Chromatography Forum

Try a lower incubation temperature and a shorter equilibrium time - you are generating high pressures in the vials by equilibrating at 105C, and 45 min gives a long time for small leaks to affect your analysis.

However, this still does not explain why the first vial is consistently higher than the others. Try running a blank vial (everything except the benzene) as the first vial in the series to eliminate the possibility of carryover.

Have you tried the experiment of filling and capping each vial just before it starts to equilibriate ?

You may be tightening the caps too much - does the top of the septum have a dent in it after you have crimped the cap ? If it does you have probably split the teflon liner, and the sample is exposed to the silicone, which is absorptive.

Peter

what kind of rubber is used in the septa?

You can overtighten vials. Are they screw-on caps? Or do you use a crimper?

I still suspect that since other solvents give you excellent reproducibility (am I wrong?) that it is the septa that is the problem.

The other issue is the exact procedure used in making the preparations.

You should prepare spiking solutions in a non=polar solvent such as DMF, including all dilutions, then add a reproducible amount of solvent (DMF) containing the benzene to the vials immediately before crimping the vials shut.

If you are not getting enough sample created by the pressuring the vial and heating the vial that you flush the sample loop completely at least three times its volume then you could also have variability. But since your FIRST vial is higher in response, it appears to be a leak OR loss by absorption by your septa.

Rod

i was using septa:agilent 5183-4477, each one have PTFE thin film.

usually last four vial peak area is similar, the first one area biggest, the second vial take second place。
i use crimper，not screw-on caps。overtighten not possible to well explain my phenomenon。

i think your opinion is reasonable to suspect septa。PTFE not suitable for trance benzene analysis？

your suggestions about prepare benzene solution using fully nonpolar solvent，is excellent。benzene is more steady in such solvent。but how to comply with USP 467（5ml water add 1ml mixed standard，now 6 components to assay，not only benzene），is a problem。maybe i should think over how to practise。also dilute process will use big volume of DMSO，or DMF，it's a waste to enviroment protection rather than now almost all water。


now i think loss by absorption of septa more reasonable for my observed phenomenon. i still confused，what i should do next。my job QC of pharmaceutical company，not my responsibility to change analytical method。

thanks very much，for ur help。

best wishes
dogcatlike

now i change my view。transfer solution process benzene loss of vaporation，may play vital role。i will try， make mixed standard solution ahead 1 hour，then transfer to vial to check。maybe newly prepared solution not steady enough，speed of vaporation decrease gradually。

as for septa adsorption，the temp 105 degree may be a hindrance to adsorption。if really cause of adsorption，maybe i should ahead of many hours to saturation the septa to get rid of the difference of six vials，due to different adsorption time。

Dogcatlike, You wrote:

"overtighten not possible to well explain my phenomenon。"

You are wrong. Sorry to be so blunt. Overtightened caps LEAK. There is a fine line between too loose and too tight. If you can barely rotate the cap and septa with great strength after crimping the cap, then it is about the correct tightness.

Back in 1995 I spent three weeks of research, checking and determining how much adjustment I needed to make to achieve a consistent and leak-proof seal. It drove me crazy. When I could get 1 microgram of benzene to reproduce at less than 0.2% RSD for 6 vials THEN I was happy with the crimping tool.

If you wish to claim you are preparing samples at concentrations prescribed by the USP specifications, fine, do so. But IN MY OPINION if I am told to put 100 ng of benzene, 1mL of DMSO, and 5mL of water into a vial, I will do it. But why should I put one into the vial first and the other second or third? As long as I put all three into the vial EXACTLY, who should care about whether the benzene is dissolved into the water or into the DMSO? And if I put the cap on the vial instantaneously after putting the last component into the vial, why should I have to wait 5 or 25 seconds before I do so? Do you see my point?

There are good reasons sometimes for following a procedure EXACTLY as written, in the order that it is written, but we are supposed to be educated chemists and use good practical scientific knowledge in what and how we make our samples.

I used to put 5 solvents into a small pre-weighed volumetric flask by weight, say: 75.03 mg, 99.10 mg, 34.76 mg, 222.12 mg, and 45.55 mg into 23mL of DMF. Then add enough DMF to prepare exactly 25mL of the 5 solvent DMF solution. Then by weighing the flask I would determine the density of the solution. When I transferred a portion from this flask I would weigh the transfer and know VERY ACCURATELY the volume I transferred and from that value, how much of each solvent I transferred as well. My transfer error was not 1% but less than 0.001%.

When operating in a reasonable range of concentrations and temperatures, headspace is more accurate than the best chemist. My samples always told me how good I was that day in making my preparations.

