Noncovalent Interactions Drive the Efficiency of Molybdenum Imido Alkylidene Catalysts for Olefin Metathesis

Article abstract of DOI:10.1021/jacs.9b04367

High-throughput experimentation and multivariate modeling allow identification of noncovalent interactions (NCIs) in monoaryloxy-pyrrolide Mo imido alkylidene metathesis catalysts prepared in situ as a key driver for high activity in a representative metathesis reaction (homodimerization of 1-nonene). Statistical univariate and multivariate modeling categorizes catalytic data from 35 phenolic ligands into two groups, depending on the substitution in the ortho position of the phenol ligand. The catalytic activity descriptor TON_1h correlates predominantly with attractive NCIs when phenols bear ortho aryl substituents and, conversely, with repulsive NCIs when the phenol has no aryl ortho substituents. Energetic span analysis is deployed to relate the observed NCI and the cycloreversion metathesis step such that aryloxide ligands with no ortho aryls mainly impact the energy of metallacyclobutane intermediates (SP/TBP isomers), whereas aryloxides with pendant ortho aryls influence the transition state energy for the cycloreversion step. While the electronic effects from the aryloxide ligands also play a role, our work outlines how NCIs may be exploited for the design of improved d⁰ metathesis catalysts.

Full text of DOI:10.1021/jacs.9b04367

Subscriber access provided by SUNY OSWEGO
Article
Non-Covalent Interactions Drive the Efficiency of
Molybdenum Imido Alkylidene Catalysts for Olefin Metathesis
Marco Antonio Barbosa Ferreira, Jordan De Jesus Silva, Samantha
Grosslight, Alexey Fedorov, Matthew S Sigman, and Christophe Copéret
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b04367 • Publication Date (Web): 10 Jun 2019
Downloaded from http://pubs.acs.org on June 10, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Non-Covalent Interactions Drive the Efficiency of Molybdenum
Imido Alkylidene Catalysts for Olefin Metathesis
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Marco A. B. Ferreira,*,†,§ Jordan De Jesus Silva,‡ Samantha Grosslight,† Alexey Fedorov,*,‡,¶
Matthew S. Sigman,*,† Christophe Copéret*,‡
†Department of Chemistry, University of Utah, 315 South 1400 East, Salt Lake City, Utah 84112, United States
‡Department of Chemistry and Applied Biosciences, ETH Zürich, Vladimir-Prelog-Weg 1-5, CH-8093 Zürich,
Switzerland
§Department of Chemistry, Federal University of São Carlos – UFSCar, Rodovia Washington Luís, km 235, SP-310,
São Carlos, São Paulo, Brazil, 13565-905
¶
Department of Mechanical and Process Engineering, ETH Zürich, Leonhardstrasse 21, CH-8092 Zürich,
Switzerland
ABSTRACT: High-throughput experimentation and multivariate modelling allow identification of non-covalent
interactions (NCIs) in monoaryloxy-pyrrolide Mo imido alkylidene metathesis catalysts prepared in situ as a key driver
for high activity in a representative cross-metathesis reaction (homodimerization of 1-nonene). Statistical univariate and
multivariate modelling categorizes catalytic data from 35 phenolic ligands into two groups, depending on the
substitution in the ortho position of the phenol ligand. Catalytic activity descriptor TON1h correlates predominantly with
attractive NCIs when ortho aryl phenols are used and, conversely, with repulsive NCIs when the phenol has no aryl ortho
substituents. Energetic span analysis is deployed to relate the observed NCI and the cycloreversion metathesis step such
that aryloxide ligands with no ortho aryls mainly impact the energy of metallacyclobutane intermediates (SP/TBP
isomers), whereas aryloxides with pendant ortho aryls influence the transition state energy for the cycloreversion step.
While the electronic effects from the aryloxide ligands also play a role, our work outlines how NCIs may be exploited for
the design of improved d0 metathesis catalysts.
INTRODUCTION
computational studies to adopt a 5-coordinate trigonal-
bipyramidal (TBP) geometry with the X – typically
more -donating – group and the metallacycle in the
equatorial plane (Scheme 1A).6 Such TBP intermediates
can however undergo a turnstile isomerization process
yielding an off-cycle MCB with a square-based pyramidal
(SP) geometry that slows down catalysis and promotes
deactivation and side-reactions.6a-d,8,9,10 The presence of
-donor ligands with varying donation properties in
these unsymmetrical catalysts raises the energy of
metallacyclobutane intermediates preventing their
overstabilization.6b-d Overall, this suggests that the
effectiveness of an alkene metathesis reaction could be
improved by minimizing the concentration of an off-
cycle SP metallacycle, however, the rational design of
Schrock-type metathesis catalysts remains challenging
and often unpredictable.
Transition-metal-catalyzed olefin metathesis has over the
last decades transformed the pharmaceutical, polymer,
and fine chemicals industries.1 Among the various
systems available, Schrock-type Mo and
W
d0
alkylidenes2 have been developed into highly active,
selective and stable metathesis catalysts, with
performance metrics that are attractive for industrial
applications.3 Recent compelling developments include
metal imido alkylidene complexes (X)(Y)M(=NR)(=CHR1)
with different X and Y ligand set (M = Mo, W; X ≠ Y, and
X is a stronger -donor), for instance the monoaryloxy
pyrrolide (MAP)2,4 and monoaryloxy chloride (MAC)5
family of catalysts, that enable the selective synthesis of
(Z)-alkenes, including chloro and fluoroalkenes. The
dissymmetry at the metal center of such catalysts was
recognized early as an important factor associated with
enhanced metathesis activity.6 Metallacyclobutanes
(MCB), the key intermediates of the Chauvin mechanism
of olefin metathesis,7 are proposed through
Herein, we describe a high-throughput experimental
(HTE) approach to identify active, in-situ-generated
molybdenum alkylidene complexes tested for
a
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 17
prototypical
cross
metathesis
reaction, the
phenols and two precursor bis-pyrrolido complexes,12
targeting both unsymmetrical MAP and symmetrical
bisaryloxide families of complexes, prepared using 1:1
1
2
3
4
5
6
7
8
homodimerization of 1-nonene, that is plagued with
deactivation pathways associated with the formation of
ethylene, and that requires a better understanding of key
parameters driving this reaction.6a,d,11 The HTE
methodology is integrated with multivariate statistical
modelling strategies that guide the search for highly
efficient catalysts and provide insights on key parameters
for further catalyst design (Scheme 1B,C). We evaluated
in situ formulations prepared from a library of 35
and
2:1
ratios
of
ArOH
and
(2,5-(Me)2-
Pyr)2Mo(=NAr)(=CHCMe2Ph), respectively [Ar = 2,6-
(iPr)2-Ph (Mo-1) and 2,6-(Me)2-Ph (Mo-2)].
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. The mechanism of olefin metathesis with d0 catalysts (A), design of the HTE study (B, C) and catalytic results
(D) for in situ formulations of Mo-1 with ArOH 1-35 in 1:1 ratio (Pyr = 2,6-dimethylpyrrolido). *Because of the
increasing amounts of isomerization products after the complete consumption of 1-nonene, TON values are reported
after 75 minutes and 135 minutes instead of 75 minutes and 501 minutes.
