EPA 545 Anatoxin-a by LCMSMS, dropping response on I.S.

Discussions about GC-MS, LC-MS, LC-FTIR, and other "coupled" analytical techniques.

8 posts Page 1 of 1
We am currently running EPA545 Anatoxin-a/Cylindrospermopsin. We are using an API-3200 MSMS with Agilent 1200 binary pump and autosampler.

The transitions used are Cylindrospermopsin 416/194, Anatoxin-a 166/149, Uracil-d4 115/98, L-Phenylalanine-d5 171/125. Uracil-d4 is the internal standard for Chlindrospermopsin and L-Phenylalanine-d5 is the internal standard for Anatoxin-a. ESI positive mode used.

The column is the Waters X-Select HSS T3 3.5um 2.1x150mm at 40C and mobile phase is step gradient of A=5.8ml Acetic Acid/L H20 B=Methanol that steps from 100%A to 75%A at 0.8 minutes with flow of 0.375ml/min.(drops back to original percentages after elution of last peak at ~5 minutes and allow to equilibrate until 15 minutes)

The problem I am having is that l-Phenylalanine-d5 will begin with a huge peak, then after about 4 injections drop to about 1/2 the response, then continues to drop over a long run of about 35 injections. This last run l-Phenylalanine-d5 started with 160251 area counts for first blank(after 5 warmup injections) then fell to 121593 area counts at first calibration standard and on down to 55155 area counts at end of run. The run was queued to run again immediately after last injection where the area counts ranged from 54158 down to 39241 over the next 35 injections. The second run passes QC limits per the method since it falls less than 50% of the average of the calibration curve.

In the first set, the Uracil-d4 ranges from 139487 to 127594 on first run and 163564 to 147904 on second run.

The Cylindrospermopsin has recoveries of near 100% for all three calibration checks ran immediately after the calibration, mid run and end of run. The Anatoxin-a however has calibration check recoveries of about 120% on first check, 165% at mid check and 185% on last check. It seems that only the L-Phenylalanine-d5 suffers from the drop in sensitivity over time, which causes the increase in recovery of the Anatoxin-a. If I run the analysis the next day with the same mobile phase and samples the L-Phenylalanine-d5 will be even more steady, but slightly less response.

If this was charging of the quads, I would think that the Anatoxin-a would suffer nearly as much as the L-Phenylalanine-d5. So I tried adding a stirbar to the mobile phase and stirring during the run and it seems to drop a little less, plus letting the mobile phase sit over night also seems to stabilize the sensitivity, which this set of numbers comes from, when starting from fresh mobile phase made just before analysis the drop on the first run will be even greater.

Has anyone seen any similar problems with an analyte like this?

I did try increasing the acetic acid in the mobile phase but it lowers sensitivity for everything so maybe I need to try less. The peaks are needle sharp except for the Uracil which spreads a little at the bottom both front and back which I think causes the small amount of response shift it has. Overall the method works great, except for the rapid loss of response for the one internal standard. I always get a reportable run, but always as the second time through which wastes a lot of analysis time.

Thanks for any ideas anyone might have.
The past is there to guide us into the future, not to dwell in.
You may have to go to the labeled target analytes. It solved my Uracil recovery problem in surface water. They are pricey but it helps solve some of these pesky problems.
Steve Reimer wrote:
You may have to go to the labeled target analytes. It solved my Uracil recovery problem in surface water. They are pricey but it helps solve some of these pesky problems.


Not sure that I can switch since these are what we were certified for UCMR4 with. Everyone else reports problems with Uracil, but that is my best performing analyte which is strange. A year ago the L-Phenylalanine-d5 was stable enough to pass in a single run, not dropping more than maybe 20%, now it won't behave at all.
The past is there to guide us into the future, not to dwell in.
An interesting update on this one.

Last week I was looking through the method and I noticed that the L-Phenylalanine-d5 transition was using a cell entrance potential of 2.5v, while two other transitions use 3.5v, another uses 9.0v and another 10.0v. I bumped the 2.5v up to 3.5v and was able to report from the first run through the batch with only about a 30% drop in sensitivity, where before it was over a 60% drop.

