The electron energy spectra of transmitted scattered electrons from free-standing films are simulated using a Monte Carlo computational approach. Elastic scattering is simulated using Mott cross-sections and inelastic scattering via discrete processes determined from dielectric function data. This allows one to simulate the secondary electrons as well as the loss peaks near the elastic (zero-loss) peak. The current study suggested a directed approach for determining the electron inelastic mean free path (IMFP) of materials at low primary electron energies. The IMFP of the reference material is not necessary for the suggested technique. The suggested technique uses the ratio between the transmitted elastic peak intensity and the background intensity of backscattered electrons. Free-standing films of Si, Cu, and Au were studied with thicknesses varying from 2 to 12 nm. Primary electron energies of 1, 3, and 5 keV were applied. The results appeared very good, with the percentage error range being between 5% and 25%. We also investigated the proportion of the first and second plasmon peak intensities to the elastic peak intensity. We believe that the latter could provide a directed method of measuring the IMFP of materials.

We simulate the electronic system of ejected electrons arising when a tip, positioned few 10 nm away from a surface, is operated in the field emission regime. We find that, by repeated quantum reflections (“quantum skipping”), electrons produced at the nanoscale primary site are able to reach the macroscopic environment surrounding the tip-surface region. We observe the hallmark of quantum skipping in an energy filtered experiment that detects the spin of the ejected electrons.

Scanning Field Emission Microscopy with Polarization Analysis was recently introduced to detect the spin polarization of electrons excited in the field emission regime of Scanning Tunnelling Microscopy. In this work, a miniature electron energy analyzer, called Bessel Box, is implemented into the Scanning Field Emission Microscope with Polarization Analysis setup. It is used to filter electrons according to their energy before they reach the spin detector. The Bessel Box allows, e.g., the spin polarization of elastically scattered electrons to be compared with the spin polarization obtained with the full energy spectrum. We use this technology to measure the local in-plane polarization signal as a function of the magnetic field B at room temperature for 10 monolayers Fe deposited on top of a W(011)-single crystal surface through a half mask (half of the surface is covered with Fe, the other half is uncovered). The spin polarization at the Fe-W crossing drops sharply from 9 % above Fe to 0 % above W(011) only if the elastically scattered electrons are selected for spin analysis. The mechanism of signal generation in Scanning Field Emission Microscope with Polarization Analysis including the formation of cascade of inelastically scattered electrons is discussed as an explanation for the different spin polarization profiles observed with and without Bessel Box (energy filtered).

In this study, we use Scanning Field Emission Microscopy (SFEM) combined with a miniature electron energy analyzer known as a Bessel box to measure electron energy spectra emitted from a sample. Previous studies using SFEM have revealed that the work function (ϕ) of the material under study has a significant role to play in the formation of the signal intensity. Hence, in order to understand the role of ϕ in greater detail, a sample of W(110) (ϕ = 5.25 eV) and a sample of Cs deposited on W(110) (ϕ ≈ 1.7 eV) were investigated. STM images show that the Cs covered surface has a speckled appearance indicating small Cs islands. The electron energy loss spectra obtained (which are the first using the Bessel box in SFEM) show differing structure in the elastic peak region. Monte Carlo (MC) simulations including quantum mechanical "bouncing" have been carried out. The results are consistent with MC simulations of the electrons escaping from the tip-sample junction.

The transmission of electrons with energies 15 keV and 30 keV through Si and Au films of 100 nm thickness each have been studied in a Scanning Transmission Electron Microscope. The electrons that were transmitted through the films were detected using a multi-annular photo-detector consisting of a central Bright Field (BF) and several Dark Field (DF) detectors. For the experiment the detector was gradually offset from the axis and the signal from the central BF detector was studied as a function of the offset distance and compared with MC simulations. The experiment showed better agreement between experiment and several different MC simulations as compared to previous results, but differences were still found particularly for low angle scattering from Si. Data from Au suggest that high energy secondary electrons contribute to the signal on the central BF detector for low primary beam energies, when the STEM detector is in its usual central position.

A miniature electron energy analyzer (Bessel box – BB) is described. The trajectory of electrons in the BB is simulated and results obtained using the BB are compared with these simulations. Its use in the Near Field Emission Scanning Electron Microscope (NFESEM) is discussed.

With the aim of improving detection and analysis of energy filtered electrons in the Scanning Field-Emission Microscope (SFEM) and of the spin polarised electrons in the SFEM with Polarisation Analysis (SFEMPA) tests are performed on a miniature electron detection unit employing a Bessel Box energy analyser. Even in conventional electron microscopes, the detection of low-energy electrons (with kinetic energies of the order of 100eV or lower) is inherently difficult due to the presence of electrostatic (and magnetic) fields in proximity of the beam-target interaction region, inhibiting the escape of these electrons and complicating the interpretation of their detected signal. The reduced dimensions of such a compact energy analyser - with a length of 1&1/2 channeltrons - consent its employment close to the sample surface, thus minimising the aforementioned fields effects. Experimental results demonstrating the capability of this analyser to collect electron spectra are discussed.