ACS Paragon Plus Environment

Page 3 of 17
Journal of the American Chemical Society
A. Proposed cataly c cycle with d0 (X)(Y)M(=E)(=CHR1) catalysts for alkene
B. In-situ-generated molybdenum alkylidene
complexes for homodimeriza on of
1-nonene (steps 1-2)
1
2
3
4
5
metathesis
coordination
[2+2] cycloaddition
cycloreversion
decoordination
R
R
X
Y
E
R
E
E
E
E
R
R
Mo
Mo
X
Mo
Y
Mo
Mo
+
+
i-Pr
i-Pr
CMe2Ph
Me
Me
CMe2Ph
X
Y
X
N
N
Y
X
Y
TBP
Mo
Mo
Pyr
Pyr
6
7
8
9
Pyr
Pyr
Mo-1
Mo-2
turnstile
isomerization
X, Y = OR, NR2, CH2R1
X stronger s-donor than Y
E = NR, O
Step 1
β-hydride
transfer
E
E
1:1 or 1:2 ratio
R
Mo-1/2
ArOH 1-35
+
by-products,
catalyst
deactivation
140 in situ formulations
Mo
Mo
X
Y
toluene, 27 ºC, 3 h
X
H
R
Y
SP
Step 2
catalyst
formulation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
C7H15
+
C7H15
C7H15
cis- and
trans-hexadec-8-ene
toluene, 27 ºC, 8 h
C. Library of phenol ligands
OH
OH
OH
OH
OH
F
OH
OH
OH
OH
OH
OH
F
Me
Me
Me
F
F
MeO
OMe
CF3
i-Pr
i-Pr
F
F
Br
Cl
F
CF3
11
1
2
3
4
5
6
7
8
9
10
OH
OH
F
OH
OH
OH
OH
OH
OH
OH
t-Bu
t-Bu
F
F
F
Cl
Cl
Ph
Ph
O2N
NO2
Ph
Ph
Ph
Ph
Br
Br
Cl
Cl
Cl
Cl
F
Cl
CF3
Me
F
Me
Cl
12
13
14
15
16
17
18
19
20
Me
Me
OH
Me
Me
OH
Me
OH
Me
Me
Me
Me
Me
OH
OH
OH
OH
OH
Br
Br
Br
Ph
Ph
Ph
Ph
Br
Br
Me
Me
OMe
OMe
Me
Me
Me
F
F
Ph
Br
Br
28
21
22
23
24
25
26
27
Br
CF3
CF3
Br
MeO
OH
Ph Ph
OH
TBDMSO
OH
Br
OH
OH
OH
OH
Ph
Ph
Ph
Ph
Br
Ph
Ph
Ph
Ph
Br
Br
Br
Br
N
N
F3C
CF3
Ph
Ph
Br
Br
29
30
31
32
33
34
35
D. Results for the cataly c homodimeriza on of 1-nonene with formula ons prepared from Mo-1 and ArOH-1-35 (1:1 ra o, see ESI
for details)
The resulting dependence of both the productive (i.e.,
calculated from the yield of the C16 products) turnover
number (TON) and turnover frequency (TOF) was
quantitatively related to structural features of aryloxy
ligands by examining their computed electronic and
steric characteristics using multivariate regression
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 17
analysis.13 The resultant statistical models provide a
Examples of outputs for productive TOFin, TON1h and
1
2
3
4
5
6
7
8
correlation between a non-covalent interaction (NCI)
exerted by phenolic ligands and productive TONs and
TOFs. Ultimately, the results could be interpreted with
respect to the stability of the TBP and SP intermediates
and respective transition states for the cycloreversion
step, furnishing guidelines for predictable catalyst
design.
TON8h obtained with 1:1 ratio of Mo-1 and ArOH 1-35
are presented in Scheme 1D (see ESI for such plots using
1:2 ratio and results with Mo-2, Figures S3-5).
Inspection of TOFin and TON1h for 1:1 and 1:2
formulations with either Mo-1 or Mo-2 reveals high
collinearity (R2 = 0.79 – 0.97) of these reactivity outputs
(Figures S13-14). In contrast, comparing TOFin and TON1h
values for 1:1 formulation relative to both arylimido
ligands uncovers noteworthy differences. In situ
formulations derived from Mo-2, a precursor with a
smaller 2,6-dimethylphenylimido ligand relative to 2,6-
diisopropylphenylimido ligand in Mo-1, exhibit
significantly reduced TOFin and TON1h values using
phenols that lack pendant aryl groups in ortho positions,
with several phenols giving inactive formulations, in
sharp contrast to Mo-1 (Scheme 1D and Figures S15).
However, phenols with pendant aryls display a high
degree of collinearity for both Mo precursors. These
observations point at significant influence of ortho aryl
substituents on generating active catalyst formulations
(vide infra). Remarkably, almost all active catalysts
present high SC16 selectivity (averages of 95% and 93% at
1h and 8h, respectively).
9
RESULTS AND DISCUSSION
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Testing in situ Prepared Mo Metathesis Catalysts in
the Homodimerization of 1-Nonene
Phenol ligands 1-35 were selected for catalytic testing
due to their broad range of electronic and steric
properties, while also possessing common characteristics
that could facilitate identification of uncorrelated
structural features essential for catalytic performance. For
simplicity, ArOH are numbered according to their
increasing molecular weight. Formulations were
prepared in situ from molybdenum bispyrrolide
precursors Mo-1 and Mo-2 and ligands 1-35
(Scheme 1B,C) by stirring 1 or 2 equiv. of the phenol with
a respective Mo-precursor in toluene for 3 hours at
27 °C. An aliquot of each mixture was combined with 1-
nonene to deliver a 1000/1 substrate-to-catalyst ratio
(0.1 mol % catalyst loading). These steps were performed
by an automated liquid handling robotic system
operated inside an inert atmosphere glovebox (see ESI
for details). The reaction mixtures were agitated for ca. 8
hours at 27 °C in open vials,14 and reaction progress was
monitored by GC after ca. 6, 16, 39, 72, 135, 258 and 501
minutes, giving conversions (X), selectivity for the
formation of hexadec-8-ene (SC16 and SC16(E/Z)), and
respective turnover numbers and turnover frequencies.
These data are presented in the ESI (Tables S3-6)
including conversion vs. time and E/Z selectivity vs. time
plots (Figures S18-159). Robustness tests were routinely
performed in triplicate with new batches of 1-nonene,
exhibiting good reproducibility (Tables S7-8). Below, the
discussion is focused on two selected responses, namely
productive TOFin and TON1h (data points collected after
ca. 6 and 72 min, respectively). The first response reflects
the initial activity of the catalyst formulation, wherein the
influence of catalyst deactivation is minimal. While no
formulation reaches complete conversion of 1-nonene
after 1 h, several formulations feature conversion
exceeding 90 %, and in general X1h > 50% at this reaction
time (Scheme 1D). Therefore, TON1h was used as a
response to evaluate the overall catalyst efficiency
comprising both activity and stability as it allows us to
differentiate each catalyst’s efficiency that is not possible
with later time points since many formulations reach
complete conversion making these measurements
insensitive for discriminating between catalysts.
Next, we sought to correlate activity and stability of in
situ formulations with the structural characteristics of the
tested phenol ligands. For brevity, we limit the discussion
below primarily to results obtained using 1:1 ratio of
Mo-1 and phenols 1-35. TOFin and TON1h values for
these tests are reasonably correlated (R2
Figure 1A), suggesting that most of the ligands are
associated with similar deactivation pathways.
= 0.67,
Considering that 1-nonene concentration decreases over
time, a drop in the reaction rate and consequently TOF1h
< TOFin were expected. To identify ligands furnishing
higher catalytic stability, we defined a deactivation
parameter as DEACT = [(TOF1h – TOFin)/TOFin]  100%
that revealed a typical range for deactivation across most
of tested phenols at 65-90%. However, bulky phenols 2,
17, 23, 27, 28, 30, and 33 give more stable formulations
with Mo-1, deactivating to a significantly lower extent
(20-55%). Hexadec-8-ene forms faster at 1 h relative to
the initial rate only for one example, the bulky phenol 11.
This result is due to a slow induction period observed at
the beginning of the reaction with 11. Thus, the DEACT
parameter identifies the most stable formulations, as well
as induction periods (observed only for ArOH-11).
In general, the (E/Z) ratios for the SC16 (E) isomer increase
with reaction progress for most formulations and
catalysts typically display SC16(E/Z)8h >4 if TON8h exceeds
ca. 850, especially for formulations with Mo-1
(Figure 1B). This is consistent with the competing
metathesis isomerization that is favored for the more
active and stable catalysts as these systems are likely to
ACS Paragon Plus Environment

Page 5 of 17
Journal of the American Chemical Society
display a E/Z ratio approaching the thermodynamic value under these conditions even with rather bulky phenols as
1
2
3
4
5
6
7
8
at higher conversions. Only the bulky ligands 26, 33, 34,
and 35 do not follow this trend and favor the formation
of cis-hexadec-8-ene, in line with previous reports
demonstrating that homocoupling3 and cross-
metathesis4 of terminal alkenes could proceed in high Z-
selectivity with bulky MAP catalysts.
