Photochemistry of chromium arene tricarbonyl complexes with tethered pyridinyl and propenyl groups: Investigations of the effect of ring size on chelate formation, structure, and linkage isomerization

Article abstract of DOI:10.1021/om400928k

Chromium arene tricarbonyl complexes with tethered pyridinyl groups [Cr{η⁶-C₆H₅(CH₂) _n(2-Py)}(CO)₃] (4-6) (2-Py = 2-pyridinyl, n = 1-3, respectively) were synthesized and irradiated to form the chelates [Cr{η⁶-C₆H₅(CH₂) _n(2-Py)-κN}(CO)₂] (7-9). Studies examined the effect of ring size and structure on chromophore λ_max, stability, and photosensitivity, which are factors important for photochromes based on linkage isomerization of tethered functional groups. The studies also include [Cr{η⁶-C₆H₅CH(2-Py)CH₂CH=CH ₂}(CO)₃] (3), which has a bifunctional tether of propenyl and pyridinyl groups, and irradiation produces the linkage isomers [Cr{η⁶-C₆H₅(CH(2-Py)CH₂CH= CH₂)-κN}(CO)₂] (1) and [Cr{η⁶-C ₆H₅(CH(2-Py)CH₂CH=CH₂) (η²-CH=CH₂)}(CO)₂] (2). X-ray crystal structures for 7-9 show that the dihedral angle between the coordinated pyridinyl groups and the phenyl-chromium centroid increases from 1 to 73 (n = 1-3, respectively). The experimental and TDDFT computed optical changes accompanying an increase in the dihedral angle are modest and not monotonic for 7-9 due to structural changes inherent in the chelate rings. An increase in Cr-N bond lengths and decrease in their bond energies were observed experimentally and computationally for the series of 7-9. The quantum yields for formation of the five-, six-, and seven-membered chelate rings during the conversion of 4-6 to 7-9, respectively, were within experimental error for that observed for conversion of 10 [Cr{η⁶-C₆H₆}(CO) ₃] with free pyridine to 11 [Cr{η⁶-C₆H ₆}(C₅H₅N-κN)(CO)₂], indicating that the product-determining step precedes chelation. The enthalpies for chelation of 4-6 to 7-9 were determined independently by photoacoustic calorimetry and DFT computations. The computationally derived mechanism for thermal isomerization of 1 to 2 indicates that the transition state is a dissociative interchange with a free energy of activation of 27.9 kcal mol ^-1 (1 → 2), a result consistent with an experimentally bistable photochrome. The results indicate which tether properties are important for optimizing photochrome performance.

Full text of DOI:10.1021/om400928k

Article
pubs.acs.org/Organometallics
Photochemistry of Chromium Arene Tricarbonyl Complexes with
Tethered Pyridinyl and Propenyl Groups: Investigations of the Effect
of Ring Size on Chelate Formation, Structure, and Linkage
Isomerization
Charles B. Duke III,† Roger G. Letterman,† Jermaine O. Johnson,† James W. Barr,† Songnan Hu,†
Charles R. Ross II,‡,§ Charles Edwin Webster,*,† and Theodore J. Burkey*,†
†Department of Chemistry, The University of Memphis, 213 Smith Chemistry Building, Memphis, Tennessee 38152-3550, United
States
‡Department of Structural Biology, St. Jude Children’s Research Hospital, 332 North Lauderdale, Memphis, Tennessee 38105-2794,
United States
S
* Supporting Information
ABSTRACT: Chromium arene tricarbonyl complexes with
tethered pyridinyl groups [Cr{η6-C6H5(CH2)n(2-Py)}(CO)3]
(4−6) (2-Py = 2-pyridinyl, n = 1−3, respectively) were
synthesized and irradiated to form the chelates [Cr{η6-
C6H5(CH2)n(2-Py)-κN}(CO)2] (7−9). Studies examined the
effect of ring size and structure on chromophore λmax, stability,
and photosensitivity, which are factors important for photo-
chromes based on linkage isomerization of tethered functional
groups. The studies also include [Cr{η6-C6H5CH(2-Py)CH2CHCH2}(CO)3] (3), which has a bifunctional tether of propenyl
and pyridinyl groups, and irradiation produces the linkage isomers [Cr{η6-C6H5(CH(2-Py)CH2CHCH2)-κN}(CO)2] (1) and
[Cr{η6-C6H5(CH(2-Py)CH2CHCH2)(η2-CHCH2)}(CO)2] (2). X-ray crystal structures for 7−9 show that the dihedral
angle between the coordinated pyridinyl groups and the phenyl-chromium centroid increases from 1 to 73° (n = 1−3,
respectively). The experimental and TDDFT computed optical changes accompanying an increase in the dihedral angle are
modest and not monotonic for 7−9 due to structural changes inherent in the chelate rings. An increase in Cr−N bond lengths
and decrease in their bond energies were observed experimentally and computationally for the series of 7−9. The quantum yields
for formation of the five-, six-, and seven-membered chelate rings during the conversion of 4−6 to 7−9, respectively, were within
experimental error for that observed for conversion of 10 [Cr{η6-C6H6}(CO)3] with free pyridine to 11 [Cr{η6-C6H6}(C5H5N-
κN)(CO)2], indicating that the product-determining step precedes chelation. The enthalpies for chelation of 4−6 to 7−9 were
determined independently by photoacoustic calorimetry and DFT computations. The computationally derived mechanism for
thermal isomerization of 1 to 2 indicates that the transition state is a dissociative interchange with a free energy of activation of
27.9 kcal mol−1 (1 → 2), a result consistent with an experimentally bistable photochrome. The results indicate which tether
properties are important for optimizing photochrome performance.
Scheme 1
INTRODUCTION
■
Recent studies have examined organometallics that undergo a
linkage isomerization as a photochromic mechanism.1,2 This is
a departure from other studies where the metal serves as a
spectator and binds to a photochromic ligand but is not
involved in bond-forming or -breaking processes.3 Our group
investigates the design of photochromes based on linkage
isomerization: in particular, systems where functional groups
tethered to a metal center undergo a photoinitiated exchange
(Scheme 1). The isomerization in this case relies on the
reversible interconversion of two chelates, each with a different
coordinated functional group and ultimately a different ring
system. The length of a tether and its composition
(heteroatom, branching, or multiple bonds) are anticipated to
affect the chelate ring properties, including the rate of chelate
ring formation, ring strain, and orientation of the chelated
functional group. Likewise, these ring effects can alter
photochrome properties and performance such as fatigue,
quantum yield, response times, and resolution of isomer
chromophores. We previously described the photochromic
behavior of the chromium arene chelates 1 and 2 (Scheme 2).
Received: September 16, 2013
Published: January 14, 2014
© 2014 American Chemical Society
485
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
This system is a bistable photochrome where two functional
groups share a tether but suffer from fatigue and a relatively
slow response time (50 ns).2
properties. By studying the conversion of 4−6 to 7−9,
respectively, we are examining models for half the photo-
chromic response: the conversion of 2 to 1.
Previous photochelation studies of tethered functional
groups (olefin and phosphite) for related chromium arene
complexes established which chain lengths resulted in thermally
stable chelates.4−7 The thermal stability of these chelates also
depends on the tethered functional group. In an effort to obtain
a comprehensive understanding of the tether properties on
photochelation, we compare 4−9 with the untethered
compounds 10 or 11 (Scheme 4) and report the effects of
Scheme 2
Scheme 4
In an effort to understand how the properties of the pyridine
chelate affect the performance of 1/2, the formation of 1 from
its parent tricarbonyl 3 was examined by time-resolved IR
spectroscopy.2 The results were compared with the ultrafast
chelation of chromium benzene tricarbonyl derivatives 4−6,
which have different pyridine tether lengths (Scheme 3). The
Scheme 3
tether length on (a) the stability of the pyridinyl chelate, (b)
Cr−N bond strengths, (c) the orientation of the pyridinyl
group and related bond angles, (d) electronic transition
energies, and (e) the quantum yield of chelate formation.
The effect of tether length on these properties is also expected
to affect the performance of photochromes based on tethered
linkage isomerization. Changing the tether length could confine
the pyridinyl group to different orientations (i.e., rotation about
the Cr−N bond and changes in N−Cr−benzene-centroid (Ct)
angle, item c), leading to changes in orbital overlap with metal-
centered orbitals. Thus, changes in bond energies and
HOMO−LUMO transitions (items a, b, and d) might be
observed. The faster rate of chelation reported for 4 could
potentially lead to a higher quantum yield (item e). For
example, we reported that the quantum yields for photo-
initiated chelation of manganese cyclopentadienyl tricarbonyl
derivatives with different tethered functional groups varied from
0.65 to 1.00.8 Unit quantum yields demand that every absorbed
photon results in CO dissociation with no cage CO
recombination.9 For comparison, the quantum yield for the
analogous bimolecular substitution of CpMn(CO)3 was only
0.65 and was found to be independent of ligand structure.10
Finally, we also examine the bifunctional derivative 3 and the
corresponding chelates 1 and 2, which have a propenyl group in
addition to a pyridinyl group (Scheme 3, n = 1).11 In particular,
we examine the mechanism for thermal isomerization between
1 and 2 (Scheme 2), since an energetically significant free
energy of activation is essential for a thermally bistable
photochrome. A comparison of this mechanism with those
for bifunctional tethered chelates where the linkage isomer-
ization has a low free energy of activation (and thus are not
bistable) provides a rational basis for the design of new bistable
organometallic photochromes.
rate of chelation following CO dissociation was observed to
decrease with tether length, and the mechanism for formation
of 7 (a single-methylene tether) was found to be different from
that of 8 and 9 (two- and three-methylene tethers,
respectively).2 Specifically, irradiation of 5 and 6 initially
formed only solvent-coordinated intermediates (<100 ps in
heptane) which later converted to chelates 8 and 9 (100 and
200 ns, respectively). In contrast, the irradiation of 4 initially
formed three complexes (<100 ps in heptane): this included
the chelate 7 as well as an agostic complex (hydrogen at the
pyridinyl C3) and a solvent-coordinated complex. These last
two transient complexes subsequently convert to 7 within 50
ns.2 Thus, only 4 formed a chelate (7) immediately in
competition with a solvent-coordinated intermediate. The
ultrafast dynamics following irradiation of 4 was found to be
nearly the same as that for 3, except 3 immediately formed a
propenyl chelate in competition with three other species: an
agostic complex with the pyridinyl group, a solvent-coordinated
complex, and the pyridinyl chelate.
Ring formation involving a tethered functional group is a
fundamental process for the forward and reverse reactions in
Schemes 1 and 2; thus, controlling the factors that affect the
rate, stability, and reversibility of ring formation is important for
photochrome performance. These properties determine to
what extent competing processes may interfere with the ring
formation (and consequently the photochromic response).
