How a Thermally Unstable Metal Hydrido Complex Can Yield High Catalytic Activity Even at Elevated Temperatures

Article abstract of DOI:10.1002/chem.201601641

Despite their instability in ethereal solvents, organotitanium hydride catalysts are successfully employed in catalysis at moderate to high temperatures (110 °C), even in the presence of alcohols. It is shown computationally (bond dissociation energy (BDE) analysis and energetic profile for regeneration) and experimentally (EPR studies and kinetic studies), with the specific example of hydrodefluorination (HDF), that despite the long standing belief, regeneration of Ti?H bonds from Ti?F bonds using silanes is endergonic. The resulting low concentration of Ti?H species is crucial for the catalytic stability of those systems. The resting state in the catalysis is a Ti?F species. The most promising silanes for regeneration are not the ones that have the strongest Si?F bond, but the ones that show the largest difference in Si?F and Si?H BDEs.

Full text of DOI:10.1002/chem.201601641

DOI: 10.1002/chem.201601641
Full Paper
&
Titanium Catalysis |Hot Paper|
How a Thermally Unstable Metal Hydrido Complex Can Yield High
Catalytic Activity Even at Elevated Temperatures
Christian Ehm,[b] Juliane Krüger,[a] and Dieter Lentz*[a]
Abstract: Despite their instability in ethereal solvents, orga-
notitanium hydride catalysts are successfully employed in
catalysis at moderate to high temperatures (1108C), even in
the presence of alcohols. It is shown computationally (bond
dissociation energy (BDE) analysis and energetic profile for
regeneration) and experimentally (EPR studies and kinetic
studies), with the specific example of hydrodefluorination
(HDF), that despite the long standing belief, regeneration of
TiÀH bonds from TiÀF bonds using silanes is endergonic.
The resulting low concentration of TiÀH species is crucial for
the catalytic stability of those systems. The resting state in
the catalysis is a TiÀF species. The most promising silanes for
regeneration are not the ones that have the strongest SiÀF
bond, but the ones that show the largest difference in SiÀF
and SiÀH BDEs.
Introduction
In the case of hydrosilylation and HDF, synthetic procedures
have been published utilizing TiIV-difluorides as catalyst precur-
sors for catalytically active TiIII-hydrides; activation is achieved
using a variety of silanes.[5] Cp2TiH (1) or Cp2TiH(THF) (1-THF)
has been described as the active species for hydrosilylation or
HDF at temperatures up to 110 8C.[5] Contrary to this, Bercaw
and Brintzinger observed ether cleavage in 1-THF even at
À308C.[2]
Organotitanium hydrides represent a highly reactive class of
compounds thought to be the active species in various catalyt-
ic processes such as silane dehydrocoupling, olefin hydrogena-
tion, hydrosilylation, and hydrodefluorination (HDF).[1]
The reactivity of such Lewis acidic compounds strongly de-
pends on their molecular structure. Titanocene hydride (1)
forms m-H bridged dimers in the solid state as well as in solu-
tion (at ca. À708C).[2] By warming a THF solution of (Cp2TiH)2
to À308C, formation of monomeric Cp2TiH(THF) is observed,
which decays rapidly via ethereal cleavage yielding the corre-
sponding alkoxy species. Decomposition of (Cp2TiH)2 occurs
even in aliphatic solvents like hexane.[2] The fluorinated deriva-
tive Cp2TiF, in contrast, is stable in solution at room tempera-
ture, and forms a trimer with a coplanar cyclic (TiÀF)3 unit.[3] It
crystallizes in the presence of THF without adduct formation.
Several other terminal titanium(III) hydrides bearing sterically
demanding Cp* or Cp’ (Cp’=(1,3-Me3C)2C5H3 or C5Me4Ph) li-
gands and/or possessing a Lewis base in the coordination
sphere of the metal are known, are thermally stable, and com-
monly exist in monomeric form.[4] Only 1 and other bis-cyclo-
pentadienyl derivatives have been successfully employed in
catalysis, however.
In the specific case of HDF, stoichiometric activation of
strong CÀF bonds with metal hydrides can be achieved with
early as well as late transition metals (TM).[6] Catalytic CÀF
bond activation is, however, often hampered by strong transi-
tion metal–fluorine bonds affecting the regeneration of the
catalyst. Silanes proved to be very effective in this regard for
both early and late TM; this is somewhat surprising because
the silane could also be dehydrocoupled or polymerized by
the catalyst.[1] In 2010, we reported the first Ti-catalyzed HDF
of fluoroalkenes (Scheme 1, updated catalytic cycle).[5] Subse-
quently, we gained valuable insight into the regioselectivity of
alkene HDF[3], as well as the regioselectivity and mechanism of
arene HDF for 1.[7] Several experimental observations indicated
that the active species is indeed 1-THF.[8] The discrepancy be-
tween the alleged stability of 1-THF under catalytic conditions
and its instability when prepared separately, point to catalyst
regeneration playing a vital role.
