In Situ Electron Microscopy Techniques Part 1 (Nanotechnology)

INTRODUCTION

For decades, electron microscopies, transmission electron microscopy especially, have been employed as an important tool in materials research studies in nanoscience and developments in nanotechnology, for their powerful analytical techniques at spatial resolution and also as a kind of microlaboratory in which dynamic studies have been conducted ”in situ.” Such in situ or real-time studies have included those involving phase transformations, mechanical deformation and failure, irradiation-related phenomena, and a host of others involving changes in physical, chemical, magnetic, and electrical structure, microstructure, and properties. This article will present a number of examples from in situ studies to illustrate how such electron microscopy-based studies may contribute, often in unique ways, to the understanding of the behavior of material systems of relevance to modern nanoscience and nanotechnology. We will also review briefly a number of analytical techniques that have been developed over the years and applied to electron microscopies and finally comment on potential impact ongoing developments, particularly the correction of electron optical aberrations, will have on in situ studies.

OVERVIEW

The initial demonstration of an in situ experiment with photographic record is often ascribed to an English-born photographer, Eadweard Muybridge. In the 1870s and 1880s, in part under the auspices of Leland Stanford, then Governor of California, Muybridge employed a series of photographic cameras triggered sequentially as a horse trotted by. The resulting sequence of photographs clearly demonstrated that the depiction of horses in motion in paintings and in the common belief of the time had been anatomically incorrect; that, rather than the horse’s gait being side to side with one pair of legs at a time, all four legs come off the ground at once as a horse trots or gallops (For example, see Ref. [1]).


An in situ experiment is somewhat like watching an entire football match, rather than trying to imagine what happened by checking the score occasionally. For systems of almost any complexity or with events occurring rapidly as in the trotting horse example, it is not always possible to correctly deduce the evolution of changes simply by observing the end results in relation to the initial and final or sometimes even selected intermediate conditions. Furthermore, a well-designed in situ experiment may be an extremely efficient way to perform a scientific or technological study.

For many years in situ experiments have been performed involving a host of different analytical tools including X-ray diffractometers and synchrotron radiation sources, optical microscopes, and electron microscopes and have involved techniques of observation and measurement in real time or near real time such as microscopies (i.e., imaging of events), diffraction, elemental microanalysis, and changes in various physical properties. This article focuses mainly on the application of transmission and scanning transmission electron microscopies (TEM and STEM), and to a lesser extent on scanning electron microscopy, for in situ experiments with particular emphasis on applications related to nanoscience and nanotechnology. The topic of related instrumentation such as scanning tunneling microscopy (STM) and atomic force microscopy (AFM) is presented elsewhere in this volume.

In situ techniques are often very powerful in yielding insight into the mechanisms of real-time behavior, such as when two or more phenomena compete in producing changes or when highly localized short-term changes are superimposed on long-term trends. A further situation in which in situ observation is essential is when elevated or cryogenic temperatures, loads, or fields are employed in a materials study, and further changes in the experimental object would occur as a consequence of returning the material to room temperature, zero load, or zero field for subsequent observation and analysis. The same may be said for all types of in situ studies, X-ray, neutron, and electron.

In situ electron microscopy studies should be regarded as complementary to similar synchrotron studies and, for very slowly varying phenomena, neutron scattering studies. The latter present a view of time-dependent phenomena, generally averaged over much larger volumes of observation. It is, however, sometimes difficult to convince people who have extensive experience with X-ray or neutron scattering studies of the value-added aspect of complimentary analyses by some form(s) of electron microscopy, including the application of in situ techniques. This complementarity of techniques is emphasized in Table 1 in terms of selected characteristics for each of these types of radiation sources.

