Index Home About Blog
From: "." <dose@worldnet.att.net>
Newsgroups: sci.chem.analytical
Subject: Re: gc water injection
Date: Thu, 11 Jul 1996 21:52:41 -0400

Bruce Hamilton wrote:
> 
> "." <dose@worldnet.att.net> wrote:
> 
> >By FAR the best way to do this is with purge and trap sampling.  It is
> >possible that you may be able just to use a static headspace sampler
> >with a more specific detector like electron capture, although EC is not
> >terrifically sensitive to methylene chloride.  You also might want to
> >check into the new microextraction methods pushed by Supelco.
>
> I assume that they only have that system available, and while less than
> ideal, it will work for MeCl2 in water at ppm levels.
> 
> >If your rinsewater has soap, dirt, etc. in it, I wouldn't direct inject
> >it into my GC columns on a dare.  Sorry, but you'll ruin your column in
> >a hurry.
> 
> Actually, I've injected far worse into mine, and recently finally managed
> to inadvertently kill a 4 year old polar column - I injected the polar folch
> fraction from bovine brains straight into the splitter :-) - after about 100
> injections the column lost resolution - given that the yellow muck just
> about made everybody vomit - I was surprised it lasted so long. That
> column would have seen well over 5000 injections, so it didn't owe
> me anything. My experience has always been that time spent
> deactivating injection ports and setting up split injections can result
> in runs of upto 500 samples without cleaning the liner.
> 
> The trick is all in the splitter - control the temperature *and* use a
> liner that will hold up insoluble material. I'm not a fan of straight-thru
> liners for dubious samples, I use the old HP design of a small amount
> of 10% OV1 packing with silanised wool. I've also used fused silica
> versions of the same liner will success. What you need to do is set
> the temperature sufficiently high to volatilise all the water and MeCl2,
> ( around 200-250C ) with a split ratio as low as possible without the
> water extinguishing the flame - about 30:1 for a hydrogen carrier
> and 5ul injection.
> 
> Then just ensure that you keep injecting standards until you get
> a consistent and linear response - autosamplers help. Most salts
> will not interfere as they accumulate at such low tempertures. You
> should get several hundred injections without trauma, provided the
> injector is kept inert by not allowing air ingress while above ambient.
> 
>        Bruce Hamilton

*********************************************************************

Well, we've all done it, but I just don't see why anyone would inject 
soap, grease, and other junk into a clean column if it were not really 
necessary.  Sometimes it is, but purge-and-trap of 1-5 ppm MeCl is a 
cakewalk. 

For the original enquirer:

If you have just a few samples, call some contract labs and let them 
sweat it out. Single analyte prices are not always too bad.

If you do it yourself without purge-and-trap, consider a C-18
solid-phase extraction to clean up your sample.  Your GC will love you.
Run perhaps 10 mL of sample through; don't let the cartridge dry out!
Elute with 2-5 mL suitable solvent, perhaps methanol if you're going to
a nonpolar capillary GC column.  (Make up test samples and run all this 
in order to make sure you're getting near 100% yield.)  Inject as many 
microliters as you need to.

If I just HAD to do direct injections, here's how I'd try it: A 
relatively cool inlet (ca. 150-180C), cool the column as much you can 
(most labs can get 35C without oven cryogenics), use a well-packed inlet 
to keep as much junk out of my column as possible, a 1 microliter 
injection with perhaps a 10:1 to 25:1 split, and a hot FID.  The flame 
should stay on (ours does on a HP 5890 II), and you should be able to 
see methylene chloride with a FID of recent vintage.  Change the inlet 
liner at the end of each day.  This will about maximize your odds of 
success (though I still hold my nose).

Hope this helps.  Good luck.



Newsgroups: sci.chem.analytical
From: B.Hamilton@irl.cri.nz (Bruce Hamilton)
Subject: Re: gc water injection
Date: Sat, 13 Jul 1996 19:39:27 GMT

"." <dose@worldnet.att.net> wrote:

>Bruce Hamilton wrote:
>> >"." <dose@worldnet.att.net> wrote:
>> >By FAR the best way to do this is with purge and trap sampling.
...
>> I assume that they only have that system available, and while less than
>> ideal, it will work for MeCl2 in water at ppm levels.
...
>Well, we've all done it, but I just don't see why anyone would inject 
>soap, grease, and other junk into a clean column if it were not really 
>necessary.  Sometimes it is, but purge-and-trap of 1-5 ppm MeCl is a 
>cakewalk. 

As I noted above, maybe the laboratory doesn't have purge and trap,
I just assumed the original poster would have used/mentioned it if 
they had one available. That's why I noted that it can be done using
what they have - it may not be an EPA-approved method, but it
has worked for us in the past.

Using a GC injection splitter correctly means that very little of the 
sample gets to the column, and long runs with an autosampler
are possible.

