Photoswitching a molecular catalyst to regulate CO2 hydrogenation

Article abstract of DOI:10.1039/c5dt01649e

Inspired by nature's ability to regulate catalysis using physiological stimuli, azobenzene was incorporated into Rh(bis)diphosphine CO₂ hydrogenation catalysts to photoinitiate structural changes to modulate the resulting catalytic activity. The rhodium bound diphosphine ligands (P(Ph₂)-CH₂-N(R)-CH₂-P(Ph₂)) contain the terminal amine of a non-natural amino acid, with the R-group being either β-alanine (β-Ala) or γ-aminobutyric acid (GABA). For both β-Ala and GABA containing complexes, the carboxylic acids of the amino acids were coupled to the amines of diaminoazobenzene, creating a complex consisting of a rhodium bound to a photo-responsive tetradentate ligand. The photo-induced cis-trans isomerization of the azobenzene-containing complexes imposes structural changes on these complexes, as evidenced by NMR studies. We found that the CO₂ hydrogenation activity for the β-Ala bound rhodium complex is 40% faster at 27°C with the light on, i.e. azobenzene in the cis-conformation (TOF = 16 s^-1) than when the complex was in the dark and the azobenzene in the trans-conformation (TOF = 11 s^-1). In contrast the γ-aminobutyric acid containing rhodium complex has the same rate (TOF ～17 s^-1) with the azobenzene in either the cis or the trans-conformation at 27°C. The corresponding (bis)diphosphine complexes without the attached azobenzene were also prepared, characterized, and catalytically tested for comparison, and have TOF's of 30 s^-1. Computational studies were undertaken to evaluate if the difference in rate between the cis- and trans-azobenzene isomers for the β-Ala bound rhodium complex were due to structural differences. These computational investigations revealed major structural changes between all cis- and trans-azobenzene structures, but only minor structural changes that would be unique to the β-Ala bound rhodium complex. We postulate that the different rates between the cis- and trans-azobenzene β-Ala bound containing rhodium complexes are due to subtle changes in the bite angle arising from steric strain due to the azobenzene-containing tetradentate ligand. This strain alters the hydricity of the subsequent rhodium hydride and consequently the rate.

Full text of DOI:10.1039/c5dt01649e

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Photoswitching a molecular catalyst to regulate
CO2 hydrogenation†
Cite this: Dalton Trans., 2015, 44,
14854
Nilusha Priyadarshani, Bojana Ginovska, J. Timothy Bays, John C. Linehan* and
Wendy J. Shaw*
Inspired by nature’s ability to regulate catalysis using physiological stimuli, azobenzene was incorporated
into Rh(bis)diphosphine CO2 hydrogenation catalysts to photoinitiate structural changes to modulate the
resulting catalytic activity. The rhodium bound diphosphine ligands (P(Ph2)-CH2-N(R)-CH2-P(Ph2))
contain the terminal amine of a non-natural amino acid, with the R-group being either β-alanine (β-Ala)
or γ-aminobutyric acid (GABA). For both β-Ala and GABA containing complexes, the carboxylic acids of
the amino acids were coupled to the amines of diaminoazobenzene, creating a complex consisting of a
rhodium bound to a photo-responsive tetradentate ligand. The photo-induced cis–trans isomerization of
the azobenzene-containing complexes imposes structural changes on these complexes, as evidenced
by NMR studies. We found that the CO2 hydrogenation activity for the β-Ala bound rhodium complex is
40% faster at 27 °C with the light on, i.e. azobenzene in the cis-conformation (TOF = 16 s−1
)
than when the complex was in the dark and the azobenzene in the trans-conformation (TOF = 11 s−1). In
contrast the γ-aminobutyric acid containing rhodium complex has the same rate (TOF ∼17 s−1) with the
azobenzene in either the cis or the trans-conformation at 27 °C. The corresponding (bis)diphosphine
complexes without the attached azobenzene were also prepared, characterized, and catalytically tested
for comparison, and have TOF’s of 30 s−1. Computational studies were undertaken to evaluate if the
difference in rate between the cis- and trans-azobenzene isomers for the β-Ala bound rhodium complex
were due to structural differences. These computational investigations revealed major structural changes
between all cis- and trans-azobenzene structures, but only minor structural changes that would be
unique to the β-Ala bound rhodium complex. We postulate that the different rates between the cis- and
trans-azobenzene β-Ala bound containing rhodium complexes are due to subtle changes in the bite
angle arising from steric strain due to the azobenzene-containing tetradentate ligand. This strain alters
the hydricity of the subsequent rhodium hydride and consequently the rate.
Received 2nd May 2015,
Accepted 22nd July 2015
DOI: 10.1039/c5dt01649e
www.rsc.org/dalton
many other examples highlight the importance of controlled
structural switching in enzymatic catalysis.1a
Introduction
Cooperation between structure and motion contributes to the
Inspired by the connection between the structure,
impressive performance of enzymes.1 Structural and mechan- dynamics, and function of these bio-catalysts, our group is
istic studies have revealed roles of dynamics in controlling sub- investigating the role of a structurally switchable outer coordi-
strate access,1 synchronizing reactions,1 changing the ligand nation sphere on molecular catalysts. Implementing structural
structure,2 or even changing the active site structure.3 For regulation into molecular catalysts is in its infancy, and many
example, a report by Dorothee Kern et al. demonstrates that previous demonstrations of this control have relied upon
the structure of adenylate kinases undergoes a dramatic con- changes in solubility to affect activity. For instance, Bergbreiter
formational change which closes some of the open sites, et al. designed thermally responsive smart catalysts by in-
excluding bulk water from the active site, and bringing the corporating poly-N-isopropyl acrylamide (pNIPAAm) into a
reactant into the active site for catalysis to occur.4 This and block copolymer which is bound to a phosphine-ligated metal
complex.5 Due to a structural change above 35 °C, pNIPAAm
becomes insoluble, allowing the catalyst to be turned off by
precipitation, and the precipitated catalyst to be recovered.
Attachment of molecular catalysts to polymers for the purpose
Pacific Northwest National Laboratory, Richland, WA 99354, USA.
E-mail: wendy.shaw@pnnl.gov, john.linehan@pnnl.gov; Fax: 509-371-6498;
Tel: 509-375-5922, 509-375-3983
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
of catalyst recoverability, with and without stimulus control,
has been a common theme.6 Using a similar approach, we
c5dt01649e
14854 | Dalton Trans., 2015, 44, 14854–14864
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
attached pNIPAAm to the well understood Rh(bis)diphosphine turally controllable outer coordination sphere. Herein, we
for hydrogenation of alkenes and demonstrated catalytic build upon our previous work by attaching non-natural amino
hydrogenation of 2-butenol.7
acids into the Rh(bis)diphosphine complex. To this we couple
Maintaining solubility while switching structure/reactivity a single azobenzene across the complex to both amino acids,
more directly mimics enzyme function and has the potential creating a photo-responsive tetradentate ligand.15 Our experi-
to enable a number of different ways to control catalysis. There mental results suggest that when the azobenzene ligand intro-
have been a number of approaches to this, for example, alter- duces steric strain, the cis–trans isomerization of azobenzene
ing the structure of a supramolecular complex to control steric regulates catalytic activity.
accessibility,8 using redox potentials to control non-catalytic
metal properties,9 and using protons to control oxidation
potentials.10
Results
Photolysis of appropriately modified molecular catalysts
has also been used as a means of controlling reactivity with an
external stimulus. While some of these complexes take advan-
Synthesis and characterization of photo-responsive
azobenzene Rh(bis)diphosphine complexes
tage of inducing charge transfers or altering luminescent pro- Two methodologies were attempted to synthesize the desired
perties of the core catalyst,11 the overwhelming majority of metal complexes. Pre-forming the intact tetra-dentate phos-
examples using light as a stimulus incorporate molecules such phine ligand followed by attachment to the metal core was
as azobenzene or stilbene, molecules which change confor- attempted first. While ligand synthesis and attachment of the
mation upon photolysis.12 One of the first azobenzene-based azobenzene bridge to the two bidentate phosphine ligands
(non-metal) photoswitchable catalysts was reported in 1981 by was accomplished, subsequent coordination of the rhodium
Ueno et al.13 In their work, azobenzene bound β-cyclodextrin metal resulted in multiple metal complexes as evident from
was active for ester hydrolysis where substrate access to the 31P NMR spectroscopy.
pocket, and consequently catalytic activity, was controlled by
The second approach was to coordinate two bidentate
the cis–trans isomerization of azobenzene. More recently, azo- ligands to the rhodium metal center and then join the ligands
benzene was attached to two zinc catalysts and the resulting with an azobenzene linker. This approach resulted in the
complex was demonstrated to have controllable reactivity for desired product. The synthesis, purification, and characteriz-
DNA cleavage, where the cis–trans isomerization changed the ation of bidentate P(Ph)2NRP(Ph)2 ligands (PNRP) and corres-
relative positioning of the zinc complexes, hindering or pro- ponding [Rh(PNRP)2]+ complexes, where R = CH2CH2COOH
moting their cooperativity.12c
(β-alanine or β-Ala) and R = CH2CH2CH2COOH (γ-aminobutyric-
Our photoswitchable catalysts build upon the Rh(bis)diphos- acid or GABA) followed a previously reported procedure,14
phine complexes we have developed for CO2 hydrogenation to result in [Rh(PNβ-Ala-OHP)2]+ and [Rh(PNGABA-OHP)2]+ in
(Fig. S1†) which contain amino acids,14 serving as attachment good yield (Scheme 1). These carboxylic acid functionalized
points for azobenzene. The development of this core complex [Rh(PNRP)2]+ metal complexes were bridged by chemically
serves as a foundation upon which we can investigate a struc- coupling them with one equivalent of 4,4′-diaminoazobenzene
Scheme 1 Synthetic scheme for azobenzene cross-linked Rh-R-Azb complexes.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 14854–14864 | 14855

View Article Online
Paper
Dalton Transactions
Fig. 1 (A) UV/Vis spectra of Rh-β-Ala-Azb (0.2 mM in N-methyl-2-pyrrolidinone (NMP)) as a function of time, tracking the cis-to-trans isomeriza-
tion after irradiation with 375 nm light. Inset shows the feature at 500 nm. (B) Time profile of the cis-to-trans isomerization of Rh-β-Ala-Azb (black)
and Rh-GABA-Azb (red) by monitoring the absorbance at 375 nm. Inset shows the relative rates of relaxation from cis-to-trans for both complexes,
indicating that Rh-β-Ala-Azb is slower.
using
1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo- wavelength of 375 nm converted trans-azobenzene to the cis-
[4,5-b]pyridinium 3-oxid hexafluorophosphate (HATU) as conformation, as evidenced by the shift of the π → π* from
the coupling reagent (Scheme 1) in the presence of an 374 nm to 264 nm, consistent with the cis-conformation
organic base such as diisopropylethyl amine (DIPEA) to yield (Fig. 1A). The n → π* (λmax ∼ 500 nm) also increased in inten-
sity, consistent with the photolytic behavior of azobenzene.16
The dark cis-to-trans isomerization rate of Rh-β-Ala-Azb and
Rh-GABA-Azb was significantly slower than for the uncom-
plexed 4,4′-diaminoazobenzene derivative, taking six to eight
hours for complete relaxation to the trans conformation
(Fig. 1B), compared to 15 minutes for the 4,4′-diaminoazo-
, annotated as (Rh-β-Ala-Azb or Rh-GABA-Azb),
in fair yield. The reaction was maintained at high dilution
(2.5 mM [Rh(PNRP)2]+). The azobenzene containing metal com-
plexes were then obtained by partial evaporation of the reaction
solvent and purified by precipitation of the complex with diethyl
ether.
benzene starting compound. The longer relaxation time is
Both azobenzene containing complexes were characterized
expected based on the restricted flexibility of azobenzene when
1
by H and 31P{1H} NMR spectroscopy and ESI-MS, all of which
it is coupled on both ends in [Rh(PNRP)2]+. The complex with
are consistent with the proposed structures (Fig. S2 and S3†).
the two carbon chain, Rh-β-Ala-Azb relaxes slightly slower than
Rh-GABA-Azb with a three carbon chain (Fig. 1B). This suggests
more steric constraint in the two carbon chain Rh-β-Ala-Azb
than in the three carbon chain Rh-GABA-Azb.
The newly formed amide bond was confirmed by 1H NMR
where the amide proton (CONH) was apparent at
10.0–10.2 ppm in H2O/THF (but not D2O/THF) (Fig. S2 and
S3†). The absence of potential byproducts was also confirmed
The structural isomerization of these complexes was also
observed by 1H NMR spectroscopy. Protons attached to the
by mass spectroscopy, which lacked corresponding masses for
complexes containing two azobenzenes, complexes with azo-
benzene attached at only one end, or complexes containing
multiple metal units (i.e. polymers).
methylene carbons between the nitrogen and phosphorus
(labeled 3 in Fig. 2) in Rh-β-Ala-Azb have inequivalent chemi-
cal shifts in the cis-conformation as opposed to the trans-
conformation, where the same aliphatic protons are equivalent
(Fig. 2). The rate of the cis-to-trans isomerization observed
from H NMR spectroscopy agrees with the rate obtained from
Isomerization studies of the azobenzene rhodium complexes
1
The UV/Vis spectrum of azobenzene shows two well separated
bands in the UV/Vis region (average λmax ∼ 375 nm (strong)
and 500 nm (weak)).16 The strong UV band arises from the π →
π* transition of the NvN bond, and the weaker absorbance
UV/Vis absorption methods (Fig. S5†).
Computational structural characterization
band in the visible region arises from the n → π* transition, To provide further insight into structural changes, compu-
with specific values being dependent upon substituents.16 tational studies of the complexes with azobenzene were evalu-
Similarly, these major absorbance bands were apparent in the ated and compared to [Rh(PNGly-OHP)2]+. The simulations show
UV/Vis spectra of Rh-β-Ala-Azb and Rh-GABA-Azb, as shown for that in all cases the dihedral angle (the angle between the two
Rh-β-Ala-Azb in Fig. 1A, with λmax = 374 nm and λmax = 371 nm planes defined by P–Rh–P and P′–Rh–P′ (Fig. S8†)) at the metal
for Rh-GABA-Azb. Irradiation of either rhodium complex at a center deviates from planarity, contrary to the crystal structure
14856 | Dalton Trans., 2015, 44, 14854–14864
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Fig. 2 1H NMR of Rh-β-Ala-Azb before (time = 0) and after (time = 7 minutes to 12 hours) 5 min of UV light irradiation. The three aliphatic proton
resonances observed for the trans-isomer (time = 0) become inequivalent upon 5 min irradiation with UV light, resulting in three pairs of aliphatic
proton resonances. This is indicative of a structural change and suggestive of an asymmetric structure due to the strain induced by the cis confor-
mation of the azobenzene.
to maintain the chair–chair conformation and the azo-bridge
did not undergo cis–trans isomerization within a given simu-
lation. Representative structures of the cis and trans-isomers
are shown in Fig. 3.
Table 1 Calculated structures for the azobenzene-containing rhodium
complexes reported in this study. Additional parameters are reported in
Table S1
∠dihedrala ∠P–Rh–P ∠P′–Rh–P′ r(Rh-N)
Complex
(°)
(°)
(°)
(Å)
Effect of the azobenzene photoswitch on the catalytic
hydrogenation of CO2
Rh-β-Ala-(cis)Azb
Rh-β-Ala-(trans)Azb
Rh-GABA-(cis)Azb
36.7 8.4
11.4 5.2
34.8 9.0
89.2 2.8 89.1 3.0
90.0 2.6 86.3 2.5
90.1 2.8 89.1 2.8
90.0 2.7 87.1 2.6
91.2 2.6 89.7 2.7
9.2 0.4
6.8 0.4
9.6 0.5
6.7 0.4
—
Each of the four metal complexes is active for CO2 hydrogen-
ation at 27 °C. N-Methylpyrrolidinone (NMP) was used as the
solvent to accommodate the limited solubility of the
azobenzene-containing complexes. Catalytic reactions of CO2
to formate were performed under 1 : 1 CO2 : H2 at 40 atm in a
view cell composed of a stainless steel block reactor with a sap-
phire window in one side to allow photolysis of the sample.
Catalytic activity of the azobenzene-containing complexes was
evaluated both during constant irradiation with a 100 watt UV
light (λmax = 365 nm) and in the dark, where the dark experi-
ments were performed irradiated, but with the window
covered, to account for any thermal heating of the reactor. The
catalytic production of formate was quantified by integrating
the formate resonance in the 1H NMR spectrum (8.7 ppm)
with respect to the formamide resonance (8.0 ppm) of DMF
added as an internal standard (see Table 2 and the Experi-
mental section for details) The base 1,8-diazobicyclo[5.4.0]-
undec-7-ene (DBU) was added to deprotonate the rhodium
dihydride intermediate formed during the catalytic cycle, as
well as to stabilize the formate from the reaction back to CO2
and H2.
Rh-GABA-(trans)Azb 10.7 5.2
[Rh(PNGly-OHP)2]+
9.8 5.7
a The dihederal is the angle between the two planes P–Rh–P and
P′–Rh–P′. r(Rh–N) is the distance between the Rh metal center and one
of the N atoms in azo group in the Azb bridge (shown in Fig. 3).
isolated for [Rh(PNGly-OHP)2]+,14 with an angle ∼10 degrees
in the [Rh(PNGly-OHP)2]+ complex and in the trans-azobenzene
isomers, and ∼35 degrees for the cis-azobenzene-isomers
(Table 1). The distances between the rhodium center and the
nitrogen of the azo-bridge are also significantly different for
the complexes in the cis- (dark) and trans- (light) confor-
mations, with the azo-bridge in the cis-isomer giving a dis-
tance of ∼9 Å, and in the trans-isomer ∼7 Å (Table 1 and
Fig. 3), in both cases too far from the rhodium to hinder
activity via a direct interaction of the azo-group with rhodium.
Changes in the bite angle (∠P–Rh–P) were measured from the
simulations, as it has been shown that small changes in this
angle can significantly alter the electronics of the metal
center.17 The pendant amines in the complexes were observed
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 14854–14864 | 14857

View Article Online
Paper
Dalton Transactions
Fig. 3 Representative structures of the (left) cis- and (right) trans-azobenzene in Rh-β-Ala-Azb, showing the chair–chair conformation maintained
by the pendant amine in both complexes, as well as the average distance of the azo-group from the rhodium.
Table 2 Rates of CO2 hydrogenation for [Rh(PNRP)2]+ and Rh-R-Azb
complexes. Rates for the azobenzene containing complexes were deter-
mined in the cis-azobenzene (light) and trans-azobenzene (dark)
conformations
Table 3 Thermodynamic parameters for the reaction of [Rh(PNRP)2]+
or Rh-R-(trans)-Azb (dark) with H2 to produce the dihydride complexes,
[H2Rh(PNRP)2]+ or H2-Rh-R-(trans)-Azb
Complex
Keq (atm−1
)
ΔGH (kcal mol−1
)
2
[M+] [DBU] Azobenzene
(mM) (mM) conformation (h−1
TOF
TONa
(TON b)
[H2Rh(dppp)2]
6.70a
0.30a
0.53
1.71
1.48
0.97
−1.10a
+0.77a
+0.37
−0.32
−0.23
+0.02
Catalyst
)
[H2Rh(PNGly-OHP)2]
[H2Rh(PNβ-Ala-OHP)2]
[H2Rh(PNGABA-OHP)2]
H2Rh-β-Ala-Azb
[Rh(PNβ-Ala-OHP)2]+ 1.5
[Rh(PNGABA-OHP)2]+ 1.5
700
800
420
420
420
420
—
—
cis
trans
cis
33
32
16
11
3
790 (890)
770 (910)
380 (600)
280 (630)
2
2
2
Rh-β-Ala-Azb
1.2
H2Rh-GABA-Azb
Rh-GABA-Azb
1.1
17 0.2 410 (710)
18 0.2 430 (720)
a Reported in ref. 14. All H2 additions were carried out in a J. Young
NMR tube at 1 atm H2. The units for Keq result from the concentration
of H2 gas which is not 1 M (i.e. non-standard state).
trans
All the reactions were carried out in N-methyl-pyrolidinone, catalyst
concentration 1.1–1.5 mM, 27 °C, either light or dark, for 24 hours. All
data are an average of three runs, except [Rh(PNβ-Ala-OHP)2]+ which is
an average of two runs. a The TON was calculated experimentally. b The
theoretical TON is based on added DBU and the formate : DBU (1.7)
due to homoassociation (see Experimental section and Fig. S6 for
details).
NMR spectra, consistent with a square pyramidal structure
containing four equivalents phosphorous atoms. Repeating
these experiments under light conditions (cis-azobenzene con-
figuration) produced the same distribution of intermediates.
Thermodynamic studies
To evaluate mechanistic steps in the catalytic cycle, stoichio-
metric studies of H2 addition and subsequent deprotonation
of all four complexes were carried out under similar solvent
conditions to those used for catalysis. The dihydrides resulting
from H2 addition (1 atm), [H2Rh(PNRP)2]+ and H2Rh-R-Azb
Discussion
Enzymes have long been known to control catalysis by altering
their dynamic structure.1a Mimicking this feature in molecular
catalysts has the potential to provide similar regulation in
molecular complexes, and could include additional advantages
such as protecting the catalyst from poisons or modulating
catalytic activity between different substrates. In this work,
we exploit the structural interconversion of azobenzene by
binding azobenzene across a rhodium center to produce a
photo-responsive tetradentate ligand, to investigate the impact
on catalytic activity.
To investigate the role of structural switching on catalytic
activity, we prepared rhodium complexes with an azobenzene-
derived (bis)diphosphine photo-responsive tetradentate
ligand, Rh-R-Azb. This ligand motif is similar to those
observed in trans-spanning ligands. The first trans-spanning
1
were identified by two doublets of triplets in the H NMR with
the typical AA′XX′M splitting pattern observed for these
complexes.14 Representative 1H and 31P{1H} NMR spectra for
Rh-β-Ala-Azb are shown in Fig. 4.18 As captured in the Keq
values, the Rh-β-Ala-Azb formed significantly more dihydride
than the non-azobenzene counterpart ([Rh(PNβ-Ala-OHP)2]+),
while the opposite trend was observed for the GABA containing
complexes (Table 3). In all cases, addition of ∼50 eq. DBU to
the above dihydride complexes quantitatively converted to a
monohydride, [HRh(PNRP)2]+, containing a single broad (too
broad to observe Rh–H coupling) resonance in the 1H NMR
spectra and a doublet due to rhodium coupling in the 31P{1H}
14858 | Dalton Trans., 2015, 44, 14854–14864
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Fig. 4 The 31P{1H} NMR spectra of Rh-β-Ala-Azb (bottom), H2Rh-β-Ala-Azb (middle) and HRh-β-Ala-Azb (top). One atm H2 was added using a
J. Young NMR tube with 10 mM complex in NMP. Upper right shows the 1H NMR of the hydride resonance for the monohydride and dihydride of
Rh-β-Ala-Azb.
ligand reported by Issleib and Hohlfeld in 1961 was based on showed rates two-times higher compared to the azobenzene
an alkyl chain bridged diphosphine ligand (C6H11)2P(CH2)n- cross-linked derivatives (Table 2), attributed to the free acidic
P(C6H11)2 which yielded distorted square planer geometries functionality as previously reported.14 This suggests several
when bound to nickel with n < 4.19 Development of trans-span- factors which may be contributing to reactivity: (1) the azobenz-
ning ligands is based on understanding and modulating the ene could change the geometry of the active site for Rh-β-Ala-
steric and electronic effects of the specific ligands.15,20 The Azb altering the electronics and/or thermodynamics; (2) the
incorporation of azobenzene into the [Rh(PNRP)2]+ complexes nitrogen groups of the azo functional group could bind to the
discussed here imposes a similar structural strain on the metal in the trans-configuration, hindering reactivity.
macrocycle formed, and additionally provides the opportunity
to toggle between two differently strained structures.
Crystal structures have the potential to reveal structural con-
tributions of the azobenzene ligands to the reactivity, but
The Rh-β-Ala-Azb complex showed a difference in CO2 unfortunately we were unable to obtain crystals of these com-
hydrogenation activity under cis- and trans-conformational plexes. To provide insight into which of these possibilities
forms. The complex with the azobenzene in the cis-confor- might be dominating the reactivity of these complexes, we
mation showed a 40% rate enhancement compared to the performed semi-empirical calculations on the azobenzene-
complex with the azobenzene in the trans-conformation, indi- containing complexes and [Rh(PNGly-OHP)2]+. These calcu-
cating that the photoinitiated structure is altering the active lations demonstrate that azobenzene is too far from the metal
site in some way to influence reactivity. Dependence upon to bind in either the cis- or trans-conformations, removing azo-
chain length is demonstrated in that the Rh-GABA-Azb benzene metal binding as a possible explanation of the differ-
complex with an additional carbon does not exhibit a differ- ences in observed rates. Significant differences in structure are
ence in rate between the cis- and trans-conformations. observed between cis- and trans-conformations, particularly in
Attempts to look at an even shorter carbon chain, i.e. glycine, the dihedral angle and the Rh–N distances, but not between
did not result in the tetradentate ligand with the azobenzene the complexes with different carbon length linkers, suggesting
bridge, but instead resulted in a complex with an azobenzene that significant structural changes are not contributing to the
attached to each glycine, preventing further evaluation of the observed catalytic behavior. Further, because large changes are
chain length (See ESI† for details).
observed in the dihedral angle between cis and trans for both
The parent rhodium complexes, [Rh(PNβ-Ala-OHP)2]+ and complexes, this suggests that reactivity for CO2 is not sensitive
[Rh(PNGABA-OHP)2]+, were also investigated for comparison and to these angular and spatial parameters.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 14854–14864 | 14859

View Article Online
Paper
Dalton Transactions
It is also possible that the azobenzene is contributing cycles before use. Tetrahydrofuran (d8-THF) from Cambridge
directly to steps in the catalytic cycle, such as H2 addition or isotope was dried over NaK and vacuum transferred. Other
deprotonation. To understand potential changes in thermo- organic solvents were dried and degassed using an Innovative
dynamics, intermediates in two of the three steps in the pro- Technology, Inc., PureSolv Solvent purification systems. UHP
1
posed catalytic cycle (Fig. S1†) were prepared. The dihydride (ultra-high purity) CO2 and H2 were purchased from Oxarc. H
was formed by exposure to H2, and deprotonation to form the and 13C{1H} NMR spectra were recorded on Varian Inova
monohydride was achieved by the addition of DBU. Rh-β-Ala- 500 MHz or VNMRS 300 MHz NMR spectrometers. All 1H
Azb and Rh-GABA-Azb have negative or zero values for ΔGH
chemical shifts are internally calibrated using the monoprotic
2
resulting in favorable H2 addition, similar to other impurity of the deuterated solvent; 13C NMR chemical shifts
[Rh(PNRP)2]+ complexes in this work and many that are are referenced to CD3CN or d8-THF. 31P{1H} NMR spectra were
reported earlier.14 Complete conversion to the hydride upon obtained on a Varian Inova 500 spectrometer and are refer-
addition of base for all complexes, as well as a lack of corre- enced externally to 85% H3PO4 (0 ppm). Mass analysis was per-
lation between the rate of H2 addition and CO2 hydrogenation, formed using a 15T Fourier transform ion cyclotron resonance
suggest that H2 addition is not the rate limiting step in the mass spectrometer (FTICR-MS) (Bruker SolariX, Billerica, MA)
mechanism, in agreement with the previously reported behav- outfitted with a standard electrospray ionization (ESI) inter-
ior.14 Importantly, the variability between the ΔGH of these face. UV-Vis spectra of azobenzene derivatives were measured
2
and previous complexes, with no correlation to activity, on Shimadzu UV-2401PC UV-Vis spectrometer. Isomerization
suggests that H2 addition and subsequent deprotonation are of these samples from trans to cis was achieved with a 1000
not contributing to the differences observed in rates for the watt xenon–mercury arc lamp, directed through a monochrom-
complexes in this study.
