the entropy- driven EDL force is a repulsive force that is often derived by solving the Possion- Boltzmann equation. For two charged surfaces, FEDLis given by: where i / q and ip2are the zeta potentials of the monolayer and the substrate, respectively; E is the dielectric constant of concentrated ammonia; EQis the vacuum permitivity; r is distance between the monolayer and the substrate surface; and LDis the Debye screening length (a measure of the double-layer thickness) given by: where I is the ionic strength of the electrolyte; kBis the Boltzmann constant; T is the temperature; NAis Avogadro’s number (~6.O22 xio23 mol’1); and e is the elementary charge. The ionic strength in the expression is given by:

[0127] I = f iZiCi, (5) where zLis the valence of the ith ion and ctis the concentration of the ith ion in the solution.

[0128] Based on the force calculations above, our proposed detaching mechanism for a vdW material 32 from a source substrate 12 in EFT comprises the following two steps (FIG. 12): (1) initial gap enlargement by steric hydration repulsion 40, (2) EDL formation and further gap enlargement by EDL repulsion 42, and (3) detachment 44. First, when a vdW material / substrate is immersed in the concentrated ammonia solution, the ammonia solution penetrates from the edges slightly into the vdW material / substrate interface gap, a tendency to balance the chemical potential inside and outside the gap. Since the initial vdW material / substrate distance is very small (~I.65 A), the molecules penetrated inside the interface to generate a strong steric hydration force 40, which enlarges the vdW gap at the interface edge, as illustrated in FIG. 12. However, the steric hydration force 40 alone does not detach vdW materials 32 due to its much faster attenuation than the vdW force [see Equations (1) and (2)]. After Attorney Docket No. mit-25634pct the initial gap enlargement by the steric hydration repulsion 40, both the vdW material 32 and the substrate 12 are negatively charged (the origins of which are explained below) and attract positive charges to their surfaces, thus two EDLs form at the interface (FIG. 2). The origins of objects being charged in a solution vary for different materials and surroundings. The solution pH is particularly important because the adsorption of protons and hydroxyl groups generates positive charges and negative charges, respectively. For Si02in concentrated ammonia, the adsorption of hydroxide ions (OH-) makes Si02negatively charged, which can be expressed by the following formula:

[0129] High pH

[0130] SiO2+ 20H- - >SiO2(OH)22. (6)

[0131] In the case of metal-oxide substrates, the metal-oxide substrates are negatively charged in concentrated ammonia by the deprotonation effect of the hydroxyl group (M-OH) on the oxide surface:

[0132] High pH

[0133] (M — OH) + OH- - > (M - 0)- + H20. (7)

[0134] For vdW materials, we suspect that they are also negatively charged with similar mechanisms. With the formation of the EDL, the long-range EDL force dominates over the vdW attraction force [see Equation (3) and FIGS. 12 and 13]; and the EDL force will further repel the vdW materials and the substrate. The EDL repulsion force 42 enlarges the interface gap effectively and ultimately detaches the vdW material 32 from the substrate 12. This strong, simultaneously formed EDL repulsion force in concentrated ammonia facilitates the success of EFT. The total repulsion force 48 resulting from the combined steric hydration repulsion 40 and EDL repulsion 42, as well as the absolute vdW attraction 46 are plotted in FIG. 13.

[0135] The above EFT mechanism is supported by the following evidence. First, a strong repulsion force is observed in the concentrated ammonia solution. Three pieces of Si02gel (Methods section, infra) were put into concentrated ammonia (NH3~ 32 wt%), KOH (~ 1 mol / L), and NH4C1 (~ 1 mol / L) solutions, respectively. When the Si02gel was immersed in concentrated ammonia, it immediately increased its volume and diffused. On the contrary, expansion was not observed in either KOH or NH4C1 solution. FIG. 14 plots the expansion length, which measures the width of the Si02gel pieces as a function of time, in ammonia solutions 54, in NH4C1 solutions 56, and in KOH solutions 58. The initial expansion speed is estimated to be around 300 mm / s. These observations indicate that a strong repulsion force is generated in concentrated ammonia, but it does not exist in the KOH or NH4C1 solutions. This repulsion force should not be caused by a chemical-reaction-induced concentration gradient, as the Si02gel dissolves more easily in the KOH solution with higher alkalinity. The intercalation effect by NH4+is also less likely because NH4C1 solution should have a Attorney Docket No. mit-25634pct higher NH4+concentration. The origin of this repulsion force should be more likely due to the EDL force. In contrast to KOH and NH4C1 solutions, concentrated ammonia is a solution with both a high pH value and weak ionic strength, which is the typical chemical environment that promotes formation of a strong EDL, as will be explained in the experiments below.

[0136] Using a M0S2 / sapphire substrate as an example case, we further investigated the impact of ionic strength and pH on EFT. FIG. 15 shows the dependence of the detaching rate on the ionic strength. The detaching rate is defined as the rate at which ammonia penetrates into the vdW / substrate interface and detaches vdW material from the substrate. We let the sample float on the liquid surface without using TRT so that the detaching rate should reveal the strength of the EDL force. The ionic strength is tuned by adding NaCl into ammonia. It turned out that when the ionic strength increases, the detaching rate drops significantly (FIGS. 15 and 16, with the half-detachment of PMMA by ammonia solution). As shown in FIGF. 16, the detachment rate was highest for pure ammonia 54, followed by 0.37 mol / L NaCl solution 56, with 0.04 mol / L NH4C1 solution 58 having the lowest detachment rate. This agrees well with our calculation, shown in FIG. 17, that the EDL force drops dramatically with increasing ionic strength [see Equations (3)"(5)J-

[0137] With the increased ionic strength I, the debye length decreases, which eventually results in the decrease of electrostatic potential ip r) = [toe~r / LD(Debye-Hiickele equation), in which ip0is the surface potential. This explains why EDL repulsion is not observed in strong alkali solutions, such as KOH, that have a much higher ion concentration than ammonia. On the other hand, FIG. 18 shows the detaching rate as a function of pH. The pH is tuned by adding HC1 into concentrated ammonia. A digital pH monitor is used to measure pH at each data point. The detaching rate drops dramatically when the pH value decreases. FIG. 19 shows the calculated dependence of the EDL force on pH. EDL force drops at decreasing pH, in agreement with the experiments. The pH influences the EDL force via zeta potential. As the pH value decreases, the zeta potential of M0S2 and sapphire both become less negative. The zeta potential of sapphire changes from negative to positive when pH is further reduced (FIG. 4). As a result, the EDL force first decreases and then changes sign into an attraction force. Thus, with the weak ionic strength and high pH value in concentrated ammonia, the EDL force is prominent, while it is not observed in KOH, NaCl, or NH4C1 solutions.

[0138] The methods for transferring a vdW material can use any of a variety of methods utilizing ammonia solution, including but not limited to: a) using ammonia solution to transfer different types of van der Waals materials from different substrates; b) using ammonia solution combined with different types of support layers for transfer, wherein Attorney Docket No. mit-25634pct the support layer can be selected from, e.g., polydimethylsiloxane (PDMS), polymethyl methacrylate (PMMA), polypropylene carbonate (PPC), polycarbonate (PC), polystyrene (PS), thermal release tape (TRT), TRT + PMMA, PDMS + PMMA, etc.; or c) any dry or wet transfer methods that utilize ammonia, regardless of the ammonia concentration.

[0139] To further distinguish the EFT process from conventional chemical etching transfer (CET), we compare the substrate damage for both of the transfer processes. When a pristine Si02 / Si substrate was immersed in concentrated ammonia at room temperature for 30 minutes, the substrate remained undamaged. However, for Si02 / Si immersed in hot KOH (70°C, 1 mol / L), a typical condition used in CET, Si02 / Si was severely damaged with a clear change of color due to the reduced oxide thickness. FIG. 20 quantitatively shows the change of oxide thickness and surface roughness, which are measured by ellipsometry and atomic force microscopy (AFM), respectively (FIGS. 22- 27, wherein the plots of FIGS. 25-27 include both calculated values 62 and measured values 64), indicating that the EFT method induces minimal damage to the substrate. Besides, the detaching rate of EFT decreases when NaCl is added to the solution. If the EFT is based on chemical etching, the transfer speed should not change because the pH value does not change. The EFT is further distinguished from CET by its successful application to gold substrates and its much faster detaching rate with a zero-etching rate (FIGS. 28-31). Therefore, chemical etching is not involved in the EFT method.

