Fine scale processing using high beam brightness - the sequel

TWI Bulletin, November - December 2010

The cathode diameter, the beam processing and potential applications

Part one of Allan Sanderson's and Nick Bagshaw's work on fine scale processing laid out the objectives of this complex task. It dealt with the equipment development, the finite element analysis, the cathode design and the beam processing results. Now part two examines the cathode diameter considerations, the processing using the electron beam and the potential applications. It also compares the relative merits of the electron beam and laser beam approaches to fine scale surface modification.

Cathode diameter considerations

The main object of the equipment modification was to achieve higher electron beam brightness by a reduction in cathode diameter. A general view of the high brightness electron beam equipment is shown in Fig.1. Previously the cathode diameter used was 0.3mm and this allowed melted track widths of down to approximately 60µm to be made. Electron back-scattered imaging indicated that the beam diameter was approximately 10µm. Empirical calculations indicated that the beam diameter for working distances of 25 and 5mm, (Fig.2) could be reduced to 5.4 and 3.4µm (for a 200µA beam) respectively, by reducing the cathode diameter to 0.1mm. The finite element calculations, however, even without including the effect of electron velocity spread, predicted a diameter of some 15µm, but this is believed to be largely a result of FEA meshing problems particularly around the

Fig.1. General view of high brightness electron beam equipment.

 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:
Fig.2a) 25mm; Fig.2b) 5mm.

Electron beam processing

In practice, the introduction of the 0.1mm cathode appears to have had a beneficial effect on the focused beam diameter even though the unfocused beam diameter in the focusing lens had not yet been fully optimised, see Fig.3 for trials at 25mm working distance. Further melt run tracks at 5mm distance on stainless steel were reduced to 20µm as can be seen in Fig.4. This is ~1/3 of the track width that was achieved in using the previous high brightness equipment with a 0.3mm diameter cathode.

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 0.5 x 0.5A:

Fig.3a) General view;

Fig.3b) 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 the case of electron beam welding, the beam forms a vapour filled cavity, displacing an appreciable volume of liquid metal that subsequently forms the weld beads. In the Surfi-Sculpt process, surface tension forces drive the liquid metal aided by the jetting action of metal vapour. In the present high brightness electron beam development the object is to provide a tool that can be used to make textures and patterns at a scale of 1 to 100µm. Moreover, if the beam brightness can be increased to a sufficiently high level then it may well be possible to vaporise metal without producing significant melting, so called ablation. Certainly some of the effects noted in this work have indicated that when the beam is highly focused metal is removed largely by evaporation, see Fig.5.

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.

Potential applications

Potential applications for micro-processing exist in many sectors. In the medical field there are a number of interesting areas including surface texturing to promote bone cell growth on implants. In this case there is a need to provide a surface condition in which the features match the scale of the cellular structure. Typically, features need to be approximately in the range 1 to 10µm in width to achieve intimate cellular bonding with the metallic implant. Medical in-body instrumentation is another area of interest where extremely small instrumentation packages need to be welded with high integrity to prevent ingress of body fluids and gases.

In the electronic industry, lab-on-chip technology is another field where the machining of microscopic grooves in metals and ceramics could benefit from the recent advances in the high brightness beam processing technology.

There may also be non-medical applications where hermetic sealing of micro-packages would be better achieved using a micro-electron beam weld.

The lubrication of moving parts relies on a constant supply of fresh oil at a micro level to reduce wear and extend component life. This is another area where micro-texturing and the micro-Surfi-Sculpt process could offer special benefits.

In the production of micro-parts in difficult-to-machine materials, such as hard metals and ceramics, electron beam micro-machining by evaporation, without significant melting, is worthy of further study. Certainly, in the manufacture of miniature electron gun cathodes from lanthanum hexaboride, the high brightness beam itself may well provide the precision machining capability required to further increase beam brightness.

Electron and laser beams comparisons

The capabilities of the high brightness electron beam should be set in the context of what laser beams can currently offer. For lasers, the minimum feature size that can be achieved is ultimately limited by diffraction effects which are dependent upon the wavelength of the laser light.

The focused spot radius limited by diffraction is given by:

r_d = 0.61 αå

Where α is the beam semi-angle in radians.

The equation applies to both electron and laser beams. The wavelength of a 60kV electron beam is 4.865 x 10^-12 (4.865 x 10^-6µm). Compared with a 1.06µm Nd:YAG laser, the 60kV electron wavelength is over 200,000 times shorter. Taking the optimum beam radius at the lens centreplane for a 60kV, 200µm beam, as shown in Fig.2a. to be 1.8mm for a working distance of 25mm provides a semi-convergence angle of arc tan(18(25+20)) = 2.29° or 4.0 x 10^-2 radians. From the equation above the diffraction limited spot radius for a 60kV electron beam would be 7.4 x 10^-5µm. The same calculation for a 1.06µm laser beam would yield a diffraction limited focus radius of 16µm.