A new form of charged particle energy analyser is proposed. It is broadly based on the 180° magnetic spectrograph, but is intended to detect charged particles moving out of the dispersion plane with a helical motion. The analyser has the capability to acquire charged particle energy spectra over a large energy range, similar to those acquired in Auger electron spectroscopy, ca. 2500 eV and large angular range, up to 90°, in parallel. These conditions are more favourable for surface analysis by electron spectroscopy at high vacuum, where for example an electron energy resolution of 0.2% to 0.5% is typical. Expressions showing how the landing positions of the charged particles on the detector vary as a function of energy and polar take off angle are determined as well as the conditions for optimum energy resolution at a range of polar take off angles. The equations reveal that in general, the device obtains the highest resolution at angles of revolution greater than 180°. The design is simple and could be easily put into practice using available material and technologies and be used to analyse the energies of electrons emitted from a sample placed in a scanning electron microscope. It can be made to function with a primary electron beam of any desired energy and could fit in to the small space between the sample and the end of an electron column. However, the device is difficult to retrofit into existing SEMs and ideally an SEM column needs to be designed to work in association with the analyser. The direction of the magnetic field of the analyser is coincident with the axis of the electron gun so that the primary beam is little influenced by the magnetic field and symmetry can be maintained in the primary beam electron column. Because the device is intended to acquire electron spectra in parallel, any movement of the primary beam on the sample because of a ramping field in the analyser is avoided. The field of view and the effect of the analyser upon the operation of the SEM are discussed. Spectra including elastic and Auger peaks reveal an energy resolution of ~4 eV at 900‐eV electron energy. Copyright © 2016 John Wiley & Sons, Ltd.

It is generally regarded as impossible to carry out Auger Electron Spectroscopy (AES) analysis in a scanning electron microscope (SEM) due to the high ambient gas pressure in an SEM. This is because standard electron energy analysers such as the Concentric Hemispherical Analyser (CHA) or Cylindrical Mirror Analyser (CMA) are devices that acquire data in a serial manner that can last up to few minutes. This is considered too slow for high vacuum (10−6mbar) and a previously cleaned surface would be re-contaminated before a spectrum could be completed. This has led to AES being traditionally carried out under ultrahigh vacuum (UHV) environment. We report on two devices for fast acquisition of AES data characterising nanoscale objects by the use of AES in an SEM.

The sensitivity of Monte Carlo estimates of backscattering coefficients η to the accuracy of their input data is examined by studying the percentage change in η due to changes of 10% and 20% in the differential elastic scattering cross-section dσ/dΩ and corresponding changes in the stopping power S(E) in the primary energy range 200-10,000 eV. To a good approximation equivalent elastic and inelastic scattering changes produce equal and opposite shifts in η, a result consistent with predictions of transport theory. For medium to high atomic numbers an x% error in the specification of either S(E) or dσ/dΩ produces a percentage change in η significantly less than x%, while at low atomic number Δη/η increases approximately linearly with ln E so that Monte Carlo predictions are then more sensitive to parameter precision at high energy.

Direct detection of low-energy electrons (500-2000 eV) with a novel back-thinned CMOS active pixel sensor (APS) is reported in this paper. The sensor was installed in a JEOL 6400F scanning electron microscope to quantitatively test its linearity and spatial resolution by a focused electron beam. The obtained results show good linearity at electron beam energy values from 500 to 2000 eV. The full-width at half-maximum spatial resolution was around 2 pixels, a limit set by charge diffusion in the epilayer. These results show that this CMOS APS sensor is appropriate for low-energy electron detection providing the benefits of direct detection for many applications.

Although the Scanning Electron Microscope (SEM) has been in existence for many decades, it cannot be yet regarded as a true quantitative instrument—certainly when applied at the nanoscale. This is due to the presence of carbonaceous deposits at the surface and a poor understanding of the emission of secondary electrons from materials. In this paper, a short review is given of some of the progress made in the efforts to improve quantification in the SEM at York. We present results which strongly suggest that the currently accepted theory, which explains why there is a correlation between the secondary electron yield and the work function of a metal, is incorrect. In addition, we show that the backscattering coefficient from materials can be strongly influenced by surface layers at low primary electron energy. Finally, we present Auger electron spectra, which have been acquired at high speed at high vacuum (10−7mbar) and thus represent a new way to determine the composition of nanostructures in the SEM.