[0140] During EFT, the source substrate is undamaged and can be reused for cycled growth, which is favorable in terms of reducing production costs. FIG. 18 shows the growth of MOS2on both an as-purchased Si02 / Si (only cleaned with acetone and isopropyl alcohol) and a reused substrate after EFT. The repeated growth shows good quality from both optical images (FIG. 18) and Raman spectroscopy (FIG. 32). Furthermore, FIG. 21 shows that carbon nanotubes (CNTs) grown on reused singlecrystalline quartz are still aligned, which means the substrate surface preserves its epitaxial relation with the nanomaterials after EFT. Such reuse of epitaxial substrates could be very important for industrial application of epitaxial growth of lowdimensional materials from a cost-effective perspective.

[0141] In addition to the etching-free nature of EFT to substrates, the vdW materials are also well-preserved, with much reduced issues, as observed in current transfer methods, such as wrinkles, cracks, polymer residues, metal ion contaminations, and defects (illustrated in FIG. 33), in comparison with vdW materials transferred via CET (see FIG. 34). FIGS. 35 and 36 show the optical-microscope images of MOCVD-grown monolayer MoS2transferred by EFT and CET, respectively. Meanwhile, FIGS. 37 and 38 show the atomic-force-microscope (AFM) images of MOCVD-grown monolayer MOS2transferred by EFT and CET, respectively. For a fair comparison, PMMA is used Attorney Docket No. mit-25634pct as a support layer for both transfers. Wrinkles and cracks are barely observed on EFT- M0S2 (see FIGS. 35 and 37). In contrast, in CET-M0S2, two types of wrinkles and cracks are observed (see FIGS. 36 and 38). One is macroscopic wrinkles 69 and cracks 74 (with a length > 10 pm) that are easily observed under an optical microscope (see FIG. 36). The other type of wrinkle is microscopic wrinkles 70 (with a length of about 1 pm), which are only observed under AFM and SEM images (FIGS. 38, 42, and 43). The macroscopic wrinkles and cracks in vdW materials can be generated when the support layer cannot provide continuous mechanical support.

[0142] While PMMA is a robust polymer, it has chemical reactions with the commonly used chemical etchants in CET (HF, hot KOH, or NaOH), which greatly weakens its mechanical strength. As evidenced by FIG. 43, when PMMA / M0S2 is immersed in hot KOH for an elongated time, cracks 74 and wrinkles 70 increase significantly due to the increased degree of chemical reaction. On the other hand, concentrated ammonia does not react with PMMA at room temperature; therefore, in EFT, the macroscopic wrinkles 69 or cracks 74 are significantly reduced. In terms of the microscopic wrinkles 70, the experimentally measured ratios of areas of micro-wrinkles 70 to the total vdW material areas are 2.33% (with a standard deviation of 0.40%) for CET-M0S2 and 0.06% (with a standard deviation of 0.08%) for EFT-M0S2, as extracted from SEM images using the ImageJ segmentation and particle analysis tools (FIGS. 44 and 45). The severe microwrinkles 70 in CET-M0S2 are likely generated by the heating and cooling step involved in CET, considering the large mismatch of the thermal expansion coefficients between M0S2 and PMMA (1.9 XIO-4 / K versus 7.6xio_6 / K).

[0143] In Supplementary Note 5, infra, we estimated the wrinkle areas that the heatingcooling process can introduce. The calculated value is ~2.27%, close to our experimental observation as well as to literature reports. The close agreement indicates that the heating-cooling process in CET is likely the cause of micro-wrinkles. In EFT, however, the detachment is done at room temperature so the microscopic wrinkles are avoided. Note that the above discussions only focus on transfer-induced wrinkles. In some cases, such as graphene grown on copper, the wrinkles have already been induced during the growth process. The EFT method cannot heal those growth-induced wrinkles, but it effectively reduces the transfer-induced wrinkles and cracks.

[0144] Second, the EFT enables clean transfer with much less polymer residues. It is usually perceived that PMMA leaves significant polymer residues on vdW materials, and cleaner support layers are being researched. However, we note that no matter what support layer is used, PMMA is usually involved in the following device fabrication steps as an EBL resist, which has led to M0S2 transistors with high performances, which raises a question: why is PMMA so dirty when used in transfer but less problematic in device fabrication? We found one of the likely reasons is that PMMA has severe Attorney Docket No. mit-25634pct chemical reactions with the commonly used etchants in CET (e.g., HF, hot KOH, hot NaOH, and FeCl3), resulting in the damage of PMMA with insoluble residues. On the other hand, the ammonia solution used in EFT does not damage PMMA at room temperature, thus leaving much less PMMA residue after an acetone rinse.

[0145] To verify this hypothesis, we floated PMMA film on ammonia and hot KOH (70 °C), respectively, for 30 minutes. The PMMA films were then transferred onto a Si02 / Si substrate. FIGS. 39 and 40 show the dark field optical images (oblique incidence with about 1 x 108lux illuminance) of the Si02 / Si substrate after PMMA removal, respectively, with ammonia and KOH. Because of the very high illuminance, even nanocontamination significantly scatters the light, becoming visible. The polymer contamination is barely observed (with a dirt area of 0.04%) on Si02for PMMA floated on ammonia, while in the hot KOH case, a great amount of polymer residue 72 (with a dirt area of 6.41%) is observed. The same observation is found when transferring MoS2. As shown in FIGS. 41-43, the EFT-MoS2(FIG. 41) is much cleaner than CET-MoS2(FIGS. 42 and 43). Note that when PMMA / MoS2is floated on hot KOH for an elongated time during CET, the MoS2becomes much dirtier with a significant number of unremovable particles, which is additional evidence of the severe chemical reaction between KOH and PMMA. Furthermore, FIGS. 46-49 show the scanning transmission electron microscope (STEM) images of EFT-MoS2under different scales. The MoS2is ultra-clean at the atomic scale despite the use of PMMA. It is apparent from this investigation that the EFT method prevents chemical damage, leading to a substantially cleaner transfer process.

[0146] Additionally, the EFT process is free of metal-ion contaminations. In CET, the existence of metal ions from KOH or NaOH makes the transfer process CMOS- incompatible, and causes unintentional doping of the as-transferred vdW materials. To understand the doping effect on MoS2caused by the transfer process, we performed a statistical study on the Raman E^g and Aigmode for EFT-MoS2and CET-MoS2(typical Raman spectra and mapping are shown in FIGS. 50 and 51). In FIG. 50, the Raman spectra for as-grown MoS284, EFT-MoS286, and CET- MoS288 are plotted for comparison. FIG. 52 shows the deconvolution of strain and doping through the correlation analysis of Aigand E2gvibration modes for as-grown MoS284, EFT-MoS286, and CET- MoS288. The E2gpeak of both EFT-MoS286 and CET-MoS288 exhibits an evident blue shift, which is attributed to the release of tensile strain induced by the CVD growth. CET-MOS288 is slightly further blue-shifted than EFT-MoS286. This result agrees with the wrinkle formation in KOH transfer that will cause compressive strain. Regarding the Aigpeak that corresponds to the doping of MoS2, while the average Aigpeak position of EFT-MoS286 and CET-MoS288 are close to each other, the variation of the Aigpeak in CET-MoS288 is much larger than in EFT-MoS286 (see the Attorney Docket No. mit-25634pct circled region in FIG. 52). This difference is attributed to the K+in the KOH solution that can be trapped at a random location on the M0S2 or substrate surface, which leads to an unintentional doping effect.

[0147] Last but not least, the transfer-induced defects are further characterized by the Raman spectra of CVD-grown carbon nano tubes (CNTs) after transfer. Here, CNTs were chosen instead of M0S2 because the D band intensity of CNTs is directly correlated with defect density. FIGS. 52 and 53 show the Raman spectra of CNTs from the same growth batch but transferred by EFT and CET (hot KOH), respectively. For the EFT method, PMMA / hBN is used as a support layer. It turns out that the hot-KOH transfer significantly increases the D peaks, while no D peak is observed in EFT-CNTs. Therefore, CET introduces a significant number of defects into vdW materials while the defects introduced by EFT are minimized.