[a] J. Krüger, Prof. Dr. D. Lentz
Bond dissociation energies are the foundation of what is
commonly known as ‘chemical intuition’. Accurate knowledge
of BDEs allows the prediction of reaction energetics, and thus
reactivity. Strong bonds are often viewed as inert or ‘hard to
activate,’ and the formation of strong bonds is seen as a poten-
tial driving force for reactions. Subsequently, it is commonly ac-
cepted that the driving force for catalyst regeneration in HDF
reactions can be the SiÀF bond strength, as SiÀF bonds are
stronger than most metal fluoride bonds.[9] Employing silanes
as hydride sources can convert Ti-F species to hydrides, thus
Institut für Chemie und Biochemie, Anorganische Chemie
Freie Universität Berlin
Fabeckstr. 34–36, 14195 Berlin (Germany)
E-mail: dieter.lentz@fu-berlin.de
[b] Dr. C. Ehm
Dipartimento di Scienze Chimiche
Università di Napoli Federico II
Via Cintia, 80126 Napoli (Italy)
Supporting information and the ORCID identification number for the corre-
sponding author of this article are available on the WWW under http://
dx.doi.org/10.1002/chem.201601641.
Chem. Eur. J. 2016, 22, 9305 – 9310
9305
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Calculated and experimentally available SiÀH, SiÀF, TiÀH and TiÀ
F bond dissociation energies for silanes and titanium species in kcal
molÀ1
.
SiÀH
R=
R2HSiÀH
Ph
R2FSiÀH
R3SiÀH
Me
91.1
Me
Ph
Me
91.9
EtO
calc.
exp.
88.3
90.6Æ1.6
93.5
91.6
95.6
93.5Æ1.2
R2FSiÀF
160.4 157.7
94.7Æ1.0
SiÀF
calc.
exp.
R2HSiÀF
R3SiÀF
156.6
159.9Æ4.8
153.8
150.4
162.5
Scheme 1. Catalytic cycle for HDF of fluorinated substrates, including the im-
portance of endergonic regeneration. Ph2SiHF (2a), Ph2SiF2 (2b), hexafluoro-
propene (3), 1,2,3,4,5-pentafluorotoluene (5), perfluorotoluene (6), and their
corresponding HDF products E(3a)- and Z(3b)- pentafluoropropene, 1,2,4,5-
tetrafluoro-3-methyltoluene (5a) and 1,2,4,5-tetrafluoro-3-(trifluoromethyl)-
benzene (6a). In donor solvents 1 exists as 1-donor and 4 as 4-donor.
TiIII
Cp2TiÀH
71.7
Cp2TiÀF
150.3
F2TiÀF
138.0
~138[a]
calc.
exp.
TiIV
Cp2HTiÀH
50.3
Cp2FTiÀH
39.4
Cp2HTiÀF
117.8
Cp2FTiÀF
94.6
F3TiÀF
116.8
109Æ5
calc.
exp.
[a] Ref. [9] gives a value of 155Æ5 kcalmolÀ1 based on Ref. [11]. The latter
yielding catalytic activity. In the following, we will show that
based on BDE arguments, regeneration of TiÀH bonds via si-
lanes is endergonic, and we will present computational and ex-
perimental evidence for the importance of a TiÀF resting state
in HDF catalysis using 1.
reference uses an outdated value for DHf(TiF3)=338.1 kcalmolÀ1 to calcu-
late the BDE, while the latest NIST recommendation is 283 kcalmolÀ1 [12]
.
This lead to a severe overestimation of the TiIIIÀF bond energy in Ref. [9]
and [11].
molÀ1 for the first H–F exchange (secondary silanes), resulting
in DEBDE(Regeneration) of +16 kcalmolÀ1 [Eq. (1c)].
Results and Discussion
BDE analysis
DEBDE ¼jDETijÀjDESij
The driving force of a reaction of the type depicted in Equa-
tion (1a) is, however, the difference in relative BDEs of the
pairs EX/EY and E’X/E’Y (E, E’, X, Y=different groups or atoms),
that is, the difference between the energies of the bonds that
are broken and newly formed.
