Liliana Capulín Flores1,2, Lucas A Paul3, Inke Siewert3, Remco Havenith1, Noé Zúñiga-Villarreal2, Edwin Otten1. 1. Stratingh Institute for Chemistry, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands. 2. Instituto de Química, Universidad Nacional Autónoma de México, Ciudad Universitaria, Circuito Exterior, 04510 México, D.F., México. 3. Universität Göttingen, Institut für Anorganische Chemie, Tammannstraße 4, D-37077 Göttingen, Germany.
Abstract
Metal complexes with ligands that coordinate via the nitrogen atom of azo (N═N) or imino (C═N) groups are of interest due to their π-acceptor properties and redox-active nature, which leads to interesting (opto)electronic properties and reactivity. Here, we describe the synthesis and characterization of rhenium(I) tricarbonyl complexes with neutral N,N-bidentate formazans, which possess both N═N and C═N fragments within the ligand backbone (Ar1-NH-N═C(R3)-N═N-Ar5). The compounds were synthesized by reacting equimolar amounts of [ReBr(CO)5] and the corresponding neutral formazan. X-ray crystallographic and spectroscopic (IR, NMR) characterization confirmed the generation of formazan-type species with the structure fac-[ReBr(CO)3(κ2-N2,N4(Ar1-N1H-N2═C(R3)-N3═N4-Ar5))]. The formazan ligand coordinates the metal center in the 'open' form, generating a five-membered chelate ring with a pendant NH arm. The electronic absorption and emission properties of these complexes are governed by the presence of low-lying π*-orbitals on the ligand as shown by DFT calculations. The high orbital mixing between the metal and ligand results in photophysical properties that contrast to those observed in fac-[ReBr(CO)3(L,L)] species with α-diimine ligands.
Metal complexes with ligands that coordinate via the nitrogen atom of azo (N═N) or imino (C═N) groups are of interest due to their π-acceptor properties and redox-active nature, which leads to interesting (opto)electronic properties and reactivity. Here, we describe the synthesis and characterization of rhenium(I) tricarbonyl complexes with neutral N,N-bidentate formazans, which possess both N═N and C═N fragments within the ligand backbone (Ar1-NH-N═C(R3)-N═N-Ar5). The compounds were synthesized by reacting equimolar amounts of [ReBr(CO)5] and the corresponding neutral formazan. X-ray crystallographic and spectroscopic (IR, NMR) characterization confirmed the generation of formazan-type species with the structure fac-[ReBr(CO)3(κ2-N2,N4(Ar1-N1H-N2═C(R3)-N3═N4-Ar5))]. The formazan ligand coordinates the metal center in the 'open' form, generating a five-membered chelate ring with a pendant NH arm. The electronic absorption and emission properties of these complexes are governed by the presence of low-lying π*-orbitals on the ligand as shown by DFT calculations. The high orbital mixing between the metal and ligand results in photophysical properties that contrast to those observed in fac-[ReBr(CO)3(L,L)] species with α-diimine ligands.
Formazans are a large family of compounds
containing the R1-NH-N=C(R3)-N=N-R5 backbone,
known for their use as analytical reagents for metal detection[1] and as cellular[2] and
textile dyes.[3] These applications are the
consequence of its well-defined redox chemistry[4] and its ability to chelate metal centers in its deprotonated
form, i.e., the delocalized formazanate anion (R1-N=N-C(R3)=N-N-R5)−. Although formazanate
coordination chemistry was first described in 1941,[5] it was not until the last decade that its study has reemerged
due to its electrochemical and optical properties.[6] A wide variety of formazanate complexes with both main
group and transition metal elements have been reported, wherein the
(anionic) ligand usually coordinates through the terminal donor sites
to form 6-membered chelates.[7] In addition
to taking advantage of the unique optoelectronic properties imparted
by formazanate ligands, recent reports show that their redox-active
nature can also be used to obtain new catalytic reactivity.[8] In contrast to complexes with anionic formazanates,
reports on coordination of the neutral formazan fragment remain scarce
to date. In 2015, our group described the first example of a formazan-type
complex,[9] in which the neutral ligand binds
Zn(C6F5)2 through one terminal and
one internal nitrogen atoms yielding a five-membered chelate (Scheme ), also described
as the ‘open’ coordination mode. It
was hypothesized that the poor basicity of the Zn-C6F5 group in the precursor allowed the isolation of the Zn-formazan
compound, as the more basic reagent ZnMe2 does result in
rapid deprotonation of the formazan NH group.[9]
Scheme 1
Synthesis of Zn Formazan Species
Metal complexes with ligands containing the
NH functionality have
gained importance in catalysis since the NH arm can serve as an anchor
for substrate recognition, thus enhancing catalyst selective and activity.[10] A proton source located at the proximity of
the metal center has been widely investigated in the proton-coupled
electron transfer reduction of small molecules relevant in energy
conversion reactions such as hydrogen evolution[11,12] and CO2 reduction.[13−15] It was proposed to modulate the
redox properties, aid in the stabilization of intermediates, or impact
