Highly Stable Molecular Borromean Rings

Article abstract of DOI:10.1002/cjoc.201700590

A series of Cp*Rh-based molecular Borromean rings (BRs) are prepared from naphthazarine or metallaligand. Some of as-synthesized BRs display high stability and are formed in high yields in solution. The reason is related to the length ratio of the long-arm linker and short-arm linker, where smaller aspect ratios of the metallarectangles promote improved stability and yields of the BRs in solution. Increasing the width of the metallaligand or pyridyl ligand hinders the formation of BRs and leads to unoccupied monomeric rectangles, which were further used as catalysts for the acyl transfer reaction between N-acetylimidazole and (4-(pyridin-4-yl)phenyl) methanol.

Full text of DOI:10.1002/cjoc.201700590

Highly Stable Molecular Borromean Rings
Ye Lu, a Yue-Jian Lin, a Zhen-Hua Li a and Guo-Xin Jin*,a
a State Key Laboratory of Molecular Engineering of Polymers, Collaborative Innovation Center of Chemistry for
Energy Materials, Department of Chemistry, Fudan University, Shanghai, 200433, P. R. China
A series of Cp*Rh-based molecular Borromean rings (BRs) are prepared from naphthazarine or metallaligand.
Some of as-synthesized BRs display high stability and are formed in high yields in solution. The reason is related to
the length ratio of the long-arm linker and short-arm linker, where smaller aspect ratios of the metallarectangles
promote improved stability and yields of the BRs in solution. Increasing the width of the metallaligand or pyridyl
ligand hinders the formation of BRs and leads to unoccupied monomeric rectangles, which were further used as cat-
alysts for the acyl transfer reaction between N-acetylimidazole and (4-(pyridin-4-yl)phenyl) methanol.
Keywords Borromean rings, Metallarectangles, Self-assembly, Supramolecular chemistry, Density functional cal-
culations
Introduction
the BRs structure, what kind of special properties could
be engineered into the BRs by altering this aspect ratio?
In fact, our original intention for using Cu(Ⅱ)
metallaligand was to construct discrete metallarectan-
gles with two coordinatively-unsaturated metal ions for
synergistic catalysis.12 Compared with metal-organic
frameworks, the unique advantage of metallarectangles
is that the distance between two catalytic centers can be
conveniently adjusted according to the size of the sub-
strate molecule.13 However, by increasing the distance
between the two open Cu(Ⅱ) ions, BRs were formed,
hampering the catalysis for larger substrates. Thus, for
this application, the size of metallarectangles should be
retained to hinder the formation of BRs. Though guest
or solvent molecules could force the BRs to convert to
the corresponding monomeric rectangle (MR) assembly,
these guest or solvent molecules would bind inside the
metallarectangles and impede further catalysis.9b
Therefore, we have to explore other methods to achieve
unoccupied MR assemblies, in order to maintain the
distance between two open Cu(Ⅱ) ions.
Herein, we reported a series of Cp*Rh-based BRs
prepared by pyridyl ligand and various binuclear pre-
cursor. Those precursor is bridged by naphthazarine,
metallaligand [Pd(opba)]2– [opba = o-phenylenebis
(oxamato)] or metallaligand [Cu(nabo)]2- [nabo =
2,3-naphthalenebis(oxamato)]. Some of these BRs dis-
play high stability and are formed in high yields in solu-
tion. The reason is perhaps related to the length ratio of
the long-arm linker and short-arm linker, where smaller
aspect ratios of the metallarectangles could promote
improved stability and yields of the BRs in solution.
In the past two decades, a wide range of topologi-
cally fascinating interlocked molecules have prepared
by coordination-driven self-assembly,1 such as Borro-
mean rings,2 Solomon links,3 star of David catenanes,4
pentafoil knots,5 and others.6 One intriguing and chal-
lenging synthetic target in this field is the family of mo-
lecular Borromean rings (BRs), which consist of three
chemically independent rings that are locked in such a
way that no two of the three rings are linked with each
other.2a, 7 The first BRs structure prepared by “all-in-one”
synthetic strategies was reported by Stoddart and
coworker in 2004, and was based on metal ion tem-
plates.8 After that, our group presented a template-free
self-assembly method for synthesizing BRs based on
Cp*Rh (Cp* = pentamethylcyclopentadienyl) and
Cu(Ⅱ) metallaligand in 2013.9 Recently, other BRs
based on template-free self-assembly are synthesized by
our group and Chi’s group, constructed by
half-sandwich metal units (Rh, Ir or Ru).10 Moreover,
we further established a stepwise separation method for
p-dihalobenzenes based on BRs.11 As research contin-
ues, a number of new problems attracted our interest.
For example, the BRs built by dihalogenated or te-
tracene-based bridge ligand showed modest stability and
yield in solution, which is reduced further by dilution,
solvent or guest molecules,10a, 11 thus, how can we pre-
pare the BRs with high stability and high yield in solu-
tion? What is the relationship between stability and
structure in BRs? Meanwhile, we established earlier that
the formation of BRs is related to the length of long-arm
linker and short-arm linker.9b However, after forming
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which
may lead to differences between this version and the Version of Record. Please cite this
article as doi: 10.1002/cjoc.201700590
This article is protected by copyright. All rights reserved.

Meanwhile, increasing the width of the metallaligand or
pyridyl ligand hindered the formation of those highly
stable BRs and led to unoccupied MRs. Thus, these
MRs were further used in catalysis on larger substrates,
for example, the acyl-transfer reaction between
N-acetylimidazole and (4-(pyridin-4-yl)phenyl) metha-
nol.
4,4'-bipyridylacetylene (L2) (40.8 mg, 0.2 mmol) in-
stead of 4,4'-bipyridine (L1). 4a was obtained as a yel-
low solid in a 92.3% yield (246.1 mg); 1H NMR (400
MHz, CD3OD ppm): δ = 8.80 (d, J=5.6 Hz, 12H, Py-H),
8.70 (d, J=5.6 Hz, 12H, Py-H), 8.33 (d, J=8.0 Hz, 6H,
obpa-H), 8.07 (d, J=8.0 Hz, 6H, obpa-H), 7.76 (d, J=5.6
Hz, 12H, Py-H), 6.98 (d, J=5.6 Hz, 12H, Py-H), 6.78 (t,
6H, obpa-H), 6.59 (t, 6H, obpa-H), 1.79 (s, 90H, Cp*),
1.74
(s,
90H,
Cp*);
Anal.
calcd
for
Experimental
C276H252F36N24O72Pd6Rh12S12: C 41.44, H 3.18, N 4.20,
found: C 41.43, H 3.15, N 4.18; ESI-MS m/z: [4a-BRs
− 3OTf]3+ calcd. 2517.24, found 2517.24; [4a-BRs −
4OTf]4+ calcd. 1850.69, found 1850.69; IR (KBr disk,
cm−1) v = 1606, 1581, 1471, 1422, 1383, 1342, 1261,
1225, 1202, 1161, 1031, 834, 759, 639, 587, 518, 502,
461.
Preparation of 4b-BRs. The synthesis of 4b-BRs
was carried out similarly to that of 4a-BRs with the use
of Na2[Cu(opba)] (71.4 mg, 0.2 mmol) instead of
K2[Pd(opba)]. 4b was obtained as a green solid in a 90.8%
Preparation of Ⅱ-BRs. A CH3OH solution of
[Cp*RhCl2]2 (124 mg, 0.2 mmol) was added to a solu-
tion of naphthazarine (38 mg, 0.2 mmol) and NaOH (16
mg, 0.4 mmol) in CH3OH (40 mL), and the suspension
was stirred at room temperature for 6 h. AgOTf (102.8
mg, 0.4 mmol) was added to the mixture and stirred for
3 h, followed by filtration to remove insoluble com-
pounds (AgCl and NaCl). 4,4'-Bipyridylacetylene (40.8
mg, 0.2 mmol) was then added to the filtrate. After the
solution was stirred at room temperature for 12 h, the
reaction mixture was concentrated to a volume of 3 mL
under reduced pressure, filtered through Celite and re-
crystallized by slow diffusion of diethyl ether into the
filtrate. A green crystalline solid was obtained in 88.6%
yield (206.7 mg); 1H NMR (400 MHz, CD3OD, ppm): δ
= 9.04 (d, J =5.6 Hz, 24H, Py-H), 7.30 (d, J =5.6 Hz,
24H, Py-H), 6.70 (s, 24H, naphthazarine-H), 1.69 (s,
180H, Cp*); Anal. Calcd for C276H252F36N12O60Rh12S12:
C 47.35, H 3.63, N 2.40, found: C 47.31, H 3.64, N 2.43;
ESI-MS m/z: [Ⅱ-BRs − 3OTf]3+ calcd. 2184.11, found
2184.11; [Ⅱ-BRs − 4OTf]4+ calcd. 1600.84, found
1600.84; IR (KBr disk, cm−1) v = 1604, 1533, 1488,
1416, 1270, 1224, 1157, 1031, 963, 854, 836, 638, 548,
518, 445.
Preparation of 3a. A CH3OH solution of
[Cp*RhCl2]2 (124 mg, 0.2 mmol) was added to a solu-
tion of K2[Pd(opba)] (86.4 mg, 0.2 mmol) in CH3OH
(40 mL), and the suspension was stirred at room tem-
perature for 1 h. AgOTf (102.8 mg, 0.4 mmol) was then
added to the mixture, and this was stirred for 6 h fol-
lowed by filtration to remove insoluble compounds
(AgCl and NaCl). 4,4'-bipyridine (L1) (31.2 mg, 0.2
mmol) was then added to the filtrate. After the solution
was stirred at room temperature for 12 h, the reaction
mixture was concentrated to a volume of 3 mL under
reduced pressure, filtered through Celite and recrystal-
lized by slow diffusion of diethyl ether into the filtrate.
An orange crystalline solid was obtained in 91.6% yield
(235.4 mg); 1H NMR (400 MHz, CD3OD and d6-DMSO,
ppm): δ = 8.70 (d, J=6.4 Hz, 8H, Py-H), 8.21 (d, J=6.4
Hz, 8H, Py-H), 8.18 (d, J=6.0 Hz, 4H, obpa-H), 7.27 (t,
4H, obpa-H), 1.72 (s, 60H, Cp*); Anal. calcd for
C84H84F12N8O24Pd2Rh4S4: C 39.25, H 3.29, N 4.36,
found: C 39.20, H 3.27, N 4.33; IR (KBr disk, cm−1) v
=1610, 1581, 1470, 1420, 1383, 1343, 1278, 1224, 1160,
1072, 1031, 819, 759, 638, 587, 517, 462.
yield
(234.3
mg);
Anal.
calcd
for
C276H252F36N24O72Cu6Rh12S12: C 42.82, H 3.28, N 4.34,
found: C 42.81, H 3.25, N 4.31; IR (KBr disk, cm−1) v =
1608, 1576, 1461, 1412, 1384, 1348, 1261, 1225, 1202,
1168, 1035, 837, 751, 639, 587, 517, 502, 462.
Preparation of 5a-BRs. The synthesis of 5a was
carried out similarly to that of 3a with the use of
1,4-di(pyridin-4-yl)benzene (L3) (46.4 mg, 0.2 mmol)
instead of 4,4'-bipyridine (L1). 5a-BRs was obtained as
1
a yellow solid in a 93.8% yield (255.4 mg); H NMR
(400 MHz, CD3OD, ppm): δ = 9.37 (d, J=5.6 Hz, 6H,
Py-H), 9.31 (d, J=5.6 Hz, 6H, Py-H), δ = 9.02 (d, J=5.6
Hz, 6H, Py-H), 8.87 (d, J=5.6 Hz, 6H, Py-H), 8.56 (d,
J=8.0 Hz, 6H, Py-H), 8.21 (d, J=5.6 Hz, 6H, Py-H),
8.09 (d, J=8.0 Hz, 6H, obpa-H), 7.82 (d, J=8.0 Hz, 6H,
obpa-H), 7.01 (d, J=5.6 Hz, 6H, Py-H), 6.83 (br, 6H,
phenyl-H), 6.82 (d, J=8.0 Hz, 6H, Py-H), 6.81 (t, 6H,
obpa-H), 6.54 (t, 6H, obpa-H), 6.41 (br, 6H, phenyl-H),
5.22 (br, 6H, phenyl-H), 5.15 (br, 6H, phenyl-H), 1.88 (s,
90H, Cp*), 1.73 (s, 90H, Cp*); Anal. calcd for
C288H276F36N24O72Pd6Rh12S12: C 42.35, H 3.41, N 4.12,
found: C 42.36, H 3.39, N 4.15; ESI-MS m/z: [5a-BRs
− 5OTf]5+ calcd. 1484.81, found 1484.80; [5a-BRs −
6OTf]6+ calcd. 1212.01, found 1212.03; IR (KBr disk,
cm−1) v =1609, 1581, 1421, 1258, 1224, 1156, 1030,
817, 756, 638, 586, 518, 460.
Preparation of 6b. The synthesis of 6b was carried
out similarly to that of 4b-BRs with the use of
1,4-bis(4-pyridyl)naphthalene (L4) (56.4 mg, 0.2 mmol)
instead of 4,4'-bipyridylacetylene (L2). 6b was obtained
as a green solid in a 90.5% yield (247.7 mg); Anal.
calcd for C104H96Cu2F12N8O24Rh4S4: C 45.64, H 3.54, N
4.09, found: C 45.66, H 3.53, N 4.10; ESI-MS m/z: [6b
− 2OTf]2+ calcd. 1219.55, found 1219.57; [6a-BRs −
4OTf]4+ calcd. 1850.69, found 1850.69; IR (KBr disk,
cm−1) v = 1617, 1587, 1471, 1419, 1278, 1224, 1161,
1031, 638, 573, 517, 463.
Preparation of 4a-BRs. The synthesis of 4a-BRs
was carried out similarly to that of 3a with the use of
Preparation of 7a-BRs. The synthesis of 7a-BRs
This article is protected by copyright. All rights reserved.

was carried out similarly to that of 3a with the use of
N,N'-di-4-pyridinyloxalamide (L5) (48.4 mg, 0.2 mmol)
instead of 4,4'-bipyridine (L1). 7a-BRs was obtained as
a yellow solid in a 91.1% yield (249.8 mg); Anal. calcd
for C264H264F36N36O84Pd6Rh12S12: C 38.54, H 3.23, N
6.13, found: C 38.55, H 3.25, N 6.14; ESI-MS m/z:
[7a-BRs − 3OTf]3+ calcd. 2593.26, found 2593.26;
[7a-BRs − 4OTf]4+ calcd. 1907.71, found 1907.70; IR
(KBr disk, cm−1) v = 1712, 1611, 1582, 1504, 1485,
1425, 1259, 1161, 1031, 639, 574, 518, 460.
Preparation of 8a-BRs. The synthesis of 8a-BRs
was carried out similarly to that of 3a with the use of
N,N'-bis(4-pyridyl)terephthalamide (L6) (63.6 mg, 0.2
mmol) instead of 4,4'-bipyridine (L1). 8a-BRs was ob-
tained as a yellow solid in a 88.9% yield (257.3 mg);
Anal. calcd for C300H288F36N36O84Pd6Rh12S12: C 41.49,
H 3.34, N 5.81, found: C 41.47, H 3.35, N 5.80; IR
(KBr disk, cm−1) v =1661, 1610, 1580, 1515, 1420,
1280, 1256, 1163, 1031, 639, 518, 460.
Preparation of 9b. The synthesis of 9b was carried
out similarly to that of 4b-BRs with the use of
N,N-bis(4-pyridyl)-1,4,5,8-naphthalenetetracarboxydiim
ide (L7) (84.0 mg, 0.2 mmol) instead of
4,4'-bipyridylacetylene (L2). 9b was obtained as a green
solid in a 92.1% yield (277.5 mg); ESI-MS m/z: [9b −
3OTf]3+ calcd. 855.03, found 855.03; Anal. calcd for
C112H92Cu2F12N12O32Rh4S4: C 44.65, H 3.08, N 5.58,
found: C 44.66, H 3.09, N 5.56; IR (KBr disk, cm−1) v =
1721, 1681, 1618, 1584, 1419, 1347, 1250, 1160, 1030,
766, 638, 517.
(264.4 mg); ESI-MS m/z: [11b-BRs − 6OTf]6+ calcd.
1269.40, found 1269.41; Anal. calcd for
C336H300Cu6F36N24O72Rh12S12: C 47.42, H 3.55, N 3.95,
found: C 47.40, H 3.58, N 3.91; IR (KBr disk, cm−1) v
=1617, 1589, 1473, 1422, 1384, 1268, 1224, 1160, 1031,
638, 573, 518, 463.
Crystallographic Details. Crystallographic data for
complexes 3a, Ⅱ-BRs, 5a-BRs, 7a-BRs were collected
at 203 K, 150 K or 173 K using a CCD-Bruker APEX
DUO system (Mo Kα, λ = 0.71073 Å). Those of 4a-BRs,
4b-BRs, 6b, 8a-BRs, 9b, 11b-BRs were collected at
173 K or 150 K using a Bruker D8 VENTURE micro-
focus X-ray source system (Cu Kα, λ = 1.54178 Å).
Indexing was performed using APEX 2 (difference vec-
tors method). Data integration and reduction were per-
formed using SaintPlus 6.01. Absorption correction was
performed by the multiscan method implemented in
SADABS. The structures were solved and refined using
SHELXTL-97. The single-crystal X-ray diffraction data
of 3a, Ⅱ-BRs, 4a-BRs, 4b-BRs, 5a-BRs, 6b, 7a-BRs,
8a-BRs, 9b and 11b-BRs have been deposited in the
Cambridge Crystallographic Data Centre under acces-
sion number CCDC: 1537126 (3a), 1537136 (4a-BRs),
1552228 (Ⅱ-BRs), 1537137 (4b-BRs), 1537141
(5a-BRs), 1537143 (6b), 1537142 (7a-BRs), 1537144
(8a-BRs), 1537139 (9b) and 1537140 (11b-BRs).
Results and discussion
The relationship between stability and structure
in BRs. Our research group has a longstanding interest
in the self-assembly of two preorganized binuclear
half-sandwich metal (Rh or Ir) molecular clips and
pyridyl ligands, which led us to attempt similar reac-
tions with Pd(Ⅱ) metallaligand in order to avoid the
paramagnetic Cu(II) nuclei used in previous studies.7, 14
Thereby, a solution of dirhodium precursor 1a based on
Preparation of 10a. The synthesis of 10a was car-
ried out similarly to that of 5a-BRs with the use of
K2[Pd(nabo)] (96.0 mg, 0.2 mmol) instead of
K2[Pd(opba)]. 10a was obtained as a yellow solid in a
1
89.5% yield (252.6 mg); H NMR (400 MHz, CD3OD
and d6-DMSO, ppm): δ = 8.64 (s, 4H, nabo-H), δ = 8.62
(d, J=5.4 Hz, 8H, Py-H), 8.00 (d, J=5.6 Hz, 8H, Py-H),
7.97 (m, 4H, obpa-H), δ = 7.95 (d, J=2.8 Hz, 8H, phe-
nyl-H), 7.58 (t, 4H, obpa-H), 1.76 (s, 60H, Cp*); Anal.
calcd for C104H96F12N8O24Pd2Rh4S4: C 44.25, H 3.43, N
3.97, found: C 44.23, H 3.45, N 3.96; IR (KBr disk,
cm−1) v =1607, 1583, 1484, 1460, 1423, 1262, 1030,
1009, 756, 638, 517, 474.
[Pd(opba)]2– (1.0 equiv) reacts with L1 (L1
=
4,4'-bipyridine) (1.0 equiv), leading to exclusive for-
mation of a tetranuclear metallarectangle 3a (Scheme 1).
Single-crystal X-ray crystallographic analysis confirmed
the structure of 3a to be a discrete MR (Figure S46) and
analyzed by NMR spectroscopy (Figure S1-S4).
The formation of BRs based on [Cu(opba)]2– is re-
lated to the length of pyridyl ligand.9a Hence, we at-
tempted to use the longer (and linear) bipyridyl ligand
L2 (L2 = 4,4'-bipyridylacetylene) to build a metal-
larectangle (4a-BRs, Scheme 1 and Figure 1) with pre-
cursor 1a. Gratifyingly, orange block crystals of 4a-BRs
were isolated in a 92.3% yield upon crystallization by
diffusion of diethyl ether into a methanol solution (in-
cluding several drops of DMSO). The single-crystal
X-ray crystallographic analysis established 4a-BRs to
have a discrete BRs structure (Figure 2a and Figure
S47a), which is further checked by ESI-MS (Figure
S60-S61). In the structure of 4a-BRs, the metallaligand
unit [Pd(opba)]2–, displays a planar structure and the
distance from the carbon atoms of the alkynyl group to
Preparation of 10b. The synthesis of 10b was car-
ried out similarly to that of 10a with the use of
Na2[Cu(nabo)] (81.0 mg, 0.2 mmol) instead of
K2[Pd(nabo)]. 10b was obtained as a green solid in a
89.1% yield (243.7 mg); ESI-MS m/z: [10b − 3OTf]3+
calcd. 763.05, found 763.05; Anal. calcd for
C104H96Cu2F12N8O24Rh4S4: C 45.64, H 3.54, N 4.09,
found: C 45.66, H 3.55, N 4.08; IR (KBr disk, cm−1) v
=1608, 1590, 1456, 1422, 1384, 1258, 1224, 1160, 1031,
638, 518, 475.
Preparation of 11b-BRs. The synthesis of 11b-BRs
was carried out similarly to that of 10b with the use of
2,6-di(pyridin-4-yl)naphthalene (L8) (56.4 mg, 0.2
mmol) instead of 1,4-bis(4-pyridyl)-naphthalene (L4).
11b-BRs was obtained as a green solid in a 93.2% yield
This article is protected by copyright. All rights reserved.

the [Pd(opba)]2- plane is ca. 3.5 Å, which is accordance
with the conventional distance of - interactions. It is
worth noting that, for 4a-BRs, the length of the long
arm minus the length of short arm is less than 7 Å,
which is double the conventional distance of - inter-
actions. Hence, in order to achieve the optimal distance
for - interactions, the long arm of 4a-BRs had to
curve inward. As is shown in Figure 2c, the short arm of
4a-BRs is ca. 10.8 Å, while the distance between two
corresponding alkynyl groups is only ca. 9.5 Å. A simi-
lar situation was also observed for 4b-BRs (Figure
S47-S49), which was built using [Cu(opba)]2– and L2
(Scheme 1).
Scheme 1. Synthesis of metallarectangles and interconversion between BRs and MR
In order to discover the relationship between the as- chloranilic acid and bromianilic acid, and then, the length
pect ratio and properties, we attempt to keep the long-arm of their short-arm linker is only ca. 7.9 Å (Figure S51).11
linker and shorten the short-arm linker. A shorter dirho-
Three kinds of BRs, 4a-BRs, Ⅱ-BRs and the BRs
dium precursor Ⅰbased on naphthazarine is used to con- based on dihalogenated ligands, have the same long-arm
struct metallarectangles with L2, and named Ⅱ-BRs linker and different short-arm linker (Table S12). At this
(Figure 1). Single-crystal X-ray crystallographic analysis point, we considered the possibility that different proper-
and ESI-MS confirmed the BRs structure of Ⅱ-BRs ties could be caused by the dissimilarity between the
(Figure 2b, S50, S62 and S63). The distance from the structures. The BRs based on dihalogenated ligands
carbon atoms of the alkynyl group to the naphthazarine showed modest stability and yield in methanol solution,
plane is ca. 3.4 Å, which is also accordance with the which is reduced further by dilution. However, Ⅱ-BRs
1
conventional distance of - interactions. In contrast to showed very different behavior. The H NMR spectra of
4a-BRs, for Ⅱ-BRs, the length of the long arm minus the Ⅱ-BRs in methanol solution displayed one groups of
length of short arm is more than 7 Å, so that the long arm metallarectangle signals (Figure S5), which is checked by
1
of Ⅱ-BRs had to curve outward to achieve the optimal 13C NMR, H-1H ROESY NMR and DOSY NMR (Fig-
distance for - interactions, the short arm of Ⅱ-BRs is ure S6-S8). Moreover, with the dilution, no new signals
ca. 8.3 Å, while the distance between two corresponding were found, even if the concentration have diluted to
alkynyl groups is ca. 9.3 Å (Figure 2d). Actually, BRs 0.5mM (Figure S9). That means there are no balance
could be obtained by further shortening the short-arm between BRs and corresponding MR in methanol solu-
linker. Very recently, we reported the BRs based on di- tion of Ⅱ-BRs and Ⅱ-BRs present stronger stability than
halogenated ligands, which is also constructed by L2, but the BRs based on dihalogenated ligand in solution.11
their dirhodium precursor is bridged by fluoranilic acid, Then, upon increasing the ratio of d6-DMSO in CD3OD,
This article is protected by copyright. All rights reserved.

a clear signal attributable to the corresponding MR, Ⅱ, the three MRs of 5a-BRs (–217.3 kcal/mol) is much
was observed (Figure S10-S11), and further confirmed by lower than that of 4a-BRs (–147.4 kcal/mol) (Table S1,
1
13C NMR, H-1H ROESY NMR and DOSY NMR (Fig- Figure S74-S75). These results further underscore our
ure S12-S14).15 When the CD3OD:d6-DMSO ratio is 5:4, contention that metallarectangles with smaller aspect ra-
Ⅱ-BRs was found to be completely transformed to the tios could help to increase the stability of BRs structures.
corresponding MR Ⅱ(Figure S10).
BRs also could be built by the interactions between
Compared with Ⅱ-BRs and the BRs based on dihalo- Cu(Ⅱ) centers of the metallaligand and carbonyl groups
genated ligands, 4a-BRs have the same long-arm linker of the pyridyl ligand. In contrast to Cu(Ⅱ), Pd(Ⅱ) centers
and longest short-arm linker. The behavior of 4a-BRs in only have four coordination sites and their axial coordi-
1
solution is also different with the former two. The H nation ability is very weak. We were also interested in
NMR spectra of 4a-BRs in methanol displayed two testing whether BRs could be formed by the interaction
groups of metallarectangle signals in a 1:1 ratio (Figure between tetracoordinate Pd(Ⅱ) centers and carbonyl
1
S15), not one group, checked by 13C NMR, and H-1H groups. More importantly, what effect does the aspect
ROESY NMR (Figure S16-S18). Furthermore, the ratio ratios of metallarectangle have on the stability of this
of the two groups of signals remained unchanged under kind of BRs? To address those problems, two kinds of
different concentrations (Figure S19), and two groups of pyridyl ligand with carbonyl group, L5 and L6 (L5 =
signals were confirmed to belong to one complex due to N,N'-di-4-pyridinyloxalamide
and
L6
=
their identical diffusion coefficients observed via DOSY N,N'-bis(4-pyridyl)terephthalamide), were used to con-
NMR (Figure S20). The reason of two group signal struct metallarectangle with precursor 1a, respectively,
maybe is due to the smaller aspect ratio and higher sym- resulting in the formation of two new compounds 7a-BRs
metry cannot be achieved in this situation. Meanwhile, (based on L5) and 8a-BRs (based on L6) (Scheme 1). The
with the decrease of aspect ratios, 4a-BRs display higher single-crystal X-ray crystallographic analysis indicated
stability than Ⅱ-BRs, when the CD3OD:d6-DMSO ratio 7a-BRs and 8a-BRs to have a discrete BRs structure
is 5:4, there is no new signal found in 4a-BRs (Figure (Figure 3, Figure S54, Figure S55 and Table S12). The
S19), 4a-BRs is immune to DMSO. This higher stability BRs structure of 7a-BRs was further confirmed by
in solution may be attributed to the length ratio of the ESI-MS (Figure S66-S67).
long-arm and short-arm linkers.
Figure 1. Synthesis and conversion of 4a-BRs, 4a, Ⅱ-BRs and Ⅱ.
In order to further prove this, the aspect ratios of the
metallarectangles were decreased again. 5a-BRs was
prepared using precursors 1a and shorter pyridyl ligands
L3 (L3 = 1,4-di(pyridin-4-yl)benzene) by the same meth-
od (Scheme 1). Single-crystal X-ray crystallographic
analysis and ESI-MS confirmed the BRs structure of
5a-BRs (Figure S52-S53 and Figure S64-S65). Similar to
4a-BRs, for 5a-BRs, the length of the long arm minus the
Figure 2. Single-crystal X-ray structures of (a) 4a-BRs, (b) II-BRs,
(c) the metallarectangle of 4a-BRs and (d) the metallarectangle of
II-BRs (N, blue; O, red; C, gray; Rh, orange; Pd, violet). Hydrogen
atoms and counter anions are omitted for clarity.
length of short arm is also less than 7 Å, the long arm of
The size of the metallarectangles of 7a-BRs is similar
to that of 5a-BRs, and the long arm of the metallarectan-
gle was also found to curve inward (Figure 3c and Table
S12). The distance between the Pd(Ⅱ) center and the
corresponding oxygen atom of the carbonyl group was
found to be ca. 3.6 Å (Figure S56). In methanol solution,
the 7a-BRs also displayed similar properties to 5a-BRs.
The spectra of 7a-BRs present two groups of metal-
larectangle signals in a 1:1 ratio (Figure S28-S31), and
these two groups of signals have the same diffusion con-
stants (Figure S32). The BRs structure of 7a-BRs is also
5a-BRs was found to curve inward, too (Figure S52 and
Table S12). Hence, the NMR spectra of 5a-BRs in
methanol also displayed two groups of metallarectangle
signals in a 1:1 ratio (Figure S21-S26), and these two
groups of signals have the same diffusion constants (Fig-
ure S23). And then, 5a-BRs also displays excellent sta-
bility in solution. The BR structure is immune to dilution,
and addition of DMSO (Figure S27). DFT binding energy
calculations were performed to further find out the dif-
ference of 4a-BRs and 5a-BRs on stability. The compu-
tational results showed that the binding energy between
This article is protected by copyright. All rights reserved.

immune to changes in concentration (Figure S33). But, long-arm and short-arm linkers also contributed to the
upon increasing the ratio of d6-DMSO in CD3OD, new formation of BRs. DFT binding energy calculations were
1
peaks were observed in H NMR spectrum, along with performed to further confirm the relationship between the
peaks from 7a-BRs, that indicated the formation of a new stability of BRs and the aspect ratio in this kind of BRs:
1
compound (7a; Figure S34). The H NMR signal of the the binding energy between the three MRs of 7a-BRs (–
new compound 7a presented an MR structure (Figure 217.5 kcal/mol) is stronger than that of 8a-BRs (–192.3
1
S34), which was confirmed by a H DOSY spectrum kcal/mol) (Table S1, Figure S76-S77). These results un-
(Figure S35). However, even in pure d6-DMSO (5.0 mM), derline that metallarectangles with smaller aspect ratios
the BRs structure of 7a-BRs cannot be completely trans- could help to increase the stability of BRs structures in
formed to the corresponding MR (Figure S34). Despite solution, both for BRs based on - stacking or those
the similar size of their metallarectangles, 5a-BRs is im- based on interactions between Pd(Ⅱ) centers and carbon-
mune to d6-DMSO, but 7a-BRs is not. This means that yl groups.
the axial position of Pd(Ⅱ) and the carbonyl group plays
a very important role in the formation of 7a-BRs.
Increasing the width of ligands. Metallarectangles
based on [Cu(opba)]2– and pyridyl ligands are promising
catalysts. However, by lengthening the distance between
the two Cu(Ⅱ) ions, BRs are formed, impeding catalysis
with larger substrates. Moreover, some of these BRs dis-
play high stability in solution. Thus, how to maintain the
distance between the two open Cu(Ⅱ) centers and achieve
an empty MR? We sought to increase the width of the
linker ligand (the details see catalytic activity studies in
supporting information).
We chose L4 (L4 = 1,4-bis(4-pyridyl)naphthalene) to
construct a metallarectangle (6b) with precursor 1b as L4
is the same length as L3, but significantly wider (Scheme
1). As predicted, the structure of 6b is a discrete MR,
which was confirmed by single-crystal X-ray crystallo-
graphic analysis (Figure 4) and ESI-MS (Figure S68). A
pair of conformational isomers were found to exist in the
crystal, denoted 6b-1 (naphthalene groups facing in the
same direction) and 6b-2 (naphthalene groups facing in
opposite
directions).
L7
(L7
=
Figure 3. Single-crystal X-ray structures of: (a) 7a-BRs, (b)
8a-BRs, (c) the metallarectangle of 7a-BRs and (d) the metal-
larectangle of 8a-BRs (N, blue; O, red; C, gray; Rh, orange; Pd,
violet). Hydrogen atoms and counter anions are omitted for clarity.
N,N-bis(4-pyridyl)-1,4,5,8-naphthalenetetracarboxydiimi
de) was also used to construct the metallarectangle 9b
under similar conditions. Similarly to L6, L7 also bears
suitable carbonyl groups for coordination, and while both
are of similar lengths, L7 is wider than L6 (Scheme 1). 9b
is a discrete MR, as confirmed by single-crystal X-ray
crystallographic analysis (Figure S58) and ESI-MS (Fig-
ure S69).
Compared with 7a-BRs, the aspect ratios of the
metallarectangles in 8a-BRs is obviously larger than that
in 7a-BRs (Figure 3d and Table S12). For 8a-BRs, the
distance between the Pd atom and the corresponding ox-
ygen of carbonyl group is only ca. 3.1 Å (Figure S57),
which is significantly shorter than this distance in
7a-BRs. This indicated that the interaction between the
Pd(Ⅱ) atom and the carbonyl group in 8a-BRs is stronger
than that in 7a-BRs. Interestingly, 8a-BRs display lower
stability in solution than 7a-BRs. The 1H NMR spectrum
of 8a-BRs is very cluttered and difficult to confidently
assign (Figure S36). However, upon increasing the ratio
of d6-DMSO to CD3OD, a clear signal attributable to the
corresponding MR, 8a, was observed (Figure S37-40).
Moreover, when the CD3OD:d6-DMSO ratio is 5:4 (5.0
mM), 8a-BRs was found to be completely transformed to
the corresponding MR 8a. However, for 7a-BRs, even if
in pure d6-DMSO (5.0 mM), the BR structure cannot be
completely transformed to corresponding MR. Further-
more, the BRs structure of 8a-BRs could be destroyed by
dilution (Figure S41). Therefore, unless the interaction
between Pd(Ⅱ) and carbanyl group, the length ratio of the
Figure 4. Single-crystal X-ray crystal structures of: (a) 6b-1 and (b)
6b-2 (N, blue; O, red; C, gray; Rh, orange; Cu, green). Hydrogen
atoms and counter anions are omitted for clarity. The distance be-
tween the Cu(Ⅱ) atom and the oxygen atom of DMF is ca. 2.2 Å,
DMF is added during the recrystallization process.
We reasoned that increasing the width of the
metallaligand could similarly hinder the formation of
BRs. Dirhodium precursors 2a and 2b were thus prepared
using [Pd(nabo)]2+ and [Cu(nabo)]2+, respectively. The
only difference between nabo and obpa is that a naphthyl
This article is protected by copyright. All rights reserved.

group replaces the phenylene group of the latter (Scheme abilities with high efficiency in the acyl transfer reaction
1). As predicted, the structure of 10b is also a discrete between NAI and (4-(pyridin-4-yl)phenyl)methanol. We
MR, which is confirmed by ESI-MS (Figure S70). In or- hope that our results will help to deepen the understand-
der to further prove the MR structure, 10a was prepared ing and awareness of complicated interlocked structure,
from 2a and L3 for NMR spectroscopic characterization. and advance the field of supramolecular assembly chem-
The NMR results of 10a showed typical signals for the istry.
MR structure, and only one group of metallarectangle
signals (Figure S42-S45).
Acknowledgement
Moreover, if the long arm was further lengthened, the
This work was supported by the National Science
BRs structure could be obtained once again, for example
Foundation of China (21531002, 21720102004), the Pro-
gram for Changjiang Scholars and Innovative Research
in the case of 11b-BRs, which is constructed from pre-
cursors 2b and L8 (L8 = 2,6-di(pyridin-4-yl)naphthalene)
Team in University (IRT-15R12) and the Shanghai Sci-
(Scheme 1). Its BRs structure is confirmed by the X-ray
ence Technology Committee (13JC1400600) and China
structures and ESI-MS (Figure S59 and S71). This fact
Postdoctoral Science Foundation (2015M581517); GXJ
further underscores the relationship between the for-
thanks the Alexander von Humboldt Foundation for a
mation of BRs and the width of the ligand.
Humboldt Research Award.
The catalytic acyl transfer reaction. After obtaining
the MR 6b and 10b, which are larger than those reported
previously,9 We attempted to use them as catalysts for the
acyl transfer reaction with larger substrates. We reasoned
that the NAI and (4-(pyridin-4-yl)phenyl)methanol sub-
strates could achieve optimal orientation and alignment
for highly efficient acyl transfer by binding within the
cavity of the metallacycle through the open Cu centers.
Furthermore, only a combination of substrates with the
right distance can span the cavity and react at an acceler-
ated rate. As is shown in Figure S72, both 6b and 10b
significantly accelerate the reaction rate, which may be
ascribed to 6b and 10b having highly favorable cavity
sizes for accommodating the two reactants. These results
also prove that the structure of 10b is indeed a discrete
MR. At the same time, the reaction rate of 6b is slightly
lower than that of 10b because of the conformational
isomer (6b-1 and 6b-2). 5b-BRs was found to ineffi-
ciently catalyze the reaction as there is no space for sub-
strate entry into the BR structure. However, the cavity of
References
1
(a) Raymo, F. M.; Stoddart, J. F. Chem. Rev. 1999, 99, 1643;
(b) Forgan, R. S.; Sauvage, J. P.; Stoddart, J. F. Chem. Rev.
2011, 111, 5434; (c) Ramirez, G. G.; Leigh, D. A.; Stephens,
A. J. Angew. Chem. Int. Ed. 2015, 54, 6110; (d) Berl, V.; Huc,
I.; Khoury, R. G.; Krische, M. J.; Lehn, J. M. Nature 2000,
407, 720; (e) Cook, T. R.; Stang, P. J. Chem. Rev. 2015, 115,
7001; (f) Smulders, M. M. J.; Riddell, I. A.; Browne, C.;
Nitschke, J. R. Chem. Soc. Rev. 2013, 42, 1728; (g) Kloster-
man, J. K.; Yamauchi, Y.; Fujita, M. Chem. Soc. Rev. 2009, 38,
1714; (h) Han, M.; Engelhard, D. M.; Clever, G. H. Chem.
Soc. Rev. 2014, 43, 1848; (i) Gianneschi, N. C.; Masar, M. S.;
Mirkin, C. A. Acc. Chem. Res. 2005, 38, 825; (j) Vogtle, F.;
Dunnwald, T.; Schmidt, T. Acc. Chem. Res. 1996, 29, 451; (k)
Xue, M.; Yang, Y.; Chi, X. D.; Yan X. Z.; Huang F. H.
Chem. Rev. 2015, 7398.
2
3
(a) Cantrill, S. J.; Chichak, K. S.; Peters, A. J.; Stoddart, J. F.
Acc. Chem. Res. 2005, 38, 1; (b) Meyer, C. D.; Forgan, R. S.;
Chichak, K. S.; Peters, A. J.; Tangchaivang, N.; Cave, G. W.
V.; Khan, S. I.; Cantrill, S. J.; Stoddart, J. F. Chem-Eur. J.
2010, 16, 12570.
(a) Song, Y. H.; Singh, N.; Jung, J.; Kim, H.; Kim, E. H.;
Cheong, H. K.; Kim, Y. ; Chi, K. W. Angew. Chem. Int. Ed.
2016, 55, 2007; (b) Schouwey, C.; Holstein, J. J.; Scopelliti,
R.; Zhurov, K. O.; Nagornov, K. O.; Tsybin, Y. O.; Smart, O.
S.; Bricogne, G.; Severin, K. Angew. Chem. Int. Ed. 2014, 53,
11261; (c) Beves, J. E.; Danon, J. J.; Leigh, D. A.; Lemonnier,
J. F.; Vitorica-Yrezabal, I. J. Angew. Chem. Int. Edit. 2015, 54,
7555; (d) Ibukuro, F.; Fujita, M.; Yamaguchi, K.; Sauvage, J.
P. J. Am. Chem. Soc. 1999, 121, 11014; (e) Prakasam, T.; Lusi,
M.; Elhabiri, M.; Platas-Iglesias, C.; Olsen, J. C.; Asfari, Z.;
Cianferani-Sanglier, S.; Debaene, F.; Charbonniere, L. J.;
Trabolsi, A. Angew. Chem. Int. Edit. 2013, 52, 9956.
3b
is
inherently
too
narrow
for
(4-(pyridin-4-yl)phenyl)methanol. These findings lead to
the postulate that the nitrogen atoms of NAI and
(4-(pyridin-4-yl)phenyl)methanol can easily coordinate to
the inner Cu centers of 6b and 10b, forming the coopera-
tively-bound intermediate with low activation energy,
thus facilitating the subsequent catalytic cycle (Figure
S73).
Conclusions
We have prepared a series of Cp*Rh-based BRs from
various binuclear precursors. Through experiments and
theoretical calculations, we found that the stability of the
BRs is associated with the length ratio of the long-arm
and short-arm linkers, with smaller aspect ratios leading
to improved stability and yield of the BRs in solution. By
increasing the width of the metallaligand or pyridyl lig-
and, the formation of BRs was hindered and MR could be
obtained. These MR have the similar sizes to the metal-
larectangles of corresponding BRs. With the appropriate
cavity size, MR 6b and 10b show remarkable catalytic
4
5
Leigh, D. A.; Pritchard, R. G.; Stephens, A. J. Nat. Chem.
2014, 6, 978.
Marcos, V. ; Stephens, A. J.; Jaramillo-Garcia, J.; Nussbaumer,
A. L.; Woltering, S. L.; Valero, A.; Lemonnier, J. F.; itori-
ca-Yrezabal, I. J. V; Leigh, D. A. Science 2016, 352, 1555.
(a) Danon, J. J.; Kruger, A.; Leigh, D. A.; Lemonnier, J. F.;
6
This article is protected by copyright. All rights reserved.

Stephens, A. J.; Vitorica-Yrezabal, I. J.; Woltering, S. L. Sci-
ence 2017, 355, 159; (b) Sawada, T.; Yamagami, M.; Ohara,
K.; Yamaguchi, K.; Fujita, M. Angew. Chem. Int. Edit. 2016,
55, 4519; (c) Fujita, M.; Fujita, N.; Ogura, K.; Yamaguchi, K.
Nature 1999, 400, 52; (d) Lee, H.; Elumalai, P.; Singh, N.;
Kim, H.; Lee, S. U.; Chi, K. W. J. Am. Chem. Soc. 2015, 137,
4674; (e) Zhu, R.; Lubben, J.; Dittrich, B.; Clever, G. H. An-
gew. Chem. Int. Edit. 2015, 54, 2796.
Han, Y. F.; Jin, G. X. Acc. Chem. Res. 2014, 47, 3571.
Chichak, K. S.; Cantrill, S. J.; Pease, A. R.; Chiu, S. H.; Cave,
G. W. V.; Atwood, J. L.; Stoddart, J. F. Science 2004, 304,
1308.
Kim, D.; Kim, D. H.; Kim, E. H.; Kim, H.; Lah, M. S.; Chi,
K. W. Dalton Trans. 2017, 46, 571; (c) Zhang, L.; Lin, L.; Liu,
D.; Lin, Y. J.; Li, Z. H.; Jin, G. X. J. Am. Chem. Soc. 2017,
139, 1653.
11 Lu, Y.; Deng, Y. X.; Lin, Y. J.; Han, Y. F.; Weng, L. H.; Li,
Z. H.; Jin, G. X. Chem 2017, 3, 110.
12 (a) Shultz, A. M.; Farha, O. K.; Hupp, J. T.; Nguyen, S. T. J.
Am. Chem. Soc. 2009, 131, 4204; (b) Yoon, H. J.; Mirkin, C.
A. J. Am. Chem. Soc. 2008, 130, 11590; (c) Yoon, H. J.; Heo,
J.; Mirkin, C. A. J. Am. Chem. Soc. 2007, 129, 14182.
13 Deria, P.; Gomez-Gualdron, D. A.; Hod, I.; Snurr, R. Q.;
Hupp, J. T.; Farha, O. K. J. Am. Chem. Soc. 2016, 138,
14449.
14 Han, Y. F.; Jin, G. X. Chem. Soc. Rev. 2014, 43, 2799.
15 Vajpayee, V.; Song, Y. H.; Yang, Y. J.; Kang, S. C.; Cook, T.
R.; Kim, D. W.; Lah, M. S.; Kim, I. S.; Wang, M.; Stang, P. J.;
Chi, K. W. Organometallics 2011, 30, 6482.
7
8
9
(a) Huang, S. L.; Lin, Y. J.; Hor, T. S. A.; Jin, G. X. J. Am.
Chem. Soc. 2013, 135, 8125; (b) Huang, S. L.; Lin, Y. J.; Li,
Z. H.; Jin, G. X. Angew. Chem. Int. Ed. 2014, 53, 11218.
10 (a) Kim, T.; Singh, N.; Oh, J.; Kim, E. H.; Jung, J.; Kim, H.;
Chi, K. W. J. Am. Chem. Soc. 2016, 138, 8368; (b) Singh, N.;
This article is protected by copyright. All rights reserved.

Entry for the Table of Contents
Page No.
Title
Highly Stable Molecular Borromean
Rings
A series of Cp*Rh-based molecular Borromean rings (BRs) are prepared from naphtha-
zarine or metallaligand. Smaller aspect ratios of the metallarectangles could promote
improved stability and yields of the BRs in solution. Increasing the width of ligand hin-
ders the formation of BRs and leads to unoccupied monomeric rectangles, which were
further used as catalysts for the acyl transfer reaction between N-acetylimidazole and
(4-(pyridin-4-yl)phenyl)methanol.
Ye Lu; Yue-Jian Lin; Zhen-Hua Li and
Guo-Xin Jin*
This article is protected by copyright. All rights reserved.