Reversible Photoinduced Shape Changes of ... - Semantic Scholar
COMMUNICATION
DOI: 10.1002/adma.200602741
Reversible Photoinduced Shape Changes of Crystalline Organic Nanorods** By Rabih O. Al-Kaysi and Christopher J. Bardeen* Mechanical devices that function on sub-micrometer length scales are expected to find applications in fields ranging from medicine to manufacturing. Biology provides many examples of nanoscale machines, for example, the adenosine triphosphate (ATP)-fueled motion of myosin along the actin filament.[1] In such systems, the motion is fueled by random encounters with high-energy chemical species in the surrounding medium and occurs stochastically, because there is no way to externally control diffusive encounters at the molecular level. An alternative is to use external energy sources (e.g., photons) to activate motion at the nanoscale. Photochemically powered mechanical motion is attractive because it does not require the presence of a second chemical species and external control of the motion can be achieved by manipulating the illumination conditions. These advantages have propelled research into photoactivated mechanical motion on the molecular level, and there have been many examples of photoactivated molecular switching, translation, and geometrical transformations.[2–4] In general, these schemes rely on intramolecular events—bond formation, cyclization, cis–trans isomerization—to drive a structural change on the order of 1 Å or less. This molecular-level change can be amplified by coupling multiple molecules together using a liquid-crystal polymeric host. Recent progress in the development of photodeformable polymer fibers and strips has led to micrometer-scale materials that exhibit light-induced shape memory effects,[5,6] reversible bending under anisotropic illumination conditions,[7–13] and photocontrol of the crosslink density.[14] Recently, we demonstrated the use of an intermolecular photochemical reaction to achieve large physical displacements in a different type of structure: organic molecular crystal nanorods.[15] By taking advantage of the well-known anthracene [4+4] photocycloaddition reaction in 9-tertbutyl-anthracene ester (9-TBAE) molecular crystal nanorods, a large (15 %) change in rod length could be generated. In most cases, the photochemical changes drive a reconstructive crystal-to-crystal phase transition that leads to crystal fragmentation and disin-
– [*] Prof. C. J. Bardeen, Dr. R. O. Al-Kaysi Department of Chemistry, University of California Riverside, CA 92521 (USA) E-mail: [email protected] [**] This work was supported by the National Science Foundation, grant number CHE-0517095, and the University of California Energy Institute, and the Central Facility for Advanced Microscopy and Microanalysis (CFAMM) at UCR.
1276
tegration.[16–18] Apparently, the high surface-to-volume ratio of the nanorods provides sufficient strain relief to avoid cracking and fragmentation during this transition. Although the expansion of the 9-TBAE rods was ca. 100 times larger than that observed in other photoisomerizable molecular crystals,[19] it was largely irreversible. If a reversible system could be found, significant amounts of useful work could be generated by repeatedly inducing such mechanical motion. In the following, we demonstrate that molecular crystal nanorods composed of a related molecule, 9-anthracene carboxylic acid (9-AC), can undergo reversible photoinduced cycling between well-defined shapes after spatially localized excitation. The kinetics of recovery, its dependence on illumination conditions, and the long-term stability of the effect over the course of multiple photocycles are all characterized. In our search for anthracene derivatives that undergo reversible photodimerization reactions, we examined 9-AC, which crystallizes in a head-to-head “cis” arrangement[20] rather than the head-to-tail “trans” arrangement common to most 9-substituted anthracenes. The 9-AC monomer bulk crystal structure (Fig. 1a and b) consists of neighboring stacks of molecules that interact via hydrogen bonding between the carboxyl groups. This crystal structure is retained in the nanorods, as can be seen from the transmission electron microscopy (TEM) image of a single 9-AC nanorod (Fig. 1c) and its electron diffraction pattern (Fig. 1c, inset). Although the cis arrangement is often assumed to prevent the [4+4] cycloaddition reaction in the solid state due to “topochemical” factors,[21] solid state NMR measurements[22] showed that 9-AC does in fact undergo the [4+4] cycloaddition reaction characteristic of anthracenes in the solid state (Fig. 1d). This photodimer is unstable at room temperature and spontaneously reverts back to the monomer state within a few minutes. We have confirmed the reversible nature of the solid-state photodimerization in our samples by monitoring the decrease and subsequent recovery of the monomer crystal fluorescence, which peaks at 505 nm in both nanorods and other crystal types. Although solid state NMR data has confirmed the molecular structure of the photodimer,[22] it does not provide information on changes in the crystal structure or unit cell dimensions. Unfortunately, attempts to obtain the crystal structure of the photodimer were unsuccessful. Even though it was possible to stabilize the photodimer using a liquid-nitrogencooled X-ray diffractometer mount, every attempt to irradiate a single 9-AC crystal of sufficient size to obtain high quality diffraction data resulted in cracks and fissures, which made the crystal unsuitable for diffraction.
