Suggestion for Valve Box Temperature

[fergman]

Hello,

So I'm analyzing a mix that is mostly nitrogen and hydrogen, with trace levels of CO, CO2, and CH4. During my experiment I'm taking measurements starting from 100 C up to 250 C.

To ensure that I'm always injecting the same quantity of gas, I need to heat the valve box. I definitely saw improvement in keeping the peak area constant with increases in injected gas temperature when I heated the valve box from ambient (20 C) to 105 C, but I know if I go higher I can do better. Variation for 105 C was 0.0074% decrease in counts per increase in gas temperature and the TCD no trend could be determined.

However, in my method, my oven regime runs at cryo (-80 C) for 3 min then heats up to 105 C at 60 C/min and holds for 1 min. I'm doing this because my column is a gaspro, which has to run at cryo to separate hydrogen. Since it doesn't seem like the insulation between the valve box and the oven is fantastic, I'd worry that too high a valve box temperature might influence the temperature in my oven.

Am I right to think that, ideally, I would want a valve box temperature that is above my highest operating temperature, so around 275 C? And is the this the kind of thing that is basically trial-and-error to see what will work best for both keeping temperature in my loop constant without negatively influencing my oven? Also I know the higher I go, the less sample is injected...you just can't win, I guess!

I know on the % level such variations in gas injection temperature won't really matter, but on the ppm level they will, which is an important level for me.

Here's some other pertinent information:

GC: Agilent 7890A w/ TCD + FID
Method of Sample Introduction: Gas comes straight from my reactor and into a 6-port sampling valve.
Sample Loop Size: 250 ul

Any valve box temperature suggestions, or alternative solutions, would be greatly appreciated.

Many many thanks,
=tom

[GasMan]

Keep the temperature as low as possible, none of your gases will condense at room temperature. If you increase the temperaure, you will inject less gas. Remember PV=nRT. Always fill your loop with sample and allow the pressure to come down to ambient for reproducible results. It is not good to inject with sample FLOWING through the loop at time of injection.

Gasman

[AICMM]

fergman,

My two cents worth.... Keep the valve box hot enough to keep water in the gas phase. If it were me, I would set it to about 125 to 150. The heater on the valve box has it's own RTD so if you get a lot of heat leaking from the GC oven, the heater in the valve box will cycle on less in order to maintain the proper temperature. Likewise for the oven. The other thing you need to be aware of is the upper operating temperature of the valves you have installed so you stay below that value.

You will have a hard time doing low ppm with a TCD and 250 uL loop. Possible with a methanizer and FID but still difficult with that loop size I should think.

Best regards.

[fergman]

Thanks for the replies.

Gasman - PV=nRT is actually the reason why I am interested in heating the valve box As my experiment progresses, the temperature of the product gas I'm making to be analyzed is increasing, so in order for me to properly quantitate according to my calibration, I have to make sure that everything I inject is at the same temperature as the calibration gas, meaning I have to do it according to my highest experimental temperature, 250 C. And, yeah, I'm still trying to get a feel for how long it takes for my loop to be filled. And I know there's a tradeoff by having a high valve box temperature...which is why I made the "can't win" comment. I thought about cooling the gas down after the reactor, but I'd need a way to measure that temperature, a way to cool it, a way to make sure my flow doesn't change. In the future that's probably what I'll do, but I'm getting ready for a conference and have a million other things that have to get done as well, so I felt the valve box would be simplest to at least get me good calibration for % level results.

AICMM - I do happen to have a methanizer with FID, and I recall from the initial set up of the instrument (I've been doing this since August) that we switched out from a 3 ml loop to this 250 ul loop because the peaks were saturating the detector (could be wrong, but I know the tech gave some reason for it). I suppose if I have to I can run a separate experiment where I use a different sized loop to get low ppm.

I've been trying valve box temperatures of 105, 190, and 275 and seeing how my counts vary as I increase my reactor temperature from 100-300 C (I'm injecting pure CO2 each time), and as expected as valve box temperature goes up, deviation of counts goes down, with 6.7% (ambient valve temp.) to 0.58% (275 valve temp.) decrease in counts for a 200 degree reactor difference for FID and 4.6% (ambient valve temp.) to 0.97% (275 valve temp.) for TCD.

