Headspace Vs Direct injection

[chromatographer1]

Approximate µL Gas Volumes of Solvents per µL injected @ inlet temperatures and pressures
Head Pressure @ 200°C Head Pressure @ 300°C
Solvent 10 psig 20 psig 30 psig 10 psig 20 psig 30 psig

EtoAc 230 168 131 286 204 158
Hexane 177 126 98 214 153 119
Isooctane 140 100 77 170 121 94
Methanol 570 406 315 691 492 382
CHCl2 360 257 200 437 311 241
MtBE 194 138 108 235 167 130
Water 1279 910 706 1548 1102 855

[IrishMike]

Hi all

I have been playing with this on and off over the last few days (other stuff getting in the way)

I ran a std of 0.1% meoh, chex, etac, acetone ipa and toluene in acetonitrile..
We have loads of acetonitrile available so I deciede to use this as the matrix..

I ran 20% and 50% mixtures of IPA, acetone, meoh in water. These were diluted 1:100 in acetonitrile and injected 2 microlitres of each sample.
The values returned for meoh were spot on however acetone and ipa were off by about 5%...20 and 50% chex and toluene in acetonitrile are also good...

Any thoughts??

Thanks for the help so far..

Mike

[Peter Apps]

Hi Mike

Getting within 5% on a first shot is not so bad, so now we need the details of what and how, temperatures, flows, split ratios, hardware etc etc.

Peter

[IrishMike]

Hi Peter

I am just back from a weeks holiday so I will be getting stuck in to this again..

I will post details during the week..

Regards

Mike

[IrishMike]

Hi All!

I have gone with the HS option as we wern't getting good results with liquid injections.
Here is the test method we are currently running:

Column : CP Wax 52 CB
Injector: 180 C
Flow 1 ml/min
Split ratio: Initial 50, 0.01 min 100, 0.02 min 50
Oven 50C, hold 1 min, to 90 C at 4 C per min hold 1 min
Detector : 200 C
Headspace parameters: Agitator: 40 C, 3 min incubation
Syringe 60 C,
30 microlitre per vial.
Injection vol: 100 microlitre

Let me know if you need anything else

Mike

[Peter Apps]

Hello Mike

Two comments; do not change the spit ratio during the injection - the only effect will be to give inconsistent gas flows in the injector, and if you have samples with a lot of water you will need to increase the agitator temperature to get complet evaporatoration of the sample.

Peter

[IrishMike]

Hi again

I am still plugging away at this and have now switched to using on column injection insted of HS. Settings as follows:

Column : CP Wax 52 CB
Injector: 180 C
Flow 0.8 ml/min
I have set the split ratio at 200 for the entire run
Oven 35C, hold 1 min, to 65 C at 2 C per min hold 1 min
Detector : 200 C

I am injecting 0.5 microlitre of 0.1%, 0.5% 1% and 5% stds.RDS were working out ok with the r2 generally around the 0.999 mark!
However when I ran a prepared mix of 20% each of IPA, ETOH,Acetone,THF, 10% Toluene and balance H2O.
I have been getting results around 2-3 % above the actual values

Any thoughts on how to get improved results??

Thanks

Mike

[Peter Apps]

Hi Mike

Probably the calibration is not completely linear over the very wide range of sample composition that you are injecting. You need a calibration that spans the concentrations of the real samples.

Is this really on-column, or is it split injection ?, 0.8 ml/min flow is appropriate for 250 um columns, which need special very thin needles for on-column.

Peter

[IrishMike]

Hi Peter..

You are right it is a split injection.

Regarding the linearity my problem is that our sample can be anything from 1 to 100% solvents with the balance water..

I never mentioned that I dilute the sample 1:100 and inject the diluted sample.

regards

Mike

[Peter Apps]

A calibration does not have to be linear to produce good results, you can fit all sorts of curves using desk-top software and go from there. Or you can do separate calibrations over shorter regions of the concentration range.

If you are diluting the samples in water you could try a genuine on-column injection if you can use a 530 micron column or retention gap you can get inlet liners that guide the tip of the needle into the top of the column. Restek does one.

Is there anything that you could use as an internal standard - i.e. something that is not going to be in the samples, and that you can resolve from what is there. ?

Peter

[IrishMike]

Hi Peter

I am diluting the samples in acetonitrile, the use of an internal std could be tricky as we are analysing waste which potentially contain anything!

As I dilute 1:100 should the calibration be in the range 0.5% - 2%??

Mike

[Peter Apps]

Hi Mike

Your standards need to duplicate your samples as closely as they can. So make up standard solutions in water, with several solvents in each standard, at concentrations that span what you get in the samples. Then dilute them in AcN at the same ratio that you use for samples. When you plot/calculate the calibration, use the concentrations in the orignal undiluted standard, not the concentration in the acetonitrile. Then when you calculate from sample peak areas you get the concentration in the sample directly without having to bother with the dilutions.

Peter

[IrishMike]

Thanks again Peter

I used AcN as diluent because we often get chex, toluene, hexane in our samples which are imiscible with water. I dilute the samples 1:100 in AcN to get the samples in the range of the standards, as most of our samples are mixtures of various solvents.

Mike

[sdegrace]

One note with headspace, we have found a dramatic improvement in the precision by using an internal standard. If you are using headspace and you can go that route I highly recommend it. I just finished designing and validating a residual solvents method where I had no choice but to use HSS (product is a water-soluble inorganic salt) and I chose to use internal standard calibration and never regretted it. The validation went extremely well and the precision of the response versus ISTD was better than 0.3% RSD (n=6) at most calibration levels. This was way better than just taking the solvent peak areas. The method was also effectively remarkably robust due to the internal standard... I could make changes which were causing huge swings in the actual peak areas and my recoveries were still bang on.

Stephen

[Peter Apps]

Hi Stephen

0.3 % ! Please tell us what instruments and conditions you used to get this.

Peter