Converting 6890 liquid injection to gas sampling

[bgrasmeder]

Has anyone ever converted an Agilent 6890 from liquid injection to gas sampling? Our current 6890 is fitted with 2 liquid injectors, an autosampler, an FID and ECD detectors. It's from an old project and the system has been sitting idle for 8 or 9 years. Almost all our work is with gases so we have no need for the GC in its current configuration. Any experiences shared would be greatly appreciated.

[GasMan]

If you can give me your serial number, I can check what was installed on your GC and then advise you on what to order to update it with gas sampling valves.

Gasman

[bgrasmeder]

The serial number is US00037208. It is equipped with two 7683 series injectors and one 7683 auto sampler on the front injector.

[GasMan]

Do you know exactley what you wish to add. By this I would mean the following:
1 How many valves do you wish to install.

2. Do you want 6 or 10 port valves.

3. You have two capillary inlets on the GC. Will you be using capillary or PLOT columns with these inlets and injecting into the capillary inlets.

4. From your first post, you mention that you will only be using gases. Will these always be in the gas phase or will you be injecting liquefied gases.

5. What size samples will you wish to inject.

Gasman

[bgrasmeder]

We have another 6890 that we use for gases. That one is set up with just one 6 port valve and seems to work pretty well. For now, we'll just use the one detector so we'll just need the one valve. There is no need for backflush.

When I say gases, I mean gas blends. We may be looking for simple hydrocarbons like methane or ethane or more complex molecules, but always mixed in a balance gas like helium, hydrogen, argon, or nitrogen.

We use some capillary columns but also use packed column. We've tried PLOT columns in the past but found we didn't get as good results as the packed column. Generally, we just use an adapter when switching from packed to capillary.

Sample loops vary in size from 5 microliter up to 10 milliliters, depending on the concentration of the sample. We may look at methane down in the ppb range, but may look at other samples near percent level in concentration. In order not to overwhelm the detector but still get adequate responses for low level analyses, we use different sample loops.

Your help is greatly appreciated with this project Mr. Gasman! If there is any other information you need, just let me know.

[GasMan]

If you wish to add one 6 port valve configured as a gas sampling valve and use packed columns, you can order the following from Agilent.

1 x G1580A Valve box with automation for one valve and heated zone.

1 x 5062-9508 Six port valve with max temp of 225°C

1 x G2740A Gas valve hardware mounting kit.

This will give you the basic hardware to mount valves on the 6890. You can of course order your valve from Valco if you wish.

There is a potential problem with the carrier gas supply to your gas sample valve. At the moment you have two capillary injection ports installed, and it is not advised to use this as a carrier gas source for packed columns. A better solution would be to install a EPC module that can control flow or pressure. To do this, you will need to order the module G2317A from Agilent. You would then need to remove one of the EPC modules that supply one of your capillary inlets and also disconnect the heater/sensor for that inlet. The module G2317A would be built into the vacant position.

Hope this helps

Gasman

[AICMM]

Mr. Grasmeder,

Speaking for myself, this conversion is the type of work I do all the time. I would suggest you think about looking to a third party to make this change over unless you are very comfortable re-plumbing.modifying GC's. That being said, I suspect this is the type of job that Gasman would also be interested in so I do not want to step on his toes here.

Three things I would add to Gasman's comments. First, you might add an Aux EPC to use packed columns and by-pass the SSL injectors completely. Second, if you get Valco valves directly, you need to keep in mind how and where you would mount these. This is not as simple as the Agilent valve oven retro-fit but does have some distinct advantages, depending out how your other instruments are configured. Third, you have ECD and FID. Not what most people think of as gas detectors unless you are interested in just hydrocarbons or something like trace nitrous. Might be better to have an HID or TCD for a gas analysis instrument instead of the ECD (HID and TCD are more generic.)

If you wish to contact me directly, I can be reached at aicmm at flash.net.

Best regards,

AICMM

[bgrasmeder]

To Gasman and AICMM, I'm truely thankful for the information. I do have some limited experience with modifying, repairing and setting up GC's, but certainly not to the extent you two have - I can tell that much. I'm fortunate that one of the other labs here is doing exactly the same thing I'm attempting, but they don't seem to have a lot of information that you've provided. I will be passing along what you have provided though! I'm also fortunate that we have a GC in my lab that was set up from a system with a liquid injection port, so I do have some frames of reference. All that being said, I'm going to give this a shot and I'm looking forward to learning more about this.

With regard to the detectors on the GC, I only plan on using the FID detector and it will be used exactly for what you said it would be used for - hydrocarbon analysis. I would love to hook up a DID in place of the ECD, but that's not going to happen right now. On the converted system in my lab, they used Valco products, but it looks clunky. I will look at the Agilent parts.

If you have any other words of wisdom, I would appreciate it. I'll post my progress from time to time.

And again, thank you very much for this very valuable information.

Sincerly,

[bgrasmeder]

I have found a vendor who is willing to convert the GC from liquid injection to gas sampling. He is proposing swapping the injectors and the autosampler tray for doing the conversion. I have not spoken to him directly yet, but we have exchanged a few emails. I've done a little research and found that these pieces of equipement are going for $4,000 or more on used equipment sites. So potentially, that's $12,000 worth of equipment in exchange for doing the conversion. I doubt the conversion would normally cost this much, but in the end, we both get what we want. I get a functional GC and he gets equipment I know he can turn around and sell at a profit.

Any comments?

[larkl]

If you can get away with leaving the sample valve at ambient temperature (no heavy components), you can just cut the carrier line going to one of the inlets and install a Valco 6 port valve with a sample loop. You'll either need to add the activator assembly or just add a handle to the valve and do it manually.

If you have the money and time, valve oven, valve, actuator is the way to go.

If you need to get in and out cheaply and you can tolerate ambient temperature valving and a manual valve actuation, just add the gas valve with handle and bolt it to the side of the GC.

[bgrasmeder]

The conversion has been done. I have a working FID and a gas sampling valve. I had someone do the conversion for me by swapping him an autosampler tray and injector. However, he ran out of time before he could get the interface working between the computer and the GC. I can get the data to EZ Chrome, and the computer recognizes when I start the GC and the computer tells it to stop. But, I can not get the new gas sampling valve to work with the software. I have the pin out diagram for the serial connection, and I've tried many different combinations, but I can't seem to get it to work. Any suggestions?

Regards,

Bob Grasmeder

[larkl]

Have you tried manually adding the valve to the GC using the keypad on the GC and then manually actuating the valve, again with the keypad? If this won't work, EZC certainly won't be able to control it.

[bgrasmeder]

Yes, the GC recognizes the VALVE BOX as an auxilary (AUX) item. I can open and close the valve (ON/OFF) manually. I've tried configuring the valve as gas sampling, ON/OFF and all of the other configurations it allows me, and it still won't work. Also, I've tried the options for COMMUNICATIONS, XON-XOFF, etc. and still nothing. I am going to try and use it as is for some analyses since the software will wait until the START button is pressed on the GC. Then I can manually stop the GC after the method is done running. Fortunately, this is not a primary instrument so it is not time critical.

Thanks for getting back to me.

Regards,

Bob.