Measurement of pressure within gas sampling loops

[Justin Hyde]

I am analysing various fuel and exhaust gases with a GC with TCD detector. I have a 10 port valve with sampling loop to introduce the samples and to reverse the column order during the analysis.

My main concern is the variability of peak areas with atmospheric pressure. (The temperature of the valve and loop is kept constant and the volume of the sample loop is fixed.)

I introduce samples from pressurised sources and also by connecting a syringe of gas to the inlet line of the sample loop.

After sample introduction I usually switch off any gas flows. The pressure in the sample loop lines should equilibriate to atmospheric pressure since the outlet from the sample loop is open to atmosphere.

I take a reading from our barometer in the lab. I can calculate adjustments in the peak areas by simple ratios with the air pressure at the time the standard gas was analysed.

My questions are...
1. Is there an alternative method of taking pressure variations into account?

2. I have thought of using a pressure gauge/transducer connected to the sample loop line(s) to measure the pressure within. Is anyone doing this? If so, can you please provide some details?

Kind regards
Justin Hyde
University of Waikato
Hamilton, New Zealand

[chromatographer1]

Justin,

Working with process GC analyzers almost all use atmosphere as the reference and use standard gas mixes to calibrate, which compensates for the difference in barometric pressure.

Measurement of air pressure with a gauge is not always accurate and I have never seen it in practice.

best wishes,

Rod

[GasMan]

Justin,

The best way that I have found to get reproducable results is to fill the loop with sample, stop the flow of sample and wait 10 to 15 seconds for the pressure in the sample loop to come down to atmospheric, and then inject. Some years ago I did an experiment over a few days of analyzing natural gas and also recording the atmospheric pressure. I could see a change on the absolute areas, but in reality it only affected the high concentration (70%) compounds.

If your calculations are being normalised to 100%, then you will see no difference. The only way that I know of that will keep the pressure constant in the sample loop, is to use a backpressure regulator on the sample out line.

Gasman

[GasMan]

Justin,

The best way that I have found to get reproducable results is to fill the loop with sample, stop the flow of sample and wait 10 to 15 seconds for the pressure in the sample loop to come down to atmospheric, and then inject. Some years ago I did an experiment over a few days of analyzing natural gas and also recording the atmospheric pressure. I could see a change on the absolute areas, but in reality it only affected the high concentration (70%) compounds.

If your calculations are being normalised to 100%, then you will see no difference. The only way that I know of that will keep the pressure constant in the sample loop, is to use a backpressure regulator on the sample out line.

Gasman

[Justin Hyde]

Hi Rod and GasMan

Thanks for your advice. Let's see if I understand correctly...

A reasonable amount of time (10-15 seconds) should be left after switching off the gas flow through the sample loop. This will allow the pressure to equalise with current atmospheric pressure. (assuming that at least one end of the sample loop is open to atmosphere)

When we use atmospheric pressure as the reference pressure, we can record atmospheric pressure at the time of injection of our 'unknown' sample. Likewise for the standard gas, at the time of injection atmospheric pressure can be recorded. Calculations based on the pressure differences can be used to compute the equivalent areas.

I need not worry about variations in atmospheric pressure too much unless there are some high concentration components. (But I should/will verify this myself)

Measurement of pressure within the lines can be inaccurate/inconsistent and not normally done in GC analysis.

I could use back-pressure regulators (pressure adjustable check valves) to maintain a constant pressure in the sample loop. I guess you would need 2 regulators/check valves, one at each end of the sample loop lines. And new samples would have to be introduced at higher-than-normal pressures.

Re normalisation... May seem a bit crazy, but I haven't got my head around this completely yet. Could you give a brief explanation please?

Cheers
Justin

[Bruce Hamilton]

I've used several methods of gas sampling...

1. Atmospheric pressure - sample line outlet from loop is free to atmosphere. OK for samples that are near to atmosphere pressure, but beware of back diffusion or inadvetent cooling of the vent line if flows are too high. Use accurate barometer ( Fortin or similar ) to apply corrections. You can use syringes and pressurised samples.

2. Bubble outlet of sampling loop through a head of 20 - 50 mm of water and stop flow, that means standards and samples aren't exposed to atmosphere and are at slightly elevated pressure, or open back vent to obtain atmosphere pressure. Method is OK for samples that don't dissolve in the water ( used to use Hg, but that went out of fashion 20 years ago ). Can be used with syringes and pressurised samples.

3. Evacuate down to <1 torr the sample loop and sampling system ( with a 1 or 2 ml sealed volume on dead side of sample loop, and very small pressure transducer between sample and sample loop ), close vacuum, open sample, record pressure. You need the small volume on the dead side of the loop to drive all polar components of samples off the walls of sample loops when evacuating. Good for samples that are below or near ambient pressure in sealed flasks. The transducer volumes are very small, and some have chemically-inert membranes. Can control amount injected by loop pressure as well as size, which is good for some research applications.

4. Moving piston sampling system. These are required for simple sampling of materials where pressure drops cause phase separation or condensation. They enable you to keep the sample and loop at the same pressure. Expensive, tedious, and it's usually easier to allow the sample to separate consistently.

You said you are measuring exhaust gases. If the samples are from HC combustion, you have to maintain your sample above the dew point, and sometimes also watch for cooling effects from gas flows through narrow orifices ( valves only cracked open, rather than fully opened ). Some sampling methods use a dilution tunnel. All samples that produce condensables with temperature or pressure changes have to be processed carefully.

Note that back-pressure regulators are usually not precise or repeatable enough to be used in gas sample loops, hence the simple head of liquid technique. You usually can not use a gauge or pressure transducer as part of the sample loop, because most of them change volume with pressure. You need a really big loop to reduce that error.

Many of the standard methods and publications for gas analysis also devote quite a lot of attention to the sampling technique, so you should use what is typical for the samples, unless you have a good reason not to.

Bruce Hamilton

[GasMan]

Justin,

Normalisation is one of the calculation methods used in GC. It can ONLY be used when all of the compounds in the sample are measured and calibrated. You calculation will give you an answer as percentage.

As a matter of interest, what GC are you using, as some of the more modern units with electronic pressure control, have an ambient pressure sensor built in and you can measure the atmospheric pressure directly in the GC.

Gasman

[Justin Hyde]

Thanks for the brief explanation on normalisation.

The GC I am using was purchased in 1999 from Perkin-Elmer. It is an AutoSystem XL with electronic PPC (proportional pressure control?). The reference gas flow does not have PPC and must be adjusted manually.

Regards
Justin

[pi3832]

How accurate do you need your analysis, and how often do you calibrate?

According to the Weather Underground over the past month the barometer readings for Tauranga (closest station to Hamilton) have varied from a max of 30.73" Hg to a min of 29.29" Hg. That's a 3.75% difference, and took a week to happen. On any given day the pressure is unlikely to change more than 1%.

If you calibrate daily, the atmospheric pressure isn't going to change enough to throw your results off much, unless there's a typhoon blowing through, in which case, DUCK!

As for waiting 10 of 15 seconds, I find that it's easier to program that pause into the injection program, if you can. I have a ball flow-meter attached to the outlet of the sample loop. When the ball drops to the bottom of the flow meter, I hit the start button on the GC. The program then waits 15 seconds before firing the valve.