>If you do it yourself without purge-and-trap, consider a C-18 
>solid-phase extraction to clean up your sample.  Your GC will love you. 
>Run perhaps 10 mL of sample through; don't let the cartridge dry out!  
>Elute with 2-5 mL suitable solvent, perhaps methanol if you're going to 
>a nonpolar capillary GC column.  (Make up test samples and run all this 
>in order to make sure you're getting near 100% yield.)  Inject as many 
>microliters as you need to.

Nowhere do you mention what internal standard you would use during
this process to measure losses, indicating that perhaps you haven't 
actually tried this with MeCl2. Sounds like a *very* dubious procedure 
to me just to remove a few non-volatile contaminants and achieve a
3-4 fold concentration, and even more dubiouss after considering the 
various partition coefficients, volatilities, and solubilities of MeCl2, water,
C18, soaps, and grease.  

When we do any sort of sample cleanup for GC, we add an internal 
standard before the cleanup, to check for losses. When performing direct
injections, we would use the normal practice of spiking some
representative samples to check the slope of the external standard
calibration, and if different, use both internal standard and standard
additions, but it depends on the analyte and contaminants . 

Our experience ( using both ECD and FID on aqueous, alkaline samples 
from scrubbers of flue gases from the catalytic destruction of
1,1,1-Trichloroethane ), has been that CHCl3, Vinylidene Chloride, MeCl2, 
and TCE in aqueous samples can be quickly, quantitatively measured 
by direct injection down to ppm levels without the use of either internal
standard or spiking, but obviously the original poster would confirm
the calibration and use the most appropriate calibration technique 
once they had established adequate sensitivity was available. I even
set up an on-line GC monitoring system on the rig using a sampling 
valve and a Pye 104 GC with 2 meter packed column to allow them
measure the 1,1,1-TCE - an unexpected decomposition product ( for
the catalyst they used ) was chlorine - and I had to change the system 
to PTFE valve and lines. Simple, scrap GCs ( like the Pye 104 ) are ideal 
for such work, when customers don't want detailed analyses and yet
monitor an experiment.  
 
[ Rest deleted, mostly similar to what I suggested, except for the
suggestion to change the packed inlet liner daily. If you have to
do that, then you conditions aren't set correctly. It's not economic
to clean, deacticate, and repack an injection liner daily for low
cost routine analyses such as MeCl2- something should be changed. ]

           Bruce Hamilton         


Newsgroups: sci.chem.analytical
From: B.Hamilton@irl.cri.nz (Bruce Hamilton)
Subject: Re: gc water injection
Date: Wed, 17 Jul 1996 10:39:04 LOCAL

In article <TADEUSZ.42.31E41EE7@CHEMISTRY.watstar.uwaterloo.ca>
 TADEUSZ@CHEMISTRY.watstar.uwaterloo.ca (Tadeusz Gorecki) writes:

>I am talking about injecting 2 microliters of water on-
>column to get good enough sensitivity. This produces about 2 mL of water 
>vapour at oven temperature. I can assure you, that under such conditions the 
>flame WILL be extinguished.

Get another GC :-). I routinely inject 2-5ul of aqueous samples,
with split ratios down to 30-40:1 without the flame going out
on both HP 5890 and HP5792 systems without the flame 
going out - using H2 carrier, but that makes little difference.
If your flame goes out, you need to review your FID gas flows,
particularly the fuel and makeup gases. 
  
Nobody is saying it is the preferred method, but it works down
to ppm levels for me, and it should for you.

        Bruce Hamilton



Newsgroups: sci.chem.analytical
From: B.Hamilton@irl.cri.nz (Bruce Hamilton)
Subject: Re: Hydrogen As GC Carrier Gas
Date: Sat, 19 Apr 1997 14:56:06 LOCAL

In article <01bc4c4e$738b4b80$eaa83dcf@glyph>
 "Lesley Woodhouse" <glyph@csolve.net> writes:

>Hydrogen is a superior carrier gas for GC analysis but there is a risk for
>explosion.  Newer HP models (5890/6890) have the capability of detecting
>leaks and will subsequently shutdown the instrument if a leak is detected. 
>However, I still do not recommend its use unless you require it for
>performance.

Neither of my 5890s have sensors. I've used hydrogen carrier on all
of my capillary GCs for over 12 years without an incident. About a
decade ago, Walt Jennings visited me, and as I was showing him 
our 0.1mm columns used for triglycerides and sterols, and the mass 
flow controllers we put upstream of our older GCs that just had pressure 
control. He said that the only hydrogen explosion that he had ever had, 
occured when he forgot to block off a 1/4" packed column injection port
fed just by a pressure regulator, and hydrogen was flooding into the GC :-).

Virtually all modern GCs have sprung loaded doors, so that if an explosion 
does occur, the over doesn't disintegrate with debris hurtling around the
laboratory. I would recommend using hydrogen all the time, unless you
have a very good reason for not using it, it's faster, provides better
resolution ( only nitrogen at 10 cm/sec beats it ), and is cheaper than 
helium.

