Chromatography Forum

I periodically redo my surrogate recoveries statistics and I always see this pattern....
Computing the statistics for CCV's always gives lower RSD's than those for the samples and method blanks. Method blanks give the highest RSD's. Since most samples tend to be clean; the RSD's computed for water samples tend to equal or be slightly lower than those of the method blanks.

Is this a trap performance related thing that surrogates recover better when there are other analytes on the trap?

Edit:
I'm comparing 307 CCVs, 341 Blanks, and 2570 water samples run over two years time.

The amount of analyte loading on a trap will influence recoveries of each analyte.

I did a study a few years ago trying to solve the problem where 1,4-Dichlorobenzene-d4 would increase in response as the calibration standards increased in concentration. It was always a problem when the curve was in the normal range of 5-200ppb, but I was working on a project that was taking the concentrations down to 0.1-20ppb and when using internal standards at 5ppb the 1,4-DCB would double in response from lowest calibrator to highest. I then ran the same curve with the internal standards at 50ppb and there was almost no change in response over the same calibration levels. If you calculated the total ng of analytes including IS/Surr for both calibrations the total mass of analytes increases almost 100 fold for the lower internal standard curve while only about 20 fold for the higher internal standard curve. I even tried another with the internal standards at 100ppb and got an even better linear response for the whole calibration since the internal standards and surrogates were the majority of the total mass and caused even less change in the total over the curve range.

I think it is a similar effect as to when in semivolatiles pentachlorophenol will be almost nonexistent in the tune shot with 4 analytes at 50ppm, but will give great response in the full CCV with every analyte at 25ppm. The more total analyte mass present the less gets left adsorbed on active sites.

I thought it might be the trap.
(1) So this is not an indication of need for new trap
(2) This is not an inlet/split ratio problem
(3) Not active site deactivation in xfer lines and inlet
(4) RSD dependent on all sections of the trap being filled.

(4) indeed seems to be a possible explanation. The RSD's of surrogates in MS/MSD samples are about the same as those for CCV's.

(3) reminds me of back in the 1980's when I used to do PCB's by GC-FID. I always made sure to run a spike and blank before I set up a new run. Otherwise the first CCV would have some failed recoveries. I reasoned that was an active site problem.

I purge with UHP nitrogen. I leave standby flow through the trap when its idle (Tekmar 3000). With some LN2 dewars I see a buildup of tetrahydrofuran on the trap because of this. So, if its been more than overnight I run a "BFB" standard before the CCV and Blank. The BFB acts as a cleanup and also conditions the P&T system to be ready for the rest of the run.

LALman wrote:
I thought it might be the trap.
(1) So this is not an indication of need for new trap
(2) This is not an inlet/split ratio problem
(3) Not active site deactivation in xfer lines and inlet
(4) RSD dependent on all sections of the trap being filled.

(4) indeed seems to be a possible explanation. The RSD's of surrogates in MS/MSD samples are about the same as those for CCV's.

(3) reminds me of back in the 1980's when I used to do PCB's by GC-FID. I always made sure to run a spike and blank before I set up a new run. Otherwise the first CCV would have some failed recoveries. I reasoned that was an active site problem.

I purge with UHP nitrogen. I leave standby flow through the trap when its idle (Tekmar 3000). With some LN2 dewars I see a buildup of tetrahydrofuran on the trap because of this. So, if its been more than overnight I run a "BFB" standard before the CCV and Blank. The BFB acts as a cleanup and also conditions the P&T system to be ready for the rest of the run.

We always start with the same three injections. Since BFB is a Surrogate we just run what would be a normal blank and call it the Tune check, then CCV, then CCB, I always figured the first shot would clean out the system instead of doing a simple trap bakeout before starting the run. Some methods even say to bake for an hour before starting, but that was when we used the Tenax/silica gel/charcoal traps, which not many have used for several years(decades).

I always fought with the surrogates or internal standards having different responses in a CCV versus those in blanks or regular samples. I was told to add methanol to the samples in the same amount that would be in the CCV, but that never worked. I tried different inlet temps, and just about ever other trick I read. I only noticed the effect I mentioned when I was doing that special ultra low level study for a client. There is a thread here a few years back about it that was several pages in fact.

One thing I did notice is the effect is less when the moisture trap on the Encon purge and trap is new. The Encon uses an empty trap at room temperature to condense out the moisture during desorb, can't remember if the Tekmar 3000 uses that or something different. I have never been able to prove it, but one theory I had is that over time the trace organics left on the moisture trap(and could happen in the analytical trap also) build up on the surface and when baked they form a char, which would result in a very thin layer of carbon that retains some of certain analytes, and when you overload it they can't adsorb, which causes those analytes to have better response when total concentrations are higher. I did reduce my moisture trap bake temperature from 260C to 180C and I see much less of the effect and the moisture traps need to be replaced less often, so that is one anecdotal test I have to back it up, but nothing scientifically proven. In a sense that would cover number 3 on your list also.