In our latest experiment, thin films deposited on a glass plate are put at the bottom of a vacuum chamber and irradiated by a Mercury lamp (500 W electrical power). The vacuum chamber is cooled down by water flowing in hollow spaces in contact with the internal surfaces of the chamber. Radiation enter the chamber through a UV-graded quartz window to allow ultraviolet (UV) irradiation of films.
After the irradiation onset, temperature of the glass plate raises at an impressive rate of 15 °C/min. Our temperature sensor saturated at 125 °C after just 7 minutes, while temperature was steadily raising.
High temperatures above 100 °C can damage films and must be avoided.
The main problem is the reduced surface contact between the glass plate and the cool bottom of the chamber, which limits the thermal exchange. Does it exist a (soft) material which is compatible to vacuum (mbar), resistant to high (kJ) UV dose, and has good thermal conductivity to improve thermal exchange between glass plate and chamber?
Check for themal interface materials with silicone base: https://www.parker.com/literature/Chomerics/Parker%20Chomerics%20Thermal%20Catalog.pdf
Also look in online electronics shops for liquid metal thermal interface materails for processors (gallium and gallium-based)
Are you able to supply a photo of the situation? What I assume - though not written with the necessary clarity: a non-flat surface of the chamber's bottom side.
Installing some metallic (eg. Al) adapter with a flat surface would be one of the easiest means, also avoiding silicone contaminations.
Eventually you will run into thermal stress issues with your glass substrate...
Dear Paolo,
at these temperatures you can try soft metall (Indium) or Field's alloy. Choose indium if your glass plates are thick enough to press an indium sheat and provide a good contact (http://www.espimetals.com/index.php/online-catalog/854-indium-foil-sheet).
If your glass plates are fragile try Field's alloy. It melts at 62 C, therefore you will get very good thermal contact between a plate and a cold bottom of your chamber.
Best regards,
Roman
Dear Andrei, thank you for your suggestion. We are considering the various items of the catalog you provided one by one. Unfortunately, liquid, gel and paste have all high contamination potential compromising the actual test performance. A solid and dry interface should be the solution, but then the issue of small effective contact surface remains.
Dear Dreher, the surface of the chamber' bottom side is flat. The main point is that, as a rule of thumb, the effective area of contact between two flat but non polished objects is about 10-4 of the surface of the smallest object. In our case, if the glass is, say, 10 cm2, the effective area of contact is 0.1 mm2. This really small contact surface is a big impedance to transfer heat from glass to the sink chamber. We would need a soft interface to increase the contact surface. However, most soft materials are not designed to work in vacuum (evaporation, sublimation, and then contamination) nor to withstand UV / visible irradiation.
Soft and thin metallic sheets could be considered. Thermal stresses of the glass substrate is a second step problem, that can be managed by choosing the appropriate material and thickness.
I second the use of indium - or if the budget is tight, try some lead sheet.
(I've used that in situations where I needed to cool a rotating object in UHV conditions)
Indium looks like suitable but it is sticky enough to prevent comfortable takedown. Maybe tin, aluminium or silver foil is more suitable. Annealed copper foil with pre-oxidized surface is also attractive.
Graphite (random oriented) is promising, it is used in gaskets as it is easy deformable and have relative high thermal conductivity. Special grade SPG40 (http://www.ceramisis.com/carbons_graphites_specs.htm ) have Youngs Modulus 11 Gpa (Indium ~10-13 Gpa) and thermal conductivity 140 W/mK (Copper - 386, Aluminium - 210, Indium - 147. Helium - 0.142 (1 atm), and much lower at reduced pressure).
PANASONIC produces Pyrolytic Graphite Sheet - flexible graphite foil with very high thermal conductivity in a-b plane - 700 to 1950 W/(m·K) (2 to 5 times as high as copper, 3 to 8 time as high as aluminum), and moderate in z-axis (15 W/(m·K)) http://www.mouser.com/pdfdocs/thermalgraphitesheets.pdf
UV-stable vacuum grease with enhanced thermal conductivity https://docs.google.com/viewer?url=patentimages.storage.googleapis.com/pdfs/US20050233912.pdf . Patent contain procedures to achive low vapor pressure of common vacuum fluids/materials.
