Nanowire facilitated transfer of sensitive TEM samples in a FIB

Ultramicroscopy 219 (2020) 113075 Contents lists available at ScienceDirect Ultramicroscopy journal homepage: www.elsevier.com/locate/ultramic Nanowire facilitated transfer of sensitive TEM samples in a FIB a,b a,b a,b a a Saleh Gorji , Ankush Kashiwar , Lakshmi S. Mantha , Robert Kruk , Ralf Witte , ⁎ ⁎ Peter Mareka, Horst Hahna,b, Christian Kübela,b,c, , Torsten Scherera,c, a b c T Institute of Nanotechnology, Karlsruhe Institute of Technology, 76344 Eggenstein-Leopoldshafen, Germany KIT-TUD-Joint Research Laboratory Nanomaterials, Technical University Darmstadt, 64287 Darmstadt, Germany Karlsruhe Nano Micro Facility, Karlsruhe Institute of Technology, 76344 Eggenstein-Leopoldshafen, Germany A R TICL E INFO A BSTR A CT Keywords: TEM sample preparation Nanomanipulator Focused ion beam Thin film transfer Lamellae lift-out FeRh We introduce a facile approach to transfer thin films and other mechanically sensitive TEM samples inside a FIB with minimal introduction of stress and bending. The method is making use of a pre-synthetized flexible freestanding Ag nanowire attached to the tip of a typical tungsten micromanipulator inside the FIB. The main advantages of this approach are the significantly reduced stress-induced bending during transfer and attachment of the TEM sample, the very short time required to attach and cut the nanowire, the operation at very low dose and ion current, and only using the e-beam for Pt deposition during the transfer of sensitive TEM samples. This results in a reduced sample preparation time and reduced exposure to the ion beam or e-beam for Pt deposition during the sample preparation and thus also reduced contamination and beam damage. The method was applied to a number of thin films and different TEM samples in order to illustrate the advantageous benefits of the concept. In particular, the technique has been successfully tested for the transfer of a thin film onto a MEMS heating chip for in situ TEM experiments. 1. Introduction Preparing an electron transparent sample for transmission electron microscopy (TEM) and transferring it to a TEM grid is one the most critical steps for successful TEM measurements of any bulk material. Focused ion beam (FIB) systems [1] are currently one of the most powerful and flexible instruments to prepare site-specific TEM samples [2]. FIB has been proven to be reliable to prepare a wide range of different TEM samples [3]. During conventional in situ FIB sample preparation from a bulk material, a lamella is prepared, lifted out by a micromanipulator, and attached to a TEM grid, followed by careful final thinning to make an electron transparent area with the desired thickness for the TEM analysis. Alternatively, ex situ lift-out (EXLO) techniques, using an optically controlled micromanipulator station outside the FIB, can be used as easy and fast approach to transfer a precut thin TEM ready sample to a suitable grid [4], where it adheres by electrostatic [5], van der Waals, and capillary forces [6]. This prevents electron/ion beam damage during the transfer process [7] and the need for gas deposition induced fixation of the sample, thus reducing potential alterations of the TEM sample. If the original sample is already electron transparent with a thickness suitable for TEM analysis, e.g. in case of a freestanding thin film or a mechanically/chemically prepared ⁎ thin disk, part of the thin area can be cut free and directly transferred to a TEM grid using the micromanipulator inside the FIB. This approach is particularly beneficial to maintain the native state of the sample by minimizing damage of the sample due to gallium ion beam cutting of the bulk specimen and minimizing contamination with gallium, platinum/tungsten and carbon. In recent years, in situ TEM experiments [8] have been taken to a new level by using microelectromechanical based systems (MEMS) with minimized drift and increased precision and stability for various in situ TEM experiments such as in situ electrical [9,10], thermal [11–14], electro thermal [15], mechanical [16,17], and in situ electromechanical [18] experiments inside TEM. However, facile, reliable and reproducible transfer of sensitive TEM samples onto the MEMS based in situ TEM chips, which maintains the native state of the material, is one of the main challenges to perform meaningful in situ TEM experiments. Ex situ lift-out approaches can be used to transfer samples onto MEMS based TEM chips and e.g. for in situ heating experiments, it has been shown that Ga ion beam effects during sample transfer inside a FIB could be avoided by an ex situ transfer [7]. However, more commonly, TEM samples are transferred by in situ lift-out inside a FIB, typically relying on the rigid tip of a tungsten micromanipulator [9,13,19], and the final thinning is carried out after attaching the sample to the Corresponding authors at: Karlsruhe Nano Micro Facility, Karlsruhe Institute of Technology, 76344 Eggenstein-Leopoldshafen, Germany. E-mail addresses: christian.kuebel@kit.edu (C. Kübel), torsten.scherer@kit.edu (T. Scherer). https://doi.org/10.1016/j.ultramic.2020.113075 Received 8 April 2020; Received in revised form 27 June 2020; Accepted 11 July 2020 Available online 15 July 2020 0304-3991/ © 2020 Published by Elsevier B.V. Ultramicroscopy 219 (2020) 113075 S. Gorji, et al. membrane of the MEMS chip. This can be problematic in terms of the mechanical forces imposed on the sample, the contamination and damage produced during FIB cutting and welding for the sample transfer and damage to the membrane during the final thinning. As an alternative, a rotating micro-gripper and an adhesive material has been used to transfer a TEM lamella to a MEMS based chip in order to prevent Pt contamination during FIB welding and damage of the membrane during the final thinning on the MEMS chip [20]. However, an adjustable rotating micro-gripper and adhesive materials are two additional requirements, and the adhesive material itself can be a source of contamination. We have previously shown that it is possible to use a stabilizing