Ground-state singlet L3Fe-(μ-N)-FeL3 and L 3Fe(NR) complexes featuring pseudotetrahedral Fe(II) centers

Article abstract of DOI:10.1021/ja0453073

Pseudotetrahedral iron(II) coordination complexes that contain bridged nitride and terminal imide linkages, and exhibit singlet ground-state electronic configurations, are described. Sodium amalgam reduction of the ferromagnetically coupled dimer, {[PhBP₃]Fe(μ-1,3-N ₃)}₂ (2) ([PhBP₃] = [PhB(CH₂PPh ₂)₃]^-), yields the diamagnetic bridging nitride species [{[PhBP₃]Fe}₂(μ-N)][Na(THF)₅] (3). The Fe-N-Fe linkage featured in the anion of 3 exhibits an unusually bent angle of approximately 135°, and the short Fe-N bond distances (Fe-N_av ≈ 1.70 A) suggest substantial Fe-N multiple bond character. The diamagnetic imide complex {[PhBP₃]Fe^∥≡N(1-Ad)} {nBu₄N} (4) has been prepared by sodium amalgam reduction of its low-spin iron(III) precursor, [PhBP₃]Fe^III≡N(1-Ad) (5). Complexes 4 and 5 have been structurally characterized, and their respective electronic structures are discussed in the context of a supporting DFT calculation. Diamagnetic 4 provides a bona fide example of a pseudotetrahedral iron(II) center in a low-spin ground-state configuration. Comparative optical data strongly suggest that dinuclear 3 is best described as containing two high-spin iron(II) centers that are strongly antiferromagnetically coupled to give rise to a singlet ground-state at room temperature.

Full text of DOI:10.1021/ja0453073

Published on Web 01/19/2005
Ground-State Singlet L3Fe-(µ-N)-FeL3 and L3Fe(NR)
Complexes Featuring Pseudotetrahedral Fe(II) Centers
Steven D. Brown and Jonas C. Peters*
Contribution from the DiVision of Chemistry and Chemical Engineering,
Arnold and Mabel Beckman Laboratories of Chemical Synthesis,
California Institute of Technology, Pasadena, California 91125
Received August 4, 2004; E-mail: jpeters@caltech.edu
Abstract: Pseudotetrahedral iron(II) coordination complexes that contain bridged nitride and terminal imide
linkages, and exhibit singlet ground-state electronic configurations, are described. Sodium amalgam
reduction of the ferromagnetically coupled dimer, {[PhBP3]Fe(µ-1,3-N3)}2 (2) ([PhBP3] ) [PhB(CH2PPh2)3]-),
yields the diamagnetic bridging nitride species [{[PhBP3]Fe}2(µ-N)][Na(THF)5] (3). The Fe-N-Fe linkage
featured in the anion of 3 exhibits an unusually bent angle of approximately 135°, and the short Fe-N
bond distances (Fe-Nav ≈ 1.70 Å) suggest substantial Fe-N multiple bond character. The diamagnetic
imide complex {[PhBP3]FeIItN(1-Ad)}{nBu4N} (4) has been prepared by sodium amalgam reduction of its
low-spin iron(III) precursor, [PhBP3]FeIIItN(1-Ad) (5). Complexes 4 and 5 have been structurally
characterized, and their respective electronic structures are discussed in the context of a supporting DFT
calculation. Diamagnetic 4 provides a bona fide example of a pseudotetrahedral iron(II) center in a low-
spin ground-state configuration. Comparative optical data strongly suggest that dinuclear 3 is best described
as containing two high-spin iron(II) centers that are strongly antiferromagnetically coupled to give rise to a
singlet ground-state at room temperature.
Introduction
examples of bimetallic Fe-(µ-N)-Fe species if one restricts
consideration to iron centers that feature coordination numbers
Low coordinate iron sites that feature terminally bound or
bridged nitride ligands may be relevant to various nitrogen
reduction schemes.1 For example, the report of an interstitial
µ6-ligand (presumably a nitride) within the MoFe-cofactor of
nitrogenase2 has drawn recent attention to Fe-N-Fe structure
types germane to biological nitrogen fixation.3 Schemes have
also been forwarded to suggest that an N2 scission process on
hot iron surfaces might lead to surface-bound iron nitrides that
are reactive toward hydrogen during the Haber-Bosch ammonia
synthesis.4 In this broad context, it is noteworthy that syntheti-
cally well-defined molecular iron nitrides, whether terminal or
bridging, are uncommon.5,8 For instance, there are presently no
lower than five. Moreover, terminal nitrides of iron, regardless
of coordination number, are exceptionally rare.8,11 Well-defined,
low coordinate iron nitrides that can be thoroughly characterized
are therefore of synthetic interest, and we have begun to
undertake their systematic development.
As a preface to the current study, our laboratory has invested
considerable effort in the synthesis and study of various
monomeric pseudotetrahedral L3Fe-Nx platforms to assess what
range of Nx-type ligands, oxidation states, and redox processes
might be accommodated by a single pseudotetrahedral iron site.
Using tris(phosphino)borate ligands as the L3 auxiliary,6 we have
observed that an L3Fe-Nx center can display a remarkable range
of redox flexibility, supporting π-acidic N27 and π-basic nitride8
(N3-) at the extremes, but also a range of intermediate Nx-type
(1) (a) Huniar, U.; Ahlrichs, R.; Coucouvanis, D. J. Am. Chem. Soc. 2004,
126, 2588. (b) Hinnemann, B.; Nørskov, J. K. J. Am. Chem. Soc. 2004,
126, 3920. (c) Lee, H.-I.; Benton, P. M. C.; Laryukhin, M.; Igarashi, R.
Y.; Dean, D. R.; Seefeldt, L. C.; Hoffman, B. M. J. Am. Chem. Soc. 2003,
125, 5604. (d) Schimpl, J.; Petrilli, H. M.; Blo¨chl, P. E. J. Am. Chem. Soc.
2003, 125, 15772. (e) Cao, Z.; Zhou, Z.; Wan, H.; Zhang, Q.; Thiel, W.
Inorg. Chem. 2003, 42, 6986. (f) Hinnemann, B.; Nørskov, J. K. J. Am.
Chem. Soc. 2003, 125, 1466. (g) Dance, I. Chem. Commun. 2003, 324.
(2) Einsle, O.; Tezcan, F. A.; Andrade, S. L. A.; Schmid, B.; Yoshida, M.;
Howard, J. B.; Rees, D. C. Science 2002, 297, 1696.
ligands that include amide,9 imide,7,10 and diazenido.7
A
characteristic feature of these L3Fe-Nx species that feature
strong Fe-Nx π-bonding is their propensity to populate a low-
spin ground-state electronic configuration. For instance, the
complexes [PhBPiPr3]FeIVtN and [PhBPiPr3]FeIIItN(1-Ad) are
characterized by one σ and two π bonds at the Fe-Nx linkage
(3) Lee, S. C.; Holm, R. H. Chem. ReV. 2004, 104, 1135.
(4) Ertl, G. Chem. Rec. 2001, 1, 33.
(5) For examples of structurally characterized Fe-N-Fe linkages, see: (a)
Ju¨stel, T.; Mu¨ller, M.; Weyhermu¨ller, T.; Kressl, C.; Bill, E.; Hildebrandt,
P.; Lengen, M.; Grodzicki, M.; Trautwein, A. X.; Nuber, B.; Wieghardt,
K. Chem.-Eur. J. 1999, 5, 793. (b) Kienast, A.; Homborg, H. Z. Anorg.
Allg. Chem. 1998, 624, 233. (c) Moubaraki, B.; Benlian, D.; Baldy, A.;
Pierrot, M. Acta Crystallogr., Sect. C 1989, 45, 393. (d) Scheidt, W. R.;
Summerville, D. A.; Cohen, I. A. J. Am. Chem. Soc. 1976, 98, 6623. (e)
Summerville, D. A.; Cohen, I. A. J. Am. Chem. Soc. 1976, 98, 1747.
(6) (a) Thomas, J. C.; Peters, J. C. Inorg. Synth. 2004, 34, 8. (b) Barney, A.
A.; Heyduk, A. F.; Nocera, D. G. Chem. Commun. 1999, 2379.
(7) Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2003, 125, 10782.
(8) Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2004, 126, 6252.
(9) Brown, S. D.; Peters, J. C. J. Am. Chem. Soc. 2004, 126, 4538.
(10) Brown, S. D.; Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2003, 125,
322-323.
(11) High-spin iron(V) terminal nitrides have been reported on the basis of
spectroscopic data at low temperatures. See: (a) Meyer, K. M.; Eckhard,
B.; Mienert, B.; Weyhermu¨ller, T.; Wieghardt, K. J. Am. Chem. Soc. 1999,
121, 4859. (b) Wagner, W.-D.; Nakamoto, K. J. Am. Chem. Soc. 1989,
111, 1590.
