Nanocuvette: A Functional Ultrathin Liquid Container for Transmission Electron Microscopy Carl Wadell,* Satoshi Inagaki, Tomiro Nakamura, Ji Shi, Yoshio Nakamura, and Takumi Sannomiya* Department of Materials Science and Engineering, School of Materials and Chemical Technologies, Tokyo Institute of Technology, 4259 Nagatsuta, Midoriku, Yokohama 226-8503 Japan S Supporting Information *
ABSTRACT: Advances in TEM techniques have spurred a renewed interest in a wide variety of research fields. A rather recent track within these endeavors is the use of TEM for in situ imaging in liquids. In this article, we show the fabrication of a liquid cell for TEM observations which we call the nanocuvette. The structure consists of a nanohole film sandwiched by carbon films, sealing liquid in the holes. The hole film can be produced using a variety of materials, tailored for the desired application. Since the fabrication is based on self-assembly, it is both cheap and straightforward. Compared to previously reported liquid cells, this structure allows for thinner liquid layers with better controlled cell structures, making it possible to achieve a high resolution even at lower acceleration voltages and electron doses. We demonstrate a resolution corresponding to an information transfer up to ∼2 nm at 100 kV for molecular imaging. Apart from the advantages arising from the thin liquid layer, the nanocuvette also enables the possibility to study liquid−solid interfaces at the side walls of the nanoholes. We illustrate the possibilities of the nanocuvette by studying several model systems: electron beam induced growth dynamics of silver nanoparticles in salt solution, polymer deposition from solution, and imaging of nonstained antibodies in solution. Finally, we show how the inclusion of a plasmonically active gold layer in the nanocuvette structure enables optical confirmation of successful liquid encapsulation prior to TEM studies. The nanocuvette provides an easily fabricated and flexible platform which can help further the understanding of reactions, processes, and conformation of molecules and atoms in liquid environments. KEYWORDS: transmission electron microscopy, liquid cell, plasmonic nanopore, colloidal lithography, nanoparticles, polymer deposition, molecular imaging
N
introduced that are thin enough for the electrons to pass through without too much scattering by the liquid and still provide sealing to separate the liquid from the vacuum surroundings.4,5 Holders for such liquid cells are nowadays commercially available. These holders utilize silicon chips with 20−50 nm SiN membrane windows on top and bottom, sandwiching the liquid, forming the cell.4,5 In this configuration, operando observations such as in situ electrochemistry are possible utilizing a specialized liquid holder.6−9 However, chipbased liquid cells are typically thick (>100 nm) and made of relatively high atomic number elements, sacrificing spatial resolution due to electron scattering. In practice, storing “large” volumes of liquid with this thin membrane separation right next
anoscale real space imaging in liquid environments is of interest in a wide range of research fields; most biological phenomena take place in aqueous conditions, and many chemical processes are undergone in the liquid phase. Compared to indirect measurement techniques, real space imaging provides good confidence in the results, as one can directly see the processes taking place. This in contrast to indirect methods where the possibility of other processes giving rise to the observed signal has to be first excluded. With the recent advancements in aberration correctors, the transmission electron microscope (TEM) is a technique that provides imaging possibilities with a resolution in the ångström range.1−3 However, imaging of liquids in the TEM has the intrinsic difficulty that the electrons are strongly scattered by liquid molecules and that the liquid needs to be separated from the vacuum environment inside the TEM. Recent chip fabrication technology and material development have overcome some of these problems. Liquid cells have been © 2017 American Chemical Society
Received: July 26, 2016 Accepted: January 30, 2017 Published: January 30, 2017 1264
DOI: 10.1021/acsnano.6b05007 ACS Nano 2017, 11, 1264−1272
Article
www.acsnano.org
Article
ACS Nano to high-vacuum systems also carries a potential risk if the membrane was to rupture by possible mistreatment or mismounting of the samples. For safer observation, a differential pumping system or seal checking system can be introduced. This however requires extra instrumentation or highly experienced users’ skill. Alternative approaches where smaller liquid volumes are contained in the sample have also been reported, such as liquid encapsulation within graphene membranes.10−12 In this case, no special hardware setup is required, although operando observation is no longer possible. The strongest advantage of such a sample-built-in systems is that it allows smaller thicknesses of the liquid layer, thus enabling higher spatial resolution. Using graphene sealing, even atomic scale observation has been realized.10 However, handling of defect-free graphene is tedious, and the liquid cell size and thickness cannot be easily controlled or confirmed. So overall, although the thickness control is critical for both chipbased and sample-built-in liquid cells, properly controlled (or measurable) liquid cells with sufficiently small thickness are currently not available. In this paper, we propose a platform for sample-built-in liquid cells, which we call “nanocuvette”. This structure has the possibility of smaller liquid thickness and can include functions to optically monitor the cell condition. The liquid cell we propose is based on the previously reported ultrathin plasmonic nanopores, but with the addition of sealing carbon membranes on both sides.13 Thanks to the small amount of liquid contained, conventional TEM systems can be used without risking the vacuum. We report here the fabrication of the nanocuvette structure as well as utilize it for some in situ observation of various systems, including: growth of silver particles in salt solution, deposition of polymer in a poly(allylamine hydrochloride) (PAH) solution, and imaging of biomolecules in a liquid environment. We also show how the liquid container condition of the nanocuvettes can be studied using a simple optical measurement.
