Contribution to the Chemistry of Boron, 136 [1] a Reinvestigation of the Existence of L,3,2-Dioxaborinium(L-|-) Cations

Contribution to the Chemistry of Boron, 136 [1] A Reinvestigation of the Existence of l,3,2-Dioxaborinium(l-|-) Cations

Chaitanya K. Narula and Heinrich Nöth* Institut für Anorganische Chemie der Universität München, Meiserstraße 1, D-8000 München 2

Z. Naturforsch. 88b, 1161-1164 (1983); received July 5, 1983

l,3,2-Dioxaborinium(l +) Salts, 2-Perchlorato-l-oxa-3-oxonia-2-borato-cyclohexadienes, 2-Triflato-l-oxa-3-oxonia-2-borato-cyclo-hexadienes

It is shown that the previously reported 1,3,2-dioxaborinium salts containing tricoordinated boron are in fact tetracoordinated coordination compounds. The same holds for the corresponding triflates.

Introduction they have to be described by formula 2 rather than The chemistry of boron cations was a field long by 1. Evidence for this structure rests also on the neglected [2]. Meanwhile, many cations containing IR spectra which show well defined bands at 1010 tetracoordinated boron have been reported [3] and and 1145 cm-1 typical for a monodentate CIO4- their chemistry has been explored. In addition there species of local C3V symmetry. In addition, a band exists scattered information relating to tri- [4-8] of medium intensity is found at 918 cm-1, which is and dicoordinated [9, 10] boron cations. Unlike also indicative of coordinated C104- of C3V symmetry carbenium ion chemistry their role in boron chem- [14]. istry is not yet well understood. This situation, however, will quickly change as their chemical FORMULAE 1 behaviour is being studied. Having evaluated some governing factors relating to the stability of dicoordinated cations of boron [9] we became interested in those of coordination number 3 because their stability should also depend 2(5,6) on mesomeric effects. Such effects could prevail in tricoordinated l,3,2-dioxa-borinium(l + ) salts 1 which were isolated as Perchlorates and hexachloro- antimonates [5-7]. One important argument for the presence of a tricoordinated boron atom in these compounds was an nB NMR signal at ~ 26 ppm which is typical for a CBO2 structural unit [11]. R1 R2 X

Results 5a C6H5 C6H5 ClOA 5b CH ClO Since the triflate group is an exceptionally good <=6H5 3 A 6a S03CF leaving group we turned our attention to the prep- C6H5 C6H5 3 aration of the triflates 6a and 6b. However, these 6bH20 C6H5 CH3 SO3CF3 compounds (Table I) contain tetracoordinated boron by UB NMR evidence. Therefore we repeated Since 5 a and 5 b show low solubility in CH2C12 n the preparation of 5 a and 5 b. Although we could their B NMR spectra were also recorded in aceto- easily obtain these compounds [12], they exhibited, nitrile solution, and again only tetracoordinated in our hands, nB NMR data for tetracoordinated boron (Table I) was observed. It is likely, there- boron in CH2CI2 or benzene solution. Therefore fore, that this solvent coordinates to 1 thus leading to 3 [14]. This assumption rests on the fact that the

* Request for reprints to Prof. Dr. H. Nöth. acetonitrile solutions show high electrical conduc- 35 0340-5087/83/1000-1161/$ 01.00/0 tivity as well as a <5 C1 shift typical of an ionic 1162 Ch. K. Narula-H. Nöth • Contribution to the Chemistry of Boron

Table I. NMR data of 1,3,2-dioxa-bora-cyclohexadiene-compounds 5 and 6; (5-values in ppm.

önB Solvent <5*9F Solvent <5*H Solvent

5a 5.6 CH3CN 8.1 CH2C12 5b 5.5 CH3CN 6.8 CH2CI2 6a 8.6 CH2CI2

6b H20 6.5 CH2CI2 — 77.09 CDCI3 2.3 (S, CH3) CDCI3 7.26-7.98 (M, C6H5) 10.95 (S H20)a C6H5B(OH)2 29.4 CH2CI2

compound ratio <5UB Intensity ratio

b C6H5B(OH)2 - BzacH 1:1 29.9 6.8 (2:1) 1:2 29.1 6.7 (1:1) 1:3 29.9 6.8 (1:1)

a This signal disappears on addition of D2O; b BzacH = benzoylacetone; solvent CH2CI2.

Perchlorate (found: = 1000 ppm, h(l/2) 65 Hz, The formation of compounds 5 and 6 seems to literature [13]: <535C1 = 1000 ppm) [13]. Evidence occur in two steps. UB NMR signals in an approx- for coordination of acetonitrile is obtained also imately 1:1 ratio are recorded on addition of the from XH NMR spectra of a 1:1 solution of 5 a and diketone to a solution of phenylboronic acid (see

CH3CN in CDCI3 which shows a 9 Hz downfield Table I), one at <511B=27, the other at 6,8 ppm. Since shift for the signal of acetonitrile although part of this intensity ratio changes in favour of the high the Perchlorate precipitates from the solution, and field signal at higher diketone ratios an equilibrium this may have changed the stoicheiometry. situation (1) is indicated. The addition of either In the solid state, however, the structure of 5 a HCIO4 or HOSO2CF3 then produces the nB NMR and 5 b is best represented by 4. The compounds signal of 5 a and 6 a respectively, according to (2a, b). hydrolyse fairly readily in CH2CI2 solution. Thus In contrast, 6b could only be isolated as a hydrate, the UB NMR signals reported previously for 5 a and and we assume that this compound has the con- 5 b are most likely due to phenylboronic acid [13]. stitution of the intermediate 7 [15]. The triflates 6a and 6b, prepared in analogy to the Perchlorates, are sharp melting, powdery solids and are more soluble in dichloromethane than the FORMULAE II corresponding Perchlorates. They show nB chemical shifts (Table I) typical for tetracoordinated boron. All NMR data for 6 a are in accord with a structure containing a coordinated triflate group in analogy to 4. Therefore, the comparison of the Perchlorates and triflates very strongly point to the conclusion that none of these compounds is truly ionic, i.e. are salts of a l,3,2-dioxaborinium(l + ) cation. How- ever, the electronegative group X = CIO4 or SO3CF3 may be readily replaced by a suitable base, (2) the compounds thus acting as a tricoordinate boron cationic species with a loosely coordinated anion. This situation thus resembles the highly reactive

compounds CpM(CO)2BF4 (M = Mo, W) [16]. 6b (X = SO3CF3) Ch. K. Narula-H. Nöth • Contribution to the Chemistry of Boron 1163

Experimental flask (100 ml) fitted with a dropping funnel and reflux condenser was charged with dibenzoyl- All manipulations were carried out under rigor- methane, phenylboronic acid and 25 ml CH2CI2. ously dry conditions in a nitrogen atmosphere using This flask was immersed in an ice-salt bath and the standard high vacuum or Schlenk tube techniques. mixture stirred magnetically. A solution of triflic Solvents were carefully dried and stored under di- acid (or HCIO4) in CH2CI2 was added dropwise, and nitrogen. Most of the starting materials are com- a yellow powder starts separating after addition of mercially available and were used as such. An- — 1/5 of the acid. Having completed the addition hydrous perchloric acid was prepared by extracting stirring was continued for another 2 h, the solid a mixture of oleum and 70% perchloric acid in filtered, washed with 20 ml of CH2CI2 and dried dichloromethane [17]. The solution obtained was in vacuo. standardized by base titration. Results are summarized in Table II. IR-spectra: Perkin-Elmer 325, Nujol/hostaflon mulls; NMR: Jeol FX 90 and Bruker WP200; C. K. N. gratefully acknowledges a Humboldt standards for «51 H: tetramethylsilane; award. We also thank Fonds der Chemischen In- HB: F3B • 0(C2H5)2 ext.; ^Cl: 1M NaC104 solution, dustrie for continous support, Dr. habil. B. Wrack- ext. meyer, Mrs. H. Bauer and Mr. D. Schlosser for General method: A two necked, round bottomed recording spectra.

Table II. Preparative and analytical results for the Perchlorates and triflates, 5 and 6.

Reactants (g(mmol)) CH2CI2 yield m.p. PhB(OH)2 Diketone Acid ml g(%) [°C]

2-Perchlorato-2,4,6-triphenyl-l-oxa-3-oxonia- .600 1.11 .498 10 1.57 llla 2-borata-cyclohexadien(3,5) (5 a) (5.00) (4.95) (4.96) (78) 2-Perchlorato-2,4-diphenyl-6-methyl-l-oxa- .600 .810 .498 10 1.48 114b 3-oxonia-2-borata-cyclohexadien(3,5) (5b) (5.00) (4.96) (4.96) (85) 2-Triflato-2,4,6-triphenyl-l-oxa-3-oxonia- .695 1.27 .850 60 2.32 94 dec. 2-borato-cyclohexadien(3,5) (6a) (5.80) (5.66) (5.67) (89) 2-Aquo-2,4-diphenyl-6-methyl-l-oxa-3-oxonia-2- .690 .940 .850 60 2.18 74 borata-cyclohexadienyl(3,5)-trifluorsulfonat (6 b) (5.75) (5.76) (5.67) (95)

Analysis (calcd (found))

Formula mol.weight C [18] H B Cl

5a C2IHI6BC106 410.6 61.43 3.93 (48.29) (4.11)

ob CI6H14BC106 348.5 55.14 4.05 3.10 10.17 (49.19) (4.01) (2.83) (10.2)

6 a C22HI6BF305S 460.2 57.42 3.50 2.35 6.97 (40.78) (3.63) (2.04) 8.23

6b C17HI6BF306S 398.1 49.06 3.88 (48.44) (4.21)

a Lit. [5] m.p. 100 °C; b Lit. [5] m.p. 104 °C.

[1] Contribution 135: K. Anton and H. Nöth, Chem. [4] H. Nöth and P. W. Fritz, Z. Anorg. Allg. Chem. Ber., in press. 322, 297 (1963). [2] Gmelin, Handbuch der Anorganischen Chemie, [5] A. T. Balaban, A. Barabas, and A. Arsene, Ergänzungswerk zur 8. Auflage, Volume 37, Bor- Tetrahedron Lett. 1964, 2721. verbindungen Part 10, p. 37ff., 1976. [6] A. T. Balaban, A. Arsene, I. Bally, A. Barabas, [3] O. R. Shitov, S. L. Joffe, V. A. Tartakovskii, and M. Paraschiv, and E. Romas, Tetrahedron Lett. S. S. Novikov, Russ. Chem. Rev. 39, 905 (1970). 1965, 3917. 1164 Ch. K. Narula-H. Nöth • Contribution to the Chemistry of Boron 1164

[7] A. Trestianu, M. Niculescu-Majewska, I. Bally, [14] The chemical shift reported for 5 a in CH3CN was A. Barabas, and A. T. Balaban, Tetrahedron 24, <5nB = 26 ppm. This values is close to duB = 28.4 2499 (1968). reported for phenylboronic acid in ethanol (M. J. S. Dewar and J. R. Jones, J. Am. Chem. [8] P. Binger and R. Köster, Chem. Ber. 108, 395 Soc. 89, ,2408 (1964); ibid. 89, 4251 (1967)), the (1975). product of hydrolysis of 5 a. (5UB values for [9] H. Nöth, R. Staudigl, and H.-U. Wagner, Inorg. 30-substituted phenyl boronic acids are found in Chem. 21, 706 (1982) and literature cited therein. the (5-range 31-25.8 (H. C. Beachall and D. W. Beistel, Inorg. Chem. 3, 1028 (1964)). We observed [10] J. Higashi, A. D. Eastman, and R. W. Parry, <5nB = 26.8 ppm for phenylboronic acid in Inorg. Chem. 21, 718 (1982). CH2C12. [11] H. Nöth and B. Wrackmeyer, Boron NMR [15] During crystallisation of compounds of type 5 Spectroscopy in "NMR Spectroscopy, Principles phenylboronic acid slightly contaminated by the and Applications", P. Diehl, E. Fluck, and R. /3-diketone could have been obtained [5]. Kosfeld (eds): Springer Verlag, Heidelberg, New [16] W. Beck and K. Schloter, Z. Naturforsch. 33b, York 1978. 1215 (1978); K. Schloter and W. Beck, ibid. 35 b, 985 (1980). [12] Attempts to obtain single crystals for X-ray structure determinations have so far been un- [17] F. Klages and P. Hegenberg, Angew. Chem. 14, successful. 902 (1962). [18] Carbon analysis gave consistently low figures [13] A. E. Wickenden and R. A. Krause, Inorg. Chem. even in the presence of V2O5 or WO3. This 4, 404 (1965). corresponds to reported results [5-7].