the kinetics due to the increased local proton concentration. Furthermore,
deprotonation of the NH group is known to modify the electronic and
geometric structure of such complexes.[16−18]Extensive research
has focused on the properties and potential
applications of fac-[L,LReX(CO)3] (L =
α-diimine) compounds in medicinal inorganic chemistry,[19,20] material science,[21,22] and catalysis.[23,24] Particularly, these compounds have shown to be good candidates for
electrochemical CO2 reduction, in which a proton source
is required either for enhancing or triggering the catalytic process.[25,26] Mainly, Re-based bipyridine systems have been reported as catalysts
for CO2 electroreduction, wherein the presence of XH (X
= O, N, C)[18,27,28] functionalizations boosts the catalytic effect or induces other
reactivity patterns (Chart ).[29−31] Key to the catalytic conversion of CO2 to CO by the well-studied bipyridine Re and Mn complexes is the
involvement of the supporting (‘redox-active’) ligand
in the reduction chemistry.[28,32−34] We hypothesize that replacing the bipyridine ligand (an aromatic
α-diimine) for a redox-active formazan ligand (formally an amino-substituted
α-azoimine) could provide an avenue to influence the potential
at which reduction of the catalyst occurs.[35,36] In addition, such ligands provide access to flexible coordination
modes (hemilability)[37] due to the presence
of four nitrogen atoms in the backbone, as well as proton-responsivity
via the NH moiety that is in close proximity to the metal center,
features that are key to the activity/selectivity of metalloenzymes
but challenging to emulate in synthetic catalysts.[38]
Chart 1
Representative Examples of Re Complexes Bearing an
XH Functionality
Herein, we report synthesis of a series of fac-Re(CO)3(formazan) complexes and investigate
their (electronic)
structures and photophysical properties.
Results and Discussion
Ligand Synthesis
Formazan ligands L1H–L4H were synthesized according to the procedure reported by
Hicks and coworkers, via aldehyde condensation with phenylhydrazine
followed by a coupling reaction with phenyldiazonium chloride in a
biphasic reaction medium (CH2Cl2/water) under
mild basic conditions.[4] Similarly, L5H was prepared using the methodology previously described
by our group, in which the coupling step is carried out in acetone/water
with NaOH as base (Scheme ).[9] In all cases, the compounds
were obtained in moderate yields after purification (27–54%).
Scheme 2
Synthesis of Formazan Ligands
Complex Synthesis
Equimolar amounts of [ReBr(CO)5] and the corresponding formazan, L1H–L4H, reacted in refluxing toluene for 1 h to afford complexes 1–4 in moderate to good yields (28–67%)
(Scheme a). In all
cases, complete conversion of the starting material was confirmed
by 1H NMR and infrared spectroscopy. The compounds are
air-stable solids with dark red color and are soluble in low to medium
polarity solvents. Complexes 2–4 were isolated
as pure materials by either recrystallization or rinsing with pentane.
A minor impurity was invariably present (1H NMR spectroscopy)
in the isolated material of 1. Attempts to further purify
the material by crystallization were unsuccessful. The reaction of
the asymmetric formazan L5H with [ReBr(CO)5] in refluxing toluene gave a mixture of two complexes (5a and 5b) based on 1H NMR spectroscopy (Scheme b), which differ
in the substituent at the NH position (Mes or Ph). Unsurprisingly,
complexes 5a/b present similar physical properties—dark
red solids soluble in low polarity solvents—that we were unable
to separate, and solution characterization data are reported below
for the mixture.
Scheme 3
General Synthesis of Compounds (a) 1–4 and (b) the Mixture of Isomers 5a and 5b
FT-IR Spectroscopy
The infrared spectra of complexes 1–4 and the mixture 5a/b feature the characteristic pattern for fac-tricarbonyl species: three intense bands in the ν(CO) carbonyl
region arising from the IR-active 3A normal vibration modes for complexes
with a C1 symmetry (Figure ). The CO stretching frequencies for compound 1 are observed at 2035, 1959, and 1923 cm–1 in CH2Cl2 solution. The effect of the para-substituent on the aromatic ring (R3) is
minimal, and compounds 2–5 show virtually
identical IR spectra.
Figure 1
CO stretching bands (ν(CO)) in the FT-IR spectra
of complexes 1–4 and the mixture 5a and 5b in CH2Cl2 solution
at rt.
CO stretching bands (ν(CO)) in the FT-IR spectra
of complexes 1–4 and the mixture 5a and 5b in CH2Cl2 solution
at rt.
NMR Studies
1H NMR spectra of 1–4 show a singlet ca. 8.5 ppm assigned to the hydrazo proton, consistent
with the presence of a neutral formazan ligand (see Figure S1a–S4a). In agreement with the presence of
an asymmetric, protonated formazan ligand, the 13C NMR
spectrum shows three distinct carbonyl resonances between 185 and
192 ppm indicative of C1 symmetry for
these Re complexes. In 13C NMR spectroscopy, the N-Ph ipso-carbon atoms attached to the formazan backbone display
distinctive chemical shifts. Unequivocal assignment of these was possible
based on an HSQC experiment where coupling between the hydrazo NH
group and one of the Ph ipso-C resonances was observed.
