J. Phys. Chem. B 2007, 111, 941-945
941
Biologically Inspired Path-Controlled Linear Locomotion of Polymer Gel in Air Songmiao Liang,†,‡ Jian Xu,*,† Lihui Weng,†,‡ Lina Zhang,*,‡ Xinglin Guo,‡ and Xiaoli Zhang‡ State Key Laboratory of Polymer Physics and Chemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100080, People’s Republic of China, and Department of Chemistry, Wuhan UniVersity, Wuhan 430072, People’s Republic of China ReceiVed: NoVember 8, 2006; In Final Form: December 11, 2006
A novel approach based on electrohydrodynamic behavior of a dielectric liquid pattern in electric field was developed to fabricate a poly(vinyl alcohol)/dimethyl sulfoxide (PVA/DMSO) gel electromechanical system. Driving experiments indicate that this system could be well-operated in air by using a direct current (DC) electric field, and the gel exhibits a long-range path-controlled snaillike or snakelike motion with a fast crawling speed of 14.4 mm/s. Some factors, such as the applied electric field and the mass of the gel on the average crawling speed of the gel at linear path and curvilinear path, are investigated. Furthermore, a transition between snaillike gaits and snakelike gaits of the gel is also further studied in this system. The mechanism analysis suggests that this path-controlled motion of the gel arises from the drag of the spatial varied shear force F originated from the electrohydrodynamic flow of the solvent in and out of the gel.
Introduction For the purpose of being able to reproduce the multifunctionality of biological muscles, the development of intelligent soft machines has attracted considerable attention in the past decades. Electroactive polymers (EAPs), such as polymer gels,1-5 conducting polymers,6-8 copolymers,9 elastomers,10,11 and ionic polymer metal composites,12,13 have emerged as potential materials for artificial muscles. These materials can be supported on or between compliant electrodes for the development of soft electromechanical systems based on the electrophoretic, piezoelectric, electroosmotic, Coulombic, or electrostrictive properties of them in an electric field. In general, the generation of forces or the output of works in these systems is primarily by two routes: angular movements and/or linear deformations. For example, devices based on polymer gels and conducting polymers have been reported to show a large range of bending angle from a few to more than 360°.4b,6 Electroelastomers based on acrylic copolymer elastomers could exhibit high electrically induced strains of up to 380% according to the expansion area.10b At this level, these materials are promising for carrying out certain functions of natural muscles in practical applications. However, for some special practical applications, there is a clear need for soft electromechanical systems with some special capability comparable to animals while not a single biological muscle. As is well-known, animals can control their motion path and employ different gaits in responding to the environmental stimulus. The introduction of such capability to an artificial electromechanical system would make it more approach biological reality and exploit an attractive application potential. Unfortunately, little effort has been paid for the development of such a soft electromechanical system by using those observed actuator materials. Although a few developed systems based on polymer gels or conducting polymers could display linear movements or inverse motion of a few millimeters to several centimeters * To whom correspondence should be addressed. E-mail:
[email protected] (J.X.);
[email protected] (L.Z.). † Chinese Academy of Sciences. ‡ Wuhan University.
in aqueous solution, the control of a real motion path in these systems has been a failure since the systems only can move or swim along a simple linear path at the regions between the electrodes.1b,14 Furthermore, with a limited response rate, the need for electrolyte solution for operation of these systems poses some very challenging packaging problems to allow a linear motion in air. Recently, researchers have developed some special biomimetic systems based on non-ionic polymer gels such as poly(vinyl alcohol)/dimethyl sulfoxide gel (PVA/DMSO) and poly(vinyl chloride)/di-n-butylphthalate (PVC/DBP).15-19 Compared with other systems, these systems exhibited some merits such as large electromechanical strain (8% in length) and fast bending rate (180° within 90 ms).15,18 Nevertheless, although creeping of PVC/DBP gel also has been studied by using a special contacted electric field, the creeping displacement and velocity were still limited at a low level.19 The long-range pathcontrolled motion of the gel remains unreachable due to the limitation of the sandwichlike electrode configuration. To resolve the above problems, our attention was focused on a simple phenomenon stated by electrohydrodynamics. It is well-known that electrohydrodynamics deals with the flow phenomenon of dielectric fluid induced by the external electric field.20,21 When a high voltage is applied across the electrodes, the electrical forces acting on a dielectric liquid containing free charge carriers (typically are ions or electrons) will draw it into the air gap of a plate capacitor or flow between the two electrodes. This interesting phenomenon has been widely applied to the studies of biophysical process,22 electrically induced pumping23 and mixing, and enhancement of heat transfer.24 In this paper, we attempt to develop such a gel-based electromechanical system which has great path-controlled capability based on this principle. Since DMSO is an incompressible dielectric liquid with high dielectric constant, we here prefer to use PVA/ DMSO gel to fabricate this novel system. Experimental Section Materials and Methods. Poly (vinyl alcohol) (>99%) with DP ) 1750 was purchased from Aldrich. Dimethyl sulfoxide
10.1021/jp0673821 CCC: $37.00 © 2007 American Chemical Society Published on Web 01/18/2007
942 J. Phys. Chem. B, Vol. 111, No. 5, 2007
Liang et al.
