Polymer-Assisted TEM Preparation
February 1, 2021 – Many new 2D materials oxidize before they reach the microscope. Researchers at Ulm University transfer ultrathin flakes onto the TEM grid with a soluble polymer instead of potassium hydroxide, keeping MnPS3, FePS3, TiSe2 and NbSe2 intact, and imaged a freestanding single layer of NbSe2 with atomic resolution in the Cc/Cs-corrected SALVE microscope at 80 kV.
A flake only a few atoms thick can be ruined before it ever reaches the microscope. Many of the most interesting new 2D materials react with air, water and chemicals within minutes. Transition metal phosphorus trisulfides (TMPTs) such as MnPS3 and FePS3 are layered magnets with promising electrical properties[1],[2] and candidates for future heterostructure devices[3], for instance combined with transition metal dichalcogenides (TMDs)[4]. Aberration-corrected high-resolution transmission electron microscopy (AC-HRTEM) can analyse and even modify such freestanding layers atom by atom[5]. But surface contamination and oxidation hide the pristine material, so everything starts with a clean sample.
Researchers at Ulm University have now developed a preparation method that keeps oxidation-sensitive flakes intact on their way to the TEM grid. It works under ambient conditions and avoids any contact with acids or bases. The team tested it on two TMPTs, MnPS3 and FePS3, and on two TMDs, 1T-TiSe2 and 2H-NbSe2, and compared the results with the method most TEM laboratories use. With the Cc/Cs-corrected SALVE microscope at 80 kV, they then imaged a freestanding single layer of NbSe2 with atomic resolution.
The trouble with potassium hydroxide
Many routes lead from a crystal to a TEM sample: transfer with a supporting polymer[6], direct transfer for graphene[7], polyimide-assisted transfer for hexagonal boron nitride[8] or liquid exfoliation[9]. Most common is the scotch-tape method with potassium hydroxide (KOH), the “KOH method”[10]. Flakes are peeled off a crystal with adhesive tape and pressed onto an oxidized silicon wafer, and the KOH solution then releases them onto the grid. This puts a flake of known thickness exactly where it is wanted. But the process runs in air and involves a highly reactive chemical, so chemical reactions and structural decomposition follow[11]. A stamping method in an oxygen-free glovebox avoids this and yields single layers[12],[13], but it struggles to transfer a freestanding flake onto a TEM grid.
A polymer instead of a base
The new method replaces the base with a polymer that dissolves in a harmless solvent (Fig. 1). A Si/SiO2 wafer is first coated with polyvinyl alcohol (PVA) or poly(methyl methacrylate) (PMMA). Flakes thinned down by adhesive tape are pressed onto this layer, and thin ones are found under the optical microscope. Their thickness can be estimated at this stage from the optical contrast, calculated with the Fresnel equations for a thin layer on the coated wafer[14], so the team knows whether a flake is a single layer or a few layers before it is transferred. According to the supplementary material, an SEM cross-section showed the PVA layer on the 90 nm oxide to be 20–30 nm thick. Because PVA and PMMA refract light almost like SiO2, the calculated contrast of a single NbSe2 layer in green light (530 nm) is 2–6%, and a single layer on such a wafer showed 5% in the green channel of the optical image[14]. Next, a TEM grid is held on the flake with a metal or glass tip and wetted from the outside with a drop of isopropyl alcohol (IPA) or water. As the liquid evaporates, it pulls the flexible carbon membrane of the grid onto the flake. Finally, the polymer is dissolved, PVA in ultrapure water and PMMA in acetone, and the grid floats free and is lifted off with tweezers. At no point does the flake touch an acid or a base. The exact recipe is given in the box below, and the instruments in the box after it.
Rough surfaces and holes
The team first tested what KOH does to bulk MnPS3. In SEM images, an untreated crystal had a smooth surface, while a crystal that had been in contact with the KOH solution had become strongly rough (Fig. 2a). The EDX spectra of both showed carbon and oxygen, which may stem from handling in air (Fig. 2b). But in the untreated crystal the oxygen peak was small compared with those of manganese, sulfur and phosphorus, while after KOH it was by far the strongest. At the same time the sulfur and phosphorus peaks had dropped relative to manganese, and a potassium peak had appeared. The KOH had oxidized the crystal and decomposed it. Cross-sections milled into the crystals (supplementary material) showed that the oxidation starts at the surface and continues into the flake, in MnPS3 as in FePS3.
The difference grew in thin flakes. A few-layer MnPS3 flake prepared with PVA was intact, with sharp edges, and its AC-HRTEM image and Fourier transform (FFT) showed a single crystal viewed along the [103] direction (Fig. 2c). A flake prepared with KOH was heavily damaged: the overview image showed a broken-up sheet, and the HRTEM image showed holes and a largely amorphous, polycrystalline structure, as the ring-like FFT confirmed (Fig. 2d). The supplementary data show the same for FePS3: after KOH, its EDX spectrum had a larger oxygen peak while the phosphorus and sulfur peaks almost vanished, and a few-layer flake prepared with KOH was strongly degraded, while one prepared with PMMA stayed crystalline along [103].
A systematic test in the supplementary material checked the liquids themselves. In KOH solution, MnPS3 and FePS3 flakes visibly deteriorated, followed for MnPS3 over up to four hours. In ultrapure water and in IPA, the liquids of the polymer method, the flakes were basically unaffected even after more than two weeks. The iron compound oxidizes much faster than the manganese one.
Oxygen in the spectrum
Could the method also protect TMDs that oxidize even more easily? The team prepared few-layer 1T-TiSe2 three ways: with KOH, with PVA and with PMMA (Fig. 3). The two polymers were used to find out whether the result depends on the polymer and its solvent, water or acetone. In the overview images, the KOH sample had a frayed, contaminated edge, while both polymer samples had sharp edges. The Cs-corrected HRTEM images at 80 kV along [001] showed the crystalline TiSe2 lattice for PVA and PMMA, but a mostly amorphous structure for KOH, as its FFT also showed at low spatial frequencies. SEM of the bulk crystals (supplementary material) told the same story: after KOH, the TiSe2 surface had become rough and covered with clusters that EDX identified as selenium, the oxygen signal had risen and the selenium signal fallen, and cross-sections showed the oxidation proceeding from the surface inwards. NbSe2 crystals formed selenium clusters as well.
EELS made the oxidation directly visible. Only the KOH sample showed an oxygen K edge starting at 530 eV[15], and its titanium L3 edge was split, a sign that Ti–O bonds had formed. In the PVA and PMMA samples, the Ti L2,3 edge kept the shape known for pristine 1T-TiSe2[16]–[18], and no oxygen edge was detected. Between PVA and PMMA there was no significant difference. How the oxide shows up in the titanium edge is explained in the box below.
A freestanding single layer
The decisive test was a single layer. The team chose 2H-NbSe2, a superconductor, whose properties depend so much on its structure that excellent samples are crucial for understanding them[22] (Fig. 4a). NbSe2 nanosheets are known to be sensitive to oxygen, which changes their electronic properties markedly[23]. Atomically resolved (S)TEM studies of single-layer NbSe2 had therefore only been possible inside a graphene-encapsulated three-layer stack, because a freestanding single layer oxidized too fast[24],[25]. With the SALVE microscope at 80 kV, the team imaged few-layer and single-layer NbSe2 prepared with PVA, both with atomic resolution and without signs of deterioration (Fig. 4b, c). In the single layer, one defect stood out: a double selenium vacancy, in which both selenium atoms of one site are missing[22]. Because no lattice of a second layer shows through the gap, it also proves that the flake is a true single layer. “Here, atomically-resolved AC-HRTEM characterization of freestanding single-layer 2H-NbSe2 is reported for the first time,” the authors write.
The box below collects all comparisons, between the two preparation methods and with earlier measurements on pristine material.
