Langmuir and Langmuir−Blodgett Films of Amphiphilic Bistable

Langmuir 2004, 20, 5809-5828

5809

Langmuir and Langmuir-Blodgett Films of Amphiphilic Bistable Rotaxanes Isaac C. Lee and Curtis W. Frank* Center on Polymer Interfaces and Macromolecular Assemblies and Department of Chemical Engineering, Stanford University, Stanford, California 94305-5025

Tohru Yamamoto, Hsian-Rong Tseng, Amar H. Flood, and J. Fraser Stoddart* The California NanoSystems Institute and Department of Chemistry and Biochemistry, University of California, Los Angeles, Los Angeles, California 90095-1569

Jan O. Jeppesen Department of Chemistry, Odense University (University of South Denmark), Campusvej 55, DK-5230, Odense M, Denmark Received November 15, 2003. In Final Form: April 7, 2004 A series of amphiphilic bistable [2]rotaxanessin which a ring-shaped component, the tetracationic cyclophane, cyclobis(paraquat-p-phenylene), has been assembled around two recognition sites, a tetrathiafulvalene (TTF) unit and a 1,5-dioxynaphthalene (DNP) ring system, situated apart at different strategic locations within the central polyether section of an amphiphilic dumbbell component that is terminated by a hydrophobic tetraarylmethane-based stopper (near the TTF unit) at one end and by a hydrophilic tetraarylmethane-based stopper (near the DNP ring system) at the other endshas been designed and synthesized. The effects of systematic changes in the constitutions of the three ethylene glycol tails (diethylene or tetraethylene glycol) and end groups (hydroxyl or methoxyl functions) attached to the hydrophilic stoppers on Langmuir film balance and surface rheology experiments at 20 °C were examined to determine the monolayer stabilities and co-conformations of the [2]rotaxanes and their free dumbbell counterparts. These experiments allow us to propose a model for the rotaxane’s structures at different surface pressures. All the [2]rotaxanes form stable Langmuir films. These films typically pass from a liquid-expanded region to a liquid-condensed region. The transition between the two regions was either directly observed or ascertained using film stability experiments. Film balance and surface rheology experiments showed that the addition of the tetracationic cyclophane component and hydroxyl end groups markedly increased the stabilities and viscoelasticity of the films.

Introduction [2]Rotaxanes comprise a class1 of mechanically interlocked molecules wherein a dumbbell components consisting of a central linear section, terminated by two bulky stopperssis encircled by a ring component. To be able to make such compounds with a modicum of efficiency, it is necessary to locate sites that recognize each other in the dumbbell and ring components.2 Then, templatedirected synthesis2,3 can be brought into play. The minimum requirement is the presence of one recognition site in the dumbbell. When two are present, however, the efficiency of the production of the [2]rotaxane can often be doubled. Moreover, if the dumbbell is constitutionally symmetrical, then the resulting [2]rotaxane will be capable of undergoing a degenerate process wherein the ring is darting back and forth between two equivalent recognition * To whom correspondence should be addressed. E-mail: [email protected]; [email protected]. (1) (a) Schill, G. Catenanes, Rotaxanes, and Knots; Academic Press: New York, 1971. (b) Molecular Catenanes, Rotaxanes and Knots; Sauvage, J.-P., Dietrich-Buchecker, C. O., Eds.; Wiley-VCH: Weinheim, 1999. (2) Stoddart, J. F.; Tseng, H. R. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4797-4800. (3) (a) Anderson, S.; Anderson, H. L.; Sanders, J. K. M. Acc. Chem. Res. 1993, 26, 469-475. (b) Schneider, J. P.; Kelly, J. W. Chem. Rev. 1995, 95, 2169-2187. (c) Raymo, F. M.; Stoddart, J. F. Pure Appl. Chem. 1996, 68, 313-322. (d) Templated Organic Synthesis; Diederich, F., Stang, P. J., Eds.; Wiley-VCH: Weinheim, 1999.

sites on the dumbbell. We call such a system a molecular shuttle.4 When constitutional asymmetry is introduced into the dumbbell, it becomes possible, in the situation where one recognition site is more favored than the other by the ring, to construct a controllable molecular shuttle.5 The main requirement is that one is able to switch off the recognition at the preferred site, leaving the ring with no choice other than to travel in the direction of the alternative recognition site and to surround it. The “jolts” that have been delivered to induce this ring movement include chemical,6 electrochemical,7 electrical,8 and optical9 stimuli. These controllable molecular shuttles, based on bistable rotaxanes, offer10 a unique opportunity to develop functioning nanosystems that include actuators, amplifiers, sensors, switches, and valves. Before such nanosystems can become reality, however, ways have to be developed to introduce bistable rotaxanes (4) (a) Anelli, P.-L.; Spencer, N.; Stoddart, J. F. J. Am. Chem. Soc. 1991, 113, 5131-5133. (b) Anelli, P.-L.; Asakawa, M.; Ashton, P. R.; Bissell, R. A.; Clavier, G.; Go´rski, R.; Kaifer, A. E.; Langford, S. J.; Mattersteig, G.; Menzer, S.; Philp, D.; Slawin, A. M. Z.; Spencer, N.; Stoddart, J. F.; Tolley, M. S.; Williams, D. J. Chem. Eur. J. 1997, 3, 1113-1135. (c) Leigh, D. A.; Troisi, A.; Zerbetto, F. Angew. Chem., Int. Ed. 2000, 39, 350-353. (d) Cao, J. G.; Fyfe, M. C. T.; Stoddart, J. F.; Cousins, G. R. L.; Glink, P. T. J. Org. Chem. 2000, 65, 1937-1946. (e) Elizarov, A. M.; Chang, T.; Chiu, S. H.; Stoddart, J. F. Org. Lett. 2002, 4, 3565-3568. (5) (a) Ashton, P. R.; Bissell, R. A.; Spencer, N.; Stoddart, J. F.; Tolley, M. S. Synlett 1992, 923-926. (b) Bissell, R. A.; Co´rdova, E.; Kaifer, A. E.; Stoddart, J. F. Nature 1994, 369, 133-137.

