By: John Cenker, University of Washington
Introduction
Under an applied mechanical stress, a crystal will tend to break at large defects in the crystal lattice. By reducing the number of layers, the number of defects reduces; consequently, thinner samples can survive larger strains. Two dimensional materials thus offer an exciting platform for studying the effects of extreme strain approaching the intrinsic strength of the material. However, designing a strain experiment capable of applying such strains in-situ at cryogenic temperatures and with optical access has remained a challenge. In this guide, I will show how to suspend 2D materials over a microscopic gap attached to a piezoelectric strain cell. Due to the small gap size, large strains up to ~ 1.7 % are achieved. The techniques used in this guide and the experimental results arising from them have been published in Nature Nanotechnology.
Parts used in this guide
- Razorbill flexure sample plate (FSP)
- silicon wafer (typically 50-100 µm thick)
- diamond scribe
- STYCAST 2850FT epoxy
- Polydimethylsiloxane (PDMS)
- optical microscope and transfer stage
- CS100 strain cell
Section I: Preparing the gap:
1. The first step in preparing the gap is to mix Stycast 2850 and glue a piece of silicon to the CS1X0 flexure sample plate (FSP). A thin layer of Stycast should be applied to both sides, and the silicon should be gently pushed down on both sides to ensure that it is flat. Tip: efficiency can be greatly boosted by preparing multiple gaps at once.
2. Break the silicon using a diamond scribe. This is done by making a scratch in the silicon over the gap in the titanium sample plate. Then, the wafer is cracked using round carbon fiber tip tweezers. This will form a ~ 3-5 µm crack in the wafer which will serve as our small gap. Since both sides were glued, the height difference across the gap is small. After breaking the silicon, make sure to blow the gap with nitrogen to remove any excess silicon dust.
3. Screw the completed gap into the strain cell using the M2 screws. It is extremely important to check the height uniformity on both sides of the gap to transfer the 2D material on top. To do this, we look at the gap under a 100x microscope and make sure both sides are at the same focus. Note: It might take several attempts to achieve good height uniformity. More reproducible height uniformity can be achieved by making the gap with the sample plate already screwed in. However, this limits the number of gaps which can be prepared at a time.
1. We begin by making PDMS in a petri dish. A relatively thick PDMS should be made (2-4 mm thick). A large rectangle is then cut from the petri dish and placed on a glass slide.
2. We prepare for exfoliation by placing 2-3 bulk crystal flakes on a piece of scotch tape and pulling it apart several (~ 4-5) times. Care should be taken so that the tape has a dense coverage, but with little overlap of the bulk crystals on top of each other.
3. The tape is then placed on top of the PDMS slab, and then quickly peeled off.
The PDMS adhered to the glass slide (left) and the exfoliated 2D material stuck onto the scotch tape (right).
The 2D material is transferred from the scotch tape to the PDMS by pressing the two face-to-face.
The 2D material adhering to the PDMS.
4. After exfoliation, the PDMS slab is searched under the optical microscope and ideal flakes are identified. For our purposes, long flakes are preferred due to the increased adhesion to the substrate. After finding the flake, the rest of the PDMS is cut away using a sharp razor blade, leaving behind a small PDMS rectangle. Note: Due to the design of the Razorbill CS100, the PDMS block must be put in the corner of the glass slide so that the PDMS can make good contact with the gapped substrate without running into the ledge of the CS100 (see below)
Transferring the sample onto the Razorbill strain cell. The above schematic for demonstrates the basic principle for our transfer. The glass slide with PDMS (commonly called the “stamp” in 2D materials community terminology) is aligned with the gap prepared in Section I of this guide, and then lowered using a motor until it comes into contact with the gap. Figure 1 shows a photograph of our transfer stage setup. The optical microscope enables accurate alignment of the exfoliated crystal with the prepared gap (Fig. 1b). Since the van der Waals material adheres to the substrate significantly more than the PDMS, it is deposited across the gap when the stamp is lifted (Fig. 1c).
Figure 1: Transferring 2D material across gap. a) Image demonstrating the transfer process. A motor (off picture left) lowers the stamp until it makes contact with the substrate. b) Microscope image of the flake with the gap in the background. Before touching the substrate, the flake is carefully aligned so that the long axis of the crystal is perpendicular to the gap. c) After raising the PDMS slab, the thin bulk ( ~ 20 nm) CrSBr flake is deposited over the micron-scale gap.
After the sample is transferred across the gap, it is immediately put into the cryostat for measurement.
Additional comments:
- The height uniformity of the gap can be easily tested using an empty stamp – i.e. a PDMS rectangle with no flake. If the height is uniform, the stamp should make contact smoothly over it, with the contact front making complete contact with both sides.
- The technique described above is the simplest and most reliable way to suspend thin bulk vdW flakes. However, in principle, many other transfer techniques can work. There is a vast literature of preparing suspended samples for transmission electron microscopy (TEM) measurements. Further development could enable strain experiments on more complex heterostructures and encapsulated samples.
- Due to the intense strain gradients which arise over the micron-scale gap, this technique is currently most suited for optical measurements, though efforts are underway to adapt it for transport measurements using lithographically patterned contacts.
- The amount of strain applied to the sample is determined by Raman spectroscopy. For more details on this process, including a universal method for calibrating the strain response of 2D materials via a strain gauge heterostructure, please see the Supplementary Information of our recent work using this technique (https://www.nature.com/articles/s41565-021-01052-6).