Search

International audience; A bewildering number of techniques have been developed for transmission electron microscopy (TEM), involving the use of ever more complex combinations of lens configurations, apertures and detector geometries. In parallel, the developments in the field of ion beam instruments have modernized sample preparation and enabled the preparation of various types of materials. However, the desired final specimen geometry is always almost the same: a thin foil of uniform thickness. Here we will show that judicious design of specimen geometry can make all the difference and that experiments can be carried out on the most basic electron microscope and in the usual imaging modes. We propose two sample preparation methods that allow the formation of controlled moiré patterns for general monocrystalline structures in cross-section and at specific sites. We developed moiré image treatment algorithms using an absolute correction of projection lens distortions of a TEM that allows strain measurements and mapping with a nanometer resolution and 10 −4 precision. Imaging and diffraction techniques in other fields may in turn benefit from this technique in perspective. Developments in transmission electron microscopy (TEM) have centered on improving instrumentation and optical configurations. Considerable effort has also been put into developing sophisticated specimen preparation methods and equipment, from the humble ion-beam polisher, through tripod polishing and ultramicrotomy, to multifunction focused-ion beam (FIB) systems. However, the desired final specimen geometry is always the same: a thin foil of uniform thickness. Here we will show that judicious design of specimen geometry can make the all the difference and that experiments can be carried out on the most basic electron microscope and in the usual imaging modes. The example we have chosen to illustrate electron microscopy by specimen design is the measurement of strain in thin-films and devices. Strain engineering is now an important feature of many research areas, from strained-silicon transistors 1 to ferroelectrics 2 and thermoelectric materials 3. It is therefore not surprising that many TEM techniques have been developed over the years to measure it accurately 4. These include quantitative analysis of high-resolution transmission electron microscope images (HRTEM) 5–7 , convergent-beam electron dif-fraction (CBED) 8 , nano-beam electron diffraction (NBED) 9,10 , dark-field off-axis electron holography (DFEH) 11 , dark-field inline holography (DIH) 12,13 and more recently STEM moiré fringes (SMF) 14,15. All these techniques have specific instrumental requirements such as aberration correctors, Lorentz lenses or electrostatic biprisms (for holography). Even DIH and CBED require an imaging energy filter faced with the complexity of the data simulation and interpretation 12,13,16,17. Our aim is not to belittle instrumental developments but to explore an alternative route to making measurements with electron microscopy that can, in some ways, be superior to existing techniques. How then can we design the specimen geometry to perform the technique on the most basic conventional TEM: by returning to the moiré imaging phenomenon known from the very beginnings of electron microscopy 18,19. Moiré patterns are a general interference phenomena that appear when two periodic arrays are superposed 20. In TEM, they usually occur by chance, when two crystals with slightly different lattice parameters or orientation are superposed along the path of the electron beam (Fig. 1). In a bright-field (BF) image, the fringe pattern results from the interference of the transmitted beam passing through the two crystals with a doubly diffracted beam in crystal II. In a dark-field (DF) image, the resulting Moiré pattern arises from an interference between two beams diffracted in crystal I and II. Moiré fringes are formed if there is a vector difference = − m g g 2 1 between the diffraction vectors I, g 1 and II, g 2. The moiré spacing, m, can therefore be linked to the strain, or orientation