hs-gc benzene

[erdno]

Dear all!

A have a 6890 agilent GC with FID detector, and it is connected to a 7694 HS.
I have to measure benzene from water and detection limit should be under 0.25 ug/l.

We are using N2 as carrier gas and DB WAX column ( 30*530*1)
The parameters are:
HS:
oven temp: 70°
transfer line: 110°
oven equ. : 25 min
shake: low
Vial pressure: 20 psi

GC parameters:
Injector temp: 250°
Oven temp: 35°for 3 min, then 4°/min to 90, 45°/min to 220 and hold for 1 min
FID temp: 300°, H2: 30 ml, Air: 450ml, Makeup (N2) 25 ml

I use 20ml vials, with 10 ml water, about 4g NaCl, with internal std.

As i know, i should use split injection, but i heard spitless can work too. A i have problem with detection limit, i havent been able to reach it yet. I am not sure, if i am using correct flows, from HS, and from GC. Since i am using N2, i need low linear velocity.
Can anybody help me please with it?

Thanks for advance!
Norbi

[Peter Apps]

Welcome to the forum.

Start with some simple arithmetic.

The detection limit you want is 250 ng/l

That is 2.5 ng in your 10 ml sample.

Assume for the sake of argument that all of the benzene migrates to the HS.

You then have 2.5 ng in the headspace at 20 psi (I am assuming that this is 20 psi above atmospheric i.e. 35 psi absolute).

Loop fill time and loop volume are operating parameters that you do not give, I assume that you do the default of filling a 1 ml loop at atmospheric pressure = 15 psi.

Of the 2.5 ng in the 10 ml headspace you get 2.5 / 10 x 15/35 = 0.11 ng.

So it is no surprise that you cannot reach the LOD you need. You can see 0.11 ng with an FID as long as the peak is sharp, but you cannot do a split injection, and the assumption of all the benzene going into the headspace is urealistically optimistic.

250 ng/l benzene in water by equilibrium HS is going to be a major challenge. Have you thought about SPME or purge and trap ?

Peter

[erdno]

Dear Peter!

I am using 1ml loop, at the moment with 0,4 min fill time. I heard much lower fill time can help to increase concentrations, but i havent tryed it yet.

Can u explain this a little please? "Of the 2.5 ng in the 10 ml headspace you get 2.5 / 10 x 15/35 = 0.11 ng."

SPME and purge and trap would be a solution, but at the moment we cant afford it.

Norbi

[dblux_]

...
I have to measure benzene from water and detection limit should be under 0.25 ug/l.
...

Not with static HS.

[Peter Apps]

Dear Peter!

I am using 1ml loop, at the moment with 0,4 min fill time. I heard much lower fill time can help to increase concentrations, but i havent tryed it yet.

Can u explain this a little please? "Of the 2.5 ng in the 10 ml headspace you get 2.5 / 10 x 15/35 = 0.11 ng."

SPME and purge and trap would be a solution, but at the moment we cant afford it.

Norbi

1 ml is one tenth of 10 ml, so of the 2.5 ng you get 0.25 ng.

But your 10 ml is at 35 psi and the 1 ml loop is at 15 psi. 10 ml at 35 psi is 23.3 ml at 15 psi. So your 2.5 ng in the headspace is in 23.3 ml at 15 psi. 2.5/23.3 = 0.11.

The absolute maximum improvement you can expect by reducing the fill time is to get 0.25 ng into your loop, but in practice you will never achieve that.

The only practical thing that is likely to help is to increase the sample volume to 18 ml so that the benzene is in a smaller gas volume.

Equilibrium HS is the wrong method for what you need to do, I'm afraid.

Peter

[erdno]

Thanks for helping!

Norbi

[AICMM]

erdno,

Three comments. 1) You might consider a different detector, one more sensitive to benzene. An AID or PID would get you lower but how much lower I am not sure without some work on my part. 2) The comment about increasing water volume is a good one. Restek has a quick technical guide on headspace that is very informative, go to their website and search headspace and it will show up under resources. 3) Manual SPMe would be very inexpensive place to start.