Am I fretting over this way too much? What are the errors of the detectors themselves? I just am trying to do this right so I can feel good about what I publish/present. I know I'd like to run low to ambient for a temp., but I also want to be sure I am quantitating the same number of moles each time.

Apologies in advance for the length and many many thanks.

[chromatographer1]

Sorry if my remarks aren't helpful but not knowing how much gas you can pass through your valve box can affect the value of this suggestion.

To keep your gas temperature constant in the sample loop it behoves you to put a heat exchanger before the valve inside the valve box. This can be a simple loop of tubing which moderates the sample gas temperature. Also by letting the gas equilibrate in the gas sample loop (using a sample shut-off) you minimize the sampling variation as the vent line from the valve backfeeds sample into the valve loop if the temperature is decreasing or allows the excess to escape the loop if the temperature is increasing before temperature equilibration is achieved.

You need to be aware that even approaching the temperature limit of the valve rotor material will increase the likelihood of rotor leakage which can cause greater variation in accuracy than almost anything else you might do.

I would recommend that you pick a temperature from 100 to 120°C and maximize the life of your valve rotor. Pick a temperature where even when the water content of your sample is maximized there will be no condensation at the chosen temperature. Keep it there as changing the operational temperature of sampling valves is quite damaging to repeatable and reliable operation unless you replace and properly cure the new rotor at the new temperature.

I have been a GC process analyzer and sample system engineer and I believe my suggestions are well founded.

best wishes,

Rodney George
consultant

[GasMan]

As Chromatographer1 mentions, you need to be very careful with your valve box temperature. Agilent supply two standard valves for their valve systems, 'low' temp and 'high' temp valves. The 'low' temp valves have a temperature range from room temp to 225'C. The 'high' temp valves have a range from 150'C to 350'C.

If you have taken a low temp valve to 275'C for any length of time, you will probably find that the rotor is defective. The valves are standard VALCO valves, the low temp version uses their 'E' rotor material and the high temp their 'T' rotor material.

Gasman

[thohry]

I once analyzed gas and the method fixed the valve box temp at 42 degrees C. One problem I was still irritated is that the valve box temp drifted with the oven temp which is 230 at max. So finally I increased the valve box temp to 110. The same results and a lot of time saved because my GC did not have cryo gas.

[fergman]

Another question I have is when I'm putting my sample through the sample loop which is being heated by the valve box, how long does it take for the temperature of the gas to equilibrate to the temperature of the valve box?

Basically what I'm doing is putting sample through the loop for about 30 seconds, shutting of the flow for about 30 seconds for pressure to go down to atmospheric, and then injecting. Will this room temperature gas be injected at 150 C (valve box temperature)? What about when my process gas will be 250 C and the valve box still 150 C? Are there approximate times for all this?

[chromatographer1]

It depends on the mass and the metal composition of the heat exchanger before your valve. and the diameter and composition of the sample line tubing.

If you are using 1/16" OD SS tubing with 1mm ID and your heat exchanger is 4-6 feet in length, 30 seconds should be fine I would guess. Given your maximum temperature is 250°, a temperature of 130-150° is not unreasonable.

best wishes,

Rodney George
consultant

[fergman]

Hey Rodney,

Thanks. Those are my specs.

Yeah I'm thinking about going with 150 myself because the error associated with it is on the 1% level, and at 275 some of my standard peaks are way too small.

Do you have any idea how the valve box cools itself when hotter gas is flowing through the sample loop?

[chromatographer1]

as the hotter sample gas enters the valve box it warms the heat exchanger which warms the air inside the valve box. As the temperature sensor inside the valve box is warmed the heater for the valve box turns off. When the loss of heat from the box into the ambient air cools the box temperature enough to be below the temperature setting of the controller, the box heater turns on and heats the box to the set temperature.

IN other words, the valve box is not insulated perfectly and if the temperature in the lab is below the temperature of the interior of the box, heat will flow out of the box.

best wishes,

Rodney George
consultant

wire the $100 to chromatographer1, thanks.