Furthermore, the functional group orientation can potentially
affect spectroscopic (via orbital overlap) and photochemical
EXPERIMENTAL SECTION
■
Instrumentation. Spectra were obtained on a 270 MHz JEOL
NMR spectrometer, a 500 MHz Varian NMR System, Mattson Galaxy
2020 FT-IR spectrometer, and an Agilent 8453 UV−vis spectropho-
tometer. Chemical shifts were referenced to residual solvent peaks as
internal standards, and coupling constant errors are typically 0.2 Hz.12
UV irradiation was supplied by a Rayonet photochemical reactor
(RPR-100) equipped with eight 300 nm lamps (RPR-3000), a single
300 nm lamp, or a nitrogen laser (PTI GL3300 337.1 nm N2 laser, 0.9
ns pulse width, 1.4 mJ/pulse). Visible irradiation was supplied by a 600
486
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
W quartz lamp with a UV and water filter. The extent of each
photochemical reaction was determined with IR or 1H NMR
spectroscopy. All organometallic solutions for actinometry, extinction
coefficients, and photochemical synthesis were prepared under
nitrogen or argon. Similarly, helium was used for photoacoustic
calorimetry.
1.89 (d, 1 H, J = 8.8 Hz, cis CH2); minor isomer (30%) δ 8.33 (m, 1
H, 6-py), 6.94 (dt, J = 7.7, 1.8 Hz, 1 H, 4-py), 6.51 (m, 2 H, 3,5-py),
5.37 (d, J = 5.4 Hz, 1 H, o-Ph), 4.79 (t, J = 6.2 Hz, 1 H, m-Ph), 4.72 (t,
J = 5.6 Hz, 1 H, m-Ph), 4.53 (d, J = 6.5 Hz, 1 H, o-Ph), 4.17 (m, 1 H,
CH), 2.97 (m, 2 H, p-Ph, cis CH2), 2.6−2.3 (m, 3 H, CH, CH2),
2.36 (d, 1 H, J = 8.4 Hz, trans CH2). 13C NMR (C6D6): major
isomer δ 248.0, 245.6, 160.7, 149.7, 135.9, 121.9, 121.7, 99.2, 96.9,
94.9, 92.2, 89.2, 80.3, 62.7, 47.5, 44.3, 40.8; minor isomer δ 244.0,
239.6, 160.2, 149.5, 135.9, 126.1, 121.8, 94.9, 92.5, 91.6, 87.5, 87.4,
81.3, 70.2, 59.1, 56.6, 48.5.
X-ray Diffraction.13 The single-crystal structures of 1−9 and 11
were determined using a Bruker AXS Proteum diffractometer with a
CCD area detector and Osmic mirror-monochromated rotating anode
Cu source (Kα radiation, ν = 1.54178 Å). Crystals coated with
paratone oil (Exxon) were mounted on a polymer loop under a
nitrogen stream at −100 °C. Global refinement of unit cells,
orientation matrix data, genetic-algorithm data collection strategy
design and collection, and data reductions were performed using
Proteum2 software.14 Structure refinement was performed using the
SHELXTL 5.1 and WinGX software packages.15,16 The SADABS
program was used to make empirical absorption corrections.17 ORTEP
graphics were created using RastTEP/Raster3D.18 Complete listings of
atom coordinates, bond distances, angles, and atomic displacement
factors may be found in the Supporting Information.
[Cr{η6-C6H5(CH(2-C5H4N)CH2CHCH2)}(CO)3] (3). n-BuLi
(0.928 mL, 1.85 mmol) was added to 4 (512 mg, 1.68 mmol) in 25
mL of Et2O at 195 K. The mixture was warmed to 273 K for 1.5 h, and
allyl bromide (236 μL, 2.7 mmol) was added. The mixture was stirred
for 1 h and filtered to provide a yellow solid upon drying in vacuo.
Recrystallization in Et2O/hexanes produced a yellow solid (430 mg,
74%) from which crystallographic quality crystals were obtained. Anal.
Found (calcd): C, 61.5 (62.6);22 H, 4.11 (4.38); N, 3.85 (4.06). H
1
NMR (C6D6): δ 8.41 (br d, J = 5 Hz, 1 H, 6-py), 7.07 (td, J = 7.7, 1.8
Hz, 1 H, 4-py), 6.89 (br d, J = 7.7 Hz, 1 H, 3-py), 6.59 (ddd, J = 7.5,
4.7, 1.1 Hz, 1 H, 5-py), 5.55 (m, 1 H, CHC), 5.23 (d, J = 6.5 Hz, 1
H, CH2C), 4.88 (m, 3 H, 3,5-Ar, CH2C), 4.37 (m, 2 H, 2,6-Ar),
4.27 (m, 1 H, 4-Ar), 3.38 (dd, J = 5.3, 4.4 Hz, 1 H, CH), 2.92 (m, 1 H,
CH2), 2.60 (m, 1 H, CH2). 13C NMR (C6D6): δ 233.5, 161.0, 149.5,
136.5, 136.0, 124.2, 122.2, 117.1, 114.2, 94.8, 94.6, 93.1, 91.2, 90.7,
51.8, 40.8.
Materials. Except where noted, chemicals were used as received
from Sigma-Aldrich, and procedures were carried out under an argon
or nitrogen atmosphere using a glovebox or standard Schlenk
techniques. The argon for the Schlenk techniques was purified with
an Ace-Burlitch Inert Atmosphere System (ACE Glass, Inc.). 10 (η6-
C6H6)Cr(CO)3 was sublimed twice at 45 °C and 25 mTorr. Pyridine
(from Mallinckrodt) was distilled after refluxing over calcium hydride
or NaK alloy under argon. THF (Fisher), hexane, and heptane were
distilled under argon after refluxing over NaK alloy. A photochromic
fulgide (3-[1-(2,5-dimethyl-3-furanyl)ethylidene]dihydro-4-(1-methyl-
ethylidene)-(3E)-2,5-furandione) was prepared by a literature method
and used as a chemical actinometer.19
2-(2-Phenylethyl)pyridine. 2-Stilbazole (500 mg, 2.76 mmol)
was added to 30 mL of absolute ethanol and 18 mg of Pd/C (10%) in
a 200 mL hydrogenation bottle. The reaction mixture was shaken with
50 psi of H2 for 2 days. After filtration through Celite and
concentration by evaporation, 507 mg of a clear liquid was recovered
(2.76 mmol, 100%). 1H NMR (CDCl3): δ 8.56 (m, 1 H, 6-py), 7.59 (t,
1 H, J = 7.7 Hz, 4-py), 7.2 (m, 2H, Ph, 3,5-py, overlap with CHCl3),
3.09 (m, 4 H, CH2CH2). 13C NMR (CDCl3): δ 160.9, 148.5, 141.3,
137.3, 128.5, 126.1, 123.6, 121.6, 77.4, 39.7, 36.0.
[Cr{η6-C6H5CH2(2-C5H4N)}(CO)3] (4). In a typical synthesis using
the Mahaffey and Pauson method,21 2-benzylpyridine (2.97 g, 17.6
mmol), Cr(CO)6 (3.86 g, 17.5 mmol), Bu2O (120 mL), and THF (12
mL) were refluxed for 48 h. The solution was concentrated by rotary
evaporation, and 4.10 g of a yellow powder was crystallized from Et2O
and hexanes at 253 K (13.43 mmol, 77%). Crystals suitable for
crystallography were grown from Et2O/hexanes at 253 K. Anal. Found
1
(calcd): C, 58.78 (59.02); H, 3.30 (3.63); N, 4.41 (4.59). H NMR
(C6D6): δ 8.34 (br d, J = 3.7 Hz, 1 H, 6-py), 7.00 (br t, J = 7.6 Hz, 1 H,
4-py), 6.71 (d, J = 7.6 Hz, 1 H, 3-py), 6.56 (m, 1 H, 5-py), 4.71 (br d, J
= 6.0 Hz, 2 H, 2,6-Ph), 4.44 (t, J = 6.2 Hz, 2 H, 3,5-Ph), 4.27 (t, J = 6.0
Hz, 1 H, 4-Ph), 3.43 (s, 2 H, CH2). 13C NMR (C6D6): δ 233.7, 158.8,
149.8, 136.4, 123.0, 121.9, 110.7, 93.8, 93.12, 90.9, 43.3.
[Cr{η6-C6H5(CH2)2(2-C5H4N)}(CO)3] (5). Similar to the synthesis of
4, Cr(CO)6 (610 mg, 2.77 mmol), 2-(2-phenylethyl)pyridine (503 mg,
2.74 mmol), 15 mL of Bu2O, and 1.5 mL of THF were refluxed for 20
h. After a black solution was concentrated in vacuo, the residue was
dissolved in Et2O, and filtration through Celite afforded a clear yellow
solution, which was concentrated by rotary evaporation. Flash
chromatography on an AnaLogix SuperFlash SF25-34g silica column
yielded 0.709 g of a yellow-orange oil (2.22 mmol, 81%). Crystals
suitable for crystallography were grown from Et2O/hexanes at 253 K.
Anal. Found (calcd): C, 59.88 (60.19); H, 4.05 (4.10); N, 4.37 (4.39).
1H NMR (C6D6): δ 8.42 (d, 1 H, J = 4.6 Hz, 6-py), 6.99 (td, 1 H, J =
7.7, 1.9 Hz, 4-py), 6.59 (m, 2 H, 3,5-py), 4.41 (m, 4 H, 2,3,5,6-Ph),
4.22 (t, 1 H, J = 3.0 Hz, 4-Ph), 2.69 (m, 2 H, CH2py), 2.54 (m, 2 H,
CH2Ph). 13C NMR (C6D6): δ 233.8, 160.1, 149.8, 136.0, 122.9, 121.4,
113.1, 93.6, 92.7, 90.2, 39.3, 34.4.
[Cr{η6-C6H5(CH(2-C5H4N)CH2CHCH2)-κN}(CO)2] (1). The tri-
carbonyl complex 3 (153 mg, 0.45 mmol) in THF (2.5 mL) and 50
mL of hexane was irradiated in a 50 mL Schlenk flask for 1 h with the
Rayonet reactor while the mixture was stirred and the headspace
purged with argon. The red solution was concentrated in vacuo, and a
portion of the crude product was recrystallized from Et2O/hexanes,
yielding 9.1 mg (6.5%) of a red solid, from which crystallographic
quality crystals were selected. Anal. Found (calcd): C, 60.58 (64.35);22
H, 4.44 (4.76); N, 4.03 (4.41). 1H NMR (C6D6): δ 8.29 (br d, J = 5.3
Hz, 1 H, 6-py), 6.47 (br t, J = 7.3 Hz, 1 H, 4-py), 5.96 (br d, J = 7.7
Hz, 1 H, 3-py), 5.83 (br t, J = 6.2 Hz, 1 H, 5-py), 5.54 (m, 1 H, =CH),
4.96 (m, 3 H, =CH2, o-Ph), 4.77 (m, 2H, m-Ph) 4.57 (d, 6.1 Hz, 1 H,
o-Ph), 2.89 (m, 2 H, p-Ph, CHC3), 2.04 (m, 2 H, CH2). 13C NMR
(C6D6): δ 245.1 (CO), 244.8 (CO), 173.5, 156.9, 135.0, 134.5, 121.7,
121.0, 117.4, 114.0, 92.4, 90.7, 90.3, 86.9, 77.7, 51.0, 36.3.
[Cr{η6-C6H5(CH2)3(2-C5H4N)}(CO)3] (6). In a procedure similar to
the synthesis of 4, Cr(CO)6 (2.00 g, 9.09 mmol), 2-(3-phenylpropyl)-
pyridine (1.78 g, 9.02 mmol), 60 mL of Bu2O, and 7 mL of THF
provided 2.67 g of a yellow solid (8.01 mmol, 89%). Crystals suitable
for crystallography were grown from Et2O/hexanes at 253 K. Anal.
[Cr{η6-C6H5(CH(2-C5H4N)CH2CHCH2)(η2-CHCH2)}(CO)2]
(2). In a synthesis similar to that for 1, 3 (152 mg, 0.44 mmol) was
irradiated in the Rayonet reactor followed by 1.5 h of irradiation with
visible light (600 W Commercial Electric tungsten halogen lamp with
water and UV filters). The solution was concentrated to 80% of its
original volume and cooled to 253 K to obtain 289 mg (20%) of a red
crystalline solid. Anal. Found (calcd): C, 63.81 (64.35);22 H, 4.27
(4.76); N, 4.15 (4.41). 1H NMR (C6D6):20 immediately following
irradiation, major isomer (70%) δ 8.37 (m, 1 H, 6-py), 6.92 (td, J =
7.7, 1.8 Hz, 1 H, 4-py), 6.53 (m, 2 H, 3,5-py), 5.52 (d, J = 6.2 Hz, 1 H,
o-Ph), 5.25 (t, J = 6.2 Hz, 1 H, m-Ph), 4.84 (t, J = 5.9 Hz, 1 H, m-Ph),
4.17 (d, J = 5.9 Hz, 1 H, o-Ph), 3.98 (tt, J = 5.9, 1.0 Hz, 1 H, p-Ph),
3.58 (m, 1 H, CH), 3.46 (dddd, J = 13, 9, 8, 1.5 Hz, 1 H, CH), 3.23
(m, 1 H, trans CH2), 2.98 (m, 1 H, CH2), 2.6−2.4 (m, 1 H, CH2),
1
Found (calcd): C, 60.84 (61.26); H, 4.17 (4.54); N, 4.08 (4.20). H
NMR (C6D6): δ 8.47 (br d, J = 4.8 Hz, 1 H, 6-Py), 7.05 (td, J = 7.7,
1.9 Hz, 1 H, 4-Py), 6.64 (m, 2 H, 3,5-Py), 4.48 (t, J = 6.4 Hz, 2 H, 3,5-
Ph), 4.36 (br d, J = 6.6 Hz, 2 H, 2,6-Ph), 4.25 (br t, J = 6.2 Hz, 1 H, 4-
Ph), 2.56 (t, J = 7.5 Hz, 2 H, CH2), 1.99 (m, 2 H, CH2), 1.73 (m, 2 H,
CH2). 13C NMR (C6D6): δ 233.8, 161.5, 149.8, 135.9, 122.7, 121.2,
113.6, 93.8, 92.6, 90.1, 37.4, 34.5, 30.9.
[Cr{η6-C6H5CH2(2-C5H4N)-κN}(CO)2] (7). 4 (203 mg, 665 μmol)
and 200 mL of hexane in a Pyrex flask were purged with argon during
irradiation with the Rayonet reactor for 3 h, and a precipitate formed
with a red solution. After the mixture was cooled to 261 K, filtration
487
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
yielded 113 mg (61%) of red needlelike crystals. Anal. Found (calcd):
C, 60.31 (60.65); H, 3.11 (4.00);22 N, 4.82 (5.05). 1H NMR (C6D6): δ
8.21 (br d, J = 5.1 Hz, 1 H, 6-Py), 6.41 (t, J = 7.7 Hz, 1.5 H, 4-Py),
5.82 (m, 2 H, 3,5-py), 4.90 (t, J = 5.8 Hz, 2 H, 3,5-Ph), 4.68 (d, J = 5.5
Hz, 2 H, 2,6-Ph), 2.82 (br t, J = 5.9 Hz, 1 H, 4-Ph), 2.72 (s, 2 H, CH2).
13C NMR (C6D6): δ 245.1, 170.9, 156.7, 134.4, 121.6, 120.8, 111.4,
91.3, 89.0, 77.4, 42.4.
Photoacoustic Calorimetry. The instrument, sample preparation,
and data collection have been described recently.26 Typical conditions
were 70 mM ferrocene in heptane and 0.020 mM 4−6 or chromium
benzene tricarbonyl with 24−280 mM pyridine for the latter. The
pressure transducer of a photoacoustic calorimeter responds to
pressure changes caused by thermal expansion and reaction volumes
that occur faster than or comparable to the response time of the
transducer.26 The amplitude of the PAC signal is expressed as a
fraction of the energy of the absorbed photon. For processes much
faster than the response time of the transducer, the relative amplitude
(ϕ1) depends on the energy of the photon (hν), the reaction enthalpy
(ΔH1), the volume change (ΔV1) associated with processes much
faster than the transducer response, the quantum yield (Φ), and the
thermoelastic properties of the solvent (χ = (MW)β/Cpρ where MW
= molecular weight, β = coefficient of thermal expansion, Cp = molar
heat capacity, and ρ = density) (eq 1).27 The amplitude of the PAC
[Cr{η6-C6H5(CH2)2(2-C5H4N)-κN}(CO)2] (8). Heptane (50 mL) and
5 (126 mg, 0.39 mmol) were irradiated for 3 h in a septum-sealed 50
mL graduated cylinder with a Rayonet reactor while purging with
argon. A red solution with precipitate was cooled to 261 K overnight
and filtered to obtain 64 mg (49%) of crystallography-quality crystals.
Anal. Found (calcd): C, 59.7 (61.9);22 H, 4.5 (4.5); N, 4.6 (4.8). H
1
NMR (10 °C, THF-d8): δ 9.23 (br d, J = 5.7 Hz, 1 H, 6-py), 7.49 (td, J
= 7.5, 1.5 Hz, 4-py), 7.15 (br d, J = 7.5 Hz, 1 H, 3-py), 6.79 (ddd, J =
7.3, 5.6, 1.4 Hz, 1 H, 5-py), 5.13 (d, J = 6.1 Hz, 2 H, 2,6-Ph), 4.66 (t, J
= 6.0 Hz, 2 H, 3,5-Ph), 3.58 (obscured by solvent, it appears at 3.69
ppm at −96 °C, t, J = 5.6 Hz, 1 H, 4-Ph; see Figure S54, Supporting
Information), 3.10 (m, 2 H, CH2), 2.28 (m, 2 H, CH2). 13C NMR
(C6D6): δ 245.2, 165.3, 158.8, 135.4, 124.5, 120.5, 106.6, 95.3, 87.9,
79.0, 42.3, 30.2.
hνχ − Φ(ΔH χ − ΔV)
1
1
φ1 =
hνχ
(1)
[Cr{η6-C6H5(CH2)3(2-C5H4N)-κN}(CO)2] (9). The synthesis was
similar to that of 8 with an isolated yield of 74%. Anal. Found (calcd):
signal for subsequent reactions on a time scale similar to the
transducer response time (ϕ2) has an analogous dependence without
the photon contribution (eq 2). Rearrangements to eqs 3 and 4
C, 61.7 (63.0);22 H, 4.4 (5.0); N, 4.5 (4.6). H NMR (C6D6): δ 9.42
1
(br d, J = 5.5 Hz, 1 H, 6-Py), 6.59 (br t, J = 7.6 Hz, 1 H, 4-Py), 6.19 (d,
J = 7.7 Hz, 1 H, 3-py), 5.98 (br t, J = 6.3 Hz, 1 H, 5-py), 4.66 (br m, 2
H, 3,5-Ph), 4.35 (br m, 2 H, 2,6-Ph), 3.25 (m, 3 H, 4-Ph, CH2), 1.38
(br m, 2 H, CH2) 1.27 (p, J = 6.4 Hz, 2H, CH2). 13C NMR (C6D6): δ
243.5, 166.3, 157.5, 135.3, 124.2, 119.3, 105.7, 92.3, 77.7, 36.1, 28.4,
24.0.
Φ(ΔH2 χ − ΔV )
2
φ2 = −
hνχ
(2)
hνχ(1 − φ1)
[Cr(η6-C6H6)(κ1-C5H5N)(CO)2] (11). On adaptation of a published
method,23 Cr(η6-C6H6)(CO)3 (10; 52 mg, 0.24 mmol) was stirred in
100 mL of heptane. The solution was filtered, transferred with excess
pyridine into a 100 mL graduated cylinder, and irradiated in the
Rayonet reactor with purging and stirring until red deposits ceased to
form. The solution was vacuum-filtered in air to give 50 mg (78%) of
red crystals after cooling at 260 K overnight.
= ΔH χ − ΔV
1
1
(3)
(4)
Φ
−hνχφ2
= ΔH2 χ − ΔV
2
Φ
Extinction Coefficients. In a typical experiment, a stock solution
of 7 (19.2 mg, 6.93 μmol) was prepared by dilution with 200 mL of
heptane (0.346 mM). Five 10 mL volumetric flasks, containing 0.5−3
g of stock solution, were weighed and diluted to the mark. For 11, the
solutions also contained 0.05 M pyridine. Absorbance measurements
were determined in triplicate. Extinction coefficients were obtained
from the slopes of concentration versus absorbance plots for solutions
of 4−9 and 11 in heptane and are reported in the Supporting
Information.
Actinometer Calibration. A typical sample of the fulgide (15 mg,
83 μM) in 25 mL of anhydrous toluene was kept in the dark. The
same sample cuvette was used for extinction coefficient and quantum
yield determinations. The fulgide solution (2.5 mL) was weighed in
the sample and reference cuvettes. The sample and reference cuvettes
were irradiated for 3−5 min in a split-beam actinometer similar to that
reported previously24 (Figure S11, Supporting Information) with the
PTI laser at 3 pulses/s. The beam was focused to approximately 7 mm
width in the cuvettes. Absorbances at 494 nm were acquired before
and after photolysis. The ratio of photons absorbed in the sample and
the reference cuvettes were determined from the change in the fulgide
absorbances. The fulgide was used with the limitations noted in the
literature.25
Quantum Yields. Sample solutions of 4−6 were prepared in
heptane similarly to those for extinction coefficient determinations.
Absorbances at 337 nm were close to 4 OD. In a typical experiment, a
2.009 g sample of a 2.31 mM solution of 6 and 2.2383 g of a 4.61 mM
solution of the fulgide were weighed in the sample and reference
cuvettes, respectively. The two solutions were irradiated using the
split-beam actinometer for 2.5 min. Absorbances were measured at 337
and 491 nm for 6 and 494 nm for the fulgide before and after
irradiation. Quantum yields were calculated as before.24 The quantum
yields for 10 were obtained with 0.5−1.4 mM 10 and 0.025−0.11 M
pyridine but never less than 50-fold excess pyridine.
demonstrate the dependence on the enthalpy and changes in volume
for each time scale. The assignments of specific reactions or processes
to ϕ1 or ϕ2 are determined from known rates of reaction and the
response time of the transducer. Previous studies suggest that a bond
dissociation results in a volume increase while bond formation results
in a volume decrease; thus, a substitution tends to result in a
cancellation in the volume changes.26 The enthalpy of ligand
substitution (CO dissociation followed by addition of tethered or
untethered pyridine) can be estimated by adding eqs 3 and 4, yielding
eq 5, where the volume terms cancel. If ΔV1 + ΔV2 = 0, then
combining eqs 3 and 4 yields
hν(1 − φ1 − φ2)
= ΔH + ΔH2 = ΔH
1
1+2
(5)
Φ
Computational Methods. Theoretical calculations have been
carried out using the Gaussian0928 implementation of PBEPBE (the
PBE exchange functional and the PBE correlation functional29) density
functional theory,30 using the default pruned fine grids for energies
(75, 302), default pruned course grids for gradients and Hessians (35,
110) (neither grid is pruned for chromium), and nondefault SCF
convergence for geometry optimizations (10−6). Single-point time-
dependent density functional theory calculations iteratively solving for
the first 50 excitations were performed on DFT-optimized geo-
metries.31
The basis set for chromium (341/341/41) was the Hay and Wadt
basis set and effective core potential (ECP) combination
(LANL2DZ)32 as modified by Couty and Hall, where the two
outermost p functions have been replaced by a (41) split of the
optimized chromium 4p function.33 An f polarization function was
added to chromium.34 The 6-31G(d′)35 basis set was used for all other
atoms. The density fitting approximation36 for the fitting of the
Coulomb potential was used for all calculations; auxiliary density-
488
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
fitting basis functions were generated automatically (by the procedure
implemented in Gaussian 09) for the specified AO basis set. Spherical
harmonic d and f functions were used throughout; that is, there are five
angular basis functions per d function and seven angular basis
functions per f function. All structures were fully optimized (except for
optimizations of constrained species), and analytical frequency
calculations were performed on all structures to ensure a zeroth-
order saddle point (a local minimum), first-order saddle point
(transition state TS), or higher-order saddle point was achieved. All
enthalpies and Gibbs free energies reported have been calculated using
standard conditions (298.15 K and 1 atm). The simulated UV−vis
spectra in Figure 5 were produced with an in-house program using
Gaussian broadening and a bandwidth of 35 nm.
RESULTS
■
Synthesis. Good yields (70−97%) of chromium arene
complexes with tethered pyridinyl groups (4−6) were obtained
using Mahaffey and Pauson’s method by refluxing phenyl-
alkylpyridines with Cr(CO)6 in Bu2O/THF.21 The correspond-
ing chelates 7−9, where the tethered pyridine replaces a carbon
monoxide, were obtained by irradiating 4−6, respectively
(Scheme 3). A benzene chromium tricarbonyl derivative with a
bifunctional tether (3) was prepared from the tethered
pyridinyl complex 4 (Scheme 5). Lithiation of 4 followed by
Figure 1. Crystal structures of 7 (red), 8 (yellow), 9 (green), and 11
(black) superimposed with respect to chromium and the carbonyl
ligands.
Scheme 5
alkylation with allyl bromide afforded 3 in good yield (77%).
UV irradiation of 3 formed chelate 1 in excess of chelate 2,
allowing selective crystallization of 1. A mixture of 1 and 2 was
irradiated with visible light to convert 1 to 2, upon which 2 was
crystallized. The preparation of 11 ((η6-benzene)dicarbonyl(η1-
pyridine)chromium) from 10 ((η6-C6H6)Cr(CO)3) was similar
to that for 7−9.23
Effect of Chelate Ring Size on Molecular Structure.
The structures of 1−9 and 11 were confirmed by proton and
carbon NMR spectroscopy23,37 and X-ray crystallography
(Figures 1−3 and Supporting Information, Figures S1−S10).
The solved structure of 7 was nearly Cs symmetric with the
pyridinyl group in an orientation that bisects the two carbonyls.
However, the lowest-energy computed gas-phase geometry for
7 has C1 symmetry, and the Cs-symmetric geometry is a low-
energy transition state that interconverts mirror images of the
two C1-symmetric enantiomers (Figure S50, Supporting
Information). The crystal structures of 7 and 11 are distinctly
different: the Cpyr−N−Cr−Ct dihedral angle (Cpyr = pyridinyl
C2 and Ct = phenyl centroid) of the untethered pyridine is
rotated 81° relative to the pyridinyl group of 7 (Table 1). The
pyridinyl group of 8 is eclipsed with one of the carbonyls, and
the pyridinyl orientation of 6 is similar to that of 11 (Figure 1).
The pyridinyl group of 1, like 7, bisects the carbonyl groups
Figure 2. ORTEP diagram of the structure of 1 in the centric space
group P21/n with 50% probability anisotropic displacement ellipsoids
and two disordered side chain representations.
(Figure 2). NMR spectra of 2 are consistent with two
diastereomers in solution (2:1) where the Re and Si faces of
the pendant propenyl group are coordinated. Only the Si
diastereomer of 2 was observed by X-ray crystallography
(Figure 3).
Computed Structures for Conformers and Diaster-
eomers of Chelates 1 and 2. Solutions of 1 or 2 in benzene
in the absence of light show no evidence of decay or
isomerization for 24 h.2 The free energy of activation required
for the thermal isomerization of 1 to 2 that accounted for
photochrome bistability was investigated computationally.
Local energy minima were found for the structures representing
1 and 2 in Scheme 6 as well as for the related conformers 12−
14. In the search for the lowest energy pathway connecting the
local minima 1 and 12, the structure of TS-1-12 (transition
state between 1 and 12) was found. Structures for other lowest
energy pathways that connect other pairs of minima in Scheme
6 likewise were found. The computed free energy of activation
489
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
investigated a structural modification to this system. When the
two methylene hydrogens of the propenyl arm were replaced
with methyl groups, the free energy of activation of the
transition state analogous to TS-13-14 was raised by 1.1 kcal
mol−1.
A computational search for stationary points between the Si
(2) and the Re (16) diastereomers revealed two transition
states and a single minimum (Scheme 7). 15 is a shallow local
Scheme 7
Figure 3. ORTEP diagram of the structure of 2 with 50% probability
anisotropic displacement ellipsoids.
Table 1. Comparison of Selected Experimental and
Computational Bond Lengths and Angles
dihedral angle
N−Cr−Ct angle
a
Cr−N length (Å)
(deg)
(deg)
minimum with an asymmetric potential well (0.8−4.6 kcal
mol−1). The computed free energy of activation for the
isomerization via TS-15-16 is 22.0 kcal mol−1.
b
exptl
comp
exptl
comp
exptl
comp
rms (Å)
7
2.108
2.133
2.155
2.105
2.089
2.111
2.137
2.079
1.2
57.7
72.7
82.6
16.3
57.7
71.6
83.9
116.8
121.0
122.3
121.0
118.1
122.1
123.1
123.7
0.3076
0.0808
0.1012
0.3298
8
Quantum Yields and Enthalpies for Chelation Follow-
ing Metal Carbonyl Dissociation. Irradiation of chromium
arene tricarbonyls 4−6 leads to CO dissociation followed by
addition of a tethered pyridinyl group to the metal. The
quantum yields for formation of 7−9 were found to be
independent of chain length (within experimental error) and
were the same as that for CO substitution by pyridine after
irradiation of 10 (Table 2).
9
11
a
b
Cpyr−N−Cr−Ct pyridine−centroid dihedral angle. Root-mean-
square deviation of the heavy atoms of the experimental and
computational structures.
for the linkage isomerization via TS-13-14 is 27.9 kcal mol−1.
Recognizing the congested nature of this interchange transition
state (TS-13-14) and the steric effects in the approach of the
interchanging pyridinyl and propenyl functional groups, we
PAC experiments for 10 with 12−124 mM pyridine revealed
two sequential pressure decays. The first decay had an
amplitude (ϕ1 measured as a fraction of a reference signal)
Scheme 6
490
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
structures for 7−9 with 35 nm line broadening to simulate the
line broadening effects of vibrations and rotations (Figure 5 and
Table 2. Comparison of Photochelation Yields, PAC
Amplitudes, and Substitution Enthalpies (kcal mol−1) for
Chromium Benzene Tricarbonyl Derivatives at 297 K
a
complex
4
5
6
10
b
c
Φ337
ϕ1
0.77(1)
0.92(3)
0
0.76(1)
0.92(5)
0
0.76(2)
0.87(2)
0
0.79(2)
0.77(3)
0.12(2)
c
ϕ2
d
e
ΔH1+2
ΔH1
ΔH2
8.8(36)
57.6
8.8(57)
57.3
14.3(25)
53.2
12.1(36)
56.5
comp
comp
−28.4
29.2
−27.5
29.8
−20.7
32.4
−30.5
26.1
comp
ΔH1+2
a
b
Uncertainties are 1σ expressed as error in last digit(s). See also ref
c
d
38. [pyridine] = 12−124 mM, k2 = [5.3(3)] × 108 M−1 s−1. In kcal
mol−1, calculations based on Φ337 = 0.77 and neglecting ΔV, see
Discussion. Relative errors (neglects error in Φ) for 4−6 are 3.3, 5.5,
e
and 2.2 kcal mol−1, respectively. ΔH1 = 25.3 kcal mol−1 and ΔH2 =
Figure 5. Simulated UV−vis spectra and computed transitions of 7−9.
