Ground-state equilibrium thermodynamics and switching kinetics of bistable [2]rotaxanes switched in solution, polymer gels, and molecular electronic devices

FULL PAPER

DOI: 10.1002/chem.200500934

Ground-State Equilibrium Thermodynamics and Switching Kinetics of

Bistable [2]Rotaxanes Switched in Solution, Polymer Gels, and Molecular

Electronic Devices

Jang Wook Choi,^[a]Amar H. Flood,^{[b, d]}David W. Steuerman,^{[a, e]}Sune Nygaard,^[c]

Adam B. Braunschweig,^[b]Nicolle N. P. Moonen,^[b]Bo W. Laursen,^{[b, f]}Yi Luo,^[a]

Erica DeIonno,^[a]Andrea J. Peters,^[b]Jan O. Jeppesen,*^[c]Ke Xu,^[a]

J. Fraser Stoddart,*^[b]and James R. Heath*^[a]

Abstract: We report on the kinetics

and ground-state thermodynamics asso-

ciated with electrochemically driven

molecular mechanical switching of

three bistable [2]rotaxanes in acetoni-

trile solution, polymer electrolyte gels,

and molecular-switch tunnel junctions

(MSTJs). For all rotaxanes a p-elec-

and this equilibrium is relatively tem-

perature independent. In the third ro-

taxane (RBPTTF⁴⁺), the TTF unit is

replaced by a p-extended analogue (a

bispyrrolotetrathiafulvalene (BPTTF)

unit), and the CBPQT⁴⁺ring encircles

almost equally both recognition sites at

equilibrium. This equilibrium exhibits

strong temperature dependence. These

thermodynamic differences were ra-

tionalized by reference to binding con-

stants obtained by isothermal titration

calorimetry for the complexation of

model guests by the CBPQT⁴⁺host in

acetonitrile. For all bistable rotaxanes,

oxidation of the TTF (BPTTF) unit is

accompanied by movement of the

CBPQT⁴⁺ring to the DNP site. Reduc-

tion back to TTF⁰(BPTTF⁰) is fol-

lowed by relaxation to the equilibrium

distribution of translational isomers.

The relaxation kinetics are strongly en-

vironmentally dependent, yet consis-

tent with a single electromechanical-

switching mechanism in acetonitrile,

polymer electrolyte gels, and MSTJs.

The ground-state equilibrium proper-

ties of all three bistable [2]rotaxanes

were reflective of molecular structure

in all environments. These results pro-

vide direct evidence for the control by

molecular structure of the electronic

properties exhibited by the MSTJs.

tron-deficient

cyclobis(paraquat-p-

phenylene) (CBPQT⁴⁺) ring compo-

nent encircles one of two recognition

sites within a dumbbell component.

Two rotaxanes (RATTF⁴⁺and RTTF⁴⁺

)

contain tetrathiafulvalene (TTF) and

1,5-dioxynaphthalene (DNP) recogni-

tion units, but different hydrophilic

stoppers. For these rotaxanes, the

CBPQT⁴⁺ring encircles predominantly

(>90%) the TTF unit at equilibrium,

Keywords: calorimetry · kinetics ·

molecular devices

thermodynamics

· rotaxanes ·

[c] S. Nygaard, Prof. J. O. Jeppesen

Department of Chemistry

[a] J. W. Choi, Dr. D. W. Steuerman, Dr. Y. Luo, E. DeIonno, K. Xu,

Prof. J. R. Heath

Odense University (University of Southern Denmark)

Campusvej 55, 5230, Odense M(Denmark)

Fax : (+45)66-15-87-80

Division of Chemistry and Chemical Engineering (127–72)

California Institute of Technology, 1200 E. California Blvd.

Pasadena, CA, 91125 (USA)

E-mail: joj@chem.sdu.dk

Fax : (+1)626-395-2355

E-mail: heath@caltech.edu

[d] Dr. A. H. Flood

Current Address: Department of Chemistry, Indiana University

800 East Kirkwood Avenue, Bloomington, IN 47405 (USA)

[b] Dr. A. H. Flood, A. B. Braunschweig, Dr. N. N. P. Moonen,

Dr. B. W. Laursen, Dr. A. J. Peters, Prof. J. F. Stoddart

California NanoSystems Institute and

[e] Dr. D. W. Steuerman

Department of Chemistry and Biochemistry

University of California, Los Angeles, 405 Hilgard Avenue

Los Angeles, CA, 90095-1569 (USA)

Fax : (+1)310-206-1843

E-mail: stoddart@chem.ucla.edu

Current Address: Department of Physics

University of California, Santa Barbara CA 93106 (USA)

[f] Dr. B. W. Laursen

Current Address: Nano-Science Center

University of Copenhagen, Universitetsparken 5

2100, København Ø (Denmark)

Chem. Eur. J. 2006, 12, 261 – 279

ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

261

Introduction

One of the goals^[1–3]of the field of molecular electronics is

to be able to control the properties of molecular-based

solid-state devices through chemical design and synthesis.

Such control has been demonstrated^[4–9]for passive devices,

the simplest of which are molecular tunnel junction resistors

consisting of a molecular monolayer, often a functionalized

alkane, sandwiched between two conductors. Several groups

have shown that the tunnel current varies exponentially

with chain length,^[6]although they have also found that

atomistic details,^[7–8]such as the packing of the chains, the

molecular alignment within the monolayer, and the nature

of the electrodes,^[9]are all important.

Figure 1. a) Structural formulas of the two translational isomers of the bi-

stable rotaxane RATTF⁴⁺corresponding to the ground-state co-confor-

mation (GSCC) and the metastable-state co-conformation (MSCC).

b) Potential-energy surface for the bistable RATTF⁴⁺, in which the

energy wells correspond to the GSCC and MSCC. The free-energy differ-

ence DG, between the wells and the free-energy barrier to relaxation,

DG^°, from the MSCC to the GSCC are defined against a normal coordi-

nate, Q, representing translation of the ring along the dumbbell compo-

nent of the [2]rotaxane.

Molecular rectifiers, typically represented by an electron-

donor–bridge–acceptor molecule extended between two

electrodes,^[10]represent a more sophisticated passive device.

Demonstrations of molecular control over current rectifica-

tion have required a substantial effort by a number of

groups,^[10–18]and have only been achieved within the past

few years. Details such as the nature of the molecule/elec-

trode interface, the donor and acceptor molecular orbital

energies, and the structure of the molecule within the

device—that is, the extension of the donor–bridge–acceptor

between the two electrodes—are all important, since rectifi-

cation can arise from many areas within a junction.^[10–18]

Active molecular electronic^[19]devices (switches) repre-

sent a significant jump in terms of molecular complexity. We

have reported previously^[20–22]on the use of electrochemical-

ly switchable, donor–acceptor, bistable [2]catenane and

[2]rotaxane molecular-switch tunnel junctions (MSTJs). As

in the case of the molecular tunnel junction resistors and

rectifiers, MSTJs also represent a highly coupled molecule/

electrode system.^[9,23–24]However, for the bistable [2]cat-

enane and [2]rotaxane switches, there are a number of ex-

perimental parameters that can be measured to correlate

molecular structure and solution-phase switching behavior

with molecular electronic-device-switching properties. These

parameters include colorimetric changes,^[25]shifts in electro-

chemical potentials,^[26–27]and temperature-dependent kinet-

ics^[25–27]for the cycling of the switch.

As an example, consider the redox-switchable [2]rotaxane

RATTF⁴⁺illustrated in Figure 1a. This bistable [2]rotaxane

is composed of p-electron-accepting cyclobis(paraquat-p-

phenylene) (CBPQT⁴⁺) ring (blue) that encircles either a

tetrathiafulvalene (TTF) unit (green) or a 1,5-dioxynaptha-

lene (DNP) unit (red), both p-electron-donating systems.

This mechanically interlocked molecular compound and

other closely related bistable rotaxanes,^[28]as well as rotax-

anes constructed from different donor–acceptor units^[29]or

from hydrogen-bonded systems^[30]and transition-metal tem-

plates,^[31]have been investigated in depth previously. Under

ambient conditions in acetonitrile, the CBPQT⁴⁺ring in

RATTF⁴⁺encircles the TTF unit preferentially (>90%)

with respect to the DNP unit. This equilibrium is described

by the DG₂₉₈change shown in Figure 1b, in which DG=

+1.6 kcalmol^ꢀ1when the CBPQT⁴⁺ring moves from the

TTF to the DNP unit. Hence, the co-conformation (CC)

with the CBPQT⁴⁺ring encircling the TTF unit is referred

to as the ground-state co-conformation (GSCC). The first

two oxidation states of RATTF⁴⁺correspond to the TTF⁰!

+

TTFC !TTF²⁺processes. Upon formation of TTFC radical

cation, Coulombic repulsion between the CBPQT⁴⁺ring

and the TTFC results^[25–28]in the translation of the ring to

+

the DNP unit. This process^[32]is fast and is believed to con-

vert all of the GSCC into the MSCC. When the TTFC⁺radi-

cal cation is reduced back to TTF⁰, the CBPQT⁴⁺ring re-

mains around the DNP unit for a period of time. This trans-

lational isomer of the GSCC is the metastable-state co-con-

formation (MSCC). Recovery of the MSCC/GSCC equilibri-

um distribution (~1:9) is an activated process. This switching

cycle can be detected by a number of experimental observa-

tions. First, the lowest oxidation potential (corresponding to

TTF⁰!TTFC⁺) of the GSCC is +490 mV, while that for the

MSCC is +310 mV. (All potentials referenced to an Ag/

AgCl electrode.) Second, the colors of GSCC- and MSCC-

dominated solutions are green and red, respectively. Thus,

electrochemistry and spectroscopy can be employed to

quantify the MSCC/GSCC ratio in such a bistable rotaxane

at any given time. Third, the (activated) relaxation of an

MSCC- back to a GSCC-dominated distribution is tempera-

ture dependent, and so the kinetic parameters may be quan-

tified through time- and temperature-dependent measure-

ments. For example, the DG^°₂₉₈for this process in the case^[27]

of RATTF⁴⁺in the solution phase is 16.2(ꢁ0.3) kcalmol^ꢀ1.

We have recently reported on the MSCC!GSCC relaxa-

tion kinetics for a number of bistable [2]catenanes and

[2]rotaxanes in several different environments, including 1)

in acetonitrile,^[27]2) in monolayers ([2]rotaxanes only)

bonded to the surfaces of Au working electrodes,^[26]and 3)

in solid-state polymer electrolytes.^[25]In the case of the ace-

262

www.chemeurj.org

ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

tonitrile and the polymer electrolyte devices, we have dem-

onstrated^[25,27]that the relaxation kinetics were sensitive to

both molecular structure and physical environment, al-

though the overall switching mechanism remains the same.

In this paper, we extend these measurements to include

MSTJ devices, as well as establishing the ground-state equi-

librium thermodynamics. Three bistable [2]rotaxanes—

namely RATTF⁴⁺, RTTF⁴⁺, and RBPTTF⁴⁺—plus the con-

trol^[33][2]rotaxane RBLOCK⁴⁺were investigated. It is evi-

dent from inspection of the structural formulas of these

These thermodynamic differences will be rationalized in this

paper by reference to binding constants obtained by isother-

mal titration calorimetry (ITC) for the complexation of

model guests containing TTF, BPTTF, and DNP units, by

the CBPQT⁴⁺host in acetonitrile at 298 K.

Previously we have hypothesized^{[20–22,24,25,35]}that the

GSCC corresponds to the low-conductance (switch-open)

state of an MSTJ, while the MSCC corresponds to the high-

conductance (switch-closed) state. This hypothesis is consis-

tent with many observations, including the shift in the oxida-

tion potential of the TTF group that correlates with the

three [2]rotaxanes shown in Figures 1 and 2 that RATTF⁴⁺

,

Figure 2. Structural formulas of the translational isomers of the bistable rotaxanes a) RTTF⁴⁺and b) RBPTTF⁴⁺both in their GSCC and MSCC.

c) Structural formula of the sterically-blocked (SEt) [2]rotaxane RBLOCK⁴⁺used in control studies.

RTTF⁴⁺and RBPTTF⁴⁺can exist at equilibrium as two

translational isomers (or co-conformations). By contrast,

RBLOCK⁴⁺has the CBPQT⁴⁺ring located exclusively

around the DNP unit as a result of the presence of the

bulky SEt group on the monopyrrolotetrathiafulvalene unit,

acting as an effective steric barrier, and thus preventing

translational isomerism. The critical difference in the molec-

ular structures between the RATTF⁴⁺and RTTF⁴⁺pair and

the RBPTTF⁴⁺lies with the replacement of the simple TTF

unit for the bispyrrolotetrathiafulvalene (BPTTF) unit.^[34]

However, all three bistable rotaxanes have slightly different

stoppers—RATTF⁴⁺bears a substituted benzylic alcohol

function and both RTTF⁴⁺and RBPTTF⁴⁺have slightly dif-

ferent hydrophilic stoppers facilitating their incorporation

into MSTJ devices. The major difference in the switching

properties between these bistable rotaxanes is that the equi-

librium MSCC/GSCC ratio (~1:9) for RATTF⁴⁺and

RTTF⁴⁺is relatively temperature independent, while the

equilibrium MSCC/GSCC ratio (~1:4 at 298 K) for

switching from the GSCC to the MSCC structure. In addi-

tion, Goddardꢀs group^[36]has found by computational meth-

ods that the MSCC structure has extended electron delocali-

zation—and thus enhanced conductivity—in comparison

with the GSCC.

The switching kinetics of RATTF⁴⁺

, , and

RTTF⁴⁺

RBPTTF⁴⁺should be relatively similar. By contrast, the

ground-state thermodynamics—and hence the temperature

dependence of the switching amplitude—should be quite

different. In this paper, we employ temperature dependent

electrochemical and current-voltage measurements to corre-

late qualitatively the thermodynamic properties of RATTF⁴⁺

in 1) acetonitrile and 2) solid-state polymer electrolytes, and

of RTTF⁴⁺in MSTJs together with RBPTTF⁴⁺across all

three environments. We also correlate quantitatively the

MSCC!GSCC relaxation kinetics in these three different

physical environments. We find that the ground-state ther-

modynamic differences between the pair of TTF-containing

rotaxanes (RATTF⁴⁺and RTTF⁴⁺) and RBPTTF⁴⁺are rel-

atively independent of physical environment, but strongly

RBPTTF⁴⁺exhibits

a strong temperature dependence.

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

263

J. F. Stoddart, J. R. Heath, J. O. Jeppesen et al.

influenced by molecular structure. We also find that al-

though the MSCC!GSCC relaxation kinetics exhibit a

strong environmental dependence in the case of all three ro-

taxanes, the switching mechanism appears to be similar for

all three compounds and is robust and consistent in all three

environments. These findings allow us to refine our initial

hypothesis such that the high-conductance (switch-closed)

state of an MSTJ still corresponds to the MSCC, but that

the low-conductance (switch-open) state is now related to

the MSCC/GSCC ratio at equilibrium. These experiments

provide a proof-of-principle for the control of molecular

structure over a key device characteristic—temperature-de-

pendent switching amplitudes in molecular electronic devi-

ces.

Results and Discussion

Molecular design: Although the bistable [2]rotaxanes

RATTF⁴⁺, RTTF⁴⁺, and RBPTTF⁴⁺all contain DNP sites,

they differ in that the first two contain a TTF unit and the

third a BPTTF. To understand how these units influence the

switching in these bistable rotaxanes, a series of model

guests were investigated for their binding with the

CBPQT⁴⁺host, as its tetrakis(hexafluorophosphate) salt, by

using ITC. The model guests are shown in Figure 3a. They

are tetrathiafulvalene (TTF) and its bispyrrolo derivative

H₂BPTTF; their diethyleneglycol-disubstituted derivatives

TTF-DEG and BPTTF-DEG;

Figure 3. a) Structural formulas for a series of model guests. b) Host–

guest complexation between the CBPQT⁴⁺host and each of the guests.

and

1,5-dioxynaphthalene

Table 1. Thermodynamic binding data^[a]corresponding to the complexation between CBPQT⁴⁺and the indi-

vidual components of the bistable rotaxanes in MeCN determined by isothermal titration microcalorimetry at

298 K^[38]in addition to solution-phase thermodynamic data of bistable rotaxanes.

[c]

(DNP-OH) and its diethylene-

glycol-disubstituted derivative

DNP-DEG. Addition of the

DEG substituents to the TTF

and DNP units is known^[37]to

enhance their binding constants

with the CBPQT⁴⁺host to the

extent that they increase by up

to two orders of magnitude. By

contrast, the binding of BPTTF

by the CBPQT⁴⁺host is al-

ready quite high and only dou-

bles.

Guest

DH^[b]

DS^[c]

DG^[d]

K_a

[kcalmol^À1

]

[calmol^À1K^À1

]

[kcalmol^À1

]

[10³m^À1

]

TTF^[e]

À10.64Æ0.12

À14.21Æ0.06

À9.00Æ0.02

À8.20Æ1.70

À16.04Æ8.11

À15.41Æ0.02

À2.82Æ1.79^[j]

À6.64Æ0.67

À18.1

À5.27Æ0.03

À7.66Æ0.07

À6.66Æ0.03

À7.17Æ0.12

À3.63Æ0.36

À6.26Æ0.04

+1.56Æ0.24

+1.11Æ0.07

6.9Æ0.18

380.0Æ22.0

70.8Æ0.98

168.0Æ17.0

0.44Æ0.13

36.4Æ0.25

TTF-DEG

H₂BPTTF^[f]

BPTTF-DEG

DNP-OH^[g]

DNP-DEG^[h]

RATTF^{4+ [i]}

RBPTTF^{4+ [i]}

À22.1

À7.9

À3.6

À41.7

À30.8

À14.7Æ6.8^[j]

À26.0Æ2.5

[a] A 0.39 mm standard solution of CBPQT⁴⁺was used for all titrations into which solutions of various concen-

trations of guest were added in 5 mL aliquots (4.7 mm TTF; 3.2 mm TTF-DEG; 5.0 mm H₂BPTTF; 2.1 mm

BPTTF-DEG; 5.4 mm DNP-OH; 3.9 mm DNP-DEG). [b] Under the constant pressure of the instrument, DH

is obtained from the heat of the reaction.^[38][c] Fits were performed using software provided by Microcal LLC

software, and the stoichiometry of all complexes was between 0.97 and 1.03 indicating a 1:1 complex was

formed. [d] Calculated from the fitted value of K_a. [e] The binding constant for the complex formed between