We use industrial grade hydrogen as carrier, passed through 
indicating drierite/molecular sieve traps, followed by an oxy-trap, 
followed by a indicating oxy-trap, followed by an activated carbon
trap. Our three capillary GCs ( two with liquid autosamplers ) are 
in use nearly 24 hours/day seven days a week - only closing
for Christmas/New Year break, at which time every year or so,
we regenerate the driers and oxy-trap. Most of our columns last 4-10
years, the trick is never to allow a column or injector to be exposed to 
air at above-ambient temperatures, and to use a packed split liner to 
collect all the junk. We try for minimal sample preparation ( grinding
up meat in chloroform and methanol, 0.45um filtering and injecting
the solution for triglyceride and sterol analysis ), and the columns
will last for years. Using industrial grade gases is also a lot cheaper 
and they tend to be fairly pure ( for detectors, the gases pass through 
drierite/molecular sieve and activated carbon traps only ). It's 
worthwhile having a purge valve on the high-pressure cylinder manifold
to purge any air that gets into the hose when changing cylinders - and
we keep all lines from cylinder to GC toggle valves pressurised at all
times ( they all have pressure relief valves at around 400psi should 
the regulator leak through - bursting disks can also be used. 

The first thing I do to new GC is to mount Swagelok on/off toggle valves 
on the carrier gas lines as close as possible to the controllers ( we 
plumb both Helium and Hydrogen to GCs, as some standard methods 
still specify He, with a 1psi non-return valve on the He to ensure H2 
never gets into the He line if somebody inadvertently opened both 
toggles at once ). I ensure all lines are 1/8" copper or SS, with 2-stage
regulators with SS diaphragms, and no plastic anywhere. Using toggle 
valves means that changing septa etc. can be easily performed without 
affecting settings.  

I have no worries at all about hydrogen leaks, we have a Matheson
( actually from Gow-Mac ) leak detector that we use to check any
new plumbing, and when people are learning how to change columns.
I tend to operate at much higher linear velocities with hydrogen so
we can perform rapid analses, eg one GC analysing free fatty acids 
currently has a 10 metre column where the solvent peak is coming 
out in 5 secs, and the total analysis time is 2.5 minutes.

So, in summary, provided the GC is designed to cope with a hydrogen
explosion ( spring-loaded door or vent ), hydrogen is an excellent
carrier gas for both manual and automated unattended operation.
There is no need for additional safety devices, but if you wish you
can get sensors, and can install mass flow controllers upstream of
the GC to limit the maximum flow to about 250 ml/min if using
pressure-controlled carrier gases that doen't have a flow controller.

           Bruce Hamilton 


Newsgroups: sci.chem.analytical
From: B.Hamilton@irl.cri.nz (Bruce Hamilton)
Subject: Re: H2 as carrier gas - safety
Date: Sun, 20 Apr 1997 19:01:21 GMT

klee <cheese1@delanet.com> wrote:

>On the topic of H2 safety:
[ possible static ignition of flame deleted ]
Hydrogen has the same minimum ignition energy in air 
( 0.0017mJ ) as acetylene, and static ignitions are
very rare.

In reality, most hydrogen lines should be permanent, and
should never need to be being broken to change cylinders. We 
always mount the regulators on wall-mounted brackets, and ensure
all downstream lines are permanent. Thus the only problem
could be when cylinders are connected and disconnected, and
thats just at the cylinder connection. 

In our case, the cylinders and high pressure manifold
are outside in a concrete-block facility, and most sites
here have the cylinders and manifolds outside and well earthed. 
Matheson sell cylinder trollies for bringing cylinders into buildings 
that have static-dissipating ( conductive ) wheels to ensure 
charge does not accumulate if cylinders are moved around buildings.
Most laboratory floor coverings these days will dissipate static.  
   
>The remedy is obvious: combustable compressed gas safety awareness and
>safe laboratory practice.  This includes accessible on/off valves,
>visible pressure gauges, flow limiters (mass flow controllers or H2
>generators with limited capacity), safety glasses, gloves ....

Better gas supply system design to eliminate the risk is also 
recommended.

>I see no incremental safety concerns with using H2 as 
>a carrier gas in addition to fuel for FIDs.

There is one that I forgot to mention in my previous post.
Ensure all split vent ( and septum purge if paranoid ) lines
are metal and lead to a location where they can be safety
vented. I use a long 1/4" copper line from each instrument
to outside the instrument room. Don't manifold the lines
from different instruments.

You can put a metal sinter on the outlet if the vent line 
does not go to a restricted ( no flames ) area. But if you
use a sinter, ensure that it is cleaned regularly to prevent
backpressure. I allow the septum purge lines ( 2.5 ml/min )
to vent into the laboratory, but if you have sensors, or a
small room with poor circulation you should also vent them,
and they can be connected to the 1/4" split vent line.

I cut the split vent line on 5890s and put a 1/8" tee in it. 
I then put a toggle valve on the line going outside and a
cap on the split vent outlet on the front of the instrument. 
This means that when I want to set split flows, I just uncap
the front, add the soap bubble meter, and close the toggle.
I suppose richer people could just buy a three-way valve to
ensure people can't get the sequence wrong and completely
block the split flow, overpressuring the injector if mass
flow control is used..

     Bruce Hamilton

Index Home About Blog