This is sort of a continuation of "Hydrophobic collapse" from Dec. 21. A more thorough lit. search uncovered the sought after refs + more.
Doyle. et al, J Chrom A, 877, 25(2000) did NOT see a phase collapse in C-18 columns using Raman. They list sundry refs, including a review on NMR: Sentell, J Chrom A, 656, 231(1993). Reading some of this material I am even more puzzled regarding phase collapse. Why should there be a strong barrier (kinetic) toward reaching the equilibrium (thermodynamic) in going from an all aueous phase to one with water + organic? Uwe Neue already confirmed that there is no practical barrier in going from 100% water to 100% organic.
Again: we never saw anything attributable to phase collapse, even though we should have, according to many statements. (A particular interesting one, where phase collapse appears apparently only after 10 min of stop flow, is found in www.iscpubs.com/pubs/glevy/aln1198.html, based on work from Przybyciel and Satangelo). What gives?

It is good to continue this discussion. If you want to, I can send you a few slides on our results (you need to tell me if you can read Freelance or Powerpoint). The phenomenon that we describe depends on the type of stationary phase that you are using. It does not happen on phases full of silanols or phases with embedded polar group. It does happen with highly coated C18's with very good endcapping and deactivation of the silanols.
I had discussed our results with Przybyciel at some point. He either did not believe me or he thought that this is an interesting phenomenon. I know that he repeated some of our results, including the ones that demonstrated that this does not happen with phases with an incorporated polar group.
Our experience is as follows: Several people in our labs carried out related experiments. They consisted of taking a column under flow from a functional state with a reasonable organic content in the mobile phase to 100% water (usually with a buffer), equilibrating the column with the 100% aqueous mobile phase, and running chromatograms under these conditions to make sure that one can get reproducible chromatography. Then the flow was stopped. As little as a minute was enough. Then the flow was started and the sample was reinjected. The retention had dropped to nearly nothing from very high (I need to look up the slides myself to give you accurate data). The problem was eliminated by rewetting the column with an organic solvent. 1 to 2 mL were enough.
Related to this is the use of the ancient radially compressed columns (although by no means the same phenomenon). These were C18 columns with a high silanol content shipped dry. In order to get the air out of the pores, one needed to wet the packing with an organic solvent, or at least with something containing a lot of organic solvent. Once wetted, this packing remained solvated. The high silanol content prevented the "hydrophobic collapse" phenomenon that we are discussing here.
There is also a publication by Engelhardt somewhere describing the shift in the retention time of unretained peaks as one goes from a high organic content of the mobile phase to 100% water and then back to a high organic content. The phenomenon was a kind of hysteresis.
These three phenomena may be closely related, but they are by no means the same. The one with the spontaneous change in retention after releasing the pressure from the column is the one that I refer to when I talk about hydrophobic collapse.

Thanks Uwe.
Some things have to be checkd here before contacting you.
It could be that it has not been seen in this laboratory, because our columns usually have to be equilibratd for long times (>30 min) on even minor solvent changes (to get rid of "dirt"). Have you, or anybody else, seen the following: Chromatography runs nicely with, lets say 20% org.; then 100% water is run on the column which brings about a "collapse" of some kind; subsequently, the 20% org. + 80% aqu. is run again now with a nearly permannt extremly low retention time of analyte? (No intervening 100% org.)
That is, can a histeresis exist which has a very important PRACTICAL consequence?

Greetings
Hans

Yes, a hysteresis can exist. The column needs to be "rewetted" with a certain amount of organic solvent to eliminate this phenomenon. We never explored in detail, how much organic you need. We just rewetted the phase with a very high organic content, or even 100% organic (not much is needed, not even a column volume). On the other hand, I am not surprised that 20% organic is not enough. Engelhardt's hysteresis experiments required a higher organic content to "rewet" the phase. We have done the collapse experiments with small amounts of organic (less than 5%) or no organic. The other thing that one can do is take a dry C18 and put it on top of water or a mixture of water and organic. You need a large amount of organic to make the stuff sink. Some people have played with this as a measure of the "hydrophobicity" of a packing.

Dear Uwe and Mueller:
Thank you very much for your help!

Hi!
I'm actually fighting with a problem witch seems to me as a pase collaps phenomenon (using a Waters Novapak column and a gradient beginning with 97% water). I' ve heart, that phase collaps could be avoided by taking an acid pH??? (a column dealer, not Waters, said so) What do you think about this?
Thank you for your opinion!

Hmm, I think the column dealer wants you to strip the coating of the packing. This will work and prevent phase collapse, but you will never be able to get another column quite like this one...
OK, seriously: In a gradient, phase collapse is not very likely, especially when you are running routine gradients. People often recommend to run a blank gradient or two before you start running your routine analysis. This helps in the reproducibility of retention times. However, this phenomenon has little to do with phase collapse, just with the fact that in a gradient run you are never in equilibrium.