Subject: Chromatograph Malfunction


Dear Sir\Miss,


I am the Senior Chemist of a lab that services the oil and gas industry in Egypt. We provide Gas Compositional Analysis, Water Analysis and such. I am writing regarding an HP 5890 Series Chromatograph Gas Analyzer.

Our GC configuration is as follows:
*Injection via gas sampling valve (25
micro liters)
*Inlet is a capillary split/splitless
*Detectors are FID & TCD in series
Carrier & makeup gas is Helium (purity
99.999%)
*Moisture & O2 traps are installed on the
carrier line
*Moisture trap is installed on the
airline of the FID
*Air source is an oil less compressor
*Two capillary columns installed:
- The first (A) is SUPELCO SPB-1 fused silica 60m in length, ID 0.32mm with a max. temperature of 320 oC
- The second (B) is RT-Msieve 13x 15m in length, ID 0.32mm with a max. temperature of 300 oC (serial # 218637)

While running our GC, I faced the following problems:
1. Nitrogen (N2) can not be separated by SPB-1 column and very poor separation of CO2 particularly if of low concentration (Bitmap #1)
To over come this problem we had to install the second column (B) parallel to the
first column (A) which gave us a good separation of N2 at low retention time, but
still poor separation of CO2. (Bitmap #6)

2. Column (B) separates the gasses (He, O2, N2, CH4, CO) but the other components are separated as a big broad peak with a long tail (Bitmap #2). This tail holds some of the other gas components to be detected affecting the area accuracy.

3. Column (A) should separate all the hydrocarbon components in gas from C1 to heavy ends. I figured a bad resolution of C1 and C2. Also separation of C1, C2, C3, and C4 have big tails ending in inaccurate area leading to misleading values (Bitmap #3 )

By changing the operation parameters (temp. & flow rates) we were able to obtain
better results (Bitmap#4) but then CO2 separation completely disappeared.

4. Sometimes a symmetric sin waves or negative (-) peaks appear on the base line of the FID and TCD respectively affecting the accuracy of the peak areas specifically those components of low concentration (Bitmap# 5)

5. Our base line of both FID & TCD is relatively high @ 180 counts for FID and 80 counts for TCD

6. The separation peak area of C1 in the (B) column is about three times higher than that of the (A) column. Why is that? (Bitmap # 6)

7. As we do not require a permanent parallel connection of both columns by polyimide sealing resin(from RESTEK), we require your advice on the best way to detect any leaking that may occur in the connection points between the two columns (i.e. What kind of leaking detection solvent should be used?)


Any advice as to how to over come these problems will be highly appreciated.
Best Regards;


Mustafa El Said M.

R & F Petroleum Consultancy & Services

#1

#2 The best solution to your problem is not to run the columns strictly in parallel, but to get another gas valve. Construct a scheme to divert flow from an initial condition (column A and B in series). As soon as C1 elutes switch col A to Fid, while B continues to TCD. You do not want to put heavies onto your MS column. That might be your problem for #4.

#3 Tailing indicates a leak problem.

#4 This might be explained as strongly retained peaks eluting from your MS column over time (for the TCD). But for the FID? Perhaps an electronics problem.

#5 High baseline might also indicate a leak/intrusion of air.

#6 If the peak is wider due to the seperation characteristics of the column, the area count will be higher. This is no problem as long as the standards exhibit the same behavior.

#7 A handheld TCD is the preferred method of leak detection.

As my understanding I think your system is HP"system II"configulation that three valves,two valves and three columns for inert gas analysis and the other valve connected with PLOT used for separate Hydrocarbon only.

1.For channel hydrocarbon you need to setting the new condition,Oven = 70-150 notmax 180 because some column inside oven lower max temp.Pls.check the leak inside the valve box may be leak between the rotor valve.

2.For channel of inert gases you need to reset the timing of valve switching by determine each column separate(see in mannual HP_NGA)may be difficult.Adjusted the restrictor of valve 3 may be help better separate O2 and N2.

3.If the separate of oxigen and nitrogen still not good.the column mol.seive should be conditioning(should be out done the other oven.

Thailand,
Agilent.Petrochemical