transfer mask to transfer mechanically sensitive samples onto push-to-pull (PTP) devices for in situ tensile experiments [21]. However, PTP devices have a rigid silicon platform, whereas the membranes of the MEMS chips are rather fragile for using and handling a transfer mask. Another approach can be to modify the conventional tungsten micromanipulator itself. There have been several efforts in that direction for nano-manipulation [22–29]. The modifications of the W micromanipulator span from simply sharpening the tip down to less than 200 nm by electrochemical etching [26] to create micro- and nano-grippers and actuators based on alloys or composite materials that exhibit a shape memory effect [22–27], or using a combined FIB-CVD (chemical vapor deposition) method to directly grow a nanomanipulator on the W micromanipulator [28,29]. Nevertheless, these modifications of the W tip for micro/nano manipulation still result in a rigid manipulator in case of sharpened tips [26] or would need significant additional synthesis/preparation or actuation equipment inside the FIB in case of the nano-tweezers and nano-actuators. Ideal would be a universal, easy to use method that minimizes preparation time, damage, and contamination of the sample while only requiring a single nanomanipulator inside the FIB. With this work, we introduce a fast and facile approach to attach a single pre-synthetized flexible freestanding Ag nanowire to the tip of a tungsten micromanipulator to transfer sensitive TEM samples inside the FIB with minimal stress-induced bending or straining of the sample and significantly reduced preparation time, contamination and beam damage. The mechanical properties of an Ag nanowire were examined and five different examples for TEM sample transfer are presented to show the potential of this approach to prepare a wide range of different in situ TEM samples and also for conventional TEM lamella lift-out. in situ lift-out. The FIB is equipped with a gas injection system (GIS) for Pt, C, and W. In this work, all Pt contacts and the patches for welding the nanowire to the W micromanipulator tip and to the TEM sample itself were made by e-beam induced decomposition of a Pt metalorganic precursor (trimethyl(methylcyclopentadienyl)platinum: C5H4CH3Pt (CH3)3) from the GIS system at 5 kV accelerating voltage and 5 nA ebeam current. Throughout this work, the focused ion beam was operated at 30 kV with a low ion-beam current of 9 and 28 pA, which was used to cut the Ag nanowire as well as the freestanding FeRh thin film [32]. However, ion-beam-induced Pt deposition was not used during any step of a sensitive TEM sample transfer in this work due to the larger contamination area, compared to e-beam-induced Pt deposition [10], and to avoid ion beam damage. An aberration (image) corrected Titan 80-300 (FEI Company) operated at 300 kV acceleration voltage, equipped with a US1000 slow-scan CCD camera (Gatan Inc.), was used for TEM experiments. In situ TEM uniaxial tensile measurements were performed on an Ag nanowire transferred onto a push-to-pull (PTP) device using a Picoindenter PI 95 straining holder (Hysitron). In situ Lorentz TEM thermal experiments were carried out using an Aduro 210 heating holder (Protochip Inc.). During the in situ Lorentz experiments the objective lens was switched off to have a magnetic field-free environment, with a stray field of around 0.01 mT measured separately using a Hall probe holder from ThermoFisher. 3. Results and discussion 3.1. Preparation and mounting a nanowire on a micromanipulator tip The first step to prepare a nanowire manipulator is to make freestanding nanowires available, ready to be picked up and to be attached to the micromanipulator with minimum effort and time consumption. One way would be to use vertically grown nanowires on a substrate [33], where one would simply pick a nanowire from vertically stacked nanowires. However, as an easily accessible alternative, we dispersed a drop of a commercially available silver nanowire suspension onto a conventional gold TEM grid (without carbon film) with a mesh size corresponding to the nanowire length which results in a large number of Ag NWs immobilized on the TEM grid (Fig. 1a). The diameter of the nanowires can be adjusted to fit the necessary strength and ductility for the sample transfer. As an example, the mechanical characteristics of a typical nanowire with 320 nm diameter are illustrated in the supplementary information (SI) (Fig. S1). The next step is to pick up the nanowire with the tungsten micromanipulator. Although it is easier with a sharp W tip (Fig. 1b), this is even possible with a thick undefined W tip (Fig. 1c). The reason for this facile pick-up in both cases is that the welding of the wire to the W tip using e-beam induced Pt deposition strongly connects the wire to the tip of the micromanipulator in only a few seconds. Subsequently, the other end of the nanowire is released by cutting off the nanowire with a low ion current (9–28 pA), depending on the diameter of the nanowire in a matter of seconds. For instance, for a nanowire with diameter of ∼320 nm, it typically takes around 4 s to cut the nanowire using a 9 pA ion current and only 1–2 s using a 28 pA current. Similarly, throughout this work, 3–6 s of e-beam induced Pt deposition was enough to weld a nanowire to a sample for reliable lift-out. Another important aspect is the flexibility of the nanowire mounting. A simple tilt and rotation adjustment of the TEM grid, from which the nanowire is picked up, enables orienting the nanowire with respect to the micromanipulator axis or the sample for dealing with special sample geometries. The possibility to rotate both the micromanipulator and the flip stage (in FEI Strata 400) provides additional degrees of freedom for a proper adjustment of a nanowire on a micromanipulator tip. Employing these degrees of freedom makes it possible to carefully align the angle of the nanowire with respect to the probe axis, before the nanowire attachment, using SEM and low-current FIB imaging. If necessary, to ensure the nanowire is aligned exactly along the probe axis, one