Challenges named by the authors
The authors name the limits of their method. Because it runs under ambient conditions, slight oxidation of the PVA and PMMA samples is to be expected as well. Even water is not entirely harmless for the most sensitive material: FePS3 flakes kept in water for two weeks developed small rust-coloured islands, a sign of slight surface oxidation. In the TEM samples, since no oxygen edge appeared in EELS, the amount of oxidation must lie at least below the detection limit of the settings used. At 0.6 eV energy resolution, the titanium oxide in the KOH sample could not be identified unambiguously. And the atomic defects seen in NbSe2 may come from oxidation during preparation as well as from the electron beam during imaging, so the method does not remove every source of damage.
Clean samples for sensitive materials
The study shows that the final step of the preparation decides whether an oxidation-sensitive 2D material can be studied in its pristine state. Replacing the base with a water- or acetone-soluble polymer removed the oxidation that the KOH method caused, for TMPTs as well as for TMDs, for few layers as well as for a single layer. “In conclusion, our new polymer-assisted TEM sample preparation method provides an enhanced sample quality, enabling atomically precise analysis of structural and chemical properties of oxidation-sensitive freestanding 2D layers,” the authors conclude. Since it needs no glovebox, it runs in ordinary air.
Resource:
Köster, J., Liang, B., Storm, A., Kaiser, U. (2021).
Polymer-assisted TEM specimen preparation method for oxidation-sensitive 2D materials.
Nanotechnology 32(7), 075704.
https://doi.org/10.1088/1361-6528/abc49e
Update – August 15, 2022
The news article of February 1, 2021 ends above. Later Ulm studies that prepared their samples with the polymer-assisted method are listed below.
Devices with defects
A study of atomic defects in 2D nanosheets for electronic and optical sensing used the PVA transfer for its TEM samples and extended it by one step, a thin copper layer sputtered onto the grid. A KOH-based transfer was out of the question there, because KOH oxidizes the copper[26] (see the news article 2D Nanosheet Device Integration).
Update – September 1, 2022
Electron beams on TMPTs
An Ulm study of how the electron beam modifies few-layer TMPTs named clean, freestanding samples made by polymer-assisted exfoliation as its prerequisite and prepared MnPS3 this way[27] (see the news article Chemical Modification of 2D TMPTs).
Update – February 20, 2023
MnPS3 under beam and heat
For a study of how electron irradiation and heating transform few-layer MnPS3, the flakes were pressed onto PVA-coated silicon and transferred to the TEM grid with distilled water[28] (see the news article Few Layer MnPS3).
Update – December 19, 2023
MnPSe3
The same route gave few-layer MnPSe3 flakes that appeared highly crystalline and free of surface contamination in the 80 kV HRTEM images[29] (see the news article Few Layer MnPSe3).
Download this news article as PDF – for reading on paper, taking notes or filing.