¼ 78:4 kcal molÀ1À62:0 kcal molÀ1
¼ 16:4 kcal molÀ1
ð1cÞ
Subsequently, the regeneration step of the catalytic cycle is
endergonic, not exergonic as previously assumed.[13] In con-
trast, regeneration for late TM HDF cycles is exergonic, as re-
ported by Macgregor and Whittlesey.[14] The overall catalytic
cycle is driven by the transformation of CÀF into CÀH and SiÀF
into SiÀH bonds since jDEC j ! jDESi j. The best substrates for
regeneration are the ones that show the largest difference in
EÀF and EÀH BDEs, not necessarily the ones that have the
strongest EÀF bonds. Alanes, for example, have weaker EÀH
bonds but almost equally strong EÀF bonds compared to si-
lanes, and have been successfully used in catalytic HDF reac-
tions.[9,15] Finally, it appears that electron-withdrawing substitu-
ents on silicon increase the energy difference between SiÀF
and SiÀH bonds (Table 2). This is supported by reports of Buch-
wald that addition of MeOH to hydrosilylation reactions of
1 enhance the rate due to formation of (MeO)xSiÀH species.[5b]
EXþE0Y ! EYþE0X
ð1aÞ
For the specific example considered here, the difference
(DEBDE) in relative TiÀH/TiÀF (jDETi j) versus SiÀH/SiÀF (jDESi j)
bond strengths has to be considered [Eq. (1b)].
DEBDE¼ BDETiÀFþBDESiÀHÀBDETiÀHÀBDESiÀF
¼ ðBDETiÀFÀBDETiÀHÞþðBDESiÀHÀBDESiÀF
¼ jDETijÀjDESij
Þ
ð1bÞ
Experimental titanium hydride bond strengths are unavaila-
ble. Therefore, we calculated the BDE data for several silanes
and Ti complexes using DFT (gasphase; Table 1).[10] Available
experimental BDE data for SiÀH, SiÀF, and TiÀF bonds are in
good agreement with DFT predictions. Note that the only
available experimental BDE for TiF3 is flawed owing to the use
of an outdated value in the original BDE calculation.[9,11,12] We
re-estimated the BDE (see Table 1).
Table 2. BDE differences for SiÀF/SiÀH substitution jDEBDE j in kcalmolÀ1
.
Interestingly, for silane/titanium(III) H–F exchange, BDE con-
siderations predict a highly endergonic reaction. The TiÀF
bond is much stronger than the TiÀH bond (by ~78 kcalmolÀ1),
while for SiÀF/SiÀH the difference amounts to ‘only’ ~62 kcal
R3SiÀ or
R2XSiÀ
Me2HSi
Ph2HSi
Me2FSi
Ph2FSi
Me3Si
EtO3Si
jDEBDE
j
62.7
62.1
66.9
66.1
64.7
66.9
Chem. Eur. J. 2016, 22, 9305 – 9310
www.chemeurj.org
9306
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Fully in line with the straightforward activation of 1a using
silanes, TiIV-mono and -dihydrides are predicted to have even
weaker TiÀH bonds than 1.[16] The TiÀF bond in TiF3 is much
stronger than in TiF4 (Table 1). Similarly, for experimentally
available TiÀC BDEs in Cp2TiR2 (R=Me, Ph) systems, the second
TiÀC dissociation energy (the TiIII bond strength) is also always
higher; the same goes for TiÀCl bonds in TiCl4.[17]
Figure 2. Molecular structure (left) and space filling model (right)[18] of 4-
trimer emphasising the shielded Ti environment. Based on data in Ref. [3].
Energetic profile of Ti-H regeneration
DFT predicts regeneration of the catalyst (4 to 1) to be sub-
stantially endergonic [Eqs. (2) and (3)] in solution (PCM model,
298 K) for both Me2SiH2 (computational model) as well as 2.
partially compensates the entropic penalty of forming the re-
generation TS. The amount of 1-THF at any given point is ex-
ceedingly low. Catalytic turnover is achieved only by HDF.[20]
Catalyst decomposition of 1-THF through THF ring opening
is predicted to have a sizable but accessible barrier of 33.2 kcal
molÀ1 with no entropic penalty, as THF is the solvent. The bar-
rier is significantly higher from 4-trimer (52.2 kcalmolÀ1), ex-
plaining the stability of the catalytic system over prolonged
times at high temperatures. HDF barriers for fluorinated sub-
strates are strongly substrate dependent, but decomposition
via THF ring opening can become competitive for less reactive
substrates.