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2007, 19, 1276–1280
c
COMMUNICATION
their subsequent contraction in the dark. The overall length change is significantly smaller than that of the previously studied 9-TBAE rods, and is estimated to be 1–3 %. These observations confirm that the crystal structure of the photodimer is different enough from that of the monomer that it can induce useful displacements on the nanoscale. Nevertheless, these changes were relatively small, and the reversibility of the effect disappeared after four illumination cycles or less. We hypothesized that we might obtain more dramatic geometry changes if we irradiated only a segment of the rod. In this case, interfacial strain between the dimer and monomer crystal phases could lead to larger shape deformations. To alleviate COOH the reversibility problem seen on bare COOH surfaces exposed to air, we left the rods λ > 300 nm suspended in acidic water. This leads to [4+4] photocycloaddition HOOC lower ambient oxygen concentrations and hopefully greater photostability. Room Temperature Suspending the rods in aqueous solution also provides them with more freedom COOH of movement. 9-AC in the solid state 9-AC head to head metastable photodimer Our second round of experiments thus concentrated on aqueous suspenFigure 1. a,b) Side and top views of the X-ray crystallographic structure of 9-AC, showing the stacksions of 9-AC nanorods that were exing and hydrogen bond interactions. c) Transmission electron microscopy (TEM) image of a 9-AC posed to spatially localized excitation. single nanorod lying diagonally across a perforated carbon TEM grid. Inset: electron diffraction patIn general, isolated rods underwent distern from a 1.5 lm region of a single nanorod, showing that it is crystalline over this distance. d) Photodimerization and dissociation reaction scheme of 9-AC. tinct shape changes (usually bending) after 2–15 s of UV radiation. Shape changes in rods that were aggregated were much more difficult to observe. Using phase contrast miAlthough we were unable to obtain an X-ray structure for croscopy (Fig. 3), we found that a single rod, irradiated in its the photodimer crystal, we could obtain an upper bound for the change in unit cell size by monitoring the expansion of incentral region, instantly bends under the influence of the light dividual rods using atomic force microscopy (AFM), as shown beam. After 2–5 min in the dark, the bent rod returns to its in Figure 2, after irradiation through the glass substrate and original shape. This sequence can then be repeated, the rod bending under illumination and then straightening in the dark. The degree of bending varies between different rods— the ca. 25° bend in Figure 3 is typical. Fluorescence microscoa b c py shows that bending is accompanied by the appearance of a blue region in the center, where the green excimer fluorescence of the monomer crystal has been replaced by the fluorescence of isolated monomers trapped in the dimerized crystal.[23] If a rod is irradiated along its entire length, the entire rod turns blue, and briefly bends or twists but then returns to its original shape without a permanent shape change. This process is also reversible, but the only net change induced by Figure 2. AFM images of twin 200 nm 9-AC nanorods deposited on a silanized glass surface. a) Before the 365 nm light is turned on, the twin the light is the slight length change observed in the AFM imrods have equal length. b) When the nanorods are illuminated with ages. 365 nm light the upper rod expands more than the lower one (slips by). In contrast to the loss of reversibility seen in the rods studc) After the light is turned off, the sample relaxes back to its original ied by AFM on air-exposed surfaces, the rods retained their shape. The AFM images were collected using intermittent tapping mode. bending ability in aqueous solution when 5 s illumination Scale bar = 500 nm.
a
b
d
Adv. Mater. 2007, 19, 1276–1280
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advmat.de
1277
Normalized Signal Intensity
COMMUNICATION
photomechanical response appears to be reasonably robust with respect to photobleaching. The kinetics and the degree of the a recovery of the 9-AC rods after photod b c excitation depends on the illumination conditions. Fluorescence recovery curves (Fig. 5) for three different periods of illumination at 325 nm (10 s, 30 s, and 60 s) show that prolonged exe f g h posure to the UV irradiation leads to slower recovery curves and greater Figure 3. Single 200 nm diameter nanorod of 9-AC (ca. 60 lm long) briefly exposed to 365 nm photobleaching. The 10 s illumination light in a 50 % solution of phosphoric acid in water. The nanorod repeatedly flexes back and forth results in almost full recovery with an (after UV illumination = panels b, d, f, h; after dark period = panels a, c, e, g). The time required to revert back is around 2 min at room temperature. The dotted circle shows the illuminated region exponential rise-time of 2.6 min, while (35 lm in diameter). Scale bar = 20.7 lm. the 60 s illumination only recovers to about 75 % of its initial value with a slower rise-time of 4.3 min. The lack of full recovery with the longer illumination times is probably periods were used, even after six cycles or more. Nevertheless, due to increased photobleaching. The slower recovery time in any photochemical process side-reactions are always a conis harder to understand, but it may be that the photodegracern. In order to measure the amount of photobleaching of dation products interfere with the reactions that reform the the monomer, we detected the loss of monomer crystal fluomonomer crystal. The data in Figure 5 show that illuminarescence after each illumination cycle. After five cycles of 15 s illumination, the amount of monomer emission has decreased by about 20 % (Fig. 4). The aqueous rods show a much great10 sec 30 sec er degree of reversibility than the polycrystalline films studied previously, where the monomer fluorescence only recovered 1.0 60 sec to 30 % of its original value after irradiation.[22] Extrapolation 0.8 of this curve leads to an estimate of the half-life of the rods of approximately 12 cycles for 15 s illumination periods, and 0.6 more than 30 cycles for 5 s illumination periods. It is likely that this could be improved by more rigorous exclusion of O2 0.4 from the sample. Electronically excited polyacenes are known 0.2 to undergo photoperoxidation,[24] which would prevent the excimer fluorescence and lead to the progressive loss of the 0.0 monomer crystal. It is worth noting that even though the fluo0 5 10 15 20 25 30 35 Time / minutes rescence recovery indicates that some of the monomer is lost upon every illumination cycle, this photodegradation did not Figure 5. Recovery curves of the normalized 9-AC nanorod monomer appear to prevent the associated shape changes. Thus, the crystal fluorescence after 325 nm illumination at time = 0 for three differ-
Normalized signal intensity
ent illumination periods: 10 s, 30 s, and 60 s. The initial fluorescence signal has been normalized to 1.0.
1.0 0.8 0.6 0.4 0.2 0.0 1
2 3 4 5 Number of Cycles
6
Figure 4. Cycling of the 9-AC monomer fluorescence after five exposures, using 325 nm laser light with an exposure time of 15 s. The initial fluorescence signal has been normalized to 1.0, and the recovery signal is recorded after 20 min in the dark.
