Browsing by Author "Fecher, Sven"

Now showing 1 - 1 of 1

Open Access
Lithium intercalation studies in 2D materials using electrolyte gating
(2022) Fecher, Sven; Smet, Jurgen H. (Dr.)
The aim of this thesis was to investigate lithium intercalation in the single van der Waals gallery of a bilayer of two-dimensional materials driven by electrochemistry. Lithium intercalation is the process of lithium ions being incorporated between the layers of a layered host material. Since these ions carry a charge, energy can be stored this way. The most known application are batteries, where upon charging, lithium is removed from the cathode, a lithium compound, and incorporated at the anode. The commercially most frequently used material for this purpose is a compound consisting primarily of graphite, the layered allotrope of carbon. The elemental building block of graphite is graphene, the two-dimensional, one atom thick allotrope of carbon. It was first isolated in experiments in 2004 by A. Geim and K. Novoselov. They were awarded the Nobel prize in 2010 for their findings. Graphene opened up a wide variety of research due to the ease of production. When thinning down graphite, which is a random process, bilayer graphene can also be isolated. It provides the thinnest possible material suitable for lithium intercalation by offering only one interlayer gap. Thus, it can be considered the most basic building unit of a battery anode. In previous studies, this material has been successfully intercalated with lithium ions using on-chip electrochemistry. By doing so, charged ions in the electrolyte are separated and they accumulate at the interfaces. In the case of a lithium containing electrolyte, lithium ions can additionally intercalate into the material when a voltage above a certain threshold is applied, which leads to doping of the host material. This method is applied in this thesis as well in order to further investigate the process of lithium intercalation into bilayer graphene. The investigation tool of choice here was the SALVE TEM, which enables in-situ recording on the atomic scale due to its outstanding resolution. To perform these TEM studies, modifications of the sample layout became necessary, which are explained in this thesis. The starting point for investigation of lithium intercalation is the standard sample layout developed earlier in the group of Jurgen Smet. It consists of a bilayer graphene flake, patterned into a Hallbar shape and contacted using metal leads. A lithium counter-electrode is used as cathode and an electrolyte is placed connecting the bilayer graphene device with the cathode. The electrolyte only covers a part of the sample in order to exclude the influence of the electrolyte on the investigated portion of the sample. It was possible to see a reversible change of the longitudinal resistance R_xx upon lithiation as well as a change in the Hall resistance R_xy, due to charge induced by the lithium ion intercalates. During measurements, some device degradation was observed after prolonged experiments resulting finally in device failure. We identified reaction of lithium metal used as counter-electrode with residual oxygen as the most likely source of this sample degradation. Therefore, lithium metal was excluded for use as counter-electrode during the sample fabrication process for the remainder of this thesis. Instead a bare Ti or Pt metal counter-electrode was used. This has the disadvantage that the voltage drop applied across the bilayer graphene electrode and the electrolyte droplet is not known, however, empirically this uncertainty can be lived with. Further samples were prepared and transferred into the TEM column in order to test if images with atomic resolution can be acquired. During these early investigations, two main challenges became obvious: The electrolyte droplet, containing hydrocarbons, seemed to outgas in the UHV atmosphere of the sample space and amorphous carbon formed in the beam illuminated regions, preventing proper imaging. The second challenge was that the bilayer graphene lattice developed defects with time because of the co-existence of the high energy electron beam and the carbon contaminants. Extremely clean sample surfaces are required for prolonged imaging. We were able to solve these issues by encapsulating the electrolyte droplet with a 200 nm thick layer of SiO_x, thermally evaporated on top of the drop. This was done with the help of a shadow mask to keep most of the bilayer graphene uncovered. This layer successfully prevented the electrolyte from outgasing and no additional amorphous carbon formed. The carbon residues from processing were removed by current-annealing the sample directly inside the TEM column. A high current is sent through the device which causes strong local heating. This reliably cleans the sample in essence by burning off carbon residuals. After these modifications to the sample fabrication procedure we were able to show that the bilayer graphene samples could still be intercalated with lithium ions and the intercalation was to a large extent reversible. Samples with the modified device layout were then used inside the TEM column to record image series with atomic scale resolution to monitor the intercalation and de-intercalation process in-situ. The images revealed that an additional lattice appears and grows with time. An analysis of the chemical elements present using EELS indicated that only the elements lithium and carbon are available. Since the graphene lattice was also feasible, we conclude that the extra crystalline lattice must consist of pure lithium. The Fourier transform of a real space image yields its image in reciprocal space and thus the lattice information can be extracted from such an image. In the specific sample three lithium crystals with slightly different rotation angles and a lattice constant of 0.31 nm, which is slightly larger than the lattice constant of bilayer graphene 0.246 nm) were identified. By masking the graphene reflexes in the FFT image and back-transforming to real space, the underlying bilayer graphene lattice can be removed to improve the visibility. Additionally, the three different crystal grain rotations can be coloured to further enhance the visibility of the growth process of these lithium crystals. Furthermore, we observed that even within one grain, regions with different contrast appear. This was attributed to a varying thickness of the lithium crystal grain. Hence, the lithium crystal grains are thicker than just one atomic layer. Attempts to measure this thickness by using low-loss EELS suggested a total thickness of 3 - 4 nm. However, the applicability of the usual expressions to convert the signal strength to thickness is questionable since for a two layer thick sample, surface plasmons dominate, normally not considered. Hence, this value only is an upper limit for the thickness. Unfortunately, no other possibility existed to determine the thickness from TEM data. Finally, we were also able to record image series during delithiation, i. e. the removal of lithium from the bilayer graphene host. Hence, lithium crystal growth is a reversible process. We focussed on the nucleation and the time development of the lithium crystal inside the bilayer graphene host. Nucleation starts at defect sites or residuals from fabrication. Typically, these are regions where the lattice is distorted and where the surface energy is higher. The crystal grains exhibit a triangular shape initially. After some time, they still are characterized by sharp edges at the growth front. The rotation angle of a lithium crystal seems to be independent from the orientation of the underlying bilayer graphene lattice. Hence, there appears to be no preferred growth direction and the FFT images show diffraction spots that form a ring. When different grains grow towards each other, they will eventually meet. Three possible scenarios can be distinguished. (1) The crystal grains overlap leading to an increase in thickness visible by a darker appearance in the real space TEM image. (2) One of the grains changes orientation in order to match the other grains orientation and therefore the grains merge into a single grain. This seems to happen only for very small grains, presumably because reorientation of a large grain is energetically too costly. (3) The crystal grains form a grain boundary between them and growth continues in different directions since the interlayer gap is already occupied. The growth rate was determined by a series of images in which the lithium crystal grains fill the entire field of view after about 6 minutes. This led to an estimated rate of 2.45 nm^2/s. TEM images were also taken on triple graphene layer samples. Lithium intercalation basically appears similar and the same mechanism seems to be active. Nucleation again occurs primarily at defects or amorphous residuals on the sample surface. The lithium crystal grains also exhibit a triangular shape and sharp edges after continued growth. Growth itself occurs in every direction without any preferred growth directions. Regions with increased contrast within one grain are visible as well pointing to regions with increased thickness.