2þ4 ! 1þ2 a DG ¼ þ15:9 kcal molÀ1
ð2Þ
ð3Þ
Me2SiH2þ4 ! 1þMe2SiHF DG ¼ þ14:3 kcal molÀ1
The reverse barrier (F–H exchange) is substantially smaller
(10 kcalmolÀ1) and regeneration is reversible. The energetic
profile for regeneration with Me2SiH2 is shown in Figure 1. The
trimer of 4 (4-trimer) is predicted to be the resting state (co-
planar cyclo-(TiÀF)3). The crystal structure as well as the space
filling model of 4-trimer is shown in Figure 2.
Recently, Raza and Braun reported reversible SiÀF bond for-
mation at FSi(OEt)3 promoted by RhI complexes.[21] In light of
our results, this indicates that reversibility in SiÀF bond forma-
tion might be more common than previously assumed, albeit
one of the strongest element–fluorine bonds known. The ki-
netic profile is sensitive to concentration changes, and a more
realistic profile for initial catalytic concentrations of catalyst
and substrates is also shown in Figure 1.
H–F exchange via aSBM TS at 4-monomer with pentacoordi-
nated Si has a barrier of 21.7 kcalmolÀ1 (at standard condi-
tions). However, an additional barrier of 4.6 kcalmolÀ1 has to
be taken into account for the dissociation of 4-trimer. Regen-
eration at 4-trimer through SBM is not possible, as the (TiÀF)3
unit is effectively shielded by the Cp rings, and a vacant coor-
dination side at Ti is only generated via dissociation.[19] This
The energetic profile has some major kinetic consequences:
1) Highly concentrated solutions are needed to achieve ac-
ceptable TOF, because regeneration TS (TS4-1) as well as
HDF TS are formed without prior adduct formation.[3,5,7]
2) Substrates with a low HDF barrier constantly remove 1-THF
from the equilibrium, shifting it in the direction of 1-THF.
3) As the silane is consumed, the forward regeneration barrier
rises, since the entropy contribution to the TS increases.
Thus, an excess of silane is needed to drive HDF reactions
to full conversion.[3,5,7]
4) Whether regeneration or HDF is the rate-limiting step de-
pends on the specific barrier for HDF.
5) When HDF is the rate limiting step, the overall barrier is
composed of the exchange part (regeneration) and the CÀ
F activation barrier.
6) The equilibrium of 1-THF and 4-trimer heavily affects over-
all barrier heights and is solvent dependent.
Figure 1. Enthalpy (bold values) and Gibbs Free Energy (italic values) profile
for regeneration of the catalyst in kcalmolÀ1 (298 K, THF). Bold profile: stan-
dard conditions, dashed profile for realistic concentrations: Me2SiH2
(5”10À1 molLÀ1)@[Cp2TiF]3 (5”10À3 molLÀ1)@Cp2TiF
(5”10À4 molLÀ1)@Cp2TiH (5”10À5 molLÀ1) and Me2SiH2 >Me2SiHF
(5”10À3 molLÀ1). Level of theory M06-2X(PCM)/TZ//M06-2X(PCM)/DZ (SBM
TS=s-bond metathesis transition state, HDF TS=hydrodefluorination transi-
tion state). Lowest HDF TS and decomposition TS are only shown as orienta-
tion.
Experimental studies
As predicted by DFT, experimental kinetics are affected by the
choice of solvent (Table 3). More basic ethers (e.g., THF) lead to
higher TOF than less basic ethers (e.g., diglyme) or even aro-
Chem. Eur. J. 2016, 22, 9305 – 9310
www.chemeurj.org
9307
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 3. Solvent influence on catalytic activity.
substrate solvent conv. 1a
2
T
t
TOF
[hÀ1
0.25 <1
1.4
0.25 530
(2 mL) [%] [mol%] [equiv] [8C] [h]
]
3
3
3
5
5
5
toluene traces
toluene 72
diglyme 79
diglyme 20
dioxane 30
0.7
2
0.6
15
15
15
1.1
1.5
1.1
5
5
5
rt
110 24
rt
110 72
110 96
110 96
1.9”10À2
2.1”10À2
7”10À2
THF
100
Figure 3. EPR spectra of the activated precatalyst (A) and of the reaction
mixture (B) at room temperature. For (A), 0.5 mL of a stock solution with
[1a]=5 mmolLÀ1 and [2]=0.6 molLÀ1 in 2 mL THF were filled in a glass
ampule (4 mm o.d.) degassed and flame sealed. (B) was prepared identically
to (A) +0.16 mmol of 6.
matic solvents. Donor solvents are not critically needed but
reduce the overall barrier. The choice of solvent can be crucial
in case of unreactive substrates like 5 to achieve full conver-
sion within reasonable reaction times.
