Catalytic N-N coupling of aryl azides to yield azoarenes via trigonal bipyramid iron-nitrene intermediates

Article abstract of DOI:10.1021/ja910224c

(Figure Presented) The reactivity of the trigonal bipyramidal iron(I) complex [SiP^iPr₃]Fe(N₂) ([SiP ^iPr₃] = (2-iPr₂PC₆H ₄)₃Si^-) toward organoazides has been examined. 1-Adamantylazide was found to coordinate the Iron center to form stable [SiP^iPr₃]Fe(η¹-N₃Ad). Aryl azides instead afforded unstable [SiP^iPr₃]Fe(N ₃Ar) species that decayed gradually to regenerate [SiP ^iPr₃]Fe(N₂) with release of azoarenes (ArN=NAr). The conversion of aryl azides to azoarenes can thus be achieved catalytically. Competitive trapping experiments strongly suggest the intermediacy of reactive nitrene complexes of the type [SiP^iPr ₃]Fe(NAr) that couple bimolecularly in the N-N bond forming step. Evidence for one such intermediate was provided by electron paramagnetic resonance spectroscopy via photolysis of [SiP^iPr₃] Fe(N₃Ar) in a frozen glass. The electronic structures of these putative nitrene intermediates have been examined by DFT methods. Copyright

Full text of DOI:10.1021/ja910224c

Published on Web 03/03/2010
Catalytic N-N Coupling of Aryl Azides To Yield Azoarenes via Trigonal
Bipyramid Iron-Nitrene Intermediates
Neal P. Mankad, Peter Mu¨ller, and Jonas C. Peters*
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
Received December 3, 2009; E-mail: jcpeters@mit.edu
1
Organoazides (N3R) are desirable reagents for nitrene transfer
reactions.1 Generally, organoazides interact with transition metal
catalysts to form transient M-N3R adduct species that in many
cases subsequently extrude N2 to yield metal-imido/nitrene (MNR)
complexes as the key reactive intermediates for substrate activation.1
Understanding the properties of metal-organoazide and metal-imido/
nitrene complexes is fundamentally important in this context.
Several iron-imido/nitrene intermediates that are unstable toward
hydrogen atom abstraction and/or intramolecular ligand oxidation
pathways have been proposed for various coordination geometries.2
We3 and others4 have shown that iron-imido complexes formed
from organoazides can be stabilized in pseudotetrahedral coordina-
tion environments. Subsequently, stable iron-imido complexes with
distorted square planar geometries,5 as well as a trigonal planar
bis(imido)iron complex,6 also have been structurally characterized.
S ) /2 ground state similar to 1 and in accord with its intense
electron paramagnetic resonance (EPR) signal with gaverage ) 2.086.
The solid-state structure of 2 (Figure 2a) is noteworthy because
isolable metal-organoazide complexes are unknown for Fe and
are rare in general.11 The η1-Nγ binding exhibited in 2 is the most
common binding mode represented in the literature.11a–f The
diazenylimido(2-) resonance structure (MdN-NdNR) is descrip-
tive of many of these complexes11a–e and results in bent N-N-N
angles (114-117°) and red-shifted ν(N3) IR bands. A recent
copper(I) 1-adamantylazide complex,11f on the other hand, is best
formulated as a redox-innocent RN3fM adduct and exhibits a linear
N-N-N angle of 173.1(3)° and a blue-shifted ν(N3) IR band.
Complex 2 has an N-N-N angle of 147(4)°, and its N3 vibration
occurs at an energy that is indistinguishable from free 1-adaman-
tylazide. Also noteworthy is the Fe-N distance in 2 (1.769(5) Å),
which is significantly shorter than that of 1 (1.817(4) Å).10 These
data collectively indicate that the electronic structure of 2 lies
somewhere between the above two limiting resonance structures.
DFT calculations12 are consistent with this hypothesis, as the
calculated spin densities (molecular sum 1.00) on the P3Fe and N3
units are 1.81 and -0.77, respectively (Figure 1c). This electronic
structure is distinct from metalloradical 1, for which 92% of the
calculated spin density resides on Fe.12
Figure 1. Qualitative d-orbital splitting diagrams for pseudotetrahedral and
trigonal bipyramidal iron-imido/nitrene complexes.
