Aluminium-containing ring systems and N-heterocycle formation via nitrile insertions into Al–N bonds

Abstract

Reactions of Me₃Al with 1,2-diaminobenzene [1,2- (H₂N)₂C₆H₄] or anthranilic acid, [1,2-(H₂N)(HO₂- C)C₆H₄], followed by treatment with acetonitrile, afford tetranuclear and hexanuclear aluminium-containing ring systems; a single crystal X-ray structure on the hexametallic product reveals the construction of quinazoline ligands arising via insertion of acetonitrile into Al–N bonds.

---

It is well over one hundred years since the trimerisation of nitriles to triazines by organometallic reagents was first noted, by Hofmann¹ using sodium, and by Frankland² in his studies on Et₂Zn. During the 1960s, Wade³ and Lappert⁴ did much to advance the understanding of nitrile binding and insertion reactions at main group centres; however, the potential of such reactions for synthesising useful heterocycles and interesting metal-containing ring systems has remained largely undeveloped.

Recently, we described the synthesis of large aluminium-containing ring systems via treatment of Me₃Al with hydrazines;⁵ these included a highly unusual octa-aluminium structural analogue of a tetrapyrrole. The key macrocycle-forming step revolves around the insertion of acetonitrile into the aluminium–nitrogen bonds of intermediate amide species. With a view to probing the generality of this approach for the synthesis of aluminium-containing macrocycles, and also to evaluate the potential of this methodology for constructing nitrogen heterocycles, we have extended the study to other classes of amine substrate. Here, we describe the reactivity of Me₃Al towards 1,2-diaminobenzene [1,2-(H₂N)₂C₆H₄] and anthranilic acid [1,2-(H₂N)(HO₂C)C₆H₄]. The former gives rise to a tetranuclear complex, the latter to an unprecedented hexanuclear species incorporating N-heterocyclic ligands.

Slow addition of a solution of Me₃Al (2 equivalents) in toluene to 1,2-(H₂N)₂C₆H₄, followed by a 12 h reflux, afforded a pale brown solution. After removal of the solvent under reduced pressure, the residue was dissolved in acetonitrile and heated to reflux for 2–3 min. Slow cooling of this solution, followed by standing at room temperature for 2 days, gave colourless needles of 1 in 40% yield (Scheme 1). The ¹H NMR spectrum‡ of 1 consists of four equal intensity singlets in the Al–methyl region (δ −0.37 to −0.75) together with a similar intensity singlet at δ 1.29 attributable to carbon-bonded methyl groups. X-Ray analysis§ reveals the product to be the C_i symmetric twelve-membered macrocyclic complex 1 comprising four aluminium atoms (two bridging and two chelating), six nitrogen atoms and two carbon atoms, The structure of this product is closely related to that obtained from the reaction of MePhNNH₂ with Me₃Al,⁵ and is not described in further detail here.

Scheme 1 Compounds 1–4 (inserted molecules of acetonitrile are shown in bold).

In the absence of acetonitrile, the 2∶1 reaction of Me₃Al and 1,2-(H₂N)₂C₆H₄ (in toluene) has been shown to afford the asymmetric complex [(Me₂Al)₂AlMe(C₆H₄(NH)₂ )₂]·AlMe₃2.⁶ This, or a closely related derivative, is the likely precursor to 1. Formation of the 12-membered ring product is brought about by insertion of two acetonitrile molecules into the Al–N bonds, in a related manner to that postulated for the reaction of hydrazines with Me₃Al/MeCN.⁵

