Most textbooks on catalysis describe experimental difficulties in terms of a ‘materials gap’ and a ‘pressure gap’. These concepts illustrate, respectively, the differences in structure and operating conditions between model systems and real-world catalysts (see Fig. 1). Conventional surface science studies, for example, may use a single crystal substrate under ultra high vacuum conditions, whilst heterogeneous industrial catalysts often comprise nano-particulate metals and high surface-area oxide supports that are held above atmospheric pressure. Fortunately, bridging the materials gap is now relatively routine: high resolution microscopy techniques can be used to identify structural features in inhomogeneous samples (see, for example, Ref. ). Of these, transmission electron microscopy (TEM) is particularly versatile, not least because of its capability for mapping chemical information right down to the atomic scale. Such TEM studies, however, still struggle to bridge the pressure gap, since they are typically conducted under high vacuum conditions. In a Letter in this issue, Boyes et al.  showcase recent developments in environmental TEM techniques that can be conducted at pressures of several Pascals whilst retaining atomic resolution and full TEM functionality. The experimental advances promise new insights into a host of catalytic and other systems under conditions that approach ambient pressures.
Environmental TEM has been under development by the York group for some time  and the main advance described now is to extend their methodology to scanning TEM (or STEM) studies: in particular, to aberration-corrected STEM, giving ångström-scale spatial resolution at elevated pressures. In contrast to their previous TEM work, which illuminated a thin specimen with a relatively broad electron beam, in STEM a focused electron probe is rastered across the sample to create an image pixel-by-pixel.
The methodology has a number of advantages. Firstly, a STEM image compiled using electrons scattered through high angles is directly-interpretable and uncomplicated by the diffraction effects that tend to dominate TEM images of crystalline materials. The pixel intensity depends linearly on the number of scattering atoms and as the square of the sample's atomic number. Thus, recent work has displayed exquisite three-dimensional and atom-resolved studies of nanoparticle surfaces  – the kind of work that would be essential to the identification of the active sites of a supported metal catalyst. Secondly, the collection of additional signals during rastering, such as x-rays or inelastically-scattered electrons, generates pixel-by-pixel spectra that provide comprehensive functional maps . Usually, such work is conducted with the sample held under high vacuum conditions, but the York group have incorporated substantial differential pumping directly within the microscope's objective lens. Their samples, which can also be heated in excess of 800 K, can now be held in pressures of a few Pascals whilst the rest of the electron optics are retained under the vacuum conditions that are essential to their operation. Curiously, this development is facilitated by the incorporation of the aberration corrector. It is usually the case that STEM resolution improves as the space around the sample diminishes (ie. as the pole-pieces of the objective lens’ magnetic circuit get closer together). However, the corrector here allows a better compromise to be made so as to ensure adequate pumping speeds about the sample whilst maintaining sub-atomic resolution. Importantly, the customised instrument still allows a wide variety of STEM imaging modalities to be used, including spectroscopic studies for future chemical mapping of catalysts. In addition, there is the potential to perform mass spectrometry of product gas species, which would enable studies of catalytic activity and performance.
Boyes et al. launch their new ‘environmental STEM’ technique with a study of platinum clusters held at 670 K and under 2 Pa of hydrogen gas. Whilst this is still below the pressure of industrial catalytic reactions, it is substantially above the pressure required to completely saturate surface and bulk adsorption sites and so can be said to cross the problematic pressure gap of previous work. Spatial resolution is demonstrated by direct, dynamic visualisation of individual Pt atoms as they migrate between nanoparticles during sintering – the first such STEM visualisation of atom migration under such conditions and an observation that would be beyond typical TEM experiments. They go on to show – by counting atoms directly – that the clusters restructure at elevated temperatures and pressures. This is precisely the kind of information that is required if active sites on realistic catalysts are to be identified.
There is much to look forward to from the new environmental STEM technique. Atomic-resolved studies of catalytic activity are clearly a major area of future research, but a variety of other exciting experiments will be feasible within the temperature and pressure range offered here. These include studies of gas-induced phase changes, sintering, embrittlement and characterisation of more delicate samples that require to be hydrated during imaging. Looking forward, one might also wonder what further advances in technique are feasible. Here, an intriguing possibility would be the incorporation of next-generation monochromated electron sources  that may enable studies of vibrational states of molecular adsorbates during catalysis: watch this space.