An interesting geometry to consider for an FeNR unit is the
trigonal bipyramid (TBP). The well-studied pseudotetrahedral
L3FeIIINR complexes feature bona fide Fe-N triple bonds resulting
from two highly destabilized, unoccupied π*FeN orbitals (Figure 1,
left).3 Introducing a ligand trans to the imido group and shifting
the Fe into the L3 plane in a TBP leads in principle to population
of the π*FeN set (Figure 1, right), thereby obliterating a significant
degree of the Fe-N multiple bonding character presumed respon-
sible for the stability of pseudotetrahedral L3FeIIINR species.3,4a
Accordingly, until recently7 metal-ligand multiply bonded species
in TBP configurations had been isolated only for d-electron counts
of 0 or 1.8 TBP systems with higher d-electron counts often
dissociate the apical ligand and distort toward the more stable
pseudotetrahedralgeometrywhenaccommodatinganaxialmetal-ligand
multiple bond.9 In this context, we sought to examine the ramifica-
tions of placing a FeNR linkage in a TBP environment. Our recent
work10 with low-valent TBP Fe complexes supported by anionic
Figure 2. Core structures of (a) 2 and (b) 6 as 50% probability ellipsoids.
(c) Spin density plot of 2 (0.002 isocontour). Selected bond lengths (Å)
and angles (deg) for 2: Fe-N1, 1.769(5); N1-N2, 1.269(13); N2-N3,
1.25(3); Fe-N1-N2, 162(3); N1-N2-N3, 147(4); N2-N3-C, 111.5(15).
For 6: Fe-N, 1.963(2).
Stoichiometric reactions of 1 with aryl azides (N3Ar) gave green-
colored solutions that persisted at -30 °C but that decayed gradually
to yield red solutions containing 1 and the corresponding azoarene
(ArNdNAr) products at room temperature (Scheme 1). Complex
1 also was found to be a catalyst for this unusual N-N coupling
reaction. With 5% catalyst loading in C6D6 solutions at 70 °C,
various aryl azides (Ar ) Ph, p-tolyl, p-C6H4OMe, Mes) were
converted to the corresponding azoarenes in moderate yields
(44-57%). The only other spectroscopically detectable organic
products were the corresponding anilines (ArNH2) as minor
tris(phosphino)silyl ligands (2-R2PC6H4)3Si- ([SiPR ]-) provided
3
a convenient entry point for such studies (Figure 1).
The addition of 1-adamantylazide to red-colored [SiPiPr3]Fe(N2)
(1) produced dark brown solutions from which the organoazide
adduct [SiPiPr3]Fe(η1-N3Ad) (2) was crystallized. A characteristic
optical band for 2 appears at 679 nm (ε ) 1100 M-1 cm-1). The
solution magnetic moment of 2 is µeff ) 2.2 µB, consistent with an
9
10.1021/ja910224c  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 4083–4085 4083

C O M M U N I C A T I O N S
byproducts (8-24%).13 To our knowledge, the only previous
example of catalytic N-N coupling to yield diazenes from
organoazides was reported as a side reaction during the catalytic
amination of C-H bonds by Co porphyrins.14 Thus, we became
interested in understanding the mechanism of this unusual process.
The green intermediates observed during the reactions between
1 and aryl azides were assigned as the aryl azide complexes
[SiPiPr3]Fe(N3Ar) (3, Ar ) Ph, p-tolyl, p-C6H4OMe, Mes, 2,6-
Et2C6H3) on the basis of in situ spectroscopic characterization and
comparison to stable organoazide complex 2. Characteristic ν(N3)
IR bands were observed at energies deviating by e3 cm-1 from
the corresponding free aryl azides, implying the presence of
coordinated azide ligands. Complexes 3 are paramagnetic, and the
p-tolyl derivative [SiPiPr3]Fe(N3Tol) (3-Tol, Tol ) p-tolyl) exhibited
an intense EPR signal with gaverage ) 2.106. Characteristic optical
bands for 3 were evident at λmax ) 608-617 nm (ε ) (est.) 1400
M-1 cm-1) and ca. 775 nm (Figure 3).
with isocyanides to yield carbodiimide products are diagnostic of
isolable FeNR species. Second, when 3-Tol was generated in the
presence of 9,10-dihydroanthracene (DHA), the resulting product
mixture contained a new paramagnetic product (6), identified
crystallographically (Figure 1b) as [SiPiPr3]Fe(NHTol) after inde-
pendent synthesis from [SiPiPr3]Fe(OTf) and LiNHTol. For com-
parison, adamantylazide adduct 2, which does not decay at a
measurable rate under analogous conditions, was found to be
unreactive toward DHA and underwent ligand substitution with
tBuNC to yield 5 and free N3Ad.
Scheme 1
Figure 3. Changes in the UV-vis spectrum during the decay of
[SiPiPr3]Fe(N3Ph) (inset: first-order plot of absorbance at 611 nm).
