Chromatography Forum

Hi, dear all

I am doing HS-GC method development and validation of residual solvents, methanol (MeOH) and isopropanol (IPA), in an API.

The RSDs for peak area of the two analytes from 6 replicat injections of the standard solution are really bad. Usually they are greater than 5%.

Method parameters are listed below:

Diluent: Water
Standard Solution: MeOH 150 μg/ml, IPA 250 μg/ml

GC: 7890A, FID Detector
Inlet temperature: 200 ℃, Detector temperature: 250 ℃
Oven Temperature: 40℃ for 5 min, 38℃/min to 230℃, hold for 8 min
Colum: DB-624, 30m*530μm*3μm
Carrier gas: Nitrogen, Flow rate: 4.8 ml/min, Split ratio: 10 to 1
Hydrogen flow rate: 40 ml/min, Air flow rate: 400 ml/min
Make-up flow rate: 35.2, Septum purge flow: 5 ml/min

HS: G1888, Oven temperature: 80℃, Loop temperature: 90℃
Tran. Line temperature: 105℃, Shake: High, Vial pressure: 15.2 psi
Equilibration time: 25 min, Vial pressurization time: 0.4 min
Loop fill time: 0.2 min, Loop equilibariton time: 0.1 min
Injection time: 1.0 min, Injection volume: 1.0 ml

A point to note is that this HS-GC instrument gives me wonderful RSD <2%, when I am runing similar methods with other diluents such as DMF and DMI. So I am keen to find out any unsuitability of the diluent (water) for the method parameters above.

Hope I have said what I want to say due to my unpolished English. Please shed some light on this.

Thanks in advance.

Terry

It's pity that ur problem hasn't been solved. Maybe u could find some useful information or help here!Good luck!

You included so much information it was a joy to see. Thanks, BUT, one detail you left out was the AMOUNT of water in your vial and the vial volume. How much alcohol by weight are you putting into the HS vial?

Now the amount of alcohol in your water solution is rather high. I routinely would have no more than 10µg of any alcohol in a vial, which corresponds to a very small volume of your sample. This size for IPA using a 9mL vial would give a signal that would be FULL SCALE on a Varian 3500 GC. See peak #5 in this sample of 0.5µg of IPA.



I would suggest you use a sample volume less than 250µL and have concentrations of less than 50µg/mL for your headspace analysis. Using small volumes of sample (less than 200 µL the optimum time to heat at 80°C should be about 8-12 minutes. Heating longer will increase your RSD values to above 2%.

best wishes,

Rod

ps I forgot to add that the higher partition of the alcohols using water is the cause of the saturation of your HS vial. There is less partition into the gas phase if DMF or DMI is used instead of water and thus you did not exceed the saturation point using those solvents.

You might be getting condensation in the sample loop. A headspace method that I worked on a while back with water as the matrix gave the best repeatability with the loop at 110, transfer line and inlet at 120.

If that is no good you could try using an internal standard (acetone?).

To Rod,

The sample size is 5 ml, and the vial volume is 20 ml.

I got 750µg of MeOH and 1250µg of IPA in one single vial with prolonged equilbration time up to 25 mins. This is not right according to the rule of thumb given by you.

I would reduce the sample size, the STD concentration and the equilibaration time in the following days.

To Modonovan:

I also would like to raise the loop and inlet temperature to avoid condension.

Many thanks to both of you

Your temperatures are reasonable. For years for many methods I never went beyond 110°C for the loop and transfer line. As long as you don't use a solvent with a higher temperature requirement (I never went past C9s and Di Methyl Acetamide) that should be adequate.

If your headspace is set at 80°C you are not likely to have condensation at 105°C from the water vapor saturated sample. You will have variability of the alcohol saturated vapor when you pressurize the vial and then sample it at 80°C. You could lower the temperature of the sample to 75°C but this will reduce the recovery at low levels.

You will be wise to reduce your sample to 1mL or less. The smaller the sample the quicker the equilibration and the more repeatable the analysis.

You might be very surprised how little your alcohol peak changes in size as you reduce the volume in the vial but increase the partition into the gas phase. I would have you try 250µL and heating for 15 minutes.

Good luck and tell us your results.

best wishes,

Rod

I am back to report the result of my work on the MeOH and IPA thing.

First, I reduced the sample size to 1ml (this is the size of the smallest pipette available in my lab), decreased the equilibration time to 12min (I ran a sequence samples with different equil. time and found 12min was enough). The peak response did not change at all. I was really surprised. Unfortunately, the RSD was still terrible.

