PMI results 316L welds

[jrupert24]

Having a problem with low nickel in the welds scanning with a XRF spectrometer. Checked pipe and it is around 10.5-11.0 and undiluted filler metal is 11.5-12 which matches the certs that were provided (we require actual chemistries). When we check the welded joints we are getting 8.75 to 10.5 with the majority of readings around 9.5. The pipe sizes are 3/4"-3" sch 40. I was expecting the values to be at least around 11 for the weld metal since there is dilution. I checked some 304L pipe and everything scanned OK. I even had one welded up in front of me and got the same results after welding. Scanner is calibrated, new battery, and we checked it on some other known parts. All reading look normal which makes me believe the scanner is working correctly. Any suggestions??

 

[metengr]

This is not uncommon, I have seen similar variation in field testing.

 

[ApC2Kp]

jrupert24,

The seeming loss of nickel would be a puzzle. A comparison of melting / boiling points might shed some understanding.
Nickel Ni 1455 deg C melting, 2730 deg C boiling,
Iron Fe 1539 deg C melting, 2740 deg C boiling,
Chromium Cr 1890 deg C melting, 2500 deg C boiling,
Molybdenum Mo 2625 deg C melting, 4800 deg C boiling.
It might explain nickel losses by spattering, or its separation into concentrated islands or strata in the weld puddle. Nickel having slightly higher density than Cr-Fe might also tend to settle with Mo to bottom of the weld puddle?

 

[stanweld]

Agree with metengr; it is not an uncommon occurance. Nickel content accuracy was especially problematic during the early days of XRF analysis using portable equipment (Texas Nuclear). Whereas you appear to be obtaining resonably accurate analysis on the fully austenitic base metals, possibly the mixed ferritic-austenitic microstructure of the deposited weld metal may be influencing the results.

 

[jrupert24]

This seems a logical explanation as it doesn't seem realistic to "lose" such a high percentage of nickel during welding. I appreciate the comments so far.

 

[EdStainless]

How are you preparing the surface? If you are shooting as welded then the surface texture may also be having an impact.

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

 

[jrupert24]

We tried shooting as-welded, wire brushed, ground, and polished. These were the techniques they said worked in the past and would get the values to fall just within spec to pass. We tried these, but does not seem to help much or at least not enough to get us in the 10 range where I could justify the low weld Ni% due to dilution.

 

[unclesyd]

Do you have a traceable standard to test?

 

[Stl63]

What's the weld filler metal type being used? ER308L?

"Saving the World One Beam at a Time"

 

[jrupert24]

Posting a follow up. I welded up another sample of 316L pipe welded with 316L weld wire (two welds). I scanned it with our XRF scanner and got about 10.5 for each. Set it to a test lab to have it scanned with their XRF scanner and it came out a 9.5 for each. I then had them perform OES and the results were 11.6 and 12.0. I think there are several factors that may be effecting the results. (1) the weld is smaller than the XRF scanner window (2) bead geometry (3) bead contour/preparation (3) microstructure???? (4) scanner algorithm for calculating Ni based on feedback. Client accepted all of the results based on the mock up chemistry.

 

[BSAM81]

Portable XRF analysers are indicative instruments only and have some limitations on accuracy compared with laboratory analysis. Results obtained will be affected by component dimensions, shape and surface finish. Results reported should be taken as a guide to material type only and shall not be relied upon as a definitive analysis. Contamination of surface can occur with dissimilar metal wire brushes used to clean test area, CS grinding disks used on SS test material, insufficient surface preparation-including removal of decarburised layer formed during welding. Geometry of the test area will also give varing results within the test. Best way to test weld material chemistry with small bore piping, is to perform a pad weld onto the parent material of around 50x50mm in size and prepare test surface on the pad to a suitable standard using correct tecniques for that material. Then if the nickle is still low in your results you can rule out test area geometry and incorrect surface preparation as resons for these lower than expected readings.

 

[EdStainless]

You don't loose Ni in the welds as much as the surface tends to become Fe rich. This is why surface prep is required.

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