REPORTS E-field ground offsets (5). We obtained direct wN E systematic limits of ≲1 mrad/s for each. We simulated the effects that contribute to fE by ˜ which allowed deliberately correlating Bz with E, us to place a ~10−2 mrad/s limit on their combined effect. Because of our slow molecular beam, relatively small applied E-fields, and small magnetic dipole moment, we do not expect any of these effects to systematically shift wN E above the 10−3 mrad/s level (10, 11). The result of this first-generation ThO measurement, d e ¼ ð−2:1 3:7stat 2:5syst Þ 10−29 e⋅cm ð4Þ NE
comes from d e ¼ −ℏw =E eff using E eff = 84 GV/cm (8, 9) and wN E = (2.6 T 4.8stat T 3.2syst) mrad/s. This sets a 90% confidence limit, jd e j < 8:7 10−29 e⋅cm
ð5Þ
that is smaller than the previous best limit by a factor of 12 (4, 5)—an improvement made possible by the use of the ThO molecule and of a cryogenic source of cold molecules for this purpose. If we were to take into account the roughly estimated 15% uncertainty on the calculated E eff (8) and assume that this represents a 1s Gaussian distribution width, the d e limit stated above would increase by about 5%. Because paramagnetic molecules are sensitive to multiple time reversal (T)– violating effects (24), our measurement should be interpreted as ℏwN E ¼ −d e E eff − W S C S , where C S is a T-violating electron-nucleon coupling and W S is a molecule-specific constant (8, 25). For the d e limit above, we assume C S = 0. Assuming instead that d e = 0 yields C S = (–1.3 T 3.0) × 10−9, corresponding to a 90% confidence limit jC S j < 5.9 × 10−9 that is smaller than the previous limit by a factor of 9 (26). A measurably large EDM requires new mechanisms for T violation, which is equivalent to combined charge-conjugation and parity (CP) violation, given the CPT invariance theorem (2). Nearly every extension to the Standard Model (27, 28) introduces new CP-violating phases fCP . It is difficult to construct mechanisms that systematically suppress fCP , so model builders typically assume sin(fCP ) ~ 1 (29). An EDM arising from new particles at energy L in an n-loop Feynman diagram will have size a n m c2 de eff e ∼k sinðfCP ÞðℏcÞ e 4p L2
ð6Þ
where aeff (about 4/137 for electroweak interactions) encodes the strength with which the electron couples to the new particles, me is the electron mass, and k ~ 0.1 to 1 is a dimensionless prefactor (2, 30, 31). In models where 1- or 2-loop diagrams produce d e , our result typically sets a bound on CP violation at energy scales L ~ 3 TeVor 1 TeV, respectively (27–29, 31). Hence, within the context of many models, our EDM limit constrains
272
CP violation up to energy scales similar to, or higher than, those explored directly at the Large Hadron Collider. References and Notes 1. P. G. H. Sandars, Phys. Lett. 14, 194–196 (1965). 2. I. B. Khriplovich, S. K. Lamoreaux, CP Violation Without Strangeness (Springer, New York, 1997). 3. E. D. Commins, D. DeMille, in Lepton Dipole Moments, B. L. Roberts, W. J. Marciano, Eds. (World Scientific, Singapore, 2010), chap. 14, pp. 519–581. 4. B. Regan, E. Commins, C. Schmidt, D. DeMille, Phys. Rev. Lett. 88, 071805 (2002). 5. J. J. Hudson et al., Nature 473, 493–496 (2011). 6. D. M. Kara et al., New J. Phys. 14, 103051 (2012). 7. M. A. Player, P. G. H. Sandars, J. Phys. B 3, 1620–1635 (1970). 8. L. V. Skripnikov, A. N. Petrov, A. V. Titov, J. Chem. Phys. 139, 221103 (2013). 9. E. R. Meyer, J. L. Bohn, Phys. Rev. A 78, 010502 (2008). 10. A. C. Vutha et al., J. Phys. B 43, 074007 (2010). 11. A. C. Vutha et al., Phys. Rev. A 84, 034502 (2011). 12. S. Bickman, P. Hamilton, Y. Jiang, D. DeMille, Phys. Rev. A 80, 023418 (2009). 13. S. Eckel, P. Hamilton, E. Kirilov, H. W. Smith, D. DeMille, Phys. Rev. A 87, 052130 (2013). 14. W. C. Campbell et al., EPJ Web of Conferences 57, 02004 (2013). 15. E. Kirilov et al., Phys. Rev. A 88, 013844 (2013). 16. S. E. Maxwell et al., Phys. Rev. Lett. 95, 173201 (2005). 17. N. R. Hutzler et al., Phys. Chem. Chem. Phys. 13, 18976 (2011). 18. N. R. Hutzler, H.-I. Lu, J. M. Doyle, Chem. Rev. 112, 4803–4827 (2012). 19. H. R. Gray, R. M. Whitley, C. R. Stroud Jr., Opt. Lett. 3, 218–220 (1978).