Next, I diluted 10ml of the original standard solution to 50ml with water. The concentration of MeOH and IPA are 30μg/ml and 50μg/ml, respectively. With HS parameter as in the first step, I just got an 1/5 peak response as above. Then guess what about the RSD. Oh yeah! It is awesome (<2%).

Thanks so much for sharing your expertise, Rod.

I get two more question to ask: Is the sample size (1ml) appropriate for the vial size (20ml)? Should I use smaller vials?

Regards,

Terry

Terry,

Glad to hear the physical laws work the same in China as they do in the rest of the world. My opinions have been discounted on this topic by others and I am very happy your results agree with my experience. It means I am not crazy (or not as crazy as popularly supposed).

I never needed 20mL vials to do my work. I found that 6 to 12 mL vials were quite adequate. The only time I needed 20mL vials were when I looked for benzene in water at 25-500 ppt by headspace using a FID and 25-100 ppb for chloroform in water with an FID because no PID, MSD, or ECD were available.

I almost never used a sample in pharmaceutical analysis that was larger than 0.25mL.

best wishes,

Rod

Hi All,

I have a similar problem as Terry here so I thought I would post it in the same thread. I am trying to develop a headspace method for methanol and t-butanol using a water as the solvent but my RSD’s are terrible (15%). My operating parameters are listed below, if you have any thoughts I would appreciate it.

Method parameters are listed below:

Diluent: Water
Standard Solution: MeOH 100 μg/ml, t-Butanol 750 μg/ml

GC: 6890, FID Detector
Inlet temperature: 140 ℃, Detector temperature: 250 ℃
Oven Temperature: 70℃ for 5.5 min, 20℃/min to 140℃, 40℃/min to 200
Colum: RTX-1301, 30m*530μm*3μm
Carrier gas: Helium, Flow rate: 3.6 ml/min, Split ratio: 5 to 1
Hydrogen flow rate: 40 ml/min, Air flow rate: 400 ml/min
Make-up flow rate: 30 ml/min,

HS: 7694, Oven temperature: 80℃, Loop temperature: 105℃
Tran. Line temperature: 105℃, Shake: High,
Equilibration time: 12 min, Vial pressurization time: 0.5 min
Loop fill time: 0.2 min, Loop equilibration time: 0.2 min
Injection time: 1.0 min,

Thanks

If your vials are not sealed properly you will have high RSDs. An issue often ignored.

1. How much sample are you putting into the vials?

2. How large are the vials?

Rod

1mL of sample into 20mL vials and the vials are sealed properly with a crimper (the lids don't move after they are crimped)

First thing to change:

Reduce your sample size from 1mL to 100 microliters. Yes, a 10 fold reduction !

Then see what happens. You might increase sample size if your RSD values are shown to be below 3%.

Your sample is very highly concentrated and you might be saturating the headspace, delaying the equilibrium.

best wishes,

Rod

ps

Second thing to check: I am not saying that this is definitely the problem, but please check.

You can overtighten the vials. They might not seal properly when heated. Been there done that.

Third thing to check: Your transfer line AND/OR the column installation in the splitter injection port.

I addition to all that Rod has said; try cutting the vial pressurize time to zero (you have enough pressure and volume in a 20 ml vial at 70C to fill the loop), increase the loop fill time (with a shorter time you might be cutting the fill while headpsace is still flowing through the loop at significant rates, making the operation vulnerable to small differences in flow and timing), eliminate the loop equilibrium (once the sample is in the loop nothing good can come of leaving it in there).

I would also look at increasing the carrier gas flow and decreasing the initial column temperature to get some benefit in terms of sharper peaks from stationary phase focussing. Also, unless you definitely have heavy components that you have to bake off, which is intrinsically unlikely with equilibrium headpsace, there is no use in punushing you column by programming up to 200 every time.

How is the headspacer connected to the inlet ?

Peter

The headspace is connected via a transfer line to the inlet.

A connection is only a connection. It can be done poorly, inadequately, badly enough to cause poor RSD values.

Personally, I don't like splitters. Why split when you can reduce the loop size and do a direct injection?

Read my paper: Analytical Chemistry June 1997

I injected upto 1mL samples directly with a 20-30 psi carrier pressure using a flow restrictor on the tail of the column of 1 to 5 meters of 0.25mm ID tubing.

best wishes,

Rodney George