When I taught a new chemist how to make HS preparations, he was usually better than I was on the very first attempt, getting EXTREMELY good values because he took his time and did VERY careful work because he was nervous in learning how to prepare them, and I was a tiny bit sloppy as I was in a hurry and only tried to get the samples made GOOD ENOUGH. So his 6 stds were less than 1% and mine were less than 2%.

Because benzene has such a HIGH partition value out of water it is a good diagnostic analyte to check how well and carefully you are preparing samples.

Good luck. I am certain success is just around the corner.

Rod

Try a lower incubation temperature and a shorter equilibrium time - you are generating high pressures in the vials by equilibrating at 105C, and 45 min gives a long time for small leaks to affect your analysis.

However, this still does not explain why the first vial is consistently higher than the others. Try running a blank vial (everything except the benzene) as the first vial in the series to eliminate the possibility of carryover.

Have you tried the experiment of filling and capping each vial just before it starts to equilibriate ?

You may be tightening the caps too much - does the top of the septum have a dent in it after you have crimped the cap ? If it does you have probably split the teflon liner, and the sample is exposed to the silicone, which is absorptive.

Peter

‘Try a lower incubation temperature and a shorter equilibrium time’, I agree with u for good rsd. But peak area may also decrease, maybe a influence to sensitivity。Mean smaller benzene peak，and carbon tetrachloride in class1 standard smaller peak，hard to pass the signal-noise-ratio requirement。That’s all my imagine，actual situation need to try to know。

The vial before the six vial is class 2A standard，no benzene in it。I see its chromatogram，corresponding position very small peak，area 0.7（before 2A is class1 standard，containing benzene）. So small that it doesn’t matter to my assay，can’t be the reason to harm rsd。

I thought 0.1ppm benzene water solution not steady，if not parallelly transfer to vial and cap，will bring more difference，maybe bigger rsd。So never try as you say。And the manner will tire person，making person in pressure，time between injections about 66min。must late for regular bus，have to extra work。

Occationally tightening the caps too much，I really found some vial septa side having dent。
What like is ideal state to cap。I thought regular dent all around the aluminum cap is perfect，this mean septa is compress，thickness lessen，tighter。Or absolutely flat at the side of aluminum cap，is best？there should be some moulage at the side of aluminum cap（I mean the position, when turn cap, fingers hold position）, for ideal operation？instinct I doesn’t think it’s decisive factor to terrible rsd.

A new idea strike me. Maybe the benzene water solution of 0.1ppm, not homogenous。Benzene density lower than water，maybe part dissolved，part in other style floating in upper bottle。When draw solution，different depth and time，contribute to bad rsd。Maybe the solution prepared keep still 1 hour before transferring to vial，helpful，although maybe loss some benzene not dissolved。

Thanks for ur attention。
Best wishes。

Dogcatlike, You wrote:

"overtighten not possible to well explain my phenomenon。"

You are wrong. Sorry to be so blunt. Overtightened caps LEAK. There is a fine line between too loose and too tight. If you can barely rotate the cap and septa with great strength after crimping the cap, then it is about the correct tightness.

Back in 1995 I spent three weeks of research, checking and determining how much adjustment I needed to make to achieve a consistent and leak-proof seal. It drove me crazy. When I could get 1 microgram of benzene to reproduce at less than 0.2% RSD for 6 vials THEN I was happy with the crimping tool.

If you wish to claim you are preparing samples at concentrations prescribed by the USP specifications, fine, do so. But IN MY OPINION if I am told to put 100 ng of benzene, 1mL of DMSO, and 5mL of water into a vial, I will do it. But why should I put one into the vial first and the other second or third? As long as I put all three into the vial EXACTLY, who should care about whether the benzene is dissolved into the water or into the DMSO? And if I put the cap on the vial instantaneously after putting the last component into the vial, why should I have to wait 5 or 25 seconds before I do so? Do you see my point?

There are good reasons sometimes for following a procedure EXACTLY as written, in the order that it is written, but we are supposed to be educated chemists and use good practical scientific knowledge in what and how we make our samples.

I used to put 5 solvents into a small pre-weighed volumetric flask by weight, say: 75.03 mg, 99.10 mg, 34.76 mg, 222.12 mg, and 45.55 mg into 23mL of DMF. Then add enough DMF to prepare exactly 25mL of the 5 solvent DMF solution. Then by weighing the flask I would determine the density of the solution. When I transferred a portion from this flask I would weigh the transfer and know VERY ACCURATELY the volume I transferred and from that value, how much of each solvent I transferred as well. My transfer error was not 1% but less than 0.001%.

When operating in a reasonable range of concentrations and temperatures, headspace is more accurate than the best chemist. My samples always told me how good I was that day in making my preparations.

When I taught a new chemist how to make HS preparations, he was usually better than I was on the very first attempt, getting EXTREMELY good values because he took his time and did VERY careful work because he was nervous in learning how to prepare them, and I was a tiny bit sloppy as I was in a hurry and only tried to get the samples made GOOD ENOUGH. So his 6 stds were less than 1% and mine were less than 2%.

