I'm not sure if lasers with such wide beams (1 cm in diameter) exist? Anyway, thanks for the hint, I'll try to google it.

highly exothermic chemical reactions

Ehab, we need steady and controllable heat release. I think chamical reactions are not good for that!

Khaoula, sorry but it's not quite clear what you mean.

Electronic beam can be very efficient: one ampere accelerated with 5 kV will do the job. However, this needs vacuum on the heated side.

Other possibilities:

Concentrating solar radiation on a clear day from 5 m² with an mirror having a focal distance of about 6 m

Plasma torch my also do the job

As in idea?

http://www.cleanlaser.de/wEnglish/produkte/high-power-cl-1000.php?navanchor=2110014

Dear Claude-Alain,

Thanks a lot for your valuable remarks.

>Electronic beam can be very efficient: one ampere accelerated with 5 kV >will do the job. However, this needs vacuum on the heated side.

Vacuum is not a problem. However, the price of such an accelerator may be a problem. I have no idea how much it may cost: 1k$, 10k$ or maybe 100k$?

>Other possibilities:

>Concentrating solar radiation on a clear day from 5 m² with an mirror >having a focal distance of about 6 m

Yes, we've been thinking about this solution. The main problem is precise tracking of the sun. There are standard tracking mounts for telescopes, but the price for payloads of about 100 kg (the approx. weight of the mirror) is over 10k$.

>Plasma torch my also do the job

We've tried to use a gas burner, but failed to create a localized and controllable heating.

Dear Ron,

Thanks for the link.

Seems that these lasers have short and powerful pulses, to get sublimation. In contrast, we need continuous beam of light.

I don't quite understand your task. It is possible to recommend plates on the basis of the silicified threads of carbon in a matrix of silicon for reception of considerable thermal power on small areas. But, unlike nichrome heaters, the power supply must provide high currents at low voltages . Our experience shows that to achieve a temperature of 1700 K at atmospheric pressure argon requires a specific power of 0.3 kW/cm2 Composite for the manufacture of the heater, we can provide as well as give recommendations for external electrical contacts. But the specific power of 2-5 kW / cm2 seems to be unrealistically high.

For the electronic beam heating, you need a heated filament, e.g in tungsten. This is heated using low voltage transformer. The primary of that transformer should be controlled by another variable transformer (e.g. VARIAC). The power needed here is a about hundred watts, and depends on the size of the filament.

The electrons emitted by the filament are then accelerated by a high voltage applied between the filament and your target to be heated. This high voltage should be provided by a 5 kW DC power supply . You can find on http://www.surplussales.com/PowerSupplies/PowerS-7.html such units for 4 to 6 thousand dollars.

Caution: Either the filament (- x kV) or the target (+ x kV) will be at high, dangerous voltage with respect to the ground! If it is the filament, the low voltage transfromer used to heat it should be well insulated between primary and secondary coils. If it is the target, it shoudl be insulated from the ground.

Dear Prof. Brantov,

Our task is as follows. Basically, we have a copper substrate. To the bottom side of the substrate we apply localized heating; from the top side we remove the heat using our advanced evaporative cooling system. Our cooling system allows to keep the surface temperature of the substrate below 130 deg. C under the heat flux of up to 1.2 kW/cm2. Our goal is to test our cooling system under much higher values of the heat flux (3-10 kW/cm2).

Im not sure if silicified threads of carbon are capable to generate such heat fluxes.

Dear Claude-Alain,

Building up such an setup seems too much complicated. Are there any alternatives which can be used out-of-the-box?

For a simple solution, I think about the plasma torch, much more powerful than a gas burner. Such torchs are used (among other uses) for argon-arc welding. Maybe you have also an oxy-acetylen torch in the neighborhood?

The difficulty is to measure the heating power applied to your cooler, since some of the power of the torch is dissipated in the room. But you can likely measure the evacuated heat!

Claude-Alain,

Determination of the amount of heat transported to the cooler is not a problem: we can measure it using 2 thermocouples embedded in the substrate.

The problem is localization of the heating. You need to transport the flame/plasma to the needed localized area of the substrate and then somehow to remove the remainings of the flame/plasma out of that area. When trying to do this in our experiments with a gas burner, the flame always died out (due to lack of the oxygen).