The resonance of the ipso-C bound to the azo group
appears at 157 ppm, which is downfield from both the ipso-C linked to the NH function located ca. 141 ppm, and the ipso-C of the C-Ar group (120–125 ppm). Similarly,
the N-Ph groups are inequivalent in the 1H NMR spectrum
also at elevated temperature (80 °C in toluene-d8), suggesting that chemical exchange by proton transfer
between the azo (C-N=N-Ph) and hydrazo (C=N-NH-Ph) fragments
does not readily occur. This was further corroborated by the absence
of exchange crosspeaks in the 2D EXSY spectrum (80 °C, mixing
time of 0.5 s, Figure S7).1H NMR analysis of the reaction mixture of complexes 5a/b indicates the generation of two main products in ratio of 0.7:1.0
with both compounds exhibiting the characteristic NH proton signal
of the neutral formazan ligand at 8.15 and 7.60 ppm, respectively
(Figure a). Minor
signals for another species were observed (<5%) but not investigated
further. For the two major products, the observation of eight unique
resonances in the aliphatic region of the 1H and 13C NMR spectra indicates that all CH3 groups are inequivalent:
each of the two products features four signals due to the CH3 substituents at the p-tolyl (1) and mesityl (3)
rings. Thus, at room temperature, the rotation around the N-Mes bond
is slow on the NMR timescale. In the most downfield part of the 13C{1H} NMR spectrum, i.e., between 180 and 200
ppm, there are six resonances that can be attributed to carbonyl ligands,
which corroborates that both 5a and 5b are
tricarbonyl rhenium complexes (Figure S6). Based on the spectroscopic data, we assign 5a and 5b as two different isomers with the composition [(L5H)Re(CO)3Br], which differ in the nature of the ‘pendant’
(non-coordinating) N-Ar group of the formazan (see Scheme b). Heating an NMR tube containing
the mixture of complexes 5a/b to 80 °C inside the
NMR spectrometer did not significantly change their molar ratio. Inspection
of 2D NMR experiments allowed the assignment of 1H and 13C spectra (see Figure S6). Identification
of the m-CH (Mes) and the m- and o-CH (p-tolyl) protons allowed establishing
of the connectivity in both of the isomeric compounds present in solution.
The 1H,13C correlations in the HMBC spectrum
between the NH fragment and the carbon atoms that are two and three
bonds away indicate that in the major isomer (5b), the
NH group is bound to a mesityl group, whereas in 5a it
is connected to a phenyl group. The greater shielding effect of mesityl
compared to the phenyl group causes the NH proton of the former to
appear at higher field (δ 7.61 ppm in 5b and 8.22
ppm in 5a).
Figure 2
(a) 1H NMR spectrum of the mixture 5a and 5b at room temperature in benzene-d6. (b) EXSY experiment at 80 °C in toluene-d8.
(a) 1H NMR spectrum of the mixture 5a and 5b at room temperature in benzene-d6. (b) EXSY experiment at 80 °C in toluene-d8.To investigate the dynamics of isomers 5a/b in solution, we collected a 1H EXSY NMR
spectrum at 80
°C in toluene-d8 (Figure b). Crosspeaks are observed
between the Mes ortho-CH3 groups within
each isomer due to rotation around the N-Mes bond but not between
isomers 5a and 5b. Whereas free formazans
undergo intramolecular proton exchange rapidly (‘tautomerization’),[39,40] the lack of exchange between 5a/b indicates
that the Re–N bonds are non-labile and coordination to the
Re center effectively blocks exchange. This is in agreement with the
data for the symmetrical derivative 4, which also does
not show exchange between the azo and hydrazo fragments (vide
supra). It should be noted however that the EXSY spectrum
does evidence exchange between the NH groups in 5a/b, which we believe to occur by an intermolecular pathway
instead. This is further corroborated by the observation of exchange
crosspeaks between the NH protons in 5a/b and residual H2O when the NMR solvent is not fully anhydrous
(Figure S8).Additional experiments
were conducted to determine whether intermolecular
proton interchange processes also place in the complexes containing
a symmetric formazan ligand. Partial H/D exchange of the NH proton
in complex 4 was achieved by mixing a CDCl3 solution of the aforementioned complex with D2O. 1H NMR analysis of the resulting mixture indicated that it
was composed of 60% 4D (deuterated product) and 40% the
non-deuterated species 4. The remaining D2O was subsequently removed by stirring over MgSO4, the 4/4D mixture was isolated and then reacted with
an equivalent of 2 in C6D6. Monitoring
the composition by 1H NMR spectroscopy showed that the
intensity of the NH resonance of 4 increased (to 72%)
in the course of 30 min with a concomitant decrease of that in 2, confirming that intermolecular proton exchange is taking
place (see Figure S9).
Structural Studies
Crystals suitable for single-crystal
X-ray diffraction were obtained from slow diffusion of pentane into
a CHCl3 solution of compounds 3 and 4, respectively. The mixture of 5a/b did not crystallize using the same method, but we were able to obtain
a microcrystalline sample from hot hexane that contained some small
needles that were suitable for X-ray crystallographic characterization.
