Sciencemadness Discussion Board

Modafinil: its preparation via Bunte salt

kmno4 - 10-1-2015 at 17:32

There are several synthetic ways to modafinil, more or less complicated, giving various yields of
final product, with reported melting point in wide range of 150 C - 168 C for pure (racemate) substance (!).
moda.bmp - 166kB
Some says it is because of various polymorphs of this sulfoxide, in a patent (US8048222) I even saw preparation of "amorphous modafinil by crystallization from xylenes". It seems to me that most of these polymorphs are just contaminated modafinil... nevermind.
Publication from Russian Chemical Bulletin vol.59, pp.91-101, 2010 (DOI: 10.1007/s11172-010-0049-8, available also on this forum) gives interesting procedure for preparation of the sulfide (2-(benzhydrylthio)acetamide), which is further oxidised to the sulfoxide (modafinil).
Presented procedures are modified versions of given ones in this paper.

Part 1: Preparation of 2-(benzhydrylthio)acetamide.

The sulfide is prepared from the Bunte salt and benzhydrol in formic acid (+ water).
Mentioned Bunte salt is prepared from chloroacetamide and sodium thiosulfate:
Cl-CH2CONH2 + Na2S2O3 → NaS(O)3S-CH2CONH2 + NaCl
This is common way for praparation of various Bunte salts: some alkyl halide + some thiosulfate. More informations about Bunte salts(=alkyl thiosulfates) can be found, for example, in "Organic chemistry of bivalent sulfur" vol.1 by E.Emmet Reid.
In original procedure, the Bunte salt is prepared in pure crystalline form (long transparent needles), with rather poor yield (~70%). This is caused by solubility matters, the reaction is practically quantitative (as checked in separate experiments).
There is no need to prepare pure Bunte salt, equally good is its solution. Contamination with NaCl makes no harm, because it is only a spectator in subsequent reaction.
For each 1 mol of sodium thiosulfate*, 1,02-1,05 mol of chloroacetamide** and 250 g of water is used.On cold, chloroacetamide floats on the surface of thiosulfate solution, on warming at 50—60°C it starts to dissolve, giving clear solution. This warm solution of the Bunte salt is used for the next step without cooling. When less water is used, chloroacetamide also dissolves, but precipitate of salts may separate. Such mixture is equally good as transparent solution.
To 1000 cm3 of formic acid, 1 mol of benzhydrol is added, then 1,2-1,4 mol of the Bunte salt (as solution) is added. This mixture is then heated at 60-70°C, during about 1 hour. Original procedure says 20 minutes, but this period can be extended to 2 hours without negative effects. Magnetic stirring is desirable, but manual stirring from time to time is also possible (in this case I would recommend 2 hours). The mixture is biphasic all the time, with oily layer on the top and acidic layer under, that is why stirring is important.
Overall net reaction:
(C6H5)2CHOH + NaS(O)3S-CH2CONH2 → (C6H5)2CH-S-CH2CONH2 + NaHSO4
(mechanistic speculations can be found in the Russian paper)
Reaction is finished without any special visible effects, still two layers are present.
The hot mixture is then added, with stirring, to ~2000 g of cold water. White substance starts to separate as powder and small conglomerates. When all settels down, the solid is filtered.
It is raw sulfide, contaning absorbed and ocluded acid. The milimeter-size conglomerates are very hard, trying to crush them to smaller parts by some rod in glass flask may cause disaster !
All the solid, without washing, is placed in mortar with some water and crushed with pestle until all turns to white mud, without larger particles. The mud is then filtered and washed several portions of water. Solubility of the sulfide in water is small, no need to worry.
Wet cake is then dried in the air at ~50°C to constant weight. It is worth to control if all(almost) larger particles dissapeared. If not - treatment in mortar should be repeated. It is done because desorption of last traces of formic acid from the sulfide may last days (sour odour), especially if dried at rt.
Dried raw sulfide is white, slightly sticky powder, with weak "devilish odour". It is caused by unreacted benzhydrol, thiobenzhydrol... etc. Weight should be similar to given in the Russian paper: 24 g of raw sulfide from 18,4 g of benzhydrol. However, the paper gives melting point of this raw product as for pure one: 109—110°C. In reality m.p. is below 100°C. This fact was confirmed in a few experiments and by few checkers.