ArOH 29, 30 and 35. Furthermore, evaluation of the
homologous series of tri- and penta-halogenated
phenols (6, 12, 13, 20, 27, and 32) with Mo-1 and Mo-2
reveals that the more electronegative ligands display
lower TON1h values (Figure 2B). The fluorophenols 6 and
12 are less electronegative than chloro- and
bromophenols (13, 20, 27, 32) and based on the above
result might be expected to feature high TON1h as well.
However, the most electropositive phenol 6 yields the
lowest TON1h in this series, possibly due to the
intermolecular ligand scrambling forming respective
bispyrrolides and bisaryloxides.15 Overall, this points to
an entangled influence of phenol structural properties on
catalyst activity, which in turn hampers rational design of
metathesis catalysts.
9
DEACT = [(TOF1h – TOFin)/TOFin] × 100%
A.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R2 = 0.67
40
90
80
70
60
50
40
30
20
10
0
31
20
0
25
22
19
-20
-40
-60
-80
-100
3
24
15
21
8
29 18
13
27
5, 7, 10, 14, 16, 26, 34
6
20
28
17
33
30
12
4
32
35
23
2
11
9
1
0
200
400
600
800
1000
TON1h
Higher E selec vity
at ini al me
B.
Z > E
a er 8 h
Higher E selec vity
at 8 h
8
7
6
5
4
3
2
1
0
R2 = 0.51
900
29
800
700
600
500
400
300
200
100
0
15
24
18
17
31
25
Figure 2. Preliminary analysis of structural features of
selected phenol ligands and catalytic activity of
formulations derived from Mo-1 and Mo-2. The
estimates for size and electronegativity were molecular
weight (MW) and natural bond orbital (NBO) charges of
the oxygen (qO of phenolate), respectively.
23
14
32
11
19
13
2
34
3
8
21
27
1
4
6
30
26
12
22
20
35
9
5
10
33
28
0
2
4
6
8
SC16(E/Z)8h
Parametrization
Figure 1. Analysis of experimental descriptors obtained
with 1:1 in situ formulations with ligands 1-35 and Mo-1.
Considering the lack of obvious trends between catalytic
activity and structural features of phenols, we chose to
apply statistical correlation tools to classify TOFin and
TON1h data using approaches evolving from the Sigman
group.16 Our general workflow to correlate these reaction
outputs to molecular descriptors of phenolic ligands
included a variety of calculated steric and electronic
molecular parameters (Figure 3A). First, a conformational
ensemble of relevant geometries was generated using
Molecular Mechanics (MM), followed by DFT geometry
optimization to identify the lowest energy conformer at
While well-defined metathesis catalysts with large
aryloxy ligands typically produce high TONs and reaction
rates,4,5,9 in situ formulations with bulky ligands 15, 26,
29, 30 and 35 do not follow a clear trend, with Mo-1
giving TON1h of 860, 110, 850, 620 and 190, respectively;
the TON1h values are also similar with Mo-2 (Figure 2A).
Notably, these high TON values highlight that the
exchange of bispyrrolido for aryloxy ligands proceeds
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 17
the M06-2X/def2-TZVP level of theory, which provides
the platform to assemble structural parameters for
further analysis. Subsequently, preliminary
identification of univariate trends and application of
multivariate linear regression analysis provide statistical
models required for interrogation of the origin of olefin
1
2
3
4
5
6
metathesis efficiency of the in situ formulations. In
particular, we highlight below the development of
statistical models based on new tailored steric probes
(calculated at B97-D3/def2-TZVP level of theory).
a
A. Multivariate Model Development Workflow
7
Steric
8
9
o
o
o
Sterimol
Percent buried volume
Out-of-plane distances
TON1h = f(xi)
1) DFT calculation and
selection of lowest
energy structure
search (MM)
OH
Regression
analysis
Conformational
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Electronic
o
o
o
o
NBO charge and energy
Polarizability
Molecular dipole
HOMO/LUMO energy
2) Parameters
collection
R
TON1h measured
B. Molecular Parameters
NBO analysis
Molecular orbitals
HOMO
Percent buried volume
Out-of-plane distance
hup
Sterimol
R
Ortho-substituion
qH
H
R
qO
B'5
O
L'
Natural charge (q)
NBO energy
B'1
hdown
OH
LUMO
O
OH
2
Å
r
Hout,sum = hleft,up + hleft,down
M
B5
+ hright,up + hright,down
Molecular dipole
B1
L
Lsum = L+L'
B1,sum = B1 + B'1
B5,sum = B5 + B'5
h’up
E
E
∗
C-O)
Full ligand
C-O)
Polarizability
B1,full
OH
OH
+
h’down
+
B5,full
+
ELP2
H'out,sum = h'up+h'down
O)
ELP1
O)
Lfull
Figure 3. Parameterization of phenol ligands: development workflow of multivariate models (A) and selected calculated
molecular parameters (B).
Descriptors of Phenol Ligands
centered on the metal atom.20 The set radius sphere of r
= 5.0 Å was selected to capture the steric influence of
the meta substituent of the phenol ring; lower r values
(3.5 and 4.0 Å) allow selectively capturing steric effects of
ortho substituents.
Electronic Descriptors. The electron density of each free
phenol ligand or phenolate anion were assessed relative
to natural bond orbital (NBO) charges of oxygen (qO),17
C‒O bonding σ(C-O) (Eσ(C-O)) and antibonding σ*(C-O)
Preliminary Correlation Analysis
(Eσ*(C-O)
)
NBO energies (Figure 3B). Additional
parameters included HOMO/LUMO energies, NBO
energies of lone pairs of oxygen ELP(O), molecular dipole
(μ), and polarizability (Pol). These values are presented in
the supporting information.
At this stage, we considered all tested phenol ligands
that had improved activity of the precursor complex Mo-
1, seeking to relate their electronic and steric properties
to experimental TOFin and TON1h values. An initial
univariate correlation matrix, represented as a color map
in Figure 4A, was generated, revealing only low R2 values
not exceeding 0.22. A more detailed analysis of the
univariate trend of TOFin with the best correlating
parameter, Lsum, showed two distinct sets of catalysts that
differ by the presence or absence of aryl substituents in
ortho positions of respective phenol ligands (Figure 4B).
Therefore, we categorized the data set into two distinct
subsets: Group A contains phenols without aryl arms in
the ortho position (1-14, 16, 19-21, 27, 32) and Group B
contains ortho-arylated phenols (15, 17-18, 22-26, 28-
31, 33-35). These improved individual univariate
correlations using several steric descriptors of phenol
ligands and TOFin and TON1h values (R2 in the range of
Steric Descriptors. The steric influence of phenol
ligands was initially assessed using Verlop’s Sterimol
parameters.18 We considered individually the two
pendant groups in the ortho position (L, B1, B5, L’, B’1, B’5),
using the sum of both as the final parameter (Figure 3B).
In a second approach, we subjected the entire phenol to
the Sterimol calculation. Additionally, we investigated
alternative steric parameters including the height above
19
a defined plane, labelled as Hout,sum and H’out,sum
.
These
parameters account for the steric effects caused by 2,6-
substituents of the phenol ring; additional parameters
are presented in ESI. Finally, the percent buried volume
(%Vbur) was determined, which is defined as the fraction
of the volume occupied by a ligand in an abstract sphere
ACS Paragon Plus Environment

Page 7 of 17
Journal of the American Chemical Society
0.5-0.9; see Figures S16-S17 in ESI). For example, a %Vbur captures the steric effect mainly of the ortho
1
2
3
4
5
6
7
8
statistically significant correlation was observed for 1:1
formulations with Mo-1 in Group A using the size of the
phenolic substituents represented by the %Vbur steric
descriptor (Figure 4C). Interestingly, a correlation of
Group B ligands with the same measure reveals an
inverted correlation, wherein the larger phenols are
associated with reduced activity. Considering that the
positions, these trends hint at the importance of NCIs
exerted by these substituents with optimal radius
spheres of r = 3.5 and 4.0 Å for Groups A and B,
respectively.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Initial correlation analysis. Correlation color map for formulations with Mo-1 (A), optimal univariate steric
correlation parameter (B), and best univariate correlation after splitting the data set into groups with and without ortho
aryl substituents in phenols (C). The Group A contains phenols without aryl arms in the ortho position, and Group B
contains ortho-arylated phenols.
