A nanomechanical device based on linear molecular motors
Tony Jun Huang, Branden Brough, and Chih-Ming Hoa)
Mechanical and Aerospace Engineering Department and the Institute for Cell Mimetic Space Exploration
420 Westwood Plaza, University of California, Los Angeles, California90095
Yi Liu, Amar H. Flood, Paul A. Bonvallet, Hsian-Rong Tseng, and J. Fraser Stoddarta)
Department of Chemistry and Biochemistry and the California NanoSystems Institute
603 Charles E. Young East Drive, University of California, Los Angeles, California90095-1596
Marko Baller and Sergei Magonov
Veeco Instruments, 112 Robin Hill Road, Santa Barbara, California93117
SUPPORTING
INFORMATION
FOR
ELECTRONIC PHYSICS AUXILIARY PUBLICATION SERVICE
(EPAPS)
S1. Experimental Procedures for Device Fabrication and Evaluation
S2. Molecular Modeling
S3. Molecular Force and Cantilever Bending Calculation
S4. Control Studies and Discussion
S5. Molecular Coverage
S1. Experimental Procedures for Device Fabrication and Evaluation
Materials: The MeCN used (i) in the UV/visible spectroscopic switching experiments and (ii) the formation of the self-assembled monolayers (SAMs) was purified using a solvent purification system (Anhydrous Engineering) in which the solvent is passed through a steel column packed with activated alumina. The Fe(ClO4)3 (Alfa Aesar) and ascorbic acid (Sigma) were used without prior purification, and solutions were prepared from MeCN and HPLC grade H2O immediately prior to use. The pH 3.2 and 3.5 buffer solutions (citrate-phosphate) were prepared from HPLC grade water. UV-visible spectra were recorded on a Varian Cary Bio100 spectrometer.
Cantilever Preparation: The silicon cantilever arrays employed in this experiment incorporate cantilever beams with a length of 500 m, a width of 100 m, a thickness of 1 m, and a spring constant of 0.02 Nm-1. They were fabricated using silicon-on-insulator (SOI) wafers. The gold coatings were prepared by thermally evaporating 1.5 nm of Ti, followed by 20 nm of Au, onto the topside of the cantilever devices. Immediately afterwards, the cantilever arrays were immersed in a 0.1 mM MeCN solution of the disulfide-tethered, bistable, redox-controlled [3]rotaxane R18+ for 48 h in order to form a SAM of the compound. Following SAM formation, the cantilevers were rinsed with MeCN to remove any unbound material.
Experimental Setup: The experimental setup is based on a Digital Instruments Scentris™ Research Tool. The cantilever array is housed in a 50 L sample chamber covered by a transparent window in order to record the cantilevers’ deflections by the beam-deflection method. Light from multiple superluminescent diodes emitting at 850 nm is coupled to an array of multimodefibers and projected onto the end of a single cantilever at a time. Uponreflection, the light is collected by a linear position-sensitive detector (PSD),and the photocurrents are converted into voltages. The position of the cantilever beam is determined within ± 0.5 nm or better. After amplification, thesignals are digitized and stored. A time-multiplexing scheme is applied by switching individual light sources on and off to illuminate only one cantilever at a time. Premixed reagent solutions (1 mM Fe(ClO4)3 and 2 mM ascorbic acid, each in HPLC grade water) from different reservoirs were sent continuously through the sample chamber into a waste container by gravitational force at a rate of 250–300 L/min. The sample reservoirs were connected to a computer-controlled six-way valve in order to minimize the disturbance of the flow when switching between the different solutions. Subsequent to the data collection, a baseline was subtracted from the data and the curves were smoothed in order to improve clarity.
Atomic Force Microscopic Measurements: A monolayer of the disulfide-tethered, bistable, redox-controlled [3]rotaxane R18+ on an atomically flat Au(111) surface (Molecular Imaging Inc.) was analyzed by an Atomic Force Microscope (AFM) (Veeco Instruments Inc., Nanoscope IIIa) in tapping mode. The surfaces were scanned at 1 Hz with 512 lines per image resolution using a 10 m × 10 m E scanner and silicon cantilevers (Veeco Instruments Inc., 120 m, tip radius 5–10 nm).
S2. Molecular Modeling
Computational modeling of R18+ was carried out for the free-body analysis and subsequent determination of the cantilever beam’s bending moment.
Three approximations were made: (1) For the sake of computational speed, only half of the dumbbell component of the molecule was modeled. (2) The ethylene glycol chains were constrained into gauche conformations with respect to the oxygen atoms. (3) The entire half-molecule was constrained to a linear conformation. A fragment of R18+, constrained to its fully elongated conformation (Fig. S1), was minimized with the AMBER* (Assisted Model Building and Energy Refinement) force field and GB/SA (Generalized Born/Surface Area) solvent model for CHCl3 as implemented in Maestro 3.0.038 module of the Schrödinger molecular modeling suite. Calculations of molecular force, molecular strain, and theoretical cantilever bending are based upon the contour length of R18+. It should be noted, however, that X-ray structural dataS1 on related pseudorotaxanes suggest that the equilibrium structure of R18+ will lie between the fully elongated structure shown in Fig. S1 and one in which the ethylene glycol chains are folded around the CBPQT4+ ring as a consequence of noncovalent [C-H···O] interactions.
S3. Molecular Force and Cantilever Bending Calculations
Fig. S2. Distance estimates based upon molecular modeling of the [3]rotaxane R28+.Estimate of Molecular Force:The mean molecular force of each individual TTF2+–CBPQT4+ interaction was estimated to be 20 pN based upon the Coulombic repulsion in a 1 mM aqueous solution (water= 80) arising from a 1 nm diameter ring, carrying four evenly dispersed positive charges, encircling a site containing two positive charges. Movement is assumed to proceed linearly up to a distance of 1.4 nm away from the starting position, as shown in Fig. S2.
Cantilever Bending Calculations: A single beam’s deflection upon contraction of the “molecular muscle” was analyzed using a quantitative model for the bistable, redox-controlled [3]rotaxane bound to a section of the beam (see Fig. 2 in main text). Oxidation of R18+produces a contraction of the inter-ring distance and correspondingly exerts a force F upon the beam. In order to simplify the free-body analysis, this force can be translated into an equivalent force F1 of equal magnitude plus a bending moment M,both acting along the neutral axis of the beam. While the force F1 results in the possible contraction of the beam, the bending moment M, with its lever arm approximated by half the thickness (D/2) of the cantilever, will result in a bending of the beam. The out-of-plane displacement wof a cantilever beam is governed by the Euler-Bernoulli Beam equation,
(1)
where Mbeam is the moment on the beam, L is the total length of the cantilever beam, E is the Young’s modulus of the cantilever, and I is the area moment of inertia of the beam’s cross section. The parameter Mbeam can be obtained from , where N is the number of molecules along the width of the beam and M is the moment generated by a single molecule, which equals half of the beam thickness multiplied by the estimated force generated (40 pN) by a single molecule. The parameters f1, f2, and f3 are constants corresponding, in turn, to the surface coverage, the idealized geometry, and the assumed random orientation of the molecule, respectively. Since only the portion between the two rings will generate a moment and contribute to bending, f2 = L1 / L2 . Here, L1 is the distance between two rings at their starting positions, and L2 is the contour length of the molecule. Considering a molecule in its extended inter-ring conformation, L1 is 4.2 nm while L2 is 7.4 nm. Since the molecules are assumed to be randomly aligned and only the force exerted along the longitudinal axis of the beam can contribute to bending, f3 equals . Based on this simplified model, the force generated by the “molecular muscle” can act against the spring-like restoring force of the cantilever beam to produce a theoretical beam displacement w of 48 nm.
S4. Control Studies and Discussion
The movement of the cantilever beams is correlated with the cycling of the oxidant and reductant solutions. In solution, these same chemical reagents cause the contraction and extension of the inter-ring distance of the model bistable, redox-controlled [3]rotaxane R28+. For the cantilever experiment, in which the molecules are mounted on a solid surface, it is important to consider other factors that can potentially contribute to the bending of the beams. Changes in solution pH, for example, have been demonstrated to cause silicon cantilevers to bend.S2 A control experiment (Fig. S3) with citrate buffer solutions, which have the same pH as the oxidant and reductant solutions (3.2 and 3.5, respectively) establish, however, that pH variations have an insignificant effect on the bending of cantilever beams, coated with a SAM of R18+. Furthermore, the direction of beam bending does not correlate in any repeatable way with the pH of the surrounding solution. Control experiments conducted in solution also suggest that the bistable, redox-controlled [3]rotaxane R18+is stable to acid for extended periods of time, indicating that pH-controlled chemical decomposition does not affect significantly the molecular SAM that coats the cantilever beams. Specifically, the model [3]rotaxane R28+ displays no apparent UV/visible and 1H NMR spectroscopic changes following exposure to trifluoroacetic acid over a two-week period.
Fig. S3. Control experiment demonstrating that cantilever beams coated with the [3]rotaxane R18+ do not display any significant deflections when exposed to the same variation in pH (3.2 and 3.5) that accompanies the addition of oxidant and reductant, respectively. The position of each of the four individual cantilevers is represented in pink, green, blue and purple, respectively. The data are baseline corrected.Other factors, such as charge repulsion and thermal effects also may contributeS3 to beam bending. Upon oxidation, the electronic charge of the disulfide-tethered [3]rotaxane R18+ changes from +8 to +12, giving rise to a change in the electrostatic interactions between neighboring surface-bound [3]rotaxane molecules. The proximity of two like-charged molecules directly side-by-side on the cantilever surface could create a repulsive force, thus leading to the beam bending downwards. Similarly, when considering thermal effects, the release of energy upon oxidation of the [3]rotaxane would cause the bimorph-like beams to bend downwards on account of the higher thermal expansion coefficient of the gold upper surface in comparison with the silicon lower surface.S4 In addition, subtle thermal effects are negligible in aqueous solutions because of the high heat capacity of water. Finally, the photothermal bending effectS5 initiated by the 850 nm detection laser beam is discounted, both on account of water’s high heat capacity and by the fact that the incident laser light with a spot size of 3 m irradiates an insignificant amount (0.02%) of the cantilever’s total surface area.
S5. Molecular Coverage
The molecular coverage of the disulfide-tethered [3]rotaxane R18+on the cantilever beams has been estimated by AFM. Although it was observed that molecular aggregates form in a few places, Fig. S4 shows that the Au surface has been well covered by a fairly uniform monolayer. The root mean square (rms) roughness is only 0.225 nm.
Fig. S4. Tapping-mode AFM image of a SAM of the disulfide-tethered [3]rotaxane R18+ on an atomically flat gold surface.Islands with a diameter of 5–8 nm are observed on the surface. Considering that the size of the rotaxane is 1 nm × 8 nm, we assume that the islands seen in the AFM image represent single [3]rotaxane molecules. From the AFM image shown in Fig. S4, we estimate the molecular coverage to be >90%.
REFERENCES
S1.P. R. Ashton, D. Philp, N. Spencer, J. F. Stoddart, and D. J. Williams, Chem. Commun. 181 (1994).
S2. J. Fritz, M. K. Baller, H. P. Lang, T. Strunz, E. Meyer, H. J. Guntherodt, E. Delamarche, C. Gerber, and J. K. Gimzewski, Langmuir16, 9694 (2000).
S3.J. Varesi, J. Lai, T. Perazzo, Z. Shi, and A. Majumbar, Appl. Phys. Lett. 71, 306 (1997).
S4. A. M. Moulin, R. J. Stephenson, and M. E. Welland, J. Vac. Sci. Technol. B 15, 590 (1997).
S5. J. R. Barnes, R. J. Stephenson, M. E. Welland, C. Gerber, and J. K. Gimzewski, Nature 372, 79 (1994).
S1