Collectively, the computational and thermodynamics azobenzene bound complexes were performed using the 100
results suggest that the azobenzene groups are not providing a watt UV lamp, UVP-B100 series with peak intensity
ator to generate light at λ = 375 nm. Catalytic studies of the
a
significant structural change, or in the energetics of the inter- (21 700 μW cm−2 at 2″ distance) at 365 nm and a distribution
mediates in the catalytic cycle. Complexes similar to these of 40 nm. (The 100 watt UV lamp was used for catalysis after
have been shown to be extremely sensitive to small changes in determining the rates (∼5 minutes) of trans-to-cis conversion
hydricity, which can arise from subtle changes in the P–Rh–P of the complexes by both lamps were similar.)
bite angle. For example, ∼10 degree changes in the bite angle
The rates of the cis-to-trans isomerization of Rh-β-Ala-Azb
for Co, Rh, Ir, Ni, Pt or Pd complexes results in ∼20 kcal mol−1 and Rh-GABA-Azb were determined by monitoring the increase
change in the hydricity, or ∼2 kcal per mol per degree.17a The in the π to π* absorption band at 365 nm after irradiation.
modest changes in rate observed here are consistent with a Rate constants for isomerization were determined by fitting
ΔΔG of 0.25 kcal mol−1 or a change in bite angle of only 0.5 the data to a first order rate equation.
degrees, a change that is within the error of our calculations.
Preparation of PNβ-Ala-OH
P
In summary, the introduction of an azobenzene across
Rh(bis)diphosphine complexes allowed us to photolytically A 250 mL Schlenk flask was loaded with 0.081 g (2.69 mmol)
control the rate of CO2 hydrogenation. This was only true in paraformaldehyde, 0.5 g (2.69 mmol) diphenylphosphine, and
the complex with the shorter chain length (Rh-β-Ala-Azb), β-alanine (0.119 g, 1.34 mmol) in 150 mL ethanol. The reaction
suggesting a steric component to the catalytic control. Compu- flask was heated to 70 °C and stirred for 12 hours. After
tational data suggest that the large changes in structure due to 12 hours the solvent was removed under reduced pressure to
cis–trans isomerization (i.e. dihedral and r(Rh–N)) are not obtain a viscous oil. The oil was dissolved in 2–3 mL of THF
responsible for the differences in rate. Consistent with both and the white product was precipitated using diethyl ether.
the experimental and computational data is that the steric The solid was filtered through a medium frit and thoroughly
rigidity is introducing subtle changes in the P–Rh–P bite angle washed with diethyl ether to obtain a white solid in 85% yield.
for Rh-β-Ala-Azb, resulting in a change in hydricity, controlling 1H NMR (500 MHz, CD3CN): δ 2.37 (br, 2H, NCH2CH2COOH),
catalysis. This behavior is inspired by enzymatic regulation of 3.14 (br, 2H, NCH2CH2COOH), 3.65 (s, 4H, PCH2N), 6.96–8.06
catalytic processes, and demonstrates that similar control is (m, 20H, P-PhH), 10.82 (br, 1H, N–CH2CH2COOH). 13C{1H}
possible for molecular complexes.
NMR (125 MHz, CDCl3): δ 30.3 (–CH2–COOH), 51.9 (N–CH2–
2
CH2), 57.7, 128.9 (s, C-Ph), 129.3 (s, C-Ph), 133.2 (d, JPC = 19
1
Hz, C-Ph), 136.8 (d, JPC = 10.7 Hz, C-Ph), 173.9 (–CH2–COOH).
31P{1H} NMR (202 MHz, d8-THF): δ −29.15. ESI MS: observed
(MH+) m/z = 486.17; calculated (MH+) m/z = 486.17. Elem. anal.
Calcd for C29H29NO2P2 + 0.5diethyl ether: C, 71.25; H, 6.56; N,
Experimental
General methods
All chemical reactions were performed under inert atmosphere 2.68; Found C, 71.10; H, 6.35; N, 2.73.
on a Schlenk line or in a glove box unless otherwise noted. All
Preparation of PNGABA-OH
P
the reagents were purchased from Sigma Aldrich and used
without further purification. DBU (1,8-diazobicyclo[5.4.0]- This ligand was prepared in a similar manner to PNβ-Ala-OHP
undec-7-ene) was degassed by repeated freeze–pump–thaw with 0.74 g (7.14 mmol) γ-aminobutyric acid (GABA), 2.67 g
14860 | Dalton Trans., 2015, 44, 14854–14864
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
(14.3 mmol) diphenylphosphine and 0.50 g paraformaldehyde calculated (M+ − OTf ) m/z = 1101.25. Elem anal. Calcd for
in 150 mL ethanol and stirred for 16 h at 71 °C. Removal of C61H62N2O7F3P4RhS + 1H2O: C, 57.69; H, 5.16; N, 2.21; Found
solvent produced an oily material. This was then dissolved in C, 57.43; H, 5.17; N, 2.22.
minimal THF (∼1–2 ml) and added dropwise to 10 ml of
diethyl ether to produce a white precipitate. The white solid
Rh-β-Ala-Azb
was isolated upon filtration and washed with excess diethyl A 250 mL schlenk flask containing 0.2 g (0.16 mmol)
ether and dried under vacuum to obtain a white solid in good [Rh(PNβ-Ala-OHP)2](OTf) and 2.2 equivalent HATU (0.124 g,
yield, ∼82%. 1H NMR (500 MHz, d8-THF): δ 1.33 (br, 2H, 0.33 mmol) in 100 mL THF was stirred and diisopropylethyl
NCH2CH2CH2COOH), 1.69 (br, 2H, NCH2CH2CH2COOH), 2.33 amine (DIPEA) (0.046 g, 0.36 mmol) was added to the flask.
(br, 2H, NCH2CH2CH2COOH), 3.29 (s, 4H, PCH2N), 7.10–7.33 This solution was stirred for 15 minutes and diaminoazo-
(m, 20H, PPhH), 8.82 (br, 1H, NCH2CH2COOH). 13C{1H} NMR benzene (0.035 g, 0.16 mmol) dissolved in 5 mL of THF was
(125 MHz, CDCl3): δ 21.9 (CH2–CH2–CH2), 30.7 (CH2–CH2– added drop wise in the dark to the above reaction solution and
CH2), 32.0 (CH2–CH2–CH2), 58.7 (P–CH2–N), 128.5 (C-Ph), stirred for 24 hours under ambient light. Removal of solvent
2
1
129.2 (C-Ph), 133.1 (d, JPC = 17.5, C-Ph), 137.9 (d, JPC = 12.5, resulted in a yellow powder which was redissolved in 2–3 mL
C-Ph). 31P{1H} NMR (202 MHz, CD3CN): δ −26.7. ESI MS: of 5–10% aqueous THF and precipitated into 15 mL of diethyl
observed (MH+) m/z = 500.19; calculated (MH+) m/z = 500.18. ether. This precipitate was filtered through a medium frit, the
Elem anal. Calcd for C30H31NO2P2 + 0.5diethyl ether: C, 71.63; yellow solid was washed thoroughly with diethyl ether followed
H, 6.76; N, 2.61; Found C, 71.66; H, 6.44; N, 2.81.
by washing with acetonitrile to obtain a bright yellow solid in
42% yield, which has a ∼50 : 50 mixture of triflate and hexa-
fluorophosphate counterions (if excess HATU was used for the
General procedure for the synthesis of [Rh(PNRP)2]+
To a solution of the PNRP ligand (1 mmol) in THF, 0.5 mmol synthesis of the above complex, total exchange of the OTf to
of Rh(COD)2OTf was added drop wise to a 20 mL glass vial PF6 was observed. The rates for CO2 hydrogenation with a
equipped with a stir bar. This solution was stirred for 12 h to single counter anion or mixed counter anions did not show a
obtain a clear yellow solution. The THF solution was concen- significant difference (Table S3†).) 1H NMR (500 MHz, 10%
trated to 1 mL under reduced pressure and was added drop D2O and d8-THF): δ 2.30 (br, 4H, NCH2CH2CONH), 2.83 (br
wise to a flask containing 10 mL diethyl ether to precipitate 4H, NCH2CH2CONH), 3.43 (s, 8H, PCH2N), 6.74–7.91 (m, 48H,
the product as a yellow suspension. The product was filtered PPhH & Azb-PhH); 13C{1H} NMR (125 MHz, 10% D2O and
through a medium frit, the solid was thoroughly washed with d8-THF)): δ 31.22, (NCH2CH2–), 33.37, (NCH2CH2–), 57.18 (br,
diethyl ether and dried to afford the Rh complex in good yield PCH2N), 119.3, 123.3, 128.2, 130.1, 131.0, 132.2, 132.8, 133.3,
(70–85%).
(aromatics) 169.9 (–CONH); 31P{1H} NMR (202 MHz, 10% D2O
and d8-THF): δ 7.93 (d, JRhP = 130.3 Hz). ESI MS: observed
(MH+ − OTf ) m/z = 1251.325; calculated (MH+ − OTf ) m/z =
[Rh(PNβ-Ala-OHP)2](OTf)
The compound was prepared following the above procedure at 1250.32. Elem anal. Calcd for C70H66F6N6O2P5Rh: C, 60.27;
a 1 mmol ligand scale to obtain a yellow solid in 70% yield. 1H H, 4.77; N, 6.02; Found C, 60.05; H, 4.90; N, 6.33.
NMR (500 MHz, CD3CN): δ 2.11 (t, 4H, JHH = 8 Hz,
Rh-GABA-Azb was prepared following a similar procedure
CH2CH2COOH), 2.66 (t, 4H, JHH = 8 Hz, NCH2CH2COOH), 3.36 as described above with 0.25 g of [Rh(PNGABA-OHP)2]+ to obtain
(s, 8H, PCH2N), 7.10–7.55 (m, 40H, PPhH), 9.16 (br, 2H, a bright yellow solid in 40% yield, with a ∼50 : 50 mixture
NCH2CH2COOH): 13C{1H} NMR (125 MHz, d6-DMSO): δ 33.9, of triflate and hexafluorophosphate counterions. 1H NMR
35.2 (–CH2CH2C(O)), 60.7 (P–CH2–N) 119.9, 123.9, 142.0, (500 MHz, 10% D2O and d8-THF)):
δ 1.52 (br 4H,
148.15 (aromatics) 169.3, (COOH); 31P{1H} NMR (202 MHz, d8- NCH2CH2CH2CONH), 2.06 (br 4H, NCH2CH2CH2CONH), 2.47
THF): δ 8.38, JRhP = 130.1 Hz. ESI MS: observed (M+ − OTf ) m/z (br 4H, NCH2CH2CH2CONH), 3.38 (s, 8H, PCH2N), 6.72–7.79
= 1073.24; calculated (M+ − OTf ) m/z = 1072.31. Elem anal. (m, 48 H, PPhH & AzbPhH); 13C{1H} NMR (125 MHz, CD3CN):
Calcd for C59H58N2O7F3P4RhS: C, 57.94; H, 4.78; N, 2.29; δ 25.7, 27.3, 33.9, (–CH2CH2CH2–), 60.9, (PCH2N–), 119.0,
Found C, 57.70; H, 4.90; N, 2.30.
119.3, 123.2, 123.4, 141.9, 147.4, 157.4, (aromatics)
171.5 (CONH); 31P{1H} NMR (202 MHz, CD3CN): δ 7.07 (d,
JRhP = 131 Hz). ESI MS: observed (MH+ − OTf ) m/z = 1278.34;
[Rh(PNGABA-OHP)2](OTf)
Following the general procedure above, [Rh(PNGABA-OHP)2]+ calculated (MH+ − OTf ) m/z = 1278.36. Elem anal. Calcd for
compound was prepared as a yellow solid in 85% yield. 1H C72H70F6N6O2P5Rh + 4THF + 3H2O: C, 60.92; H, 5.34; N, 4.79;
NMR (500 MHz, d8-THF): δ 1.28 (br, 4H, NCH2CH2CH2COOH), Found C, 60.75; H, 5.07; N, 4.63.
1.64 (br, 4H, NCH2CH2CH2COOH), 2.28 (br, 4H,
General procedure for catalysis
NCH2CH2CH2COOH), 3.24 (s, 8H, PCH2N), 6.92–7.49 (m, 40H,
PPhH; 31P{1H} NMR (202 MHz, CD3CN): δ 8.43 (d, JRhP = 131.3 The high-pressure view cell, previously reported,21 having a
Hz). 13C{1H} NMR (125 MHz, 10% D2O and d8-THF)): δ 21.2, sapphire optical window with a teflon o-ring seal and one gas
(NCH2CH2CH2–), 29.5, (NCH2CH2–CH2–), 33.4, (NCH2CH2- port, was used for all batch CO2-hydrogenation reactions.
CH2–), 55.0 (br, PCH2N), 126.1, 126.3, 131.9, 142.1 (aromatics) The cell was sufficient to contain the 40 atm of CO2/H2 used
175.1 (–CONH) ESI MS: observed (M+ − OTf ) m/z = 1101.27; for catalysis at room temperature. This window was used to
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 14854–14864 | 14861

View Article Online
Paper
Dalton Transactions
irradiate the sample with UV light to control the conformation concentration of DBU in NMP and monitoring the required
of the azobenzene in the complex and to provide a means of amount of formic acid to consume the DBU (Fig. S6†).
viewing the reaction solution during catalysis.
Thermodynamic studies
The reaction solution was prepared in a glove box under N2
atmosphere and the catalyst and DBU base were dissolved in
1.0 mL of N-methylpyrrolidinone (NMP) which was then added
to the high-pressure view cell. The volume was then adjusted
to a total volume between 3.0 mL to 3.5 mL with NMP to result
in final catalyst concentrations of 1.1 mM–1.5 mM and final
DBU concentrations of 410–650 mM. No difference in rate was
observed for a given catalyst under similar conditions within
this concentration range. The reaction cell was connected to a
model 500D ISCO syringe pump apparatus via PEEK tubing,
and an N2 environment in the tubing were eliminated by
applying a pseudo vacuum (connected to the vacuum when
active vacuum valve is closed).14 Then the 1 : 1 CO2/H2 gas
mixture was delivered at a constant pressure of 40 atm using
Addition of H2 to the [Rh(PNRP)2]+ complexes was carried out
in J. Young NMR tubes. The metal complex was dissolved in
400 µL of NMP and added to the NMR tube in a N2 atmos-
phere. On a gas manifold, the N2 atmosphere in the NMR tube
was removed by exposing it to the static vacuum in the mani-
fold three times. This involved pulling the vacuum on the gas
manifold with the J. Young tube closed, isolating the manifold
from the vacuum pump, and then opening the J. Young tube
to the static vacuum in the manifold. Hydrogen was added
similarly by exposing the solution in the J. Young tube to
1 atm hydrogen three times, with one minute of vortex mixing
between each addition.
an ISCO syringe pump. Following this addition, the reaction Computational studies
proceeded for 24 hours with stirring to allow for gas mixing
Stable structures were optimized starting from the crystal
with the liquid phase during catalysis. UV light was kept
turned “ON” even for dark reaction conditions where the view-
port was closed with a black aluminum foil to account for any
thermal fluctuations in the “dark” and “light” states. The
temperature of the pressure vessel was monitored using an
external thermocouple, and recorded to be 27 1 °C in all
catalytic reactions, both light and dark. A 10–20% loss of the
complex in the trans-conformation was observed after 24 hours
under catalytic conditions, performed under either light or
dark conditions.
structures of [Rh(PNGly-OHP)2]+ using a semi-empirical level of
theory (PM6).22 The lowest energy structure was selected (see
ESI of ref. 14) and molecular dynamics studies were carried
out in the gas phase to explore the structural parameters of the
cis-azobenzene and trans-azobenzene within the complexes,
for both the Rh-β-Ala-Azb and Rh-GABA-Azb. All simulations
were performed using the CP2K program23 and trajectories
were generated using the NVT ensemble at 300 K. Data were
collected for 75 ps of simulation time. The initial structures in
the simulations were in the chair–chair configuration, as that
configuration was found to be most stable.
Determination of TONs
After 24 hours, the view cell and PEEK tubing were dis-
connected from the CO2 : H2 gas inlet and the pressure was
released in the glove box. 300 µL aliquots of the reaction solu-
tion were placed in an NMR tube to which 50 µL aqueous 10%
N,N-dimtheylformamide (DMF) was added. The determination
of TONs for the catalytic runs were carried out by integrating
the formate resonance with respect to the formamide proton
of DMF in the 1H NMR spectrum which was run with an acqui-
sition time of 2 s, a 1 s relaxation delay, and 100 scans. The
TOF’s reported in Table 2 were calculated for reactions after
24 h using the following equation:
Acknowledgements
This work was supported by the US Department of Energy,
Office of Science, Office of Basic Energy Sciences, Division of
Chemical Sciences, Geosciences & Biosciences. Pacific North-
west National Laboratory (PNNL) is a multiprogram national
laboratory operated for the DOE by Battelle. A portion of this
research was performed using EMSL, a national scientific user
facility sponsored by the Department of Energy’s Office of Bio-
logical and Environmental Research and located at Pacific
Northwest National Laboratory.
½formateꢀ
TOF ¼
½catalystꢀ ꢁ 24ðhÞ
References
and the value reported is an average of three experiments
run under identical conditions, with the exception of
[Rh(PNβ-Ala-OHP)2]+ which is an average of only two experi-
ments. The theoretical TON was calculated using the following
equation:
1 (a) A. Gora, J. Brezovsky and J. Damborsky, Gates of
enzymes, Chem. Rev., 2013, 113(8), 5871–5923; (b) J. Lee
and N. M. Goodey, Catalytic contributions from remote
regions of enzyme structure, Chem. Rev., 2011, 111(12),
7595–7624.
2 (a) J. C. Boyington, V. N. Gladyshev, S. V. Khangulov,
T. C. Stadtman and P. D. Sun, Crystal structure of formate
dehydrogenase H: catalysis involving Mo, molybdopterin,
selenocysteine, and an Fe4S4 cluster, Science, 1997, 275(5304),
½DBUꢀ ꢁ 1:7
TON ¼
½catalystꢀ
where the formate : DBU homoassociation was found to be
∼1.7, which was determined by adding formic acid to a known
14862 | Dalton Trans., 2015, 44, 14854–14864
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
1305–1308; (b) H. C. Raaijmakers and M. J. Romao,
Formate-reduced E. coli formate dehydrogenase H: The
amide-based pincer platinum complexes, Organometallics,
2009, 28, 3307–3310.
reinterpretation of the crystal structure suggests a new 12 (a) B. M. Neilson and C. W. Bielawski, Illuminating Photo-
reaction mechanism, J. Biol. Inorg. Chem., 2006, 11(7),
849–854.
switchable Catalysis, ACS Catal., 2013, 3(8), 1874–1885;
(b) S. Umeki, S. Kume and H. Nishihara, Switching of
molecular insertion in a cyclic molecule via photo- and
thermal isomerization, Inorg. Chem., 2011, 50, 4925–4933;
(c) A. Panja, t. Matsuo, S. Magao and S. Hirota, DNA
cleavage by the photocontrolled cooperation of ZnII centers
in an azobenzene-linked dizinc complex, Inorg. Chem.,
2011, 50, 11437–11445; (d) J. Yin, G.-A. Yu, J. Guan, F. Mei
and S. H. Liu, Synthesis and properties of conjugated
bimetallic ruthenium complexes with s,s-bridging azoben-
zene chains, J. Organomet. Chem., 2005, 690, 4265–4271;
(e) Y.-X. Yuan, Y. Chen, Y.-C. Wang, C.-Y. Su, S.-M. Liang,
H. Chao and L.-N. Ji, Redox responsive luminescent switch
based on a ruthenium(II) complex [Ru(bpy)2(PAIDH)]2+,
Inorg. Chem. Commun., 2008, 11, 1048–1050; (f) Y.