20. G. J. Feldman, R. D. Cousins, Phys. Rev. D 57, 3873–3889 (1998). 21. J. H. Curtiss, Ann. Math. Stat. 12, 409–421 (1941). 22. S. Eisenbach, H. Lotem, Proc. SPIE 1972, 19 (1993). 23. H. G. Berry, G. Gabrielse, A. E. Livingston, Appl. Opt. 16, 3200–3205 (1977). 24. M. G. Kozlov, L. N. Labzowsky, J. Phys. At. Mol. Opt. Phys. 28, 1933–1961 (1995). 25. V. A. Dzuba, V. V. Flambaum, C. Harabati, Phys. Rev. A 84, 052108 (2011). 26. W. C. Griffith et al., Phys. Rev. Lett. 102, 101601 (2009). 27. S. Barr, Int. J. Mod. Phys. A 08, 209–236 (1993). 28. M. Pospelov, A. Ritz, Ann. Phys. 318, 119–169 (2005). 29. J. Engel, M. J. Ramsey-Musolf, U. van Kolck, Prog. Part. Nucl. Phys. 71, 21–74 (2013). 30. N. Fortson, P. Sandars, S. Barr, Phys. Today 56, 33 (2003). 31. W. Bernreuther, M. Suzuki, Rev. Mod. Phys. 63, 313–340 (1991). Acknowledgments: Supported by NSF and the Precision Measurement Grants Program of the National Institute of Standards and Technology. We thank M. Reece and M. Schwartz for discussions and S. Cotreau, J. MacArthur, and S. Sansone for technical support. P.W.H. was supported in part by the Office of Science Graduate Fellowship Program, U.S. Department of Energy. The authors declare no competing financial interests.
Supplementary Materials www.sciencemag.org/content/343/6168/269/suppl/DC1 Materials and Methods Fig. S1 Table S1 References (32–36) 7 November 2013; accepted 9 December 2013 Published online 19 December 2013; 10.1126/science.1248213
Single-Crystal Linear Polymers Through Visible Light–Triggered Topochemical Quantitative Polymerization Letian Dou,1,2,3 Yonghao Zheng,1,4 Xiaoqin Shen,1 Guang Wu,5 Kirk Fields,6 Wan-Ching Hsu,2,3 Huanping Zhou,2,3 Yang Yang,2,3† Fred Wudl1,4,5*† One of the challenges in polymer science has been to prepare large-polymer single crystals. We demonstrate a visible light–triggered quantitative topochemical polymerization reaction based on a conjugated dye molecule. Macroscopic-size, high-quality polymer single crystals are obtained. Polymerization is not limited to single crystals, but can also be achieved in highly concentrated solution or semicrystalline thin films. In addition, we show that the polymer decomposes to monomer upon thermolysis, which indicates that the polymerization-depolymerization process is reversible. The physical properties of the polymer crystals enable us to isolate single-polymer strands via mechanical exfoliation, which makes it possible to study individual, long polymer chains. btaining single-crystalline materials is of importance in chemistry, physics, and materials science because it enables not only a fundamental understanding of the nature of the materials through structure-function correlations but also provides a wide range of advanced applications (1–3). Different from inorganic compounds or organic small molecules, polymers tend to form amorphous or semicrystalline phases because of entanglements of the long and flexible backbone (4, 5). Preparing large-size polymer single crystals remains a challenge in polymer
O
17 JANUARY 2014
VOL 343
SCIENCE
science (6–8). Topochemical polymerization, a process whereby the confinement and preorganization of the solid state forces a chemical reaction to proceed with a minimum amount of atomic and molecular movement, has provided a promising solution (9, 10). Hasegawa et al. reported topochemical polymerization reactions of diolefin-related compounds (11, 12) and Wegner discovered the polymerization of the 1,4disubstituted-1,3-diacetylene single crystals by heating or high-energy photon irradiation (13). It was found that, if the reactive monomers are
www.sciencemag.org
REPORTS preorganized at a distance commensurate with the repeat distance in the final polymer, polymerization ensued (13, 14). Numerous studies have been carried out to understand the mechanism and explore applications (15–19). The metallic and superconducting polymer (SN)x was prepared through the thermal polymerization of S2N2 crystals. (20). More recently, the topochemical polymerization of diene (21–23) and quinodimethane (24, 25) compounds were 1
California NanoSystems Institute, University of California, Santa Barbara, CA 93106, USA. 2Department of Materials Science and Engineering, University of California, Los Angeles, CA 90095, USA. 3California NanoSystems Institute, University of California, Los Angeles, CA 90095, USA. 4Advanced Institute for Materials Research at Tohoku University, a World Premier International Research Centers (WPI) Initiative Program (WPI-AIMR) and joint center with the California NanoSystems Institute, University of California, Santa Barbara, CA 93106, USA. 5Department of Materials, University of California, Santa Barbara, CA 93106, USA. 6Department of Mechanical Engineering, University of California, Santa Barbara, CA 93106, USA. *Dedicated to M. Prato on the occasion of his 60th birthday. †Corresponding author. E-mail:
[email protected] (F.W.);
[email protected] (Y.Y.)
A
studied by Matsumoto et al. and Itoh et al., respectively. Although ultraviolet (UV) light–mediated reversibility of polymerization has been reported (26), in most cases, a reversible polymerization– thermal depolymerization process cannot be directly observed, and more important, polymerization is not quantitative. It is not possible to form large single crystals under irradiation because the justformed polymer at the crystal surface layer prevents penetration of light through the bulk. In addition, most of the reaction can happen only in crystals where the monomers are aligned perfectly (9–26). We demonstrate a topochemical polymerization reaction of two conjugated dye molecules based on [2,2′-bi-1H-indene]-1,1′-dione-3,3′-diyl dialkylcarboxylate. The alkyl side chains are found to play an important role in the molecular packing and, hence, topochemical reactivity. Because of the relatively small optical bandgap of the precursor solids (~2.2 eV), visible light can be used to induce the polymerization (27–29). The polymerization occurs, not only in the single
O
R O
O O
(i), (ii)
O
(iii) O
OH O OH
O
O
O
∗
G
F
O
R
∗
O
BIT-OH2
R=n-C 6H 13, BIT-Hep2 R=n-C 8H 17, BIT-Non 2 R=2-ethylpentyl, BIT-EH2
n
R
R=CH 3, BIT-Ac 2
R=n-C6H13 , PBIT-Hep2 R=n-C 8H17, PBIT-Non2
D
B
O
(iv) O
O O
R
O
O HO
+ HO
O
crystalline state but also in highly concentrated solution or semicrystalline thin films. Moreover, the topochemical polymerization-decomposition process is reversible under certain conditions. The dye molecule [2,2′-bi-1H-indene]-3,3′dihydroxy-1,1′-dione (BIT-OH2) was synthesized and characterized in 1898 by Gabriel and Leupold (30), using inexpensive starting materials via a simple one-step reaction. In more recent times, this dye has not been widely investigated, and only few derivatives that showed liquid crystallinity and photochromism have been reported (31–33). In this work, we functionalize the BITOH2 with a series of alkylcarboxylates on the hydroxyl groups in the 3 and 3′ positions and demonstrate that topochemical polymerization occurs under exposure to visible light. The synthetic routes for four derivatives—namely, BITAc2, BIT-Hep2, BIT-Non2, and BIT-EH2 (acetate, heptanoate, nonanoate, and 2-ethylhexanoate side chains, respectively)—and their corresponding polymers (PBIT-Hep2 and PBIT-Non2) are shown in Fig. 1A. BIT-OH2 was synthesized according to Gabriel and Leupold (30). All the side
H
I
dC-C
dC-C
J
C
E
dC-C
K
Fig. 1. Synthesis and crystal structure characterization of single crystalline monomers and polymers. (A) Synthetic route to monomers and polymers. Reaction conditions: (i) melting reaction, potassium acetate, 220°C, 2 hours; (ii) methanol, sodium, 60°C, 30 min; (iii) acyl chloride, triethyl amine, dichloromethane, 0°C to room temperature, 3 hours; and (iv) where ħ is Planck’s constant, n is the frequency, and l is the wavelength of the photon’s radiation, ħn, l ~ 500 nm. (B) Crystal structure of BIT-Hep2. (C) Crystal structure of PBIT-Hep2. (D) Crystal structure of BIT-Ac2. (E) Crystal structure of BIT-EH2. Long side chains and hydrogen atoms are removed for clarity. The oxygen atoms at 3 and 3′ position are labeled using purple color to distinguish them from the one at 1 and 1′ positions. (F) Image of BIT-Hep2 (scale bar, 1.0 cm). (G) Image of PBIT-Hep2 (scale bar, 1.0 cm). (H) Optical microscope image of BIT-Hep2 (scale bar, 0.5 mm). (I) Optical microscope image of PBIT-Hep2 (scale bar, 0.5 mm). (J) SEM image of BIT-Hep2 (scale bar, 20 mm). (K) SEM image of PBITHep2 (scale bar, 20 mm). www.sciencemag.org
SCIENCE
VOL 343
17 JANUARY 2014
273
REPORTS Table 1. Single-crystal x-ray diffraction data of monomers and polymers (see table S2). Values in parentheses for cell lengths and cell angles are estimated standard deviations. Monomers and polymers
Space group
BIT-Ac2
C 2/c
BIT-Hep2
P 21/c
PBIT-Hep2
P 21/c
BIT-Non2
P 21/c
PBIT-Non2
P 21/c
BIT-EH2
Pnna
BIT-OH2
P2(1)2(1)2(1)
Cell lengths (Å) a 17.940 (7) b 4.8376 (17) c 19.458 (7) a 14.014 (2) b 4.8850 (7) c 19.589 (3) a 14.177 (5) b 4.8325 (17) c 19.159 (7) a 15.949 (3) b 4.9529 (8) c 19.403 (3) a 16.259 (2) b 4.8471 (6) c 19.078 (3) a 7.0782 (6) b 28.764 (2) c 14.618 (11) a 3.6817 (8) b 12.291 (3) c 27.527 (6)
chains were added by using the same conditions, and all the monomers (BIT-Ac2, BIT-Hep2, BIT-Non2, and BIT-EH2) were obtained in ~60% yield. The detailed synthesis and molecular structure characterization can be found in figs. S1 and S2 (34). All the monomers are orange in solution, as well as in the solid state, and exhibit similar absorption onset at ~550 nm [ultravioletvisible (UV-vis) absorption spectra of the monomers are shown in fig. S3]. The absorption band from 350 to 550 nm is due to the p-electron delocalization and intramolecular donor-acceptor interactions (the alkylcarboxylate groups at 3,3′ positions are weak donors, and carbonyl groups at 1,1′ positions are strong acceptors). Single crystals of monomers were obtained from dichloromethane (DCM)–ethanol solutions by slow evaporation. The orange crystals of BIT-Hep2 and BIT-Non2 (with long linear side chains) turned to light yellow crystals when exposed to sun light (in ~1 hour) or irradiated under a highpressure sodium lamp (in ~10 min) and became insoluble in all common organic solvents. X-ray diffraction analysis of the light yellow crystals revealed the solids to be polymer single crystals. The crystal structure characterization of monomers and polymers is shown in Fig. 1, B to K, and Table 1. Because BIT-Hep2 and BITNon2 have the same space group, have similar unit cells, and have similar reactivity toward polymerization, only BIT-Hep2 is shown. The packing of BIT-Hep2 and PBIT-Hep2 in the crystals is shown in Fig. 1, B and C, respectively (side chains are removed for clarity). It can be seen that, after polymerization, the carbon atoms in the 3,3′ positions changed from sp2 to sp3 hybridization, and there is a single bond formed
274
Cell angles (°) a 90.00 b 98.650 (10) g 90.00 a 90.00 b 100.086 (11) g 90.00 a 90.00 b 104.422 (9) g 90.00 a 90.00 b 93.527 (5) g 90.00 a 90.00 b 98.974 (6) g 90.00 a 90.00 b 90.00 g 90.00 a 90.00 b 90.00 g 90.00
Cell volume (Å)
dC-C (Å)
1669.5 (11) (Z = 4)
6.612
4.498
4.1
1320.3 (3) (Z = 2)
3.228
3.400
6.0
1271.2 (8) (Z = 1)
1.590
3.229
5.2
1529.9 (5) (Z = 2)
3.258
3.387
4.0
1485.1 (3) (Z = 1)
1.599
3.200
4.8
2976.2 (4) (Z = 4)
5.561
3.306
16.7
1245.6 (5) (Z = 4)
5.201
3.368
4.1
between the 3 and 3′ positions of adjacent monomers. The length of the bond is 1.590 Å, slightly larger than that of a conventional C – C single bond (~1.54 Å) (35). The polymer crystal shares the same space group with the monomer, slightly larger cell angle b (104.422 versus 100.086 degrees), slightly smaller cell volume and slightly smaller p – p stacking distance (dp-p). The perfect alignment of the monomers in the crystals and the short distance of the two active carbon atoms (dC-C, ~3.2 to ~3.3 Å) are necessary to ensure that topochemical reaction will occur with concomitant high-quality polymer crystal formation (R factor of ~5%) for both BIT-Hep2 and BITNon2. As shown in Fig. 1, D and E, the crystal packing and space group of BIT-Ac2, with very short side chains, and BIT-EH2 with long and branched side chains, are very different from BITHep2 or BIT-Non2. For BIT-Ac2, the shortest side chains lead to the largest dC-C of 6.6 Å and dp-p of 4.5 Å. For BIT-EH2, the 2-ethylhexyl groups lead to significant steric hindrance to close packing, which results in a ~30° twisting of the normally coplanar pentagon planes. After many trials, only low-quality crystals (R factor of 16.8%) were obtained for BIT-EH2. Because BIT-Ac2 and BIT-EH2 show larger dC-C than the typical value for topochemical polymerization (~4 to 5 Å) (9, 10), no polymerization reaction is observed. BIT-OH2 with two hydroxyl groups is also inactive, probably because of the unfavorable packing (see Table 1 and supplementary materials for the crystal structure data) (34). The self-assembly effect of nonaromatic alkyl side chains plays a critical role in determining the molecular packing and reactivity. Oak Ridge Thermal Ellipsoid Plots (ORTEP) of BIT-Hep2, BIT-Non2, PBIT-Hep2, and PBIT-Non2
17 JANUARY 2014
VOL 343
SCIENCE
d
p-p
(Å)
R factor (%)