Similar thermal grease and gel is produced by AiT https://www.aitechnology.com/products/thermalinterface/thermgel/ - please contact to talk about details
http://web.ornl.gov/info/reports/1976/3445605506342.pdf
https://na.industrial.panasonic.com/products/circuit-thermal-protection/thermal-protection/pyrolytic-graphite-sheet-pgs
Ciao Paolo,
I have a similar problem in one of our HV chambers. After some web search we decided it might be worth a try to cool our specimen using gas. in our case we would firmly clamp the specimen to a plate having several through-holes through which a noble gas is let in. Sure, your pressure level in the chamber will rise, but since it is a noble gas it might not be an issue. Depending on your process you may also try N2, which should have in theory a better performance in cooling, being a molecular gas.
Please be aware though, that the above is not an experience report, but just an idea.
Dear Paolo, I have read all the suggestions above. As for me – the soft metal intermediate layer (together with surfaces polishing and pressure application to make them closer to each other) sounds good, as well as the cold gas cooling. And may I suggest another possibility:
Maybe it will be possible in your installation to attach (with some holder) the glass plate with the film not directly to the chamber bottom but to a Peltier – Thermoelectric Cooler Module, then to attach the opposite side of the module to the surface cooled by water and to let the module work…
First of all, I wish to thank everyone took his time to read and answer my question. Me and my colleagues are in debt with all of you. Every proposed solution has its pro and contra.
Probably I did not stress enough one of the boundary conditions, namely everything we put in the vacuum chamber will be exposed to huge dose of infrared, visible and UV radiation, 24 hour a day, seven days a week, for at least three consecutive weeks. Only considering UV, we are speaking of kJ dose, that one of visible will be much more. As an example, every black material would absorb a very huge energy from the visible part of the lamp spectrum, possibly changing its state and/or characteristics, and therefore is not suitable for our set up.
A soft-metal interface having a high thermal conductivity, to be sandwiched between glass plate and sink chamber, is a good solution only if we can uniformly and tightly press the "sandwich" in order to increase the tree contact surfaces. The weight of the glass plate is not sufficient alone, films above the glass must be open to radiation (cannot be covered) and clamping just the perimeter of the glass does not solve the uniformity issue.
The ideal solution does not exist, probably. However, among those proposed, using a noble gas like He offers some apparent advantages: it is easy to implement, surely increases the heat exchange between glass and chamber, it is cheap, and does not affect irradiation, as UV is not absorbed by He (UV is partially absorbed by N2).
In the next days we will modify the chamber to allow He injection and will be back here to report the temperature measurement results.
Many thanks, Christoph!
If you don't care about the contamination of the glass, any low melting alloy (rose's, Wood's etc. should work. May or may not contaminate the inner surface ot the chamber's bottom, but does not evaporate.
Ga and In are a bit tricky as they will also stick to the glass itself, 'wetting' it.
What you may or may not have ignored until now: the glass itself is a rather good thermal isolator. Could you use another substrate like Al2O3 which is more stable in mechanical terms - still not being the best thermal conductor. Or some silicon wafer with a SiO2 top layer - adding mechanical stability to a very thin - thus comparably well conducting - 'glass' layer.
We'll watch your results
Dear Paolo,
In thin film technology & semiconductor industry temperature control of the substrate always has been an issue. The more advanced options are electrostatic clamping making use of an ESD chuck (expensive) or mechanical clamping and backside cooling with gas (He is mostly used in RIE etching). A very simple trick I used in R&D is applying vacuum grease to stick the glass plate to a high heat capacity substrate holder (e.g. Al). The grease does not degass and contaminate your vacuumsystem (very low vapour pressure), gives very good thermal contact and can easily be removed after the processing.
Dear Paolo
we had a similar problem (although opposite in goal) trying to maintain glass substrate with a:Si films at 500 C during laser recrystallisation in vacuum. Our solution was to let pure helium flow from under the glass substate through a controlling valve and forming a sort of inert gas cushion capable of heat flow. This of course reduced vacuum, but we could find a good compromise with substrate temperature control. It worked for heating, don't know for cooling...
I did update my hit answer (thanks to all upvoters) - maybe additional info can help for current challenge or interesting for related applications. According to preliminary evaluation, helium cooling can also work. 100W removal need helium minimal $8,000 per week (GAZPROM price $50L/kg for l-He, helium input at 25oC and some atm pressure by slit gap, output at 100oC in vacuum without recycling). With recycling using available equipment , helium cooling is highly competitive.
