In performing simultaneous k-ratio measurements on multiple WDS and EDS spectrometers, we noticed some issues comparing the WDS k-ratios with the EDS k-ratios in several materials. You may have seen some of these issues in these posts from some time ago:
https://probesoftware.com/smf/index.php?topic=1569.msg12228#msg12228This is after Thermo fixed the Portal interface bugs relating to the definition of the detector type, and other detector parameters in v. 2.10 of Pathfinder:
' Only available in PathFinder 2.10 or higher (ThermoNSSVersionNumber >= 6)
Private Declare Function TEVB_SpectrumGetDetectorSubtype Lib "TEPortal.dll" () As Long
Private Declare Sub TEVB_SpectrumSetDetectorSubtype Lib "TEPortal.dll" (ByVal lDetectorSubtype As Long)
Private Declare Function TEVB_SpectrumGetTimeConstant_ns Lib "TEPortal.dll" () As Long
Private Declare Sub TEVB_SpectrumSetTimeConstant_ns Lib "TEPortal.dll" (ByVal lTimeConstant_ns As Long)
Private Declare Function TEVB_SpectrumGetZeroWidth_eV Lib "TEPortal.dll" () As Double
Private Declare Sub TEVB_SpectrumSetZeroWidth_eV Lib "TEPortal.dll" (ByVal dZeroWidth_eV As Double)
More recently we saw similar discrepencies for the P Ka in GaP and InP k-ratios, so I contacted Nicholas Ritchie at NIST and asked if he would look into this and he responded "yes", so I sent him the spectra for the Gap and InP standards. Plotting up the WDS k-ratios along with the results from Thermo and Nicholas' DTSA2 we obtained this plot:
So the net intensities from DTSA agree with the WDS k-ratios, but the net intensities from Thermo Pathfinder do not. Meanwhile Gareth Seward noticed that while Pathfinder gave incorrect results through the Thermo Portal interface, it gave the correct results when processing the spectra through the Pathfinder application!
Gareth then proceeded to track down the problem (we think) and that is that when the get net intensities function is called through the Thermo Portal interface it resets the sum peak and escape peak correction *off*, even if they are set in the Pathfinder application. Here is what Gareth wrote:
The variables that sets Esc and Sum peak processing on (1) and off (0) for a given spectrum are stored with individual .EMSA files as they are created/modified
##ESCPKSREM : 1
##SUMPKSREM : 1
When an external spectrum is sent through the API using TEVB_PutSpectrum the data is written to the 'temporary' file:
C:\ProgramData\Thermo Scientific\NSS\NSS Libraries \Results.emsa
This file contains the info that was passed through the API prior to TEVB_PutSpectrum (e.g. ##TIMECNSTNT 
, but it is also populated by many other parameter values that are seemingly hardcoded/default within Pathfinder and are not appropriate to the hardware on which the spectrum was collected, or the current hardware. These parameters include for example ##CONTMATL : which, by inspection, seems to be pertinent to esc/sum peak processing. Others include:
##CRYSTMATL ##CRYSTAREA ##CRYSTTHICK ##WINTYPE ##WINDCOAT ##WINDTHICK : 0.040
Basically, when the get net intensities function is called through the Thermo Portal interface, Pathfinder writes the passed spectrum to a temporary EMSA file, but *without* many critical parameters (recall the missing detector type, etc., parameters that were added to the Portal interface last year by Thermo).
But the really weird thing is that Nicholas said he did not perform an escape peak or sum peak correction in DTSA, so perhaps there are other parameters that are not being properly set through the Thermo Portal interface as Gareth mentions in the quoted text above.
Anyway, I've discussed this with both Steve Seddio and David Rohde and they have said there is nothing they can do about this even through they are currently still selling the Pathfinder EDS software!
So I would humbly ask any of you who either have a Thermo Pathfinder system (on your EPMA instrument), or are considering a Thermo Pathfinder EDS system (for your SEM or other instrument), to please contact Thermo and ask them to fix these missing parameters when making calls through the Thermo Portal interface.
Hey, if you don't ask, you already know the answer!