Today I bumped all of the 3.5v settings up to 4.0v and right now it has run through the calibration with a smaller sensitivity drop than last week. I will see how it does through the whole run, but hoping that will fix it.

The setting was determined by letting the instrument automatically optimize on each analyte while being infused. The results were stable until just before we stopped getting samples last fall ( the program does not take samples during December through February). I am wondering if maybe the voltage controller is having difficulty maintain the lower voltages but can hold stead at higher levels.
The past is there to guide us into the future, not to dwell in.
Thank you for posting your progress! So few people do so, and yet it vastly increases the value of this (already valuable) message board. And congratulations on spotting that change!
lmh wrote:
Thank you for posting your progress! So few people do so, and yet it vastly increases the value of this (already valuable) message board. And congratulations on spotting that change!


I agree, posting here isn't just to solve our problems but to help others who may have the same problems later on.

An update

I ran about 30 injections on this last run and the internal standard area counts remained steady throughout the entire run, instead of dropping 50-60% within the first 8-10 injections. Actually there was a slight increase in sensitivity at the end of the run but I believe that is due to the dropping building temperature during the night. Total deviation was probably less than 10% above and below the average, which is more like it was a year ago.

I can only think that for some reason the instrument is having difficulty maintaining the lower voltage setting at the collision cell entrance for some reason. Every other change I made had little to no effect at all.
The past is there to guide us into the future, not to dwell in.
This perhaps isn't relevant in your scenario, but something I've noticed about automatic parameter setting (in my case with a Waters Xevo TQS): In the "good old days" a lot of parameters seem to have been quite critical, with sensitivity rising to a peak and falling again fairly sharply. Manufacturers have worked hard to make their systems more robust, so some of the parameters that used to do this, now rise to a plateau and hang around with little change for a large range of voltages. The cone voltage on the Xevo is an example. That's fine, but it means the actual voltage selected by autotune software is largely random. It's at the biggest noise-spike on a plateau. If you're unlucky, it's right up one end, which I presume means if anything happens to the instrument that shifts the curve slightly (perhaps some dirt accumulating on a lens over a run??) then you're plunging down the side of the flat-topped hill, whereas a human would instinctively ignore the noise and choose somewhere in the middle of the flat hill.
lmh wrote:
This perhaps isn't relevant in your scenario, but something I've noticed about automatic parameter setting (in my case with a Waters Xevo TQS): In the "good old days" a lot of parameters seem to have been quite critical, with sensitivity rising to a peak and falling again fairly sharply. Manufacturers have worked hard to make their systems more robust, so some of the parameters that used to do this, now rise to a plateau and hang around with little change for a large range of voltages. The cone voltage on the Xevo is an example. That's fine, but it means the actual voltage selected by autotune software is largely random. It's at the biggest noise-spike on a plateau. If you're unlucky, it's right up one end, which I presume means if anything happens to the instrument that shifts the curve slightly (perhaps some dirt accumulating on a lens over a run??) then you're plunging down the side of the flat-topped hill, whereas a human would instinctively ignore the noise and choose somewhere in the middle of the flat hill.


I normally review what the instrument does automatically so as to avoid such problems. It worked for a while but now seems to have drifted, but for a 2006 model instrument is still runs quite well. It probably is due to have the rails pulled and everything cleaned again, has been several years.
The past is there to guide us into the future, not to dwell in.
8 posts Page 1 of 1

Who is online

In total there are 6 users online :: 0 registered, 0 hidden and 6 guests (based on users active over the past 5 minutes)
Most users ever online was 599 on Tue Sep 18, 2018 9:27 am

Users browsing this forum: No registered users and 6 guests

Latest Blog Posts from Separation Science

Separation Science offers free learning from the experts covering methods, applications, webinars, eSeminars, videos, tutorials for users of liquid chromatography, gas chromatography, mass spectrometry, sample preparation and related analytical techniques.

Subscribe to our eNewsletter with daily, weekly or monthly updates: Food, Environmental, (Bio)Pharmaceutical, Bioclinical, Liquid Chromatography, Gas Chromatography and Mass Spectrometry.

Liquid Chromatography

Gas Chromatography

Mass Spectrometry