The decay profiles of 3 were followed by UV-vis spectroscopy
at room temperature. In all cases studied, clean isosbestic behavior
was observed (Figure 3), indicating that no intermediates were
accumulating during the conversion of 3 to ArNdNAr and 1. The
decay of 3 was found to be clean and first order (Figure 3). The
rate of decay of the aryl azide complexes was relatively insensitive
to the identity of the aryl substituent, as the t1/2 of the complexes
spanned a relatively small range (1.2-3.2 h at 23 °C) even as the
electronics (Ar ) Ph, p-tolyl, p-C6H4OMe) and sterics (Ar ) Mes,
2,6-Et2C6H3) were varied.
It is likely that complexes 3 decayed to yield reactive
[SiPiPr3]Fe(NAr) intermediates (4) prior to N-N coupling, akin to
first-order conversions of Cp2Ta(Me)(N3Ar) complexes to
Cp2Ta(Me)(NAr) studied by Bergman.11a,b Important to note is our
recent isolation and structural characterization of the Ru analogues
of 4, [SiPiPr3]Ru(NAr).15 Though the corresponding Fe complexes
4 are too reactive to permit isolation and thorough characterization,
it appears that the p-tolyl derivative [SiPiPr3]Fe(NTol) (4-Tol) can
be observed by EPR spectroscopy in a frozen glass. Photolysis of
3-Tol in frozen 2-methyltetrahydofuran over 1 h at 77 K resulted
in a color change from green to red-brown. The disappearance of
the EPR signal for 3-Tol (Figure 4b) was accompanied by the
appearance of a new EPR signal (Figure 4c) that we presume to
correspond to 4-Tol. A simulation of this signal was obtained with
the parameters (gx, gy, gz) ) (1.990, 2.032, 2.098) and (AxP, AyP,
AzP) ) (55, 40, 50 MHz) (Figure 4d), suggesting that imido complex
4-Tol possesses an S ) 1/2 ground state. Upon thawing of the glass,
intermediate 4-Tol rapidly converted to 1 (Figure 4e) and
TolNdNTol.
Careful analysis of product distributions in these trapping
experiments proved informative with regard to the mechanism of
N-N bond formation. The ratio of azotoluene to anthracene
decreased linearly when decreasing the concentration of 1 while
holding the concentrations of N3Tol and DHA constant (Supporting
Information (SI), Table S5). If one assumes that only one Fe center
is involved in the hydrogen atom transfer (HAT) reaction with
DHA, this result implies the involvement of two Fe centers in the
N-N bond forming step. On the other hand, the azotoluene/
anthracene ratio was insensitive to varying the concentration of
N3Tol while holding the concentrations of 1 and DHA constant
(see SI). Though mechanisms involving the reactions of metal-
imido species with excess organoazides have been discovered for
stoichiometric generation of diazenes,18,19 we propose for the
present system that the mechanism most consistent with these
product distributions is the bimolecular coupling of Fe-nitrene
species 4. Stoichiometric imido-imido coupling has been observed
recently.20
This finding is significant not only in the context of catalytic
nitrene transfer reactions, but also because nitrene-nitrene coupling
is conceptually related to oxo-oxo coupling, which may be relevant
to O-O bond forming processes for some water splitting catalysts.21
It has been suggested that oxygen-centered radical character is
crucial for such reactions to proceed.21 We hence chose to examine
the model complex [SiPMe3]Fe(NPh) (7) by DFT methods12 to
explore the degree of spin character carried by the NPh moeity.
The low spin, S ) 1/2 state (7-LS) was calculated to be the ground
state of 7, consistent with the EPR spectroscopy of 4-Tol and the
observed ground state of [SiPiPr3]Ru[N(p-C6H4CF3)].15 The opti-
mized geometry of 7-LS falls intermediate between TBP and square
pyramidal (τ ) 0.49)22 and is similar to the crystallographically
determined structure of [SiPiPr3]Ru[N(p-C6H4CF3)] (τ ) 0.54). The
Two chemical transformations further support the existence of
the proposed FeNAr intermediates (Scheme 1). First, when 3-Tol
t
was generated in the presence of BuNC, the resulting product
mixture contained TolNdNTol, 1, tBuNdCdNTol, and
[SiPiPr3]Fe(CNtBu) (5). Stoichiometric16 and catalytic17 reactions
9
4084 J. AM. CHEM. SOC. VOL. 132, NO. 12, 2010

C O M M U N I C A T I O N S
optimized Fe-N distance in 7-LS (1.703 Å) is much shorter than
that of anilido complex 6 (1.963(2) Å). Structurally characterized
Fe-NR bond distances range from 1.61-1.66 Å for tetrahedral
systems3,4 to 1.70-1.72 Å for distorted square planar systems.5
Low-spin [SiPiPr3]Ru[N(p-C6H4CF3)] has experimental and com-
putational characteristics implying significant radical delocalization
through the RuNAr π-system.15 Such delocalization is less obvious
for 7-LS, which has calculated spin densities (molecular sum 1.00)
on the Fe center and NPh unit of 0.89 and 0.16, respectively, with
very little spin density on the N atom itself (0.01). Noteworthy,
however, are the short N-C (1.341 Å) and alternating C-C bond
lengths in the calculated NPh unit of 7-LS that suggest a significant
quinoidal resonance contributor (see SI).