Similar treatment of anthranilic acid, [1,2-(NH₂)- (HO₂C)C₆H₄] (twice sublimed) with 2 equivalents of a toluene solution of Me₃Al affords, after work-up in acetonitrile, large yellow prisms for which the ¹H NMR spectrum possesses nine distinct singlets in the Al–methyl region. The X-ray analysis§ of the product revealed the chiral trimeric hexanuclear complex [(Me₂AlL)(MeAl)(μ₃-O)(μ-O)]}₃3 (L = quinazoline) shown in Fig. 1. The central Al₃O₃ ring has a twisted boat conformation, whereas the three attached Al₂O₂CN rings each have a folded envelope geometry. All six aluminium atoms exhibit marked departures from ideal tetrahedral geometry with angles in the range 100.9(2)–124.4(3)°. Although not possessing strict C₃ symmetry [the methyls on Al(2) and Al(4) are ‘up’, whilst that on Al(6) is ‘down’], the pattern of bonding throughout the structure is essentially three-fold symmetric. It is interesting to note that whilst all of the Al–O bond lengths within the central Al₃O₃ ring and also those to O(2), O(4) and O(6) are all essentially the same [1.770(4)–1.794(4) Å], those to Al(1), Al(3) and Al(5) are all significantly longer [1.811(4)–1.827(4) Å]. As 3 is not the product of a chiral synthesis, the presence in the crystals of molecules of only one chirality is a consequence of spontaneous resolution upon crystallisation.

Fig. 1 The molecular structure of 3. Selected bond lengths (Å); Al(1)–O(1) 1.811(4), Al(1)–N(2) 2.021(5), Al(2)–O(1) 1.785(4), Al(2)–O(3) 1.788(4), Al(2)–O(2) 1.779(4), Al(3)–O(3) 1.827(4), Al(3)–N(22) 2.034(5), Al(4)–O(3) 1.770(4), Al(4)–O(4) 1.783(5), Al(4)–O(5) 1.785(4), Al(5)–O(5) 1.825(4), Al(5)–N(42) 2.015, Al(6)–O(1) 1.784(4), Al(6)–O(5) 1.794(4), Al(6)–O(6) 1.785(5).

Whereas the formation of 1 can be rationalised in terms of acetonitrile insertion into the aluminium–nitrogen bonds of 2, the carboxylic acid group of anthranilic acid contributes oxygen atoms to the Al₃O₃ core around which the quinazoline ligands are clustered. The reaction reproducibly affords 3 in ca. 45% yield; its hydrolysis then readily releases the 2-methyl-4(3H)-quinazolinone heterocycle 4 (identified by comparison of NMR data with those of an authentic sample) in excellent yield. We note that the formation of N-heterocycles by this methodology is related to the intramolecular Ritter reaction^7–9 in which nitrilium salts, generated in the presence of Friedel–Crafts reagents, react with a second nitrile molecule to give quinazoline ring systems.⁹

Future studies will focus on further exploiting nitrile insertions into aluminium (and gallium) bonds to access unusual inorganic ring systems and nitrogen heterocycle products.

Acknowledgements

The EPSRC and the Leverhulme Trust (for a Research Fellowship to C. R.) are thanked for financial support. Professor Charles Rees is thanked for helpful discussions.

Notes and References

1. A. W. Hofmann, Chem. Ber., 1868, 1, 194.
2. E. Frankland and J. C. Evans, J. Chem. Soc., 1880, 37, 563.
3. H. J. Emeléus and K. Wade, J. Chem. Soc., 1960, 2614; J. E. Lloyd and K. Wade, J. Chem. Soc., 1964, 1649; J. E. Lloyd and K. Wade, J. Chem. Soc., 1965, 2662; J. R. Jennings, J. E. Lloyd and K. Wade, J. Chem. Soc., 1965, 5083; I. Pattinson and K. Wade, J. Chem. Soc., 1968, 57; J. R. Jennings and K. Wade, J. Chem. Soc. A, 1968, 1946.
4. W. Gerrard, M. F. Lappert and J. W. Wallis, J. Chem. Soc., 1960, 2178; M. F. Lappert and B. Prokai, Adv. Organomet. Chem., 1967, 5, 225 and references therein..
5. V. C. Gibson, C. Redshaw, A. J. P. White and D. J. Williams, Angew. Chem., Int. Ed., 1999, 961.
6. R. L. Wells, H. Rahbarnoohi, P. B. Glaser, L. M. Liable-Sands and A. L. Rheingold, Organometallics, 1996, 15, 3204.
7. J. J. Ritter and P. P. Minieri, J. Am. Chem. Soc., 1948, 70, 4045; J. J. Ritter and J. Kalish, J. Am. Chem. Soc., 1948, 70, 4048.
8. L. I. Krimen and D. J. Cota, Org. React. (N.Y.), 1969, 17, 213.
9. R. Bishop, in Comprehensive Organic Synthesis, ed. B. M. Trost and I. Fleming, Pergamon,Oxford, UK, 1991, vol. 6 and references therein..