Dear Tillmann,

Yes, it's a good idea. However, I think that this solution has similar limitation as nichrome-wire-based heater, namely, huge temperature gradient along the rod. For example, the temperature gradient along the tungsten rod of 1 cm2 cross-section area under the heat flux of 5 kW/cm2, will be 500 K per mm! This means that tungsten rod with length of 7 mm or longer will melt. I doubt that the use of induction heating will allow to generate heat flux of 5 kW/cm2 in the rod less than 7 mm long.

Did you try with an oxy-acetylen or oxy-hydrogen welding torch? These can heat a a localised area about 1 cm². About 10 kW is achieved by burning one kg of acetylen only per hour.

Claude-Alain,

Yes, a welding torch seems more promising than a gas burner . We gonna try it. Thanks!

Can you reduce the size of the required surface area from 1cm2? This would allow you to use a smaller heat source (such as laser) but still get the same heat flux. Also, many pulsed lasers can produce MW/cm2 per pulse depending on the focusing and can also pulse very rapidly (kHz repetition frequency), so you could still get a continuous heating effect (similar to PWM for voltage)

William,

We can slightly reduce the surface area, say down to 0.25 cm2, but not more.

Yes, principle of PWM is applicable, provided that the power of a single pulse is not enough for melting/vaporization of the substrate.

I don't know if such lasers exist - providing uniform heating of 0.25-1 cm2 surface area with heat flux of 2-5 kW/cm2?

I believe some industrial lasers used for laser welding or cutting can deliver up to 5 kW, so they do exist. But they're big (and expensive).

As far as I know, all industrial lasers used for laser welding/cutting have the beam diameter less than 1 mm, while we need the beam diameter of around 10 mm.

I don't know if easy solutions exist allowing to increase the beam diameter up to ~10 mm.

You could use some diverging optics.

I doubt any conventional optics can cope with the beam power of 5 kW. Even 1% of the light adsorption in the optical element will result in 50 W of heat generation localized at the surface of the element with the area of less than 1 mm2.

Most laser cutters/welders already have collimating and focusing optics in the delivery head, and a glass cover to protect the lenses from spatter.

Dear all,

Thank you very much for your useful hints.

It seems that a powerful laser is the most feasible solution for my task.

I've asked a new specific question:

https://www.researchgate.net/post/A_powerfull_laser_2-5_kW_is_needed_for_intense_uniform_heating_of_the_copper_substrate_with_area_of_1_cm2_Any_suggestions

Thanks for your recommendation Dmitry. Seems I put you on the right track. Good luck with your further search.

Best regards,

Ron

Dmitry: My first thought was to move to a laser approach. But that might not be practical or cost effective. I believe it can be done though. It isn't a problem finding a CW laser capable of producing flux densities of 2kW/cm2 (~GW/cm2 are not uncommon). But the total power is relatively low because the laser cross section is narrow. But there are other issues that complicate an optical approach like reflectivity, absorbance, wavelength, etc.. Let me say finally it's quite easy to expand a 1mm diameter high power beam to 1cm diameter. Laser induced optical damage is normally not an issue (with careful design) since the damage threshold is typically quite high, in excess of 10's of MW/cm2. I like everybody's gas torch or plasma torch ideas in some ways. They're certainly capable of generating huge thermal powers. And they're relatively cheap. But lack of controllability is a serious drawback. But I think you should be able to do it using a resistive wire approach. I think that approach is not working for you right now because you might have a design problem. And I don't think the problem is with the material limitations of NiCr wire.

Might I suggest you generate the heat in a relatively large surface area, say 10cm2 or more, then use a tapered thermal conductor to concentrate the heat flux? Say you're using a 10:1 'compression' ratio. That means you should generate just 500W/cm2 initial power density to create a heat flux of 5kW/cm2. If you do some careful thermal design, you should be able to accomplish your goal. Anyway, that's the general idea. If you have access to any kind of Finite Element Modeling software, you might be able to further optimize the design. For example, the taper could more effectively be adjusted to an exponential rather than simple linear cross section. Also, with a FEM analysis, you might discover you can use a material other than copper as the thermal concentrator, one that has a higher melting point. You'll also be able to optimize the length of the concentrator. Finally, don't forget about the thermal resistance at the interface itself. Just having a material discontinuity can induce a large temperature rise. You'll need to polish the surfaces at the interface to a mirror like finish and apply a lot of pressure. Even so, I'd recommend some kind of thermally conductive intermediate material. Normally this would be done with 'thermal grease' (and there are a lot of choices), but at 130C as you say in your paper, that might have problems. Instead, perhaps a very thin (~25um) Indium foil? It won't melt at ~130C, but it is very soft and pliable, even at room temperature. This could significantly lower the thermal resistance at the interface. Best of luck Dmitry! You're project sounds interesting. Pat

Dear Pat,

Many thanks for your answer.

