Size-complementary rotaxane cross-linking for the stabilization and degradation of a supramolecular network

Article abstract of DOI:10.1002/anie.201008020

Break it down: Gels formed from rotaxane cross-linkers with end groups that are size-complementary to the macrocyclic cavity of wheel components (see picture) were prepared. The network structure was maintained in polar organic solvents or in the presence of a base to prevent hydrogen bonding. Anion exchange enabled the selective and efficient de-cross-linking of the gels. Copyright

Full text of DOI:10.1002/anie.201008020

Communications
DOI: 10.1002/anie.201008020
Supramolecular Gels
Size-Complementary Rotaxane Cross-Linking for the Stabilization and
Degradation of a Supramolecular Network**
Yasuhiro Kohsaka, Kazuko Nakazono, Yasuhito Koyama, Shigeo Asai, and Toshikazu Takata*
As cross-linked polymers are widely used fundamental
materials, their chemical recycling becomes a quite important
issue.[1] A promising strategies is to utilize reversible cross-
linking based on dynamic covalent chemistry (DCC).[2,3] For
example, Chen et al. reported the efficient thermal reversi-
bility of cross-linking by Diels–Alder reaction.[4] On the other
hand, supramolecular gels formed by intermolecular inter-
actions, such as hydrogen bonds,[2] can be efficiently de-cross-
linked by specific stimuli without destruction of covalent
bonds.[2,3d,5–7] However, supramolecular gels are naturally
stable only under limited conditions that are capable of
keeping the intermolecular interactions strong. A polyrot-
axane network (PRN) is a supramolecular gel stabilized by
not only intermolecular interactions but also mechanical
restriction.[8] We previously reported a reversibly cross-link-
able PRN based on DCC, consisting of poly(crown ether)
backbone and bis(ammonium) cross-linker possessing a
disulfide bond in its center (Figure 1a).[9] The work gave a
new impulse for the chemical recycling of cross-linked
polymers, but it suffered from the disadvantage that it is a
sluggish reaction. Therefore, we have developed a novel
approach that enables the efficient de-cross-linking of PRNs
Figure 1. Strategy for de-cross-linking a PRN using a) a reversible
cleavage of disulfide bond and b) characteristics of a rotaxane cross-
link consisting of size-complementary components.
without any cleavage of covalent bonds. Recently, we have
found a novel procedure in which the axle component as a
cross-linker has an end group the size of which is comple-
mentary to the macrocycle cavity placed on the trunk polymer
(Figure 1b). The end groups provide an energy barrier to slow
the dissociation, thereby kinetically stabilizing the rotaxane
skeleton.[10] Thus, the PRN stabilizes the network structure
under normal conditions, but it can be de-cross-linked when
certain conditions, such as those that accelerate the dissoci-
ation of the rotaxane skeletons, are satisfied. Because the de-
cross-linking can be achieved without breaking the covalent
bonds, the PRN is selectively degraded so as to not damage
the trunk polymer. Furthermore, the stability and de-cross-
linking capability of PRNs can be adjusted by the size of the
end groups of the axle components. Herein we describe the
concept of novel de-cross-linkable network polymers that
utilize the size-complementary effect of the rotaxane cross-
links.
First, we investigated a model for the size-complementary
effect using crown ether/ammonium salt [2]rotaxanes 1
(Scheme 1). According to previous reports,[10–12] roxaxanes 1
with suitable end groups (R = cyclohexyl, tBu, 4-tBuC6H4)
were sufficiently stable to maintain its threaded structure
owing to both the bulky end groups and hydrogen bonds
between dibenzo [24]crown-8 ether (DB24C8) and the
ammonium group.[13] However, 1 dissociated into two parts,
axle and wheel, when the hydrogen bonds were disturbed by a
stimulus, in accordance with reported results (R = 4-
tBuC6H4).[12] Details are summarized in Table 1. As we
envisioned, the dissociation rate depends on both the size of
the end group and the kind of the external stimulus, as
discussed below.
