Matrix Sample Discrimination in GC HS

[lemonleslie]

I have a GC headspace method which is giving me great precision (area %RSD <2) but recoveries in spiked experiments of 130% (in which the sample is spiked with the standard solution to eliminate any error due to different preparations). The standard is prepared at 0.15% w/w methanol and ethanol (0.15 mg/mL). I am absolutely stumped on where to proceed with this headspace method. The method is exhibiting a matrix effect where the response is higher in the samples than the standard.


G1888 Headspace Sampler
Vial pressure 0 psi & 10 psi with same result (supplied by GC EPC Aux)
Headspace oven 80 C
Loop temp 105 C
Transfer line temp 110 C
Equilibration time 15 min, high shake
Vial pressurization time 0.5 min.
Loop fill time 0.15 min.
Loop equilibration time 0.10 min.
Injection time 0.5 min.

Column
Restek Stabilwax 60 m x 0.32 mm x 1.0 µm

6890N GC
Injection Port Split/Splitless
Temperature 140 C
Split ratio 41:1
carrier gas Helium
Column flow 1.6 ml/min (resulting inlet pressure of 17 psi)
Detector FID, 240 C
GC oven program
initial temperature 100 C (isocratic)
GC cycle time 15 min.

diluent 75/25 ACN/Water, 5 mL
sample desloratadine

analytes
methanol and ethanol

[chromatographer1]

lemonleslie

Forgive my soapbox response, but after dealing with issues like yours with management, the solution became clear and undeniable. Never use external standards for quantifying analytes by headspace.

It takes a little longer (not always, do you ever run multiple preps of ext stds?) but it removes all issues of linearity and accuracy, and sample homogeneity if you use standard addition methods to measure your impurities.

Simply, make three preparations of your sample.

Add your standard solution once, twice, and thrice (or not at all). Add the standard solvent to the same three preparations, twice, once, and none, (or three times).

Example: 100 mg of drug weighed out into each of 3 vials.

Add 100µL of Standard solution: solvents (1mg/mL of each) in DMF
to first vial.

Add 200µL of Standard solution to second vial.

Add 300 µL of Standard solution to third vial. (or add 0µL of Std Sol.)

Then add 200µL of DMF to first vial.

100 µL to second vial

and 0µL of DMF to third vial (or add 300 µL of DMF to third vial)

Now run the 3 preparations. Say this takes 1 hour (20 minutes each)

if you ran one sample and one Std by external measurement it would take 40 minutes, and another 20 minutes arguing whether the results were good or not.

But using standard addition method, you immediately will demonstrate linearity of the method and homogeneity of the sample. Two important concerns about which you will never have solid answers if running only one preparation and one standard.

Plus if the results are non-linear you know you must determine the cause of the problem, but at least you KNOW you have a problem. Running an external std method will NOT give you that information.

Auditors love to see SOLID and UNDENIABLE science when they review research. Standard addition methods avoid so much controversy.

I advocate that you always use this technique unless you are sample limited. (been there, done that when I had to)

best wishes,

Rod

[krickos]

Most certainly agree with Rod.

Your sample is causing a matrix effect also called "activity coefficient" in some litterature.

ACN as sample/stad diluent can also cause you trouble with septa, make sure you have picked the right one:
http://www.chem.agilent.com/en-US/Produ ... ction.aspx

If you change diluent to DMF, do not forget to raise temp in loop and transfer line to boiling pont of DMF ie something like 150/160°C too keep loop free of DMF residues.

[chromatographer1]

krickos' useful remark:

"If you change diluent to DMF, do not forget to raise temp in loop and transfer line to boiling pont of DMF ie something like 150/160°C too keep loop free of DMF residues."

is especially important if your sample path contains exposed metal and is not composed of inert fused silica or is not fused silica/glass coated.

Rodney George