4670 Willow Rd #100
Pleasanton CA 94588

Frequently Asked Questions about Dynamic and Ultrafast TEM

What is TEM?

TEM stands for either “transmission electron microscope” (an instrument) or “transmission electron microscopy” (a field or general technique), depending on the context. A modern TEM is one of the most powerful tools ever developed for the atomic-resolution study of biological and engineering materials. A TEM focuses a high-energy electron beam through a sample to capture images, diffraction patterns, and spectra, revealing detailed information about atomic structure, nano- and micro-structure, composition, chemical bonding, and even optoelectronic properties.

What is in situ TEM?

While TEM measurements of static samples provide an enormous amount of information (including much of what we know about the nanoscale structure of materials), sometimes there’s just no substitute for watching the full time evolution of a process as it occurs. In situ TEM means doing experiments inside the microscope itself, even while you’re taking high-resolution images or spectra or diffraction patterns of the unfolding process. The field of in situ TEM has advanced greatly in recent years, including the use of heat, focused laser beams, electric currents, mechanical stresses, and flowing gases and liquids inside the microscope.

What is time-resolved TEM?

Time-resolved TEM, which includes DTEM and UTEM (see below), is a general term for in situ TEM with time resolution much better than the millisecond-scale resolution achievable by conventional techniques. Typically this means replacing the continuous-wave (CW) moderate currents of conventional TEM with short pulses of much higher current, typically many milliamperes or more at peak.

Why is time-resolved TEM important?

Because small things change fast. For example, an entire nanoparticle can undergo a reversible phase transformation in nanoseconds or even picoseconds. Conventional time resolution can only show you the starting and ending points of such a transformation, and you’ll miss out on the most interesting details. Does the transformation always start at the same point? How fast does the interface move? Does it immediately go to its final phase, or is there a metastable phase in between? Time-resolved TEM reveals the inner workings of materials on the nanoscale in ways that no other technique can. Such capabilities will only rise in importance as more and more nanoscale technology comes into being.

Why not just use a faster camera?

Because for the most part you won’t see anything. For example, a one-nanosecond exposure with a typical TEM beam current in the nanoampere range will only cause a few electrons to arrive at the camera. So it generally is necessary to increase the beam current considerably compared to conventional TEM. Also, it’s very hard to build a camera that can take images much faster than a few thousand times per second. Nanometer- and micrometer-scale processes tend to have characteristic times ranging from picoseconds to microseconds, so millisecond-scale resolution is often no better than no time resolution at all.

Why pulses? Why not just run many milliamperes of current all the time?

Because you’ll melt the electron gun! Part of the reason TEM can give such high resolution is spatial coherence, which means the size of the electron gun is limited. Running so much current through such a small amount of material dissipates a lot of heat, and the gun is a tiny piece of metal sitting isolated inside a vacuum system. It would melt pretty fast! That said, there are concepts for electron gun designs and data acquisition systems that can get around this problem. The next 10 years of instrument/technique development are looking to be pretty exciting.

How do we generate these electron pulses?

The only well-developed technology that can produce the required beam current, spatial coherence, and short pulse duration is photoemission (and some variations, such as photo-assisted field emission). This means sending a pulsed ultraviolet laser (the “cathode laser”) up to the electron gun, which then liberates anywhere from 1 to many billions of electrons per pulse. The exposure time is then the duration of the electron pulse, which (because of electron beam dynamics, including the fact that electrons repel each other) is a little longer than the duration of the laser pulse.

How do you make your electron pulse come at the right time to catch the process in the act?

We use a second pulsed laser aimed at the sample: the “sample drive laser.” This either heats up or optically stimulates the tiny area we’re looking at in the microscope and initiates the process we’re interested in—typically a chemical reaction, phase transformation, microstructural change, or electronic excitation. The time lag between the sample drive laser and the cathode laser sets the time of the observation. Repeating experiments with different time lags lets us build up a story of how the process evolves in time.

What is DTEM?

Dynamic Transmission Electron Microscopy (DTEM) refers to time-resolved TEM applied to irreversible physical processes. Such a process can’t be reset to its initial state, so you only get one chance to catch it in the act before you have to move to a new position on the sample and try again. This means there must be enough electrons (typically many millions to a few billion) in a single pulse to capture all of the information you want. This is referred to as the “single shot” approach to time-resolved TEM.

What is Movie-Mode DTEM?

Movie-Mode DTEM is an advanced variation of single-shot DTEM that sends a rapid train of up to 16 cathode laser pulses rather than just a single pulse, each of which produces enough electrons to capture a complete image. This lets you capture not just one but 16 images in a span of time as short as half a microsecond. So instead of building up a story of how the process typically evolves from a series of independent experiments, Movie Mode DTEM captures the complete history from a single experiment. This is a huge advantage for the many processes that never unfold exactly the same way twice.

Don't all of those images overlap? How do you get a multi-frame movie?

