Chromatography Forum

Dear Reader,

I appear to be having some problems with the precision of a SIM method for the separation of 6 compounds + ISTD. The method I am using is as follows:

System – Agilent 7000d
Column – HP5 MS-UI, 30m, 0.25/0.25
Run Solvent – IPA
Injection Volume – 1ul – This equates to a vapour cloud of around 50% liner volume so backflash unlikely
Injection Parameters – 6 x prewashes (Solvent A and B), 3 sample washes, 4 sample pumps. No dwell time, draw at 300ul/min and dispense at 6000ul/min
Injection type – Splitless
Injector – Either 180C for 30s then ramp at 900C/min to 350C followed by purge at 1.33 mins or 180C isothermal – Former is slightly better overall
Flow (He) – 1.2ml/min
Oven – 50C hold for 3mins, ramp 15C/min to 150C hold for 2 mins, then ramp 40C/min to 265C hold for 8 mins.
Transfer Line – 250C (280 results in even poorer RSD!)
Source Temp – 230C
Quads – 150C

Overall, the chromatography is rock solid with great peak shapes, good separation and RT RSDs of 0.01% or less! However, peak area precision is poor and I have been trying to work through the problem with little success. For troubleshooting, I am injecting about 10PPM of each component and my within batch variation (5 sequential runs) varies from around 3% up to >8% depending upon the compound. Inter batch variation is pretty big with RSDs of up to 27% which obviously makes the method unworkable!

It seems to me that this is a problem that lies either with the ALS or with the injection, but am willing to be convinced otherwise. Any suggestions on dealing with this?

Kind Regards


TD2

Is it 5 shots from the same vial that gives 3-8% or 5 separate vials?

Second batch with 27%, is that another fresh set of vials or reshooting the original vials?

Are the target analytes volatile with boiling points less than 100C?

Hi James,

Many thanks for the response and of course for all your work on the forum

Annoyingly, all 5 samples come from the same aliquot. My diluted stock was combined with ISTD in a tube which was well mixed using a vortex mixer and then pipetted into 5 separate tubes. These were run sequentially with 2 blanks prior to the samples and a blank between each sample on a GC which has been set at method parameters for a number of hours before running. All vials are 'Surestart level 3' screw cap, so the vial seals should be pretty consistent.

I repeated this same process (with a fresh aliquot) over a number of days (Day 0, 3, 7, 11) and this yielded the RSD figure of 27%. I would love to be able to say that the result is down to sample degradation, but the compounds are pretty stable and the overall trend in peak areas is not a negative one if there is an actual trend!

In terms of BPs, the analytes are not especially volatile with BPs in the 160-350C region.

I have of course considered whether this inconsistency could be down to my pipetting, but all transfers are made with pipettes I have personally calibrated (specifically for my solvent, with pre-wetted tips and in reverse mode) and I spent 10 years in molecular biology, so I would like to think that pipetting is less likely to be the weak link in all this!

Many thanks for your time

Kind Regards

TD2

You mention area RSD is high. How about calculated concentration? If these are solid it would definitely indicate an injection issue, either inconsistent injection volume or inconsistent extract transfer to the head of the column.

You may consider trying a different syringe; I prefer ones from Agilent with the PTFE tipped plunger as these give me more consistent injections for longer period of time.

Another relatively easy troubleshooting idea I've used before is to take the splitless method and turn it into a 5:1 split, run a bunch of injections and compare RSDs. If they've improved dramatically consider trying pulsed splitless.

Hi CJM,

Many thanks for the response and in particular, for getting me to think about pulsed splitless - More on this shortly!

As it turns out, there is a very clear (and embarrassing) reason for the lack of precision that I described in my previous e-mail. When it happened, the first thing I thought of was liner activation but dismissed this as the liner and the septum were brand new; except the liner wasn't! In our lab, all instrumental maintenance and anything requiring hardware interaction must done by lab techs and when I asked for the instrument to be set up, it was fitted with an old liner which was full of black detritus from septum particles... When I discovered this, the new liner was promptly fitted and all appears to be OK, with RSDs of around 1% or less.

Asa result of your response, I did however give PP injection a go and it works very well. I had tried it in the past with limited success (my fault for lacking the knowledge at the time), but this time, it has provided results that are comparable with my PTV method (similar response and similar RSD) with the added benefit of the option of being able to near quadruple injection volume if required and that's with an unoptimised 'guesswork' method!

PP may well open up a world of potential possibilities for which I am most grateful!

Kind Regards

TD2

Glad you discovered the problem.

Pulsed splitless or pulsed split modes are what I often use, especially if the solvent has a lot of expansion upon injection. Compressing the solvent plug helps avoid backflash but also tightens up the peak shapes and often I see less tailing with the solvent peak.