Triphenylpyrylium bisulfate

Pyrylium is the name of a cationic heterocycle, a benzene with one =CH- replaced by a positively charged oxygen =O(+)-. Despite looking quite unstable, it is actually part of many plant dyes, particularly ones that have a flavylium structure. The oxygen in the pyrylium salt can be replaced by various nucleophiles and thus offers great synthetic applicability. One famous example is the synthesis of azulene and azulene derivatives (although this derivative presumably doesn't work for that).

2,4,6-Triphenylpyrylium bisulfate
There are many salts of the 2,4,6-triphenylpyrylium cation known. The most common might be the tetrafluoroborate and the perchlorate.^[1] However HBF₄ or BF₃ are difficult to acquire and quite dangerous, and HClO₄ is as well, and on top of that the pyrylium perchlorates are potentially explosive. A good alternative is the bisulfate.

This synthesis is based on a US patent.^[2] There are procedures that first prepare chalcone from acetophenone and benzaldehyde^[3] and then react that with more acetophenone and sulfuric acid.^[4] The yield given is somewhat variable, and while this two step procedure might save a little bit on sulfuric acid, the overall yield based on acetophenone and benzaldehyde is almost the same as the one-pot reaction.


Fig. 1: Reaction scheme.

Synthesis:
A 250 mL round bottom, single neck flask was placed in an ice bath and a stir bar, 22.3 g (0.186 mol, 1.47 eq.) acetophenone, and 13.4 g (0.126 mol, 1.00 eq.) benzaldehyde (Note 1) were added and precooled. From an addition funnel 20.0 g (95-97 %, ca. 0.194 mol, ca. 1.53 eq.) of sulfuric acid were added during 14-15 minutes. The solution changes from colorless to yellow to red and becomes somewhat viscous.


Fig. 2: Color progression.

A drying tube with CaCl₂ was placed on top and the flask was placed in an oil bath, the oil was then heated to 100-105 °C. The solution becomes redder and less viscous during the heating. Heating is continued for 6 hours. The flask was then taken out, left to cool somewhat and 120 mL of water were added. The mixture is then boiled briefly and a small amount of oil is seen on top of the water layer. It is then gravity filtered through some filter paper containing a small amount of celite. The mixture is then stirred at room temperature until crystallisation begins and placed over night in the fridge. Then it was filtered with suction and washed with ca. 50 mL water. It was then dried at 65 °C for 10 h.



Fig. 3: Workup sequence.

Yield: 15.02 g (0.037 mol, 60 % of the theory, literature: 13.9-15.1 g) of slightly inhomogenous yellow-orange product. (Note 2)

From the various filtrates could be precipitated with 14 % sulfuric acid an additional 0.4 g of product.

Notes:
1. According to the patent,^[2] the benzaldehyde should be free of benzoic acid. My benzaldehyde was a several years old commercial sample, which however after shaking with water gave a pH of 7 in the aq. phase, and was used without further purification.
2. According to the literature, the product can be further purified by washing with acetone and recrystallizing from ethanol or ethanol containing some sulfuric acid.^[4] The product in this state worked fine for making a solvatochromic betain dye and for making 2,4,6-triphenylnitrobenzene, however.

Attempted purification:
First let me state that this was a failure and I do not recommend it. I was unable to get the bisulfate back and had to convert it into the less soluble perchlorate in the end. The raw product was used successfully in other reactions, so I'd just use it as is. Anyways, here is what I tried:
8.25 g of the crude bisulfate was dissolved with 150 mL of boiling ethanol, then 3 mL of 40 % sulfuric acid was added, which seemed to facillitate dissolution. No crystals were obtained after cooling. Neither addition of water nor scraping or cooling induced crystallisation. The solution was boiled down further to a volume of 30-35 mL, in the freezer some crystals started forming. They were filtered off and were a lemon yellow color, but during drying it turned orange in some spots and all in all looked like before. The amount recovered was 4.52 g.

From the residual solutions the perchlorate was precipitated by addition of a sodium perchlorate solution (in excess), which immediately precipitated a fine lemon yellow precipitate. After air drying for several weeks, the mass was 2.51 g.


