Negative Peaks

[kahmark]

I recently returned to the Gas Chrom scene. My first job is to get an SRI 8610C (ECD and FID) up and running. We use our own sampling board with two Valco 8 port injectors. After changing a gas cylinder (N2 carrier), negative peaks started to occur on the ECD only with some noise at each interval on the FID. They last for 1000's of minutes, sometimes disappear, and then reappear for several thousand minutes. The interval between peaks varies between 4 and 9 minutes. Our oven is set at 61, ECD 320, FID 300. We tried to isolate the peaks, they would disappear then appear again later.

We bypassed the entire system and only checked the makeup gas (45 ml/min). Peaks still there. We then added a flow controller betweeen the tank cylinder and the makeup. The peaks went away for a 1/2 day and now appear. I can only think to change the regulator (we already changed the tank twice) or look into the ECD. We did not get the 1000MOhm resistor to test the electronics.

HELP!

Kev

[sassman]

We had this problem before twice with our ECD. Both times it went away after we had the ECD cleaned.

[Peter Apps]

ECDs have a fixed lifetime regardless of whether or not they are used. Yours might be feeling its age. You could also try a gas scrubber on the make up and carrier gasses.

Peter

[kahmark]

This ECD is about four months old. However it is out in the field in a trailer. We are sampling air for greenhouse gases, that is it.

Seems strange that the ECD is already due for a cleaning but everything points to that. Any other avenues?

Thanks for your help guys.

[Bruce Hamilton]

If your detector is dirty, I'd also want to ensure that whatever contaminated the detector isn't going to kill the next one. Maybe look at your cylinder chnaging regimes.

I'd also really want to ensure that the problem is a dirty ECD detector, as you are also seeing an FID problem ocassionally. The shape of the peak may give some indication - ensure your integration sampling frequency is high, and see if the peaks syncronise on both detectors, in which case I'd be looking at any valve switching causing electronic or pneumatic strangeness.

Please keep having fun,

Bruce Hamilton

[kahmark]

Hi Bruce,

Can you elaborate on the cylinder regime? What process do you go through when changing out your cylinders? We did change out the cylinder while continuous sampling was being done. We turned down the ECD temp, turned the ECD off, the gas was out for a minute or two.

Anyway, it is the irregularity of these negative peaks that puzzles me. They come for several thousand minutes and then disappear for many more at random. Not sure what it is......

I tracked the temp during negative peaks and saw no changes. I isolated the stream flow from the tank, through the carrier gas line, straight to the detector and still these random negative peaks occured. They will be about a minute in length and 4-9 minutes between. They usually start at a baseline rise which suggests contamination.

Frustrating.

Thanks, Kevin

[Bruce Hamilton]

I assume that you have some purifiers in your gas lines, and that the lines are always under pressure.

If not, then a new cylinder could displace accumulated wall-bound material from throughout the system, including columns, so the column exit should not be connnected to the detector until the instrument has been through a conditioning cycle after startup.

Just block the detector column inlet and condition the column with gases flowing before connecting to the detector. Consider purchasing gas purifiers if you don't have them.

My system had a single stage SS-diaphragm regulator on the wall near the cylinder, and two stage SS-diaphragm regulator just before the instrument.

The hose from cylinder to the regulator had a bleed needle valve ( caution - very high pressure. Downstream of the first regulator would be preferred these days for safety reasons ) - with line and gas purifer volumes suffiecient for several minutes of instrument running when the source gas was closed off.

The first regulator was closed off, the cylinder valve opened and closed, the bleed valve opened to clear the line, then the cylinder valve opened again and the instrument regulator valve also opened.

In your case, I would be looking closely at the peak shape. and timing, just to ensure your instrument isn't picking up electrical interference

BRuce Hamilton

[kahmark]

Bruce and others,

These negative peaks come and go over time. They show up for hundreds of minutes and then disappear for many more. The peak shape resembles either a sawtooth or a sharks tooth. Not your typical looking positive peak.

Their is no rythym. Peculiar. At times the baseline rises before the peaks begin, other times it is not so evident. My next move is to bring the GC ino the lab and run it with the makeup gas only and observe for several days.

We have an ascarite purifier in place prior to the column on the ECD only.

If anyone else has a cylinder change protocol or procedure I'd like to see it! kahmark@msu.edu Also, could anyone recommend source info on the ECD detector. We have an SRI instrument.

Thanks again for your help and keep your thoughts coming.

Kevin

[AICMM]

kahmark,

To be honest, this sounds more like electronics than chromatography. Do you have access to an external electrometer and do you know how your ECD is biased? If you have an external electrometer/picoammeter you could get your signal from this and determine if the peaks you are seeing are due to the GC electronics or not. Most ECD's run nA of current typically with low bias voltages on the order of 5-35 Volts. Most of them also run pulsed bias (constant current) but they don't have to be run that way for troubleshooting.