Table 3. Comparison of Computed Wavelengths (λcomp, nm)
−13.2 kcal mol−1 assuming ΔV1 = ΔV2 = 0.
that was independent of the pyridine concentration and did not
undergo a phase shift relative to the reference signal. This decay
is assigned to processes faster than the transducer response
time (<100 ns). A second decay had an amplitude (ϕ2) with a
phase shift and therefore a lifetime that was dependent on the
pyridine concentration, and the amplitude was different from
those produced by other dispersed Lewis bases (tetrahydrofur-
an, triethylsilane, acetonitrile, and triethyl phosphite).39 In
contrast, only a single heat decay without phase shift was
observed for 4−6 (i.e., ϕ2 = 0). Within experimental error, the
sum of the signal amplitudes (ϕ1 + ϕ2) was independent of
chain length (4−6) or the lack of pyridinyl tethering (11)
(Table 2).
of Transition Maxima as a Function of Constrained Cpyr
N−Cr−Ct Dihedral Angles
−
a
chelate
11
17
13°
557 nm
60°
529 nm
73°
536 nm
(85°)
18
12°
563 nm
59°
523 nm
72°
521 nm
(83°)
19
12°
565 nm
59°
517 nm
72°
517 nm
(81°)
7 (16°)
535 nm
8 (58°)
560 nm
9 (72°)
555 nm
14°
555 nm
59°
522 nm
73°
525 nm
(86°)
Effect of Tether Length/Pyridine Orientation on
Electronic Spectra. There are small differences in intensity
and wavelength of local maxima (λmax) in the two regions of
350 and 500 nm (Figure 4)40 as the orientation of the
502 nm
505 nm
530 nm
534 nm
a
Maxima between 400 and 600 nm are reported. DFT geometry
optimized, the angles in parentheses are unconstrained Cpyr−N−Cr−
Ct.
Table 3). For comparison, vertical markers locate the individual
computed electronic transitions. The simulated UV−vis spectra
for 7−9 have peaks in the 400−600 nm region42 where the λmax
value occurs at 527 nm for 7 but is red-shifted 27 nm to 554
nm for 8 and then blue-shifted 17 nm back to 537 nm for 9.
TDDFT computations for 11 were obtained for a range of
constrained Cpyr−N−Cr−Ct dihedral angles from 0 to 88°.43
The maximum computed electronic transitions in the 400−600
nm region of the spectrum blue-shifts 53 nm (555 to 502 nm,
respectively, Table 3). Furthermore, the composition of this
peak changes from a predominantly HOMO → LUMO
transition (0 to 45°) to a transition with a mixture of
components (45 to 88°).44 A Fenske−Hall molecular orbital
analysis45 of 11 shows the orientation of the HOMO of the
arene−Cr−dicarbonyl fragment (comprised primarily of Cr d
orbitals) with the LUMO of the pyridine fragment (comprised
primarily of pyridine π* orbitals; see Figure S51, Supporting
Information).
Figure 4. Experimental UV−vis spectra of 7−9 in heptane.
coordinated pyridine for 7−9 varies by over 80° (Figure 1 and
Table 1). One notable difference is a weak band near 350 nm in
the spectra of 8 and 9 that appears to decrease and move to 334
nm for 7. A second difference is the red shift from the λmax
value of 498 nm for 7 to 507 nm for 8 and a blue shift to 490
nm for 9.41
Using time-dependent density functional theory (TDDFT),
simulated electronic spectra were derived from the computed
electronic transitions and oscillator strengths of optimized
Simulated spectra were obtained from computed transitions
for three methylated derivatives of 11: 17 (a toluene
derivative),46 18 (a 2-methylpyridine derivative),47 and 19
(toluene and 2-methylpyridine derivative) (see Scheme 8).48
For comparison, the wavelength of computed transition
maximum (λcomp) increases (25 nm) and then decreases (5
nm) upon increasing the Cpyr−N−Cr−Ct dihedral angle to 58°
491
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
1911 cm−1 with a subtle shoulder near 1903 cm−1, while only
two peaks are observed for 6 at 1978 and 1909 cm−1 without a
shoulder for the latter. However, the apparent half-width of the
peak at 1909 cm−1 is much larger than that for the peak at 1978
cm−1. Thus, the separation of the two low-frequency peaks
decreases as the chain length increases for 4−6. After
irradiation of 4−6, in each case two new peaks assigned to
7−9 respectively appear at lower frequencies as the chain
length increases (Figure 7 and Table 5).
Scheme 8
(7 to 8) and then to 72° (7 to 9) (Table 3).49 In contrast, the
λcomp values of untethered analogues 11 and 17−19 all blue-
shift 30−50 nm over a similar dihedral angle range. The λcomp
values for 11 and 17−19 at dihedral angles comparable to that
for 7 (16°) are red-shifted 20−30 nm. On the other hand, for
dihedral angles comparable to that for 9 (72°), the transitions
for 11 and 17−19 are blue-shifted 19−38 nm.
Infrared Spectra. Depending on the pyridinyl tether chain
length, the tricarbonyl derivatives 4−6 exhibit two to three
peaks in the metal carbonyl region of infrared spectra. Peaks for
4 are observed at 1978 and 1912 with a shoulder at 1898 cm−1,
(Figure 6, Table 4). Two peaks are observed for 5 at 1978 and
Figure 7. FTIR difference spectra before and after UV irradiation of
4−6 in heptane (0.82, 1.1, and 0.96 mM, respectively).
Table 5. CO Stretching Frequencies of Chromium Benzene
Dicarbonyl Derivatives in Heptane
CO stretching frequency (cm−1
)
a
exptl
comp
sym
complex
sym
asym
1852
asym
1906
[Cr{η6-C6H5CH2(2-C5H4N)-κN}
(CO)2] (7)
1907
1941
1932
1929
1941
1959
[Cr{η6-C6H5(CH2)2(2-C5H4N)-κN}
(CO)2] (8)
1900
1898
1908
1935
1902
1844
1842
1852
1880
1850
1897
1896
1906
1924
[Cr{η6-C6H5(CH2)3(2-C5H4N)-κN}
(CO)2] (9)
[Cr{η6-C6H5(CH(2-C5H4N)-κN-
b
CH2CHCH2)}(CO)2] (1)
Figure 6. FTIR spectra of 4−6 in heptane (0.82, 1.1, and 0.96 mM,
respectively).
[Cr{η6-C6H5(CH(2-C5H4N)CH2-η2-
b
CHCH2)}(CO)2] (2)
(η6-C6H6)Cr(CO)2(pyridine) (11)
1937
1931
1905
1896
Table 4. CO Stretching Frequencies of Chromium Benzene
Tricarbonyl Derivatives in Heptane
(η6-C6H6)Cr(CO)2(2-methylpyridine)
(18)
a
b
CO stretching frequency (cm−1
)
Gas-phase computed harmonic frequencies. Reference 2.
a
exptl
comp
Increasing the length of the pyridinyl tether could inductively
complex
sym
asym
sym
asym
change the electron density at the metal center and therefore
alter the metal CO stretching frequencies. This relationship was
investigated by comparing the computed metal CO frequencies
of 11 with those for 18, where a methyl group is added to the
pyridine group of 11 (Table 5).52 There is a red shift in the
experimental and computed metal carbonyl frequencies as the
length of the pyridine tether increases; the frequencies for the
“untethered” 11 are between those of 7 and 8, and those of the
“untethered” 18 are between those of 8 and 9.
Computed values for bond distances (Cr−N, Cr−Ct, and
CO) and CO stretching frequencies were determined for 11 at
19 different dihedral angles from 0 to 88° and for 18 (which has
a dihedral angle of 83°). Representative data are summarized in
Table 6.53 As the dihedral angle increases from 0 to 83° for 11,
the Cr−N bond and Cr−Ct lengths decrease while the CO
bond length is essentially unchanged. The CO stretching
[Cr{η6-C6H5CH2(2-C5H4N)}(CO)3] 1978 1912,
1996 1948,
1939
(4)
1898
[Cr{η6-C6H5(CH2)2(2-C5H4N)}
(CO)3] (5)
[Cr{η6-C6H5(CH2)3(2-C5H4N)}
(CO)3] (6)
1978 1911
1978 1909
1977 1912
1996 1947,
1942
1995 1946,
1938
[Cr{η6-C6H5(CH(2-C5H4N)
CH2CHCH2)}(CO)3] (3)
1995 1948,
1936
b
(η6-C6H6)Cr(CO)3 (10)
1983 1914
1977 1906
2004 1952
n.d. n.d.
(η6-C6H5(CH2)2CHCH2)
c
Cr(CO)3
a
b
Gas-phase harmonic frequencies. Literature comparison: 1983/1916
cm−1 in heptane and 1983/1915 cm−1 in cyclohexane.50,51 In
c
cyclohexane.11
492
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
most similar to that of 9, the Cr−N distance of 11 is most
similar to that for 7 (Table 1). Computed Cr−N bond lengths
and dihedral angles reveal a similar trend. We conclude the
longer Cr−N bond length of 9 indicates a weaker bond caused
by greater strain associated with the larger chelate ring.
Table 6. Computed Bond Lengths (Cr−N, Cr−Ct, and CO)
and CO Stretching Frequencies for 11 at Representative
Cpyr−N−Cr−Ct Dihedral Angles in Comparison to Those of
a
18
dihedral
angle (deg)
Cr−N bond
Cr−Ct
C−O bond
νCO
A comparison of the heavy-atom positions for the X-ray
crystal structure of 7 (nearly Cs symmetric) with those of the
computed Cs-symmetric structure (Figure S50, Supporting
Information) gives an RMSD (0.054 Å) that is lower than the
RMSD of the same crystal structure with the computed lowest
energy C1-symmetric structure (0.308 Å). The C1-symmetric
structure of 7 is computed to be 1.3 kcal mol−1 lower in energy
than the Cs-symmetric structure, suggesting that a C1-symmetry
structure should be observed (Figure S50, Supporting
Information). The small energy difference between the C1-
and Cs-symmetric structures could be altered by packing
forces.55 Upon examination of the Fenske−Hall molecular
orbital analysis of 11, a decrease in the dihedral angle about the
Cr−N axis reduces the overlap of these orbitals, leading to a
decrease in back-bonding from the metal center to the pyridine
π* and therefore a longer Cr−N bond length (see Table S63,
Supporting Information).
length (Å)
length (Å)
length (Å)
(cm−1
)
11
11
11
18
0
40
83
83
2.138
2.102
2.073
2.127
1.691
1.694
1.679
1.685
1.179, 1.179 1935,
1901
1.177, 1.181 1938,
1898
1.177, 1.178 1939,
1906
1.178, 1.180 1931,
1896
a
For additional data see ref 52.
frequencies decrease at intermediate dihedral angles and then
become larger at 83° than they were initially at 0°. A
comparison of 11 and 18 at a dihedral angle of 83° shows
greater bond lengths and lower CO stretching frequencies for
18.
