High beam brightness aids fine scale processing
TWI Bulletin, September - October 2010
Allan Sanderson. After joining TWI (then BWRA) in 1966, Allan pioneered the development of high power EB equipment and subsequently Reduced Pressure EBW. He has served as Section Leader, Head of the Department and Technology Manager prior to his present position of Technology Fellow, EB Group. Currently he is engaged in the development of mobile sliding seal devices for large scale EBW fabrication projects. Over the decades, one of his special activities has been the design and development of a wide range of gun column and electron gun systems, many of which employ an innovative radio frequency cathode heating method.
Nick Bagshaw joined the Structural Integrity Technology Group, TWI, as Project Leader in August 2001; his current position is Principal Project Leader working in the Electron Beam Processes Group. He has experience in weld process development using experimental and numerical methods, contributing to pioneering work in the field of weld process modelling at TWI to predict weld residual stresses and distortion. Before joining TWI, Nick was awarded an MSc in Structural Integrity and a PhD in on leak rates through narrow cracks.
Power beams are widely used within manufacturing industry for welding, melting, texturing, drilling and thermally transforming substrates. Electron beam (EB) and laser beam (LB) equipment have developed progressively over the past five decades. As authors Allan Sanderson and Nick Bagshaw explain higher power levels are now available for processing at higher speed, or for welding thicker section material. Also equipment has been devised to produce high quality beams that can be projected as far as half a metre by fibre lasers, whilst retaining a focused spot diameter of only 0.3mm.
In preceding work, TWI carried out design of equipment to address a gap in electron beam processing that existed between EB low power welding and EB lithography. Electron beams were anticipated to offer a unique potential for power beam processing at scales below 10µm that laser beams could not easily address because of diffraction limitations. Equipment was initially developed with relatively high acceleration potential (150kV, see Figure 1) but, to process at reduced scale, lower acceleration potential (60kV) was used. This produced slower electrons that penetrated less into the solid material being processed. Consequently, with optimisation of the equipment design to maintain a fine beam focal spot, the beam power could be deposited into a smaller volume of material in the workpiece, and a minimum beam diameter of 50µm was achieved. Processing capability was explored for welding, drilling and texturing of metallic and silicon substrates.
Fig.1. General view of high brightness electron beam equipment
In other earlier work TWI developed a surface modification technique called Surfi-Sculpt®, commonly using adapted designs of electron beam welding equipment operated at powers of approximately 1kW. The process used an electron or laser beam to melt and manipulate material to form surface features typically of dimensions of a few millimetres.
This process makes use of the surface tension forces present in molten metal where there is a temperature gradient; molten metal streams from the hotter regions to cooler regions driven by the higher surface tension levels in the cooler regions. The process is also aided by vapour pressure forces created by the jetting action of metal atoms emitted by the liquid pool. The same driving forces play a key role in EB and LB welding where molten metal is seen to stream from the extremely hot weld pool into the wake of the beam, forming a protruding bead.
In the case of Surfi-Sculpt the repeated scanning from any one end of a line to the other progressively displaces metal producing protrusions and depressions. The molten metal is always displaced towards the start of the line scan. The process can be carried out very rapidly and can produce a wide variety of feature shapes. Many potential applications could be addressed, particularly in the medical sector, if the Surfi-Sculpt process could be used to produce much smaller features of size of 5-100µm, which may be possible to create if the beam spot size was reduced further.
This article outlines further developments of high brightness EB equipment at TWI, and describes the achievement of novel processing (ie texturing, Surfi-Sculpt and drilling) of stainless steel, at scales as fine as 20µm. The work has been carried out to demonstrate equipment capability and to allow assessment of the suitability of these processes for a range of potential applications including micro-electronics, micro-fluidics, medical device encapsulation and creation of bio-compatible surfaces.
Objectives
The main objectives of this work were to:
- Improve electron gun design and reduce cathode diameter for higher beam brightness at focus.
- Generate a high brightness beam, with a focused spot size of less than 10µm, operating at an accelerating potential of 60kV.