Supporting Information Available: Synthetic, spectroscopic, crys-
tallographic, and computational details. This material is available free
of charge via the Internet at http://pubs.acs.org.
References
(1) Cenini, S.; Gallo, E.; Caselli, A.; Ragaini, F.; Fantauzzi, S.; Piangiolino,
C. Coord. Chem. ReV. 2006, 250, 1234.
(2) (a) Eckert, N. A.; Vaddadi, S.; Stoian, S.; Flaschenriem, C. J.; Cundari,
T. R.; Holland, P. L. Angew. Chem., Int. Ed. 2006, 45, 6868. (b) Lucas,
R. L.; Powell, D. R.; Borovik, A. S. J. Am. Chem. Soc. 2005, 127, 11596.
(c) Jensen, M. P.; Mehn, M. P.; Que, L., Jr. Angew. Chem., Int. Ed. 2003,
42, 4357. (d) King, E. R.; Betley, T. A. Inorg. Chem. 2009, 48, 2361.
(3) (a) Brown, S. D.; Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2003, 125,
322. (b) Brown, S. D.; Peters, J. C. J. Am. Chem. Soc. 2005, 127, 1913.
(c) Thomas, C. M.; Mankad, N. P.; Peters, J. C. J. Am. Chem. Soc. 2006,
128, 4956.
(4) (a) Nieto, I.; Ding, F.; Bontchev, R. P.; Wang, H.; Smith, J. M. J. Am.
Chem. Soc. 2008, 130, 2716. (b) Verma, A. K.; Nazif, T. N.; Achim, C.;
Lee, S. C. J. Am. Chem. Soc. 2000, 122, 11013.
(5) Bart, S. C.; Lobkovsky, E.; Bill, E.; Chirik, P. J. J. Am. Chem. Soc. 2006,
128, 5302.
(6) Ni, C.; Fettinger, J. C.; Long, G. J.; Brynda, M.; Power, P. P. Chem.
Commun. 2008, 45, 6045.
(7) England, J.; Martinho, M.; Farquhar, E. R.; Frisch, J. R.; Bominaar, E. L.;
Mu¨nck, E.; Que, L., Jr. Angew. Chem., Int. Ed. 2009, 48, 3622.
(8) Betley, T. A.; Wu, Q.; Van Voorhis, T.; Nocera, D. G. Inorg. Chem. 2008,
47, 1849.
(9) (a) Hu, X.; Meyer, K. J. Am. Chem. Soc. 2004, 126, 16322. (b) Vogel, C.;
Heinemann, F. W.; Sutter, J.; Anthon, C.; Meyer, K. Angew. Chem., Int.
Ed. 2008, 47, 2681.
(10) Mankad, N. P.; Whited, M. T.; Peters, J. C. Angew. Chem., Int. Ed. 2007,
46, 5768.
(11) (a) Proulx, G.; Bergman, R. G. J. Am. Chem. Soc. 1995, 117, 6382. (b)
Proulx, G.; Bergman, R. G. Organometallics 1996, 15, 684. (c) Fickes,
M. G.; Davis, W. M.; Cummins, C. C. J. Am. Chem. Soc. 1995, 117, 6384.
(d) Guillemot, G.; Solari, E.; Floriani, C.; Rozzoli, C. Organometallics
2001, 20, 607. (e) Hanna, T. A.; Baranger, A. M.; Bergman, R. G. Angew.
Chem., Int. Ed. Engl. 1996, 35, 653. (f) Dias, H. V. R.; Polach, S. A.;
Goh, S.-K.; Archibong, D. F.; Marynick, D. S. Inorg. Chem. 2000, 39,
3894. (g) Barz, M.; Herdtweck, E.; Thiel, W. R. Angew. Chem., Int. Ed.
