Sciencemadness Discussion Board

Benzylation of benzene with benzyl alcohol

kmno4 - 20-7-2011 at 13:05

I am looking for informations about benzylation of PhH with PhCH2OH and H2SO4 as "catalyst".
I have read old article from Monatshefte für Chemie (DOI: 10.1007/BF01521832) but they use surrealistic proportions:
300 cm3 of 70% H2SO4
20g of benzyl alcohol
30g of benzene
(Desired product is of course diphenylmethane)
Is it possible to use ~equimolar proportions alcohol-acid ?
If not, why ?

Nicodem - 20-7-2011 at 14:05

I would say the problem of using alcohols in the Friedel-Crafts alkylation is the water formed. Water is many magnitudes more nucleophilic than benzene and will thus compete in the reaction, slowing it down to a useless rate. Overstochimetric amounts of H2SO4, ZnCl2 or polyphosphoric acid are therefore commonly used.

It should be, and it is possible to use catalytic amounts of the acid if employing a Dean-Stark trap. If using triflic acid, tungstophosphoric acid or other such superacids probably less than a mol% should be enough. Apparently, Amberlyst works as a catalyst (DOI:10.1016/0923-1137(95)00019-F), and of course Nafion as well (DOI: 10.1021/jo00006a023, Bull. Korean Chem. Soc., 12, 463). Tosylic acid works as well with the Dean-Stark, but it requires larger amounts of this catalyst, about 50 mol% for a rapid conversion (DOI: 10.1021/ja01159a079). I guess you can use 10 mol% or even less, but the reaction will take much longer and some dibenzyl eter might remain unconverted. The good thing about the Dean-Stark is that you can monitor the reaction conversion in real time.

kmno4 - 20-7-2011 at 14:53

Thanks Nicodem !
Dean-Stark trap sounds good, but it is currently unavailable for me :mad:
However, formation diphenylmethane is practically irreversible reaction and even dibenzyl ether can be used for benzylation.
I am just surprised amount of acid solution from paper given earlier.
PhCH2OH is 0,19 moles, PhH is 0,38 moles
+ 3,5 moles of H2SO4 and 8 moles of water, and they still can obtain benzylation products !!
(amount of water is then practically constant)
It is amazing....
I am asking because I got very low yields of product, made with much less amount of H2SO4 (alcohol/benzene was 1/5 [mol/mol])

smuv - 20-7-2011 at 19:22

I wonder if you could do this with more of a catalytic amount of acid, if you set it up to be biphasic with a more organo-soluble acid. The solubility of water in benzene is very low (as opposed to the opposite) so this may help drive to equilibrium. Also, you can still use Lewis acids like FeCl3 or ZnCl2 for alcohols, I think something like 2eq Lewis acid, 1 eq benzyl alcohol excess benzene at reflux could accomplish the transformation (say lewis acid is AlCl3, ZnCl2, FeCl3, TiCl4).

Also...if you have access to lewis acids, you can easily make diphenylmethane from 1eq Benzene, 1eq AlCl3 in an excess of DCM. This works well even at ice bath temps. ZnCl2 doesn't work, IDK about other acids.

[Edited on 7-21-2011 by smuv]

kmno4 - 21-7-2011 at 00:19

H2SO4 is my choice. I have read about many variants of this reaction, with AlCl3, BF3 etc. Yields are very good.
Unfortunaley, during work-up, I get emulsions, both in organic and water phase, separating in hours :mad:
From the last experiment (30g benz. + 7,5 g benz.alc. + 1,5 cm3 H2SO4 + 3 hours reflux) I have turbid mixture with ~1 cm3 of clear dense layer. It is till turbid after 10 hours standing. I am going to reflux it (+ stirring) . If it will not help - another 0,5 cm3 portion of H2SO4. We will see ;)

smuv - 21-7-2011 at 21:06

What was the work-up? Did you dilute the reaction mix with water. I am surprised you are getting an emulsion.

