by
lmh » Thu Jan 05, 2017 2:14 pm
Sorry if I've posted on this before; this question looks vaguely familiar from a discussion last autumn.
While there's nothing wrong with post-column splitting, I'm not sure I understand quite why you need it in this case. Any MS should cope with 200uL/min, and I can't imagine that your back pressure is currently high enough to be a problem for any LC pump.
If you are aiming to flow-split because your pump could actually run a lot faster than 200uL/min with that column, but your MS cannot cope with more than 200uL/min, then you could consider whether you can use a 2mm column. This will have 5 times the back-pressure (as it has 1/5th of the cross sectional area) so the equivalent flow for the existing method would be a mere 40uL/min. If you can pump faster than this, you should be able to speed up the method.
I say "should" because at these slow flow rates, dead volume of the system can be a serious problem. If your injection peak from the 200uL/min run is at 1 minute and you are looking at an analyte at 3 minutes for example, then at 40uL/min (keeping existing linear flow rate) the injection peak is likely to be around 5 minutes; if you speed up your run to 80uL/min then the injection peak will be at 2.5minutes, and the analyte peak probably a minute later (you've doubled the linear velocity of solvent so it should move down the column twice as fast), so although you've speeded up column flow, the overall method has taken longer!
If you must flow-split, then I'd ignore the commercial flow-splitter, and make your own, with a tee-piece as you describe, adjusting the relative flows in each of its arms by adding lengths of fine tubing. You can also put an adjustable back-pressure regulator in one of the arms (this gives you control over the split ratio without adding a lot of extra dead volume, as a long length of tubing will). Adjustable back pressure regulators are a lot cheaper than adjustable flow-splitters, and useful to have anyway.
All passive flow splitters, even the commercial ones, are merely a couple of resistances, so if you pay for an expensive commercial version, largely you are paying for a calibrated split ratio under defined conditions. When you attach it to other detectors, the split ratio will change anyway, because you are adding extra resistances to the two arms of the split. Therefore even if you do pay for the commercial one, you'll still have to measure your split ratio.
Generally changing split ratio isn't a major problem through a gradient run as the viscosity in each arm will be approximately the same even if it changes through the gradient.
