by
lmh » Thu Dec 17, 2015 1:32 pm
They're so different. It's hard to compare even the purchase cost, because the price of either a basic GC or a basic LC can vary nearly 2-fold depending on how very basic it is. The LC person has to decide whether they want UPLC, which will make the purchase cost higher, but it's still "basic" in that it won't do anything a normal HPLC can't do - it just does it faster - and that means that although the purchase cost is higher, the cost per sample can be quite a bit lower. You can't write UPLC off as an expensive luxury. The LC person also has to decide what detectors are necessary. Meanwhile the GC person needs to think about what inlet(s) to order. LC and GC people both need to decide what detectors are necessary, although FID for GC and PDA for LC are the default options, and both very versatile. In effect, saying that GC is cheaper than LC is like saying women are shorter than men. In general, yes, but the standard deviations are so high that there are sufficient exceptions for it to be unsafe to assume it's true when making a decision.
Also, the unreliabilities are different. I'd agree that a GC probably has a longer time between failures, but the most frequent failures of LC systems are often fairly trivial failures, like check-valve problems, which can be solved by a lab technician fairly easily. GC failures, although rare, tend to be fairly catastrophic events that need an engineer visit.
As you know, also, the methods are going to be radically different, and this affects the cost of use. There are so many small molecules that can be run directly by LC, but require derivatisation for GC. The cost of analysis therefore has to take into consideration the technician time (or automated equipment) necessary to carry out derivatisation, as well as the cost of derivatising reagents (and the cost of re-runs when it all goes wrong).
My feeling is that it's impossible to compare the cost of ownership unless you mean it literally: the cost of having an instrument sitting on a bench (and this is rather meaningless). It's better to compare the cost of analysing things, and there I have to agree with everyone so far: the important question is what things you need to analyse.