Catalyst Behavior in Metal-Catalyzed Carbonyl-Olefin Metathesis

Article abstract of DOI:10.1021/jacs.9b02613

Iron(III)-catalyzed carbonyl-olefin ring-closing metathesis employs reactivity not typically observed in Lewis acid-catalyzed reactions. In converting a ketone with a pendant olefin into a cycloalkene and a simple carbonyl byproduct, the reaction requires the Lewis acid catalyst to differentiate between the carbonyl of the substrate and that of the byproduct. It is necessary to determine how this solution interaction imparts the desired reactivity to best employ this method. Herein, we report detailed kinetic, spectroscopic, and colligative measurements applied toward the identification of the solution structures of the active Fe(III) and Ga(III) carbonyl-olefin metathesis catalysts. These data are consistent with formation of Lewis acid-carbonyl pairs for both metal systems under stoichiometric conditions. However, they diverge in the presence of higher equivalents of carbonyl, with Fe(III) forming highly ligated complexes, and no observed change for Ga(III). These findings are consistent with the resting state identity of the Fe(III) metathesis catalyst changing over the course of the reaction.

Full text of DOI:10.1021/jacs.9b02613

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Article
Catalyst Behavior in Metal-Catalyzed Carbonyl-Olefin Metathesis
Carly S. Hanson, Mary C. Psaltakis, Janiel J. Cortes, and James J. Devery, III
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b02613 • Publication Date (Web): 05 Jul 2019
Downloaded from http://pubs.acs.org on July 5, 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 11
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Catalyst Behavior in Metal-Catalyzed Carbonyl-Olefin Metath-
esis.
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
Carly S. Hanson, Mary C. Psaltakis, Janiel J. Cortes, and James J. Devery, III*
Department of Chemistry & Biochemistry, Loyola University Chicago, Flanner Hall, 1068 W Sheridan Road, Chicago, IL
60660, United States
ABSTRACT: Iron(III)-catalyzed carbonyl-olefin ring-closing metathesis employs reactivity not typically observed in Lewis acid-
catalyzed reactions. In converting a ketone with a pendant olefin into a cycloalkene and a simple carbonyl byproduct, the reaction
requires the Lewis acid catalyst to differentiate between the carbonyl of the substrate and that of the byproduct. It is necessary to
determine how this solution interaction imparts the desired reactivity in order to best employ this method. Herein, we report detailed
kinetic, spectroscopic, and colligative measurements applied towards the identification of the solution structures of the active Fe(III)
and Ga(III) carbonyl-olefin metathesis catalysts. These data are consistent with formation of Lewis acid-carbonyl pairs for both metal
systems under stoichiometric conditions. However, they diverge in the presence of higher equivalents of carbonyl, with Fe(III) form-
ing highly ligated complexes, and no observed change for Ga(III). These findings are consistent with the resting state identity of the
Fe(III) metathesis catalyst changing over the course of the reaction.
O
A
Introduction
FeCl3 (5 mol%)
DCE, rt
O
Me
Me
+
Ph
The interactions between Lewis acids and carbonyls have
played a significant role in the construction of important mole-
cules.1 While a great deal of insight has been gained regarding
classical stoichiometric regimes, like the Friedel-Crafts reac-
tion, more discoveries continue to be made about the complex-
ities of these interactions between carbonyls and Lewis acids in
catalytic systems. In particular, the new reactivity observed in
Lewis acid-catalyzed carbonyl-olefin metathesis demonstrates
that a comprehensive understanding of the interactions between
these classical Lewis pairs remains incomplete. Representing a
powerful reaction manifold for the production of C=C bonds
from functional groups that are broadly utilized in the construc-
tion of complex molecules,2 the Fe(III)-catalyzed process de-
veloped by Schindler and coworkers3 triggered a series of syn-
thetic developments that expand the use of Fe(III),4 employ
Ga(III),5 I2,6 as well as employing Brønsted acids.7 This method
has proven useful in the synthesis of many cyclic scaffolds via
ring closing, including di- and trisubstituted cyclopentenes and
cyclohexenes, polycyclic aromatic hydrocarbons, 2,5-dihydro-
pyrroles, as well as ring-opening metathesis and cross metathe-
sis. One of the major benefits of the transformation is the sim-
plicity of execution, requiring only catalyst, solvent, and a car-
bonyl-olefin pair: a concise list of variables for mechanistic
analysis. However, one question remains unexplored: how does
the catalyst distinguish between the substrate carbonyl and
product carbonyl? We report herein the solution behavior of the
Lewis acid catalyst in the presence of a model substrate and typ-
ical carbonyl byproducts. The application of in situ infrared
spectroscopy displays formation of Lewis acid-dependent ag-
gregates, which competitively inhibit the metathesis cycle.
These mechanistic findings provide insight into procedural
modifications to facilitate the conversion of recalcitrant metath-
esis substrates.
Me Me
Ph
R
R
1
2
3
B
O
O
Me
Me
Me Me
Ph
3
1
R
FeCl3
carbonyl
exchange
FeCl3
O
Ph
O
R
Me
Me
2
Me Me
Ph
R
7
4
retro-[2+2]
Me
Me
Me
Me
Cl3Fe
O
Cl3Fe
O
Ph
Ph
asynchronous
concerted
[2+2]-cycloaddition
R
6
R
5
Ph
C
D
FeCl3 (20 mol%)
allylTMS (5 equiv)
N
Ts
O
Ph
N
enhance carbonyl
exchange
Ph
Ts
Ph
9, 91%
10, 90%
8
O
FeCl3 (5 mol%)
DCE, rt
O
R1
+
Ph
R1 R2
Ph
R2
CO2Et
CO2Et
12
11a, R1 = R2 = Me
3, acetone
13, benzaldehyde
14, acetophenone
a, 99%
b, 60%
c, 49%
11b, R1 = Ph, R2 = H
11c, R1 = Ph, R2 = Me
Figure 1. Fe(III)-catalyzed carbonyl-olefin metathesis (A). Cata-
lytic cycle (B). Additive-facilitated metathesis (C). Byproduct-
linked diminished yields (D).
Previous efforts from our lab, working alongside the
Schindler and Zimmerman labs, focused on the determination
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 11
A
of the operating catalytic cycle that results in formation of di-
O
substituted cyclopentene 2 and acetone (3, Figure 1A).8 We pro-
posed that iron(III) chloride forms coordination complex 4, re-
sulting in an interaction that activates the carbon-oxygen double
bond. The activated complex then undergoes an asynchronous,
concerted [2+2]-cycloaddition to form oxetane-complex 5 as
the turnover-limiting step. Fe(III)-mediated retro-[2+2] cy-
cloaddition yields cyclopentene 2 and Fe(III) complex (7). Car-
bonyl exchange allows for the subsequent catalytic turnover.
MCl3 (5 mol%)
1
2
3
4
5
6
7
8
O
Me
Me
+
Ph
DCE (0.025 M), 0 ºC
Me Me
Ph
Me
Me
16
3
15, (1 equiv)
M = Fe or Ga
B
20
The final step of the cycle is critical for the success of the
catalytic mechanism, suggesting that product inhibition is
likely, under reaction conditions. Indeed, Li and coworkers re-
ported the use of an additive to facilitate the formation of dihy-
dropyrroles (9) from N-cinnamyl glycine derivatives (8, Figure
1C).4a Utilizing a styrenyl olefin partner that results in the for-
mation of benzaldehyde as the carbonyl byproduct, they ob-
served that product formation required the addition of super-
stoichiometric allyltrimethylsilane. This additive facilitated the
formation of the desired heterocycle, as well as diallylated 10.9
Similarly, we observed that systems that formed benzaldehyde
or acetophenone as the byproduct provided diminished yields
(Figure 1D). In addition to these potential effects of byproduct,
we have also demonstrated that the addition of exogenous
Lewis bases to the reaction mixture eliminates metathesis reac-
tivity.8 Further, Schindler, Zimmerman, and coworkers showed
that Lewis basic moieties within the substrate have the potential
to inhibit the preferred reactivity.4c These collected observa-
tions are consistent with byproduct inhibition occurring via in-
efficient carbonyl exchange.