9
10.1021/ja0453073 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 1913-1923
1913

A R T I C L E S
Scheme 1
Brown and Peters
and exhibit S ) 0 and S ) 1/2 ground states, respectively, despite
reveals a singlet at δ 44.5 ppm, consistent with the formation
of a C3-symmetric solvento adduct that we formulate as
{[PhBP3]Fe(CH3CN)3}{Cl}. A related species, {[PhBP3]Fe-
(CH3CN)3}{PF6}, has been isolated and characterized.16 The
1H NMR spectrum (CD3CN) reveals, in addition to the major
diamagnetic product, the presence of a minor paramagnetic
species that we tentatively formulate as [PhBP3]Fe(CH3CN)(Cl).
A similar five-coordinate iron-containing species has been
characterized with the [PhBPiPr3] ligand.12b Upon addition of 1
equiv of [Na][N3] to a red acetonitrile solution of 1, the gradual
precipitation of purple solids occurs over a period of hours.
Extraction of these solids with copious amounts of benzene and
subsequent filtration through a pad of Celite yields {[PhBP3]-
Fe(µ-1,3-N3)}2 (2) as a reddish-brown solid in 85% yield.
Installation of a terminal azide functionality is readily confirmed
by solution-state infrared spectroscopy due to the presence of a
the fact that the iron centers are pseudotetrahedral ([PhBPiPr
]
3
) [PhB(CH2PiPr2)3]-).12
In this report, we describe a new bimetallic L3Fe-Nx-FeL3
complex in which the L3 scaffold is the parent [PhBP3] ligand
([PhBP3] ) [PhB(CH2PPh2)3]-)13 and the Nx ligand is a bridging
nitride group. As alluded to above, previously prepared bridged
nitride diiron systems have featured five- or six-coordinate iron
centers in relatively high oxidation states (e.g., FeIII(µ-N)FeIV,
FeIV(µ-N)FeIV).5 The system described herein, [{[PhBP3]Fe}2-
(µ-N)][Na(THF)5] (3), is distinct from these complexes with
respect to the coordination number and the formal oxidation
state of each of its iron centers. The divalent iron sites in 3 are
pseudotetrahedral iron(II), and the complex is diamagnetic in
solution even at room temperature. A related mononuclear Fe(II)
complex can be prepared that features a bona fide low-spin
ground-state, {[PhBP3]FeIItN(1-Ad)}{nBu4N} (4). However,
comparative spectroscopic data indicate that 3 is best described
by two high-spin Fe(II) centers that exhibit such strong
antiferromagnetic coupling that a singlet state is exclusively
populated at 293 K.
strong vibration at 2077 cm-1 17
Repeating the synthesis with
.
[Na][15NNN] yields 2-15N in which the azide vibration is shifted
to 2066 cm-1. Dilute solutions of 2 are transparent yellow in
color with optical data (λ ) 410 nm, ꢀ ) 700 M-1 cm-1) similar
to that observed for the monomeric chloride 1. Evans method18
magnetic measurements on crystalline samples of 2 dissolved
in C6D6 repeatedly afforded magnetic moments of 4.50((0.04)
µΒ, a value somewhat below the spin-only value of 4.89 µΒ for
four unpaired electrons. Other pseudotetrahedral [PhBPiPr3]Fe(II)
complexes, for example, [PhBPiPr3]Fe-Me, also exhibit moments
below the spin-only value.19 We note that a monomer-dimer
equilibrium process occurs for 2 in solution, as reflected by a
color change from yellow to red upon cooling, and the dimeric
form has been characterized in the solid-state by X-ray diffrac-
tion and SQUID magnetometry.
Results and Discussion
Whereas the reaction between [PhBPiPr3]FeCl and [Li][dbabh]14
(dbabh ) 2,3:5,6-dibenzo-7-aza bicycle[2.2.1]hepta-2,5-diene)
cleanly generates an amide intermediate, [PhBPiPr3]Fe(dbabh),
that subsequently decays to [PhBPiPr3]FeIVtN,8 the reaction
between [PhBP3]FeCl (1)10 and [Li][dbabh] is ill-defined and
no evidence for a terminally bound nitride species is observed.
To examine N-atom transfer to this latter system, we focused
on the installation of azide (N3-) as a suitable N3- source.15
Access to the required iron(II) azide precursor is readily
accomplished via metathesis between 1 and sodium azide in
acetonitrile (Scheme 1). Dissolution of yellow, paramagnetic 1
in CH3CN gives rise to a red solution in which complete
consumption of 1 is evident by 1H and 31P{1H} NMR
spectroscopy. The 31P{1H} NMR spectrum of this solution
Red crystals of 2 suitable for an X-ray diffraction experiment
were grown via a benzene/petroleum ether vapor diffusion
chamber, and its solid-state structure is shown in Figure 1. Azide
2 crystallizes in the triclinic crystal system with a benzene
solvent molecule (omitted from Figure 1 for clarity) and features
a dimeric Fe2(µ-1,3-N3)2 core that possesses a crystallographic
center of symmetry.20 The coordination geometry of each iron
center is therefore identical and is best described as distorted
square pyramidal (τ ) 0.4)21 in which N1, N3, P2, and P3
comprise the basal plane. As a result of the vacant coordination
site trans to P1, the Fe-P1 bond distance of 2.1610(5) Å is
(12) For other examples of chemistry featuring [PhBPiPr3], see: (a) Turculet,
L.; Feldman, J. D.; Tilley, T. D. Organometallics 2004, 23, 2488. (b) Betley,
T. A.; Peters, J. C. Inorg. Chem. 2003, 42, 5074. (c) Turculet, L.; Feldman,
J. D.; Tilley, T. D. Organometallics 2003, 22, 4627.
(13) For other examples of chemistry featuring [PhBP3], see: (a) Jenkins, D.
M.; Peters, J. C. J. Am. Chem. Soc. 2003, 125, 11162. (b) Jenkins, D. M.;
Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2002, 124, 11238. (c) Jenkins,
D. M.; Di Bilio, A. J.; Allen, M. J.; Betley, T. A.; Peters, J. C. J. Am.
Chem. Soc. 2002, 124, 15336. (d) Feldman, J. D.; Peters, J. C.; Tilley, T.
D. Organometallics 2002, 21, 4065. (e) Feldman, J. D.; Peters, J. C.; Tilley,
T. D. Organometallics 2002, 21, 4050.
(16) See the Supporting Information for the synthesis and characterization of
{[PhBP3]Fe(CH3CN)3}{PF6}.
(14) (a) Mindiola, D. J.; Cummins, C. C. Angew. Chem., Int. Ed. 1998, 37,
945. (b) Carpino, L. A.; Padykula, R. E.; Barr, D. E.; Hall, F. H.; Krause,
J. G.; Dufresne, R. F.; Thoman, C. J. J. Org. Chem. 1988, 53, 2565.
(15) (a) Du Bois, J.; Tomooka, C. S.; Hong, J.; Carreira, E. M. Acc. Chem.
Res. 1997, 30, 364. (b) Nugent, W. A.; Mayer, J. M. Metal-Ligand Multiple
Bonds; Wiley & Sons: New York, 1988; p 76. (c) LaMonica, G.; Cenini,
S. J. Chem. Soc., Dalton Trans. 1980, 1145.
(17) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination
Compounds. Part B: Applications in Coordination, Organometallic, and
Bioinorganic Chemistry, 5th ed.; Wiley & Sons: New York, 1997; p 124.
(18) (a) Sur, S. K. J. Magn. Reson. 1989, 82, 169. (b) Evans, D. F. J. Chem.
Soc. 1959, 2003.
(19) Daida, E. J.; Peters, J. C. Inorg. Chem. 2004, 43, 7474. Also see ref 12c
for a [PhBPiPr3]FeII-SiR3 example.
9
1914 J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005

Pseudotetrahedral Iron(II) Coordination Complexes
A R T I C L E S
Figure 1. Displacement ellipsoid representation (50%) of 2. A benzene
solvent molecule has been omitted for clarity. Selected bond distances (Å)
and angles (deg): Fe-N1, 2.0107(15); Fe-N3, 1.9516(16); N1-N2,
1.181(2); N2-N3A, 1.175(2); Fe-P1, 2.1610(5); Fe-P2, 2.2279(6); Fe-
P3, 2.2101(6); P1-Fe-P2, 89.37(2); P1-Fe-P3, 92.23(2); P1-Fe-N1,
94.52(5); P1-Fe-N3, 116.91(5); P2-Fe-P3, 89.88(2); N3-Fe-N1,
87.06(6); N1-Fe-P2, 175.64(5); N1-Fe-P3, 91.92(5); P3-Fe-N3,
150.84(5).
slightly shorter than that observed for Fe-P2 (2.2279(6) Å) and
Fe-P3 (2.2101(6) Å). The N-N bond distances in each of the
azide bridges are nearly identical at 1.181(2) Å (N1-N2 and
N1A-N2A) and 1.175(2) Å (N2-N3A and N2A-N3).
Solid-state magnetic susceptibility data for 2 were obtained
from 4 to 290 K by SQUID magnetometry and are plotted, per
dimeric unit, in Figure 2. The plot of øMT verses temperature
(Figure 2A) is consistent with ferromagnetic coupling as a
marked increase in the value of øMT is observed as the sample
is cooled. A maximum value of 8.31 cm3 K mol-1 is observed
at 40 K, and further cooling of the sample results in a decrease
in the value of øMT which is likely attributable to zero-field
splitting.22 The plot of µeff versus temperature (Figure 2B)
establishes a room-temperature magnetic moment of 4.3 µB for
2, while a maximum value of 8.2 µB is observed at 40 K.
As shown in Scheme 1, the chemical reduction of 2 is readily
accomplished in THF with 1 equiv of a 0.2 wt % Na/Hg
amalgam. A darkening of the reaction solution is accompanied
by the effervescence of gas almost immediately upon exposure
of 2 to the amalgam. After the mixture was stirred at room
temperature for 8 h, the major product of the reaction,
[{[PhBP3]Fe}2(µ-N)][Na(THF)5] (3), was isolated as a brown
solid in ca. 50% yield. Crystals of 3 suitable for X-ray diffraction
analysis were grown via a THF/hexanes vapor diffusion
chamber, and its solid-state structure is shown in Figure 3A.
Compound 3 crystallizes in the monoclinic crystal system with
a free molecule of THF (omitted from Figure 3 for clarity).
Figure 2. SQUID magnetometry data for 2 plotted per dimeric unit as (A)
øMT versus temperature and (B) µeff versus temperature.
The nitride ligand was refined in two positions (Figure 3B and
C) and modeled satisfactorily with a population ratio of 1.86 to
1. Repeating the X-ray diffraction experiments on several
independently synthesized and crystallized samples consistently
afforded data in which the nitride ligand was located in two
positions, including samples in which the sodium cation had
been exchanged for a nBu4N+ cation. Combustion analysis data
for 3 is consistent with a single µ-N ligand.
In contrast to precursor 2, each iron center in the anion of 3
is crystallographically unique and exhibits a geometry that is
best described as distorted tetrahedral. The relatively short Fe-
Nav bond distance of 1.70 Å for both positions of the nitride
ligand is consistent with multiple bonding character.7,10,23 One
of the dramatic structural differences between 3 and other
literature examples of diiron µ-nitride systems concerns the Fe-
N-Fe bond angle of 3, which averages 135° for either position
of the nitride ligand.24 All previous examples of structurally
characterized bridging iron nitrides feature linear Fe-N-Fe
linkages.5 In fact, to the best of our knowledge, the only other
report of a mono-bridged bimetallic nitride species featuring
an appreciably bent M-N-M linkage is that from Floriani and
co-workers in which the anion of [{p-tBu-calix[4]-(O)4}2Nb2-
(µ-N)][Na(TMEDA)2] (TMEDA ) N,N,N′,N′-tetramethyleth-
ylenediamine) was reported to have a bond angle of 145.2(1)°.25
(20) To the best of our knowledge, examples of µ-(1,3) bridging azide systems
for iron have yet to be reported. For examples of µ-(1,1) systems, see: (a)
Clemente-Juan, J. M.; Mackiewicz, C.; Verelst, M.; Dahan, F.; Bousseksou,
A.; Sanakis, Y.; Tuchagues, J.-P. Inorg. Chem. 2002, 41, 1478. (b) Reddy,
K. R.; Rajasekharan, M. V.; Tuchagues, J.-P. Inorg. Chem. 1998, 37, 5978.
(c) De Munno, G.; Poerio, T.; Viau, G.; Julve, M.; Lloret, F. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1459. (d) De Munno, G.; Poerio, T.; Viau, G.;
Julve, M.; Lloret, F. Chem. Commun. 1996, 2587.
(21) A τ value of 1 corresponds to an ideal trigonal bipyramidal geometry, while
a τ value of 0 corresponds to an ideal square pyramidal geometry. See:
Addison, A. W.; Rao, T. N.; Reedijk, J.; van Rin, J.; Verschoor, G. C. J.
Chem. Soc., Dalton Trans. 1984, 1349.
(22) The decrease in øMT below 16 K for [R-(Me3tacn)2(cyclam)NiMo2(CN)6]I2,
which features a ferromagnetically coupled S ) 4 electronic configuration,
was also attributed to zero-field splitting. See: Shores, M. P.; Sokol, J. J.;
Long, J. R. J. Am. Chem. Soc. 2002, 124, 2279. For a more detailed
discussion concerning the magnetism of differous bridged systems, see:
Hendrich, M. P.; Day, E. P.; Wang, C.-P.; Synder, B. S.; Holm, R. H.;
Mu¨nck, E. Inorg. Chem. 1994, 33, 2848 and references therein.
(23) Verma, A. K.; Nazif, T. N.; Achim, C.; Lee, S. C. J. Am. Chem. Soc. 2000,
122, 11013.
(24) We have considered the presence of a (µ-NH) functionality instead of a
(µ-N) ligand. However, IR studies of 3 as both a solution (THF) and a
solid (Nujol, KBr pellet) provide no evidence for an N-H vibration.
Additionally, it is not possible to assign formal iron oxidation states with
a (µ-NH) ligand and still maintain a diamagnetic manifold while accounting
for all charges.
(25) Caselli, A.; Solari, E.; Scopelliti, R.; Floriani, C.; Re, N.; Rizzoli, C.; Chiesi-
Villa, A. J. Am. Chem. Soc. 2000, 122, 3652.
9
J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005 1915

A R T I C L E S
Brown and Peters
Figure 3. (A) Displacement ellipsoid representation (50%) of 3 showing only the more populated position of the disordered nitride ligand. A free molecule
of THF has been omitted for clarity. Selected bond distances (Å) and angles (deg): Fe1-N1A, 1.675(5); Fe2-N1A, 1.705(5); Fe1-P1, 2.2295(10); Fe1-
P2, 2.2338(10); Fe-P3, 2.2127(10); Fe2-P4, 2.2299(10); Fe2-P5, 2.2106(10); Fe2-P6, 2.2534(10); Fe1-N1A-Fe2, 135.9(3); P1-Fe1-P2, 92.10(4);
P1-Fe-P3, 88.83(4); P2-Fe1-P3, 90.79(4); P2-Fe1-N1A, 143.93(16); P1-Fe1-N1A, 112.69(14); P3-Fe1-N1A, 114.45(15). (B) Side view of the
Fe1-N-Fe2 core of 3 showing both positions of the disordered nitride ligand. (C) Top view of the Fe1-N-Fe2 core of 3 showing both positions of the
disordered nitride ligand.
This type of bent structure is reminiscent of R3PdNdPR3
cations, which feature bridging nitride ligands sandwiched by
four-coordinate phosphorus centers. Examples of R3PdNdPR3
cations in which the P-N-P bond angle is approximately 135°
and the two N-P bond distances are similar, but crystallo-
graphically distinct, have been reported.26 Direct Fe-Fe bonding
interactions within 3 are unlikely given their separation by ca.
3.13 Å.27
the assignment of a diamagnetic ground state, nitride 3 is X-band
EPR silent at 4 K as a glassy 1-methyltetrahydrofuran solution.
Two limiting electronic descriptions for complex 3 need to
be considered. It can be electronically described by two localized
low-spin S ) 0 Fe(II) centers, or alternatively as a fully
delocalized system that features two high-spin S ) 2 Fe(II)
centers that are sufficiently antiferromagnetically coupled that
a singlet ground state is populated even at room temperature.
In accord with the latter description, Holland, Mu¨nck, and co-
workers recently reported a sulfido-bridged diiron(II) compound
Characteristic NMR resonances for triply recrystallized
samples of nitride 3 are observed by 1H, 31P{1H}, and 15N NMR
spectroscopy (a 50% 15N-enriched sample of 3 may be generated
from the reduction of 2-15N), consistent with the assignment of
a singlet ground state. The 31P{1H} spectrum features a singlet
at δ 58.0 ppm, and the 1H NMR spectrum in THF-d8 displays,
in addition to several equivalents of THF due to the sodium
cation, one set of diamagnetic resonances corresponding to the
[PhBP3] ligand. A singlet at δ 801 ppm (vs NH3 at δ 0 ppm) in
the 15N NMR spectrum of 3 confirms the presence of the nitride
functionality. We were unable to resolve 15N-31P coupling,
similar to the case of the terminal nitride derivative [PhBPiPr3]-
FeIVtN8 where an 15N NMR signal was observed at δ 952 ppm.