We also tested triethoxysilane (2c) as a hydride source in
the HDF reaction of 3 (Table 4). A color change to dark blue
upon addition of 2c indicates successful reduction of the pre-
catalyst, however, after 15 min at room temperature, no HDF
product was observed. Activation of the precatalyst with 2 fol-
lowed by cooling to À358C and subsequent addition of 3 and
2c, yields a TOF of 12 hÀ1, surpassing the TOF of 0.7 hÀ1 at
À258C achieved in kinetic experiments with 2 alone. We tenta-
tively conclude that alkoxysilanes lower the regeneration barri-
er as predicted by BDE arguments, but accelerate catalyst de-
composition to alkoxytitanium compounds.
solved.[24] The EPR spectrum (B) remains unchanged over three
days (complete conversion of 6).
Preliminary kinetic studies (À258C to À158C) under pseudo-
zeroth order conditions (large excess of 2 and 3, max. 10%
conversion monitored) yielded activation parameters of DH=
20 kcalmolÀ1 and DS=7.6 calmolÀ1 KÀ1 (Figure 4). The entropy
contribution is small.
Table 4. Silane influence on catalytic activity.
substrate
solvent
(2 mL)
conv. 1a
2c
T
t
TOF
[%]
[mol%] [equiv] [8C]
[h]
[hÀ1
]
3
3
3
diglyme
diglyme
diglyme 13
0
2
3.0
1.1
1.1
1.3
1.0[a]
1.0[a]
rt 0.25
0
2
12
À70
À35
1
1
[a] Plus 1.2 equiv of 2.
Figure 4. Eyring plot for catalytic HDF of 3 at À15, À20, and À258C.
EPR spectra of the catalytic reaction mixture are shown in
Figure 3, and were taken before (A) and after addition of the
perfluorinated substrate (B). HDF of perfluorotoluene (6) is
slow enough to allow a catalytic EPR study at room tempera-
ture. Prior to the reaction, a featureless broad singlet is ob-
served. This is consistent with a Ti-alkoxide (Cp2TiÀOnBu) as
the major decomposition product of 1-THF.[2] The g-value of
1.987 is in line with literature data for Cp2TiÀOMe,[2b] and im-
plies the intermediate presence of 1-THF.[22] Furthermore,
a smaller signal is buried under the large signal from Cp2TiÀ
OnBu, and this might very well correspond to a doublet from
residual amounts of 1-THF or a silyl species (doublet coupling).
Dimeric TiIII/TiIV (higher order coupling) species cannot be re-
sponsible for the is resonance.
Note that while trimer dissociation can compensate for the
entropic penalty of forming a TS out of two independent mol-
ecules, release of THF from 1-THF into the THF solvent should
not lead to pronounced entropy compensation (Scheme 2).[25]
This is fully in line with 4-trimer and not 1-THF being the rest-
ing state of catalysis.
Independently prepared (Cp2TiH)2 (1-dimer) is not stable in
the presence of fluorinated silanes (Scheme 3), and yields 4-
During the reaction (B), a broad singlet with a g-value of
1.980, comparable to 1.981 for Cp2TiF(PMe3), is observed likely
corresponding to either 4 or 4-trimer.[23] Similar to other litera-
ture examples, the fluorine coupling is too small to be re-
Scheme 2. regeneration TS (top), and HDF TS (bottom).
Chem. Eur. J. 2016, 22, 9305 – 9310
www.chemeurj.org
9308
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
entropy, 298 K, 1 bar) for all species considered. Final single-point
energies (SP) were calculated at the M06–2X(PCM)[29] level of
theory employing triple-z Dunning basis sets (cc-pVTZ) from the
EMSL basis set exchange library,[30] to minimize BSSE contribu-
tions.[31] BDE calculations were performed similarly but without sol-
vent corrections.
Scheme 3. Reaction of 1-dimer with 2b.
Techniques: All reactions and manipulations were carried out in
pre-dried glassware under an argon atmosphere by using standard
Schlenk-type and vacuum-line techniques, or by working in an
argon-filled glovebox. The amount of gaseous compounds was de-
termined using pVT technique or by condensing the gas into
a weighted J. Young valve flask.
trimer (identified by EPR) and oligomeric Si species (indicated
by the disappearance of fluorosilane 2b and appearance of
several new signals in the polysilane 29Si-NMR region).
