Skip to content

Diamond Nanowire Lift-Out Sample Preparation for TEM Analysis and Mechanical Experiments

FieldValue
JournalMicroscopy and Microanalysis
Published2026-07-01
AuthorsTymofii S. Pieshkov, C Li, Bongki Shin, Jun Lou, Yimo Han
DOI10.1093/mam/ozag053.124

Diamond nanowires measuring about 3 to 4 µm in length and 200 to 300 nm in diameter were transferred from a reactive ion etched polycrystalline diamond substrate. The transfers used a manipulator inside a scanning electron microscope and placed the wires on devices for conductive AFM, STEM, and in situ SEM mechanical measurements. The method avoids the complex assembly equipment used in some prior nanodevice work and supports direct placement on test structures.

The starting material was diamond nanowires produced by reactive ion etching from a polycrystalline diamond substrate described in separate work. The authors placed the substrate in isopropyl alcohol and sonicated it for one hour to break down the nanowires. They drop cast the resulting solution onto Si wafers before putting the samples in an SEM.

Preliminary TEM inspection gave lengths of about 3 to 4 µm and diameters of about 200 to 300 nm. Selected area electron diffraction was assigned to the (111) orientation. An HRTEM image of a tip showed that the single-crystal structure extended to the thinnest and most fragile region. The tip also carried a thin amorphous carbon layer. These observations support the authors’ description of the nanowires as predominantly single crystalline, including at their tips.

The paper does not report post-transfer dimensions. It therefore does not quantify whether welding or bending changed the wire geometry after the initial TEM characterization.

Cross-sectional lamella lift-out is widely used for TEM preparation because it preserves the cross section of flakes and multilayer devices. The authors note that this approach suits 2D materials grown as flakes and multilayered devices made by electron beam lithography. Nanowires present a different problem.

A wire can be broken from a surface with an SEM manipulator, but beam charge can then sweep it off the screen. Adding a Pt protective layer and glue for a lamella also leaves the wire unstable during thinning. The wire can move under the beam, and the authors report that this causes the lamella to break down.

The paper contrasts this behavior with prior nanodevice assembly that uses complex and expensive machines for precise transfer. Their goal was to use manipulators such as Omniprobe and EasyLift units that are common in SEMs dedicated to TEM sample preparation. The reported workflow therefore targets equipment already associated with routine electron microscopy preparation.

After drop casting from isopropyl alcohol, every nanowire was embedded in a crust that prevented movement and transfer. Similar crust formations appeared on Si wafers, Au, Ti, and glass slides. The authors suspect that the solvent picked up dust from the air. As the solution slowly evaporated, surface tension was proposed to concentrate the dust around the nanowires. This explanation remained a suspicion because the crust was not chemically identified.

The observation across the tested substrates shows that the crust was not limited to the initial Si surface. It does not establish why the crust formed, so solvent choice remains an open part of the preparation method.

The nanowire could be freed by pushing it with the needle normally to its length. The freed wire sometimes flew away or attached to the manipulator through electrostatic force. Au and Ti coatings were used on substrates to diminish charging under the electron beam. Deionized water was proposed as a cleaner solvent that would pick up less dust, but the paper reports no test of that change.

Once freed, the nanowire could be picked up with the manipulator. Pt or C deposition through a gas injection system was used for welding. The authors preferred electron beam assisted deposition because it was more precise when depositing smaller amounts of material.

They warn that focused ion beam welding can deposit material around the pattern area and uses higher energies. This can promote an unwanted bond to the source substrate, prevent further transfer, or damage the surface. Electron beam assisted welding was therefore the authors’ preferred route when the wire needed to remain movable.

After pickup, the manipulator could move the wire to a target substrate and weld it using Pt, W, or C. If the nanowires were drop cast directly on the target, the manipulator could move them to a designated area. The authors also describe bending a wire to a chosen angle while keeping its other side welded for mechanical, electrical, or magnetic testing.

This arrangement gives the wire a fixed mechanical and electrical connection to the test structure while leaving part of its geometry accessible to a probe or microscope.

The reported placements support several downstream experiments. A diamond nanowire was transferred to a push-to-pull device for strain testing. Other wires were placed on the SiN membrane of a TEM heating chip for in situ TEM heating. Nanowires were also transferred to Au-coated Si for atomic force microscopy bending and conductivity measurements.

The figures show wires with different sizes and geometries on Au-coated Si. Au and Ti coatings were used where needed to reduce electron beam induced charging. The transferred structures then supported STEM analysis and in situ SEM mechanical testing.

The paper documents successful placement for these uses, but it does not give force displacement curves, resistance values, heating temperatures, or other device performance metrics. It also omits a transfer success rate and the time required per transfer. The result is a demonstrated preparation and placement method, not a quantified process window.

The workflow preserves a freestanding wire rather than requiring it to remain stable through cross-sectional lamella thinning. That change allows the wire to be bent, welded to a test structure, and used in several measurement configurations. The method also removes the need for the complex assembly machines described in prior work, but shifts the burden to release and manipulator alignment.

The authors propose that the transfer method could apply to Si, GaN, and other nanowires prepared in the same way. Only diamond nanowires were demonstrated in this report, so the extension remains untested.

Several handling issues remain. The crust source is inferred rather than confirmed, deionized water is an untested proposed solvent, and electrostatic attachment can remove a wire after release. Focused ion beam welding also carries the risk of unwanted bonding or surface damage. For device engineers, the immediate use is controlled placement of reactive ion etched diamond nanowires on compatible test chips. Quantitative transfer reliability remains unresolved in the reported data.

