The Blumlein could feed a tube to emit nanosecond X rays pulses
. The impedance variation of a line shall provide the high voltage. Sketch here under if you're logged in.
start with +40kV as an example. Al2
(K=9 and 30MV/m) between r=50mm and R=53mm, 100mm long, stores 0.67J at 1.2ohm line impedance, with 2ns round trip. When the first line is about discharged, the second gap fires, creating a -20kV 17kA step at the start of the second line.
The second line tapers to r=8mm R=29mm in vacuum for 77ohm line impedance. Gaussian impedance transitions preserve pulse shapes, ask your local RF specialist. The step has -160kV amplitude, so from initial +40kV it feeds the tube with -120kV 2kA over 2ns
Fields are amplified where the dielectric leaves the surrounding ground. The lower permittivity material shall alleviate this. Silicone rubber at least coexists friendly with ceramics.
needs about 15bar to fire at 40kV over 3mm. Its light (337nm=3.68eV) lets LaB6
(2.70eV) and CeB6
emit reported 7ppm electron per photon, enough to seed the avalanche. If the borides survive, this would favour the quick photoemission breakdown. ThO2
(3.03eV) or La2
(2.71eV) filling W would be more durable, but they will see few photons.
Traces of a gas more easily ionized favours the photoionization breakdown. This worsens the gaps' firing voltage; would it accelerate the breakdown?
Electrodes of different radius favour a clean firing. The proper polarization is opposite to an avalanche particle detector... The odds through reasoning being 50-50, just experiment which polarization.
Mirrors shall favour photo-everything breakdown, while a screen separates the gaps.
To emit X rays
, the ancestor machine (with a Marx generator instead of the Blumlein) described in US Patent 6,166,459 has whiskers at the cathode, whose vaporization due to field emission current makes a plasma able to carry the current. Maybe the description of a natural process where the whiskers reappear at each cycle.
This process works, it may even be unavoidable, and must be considered for the present attempt. I describe now a strict field emission alternative, which is uncertain since vacuum breakdown isn't well understood.
The D=2mm cold cathode has 400 photolithographic cones of r=7µm to R=50µm tip to base. I take Cu (4.65eV), but Mo or W are possible. 120kV split as 92kV in straight 1.3mm (adjust as cones wear out) and 28kV in 400 conical paths of pi steradians achieving 3.5GV/m at the r=7µm tips. Fowler-Nordheim lets each emit 5A. Very thin LaB6
at the tips would ease emission but must be less durable. La2
powder in W would need grains far under 14µm or be a film.
5A drop 10mV in each copper cone, and the heat spans only 500nm in 2ns. I consider electrons cool the cones: they enter at 26meV and exit at 26+10meV. Since the hotter electrons evade more easily, they could carry even more heat away. At peak 32mA/µm2
, electromigration isn't a worry.
The anode has a 40µm W film to stop the electrons. It could be 238
U. The film could be just thinner than 29µm to improve the X-ray spectrum. The electron pulse releasing 0.5J in d=2.5mm *20µm would heat the film momentarily by 2000K, but heat exits the 20µm within the 2ns. A 2mm Al (or Cu) window hold in a cone can evacuate mean 150W heat - or make it thicker, possibly of Be, or cool by fluid. 150W permit 300Hz sustained repetition rate.
Converting 1% in photons, of which 1/10 * 1/30 have the proper energy and direction, gives 109
photons per bunch. An object absorbing 9/10 and a 100,000 pixels detector 1/4 efficient leave 250 photons per pixel.
The design is adaptable
somewhat. Spike radii at the cold cathode adjust the breakdown voltage and the anode distance. Other pulse energies and voltages are possible, and a switchable gamma source welcome; insulating 1MV shortly in vacuum isn't that bulky.
The lines are less adaptable. Twice the 77ohm is already difficult, but crimping achieves less than 1.2ohm. More than 40kV is possible. Ferrite cores are not obvious.
for this source? Imaging of course, and 300 frames per second are by far not the limit. But what more?
- Study how vacuum insulation breaks down over a short time.
- Observe X-ray fluorescence, with a strong 100keV quick source.
- In backscattering or fluorescence imaging, compensate the illumination versus the distance, as previously described for light. Vary the gain through the detector's or photomultiplier's voltage.
- Any useful as the electron source of an accelerator?
Marc Schaefer, aka Enthalpy