can inspect 2. Experimental procedure A drop of Ag nanowires in isopropyl alcohol (0.5% suspension with an average diameter of 120–150 nm and 20–50 µm in length, Aldrich) was dispersed on a hexagonal gold TEM grid (400 mesh, Plano) to immobilize the nanowires and provide freestanding nanowires for convenient pick-up. A 16 nm thick FeRh magnetic thin film was epitaxially grown by co-evaporation on a MgO substrate using a customized mini molecular beam epitaxy (MBE) system [30] followed by a subsequent post-annealing. The single-crystalline FeRh thin film contains Rh-rich precipitates ranging in size approximately from less than 10 nm to a couple of hundred nanometers. The freestanding FeRh thin film was prepared by etching the MgO substrate using a 0.3 molar solution of ethylenediaminetetraacetic acid (EDTA) at pH 4 at ∼75 °C [31]. The etched-off freestanding film was collected using an Au TEM grid (200 mesh, Quantifoil, R 1.2/1.3) covered with a patterned (circular) holey carbon foil ∼12 nm thick, with 1.2 µm hole size and 1.3 µm spacing. Magnetic characterization of the freestanding film and the FeRh film on the MgO substrate was performed using a MPMS XL Quantum Design SQUID (Superconducting Quantum Interference Device) magnetometer. The results of the magnetic characterization are presented in the SI. A Strata 400S dual-beam FIB system (FEI Company) equipped with an OmniProbe 100 micromanipulator (Oxford Instruments) with a tungsten tip was used for the nanowire manipulator fabrication and the 2 Ultramicroscopy 219 (2020) 113075 S. Gorji, et al. Fig. 1. (a) Overview of a freestanding Ag nanowire on a hexagonal TEM grid, and nanowire pick-up using (b) a sharp and (c) a thick tungsten micromanipulator. the angle of the attached nanowire in FIB/SEM while rotating the probe. For a detailed description of the concentric alignment procedure, see the SI. After a desirable attachment condition is met, the nanowire manipulator is ready to be used for a TEM sample transfer. Another advantage of using the nanowire manipulator for TEM sample transfer is the possibility to store the nanowire used for one transfer and re-use it for later sample transfers. Fig. 2 shows the process in which a nanowire is picked up from a TEM grid and stored on a Cu grid inside the FIB on a flip stage for later usage. The storing and retrieval procedure is quick and can be done in a couple of minutes. During this process, first the nanowire on the W tip approaches a Cu grid for storage (Fig. 2a-b). It only takes a few seconds of e-beam-induced Pt deposition to attach the nanowire to a Cu grid (Fig. 2c), and a few seconds of low ion-current of to cut off the nanowire from the W tip (Fig. 2d) and store it for later usage (Fig. 2e). Moreover, as can be seen from Fig. 2f, it is possible to further thin the tip of the nanowire down to around 30 nm, convenient for picking up extremely small samples. 3.2. Transferring large area FeRh thin films for in situ TEM To illustrate the beneficial possibilities the Ag nanowire modified micromanipulator offers, we present the rapid transfer of a 16 nm thick FeRh thin film (placed on a holey carbon TEM grid) onto a MEMS chip for in situ thermal testing. The FeRh thin film preparation steps are shown in Fig. 3a and the details are provided in the experimental procedure subsection. All attempts with a conventional micromanipulator transfer of this film failed. The scientific motivation for this study are the interesting magnetic properties of nearly equi-atomic FeRh, which is antiferromagnetic (AF) at room temperature and undergoes a first-order magneto-structural transition to a ferromagnetic (FM) phase at around 78–87 °C with potential for spintronics [34] and magnetocaloric applications [35,36]. The transition temperature is very sensitive to compositional variations, chemical homogeneity, interfacial straining (induced by the substrate or the capping layer) and the thickness of the FeRh thin film [37] as well as anti-site defects, which has been used to finely tune the AF-FM transition temperature [38]. Damage of the film by Ga ion implantation during FIB preparation or Fig. 2. Storing a nanowire on a TEM grid inside the FIB: (a) Overview of the nanowire on the W tip approaching a Cu grid for storage, (b) magnified view of the dashed square in (a) showing a 20 µm long Ag nanowire attached to a W micromanipulator, (c) nanowire in contact and attached to the Cu grid with a few seconds of e-beam induced Pt deposition, (d) cutting the nanowire close to the W tip using a few seconds of low ion beam current of 9 pA, (e) nanowire attached to the Cu grid and stored for later usage, (f) and nanowire sharpened. 3 Ultramicroscopy 219 (2020) 113075 S. Gorji, et al. Fig. 3. Preparation and transfer of a FeRh thin film to a MEMS chip for in situ TEM experiments: (a) Preparation steps of the freestanding FeRh thin film for the nanowire–assisted transfer, where FeRh epitaxially grown on a MgO substrate, is chemically etched by EDTA solution (step 1) and subsequently placed on a carbon coated TEM grid (step 2). (b) Overview of the freestanding FeRh film on a TEM grid after four cuts have been made (white arrows) and after marking the film with square holes as reference points (red arrows), (c) after the nanowire is attached and the film is cut free, (d) after transfer of the FeRh thin film to the MEMS chip, held by the Ag nanowire above the MEMS device, (e) cutting the nanowire, (f) the FeRh film fully attached to the membrane and (g) e-beam-induced Pt deposition for contacting the film on the W electrodes of the MEMS chip (at 52° tilt view). mechanically induced defects will also shift this transition temperature. The damage is very significant when the FIB is directly applied to cut the thin the film for a cross-section or plan-view sample preparation [39]. There is also a possibility of lateral diffusion of the ions during the cutting process, which is unavoidable [40]. Finally, because of the Gaussian profile of the focused ion beam, the beam tails will cause damage even significantly outside the selected area to be cut [41,42]. This implies, that even by