This was identified as isomer 5b, in which the sterically
most demanding Mes group is situated at the non-coordinated N atom
of the ligand; the solid-state structure observed for this material
is consistent with the major species in solution by NMR spectroscopy.
Analysis of the molecular structures of 3, 4, and 5b shows that the three compounds are isostructural
(see Figure and Table for pertinent bond
lengths and angles). The geometry around the metal center is pseudo-octahedral
with the carbonyl ligands in a facial arrangement. The formazan fragment
coordinates in a bidentate fashion through atoms N1 and N3, generating
a five-membered chelate. Coordination of the neutral formazan is scarce,
only observed in the complex [L2H]Zn(C6F5)2 previously reported by our group.[9] The formazan bite angles are virtually identical
in the three complexes (3 = 73.04(9)°, 4 = 73.0(2)°, 5b = 72.9(1)°) and somewhat smaller
than the bite angle reported for the [L2H]Zn(C6F5)2 complex (74.23(13)°). The C7–N2
and C7–N3 bond lengths are different from each other, the magnitude
of the C7−N3 bond lies in between the typical values for C–N
single and double bonds (−C(sp2)–N–
= 1.355 Å; −C(sp2)=N– = 1.279
Å), while the C7−N2 bond length indicates a single bond
character. The N1–N2 bond length is longer than a N=N
double bond (−N=N– = 1.240 Å) and smaller
than a N–N single bond (−N–N– = 1.425
Å).[41] The metallacycle is not fully
planar as the Re atom is displaced out of the ligand plane (N1–N2–C7–N3)
by 0.213–0.393 Å. The dihedral angle between the ligand
plane and a phenyl group in the R1 position is similar
in complexes 3 (50.67°) and 4 (51.12°).
Changing the R1 substituent for the bulkier mesityl group
(5b) causes a rotation out of the ligand plane by almost
30° resulting in a dihedral angle of 79.52° that prevents
steric interactions between the Mes substituent and the equatorial
CO ligand. The structure indicates that rotation around the N-Mes
bond cannot occur freely due to these steric interactions, which is
in agreement with the solution NMR data discussed above. The Re1–N1
bond length to the azo moiety is virtually the same in the three complexes
(2.099–2.126 Å) but it is shorter than the Re–N(azo)
bond length reported for the related [ReBr(CO)3(azopyridine)][42] complex (2.156(3) Å). The Re1–N3
bond lengths are in accordance with the typical Re–N(imine)
bond distances (2.173–2.185 Å in compounds 3–5vs 2.173(3) Å in [ReBr(CO)3(6-methoxipyridine-2-yl)-N-(2-methylthiophenyl)methanimine)],
respectively).[43] The unusually short Re–N1(azo)
bond length reflects that π-backdonation from the Re center
to the azo group is more pronounced in the formazan species than in
azopyridine complexes.[44] The π-acceptor
capabilities of the azo ligand are also reflected in the Re–carbonyl
bond lengths. The Re–C20 bond length, trans to the azo group, is longer than the Re–C21 bond length.
This is consistent with the considerable π-acidity of the azo
group,[45,46] which appears to be more significant in
our formazan complexes than in the corresponding azopyridine analogues.[42]
Figure 3
Molecular structures of 3 (left), 4 (middle),
and 5b (right) showing 50% of ellipsoids.
Table 1
Selected Metrical Parameters for 3, 4, and 5b (Bond Lengths in Å,
Angles in °)
3
4
5b
Re1–Br1
2.5977(3)
2.6236(7)
2.5946(6)
Re1–N1
2.126(2)
2.099(6)
2.122(3)
Re1–N3
2.173(3)
2.185(5)
2.174(4)
Re1–C20
1.955(3)
1.955(6)
1.957(5)
Re1–C21
1.921(3)
1.919(5)
1.916(6)
Re1–C22
1.941(2)
1.918(6)
1.964(4)
N1–N2
1.298(3)
1.291(7)
1.293(5)
C7–N2
1.363(4)
1.382(7)
1.364(8)
C7–N3
1.330(4)
1.319(8)
1.326(5)
N3–N4
1.343(4)
1.324(8)
1.343(7)
Br1–Re1–C22
178.02(9)
176.8(2)
178.6(2)
N1–Re1–C20
170.3(1)
168.7(2)
170.2(2)
N3–Re1–C21
166.5(1)
169.7(2)
170.3(2)
N1–Re1–N3
73.04(9)
73.0(2)
72.9(1)
Molecular structures of 3 (left), 4 (middle),
and 5b (right) showing 50% of ellipsoids.