kmno4 - 22-7-2011 at 11:25

I added additionally 2 x 0,5 cm3 of H2SO4, but nothing interesting happened. Just larger yellow-red layer of acid at the bottom.....
After long-lasting refluxing I mixed it with water. Again, slow separating emulsions. Water layer was separated and organic layer mixed again with water.... and again water. Total 4 washings and the last water layer was clear. Organic layer was turbid, with some strange foam (?) between organic and remaining ~1cm3 of water layer. This "foam" during ~12 hours standing did not change.
I took it all to the flask and distilled. About 25g of benzene + 1cm3 of water was recovered as distillate. In the end of removing benzene, something was happening in flask and water appeared again. Temp. of vapours was not stable, in the end I had 75-82 C.
After cooling, I got some shit in the flask: white-yellow foamy emulsion, with water. Droplets on the walls is also water.
Typical product :P
I suspect that benzyl acohol is converted to other products, instead of benzylation of benzene. Sample of product with added CrO3 + H2SO4 gives an once benzaldehyde smell. It seems also that alcohol is still present in mixture. Or maybe it is dibenzyl ether ?
Without Dean-stark trap it is hard to say anything about this reaction.
I added photo of flask with "product"

bbb.JPG - 21kB

smuv - 22-7-2011 at 21:45

You need to dry your organics before distillation. Also it might help to use a solvent for the extraction. The benzyl alcohol is probably the source of some of your problems, IIRC its an emulsifier.

BTW, did you stir this rxn?

Bolt - 22-7-2011 at 22:14

Can you run TLC? You will be able to see the benzyl alcohol spot disappear. The separation of an alcohol and an unsaturated hydrocarbon might be achieved using hexane as an eluent. You may be able to see impurities develop.

kmno4 - 23-7-2011 at 02:14

Quote: Originally posted by smuv  
You need to dry your organics before distillation. Also it might help to use a solvent for the extraction. The benzyl alcohol is probably the source of some of your problems, IIRC its an emulsifier.

BTW, did you stir this rxn?

I distilled all because it contained ~25g of benzene and ~1 g of water and water should have passed with benzene as azeotrope.
And it indeed happens, but water appears again in the end of distillation.
Some dehydration reaction(s) must take place at higher temperatures in flask. I realise that benzyl alcohol can act as emulsifier, but is shoud dissapear from reaction, even if not converted to diphenymethane. It should give some polycondensation products under this conditions. That is why I washed reaction mixtere with few large amounts of water - to remove remaining alcohol and - possibly - its "sulphonic products", being much better emulsifiers than alcohol itself.
Of course, during refluxing it was magnetically stirred.
Unfortunately it is not so simple and possibly alcohol is still present in my reaction mixture. I should be able to removed it by anh. CaCl2 - it is under investigation. But it does not look good :mad:
I hoped to find some application for my "useless" benzyl alcohol, but my procedure seems to be not good.
TLC yes, but not at this stage, where product looks very shitty at a glace.

smuv - 23-7-2011 at 19:56

Maybe a base wash. Hopefully this will avoid decomposition during distillation.

I don't know about benzenesulfonic acid being a good emulsifier though, it is way too polar to be one, only the much longer chain sulfonic acids are good emulsifiers.

[Edited on 7-24-2011 by smuv]

S.C. Wack - 23-7-2011 at 20:53

Cumming (library) has a couple routes. One uses even more H2SO4, and acetic acid, less benzene, and 25% yield. The one using benzyl chloride looks much more attractive.

Nicodem - 24-7-2011 at 07:18

Perhaps the emulsion is only the consequence of polymeric crap that precipitates and prevents the droplets coalescing. In such case vacuum filtration of the whole water diluted mess trough celite or silicagel should help. Alternatively, adding some CH2Cl2 might dissolve the crap and allow a proper phase separation needed for the washes prior the distillation.
If everything else fails, try with the extraction with CH2Cl2 from the reaction mixture without diluting it first (yes, extract directly from H2SO4), let the extract stir with Na2CO3, filter and vacuum distil. You might find it impossible to efficiently separate diphenylmethane from dibenzyl ether without fractionation with a distillation column.

kmno4 - 24-7-2011 at 14:42

I tried base washing at first, in earlier experiment. It caused nothing, but larger emulsion.
I added some CaCl2 and got two layers: turbid yellowish organic and CaCl2 sol. ones. More CaCl2 gave nothing, only crystals of hydrates appeared in CaCl2 layer. I can see now that this turbidity very, very slowly dissapears. It seems that it is not caused by water or any ionic substance.
Organic layer has hard to describe smell, something flowery + benzyl chloride, but without tears and gives blue fluorescence under UV. I cannot sense benzyl alcohol , but it means nothing (because of its very weak smell).

