Sciencemadness Discussion Board

Has anyone here attempted to build a DIY HPLC contraption?

FireLion3 - 18-9-2014 at 13:05

At the most basic level, would it be plausible to build a unit that circulates a solvent stream continually through a chromatography tube. With enough passes the compounds should become quite separated. Without detectors the only real trouble would be trying to locate each compound in the stream, but I don't imagine that would be too difficult with a variety of techniques.

Forgive me if this has already been discussed to death. I only just started learning about chromatography and it seems fairly straight forward. I don't immediately see any reason why a small scale cheap DIY automatic circulating unit could not be built that would be fairly efficient.



Such a system could be tested as above, for instance, with various ink, at a basic level.

aga - 18-9-2014 at 13:18

Knowing the gas pressure, and pulsing it for a known time between readings would be good.

What detection techniques do you have in mind, assuming that it's not colours ?

FireLion3 - 18-9-2014 at 13:26

That is still one of the things I am working out. I literally only just thought this up like an hour ago. These little issues, while unresolved at the moment, can be worked out.

It really depends on what the compounds are. For instance, after it circulates for a while, you could take TLC samples every little bit until you find out where the 'blank regions' are between your compounds. This could be incredibly tedious, but it could be viable with practice, depending on the amount of substrate to solvent, and other variables.

Tonight I will be researching a bit more into the various types of detectors commonly employed. I'm sure it wouldn't be difficult to incorporate some [amateur variation] of them into a home system.

[Edited on 18-9-2014 by FireLion3]

forgottenpassword - 18-9-2014 at 13:30

If you are using TLC to assess the effectiveness of the separation, why not just use TLC instead of HPLC? What benefits do you think HPLC has over TLC?

packetforger - 18-9-2014 at 13:51

UV sources and detectors, while a bit expensive, could be used. Or even just a wide spectrum light source and a sensor on the other side - if the absorption/transmission changes during the run, flag it as a peak for however long.

I actually prefer HPLC to GC, despite GC supposedly being more efficient, as it means it is possible to collect all the individual fractions that come off at the end for further analysis, wheras with GC recovery of samples is impossible (as far as I am aware/in my experience).

aga - 18-9-2014 at 14:08

Clearly it is an Idea in Motion.
I think it has merits.

First is that Pulsed, then scanned, then pulsed again will give more data.

Pulsing would apply to gas or liquid. TLC maybe not.

The Sensors are the Biggie, as always: what to detect ?

phlogiston - 18-9-2014 at 14:26

What you are describing is just LC, very useful indeed and easily done with as little equipment as a glass tube (and optionally a pump if gravity is too slow). To call it HPLC at this level of implementation would be a stretch unless you can get a very well performing solid phase material, pack it properly and achieve the required high pressures, yet constant non-pulsing flow rates.

Ofcourse, even a primitive setup would be cool and useful, it is just a matter of semantics.

Some other things to consider:
1. what are you going to use as the stationary phase?
2. how to load the sample. e.g. improvise a loop valve?

[Edited on 18-9-2014 by phlogiston]

aga - 18-9-2014 at 14:41

Gravitic, controlled flow or pulsed, how to detect ?

That is the main problem isn't it ?

Edit:

Pardon if this is just stupidity.

Surely working Against Gravity would give more Control, and Data.

Controlled Upwards pressure, pulsed, would give both speed of separation, and rate of regress (after the pulse ends) rather than just how fast it descends the column in a gravitic system.

2 data points per pulse of Upwards pressure.

Pulses can be per hour, minute, second, microsecond etc

[Edited on 18-9-2014 by aga]

FireLion3 - 18-9-2014 at 17:51

I don't think the pressures need to be ridiculously high. Of course the effeciency will drop, but even a moderate amount of pressure should allow for quite pure sample isolation. Remember, we are not talking about thousand dollar equipment here, but improvised equipment that could do the job quite nicely for any amateur scale.
Quote: Originally posted by forgottenpassword  
If you are using TLC to assess the effectiveness of the separation, why not just use TLC instead of HPLC? What benefits do you think HPLC has over TLC?


I'm no expert, but can TLC be scaled up beyond mmol batches?

Quote: Originally posted by packetforger  
UV sources and detectors, while a bit expensive, could be used. Or even just a wide spectrum light source and a sensor on the other side - if the absorption/transmission changes during the run, flag it as a peak for however long.