The paper’s reactive ion etching process starts with a polycrystalline diamond substrate, and its SEM transfer work uses conductive Au-coated destination surfaces. The diamond materials catalog from 6CCVD lists single-crystal and polycrystalline diamond grown by MPCVD, including plates and wafers up to 125 mm, with metallization in Au, Pt, Pd, Ti, W, or Cu. Engineering support can be used to discuss custom substrate and coating requirements for TEM preparation and mechanical test structures.

View Original Abstract

In the field of scanning electron microscopy (SEM)-related instruments transmission electron microscopy (TEM) sample preparation in the form of cross-sectional cut with the lamella lift out has already earned a major place in all TEM related experiments [1]. Most of the 2D materials are usually grown as flakes and looking at their cross-sections in this way is a reasonable approach. Similarly, researchers can analyze cross-sections of various multilayered devices, such as field-effect transistors, made via e-beam lithography [2]. However, when we have a task of transferring and analyzing nanowires, the same approach gets more complex. It is very easy to break a wire off a surface with an SEM manipulator for later transfer [3], however, it is also easy for it to be swiped off the screen in the next moment because of the flowing electron and ion beam charges. If we were to make the transfer similar to lamella preparation by introducing Pt deposition as a protective layer and to “glue” wires together, the wire itself will not be stable during thinning which would cause lamella to break down. There are reports of researchers using very complex and expensive machines to assemble nanodevices which to make the transfer precise [4]. At the same time, our goal was to develop a method that can be used with easily accessible equipment like Omniprobe or EasyLift manipulators that are usually present in each SEM, dedicated to TEM sample preparation. In this work, we are describing some of the challenges (and propose solutions) that we encountered while trying to transfer diamond nanowires from the reactive ion etched (RIE) substrate to various target substrates. We also demonstrate how the results of transfers were used in conductive AFM measurements, STEM analysis and in-situ SEM mechanical testing. Initially, diamond nanowires were acquired via reactive ion etching method from polycrystalline diamond substrate that is described in a separate work [5]. Fig. 1a shows the steps used for initial sample preparation of diamond nanowires and how they were transferred to Si wafer before being put inside the SEM. Diamond substrate was put inside the isopropyl alcohol (IPA) solution and sonicated for one hour to break down the nanowires. Fig. 1b shows the nanowires after RIE and before sonication. Next, we drop casted the solution onto Si wafers. Preliminary characterization with the TEM (Fig. 1d) shows the geometry and size of the diamond nanowires with their length being ∼3 - 4 µm and diameter being ∼200 - 300 nm. Diamond nanowires are predominantly single crystalline in nature as we can see from the selected area electron diffraction (SAED) image on Fig. 1e and even at their tips (where the wire is the thinnest and most fragile) they show single crystalline morphology with a thin amorphous carbon layer coating (Fig. 1c). When performing the transfer with manipulator needle inside the SEM, we encountered a range of problems, specific to morphology and size of our sample. After the drop casting of IPA solution, we discovered that every nanowire ends up entrenched in a crust that prevents its further movement and transfer. We suspect that the crust is being formed from the dust that IPA solution picks up from the air. As it slowly evaporates this dust accumulates around the nanowires due to surface tension. The only way we were able to free the wire is by pushing it with the needle normally to its length. Sometimes after the nanowire is freed, it flies away as is shown on Fig. 2a or gets attached to the manipulator due to electrostatic force. We tested this approach on various substrates such as Si wafer, Au, Ti and glass slide, and we encountered similar crust formations on each. Au and Ti coating was used to diminish the substrate charging inside the SEM due to electron beam. One of the possible solutions for the crust formation would be using deionized water as a solvent which would diminish the amount of dust picked up by the solution. Once the nanowire is freed, it can be picked up with manipulator. Pt or C deposition via gas injection system can be used for welding, preferably with the help of electron beam since it is more precise at smaller amounts of deposited material. If focused ion beam is used for welding, some amount of welding material will be deposited around the pattern area and, since it also uses higher energies, it would be easier for the manipulator and the nanowire to be welded to the substrate, preventing further transfer and/or damaging the surface in the process. Next, the manipulator with the wire can be moved to the region of interest on a different substrate and welded to it in the same way with Pt, W or C. Alternatively, if nanowires are drop casted directly on the target substrates, they can be moved around by the manipulator to the designated area. They can also be bent at specific angles, keeping the other side welded for mechanical, electrical or magnetic testing. By using this method, we managed to transfer diamond nanowires with different geometries to a range of substrates: for strain testing using push-to-pull device (Fig. 2c), in-situ TEM heating using heating chips (Fig. 2d), and atomic force microscopy bending and conductivity measurements on top of Au coated Si wafer (Fig. 2e). Even though, this method has been used on diamond nanowires, it can be applicable for the transfer of Si, GaN and other types of nanowires prepared in the same way [6]. Sample preparation and characterization of diamond nanowires. a) steps of sample preparation of diamond nanowires from RIE to SEM manipulator transfer, b) SEM image of diamond nanowires after RIE, c) HRTEM image of the tip of the nanowire, d) TEM image of the wire after sonication, e) SAED of the (111) orientation of nanowire Problems, transfer and results of in-situ SEM nanowire transfers. a) surface crust that was formed around nanowire after drop casting from IPA solution, b) freed from crust wire, being picked up by the manipulator, c) diamond nanowire, transferred to push-to-pull device for mechanical transfer, d) nanowires, transferred on the SiN membrane of the TEM heating chip, e) two SEM images with diamond nanowires of different size transferred on Au coated Si wafer