Table 1 Comparison of selected (approximate) characteristics of state-of-the-art neutron, x-ray, and electron sources employed for materials research 

Type of radiation

Source brightness (particles/cm2/sr/eV)

Elastic mean-free path (nm)

Absorption length (nm)

Minimum

probe size (nm)

Spatial resolution (nm)

Neutrons (SNS)

1014

107

108

106

106

X-rays (APS)

1026

103

105

102

102

Electrons (S)TEM

1029

101

102

10-1

10-1

The elastic mean free paths for neutrons and X-rays are large compared to that for electrons, which means that elastic scattering, for example, is relatively inefficient in the former requiring very high intensities or long data acquisition times to achieve reasonable spatial resolutions. On the other hand, small absorption lengths for electrons require very thin specimens, but because electrons are readily focused by electromagnetic and electrostatic lenses, very fine, intense probes may be formed which, coupled with the thin specimen, result in excellent spatial resolution. The bottom line is simply that these related techniques can and should be more often understood as complementary rather than competitive.

EXAMPLES OF IN SITU ELECTRON MICROSCOPY STUDIES

It may be easier to grasp the value of applying in situ techniques in electron microscopy by considering a variety of representative examples. By no means will these form an exhaustive compilation of such studies. The examples presented are brief excerpts from in situ TEM or STEM studies. However, some of these same ideas may be adapted or have been adapted for in situ SEM studies. Furthermore, variable pressure or environmental SEMs offer other altogether unique aspects for in situ experiments.

Magnetic Studies

The techniques for imaging and analysis of magnetic structures in TEM have been developed over many decades. These include Fresnel, Foucault, electron holographic, and differential phase contrast imaging. The

Fresnel technique is an off-focus technique which, for example, images magnetic domain walls as bright or dark lines due to an excess or deficiency of local image brightness. Foucault imaging is a dark field technique (i.e., employing scattered rather than transmitted electrons) in which the objective aperture selects a segment of a split deflected beam, which is due to different local magnetization directions within the imaged area. Electron holography is a field-imaging technique which results from the interaction of bright field specimen waves with a reference wave usually from an adjacent open area of specimen. Differential phase contrast imaging is a STEM technique employing position-sensitive detection.

For an in situ TEM investigation of the effect size of elliptical disk-shaped particles of a Permalloy (Ni+19 at.% Fe), arrays of such particles 8 nm thick were prepared employing electron beam lithography and a 30-nm Si3N4 support film. Four such particles are shown in Fig. 1, whose aspect ratios (major axis dimension/minor axis dimension) are 2:1 and whose major axes vary from 3.4 to 5.0 mm. This system of particles has been magnetically saturated in the direction indicated by the large arrow and then is subjected to an increasing applied field in the opposite sense (smaller arrows) as the specimen is imaged using the Fresnel technique (also known as Lorentz microscopy), the resulting contrast of which is sensitive to local variations of a particle’s magnetic induction and its gradients in the horizontal plane. The interaction of such variations with the incident electrons produces complementary image contrast when the objective lens is under- or overfocused. The series of images in Fig. 1 shows the dependence and sequential details of the reverse magnetization processes in the four particles. Image defocus is constant. While the geometrical demagnetization factor is the same for the four disks (constant aspect ratio), the observed differences for a given applied field are due, in part, to the role of the resulting demagnetizing fields associated with the different intrinsic distributions of magnetic moment within each particle due to size (the total energy of individual particles must be minimized). In this experiment a commercially available Lorentz objective lens is employed in which the usual field of the objective may be switched off and a pair of minilenses is employed in its place, whose fields can be set to cancel at the specimen position; a special specimen holder incorporates a set of miniature coils to produce the required variable magnetic field in the plane of the specimen.[2] In such a situation, image resolution is limited to several nanometers (Josef Zweck, personal communication, 2003).

