here we have two syntheses:
- a mixture of phloroglucinol, pyrocatechol and resorcinol by the action of molten NaOH on phenol
- phloroglucinol by the action of molten NaOH on benzenetrisulfonic acid
phloroglucinol is a precursor for TMA-6 and pyrocatechol for the methylenedioxy family.
the article was digged up by lugh who also corrected my most blatant mistakes,
so give him the credits
Berichte d. D. chem. Gesellschaft Jahrg. XII
104. L. Barth and J. Schreder: On the effect of molten sodium hydroxide on phenol and the synthesis of phloroglucinol
Some years ago, as one of us studied the effect of molten potassium on benzoic acid, he discovered that when replacing the potassium by sodium the result was different. This discovery was not pursued. Recently, when we explored the products given by phenol in molten potassium, the idea to explore the action of molten sodium on phenol suggested itself; the results obtained are surprising and show that, in this case too, potassium and sodium have a blatantly different effects, like in other experiments performed by Kolbe, Ost, and others.
When one melts phenol with 6 equivalents commercial sodium hydrate(1), one discovers that even at high temperature the thus formed phenol-sodium swims in the form of an oily layer on the molten sodium hydrate. But little by little foaming of hydrogen starts, which becomes stronger and transforms the melt into a brown homogeneous fine-bubbled mass. After some time, when the foam starts to sink, the fire is removed. The melt is poured into dilute sulfuric acid and the notable amount of black crumbly mass is separated by filtration. The filtrate is extracted multiple times with ether and yields after distillation of the ether a syrupy mass which forms crystals after a little while. It's weight was about 20% of that of the used phenol. In order to remove potential acids, the mass is diluted with water and washed with barium carbonate. After newly extractions with ether, nearly everything went into the ether and only very small amounts of barium salts were left in the aqueous phase. The latter was coloured dark brown and worked up like described later.
After distillation of the ether and standing for several days, the etherical solution of the main products precipitated many crystals which where vacuum filtered using a Bunsen pump. The crystals such obtained where purified by multiple recrystallisation. The first crop of plate shaped crystals was contaminated with a small amount of fine, flat needles, whereas later crystallisations were free of them. The needles couldn't be obtained in there pure form even after multiple recrystallisations and there were so few of them that analysis was not possible. The major product, which crystallised in form of the mentioned plates was determined to be phloroglucinol. All physical properties and the composition of the material prove this to the fullest. The aqueous solution becomes blue-violet on addition of ferric (or ferrous) chloride and shows the sensitive reaction, quoted by Weselsky(1b) and more recently by Wiesner(2b), with exceptional precision. The taste is pure sweet. It contains 2 molecules crystal water and gave on analysis the following numbers:
Found Computed for C6H6O3
C 56.78 56.80 57.14
H 4.60 4.79 4.76
Computed for C6H6O2 + 2H2O
H20 22.47 22.32 22.22
For the purpose of comparison we synthesised very pure phloroglucinol from Maolurin and compared it very thoroughly with the one made from phenol. They were exactly identical. Older references site the melting point of phloroglucin as 220°, we found for our preparation as well as for the totally clean and colourless one from Maolurin a melting point of 206° (uncorr.) (3b) and noted this for the correction of the older references.
The filtrate from the vacuum filtration of the phloroglucinol was steam distilled from a retort until the residue got a dark brown colour. With this method a separation of the left over substances was not possible. To make a long story short, apart from unreacted phenol the distillate contains lots of pyrocatechol and resorcinol, the residue contained the same compounds and considerable amounts of phloroglucinol. Moreover the residue contains a viscous not crystallising oil that distills only beyond the boiling point of mercury. By precipitation with lead acetate and fractional crystallisation, pyrocatechol, resorcinol and phloroglucinol were separated. The cleaned dihydroxybenzenes where identified by melting point, reaction with iron and more. Further each of the compounds was analysed.
1) Compound precipitable with lead acetate (pyrocatechol), crystallises without water in rhombic prisms, melting point 102°. The analysis gave the following numbers:
Found Computed for C6H6O2
C 65.25 65.45
H 5.39 5.45
2) Compound not precipitated by lead acetate and thoroughly separated from the phloroglucinol (resorcinol) with melting point 109° shows the following composition:
Found Computed for C6H6O2
C 65.45 65.45
H 5.16 5.45
The above mentioned brown residue of the steam distillation was standing a long time until the crystallisable compounds precipitated and could be removed. Finally the residue was distilled. From the fraction boiling up to 300° pyrocatechol and resorcinol could easily be isolated, then beyond 360° distilled an oil that became very thick and viscous on cooling. This oil could be separated using lead acetate in a precipitable and a non precipitable part. After unleading and evaporating to dryness both parts were redistilled and again yielded viscous oils without any trace of crystallisation. Pyrocatechol was identifiable in one part, resorcinol in the other. A complete separation was not possible. The oils are soluble in water, only the part that was not precipitable by lead acetate contained a little amount of water insoluble compounds. Because of the presence of pyrocatechol and resorcinol and the inability to crystallise those, we refrained from the analysis of the composition. We suspected that these oils are diphenols, probably different from those in the potassium melt. For further evidence we only tried the zinc powder reaction. Indeed both oils gave a certain amount of diphenyl, but also other hydrocarbons. Especially the oil which was precipitable with lead acetate gave a residue with a much higher melting point than diphenyl. Due to the little amount, a separation was not possible. Anyway, the property of these high boiling compounds being water soluble is highly remarkable. The total amount of them was 2%.