10.1021/la0361518 CCC: $27.50 © 2004 American Chemical Society Published on Web 06/03/2004

5810

Langmuir, Vol. 20, No. 14, 2004

onto surfaces11,12 and into solid-state devices,8,13 such as two-dimensional electronic circuits. The construction of such devices has relied, for the most part, upon introducing amphiphilicity14 into redox-switchable [2]rotaxanes so that they can be self-organized as Langmuir films and then transferred as molecular monolayers into the device configuration using the Langmuir-Blodgett (LB) technique.15 A molecular switch tunnel junction (MSTJ) has been fabricated8 by sandwiching self-organized monolayers of amphiphilic bistable [2]rotaxanes between polysilicon wires and top electrodes of titanium topped with aluminum. The devices have been shown to exhibit hysteretic current-voltage responses on account of marked differences in conductance between the ground (low) and metastable (high) states of a range of amphiphilic bistable [2]rotaxanes that have been employed in their fabrication. At this time, aside from obtaining evidence for the electromechanical mechanism that has been postulated to account for the operation of these mechanical molecular switches, there is a real need to elucidate the nature of the superstructures associated with their Langmuir monolayers and LB films. This paper examines how systematic changes to the hydrophilic stopper of amphiphilic bistable [2]rotaxanes affect Langmuir and LB film deposition. The quality of a (6) (a) Co´rdova, E.; Bissell, R. A.; Spencer, N.; Ashton, P. R.; Stoddart, J. F.; Kaifer, A. E. J. Org. Chem. 1993, 58, 6550-6552. (b) Devonport, W.; Blower, M. A.; Bryce, M. R.; Goldenberg, L. M. J. Org. Chem. 1997, 62, 885-887. (c) Ashton, P. R.; Ballardini, R.; Balzani, V.; Go´mez-Lo´pez, M.; Lawrence, S. E.; Martı´nez-Diaz, M. V.; Montalti, M.; Piersanti, A.; Prodi, L.; Stoddart, J. F.; Williams, D. J. J. Am. Chem. Soc. 1997, 119, 10641-10651. (d) Ashton, P. R.; Ballardini, R.; Balzani, V.; Baxter, I.; Credi, A.; Fyfe, M. C. T.; Gandolfi, M. T.; Go´mez-Lo´pez, M.; Martı´nezDiaz, M. V.; Piersanti, A.; Spencer, N.; Stoddart, J. F.; Venturi, M.; White, A. J. P.; Williams, D. J. J. Am. Chem. Soc. 1998, 120, 1193211942. (e) Elizarov, A. M.; Chiu, S. H.; Stoddart, J. F. J. Org. Chem. 2002, 67, 9175-9181. (f) Tseng, H. R.; Vignon, S. A.; Stoddart, J. F. Angew. Chem., Int. Ed. 2003, 42, 1491-1495. (7) (a) Raehm, L.; Kern, J. M.; Sauvage, J. P. Chem. Eur. J. 1999, 5, 3310-3317. (b) Ballardini, R.; Balzani, V.; Dehaen, W.; Dell’Erba, A. E.; Raymo, F. M.; Stoddart, J. F.; Venturi, M. Eur. J. Org. Chem. 2000, 591-602. (c) Collin, J.-P.; Kern, J.-M.; Raehm, L.; Sauvage, J.-P. In Molecular Switches; Feringa, B. L., Ed.; Wiley-VCH: Weinheim, 2001; pp 249-280. (d) Colasson, B. X.; Dietrich-Buchecker, C.; Jimenez-Molero, M. C.; Sauvage, J. P. J. Phys. Org. Chem. 2002, 15, 476-483. (8) Luo, Y.; Collier, C. P.; Jeppesen, J. O.; Nielsen, K. A.; DeIonno, E.; Ho, G.; Perkins, J.; Tseng, H. R.; Yamamoto, T.; Stoddart, J. F.; Heath, J. R. ChemPhysChem 2002, 3, 519-525. (9) (a) Ashton, P. R.; Ballardini, R.; Balzani, V.; Credi, A.; Dress, K. R.; Ishow, E.; Kleverlaan, C. J.; Kocian, O.; Preece, J. A.; Spencer, N.; Stoddart, J. F.; Venturi, M.; Wenger, S. Chem. Eur. J. 2000, 6, 35583574. (b) Brouwer, A. M.; Frochot, C.; Gatti, F. G.; Leigh, D. A.; Mottier, L.; Paolucci, F.; Roffia, S.; Wurpel, G. W. H. Science 2001, 291, 21242128. (c) Gatti, F. G.; Leo´n, S.; Wong, J. K. Y.; Bottari, G.; Altieri, A.; Morales, M. A. F.; Teat, S. J.; Frochot, C.; Leigh, D. A.; Brouwer, A. M.; Zerbetto, F. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 10-14. (10) (a) Stoddart, J. F. Chem. Aust. 1992, 59, 576-577 and 581. (b) Go´mez-Lo´pez, M.; Preece, J. A.; Stoddart, J. F. Nanotechnology 1996, 7, 183-192. (c) Balzani, V.; Go´mez-Lo´pez, M.; Stoddart, J. F. Acc. Chem. Res. 1998, 31, 405-414. (d) Balzani, V.; Credi, A.; Raymo, F. M.; Stoddart, J. F. Angew. Chem., Int. Ed. 2000, 39, 3349-3391. (e) Balzani, V.; Credi, A.; Venturi, M. Chem. Eur. J. 2002, 8, 5524-5532. (f) Balzani, V.; Credi, A.; Venturi, M. Molecular Devices and Machines-A Journey into the Nano World; Wiley-VCH: Weinheim, 2003. (11) Huang, T. J.; Flood, A.; Chu, C.-W.; Kang, S.; Guo, T.-F.; Yamamoto, T.; Tseng, H.-R.; Yu, B.-D.; Yang, Y.; Stoddart, J. F.; Ho, C.-M. IEEE-NANO 2003, 2, 698-701. (12) Tseng, H.-R.; Wu, D.; Fang, N. X.; Zhang, X.; Stoddart, J. F. ChemPhysChem 2004, 5, 111-116. (13) Yu, H.; Luo, Y.; Beverly, K.; Tseng, H.-R.; Stoddart, J. F.; Heath, J. R. Angew. Chem., Int. Ed. 2003, 42, 5706-5711. (14) (a) Jeppesen, J. O.; Perkins, J.; Becher, J.; Stoddart, J. F. Angew. Chem., Int. Ed. 2001, 40, 1216-1221. (b) Jeppesen, J. O.; Nielsen, K. A.; Perkins, J.; Vignon, S. A.; Di Fabio, A.; Ballardini, R.; Becher, J.; Stoddart, J. F. Chem. Eur. J. 2003, 9, 2982-3007. (c) Tseng, H. R.; Vignon, S. A.; Celestre, P. C.; Perkins, J.; Di Fabio, A.; Ballardini, R.; Gandolfi, M. T.; Venturi, M.; Balzani, V.; Stoddart, J. F. Chem. Eur. J. 2004, 10, 155-172. (d) Yamamoto, T.; Tseng, H. R.; Stoddart, J. F.; Balzani, V.; Credi, A.; Marchioni, F.; Venturi, M. Collect. Czech. Chem. Commun. 2003, 68, 1488-1514.

