Tricyanomethane and its Salts with Nitrogen Bases: A Correction of Sixteen Reports

Abstract

A series of 16 articles, dealing with formation of tricyanomethanide salts from nitrogen bases (amines) and tricyanomethane, turned out to be wrong because no tricyanomethane or similar compounds are used and no tricyanomethanide salts are prepared. This statement is based mainly on the recently published spectroscopic data and depicted spectra, which are completely incompatible with the claimed structures. Some of the tricyanomethanide salts are now synthesized from nitrogen heterocycles and tricyanomethane or other precursors. The corresponding products show plausible spectroscopic data and physical properties, which are entirely different to those previously reported.

Key words

exclusion of structures - spectroscopic characterization - tricyanomethane - tricyanomethanide salts - unstable compounds

Scheme 1 Reported syntheses of cyanoform (5)

First attempts to prepare tricyanomethane (5), also called cyanoform, were reported as early as 1864,[1] but the syntheses could not be reproduced by other authors,[1c] [2] and later it turned out that 5, even if formed, would not have been able to survive the reaction conditions employed. In 1896, Schmidtmann treated sodium tricyanomethanide (1a) with dilute sulfuric acid, and, after addition of diethyl ether, he obtained a three-phase system, which included a greenish yellow middle layer 2 (Scheme [1]).[3] This layer was claimed to contain 5, however, first experiments to remove the solvents did not lead to characterizable products.[4] In 1963, Trofimenko utilized 1b, repeated the synthesis of 2, and called the product ‘aquoethereal cyanoform’.[5] He was successful in liberating 2 from the solvents by rapid evaporation and sublimation and obtained unstable white crystals, to which he erroneously assigned the structure of ketenimine 3. By treating 2 with tertiary amines, such as triethylamine or pyridine, or sterically shielded primary amines, Trofimenko et al. prepared ‘amine salts of tricyanomethane’.[6] Dunitz et al. performed rapid evaporation of 2 without sublimation and isolated single crystals of hydronium tricyanomethanide (4) as confirmed by crystallographic X-ray diffraction studies.[7] In 2017, it turned out that ­Trofimenko’s experiment did not lead to 3 because rapid evaporation and sublimation of 2 actually resulted in the isolation of a mixture of 4 and 5.[8] Single crystals of cyanoform (5) were conveniently available under special conditions of vacuum sublimation and allowed structure verification by X-ray diffraction. Such single crystals could be handled at room temperature for a short time if moisture was excluded. However, 5 could not be characterized in solution by NMR spectroscopy at ambient temperature because of rapid decomposition.[9] Furthermore, 5 was easily transformed into 4, even when only trace amounts of water were present. In particular, mixtures of 4 and 5 underwent dynamic processes leading to extremely broad NMR signals at temperatures between −50 and 0 °C.[9] Consequently, even lower temperatures were necessary and resulted in sharp ¹³C NMR signals with δ = 108.3 (CN) and 18.7 ppm (central carbon) for pure 5 in THF-d ₈ at −70 °C.[8]

Tricyanomethane (5) was not only generated from 1a or 1b via 2, but also by treating 1c with pure, dry hydrogen sulfide,[10] or alternatively by reacting 1d with an excess of anhydrous hydrogen fluoride,[11] and finally by photolysis or short-time thermolysis of azide 7.[8] In the case of precursor 1d, the products 5 and 6 could not be separated. Nevertheless, convincing spectroscopic characterization of 5 was possible.[11] When 7 was photolyzed in solution, the 2H-azirine 8 was formed besides 5, whereas thermolysis of 7 led to 5 and small amounts of 4 only. Irradiation of 7 isolated in an argon matrix did not produce 5, but ketenimine 3 and azirine 8 were formed instead.[8]

The chemistry of tricyanomethane (5) has been intensively investigated by many chemists in the course of more than 150 years.[4] Its gas-phase structure was analyzed by microwave spectroscopy[10] and photoelectron spectroscopy,[12] and the relative stabilities, spectroscopic features, and isomerization reactions of 3, 5, and other C₄HN₃ species were studied using quantum-chemical methods.[13] Especially, the acidic properties of 3 and 5 were discussed in detail.[5] [14] Thus, 5 is mentioned in textbooks of organic chemistry as one of the strongest carbon acids.