[*] Dr. Y. Kohsaka, Dr. K. Nakazono, Dr. Y. Koyama, Prof. Dr. S. Asai,
Prof. Dr. T. Takata
Department of Organic and Polymeric Materials
Tokyo Institute of Technology
2-12-1 (H-126), O-okayama, Meguro-ku, Tokyo 152-8552 (Japan)
Fax: (+81)3-5734-2888
E-mail: ttakata@polymer.titech.ac.jp
[**] This work was financially supported by a Grant-in-Aid for Scientific
Research (No. 21106508, Soft Interfaces) from MEXT (Japan). JSPS
Fellowships for Young Scientists (Y.K.) and the Circle for the
Promotion of Science and Engineering (Y.K.) are gratefully
acknowledged. We thank Dr. Shuichi Akasaka for his help in the
dynamic mechanical analysis.
An investigation into the dissociation behavior of 1 in
dimethylsulfoxide (DMSO), a polar solvent that disturbed the
hydrogen bonds, showed the decomposition of only 1a among
three derivatives. The dissociation rate of 1a to DB24C8 and
2a obeyed first-order kinetics. The half-life (t1/2) was esti-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201008020.
4872
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4872 –4875

mated at 5.8 h.[13] A second stimulus was the addition of a base
for the neutralization of the ammonium group. The addition
of triethylamine (Et3N) caused prompt dissociation of 1a (t1/
2 = 2.6 h), even in CH3NO2 with a low donor number.[14,15]
While 1b was slowly dissociating (t1/2 = 18 d), the dissociation
of 1c was very slow to measure. It was found that 1b and 1c
are kinetically much more stable than 1a.
A third stimulus was the counteranion exchange of the
ammonium salt with a fluoride ion that markedly decreases
the acidity of the ammonium moiety. Recently, we found that
the counteranion exchange of a similar rotaxane with
tetrabutylammonium fluoride (TBAF) readily afforded the
corresponding neutral rotaxane.[16] The addition of TBAF led
to a rapid dissociation of all rotaxanes (1a–c). The dissoci-
ation of 1a was very rapid, while 1b smoothly dissociated
(t1/2 = 0.70 h). Furthermore, even the most stable rotaxane,
1c, dissociated slowly (t1/2 = 88 h). Thus, the present model
clearly indicates that the bulkiness of the end group and the
extent of the stimulus disturbing the hydrogen bonds facilitate
the dissociation, which can be easily achieved by controlling
these two variables to adjust the desired decomposition time.
Recognizing the results of the above model, we examined
the de-cross-linking capability of PRN 6 with the size-
complementary end groups at the cross-linking points. Poly-
(DB24C8) 3, with an average molecular weight Mn of 4000
and polydispersity index (PDI) of 1.35, was prepared accord-
ing to our previous report.[17] PRN 3 was treated with sec-
ammonium salts 4 in CHCl3 to give the corresponding
pseudo(polyrotaxane) 5,[17] which was cross-linked with a
diisocyanate (MDI) to quantitatively yield 6 (Scheme 2).[17b]
To equalize the degree of cross-linking, the complexation rate
of the obtained pseudorotaxanes were adjusted to about
20%.[13,17] The thermal properties of 6 are shown in Table 2.
The structures of 6a–c showed similar glass transition
temperatures (Tg,DSC) measured by differential scanning
calorimeter (DSC) at about 58C, while the 5% weight loss
Scheme 1. Dissociation reactions of model [2]rotaxane 1. Reagents and
conditions: a) [D6]DMSO, 258C, b) Et3N, CD3NO2, 608C, or
c) TBAF·3H2O, [D6]DMSO, 608C. DMSO=dimethylsulfoxide, TBAF=
tetrabutylammonium fluoride.