[0148] To study the influence of the transfer to device performances, we transferred two pieces of MOCVD-grown M0S2 (from the same growth) using EFT and CET methods, respectively, and fabricated back-gated field effect transistors (FETs) with identical device structures, including nickel contacts 90 on a monolayer of M0S2, which is on a 7.2-nm-thick layer of Hf02, which is on a P++ silicon gate layer 94 (FIG. 54, see the Methods section for details). The channel length was set at 1 pm so that the resistance was dominated by channel material instead of contact. The transfer characteristic curves for both EFT-M0S2 86 and CET-M0S2 88 FETs are shown in FIG. 55, and the output curve is shown in FIG. 56. The on-current (Ion) distributions are plotted in the box plots in FIG. 57. The EFT-M0S2 FETs exhibit 290% higher average Ion than CET- M0S2 FETs (17.9 ± 9.0 pA / pm vs. 4.6 ± 2.5 pA / pm), while the relative standard deviations of Ion are close to each other (defined as sion / <Ion>10, where sion and <Ion> are the standard deviation and mean value of Ion, respectively, and similarly hereafter). The threshold voltages, VT, box plots are shown in FIG. 58. While the average VT is similar. The relative standard deviation of V in EFT-M0S2 is 46% less than that in CET-M0S2 (0.80 ± 0.15 V versus 0.81 ± 0.28 V). To avoid the variations induced by material growth or device fabrication, we repeated the above comparison using M0S2 grown by CVD and other device structures in controlled experiments (FIGS. 59-66). Within each of the two batches of devices here (batch #1 is shown in FIGS. 59-62, while batch #2 is shown in FIGS. 63-66); everything is kept the same, except the transfer method (EFT versus CET). The differences between batch #1 and #2 are the growth recipe of M0S2 and the device geometry.

[0149] They all show similar trends of enhanced Ion and reduced VT variations for the EFT method. The enhanced Ion is attributed to the minimal wrinkles, contaminations, and defects in EFT-M0S2 (FIGS. 35-40, 46-49, 52, and 53), leading to reduced scattering of electron transport and thus higher mobility. The reduced VT variation is Attorney Docket No. mit-25634pct credited to zero metal ions (K+, Na+, etc.) in ammonia solution in EFT. In CET, the metal ions (K+or Na+) in the alkaline solution have a chance to stay on the vdW materials and cause random local doping. This doping leads to larger Vr variations, which also agrees with Raman statistics results. Furthermore, high-performance devices can be achieved by combining EFT and semi -metal contacts. FIG. 67 shows the transfer-length-method (TLM) devices fabricated with EFT-M0S276 and bismuth contacts 90 (see the Methods section, infra, for details). The contact resistance extracted from the TLM shows an ultra-low contact resistance of ~115 £1 and a high mobility of 36.8 cm^S’1. Therefore, the EFT method enables the production of devices with a higher on-current and a smaller variation in threshold voltage that show better device performances.

[0150] The benchmarks of transfer yield of transferring M0S2 from Si02, sapphire, and gold tape via the EFT method 102 are plotted in FIG. 73, via chemical etching 104, via hydrophilicity 106, via metal or adhesive-polymer assist 108, and via ice assist no. The transfer yields for all methods, except ice-assisted transfer no, are extracted from device measurements based on whether the device is electrically measurable or not. For ice-assisted transfer 110, the yield is based on optical-microscope images. The EFT and electrobubbling transfer methods are both CMOS-compatible.

[0151] The minimal damage to vdW materials is also evidenced by the batch fabrication of M0S2 devices over large areas. We fabricated 24 device sets at different locations. In each device set, seven back-gated FETs were included. FIG. 71 shows the transfer curves of one device set and an image of the device set (inset). FIG. 74 shows the Ion map of all the device sets, including those formed using the following transfer methods: EFT 102, chemical etching 104, hydrophilicity 106, and electrobubbling 112. All the devices work with high on-current, and no open-circuit device is observed, indicating that the M0S2 is well-preserved after transfer. The transfer yield (defined as the working device number divided by the total device number) is 100%.

[0152] Only one recent report also shows 100% transfer yield, where the M0S2 was directly delaminated by water. This indicates that the EFT and water can both achieve damage-free transfer. However, as discussed in the Methods section, infra, water transfer was only found applicable in several limited cases and did not work for most of the situations evaluated. The transfer from a Si02 / Si source substrate of a MOCVD M0S2 (#1) layer with a PMMA support layer by water was successful. The "#1" means growth recipe 1 with a growth temperature of 700 °C, 1 standard cubic centimeter per minute (seem) hydrogen, and the use of a seeding promotor (NaCl). On the other hand, attempts to transfer the following vdW materials from a Si02 / Si substrate by water using a PMMA support layer failed to produce detachment of the vdW material from the source substrate: (a) MOCVD M0S2 (#1, after previous transfer to an untreated Attorney Docket No. mit-25634pct

[0153] SiOs / Si substrate), (b) MOCVD M0S2 (#1, after previous transfer to a hydrophilic Si02 / Si substrate, wherein the Si02 / Si substrate was treated with oxygen plasma or with a solution of sulfuric acid and hydrogen peroxide), (c) MOCVD MoS2(#2), (d) exfoliated MoS2, and (e) MOCVD WSe2; we also were unable to detach the WSe2using polystyrene and water, though we were able to detach it using EDL repulsion force. The "#2" means growth recipe 2 with a growth temperature 400 °C, o seem hydrogen, and no use of seeding promotor (NaCl).

[0154] We further benchmark EFT with existing large-scale transfer methods in terms of industrially relevant factors, including scalability, cost-effectiveness, yield, time- efficiency, and versatility. A more comprehensive benchmark is shown in Table 1, below. The EFT method proposed in this work simultaneously overcomes a variety of bottlenecks in the vdW materials transfer process. Moreover, EFT is envisioned to be combined with other state-of-the-art transfer techniques. For instance, the EFT method can be easily combined with other support layers, such as polycarbonate (PC), for ultra- clean transfer. Additionally, the EFT-detached vdW materials can be taken into high vacuum to construct bubble-free and adsorbent-free interfaces. Since it introduces minimal damage to both substrates and vdW materials, while being fast, cost-effective, CMOS-compatible, and widely applicable, it is suitable for producing large-scale nanoelectronics for practical industrial applications. Attorney Docket No. mit-25634pct

[0155] Table i: benchmark of transfer methods for large-scale vdW materials.

[0156] Methods:

[0157] Electrical double-layer force-assisted transfer (EFT)

[0158] In the EFT method, the detachment of vdW materials from substrates is realized by simply putting the sample in concentrated ammonia (~32 weight-%) solution, in which the EDL repulsion force is generated spontaneously. This process can be performed with various support layers. Here, we describe the process flow for three commonly used supports, including (1) polymethyl methacrylate (PMMA) + thermal release tape (TRT), (2) polymethyl methacrylate (PMMA) only, and (3) polydimethylsiloxane (PDMS). It should be applicable to other supports with similar principles as well. The schematic and photos of the process flows are shown in FIGS. 75 and 76.

[0159] In the alternative CET and EFT processes shown in FIG. 75, (a) a vdW material 32 is first grown on a source substrate 12. (b) A support layer 30 of PMMA is spin- coated onto the source substrate 12 and the vdW material 32 and then baked at 7O°C for 5 minutes. The CET and EFT processes diverge at step (d), wherein the PMMA support layer 30 and the attached vdW material 32 are removed via the CET process in hot KOH solution 58 (1 mol / L, at ~7O°C ), wherein the detachment rate is less than 0.1 mm / s; in the EFT process, the PMMA support layer 30 and the attached vdW material 32 are detached in a 32 wt-% solution 54 of ammonia in water at room temperature, wherein the detachment rate is about 1 mm / s. In step (e), the CET-detached combined support layer and vdW material 16 are rinsed in stagnant de-ionized (DI) water 114, while the EFT-detached combined support layer and vdW material 16 are rinsed in flowing DI water 14. In step (f), both the CET-detached and EFT-detached combined support layer Attorney Docket No. mit-25634pct and vdW material 16 are attached to a target substrate 34 to place the vdW material 32 in contact with the target substrate 34. In step (g), in both processes, the target substrate 34 with the attached vdW material 32 and support layer 30, collectively labeled 118, are baked at 50°C for 30 minutes in a vacuum in a vacuum oven 116. Pressure can be applied to the laminate structure in the baking process to promote intimate contact between the layers. Lastly, in step (h), the PMMA support layer 30 is removed by applying acetone and isopropyl alcohol (IPA) to produce the target substrate 34 with the adhered and uncovered vdW material 32.

