Resolving 304 SS chemical analysis disputes

[KM0818]

Hi, all. I would appreciate input on this.

We purchase stainless steel components regularly and have an incoming inspection procedure that combines sampling with PMI method and outside lab tests using OES.

Just recently, for the first time in a few years, we began repeatedly receiving batches of products that are low in Cr and oddly high in Ni (only around 9%-9.5%, but it is unusual for the bar products). The supplier disputes, or at least questions, our results and we have been going back and forth with additional tests, statistical analyses of results, etc. By controlling the test specimens better (trying to eliminate piece-to-piece and within piece variation) we have reduced the amount of discrepancy, but not enough.

For the most part, we see the OES results falling within the stated (and verified) error range of the PMI results. But the supplier results, including their additional tests by outside labs, sometimes still don't match ours.

This supplier is in China and we are seeing the differences in cast products and in bar products from a supposedly very reputable steel bar supplier. We did find some material control problems at the factory, but even when we carefully control the exact pieces tested by our labs and theirs (ship and test the same pieces) we have some discrepancies.

To bolster their argument, the supplier points out that the bar supplier would not use such high Ni levels as seen in our tests, but not seen in theirs.

That's the background. Now the latest question, for which I could not get an answer from the testing technician at the lab, but probably can from the metallurgist. Does anyone know the accuracy and precision of the OES method? The lab has PM on the machine twice per year, but the technician could not tell me what accuracy or precision they test to. (Again, I'm pretty sure the metallurgist will be able to tell me, but I haven't asked yet).

Second question: The China supplier has sent samples to multiple labs, including a quasi-government or government lab, and reported results. Their current position, though they are seemingly being cooperative in the investigation, is that the "chemical" methods they are using are recognized as the standard method in China and that the OES method is more for high production and perhaps not as reliable, but is improving in China. I have done some Google searching for comparative accuracy and precision of the methods, but I haven't been able to nail it down.

Can anyone enlighten me a little on the accuracies and perhaps suggestions for the best way to proceed on this? Thanks.

 

[Metaljon]

For a modern OES instrument it would be expected an accuracy of +/-0.5% of the actual analysis. For Cr at 18wt% we would expect all results to be within the levels 17.91 to 18.09 wt%. This assumes that the instrument is properly setup and calibrated and the sample tested being fully homogeneous. For our samples, we always run samples of similar alloy at the same time and then use a standardisation to check for accuracy. You can do the same by purchasing a chemical standard and running at the same time. This will show up errors in analysis and is a good QA check.

In general, I have found the analysis from Chinese companies is not very good. There is also a lot of bogus testing and false certification of products in China. Cast samples will always provide source of increased composition variability.

For bar and flat product we also find problems with Chinese product with carbon contamination rolled into surface, sensitisation of the SS, dimensional mis-tolerance due to heavy surface grinding and use of bog to cover up defects and spray with silver paint.

 

[KM0818]

Thanks, Metaljon

 

[EdStainless]

Teh accuracy will depend on the reference standards being used. Many people are relying on computer models and using fewer real physical standards.
That is the real test, not what they get on your samples but what standards they are using and what values they get when they re-analyze them.

One issue that I have encountered in China is that the various functions within the company are independent. While we would at first think that this is good it actually turns out that each group is only looking out for themselves and no one is taking care of the customer.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[arunmrao]

For a test check,submit a sample for wet analysis and compare with OES results. Have you factored for the permissible deviations between product analysis and bath/melt analysis.

All disputes could be resolved by undertaking wet analysis at both ends for the same sample and then comparing the results. I have gone through this process in the past.

 

[EdStainless]

"WET"? You are as old as I am.
Today you will find OES, XRF, ICP, AA, and a few others, but no one will do direct chemistry any more.

The check limits is a good comment. There can (and usually is) a fair amount of scatter in check chemistry. This is why they are allowed to be outside the original specification ranges.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[KM0818]

Except for the supposed certification from the mill for the bars and the original certifications from the casting supplier, all tests have been done on product.

Maybe they are still doing "wet" analysis in China. Here is a recent explanation I got from the guy who is trying to resolve this on the China end:

"They may Chemical analyse is the most accurate method which it is used for finally arbitrament in official lab , results with Legal effect ; detail is : drill and pick material scrap from sample part and dissolve in chemical reagents and analyse different components (I don't know exact steps) ; this methods need long time and rich skill/experience to operate so is not offten used at workshop or production line"

Is this describing WET analysis? Is it really more accurate than the other methods?

(I did not have the chance to call the metallurgist at the domestic lab today).

Thanks.

 

[arunmrao]

He is describing the wet analysis process. It is definitely accurate,as it is based on pure chemistry,however,the standard solutions and chemicals that are used for the procedure must be carefully selected and must be of analaR grade.

Ed,I am not old,but ancient,I learnt my job in a company doing wet analysis for 6 months every shift.

 

[Guest102023]

PMI is NOT chemical analysis, although many testing companies market it as such. PMI is a cash cow for these companies, and it is often done by technical illiterates who blithely write down all the figures the machine spits out, significant or not (although I have seen licensed engineers routinely sign off on the reports).

It is for 'alloy identification' and sorting only. It has also become a substitute for meaningful quality control, speaking to the shortage of competence in the metal trading business as well as a lack of trust from Owners. On piping projects, for instance, the same component may be PMI tested three times in three weeks before going into service.

 

[EdStainless]

Dissolving cuttings in acid is the precursor for half a dozen different methods.
They told you the key point, that they rely on experience.

Well done XRF or ICP will be as good as any wet methods. In fact ICP has become the basis of most all modern high resolution analysis.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube