Investigations into the Mechanism of Inter- And Intramolecular Iron-Catalyzed [2 + 2] Cycloaddition of Alkenes

Article abstract of DOI:10.1021/jacs.0c00250

Mechanistic studies are reported on the inter- and intramolecular [2 + 2] alkene cycloadditions to form cyclobutanes promoted by (^tricPDI)Fe(N₂) (^tricPDI = 2,6-(2,4,6-tricyclopentyl)C₆H₂N = CMe)₂

Full text of DOI:10.1021/jacs.0c00250

Subscriber access provided by La Trobe University Library
Article
Investigations into the Mechanism of Inter- and
Intramolecular Iron-Catalyzed [2+2] Cycloaddition of Alkenes
Matthew V. Joannou, Jordan M Hoyt, and Paul J. Chirik
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.0c00250 • Publication Date (Web): 20 Feb 2020
Downloaded from pubs.acs.org on February 22, 2020
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 18
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Investigations into the Mechanism of Inter- and Intramolecu-
lar Iron-Catalyzed [2+2] Cycloaddition of Alkenes
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
Matthew V. Joannou, Jordan M. Hoyt, Paul J. Chirik*
Department of Chemistry, Princeton University, Princeton, New Jersey 08544, United States of America
ABSTRACT: Mechanistic studies are reported on the inter- and intramolecular [2+2] alkene cycloadditions to form cyclobutanes promot-
ed by (tricPDI)Fe(N2) (tricPDI = 2,6-(2,4,6-tricyclopentyl)C6H2N=CMe)2C5H3N). A combination of kinetic measurements, freeze-quench
57Fe Mössbauer and infrared spectroscopic measurements, deuterium labeling studies, natural abundance 13C KIE studies, and isolation and
characterization of catalytically relevant intermediates were used to gain insight into the mechanism of both inter- and intramolecular [2+2]
cycloaddition reactions. For the stereo- and regioselective [2+2] cycloaddition of 1-octene to form trans-1,2-dihexylcyclobutane, a first-order
dependence on both iron complex and alkene was measured as well as an inverse dependence on N2 pressure. Both 57Fe Mössbauer and infra-
red spectroscopic measurements identified (tricPDI)Fe(N2)(η2-1-octene) as the catalyst resting state. Rate-determining association of 1-
octene to (tricPDI)Fe(η2-1-octene) accounts for the first order dependence of alkene and the inverse dependence on N2. Heavy atom 13C/12C
kinetic isotope effects near unity also support post rate-determining C–C bond formation. By contrast, the intramolecular iron-catalyzed
[2+2] cycloaddition of 1,7-octadiene yielded cis-bicyclo[4.2.0]octane in 92:8 d.r. and a first order dependence on the iron precursor and
zeroth order behavior in both diene and N2 pressure were measured. A pyridine(diimine) iron trans-bimetallacycle was identified as the cata-
lyst resting state and was isolated and characterized by X-ray diffraction, 1H NMR and 57Fe Mössbauer spectroscopies. Dissolution of the iron
trans-bimetallacycle in benzene-d6 produced predominantly the cis-cyclobutane product, establishing interconversion between the trans and
cis metallacycles during the catalytic reaction and consistent with a Curtin-Hammett kinetic regime. A primary 13C/12C kinetic isotope effect
of 1.022(4) was measured at 23 ºC, consistent with rate-determining unimolecular reductive elimination to form the cyclobutane product.
Despite complications from competing cyclometallation of chelate aryl substituents, deuterium labeling experiments were consistent with
unimolecular C–C reductive elimination that occurred either by a concerted pathway or a radical rebound sequence that is faster than C–C
bond rotation.
cycloadditions, along with isolated and proposed
metallacyclic intermediates. E = NR’, CR’2; G = alkyl,
benzyl, heterobenzyl.
∎ INTRODUCTION
The synthesis of strained carbocycles is a challenge in organic
chemistry with a host of potential applications in polymers, phar-
1
maceuticals, fine chemicals, and fragrances. In particular, the struc-
tural rigidity and strain in cyclobutanes offer a range of possible
applications and the development and understanding of new cata-
lytic processes for their synthesis is highly desirable to the synthetic
community. Traditional routes to cyclobutanes from olefins em-
ploy radical- and photochemical-based methods designed to over-
come the orbital symmetry constraints of thermally forbidden cy-
2
cloadditions. The substrates required for these reactions, however,
are highly activated (α,β-unsaturated esters, allenes, etc.) and as a
consequence can limit the scope and applicability of the transfor-
mation. The transition metal-catalyzed formal [2+2] cycloaddi-
tions of unactivated feedstock α-olefins have emerged as an attrac-
3
tive, atom-economical route to these valuable 4-membered rings.
A distinct advantage over radical-based processes is the opportuni-
ty for catalyst control, enabling the rational modification of chemo-
and regioselectivity resulting in the precision synthesis of a variety
of cyclic products.
As presented in Scheme 1, our group has developed and studied
several [2+2] cycloadditions of olefins to generate cyclobutanes,
utilizing iron- and cobalt-catalysts bearing redox-active pyri-
4
dine(diimine) (PDI) ligands. For iron, both intra- and intermolec-
ular reactions have been demonstrated in alkene-alkene and diene-
alkene [2+2] cycloaddition reactions. While both combinations of
coupling partners yielded 4-membered ring products, the observed
organometallic intermediates formed during these reactions have
different geometries, spin states and coordination numbers, raising
1
Scheme 1. Previously reported pyridine(diimine)
iron-catalyzed alkene-alkene and alkene-diene [2+2]
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 18
questions about the nature of the reductive elimination and
of product from the starting material, as well as reliable kinetic
C(sp3)–C(sp3) bond-forming step.
measurements by analysis of reaction aliquots. 1-Octene also has an
accessible melting-point (-101 °C) and glass-transition at that melt-
ing point which aids in the preparation of samples for freeze-
1
2
3
4
5
6
7
8
Previous studies on intramolecular alkene-alkene [2+2] cycload-
dition resulted in isolation of catalytically competent S = 1 pyri-
dine(diimine) iron bis(alkene) and metallacycle complexes
57
,
7
6
quench Fe Mössbauer studies. As depicted in Figure 1, with 1
mol
%
(
tricPDI)Fe(N2) in neat 1-octene, trans-1,2-
5
(Scheme 1, right). Structural, spectroscopic, and computational
dihexylcyclobutane was formed in 31% yield in >99% selectivity.
The density functional theory (DFT)-computed thermodynamic
properties of the [2+2] reaction establish a near thermoneutral
reaction under standard conditions (B3LYP level of theory, see
Supporting Information for more computational details). While
highly enthalpically driven (ΔH° = -14.4 kcal·mol-1), the large en-
tropic penalty of the reaction (ΔS° = -46.4 cal·mol-1·K-1) results in
an overall ΔG° of -0.7 kcal·mol-1.
data supported the presence of the mono-reduced form of the che-
late ([PDI]1-) as the iron cycled between the two intermediates,
consistent with an Fe(I)–Fe(III) redox cycle. For butadiene-
ethylene coupling, an S = 0 pyridine(diimine) iron alkyl-allyl metal-
lacycle was isolated and crystallographically characterized where
the chelate is in its closed-shell dianionic form (Scheme 1, right).4c
The isolated iron metallacycle did not undergo C–C reductive
elimination at ambient temperature in solution, however upon
addition of butadiene or carbon monoxide, vinyl cyclobutane was
generated, consistent with “ligand-induced” reductive elimina-
tion.4d
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
Modification of the aryl groups on the pyridine(diimine) iron di-
nitrogen complex enabled the highly chemo-, regio- and diastere-
oselective intermolecular cycloaddition of unactivated alkenes. In
contrast to the intramolecular cases where cis stereochemistry was
observed, the intermolecular [2+2] cycloaddition of unactivatived
terminal alkenes promoted by (tricPDI)Fe(N2) (tricPDI = 2,6-
(2,4,6-tricyclopentyl)C6H2N=CMe)2C5H3N) produced exclusively
1,2-trans products (Scheme 1, left).4c The cyclopentyl-subtituted
aryl variant of the pyridine(diimine) ligand was used to suppress
transfer hydrogenation of the alkene observed previously with the
isopropyl variant, (iPrPDI)Fe(N2)2, arising from cyclometallation of
the aryl substituents.4c Aside from these phenomenological obser-
vations, little is known about the mechanistic details including the
nature of the active species, catalyst deactivation pathways, and
turnover-limiting step of these unique cycloaddition reactions.
Ultimately such information will inform design of next generation
catalysts with broader scope, lifetime, and activity. Here we de-
scribe a comprehensive mechanistic study into iron-catalyzed inter-
and intramolecular [2+2] cycloadditions of unactivated alkenes
(Scheme 2).
Scheme 2. Reactions investigated in this article and
associated mechanistic questions.
∎ RESULTS AND DISCUSSION (INTERMOLECULAR)
Kinetic Studies. Our investigations commenced with studies
into the iron-catalyzed [2+2] cycloaddition of 1-octene as a repre-
sentative alkene. In addition to being a readily available α-olefin, 1-
octene also has several advantageous physical properties including
a mid-range boiling point of 122 °C that allows for facile separation
2
ACS Paragon Plus Environment

Page 3 of 18
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
Figure 1. Kinetic analysis of iron-catalyzed [2+2] cycloaddition of 1-octene. A. Time course of the reaction from 0 to 48 hours. B. Initial
rates plots for determining the order in 1-octene. C. Initial rates plots for determining the order in catalyst. 1 mol % Fe is equal to 0.5 M in
(
tricPDI)Fe(N2). a = unknown order, yet to be determined. “Yield” indicates isolated yield of the product, whereas “conversion” indicates
conversion to product as determined by gas chromatography.
To determine the overall reaction order and rate-determining
step, kinetic analysis on the cyclodimerization of 1-octene was per-
formed. The progress of iron-catalyzed [2+2] cycloaddition of 1-
octene ([(tricPDI)Fe(N2)] = 0.5 M, [1-octene] = 50 M) was
monitored by gas chromatography over the course of 48 hours at
23 °C. A representative reaction time course is presented in Figure
1A and establishes well-behaved kinetic behavior at early conver-
sion with no induction period for the onset of catalysis. After ap-
proximately 20 turnovers, the rate of reaction begins to slow, lead-
ing to almost no catalysis after 24 hours, suggesting catalyst deacti-
vation and ultimately death. Because of the well-behaved kinetics
observed at early conversion, the order in iron catalyst and alkene
were determined using the method of initial rates (Figure 1B, 1C).
The data support a first order process in both the iron precatalyst
and 1-octene for intermolecular iron-catalyzed [2+2] cycloaddition
of alkenes.
understanding of dinitrogen coordination properties of the iron
precatalyst was required.
8
As has been reported previously, PDI iron dinitrogen complexes
often exist as equilibrium mixtures of mono- and bis-dinitrogen
species in both solution and in the solid state (Figure 2A). A com-
1
bination of H NMR, infrared, and 57Fe Mössbauer spectroscopies
were
tricPDI)Fe(N2):(tricPDI)Fe(N2)2 under various conditions.
Figure 2B presents the zero-field 57Fe Mössbauer spectrum of
tricPDI)Fe(N2)n in the solid state at 80 K. The majority of the
used
to
determine
the
ratio
of
(
(
spectrum (90%) corresponds to a quadrupole doublet with an iso-
mer shift (δ) of 0.39 mm/s and an absolute quadrupole splitting
(│ΔEQ│) of 2.07 mm/s, while the minor doublet has parameters δ
= 0.35 mm/s and │ΔEQ│ = 0.33 mm/s, corresponding to the bis-
9
dinitrogen complex. The spectroscopic data support the four-
coordinate, iron mono-dinitrogen complex as the major compound
in the solid state. The solid-state (KBr) infrared spectrum (Figure
2C, bottom) of (tricPDI)Fe(N2)n exhibits a strong band at 2036
cm-1, assigned as the N–N stretch of the coordinated N2 ligand of
the four-coordinate complex. A second, less intense band at 2125
cm-1 was also observed and assigned as the antisymmetric stretch
corresponding to (tricPDI)Fe(N2)2. The infrared data support
the same ratio (90:10) of iron compounds as determined by 57Fe
Mössbauer spectroscopy at 1 atm of N2. The ratio of four- to five-
coordinate iron dinitrogen compounds was also studied in ben-
Determination of Reaction Order in N2. The presence of
coordinated dinitrogen on the iron precatalyst suggested that re-
moving it from the reaction may facilitate substrate coordination
and improve overall catalytic performance. An inverse relationship
between pressure and conversion to product after 48 hours of reac-
tion was observed: 0 atm N2 – 44% conv., 1 atm N2 – 31% conv., 4
atm N2 – 18% conv. (Figure 1, top). To determine whether N2
pressure is a component of the Fe-catalyzed cyclodimerization rate
law and to accurately assign its reaction order (i.e. the precise con-
centration of N2 at the outset of any reaction), a more in-depth
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
zene-d6 solution at 1 atm of N2. The solution infrared spectrum
Page 4 of 18
1
2
3
4
5
6
7
8
(Figure 2C, top) exhibits symmetric and anti-symmetric N2
stretching bands for (tricPDI)Fe(N2)2 at 2125 and 2057 cm-1,
respectively. The broad line width of the band at 2057 cm-1 ob-
scures the N2 band for (tricPDI)Fe(N2) at 2038 cm-1. From these
data, at 1 atm N2 in benzene at 23 °C a value of Keq = 2.1 was estab-
lished for the interconversion of the mono- and bis-dinitrogen
10
compounds. This ratio was further confirmed by NMR spectros-
copy where the benzene-d6 1H NMR spectrum of
(
tricPDI)Fe(N2)n displayed broadened signals for the mono- and
9
bis-dinitrogen complexes in a 1:2.1 ratio, respectively. Degassing a
benzene-d6 solution of (tricPDI)Fe(N2)n leads to sharpening of
the signals and >98% conversion to (tricPDI)Fe(N2), as judged
by 1H NMR spectroscopy.
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 3. Kinetic analysis of iron-catalyzed [2+2] cycloaddition
of propylene at 0 and 1 atm N2. A. Time course of the reaction
from 0 to 48 hours at 0 and 1 atm N2. B. Initial rates plots for de-
termining the order in N2. 1 mol % [Fe] = 0.012 M. [N2 – 1 atm] =
0.026 M, [N2 – 0 atm] = 0.012 M (ratio of 2.15:1, initial rate ratio
found = 2.19:1).
Natural Abundance 13C KIE Measurements. As there are
several mechanistic possibilities consistent with the observed rate
law, measurement of kinetic isotope effects may shed light on irre-
versible bond-forming or -breaking processes during the reaction.
Using the method described by Singleton and previously applied by
,
11 12
our group in the iron-catalyzed [4+4] cycloaddition of dienes,
the 13C KIE for formation of trans-1,2-dihexylcyclobutane was
measured (Figure 4). Values near unity were obtained at all posi-
tions within the 4-membered ring and alkyl chains, consistent with
no heavy atom KIE and the absence of Fe–C or C–C bond-forming
or -breaking events in the first irreversible step of the reaction.
Figure
2.
Determining
the
ratio
of
(
tricPDI)Fe(N2):(tricPDI)Fe(N2)2 in solution and in the solid
state at 1 atm N2. A. Equilibrium mixture of mono- and bis-
dinitrogen complexes. B. Zero-field 57Fe Mössbauer spectrum of
(
tricPDI)Fe(N2)n at 80 K in the solid state. C. Stacked IR spectra
of (tricPDI)Fe(N2)n in benzene (top) and dispersed in a KBr
pellet (bottom). Both samples were prepared at and spectra ac-
quired at 23 °C and 1 atm N2.
To avoid experimental complications from monitoring the [2+2]
cycloaddition of 1-octene at different pressures of N2, propylene
was used as the alkene substrate as neat conditions were not re-
quired to obtained appreciable yields and the reaction was readily
monitored by 1H NMR spectroscopy over the course of 48 hours at
23 °C in benzene-d6. Time courses for the reaction at both 0 and 1
atm of N2 are presented in Figure 3A. As with 1-octene, the reaction
displays well-behaved kinetic behavior at early conversion, followed
by a decrease in rate as the reaction proceeds. Interestingly, while
the initial rate of reaction at 0 atm N2 is higher, the rate of catalyst
deactivation increased without exogenous N2, as both the 0 and 1
atm N2 reactions reach the same conversion after 24 hours (71%
conversion). This suggests that catalyst deactivation could involve
N2 displacement, rather than substrate inhibition (vide infra). Uti-
lizing the method of initial rates (Figure 3B), the order in N2 was
determined to be -1 overall, corroborating the inverse relationship
between conversion and N2 pressure observed in the [2+2] cy-
cloaddition of 1-octene.
Figure 4. Results of natural abundance 13C KIE analysis for iron-
catalyzed [2+2] cycloaddition of 1-octene.
Observation of Catalyst Resting State and Degrada-
tion Species. To gain additional information about the catalytic
cycle and iron speciation throughout the intermolecular [2+2]
reaction, isolation of catalytically relevant organometallic interme-
diates was explored. Due to the high lipophilicity of tricPDI and the
corresponding iron complexes, attempts to obtain single crystals
suitable for X-ray diffraction studies have been unsuccessful. Anal-
1
ysis by H NMR spectroscopy was similarly uninformative, as any
spectra acquired under conditions similar to the catalytic cycloaddi-
tion reaction (i.e. (tricPDI)Fe(N2) in the presence of excess 1-
octene or propylene) produced an array of signals that could not be
deconvoluted. Freeze-quench 57Fe Mössbauer spectroscopy was
therefore used to determine the number and nature of iron com-
pounds formed under catalytic conditions. In a typical procedure,
the [2+2] cycloaddition of 1-octene was stirred for 10 minutes
(approximately 5% conversion) and the sample rapidly frozen at 77
4
ACS Paragon Plus Environment

Page 5 of 18
Journal of the American Chemical Society
K and analyzed by 57Fe Mössbauer spectroscopy (Figure 5). Be-
cause there is no induction period in the catalytic reaction, this
1
2
3
4
early time point was chosen to avoid potential signal contamination
from deactivation byproducts that form at higher conversion. A
quadrupole doublet was observed with the parameters (δ = 0.37
5
mm/s; │ΔEQ│ = a
0.66 mm/s)13 being consistent with
(PDI)Fe(L)(N2) compound where L can be N2, PR3, or an al-
kene.14 These complexes are highly covalent and the electronic
structure at iron is best be described as a hybrid between low-spin
Fe(0) and Fe(II) coordinated to a closed-shell PDI ligand. The
ambiguity of oxidation state is a result of the redox non-innocence
of the PDI ligand and the high degree of π-back donation from
6
7
8
9
To investigate the mechanism of ligand dehydrogenation and to
understand the coordination of (dehydro-tricPDI) to iron during
the catalytic reaction, freeze-quench 57Fe Mössbauer experiments
were performed. The iron-catalyzed [2+2] reaction was conducted
for 8 hours and rapidly frozen at 77 K and then analyzed by 57Fe
Mössbauer spectroscopy (Figure 6). Two overlapping quadrupole
doublets were observed in a 61:39 ratio with parameters δ = 0.32
mm/s and │ΔEQ│ = 1.94 mm/s and δ = 0.33 mm/s and │ΔEQ│ =
0.92 mm/s, respectively. The latter signal has parameters strikingly
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
iron.9, These data, along with the rate law data, support
15
(
tricPDI)Fe(N2)(η2-1-octene) as the catalyst resting state at
16
early conversion to cyclobutane product.
To corroborate both the electronic structure and spectroscopic
data of (tricPDI)Fe(N2)(η2-1-octene), full molecule DFT
calculations were performed. At the B3LYP level of theory, the
lowest energy electronic configuration converged to a closed-shell,
S = 0 solution, reflecting the structure of the proposed resting state.
The DFT calculated 57Fe Mössbauer parameters of δ = 0.45 mm/s
and ΔEQ = 0.81 mm/s are in good agreement with the experimental
findings and within the generally accepted error range.17
similar
(
to
(PDI)Fe(L)(N2)
compounds
and
tricPDI)Fe(N2)(η2-1-octene), suggesting that the doublet
corresponds to (dehydro-tricPDI)Fe(N2), which likely exists as
a mixture of two olefin regioisomers.19 The major quadrupole dou-
blet has parameters that suggest an intermediate spin Fe(II) or
Fe(III) compound with an asymmetric ligand field, likely
(PDI)FeX2.20 While an olefin-derived iron metallacycle is a possi-
bility, it is unlikely given that little productive catalysis occurred at
this conversion. The DFT-computed (B3LYP level of theory) 57Fe
Mössbauer parameters (δ = 0.25 and ΔEQ = 2.04 mm/s) for the
putative cyclometallated iron hydride, (cyclomet-tricPDI)Fe-H
with a BS(3,1) electronic configuration reproduce the experimental
data. This compound, which likely exists as a mixture of isomers, is
possibly an intermediate en route to dehydrogenation of the cyclo-
pentyl ring.
Figure 5. Freeze-quench zero-field 57Fe Mössbauer spectrum of
iron-catalyzed [2+2] reaction after 10 minutes of reaction.
To understand the origin of catalyst deactivation, similar analyti-
cal experiments were conducted to identify the iron products. Fol-
lowing the [2+2] cycloaddition reaction, the ligand was removed
from iron by addition of water and the resulting isolated solid ana-
1
lyzed by H NMR spectroscopy. As illustrated in Scheme 3, one of
the cyclopentyl aryl substituents was converted to a cyclopentenyl
group, establishing dehydrogenation of the ligand to (dehydro-
tricPDI). A 1:1 mixture of olefin regioisomers (1-cyclopentenyl:2-
cyclopentenyl) along with a stoichiometric amount of octane (with
respect to the iron precatalyst) was also observed at the end of the
reaction (the latter confirmed by gas chromatographic analysis of
the reaction mixture), accounting for the loss of one equivalent of
H2 from the ligand. This indicates that (tricPDI)Fe(N2), in analo-
gy to other (PDI)Fe(N2) complexes, ultimately undergoes transfer
Figure 6. A. Freeze-quench zero-field 57Fe Mössbauer spectrum
of iron-catalyzed [2+2] reaction after 8 hours. B. Iron species in
the spectrum (likely a mixture of regioisomers).
To confirm the identity of both the catalyst resting state and its
degradation products, the [2+2] cycloaddition of 1-octene was also
monitored by infrared spectroscopy as the N-N stretching region is
diagnostic. As presented in Figure 7, aliquots of the iron-catalyzed
[2+2] reaction were taken at 10 minutes, 4 hours, and 18 hours,
and the solution samples analyzed by IR spectroscopy. At the 10
minute interval, a single band at 2117 cm-1 was observed assigned
to (tricPDI)Fe(N2)(η2-1-octene). At 4 hours, two new stretch-
es at 2096 and 2085 cm-1 were observed, likely arising from the two
5
hydrogenation from ligand to α-olefin, although the rate of dehy-
drogenation is substantially slower for (tricPDI)Fe(N2), which is
in line with the observed [2+2] reactivity and slower catalyst deg-
radation.4d,
18
Scheme 3. Isolation of dehydrogenated PDI ligand.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 18
regioisomers of (dehydro-tricPDI)Fe(N2). Finally, after 18
hours of reaction, the stretch at 2117 cm-1 was absent but the bands
at 2096 and 2085 cm-1 persisted supporting (tricPDI)Fe(N2)(η2-
1-octene) as the catalyst resting state that over time converts to
the deactivated product, (dehydro-tricPDI)Fe(N2) by dehydro-
genation of a cyclopentyl substituent.
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
Proposed Catalytic Cycle. Based on a combination of the
data disclosed in the previous section, a catalytic cycle for iron-
catalyzed intermolecular [2+2] cycloaddition of α-olefins is pro-
posed (Scheme 4). The lack of any detectable natural abundance
13C KIE at any cyclobutane carbon of the product indicates that
neither oxidative cyclization nor reductive elimination are the first
irreversible steps, and therefore, potentially not rate-limiting. This
suggests that ligand substitution, dissociation, or association are
viable candidates for the rate-determining step. The freeze-quench
57Fe Mössbauer and IR monitoring studies support
(
tricPDI)Fe(N2)(η2-1-octene) as the catalyst resting state,
consistent with rapid coordination of the first alkene to the precata-
lyst. Using the rapid pre-equilibrium approximation (assuming that
K2 > k3) a rate law derived from the reaction mechanism presented
in Scheme 4 matches that obtained from the kinetic data (rate =
kobs[Fe]1[1-octene]1[N2]-1).
Figure 7. Top, progression of PDI iron speciation over the
course of the iron-catalyzed [2+2] reaction. Bottom, solution in-
frared spectrum (1-octene, 23 °C) throughout the reaction (N2
region).
As presented above, catalyst degradation, not inherently slow re-
activity or unfavorable thermodynamics, was the cause for the low-
er yields of certain substrates in the iron-catalyzed alkene cyclodi-
merization. Support for this hypothesis was obtained from catalyst
loading studies (Table 1). With 3 mol % (tricPDI)Fe(N2), α-
olefins including propylene, 1-hexene, and 1-octene underwent
[2+2] cycloaddition in >98% yield and >99% regio- and diastere-
oselectivity after 24 hours. 4-Methyl-1-pentene, allylbenzene and
homoallylbenzene similarly produced the corresponding trans-
cyclobutanes in high yields (92%, 97%, and >98%, respectively)
with >99% regio- and diastereoselectivity after 24 hours. As the
reaction is first order in iron precatalyst, the increased catalyst load-
ing likely increases the rate of reaction to surpass the rate of catalyst
degradation by ligand dehydrogenation and allow for complete
conversion of the olefins to the corresponding substituted cyclobu-
tane.
Table 1. Iron-catalyzed [2+2] cycloaddition of alpha-olefins at
two different catalyst loadings. *Indicates values taken from refer-
ence 4c.
Scheme 4. Proposed catalytic cycle for the iron-catalyzed intermolecular [2+2] cycloaddition of α-olefins. Rds =
rate-determining step. Ar = 2,4,6-tricyclopentylaryl. R = hexyl. k5 and k5’ are possible turnover pathways of the
cycle.
6
ACS Paragon Plus Environment

Page 7 of 18
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
The reaction commences with formation of the catalyst resting
fore be unable to mix with an antisymmetric PDI orbital. However,
the presented dz2 orbital is the corresponding antibonding interac-
tion of the parent Fe dz2 orbital and the alkene π-orbital. This de-
forms the presented dz2 orbital towards B2 symmetry (causing the
lobe opposite the alkene to enlarge, while the lobe pointed towards
the alkene is diminished) and allows for mixing. The overall low
symmetry of the molecule also aids in this mixing. The lowest oc-
cupied molecular orbital (LUMO) is the corresponding antibond-
ing interaction between dz2 and PDI π*, of which dz2 only consti-
tutes 17% of the total molecular orbital. Because the low energy,
state (tricPDI)Fe(N2)(η2-1-octene) followed by unimolecular
dissociation of N2 from this complex to form (tricPDI)Fe(η2-1-
octene). This equilibrium process likely involves a spin-transition
from S = 0 to S = 1 and precedes rate-determining association of 1-
octene to the four-coordinate complex to form (tricPDI)Fe(η2-1-
octene)2 where the PDI ligand has dissociated an imine arm and
2
21
is bound κ . This accounts for both the inverse order in N2 con-
centration and the first-order dependence in 1-octene. The subse-
quent steps along the catalytic cycle are fast and likely include:
oxidative cyclization to produce the ferrocyclopentane and reduc-
tive elimination of the resulting metallacycle to yield the cyclobu-
tane product. Due to the high coordination affinity of N2 to
empty
(
orbital
required
for
ligand
association
at
tricPDI)Fe(N2)(η2-1-octene) is largely ligand-based (instead
of metal-based), combined with the fact that it is an 18-electron
complex, the catalyst resting state likely cannot undergo associative
ligand substitution. The most likely productive process is dissocia-
tion of N2, consistent with the observed rate law.
(
tricPDI)Fe, it is difficult to know precisely how the catalyst turns
over. Following reductive elimination, the coordinatively unsatu-
rated iron catalyst can either coordinate 1-octene (Scheme 4,
pathway with rate constant k5), or uptake N2 (Scheme 4, pathway
with rate constants k5’ and k1).9 While both routes are possible from
a thermodynamic standpoint (see Figure S36 in the Supporting
Information for details), it is ambiguous whether the catalyst rest-
ing state is on- or off-cycle, although N2 pressure likely affects
which turnover pathway is preferred (i.e. k5 for low N2 pressure and
k5’ for high N2 pressure).
To better understand the persistence of (tricPDI)Fe(N2)(η2-1-
octene) and why direct olefin association to it is not observed, a
more thorough analysis of the electronic structure of the catalyst
resting state was performed. The DFT-calculated structure of
(
tricPDI)Fe(N2)(η2-1-octene) supports a low-spin iron center
bound to a doubly-reduced, redox non-innocent PDI ligand. The
oxidation state ambiguity at iron is due to the strong π-back dona-
tion from iron to PDI.9 Indeed, the highest occupied molecular
orbital (HOMO) of the complex is an ad-mixture of a dz2-type
orbital localized on iron and a PDI π* orbital of B2 symmetry (Fig-
ure 8). Normally dz2 is a totally symmetric orbital and would there-
Figure 8. Qualitative frontier molecular orbital diagram of
(
tricPDI)Fe(N2)(η2-1-octene) along with representations of
the HOMO and LUMO of the complex.
While additional details of N2 dissociation from
(
tricPDI)Fe(N2)(η2-1-octene) are difficult to ascertain (as it
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 18
of deuterated isotopologues and isotopomers of the α,ω-diene
can involve a number of different steps, including imine dissocia-
1
2
3
4
5
6
7
8
tion before, after, or during spin transition), the net result of any
pathway is formation of (tricPDI)Fe(η2-1-octene). DFT calcu-
lations on (dicPDI)Fe(η2-propylene) were performed with
B3LYP level of theory to better understand this four-coordinate
complex (the octyl group was truncated to a methyl group and the
tricyclopentylaryl groups were truncated to dicyclopentylaryl
groups). The complex converged to a broken symmetry (1,1) solu-
tion where an S = ½ Fe(I) center is ferromagnetically coupled to an
unpaired spin on the singly-reduced PDI ligand. The magnetic
orbitals of (dicPDI)Fe(η2-propylene) are dz2 and (dxz+PDI)*,
which have a spatial overlap of S = 0.04, a remarkably low value due
to the mismatched symmetry of the two orbitals, preventing effec-
more convenient and synthetically feasible (vide infra). In addition,
1,7-octadiene contains no heteroatoms to potentially complicate
the coordination chemistry around iron. In the presence of 5 mol%
(
tricPDI)Fe(N2) in benzene-d6, 1,7-octadiene underwent [2+2]
cycloaddition to form cis-bicyclo[4.2.0]octane in >98% conversion
and 92:8 d.r. after 9 hours at 23 °C (Figure 9, top). The major cis
diastereomer is opposite the trans configuration observed from the
intermolecular reaction. Thermodynamic parameters for the reac-
tion, ΔG° = -8.9 kcal·mol-1, ΔH° = -14.4 kcal·mol-1 and ΔS° = -14.7
cal·mol-1·K-1 were computed by DFT (B3LYP level of theory)
(Figure 9, top), where the entropic penalty of the unimolecular
cycloaddition is significantly lower than the intermolecular variant.
Furthermore, upon completion of the reaction, addition of more
α,ω-diene re-initiated catalysis, demonstrating that deactivation of
the iron compound is less pronounced with 1,7-octadiene as com-
pared to 1-octene.
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
22
tive overlap and coupling. The relative energies of the singlet and
triplet states of this complex differ by only by 2.1 kcal·mol-1, favor-
ing the triplet configuration. Due to this narrow singlet-triplet gap,
it is difficult to accurately determine which spin state is lowest in
23
energy and whether spin-crossover occurs explicitly before or after
N2 dissociation, the latter scenario occurring via facile thermal ac-
cess of the triplet state, not necessarily via spin-orbit coupling (see
Supporting Information for preliminary computational studies on
this process).
Upon formation of
(
tricPDI)Fe(η2-1-octene), the 4-
coordinate complex can either react productively with an alkene or
unproductively with dinitrogen (and re-form the catalyst resting
state). As mentioned above, N2 likely has a higher binding constant
for (tricPDI)Fe than an alkene, corroborated by the observed in-
verse order in [N2] for the [2+2] cycloaddition reaction. This re-
sults in association of 1-octene to (tricPDI)Fe(η2-1-octene)
becoming the rate-determining step of the reaction. Additionally,
the only molecular orbital of appropriate symmetry and energy for
ligand association in (tricPDI)Fe(η2-1-octene) is dz2, however
this orbital is partially filled, which likely increases the kinetic barri-
er for 1-octene addition and further contributes to the step being
rate-limiting.
Because C–C and Fe–C bond formation occur after the rate-
determining step, it is difficult to experimentally determine the
mechanism and molecularity of reductive elimination and oxidative
cyclization in the intermolecular cycloaddition. It does, however,
explain why intermolecular alkene-alkene cycloadditions are less
efficient and more plagued by competing catalyst decomposition
than those with conjugated and α,ω-dienes. Formation of the bis-
olefin iron complex (which leads to the initial C–C bond for-
mation) becomes the highest barrier of the reaction, manifesting
itself in rate-determining addition of 1-octene to (tricPDI)Fe(η2-
1-octene). This brings the rate-constants of other pathways such
as N2 association and ligand C-H activation closer to that of C–C
bond formation and leads to the observed rapid ligand dehydro-
genation and inverse order in N2. Improvements to alkene-alkene
iron cycloaddition methodologies will include development of N2-
free reduced iron precatalysts, as well as increasing the binding
affinity of alkenes over other ligands.
∎ RESULTS AND DISCUSSION (INTRAMOLECULAR)
Representative Intramolecular Reaction. Investigations
into the iron-catalyzed intramolecular [2+2] cycloadditions of
alkenes were conducted with 1,7-octadiene as a representative
substrate. As with the selection of 1-octene for the intermolecular
variant, a boiling point of 117 °C makes the synthesis and isolation
8
ACS Paragon Plus Environment

Page 9 of 18
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
Figure 9. Kinetic analysis of iron-catalyzed intramolecular [2+2] cycloaddition of 1,7-octadiene. A. Time course of the reaction from 0 to
9 hours at 23 °C. B. Initial rates plots for determining the order in iron catalyst. C. Initial rates plots for determining the order in 1,7-
octadiene. D. Initial rates plots for determining the order in N2. y = 0 or 1 atm of N2.
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
Kinetic Analysis. The rate law for the iron-catalyzed [2+2]
cycloaddition of 1,7-octadiene was determined to gain mechanistic
insight and for comparison to the intermolecular case. The catalytic
produced the bicycle in similar yield and diastereoselectivity. Stir-
ring a hexanes solution of (dicPDI)Fe(N2)2 with excess 1,7-
octadiene produced a red solid that, after washing with pentane,
afforded (dicPDI)Fe(bimetallacyclo[4.3.0]nonane) in 41%
yield (Scheme 5). The complex was characterized by a combina-
tion of solution-state magnetic measurements, X-ray diffraction,
and 1H NMR spectroscopy.
1
reaction was monitored by H NMR spectroscopy over the course
of 9 hours at 23 °C ([(tricPDI)Fe(N2)] = 0.043 M, [1,7-
octadiene] = 0.86 M) and the time course presented in Figure 9A.
The data support a well-behaved kinetic profile throughout the
reaction with no evidence for an induction period or catalyst deac-
tivation. The rate of the reaction is essentially constant throughout
and only drops off towards the end of the reaction (low substrate
concentration), suggesting saturation behavior with the α,ω-diene.
The method of initial rates was used to determine the order in iron
precatalyst, diene and nitrogen pressure (Figures 9B-D). The cata-
lytic reaction was found to be first order in iron precatalyst and
Scheme
5.
Synthesis
of
(
dicPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane).
24
zeroth order in 1,7-octadiene and N2 pressure. The observed ze-
roth order in diene supports saturation kinetics and a catalyst-
25
Single crystals suitable for X-ray diffraction studies were obtained
from a concentrated pentane solution at -35 °C over 24 hours. A
representation of the solid-state structure is presented in Figure 10,
where the geometry at iron is best described as a distorted square-
based pyramid (τ5 = 0.35) with the three nitrogen atoms of the PDI
ligand and the carbon trans to the pyridine comprising the basal
plane of the complex. The Cimine–Nimine and Cipso–Cimine bond dis-
tances are established metrics for determining the redox state of the
substrate complex as the resting state.
Isolation and Characterization of Catalytically Rele-
vant Iron Metallacycles. Because catalyst deactivation was
minimal, the isolation of catalytically relevant intermediates was
pursued. Attempts to obtain X-ray quality crystals from mixtures of
(
tricPDI)Fe(N2) and 1,7-octadiene were unsuccessful as the
resulting iron complex(es) were too soluble in common organic
solvents. To reduce lipophilicity and improve crystallinity, the
pyridine(dimine)
ligand.27
For
(
dicPDI)Fe(trans-
dicPDI
ligand
(
dicPDI
=
2,6-(2,6-
bimetallacyclo[4.3.0]nonane), Cimine–Nimine distances of
1.331(2) and 1.329(2) Å and Cipso–Cimine values of 1.427(2) and
1.430(2) Å are consistent with a singly-reduced form of the chelate,
supporting an Fe(III) oxidation state. The data were of sufficient
9
26
dicyclopentyl)C6H3N=CMe)2C5H3N) was used instead. The re-
sulting iron dinitrogen complex, (dicPDI)Fe(N2)2, was catalyti-
cally competent for the [2+2] cycloaddition of 1,7-octadiene and
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 18
quality that the hydrogen atoms of the bimetallacycle were located protons alpha to iron, due to enhanced spin polarization at these
positions.29 The observed number of peaks is consistent with a Cs
symmetric molecule, suggesting the bimetallacycle undergoes rapid
ring flips on the NMR time scale. A solution-state magnetic mo-
ment of 2.9(1) μB was measured at 23 °C (Evans’ method, benzene-
d6), consistent with two unpaired electrons and an S = 1 ground
state. The magnetic data, in combination with metrical data from
1
2
3
4
5
6
7
8
freely and refined, allowing for unambiguous assignment of the
stereochemistry of the bimetallacycle. Notably, the hydrogens are
in a trans configuration, the opposite orientation of the diastere-
omer obtained from the catalytic reaction. Grubbs and co-workers
reported nickel bimetallacyclo[4.3.0]nonane complexes that inter-
converted between the cis and trans isomers of the bimetallacyle at
28
X-ray
diffraction
studies
support
(
dicPDI)Fe(trans-
0 °C, suggesting that isomerization of metallacyclic intermediates
bimetallacyclo[4.3.0]nonane) as an intermediate spin
Fe(III) (SFe = 3/2) compound ligated to a singly reduced pyri-
dine(diimine) chelate, leading to the observed S = 1 ground state.
in the current iron-catalyzed intramolecular [2+2] cycloaddition
reaction may also be operative (vide infra). To reduce strain and
accommodate the trans stereochemical arrangement, the cyclohex-
ane ring of the bicycle adopts a chair conformation, while the
metallacyclopentane ring adopts a half-chair conformation.
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
Although
isolation
of
(
tricPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane) was difficult, the 1H NMR
spectrum and solution magnetic moment measurements are similar
to those of the (dicPDI) variant.
Scheme 6.
bicyclo[4.2.0]octane
Rapid reductive elimination of cis-
from
(
dicPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane) in solution at 23 °C.
To gain insight into the electronic structure of
(
dicPDI)Fe(trans-bimetallacyclo[4.3.0]nonane),
full-
molecule DFT studies were conducted using the B3LYP level of
theory. The complex converged to a broken symmetry (3,1) solu-
tion where the resulting Mulliken spin density plot was also gener-
ated and supports an intermediate-spin Fe(III) center antiferro-
magnetically coupled to
a spin delocalized over the pyri-
dine(diimine) chelate with minor contributions from the Fe–C
orbitals of the metallacycle (Figure 11). Because d-orbitals make
such a significant contribution to the two magnetically coupled
orbitals of the molecule, there is both positive and negative spin
density at iron. Examination of the qualitative d-orbital splitting
diagram identified the magnetically coupled orbitals as both having
dyz parentage, with the spin up orbital containing significant contri-
butions from a PDI π* orbital (due to strong π-back donation) and
the spin down orbital containing significant contributions from a
dyz orbital mixed with Fe – C σ bonding orbitals of the metallacycle
(Figure 12).
Figure 10. Representation of the solid-state structure of
(
dicPDI)Fe(trans-bimetallacyclo[4.3.0]nonane)
with
thermal ellipsoids at 30% probability. Hydrogen atoms (except for
those at the ring juncture of the bimetallacycle) omitted for clarity.
Selected bond distances (Å) and angles (deg): Fe1 – N1 2.040(1),
Fe1 – N2 1.908(1), Fe1 – N3 2.038(1), N1 – C2 1.331(2), N3 –
C5 1.329(2), C2 – C3 1.427(2), C4 – C5 1.430(2), Fe1 – C10
2.024(1), Fe1 – C17 2.047(1); N1 – Fe1 – N2 78.09(4), N1 – Fe1
– N3 145.65(4), N2 – Fe1 – N3 77.73(4), N2 – Fe1 – C10
166.51(5), N1 – Fe1 – C10 99.17(5), N3 – Fe1 – C10 98.69(5),
N2 – Fe1 – C17 108.78(5), C10 – Fe1 – C17 84.69(5).
1
Attempts to obtain a H NMR spectrum of (dicPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane) in benzene-d6 were unsuccess-
ful, as the complex underwent rapid reductive elimination to form
cis-bicyclo[4.2.0]octane in 93:7 d.r. along with (dicPDI)Fe(N2)2
(Scheme 6). These observations support conversion of the ob-
served trans-metallocycle to the more reactive cis isomer which has
a lower barrier for C(sp3)–C(sp3) reductive elimination. In the
presence of a slight excess of 1,7-octadiene, a single paramagnetic
species with a signal range from -20 to 120 ppm was observed along
with the α,ω-diene and product. The most isotropically shifted
signals correspond to protons located on the pyridine and imine
methyl groups of the pyridine(diimine) chelate and the methylene
10
ACS Paragon Plus Environment

Page 11 of 18
Journal of the American Chemical Society
iPr(TB)PDI
(
iPr(TB)PDI
=
2,6-(2,6-
1
2
3
4
(diisopropyl)C6H3N=C(CH2)3)2(C5HN)) was explored. Previous
work has demonstrated that bis-olefin and metallacycle iron com-
plexes of this chelate are more crystalline and less reactive than
untethered variants.5 Using conditions similar to those used for the
synthesis of the dicPDI variant,
(
iPr(TB)PDI)Fe(trans-
5
bimetallacyclo[4.3.0]nonane) was isolated as a red solid in
72% yield (Scheme 7). The complex was stable in solution for ex-
tended periods of time (>48 hours) and the benzene-d6 NMR spec-
6
7
8
9
10
11
12
13
14
15
16
17
18
19
trum was remarkably similar to
(
dicPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane), containing 19 paramagneti-
cally shifted signals from -20 to 150 ppm, supporting analogous
solution-state structure for the dicPDI and iPr(TB)PDI variants. A solu-
tion magnetic moment of 2.9(1) μB, was measured at 23 °C (Evans’
method) supporting an S = 1 ground state.
Scheme
7.
Synthesis
of
(
iPr(TB)PDI)Fe(trans-
bimetallacyclo[4.3.0]nonane).
Figure 11. Mulliken spin density plot of (dicPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane)
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
Single crystals suitable for X-ray diffraction studies were obtained
from a concentrated pentane solution at -35 °C over 24 hours. The
solid-state
structure
of
(
iPr(TB)PDI)Fe(trans-
bimetallacyclo[4.3.0]nonane) supports a distorted square-
based pyramid geometry at iron (τ5 = 0.38) and is analogous to the
dicPDI variant (see Supporting Information for a representation of
the solid-state structure). All these data support the assertion that,
aside from its reduced reactivity, both
dicPDI)Fe(bimetallacyclo[4.3.0]nonane) have identical
solution-state, solid-state, and electronic structures. The zero-field
57Fe
Mössbauer spectrum of
iPr(TB)PDI)Fe(bimetallacyclo[4.3.0]nonane) was record-
(
iPr(TB)PDI) and
(
(
ed at 80 K in the solid state and is presented in Figure 13. An iso-
mer shift of 0.31 mm/s and a quadrupole splitting of 2.83 mm/s
were obtained, parameters that correlate well with the calculated
values found for (dicPDI)Fe(bimetallacyclo[4.3.0]nonane),
as well as other pyridine(diimine) iron metallacycles previously
reported.5
Figure 12.
Qualitative d-orbital splitting diagram for
(
dicPDI)Fe(trans-bimetallacyclo[4.3.0]nonane).
Due
to
the
propensity
of
(
dicPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane) to undergo reductive elimina-
tion and the scarcity of (dicPDI), sufficient quantities of the com-
pound for a zero-field 57Fe Mössbauer spectrum were not available.
DFT-generated 57Fe Mössbauer parameters of δ = 0.34 mm/s and
ΔEQ = 2.14 mm/s are in agreement with PDI iron metallacycles.5
To ensure accuracy of these DFT calculations and to investigate
whether
other
PDI
ligands
support
trans-
bimetallacyclo[4.3.0]nonanes (or whether dicPDI is unique in this
regard), the synthesis of the corresponding metallacycle with
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 18
Figure 13. Solid-state zero-field 57Fe Mössbauer spectrum of
iPr(TB)PDI)Fe(trans-bimetallacyclo[4.3.0]nonane).
Acquired at 80 K.
As stated
bimetallacyclo[4.3.0]nonane) is stable in benzene-d6 with no
evidence for reductive elimination. It is likely that introduction of
alkyl substitution at the 3 and 5 positions of the pyridine ring gen-
erates a more electron rich iron center, raising the barrier for
C(sp3)–C(sp3) reductive elimination. To induce C–C bond for-
1
2
3
4
5
6
7
8
(
previously,
(
iPr(TB)PDI)Fe(trans-
mation,
(
a
benzene-d6
solution
of
9
Proposed Catalytic Cycle. The combination of the kinetic
iPr(TB)PDI)Fe(bimetallacyclo[4.3.0]nonane) was exposed
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
data, identification of the catalyst resting state, and a primary 13C
kinetic isotope effect, support the catalytic cycle proposed in
Scheme 10. As the rate law has no dependence on diene concentra-
tion nor N2 pressure, this suggests that neither is involved in the
isomerization of the bimetallacycle nor reductive elimination and
that both processes are unimolecular. Any substitution reactions or
equilibria that involve 1,7-octadiene or N2 occur before formation
of the catalyst resting state and are, therefore, kinetically irrelevant.
Using the steady-state approximation and saturation kinetics (i.e.
[Fe] = [catalyst resting state]), a rate law for the mechanism pre-
sented in Scheme 10 was derived and matches that obtained from
the kinetic analyses. (tricPDI)Fe(N2) rapidly reacts with 1,7-
octadiene to form (tricPDI)Fe(η2:η2-1,7-octadiene), which
to 1 atm of CO. An immediate color change to blue was observed
with concomitant formation of trans-bicyclo[4.2.0]octane in >99:1
d.r. (Scheme 8). The higher selectivity of the stoichiometric C–C
bond-forming reaction as compared to the catalytic variant (92:8
d.r.) suggests that CO-induced reductive elimination is more rapid
than unimolecular reductive elimination of the bimetallcycle during
the catalytic [2+2] cycloaddition reaction and prevents any isomer-
ization of the bimetallacycle.
Scheme 8. Addition of CO to
bimetallacyclo[4.3.0]nonane).
(
iPr(TB)PDI)Fe(trans-
undergoes oxidative cyclization to produce
(
tricPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane), the catalyst resting state. An
equilibrium between trans- and cis-diastereomers of the bimetal-
lacycle precedes rate-limiting, unimolecular reductive elimination
to furnish predominantly the cis-product (due to k6 >> k5). Rapid
recombination of the iron catalyst and substrate regenerate the
pyridine(diimine) bis-olefin complex and turn over the cycle.
To gain additional information about the product- and selectivity-
determining step of the intramolecular [2+2] reaction, the 13C/12C
kinetic isotope effect was measured for the formation of natural
abundance cis-bicyclo[4.2.0]octane. Because the resting state of
the catalyst is the trans-bimetallacycle, there are two possibilities
for the first irreversible step of the reaction (Scheme 9, left). The
first is irreversible isomerization from trans- to cis-bimetallacycle
followed by rapid reductive elimination, giving rise to a small KIE
30
(either primary or inverse) at all carbons of the cyclobutane. The
second possibility is rapid equilibration of the trans- and cis-
bimetallacycles followed by irreversible reductive elimination,
where a primary KIE at the terminal cyclobutane carbons would be
expected.11b Indeed, a primary 13C kinetic isotope effect of 1.022(4)
was measured at 23 °C for the terminal cyclobutane carbon of cis-
bicyclo[4.2.0]octane, supporting unimolecular reductive elimina-
tion that is both irreversible and selectivity-determining (Scenario
2). No other significant KIE values were measured at any other
carbon of the bicycle. Rapid and reversible equilibration of the
trans- and cis-bimetallacyles before irreversible reductive elimina-
tion and possible Curtin-Hammett kinetics are also consistent with
the observed diastereoselectivity of the reaction.
Scheme 9. Possible kinetic isotope scenarios for dif-
ferent irreversible steps in the iron-catalyzed [2+2]
cycloaddition of 1,7-octadiene (left). Observed 13C
KIE values for cis-bicyclo[4.2.0]octane (right).
12
ACS Paragon Plus Environment

Page 13 of 18
1
Journal of the American Chemical Society
Scheme 10. Proposed catalytic cycle for the iron-catalyzed intramolecular [2+2] of 1,7-octadiene. rds = rate-
2
determining step. Ar = 2,4,6-tricyclopentylaryl.
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
Due to the chelate effect, 1,7-octadiene has a stronger coordina-
tion affinity to [(tricPDI)Fe] than 1-octene, significantly lowering
the barrier for N2 displacement from (tricPDI)Fe(N2) and alkene
association to a putative (tricPDI)Fe(η2-1,7-octadiene), mak-
ing the corresponding rate constants to these processes kinetically
inconsequential. This causes a shift in the catalyst resting state and
the rate-determining step of the reaction. This phenomenon also
explains the reduced catalyst decomposition rates observed in in-
tramolecular [2+2] reactions: the iron catalyst is more likely to
exists as a 1,7-octadiene adduct (either as the bis-olefin complex or
the bimetallacycle), than interact with a cyclopentyl group on the
ligand and undergo dehydrogenation.
tained for the [2+2] cycloaddition of 1,7-octadiene, the total barri-
er for reductive elimination from the trans-metallacycle (the cata-
lyst resting state) to the cis-product can be obtained and is equal to
23.1 kcal·mol-1 (at 23°C). This value does not correspond to ΔG‡
,
cis
but the sum of ΔG‡ and the energy difference between the trans-
cis
and the cis-bimetallacycles (ΔGmet). Using DFT methods (B3LYP
level of theory), ΔGmet was calculated to be 3.7 kcal·mol-1, with the
trans-diastereomer having the lower ground state energy, reflecting
the observed isolation of only
(
dicPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane). Qualitatively, the cis-
bimetallacycle should be higher in energy than the trans-isomer as
it contains two additional gauche butane interactions within the
cyclohexane ring (each worth approximately 0.9 kcal·mol-1), in
32
Another question that arises in the proposed catalytic cycle is, why
is the resting state of the intramolecular [2+2] cycloaddition the
trans-bimetallacycle, while the product is predominantly the corre-
sponding cis-bicycle? As the proposed catalytic cycle includes the
equilibration of two diastereomeric bimetallacycles before the rate-
determining step and a distribution of both products is formed
addition to the steric interaction with one of the 2,4,6-
tricyclopentylaryl groups on the pyridine(diimine) ligand (none of
which are present in the trans-isomer). Applying both the experi-
mental and computational values, ΔG‡ = 19.4 kcal·mol-1 and
cis
ΔG‡ = 24.5 kcal·mol-1 were obtained and indicate that while the
trans
trans-bimetallacycle is lower in energy, the transition state for re-
ductive elimination from the cis-bimetallacycle is lower in energy,
explaining the observed preference for formation of cis-
31
(92:8 d.r.), a Curtin-Hammett scenario is likely to exist. A reaction
coordinate diagram describing this Curtin-Hammett scheme: in-
terconversion of trans- and cis-bimetallacycles and their respective
reductive elimination pathways to form their respective bicycles, is
presented in Scheme 11 (the equilibration of trans- and cis-
bimetallacycles is simplified to a single step). ΔΔG‡ can be calculat-
ed from the diastereomeric ratio of bicyclo[4.2.0]octane formed
during the reaction and is equal to 1.4 kcal·mol-1 favoring the cis
diastereomer. Utilizing the Eyring equation and kinetic data ob-
bicyclo[4.2.0]octane. During the synthesis of (dicPDI)Fe(trans-
33
bimetallacyclo[4.3.0]nonane), the equilibrium between the
trans- and cis-bimetallacycles is likely perturbed by the low solubili-
ty of the trans-isomer in hexanes, which precipitates the product,
slows isomerization/reductive elimination, and allows for isolation.
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 18
association to (tricPDI)Fe(η2-1-octene), we can surmise that it
is analogous to the intramolecular case. Attempts to calculate a
potential ligand-induced (N2, alkene, α,ω-diene, etc.) reductive
elimination pathway from PDI iron metallacycles or bimetallacy-
cles was unsuccessful, as no global minima were located. In the next
section, experiments are detailed to probe the mechanism of reduc-
tive elimination.
Scheme 11. Curtin-Hammett scenario for the reduc-
1
2
3
4
5
6
tive
elimination
of
bicyclo[4.2.0]octane
from
(
(
dicPDI)Fe(bimetallacyclo[4.3.0]nonane).
dicPDI)Fe.
“Fe”
=
7
8
9
∎ STUDIES ON REDUCTIVE ELIMINATION
Investigations into the Reductive Elimination Step of
[2+2] Cycloadditions. To better understand the pathway for
C(sp3)–C(sp3) reductive elimination in both inter- and intramo-
lecular [2+2] cycloadditions, experiments were devised that would
allow probes of a stepwise or concerted process. A “stepwise”
mechanism is defined as initiating with Fe–C bond homolysis,
generation of a carbon-centered radical followed by radical recom-
bination to both cleave the second Fe–C bond and form the cyclo-
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
34
butane product. As proposed in Scheme 12, generating metallacy-
cles from stereodefined, deuterated olefins allows for the differenti-
ation of both pathways. In a concerted pathway, any relative stereo-
chemistry between the deuteria and other groups on the ring
should be retained in the cyclobutane product and a single isoto-
pomer should be observed. Conversely, if a stepwise mechanism is
operative, generation of carbon-centered radicals would epimerize
the stereocenters and a mixture of isotopomers would be observed.
To perform these experiments, both 1-deutero-E-1-octene and 1,8-
dideutero-E,E-1,7-octadiene were synthesized and subjected to
catalytic cycloaddition conditions.
Scheme 12. A. Summary of inter- and intramolecular
iron-catalyzed [2+2] cycloadditions examined in this
article. B. Possible strategy for distinguishing between
reductive elimination pathways.
When comparing the inter- to the intramolecular iron-catalyzed
[2+2] cycloadditions, the most obvious difference is the change in
diastereoselectivity of the 1,2-cyclobutane: the intermolecular reac-
tion is trans-selective while the intramolecular is cis-selective. As
there is less experimental data on the product-forming step of the
intermolecular [2+2] cycloaddition reaction, we currently do not
know the precise origin of this change in selectivity. While the
DFT-computed (B3LYP) ground-state energy difference between
cis-
vs.
trans-(dicPDI)Fe(2,3-
dimethylmetallacyclopentane) is 6.7 kcal·mol-1 favoring the
trans-isomer (an even more pronounced preference than in the
intramolecular case), this is by no means definitive as it reveals
nothing about the difference in transition state energies for reduc-
tive elimination of the two metallacycles. A high barrier for trans-
to-cis metallacycle isomerization is a possible explanation for the
observed exclusive trans-selectivity, while another is that the rela-
tive stereochemistry of the metallacycle (and therefore the prod-
uct) is determined during the rate-determining step of the reaction:
olefin coordination to (tricPDI)Fe(η2-alkene). The trans-
selectivity could then be explained via a straightforward steric ar-
gument, as the incoming alkene would coordinate to give a trans
arrangement of the olefin R-groups to minimize interactions with
each other and the imine aryl groups on the PDI ligand.
Deuterium Labelling Studies: 1-octene. Exposure of 1-
deutero-E-1-octene to the intermolecular [2+2] cycloaddition
conditions produced trans-1,2-dihexylcyclobutane in 23% yield and
a combination of 1H and 2H NMR and mass spectrometry (GC-MS
in EI mode) was used to determine the distribution of deuterium in
the product as well as any isotopomers that formed in the reaction
(results presented in Figure 14). The product was a complex mix-
ture of isotopomers with deuterium incorporated statistically
around all carbon positions of the cyclobutane ring, with no incor-
poration along the alkyl chain of the product. Mass spectrometry
In the [2+2] cycloaddition of 1,7-octadiene, reductive elimination
is likely unimolecular and therefore, not ligand-induced, as was
reported for the PDI iron-catalyzed cross [2+2] cycloaddition of
dienes and alkenes.4d Additionally, while we cannot know the pre-
cise molecularity of reductive elimination in the [2+2] cycloaddi-
tion of 1-octene, as this step occurs after rate-determining olefin
ACS Paragon Plus Environment

Page 15 of 18
Journal of the American Chemical Society
established that the product was also a mixture of isotopologues,
Scheme 13.
Addition of 1-deutero-E-1-octene to
(
tricPDI)Fe(N2).
1
2
3
4
5
6
7
8
with only 29% of the material corresponding to the dideutero
product (the expected isotopologue) and the rest being a mixture
from d0-d5-1,2-dihexylcyclobutane. As stated previously in this
article, ligand cyclometallation is a likely initiation pathway for
catalyst degradation and could also be the source of deuterium
scrambling. To confirm this hypothesis, (dehydro-tricPDI) was
also isolated from the reaction mixture and examined by a combi-
nation of 2H NMR spectroscopy and high-resolution mass spec-
trometry (LC-MS, ESI mode). Indeed, the ligand was found to be
a complex mixture of isotopologues ranging from d0 to d9. D-
incorporation was also observed at multiple positions across both
the cyclopentyl and cyclopentenyl groups of (dehydro-tricPDI).
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
Deuterium Labelling Studies: α,ω-dienes. A similar deu-
terium labelling study was performed for the iron-catalyzed intra-
molecular [2+2] cycloaddition of 1,8-dideutero-E,E-1,7-octadiene.
Under the standard reaction conditions, cis-bicyclo[4.2.0]octane
was formed in 75% yield as a mixture of both isotopologues and
isotopomers. Analysis by 2H NMR spectroscopy revealed that
deuterium was distributed statistically over all cyclobutane carbons,
similar to the intermolecular case (Figure 15). Mass spectrometry
(GC-MS, EI mode) also indicated that a mixture of isotopologues
was also formed with only 37% of the total product corresponding
to the dideutero product. (tricPDI) isolated following the catalytic
reaction was also a mixture of isotopologues and isotopomers as
2
judged by a combination of H NMR spectroscopy and mass spec-
trometry (LC-MS, ESI mode). Although dehydrogenation of the
aryl substituents was markedly slower than in the intermolecular
case, reversible cyclometallation still occurred at a similar rate for
both the intra- and intermolecular [2+2] cycloadditions, leading to
almost identical deuterium scrambling patterns on both the prod-
ucts and the recovered ligands.
Figure 14. Iron-catalyzed [2+2] cycloaddition of 1-deutero-E-
1-octene and techniques to determine number of isotopologues
and isotopomers. The mass spectrum histograms are representa-
tions of the originals, which can be found in the Supporting Infor-
mation.
To probe the relative rate of cyclometallation with
(
tricPDI)Fe(N2), 20 equivalents of 1-deutero-E-1-octene were
added to the iron precatalyst and the reaction monitored by 1H
NMR spectroscopy (Scheme 13). After 10 minutes at 23 °C, deu-
terium scrambling was observed at both the 1- and 2-positions of
35
the olefin, leading to a statistical mixture after 2 hours. While diffi-
1
cult to distinguish by H NMR spectroscopy, the presence of up to
d5-1,2-dihexylcyclobutane indicates that d2- and d3-1-octene are
likely also formed. This NMR experiment and the previous deuter-
ium labelling study indicated that cyclometallation occurred on the
same time-scale as productive [2+2] cycloaddition, and that it is
likely reversible, accounting for deuterium incorporation at multi-
ple positions of both the PDI ligand and the alkene. This also
demonstrates that tracking deuterium incorporation from starting
material to product cannot be used to deconvolute the mechanism
of C(sp3)–C(sp3) reductive elimination in intermolecular [2+2]
cycloadditions, as it is immediately complicated by ever-present
ligand-cyclometallation.
Figure 15. Iron-catalyzed [2+2] cycloaddition of 1,8-dideutero-
E,E-1,7-octadiene and techniques to determine number of isotopo-
logues and isotopomers. The mass spectrum histograms are repre-
sentations of the originals, which can be found in the Supporting
Information.
Reaction of other diene substrates was explored to evaluate
whether ligand cyclometallation could be suppressed by increasing
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 18
the rate of [2+2] cycloaddition. In the presence of 5 mol %
tricPDI)Fe(N2), the [2+2] cycloaddition of N,N-bis(3-deutero-
Comprehensive experimental and computational studies have
1
2
3
4
5
6
7
8
(
been conducted to gain insight into the mechanism of both inter-
and intramolecular iron-catalyzed [2+2] cycloadditions of alkenes
and to determine similarities and differences between the two pro-
cesses. For the iron-catalyzed [2+2] cycloaddition of 1-octene, a
reaction first order in both iron precatalyst and alkene but inverse
first order with respect to dinitrogen was observed. The coordina-
tively saturated iron alkene complex, (tricPDI)Fe(N2)(η2-1-
octene) was established as the catalyst resting state via a combina-
tion of infrared and freeze-quench 57Fe Mössbauer spectroscopy.
These observations support slow release and rate-determining
association of the second alkene substrate to a four-coordinate
Z-allyl)-4-fluoroaniline was complete within 2 hours at 23 °C, sig-
nificantly faster than the cycloaddition of 1,7-octadiene (Figure
16). The degree of deuterium scrambling was also significantly
reduced as only 19% of the total deuterium was incorporated at the
substituted position of the cyclobutane ring. A 90:10 mixture of
only two isotopomers was observed by 2H NMR spectroscopy,
significantly reduced from the complex mixture observed from the
cycloaddition products of 1-deutero-E-1-octene and 1,8-dideutero-
E,E-1,7-octadiene. As reported previously,4a the [2+2] cycloaddi-
tion of N,N-(3-deutero-Z-allyl)-tert-butylamine was complete in
less than 5 minutes at 23 °C and the product was obtained as a
single isotopomer and isotopologue. Due to the increased rate of
reaction, cyclometallation was likely too slow to compete with cy-
cloaddition and no scrambling was observed. Based on this data, it
is likely that C(sp3)–C(sp3) reductive elimination occurs either by
a concerted mechanism or a stepwise mechanism where the rate of
radical recombination is significantly faster than the rate of C–C
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
(
(
tricPDI)Fe alkene complex. Because catalyst turnover to either
tricPDI)Fe(N2) or (tricPDI)Fe(η2-alkene) is possible, it is
currently unknown whether the catalyst resting state is on- or off-
cycle. Competing cyclometalation of the aryl substituents of the
chelate resulted in catalyst deactivation and ultimately death, aris-
ing from dehydrogenation of a cyclopentyl group.
For the intramolecular [2+2] cycloaddition of 1,7-octadiene, cis-
bicyclo[4.2.0]octane was produced in >98% conversion and 92:8
d.r. A reaction first order in iron precatalyst and zeroth order in
α,ω-diene and dinitrogen was observed, distinct from the intermo-
36
bond rotation.
lecular
case.
The
iron
metallacycle,
(
dicPDI)Fe(bimetallacyclo[4.3.0]nonane) was found to be
the catalyst resting state and, in combination with 12C/13C KIE
measurements, underwent rate determining, unimolecular reduc-
tive elimination to form the bicyclic alkane product. Unlike the
intermolecular case, the catalytic cycle likely remains on the S = 1
spin surface, a consequence of the chelate effect and the higher
binding affinity of 1,7-octadiene compared to dinitrogen and 1-
octene. While the observed iron trans-bimetallacycle is lower in
energy, the transition state for reductive elimination of the cis-
metallacycle is lower in energy, leading to preferential formation of
the cis-bicycle. In both inter- and intramolecular reactions, compet-
ing cyclometalation of the aryl-substituents complicated stere-
ochemcial determination of the C–C bond-forming step. In cases
where the intramolecular catalytic reaction was sufficiently rapid,
deuterium labeling supported a concerted pathway for reductive
elimination or rapid radical rebound faster than C–C bond rota-
tion. Future work will focus on the application of these mechanistic
findings to next generation iron cycloaddition catalysts with im-
proved scope and activity.
Figure 16. Iron-catalyzed [2+2] cycloaddition of N,N-bis(3-
deutero-Z-allyl)-4-fluoroaniline and techniques to determine num-
ber of isotopologues and isotopomers. The mass spectrum histo-
grams are representations of the originals, which can be found in
the Supporting Information.
ASSOCIATED CONTENT
Computational studies on C–C bond forming processes in PDI
iron-catalyzed cycloaddition reactions have been reported previ-
The Supporting Information is available free of charge via the Internet
at http://pubs.acs.org and contains additional experimental details,
characterization data including NMR spectra, Mössbauer spectra, mass
spectrometry, and computational details and data.
37
ously by Chen and co-workers and we wish to address some of
their findings. Our experimental results on the symmetry allowed,
unimolecular reductive elimination of S = 1 PDI ferracyclopentanes
are in line with lowest energy reaction coordinate obtained by
Chen and co-workers for this pathway. The calculated transition
state energy for the reductive elimination process studied by Chen
and co-workers (24.4 kcal·mol-1) is also nearly identical to the ex-
perimentally determined transition state energy of 24.5 kcal·mol-1
for reductive elimination of the trans-bicycle from
Crystallographic
information
for
(
dicPDI)Fe(trans-
bimetallacyclo[4.3.0]nonane),
(
iPr,TBPDI)Fe(bimetallacyclo[4.3.0]nonane)
and (Et-
M ePDI)Fe(η2-allylbenzene)(N2) (CIF).
AUTHOR INFORMATION
(
dicPDI)Fe(trans-bimetallacyclo[4.3.0]nonane), further
corroborating both our group’s and Chen’s computational results
Corresponding Author
on [2+2] alkene dimerization.
* Email for P.J.C. pchirik@princeton.edu
ORCID
∎ CONCLUSIONS
Matthew V. Joannou: 0000-0002-0079-7107
ACS Paragon Plus Environment

Page 17 of 18
Journal of the American Chemical Society
Paul J. Chirik: 0000-0001-8473-2898
1
9
Stieber, S. C. E.; Milsmann, C.; Hoyt, J. M.; Turner, Z. R.; Finkelstein, K.
2
3
4
5
6
7
8
9
Notes
D.; Wieghardt, K.; DeBeer, S.; Chirik, P. J. Bis(Imino)Pyridine Iron Dini-
trogen Compounds Revisited: Differences in Electronic Structure Between
Four- and Five-Coordinate Derivatives. Inorg. Chem. 2012, 51, 3770–
3785.
The authors declare no competing financial interest.
ACKNOWLEDGMENT
We thank Firmenich SA for financial support.
10
For the calculation of Keq, it was assumed that the symmetric and anti-
symmetric N2 stretches of (tricPDI)Fe(N2)2 are of equal intensity as is
reported for (iPrPDI)Fe(N2)2 (ref. 8).
REFERENCES
11
(a) Singleton, D. A.; Thomas, A. A. High-Precision Simultaneous De-
termination of Multiple Small Kinetic Isotope Effects at Natural Abun-
dance. J. Am. Chem. Soc. 1995, 117, 9357–9358. (b) Frantz, D. E.; Single-
ton, D. A.; Snyder, J. P. 13C Kinetic Isotope Effects for the Addition of Lith-
ium Dibutylcuprate to Cyclohexenone. Reductive Elimination Is Rate-
Determining. J. Am. Chem. Soc. 1997, 119, 3383–3384.
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
1
(a) Namyslo, J. C.; Kaufmann, D. E. The Application of Cyclobutane
Derivatives in Organic Synthesis. Chem. Rev. 2003, 103, 1485–1538. (b)
The Chemistry of Cyclobutanes, Parts 1-2; Rappoport, Z., Ed.; Wiley &
Sons Ltd: Weinheim, Germany, 2005.
12
2
Kennedy, C. R.; Zhong, H.; Macaulay, R. L.; Chirik, P. J. Regio- and
(a) Wang, M.; Lu, P. Catalytic Approaches to Assemble Cyclobutane
Diastereoselective Iron-Catalyzed [4+4] Cycloaddition of 1,3-Dienes. J.
Am. Chem. Soc. 2019, 141, 8557-8573
Motifs in Natural Product Synthesis. Organic Chemistry Frontiers 2018,
5, 254–259. (b) Erman, M. B.; Kane, B. J. Chemistry Around Pinene and
Pinane: A Facile Synthesis of Cyclobutanes and Oxatricyclo-Derivative of
Pinane from Cis- and Trans-Pinanols. Chemistry & Biodiversity 2008, 5,
910–919. (c) Du, J.; Yoon, T. P. Crossed Intermolecular [2+2] Cycloaddi-
tions of Acyclic Enones via Visible Light Photocatalysis. J. Am. Chem. Soc.
2009, 131, 14604–14605. (d) Ischay, M. A.; Anzovino, M. E.; Du, J.;
Yoon, T. P. Efficient Visible Light Photocatalysis of [2+2] Enone Cycload-
ditions. J. Am. Chem. Soc. 2008, 130, 12886–12887. (e) Wang, Y.-M.;
Bruno, N. C.; Placeres, Á. L.; Zhu, S.; Buchwald, S. L. Enantioselective
Synthesis of Carbo- and Heterocycles through a CuH-Catalyzed Hydroal-
kylation Approach. J. Am. Chem. Soc. 2015, 137, 10524–10527. (f) Pop-
lata, S.; Tröster, A.; Zou, Y.-Q.; Bach, T. Recent Advances in the Synthesis
of Cyclobutanes by Olefin [2ꢀ+ꢀ2] Photocycloaddition Reactions. Chem.
Rev. 2016, 116, 9748–9815. (g) Ito, H.; Toyoda, T.; Sawamura, M. Stere-
ospecific Synthesis of Cyclobutylboronates through Copper(I)-Catalyzed
Reaction of Homoallylic Sulfonates and a Diboron Derivative. J. Am.
Chem. Soc. 2010, 132, 5990–5992. (h) Conner, M. L.; Brown, M. K.
Synthesis of 1,3-Substituted Cyclobutanes by Allenoate-Alkene [2 + 2]
Cycloaddition. J. Org. Chem. 2016, 81, 8050–8060.
13
If the reaction is mixed thoroughly before freezing to 77 K, a 50:50 mix-
ture of (tricPDI)Fe(N2) and (tricPDI)Fe(N2)(η2-1-octene) is ob-
served in the zero-field 57Fe Mössbauer spectrum. The precatalyst is not
completely soluble in 1-octene, which is the likely source of this observa-
tion. The other quadrupole doublet observed in Figure 5 accounts for only
8% of the total iron content of the sample. It is currently unknown whether
this signal corresponds to another iron species generated during the cy-
cloaddition reaction or a small amount of decomposition formed during
sample preparation.
14
Bart, S. C.; Lobkovsky, E.; Bill, E.; Wieghardt, K.; Chirik, P. J. Neutral-
Ligand Complexes of Bis(Imino)Pyridine Iron: Synthesis, Structure, and
Spectroscopy. Inorg. Chem. 2007, 46, 7055–7063.
15
Chirik, P. J. Preface: Forum on Redox-Active Ligands. Inorg. Chem.
2011, 50, 9737–9740.
16
Because it was difficult to obtain crystals suitable for X-ray diffraction of
(
tricPDI)Fe(η2-1-octene)(N2), other olefins and PDI ligand combina-
tions were tested to improve crystallinity of an olefin-N2 complex. Several
single crystals of (Et-M ePDI)Fe(η2-allylbenzene)(N2) were obtained
from [(Et-M ePDI)Fe(N2)]2(μ-N2) in a mixture of pentane and al-
lylbenzene in a -35 °C freezer. Large amounts of this compound could not
be isolated for thorough characterization, but the X-ray structure, which is
of high quality, lends support for assignment of the Mössbauer spectrum in
Figure 5 as well as starting points for DFT calculations (see Supporting
Information for a depiction of the solid-state structure).
3
Chirik, P. J. Carbon–Carbon Bond Formation in a Weak Ligand Field:
Leveraging Open-Shell First-Row Transition-Metal Catalysts. Angew.
Chem. Int. Ed. 2017, 56, 5170–5181.
4
(a) Bouwkamp, M. W.; Bowman, A. C.; Lobkovsky, E.; Chirik, P. J. Iron-
Catalyzed [2π + 2π] Cycloaddition of α,ω-Dienes: The Importance of
Redox-Active Supporting Ligands. J. Am. Chem. Soc. 2006, 128, 13340–
13341. (b) Schmidt, V. A.; Hoyt, J. M.; Margulieux, G. W.; Chirik, P. J.
Cobalt-Catalyzed [2π + 2π] Cycloadditions of Alkenes: Scope, Mechanism,
and Elucidation of Electronic Structure of Catalytic Intermediates. J. Am.
Chem. Soc. 2015, 137, 7903–7914. (c) Hoyt, J. M.; Schmidt, V. A.; Ton-
dreau, A. M.; Chirik, P. J. Iron-Catalyzed Intermolecular [2+2] Cycloaddi-
tions of Unactivated Alkenes. Science 2015, 349, 960–963.
17
Römelt, M.; Ye, S.; Neese, F. Calibration of Modern Density Functional
Theory Methods for the Prediction of 57Fe Mössbauer Isomer Shifts:
Meta-GGA and Double-Hybrid Functionals. Inorg. Chem. 2009, 48, 784–
785.
18
Trovitch, R. J.; Lobkovsky, E.; Chirik, P. J. Bis(Imino)Pyridine Iron
Alkyls Containing β-Hydrogens: Synthesis, Evaluation of Kinetic Stability,
and Decomposition Pathways Involving Chelate Participation. J. Am.
Chem. Soc. 2008, 130, 11631–11640.
(d) Russell, S. K.; Lobkovsky, E.; Chirik, P. J. Iron-Catalyzed Intermolecu-
lar [2π + 2π] Cycloaddition. J. Am. Chem. Soc. 2011, 133, 8858–8861.
19
DFT simulated 57Fe Mössbauer parameters for both regioisomers of
5
Hoyt, J. M.; Sylvester, K. T.; Semproni, S. P.; Chirik, P. J. Synthesis and
(dehydro-tricPDI)Fe(N2) matched those obtained experimentally and
are presented in the Supporting Information. Additionally, while small
amounts of the cyclobutane product are still being formed after 8 hours of
reaction, the diminished signal for the catalyst resting state is likely ob-
scured by the signal for (dehydro-tricPDI)Fe(N2).
Electronic Structure of Bis(Imino)Pyridine Iron Metallacyclic Intermedi-
ates in Iron-Catalyzed Cyclization Reactions. J. Am. Chem. Soc. 2013,
135, 4862–4877.
6
(a) Scott, D. R.; Allison, J. B. Solvent Glasses for Low Temperature Spec-
troscopic Studies. The Journal of Physical Chemistry 1962, 66, 561–562.
(b) Greenspan, H.; Fischer, E. Viscosity of Glass-Forming Solvent Mix-
tures at Low Temperatures. The Journal of Physical Chemistry 1965, 69,
2466–2469.
20
P. Gütlich, E. Bill, A.X. Trautwein, Mössbauer Spectroscopy and Transi-
tion Metal Chemistry, Springer-Verlag, Berlin Heidelberg, 2011.
21
Optimization of (dicPDI)Fe(η2-propylene)2 in a BS(3,1), S = 1 configura-
7
tion led to dissociation of one imine arm of the PDI ligand. This has been
previously observed both experimentally and computationally in reference
N.N. Greenwood, T.C. Gibb, Mössbauer Spectroscopy, Springer Nether-
lands, Dordrecht, 1971.
8
5.
Bart, S. C.; Lobkovsky, E.; Chirik, P. J. Preparation and Molecular and
22
Low spatial orbital overlap of this type has been calculated previously in
Electronic Structures of Iron(0) Dinitrogen and Silane Complexes and
Their Application to Catalytic Hydrogenation and Hydrosilation. J. Am.
Chem. Soc. 2004, 126, 13794–13807.
PDI iron complexes, especially those with broken symmetry electronic
configurations such as (PDI)Fe(N2). See reference 9.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 18
1
2
3
4
5
6
7
8
23 (a) Hopmann, K. H. How Accurate is DFT for Iridium-Mediated Chem-
istry? Organometallics 2016, 35, 3795-3807. (b) Wodrich, M.D.; Cormin-
boeuf, C.; Schreiner, P.R.; Fokin, A.A.; von Ragué Schleyer, P. How Accu-
rate are DFT Treatments of Organic Energies. Org. Lett. 2007, 9, 1851-
1854.
24
The Fe-catalyzed [2+2] cycloaddition of 1,7-octadiene conducted at 4
atm N2 reaches 50% conversion in 2.5 hours and complete conversion in 9
hours (identical to the 0 and 1 atm experiments), which further supports
that the reaction is zeroth order in [N2].
9
25
(a) Inorganic and Organometallic Reaction Mechanisms. Atwood, J. D.
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
Wiley-VCH Inc., 1997. (b) Blackmond, D. G. Reaction Progress Kinetic
Analysis: A Powerful Methodology for Mechanistic Studies of Complex
Catalytic Reactions. Angew. Chem. Int. Ed. 2005, 44, 4302–4320.
26 Small quantities of 2,6-dicyclopentylaniline were initially purchased from
Sigma-Aldrich but it has since been discontinued. tricPDI is more readily
accessible due to its straightforward synthesis beginning with Friedel-Crafts
tri-alkylation of benzene with cyclopentyl bromide.
27
Bart, S. C.; Chłopek, K.; Bill, E.; Bouwkamp, M. W.; Lobkovsky, E.;
Neese, F.; Wieghardt, K.; Chirik, P. J. Electronic Structure of
Bis(Imino)Pyridine Iron Dichloride, Monochloride, and Neutral Ligand
Complexes: A Combined Structural, Spectroscopic, and Computational
Study. J. Am. Chem. Soc. 2006, 128, 13901–13912.
28
Grubbs, R. H.; Miyashita, A. Preparation and Isomerization Reactions of
2-Nickelahydrindanes. Journal of Organometallic Chemistry 1978, 161,
371–380.
29
Solution NMR of Paramagnetic Molecules. Bertini, I.; Luchinat, C.;
Parigi, G., Ed.; Elsevier Science: Amsterdam, The Netherlands, 2001.
30
Because the exact process accounting for the isomerization from trans- to
cis-bimetallcycle is unknown (and likely involves several distinct steps),
either a primary or inverse 13C KIE could be observed at all cyclobutane
carbons of the product. Although, the magnitude is likely to be small as the
effect will be averaged over 4 cyclobutane carbons.
31
Modern Physical Organic Chemistry. Anslyn, E. V.; Dougherty, D. A.
University Science Books, 2007.
32 Advanced Organic Chemistry, Part A: Structure and Mechanisms. Carey,
F. A.; Sundberg, R. J. Springer Science and Business Media, LLC, 2007.
33
The cis-product is also 5 kcal·mol-1 lower in energy than the trans. While
the relative stability of the products does not dictate the distribution of
products in a Curtin-Hammett scenario, it is still worthwhile to mention.
This calculation also accurately matches the energy difference between a
twist-boat and boat cyclohexane conformation (4.5 kcal·mol-1) (cf. refer-
ence 27)
34
Lau, W.; Huffman, J. C.; Kochi, J. K. Electrochemical Oxidation-
Reduction of Organometallic Complexes. Effect of the Oxidation State on
the Pathways for Reductive Elimination of Dialkyliron Complexes. Organ-
ometallics 1982, 1, 155–169.
35
After 18 hours at 23 °C, small amounts of isomerized 1-octene were also
observed.
36
The barrier for C–C bond rotation about ethane has been calculated to
be 2.9 kcal·mol-1, while that for ethyl radical is 0.8 kcal·mol-1. This suggests
that the rate of radical recombination in the stepwise mechanism could be
>1010 s-1, which would likely be kinetically indistinguishable from a concert-
ed reductive elimination mechanism. (a) Pacansky, J.; Coufal, H. On the
Barrier for Rotation about the C–C Bond of the Ethyl Radical. The Journal
of Chemical Physics 1980, 72, 5285–5286. (b) Edidin, M. A.; Ratto, T. V.;
Longo, M. L.; Sturtevant, J. M.; Sandhoff, K.; Simons, K.; Vogel, H.; Nel-
son, W. J. Ultrafast Carbon-Carbon Single-Bond Rotational Isomerization
in Room-Temperature Solution. Science 2006, 313, 1951-1954.
37
Hu, L.; Chen, H. Substrate-Dependent Two-State Reactivity in Iron-
Catalyzed Alkene [2+2] Cycloaddition Reactions. J. Am. Chem. Soc.
2017, 139, 15564–15567.
For Table of Contents Use
ACS Paragon Plus Environment