I am trying to order Dean-Stark trap, but in this damned country I live, email exchange is very long procedure :mad: Just they do not respect the clients.
The Korean paper mentioned by Nicodem looks very sweety....
D-S trap would be much better than silica gel from that paper.

[Edited on 24-7-2011 by kmno4]

kmno4 - 26-7-2011 at 01:00

I have just distilled my "product".
Distillaton has started at ~210 C (not counting few drops of colored benzene passing first) with continuously rising temperature.
At ~270 C distillation has been stopped.
Input: ~14 cm3, ~3cm3 of distillate, remaining ~10 cm3 in flask.
No diphenylmethane :D
Finally I ordered D-S trap.

Nicodem - 26-7-2011 at 10:15

While you wait for the DS trap, you could use your benzyl alcohol in other reactions, possibly requiring less tedious work up. You can use benzyl alcohol in H2SO4 to N-benzylate acetonitrile or some other relatively non-basic nucleophile via SN1. Just as an example, with acetonitrile you would get N-benzylacetamide which can be hydrolyzed to benzyl amine.

kmno4 - 26-7-2011 at 14:28

Nicodem - bowings to the clairvoyance.
I have read about it recently (some JOC or JACS paper). This is next in the queue.

smuv - 26-7-2011 at 20:25

AKA ritter reaction, to help you look up refs...

kmno4 - 2-8-2011 at 04:53

I am still waiting for damned D-S trap :mad:
But in the meantime, I did some literature and experimental researches.
I turns out, that sulfuric acid cannot be used in this reaction when its concentration exeedes some value, probably 70-80%.
If concentration is higher, it forces benzyl alcohol to alkylate itself in the ring (well known "polybenzyl" ). Apparently, mechanism of reaction changes if more concentrated H2SO4 is used.
In the case of H2SO4, D-S trap is useless.
I am going to use p-toluenesulfonic acid or some H-resin.
Simple preparation of such resin from polystyrene and H2SO4 can be found in J.Chem.Ed. paper (DOI: 10.1021/ed066p613) and many other papers.
Also H3PO4 seems to be good candidate. 75% acid gives almost pure dibenzyl ether (strongly refractive liquid), reaction mixure is almost colourless and no problems with emulsions.
(emulsions are also mentioned in old patent (GB446450A) - they use H2SO4 "diluted" with ethyl alcohol (of course it gives ethyl sulfuric acid)).
It seems that these emulsions are really caused by Nicodem's polymeric crap. "Polybenzyl" is just benzyl alcohol polycondensate (with low molecular weight and OH groups attached at the "ends" of molecule)

Nicodem - 2-8-2011 at 10:25

Quote: Originally posted by kmno4  
I turns out, that sulfuric acid cannot be used in this reaction when its concentration exeedes some value, probably 70-80%.
If concentration is higher, it forces benzyl alcohol to alkylate itself in the ring (well known "polybenzyl" ). Apparently, mechanism of reaction changes if more concentrated H2SO4 is used.

That's quite unusual. It might have to do with phase partition behaviour, else I can not explain why the selectivity for self alkylation would be increased in more concentrated H2SO4. Benzyl alcohol should be better partitioned in conc. H2SO4 than benzene, so it can selfalkylate in the H2SO4 phase (assuming there are two phases at all). Though, I don't know how come 70% H2SO4 would so radically change this. There must be some other explanation. Can you post the reference where this was reported?
Was your reaction with conc. H2SO4 a solution or was it still biphasic? (I assume it was concentrated H2SO4 that you used?)
At least this explains the need for such huge excess of H2SO4.

kmno4 - 2-8-2011 at 14:09

The only true reference is in my flask. I could not find anything about BA in available to me literature.
But there are papers, for example (DOI):
about changes in products, in range of 70...80-98% H2SO4 used.
For some reasons in Monatshefte für Chemie paper author use 70% not 96-98% (=concentrated) H2SO4. Also commonly known is effect of mixing H2SO4 with BA (=C6H5CH2OH).
In my experiments I used conc. H2SO4 (about 96,5% from old titration). On 10 g of benzene scale, adding less than 0,5 cm3 H2SO4, with cooling, gave clear solution. During reaction mixture became non-homogeneous, with milky layers after standing. Refluxing gave more colored mixtures than not refluxed ones, but "products" were ~ the same.
Solubility of benzene in 70% H2SO4 is less than 1g per 1000cm3 of acid, but starts to rise above "magic" value 75% (DOI: 10.1002/recl.19650840414)

kmno4 - 6-10-2011 at 00:01

Finally - I used up my ~300g of my benzyl alcochol for experiments (using D-S trap) :D
Results will be posted soon.