minimizing Ga ion beam exposure, by refraining from direct FIB imaging of the area of interest, some degree of local Ga ion beam damage during the sample preparation is inevitable. Therefore, in this study, we cut and transferred large-area freestanding FeRh thin films to MEMS chip membranes to perform in situ heating experiments. The large-area freestanding FeRh film was prepared by etching the MgO substrate with EDTA and collecting the free-floating film on a holey carbon coated TEM grid. This completely avoids any mechanical or ion beam preparation and thus does not create additional defects in the thin film. SQUID magnetometry experiments (Fig. S2 in SI) were performed on the FeRh film on the MgO substrate and the chemically etched freestanding FeRh film to confirm that the film was transferred intact. Based on the magnetization versus temperature plots, there was no indication of any damage to the film due to the etching and transfer. The AF to FM transition during heating starts at 78–87 °C and is complete at 146 °C, where the film is fully FM. The transition temperature of the freestanding film is shifted by around 15 °C to higher temperatures compared to the MgO supported film as the 0.3% epitaxial strain on the MgO is released in the freestanding film and indicates a low density of defects generated during the preparation. However, as the film is only around 16 nm thick and contains Rh-rich precipitates leading to locally strained areas, it tends to fold after etching from the substrate as is visible in Fig 3b. Nevertheless, the main area of interest (marked in Fig 3b) is a single layer FeRh film freestanding on the holey carbon film (1.2 µm hole size). This area, unaffected by bending, was placed on the e-beam transparent region (marked circle in Fig. 3f) of the SiN membrane of the MEMS chip. To ensure that the FeRh thin film is not bending further during the cutting process, after the nanowire is attached, we used an approach, where four cuts are made in the film, only leaving the corners of the film attached to the main film (Fig. 3b). Right after the last cut has been performed, the film stays completely straight only supported by the nanowire (Fig. 3c). Moreover, since the FeRh thin film is very sensitive to Ga ion beam damage changing the magnetic properties, the entire transfer process was carried out only using the 4 Ultramicroscopy 219 (2020) 113075 S. Gorji, et al. electron beam, whereas the film was never imaged with the ion beam; only the areas to be cut were ever exposed to the ion beam. In order to facilitate this, the exact position of the coincidence point was determined by carefully aligning the eucentric height at 52° tilt and the coincidence point of the FIB and SEM at zero tilt at another location at the same height. The area of interest was then cut only using the SEM for navigation. As the cut is made at zero degree tilt and the incident ion beam is not perpendicular to the area to be cut, an additional step is required to mark and locate the exact area of interest. Small square marks (1 × 1 µm2 square holes) were made using low ion current FIB and imaged in the SEM. Their relative distance is measured using the SEM to see the location of the drilled square (because the sample is tilted in FIB view). This drilled square was then used as a reference point for cutting, relying on the accuracy of the FIB-SEM coincidence point. Thus, we were able to prepare the area of interest without ever imaging it by FIB minimizing ion-beam damage and Ga implantation. After the successful preparation procedure, a big square (∼35 × 35 µm2) was cut and the nanowire attached by a few seconds of e-beam-induced Pt deposition. The square was subsequently transferred (Fig. 3d) and placed on the active part of the MEMS chip. During the transfer, the flexibility of the Ag nanowire prevents bending of the sample or the membrane of the MEMS chip. It can be seen that the film attached to the nanowire is floating above the MEMS chip (Fig. 3d). After the transfer to the chip (Fig. 3f) it keeps its shape and there is no noticeable bending of the film compared to the initial stage (Fig. 3b). As soon as the thin film gets close enough to the supporting membrane of the MEMS chip, the thin film is attracted and adheres to the membrane of the MEMS chip by Casimir and van der Waals forces [22,43]. As the film adheres to the membrane, small Pt patches are deposited on four sides of the film to prevent the film from moving before cutting the nanowire. The nanowire is then quickly (in a few seconds) cut with a low ion beam current of around 9 pA. When cutting the nanowire, because of the height difference between the film and the area of the nanowire to be cut (in FIB view), the FeRh thin film is not directly exposed to the ion beam (Fig. 3e), but just a small area of the membrane is cut by the ion beam. The red arrow in Fig. 3e points to the small cut in the membrane, where the nanowire was cut, which is far from the FeRh area of interest. Moreover, a few seconds of low ion beam current exposure was sufficient to cut the wire, not causing significant damage of the membrane. As a last step, the thin film was electrically contacted to the W electrodes of the MEMS chip using e-beam-induced Pt deposition (Fig. 3g). As the freestanding FeRh thin film was transferred together with the supporting holey carbon foil, two areas in Fig. 3f-g can be distinguished and are marked in the figure: part of the FeRh film sits on the carbon support foil and in the small circular areas (1.2 µm hole diameter) the film is freestanding across the holes of the carbon film. This created a visible pattern, as parts of FeRh film are bending into the holes of the carbon film (Fig. 3g). Furthermore, because of the W electrodes, the film tends to slightly bend toward the central region between the electrodes. Fig. S3 in SI contains a magnified SEM image showing this slight inward-bent and the pattern created by the holey carbon circular holes, and the corresponding TEM image before and after the heating experiment. To make sure that the film was not noticeably damaged during the nanowire-facilitated