I actually prefer HPLC to GC, despite GC supposedly being more efficient, as it means it is possible to collect all the individual fractions that come off at the end for further analysis, wheras with GC recovery of samples is impossible (as far as I am aware/in my experience).


Interesting thought, definitely dooable if you can find some reasonably priced detectors. Uv sources are no problem.

You can take this a step further and hook it all in to a PC, and have the PC record the data from the device per second, and program it to control the pump. If you can get the device, then interfacing it to the PC almost costs nothing.
Quote: Originally posted by phlogiston  
What you are describing is just LC, very useful indeed and easily done with as little equipment as a glass tube (and optionally a pump if gravity is too slow). To call it HPLC at this level of implementation would be a stretch unless you can get a very well performing solid phase material, pack it properly and achieve the required high pressures, yet constant non-pulsing flow rates.

Ofcourse, even a primitive setup would be cool and useful, it is just a matter of semantics.




As I understand it, HPLC is far more effecient than normal LC in a manual glass tube because it can do dozens of cycles in minutes, continually. If you were to do this amount of cycles/separations on a normal glass tube you would be doing it for days.
Quote: Originally posted by phlogiston  

Some other things to consider:
1. what are you going to use as the stationary phase?
2. how to load the sample. e.g. improvise a loop valve?

1 - The solid phase, could be experimented with, but I imagine you could use any number of things. You would just set the column up as you would a normal LC before each separation you do, and go from there.

2- something like that, This would not be hard to implement.

What I am most interested in at the moment is Chiral LC, in which the solid phase is some sort of chiral component that does not dissolve in the solvent, and will thus separate racemic mixtures.

Quote: Originally posted by aga  
Gravitic, controlled flow or pulsed, how to detect ?
2 data points per pulse of Upwards pressure.

Pulses can be per hour, minute, second, microsecond etc


What is the benefit of pulsing again? If you can get a UV sensor linked in there and hook that to a computer to chart the data, you could pretty easily be able to tell when the mixtures when the separation is efficient, and even get it to stop after that point.

Use a solvent with minimal UV intereaction, the UV graph on the computer after separation is good may look like this.

full uv ...... item detected.....full uv..... item detected.

Or something of that nature. Each compound would have a slightly different reading depending on your measurement. You could put the detector just before the drainage valve of some kind, and make it drain specifically until the item is no longer detected by the UV. You could then set up another container, and tell it to run through until it hits the next item, open the valve, pump until that item is no longer read....voila.... 2 perfect separations.

UV would have limited applications but if you can get a UV sensor and hook it to a computer, you could have a significantly useful set up starting here.

smaerd - 18-9-2014 at 19:34

UV is great for conjugated molecules. You're method of detection should be based on the type of chemistry you would like to do. Building a UV detector would be quite simple so long as absolutely stable intensity readings were not a neccessity.

HPLC is a bit of a pain to engineer at home. At least reaching the specifications of a commercial instrument would not be easy. For example, take a look at the pump heads on those puppies? Believe they often utilize sapphire for it's material properties. Often times scrapped HPLC's will be found with said sapphires removed :mad:

Another issue that seems trivial but can be mind-bending to concoct a work-around for is, Rheodynes. Unfortunately they are pretty expensive. Sure classic valves rated for the conditions used and a net-work of tubing could be used, but that complicates the instrument and limits fun idea's like automation.

The other thing is you know, building a pressure system that flows in solvents/possible corrosives. Not the easiest thing to do cheaply/'safely'. And all the small details like sparging solutions and filtering them, useable fluid flow etc. Do you want gradient elution, etc.


I concocted plans for a 'medium performance' chromatography apparatus that could automate fraction collection and other things. Unfortunately that is on the back burner as I have way too much to do right now. I could try and dig up the plans for said device but plans are plans, and I think you can come up with something equivalent without too much effort. Funnily enough my interest in this was also driven toward to chiral resolution. Chiral chromatography isn't really for the uninitiated though, so definitely read up if you haven't already. The idea was even if a poor resolution was initially obtained I could automate it to cycle back through a column. Kind of like a half-ass simulated moving bed column. If the analyte does not irreversibly bond to the stationary phase, one can actually separate a decent amount of material on a small scale many times quite quickly. Pretty surprising to be honest :).