Lee et al.

LB film deposition is dependent on a variety of processing conditions, e.g., temperature, compression rate, and subphase salt concentration. Prior to carrying out a detailed investigation of LB film deposition of any given molecule, it is necessary to understand the equilibrium and dynamic properties of Langmuir films composed of these same molecules. Surface pressure versus mean molecular area (π-A) isotherms obtained near equilibrium conditions are a proven technique to obtain information on phase transitions, molecular conformations, and stabilities in the film as a function of the area per molecule.16 Measuring surface viscoelastic parameters of Langmuir films provides another means of relating surface properties to molecular structure. Previous work has shown that monitoring the surface viscoelastic properties versus lateral compression can help identify subtle changes in chemical structure.17-19 Rotaxanes are composed of interactive subunits, each with differing affinities for water and for each other. The dumbbell-shaped components contain a tetrathiafulvalene (TTF) unit, a 1,5-dioxynaphthalene (DNP) ring system, and two tetraarylmethane-based stoppers. Ethylene glycol (EG) chains provide the linkages between the interactive subunits and also serve as the hydrophilic tails on one of the stoppers. The ring-shaped component is the tetracationic cyclophane, cyclobis(paraquat-p-phenylene) (CBPQT4+), which resides preferentially on the TTF unit. The complexity of the compounds investigated can be illustrated by briefly reviewing the Langmuir monolayer behavior of the rotaxane subunits. TTF is a common donor molecule in charge-transfer complexes, and its derivatives have been extensively studied at the air-water interface, with the TTF unit acting as the hydrophilic group.16,20-22 The large aromatic surfaces of the DNP ring system and the tetraarylmethane stoppers are based on aromatic groups and should function as hydrophobic units. CBPQT4+ has a high ionic character, is soluble in water as the chloride salt, and is likely to be hydrated at the air-water interface.23,24 Oligo(EG), though a weak amphiphile, has been used as a hydrophilic anchor.25 Poly(ethylene glycol) has been shown to lie flat at the airwater interface until it is forced into the subphase by compression.18,26-30 The tetracationic cyclophane and the DNP ring system form a donor-acceptor pair, which may (15) Collier, C. P.; Jeppesen, J. O.; Luo, Y.; Perkins, J.; Wong, E. W.; Heath, J. R.; Stoddart, J. F. J. Am. Chem. Soc. 2001, 123, 1263212641. (16) Petty, M. C. Langmuir-Blodgett Films: An Introduction; Cambridge University Press: Cambridge, 1996. (17) Brooks, C. F.; Fuller, G. G.; Frank, C. W.; Robertson, C. R. Langmuir 1999, 15, 2450-2459. (18) Naumann, C. A.; Brooks, C. F.; Fuller, G. G.; Knoll, W.; Frank, C. W. Langmuir 1999, 15, 7752-7761. (19) Naumann, C. A.; Brooks, C. F.; Fuller, G. G.; Lehmann, T.; Ruhe, J.; Knoll, W.; Kuhn, P.; Nuyken, O.; Frank, C. W. Langmuir 2001, 17, 2801-2806. (20) Richard, J.; Vandevyver, M.; Barraud, A.; Morand, J. P.; Lapouyade, R.; Delhaes, P.; Jacquinot, J. F.; Roulliay, M. J. Chem. Soc., Chem. Commun. 1988, 754-756. (21) Dhindsa, A. S.; Pearson, C.; Bryce, M. R.; Petty, M. C. J. Phys. D: Appl. Phys. 1989, 22, 1586-1590. (22) Dhindsa, A. S.; Cooke, G.; Lerstrup, K.; Bechgaard, K.; Bryce, M. R.; Petty, M. C. Chem. Mater. 1992, 4, 720-723. (23) Ahuja, R. C.; Caruso, P. L.; Mobius, D.; Wildburg, G.; Ringsdorf, H.; Philp, D.; Preece, J. A.; Stoddart, J. F. Langmuir 1993, 9, 15341544. (24) Asakawa, M.; Higuchi, M.; Mattersteig, G.; Nakamura, T.; Pease, A. R.; Raymo, F. M.; Shimizu, T.; Stoddart, J. F. Adv. Mater. 2000, 12, 1099-1102. (25) Kampf, J. P.; Frank, C. W.; Malmstrom, E. E.; Hawker, C. J. Langmuir 1999, 15, 227-233. (26) Shuler, R. L.; Zisman, W. A. J. Phys. Chem. 1970, 74, 15231534. (27) Kawaguchi, M.; Komatsu, S.; Matsuzumi, M.; Takahashi, A. J. Colloid Interface Sci. 1984, 102, 356-360.