Because benzene has such a HIGH partition value out of water it is a good diagnostic analyte to check how well and carefully you are preparing samples.

Good luck. I am certain success is just around the corner.

Rod

you are so kind to share your experience to me。I’m very appreciated。

"over tighten not possible to well explain my phenomenon。" phenomenon ，I mean the first vial benzene peak area several times bigger than the last，the second vial also bigger，but smaller than the first，it’s very often。I know loose may leak，so only over tighten will happen to me。It’s impossible that always first vial loose or over tighten，not happen at the other vial。It’s without a doubt that cap operation is important。I think I’m full of strength, maybe I prone to over tighten, for girl maybe prone to loose. I think hard to curb. About how to use crimping tool well，I hope get more experience from u。

I admire your spirits in study to make puzzle clear。

I see your point，and I can’t agree more。But as a QC analysist，I have no right to change a method constructed by Research Department。I will say my opinion to them as reference。

Tomorrow I will go to Research Department to discuss how to do，next step。

Thanks for ur attention，dear Rod。
Best wishes。
Dogcatlike

I may have missed it, but have you tried waiting to prepare each sample until just before it goes into the oven?

If the issue is due to leaking while sitting in the tray before going into the oven, then this will confirm or disprove it.

I may have missed it, but have you tried waiting to prepare each sample until just before it goes into the oven?

If the issue is due to leaking while sitting in the tray before going into the oven, then this will confirm or disprove it.

i'm afraid that the benzene 0.1ppm water solution not steady，concentration vary by time，even if store cautiously as possible as you can。so i never try as u say。
but one time，i put the six capped vial in freezer（2-8degree） ahead，then when run to each vial position，i put the corresponding vial on the tray。still bad rsd。
i have try two different kind of septa and aluminum cap，no difference。i have realized the importance of crimper operation，and i'm cautious all the time。

I may have missed it, but have you tried waiting to prepare each sample until just before it goes into the oven?

If the issue is due to leaking while sitting in the tray before going into the oven, then this will confirm or disprove it.

i'm afraid that the benzene 0.1ppm water solution not steady，concentration vary by time，even if store cautiously as possible as you can。so i never try as u say。
but one time，i put the six capped vial in freezer（2-8degree） ahead，then when run to each vial position，i put the corresponding vial on the tray。still bad rsd。
i have try two different kind of septa and aluminum cap，no difference。i have realized the importance of crimper operation，and i'm cautious all the time。

I think, except water and benzene, you have other components. Benzene almost insoluble in water. Other components, if present, affect benzene peak response. Prepare solution of water and benzene only and you should improve reproducibility

Two points:

You are right in that other non-aqueous components can affect the partition of benzene from an aqueous matrix.

One should NEVER try to make a solution of benzene in water alone.

WHY?

As you, chemist10, have stated before, benzene has a very low solubility in water, and even if you keep the water cold enough so the benzene does not flee immediately from the aqueous phase it is nearly impossible to sample such solution where you get two consistent concentrations of benzene. Time and warming the temperature of the benzene solution so that it reaches room temperature allows the benzene to escape from the water phase so that its concentration in water approaches zero. How can one sample consistently a moving target?

Reproducibility will suffer. Hmmm Have you heard any complaints about this problem on the forum?

dogcatlike,

does this ring a bell for you?

best wishes,

Rod

Dear Rod,

I disagree with you, that Benzene does not work in water. I think, at low concentration, water is the best solvent for Benzene for GC HS analisis.Using crasy mixure of water and DMSO or water and DMF as a solvent creates tons of problems. But it is just very hard to prove on the forum.

Mark

Oh, it CAN work, but special handling is necessary.

For some it will not work.

One can demonstrate a technique proving that chemist or lab can make it work. I have done this.

BUT

there are some who write the forum saying they have problems with it.

I believe them. But it is hard to teach handling techniques through an internet forum.

And I think we agree on this, Mark.

Rod

Headspace method for Benzene determination works in water very well without any special care. I developed and validated few methods by myself. Reproducibility is within 2-3%. Benzene dissolves in water good enough to prepare 2 ppm and, of course, 0.1 ppm solutions. Check Merck index for solubility.
But you have right to believe to some one who sad it does not.
And yes, USP method for Benzene will not work, because of presence of DMSO in the system. It will screw up linearity, recovery and may be reproducibility.

I have had success with Restek Residual Solvent Standards(in DMSO). They are made at a concentration level where they equal 1/20 the USP limit when diluted 1 to 100 for Class 2 solvents or 1 to 100000 for Class 1 solvents. You must add the 9 mL of DMSO to the initial Class 1 dilution.