kmno4 - 26-10-2011 at 10:08

All preparations were conducted with ~100 g of benzene (BN) and 15 g of benzyl alcohol (BA), with aid of Dean-Stark trap. Theoretical amount of mater, corresponding to 100 % (=23 g) yield of diphenylmethane (DPM) is 2,5 g. Sulfuric acid (SA) was used for all preparation.

4,8 g of 85% SA (arbitrary amout and concentration) was added and mixture was refluxed (with pieces of porcelain only) without stirring. Water slowly, very slowly collects in the trap:
2 hours (h) &#8594 1,4 g of water
8 h &#8594 1,9 g
24 h &#8594 2,6 g
....and a little longer &#8594 2,7 g
About 4,8 g of red coloured SA was carefuly pipeted out, ~0,5 g of MgO was added and BN distilled off (to ~100 C in vapour) .
After separating of MgO, about 20g of residue as orange oil was left.

9,6 g of 85% SA (rest the same as in 1)
12 h &#8594 1,3 g
27 h &#8594 2,2 g
10,2 g of red SA, about 20 g of residue.

The same as 1. but magnetically stirred, without porcelain. Subsequent experiments were conducted only with magn. stirring.
0,5 h &#8594 1,3 g
2 h &#8594 1,6
9 h &#8594 2,8 g
4,5 g of red SA, about 20 g of residue.

Residues were combined and distilled:
To ~110 C some (about 5-7 cm3) stinky fraction (but mainly benzene) collects, temperature slowly rises to ~140 C and suddenly jumps to ~250 C. DPM starts to collect (from ~260 C) but temperature does not hold at 263 C (B.P. of DPM) but slowly and continuously rises. Distillation was continued until temp. reached 290 C. Even at this temp. some DPM collects.
About 16 g of clear, water-like DPM fraction was collected and about 34 g of residue - viscous, orange, nicely smelling oil.
Yield of DPM: about 20% - not worth my work and chemicals.

Which way to go ?

Equimolar (counting on BA) amout of 85% SA was used.
10 h &#8594 2,2 g
13 h &#8594 2,5 g
15 h &#8594 2,7 g
16 g of residue, looking very shitty - wrong way.


The same molar amount of 75% SA (=5,45 g) was used instead of 85%. The rest as in experiments 1-3
0,5 h &#8594 1,5 g
1 h &#8594 2 g
2 h &#8594 2,6 g
Colours are similar but less orange than earlier; 5 g of red SA; BN was distilled off a little deeper (vap. temp. 110 C); 19-20 g of residue

The same as 5, but refluxing was longer:
10 h &#8594 3,2 g
Appearently some water is removed from 75% SA
19 g of residue, as earlier.

Residues from 5 and 6 were combined, distilled.. etc.
Yield of DPM: damned 20 %

75% H3PO4 instead of H2SO4 (molar amount the same)
2 h &#8594 2 g
6 h &#8594 3,4 g
Amount of pipeted off H3PO4 was small, less than amount of pure H3PO4 in 75% solution taken to reaction. Dehydration goes to more condensated phosphoric acids, H4P2O7 etc.
18 g of "white" residue, not orange. Possibly it is impure dibenzyl ether - soluble in hot C2H5OH and separates on cooling.


Three drops of conc.SA were added. Almost nothing, small amount of water collects...
Very slow reaction. Additional 3 drops of ~70 % SA... nothing.... next 3 drops...
10 h &#8594 0,5 g
... another 3.... another ones...
18 h &#8594 1,4 g
After adding ~25 drops of acid reaction is much faster.
20 h &#8594 2,4 g
Refluxing mixture is homorgenous, but when it comes to the end (~2,5 g of water) acid starts to separate as red layer appears at bottom of flask.
Next tandard procedures...
21 g of orange-yeloow residue. More shit or more desired product ??