Fig. 4: Perchlorate salt in suspension.

It was also attempted to basify a sample with KOH (orange and clumpy) and extraction with ethyl acetate, after separating the organic layer it was reacidified with dilute sulfuric acid and the ethyl acetate boiled away, no crystallisation occured but fine yellow crystals were obtained on addition of ice. They had a visibly impure appearance however and a tendency to turn back into an oil, so this method was also discarded.

Analytical results:
TLC was attempted in hexane/ethyl acetate 2/1 (A) as well as acetonitrile/water/potassium nitrate 10 mL + 1 mL + 30 mg (B), once with the triphenylpyrylium bisulfate spotted in ethanol and once spotted in ethanol (1) with a few drops of 40 % sulfuric acid added (2). There were problems with heavy streaking and the addition of acid drastically changed the elution behaviour, so there seems to be some reversible decomposition going on. The yellow spots (with blue fluorescence) turned first orange and then somewhat faded in an atmosphere of ammonia and their fluorescence disappeared. The results for the different combinations are listed below:
A-1: Two spots at hR_f = 72 and 87, both with blue fluorescence under 365 nm UV and yellow appearance under visible light.
A-2: Everything remained on the starting line (hR_f = 0).
B-1: Two spots at hR_f = 62 and 99 (heavy streaking), along with some residue on the starting line.
B-2: Two spots at hR_f = 0 and 57 (main part) with some smaller amount running with the solvent front.

Overall, I wouldn't put too much weight on the TLCs. They were more confusing than actually helpful. Nonetheless, that the product actually is the triphenylpyrylium, is certain.

Literature:
[1] - (a) NI2, "Synthese von 2,4,6-Triphenylpyryliumtetrafluoroborat", Illumina-Chemie 2013, https://illumina-chemie.org/viewtopic.php?p=51498, (b) K. Dimroth, C. Reichardt, K. Vogel, "2,4,6-Triphenylpyrylium tetrafluoroborate", Org. Synth. 1969, 49, 121, https://doi.org/10.15227/orgsyn.049.0121
[2] - T. C. Chadwick, "Water-soluble tetraphenylpyridinium salt for use as an analytical reagents", US4150233, https://patents.google.com/patent/US4150233A/en
[3] - E. P. Kohler, H. M. Chadwell, "Benzalacetophenone", Org. Synth. 1922, 2, 1, https://doi.org/10.15227/orgsyn.002.0001
[4] - (a) T. C. Chadwick, "Gravimetric determination of gold and platinum using 2,4,6-triphenylpyrylium chloride", Anal. Chem. 1974, 46, 9, 1326-1328, https://doi.org/10.1021/ac60345a017, (b) A. Dinculescu, A. T. Balaban, "2,4,6-Triphenylpyrylium hydrogen sulfate", Org. Prep. Proced. Int. 1984, 16, 6, 407-410, https://doi.org/10.1080/00304948409458669

Diachrinic I found an experiment which uses your product as a substrate to synthesize 2,4,6-triphenylpyridine.
In RBF equipped with reflux condenser heat 4 g of 2,4,6-Triphenylpyrylium perchlorate (I hope your hydrogensulfate would work too, just maybe adjust the amount of ammonium acetate?), 7-8 g of ammonium acetate and 30 ml of glacial acetic acid for 30 minutes. After cooling pour the mixture into 150 ml of water. After few hours filter, wash few times with water, air dry or in vacuum desiccator. White crystals m.p. 138 C (from alcohol). Yield 2,7 g (90%).
It is a russian book about heterocycles
http://www.newlibrary.ru/genre/nauka/himija/organicheskie_si...
Пожарский А.Ф., Анисимова В.А., Цупак Е.Б.
Практические работы по химии гетероциклов
http://www.newlibrary.ru/read/pozharskii_a_f___anisimova_v_a...
you should correct the filename after downloading, add it djvu extension and then use a tool to read this very old format



[Edited on 13-3-2026 by Fery]