Effect of endergonic regeneration on catalyst stability
Chemicals: All solvents were distilled from sodium or potassium,
and stored over sodium/potassium alloy. 1-Dimer was prepared in
a 50 mL J. Young valve flask similar to the literature procedure.[2b]
1a, 2, 5, and 6 were purchased from abcr GmbH, Karlsruhe. 2, 5, 6
were distilled over CaH2 and 2 was stored in the glove box. Ph2SiF2
was purified by vacuum distillation at room temperature. 3 (Solvay
Fluor) and 2c (Prof. Braun, HU Berlin) were obtained free of
charge, 3 was used as received and 2c was stored in the glove
box.
The kinetic profile for regeneration provides an explanation for
seemingly contradictory experimental observations by us and
Bercaw. Catalytic HDF reactions at RT or higher temperatures
are possible in THF (or other ethereal solvents) with the active
species being 1-THF because it is only present in small
amounts. The stabilizing effect of endergonic regeneration ex-
tends beyond the HDF example presented here. Buchwald’s
catalyst activation protocol for hydrosilylation, for example,
should yield 4 or 4-trimer in the absence of a substrate, and
could explain the high stability of ‘1-THF’ even in the presence
of methanol.[3b] Under catalytic conditions, the amount of fluo-
rosilanes from activation will always be much higher than the
amount of free 1-THF.
Catalytic hydrodefluorination: Substrates, conditions, conversion
and TOF are listed in Tables 3 and 4. A single-necked flask
equipped with a J. Young valve was charged with 1a, 2 (and 2c)
and solvent, and a change in color from yellow to purple to green
was observed immediately. The substrate was added with a syringe
or condensed into the reaction mixture. Subsequently, the flask
was degassed and the corresponding conditions were applied, re-
spectively (cf. Tables 3 and 4). The conversion of the substrates was
determined from NMR spectra by integration of product resonan-
ces versus the internal standard (fluorobenzene). The products
were identified by 19F NMR spectroscopy ([D6]benzene), using avail-
able literature data for 3a,[32] 3b,[32] and 5a.[33]
Conclusion
The driving force for catalytic HDF reactions is the BDE differ-
ence between CÀF/CÀH and SiÀF/SiÀH, respectively, because
regeneration is endergonic. DFT predictions are in very good
agreement with experimental results. The resting state in HDF
catalysis of 1 is the Ti-F species 4-trimer. The results presented
here emphasize that catalyst regeneration in a catalytic cycle
does not necessarily have to be exergonic, as long as the over-
all cycle is exergonic. Endergonic regeneration can actually be
beneficial to catalyst stability of very reactive catalysts. The
rate-limiting step in such cases consists of a catalyst regenera-
tion and CÀF bond activation part.
Kinetic experiments: For a stock solution, 1a (5.5 mg, 25.5 mmol),
2 (550 mL, 3.0 mmol) and [D8]THF (2 mL) were mixed in the glove-
box and stored at À408C. For each sample, 200 mL stock solution
was filled in a glass ampule (4 mm outer diameter (o.d.)), degassed,
and 3 (115 mg, 0.7 mmol) was added by vacuum transfer; the
ampule was flame sealed and subsequently immersed into a cool-
ing bath at À788C. To determine G/H/S, the reaction progress was
monitored by 19F NMR spectroscopy at À15, À20, and À258C, re-
spectively. The conversion of the substrate was determined from
19F NMR spectra by integration of product and starting material
resonances (detailed data is given in the Supporting Information).