1278
www.advmat.de
tion conditions play an important role in both the rate and the ultimate degree of reversibility in the 9-AC rods. Further work is required to fully characterize both the primary photoreaction and the role of possible side reactions. It is important to note that all three illumination periods gave rise to the same amount of shape change in a particular rod, suggesting that to extend the lifetime of the nanorod mechanical response, the period of UV illumination should be minimized. We have experimentally established that localized UV excitation can repeatedly induce reversible shape changes in crystalline 9-AC nanorods. The question remains as to the mechanism of these shape changes. Carbon nanotube bundles can deform under uniform visible illumination,[25] but this re-
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2007, 19, 1276–1280
Adv. Mater. 2007, 19, 1276–1280
improved understanding of the bending mechanism is clearly necessary if this phenomenon is to be harnessed for technological applications, the generality of the porous Al2O3 template method suggests that crystalline rods of even smaller diameters may be synthesized,[35] leading to the development of truly nanoscopic photoactuators.
COMMUNICATION
sponse has been attributed to electrostatic interactions due the build-up of photogenerated charges, and disappears when the illumination is removed. It has been shown theoretically that simple free expansion can lead to large temporary deformations of elastic filaments, although this does not lead to a permanent change in shape.[26] Most photodeformable polymers undergo bending in response to asymmetric illumination conditions, and thus most models of electrochemical bending assume a concentration gradient of reacted species.[27–29] For photoinduced bending, the exponential attenuation of light in the material leads to a gradient in photogenerated species parallel to the light propagation direction, which in turn leads to elastic deformation.[30] Given a peak value of the molecular absorption coefficient of ca. 104 L M–1 cm[22] and a density of 1.15 g cm–3, we can estimate the maximum absorption in the 200 nm rods to be ca. 1.0, that is, 10 % of the exciting light is transmitted. This value is large enough to permit the creation of a vertical gradient leading to deformation.[30] However, the fact that bending only occurs when a subsection of the rod is illuminated indicates that a gradient perpendicular to the light propagation must also play a role in the bending. While the exact mechanism requires further investigation, the bending may result from asymmetric strain at the interface between regions of dimerized and monomeric 9-AC molecules, possibly enhanced by the strong hydrogen bond interactions between carboxylic acid groups in these regions. This type of mechanism would suggest that the bending in Figure 3 should be localized at the boundaries of the circular illuminated region outlined in Figure 3a. The fact that the bend is not localized at the edges may reflect a gradient of light intensity in the illuminated region due to aperture effects, so the interface between monomer and dimer regions is not as sharp as suggested by the circle. In summary, we have shown how large shape changes in crystalline organic nanorods can be reversibly induced by localized photoexcitation. We have characterized the kinetics of the reversible photodimerization and showed that the shape changes persist over multiple illumination cycles. This work reinforces our earlier conclusions that nanometer-scale molecular crystal structures can alleviate the strain-induced disruption and fragmentation often associated with photochemical changes occurring in larger (micrometer to millimeter-scale) crystals. These crystalline photomechanical systems may have several important advantages relative to polymeric and liquid crystalline actuators. First, the typical organic molecular crystal has a larger elastic modulus (e.g., that of anthracene is ca. 8 GPa)[31,32] than any of the liquid crystalline systems (ca. 1 GPa)[9] examined so far. A higher elastic modulus results in a greater amount of force generated for a given displacement. Second, the 200 nm diameter rods studied in this work are considerably smaller than the micrometer-scale polymer structures typically studied. Although the size dependence of the elastic modulus is only beginning to be investigated, preliminary indications are that nanorods and nanowires are capable of retaining or even surpassing the mechanical properties of the bulk material.[33,34] Although an
Experimental Synthesis of 9-AC Nanorods: 9-anthracene carboxylic acid (9-AC, Aldrich 99 %) was recrystallized two times from spectroscopic grade toluene to yield pure yellow crystalline needles. The 9-AC nanorods were made in a manner similar to that reported previously [36]. 12 mg of 9-AC was dissolved in 0.2 mL of dry spectroscopic grade tetrahydrofuran (THF). The solution was then gently deposited on both sides of an anodized alumina Anodisc (Whatman Anodisc-13, 200 nm pore diameter, with average membrane thickness of 60 lm, and pore density of 109 cm–2) and the excess 9-AC was gently wiped off the surface. The loaded Anodisc was placed on a specially designed Teflon holder that holds the Anodisc from the polymer support ring. The Anodisc/ Teflon holder was placed in a glass jar containing 2 mL of dry tetrahydrofuran (THF), and a small vial containing phosphorous pentoxide to prevent water contamination. The jar was tightly covered and a long needle was inserted through the lid to the bottom of the jar. Another needle was inserted in the cover coupled with a gas bubbler. Argon gas was gently introduced into the jar at a rate of 1 bubble every 8 to 10 s. The slow room-temperature solvent vapor annealing causes the 9-AC in the Anodisc to dissolve in the solvent vapors and slowly crystallize in the 200 nm channels after all the solvent has evaporated, forming crystalline nanorods. After all the solvent had evaporated under the slow stream of Argon gas (ca. 48 h), the annealed Anodisc surface was polished using 1500 grit sandpaper to remove excess solid 9-AC and the Anodisc was dissolved in 50 % aqueous phosphoric acid to form a 9-AC nanorod suspension. The nanorods were kept in acidic solution to suppress the dissolution of the 9-AC in water. Atomic Force Microscopy (AFM) Measurements: A drop of the nanorod suspension in aqueous phosphoric acid was gently swirled on a silanized microscope slide, yielding a well-dispersed layer of 9-AC nanorod clusters. The silanized piece of glass was gently washed with deionized water to remove any acid residue and dried at room temperature. The samples were scanned using intermittent contact (tapping) mode to minimize perturbation of the sample surface. The scanning rate was 2 Hz, for a total scan time of 7 min. The 9-AC nanorods were exposed to 365 nm light to induce the photoreaction and scanned while irradiated in order to sustain the photodimerization and minimize the reverse reaction. Optical Microscopy Measurements: A drop of the nanorod suspension was sandwiched between two microscope cover slips for optical microscopy measurements. The Olympus IX-70 inverted microscope was coupled to both a charge-coupled device (CCD) camera for imaging and a photomultiplier tube for measuring fluorescence intensities. The relative amount of monomer crystal present was monitored by detecting its characteristic excimer fluorescence using a 510 nm longpass filter. Shape changes were monitored using phase contrast illumination with a tungsten lamp to avoid exposing the rods to additional UV radiation. Spatially localized excitation was obtained using either a Hg lamp (365 nm excitation) apertured to a 35 lm spot, or a HeCd laser beam (325 nm excitation) focused to a 15 lm spot. Both types of excitation induced shape changes. The average intensity at the sample plane was similar for both types of illumination, about 1 W cm–2. In all cases, the rods were illuminated on one side, from below.