The downfield nature of the nitride resonance in comparison
with other mono-bridged systems28 likely reflects a paramagnetic
contribution to the chemical shift from the anti-ferromagnetically
coupled high-spin ferrous centers (vide infra). In accord with
1
that features paramagnetic solution chemical shifts in the H
NMR spectrum at room temperature, but is assigned to an S )
0 ground state due to antiferromagnetic coupling on the basis
of low-temperature Mo¨ssbauer data.29 In addition, Wieghardt,
Trautwein, and co-workers have described an interesting FeIV(µ-
N)FeIV complex that is diamagnetic in solution at room
temperature, suggesting strong antiferromagnetic coupling and
a lower |J| value limit of 250 cm-1 5a
.
To further examine this issue, we prepared a mononuclear
d6 iron(II) complex exhibiting a related pseudotetrahedral
geometry with one π-bonded Nx type linkage for comparison.
This complex, {[PhBP3]FeIItN(1-Ad)}{nBu4N} (4), is diamag-
netic and clearly establishes that an isolated pseudotetrahedral
Fe(II) center can adopt a low-spin ground-state electronic
configuration. Complex 4 was easily prepared by Na/Hg
1
amalgam reduction of S )
/
[PhBP3]FeIIItN(1-Ad) (5),
2
followed by the addition of [nBu4N][Br]. The trivalent imide
precursor 5 was itself readily prepared by the addition of
1-adamantyl azide to [PhBP3]Fe(PPh3), a procedure analogous
to that we used in the preparation of the related imide
(26) (a) Glidewell, C.; Holden, H. D. J. Organomet. Chem. 1982, 226, 171. (b)
Canziani, F.; Garlaschelli, L.; Malatesta, M. C.; Albinati, A. J. Chem. Soc.,
Dalton Trans. 1981, 2395.
(27) (a) Shannon, R. D. Acta. Crystallogr. 1976, A32, 751. (b) Shannon, R. D.;
Prewitt, C. T. Acta. Crystallogr. 1969, B25, 925.
(28) See, for example: (a) Cherry, J.-P. F.; Stephens, F. H.; Johnson, M. J. A.;
Diaconescu, P. L.; Cummins, C. C. Inorg. Chem. 2001, 40, 6860. (b) Tsai,
Y.-C.; Johnson, M. J. A.; Mindiola, D. J.; Cummins, C. C. J. Am. Chem.
Soc. 1999, 121, 10426. (c) Banaszak Holl, M. M.; Kersting, M.; Pendley,
B. D.; Wolczanski, P. T. Inorg. Chem. 1990, 29, 1518.
(29) Vela, J.; Stoian, S.; Flaschenriem, C. J.; Mu¨nck, E.; Holland, P. L. J. Am.
Chem. Soc. 2004, 126, 4522.
9
1916 J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005

Pseudotetrahedral Iron(II) Coordination Complexes
A R T I C L E S
anomaly to π-back-bonding. These explanations notwithstand-
ing, the magnitude between the respective Fe-P bond distances
in 4 and 5 is noteworthy.
The average of the Fe-N bond distances in 3 is ap-
proximately 0.05 Å longer than that observed for 4 and 5.
Additionally, the Fe-P bond distances observed for 3 (Fe-
Pavg. ) 2.228 Å) are intermediate to those observed for the
structures of 4 (Fe-Pavg. ) 2.142 Å) and 5 (Fe-Pavg. ) 2.254
Å). By contrast, the recently communicated high-spin Fe(II)
amide complex [PhBP3]Fe(N(H)-p-tolyl)9 features diminished
π-bonding character at the Fe-N linkage in comparison to 3,
4, and 5 and exhibits expanded Fe-N and Fe-P distances due
to its preferred S ) 2 ground-state electronic configuration
(Fe-N ) 1.913 (2) Å, Fe-Pavg. ) 2.424 Å). On the basis of
these structural considerations, the tabulation of charges for 3,
and the fact that it exhibits a diamagnetic ground state, it might
seem most reasonable to assign each of the iron centers of 3 as
localized and low-spin Fe(II). A comparison of the electronic
absorption spectra for compounds 1, 3, and 4, however, seems
to suggest that this is not the best description.
Electronic absorption spectra for compounds 1, 3, and 4 were
collected from 350 to 2000 nm at room temperature in THF-
d8, and the resulting spectra are shown in Figure 5. Chloride 1
(Figure 5, inset), which represents a classic example of high-
spin iron(II), exhibits an absorption feature at 1285 nm (7782
cm-1, ꢀ ) 185 M-1 cm-1) that is reflective of its expected low-
energy d-d transitions.33 The d-d transitions of similar energy
are not expected for a low-spin iron(II) center due to a large
HOMO-LUMO gap, nor are charge-transfer transitions, which
would be expected to occur at much higher energy. Indeed, no
d-d type or charge-transfer type transitions are observed in the
NIR region of the spectrum for the rigorously low-spin Fe(II)
imide 4; a feature at ca. 510 nm (19 608 cm-1, ꢀ ) 2112 M-1
cm-1) is likely reflective of an LMCT transition. Inspection of
the NIR region of the spectrum for complex 3, however, reveals
a very broad and intense feature centered at approximately 1120
nm (8929 cm-1, ꢀ ) 3720 M-1 cm-1). This feature, which we
assume arises from a charge-transfer transition associated with
the bridged nitride ligand, displays an easily discernible shoulder
on its low-energy side that reflects the low-energy d-d
transitions characteristic of high-spin Fe(II) and observed in 1.
Curve-fitting of the NIR data for 3 from 850 to 2000 nm using
GRAMS v.3.2 (Thermo Galactic, see Supporting Information)
resolves the two features into a transition centered at 1120 nm
(8929 cm-1, ꢀ ) 2740 M-1 cm-1) and one centered at 1266
nm (7900 cm-1, ꢀ ) 943 M-1 cm-1).
Figure 4. Solid-state molecular structures of 4 (left) and 5 (right) showing
50% displacement ellipsoid representations for the P3FetN(1-Ad) core.
For 4, a THF solvent molecule and the nBu4N+ cation have been removed
for clarity. Selected bond lengths and angles are highlighted within the
figure.
[PhBP3]FeIIItN(p-tolyl).10 Complex 4 represents what is, to the
best of our knowledge, the only reported example of a d6 iron
imide,30 and is moreover the first example of a monomeric,
pseudotetrahedral S ) 0 Fe(II) complex.
The solid-state structures of both 4 and 5 have been
determined for comparison with one another and with nitride
3, and their core structures are shown in Figure 4 (for complete
structures of 4 and 5, see the Supporting Information). Most
striking is the overall similarity of the structures of 4 and 5 at
the Fe1-N1-C46 linkage, with respect to both the Fe1-N1
bond distance and the Fe1-N1-C46 bond angle, despite their
different spin states. Given this similarity, a structural distinction
worth noting is that the average Fe-P bond distances in the
divalent anion of 4 are ca. 0.1 Å shorter than the average of the
Fe-P bond distances for neutral, trivalent 5. This difference is
somewhat counterintuitive and may reflect the presence of an
unpaired spin in the case of 5, although this explanation is
unsatisfying given the predominantly nonbonding nature of the
2
2
,
orbital that would house the unpaired electron (dxy or dx -y
vide infra). An alternative possibility to further consider is that
the anion of 4 more strongly π-back-bonds into σ* orbitals of
the appropriate symmetry from the [PhBP3] ligand than for the
case of 5 by virtue of its d6 (versus d5) electronic configuration.31
This scenario would give rise to shorter Fe-P interactions. Also
consistent with a π-back-bonding argument is the observation
that the P-C bonds of anionic 4 are, on average, 0.016 Å longer
than those for neutral 5. A similar phenomenon has been
observed for the complex [(PP3)Fe(CtCPh)][BPh4] (PP3 )
P(CH2CH2PPh2)3),32 in which the one-electron reduction of this
species induces a contraction of the Fe-P bond distances by
an average of 0.09 Å. The authors also attributed this structural
It is interesting to contemplate the cause of the low-spin
configurations in complexes 4 and 5. According to the d-orbital
splitting scheme that we have previously used to describe the
d6 cobalt imide complex [PhBP3]CotN(p-tolyl),13b an appropri-
ate electronic configuration for the simplest case, d6 4, is
2
2
2 4
0
2
(dz ) (xy, x -y ) (xz,yz) . Under idealized C3V symmetry, these
2
d-orbitals transform as a1 + 2e. The dz orbital is expected to
lie low in energy, very close to the nearly degenerate pair of
d-orbitals that are oriented perpendicular to the Fe-N bond
vector. A DFT calculation on the slightly simplified complex
{[PhBP3]FetN(tBu)}- provides a minimized geometrical (Fig-
(30) For an example of a d6 iridium imide, see: Glueck, D. S.; Wu, J.; Hollander,
F. J.; Bergman, R. G. J. Am. Chem. Soc. 1991, 113, 2041.
(31) Orpen, A. G.; Connelly, N. G. Organometallics 1990, 9, 1206.
(32) Bianchini, C.; Laschi, F.; Masi, D.; Ottaviani, F. M.; Pastor, A.; Peruzzini,
M.; Zanello, P.; Zanobini, F. J. Am. Chem. Soc. 1993, 115, 2723.