Yamamoto, H. Nakamura and J.-F. Ma, Preparation
and characterization of ruthenium(II), rhodium(III) and
iridium(III) complexes of isocyanide bearing the azo group,
J. Organomet. Chem., 2001, 640, 10–20.
3 (a) K. A. Henzler-Wildman, V. Thai, M. Lei, M. Ott,
M. Wolf-Watz, T. Fenn, E. Pozharski, M. A. Wilson,
G. A. Petsko, M. Karplus, C. G. Hubner and D. Kern,
Intrinsic motions along an enzymatic reaction trajectory,
Nature, 2007, 450(7171), 838–844; (b) E. Z. Eisenmesser,
O. Millet, W. Labeikovsky, D. M. Korzhnev, M. Wolf-Watz,
D. A. Bosco, J. J. Skalicky, L. E. Kay and D. Kern, Intrinsic
dynamics of an enzyme underlies catalysis, Nature, 2005,
438(7064), 117–121.
4 K. A. Henzler-Wildman, M. Lei, V. Thai, S. J. Kerns,
M. Karplus and D. Kern, A hierarchy of timescales in
protein dynamics is linked to enzyme catalysis, Nature,
2007, 450(7171), 913–916.
5 D. E. Bergbreiter, Using Soluble Polymers To Recover
Catalysts and Ligands, Chem. Rev., 2002, 102(10), 3345–
3384.
6 (a) C. Baleizao and H. Garcia, Chiral salen complexes: an
overview to recoverable and reusable homogeneous and 13 A. Ueno, K. Takahashi and T. Osa, Photocontrol of catalytic
heterogeneous catalysts, Chem. Rev., 2006, 106, 3987–4043;
(b) T. J. Dickerson, N. N. Reed and K. D. Janda, Soluble
activity of capped cyclodextrin, J. Chem. Soc., Chem.
Commun., 1981, (3), 94–96.
polymers as scaffolds for recoverable catalysts and reagents, 14 J. T. Bays, N. Priyadarshani, M. S. Jeletic, E. B. Hulley,
Chem. Rev., 2002, 102, 3325–3344; (c) N. Madhavan,
C. W. Jones and M. Weck, Rational approach to polymer-
supported catalysts: synergy between catalytic reaction
mechanism and polymer design, Acc. Chem. Res., 2008, 41,
1153–1165.
7 (a) W. J. Shaw, Y. Chen, J. Fulton, J. Linehan, A. Gutowska
and T. Bitterwolf, Structural evolution of a recoverable
rhodium hydrogenation catalyst, J. Organomet. Chem., 2008,
D. L. Miller, J. C. Linehan and W. J. Shaw, The influence of
the second and outer coordination spheres on Rh(diphos-
phine)2 CO2 hydrogenation catalysts, ACS Catal., 2014,
4(10), 3663–3670.
15 C. A. Bessel, P. Aggarwal, A. C. Marschilok and
K. J. Takeuchi, Transition-Metal Complexes Containing
trans-Spanning Diphosphine Ligands, Chem. Rev., 2001,
101(4), 1031–1066.
693(12), 2111–2118; (b) W. J. Shaw, J. C. Linehan, 16 H. M. D. Bandara and S. C. Burdette, Photoisomerization
A. Gutowska, D. Newell, T. Bitterwolf, J. L. Fulton, Y. Chen
and C. F. Windisch, Synthesis and characterization of
in different classes of azobenzene, Chem. Soc. Rev., 2012,
41(5), 1809–1825.
a recoverable rhodium catalyst with a stimulus sensitive 17 (a) X.-J. Qi, Y. Fu, L. Liu and Q.-X. Guo, Ab initio calcu-
polymer ligand, Inorg. Chem. Commun., 2005, 8(10), 894–
896.
8 N. C. Gianneschi, P. A. Bertin, S. T. Nguyen, C. A. Mirkin,
L. N. Zakharov and A. L. Rheingold, A Supramolecular
Approach to an Allosteric Catalyst, J. Am. Chem. Soc., 2003,
125(35), 10508–10509.
9 L. Fabbrizzi, M. Licchelli and P. Pallavicini, Transition
metals as switches, Acc. Chem. Res., 1999, 32, 846–853.
10 M.-a. Haga, T. Takasugi, A. Tomie, m. Ishizuya, T. Yamada,
M. D. Hossain and M. Inoue, Molecular design of a proton-
induced molecular switch based on rod-shaped Ru dinuc-
lations fo thermodynamic hydricities of transition-metal
hydrides in acetonitrile, Organometallics, 2007, 26, 4197–
4203; (b) X.-J. Qi, L. Liu, Y. Fu and Q.-X. Guo, Ab Initio
Calculations of pKa Values of Transition-Metal Hydrides in
Acetonitrile, Organometallics, 2006, 25(25), 5879–5886;
(c) M. R. Nimlos, C. H. Chang, C. J. Curtis, A. Miedaner,
H. M. Pilath and D. L. Dubois, Calculated Hydride Donor
Abilities of Five-Coordinate Transition Metal Hydrides
[HM(Diphosphine) 2] +(M = Ni, Pd, Pt) as a Function of
the Bite Angle and Twist Angle of Diphosphine Ligands,
Organometallics, 2008, 27, 2715–2722.
lear complexes with bis-tridentate 2,6-bis(benzimidazol-2- 18 In addition to the expected dihydride species, in some
yl)pyridine derivatives, Dalton Trans., 2003, 2069–2079.
11 (a) G. Knor, Bionic catalyst design: a photochemical
approach to artificial enzyme function, ChemBioChem,
2001, 7/8, 593–596; (b) K. Okamoto, T. Yamamoto,
instances, an additional unidentified species was observed
along with the Rh-R-(trans)-Azb complex (before irradiation)
that slowly disappeared upon conversion to the cis con-
figuration with exposure to UV light (Fig. S7†).
M. Akita, A. Wada and T. Kanbara, Chemical stimuli 19 K. Issleib and G. Hohlfeld, Beiträge zur Komplexchemie
induced phosphorescence modulation of secondary thio-
der Phosphine und Phosphinoxyde. X. Schwermetallkom-
This journal is © The Royal Society of Chemistry 2015
Dalton Trans., 2015, 44, 14854–14864 | 14863

View Article Online
Paper
Dalton Transactions
plexe ditertiärer Phosphine, Z. Anorg. Allg. Chem., 1961, 312 21 D. M. Pfund, J. G. Darab, J. L. Fulton and Y. Ma, An
(3–4), 169–179.
XAFS Study of Strontium Ions and Krypton in
Supercritical Water, J. Phys. Chem., 1994, 98(50), 13102–
13107.
20 (a) K. Skopek, M. C. Hershberger and J. A. Gladysz,
Gyroscopes and the chemical literature: 1852–2002, Coord.
Chem. Rev., 2007, 251(13–14), 1723–1733; (b) K. Skopek, 22 J. Stewart, J. Mol. Model, 2007, 13, 1173–1213.
M. C. Hershberger and J. A. Gladysz, Gyroscopes and the 23 J. K. M. Vande Vondele, F. Mohamed, M. Parrinello,
chemical literature (1852–2002), Coord. Chem. Rev., 2007,
251, 1723–1733.
T. Chassaing and J. Hutter, Comput. Phys. Commun., 2005,
167, 103.
14864 | Dalton Trans., 2015, 44, 14854–14864
This journal is © The Royal Society of Chemistry 2015