Amphiphilic Bistable Rotaxanes

Langmuir, Vol. 20, No. 14, 2004 5811

Chart 1. Molecular Formulas of the Amphiphilic Bistable [2]Rotaxanes and Their Dumbbell Counterparts

stabilize folded conformations.14b-d Thus, since a variety of hydrophilic and hydrophobic subunits are alternated along the rotaxane, complex amphiphilic behavior and molecular conformations are expected. Finally, after understanding the effects of molecular structure on film balance behavior and surface viscoelasticity, several processing parameters are investigated using a subset of the amphiphilic bistable [2]rotaxanes. We examine the effects of varying compression rate and subphase temperature, given that π-A isotherms of macromolecular amphiphiles often show strong dependence on these parameters.25,31-34 The stability of the [2]rotaxane monolayers with respect to time, measured using creep and compression-expansion experiments, is informative regarding monolayer phases and is also crucial for successful LB film preparation. The effects of using different spreading solvents and subphase salt concentrations are also examined. The design of these materials represented an attempt14d to introduce, with respect to earlier generations14a-c of such molecular switches, increased redox stability in relation to the chemical and (28) Kuzmenka, D. J.; Granick, S. Macromolecules 1988, 21, 779782. (29) Sauer, B. B.; Yu, H.; Yazdanian, M.; Zografi, G.; Kim, M. W. Macromolecules 1989, 22, 2332-2337. (30) Xu, Z.; Holland, N. B.; Marchant, R. E. Langmuir 2001, 23, 377-383. (31) Adams, J.; Buske, A.; Duran, R. S. Macromolecules 1993, 24, 2871-2877. (32) Cheng, Q.; Stevens, R. C. Chem. Phys. Lipids 1997, 30, 41-53. (33) Kampf, J. P.; Frank, C. W.; Malmstrom, E. E.; Hawker, C. J. Science 1999, 283, 1730-1733. (34) Gourier, C.; Knobler, C. M.; Daillant, J.; Chatenay, D. Langmuir 2002, 26, 9434-9440.

electrochemical oxidation employed to induce the cyclophane to move away from the TTF unit to the DNP ring system on the dumbbell component. Experimental Section Design and Materials. The investigation was limited to a manageable number of compounds that were expected to exhibit differences in their Langmuir and LB film properties as a result of specific differences in the molecule design. Chart 1 shows the molecular formulas of the [2]rotaxanes and their corresponding dumbbell precursors. Figure 1 shows a three-dimensional representation of MeO-TEGD. Both the rotaxanes and their dumbbell precursors differ in the constitution35 of the hydrophilic

Figure 1. Three-dimensional representation of MeO-TEGD (Chem3D Ultra 6.0). Darker segments represent hydrophilic subunits, and light segments represent hydrophobic subunits.

5812

Langmuir, Vol. 20, No. 14, 2004

Lee et al.

Table 1. Molecular Weight, Estimated Transition Temperature Ts, Average LB Film Thickness at πtransfer ) 30 mN/m from Ellipsometry, and Isothermal Compressibility C at π ) 10 mN/m compound MeO-DEGR4+ MeO-DEGD MeO-TEGR4+ MeO-TEGD HO-DEGR4+ HO-DEGD HO-TEGR4+ HO-TEGD iso-MeO-DEGR4+ iso-MeO-DEGD

molecular Ts thickness weight [°C] [Å] 2959 1858 3223 2123 2903 1802 3167 2067 2959 1858

30 15 Gs′′, with both moduli independent of frequency (exponent ≈ 0). Figure 5 shows that the HO-DEGR4+ monolayer approaches the elasticdominated regime as the surface pressure increases but remains viscoelastic at the surface pressures investigated.

5820

Langmuir, Vol. 20, No. 14, 2004

Lee et al.

Figure 4. Dynamic surface storage modulus Gs′ and loss modulus Gs′′ versus area per molecule for (a) MeO-DEGR4+, (b) MeODEGD, (c) MeO-TEGR4+, (d) MeO-TEGD, (e) HO-DEGR4+, (f) HO-DEGD, (g) HO-TEGR4+, (h) HO-TEGD, (i) iso-MeODEGR4+, and (j) iso-MeO-DEGD.

This implies that there is no formation of physical networks or structure in the film that might store elastic energy in the regions studied. The monolayer at pressures greater than 40 mN/m becomes too viscous to be investigated. This demonstrates the dramatic increase in surface viscosity in these molecules. Looking at the threshold mean molecular area at which Gs′′ equals 0.01 mN/m, (Figure 6a), we can see that

molecules with DEG tails reach that value at higher A than molecules with TEG tails. This observed difference implies that although molecules with TEG are larger and have an appreciable pressure at higher A, molecules with DEG can have stronger intermolecular interactions and greater resistance to shear deformation than molecules with TEG tails. The OH end group leads to higher surface loss moduli over the OMe end group as a result of stronger

Amphiphilic Bistable Rotaxanes

Langmuir, Vol. 20, No. 14, 2004 5821

Figure 5. Gs′ and Gs′′ versus ω for HO-DEGR4+ at (a) π ) 10 mN/m, (b) π ) 15 mN/m, and (c) π ) 39 mN/m.