Reaction of 1-dimer with 2b: 1-Dimer (150 mg, 0.422 mmol/
0.843 mmol monomeric) was dissolved in [D8]toluene (2 mL) at
room temperature. Upon addition of 2b (88.2 mL, 0.461 mmol) the
color changed from violet to green. 0.4 mL of the reaction mixture
was used for an NMR experiment. In the 29Si NMR spectra
([D8]toluene), the initial resonance of Ph2SiF2 at À26.9 ppm disap-
peared and four new resonances (À33.09, À33.45, À36.8, and
À37.0 ppm) in the disilane/polysilane region developed.[34] The sol-
vent of the residual mixture was removed under vacuum and an
EPR spectra was recorded. EPR (toluene): broad singlet at g=
1.976(9), identical to the literature data.[24]
Experimental Section
Calculations: All structures were fully optimized at the M06–
2X(PCM)[26]/6–31+(2d,p) level using Gaussian 09[10] coupled to an
external optimizer (PQS)[27] instead of the internal Gaussian opti-
mizer, using an ultrafine grid (Int(Grid=ultrafine)) and standard
SCF convergence quality settings (Scf=tight) for Gaussian single-
point calculations. The nature of each stationary point was checked
with an analytical second-derivative calculation (no imaginary fre-
quency for minima, exactly one imaginary frequency for transition
states, corresponding to the reaction coordinate) and the accuracy
of the TS was confirmed with IRC scans. S2 values for all doublet
species are below 0.77. Solvent influence of polar solvents (THF,
e=4.24) was modeled explicitly, using the polarizable continuum
model (PCM) implemented in the Gaussian 09 software suite. Tran-
sition states were located using a suitable guess and the Berny al-
gorithm (Opt=TS).[28] Vibrational analysis data derived at this level
of theory were used to calculate thermal corrections (enthalpy and
Acknowledgements
The authors would like to thank Prof. P. H. M Budzelaar (U-
Manitoba, Winnipeg) for generous donation of CPU time and
valuable discussions on the manuscript. Support from the
Chem. Eur. J. 2016, 22, 9305 – 9310
www.chemeurj.org
9309
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[11] K. F. Zmbov, J. L. Margrave, J. Phys. Chem. 1967, 71, 2893–2895.
[12] M. W. Chase, C. A. Daives, J. R. Downey, D. J. Frurip, R. A. McDonald,
A. N. Syverud, in JANAF Thermochemical Tables, Third Edition JPCRD 14,
Supplement No. 1, 1985.
Deutsche Forschungsgemeinschaft (DFG) is gratefully
acknowledged.
[13] We initially considered that TiIII-systems might pose some problems for
DFT in predicting accurate BDEs. However, for the exchange process,
only DEBDE matters, and though absolute accuracy in BDE might be off,
relative accuracy is usually more reliable due to fortuitous error cancel-
lation.
Keywords: catalyst regeneration
calculations · homogeneous catalysis · hydrodefluorination ·
metal hydride
·
density functional
[1] a) A. J. Hoskin, D. W. Stephan, Coord. Chem. Rev. 2002, 233–234, 107;
b) J. F. Harrod, Coord. Chem. Rev. 2000, 206–207, 493.
[2] a) H. Brintzinger, J. Am. Chem. Soc. 1967, 89, 6871–6877; b) J. E. Bercaw,
H. Brintzinger, J. Am. Chem. Soc. 1969, 91, 7301–7306.
[14] J. A. Panetier, S. A. Macgregor, M. K. Whittlesey, Angew. Chem. Int. Ed.
2011, 50, 2783–2786; Angew. Chem. 2011, 123, 2835–2838.
[15] a) U. Jäger-Fiedler, M. Klahn, P. Arndt, W. Baumann, A. Spannenberg,
V. V. Burkalov, U. Rosenthal, J. Mol. Catal. A 2007, 261, 184–189; b) S.
Yow, S. J. Gates, A. J. P. White, M. R. Crimmin, Angew. Chem. Int. Ed.
2012, 51, 12559–12563; Angew. Chem. 2012, 124, 12727–12731.
[16] a) E. Samuel, J. F. Harrod, J. Am. Chem. Soc. 1984, 106, 1859–1860;
b) C. T. Aitken, J. F. Harrod, E. Samuel, J. Am. Chem. Soc. 1986, 108,
4059–4066.
[17] J. A. M. Simoes, J. L. Beauchamp, Chem. Rev. 1990, 90, 629–688.
[18] Diamond - Crystal and Molecular Structure Visualization Crystal Impact -
Dr. H. Putz & Dr. K. Brandenburg GbR, Kreuzherrenstr. 102, 53227 Bonn,
Germany.
[3] M. F. Kuehnel, P. Holstein, M. Kliche, J. Krüger, S. Matthies, D. Nitsch, J.
Schutt, M. Sparenberg, D. Lentz, Chem. Eur. J. 2012, 18, 10701–10714.
[4] a) W. W. Lukens, P. T. Matsunaga, R. A. Andersen, Organometallics 1998,
17, 5240–5247; b) G. A. Luinstra, J. H. Teuben, J. Chem. Soc. Chem.