Best regards,

AICMM

[rb6banjo]

You should have success with the SPME option. I have a GC set up that I could easily try it. I am able to see 0.22 ppbv (ng/mL) of benzene in water (no salt, just heated headspace at 60 °C).

http://s1285.photobucket.com/user/rb6ba ... 0.jpg.html

I used a carboxen/pdms/dvb-2cm fiber. 30 minutes of sampling the headspace after 15 minute preheat at 60 °C. The injection was 2:1 split! I did not try to optimize the separation. The bad thing about SPME is that it is generally so sensitive that your blanks are not "clean".

Good luck.

All instrument conditions:

GC: HP 5890 Series II
Column: Zebron-Wax (Phenomenex), 30 m x 0.53 mm x 1.0 µm
Inlet: 230 °C, 2:1 split, helium carrier gas at 2.5 psig
Oven: 40 °C for 2 min., 40-230 at 10 °C/min., 230 °C for 14 min.
Det.: FID at 230 °C, 30:300:25 H2:Air:Helium makeup

[MSCHemist]

Even with a Mass spec and SIM I doubt you can reach 250ppb with static headspace. This seems like a classic environmental Purge and trap method or SPME might do. Check EPA methods.

[erdno]

Thank u for helping me.

I have 1 more question about this.

Static headspace with gc and still benzene.

Can somebody verify it, if i calculate it good?

HS oven temp: 70 ° so k=1,71
If i would use 18 ml water, then the headspace is 4ml (22 ml vials).
The concencration in the sample is 0,25 ng/l

So with this, the concentraion in the headspace after equilibrium (25 min) should be: 0,25/(1,71+4/18)=0,1293 ng/ml

This means with 4ml headspace i have 4*0,1293=0,5175 ng in headspace.

15 PSI in loop and 5+15 PSI (i hope, with 18 ml/22 ml 5 PSI extra pressure should be enough ) in headspace with 1 ml loop volume and 70 ° ove temp and 85 ° loop temp
I can inject: 1ml *15 PSI/20 PSI* (70°+273°)/(85°+273°)=0,7186ml

So with this, i can reach: 0,7186ml*0,1293 ng/ml= 9,23*10 ^-2 ng

How can i calculate with inorganic salt effect? I am using NaCl at the moment, which means about *4 concentration.

Thx for helping me!
Norbi

[Peter Apps]

Hi Norbi

It's late on Friday afternoon, and you start off with 0.25 ng/l when you mean 0.25 ng/ml.

Salt does not have much effect on non-polar things like benzene, I would leave it out.

No matter what the calculations say, you need to see a peak on a chromatogram. Next step is to fire up the instrument and give it a try. Be sure to tell us if it works.

Peter

[rb6banjo]

I think your biggest issue is that 18 mL might be too much sample. You don't want liquid to get into your headspace unit. Please read up on how deep the needle goes into the vial before you select your maximum sample size.

[mckrause]

Not trying to be contrary, but salt has a HUGE effect in static HS. We add 10% (1 mL) of saturated salt solution for every 10 mL of sample.

You need to equilibrate at approximately 70 C. We have found that to be the optimum for benzene.

You MUST use split injection for HS work. Splitless will not work. We split approximately 15:1.

We can see 0.10 ug/L benzene in SIM mode using HS-GC-MS. You will not come anywhere close to that using FID.

[rb6banjo]

I'll agree with you that salt can have a big effect but my experience has been like what Peter said. For nonpolar materials, not so much. I've seen much larger effects on things that are very water soluble. You're not going to get a "leaps and bounds" better detection limit for benzene just because you add salt.

[RogerBardsley]

Hi All,
I typically read these request, but because I am an Applications Chemist for Teledyne Tekmar, I typically reply only if it involves our HT3 or Versa headspace instrument. (I wish I was young enough to figure out how to put smilies in here) I am taking this opportunity to point out to a reply by MSCHemist on Jan 13 that I was able to detect 50 ppt of benzene in water with GC/MS in SIM with the Versa Static Headspace Vial Sampler. The app note is on our website.

Best Regards
Roger Bardsley
Applications Chemist
Teledyne Tekmar