The benefit of medium performance is, less pressure ''issues'' in the design. (I never down-play pressure hazards, and try to avoid it at all costs in projects). Capacity of analyte can be appreciable. Stationary phase is cheap(not $1,000 columns that load 10mg of sample). Last but not least, total COST.

When you look into it, it's often a lot cheaper to second hand purchase an instrument such as an HPLC rather than reinvent the wheel. Then again, I'd love to see some ingenuity put toward such a project. So don't be afraid to think outside the box here ;).

[Edited on 19-9-2014 by smaerd]

Edit - ultimately I shifted my interest in the project toward lab on a chip. It's most cost effective, obnoxious to manufacture considering the tools to construct the tools need to be home-made(in progress). It's also micro/nanoscale, but the wonderful thing about lab on a chip is processes can be run in parallel without too much issue and laminar flow concerns are largely eliminated without much effort. More wonderfully, tiny bits of stationary phase. Check out UHPLC, that's how they're doing this stuff now days :). What I wouldn't give for $1,000 spending cash and 10 hours spare a week for this stuff...

[Edited on 19-9-2014 by smaerd]

FireLion3 - 18-9-2014 at 20:35

These are all valid points. Are there reasonably priced HPLC machines on the market? All the ones I immediately saw on google were $5000-$55,000. You mentioned one for $1000? Still not cheap, but if I really wanted one that is a reasonable goal to save my money for.

[Edited on 19-9-2014 by FireLion3]

phlogiston - 19-9-2014 at 01:12

Quote: Originally posted by FireLion3  
I don't think the pressures need to be ridiculously high. Of course the effeciency will drop, but even a moderate amount of pressure should allow for quite pure sample isolation. Remember, we are not talking about thousand dollar equipment here, but improvised equipment that could do the job quite nicely for any amateur scale.


Ask yourself why a pump is needed at all.... Have a look at the van Deemter equation if you have not heard of it. The flow rate is set to the optimal value to get the best separation. It will mainly depend on the particle size of your solid phase.

The reason pressures are so high in HPLC is because the solid phase particles are very small. This results in good separations, but to achieve the optimal flow rate of the mobile phase, high pressure is needed. So, in your setup the required pressure depends nearly entirely on what you will be using as a solid phase and your column dimensions.

Generally, you want a pulse free flow to avoid fluctutations in the chromatogram baseline and symmetric, sharp peaks.

As an alternative option for chiral separations, you might want to consider that you can also derivatise your compound with a chiral reagent and separate the resulting diastereoisomer on a non-chiral column. Or use a chiral modifier in your mobile phase.

[Edited on 19-9-2014 by phlogiston]

aga - 19-9-2014 at 10:40

Somehow i think this idea is do-able.

I'm definitely out of my depth with the physics going on though.

Can anyone suggest a simple experiment with readily available stuff that demonstrates the effects ?

Dr.Bob - 19-9-2014 at 10:44

If you want to do LC, the simplest way would be to start with a simple gravity column, which costs almost nothing, a glass tube, preferably with a frit and a stopcock, but even a pipette and some glass wool will do. That will work with simple separations, like using silica gel or alumina to separate the dyes in Kool Aid, using cellulose to separate some biologicals, and using ion exchange resins to separate ionics apart.

If you go up to flash Chrom, which is just the same column with a simple way to pressurize it, you can get better and faster separations of organics on silica gel or alumina, bu8t chiral separations are very difficult, both in terms of finding conditions that separate two isomers, and in terms of detecting them. I would always start with something colored, as that will let you see what is happening, so you will learn how to pack the columns (you will see voids and cracks in the packing material easier), and you don't need a fancy detector. Or invest in a test tube rack and lots of tubes, collect all of the fractions (~1/20th of the total volume per tube is a good starting point) and then either let them evaporate or spot them on a clean piece of glass and look for evaporation residue. Very few solids separate organics as well as silica gel or alumina, which is why they are used so much.

