Getting secondary beams and broken crucibles when trying to e-beam evaporate Ag in alumina crucibles
June 8, 2012
I suspect the secondary beams are related to the crucible charging and deflecting the primary beam.
Crucible cracking may be due to electrical stress (caused by the e-beam finding a high resistance in its return path to the power supply) or mechanical stress (caused by thermal expansion coefficient differences between silver and alumina as the crucible cools).
Try using an electrically conducting crucible such as Fabmate or a refractory metal (Mo, Ta, W)
Where can I source Technical Info such as "Thickness of deposition materials to coat a substrate"
June 1, 2012
We would be pleased to send an Excel spreadsheet that estimates the volume and mass of elements and (some) compounds required to make given number of films, with a given thickness, at a given throw distance.
This Excel sheet is available at no charge by emailing firstname.lastname@example.org with the subject Film Mass. Make sure to include your name, postal address, email address, and phone number. Without that information, regrettably, we cannot send the spreadsheet.
My CF viewports need cleaning for UHV (<10E-11 torr). Can I use Deconex 15-PF soap?
May 25, 2012
The best method for cleaning a viewport depends on what must be removed from the surface.
If the viewport is just covered in finger-grease or dust, then ultrasonic and dilute Deconex PF15 should be adequate (for solution details see http://www.borer.ch/pdf_files.php?type=prodinfo&file=pdf_files/Laboratory/english/Product%20datasheets/deconex_15_PF-x_EN.pdf). Standard CF viewports do not contain any materials that will be attacked by alkali in this product. But remember to flow rinse (and ultrasonic rinse many times) with DI water to remove all traces of the surfactant.
If the viewport is coated with materials from thin film deposition processes, then you need to know what those materials are and take appropriate steps to dissolve or mechanically remove them. Such films typically have extremely good adhesion and just using ultrasonic cleaning with detergent and water is a waste of time.
Want to transfer LOX and LH2 through a test coupling inside a vacuum chamber. What about leaks?
May 18, 2012
If the vacuum chamber surrounding the joint is capable of reaching 1E-6 Torr, one way of detecting a leak during liquid cryogen transfer is to use a residual gas analyzer (RGA). Without leaks, a vacuum chamber at 1E-6 Torr usually shows H2O (vapor), H2, N2, CO, CO2, and perhaps some 'hydrocarbons' (HCs).
At 1E-6 Torr there is a little O2 in the ‘normal’ spectrum. If O2 is leaking from the coupling it will show up as an increased amplitude in the peak at mass 32. Interestingly, there are very few materials (other than O2) that produce mass 32 so an O2 leak is easy to detect.
On the other hand, at 1E-6 Torr, H2 at mass 2 is often a major peak in the residual gas. My suggestion is: before starting the cryogen run, make a background spectrum by first blowing say N2 or Ar through the coupling at room temperature. Then, when the LH2 flow is started, you should be able to detect an increase in the mass 2 peak amplitude if any H2 is leaking from the coupling.
There are a few compounds that also give mass 2 ions. They include H2O vapor and HCs. But your application has an interesting side benefit. At LH2 temperatures, the exposed joint will act as an excellent cryo pump for both H2O and HCs. They will freeze out on the coupling and contribute almost nothing to the mass 2 peak.
What is the best o-ring compound for a space application?
May 11, 2012
At Temperature range:- -55?C to 100?C
I suggest you go to http://www.parker.com/literature/ORD%205700%20Parker_O-Ring_Handbook.pdf and search the document for < low temperature >.
I regard Parker's information as highly reliable and it indicates (page 2-5) Viton for static seals is limited to -26?C (page 2-8) The table suggests fluorosilicone rubbers and 'low temperature' butyl rubbers may be good from a little below -50?C and +100?C. Note, those specs almost certainly require the o-rings to be specially 'compounded' which is OK if you are buying a million of them . . . but just a couple? Hmmm!
Does a needle valve shut off when your O2 free glovebox becomes contaminated?
May 4, 2012
I am not aware of any fine control needle valve that ‘shuts off’. The needle mechanism is for flow control and when fully closed the valve still has a very low flow. The most common failure mode for these valves is breaking the needle by over-torquing the handle trying to stop the leak.
One way to prevent the through leak is to add a genuine shut-off valve in series. But this introduces its own issues. Is it better to:
Put the shut-off valve upstream of the needle valve? (Problem – it takes forever to pump out the gas in the volume between shut-off and needle valves)
Put the shut-off downstream of the needle valve? (Problem – the volume between needle and shut-off valves will probably reach cylinder regulator pressure and there will be a gas surge into the chamber when the shut-off is opened.
