If you don't know what an ELSD is, you don't need this article.
What I'm posting now is apparently the text from a PowerPoint presentation. I will attempt, later to post that presentation, but I that won't happen soon.
As far as I can tell, the governing equation was first described in references 3, 11 & 12 of the following article. (I have not reviewed these references myself.)
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwiv8LzktbrtAhXkGVkFHQ_3D0YQFjAEegQIAhAC&url=http%3A%2F%2Ffiles.pharmtech.com%2Falfresco_images%2Fpharma%2F2014%2F08%2F22%2F47cf43ca-509f-4caa-bc4e-22e9327cb692%2Farticle-47789.pdf&usg=AOvVaw01Js6v0rxnSvUnBaWPf0LM
_______________________________________________________________________________________
The ELSD
· The evaporative light-scattering detector (“ELSD” and) its operation
· The low-temperature accessory (“LTA”)
Quantification
The ELSD responds to the mass ("M") of an analyte that enters the detector, but gives non-linear response to this mass. The governing equation for the response ("R") of an ELSD is as follows:
R = a ·Mb
There are many examples in the literature of analysts limiting the region over which they calibrate, or using polynomial and other "fits" of the R vs M standardization data. These can work under the specified conditions, but it is much more useful to "linearize" the relationshop of R to M to enable better accuracy and some degree of accuracy when the mass of analyte falls outside of the range of the standards (i.e., extrapolation).
The Calibration/Standardization Procedure
· Linearize the governing equation
R = a ·Mb
log(R) = log(a) + b ·log(M)
· "Calibrate" (to find b) using a geometric series of
injection volumes of a single standard solution.
· "Calibrate" using a geometric serial dilution of a
standard solution
· Plot log(R) versus log(M). The slope of the line is b.
Specific response ("SR") and its use in quantification
· For linear detectors, one can determine a meaningful specific response ("SR") of the analyte:
SR = R ¸ M
· One can then base one
Pspl = (SRspl / SRstd) · Pstd
This is not useful for ELSD due to the non-linearity.
The Calibration/Standardization Procedure, continued
· Linearized specific response and its use in quantification
R = a · Mb
R ¸ Mb = a
LSR = R1/b ¸ M = a1/b
· The LSR is used to standardize the sample against the standards
Pspl = (LSRspl ¸ LSRstd) · Pstd
The Calibration/Standardization Procedure, Summarized
· "Calibrate" using a geometric series of injection volumes of a single standard solution, or using a geometric-serial dilution of a standard solution and a single injection volume.
· Plot log(R) versus log(M). The slope of the line is b.
log(R) = log(a) + b ·log(M)
· Calculate the LSR of samples and standards, and "standardize" the samples against the standards.
LSR = R1/b ¸ M = a1/b
Pspl = (LSRspl ¸ LSRstd) · Pstd
The Process Problem
· The process involved the use of PPG 2000 in one step, followed by a methanol extraction, and by other chemical and clean-up steps.
· The question arose whether PPG would carry through the process or be eliminated by the existing clean-up steps
· HPLC with ELSD was selected as the means to quantify the PPG in the process streams.
The Chromatography
· A 15-cm C18 column was used with an acetonitrile-water gradient.
· The conditions used were as follows
ELSD without LTA ELSD with LTA
Flow Rate: 1 mL/min Flow Rate: 3 mL/min
Time Acetonitrile Time Acetonitrile
0 min 10% 0 min 10%
2 min " 0.7 min "
7 min 40% 2.3 min 40%
63 min 96% 21 min 96%
100 min " 33.3 min "
ELSD 95°C, 3 L/min N2 ELSD 40°C, 2 L/min N2
The Chromatograms
· Show figures 2 and 10 from the method.
Interferences in the Chromatograms
· Interference seen in process stream samples
· Show figure 4 from the method.
· Because of this interference, it was impossible to integrate all the peaks together and thereby quantify the samples versus a standar PPG-2000.
Quantification versus the Same Lot of PPG
· PPG-2000 from the same lot as used in the process was obtained for use as a standard.
· Show figure 6 from the method.
· Since the fraction of the PPG represented by a given peak is the same in both samples and standard, any single peak could be used for the quantitation.
· In fact, a few peaks were used, and the results averaged.
"Calibration" of PPG-2000
· Show figure 7 from the method.
"Calibration" of PPG-2000 (continued)
· If we choose one injection volume (or serially-diluted standard solution) as the reference, and call its R and M, R° and M°, then the governing equation
R = a ·Mb
can be rewritten as
R ¸ R° = (M ¸ M°)b
or
log(R ¸ R°) = b·log(M ¸ M°)
· Plotting log(R ¸ R°) versus log(M ¸ M°) will give a single line for all PPG oligomers.
· This increases the accuracy of our estimation of b.
The Calibration/Standardization Procedure for PPG-2000
· "Calibrate" using a geometric series of injection volumes of a single standard solution, or using a geometric-serial dilution of a standard solution and a single injection volume.
· Plot log(R ¸ R°) versus log(M ¸ M°) for all oligomers for which there are no interferences The slope of the line is b.
log(R) = log(a) + b ·log(M)
· Calculate the LSR of samples and standards for each of the largest few peaks, and "standardize" the samples against the standards.
LSR = R1/b ¸ M = a1/b
Pspl = (LSRspl ¸ LSRstd) · Pstd
Use of the Method for Process Streams Potentially Containing PPG-2000
· For several of the streams, the method worked fine.
· For some streams, the "PPG" oligomer peaks seemed to be shifted from their proper retention times.
· The process chemist recognized the problem: In these streams, the PPG had been derivatized along with the precursor to the API.
· When the derivatization was reversed by hydrolysis, these streams also showed the proper pattern of PPG oliomers.
· Conclusion: PPG carried into the process, but was removed by the existing purifications scheme.
Hindsight
· This work was exploratory and taught us much about the use of the ELSD.
· These results demonstrated that a general assumption of linearity is not valid for the ELSD.
· In a well-behaved system, only two standard injections should be needed to establish the value of b.
· Quantitation by calibration/standardization was complicated by the fact that XChrom software could not perform a log transformation of the data.
· It may, in fact, be better to use the same serial dilutions or variable-volume injections as in this method, then do point-to-point calibration (assuming linearity between points).
Quantification of PPG-2000 without Identical Lot as Standard
· Sucessfully separate all PPG oligomers from all other stream components.
· Modify chromatography to elute all PPG oligomers as a single peak.
· Take advantage of ±Gaussian distribution of oligomers in mixture, and estimate total from quantities of a few, compared to a well-characterized standard.