UV–vis Spectroscopy
The electronic spectra of
complexes 1–4 and the mixture of complexes 5a/5b were measured at 25 °C in toluene
(c ≈ 10–5 M), Figure . Their corresponding data
are summarized in Table . Compounds 1–4 show similar features: an intense
band in the range of 490–510 nm with maximum molar absorptivities
from 18,000 to 28,000 M–1 cm–1. DFT calculations and comparison to literature data allows us to
assign this band to electronic excitations that are Re(dπ) → azo(π*)[42] MLCT and formazan
π → π* in nature.[9] A
band of lower intensity in the range of 330–400 nm (ε
=15,000–16,500 M–1 cm–1) is observed in all compounds. While bands at similar energies are
typically assigned to metal-to-ligand charge transfer in related compounds,[47,48] time-dependent DFT calculations for 4 instead indicate
little involvement of the Re d-orbitals in this transition (vide infra). Overall, the λmax values of 1, 2, and 4 do not differ significantly,
suggesting that the substituent at the para-position
of the N-Ar rings has little influence on the energy of the electronic
transitions, which is in line with the notion that the acceptor orbital
in these transitions is a formazan π*-orbital that is relatively
insensitive to the para-substituent. In the case
of complex 3, λmax is slightly red-shifted
(509 nm) compared to complex 1 (490 nm), indicating that
the involvement of the π-donating OMe group on the conjugated
system is noticeable. When the spectrum of 4 was recorded in acetonitrile, a modest hypsochromic shift was observed
(λmaxAcCN = 483 nm (Figure a); λmaxToluene = 495 nm), showing that these species manifest a small, negative
solvatochromism. Comparing λmax values to those reported
for complexes with anionic formazanate ligands,[9,49,50] the absorption maxima in 1–4 are blue-shifted due to a smaller extent of π-conjugation
within the backbone of the neutral ligands compared to the fully delocalized
anions. The mixture of complexes 5a/b features
two intense bands at 520 and 452 nm derived from the MLCT and π–π*
formazan electronic transitions and a shoulder at 350 nm. Overall,
the influence of the substituents on the lowest energy band is more
pronounced when they are located at the N=N and NH formazan
positions, similar to what was observed in complexes with anionic
formazanate ligands.[6] Clearly, the nature
of electronic absorptions for the two isomers 5a/b is quite distinctive, as is manifested by the significant
shift in λmax.
Figure 4
Absorption spectra for compounds 1–4 and 5a/b in toluene solution.
Table 2
UV–vis Absorption Data
compound
λmax (nm)
ε (M–1·cm–1)
λmax (nm)
ε (M–1·cm–1)
1
356
16,500
490
18,100
2
353
15,300
485
19,400
3
375
10,500
509
14,100
4
353
15,300
495
27,000
5a/b
452
∼520a
Approximate band position due to
its appearance as a shoulder on the more intense absorption of the
other isomer.
Figure 6
(a) Normalized absorption and emission
spectra of 4 in MeCN recorded at λexc = 320 nm; (b) excitation
spectrum of 4 in MeCN. The data were collected at room
temperature.
Absorption spectra for compounds 1–4 and 5a/b in toluene solution.Approximate band position due to
its appearance as a shoulder on the more intense absorption of the
other isomer.
Density Functional Theory (DFT) Calculations
For representative
complexes 4 and 5a/b, geometry
optimizations were carried out in the ground state using density functional
theory (DFT; MN15L[51] functional and def2-TZVP[52] basis set) using the crystallographic coordinates
as a starting point. The geometries were confirmed to be minima on
the potential energy surface by frequency calculations (no imaginary
frequencies); the resulting structures are in good agreement with
the metrical parameters obtained from X-ray diffraction (complexes 4 and 5b), albeit that the Re–Br and hydrazo
N–N bonds are slightly overestimated (Tables S2 and S3). Analysis of the frontier orbitals at the optimized
geometry of 4 showed that the HOMO is mainly localized
on the [ReBr(CO)3] core and is composed of a Re dπ orbital that is antibonding with a bromine p orbital and π-bonding with the CO ligand located trans to Br. The HOMO level also contains some ligand character
(the hydrazo-phenyl fragment). On the other hand, the LUMO is primarily
a π*-orbital of the formazan framework, with a minor Re 5d character (Figure S10).The optimized structures of 5a and 5b are
overall similar, but the variation in the position of the Mes group
(on the azo or hydrazo N-atom, respectively)
leads to somewhat different frontier orbitals. While those of 5b are similar to 4, the HOMO of 5a has noticeably smaller formazan contribution (Figures S12 and S14). The relative stability between the isomeric
forms 5a and 5b was also evaluated based
on these DFT calculations. Using the gas phase geometries, the Gibbs
free energy difference between both compounds was computed in toluene
solution using the solvation energies from SMD calculations and found
to be 1.6 kcal/mol at room temperature, with 5b being
the most stable isomer. Qualitatively, the trend in relative stability
is consistent with our empirical data since compound 5b is the predominant species in the reaction mixture according to
the NMR integration. It should be noted that it is also possible that
the 5a/5b ratio found experimentally is
kinetically controlled as no interconversion between both isomers
was observed.Time-dependent density functional theory (TDDFT)
calculations were
carried out on complex 4 as a representative example.
Relevant excitations were analyzed in more detail using natural transition
orbital calculations (NTOs) to provide insight into their nature.