[0160] 1) EFT with TRT+ PMMA

[0161] As shown in FIG. 76, a layer of PMMA 122 (A4 950) is spin-coated [3,000 revolutions per minute (rpm) for 60 seconds] onto as-grown vdW materials 32 to serve as a support layer. The PMMA / vdW materials 16 are baked at 70 °C for 5 minutes. After baking, the edges of PMMA / vdW materials 16 are scratched with tweezers. A TRT 120 is attached to the PMMA / vdW materials 16 on the source substrate 12. Then, the stack is fully immersed in concentrated ammonia (32 weight-%) at room temperature. The TRT / PMMA / vdW 16 is then gently peeled off of the source substrate 12 by tweezers, enabled by the EDL repulsion force. Following that removal, the TRT / PMMA- supported vdW material 16 is rinsed with running deionized (DI) water, dried by a nitrogen gun, and attached to the target substrate 34. To remove interface adsorbents, the stack is placed on a hotplate (at 50°C) in a vacuum (at 1 Torr) for 15 minutes, followed by a hot press at ioo°C for 20 minutes. After the pressure is released and the chamber is vented, the thermal release tape will detach at 100 °C. Finally, the PMMA 122 is removed with acetone and isopropyl alcohol (IPA), leaving the target substrate 34 coated with the vdW material 16.

[0162] The use of a TRT 120 provides an additional peeling force that can help the detachment of films with stronger adhesion. In experimental exemplifications, the PMMA 122 baked at i8o°C was detached from a Si02 / Si substrate 12 using TRT / PMMA as the support layer 30. When PMMA 122 was used without TRT 120 (in ammonia at i8o°C), the detachment of the PMMA 122 did not happen. This is because the i8o°C- baked PMMA 122 has strong adhesion with Si02, which cannot be overcome by EDL repulsion force alone.

[0163] Detachment and transfer were, however, successful in the following contexts: (a) detachment of PMMA (baked at 70°C) in ammonia and its transfer to a Si02 / Si substrate, (b) detachment of PMMA (baked at i8o°C) bonded to a TRT in ammonia and transfer of the PMMA and TRT to a Si02 / Si substrate, and (c) detachment of PMMA bonded (baked at i8o°C) to a TRT in air and transfer of the PMMA and TRT to a Si02 / Si substrate. Attorney Docket No. mit-25634pct

[0164] The TRT enables an external manual detaching force to be applied together with the EDL repulsion force, leading to the successful detachment of the PMMA from the substrate. On the other hand, if the PMMA is baked at 70°C, the detachment of PMMA happened even without the use of TRT [as in (a), above] because the adhesion is weak and could be easily overcome by EDL repulsion force alone.

[0165] When trying to peel a TRT / PMMA / vdW / substrate directly in air, unless the adhesion between the film and substrate is extremely weak, such as the MOCVD-grown M0S2 on a Si02substrate, detachment happened between the TRT and PMMA, after the transfer, which means both PMMA and vdW materials remained on the substrate. However, when ammonia is used to generate EDL repulsion, the successful detachment of film from the substrate is realized.

[0166] Detachment of TRT 120 and PMMA 122 layers, wherein the TRT 120 is not in intimate contact with the PMMA 122, is shown in FIG. 78, where a bubble 124 is shown between these layers. At those places that are not closely adhered, the attraction force between the PMMA 122 and the substrate 34 is larger than that between the PMMA 122 and the TRT 120. Therefore, the detachment happens between the PMMA 122 and TRT 120 while the PMMA 122 remains on the original substrate. This result is similar to the observation in graphene transfer by EFT. Due to the fact that graphene is only partially decoupled, some places still strongly adhere to the growth copper substrates and will remain on the original substrate.

[0167] 2) EFT with PMMA only

[0168] A layer of PMMA (A4 950) is spin-coated (at 3,000 rpm for 60 seconds) onto as- grown vdW materials to serve as a support layer. The PMMA / vdW materials are baked at 70°C for 5 minutes. After baking, the edges of PMMA / vdW materials are scratched with tweezers to allow solution penetration. Then, the PMMA / vdW materials are gently put into concentrated ammonia (32 weight-%) with an inclined angle (~30°) at room temperature. The PMMA / vdW materials will be immediately detached from the substrates and will float on the solution surface. Following that, the PMMA-supported vdW materials are rinsed with running DI water. The customized setup is applied to enable the rinse without taking the PMMA / vdW out from the solution surface so that the potential risk of causing cracks or folds is avoided. The PMMA / vdW materials are then fished out using the target substrate, and we wait for it to dry naturally. This step is followed by a vacuum bake (at 1 Torr and at 50°C) for 30 minutes to fully remove the water residues. Finally, the PMMA is removed with acetone and isopropanol (IPA).

[0169] For the success of EFT with only PMMA, there are a couple of things to which attention should be paid. When spin-coating the PMMA, a higher bake temperature (>12O°C) and long bake time (>30 minutes) are avoided because they can increase the Attorney Docket No. mit-25634pct adhesion of PMMA to the substrate and obstruct the penetration of ammonia into the interface, which will significantly slow down the transfer speed or even result in failure.

[0170] 3) EFT with PDMS

[0171] As shown in FIG. 77, a PDMS film 123 (PF-40X40-0015-X0 gel film from Gel- Pak) is attached to as-grown vdW materials 32 to serve as a support layer 30. A gentle pressure is applied for 2 minutes to make intimate contact. Then, the PDMS / vdW- material 16 attached to the source substrate 12 is fully immersed into concentrated ammonia 54 (32 weight-% in water) at room temperature. The PDMS layer 123 is slowly peeled off from the substrate by tweezers. Due to EDL force, the vdW materials 32 will be detached together with the PDMS layer 123. Following that detachment, the PDMS / vdW material 32 is rinsed with running de-ionized (DI) water 114, dried by nitrogen gun, and attached to the target substrate 34. The target substrate 34 is pretreated with UV ozone to be more sticky. Finally, the PDMS layer 123 is directly peeled off the target substrate 34 in air. Due to the much smaller adhesion energy between PDMS 123 and vdW material 32 as compared to the adhesion energy between the vdW material 32 and the target substrate 34, only PDMS 123 will be detached, and the vdW material 32 will stay on the target substrate 34. Note that this EFT-with-PDMS method only works for vdW materials 32 with a low adhesion energy on a source substrate 12, such as M0S2. Otherwise, during detachment, only PDMS 123 will be detached due to its small adhesion energy with vdW materials 32.

[0172] Process flow for chemical-etching-assisted transfer (CET) with hot KOH:

[0173] The chemical-etching-assisted transfer (CET) process is carried out following similar steps as are used for EFT with PMMA only (see the Methods section, described above). The key difference between CET and EFT lies in the detaching step. In CET, the PMMA / vdW material is immersed in hot KOH (1 mol / L, 70°C), and >30 minutes is required for complete detachment. The concentration and temperature of KOH are chosen based on our detachment rate measurement shown in FIG. 31. We found that when KOH is used, the fastest detachment rate was realized when the concentration was ~1 mol / L and when the temperature was ~70°C.

[0174] Influence of adhesion energy on EFT:

[0175] Based on this Derjaguin, Landau, Verwey, and Overbeek (DLVO) framework, we further calculated the influence of adhesion energy on transfer (FIGS. 12, 13, and 91). For many vdW / substrates, the adhesion energy is less than 500 mJ / m2, and EFT can be successfully employed because of the FEDL being larger than the -aw. However, when adhesion energy is larger than 1000 mJ / m2, which is often the case when additional attractive force other than the vdW attraction force is involved, EFT may fail. One typical example that is of practical importance is CVD-grown graphene on copper, Attorney Docket No. mit-25634pct which is reported to have an incredibly high adhesion energy of over 6,000 mJ / m2, in which case, according to FIG. 91, EFT will fail. Experimentally, we found the prediction to be true that the as-grown graphene / copper could not be directly detached by EFT. However, when we decouple the graphene from copper, as described in the next paragraph, which effectively reduces the adhesion energy, the graphene can be successfully peeled off the copper using the EDL repulsion force. This result indicates that the adhesion energy can be taken as a reference for evaluating the effectiveness of EFT.

[0176] Application of EFT to graphene / copper and hBN / copper - an additional decoupling step needed:

[0177] CVD graphene (also hBN) grown on copper can be successfully transferred by EFT with PMMA+TRT when graphene (or hBN) is decoupled from copper by either natural oxidation (by exposing to air for >4 months), water oxidation, or exposure to hydrogen plasma. The most effective way that we currently find to achieve this decoupling is water oxidation. The water oxidation is performed by immersing the as- grown graphene / copper (or hBN / copper) in water for 46 hours at room temperature to oxidize the copper interface and to reduce the adhesion energy. After being taken out and dried, the graphene (or hBN) can be successfully detached using EFT with PMMA+TRT (see FIG. 76) following the steps shown in the Methods section, above. Note that it is difficult to completely decouple graphene from copper by post treatments, and the places that are not decoupled cannot be transferred by EFT. Currently, up to 80% graphene can be transferred by water oxidation and EFT. Further optimization may fully decouple the graphene from copper.