How do I get an efficient pumping speed using an inlet filter on a system with particles?
April 27, 2012
Many points can be made about particles and filters:
For particles to move in a gas flow means the gas pressure is high enough to entrain/move them. (It is the gas equivalent of Brownian motion.)
A good filter will trap some fraction of the particles it encounters but rarely all.
Filters are often specified as capable of trapping down to a certain particle diameter. Below that diameter trapping is questionable.
There is no way around it. Inserting an inlet filter introduces an extra conductance element which will reduce the effective pumping speed (EPS) from the chamber.
The EPS reduction can be measured by determining the upstream pressure (produced by a known constant gas flow) with and without the filter element in place.
I strongly recommend you add an oil filtration unit to your pump which may remove some fraction of the particles that bypass in inlet filter.
Ever wondered how to mount a polycarbonate cylinder onto a flat metal flange?
April 20, 2012
In my opinion the best information source for o-ring seals is Parker Hannifin Corp. That company's handbook is available in pdf form so search the net for "Parker O-Ring Handbook".
It used to be (and may still be) printed which, if you are a heavy user, is actually the faster format in which to find stuff. You disagree? OK, what’s the algorithm for pages riffling under one’s right thumb?
Will a standard, Cu gasket CF flange leak at -273degC with 50psi internal He pressure?
April 16, 2012
We are aware that some researcher do indeed use ConFlat®) flanges with high internal pressures. But we decided many years ago this practice is so potentially dangerous (and un-testable) that we could not provide any information about using vacuum flanges at internal pressures in excess of 1 atmosphere absolute.
I write 'un-testable' since one issue of pressurized flanges is, obviously, the strength of the bolts. In our experience many people make CF flanged joints with 'available' bolts, without regard to tensile strength, thermal expansion coefficient, or frangibility. Cooling the bolts, flanges, and gaskets to temperatures close to absolute zero may significantly change the room temperature values of these any other parameters, increasing the uncertainty and potential danger.
Can I exhaust a rotary vane pump into the room air?
April 5, 2012
While exhausting a vacuum pump into the room can be (and is frequently) done, it is NOT advisable for at least two reasons:
(1) Any gases injected into the chamber, of course, appear at the pump's exhaust (at atmospheric pressure). If these gases are toxic (H2S) or spontaneously flammable (SiH4, PH4), checking the MSDSs should alert you to the dangers. But what about gases like H2 or CH4 both of which, when mixed with air, are not particularly toxic but are highly explosive? Or what about 100% O2? At atmospheric pressure it makes many, otherwise innocuous, materials highly flammable or explosive (including the pump oil vapor).
[I’ve written elsewhere, I once worked with a lab clown. His favorite trick was to dunk a cotton wool ball in LOX (which we had in this lab), drop the ball onto the floor and then drop a lighted match on it. In the 100% O2 atmosphere, the cotton wool ball exploded!]
(2) When evacuating the chamber from atmospheric pressure, an oil sealed pump initially blows a 'mist' out of the exhaust. It is really an aerosol of the pump oil. While I'm don’t know the health hazard associated with breathing hydrocarbon oil aerosols, I have a VERY strict policy about not wanting my colleagues, or myself, to be experimental test animals.
My advice is, depending on local, state, and federal requirements, have the pump's exhaust port: plumbed to a fume hood; plumbed to an exhaust gas abatement system; or plumbed out of the building. If gas abatement isn't needed but you are pumping explosive gases, make sure the exhaust line is diluted with a large flow of diluent gas (N2).
Trying to mount a calibrated Ne leak on my 40L/s ion pumped chamber?
March 30, 2012
If this ion pump system exists, unless the pump is a 'noble diode' (also called a differential ion DI pump with one Ti plate and one Ta plate) type, its pumping speed for neon may be almost zero. According to http://www.cientificosaficionados.com/libros/CERN/vacio3-CERN.pdf even if it is a noble diode the pumping speed for Ne might be 15 - 20% of the given N2 speed.
I don't recall ever seeing data on the life-time of a pump at a (relatively) high Ne pressure. For N2 a typical lifetime at 1E-6 mbar is quoted as ~30,000 hours, meaning at 1E-5 mbar the lifetime is 3,000 hours (about 18 weeks). But that is for N2 . . . Ne could be less since its pumping mechanism is burial, not reaction with Ti.
