Hi,

I'm running GPC analysis on a Hewlett Packard PLGel Mixed-B column. My mobile phase is entirely acetone, and my columns are dying every three months or so. My mobile phase and sample are filtered--is the column dying simply due to acetone interaction with the packing? Or is there other evil at work here?

Suggestions appreciated,

RSS

Contact Hewlett Packard by e-mail, telephone, fax, snail-mail or through their web page. They should be able to provide information concerning the column you are using and its ability to handle acetone as mobile phase. Also, with gel columns, high pressures can cause problems. Gel column packing can be compressed if run at too high a pressure. The pressure that a gel column can sustain "MAY" be different for different eluents. It "MAY" be able to take higher pressures for aqueous eluents than for certain organic eluents. The manufacturer or distributor should be able to provide the information you require.

Hi.

Can you elaborate on what you mean by dying? PL Gel Mixed B is actually made by Polymer Labs in Amherst, MA not HP. Were your columns originally packed in acetone? I suspect that you are shocking the columns by not leaving a constant flow through them at all times, say 1 ml/minute. Acetone will evaporate and crack the polymer beads that make up the column packing if the column is stored for an extended period. At start up, you should slowly ramp the flow up to 1 ml per minute, if you go above 1 ml per minute you are likely disturbing the packing bed and the column will no longer function properly. I would suggest using a secondary pump to keep constant flow through the column when not in use, after all I would guess you have about 3 of them (pumps). I suspect you are analyzing cellulose acetate with an HP1100 and Wyatt light scattering detector, no? Try contacting the person who originally set up the instrument........

Hi all,

Thanks for your assistance--Columns are switched from the original packing solvent (toluene) to acetone when put in use, and they are ramped slowly & gently and run at a maximum of 1 mL/min. They are not constantly run. I have spoken with both companies (PL and Agilent) and have gotten little information thus far "We'll get back to you..." I will try to get a hold of the originator of the system =) Dying means my pressure looks fine and dandy but my chromatograms start losing resolution and repeatability.

Thank you!

I agree with RRC, sounds like what I call column collapse. Keeping a constant flow and pressure might help prevent the problem. Have you considered a solvent recycle system. During analysis the mobile phase is directed to waste. When not in use for prolonged time periods it is directed back to the eluent feed bottle. Cost effective and safe if the unit has built in detector control which ensures no contamination of the eluent.

The mixed B LS (make sure you order the LS as they "shed less") should last you over a year or longer with proper care. I've been in that situation before when the resolution falls apart, and I actually cut a column open and inspected the packing with electron microscopy to find that the SDVB packing material was severly cracked. At that time, I started packing the columns with toluene and storing them as opposed to storing packed in acetone. PL "experts" told me I couldn't do this but the fact is they had never tried and it worked fine (they just loved selling me $4K worth of columns every few months). I later found that keeping constant flow through the columns helped tremendously and also kept the baseline stable for LS work. I strongly reccomend seperating your system into two seperate systems, purchase the Wyatt DRI(end of year clearance ca. $16K) and another vac station (ca. $1.8K) and letting the system recycle when not performing analysis.

Crosslinked polystyrene/divinylbenzene packing materials (like that used in PLgel columns) are exceptionally stable. However, compounds that contain aromatic rings or atoms with unshared electron pairs such as N or O can be strongly retained, leading to loss of resolution, peak tailing or, in extreme cases, complete retention of the analyte. Sometimes this phenomena is not noticeable until the columns have been used awhile. Inclusion of 2% triethylamine or another electron rich modifier in the mobile phase may eliminate this problem.

You might want to try this experiment. Take an old, bad column (if you still have one around) and purge it with acetone/TEA 98/2 for about 60 - 90 minutes at 1 mL/min, then inject your sample diluted in acetone/TEA. If your lucky, you may find the original resolution (or better) has been restored.

Best Regards,

Michael Hinsberg
http://www.hinsbarlabs.com

I agree with CS and RRC, you should use a constant flow through the columns at all times. If you are doing light scattering with multiangle diode detection; as RRC assumes, this will keep your system at equillibrium also. The packing material is not as "exceptionally stable" as hinsbarlab would lead you to believe, especially in GPC applications. They are very sensitive to pressure changes and drying out, again as CS pointed out. If the cellulose acetate inference is true and you have lost resolution (cellulose acetate should have on peak, at least in THF), then you need not worry about anything being retained on the packing material so I would neglect the TEA experiments as they will not work for reviving GPC columns anyway.