transfer to the MEMS chip, in situ heating Lorentz TEM measurements were performed to evaluate and check for any changes to the AF to FM transition temperature based on the formation of FM domain walls. Overview and magnified images of the transferred FeRh thin film on the MEMS chip are shown in Fig. 4. The aforementioned two areas in the SEM images (Fig. 3f–g) are also distinguishable and marked in Fig 4b (the FeRh film that sits on carbon foil and over the holes in the carbon foil). The FM domains start to become visible in the Lorentz images of the FeRh film over the holes at around 73 °C and starting around 90–97 °C (Fig. 4c–f) in the part of the FeRh film on the carbon. In Fig. 4c–f we selected an area of the FeRh film on the carbon foil with clear contrast for visualization of the domain wall. To get a better sense of the overall transition and the time/temperature delay of AF-FM transformation between these two areas, a video clip of a heating series with a heating rate of 0.3 °C/min is presented in Fig. S4 in the SI. Fig. 4c is an under-focused Lorentz TEM image at 97 °C where no nucleation and development of the FM state had started and therefore no bright/dark contrast (representing the magnetic domain walls) is manifested on parts of the FeRh film on carbon foil. After the transition is complete with the film fully in the FM state, at around 150 °C, three images were taken with the electron beam in focus (Fig. 4d), over-focused (Fig. 4e), and under-focused (Fig. 4f). The dark and bright lines in the over- and under-focused images (Fig. 4d, e) represent a fully shaped magnetic domain wall that disappear when the sample is in focus (Fig. 4d). This domain wall demonstrates the boundaries of two adjacent domains with different magnetization direction and is part of a complex magnetic microstructure of FM domains on the FeRh film. The complexity of FM domain structure originates from the first-order nature of the AF-FM transition due to the coexistence of AF and FM domains over a wider temperature region, where in addition structural defects influence the nucleation of FM domains as a new phase within the AF domains [44]. In this particular thin film, Rh-rich precipitates (dark contrast dots inside the red rectangle in Fig. 4d) act as nucleation points, where magnetic vortices are formed and eventually influence the final shape and magnetization orientation of the FM domains. We observed and tracked the nucleation and formation of the magnetic domain locally to the point where they merge and form the final magnetic microstructure at the end of the transition to the FM state. The complete nucleation and transformation steps are shown in the video clip in Fig. S4 in the SI. In situ heating Lorentz TEM results confirm that the AF-FM transformation of the areas of the FeRh thin film sitting on the carbon foil are in a relatively good agreement with the transition temperature range of the pristine freestanding film in the magnetization versus temperature (MvT) results in Fig. S2 (SI) exhibiting only a slight elevated transition temperature. However, the transformation of parts of the thin film sitting on the carbon foil holes (freestanding FeRh film) at a lower temperature (∼ 8 K compared with the MvT curve) and the consequent time/temperature delay is mainly because of the straining effect. In part of the film over the holes that are stretched inward (tensile strained) the transition temperature shifts to a lower temperature as strain favors the FM phase. We observed this reduced transition temperature only on these strained areas of the FeRh film, which was to some extent expected as discussed earlier, when explaining the sensitivity of the transition temperature to the strain effect. The presence of the amorphous carbon foil itself (with a low thermal conductivity) can possibly add to this time/temperature delay between the two areas and would result in the freestanding FeRh film over the holes to have an even lower transition temperature compared with the film on carbon foil, in addition to the strain effect. Finally, by taking into account the overall 5% temperature accuracy of the MEMS chips based on their calibration files, the presented transition temperature values of the FeRh film on carbon foil can be consistently correlated with the MvT plot values. 3.3. Additional advantageous applications of nanowire manipulator In addition to the nanowire-facilitated transfer of the FeRh thin film onto a MEMS membrane, where the main challenge is handling a sensitive thin film with minimal ion beam exposure, there are other aspects that make a TEM sample transfer challenging. For example, dealing with irregular geometries, preventing mechanical deformation during placement and bigger/heavier samples. In order to address these challenges and to show how using a flexible yet strong nanowire improves the chances of a successful transfer with minimized damage, contamination, and reduced time for critical steps during the transfer process, selected examples are presented below. The examples include a rigid sample with irregular geometry (varying thickness and shape) 5 Ultramicroscopy 219 (2020) 113075 S. Gorji, et al. Fig. 4. (a) SEM and (b) TEM overview image of the FeRh thin film on a MEMS chip, (c) under-focused Lorentz TEM image at 97 °C. After heating to 150° (above the transition temperature), (d) in-focus, (e) over-focused and (f) under-focused Lorentz images show the FeRh in ferromagnetic state with the dark and bright lines representing a magnetic domain wall in (e) and (f). transferred to a push-to-pull (PTP) device (Fig. 5a–c) and to a MEMS heating chip (Fig. 5d–e), a heavy lamella with an already thinned region (window) transferred to a MEMS chip (Fig. 5f–g), and a conventional lamella lift-out (Fig. 5h–j). The first example shown in Fig. 5a–c (red box) is a thin sample (a MnFeCoNiCr high entropy alloy) prepared by PIPS, right before and after attaching the nanowire, and transfer to a PTP device. To avoid imaging the whole area of the sample with the ion beam, small squared marks are made by the ion beam, similar to the FeRh thin film