Figure 1. Schematic illustration of nanocuvette structure (a) and fabrication procedure (b−e). On a substrate with a sacrificial Al and bottom carbon layers, colloidal lithography is used to produce a negative mask layer onto which the desired wall layers are deposited. After lift off of the colloid mask, the structure is transferred to a TEM supporting grid by etching of the sacrificial layer. The nanocuvette structure is finally sealed by a carbon membrane floated on the surface of the desired solution.
RESULTS AND DISCUSSION The fabrication process of the nanocuvette structure is schematically illustrated in Figure 1. In short, colloidal lithography is carried out on a substrate with a sacrificial Al and bottom carbon layers. The colloids form a layer with shortrange order (SRO) on top of the substrate and is subsequently used as a negative mask for the deposition of the wall layers of the nanocuvette structure. One of the strengths of the nanocuvette structure is the flexibility in the wall materials that can be used, depending on the desired application. In this work, we use three different wall layer configurations: AlN single layer, C single layer, or an AlN/Au/AlN multilayer. The benefits of these different structures will be shown later. After removal (lift off) of the colloid mask, the sacrificial Al layer is etched away, and the structure is transferred to a TEM supporting grid. The nanocuvette structure on the TEM grid is then used to fish up a carbon membrane floated on the desired solution, sealing the solution in the nanoholes, completing the structure. For a detailed description of the process, we refer the reader to the Methods section. We first confirmed the successful fabrication of a nanocuvette structure, especially the carbon sealing, by combined scanning electron microscopy-scanning transmission electron microscopy (SEM-STEM). Figure 2 shows the STEM bright-field (BF) image and SEM secondary electron (SE) image of a carbon sealed AlN nanocuvette sample with a partially ruptured
Figure 2. Carbon capped AlN nanocuvette structure where the top carbon sealing was partially ruptured (bright contrast in panel (b)). (a) STEM-BF and (b) SEM-SE images. The white dots in panel (a) correspond to the cuvettes, which are also slightly visible as dark dots in panel (b). The images confirm the successful application of the top carbon sealing layer. Both images were recorded at 30 kV in a combined STEM/SEM. The cuvette diameter is 100 nm.
top carbon sealing. The STEM-BF shows uniformly distributed cuvettes with short-range order (SRO) within the film. A brighter contrast region can be seen in the middle of the image. Looking at the SEM-SE image, it is evident that this brighter region was caused by a rupture in the top carbon layer. The absence of the carbon layer in this region allows more SE production from the heavier elements, and also due to a lower 1265
DOI: 10.1021/acsnano.6b05007 ACS Nano 2017, 11, 1264−1272
Article
ACS Nano
bubble formation reaction, allowing for higher electron doses.16,17 Local conductivity of the specimen also seems to play a role, as there were spots on the samples where bubbles did not form. Similar observations of bubble-like contrast formation have previously been reported.12,18 Inayoshi and Minoda reported a total threshold dose for bubble-like contrast formation at around 500 electrons nm−2 for carbon-based liquid cells.12 Considering the irradiation time, the total threshold dose they report is on the same order as what we observe for our nanocuvettes. We note, however, that we expect a lower threshold for the nanocuvettes, since the volume of liquid contained in a single compartment is smaller in our nanocuvettes (typically