interactions with the water subphase. Molecules with cyclophanes, i.e., rotaxanes, also have greater viscoelastic interactions at higher A, which is expected given that cyclophanes add size and hydrophilic interactions to the molecule. Figure 6b,c shows Gs′ and Gs′′ at π ) 20 mN/m for all the molecules; it is clear that HO-DEGD is the most viscoelastic molecule. The trends in surface viscoelasticity observed in Figure 6a are repeated: OH end groups and DEG tails lead to greater surface moduli than OMe end groups and TEG tails. The trends related to the cyclophane are more complicated; for molecules with DEG tails, those that have no cyclophane, the dumbbells, are more viscoelastic than their corresponding [2]rotaxanes. For molecules with TEG tails, the viscoelasticity is similar in magnitude. The observations can be rationalized from the fact that dumbbells, without the cyclophane, must be compressed to smaller A for the same increase in pressure; hence, although the pressure is low, the density of molecules can be high, which leads to high viscoelastic interactions. Returning to the relaxation phenomena discussed above, the extent of relaxation can be seen by plotting tan(δ) ) Gs′′/Gs′ versus area per molecule for each compound (not shown). Maxima in tan(δ) correspond in location to the dips in Gs′ observed in Figure 4. The magnitude of tan(δ) is greater for molecules without a cyclophane, suggesting that the cyclophane adds stiffness to the molecules. Also, molecules with OH end groups have stronger interactions with the subphase than molecules with OMe end groups, so they appear to add more stiffness to the monolayer. MeO-DEGD has the largest relaxation, in part because it has OMe end groups and no cyclophane. After the peak in tan(δ), Gs′ and Gs′′ increase monotonically for most of the compounds. HO-DEGR4+ and HO-DEGD (Figure 4e,f) go through a 1000-fold increase

in Gs′′, and HO-TEGR4+ has a 10 000-fold increase in Gs′′. We were unable to probe the LC region for most of the compounds as the surface viscosity increased beyond the detection limit of the interfacial stress rheometer. In other cases, the measurement was stopped due to monolayer collapse (MeO-DEGR4+ and MeO-DEGD in Figure 4c,d) or instability (MeO-DEGR4+ in Figure 4a, iso-MeODEGR4+ in Figure 4i). Compression Rate and Temperature Dependence. To investigate further for evidence of phases in the Langmuir films, the dependence of the π-A isotherms of MeO-DEGR4+, MeO-DEGD, MeO-TEGR4+, HODEGR4+, and HO-DEGD on the compression rate has been examined. Figure 7a displays isotherms for MeODEGR4+ at two different compression rates. The isotherms diverge at the transition pressure: measured at 10 mm2/ s, the projected area of the LC region is 70 Å2; measured at 100 mm2/s, the projected area of the LC region is 85 Å2. At faster compression speeds, the monolayer cannot relax completely in the LC region and the transition is suppressed. Thus, the effects of compression rate on features in the isotherm can be used to investigate rearrangements in the monolayer. The effects of compression rate can also be used to investigate conformational rearrangements that are less than apparent in the isotherms. For MeO-DEGD (Figure 7b), the isotherms obtained at 10 and 100 mm2/s diverge at the shoulder at 16 mN/m, implying that a conformational rearrangement occurs at a low surface pressure. The isotherm for HO-DEGR4+ (Figure 7d) taken at 25 mm2/s shows only a slight kink at moderate pressure, but when compared to the isotherm taken at 100 mm2/s, a conformational rearrangement can be pinpointed from the divergence in the two isotherms. (This conformational rearrangement is analogous to the transition at 37 mN/m seen in the MeO-DEGR4+ isotherm.) The compression

5822

Langmuir, Vol. 20, No. 14, 2004

Lee et al.

Figure 6. (a) Threshold mean molecular area at which Gs′′ ) 0.01 mN/m, (b) Gs′ at π ) 20 mN/m, and (c) G′′ at π ) 20 mN/m for the [2]rotaxanes.

rate effects on the isotherms for HO-DEGD (Figure 7e) reveal rearrangements as the monolayer transitions from the gaseous region to the LC region. For MeO-TEGR4+ (Figure 7c) the two isotherms, taken at different compression rates, approach each other asymptotically at high mean molecular area, suggesting that molecular conformational rearrangements are occurring for MeO-TEGR4+ at large mean molecular area. The dependence of isotherm behavior on the subphase temperature for MeO-DEGR4+, MeO-DEGD, MeOTEGR4+, HO-DEGR4+, HO-DEGD, and HO-TEGD was also examined (Figure 8). An increase in temperature gives the same physical manifestation as a decrease in compression rate; with more thermal energy, relaxations in the monolayer are accelerated. At monolayer collapse, as there is an energy barrier before the molecules can leave the interface and form a multilayer, an increase in temperature facilitates domain formation, depresses collapse pressures, and broadens the transition region. Likewise, lowering the subphase temperature can increase stability with respect to π. The effect of temperature on the isotherms of MeO-DEGR4+ (Figure 8a) demonstrates this phenomenon. In Figure 8b, the collapse at 32 mN/m in the isotherm for MeO-DEGD is completely suppressed at 10 °C and the LC region is observed. The effect of temperature on the isotherms of MeODEGR4+ and MeO-DEGD leads us to define a transition temperature Ts below which the LC region is observed and collapse is suppressed and above which collapse is observed and the LC region is nonexistent. Estimated values for Ts are given in Table 1 for qualitative comparison

between the compounds. Ts is greater for MeO-DEGR4+ than for MeO-DEGD, which is consistent with our observation that MeO-DEGR4+ is more stable than MeODEGD with respect to monolayer collapse. In the case of MeO-TEGR4+ (Figure 8c), Ts is below the investigated temperature range, while Ts for HO-DEGR4+ (Figure 8d) is above the investigated temperature range. Ts is not estimated for HO-DEGD (Figure 8e) since a direct LEto-LC transition is not observed. In summary, the effects of the cyclophane are to increase Ts during the phase transition at moderate pressure and the end group OH increases Ts over the end group OMe. It also appears that DEG tails increase Ts over TEG tails, implying that DEG tails produce more stable monolayers. Film Stability. The π-A isotherms are measured using constant barrier compression and therefore cannot be regarded as equilibrium values;44 however, inferences about the LE and LC regions can be made. We investigate the stability of Langmuir films by measuring the magnitude of supramolecular reorganizationsand conformational/co-conformational readjustment in the moleculesswhen π is held constant. Figure 9 shows the results of creep tests for MeO-DEGR4+, MeO-DEGD, HO-DEGR4+, HO-DEGD, and HO-TEGR4+. MeODEGR4+ (Figure 9a) is reasonably stable for surface pressures below the transition, with only 10% area loss after 1 h at 30 mN/m, but is unstable above it, with 35% area loss after 1 h. It is likely that the area loss arises (44) Roberts, G. Langmuir-Blodgett Films; Plenum Press: New York, 1990.