There are a number of useful although sometimes less quantitative examples of in situ magnetic studies in which magnetic structure changes of a specimen are achieved by tilting the specimen in the objective lens field, which results in an increasing component of applied field in the plane of the specimen. An example of this is the following study of magnetization reversal in a series of patterned interconnected rings as shown in Fig. 2. Again the imaging technique involves a form of Lorentz microscopy but this time in a scanning transmission electron microscope (STEM). The material is sputter-deposited Co on a thin silicon nitride substrate. In the actual experiment the specimen was tilted in 1° increments about an axis parallel to the ring array. Fig. 2 shows Lorentz images at four stages in the magnetization reversal process. Many essential features of this process may be deduced unambiguously from the series of images; for example, the white and dark lines at the ring intersections in Fig. 2a and d are 90° domain walls (Nestor Zaluzec and Natali Metlushko, personal communication, 2003).

Magnetization reversal in elliptical disks 30 nm thick and major axes ranging from 3.4 to 5.0 mm; Ni+19 at.% Fe and constant aspect ratios 2:1. Following saturation in one easy direction (large arrow), magnetization reversal process under oppositely directed applied field (small arrows) depends on size, the larger disks reversing at smaller fields. Lorentz TEM.

Fig. 1 Magnetization reversal in elliptical disks 30 nm thick and major axes ranging from 3.4 to 5.0 mm; Ni+19 at.% Fe and constant aspect ratios 2:1. Following saturation in one easy direction (large arrow), magnetization reversal process under oppositely directed applied field (small arrows) depends on size, the larger disks reversing at smaller fields. Lorentz TEM.

Reversal of the magnetic domain state of a row of sputter-deposited Co rings on silicon nitride. In-plane field component normal to row increases from (a) to (d), reversing magnetic state. Lorentz STEM.

Fig. 2 Reversal of the magnetic domain state of a row of sputter-deposited Co rings on silicon nitride. In-plane field component normal to row increases from (a) to (d), reversing magnetic state. Lorentz STEM.

Electrical Property Studies

Studies analogous to those for magnetic materials may also be performed in situ for ferro- and ferrielectric material systems, which are in some ways more straightforward because the applied fields are electric rather than magnetic fields. The magnetic field of the objective lens has little effect on electric polarization processes.

Another type of electrical property study with simultaneous TEM is a quantized conductance study of gold wires of nanometer and subnanometer cross section (quantum point contacts). In this instance a miniaturized STM was incorporated as part of the specimen holder of an ultrahigh vacuum high-resolution TEM. By means of the shear-type piezoelectric positioner, a sharpened gold tip was brought into contact with a gold island and slowly withdrawn at constant speed under computer control. Structural changes were observed continuously and video recorded as indicated in Fig. 3a-f, and electrical conductance was measured simultaneously as shown in Fig. 3g. In the same work a single row of four gold atoms (the ultimate wire) parallel to [100] was formed whose conductance was ~ 13 kQ~1 corresponding to 2e2/h, the quantized conduction unit. In addition the conductance of a double row was twice this value, showing that equipartition holds for electron transport in these quantum systems. The high-resolution microscopy was clearly essential to the analysis of the conductance data.[3]

Mechanical Property Studies

The deformation and fracture behavior of metals and alloys represents one of the most common applications for in situ microscopy studies since the 1960s, often having been performed in high-voltage TEM because of greater thickness for observation afforded by the higher energy electrons. In our context, the goal of such studies is to determine the fundamental mechanisms controlling the mechanical response of nanostructured systems and to correlate it with the observed mechanical behavior. For example, to achieve this goal a novel MEMS microtensile test device has been developed which allows measurement of the applied load and sample displacement with concurrent observation of the deformation mechanisms. The device is such that it can be used in a wide range of instruments, although it has been used primarily in SEM and TEM studies. The device consists of a thin dogbone-shaped test specimen that is patterned and cofabricated with microforce and displacement sensors and supports made of silicon using standard photolithographic and deep reactive ion etching techniques. The support beams ensure true uniaxial loading. Fig. 4 describes the fabricated device. The specimen is freestanding between the beams, allowing its inherent response to loading to be determined, including details of the deformation and fracture mechanisms. Because the specimen is integrally fabricated with the loading device structure, issues of attaching it to the test frame and aligning it to ensure uniaxial loading are mitigated.[4] Fig. 5 is a series of TEM micrographs showing the glide of a dislocation from the interior of a 200-nm-thick grain to a grain boundary into which it is absorbed, leaving no trace. This demonstrates in a very elementary way the need for such dynamic in situ studies (Ian Robertson, private communication, 2003).