DISCUSSION
Enthalpies of Substitution. The enthalpy of CO
substitution by tethered and untethered pyridine was examined
by photoacoustic calorimetry and computational methods.
Since previous studies of 4−6 demonstrate that ring formation
following irradiation is faster than the time response of our 1
MHz transducer (ca. 100 ns),2 both substitution steps (CO
dissociation and pyridinyl addition) should contribute to ϕ1.
Consistent with these results, a single component (ϕ1) with no
phase shift was detected in the PAC signal following irradiation
of 4−6; therefore, there is no “slow” PAC component (ϕ2).
The amplitudes of the signals in each case were larger than the
ϕ1 value observed for 10 (Table 2), where pyridine addition at
low concentrations contributes to only ϕ2 and not ϕ1. These
observations indicate pyridine addition to the metal is much
slower than CO dissociation and occurs on a time scale that
produces a slow component (ϕ2) of the signal. The ΔH1+2
value in Table 2 for the substitution of CO by a pyridinyl group
increases with the tether chain length, although the results are
within experimental errors.56 The computed enthalpies for
ΔH1+2 likewise show an increase with chain length and illustrate
that the effect is primarily a decrease in the Cr−N bond energy.
We conclude that increasing the tether length introduces very
little strain, at least in the case of a tethered pyridinyl group.
The experimental results for ΔH1+2 are about 20 kcal mol−1
less endothermic than the gas-phase computational results.
Examination of Table 2 for 10 indicates the difference can be
attributed to differences in the determination of enthalpy of CO
dissociation. The ΔH1comp and ΔH2comp values are equivalent to
gas-phase bond dissociation energies and cannot be directly
compared with ΔH1 and ΔH2 from PAC determinations, where
coordination of the solvent is particularly important: heptane
displaces CO in the first step and is displaced by pyridine in the
second step. Previous results indicate that the metal−heptane
bond should be about 12−15 kcal mol−1.26 For the moment,
assume ΔV1 = ΔV2 = 0 for eqs 3 and 4; then the PAC
determination leads to ΔH1 = 25.3 kcal mol−1 and ΔH2 =
−13.2 kcal mol−1 for 10. Accounting for a 15 kcal mol−1
contribution of bonding to heptane results in bond dissociation
energies in solution of 40 kcal mol−1 for Cr-CO and 28 kcal
■
The chain length of a functional group tethered to a metal can
potentially affect the rate and yield of chelate formation, metal−
functional group bond energies, bond lengths, and orientation.
We previously reported that irradiation of 5 and 6 in heptane
initially produces only heptane-coordinated intermediates
(<100 ps) which convert to the corresponding chelates 8 and
9 in 100−200 ns.2 On the other hand, within 100 ps of
irradiation 4 forms both the N-chelate 7 and an agostic chelate
in competition with a heptane-coordinated intermediate. The
heptane-coordinated intermediate and the agostic complex
subsequently convert to the N-chelate 7 within 50 ns. In spite
of these mechanistic differences, we find in the current study
that the quantum yield for N-chelate formation is independent
of chain lengths for 4−6 and is the same as that for the
photosubstitution of (η6-C6H6)Cr(CO)3 with a dispersed Lewis
base (Table 2). Together, these results indicate that the
product-determining step precedes chelate formation (<100 ps
for 4). These results are in contrast to those for (η5-
C5H4R)Mn(CO)3, where the quantum yields are found to
depend on the side chain structure (R) and in some cases unit
quantum yields are observed.8
Some combination of two mechanisms may account for the
quantum yield independence of chain length. First, the
quantum yield is determined by the partitioning of the excited
states between CO dissociation (0.77) and relaxation back
(0.23) to the ground state (no luminescence has been observed
for chromium arene tricarbonyls). Second, the yield for CO
dissociation is higher than 0.77, but its recombination is
unfettered by a tethered pyridinyl group which behaves
kinetically no differently than the solvent as a barrier to cage
escape. The first mechanism is consistent with the conclusions
of recent time-resolved studies which indicate that the yield of
product is determined by the population distribution of excited
states.54
Orientation of Coordinated Pyridine. The crystal
structures of 7−9 shown in Figure 1 demonstrate the effects
of chelate ring size: the phenyl ring progressively rotates
relative to the Cr−N bond, and the Cpyr−N−Cr−Ct dihedral
angle increases approaching the conformation of 11 which has a
dihedral angle near 90°. Although the dihedral angle of 11 is
comp
mol−1 for Cr−N. A comparison of these values with ΔH1
and ΔH2comp in Table 2 suggests the difference between ΔH1+2
493
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
and ΔH1+2comp can be attributed largely to the CO dissociation,
suggesting that the contribution of heptane coordination does
not account for the discrepancies. If the assumption that ΔV is
negligible is incorrect, the magnitude of reaction enthalpies and
bond energies might be underestimated from PAC determi-
nations. By using eq 557 and the difference between ΔH1+2 and
ΔH1+2comp of 14 kcal mol−1 (Table 2), 48 mL mol−1 is obtained
for ΔV1+2 for the pyridine substitution with 10. Typically the
breaking of a single bond affords a volume change of 10 mL
mol−1;58 and for comparison, the molecular volume of pyridine
is only 81 mL mol−1. Such a large reaction volume is
group. The increase in bond length for 7−9 also correlates with
comp
a decrease in bond energy reflected in both ΔH1+2 and ΔH1
(Table 2).63 All of these results point to the counterintuitive
conclusion that the more electron donating pyridinyl groups
have weaker Cr−N bonds, and it does not appear to be due to a
strain effect!64 On the other hand, we believe the decrease in
CO stretching frequencies observed for 7−9 in Table 5 is
dominated by changes in orbital overlap as the dihedral angle
increases. Therefore, back-bonding increases as the dihedral
angle increases, resulting in the decrease in CO stretching
frequencies. This result is consistent with an increase in CO
stretching frequency observed for 11, where electron density
has been removed relative to 9, and return to a lower CO
stretching frequency for 18, where electron density has been
restored (Table 5).
The trend in the experimental λmax values for 7−9 is faithfully
reflected in the computed maximum transition (Figures 4 and 5
and Table 3). The lack of a consistent increase in wavelength
with methyl substitution on 11 for a constrained dihedral angle
(for example, 11, 17, 18, and 19 constrained at 59, 60, 59, and
59°, respectively; see Table 3) indicates that electron density is
not a significant factor for the λmax changes in 7−9. There is a
significant decrease in wavelength with increase in dihedral
angle for all untethered analogues (for example, 17 at 13, 60,
and 73° in Table 3), but this trend is the opposite direction
observed for 7−9. The only physical parameter that correlates
with the optical changes for 7−9 is the N−Cr−Ct angle (Table
1), which suggests either a decrease in ground state and excited
state energy differences or more likely subtle changes in the
transition contributions (e.g., less HOMO−LUMO and more
HOMO-1−LUMO in 7 versus 5).
Lowest-Energy Thermal Isomerization Pathways.
Previously we prepared manganese cyclopentadienyl complexes
with two functional groups tethered to the cyclopentadienyl
ring where one group is chelated to the metal.1b In one
example, an irreversible thermal linkage isomerization between
a side-chain carbonyl group and a pyridinyl group occurs in less
than 5 min. In another example, a thermal linkage isomerization
occurs irreversibly between a pyridinyl group and a thioamide
group: this result is in spite of the fact that, when the thioamide
group is replaced by a nitrile group, the pyridinyl chelate is
stable indefinitely. Clearly, the intrinsic manganese−pyridinyl
bond strength is unchanged, but the tether provides a low-
energy pathway for linkage isomerization, which is observed
only when the less stable isomer can be prepared rapidly. This
isomerization cannot involve a simple Mn−N bond dissocia-
tion, because the bond strength is too large to break in a
transition state where the activation energy corresponds to a
few minutes. We concluded that the metal either walks along
the tether between the two functional groups or there is a
concerted substitution of the functional groups, thus avoiding
the barrier of a complete dissociation of the tethered functional
groups from the metal. It is therefore imperative that we
understand potential mechanisms that may result in a thermal
linkage isomerization if we wish to prepare bistable photo-
chromes based on a photolytic linkage isomerization.
unprecedented and unrealistic. Likewise, the difference between
comp
ΔH1 and ΔH1
of 16.5 kcal mol−1 (includes 15 kcal mol−1
correction for bond to heptane; vide supra) requires a volume
of 56 mL mol−1 for CO substitution by heptane to account for
this difference. This volume also seems unreasonably large, and
neglect of reaction volume does not satisfactorily explain the
difference between the computational value of 56.5 kcal mol−1
(ΔH1comp) and the PAC value of 40 kcal mol−1 (ΔH1 + heptane
bond energy). These calculations confirm that the discrepancy
between the computational and PAC results is largely in the
first step. The thermal contribution of solvation (non-
coordinating) of reactants and products has not been
reconciled, but these contributions typically cancel each other
in hydrocarbon solutions.59
Infrared Spectra for Tricarbonyl Compounds. Com-
puted values of metal CO frequencies (Tables 4 and 5) are
universally greater than the experimental results, a difference
that is attributable to the neglect of anharmonicity.60
Nevertheless, the variation of computed vibrational frequencies
with structure is entirely consistent with the experimental
results.61 While the symmetrical stretches for 4−6 are
independent of chain length (1978 cm−1), the asymmetric
stretches are essentially degenerate for 6 but gradually separate
and are most distinct for 4 (Figure 6). A loss of asymmetric
stretch degeneracy was observed previously for benzene-
substituted derivatives of 1062 and more recently for (η5-
C5H5R)Mn(CO)3.1b Due to the insulating nature of the side
chain, we believe that this loss of degeneracy is a through-space
electronic interaction. The shift to lower frequencies when CO
is substituted by pyridinyl or propenyl functional groups is
consistent with a greater electron donation to the metal center
and concomitant back-bonding to the metal carbonyls.