TTF and CBPQT⁴⁺, previously measured in MeCN by the ¹H NMR single-point method, was determined to

The enthalpic contribution

DH to the binding affinity K_a

between DNP-DEG and the

CBPQT⁴⁺host is similar

(Table 1) to that for TTF-DEG, be 8000m^{À1 [39a]}

,

and was found to be 10000m^À1by the UV/Vis dilution method.^[39b][f] The binding constant for

the complex formed between H₂BPTTF and CBPQT⁴⁺, previously measured in Me₂CO by the UV/Vis dilu-

tion method, was determined to be 12000m^{À1 [40]}

[g] The binding constant for the complex formed between

DNP-OH and CBPQT⁴⁺, previously measured in MeCN by the UV/Vis dilution method, was determined to

be 990m^{À1 [41]}[h] The binding constant for the complex formed between DNP-DEG and CBPQT⁴⁺, previously

measured in MeCN by the UV/Vis dilution method, was determined to be 25400m^{À1 [41]}

[i] The given thermo-

but it is almost double that for

BPTTF-DEG. This larger dif-

ference between the enthalpy

changes DH of association for

the two complexes is also repre-

sented in the bistable rotaxanes

by the enthalpy change DH of

equilibrium associated with the

.

dynamic values for RATTF⁴⁺and RBPTTF⁴⁺were obtained by the variable temperature CV measurements,

and they correspond to thermodynamic differences between the MSCC and GSCC; that is, the changes in H,

S, and G when the CBPQT⁴⁺ring moves from the TTF(BPTTF) unit to the DNP unit. [j] The linear fit to

DG/T vs 1/T for RATTF⁴⁺produced a low R²of 0.4 because the DG for RATTF⁴⁺was reasonably insensitive

to temperature changes and therefore the data obtained reflects the standard error from the CV measure-

affinity of the CBPQT⁴⁺ring ments.

264

www.chemeurj.org

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

for the two recognition units.

Correspondingly, the bistable

rotaxane RBPTTF⁴⁺(À7.2 to

À6.6 kcalmol^À1for association

and equilibrium, respectively)

shows a much higher DH than

RATTF⁴⁺

(À1.2

to

À2.8 kcalmol^À1, respectively).

The direct consequence of this

large DH difference of associa-

tion between the complexes of

the CBPQT⁴⁺host with DNP-

DEG and BPTTF-DEG guests

is that the MSCC/GSCC ratio

for RBPTTF⁴⁺exhibits

a

strong temperature depend-

ence, such that the ratio

changes from 0.73 at 262 K to

0.25 at 284 K.^[42]Moreover, this

variable ratio should be detect-

able in all three environments.

In the solution phase and poly-

mer gels, the MSCC/GSCC

ratio can be quantified directly

through CV measurements. In

the MSTJs, the temperature-de-

pendent MSCC/GSCC ratio

Scheme 1. Synthesis of the semi-dumbbell compound 13.

should be reflected in a temper-

ature-dependent switching amplitude. By contrast with

RBPTTF⁴⁺, the smaller DH difference for the binding of

the CBPQT⁴⁺ring to the TTF and DNP units should favor

a relatively temperature-independent MSCC/GSCC ratio in

RATTF⁴⁺, with the GSCC remaining the dominant co-con-

formation at all temperatures and in all environments, a sit-

uation that is indeed observed. Irrespective of these differ-

ences in the ground-state thermodynamics, for both

RATTF⁴⁺and RBPTTF⁴⁺the actual electrochemically

driven switching mechanism should be the same.

10 (70%), which was tosylated by using TsCl in CH₂Cl₂af-

fording 11 in 81% yield. Subsequently, 11 was treated with

the DNP derivative^[28d]12 under alkylation conditions

(K₂CO₃/LiBr/MeCN) affording the BPTTF derivative 13 in

60% yield, which on treatment with TsOH in THF/EtOH,

gave (Scheme 2) the alcohol 14 in 56% yield. The free hy-

droxyl function in compound 14 was thereafter converted to

a tosylate group in 98% yield (14!15) and then to a thio-

cyanate group in 97% yield (15!16). The thiocyanate

group was reduced in situ with NaBH₄, and the resulting thi-

olate was subsequently coupled with the hydrophilic chlo-

ride^[21]17 in THF/EtOH to give the dumbbell 1 in 68%

yield. Finally, the [2]rotaxane RBPTTF·4PF₆was self-assem-

bled (Scheme 3) under high-pressure conditions by using the

dumbbell compound 1 as the template for the formation of

the encircling CBPQT⁴⁺tetracation; the [2]rotaxane

RBPTTF·4PF₆was isolated in 47% yield from a mixture of

the dumbbell compound 1, the dicationic precursor^[45]

2·2PF₆, and the dibromide 3 after they had been subjected

to a 10 kbar pressure in DMF at room temperature for three

days.

Synthesis and characterization: The synthesis of the bistable

rotaxanes RATTF⁴⁺(Figure 1) and RTTF⁴⁺(Figure 2) have

been described elsewhere.^[25,28d]UV-visible and ¹H NM R

spectroscopy indicate that both these bistable rotaxanes

exist predominantly (at least 90%) in the GSCC.

Synthesis of amphiphilic [2]rotaxane RBPTTF·4PF₆: The

[2]rotaxane RBPTTF·4PF₆was synthesized from precursors

1–3 according to the routes outlined sequentially in

Schemes 1–3. Alkylation of H₂BPTTF^[43](4) with 2-[2-(2-

iodoethoxy)ethoxy]tetrahydropyran^[44](5) in DMF gave the

BPTTF derivative 6 in 67% yield (Scheme 1). Removal of

the THP-protecting groups with p-toluenesulfonic acid

(TsOH) gave the diol 7 in 67% yield. The monotosylate 8

was obtained in 22% yield by reaction of the diol 7 with

one equivalent of p-toluenesulfonyl chloride (TsCl). Alkyla-

RBPTTF⁴⁺exists as a mixture of the two possible isomers

in which the CBPQT⁴⁺ring is located around the BPTTF

unit in the GSCC and around the DNP unit in the MSCC.

These two isomers give rise to characteristic charge-transfer

(CT) absorption bands centered on 825 (GSCC) and 550 nm

(MSCC), respectively, in the electronic absorption spectrum.

The equilibrium population ratio can be determined from

tion of the hydrophobic tetraarylmethane-based stopper^[21]

with 8 in MeCN in the presence of K₂CO₃gave the alcohol

9

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

265

J. F. Stoddart, J. R. Heath, J. O. Jeppesen et al.

Scheme 2. Synthesis of the dumbbell compound 1.

Synthesis of the amphiphilic

[2]rotaxane

RBLOCK·4PF₆:

The [2]rotaxane RBLOCK·4PF₆

was synthesized according to

the routes outlined sequentially

in Schemes 4 and 5. A solution

of 4,5-bis(2-cyanoethylthio)-1,3-

dithiole-2-thione^[46](19) in THF

was treated with one equivalent

of NaOMe. This procedure gen-

erated the monothiolate, which

was alkylated with EtI affording

compound 20 in 92% yield.

Cross-coupling of 5-tosyl-(1,3)-

dithiolo[4,5-c]pyrrole-2-one^[43]

(21) with three equivalents of

the thione 20 in neat (EtO)₃P

gave (Scheme 4) the MPTTF

derivative 22 (74%) in gram

quantities after column chroma-

tography. The iodide^[28h]18 was

coupled with the MPTTF build-

ing block 22, following its in

situ deprotection with one

equivalent of CsOH·H₂O to

Scheme 3. Synthesis of the bistable [2]rotaxane RBPTTF·4PF₆.

give 23 in 87% yield. The tosyl

the

¹H NMR

spectrum

(400 MHz)

recorded

on

protecting group on the MPTTF unit was removed in good

yield (87%) using NaOMe in a THF/MeOH mixture. The

resultant pyrrole nitrogen atom in 24 was alkylated with

compound 25,^[47]affording the chloride 26 in 75% yield. The

chloride in compound 26 was initially converted to (26!27)

an iodide in 99% yield and then to (27!28) a thiocyanate

group in 99% yield. The thiocyanate group was reduced in

situ with NaBH₄, and the resulting thiolate was subsequently

coupled with the hydrophilic chloride^[21]17 in THF to give

RBPTTF·4PF₆, since several protons in the dumbbell com-

ponent give rise to two sets of signals, one for each of the

two isomers. From integration of the resonances associated

with the benzylic protons in the hydrophilic stoppers, the

MSCC/GSCC equilibrium population ratio was found to be

1:3 at 295 K in CD₃CN.

266

www.chemeurj.org

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

Scheme 4. Synthesis of the semi-dumbbell compound 28.

the dumbbell 29 in 78% yield. Finally, the [2]rotaxane

RBLOCK·4PF₆was self-assembled (Scheme 5) under high-

pressure conditions by using the dumbbell compound 29 as

the template for the formation of the encircling CBPQT⁴⁺

tetracation; the [2]rotaxane

RBLOCK·4PF₆was isolated in

41% yield from a mixture of

the dumbbell compound 29, the

dicationic precursor^[45]2·2PF₆,

and the dibromide 3 after they

had been subjected to a 10 kbar

pressure in DMF at room tem-

perature for three days.