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Received: November 30, 2006 Revised: February 21, 2007 Published online: April 4, 2007
www.advmat.de
1279
COMMUNICATION
– [1] A. Yildiz, J. N. Forkey, S. A. McKinney, T. Ha, Y. E. Goldman, P. R. Selvin, Science 2003, 300, 2061. [2] B. L. Feringa, R. A. van Delden, N. Koumura, E. M. Geertsema, Chem. Rev. 2000, 100, 1789. [3] K. Kinbara, T. Aida, Chem. Rev. 2005, 105, 1377. [4] V. Balzani, M. Clemente-Leon, A. Credi, B. Ferrer, M. Venturi, A. H. Flood, J. F. Stoddart, Proc. Natl. Acad. Sci. USA 2006, 103, 1178. [5] H. Jiang, S. Kelch, A. Lendlein, Adv. Mater. 2006, 18, 1471. [6] A. Lendlein, H. Jiang, O. Junger, R. Langer, Nature 2005, 434, 879. [7] H. Finkelmann, E. Nishikawa, G. G. Pereira, M. Warner, Phys. Rev. Lett. 2001, 87, 015 501/1. [8] M. Camacho-Lopez, H. Finkelmann, P. Palffy-Muhoray, M. Shelley, Nat. Mater. 2004, 3, 307. [9] K. D. Harris, R. Cuypers, P. Scheibe, C. L. van Oosten, C. W. M. Bastiaansen, J. Lub, D. J. Broer, J. Mater. Chem. 2005, 15, 5043. [10] M.-H. Li, P. Keller, B. Li, X. Wang, M. Brunet, Adv. Mater. 2003, 15, 569. [11] A. Athanassiou, M. Kalyva, K. Lakiotaki, S. Georgiou, C. Fotakis, Adv. Mater. 2005, 17, 988. [12] S. V. Ahir, E. M. Terentjev, Nat. Mater. 2005, 4, 491. [13] S. Bian, D. Robinson, M. G. Kuzyk, J. Opt. Soc. Am. B 2006, 23, 697. [14] T. F. Scott, R. B. Draughon, C. N. Bowman, Adv. Mater. 2006, 18, 2128. [15] R. O. Al-Kaysi, A. M. Muller, C. J. Bardeen, J. Am. Chem. Soc. 2006, 128, 15 938.
[16] G. Kaupp, Angew. Chem. Int. Ed. Engl. 1992, 31, 595. [17] A. E. Keating, M. A. Garcia-Garibay, in Organic and Inorganic Photochemistry, Vol. 2 (Eds: V. Ramamurthy, K. S. Schanze), Marcel Dekker, New York 1998, p. 195. [18] I. Turowska-Tyrk, J. Phys. Org. Chem. 2004, 17, 837. [19] M. Irie, S. Kobatake, M. Horichi, Science 2001, 291, 1769. [20] L. J. Fitzgerald, R. E. Gerkin, Acta. Cryst. C 1997, 53, 71. [21] M. D. Cohen, Angew. Chem. Int. Ed. Engl. 1975, 14, 386. [22] Y. Ito, H. Fujita, J. Org. Chem. 1996, 61, 5677. [23] R. M. MacFarlane, M. R. Philpott, Chem. Phys. Lett. 1976, 41, 33. [24] B. Stevens, S. R. Perez, J. A. Ors, J. Am. Chem. Soc. 1974, 96, 6846. [25] Y. Zhang, S. Iijima, Phys. Rev. Lett. 1999, 82, 3472. [26] A. J. Spakowitz, Z. G. Wang, Phys. Rev. E 2001, 64, 061 802. [27] E. W. H. Jager, E. Smela, O. Inganäs, Science 2000, 290, 1540. [28] G. Alici, B. Mui, C. Cook, Sens. Actuators A 2006, 126, 396. [29] P. Metz, B. Alici, G. M. Spinks, Sens. Actuators A 2006, 130–131, 1. [30] M. Warner, L. Mahadevan, Phys. Rev. Lett. 2004, 92, 134 302/1. [31] A. Bondi, J. Phys. Chem. Solids 1967, 28, 649. [32] R. J. Roberts, R. C. Rowe, P. York, Powder Technol. 1991, 65, 139. [33] E. W. Wong, P. E. Sheehan, C. M. Lieber, Science 1997, 277, 1971. [34] H. Liang, M. Upmanyu, H. Huang, Phys. Rev. B 2005, 71, 241 403/1. [35] C. R. Martin, Science 1994, 266, 1961. [36] R. O. Al-Kaysi, C. J. Bardeen, Chem. Commun. 2006, 1224.