- Demonstrate smaller scale processing.
High brightness equipment development
Gun column design
High beam brightness is achieved when the beam intensity is high and the convergence angle is small. The brightness of an electron beam is established at the cathode emitter and subsequently by the quality of the gun electrodes that shape the beam. Mathematically, brightness is usually defined as the power density at focus divided by the solid angle (measured in steradians (sr)) of convergence eg kWm -2sr-1.
After the anode, as the beam passes through focusing and deflection optics, the beam brightness is generally reduced; it cannot be subsequently increased by electron optical devices. For a given well designed system the product of beam radius and the angle of convergence stays approximately constant. Beam intensity can be increased at the focus by increasing the angle but this can be counter-productive because it reduces the depth of focus and the amount of energy that can be concentrated into the volume encompassing the most intense portion of the beam.
Empirical beam calculations
The focused beam diameter for the high brightness machine is predominantly controlled by the thermal velocity spread of the emitted electrons and lens aberration. Lens aberration is a function of the lens design, in particular the lens bore, pole-piece gap and the radius of the electron beam at the lens centre-plane. The larger the radius of the electron beam the greater the beam aberration. On the other hand, the degree of spreading of the electron beam, due to the fundamental way in which the electrons emerge from the cathode surface, is a function of the beam convergence angle as well as the cathode radius, temperature and accelerating voltage.
Space charge is a further consideration, but in this context the beam current is so small, and the angle of convergence relatively high, that this effect is negligible, especially in the presence of positive ions created in the material processing environment. These will tend to neutralise any small mutual repulsion effects between electrons as they come into focus.
Thus the focused spot size is controlled by the balance between the need to have sufficient beam convergence to combat thermal velocity spreading effects and lens aberration caused by excessive beam diameter at the lens centre-plane.
Figure 2a shows how these factors interplay for a 0.1mm diameter cathode and a working distance of 25mm and indicates the optimum beam radius at the lens centre-plane for the smallest focused beam radius. At an accelerating voltage of 60kV and for the existing focusing lens, the minimum focused spot radius of 2.9µm occurs when the radius at the lens centre-plane is 1.8mm for a 200µA beam. There is little change in this figure over the beam current range of from 10 to 200µA, although the spot radius would be expected to decrease to 2.7µm at the lower current level due to the reduced cathode temperature required (because of reduced thermal velocity spread).
Fig.2. Effect of beam radius (rp) on the thermal velocity radius (rt), the lens aberration radius (rc) and the root mean square radius (rms) for a 60kV, 200µA beam, from 0.1mm diameter cathode, at working distances of: a) 25mm; b) 5mm
If the working distance is reduced to 5mm, (see Fig.2b) this produces a significant reduction in the focused spot radius, ie 1.9 cf 2.9µm and 1.7 cf 2.7µm, for 200 and 10µA beams, respectively.
Finite element analysis
General approach
The finite element (FE) method was used to investigate alternative EB gun electrode geometries. This involved three main aspects, viz:
- Optimisation of the electrode geometry to give the lowest beam distortion.
- Confirmation of the optimum beam profile giving the minimum focused spot diameter.
- Beam current calculation to set a sensible limit on cathode temperature.
A single model of the entire gun column was used. Magnetic component flux densities were first determined for the electromagnetic focusing lens element taking into account the change in permeability with magnetic field strength whilst using a coil current known from empirical calculations to produce the required focal length. The r (radial) and z (axial) components of magnetic flux density were extracted and stored. The gun electrodes were then optimised to produce sufficient beam current at an acceptable cathode temperature, with the correct divergence (as determined by the empirical calculations above). When a beam of the correct profile and current level was achieved, the stored flux density components were re-introduced into the problem to produce a combined electrostatic, space charge and magnetic focusing solution. The gun was modelled with an accelerating potential of 60kV as a temperature limited diode.