Engl. 1998, 37, 2262. (h) Albertin, G.; Antoniutti, S.; Baldan, D.; Castro,
J.; Garcia-Fontan, S. Inorg. Chem. 2008, 47, 742. (i) Waterman, R.;
Hillhouse, G. L. J. Am. Chem. Soc. 2008, 130, 12628.
(12) B3LYP, 6-31+G* for minimizations, 6-311++G** for final energy
calculations. Spin densities were calculated using Mu¨lliken population
analyses (see Supporting Information for further details).
(13) Though the complete mass balance was not accounted for in these product
mixtures, we note that dark highly insoluble material was evident upon
scale-up, suggesting the presence of oligomeric or polymeric material.
(14) Ragaini, F.; Penoni, A.; Gallo, E.; Tollari, S.; Gotti, C. L.; Lapadula, M.;
Mangioni, E.; Cenini, S. Chem.sEur. J. 2003, 9, 249.
(15) Takaoka, A.; Gerber, L. C. H.; Peters, J. C., submitted for publication.
(16) Brown, S. D. Ph.D. Thesis. California Institute of Technology, May, 2005.
(17) Cowley, R. E.; Eckert, N. A.; Elha¨ık, J.; Holland, P. L. Chem. Commun.
2009, 1760.
(18) (a) Hansert, B.; Vahrenkamp, H. J. Organomet. Chem. 1993, 459, 265. (b)
Harrold, N. D.; Waterman, R.; Hillhouse, G. L.; Cundari, T. R. J. Am.
Chem. Soc. 2009, 131, 12872.
(19) See Supporting Information for data regarding crossover experiments
yielding mixed diazenes.
Figure 4. X-band EPR spectra (77 K, 2-methyltetrahydrofuran) of (a) 1,
(b) 3-Tol, (c) 4-Tol produced by photolysis of a frozen glass of 3-Tol, (d)
simulated spectrum of 4-Tol, and (e) 1 produced upon warming and
refreezing 4-Tol; spin density plots (0.002 isocontours) of (f) 7-LS and (g)
7-IS.12
Interestingly, the intermediate spin, S ) 3/2 state (7-IS) of 7 was
higher in energy than 7-LS by only 2.8 kcal/mol. The calculated
spin densities for 7-IS (molecular sum 3.00) on the Fe, N, and Ph
units are 2.00, 0.82, and 0.15, respectively, indicating some radical
character for the imido ligand (Figure 4).12 Considering the
likelihood that, based on our DFT studies of 7, the [SiPiPr3]Fe(NAr)
intermediates 4 also possess low-lying excited states, the possibility
of two-state reactivity23 for 4 could account for the strikingly rich
redox chemistry that enables 1-electron (HAT), 2-electron (nitrene
transfer), and 4-electron (bimolecular coupling) redox processes
and merits further study.
(20) (a) Yiu, S.-M.; Lam, W. W. Y.; Ho, C.-M.; Lau, T.-C. J. Am. Chem. Soc.
2007, 129, 803. (b) Zarkesh, R. A.; Ziller, J. W.; Heyduk, A. F. Angew.
Chem., Int. Ed. 2008, 47, 4715. (c) Mansuy, D.; Battioni, P.; Mahy, J. P.
J. Am. Chem. Soc. 1982, 104, 4487.
(21) Yang, X.; Baik, M.-H. J. Am. Chem. Soc. 2006, 128, 7476.
(22) τ
) 0 for an ideal square pyramid and 1 for an ideal trigonal
bipyramid: Addison, A. W.; Rao, T. N.; Reedijk, J.; van Rijn, J.; Verschoor,
G. C. J. Chem. Soc., Dalton Trans. 1984, 1349.
(23) (a) Klinker, E. J.; Shaik, S.; Hirao, H.; Que, L., Jr. Angew. Chem., Int. Ed.
2009, 48, 1291. (b) Hirao, H.; Kumar, D.; Que, L., Jr.; Shaik, S. J. Am.
Chem. Soc. 2006, 128, 8590. (c) Dhuri, S. N.; Seo, M. S.; Lee, Y.-M.;
Hirao, H.; Wang, Y.; Nam, W.; Shaik, S. Angew. Chem., Int. Ed. 2008,
47, 3356.
Acknowledgment. This work was funded by the NIH (GM-
070757). Tim Kowalczyk and Prof. Seth Brown provided helpful
discussions.
JA910224C
9
J. AM. CHEM. SOC. VOL. 132, NO. 12, 2010 4085