And to be able to do much LC, you really need to understand TLC, for finding the optimum conditions to run the column in, checking fractions, and learning the techniques cheaply. That is also fairly cheap, you can get a simple chrom. jar (or use any glass jar of the right size) and some small TLC plates (2.5 x 7.5 or 5 x 10 cm are common sizes) with silica gel and a fluorescent dye and you can learn TLC in a few days. There are other packing materials that work with some solvents, but they tend to be very expensive, and chiral columns cost a huge amount, in most cases. There are cheaper ways to do many chiral separations, like salts with other chiral compounds, and picking apart crystals with tweezers... :-)

Most HPLC uses reverse phase systems on C18 columns, mostly water and acetonitrile or methanol solvents. HPLC does not work as well with organic phases, as the columns tend to gum up or get clogged faster under the higher pressures with silica gel.

If anyone wants to do this, I have some of the pieces necessary for TLC, LC, flash, and even some HPLC columns.

unionised - 19-9-2014 at 15:07

There was always a debate whether the "HP" in HPLC stood for high pressure, or high performance.
I don't see any home-made kit running at 2 or 3 tons per square inch or offering the ten thousand theoretical plates of separation you get from a commercial system; so DIY HPLC is, I think, a non- starter.

smaerd - 19-9-2014 at 17:12

The $1000 reference was for putting into a project not an HPLC. I have seen some old waters HPLC pumps online circa 1980 for around 200-300$.


I tend to do half-ass flash chrom at home. Either I use a syringe to push air into a column(depress plunger and watch the column piss), or a small aquarium pump(~3psi). Which I must mention I would not be doing without the help of Dr. Bob helping me acquire beautiful columns, and brand new TLC plates :) . Granted I've only run a few columns at home. Actually I taught a class-mate how to pack a chromatography column because she had never done it before. Benefits of a home-lab!

I was told it stands for high performance and that high pressure is a misnomer.

I definitely agree with unionised on HPLC not being too attainable for at home. Shifting the discussion toward 'medium' performance, simulated moving bed, or automated chromatography though seems to have great merit for an amateur chemist. Or just someone interested in separation science :].

smaerd - 19-9-2014 at 17:20

Okay, fine I'll show you a little secret gem I found on the web because UV detectors were mentioned

Arduino based UV detector for column chrom.

http://waveandsignal.blogspot.com/

Respect for the author! Battling noise in a DIY instrument is always something worthy of a hat-tip. edit- speculative but I think a lot of his noise issues had to do with things like inductive effects from oh I dunno say a vacuum pump running, fume hood fan's, etc. Doesn't look like he filtered the circuit at all either, just straight up detector to analog in. They did mention smoothing the data though. Could be wrong(often am).


[Edited on 20-9-2014 by smaerd]

phlogiston - 20-9-2014 at 01:10

Another source may be the fluctuating light from fluorescent lighting in his UV cell. His dominant source of interference seems to be at mains frequency.

Interesting effort indeed. Probably, for most DIY scientist, acquiring a quartz cuvette will present problems, making the interesting UV part of the spectrum unavailable.

S.C. Wack - 20-9-2014 at 04:37

Quote: Originally posted by smaerd  

I definitely agree with unionised on HPLC not being too attainable for at home.


Sorry for extracting your quote in this fashion, but BTW I've had the opportunity to pass on being the high bidder on probably every GC, (and some LC parts) the university had from 1970-2010, and at the last auction, there were I think HP/Agilent 1100 HPLC modules that went unsold. It was a shame I couldn't identify even the function of these pieces at the time because they could have been sold for hundreds at least. They were tossed in the trash, and the U doesn't auction non-functional equipment. I mention this not because that kind of HPLC is particularly useful at home other than fund-raising (software and solvents not being included), but because of the complete disinterest/bafflement from several who normally pay good money for scientific equipment they know nothing about, and don't hesitate to drop $50,000 in one auction. A GC went for not more than $200 once. Otherwise, no one ever bids on GC and LC equipment. Ditto anything biological/DNA/protein, even though these pros spend all day googling values on the phone.

smaerd - 20-9-2014 at 05:15

Phlogiston there are quartz cuvettes on ebay from china for $10. His flow-cell was more involved but still, very possible.

Yea S.C. Whack my quote implied Building an HPLC from scratch would not be attainable. Frankensteining parts or fixing a damaged instrument is definitely possible. Like I said I have seen functioning older HPLC's for 200-300$ on auction websites.

Ex:

http://www.ebay.com/itm/WATERS-6000A-HPLC-SOLVENT-PUMP-/2313...

[Edited on 20-9-2014 by smaerd]