______________________
1280
www.advmat.de
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2007, 19, 1276–1280
DOI: 10.1002/adma.200602741
Reversible Photoinduced Shape Changes of Crystalline Organic Nanorods** By Rabih O. Al-Kaysi and Christopher J. Bardeen* Mechanical devices that function on sub-micrometer length scales are expected to find applications in fields ranging from medicine to manufacturing. Biology provides many examples of nanoscale machines, for example, the adenosine triphosphate (ATP)-fueled motion of myosin along the actin filament.[1] In such systems, the motion is fueled by random encounters with high-energy chemical species in the surrounding medium and occurs stochastically, because there is no way to externally control diffusive encounters at the molecular level. An alternative is to use external energy sources (e.g., photons) to activate motion at the nanoscale. Photochemically powered mechanical motion is attractive because it does not require the presence of a second chemical species and external control of the motion can be achieved by manipulating the illumination conditions. These advantages have propelled research into photoactivated mechanical motion on the molecular level, and there have been many examples of photoactivated molecular switching, translation, and geometrical transformations.[2–4] In general, these schemes rely on intramolecular events—bond formation, cyclization, cis–trans isomerization—to drive a structural change on the order of 1 Å or less. This molecular-level change can be amplified by coupling multiple molecules together using a liquid-crystal polymeric host. Recent progress in the development of photodeformable polymer fibers and strips has led to micrometer-scale materials that exhibit light-induced shape memory effects,[5,6] reversible bending under anisotropic illumination conditions,[7–13] and photocontrol of the crosslink density.[14] Recently, we demonstrated the use of an intermolecular photochemical reaction to achieve large physical displacements in a different type of structure: organic molecular crystal nanorods.[15] By taking advantage of the well-known anthracene [4+4] photocycloaddition reaction in 9-tertbutyl-anthracene ester (9-TBAE) molecular crystal nanorods, a large (15 %) change in rod length could be generated. In most cases, the photochemical changes drive a reconstructive crystal-to-crystal phase transition that leads to crystal fragmentation and disin-
– [*] Prof. C. J. Bardeen, Dr. R. O. Al-Kaysi Department of Chemistry, University of California Riverside, CA 92521 (USA) E-mail: [email protected] [**] This work was supported by the National Science Foundation, grant number CHE-0517095, and the University of California Energy Institute, and the Central Facility for Advanced Microscopy and Microanalysis (CFAMM) at UCR.
1276
tegration.[16–18] Apparently, the high surface-to-volume ratio of the nanorods provides sufficient strain relief to avoid cracking and fragmentation during this transition. Although the expansion of the 9-TBAE rods was ca. 100 times larger than that observed in other photoisomerizable molecular crystals,[19] it was largely irreversible. If a reversible system could be found, significant amounts of useful work could be generated by repeatedly inducing such mechanical motion. In the following, we demonstrate that molecular crystal nanorods composed of a related molecule, 9-anthracene carboxylic acid (9-AC), can undergo reversible photoinduced cycling between well-defined shapes after spatially localized excitation. The kinetics of recovery, its dependence on illumination conditions, and the long-term stability of the effect over the course of multiple photocycles are all characterized. In our search for anthracene derivatives that undergo reversible photodimerization reactions, we examined 9-AC, which crystallizes in a head-to-head “cis” arrangement[20] rather than the head-to-tail “trans” arrangement common to most 9-substituted anthracenes. The 9-AC monomer bulk crystal structure (Fig. 1a and b) consists of neighboring stacks of molecules that interact via hydrogen bonding between the carboxyl groups. This crystal structure is retained in the nanorods, as can be seen from the transmission electron microscopy (TEM) image of a single 9-AC nanorod (Fig. 1c) and its electron diffraction pattern (Fig. 1c, inset). Although the cis arrangement is often assumed to prevent the [4+4] cycloaddition reaction in the solid state due to “topochemical” factors,[21] solid state NMR measurements[22] showed that 9-AC does in fact undergo the [4+4] cycloaddition reaction characteristic of anthracenes in the solid state (Fig. 1d). This photodimer is unstable at room temperature and spontaneously reverts back to the monomer state within a few minutes. We have confirmed the reversible nature of the solid-state photodimerization in our samples by monitoring the decrease and subsequent recovery of the monomer crystal fluorescence, which peaks at 505 nm in both nanorods and other crystal types. Although solid state NMR data has confirmed the molecular structure of the photodimer,[22] it does not provide information on changes in the crystal structure or unit cell dimensions. Unfortunately, attempts to obtain the crystal structure of the photodimer were unsuccessful. Even though it was possible to stabilize the photodimer using a liquid-nitrogencooled X-ray diffractometer mount, every attempt to irradiate a single 9-AC crystal of sufficient size to obtain high quality diffraction data resulted in cracks and fissures, which made the crystal unsuitable for diffraction.