The 'equilibrium pressure' (EP) is easily calculated from Gas In = Gas Out. Ignoring all other gas load sources, the mass flow into the system is the calibrated leak's rate 1E-4 mbar.liter.sec. The mass flow out of the system is (at best) 0.2 x 40 x EP mbar.liter/sec or 8 mbar.liter/sec. So the EP is given 8 * EP = 1E-4 mbar, or EP = 1.25E-5 mbar.
Being cautious, let me emphasize, I've taken the max values quoted. That is a 40 L/sec DI ion pump pumping a Ne leak of 1E-4 mbar.L/sec will give a base chamber pressure of 1.25E-5 mbar. Anything other than those ‘max value’ conditions will produce a higher base pressure.
Trying to deposit Calcium Carbonate on a surface by PVD to mimic an egg shell
March 23, 2012
Calcium carbonate (CaCO3) loses CO2 at 800 - 900?C producing 'quick lime' (CaO). Attempting to evaporate or sputter CaCO3 will convert some fraction (of the vaporizing material) to CaO which will contaminate the film, perhaps to a considerable extent,.
While CaO will convert to CaCO3 by the addition of CO2, there are issues including CO2's diffusion from the film's surface into its interior. I don't know anything about the rate of such processes so can't comment if it will be fast or complete.
In my opinion, attempting to make a CaCO3 layer by thin film deposition techniques is the wrong approach. Have you considered wet chemistry methods such as blowing CO2 into a solution of a (soluble) calcium salt? Or having a thin layer of powdered CaCO3 compressed (but not heated)
Having little success thermally evaporating Ti and Ni?
March 16, 2012
Both Ti and Ni are particularly difficult to thermally evaporate.
A material's 'evaporation temperature' is often regarded as that needed for its equilibrium vapor pressure to reach 1E-2 Torr. At that vapor pressure, the deposition rate on a substrate in a system of 'normal' geometry is good/high. For Ti that temperature is ~1750?C and Ni ~1510?C at which temperatures both are liquid
If thermal evaporation is attempted from a refractory boat, these liquid metals quickly alloy with the boat, destroying its electrical and mechanical properties (the boat usually cracks and falls apart)
Using a thermal evaporation crucible involves heat transfer across a number of interfaces (from heater to crucible’s exterior; through crucible; from crucible interior to evaporant’s surface; through evaporant to its evaporating surface). For the evaporating surface to reach evaporation temperature, the heater/crucible must be at a higher (perhaps much higher) temperature. And there is the addition problem: liquid Ti (and to only a lesser extent, liquid Ni) is a 'universal solvent' that reacts with and destroys most (all?) crucible materials.
A significant portion of the applied power is simply radiated to from the heater crucible outer surfaces to the cold chamber walls, which is why high input powers required and there are high thermal loads on the substrate.
One of our researchers had success using Ti pellets in a thick W boat as a 'one time' evaporation source. He first preheated the empty boat to the temperatures he would later use for evaporation. As I recall he was occasionally able to achieve 1000 angstrom thick films (using a short throw distance) before the boat fell apart. But note he was unconcerned about the substrate's thermal load.
One final point, to reduce power requirements for the source and reduce the substrate’s thermal load, appropriately placed thermal shielding (made of spaced, polished tantalum sheets) to reflect thermal radiation back to the heater/crucible or boat is a good idea!
Can I sputter or evaporate brass into a thin film on a substrate?
March 9, 2012
First, a little about sputtering and evaporation. They are different processes using different physical mechanisms to atomize the bulk's components.
Sputtering atomizes a solid (cooled) target by Ar ion bombardment. At a basic level, the rate at which an element sputters depends on its mole fraction on the target's surface and on its sputter yield - the number of atoms entering the gas phase for each ion hitting the target.
The National Physics Lab site lists elemental sputter yields showing Zn is ~4.6 and Cu ~1.75. That is, a 50/50 atomic mix on the brass surface would sputter Zn roughly 4.6/1.75 times faster than Cu so the vapor landing on the substrate is initially Zn-rich. But after sputtering a while the Zn’s mole fraction is reduced and the target will sputter with ~50/50 ratio . . . with one large proviso. The target must be cooled so Zn atoms don’t diffuse through the solid brass to the surface.