According to the calculations, the three lowest-energy transitions
in 4 (λcalc = 645, 539 and 526 nm) all
have small oscillator strength and involve transitions from orbitals
centered on the [ReBr(CO)3] fragment (π*(Re-Br) and
(π(Re-CO)) into the formazan π*-orbital. The fourth excited
state, with the highest oscillator strength in the visible range (λcalc = 487 nm; λmax,exp = 495 nm, Figure S11a-b), has a more pronounced formazan
(intraligand) π–π* character, involving an occupied
azo π-orbital as the donor (see Table S4 for the corresponding NTO), but also here the contribution of metal-based
orbitals is still clearly noticeable. Thus, in all excitations in
the visible range, there is extensive mixing between the metal and
ligand orbitals in the ground and excited states, which results in
electronic transitions of mixed nature: all show contributions from
MLCT Re(dπ) → formazan(π*), LLCT Br(p) → formazan(π*), and ILCT azo(π) → formazan(π*)
excitations. The strong metal–ligand orbital mixing results
in reduced charge transfer character in the MLCT bands, which is reflected
in the minor influence of solvent polarity (toluene vs acetonitrile) on the empirical electronic absorption spectrum (vide supra).[53] Similar to the
absorptions in the visible range, analysis of the NTOs of the higher
energy transitions shows that these involve the formazan π*-orbital
as the acceptor and are also highly mixed in character.TDDFT
calculations were performed on the optimized structures of
complexes 5a/b to understand the impact of the pendant
R1/R5 arms on the electronic transitions. The
intense low-energy absorption is computed to be shifted to higher
energy for 5a (444 nm) compared to 5b (489
nm), see Figures S13 and S14, respectively.
This is in agreement with the empirical UV/vis spectrum, which shows
two distinct bands at 452 and 520 nm for the 5a/b mixture. As in 4, the natural transition orbital
pair for the main low-energy excitation in 5b consists
of a ‘hole’ NTO on the [ReBr(CO)3] core,
whereas the excited electron (‘particle’ NTO) consists
primarily of the π* formazan orbital. A comparison of the NTOs
for 5a and 5b shows that the main difference
between the two isomers is found in the hole NTO (Figure ), which has a higher formazan
contribution in 5a. Based on the optimized geometries,
the orientation of the azo-NAr ring changes upon swapping the aryl
groups on the nitrogen atoms (Ph/Mes): the angle between the plane
defined by the five-membered chelate ring and the Mes-substituent
is 73.77° in 5a, whereas the corresponding angle
with the Ph-substituent in 5b is only 39.28°. To
test our hypothesis that the orientation of the azo-NAr group has
a major impact on the spectral properties, we took the geometry of 5b and rotated the N-Ph group out of the ligand plane to be
in the same orientation as the N-Mes group in 5a. This
structure is labeled 5b_rot. The main visible band in
the TDDFT spectrum calculated at the 5b_rot geometry
is blue-shifted by 30 nm (1342 cm–1) compared to 5b, but the other transitions remain at similar energies (Figure S17). An analysis of the orbital mixing
between the azo-NAr ring and the rest of the ligand π-system
confirms that rotating the Ar ring out of the plane disrupts conjugation
(Table S7), and thus we conclude that this
is responsible for the spectral shift observed.
Figure 5
Natural transition orbitals
for the highest-intensity excitation
in the visible (isovalue = 0.05) for (a) 5a and (b) 5b represented as a hole → electron.
Natural transition orbitals
for the highest-intensity excitation
in the visible (isovalue = 0.05) for (a) 5a and (b) 5b represented as a hole → electron.
Luminescence Spectroscopy
The emission spectrum of 4 measured at room temperature in acetonitrile using an excitation
wavelength of 320 nm shows a broad featureless emission band at 380
nm (Figure a). The excited state showed a monoexponential decay
(λex = 370 nm) with a lifetime (τ) of 3.69
ns under a N2 atmosphere, which does not appreciably change
under O2 (τ = 3.60 ns) (see Figure S19). Furthermore, the excitation spectrum (λem = 380 nm) exhibits a broad band centered at 310 nm (Figure b). In contrast to the majority
of fac-[ReX(CO)3(L,L)] complexes with
bidentate N-donor ligands (e.g., α-diimines), which typically
show emission at higher wavelengths (400–600 nm),[54,55] this data shows that the triplet (metal–ligand or ligand-centered)[56] excited states typical for the photoluminescence
of fac-[ReX(CO)3(L,L)] compounds[57] are non-emissive in formazan Re(I) species.
This also stands in contrast to complexes with anionic formazanate
ligands, which show highly tunable emission with large Stokes shifts
at much lower energies.[58−60](a) Normalized absorption and emission
spectra of 4 in MeCN recorded at λexc = 320 nm; (b) excitation
spectrum of 4 in MeCN. The data were collected at room
temperature.