Scheme 2 Tricyanomethanide salts 9a–i and compound 9j reported by M. A. Zolfigol et al.[15]

Quite recently, a series of 16 articles by M. A. Zolfigol et al. presented the reaction of tricyanomethane (5) with amine bases to form the corresponding tricyanomethanide salts 9a,[15a] [f] [g] 9b,[15i,o] 9c,[15c] [h] 9d,[15d] [p] 9e,[15m] 9f,[15n] and 9g;[15j] additionally, the magnetic silica-coated nanoparticles 9h [15b] [e] and the phthalocyanine-based vanadium derivative 9i [15l] were described and the synthesis of 9j [15k] from 5 and chlorosulfonic acid was reported (Scheme [2]). The compounds 9a–j were utilized as catalysts for various condensation reactions.[15]

The following discrepancies are characteristics of all the 16 articles:

a) There are no pieces of information or references to the origin of 5.[15] Consequently, the readers might erroneously assume that it is a commercial product. Nevertheless, nonsensical IR and NMR spectra of 5 are depicted in the publications or in the corresponding supplementary data.[15a] [c] [d] [g] ^, [15`] [j] [k] [l] [m] [n] [o]

b) The shown IR spectra of alleged 5 do not present any signal of the C–H stretching vibration.[15a] [c] [d] ^, [15`] [j] [k] [l] [m] [n] [o] This contrasts with the IR data of pure 5 isolated in an argon matrix,[8] which indicated a stronger signal for C–H at 2927.6 cm^–1 if compared with that of C≡N at 2268.1 cm^–1. The depicted IR spectrum of the substance claimed to possess the structure of 9h does not include any signal of an aliphatic or aromatic C–H stretching mode, although the corresponding C≡N signal proves to be observable.[15b] [e]

c) The shown NMR spectra of supposed 5 were measured with solutions in D₂O,[15g] [i] [j] [l] [m] despite the fact that 5 is extremely sensitive to water (see above),[8,9] which excludes NMR characterization of 5 using this solvent. Moreover, the ¹³C NMR signals of 5 are claimed to appear at δ = 77.9 and 165.25 ppm.[15g] [i] [j] [l] [m] These data are not only completely different to the established chemical shifts of 5 (see above);[8] [11] they also stand in flagrant contradiction to general rules, which describe the chemical shifts of cyano groups (δ = 110–125 ppm) and their shielding effects on adjacent carbon atoms.[16] The reported[15g] [i] [j] [l] [m] data of 5 are also incompatible with the known[8] ¹³C NMR data of 4.

d) The ¹³C NMR data of presumed tricyanomethanide salts 9a–g [15a] [c] [d] ^, [15`] [g] [h] [i] [j] <sup>,</sup> [15`] [n] [o] are totally incompatible[17] with the corresponding chemical shifts of similar or identical salts, which are commercially available or reported in the literature.[18] For example, the δ values of the cyano groups are allegedly found at 70.0 for 9d,[15d] 74.8 (9a),[15a] 136.3 (9f),[15n] 152.6 (9e),[15m] and 166.4 ppm (9b).[15i] The central carbon of the tricyanomethanide unit was characterized with δ = 41.9 for 9g,[15j] 50.1 (9c),[15c] 63.0 (9d),[15d] 69.6 (9b),[15i] 120.9 (9f),[15n] and 130.1 ppm (9e).[15m] Clearly, these data are far away from plausible chemical shifts of cyano groups, and the high-field signal of the central carbon atom (δ = 4–6 ppm) is missing in all cases of tricyanomethanide salts 9.[15] Moreover, it is remarkable that entirely different chemical shifts were presented for the same tricyanomethanide species in distinct salts, δ = 70.0–166.4 ppm for the cyano groups and δ = 41.9–130.1 ppm for the central carbon.

We prepared 9a and 9c from single crystals of 5 and 1-methylimidazole and 4,4′-bipyridine, respectively, as well as 9b and 9d from 2 and imidazole or 2,2′-bipyridine.[19] Furthermore, we conveniently synthesized 9a–d from 1b and the corresponding amine hydrochlorides by salt metathesis reactions using a procedure which was analogously utilized in the literature.[18] [20] The ¹³C NMR spectra of the products indicated signals with δ = 120.5 (cyano groups) and 4.8 ppm (central carbon of the tricyanomethanide unit) in all four cases.[19] Thus, we obtained plausible ¹³C NMR data of tricyanomethanide salts 9a–d that are very close to those of similar compounds, reported in the literature,[18] [20] but differ drastically from the data published by Zolfigol et al.[15] [19]