Kinetics and thermodynamics

of switching in solution and in

polymer electrolytes: We have

previously demonstrated that

the first oxidation potentials of

bistable rotaxanes can be utiliz-

ed to quantify the MSCC/

GSCC ratios in the solution

phase,^[27]for monolayers assem-

bled onto Au surfaces,^[26]and

for polymer electrolyte gels.^[25]

In this section, we report on a

set of similar variable time and

temperature cyclic voltammetry

(VTTCV) measurements in so-

lution and polymer environ-

ments to probe the thermody-

namics of the MSCC/GSCC

Scheme 5. Synthesis of the [2]rotaxane RBLOCK·4PF₆.

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

267

J. F. Stoddart, J. R. Heath, J. O. Jeppesen et al.

equilibrium ratios for RATTF⁴⁺and RBPTTF⁴⁺. From

these measurements, we can extract free energy differences

(DG from Figure 1b) between the two co-conformations.

We also utilized VTTCV to quantify the kinetics (DG^ꢀ

from Figure 1b) associated with the recovery of the equilib-

rium MSCC/GSCC distribution for RBPTTF⁴⁺and

RATTF⁴⁺. The relaxation kinetics for [2]rotaxane RATTF⁴⁺

and for related TTF-based rotaxanes were thoroughly inves-

tigated in recent papers,^[25–27]while the equivalent VTTCV

measurements for RBPTTF⁴⁺are reported here for the first

time.

The VTTCV measurements were carried out as follows:

two CV cycles were collected in succession, starting with the

system at equilibrium. This first CV cycle displays peaks

that can be assigned to the resting state populations of the

MSCC and GSCC, since the first oxidation potential of the

TTF (BPTTF) group is sensitive to whether or not it is en-

circled by the CBPQT⁴⁺ring. The second cycle, if collected

quickly enough, records a shift in the equilibrium population

towards the one dominated by the MSCC. This shift is re-

flected in an increase in the area of the peak assigned to the

MSCC, in coincidence with a decrease in the area for the

GSCC peak. By controlling the time between the first and

second CV cycles, and the temperature of the experiment,

the kinetic parameters associated with the recovery of the

MSCC/GSCC equilibrium ratio can be quantified.

We first focus on utilizing VTTCV to probe the MSCC/

GSCC population ratio at thermal equilibrium. The CVs of

RBPTTF⁴⁺in the solution phase exhibit

a peak at

+

+530 mV, which corresponds to the BPTTF!BPTTFC oxi-

dation of the proportion of the bistable rotaxane that exists

in the MSCC (Figure 4a). The smaller peak at +680 mV

corresponds to the BPTTF!BPTTFC⁺oxidation of the

GSCC. The larger peak at +780 mV corresponds to the

+

second oxidation (BPTTFC !BPTTF²⁺). This second oxida-

tion is independent of the co-conformation, since once the

+

BPTTFC is formed, the CBPQT⁴⁺ring moves to the DNP

unit. The MSCC/GSCC population was thus measured as a

function of temperature. For RBPTTF⁴⁺, decreasing the

temperature led to a significant increase in the MSCC/

GSCC population ratio. The ratio, for example, increases

(Figure 4a) more than two-fold (from around 0.3 to 0.7) as

the temperature is decreased from 284 to 262 K. By compar-

ison, for RATTF⁴⁺, the MSCC/GSCC population ratio does

not deviate significantly from 0.1, even when the rotaxane is

probed (Figure 4b) over a broader temperature range (248–

283 K).

Figure 4. The first CV cycles of a) RBPTTF⁴⁺recorded at 262 and 284 K,

b) RATTF⁴⁺recorded at 248 and 283 K, and c) RBLOCK⁴⁺recorded at

295 K (MeCN/0.1m TBAPF₆/200 mVs^À1). The peak assigned to the

MSCC at approximately +500 mV for RBPTTF⁴⁺fluctuates more than

for RATTF⁴⁺across different temperature ranges. The simple, dumbbell-

like CV for RBLOCK⁴⁺, displaying a full-intensity MSCC peak at about

500 mV, verifies that the CBPQT⁴⁺ring is sterically blocked.

The relative temperature dependences of the MSCC/

GSCC ratios for RBPTTF⁴⁺and for RATTF⁴⁺are consis-

tent with the ITC investigations of the complexation of the

CBPQT⁴⁺host with the individual BPTTF-DEG, TTF-

DEG, and DNP-DEG guests that were discussed above and

presented in Table 1. Translating the behavior of the guests

to what might be predicted for the two bistable [2]rotaxanes,

one expects that the enthalpic contribution DH=

(H_MSCCÀH_GSCC) should be significantly less than zero for

RBPTTF⁴⁺. By comparison, the corresponding DH for

RATTF⁴⁺should be much closer to zero. As a consequence,

the MSCC/GSCC ratio for RBPTTF⁴⁺varies^[42]more readi-

ly with temperature.

Although it is not so straightforward to interpret, the long

and flexible diethylene glycol chains appear to have an

impact on the binding K_aand therefore the population

ratios of the bistable rotaxanes. The DEG chains enhance

(Table 1) the binding affinity for each of the three guests

268

www.chemeurj.org

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

with the CBPQT⁴⁺host, but they do so by influencing the

DH and DS of each component differently. For TTF-DEG,

the DEG chains leads to better enthalpy, but worse entropy.

However, for the DNP-DEG and BPTTF-DEG guests, it is

the opposite with the entropy contribution favoring binding

and enthalpy disfavoring it, albeit only mildly so. Further-

more, it is known that when these DEG chains are attached

to DNP and TTF units they are capable of wrapping them-

selves around the CBPQT⁴⁺ring in order to acquire stabi-

[37]

À

lizing, noncovalent C H···O interactions.

Consequently,

the significant enhancement of the enthalpic contribution to

the complexation between TTF-DEG and the CBQPT⁴⁺

host by the DEG chains brings its DH to within a few

kcalmol^À1of the DNP-DEG guest, leading to a relatively

temperature-insensitive MSCC/GSCC ratio for the rotaxane

RATTF⁴⁺. However, the DEG chains have little effect on

the DH contribution to complexation of the DNP-DEG and

BPTTF-DEG guests by the CBPQT⁴⁺host, such that they

maintain their large and significant differences in enthalpy

within the RBPTTF⁴⁺, leading to the rotaxaneꢀs correspond-

ingly large sensitivity of the population ratios to tempera-

ture. The DEG chains are thus an essential factor influenc-

ing the temperature sensitivities of the MSCC/GSCC popu-

lation ratios of these bistable rotaxanes. The observation

from the electrochemical studies in the solution phase and

in the polymer matrix provide a view of both RATTF⁴⁺and

RBPTTF⁴⁺that is completely consistent with the ITC meas-

urements on the subunits of the rotaxanes. It is also consis-

tent with the molecular structure differences between these

two switches.

The relaxation kinetics and thermodynamics associated

with the free-energy barrier (DG^ꢀ) for relaxation from the

MSCC to the GSCC for RATTF⁴⁺and RBPTTF⁴⁺were

also analyzed quantitatively. The viscosity of the acetonitrile

phase and polymer gel were about 3.5 cP and 50000 cP at

298 K, respectively. This large increase in viscosity is reflect-

ed in the slower first-order decay kinetics for RBPTTF⁴⁺as

measured by VTTCV. Data for acetonitrile and the polymer

gel are presented in Figure 5a and b, respectively. In addi-

tion to the viscosity effects, these plots also reveal how the

thermally activated relaxation rates drop as the temperature

is lowered. Itꢀs instructive to notice that both the MSCC and

GSCC are at significant concentrations under equilibrium

conditions for RBPTTF⁴⁺, especially at lower temperatures.

The implication is that the reverse reaction GSCC!MSCC

is occurring at a rate comparable to that of the forward re-

action. In analyzing the relaxation kinetics, both processes

should be taken into consideration. Thus, for the equilibri-

um reaction given in Equation (1) the formula in Equa-

tion (2) is readily obtained (cf. Experimental Section), in

which x_t=N_MSCC/N_Totalis the MSCC population ratio at time

t, x₀=x_t=0, and x_eq=x_t!1is the MSCC population ratio at

final equilibrium.

Figure 5. Fitted exponential decay curves and time constants (t) obtained

from the CV data for RBPTTF⁴⁺measured at various scan rates for each

temperature in a) solution and b) polymer phases.

Experimental relaxation data were thus fitted with this

formula to obtain the decay time constant t, and accordingly

the rate constant for the forward reaction k₁=(1Àx_eq)/t.

Note that when x_eqis small (i.e., for the case of R(A)TTF⁴⁺),

the formula naturally reduces to the more familiar formula

for a simple first-order reaction. The parameters DG^ꢀ, DH^ꢀ,

DS^ꢀ, and E_awere then fitted from the temperature depend-

ence of k₁.^[49,50]The kinetic data are summarized in Table 2

alongside values for R(A)TTF⁴⁺. Note that the MSCC/

GSCC population ratio for RBPTTF⁴⁺, as measured at long

times (at equilibrium), shows (Figure 5) a significant sensi-

tivity to temperature that is consistent with the thermody-

namic descriptions and data for the host–guest complexa-

tion. By contrast, RATTF⁴⁺displays only a small thermal

sensitivity in both the polymer and solution phase environ-

ments.