Table 1: Kinetic parameters for the dissociation of rotaxanes 1a–c.[a]
End
[D6]DMSO
Et3N/CD3NO2 TBAF/[D6]DMSO
608C[d] 608C[e]
koff [sÀ1 koff [sÀ1
t1/2 t1/2
group[b]
258C[c]
koff [sÀ1
]
t1/2
]
]
[f]
[f]
1a Cy
1b tBu
1c 4-tBuC6H4
3.3ꢀ10À5 5.8 h 7.3ꢀ10À5 2.6 h
–
–
[g]
–
–
4.4ꢀ10À7 18 d 2.7ꢀ10À4 0.70 h
[g]
[g]
–
2.2ꢀ10À6 88 h
[a] Determined from 1H NMR analysis. Details are given in the Support-
ing Information. [b] Cy=cyclohexyl, Bu=butyl. [c] [1]=10 mm. [d] [1]=
10 mm, [Et3N]=100 mm. [e] [1]=10 mm, [TBAF]=30 mm. [f] The dis-
sociation was very rapid to give a spectrum at the intermediate state.
[g] No dissociation was observed after 24 h.
Scheme 2. Preparation of PRNs 6.
Angew. Chem. Int. Ed. 2011, 50, 4872 –4875
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4873

Communications
Table 2: Thermal properties of PRNs 6.
[a]
[b]
[c]
PRN
End group
Td5 [8C]
Tg,DSC [8C]
Tg,DMA [8C]
[d]
3
–
Cy
279
202
160
165
11.3
5.0
5.2
–
6a
6b
6c
12.0
14.0
12.0
tBu
4-tBuC6H4
3.8
[a] 5% Weight-loss temperature by thermogravimetric analysis (TGA).
[b] Glass transition temperature measured by differential scanning
calorimetry (DSC). [c] Glass transition temperature by dynamic mechan-
ical analysis (DMA). [d] Not measured.
temperature (Td5) of 6a–c was lower than that of 3. The fact
that DSC profiles did not show any peak up to Td5 suggested
that the rotaxane cross-link in 6 was stable to heating. The
dynamic mechanical analysis (DMA) showed similar glass
transition temperatures (Tg,DMA) to Tg,DSC. Furthermore, the
plateau regions of tensile storage modulus (E’) between 50–
1108C clearly indicated the network structure of 6.[13]
Figure 2. The time-dependent weight loss of the insoluble part of 6
under the following conditions: a) DMF, 258C; b) Et3N (100 mm),
*
CH3NO2 , 608C; c) TBAF (30 mm), DMF, 608C. 6a, + 6b, ꢀ 6c.
DW/W0 =(weight of dried gel)/(weight of initial dried gel)ꢀ100%.
The dissociation abilities of the various structures of 6
were examined in a manner similar to those of PRN 1
(Scheme 3). Figure 2 shows the time-dependent weight loss of
6 under various conditions. Cyclohexyl-terminated 6a was
gradually de-cross-linked in DMF at 258C and finally gave a
completely homogeneous solution after 90 min (Figure 2a).
Compound 6a was also partially de-cross-linked to afford a
soluble part in other less polar solvents.[13] On the other hand,
tBu-terminated 6b and 4-tBuC6H4-terminated 6c were insolu-
ble in any solvents but were only swollen, suggesting their
kinetic stability. Treatment of 6a and 6b with Et3N in
CH3NO2 slowly promoted the de-cross-linking reaction, and
they were finally de-cross-linked after 60 min and 360 min,
respectively (Figure 2b). In contrast, 6c maintained the cross-
linked structure under these conditions. Furthermore, 6a was
perfectly dissociated within 20 min by the treatment with
TBAF (Figure 2c). For 6b, it took 75 minutes to achieve
complete de-cross-linking. The drastic change in solubility is
seen in Figure 3.
Although 6c afforded little solubility after treatment for
90 minutes with TBAF, 6c slowly decomposed to give a
completely homogeneous solution after 96 h. These results
were in good agreement with the results of the models using
Scheme 3. De-cross-linking reactions of PRN 6. Reagents and conditions: a) [D6]DMSO, 258C, b) Et3N, CD3NO2, 608C, or c) TBAF·3H2O,
[D6]DMSO, 608C.
4874 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4872 –4875

Keywords: gels · kinetic stabilization · networks · rotaxanes ·
.
supramolecular chemistry
[1] a) F. Sanda, T. Endo, Eco Ind. 2001, 6, 18 – 26; b) F. Sanda, T.
Endo, Polym. Recycl. 1998, 3, 159 – 163; c) T. Endo, F. Sanda,
Angew. Makromol. Chem. 1996, 240, 171 – 180.
[2] For a related review, see: C. J. Kloxin, T. F. Scott, B. J. Adima,
C. N. Bowman, Macromolecules 2010, 43, 2643 – 2653.
[3] For related reviews, see: a) T. Maeda, H. Otsuka, A. Takahara,
Prog. Polym. Sci. 2009, 31, 581 – 604; b) T. Endo, A. Sudo, J.
Polym. Sci. Part A 2009, 47, 4847 – 4858; c) S. D. Bergman, F.
Wudl, J. Mater. Chem. 2008, 18, 41 – 62; d) J.-M. Lehn, Prog.
Polym. Sci. 2005, 30, 814 – 831.
Figure 3. Photographic images of PRNs 6b: a) In the swollen state in
CH3NO2; b) after treatment with TBAF/DMF.
[4] X. X. Chen, M. A. Dam, K. Ono, A. Mal, H. B. Shen, S. R. Nutt,
K. SHeran, F. Wudl, Science 2002, 295, 1698 – 1702.
[5] For related reviews, see: a) W. H. Binder, R. Zirbs in Hydrogen
Bonded Polymers (Ed.: W. Binder), Springer, Berlin, 2007,
pp. 1 – 71; b) G. ten Brinke, J. Ruokolainen, O. Ikkala in
Hydrogen Bonded Polymers (Ed.: W. Binder), Springer,
Berlin, 2007, pp. 113 – 178.
[2]rotaxane 1: The stability of PRNs was controllable as
discussed above. The fact that 6a cannot maintain its network
structure in any solvents indicates that a cyclohexyl group is
too small to kinetically stabilize the rotaxane skeleton, and 6a
is not much more than a conventional supramolecular
network (polypseudorotaxane network). On the other hand,
6b and 6c, stabilized by a bulker end-group than the
cyclohexyl group, showed both good stability and suitable
degradability.
To evaluate recyclability, the recovery ratio of the starting
linear polymer poly(crown ether) 3’ was evaluated. From the
homogeneous solution of 6a obtained by dissolving in DMF,
3’ was recovered as a white powder in 98% yield. Similarly, a
100% yield of 3ꢀ was obtained from mixtures of 6b and
TBAF. The very high recovery ratios of the trunk polymer
strongly indicate the high efficiency of the dissociation of the
present non-covalent rotaxane cross-link. The 1H NMR
spectrum of recovered 3’ showed no signal other than those
assignable to 3.[13] Therefore, the high recovery yield of pure 3’
clearly reveals the occurrence of the selective dissociation of
the rotaxane skeleton without any destruction or damage to
the polymer backbone.
In conclusion, we demonstrated a new design for degrad-
able cross-linked polymers characteristic of rotaxane cross-
links, which have sufficient stability and de-cross-linking
capability depending on the structure of the cross-linker and
the dissociation conditions. The PRN, and especially 6b and
6c, were very similar to the covalent form, thus showing the
high stability as the polymer material, but it is weak relative to
the hydrogen-bond-based supramolecular form, as indicated
by the prompt de-cross-linking by the impediment of hydro-
gen bonds. Furthermore, the high recovery yield of the trunk
polymer shows the potential usefulness of this concept as a
way to recycle cross-linked polymers.