In fact, we do use a tapered thermal concentrator with the area decreasing from 7 to 1 cm2. Please find attached a sketch of the actual heating element we used in the experiment (in the paper we pesented heating element with an older design).

Careful analysis using FEM will certainly provide more optimal profile, but I don't think it will allow us to reach the heat flux higher than ~2 kW/cm2.

The main problem is high temperature gradients along the concentrator (even for copper). For example, the temperature gradient along the copper rod with 1 cm2 cross-section area under the heat flux of 2 kW/cm2 is 500 K per cm!

I'm not sure if we have any thermal resistance at the interface, as we have direct contact of liquid (water) with copper surface.

What is important to take into account, is the thermal resistance in the contact of the nichrome wire with the copper concentrator. In this contact we can easily have temperature drop of several hundred of K (at 1 kW/cm2 the nichrome wire is always in red-hot condition, while the copper concentrator was never seen at such condition).

I think you're right about the thermal resistance between the wire and the copper block. But the energy density is an important design parameter. If 7:1 is not getting you there, then 14:1 might. I'm not sure that a large temperature gradient is a bad thing. Indeed, you need it in order to drive the heat flux.

This problem is rather like the inverse of a problem I've been working on where I need to cool rather than heat. In my application, I machined a copper block in two halves and then machined a long serpentine path. Then I soldered the copper blocks back together. This made a rather primitive heat exchanger. Then I sucked cryogenic nitrogen gas through the block, cooling the block to liquid nitrogen temperatures fairly quickly. Just after the block, I used a simple coiled copper tube heat exchanger to warm the cryogenic nitrogen back to room temperature so as to avoid low temperatures at the pump.

In your case, perhaps you can circulate hot nitrogen through the block. It would be easy to pump the nitrogen through an oven with bare heating elements for very high thermal transfer to the gas. The melting point of nichrome is about 1400C, so it should be easy to create 1000C -1200C hot nitrogen. Copper melts at about 2000C, so there's no problem there. You can avoid a pump entirely by using a high pressure gas tank. If you use an open cycle, you can just dump the hot nitrogen back into the ambient environment.

My first thoughts along this 'working fluid' approach was to use a molten salt, similar to designs of nuclear reactors. But that's not a practical approach in this case. Then I thought that using nitrogen as a working fluid would probably work OK. I guess the issue is really about the heat capacity of nitrogen gas (pressure dependent!), and establishing an adequate flow rate. But it seems possible to me.

There are other gases too, argon, helium, etc. But those would require closing the flow loop and greater complexity. I guess it depends on if you need very long operation. If your tests run only for a short period, an open loop might be best.

Pat

Dear Pat,

High temperature gradient is bad because it induces high temperature rise along the copper block, which, in turn, induces nichrome wire burning and melting (provided the block is sufficiently long).

The solution with hot gas circulation is possible I think, but it's a bit too much complicated. The problem with nichrome is that it starts to slowly burn (due to oxidation) at the temp. ~ 800 deg. C, i.e. much lower than its melting point.

It is obviously not an easy task. I am afraid there is no ready to use solution.

I would imagine a sort of gas torch (acetylene) heating a piece of high temperature (alumina) ceramic ... .

Dmitry:

Excessive temperature gradient is bad, I agree. But without the gradient, no heat flux. That's just physics. If you want a large heat flux, you need a large gradient. The issue actually is managing the implications of that gradient. I hadn't thought of the nichrome actually oxidizing though. Good point.

Using a working fluid, like N2 side steps some (if not all) of those implications. If the nichrome wire is inside a nitrogen filled oven, no oxidation is possible because there's no oxygen. So, your nichrome wire is more stable at higher temperatures. Hence, you can run the nichrome wire at higher temperatures. As any thermodynamicist will tell you, the efficiency of a heat engine depends on the difference between the hot and cold regions. So, in this case, the hotter the better.

About complexity. In the scheme described, there's no moving parts, aside from the nitrogen gas itself. You will have to machine a serpentine path through the copper block though. And, although I used solder to join the two halves together in my design, that's not possible for your application. Even so, if the surfaces are reasonably flat, and you use a metal gasket (talk to any HiVAC person), the halves can be simply bolted together, producing a high quality seal. Even so, the pressures should be relatively low, a few atm. at most. All those things are very low tech, which is what you want.

