Silicon Carbide Thermal Shock Parameter: 167C Explained

[Cyrus]

I am looking for a material to use, namely Silicon Carbide, and it gives a "Thermal Shock Paramter - 167C"

Obviously thermal shock is going to be failure when there is a rapid temperature change, but what does 167C mean? It's not a rate of cooling, just a temperature. I can't find anything on it in my book or in google. I am guessing it might be a value like a Hardness value set against a standard?

http://www.ceradyne.com/Uploads/Silicon_Carbide_data_sheet_10-03.pdf

I tried to call them but there closed for the rest of the week.

 

[nazzard]

I'm not familiar with the subject but I've found a lecture by Steve Roberts (Oxford) that might be helpful:

http://www-sgrgroup.materials.ox.ac.uk/lectures/ceramics.html

Take a look at "Handout 3: Crack growth and thermal shock", page 3.

Regards,

nazzard

 

[Astronuc]

I am looking for a material to use, namely Silicon Carbide, and it gives a "Thermal Shock Paramter - 167C"

Obviously thermal shock is going to be failure when there is a rapid temperature change, but what does 167C mean? It's not a rate of cooling, just a temperature. I can't find anything on it in my book or in google. I am guessing it might be a value like a Hardness value set against a standard?

http://www.ceradyne.com/Uploads/Silicon_Carbide_data_sheet_10-03.pdf

I tried to call them but there closed for the rest of the week.

Are you looking for impact toughness? or fracture toughness? both of which change with temperature.

Thermal gradients induce localize stress fields, with the hotter portions normally under compression, while the cooler areas are under tension.

 

[Cyrus]

I want to know what does this "thermal shock parameter" mean.

This is for my Heat Transfer Project, I am making a test rig to simulate an Intel Chip to test spray cooling.

If I suddenly turn on the spray cooling I might thermally shock the Silicon Carbide and destory it, so I want to know if I have to slowly turn on the spray cooling to replicate a qasi-equilirbium process to avoid thermal shock.

 

[Astronuc]

Thermal Shock Parameter = [Strength * (1-Poisson’s Ratio)] / (Elastic Modulus * Thermal Expansion Coeff.)
from - http://ceradyne.com/uploads/SiN%20Industrial%20Insert.pdf

Also see page 14 of http://www.ecm.auckland.ac.nz/course/phys130/ceramics.pdf

It is apparently related to fracture toughness, but I could only find an abstract that mentioned that fact. I think the parameter may be discussed in Callister's book on materials, but I'll have to dig it up.

 

[Cyrus]

I saw that at the bottom, but it does not give me a 'feel' for what it really means.

As of now, its just some equation that does not tell me anything in terms of what will happen when I spray cool my simulated chip surface.

Hmm, I just saw this in your second link:

THERMAL SHOCK
Thermal shock refers to the thermal stresses that occur in a component as a result of
exposure to a temperature difference between the surface and interior or between
various regions of the component.
The peak stress typically occurs at the surface during cooling according to the
following equation:
sT = ( E a DT ) / (1-u)
When selecting a ceramic material for an application where thermal shock is expected
to be a problem, calculation of the appropriate thermal shock parameter (R) (or the
maximum allowable temperature difference, DTmax ) for various candidate materials
may be useful.

Crap, that's not what I wanted to hear. I have a temperature distribution of 70C to 903C over a span of 0.5cm. This thing will shatter. I'm going to have to find a new material.

 

[Astronuc]

Is 903°C correct? In a chip? That's awfully hot.

 

[Cyrus]

It's correct. The surface temperature of the chip is at 70C, which is the correct operating temperature.

The temperature distribution results in the bottom surface of the chip to be 903C, but I don't care what the temperatures are at the bottom, that's not what's being simulated. But I do need to watch out for the bottom temperature because of (i) melting or (ii) thermal shock.

Because of (ii), I can't use this material anymore. I need a resistivity of 1-10 ohm-cm. I am looking at matweb for materials right now, but most don't have the right combination of properties.

This project is pushing the limits of current technology. We are putting 2000W through a 1cm^2 chip area. Very very difficult.

You have callisters book too. That book is a good paper weight.

 

[Cyrus]

Ah yes, I played around with the numbers and material selection and I think I got a number that will work. The temp diff is only 40C, and so shock is no longer an issue. But now the material is very VERY thin, on the order of 0.8mm thick. (but still silicon carbide)

Fantabulous.

 

[Gokul43201]

This is crazy...903C!

If nothing else (melting the packaging, breaking stuff from differential thermal expansion, etc.) do you know how pathetically small electron mobilities in Si become at temperatures like that? Your chip will be at least several tens of times slower than it woulb be, close to room temperature.

 

[Cyrus]

Cyrus..."fill in blank" is thy name!

You don't know all the details yet!

Don't worry, the chip is NOT at 903C. The Sillicon Carbide I was modeling the chip as, FOR THE SURFACE ONLY! Was at 70C. Due to internal heat generation and surface convection (h=400,000), the bottom temp would be at 903C. This is PERFECTLY FINE, because I AM NOT modeling the BOTTOM OF THE CHIP!

Anywho, don't sweat it. I will give you all a post with all the details once I am done with the project, as for now, I don't have the time to do all that, sorry.

Just trust me, it DOES Work.

breaking stuff from differential thermal expansion, etc.

Yes, based on that thermal shock parameter, it was a problem, but I have worked a way around it. (I think) I had to make the chip 0.8mm thick, but now the temperature differnce is only 40C. No more shock issues.

 

[Cyrus]

If you want some background on Spray Cooling, here is my Professors website.

http://www.enme.umd.edu/~kimjh/

 

[Cyrus]

Another question, anyone know of a material that I can use as an insulator? It has to be very thin, something I can spray on as a film on the order of nanmeters thick, if possible.

 

[Cyrus]

Crap, anyone know of a thermal paste/grease that is good for temperatures in the range of 800C?