e) Standard mass spectra (EI, 70 eV)[15j] [n] were also utilized by Zolfigol et al. to analyze alleged compounds 9. However, the depicted spectra indicated a very high number of signals, and the characteristic signals were not extractable without strong doubts; nevertheless, the desired [M]⁺ signal with low intensity was always picked out.[15a] [c] [d] ^, [15`] [g] [h] ^, [15j] [k] [m] [n] In the cases of 9a,[15a] [f] [g] 9c,[15c] [h] 9d,[15d] 9e,[15m] 9f,[15n] and 9g,[15j] however, the data include the corresponding signal of [M]⁺, which is based on the total mass of the salt, i.e., the sum of the (di)cation mass and the mass of the anion(s). But detection of cations and anions at the same time is impossible as shown generally[21] and also for similar imidazolium, pyridinium, and ammonium tricyanomethanides.[18]

f) Not only spectroscopic data of 9 but also physical properties, presented by Zolfigol et al.,[15] differ extensively from data reported by other authors. Whereas Trofimenko described 9f as an oil,[6] Zolfigol et al.[15n] claimed isolation of 9f as a yellow solid with a melting point of 136–138 °C. We obtained 9a as a yellow oil with a melting point slightly below room temperature;[19] in great contrast, Zolfigol et al. characterized 9a as a pale-pink solid with m.p. > 300 °C.[15a] [f] [g] While we measured a melting point of 210–215 °C for 9c,[19] Zolfigol et al. described the same compound with m.p. > 300 °C.[15c] Finally, we established that 2 or pure 5 did not react with the very weak base triphenylamine. But Zolfigol et al. claimed a conversion of these precursors to produce 9g with 97% yield.[15j]

In summary, it turned out without any doubt that ­Zolfigol et al.[15] [22] did not utilize tricyanomethane (5) nor salts of type 1 or compounds 2 or 4 as substrate in their syntheses [see points b) and c)], and they did not prepare any tricyanomethanide salt 9 or product with a tricyanomethyl group [points d), e), and f)]. Moreover, their real experiments cannot be clarified because they did not give any information about the origin of the presumed tricyanomethane (5) [see point a)]. Consequently, all described[15] [22] condensation reactions, catalyzed by alleged tricyanomethanide salts 9, are totally unclear and cannot be reproduced by other researchers. It is therefore clear that fundamental corrections are necessary for all 16 articles, or these reports should be withdrawn.

Acknowledgment

We thank Dr. Nader G. Khaligh for fruitful discussions, Jana ­Buschmann, Philipp Denkmann, and Christopher Mohrig for some experiments with 5 and 9, and Dr. Andreas Ihle for his help in connection with the manuscript.

• References and Notes

• 1a Fairly T. Ann. Chem. Pharm. 1864; Supl. 3, 371
• 1b Pfankuch F. J. Prakt. Chem. 1871; 4: 35
• 1c Pfankuch F. J. Prakt. Chem. 1873; 6: 97
• 2a Claus A. Ber. Dtsch. Chem. Ges. 1876; 9: 223
• 2b Claus A. Justus Liebigs Ann. Chem. 1878; 191: 33
• 3 Schmidtmann H. Ber. Dtsch. Chem. Ges. 1896; 29: 1168
• 4 For a more detailed description of the cyanoform story, see Ref. 7 and 8.
• 5 Trofimenko S. J. Org. Chem. 1963; 28: 217
• 6 Trofimenko S, Little EL. Jr, Mower HF. J. Org. Chem. 1962; 27: 433
• 7 Šišak D, McCusker LB, Buckl A, Wuitschik G, Wu Y.-L, Schweizer WB, Dunitz JD. Chem. Eur. J. 2010; 16: 7224
• 8 Banert K, Chityala M, Hagedorn M, Beckers H, Stüker T, Riedel S, Rüffer T, Lang H. Angew. Chem. Int. Ed. 2017; 56: 9582
• 9 The special conditions that are necessary to detect NMR signals of 4 and 5, the stability, and the decomposition of these compounds are explained in detail in Ref. 8.
• 10 Bak B, Svanholt H. J. Mol. Struct. 1977; 37: 153
• 11 Soltner T, Häusler J, Kornath AJ. Angew. Chem. Int. Ed. 2015; 54: 13775
• 12 Bock H, Dammel R. Z. Naturforsch., B 1987; 42: 315
• 13a Szczepaniak M, Moc J. J. Phys. Chem. A 2017; 121: 1319
• 13b Brand H, Liebman JF, Schulz A. Eur. J. Org. Chem. 2008; 4665
• 13c Bak B, Bjorkman C. J. Mol. Struct. 1975; 25: 131
• 13d Stoyanov SS, Tsenov JA, Binev IG. Asian Chem. Lett. 2009; 13: 155
• 13e Elroby SA. Chem. Cent. J. 2016; 10: 20
• 14a Hantzsch A, Osswald G. Ber. Dtsch. Chem. Ges. 1899; 32: 641
• 14b Boyd RH. J. Phys. Chem. 1963; 67: 737
• 14c Raamat E, Kaupmees K, Ovsjannikov G, Trummal A, Kütt A, Saame J, Koppel I, Kaljurand I, Lipping L, Rodima T, Pihl V, Koppel IA, Leito I. J. Phys. Org. Chem. 2013; 26: 162
• 14d Kütt A, Rodima T, Saame J, Raamat E, Mäemets V, Kaljurand I, Koppel IA, Garlyauskayte RY, Yagupolskii YL, Yagupolskii LM, Bernhardt E, Willner H, Leito I. J. Org. Chem. 2011; 76: 391
• 15a Zolfigol MA, Baghery S, Moosavi-Zare AR, Vahdat SM, Alinezhad H, Norouzi M. RSC Adv. 2015; 5: 45027
• 15b Zolfigol MA, Yarie M. RSC Adv. 2015; 5: 103617
• 15c Zolfigol MA, Yarie M, Baghery S. Synlett 2016; 27: 1418
• 15d Zolfigol MA, Mansouri N, Baghery S. Synlett 2016; 27: 1511
• 15e Zolfigol MA, Kiafar M, Yarie M, Taherpour AA, Saeidi-Rad M. RSC Adv. 2016; 6: 50100
• 15f Zolfigol MA, Bahrami-Nejad N, Baghery S. J. Mol. Liq. 2016; 218: 558
• 15g Zolfigol MA, Bahrami-Nejad N, Afsharnadery F, Baghery S. J. Mol. Liq. 2016; 221: 851
• 15h Zolfigol MA, Yarie M, Baghery S. J. Mol. Liq. 2016; 222: 923
• 15i Zolfigol MA, Yarie M, Baghery S, Khoshnood A, Alonso DA. Res. Chem. Intermed. 2017; 43: 3291
• 15j Baghery S, Zolfigol MA, Schirhagl R, Hasani M. Synlett 2017; 28: 1173
• 15k Zolfigol MA, Khazaei A, Alaie S, Baghery S. Can. J. Chem. 2017; 95: 560
• 15l Baghery S, Zolfigol MA, Safaiee M, Alonso DA, Khoshnood A. Appl. Organomet. Chem. 2017; 31: e3775
• 15m Baghery S, Zolfigol MA, Schirhagl R, Hasani M. Catal. Lett. 2017; 147: 2083
• 15n Baghery S, Zolfigol MA, Maleki F. New J. Chem. 2017; 41: 9276
• 15o Yarie M, Zolfigol MA, Babaee S, Baghery S, Alonso DA, Khoshnood A. Res. Chem. Intermed. 2018; 44: 2839
• 15p Mansouri N, Baghery S, Zolfigol MA. Nanochem. Res. 2018; 3: 170
• 16 Kalinowski H.-O, Berger S, Braun S. ¹³C NMR Spektroskopie . Thieme; Stuttgart: 1984: 227
• 17 The ¹³C NMR spectrum of 9j depicted in Ref. 15k and that shown in the corresponding supplementary data are quite different, but both spectra are clearly incompatible with the supposed structure of 9j.
• 18 For some examples, see: Marszalek M, Fei Z, Zhu D.-R, Scopelliti R, Dyson PJ, Zakeeruddin SM, Grätzel M. Inorg. Chem. 2011; 50: 11561
• 19 For details, see the Supporting Information.
• 20 For an alternative synthesis of 9a, see: Kubo D, Ueda T, Tahara H, Mizuta K, Hashimoto H, Kasahara T. US 20080083626A1, 2008
• 21 Hesse M, Meier H, Zeeh B, Bienz S, Bigler L, Fox T. Spektroskopische Methoden in der Organischen Chemie, 9th ed. Thieme; Stuttgart: 2016: 277
• 22 Quite recently, another report on 9d, which was claimed to result from alleged 5, was published, see: Chehardoli G, Mansouri N. Iran. Chem. Commun. 2018; 6: 450