are shown in fig. S4. The carbon atoms at 3,3′ positions in polymers have smaller thermal displacement, probably because they are bonded to four atoms in the polymers, whereas they bond only three atoms in the monomers (the polymer increases the rigidity of the atoms in question) (34). In a study of the wavelength dependence of the reaction, the monomer crystals (BIT-Hep2 and BIT-Non2) were irradiated with 254-, 365-, and 500-nm light for 1 hour. The results showed that 356- and 500-nm light effected polymerization. It was also found that heating the orange crystals in the dark (from room temperature to 200°C) did not initiate polymerization. Therefore, the absorption band from 350 to 550 nm is responsible for the topochemical reactivity. The polymerization mechanism is believed to be similar to that reported for the diene and the quinodimethane compounds (21–25), except that the relatively long-lived excited state is created by visible and not UV light. At the same time, attempted polymerization by a free radical initiator, azobisisobutyronitrile, in a concentrated toluene solution in the dark, afforded only ~5% yield. This result, coupled with the result of photolysis of concentrated solution demands that the polymerization proceeds through a diradical and is not a simple free radical polymerization. This situation would be analogous to diacetylene polymerization where the propagating species is a carbene (14, 16) and not a free radical. We examined the yield of the reaction by extraction of the polymer crystals with DCM, as well as chloroform, for several hours. Evaporation of the solvent afforded no residue. This is in contrast to most previous cases (11–26, 36). The polymerization process described here is quantitative (>99% yield) for
www.sciencemag.org
REPORTS A
B 0.25
Experimetal Simulated
Absorbance (AU)
Intensity (a. u.)
6000
C
4000
2000
(1,0,0) (2,0,0) (2,1,-2)
0 20
30
2 theta (degree)
40
Fig. 2. Characterization of PBIT-Hep2 synthesized from concentrated solution and thin film. (A) Powder x-ray diffraction data from experiments and simulation. a.u., arbitrary units. (B) SEM image (scale bar, 100 mm)
both BIT-Hep2 and BIT-Non2. The excellent conversion yield is attributed to the high reactivity of the monomer crystals and the fact that the polymers do not absorb the wavelength (450 to 550 nm) that is required for the reaction. Images of the BIT-Hep2 crystals (orange) and PBIT-Hep2 crystals (light yellow) are shown in Fig. 1, F and G, respectively. The average length is about 1.5 cm (maximum ~ 1.9 cm), and the average width is about 0.6 mm (maximum ~ 1.1 mm). Optical microscope images of the same crystals before and after polymerization are shown in Fig. 1, H and I. The gross morphology of the polymers is the same as the monomers. The scanning electron microscope (SEM) images of a BIT-Hep2 and PBIT-Hep2 crystal are shown in Fig. 1, J and K, respectively. The monomer and polymer show long and narrow straight domains along the crystal but no obvious “grain boundary” in the long-axis direction. The surface morphology of the polymer crystal is smoother than that of the monomer crystal, probably because of the slight changes of molecular orientation in the crystals. SEM images of the same crystal before and after polymerization can be found in fig. S5. The orientation of the polymer backbone is determined to be along the long axis of the crystal (see fig. S6 for more details), and the molecular mass is estimated to be extremely large. For example, an ideal 5-mm-long polymer in the crystal contains ~107 monomers (the distance between two monomers is ~5 Å) and the molecular mass is ~6 × 109 daltons (37). A closer examination of Fig. 1C and fig. S6 reveals that, in the polymerization process, two asymmetric centers per monomer unit are created. Further, the two asymmetric centers are of opposite configuration. As a result, each monomer unit is of meso configuration and the polymer is not expected to be optically active (38). Because of the high reactivity of BIT-Hep2, other polymerization conditions, including solution and thin film, were examined with the aim of synthesizing PBIT-Hep2 with different dimensions
BIT-Hep2 film Intermediate state PBIT-Hep2 film
0.15 0.10 0.05 0.00 300
50
Fig. 3. Representative tensile strain-stress curves of polymer single crystals. The initial length of the tested crystals (between two gauge length markers) is ~3.0 mm. Average yield strength = 102 MPa. Average tensile strength = 118 MPa. Estimated elastic strain = 0.8%. Estimated average Young’s modulus = 15 GPa.
400
500
Wavelength (nm)
600
and optical microscope image (inset). (C) UV-vis absorption spectra of BITHep2 and PBIT-Hep2 in thin films and corresponding real images (inset). AU, absorbance units.