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
10
0
0
1000
2000
3000
Time (s)
Figure 2. Metal halide-catalyzed carbonyl-olefin metathesis of 15
(A). FeCl3-mediated system (■), GaCl3-mediated system (●),
FeCl3-mediated system with 0.5 equiv 3 (□), GaCl3-mediated sys-
tem with 0.5 equiv 3 (○) (B) Error bars omitted for clarity.11
impacted by the presence of byproduct compared to Ga(III). In-
triguingly, the slopes of the decays suggest that the Ga-cata-
lyzed system appears to approach a constant rate whether or not
the reaction is initiated in the presence of 3; whereas, this ob-
servation is not present in the Fe reaction, suggesting different
affinities for the interactions of Fe and Ga with 3 and 15.
Results and Discussion
Kinetic Analysis: Our initial efforts to elucidate the presence of
byproduct inhibition began with observation of the metathesis
reaction under synthetically relevant conditions. Using the re-
action defined in Figure 2A, we employed aromatic ketone 15
as the substrate for the metathesis reaction in DCE.10 The ex-
traction of kinetic information occurred by monitoring the [15]
via reversed-phase ultra-performance liquid chromatography
coupled with a transmission UV/vis detector. We performed re-
actions by first combining a metal halide with DCE. Then, ca-
talysis was initiated via the addition of 15 to the mixture. To
obtain a baseline for this particular reaction, we examined an
FeCl3-catalyzed system (■, Figure 2B), as well as a GaCl3-cat-
alyzed system (●, Figure 2B). The Fe(III) reaction displays a
significantly faster rate than Ga(III), as we have previously re-
ported.8
Spectroscopic Investigation: The contrasting observations we
confronted for the FeCl3- and GaCl3-catalyzed systems necessi-
tated an in-depth examination of the interactions of these cata-
lysts with carbonyls. These interactions have played a signifi-
cant role in the construction of important molecules.1 Because
of their widespread application, significant effort has been de-
voted to characterizing the behavior of Lewis acids and bases,
relying heavily on infrared (IR) spectroscopy to elucidate the
discrete structure of Lewis pairs, and to utilize IR as a tool for
the determination of Lewis acidity.1d,12 In particular, the Susz
lab studied the stoichiometric coordinating interactions of
Lewis acids and carbonyls in great detail in the solid state. They
employed IR and elemental analysis to determine the composi-
tion of neat, 1:1 mixtures of Lewis pairs, yielding a great deal
of structural information about the interactions of simple ke-
tones and aldehydes with a range of Lewis acids. These data
form a solid foundation for the spectroscopic behavior of 3, ben-
zaldehyde (13), and acetophenone (14) in the presence of
AlCl3,13 TiCl4,14 BF3,15 and FeCl3.15
Next, we initiated the reaction in the presence of 0.5 equiv 3
with respect to 15. This modification was accomplished by pre-
mixing 3 with the salt/DCE slurry prior to addition of 15. In the
presence of 0.5 equiv added byproduct 3, we observe significant
inhibition of catalytic activity for the Fe(III)-mediated system,
decreasing the reaction to a rate slower than the GaCl3 process.
Similarly, we observe a decrease in rate when the GaCl3-cata-
lyzed reaction is initiated in the presence of 3; however, the de-
crease is less significant compared to that observed when Fe(III)
is employed. We continued our examination of the impact of 3
on the rate of reaction, probing 0.2 equiv 3 with respect to 15.11
In both systems, we observe a concentration-dependent delete-
rious effect on the rate of reaction arising from the presence of
3; however, in neither case is this decrease consistent with a
whole number rate order. Collectively, these data are consistent
with catalyst behavior being impacted by the identity of the
Lewis acid. Under typical reaction conditions for 15, Fe(III) is
a much more efficient catalyst. However, its rate is significantly
With this wealth of spectroscopic information in the solid
state available as a starting point, we began to examine the so-
lution interactions of GaCl3 and FeCl3 in combination with 3,
13, and 14. Using solution IR, we posited that we would be able
to compare the relative amounts of free carbonyl compound
with the complex formed between the carbonyl and Lewis acid.
Under anhydrous conditions, GaCl3 and FeCl3 have very differ-
ent solubilities, with GaCl3 rapidly dissolving and FeCl3 being
largely insoluble in DCE. We prepared mixtures of these salts
to which we titrated carbonyl incrementally. Our spectroscopic
investigation began with the examination of the interaction be-
tween GaCl3 and 3. Intriguingly, between 0 and 1 equiv 3 added
to the solution, we observe no unbound 3. In the carbonyl region
ACS Paragon Plus Environment

Page 3 of 11
Journal of the American Chemical Society
0-1 equiv carbonyl
1
2
3
4
5
6
0.4
0.3
0.2
0.1
0.0
0.4
0.3
0.2
0.1
0.0
0.4
C
A
B
GaCl3
GaCl3
GaCl3
O
0.3
0.2
0.1
0.0
O
O
Me Me
Ph
H
Ph Me
GaCl3
17
18
19
7
8
1700
1600
1500
1800
1700
1600
1500
1800
1700
1600
1500
9
Wavenumber (cm-1)
Wavenumber (cm-1)
Wavenumber (cm-1)
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
0.4
0.3
0.2
0.1
0.0
0.4
0.3
0.2
0.1
0.0
0.4
0.3
0.2
0.1
0.0
FeCl3
FeCl3
FeCl3
D
E
F
O
O
O
Ph
20
H
Ph Me
Me Me
21
7
FeCl3
1800
1700
1600
-1
1500
1800
1700
1600
1500
1700
1600
1500
Wavenumber (cm-1)
Wavenumber (cm-1)
Wavenumber (cm
)
> 1 equiv carbonyl
0.4
0.3
0.2
0.1
0.0
0.4
0.3
0.2
0.1
0.0
0.4
G
H
I
0.3
0.2
0.1
0.0
GaCl3
1700
1600
1500
1800
1700
1600
1500
1800
1700
1600
1500
Wavenumber (cm-1)
Wavenumber (cm-1)
Wavenumber (cm-1)
0.4
0.4
0.3
0.2
0.1
0.0
0.4
0.3
0.2
0.1
0.0
J
K
L
0.3
0.2
0.1
0.0
FeCl3
1800
1700
1600
1500
1700
1600
1500
1800
1700
Wavenumber (cm-1)
1600
1500
Wavenumber (cm-1)
Wavenumber (cm-1)
Figure 3. Solution IR data for titrations of GaCl3 (1 mmol in 6 mL DCE) and FeCl3 (2 mmol in 12 mL DCE) with 0-1 equiv 3 (A and D),
13 (B and E), and 14 (C and F), as well as >1 equiv 3 (G and J), 13 (H and K), and 14 (I and L). Titrations proceed from red to violet with
increasing amounts of titrant. (A: [3] = 0 M, 0.023 M, 0.045 M, 0.067 M, 0.112 M, 0.156 M. B: [13] = 0 M, 0.049 M, 0.081 M, 0.097 M,
0.129 M, 0.160 M. C: [14] = 0 M, 0.029 M, 0.057 M, 0.085 M, 0.113 M, 0.155 M. D: [3] = 0 M, 0.034 M, 0.067 M, 0.101 M, 0.134 M,
0.178 M. E: [13] = 0 M, 0.033 M, 0.057 M, 0.081 M, 0.105 M, 0.129 M, 0.161 M. F: [14] = 0 M, 0.021 M, 0.043 M, 0.071 M, 0.113 M,
0.168 M. G: [3] = 0.222 M, 0.287 M, 0.351 M, 0.415 M, 0.499 M, 0.786 M. H: [13] = 0.177 M, 0.224 M, 0.270 M, 0.316 M, 0.392 M, 0.482
M. I: [14] = 0.196 M, 0.277 M, 0.343 M, 0.473 M, 0.599 M, 0.721 M. J: [3] = 0.233 M, 0.265 M, 0.319 M, 0.383 M, 0.447 M, 0.593 M. K:
[13] = 0.177 M, 0.224 M, 0.270 M, 0.316 M, 0.541 M, 0.824 M. L: [14] = 0.182 M, 0.263 M, 0.330 M, 0.460 M, 0.523 M, 0.586 M.).11
of the spectrum, we observe exclusive formation of a signal at
1630 cm-1 (Figure 3A). When a similar titration was performed
with FeCl3, again no unbound 3 was observed, with exclusive
formation of a vibration at 1633 cm-1 (Figure 3D). Importantly,
the FeCl3 system remains heterogeneous until 1 equiv 3 is pre-
sent. We performed an analogous titration with 13 and GaCl3
and observed exclusive formation of a single species with ab-
sorbances at 1610, 1596, and 1573 cm-1 (Figure 3B). The corre-
sponding titration into an FeCl3 slurry resulted in the formation
of a species with vibrations at 1610, 1592, and 1569 cm-1 (Fig-
ure 3E). Again, the FeCl3 system remains heterogeneous until 1
equiv 13 is present in solution with respect to metal halide.
When the titrant is 14, GaCl3 yields 1603, 1588, and 1563 cm-1
(Figure 3C), while FeCl3 provides 1603, 1589, and 1558 cm-1
(Figure 3F). The presence of 1 equiv 14 with respect to FeCl3
generates a homogeneous solution.
In their study of the interactions of Lewis acids and bases, the
Susz lab prepared a number of 1:1 complexes of Lewis acids
and carbonyls as solids. They examined these structures via el-
emental analysis to determine their composition and IR to de-
termine the manner of interaction between acid and base. They
report a 1:1 complex between 3 and BF3 (1640 cm-1),15a as well
as with 3 and TiCl4 (1625 cm-1).14b Further, Greenwood meas-
ured the heat of formation of a 1:1 GaCl3-3 complex at -15.3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 11
1, 21 = 1558 cm-1) by multiplying the signal by the volume pre-
sent in each measurement yielding eq. 1:
0.0010
0.0008
0.0006
0.0004
0.0002
0.0000
A
1
2
3
4
5
6
7
ꢀꢁ = ꢂꢃꢄ
(1)
where 1) both absorbance (A) and volume (V) are measurable
terms; 2) molar absorptivity (e) and pathlength (l) are constant,
allowing 3) number of moles (n) to be examined. Using eq. 1,
we can examine the observed amount of each component as a
function of equivalents of carbonyl added (Figure 4).
17
3
0
1
2
3
3 Equiv
8
9
When we consider the addition of 3 to GaCl3, we see several
key observations (Figure 4A). The amount of 17 (●) increases
proportionately with the amount of 3 from 0 equiv 3 until ≈1
equiv 3 has been added. After an equivalent amount of 3 with
respect to GaCl3 is present, no significant change in amount of
17 occurs, concomitant with observation of 3 in solution (■).
Analogous features are observed for the addition of 13 to GaCl3
and 14 to GaCl3.11 Similarly, the titrations of carbonyls into the
FeCl3 mixture display proportional growth in the amount of 7
(Figure 4B), 20, and 21.11 In all three cases, the amount of 1:1
Lewis acid-carbonyl complex increases until 1 equiv carbonyl
is present in solution. However, after the addition of 1 equiv
carbonyl, the Fe-containing systems diverge from their Ga-con-
taining counterparts. In all three cases, the amount of coordina-
tion complex decreases with increasing equivalents of carbonyl.
0.0025
0.0020
0.0015
0.0010
0.0005
0.0000
B
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
7
3 observed
3 theoretical
0
1
2
3
3 (equiv)
Figure 4. Analysis of components: 3 and 17 (A) as well as 3,
theoretical 3 after 1 equiv, and 7 (B).11
kcal mol-1,16 consistent with our observed absence of unbound
3. These precedents are consistent with our observations be-
tween 3 and both GaCl3 and FeCl3, suggesting that between 0
and 1 equiv 3, a 1:1 coordination complex forms exclusively
between 3 and GaCl3 (17), as well as 3 and FeCl3 (7). The sim-
ilar behavior observed in the systems employing 13 and 14 is
suggestive of an analogous complexation, where 13 forms ben-
zaldehyde-GaCl3 complex 18 and benzaldehyde-FeCl3 complex
20. The corresponding addition of 14 forms acetophenone-
GaCl3 complex 19 and acetophenone-FeCl3 complex 21 with
GaCl3 and FeCl3, respectively. Our observation of 21 is con-
sistent with the solid state IR reported by the Susz lab.15b Fur-
ther, Kochi and coworkers reported a crystal structure of an
analogous GaCl3 complex with 4-fluorobenzoyl chloride.17
At this point, it is important to consider our collected obser-
vations from all six titrations. 1) Under anhydrous conditions,
GaCl3 is soluble in DCE while FeCl3 is not. 2) The solubility of
FeCl3 increases with increasing amount of added carbonyl, re-
gardless of identity. 3) Upon the addition of 1 equiv carbonyl,
all three Fe systems achieve homogeneity. 4) Solution IR of all
six systems displays vibrations consistent with the formation of
1:1 Lewis acid-carbonyl coordination complexes, similar to the
IR spectra reported by Susz for identical/analogous structures.
5) In all cases, no uncoordinated carbonyl is observed until the
first equivalence point is reached. Collectively, these observa-
tions suggest that, beginning at 1 equiv carbonyl added, the sig-
nal of uncoordinated titrant should be representative of the total
amount added. The amounts of 3, 13, and 14 observed in each
of the titrations are consistent with this statement, suggesting
that once each GaCl3 molecule is ligated by a molecule of car-
bonyl, no further measurable interaction occurs. Alternatively,
the suggestion that all carbonyl added after the equivalence
point should be observable does not hold true for the FeCl3 sys-
tems. When the titration of FeCl3 with 3 is considered, we know
what the signal of 3 should be based on observation in the ab-
sence of FeCl3 (Figure 4B).11 The amount of 3 observed (■) is
less than expected (□). This observation is similarly true for the
observed amounts of 13 as well as the observed amounts of 14.11
When these observations are taken in context with our data that
include the formation of new spectral features concomitant with
decrease in the intensity of coordination complex, they suggest
that 7, 20, and 21 are being consumed and converted to alternate
complexes.
Interestingly, when the addition of the carbonyl proceeds be-
yond 1 equiv with respect to the metal halide, the behavior of
the two systems diverges. For GaCl3, when >1 equiv 3, 13, and
14 are added to the solution, we observe the presence of un-
bound carbonyl: 1714 cm-1 for 3 (Figure 3G), 1704 cm-1 for 13
(Figure 3H), and 1685 cm-1 for 14 (Figure 3I). When this same
range of 3 is added to FeCl3, more complex spectra are obtained.
When 3 is added beyond 1 equiv, we observe an isosbestic
point at 1648 cm-1 (Figure 3J). The C=O vibration of free 3 is
observed at 1714 cm-1; however, we also observe the signal at
1644 cm-1 decrease in intensity while a new signal at 1663 cm-
1
forms. Similarly, when 13 is added beyond 1 equiv, an isos-
bestic point is observed at 1574 cm-1 (Figure 3K). Simultane-
ously, as the intensity of the signal for free 13 grows (1704 cm-
1), the signal at 1569 cm-1 decreases while 1577 cm-1 and 1626
cm-1 grow. For the addition of superstoichiometric 14 to FeCl3,
an isosbestic point is present at 1566 cm-1 (Figure 3L). The vi-
bration for 21 at 1558 cm-1 decreases as more 14 is observed at
1685 cm-1. Further, the range between 1670 and 1610 cm-1
grows.
For the FeCl3 systems, once the equivalence point is reached,
the amount of 1:1 coordination complex decreases with increas-
ing equivalents of carbonyl. The data presented so far are con-
sistent with the amount of carbonyl added being equivalent to
the amount of 1:1 coordination complex formed for the titration
range from 0 to 1 equiv carbonyl. Additionally, the maximal
amount of complex is defined by the moles of metal halide pre-
sent (CMAX = 2 mmol FeCl3). With this relationship, we can treat
the moles of carbonyl added (CADD) as equal to the moles of
coordination complex (CCOORD) to develop a Beer-Lambert re-
lationship with absorbance.11 We can use this relationship
To gain more insight into the difference in behavior of GaCl3
and FeCl3, we examined the amounts of each component pre-
sent in solution with respect to the equivalents of carbonyl
added. Because data are collected via titration, dilution is a fac-
tor for which we must account. We accomplish this task through
normalization of the absorbance of the lmax of each component
(3 = 1714 cm-1, 7 = 1644 cm-1, 13 = 1704 cm-1, 14 = 1685 cm-1,
17 = 1633 cm-1, 18 = 1573 cm-1, 19 = 1563 cm-1, 20 = 1569 cm-
ACS Paragon Plus Environment

Page 5 of 11
Journal of the American Chemical Society
0.0020
0.0015
0.0010
0.0005
0.0000
1500
1250
A
B
1
2
3
4
5
6
7
8
9
Titrant
Slope
1000
750
500
250
0
3
3.54 ± 0.04
2.88 ± 0.05
3.67 ± 0.08
13
14
FeCl3
GaCl3
0.0000 0.0002 0.0004 0.0006 0.0008
7 CMAX-CCOORD (mol)
0
1
2
3
4
5
6
Figure 5. Undetected carbonyl CND vs. consumed coordination
complex for 3 and 7 (A), and the slopes of the correlations for 3
and 7, 13 and 20, as well as 14 and 21 (B).11
3 (equiv)
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 7. Conductivity of FeCl3 (2 mmol in 12 mL DCE) and
GaCl3 (2 mmol in 12 mL DCE) with increasing amounts of 3.11
O
yCl-
O
FeCl3
FeCl(3-y)
Ph
Ph
H
O
Me Me
Me Me
O
O
H
FeCl3
R1 R2
Ph
O
R1 R2
O
1+x
FeIII
>1 equiv
22
23
0 - 1 equiv
H
Cl
O
Cl
Figure 6. Proposed solution behavior of FeCl3 in the presence of
carbonyl byproducts in DCE.
Ph
H
24
observed in the 0-1 equiv region to determine CCOORD for obser-
vations >1 equiv carbonyl via absorbance. This amount can be
compared to the theoretical amount of complex to determine the
amount consumed by excess carbonyl. Additionally, some
amount of the carbonyl added is not detected spectroscopically
as either CCOORD or unbound (COBS). Specifically, we can track
the as yet undetected carbonyl in the mass balance in eq. 2:
Figure 8. Solid state structure of 24. Thermal ellipsoids at 50%
-
probability; hydrogen atoms and counter anions (FeCl4 ) omitted
for clarity. Color key: orange = Fe, red = O, gray = C, green = Cl.11
Further, detailed 71Ga NMR experiments by Novikov, Tomilov,
and coworkers show the addition of three chelating ligands to a
Ga(III) center, resulting in the formation of ion pairs.18 If our
GaCl3 systems behave analogously, we should observe the ef-
fects of ion pair formation.
CADD = COBS + CCOORD + CND
(2)
where CND is the moles of carbonyl not detected. Using the re-
gion >1 equiv carbonyl added, we can plot CND as a function of
the moles of complex that have been consumed (CMAX
Colligative Measurements: To probe the interactions of FeCl3
and GaCl3 in DCE with superstoichiometric amounts of 3, 13,
and 14, we examined the conductivity (k) with increasing
amounts of carbonyl (Figure 7). These measurements were ac-
complished via titration using identical concentrations to those
employed for our IR investigation. Beginning with GaCl3 and 3
(■, Figure 7A), we see a slight increase in k from 0-5 equiv 3 to
a value of 121 µS cm-1. For FeCl3 and 3 (●), we see analogous
behavior from 0-1 equiv added 3. At the equivalence point, we
see a conductivity of 96 µS cm-1 for the FeCl3 system. At 2 equiv
3, k increases to 733 µS cm-1, which continues up to 1244 µS
cm-1 at 5 equiv 3. We obtained analogous results for titrations
of 13 and 14. Titration of GaCl3 with 13 displays little increase
in k over the course of the titration, achieving a value of 59 µS
cm-1 at 5 equiv 13. The addition of 13 to FeCl3 displays negli-
gible conductivity from 0-1 equiv 13, while rapidly increasing
to 1247 µS cm-1 at 2 equiv 13. Lastly, examination of the com-
bination of both metal halides with 14 yields results proximal
to the titrations with 3. The GaCl3 titration displays a marginal
increase in conductivity; whereas, the addition of 14 to FeCl3
achieves a k of 796 µS cm-1 at 2 equiv 14, which increases to
1223 µS cm-1 at 5 equiv added 14.
–
CCOORD). Indeed, when these values are compared, we observe
a linear relationship between the moles of undetected carbonyl
and the moles of coordination complex 7 that have been con-
sumed (Figure 5). We observe correlation between the amount
of carbonyl for which we cannot account and the amount of
complex that is consumed with similar results for 13 and 14.11
Importantly, both axes have units in moles, suggesting that the
slope of the line is related to the moles of carbonyl necessary to
consume one mole of 1:1 Lewis acid-carbonyl coordination
complex. The relationship between 3 and 7 suggests that 3-4
equiv 3 are required to consume 1 equiv 7 (Figure 5). Similar
analysis of 13 suggests that ≈3 equiv 13 are required to consume
1 equiv 20, and the relationship between 14 and 21 suggests that
3-4 equiv 14 are required to consume 1 equiv 21.11
If we consider the structural ramifications of the correlations
in Figure 5, they collectively provide a great deal of insight into
the solution behavior of FeCl3 in the presence of carbonyls.
When FeCl3 is exposed to a stoichiometric amount of carbonyl
compound, the classical solution structures for Lewis-acid-me-
diated systems form, comprised of one molecule of FeCl3 and
one molecule of carbonyl compound (22, Figure 6). However,
in the presence of a superstoichiometric amount of carbonyl,
complexes form via the addition of further equivalents of car-
bonyl compounds. As a result, some population of carbonyl-li-
gated complexes will exist in solution with different degrees of
coordination by the carbonyl (23). When x is consistent with
the slopes observed in Figure 5 (slope ≥ 3), this process will
form structures where some number (y) of the chloride ligands
are displaced to the outer sphere. If this process is occurring,
one or more colligative properties of the solution will change.
Detailed examination of these results yields several key in-
sights. We continue to observe analogous behavior between the
GaCl3 systems and the FeCl3 systems for <1 equiv added car-
bonyl. Similar to IR, the behavior of both systems diverge upon
the addition of ≥1 equiv carbonyl, with GaCl3 displaying mar-
ginal if not negligible changes in properties, while FeCl3 dis-
plays properties consistent with the formation of ion pairs (x≥3
and y≥1, Figure 6). Importantly, the observed behavior is incon-
sistent with the complexes reported by Novikov, Tomilov, and
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 11
1
2
3
4
5
6
7
8
0.4
0.3
0.2
0.1
0.0
0.4
0.0005
A
B
C
FeCl3
26
25
O
O
0.0004
0.0003
0.0002
0.0001
0.0000
0.3
0.2
0.1
0.0
Me
Me
Me
Me
Ph
Ph
Me
Me
Me
FeCl3
Me
25
26
1800
1700
1600
1500
1800
1700
1600
1500
0
1
2
3
Wavenumber (cm-1
)
Wavenumber (cm-1
)
25 (equiv)
9
0.4
0.3
0.2
0.1
0.0
0.0005
0.0004
0.0003
0.0002
0.0001
0.0000
0.4
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
D
E
F
GaCl3
27
25
O
O
0.3
0.2
0.1
0.0
Me
Me
Me
Ph
Ph
Me
Me
Me
Me
GaCl3
Me
25
27
1800
1700
1600
1500
1800
1700
1600
1500
0
1
2
3
Wavenumber (cm-1
)
25 (equiv)
Wavenumber (cm-1
)
Figure 9. Solution IR data for titrations of FeCl3 (0.5 mmol in 3 mL DCE) and GaCl3 (0.5 mmol in 3 mL DCE) with 0-1 equiv 25 (A and
D), as well as >1 equiv 25 (B and E) Titrations proceed from red to violet with increasing amounts of titrant. Analysis of components: 25
and 26 (C); 25 and 27 (F). (A: [25] = 0 M, 0.015 M, 0.030 M, 0.045 M, 0.074 M, 0.089 M, 0.132 M. B: [25] = 0.146 M, 0.202 M, 0.310 M,
0.361 M, 0.437 M, 0.580 M. D: [25] = 0 M, 0.030 M, 0.060 M, 0.089 M, 0.118 M, 0.146 M. E: [25] = 0.202 M, 0.243 M, 0.283 M, 0.323
M, 0.361 M, 0.486 M, 0.510 M.).11
coworkers, suggesting that displacement of the counterions
from Ga(III) may require chelating carbonyl ligands.18
ketone 15 displays an order of 1.13 ± 0.03,11 which is consistent
with other reports on reactions of aromatic ketones.4e,8
Crystallographic Investigation: To gain further support for the
solution structures that are possible in the reaction in Figure 2A,
we turned to X-ray crystallography.19 Indeed, when FeCl3 is
combined in DCE in the presence of excess 13 with pentane-
assisted precipitation, we are able to resolve a crystal structure
for FeCl3 (24, Figure 8).11 We observe an octahedral, Fe-cen-
tered complex, showing that 13 displaces one chloride anion in
24. Importantly, the observed number of molecules of 13 bound
to the Fe(III) center is consistent with our analysis in Figure 5,
where three additional equivalents of 13 add to complex 20. The
presence of four molecules of 13 as well as two chloride anions
is consistent with our conductance data, where solution conduc-
tivity increases because of displacement of chloride to the outer
sphere. This higher order of coordination is consistent with
Denmark and coworkers’ report on interactions of SnCl4 with
aldehydes.20 It is important to point out that this structure repre-
sents a single possible coordination complex in this system, and
that it is simply the structure that precipitates under forcing con-
ditions. Further, we do not observe precipitation under reaction
or titration conditions, and our titration data are consistent with
complete formation of 1:1 coordination complex followed by
subsequent addition to form a more highly ligated species.
Substrate Analysis: Having characterized the interactions of
typical carbonyl-olefin metathesis byproducts in detail, we next
examined the interactions of substrate and metal halide. Be-
cause of the rapidity of reaction onset when either FeCl3 or
GaCl3 are combined with 15, we examined the Lewis acid-car-
bonyl interactions of 25: the structural analogue of the substrate,
but with the olefin partner hydrogenated.16 We applied our ti-
tration protocol to this model and found little difference in be-
havior between the Fe(III) and Ga(III) systems (Figure 9). Be-
tween 0 and 1 equiv 25 added to the solution, we observe no
unbound 25 in either system, consistent with our observations
for simple carbonyls. In the carbonyl region of the Fe(III) spec-
trum, we observe the formation of signals at 1551, 1584, and
1596 cm-1 (Figure 9A). Again, the FeCl3 system remains heter-
ogeneous until 1 equiv 25 is present. Equivalent spectroscopic
features are present with GaCl3, with formation of vibrations at
1558 and 1584 cm-1 (Figure 9D). Interestingly, when the addi-
tion of 25 proceeds beyond 1 equiv, we no longer observe the
divergent behavior seen with simple carbonyls (Figures 9B and
9E). In both systems, the initial vibrations remain unchanged at
higher equivalents of 25, and we observe the carbonyl of 25 at
1670 cm-1. These spectral features can be seen in greater clarity
via analysis of the system components (Figures 9C and 9F). In
both systems, we observe growth of the coordination complex
to a maximum at approximately 1 equiv 25, at which point 25
is observed. The appearance of 25 at lower equivalents in Fig-
ure 9C suggests a different binding affinity between the sub-
strate and FeCl3 and the substrate and GaCl3. Importantly,
highly ligated complexes are not observed.
-
Taken together, these observations suggest that the FeCl4 ob-
served in the crystal structure is an artifact of the equilibria in-
volved in precipitation. In order to form FeCl4 in our titrations,
2 equiv byproduct would be required to consume 2 equiv 1:1
complex (1 mol 13 per 1 mol 20). Our analysis in Figure 5 is
consistent with 3 equiv of byproduct consuming 1 equiv 1:1
-
-
complex (3 mol 13 per 1 mol 20). Lastly, formation of FeCl4
could be consistent with the superelectrophilic homo dimers re-
ported by Schindler, Sigman, and Zimmerman.4e In their report,
they observe second order kinetics with respect to FeCl3 when
employing aliphatic ketones as substrates; whereas, aromatic
ketones display first order kinetics with respect to FeCl3, which
is inconsistent with the superelectrophile-mediated process. Ex-
amination of the rate order of FeCl3 for the reaction of aromatic
Carbonyl Exchange: Lastly, we sought to examine the ability
of the substrate to displace byproduct in order to access the Fe-
center in the carbonyl exchange step. To this end, we performed
a titration, in which we preformed 1:1 complex 7, and then
added increasing amounts of 25. When 25 is added to 7 from 0-
1 equiv, 7 is consumed, free 3 appears, free 25 can be observed,
as well as 1:1 complex 26 (Figure 10A). When 25 is added from
ACS Paragon Plus Environment

Page 7 of 11
Journal of the American Chemical Society
0.3
0.2
0.1
0.0
0.3
A
C
1
2
3
4
5
6
7
8
25
0.2
7
25
28
26
3
3
0.1
Me
Cl-
Me
Me
Me
0.0
1800
O
Me
1800
1700
1600
1500
1700
1600
1500
Me
O
O
O
Wavenumber (cm-1
)
Wavenumber (cm-1
)
Fe
Me
Cl
Me
9
25
0.3
0.2
0.1
0.0
0.3
0.2
0.1
0.0
Cl
25
B
10
11
12
13
14
15
16
17
18
D
28
3
3
28
7
26
1800
1700
1600
1500
1800
1700
1600
1500
Wavenumber (cm-1
)
Wavenumber (cm-1
)
Figure 10. Solution IR data for titrations of 1:1 acetone-FeCl3 (0.5 mmol in 3 mL DCE) with 0-1 equiv 25 (A) as well as >1 equiv 25 (B).
Solution IR data for titrations of 2:1 acetone-FeCl3 (0.5 mmol in 3 mL DCE) with 0-1 equiv 25 (C), as well as >1 equiv 25 (D). Titrations
proceed from red to violet with increasing amounts of titrant. (A: [25] = 0 M, 0.045 M, 0.074 M, 0.103 M, 0.123 M, 0.174 M. B: [25] =
0.202 M, 0.257 M, 0.323 M, 0.374 M, 0.474 M, 0.591 M. C: [25] = 0 M, 0.045 M, 0.074 M, 0.103 M, 0.131 M, 0.174 M. D: [25] = 0.201
M, 0.256 M, 0.321 M, 0.373 M, 0.471 M, 0.589 M.).11
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
O
O
O
Me
Me
Me
Me Me
Me Me
Cl-
Ph
3
Me
Me
3
FeCl3
1
O
R
Me
O
Me
O
O
Me
Me
Fe
Ph
Me
Cl
O
R
Cl
O
4
Me
+
Ph
Me
Cl-
Me Me
FeCl3
Me
R
Me
Me
Me
3
O
29
Me
O
primary
cycle
secondary
cycle
Me
O
O
O
O
-
Fe
Me Me
Me
Me
Me Me
Cl
Cl-
Me
Me
7
3
Me
O
Cl
Me
reversible aggregation
28
Me
Me
R
Me
O
O
O
Cl3Fe
Fe
O
Me
Cl
Me
Ph
Me
Ph
Cl
Ph
30
R
6
R
2
Figure 11. Final mechanistic proposal.
1-3 equiv, more 26 is formed, more 7 is consumed, and more
free 3 and free 25 are observed (Figure 10B). These data suggest
that the substrate is capable of displacing the byproduct present
in 1:1 byproduct complex 7 to form 1:1 substrate complex 26.
Alternatively, when the reaction is initiated with 0.5 equiv 3
with respect to substrate, the reaction begins at 10D, with far
less substrate able to bind exclusively to FeCl3. This shift in Fe
structure to the highly ligated system coincides with decrease in
rate and even termination of reactivity.
We observe disparate behavior when the titration begins with
highly ligated byproduct complex 28. When 25 is added to this
complex from 0-1 equiv, the most significant change in the
spectrum is an increase in the signal for unbound 25 (Figure
10C). There are minimal changes in the intensity of 28, trace
amounts of 26 are observed, and the signal for free 3 remains
unchanged. When 25 is added from 1-3 equiv, again, the most
significant change is the addition of unbound 25. A small in-
crease in free 3 is present, and a few vibrations appear in the
region where 1:1 complex 26 appears. However, the lmax in this
region is not consistent with the signal observed in Figure 9A,
and may represent a different structure. Lastly, we can explain
the behavior of the FeCl3 catalyst as the reaction progresses
with these data. At low conversion, Figure 10B represents the
system at high concentration of substrate and low concentration
of byproduct. Whereas, Figure 10C represents catalyst condi-
tions from 50% conversion to termination, resulting in 28.
Final Mechanistic Proposal: Our kinetic, spectroscopic, con-
ductance, and crystallographic investigations yield the follow-
ing results: 1) The rate of catalytic turnover is decreased by the
byproduct when FeCl3 is the catalyst. 2) This interaction in the
GaCl3 system is not nearly as pronounced. 3) Titration of both
GaCl3 and FeCl3 in DCE with 0-1 equiv 3, 13, and 14 all result
in exclusive formation of a 1:1 Lewis acid-carbonyl coordina-
tion complex, consistent with previous reports. 4) When >1
equiv 3, 13, or 14 are added to GaCl3, no effect is detected spec-
troscopically or via conductance. 5) When >1 equiv 3, 13, or 14
are added to FeCl3, the initial complex is consumed and is con-
verted to an alternative species detected by IR, consistent with
the addition of ≥3 additional equivalents of carbonyl com-
pound. 6) When FeCl3 in DCE is exposed to >1 equiv of 3, 13,
or 14, a marked increase in conductivity is observed, consistent
with the formation of solvent-separated ion pairs. 7) X-ray
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 11
A
to 29. In the former case, the decrease in [4] will lead to a lower
turnover frequency. In the case of the secondary cycle, the de-
O
FTs
N
MCl3 (50 mol%)
1
2
3
4
5
6
7
8
O
Ph
N
FTs
Ph
+
Ph
creased rate may result from an increase in the degree of coor-
dination to the Lewis acid, which is consistent with the report
of Denmark and coworkers who showed that the structure of the
Lewis acid-carbonyl complex can affect the reactivity of the
carbonyl.20 By attaching the carbonyl to sterically encumbered
catalyst 28, the turnover-limiting [2+2]-cycloaddition becomes
more difficult by inhibiting association of the pendant olefin
with the carbonyl. This steric inhibition is analogous to the in-
hibition observed in SmI2-mediated ketyl-olefin 5-exo-trig cy-
clizations reported by Flowers and coworkers.21 Lastly, this re-
sult is consistent with our previous observations that the pres-
ence of exogenous Lewis bases inhibits product formation.8
DCE (0.010 M), 0 ºC
Ph
H
Ph
Ph
32
13
31, (1 equiv)
M = Fe or Ga
10
B
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
5
To test the explanatory power of our model for Lewis acid
behavior, we examined a system reported by Schindler and
coworkers that produces dihydropyrrole 33 and 13 (Figure
12A).4c This system is analogous to the work of Li and cowork-
ers4a with two key modifications appropriate for the testing of
our supposition: 1) the system does not require allyltrime-
thylsilane to eliminate the byproduct, and 2) the application of
the FTs group decreases the likelihood of interaction of this pro-
tecting group with the Lewis acid mediator. As a result, these
conditions allow us to see the competition between the substrate
carbonyl and the byproduct carbonyl.
0
0
1000
2000
3000
4000
Time (s)
Figure 12. Metal halide-mediated carbonyl-olefin metathesis of 31
(A). GaCl3-mediated system (●), FeCl3-mediated system (■), 1:1
13:FeCl3-mediated (□), 1:1 13:GaCl3-mediated system (○) (B).11
We examined the FeCl3-catalyzed system (■, Figure 12B), as
well as a GaCl3-catalyzed system (●, Figure 12B). Intriguingly,
GaCl3 displays an initial rate faster than that of the FeCl3-medi-
ated reaction, while the FeCl3 system maintains a higher rate at
high conversion of substrate. Next, we initiated the reaction in
the presence of 1 equiv 13 with respect to the metal halide. In
the presence of added byproduct 13, a significant decrease in
rate occurs for the Fe(III)-mediated system, while no reaction is
observed with GaCl3 even after 4 h. This system requires an or-
der of magnitude more Lewis acid than that of Figure 2, sug-
gesting that substrate 31 has a much lower affinity for each
Lewis acid than substrate 15. This lower relative affinity makes
byproduct inhibition significantly more pronounced. Further,
this inhibition is so pronounced in the Ga-mediated system that
reactivity is prevented, while FeCl3 is still capable of carrying
out the reaction in the presence of 1 equiv 13. The GaCl3 result
is consistent with direct competition between the substrate and
byproduct; whereas, the FeCl3 reaction is still able to proceed
because multiple carbonyls are capable of accessing the metal
center.
crystallographic data demonstrate that an octahedral Fe-cen-
tered complex can form in solution and displace the original
chloride ligands. 8) The highly ligated structures are not ob-
served for titrations of a substrate model, displaying only 1:1
complex. These observations allow us to address the question
we raised via our kinetic experiments: Why are different behav-
iors observed for each Lewis acid in the presence of byproduct?
At all equivalents of carbonyl examined with respect to
GaCl3, we observe no additional interaction beyond the for-
mation of a 1:1 Lewis acid-carbonyl complex, consistent with
the classical mechanisms drawn for Lewis-acid-mediated reac-
tions. The classical Lewis-acid-mediated model is true for stoi-
chiometric interactions between FeCl3 and the carbonyls exam-
ined (0-1 equiv). However, this model stops being representa-
tive of the solution structure of the Lewis acid at superstoichio-
metric loadings of byproduct carbonyl (>1 equiv). A highly li-
gated species results when large excesses of byproduct carbonyl
are present with respect to FeCl3. In our study, we examined
loadings of carbonyl up to 5 equiv, which would be consistent
with a 20 mol% loading of metal halide with respect to car-
bonyl. With respect to the metathesis reaction, 20 equiv of by-
product are present at the end of a reaction employing 5 mol%
catalyst. In the Ga-catalyzed system, 1:1 interactions are all that
occur under reaction conditions, suggesting Ga(III) follows the
primary cycle of the metathesis reaction (Figure 11), and that
byproduct inhibition will arise when the byproduct has a bind-
ing affinity for the Lewis acid capable of outcompeting the sub-
strate. Alternatively, the behavior of FeCl3 changes over the
course of the reaction. At low turnovers, the primary cycle effi-
ciently converts substrate 1 to product 2. However, as the con-
centration of byproduct increases, the Fe center of 1:1 complex
7 is ligated by additional byproduct carbonyls, forming highly
ligated complexes like 28, outside of the primary cycle. Car-
bonyl exchange data suggest that at high substrate concentra-
tions relative to 3, byproduct can be displaced from the complex
(Figure 10D). As a result, either substrate can still form 4 by
displacing multiple byproduct ligands, or 28 can be converted
Lastly, this mechanistic proposal facilitates the adaptation of
new procedures by synthetic chemists attempting to employ the
metathesis reaction to recalcitrant substrates. The benzalde-
hyde-producing transformation in Figure 12A is facilitated by
an increase in the loading of Lewis acid.4c By increasing the ra-
tio of FeCl3 to byproduct, the reversible aggregation process fa-
vors the 1:1 Lewis acid-carbonyl complex. As a result, the me-
tathesis reaction can remain in the primary cycle until higher
conversions of starting material. Alternatively, reports by the
Li4a and Schindler4e labs have addressed benzaldehyde-medi-
ated inhibition by the addition of allyltrimethylsilane to chemi-
cally remove benzaldehyde from the system. To address ace-
tone-mediated inhibition, a similar increase in the loading of
Lewis acid can shift the equilibrium to favor the primary cycle.
