44 I don’t like to run columns and I generally don’t. My syntheses are so bad assed, I generally work harder at figuring out purification methods that don’t require silica because columns are very, very wasteful and it hurts the earth and Al Gore’s feewings. That being said, most of my shit is purified on a column. But I’m both smart and lazy so I’ve used technology to make my life 900% easier by running flash chromatography once a month. My other column is a ghetto MPLC and I LOVE IT OMG LOL!!!!1!1one!.
An MPLC is the “medium pressure” variant of an HPLC, but they really couldn’t be further from HPLCs. Whereas HPLCs are unmistakably scientific instrumentation, MPLCs never are. They’re simple purification machines and, as such, they shouldn’t cost $6000. (a bargain when you consider the $40,000 an HPLC costs) I provide here the means for you to build your own MPLC for purifying organic goodies for under $1500. If you can’t swing that, you should get another lab. Since I’ve run all kinds of things on this set up, I can promise you that if you can separate it on a flash column, you can separate it better on an MPLC. I list vendors here, but you can buy from anyone. These are just the guys I buy from and I’m not getting any money by mentioning their names (though I should) so I can be candidly honest about all of them.
MPLCs can be divided up into 3 parts: The part that runs the solvent, the place where your shit is separated and the part that collects the solvent as it comes off the column. The first thing you need to consider is a pump and a solvent system. I
run a lot of binary gradients, but that doesn’t mean I need to buy two pumps. All it means is I need to buy two bottles, one of which has an awesome screwcap top. The pump I have connected it to is an FMI “Q” Pump. You can find out more about these guys here at their website. The pumps can produce a lot of pulsation, but that’s taken care of down the line. This pump is a cheap one and I’ve let it run dry over a weekend and it still keeps chugging. It’s grade A good shit. The whole thing is about $500. You can control the flow rate by twisting a little knob. It not only gets pretty fast, but you can even make it go backwards, which I have yet to find a use for. That bottle is obviously from an Acros bottle and that PFTE cap was purchased from RT Scientific. Essentially, as a vacuum is created in the brown bottle, it pulls solvent from the clear bottle. This is how I get my gradient. It’s very much a true gradient – as much so as a 50mL prep HPLC binary system can produce. That silver thing the brown bottle is sitting on is an old shitty stir plate, of which most labs have a shitty one lying around.
That’s the first part. The second part involves the actual purification and sample introduction. That’s actually even cheaper. As you need to purchase an $8 disposable column from Ana-logix (you can ask for a free sample pack from a sales rep), a $30 gas tight syringe, a $70 3-way lure lock and a small glass loading column, also available from RT Scientific for something like $100.
The whole assembly is connected with about $20 worth of tubing and I’ve run more than 300 liters of chloroform (seriously) through this set up. Nothing will touch it. The pulse suppressor is probably necessary. You can buy it from FMI (see that link above) for something like $250. That’s a lot of money, but separation is generally unaffected if you keep the pump speed low. If you can afford it, get it. So, that’s pretty much all you need if you want to collect by hand and you’re still under $1000. Collecting by hand is awfully gay and sort of defeats the point of “automation” since I find nothing more enjoyable than setting up my MPLC and going out to lunch or writing a blog post (the MPLC is indeed pumping as I write this). But this isn’t such a big deal. While fraction collectors are notoriously over priced, Ebay always has good deals on them. Indeed, a quick search turned up 66 different items. The best part is, if you pay for it with your own money you can tell people to fuck off if they want to use it and you can take it with you to your next job/postdoc/whatever. I didn’t have to buy my own, but life without it would be the sucks. Indeed, here is an Ebay store selling the older version of what I’m using right now for only $250.
So, there you have it. Automated for under $1200. You can think of it in terms of saving money in solvent and silica (you won’t) or think of it in terms of saving time.
Now… you could be asking yourself “But Kyle, you’re rich and famous. Why are you using such a ghetto-assed-fuggly MPLC set up like that?” To which I would respond: Because. I built it from NOTHING AND MADE IT PERFECT. And no one will use something that looks like a bunch of shit hobbled together so I pretty much get it 100% of the time, even though it kicks so much ass its feet are forever covered in shit.
Back to top
How much sample do you separate at a time?
Typically around 300mg. I’ve done between 10g and 40mg on them. Bigger and smaller than that and I use flash.
I’m very much impressed, Kyle. Now it looks like I have a new weekend project.
i want one and i hardly ever separate things!
Ghetto fabulous.
What kind of pressure are you running at?
depends on the column. RT scientific packs glass MM columns that I push to 60 PSI. Those things are ultra-awesome as I can wash them with METHANOL and reuse them about 10-20 times if I need to. (yes. wash them with pure methanol 10 to 20 times). The analogix columns I run about 10-30psi, depending on the silica.
That reminds me of something I’ve wondered. What does methanol do to silica gel? It can’t be that silica is actually soluble in it, right? But it sure looks that way–or at least there’s something milky and sort of opalescent that methanol pulls out of silica. What’s that about?
