304 and thermal expansion vs. % cold work

With a lot of cold work you can get a change in modulus. Are you sure that this isn't related to residual stress issues?

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

According to my reference material, “Mechanical Metallurgy” by G. Dieter, page 334, and I quote;

"The modulus of elasticity is determined by the binding forces between atoms. Since these forces cannot be changed without changing the basic nature of the material, it follows that the modulus of elasticity is one of the most structure-insensitive of the mechanical properties. It is only slightly affected by alloying additions, heat treatment, or cold work."

You did not mention how much change in the elastic modulus or expansion you observed, nor how you tested the diaphragm to obtain these values. I would imagine there could be such a small change that it is within measurement error.

I would certainly think that your measurement of strain in the diaphragm would be more affected by cold work (or residual stress) as edstainless mentioned.

When you cold work 304 you transform some percentage of austenite to martensite which, as BCC (HCP) has a lower coefficient of thermal expansion than austenite.
Modulus is determined by composition. Thermal expansion, and thermal conductivity, is not.

Michael McGuire

Thanks for the response. I think my original posting lacked some key details.

Both the sensor body and the diaphragm are made from 304. When assembled the diaphragm and body can be thought of like a snare drum, which is radially tensioned to “tune” for the desired sound(or motion and pressure range). The diaphragm's movement under pressure load is very accurately measured with our custom electronics. Typical motion is 0.002" at room temperature.

Diaphragms that have been annealed before assembly results in motion at elevated temps (150oF) within 1% of room temp motion. If the diaphragm is not heat treated before assembly the motion is less. The reduced motion can be 10%-50% less, and is proportional to the amount of cold work., which has varied from 50%-75%.

Mcguire’s response about lower coef of thermal expansion has been our assumption, but I’ve not seen any thing that directly related %cold and thermal coef. Does anyone know of a reference that would help clarify the issue further?

The goal is to order material that does not need to be heat-treated. Since the material is very expensive and requires large batches to be ordered, we prefer not to continue our trial and error method and would like to nail this down before our next order.

You can go to Allegheny Ludlum's web site and look on one of their "Blue Sheets". E.g.You will have a large intrinsic variability since the amount of strain-formed martensite varies with individual heat chemistry and temperature at which cold work is performed. You will never have the same diaphram twice!
If you avoid the strain-induced martensite by using 305, you will have much better reproducibility.
Even with this you will have better reproducibility of mechanical properties if you stress relieve the material because of the Bauschinger effect lowering the yield strength. Maybe it can be supplied stretcher leveled which will help in this regard.

Michael McGuire

The tensile modulus measured on real samples is a function of cold work levels for heavily cold worked material. These curves aren't sraight, they don't pass through zero, they aren't reversable, and little things like Poisson's ratio get in the way.

That aside, I agree that varying degrees of transformation are probably to blame. Even if there is no measurable level of transformation it is likely that there is some localized trans.
305 or 310 would be good choices. You do need to know how the material was rolled and flattened. I don't know why you would care if the material was roller leveled vs strech leveled, but if you compare material processed the two different ways you will find that they appear to have different yield strengths.

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

As I've posted before I manufactured rupture disks for in house use from Austenitic SS. The overcome the variability in the bursting pressure the one thing that had the most impact on the final bursting strength was that I partially annealed, a few minutes at 1500?F to 1550?F, the 7/8" dia. blanks 0.008" tk. After prebulging the disks I repeated the heat treatment which dropped the variability at bursting to essentially zero (±25 #) at my desired rupture pressure of 6750 psig.

Another little tidbit is the variability in thickness of the shim stock that I used or in fact any type stock that I looked at. You could see a variability in your thickness depending on what precision you purchase. I took the route of buying, say ten rolls of the same lot of shim stock and sorting the material myself and making small adjustments in the prebulge pressure, up or down, depending on the actual thickness. You can rest assured that the manufacturer will hang on the outside of the tolerance range . Made possible by better measuring equipment.

There was a measurable variance of the apparent grain size in the material as received. This was observed while looking for grain growth from my heat treatment. At the same time I observed the possibility of different amounts of cold work, evidenced by twining, involved in the manufacture of the material. I saw this variability in microstructure on all of the stock that I tried.

From my experience in dealing with thin cold worked SS I don’t think you will get your required precision without at least the above heat treatment or a full anneal.

There was a lot of good information that I could have gleaned from my work but as a side job I couldn’t justify the time and effort. As these disk were so much better than the existing disks that once I filled everyone’s private stash and we only used 6 on each of 5 units I worked myself out of job real quick.