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2007, 19, 1276–1280
c
COMMUNICATION
their subsequent contraction in the dark. The overall length change is significantly smaller than that of the previously studied 9-TBAE rods, and is estimated to be 1–3 %. These observations confirm that the crystal structure of the photodimer is different enough from that of the monomer that it can induce useful displacements on the nanoscale. Nevertheless, these changes were relatively small, and the reversibility of the effect disappeared after four illumination cycles or less. We hypothesized that we might obtain more dramatic geometry changes if we irradiated only a segment of the rod. In this case, interfacial strain between the dimer and monomer crystal phases could lead to larger shape deformations. To alleviate COOH the reversibility problem seen on bare COOH surfaces exposed to air, we left the rods λ > 300 nm suspended in acidic water. This leads to [4+4] photocycloaddition HOOC lower ambient oxygen concentrations and hopefully greater photostability. Room Temperature Suspending the rods in aqueous solution also provides them with more freedom COOH of movement. 9-AC in the solid state 9-AC head to head metastable photodimer Our second round of experiments thus concentrated on aqueous suspenFigure 1. a,b) Side and top views of the X-ray crystallographic structure of 9-AC, showing the stacksions of 9-AC nanorods that were exing and hydrogen bond interactions. c) Transmission electron microscopy (TEM) image of a 9-AC posed to spatially localized excitation. single nanorod lying diagonally across a perforated carbon TEM grid. Inset: electron diffraction patIn general, isolated rods underwent distern from a 1.5 lm region of a single nanorod, showing that it is crystalline over this distance. d) Photodimerization and dissociation reaction scheme of 9-AC. tinct shape changes (usually bending) after 2–15 s of UV radiation. Shape changes in rods that were aggregated were much more difficult to observe. Using phase contrast miAlthough we were unable to obtain an X-ray structure for croscopy (Fig. 3), we found that a single rod, irradiated in its the photodimer crystal, we could obtain an upper bound for the change in unit cell size by monitoring the expansion of incentral region, instantly bends under the influence of the light dividual rods using atomic force microscopy (AFM), as shown beam. After 2–5 min in the dark, the bent rod returns to its in Figure 2, after irradiation through the glass substrate and original shape. This sequence can then be repeated, the rod bending under illumination and then straightening in the dark. The degree of bending varies between different rods— the ca. 25° bend in Figure 3 is typical. Fluorescence microscoa b c py shows that bending is accompanied by the appearance of a blue region in the center, where the green excimer fluorescence of the monomer crystal has been replaced by the fluorescence of isolated monomers trapped in the dimerized crystal.[23] If a rod is irradiated along its entire length, the entire rod turns blue, and briefly bends or twists but then returns to its original shape without a permanent shape change. This process is also reversible, but the only net change induced by Figure 2. AFM images of twin 200 nm 9-AC nanorods deposited on a silanized glass surface. a) Before the 365 nm light is turned on, the twin the light is the slight length change observed in the AFM imrods have equal length. b) When the nanorods are illuminated with ages. 365 nm light the upper rod expands more than the lower one (slips by). In contrast to the loss of reversibility seen in the rods studc) After the light is turned off, the sample relaxes back to its original ied by AFM on air-exposed surfaces, the rods retained their shape. The AFM images were collected using intermittent tapping mode. bending ability in aqueous solution when 5 s illumination Scale bar = 500 nm.
a
b
d
Adv. Mater. 2007, 19, 1276–1280
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advmat.de
1277
Normalized Signal Intensity
COMMUNICATION
photomechanical response appears to be reasonably robust with respect to photobleaching. The kinetics and the degree of the a recovery of the 9-AC rods after photod b c excitation depends on the illumination conditions. Fluorescence recovery curves (Fig. 5) for three different periods of illumination at 325 nm (10 s, 30 s, and 60 s) show that prolonged exe f g h posure to the UV irradiation leads to slower recovery curves and greater Figure 3. Single 200 nm diameter nanorod of 9-AC (ca. 60 lm long) briefly exposed to 365 nm photobleaching. The 10 s illumination light in a 50 % solution of phosphoric acid in water. The nanorod repeatedly flexes back and forth results in almost full recovery with an (after UV illumination = panels b, d, f, h; after dark period = panels a, c, e, g). The time required to revert back is around 2 min at room temperature. The dotted circle shows the illuminated region exponential rise-time of 2.6 min, while (35 lm in diameter). Scale bar = 20.7 lm. the 60 s illumination only recovers to about 75 % of its initial value with a slower rise-time of 4.3 min. The lack of full recovery with the longer illumination times is probably periods were used, even after six cycles or more. Nevertheless, due to increased photobleaching. The slower recovery time in any photochemical process side-reactions are always a conis harder to understand, but it may be that the photodegracern. In order to measure the amount of photobleaching of dation products interfere with the reactions that reform the the monomer, we detected the loss of monomer crystal fluomonomer crystal. The data in Figure 5 show that illuminarescence after each illumination cycle. After five cycles of 15 s illumination, the amount of monomer emission has decreased by about 20 % (Fig. 4). The aqueous rods show a much great10 sec 30 sec er degree of reversibility than the polycrystalline films studied previously, where the monomer fluorescence only recovered 1.0 60 sec to 30 % of its original value after irradiation.[22] Extrapolation 0.8 of this curve leads to an estimate of the half-life of the rods of approximately 12 cycles for 15 s illumination periods, and 0.6 more than 30 cycles for 5 s illumination periods. It is likely that this could be improved by more rigorous exclusion of O2 0.4 from the sample. Electronically excited polyacenes are known 0.2 to undergo photoperoxidation,[24] which would prevent the excimer fluorescence and lead to the progressive loss of the 0.0 monomer crystal. It is worth noting that even though the fluo0 5 10 15 20 25 30 35 Time / minutes rescence recovery indicates that some of the monomer is lost upon every illumination cycle, this photodegradation did not Figure 5. Recovery curves of the normalized 9-AC nanorod monomer appear to prevent the associated shape changes. Thus, the crystal fluorescence after 325 nm illumination at time = 0 for three differ-
Normalized signal intensity
ent illumination periods: 10 s, 30 s, and 60 s. The initial fluorescence signal has been normalized to 1.0.
1.0 0.8 0.6 0.4 0.2 0.0 1
2 3 4 5 Number of Cycles
6
Figure 4. Cycling of the 9-AC monomer fluorescence after five exposures, using 325 nm laser light with an exposure time of 15 s. The initial fluorescence signal has been normalized to 1.0, and the recovery signal is recorded after 20 min in the dark.