Alternatively, we have found that increasing reaction tempera-
ture can improve turnover of the catalyst.4e For example, the re-
action depicted in Figure 2A is performed at 0 ºC and reaches
completion in about 20 min. When the same reaction is
ACS Paragon Plus Environment

Page 9 of 11
Journal of the American Chemical Society
performed in the presence of 0.5 equiv 3 with respect to sub-
1
2
3
4
Notes
strate at 0 ºC, the reaction terminates early. When the 3-contain-
ing system is heated to 40 ºC, the reaction reaches full conver-
sion in 20 min.
The authors declare no competing financial interest.
ACKNOWLEDGMENT
Conclusion
We thank Loyola University Chicago, Merck & Co., Inc., and the
NIH/National Institute of General Medical Sciences (GM128126)
for financial support. This work made use of the Jerome B. Cohen
X-Ray Diffraction Facility supported by the MRSEC program of
the National Science Foundation (DMR-1720139) at the Materials
Research Center of Northwestern University and the Soft and Hy-
brid Nanotechnology Experimental (SHyNE) Resource (NSF
ECCS-1542205). We are very grateful to Prof. Wei-Tsung Lee, Ms.
Adriana Lugosan, Mrs. Loretta Devery, and Mr. James Devery for
helpful discussions and suggestions. We also thank the reviewers
for their insightful comments on an earlier version of this paper.
5
6
7
8
The solution structures of metathesis-active catalysts were in-
vestigated on the basis of kinetic, spectroscopic, colligative, and
crystallographic experiments. These data have given us insight
into the divergent kinetic behavior of GaCl3 and FeCl3 as cata-
lysts in DCE. GaCl3 interacts with carbonyls through a classical
Lewis acid-Lewis base interaction, forming a 1:1 coordination
complex, regardless of relative amount of carbonyl. Con-
versely, FeCl3 does not only exist as a 1:1 coordination complex
when employed as a catalyst, but rather can be reversibly ligated
by multiple molecules of byproduct, while potentially remain-
ing catalytically active when substrate-binding affinity is high.
The presence of alternative Lewis bases in addition to the sub-
strate carbonyl inhibits the turnover-limiting [2+2]-cycloaddi-
tion that yields product. This work describing the solution struc-
tures for catalysis of aromatic carbonyls, in concert with the re-
cent report of Fe(III) dimers by Schindler, Sigman, and Zim-
merman for the reaction of aliphatic carbonyls,4e indicates that
significant consideration of solution structures allows for a
more complete understanding of reaction behavior in catalytic
systems. Indeed, the highly ligated Fe(III) complexes we de-
scribe likely have a more pronounced inhibitory effect on the
formation of the superelectrophilic iron dimers, evidenced by
the need for elevated temperatures in some cases or the addition
of allyltrimethylsilane in benzaldeyde producing substrates.
When alternatives to chlorinated solvents are considered, fur-
ther complications arise, ranging from inhibition of reactivity
with Lewis basic solvents8 to the trapping of metathesis inter-
mediates in more lipophilic solvents.4b,11 We are currently in-
vestigating the complexities of solvent interactions to map the
diverse array of solution structures. These considerations are
not only important for reaction design and catalyst selection, but
also for computational analysis of reaction intermediates and
transition states. Further, the development of ring-opening and
cross carbonyl-olefin metathesis reactions requires the ability of
the catalyst to adequately differentiate between substrate and
product carbonyls. We have begun studies of other metathesis-
active Lewis acids to ascertain the full scope of this effect. Fur-
ther, the described byproduct inhibition is likely a factor in
many carbonyl-based FeCl3-catalyzed reactions beyond car-
bonyl-olefin metathesis. We are currently examining alternative
systems to understand the impact of the solution structures ac-
cessible to FeCl3 in catalytic regimes.
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
REFERENCES
(1) a) NATO ASI Series, Ser. C: Mathematical and Physical Sci-
ences: Selectivities in Lewis Acid Promoted Reactions., Kluwer, 1989,
vol. 289. b) Santelli, M.; Pons, J.-M. Lewis Acids and Selectivity in Or-
ganic Synthesis, CRC Press, Boca Raton, 1996. c) Yamamoto, H. Ed.,
Lewis Acids in Organic Synthesis, Wiley-VCH, Weinheimꢀ; New York,
2000. d) Satchell, D. P. N.; Satchell, R. S. Quantitative Aspects of the
Lewis Acidity of Covalent Metal Halides and Their Organo Deriva-
tives. Chem. Rev., 1969, 69, 251–278.
(2) a) Ludwig, J. R.; Schindler, C. S. Lewis Acid Catalyzed Car-
bonyl–Olefin Metathesis. Synlett, 2017, 28, 1501–1509. b) Becker, M.
R.; Watson, R. B.; Schindler, C. S. Beyond Olefins: New Metathesis
Directions for Synthesis. Chem. Soc. Rev., 2018, 47, 7867-7881.
(3) Ludwig, J. R.; Zimmerman, P. M.; Gianino J. B.; Schindler, C.
S. Iron(III)-Catalysed Carbonyl–Olefin Metathesis. Nature, 2016, 533,
374–379.
(4) a) Ma, L.; Li, W.; Xi, H.; Bai, X.; Ma, E.; Yan, X.; Li, Z. FeCl3-
Catalyzed Ring-Closing Carbonyl-Olefin Metathesis. Angew. Chem.
Int. Ed., 2016, 55, 10410–10413. b) McAtee, C. C.; Riehl P. S.;
Schindler, C. S. Polycyclic Aromatic Hydrocarbons via Iron(III)-
Catalyzed Carbonyl–Olefin Metathesis. J. Am. Chem. Soc., 2017, 139,
2960–2963. c) Groso, E. J.; Golonka, A. N.; Harding, R. A.; Alexander,
B. W.; Sodano T. M.; Schindler, C. S. 3-Aryl-2,5-Dihydropyrroles via
Catalytic Carbonyl-Olefin Metathesis. ACS Catal., 2018, 8, 2006–
2011. d) Ludwig, J. R.; Watson, R. B.; Nasrallah, D. J.; Gianino, J. B.;
Zimmerman, P. M.; Wiscons, R. A.; Schindler, C. S. Interrupted Car-
bonyl-Olefin Metathesis via Oxygen Atom Transfer. Science, 2018,
361, 1363-1369. e) Albright, H.; Riehl, P. S.; McAtee, C. C.; Reid, J.
P; Ludwig, J. R.; Karp, L. A.; Zimmerman, P. M.; Sigman, M. S.;
Schindler, C. S. Catalytic Carbonyl-Olefin Metathesis of Aliphatic Ke-
tones: Iron(III) Homo-Dimers as Lewis Acidic Superelectrophiles. J.
Am. Chem. Soc. 2019, 141, 1690–1700.
(5) Albright, H.; Vonesh, H. L.; Becker, M. R.; Alexander, B. W.;
Ludwig, J. R.; Wiscons, R. A.; Schindler, C. S. GaCl3-Catalyzed Ring-
Opening Carbonyl–Olefin Metathesis. Org. Lett., 2018, 20, 4954–
4958.
(6) Tran, U. P. N.; Oss, G.; Breugst, M.; Detmar, E.; Pace, D. P.;
Liyanto, K.; Nguyen, T. V. Carbonyl-Olefin Metathesis Catalyzed by
Molecular Iodine. ACS Catal. 2019, 9, 912–919.
(7) a) Catti, L.; Tiefenbacher, K. Brønsted Acid-Catalyzed Car-
bonyl-Olefin Metathesis inside a Self-Assembled Supramolecular
Host. Angew. Chem. Int. Ed., 2018, 57, 14589-14592. b) Tran, U. P.
N.; Oss, G.; Pace, D. P.; Ho J.; Nguyen, T. V. Tropylium-Promoted
Carbonyl-Olefin Metathesis Reactions. Chem. Sci., 2018, 9, 5145–
5151.
(8) Ludwig, J. R.; Phan, S.; McAtee, C. C.; Zimmerman, P. M.;
Devery, J. J., III; Schindler, C. S. Mechanistic Investigations of the
Iron(III)-Catalyzed Carbonyl-Olefin Metathesis Reaction. J. Am.
Chem. Soc., 2017, 139, 10832–10842.
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website.
Experimental data (PDF)
Data for C28H24O4Fe2Cl6 (CIF)
AUTHOR INFORMATION
Corresponding Author
* jdevery@luc.edu
(9) Hosomi, A.; Sakurai, H. Syntheses of γ,δ-Unsaturated Alcohols
from Allylsilanes and Carbonyl Compounds in the Presence of Tita-
nium Tetrachloride. Tetrahedron Lett., 1976, 17, 1295–1298.
ORCID
Carly S. Hanson: 0000-0002-4288-1219
James J. Devery, III: 0000-0003-3124-465X
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 11
(10) Becker, M. R.; Rykaczewski, K. A.; Ludwig, J. R.; Schindler,
BF3, FeCl3, ZnCl2 et AlCl3 et Nature de la Liaison Oxygène-métal.
C. S. Carbonyl-Olefin Metathesis for the Synthesis of Cyclic Olefins.
Org. Synth., 2018, 95, 472–485.
(11) See Supporting Information for more details.
Helv. Chim. Acta, 1958, 41, 1332–1341.
1
2
3
4
5
6
7
8
(16) a) Greenwood, N. N. The Ionic Properties and Thermochemis-
try of Addition Compounds of Gallium Trichloride and Tribromide. J.
Inorg. Nucl. Chem., 1958, 8, 234–240. b) Greenwood, N. N.; Perkins,
P. G. 68. Thermochemistry of the Systems Which Boron Trichloride
and Gallium Trichloride Form with Acetone and Acetyl Chloride, and
the Heat of Solution of Boron Tribromide in Acetone. J. Chem. Soc.,
1960, 356.
(17) Davlieva, M. G.; Lindeman, S. V.; Neretin, I. S.; Kochi, J. K.
Isolation, X-ray Structures, and Electronic Spectra of Reactive Inter-
mediates in Friedel−Crafts Acylations. J. Org. Chem., 2005, 70, 4013–
4021.
(18) a) Novikov, R. A.; Potapov, K. V.; Chistikov, D. N.; Tarasova,
A. V.; Grigoriev, M. S.; Timofeev, V. P.; Tomilov, Y. V. Synthesis and
Structures of Cyclopropanedicarboxylate Gallium Complexes. Organ-
ometallics, 2015, 34, 4238–4250. b) Novikov, R. A.; Denisov, D. A.;
Potapov, K. V.; Tkachev, Y. V.; Shulishov, E. V.; Tomilov, Y. V. Ionic
Ga-Complexes of Alkylidene- and Arylmethylidenemalonates and
Their Reactions with Acetylenes: An In-Depth Look into the Mecha-
nism of the Occurring Gallium Chemistry. J. Am. Chem. Soc., 2018,
140, 14381-14390.
(19) For a review of solid state structures of Lewis acids and carbon-
yls: Shambayati, S.; Crowe, W. E.; Schreiber, S. L. On the Confor-
mation and Structure of Organometal Complexes in the Solid State:
Two Studies Relevant to Chemical Synthesis. Angew. Chem. Int. Ed.,
1990, 29, 256–272.
(12) a) Mohammad, A.; Satchell, D. P. N.; Satchell, R. S. Quantita-
tive Aspects of Lewis Acidity. Part VIII. The Validity of Infrared Car-
bonyl Shifts as Measures of Lewis Acid Strength. The Interaction of
Lewis Acids and Phenalen-1-one(perinaphthenone). J. Chem. Soc. B
Phys. Org., 1967, 723–725. b) Satchell, D. P. N.; Satchell, R. S. Quan-
titative Aspects of Lewis Acidity. Q. Rev. Chem. Soc., 1971, 25, 171–
199. c) Driessen, W. L.; Groeneveld, W. L. Complexes with Ligands
Containing the Carbonyl Group. Part I: Complexes with Acetone of
Some Divalent Metals Containing Tetrachloro-ferrate(III) and -In-
date(III) Anions. Recl. Trav. Chim. Pays-Bas, 1969, 88, 977–988. d)
Driessen, W. L.; Groeneveld, W. L. Complexes with Ligands Contain-
ing the Carbonyl Group. Part III: Metal (II) Acetaldehyde, Propionalde-
hyde and Benzaldehyde Solvates. Recl. Trav. Chim. Pays-Bas, 1971,
90, 87–96. e) Driessen, W. L.; Groeneveld, W. L. Complexes with Lig-
ands Containing the Carbonyl Group. Part IV Metal(II) Butanone, Ace-
tophenone, and Chloroacetone Solvates. Recl. Trav. Chim. Pays-Bas,
1971, 90, 258–264.
(13) Susz, B. P.; Cooke, I. Etude du Spectre Infrarouge de Com-
plexes Formés par les Halogénures d'Aluminium. I. Complexes de
l'Acétophénone et de la Benzophénone; Comparaison avec les Spectres
de Divers Compossé Organiques Présentant la Liaison Aluminium-ox-
ygène. Helv. Chim. Acta, 1954, 37, 1273–1280.
(14) a) Susz, B. P.; Lachavanne, A. Etude de Composés d'Addition
des Acides de Lewis. V. Absorption Infrarouge de Composés d'Addi-
tion Formés par le Chlorure de Titane avec la Benzophénone et
l'Acétophénone. Helv. Chim. Acta, 1958, 41, 634–636. b) Cassimatis,
D.; Susz, B. P. Etude des Composés d'Addition des Acides de LEWIS.
XI. Spectre d'Absorption Infrarouge du Composé d'Addition
CH3COCH3, TiCl4. Helv. Chim. Acta, 1960, 43, 852–862. c) Weber,
R.; Susz, B. P. Etude des Composés d'Addition des Acides de Lewis
XXX. Note Préliminaire Sur les Composés d'Addition Entre Aldéhydes
et TiCl4. Helv. Chim. Acta, 1967, 50, 2226–2232.
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
(20) a) Denmark, S. E.; Henke, B. R.; Weber, E. SnCl4(4-tert-
BuC6H4CHO)2. X-ray Crystal Structure, Solution NMR, and Implica-
tions for Reactions at Complexed Carbonyls. J. Am. Chem. Soc., 1987,
109, 2512–2514. b) Denmark, S. E.; Almstead, N. G. Spectroscopic
Studies on the Structure and Conformation of Lewis Acid-Aldehyde
Complexes. J. Am. Chem. Soc., 1993, 115, 3133–3139.
(21) Sadasivam, D. V.; Antharjanam, P. K. S.; Prasad E.; Flowers,
R. A., II Mechanistic Study of Samarium Diiodide-HMPA Initiated 5-
exo-trig Ketyl−Olefin Coupling: The Role of HMPA in Post-Electron
Transfer Steps. J. Am. Chem. Soc., 2008, 130, 7228–7229.
(15) a) Chalandon, P.; Susz, B. P. Etude de Composés d'Addition
des Acides de Lewis. VI Spectre d'Absorption Infrarouge de l'Acétone-
trifluorure de Bore; Apectre d'Absorption Infrarouge et Moment de
Dipǒle du Di-propyl-cétone-trifluorure de Bore. Helv. Chim. Acta,
1958, 41, 697–704. b) Susz B. P.; Chalandon, P. Etude de Composés
d'Addition des Acides de Lewis. IX. - Spectres d'Absorption Infrarouge
des Composés Formés par la Benzophénone et l'Acétophénone avec
ACS Paragon Plus Environment

Page 11 of 11
Journal of the American Chemical Society
O
Me Me
O
1
2
3
FeCl3
Me
O
Ph
Me
Me
R
Me
Ph
R
Me
O
Cl-
Me
Me
Me
+
4
O
Me Me
Me
FeCl3
Me
O
O
O
primary
cycle
5
6
O
Fe
O
Me
Cl
Me
-
Me Me
Cl
Me Me
7
8
9
byproduct inhibition
Me
Me
R
Cl3Fe
O
Ph
Ph
R
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
11