Reciprocating Power Generation in a Chemically Driven Synthetic

S3 7HF, U.K., CCLRC Daresbury Laboratory, Warrington WA4 4AD, U.K.,. DUBBLE CRG, ESRF, 6 rue Jules Horowitz, BP 220, F-38043 Grenoble Ce´dex 9,...
0 downloads 0 Views 240KB Size
Download PDF
NANO LETTERS

Reciprocating Power Generation in a Chemically Driven Synthetic Muscle

2006 Vol. 6, No. 1 73-77

Jonathan R. Howse,† Paul Topham,† Colin J. Crook,† Anthony J. Gleeson,‡ Wim Bras,§ Richard A. L. Jones,| and Anthony J. Ryan*,† The Polymer Centre, Department of Chemistry, The UniVersity of Sheffield, Sheffield S3 7HF, U.K., CCLRC Daresbury Laboratory, Warrington WA4 4AD, U.K., DUBBLE CRG, ESRF, 6 rue Jules Horowitz, BP 220, F-38043 Grenoble Ce´ dex 9, France, and The Polymer Centre, Department of Physics & Astronomy, The UniVersity of Sheffield, Sheffield S3 7RH, U.K. Received October 18, 2005; Revised Manuscript Received November 29, 2005

ABSTRACT A scalable synthetic muscle has been constructed that transducts nanoscale molecular shape changes into macroscopic motion. The working material, which deforms affinely in response to a pH stimulus, is a self-assembled block copolymer comprising nanoscopic hydrophobic domains in a weak polyacid matrix. A device has been assembled where the muscle does work on a cantilever and the force generated has been measured. When coupled to a chemical oscillator this provides a free running chemical motor that generates a peak power of 20 mW kg-1 by the serial addition of 10 nm shape changes that scales over 5 orders of magnitude. It is the nanostructured nature of the gel that gives rise to the affine deformation and results in a robust working material for the construction of scalable muscle devices.

Motility is a nearly ubiquitous feature of living systems and is driven by the ability directly to convert chemical energy into mechanical work. Synthetic polyelectrolytes provide a simple analogy to this through the volume change associated with ionization; this effect was used to build a single stroke “artificial muscle” more than fifty years ago. Macroscopic applications of such stimulus responsive gels are limited, however, because the volume change relies on mass transport of solvent and is limited by the gel-response rate. For this reason, more recent efforts to make synthetic muscles have concentrated on actuation by temperature, light, or electric fields. Recognizing the fundamental limitation of chemical actuation, we have constructed a scalable responsive-gel from a robust, self-assembled block copolymer comprising hydrophobic, glassy end-blocks and a weak polyacid mid-block. The gels deform affinely in response to a pH stimulus with a volume change of 3. When coupled to a chemical oscillator, this provides a free running chemical motor that generates a peak power of 20 mW kg-1 by the serial addition of 10 nm shape changes that scales over 5 orders of magnitude to provide reciprocating macroscopic motion. Taking a global view, a muscle is a device for translating chemical energy into mechanical work under isothermal * Corresponding author. E-mail: [email protected]. † The Polymer Centre Department of Chemistry, The University of Sheffield. ‡ CCLRC Daresbury Laboratory. § DUBBLE CRG, ESRF. | The Polymer Centre, Department of Physics & Astronomy, The University of Sheffield. 10.1021/nl0520617 CCC: $33.50 Published on Web 12/10/2005

© 2006 American Chemical Society

conditions.1 The most common way of achieving this couples a molecule that can change its shape in response to a chemical stimulus with a cyclic chemical change. Mammalian muscle,2,3 with its exquisite structure and function, offers the most familiar example from Nature of this energy conversion process, but the existence of many other types of biological motors shows that there can be many different implementations in detail of the general principle. The microtubule-based kinesin motors and actin-based myosin motors generate motions associated with intracellular trafficking, cell division, and muscle contraction.2 Modular design has given rise to a remarkable diversity of kinesin and myosin motors whose motile properties are optimized for performing distinct biological functions. But not all biological movements are caused by molecular motors sliding along filaments or tubules.3 Just as springs and ratchets can store or release energy and rectify motion in physical systems, their analogues can perform similar functions in biological systems. The energy of biological springs is derived from hydrolysis of a nucleotide or the binding of a ligand, whereas biological ratchets are powered by Brownian movements of polymerizing filaments. However, the viscous and fluctuating cellular environment and the mechanochemistry of soft biological systems constrain the modes of motion generated and the mechanisms for energy storage, control, and release.1-3 Early attempts4 to build a synthetic muscle were based on the volume change that occurs in polyelectrolyte networks