Kinetics and thermodynamics of molecular-switch tunnel

junctions: The MSTJs investigated here contained a mono-

layer of the amphiphilic bistable rotaxanes RTTF⁴⁺or

RBPTTF⁴⁺, or the sterically blocked metastable-like rotax-

ane RBLOCK⁴⁺, sandwiched between an n-type poly-

k₁

MSCC

GSCC

H

ð1Þ

ð2Þ

G

k₂

x_t¼ x_eqþ ðx₀Àx_eqÞ exp½Àk₁t=ð1Àx_eqÞ

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

269

J. F. Stoddart, J. R. Heath, J. O. Jeppesen et al.

Table 2. Kinetics data for the relaxation from the MSCC to the GSCC for RBPTTF⁴⁺and the free-energy barriers for RATTF⁴⁺and RTTF⁴⁺. Data for

solution, polymer and MSTJ were obtained from variable temperature CVs and from measurements of the relaxation of a MSTJ from the high to the

low conductance state.^[49,50]

t₂₉₈

k₂₉₈

DG^ꢀ₂₉₈

DH^ꢀ

DS^ꢀ

E_a

DG^ꢀ₂₉₈

[s]

[s^À1

]

[kcalmol^À1

]

[kcalmol^À1

]

[calmol^À1K^À1

]

[kcalmol^À1

]

RATTF⁴⁺

RTTF⁴⁺

solution^[a]

polymer^[b]

MSTJ

1.26Æ0.10

10.2Æ0.12

624Æ82

0.69Æ0.05

17.69Æ0.05

19.15Æ0.01

21.7Æ0.1

8.4Æ0.5

8.4Æ1.1

16.1Æ1.4

À31.0Æ1.7

À36.0Æ3.4

À18.7Æ4.1

9.0Æ0.5

9.0Æ1.0

16.7Æ1.3

16.2Æ0.3

18.1Æ0.2

–

0.059Æ0.001

(8.4Æ0.8)10^À4

22.21Æ0.04

[a] Solution-phase data was obtained for 1 mm samples dissolved in MeCN (0.1m TBAPF₆) using a glassy carbon working electrode. All potentials were

referenced to a Ag/AgCl reference electrode.^[27][b] Polymer-phase data was obtained in a polymer matrix—w:w:w:w ratios of 70:7:20:3 for MeCN/poly-

methylmethacrylate/propylene carbonate/LiClO₄. The sample was spread onto three lithographically patterned Pt electrodes (50 nm) on top of Ti

(10 nm) (working, counter, reference).^[25]The DH^ꢀand DS^ꢀwere obtained from an average of many devices, while the Eyring plot in Figure 8b repre-

sents just one device.

(silicon) bottom electrode (passivated with the native oxide)

and a metallic top electrode. The detailed procedures relat-

ing to the fabrication and operation of these devices have

been previously reported.^[20–22]Briefly, the molecules are

prepared as a Langmuir–Blodgett film and then transferred

as a compressed Langmuir monolayer (p=30 mNm^À1) onto

a substrate pre-patterned with poly-silicon electrodes. A

thin 10 nm Ti adhesion layer, followed by a thicker 200 nm

top Al layer is evaporated through a shadow mask by using

e-beam evaporation to form the top electrodes. During this

step, the substrate is held at room temperature at a source–

sample distance of ~0.7 m. This procedure ensures that little

or no substrate heating from the e-beam evaporation source

occurs. The e-beam evaporation was processed at the depo-

sition rate of 1–2 s^À1under high vacuum (~510^À7Torr).

For all experiments reported here, the bottom electrodes

were 5 mm wide, n-type (doping level ~510^À19cm^À3) poly-

Si, while the top electrodes were 10 mm wide. Each fabrica-

tion run produced approximately 100 MSTJ devices, and the

results presented here were consistently observed in multi-

ple devices across multiple fabrication runs, with tempera-

ture-dependent data collected in random sequence. More

than 90% of the MSTJ devices displayed consistent and re-

producible temperature dependence. The operational char-

acteristics of MSTJs containing bistable catenanes and ro-

taxanes, but patterned at both larger and also much smaller

dimensions, have been reported.^[22,24]

CV measurements are not possible for MSTJs, but there

are other electronic measurements that can be carried out

to assess both the thermodynamic and kinetic properties of

the bistable rotaxanes within the devices. Our hypothesis—

for both bistable catenanes and bistable rotaxanes—has

been refined such that the MSCC represents the high-con-

ducting, switch-closed state of the device, while the MSCC/

GSCC ratio at equilibrium represents the low-conducting,

switch-open state. For an MSCC-dominated system, regard-

less of environment, reduction of the CBPQT⁴⁺ring pro-

vides a rapid route towards recovering the equilibrium

MSCC/GSCC distribution.^[20,26]In the absence of such a re-

duction step, a device in the high-conductance state will

relax to the equilibrium MSCC/GSCC ratio, according to a

timescale described by DG^ꢀ(Figure 1b). From a practical

point of view (i.e., for memory devices), this relaxation pro-

cess correlates to the volatility, or memory-retention charac-

teristics, of the device. The volatility can be quantified by

measuring the temperature dependence of the decay of the

switch-closed, high-conductance state of a device back to

the switch-open state.

The equilibrium thermodynamic properties of the devices

can also be inferred within MSTJs by considering that the

high- and low-conductance states of the devices correlate

with different MSCC/GSCC ratios. Thus, the temperature-

dependent switching amplitude, normalized against the tem-

perature-dependent transport characteristics of an MSTJ,

opens a window into the thermodynamics of the molecules

within the junction. Such a measurement provides a qualita-

tive picture that can be compared against quantitative

VTTCV measurements of the MSCC/GSCC ratios in other

environments.

Measurements of the bistable character of MSTJs contain-

ing RTTF⁴⁺, RBTTF⁴⁺, and the RBLOCK⁴⁺control rotax-

ane are shown in Figure 6a and b. This type of data is called

a remnant molecular signature,^[20–22]and represents a nearly

capacitance-free map of the hysteretic response of an MSTJ.

Briefly, the x axis of a remnant molecular signature plot cor-

relates to a value of a voltage pulse that is applied across

the junction. A train of voltage pulses, starting at 0 V and

following the direction of the arrows shown on the plots, is

applied to the MSTJ, and, after each voltage pulse, the cur-

rent through the MSTJ is monitored at +0.1 V.^[51]The re-

sulting normalized current is represented on the y axis.

These hysteresis loops not only provide a key indicator that

the MSTJs can be switched reversibly between the high- and

low-conducting states, but they also qualitatively reflect the

ground-state MSCC/GSCC ratio, since the switching ampli-

tude is sensitive to this ratio. For the high-conductance state,

in which the entire population has been converted into the

MSCC, the maximum current is controlled by the intrinsic

conductance properties of this co-conformation. However,

for the low-conductance state, the minimum conduction is

not only controlled by the intrinsic properties of the GSCC

but also by the MSCC/GSCC ratio—a factor under thermo-

dynamic control. For instance, at 295 K RBPTTF⁴⁺and

RBLOCK⁴⁺do not appear to be “good” switches, while the

switching amplitude of RTTF⁴⁺is about a factor of 8. At

320 K, the small hysteretic response for RBLOCK⁴⁺dimin-

270

www.chemeurj.org

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

sured in the low conductance state) of over 3. This enhanced

switching amplitude presumably reflects a smaller MSCC/

GSCC equilibrium ratio at the higher temperature, and is

consistent with what is observed for the solution and poly-

mer phase measurements for RBPTTF⁴⁺. The switching am-

plitude of RTTF⁴⁺remains fairly constant across this tem-

perature range, consistent again with measurements in the

other environments. The MSCC!GSCC relaxation kinetics

can be monitored by measuring the time-dependence of the

decay of the high-conductance to the low-conductance state;

the data for all three amphiphilic rotaxanes at 295 K is pre-

sented in Figure 6c.

The high- to low-conductance decay of all three rotaxanes

exhibited different temperature dependences. While MSTJs

fabricated from RBPTTF⁴⁺and RTTF⁴⁺show strong tem-

perature dependences—as the temperature was increased

from 295 K to 320 K, the 1/e decay time decreased by fac-

tors of 6–7 for those rotaxanes (Figure 7)—RBLOCK⁴⁺ex-

hibited a much weaker temperature dependence. MSTJs

fabricated from RBLOCK⁴⁺were investigated over

a

Figure 6. Switching responses of three rotaxanes within MSTJs. a) and b):

Remnant molecular signature traces of the hysteretic switching responses.

The arrows indicate the direction of the voltage sweep, and all currents

were recorded at +0.1 V. The y axis current was normalized by setting

the initial (low-conductance state) current to 1. Note that the response of

RBPTTF⁴⁺increases in amplitude at higher temperature, reflecting a de-

creasing MSCC/GSCC equilibrium ratio, while RTTF⁴⁺is relatively con-

stant. There is

a finite amount of field-induced polarization in

RBLOCK⁴⁺that is almost undetectable at 320 K. c) Relaxation of

MSTJs from high-to-low conducting states recorded at 295 K. The char-

acteristic relaxation times are: RTTF⁴⁺=3450 s; RBPTTF⁴⁺=660 s;

RBLOCK⁴⁺=60 s.

ishes further, but the hysteresis loop of RBPTTF⁴⁺opens

up to yield a switching amplitude (i.e., the current measured

in the high conductance state divided by the current mea-

Figure 7. Decay curves of a) RTTF⁴⁺and b) RBPTTF⁴⁺MSTJs recorded

as a function of temperature. Note that the normalized switching ampli-

tude of RBPTTF⁴⁺exhibits a strong temperature dependence.