Quantum conductance of [110] rows of gold atoms. (a)-(e) Gold tip is withdrawn from gold sample, causing continuous reduction in the number of vertical gold atom rows to fracture at (f), with simultaneous conductance measurement. Note dislocation in (a). At (e) conductance is ~ 2 x (13 kfl)~ \ two times unit conductance (unit conductance = 2 e2/h). (g) Quantized conductance measured during withdrawal of the gold tip [as in (d)-(f)].

Fig. 3 Quantum conductance of [110] rows of gold atoms. (a)-(e) Gold tip is withdrawn from gold sample, causing continuous reduction in the number of vertical gold atom rows to fracture at (f), with simultaneous conductance measurement. Note dislocation in (a). At (e) conductance is ~ 2 x (13 kfl)~ \ two times unit conductance (unit conductance = 2 e2/h). (g) Quantized conductance measured during withdrawal of the gold tip [as in (d)-(f)].

Schematic drawing and SEM micrograph of tensile testing chip.

Fig. 4 Schematic drawing and SEM micrograph of tensile testing chip.

Dislocation motion and incorporation into a grain boundary in an Al crystal under increasing tension in device described in Fig. 4.

Fig. 5 Dislocation motion and incorporation into a grain boundary in an Al crystal under increasing tension in device described in Fig. 4.

Electric-field-induced mechanical resonance in a carbon nanotube. (a) The nanotube is quasi-stationary (thermal vibration only). (b) The first and (c) second harmonic resonance is induced by high-frequency electric fields (MHz range).

Fig. 6 Electric-field-induced mechanical resonance in a carbon nanotube. (a) The nanotube is quasi-stationary (thermal vibration only). (b) The first and (c) second harmonic resonance is induced by high-frequency electric fields (MHz range).

An example, quite different from the first, of the determination of elastic properties of carbon nanotubes by in situ TEM is represented in Fig. 6.[5] In this case the long carbon nanotube shown is excited to mechanical resonance by a high-frequency electric field imposed between it and a nearby positionable counterelectrode. The first and second harmonics are induced by tuning the frequency. The shape of the nanotube at resonance corresponds closely to that expected for a cantilevered uniform beam, which allows its bending modulus to be determined from the dimensions of the nanotube and the driving frequency at resonance. For nanotubes produced by arc-discharge, it is found that this modulus varies from about 1.2 TPa (like the value for diamond) for nanotubes smaller than 8 nm to as little as 0.2 TPa for those with diameters larger than 30 nm. Static electric fields which induce bending of carbon nanotubes in situ may also be employed for elastic modulus determin ation.[6]

Finally, Fig. 7a describes a TEM/STEM specimen holder for nanoindentation of a thin foil or film, which, like that employed in the vibrating carbon nanotube example and the earlier quantum contact example, involves piezoelectric positioning of a part of the experimental apparatus (the indenter, the counterelectrode, or the gold tip). Fig. 7b-e shows TEM micrographs of dislocation activity during the indention process.[7,8]