Tether Length Effect on Chelate Bond Lengths, Bond
Strengths, and IR and Electronic Spectra. The Cpyr−N−
Cr−Ct dihedral angles increase monotonically with the tether
length, but bond lengths, bond strengths, and spectra generally
do not display trends simply related to the dihedral angle. The
Cr−N bond length decreases as the constrained dihedral
increases for 11 (Table 6). This result suggests that the Cr−N
bond strength increases with the dihedral angle. In contrast, the
bond length increases for tethered pyridinyl groups as the
dihedral angle increases (Table 1). The trend reversal is not an
artifact of computed results, since the trend for 7−9 is the same
for both experimental and computed bond lengths. We
conclude the trend for 7−9 is not a function of dihedral
angle but of chain length. The longer chain should have an
electron-donating effect, which is modeled by 18 and 11 at 83°
(Table 6), where there is no ring strain. The addition of the
methyl group increases the bond length by 0.054 Å. Adding this
value to the computed bond length of 7 (2.089 Å) yields 2.143
Å, which is in reasonable agreement with 2.137 Å for 9,
considering the crude approximation of simply adding a methyl
In the current study, the computational investigation of the
linkage isomerization of 1 to 2 (Scheme 6) indicates that the
lowest free energy of activation to isomerization is 28 kcal
mol−1. This high free energy of activation is consistent with the
experimental observation that 1 or 2 do not isomerize to each
other at room temperature except during irradiation.2 The
computed energies for conformers of 1 and 2 (Scheme 6)
494
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
suggest the lowest energy isomerization proceeds via four steps
in a pathway that includes a dissociative interchange where
there is a rotation about the bond between the phenyl group
and the adjacent chiral carbon. In the transition state TS-13-14
(Scheme 6), the metal interacts with both functional groups
simultaneously (Cr−N is 3.228 Å and Cr−vinyl centroid is
3.481 Å), and the metal does not bond to any part of the bridge
between the functional groups. Thus, the high free energy of
activation is due to the nearly complete breaking of the bonds
to the functional groups forced by the steric requirements of
the vinyl group: the π electrons do not extend as far from the
nucleus as those of lone pairs and demand uninhibited space
around the metal, because the four groups attached to the π
bond are coplanar. Thus, the approach of the π bond to the
metal is expected to be sterically inhibited for an interchange.
Indeed, the computed transition state is 1.1 kcal mol−1 higher
when the two methylene hydrogens of the propenyl arm are
replaced with methyl groups. It is instructive to compare the
current mechanism for linkage isomerization with those of two
manganese photochromes.1b Notably, neither photochrome is
bistable, and the tethering of the functional groups is distinctly
different from that for the 1/3 photochrome. The metal center
of one photochrome can “walk” along π bonds between the
tether functional groups, providing a low-energy nondissocia-
tive pathway. In contrast, the tether of the other photochrome
forces the uncoordinated function group so close to the metal
that it promotes a low-energy associative interchange
mechanism. Thus, bistable photochromes based on a linkage
isomerization requires the positioning of functional groups that
allows photoisomerization without thermal isomerization.
The experimental NMR spectrum of the propenyl chelate is
consistent with two diastereomers in solution (2 and 16,
Scheme 7; see Figures S60−S63, Supporting Information).
Given the anisotropic nature of aromatic solvents, the
assignments of these diastereomers are in reasonable agreement
with computed NMR spectra (see Figure S49 and Table S70,
Supporting Information). However, a comparison of the
computed and experimental NMR spectra suggests that the
crystal structure likely corresponds to the isomer in lower initial
yield. It is interesting that the computed structure of 9
corresponds to the crystal structure, and its lower energy
(Scheme 7) may contribute to the fact that it was the only
isomer crystallized. The 22.0 kcal mol−1 free energy of
activation is low enough to allow isomerization of the
diastereomers at room temperature over time but high enough
that this isomerization is slow enough to allow the observation
of both diastereomers by NMR spectroscopy. These computa-
tional results are completely consistent with our experimental
results and provide a strategy for creating bistable photo-
chromes by designing high-energy transition states.
The dihedral angle defined by the phenyl centroid, chromium,
pyridinyl nitrogen, and the pyridinyl C2 changes by 80° as the
chain lengthens. This change is accompanied by an
experimental and computed decrease in CO bond stretching
frequencies consistent with an increase in electron density on
the metal center, attributed to a change in ligand−metal orbital
overlap. The computed and experimental Cr−N bond distances
for the chelates 7−9 increase with chain length while the
corresponding bond energies decrease. The effect is attributed
to ring strain. Previous time-resolved studies of this series of
chromium arene tricarbonyl complexes with tethered functional
groups in heptane revealed that photosubstitution of CO by a
pendant group occurs within 200 ns, supporting the kinetics
observed for current PAC studies. In each case for 3 and 4−6, a
heptane-coordinated intermediate is formed before the chelate
formation is complete. Only in the case of 3 and 4 did chelation
occur directly in competition with heptane coordination.
Studies are underway to further characterize the reaction
coordinate and dynamics of chelate ring inversion for
chromium and manganese complexes.
The results indicate that there is little reason to believe the
photochrome properties of 1/2 would materially improve if the
pyridinyl group had a longer chain. There is only a small change
in the chromophore absorbance or λmax and metal−nitrogen
bond energies, and there is no measurable difference in
quantum yields for chelation. There is a notable change in the
pyridinyl group position (pyridinyl dihedral angle versus Cr−
phenyl centroid), and this change could have an effect on the
trajectory of pyridinyl dissociation, recombination, and
efficiency of replacement by its companion functional group.
The best way to investigate this phenomenon is by time-
resolved spectroscopy and quantum yields. These studies are
limited by the solubility of the chelate in alkane solvents. The
dynamics and chemistry are likely to be different in polar
solvents, and we are avoiding this complication by making
alkane-soluble analogues. An ultimate goal is the design of a
linkage isomerization that is so fast that interactions with a
solvent or matrix are negligible during ring formation. We have
reported some success in this regard.65
ASSOCIATED CONTENT
* Supporting Information
■
S
Crystallographic data, experimental UV−vis and NMR spectra,
actinometry apparatus and data. Computational results: UV−
vis, IR, and NMR transitions; simulated UV−vis and NMR
spectra; and higher energy pathways. Complete listing of atom
coordinates of all optimized structures. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
SUMMARY AND CONCLUSIONS
■
*E-mail: tburkey@memphis.edu (T.J.B.); cewebstr@memphis.
edu (C.E.W.).
This study is the first report of the synthesis of chromium arene
tricarbonyl complexes with tethered pyridinyl groups and the
corresponding dicarbonyl chelates. The effect of chain length
was examined for parent tricarbonyls as well as the pyridinyl
chelates. Quantum yields for chelate formation from the parent
tricarbonyl were independent of chain length and the
mechanism of chelation. The equivalent yields indicate that
the product-determining step precedes chelation, where
chelates with longer tethers (5 and 6) came exclusively from
solvent-coordinated intermediates, while the chelate with a
short tether (4) was formed from three different intermediates.2
Notes
The authors declare no competing financial interest.
§Deceased.
ACKNOWLEDGMENTS
■
This work was partially supported by the National Science
Foundation under Grant Nos. CHE-0227475 (T.J.B.) and
CHE-0911528 (T.J.B. and C.E.W.). Computational work was
performed on resources at the University of Memphis High-
495
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
(21) Mahaffy, C.; Pauson, P. L. Inorg. Synth. 1990, 28, 136−140.
(22) The elemental analysis for some complexes yields poor results
even with resubmitted and further purified samples. While structures
were established by single-crystal X-ray diffraction, the estimates of the
purity of these samples along with NMR spectra are reported in the
Supporting Information.
(23) Spectral properties for 11 are consistent with published results.
See for leading references on spectral properties: Wrighton, M. S.;
Haverty, J. L. Z. Naturforsch. 1975, 30B, 254−258.
(24) Pang, Z.; Burkey, T. J.; Johnston, R. F. Organometallics 1997, 16,
120−123.
(25) (a) Uhlmann, E.; Gauglitz, G. J. Photochem. Photobiol. A. Chem.
1996, 98, 45−49. (b) Guo, Z.; Wang, G.; Tang, Y.; Song, X. J.
Photochem. Photobiol. A 1995, 88, 31−34. (c) Boule, P.; Pilichowski, J.
F. J. Photochem. Photobiol. A: Chem. 1993, 71, 51−53.
(26) Gittermann, S. M.; Letterman, R. G.; Jiao, T.-J.; Leu, G.-L.;
DeYonker, N. J.; Webster, C. E.; Burkey, T. J. J. Phys. Chem. A 2011,
115, 9004−9013.
(27) Rudzki, J.; Libertini, L. J.; Small, E. W. Biophys. Chem. 1992, 42,
29−48.
Performance Computing Facility. Additional support of this
research by the Cancer Center Support CORE Grant, P30 CA-
21765 and the American Lebanese Syrian Associated Charities
(ALSAC) is gratefully acknowledged by C.R.R.
REFERENCES
■
(1) (a) To, T. T.; Barnes, C. E.; Burkey, T. J. Organometallics 2004,
23, 2708−2714. (b) To, T. T.; Duke, C. B., III; Junker, C. S.; O’Brien,
C.; Ross, C. R., II; Barnes, C. E.; Webster, C. E.; Burkey, T. J.
Organometallics 2008, 27, 289−296. (c) Heilweil, E. J.; Johnson, J. O.;
Mosley, K. L.; Lubet, P. P.; Webster, C. E.; Burkey, T. J.
Organometallics 2011, 30, 5611−5619. (d) Kelbysheva, E. S.;
Ezernitskaya, M. G.; Strelkova, T. V.; Borisov, Y. A.; Smol’yakov, A.
F.; Starikova, Z. A.; Dolgushin, F. M.; Rodionov, A. N.; Lokshin, B. V.;
Loim, N. M. Organometallics 2011, 30, 4342−4353.
(2) To, T. T.; Heilweil, E. J.; Duke, C. B., III; Ruddick, K. R.;
Webster, C. E.; Burkey, T. J. J. Phys. Chem. A 2009, 113, 2666−2676.
(3) Akita, M. Organometallics 2011, 30, 43−51.
(4) Trahanovsky, W. S.; Hall, R. A. J. Organomet. Chem. 1975, 96,
71−82.
(5) Donnini, G. P.; Shaver, A. Can. J. Chem. 1978, 56, 1477−1481.
(6) Nesmeyanov, A. N.; Krivykh, V. V.; Petrovskii, P. V.; Kaganovich,
V. S.; Rybinskaya, M. L. J. Organomet. Chem. 1978, 162, 323−342.
(7) (a) Nesmeyanov, A. N.; Krivykh, V. V.; Panosyan, G. A.;
Petrovskii, P. V.; Rybinskaya, M. I. J. Organomet. Chem. 1979, 164,
167−175. (b) Nesmeyanov, A. N.; Krivykh, V. V.; Rybinskaya, M. I. J.
Organomet. Chem. 1979, 164, 159−165.
(8) Jiao, T.; Pang, Z.; Burkey, T. J.; Johnston, R. F.; Heimer, T. A.;
Kleiman, V. D.; Heilweil, E. J. J. Am. Chem. Soc. 1999, 121, 4618−
4624.
(9) (a) Perutz, R. T.; Turner, J. J. J. Am. Chem. Soc. 1975, 97, 4791−
4800. (b) Schwartz, B. J.; King, J. C.; Zhang, J. K.; Harris, C. B. Chem.