-
Lee, J.-U., Lee, S., Ryoo, J. H., Kang, S., Kim, T. Y., Kim, P., Park, C.-H., Park, J.-G., Cheong, H. (2016). Ising-Type Magnetic Ordering in Atomically Thin FePS3. Nano Lett., 16, 7433–7438. https://doi.org/10.1021/acs.nanolett.6b03052
-
Lee, S., Choi, K.-Y., Lee, S., Park, B. H., Park, J.-G. (2016). Tunneling transport of mono- and few-layers magnetic van der Waals MnPS3. APL Mater., 4, 086108. https://doi.org/10.1063/1.4961211
-
Gibertini, M., Koperski, M., Morpurgo, A. F., Novoselov, K. S. (2019). Magnetic 2D Materials and Heterostructures. Nat. Nanotechnol., 14, 408–419. https://doi.org/10.1038/s41565-019-0438-6
-
Huang, H., Song, J., Yu, D., Hao, Y., Wang, Y., Peng, S. (2020). Few-Layer FePS3 Decorated With Thin MoS2 Nanosheets for Efficient Hydrogen Evolution Reaction in Alkaline and Acidic Media. Appl. Surf. Sci., 525, 146623. https://doi.org/10.1016/j.apsusc.2020.146623
-
Zhao, X., Kotakoski, J., Meyer, J. C., Sutter, E., Sutter, P., Krasheninnikov, A. V., Kaiser, U., Zhou, W. (2017). Engineering and Modifying Two-Dimensional Materials by Electron Beams. MRS Bull., 42, 667–676. https://doi.org/10.1557/mrs.2017.184
-
Lin, J., Lin, Y.-C., Wang, X., Xie, L., Suenaga, K. (2016). Gentle Transfer Method for Water- and Acid/Alkali-Sensitive 2D Materials for (S)TEM Study. APL Mater., 4, 116108. https://doi.org/10.1063/1.4967938
-
Regan, W., Alem, N., Alemán, B., Geng, B., Girit, Ç., Maserati, L., Wang, F., Crommie, M., Zettl, A. (2010). A Direct Transfer of Layer-Area Graphene. Appl. Phys. Lett., 96, 113102. https://doi.org/10.1063/1.3337091
-
Pacilé, D., Meyer, J. C., Girit, Ç. Ö., Zettl, A. (2008). The Two-Dimensional Phase of Boron Nitride: Few-Atomic-Layer Sheets and Suspended Membranes. Appl. Phys. Lett., 92, 133107. https://doi.org/10.1063/1.2903702
-
Nicolosi, V., Chhowalla, M., Kanatzidis, M. G., Strano, M. S., Coleman, J. N. (2013). Liquid Exfoliation of Layered Materials. Science, 340, 1226419. https://doi.org/10.1126/science.1226419
-
Meyer, J. C., Girit, C. O., Crommie, M. F., Zettl, A. (2008). Hydrocarbon Lithography on Graphene Membranes. Appl. Phys. Lett., 92, 123110. https://doi.org/10.1063/1.2901147
-
Her, M., Beams, R., Novotny, L. (2013). Graphene Transfer With Reduced Residue. Phys. Lett. A, 377, 1455–1458. https://doi.org/10.1016/j.physleta.2013.04.015
-
Zomer, P. J., Dash, S. P., Tombros, N., van Wees, B. J. (2011). A Transfer Technique for High Mobility Graphene Devices on Commercially Available Hexagonal Boron Nitride. Appl. Phys. Lett., 99, 232104. https://doi.org/10.1063/1.3665405
-
Frisenda, R., Navarro-Moratalla, E., Gant, P., Pérez De Lara, D., Jarillo-Herrero, P., Gorbachev, R. V., Castellanos-Gomez, A. (2018). Recent progress in the assembly of nanodevices and van der Waals heterostructures by deterministic placement of 2D materials. Chem. Soc. Rev., 47, 53–68. https://doi.org/10.1039/C7CS00556C
-
Blake, P., Hill, E. W., Neto, A. H. C., Novoselov, K. S., Jiang, D., Yang, R., Booth, T. J., Geim, A. K. (2007). Making Graphene Visible. Appl. Phys. Lett., 91, 063124. https://doi.org/10.1063/1.2768624
-
Bastow, T. J., Botton, G. A., Etheridge, J., Smith, M. E., Whitfield, H. J. (1999). A Study of Li2TiOSiO4 and Li2TiOGeO4 by X-Ray Powder and Electron Single-Crystal Diffraction, 17O MAS NMR and O K-Edge and Ti L2,3-Edge EELS. Acta Crystallogr. A, 55, 127–132. https://doi.org/10.1107/S0108767398006990
