Hi,
I am currently running a USP <467> protocol which monitors the recovery of: methanol, ethanol, acetone, isopropyl alcohol, hexanes, and toluene. I make a standard solution using DMAC with known amounts of solvent and use the same solution to spike my samples and monitor the recovery of the solvents in the spiked sample solutions. The original procedure used an Agilent 7697A headspace sampler while my company has an Agilent G1888A headspace sampler.
Both GC's have the same method:
DB-624 column
Injector temperature: 180C
Split ratio: 1:10
Detector temperature: 250C (FID)
Carrier gas: 3.0 mL/min
Hydrogen flow: 30.0 mL/min
Air flow: 300 mL/min
Makeup flow: (N2) 25 mL/min
Run time: 20 min
The headspace parameters from the Agilent 7697A headspace sampler is as follows:
Oven Temperature: 80C
Loop Temperature: 100C
Transfer line Temperature: 110C
GC Cycle time: 35 min
Vial equilibration time: 20 min
Fill pressure: 15 psi
Fill time: 1 min
Injection time: 1 min
Loop final pressure: 10 psi
Loop equilibration time: 0.2 min
Vial pressurization time: 1 min
The headspace parameters that I adopted on our Agilent G1888A headspace is as follows:
Oven Temperature: 80C
Loop Temperature: 100C
Transfer line Temperature: 110C
GC Cycle time: 35 min
Vial equilibration time: 20 min
Fill pressure: 15 psi
Fill time: 1 min
Injection time: 1 min
Loop equilibration time: 0.2 min
Vial pressurization time: 1 min
Shake setting: high
Recoveries for all solvents but acetone are around 100%. However, the acetone recovery is consistently around 70%.
My initial thoughts are that the vial pressurization time between the headspace samplers are different, as the newer headspace model allows the feature where you can set the loop final pressure. I've tried running the test at 0 min, 0.2 min, 0.5 min, and 1 min for the vial pressurization time. All but the 0 min time give a recovery of 70% of acetone recovery. However, at 0 min, all the other solvents are spiked to 120% recovery. I've ruled out evaporation issues, liner issues (straight vs tapered, wool vs. just glass), diluent (DMAC) issues, and have tried this on two separate GC systems.
Do you think that the protocol conversion between the two headspace samplers could result in such a difference? The original method validation report has an 85% average recovery of acetone. I've also run this before in July and got acetone recoveries of about 83-85%. I'm stumped as to where the low acetone recovery is coming from. Do you have any thoughts?