Problem with carryover

[Aubry]

Hello Everyone,

First off I am new to this forum as well as new to troubleshooting for our instrument. I have only been trained on how to operate the instrument, not fix any problems that arrise, so please keep that in mind. I will try to do my best to provide as much info about our setup as I can but there are certain aspects that we need to withhold to prevent the possibility of our competetors obtaining and using.

Our setup: At our lab we run natural gas samples for petroleum companies, with typical samples containing hydrocarbons from C1 to C10, as well as He, H2, O2, N2 and H2S. Our instrument is a modified Agilent 7890A. There are 4 valves added to our instrument (3 with 10 ports, 1 with 6 ports), that lead to a series of columns. When injecting a sample we flow the pressurized sample gas from a metal sample cylinder through a 1/16" sample in line. The gas first passes through a 7 micron filter that is housed in the sample in line, to prevent any solid particles from being introduced into the valve system. The gas then flows through one of the ports of the first valve, through a sample loop, out of one of the other ports and sent to another valve through a series of 1/16" lines. The sample arrives at the next valve with another sample loop and then out again to the next valve. Once the gas has passed through all the valves and sample loops, the gas then travels trough a line to a bottle of scrubber solution, where the line is submurged in the solution. When we are injecting a gas we can roughly see how much pressure is going through the system by the amount of bubbling. After roughly 10 seconds of flow, we will close of the sample cylinder valve and stop the flow of gas, wait for the bubbling to stop (which indicates that the pressure in the injection system is now at atmospheric pressure) and hit the start button on the GC. This causes the valves to switch by the actuators to the on position and the sample gas in all the sample loops are introduced on to the columns with the flow of the carrier gases.

The problem were are having is with carryover. At the start of each day the GC has been allowed to sit for roughly 14 hours. We typically run our hydrocarbon standard first thing in the morning to flush out any contaminants. At the end of that first run we always notice we have very high air contamination, with high levels of O2 and N2. Our standard has 2.00% O2 and 4.99% N2 but our results show about 1% higher for each, and as a result the values for our hydrocarbons come out lower. If we run the standard again we will get a significant decrease in O2 and N2 values which usually matches the true values much closer, but usually it is still a little too high. It usually takes about 5 runs to get our results to stabalize and get the 2 consecutive runs that match that we need to calibrate, which is more than we should allow because it eats up a large portion of our day.

We are trying to figure out why it takes so long to flush out all the air that seeps into our injection system over night but haven't had any luck. We have try letting the standard flow for longer and it still doesn't get rid of it effectively. We have tried to find a leak at every point in our injection system with a leak detector and haven't been able to find one. Also, it takes at least 2 runs to clear out all the N2 if we inject samples with high N2 (which sometimes can have N2 over 80%) so it appears that carryover, in general, seems to be the problem. Our samples and standards seem to flow very quickly through the injection system, from what we can see from the bubbling in the scrubber so I doubt its a blockage issue; any old air contamination should be expelled very quickly, I would think.

If any of you have any helpful ideas or possible solutions we would be glad to hear it. If you need any other info please ask and if we are allowed to provide it we will.

Thank you.

[chromatographer1]

There are several possibilities, more than one may be in play.

First, there may be leaks in the fittings connecting all the lines that convey the gas standard.

Second, your valves may have ports of 0.030 or 0.020". If your lines are of a larger diameter (and they probably are) then a vortex of the gas is occurring which means it may take longer to purge the system than you are doing now. In fact you yourself state that it takes 5 runs (purging of the sample through the system) before you get a proper sampling of the gas std.

Third, your rotors may be worn in your valves which may be leaking air into the sample system.

Fourth, your columns may be leaking air into the system are they are put into trap mode and back into flow mode during the operation of the valves.

I think it most likely that #2 is the cause of your problem, but in any case at night I would switch my sample line from my gas standard to feed carrier gas through your sample system instead of natural gas and continue to run without collecting data. Then in the morning purge your sample system with your gas std and see if your problem disappears.

best wishes,

Rod

[Aubry]

Thanks for your input Rod. I think you may be on to something with your second possibility. I checked the inner diameter of our tubing and compared it to the diameter of the port hole in one of our backup valves and you were correct in your assumption, the tubing is wider. When our lab supervisor did the plumbing for our valve system he selected the larger ID tubing because it was more flexable which made things much easier to put into place since the valve box is so small, and there were dozens of lines to install. Hopefully in the next few weeks we have the opportunity to shut everything down and replumb it all.

Thanks again.

[chromatographer1]

One must watch these lab supervisors. They can be troublesome.

Yet every lab seems to have one or more. They can take all the fun out of life. Carry on.

good luck,

Rod