[0178] Measurement of detachment rate in EFT using only PMMA as support and synthesis of silica gel:

[0179] The detachment rate in FIG. 15 is defined as the speed at which ammonia solution penetrates into the interface and detaches the vdW material from the substrate. The detachment rate is used to estimate the magnitude of EDL force. The PMMA / substrate floats on the ammonia, and the detaching happens by itself. No external force is applied.

[0180] With regard to the silica gel that was used in the experiment where it was immersed in ammonia, KOH, and NH4C1, the silica gel was made from the chemical reaction between sodium silicate and ammonia.

[0181] Comparison between EFT and water transfer:

[0182] It has been reported that M0S2 can be peeled off by directly immersing it in water or ultrasonication in water. However, we found that the water transfer is not reliable in many situations with its viability highly dependent on the growth recipes, the Attorney Docket No. mit-25634pct type of vdW materials, and the substrates. As noted above, while the M0S2 grown by recipe #1 can be peeled off by water, it cannot be peeled off again after being transferred onto a new substrate, regardless of the hydrophilicity of the new substrate. In addition, neither the M0S2 grown using recipe #2, the exfoliated M0S2, nor the M0CVD-WSe2can be peeled off by water. Water transfer also fails to transfer other vdW materials, such as CNTs, h-BN, and graphene. EFT, on the other hand, is applicable to all these cases. One of the potential reasons could be that the substrate becomes hydrophilic after growth (possibly due to the surface reaction of sulfur passivation), and the capillary force enables the water transfer. Since the as-transferred M0S2 will no longer be detached from the substrate in water, water can be used to rinse and clean the as- transferred M0S2 without detaching the M0S2.

[0183] More comparison between EDL repulsion and chemical etching:

[0184] In addition to the ion sensitivity of EFT illustrated in FIGS. 15-17, additional differences between EDL repulsion and chemical etching include: 1) PMMA could be detached from gold by EFT but not by KOH because gold is not reactive with either ammonia or KOH, and the successful detachment confirms that EFT is not enabled by etching; and 2) the detachment rate of EFT is much faster than the detachment rate of CET, while being etching-free, as shown in FIGS. 40 and 41. If a transfer method is based on chemical etching, its speed should be positively correlated to the etching rate, which is the case of CET. However, for ammonia, EFT achieves a much faster detachment rate with the etching rate near zero, further distinguishing EFT from CET.

[0185] Calculation parameters used in DLVO theory and steric hydration force:

[0186] The calculations of DLVO shown in FIGS. 12, 13, and 91 are mainly based on MoS2 / sapphire, which is a widely studied system, so experimental measured parameters are accessible in the literature. For MoS2 / sapphire, the measured adhesion energy, r0= t / vdw(r0) = 101 mJ / m. The other parameters used include: r0= 1.65 A, which is taken as the “universal contact distance44found for many vdW systems; i / T = 40 mV and i / 2= 60 mV, which are the zeta potential of M0S2 and sapphire, respectively, when pH is ~12 (see Supplementary Note 1, infra, for discussions of zeta potential); ammonia concentration = 32 weight-%; kb= 1.71 x 105, which is the ionization constant of ammonia; T = 300 K; and E = 31.6, which is the dielectric constant of concentrated ammonia solution. The E for 32 weight-% ammonia was not found, so the value for 25 weight-% ammonia was used instead. The calculated Debye length with the above parameters is LD~1.37 nm. For the calculation in FIG. 17, FEDL is calculated for a different I, while r and pH are fixed (r = 3 nm and pH = 12). For the calculation in FIG. 19, FEDL is calculated for different pH values, while r and I are fixed (r = 3 nm and I = 0.02 mol / L). At different pH values, the zeta potential of the film and substrate (M0S2 Attorney Docket No. mit-25634pct and sapphire in this case) are extracted from the literature to calculate FEDL. We note that the DLVO formula only strictly holds for | | < 25 mV, and the zeta potential value is dependent on multiple factors and has variations. However, we use them here to illustrate the mechanism, realizing that any numbers generated will be approximate. With regard to the steric hydration force, the parameters used for the plots are C = 30 0.236 nm.

[0187] MOCVD or CVD growth of M0S2, CNTs, graphene, and hBN:

[0188] The wafer-scale continuous monolayer M0S2 used in this study was grown using metal-organic chemical vapor deposition. The growth was carried out in a 3-inch hot- wall quartz tube furnace. Molybdenum hexacarbonyl [Mo(CO)e, MHC, Sigma Aldrich, 99.9% purity] and diethyl sulfide [(C2H5)2S, DES, Sigma Aldrich, 98% purity] were used as chemical precursors for molybdenum and sulfur. They entered the furnace in the gas phase due to their high equilibrium vapor pressure near room temperature. The growth was performed at a temperature between 45O-7OO°C. The flow rates were 5 standard cubic centimeters per minute (seem) for MHC, 0.6 seem for DES, o or 1 seem for H2, and 800 seem for Ar, wherein the flow rates were regulated by mass flow controllers (MFC). The growth time for continuous monolayer MoS2was 5 hours on Si02. NaCl was loaded in the upstream region to act as a nucleation suppressant during the growth.

[0189] CVD growth of monolayer M0S2 using solid sources:

[0190] In this method, monolayer MoS2was prepared by solid-source CVD at atmospheric pressure. We utilized molybdenum trioxide (MoO3) powders, sulfur (S) powders, and perylene-3,4,9,io-tetracarboxylic acid tetrapotassium salt (PTAS) as the molybdenum precursor, sulfur precursor, and growth promoter, respectively. In the synthesis process, 10 mg of MoO3powders were first placed in a porcelain boat. Then, a clean Si02 / Si substrate was suspended on the top of the porcelain boat between two pieces of Si02 / Si wafers that had been coated with PTAS molecules, with the polished surface facing down. The porcelain boat was put in the middle of a i-inch tube furnace, and another boat containing sulfur powders was placed upstream. Subsequently, the furnace was ramped up to 65O°C and held for 3 minutes, and sulfur powders were heated up to 165°C during the growth process. 20 seem of argon was used as the carrier gas.

[0191] CVD growth of monolayer M0S2 using liquid-phase precursors:

[0192] In this method, monolayer MoS2was prepared by liquid-phase precursor- assisted CVD at atmospheric pressure. It involved the use of a solution containing both the molybdenum precursor and growth promoter. Specifically, 25 mg of MoO3and 25 mg of potassium iodide (KI) were dissolved together in 20 mL of ammonia (NH40H), and this solution was then spin-coated onto a clean Si02 / Si substrate. Subsequently, the Attorney Docket No. mit-25634pct precursor-coated Si02 / Si was put in the middle of a i-inch tube furnace, and another boat containing sulfur powders was placed upstream. In the synthesis process, the furnace was ramped up to 700 °C and held for 5 minutes, and sulfur powders were heated up to 165°C simultaneously. 20 seem of argon was used as the carrier gas.

[0193] Characterization of vdW materials and substrates:

[0194] The structure and composition of vdW materials before and after transfer were characterized by a variety of tools. The surface topography was measured using an atomic force microscope (AFM, CYPHER VRS AFM from Oxford Instruments Asylum Research). The Raman measurements were performed using Raman spectroscopy (INVIA confocal Raman microscope from Renishaw Inc.) with an excitation wavelength of 532 nm and a grating of 2400 mm-1. The laser power on the sample is 1 mW, and the accumulation time is 10 seconds (s). The morphology of the vdW materials was characterized using an optical microscope and a scanning electronic microscope (GEMINISEM field emission SEM from ZEISS Microscopy; 3 kV). The atomic scaling imaging was conducted using a high-resolution scanning transmission electron microscopy (TITAN THEMIS Z G3 Cs-Corrected S / TEM from ThermoFisher Scientific; 60 kV). Before the STEM imaging, the M0S2 was transferred onto the TEM grid and annealed under an argon atmosphere at 200°C for 2 days.

