Making molecular-necklaces from rotaxanes

Article abstract of DOI:10.1016/S0040-4020(01)01110-3

During an attempt to synthesize polyrotaxanes by Wittig step-growth polymerization between a dibenzylic bis(triphenylphosphonium)-stoppered [2]rotaxane, with an NH₂⁺ recognition site encircled by a DB24C8 ring, and appropriate derivatives of terephthaldehyde carrying bulky groups, large macrocyclic compounds with the topologies of catenanes and molecular-necklaces, are formed.

Full text of DOI:10.1016/S0040-4020(01)01110-3

TETRAHEDRON
Pergamon
Tetrahedron 58 (2002) 807±814
Making molecular-necklaces from rotaxanes
Sheng-Hsien Chiu,a Stuart J. Rowan,a,² Stuart J. Cantrill,a,³ Ludek Ridvan,a Peter R. Ashton,b
Robin L. Garrella and J. Fraser Stoddarta,p
aDepartment of Chemistry and Biochemistry, University of California, Los Angeles, 405 Hilgard Avenue, Los Angeles, CA 90095-1569, USA
bSchool of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
Received 19 June 2001; revised 14 October 2001; accepted 15 October 2001
AbstractÐDuring an attempt to synthesize polyrotaxanes by Wittig step-growth polymerization between a dibenzylic bis(triphenylpho-
sphonium)-stoppered [2]rotaxane, with an NH21 recognition site encircled by a DB24C8 ring, and appropriate derivatives of terephthalde-
hyde carrying bulky groups, large macrocyclic compounds with the topologies of catenanes and molecular-necklaces, are formed. q 2002
Elsevier Science Ltd. All rights reserved.
1. Introduction
molecular machines.9 Indeed, appropriately crafted [2]-
rotaxanes have provided the basis for constructing molec-
ular shuttles10 and then subsequently turning them into
molecular switches11 that can ultimately be fabricated into
molecular electronic devices.12
Interlocked molecules1 are comprised of two or more
mechanically bonded components that are not covalently
linked to each other. The applications of the concept of
self-assembly2 and the fundamental principles of molecular
recognition3 in the context of supramolecular assistance to
covalent synthesis4 has led to highly ef®cient template-
directed procedures5 for constructing interlocked molecular
compounds. RotaxanesÐmolecules in which one or more6
macrocyclic components are trapped around (a) recognition
site(s) along the rod section of a dumbbell-shaped compo-
nent by bulky terminal groups which are usually referred to
as stoppersÐrepresent one of the two major subsets7 of
such compounds. In recent times, [2]rotaxanes6Ðand
indeed catenanes7 as wellÐhave aroused a lot of interest
in the wider scienti®c community simply because the
molecular recognition that exists between their two compo-
nents can be turned off and on at will by chemical or elec-
trochemical or photochemical means,8 thus making it
possible for the components to be stimulated to move
relative to each other in a highly controlled manner. The
fact that we have learned how to control the relative move-
ments of the components in appropriately self-assembled
[2]rotaxanes6 by switching off and on noncovalent bonds
between the macrocyclic and dumbbell-shaped components
has inspired the design and synthesis of numerous arti®cial
The successful performance of rotaxane molecules as
switches and devices at the lower end of the nanometer
scale prompts the obvious questionÐcould such well-
de®ned interlocking and the possibility of controlling
component recognition and geometry also have a profound
effect further up the nanometer scale in the macromolecular
world? The only way to answer this question is to make
interlocked macromolecules. Consequently, the progression
from rotaxanes to polyrotaxanes13 has been a rapid one,
resulting in the design and construction of a diverse range
of interlocked macromolecular compounds.14
The post-assembly modi®cation of a triphenylphospho-
nium-stoppered [2]rotaxane,15 using conventional Wittig
chemistry, has been employed recently by us,16 not only
to convert this `reactive' [2]rotaxane, without any loss of
its mechanical integrity, into other `inert' [2]rotaxanes, but
also to construct, with complete supramolecular control, a
degenerate molecular shuttle as well as a branched [4]rotax-
ane. The principle that is being exercised and exploited in
this chemistry is shown schematically in Fig. 1 for the
Keywords: molecular-necklaces; molecular recognition; rotaxanes; self-
assembly; Wittig reaction.
p
Corresponding author. Tel.: 11-310-206-7078; fax: 11-310-206-1843;
e-mail: stoddart@chem.ucla.edu
Present address: Department of Macromolecular Science, Case Western
²
Reserve University, 2100 Adelbert Road, Cleveland, OH 44106-7202,
USA.
Present address: Division of Chemistry and Chemical Engineering,
³
California Institute of Technology, 1200 East California Boulevard,
Pasadena, CA 91125, USA.
Figure 1. A schematic representation of the conversion of a `reactive'
[2]rotaxane into an `inert' [2]rotaxane.
0040±4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)01110-3

808
S.-H. Chiu et al. / Tetrahedron 58 (2002) 807±814
Figure 2. Monomers for the proposed syntheses of polyrotaxanes by Wittig
step-growth polymerization. The AB monomer is designed to form daisy-
chain polyrotaxanes whereas the AA and BB monomers should form linear
polyrotaxanes carrying regularly spaced entrapped macrocycles.
Figure 4. Schematic representations of the [3]catenane IV and the molec-
ular-necklaces V and VI formed upon reaction of the AA and BB
monomers at low concentrations.
conversion of a `reactive' [2]rotaxane into an `inert' one.
Essentially, it involves the replacement of a surrogate
stopper in the `reactive' [2]rotaxane I to afford a new
`inert' [2]rotaxane III. The judicious selection of an appro-
priate exchange reactionÐone, like the Wittig reaction,
where a concerted mechanism operatesÐensures the
integrity of the mechanical bond throughout the transforma-
tion, i.e. in such a process, the intermediate II retains bulky
stoppers at both ends of the dumbbell-shaped component.