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

271

J. F. Stoddart, J. R. Heath, J. O. Jeppesen et al.

broader temperature range (295–383 K) and the characteris-

tic relaxation time decreased by only a factor of 2 or so over

this entire range. This decay-rate data fitted well to a 1/T

plot (R²>0.99), which is at least consistent with existing

models for dielectric relaxation,^[52,53]although measurements

over an even broader temperature range would be required

to establish this relationship more firmly. In any case,

MSTJs fabricated from RBLOCK⁴⁺were poor switches at

all temperatures investigated, and the small switching re-

sponse that could be recorded exhibited a very different and

much less-pronounced temperature dependence, in compari-

son to MSTJs fabricated from RBPTTF⁴⁺and RTTF⁴⁺

.

The switching amplitude can be recorded by either meas-

uring the amplitude of the hysteresis loops from the rem-

nant molecular signature data, or by measuring the time-de-

pendent decay of the high- to the low-conductance state.

Any molecular electronic junction for which charge trans-

port is not strictly a quantum mechanical tunneling process

will exhibit a strong temperature-dependent conductance,

that is, charge transport is thermally activated, and the rate

of transport increases with increasing temperature. This is

the case for all three of the amphiphilic rotaxanes investigat-

ed here. However, this temperature-dependent component

should depend only weakly upon molecular structure, espe-

cially for molecules that are as similar as RTTF⁴⁺

,

RBPTTF⁴⁺, and RBLOCK⁴⁺, and should not be particularly

sensitive to the MSCC/GSCC ratio within a device. Thus, we

remove this component of the temperature dependence by

normalizing the switching amplitude to the initial current

value, measured at t=0 after placing the switch into the

high conductance state. The hypothesis is that the (normal-

ized) current at long times—that is, when the system has

reached equilibrium—divided by the t=0 current should

correlate^[54]qualitatively with the MSCC/GSCC ratio. In

Figure 7, we present such normalized decay curves, for vari-

ous temperatures, for both RTTF⁴⁺and RBPTTF⁴⁺. Note

two things about the data of Figure 7. First, the curves clear-

ly represent activated processes, since, for both bistable ro-

taxanes, the relaxation times decrease rapidly with increas-

ing temperature. Second, the switching amplitude for

RTTF⁴⁺is relatively temperature independent, exhibiting

almost an order-of-magnitude difference in the (normalized)

current change between the high- and low-conductance

states for all temperatures. By contrast, the switching ampli-

tude for RBPTTF⁴⁺exhibits a strong temperature depen-

dence over the same range. This observation is consistent

with the remnant molecular signature data presented in

Figure 6. Also, it is consistent with the behavior of the corre-

sponding bistable rotaxanes (RATTF⁴⁺and RBPTTF⁴⁺) in

the other environments, as well as the ITC data obtained

from host-guest complexation experiments.

Figure 8. a) The temperature-dependent GSCC/MSCC equilibria for all

three environments are presented. Solution and polymer phase data

(N_MSCC/N_Total) were recorded for RATTF⁴⁺and RBPTTF⁴⁺and are

based upon quantitative electrochemical measurements of the MSCC/

GSCC ratios. The MSTJ data, which were recorded for RTTF⁴⁺and

RBPTTF⁴⁺

(I_OPEN/I_CLOSED), and represent a qualitative measurement of the N_MSCC

, show the temperature-dependent switching amplitude

/

N_Totalratio, based upon the proposed switching mechanism. Note that the

large (enthalpically driven) temperature dependence for RBPTTF⁴⁺, and

the relative temperature independence of RATTF⁴⁺and RTTF⁴⁺

(R(A)TTF⁴⁺) is reflected in all environments. b) Eyring plots of the

MSCC!GSCC (or high!low-conducting MSTJ) relaxation process, for

all three environments.

the temperature-dependent switching amplitude I_OPEN

/

I_CLOSEDprovides for qualitative comparison^[54]with the other

environments. For the relaxation kinetics, data for the two

TTF-containing rotaxanes (R(A)TTF⁴⁺and RBPTTF⁴⁺

)

are plotted in the form of Eyring plots, in order to quantify

(Table 2) DG^ꢀ, DH^ꢀ, and DS^ꢀin all three environments.^[49,50]

We first consider the kinetic data given in Figure 8b and

Table 2. For the case of RBPTTF⁴⁺, the free-energy barrier

(DG^ꢀ) to relaxation at 298 K increases from 17.7 to 19.2 to

21.7 kcalmol^À1upon moving from acetonitrile to polymer

gels to MSTJs. For R(A)TTF⁴⁺, the situation is qualitatively

similar. Both rotaxanes exhibit an increase in the energy

barrier DG^ꢀfrom the solution to polymer phase by between

The temperature-dependent thermodynamic and relaxa-

tion kinetic data for all environments are presented in Fig-

ure 8a and b, respectively. In Figure 8a, we have plotted the

temperature-dependent ratios as N_MSCC/N_TOTAL, quantitative-

ly measured in the solution-phase and polymer environ-

ments. For the MSTJs, this ratio cannot be quantified, but

272

www.chemeurj.org

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

1 and 2 kcalmol^À1. However, the DG^ꢀincrease in moving

from the polymer to the MSTJ is significantly larger for

R(A)TTF⁴⁺than for RBPTTF⁴⁺(4.1 vs 2.5 kcalmol^À1). This

difference may be related to the differences in packing be-

tween the Langmuir monolayers of the amphiphilic rotax-

anes. Both monolayers were transferred onto the electrode-

patterned substrate at a pressure of 30 mNm^À1. However,

to validate the proposed switching mechanism and its uni-

versality, but also to correlate switching speeds with the

nature of the environment. The trends in the kinetics and

the validity of the switching mechanism hold true for both

classes of bistable rotaxanes. Nevertheless, we are still able

to differentiate them by following a detailed thermodynamic

investigation of their thermally equilibrated states. By re-

placing the TTF unit in the bistable rotaxanes with a

BPTTF unit, the equilibrium MSCC/GSCC population ratio,

which influences the low-conductance state in MSTJs and

the temperature sensitivity of this ratio, was altered consid-

erably. Correspondingly, the switching amplitude between

the high-conductance state and the now thermally sensitive

low-conductance state changes significantly with tempera-

ture. Binding constant measurements for the complexation

of model guests with the CBPQT⁴⁺host verify that the pop-

ulation ratio and its temperature sensitivity are directly re-

lated to the different binding strengths of the DEG-disubsti-

tuted TTF and BPTTF units. Enthalpy is found to play a

crucial role in determining the temperature dependence of

these binding strengths. The observation of a mode of ther-

modynamic behavior that is universal regardless of environ-

ment validates the hypothesis that at least certain key op-

erational characteristics of the MSTJs are under the control

of chemical synthesis. This realization represents a key ele-

ment in the emerging paradigm of molecular electronics.

2

the RTTF⁴⁺rotaxanes occupy 92Æ3 per molecule, while

2

the RBPTTF⁴⁺rotaxanes occupy 122Æ5 per molecule.

Thus, the packing of RBPTTF⁴⁺is influenced by a combina-

tion of the high MSCC/GSCC ratio and the bulkier hydro-

philic stopper. These differences lead to a 30% increase in

the area per molecule over a similarly compressed film of

RTTF⁴⁺. Nevertheless, for both amphiphilic bistable rotax-

anes, the data in Figures 6 and 7 indicate a qualitatively sim-

ilar switching mechanism, regardless of physical environ-

ment.

The thermodynamic data given in Figure 8a are apparent-

ly more reflective of the structural differences between

R(A)TTF⁴⁺and RBPTTF⁴⁺, rather than the physical envi-

ronment of these molecules. In all environments, RBPTTF⁴⁺

exhibits a strongly temperature-dependent switching ampli-

tude that can be related back to the temperature depen-

dence of the MSCC/GSCC ratio. In turn, this behavior can

be connected to the free-energy difference between the two

host–guest complexes, BPTTF-DEGꢀCBPQT⁴⁺and DNP-

DEGꢀCBPQT⁴⁺, and the fact that the enthalpic contribu-

tion to the free energy is very different for these two com-

plexes. The temperature dependence of the MSCC/GSCC

ratio of RBPTTF⁴⁺is slightly more pronounced for the so-

lution and polymer environments than for the MSTJ. This is

likely due to the fact that the MSTJ constitutes a more steri-

cally crowded environment. Nevertheless, the degree to

which the free-energy landscape of the bistable RBPTTF⁴⁺

is reflected in the properties of this molecule, regardless of

environment, is striking.