If one ignores the black amorphous compound that precipitates during the acidification, and only takes into account the easily characterisable compounds, one notes that the oxidation products far outweigh the condensation products. While we can only give approximate relative amounts of the oxidation products, due to the bad separation, we believe that we are not far from the truth when we state for the mixture of pyrocatechol, resorcinol and phloroglucinol a total amount of 10%-15% which is about evenly distributed among those.
The little amount of acids that were extracted as barytes, were extracted with ether after decomposition with sulfuric acid. After distillation of the ether and dissolving the residue in water, one obtains a brown, slowly decomposing solution.
When adding lead acetate, this solution gives a dark precipitate which, after removal of the lead with hydrogen sulfide, gives a dark, not discolourable liquid, which gave on evaporating black crusts. Because the acid was not obtainable in this way and alkali and earth alkali got even more dark, it was neutralised with cadmium carbonate and the filtrate dried under vacuum. But thus as well one only obtained an amorphous, black-brown residue which seemed not suitable for analysis. The amount was very small (0.7g - 0.5g from 2000g phenol). The part that was not precipitable by lead acetate was unleaded and evaporated in order to remove the acetic acid. When the small residue (at most 0.5g - 0.6g) was heated further on water bath, a whiff of fine needles was noted on the edge off the dish. The aqueous solution gave a positive phloroglucinol reaction, which was explained by the fact that the last traces of phloroglucinol can only hardly be extracted with ether from neutral or slightly basic aqueous solutions. Separating the acid from the phloroglucinol was not possible with such small amounts.
The whole treatment of the original ether extract with barium carbonate thus seems unnecessary, because it is an additional steps and only seems to reduce the yield of the air sensitive polyphenols.
The initially mentioned black crumbly mass which precipitates when pouring the original melt into sulfuric acid, amounts to 40% of the used phenol. On further melting with sodium hydroxide it gives little amounts of the described products. The bigger part precipitates already during the melting in black, cloggy lumps, which stay undissolved during acidification. This charred looking mass dissolves to big parts in alkali, but is in no way transformable into a form suited for analysis. On heating with zinc dust some diphenyl is formed.
Like previously noted, the reaction of sodium hydroxide melt on phenol seems to be mostly an oxidation analogous to the formation of paraoxybenzoic acid from benzoic acid, of diverse purpurines from certain bioxyantrachinons, etc. What's new, at least as far as we know, is the simultaneous introduction of two hydroxyls shown by the formation of phloroglucinol.
The interesting fact about this synthesis of phloroglucinol from phenol, apart from maybe being a nice way of producing this only hardly obtainable compound, is that it's the first direct proof of the aromatic nature of this so very common compound in the plant world.
We can complete this proof by another synthesis of phloroglucinol which is, from a theoretical point of view, even more interesting and again shows the difference of the effects of potassium and sodium.
In benzenetrisulfonic acid, a substance described by Senhofer, when reacted with potassium melt, as described by the same(1c), with increasing temperature, first one, then two SHO3 groups are replaced with hydroxyl groups. One thus obtains partly a new (beta)-phenoldisulfonic acid, and partly a dihydroxybenzenemonosulfonic acid; the third sulfuric acid rest could not be replaced by a hydroxyl group in potassium melt, the substance was nearly entirely scorched.
A sodium hydroxide melt behaves differently. When the trisulfonic acid is heated with an excess of sodium hydroxide, in a little while foaming sets in, which gradually gets stronger, so that heat must be controlled in order to avoid a boil over. The melt is maintained for a short time, and after cooling the white cake is poured into dilute sulfuric acid. The mass dissolves nearly without any clouding and with development of copious amounts of sulfurous acid. After cooling the precipitated sodium sulfate is filtered off and the filtrate is extracted multiple times with ether. After evaporation of the ether one is left with a yellowish crystalline mass which is purified by recrystallisation and discolouration with animal charcoal. The resulting crystals were colourless, had the form of phloroglucinol, the sweet taste, all the characteristic reactions as well as a melting point of 206°. When air dried they contained 2 molecules crystal water which disappears at 100°. Analysis gave the following numbers:
Found Computed for C6H6O2
C 56.98 57.17 57.14
H 4.97 4.90 4.76
Computed for C6H6O2 + 2H2O
H2O 22.27 22.23 22.22
The yield of phloroglucinol is 25% - 30% of theoretical. Other solid products could not be isolated. It seems that sodium hydroxide replaces all sulfo groups at the same time by hydroxyl groups although a big amount of the trisulfonic acid is burnt during the energetic reaction.
The disclosed results demand an in-depth research of the sodium melt with the prospect of new and interesting facts. - The experiments will be continued.
Vienna, 1st University laboratory.
1) We used ample silver dishes and 100g phenol per melt.
1b) These Berichte XI, 216.
2b) Sitzungsber. der Wiener Akademie B. 77, 1. Abth. Januarheft 1878
3b) Not entirely pure phloroglucinol darkens at the quoted temperature, gets soft and an exact melting point is not determinable. Sometimes it seems to only melt completely at 210°-212°.
1c) Sitzungsber. d. k. Akad. Wien Bd. 78, Abth. 2, Oktoberheft 1878