Purification of the raw sulfide.

5 g of the product was dissolved in 10 cm3 of toluene on hot. Clear, slightly brown solution should be obtained. When cooled down, all turns to white mass. It was vacuum filtered and washed with small amount of cold toluene (or heptane, hexane...). The most possibly the sulfide gives some solvate (or similar molecular compound) with toluene, because obtained crystals changes during drying (rt. or higher) and practically turns to powder (but crystalline under microscope). This purified sulfide*** is brighter than input one, practically without any odour and melting around 110°C. Yield should be better than 90% of mass of the raw sulfide (in my case it was 92%).

The final purification.

Few solvents were tested for crystallization purposes (ethyl/buthyl acetate, acetone, methylisobutyl ketone, isopropanol). The best effect gave isopropanol (IPA). 5 g of the sulfide is dissolved in 10 cm3 of IPA on hot, then solution is slowly cooled down to rt. Pure sulfide separates as nice, white glistering soft needles, looking similar to crystals of benzhydrol. After suction, remaining IPA is used for another crystallization (small amount of fresh IPA is added to obtain ~10 cm3****), wet sulfide is dried at 30-50°C, untill odour of IPA dissapears and weight becomes constant. Obtanined pure 2-(benzhydrylthio)acetamide is now ready for the oxidation step.

* because only 2-5% mol excess of thiosulfate is used, it should have composition close to Na2S2O3•5H2O. If partly dehydrated product (or of poor quality) is at hand, it should be recrystallized from water to obtain pentahydrate.
** it is prapared according to OS procedure (DOI: 10.15227/orgsyn.007.0016). Original molar proportion of ammonia to chloroester is about 3:1, but can be reduced to 2:1 without large loss in the yield of chloroacetamide. Reaction of NH3(aq) with ethyl chloroacetate can be conducted in a stoppered bottle with shaking from time to time.
Important: it is not pointed out, but chloroacetamide sublimes at elevated temperatures, so drying to constant weight should be conducted at room temperature (max 40°C).
*** it is desired to repeat this crystallization from toluene one more time: wet sulfide is again dissolved in 7-8 cm3 of hot toluene, cooled .... etc. It is recommended especially if somebody is going to eat prepared later modafinil. The purer substrate, the purer product.
**** filtrates are combined, IPA is then distilled off and kept for preparation of the catalyst for oxidation. Remaining impure sulfide is not wort further purification and can be rejected, its amount is (should be) less than 10% of weight of input sulfide.

On the picture: sample of IPA purified 2-(benzhydrylthio)acetamide (natural light, no flash)

sulfide.JPG - 241kB

Part 2: Oxidation of 2-(benzhydrylthio)acetamide to the sulfoxide - the write up is under preparation.

Nicodem - 13-1-2015 at 13:44

Thanks for your contribution!
Quote: Originally posted by kmno4  
Some says it is because of various polymorphs of this sulfoxide, in a patent (US8048222) I even saw preparation of "amorphous modafinil by crystallization from xylenes".

That is a very nice oxymoron, worth remembering. But then, even the phrase "amorphous form", which is so commonly used, is by itself an oxymoron.
It seems to me that most of these polymorphs are just contaminated modafinil... nevermind.

Generally, if a non-originator company goes trough the trouble of filing a polymorph patent, then the polymorph likely exists, though sometimes it is actually a hydrate or a solvate that gives a different powder XRD diffractogram, but is not a true polymorph (the patent examiners only care about the XRD characterization, so they can easily be mislead). But I can agree that there can be a lot of questionable things in patents on the polymorphs of APIs.
Contamination with NaCl makes no harm, because it is only a spectator in subsequent reaction.