Conclusions
We described the straightforward synthesis
of the first complexes
bearing the neutral formazan ligand toward a group 7 metal. The fac-(CO)3ReBr(formazan) complexes obtained contain
a five-membered chelate ring, in which the ligand binds via the nitrogen
atoms of the azo (N=N-Ar) and hydrazo (C=N-NHAr) groups, which leaves a pendant acidic (exchangeable)
NH moiety in close proximity to the metal center. Structural and spectroscopic
data demonstrate that the formazan ligand is tightly bound to the
metal center, which is due to the strong π-acceptor character
of the ligand. The effect of ligand substituents on the properties
of the complexes is only minor, but the preparation of an asymmetric
derivative with a N-Mes substituent demonstrates that changes in the
sterics shift the electronic absorption spectrum due to changes in
the conjugation within the ligand. Computational studies at the DFT
level confirm a high degree of covalency in the metal-formazan interaction
and highly mixed metal–ligand character of the frontier orbitals,
which is sensitive to the degree of conjugation within the ligand
as demonstrated by sterically switching ‘off’ π-interactions
in the derivative with a N-Mes group (compound 5a). Unlike
many fac-[ReX(CO)3(L,L)] compounds (L,L
= α-diimine ligands) reported in the literature, our formazan
complexes are only weakly luminescent in the near-UV (λem = 380 nm), and emission from the lower-energy excited states
is not observed. In ongoing work, we are investigating the possibility
of using the proton-responsive nature of the NH group (i.e., formazan/formazanate
interconversion) in ‘cooperative’ reactivity of this
type of complexes.
Experimental Section
General Considerations
All work—except ligand
synthesis—was conducted under a nitrogen atmosphere using conventional
Schlenk and vacuum-line techniques. Pentane and toluene (Aldrich,
anhydrous. 99.8%) were passed over columns of Al2O3 (Fluka) and BASF R3-11-supported Cu oxygen scavengers. [ReBr(CO)5] was prepared according to the published procedures[61] from [Re2(CO)10] (Aldrich,
98%) and Br2 (Aldrich, 98%). The ligands 1,5-diphenyl-3-p-tolylformazan (L2H), 1,5-diphenyl-3-p-methoxyphenylformazan (L3H), and 1,5-diphenyl-3-p-flourophenylformazan (L4H) were synthesized
using the methodology reported by Hicks and coworkers.[4] Particularly, 5-mesityl-1-phenyl-3-p-tolylformazan
(L5H) was prepared using a modified procedure published
by our group.[9] 1,3,5-Triphenylformazan
(L1H, TCI, 92%), p-tolualdehyde (Aldrich
97%), 4-methoxybenzaldehyde (Aldrich, 98%), 4-fluorobenzaldehyde (Aldrich,
98%), phenylhydrazine (Aldrich, 99%), aniline (Aldrich 99%), sodium
nitrite (Aldrich, 99%), sodium carbonate (Aldrich, 99.5%), and [NBu4]Br were used as received. CHCl3 (Aldrich, 99%)
and CDCl3 (Aldrich, 99.8 atom %D) were used without further
purification.NMR spectra were measured on Mercury 400, Varian
Inova 500, or Bruker 600 MHz spectrometers. Residual solvent signals
were used as internal reference for 1H and 13C spectra and reported in ppm relative to TMS (0 ppm). Complete assignments
were based on two-dimensional experiments (COSY, HSQC, HMBC) using
standard pulse sequences. FT-IR spectra were collected in DCM solution
on a JASCO 4700 series FT-IR spectrometer in transmission mode using
a liquid cell with CaF2 windows. UV–vis spectra
were recorded in toluene solution on an Agilent Technologies 8453
UV–vis spectrophotometer. Luminescence spectroscopy were measured
on a Fluorolog-3 spectrometer from HORIBA Jobin Yvon.X-ray
diffraction data were collected at 100 K on a Bruker D8 Venture
diffractometer with a Mo Kα (λ = 0.71073 Å) (compounds 3 and 4) or Cu Kα (λ = 1.54178 Å)
(compound 5b) radiation source. Crystal structures were
refined using the SHELXL[62] software (Table ). Non-hydrogen atoms
were refined anisotropically.