Why are you using acetone? My company is one of the largest producers of cellulose acetate(s) and all of our GPC work is done in THF. Acetone will not dissolve the mono- or tri- acetates as well as high MW di-acetates and this may be plugging your column. Flush the old columns with toluene and/or THF, then try them again. Also make sure you are using a guard column that you change regularly.

Hope this helps.

:) Can you guess who chromgod works for? I can.

Chromgod- DMK gives you added information over THF or Toluene. If the samples are filtered through 0.02 glass ufiber, the other acetates don't present a problem.

Ok, who do I work for then cs?

I am curious as to why you would want to filter out the other acetates. Are you sure you are getting a true Mw profile. I don't believe you are.......

Wow--I think I've gained more knowledge in the last few postings than at the GPC training session. Thank you for your feedback; I'm ever interested in further discussion.

To answer the queries regarding methodology: the reason we're using acetone to look at the acetate MWD is because that's the solvent that acetate is spun in. I'm very interested in how acetate behaves in acetone. While I'm sure running in other solvents gives you information as well, I'm thinking that running in the same solvent the acetate is spun in would be more useful. While other acetates don't fully dissolve, if they are sufficiently solvated they will come through and show up at the exclusion volume as a prehump, visible in the LS chromatogram. Daicel did some nice published work in this area.

I'm confused about the usefulness of a guard column, since pressure buildup is not an issue. I was under the impression that guard columns are useful in GPC if you're building up particulate matter (in which case why don't you just filter), but if the packing material is actually dying...

More and more interesting. I'll follow the suggestion to constantly run solvent through my column and hope to see more longevity. Great discussion, keep it coming...

My first response was to suggest contacting the column manufacturer, distributor or supplier. One of them must have viewed this conversation and assumed I was the originator. Since I supplied my e-mail address he contacted me. I hate to say, "I told you so." but, I did suggest it. Following are the comments from the Agilent representative.

The most likely cause of your problem is contamination of your acetone with traces of water. Strong hydrogen bonding solvents such as water, alcohols and amines dissolve the column packing. This will cause broadening of the peaks due to a void appearing at the head of the column. The cure is simple;
store your acetone in sealed glass bottles and put some molesieve into the bottle. It will adsorb the water very quickly. The acetone should be filtered just before use to eliminate particles that might be shed by the molesieve. Minimize exposure to the atmosphere. I hope this helps.

Okay, so, like most males, I do like to say, "I told you so". Always go to the source. Column and Chromatograph manufacturers love to help us.

Thanks to;
Cliff Woodward, Ph.D.
Technical Support (Bio)Chemist,
Americas Support & Marketing Center,
Chemical Analysis Group
Agilent Technologies, Inc

Agreed, this is an interesting chain, but one thing has not been elucidated: why use such a touchy column? I am sure there are some silica columns with much higher stability and even better resolution.

I am familiar with the Diacel paper and the so called "prehump". We tried to correlate that to spinning performance and found it to be totally unrelated. I might add, we spin millions of tons of this stuff each year, so we are very interested in the same things you are. Also, the mono- and tri- acetates DO NOT partially solvate as you imagine. Careful rheological studies have nulified the Diacel theory on this. As proof of my arguement (and I am only trying to help here), try running the acetates in THF or Toluene. Then compare your Mw profile. You will notice something very disturbing when you compare to your acetone runs. When you figure out what it is, get back with me and I'll explain.

The comment on the water in DMK is interesting, but it doesn't hold water(couldn't resist). The packing is x-linked SDVB, which they also pack in RP columns like the Hamilton PRP's. It is resistant to attack by water and most anything else you can throw at it and has an exceptional pH range. In fact, the water that's present plays a big role in the resolution you are seeing and is the reason....... Well, do the THF experiment and I'll finish that sentence.

DO use a guard column.......

Can you tell me what dN/dC value you are using in your LS calculations. The Diacel paper is wrong......so I hope you aren't using that one.

A comment on the silica suggested...... Have you read this discusion entirely? Try again, and I hope you'll understand why silica can't be used.

Oh, the comment from Agilent is why I don't use Agilent equipment anymore and don't call their technical support with the equipment I do have. I would make sure I check the validity of the statements made by the Agilent person before I start saying "I told you so". It is doubtful that anyone is familiar with the intricacies of this work, unless they have done it. By the way, Agilent does not make their own GPC columns and most of their best people only know what they have been trained to know....

with regards

Sorry, but I have to disagree with some of the comments posted here.

We were faced with a virtually indentical problem several years ago in a GPC application. Column resolution declined dramatically after only a couple of months use running THF in PS/DVB columns. The addition of 2% TEA as a modifer not only eliminated the problem, but also restored the resolution of all the old columns we had to better than new. Can't say if this will work for RSS, but it's certainly worth a try.