Thermodynamically: Just consider for a moment, how much heat you can transport through a 6 mm ID tube? It's a rather straight forward (first order) calculation. I'm pretty sure you'll find that many kW can be transported from the oven to your copper block. It all depends on flow rate, density and temperature, heat capacity, etc. of course. You might first ask the related question: If I put an infinite heat sink at 1000oC up against the copper block, what's the heat flux (kW)? How must you change your copper block design to allow for your required 5kW flux to take place?

Just some thoughts ...

Pat

heat flux is, together with other factors, directly proportionate to material thickness ........ so your heat input mechanism, requiring a rod, rather than .05mm copper sheet, is self defeating.

Apologies the negativity

Dear Leonid,

It's not an easy task indeed, taking into account that the heat flux at the surface of the sun is about 6 kW/cm2.

The problem with gas tourch is how to make the heating localized. You need to transport the flame/plasma to the needed localized area of the substrate and then somehow to remove the burned gases out of that area. When trying to do this in our experiments with a gas burner, the flame always died out (due to lack of the oxygen).

The gas or acetylen torches have their own oxygen. Of corse the burned gases should be evacuated. IN normal use of such torches, the gases are evacuated to the sides of the flame.

Dear Derek,

Agreed. The thickness of the substrate is of crucial importance. A 0.05 mm thick copper sheet is too thin for incorporation of thermocouples. But a copper plate with thickness of several mm is probably OK.

Dear Pat,

Yes, the temperature gradient in the substrate is proportional to the heat flux, so we cannot reduce it. However we can reduce the temperature rise across the substrate, reducing the substrate thickness or increasing the substrate conductivity. With diamond rod it would be not a problem to get 2-3 kW/cm2 using nichrome wire :)

True indeed. Diamond just happens to be the best thermal conductor. If you've got the money, have at it! I've used diamond substrates in some of my work as a microwave IC design engineer. Synthetic, of course. But hooh pricey! Only the smallest of structures are practical, really. Still, no reason not to explore every option.

Best of luck!

I have already had to write a comment on this protracted debate. From the very beginning was incomprehensible task of its author. I was always busy in the field of development of high-temperature heaters up to 2500 K. all values of specific power never exceeded values of 1 kW/cm-2, and it was unattainable. I propose to consider the situation with the Sun, the color temperature of its surface is 5800 K. At full power radiation of 3.9*10+26 W and a radius of 7*10+8 m, the power density of the Sun is less than 6 kW cm-2 , which corresponds to the specified number. I can only express surprise attempts to reach these parameters when using resistance heaters of nichrome.

Dear Pat,

Diamond is too much expensive, of course.

Except from diamond, nanomaterials such as graphene and carbone nanotubes are also known to have extremely high conductivity. I wonder, whether it's possible to build a high-conductive substrate with the use of nanomaterials? Perhaps, a mixture of copper with carbon nanotubes?

Dear S. K. Brantov,

The key point here is the intensity of heat removal from the heated surface.

From the surface of the sun the heat is removed via radiation. In our case the main heat transfer mechanism is evaporation which is much more effective than radiation. Now we easily remove 1.2 kW/cm2 keeping the temperature of the substrate below 130 deg C. (the heat transfer coefficient is about 400 000 W/mK).

Nichrome melting point is over 1000 deg. C. This means that potential of nichrome-based heaters for our cooling system is far to be exhausted. I think that we are able to remove several kW/cm2 even using a nichrome-based heater, provided the construction of the heater is optimal. However, the use of a powerful laser is probably more convenient and feasible solution.

Dmitry:

It is tempting to look at ever more exotic materials and techniques. But you have to temper that with practicality, if you ever actually want to do something. You might have someone in the background with 'deep pockets' and if so, that's great. But I don't think that's the case. Take another, closer, look at the working fluid approach I described earlier. There are many reasons why virtually every hydrocarbon or nuclear power plant on the planet uses that method. Perhaps what I described isn't exactly what you need. But I'm certain that approach can solve your problem and do it in a practical way. After all, you did ask for an 'engineering solution' and that includes practicality.

Using a laser was my very first thought, as I've said. But I work a lot with lasers a lot too. In pulsed operation, it's easy to get many GW/cm2 energy density. But only for a few fs or ps. Recently, the average power of CW lasers does reach into the kW region (thanks Uncle Sam?). However, these devices tend to be extremely expensive. Other issues, like absorbance and/or reflectivity of the target will also affect how efficiently you can couple all that optical power into thermal energy within the target. So now you've yet another whole host of engineering issues to deal with that are really just tangential to your original problem.