1278
www.advmat.de
tion conditions play an important role in both the rate and the ultimate degree of reversibility in the 9-AC rods. Further work is required to fully characterize both the primary photoreaction and the role of possible side reactions. It is important to note that all three illumination periods gave rise to the same amount of shape change in a particular rod, suggesting that to extend the lifetime of the nanorod mechanical response, the period of UV illumination should be minimized. We have experimentally established that localized UV excitation can repeatedly induce reversible shape changes in crystalline 9-AC nanorods. The question remains as to the mechanism of these shape changes. Carbon nanotube bundles can deform under uniform visible illumination,[25] but this re-
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2007, 19, 1276–1280
Adv. Mater. 2007, 19, 1276–1280
improved understanding of the bending mechanism is clearly necessary if this phenomenon is to be harnessed for technological applications, the generality of the porous Al2O3 template method suggests that crystalline rods of even smaller diameters may be synthesized,[35] leading to the development of truly nanoscopic photoactuators.
COMMUNICATION
sponse has been attributed to electrostatic interactions due the build-up of photogenerated charges, and disappears when the illumination is removed. It has been shown theoretically that simple free expansion can lead to large temporary deformations of elastic filaments, although this does not lead to a permanent change in shape.[26] Most photodeformable polymers undergo bending in response to asymmetric illumination conditions, and thus most models of electrochemical bending assume a concentration gradient of reacted species.[27–29] For photoinduced bending, the exponential attenuation of light in the material leads to a gradient in photogenerated species parallel to the light propagation direction, which in turn leads to elastic deformation.[30] Given a peak value of the molecular absorption coefficient of ca. 104 L M–1 cm[22] and a density of 1.15 g cm–3, we can estimate the maximum absorption in the 200 nm rods to be ca. 1.0, that is, 10 % of the exciting light is transmitted. This value is large enough to permit the creation of a vertical gradient leading to deformation.[30] However, the fact that bending only occurs when a subsection of the rod is illuminated indicates that a gradient perpendicular to the light propagation must also play a role in the bending. While the exact mechanism requires further investigation, the bending may result from asymmetric strain at the interface between regions of dimerized and monomeric 9-AC molecules, possibly enhanced by the strong hydrogen bond interactions between carboxylic acid groups in these regions. This type of mechanism would suggest that the bending in Figure 3 should be localized at the boundaries of the circular illuminated region outlined in Figure 3a. The fact that the bend is not localized at the edges may reflect a gradient of light intensity in the illuminated region due to aperture effects, so the interface between monomer and dimer regions is not as sharp as suggested by the circle. In summary, we have shown how large shape changes in crystalline organic nanorods can be reversibly induced by localized photoexcitation. We have characterized the kinetics of the reversible photodimerization and showed that the shape changes persist over multiple illumination cycles. This work reinforces our earlier conclusions that nanometer-scale molecular crystal structures can alleviate the strain-induced disruption and fragmentation often associated with photochemical changes occurring in larger (micrometer to millimeter-scale) crystals. These crystalline photomechanical systems may have several important advantages relative to polymeric and liquid crystalline actuators. First, the typical organic molecular crystal has a larger elastic modulus (e.g., that of anthracene is ca. 8 GPa)[31,32] than any of the liquid crystalline systems (ca. 1 GPa)[9] examined so far. A higher elastic modulus results in a greater amount of force generated for a given displacement. Second, the 200 nm diameter rods studied in this work are considerably smaller than the micrometer-scale polymer structures typically studied. Although the size dependence of the elastic modulus is only beginning to be investigated, preliminary indications are that nanorods and nanowires are capable of retaining or even surpassing the mechanical properties of the bulk material.[33,34] Although an
Experimental Synthesis of 9-AC Nanorods: 9-anthracene carboxylic acid (9-AC, Aldrich 99 %) was recrystallized two times from spectroscopic grade toluene to yield pure yellow crystalline needles. The 9-AC nanorods were made in a manner similar to that reported previously [36]. 12 mg of 9-AC was dissolved in 0.2 mL of dry spectroscopic grade tetrahydrofuran (THF). The solution was then gently deposited on both sides of an anodized alumina Anodisc (Whatman Anodisc-13, 200 nm pore diameter, with average membrane thickness of 60 lm, and pore density of 109 cm–2) and the excess 9-AC was gently wiped off the surface. The loaded Anodisc was placed on a specially designed Teflon holder that holds the Anodisc from the polymer support ring. The Anodisc/ Teflon holder was placed in a glass jar containing 2 mL of dry tetrahydrofuran (THF), and a small vial containing phosphorous pentoxide to prevent water contamination. The jar was tightly covered and a long needle was inserted through the lid to the bottom of the jar. Another needle was inserted in the cover coupled with a gas bubbler. Argon gas was gently introduced into the jar at a rate of 1 bubble every 8 to 10 s. The slow room-temperature solvent vapor annealing causes the 9-AC in the Anodisc to dissolve in the solvent vapors and slowly crystallize in the 200 nm channels after all the solvent has evaporated, forming crystalline nanorods. After all the solvent had evaporated under the slow stream of Argon gas (ca. 48 h), the annealed Anodisc surface was polished using 1500 grit sandpaper to remove excess solid 9-AC and the Anodisc was dissolved in 50 % aqueous phosphoric acid to form a 9-AC nanorod suspension. The nanorods were kept in acidic solution to suppress the dissolution of the 9-AC in water. Atomic Force Microscopy (AFM) Measurements: A drop of the nanorod suspension in aqueous phosphoric acid was gently swirled on a silanized microscope slide, yielding a well-dispersed layer of 9-AC nanorod clusters. The silanized piece of glass was gently washed with deionized water to remove any acid residue and dried at room temperature. The samples were scanned using intermittent contact (tapping) mode to minimize perturbation of the sample surface. The scanning rate was 2 Hz, for a total scan time of 7 min. The 9-AC nanorods were exposed to 365 nm light to induce the photoreaction and scanned while irradiated in order to sustain the photodimerization and minimize the reverse reaction. Optical Microscopy Measurements: A drop of the nanorod suspension was sandwiched between two microscope cover slips for optical microscopy measurements. The Olympus IX-70 inverted microscope was coupled to both a charge-coupled device (CCD) camera for imaging and a photomultiplier tube for measuring fluorescence intensities. The relative amount of monomer crystal present was monitored by detecting its characteristic excimer fluorescence using a 510 nm longpass filter. Shape changes were monitored using phase contrast illumination with a tungsten lamp to avoid exposing the rods to additional UV radiation. Spatially localized excitation was obtained using either a Hg lamp (365 nm excitation) apertured to a 35 lm spot, or a HeCd laser beam (325 nm excitation) focused to a 15 lm spot. Both types of excitation induced shape changes. The average intensity at the sample plane was similar for both types of illumination, about 1 W cm–2. In all cases, the rods were illuminated on one side, from below.