The beam quality or beam brightness is largely governed by the geometric detail in the near vicinity of the cathode. It has been shown that the highest brightness for a given cathode radius and current level is obtained from diode guns in which the electron trajectories are essentially laminar. This type of flow is generally only achieved when there are no discontinuities between the cathode emission face and the metal surrounding the cathode. In the case of a diode gun, since the cathode emitter is embedded in a closely fitting cathode carrier, which in turn is surrounded by a cathode shield electrode at the same potential, there are also no extraneous electrical fields to exacerbate the effect of any physical discontinuities. For triode guns, because the grid electrode is not only physically separated from the cathode emitter but also at a more negative potential, the combination produces significant beam distortion which varies with grid potential and beam current.
Cathode design
Previously, the smallest cathode successfully manufactured and used in the high brightness machine was 0.3mm diameter. In order to increase beam brightness, a 0.1mm diameter cathode was used in empirical calculations and subsequently in the finite element analysis (FEA). Attempts were then made to manufacture the proposed cathode. Some compromises were necessary to facilitate component machining and assembly. The 0.1mm cathode was installed in the high brightness machine and measurements made of the unfocused beam. This indicated that further adjustments of the gun electrodes were required in order to approach the optimum focused beam brightness.
Beam processing results
Texturing of stainless steel
In recent TWI Core Research Programme work, using a 0.3mm diameter lanthanum hexaboride cathode, it had proved possible to make track and cut features in stainless steel down to dimensions of approximately 60µm. Melt track features were made in the work reported here on stainless steel so that comparisons could be made. All of the work was conducted on as received bright 2mm thick 304L stainless steel 26 x 26mm coupons. A fixed accelerating voltage of 60kV was used throughout and the beam focus was optimised by achieving the brightest visible moving trace when impacting at low current on the specimen. Attempts to focus an undeflected beam on stainless steel, as might be used in EB welding practice, were found to be problematic, since the beam rapidly drilled into the surface even at beam current levels below 10µA.
A 16 line raster, previously used to create fine-scale fin structures on stainless steel, was used as a means of assessing the improvement in beam intensity. It was found that the beam could produce finer features than previously, and under certain conditions the intensity was sufficient to promote strong vapourisation with minimal melting of the material. Consequently, it proved possible not only to make extremely fine features but to machine nominally square holes with dimensions of as little as 60 x 60µm.
At a working distance of 25mm, measured from the extremity of the gun column, the 16 line raster typically produced a melted track pattern as shown in Figure 3. The individual tracks were some 30µm in width for a scan area of 600 x 600µm and were virtually touching at the start of the line scan, Figure 3b. Reducing working distance to 5mm increased beam intensity as predicted by the plots shown in Figure 2. This resulted in even finer tracks 20µm wide. Remarkably the beam intensity was so high that only 1.1W of beam power was required to produce these sharp edged features, see Figure 4.
Fig.3. SEM images of melted tracks made at 25mm working distance made with a 16 line, 1000, points/line 10Hz raster, for 30s at 25µA. and scan amplitudes of 0.5A x 0.5A:a) General view;
b) Detailed image of track extremities and protrusions
Fig.4. SEM images of melted tracks made at 5mm working distance with a 16 line, 1000 points/line 10Hz raster, for 30s at 19µA and scan amplitudes of 1.0 x 1.0A.
In contrast, when the same deflection pattern raster was applied at a beam current level of 100µA for 30 seconds the beam vapourised a relatively clean edged square hole, Figure 5. However, there was evidence of re-condensing of the vaporised material around the hole periphery and retained melted material.
Fig.5. SEM image of square hole made at 5mm working distance with a 16 line, 1000 points/line 10Hz raster and scan amplitudes of 0.5 x 0.5A. with a beam current of 100µA applied for 30s
End of part I. Part II of Allan Sanderson's and Nick Bagshaw's work on fine scale material processing using a high brightness electron beam continues in the next edition of Bulletin. It will discuss the cathode diameter considerations, the processing of the electron beam, the potential applications, and also compare the relative merits of the electron beam and the laser beam.