Dear Pat,

I agree, the use of laser may cause a pile of extra issues.

Your suggestion with working fluid seems very promising indeed.

In fact, we did succesfully use liquids to heat substrates in some of our experinments (with small heat fluxes). The liquid was water with temperature below 90 deg. C. Taking another liquid (or gas) with temperature of about 1000 dec C. will give the heat flux of 5 kW/cm2 for copper substrate of 7 mm thick. Everything looks very realistic. The main problem is how to generate the flow of a liquid with temp. of 1000 deg. C. I would highly appreciate if you provide any thoughts/references.

Dmitry:

Well, my original suggestion was to make the system 'open loop'. That is, simply use one or more high pressure nitrogen tanks (i.e., 1500PSI or 100atm, pretty common) as a supply. Pick some kind of system operating pressure, the higher the better (for density/heat transport considerations), but not so high as to cause technical problems with sealing the system against leaks. Another consideration is that the higher the pressure, the harder it will be to establish a given flow rate. You might need multiple tanks and a manifold of some sort. I think you could probably express all this as some kind of differential equation or perhaps even an algebraic equation. Take the derivative and find the optimum balance of pressure/flow rate/ heat transport. Of course, none of that is really required. Personally, I'd just pick a pressure, like 2atm and calculate the required flow rate to give the required heat transport plus some margin of error. Don't forget you'll have losses to account for as well.

Closing the loop is attractive, but adds a whole new layer of complexity and money. I would have to say that unless you need extended operation, like hours at a time, closing the loop might not be desirable. Even so, closing the loop means finding a pump capable of producing the desired flow rate and operation at the desired pressure. That number may, or may not, be a very high value. I'm sure you've already seen that it's possible to get very high heat transport even through a 5mm or 6mm ID (1/4" ID ... ) tube. However, if you choose to close the loop, be sure to put a heat exchanger just before the circulation pump. You'll want to dump all that excess heat to the environment and lower the gas temperature before you run it through the compressor. The heat exchanger could easily be a simple helical coil and a fan. (I use something similar in my cryogenic project, but in the opposite direction.)

Finally, it occurred to me the other day is that what you want is the inverse of your cooling device. And, you're using a cooling fluid (a liquid) there, aren't you? In you're cooling device, you're pushing the limits of materials and thermodynamics. Here, though, you're free to pick a more practical, well established approach.

dear Dimitry

I am sure you have developped your cooler for a given purpose. Is it not possible to use the heat source you developped the cooler for?

Dear Pat,

Thanks a lot for this info! We'll discuss your suggestion.

Dear Claude-Alain,

From the very beginning we were motivated by the problem of cooling of high performance microelectronics. The maximum value of the heat flux in modern chips doesn't exceed several hundred of W/cm2. For our experiments we need much more.

Now we mostly focused on fundamental aspects of the problem. Under such extreme heat fluxes (~ 1 kW/cm2) the behaviour of thin liquid films was found to be highly unexpectable.

Why not try a gas burner of propane-butane or acetylene?

Dear Rashko,

This solution has been already discussed in this thread.

I copy my answer:

" The problem with a gas burner is how to make the heating localized. You need to transport the flame to the needed localized area of the substrate and then somehow to remove the burned gases out of that area. When trying to do this in our experiments with a gas burner, the flame always died out (due to lack of the oxygen). "

I think using gas-oxygen torch is very simple and applicable tool to produce concentrated localized heat flux at specified point with easy control.

good luck

Dear Amar,

Thanks. Yes, we'll consider this solution.

Dear Dmitry, please look at paper title

"Enhancement of Critical Heat Flux From High Power Microelectronic Heat Sources in a Flow Channel " for Prof Issam Mudawar 1990

I see in this paper in Heat Source Design method an attractive method to get high heat flux for their test chip with surface modification in order to get CHF point. Regards

Dear Mohammed,

Thanks for your reply. Their heater generates 360 W/cm2 only. We need 10 times more.

We can easily reach values of the heat flux in the range of 2-5 kW/cm2 on our experimental setup by releasing pulse electric current. In the case of existing that huge amount of the heat released into the liquid, it is only the matter of time when boiling crisis will begin.That's the reason why we have to use an automatic control system with time span 2ms and deal with the unsteady conditions for the extremely high heat fluxes. Of course, the thickness of the heater is 1 mm due to the burnout problem. The whole experimental setup has the electric power above 1.5 MW. By manipulating with the heater properties (thickness, material) and level of the power pulse we can reach most the scenarios of the real applications.