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Received: November 30, 2006 Revised: February 21, 2007 Published online: April 4, 2007
www.advmat.de
1279
COMMUNICATION
– [1] A. Yildiz, J. N. Forkey, S. A. McKinney, T. Ha, Y. E. Goldman, P. R. Selvin, Science 2003, 300, 2061. [2] B. L. Feringa, R. A. van Delden, N. Koumura, E. M. Geertsema, Chem. Rev. 2000, 100, 1789. [3] K. Kinbara, T. Aida, Chem. Rev. 2005, 105, 1377. [4] V. Balzani, M. Clemente-Leon, A. Credi, B. Ferrer, M. Venturi, A. H. Flood, J. F. Stoddart, Proc. Natl. Acad. Sci. USA 2006, 103, 1178. [5] H. Jiang, S. Kelch, A. Lendlein, Adv. Mater. 2006, 18, 1471. [6] A. Lendlein, H. Jiang, O. Junger, R. Langer, Nature 2005, 434, 879. [7] H. Finkelmann, E. Nishikawa, G. G. Pereira, M. Warner, Phys. Rev. Lett. 2001, 87, 015 501/1. [8] M. Camacho-Lopez, H. Finkelmann, P. Palffy-Muhoray, M. Shelley, Nat. Mater. 2004, 3, 307. [9] K. D. Harris, R. Cuypers, P. Scheibe, C. L. van Oosten, C. W. M. Bastiaansen, J. Lub, D. J. Broer, J. Mater. Chem. 2005, 15, 5043. [10] M.-H. Li, P. Keller, B. Li, X. Wang, M. Brunet, Adv. Mater. 2003, 15, 569. [11] A. Athanassiou, M. Kalyva, K. Lakiotaki, S. Georgiou, C. Fotakis, Adv. Mater. 2005, 17, 988. [12] S. V. Ahir, E. M. Terentjev, Nat. Mater. 2005, 4, 491. [13] S. Bian, D. Robinson, M. G. Kuzyk, J. Opt. Soc. Am. B 2006, 23, 697. [14] T. F. Scott, R. B. Draughon, C. N. Bowman, Adv. Mater. 2006, 18, 2128. [15] R. O. Al-Kaysi, A. M. Muller, C. J. Bardeen, J. Am. Chem. Soc. 2006, 128, 15 938.
[16] G. Kaupp, Angew. Chem. Int. Ed. Engl. 1992, 31, 595. [17] A. E. Keating, M. A. Garcia-Garibay, in Organic and Inorganic Photochemistry, Vol. 2 (Eds: V. Ramamurthy, K. S. Schanze), Marcel Dekker, New York 1998, p. 195. [18] I. Turowska-Tyrk, J. Phys. Org. Chem. 2004, 17, 837. [19] M. Irie, S. Kobatake, M. Horichi, Science 2001, 291, 1769. [20] L. J. Fitzgerald, R. E. Gerkin, Acta. Cryst. C 1997, 53, 71. [21] M. D. Cohen, Angew. Chem. Int. Ed. Engl. 1975, 14, 386. [22] Y. Ito, H. Fujita, J. Org. Chem. 1996, 61, 5677. [23] R. M. MacFarlane, M. R. Philpott, Chem. Phys. Lett. 1976, 41, 33. [24] B. Stevens, S. R. Perez, J. A. Ors, J. Am. Chem. Soc. 1974, 96, 6846. [25] Y. Zhang, S. Iijima, Phys. Rev. Lett. 1999, 82, 3472. [26] A. J. Spakowitz, Z. G. Wang, Phys. Rev. E 2001, 64, 061 802. [27] E. W. H. Jager, E. Smela, O. Inganäs, Science 2000, 290, 1540. [28] G. Alici, B. Mui, C. Cook, Sens. Actuators A 2006, 126, 396. [29] P. Metz, B. Alici, G. M. Spinks, Sens. Actuators A 2006, 130–131, 1. [30] M. Warner, L. Mahadevan, Phys. Rev. Lett. 2004, 92, 134 302/1. [31] A. Bondi, J. Phys. Chem. Solids 1967, 28, 649. [32] R. J. Roberts, R. C. Rowe, P. York, Powder Technol. 1991, 65, 139. [33] E. W. Wong, P. E. Sheehan, C. M. Lieber, Science 1997, 277, 1971. [34] H. Liang, M. Upmanyu, H. Huang, Phys. Rev. B 2005, 71, 241 403/1. [35] C. R. Martin, Science 1994, 266, 1961. [36] R. O. Al-Kaysi, C. J. Bardeen, Chem. Commun. 2006, 1224.
______________________
1280
www.advmat.de
© 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Mater. 2007, 19, 1276–1280