[0195] Fabrication and characterization of back-gated M0S2 field-effect transistors:

[0196] For the back-gated M0S2 transistors, shown in FIGS. 54, 55, 57, and 58, Hf0292 (7.2-nm thick) deposited via ALD on a heavily doped silicon substrate 94 was used as the gate dielectric, while the substrate 94 was used as the gate electrode. Electron-beam lithography (EBL) and electron-beam (e-beam) evaporation (at ~3 x io-6torr) followed by a lift-off process were used to pattern and deposit Au (50-nm thick) as the alignment markers. A continuous monolayer of MOCVD-grown M0S2 film was transferred on top of two identical substrates by EFT and CET, respectively. Then EBL and reactive-ion etching (RIE) using oxygen plasma were used to define the device area of interest. EBL and e-beam evaporation (at ~3 x io-6torr) followed by a lift-off process were used to pattern and deposit Ni (80-nm thick) as the source / drain contacts. PMMA A2 (with a molecular weight of 950,000) was spin-coated on top of the 200-mm wafer at 3,000 revolutions per minute for 60 seconds as the passivation layer of the back-gated M0S2 transistors. The sample was heated at i8o°C for 3 minutes before the electrical measurements. The other devices with Ni contacts 90 share identical steps except for the difference in the gate dielectric (FIGS. 59-66 and 71-73 used 300-nm-thick Si0292 on a heavily doped silicon substrate 94) and the M0S2 76 growth recipe (FIGS. 59-65 used CVD grown M0S2 by liquid or solid precursors, and FIG. 71-73 used MOCVD- Attorney Docket No. mit-25634pct grown M0S2). All electrical measurements described in this paragraph were carried out in air at room temperature using a semiconductor parameter analyzer (4155C analyzer from Agilent Technologies) and signal generators. Abetter electrical performance is expected if the vacuum level of the metal evaporation chamber can be improved.

[0197] For the bismuth (Bi) contact 90 M0S2 transistor shown in FIGS. 67-70, 10 nm of Ti followed by 50 nm of Au were deposited by e-beam evaporation as local bottom gates 96. Hf02(10 nm) deposited via ALD on a heavily doped silicon substrate 94 was used as the gate dielectric. Then, the MOCVD-grown M0S276 was transferred by EFT onto the local bottom gates 96. EBL and reactive-ion etching (RIE) using oxygen plasma were used to define the device area of interest. Molecular beam epitaxy followed by a lift-off process was used to pattern and deposit bismuth (20 nm) as the source / drain contacts 90, and e-beam evaporation was used to deposit 50 nm of Au as a capping layer, followed by a lift-off process. The measurement is conducted in a rough vacuum environment (i.8xio-1Torr) using the semiconductor parameter analyzer and signal generators.

[0198] Supplementary Note 1: A short introduction to zeta potential

[0199] As shown in FIG. 1, zeta potential is defined as the potential at the slipping plane of the EDL. A slipping plane is defined at the boundary between the liquid “fixed” to the surface and the mobile fluid. When a charged particle moves in a liquid medium, part of the diffuse layer will move with it. Although the plane of shear is not located right at the surface, the magnitude of zeta potential is positively correlated with the magnitude of the surface potential as long as no ion adsorption of condensation occurs. Since zeta potential is easily measured experimentally, it is usually used instead of surface potential or surface charge to estimate the magnitude of EDL force. In addition, the sign of zeta potential reveals the type of charge of EDL. A negative zeta potential is indicative of a surface being negatively charged, which will attract positive counterions surrounding it to form EDL.

[0200] Supplementary Note 2: Determining whether EDL force is attractive or repulsive for two charged surfaces

[0201] For two dissimilar surfaces facing each other, the EDL force can be either attractive or repulsive, which is dependent on the zeta potentials of the two objects (i / q, >2), the distance between the surfaces (r) and Debye length (Ln). In the context of our study, the EDL force is repulsive as long as the zeta potential for both surfaces is negative. A more detailed discussion is provided below.

[0202] First, we write down the expressions for EDL force for two dissimilar surfaces. It is impossible to find a simple analytical expression for EDL energy (and force) using the full Poisson-Boltzmann equation. Using the linearized Poisson-Boltzmann equation Attorney Docket No. mit-25634pct under the boundary condition of a constant potential, the free energy of EDL interaction per unit area, UEDL (unit is given by the “Hogg-Healy-Fuerstenau" equation as: which leads to the force per area, FEDL, to be: where r is the distance between the two surfaces, LD is the Debye length, i2are the zeta potentials, E is the dielectric constant of the solution, and E0is the vacuum permitivity.

[0203] The approximate expression for the boundary condition of constant charge is more complicated. The expression that is probably simplest and also reasonably accurate for 1:1 electrolytes is given by: where r is the distance between the two surfaces, LD is the Debye length, i2are the zeta potentials, pVjis the ion concentration in the solution, z is the valence of the ions in the solution (here, one type of valence is assumed for simplicity), e is the elementary charge, k is the Boltzmann constant, and Tis temperature.

[0204] When the separation distance is large (r / LD» 1, in our case, LD= 1.37 nm), both Equation (1) and Equation (2) reduce to: which is essentially Equation (3) from the main text, above. In this equation, the EDL force is repulsive when and ip2have the same sign. The EDL force will become attractive when the signs are opposite. Note that the surface potential is obtainable experimentally. Zeta potential is usually used to approximate the surface potential. Therefore, in our case, the force is repulsive as long as the zeta potentials are both negative.

[0205] When the separation distance is very small (r 0), the situation becomes complicated. First, under constant potential assumption, Equation (1) reduces to: Attorney Docket No. mit-25634pct

[0206] (12)

[0207] LD2rz

[0208] In this equation, the FEDLis always negative, which means the force is attractive regardless of zeta potential.

[0209] However, if we look into the constant charge assumption of Equation (2), at a small distance limit FEDL (r -> 0), it reduces to:

[0210] In this equation, the FEDLis always positive, which means the force is repulsive regardless of zeta potential.

[0211] It can be seen that when r 0, the constant charge and constant potential assumption result in opposite conclusions. Indeed, when r 0, both assumptions are inaccurate because there will be charge regulation, especially when the two surfaces are different. The counter ions may recombine with the surface charge, thus reducing the surface change and potential. The surfaces may also have ion-exchangeable sites and the charges may move and redistribute as the surfaces come into contact. When we consider the FEDL force, because the steric hydration force will dominate below LD, we only plot the FEDL force in the range r > LDin FIGS. 12 and 13.

[0212] To conclude, in the context of this work, r > LD, and FEDLcan be fairly estimated using Equation (3). When the zeta potential of vdW materials and substrates are both negative, FEDL> o. The EDL force is repulsive.

[0213] Supplementary Note 3: Derivation of vdW force

[0214] We first derive the expression for Uvdwand Fvdw. The van der Waals energy between two molecules is often expressed by Lennard- Jones potential, WLJ where C and C2are two constants for the long-range attraction force and short-range repulsive force, respectively. Noted in our case, the interaction is between an atomically thin monolayer and a thick 3D substrate. Thus, the total vdW energy, Uvdw, can be obtained by integrating all the atoms of the film and substrate: where p2Drepresents the number of atoms per unit area of the 2D film, i42o is the 2D film area, p3Dis the number of atoms per unit volume of the substrate, and V3Dis the substrate volume. Thus, Uvdwcan be yielded via the following equation: Attorney Docket No. mit-25634pct where r is the distance between the monolayer and the substrate surface, r0is the equilibrium separation, and r0is the interfacial adhesion energy per unit area. Since the attraction force is dominant when r > r0, the repulsive vdW force can be neglected.

[0215] Therefore, the vdW attraction energy, Uvdw, is calculated as follows:

[0216] The vdW force, Fvdw, can be derived from the derivative of Uvdw,

[0217] Supplementary note 4: Paths for surface separation - why force should be considered instead of energy

[0218] In the main text (FIGS. 12-15 and 17-21), we assume the vdW materials and substrates are two parallel plates and consider the force and energy required to detach them from each other. This actually overestimates the force required for detachment. This is because the adhesion energy for a certain vdW / substrate is constant, but the peeling force can vary significantly depending on the path of detachment. As shown in the schematic below, while the total adhesion energy that needs to be overcome is the same, the detachment force needed in FIG. 94 is much smaller than the case in FIG. 93 due to a longer work distance. The EFT is the second case because we are peeling off the force starting from the edge. Therefore, the force discussion in the main text should be giving the upper bound. The real detachment by EDL force will be even easier.