I stand by my comment that PS/DVB columns are exceptionally stable. We've been using the same column set for 4 years without problems. I've abused the columns by switching solvents quickly, pressure shocking, and even replacing the frits without loss of resolution. To be fair though, we're using a 10^3 angstrom packing. Mixed bed columns are inherently less stable.

Water tends to be strongly retained on the columns and can lead to loss of resolution (we've done the experiments), although I don't believe it's due to dissolution of the packing material. If HP's comment were correct then our packing, run in THF/TEA, would have dissolved long ago.

Question for RSS. Have you opened up one of bad columns to check for a void? In my experience, void formation usually coincides with a pressure increase. Since you're not seeing any increased pressure I would guess that the packing is fine, however it's worth a look. Acetone is a low swell solvent whereas toluene is a high swell solvent so switching from one to the other could cause void formation. If it really is column collapse, you may want to special order your columns packed in acetone, many manufacturers will do this.

In my opinion, guard columns are of limited usefulness in most applications. Frankly, I never use them. I do recommend using a pre-column filter though.

Best Regards,

Michael Hinsberg
http://www.hinsbarlabs.com

I have this sinking feeling that it's going to take me 100 years before I really know anything about LC...OK. I still do not see why using a guard column is useful in GPC. So I'll probably hold off on instituting that until I understand better why I'd do it. I am very appreciative of all the feedback, so thank you. My dn/dc value is off a table of dn/dc: 0.108, not from the Daicel articles.
"We spin millions of tons of this stuff each year, so we are very interested in the same things you are." OK Chromgod, perhaps you are a deity of sorts. I won't make sacrifice any goats just yet though. I will give the THF a whirl when I have time (hopefully by next year!).
To H. Jamieson--=) I do think your "I told you so" was a bit early. If the presence of 0.23% water in the HPLC grade acetone were going to interfere with the column packing badly enough to kill it in three months...that's pretty sad. And even if that were right (I'm highly skeptical), I *did* try to contact both Agilent and PolymerLabs. Maybe I need to go back to charm school so I can get a faster response time...

OK, I read everything again and still don´t see why a silica gel column can´t be used. The above also seems to indicate that we ALL need to learn here and there. So . . . ., if one can do proteins on a silica column why not cellulose? Of course, I am talking about surface modified silica.

Let me repeat, if you don't have specific experience with the EXACT analysis, then your disagreements are without merit. "Mixed bed columns are inherently less stable", is the key point, they are much more sensitive than a non-linear column. Other than RSS, is anyone using a pl mixed B LS? Is anyone analyzing CA? I thought not. I afford knowledge based on EXACT experiences with this analysis. Suggestions I'm sure are welcome, but why not agree to not disagree with EXACT experience?

Why use a guard column? My wonderful HP1100 once started shedding material from a pump seal and ruined a new set of phenogels(4), before my technician realized there was a problem. Had she used the guard columns as I instructed, this would not have happened. So I pose the question to you; why not?

The poster that noted the cracked packing material hit the nail on the head, this is your problem and your remedy of constant flow will serve you well. Don't be confused by some of the other posts.

Charm school won't do you any good with Agilent, I'm suprised that PL didn't offer you immediate help, its most likely due to the holidays and people on vacation, they have an adequate tech staff.

Hi again all. Aren't open forums great. What a tremendous amount of info in one thread.

I have to agree, although seemingly arrogant, chromdog's points are all valid. I've also done the same analysis, the same in more ways than you know. Water in acetone has nothing to do with it. Running in acetone is much different than THF, but I do know it works and have seen data to the contrary to yours. I'm curious as to why you believe the Diacel paper is so flawed. Is it because they used only the right angle detector for their work? Or, is it because they are a competitor?

By the way chromdog, hows the weather up off Wilcox Drive?

I really don't think there's any need to get snippy--as knowledgeable as you might be, you can always learn more. At least I hope so, life would get awfully boring otherwise.

Guard column--at the risk of sounding snippy myself, if the failed seal ruined 4 full-length columns, why would the guard column have helped? Your first column should have served as a substitute guard column and saved the last three--unless you're running them in parallel, which is unlikely! And to the query of why not--because I still cannot see a clear reason as to why they're helpful unless you have a good sign to indicate when to change them out. My question is Why?

I'm also unclear what the problem is with running silica gel, I just know we have not been using that packing. I'll let you know when I find out.
And as far as TEA is concerned, it's on my ever-expanding list of things to read about--thank you.