I’ve tested this with various samples of silica and have never found methanol to actually move the silica. It could be the type of silica we use, but even when I flush my flash columns with pure methanol to remove shit stuck to the top, I see nothing but methanol come out of the bottom.
lemme see you purify 5g of material on that thing!
That’s pretty easy to do, actually. Much over 10g and it gets difficult, so I’ll switch to flash, but in principle I could load as much as I wanted to.
Now all you need to pass the collecting end of the tube taped under a uv lamp (for aromatics and other uv comps).
it should read … to do is to pass the ….
Thats pretty nice set up. How can you run so much sample on that small column?
That column handles up to 1g, in theory, but I’d never run more than 300mg on it. You would need a much larger column to run things bigger, but they’re all around $10-30 – depending on the quality of silica, the manufacturer, etc.
The one thing you may need is adjustable UV detector and a fraction collector.
What kind of change in solvent mixture does your gradient run? Are we talking 0-50% or a bigger/smaller change? How precise is the control on that?
Well, it is whatever you want to make it, I guess. Just fill your second bottle up with your target – if I want to ramp it up to 5% methanol in chloroform, i’d fill the brown bottle up with 500mL of chloroform and fill the clear bottle up w/ 1000mL of my 5% mixture. The collection time is adjusted by the flow rate, so from there you can figure out how much and how quickly you’re ramping it up. You can do the math on an excel sheet, if you were super curious, but it essentially tangentially approaches 5%.
Of course you could do it like you would do a flash column, and just open the bottle and add your gradient in from a volumetric flask, but since I have the fraction collector attached, I typically let the column run very slow and even overnight so I get maximum separation.
Exactly what I wanted to hear. Looks like someone is going to be making a few equipment purchases in the very near future.
We use this same principle in biology to cast gradient gels. However, most people these days buy their gradient gels for $10 a pop from invitrogen (doing it by hand nets a $2 reagent cost and a time for grad student cost, which is zero), so the art of doing this is lost.
How precise is the control on that?
The answer is: Good enough.
Seriously, let’s say the gradient oscillates up and down slightly on the way up, even taking account the mixing in the second beaker. It’s not going to be a big deal because over the length of your column, your product will emerge as a stretched normal distribution, and you can bet that the sigma for this distribution will be way more than any oscillations that you observe in your gradient, not to mention the fact that your fractions will probably be in the 5-10 ml volume range anyway.
Grad student time cost is zero? Hardly. The cost of a grad student is typically more than a postdoc (think >$60,000 when tuition and fringe benefits are paid).
Assuming a 60 hour work week (which seems standard these days, right?), even if the gel takes 30 minutes to make (that is roughly $10 in labor costs), buying a pre-cast gel makes a lot more sense.
When the tuition and fringe benifits are paid, a grad student will not cost more than 45K, which is the absolute upper limit. Tuition is paid in state in most cases no matter where the grad student is from.
I’m guessing you’re at Fleece U. I really hate that place.
Benefits/management costs are substantially more than salary in most places – for example (we have a good health plan) the cost of health care is estimated at $7K/employee. Grad students sort of get vacation and sick leave (but not really), but they still have to pay SS and Medicare on salaries, and some health care. If you make $20K (gross) as a GS, there is at least $10-20K in tuition to go along with it, for $35-45K before benefits. $45K isn’t going to be an absolute upper limit anywhere.
(I think in the private sector, the expectation is that a position actually costs the employer something like 2.5X salary – if that rule held, GS would cost at least $50K, not counting tuition.)
Well, I pay $58,000+ per graduate student. At a top-20 public research university.
The breakdown is: salary ($28,000) + benefits ($12,500) + tuition ($17,600).
This is very similiar to where I did my postdoc (top-5 public institution). I guess if you still disagree, you might find out the truth someday.
Yeah, that sounds like Fleece U. Although, I wouldn’t mind the 28K salary. The tuition that they charge here is only 7.6K and the benefits don’t include health or dental, so they save on that as well; the grad students pay for it. The phantom benefits (for god knows what) are not anywhere close to 12.5K.
Why would I disagree with you? Obviously I was wrong because I was universalizing my experience at the current place, which wants me to live in abject poverty in one of the most expensive metro areas in the entire country. At least the profs save some on tuition. This makes me kind of hate the place even more. Even more than the usual last year graduate student hate for their university.
Working at a low paid job for the last few years (you don’t learn much after the first year unless you sit in on classes or make an effort to learn other techniques — which many ’students’ don’t) has made it very hard for me to justify working more than 40 hours a week. The biggest reason is that I save money by being in the lab and not outside of it. But I can also play video games and the end effect is the same.
Sorry… the salary figure you put up has made me very, very bitter.
You shouldn’t be. $27,000 isn’t enough to live off of in Berkeley. It may be more than enough in Boise, Bloomington or Eugene (or even Atlanta, South Bend and Madison), but you’ll have no extra cash to do anything in Berkeley or Boston aside from eating Raman and working in the lab.
This is way fucking sweet. I dreamed about these at one point.
do you have a specific blue print on how to set this up. You post was very informative but I am a little slow. If you don’t can you point me towards a place I might find one? Thank you.