140 120 100
Stress (MPa)
10
0.20
80 60 40 20 0 0.00
0.01
0.02
0.03
0.04
0.05
Normalized displacement (mm/mm)
and morphology. For the reaction in solution, a concentrated toluene solution (~ 350 mg/ml) of BIT-Hep2 was heated to 50°C (to keep it fully dissolved) under a high-pressure sodium lamp and stirred for 6 hours. Powderlike fine crystals of PBIT-Hep2 were obtained with an isolated yield of ~60%. The experimental and simulated x-ray powder diffraction profiles of the solution-generated polymers are shown in Fig. 2A (simulation is based on single-crystal x-ray diffraction data). Several peaks such as (1,0,0), (2,0,0), and (2,1,–2) in the experimental data match well with the simulated profile. The flat background in the experimental profile indicates the crystallinity is very high for the solutiongenerated polymers. The SEM image and optical microscope image (inset) of the powderlike polymers are collected in Fig. 2B. Different from highly ordered single crystals (Fig. 1J), disordered tiny fibers with diameters of ~1 to ~10 mm can be seen. The highly crystalline polymer nanocrystal fibers obtained here represent an interesting morphological feature.
www.sciencemag.org
SCIENCE
VOL 343
The topochemical polymerization of BITHep2 in thin film was examined to demonstrate the solution processability that is needed for use in flexible plastic electronics. A thin layer of BIT-Hep2 (~60 nm) was spin-cast on a glass substrate from chloroform solution. The orange-colored thin film was irradiated with a high-pressure sodium lamp, which resulted in a colorless film in 10 min. The UV-vis absorption spectra of the thin films before and after polymerization, as well as the images of the orange-colored monomer film and colorless polymer film, are shown in Fig. 2C. After polymerization, the absorption band in the visible range disappeared, and the absorption below 350 nm became stronger. As shown in fig. S7, the x-ray diffraction profile of the polymer thin film consists of broader and weaker peaks than the signal of the polymer powders generated from solution. The average size of the crystallites in the thin film was calculated to be around 40 nm (using the Scherrer equation), much smaller than the polymer powders shown in Fig. 2B. These results indicate that it is possible to coat high-
17 JANUARY 2014
275
REPORTS A
B
C
D h = 1.4 nm
E Fig. 4. Characterization of single and multiple polymer chains obtained by mechanical exfoliation. (A to C) SEM images of single and multiple polymer chains on Si/SiO2 (scale bars, 50 nm). A thin layer of gold (4 nm) was deposited on the sample for better conductivity. (D) 3D AFM image of a few polymer chains. The height of the polymer chain is 1.4 nm, corresponding to a single-layer molecule. (E and F) TEM images of single and multiple polymer chains (scale bars, 100 nm).
quality monomer thin films by solution processing and then convert them to nanocrystalline polymer thin films with visible light. The reaction is thermally reversible. The optical microscope and polarized optical microscope (POM) images of PBIT-Hep2 crystals on a hot stage are shown in fig. S8 (34). When the polymer was heated to 195°C, it turned back to orange, and the orange crystal melted at 205°C (fig. S8, A, B, and C). The highly orientated crystals show bright colors in the POM images (fig. S8, D and E), but the isotropic liquid is dark (fig. S8F). The two processes at 195°C and 205°C are endothermic, as determined by differential scanning calorimetry (see fig. S9). When the orange crystals obtained at 195°C were exposed to visible light, however, they changed back to light yellow, which meant that the reaction is multiply reversible (see fig. S10 for the optical microscope images of the reversibility). The molecular structure of the monomer recovered by decomposition of the polymer was confirmed by nuclear magnetic resonance, and the result is shown in fig. S11. The decomposition process of a polymer crystal to monomer crystal is shown in fig. S8, G to J. It was found that the decomposition happens along the crystal length, which further confirms that the polymer backbone is along the crystal long-axis direction. The thermal decomposition of the polymers is probably due to the relatively longer carbon-carbon single-bond length (~1.59 Å) between two adjacent monomers, compared with the conventional bond length (the polymer chains are under slight tensile stress). Therefore, the bond should be easier to break at elevated temperature or under high-energy photon irradiation. Thermogravimetric analysis of the monomer BIT-Hep2 shows that it decomposes
276
F
d = ~ 3 nm
