Angewandte
Communications
Chemie
Catalyst temporal characteristics were next investigated in
a series of three hydrogenation cycles with Cp*ZrBz2/ZrS as
the catalyst. After each cycle, all volatiles were vacuum-
transferred from the reactor to a À788C trap for further
analysis. The supported catalyst was then dried in vacuo for
1 h before substrate addition for the next cycle. For each cycle
(for 2nd cycle data see Table 1, entry 2; for 3rd cycle data see
Table S4), negligible variation in product cis-selectivity
(> 99%) is observed, suggesting that one benzyl group
remains bound to the ZrIV center. The decrease in o-xylene
hydrogenation TOF[18] in succeeding cycles likely reflects
small trace impurities.[13]
The structure of 13C-enriched Cp*Zr(13CH2Ph)2/ZrS was
investigated by 13C CPMAS NMR before and after o-xylene
hydrogenations (Figure S13). The spectrum before catalytic
runs shows four major resonances at d = 128.0, 125.0, 74.0,
and 11.2 ppm, assigned, respectively to aromatic carbons, Zr-
CH2-Ph, Cp* framework, the 13C-enriched Zr-13CH2-Ph
group, and Cp*-Me carbon atoms. The downfield shifted
broad signal at d = 74.0 ppm clearly indicates the formation of
a “cation-like” electron-deficient organozirconium surface
species as evidenced by comparison to model ion pair
Figure 3. Energy profile (kcalmolÀ1) of the first o-xylene hydrogenation
subcycle at Cp*Zr(H)Bz/ZrS (see Figure S3 for the entire profile). The
ZrS surface is omitted for clarity.
+
À
Zr(CH2Ph)3 -B(CH2Ph)(C6F5)3
neutrally charged Zr(CH2Ph)3[(CH3C)CO]
(d = 74.8 ppm)[19] versus
(d =
shown in Figure 3 indicates that o-xylene coordination (1) is
slightly exergonic (DG = À0.7 kcalmolÀ1) and the first H-
transfer step is barrierless, leading to slightly endergonic
intermediate 2. The energetic demands of the first H2 addition
primarily reflect the entropic contribution associated with the
H2 activation/binding. The subsequent Zr-C hydrogenolysis
leads to dimethyl-hexadiene complex 3 that spontaneously
evolves to stable intermediate 4 via a barrierless second H-
transfer. The second and third H2 additions occur in a similar
manner, but lie at lower energies. An energetic span model[26]
was next used to probe the energetic difference (Table S1)
between the coordinated arene species (TOF determining
intermediate, TDI) and the first Zr-C hydrogenolysis tran-
sition state (TOF determining transition state, TDTS). Note
that xylene coordination lies outside the TDI and TDTS
range and only one H2 addition lies within the TDI—TDTS
range, indicating a kinetic rate law zero-order in [xylene] and
first-order in [H2], in good agreement with experiment. The
computed energetic barriers follow the trend o-xylene < m-
xylene < p-xylene (27.9, 28.7, and 34.4 kcalmolÀ1), respec-
tively, again in good agreement with experiment (Tables 1 and
S1). The relative TDI and TDTS stability found for the o- m-
and p-xylenes reflects an interplay of steric and electronic
effects.
Exclusive formation of all-cis isomers with Cp*Zr(H)Bz2/
ZrS but not with Cp*ZrH2/ZrS (Figure 1) suggests that
coordinated substrate p-face exchange is insignificant in the
former on the catalytic timescale. Two arene face exchange
pathways are conceivable: 1) Zr-substrate dissociation/re-
coordination (note that that free dienes and olefins are not
detected during the course of these reactions). 2) Face
exchange within the Zr coordination sphere.[27] To probe the
energetics of dissociation/re-coordination, DFT analysis
focused on the dissociation energetics of xylene and olefinic
intermediates present during the hydrogenation cycle for
Cp*ZrH2/ZrS versus Cp*Zr(H)Bz/ZrS. It is found that all
species bind more strongly to Cp*ZrH2/ZrS than to Cp*Zr-
65.7 ppm).[20] After exposure to reaction conditions (o-
xylene, 1 atm H2) for 1 hour, partial hydrogenolysis (40 Æ
5%, Figure S13) of the 13C-enriched Zr-CH2-Ph moieties is
observed, in accord with formation of catalytically active
cationic Cp*Zr(H)Bz/ZrS.[21]
Kinetic analysis of Cp*ZrBz2/ZrS-catalyzed o-xylene
hydrogenation at 258C indicates a rate law first-order in
[Zr], first-order in [H2], and approximately zero-order in [o-
xylene] over a broad range of arene concentrations.[22] Kinetic
measurements on Table 1, entry 1 between 0 and 458C at
constant 7 atm of H2 with rapid (> 1450 rpm) stirring to
minimize mass transfer effects (Table S5) combined with
a standard Eyring/Arrhenius analysis yields DH° = 4.8 Æ
0.5 kcalmolÀ1, DS° = À58.7 Æ 1.9 eu, and Ea = 5.3 Æ 0.6 kcal
molÀ1, suggesting a highly organized transition state (large
negative DS°), typical of many d0/fn-centered catalytic
processes,[23] and an intermolecular turnover-limiting step—
the first H2 delivery. The observed Ea (5.3 kcalmolÀ1) is
somewhat lower than those for conventional supported metal
nanoparticle o-xylene hydrogenation catalysts (Ea = 7.4–
14.9 kcalmolÀ1).[24]
DFT modeling was next carried out to gain further insight
into the remarkable cis-selectivity of the Cp*Zr(H)Bz/ZrS-
catalyzed hydrogenation and the reaction kinetics.[13] The ZrS
surface model consists of a 4 2 slab (12.85 14.50 30 )
built starting from the (101) plane cut from the optimized
bulk structure of tetragonal zirconia (See SI for more
details).[25] The six layer surface was saturated with four
dissociated pyrosulfuric acid and four undissociated water
molecules [S2O7À, 2H+, H2O]. The lowest oxygen layer was
fixed during the geometry optimization to simulate bulk
constraints. The most stable anionic surface was found by
removing the most acidic hydrogen from the surface. Only
one organozirconium cation was placed on the anionic surface
to avoid lateral interactions. The Gibbs free energy profile
Angew. Chem. Int. Ed. 2016, 55, 5263 –5267
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5265