Towards reliable electrical measurements of superconducting devices inside a transmission electron microscope

TL;DR

This work demonstrates operando electrical transport measurements of superconducting NbN devices inside a transmission electron microscope using a continuous-flow liquid-helium cryostat. By implementing a modified cryo-shield and carefully controlling electron-beam and objective-lens fields, the authors show that the specimen temperature can approach $8$–$9$ K, close to NbN's $T_{ m c}$, and that thermal radiation dominates unless shielding is optimized. They reveal that electron-beam heating and magnetic-field excitation perturb superconductivity near $T_{ m c}$, while detailed radiative-heat calculations quantify the benefits of reduced imaging apertures for minimizing heating. The study establishes a platform for correlative, low-temperature TEM experiments that combine structural, spectroscopic, and transport measurements to probe the microscopic origins of superconductivity and other quantum phenomena.

Abstract

Correlating structure with electronic functionality is central to the engineering of quantum materials and devices whose properties depend sensitively on disorder. Transmission electron microscopy (TEM) offers high spatial resolution together with access to structural, electronic, and magnetic degrees of freedom. However, electrical transport measurements on functional quantum devices remain rare, particularly at liquid helium temperature. Here, we demonstrate electrical transport measurements of niobium nitride (NbN) devices inside a TEM using a continuous-flow liquid-helium-cooled sample holder. By optimizing a thermal radiation shield to limit radiation from the nearby pole pieces of the objective lens, we achieve an estimated base sample temperature of 8-9 K, as inferred from the superconducting transition temperatures of our devices. We find that both electron beam imaging and the magnetic field of the objective lens perturb the superconducting state, because the base sample temperature is close to the superconducting transition temperature of NbN. Finally, we perform calculations that underscore the importance of cryo-shielding for minimizing thermal radiation onto the device. This capability enables correlative low-temperature TEM studies, in which structural, spectroscopic, and electrical transport data can be obtained from the same device, thereby providing a platform for probing the microscopic origins of quantum phenomena.

Figures (5)

• Figure 1: Measurement setup overview.(a) Photograph of the helium dewar, helium transfer line inserted into the back of the TEM holder, and the arm of the damping stage that supports the helium transfer line. (b) Photograph of the temperature sensor readout, lock-in amplifier, and grounding box. A diagram of the electrical measurement setup is given in Fig. S2. In (a, b) the background hasbeen blacked out to emphasize the equipment used in this work. (c) SEM image of NbN device 1. The numbering of the electrodes is indicated on the image. (d) TEM sample holder with opened tip. The grid cartridge with the TEM grid has been mounted onto the holder. The location of the thermometer is indicated. The TEM sample holder is placed on a holder stand. The inset shows a diagram of the TEM grid, the grid cartridge, and the wire bonds connecting both together. (e, f) Photographs of the TEM sample holder with the tip closed and encased in the (e) regular and (f) modified cryo-shield.
• Figure 2: Resistance measurements of superconducting NbN devices.(a) Resistance plotted against thermometer temperature for device 1. The measurements were performed in the TEM in LTEM mode without electron beam illumination. The black and green data were obtained using the modified and regular cryo-shield, and using test currents of 500 and 100 nA, respectively. The red data were obtained from a separate PPMS measurement. The black and green data were measured during warming and cooling, respectively. The rate of warming can be controlled much more precisely using small over-pressures (0-100 mbar) in the helium dewar. The small number of green data between 7-12 K is due to the fast cooling rate in this temperature range. The fast cooling rate does not affect the presence of a superconducting transition, but would result in an inaccurate measurement of $T_{\mathrm{c}}$ since the thermometer and specimen are not in thermal equilibrium. (b) Resistance plotted against thermometer temperature for NbN device 2. The measurements were performed in a vacuum chamber. The black and green data were obtained with the modified and regular cryo-shield, respectively, using a test current of 100 nA. In this case, the cryo-shield was modified by covering the imaging aperture of the cryo-shield entirely with aluminium. The red data were obtained from a separate PPMS measurement.
• Figure 3: Effect of TEM imaging.(a) Resistance plotted against time during an experiment where the beam is initially blanked for 15 s, and then the beam current is decreased every 15 s by increasing the spot size. The device becomes superconducting between each spot size change because the electron beam is briefly blanked during the change. The electron beam current density is indicated on the figure. The experiment was performed at a thermometer temperature of 6.1 K. (b) Resistance plotted against electron beam current density for four different experiments performed at slightly different thermometer temperatures between 5.9 and 6.6 K. During each experiment, the temperature was stable with a standard deviation $\sim$0.01 K for on the thermometer readings. The inset shows a TEM image of the device.
• Figure 4: Effect of magnetic field.(a) Resistance plotted against time during an experiment where the electron beam is blanked, and the magnetic field is increased in steps of 100 mT every 10 s as indicated on the plot. The thermometer temperature was 5.9 K during the measurement. (b) Average resistance plotted against magnetic field for two different experiments performed at thermometer readings of $T$=5.9 and 6.4 K. The error bars (the standard deviation of the measurements) are larger for the red data compared to the black data because test currents of 100 and 500 nA were used for the red and black data, respectively.
• Figure 5: Thermal radiation heat loads. The red and blue curves show the net radiative power (direct line-of-sight) on the device, $P_{\mathrm{dir}}$, calculated using Eq. 4, for $T_{\mathrm{sh}}=30$ K (blue) and 50 K (red). The radiative power transmitted through the imaging apertures, $P_{\mathrm{in}}$, is plotted in black. The inset shows a diagram of the model. The horizontal dotted line indicates the direct line-of-sight contribution without a cryo-shield, $P_{\mathrm{dir,noshield}}$.