Beam Deceleration is a relatively simple method to reduce electron beam energy and improve imaging parameters such as resolution and contrast. The scanning electron microscope (SEM) uses a sharply focused electron beam to probe the specimen surface. The energy of the electrons forming such a probe is determined by the electrical potential of the electron source, referred to as accelerating voltage or high voltage (HV). No matter how many times the electrons are accelerated or decelerated inside the column, they leave the column with an energy corresponding to the high voltage. The high voltage is usually controllable within a range of 200 V to 30 kV for most commercially available SEMs, allowing the operator to select the electron beam energy suitable for the application. Imaging with very low electron beam energy has great importance, which is illustrated by SEM instrumentation development over the last few decades [1–2]. Low voltage microscopy is a topic discussed at most microscopy-related conferences these days, but generally, it is approached with an immersion lens and field emission gun (FEG) SEM system because of the better beam current densities. However, beam deceleration is also a means to bring low kV improvement to SEMs with thermionic electron sources.

Extended abstract of a paper presented at Microscopy and Microanalysis 2010 in Portland, Oregon, USA, August 1 – August 5, 2010.

Techniques to vary the landing energy have been used to achieve very low surface energy imaging over decades [1,2], however this technique can be applied to yield other image quality improvements at a range of landing energies such as lowering the signal detection limit, improving resolution and general contrast enhancement. Applying a bias voltage to a sample to decelerate the beam can improve resolution which is a key component of image quality. This can be achieved by slowing the electrons at the point of interaction to a speed equal to the difference in potential between the primary beam energy and the sample bias voltage. In this way a higher voltage electron beam can be used to yield a smaller interaction area. Typically this well known principle (also called cathode lens mode) has been applied to decrease interaction volumes for surface characterization at low energies [1, 2] however this same principle can be used to increase image resolution and quality. Image quality is affected in a number of ways by beam deceleration including a lowering of signal detection limits and improvement in contrast. Particularly, imaging at lower accelerating voltages, beam deceleration can improve the ability to detect signal from conductive and semi-conductive samples. This is illustrated in figure 1 by a 500 eV image of a palladium solder/copper alloy (landing energy 500 eV/ primary beam energy 3.5 kV with stage bias of 3kV). Clearly surface details are more distinct, surface feature contrast is more visible, and the image quality is much improved. Without the beam deceleration the detection efficiency of backscattered electron (BSE) detector is very poor at 500 eV therefore the detection limit is lowered. Improved image quality can be a direct result of the angular selection of signal that is possible with beam deceleration across a wide variety of accelerating voltages. Changing the bias voltage on the sample allows angular selection of signal and thus additional contrast mechanisms to be observed. This is illustrated in figure 2 by comparing images of microcrystalline diamond deposited on silicon substrate without using the beam deceleration to one at the same landing energy (3 keV) with the stage bias of 3 kV. Surface details are more distinct with the beam deceleration and the image quality is much improved while the contrast from atomic differences is lessened as the high angle BSE are moved in line with the primary beam axis.

Angular segmentation of detected signal in a SEM can be used to segment topographic and compositional contrast as has been shown with concentric ring solid state detection [1]. Subdividing the topographic contrast further can yield additional information on the direction of sample surfaces and allow clear interpretation of topography. Signal intensity is influenced by both sample topography and composition and the ability to separate one from the other may be useful in estimating scale of topography.Solid state detector diodes have been introduced with the ability to address concentric segment rings around the beam exit to separate high angle and low angle electrons (as shown in Figure 1). This works quite well to distinguish between composition and density on the inner rings closest to the beam and topographic information which is more prominent on the outer rings [1]. What is perhaps missing is the relative direction of the surface as the shading is uniform from one side of the image to the other. Conventional segmentation of the detector such that there are 4 segments does not allow removal of atomic number effects on the topographic information as the inner (composition) and outer (topographic) information are both within a given signal collection area. Since BSE distribution is dependent on both composition and topography, the effect of a high atomic number region with topography may be exaggerated in respect to an adjacent low atomic number region due to the combined signal brightness and contrast. This direct result is used for calculating height of topography based on contrast and brightness differences. The inability to separate the brightness due to composition could skew results if brightness and contrast are used as the determining factor for height, even when this is -axis th segment signal and this could be a contributing factor for why this data is not so accurate [2].Subdividing the outer rings of concentric ring detector segments into three 60 degree arch segments, eliminates the composition information and allows characterization of directional information from topographic content. An example of this is shown in Figure 2, where images from the inner ring and the three segments are shown. Shadowing is based on the direction of the sample surface.Directional signal segmentation is an emerging way to get new sample surface information. Angular filtering has the potential to create more accurate surface height representations in SEM imaging where composition can be removed from the equation or conversely used to refine the signal interpretation to yield better estimations because of the compositional signal differences. New ability to separate signal provides valuable insight into topographic information and presents new opportunity to use different, direction-rich data for surface mapping and height measurements.

Extended abstract of a paper presented at Microscopy and Microanalysis 2012 in Phoenix, Arizona, USA, July 29 – August 2, 2012.