at, or above, 200°C (see fig. S12). Therefore, fig. S8C shows the decomposed monomer instead of the pure melt. A summary of the polymerization and decomposition conditions can be found in table S1. Preliminary investigations show that the mechanical properties are dominated by very weak interchain van der Waals attraction. During the tensile stress-strain measurements (see fig. S13 for experimental set-up) (34), it was found that the polymer crystals behave as a unidirectional, fiber-reinforced composite, with extreme strength in the fiber direction and basically none in the transverse direction. This behavior resulted in the specimens splitting into individual strands in the grips, sliding with respect to each other. At maximum stress, no final “fracture” occurred. After the point of maximum stress, individual strands failed similar to the way rope fails, such that the stress decreased as the number of failed strands climbed (see fig. S14 and movie S1 for details). This behavior yields a stress-strain curve that exhibits a “toughening” of the material. Despite many difficulties during the measurements (see supplementary materials), the average Young’s modulus could be estimated to be around 15 GPa for several polymer crystal samples. The stressstrain curves of three representative samples are shown in Fig. 3. The result is lower than a typical polydiacetylene polymer (39). One possible reason is, again, the larger carbon-carbon singlebond length (~1.59 Å) between bonded monomers. Another possibility is that the value is underestimated because of the sliding effect (along the crystal long axis) between the polymer chains because the intermolecular interactions are very weak (which leads to overestimated strain). Based on these properties, we are able to isolate a single and/or multi-
17 JANUARY 2014
VOL 343
SCIENCE
d = ~ 6 nm
ple unentangled rigid polymer chain by simple mechanical exfoliation (similar to graphene). To characterize the mechanically exfoliated polymer chains, several microscopy technologies, such as SEM, atomic force microscopy (AFM), and transmission electron microscopy (TEM) were used, and the results are shown in Fig. 4, A to C; D; and E and F; respectively). Some very tiny polymer fibers with diameters around 10 to 15 nm are shown in Fig. 4, A to C. The thinnest one that we can find is 10.7 nm. Considering that ~4-nm-thick gold was deposited on the sample (to ensure good conductivity for SEM measurements), the 10-nm one should contain only one, at most two, polymer chains. The threedimensional (3D) AFM image of few polymer chains is shown in Fig. 4D. The height of the observed chain is measured to be 1.4 nm, which is in accordance with the thickness of one polymer layer. TEM images of several polymer chains are shown in Fig. 4, E and F. The diameters of the chains labeled are around 3 and 6 nm, which indicates single and double polymer chains, respectively. The mechanical exfoliation relies on the highly ordered alignment of the polymer chains in the single crystal and the weak interchain interactions. We demonstrated selective visible light– triggered topochemical polymerization in single crystals, concentrated solutions, and thin films, as well as the reverse process—topochemical decomposition from polymers to monomers, of two bi-indene-dione derivatives. The simple and cost-effective chemistry presented here suggests that dihydroxy-bi-indene-dione is a versatile platform for further study. The properties of a polymeric solid consisting entirely of nonentangled, straight, infinitely long chains are influenced by
www.sciencemag.org
REPORTS the absence of entanglements and relatively weak van der Waals attractive interactions. Just as graphite is made up of two-dimensional graphene sheets held together by weak van der Waals forces, PBIT-Hep2 and PBIT-Non2 are made up of onedimensional chains held together by the same weak forces, mechanically, but not electronically, akin to “one-dimensional graphite.” References and Notes 1. E. M. Landau, M. Levanon, L. Leiserowitz, M. Lahav, J. Sagiv, Nature 318, 353–356 (1985). 2. Y. Yin et al., Science 304, 711–714 (2004). 3. S. Mann, G. A. Ozin, Nature 382, 313–318 (1996). 4. P. J. Flory, J. Chem. Phys. 10, 51–61 (1942). 5. M. L. Huggins, J. Chem. Phys. 9, 440 (1941). 6. E. W. Fischer, G. F. Schmidt, Angew. Chem. Int. Ed. Engl. 1, 488–499 (1962). 7. X. Liu et al., Science 307, 1763–1766 (2005). 8. J. Xu, Y. Ma, W. Hu, M. Rehahn, G. Reiter, Nat. Mater. 8, 348–353 (2009). 9. M. Hasegawa, Chem. Rev. 83, 507–518 (1983). 10. K. Biradha, R. Santra, Chem. Soc. Rev. 42, 950–967 (2013). 11. M. Hasegawa, Y. Suzuki, J. Polym. Sci. Part B 5, 813–815 (1967). 12. M. Hasegawa, Adv. Phys. Org. Chem. 30, 117–171 (1995). 13. G. Wegner, Z. Naturforsch. B 24, 824–832 (1969). 14. G. Wegner, Pure Appl. Chem. 49, 443–454 (1977). 15. D. Bloor, L. Koski, G. C. Stevens, J. Mater. Sci. 10, 1689–1696 (1975). 16. H. Eichele, M. Schwoerer, R. Huber, D. Bloor, Chem. Phys. Lett. 42, 342–346 (1976). 17. Y. Lu et al., Nature 410, 913–917 (2001). 18. J. W. Lauher, F. W. Fowler, N. S. Goroff, Acc. Chem. Res. 41, 1215–1229 (2008).