when their ionization state changes. On a macroscopic level, these systems are of limited practical use because the response rate is limited by diffusion and inhomogeneous stresses caused by expansion often lead to fracture of the network. The problems associated with responsive gels were elegantly demonstrated by Tanaka.5 Interestingly, these problems become much less severe as the length scale of the system is reduced, and Beebe et al.6 have demonstrated functional valves for microfluidic devices based on micrometer scale structures from responsive hydrogels. The emphasis in recent work in artificial muscles has, however, largely switched to systems which use either thermal or electrical actuation via electromechanical stress or electro-osmosis. While this has grown into a dynamic field with potentially useful results,7 this approach does fundamentally depart from the biological model in that chemical energy is not directly converted into mechanical work. In our approach we have returned to the original idea of exploiting the coil-globule transition of a polyelectrolyte in solution in response to a change of ionization state. To this classical basis we add two additional ingredients. To go beyond a simple one-off actuation step and achieve a free running motor, we couple the molecular response to an oscillating chemical reaction.8 To overcome the inevitable problems of inhomogeneous stresses in macroscopic, chemically cross-linked gels, we use a self-assembled structure in which the active polymer strands are held in a regular, nanoscale array. This also allows us to directly measure the size of the individual polymer chains as they respond to the changing chemical conditions. We have been able to measure both the mechanical and structural response, demonstrating that the system can generate a specific power of 20 mW kg-1 and deforms affinely with a 10 nm molecular shape change that scales over 5 orders of magnitude. At the macroscopic level, a prototype of a free-running chemical motor is provided in pioneering studies carried out by Yoshida and co-workers.9-12 They have coupled the collapse phenomenon in macroscopic responsive gels with oscillating chemical reactions (such as the BelousovZhabotinsky, BZ, reaction) to create conditions where “pseudo”-nonequilibrium systems which maintain rhythmical oscillations can be demonstrated, in both quiescent10 and continuously stirred11 reactors. The ruthenium complex of the BZ reaction was introduced as a functional group into poly(N-isopropylacrylamide) (PNIPAM), which is a temperature-sensitive polymer.9 The ruthenium group plays its part in the BZ reaction, and the oxidation state of the catalyst changes the collapse temperature of the gel.11 The result is, at intermediate temperature, a gel whose shape oscillated (by a factor of 2 in volume) in a BZ reaction, providing an elegant demonstration of oscillation in a polymer gel. This system, however, is limited by the concentration of the catalyst which has to remain relatively small12 and suffers the lack of scalability and inherent fragility associated with covalently cross-linked systems. The use of a self-assembled system was suggested in the context of a thermally actuated gel by de Gennes.13 A version of this scheme, in which an acrylate triblock copolymer self-assembles to form alternating 74

Figure 1. (a) A schematic illustration of the block copolymer structure in the neutral form (acid) and the ionized form (base) with, below, a sketch of the single molecule conformation and the gel morphology. The gel morphology and macroscopic structures in the collapsed and expanded states are characterized by smallangle X-ray scattering patterns (b) and the photographs (c), respectively.

layers of a rubber and a thermotropic liquid crystal, has been experimentally realized14 and shows reversible expansion and contraction of 18% in length over a temperature interval of 120 K. Our working material is a triblock copolymer with hydrophobic end-blocks of poly(methyl methacrylate) and a mid-block of poly(methacrylic acid), see Figure 1. The polymer was made by anionic polymerization,15 has a molecular weight of 155 kg mol-1, a polydispersity of 1.04, and a hydrophobe volume fraction of 0.16. The bulk morphology of block copolymers is well predicted by selfconsistent field theory16 and the equilibrium structure of this polymer should be a body-centered cubic array of PMMA spheres in a PMAA matrix. To prepare films a 40% polymer solution in a 85:15 methanol/acetone mixture was cast onto a slide. The solvent was slowly removed, and an elastic film was formed prior to complete drying. Both polymers have a high glass transition so the film was transferred into water to remove the remaining solvent by osmosis; water acts as a plasticizer for the PMAA matrix yielding an elastic film with physical cross-links of PMMA. Nano Lett., Vol. 6, No. 1, 2006

Figure 2. (a) The response of the macroscopic structure (gel length) (i) and gel morphology (SAXS d spacing) (iii) to the variation in pH (ii) induced by the bromate/sulfite/ferrocyanide oscillating reaction. (b) Time course of the relative size change for the morphology and the macroscopic size (dt/dt)0). (c) The rate of expansion (∆d/∆t) plotted with the pH during the time course of one oscillation.

A schematic diagram of single molecules in Figure 1a shows the bridging conformation where the hydrophobic chain ends are trapped in glassy hydrophobic domains. Figure 1b presents synchrotron small-angle X-ray scattering (SAXS)15,16 from these materials with a single peak arising from a liquid-like structure factor rather than the predicted16 equilibrium cubic structure. That the glassy domains had no long range order was confirmed by atomic force microscopy (AFM) (see Figure 3), and this is represented in their liquidlike packing in the morphology cartoon of Figure 1a. The photographic panels on the left of Figure 1c show the macroscopic size of film in collapsed and expanded states. At low pH the polyacid is protonated and neutral, water is a poor solvent, and the matrix polymer collapses. SAXS identifies a characteristic length scale of 28 nm, which corresponds to glassy domains with a diameter of ∼2 nm having an aggregation number of Nagg ∼ 100. At high pH the polyacid is ionized, water is a good solvent, and there is electrostatic repulsion of like charges which causes the matrix polymer to expand. SAXS identifies a characteristic length scale of 39 nm. The size, shape, and Nagg of the glassy domains is unaffected by the change in conformation of the mid-block, but the interdomain spacing increases by 11 nm. These changes in gel morphology are also illustrated schematically in Figure 1a. To construct a free-running motor, the polymer film was placed in an oscillating reaction. The Landolt pH-oscillator is based on a bromate/sulfite/ferrocyanide reaction and has a room-temperature period of 20 min with a range of 3.1 < pH < 7.0. The reaction is a development of a range of pH oscillators described by Rabai, Orban, and Epstein8a and was conducted in an open continuously stirred tank reactor (CSTR), with a peristaltic pump supplying the feed solutions Nano Lett., Vol. 6, No. 1, 2006

of potassium bromate, sodium sulfite, potassium ferrocyanide, and sulfuric acid and also pumping out bulk solution to keep the volume in the CSTR constant. The optimum flow rate for reliable oscillations was found to be 29 µL s-1 with concentrations [SO32-] ) 0.075 M, [BrO3-] ) 0.065 M, [Fe(CN)64-] ) 0.02 M, and [H2SO4] ) 0.01 M. A typical pH-time course is illustrated in the central panel of Figure 2a. Figure 3a illustrates the special cell designed to study the pH response of the block copolymer. A nylon cell was fitted with two sets of orthogonal windows such that the macroscopic size of the gel could be captured by an inspection microscope and the synchrotron X-ray beam could be used to perform real-time SAXS on beamlines 16.1 at the CLRC Daresbury facility17 and BM26 DUBBLE at the ESRF.18 The solutions for the oscillating chemical reaction were supplied via a peristaltic pump, and pH was recorded using a combined microelectrode (Mettler Toledo) connected to a PC situated outside of the experimental area. SAXS images were collected at 60 s per frame for a total of 160 frames while situated within the oscillating chemical reaction. The time course of the characteristic length scales in the gel are shown in Figure 2a for macroscopic (upper graph) and molecular (lower graph) dimensions. As the pH oscillates, the gel flips from expansion to contraction as it passes pH 5.5 (the pKa of poly(methacrylic acid)). The gel continues to expand, through mass transport of solvent, throughout the whole 10 min period at high pH; contraction, however, is significantly faster and there is a period of constant length prior to expansion. The molecular size, as determined by the separation of the hydrophobic clusters, closely tracks the sample size. Figure 2b shows quite clearly that the gel 75

Figure 3. (a) A schematic diagram of the experimental apparatus for simultaneously measuring pH morphology and force generation. In this device the block copolymer acts as a muscle by doing macroscopic work and bending a cantilever. (b) An AFM micrograph of the dry polymer structure with inserts showing the FFT of the structure (i) and the SAXS patterns of the stretched (ii) gel. (c) The response to the variation in pH (i) of the gel length (ii) and the force acting on the cantilever (iii) to the specific power generated (iv) by the gel.

behaves affinely over the whole range of length scales from the molecular to the macroscopic, and Figure 2c shows that the maximum rate of change of the molecular dimensions is the same in both expansion and contraction. This object, fabricated from self-assembling macromolecules, affinely changes shape in response to its environment over many cycles. That is, the serial addition of many molecular length changes makes the macroscopic length change. A movie, combining images of gel expansion and contraction with contemporaneous SAXS patterns, a cartoon of the morphology, and the time course of the pH, is available as Supporting Information. The specific power was measured by constructing a muscle device where the gel does work against a soft cantilever (0.47 N m-1) while being driven by the oscillating reaction. The orientation and relaxation of the structure were measured contemporaneously by SAXS, and the experimental setup is shown in Figure 3a. The cantilever spring was first calibrated for its force constant, and then the deflection of the cantilever was measured by a reflected laser beam. Thus the molecular structure, macroscopic deformation, and force could be measured. Figure 3b is an AFM tapping mode image of the collapsed block copolymer film, and the insets are the Fourier transform of the image (i) (cf. Figure 1 SAXS patterns) and the SAXS patterns of the film as it is stretched by the spring at pH 3; that the cantilever spring deforms the molecules and the gel structure is apparent from the SAXS pattern changing from a circle for the free gel to an ellipse indicating an increase in the length scale in the vertical axis. 76

Figure 3c is a plot of the time course of the pH, length of the gel, l, the force, f, on the cantilever spring, and the derived quantity the instantaneous specific power, Ps. The specific power is the power per unit mass obtained by normalizing to the mass of the swollen gel and provides a lower boundary estimate of Ps ) 20 mW kg-1. As a figure of merit, the specific power has been used previously2 to compare the performance of various cellular engines to thermal energy (kBT), and an automobile engine ∼300 W kg-1. The calculations of specific power are based on the molecular weight of the smallest unit. This favors the molecular motors and polymerization-based engines, which are more powerful than the cellular structures in which they are found, and can be seen in the comparison between myosin at 20 kW kg-1 and striated muscle at 200 W kg-1. Strictly speaking, one ought to compare values of the specific power averaged over a complete motor cycle rather than an instantaneous specific power, but the latter value sets the order of magnitude. The block copolymer based muscle is obviously rather weak, a million times less powerful than myosin, 10000 times weaker than striated muscle, and comparable to a eukaryotic spindle at 30 mW kg-1. Moreover the efficiency, calculated as the amount of chemical energy introduced to the oscillating reaction divided by the work done on the spring, is vanishingly small because the chemical reaction takes place in the whole volume of the cell. This number improves significantly, to ∼0.1, when the energy input is estimated from the volume of the gel. Seen in the context of a prototype for a polymer-based molecular engine this is, however, rather encouraging. Nano Lett., Vol. 6, No. 1, 2006

The volume transition of responsive gels represents a direct, macroscopic manifestation of the conformational response of the individual molecules making up the gel.4 Responsive gels have in the past been enthusiastically greeted as candidates for a new generation of intelligent materials with sensor, processor, and actuator functions. A wide variety of different stimuli-responsive gels have been developed for specific applications such as drug release and actuators for artificial muscle.5 However, macroscopic applications of stimulus responsive gels have not developed as fast as was initially anticipated. The reason for this is the fact that diffusion limits the rate at which the gel can respond to changes in environment, and that also applies to the example herein. The lack of robustness and scalability, the Achilles heel of conventional cross-linked polymer, are removed by the use of a triblock copolymer that gives homogeneous distribution of fixed cross-links, by self-assembly, and deforms affinely due to its isotropic structure. Coupling this structure to an oscillating chemical reaction generates a free running motor in which the reciprocating power stroke has been quantified. The unique aspect of having a gel structured on the nanoscopic level is that the homogeneous nature of the shape change at the nanoscale means that the device actually works. Conventionally cross-linked materials do not normally respond affinely due to local structural inhomogenieties that are a relic of the cross-linking process.19 This leads to inhomogeneous swelling and a local build-up of stress such that cross-linked polymers that undergo swellingdeswelling cycles tear themselves apart.4,5,19 We have demonstrated nanostructured gels that lead to affine deformation over 6 orders of magnitude in length scale. There are three obvious candidates for increasing the specific power of these polymer-based devices. First the material could be made stiffer so larger forces are developed. Second, the repetition rate of the oscillating reaction can be reduced, with a corresponding increase in dpH/dt. Third, the diffusion length can be altered by fabricating a device made from microfibers of responsive gel. The latter is the most likely candidate for improvement in the specific power. A change in dimensions of the gel of a factor of 100, by making microfibers, could yield an improvement in the rate of length change of 104, and we are currently exploring routes to do this, making block copolymer devices that could compete with striated muscle in specific power if not efficiency.

Nano Lett., Vol. 6, No. 1, 2006

In summary, we have constructed a synthetic muscle device by combining a scalable responsive gel comprising a robust, self-assembled, nanostructured block copolymer with a chemical oscillator. The gels deform affinely in response to a reciprocating pH stimulus with a volume change of 3. This is a significant advance providing a free running chemical motor that generates a peak power of 20 mW kg-1. It is completely scalable, as the mechanism of operation is the serial addition of molecular shape changes, and can provide reciprocating motion over 6 orders of magnitude in length scale, from nanometers to millimeters. Acknowledgment. This work was funded by EPSRC Grants (GR/T11562, GR/S47496, GR/R77544) and by ICI plc. Supporting Information Available: A video showing images of gel expansion and contraction with contemporaneous SAXS patterns, a drawing of the morphology, and the time course of the pH. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Jones, R. A. L. Soft Machines; Oxford University Press: Oxford, U.K., 2004. (2) Vale, R. D.; Milligan, R. A. Science 2000, 288, 88. (3) Mahadevan, M.; Matsudaira, P. Science 2000, 288, 95. (4) Kuhn, W. Experientia 1949, 5, 318. Katchalsky, A. Experientia 1949, 5, 319. Katchalsky, A.; Sussman, M. V. Science 1970, 167, 45. (5) Shibayama, M.; Tanaka, T. AdV. Polym. Sci. 1993, 109, 1. (6) Beebe, D. J.; et al. Nature 2000, 404, 588. (7) See, for example: Baughman, R. H. Science 2005, 308, 63. Shahinpoor, M.; Kim, K. J. Smart Mater. Struct. 2004, 13, 1362. (8) Rabai, G.; Orban, M.; Epstein, I. R. Acc. Chem. Res. 1990, 23, 258. (9) Yoshida, R.; Takahashi, T.; Yamaguchi, T.; Ichijo, H. J. Am. Chem. Soc. 1996, 118, 5134. (10) Yoshida, R.; Kokufuta, E.; Yamaguchi, T. Chaos 1999, 9, 260. (11) Yoshida, R.; Otoshi, G.; Yamaguchi, T.; Kokufuta, E. J. Phys. Chem. A 2001, 105, 3667. (12) Yoshida, R.; Takei, K.; Yamaguchi, T. Macromolecules 2003, 36, 1759. (13) de Gennes, P.-G. C. R. Acad. Sci., Ser. IIb 1975, 281, 101. (14) Li, M.-H.; et al. AdV. Mater. 2005, 16, 1922. (15) Ryan, A. J.; et al. Faraday Discuss. 2005, 128, 55. (16) Matsen, M. W.; Bates, F. S. Macromolecules 1996, 29, 7641. (17) Fairclough, J. P. A.; et al. Polymer 1999, 41, 2577. (18) Bras, W.; et al. J. Appl. Crystallogr. 2003, 36, 791. (19) Polymer Networks; Stepto, R. F. T., Ed.; Springer: Berlin, Germany, 1998.

NL0520617

77