transfer. The sample is subsequently patterned for an in situ tensile experiment in the TEM (Fig. 5c). The PIPS prepared sample has an irregular geometry with a relatively small usable area of interest, which requires attention when cutting and placing it on the device. Moreover, as the sample is transferred to a rather rigid PTP device, any strain during the placement on the PTP would pre-deform the sample and thus potentially modify the sample and the mechanical test. Previously, a stabilizing transfer mask was used for transferring sensitive thin films to PTP devices [21], but this approach cannot be used here as the geometry of the PIPS sample is wedge-like at the edge, which does not allow for a proper alignment of the transfer mask. However, mechanical deformation can be reduced significantly by using a flexible nanowire, which deforms preferentially. Furthermore, using a nanowire instead of a transfer mask dramatically reduces the damage/contamination and the preparation time during each step of the transfer. Fig. 5d–e (green box) shows the nanowire-assisted transfer of another PIPS prepared MnFeCoNiCu high entropy alloy to a MEMS based in situ heating chip. Here, a small piece of the sample from the edge of the PIPS milled area is cut and attached to the nanowire. A small rectangular hole is prepared in the heating chip, to enable support-free TEM imaging, and the electron transparent region of the sample is positioned carefully across the hole. In this case, the area of interest is close to the contact region. This is limiting the freedom for contacting and requires fast and low-dose operation during transfer and contacting to prevent damage of the area of interest and the MEMS membrane. Therefore, nanowire transfer is probably the only viable approach. The example shown in Fig. 5f–g (blue box) illustrated the transfer of a relatively heavy lamella with an already pre-thinned TEM window to a MEMS based heating chip. Typically, the lamella is first transferred and the final thinning/cleaning is performed with the sample already placed on the membrane [9]. That results in damage of the membrane during the final sample cleaning, which would be critical in case of an in situ liquid or gas cell setup. Here, the flexibility of the nanowire reduces the likelihood for rupture of the MEMS membrane when lowering the specimen on to the membrane without imaging in the FIB to estimate the distance (only using SEM imaging and cutting the nanowire outside the area of interest). Furthermore, Pt deposition induced contamination of the thin TEM window is reduced with the nanowire transfer, thereby reducing the need for post-deposition cleaning of the TEM sample. Alternatively, if available, the pre-thinned lamella could be transferred by ex situ lift-out, to prevent ion-beam and e-beam exposure of the membrane altogether [4,7]. In order to illustrate that the nanowire-transfer is a universal method, we further tested the reliability for conventional TEM lamella transfer. Fig. 5h–j (purple box) shows a cross-sectional TEM lamella liftout procedure of a multilayer STO, SAO and LSMO thin films system. The preparation steps include attaching the lamella to a nanowire, cutting the lamella free (Fig. 5h), and subsequent transfer to a Cu TEM grid (Fig. 5i–j). Unlike in all previous examples, here we used a combination of ion- and e-beam imaging and welding analogous to any standard in situ lift-out preparation to ensure that the nanowire is durable during conventional in situ lift-out. However, the e-beam deposited Pt weld proved sufficiently strong to hold any specimen, e.g., see Fig. 5g, and using ion-beam deposited Pt is in principle not necessary. The nanowire-facilitated transfer proves to be suitable even for heavy lamellae, with the advantage of reduced transfer time and minimized mechanical damage and contamination during the transfer. So far, we have used this technique to successfully transfer a large number of samples with an excellent success rate. In addition, the common problem of sharpening or replacing the micromanipulator tip for routine lift-out preparation can be overcome by using the nanowire as a reusable, easily replaced extension. Cutting and welding a nanowire takes as little as 1–4 s and 3–6 s, and there is no 6 Ultramicroscopy 219 (2020) 113075 S. Gorji, et al. Fig. 5. SEM micrographs of four additional examples for nanowire assisted TEM sample transfer. Example 1 (a-c, red box) shows (a) a PIPS prepared thin sample before attachment to the nanowire, (b) after attaching the nanowire and transfer to a PTP device and (c) after attaching it to a PTP device and patterning for an in situ tensile experiment. The red arrows in (a) are ion milled marks prepared to locate the cutting area without imaging the sample by the ion beam. Example 2 (d-e, green box) shows (d) a PIPS prepared thin sample at the very edge of a hole after attaching the nanowire to the lower left corner, far from the active transparent region and (e) after precisely placing it on a precut region on a MEMS heating chip. Example 3 (f-g, blue box) shows a thinned lamella (f) before and (g) after transferr to a MEMS heating chip while welding the nanowire to the upper left corner to minimize the ion-beam and Pt contamination of the thinned lamella. Example 4 (h-j, purple box) illustrates a traditional TEM lamella lift-out (h) after attaching to the nanowire connected to a thick W probe, (i) FIB image before attaching it to a Cu TEM grid, and (j) SEM image after attaching it to the TEM grid and cutting the nanowire. need to switch from low ion current (for imaging) to higher ion current (often used for cutting), which requires additional time for switching and stabilization. Therefore, the above-mentioned reduced times decrease the overall preparation time when using a nanowire manipulator instead of a conventional W micromanipulator. Finally, the commercial availability of nanowires with different dimensions can be used to adapt to various lamella sizes while maintaining the benefits of the flexibility and strength of a nanowire compared to a stiff micromanipulator. The presented examples unequivocally illustrate the high degree of versatility of the nanowire-assisted transfer technique and its great potential to be used for preparing a variety of sensitive and conventional TEM samples. but only employing a single flexible nanowire that can be stored and reused after each transfer. As a proof of concept, we presented in situ heating TEM results of the AF to FM transition in a FeRh thin film transferred on a MEMS-based chip. The film could be transferred without introducing the typical ion beam damage induced strong reduction of the magnetic transition temperature. We showed that using a nanowire significantly reduces the stress-induced bending of the sample and the MEMS membrane during transfer and attachment of the TEM samples, and how it extremely shortens the time needed to attach (by ebeam-induced Pt) and cut the nanowire with a low ion current. Declaration of Competing Interest 4. Conclusion None. We have introduced a facile nanowire-facilitated transfer method for sensitive TEM samples inside the FIB, featuring a reduced preparation time and minimized ion beam damage, breaking or bending of the sample, and FIB-induced Pt contamination. This was achieved without using any additional nano-manipulation setups inside the FIB, Acknowledgement S. Gorji and A. Kashiwar acknowledge the Doctoral Scholarship provided by the DAAD. L. S. Mantha greatly appreciates the FIB support 7 Ultramicroscopy 219 (2020) 113075 S. Gorji, et al. by Krishna Kanth Neelisetty. [23] Supplementary material Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.ultramic.2020.113075. [24] References [25] [1] T.T. Mueller, E.W. Tsong, Field Ion Microscopy Principles and Applications, Am. Elsevier, New York, 1969. [2] C.a Volkert, a.M. Minor, G. Editors, B. May, Focused Ion Beam Microscopy and Micromachining 32 MRS Bull, 2007, pp. 389–399. [3] D. Tomus, H.P. Ng, In situ lift-out dedicated techniques using FIB-SEM system for TEM specimen preparation, Micron 44 (2013) 115–119, https://doi.org/10.1016/j. micron.2012.05.006. [4] L.A. Giannuzzi, Z. Yu, D. Yin, M.P. Harmer, Q. Xu, N.S. Smith, L. Chan, J. Hiller, D. Hess, T. Clark, Theory and new applications of ex situ lift out, Microsc. Microanal. 21 (2015) 1034–1048, https://doi.org/10.1017/S1431927615013720. [5] L.A. Giannuzzi, J.L. Drown, S.R. Brown, R.B. Irwin, F.A. Stevie, Focused ion beam milling and micromanipulation lift-out for site specific cross-section tem specimen preparation, MRS Proc. 480 (1997) 19, https://doi.org/10.1557/PROC-480-19. [6] W. Zesch, M. Brunner, A. Weber, Vacuum tool for handling microobjects with a nanorobot, Proc. Int. Conf. Robot. Autom. IEEE, 1997, pp. 1761–1766. [7] X. Cen, K. van Benthem, Ion beam heating of kinetically constrained nanomaterials, Ultramicroscopy 186 (2018) 30–34, https://doi.org/10.1016/j.ultramic.2017.12. 005. [8] H. Zheng, Y. Zhu, Perspectives on in situ electron microscopy, Ultramicroscopy 180 (2017) 188–196, https://doi.org/10.1016/j.ultramic.2017.03.022. [9] A. Zintler, U. Kunz, Y. Pivak, S.U. Sharath, S. Vogel, E. Hildebrandt, H.J. Kleebe, L. Alff, L. Molina-Luna, FIB based fabrication of an operative Pt/HfO2/TiN device for resistive switching inside a transmission electron microscope, Ultramicroscopy 181 (2017) 144–149, https://doi.org/10.1016/j.ultramic.2017.04.008. [10] M. Hammad Fawey, V.S.K. Chakravadhanula, M.A. Reddy, C. Rongeat, T. Scherer, H. Hahn, M. Fichtner, C. Kübel, In situ TEM studies of micron-sized all-solid-state fluoride ion batteries: preparation, prospects, and challenges, Microsc. Res. Technol. 79 (2016) 615–624, https://doi.org/10.1002/jemt.22675. [11] J.E. Wittig, J. Bentley, L.F. Allard, J. Bentley, In situ investigation of ordering phase transformations in FePt magnetic nanoparticles, Ultramicroscopy 176 (2017) 218–232, https://doi.org/10.1016/j.ultramic.2016.11.025. [12] N.P. Young, M.A. van Huis, H.W. Zandbergen, H. Xu, A.I. Kirkland, Transformations of gold nanoparticles investigated using variable temperature high-resolution transmission electron microscopy, Ultramicroscopy 110 (2010) 506–516, https:// doi.org/10.1016/j.ultramic.2009.12.010. [13] M. Duchamp, Q. Xu, R.E. Dunin-Borkowski, Convenient preparation of high-quality specimens for annealing experiments in the transmission electron microscope, Microsc. Microanal. 20 (2014) 1638–1645, https://doi.org/10.1017/ s1431927614013476. [14] J.T. van Omme, M. Zakhozheva, R.G. Spruit, M. Sholkina, H.H. Pérez Garza, Advanced microheater for in situ transmission electron microscopy; enabling unexplored analytical studies and extreme spatial stability, Ultramicroscopy 192 (2018) 14–20, https://doi.org/10.1016/J.ULTRAMIC.2018.05.005. [15] H. Hugo, P. Garza, Y. Pivak, L.M. Luna, J.T. Van Omme, R.G. Spruit, M. Sholkina, M. Pen, Q. Xu, MEMS-based sample carriers for simultaneous heating and biasing experiments: a platform for in-situ TEM analysis, 2017 19th Int. Conf. Solid-State Sensors, Actuators Microsystems, 2017, pp. 2155–2158, , https://doi.org/10.1109/ TRANSDUCERS.2017.7994502. [16] R.A. Bernal, R. Ramachandramoorthy, H.D. Espinosa, Double-tilt in situ TEM holder with multiple electrical contacts and its application in MEMS-based mechanical testing of nanomaterials, Ultramicroscopy 156 (2015) 23–28, https://doi.org/10. 1016/j.ultramic.2015.04.017. [17] A. Kobler, A. Kashiwar, H. Hahn, Combination of in situ straining and ACOM TEM : a novel method for analysis of plastic deformation of nanocrystalline metals, Ultramicroscopy 128 (2013) 68–81, https://doi.org/10.1016/j.ultramic.2012.12. 019. [18] R.A. Bernal, T. Filleter, J.G. Connell, K. Sohn, J. Huang, L.J. Lauhon, H.D. Espinosa, In situ electron microscopy four-point electromechanical characterization of freestanding metallic and semiconducting nanowires, Small 10 (2013) 725–733, https://doi.org/10.1002/smll.201300736. [19] Y. Pivak, H.H. Perez-Garza, A. Zintler, L. Molina-Luna, Electrical characterization and failure analysis using operando TEM, Conf. Proc. from Int. Symp. Test. Fail. Anal. 2017-Novem, 2017. [20] M. Canavan, D. Daly, A. Rummel, E.K. McCarthy, C. McAuley, V. Nicolosi, Novel insitu lamella fabrication technique for in-situ TEM, Ultramicroscopy 190 (2018) 21–29, https://doi.org/10.1016/j.ultramic.2018.03.024. [21] A. Kobler, C. Brandl, H. Hahn, C. Kübel, In situ observation of deformation processes in nanocrystalline face-centered cubic metals, Beilstein J. Nanotechnol. 7 (2016) 572–580, https://doi.org/10.3762/bjnano.7.50. [22] Z. Song, Z. Wang, L. Liu, L. Li, V. Koledov, P. Lega, S. Von Gratovsky, D. Kuchin, A. Irzhak, Interaction forces on nanoscale: manipulator-object-surface, 2018 IEEE [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] 8 Int. Conf. Manip. Manuf. Meas. Nanoscale, 3M-NANO 2018 - Proc. 2018, pp. 110–113, , https://doi.org/10.1109/3M-NANO.2018.8552236. A.V. Shelyakov, N.N. Sitnikov, V.V. Koledov, D.S. Kuchin, A.I. Irzhak, N.Y. Tabachkova, Melt-spun thin ribbons of shape memory TiNiCu alloy for micromechanical applications, Int. J. Smart Nano Mater. 2 (2011) 68–77, https://doi. org/10.1080/19475411.2011.567305. D.S. Kuchin, P.V. Lega, A.P. Orlov, V.V. Koledov, A.V. Irzhak, The smallest & the fastest shape memory alloy actuator for micro & nanorobotics, Int. Conf. Manip. Autom. Robot. Small Scales, MARSS 2017 - Proc. 2017, pp. 2–5, , https://doi.org/ 10.1109/MARSS.2017.8001932. J. Lee, S. Kim, Manufacture of a nanotweezer using a length controlled CNT arm, Sens. Actuators A Phys. 120 (2005) 193–198, https://doi.org/10.1016/j.sna.2004. 11.012. X. Wang, X. Liu, H. Zhao, M. Chang, L. Chen, Fabrication of long and smooth tungsten probes for nano-manipulation, Elektron. Ir Elektrotech. 21 (2015) 17–22, https://doi.org/10.5755/j01.eee.21.4.12775. V. Koledov, Nano-manipulation, nano-manufacturing, nano-measurements by new smart material-based mechanical nanotools, 2018 IEEE Int. Conf. Manip. Manuf. Meas. Nanoscale, 2018, pp. 171–176. R. Kometani, S. Ishihara, Nanoelectromechanical device fabrications by 3-D nanotechnology using focused-ion beams, Sci. Technol. Adv. Mater. 10 (2009), https:// doi.org/10.1088/1468-6996/10/3/034501. H. Zhao, M. Chang, X. Liu, J.L. Gabayno, H.T. Chen, Design and implementation of shape memory alloy-actuated nanotweezers for nanoassembly, J. Micromech. Microeng. 24 (2014), https://doi.org/10.1088/0960-1317/24/9/095012. R. Witte, R. Kruk, M.E. Gruner, R.A. Brand, D. Wang, S. Schlabach, A. Beck, V. Provenzano, R. Pentcheva, H. Wende, H. Hahn, Tailoring magnetic frustration in strained epitaxial FeRh films, Phys. Rev. B - Condens. Matter Mater. Phys. 93 (2016) 1–9, https://doi.org/10.1103/PhysRevB.93.104416. T. Edler, S.G. Mayr, Film lift-off from MgO: freestanding single crystalline Fe-Pd films suitable for magnetic shape memory actuation – and beyond, Adv. Mater. 22 (2010) 4969–4972, https://doi.org/10.1002/adma.201002183. M.S. Gorji, D. Wang, R. Witte, X. Mu, R. Kruk, C. Kübel, H. Hahn, In situ Lorentz transmission electron microscopy of FeRh thin films, Microsc. Microanal. 24 (2018) 934–935, https://doi.org/10.1017/s1431927618005160. Y. Chun-Nuan, C. Guan-Ying, M. Xiao-Liang, F. Fang, X. Xiao-Yan, C. Guo-Rong, S. Da-Lin, Growth mechanism of vertically aligned Ag(TCNQ) nanowires, Chin. Phys. Lett. 21 (2004) 1787–1790, https://doi.org/10.1088/0256-307X/21/9/031. X. Marti, I. Fina, C. Frontera, J. Liu, P. Wadley, Q. He, R.J. Paull, J.D. Clarkson, J. Kudrnovský, I. Turek, J. Kuneš, D. Yi, J. Chu, C.T. Nelson, L. You, E. Arenholz, S. Salahuddin, J. Fontcuberta, T. Jungwirth, R. Ramesh, Room-temperature antiferromagnetic memory resistor, Nat. Mater. 13 (2014) 367–374, https://doi.org/ 10.1038/nmat3861. J. Liu, T. Gottschall, K.P. Skokov, J.D. Moore, O. Gutfleisch, Giant magnetocaloric effect driven by structural transitions, Nat. Mater. 11 (2012) 620–626, https://doi. org/10.1038/nmat3334. M. Wolloch, M.E. Gruner, W. Keune, P. Mohn, J. Redinger, F. Hofer, D. Suess, R. Podloucky, J. Landers, S. Salamon, F. Scheibel, D. Spoddig, R. Witte, B. Roldan Cuenya, O. Gutfleisch, M.Y. Hu, J. Zhao, T. Toellner, E.E. Alp, M. Siewert, P. Entel, R. Pentcheva, H. Wende, Impact of lattice dynamics on the phase stability of metamagnetic FeRh: bulk and thin films, Phys. Rev. B 94 (2016) 1–17, https://doi.org/ 10.1103/PhysRevB.94.174435. I. Suzuki, T. Koike, M. Itoh, T. Taniyama, T. Sato, Stability of ferromagnetic state of epitaxially grown ordered FeRh thin films, J. Appl. Phys. 105 (2009) 103–106, https://doi.org/10.1063/1.3054386. K. Aikoh, S. Kosugi, T. Matsui, A. Iwase, Quantitative control of magnetic ordering in FeRh thin films using 30 keV Ga ion irradiation from a focused ion beam system, J. Appl. Phys. 109 (2011), https://doi.org/10.1063/1.3549440. T.P. Almeida, R. Temple, J. Massey, K. Fallon, D. McGrouther, T. Moore, C.H. Marrows, S. McVitie, Quantitative TEM imaging of the magnetostructural and phase transitions in FeRh thin film systems, Sci. Rep. 7 (2017) 1–11, https://doi. org/10.1038/s41598-017-18194-0. M. Rommel, G. Spoldi, V. Yanev, S. Beuer, B. Amon, J. Jambreck, S. Petersen, A.J. Bauer, L. Frey, Comprehensive study of focused ion beam induced lateral damage in silicon by scanning probe microscopy techniques, J. Vac. Sci. Technol. B Nanotechnol. Microelectron. Mater. Process. Meas. Phenom. 28 (2010) 595–607, https://doi.org/10.1116/1.3431085. F. Stumpf, A.A. Abu Quba, P. Singer, M. Rumler, N. Cherkashin, S. SchammChardon, R. Cours, M. Rommel, Detailed characterisation of focused ion beam induced lateral damage on silicon carbide samples by electrical scanning probe microscopy and transmission electron microscopy, J. Appl. Phys. 123 (2018), https:// doi.org/10.1063/1.5022558. Z. Liao, T. Zhang, M. Gall, A. Dianat, R. Rosenkranz, R. Jordan, G. Cuniberti, E. Zschech, Lateral damage in graphene carved by high energy focused gallium ion beams, Appl. Phys. Lett. (2015), https://doi.org/10.1063/1.4926647. L.A. Giannuzzi, T. Clark, Optimizing Van der Waals forces for FIB ex situ lift out, Microsc. Microanal. 23 (2017) 306–307, https://doi.org/10.1017/ s1431927617002215. C. Baldasseroni, C. Bordel, A.X. Gray, A.M. Kaiser, F. Kronast, J. Herrero-Albillos, C.M. Schneider, C.S. Fadley, F. Hellman, Temperature-driven nucleation of ferromagnetic domains in FeRh thin films, Appl. Phys. Lett. 100 (2012), https://doi.org/ 10.1063/1.4730957.