Amphiphilic Bistable Rotaxanes

Langmuir, Vol. 20, No. 14, 2004 5823

Figure 7. π-A isotherms recorded at 20 °C and at two different compression rates for (a) MeO-DEGR4+, (b) MeO-DEGD, (c) HO-DEGR4+, (d) HO-DEGD, and (e) MeO-TEGR4+.

from molecular rearrangements associated with the phase transition. Likewise, MeO-DEGD, HO-DEGR4+, and HO-DEGD are stable at 30 mN/m, with less than 10% area loss. The HO-DEGR4+ monolayer (Figure 9c) decreases in stability as the surface pressure is increased beyond the transition pressure, but the HO-DEGD monolayer (Figure 9d) remains relatively stable at high pressure. The HO-TEGR4+ monolayer (Figure 9e) is more stable than the HO-DEGR4+ monolayer at 40 mN/m but has similar stability at 50 mN/m. Also, in comparing the creep of MeO-DEGR4+ with HO-DEGR4+ at 40 mN/m, it is clear that if a plateau or even a shoulder is apparent in the isotherm (compare parts a and c of Figure 2), then this will lead to a rapid loss of area due to some form of phase coexistence. Figure 9f shows the creep of a monolayer of HO-DEGR4+ as a function of the compression speed prior to the target pressure. Area loss is reduced when the compression rate is decreased. This observation verifies that isotherms obtained at slower compression rates are better approximations of an equilibrium isotherm. Compression-expansion experiments give an indication of the reversibility of the isotherm of a monolayer. At 30

mN/m (Figure 10a), the MeO-DEGR4+ isotherm is reversible; at 50 mN/m (Figure 10b), the presence of hysteresis indicates that isotherm behavior in the LC region is not reversible. The irreversibility in the isotherm originates from passing through a transition region from the stable LE phase to the less stable LC phase. The compression-expansion cycles for MeO-DEGD and HODEGR4+ at 30 and 50 mN/m (Figure 10c,f) demonstrate the same behavior: reversibility at low pressure, irreversibility at high pressure. Remarkably, the HODEGD isotherm is reversible at both 30 and 50 mN/m (Figure 10g,h), suggesting that the monolayer has no transition from LE to LC phase between 30 and 50 mN/m. Figure 10i demonstrates that compression-expansion of HO-TEGR4+ up to monolayer collapse and back again yields substantial hysteresis. Ellipsometry and Surface Force Microscopy. LB films of all compounds were prepared on double-polished silicon substrates. Table 1 summarizes the result of ellipsometry measurements for LB films transferred at 30 mN/m. For comparison, the thicknesses (21-39 Å) are much smaller than the contour lengths of the rotaxane molecules, which are in the range of 65-75 Å (Table 2).

5824

Langmuir, Vol. 20, No. 14, 2004

Lee et al.

Figure 8. π-A isotherms recorded at 10, 20, and 30 °C for (a) MeO-DEGR4+, (b) MeO-DEGD, (c) MeO-TEGR4+, and (d) HODEGR4+, and (e) HO-DEGD, and (f) isotherms recorded at 10 and 20 °C for HO-TEGD.

The molecules most likely display some degree of conformational folding and/or lie at an angle other than perpendicular to the substrate. Although the effects of salt concentration and spreading solvent on LB films of MeO-DEGR4+, MeO-DEGD, MeO-TEGR4+, and MeOTEGD were investigated, LB film thickness is mainly dependent on the transfer pressure. The increased thickness of the MeO-DEGR4+ LB film at 50 mN/m is probably a consequence of multilayer domains transferred onto the substrate and is further evidence that the LC region for this molecule is unstable. It appears that the dumbbellonly counterparts yield thicker films; they generally are denser in surface concentration than the [2]rotaxanes at transfer pressures. Scanning force microscopy (SFM) measurements of the LB films were performed in tapping mode. The images reveal relatively featureless and smooth morphologies for the films of HO-DEGR4+ transferred at 30 mN/m. All the [2]rotaxane molecules were transferred at 30 mN/m, and all showed smooth and featureless morphologies, implying that the molecular design is not crucial to overall film quality, although it may affect the orientation of the molecules. This observation could indicate a smooth

monolayer with very few multilayer domains. However, large domains (5000 nm2) reminiscent of dewetting phenomena are observed for films transferred at 50 mN/ m. The prescence of these domains are consistent with our previous observation that the Langmuir films at higher pressures are unstable and do not yield smooth LB films. We observed that subphase ion concentration and spreading solvent used were not important factors in determining film quality. Discussion Molecular Design Issues. The complexity of the molecules studied in this investigation is evident from inspection (Chart 1) of the constitutions of both the rotaxanes and their dumbbell counterparts. The central portions of all the molecules contain (interactive) components that are both hydrophobic and hydrophilic. Thus, it is not surprising that complex Langmuir isotherms arise as a consequence of the molecular structure and associated conformational equilibria. (As an example of the complexity added by the TTF units and DNP ring systems, switching their internal order as a means to increase segregation of the hydrophilic components from the

Amphiphilic Bistable Rotaxanes

Langmuir, Vol. 20, No. 14, 2004 5825

Figure 9. Dilational creep measurements recorded at 20 °C for (a) MeO-DEGR4+ at π ) 30, 40, and 50 mN/m, (b) MeO-DEGD at π ) 30 mN/m, (c) HO-DEGR4+ at π ) 30, 40, and 50 mN/m, (d) HO-DEGD at π ) 30, 40, and 50 mN/m, (e) HO-TEGR4+ at π ) 40 and 50 mN/m, and (f) HO-DEGR4+ versus compression rate (π ) 30 mN/m).