In a similar way, the temperature independent switching

amplitude of R(A)TTF⁴⁺can also be rationalized within a

self-consistent picture that connects across all environments

as well as to the free-energy differences between the TTF-

DEGꢀCBPQT⁴⁺and DNP-DEGꢀCBPQT⁴⁺host–guest

complexes. From the point of view of an MSTJ-based

memory device, RTTF⁴⁺constitutes a much superior switch

than does RBPTTF⁴⁺. First, it exhibits a stable switching

amplitude over a reasonably broad temperature range.

Second, an RTTF⁴⁺-based MSTJ remains in the high-con-

ducting (MSCC-dominated) state five times longer than an

RBPTTF⁴⁺-based MSTJ at 295 K, and ten times longer at

320 K, implying a less volatile (and more useful) switch.

Experimental Section

General methods: Chemicals were purchased from Aldrich and were

used as received, unless indicated otherwise. Bis(pyrrolo[3,4-d])tetrathia-

fulvalene^[43](4; Scheme 1), 2-(2-iododethoxy)-ethyl-p-toluenesulfonate^[44]

(5; Scheme 1), 4-[bis(4-tert-butylphenyl)(4-ethylphenyl)methyl]phenol^[21]

(9; Scheme 1), compound^[28d]12 (Scheme 1), the chloride^[21]17 (Scheme 2

and 4), 1,1’’-[1,4-phenylenebis(methylene)]bis(4,4’-bipyridin-1-ium) bis-

(hexafluorophosphate)^[45](16·2PF₆) (Schemes 3 and 5), the iodide^[28h]18

(Scheme 4),

4,5-bis(2-cyanoethylthio)-1,3-dithiole-2-thione^[46]

(19;

Scheme 4), 5-tosyl-(1,3)-dithiolo-[4,5-c]pyrrole-2-one^[43](21; Scheme 4),

and 2-(2-chloroethoxy)-ethyl-p-toluenesulfonate^[47](25; Scheme 4) were

all prepared according to literature procedures. Solvents were dried ac-

cording to literature procedures.^[48]All reactions were carried out under

an anhydrous nitrogen atmosphere. High-pressure experiments were car-

ried out in a teflon tube on a Psika high-pressure apparatus. Thin-layer

chromatography (TLC) was carried out by using aluminium sheets pre-

coated with silica gel 60F (Merck 5554). The plates were inspected under

UV light and, if required, developed in I₂vapor. Column chromatogra-

phy was carried out using silica gel 60F (Merck 9385, 0.040–0.063 mm).

Deactivated SiO₂was prepared by stirring the silica gel in CH₂Cl₂con-

taining 2% Et₃N for 10 min before it was filtered, washed with CH₂Cl₂,

and dried. Melting points were determined on a Büchi melting point ap-

paratus and are uncorrected. ¹H NMR spectra were recorded at room

temperature on a Bruker ARX500 spectrometer (500 MHz), Bruker

ARX400 spectrometer (400 MHz), or on a Gemini-300BB instrument

(300 MHz), with residual solvent as the internal standard. ¹³C NMR spec-

tra were recorded at room temperature on a Gemini-300BB instrument

(75 MHz), with residual solvent as the internal standard. ¹⁹F NMR spec-

tra were recorded at room temperature on a Bruker ARX400 spectrome-

ter (376 MHz), while ³¹P NMR spectra were recorded at room tempera-

ture on Bruker ARX400 instrument (161 MHz). All chemical shifts are

quoted on a d scale, and all coupling constants (J) are expressed in Hertz

(Hz). The following abbreviations are used in listing the NMR spectra:

Conclusion

We have investigated two classes of bistable rotaxanes—one

containing a TTF unit and the other a BPTTF unit—across

different environments. Quantifying the relaxation rates in

one critical step of the switching cycle enables us, not only

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

273

J. F. Stoddart, J. R. Heath, J. O. Jeppesen et al.

s=singlet, d=doublet, t=triplet, q=quartet, brs=broad singlet, and

m=multiplet. Samples were prepared in CDCl₃, CD₃COCD₃, or

CD₃SOCD₃purchased from Cambridge Isotope Labs. Electron impact

ionization mass spectrometry (EI-MS) was performed on a Varian MAT

311 A instrument and matrix-assisted laser-desorption/ionization time-of-

flight mass spectrometry (MALDI-TOF-MS) was performed on a Kratos

Kompact MALDI-TOF instrument, utilizing a 2,5-dihydroxybenzoic acid

matrix, high-resolution Fourier Transform matrix-assisted laser-desorp-

tion/ionisation mass spectrometry (HiRes-FT-MALDI-MS) was per-

formed on an IonSpec 4.7 tesla Ultima Fourier Transform mass spectrom-

eter, utilizing a 2,5-dihydroxybenzoic acid (DHP) matrix, while electro-

spray mass spectra (ES-MS) were obtained from a from a Sciex API III⁺

mass spectrometer. Infrared (IR) spectra were recorded on a Perkin–

Elmer 580 spectrophotometer. UV-visible spectra were recorded at room

temperature on a Shimadzu UV-160 instrument. Microanalyses were per-

formed by the Atlantic Microlab, Inc., Atlanta, Georgia.

Compound 10: A solution of the monotosylate 8 (0.37 g, 0.60 mmol) and

9 (0.86 g, 1.80 mmol) in anhydrous MeCN (50 mL) containing K₂CO₃

(0.50 g, 3.6 mmol), LiBr (10 mg, cat.) and [18]crown-6 (~10 mg, cat.), was

heated at 758C for 20 h. After cooling down to room temperature, the re-

action mixture was filtered and the residue washed thoroughly with

MeCN (20 mL). The combined organic phase filtrate was concentrated in

vacuo and the yellow residue was purified by column chromatography

(deactivated SiO₂: CH₂Cl₂/Me₂CO 97:3). The yellow band (R_f=0.2) was

collected and the solvent evaporated to give compound 10 (0.38 g, 70%)

as

a

yellow foam. ¹H NMR (300 MHz, CD₃SOCD₃): d=1.17 (t,

³J(H,H)=7.6 Hz, 3H), 1.26 (s, 18H), 2.53 (q, ³J(H,H)=7.6 Hz, 2H),

3.39–3.49 (m, 4H), 3.62–3.71 (m, 6H), 3.99–4.02 (m, 6H), 4.59 (t, ³J-

(H,H)=5.2 Hz, 1H), 6.80–6.84 (m, 6H), 7.01–7.12 (m, 10H), 7.28–

7.31 ppm (m, 4H); MS(EI): m/z (%): 917 (55) [M⁺], 105 (100); elemen-

tal analysis calcd (%) for C₅₃H₆₀N₂O₄S₄: C 69.39, H 3.05, N 6.59, S 13.98;

found: C 69.63, H 2.91, N 6.63, S 13.73.

Compound 11: A solution of compound 10 (0.38 g, 0.41 mmol), TsCl

(0.16 g, 0.82 mmol), Et₃N (0.5 mL, 0.35 g, 3.3 mmol), and DMAP

(~10 mg, cat.) in anhydrous CH₂Cl₂(150 mL) was stirred at room tem-

perature for 20 h. Al₂O₃(10 g, Brockmann 1, neutral) was added, where-

upon the solvent was removed and the residue was purified by column

chromatography (deactivated SiO₂: CH₂Cl₂/Me₂CO 99:1). The yellow

band (R_f=0.15) was collected and the solvent evaporated to give com-

Compound 6: Compound 4 (0.80 g, 2.83 mmol) was dissolved in anhy-

drous DMF (30 mL), cooled to 08C, and degassed (N₂, 10 min) before

the iodide 5 (2.50 g, 8.33 mmol) followed by NaH (0.80 g of a 60% sus-

pension in mineral oil, 20.0 mmol) was added to the yellow solution. The

reaction mixture was stirred for 3 h at 08C, whereupon the reaction mix-

ture was diluted with CH₂Cl₂(500 mL), washed with brine (10150 mL),

and dried (MgSO₄). Removal of the solvent gave a brown oil that was

purified by column chromatography (deactivated SiO₂: CH₂Cl₂/MeOH

19:1). The broad yellow band (R_f=0.6) was collected and concentrated,

pound 11 (0.35 g, 81%) as

a

yellow foam. ¹H NMR (300 MHz,

CD₃COCD₃): d=1.18 (t, ³J(H,H)=7.5 Hz, 3H), 1.29 (s, 18H), 2.43 (s,

3H), 2.60 (q, ³J(H,H)=7.5 Hz, 2H), 3.61–3.66 (m, 4H), 3.75–3.79 (m,

4H), 3.98–4.16 (m, 8H), 6.64 (s, 2H), 6.73 (s, 2H), 6.81 (d, ³J(H,H)=

9.0 Hz, 2H), 7.09–7.15 (m, 10H), 7.28–7.32 (m, 4H), 7.45 (d, ³J(H,H)=

8.4 Hz, 2H), 7.78 ppm (d, ³J(H,H)=8.4 Hz, 2H); MS(FT-MALDI): m/z

(%): 1093 (2) [M⁺+Na], 1070 (100) [M⁺], 921 (15); elemental analysis

calcd (%) for C₆₀H₆₆N₂O₆S₅: C 67.26, H 6.21, N 2.61, S 14.96; found: C

65.78, H 6.24, N 2.36, S 14.91.