The same as 8, 85% acid added in portions, firts one was 18 drops (0,68 g)
2,5 h &#8594 0,3 g
Effects very similar to 8.
A little more than 23 g of coloured residue.

The same, but 6 drops of conc. SA were added at the beginning, and wait...
11 h &#8594 0,1 g
More acid was added in subsequent portions.
22g of coloured residue.

Residues were combined and distilled. Some (3-4 cm3) stinky forerun, temperature jumps to 260 C and practically stops at 262 C . Mixture foams terribly in flask, no more than 1 drop of distillate can be taken per 10 seconds - foam keeps milimeters from the inlet of air cooler. Temperature slowly rises, foam dissapears, distillation ceases. At 285 C almost nothing, one drop per minute.
Yield of DPM: 57 % - not very good but better than 20 %....
Residue in flask is orage coloured but, after cooling, almost odourless.

Experiments 8-10 were repeated (11-13) with different amounts of starting acid, up to 25 drops.
Yield was exactly the same: 57%
Total amount of added acid was less than 1g and refluxing mixture was homogenous.
Smaller amounts give very slow reaction, larger give non-homogenous mixture and much less amount of DPM.
Striking thing - amount of SA is very critical and it does not matter if it is concentrated or 75%.
Possibly yield would be even better if ratio BN/BA is higher (as in mentioned earlier papers with p-TsOH methode).
I did not try TsOH because its cost is comparable with DPM and its separation requires additional time and work.

There was no tars, carcoal-residues etc. during DPM distillation. Polymeric substance (residue) is thermally stable up to at least 300 C. It may also contain various dibenzylbenzenes, but I had no
conditions to investigate it.

Foto 1, from left:
Raw DPM (liquid, slightly overcooled) from 8-13, residue from distilling off DPM, raw DPM (crystallised) from 1-3,5,6 and residue from distillations.


Foto 2
Nice needle-like crystals in partially molten DPM. Very similar to crystals of KNO3.


Nicodem - 5-11-2011 at 10:26

You did a very nice reaction conditions screening. It demonstrates that even with extremely limited analytical tools, one can still do some interesting research. It is truly a pity you don't have access to a GC (I assume), as that would explain a lot more. It is quite possible that with the more concentrated H2SO4 there is a delay in water accumulation (due to the extreme hygroscopicity of H2SO4) and that the peak of conversion to diphenylmethane is actually achieved relatively fast, but further reflux to achieve the theoretical amount of water only leads to crap formation due to Friedel-Crafts disproportionation / oligomerisation (thus reduced yields).

You can also do some related alkylations of benzene or toluene with your setup. For example, you can try a t-butylation with t-butanol or MTBE (t-butylations of toluene should be fairly para selective, but I have not checked the literature).

Do you have any further plans for diphenylmethane? You can try an alpha-chlorination using the so much available trichloroisocianuric acid, or you can try an oxidation to benzophenone with diluted HNO3 or KMnO4, etc. I would consider it quite an interesting starting material to play with, though electrophilic aromatic substitutions would be a bad choice due to lack of selectivity (two equal phenyls, orto/para selectivity).

kmno4 - 5-11-2011 at 11:46

Quote: Originally posted by Nicodem  

Do you have any further plans for diphenylmethane?

Yes, that is why I bodered doing this at all. The goal was relatively cheap and pure DPM as starting material. ~60 % yield is not bad, especially if its separation is easy - no messy extractions... etc.
Next step is benzophenone.
I am going to create another topic: "Oxidation of diphenylmethane to benzophenone", and post some experimental results. Unfortunately - not very interesting :P

I tried experiment with n-buthyl alcohol.
Reaction is not similar.... and I had big problems with recovery of benzene from mixture :mad:
It is because I got some "mysterious" benzene azeotrope with b.p. 79,0 C. I disstiled it 3 times - from the very beginning to the very end: damned 79 C not 80 C.... Nevermind.

kmno4 - 6-11-2011 at 10:48

Quote: Originally posted by Nicodem  
It is quite possible that with the more concentrated H2SO4 there is a delay in water accumulation (due to the extreme hygroscopicity of H2SO4) and that the peak of conversion to diphenylmethane is actually achieved relatively fast, but further reflux to achieve the theoretical amount of water only leads to crap formation due to Friedel-Crafts disproportionation / oligomerisation (thus reduced yields).