19. A. Sun, J. W. Lauher, N. S. Goroff, Science 312, 1030–1034 (2006). 20. A. G. MacDiarmid et al., J. Chem. Soc. Chem. Commun. 1975, 476–477 (1975). 21. A. Matsumoto, T. Matsumura, S. Aoki, Macromolecules 29, 423–432 (1996). 22. A. Matsumoto, T. Odani, K. Sada, M. Miyata, K. Tashiro, Nature 405, 328–330 (2000). 23. A. Matsumoto et al., J. Am. Chem. Soc. 124, 8891–8902 (2002). 24. T. Itoh et al., Angew. Chem. Int. Ed. 41, 4306–4309 (2002). 25. S. Nomura et al., J. Am. Chem. Soc. 126, 2035–2041 (2004). 26. P. Johnston, C. Braybrook, K. Saito, Chem. Sci. 3, 2301–2306 (2012). 27. So far, only a few topochemical polymerization reactions can be induced by visible light. For example, the distyryl pyrazine derivatives are sensitive to purple light. See references (28) and (29) for details. 28. M. Hasegawa, Y. Suzuki, F. Suzuki, H. Nakanishi, J. Polym. Sci. A1 7, 743–752 (1969). 29. J. Swiatkiewicz, P. N. Prasad, J. Polym. Sci., Polym. Phys. Ed. 22, 1417–1429 (1984). 30. S. Gabriel, E. Leupold, Ber. Dtsch. Chem. Ges. 31, 1159–1174 (1898). 31. V. Khodorkovsky, A. Ellern, O. Neilands, Tetrahedron Lett. 35, 2955–2958 (1994). 32. K. Tanaka, F. Toda, J. Chem. Soc., Perkin Trans. 1 1, 873–874 (2000). 33. J. Han et al., New J. Chem. 31, 543–548 (2007). 34. Materials and methods are available as supplementary materials on Science Online. 35. R. T. Morrison, R. N. Boyd, R. K. Boyd, Organic Chemistry (Benjamin Cummings, Redwood City, CA, ed. 6, 1992). 36. Although sometimes it is claimed that polymerizations of diacetylenes go to completion, the resulting solids are always extracted after polymerization, and substantial amounts of monomer and other low-molecular-weight material are isolated from the extracts.
Nonenzymatic Sugar Production from Biomass Using Biomass-Derived g-Valerolactone Jeremy S. Luterbacher, Jacqueline M. Rand, David Martin Alonso, Jeehoon Han, J. Tyler Youngquist, Christos T. Maravelias, Brian F. Pfleger, James A. Dumesic* Widespread production of biomass-derived fuels and chemicals will require cost-effective processes for breaking down cellulose and hemicellulose into their constituent sugars. Here, we report laboratory-scale production of soluble carbohydrates from corn stover, hardwood, and softwood at high yields (70 to 90%) in a solvent mixture of biomass-derived g-valerolactone (GVL), water, and dilute acid (0.05 weight percent H2SO4). GVL promotes thermocatalytic saccharification through complete solubilization of the biomass, including the lignin fraction. The carbohydrates can be recovered and concentrated (up to 127 grams per liter) by extraction from GVL into an aqueous phase by addition of NaCl or liquid CO2. This strategy is well suited for catalytic upgrading to furans or fermentative upgrading to ethanol at high titers and near theoretical yield. We estimate through preliminary techno-economic modeling that the overall process could be cost-competitive for ethanol production, with biomass pretreatment followed by enzymatic hydrolysis. iomass is emerging as a possible renewable alternative to petroleum-based resources in light of increasing environmental, economic, and political difficulties associated with fossil fuel extraction and use. Accordingly,
B
Department of Chemical and Biological Engineering, University of Wisconsin–Madison, Madison, WI 53706, USA. *Corresponding author. E-mail:
[email protected]
biomass-derived sugars have been presented as intermediates for the production of renewable fuels (1–3) and chemicals (4–6). However, producing water-soluble carbohydrates from lignocellulosic biomass requires cleaving ether bonds in hemicellulose (primarily xylan) and cellulose (glucan) chains while minimizing further degradation of the resulting C5 and C6 sugars (primarily xylose and glucose) to insoluble degradation
www.sciencemag.org
SCIENCE
VOL 343
37. Because of the insolubility of the polymers, we were not able to measure the molecular masses using conventional methods. Because we did not observe “grain boundries” along the crystals in the SEM images, and for purposes of estimation, we assumed the length of the polymer is equal to the length of the crystal. 38. We thank the editorial staff for bringing up the question of stereochemistry. 39. C. Galiotis, R. J. Young, Polymer 24, 1023–1030 (1983). Acknowledgments: L.D. was supported by NSF (grant no. DMR-1210893) and the Link Foundation Energy Fellowship. Y.Z. and X.S. were supported by the Center for Energy Efficient Materials, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under award number DE-SC0001009 and Department of Energy (grant no. DE-FG02-08ER46535). We thank M. Cornish and W. Cui at University of California, Santa Barbara, for the help with SEM and nuclear magnetic resonance measurements, respectively. We thank Y. Liu, C.-C. Chen, and C.-J. Hsu at University of California, Los Angeles, for the help with AFM, absorption, and thermogravimetric analysis measurements, respectively. Metrical parameters for the structures of BIT-Ac2, BIT-EH2, BIT-Hep2, BIT-Non2, BIT-OH2, PBIT-Hep2, and PBIT-Non2 are available free of charge from the Cambridge Crystallographic Data Centre under reference numbers CCDC-976926, 976927, 976928, 976929, 976930, 976931, and 976932, respectively.
Supplementary Materials www.sciencemag.org/content/343/6168/272/suppl/DC1 Materials and Methods Figs. S1 to S14 Tables S1 and S2 References Movie S1 12 September 2013; accepted 16 December 2013 10.1126/science.1245875
products. Unfortunately, in aqueous solutions containing low acid concentrations [