---

Footnotes

† Present address: School of Chemical Sciences, University of East Anglia, Norwich, UK NR4 7TJ.

‡ Satisfactory microanalyses have been obtained.

For 1: ¹H NMR (400 MHz, C₆D₆, 298 K), δ −0.75, −0.48, −0.42, −0.37 (4 × s, 8 × 3H, AlMe), 1.29 (s, 2 × 3H, CMe), 6.36–7.27 (3 × m, 8H, aryl H), NH not seen. ¹³C NMR (100.6 MHz, C₆D₆, 298 K), δ −12.24, −12.81 (br, AlMe), −8.91 (br, AlMe), −5.99 (br, AlMe). IR: ν(N–H) 3247 cm⁻¹.

For 3: ¹H NMR (400 MHz, CDCl₃, 298 K), δ −0.87, −0.82, −0.59, −0.47, −0.46, −0.45, −0.24, −0.23, −0.22 (9 × s, each 3H, AlMe), 2.79 (s, 3H, Me-quin), 2.88 (s, 3H, Me-quin), 2.89 (s, 3H, Me-quin), 7.61 (m, 3H, quinH), 7.82 (m, 3H, quinH), 7.92 (m, 3H, quinH), 8.35 (m, 3H, quinH). ¹³C NMR (100.6 MHz, CDCl₃, 298 K), δ −12.24, −11.56 (2 × m, 3 × AlMe), −7.22, −6.79, −6.45, −5.43, −5.10, −4.57 (6 × s, 3 × AlMe₂), 25.53, 25.82, 25.93 (3 × s, Me-quin), 117.15 (m, 3 × aryl C), 125.28–127.61 (overlapping m, 9 × aryl C), 136.33 (m, 3 × aryl C), 150.98 (m, 3 × aryl C), 158.45 (m, 3 × aryl C), 169.31 (m, 3 × aryl C). EI-MS: m/z: 807 (M⁺ − Me). IR: ν(μ₃-O)Al₃ 800 cm⁻¹.

§ Crystal data: For 1: C₂₄H₄₂N₆Al₄, M = 522.6, orthorhombic, space group Pbca (no. 61), a = 8.717(1), b = 16.868(1), c = 20.416(2) Å, V = 3002.0(4) ų, Z = 4 (the complex has crystallographic C_i symmetry), D_c = 1.156 g cm⁻³, μ(Cu-Kα) = 16.1 cm⁻¹, F(000) = 1120, T = 183 K; clear prisms, 0.27 × 0.23 × 0.13 mm, Siemens P4/RA diffractometer, ω-scans, 2182 independent reflections. The structure was solved by direct methods and the non-hydrogen atoms were refined anisotropically using full-matrix least squares based on F² to give R₁ = 0.069, wR₂ = 0.173 for 1479 independent observed reflections [|F_o| > 4σ(|F_o|), 2θ < 120°] and 162 parameters.

For 3: C₃₆H₄₈N₆O₆Al₆, M = 822.7, orthorhombic, space group P2₁2₁2₁ (no. 19), a = 12.152(1), b = 16.054(1), c = 22.310(1) Å, V = 4352.3(6) ų, Z = 4, D_c = 1.256 g cm⁻³, μ(Cu-Kα) = 17.9 cm⁻¹, F(000) = 1728, T = 173 K; yellow rhombs, 0.50 × 0.23 × 0.23 mm, Siemens P4/RA diffractometer, ω-scans, 4017 independent reflections. The structure was solved by direct methods and the non-hydrogen atoms were refined anisotropically using full-matrix least squares based on F² to give R₁ = 0.055, wR₂ = 0.133 for 3483 independent observed reflections [|F_o| > 4σ(|F_o|), 2θ < 128°] and 488 parameters. The absolute structure of 3 was determined by use of the Flack parameter which refined to a value of −0.07(7).

CCDC 182/1855. See http://www.rsc.org/suppdata/cc/b0/b007810g/ for crystallographic files in .cif format.

---