Thanks for all suggestions...I'll keep y'all posted. =)

RSS

At the risk of sounding condescending, some may feel inadequacies in their knowledge of this material causing them to construe my comments as "snippy". Not ment to be taken in that manner. I assume from some of your comments you have not been doing GPC for any length of time, which is ok. I had assumed you had asked questions in order to learn about your specific analysis, not hear guesses. At some point one has to seperate experience from blind theory, it will help you get your work done much quicker. With that said, I'll vacate any furthur assistance with this matter after the comments below.

Please call up all column manufacturers and ask for a silica equivalent to the PL mixed B. Don't be offended if some of them laugh at you. Maybe one of the other participants can pack them for you.

My columns were ruined because I WAS running in parallel ("not likely"??? good guess!) for particle size characterization studies. Far to complex for this forum. Pressure changes are a good indicator when it is time to change a guard column. A precolumn filter may serve you equally as well (hinsbarlab comment). In my many years, I simply prefer guards. Get PL's rec. on this as well.

cs- good one, the weather is great!

OK. But just to clarify, comments like "Maybe one of the other participants can pack them for you." *are* snippy.

Thank you for your assistance, much obliged.

RSS

For what it is worth, we folks at instrument companies do not know all the answers; nor do we pretend to. Many times we have to make guesses based on amount of info supplied and we are frequently wrong on the details; i.e., Chromgod's comments seem to be based on experience with this particular assay. I tend to trust that type of advice much more than what I gave above. I also mis-stated something; hydrogen bonding solvents do not dissolve the packing; they shrink it. Neither we nor Polymer Labs. knows if this is reversible. I suspect that in many cases it is, based on experiences I had in grad school with other types of polymer based columns. A procedure to do this - don't know one. I think that a long term equilibration in the shipping solvent(toluene) is likely to be the simplest and maybe best. Silica based GPC columns can be useful but they are more likely to have 'inappropriate' interactions with many polar groups; i.e., they adsorb them to some degree. Tricky to use. There is an offering from Zorbax (PSM) which we make; but there is not much we can offer in the way of application support unless I can find one of the inventors who might have had some experience with this assay.

Well, my interest in this is presently academic, but if we all seek new solutions only by finding an equivalent column we would still be pouring our samples through sand.
It seems that there is no way around naming companies to give an example: Polymer Standards Service (Mainz) is no longer laughing after their polymer column produced almost identical peaks for a monoclonal antibody (MW = ~150 000) and its Fab fragment (MW = ~100 000), while TosoHaas´s TSKgel
SuperSW 3000 separated the two almost to baseline.
I am just trying to find out why all polymer columns, GPC or RP, that I have used in the past have flunked in regard to my analytical problems.

Hi again.

Chromgod- If you don't mind replying again, could you tell me why you disagree with Diacel? I've measured the dN/dC of CA in both acetone and THF by incremental concentration and got virtually the same results as theirs. Why do you believe 0.108, which is what they got as well, to be wrong. Incidentally, I've reversed the LS equation and assumed 100% mass recovery and came up with the same number. So where's the basis for saying that number is wrong? Not trying to argue or dismiss your statement, but I really would like to know what I did wrong. My interest is purely acedemic, as well, I have no use for the analysis now.

Also, on your comments regarding THF instead of acetone: Since the LS detector is absolute, the solvent used does not matter even if the resolution from the column is different. The only difference in the LS (RDG) equation is the Rayleigh ratio of the differing solvents. The Mw profile may look different, but the statistical averages are the same. I'd therefor suggest forgetting the THF or toluene experiments. And, oh yeah, I do this type of work (LS) everyday and have done so for the last two years.....

Ah, cs, you are much too basic in your approach. Have you studied polymer chemistry? I would suggest a refresher course, especially on polymer behavior in solution. Your work is close, but try some experiments of your own and look outside the box instead of relying on what the Wyatt folks have taught you.

Silica - You obtain better resolution because silica packs much more efficiently than SDVB. But, it is not compatible with this analysis. Agreed is the fact the the TSK SW series are excellent columns.

Is it really incompatible, or is it like RSS stated, it has not been tried? Acetone elutes immediatly after water on the SW 3000 so there is not much interaction here.
If all commercial silica columns are really incompatible then it seems to be time to develop one which is compatible.

Now another question has popped up (this is not academic to me, as it is very important to protein work): Why use static LC when dynamic LC or RI+Viscosity+LS supposedly gets out more info like Rh.....?