A new approach to the acquisition of Auger electron spectra is introduced. Electrons emitted from a sample illuminated by a primary electron beam are dispersed by a magnetic field which immerses both sample and electron energy analyser. The analyser is broadly based on the 180° magnetic spectrometer, but can acquire spectra with good energy resolution for electrons with a significant component of velocity parallel to the magnetic field. An Active Pixel Sensor is used to acquire the electron spectrum without the use of a microchannel plate as in most currently used analysers. An example spectrum of an elastic peak is given.

The determination of the dopant concentration in semiconductors at the nanometre scale is an important technological goal. One possible approach is to use the dopant contrast (DC) effect in low voltage scanning electron microscopy (LVSEM). The DC effect normally causes p-doped semiconducting material to appear brighter than n-doped in a LVSEM and the contrast is dependent on the dopant concentration. In this report we have studied highly doped p- and n-type silicon with thick oxide layers and after the oxide layer was removed by dipping in diluted HF. DC was measured as a function of primary beam voltage. It was found that the DC could be reversed (i.e. n-type brighter than p-type) at higher primary beam voltages for the thicker oxide, but such reversals were not apparent for the thinner oxide. By comparing with results from different electron emission techniques such as photoemission and field emission it can be concluded that oxygen plays an important role in DC not just for monolayer thicknesses as previously recognised, but also for much thicker oxides.

The secondary electron and backscattered electron coefficients have been measured as a function of primary beam energy for as-inserted and cleaned pure element samples. Clearly, the effect of cleaning samples makes a significant effect on both these key measurements needed for understanding the electron transport measurements in scannng electron microscopy and a number of other technologies. The results from the cleaned samples suggest that the currently accepted theory for secondary electron emission (SEE) of Baroody does not take account of an important physical effect. We propose that the SEE in transition metals is mainly controlled by the inelastic mean free path (IMFP) of the secondary electrons. In combination with current theories on the transport of hot electrons in transition metals, where sensitivity to the density of empty d states is important, the apparent correlation of the work function with SEE can be explained. The effect of errors in the electron elastic scattering cross-section and the electron stopping power on the estimates of backscattered electron coefficient, η, are explored for the case of Cu. It is found that percentage errors in one parameter (e.g. stopping power) cause very similar changes in η as equal but opposite percentage errors in the other parameter (e.g. elastic scattering cross-section).

Patterned Si surfaces, p- and n-type doped, were examined for different secondary electron yield (contrast between ptype and n-type regions) under the electron beam of a scanning electron microscope. The contrast as a function of primary beam energy was studied for samples with a thick oxide layer and with the layer removed using an HF solution. It was found that the contrast between p- and n- type areas reversed on the samples with a thick oxide layer as the primary beam energy was increased. However, after the oxide layer was removed, the contrast reversal was no longer apparent. In addition, it was also found that regions on a patterned Si sample could reverse in contrast when the scan speed of the electron beam was changed. The various competing theories describing the dopant contrast effect of doped semiconductors are discussed and compared to the results reported here and elsewhere in the literature. It is concluded that oxygen at sub-monolayer coverage through to thick films plays an important role in the dopant contrast effect. However, adventitious carbon is equally important where a metal-oxide-semiconductor structure could exist with the presence of these two materials. Results from the literature using other techniques such as photoemission and field emission are also considered and it is found that these studies give results which are inconsistent with several of the current theories which attempt to explain the dopant contrast effect.

The scanning electron microscope (SEM) has been an important instrument that underpinned many of the advances made in science and engineering over the last 50 years or so. This instrument continues to play the same crucial role today in many disciplines thanks to several advantageous properties of this instrument, in comparison to others, which include: ease of sample preparation, simple instrument operation, moderately high spatial resolution approaching 1 nm, offered by a number of manufacturers of modern SEMs. The latter property has been mainly due to recent advances made in the electron optics of SEM cathodes and detectors. These properties have placed the SEM at the heart of semiconductor fabrications; as an aide in research and development, and as a metrology tool on fabrication lines. The continuous challenge to conform to the Moore's law which means ever shrinking dimensions on electronic devices, places additional demands on the instrument capabilities and its use in this environment. In addition, the assessment of nano- structures and nano-materials is another field that requires the use of the SEM. Metrology, in the wider sense of scale and composition, remains to be the challenge to SEMs, however, fortunately, the basic physics of electron-solid interaction that takes place in the SEM is quite advanced and well understood. It is time to revisit the electron-solid interaction in the SEM and to see if new detectors and procedures should now be adopted to maximise the benefits gained in inspecting a solid sample.

The secondary electron (SE) yield, delta, was measured from 24 different elements at low primary beam energy (250-5,000 eV). Surface contamination affects the intensity of delta but not its variation with primary electron energy. The experiments suggest that the mean free path of SEs varies across the d bands of transition metals in agreement with theory. Monte Carlo simulations suggest that surface plasmons may need to be included for improved agreement with experiment.