Very very sexy. I have a passionate burning hatred for flash columns, but there’s few ways around them that get things so squeaky clean.
Im thinking of making a set up like this. I already have a pump. Do you know the part number or model for those columns that you use?
Hmm… I’ll have to do a more thorough post, I guess. Possibly next week.
You may be able to modify this and get a MPLC for less than $1000.
http://cgi.ebay.com/PHARMACIA-.....18Q2el1247
I have been running MPLC like this since I was a grad student in the early 80’s and since I am on the faculty of an undergraduate institution, cost issues are definitely relevant. If you want to go really cheap, there is a line of columns called Omni Columns (I think they are made by Rainin)which are repackable with flash grade silica gel, so you can put your crappiest stuff on the column without worries – if you make a mess out of the column, you can just repack! Also if you are wondering “why the reverse setting on the pump” – here’s why: If you are too cheap to buy a syringe to load your stuff on the column,you can leave a long enough piece of tubing between the pump and the column head, you can CAREFULLY suck your compound in solution into the tubing after unhooking the tubing from the top of the column, then reattach, and load your compound on with the pump set on forward. Note – you have to be careful doing this, as you do NOT want your compound gunking up the pump head, as it will make all kinds of horrible nasty grinding noises.
I might also add a plug for old Gilson fraction collectors, such as the GILSON FC-100 – these things are absolutely bulletproof – I have never had one break, and I have a few kicking around which are probably 30 years old. The design is such that if the organic solvents miss the test tube, the working parts underneath are fully protected, hence the long life.
I don’t know which is more awesome, that you only have to run columns once a month or that you built a column that’s cheap and efficient and you don’t have to wait to use it.
Our lab has a fraction collector made from an ancient XY plotter rescued from the trash. One of our past PhD students wrote a simple program to run it – it’s fully adjustable and can take anything from test tubes to Schott bottles. Cost: nigh on zero. It would fit in great with your setup!
This is absolutely indispensable. I’m building one as we type. I had one question. If I was going to pack my own columns how might I go about doing it in order to achieve and best separation. I’ve packed it wet and dry. no matter how I pack it (with 0-60 um silica) the pressure compacts the silica and I end up with a gap where my loading band becomes ridiculously large?
if you load the material onto the column at 0% A, and flow through a lot, then the “width” of pressure at the top should not depend on diffusion induced by gaps at the top, unless your compound is significantly mobile in your A phase.
wow, this makes no sense at all. I’m not going to post advice on columns ever again, at 1:00 in the morning.
Tap the side of the column as you slowly dry-fill it to release air. Then overfill it moderately. If you have empty head space after running solvent through it, open it up and top it off.
That depends upon the columns you are using. If you can rig the bottom of the column up to a vacuum pump, you can pull the silica down (of course, your plug should be nice and snug and this is obviously a dry packing method). Adding about a gram of silica at a time to the column is a decent method to make sure it is optimally packed. Of course, you can plug the bottom into an aspirator in your hood if you don’t have a vacuum handy.
The other method is to simply tamp the silica down with a stir bar. Smacking the column works well for flash, but not for MPLC. For MPLC, you’ll need more rigorous packing conditions. The gap at the top should be filled with sand to keep the silica from sliding around when the column is not in use. It should not be a big gap – certainly no more sand than you would use on a flash column (some people go crazy with the sand… it’s almost unneeded.) Also, keep in mind that you’ll want to run your solvent in the same direction you packed.
I knew a guy from the Hoye lab (where he said they use MPLC routintely) that packed his columns dry using a vibrating Dremel tool. He would rig it up so that the tool was at the bottom of an empty column, and then he would sift in the silica at the top, while the column was vibrating. It’s similar to tapping a flash column while you dry pack it.
A Dremel tool eh… *nudge nudge wink wink say no more*
I worked for Hoye as an undergrad. He loved his MPLC, and purchased enough used HPLC equipment off of Labx that the MPLC/researcher ratio was like 1:2. I remember being so proud the first time I packed a column and got like twice as many plates as the grad student I was working for (that number may be exaggerated, it’s been a while). The Dremel tool works great, but you need to put a cork on the top so it doesn’t eat the bottom of your column. We used 18-32 micron silica to get better-than-flash separation. Honestly, I never had to run a gradient even to separate my melty-snowman spots.
You can buy columns from Ace that are admittedly expensive, but can take up to 25 g or more for those of you looking for more mass-throughput. I’ve hand-packed one of the 5 cm columns with a dremel tool. It took forever, but it gave amazing separation.
I’m happy to see the chrom equipment purchased off LabX worked out well for you in this application of it, For those of you who are not familier with LabX it is the online market place where you can find new and [i am not your Google whore] from over, right now at this time, 140,000 items. The link is http://www.labx.com and its all free for buyers to use.
Sellers have to pay $30 to list an item if you have some extras sitting around you dont need anymore. If you cant find the exact component you need make a free wanted ad and you will get some good response. Good Luck on MPLC,
Cheers, John.