Standard Inconsistencies using Headspace Analysis

[Jdougherty88]

Hello,

Currently I am having very inconsistent results with my standards when performing headspace analysis.

A stock solution containing 5 analytes (let's call them A, B, C, D, & E) in solvent is prepared. The stock solution is diluted directly into headspace vials, these vials are then analyzed as standards. Three vials of six different concentrations are prepared. The standard curves for analytes A, B, & D are consistent. This leads me to believe that I am not having a standard preparation problem, since all of the analytes are mixed together. However, the standard curves for C & E are very inconsistent.

For example, the three injections at 60 ppm may give me the following results:

Analyte - Injection 1 (area) - Injection 2 (area)- Injection 3 (area)
A - 50000- 52000- 48000
B - 60000- 60500- 59500
C - 30000- 21000- 37500
D - 55000- 58000- 53000
E - 40000- 48500- 32500

You can see that the %RSD for C & E would be very high, while the %RSD for A, B & D are low.

Has anyone ever seen this type of problem and can you offer any solutions/suggestions? Thank you!

[rb6banjo]

Chemically, what are C and E with respect to A, B, and D? What I'm getting is are A, B, and D hydrocarbons and C and E alcohols? That type of thing.

[Jdougherty88]

A - methanol
B - ethyl ether
C - ethylenediamine
D - toluene
E - 2-ethyl-2-oxazoline

I am using a Restek Rtx-Volatile Amine Column and FID detector

[rb6banjo]

Did your method behave before and not now?

Of that list, I see C and E as being the biggest problem children. Basic materials (amines, oxazolines, etc.) are some of the worst analytes to determine chromatographically. They adsorb to everything. If your surfaces are not well conditioned (transfer lines, etc.) it's really tough to get them through to the detector. You could try to run higher concentrations of those 2, multiple times, to see if your precision gets better as you put more of those analytes through your system.

I would expect the other 3 to be much more well behaved than the 2.

[Jdougherty88]

those 2 chemicals have never behaved for me.

I will try to run higher concentrations.

[Jdougherty88]

Would it be possible to add an additional analyte to compete for the adsorption onto the system, thus freeing the analytes of interest to make their way to the detector?

[krickos]

Would it be possible to add an additional analyte to compete for the adsorption onto the system, thus freeing the analytes of interest to make their way to the detector?

Not commonly. Taking Control of pH of sample solution could help to stabilize headspace equilibration and thus precision.

But the common issue with higer boling Active compunds require Another approach.
1. Use an inert materials in Contact as possible, Ni material poorer that silcosteel if using Agilent, Think silosteel is standard nowdays.

2. Most common mistake. Temperature in valve-loops-transferlines do not exceed the boling Point on your analytes. Causes interaction with Surface.

3. Related to 2. People forget the sample diluent, if you use DMSO high boling point you should increase temp as mentioned above, residual DMF, DMAA, DMSO in your lines may cause decreased sensitivity and precision.

[Jdougherty88]

I am using a Shimadzu GC2010-Plus.

Injection port temp: 180 C

This is above all of the analytes' and the solvent's boiling points. N,N-dimethylacetamide (DMAc) is the solvent (bp = 166 C).

The auto sampler (CTC Pal Headspace) incubates for 5 minutes at 140 C. The syringe, set to 140 C, then picks up 1 mL of the sample and makes the injection. The syringe's maximum temperature is only up to 150 C, I don't want to push the envelope.

I am not sure what type of materials Shimadzu uses in their GC, but I'm betting that this interaction is the reason for the inconsistent results.

[dlbenach]

Two simple things to look at if you haven't already....
1) Syringe flush time. Maybe increasing it would help.
2) Syringe fill strokes. Having at least 1 if you do not have any may help

These probrably won't solve your problem but I would hate to overlook something simple.