Analysis of Outliers and Control Experiments
of the phenol, increasing the distance between the CH3
group and the metal center. In order to obtain a more
reliable %Vbur value of 11, we fixed the dihedral
CA‒CAr‒C‒H angle to 0°, which allows this data point to
be accurately incorporated into the correlation. However,
11 remains an outlier in TOFin correlation due to the slow
induction period, as identified by the DEACT parameter
(Figure 1A).
Some outliers were excluded from the correlations
presented in Figure 4C. While 2,6-(i-Pr)2-PhOH (11) yields
a formulation with a comparable productive TON1h to
2,6-(Me)2-PhOH (2) (500 and 540, respectively), the initial
calculation of %Vbur is likely overestimated for 11, which
we attribute to conformational differences of this
aryloxide in the Mo complex and the free ligand. The
conformer is shown in Figure S7 in the supporting
information, and has one of the C‒Me bonds eclipsed by
the aromatic ring, driving the second methyl group
towards the oxygen of the phenolate, thereby
overestimating the steric effect of the i-Pr group in %Vbur
parameter. In contrast, in 11 or related complexes
Mo(NAr)(CH2CH2CH2)(Me2Pyr)L,6c the C‒H bond is
eclipsed by the aromatic ring and directed to the oxygen
Others outliers were the pentabromophenol 32, and the
[2,6-(3,5-(CF3)2-Ph)2-PhOH] 33, which combines rather
large substituents with significant electronic perturbation
on the aromatic system. This strongly suggests the
required inclusion of electronic terms in the statistical
modeling.
A special group of outliers with lower activity was found
to directly relate to slow ligand exchange when forming
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 17
the active catalysts. For example, combining Mo-1 and
2,6-(t-Bu)2-PhOH (14) of Group A leaves Mo-1 unreacted
at room temperature according to H NMR (see Figure
S6 in ESI for more details). This is presumably due to the
large size of this phenol retarding the substitution
reaction. For similar reasons, the complete ligand
exchange of Mo-1 with 2,6-(Mes)2- derivatives 26 (2,6-
correlation between TOFin and qO (R2 = 0.93, Figure 5C).
1
2
3
4
5
6
7
8
Similarly, a good correlation (R2 = 0.83) between TON1h
and C‒O bonding σ(C-O) NBO energies (Eσ(C-O); directly
associated with the electronic density of the aromatic
ring, with less negative values representing electron rich
aromatic rings, see Table S14 of ESI) was observed.
1
Overall, the data suggest that electronic influences are
more substantial at the initial phase of the reaction, while
the effect of ligand size is influential as the reaction
progresses, ultimately impacting catalyst stability as
assessed by TON1h. This is found for both Groups A and
B, confirmed by the higher collinearity between TON1h
and %VBur (Figure 4C).
(Mes)2-PhOH) and 28 (2,6-(Mes)2-4-F-PhOH) of Group B
9
requires longer reaction time (Figure S6). Likewise, the
bulky ligand 34 [2,6-(2,5-Ph-Pyr)2-PhOH] showed the
lowest catalytic performance in Group B giving <5% of
the ligand exchange product after ca. 1 day.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Finally, two particular cases of low-activity catalysts were
identified in both groups and involved 2,6-(OMe)2-PhOH
(7) and 2,6-(2-OMe-Ph)2-PhOH (23) phenols. Both yield a
stable 1:1 complex as determined by in situ NMR
experiments, avoiding the exchange of a second aryloxy
ligand when a second equivalent of 7 or 23 was added
(Figure S6). Therefore, their low activities are presumably
due the coordination between the methoxy substituent
on the ligand and the metal center (MeO–Mo
interaction).
Development of Steric Probes for Non-Covalent
Interactions (NCIs)
The rational development of catalysts21 has often relied
on the concept of steric repulsion. More recently, the
importance of stabilizing non-covalent substrate–catalyst
interactions has been recognized as central for the
geometrical preorganization of transition states.22
Although NCIs forces are individually relatively weak,
they rapidly become important for molecular structures
of increasing size.23 Multivariate statistical modeling
using NCIs as parameter has been reported.23
Electronic Correlations of Ortho-Isosteric Ligands
To explore the electronic impact of ligand 32 and related
phenols in our correlations, we evaluated a series of
phenols with the same ortho substituents and selected
phenols with the 2,6-(Br)2- and 2,6-(Ph)2- ligands (19, 21,
27, 32 from Group A, and 15, 17, 18, 24, 25, 29, 31 from
Group B). Figure 5 depicts good univariate inter-
correlations found for these sub-classes of phenols.
Analysis of 2,6-(Br)2- ligands demonstrates an excellent
correlation of TOFin and the NBO charge of the phenol
oxygen (qO; greater negative values represents a more
electron rich aromatic ring, see Table S14 of SI)
associating a reduction in TOFin with an increase in
electronegativity (Figure 5A). The latter could also be
interpreted as decrease in -donating properties of the
aryloxy ligand, which is expected to stabilize the TBP
intermediate relative to the off-cycle SP isomer, as was
already discussed above. Similarly, a good correlation is
observed for the TON1h response with the Pol parameter
(R2 = 0.97) (Figure 5B), which may be understood as a
hybrid descriptor, and potentially directly related to
dispersive stabilizing forces. It is noteworthy that more
electronegative substituents in the aromatic system raise
the polarization of the π-system, which in this case
lowers catalytic performance (TOFin).
As suggested by the initial analysis detailed above, the
correlations for Mo-1 formulations as well as inferior
performance of formulation derived from the smaller
Mo-2 precursor advocate that NCIs may be influencing
the catalytic activity. We have therefore focused on
developing probes more specific to NCI interactions and
that can be used for statistical analyses. On the basis of
reports by Bohm and Exner,24 we exploited
homodesmotic reactions25 to derive NCI sensitive
parameters using only one half of the phenol ligand to
reduce the conformational complexity (Figure 6A). For
non-symmetric ligands, we considered the sum of both
sides of the phenol. All probes were normalized to Ph,
that is defined as ΔE[Ph] = 0.
Our first probe, ΔENCI-A defined as presented in Figure 6A,
aimed to capture the NCI exerted by the pendant groups
in the ortho position of ArOH (Figure 6A). It should be
noted though that the close proximity between the
ortho-aryl substituents and the adamantyl group likely
compromises reliability of this parameter for bulkier
substituents like t-Bu, delivering values significantly
higher than expected (Table S13). Thus, a second probe
was designed, which avoids these issues. According to
previous studies using 1,8-diarylnaphthalenes in this
context,26 the stabilizing π-stacking and CX---π (X = H,
halogen) interactions exerted in the Mo complex by
ligands with pendant aryl groups may be captured (ΔENCI-
Likewise, analysis of the 2,6-(Ph)2- series of phenols
reveals a divergent scenario. These phenols demonstrate
that increasing the electronegativity (weaker -donor
ligands) enhances the rates as defined by a strong
ACS Paragon Plus Environment

Page 9 of 17
Journal of the American Chemical Society
B, Figure 6A). This probe should provide a good balance
between repulsive and attractive NCIs.
Figure 5. Univariate correlation of electronic descriptors
for selected 1:1 formulations using Mo-1.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 17
A. Non-covalent interaction probes
C. Identification of slow ligand exchange ligands using NCI probes
1
∆ENCI-A
slow
2
3
4
5
6
7
8
H
O
H
O
H
O
H
O
OH
R
R
G
R
R
R
R
+
+
H
Ad
Ad
H
H
H
H
H
N
N
N
G
G
X
X
Pyr-H
(Pyr)2
X
Mo
Mo
Mo
Pyr
X
Pyr
Pyr
I-(X)NCI-A
∆ENCI-A = [∆E-II-(X)NCI-A + ∆E-II-(Ph)NCI-A] – [∆E-I-(X)NCI-A+ ∆E-I-(Ph)NCI-A
∆ENCI-B
I-(Ph)NCI-A
II-(X)NCI-A
II-(Ph)NCI-A
O
O
H
X
X
]
∆EENCI-A (37.0)
∆EENCI-B (> 7.0)
9
Me
Me
OH
Me
Me
X
Me
Me
OH
Ph
Ph
OH
X
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OH
Me
Me
t
-Bu
t-Bu
+
+
N
N
Me
Me
Ph
Ph
Me
Me
F
I-(X)NCI-B
I-(Ph)NCI-B
II-(X)NCI-B
II-(Ph)NCI-B
14
26
28
34
∆ENCI-B = [∆E-II-(X)NCI-B + ∆E-II-(Ph)NCI-B] – [∆E-I-(X)NCI-B+ ∆E-I-(Ph)NCI-B
]
B. Correlation of TON1h of Mo-1 (1:1 formulation, Group A)
D. Correlation of TON1h of Mo-1 (1:1 formulation, Group B)
R2 = 0.72
18
15
22
25
31
24
29
17
Br
RO
OH
33
30
Br
28
35
Slow
30. R = Me
35. R = TBDMS
Ligand
Exchange
23
26
34
Figure 6. Development of non-covalent interaction probes (A); univariate correlation of TON1h and steric probes for 1:1
formulations with Mo-1 using ΔENCI-A for Group A phenols (B), and ΔENCI-B for Group B phenols (C, D). Ligand exchange
reaction and clustering of slow exchange ligands by using the non-covalent interaction probes (C).
A good correlation (R2 = 0.80) is found between TON1h
(1:1 formulations with Mo-1) of Group A, and the ΔENCI-A
probe (Figure 6B). Furthermore, good inter-correlations
are observed between ΔENCI-A (Group A) and traditional
steric parameters (B1,sum, %Vbur, Lsum give R2 > 0.9), while
the quality of correlation is moderate to poor with
electronic parameters (R2 = 0.52, 0.39 and < 0.07 for Pol,
Eσ*(C-O), qO and ELP(O1), respectively). This suggests that
this probe is describing mainly repulsive NCIs (Pauli
repulsion). Note that most ligands belonging to Group A
gave inactive formulations with the Mo-2 precursor. This
is presumably due to the smaller size of the 2,6-
dimethylimido ligand in Mo-2 compared to 2,6-
diisopropylimido ligand in Mo-1, which could favor
protonation of the alkylidene ligands during exchange11d
and Mo-2 (Figure 6C). One could anticipate that the
ligand exchange process proceeds via an associative
mechanism similar to that of the olefin coordination to
the TBP Mo alkylidene species. The equilibrium between
the phenol and the bis-pyrrolide complex (Mo-1 or Mo-
2) with the resulting adduct will be dependent on ligand
size. The 2,6-(t-Bu)2-PhOH (14) of Group A does not
undergo ligand exchange with Mo-1 and features a Δ
ENCI-A value of 37.0, significantly higher than 2,6-(i-Pr)2-
PhOH (11) (7.0). The Δ ENCI-B parameter also has
demonstrated utility in discriminate slow exchanging
ligands of Group B (Figure 6C), identifying bulky ligands
(26, and 34) with ΔENCI-B > 7.0 (Figure 6D).
Next, we investigated the applicability of the Δ ENCI-B
parameter to describe the reactivity of in situ
formulations with Group B ligands, showing an inverted
trend and good correlation with TON1h for phenols in
or
a
faster deactivation pathway due to easier
dimerization27 (see Tables S5 and S6).
We assessed if the new steric probes can be applied to
identify phenols undergoing slow exchange with Mo-1
Group B (R2 = 0.79, Figure 6D). Phenols yielding TON1h
>
650 feature negative Δ ENCI-B values suggesting the
ACS Paragon Plus Environment

Page 11 of 17
Journal of the American Chemical Society
importance of stabilizing π-stacking and CX---π (X = H, one-out (LOO) and k-fold methods), yielding good
1
2
3
4
5
6
7
8
halogen) interactions. However, the subgroup of biaryl
ligands, phenols 30 and 35 (Figure 6D) are structurally
too different from monoaryl ligands of Group B (and
feature an ortho-Br substituent) and therefore these
phenols did not correlate well with Δ ENCI-B, despite a
qualitative agreement in their observed reactivities.
Examining the inter-correlations of the ΔENCI-B descriptor
with the other parameter reveals that it has a hybrid
character (R2 = 0.6 for B1,sum and %Vbur; 0.2 for LUMO and
qH parameters; 0.1 for ELP(O1)-fenolate and qO).
scores for all cases consistent with a well-validated
model.14 The trained models from normalized descriptors
gave coefficients that revealed the significance of each of
represented effects.
ArOH ligands 26 and 31
Well-defined MAPs from ligands 26 and 31 and Mo-1
Me
Me
OH
Me
Me
9
i-Pr
i-Pr
i-Pr
i-Pr
N
Me
Me
CMe2Ph
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
CMe2Ph
Mo
Mo
Pyr
26
Pyr
O
O
Ph
Ph
Ph
OH
Previous studies correlated the rotational barrier about
the aryl-naphthalene bond in 1,6-diarylnaphthalenes to
σpara parameter, and interpreted this in terms of NCI
between the two aryl units.26 Ultimately, the increase of
the substituent size causes a strong repulsive interaction,
leading to an increase of ΔENCI-B and corresponding low
TON1h values. Notably, the ΔENCI-B parameter does not
correlate with phenols in Group A. To summarize, the
univariate correlations support our initial hypothesis
regarding distinct NCI regimes between phenols of
Groups A and B, and the importance of the ortho aryl
substitution in ArOH ligands.
Mes
Mes
Ph
Ph
Ph
Ph
Ph
Br
Br
MAP from ArOH 26
MAP from ArOH 31
31
Aiming to experimentally illustrate the importance of NCI
in our catalysts,28 we found that reaction of
trifluorophenol 6 and Mo-1 in 1:2 ratio yields a dimeric
bisaryloxide complex with relatively short contacts
between the fluorines and the isopropyl hydrogen atoms
(2.4-2.5 Å, Figure S166. These F–H interactions are likely
the driving force for the dimer stability, confirming the
relevance of NCI for systems described in this work.
Figure 7. Comparison of 1:1 in situ versus well-defined
catalysts formulations of Mo-1 and ligands ArOH 31 and
26 for homodimerization cross-metathesis reaction of 1-
nonene.
In addition, we compared performances of selected 1:1
in situ formulations of Mo-1 to the respective well-
defined MAP complexes and selected the most active
ligand ArOH 31 and the slow exchanging ligand phenol
26 (Table S7, Figure 7). The catalytic data displays only a
minor difference in TOFin and TON1h values for catalysts
derived from ArOH 31 showing a modest increase for
TOFin with the well-defined MAP catalyst. The
performance of the well-defined catalyst obtained from
the slowly exchanging ligand ArOH 26 is improved to a
slightly higher extent (Figure 7). These results validate
the use of catalytic data derived from in situ formulations
to estimate activity of the respective well-defined MAP
complexes.
The multivariate regression analysis for the Group A
produced models that describe essentially the same
effect for TOFin and TON1h (Figure 8A,B) featuring one
steric parameter and one hybrid interaction term, despite
failing to incorporate phenols 2 and 11 in the TOFin
response. These similarities are consistent with our
previous observation that the electronic effect of tested
phenols has a minor but not negligible influence on the
measured values of TOFin and TON1h. The steric
descriptors Δ ENCI-A and %Vbur used, respectively, for
modelling of TOFin and TON1h describe largely the same
effect. The inter-correlation between the Δ ENCI-A and
%Vbur parameters is R2 = 0.97. The inclusion of hybrid
stereoelectronic descriptors in interaction terms for TOFin
Multivariate Regression Analysis
(LUMO and B5,full) and TON1h (Polphenolate and Eσ*(C-O)
)
As the next step, multivariate linear regression analysis
was performed to investigate how the steric and
electronic parameters of phenolic ligands could
cooperatively impact the TOFin and TON1h responses for
Group A and B (Figure 8). We tested the consistency of
our models using internal-validation techniques (leave-
describes the properties of the entire ligand associated
to the their polarizability (Polphenolate, and B5,full are directly
related to polarization). The LUMO and Eσ*(C-O)
parameters reflect the perturbation of the electron
density by pendant substituents with different electronic
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 17
properties. However, these terms appear not to be
related to the σ-donor ability of the aryloxy ligand, as
revealed in Figure 5. It was expected that weaker
σ-donor ligands favor TBP- over SP-geometries,11
1
2
3
increasing the activity and stability of the complex by
their lower trans-influence.29
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Multivariate linear regression model to predict TOFin and TON1h for Group A (A, B) and Group B (C, D).
However, the performance of catalysts bearing 2,6-(Br)2-
substituted phenols in Group A do not fall into this
category as revealed by the correlation between TOFin
and qo (Figure 5A).30 Therefore, these descriptors
composing the interaction terms define the stabilizing
dispersive forces associated with lower catalytic activity,
as exemplified by of the poor performance of
pentabromophenol 32.
(Ph)2- ligands of Group B (Figure 5A). However, this
feature can be modulated with the increasing size of the
ortho substituent, and the steric terms H’out,sum and ΔENCI-B
introduced in the model reflect the reduction of the
activity by the increase of ligand size.
In evaluating the TON1h response, a model was obtained,
composed by the electronic parameter ELP(O2), and the
most significant repulsive interaction term (%Vbur(5 Å) and
Hout,sum) (Figure 8D). The electronic term describes the σ-
donor ability of the phenolate oxygen, while the NCI
term describes the reduction of activity with the increase
of ligand size. Particularly, this model highlights the
importance of the substituent size in the stability of the
formulation of Group B, in which ligands of comparable
size presented similar TON1h.
Next, ligands of Group B were subjected to the statistical
modeling. The correlation found for the TOFin response is
notably different to that of Group A. Its significant
coefficient in the interaction term (HOMOphenolate and Δ
ENCI-B) indicates that it is the defining parameter of the
model (Figure 8C), and can be viewed as the σ-donor
ability of the phenolate oxygen, described by
HOMOphenolate, as a function of the NCI interactions of the
aryl pendants (ΔENCI-B). This interaction term is consistent
with our empirical observations that weaker σ-donors
produced higher catalytic activity, as illustrated by 2,6-
Analysis
ACS Paragon Plus Environment

Page 13 of 17
Journal of the American Chemical Society
To better understand the impact of NCI effects within to Group A (Pauli repulsion). The most active catalysts of
1
2
3
4
5
6
7
8
phenols in the cross metathesis catalytic cycle, we
applied a kinetic model similar to that of Kozuch and
Shaik, who defined how the TOF of a catalytic system can
be calculated from the energy span (δE, in our case ΔΔG)
between the most stable intermediate and the highest
energy transition state of a catalytic cycle, to a small
subset of ligands of Group A (4, 6, 13, 20, 27).31 Thus,
energies of TBP and SP metallacycle geometries,
recognized as the resting state intermediates of olefin
cross metathesis, were calculated revealing a good
correlation between TOFin and the lowest energy
metallacycles (Figure 9A). Steric effects destabilize both
penta-coordinated metallacycle intermediates, reducing
the energy span (δE) of the overall reaction and
therefore increasing the catalytic activity (reaction rate).
However, very bulky (aryloxide) ligands destabilize the
TBP isomers less than off-cycle SP isomers because
bulkier aryloxide ligands are better accommodated in the
apical position of TBP intermediates for most imido
ligands.32 This preference for the TBP geometry with
increasing steric properties of the aryloxide ligands is
evident for both Groups A and B of calculated complexes
(Table S16) and is in line with trends revealed by the
DEACT parameter, where more stable catalysts are
associated with large aryloxide ligands. Alternatively, the
energy of the rate-determining transition state of the
catalytic cycle correlates with the stability of these
intermediates. Previous calculations revealed that the
cycloreversion step has the highest Gibbs free energy
barrier on the productive pathway of olefin metathesis.6
Group B feature higher stabilizing NCIs (ΔENCI-B <0,
Figure 6C), which possibly reduce the energy barrier of
the cycloreversion step. In contrast to ligands of Group
A, the ligands of Group B are large enough to
significantly influence the energy of the transition state
of the cycloreversion step through the NCIs. These NCIs
are increased by pendant -Ph (meta and para) and -Br
(para) substituents in 2,6-(Ph)2-ArOH. However, the
increase of sterics by bulkier ortho-pendants leads to a
drastic drop in catalytic performance, as also discussed
before for some bulkier phenols with ΔENCI-B >0, where
the ligand exchange step is too slow.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To test both hypotheses, we used previous
computational results involving the productive pathway
of olefin metathesis for MAP Mo and W d0 alkylidenes
catalysts,6c and found a strong correlation (R2 = 0.82)
between the calculated relative energies of TBP-isomers
of W and Mo complexes with different alkoxy ligands
and energy barriers of the cycloreversion step ΔΔG‡. That
said, no correlation is found using the transition state
energy ΔG‡ (R2 = 0.05) (Figure 9B). We speculate that the
size of substituents in aryloxy ligands could influence, via
NCI effects, the cycloreversion transition state to a much
less extent than the TBP/SP geometries, due to its late
character, explaining the previous correlations. We
conclude that for Group A, the changes caused by NCI in
the energy span (δE) between the cycloreversion step
and the resting state intermediates have the main
contribution from the energies of metallacyclobutanes.
As for Group B, two selected phenols (15 and 26) did not
follow the trend observed for TBP-SP metallacycles of
Group A. This result is not unexpected considering the
inverse linear trend between the two groups (Figure 4C).
The aryl arms on phenols of Group B generate CX---π (X
= H, halogen) NCI, which are of distinct nature compared
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 17
1
2
CONCLUSIONS
3
4
We have demonstrated that the efficiency of the MAP
Mo imido alkylidene metathesis catalysts depends on the
NCIs exerted by the aryloxy substituent ligands. Using
statistical tools, we categorized the data set of 35
phenols into two groups and provided evidence that
NCIs define catalysis: Group A containing phenols
without aryl arms in the ortho position (1-14, 16, 19-21,
27, 32), and Group B with ortho-arylated phenols (15,
17-18, 22-26, 28-31, 33-35). Overall catalytic
performances of in situ prepared MAP catalysts with
ligands of both Group A and B (as evaluated by TON1h) is
dominated, respectively, by repulsive and attractive
interactions (NCI), showing opposite trends with the
increase of the ortho pendant substituent size. Energetic
span analysis suggests an intrinsic relationship between
NCIs and the cycloreversion step. Specifically group A
ligands influence the energy of SP/TPB resting states
intermediates, while Group B ligands mainly impact the
transition state energy for the cycloreversion step. The
initial rates (TOFin) are also influenced by an electronic
effect of phenol ligands that is more pronounced for 2,6-
(Br)2- derivatives and 2,6-(Ph)2- derivatives, although with
different trends: 2,6-(Br)2- phenol derivatives display
decreasing initial rates with increasing electronegativity
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(reduced
σ donation ability), but the opposite is
observed for 2,6-(Ph)2- derivatives. While research efforts
in ligand design for MAP Mo imido alkylidene metathesis
catalysts have so far mainly focused on the -donation
ability of ligands, this work uncovers that the NCI is the
key driver for high catalyst activity in the cross-
metathesis reaction with d0 catalysts. To conclude,
although this work focused on a homodimerization
cross-metathesis reaction, we believe that the main
conclusions can guide catalyst selection in other cross-
metatheses reactions and related macrocyclic ring-
closing metatheses reactions.
AUTHOR INFORMATION
Corresponding Author
marco.ferreira@ufscar.br
fedoroal@ethz.ch
matt.sigman@utah.edu
ccoperet@ethz.ch
Figure 9. (A) Linear correlation between TOFin and TON1h
with e−|ΔG| where ΔG of the lowest energy conformer was
considered. Pyr = 2,5-dimethylpyrrolide. (B) Univariate
correlations between the cycloreversion step energies
and the TBP metallacycle. For panel B, data taken from
ref 6c.
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version
of the manuscript.
ACS Paragon Plus Environment

Page 15 of 17
Journal of the American Chemical Society
Monfort, X.; Copeŕet, C.; Eisenstein, O. Metallacyclobutanes
ACKNOWLEDGMENT
from Schrock-Type d0 Metal Alkylidene Catalysts: Structural
Preferences and Consequences in Alkene Metathesis.
Organometallics, 2015, 34, 1668. (d) Poater, A; Solans-
Monfort, X; Clot, E.; Copéret, C.; Eisenstein, O.
Understanding d0-Olefin Metathesis Catalysts:ꢀ Which
Metal, Which Ligands? J. Am. Chem. Soc. 2007, 129, 8207.
7. (a) Hérisson, J. L.; Chauvin, Y. Catalyse de transformation
des oléfines par les complexes du tungstène. II.
Télomérisation des oléfines cycliques en présence
d'oléfines acycliques. Makromol. Chem. 1971, 141, 161. (b)
Chauvin, Y. Olefin Metathesis: The Early Days (Nobel
Lecture). Angew. Chem., Int. Ed. 2006, 45, 3740.
1
2
3
4
5
6
7
8
The authors are grateful to the Scientific Equipment
Program of ETH Zürich and the SNSF (R’Equip grant
206021_150709/1) for financial support of the high
throughput catalyst screening facility (HTE@ETH). M.A.B.F
thanks Fundação de Amparo à Pesquisa do Estado de São
Paulo (FAPESP, proc. n° 17/13306-1, 15/08541-6 and
14/50249-8) for financial support. J.D.J.S. was supported by
the National Research Fund, Luxembourg (AFR Individual
Ph.D. Grant 12516655). M.S.S. thanks the National Science
Foundation (CHE-1763436) for support. The support and
resources from the Center for High Performance Computing
at the University of Utah are gratefully acknowledged. The
authors thank XiMo Hungary Ltd for donation of Mo-1,2
and selected phenol ligands.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
8. Gordon, C. P.; Yamamoto, K.; Liao, W.-C.; Allouche, F.;
Andersen, R. A.; Copéret, C.; Raynaud, C.; Eisenstein, O.
Metathesis Activity Encoded in the Metallacyclobutane
Carbon-13 NMR Chemical Shift Tensors. ACS Cent. Sci.
2017, 3, 759.
9. Solans-Monfort, X.; Copeŕet, C.; Eisenstein, O. Shutting
Down Secondary Reaction Pathways: The Essential Role of
the Pyrrolyl Ligand in Improving Silica Supported d0-ML4
Alkene Metathesis Catalysts from DFT Calculations. J. Am.
Chem. Soc. 2010, 132, 7750.
10. Leduc, A.-M.; Salameh, A.; Soulivong, D.; Chabanas, M.;
Basset, J.-M.; Copéret, C.; Solans-Monfort, X.; Clot, E.;
Eisenstein, O.; Böhm, V. P. W.; Röper, M. β-H Transfer from
the Metallacyclobutane: A Key Step in the Deactivation and
Byproduct Formation for the Well-Defined Silica-
Supported Rhenium Alkylidene Alkene Metathesis Catalyst.
J. Am. Chem. Soc., 2008, 130, 6288.
11. (a) Jiang, A. J.; Simpson, J. H.; Müller, P.; Schrock, R. R.
Fundamental Studies of Tungsten Alkylidene Imido
Monoalkoxidepyrrolide Complexes. J. Am. Chem. Soc.
2009, 131, 7770. (b) Flook, M. M.; Jiang, A. J.; Schrock, R. R.;
Müller, P.; Hoveyda, A. H. Z-Selective Olefin Metathesis
SUPORTING INFORMATION
Experimental procedures, catalytic data, plots of
conversion vs. time and selectivity vs. time for catalytic
experiments, NMR spectra, modeling details, and
Cartesian coordinates.
REFERENCES
1. (a) Grubbs, R. H., Ed.; Handbook of Metathesis; Wiley-VCH:
Weinheim, Germany, 2014. (b) Higman, C. S.; Lummiss, J. A.
M.; Fogg, D. E. Olefin Metathesis at the Dawn of
Implementation in Pharmaceutical and Specialty-Chemicals
Manufacturing. Angew. Chem., Int. Ed. 2016, 55, 3552.
2. Jiang, A. J.; Zhao, Y.; Schrock, R. R.; Hoveyda, A. H. Highly Z-
Selective Metathesis Homocoupling of Terminal Olefins. J.
Am. Chem. Soc. 2009, 131, 16630.
3. (a) Copéret, C.; Allouche, F.; Chan, K. W.; Conley, M. P. D.;
Murielle F.; Fedorov, A.; Moroz, I. B.; Mougel, V.; Pucino, M.;
Searles, K.; Yamamoto, K.; Zhizhko, P. A. Bridging the Gap
between Industrial and Well-Defined Supported Catalysts.
Angew. Chem. Int. Ed. 2018, 57, 6398. (b) Hoveyda, A. H.
Evolution of Catalytic Stereoselective Olefin Metathesis:
Processes
Catalyzed
by
a
Molybdenum
Hexaisopropylterphenoxide Monopyrrolide Complex. J.
Am. Chem. Soc. 2009, 131, 7962. (c) Schrock, R. R.; Jiang, A.
J.; Marinescu, S. C.; Simpson, J. H.; Möller, P. Fundamental
Studies of Molybdenum and Tungsten Methylidene and
Metallacyclobutane Complexes. Organometallics 2010, 29,
5241. (d) Mougel, V.; Copéret, C. Magnitude and
consequences of OR ligand σ -donation on alkene
From
Ancillary
Transformation
to
Purveyor
of
Stereochemical Identity. J. Org. Chem., 2014, 79, 4763.
4. Meek1, S. J.; O’Brien, R. V.; Llaverial, J.; Schrock, R. R.;
Hoveyda, A. H. Catalytic Z-selective olefin cross-metathesis
for natural product synthesis. Nature, 2011, 471, 461.
5. (a) Koh, M. J.; Nguyen, T. T.; Lam, J. K.; Torker, S.; Hyvl, J.;
Schrock, R. R.; Hoveyda, A. H. Molybdenum chloride
catalysts for Z-selective olefin metathesis reactions. Nature
2017, 542, 80. (b) Lam, J. K.; Zhu, C.; Bukhryakov, K. V.;
Müller, P.; Hoveyda, A. H.; Schrock, R. R. Synthesis and
Evaluation of Molybdenum and Tungsten Monoaryloxide
Halide Alkylidene Complexes for Z-Selective Cross-
Metathesis of Cyclooctene and Z-1,2-Dichloroethylene. J.
Am. Chem. Soc. 2016, 138, 15774.
metathesis
activity
in
d0
silica
supported
(≡SiO)W(NAr)(=CHtBu)(OR) catalysts. Chem. Sci. 2014, 5,
2475.
12. (a) Hock, A. S.; Schrock, R. R.; Hoveyda, A. H. Dipyrrolyl
Precursors to Bisalkoxide Molybdenum Olefin Metathesis
Catalysts. J. Am. Chem. Soc. 2006, 128, 16373. (b) Singh, R.;
Schrock, R. R.; Müller, P.; Hoveyda, A. H. Synthesis of
Monoalkoxide
Monopyrrolyl
Complexes
Mo(NR)(CHR‘)(OR‘ꢀ‘)(pyrrolyl): Enyne Metathesis with High
Oxidation State Catalysts. J. Am. Chem. Soc. 2007, 129,
12654. (c) Jiang, A. J.; Zhao, Y.; Schrock, R. R.; Hoveyda, A.
H. Highly Z-Selective Metathesis Homocoupling of
Terminal Olefins. J. Am. Chem. Soc. 2009, 131, 16630.
6. (a) Solans-Monfort, X.; Clot, E.; Copeŕet, C.; Eisenstein, O. d0
Re-Based
Olefin
Metathesis
Catalysts,
Re(≡CR)(=CHR)(X)(Y):ꢀ The Key Role of X and Y Ligands for
Efficient Active Sites. J. Am. Chem. Soc. 2005, 127, 14015.
(b) Solans-Monfort, X.; Copeŕet, C.; Eisenstein, O. Oxo vs
Imido Alkylidene d0-Metal Species: How and Why Do They
Differ in Structure, Activity, and Efficiency in Alkene
Metathesis? Organometallics, 2012, 31, 6812. (c) Solans-
13. (a) Mougel, V.; Santiago, C. B.; Zhizhko, P. A.; Bess, E. N.;
Varga, J.; Frater, G.; Sigman, M. S.; Copeŕet, C.
Quantitatively Analyzing Metathesis Catalyst Activity and
Structural Features in Silica-Supported Tungsten Imido–
Alkylidene Complexes. J. Am. Chem. Soc. 2015, 137, 6699.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 17
(b) Engl, P. S.; Santiago, C. B.; Gordon, C. P.; Liao, W.-C.;
23. (a) Neel, A. J.; Hilton, M. J.; Sigman, M. S.; Toste, F. D.
1
2
3
4
5
6
7
8
Fedorov, A., Copeŕet, C.; Sigman, M. S.; Togni, A. Exploiting
and Understanding the Selectivity of Ru-N-Heterocyclic
Carbene Metathesis Catalysts for the Ethenolysis of Cyclic
Olefins to α,ω-Dienes. J. Am. Chem. Soc. 2017, 139, 13117.
14. Zhizhko, P. A.; Mougel, V.; Silva, J. J. Copéret, C.
Benchmarked Intrinsic Olefin Metathesis Activity: Mo vs. W.
Helv. Chim. Acta 2018, 101, e1700302.
15. Singh, R. Cyclopropene Polymerization and Enyne
Metathesis Catalyzed by High Oxidation State
Molybdenum Alkylidenes. PhD thesis. Massachusetts
Institute of Technology, 2008.
16. (a) Guo, J.-Y.; Minko, Y.; Santiago, C. B.; Sigman, M. S.
Developing Comprehensive Computational Parameter Sets
To Describe the Performance of Pyridine-Oxazoline and
Related Ligands. ACS Catal. 2017, 7, 4144. (b) Santiago, C.
B.; Guo, J.-Y.; Sigman, M. S. Predictive and mechanistic
multivariate linear regression models for reaction
development. Chem. Sci., 2018, 9, 2398.
17. (a) Santiago, C. B.; Milo, A.; Sigman, M. S. Developing a
Modern Approach To Account for Steric Effects in
Hammett-Type Correlations. J. Am. Chem. Soc. 2016, 138,
13424. (b) Hollingsworth, C. A.; Seybold, P. G.; Hadad, C. M.
Substituent effects on the electronic structure and pKa of
benzoic acid. Int. J. Quantum Chem. 2002, 90, 1396.
18. Verloop, A. in Drug Design, E. J. Ariens, Ed. (Academic
Press, 1976), vol. 3, pp. 133–187.
19. Sevov, C. S.; Hickey, D. P.; Cook, M. E.; Robinson, S. G.;
Barnett, S.; Minteer, S. D.; Sigman, M. S.; Sanford, M. S.
Physical Organic Approach to Persistent, Cyclable, Low-
Potential Electrolytes for Flow Battery Applications. J. Am.
Chem. Soc., 2017, 139, 2924.
20. (a) Hillier, A. C.; Sommer, W. J.; Yong, B. S.; Petersen, J. L.;
Cavallo, L.; Nolan, S. P. A Combined Experimental and
Theoretical Study Examining the Binding of N-Heterocyclic
Carbenes (NHC) to the Cp*RuCl (Cp* = η5-C5Me5) Moiety:ꢀ
Insight into Stereoelectronic Differences between
Unsaturated and Saturated NHC Ligands. Organometallics
2003, 22, 4322. (b) Clavier, H.; Nolan, S. P. Percent buried
volume for phosphine and N-heterocyclic carbene ligands:
steric properties in organometallic chemistry. Chem.
Commun. 2010, 46, 841.
Exploiting non-covalent π interactions for catalyst design.
Nature, 2017, 543, 637. (b) Wagner, J. P.; Schreiner, P. R.
London Dispersion in Molecular Chemistry—Reconsidering
Steric Effects. Angew. Chem. Int. Ed. 2015, 54, 12274.
24. Bohm, S.; Exner, O. Steric effects of polar substituents
evaluated in terms of energy by means of isodesmic
reactions. Org. Biomol. Chem. 2008, 6, 1092.
25. Wheeler, S. E.; Houk, K. N.; Schleyer, P. v. R.; Allen, W. D. A
Hierarchy
of
Homodesmotic
Reactions
for
9
Thermochemistry. J. Am. Chem. Soc., 2009, 131, 2547.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
26. Cozzi, F.; Cinquini, M.; Annuziata, R.; Dwyer, T.; Siegel, J.
Polar/.pi. interactions between stacked aryls in 1,8-
diarylnaphthalenes. J. Am. Chem. Soc. 1992, 114, 5729.
27. (a) Robbins, J.; Bazan, G. C.; Murdzek, J. S.; O'Regan, M. B.;
Schrock, R. R. Reduction of molybdenum imido-alkylidene
complexes in the presence of olefins to give
molybdenum(IV) complexes. Organometallics 1991, 10,
2902. (b) Lopez, L. P. H.; Schrock, R. R.; Müller, P. Dimers
that Contain Unbridged W(IV)/W(IV) Double Bonds.
Organometallics 2006, 25, 1978.
28. Sinha, A.; Lopez, L. P. H.; Schrock, R. R.; Hock, A. S.; Müller,
P. Reactions of M(N-2,6-i-Pr2C6H3)(CHR)(CH2R‘)2 (M = Mo,
W) Complexes with Alcohols To Give Olefin Metathesis
Catalysts
of
the
Type
M(N-2,6-i-
Pr2C6H3)(CHR)(CH2R‘)(OR‘ꢀ‘). Organometallics 2006, 25,
1412.
29. Coe, B. J.; Glenwright, S. J. Trans-effects in octahedral
transition metal complexes. Coord. Chem. Rev. 2000, 203,
5.
30. The interaction between ortho-Br substituent in an aryloxy
ligand and Mo (Mo---Br) was demonstrated, see: Meek, S.
J.; Malcolmson, S. J.; Li, B.; Schrock, R. R.; Hoveyda, A. H.
The
Significance
of
Degenerate
Processes
to
Enantioselective Olefin Metathesis Reactions Promoted by
Stereogenic-at-Mo Complexes. J. Am. Chem. Soc. 2009,
131, 16407.
31. Kozuch, S.; Shaik, S.
A Combined Kinetic−Quantum
Mechanical Model for Assessment of Catalytic Cycles:ꢀ
Application to Cross-Coupling and Heck Reactions. J. Am.
Chem. Soc. 2006, 128, 3355.
32. (a) Gerber, L. C. H.; Schrock, R. R.; Müller, P. Molybdenum
and Tungsten Monoalkoxide Pyrrolide (MAP) Alkylidene
21. Toste, F. D.; Sigman, M. S.; Miller, S. J. Pursuit of
Noncovalent Interactions for Strategic Site-Selective
Catalysis. Acc. Chem. Res. 2017, 50, 609.
Complexes That Contain
a
2,6-Dimesitylphenylimido
Ligand. Organometallics 2013, 32, 2373. (b) Axtell, J. C.;
Schrock, R. R.; Müller, P.; Hoveyda, A. H. Synthesis of
Molybdenum and Tungsten Alkylidene Complexes That
Contain the 2,6-Bis(2,4,6-triisopropylphenyl)phenylimido
(NHIPT) Ligand. Organometallics 2015, 34, 2110.
22. (a) Orlandi, M.; Hilton, M. J.; Yamamoto, E.; Toste, F. D.;
Sigman, M. S. Mechanistic Investigations of the Pd(0)-
Catalyzed Enantioselective 1,1-Diarylation of Benzyl
Acrylates. J. Am. Chem. Soc. 2017, 139, 12688. (b) Orlandi,
M.; Coelho, J. A. S.; Hilton, M. J.; Toste, F. D.; Sigman, M. S.
Parametrization of Non-covalent Interactions for Transition
State Interrogation Applied to Asymmetric Catalysis. J. Am.
Chem. Soc., 2017, 139, 6803.
ACS Paragon Plus Environment

Page 17 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17
ACS Paragon Plus Environment