Commun. 1990, 1470–1471; c) M. D. Walter, C. D. Sofield, R. A. Ander-
sen, Organometallics 2008, 27, 2959–2970; d) J. M. de Wolf, A. Meet-
sma, J. H. Teuben, Organometallics 1995, 14, 5466–5468; e) J. W. Pattia-
sina, F. van Bolhuis, J. H. Teuben, Angew. Chem. Int. Ed. Engl. 1987, 26,
330–331; Angew. Chem. 1987, 99, 342–343; f) S. R. Frerichs, B. K. Stein,
J. E. Ellis, J. Am. Chem. Soc. 1987, 109, 5558–5560; g) Y. You, S. R.
Wilson, G. S. Girolami, Organometallics 1994, 13, 4655–4657.
[19] Formation of mixed TiÀF/TiÀH oligomers appears possible but does not
affect the energetics of regeneration.
[5] a) M. F. Kühnel, D. Lentz, Angew. Chem. Int. Ed. 2010, 49, 2933–2936;
Angew. Chem. 2010, 122, 2995–2998; b) X. Verdaguer, U. E. W. Lange,
M. T. Reding, S. L. Buchwald, J. Am. Chem. Soc. 1996, 118, 6784–6785;
c) G. Podolan, D. Lentz, H.-U. Reißig, Angew. Chem. Int. Ed. 2013, 52,
9491–9494; Angew. Chem. 2013, 125, 9669–9672.
[6] a) M. F. Kuehnel, D. Lentz, T. Braun, Angew. Chem. Int. Ed. 2013, 52,
3328–3348; Angew. Chem. 2013, 125, 3412–3433; b) M. Klahn, U.
Rosenthal, Organometallics 2012, 31, 1235–1244.
[7] A manuscript focusing on the studies of HDF barriers for fluorinated
arenes and olefins is currently under preparation and will be published
in due course.
[8] The stoichiometric reaction of 1-THF with 3 yielded 3a and 3b in the
same ratio in the catalytic reaction (E/Z=0.67).
[20] Details on HDF mechanisms will be published in due course.
[21] A. L. Raza, T. Braun, Chem. Sci. 2015, 6, 4255–4260.
[22] It appears plausible that Cp2TiÀOBu is the initial resting state if the sub-
strate is not added fast enough. During catalysis, however, the concen-
tration of SiÀF@SiÀOBu.
[23] L. Hao, J. F. Harrod, Inorg. Chem. Commun. 1999, 2, 191–194.
[24] L. Bai, F. Liu, J. Wang, Sci. China Ser. B 1991, 34, 257–264.
[25] C. Ehm, R. Cipullo, P. H. M. Budzelaar, V. Busico, Dalton Trans. 2016, 45,
6847–6855.
[26] Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215–241.
[27] a) J. Baker, PQS, 2.4 Parallel Quantum Solutions, Fayetteville, AR, 2001;
b) J. Baker, J. Comput. Chem. 1986, 7, 385–395.
[28] The Berny algorithm was never completely published; see the Gaussian
documentation for details, ref.[10].
[29] a) R. Raucoules, T. de Bruin, P. Raybaud, C. Adamo, Organometallics
2009, 28, 5358–5367; b) S. Tobisch, T. Ziegler, J. Am. Chem. Soc. 2004,
126, 9059–9071.
[30] a) D. Feller, J. Comput. Chem. 1996, 17, 1571; b) K. L. Schuchardt, B. T.
Didier, T. Elsethagen, L. Sun, V. Gurumoorthi, J. Chase, J. Li, T. L. Windus,
J. Chem. Inf. Model. 2007, 47, 1045.
[31] S. F. Boys, F. Bernardi, Mol. Phys. 1970, 19, 553.
[32] H. Koroniak, K. W. Palmer, W. R. Dolbier, H.-Q. Zhang, Magn. Reson.
Chem. 1993, 31, 748–751.
[33] A. A. Shtark, Yu. V. Pozdnyakovich, V. D. Shteingarts, Zh. Org. Khim.
1977, 13, 1671–1679.
[34] H. Sçllradl, E. Hengge, J. Organomet. Chem. 1983, 243, 257–269.
[9] Y.-R. Luo in Comprehensive handbook of chemical bond energies, CRC
Press, Boca Raton, 2007.
[10] a) Further details in the Supporting Information; b) Gaussian 09, Revi-
sion B.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Pe-
tersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J.
Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R.
Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H.
Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bear-
park, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J.
Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V.
Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Moro-
kuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dap-
prich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski,
D. J. Fox, Gaussian, Inc. Wallingford CT, 2009.
Received: April 8, 2016
Published online on June 3, 2016
Chem. Eur. J. 2016, 22, 9305 – 9310
www.chemeurj.org
9310
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim