In Situ Structural Determination of a Homogeneous Ruthenium Racemization Catalyst and Its Activated Intermediates Using X-Ray Absorption Spectroscopy

Article abstract of DOI:10.1002/chem.201905479

The activation process of a known Ru-catalyst, dicarbonyl(pentaphenylcyclopentadienyl)ruthenium chloride, has been studied in detail using time resolved in situ X-ray absorption spectroscopy. The data provide bond lengths of the species involved in the process as well as information about bond formation and bond breaking. On addition of potassium tert-butoxide, the catalyst is activated and an alkoxide complex is formed. The catalyst activation proceeds via a key acyl intermediate, which gives rise to a complete structural change in the coordination environment around the Ru atom. The rate of activation for the different catalysts was found to be highly dependent on the electronic properties of the cyclopentadienyl ligand. During catalytic racemization of 1-phenylethanol a fast-dynamic equilibrium was observed.

Full text of DOI:10.1002/chem.201905479

A Journal of
Accepted Article
Title: In Situ Structural Determination of a Homogeneous Ruthenium
Racemization Catalyst and its Activated Intermediates using X-
ray Absorption Spectroscopy
Authors: Jan-E. Bäckvall, Karl Gustafson, Arnar Guðmundsson, Eva
Bajnóczi, Ning Yuan, Xiaodong Zou, and Ingmar Persson
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201905479
Link to VoR: http://dx.doi.org/10.1002/chem.201905479
Supported by

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
In Situ Structural Determination of a Homogeneous Ruthenium
Racemization Catalyst and its Activated Intermediates using X-
ray Absorption Spectroscopy
Karl P. J. Gustafson,†,[a],[b],‡ Arnar Guðmundsson,†,[a] Éva G. Bajnóczi,†,[c],& Ning Yuan,[b],[c] Xiaodong
Zou,*,[b] Ingmar Persson*,[c] and Jan-E. Bäckvall*,[a]
Abstract: The activation process of
a
known Ru-catalyst,
specific bond lengths of transient metal intermediates using
methods such as single-crystal X-ray diffraction.
dicarbonyl(pentaphenylcyclopentadienyl)ruthenium chloride, has
been studied in detail using time resolved in situ X-ray absorption
spectroscopy. The data provide bond distances of the species
involved in the process as well as information about bond formation
and bond breaking. On addition of potassium tert-butoxide, the
catalyst is activated and an alkoxide complex is formed. The catalyst
activation proceeds via a key acyl intermediate, which gives rise to a
complete structural change in the coordination environment around
the Ru atom. The rate of activation for the different catalysts was
found to be highly dependent on the electronic properties of the
cyclopentadienyl ligand. During catalytic racemization of 1-
phenylethanol a fast dynamic equilibrium was observed.
X-ray absorption spectroscopy (XAS) is an element-specific
technique used to determine the local structure around an
absorbing element.[1] Unlike single-crystal X-ray diffraction, XAS
can be applied to systems in all forms of aggregation and
requires only an elemental concentration in the millimolar range.
In situ/operando XAS has predominantly been applied to
heterogeneous catalysts, and only limited examples of
homogeneous metal catalysts have been reported.[2–7] The
group of Berry utilized XAS in combination with other
spectroscopic techniques to strengthen the “push-pull”
mechanism in the bi-rhodium-catalysed carbene and nitrene
transfer reactions, and this mechanism was later confirmed by
the group of Fürstner.[8,9] In seminal work by the group of Lei on
copper-mediated cross-couplings, XAS was used to show that
Cu(I) species were the catalytically active intermediates in the
mixture, while Cu(II)- and Cu(0)-species acted as
spectators.[10,11] XAS has also been applied to other elements in
attempts to elucidate catalytic species.[5,12–16] In the few reports
that exist, to the best of our knowledge, the emphasis lies either
on the X-ray absorption near edge structure (XANES) in an
operando setup[17] or on accumulated intermediates.[18,19]
Alternatively, studies are performed on very simple systems with
high symmetry or under conditions that do not provide time-
resolved structural information of enough quality. Herein, we
report an in situ XAS that allows for the structural determination
of a transient metal intermediate in a ruthenium-catalysed
reaction.
Introduction
Knowledge about catalyst activation is crucial for the
development of new catalysts and catalytic processes. When
studying the activation mechanism of a transition metal complex
one is often limited to indirect methods. Reaction intermediates
are generally transient, but in some cases it is possible to
accumulate them for characterization by manipulating the
reaction parameters. However, it is rarely possible to obtain
[a]
Dr. Karl P. J. Gustafson, Arnar Guðmundsson, Prof. Jan-E.
Bäckvall*
Department of Organic Chemistry, Arrhenius Laboratory
Stockholm University
SE-106 91 Stockholm, Sweden
Metal-carbonyls are very common in modern coordination
and organometallic chemistry, and they are frequently found as
reagents or catalysts in organic synthesis.[20] Dicarbonyl
(pentaphenylcyclopentadienyl)ruthenium chloride (1a) is formally
a transfer hydrogenation catalyst, which has found numerous
applications as racemization catalyst in dynamic kinetic
E-mail: jeb@organ.su.se
[b]
[c]
Dr. Karl P. J. Gustafson, Dr. Ning Yuan, Prof. Xiaodong Zou*
Department of Materials and Environmental Chemistry, Arrhenius
Laboratory
Stockholm University
SE-106 91 Stockholm, Sweden
resolution (DKR) of
a
wide range of alcohols.[21–27] The
Dr. Éva G. Bajnóczi, Dr.Ning Yuan, Prof. Ingmar Persson*
Department of Molecular Sciences
mechanism for the hydride transfer of alcohols with catalyst 1
and related complexes has long been a topic of discussion and
dispute.[28–37] The current proposed mechanism starts with
activation of the Ru chloride complex 1 by potassium tert-
butoxide, to give tert-butoxide complex 3 (Scheme 1).[21,33,38]
Swedish University of Agricultural Sciences
P.O.Box 7015, SE-750 07 Uppsala, Sweden
Karl P. J. Gustafson, Arnar Guðmundsson and Éva G. Bajnóczi
contributed equally to this work.
Present address: Borregaard, P.O Box 162, N-1701, Sarpsborg,
Norway.
†
‡
Upon the addition of
a sec-alcohol, an alcohol-alkoxide
exchange takes place and a new alkoxide complex 4 is formed.
Since 4 is an 18-electron complex, a vacant site on Ru is
required for the subsequent β-hydride elimination step, where
the hydride is abstracted from the alkoxide. This vacant site was
&
Present address: Wigner Research Centre for Physics, Budapest,
Hungary
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
proposed to be generated through a CO dissociation on the
basis of DFT calculations,[32] and this pathway was later
confirmed by 13CO exchange studies.[39] The CO dissociation
enables the subsequent β-hydride elimination to occur, which
gives complex Int-1 (Scheme 1). Hydride re-addition to the
ketone followed by CO coordination would produce the racemic
ruthenium alkoxide complex 4. Release of the alcohol via
another alcohol-alkoxide exchange closes the catalytic cycle.
toluene.[33] Tetrahydrofuran (THF), previously utilized in a DKR
study with 1a,[43] is a solvent in which the metal complex 1a-c
has an increased solubility. A higher concentration of the
reactants provides an increased signal to noise ratio, which is
required when using the custom-made glass reactor in an in
situ/operando XAS setup. Initial studies on monitoring the
reaction in THF with in situ IR and 13C-NMR confirmed that the
structural changes follow the same pattern as those reported in
toluene[33] (IR spectrum is shown in Figure 1). The main
difference between the IR spectra in THF and toluene is that the
lifetime of intermediate 2a (characteristic peak at 1933 cm−1) is
significantly prolonged in THF compared to in toluene. In toluene,
full conversion to 3a took less than 5 min, whereas in THF the
same reaction took about 2 h. To confirm that the reaction
proceeds via the same mechanism in the two different solvents,
and that THF only stabilizes intermediate complex 2a, the
reaction rate was measured at varying concentrations. From
these studies it became clear that the chloride abstraction
follows first order kinetics (see Figure S1 and S2). The longer
lifetime in THF for anion 2a compared to that in toluene can be
explained by the stronger solvation of the charged intermediate
by the polar THF compared to the apolar toluene. When adding
the substrate, 1-phenylethanol, to the activated complex 3a, the
IR signals of the carbonyls of 3a slightly blue-shifted immediately
and the new complex 4a formed instantaneously (Figure S3).
Scheme 1. Proposed mechanism of catalyst 1a for racemization of sec-
alcohols.
In an early computational work, catalyst activation and
alcohol-alkoxide exchange were proposed to proceed via acyl
intermediate 2a (Scheme 1).[32] In
a subsequent report,
experimental evidence for this intermediate acyl complex during
activation of the catalyst was provided through the use of in situ
FT-IR spectroscopy under cryogenic conditions.[33] Low
temperature 13C-NMR was consistent with the proposed acyl
intermediate.[33] An XAS study of catalyst 1a and its
intermediates would provide important structural information of
the environment around the Ru, which has not been possible to
obtain before with previous methods.
To reach this goal a temperature and stirring controlled in
situ/operando reactor[40] has been applied in the present study.
The in situ/operando reactor was recently successfully used to
obtain structural information on in-house developed
heterogeneous catalysts in operando studies.[41,42]
The aim of the present work has been to study the activation
process of RuII catalysts 1a-c utilizing synchrotron radiation in an
in situ XAS setup using in situ/operando reactor.
Figure 1. Time resolved in situ IR spectrum of the activation of 1a by tert-
BuOK in THF, focused on 2090–1912 cm−1 over the time.
In the 13C-NMR recorded during the activation of catalyst 1a,
the characteristic carbonyl peaks of 2a were observed in THF-d8
at 209.0 and 208.3 ppm. This observation rules out the
Results and Discussion
possibility that 2a is
a
complex formed via simple Cl−
In situ IR
dissociation.
In the FT-IR spectrum of 2a in toluene[33] the acyl carbonyl
Activation of catalyst 1a as well as the in situ IR studies of
acyl intermediate 2a have previously only been performed in
peak was observed at 1596 cm−1. The 13C-NMR and the FT-IR
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
of 2a are not compatible with two equivalent carbon monoxide
ligands.
absorption edge. This observation verifies that the oxidation
state of the Ru catalyst is +II. The fact that the edge energy did
not change significantly neither during the activation, nor during
the catalysis suggests that the oxidation state remained
unchanged during the whole process.[46] However, a slight edge
shift of 0.6 eV is shown towards lower energies upon the
addition of the t-BuOK, while this shift disappears towards
reaching the activated state (see Figure S5). This slight edge
shift towards lower energies can be explained by the “loss” of
one CO ligand (transforms into an acyl group), therefore the
decrease of electron back-donation from the ruthenium, leading
to a slightly higher electron density around the ruthenium. The
XANES spectra of catalysts 1a, 1b and 1c show similar features
(see Figure 2 for 1a and 1b and Figure S7 for 1c), suggesting
that even though the electronic properties of the substituted
cyclopentadienyl (Cp) ligand change from electron-rich to
electron-deficient, the aryl groups on the Cp ligand are too far
from the RuII centre to influence its local geometry.