-
Qiao, Q., Zhou, S., Tao, J., Zheng, J.-C., Wu, L., Ciocys, S. T., Iavarone, M., Srolovitz, D. J., Karapetrov, G., Zhu, Y. (2017). Anisotropic Charge Density Wave in Layered 1T-TiSe2. Phys. Rev. Mater., 1, 054002. https://doi.org/10.1103/PhysRevMaterials.1.054002
-
Kobayashi, K., Yasuda, H. (2018). Formation of a Superstructure in 1T-TiSe2 Induced at Room Temperature by Electron Beam Irradiation. Mater. Res. Express, 5, 085006. https://doi.org/10.1088/2053-1591/aad222
-
Monney, C., Zhou, K. J., Cercellier, H., et al. (2012). Mapping of Electron-Hole Excitations in the Charge-Density-Wave System 1T-TiSe2 Using Resonant Inelastic X-Ray Scattering. Phys. Rev. Lett., 109, 047401. https://doi.org/10.1103/PhysRevLett.109.047401
-
Fronzoni, G., De Francesco, R., Stener, M., Causà, M. (2006). X-Ray Absorption Spectroscopy of Titanium Oxide by Time Dependent Density Functional Calculations. J. Phys. Chem. B, 110, 9899–9907. https://doi.org/10.1021/jp057353a
-
Sun, L., Chen, C., Zhang, Q., et al. (2017). Suppression of the Charge Density Wave State in Two‐Dimensional 1T‐TiSe2 by Atmospheric Oxidation. Angew. Chem. Int. Ed., 56, 8981–8985. https://doi.org/10.1002/anie.201612605
-
de la Peña, F., Berger, M.-H., Hochepied, J.-F., Dynys, F., Stephan, O., Walls, M. (2011). Mapping Titanium and Tin Oxide Phases Using EELS: An Application of Independent Component Analysis. Ultramicroscopy, 111, 169–176. https://doi.org/10.1016/j.ultramic.2010.10.001
-
Xi, X., Zhao, L., Wang, Z., Berger, H., Forró, L., Shan, J., Mak, K. F. (2015). Strongly Enhanced Charge-Density-Wave Order in Monolayer NbSe2. Nat. Nanotechnol., 10, 765–769. https://doi.org/10.1038/nnano.2015.143
-
Park, H., Kim, J. Y., Oh, J. Y., Lee, T. I. (2020). Long-Term Stable NbSe2 Nanosheet Aqueous Ink for Printable Electronics. Appl. Surf. Sci., 504, 144342. https://doi.org/10.1016/j.apsusc.2019.144342
-
Wang, H., Huang, X., Lin, J., et al. (2017). High-Quality Monolayer Superconductor NbSe2 Grown by Chemical Vapour Deposition. Nat. Commun., 8, 394. https://doi.org/10.1038/s41467-017-00427-5
-
Nguyen, L., Komsa, H.-P., Khestanova, E., et al. (2017). Atomic Defects and Doping of Monolayer NbSe2. ACS Nano, 11, 2894–2904. https://doi.org/10.1021/acsnano.6b08036
-
Quincke, M., Lehnert, T., Keren, I., et al. (2022). Transmission-Electron-Microscopy-Generated Atomic Defects in Two-Dimensional Nanosheets and Their Integration in Devices for Electronic and Optical Sensing. ACS Appl. Nano Mater., 5, 11429–11436. https://doi.org/10.1021/acsanm.2c02491
-
Köster, J., Storm, A., Ghorbani-Asl, M., et al. (2022). Structural and Chemical Modifications of Few-Layer Transition Metal Phosphorous Trisulfides by Electron Irradiation. J. Phys. Chem. C, 126, 15446–15455. https://doi.org/10.1021/acs.jpcc.2c03800
-
Storm, A., Köster, J., Ghorbani-Asl, M., et al. (2023). Electron-Beam- and Thermal-Annealing-Induced Structural Transformations in Few-Layer MnPS3. ACS Nano, 17, 4250–4260. https://doi.org/10.1021/acsnano.2c05895
-
Storm, A., Köster, J., Ghorbani-Asl, M., et al. (2023). Structural Transformations in Few-Layer MnPSe3 Stimulated by Thermal Annealing and Electron Irradiation. J. Phys. Chem. C, 127, 24713–24723. https://doi.org/10.1021/acs.jpcc.3c07112