[0219] Two different paths for the surface to separate are shown in FIGS. 93 (straight release) and 94 (peeling), involving the same net change in energy but involving very different repulsion forces. In general, since APE = Fdx, the force can be large or small depending on the path taken, which is prescribed by x.

[0220] Supplementary note 5: Calculation of wrinkles area percentage generated by CET

[0221] The coefficients for thermal expansion for PMMA and M0S2 are 1.9*10-4 / °C and 7.6*IO-6 / °C, respectively. Assume the sample is heated up from 20°C to 80 °C in hot KOH solution and cooled back to 20°C. During the heating-up process, the PMMA expands, but M0S2 does not have obvious expansion due to its much smaller thermal expansion coefficient. However, during the cooling process, as PMMA shrinks, the M0S2 has to shrink together with PMMA, causing the formation of wrinkles on M0S2. The wrinkle area percentage can be calculated by studying the change of M0S2 area. As shown in the figure below, suppose the PMMA and M0S2 are both square shapes with Attorney Docket No. mit-25634pct side lengths, L and x, respectively. During the heat up, the PMMA expands in-plane in an isotropic manner with side length becoming L + AL, while the length of M0S2 remains x. During the cool-down process, the PMMA shrinks back into length, L, and the reduction of M0S2 side length, o, is denoted as Ax. The shrinkages of PMMA and M0S2 are proportional to each other:

[0222] L + AL x

[0223] L x — Ax '

[0224] Then, Ax is:

[0225] AL

[0226] Ax = - TV x .

[0227] L + AL

[0228] The area of M0S2 before and after the heating-cooling process will be:

[0229] So= x2, S = (x — Ax)2. (21)

[0230] The wrinkled area, AS, is:

[0231] Because: 1.9 * 10 “4* (80°C - 20°C) = 0.0114, (23) substituting Equation (23) into Equation (20) provides:

[0232] Ax AL

[0233] = 0.0113. (24) x L + AL

[0234] The wrinkled area percentage is thus:

[0235] AS = 2.27%. (25)

[0236] Enhanced Electrical Performances:

[0237] As the most direct and compelling evidence of minimal contamination and damage, we have achieved superior electrical performance in EDL-transferred devices. Using MOCVD-grown monolayer M0S2 as a typical example, we fabricated back-gated field effect transistors with both long and short channel lengths, Lch (1 pm and 25 nm, respectively). FIG. 97 shows the schematic of the device structure and an SEM image of a typical short channel device. All the devices are transferred by the EFT method unless specifically mentioned. FIG. 98 shows the transfer characteristics (drain current, Id, versus gate-to-source voltage, Vgs) of a long channel FET at both linear region (sourcedrain voltage Vds = 0.05 V 176) and saturation region (source-to-drain voltage, Vds = 1 V 174). It shows enhancement mode n-type transistor behavior with an on-off ratio > 107 and threshold voltage, Vth = ~o.62 V. Notably, the subthreshold swing (SS) is 68.4 Attorney Docket No. mit-25634pct mV / dec and the average hysteresis, AVn, is only 11.1 mV (both extracted over one order of magnitude), which approach the theoretical limitation. These two parameters are extremely sensitive to interfacial traps. The ideal SS and near zero hysteresis indicate an atomically clean interface after EFT transfer, and it is constantly observed in both long- channel and short-channel devices (best SS = ~65-9 mV / dec, best AVn= mV).

[0238] Another device parameter that is sensitive to gate-channel interface cleanness is drain-induced-barrier lowering (DIBL), one of the most critical parameters for evaluating short-channel effects in advanced node transistors. We observe a near-zero DIBL (~7 mV / V, extracted over one order of magnitude) in short-channel devices with 25-nm channel lengths (FIG. 99). Such high gate efficiency can only be achieved with an atomically clean channel-dielectric interface. Additionally, our short channel FETs exhibit ultra-high on-current (Ion). When nickel is used for the contact, the highest Ion is 570 pA / m at Vds = 1 V (FIG. 100), and the average Ion from 50 devices is 428.9 ± 89.3 pA / pm (relative standard deviation 20.9%). When bismuth is used for the contact, the Ion further boosts to 1.3 mA / pm (at Vds = 1 V), as shown in FIG. 106.

[0239] Next, we evaluate the yield of EFT transfer by fabricating large-scale FET arrays (300 devices in total from two batches; each batch covered a 10 x 10 mm2area). The transfer yield is defined as the percentage of devices that are electrically functioning. As shown in FIG. 101, our transfer yield is 100% in both short-channel and long-channel devices with the device variation statistics shown in Table 2, below.

[0240] Table 2: Device variation statistics of the devices shown in FIG. 101 Batch #1 (50 shortchannel devices and 150 long-channel devices).

[0241] This degree of transfer success can only be obtained when the transfer-induced damages, such as cracks and wrinkles, are minimized. We further conduct control experiments comparing EFT-transf erred M0S2 with water-transferred and KOH- transferred ones (denoted as EDL-M0S2, H2O-MOS2, and K0H-M0S2). As shown in FIG. 102, the EDL-M0S2 has comparable device performances as H2O-MOS2 with even higher on-current and transconductance, and they both significantly exceed that of Attorney Docket No. mit-25634pct

[0242] K0H-M0S2. The EFT transfer also exhibits considerably lower device variations than K0H-M0S2, which makes it promising for manufacturing applications. The field-effect mobility extracted from a typical EFT-M0S2 device is 59.4473 cm2^1^1using the ‘Y function’ method in the linear region. FIG. 103 benchmarks the transfer yield for EFT 102 versus mobility with other transfer methods, including water-based transfer 178, ice-assisted transfer no, chemical etching 104, and dry exfoliation 180. Only one recent report also shows 100% transfer yield, where the M0S2 was directly delaminated by water. These indicate that the EFT and water can both achieve damage-free transfer. However, as discussed in the Methods section, water-based transfer has a much narrower range of application than EFT transfer, and it does not work for most of the vdW materials and substrates. With regard to mobility, we note our mobility is lower than the current state-of-the-art using single-crystalline M0S2. It is more likely limited by the quality of as-grown materials rather than the transfer process, considering that our MOCVD-grown M0S2 is polycrystalline with a small grain size (1 pm).

[0243] FIGS. 104 and 105 present the benchmark of hysteresis and mJ / m2subthreshold swing (SS), respectively, against literature using similar device structures and materials. SS and hysteresis are two important parameters that affect the low-power operation of transistors and device reliability. However, achieving ideal SS and hysteresis in low dimensional transistors is challenging, especially for highly scaled devices where the gate control becomes increasingly difficult. With the greatly improved interface cleanness, the EFT transfer enables exceptional hysteresis and SS performances in both long- and short-channel transistors, exceeding the current state-of-the-art. For energyefficient computation, it is critical to pursue high current density at a low overdrive voltage in highly-scaled devices. The combination of EFT transfer and bismuth contact enables us to reach the upper bound of M0S2 transistor scaling, with the highest on- current density among the literature at Vds = 1V (FIG. 106). FIG. 107 benchmarks the on-current of M0S2 FETs at a constant Vds = 1V. These indicate the strong potential of EFT transfer for high-performance industrial applications targeting highly scaled devices. Additionally, ammonia solution has already been widely used in semiconductor fabrication (standard clean 1 solution, SC-1), which makes the EFT transfer process easier to be adopted in the semiconductor industry. Table 1 and FIG. 74 further benchmark EFT transfer with existing large-scale transfer methods in terms of industrial relevant factors, the values of which are either adapted from literature or estimated quantitatively (see Supplementary Note 5 for details). The EFT method proposed in this work simultaneously overcomes a variety of bottlenecks in the vdW materials transfer, thus making it promising for future vdW integration. In the future, due to the simplicity of EFT transfer, it is envisioned to be easily combined with other state-of-the-art transfer techniques. For instance, the EFT transfer can be combined Attorney Docket No. mit-25634pct with other support layers, such as PPC, for ultra-clean transfer. Additionally, the EFT- detached vdW materials can be loaded into high vacuum to construct bubble-free and adsorbents-free interfaces. Because it introduces minimal damages to both substrates and vdW materials, while being fast, cost-effective, and widely applicable, it is suitable for large-scale nanoelectronics towards practically industrial applications.

[0244] Accordingly, high-performance metal-oxide semiconductor field-effect- transistors (MOSFETs) produced via the transfer methods described herein can offer high performance, including: a) near-zero hysteresis, b) near-ideal sub threshold swing, c) high on-state current, d) high on-off ratio, e) near zero drain induced barrier lowering (DIBL), f) ideal threshold voltage, and g) high yield.