Catalysis and Other Reaction Studies

Catalysis is responsible for the commercial production of countless organic substances, the catalytic particles responsible by necessity being in the nanosize regime. One of the aspects of this topic relevant to in situ microscopy studies, for example, involves ensuring that a catalyst remains in the nanosize regime under conditions simulated in the microscope for those to be encountered in production of a particular product. There have been two quite different approaches to allow such studies to be conducted in situ in TEM or STEM, one involving the design of environmental specimen holders which contain the entire experiment, environment and all, used in conjunction with an unmodified microscope, and the other, the design of a dedicated environmental microscope into which the desired chemical environment is introduced in the presence of a specimen holder which may provide independent capabilities such as heating. In the case of the environmental holder concept, windows with adequate electron transparency must be employed to isolate the experimental environment from the vacuum of the microscope, which may severely limit the pressures employed as well as instrumental spatial resolution. In an environmental microscope, as in older environmental cells that were inserted between the objective poles, usually of high-voltage electron microscopes (HVEMs), a pair of apertures, one pair before and the other after the specimen holder, serve this isolation function, there being differential pumping within each aperture pair. Alignment of the apertures is critical. The success of the dedicated in situ microscope concept, especially since about 1995, attests to its experimental versatility and economic advantage for catalysis research in both industry and university.

 (a) Schematic drawing of dedicated in situ nanoindentation holder for TEM/STEM. (b) Bright field transmission electron micrograph showing indenter tip and single grain of aluminum thin film on silicon. (c-e) Sequence of dark field images showing introduction and glide of dislocations into the aluminum grain during indentation (indenter no longer visible in dark field).

Fig. 7 (a) Schematic drawing of dedicated in situ nanoindentation holder for TEM/STEM. (b) Bright field transmission electron micrograph showing indenter tip and single grain of aluminum thin film on silicon. (c-e) Sequence of dark field images showing introduction and glide of dislocations into the aluminum grain during indentation (indenter no longer visible in dark field).

In situ real-time EHREM dynamic observations of Ru on titania xerogel catalyst in hydrogen environment recorded (a) at room temperature and subsequently during reaction at about 280°C for 2 hr.

Fig. 8 In situ real-time EHREM dynamic observations of Ru on titania xerogel catalyst in hydrogen environment recorded (a) at room temperature and subsequently during reaction at about 280°C for 2 hr.

The first of two short examples of its application is from a study of the stability of Ru particles in TiO2 during the hydrogenation of adiponitrile during the production of Nylon 6,6. Fig. 8 shows the catalyst (the 2-3-nm Ru particles) in hydrogen at 20°C and at 280°C after 2 hr, demonstrating no obvious reaction between the Ru particles and TiO2 support. Such reaction associated with reduction by hydrogen is known for this important commercial system above 500°C. This type of in situ finding has important implications for the utility and performance of such catalysts in the selective hydrogena-tion reaction.

Wet-ETEM of in situ polymerization. (a) Co-Ru on titania catalyst system particle at m (supported on sample grid G) with reactant solution from which (b) polymerization over the catalyst occurs at about 188°C.

Fig. 9 Wet-ETEM of in situ polymerization. (a) Co-Ru on titania catalyst system particle at m (supported on sample grid G) with reactant solution from which (b) polymerization over the catalyst occurs at about 188°C.

Schematic drawing of a closed system electrochemical cell for insertion in a TEM/STEM specimen holder.

Fig. 10 Schematic drawing of a closed system electrochemical cell for insertion in a TEM/STEM specimen holder.

The second example illustrates the in situ observation of catalyst-promoted polymerization and involves the same environmental TEM/STEM as the first example and a heating holder incorporating a windowless wet cell (the microscope provides the atmosphere and the wet cell, a microliter or so of fluid in contact with catalyst). In the case shown in Fig. 9 the catalyst/support particle (m), composed of a Ru-Co alloy on TiO2, is immersed in a solution of hexamethylene diamine and adipic acid, at ~ 188°C. The polymerization reaction to the polyamide is observed to occur at the interface with the particle seen in Fig. 9b. The observations demonstrate a stable catalyst and the formation of an essentially clean polymer structure. In this example a specialized specimen holder was employed. An example of a cell designed for studies of electrochemical reactions such as electroplating is shown in Fig. 10. The cell simply inserts in a holder for insertion in TEM or STEM (Frances Ross, private communication, 2003).

Next post:

Previous post: