Dirhodium catalysts that bear redox noninnocent chelating dicarboxylate ligands and their performance in intra-and intermolecular C-H amination

Article abstract of DOI:10.1002/ejic.201100814

We report two new analogues of the well-known C-H amination catalyst [Rh₂(esp)₂] (1) (esp = α,α,α′, α′-tetramethyl-1,3-benzenedipropanoate) that bear redox-active supporting ligands that are structurally similar to esp. The redox-active ligands are 2-[3-(1-carboxy-1-methylethoxy)phenoxy]-2-methylpropanoic acid (H₂L1) and (3-methoxycarbonyl-2,5-di-tert-butylphenoxy)ethanoic acid (H₂L2), which react with Rh₂(OAc)₄ to form the catalysts [Rh₂(L1)₂] (2) and [Rh₂(L2) ₂] (3). Both 2 and 3 have been characterized by X-ray crystallography and cyclic voltammetry, inter alia. Compounds 2 and 3 are structurally similar to 1 but show more complex electrochemical features. Whereas 1 has a single reversible redox wave that corresponds to the Rh₂^II,II/ Rh₂^II,III couple, 2 and 3 show multiple oxidations that are characteristic of ligand-centered oxidation. Catalysts 1, 2, and 3 perform well in a model intramolecular C-H amination reaction, and all three catalysts perform equally well during the first four hours of a model intermolecular reaction. After this point, 2 and 3 cease to function, whereas 1 continues to be active. These results support the hypothesis that intermolecular C-H amination utilizes two distinct mechanisms: (1) a nitrene interception/insertion mechanism that is fast but ceases to be operative after four hours, and (2) a one-electron mechanism that is more robust over extended time periods, but requires the catalyst to be able to undergo Rh₂-centered oxidation. Copyright

Full text of DOI:10.1002/ejic.201100814

FULL PAPER
DOI: 10.1002/ejic.201100814
Dirhodium Catalysts That Bear Redox Noninnocent Chelating Dicarboxylate
Ligands and Their Performance in Intra- and Intermolecular C–H Amination
Katherine P. Kornecki[a] and John F. Berry*[a]
Keywords: Rhodium / Homogeneous catalysis / Amination / Redox chemistry / Chelates
We report two new analogues of the well-known C–H amin-
ation catalyst [Rh2(esp)2] (1) (esp = α,α,αЈ,αЈ-tetramethyl-1,3-
benzenedipropanoate) that bear redox-active supporting li-
gands that are structurally similar to esp. The redox-active
ligands are 2-[3-(1-carboxy-1-methylethoxy)phenoxy]-2-
methylpropanoic acid (H2L1) and (3-methoxycarbonyl-2,5-
di-tert-butylphenoxy)ethanoic acid (H2L2), which react with
Rh2(OAc)4 to form the catalysts [Rh2(L1)2] (2) and [Rh2(L2)2]
(3). Both 2 and 3 have been characterized by X-ray crystal-
lography and cyclic voltammetry, inter alia. Compounds 2
and 3 are structurally similar to 1 but show more complex
and 3 show multiple oxidations that are characteristic of li-
gand-centered oxidation. Catalysts 1, 2, and 3 perform well
in a model intramolecular C–H amination reaction, and all
three catalysts perform equally well during the first four
hours of a model intermolecular reaction. After this point, 2
and 3 cease to function, whereas 1 continues to be active.
These results support the hypothesis that intermolecular C–
H amination utilizes two distinct mechanisms: (1) a nitrene
interception/insertion mechanism that is fast but ceases to be
operative after four hours, and (2) a one-electron mechanism
that is more robust over extended time periods, but requires
the catalyst to be able to undergo Rh2-centered oxidation.
electrochemical features. Whereas 1 has a single reversible
II,III
redox wave that corresponds to the Rh2II,II/Rh2
couple, 2
Introduction
Catalytic C–H amination by means of nitrene transfer
mediated by dirhodium paddlewheel compounds has be-
come an important synthetic tool.[1] This reaction is out-
lined in Scheme 1, and involves oxidative transformation of
a nitrogen-containing substrate (typically a sulfamate or
carbamate ester) into a nitrene equivalent, with subsequent
insertion of this nitrene into a substrate C–H bond. The
nitrene insertion can either be intramolecular, thereby yield-
ing cyclized products, or intermolecular. Whereas intramo-
lecular reactions are well established, it has been shown that
not every Rh2 catalyst can efficiently accomplish intermo-
lecular reactions.[2] [Rh2(esp)2] (1; esp = α,α,αЈ,αЈ-tetra-
methyl-1,3-benzenedipropanoate) has been shown to be the
best catalyst for the intermolecular amination transforma-
tion,[3] a fact that has been attributed to the added stability
of the catalyst due to the chelate effect of the bridging di-
carboxylate ligands. Current mechanistic information on in-
tramolecular C–H amination by 1 is consistent with the
mechanism shown on the left side of Scheme 1 whereby an
iminoiodinane intermediate [formed in the rate-limiting re-
action of sulfamate ester with PhI(OAc)2] transfers the ni-
[a] Department of Chemistry,
University of Wisconsin, Madison
1101 University Avenue, Madison, WI 53706, USA
Fax: +1-608-262-6143
Scheme 1. Proposed catalytic cycle for dirhodium-catalyzed C–H
amination: the inner mechanism involves direct nitrene transfer,
and the outer mechanism (with amido intermediate B) corresponds
to the one-electron pathway that dominates the intermolecular
amination mechanism.
E-mail: berry@chem.wisc.edu
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201100814 or from
the author.
562
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 562–568

Dirhodium Catalysts with Dicarboxylate Ligands
trene to the dirhodium catalyst, thereby yielding the dirho-
dium nitrene intermediate A.[2] Intermediate A is then able
to insert the nitrene directly into substrate C–H bonds to
reform the original catalyst. For intermolecular C–H amin-
ation reactions, the iminoiodinane transfer mechanism ap-
pears to account for only around 30% of the product for-
mation. The majority of the product has been proposed to
form by means of the outer mechanism in Scheme 1, which
involves single-electron redox steps.[4] The formation of C–
H amination products by using the one-electron oxidant
Ce4+ verifies the one-electron nature of this mechanism.[4]
Redox noninnocent ligands that are strongly coupled to
a transition-metal center have been employed with success
in catalytic reactions that involve redox transformations.[5]
To the best of our knowledge, redox noninnocence has
never been studied in the context of dirhodium complexes,
or the possible implications to catalysis. Ferrocene-based re-
dox auxiliaries in dirhodium complexes have been reported,
but only their electrochemistry has been studied.[6] Because
of the one-electron transformations involved in the outer
mechanism for intermolecular C–H amination catalyzed by
Rh2 complexes, we have prepared Rh2 catalysts that bear
redox noninnocent ligands and describe their performance
in intra- and intermolecular amination in this article.
Results and Discussion
Synthesis
Both H2L1 and H2L2 were previously reported by
Bonar-Law and co-workers,[7] and were prepared similarly
herein with slight modifications to the synthesis of H2L2
(outlined in the Exp. Section). Treating the dicarboxylate
ligands with dirhodium tetraacetate at 150 °C in dichloro-
benzene affords the loss of acetic acid and the formation of
complexes 1–3 in good yields (60–75%). Complex forma-
1
tion and purity were established by H NMR spectroscopy,
MALDI-MS, and elemental analysis. Crystals suitable for
X-ray diffraction were obtained for complexes 2 and 3 as
their diaquo and bis-acetone adducts, respectively (see Fig-
ures 1 and 2, vide infra). The axial ligands were removed
under vacuum prior to the use of these compounds as cata-
lysts.
Since the chelate effect is touted as the reason for the
increased stability and selectivity of 1 as an intermolecular
amination catalyst, we decided to study catalysts that are
close structural analogues of 1. The ligands L1 {2-[3-(1-
carboxy-1-methylethoxy)phenoxy]-2-methylpropanoate}
and L2 [(3-methoxycarbonyl-2,5-di-tert-butylphenoxy)eth-
anoate] (previously reported in the context of metallomac-
rocycle assembly)[7] have the same backbone chain-length
as esp; however, the backbones of L1 and L2 contain a
redox-active resorcinol-derived component (Scheme 2).
Like the more well-known 1,2- or 1,4-dialkoxy-substituted
benzenes,[8] 1,3-dialkoxybenzenes may be oxidized to the
corresponding radical cations, as shown in Scheme 2.[9] The
meta-disubstituted radical species are significantly less
stable than their ortho or para congeners.[9,10] To the best of
our knowledge, the meta-dialkoxybenzene motif has not yet
been investigated as a redox-noninnocent ligand in coordi-
nation complexes, which further prompted this study.
Figure 1. Thermal ellipsoid plot of catalyst 2·2H2O, with thermal
ellipsoids drawn at the 30% probability level. Hydrogen atoms are
omitted for clarity.
Figure 2. Thermal ellipsoid plot for 3·2acetone, with thermal ellip-
soids drawn at the 30% probability level. Hydrogen atoms are omit-
ted for clarity.[12]
Crystallography
Crystal data for 2 and 3 are given in Table 1. The solid-
state structures of 1, 2, and 3 are similar and show chelation
of the dicarboxylate ligands to the metal–metal bonded di-
rhodium core. The Rh–Rh distances in all three catalysts
range from 2.3817(9) to 2.3910(6) Å, which are effectively
the same (see Table 2) and fall within the normal range for
II,II
bond lengths in simpler Rh2
carboxylate complexes.[11]
Scheme 2. Chelating dicarboxylate ligands and the redox capabili-
ties of resorcinol-based compounds.
The Rh–O bond lengths to the carboxylate ligands and ax-
Eur. J. Inorg. Chem. 2012, 562–568
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
563

K. P. Kornecki, J. F. Berry
FULL PAPER
ial ligands are all normal. An interesting difference between dialkoxybenzene unit. The irreversibility of the oxidations
the three structures presents itself in a comparison of the is consistent with a fast decomposition process that occurs
O
bound carboxylate–C–C–Ophenolate torsion angles. There are after the radical cation is generated.
two of these torsion angles in 1 that are crystallographically The majority of dirhodium carboxylate paddlewheel
inequivalent; on one side of the bound carboxylate it is complexes exhibit one reversible redox event that corre-
II,III
57.301°, and it is 59.238° on the other, which are very close sponds to the Rh2II,II/Rh2
couple.[13] [Rh2(esp)2] (1) ex-
values and may not be considered to be significantly dif- hibits a perfectly reversible Rh-centered redox couple at
ferent in a chemical sense. This torsion angle in 3 is signifi- E1/2 = 0.82 V (versus Fc/Fc+ in dichloromethane). Chemical
cantly smaller at only 18.79°. Interestingly, 2 differs from 1 oxidation and spectroelectrochemistry results indicate that
and 3 in that its two crystallographically inequivalent small- this wave is centered at the Rh2 unit and is assigned as the
est torsion angles on either side of the bound dicarboxylate Rh2II,II/Rh2II,III redox couple.[4] Compounds 2 and 3 feature
are very different: 99.341 and 37.912°. If this difference in more complexity in their cyclic voltammograms than does
torsion angles is retained in solution, the nearly 90° torsion 1. In contrast to the single, reversible wave observed for 1,
angle could play a role in facilitating hyperconjugation ef- compound 2 displays multiple irreversible features at
fects between the oxygen atom of the meta-dialkoxybenzene Ͼ1.2 V. The irreversibility of these features and their higher
backbone and the carboxylate π orbitals, thus allowing for redox potential than 1 suggest that these oxidation events
II,III
direct electronic communication to the rhodium center.
have a different origin than the Rh2II,II/Rh2
couple of
1. Since these redox events closely mirror the irreversible
electrochemical behavior of the H2L1 ligand, it is reason-
able to assign these as oxidations of the meta-dialkoxyphen-
ylene moiety of the ligand backbone. Much like in the case
of the free ligand itself, the radical cationic species that re-
sult from these oxidations are unstable and undergo a fast
chemical reaction, likely radical/radical coupling, that ren-
ders the electrochemical signal irreversible. Like 2, com-
pound 3 shows multiple redox waves in its cyclic voltammo-
gram. Unlike the irreversible behavior of 2, the redox waves
of 3 are reversible and appear at lower potentials. The CV
of 3 shows a reversible two-electron wave at 0.97 V, and a
further reversible one-electron wave at 1.3 V.
Table 1. Crystallographic parameters for 2 and 3.
Compound
2·2H2O
3·2acetone
¯
P1
Space group
I4/m
Crystal system
a [Å]
b [Å]
c [Å]
α [°]
β [°]
triclinic
tetragonal
19.3634(5)
19.3634(5)
12.1982(4)
90
90
90
12.520(3)
12.637(3)
14.577(4)
91.372(4)
100.240(4)
101.962(4)
2215.8(10)
2
γ [°]
V [Å3]
4573.69(2)
4
Z
R1, wR2 [(I) Ͼ 2σ(I)]
R1, wR2 (all data)
0.0398, 0.0911
0.0510, 0.0995
0.0302, 0.0736
0.0446, 0.0793
It is well established that phenoxyl radicals and phenoxyl
radical complexes can be stabilized when bulky substituents
protect the ortho and para positions of the aryl ring.[14] In
a similar manner, addition of two tBu substituents to the
aryl ring of the ligands of 2 so as to form 3 leads to a
greater stabilization of the corresponding radical cation
[L2]·+. This stabilization is both a thermodynamic effect (re-
flected in the more accessible oxidation potential of 3, 0.9 V,
than 2, Ͼ1.2 V) and a kinetic effect (the reversibility of the
redox waves for 3 indicates that the chemical process re-
sponsible for rendering the 2 waves irreversible is now
Table 2. Selected bond lengths and angles in 1, 2, and 3.
Compound
1·2acetone[16]
2·2H2O
3·2acetone
Rh–Rh [Å]
Rh–Ocarboxylate [Å]
Rh–Oaxial [Å]
2.3817(9)
2.0386(18)
2.3042(19)
2.3873(6)
2.0355(2)
2.298(7)
2.3910(6)
2.036(2)
2.286(4)
O–C–C–O (torsion) [°] 57.301, 59.238 37.912, 99.341 18.79
Electrochemistry
To test the redox activity of ligands L1 and L2, cyclic
voltammetric measurements were performed (Figures S1
and S2 in the Supporting Information). The H2esp ligand,
which lacks the meta-dialkoxybenzene moiety, shows no re-
dox behavior in THF up to 2.1 V versus Fc/Fc+. Under
reducing conditions, an irreversible wave at –1.4 V is ob-
served. Similar reductive waves are observed in THF for
H2L1 and H2L2, and it is thus reasonable to conclude that
these waves correspond to electrochemical reduction and
subsequent decarboxylation of the carboxylic acid function-
alities. In addition to these reductive events, H2L1 and
H2L2 show waves at positive potentials that correspond to
oxidations that are absent in H2esp. For H2L1, multiple
irreversible oxidation events are observed at potentials
Ͼ1.2 V. In contrast, two discrete irreversible waves are ob-
served for H2L2 at 0.9 and 1.1 V. The oxidations in H2L1
and H2L2 are safely assigned to the oxidation of the meta- Figure 3. Redox features of chelate catalysts 1, 2, and 3.
564 Eur. J. Inorg. Chem. 2012, 562–568
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Dirhodium Catalysts with Dicarboxylate Ligands
slower than the scan time of the cyclic voltammogram) complement the report from Du Bois that [Rh2(S-biTISP)2]
(Figure 3). It is not possible to tell if one of the multiple {S-biTISP
= 1,3-[N,NЈ-bis(2,4,6-triisopropylbenzenesulf-
redox waves displayed by 2 or 3 is centered at the Rh2 unit. onyl)-(2S,2ЈS),(5R,5ЈR)-prolinate]benzene}, despite having
However, in comparing the electrochemical data of 1, 2, chelating dicarboxylate ligands, is a poor intermolecular C–
II,III
and 3 we may note that any Rh2II,II/Rh2
oxidations of H amination catalyst.[2] Due to the structural analogy be-
2 or 3 occur at substantially higher potential than in 1, and tween 1, 2, and 3, the poor catalytic performance of the
that such an oxidation for 2 is not reversible. Given the latter two can be attributed to electronic rather than steric
general structural similarities between 1, 2, and 3, it is not properties. Another important piece of information about
II,III
obvious why the Rh2II,II/Rh2
couple would be less ac- these reactions is that the brilliant red color ascribed to
cessible in the latter two compounds, but this is certainly one-electron oxidized Rh2II,III species appears when catalyst
what happens.
1 is used, but is absent for reactions that involve 3, and
appears but is short-lived for reactions that involve 2.
Catalysis
As mentioned above, catalytic amination of C–H bonds
is one of the primary applications of dirhodium complexes
in which their redox chemistry is proposed to play an im-
portant role. Thus the catalytic activity of 1, 2, and 3 in
intra- and intermolecular C–H amination reactions is of
interest. The reaction shown in Scheme 3 has been a useful
test reaction for the performance of dirhodium catalysts in
intramolecular C–H aminations that target the benzylic C–
H position.[15] Simple Rh2 carboxylates such as [Rh2(oct)4]
(oct = octanoate) perform well as catalysts for this reaction
with reported yields of approximately 84%.[15] Thus, chelat-
ing dicarboxylate ligands such as esp are not absolutely nec-
essary for this transformation. Catalysts 1, 2, and 3 never-
theless perform intramolecular amination very well, par-
ticularly 1 and 3. It is not entirely obvious why 2 gives a
lower yield; however, catalyst 2 is not recoverable after the
course of the reaction, thereby implying that radical reac-
tions at the ligand may initiate catalyst degradation.
Scheme 4. Intermolecular reactivity of chelate catalysts. Percent
conversion based on H NMR spectroscopic integration versus an
internal standard.
1
The catalytic results presented here can be rationalized
in terms of the mechanism shown in Scheme 1 and provide
further evidence to support this mechanism. Mechanistic
data[15] on intramolecular C–H amination reactions are
consistent with nitrene transfer from an iminoiodinane to
the dirhodium catalyst, thereby forming the dirhodium ni-
trene complex A, which in turn is extremely electrophilic[16]
and inserts the nitrene into the substrate C–H bond. This
nitrene interception/insertion mechanism does not require
any change in the oxidation state of the Rh2 core, as long
as A is considered to be an adduct of a neutral nitrene to
II,II
an Rh2
carboxylate core.[17]
At a 2 mol-% catalyst loading in the intramolecular cycli-
zation reaction (Scheme 3), catalysts 1, 2, and 3 all seem to
perform equally well and afford high yields of the product
(Ͼ75%).
Magnesium oxide is an important additive that can affect
the outcome of the reaction.[16] Table 3 gives a comparison
of the yields of intramolecular C–H amination reactions
Scheme 3. Intramolecular reactivity of chelate catalysts. Percent
conversion based on 1H NMR spectroscopic integration versus
concentration of uncyclized substrate.
To test the catalytic competence of 1–3 under intermo- catalyzed by 1 mol-% of 1, 2, and 3 (listed as turnover num-
lecular C–H amination conditions, conversion of ethyl ben- bers) for the first 24 h of the reaction in the presence or
zene to the corresponding amination product by using absence of MgO. In general, catalyst performance is en-
H2NTces (Tces = trichloroethylsulfamate) and PhI(OAc)2 hanced by this additive. We may expect that the function of
was used as a representative amination reaction (Scheme 4). MgO is to neutralize the acetic acid reaction byproduct,
Ethylbenzene is a particularly appropriate substrate since it and that excessive amounts of acetic acid may be an impor-
presents benzylic C–H bonds similar to those in the intra- tant cause of catalyst arrest, particularly for 2 and 3. In-
molecular substrate described above. At four hours under deed, when no MgO is present, the initial green color of the
these conditions, the three catalysts show comparable ac- dirhodium catalysts 2 and 3 quickly dissipates. Instead, the
tivity. However after this short timespan, 2 and 3 cease to color of the reaction mixture becomes yellow, reminiscent
function, whereas 1 continues to perform until a significant of mononuclear RhIII species. This is not the case when 1
portion of the ethyl benzene is consumed. These results is used as a catalyst, thereby indicating that some facet of
Eur. J. Inorg. Chem. 2012, 562–568
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
565

K. P. Kornecki, J. F. Berry
FULL PAPER
its structure makes it more robust under the reaction condi- meta-dialkoxybenzene ligand fragment. Neither 2 nor 3 is
tions. In fact, when 1 equiv. of 1 is treated with 2 equiv. active in C–H amination beyond the initial four-hour
each of a sulfamate ester and PhI(OAc)2 in dichlorometh- period dominated by the nitrene interception/insertion
ane, crystalline 1·2HOAc (see Figure S4 in the Supporting mechanism. There are two possible explanations for this ob-
Information) can be isolated after several days, which indi- servation that are consistent with the working mechanistic
cates that the catalyst is unchanged by acidic conditions. hypothesis. First, it is possible that the one-electron oxi-
Acetic acid is clearly more detrimental to catalysts 2 and 3, dation of 2 or 3 occurs at such a high potential that these
possibly implicating a protonation of the resorcinol-derived catalysts cannot be oxidized to B under the reaction condi-
O atoms and subsequent destruction of the ligand, which is tions. The other possibility is that 2 or 3 can be oxidized
not feasible when 1 is used. Notwithstanding, when MgO under the reaction conditions, but that this oxidation is cen-
is present, the reaction mixture will maintain its initial tered on the chelating dicarboxylate ligand rather than on
green color for 2 and 3, thus allowing all three catalysts to the Rh2 unit. If the ligand oxidation is strongly coupled to
remain active until complete product formation is achieved, the metal center, we may anticipate that B could still be
which occurs at around 20 h for catalyst 1, and after ap- formed under these conditions. However, there is no π-con-
proximately 50 h for catalysts 2 and 3.
jugation pathway between the meta-dialkoxybenzene group
and the Rh2 center and a hyperconjugation pathway is
doubtful, so the electronic coupling here is anticipated to
Table 3. Turnover numbers (TON) for the intramolecular cycliza-
tion depicted in Scheme 3 using 1 mol-% catalyst loading after
24 h.
be weak, thereby resulting in a one-electron oxidized species
II,II
that resembles an uncoupled Rh2
unit appended to an
Catalyst
TON (3 equiv. MgO)
TON (MgO-free)
organic radical cation. Such a species would not be chemi-
cally equivalent to B, and therefore may not take part in
the one-electron C–H amination mechanism. There is some
indication that both of these possibilities are, in fact, occur-
ring. When 2 is used as a catalyst, the red color characteris-
tic of B appears briefly at the beginning of the reaction,
1
2
3
99
91
56
60
86[a]
69[a]
[a] Ͼ 90 after about 50 h.
We[4] and Du Bois[2,3] have noted that the nature of inter- then disappears. This observation is consistent with the first
molecular C–H amination catalyzed by 1 is different at the possibility outlined above in which initially the concentra-
beginning of the reaction than during the later stages of the tion of oxidant is high enough that B can be produced,
reaction. Our current hypothesis for this behavior is that but as the concentration of the oxidant wanes, the reaction
intermolecular C–H amination initially occurs by means of mixture no longer has sufficient oxidizing power to utilize
the interception/insertion mechanism described above, but the one-electron mechanism. In contrast, when 3 is used as
after conversion of roughly 30% of the substrate this a catalyst, the red color of B is never observed. Thus, if 3
mechanism becomes inactive and the reaction continues by is oxidized under these conditions, then the oxidation must
following a slower one-electron mechanism. This mecha- be ligand-centered and there must be little electronic cou-
nism, which dominates product formation late in the reac- pling between the meta-dialkoxybenzene ligand fragment
tion, involves oxidation of 1 in the presence of sulfamate and the Rh2 center. We interpret these results to mean that
II,III
ester substrate to yield a one-electron oxidized Rh2
-
Rh2 centered oxidation is necessary for the success of the
amido-type species B. Species B has a brilliant red color one-electron intermolecular mechanism, and thereby for
and is observable as an intermediate in the reaction. It un- complete conversion of substrate.
dergoes a further one-electron oxidation to yield the nitrene
intermediate A, which inserts the nitrene into the substrate
C–H bond.
Conclusion
The hypothesis that the nitrene interception/insertion
mechanism and the one-electron mechanism operate con-
For the first time, we have introduced redox noninnocent
temporaneously is supported by results we report here. ligands as supporting ligands for metal–metal bonded com-
Since the nitrene interception/insertion mechanism does not pounds specifically to investigate their catalytic behavior.
involve Rh2-centered oxidation, we expect catalysts 1, 2, The meta-dialkoxybenzene motif in ligands L1 and L2 has
and 3 to perform equally well in the first stage of the cata- been shown to engender the corresponding dirhodium com-
lytic reaction, just as the three catalysts all perform well in plexes 2 and 3 with complex ligand-centered redox proper-
intramolecular C–H amination. Indeed, this is what hap- ties that are absent from the structurally analogous 1. By
pens. In the first four hours of the reaction, catalysts 1, 2, assessing the differences in performance for catalysts 1–3, it
and 3 convert around 30% of the ethylbenzene to product. is evident that rhodium-centered oxidation is essential to
The one-electron mechanism for intermolecular C–H amin- the performance of dirhodium carboxylate catalysts in
ation requires the catalyst to be oxidized to form B. Both 2 intermolecular C–H amination. Catalyst 1, which has a re-
and 3 are more difficult to oxidize than 1 by at least versible rhodium-centered redox couple, can access inter-
100 mV. Moreover, one-electron oxidation of 2 or 3 does mediate B, an Rh2II,III-amido species; catalysts 2 and 3 are
not necessarily involve oxidation of the Rh2 center from more difficult to oxidize and are not successful in this one-
Rh2II,II to Rh2II,III, but may instead involve oxidation of the electron mechanistic regime.
566
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 562–568

Dirhodium Catalysts with Dicarboxylate Ligands
Compound 1: Characterization data matched values previously re-
Experimental Section
ported by Du Bois and co-workers.[18]
General: All reagents were obtained commercially unless otherwise
noted. Reactions were performed using oven-dried glassware under
an atmosphere of nitrogen, either in a glove box or using Schlenk
techniques. Dichloromethane was dried with CaH2 and distilled
before use. [D6]Benzene was dried on an activated alumina column
prior to use. All other solvents were collected anhydrous from a
Vacuum Atmospheres solvent system. The structures of known
compounds were confirmed by 1H NMR spectroscopy and ESI-
1
Compound 2: H NMR (CDCl3 + 3 vol.-% [D6]acetone): δ = 7.033
(t, 2 H), 6.464 (dd, 4 H), 6.108 (t, 2 H), 1.358 (s, 24 H) ppm.
MALDI-MS: m/z calcd. 766.000; found 765.910. IR: ν = 2978.91,
˜
2962.29, 1727.29, 1596.67, 1201.36, 961.30 cm–1. Elemental analysis
(2 was dried at 100 °C under vacuum to afford the complex with
no axial ligation): calcd. C 43.88, H 4.21, N 0.00; found C 43.83,
H 4.55, N 0.09. Crystals suitable for X-ray diffraction were ob-
tained by dissolving the purified material in hot dichlorobenzene
and allowing the solution to cool slowly (around 70 h).
1
MS. H NMR spectra were collected on a 300 MHz Bruker spec-
trometer at room temperature.
Compound 3: 1H NMR ([D6]acetone + 3 vol.-% CDCl3): δ = 7.137
(s, 2 H), 5.202 (s, 2 H), 4.614 (s, 8 H), 1.343 (s, 36 H) ppm. MALDI-
Electrochemistry: All electrochemistry experiments were conducted
under a nitrogen atmosphere in solutions (10 mL, 0.1 m) of tetra-
butylammonium hexafluorophosphate in freshly distilled dichloro-
methane with a 0.001 m analyte concentration. The reference elec-
trode consisted of a silver wire immersed in a 10 mm silver nitrate
solution contained by a Vycor tip. The auxiliary electrode was a
platinum wire. For cyclic voltammetry, data was referenced to the
ferrocene/ferrocenium redox couple, and the working electrode was
made of glassy carbon.
MS: m/z = calcd. 878.126; found 877.980. IR: ν = 2958.67, 2914.94,
˜
2873.92, 1592.91, 1414.18, 919.83 cm–1. Elemental analysis (3 was
dried at 100 °C under vacuum to afford the complex with no axial
ligation); calcd. C 49.21, H 5.51, N 0.00; found C 49.28, H 5.79,
N 0.13. Crystals suitable for X-ray diffraction were obtained by
dissolving the purified material in a 1:1 mixture of acetone and
dichloromethane followed by slow evaporation of the solvents.
Compound 1·2HOAc: Crystals suitable for X-ray diffraction were
isolated from the treatment of 1 (30 mg, 0.039 mmol) with trichlo-
roethylsulfamate (18 mg, 0.078 mmol) and PhI(OAc)2 (25 mg,
0.078 mmol) in dichloromethane (5 mL). The red reaction mixture
was layered with hexanes. After several days, blue-green crystals
were harvested.
Ligands: H2esp was prepared according to the synthesis described
by Du Bois and co-workers.[18] H2L1 was prepared according to
the synthesis reported by Bonar-Law and co-workers.[7] H2L2 was
prepared similarly to the synthesis described by Bonar-Law and co-
workers.[7] The differences are outlined below.
Synthesis of H2L2: Et2L2: A 100 mL Schlenk flask was charged
with potassium carbonate (1.6 g, 11.6 mmol) and 4,6-di-tert-butyl-
resorcinol (1.04 g, 4.7 mmol). Acetonitrile (50 mL) was added into
the reaction, followed by the addition of α-iodoethyl acetate
(1.1 mL, 9.3 mmol). The reaction was stirred at room temperature
for 40 h. Solvent was removed in vacuo, and the residue was dis-
solved in dichloromethane. Particulates were filtered from the
dichloromethane solution, which was subsequently concentrated.
Purification was achieved by means of column chromatography on
silica gel with gradient elution from 5–20% ethyl acetate in hexanes
to yield a clear crystalline solid (0.9 g, 24%), the ethyl ester, Et2L2.
1H NMR (CDCl3): δ = 7.229 (s, 1 H), 6.199 (s, 1 H), 4.576 (s, 4
H), 4.281 (q, 4 H), 1.387 (s, 18 H), 1.315 (t, 6 H) ppm. ESI/EMM:
m/z calcd. 417.2248 [M + Na]+; found 417.2243. H2L2: NaOH
(7 mL, 1 m), ethanol (5 mL), and acetone (2 mL) were added into
CCDC-837373 (for 2), -837374 (for 3), and -846682 (for 1·2HOAc)
contain the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif.
Amination Reactions: Each of the three catalysts was tested in a
prototypical intermolecular C–H amination reaction, as well as an
intramolecular cyclization. For the intramolecular reaction, a
prototypical reaction using the substrate depicted in Scheme 3 was
performed in dichloromethane with the addition of 1.1 equiv. of
hypervalent iodine oxidant, as reported in the literature.[15] The in-
tramolecular reaction was typically complete after 4 h, as moni-
1
tored by TLC. TONs were measured internally by H NMR spec-
troscopy using a 1 mol-% catalyst loading over a 24 h period and
by conducting the reaction described above in CD2Cl2. Product
concentrations were determined against an internal cyclooctane
standard.
a
25 mL round-bottomed flask charged with diester (0.4 g,
1.0 mmol). The flask was equipped with a reflux condenser and
heated to 70 °C for 12 h. Any remaining ethanol was removed by
rotary evaporation. A solution of 1.0 m HCl was added to the re-
maining reaction mixture, thereby resulting in a creamy white pre-
cipitate that was extracted into ethyl acetate (50 mL three times).
Intermolecular reactions were performed in deuterated solvent and
product formation was monitored over the course of 24 h in the
presence of two equivalents of hypervalent iodine oxidant, as re-
The organic layer was dried with sodium sulfate, filtered, and con-
centrated. An off-white solid was isolated (0.31 g, 96%). H NMR ported previously.[4] Final product conversion was done on the ba-
1
1
([D6]acetone): δ = 7.212 (s, 1 H), 6.563 (s, 1 H), 4.730 (s, 4 H),
1.392 (s, 18 H) ppm. 13C NMR ([D6]acetone): δ = 29.79, 34.32,
65.11, 99.24, 124.88, 129.78, 155.57, 169.55 ppm. ESI/EMM: m/z
calcd. 337.1656 [M – H]–; found 337.1657.
sis of H NMR spectroscopic integration.
Supporting Information (see footnote on the first page of this arti-
cle): Figures S1 and S2 show the cyclic voltammograms for H2L1
and H2L2; and Figures S3 and S4 show the crystal structures of 3
and 1·2HOAc.
Synthesis of Dirhodium Chelate Complexes 1, 2, and 3:
Dirhodium tetraacetate·2CH3OH (100 mg, 0.197 mmol, 1 equiv.)
and chelate ligand (2.5 equiv.) were added to an Erlenmeyer flask
with anhydrous dichlorobenzene (30 mL). The flask was heated to
150 °C for 4 h, then allowed to cool completely. The solvent was
removed by rotary evaporation, and the resulting green residue was
subjected to chromatography on silica gel, gradient elution with
acetone in dichloromethane, 0–25%. The complexes were afforded
in 60–75% yields as microcrystalline green/green-blue solids.
Acknowledgments
We thank the Chemical Sciences, Geosciences, and Biosciences Di-
vision, Office of Basic Energy Sciences, Office of Science, U.S. De-
partment of Energy for support (grant number DE-FG02-
10ER16204).
Eur. J. Inorg. Chem. 2012, 562–568
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
567

K. P. Kornecki, J. F. Berry
FULL PAPER
[10] F. Marchetti, G. Pampaloni, C. Pinzino, J. Organomet. Chem.
2011, 696, 1294–1300.
[1]
a) D. N. Zatalan, J. Du Bois, Top. Curr. Chem. 2010, 292, 347–
378; b) H. M. L. Davies, J. R. Manning, Nature 2008, 451, 417–
424; c) H. M. L. Davies, M. S. Long, Angew. Chem. 2005, 117,
3584–3586; Angew. Chem. Int. Ed. 2005, 44, 3518–3520.
K. W. Fiori, J. Du Bois, J. Am. Chem. Soc. 2007, 129, 562–568.
D. N. Zatalan, J. Du Bois, J. Am. Chem. Soc. 2009, 131, 7558–
7559.
[11] H. T. Chifotides, K. R. Dunbar, in: Multiple Bonds Between
Metal Atoms (Eds.: F. A. Cotton, C. A. Murillo, R. A. Wal-
ton), Springer, New York, 2005, pp. 465–567.
[12] The structure of 3 crystallizes with both acetone and dichloro-
methane as axial ligands with disorder 84% acetone/16%
dichloromethane. The CH2Cl2 ligand was omitted from the
thermal ellipsoid plot for clarity, and is shown in Figure S3 in
the Supporting Information.
[2]
[3]
[4]
K. P. Kornecki, J. F. Berry, Chem. Eur. J. 2011, 17, 5827–5832.
[5] a) P. J. Chirik, K. Wieghardt, Science 2010, 327, 794–795; b)
[13] K. Das, K. M. Kadish, J. L. Bear, Inorg. Chem. 1978, 17, 930–
W. Kaim, Inorg. Chem. 2011, 50, 9752–9765.
934.
[6] F. Estevan, P. Lahuerta, J. Latorre, E. Peris, S. Garcia-Granda,
F. Gomez-Beltran, A. Aguirre, M. A. Salvado, J. Chem. Soc.,
Dalton Trans. 1993, 1681–1688.
[7] J. Bickley, R. Bonar-Law, T. McGrath, N. Singh, A. Steiner,
New J. Chem. 2004, 28, 425–433.
[8] J. K. Kochi, Angew. Chem. 1988, 100, 1331; Angew. Chem. Int.
Ed. Engl. 1988, 27, 1227–1266, and references therein.
[9] a) M. Jonsson, J. Lind, T. Reitberger, T. E. Eriksen, J. Phys.
Chem. 1993, 97, 11278–11282; b) J. Pavlinac, M. Zupan, S.
Stavber, Org. Biomol. Chem. 2007, 5, 699–707; c) J. Pavlinac,
M. Zupan, S. Stavber, J. Org. Chem. 2006, 71, 1027–1032.
[14] P. Chaudhuri, K. Wieghardt, Prog. Inorg. Chem. 2001, 50, 151–
216.
[15] C. G. Espino, P. M. Wehn, J. Chow, J. Du Bois, J. Am. Chem.
Soc. 2001, 123, 635–6936.
[16] K. W. Fiori, C. G. Espino, B. H. Brodsy, J. Du Bois, Tetrahe-
dron 2009, 65, 3042–3051.
[17] J. F. Berry, Dalton Trans. 2011, DOI: 10.1039/C1DT11434D.
[18] C. G. Espino, K. W. Fiori, M. Kim, J. Du Bois, J. Am. Chem.
Soc. 2004, 126, 15378–15379.
Received: August 1, 2011
Published Online: November 17, 2011
568
www.eurjic.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2012, 562–568