Using external electrometer is how I run detectors quite often so if you need more info I'll be happy to provide whatever assistance I can.

Best regards.

[kahmark]

Thank you for that reply. We do not have an external electrometer. We did take a 1000MOhm resistor and run an electronics test between the anode and cathode connectors on the board. We observed the trace over time and it was flat. No electronic noise whatsoever. This led me to believe it was not electronic in nature. But if you have more ideas along the electronic interference line let me know.

We used packed columns in this setup. Again, we tested the makeup gas only and still the neg. peaks were noticable.

I noticed today that the neg. peaks observed are decreasing over time for this particular run (7000 minutes). This run has lasted 4000 minutes so far.


Thanks and keep feeding me suggestions

Kevin

[Bruce Hamilton]

I agree with AICMM, this sounds more like an electrical, mechanical, or thermal problem, rather than chromatographic. That doesn't exclude a faulty ECD cell, unfortunately, but that's doubtful unless the environment is aggressive ( vibration, dusty, large thermal swings etc )

As well as his excellent suggestions, I'd want to ensure there was no strange change in air movement nearby, no "noisy" items on the electricaql circuit eg cycling heaters or coolers ( fridge, motors ), and no sources of vibration. You can always just point a fan at the various instrument air inlets and see if that affects the noise.

I'd look carefully at all electronic connections to ensure that instrument fan vibrations aren't causing a poor connection, which can also trigger cyclic sawtooth patterns.

Moving it to new environment should also greatly help problem solving.

Please keep having fun,

Bruce Hamilton

[kahmark]

We do have significant vibration via an air compressor and small KNF micropump. We also have a Pelonis ceramic disc heater in the trailer. I could try to disconnect the heater.

Can you explain more about vibrational effects?

The biggest issue is that it started to occur regularly at the time of the cylinder change expounded by the irregularity of the neg peaks.

Thanks again for your imput and I'll keep enjoying the ride.

Kevin

[GasMan]

ECD detectors are flow sensitive, so any change of flow in the detector will be seen. I would check your pressure regulators, it is possible that you have a diaphragm sticking.What can happen is that the diaphragm can stick in position, the pressure rises enough to 'unstick' the diaphragm and it returns to it's original position. This can repeat itself, or if you have dirt inside the regulator, it can dissappear if the dirt moves.

You should also look at the exit of the ECD. In the US you are supposed to have the exit of the ECD going to a fume exhaust, or something similar. Many years ago I was called out to a customer complaining of a noisey baseline on their GC. On my visit a few days later, there was no problem. A few days later the problem was back, and I was able to visit on the same day. The customer was on the about the tenth floor of a high rise building, and the exit of the ECD was hanging out of a window. It was a windy day, and the pressure changes on the side of the building were enough to change the flow in the detector.

Gasman

[kahmark]

Thanks Gasman.

The ECD exits back into the oven on this particular GC (SRI). It is interesting that when we lift the oven lid, the ECD changes, some due to temp changes, maybe some due to the pressure change. What if I added tubing to allow the ECD to vent outside the oven?

We changed the cylinder regulator once, it got better, then the troughs came back, then went away, etc. All with no order, only randomness.

Thanks,

[Bruce Hamilton]

If you changed the cylinder regulator and the problem changed, but then reappeared, what action may have solved the problem?. I'd want to investigate that issue closely.

Does the problem go away if you turn the gas off and then back on. That would suggest a pneumatic cause, such as suggested by Gasman.

What sort of flow control do you have for the Detctor gases. If there are single stage regulators on the detector gases on the instrument ( rather than needle valves ), one of those could be misbehaving, or affected by the ambient environment.

I would certainly want to vent the ECD into somewhere other than the oven - at least for testing, and all ECD detector vent lines should follow manufacturer's recommendation. Just ensure that you follow their recommendation for any vent line internal diameter and length as well.

Are you sure the negative peaks don't correlate with some change in ambient conditions - ensure the conditions are consistent. If there is anything that cycles, such as air conditioners, ensure the negative peaks don't correlate with ambient temp/press/humidity changes.

One common device that is often overlooked is the printer, many of which have heaters, and some of which cause effects on the electrical power circuits.

Vibration can make intermittant connections misbehave by slowly increasing electrical resistance, and then suddenly return to the original state. All sorts of peak shapes can occur, and minor thermal changes or other environmental changes can easily exacerbate the problem.

Also, if the detector is not well insulated from the oven, you could try putting a little alumina wool in the oven below the detector to ascertain if the problem is thermal from the oven.

The obvious question, belatedly asked, is has the serviceperson for instriment manufacturer offered any suggestions. If it's a problem with the instrument, they may have encountered it previously.

Bruce Hamilton