[0245] In describing embodiments herein, specific terminology is used for the sake of clarity. For the purpose of description, specific terms are intended to at least include technical and functional equivalents that operate in a similar manner to accomplish a similar result. Additionally, in some instances where a particular embodiment includes a plurality of system elements or method steps, those elements or steps may be replaced with a single element or step. Likewise, a single element or step maybe replaced with a plurality of elements or steps that serve the same purpose. Further, where parameters for various properties or other values are specified herein for embodiments, those parameters or values can be adjusted up or down by 1 / 100th, 1 / 50th, 1 / 20th, 1 / 10th, 1 / 5th, 173rd, 1 / 2, 2 / 3rd, 3 / 4* 4 / 5th, 9 / ioth, 19 / 20*, 49 / 50*, 99 / 100*, etc. (or up by a factor of 1, 2, 3, 4, 5, 6, 8, 10, 20, 50, 100, etc.), or by rounded-off approximations thereof or within a range of the specified parameter up to or down to any of the variations specified above (e.g., for a specified parameter of 100 and a variation of 1 / 100th, the value of the parameter maybe in a range from 0.99 to 1.01), unless otherwise specified. Further still, where methods are recited and where steps / stages are recited in a particular order— with or without sequenced prefacing characters added for ease of reference— the steps / stages are not to be interpreted as being temporally limited to the order in which they are recited unless otherwise specified or implied by the terms and phrasing.

[0246] Additional examples consistent with the present teachings are set out in the following numbered clauses:

[0247] 1. A method for transferring a van der Waals material, comprising: growing or transferring a van der Waals material on or onto a source substrate; adhering a support layer to the van der Waals material; then immersing the source substrate with the van der Waals material in a weak base solution, producing an electrical double layer at and between surfaces of (a) the van der Waals material and (b) the source substrate, wherein the electrical Attorney Docket No. mit-25634pct double layers generate a repulsive force that detaches the van der Waals material from the source substrate; separating the support layer with the adhered van der Waals material from the source substrate; then applying the support layer with the adhered van der Waals material to a target substrate; and then removing the support layer from the van der Waals material, leaving the van der Waals material adhered to the target substrate.

[0248] 2. The method of clause i, wherein the van der Waals material comprises at least one of transitional metal dichalcogenide, carbon nanotube, graphene, or hexagonal boron nitride.

[0249] 3. The method of clause 1 or 2, wherein the weak base solution comprises ammonia dissolved in water.

[0250] 4. The method of clause 3, wherein the production of the electrical double layer comprises: ammonia molecules reacting with water to produce an alkali environment comprising NH4+ions and OH’ ions; at least one surface of the source substrate becoming negatively charged in the alkali environment via adsorptions or chemical reactions; and positive ions in the ammonia solution being drawn to the negatively charged source substrate surface to form the electrical double layer.

[0251] 5. The method of clause 3 or 4, wherein the production of the electrical double layer comprises producing negative-charge layers at the surfaces of the van der Waals material and the source substrate via immersing the source substrate and the van der Waals material in the ammonia solution.

[0252] 6. The method of any of clauses 3-5, wherein the ammonia solution has a pH between 11 and 13.

[0253] 7. The method of clause 6, wherein the ammonia solution has an optimized pH of about 12.

[0254] 8. The method of any of clauses 3-7, wherein the concentration of ammonia in the ammonia solution is greater than 10 weight %.

[0255] 9. The method of any of clauses 1-8, wherein the source substrate comprises a metal, an oxide, or a nitride.

[0256] 10. The method of clause 9, wherein the source substrate comprises silicon.

[0257] 11. The method of clause 10, wherein the source substrate further comprises silicon dioxide.

[0258] 12. The method of any of clauses 1-13, wherein the van der Waals material adhered to the target substrate are components of a CMOS semiconductor device. Attorney Docket No. mit-25634pct

[0259] 13. The method of any of clauses 1-13, wherein the support layer with the adhered van der Waals material is separated from the support layer in less than 5 seconds.

[0260] 14. The method of any of clauses 1-13, wherein the van der Waals material is preserved through the method to be substantially free of wrinkles, cracks, transfer-induced defects, and metal-ion contaminations.

[0261] 15. The method of any of clauses 1-14, comprising reusing the same source substrate to perform a plurality of iterations of the method.

[0262] 16. A device comprising a two-dimensional layer comprising a van der Waals material, produced via the methods of any of clauses 1-15, wherein the van der Waals material is substantially free of wrinkles, cracks, transfer-induced defects, and metal-ion contaminations.

[0263] 17. A device of clause 16, wherein the device is at least one of an electronic device, an optoelectronic device, an optical device, or a quantum-qubit device.

[0264] While this invention has been shown and described with references to particular embodiments thereof, those skilled in the art will understand that various substitutions and alterations in form and details may be made therein without departing from the scope of the invention. Further still, other aspects, functions, and advantages are also within the scope of the invention; and all embodiments of the invention need not necessarily achieve all of the advantages or possess all of the characteristics described above. Additionally, steps, elements, and features discussed herein in connection with one embodiment can likewise be used in conjunction with other embodiments. The contents of references, including reference texts, journal articles, patents, patent applications, etc., cited throughout the text are hereby incorporated by reference in their entirety for all purposes; and all appropriate combinations of embodiments, features, characterizations, and methods from these references and the present disclosure may be included in embodiments of this invention. Furthermore, the components and steps identified in the Background section are integral to this disclosure and can be used in conjunction with or substituted for components and steps described elsewhere in the disclosure within the scope of the invention.

Claims

Attorney Docket No. mit-25634pctCLAIMSWhat is claimed is:

1. A method for transferring a van der Waals material, comprising: growing or transferring a van der Waals material on or onto a source substrate; adhering a support layer to the van der Waals material; then immersing the source substrate with the van der Waals material in a weak base solution, producing an electrical double layer at and between surfaces of (a) the van der Waals material and (b) the source substrate, wherein the electrical double layers generate a repulsive force that detaches the van der Waals material from the source substrate; separating the support layer with the adhered van der Waals material from the source substrate; then applying the support layer with the adhered van der Waals material to a target substrate; and then removing the support layer from the van der Waals material, leaving the van der Waals material adhered to the target substrate.

2. The method of claim i, wherein the van der Waals material comprises at least one of transitional metal dichalcogenide, carbon nanotube, graphene, or hexagonal boron nitride.

3. The method of claim 1, wherein the weak base solution comprises ammonia dissolved in water.

4. The method of claim 3, wherein the production of the electrical double layer comprises: ammonia molecules reacting with water to produce an alkali environment comprising NH4+ions and OH’ ions; at least one surface of the source substrate becoming negatively charged in the alkali environment via adsorptions or chemical reactions; and positive ions in the ammonia solution being drawn to the negatively charged source substrate surface to form the electrical double layer.

5. The method of claim 3, wherein the production of the electrical double layer comprises producing negative-charge layers at the surfaces of the van der Waals material and the source substrate via immersing the source substrate and the van der Waals material in the ammonia solution.Attorney Docket No. mit-25634pct6. The method of claim 3, wherein the ammonia solution has a pH between 11 and 13.

7. The method of claim 6, wherein the ammonia solution has an optimized pH of about 12.

8. The method of claim 3, wherein the concentration of ammonia in the ammonia solution is greater than 10 weight %.

9. The method of claim 1, wherein the source substrate comprises a metal, an oxide, or a nitride.

10. The method of claim 9, wherein the source substrate comprises silicon.

11. The method of claim 10, wherein the source substrate further comprises silicon dioxide.

12. The method of claim 1, wherein the van der Waals material adhered to the target substrate are components of a CMOS semiconductor device.

13. The method of claim 1, wherein the support layer with the adhered van der Waals material is separated from the support layer in less than 5 seconds.

14. The method of claim 1, wherein the van der Waals material is preserved through the method to be substantially free of wrinkles, cracks, transfer-induced defects, and metal-ion contaminations.

15. The method of claim 1, comprising reusing the same source substrate to perform a plurality of iterations of the method.

16. A device comprising a two-dimensional layer comprising a van der Waals material, produced via the method of claim 1, wherein the van der Waals material is substantially free of wrinkles, cracks, transfer-induced defects, and metal-ion contaminations.

17. A device of claim 16, wherein the device is at least one of an electronic device, an optoelectronic device, an optical device, or a quantum-qubit device.

Images