Coupled columns - why not?

[MaxSpecs]

There are situations when coupled columns are the most effective way to deliver adequate separation
( save proper 2D).
Think e.g. of samples with bimodal distribution of properties where common selectivity and retention is impossible to find. You can often easily top up the first column with a one of radically different selectivity and find a sweet spot for separation. There are even software approaches to help with that.
So with that why do analytical leads developing regulated methods avoid column coupling in favour of an entirely separate secondary method if a single common method cannot be found?
What is the GMP or business rationale for that?
After all this is not a proper 2D and so can be run or any standard kit really.
Robustness is of course somewhat impacted by coupling (additional variability) by is development of an entirely new, secondary method the best answer?

[tom jupille]

What is the GMP or business rationale for that?

The short answer: inertia.

The long answer: the "grandma's ham" syndrome:
I originally heard this story (parable) from a client when I was teaching an HPLC course for his group at a (now long-absorbed) pharma company in the mid-90s. This was in the context of why it was so hard to get his people to try better approaches to method development. He told it in the first-person, so I'll re-tell that way (as best I can remember):

"My wife has a recipe for 'Christmas ham' that's been in her family for generations. It involves cutting the ham in half lengthwise and roasting it in a shallow pan. Some time back, I was helping her cut the ham when I asked her 'Honey, what's the reason for going to all the trouble of cutting the ham this way?'.

'I don't know' she replied. 'I got the recipe from my mother and we've always done it that way. Tell you what, though. Mom's coming for dinner tomorrow, so why don't you ask her?'

The next day at dinner I raised the question with my mother-in-law and she replied "I really don't know. I got the recipe from my mother and we've always done it that way. We're going to the nursing home later to see her, so why don't you ask her then?'

So after dinner we all drove over to the nursing home to see my wife's grandmother, who was recuperating from a broken hip after she fell off her motorcycle. During a lull in the conversation, I slipped in the question 'Hey grandma, in the family ham recipe, why do you have to cut the ham in half in that awkward way?'

At which point she fixed me with a "you idiot" stare (I was never her favorite grandson-in law) and replied 'Well, Bill, back in the Depression when I first did that recipe we were kinda poor and the oven in the house we lived in was too small to hold the whole ham, so we had to cut it that way in order to fit.'

Sixty years later, her granddaughter was still going to the trouble of cutting the ham in half because 'We've always done it that way.' "

[MaxSpecs]

Ha! Brilliant. Thanks for your humorous reply.
Inertia is also probably a derivative of tight regulation which disfavours deviations from known patterns.

[Andy Alpert]

For what it's worth: I get an e-mail annually from the United States Pharmacopoeia, soliciting suggestions for new types of chromatography columns to list in their list of known/approved methods. They've followed through and added several categories that I recommended.

[James_Ball]

Tom,

I heard the same story except it was a beef roast and the ends were cut off because the only pan they had was too short

The classics never loose their message.

I had to couple two IC columns together before to get separation between bromate and chlorate. Neither type would quite separate them but the two in series would. As long as nobody asked I just kept the column holder closed and nobody knew the difference. There was nothing in EPA 300.1 that forbid it, so why not.

[Vlad Orlovsky]

You can easily achieve all this with mixed-mode columns. They are designed to provide 2-D approach for separation of various compounds.
Here is a simple method for chlorate and bromate ions, you can have a shorter column (50-100 mm) and get baseline separation in 2 minutes:
https://helixchrom.com/compounds/bromat ... ode-column

[lmh]

To be fair, particularly if you're using a gradient method, method development on 2 columns in series can be a bit mind-bending. Gradient methods usually benefit from the fact that the back of the peak is in a stronger solvent than the front, so it's trying to move faster, which keeps the peak tight as a sharp band. If a peak then hits the second column in conditions where it's barely binding, there is no band-tightening effect, and the second column merely represents a very large dead volume for diffusion, so if you're separating a lot of components, you can easily get into a situation where the cost of separating two of them becomes hideous peak-shape on a third. Incidentally, if you've already flattened the gradient beyond the point of common sense, in hopes of separating the two difficult components, then the band-tightening is already minimalistic anyway. I've done it, but it's not something I'm proud of, and not something of which I'd like to make a habit.

[Vlad Orlovsky]

With a proper mixed-mode column and mobile phase, you can develop an isocratic 2-D approach which is going to be much shorter that 2 columns connected.