With the idea of extending the Wittig chemistry into a
macromolecular setting, three monomers (Fig. 2) have
been synthesized15,17,18 that should allow access to at least
two different polyrotaxane architectures. The AB monomer
was designed to form daisy-chain19 polyrotaxanes,18
whereas the AA and BB monomers should form linear poly-
rotaxanes carrying regularly spaced, entrapped macro-
cycles. We were well aware of the possibility, however,
particularly at relatively low monomer concentrations, of
obtaining cyclic, as well as acyclic, products from the
proposed reactions (Fig. 3). The reactions of AB monomers
to give both cyclic and acyclic daisy-chain19 oligomers have
been described and discussed elsewhere.17,18 Here, we
report the outcome of reactions between the AA and BB
monomers, showing that, while the major product is indeed
polymeric material, at high monomer concentrations (typi-
cally 100 mM or greater), side products are also formed at
lower monomer concentrations (10 mM or less, typically).
2. Results and discussion
2.1. Synthetic strategy
In our quest to synthesize polyrotaxanes by Wittig step-
growth polymerization, we sought to identify a dialdehyde
that would also act as a stopper following the ®rst of the two
consecutive Wittig reactions, otherwise the macrocycles
would simply be released from the rods to become free in
solution.16,21 For reasons associated with its symmetry and
steric bulkÐand also the ability to be able to modify it in
the products isolated at the end of the reactionÐwe chose to
employ (Fig. 2) the benzyloxymethoxy (BOM) protected
dialdehyde BB for reaction with the bis(triphenylphospho-
nium)-stoppered [2]rotaxane AA already reported by us.15
2.2. Isolation and characterization of a [4]molecular-
necklace
The BOM-protected dialdehyde BB was prepared (Scheme
1) from diethyl 2,5-dihydroxyterephthalate in 72% overall
yield in three steps: (i) protection (BOMCl/NaH/DMF) of
the two phenolic hydroxyl groups to give 1, (ii) reduction
(LiAlH4/THF) of 1 to afford the diol 2, and (iii) Swern
oxidation of 2 to yield the monomer BB. The presence of
two BOM protecting groups in the monomer assisted greatly
in both the separation and subsequent puri®cation of the
Wittig reaction productsÐand also in their eventual charac-
terization where the appearance of two singlets for the
benzylic methylene and dioxymethylene protons in 1H
NMR spectra is indicative or otherwise of the formation
of cyclic products following reduction of the carbon±carbon
double bonds formed during the Wittig ole®nations. At a
These observations demonstrate that a `reactive' [2]rotax-
ane can be converted (Fig. 4) into a [3]catenane IV and
mechanically-interlocked compounds, e.g. V and VI,
reminiscent of molecular-necklaces.20
Figure 3. Schematic representations depicting the possible cyclic or acyclic
components that can result upon Wittig step-growth polymerization of:
(a) the AB monomer to give either cyclic or acyclic daisy-chains, and
(b) the AA and BB monomers to give either: (i) catenanes and molec-
ular-necklaces, or (ii) linear polyrotaxanes, bearing regularly spaced,
1
entrapped macrocycles. Note that the NH2 centers are generated in the
products following acidic work-ups.
Scheme 1. The synthesis of the BB monomer starting from diethyl 2,5-
dihydroxyterephthalate.

S.-H. Chiu et al. / Tetrahedron 58 (2002) 807±814
809
Scheme 2. The synthesis and isolation procedure employed to separate the BOM-protected [4]molecular-necklace 4-3H´3PF6 from the polymeric material.
concentration of 9 mM, the monomer BB was added
(Scheme 2) to a mixture of 1 equiv. of the AA monomer
in CH2Cl2 in the presence of an excess of NaH as base. The
reaction mixture was stirred at room temperature for one
week. After work-up and counterion exchange (NH4PF6/
H2O), a yellow solid was obtained. Since it was not improb-
able that some macrocyclization might have accompanied
polymerization, it was expected that the resulting product
might contain [3]catenanes and molecular-necklaces, in
addition to oligo- and polyrotaxanes, all as complex
mixtures of diastereoisomers because of the existence of
E/Z isomerism associated with all the newly formed
carbon±carbon double bonds. Indeed, the identi®cation of
peaks at m/z 2235 and 3423 for [M2PF6]1 ions in the fast-
atom bombardment (FAB) mass spectrum of this yellow
solid suggested the presence of [3]catenanes and [4]molec-
ular-necklaces, respectively. Isolation of these macrocycles
directly from this reaction mixture by HPLC was imprac-
tical, not only because they are minor products, but also
because of their existence in numerous diastereoisomeric
forms. However, in principle, the isomeric mixtures can
all be converted into single macrocycles by hydrogenation
of their ole®nic bonds. Hydrogenation employing H2 and
Pd/C in THF, however, was found to result in the concomi-
to cleavage of the backbones of the polyrotaxanes and also
of the macrocycles associated with the [3]catenanes and
molecular-necklaces. This competing reaction was evi-
denced by the large amount of free DB24C8 that could be
isolated from reaction mixtures. Adams' catalyst (PtO2) was
much more selective. Hydrogenolysis did not become a
serious competitor, unless samples were hydrogenated for
long periods (.6 h) of time. Also, hydrogenation did not
proceed well without chromatographic puri®cation of the
crude reaction mixture on silica gel. By employing FAB
mass spectrometry in a screening capacity, we found that
column chromatography on SiO2 using MeOH/CH2Cl2
(2:98) as eluant is the most reliable way to separate cyclic
products from most of the polymeric material. The less polar
[3]catenanes and molecular-necklaces, and some oligomeric
material, were isolated with this eluent. The more polar
products, which were eluted with MeOH/MeCN (1:1) and
then with pure MeCN, were found to contain none of the
macrocycles. The yellow solid, which was isolated from the
less polar fraction was dissolved in THFand hydrogenated
over Adams' catalyst. Since the product at this stage was
still a horrendous mixture of macrocycles and cyclic
oligomers, the `end-point' of the hydrogenation was not
easy to recognize either by TLC or by 1H NMR
spectroscopy. We discovered, however, that provided
1
tant hydrogenolysis of benzylic CH2±NH2 bonds, leading

810
S.-H. Chiu et al. / Tetrahedron 58 (2002) 807±814
following ¯ash chromatography on silica gel, contains
compounds of a much smaller molecular weight than
4-3H´3PF6 and so this BOM-protected [4]molecular-
necklace can be separated easily from the remainder of
the less polar fraction by GPC since it is the ®rst compound
to be eluted from the column.