Table 3
Crystallographic Data for 3, 4, and 5b
3
4
5b
chemical formula
C23H18BrN4O4Re
C22H15BrFN4O3Re
C26H24BrN4O3Re
Mr
680.52
668.49
706.6
cryst syst
triclinic
monoclinic
triclinic
color, habit
purple, needle
purple, needle
purple, block
size (nm)
0.42 ×0.13× 0.04
0.70 × 0.170 ×
0.060
0.40 × 0.33 × 0.20
space group
P1̅ (No.2)
P21/n (No.14)
P1̅ (No.2)
a (Å)
9.0463(6)
10.095(2)
9.9235(12)
b (Å)
11.0459(7)
18.586(6)
10.9196(13)
c (Å)
12.7033(8)
12.219(4)
13.0134(15)
α
(deg)
107.379(2)
90
77.085(4)
β (deg)
94.935(3)
107.191(8)
72.218(4)
γ
(deg)
103.227(2)
90
75.163(4)
V (Å3)
1162.89(13)
2190.1(11)
1281.7(3)
Z
2
4
2
ρcalc (g·cm–3)
2.302
2.027
1.831
radiation, λ
(Å)
Mo, Kα, 0.71073
Mo, Kα,
0.71073
Cu, Kα, 1.54178
μ(Mo, Kα) (mm–1)
8.777
7.413
11.384
F(000)
750
1272
684
temp (K)
100(2)
100(2)
100(2)
θ
range (deg)
3.028–27.191
3.044–27.200
3.612–70.304
data collected
(h, k, l)
–11:11, −14:14, −16:16
–12:12,
−23:23, −15:15
–12:12, −13:13,
−15:15
no. of reflns collected
45,523
27,390
21,942
no. of indep reflns
5166
4818
4526
obsd reflns Fo ≥ 2.0σ(Fo)
4966
4288
4396
R(F) [obsd reflns] (%)
1.78
2.81
2.19
Rw(F2) [all reflns] (%)
5.24
6.13
5.51
GOF
1.051
1.192
1.119
weighting a, b
0.03300, 1.26140
0.0000, 12.4515
0.0000, 2.9889
params refined
303
293
324
min, max residual densities
–1.04, 1.32
–1.61, 1.82
–0.83, 0.85
Computational Details
Density functional theory (DFT)
calculations were carried out in Gaussian 16 Revision C.02[63] software and visualized using Gaussview 6[64] or Avogadro.[65] Geometry
optimizations in the ground state were performed in the gas phase
at a MN15L[51] level of theory combined with
a triple ζ-basis set: def2tzvp.[52] The carbonyl frequency values were scaled using a factor of 0.9578[66] (see Table S1). TDDFT
calculations were performed on the optimized structures at the CAM-B3LYP/def2tzvp
level of theory. The solvent effect was simulated using the continuum
polarized model (CPM).[67]
Procedure for the Synthesis of Complexes 1–4
Equimolar amounts of [ReBr(CO)5] and
the corresponding ligand were poured into a two-necked round bottom
flask and dissolved in 20 mL of toluene. The reaction was heated up
at reflux for 1 h observing that the mixture darkened upon completion.
The solvent was evaporated to dryness. Specific details for the purification
of the entitled complexes are mentioned below.
1 (C22H16BrN4O3Re)
[ReBr(CO)5] (0.0934 g, 0.230 mmol), L1H (0.06938 g, 0.230 mmol). Work-up: 5 mL of pentane was
added to the mixture, and the crude was stirred for 30 min allowing
the formation of a dark-crimson solid material. The compound was filtered
out and rinsed with pentane (3 × 5 mL). (41.8 mg, 27.9%). 1H NMR (CDCl3, 25 °C, 400 MHz) δ/ppm:
7.30 (d, 2 H,[3]J = 8 Hz,
Ph-NH o-H), 7.39 (t, 1 H,[3]J = 8 Hz, Ph-NH p-H), 7.45–7.57
(m, 5H, Ph-NH m-H, Ph-N=N m-H, Ph-NC p-H), 7.64 (m, 3H, Ph-NC m-H, Ph-N=N p-H,) 7.84 (d, 2 H,[3]J = 8 Hz, Ph-NC o-H), 7.89 (d,
2 H,[3]J = 8 Hz, Ph-N=N o-H), 8.58 (s, 1 H, NH). 13C{1H} NMR
(CDCl3, 25 °C, 150 MHz) δ/ppm: 123.21 (Ph-NH o-CH), 123.95 (Ph-N=N o-CH), 127.90
(Ph-NH p-CH), 128.48 (Ph-CN ipso-C), 128.93 (Ph-CN o-CH), 129.31 (Ph-N=N m-CH), 129.39 (Ph-NH m-CH), 130.10 (Ph-CN m-CH), 131.69 (Ph-N=N p-CH), 131.86
(Ph-CN p-CH), 140.83 (Ph-NH ipso-C), 157.20 (Ph-N=N ipso-C), 164.65 (NCN
C), 185.33 (CO trans Br C), 192.35 (CO trans Ph-NH-N C), 192.89 (CO trans Ph-N=N C).
IR(CH2Cl2) ν(CO)/cm–1: 2035(s), 1959(s), 1923(s). MS (FAB+) (m/z): [MH + 2]+ = 653, [MH]+ = 651,
[MH-CO]+ = 623, [MH-3CO]+ = 567. HRMS (ESI +)
(m/z): Calcd. for [MH] + = 651.004167. Found = 651.00337. [MH-3CO]+ = 567.01942.
Found = 567.01817.
Authors: Ryan R Maar; Stephanie M Barbon; Neha Sharma; Hilary Groom; Leonard G Luyt; Joe B Gilroy Journal: Chemistry Date: 2015-09-23 Impact factor: 5.236
Authors: Gerald F Manbeck; James T Muckerman; David J Szalda; Yuichiro Himeda; Etsuko Fujita Journal: J Phys Chem B Date: 2015-03-02 Impact factor: 2.991
Authors: Ryan R Maar; Ruizhong Zhang; David G Stephens; Zhifeng Ding; Joe B Gilroy Journal: Angew Chem Int Ed Engl Date: 2018-12-20 Impact factor: 15.336