Cl− dissociation would in principle be possible to detect via Cl
K-edge measurements.[44] However, in light of the evidence
against Cl− dissociation (vide infra and supra), and the fact that
Cl K-edge EXAFS are not possible to perform to evaluate
distances around Cl at the concentrations available in the
systems under study as transmission experiments which are
required for a correct interpretation,[45] such experiments have
not been performed.
XANES
The XANES spectra of catalyst 1a in both toluene and THF
are identical, confirming that the local structure around Ru is
preserved regardless of the applied solvent (see Figure S6). The
K-edge was observed at 22125 eV, determined from the first
inflection point of the edge using the first derivative of the
Figure 2. a) The normalized XANES spectra of catalyst 1a dissolved in THF during the activation process recorded every 5 minutes after the addition of t-BuOK
(the measurements were started within 1 min after the addition of t-BuOK) and after the addition of 1-phenylethanol (added after 65 min) as substrate (dashed
line). b) The normalized XANES spectra of catalyst 1b dissolved in THF, during the activation process recorded every 5 minutes after the addition of t-BuOK (the
measurements were started within 1 min after the addition of t-BuOK) and after the addition of 1-phenylethanol (added after 50 min) as substrate (dashed line).
The XANES spectra showed an instantaneous change after
the addition of the activator potassium tert-butoxide (t-BuOK).
The stronger peak around 22135 eV showed a slight loss in
intensity as well as a shift towards higher energy. However, the
most noteworthy change in the XANES region is the complete
disappearance of the peak around 22158 eV after 5 min (Figure
2a, cf. spectrum 1a-0 min and 1a-5 min), which suggests
considerable alteration of the environment around Ru. This peak
regains its former shape after approximately 45 minutes, after
which no further change was observed in the XANES spectrum.
The intensity of the peak around 22158 eV never reached that of
the original catalyst After 65 min, 1-phenylethanol was added,
which led to no significant change of the XANES spectum
(dashed line, Fig. 2a), suggesting that the local structure of
complex 3a and 4a is very similar, but not necessarily identical
to that of the initial catalyst 1a.
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
The same features in the XANES spectra were observed for
catalyst 1b, although at an increased reaction rate compared to
that of 1a, with no more changes being observed after ~25
minutes (see Figure 2b). After 50 min, 1-phenylethanol was
added (dashed line). The results show that the activation
process is considerably faster when the aryl groups on the Cp
ligand are changed from C6H5 to the electron-donating 4-MeO-
C6H4, as in the case of 1b. On the other hand, electron-
withdrawing aryl groups on the Cp ligand, as in the case of 1c,
slow down the process significantly.[47] The spectra of catalyst 1c
collected at 5 minutes and after 24 h after starting the activation
process are almost identical (see Figure S7). The spectra at
5 minutes are almost identical for all three catalysts, implying
that an intermediate state is quickly formed, which is
subsequently transformed into the activated catalyst. The rate of
the latter step is strongly dependent on the electronic properties
of the substituted Cp ligand.
The characteristic features of XANES-spectra, and the
increasing intensity of the peak at 22160 eV for catalyst 1a and
1b can be used to extract rates and to calculate the approximate
half-life for the complexes during activation. Figure 3 shows the
normalized intensity values measured at this energy as a
function of time. From these curves, the half-life time (t1/2) of the
intermediates of the activation process was obtained and found
to be 28±6 and 6±0.5 minutes for catalysts 1a and 1b,
respectively. Hence the activation process for 1b is almost 5
times faster than that of catalyst 1a. Note that these values are
only estimated ones, as other geometrical changes also can
have an effect at this energy, but as the activation mechanism is
the same in all cases, they play the same role in all
measurements, allowing a viable comparison. The reaction with
catalyst 1a monitored by in situ IR show a similar rate as that
observed in the XANES measurements. The slight difference
between the two measurements may be due to that the
synchrotron irradiation changes the kinetics of the reaction.
Complex 1a dissolved in THF shows only a marginal pre-
edge structural change in the XANES spectrum (see Supporting
t
Information, Figure S9a). After addition of the activator, BuOK,
the pre-edge feature slightly shifts and becomes more distinct,
indicating that the symmetry decreases around the RuII centre
as intermediate 2a is formed. Interestingly, over time the pre-
edge feature decreases again, suggesting that the activated
complex 3a shows a higher symmetry than 2a. The same trend
is seen for catalyst 1b (see Supporting Information, Figure S9b).
X-ray absorption fine structure (EXAFS) studies
The coordination environment of ruthenium during the
course of the activation was further elucidated by analysing the
EXAFS spectra. The Fourier transformed (FT) EXAFS spectra of
the dissolved catalyst 1a in both THF and toluene (see Figure 4
and Figure S10, respectively), show no sign of change
compared to those of the solid structure, confirming that the
structure of the complex is maintained during the solvation
process (Supplementary Information Table S1). When the
catalyst is dissolved, its bond lengths are slightly longer than in
the solid state. Notably, the distances are a bit longer in THF
than in toluene, which is probably due to stronger solvation in
THF as it is a more polar solvent. The FT-spectrum of the
dissolved catalyst 1a in THF can be described with two distinct
peaks (see Figure 4). The most distinct peak at around 1.8 Å
(not phase corrected) corresponds to the interatomic distances
of the 2 C atoms of the CO’s, the 5 C atoms of the Cp ring, and
the Cl coordinated to RuII. The refined bond distances are
1.906 Å 2.240 Å and 2.436 Å, respectively (see Table 1). All
bond distances and their corresponding Debye-Waller
coefficients are summarized in Table 1. These bond distances
agree well with previously reported single crystal data of
complex 1a.Error! Bookmark not defined. The second distinct
peak at approximately 2.5 Å (not phase corrected) corresponds
to the Ru∙∙∙O and Ru–C–O multiple scatterings (MS) of the linear
carbon monoxides. These MS events are intensified due to the
focusing effect, which plays an important role when two atoms in
a coordinating ligand, or part of a larger ligand, form a linear or
close to linear M-L-L’ structural motif with the absorbing
atom.[48,49] The refined distances for the Ru∙∙∙O of the carbonyls
are 3.045 Å.
The spectra collected during the activation process of
catalyst 1a are given in Figure 4a. The spectra of the catalyst
were treated individually, and only representative spectra are
presented in Figure 4b, while their refined models are presented
in Table 1. It is clearly visible in the EXAFS spectra that after
only 5 minutes there is an immediate and prominent change in
the local structure around RuII (1a-5min in Figure 4a and b) as
well as an overall loss of intensity. The second main peak, in
particular, exhibited
a significant drop in intensity, which
indicates that the number of coordinating carbon monoxides has
decreased. The distinct separation between the two peaks has
also disappeared, implying that a new signal has appeared
around 2.7 Å. The best fitted model contains five Ru–C
distances with an average of 2.269 Å, corresponding to the Cp
ring, one Ru–Cl bond with a distance at 2.346 Å two Ru∙∙∙O
Figure 3. Normalized intensity values measured in the XANES at the
characteristic peak at 22160 eV as a function of time during the activation
process of catalyst 1a and 1b.
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
−
distances of around 2.697 Å and two short Ru–C bond distances
of around 1.859 Å (Table 1). However, the second major peak
can be fitted with the MS contribution from only one CO. These
distances can be interpreted in such a way that the six
coordination sites around RuII are occupied by the Cp ring
(hapticity of 5, three coordination sites on Ru), one Cl, one CO
and one acyl-ester as in intermediate 2a. The Debye-Waller
factors are only somewhat larger than the values of 1a, which
implies that there is one dominant species in the reaction system
after 5 minutes with a very minor contribution from a second
species.
The shortening of the Ru-Cl bond length from 2.44 to 2.35 Å
on going from 1a to 2a is most likely due to the fact that one CO
is reacted into a carboalkoxy group (CO2tBu). With this new
ligand the back-donation to Ru is decreased, which leads to a
significantly shorter (about 0.08 Å) Ru-Cl bond.. A similar
shortening affect was observed in the Ru-Cl bond distance at a
comparison of (S-dmso)3RuIICl3 , and O4RuCl2 complexes (see
more detailed discussion in the Supporting Information, Tables
S4a-c). Furthermore, no complex with the composition RuCl(O-
ligand)(CO)(Cp) have been reported indicating the labile
character of such complexes.
During the activation process, the most prominent peak
centered around 1.8 Å (not phase corrected) moves towards
slightly shorter distances. The second main peak, which
corresponds to the CO’s, increases in intensity and the distinct
separation between the two peaks reappears when the activated
state 3a is reached (1a-55min in Figure 4a and b). The main
peak has shifted a few tenths of an Ångström towards shorter
distances, which suggests that the chloride is lost during the
activation. The best fit for the active catalyst corresponds to a
coordination environment with the Cp ring, two COs, and the
tert-BuO bonded to the RuII (3a), see Table 1 for the refined
values.
Figure 4. a) Fourier transformed EXAFS spectra of catalysts 1a over time dissolved in THF (1a in THF), during the course of activation. b) Fourier transform of
the EXAFS spectra of catalyst 1a dissolved in THF (1a-0min), and at 5, and 55 minutes after the addition of the t-BuOK (1a-5 min, and 1a-55 min, respectively). c)
Fourier transform of the EXAFS spectra of the activated catalyst 1a (1a-55 min) and after the addition of 1-phenylethanol (1a-s). k-range 2 – 10 Å–1, no phase
correction (ΔR ~ 0.5 Å).
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
Figure 5. a) Fourier transform of the EXAFS spectra of catalyst 1b dissolved in THF (1b-0min), and at 5 and 45 minutes after the addition of t-BuOK (1b-5 min
and 1b-45 min, respectively). The inset shows Fourier transform of the EXAFS spectra of the activated catalyst 1b (1a-45 min) and after the addition of 1-
phenylethanol (1b-s). k-range 2 – 10 Å–1, no phase correction (ΔR ~ 0.5 Å). b) Fourier transform of the EXAFS spectra of catalyst 1c dissolved in THF (1c-0 min),
at 5 minutes and 24 hours after the addition of t-BuOK (1c-5 min and 1c-24 h, respectively), k-range 2 – 10 Å–1, no phase correction (ΔR ~ 0.5 Å).
1
2
3
4
5
6
7
8
9
Catalyst 1b, with an electron-donating cyclopentadienyl ring,
exhibits the same spectral patterns. Since the reaction of 1b is
significantly faster than that of 1a, the first spectrum collected
during the activation (Figure 5, 1b-5 min) is already a mixture of
species.
The results of the model fitting strengthen this assumption
(Supplementary Information Table S2). In order to obtain a good
fit, the use of a mixture of possible coordinating moieties is
necessary, including several factors like broken coordination
10 numbers of the chloride ions, the carbon monoxides, the acyl-
11 group and even the tert-BuO group. The refined Debye-Waller
12 factors are fairly large, which is consistent with a much faster
13 reaction leading to mixed species. After the reaction had
14 reached the activated complex 3b of catalyst 1b (Figure 5, 1b-45
15 min), it can be described with the exact same model as for
16 catalyst 1a (Table S2).
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
Table 1. The k space fitted models for the Ru K-edge EXAFS measurement preformed on catalyst 1a in THF and at different times during the reaction. Number of
distances (N), mean distances (R/Å) and Debye-Waller factor (σ2/Å2). The fitted parameters for the multiple scattering paths are omitted for the sake of clarity, and
they can be found in Table S1 in the Supporting Information.
1a
2a
3a
Structure
Bond
N
5
R
σ2
1a-0min (1a)
2.240(8)
0.0016(9)
2
2
1
5
1.906(5)
3.045(6)
2.436(8)
2.269(11)
0.0018(6)
0.0062(5)
0.0044(5)
0.0089(9)
1a-5min (2a)
2a
1
1.859(12)
3.007(18)
0.0125(18)
0.0093(9)
2b
2.697(8)
0.00361(8)
1
5
2.346(6)
0.0062(9)
1a-55min (3a)
2.278(15)
0.0039(12)
2
2
1
1.894(5)
3.025(12)
2.095(9)
0.0057(7)
0.0086(4)
0.0026(9)
[a] It is not possible to distinguish between the two Ru-C bond distances. The value given is an average of the two Ru-C distances.
[b] It is not possible to distinguish between the two Ru-O distances. The value given is an average of the two Ru-O distances.
During the course of the synchrotron experiments, pre-
catalyst 1c did not reach the active state 3c. It seems that the
whole activation process is arrested at the intermediate state 2c
as shown in Figure 5b. Even after 24 h there is no significant
change in the structure around RuII (Figure 5b and
Supplementary Information Table S3).
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
Experimental methods XAS
Substrate addition
The Ru K-edge EXAFS measurements of catalyst 1a in the solid
state and dissolved in toluene were performed at a wiggler-based beam
line 4-1 at the Stanford Synchrotron Radiation Lightsource (SSRL),
Stanford, USA, operating at 3.0 GeV and 300 mA (top-up mode). The X-
rays were monochromatized by a Si[220] double crystal monochromator.
The second monochromator crystal was detuned to reflect 75% of the
maximum intensity at the end of the scans in order to minimize the higher
order harmonics. All the other ex situ as well as in situ spectra were
collected at the undulator based P64 beam line, equipped with a Si[311]
double crystal monochromator, at the Petra III Extension, Hamburg,
Germany, operating at 6.0 GeV and 100 mA (top-up mode). The energy
was internally calibrated in both cases with a metallic Ru foil, with the first
inflection point of the K-edge spectrum assigned as 22117 eV.[50]
1-Phenylethanol was added to activated complexes 3a and
3b, as substrate for the racemization. Following the addition, the
XANES spectra of 4a and 4b resembled each other (Figure 2,
dashed lines), as did their respective Fourier transforms (Error!
Reference source not found.4c and inset in Figure 5a). When
fitting the refined data, it becomes apparent that the hexa-
coordinated structure is maintained in an octahedral
configuration around RuII and the coordination of the Cp ring and
the two CO ligands are preserved during catalysis. The last
coordination site cannot be identified with certainty. It is most
probably a loosely bound oxygen based on the XANES
spectrum, although the EXAFS contribution is on the level of
noise. The lack of clear identification of the alkoxide ligand in
complex 4a-c suggests that the systems are in dynamic
equilibria with fast interconversions of substrate molecules
For the ex situ measurements, three spectra were collected per
sample and they were averaged after energy calibration, while the
spectra in the in situ measurements were treated individually. If not noted
differently, the scanning rate was 5 min/scan in the -200 - +800 eV
energy range around the Ru K-edge. The measurement of the solid was
performed in transmission mode, while the solution was detected in
fluorescence mode using a Lytle detector. The 1a catalyst in solid state
measured at SSRL was diluted with approximately the same amount of
boron nitride (BN, Merck), while the toluene solution was contained in a
sample made of titanium frame, a Teflon spacer and 6 mm polypropylene
X-ray film as windows. The measurements at Petra III were performed in
a custom made temperature controlled reactor with a glass test tube with
1 mm thick glass walls with the beam hitting as perpendicular as possible.
The measurements were performed at ambient room temperature, ca
25°C. The reactor is described in detail elsewhere.[40] At Petra III, all
spectra were collected in transmission mode using ion-chambers as
detectors filled with appropriate gas mixtures. All data treatment (energy
calibration, averaging, pre-edge subtraction, spline fitting and removal,
normalization and Fourier transformation) were performed with
EXAFSPAK program package.[51] The k3-weighted experimental data
were fitted by refining the structural parameters using Marquardt non-
linear least-square fitting algorithm within the EXAFSPAK package. The
following parameters were refined: Mean interatomic distance (R),
Debye-Waller factor coefficients (σ2), amplitude reduction factor (S02),
and the threshold energy (E0). The number of neighbouring atoms were
held constant during the fitting process due to strong correlation between
the number of distances and Debye-Waller coefficients. The theoretical
phases and amplitudes were calculated with the FEFF7 program.[52]
Conclusions
In conclusion, XAS was used to obtain detailed structural
information in solution of catalysts 1a-c, the proposed acyl
intermediates 2a-c, as well as the activated alkoxide complexes
3a and 3b. Different rates of activation were observed for the
different catalysts, which was highly dependent on the electronic
properties of the Cp ligand. In the case of 1c, the strong
electron-withdrawing character of the Cp ligand inhibited the
activation process. During the catalytic racemization of 1-
phenylethanol,
a fast dynamic equilibrium is established
between the two enantiomers of 4a-b and the ketone int-1
resulting in less resolved spectra for 4a and 4b. Therefore, only
an average structure was obtained for these compounds.
Information gained from this study may aid in the
development of new catalysts and to understand the different
modes of activation of different catalysts.
Experimental Section
General experimental procedure
The standard deviations for the refined parameters were obtained
from k3-weighted least-squares refinements of the EXAFS function (k),
and did not include systematic errors of the measurements. These
statistical error values allowed reasonable comparisons e.g. of the
significance when comparing relative shifts in distances. However, the
variations in the refined parameters, including the shift in the E0 value (for
which k = 0), using different models and data ranges, indicated that the
absolute accuracy of the distances given for the separate complexes is
within 0.005 to 0.02 Å for well-defined interactions. The “standard
deviations” given in the text have been increased accordingly to include
estimated additional effects of systematic errors.
1H-NMR and 13C-NMR were recorded at 400 MHz and 125 MHz,
respectively. Chemical shifts are reported in ppm, using the residual
solvent peak in CDCl3 (H 7.26 or C 77.16 ) or THF-d8 (H 3.58 or C
67.21 ) as internal standard. All solvents were dried prior to use, stored
over activated 4Å molecular sieves under 6.0 argon. Chemicals other
than the catalysts were purchased from commercial sources and dried
before use. Reactions were monitored using aluminium-packed plates
(1.5 Å, 5 cm) which are pre-coated with silica gel (0.25 mm). UV light or
KMnO4 was used for visualization. Column chromatography was
performed using silica gel (particle size 40-63 m, mesh size 230-400,
pore size 60 Å). Reactions at the synchrotron were performed in a
glovebox under argon atmosphere. In situ IR was measured using
ReactIR with a zirconium probe in a Schlenk flask which allowed easy
connection with the ReactIR probe using a NS 19 fitting and the reactions
were performed under 6.0 argon. Ruthenium catalysts 1a-c were
prepared according to a literature procedure.[47] Caution should be taken
when distilling the dicyclopentadiene for the synthesis of the ligand for 1c.
General procedure for in situ XAS
A flame dried vial was placed while hot directly into antechamber of
the glovebox and placed under vacuum, then transferred into the
glovebox after cooling down. Catalyst 1 (0.10 mmol) was dissolved in
1.8 ml of dry THF and the reaction vial was sealed. XAS spectra were
collected on catalyst 1, the reaction vial was placed back in the glovebox
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
and t-BuOK (1 M in THF, 100 L, 0.1 mmol) was added. The reaction
vessel was placed back into the beam and the structural changes were
monitored over time.
[6]
[7]
[8]
G. J. Sherborne, B. N. Nguyen, Chem. Cent. J. 2015, 9, 37.
D. Koziej, S. DeBeer, Chem. Mater. 2017, 29, 7051–7053.
K. P. Kornecki, J. F. Briones, V. Boyarskikh, F. Fullilove, J.
Autschbach, K. E. Schrote, K. M. Lancaster, H. M. L. Davies, J. F.
Berry, Science (80-. ). 2013, 342, 351–354.
Experimental procedure for in situ IR measurement of the activation
of 1a
[9]
C. Werlé, R. Goddard, A. Fürstner, Angew. Chemie Int. Ed. 2015,
54, 15452–15456.
Ruthenium catalyst 1a (203 mg, 0.32 mmol) was dissolved in 5 ml of
dry THF at room temperature using a flame dried 100 mL Schlenk flask
[10]
[11]
[12]
Q. Lu, J. Zhang, P. Peng, G. Zhang, Z. Huang, H. Yi, J. T. Miller, A.
Lei, Chem. Sci. 2015, 6, 4851–4854.
under argon and a series of spectra were recorded (2049 and 2001 cm−1
)
before addition of t-BuOK (1 M in THF, 320 L, 0.32 mmol) and the
activation was followed in an operando fashion with 15 sec scan interval.
Upon addition of the t-BuOK the solution turned red and the two previous
CO signals disappeared, with a new dominant CO signal at 1933 cm−1
being observed from complex 2a. Gradually over time, the activated
complex 3a could be observed with the corresponding CO signals at
2021 and 1964 cm−1. After the complete disappearance of 2a, the
substrate, 1-phenylethanol, was added (100 L, 0.82 mmol) and alkoxide
3a’ formed instantaneously without an apparent intermediate (CO peaks
2026 and 1971 cm−1).
C. He, G. Zhang, J. Ke, H. Zhang, J. T. Miller, A. J. Kropf, A. Lei, J.
Am. Chem. Soc. 2013, 135, 488–493.
J. Rabeah, M. Bauer, W. Baumann, A. E. C. McConnell, W. F.
Gabrielli, P. B. Webb, D. Selent, A. Brückner, ACS Catal. 2013, 3,
95–102.
[13]
A. S. K. Hashmi, C. Lothschütz, M. Ackermann, R. Doepp, S.
Anantharaman, B. Marchetti, H. Bertagnolli, F. Rominger, Chem. - A
Eur. J. 2010, 16, 8012–8019.
[14]
[15]
H. Asakura, T. Shishido, T. Tanaka, J. Phys. Chem. A 2012, 116,
4029–4034.
A. R. Corcos, O. Villanueva, R. C. Walroth, S. K. Sharma, J. Bacsa,
K. M. Lancaster, C. E. MacBeth, J. F. Berry, J. Am. Chem. Soc.
2016, 138, 1796–1799.
Acknowledgements
Financial support from The Swedish Research Council (2016-
03897, 2017-04321), the Berzelii Center EXSELENT, and the
Knut and Alice Wallenberg Foundation (CATSS, KAW
2016.0072) is gratefully acknowledged. Parts of this research
were carried out at PETRA III at Deutsches Elektronen-
Synchrotron (DESY), a member of the Helmholtz Association
(HGF). We would like to thank Vadim Murzin and Wolfgang
Caliebe for assistance with using the P-64 beamline. Another
part of this research was performed at the Stanford Synchrotron
Radiation Lightsource, SLAC National Accelerator Laboratory, is
supported by the U.S. Department of Energy, Office of Science,
Office of Basic Energy Sciences under Contract No. DE-AC02-
76SF00515. The SSRL Structural Molecular Biology Program is
supported by the DOE Office of Biological and Environmental
Research, and by the National Institutes of Health, National
Institute of General Medical Sciences (including P41GM103393).
The contents of this publication are solely the responsibility of
the authors and do not necessarily represent the official views of
NIGMS or NIH. We would also like to express our appreciation
to Niclas Heidenreich, A. Ken Inge and Sebastian Leubner for
their assistance with the reactor.
[16]
[17]
[18]
S. K. Sharma, P. S. May, M. B. Jones, S. Lense, K. I. Hardcastle, C.
E. MacBeth, Chem. Commun. 2011, 47, 1827–1829.
D. Lebedev, Y. Pineda-Galvan, Y. Tokimaru, A. Fedorov, N. Kaeffer,
C. Copéret, Y. Pushkar, J. Am. Chem. Soc. 2018, 140, 451–458.
D. Moonshiram, J. W. Jurss, J. J. Concepcion, T. Zakharova, I.
Alperovich, T. J. Meyer, Y. Pushkar, J. Am. Chem. Soc. 2012, 134,
4625–4636.
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
Y. Pushkar, D. Moonshiram, V. Purohit, L. Yan, I. Alperovich, J. Am.
Chem. Soc. 2014, 136, 11938–11945.
M. Zhou, L. Andrews, C. W. Bauschlicher, Chem. Rev. 2001, 101,
1931–1962.
B. Martín-Matute, M. Edin, K. Bogár, J.-E. Bäckvall, Angew. Chemie
Int. Ed. 2004, 43, 6535–6539.
B. Martín-Matute, M. Edin, K. Bogár, F. B. Kaynak, J.-E. Bäckvall, J.
Am. Chem. Soc. 2005, 127, 8817–8825.
R. Lihammar, R. Millet, J.-E. Bäckvall, Adv. Synth. Catal. 2011, 353,
2321–2327.
R. Lihammar, R. Millet, J.-E. Bäckvall, J. Org. Chem. 2013, 78,
12114–12120.
R. Lihammar, J. Rönnols, G. Widmalm, J.-E. Bäckvall, Chem. - A
Eur. J. 2014, 20, 14756–14762.
Keywords: in situ EXAFS spectroscopy• homogeneous
catalysis • racemization • ruthenium
A. Trꢀff, R. Lihammar, J.-E. Bꢀckvall, J. Org. Chem. 2011, 76,
3917–3921.
S. Y. Lee, J. M. Murphy, A. Ukai, G. C. Fu, J. Am. Chem. Soc. 2012,
134, 15149–15153.
[1]
[2]
P. Glatzel, U. Bergmann, Coord. Chem. Rev. 2005, 249, 65–95.
S. Bordiga, E. Groppo, G. Agostini, J. A. van Bokhoven, C. Lamberti,
Chem. Rev. 2013, 113, 1736–1850.
C. P. Casey, S. W. Singer, D. R. Powell, R. K. Hayashi, M. Kavana,
J. Am. Chem. Soc. 2001, 123, 1090–1100.
[3]
[4]
[5]
J. Singh, C. Lamberti, J. A. van Bokhoven, Chem. Soc. Rev. 2010,
39, 4754.
J. S. M. Samec, A. H. Éll, J. B. Åberg, T. Privalov, L. Eriksson, J.-E.
Bäckvall, J. Am. Chem. Soc. 2006, 128, 14293–14305.
C. P. Casey, T. B. Clark, I. A. Guzei, J. Am. Chem. Soc. 2007, 129,
11821–11827.
S. Bordiga, F. Bonino, K. P. Lillerud, C. Lamberti, Chem. Soc. Rev.
2010, 39, 4885.
S. N. MacMillan, K. M. Lancaster, ACS Catal. 2017, 7, 1776–1791.
B. Stewart, J. Nyhlen, B. Martín-Matute, J.-E. Bäckvall, T. Privalov,
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
Dalt. Trans. 2013, 42, 927–934.
[43]
L. Borén, B. Martín-Matute, Y. Xu, A. Córdova, J.-E. Bäckvall, Chem.
- A Eur. J. 2006, 12, 225–232.
[32]
[33]
[34]
[35]
J. Nyhlen, T. Privalov, J.-E. Bäckvall, Chem. - A Eur. J. 2009, 15,
5220–5229.
[44]
[45]
L. Yan, R. Zong, Y. Pushkar, J. Catal. 2015, 330, 255–260.
I. Persson, M. Trublet, W. Klysubun, J. Phys. Chem. A 2018, 122,
7413–7420.
J. B. Åberg, J. Nyhlꢁn, B. Martꢂn-Matute, T. Privalov, J.-E. Bꢀckvall,
J. Am. Chem. Soc. 2009, 131, 9500–9501.
M. C. Warner, O. Verho, J. E. Bäckvall, J. Am. Chem. Soc. 2011,
133, 2820–2823.
[46]
[47]
K. Mori, M. Kawashima, K. Kagohara, H. Yamashita, J. Phys. Chem.
C 2008, 112, 19449–19455.
G. A. I. Moustafa, Y. Oki, S. Akai, Angew. Chemie Int. Ed. 2018, 57,
10278–10282.
O. Verho, E. V. Johnston, E. Karlsson, J.-E. Bäckvall, Chem. - A Eur.
J. 2011, 17, 11216–11222.
[36]
[37]
D. G. Gusev, D. M. Spasyuk, ACS Catal. 2018, 8, 6851–6861.
B. L. Conley, M. K. Pennington-Boggio, E. Boz, T. J. Williams,
Chem. Rev. 2010, 110, 2294–2312.
[48]
[49]
[50]
P. A. Lee, J. B. Pendry, Phys. Rev. B 1975, 11, 2795–2811.
C. A. Ashley, S. Doniach, Phys. Rev. B 1975, 11, 1279–1288.
A. C. Thompson, J. Kirz, D. T. Attwood, E. M. Gullikson, M. R.
Howells, J. B. Kortright, Y. Liu, A. L. Robinson, H. Underwood,
James, G. P. Williams, et al., Xdb-New.Pdf, Berkeley, California,
2009.
[38]
[39]
[40]
[41]
[42]
B. Martín-Matute, J. B. Åberg, M. Edin, J.-E. Bäckvall, Chem. - A
Eur. J. 2007, 13, 6063–6072.
M. C. Warner, O. Verho, J.-E. Bꢀckvall, J. Am. Chem. Soc. 2011,
133, 2820–2823.
[51]
[52]
G. N. George, I. J. Pickering, — a suite of computer programs for
analysis of X-ray absorption spectra. Stanford Synchrotron
Radiation Laboratory, Stanford, Ca, USA, 2000.
N. Heidenreich, U. Rütt, M. Köppen, A. K. Inge, S. Beier, A.-C.
Dippel, R. Suren, N. Stock, Rev. Sci. Instrum. 2017, 88, 104102.
A. Bruneau, K. P. J. Gustafson, N. Yuan, C.-W. Tai, I. Persson, X.
Zou, J.-E. Bäckvall, Chem. - A Eur. J. 2017, 23, 12886–12891.
N. Yuan, V. Pascanu, Z. Huang, A. Valiente, N. Heidenreich, S.
Leubner, A. K. Inge, J. Gaar, N. Stock, I. Persson, et al., J. Am.
Chem. Soc. 2018, 140, 8206–8217.
S. I. Zabinsky, J. J. Rehr, A. Ankudinov, R. C. Albers, M. J. Eller,
Phys. Rev. B 1995, 52, 2995–3009.
This article is protected by copyright. All rights reserved.

10.1002/chem.201905479
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Text for Table of Contents
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
((Insert TOC Graphic here: max.
width: 5.5 cm; max. height: 5.0 cm))
Layout 2:
FULL PAPER
Karl P. J. Gustafson, Arnar
Guðmundsson, Éva G. Bajnóczi, Ning
Yuan, Xiaodong Zou, Ingmar Persson*
and Jan-E. Bäckvall*
Page No. – Page No.
In Situ Structural Determination of a
Homogeneous Ruthenium
Racemization Catalyst and its
Activated Intermediates using X-ray
Absorption Spectroscopy
The activation process of a known Ru-catalyst, dicarbonyl(pentaphenylcyclopenta-
dienyl)ruthenium chloride, is studied in detail using time resolved in situ X-ray
absorption spectroscopy. The data provide bond distances of the species involved
in the process such as the acyl type key intermediate as well as information about
bond formation and bond breaking during the activation process.
This article is protected by copyright. All rights reserved.