2+
(33) For example, the 5E f 5T2 transition for Fe(HMPA)4 (HMPA )
hexamethylphosphoramide) occurs at 6950 cm-1. See: Lever, A. B. P.
Inorganic Electronic Spectroscopy, 2nd ed.; Elsevier: Amsterdam, 1984;
p 461.
9
J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005 1917

A R T I C L E S
Brown and Peters
Figure 5. Electronic absorption spectrum of 1 (inset, blue), 3 (green), and 4 (red) in THF-d8.
Figure 6. (A) Theoretically predicted geometry and (B) electronic structure (DFT, JAGUAR 5.0, B3LYP/LACVP**) for S ) 0 {[PhBP3]FetN(tBu)}-. (C)
Lobal representations of the highest occupied molecular orbital (HOMO), the orbital of dz2 parentage (HOMO - 2), and the dxz orbital (LUMO + 1).
Structural parameters: Fe-P ) 2.209, 2.213, 2.213 Å; Fe-N ) 1.651 Å; Fe-N-C ) 179.45°; N-Fe-P ) 125.08°, 125.39°, 125.63°; P-Fe-P ) 89.72°,
89.88°, 89.97°.
ure 6A) and electronic structure picture fully consistent with
this view. As shown in Figure 6B, a two-over-three splitting
µB). Because the electron is removed from an essentially
nonbonding orbital, little structural change is expected at the
Fe-N-C linkage, fully consistent with the solid-state structures
of 4 and 5. That the Fe-P bond distances in fact contract on
reducing 5 to 4 is not as readily explained (vide supra).
The comparative cyclic voltammetry of 3 and 5 is of obvious
interest and has been studied in THF (Fc/Fc+, 0.30 M
[nBu4N][PF6], 50 mV/s; Fc ) ferrocene). The data acquired are
shown in Figure 7. Inspection of the electrochemical response
of 3 reveals an irreversible oxidative wave at -350 mV, a fully
reversible redox process centered at -1340 mV, and another
redox process at -2520 mV that appears to be quasi-reversible.
2
2
2
diagram is obtained in which dz , dxy, and dx -y are nearly
degenerate and nonbonding, and a pair of unoccupied dxz, dyz
orbitals that are both σ* and π* in character lie at much higher
2
energy. The calculation predicts the orbital of dz parentage to
lie lowest in energy (HOMO - 2), and the HOMO is therefore
2
2
predicted to be either dxy or dx -y . Lobal representations for
the HOMO, HOMO - 2, and LUMO + 1 orbitals are shown
in Figure 6C. Hypothetical removal of one electron from this
system to generate Fe(III) is expected to give rise to a doublet
ground-state species, which is observed for imide 5 (µeff ) 1.98
9
1918 J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005

Pseudotetrahedral Iron(II) Coordination Complexes
A R T I C L E S
and Cohen in the context of the photolysis products of a
tetrphenylporphinatoiron(III) azide,5e and also by Wieghardt,
Trautwein, and co-workers.5a We note that repeating the
reduction of 2 in the presence of excess PPh3 results in the
formation of a substantial quantity of the previously reported
Fe(I) complex [PhBP3]Fe(PPh3),10 in addition to nitride 3.
Because complex 2 itself does not react with PPh3, it appears
likely that PPh3 traps an Fe(I) source that is generated in-situ.
We can therefore put forth the notion that “{[PhBP3]-
FeIII(N)}-” rapidly condenses with an “[PhBP3]FeI(solv)” spe-
cies to form isolable 3, but the speculative nature of this
suggestion needs to be underscored. One way to identify
{[PhBP3]FeIII(N)}- as an intermediate would involve trapping
it by addition of an appropriate electrophile. Accordingly, we
have tried to reduce the azide precursor 2 in the presence of
trimethylsilyl chloride. While we do not know the nature of
the reaction products, it is quite clear from spectroscopic data
that [PhBP3]FeIII(NSiMe3), the simplest product we might expect
to form via metathesis (and elimination of NaCl and N2), is not
among these products.
Figure 7. Cyclic voltammetry of [{[PhBP3]Fe}2(µ-N)][Na(THF)5] (3),
bottom, and [PhBP3]FeIIItN(1-Ad) (5), top. Experimental parameters: 2.0
mM analyte, 0.30 M [nBu4N][PF6] electrolyte, scan rate ) 50 mV/s.
The two redox events observed at the more positive potentials
are strikingly similar to those observed for 5. In a previous
communication, we have reported that a reversible FeIII/II redox
process is available for [PhBP3]FeIIItN(p-tolyl),10 and the
reversible wave centered at -1315 mV in the cyclic voltam-
mogram of 5 is consistent with this picture and with the fact
that this species can be chemically reduced to produce anionic
4. The irreversible wave in the cyclic voltammogram of 5 has
a peak maximum at -130 mV. We presume that this step
represents an oxidation to generate a highly unstable “[PhBP3]-
FeIVtN(1-Ad)+” species, or constitutes an oxidation of the
borate ligand. We suggest that the reversible wave centered at
-1.3 V in the cyclic voltammogram of 3 corresponds to an
Fe(III)Fe(II)+/Fe(II)Fe(II) redox process. Several assignments
are possible for the irreversible wave at -350 mV. It may
correspond to an Fe(III)/Fe(IV) redox event at a single iron
center consistent with the picture put forth for 5, to an
Fe(III)Fe(II)+/Fe(III)Fe(III)2+ redox process, or to oxidative
degradation of a borate ligand. The reduction wave at low
potential (-2.5 V) is interesting and is not observed for complex
5. This wave is suggestive of an Fe(II)Fe(II)/Fe(II)Fe(I)- redox
process in which the resulting reduced iron species appears to
be relatively unstable under the conditions of the electrochemical
experiment, as the anodic wave decreases and eventually
vanishes when this redox event is isolated and cycled multiple
times. Synthetic examination of this system using chemical
reductants and oxidants may offer a better understanding of these
redox events.
An intriguing possibility was that addition of CO would lead
to disproportionation of the system with liberation of [PhBP3]-
Fe(CO)210 and the elusive nitride species “{[PhBP3]FeIII(N)}-”
to which we have alluded to above. Exposure of 3 to an
atmosphere of CO results in its instantaneous consumption
(Scheme 3), but the major product of the reaction proved to be
the FeI dicarbonyl, [PhBP3]Fe(CO)2. A second iron-containing
product, present in ca. 25% yield, was identified as the
diamagnetic iron(II) isocyanate complex [PhBP3]Fe(CO)2(NCO)
(6), a presumed product of nitride capture. The identity of
diamagnetic 6 was readily ascertained by comparison of the
crude spectral data (IR, 1H, and 31P{1H}) from the reaction with
that of an independently prepared sample of 6 generated by the
reaction between [PhBP3]Fe(CO)2Cl (7) and [K][NCO]. We
have not identified the fate of the sodium cation in this reaction,
but it is worth noting that, based upon the electronic structure
considerations mentioned above for “{[PhBP3]FeIII(N)}-”, we
would expect it to be a powerful reductant if generated as an
intermediate. Its oxidation product, [PhBP3]FeIVtN, should be
susceptible to attack by CO to produce isolable 4.35 While it is
unclear which species serves as the formal electron acceptor in
the production of 6, [PhBP3]Fe(CO)2 and 6 are the only iron-
containing products observed in the reaction.
The mechanism concerning the conversion of 2 to 3 deserves
some consideration. Due to the heterogeneous reaction and the
paramagnetic nature of the precursor complex and presumable
intermediates, it is a difficult system to probe experimentally.
As shown in Scheme 2, we presume that Na/Hg amalgam
reduction of 2 to its corresponding anion releases N2 gas,
consistent with the observation of rapid effervescence. A
plausible iron byproduct from N2 release is the anionic nitride
species “{[PhBP3]FeIII(N)}-”. Based upon DFT calculations we
have used to explore the electronic nature of a structurally related
species, [PhBPiPr3]FeIVtN,8 a single unpaired electron would
populate a high-lying a1 orbital featuring significant Fe-N σ*
character in {[PhBP3]FeIII(N)}-. This electronic situation should
render the putative species highly reactive. It is therefore
reasonable to suspect that if generated intermittently, such a
species might attack the iron center of a secondary molecule of
2 to displace azide, or might attack the azide ligand itself.34
Both of these scenarios have been suggested by Summerville
While more comprehensive reactivity studies are in the offing,
we wish to note that the (µ-N) ligand of 3 can serve as a source
of NH3 upon exposure to acid (eq 1). Thus, the addition of 3
equiv of HCl to a THF solution of 3 at -35 °C released NH3
in good yield (two independent experiments provided NH3 in
yields of 80% and 95%),8,36 along with the chloride complex 1
as the predominant iron-containing byproduct (>80%). The yield
of NH3 was determined by vacuum transfer of the reaction
volatiles into an ethereal solution containing HCl, which results
(34) For an example of a bridged N4 ligand, see: Mass, V. W.; Kujanek, R.;
Baum, G.; Dehnicke, K. Angew. Chem., Int. Ed. Engl. 1984, 23, 149. For
an example of a terminal N4 ligand, see: Huynh, M. H. V.; Baker, R. T.;
Jameson, D. L.; Labouriau, A.; Meyer, T. J. J. Am. Chem. Soc. 2002, 124,
4580.
