Problem with HP 1050 MWD at 198nm

[jtmarten]

All,
I've done some analysis of plating bath organic additives on my manual HPLC system at 198nm with no problems (TSP 3500 pump, ABI783A detector, 4.6x150 C18 column, 20uL sample volume, 1ml/min flow rate).

I set up a sequence on my 1050 system, but the signal at 198nm is bouncing significantly just flowing the mobile phase. I ran many bath samples in the past on the 1050 using 258nm when measuring a different additive, so I added that as a reference signal. The 258nm signal is a flat zero while the 198 bounces like crazy. Mobile phase for both systems is 5% acetic.
The chem supplier said an alternate wavelength is 269nm, which I ran for standard 1 and had a flat zero signal and a beautiful peak at 6.5min, so I will likely use 269nm.

But, just wondering why the signal would bounce so significantly at 198nm. I know it's close to the limit, but should still be accurate at that wavelength. Internal diagnostics report no issues.

[vmu]

The 258nm signal is a flat zero while the 198 bounces like crazy. Mobile phase for both systems is 5% acetic.

Acetic acid absorbs UV light at 198 nm. This leads to the high (or even overloaded at 5 % AcOH) background signal and the high signal noise. WL 198 nm is absolutely unsuitable for your mobile phase. Did you see any analyte peaks at 198 nm with this MP on the first instrument or did you see just a straight baseline without any peaks and noise?

[jtmarten]

Thank you for the info!

The MP signal on the manual system was noisier at 198nm than I usually saw at 258nm, looked like shag carpet once I shot a sample (baths have other organics and likely a significant conc of breakdown contaminants). I sent the remaining raw sample solutions to the mfg for analysis; my results differed by 0.01-0.06%.

Acetonitrile looks to be a better MP for use at 198nm; minimal interference assuming a high enough signal from the sample analyte.

[Consumer Products Guy]

Hmmm - first thing I've always done when I had a noisy baseline was shut off the pump, to see if the detector noise decreased.

[DR]

Agreeing w/ VMU - 198nm is below the UV cutoffs assigned to many mobile phase modifiers and some solvents.
In fact, air can contribute to absorptive noise at that wavelength which is why many UV detectors have a little barbed nipple somewhere on their shell - it's for nitrogen sparging of the optical bench.
In theory, "everything" absorbs at lower wavelengths, which is great for detection of unknowns in stability indicating methods, but 198nm is just too low for most practical applications.

[Multidimensional]

This is a training issue, not an instrument problem. As already noted, using an HPLC method at UV = 198 nm (198 nm has little value as no selectivity can be obtained at such a low UV wavelength where everything absorbs) coupled to a mobile phase containing acetic acid (absorbs strongly at low UV wavelengths) is by itself, invalid and does not confirm to good chromatography practices. The idea that the use of the HP 1050 system may be to blame for any instability or problems is irrational as the instrument is not the issue. Setting up and using the system with a wavelength of 198nm is. Additionally, you used the word "Reference Wavelength" in your post. *I hope that you are not turning on the MWD's "[b]Reference Wavelength" feature as that will subtract all data collected at the "Reference"value from your original signal (destroys the original data forever!).[/b] Never turn on the "Reference Wavelength" and please get some formal training before using any of the advanced features.

To develop high quality HPLC methods using the instruments you have available, ask your employer to higher a professional chromatographer. Please contact someone with demonstrated industrial experience in developing HPLC methods, on the systems you use, so a valid method can be developed that is selective for your analytes of interest and follows good chromatography guidelines.

[Multidimensional]

REFERENCE WAVELENGTHS (as used in HPLC UV/VIS): https://hplctips.blogspot.com/2011/03/r ... -hplc.html

[Multidimensional]

Proper Wavelength Selection for HPLC Method Development (or Purity Determination); https://hplctips.blogspot.com/2013/08/w ... ethod.html

[jtmarten]

This is a training issue, not an instrument problem. As already noted, using an HPLC method at UV = 198 nm (198 nm has little value as no selectivity can be obtained at such a low UV wavelength where everything absorbs) coupled to a mobile phase containing acetic acid (absorbs strongly at low UV wavelengths) is by itself, invalid and does not confirm to good chromatography practices. The idea that the use of the HP 1050 system may be to blame for any instability or problems is irrational as the instrument is not the issue. Setting up and using the system with a wavelength of 198nm is. Additionally, you used the word "Reference Wavelength" in your post. *I hope that you are not turning on the MWD's "[b]Reference Wavelength" feature as that will subtract all data collected at the "Reference"value from your original signal (destroys the original data forever!).[/b] Never turn on the "Reference Wavelength" and please get some formal training before using any of the advanced features.

To develop high quality HPLC methods using the instruments you have available, ask your employer to higher a professional chromatographer. Please contact someone with demonstrated industrial experience in developing HPLC methods, on the systems you use, so a valid method can be developed that is selective for your analytes of interest and follows good chromatography guidelines.

Thanks for the great info!