Figure 1. Sketch of the experimental setup.
was purchased from Beijing Yili Chemical Co. Ltd. and was re-distilled before use. PVA gel swollen by DMSO was prepared by endowing chemical cross-links in physically cross-linked PVA gel. A 10.0 g amount of PVA was dissolved in 90.0 g of mixture solvent containing 72.0 wt % DMSO and 18.0 wt % water at the solution boiling temperature. The resulting solution then was poured into a special glass mask and allowed to cool at -20 °C for 12 h. After the PVA solution formed into physically cross-linked gel, the resulting gel stack was immersed in 50.0 mL of 0.07 wt % glutaraldehyde solution at 5 °C for about 8 h. Adjusting the solution with hydrochloric acid to the levels of pH 2-3 and heating at 30 °C for 1 h, the gel was turned into a chemically crosslinked one. Finally, the chemical cross-linked gel block was immersed in methanol for de-swelling and then in DMSO for re-swelling, repeating this de-swelling and re-swelling process several times until the gel block was swollen fully by DMSO. The experimental setup is shown schematically in Figure 1. Two thin gold electrodes (length × width × thickness is 15.0 mm × 12.0 mm × 0.5 mm) were fixed on the glass substrate with an insulating adhesive film (the thickness is about 100.0 µm) and connected to opposite voltage terminals. The space between the electrodes is 20.0 mm. A DW 300-1 source was applied as the DC power. An ampere meter (FT 9801) with a precision of 0.01 µA was used to measure the microcurrent between the two electrodes when a voltage was applied across them. Prior to the driving test, a DMSO liquid line with designed pattern was cast on the glass substrate. The two ends of the line are connected with the two opposite electrodes, respectively. The width and thickness of the DMSO line is about 2.0 mm and 10.0 µm. In the driving process, a gel strip with designed size was located on the end of the DMSO line adjacent to the anode. By applying a DC electric field, the whole motion process was captured by a video camera (Wat-221S) from a top view at 25 frames/s. The images were digitized and analyzed using motion analysis software (CAPFIT). All driving experiments were directly performed in air under ambient circumstance of T ) 288 K and relative humidity of H ) 23.6%. The mass loss (u) of the gel samples within 120 s under a certain applied electric field is investigated. It was approximately calculated by u ) {(m1 - m2)/m1} × 100%. Here, m1 and m2 are the mass of the gel samples before the electric field was applied and after the electric field was applied, respectively. All of the gel samples (size, 10.0 mm × 10.0 mm × 4.0 mm) used in this measurement have a lower average swelling degree of 97.7 wt %. Results and Discussion Figure 2 shows the motion of the gel along a “Ω” mode path in the applied electric field of 300 V/mm. The real locomotion
Figure 2. “Ω” motion path of the gel in a DC electric field of 300 V/mm out of the two gold electrodes. The radius and length of the gel are 1.5 mm and 6.0 mm, respectively.