Dear Roberto and Andrei,
Is there another reason apart from economical to use He instead of, e.g. Ar?
Dear Paolo,
do you have any information on the UV emission transmission of your UV-graded quartz window? I think it is not transparent for deep UV emission (shorter than 200 nm). Therefore you can use nitrogen gas. Moreover, liquid nitrogen is much cheaper than liquid helium.
Best regards,
Roman
@Cristoph -> Liquid helium is ~300x more costly in comparison to argon (compressed gas is less suitable because of higher cost and inconvenience of handling in such application). But helium is UV transparent and do not produce reactive species (in contrary to other gases) and was used as background for UV-spectra measurement for argon: Kouichi Yoshino, "Absorption Spectrum of the Argon Atom in the Vacuum-Ultraviolet Region," J. Opt. Soc. Am. 60, 1220-1229 (1970)
https://www.osapublishing.org/josa/abstract.cfm?URI=josa-60-9-1220 , http://sci-hub.bz/10.1364/JOSA.60.001220
http://iopscience.iop.org/article/10.1088/1742-6596/635/9/092087/pdf
How gas ionization have influence on Paolo's process - I have nothing in mind, maybe it cause reactive etching. Hydrogen gas have 2x higher heat capacity (over He) , 3x lower viscosity and 25x lower cost - it should be best gas in cooling use
Dear Christoph,
we used He mainly because it has a good thermal conductivity and we had vacuum system issues that forced us to minimize gas flow. But, to be honest, when it worked we didn't try other solutions. Maybe Ar can work as well
http://www.engineersedge.com/heat_transfer/thermal-conductivity-gases.htm
Roberto, Andrei,
Thank you for your reply. Andrei, yes, I forgot about the UV absorption in this specific case. Thanks again!
Dear Andrei,
thanks for good references on Ar atom absorption in VUV range. Really these works about emission with wavelengths shorter than 100 nm. Such emission can't penetrate into the vacuum chamber through a quartz window. A UV-graded quartz cuts off any emission below 160 nm -
https://www.pgo-online.com/intl/katalog/curves/quartz-glass-transmission.html
Therefore, argon gas is a convenient candidate to replace rather expensive helium gas.
H2 gas is not applicable because a hydrogen molecule has the dissociation energy value about 4.5 eV and many strong absorption bands within the wavelength range under discussion.
Best regards,
Roman
@Roman -> H2 is best cooling gas, but I never proposed it for Paolo's process.
Argon is cheap but have 1/10 of thermal capacity and thermal conductivity from helium ( http://www.engineeringtoolbox.com/specific-heat-capacity-gases-d_159.html ,http://www.engineersedge.com/heat_transfer/thermal-conductivity-gases.htm ). Argon is also10x heavier - may be it mutually compensated at equal pressure and intensive mixing. In any case, gas managment in vacuum chamber is much more complex than simple pad from appropriative material. IMHO
I understand (and support) your idea about UV shielding of guartz, moreover, irradiated film is deposited on ordinary glass (strong UV filter, cutoff~ 280nm). It is mean that thermal interface materail (on opposite side of glass) should be low irradiated = should be stable (acrylic-based pads seems to be problematic, but PDMS- and perfluoropolyethers-based thermal interface materials must be OK).
Roman, our lamp is a 500 W electrical power mercury lamp. Its emission spectrum has several lines overlapped to a continuum. The shortest wavelength line is at 250 nm, and actually it is the less intense one. As a consequence, when choosing a gas, its absorption spectra in the VUV is not relevant in our case. We tried to find spectral absorption of Ar at UV wavelengths, without success. However, Ar is the component of several excimer lasers, among them the most used is ArF emitting at 193 nm, and as a consequence we guess Ar shouldn't exhibit strong absorption in the UV.
O'k Paolo,
you can also try N2 gas. "Dry" N2 gas from evaporating liquid nitrogen (the H2O content is very low in this case) will be very convenient due to a rather low price.
Roman
Dear All, yesterday late in the afternoon we filled the chamber with 1.1 bar He. No leakage was detected overnight. This means the small 0.1 bar overpressure is maintained well by ordinary o-ring.
Today we have performed temperature measurements vs. time in the chamber filled with 1,1 bar He (static, no flow). We decided to use a small overpressure of He to avoid accidental air (oxygen) contamination.