In principle, the chloride anion can participate and react with the benzhydryl carbocation. Since in such polar solvents like aq. formic acid this is a reversible reaction, it does not matter (similarly as for HCOOH and H2O). However, in practice, the chloride anion is highly solvated in such media and therefore much less reactive than the mercaptoacetamide.
For each 1 mol of sodium thiosulfate*, 1,02-1,05 mol of chloroacetamide** and 250 g of water is used.

I assume the reagent in excess is the thiosulfate, the above being an error and the correct ratio being in the notes?
(mechanistic speculations can be found in the Russian paper)

I'm sceptical on their proposal of the SN1 directly on the Bunte salt rather than on the mercaptoacetamide which can easily form by acid hydrolysis under the reaction conditions (DOI: 10.1021/jo01039a019). I'll have to read their arguments when I'll have the time. Nevertheless, their direct use of Bunte salts in such thiether synthesis is an elegant solution to the problem of stench of thiols and their sensitivity to oxidation to disufides. In addition, the reaction forms its own catalyst as a reaction byproduct.

CuReUS - 14-1-2015 at 00:23

nicodem,I don' t get the oxymoron thing that you are saying :( ,could you explain that ?
also what is the catalyst that is formed as by-product ? I am a little confused :(
here is a thread on nootropics for furthur reading

[Edited on 15-1-2015 by CuReUS]

careysub - 14-1-2015 at 04:23

Quote: Originally posted by CuReUS  
nicodem,I don' t get the oxymoron thing that you are saying :( ,could you explain that ?

This part I can answer.

An oxymoron is a contradiction in terms.

"crystallization" refers to the formation of crystals while "amorphous" means lacking in definite form (no crystals).

kmno4 - 14-1-2015 at 13:00

Quote: Originally posted by Nicodem  

For each 1 mol of sodium thiosulfate*, 1,02-1,05 mol of chloroacetamide** and 250 g of water is used.

I assume the reagent in excess is the thiosulfate, the above being an error and the correct ratio being in the notes?

Thanks for careful reading my scribble ;)
Thiosulfate is limiting reagent, chloroacetamide is in excess.
The note should have started like this:
* because only 2-5% mol excess of chloroacetamide is used, sodium thiosulfate should have composition close to....
I think that excess of chloroamide does less harm than excess of thiosulfate during reaction in formic acid.
By the way, It would be interesting to check if benzhydryl formate (or acetate) reacts with chloroacetamide in buffered (pH ~7) water solution.
Thiobenzhydrol in ~5% amount was detected in raw sulfide, at least in one case (because only one was analysed)

[Edited on 14-1-2015 by kmno4]

