I have been making water-soluble polymers with biomedical applications for the last 16 months and it is quite satisfying: Our macromolecules are usually well behaved – they extract into organic phase. They precipitate as a snow-white fluffy crystalline solid, on a kilo scale. They even have beautiful NMR spectra. Unfortunately, such was not the case with the frothy mixture in the picture. I had to isolate the material from a solution in concentrated HCl (0.3L), with extra sludge of inorganic salts and assorted gunk that included gram quantity of dimethyl sulfide.
The usual process would be: dilute, filtrer, dialyze. But dialysis is a slow and rather frustrating business and we don’t even have bags giant enough for removing few mols of salts and HCl. So I was delighted to learn that tangential flow filtration is a turbo-alternative to a dialysis – instead of steeping a swollen dialysis sausage bag (that can burst overnight) for days and waiting for the diffusion to run its course, the tangential flow filtration setup visibly labors for you: the pump pushes the mixture against a semi-permeable membrane, water and other small molecular weight material leak out, the macromolecular fraction stays in. The purification is done in few hours.
The peristaltic pump in the picture circulates the crude mixture at moderate pressure and high flow rate (20 psi, 1.7 L/min) from the beaker to bottom of the column; the stuff that flows out at the top is fed back into the beaker in a closed loop. The column consist of a bunch of spaghettini-like capillaries that are coated with a semipermeable membrane. The spaghettini are housed in a plastic pipe casing. It is inside these capillaries that the mixture rushes through at high speed over and over again – water and small molecule material that leaks out through the walls of the capillaries collect in the casing and flow into waste (the sidearm and the transparent bottle). One has to keep adding water into the beaker quite often because with a good column + pumping rate/pressure the mixture gets concentrated rather quickly.
The time to end the purification is when chromatography (GPC) can no longer detect small-molecular weight impurities. Of course with a whopping excess of HCl at the beginning, one doesn’t need to run GPC to confirm that all low-molecular weight material is gone – a pH paper will tell you that. (A sniff test for dimethylsulfide presence is also fast … and revolting…)
Org Prep Daily 
I love UF. works great in the lab, and scaleable too. A conductivity meter is another useful way to detect the end-point of the desalting, we use an in-line conductivity cell that measures the permeate.
Comment by Handles — October 30, 2012 @ 1:55 am
It is a clever technique, but pretty expensive as these things go. For many polymer chemists, if you have to purify your product at all, you won’t be able to make a profit on it 🙂
Comment by Joe Q. — October 30, 2012 @ 9:12 am
I agree, the large UF columns can be costly but they are reusable. The rest is just a pump and lots of water. And if you are making polymers for potential therapeutic applications, even hundred bucks per gram is reasonable.
Comment by milkshake — October 30, 2012 @ 2:32 pm
What’s the make and model of your unit? I might look into ultrafiltration for my purposes. Do you have any experience with other brands? Thanks.
Comment by Bend — October 30, 2012 @ 6:50 pm
The setup is made by SpectrumLabs.com (they make the columns too) and the model name is “KrosFlo Research II i-TFF system”
I don’t have any comparison, from the few uses so far I liked it but there were some small bugs (a leak in a tubing connector, one defective new column that leaked so we returned it, the delayed and less-than-smooth response of the the control box for transmembrane pressure – which can cause pressure overshot and pressure error shutdown if you ramp up the pressure too fast). My colleague who is using the instrument almost daily really likes the system though
I suppose before you start you want to give the column a good wash first (especially with new a column, it comes stabilized with preservatives); pumping pure water through the system will tell you if there are any leaks
Comment by milkshake — October 30, 2012 @ 7:17 pm
I can recommend Pall: http://www.pall.com/main/Biopharmaceuticals/Product.page?id=33893
The minimate cartridges are very reproducible and give predictable scale-up.
Comment by Handles — October 30, 2012 @ 7:21 pm
My lab still uses dialysis bags for most work, with pretty good results. I recently bought an Amicon stirred cell: http://www.millipore.com/catalogue/module/c3259 and the results are often fantastic. Pressure is provided by an N2 line and a stirred solution is pushed through a regenerated cellulose membrane. This setup doesn’t work well with solvents, but is amazing if you have to do buffer exchange for GPC.
Comment by Mat — October 31, 2012 @ 1:52 pm
I’m a little confused by the picture. I can see that the solution goes from the beaker with the orange solution to the pump to the bottom of the column. And there is a hose from the top of the column to a waste bottle. But where does your polymer go from the bottom of the column (where it is presumably forced to stay there due to the filters inside the column)? Does it accumulate at the bottom of the column or what? And where is the left hose in the orange solution coming from?
Comment by Ron — November 7, 2012 @ 4:02 am
there is an outlet at the very top which is partially obscured from the view, a tubing from the top feeds the stuff back to the beaker. The top and the bottom are interconneceted by the pack of semipermeable capillaries through which the mixture circulates at high speed. In addition, there is also a side-arm outlet on the top left which lets out the waste that leaks out from the capillaries and pools outside, in the casing tube that houses the capillaries. Inside the capilaries the good stuff is circulating isolated, in the space outside is the urine gagarine
Comment by milkshake — November 7, 2012 @ 6:56 am
You should check out tangential flow filtration from millipore. They sell a nice set up that can purify and concentrate 500 ml of solution. We use it frequently for polymer and protein polymer purification .
Comment by Curryworks — November 26, 2012 @ 11:01 pm
I think that system is tangential flow filtration system, according to spectrum labs’ website. Ultrafiltration is when you use gas pressure to push a solution through a filter membrane with a given molecular weight cutoff.
I tend to work on smaller scales (10-50 mg), so disposable spin filters from Amicon and Pall have been my mainstay. They work like ultrafiltration, but instead use the centrifugal force of being spun in a centrifuge. This means after every spin (10 to 45 minutes, depending on brand and MWCO) you need to pour out the old solution and add fresh wash solution. Amicon filters have better body design, but Pall filters need shorter spin times and sometimes seem to be more efficient in removing the small molecule impurities.
Comment by sciencemonkey — February 1, 2013 @ 3:36 pm
Hello milkshake, quick question for you: are boronic acid pinacol esters stable in methanol? I suspect one of the highly polar pinacol esters I’ll be making will only nudge if I use a DCM/MeOH system, and the last thing I want is to have it decompose while it’s on the column. Thanks.
Comment by Student — March 16, 2013 @ 2:47 pm
yes, pinacolates should be stable on silica and even in a mix with MeOH. But they are not stable in the presence of diluted acid – on reverse phase HPLC run in acetonitrile + water + 0.1% of trifluoroacetic acid they do fall apart. The easiest way to tell whether or not it will survive is to test it on TLC – you should be able to observe the pinacolate ester decomposition to boronic acid in form of tailing, and the liberated pinacol should be easy to detect with permanganate
Comment by milkshake — March 17, 2013 @ 2:47 am
Hello Milkshake,
can u please tell me how to quench LiAlH4 ? the unused up LiAlH4 will be filtered ? and how to quench that that filtered LiALH4 solids ??? any good ideas??
If i just take that powder and all of it to propnol– will it catch fire ??
Comment by babyelephant — March 26, 2013 @ 2:56 am
The usual 1x +1x(15% NaOH)+ 3x quench procedure by Fieser for THF as a solvent goes like this: Place the flask with the reaction mixture (in THF as a solvent) on ambient water cooling bath, with efficient stirring add water DROPWISE, for each used 1 g of LiAlH4 add 1 mL of water. This is accompanied by hydrogen gas evolution and boiling because the quench is so exothermic, some THF will be lost by evaporation. (You can also pre-cool the flask on ice bath but doing the quench on ambient bath actually tends to produce a little nicer, more filtrable solid cake of Al salts). When all water is added, continue the quench with 15% aqueous NaOH, again 1 mL of 15% NaOH for each 1g of LiALH4 used. Finally, complete the quench with 3mL of water for each 1g of LiAlH4 (the gas evolution and exotherm stops already so the water addition can be faster at this point). Stir for 20 minutes on ambient water bath, filter, wash the salt cake on the Buchner funnel with THF. Your product should be in the filtrates but it is worth saving the salt cake until you are finished with workup and isolation – in some cases the product can get stuck on the salt cake (this happens with substrates that have acidic groups, i.e. free thiol or phenol groups)
Comment by milkshake — March 26, 2013 @ 9:16 am
Milkshake; what is the lowest molecular weight cut off made for this system. I have a salty solution of 1400 g/mol I am trying to desalt.
Comment by okemist — May 22, 2013 @ 3:17 pm
that is a bit too low. I think the lowest MW cutoff column they have is 3000 (nominal value) and it has been my experience that when we put some of our polymers with Mn average around 5000 on the 3kDa column we still lose quite a lot. (It probably also depends on the macromolecule structure). I think you may need to go with dialysis
Comment by milkshake — May 22, 2013 @ 3:20 pm