[WalterFaber]

Hi,

I'm currently working on the synthesis of some 4,5-disubstituted acridine derivatives with different aminomethylene groups. The starting material is 4,5-bis(bromomethyl)acridine that is reacted with a secondary amine and K₂CO₃ in MeCN or CHCl₃.

The reactions themselves work reasonably well. After workup (filtration and removal of the solvent) an NMR of the crude product usually shows complete consumption of the 4,5-bis(bromomethyl)acridine but often some unidentifiable side products are present. As I want to do fluorescence studies with the final products they have to be absolutely free of impurities. However I have great trouble with the purification.

I tried silica gel columns in most cases but the yields were terrible and separation was bad despite the TLC looking ok. Wet loading usually is not possible because of the limited solubility of my substances and dry loading doesn't work so well (the compounds seem to stick to the silica afterwards). I tried dry loading on celite (washed with acid and DI H₂O) and it seems to help. Nonetheless separation is bad.

I also tried some columns with basic alumina but they didn't go much better. We only have TLC plates with neutral alumina and I think the separation results I get on those aren't really comparable to the basic alumina columns.
Is it possible to modify the TLC plates beforehand to make them more similar to the column material? Perhaps by treatment with NEt₃?
Also neither my coworkers nor me have much experience with alumina compared to silica. How do you estimate the amount of alumina needed for a certain separation? Can I use the same rules of thumb as for silica (rf of the product ~0.3, diameter of the column depends on the amount of crude product, the stationary phase is packed ~15 cm high, flow rate of 5 cm/minute)?

Any help would be greatly appreciated!

[phth]

dean stark right? Try C18 silica, it maybe cheaper to test with a TLC/prep (to check degredeation) plate or HPLC first.  The chemicals will be farther away from the silica, and your compound does not have to be soluble in the solvent system for the separation to work.  Often TLC and column may not be the same sized beads.  The reaction of radicals, cyclohexyldienyl adduct with nitrogen can happen-->tautomer immine-->substituted enamine-->harder separation and degredation.  Try doing it dark and/or with a high energy radical trap such as butylated hydroxy toluene (catalytic %) which is cheap to buy and easy to make.

[phth]

[discodermolide]

Why not try and crystallise them?

[WalterFaber]

I will give it a shot with the next few compounds. Most of the substances I synthesized so far really stretch the definition of a solid. Even if they're completely pure by NMR they are very gummy oils.

[phth]

Ok.  Make them salts, and then try.

[WalterFaber]

Ok.  Make them salts, and then try.

What salts would you propose? When I started the project I thought about precipitating them as hydrochlorides but both of the starting materials (and any acridine-related side products) form hydrochlorides as well.

[phth]

You can get a multi pka curve like an isoelectric curve.  Selective precipitation.  Using a pH meter you will observe the multiple buffer reigions.  You can determine % product % impurity this way. may need NMR

[WalterFaber]

OK, I'll have to ask around if there is a good pH meter somewhere in our department. Usually they're only used by our first year students and really inaccurate overall.

[phth]

Yes, They are, but what comes to my head is  titration=concentration.  NMR is very often used to determine pKa's.  you just gotta normailize solutions and take careful quantitiative NMR samples..   People also use this info for nanopartical avg size±u but XRD gives a distribution of sizes, for example.  AKA LEFR. Very low uncertainty.

[phth]

The power of NMR is the impurity peaks will go out of phase as the precipitate, so you then can infer by double Fourier transforming what is starting to go into the noise.

[Dan]

I tried silica gel columns in most cases but the yields were terrible and separation was bad despite the TLC looking ok. Wet loading usually is not possible because of the limited solubility of my substances

For highly crystalline compounds I find toluene- or DCM- or chloroform-based eluents much more effective. Hexane/petrol has very poor solvating ability, and if the eluent cannot dissolve the compound well enough, you will get streaking and little or no separation.

Toluene/acetone mixtures are pretty much bulletproof (but you need a good pump afterwards).

[WalterFaber]

Toluene/acetone mixtures are pretty much bulletproof (but you need a good pump afterwards).

I'll definitely try these mixtures. I recently used toluene/EtOAc for one of my precursors (an N-pivaloyl anilide) and it worked really well.
Removal of the toluene is quite unproblematic, we have Schlenk lines with good pumps in every fume hood.

Are there any problems regarding acid-catalyzed side reactions of acetone when it comes into contact with silica?

[Dan]

Are there any problems regarding acid-catalyzed side reactions of acetone when it comes into contact with silica?

No

[WalterFaber]

OK so yesterday I purified the compound I'm currently working on.

I did the synthesis twice before. The first time around I used a basic alumina column with Et₂O for purification. Separation wasn't that great and the yield of pure product was low. In the second run I used a silica column with Et₂O/pentane 4:1. Again it was a difficult separation but the yield was better.

Yesterday I tried the toluene/acetone suggestion. In all the mixtures I tried the product and excess starting amine ran very close. However toluene/EtOAc did the trick. The product has an Rf of 0.3 in a 5:1 mixture of these solvents.
I did a gradient elution starting with a 9:1 mixture going to 8:1, 7:1 etc. after each column volume and got a clean separation from the starting material. There still was some streaking, so I will try adding NEt₃ in future runs, but the toluene worked quite nicely!