kmno4 - 1-3-2015 at 23:58

Part 2: Oxidation of 2-(benzhydrylthio)acetamide to the sulfoxide

Among preparative methods of oxidation of sulfides to sulfoxides (e.g.[1]), oxidation by H2O2 is a method of choice because of its simplicity, cheapness and giving relatively high yields of sulfoxides under mild conditions.
It seems (from literature research) that the most common medium for oxidation of 2-(benzhydrylthio)acetamide to modafinil is acetic acid. The Russian paper gives formic acid as alternative for acetic acid, because "no sulfone is detected". Interesting alternative could be methanol, as it is more "amateur-friendly" than mentioned acids.
Procedure from the Russian paper was repeated<sup>1</sup> with 5 g of the sulfide: about 5 g of white powder was obtained, melting point was in range 155-158°C, not very sharply.
Oxidation in AcOH (e.g.[3]), was conducted in earlier experiments and the results were similar to given above.
Procedure from [2], utilizing methanol looks promissing and was tried with very good results.
For example: 5 g of sulfide is dissolved in 40 cm3 of methanol with 5 cm3 of catalyst<sup>2</sup> (isopropanol+sulfuric acid, in proportion 6 g of IPA + 0,28 g of H2SO4). After dissolution, the solution is filtered<sup>3</sup> and 3,9 g of 35% H2O2 is added<sup>4</sup> with stirring. The homogeneous mixture is then set on mag. stirrer at room temperature. There is small heat effect after peroxide addition, but there is no need to do anything with it. After ~2 hours<sup>5</sup>, white solid starts to precipitate (modafinil). Stirring is continued for ~12 hours (24 h of stirring (at r.t.) is not harmful).
The white suspension is diluted with ~50-70 g of water and filtered<sup>6,7</sup>, the precipitate is washed with water and dried. The yield is about 4,5 g of white, odoureless powder. It is modafinil, pure enough (sharp melting point at 160-161°C) to be eaten. In order to prepare purer product, all prepared earlier samples (about ~25 g total) were combined and mixed to obtain uniform composition.
5 g of this crude modafinil is dissolved in ~75 cm3 of hot methanol (from recovery), cooled, crystals separated, the filtrate is used for dissolving another 5 g... etc. Yield of purified modafinil is better than 90% of weight of crude material. It is starting material for final crystallizations.
Some tests of solubility of modafinil in IPA and MeOH were performed. The 5 g sample dissolves in 75 cm3 of ~70°C IPA, yielding on cooling 4,5 g of crystals. Another sample, dissolved in 53 cm3 of MeOH at ~60°C, gave after cooling 4,1 g of crystals. The crystals (foto is taken) look like small pieces of glass (in both cases), are rather hard and melt sharply at 161-162°C, with decomposition<sup>8</sup>.
To check purity, prepared crystals of modafinli were dissolved in methanol on hot (3,5 g /100ml), cooled, crystals separated, remaining solution was distilled to obtain ~30 cm3 of metanol, cooled...etc. Four fractions of crystals were obtained, in each case m.p. was 161-162°C.
When recrystallized from MeOH/H2O mixture<sup>9</sup>, modafinil separates as soft, small crystalline plates (foto is taken) with m.p. also 161-162°C. No higer m.p. could be obtained, in [3] the authors also give 161-162°C for their product.

Seems that thermal stability (in air) of modafinil is not large (oxidation ?). Samples boiled with water for 1-2 hours, after filtering are very slightly brownish, however with correct m.p. So, any thermal treatment of modafinil should be conducted as short as possible. On melting, modafinil turns brown and does not solidify back (it does, but after a longer time and lower temp.), with sulfur-like odour. On further heating of molten sulfoxide, it turns to red-brown liquid, tar-like at r.t. with SO2 odour. This sticky mass is easily soluble in water (especially hot), giving brownish solution.

<sup>1</sup> several oxidation in formic acid was done, with few modifications in time of reaction, amount of added water or H2O2 solution (temp. of reaction was kept as given, no more than ~10°C), but resulting product invariably melts around ~156°C
<sup>2</sup> it is unclear for me, why a mixture MeOH-H2SO4 would not work (somehow it was not tested), because in every case the "true" catalyst is some alkyl hydrogen sulfate
<sup>3</sup> it is the first filtration of solution of the sulfide, so all insoluble matters (dust, fluff of all kinds... etc) collected so far on the sulfide are removed right in this step
<sup>4</sup> any concentration above ~15% is good
<sup>5</sup> if more methanol is used, H2O2 is more diluted or temperature of reaction mixture is lower,
given period can be longer than 2 hours (in one experiment it was ~6 hours)
<sup>6</sup> when oxiadation is finished, modafinil can be just filtered (the purest, the lowest yield), filtered after addition of water, as given above (almost the purest, good yield), diluted with water, calculated amount of Na2SO3 added and most of methanol is removed by simple distillation (temp.of vap.~80°C). This gives the best yield (5 g of dried product from 5 g of starting sulfide), correct m.p., but it becomes odoureless after longer drying and may be very slightly darker than snow-white. I have impression that it gives more sulfurous aftertaste than modafinil prepared on cold.
<sup>7</sup> filtrates are combined and distilled (until temp. of vap. reaches ~80°C) to recover most of methanol. The distillate is a mixture of methanol with water and isopropanol, it is also a little stinky.
Small amount of NaOH is added (0,1-0,5 g) and methanol is distilled off again (until ~75°C). All strange odours dissapear, prepared MeOH containing small amount of H2O and IPA is used for subsequent crystallizations (or another oxidation)
<sup>8</sup> because of thermal instability, time of heating of the sample in m.p. apparatus should be as short as possible. Heating from r.t. to the melting point during ~20 minutes, gives m.p. always 1-2 deg. lower comparing to values obtained when the sample is inserted at ~155°C (time of heating to m.p. is then no longer than ~5 minutes)
<sup>9</sup> e.g. 2 g of the sulfoxide in hot 30 cm3 of H2O + 30 cm3 of MeOH

Part 3: Oxidation of modafinil to the sulfone.

Few attempts were made to prepare the legendary modafinil sulfone, in order to check its basic physiacal properities.
According to the procedure from the paper [3], 2,50 g of modafinil was dissolved in acetic acid, 35% H2O2 was added and the mixture was kept ~16h at r.t.
During this period nothig happened, the mixture was transparent and no heat effect was noticable after addition of H2O2.
Then water was added in portions (few first portions caused nothing but clear solution), and precipitated white mud was filtered and washed (water). Melting point was measured and surprice: ~160°C. It indicates almost unchanged modafinil.
Recovered white powder (~2,4 g) was split into two portions. The first portion was dissolved in 6 cm3 of acetic acid, the second one was dissolved in 6 cm3 of 80% formic acid. To both solutions, 1,5 molar excess of 35% H2O2 was added. Acetic acid solution behaved as given above, the formic solution behaved differently.
After several minutes, HCOOH solution was getting warm and some solid started separating, no oxygen bubbles were observed.
The mixture was set aside for few hours. After this time, the mixture was semi-solid but movable.
Then water was added, the solid filtered and washed. About 1,2 g of white powder was obtained. Measured m.p. was about 195°C -it indicates desired sulfone.
To the remaing acetic solution, 3 drops of 75% H3PO4 were added to test some concept.
After ~2 hours some solid started to separate, but in much smaller rate comparing to the formic acid experiment. It was put aside for several hours and then treated like formic acid mixure.
Measured m.p was lower than given above (and not so sharp), with amount of ~1 g. In both cases it was hard to remove traces of acids, dried material kept in a jar slowly becomes sour smelling.
Prepared sulfones were combined and crystallized from hot methanol (~100 cm3).
During slow cooling, the sulfone starts saparating in form of aggregates of small needles.
Filtering and drying gave ~1,8 g of pure sulfone, m.p. 199-200°C, without decomposition. It is white, fluffy powder (foto is taken), without odour and taste (possibly because of very small solubility in water). Its solubility in common solvents is much lower than the corresponding sulfide or sulfoxide.

[1] The chemistry of sulphones and sulphoxide from Patai series, gives a number of oxidation methods.
[2] Synthesis ; 1990(10): 937-938 (DOI: 10.1055/s-1990-27059),also patent nr. US8048222
[3] Neurochemical Research Vol. 29, Iss. 8, pp 1481-1486 (DOI: 10.1023/B:NERE.0000029559.20581.1a)

On the picture, from left: modafinil from MeOH (2g), modafinl from MeOH/H2O (2g), sulfone from MeOH (1,8g)

fot_3.JPG - 154kB

Metacelsus - 5-4-2017 at 16:13

Do not try to substitute glacial acetic acid for formic acid when reacting the benzhydrol with the Bunte salt. It does not work, and the benzhydrol remains unreacted. I expect this is because acetic acid is roughly an order of magnitude less acidic.

Cryolite. - 5-4-2017 at 18:31

If you read the paper, it clearly states that the concentration of benzyhydryl cations in solution with acetic acid is far too low for the reaction to work...

Metacelsus - 5-4-2017 at 19:45

Yes, I have read the paper. However, it is important to read papers before performing reactions, not afterwards when one is trying to figure out why a reaction failed. :(

I thought it would be a good idea to make a "public service announcement" to prevent others from making similar silly mistakes.