(35) Consistent with this suggestion is that the well-defined Fe(IV) nitride species
[PhBPiPr3]FetN (ref 8) is reduced by CO in benzene to produce a mixture
of two diamagnetic iron(II) isocyanates: [PhBPiPr3]Fe(NCO)(CO) and
[PhBPiPr3]Fe(NCO)(CO)2. This transformation and further reactivity studies
pertaining to [PhBPiPr3]FetN will be elaborated in due course.
(36) Yandulov, D. V.; Schrock, R. R. J. Am. Chem. Soc. 2002, 124, 6252.
9
J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005 1919

A R T I C L E S
Scheme 2
Brown and Peters
Scheme 3
the FeMo-cofactor of nitrogenase contains an interstitial light
atom, presumably a nitride, surrounded by six pseudotetrahedral
iron centers.2 While this revised cofactor structure does little
to clarify the mechanism of biological N2 reduction, it at least
calls attention to a possible role for relatively low-valent, low-
coordinate Fe-(µ-N)-Fe linkages within the active site of the
enzyme.3 Likewise, similar structural motifs may be relevant
to the industrial synthesis of ammonia mediated by catalytically
active, low-valent iron sites.4 While these suggestions remain
highly speculative due to the absence of structural data during
catalytic turnover conditions, the bimetallic nitride we have
described at least provides chemical support for well-defined,
low-valent µ-iron nitride systems. Specifically, our study shows
that low-valent iron(II) centers are able to support bridging µ2-
nitride ligands, in this case with pseudotetrahedral systems
supported by relatively soft phosphine donors. The µ2-nitride
functionality can serve a structural role under reducing chemical
conditions. The cyclic voltammetry of 3, shown in Figure 7,
adequately illustrates this point by revealing a pseudo-reversible
FeIIFeII/FeIIFeI reduction process at low potential. It is of further
note that the µ2-nitride ligand can serve as a source of NH3
upon addition of protons.
in the precipitation of ammonium chloride. Subsequent isolation
and quantification of the [NH4][Cl] salt was accomplished by
NMR spectroscopy in DMSO-d6 using a ferrocene integration
standard.8 It is of future interest to add fewer proton equivalents,
and perhaps H-atom equivalents (H+/e-) to 3 to try to elucidate
the other protonated forms of the system prior to NH3 release
(e.g., FeII(µ-NH)FeII and FeII-NH2 + FeIICl).
Conceptually replacing one of the “[PhBP3]Fe(II)” units in
the anion of 3 with an organic moiety results in a mononuclear,
anionic iron(II) imide. We have described an example of such
a species herein, {[PhBP3]FeIItN(1-Ad)}{nBu4N} (4), that
features a linear Fe-N-C bond angle, a short Fe-N bond
length, and a diamagnetic ground state. These features are
collectively indicative of significant Fe-N multiple bonding
character in the complex. The DFT calculation presented for
this system corroborates the d-orbital splitting scheme that we
have proposed previously,10 placing three low-lying and largely
3 + 3HCl f 2[PhBP3]FeCl (1) + NH3 + NaCl (1)
Concluding Remarks
The bimetallic bridged nitride complex 3 is unique in several
regards. Foremost among these are the low coordination number
and the low oxidation state of each of its two iron centers, and
its severely bent Fe-N-Fe linkage. These features are un-
precedented for bimetallic bridged iron nitrides and are very
uncommon to bimetallic bridged nitrides more generally.
Spectroscopic data for 3 suggest the presence of two high-spin
ferrous centers that are sufficiently coupled to provide a singlet
ground-state even at room temperature.
2
2
2
nonbonding orbitals (dx -y , dxy, dz ) at similar energy and well
separated from two high-lying orbitals (dxz, dyz) with significant
σ and π antibonding contributions (see Figure 8). This orbital
arrangement favors a d6 electronic configuration, and addition-
ally accommodates a d5 configuration, as demonstrated by the
relative stability of doublet imides such as complex 5. It is
therefore evident that the imide ligand confers a low-spin
electronic configuration to both pseudotetrahedral iron(II) and
iron(III) centers, at least for [BP3]Fe systems. As we have
recently shown, transforming the imide ligand to a terminal
While bridged iron nitride species are relatively rare, they
may be important to nitrogen fixation schemes. As mentioned
at the outset, it is interesting to recall the recent finding that
9
1920 J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005

Pseudotetrahedral Iron(II) Coordination Complexes
A R T I C L E S
Figure 8. (left) Qualitative d-orbital splitting diagrams illustrating the predicted ground-state electronic structures of various [BP3]FetNx species for d4, d5,
6
2
and d configurations. Note that for the imide species an orbital of a1 symmetry with dz character lies in the nonbonding region, very close in energy to
orbitals with dx -y and dxy character. Therefore, d6 and d5 configurations are relatively stable. In contrast, the a1 orbital is strongly destabilized for a terminal
nitride species, [BP3]FetN. As a result, the d4 configuration is relatively stable, whereas the d5 configuration is destabilized due to population of a high-
lying orbital. (right) Comparison of the d-orbital splitting diagrams anticipated under idealized three-fold and four-fold symmetry. Whereas an FetE triple
bond best describes the configurations shown under three-fold symmetry, an FedE double bond best describes the configurations shown under four-fold
symmetry.
2
2
Mercury-300 NMR spectrometer was used to record 1H NMR and 31
P
nitride ligand significantly alters the d-orbital splitting scheme.8
In the latter case, the a1 orbital of dz character is strongly
2
NMR spectra at ambient temperature. 1H chemical shifts were
referenced to residual solvent, while 31P chemical shifts were referenced
to 85% H3PO4 at δ 0 ppm. 11B NMR data were acquired on a JEOL
400 MHz spectrometer, and chemical shifts were referenced to neat
BF3‚OEt2 at δ 0 ppm. 15N NMR data were acquired on an Inova 500
MHz spectrometer, and chemical shifts were referenced to CH3NO2
(380 ppm relative to liquid ammonia at 0 ppm). MS data for samples
were obtained by injecting a benzene solution into a Hewlett-Packard
1100MSD mass spectrometer equipped with an electrospray (ES)
ionization chamber. UV/vis/NIR measurements were taken in THF-d8
on a Cary 500 UV/vis/NIR spectrophotometer using a 0.1 cm quartz
cell with a Teflon stopper. IR measurements were obtained with a KBr
solution cell using a Bio-Rad Excalibur FTS 3000 spectrometer
controlled by Bio-Rad Merlin Software (v. 2.97) set at 4 cm-1
resolution. X-ray diffraction studies were carried out in the Beckman
Institute Crystallographic Facility on a Bruker Smart 1000 CCD
diffractometer.
destabilized due to favorable directional overlap with a 2pz
orbital at the nitride N-atom. This arrangement favors the d4
configuration, in accord with our successful characterization of
[PhBPiPr3]FeIVtN. Addition of one electron to this latter species
produces the hypothetical d5 anion “[PhBPiPr3]Fe(N)-”, which
would be characterized by a double (or partial triple) bond due
to the population of a strongly antibonding orbital by one
electron (see Figure 8). Needless to say, such a species should
be a potent reductant in its own right, and it is therefore not
surprising that the reduction of azide 2 ultimately leads to an
anionic bridged nitride rather than an anionic terminal nitride
species. We emphasize that for the d6 and d5 imides, and for
the d4 nitride, the Fe-N triple bond formulation is appropriate
under three-fold symmetry. This situation sharply contrasts four-
fold symmetric FedE species (E ) O, N),37,38 where the
population of π* antibonding orbitals prohibits the formation
of a bona fide triple bond (Figure 8).
Magnetic Measurements. Measurements were recorded using a
Quantum Designs SQUID magnetometer running MPMSR2 software
(Magnetic Property Measurement System Revision 2). Data were
recorded at 5000 G. Samples were suspended in the magnetometer in
a clear plastic straw sealed under nitrogen with Lilly No. 4 gel caps.
Loaded samples were centered within the magnetometer using the DC
centering scan at 35 K and 5000 G. Data were acquired at 4-10 K
(one data point every 2 K), 10-60 K (one data point every 5 K), and
60-290 K (one data point every 10 K). The magnetic susceptibility
was adjusted for diamagnetic contributions using the constitutive
corrections of Pascal’s constants. The molar magnetic susceptibility
(øm) was calculated by converting the calculated magnetic susceptibility
(ø) obtained from the magnetometer to a molar susceptibility (using
the multiplication factor {(molecular weight)/[(sample weight)*(field
strength)]}). Effective magnetic moments were calculated using eq 2.
