Letter Cite This: Nano Lett. 2019, 19, 4435−4441
pubs.acs.org/NanoLett
Direct Measurement of the Electron Beam Spatial Intensity Profile via Carbon Nanotube Tomography Matthew D. Zotta, Sharadh Jois, Prathamesh Dhakras, Miguel Rodriguez, and Ji Ung Lee* Colleges of Nanoscale Science and Engineering, SUNY-Polytechnic Institute, Albany, New York 12203, United States
Downloaded via GUILFORD COLG on July 24, 2019 at 00:40:34 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
S Supporting Information *
ABSTRACT: Electron microscopes are ubiquitous across the scientific landscape and have been improved to achieve ever smaller beam spots, a key parameter that determines the instrument’s resolution. However, the traditional techniques to characterize the electron beam have limited effectiveness for today’s instruments. Consequently, there is an ongoing need to develop detection technologies that can potentially measure the smallest electron beam, which is valuable for the continual advancement of microscope performance. We report on a new electron beam detector based on a single-wall carbon nanotube. The nanotubes are atomically smooth, have a well-defined diameter that is similar in size to the finest electron probes, and can be used to directly measure the beam profile. Additionally, by rotating the nanotube in a plane perpendicular to the beam path and scanning the beam at different angles, we can apply tomographic reconstruction techniques to determine the spatial intensity profile of an electron beam accurately. KEYWORDS: Single-wall carbon nanotube, computed tomography, scanning electron microscopy
S
electron beam profile of an in-line critical dimension (CD) SEM.14 A comprehensive wave-optical model for the electron beam in lithography system has been developed to account for the pitfalls of more rudimentary methods.15 Image-based methods exist where a convolution model is implemented. A calibration image is modeled as the convolution of an unknown beam shape and a well-known test structure (i.e., a transmission electron microscope (TEM) image of gold particle islands;16 an electron beam lithography (EBL) test pattern;17 or highly uniform, spherical gold nanoparticles18). Deconvolution methods can then be used to reconstruct the beam shape from the calibration image provided the test structure is well-known16−18 or can be reasonably simulated.17,18 Here, we use a single-walled carbon nanotube (SWCNT) as the smallest detector that can directly profile the beam. We accomplish this by measuring the current in the nanotube as the electron beam scans across it. We place the SWCNT device shown in Figure 1A inside an SEM equipped with nanoprobes to initially measure the distribution of electrons in one scan direction. By scanning the beam at different angles to the nanotube, we reconstruct the beam profile in a manner similar to medical reconstruction in X-ray computed tomography (CT).19,20 A SWCNT has the advantage of being atomically smooth and its diameter can be 0.4 nm or smaller.21−23 While these
canning electron microscopes (SEMs) and scanning transmission electron microscopes (STEMs) are critical to advancing nearly all scientific disciplines including biological sciences, physical sciences, and engineering.1−3 These tools scan a focused beam of electrons across a sample plane to image objects down to nanometer and subnanometer scales. Determining the size and shape of the electron beam at the sample surface can help quantify and improve the resolution limits of these tools.4,5 The methods used to measure the electron beam profile, however, have not kept up with the capabilities of the instruments. A wide range of techniques is available to measure and determine the electron intensity profile,5−18 with verifiable resolution down to only several nanometers, while the actual beam is believed to be less than a nanometer in high-resolution instruments. The most common approach to measure the beam profile uses a knife-edge (KE).5−13 In the KE technique, an electron beam scans over a sharp edge, such as the cleaved edge of silicon or thin tungsten wire, and one measures either the transmitted5−12 or emitted13 signal as a function of position. The KE method, however, becomes increasingly unreliable as the beam spot approaches the nanometer scale because there is no ideal specimen with an atomically sharp edge.8−10,16−18 Furthermore, transmission or scattering off the KE corner obscures the beam profile7,8,12,16−18 making it appear broader than it is. Wavelet analysis of the profile resulting from various defocused line-scans across a simulated silicon nitride stepedge on pure silicon has been used to estimate the defocused © 2019 American Chemical Society
Received: March 25, 2019 Revised: June 13, 2019 Published: June 16, 2019 4435
DOI: 10.1021/acs.nanolett.9b01228 Nano Lett. 2019, 19, 4435−4441
Letter
Nano Letters
To understand the mechanism of the current generation in the nanotubes, we measure the currents from both electrodes that contact the nanotube as the beam scans perpendicularly across the nanotube. If the beam excites electrons and holes in the nanotube, the source and drain contacts will measure equal and opposite currents. Instead, in both metallic and semiconducting nanotubes, we observe a sharp peak in the transient current data, which is consistent with electrons being injected into the nanotube from both contacts. In the discussion that follows, we show that these currents are in response to plasmons that decay in both directions from the point of impact and emit secondary electrons along the nanotube.26 The raw data are shown in Figure 2A. The x-axis is converted from temporal to spatial units by using the scan rate of the electron beam in units of nm/s. From these data, different components of the measured signal can be observed. First, a small voltage difference of few microvolts that results from the preamplifiers induces a small, time-varying current in the nanotube. We call this current IDC. By summing the source and drain current, IDC is eliminated. The result (ISUM) shows an extremely low noise current with resolution in the fA range. ISUM manifests an initial drop then rise in the current as the beam approached the nanotube. When the beam strikes the nanotube, the current drops sharply and recovers to baseline as shown in Figure 2A. The beam is scanned in a line-scan path such that the beam rescans the same line until the beam is blanked. This results in many peaks for a given measurement. ISUM consists of two components: the dominant plasmon response (IP) induced directly by the electron beam and a smaller secondary electron (SE) response (ISE) that is due to the electron beam interaction with the SiO2 substrate at the bottom of the trench. SEs are low energy electrons (