I'm staring at chromatograms of CA run in THF, acetone, and DMF/LiBr--separate runs, of course. And I'm sorry to say, Chromgod, that I'm still befuddled as to the THF clue. Toss me a line if you still think I'm worthy of your knowledge...

chromdork,

I had a college professor like you once, I think it was grad level polymer chemistry. Maybe if he'd have been more helpful I'd have learned something..........

Mueller,

I think the silica idea has important merit, probably hasn't been tried and there is nothing in the lit, what do I know though....

I do have to warn against using dynamic LC from vendors like Viscotek. The analysis is loaded with error and you are forced to make way too many assumptions.....

Silica has been around for a long time. It is available in multiple pore sizes from at least one vendor with a good pore volume for GPC. If this were such an easy thing to use in SEC, everybody would use it - no swelling or shrinking, no softness of the packing.... So why isn't everybody using it?

Chromgod checking in.... If you, RSS, are looking at runs in THF, DMK, and DMF/LiBr (??!) and see no difference, then you have far more problems than I can help you with.

For those insistant on silica for GPC, please make a lit review of the subject, your questions will then be answered.

I didn't say I saw no difference; I said I was befuddled as to your clue. And I undoubtedly have far more problems than you can help me with, but that's no compliment to either of us.

In the THF run data I have (the runs are not mine; I will remedy that eventually) I still have a gel fraction and a peak for cellulose acetate. The molecular weight values are certainly different, but that's not a big surprise since it's in a different solvent. Do you have any literature you might reference to refute the Daicel papers? Or anything more specific than "if you had a brain you'd realize this is all wrong"? I'm paraphrasing earlier comments. :P I'm willing to consider that some of this work aimed in the wrong direction, but not if I have no reason for considering it.

RSS

RSS – I also have an interest in looking at cellulose di-acetate with an acetone-based GPC separation. About 11 years ago, I looked at some experimental samples with this technique, and saw some interesting results related to various manufacturing conditions. I’ve been watch and reading the literature on topics related to the solubility of CA since then. Recently, I’ve decided to set the equipment up again and explore how this method might help me understand the characteristics of our polymer. My interest is in studying what is not well dissolved in the spinning solvent, and exploring what conditions we could use to make these insolubles less of a problem.

I think what Chromgod was expecting you to see was odd peaks (e.g., prehumps) in the acetone compared to THF, because of the weak nature of acetone as a solvent. The presence of odd peaks would basically invalidate any molecular weight measurements.

Chromgod – I’m quite interested in your experiences on prehump measurements and the spinability of cellulose acetate. Was any of this work published? I can understand your results in which the prehump data was unrelated to spinability, since so may things can influence spinability it is a challenging experiment perform. I’m interested in additional processing steps, such as filtration and crimping, and the influence of insolubles. Any thing you can share?

CS – It’s been snowing lately on Wilcox Drive.

Why do odd peaks invalidate molecular weight measurements? Absolute perhaps, but I'm looking for relative correlations. The prehumps are present in very tiny quantity; undetectable without the light scattering equipment. Not to mention that I still see a prehump in the THF chromatograms. Is that atypical?

This thread is still alive? I must admit a little confusion. I thought that RSS was working for a company up in the eastern section of TN, but obviously he is not. (Obvious too, there are others that are.) It amazes me that this thread has generated such profound interest on a subject I thought to be fairly obscure. Interesting that we have competitors conversing in an open forum; be careful.

I have not published on this subject as it is not worthy of publication. The whole meat of the subject lies in the way everyone is doing science today, that is without full and clear understanding of what they are doing. Isolate your "prehumps" through preparative means and have the material characterized. Only then will you most likely dismiss the correlation with spinning performance. You may want to also look in the literature thoroughly, most of the info I have shared is already there. I'm sure there are a few posters that are very familiar with a specific author that worked for a large CA producer several years back.

You can also see the pre-humps using conventional chromatography and a simple RI or UV detector. (There is another excellent hint in that previous statement).

SAW- I see an engineer has entered the thread that is familiar with processing steps.... I'll answer your statement by indicating that one must look to other methods of polymer characterization besides exculsion chromatography if insoubles are the issue. Insolubles have much to do with raw material sources and reaction conditions, as I'm sure you already know. We found we could improve little on those so we looked at improving solubility with the addition of other materials to the processing solvent(s) in small amounts to improve in the processing arena. It worked.

Good luck to all.

Greetings, Chromboy.

It's "she". And I'm a supplier. Which you would have known already if you were as clever as you think you are.
I appreciate your contributions. Still I remain skeptical of your convictions, but apparently that's simply due to my ignorance of the fundamentals.

Yours,
RSS