Of course, for laser and electron beam, the spot radius could be reduced by increasing the beam semi-angle by using a shorter focal length lens but this would limit the depth of focus. On the other hand, the laser beam spot size can be affected by the quality of the equipment, optics and beam delivery system.

In the case of electron beams, diffraction is not the limiting factor, since for the comparative case selected, Fig.2a suggests a limiting beam radius of 2.9µm. Increasing beam semi-angle of divergence from the source for the given focusing lens would increase this radius because of lens aberration and decreasing the divergence angle would also increase the focused spot radius because of the thermal velocity factor.

Again, in the laser case the optimum spot radius could be reduced by employing say an excimer laser with a wavelength of 0.222 to 3.35µm or a copper vapour laser with a wave length of 0.578 to 8.7µm. These comparative data are summarised in the table.

Table: Comparison of electron and laser beam limiting radii for a semi-convergence angle of 2.29 degrees

In practice the relative processing capabilities of lasers and high brightness electron beams also need to take into account many other factors such as material properties, including reflectivity and absorption characteristics. More importantly laser beams are more readily pulsed to give high peak power for short duration periods. Although average power levels are typically less than 100W for a copper vapour laser, peak powers can be as high as 500kW for 25ns with high repetition frequencies (2-50kHz). Combined with a small spot size, peak power density can be as high as 10GW/mm2

This enables extremely small cuts, holes and machined features to be achieved in a wide variety of materials down to 1µm wide, with depths of up to 2mm and accuracies in some cases better than ±0.25µm. On the other hand, electron beams can be more readily deflected at very high speed to produce complex patterns and may well be better suited to producing small melted tracks, Surfi-Sculpt features and micro-welds. It is in these areas, in which the process depends on heat build up in the material to achieve highly controlled melting, that high brightness EB processing is expected to find initial applications. As presently configured, the high brightness electron beam equipment does not incorporate a pulsing capability, but it is anticipated that one will be added in the near future.

The use of a triode electron gun would allow more precise application of beam power. Discrete high energy packages could be delivered in pulses to control pattern start and stop times. It would also be essential for machining where short high energy pulses are required to create material ablation without significant melting. This would promote cleaner drill holes, cut edges and profiles.

Further improvements are also expected to be brought about by optimisation of the source divergence angle and cathode design. Also, increasing beam accelerating voltage could be beneficial for certain material types and processing speeds where electron penetration in the solid material is not the limiting factor.

Conclusions

By using high frequency beam deflection techniques it is possible to manipulate molten metal in a wide variety of ways. In order to access the effectiveness of the beam produced by the 0.1mm diameter cathode and new gun structure, a specific deflection pattern was employed. This consisted of 16 lines of 1000 points which were repeated at a frequency of 10Hz. Previously this pattern, in conjunction with a beam produced from a 0.3mm cathode created an array of fin protrusions on stainless steel, where the minimum melt track was approximately 60µm. In this work it proved possible, using the same pattern to reduce the feature size on stainless steel down to 20µm. In addition it is encouraging to note that molten metal can still be manipulated by surface tension and vapour pressure forces at this fine-scale. It is believed this represents a significant level of improvement, moving the micro-processing capability much closer to that which is required for application in the medical and electronic fields.

• The focused spot size achievable using an electron beam for beam processing is now approaching that of the finest laser beams.
  
  
• Features of less than 20µm can now be produced.
  
  
• Processing trials on 2mm thick stainless steel have demonstrated that the beam can be used to drill and weld material and to produce features using the Surfi-Sculpt process with dimensions of ~20µm. This may open upapplication areas in medical and electronic device manufacture as well as other specialised, fine-scale processing tasks.
  
  
• Further improvements are possible by further cathode and gun refinements.
  
  
• Beam pulsing would be expected to increase further fine-scale drilling, cutting and machining quality.

Recommendations

Industrial manufacturers should consider the use of high brightness EB for the following specific applications:

• Material surfacing for features of down to 50µm.
• Micro-welding with weld widths down to 20µm.
• Material drilling and cutting for features of less than 20µm.

Acknowledgement

This work was funded by Industrial Members of TWI, as a part of the Core Research Programme. The authors are grateful to colleagues in the EB team for their advice and in particular to Matthew Robson for manufacture of the precision gun components.