Evaporation is the 'atomization' of solid/liquid bulk by raising the temperature so an atom’s thermal energy overcomes its binding energy to the solid/liquid bulk. The rate at which a specific element evaporates is determined by its vapor pressure at the surface’s temperature (and, to some extent, on it mole fraction on the solid/liquid surface - but diffusion in solid/liquid state quickly changes that).
Using the equilibrium vapor pressure of the elements charts and choosing an arbitrary temperature, say 800?C, Zn's VP is ~200 Torr while Cu's VP is 2E-7 Torr. Obviously the vapor landing on any substrate will create a Zn film very slightly contaminated with Cu.
Other PVD methods that may convert brass from bulk to films with little change in composition are pulse laser deposition and flash evaporation. Both methods raise the bulk material’s temperature so fast and to such a high value that the huge difference in the equilibrium vapor pressures are inconsequential since there is no ‘equilibrium. Both elements rapidly vaporize.
Sputtering 50 nm SiO2 layers on plastic. Plastic a little darker after deposition. Why?
March 2, 2012
To me, there are at least four possible reasons why the combined plastic/SiO2 film may appears darker:
- The SiO2 film's morphology is poor with very small crystallite size, many grain boundaries, defects in the crystallites, voids caused by 'quenched' deposition of SiO2 on a (presumably) room temperature plastic substrate
- SiO2 film's is oxygen deficient, although I can’t claim that necessarily causes the film to darken
- the plastic film is reacting to thermal radiation with incipient pyrolysis, carbonization of pump fluids, continuation of the cross-linking process, or just plain chemical reaction with residual contaminants
- F-center formation in the plastic: again, I don't know if this mechanism exists in plastic films but if a 'diode' or 'unbalanced magnetron' source is being used, the argon ions and/or electrons hitting the film might induce colored f-centers.
Tests worth trying:
- Coat a glass micoscope slide and see is if it too darkens
- Potentially improve the SiO2 morphology by significantly reducing the deposition rate by reducing power or increasing throw distance
- Estimate the temperature of the substrate film's backing plate after deposition. Then heat the plastic (in air, initially) to that same temperature and watch for color changes
- If additional gas flow equipment available, add say 15/20% O2 to the argon.
I'm making a copper heat sink for UHV. Is welding needed or is brazing/soldering OK?
February 24, 2012
Depends on whether: (a) you are connecting copper-copper; or (b) copper to something else; and (c) the maximum temperature of the heat sink. For copper-copper welding is best but difficult.
Brazing is OK with a very big BUT: you must use brazes that do NOT contain cadmium, phosphorus, or similar high vapor pressure elements that are added to lower the brazing temperature. Often people think 'silver solder' must be OK because of the name. But beware! It is not a description of the contents which can, again, include high vapor pressure components.
One of the best sources for brazing information is a book from a bygone age "Handbook of Electron Tube and Vacuum Tehniques" by Fred Rosebury. It gives the composition of common brazing materials and helps with the selection of brazes for bulk materials.
Regularly lead-tin solder is a no-no for UHV. However, if the application is for a low temperature heat sink, indium might work. It melts at 156 deg C, has a very low vapor pressure, and wets copper when means it solders copper just fine. But watch out if other metals since indium may not wet them or may alloy with them sometime causing structural failure.
When I cycle my 240V heater block between 10E-3 Torr and atmosphere the wires arc. Why?
February 17, 2012
If you are not familiar with Paschen’s Law, read http://en.wikipedia.org/wiki/Paschen%27s_law It shows how a gas’s breakdown voltage is high at low pressures and at near-atmospheric pressures, but in the middle is very iffy. The chart is a bit weird to use since the x-axis is pressure x distance between surfaces at different potentials, presumably between the 240 V live wire and the neutral wire or any grounded surface.
In this case there's an additional complication since 240VAC is the rms value and the peak voltage will be (if I recall my AC stuff correctly) 240/0.707 or 340 volts
What is a 'virtual leak'?
February 10, 2012
A virtual leak is any volume that will reach atmospheric pressure when the vessel is vented but has an extremely low gas conductance to the vacuum vessel. The classic example is a blind tapped hole. The conical shape at the bottom (created by the drill bit) has a relatively large volume.
If a regular bolt is used, the gas conductance path up the mated threads to the vacuum vessel is so low that, under high vacuum, gas leaks out very slowly. This adds to what is called the vessel's 'gas load'.
For a chamber with a fixed 'effective pumping speed' (and most are) the gas load determines the base pressure. Double the gas load and the base pressure doubles and few vacuum users are happy with that.