Figure 3. Schematic illustration of the gel with snaillike gait (left) and slide gait (right). “b” is a random dot on the gel.
path of the gel here is about 50.0 mm. We prepared an “artificial snail”, which is nothing more than a long cylindrical PVADMSO gel rod with a radius of 1.5 mm and a length of 6.0 mm located on the dielectric pattern. The swelling degree is 98.7 wt %. The result indicates that when the applied electric field exceeds a certain critical driving threshold (Ecg), the gel can crawl quickly from anode to cathode along the designed path. The average crawling speed (V) can reach 4.8 mm/s under the DC electric field of E ) 300 V/mm. Here, Ecg is defined as the minimum threshold of the applied electric field to drive the gel to move. During the driving process, the microcurrent with an average strength of 26.7 µA could be found, indicating the transport of charge carriers in the dielectric pattern.21 Video images captured from a top view indicate that the gel displays a snaillike gait with a periodic elongation and shrinkage of its body. At a purely physical level, two factors, the drag of the solvent and the flexible and soft character of the gel’s body, can be attributed to this motion gait. As schematically shown in Figure 3, left (AfBfC), when the gel moves at the linear part of the path, there is an axial elongation of the gel first along the flowing direction of the solvent due to the drag of the unidirectional shear force (F). For a small deformation (Figure 3 (left (B))), this behavior of the gel can be well-approximated by a linear Hook’s law so that the elastic force H ) S dx, where S is Young’s modulus of the gel and dx is the local strain of the polymer network. With the proceeding of the elongation of the gel’s body, the increasing elastic force of the polymer network would result in shrinkage of the gel’s body along the motion direction when it exceeds the local frictional force (f) between the gel and the substrate (Figure 3 (left (C))). As a result, one step is completed and the gel moves forward. The same gait would be possessed by the gel in the following motion as a result of the synergistic effect of H, f, and F. At an
Linear Locomotion of Polymer Gel in Air experimental biological level, the crawling distance from A to C was called a step length. It is worthy to note that the gel would exhibit a more complex and attractive motion state at a curvilinear part of the path, since the presence of a nonlinear variation drag force originated from the nonlinear flow of the solvent. Other than the periodic elongation and contraction, a bending deformation in the body of the gel also could be observed. This bending deformation was induced by an elongation discrepancy between each side of the gel’s body. That is, once the gel changed its motion direction, one side of the gel would be elongated more than the other by the dragging force F so that caused a bending deformation of the gel. This is wellaccorded with a ubiquitous phenomenon existing in terrestrial limbless animals.25-27 However, our results demonstrate that this snaillike motion only can be obtained by using the gel with a higher swelling degree, since the body of the gel in this case is easily elongated by the drag force before a slide motion occurs. For a strong cross-linked gel, experimental results demonstrate that, if a proper friction is kept between the gel and the substrate, another crawling gait, a sliding gait, can be obtained, as shown in Figure 3, right (A′fB′fC′). In this gait, F cannot elongate the body of the gel and induce a deformation of the polymer matrix before the sliding motion occurs. Therefore, the locomotion of the gel is only a simple and pure slide on the substrate. However, whether or not the snaillike or the pure slide gait is chosen by this system depends on a number of factors such as the interfacial properties between the gel and the substrate, the geometry and dimension of the gel, and the swelling degree of the polymer gels. To fabricate an ideal gel system which can move with a snaillike gait, the experimental results indicate that the proper range of the swelling degree for the gel is 98.599.2 wt %, while the body of the gel with Φ < 98.0 wt % is very difficult to elongate due to the large Young’s modulus, which favors of the pure slide gait on the glass substrate. Although the snaillike motion of the gel can be easily achieved by using this approach, investigation of the influence of some factors on Ecg is very important for a more precise control of the motion. On the basis of the definition of Ecg, it is reasonable to deduce that the critical driving electric field in this system strongly depends on the frictional force f. A smaller f will require a lower critical electric field and thus will improve the driven efficiency of the gel. Therefore, any factors that can decrease the friction force will decrease the critical driving electric field. Earlier studies have demonstrated that the friction force between the polymer gel and the substrate is strongly affected by many factors, such as the swelling degree of polymer gel (Φ) and the apparent contact area (A) between the polymer gel and substrate.28-30 To explore the influence of Φ on Ecg only, all studied samples have the same size as described in Figure. 2. The critical driving electric field as a function of the swelling degree is presented in Figure 4. The result suggests that an increase in the swelling degree of PVA/DMSO gel will lead to a decrease of the critical electric field. To investigate the relationship between the average crawling speed and the applied electric field, presented in Figure 5a,b is the average speed measured versus the applied electric field at linear path and curvilinear path for the same gel, respectively. The average crawling speed in this paper takes the form V ) S/t; here, S represents the length of the path that the gel crawls within time t. The gels used in this experiment have a cuboidlike shape with a size of 6.0 mm × 2.0 mm × 1.0 mm (length × width × height). The curvilinear path is the same as that shown in Figure 2. The results elucidate that both of the average
J. Phys. Chem. B, Vol. 111, No. 5, 2007 943
Figure 4. Relationship between the critical electric field and the swelling degree of the gel.
Figure 5. Average crawling speed of the gel (a) at the linear part of the path and (b) at the curvilinear part of the path as a function of the applied electric field.
crawling speeds increase with the increase of the electric field rapidly, which can be attributed to the increase of the dragging force as the electric field is increased. We also find that, at linear path, the average crawling speed even approaches V ) 14.4 mm/s under the electric field of E ) 400 V/mm. This value is larger than that of V ) 5.5 mm/s at the curvilinear path. This can be well-understood from the angle of the path to the applied electric field and the change in the curvature of the path. As shown in Figure 2, the curvature of this path is approximately calculated ranging from 0.11 to 0.27 mm-1 and the angle to the applied electric field ranges about from 0 to 133°. In comparison with that in the linear path, when the gel crawls at the curvilinear path, the presence of the angle will lead to a decrease of the effective electrostatic force that acts on the charge carriers in the solvent and therefore decrease the flow velocity of the solvent inside and outside the gel.21 According to hydrodynamic theory, this decrease in the flow velocity would produce a smaller dragging force, which then results in a lower crawling speed of the gel. On the other hand, the change of the motion state caused by the change in the curvature of the path during the motion process is also responsible for the result. That is, once the gel moves at the curvilinear path with larger curvature, it would be subjected to a larger hydrodynamic resistance than that at a linear path. The transformation of the motion state itself from the purely axial motion to the two-
944 J. Phys. Chem. B, Vol. 111, No. 5, 2007
Liang et al.
Figure 6. Average crawling speed of the gel as a function of its mass.
dimensional motion further limited the average crawling speed at the curvilinear path. In succession, we turn to consider the influence of the mass of the gel (m) on its average crawling speed at a linear path. Each gel has the same contact area of 5.0 mm × 3.0 mm with the glass substrate and the same swelling degree of 98.7 wt %. The change of the mass is obtained by a change in the height of the gel. This enables us to neglect the effect of the apparent contact area and the swelling degree on the friction force between the gel and the substrate according to the theory fabricated by Gong.28 By using an experimental electric field of 300 V/mm, as shown in Figure 6, the average crawling speed of the gel at linear path decreases with the increase of the gel mass, which can be approximately described by a scaling relationship V ∼ 1/m. Combined with Newtonian’s motion law, this scaling relationship is mainly due to the linear increase of the friction force with the increase of the gel mass.28-30 Furthermore, by designing a special dielectric liquid pattern comprising a linear path and a wavelike path (Figure 7a), our approach seems to be endowed with a special capability to realize the transition between the snaillike gait and the snakelike gait. As presented in Figure 7b, when an electric field exceeding Ecg was applied, guided by the flow of the designed fluid pattern, the gel can move with the snaillike gait along the linear path first and then transform into the snakelike gait when moving at a wavelike path. Reversible transition is achievable by changing the direction of the external electric field. In contrast to the method developed by Mahadevan, the natural transition between different gaits by using our approach is easier realize, since it does not need to design complex scales on the substrate.31 However, to obtain an ideal snakelike gait at the wavelike path, the wavelength of the wavelike path (λ) must be smaller than the length of the gel (a). On the basis of the above results, it is clear that the pathcontrolled biomimetic locomotion of the gel arises from the flow of DMSO inside and outside of the gel. For polymer gels with unique network structure, the flow of dielectric solvent inside the polymer network from the anode to the cathode has been proved as the case under DC electric field and which has been used to achieve an angular movement.18 Figure 8 shows the mass loss (u) of the poly(vinyl alcohol)/dimethyl sulfoxide gel within 120 s as a function of the applied electric field. The result indicates that the mass loss of the gel increases with increasing applied electric field, implying the flow velocity of DMSO in the polymer network increases with the increasing strength of the applied electric field. Hydrodynamic theory has stated that this flow of the solvent in the polymer network would give rise to a shear force or dragging force (F) whose vector direction is along the tangential direction of the liquid flow. As shown in Figure 9(bottom), the shear force in this system can be divided into F1 originated from the flow of solvent inside the polymer gel and F2 caused by the flow of the solvent pattern. When the
Figure 7. (a) Sketch of the designed dielectric liquid pattern; (b) transition between the snaillike gait and the snakelike gait in this pattern.
external electric field was applied across the electrodes, the flow of the solvent filament occurs first between the two electrodes, but no motion of the gel takes place in this case. After about several seconds, a rapid jetting of DMSO from the gel can be observed and leads to a rapid motion of the gel. This suggests that the locomotion of the gel is mainly arising from the drag of F1 compared with F2. Well, then, how can the system exhibit a novel path-controlled locomotion in the applied electric field? At the hydrodynamic level, this can be well-understood from the inducing flow of the DMSO line, as shown in Figure 9 (top).
Linear Locomotion of Polymer Gel in Air
J. Phys. Chem. B, Vol. 111, No. 5, 2007 945 gel. However, although some limitations such as the need of high electric field and the evaporation of solvent in air still exist in this system, they could be overcome to some extent through adjusting the swelling degree of the gel and designing the gels with nonvolatile solvent. We expect the technique we have used in this study could serve as a general motif to fabricate such an electromechanical system, which is suitable for more diverse applications in the bionic field.
Figure 8. Relationship between the applied electric field and the mass loss of the gel samples within 120 s.
Acknowledgment. We acknowledge The National Natural Science Foundation of China (Grant Nos. 50373049, 50425312, and 50521302) and the Innovation Project of CAS for financial support. References and Notes
Figure 9. Schematic illustration of the mechanism of the pathcontrolled motion.
Although the single inducing flow of the DMSO line cannot provide a dragging force enough to drive the motion of the gel, it is still a key parameter for yielding a spatial varied dragging force in the system. As a result, under the drag of the spatial varied force, the spatial change in its crawling direction would cause a motion of the gel along the flowing path of the dielectric pattern. This is the basis for the origin of the novel pathcontrolled capability in this system. Another important function of the DMSO line here that should be pointed out is that it provides an ideal continuum channel for the transportation of the injected charges in the electric field. By applying a DC electric field, the charges injected from the electrodes would transport in DMSO under the electrostatic force and then initiate the so-called ion-dragging flow of DMSO inside and outside of the gel. Conclusions At methodological level, a novel approach for fabricating such an electromechanical system with great path-controlled capability has been developed successfully by employing the dielectric liquid pattern. In this way, poly(vinyl alcohol)/dimethyl sufoxide gel could display a path-controlled snaillike or snakelike longrange locomotion on the glass substrate when an electric field exceeding Ecg was applied. The measurement results indicated that the crawling speed of the gel increased with the strength of the applied electric field and even could reach 14.4 mm/s at linear path under the electric field of 400 V/mm. The swelling degree of the gel has great influence not only on the critical driving electric field but also on the crawling gait of the gel. In contrast to those electromechanical systems developed earlier, a natural transition between the snaillike gait and the snakelike gait of the gel in the electric field was first obtained through design of a special dielectric liquid path. Analysis of the mechanism indicated that the path-controlled locomotion of this system is due to the drag of the spatial varied shear force F originated from the flow of DMSO inside and outside of the