After switching the lamp on, temperature of the glass sample raised from 17 °C to 34 °C in 6 minutes, to stay constant in the following 40 minutes, until we switched the lamp off. Temperature dropped down back to 17 °C in a time faster than the 6 minutes "risetime" of temperature. Pressure remained constant at 1,1 bar during the test, and
We can conclude that static He at 1 bar is effective in transferring heat from the sample to the internal surfaces of the chamber, which are cooled by flowing tap water at 17°C.
Now we can start the 3 weeks no-stop irradiation of films deposited on glass.
Further comments and suggestions are welcome.
Thanks everyone
Paolo
Heat removal with helium at normal pressure is not news - it is used in high-speed HDD's for last years. Right question should sound ~ "How to remove heat upon strong UV-C radiation" - maybe remark (P.S.) below of initial question is reasonable - to avoid misunderstanding of real problem.
I was sure before last Paolo's message that vacuum is essential for current challenge. Information from topic is indexed by search engines, so no need to remove initial question and discussion - maybe some people will really need solutions for vacuum work.
Andrei, vacuum is preferred in our experiment, in order to avoid any interference of impurities (always present in every gas) as well as absorption of UV radiation, as even the smallest absorption would increase the total irradiation time to reach the wished total UV dose on films. However, once understood from discussion that it is very difficult to find a practical / low cost / UV sustainable solution, we have chosen the lesser of the evils: He gas at 1 bar does not interact too much with UV with respect to refraction, absorption, convection, and offers a cheap and effective high thermal exchange with sink chamber.
Obviously, our solution is not a universal solution: as an example, if the internal surfaces of the chamber wouldn't be cooled by flowing water, probably the final temperature might be much higher than 34 °C we have achieved.
Paulo,
Rethinking the problem.
Why not cool just the film deposition surface, rather than cooling the whole chamber?
Make a feedthrough.
In the past I have made, using CF and KF fittings, fluid loops that carried water and even liquid nitrogen to cooled stages in vacuum systems. (I've even built a water-cooled loop so that I could operate a Heraeus deuterium lamp *inside* a vacuum chamber,,,)
A feedthrough can be made from a loop of copper pipe brazed or even soft-soldered to the drilled stainless flange - and then attach the deposition plate to the loop (soft solder, indium, etc)
Bolt loop fitting to chamber, run chilled water, deposit film, turn on Hg lamp.
James, rethinking the problem is always useful and welcome. To this end, I re-read the 29 answers given by highly qualified Colleagues who participated to this discussion, one by one. In summary, the answer to my question "Does it exist a (soft) material which is compatible to vacuum (mbar), resistant to high (kJ) UV dose, and has good thermal conductivity to improve thermal exchange between glass plate and chamber'" is No, probably it does not exist. Most of proposed solutions are not suitable to kJ exposure to UV, and/or potentially create contamination, and/or are very expensive, or even time-expensive. As in every project, we too have sharp deadline, unfortunately.
Cooling the film substrate by fluid loops as you suggest is possible but it is time expensive, and possibly does not solve the problem of a good thermal exchange due to the small contact surface between the glass substrate and the loop (the 10-4 law I mentioned in a previous comment).
As an update, we are at the third day of exposure, the substrates temperature ranges between 34 °C and 28 °C depending on the substrate position on the bottom of the chamber. We never refill the 1,1 bar He.
Paolo,
Time is indeed the enemy. If, or when, time permits, I am happy to describe in more detail my adventures with cooling fragile glass objects (D2 lamps) in vacuo.
I wish you well - you seem to have found a very attractive solution. It would be worth writing your progress up, perhaps, for a UHV conference where such solutions would find many receptive ears.
You are welcome , Utpal. I am following your nice works since the time of the Self Filtering Unstable Resonators, a long ago, do you remember? All the best for your present endeavors,
Paolo
Dear All, I have just posted in RG the full text (in Italian) of a detailed (27 pages) technical report on the UV irradiation of polyimide films which was the main frame of this question.
In the report I have recognized the help of all of you to clarify some issues.
The report is in Italian, the abstract in Italian and English. At the moment I do not have much time to translate the 27 pages in English. In the future it could be reformulated and summarized as a Proceedings paper, provided someone is interested in our results.
Thanks everybody
Paolo
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages