Stable Cu+, Ag+ complexes of aza-bridged macrocyclic molecules: Structure and chemical properties

Article abstract of DOI:10.1039/a604966d

The cage type compounds 1 and 2 form stable Cu⁺ or Ag⁺ complexes, which have been employed for the preparation of cation-free host molecules. A reaction between the potassium complex K⁺?1 and a Cu^II salt generates a Cu^I complex. The Cu^II/Cu^I redox potential is observed at +0.43 V (vs. SCE) in the cyclic voltammetry, which shows that the Cu⁺ state is stabilized by its rigid molecular skeleton and spatially fixed coordination sites. A reaction between Ag⁺ and K⁺?2 yields the dinuclear complex 2Ag⁺?2, which has a short Ag⁺ ... Ag⁺ distance (2.78 A). The halide anions (Cl^-, Br^-, I^-) remove one Ag⁺ from 2Ag⁺?2 to give Ag⁺?2, but further demetallation does not occur. CV measurements show that these silver complexes are electrochemically stable. Both silver complexes are stable to sunlight. The first preparations of guest-free hosts have been achieved by treating Cu⁺?1 or 2Ag⁺?2 with CN^-. Inclusions of neutral guests (NH₃, BH₃) have been attempted using these guest-free hosts.

Full text of DOI:10.1039/a604966d

View Article Online / Journal Homepage / Table of Contents for this issue
Stable Cu+, Ag+ complexes of aza-bridged macrocyclic molecules:
structure and chemical properties
Hiroyuki Takemura,*,a Noriyoshi Kon,b Keita Tani,c Kô Takehara,a
Junko Kimoto,b Teruo Shinmyozud and Takahiko Inazu*,b
a Department of Chemistry, Faculty of Science, Kyushu University,
Ropponmatsu 4-2-1, Chuo-ku, Fukuoka, 810 Japan
b Department of Chemistry, Faculty of Science, Kyushu University, Hakozaki 6-10-1,
Higashi-ku, Fukuoka, 812 Japan
c Division of Natural Science, Osaka Kyoiku University, Kashiwara, Osaka, 582 Japan
d Institute for Fundamental Research of Organic Chemistry, Kyushu University,
Hakozaki 6-10-1, Higashi-ku, Fukuoka, 812 Japan
The cage type compounds 1 and 2 form stable Cu+ or Ag+ complexes, which have been employed for the
preparation of cation-free host molecules.
A reaction between the potassium complex K +⊂1 and a CuII salt generates a CuI complex. The CuII/CuI
redox potential is observed at +0.43 V (vs. SCE) in the cyclic voltammetry, which shows that the Cu+
state is stabilized by its rigid molecular skeleton and spatially fixed coordination sites. A reaction between
Ag+ and K +⊂2 yields the dinuclear complex 2Ag+⊂2, which has a short Ag+ ؒ ؒ ؒ Ag+ distance (2.78 Å). The
halide anions (ClϪ, BrϪ, IϪ) remove one Ag+ from 2Ag+⊂2 to give Ag+⊂2, but further demetallation does
not occur. CV measurements show that these silver complexes are electrochemically stable. Both
silver complexes are stable to sunlight. The first preparations of guest-free hosts have been achieved by
treating Cu+⊂1 or 2Ag+⊂2 with CN Ϫ. Inclusions of neutral guests (N H 3, BH 3) have been attempted using
these guest-free hosts.
two ways. A reaction between K+⊂1 and Cu(MeCN)4PF6
Introduction
Ϫ
afforded Cu+⊂1ؒPF6 as yellow crystals in 51% yield (Scheme
In earlier reports, we showed that the host molecules 1 and 2
formed stable cation complexes with alkali metal cations and a
proton. For example, log Ka of K+, NH4+⊂1 were estimated to
be 5.2 and 6.3 in [2H6]-DMSO, respectively. In CDCl3, the sta-
bility of K+⊂1 is very high (log Ka = 15.1).1 These cation-bound
hosts are very stable and the cations could not be removed in
spite of many attempts. These compounds have electron-rich
cavities because of the convergence of the nitrogen lone pairs.
The driving forces for the cation inclusions are ion-dipole inter-
actions. Also, the nitrogen lone pair is a good binding site for
hydrogen bonds and coordinate bonds. Some neutral chemical
species which form hydrogen bonds or coordinate bonds are
considered to be possible guests. For example, NH3 forms
hydrogen bonds with itself or with amines,2 while BH3 (B2H6)
forms stable complexes with a wide variety of electron donors.
We expected that these hosts 1 and 2 would form inclusion
complexes with such neutral chemical species. However, many
attempts to prepare the inclusion complexes starting from their
cation complexes were unfruitful. Thus, it was concluded that
guest-free hosts were necessary to investigate the interactions
with these guest molecules. In order to prepare the guest-free 1
and 2, we employed the reactions between heavy metal com-
plexes of the hosts and CNϪ. Furthermore, stabilization of
heavy metal complexes is expected because of their structural
features: they have spatially fixed tetrahedral or octahedral
coordination sites and the cavities are surrounded by the six
aromatic rings. We report here aerially and photo- and electro-
chemically stable Cu+ or Ag+ complexes and their reactions
with halides and cyanide ions.
1). Alternatively, a reaction with the CuII salt also generated
Cu+⊂1 in 52% yield: refluxing the methanol solution containing
equal amounts of Cu(MeCO2)2 and K+⊂1ؒBrϪ for 12 h afforded
a mixture of Cu2+⊂1 and Cu+⊂1 as green and yellow materials,
respectively. Cu+⊂1 was easily isolated from the reaction mix-
ture by column chromatography on silica gel. Excess of
Cu(MeCO2)2 decreased the yield of Cu+⊂1. The geometry of
donor sites (tetrahedral and quasi-planar) of 1 allows both Cu+
and Cu2+ coordinations, thus, planar Cu2+ can be converted into
tetrahedral coordination and simultaneously reduced to Cu+.3
Unfortunately, the crystallographic analysis of Cu+⊂1ؒPF6Ϫ did
not produce good results. However, it appeared that a bridge-
head nitrogen and two pyridine rings coordinate to the Cu+ in a
distorted tetrahedral manner and that the position of the Cu+
ion is not at the centre of the cavity but near the coordinating
nitrogen atoms. Further attempts to obtain satisfactory results
from crystallographic analysis are being made. The octahedral
coordination type host 2 also formed a Cu+ complex, but less
stable than Cu+⊂1. Therefore, the ligand 1 was employed for
investigations of the Cu+ complex.4
In the electronic spectrum of the complex, the characteristic
MLCT (metal-to-ligand charge-transfer) transition bands were
observed (Fig. 1). The relatively strong band at 259 nm (log
ε = 4.22) is assigned to the π–π* transition and the bands at 364
nm (log ε = 3.52) and 440 nm (log ε = 2.50) are assigned to the
MLCT bands.5
Cu+⊂1 is stable under aerobic conditions both in solution
Ϫ
and in the solid state. The cyclic voltammogram of Cu+⊂1ؒPF6
in acetonitrile is shown in Fig. 2. A reversible wave was
observed and an oxidation potential peak appeared at +0.16 V
(vs. Ag/AgCl, 100 mV sϪ1; +0.43 V vs. SCE). It is noteworthy
that even at a low cathode potential (~ Ϫ1.8 V), precipitation of
metal copper did not occur. Stable Cu+ states are reported in an
oligo-pyridine-Cu+ complex (+0.42 V vs. SCE) and Cu+ caten-
Results and discussion
(i) Cu+ complex of 1
The preparation of the Cu+ complex of 1 could be achieved in
J. Chem. Soc., Perkin Trans. 1, 1997
239

View Article Online
Table 1 Selected bond lengths and angles
K+ 1 • Br –
Bond lengths (Å)
Bond angles (Њ)
Cu(CH3CN)4PF6
or
MeOH
Cu(OAc)2
Ag(1)᎐Ag(2)
Ag(1)᎐N(1)
Ag(1)᎐N(2)
Ag(1)᎐N(10)
Ag(2)᎐N(3)
Ag(2)᎐N(4)
Ag(2)᎐N(9)
2.780(6)
2.441(2)
2.258(2)
2.348(2)
2.379(2)
2.428(2)
2.448(2)
Ag(1)᎐Ag(2)᎐N(3)
Ag(1)᎐Ag(2)᎐N(4)
Ag(1)᎐Ag(2)᎐N(9)
N(1)᎐Ag(1)᎐N(2)
N(1)᎐Ag(1)᎐N(10)
N(2)᎐Ag(1)᎐N(10)
Ag(2)᎐Ag(1)᎐N(1)
Ag(2)᎐Ag(1)᎐N(2)
Ag(2)᎐Ag(1)᎐N(10)
N(3)᎐Ag(2)᎐N(4)
N(3)᎐Ag(2)᎐N(9)
N(4)᎐Ag(2)᎐N(9)
148.46(6)
96.95(6)
141.38(6)
117.97(8)
71.40(8)
156.62(8)
124.78(6)
84.31(6)
108.46(6)
104.15(8)
69.07(8)
70.35(8)
N
N
N
N
Cu+
• X –
N
N
N
N
Cu+ 1 • X –
–
X – = PF6 or AcO –
Bu4NCN
N
N
N
N
N
N
Ϫ
Fig. 2 Cyclic voltammogram of Cu+⊂1ؒPF6 in MeCN. Support
electrolyte = 0.1  Et4NClO4, working electrode = Pt disk, reference
electrode = Ag/0.01  AgNO3, counter electrode; Pt wire, scan rate; 100
N
N
mV sϪ1
.
guest-free 1
that two Ag+ ions are contained in the complexes. The cavity
size of 2 is not large enough to accommodate two silver ions
(Ag+: 1.31 Å in radius, cavity size: ca. 1.9 Å in radius).
Scheme 1 Preparations of Cu+⊂1 and guest-free 1
Since, in earlier experiments, 1:1 complexes with alkali metal
ions (Li+–Cs+) were obtained 1d this complex in the early stage
Ϫ
ؒ
of the study, was thought to have the formula Ag+⊂2ؒBF4
AgBF4. However, cyclic voltammetry showed no ligand-free
Ag+ in the solution. Therefore, in order to clarify the structure,
an X-ray crystallographic analysis was carried out, revealing
that, although the size of Ag+ (1.31 Å in radius) is similar to
that of the K+ ion (1.33 Å in radius), two silver ions are actually
incorporated into the cavity of 2. The ligand structure of 2 is
considerably distorted and produces enough space for the two
silver ions: each of the two silver ions is coordinated to a bridge-
head nitrogen and two pyridine nitrogens in pyramidal geom-
etries (Fig. 3). Representative bond lengths and dihedral angles
are listed in Table 1 and the coordination geometries around
silver ions are illustrated in Fig. 4.
The Ag+᎐Ag+ distance is 2.78 Å, which is short, compared to
some of the reported dinuclear silver complexes or metallic
silver (2.889 Å).7
Fig. 1 Electronic spectrum of Cu+⊂1ؒPF6Ϫ in MeCN
By using molecular orbital calculations, Cotton et al. con-
cluded that there is no Ag+᎐Ag+ bond even in the dinuclear com-
plex, which has a short Ag+᎐Ag+ distance (2.705 Å).8 Jennische
and Hesse pointed out that the Ag+᎐Ag+ distance is determined
not by the bond but by the ligand structures.9 However, in our
ate (+0.565 V vs. SCE).5a,c Although tris(2-pyridyl)methylamine
is equivalent to the coordination site of 1, its Cu+ complex is
reported to be unstable.6 In the case of Cu+⊂1, the Cu+ state is
stabilized by its bulky molecular skeleton and spatially fixed
coordination geometry.
Ϫ
experiments, the complex Ag+⊂2ؒBF4 was not produced by a
reaction between K+⊂2ؒBrϪ and equimolar amounts of AgBF4:
the product was instead a mixture of 2Ag+⊂2ؒ(BF4Ϫ)2 and the
K+ complex. Furthermore, the host incorporates one Cu+,
although Cu+ (0.96 Å) is smaller than Ag+ (1.31 Å) and both ions
have tetrahedral coordination structures.
(ii) Structure and reactions of the Ag+ complex of 2
The results of elemental analyses of the complexes obtained
from reactions between K+⊂2 and AgNO3 or AgBF4 showed
240
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
Fig. 3 X-Ray structure of 2Ag+⊂2ؒ(BF4Ϫ)2. Hydrogen atoms are omitted for clarity. The numbering of the atoms is different from that of the
standard nomenclature. The bond between two Ag atoms shown here is not an actual bond.
Ϫ
Table 2 Crystallographic data for 2Ag+⊂2ؒ(BF4
)
2
Formula
Fw
C42H42N10Ag2B2F8
1076.20
Crystal shape
Column
Crystal dimensions, mm
0.30 × 0.10 × 0.25
Triclinic
Crystal system
Lattice parameters
a/Å
b/Å
c/Å
13.357(3)
14.787(3)
12.726(3)
¯
P1
2
2069.2(9)
1.727
288
Space group
Z
V/Å3
Dc/g cmϪ3
T/K
Radiation (λ, Å)
Mo-Kα (0.710 69)
10.27
µ/cmϪ1
R (Rw) (%)
0.032 (0.034)
Fig. 4 Bond lengths and angles for Ag+ ions and coordinated nitrogen
atoms. The structure was produced by the Chem 3D software on the
basis of the crystallographic data.
It is unclear why the host incorporates two Ag+ ions simul-
taneously in its cavity by consuming the free energy which dis-
torts the host structure. So far, we do not know whether these
results point to the existence of Ag+᎐Ag+ interactions or not.
In the 1H NMR spectrum, the appearance of signals is very
similar to that of K+⊂2 and each signal is sharp at room tem-
perature: the molecular movement of the host and the intra-
cavity Ag+ exchange are rapid in the solution, while distortion of
the structure could not be observed. However, at the low temper-
atures, both the movements of the host and the Ag+ exchanges
are frozen and the signals split into complex patterns (Fig. 5).
The silver complex Ag+⊂2 was obtained by treating 2Ag+⊂2
with halide ions. However, Ag+⊂2 was unexpectedly stable
toward halide ions. The reactions between 2Ag+⊂2ؒ(NO3Ϫ)2 and
excess of Bu4NX (X = ClϪ, BrϪ, IϪ) were monitored by means
2Ag+⊂2 was removed by halide ions and a small amount of
H+⊂2 was generated as a by-product. However, further de-
metallation did not occur, even when the solution was heated.
Ϫ
After the addition of Bu4NI into a solution of 2Ag+⊂2ؒ(NO3
)
2
in CD3OD, the aromatic protons shifted to a higher field (~ 0.2
ppm) and the benzyl protons slightly shifted to higher field
(0.02 ppm). Since further information was not obtained by the
spectra, we attempted to isolate the products. Treatment of
2Ag+⊂2ؒ(BF4Ϫ)2 with Me4NBr in methanol at room tempera-
ture gave Ag+⊂2ؒBrϪ as stable colourless crystals. This complex
was then isolated and fully characterized on the basis of its 1H
NMR and FAB mass spectra and elemental analysis (Scheme
2). According to the results of the X-ray analysis described
1
of H NMR spectroscopy (Fig. 6). One of the silver ions of
J. Chem. Soc., Perkin Trans. 1, 1997
241

View Article Online
Fig. 6 1H NMR (270 MHz) spectral changes of 2Ag+⊂2ؒ(NO3Ϫ)2 in
CD3OD by addition of Bu4NI and irradiation of light; (a)
2Ag+⊂2ؒ(NO3Ϫ)2 in CD3OD; (b) after addition of Bu4NI into (a) ; (c)
after irradiation of high pressure-Hg lamp to (b) for 1 h. * Shows the
signal of H+⊂2.
Ϫ1.20 V (P3). A peak at Ϫ1.01 V (P2) corresponds to the
absorption of the complex on the electrode. A redox couple,
Ag+/Ag0 was observed at Ϫ0.44 V (P1) and +0.00 V (P4) after
the reaction (1) took place at Ϫ1.20 V. The couple Ag+/Ag0
grows as the cycle is repeated. This is ascribed to the silver metal
precipitation, resulting from the reaction (1), on the electrode.
On the other hand, the reaction (2) does not take place even at
Fig. 5 VT NMR of 2Ag+⊂2ؒ(BF4Ϫ)2 in [2H4]methanol
above, the structure of 2Ag+⊂2 is distorted and one of the two
silver ions can interact with a bromide ion. The strain may be a
driving force for the removal of one silver ion. After the
removal of Ag+, the host’s structure changes so as to accom-
modate one silver ion with six pyridine rings and so it is pro-
tected from further attacks of the halide anions. The
Ag+ ؒ ؒ ؒ BrϪ distance in this system could not be estimated
because Ag+⊂2ؒBrϪ crystals suitable for X-ray crystallographic
analysis were not obtained.
Ag+⊂2 → Ag0 + 2
(2)
the low cathode potential (~Ϫ1.5 V). This is clearly shown by
Ϫ
the CV measurement of Ag+⊂2ؒBF4 under the same condi-
tions [Fig. (7b)]. Also in this case, the peak corresponding to the
reaction (2) was not observed.
Both the solids and the solutions of 2Ag+⊂2 and Ag+⊂2ؒBrϪ
are stable toward sunlight: precipitation of silver by exposure to
direct sunlight for at least 6 h could not be observed. Methanol
solutions containing 2Ag+⊂2ؒ(NO3Ϫ)2 and excess amounts of IϪ
or ClϪ ions were also stable. Irradiation of the solution with a
high-pressure Hg lamp (100 W) for 1 h gave radical-mediated
products: the NMR spectra showed that some parts of the
complex decomposed, but characterization of the products
could not be achieved (Fig. 6).
According to the results of CV measurements, the Ag+ ions
of these two complexes are stabilized by the ligand structure.
An Ag+ ion of 2Ag+⊂2 can be removed only at the low cathode
potential (Ϫ1.20 V), producing Ag+⊂2. After the removal of
one Ag+, further demetallation reaction does not occur, at least
at the cathode potential of Ϫ1.5 V.
As a result, insoluble and photosensitive AgX (X = Cl, Br, I)
becomes soluble and stable towards light because of effective
shielding of Ag+ and ion pair separation by the host. Further-
more, accommodation of Ag+ by the large and rigid host’s skel-
eton greatly stabilizes the redox properties of Ag+.
1
Unlike those of 2Ag+⊂2, the H NMR spectra of Ag+⊂2 at
low temperatures (~Ϫ90 ЊC) showed that the ligand structure is
not distorted. The spectrum is quite similar to that of K+⊂2 at
low temperatures. Hence, the Ag+ ion of Ag+⊂2 is located at the
centre of the cavity and the whole structure is the same as that
of K+⊂2.
(iii) Guest-free 1 and 2
In a previous report, we described the high cation affinities of
compounds 1 and 2. A proton could not be removed from
H+⊂1 and H+⊂2 even using strong bases.1 Other attempts to
obtain guest-free 1 and 2 were unsuccessful. For example, a
solution of K+⊂2 in water/MeOH = 1:1 (v:v) was heated in a
sealed tube at 150 ЊC for 7 days.1d,10 In spite of these severe
conditions, starting material was recovered unchanged and no
metal-free ligand was detected. Therefore, we employed alter-
native methods in order to obtain guest-free 1 and 2. Sauvage et
Cyclic voltammetry of 2Ag+⊂2ؒ(NO3Ϫ)2 was carried out in
acetonitrile (Fig. 7). No redox waves appeared in the range of
Ϫ0.75–+0.5 V, but irreversible waves were observed in the range
of Ϫ1.5 V–+0.5 V (Ag/0.01  AgNO3 as a reference electrode).
The peak that corresponds to the reaction (1) appeared at
2Ag+⊂2 → Ag0 + Ag+⊂2
(1)
242
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
N
N
N
N
AgBF4
–
K+ 2 • Br–
N
2 BF4
N
•
N
N
N
Ag+
–
2Ag+ 2 • (BF4
)
2
Et4NCN
Me4NBr
N
N
N
N
N
N
N
N
N
N
N
N
Br–
N
N
•
N
N
N
N
N
N
Ag+ 2 • Br
guest-free 2
Scheme 2 Preparations of 2Ag+⊂2ؒ(BF4Ϫ)2, Ag+⊂2ؒBrϪ and guest-free 2
al. succeeded in preparing metal-free catenane from reactions
between Cu+ catenates and KCN or Me4NCN. Lehn et al. syn-
thesized a macrobicyclic cryptand starting from its Ag+ com-
plex and H2S.11 These methods were applied to the preparation
(iv) Attempts to include NH3 and BH3 molecules
A water complex H2O⊂2 was characterized by means of mass
and NMR spectroscopy as described in a previous report.1a In
the next stage of our investigation, we were interested in the
inclusion complexes, NH3⊂L and BH3⊂L (L = 1, 2). A
cryptand–borane complex was reported in which borane mol-
ecules coordinate from the outside of the cryptand, but in gen-
eral, borane inclusion complexes are unknown.15 Several
attempts to prepare inclusion complexes through reactions
of the cation-free compounds 1 and 2. However, the introduc-
Ϫ
tion of H2S gas into an aqueous solution of 2Ag+⊂2ؒ(NO3
)
2
immediately generated nH+⊂2 (n = 1, 2 or 3): protons liberated
by the reaction of H2S and Ag+ were captured by 2.12 Proton-
ation of 2 occurred even in the presence of a base, Me4NOH.
When KCN was employed for the demetallation of Cu+⊂1 and
Ag+⊂2, the complexes K+⊂1, 2 were obtained instead of the
cation-free 1 and 2. Similar occurrences were mentioned in
connection with the synthesis of cation-free catenane.11a
Tetraalkylammonium cyanide would be a suitable reagent to
between NH4+⊂L and OHϪ (subtraction of H+ from NH4 ),
+
K+⊂L and ButNH2ؒBH3 (replacement of K+ by BH3), H+⊂L
Ϫ
and BH4 (complexation after deprotonation by BH4Ϫ) were
unsuccessful. Therefore, it was concluded that direct reactions
between neutral guests and the guest-free hosts should be car-
ried out.
Ϫ
achieve the desired goal. A reaction between Cu+⊂1ؒPF6 and
Bu4NCN proceeded cleanly and quantitatively: the yellow col-
our immediately disappeared when a solution of Cu+⊂1 was
treated with Bu4NCN (Scheme 1). Metal-free 1 was precipitated
from the reaction mixture as colourless crystals. The 1H NMR
spectra drastically changed after the addition of CNϪ: the ben-
zene inner protons of 1 shifted from 8.73 to 8.16 ppm and the
The guest-free 2 was dissolved in dry liquid ammonia and the
solution was left for 2 h. After the ammonia had evaporated, a
white powder remained, which was shown to be NH4+⊂2 from
its NMR spectrum. There are two possible ways to produce
NH4+⊂2. A proton transfer from trace amounts of water to a
possible intermediate inclusion complex NH3⊂2 may produce
NH4+⊂2. NH3⊂2 is thought to be a stronger base than 2 itself
because of the formation of favourable N ؒ ؒ ؒ H+ ؒ ؒ ؒ N bonds.
relatively broad AB quartet of the methylene became sharp. As
Ϫ
in the case of Cu+⊂1, the two silver ions of 2Ag+⊂2ؒ(BF4
)
2
could be removed simultaneously by Et4NCN in MeCN
(Scheme 2).
H2O
2 + NH3 → [NH3⊂2]
NH4+⊂2ؒOHϪ
Both guest-free 1 and 2 are relatively labile when exposed to
air: they absorb moisture and/or CO2, yielding less soluble
materials. The NMR spectra showed that more than three kinds
of protonated species are produced. Previously, we reported
that the bond isomer of 2 easily absorbed water molecules, pro-
ducing a hydrated material.13 However, in the case of com-
pounds 1 and 2, no hydration occurred, instead, protons were
subtracted from water.14 The reactions of 2 with water in a
toluene or a pyridine solution at 100 ЊC were monitored by 1H
NMR spectroscopy. After a week, compound 2 was completely
consumed and the generation of protonated species and a water
inclusion complex were observed.
The counter anion is a hydroxide ion, because the addition of
a small amount of water to a CDCl3 solution of the product
resulted in the remarkable growth of CHCl3 signal, which is
caused by the following reaction.
NH4+⊂ 2ؒOHϪ
CDCl3 + H2O
CHCl3 + HDO
+
Another possibility is to trap the NH4 ion from the equi-
librium mixture:
+
NH3 + H2O
NH4 + OHϪ (K = 1.79 × 10Ϫ5)
J. Chem. Soc., Perkin Trans. 1, 1997
243

View Article Online
sponding to the borane inclusion complex were not detected.
The resultant insoluble material thought to be a borane adduct,
2ؒnBH3 in which borane molecules coordinate from the outside
of the host molecule. In order to remove the excess of BH3 and
dissolve the adduct, the precipitates were dispersed in [2H8]-
THF and heated at 50 ЊC for 2 weeks, but the mixture did
not change. Thus, the structural information could not be
obtained so far. Quenching the excess of BH3 of the
adduct with methanol produced slow recovery of guest-free 2
with nH+⊂2 (n = 1–3) as by-products.
Experimental
All mps were measured in Ar sealed tubes and were un-
corrected. The 1H NMR spectra were recorded on a JEOL GSX
270 (270 MHz) spectrometer and UV/VIS spectra were recorded
on a Shimadzu UV-2200 spectrometer. Cyclic voltammetry was
performed using a Fuso Model 311 Polarograph with a Model
321 potential-sweep unit at 25 ± 0.1 ЊC. A Pt disk was used as
the working electrode. An Ag/0.01  AgNO3 /0.1  Et4NClO4
(MeCN) and a Pt wire were used as the reference electrode and
the counter electrode, respectively. Support electrolyte
(Et4NClO4, Tokyo Kasei, special analysis grade) was dried in
vacuo at 50 ЊC for several hours.
Ϫ
X-Ray crystallographic analysis of 2Ag+⊂2ؒ(BF4
)
2
Crystal data: C42H42N10Ag2B2F8 (formula weight 1076.20),
colourless column, triclinic (recrystallized from methanol),
crystal dimensions 0.30 × 0.10 × 0.25 mm, Dc l.727 g cmϪ3
,
¯
space group P1 (# 2): a = 13.357(3), b = 14.787(3), c = 12.726(3)
Å, α = 112.51(2)o, β = 116.53(2)o, γ = 75.88(2)o, V = 2069.2(9)
Å3, Z = 2, F000 = 1080. All measurements were performed on a
Rigaku AFC7R diffractometer with graphite monochromated
Mo-Kαradiation (λ = 0.710 69Å), µ = 10.27cmϪ1. Thedata were
collected at a temperature of 15 ± 1 ЊC using the ω–2θ scan
technique to a maximum 2θ value of 55.0o. Of the 9914 reflec-
tions which were collected, 9505 were unique (Rint = 0.009). The
data were corrected for Lorentz and polarization effects. The
structure was solved following the heavy-atom Patterson
methods and expanded using Fourier techniques. The non-
hydrogen atoms were refined anisotropically. The hydrogen
atom coordinates were refined but their isotropic total back-
ground counts were held fixed. The final cycle of full-matrix
least-squares refinement was based on 8241 observed reflections
[I > 3.00σ(I)] and 704 variable parameters. The final R factors
Ϫ
Fig.
7
Cyclic voltammogram of (a) 2Ag+⊂2ؒ(NO3
)
and (b)
2
Ag+⊂2ؒBF4Ϫ in 0.1  Et4NClO4–MeCN. Support electrode; Et4NClO4,
working electrode; Pt disk, reference electrode; Ag/0.01  AgNO3,
counter electrode; Pt wire, scan rate; 100 mV sϪ1
.
In another experiment, dry NH3 gas was introduced into a
[2H8]toluene solution of 2 and the NMR sample tube was
sealed with Ar gas. The NH3 proton signal appeared at 0.02,
Ϫ0.14 and Ϫ0.30 ppm (J = 43 Hz) as a relatively broad triplet in
[2H8]toluene (the methyl proton of toluene served as an internal
reference; PhCHD2, 2.1 ppm), but no spectral changes were
observed. After 24 h at room temperature, a white powder was
precipitated and the signals of 2 disappeared. The material,
which was soluble in CDCl3, was revealed to be NH4+⊂2ؒOHϪ.
Similar results were obtained in the reaction between 1 and
NH3.
were R = Σ |F | Ϫ |F | /Σ|F | = 0.032,
Rw = [Σw(|Fo| Ϫ |Fc|)2/
|
|
o
c
o
¹
2

²
ΣwFo ] = 0.034. The maximum and minimum peaks on the
final difference Fourier map are 1.26 and Ϫ0.68 e ÅϪ3, respec-
tively. All calculations were performed using the teXscan17
crystallographic software package from Molecular Structure
Corporation. Detailed crystallographic results have been
deposited with the Cambridge Crystallographic Data Centre.†
Any requests for this material should be accompanied by
a full bibliographic citation together with the reference
number 207/71.
In these experiments, the intermediate species (NH3⊂2) was
not found by NMR spectroscopy. Although all of the experi-
ments were carried out under dry conditions, trace amounts of
+
water lead to the formation of NH4 complexes. Lehn et al.,
also pointed out that the NH3 cryptate is a short-lived species.16
According to the results, hydrogen bond interactions between
NH3 and the hosts L were too weak to form a stable inclusion
+
complex, but stronger cation-dipole interactions L ؒ ؒ ؒ NH4
Preparation of Cu+⊂1
lead to the formation of NH4+⊂2 in the presence of trace
amounts of water.
Ϫ
Cu+⊂1ؒPF6 from the CuI salt. The potassium complex
K+⊂1ؒBrϪ (37 mg, 0.046 mmol) and Cu(MeCN)4ؒPF6 (25.8 mg,
0.069 mmol) were dissolved in acetonitrile (5 cm3) and the solu-
tion was stored under an Ar atmosphere for 1 week. After con-
centration of the solution, precipitated yellow granules were
collected and washed with water. The complex was dissolved in
a minimum amount of methanol and slowly passed through an
anion exchange column (Muromac 1-X8, PF6Ϫ form). The solu-
tion was evaporated to dryness and column chromatographed
on silica gel (CH2Cl2–MeOH = 98:2, v/v). Recrystallization of
Pyridine–borane or amine–borane complexes are well-
known as stable substances. The host 2 is expected to be a good
and suitable ligand for borane because six pyridine lone pairs
are converging into the cavity. A reaction between ButNH2ؒBH3
and 2 in [2H8]toluene at 100 ЊC for 1 week did not occur since
ButNH2ؒBH3 is very stable. Instead of the amine complex, an
Me2SؒBH3 complex was employed. Addition of 2 to a solution
of Me2SؒBH3 in [2H8]toluene immediately yielded a colourless
precipitate. The intensity of the BH3 quartet signal which
appears at 2.65, 2.25, 1.86 and 1.46 ppm (J = 106 Hz)
decreased. Although the mixture was heated at 100 ЊC for a
week, its NMR spectrum showed no change and signals corre-
† For details of the scheme, see Instructions for Authors (1997), J.
Chem. Soc., Perkin Trans. 1, 1997, Issue 1.
244
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
the resultant powder from EtOH–CH2Cl2 afforded yellow
granules (21 mg, 51%), mp >258Њ C (decomp.) (Found: C, 55.8;
H, 5.2; N, 11.6. C44H44N8ؒCuPF6ؒ1/2CH2Cl2ؒH2O requires C,
56.03; H, 4.97; N, 11.75%). The quantities of CH2Cl2 and H2O
were confirmed by means of the 1H NMR spectrum.
λmax(MeCN)/nm 208 (ε 54 900 dm3 molϪ1 cmϪ1), 259 (16 500)
and 364 (3300); δH (270 MHz; CDCl3; Me4Si) 8.73 (2 H, s,
benzene-H), 7.70, 7.68, 7.65 (4 H, t, J 7, Py-H), 7.29, 7.26 (8 H,
d, J 6, Py-H), 7.17–7.14 (6 H, m, benzene-H), 3.76, 3.54 (16 H,
ABq, J 13, pyridine-CHAHB) and 3.52 (8 H, s, benzene-CH2);
m/z (FAB) 747 (M + 63Cu+, 99%) and 749 (M + 65Cu+, 63%).
Cu+⊂1ؒBrϪ from the CuII salt. The K+⊂1ؒBrϪ (12.2 mg, 0.015
mmol) and Cu(MeCO2)2ؒ6H2O (4.2 mg, 0.021 mmol) were dis-
solved in methanol (5 cm3) and the solution was heated under
reflux overnight. After the concentration of the solution, pre-
cipitated yellow granules were collected and washed with water.
The complex was dissolved in a minimum amount of methanol
and slowly passed through an anion exchange column
(Muromac 1-X8, BrϪ form). The solution was evaporated to
dryness and column chromatographed on silica gel (CH2Cl2–
MeOH = 98:2, v/v). Recrystallization of the resultant powder
from methanol afforded yellow granules (6.5 mg, 52%). 1H
NMR and FAB mass spectra of this material completely
from K+⊂1ؒBrϪ) of 1, mp 243–243.3 ЊC (Found: C, 77.1; H, 6.5;
N, 16.3. C44H44N8 requires C, 77.16; H, 6.48; N, 16.36%);
δH (270 MHz; CDCl3; Me4Si) 8.15 (2 H, s, benzene-H), 7.41,
7.38, 7.35 (4 H, t, J 8, Py-H), 7.01, 6.98 (8 H, d, J 8, Py-H),
7.25–7.21 (6 H, m, benzene-H), 3.78, 3.61 (16 H, ABq, J 12,
pyridine CHAHB) and 3.65 (8 H, s, benzene-CH2); m/z (FAB)
685 (M + H+, 51%).
Preparation of guest-free 2
To a solution of the silver complex 2Ag+⊂2ؒ(BF4Ϫ)2 (28.0 mg,
0.026 mmol) in MeCN (15 cm3) was added Et4NCN (42 mg,
0.27 mmol) and the solution was stirred under an Ar atmos-
phere at room temperature for 2 h. The precipitate was collected
and recrystallized from CH2Cl2–MeCN to give colourless nee-
dles (10.2 mg, 55%), mp 256 ЊC (Found: C, 73.3; H, 6.1; N, 20.1.
C42H42N10 requires C, 73.44; H, 6.16; N, 20.39); δH (270 MHz;
CDCl3; Me4Si) 7.45, 7.42, 7.39 (6 H, t, J 8, Py-H), 7.04, 7.01 (12
H, d, J 8, Py-H) and 3.91 (24 H, s, CH2); m/z (FAB) 687
(M + H+, 100%).
The reaction between 2 and NH3
Compound 2 (12 mg, 0.017 mmol) was introduced into a flame-
dried flask under Ar gas; the flask was then cooled to Ϫ80 ЊC.
Dry ammonia gas (distilled from sodium–liquid ammonia mix-
ture) was introduced into the flask. After 1 cm3 of liquid ammo-
nia had been condensed, the mixture was stirred for 2 h under an
Ar atmosphere. Alternatively, compound 2 (5 mg, 7.3 × 10Ϫ3
mmol) was dissolved in dry [2H8]toluene and the ammonia gas
was allowed to bubble through it for 5 min. The NMR tube was
sealed with Ar gas and the spectra were recorded.
Ϫ
coincide with that of Cu+⊂1ؒPF6 which is obtained from the
CuI salt.
–
–
Preparation of 2Ag+⊂2ؒ(BF4 )2 and 2Ag+⊂2ؒ(NO3 )2ؒH2O
The potassium complex K+⊂2ؒBrϪ (25.8 mg, 0.026 mmol) and
AgBF4 (22.1 mg, 0.11 mmol) was dissolved in methanol (20
cm3) and the solution was stirred at room temperature over-
night. It was then passed through a Celite column and evapor-
ated. The resultant powder was recrystallized twice from
Acknowledgements
o
methanol to give colourless needles (27 mg, 78%) mp >284 C
(decomp.) (Found: C, 46.8; H, 4.0; N, 13.0. C42H42N10ؒ2AgBF4
requires C, 46.87; H, 3.93; N, 13.01%); δH (270 MHz; CD3OD;
Me4Si) 7.76, 7.73, 7.70 (6 H, t, J 8, Py-H), 7.29, 7.26 (12 H, d,
J 8, Py-H) and 3.93 (24 H, s, CH2); m/z (FAB) 991 (M +
This research was supported by
Encouragement of Young Scientists (No. 05740398) provided
by the Ministry of Education, Science and Culture, Japan.
a
Grant-in-Aid for
2
109Ag+ + BF4Ϫ, 5%), 989 (M + 107Ag+ + 109Ag+ + BF4Ϫ, 10%),
987 (M + 2 107Ag+ + BF4Ϫ, 6%), 795 (M + 109Ag+, 13%) and 793
(M + 107Ag+, 16%).
References
1 (a) H. Takemura, T. Shinmyozu and T. Inazu, J. Am. Chem. Soc.,
1991, 113, 1323; (b) H. Takemura, T. Shinmyozu and T. Inazu,
Tetrahedron Lett., 1988, 29, 1789; (c) H. Takemura, T. Hirakawa,
T. Shinmyozu and T. Inazu, Tetrahedron Lett., 1984, 25, 5053; (d)
H. Takemura, T. Shinmyozu and T. Inazu, Coord. Chem. Rev., in
press; (e) H. Takemura, Ph.D. Thesis, Kyushu University, 1992.
2 (a) P. A. Kollman and L. C. Allen, Chem. Rev., 1972, 72, 283; (b)
P. A. Kollman and L. C. Allen, J. Am. Chem. Soc., 1971, 93, 4991;
(c) H. J. Berthold, W. Preibsch and E. Vonholdt, Angew. Chem., Int.
Ed. Engl., 1988, 27, 1524.
–
2Ag+⊂2ؒ(NO3 )2ؒH2O was prepared by the reaction between
K+⊂2ؒBrϪ and AgNO3. The procedure is the same as that of
2Ag+⊂2ؒ(BF4Ϫ)2, mp >274 ЊC (decomp.) (Found: C, 48.3; H,
4.3; N, 16.0. C42H42N10ؒ2AgNO3ؒH2O requires C, 48.29; H,
4.25; N, 16.09%).
Preparation of Ag+⊂2ؒBrϪؒH2O
The silver complex 2Ag+⊂2ؒ(BF4Ϫ)2 (42.1 mg, 0.039 mmol) and
Me4NBr (39.5 mg, 0.26 mmol) were dissolved in methanol (30
cm3) and the solution was stirred at room temperature over-
night. It was then passed through a Celite column and evapor-
ated. The resultant powder was recrystallized twice from
MeOH–AcOEt to give colourless needles (25.4 mg, 73%),
mp >243 ЊC (decomp.) (Found: C, 56.6; H, 4.9; N, 15.6.
C42H42N10ؒAgBrؒH2O requires C, 56.51; H, 4.74; N, 15.69%);
δH (270 MHz; CD3OD; Me4Si) 7.48, 7.45, 7.42 (6 H, t, J 8, Py-H),
7.05, 7.02 (12 H, d, J 8, Py-H) and 3.88 (24 H, s, CH2); m/z (FAB)
793 (M + 107Ag+, 100%) and 795 (M + 109Ag+, 95%).
3 In this case, an oxidation reaction takes place, but we did not clarify
which materials were oxidized.
4 A benzene analogue of 1 and 2 (compound 3, Fig. 8) showed no
inclusion ability towards alkali metal ions but formed a Cu+ com-
plex. The cavity has poor ability as a ligand because the cavity size
of 3 is 1.8 Å in diameter (half that of 2) and has only four bridge-
head nitrogens as donor atoms. The reaction between 3 and
Cu(MeCN)4PF6 in MeCN–CH2Cl2 gave colourless crystals. The
complex is unstable and easy to decompose under aerobic condi-
tions. Only FAB mass and NMR spectra supported the formation of
the complex. Exposure of the complex to air gradually generates
metal-free 3 and an inorganic CuII salt. A silver complex of 3 was
Preparation of guest-free 1
N
A reaction mixture obtained from K+⊂1ؒBrϪ (74 mg, 0.092
mmol) and Cu(MeCN)4ؒPF6 (53 mg, 0.14 mmol) was column
chromatographed on silica gel (CH2Cl2–MeOH = 95:5, v/v).
Ϫ
The Cu+⊂1ؒPF6 thus obtained was dissolved in MeCN (10
N
cm3) and the solution of Bu4NCN in MeCN was added to it
until the yellow colour of the solution disappeared. After 1 h,
the solution was concentrated to 1 cm3 and precipitated colour-
less crystals were filtered off and washed with a small amount
of MeCN and then with water. Recrystallization of the crystals
from MeCN–CH2Cl2 gave colourless granules (43.4 mg, 80%
N
N
Fig. 8 Structure of compound 3
J. Chem. Soc., Perkin Trans. 1, 1997
245

View Article Online
(n = 1, 2 and 3) were observed. The results were used for the
determination of the number of protons captured in the cavity of 2.
13 H. Takemura, S. Osada, T. Shinmyozu and T. Inazu, J. Chem. Soc.,
Perkin Trans. 1, 1996, 277.
also formed by the reaction between 3 and AgBF4, but this complex
was less stable than the Cu+ complex. Further identification or char-
acterization of the Cu+ؒ3 and Ag+ؒ3 were not carried out because of
their lability.
5 (a) C. Dietrich-Buchecker, J.-P. Sauvage and J.-M. Kern, J. Am.
Chem. Soc., 1989, 111, 7791; (b) K. T. Potts, M. Keshavarz-K,
F. S. Tham, H. D. Abruña and C. R. Arana, Inorg. Chem., 1993, 32,
4422; (c) K. T. Potts, M. Keshavarz-K, F. S. Tham, H. D. Abruña
and C. Arana, Inorg. Chem., 1993, 32, 4436.
6 K. D. Karlin, J. C. Hayes, S. Juen, J. P. Hutchinson and J. Zubieta,
Inorg. Chem., 1982, 21, 4106.
7 (a) C. E. Housecroft, Coord. Chem. Rev., 1992, 115, 141; (b)
M. Munakata, M. Maekawa, S. Kitagawa and M. Adachi, Inorg.
Chim. Acta, 1990, 167, 181.
8 F. A. Cotton, X. Feng, M. Matusz and R. Poli, J. Am. Chem. Soc.,
1988, 110, 7077.
9 P. Jennische and R. Hesse, Acta Chem. Scand., 1971, 25, 423.
10 D. J. Cram, T. Kaneda, R. C. Helgeson and G. M. Lein, J. Am.
Chem. Soc., 1979, 101, 6752.
14 In order to test the basicity of the ligand, a reaction between
guest-free 2 and p-TsNH2 in dry DMSO was monitored by 1H
NMR spectroscopy (pKa = 15.58 in DMSO; K. Izutsu, Acid–Base
Dissociation Constants in Dipolar Aprotic Solvents, Chemical Data
Series No. 35, IUPAC, Blackwell Scientific Publications, Oxford,
1990, p. 89). The reaction proceeded very slowly, and broad signals
appeared after 16 days. The spectrum at 100 ЊC showed well-
resolved sharp signals which correspond to 3H+⊂2, 2H+⊂2, H+⊂2
and the starting material 2. According to the result, the proton
cryptate H+⊂2 is also a strong base itself and further protonation
reactions occur.
15 B. Metz, D. Moras and R. Weiss, J. Chem. Soc., Perkin Trans. 2,
1976, 423.
16 E. Graf, J.-P. Kintzinger, J.-M. Lehn and J. LeMoigne, J. Am. Chem.
Soc., 1982, 104, 1672.
11 (a) C. O. Dietrich-Bucheker and J.-P. Sauvage, J. Am. Chem. Soc.,
1984, 106, 3043; (b) C. Dietrich-Bucheker and J.-P. Sauvage,
Tetrahedron, 1990, 46, 503; (c) J. C. Rodriguez-Ubis, B. Alpha,
D. Plancherel and J.-M. Lehn, Helv. Chim. Acta, 1984, 67, 2264.
12 In a separate experiment, compound 2 was titrated with picric acid.
Stoichiometry and shift values of the methylene signals of nH+⊂2
17 teXsan: Crystal Structure Analysis Package, Molecular Structure
Corporation (1985 and 1992).
Paper 6/04966D
Received 15th July 1996
Accepted 14th October 1996
246
J. Chem. Soc., Perkin Trans. 1, 1997