Phys. Lett. 1993, 203, 503−508. (c) Lian, T.; Bromberg, S. E.; Asplund,
M. C.; Yang, H.; Harris, C. B. J. Phys. Chem. 1996, 100, 11994−12001.
(10) Giordano, P. J.; Wrighton, M. S. Inorg. Chem. 1977, 16, 160−
166. See also ref 8.
(28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
Revision B.01; Gaussian, Inc., Wallingford, CT, 2009.
(29) (a) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996,
77, 3865−3868. (b) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev.
Lett. 1997, 78, 1396.
(30) Parr, R. G.; Yang, W. Density Functional Theory of Atoms and
Molecules; Oxford University Press: New York, 1989.
(31) Marques, M. A. L.; Gross, E. K. U. Annu. Rev. Phys. Chem. 2004,
55, 427−455.
(32) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652.
(33) Couty, M.; Hall, M. B. J. Comput. Chem. 1996, 17, 1359−1370.
(34) Ehlers, A. W.; Bohme, M.; Dapprich, S.; Gobbi, A.; Hollwarth,
A.; Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G.
Chem. Phys. Lett. 1993, 208, 111−114.
(11) Tethered olefins for chromium arene tricarbonyl derivatives and
subsequent chelation have been reported previously: Nesmeyanov, A.
N.; Rybinskaya, M. I.; Krivykh, V. V.; Kaganovich, V. S. J. Organomet.
Chem. 1975, 93, C8−C10. See also ref 6.
1
(12) 7.16 and 128.06 ppm for H and 13C in benzene-d5, 3.58 and
67.21 ppm for 1H and 13C in THF-d7: Fulmer, G. R.; Miller, A. J. M.;
Sherden, N. H.; Gottlieb, H. E.; Nudelman, A.; Stoltz, B. M.; Bercaw, J.
E.; Goldberg, K. I. Organometallics 2010, 29, 2176−2179.
(13) To, T. T.; Duke, C. B., III; Junker, C. S.; O’Brien, C.; Ross, C.
R., II; Barnes, C. E.; Webster, C. E.; Burkey, T. J. Organometallics
2008, 27, 289−296.
(35) The 6-31G(d′) basis sets have the exponent for the d
polarization functions for C, N, and O taken from the 6-311G(d) basis
sets, instead of the original arbitrarily assigned values of 0.8 used in the
6-31G(d) basis sets: (a) Hariharan, P. C.; Pople, J. A. Theor. Chim
Acta 1973, 28, 213−222. (b) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J.
Chem. Phys. 1972, 56, 2257−2261. (c) Foresman, J. B.; Frisch, Æ.
Exploring Chemistry with Electronic Structure Methods, 2nd Ed.;
Gaussian, Inc.: Pittsburgh, PA; 1996, p 110.
(14) PROTEUM2 Version 1.0; Bruker AXS, Madison, WI, USA, 2005.
(15) Sheldrick, G. M. SHELXTL 5.1: A Program for the Refinement of
Crystal Structures; University of Gottingen, Gottingen, Germany, 1997.
(16) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837−838.
(17) Sheldrick, G. M. SADABS; University of Gottingen, Gottingen,
Germany, 1996.
(18) ORTEP3 for Windows: Farrugia, L. J. J. Appl. Crystallogr. 1997,
30, 565. (b) Merritt, E. A.; Bacon, D. J. Methods Enzymol. 1997, 277,
505−524.
(19) Darcy, P. J.; Heller, H. G.; Strydorn, P. J.; Whittall, J. J. Chem.
Soc., Perkin Trans. 1 1981, 202−205.
(20) Where peaks of major and minor isomers overlap, the area
contribution of each isomer has been apportioned on the basis of
peaks that do not overlap (see Figures S60−S63, Supporting
Information). For example, relative areas are 1.00:0.76:0.30 at δ 8.37
ppm (6-py, both major and minor 2 isomers), 5.52 ppm (o-Ph, only
major 2 isomer), and 4.53 ppm (CH2 cis, only minor 2 isomer).
Therefore, the peak at δ 8.37 ppm is estimated to be approximately
70% major and 30% minor isomer, respectively. The areas for peaks of
each isomer are reported relative to the isomer contribution at δ 8.37
ppm.
(36) (a) Dunlap, B. I. J. Chem. Phys. 1983, 78, 3140−3142.
(b) Dunlap, B. I. J. Mol. Struct. (THEOCHEM) 2000, 529, 37−40.
(c) Dunlap, B. I.; Connolly, J. W. D.; Sabin, J. R. J. Chem. Phys. 1979,
71, 3396−3402. (d) Dunlap, B. I.; Connolly, J. W. D.; Sabin, J. R. J.
Chem. Phys. 1979, 71, 4993−4999.
(37) NMR assignments are similar to those reported for (η6-
C6H5(CH2)2-η2-CHCH2)Cr(CO)2.6
(38) Values of 0.81 0.03 in isooctane and 0.64 0.01 in benzene
upon 313 nm irradiation have been previously reported in ref 22.
(39) Supporting Information, Table S38.
(40) Figures S13 and S14 in the Supporting Information.
(41) As expected, irradiation of 3 produces a mixture of 1 and 2 and
therefore a spectrum very different from those of 7−9. We previously
496
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497

Organometallics
Article
demonstrated that visible light transforms 1 to 2 and UV light
transforms 2 to 1.2
(42) Figures S15, S16, S17, S18a, and S18b in the Supporting
Therefore, the relative errors are smaller and the differences and
increase in enthalpies are significant.
(64) The Cr−pyridine bond enthaplies (corresponding to ΔH2comp in
Table 2) for 18 and the 4-methylpyridine derivative of 11 are −28.0
and −30.6 kcal/mol, respectively.
Information.
(43) Figures S19 and S23−S41 in the Supporting Information.
(44) Table S44-S62 in Supporting Information.
(65) To, T. T.; Heilweil, E. J. J. Phys. Chem. A 2007, 111, 8047−8049.
(45) (a) Hall, M. B.; Fenske, R. F. Inorg. Chem. 1972, 11, 768−775.
(b) Webster, C. E.; Hall, M. B. Forty Years of Fenske-Hall Molecular
Orbital Theory In Theory and Applications of Computational Chemistry:
The First 40 Years, A Volume of Technical and Historical Perspectives;
Dykstra, C. E., Frenking, G., Kim, K. S., Scuseria, G., Eds.; Elsevier:
New York, 2005; pp 1143−1165.
(46) Figures S20, S42, and S45 in the Supporting Information.
(47) Figures S21, S42, and S46 in the Supporting Information.
(48) Figures S43, S44, S47, and S48 in the Supporting Information.
(49) Figures S15−S18 and Table S71 in the Supporting Information.
Note that the values for 7−9 differ frrom those for Figure 5, where the
λmax value corresponds to simulated spectra, not a specific computed
transition.
(50) Alamiry, M. A. H.; Boyle, N. M.; Brookes, C. M.; George, M.
W.; Long, C.; Portius, P.; Pryce, M. T.; Ronayne, K. L.; Sun, X.-Z.;
Towrie, M.; Vuong, K. Q. Organometallics 2009, 28, 1461−1468.
(51) Creavens, B. S.; George, M. W.; Ginzburg, A. G.; Hughes, C.;
Kelly, J. M.; Long, C.; McGrath, K. M.; Pryce, M. T. Organometallics
1993, 12, 3127−3131.
(52) Tables S64 and Table S65 in the Supporting Information.
(53) A complete listing can be found in Tables S63−S65 in the
Supporting Information.
(54) (a) Clark, I. P.; George, M. W.; Greetham, G. M.; Harvey, E. C.;
Long, C.; Manton, J. C.; Pryce, M. T. J. Phys. Chem. A 2010, 114,
11425−11431. (b) Clark, I. P.; George, M. W.; Greetham, G. M.;
Harvey, E. C.; Long, C.; Manton, J. C.; Pryce, M. T. J. Phys. Chem. A
2011, 115, 2985−2993.
(55) This small energy difference between the C1 and Cs geometries
could be overcome by crystal-packing effects. The center to center
distance of a chromium-coordinated phenyl ring and the pyridinyl ring
of another molecule of 7 (in a neighboring unit cell) is 4.95 Å. This
packing of a phenyl ring with a pyridinyl ring is quite reminiscent of a
T-shaped benzene dimer which has an interaction energy of at least 2
kcal mol−1. There are two of these interactions per molecule of 7
because the two interacting molecules make a Ci-symmetric pair.
(56) The longer Cr−N bond length for 9 versus that for 7 (Table 1)
suggests that the more endothermic PAC result for 6 (ΔH1+2) may not
be just random error and is due to a slightly weaker Cr−N bond.
(57) χ = 3.4 mL/kcal for heptane.
(58) (a) van Eldik, R.; Asano, T.; le Noble, W. J. Chem. Rev. 1989, I,
549−688. (b) Drljaca, A.; Hubbard, C. D.; van Eldik, R.; Asano, T.;
Basilevsky, M. V.; le Noble, W. J. Chem. Rev. 1998, 98, 2167−2289.
(59) (a) Bryndza, H. E.; Fong, L. K.; Paciello, R. A.; Tam, W.;
Bercaw, J. E. J. Am. Chem. Soc. 1987, 109, 1444−1456. (b) Bruno, J.
W.; Marks, T. J.; Morss, L. R. J. Am. Chem. Soc. 1983, 105, 6824−6832.
(c) Gonzales, A. A.; Zang, K.; Nolan, S. P.; de la Vega, R. L.; Mukerjee,
S. L.; Hoff, C. D.; Kubas, G. J. Organometallics 1988, 12, 2429−2435.
(60) (a) Neugebauer, J.; Hess, B. A. J. Chem. Phys. 2003, 118, 7215−
7225. (b) Jonas, V.; Thiel, W. J. Phys. Chem. A 1999, 103, 1381−1393.
(c) Jonas, V.; Thiel, W. Organometallics 1998, 17, 353−360. (d) Jonas,
V.; Thiel, W. J. Chem. Phys. 1996, 105, 3636−3648. (e) Jonas, V.;
Thiel, W. J. Chem. Phys. 1995, 105, 8474−8484.
(61) Note that in each case computed values for CO wavenumbers
(Tables 3 and 4) are greater than the experimental frequencies because
the absolute values of computed harmonic stretching frequencies (ωe)
from DFT are often at higher energy in comparison to experimental
stretching frequencies (νe), which include anharmonicity (ωeχe, ωeye,
etc.).60 Nonetheless, the trends in the computed wavenumbers are the
same as for the experimental wavenumbers.
(62) van Meurs, F.; Baas, J. M. A.; van Bekkum, H. J. Organomet.
Chem. 1976, 113, 353−359.
(63) While the experimental PAC errors are significant, a portion
(about 2 kcal/mol) is due to the uncertainty in the quantum yields.
497
dx.doi.org/10.1021/om400928k | Organometallics 2014, 33, 485−497