They would overlap, except the Movie Mode system includes a fast electrostatic deflector after the sample that directs each image onto a different part of a large camera, in a 4×4 array. You can then digitally segment the camera image into a 16-frame movie after the experiment is over. This requires precise synchronization and timing control, since the deflectors have to switch states as quickly as possible between the electron pulses in order to avoid blurring the images.

Is 16 frames enough for typical DTEM applications?

Yes. The prototype Movie Mode DTEM instrument at Lawrence Livermore National Laboratory was only set up to produce 9 frames, and even then, we found that that was more than sufficient for the experiments we were interested in. Remember that we’re watching a rapid, irreversible process that’s being triggered at a precisely known time—so there’s no need to roll the shutters continuously and wait for something to happen, since the vast majority of the time, the sample will be sitting still. The nanosecond-to-microsecond-scale processes that you need a DTEM to see almost never take more than about 20 to 50 microseconds to unfold to their final states, and more typically take only 1 to 2 microseconds. So we’ve found that, in practice, 16 frames turns out to be plenty. Even so, we’re working on ideas for increasing that number and even extending movie-mode to a continuous-acquisition system with unlimited frames.

Do you need a fast camera to do Movie-Mode DTEM?

No, you don’t. You can use just about any TEM camera for this, provided it has enough pixels and a small enough point spread function for each image in the 4×4 array to look good. The time resolution comes in the brevity of the laser pulses and the speed of the electrostatic deflector. The camera itself is typically set to a roughly one-second exposure, and it captures all the electrons that strike the camera during that time. The camera doesn’t even “know” about the time resolution. In fact the repetition rate of the camera doesn’t really even come into play.

What kind of laser produces that Movie-Mode pulse train?

A very special one. Because experiments require a variety of different time resolutions and exposure levels, and because the electron gun performance is much, much better for pulses that have flat temporal profiles, Movie Mode DTEM uses a laser system based on an arbitrary waveform generator (AWG). Such a laser can put out virtually any temporal pattern of pulses and, moreover, can produce optimally shaped pulses so that the electron pulses have flat temporal profiles. Further, each pulse has to have enough ultraviolet photons to produce roughly one billion electrons at the cathode, which means that the laser has to be quite powerful. All of this is controlled in a central computer program that coordinates the pulse trains, deflector switches, and drive laser.

Are the arbitrary temporal profiles really so important?

Yes, they really are. A lot of Movie-Mode DTEM experiments use rather long laser pulses, sometimes even as long as one microsecond. Given the nonlinearity and hysteresis in complex high-energy pulsed lasers, a pulse that’s shaped as a one-microsecond flat-top profile will look like absolutely nothing of the kind after it’s gone through a few stages of laser amplification and frequency conversion. Typically you won’t really have a microsecond pulse at all at the end. Rather, you’ll have a nanosecond-scale pulse with a long tail and an extremely intense initial rise that threatens to damage optics, not to mention the electron gun, unless you reduce the overall intensity considerably. As a result you get poor total electron counts and very poor effective source brightness.

Is DTEM compatible with other in situ techniques?

Yes, if the driving forces can be triggered on the appropriate time scales. DTEM has been combined with liquid and gas environmental systems. It has also been combined with resistive heating, for example preheating the entire sample to some initial condition and then using the sample drive laser to quickly heat a small region of it just above the temperature where the process of interest occurs. There are many potential combinations that can be developed for specific applications.

What is stroboscopic TEM (also called Ultrafast TEM or UTEM)?

UTEM is a technique for studying reversible processes, for example electronic/plasmonic/photonic excitations and certain kinds of phase transformations, in a TEM. Because the process can be repeated many times, each image (or spectrum, or diffraction pattern) can be produced by accumulating signal from many thousands or millions of individual cycles. This allows measurements with atomic spatial resolution and picosecond temporal resolution.

What are some typical applications of DTEM and UTEM?

DTEM excels at the kinds of processes often studied in materials science and materials chemistry, including solid- and liquid-state chemical reactions, microstructural evolution, irreversible phase transformations and their effects on microstructure, and nanoparticle sintering. UTEM is typically aimed more at applications with a materials physics flavor, including plasmonics and reversible phase transformations governed largely by electronic processes (such as charge density wave formation and ferroelectric transformations). The field is still young, and many more applications—especially in biological sciences—are expected to emerge in coming years.

Can a single instrument be configured for both DTEM and UTEM?

Yes! In fact, apart from the choice of lasers and the details of the experimental procedures, the two techniques have so much in common that a single instrument, called a “UDTEM” for “Ultrafast and Dynamic Transmission Electron Microscope,” can achieve excellent performance in both operating modes.

Where can I find more information?

Visit www.ides-inc.com, or contact us at info@ides-inc.com.

Who wrote this FAQ list?

Bryan Reed, Chief Technology Officer of IDES. You can contact me at bryan@ides-inc.com.