I did not use the term 'reference wavelength', I wrote 'reference signal'; I used an MP signal at 258nm which yields a nice, flat, zero signal as a comparison to the bouncing 198nm signal.
I am 'the employer' in this case. These samples are run on one of my LC systems at home as I do not have an HPLC in my day-job lab; working on that. Current methods are UV/Vis; one method based on a ratio of standard to sample, the other utilizes a 'fudge factor' to account for inaccuracies in the ratio method. Neither are remotely accurate.
Hiring a professional chromatographer is a fine suggestion, but would be a significant waste of salary, and a waste of their developed skillset. Our analyte is a singular organic additive which is measured on a frequent basis. Same method, day in, day out. Additives are replaced with a new product very infrequently due to the broad scope of environmental testing required to validate and approve the 'new' process.
The 198nm is specified by the mfg as the WOI, I have no control over that. The alternate 269nm wavelength provides very nice chromatograms, but calculated concentrations are 15-20% lower than reported at 198nm. I'll run both wavelengths and track the concentration deltas, if consistent enough I can then run at 269nm and apply a calculated factor for the final concentration, either manually or built into the method parameters.

[TylerSmith123]

Hi JT!

I just read over your last comment and wanted to add this to hopefully help you out! If you're calculating the concentrations based off your initial method they will be wildly off just because your compound likely doesn't absorb light at the same intensity at 269 nm than it does at 198. In fact, you would need to run a few standard samples and make an individual calibration curve for that wavelength. However, it should be extremely simple to do overall. I assume that your analyte absorbs just slightly less at this wavelength than your original wavelength however, if you calculated your original calibration curve with this bumpy baseline issue (really just a normal thing in chromatography) you will be way off. In fact, if you validated anything with the 198 nm it may need to be reassessed for quality purposes.
What does "the mfg as the WOI" mean? Whatever and whoever it is, you need to tell them that their acceptance criteria at 198 nm is flawed! It will make all of your lives easier to simply scan this organic additive with your PDA, pick a peak-maximum, and have your method developed around identifying that (and just to be specific-- cut off any wavelength below 220nm just to be safe). Or you could look up a method for this online, but I am not sure what bath additives are in all honesty. Also, I am confused on if this is a single analyte that you are extracting and then testing, or if you're just shooting the bath additives and looking for a single analyte (you mentioned some degradation products and other organic molecules but did not elaborate on that too much).

Also, to potentially save you some cash, you may want to experiment with a lower percentage of acid in your MPs. I hope your house is very well ventilated to accommodate a few LCs, their solvents and wastes! Stay Safe!

TS

[jtmarten]

Hi JT!


What does "the mfg as the WOI" mean? Whatever and whoever it is, you need to tell them that their acceptance criteria at 198 nm is flawed! It will make all of your lives easier to simply scan this organic additive with your PDA, pick a peak-maximum, and have your method developed around identifying that (and just to be specific-- cut off any wavelength below 220nm just to be safe). Or you could look up a method for this online, but I am not sure what bath additives are in all honesty. Also, I am confused on if this is a single analyte that you are extracting and then testing, or if you're just shooting the bath additives and looking for a single analyte (you mentioned some degradation products and other organic molecules but did not elaborate on that too much).

Also, to potentially save you some cash, you may want to experiment with a lower percentage of acid in your MPs. I hope your house is very well ventilated to accommodate a few LCs, their solvents and wastes! Stay Safe!

TS

Thanks TS!
If I end up needing to use 269nm, I will start looking at developing the factor after 20 runs, then periodically validate.

"WOI" = wavelength of interest

For sure; reagents are certainly not cheap!



I sourced a liter of ACN and prepared a liter of 25% for the MP. The 198 signal was still bouncing (by bouncing I mean completely vertical signal lines). Monitoring 258nm and 269nm - they are both nice, flat signal lines.

I also monitored the 198nm signal on the MWD-B and C; same results.

[jtmarten]

An update:
The 25% ACN performed better than the 5% acetic, but still had significant interference since the ACN has some absorbance at 198nm. The ACN performed fairly well using standards that were undiluted (large peaks but not usable to determine conc, just another test), but samples from the plating baths need to be diluted a min 100:1, as are my normal standards.
I switched to an 80/20 H2O/MeOH with 1.5ml/L H3PO4. I now get a fairly nice baseline which does look like shag carpet, but the absorbance is very low by comparison. I just need to adjust my run time a bit and should have a usable method.

[DR]

If the UV spectra for the additives you're looking for have any distinct local maxima, you might be able to ditch HPLC in favor of a UV/Vis method instead.
Lots less maintenance, less time checking samples, far fewer chemicals to buy... whether this would work is a function of whether the other mixture components absorb too strongly at a given wavelength.

[jtmarten]

I really wish UV/Vis would work. The mfg has two methods for UV/Vis; one is the general ratio method for standard and sample - the problem is the additive breaks down with use into components which absorb at the same wavelength. The other method involves a spreadsheet with a 'fudge factor' to compensate for the ratio method issue. The problem -this is valid only for the bath conditions under which it was created; bath conditions for the organics change dramatically over time based on additions, consumption/breakdown of the additives, and carbon treating to remove the organic contaminants.
It sucks trying to run a method at 198nm, but it is far more accurate than the UV/Vis methods. There is an organic extraction/separation method as well, but it also lacks accuracy.

[jtmarten]

Last night I was able to get a pretty good run with the MeOH. I only had time to get a blank, three standards, and one sample processed. When I input the values for the standards into my spreadsheet, the cal line has an R2 value of 0.9994, and the one sample value looks to be accurate based on values from my manual LC system. The R2 is slightly lower than values I generally net with my manual LC system, which is in the 0.9996-1 range, but it is far better than with either of the previous MPs.