hydrophobic components resulted in only a slight increase in the collapse pressure and the hydrophilic anchoring of iso-MeO-DEGD.) Nevertheless, the π-A isotherms for all the compounds have similar features that can be interpreted in the context of their molecular structures. Most isotherms show a liquid-expanded region followed by a transition at moderate surface pressure. The midpressure transition can be characterized by the pressure πtransition, which represents the stability of the LE region at the air-water interface with respect to a molecule’s amphiphilic balance. At πtransition, the monolayer either collapses or continues to a liquid-condensed region. This phase extends until the collapse point of the LC region. Dilational creep measurements confirmed that the LC region was generally unstable for the molecules investigated. We examine the properties of the LE region for the rotaxanes using both π-A isotherm and surface rheology data. In general, greater hydrophilic strength leads to a higher stability in surface pressure. The following trends are apparent: Cyclophane. From the π-A isotherm data, at 20 °C all rotaxanes bearing the tetracationic cyclophane have

collapse pressures near 70 mN/m, except MeO-TEGR4+. The collapse pressures of the dumbbell counterparts are lower and can be ordered: HO-DEGD > HO-TEGD > MeO-TEGD > iso-MeO-DEGD > MeO-DEGD. The cyclophane is the most important factor in determining the collapse pressure, so it is likely that it is the molecular design feature that adds the most hydrophilicity to the molecules. It was also observed that the cyclophane contributes area per molecule. The surface rheology data suggests that at high A the cyclophane is important in interacting not only with the subphase through better hydrophilicity but also neighboring rotaxane molecules through increased dispersive interactions. The cyclophane also adds stiffness to the monolayer. At a given moderate pressure, the dumbbell-only counterparts may exhibit a higher viscoelasticity; this observation can be explained by increased density at the same pressure. End Group. The OH end group is more hydrophilic than the OMe end group. The OMe end group also tends to give larger footprints in the gaseous and LE regions. This observation is consistent with the OMe end group adding one methyl group in length on each tail over a OH end group. It is unclear whether the length of one bond is the

5826

Langmuir, Vol. 20, No. 14, 2004

Lee et al.

Figure 10. Compression-expansion cycles recorded at 20 °C for (a) MeO-DEGR4+ to π ) 30 mN/m, (b) MeO-DEGR4+ to π ) 50 mN/m, (c) MeO-DEGD to π ) 30 mN/m, (d) MeO-DEGD to π ) 35 mN/m, (e) HO-DEGR4+ to π ) 30 mN/m, (f) HO-DEGR4+ to π ) 50 mN/m, (g) HO-DEGD to π ) 30 mN/m, (h) HO-DEGD to π ) 50 mN/m, and (i) HO-TEGR4+ to π ) 65 mN/m. Table 2. Dimensions of compound subunitsa subunit

contour L [Å]

MeO-DEG tail MeO-TEG tail HO-DEG tail HO-TEG tail Hydrophilic stopper - DNP DNP - TTF TTF - Hydrophobic stopper Cyclophane width Cyclophane length

14.4 21.2 13.1 20.0 16.8 15.7 21.6 6.9 10.1

Contour lengths of all-trans configuration. a From Chem3D Ultra 6.0.

primary reason for the footprint differential, as the likely conformations of the tails are unknown. Surface rheology data support the conclusion that OH end groups add viscoelasticity and hydrophilicity. Tail Length. Molecules with TEG tails have increased size and display π-A isotherms that are independent of the cyclophane’s presence in the LE region, while molecules with DEG tails have differing isotherms depending on the presence of the cyclophane. TEG tails give larger footprints at high A, and TEG tails dominate the LE

region’s behavior over the presence or absence of a cyclophane. On the other hand, for DEG tails, the cyclophane affects the isotherm morphology in the LE region. When the tail is DEG, the influence of the cyclophane dominates in the LE region, but when the tail is TEG, it is the tail that dominates the LE region behavior. This interpretation implies that the tail length and the cyclophane can be competing factors in the final molecular conformation in the LE region. One possible model is that DEG tails are short and do not interact with neighboring molecules until A decreases. The addition of the cyclophane adds significant bulk and area that is comparable to the DEG tail size. TEG tails are longer and interact with neighboring molecules at higher A; the cyclophane is less significant sterically in comparison with the TEG tails. The surface rheology results show, on the other hand, that DEG tails lead to greater viscoelasticity than TEG tails. One possible explanation is that molecules with TEG tails, while larger in size, are less strongly associated with the interface. This situation could arise through either the presence of weaker hydrophilic interactions with the water or a shallower conformation at the interface. Molecules with DEG tails may be more upright at the

Amphiphilic Bistable Rotaxanes

Langmuir, Vol. 20, No. 14, 2004 5827

Figure 11. Idealized representations of the molecular conformations of the rotaxane iso-MeO DEGR4+ based on the two models proposed to account for the presence of a liquid-expanded phase that undergoes a phase transition to a liquid-condensed phase. (a) Model 1 relies primarily on hydrophilic anchoring of the cyclophane to the subphase, whereas (b) model 2 relies on the cyclophane’s steric bulk.

interface, leading to smaller areas but stronger interactions with the subphase. Internal Ordering. Altering the internal order of the TTF unit and DNP ring system slightly increases the hydrophilicity of MeO-DEGR4+, as shown by the increased collapse pressure. Also, iso-MeO-DEGD is shifted to lower A, which could be explained by a more upright or packed conformation for the molecule due to better alignment of hydrophilic groups with the subphase. Another tool for examining the state of the rotaxane monolayers in the LE region is the isothermal compressibility. The compressibility, C, is given by the relationship

C)-

1 ∂A A ∂π

( )

T

where A is the mean molecular area and the derivative is the instantaneous slope of the isotherm at constant temperature. Table 1 shows the compressibility of the molecules at π ) 10 mN/m. The compressibility at 10 mN/m increases when the cyclophane is absent and when TEG tails are present; this implies that the cyclophane adds stiffness to the monolayer and that at low π molecules with TEG tails are more expanded and compressible than molecules with DEG tails. This interpretation for the differences observed in the compressibility between the tail lengths in [2]rotaxanes and their dumbbell counterparts could support the model proposed to explain the higher viscoelasticity for DEG tails at large A. The film stability, compression rate, and temperature dependence data are internally consistent with the surface pressure and surface rheology results. The LC region was shown to be generally unstable. Molecules with OH end groups and cyclophanes tend to have LC regions or higher

transition pressures. From examining the effects of temperature and the role of Ts on the LE-to-LC transition, we determined that molecules with the cyclophane and OH end groups have greater stability and that molecules with DEG tails are more likely to give rise to a transition to the LC region. The conformational equilibrium of the molecules is unknown; however, it appears that πtransition and the overall hydrophilicity is increased by the presence of cyclophanes and OH end groups. The role of the tail length is less clear; DEG tails seem to provide greater viscoelasticity at high A. The best candidates for stability in the LE region for [2]rotaxanes should be HO-DEGR4+ and HO-TEGR4+. Transferred films had similar thicknesses and featureless morphologies when examined using SFM. The isotherm behavior of MeO-TEGR4+ and MeOTEGD appears to be anomalous in the context of the other compounds. The presence of a highly hydrophilic tetracationic cyclophane should increase the collapse pressure in MeO-TEGR4+ over MeO-TEGD. The behavior of these two compounds is not consistent with that displayed by the remaining series of compounds. Further investigation would be necessary in order to determine the nature of the anomalous isotherms. Molecular Conformation. From the analyses on the influence of the cyclophane and the tail’s length on the Langmuir isotherms of rotaxanes and their dumbbell counterparts, two properties have been identified that may control the molecular conformations within the films. The marked hydrophilicity observed for the rotaxanes may reflect the cyclophane’s close association with the subphase leading to a folded or tilted conformation (for example, see Figure 11a). In this way, the cyclophane competes effectively with the shorter DEG tails until the transition

5828

Langmuir, Vol. 20, No. 14, 2004

region, at which point the rotaxanes unfold. Such a transition is not observed in the case with the longer TEG tails, wherein it is the tail-to-tail intermolecular interactions that dominate, and so, the molecule’s conformation is less affected by the cyclophane. Alternatively, the cyclophane’s steric bulk may prevent (Figure 11b) the shorter DEG tails from forming cohesive intermolecular interactions until a sufficient pressure is applied, πtransition, at which point their proximity finally pulls the rotaxanes together into the LC phase. However, for the rotaxane with the TEG tails, the tail-to-tail interactions are present from the onset of the LE phase; hence, no transition region is observed. These two models imply two different possible conformations in the LE phase as a consequence of two different physical phenomena with two different mechanical processes occurring at the LE-to-LC phase transition. Nevertheless, it is clear that the cyclophane competes effectively with the shorter tail to direct the film’s behavior in the LE phase, whereas the cyclophane has a negligible influence on the rotaxanes with the longer tail. Molecular Conformations in Langmuir Monolayers of the Bistable Amphiphilic [2]Rotaxane MeO-DEGR4+. To date, LB monolayers of the [2]rotaxane MeO-DEGR4+ have been utilized8 in MSTJ devices. In the fabrication of these devices, monolayers were transferred at surface pressures (π ) 30 mN/m) below the transition region. The data collected and analyses discussed in this work allow us to propose an averaged molecular conformation of the bistable [2]rotaxane MeO-DEGR4+ used in MSTJ devices. At 30 mN/m, the film thickness is 28 Å and the area per molecule is 90 Å2. These two observations and the presence of the short amphiphilic tail suggest that the rotaxane is not fully elongated and that it conforms to model 1. Such an assignment suggests that at pressures which are used for LB deposition in MSTJ devices, the rotaxane’s backbone lies at an acute angle with respect to the air-water interface and the cyclophane may lie at or close to the air-water interface. This averaged molecular conformation for the liquid-expanded phase proposed in model 1

Lee et al.

has important consequences for understanding both (i) the mechanism of switching to the ON and OFF states at (2 V, respectively, and (ii) the possible modes of electrical transport measured at +0.1 V across the device once the ON and OFF states have been formed. Conclusions The tetracationic cyclophane present in the rotaxanes imposes marked stability and increased intermolecular interactions and surface viscoelasticity upon the films, as do the OH end groups on the molecule’s hydrophilic tails. The additional intermolecular interactions associated with the cyclophane are believed to compete with traditional near-equilibrium thermodynamic driving forces that direct the formation of self-organized Langmuir monolayers. Systematic changes in the constitutions of the hydrophilic stoppers in amphiphilic bistable [2]rotaxanes do not affect the deposition at π ) 30 mN/m of smooth LB films. Highquality Langmuir and LB films have been obtained from monolayers prepared near equilibrium using slightly elevated temperatures or slower compression rates (10 mm/s) and at a dipping pressure (30 mN/m) below the phase transition. Isotherm and dimensional analyses of the Langmuir and LB films suggest that the rotaxanes exist in a conformation with their backbone at an acute angle to the air or Si surfaces, respectively. Such molecular superstructures have important implications for molecular electronics devices based on bistable amphiphilic [2]rotaxanes. Acknowledgment. I.C.L. and C.W.F. acknowledge support from Hewlett-Packard through the Stanford Center for Integrated Systems. This research was supported at the University of California, Los Angeles, by the Defense Advanced Research Projects Agency (DARPA), and the National Science Foundation. T.Y. thanks Showa Denko K K for a Visiting Scholar Fellowship. J.O.J. thanks the Carlsbergfondet and the Familien Hede Nielsens Fond for financial support. LA0361518