Although we devoted a lot of effort trying to isolate other
cyclic homologues, only a trace amount of a compound,
which was eluted from the silica gel column before
4-3H´3PF6 was isolated in just suf®cient quantities
1
(,1 mg) to obtain a H NMR spectrum and a high reso-
lution FAB mass spectrum. These data are consistent with
this compound being the BOM-protected [3]catenane
3-2H´2PF6. Additionally, the presence of a triply charged
peak at m/z 1447 for [M23PF6]31 in the electrospray mass
spectrum of the crude hydrogenation product suggests the
existence of some of the homologous [5]molecular-necklace
5-4H´4PF6. The amount present in the crude product was
very small, however, and all our efforts to isolate it and
characterize it have been unsuccessful.
Figure 5. GPC Traces (Styragel column, CH2Cl2) of: (a) the less-polar
fraction following hydrogenation (see Scheme 2) and (b) the BOM-
protected [4]molecular-necklace 4-3H´3PF6 isolated after silica gel
chromatography.
hydrogenation is terminated right after the solution turns
colorless from its original bright yellow, then undesired
hydrogenolysis is minimized. After work-up was complete,
FAB mass spectrometry suggested that the BOM-protected
[4]molecular-necklace was the major cyclic product. It was
isolated as a pure compound in 4% yield22 following
chromatography on SiO2 using MeCN/CH2Cl2 (1:9) as
eluent. GPC traces of the less polar fraction, following
hydrogenation, and the pure BOM-protected [4]molecular-
necklace 4-3H´3PF6, isolated after column chromatography
on silica gel are shown in Fig. 5. The less polar fraction,
The relative simplicity of the 1H NMR spectrum (Fig. 6a) of
4-3H´3PF6 is consistent with the averaged D3h symmetry of
the BOM-protected [4]molecular-necklace.23 As expected,
the signals for the protons of the bismethylene groups that
result from hydrogenation of ole®nic double bonds appear
as a symmetrical multiplet in the range d 2.70±2.85. The
multiplets for the a-, b- and g-OCH2 protons on the
DB24C8 ring, which are centered on d 3.90, 3.60, and
3.43, respectively, are shifted signi®cantly up®eld from
Figure 6. 1H NMR spectra (400 MHz, CD3CN, 298 K) of: (a) the BOM-protected [4]molecular-necklace 4-3H´3PF6, (b) the [4]molecular-necklace
6-3H´3PF6, and (c) the permethylated [4]molecular-necklace 7-3H´3PF6.

S.-H. Chiu et al. / Tetrahedron 58 (2002) 807±814
811
and O-methylene protons, respectivelyÐand in the range d
7.21±7.31 for the phenyl ring protons.
The BOM protecting groups in 4±3H´3PF6 were cleaved
(Scheme 3) by heating a solution of the [4]molecular-
necklace in 5% HCl(aq)/THF(5:95) under re¯ux overnight.
The fully deprotected [4]molecular-necklace 6-3H´3PF6
was isolated (Scheme 3) in 70% yield after counterion
1
exchange (NH4PF6/H2O). The H NMR spectrum (Fig. 6b)
of 6-3H´3PF6 is similar to that (Fig. 6a) of the BOM-
protected [4]molecular-necklaceÐparticularly in relation
to the signals for the DB24C8 rings and the dibenzylammo-
nium subunits, suggesting little change if any around the
recognition sites. Apart from the disappearance of signals
for the BOM-protecting groups and the appearance of a
broad singlet at d 6.19 for the hydroxyl group protons,
there is an up®eld shift of the sharp singlet for the
hydroquinone ring protons from d 6.88 (protected) to 6.46
(deprotected), suggesting that deprotection occurs cleanly
without damage to the molecules. However, the deprotected
[4]molecular-necklace is not all that stable: freshly prepared
6-3H´3PF6 decomposes, even when not in solution, at
ambient temperatures in a matter of days. The gradual
1
appearance of signals for free DB24C8 in the H NMR
spectrum suggests that the large ammonium-ion-containing
macrocycle is being cleaved during the decomposition
process which we suspect might involve initially oxidation
of the substituted hydroquinone rings. This process has not
been investigated any further.
Scheme 3. The deprotection of the BOM-protected [4]molecular-necklace
4±3H´3PF6 to give the [4]molecular-necklace 6-3H´3PF6-and the structural
formula of the permethylated [4]molecular-necklace 7-3H´3PF6.
their d values (4.24, 3.75, and 3.61, respectively) in free
DB24C8, in keeping with the rings being mechanically
1
2.3. Synthesis of the permethylated [4]molecular-
necklace and its characterization
interlocked around the NH2 centers to which they are
hydrogen bonded by means of [N1±H´´´O], and, presu-
mably also, [C±H´´´O] interactions.24 Diagnostic of the
fact that this recognition motif is operative25 is the highly
characteristic multiplet centered on d 4.51 for the protons on
the benzylic methylene groups on both (equivalent) sides of
In order to establish the generality of the synthetic approach,
which we have discovered for making molecular-necklaces
from rotaxanes, we decided to replace the BOM-protected
dialdehyde in the synthesis with 2,5-dimethoxyterephthal-
dehyde. Using identical reaction conditions, the permethyl-
ated [4]molecular-necklace 7-3H´3PF6 was isolated in 2%
yield after following the same puri®cation procedure as
before, i.e. subjecting the crude product to ¯ash chroma-
tography, followed by hydrogenation of the appropriate
fraction and then using column chromatography to obtain
1
the NH2 centers. The singlet resonating at d 6.88 and the
AA0BB0 system with d values of 6.93 and 7.13 arise from
protons on the hydroquinone and p-xylyl units, respectively.
The protons on the catechol units of the DB24C8 ring
appear as an AA0BB0 system in the range of d from 6.60
through to 6.80. The BOM-protecting groups are contri-
buting singlets at d 4.85 and 5.44 for the benzylic methylene
1
7-3H´3PF6 as a pure compound (Scheme 3). The H NMR
spectrum (Fig. 6c) of 7-3H´3PF6 is similar to that (Fig. 6a) of
4-3H´3PF6 except that signals for the BOM group protons
are absent and replaced by a singlet for the methoxy protons
at d 3.66. There is also an up®eld shift of the singlet (d 6.67)
for the hydroquinone ring protons relative to that at d 6.88 in
the spectrum of the BOM-protected [4]molecular-necklace.
The FAB mass spectrum (Fig. 7) of 7-3H´3PF6 revealed
intense peaks at m/z 2508, 1328, and 1255 corresponding
to [M22H23PF6]1, [M22PF6]21, and [M2H2 3PF6]21
,
respectively.
3. Conclusions
Although we set out to try and convert a [2]rotaxane with
two potentially reactive benzylic triphenylphosphonium
stoppers into polyrotaxanes using Wittig step-growth poly-
merizations with bulky aromatic dialdehydes, we ended up
isolating small amounts of large interlocked macrocyclesÐ
Figure 7. FAB mass spectrum of the permethylated [4]molecular-necklace
7-3H´3PF6.

812
S.-H. Chiu et al. / Tetrahedron 58 (2002) 807±814
in particular, a protected [4]molecular-necklace derivative.
While the overall yields of these large macrocycles contain-
ing localized catenated rings are small, the chemistryÐ
rather complicated, as it is, from beginning to endÐ
works amazingly well. At this stage, we can claim proof
of principle for a synthetic procedure that involves recog-
nition-directed assembly early on and retains molecular
recognition features throughout for making molecular-
necklaces out of rotaxanes.
portions during 1 h to diethyl 2,5-bis(benzyloxymethoxy)-
terephthalate (1) (1.45 g, 2.9 mmol) disssolved in THF
(30 mL). The reaction mixture was heated under re¯ux for
3 h and stirred at room temperature for 18 h. After the reac-
tion had been quenched by slowly adding H2O (10 mL),
anhydrous MgSO4 (10 g) was added. The mixture was
®ltered and concentrated. Puri®cation by column chroma-
tography (SiO2) using MeOH/CH2Cl2 (1:19) as eluent gave
2,5-bis(benzyloxymethoxy)-1,4-bis(hydroxymethyl)benzene
1
(2) as a white solid (1.05 g, 88%). H NMR (400 MHz,
CD3CN): d 3.15 (t, Jˆ5.9 Hz, 2H), 4.57 (d, Jˆ5.9 Hz,
4H), 4.71 (s, 4H), 5.28 (s, 4H), 7.17 (s, 2H), 7.33 (s,
10H); 13C NMR (100 MHz, CD3CN): d 60.0, 71.0, 94.2,
115.5, 128.7, 128.9, 129.4, 131.8, 138.9, 150.1; HRMS
(FAB): calcd 410.1729 for C24H26O6 found: 410.1736.
4. Experimental
4.1. Methods and materials
All glassware, stirrer-bars, syringes, and needles were either
oven- or ¯ame-dried prior to use. Reactions were carried out
under a N2 or Ar atmosphere. All starting materials and
reagents, unless otherwise indicated, were purchased from
commercial sources. Anhydrous CH2Cl2 and MeCN were
obtained by distillation from CaH2 under N2, and anhydrous
4.1.3. 2,5-Dibenzyloxymethoxyterephthaldehyde BB. To
a round-bottomed ¯ask charged with oxalyl chloride
(0.74 mL, 8.47 mmol) and CH2Cl2 (25 mL) and cooled
down to 2788C was added DMSO (1.0 mL, 14.0 mmol)
and the reaction mixture was stirred at 2788C for 30 min.
2,5-Bis(benzyloxymethoxy)-1,4-bis(hydroxymethyl)benzene
(2) (0.9 g, 2.2 mmol) in CH2Cl2 (15 mL) was added slowly
into the reaction mixture using a syringe. The reaction
mixture was stirred at 2788C for 40 min. Triethylamine
(3.5 mL, 25 mmol) was then added to the reaction mixture
and it was slowly warmed up to room temperature. The
mixture was partitioned between CH2Cl2 (200 mL) and
H2O (200 mL). The organic layer was separated and washed
with H2O (3£100 mL), dried (MgSO4) and concentrated.
Puri®cation by column chromatography (SiO2) using
EtOAc/hexane (3:17) as eluent gave 2,5-dibenzyloxy-
methoxyterephthaldehyde BB as a light yellow solid
THFby distillation from Na/Ph CO under N2. Thin layer
2
chromatography (TLC) was performed on Analtech 0.25-
mm silica gel GHLF. Column chromatography was carried
out using silica gel 60F(Merck 9385, 0.040±0.063 mm).
GPC spectra were recorded using a Waters Styragel column
HR3, 7.8£300 mm2 with CH2Cl2 as the eluent at 258C at a
¯ow rate of 1.0 mL/min. 1H NMR spectra were recorded at
360 or 400 MHz. The deuterated solvent was used as the
lock while either TMS or the solvent's residual protons were
employed as the internal standard. Chemical shifts are
reported as d values in parts per million. Multiplicities are
given as s (singlet), d (doublet), t (triplet), q (quartet), m
(mutiplet), and br (broad). Coupling constants are in Hertz.
13C NMR spectra were recorded at 90 or 100 MHz. Fast-
atom bombardment (FAB) mass spectrometry was per-
formed using either a Kr or Cs primary atom beam in
conjunction with a m-nitrobenzyl alcohol matrix.
1
(0.85 g, 95%). H NMR (400 MHz, CD3CN): d 4.74 (s,
4H), 5.44 (s, 4H), 7.29 (s, 10H), 7.60 (s, 2H), 10.37 (s,
2H); 13C NMR (100 MHz, CD3CN): d 71.8, 94.3, 115.3,
128.8, 129.0, 129.4, 131.0, 138.5, 154.7, 190.0; HRMS
(FAB): calcd 406.1416 for C24H22O6 found: 406.1411.
4.1.4. BOM-protected interlocked molecular compounds
3-2H´2PF6, 4-3H´3PF6, 5-4H´4PF6 and oligorotaxanes.
CH2Cl2 (100 mL) was added to a round-bottom ¯ask
charged with the bis(triphenylphosphonium)-stoppered
[2]rotaxane AA monomer (1.87 g, 1.15 mmol) and NaH
(0.11 g, 4.58 mmol). 2,5-Bis(benzyloxymethoxy)terephth-
aldehyde, the BB monomer (0.45 g, 1.11 mmol), dissolved
in CH2Cl2 (20 mL), was added via an addition funnel during
24 h to the reaction ¯ask. The reaction mixture was stirred at
room temperature for 7 days. It was quenched by the
slow addition of MeOH (5 mL), and then the solution
was concentrated under reduced pressure. The solid was
suspended in MeCN and saturated aqueous NH4PF6 was
added until the solid dissolved. The MeCN was then evapo-
rated and the resulting precipitate was collected and washed
with H2O. The yellow residue was subjected to ¯ash
chromatography (SiO2; MeOH/CH2Cl2 (2:98)) and the
yellow, less polar fraction was collected and concentrated
down to a yellow solid (259 mg). The more polar fraction
was eluted with MeOH/CH2Cl2 (1:1) and was concentrated
to give a yellow solid. Oligorotaxanes were obtained by
Soxhlet extraction of the more polar yellow solid with
MeOH for 3 days, before drying the solid under vacuum
(0.68 g, ca. 40%). The less polar, yellow solid was then
4.1.1. Diethyl 2,5-bis(benzyloxymethoxy)terephthalate
(1). Diethyl 2,5-dihydroxyterephthalate (2.0 g, 7.9 mmol)
and NaH (0.56 g, 23 mmol) were cooled down to 08C in a
round-bottomed ¯ask. DMF(50 mL) was added slowly and
the reaction mixture was stirred at 08C for 1 h. Benzyloxy-
methyl chloride (3.6 mL) was added to the reaction mixture
dropwise via syringe during 30 min. The solution was
slowly warmed up to room temperature and stirred for
18 h. MeOH (3 mL) was added to quench the reaction.
The mixture was partitioned between CH2Cl2 (200 mL)
and H2O (200 mL). The organic layer was separated and
washed with 5% HCl(aq) (3£100 mL), dried (MgSO4) and
concentrated. Puri®cation by column chromatography
(SiO2) using EtOAc/hexane (1:4) as eluent gave diethyl
2,5-bis(benzyloxymethoxy)terephthalate (1) (3.35 g, 86%).
1H NMR (400 MHz, CD3CN): d 1.32 (t, Jˆ7.1 Hz, 6H),
4.31 (q, Jˆ7.1 Hz, 4H), 4.73 (s, 4H), 5.29 (s, 4H), 7.32 (s,
10H), 7.55 (s, 2H); 13C NMR (100 MHz, CD3CN): d 14.6,
62.3, 71.1, 94.1, 120.4, 127.4, 128.8, 128.9, 129.3, 138.5,
151.1, 166.1; HRMS (FAB): calcd 494.1941 for C28H30O8
found: 494.1940.
4.1.2. 2,5-Bis(benzyloxymethoxy)-1,4-bis(hydroxymethyl)-
benzene (2). LiAlH4 (1.1 g, 28.9 mmol) was added in small

S.-H. Chiu et al. / Tetrahedron 58 (2002) 807±814
813
dissolved in THF(10 mL), PtO 2 (30 mg) was added and H2
gas was bubbled through the solution for 10 min. The
mixture was then stirred under H2 (balloon) until the solu-
tion became colorless, whereupon it was concentrated under
reduced pressure. The solid was suspended in MeCN and
saturated aqueous NH4PF6 was added until the solid
dissolved. The MeCN was evaporated and the resulting
precipitate was collected, washed with H2O and puri®ed
by column chromatography (SiO2; MeCN/CH2Cl2 (1:9)).
The BOM-protected [4]molecular-necklace 4-3H´3PF6
was obtained as a white solid (49 mg, 4%); 1H NMR
(400 MHz, CD3CN): dˆ2.70±2.80 (m, 24H), 3.43 (s,
24H), 3.60 (m, 24H), 3.90 (m, 24H), 4.53 (m,12H), 4.85
(s, 12H), 5.44 (s, 12H), 6.60±6.70 (m, 12H), 6.70±6.80
(m, 12H), 6.88 (s, 6H), 6.93 (d, Jˆ8 Hz, 12H), 7.12 (d,
Jˆ8 Hz, 12H), 7.20±7.40 (br, 36H); (400 MHz, CD2Cl2):
d 2.83 (s, 24H), 3.38 (s, 24H), 3.65 (s, 24H), 4.00 (s, 24H),
4.51 (s,12H), 4.70 (s, 12H), 5.20 (s, 12H), 6.65±6.75 (m,
12H), 6.75±6.85 (m, 12H), 6.91 (s, 6H), 7.01 (d, Jˆ8 Hz,
12H), 7.13 (d, Jˆ8 Hz, 12H), 7.30±7.50 (br, 36H); 13C
NMR (100 MHz, CD3CN): d 31.8, 35.9, 53.0, 68.8, 70.9,
71.0, 71.4, 94.7, 113.4, 122.2, 128.7, 129.4, 129.5, 130.3,
130.4, 130.5, 138.8, 143.8, 148.4, 151.0; (100 MHz,
CD2Cl2): d 31.6, 35.7, 52.7, 68.6, 70.5, 70.9, 71.0, 94.1,
113.0, 117.2, 121.9, 128.2, 128.8, 129.1, 129.6, 129.6,
129.8, 138.0, 143.5, 147.9, 150.5; HRMS (FAB): calcd
3437.4984 for C192H222N3O36P2F12 [M2PF6]1; found:
3437.4939. The BOM-protected [3]catenane 3-2H´2PF6
was isolated from the fraction that eluted before the
[4]molecular-necklace. Data for 3-2H´2PF6: 1H NMR
(360 MHz, CD3CN): d 2.70±2.80 (br, 16H), 3.34 (s,
16H), 3.52 (m, 16H), 3.87 (m, 16H), 4.50 (m, 8H), 4.68
(s, 8H), 5.44 (s, 8H), 6.60±6.70 (m, 8H), 6.70±6.80 (m,
8H), 6.91 (s, 4H), 6.94 (d, Jˆ8 Hz, 8H), 7.12 (d, Jˆ8 Hz,
8H), 7.20±7.40 (br, 24H); HRMS (FAB): calcd 2243.0097
for C128H148N2O24PF6 [M2PF6]1; found: 2243.0106. The
existence of the [5]molecular-necklace 5-4H´4PF6 was
con®rmed by electrospray mass spectrometry of the hydro-
genated less polar fraction. Data for 5-4H´4PF6: MS
ane AA and 2,5-dimethoxyterephthaldehyde under identical
experimental conditions to those reported earlier for
compound 4-3H´3PF6. Data for 7-3H´3PF6: 1H NMR
(400 MHz, CD3CN): d 2.75 (s, 24H), 3.43 (s, 24H), 3.67
(m, 42H), 3.93 (m, 24H), 4.54 (m, 12H), 6.67 (s, 6H), 6.67±
6.73 (m, 12H), 6.75±6.85 (m, 12H), 6.96 (d, Jˆ8 Hz, 12H),
7.15 (d, Jˆ8 Hz, 12H), 7.35 (br, 6H); 13C NMR (100 MHz,
CD3CN): d 31.8, 35.6, 53.1, 56.7, 68.9, 71.0, 71.4, 113.5,
114.1, 122.2, 128.8, 129.5, 130.3, 130.4, 144.0, 148.5,
152.3; MS (FAB): m/z 2508 [M22H23PF6]1, 2206 [M2
H22PF62DB24C8]1, 2060 [M22H23PF62DB24C8]1,
1611 [M22H23PF622DB24C8]1, 1328 [M22PF6]21
,
1255 [M2H23PF6]21, 1030 [M2H23PF62DB24C8]21
.
Acknowledgements
We thank the Petroleum Research Fund, administered by
the American Chemical Society, for generous ®nancial
support.
References
1. (a) Catenanes, Rotaxanes and Knots; Schill, G., Ed.;
Academic: New York, 1971; p 23. (b) Molecular Catenanes,
Rotaxanes and Knots; Sauvage, J.-P., Dietrich-Buchecker, C.,
Eds.; VCH-Wiley: Weinheim, 1999. (c) Amabilino, D. B.;
Stoddart, J. F. Chem. Rev. 1995, 95, 2725±2828. (d) Breault,
G. A.; Hunter, C. A.; Mayers, P. C. Tetrahedron 1999, 55,
5265±5293.
2. (a) Lindsey, J. S. New J. Chem. 1991, 15, 153±180. (b) Philp,
D.; Stoddart, J. F. Angew. Chem., Int. Ed. Engl. 1996, 35,
1155±1196. (c) Lindoy, L. F.; Atkinson, I. M. In Self-
Assembly in Supramolecular Systems; Stoddart, J. F., Ed.;
Royal Society of Chemistry: Cambridge, 2000.
3. (a) Tetrahedron Symposia-in-Print on Molecular Recognition:
Tetrahedron, 1995, 51, 343±648. (b) See Special Issue on
Molecular Recognition: Chem. Rev., 1997, 97, 1231±1734.
4. Fyfe, M. C. T.; Stoddart, J. F. Acc. Chem. Res. 1997, 30, 393±
401.
(Electrospray): 1447 for [M23PF6]31
.
4.1.5. [4]Molecular-necklace 6-3H´3PF6. The BOM-
protected [4]molecular-necklace 4-3H´3PF6 (50 mg, 14
mmol) was dissolved in 5% HCl(aq)/THF(1:19; 5 mL) and
the solution was heated under re¯ux overnight. The mixture
was partitioned between CH2Cl2 (50 mL) and H2O (50 mL).
The organic phase was dried (MgSO4) and concentrated.
The residue was puri®ed by column chromatography
(SiO2; MeOH/CH2Cl2 (1:19)) and the [4]molecular-neck-
lace 6-3H´3PF6 (30 mg, 70%) was recovered as a white
solid; 1H NMR (400 MHz, CD3CN): d 2.60±2.80 (m,
24H), 3.43 (m, 24H), 3.67 (m, 24H), 3.97 (m, 24H), 4.54
(m, 12H), 6.19 (s, 6H), 6.46 (s, 6H), 6.70±6.80 (m, 12H),
6.80±6.90 (m, 12H), 6.96 (d, Jˆ8 Hz, 12H), 7.16 (d, Jˆ
8 Hz, 12H), 7.40±7.50 (br, 6H); 13C NMR (100 MHz,
CD3CN): d 31.6, 35.5, 53.1, 68.9, 71.0, 71.4, 113.5,
117.5, 122.2, 126.8,129.5, 130.3, 130.4, 143.9, 148.5,
148.6; MS (FAB): m/z 2716 [M2PF6]1, 2570 [M2H2
2PF6]1.
5. (a) Anderson, S.; Anderson, H. L.; Sanders, J. K. M. Acc.
È
Chem. Res. 1993, 26, 469±475. (b) Hoss, R.; Vogtle, F.
Angew. Chem., Int. Ed. Engl. 1994, 33, 375±384. (c) Hubin,
T. J.; Kolchinski, A. G.; Vance, A. L.; Busch, D. L. Adv.
Supramol. Chem. 1999, 5, 237±357.
6. The simplest rotaxane is a [2]rotaxane±a molecule composed
of a macrocyclic component which encircles a dumbbell-
shaped component.
7. The other well-known example of a mechanically interlocked
molecule is a catenaneÐa molecule composed of two or more
interlocked macrocyclic components.
8. Ballardini, R.; Balzani, V.; Credi, A.; Gandol®, M. T.;
Venturi, M. Acc. Chem. Res. 2001, 34, 445±455.
9. Balzani, V.; Credi, A.; Raymo, F. M.; Stoddart, J. F. Angew.
Chem., Int. Ed. 2000, 39, 3348±3391.
10. (a) Anelli, P.-L.; Spencer, N.; Stoddart, J. F. J. Am. Chem. Soc.
1991, 113, 5131±5133. (b) Anelli, P.-L.; Asakawa, M.;
Â
Ashton, P. R.; Bissell, R. A.; Clavier, G.; Gorski, R.; Kaifer,
A. E.; Langford, S. J.; Mattersteig, G.; Menzer, S.; Philp, D.;
Slawin, A. M. Z.; Spencer, N.; Stoddart, J. F.; Tolley, M. S.;
Williams, D. J. Chem. Eur. J. 1997, 3, 1113±1135.
4.1.6. Permethylated [4]molecular-necklace 7-3H´3PF6.
The permethylated [4]molecular-necklace 7-3H´3PF6 was
isolated in 2% yield from the macrocyclization of [2]rotax-
11. (a) Ashton, P. R.; Ballardini, R.; Balzani, V.; Credi, A.; Dress,

814
S.-H. Chiu et al. / Tetrahedron 58 (2002) 807±814
R.; Ishow, E.; Kocian, O.; Preece, J. A.; Spencer, N.; Stoddart,
J. F.; Venturi, M.; Wenger, S. Chem. Eur. J. 2000, 6, 3558±
3574. (b) Brouwer, A. M.; Frochot, C.; Gatti, F. G.; Leigh,
D. A.; Mottier, L.; Paolucci, F.; Rof®a, S.; Wurpel, G. W. H.
Science 2001, 291, 2124±2128.
DB24C8 ring dissociates from what is effectively a semirotax-
1
ane because, not only is the NH2 recognition site switched
off, but also the formyl-terminated phenylene ring of the
half-dumbbell component is not large enough to prevent
dethreading.
22. At face value, a 4% overall yield of the BOM-protected
[4]molecular-necklace, starting from the AA and BB mono-
mers, is nothing to write home about. However, if we assume
that a total of 12 reactions (®ve inter- and one intramolecular
Wittig ole®nations, followed by six carbon±carbon double
bond hydrogenations) are involved, then a back-of-the-
envelope calculation would suggest that an average yield of
76% is being attained in each of these dozen reactions. The
challenge remains, howeverÐhow do we do better? The
possibility that the proportions of macrocycles, as distinct
from polymeric materials, could be increased by carrying
out the multiple Wittig reactions at even lower reactant
concentrations, but such conditionsÐnamely, 7 and even
2 mMÐwere employed without there being an improvement
in the yields of 3-2H´2PF6, 4-3H´3PF6, and 5-4H´4PF6. Also,
when a reaction was carried out with the AA and BB
monomers, using initial monomer concentrations of 100 mM
each, FAB mass spectrometry of the crude reaction mixture
indicated the presence of more high molecular weight compo-
nentsÐpresumably acyclic oligomersÐand considerable less
of the macrocycles. So far, all our numerous attempts to
isolate high molecular weight polyrotaxanes of low polydis-
persities and uniform primary structures have been to no avail.
So instead of trying to hydrogenate the ole®nic double bonds
with higher molecular weight fractionsÐfor even a very small
amount of concomitant hydrogenolysis of the repeating
benzylic ammonium functions is going to be extremely detri-
mental to the polymer primary structureÐwe decided to try to
simplify the situation by catalyzing the isomerization of any
cis double bonds to trans ones using iodine as the catalyst.
(For a recent example where this isomerization works well,
following a Wittig step-growth polymerization, see: Star, A.;
Stoddart, J. F.; Steuerman, D.; Diehl, M.; Boukai, A.; Wong,
E. W.; Yang, X.; Chung, S.-W.; Choi, H.; Heath, J. R. Angew.
Chem., Int. Ed. 2001, 40, 1721±1725. However, the high
molecular weight material was not soluble in the preferred
solvents (PhMe or PhH) and in those solvents (e.g. CH2Cl2)
or solvent mixtures (e.g. CH2Cl2/THF) where it was soluble,
isomerization did not occur.
12. (a) Collier, C. P.; Wong, E. W.; Belohradsky, M.; Raymo,
F. M.; Stoddart, J. F.; Kuekes, P. J.; Williams, R. S.; Heath,
J. R. Science 1999, 285, 391±394. (b) Wong, E. W.; Collier,
C. P.; Belohradsky, M.; Raymo, F. M.; Stoddart, J. F.; Heath,
J. R. J. Am. Chem. Soc. 2000, 122, 5831±5840. (c) Collier,
C. P.; Mattersteig, G.; Wong, E. W.; Luo, Y.; Beverly, K.;
Sampaio, J.; Raymo, F. M.; Stoddart, J. F.; Heath, J. R.
Science 2000, 289, 1172±1175. (d) Pease, A. R.; Jeppesen,
J. O.; Stoddart, J. F.; Luo, Y.; Collier, C. P.; Heath, J. R. Acc.
Chem. Res. 2001, 34, 433±444.
13. (a) Amabilino, D. B.; Parsons, I. W.; Stoddart, J. F. Trends
Polym. Sci. 1994, 2, 146±152. (b) Harada, A. Mater. Sci.
Technol. 1999, 20, 485±512. (c) Park, K.-M.; Heo, J.; Roh,
S.-G.; Jeon, Y.-M.; Whang, D.; Kim, K. Mol. Cryst. Liq.
Cryst. Sci. Technol., Sect. A 1999, 327, 65±70. (d) Raymo,
F. M.; Stoddart, J. F. In Supramolecular Polymers; Ciferri, A.,
Ed.; Marcel Dekker: New York, 2000; pp 323±357.
(e) Taylor, P. N.; O'Connell, M. J.; McNeill, L. A.; Hall,
M. J.; Aplin, R. T.; Anderson, H. L. Angew. Chem., Int. Ed.
2000, 39, 3456±3460. (f) Gibson, H. W.; Hamilton, L.;
Yamaguchi, N. Polym. Adv. Technol. 2000, 11, 791±797.
(g) Takata, T.; Kawasaki, H.; Kihara, N.; Furusho, Y.
Macromolecules 2001, 34, 5449±5456.
14. Raymo, F. M.; Stoddart, J. F. Chem. Rev. 1999, 99, 1643±
1663 and references therein.
15. Rowan, S. J.; Cantrill, S. J.; Stoddart, J. F. Org. Lett. 1999, 1,
129±132.
16. Rowan, S. J.; Stoddart, J. F. J. Am. Chem. Soc. 2000, 122,
164±165.
17. Rowan, S. J.; Cantrill, S. J.; Stoddart, J. F.; White, A. J. P.;
Williams, D. J. Org. Lett. 2000, 2, 759±762.
18. Stoddart, J. F.; Rowan, S. J.; Chiu, S.-H.; Cantrill, S. J.;
Ridvan, L.; Sivakova, S. Am. Chem. Soc., Polym. Mater.
Sci. Eng. Prepr. 2001, 84, 148±149.
19. Previously, we have recommended that the term daisy-chain is
used to describe an interwoven chainÐusually made of
¯owersÐin which each monomer unit acts as a donor and
an acceptor for a threading interaction. See: Ashton, P. R.;
Baxter, I.; Cantrill, S. J.; Fyfe, M. C. T.; Glink, P. T.; Stoddart,
J. F.; White, A. J. P.; Williams, D. J. Angew. Chem., Int. Ed.
1998, 37, 1294±1297.
23. The 1H NMR spectrum (also recorded in CD3CN) of the
BOM-protected [3]catenane 3±2H´2PF6 is almost identical
with that (Fig. 6a) of the BOM-protected [4]molecular-neck-
lace. The most signi®cant differences are evident in the signals
for the a-, b-, and g-OCH2 protons in the DB24C8 rings which
resonate as multiplets centered on d 3.87, 3.52 and 3.34Ðcf.
the d values of 3.90, 3.60 and 3.43 for 4-3H´3PF6. These
chemical shift differences are important in so far as they tell
us that the BOM-protected [4]molecular-necklace was
isolated as a pure compound and is not contaminated with
the BOM-protected [3]catenane.
20. Kim and coworkers have de®ned a molecular-necklace as a
cyclic oligorotaxane in which a number of small rings are
threaded onto a large ring. See: (a) Whang, D.; Park, K.-M.;
Heo, J.; Ashton, P. R.; Kim, K. J. Am. Chem. Soc. 1998, 120,
4899±4900. For other examples of multiring interlocked
systems, see: (b) Bitsch, F.; Hegy, G.; Dietrich-Buchecker,
C.; Leize, E.; Sauvage, J.-P.; van Dorsselaer, A. New J.
Chem. 1994, 18, 801±807. (c) Amabilino, D. B.; Ashton,
P. R.; Balzani, V.; Boyd, S. E.; Credi, A.; Lee, J. Y.; Menzer,
S.; Stoddart, J. F.; Venturi, M.; Williams, D. J. J. Am. Chem.
Soc. 1998, 120, 4295±4307.
24. Cantrill, S. J.; Pease, A. R.; Stoddart, J. F. J Chem. Soc.,
Dalton Trans. 2000, 3715±3734.
25. Ashton, P. R.; Chrystal, E. J. T.; Glink, P. T.; Menzer, S.;
Schiavo, C.; Spencer, N.; Stoddart, J. F.; Tasker, P. A.;
White, A. J. P.; Williams, D. J. Chem. Eur. J. 1996, 2, 709±
728.
21. We have shown (Ref. 16) that when unsubstituted terephthal-
dehyde is employed in Wittig-style covalent modi®cations on
benzylic triphenylphosphonium-stoppered rotaxanes, the