Solution magnetic moments were measured using the Evans method.18
Experimental Section
All manipulations were carried out using standard Schlenk or
glovebox techniques under a dinitrogen atmosphere. Unless otherwise
noted, solvents were deoxygenated and dried by thorough sparging with
N2 gas followed by passage through an activated alumina column.
Nonhalogenated solvents other than acetonitrile were tested with a
standard purple solution of sodium benzophenone ketyl in tetrahydro-
furan to confirm effective oxygen and moisture removal. All reagents
were purchased from commercial vendors and used without further
purification unless otherwise stated. [Na][N3], [Na][15NNN] (Cambridge
Isotope Laboratories), and [K][NCO] were dried under reduced pressure
at 120 °C overnight prior to use. The compounds [PhBP3]FeCl (1) and
[PhBP3]Fe(PPh3) were prepared according to literature procedures.10
Elemental analyses were performed by Desert Analytics, Tucson, AZ.
Deuterated THF and benzene were purchased from Cambridge Isotope
Laboratories, Inc. and were degassed and dried over fine alumina (THF)
or activated 3 Å molecular sieves (benzene) prior to use. A Varian
µeff ) sqrt(7.997ømT)
(2)
Electrochemistry. Electrochemical measurements were carried out
in a glovebox under a dinitrogen atmosphere in a one-compartment
cell using a BAS model 100/W electrochemical analyzer. A glassy
carbon electrode and platinum wire were used as the working and
auxiliary electrodes, respectively. The reference electrode was Ag/
AgNO3 in THF. Solutions (THF) of electrolyte (0.3 M tetra-n-
butylammonium hexafluorophosphate) and analyte (2 mM) were also
prepared in a glovebox.
(37) Rohde, J. U.; In, J.-H.; Lim, M. H.; Brennessel, W. W.; Bukowski, M. R.;
Stubna, A.; Mu¨nck, E.; Nam, W.; Que, L., Jr. Science 2003, 299, 1037.
See also ref 11.
(38) Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr. Chem. ReV. 2004,
104, 939.
9
J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005 1921

A R T I C L E S
Brown and Peters
Table 1. X-ray Diffraction Experimental Details for {[PhBP3]FeN3}2 (2), [{[PhBP3]Fe}2N][Na(THF)5] (3), {[PhBP3]FetN(1-Ad)}{nBu4N} (4),
and [PhBP3]FetN(1-Ad) (5)
monomer 2
‚
C6H6
3‚
THF
4
‚THF
5
chemical formula
fw
T (°C)
λ (Å)
a (Å)
C51H47BFeN3P3
861.49
-177
0.71073
13.0738(10)
14.2209(11)
14.6878(12)
118.2270(10)
95.2290(10)
111.0900(10)
2128.4(3)
P-1
C
114H130B2Fe2NNaO6P6
C75H100BFeN2OP3
1205.14
-177
C55H56BFeNP3
890.58
-177
0.71073
16.4143(15)
14.1225(13)
20.7686(19)
90
109.186(2)
90
4547.0(7)
P2(1)/n
4
1952.32
-177
0.71073
16.7910(8)
13.7352(6)
43.299(2)
90
93.6530(10)
90
9965.8(8)
P2(1)/c
4
0.71073
12.173(3)
14.051(4)
22.131(6)
71.610(4)
79.790(4)
68.373(4)
3331.1(14)
P-1
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
space group
Z
2
2
D
calc (g/cm3)
µ (cm-1
R1, wR2a (I > 2σ(I))
1.344
5.08
0.0396, 0.0858
1.301
4.49
0.0650, 0.0940
1.202
3.44
0.0470, 0.0972
1.301
4.76
0.0489, 0.0982
)
a R1 ) ∑||Fo| - |Fc||/∑|Fo|, wR2 ) {∑[w(Fo - Fc )2]/∑[w(Fo )2]}1/2
.
2
2
2
X-ray Crystallography. X-ray quality crystals were grown as
indicated in the experimental procedures per individual complex. The
crystals were mounted on a glass fiber with Paratone N oil, and data
were collected on a Bruker SMART 1000 diffractometer with a CCD
area detector under a stream of dinitrogen. Structures were determined
using direct methods with standard Fourier techniques using the Bruker
AXS software package. In some cases, Patterson maps were used in
place of the direct methods procedure. Table 1 includes the X-ray
diffraction experimental details, while the full crystallographic tables
are included in the Supporting Information.
DFT. A hybrid density functional calculation was performed for
{[PhBP3]FetN(tBu)}- using the Jaguar package (version 5.0, release
20).39 The calculation employed B3LYP with LACVP** (LACVP**++
for B) as the basis set.40 A geometry optimization was carried out using
the X-ray coordinates for 5 (with all of the carbon atoms of the
adamantyl group removed except for those directly attached to the imide
nitrogen atom) as the initial HF guess. No symmetry constraints were
imposed, and the calculation was performed assuming a singlet
electronic ground state.
pressure. The crude solids were extracted with a fresh aliquot of THF
(40 mL) and filtered over Celite. Volatiles were again removed under
reduced pressure, and the resulting solids were collected on a sintered
glass frit and washed with copious amounts of benzene (∼5 × 30 mL)
followed by petroleum ether (2 × 15 mL), and dried to obtain 0.546 g
(52%) of a brown solid. X-ray quality crystals were grown from a THF/
hexanes vapor diffusion chamber, while samples for elemental and 1H
NMR analysis were repeatedly crystallized from THF/petroleum ether.
1H NMR (THF-d8, 300 MHz): δ 7.72 (d, J ) 6.0 Hz, 4H); 7.44 (br s,
24H); 7.12 (t, J ) 7.2 Hz, 4H); 6.88 (t, J ) 7.2 Hz, 2H); 6.70 (t, J )
7.5 Hz, 12H); 6.53 (t, J ) 7.5 Hz, 24H); 3.60 (m, 20H, overlaps with
residual solvent peaks); 1.78 (m, 20H, overlaps with residual solvent
peaks); 1.56 (br s, 12H). 31P{1H} NMR (THF-d8, 121.4 MHz): δ 58.0
(s). UV/vis/NIR (THF-d8) λ, nm (ꢀ, M-1 cm-1): 1266 (943), 1120
(2740), 565 (11 520), 445 (20 530), 390 (21 300). Anal. Calcd for
C110H122B2Fe2NNaO5P6: C, 70.26; H, 6.54; N, 0.74. Found: C, 70.06;
H, 6.35; N, 0.82. A sample of 50% 15N-enriched 3 was synthesized
using an analogous synthetic procedure with 2-15N. 15N NMR (THF,
50.751 MHz): δ 801 (s).
Synthesis of {[PhBP3]FetN(1-Ad)}{nBu4N}, 4. Compound 5
(0.168 g, 0.189 mmol) was dissolved in THF (5 mL) with stirring. A
0.20 wt % Na/Hg amalgam (0.0043 g of Na dissolved in ca. 2.2 g of
Hg) was added to the THF solution of 5, and the reaction was allowed
to stir at room temperature for 2 h. The reaction solution was then
transferred from the amalgam and stirred in a vial containing
[nBu4N][Br] (0.0606 g, 0.189 mmol) for 4 h. After this time, the volatiles
were removed under reduced pressure and the crude solids were
extracted with a fresh aliquot of THF (20 mL) and filtered over a pad
of Celite. Volatiles were again removed in vacuo, and the resulting
brown solids were washed with benzene (3 × 10 mL) and dried to
obtain 0.141 g (66%) of 4. Crystals suitable for an X-ray diffraction
study were grown via vapor diffusion of petroleum ether into a THF
Synthesis of {[PhBP3]FeN3}2, 2. [PhBP3]FeCl (1.00 g, 1.28 mmol)
was dissolved in acetonitrile (10 mL) with stirring and was added to a
stirring acetonitrile (5 mL) slurry of [Na][N3] (0.0837 g, 1.28 mmol).
The initial red color of the reaction gradually fades as a purple solid
precipitates out of solution. After 24 h, volatiles were removed under
reduced pressure and the solids were extracted with benzene (50 mL),
filtered over Celite, and lyophilized. The resulting solid was isolated
on a sintered glass frit, washed with petroleum ether (3 × 30 mL), and
dried to yield 0.861 g (85%) of a reddish-brown powder. X-ray quality
crystals were grown via vapor diffusion of petroleum ether into a
benzene solution. IR (KBr/C6H6): ν(N3) ) 2077 cm-1. UV-vis (C6H6)
λ
max, nm (ꢀ, M-1 cm-1): 410 (700). SQUID: µeff ) 4.3 µB at 296 K.