(1) (a) Gong, J. P.; Nitta, T.; Osada, Y. J. Phys. Chem. 1994, 98, 9583. (b) Osada, Y.; Okuzaki, H.; Hori, H. Nature 1992, 355, 242. (c) Liu, Z. S.; Calvert, P. AdV. Mater. 2000, 12, 288. (d) Calvert, P.; Liu, Z. Acta Metall. Mater. 1998, 46, 2565. (2) Tanaka, T.; Nishio, I.; Sun, S. T.; Ueno-Nishio, S. Science 1982, 218, 467. (3) (a)Yoshida, R.; Uchida, K.; Kaneko, Y.; Sakai, K.; Kikuchi, A.; Sakurai, Y.; Okano, T. Nature 1995, 374, 240. (b) Zrı´nyi, M.; Fehe´r, J.; Filipcsei, G. Macromolecules 2000, 33, 5751. (c) Fehe´r, J.; Filipcsei, G.; Szalma, J.; Zrı´nyi, M. Colloids Surf., A 2001, 183-185, 505. (4) (a) Kwon, I. C.; Bae, Y. H.; Kim, S. W. Nature 1991, 354, 291. (b) Kim, S. J.; Kim, H., II; Shin, S. R.; Kim, S. I. J. Appl. Polym. Sci. 2004, 92, 915. (c) Kim, S. J.; Park, S. J.; Kim, I. Y.; Shin, M. S.; Kim, S. I. J. Appl. Polym. Sci. 2002, 86, 2285. (5) Frank, S.; Lauterbur, P. C. Nature 1993, 363, 334. (6) (a) Otero, T. F.; Corte´s, M. T. AdV. Mater. 2003, 15, 279. (b) Otero, T. F.; Grande, H.; RodrIÄguez, J. J. Phys. Org. Chem. 1996, 9, 381. (7) Jager, E. W. H.; Smela, E.; Ingana¨s, O. Science 2000, 290, 1540. (8) Takashima, W.; Kaucko, M.; Kaneto, K.; Macdiarmid, A. G. Synth. Met. 1995, 71, 2265. (9) Zhang, Q. M.; Bharti, V.; Zhao, X. Science 1998, 280, 2101. (10) (a) Pelrine, R.; Kornbluh, R.; Pei, Q. B.; Joseph, J. Science 2000, 287, 836. (b) Ha, S. M.; Yuan, W.; Pei, Q.; Pelrine, R.; Stanford, S. AdV. Mater. 2006, 18, 887. (11) Zhang, X.; Lo¨we, C.; Wissler, M.; Ja¨hne, B.; Kovacs, G. AdV. Eng. Mater. 2005, 7, 361. (12) (a) Shahinpoor, M.; Kim, K. J. Smart Mater. Struct. 2001, 10, 819. (b) Shahinpoor, M.; Kim, K. J. Smart Mater. Struct. 2005, 14, 197. (13) Shahinpoor, M. Electrochim. Acta 2003, 48, 2343. (14) Otake, M.; Kagami, Y.; Inada, M.; Inoue, H. Rob. Auton. Syst. 2002, 40, 185. (15) Hirai, T.; Nemoto, H.; Hirai, M.; Hayashi, S. J. Appl. Polym. Sci. 1994, 53, 79. (16) Hirai, T.; Zheng, J. M.; Watanabe, M. Electroactive Polymer Actuators and Devices; Smart Struct. Mater. 1999 1999, 3669, 209. (17) Hirai, T.; Nemoto, H.; Suzuki, T.; Hayashi, S.; Hirai, M. Intell. Mater. Syst. Struct. 1993, 4, 277. (18) Hirai, T.; Zheng, J.; Watanabe, M. Mat. Res. Soc. Symp. Proc. 2000, 600, 267. (19) Uddin, Z.; Watanabe, M.; Shirai, H.; Hirai, T. J. Polym. Sci., Part B: Polym. Phys. 2003, 41, 2119. (20) Saville, D. A. Annu ReV. Fluid Mech. 1997, 29, 27. (21) (a) Melcher, J. R.; Taylor, G. I. Ann ReV. Fluid Mech. 1969, 1, 111. (b) Castellanos, A. Electrohydrodynamics; Springer: Wien, Austria, NewYork, NY, 1998; pp 22-32. (22) Castan˜eda-Priego, R.; von Gru¨nberg, H. H.; Kollmann, M. J. Phys.: Condens. Matter. 2004, 16, 1. (23) Crowley, J. M.; Wright, G.; Chato, J. C. IEEE Trans. Ind. Appl. 1990, 26, 42. (24) Seyed-Yagoobi, J. J. Electrostat. 2005, 63, 861. (25) Mizutani, K.; Ogawa, H.; Saito, J.; Oka, K. J. Exp. Biol. 2002, 205, 265. (26) Jayne, B. C. J. Exp. Biol. 1988, 140, 1. (27) Alexander, R. M. Principles of Animal Locomotion; Princeton University Press: Princeton, NJ, 2003. (28) Gong, J. P.; Osada, Y. Prog. Polym. Sci. 2002, 27, 3. (29) Gong, J. P; Iwasaki, Y.; Osada, Y. J. Phys. Chem. B 1999, 103, 6001. (30) Kagata, G.; Gong, J. P.; Osada, Y. J. Phys. Chem. B 2002, 106, 4596. (31) Mahadevan, L.; Daniel, S.; Chaudhury, M. K. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 23.
