Help: Splitting Flows Between MS and ECD

[rubyfan]

Periodically, I am required to do dual ECD/MS detection using our Varian 3800 system. Since we don't do this routinely, I simply use a Y splitter from Restek. Here are my questions/concerns.

Since 90% of my work is standard GC-MS, I have my system setup with a press tight connector that joins the analytical column (15 meter, 0.25mm ID VF-5MS) to a guard column (~0.75 meter, 0.25mm ID) inserted inside the mass spec. When I need to do dual ECD/MS testing, I install a Y splitter. Leading to the ECD is a short section of 0.25mm ID column, and from the other branch of the Y splitter I install a short section of 0.10mm ID guard column that I connect via presstight to the guard column inserted into the MS. I know this sounds rather shady, but the system seems to work ok. What I have noticed is that the peaks may be off by as much as 2-5 seconds, however.

I have noticed that SGE sells a splitter system, but a fellow chemist has tried using this repeatedly with our Agilent GC-MS and it seems the splitter system never works well. I prefer to keep things simpler, that's why I've used the Y splitter.

What I'd like to know is: what are suitable lengths & diameters of guard columns to use to make a Y splitter divert pretty close to 50/50? After reading SGE's website, I see they use 0.15mm ID guard column to the mass spec and 0.42mm ID guard column to the FID/ECD. Does anyone know what the lengths of these columns are? Or does anyone have another combination of lengths/diameters that seems to work? I would be happy to permanently remove the 0.25mm ID guard column leading to the MS and replace it with 0.15 or 0.18 mm ID column.

Any thoughts would be greatly appreciated. Thanks!!

-Aaron

[che313]

well this is an engineering isue rather than chemistry.

when u use Y spiliteer u'll not get 50/50 split unless your output diameter are same. the length have no any effect.

[Ron]

A 2 to 5 second retention time difference with a split setup is good. It is next to impossible to match retention times exactly, unless you are willing to spend days trimming columns and testing to get a retention time match closer than that. If you can demonstrate that the shift is constant, there shouldn't be a problem with peak identification.