Evans method (C6D6): 4.50 µB. Anal. Calcd for C45H41BFeN3P3: C,
68.99; H, 5.28; N, 5.36. Found: C, 68.66; H, 5.33; N, 5.34. For 2-15N:
1
solution. H NMR (THF-d8, 300 MHz): δ 7.79 (br s, 12H); 7.57 (m,
an analogous synthetic procedure employing [Na][15NNN] yielded 15
N
2H); 7.04 (t, J ) 7.2 Hz, 2H); 6.67 (m, 18H); 3.07 (m, 8H); 2.15 (s,
6H); 2.07 (s, 3H); 1.78 (m, 6H); 1.57 (m, 8H); 1.35 (m, 8H); 0.98 (t,
J ) 7.2 Hz, 12H, overlaps with ligand methylene resonance at δ 0.94);
0.94 (s, 6H, overlaps with nBu4N+ resonance at δ 0.98). 31P{1H} NMR
(THF-d8, 121.4 MHz): δ 95.0 (s). 11B{1H} NMR (THF-d8, 128.3
MHz): δ -13.5 (s). UV/vis/NIR (THF-d8) λ, nm (ꢀ, M-1 cm-1): 600
(815), 510 (2112). Anal. Calcd for C71H92BFeN2P3: C, 75.26; H, 8.18;
N, 2.47. Found: C, 75.15; H, 8.31; N, 2.78.
labeled 2. IR (KBr/C6H6): ν(N3) ) 2066 cm-1
.
Synthesis of [{[PhBP3]Fe}2N][Na(THF)5], 3. Compound 2 (0.874
g, 1.12 mmol) was dissolved in THF (15 mL) with stirring. To this
solution was added 1 equiv of sodium in the form of a 0.20 wt %
Na/Hg amalgam (0.0256 g of Na dissolved in 13.5 g of Hg). The
addition of the amalgam resulted in the vigorous effervescence of N2
as the color of the reaction solution darkened. After ∼5 min, the
effervescence had subsided and the reaction was capped and allowed
to stir at room temperature for 8 h. After this time, the reaction was
decanted from the amalgam and volatiles were removed under reduced
Synthesis of [PhBP3]FetN(1-Ad), 5. [PhBP3]Fe(PPh3) (0.100 g,
0.0996 mmol) was added to benzene (5 mL) with stirring. A benzene
solution (1 mL) of 1-azidoadamantane (0.0353 g, 0.199 mmol) was
added dropwise, during which time the reaction changed color from
orange to brown. After 12 h, volatiles were removed under reduced
(39) Jaguar 5.0, Schrodinger, LLC, Portland, Oregon, 2002.
(40) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785-789.
9
1922 J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005

Pseudotetrahedral Iron(II) Coordination Complexes
A R T I C L E S
pressure. The resulting crude solids were washed with petroleum ether
(3 × 20 mL) and dried under reduced pressure to yield 5 as a brown
solid (0.071 g, 80%). X-ray quality crystals were grown via vapor
frozen, and the N2 atmosphere was evacuated and replaced with 1 atm
of CO. The NMR tube was shaken periodically over a period of 10
min at which time a crude 31P NMR spectrum confirmed the presence
of 6. After this time, volatiles were removed under reduced pressure
1
diffusion of petroleum ether into a benzene solution. H NMR (C6D6,
1
300 MHz): δ 23.1 (br, s); 16.8 (br, s); 14.9 (s); 10.6 (t, J ) 6.3 Hz);
9.61 (t, J ) 7.2 Hz); 7.99 (d, J ) 12.3 Hz); 6.50 (d, J ) 11.1 Hz);
5.01 (d, J ) 7.2 Hz); 3.48 (s); 2.48 (t, J ) 6.3 Hz); -3.34 (br, s).
UV-vis (C6H6) λ, nm (ꢀ, M-1 cm-1): 422 (1600); 510 (1050). Evans
method (C6D6): 1.98 µB. Anal. Calcd for C55H56BFeNP3: C, 74.17;
H, 6.34; N, 1.57. Found: C, 74.13; H, 6.26; N, 1.72.
and both H NMR (C6D6) and IR analysis (KBr/C6H6) confirmed the
presence of [PhBP3]Fe(CO)210 as the major product and 6 as the minor
product. Addition of a hexamethylbenzene internal standard to this 1H
NMR sample demonstrated that 6 was produced in ca. 25% yield.
The Reaction of 3 with HCl. Nitride 3 (0.030 g, 0.0160 mmol)
was dissolved in 1 mL of THF in a sealable NMR tube. The solution
was cooled to -35 °C, and 3 equiv of HCl (46 µL of a 1 M solution
in Et2O diluted in 0.25 mL of THF) cooled to -35 °C was added. The
reaction was then allowed to warm to room temperature. After 30 min,
the reaction volatiles were vacuum transferred into 5 mL of a 1 M
solution of HCl in Et2O. Once the transfer was complete, the HCl
solution was stirred at room temperature for about 30 min. At this point,
a small amount of precipitate was evident and the volatiles were
removed under reduced pressure. The resulting residue was dissolved
in DMSO-d6, and the presence of [NH4][Cl] was confirmed by a triplet
at δ 7.15 ppm (t, J ) 51 Hz, 4H) which was integrated versus an internal
ferrocene standard.8 Two independent runs provided an 80% and a 95%
yield of NH3, respectively. The remaining solids from the original
reaction were dissolved in benzene-d6, and chloride 1 was observed to
be the major (>80%) iron-containing product by 1H NMR spectroscopy.
Synthesis of [PhBP3]Fe(CO)2(NCO), 6. Compound 7 (0.050 g,
0.060 mmol) and [K][NCO] (0.049 g, 0.60 mmol) were stirred in THF
(5 mL) for 24 h. After this time, volatiles were removed under reduced
pressure and the crude solids were extracted with benzene (5 mL),
filtered over Celite, and lyophilized to yield 0.038 g (75%) of a yellow
1
solid. H NMR (C6D6, 300 MHz): δ 7.84 (d, J ) 7.2 Hz, 2H); 7.69
(br s, 8H); 7.55 (t, J ) 7.2 Hz, 2H); 7.36 (d, J ) 7.2 Hz, 1H); 7.10 (m,
4H); 6.86 (br s, 14H); 6.72 (m, 4H); 1.83 (br d, J ) 60 Hz, 4H); 1.56
(d, J ) 14 Hz, 2H). 31P{1H} NMR (C6D6, 121.4 MHz): δ 50.4 (t, J )
58 Hz, 1P); 27.3 (d, J ) 58 Hz, 2P). IR (KBr/C6H6): ν(CO) ) 2043,
1999 cm-1; ν(NCO) ) 2195, 2224 cm-1. Anal. Calcd for C48H41BFe-
NO3P3: C, 68.68; H, 4.92; N, 1.67. Found: C, 69.01; H, 5.23; N, 2.01.
Synthesis of [PhBP3]Fe(CO)2Cl, 7. Chloride 1 (0.200 g, 0.257
mmol) was stirred in benzene (10 mL) in a 50 mL Schlenk flask. The
solution was frozen, and the atmosphere was evacuated and replaced
with 1 atm of CO. The reaction was then allowed to warm to room
temperature and stirred for an additional 30 min. After this time, the
volatiles were removed under reduced pressure yielding an essentially
Acknowledgment. Financial support from the NSF (CHE-
0132216), the DOE (PECASE), and the Alfred P. Sloan
Foundation is gratefully acknowledged. We thank the Beckman
Institute (Caltech) for use of the SQUID magnetometer, and
Larry M. Henling and Theodore A. Betley for assistance with
crystallographic studies. The Zewail group provided access to
a Cary 500 UV/vis/NIR spectrophotometer.
1
quantitative amount of a golden solid. H NMR (C6D6, 300 MHz): δ
7.87 (m, 10H); 7.56 (t, J ) 7.2 Hz, 2H); 7.35 (t, J ) 6.0 Hz, 1H); 7.20
(m, J ) 7.2 Hz, 4H); 6.84 (m, 14H); 6.74 (m, 4H); 1.91 (br d, J ) 57
Hz, 4H); 1.62 (d, J ) 15 Hz, 2H). 31P{1H} NMR (C6D6, 121.4 MHz):
δ 54.7 (t, J ) 53 Hz, 1P); 25.2 (d, J ) 53 Hz, 2P). 11B{1H} NMR
(C6D6, 128.3 MHz): δ -13.9 (s). IR (KBr/C6H6): ν(CO) ) 2039, 1997
cm-1. ES-MS: calcd for C47H41BClFeO2P3 (M)+, 833 m/z; found (M
- Cl)+, 797 m/z. Anal. Calcd for C47H41BClFeO2P3: C, 67.78; H, 4.96.
Found: C, 67.45; H, 5.13.
Supporting Information Available: Complete crystallo-
graphic details and supporting data (PDF, CIF). This material
is available free of charge via the Internet at http://pubs.acs.org.
The Reaction of 3 with CO. Nitride 3 (0.0083 g, 0.004 mmol) was
dissolved in THF in a sealable NMR tube. The resulting solution was
JA0453073
9
J. AM. CHEM. SOC. VOL. 127, NO. 6, 2005 1923