Анатолий,

Мы работаем в стационаре и легко отводим 1 кВт с площади 1 см2. Я думаю отведем и 2-3 кВт, но пока не можем такие потоки сгенерировать.

Нестационар это немного другая задача, но можно попробовать.

Какая у вас площадь нагрева?

Чем у вас ограничивается максимальная продолжительность импульса? Если мы обеспечим огромные коэффициенты теплоотдачи к теплоносителю так, что температура стенки не будет превышать 130 град С, то можно организовать продолжительный нагрев плотностью 2-3 кВт/см2?

Please check this methodology in this research;

Article Heat transfer characteristics of some oils used for engine cooling

Many thanks!

Dear Dmitry,

I was able to talk to Mr. Kabov at the 10 ECI conference in Nagasaki and you are doing really great research!

Monochrystalline C is not that expensive as discussed. Please contact diamond-materials http://www.diamond-materials.com/EN/index.htm. I do not think 10x10x0.5mm will exceed 2000€. ~1500W/(m*K)@100°C

With a transparent metal heater you will get a far better picture of the evaporating microdroplets with backlight illumination!

Good luck :)

Dear Marek,

Thanks for this useful info!

If the elecric heater is burnt out at the heatflux of around 1kw/cm2 how researcher confirmed that the cooling system is capable to remove more than 1 kw of power from the area of 1 cm2 ?

However, laser beams can provide higher heat flux than electric heater but quantity may not be defined...

Dear Mukul Ray, this is our expectation, based on the experimental trend we obtained.

Maybe an oxyacetylene flame from the back side (1700 W/cm^2 reported here from 10 mm away)Article Heat flux mapping of oxyacetylene flames and their use to ch...

Dear Alexander,

Many thanks for the ref.

As I worked on high power LED modules, the most powerful one had a heat flux of about 400 W/cm² and there was no problem with the substrate or the conductors on the substrate. The LED-junctions of the LED chips attached to the substrate were kept below 125°C.

The substrate was 0.63 mm thick AlN ceramics with silver thick film conductors. I would expect, that it should be possible to make a heater with these components for at least the lower end of the desired heat flux.

A crucial point will be the thermal connection between heater and cooler. I used a room temperature liquid gallium alloy as TIM. To use that, the copper surface (and all other metal surfaces) must be protected by a layer of TiN (deposited by sputtering) for example, as the gallium alloy corrodes almost every metal and solidifies. Small amounts of this liquid metal as TIM can be found at various PC tuning shops.

Dear Marc Schneider,

Many thanks for your advice. Could you please provide any refs?

We have generated heat fluxes of a little over 6kW/cm^2 with sputtered platinum heaters which double as RTDs for microfluidic cooling work. Here are a couple of examples which used this type of design.

http://www.bakirlab.gatech.edu/papers/Combined%20finned%20microgap%20with%20dedicated%20extreme-microgap%20hotspot%20flow%20for%20high%20performance%20thermal%20management.pdf

Article Integrated Circuit Cooling Using Heterogeneous Micropin-Fin ...

Dear Thomas,

Many thanks for this interesting info.

I'm none too sure what you are trying to achieve or why you need to achieve it, but you may find the following comment of interest.

In the USA, in the late 1960's or 1970's, NASA I believe set up a supersonic wind tunnel and using air, generated a stationary shock disk across which there was a significant temperature rise. Into this disk, they injected a stream of hydrogen which ignited within the shock liberating a large amount of energy and close to the shock a very high heat flux. The disc was about 1.0 cm diameter. The reports are in open literature, I believe.

Stephen

Great research @ GeorgiaTech! :)

good morning Marek, Dmitri and others,

usually I become slightly irritated with some of the monosyllabic responses received here but my own last response was not very illuminating. My excuse is that my reference works are temporarily unavailable.

I like the imogi but why not a smiley face? and which Tech.?

Regards, Stephen.

Dear Stephen,

Many thanks for your answer. Unfortunatelly, I couldn't find any info.

Any refs would be highly appreciated.

Dear Marek,

If you mean works of Prof. Fedorov @ GeorgiaTech, they use a tiny heating area, so the total power is not very high.

We need total power of 3-5 kW.