I have different impressions. Seems that first , small amounts of H2SO4 gives only (mono)benzyl sulfate + corresponding, small amount of water and nothing more. At some critical benzyl sulfate concentration, something happens and benzylation starts, with releasing H2SO4. It is just speculation.
Some papers claim isolation of similar benzyl tosylate if TsOH is used instead of H2SO4. However, it does not explain this "critical" amount of H2SO4.
Beside, rate of water formation at initial stage (few drops of acid added) does not depend on concentration (at least in 75-96% range) of this acid.

Nicodem - 6-11-2011 at 13:26

I will confide to your subjective impressions, due to lack of chromatographic evidence.

Benzyl tosylate should indeed form from benzyl alcohol and TsOH under such reactions conditions via the same SN1 mechanism that gives also diphenylmethane, but that ester is not the most active benzylating species. The benzylating reagent is the protonated alcohol and the transient carbocation it can form, as is generally the case with Friedel-Crafts alkylations using alcohols. This is why only benzylic, allylic and t-alcohols can be used (exceptionally s-alcohols work as well, as is the case in the synthesis of cumene from benzene and isopropanol). Just refluxing benzyl tosylate in absolutely dry benzene is unlikely to give you much diphenylmethane in the same reaction time as would benzyl alcohol with TsOH catalysis. All this is said just in case you believe that benzylation necessarily occurs via a sulfate or sulfonate ester. This is why your experiment with n-butanol gave no desired product, even though under the applied reaction conditions the formation of butyl sulfates is possible (but the formation of a carbocation under these conditions is not possible).

How about posting a detailed report about your synthesis of benzophenone, complete with references, also in the Prepublication section?

kmno4 - 6-11-2011 at 15:55

Oxidation of DPM to benzophenone will be posted later, because I still cannot find good procedure. I have just made some SciFinder researches but among 100 references I do not see any good one :(
Good = conversion not less than 90%, selectivity for benzophenone not less than 95%, cheap and fast.
The simplest methode - 10-20% HNO3 under long lasting reflux gives yellow mass. It is said that such oxidation gives only benzophenone and alpha-nitro-DPM (reference unavailable and in Russian...)
CrO3 in AcOH seems to be OK, but rather large amounts of acetic acid are used for relatively small amounts of DPM (ex. 50 ml AcOH per 5 g DPM). I would try KMnO4 in AcOH, but I am affraid to get BUM instead of oxidation. In boiling acetone KMnO4 gives selectivity for benzophenone near 100%, but conversion DPM is only 60% (...who knows why ?). Besides such oxidation has unknown (for me) stoichiometry.

Back to benzylation.
I did not say that that with n-butanol I got nothing.
It was just a trial run, to see what happens with such a primary alcohol. I would choose EtOH, but obviously it is not recomended (beacuse of experimental reasons).
H2SO4 (conc.) was adding in portions, one portion was about 10 drops. Some acid was added, some water collected and nothing more. More acid - more water.. Finally I stopped counting amount of acid.... I got some high boiling residue with toluene odour. I do not know what was that, I would have to repeat this experiment with known amount acid added.
As I said, recovered benzene had b.p. 79 C. To remove unknown impurity, I tried few "tricks". The only thing that worked was aqueous CrO3 + H2SO4 (with mag.stirring). Impure benzene became slightly turbid, then yellow, orange and... colourless and clear :o
Interesting - there was no butyric acid smell, so this impurity was not butanol or its ether.
Redistilling of this purified benzene gave normal b.p. 80 C.
I will try this one more time ( or two to collect more eventual products) with 20g of alcohol per 100g of benzene and slightly more than 1 molar equivalent of H2SO4.

Satan - 6-11-2011 at 16:22

I have an idea about oxidation of DPM. Exhaustive chlorination on -CH<sub>2</sub>-, and then hydrolysis to ketone with aqueous base. Slow addition of TCCA in CCl<sub>4</sub> (or DCM etc.) might work - just an extrapolation of other chlorinations involving benzylic carbon :cool:

kmno4 - 7-11-2011 at 13:48

If my experiments with direct oxidation will fail, then I will try this or similar methode. In this way benzophenone is prepared in OS procedure.
But now I have a few strong oxidants for testing......