[6] For related reviews, see: a) M.-O. M. Piepenbrock, G. O. Lloyd,
N. Clarke, J. W. Steed, Chem. Rev. 2010, 110, 1960 – 2004; b) M.
Suzuki, K. Hanabusa, Chem. Soc. Rev. 2009, 38, 967 – 975; c) H.
Maeda, Chem. Eur. J. 2008, 14, 11274 – 11282; d) B. M. McKen-
zie, S. J. Rowan in Molecular Recognition and Polymers: Control
of Polymer Structure and Self-Assembly (Eds: V. M. Rotello, S.
Thayuvanavan), Wiley, Hoboken, 2008, pp. 157 – 178.
[7] a) O. Kretschmann, S. W. Choi, M. Miyauchi, I. Tomatsu, A.
Harada, H. Ritter, Angew. Chem. 2006, 118, 4468 – 4472; Angew.
Chem. Int. Ed. 2006, 45, 4361 – 4365; b) A. Hashidzume, T.
Tomaru, A. Harada, Polymer 2006, 47, 6011 – 6027; c) S.
Tamesue, Y. Takashima, H. Yamaguchi, S. Shinkai, A. Harada,
Angew. Chem. 2010, 122, 7623 – 7626; Angew. Chem. Int. Ed.
2010, 49, 7461 – 7464.
[8] For related reviews, see: a) K. Mayumi, K. Ito, Polymer 2010, 51,
959 – 967; b) T. Takata, in Bottom-Up Nanofabrication Supra-
molecules, Self-Assemblies, and Organized Films, Vol. 2 (Eds.: K.
Ariga, H. S. Nalwa), American Scientific Publications, New
York, 2009, pp. 312 – 328; c) G. Wenz, B.-H. Han, A. Mueller,
Chem. Rev. 2006, 106, 782 – 817; d) T. Takata, Polym. J. 2006, 38,
1 – 20; e) A. Harada, J. Polym. Sci. Part A 2006, 44, 5113 – 5119;
f) F. Huang, H. W. Gibson, Prog. Polym. Sci. 2005, 44, 982 – 1018.
[9] a) T. Oku, Y. Furusho, T. Takata, Angew. Chem. 2004, 116, 984 –
987; Angew. Chem. Int. Ed. 2004, 43, 966 – 969; b) T. Bilig, T.
Oku, Y. Furusho, Y. Koyama, S. Asai, T. Takata, Macromolecules
2008, 41, 8496 – 8503.
[10] P. R. Ashton, I. Baxter, M. C. T. Fyfe, F. M. Raymo, N. Spencer,
J. F. Stoddart, A. J. P. White, D. J. Williams, J. Am. Chem. Soc.
1998, 120, 2297 – 2307.
[11] Y. Makita, N. Kihara, T. Takata, J. Org. Chem. 2008, 73, 9245 –
9250.
[12] Y. Tachibana, N. Kihara, Y. Furusho, T. Takata, Org. Lett. 2004,
6, 4507 – 4509.
[13] See the Supporting Information.
[14] Y. Furusho, H. Sasabe, D. Natsui, K. Murakawa, T. Takata, T.
Harada, Bull. Chem. Soc. Jpn. 2004, 77, 179 – 185.
[15] Y. Furusho, T. Oku, T. Hasegawa, A. Tsuboi, N. Kihara, T.
Takata, Chem. Eur. J. 2003, 9, 2895 – 2903.
[16] K. Nakazono, T. Takata, Chem. Eur. J. 2010, 16, 13783 – 13794.
[17] a) T. Takata, Y. Kohsaka, G. Konishi, Chem. Lett. 2007, 36, 292 –
293; b) Y. Kohsaka, G. Konishi, T. Takata, Polym. J. 2007, 39,
861 – 873.
Received: December 20, 2010
Revised: February 25, 2011
Published online: April 27, 2011
Angew. Chem. Int. Ed. 2011, 50, 4872 –4875
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4875