1

affording compound 6 (1.19 g, 67%) as a yellow oil. H NMR (300 MHz,

CD₃SOCD₃): d=1.40–1.80 (m, 12H), 3.40–4.00 (m, 20H), 4.54 (brs, 2H),

6.82 ppm (s, 4H); ¹³C NMR (75 MHz, CD₃SOCD₃): d=19.6, 25.6, 30.8,

50.3, 61.7, 66.4, 70.1, 70.7, 98.5, 114.2, 117.2, 119.5 ppm; MS(EI): m/z

(%): 626 (24) [M⁺], 542 (18), 458 (10); elemental analysis calcd (%) for

C₂₈H₃₈N₂O₆S₄: C 53.65, H 6.11, N 4.47; found: C 53.78, H 6.09, N 4.43.

Compound 7: A solution of compound 6 (1.14 g, 1.82 mmol) in THF/

EtOH (50 mL, 1:1 v/v) was degassed (N₂, 10 min) before TsOH·H₂O

(~10 mg, cat.) was added. The yellow solution was stirred for 20 h at

room temperature, whereupon it was diluted with CH₂Cl₂(100 mL). The

combined organic phase was washed with a saturated aqueous NaHCO₃

solution (200 mL), H₂O (300 mL) and dried (MgSO₄). Concentration in

vacuo gave a yellow powder, which was subjected to column chromatog-

raphy (deactivated SiO₂: CH₂Cl₂/MeOH 24:1). The greenish yellow band

(R_f=0.3) was collected and the solvent evaporated to give compound 7

(0.56 g, 67%) as a yellow powder. M.p. 138–1398C; ¹H NMR (300 MHz,

CD₃SOCD₃): d=3.39–3.42 (m, 8H), 3.64 (t, ³J(H,H)=5.2 Hz, 4H), 4.00

(t, ³J(H,H)=5.2 Hz, 4H), 4.59 (t, ³J(H,H)=5.2 Hz, 2H), 6.82 ppm (s,

4H); ¹³C NMR (75 MHz, CD₃SOCD₃): d=49.8, 60.2, 70.2, 72.2, 113.7,

116.7, 118.9 ppm; MS(MALDI-TOF): m/z (%): 458 (100) [M⁺]; elemen-

tal analysis calcd (%) for C₁₈H₂₂N₂O₄S₄: C 47.14, H 4.83, N 6.11, S 27.97;

found: C 47.04, H 4.83, N 6.08, S 27.73.

Compound 13: A solution of the tosylate 11 (0.64 g, 0.60 mmol) and 12

(0.26 g, 0.79 mmol) in anhydrous MeCN (50 mL) containing K₂CO₃

(0.34 g, 2.4 mmol), LiBr (10 mg, cat.) and [18]crown-6 (~10 mg, cat.) was

heated under reflux for 20 h. After cooling down to room temperature,

the reaction mixture was filtered and the residue washed with MeCN (2

50 mL). The combined organic phase filtrate was concentrated in vacuo

and the yellow oily residue was purified by column chromatography (de-

activated SiO₂: CH₂Cl₂/EtOH 97:3). The yellow band was collected and

the solvent evaporated affording compound 13 (0.44 g, 60%) as a yellow

foam. ¹H NMR (300 MHz, CD₃COCD₃): d=1.20 (t, ³J(H,H)=7.6 Hz,

3H), 1.29 (s, 18H), 1.49–1.53 (m, 6H), 2.60 (q, ³J(H,H)=7.6 Hz, 2H),

3.37–3.48 (m, 1H), 3.54–3.64 (m, 1H), 3.75–4.00 (m, 14H), 4.08–4.13 (m,

6H), 4.29–4.32 (m, 4H), 4.63 (brs, 1H), 6.76 (s, 2H), 6.77 (s, 2H), 6.84

(d, ³J(H,H)=8.9 Hz, 2H), 6.94–6.97 (m, 2H), 7.09–7.15 (m, 10H), 7.30–

7.44 (m, 6H), 7.80 (d, ³J(H,H)=8.4 Hz, 1H), 7.85 ppm (d, ³J(H,H)=

8.4 Hz, 1H); MS(FT-MALDI): m/z (%): 1269 (10) [M⁺+K], 1253 (10)

[M⁺+Na], 1230 (100) [M⁺].

Compound 8: TsCl (0.57 g, 2.99 mmol) dissolved in anhydrous CH₂Cl₂

(30 mL) was added dropwise over 20–30 min to an ice-cooled solution of

the diol 7 (1.30 g, 2.83 mmol), Et₃N (2 mL, 1.5 g, 14 mmol), and 4-di-

methylaminopyridine (DMAP; ~10 mg, cat.) in anhydrous CH₂Cl₂

(90 mL). The reaction mixture was stirred for 20 h (08C to RT), where-

upon Al₂O₃(10 g, Brockmann 1, neutral) was added and the solvent re-

moved. The resulting green powder was directly subjected to column

chromatography (deactivated SiO₂) and the bistosylate (0.90 g, 41%)

was eluted with CH₂Cl₂, whereupon the eluent was changed to

CH₂Cl₂/MeOH (99:1) and the yellow band (R_f=0.5) containing the de-

sired monotosylate was collected and concentrated to give compound 8

(0.38 g, 22%) as a yellow solid. Finally, the starting material 7 (0.45 g,

34%) was eluted CH₂Cl₂/MeOH (23:2). ¹H NMR (300 MHz,

CD₃SOCD₃): d=2.42 (s, 3H), 3.36–3.66 (m, 10H), 3.93–4.11 (m, 6H),

4.59 (t, ³J(H,H)=5.2 Hz, 1H), 6.74 (s, 2H), 6.83 (s, 2H), 7.46 (d,

³J(H,H)=8.0 Hz, 2H), 7.75 ppm (d, ³J(H,H)=8.0 Hz, 2H); ¹³C NM R

(75 MHz, CD₃SOCD₃): d=21.1, 49.5, 49.7, 60.2, 67.7, 69.8, 70.1, 70.2,

72.1, 113.6, 113.7, 116.7, 116.8, 118.9, 127.6, 130.1, 132.5, 144.9 ppm (one

line is missing/overlapping); MS(MALDI-TOF): m/z (%): 612 (100)

[M⁺]; elemental analysis calcd (%) for C₂₅H₂₈N₂O₆S₅: C 49.00, H 4.61, N

4.57, S 26.16; found: C 48.83, H 4.66, N 4.67, S 25.97.

Compound 14: A solution of compound 13 (0.40 g, 0.32 mmol) in THF/

EtOH (40 mL, 1:1 v/v) was degassed (N₂, 10 min) before TsOH·H₂O

(~10 mg, cat.) was added. The yellow solution was stirred for 16 h at

room temperature, whereupon it was diluted with CH₂Cl₂(50 mL). The

combined organic phase was washed with a saturated aqueous NaHCO₃

solution (50 mL) and H₂O (50 mL), and dried (MgSO₄). Concentration in

vacuo gave a yellow oil, which was subjected to column chromatography

(deactivated SiO₂: CH₂Cl₂/EtOAc 1:1). The yellow band (R_f=0.4) was

collected and the solvent evaporated to give compound 14 (0.21 g, 56%)

as

a

yellow foam. ¹H NMR (300 MHz, CD₃COCD₃): d=1.24 (t,

³J(H,H)=7.6 Hz, 3H), 1.33 (s, 18H), 2.64 (q, ³J(H,H)=7.6 Hz, 2H),

3.50–3.75 (m, 5H), 3.82–3.86 (m, 4H), 3.91–3.94 (m, 2H), 3.97–4.03 (m,

4H), 4.08–4.19 (m, 6H), 4.32–4.36 (m, 4H), 6.80 (s, 2H), 6.81 (s, 2H),

6.89 (d, ³J(H,H)=8.9 Hz, 2H), 6.99–7.01 (m, 2H), 7.13–7.20 (m, 10H),

7.33–7.48 (m, 6H), 7.83–7.89 ppm (m, 2H); MS(FT-MALDI): m/z (%):

1185 (5) [M⁺+K], 1169 (20) [M⁺+Na], 1146 (100) [M⁺].

Compound 15: A solution of compound 14 (0.20 g, 0.17 mmol), TsCl

(0.068 g, 0.35 mmol), Et₃N (0.2 mL, 0.14 g, 1.4 mmol), and DMAP

(~10 mg, cat.) in anhydrous CH₂Cl₂(50 mL) was stirred at room temper-

274

www.chemeurj.org

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

ature for 20 h, whereupon the solvent was removed and the yellow solid

was purified by column chromatography (deactivated SiO₂: CH₂Cl₂/

EtOAc 19:1). The yellow band (R_f=0.6) was collected and the solvent

S₅·2H₂O: C 57.25, H 5.40, N 2.74, S 5.23; found: C 57.05, H 5.20, N 2.82,

S 5.04.

2-Cyanoethylthio-5-ethylthio-1,3-dithiole-2-thione (20):

A solution of

evaporated to give compound 15 (0.22 g, 98%) as

a yellow foam.

compound 19 (6.09 g, 20.0 mmol) in anhydrous MeCN (150 mL) was de-

gassed (N₂, 5 min) before a solution of NaOMe (7.6 mL of a 2.75m solu-

tion in MeOH, 20.9 mmol) was added dropwise to the yellow solution by

means of a syringe over a period of 45 min at room temperature. The red

mixture was stirred for 15 min, whereupon EtI (3.9 mL, 7.70 g,

49.5 mmol) was added in one portion and the reaction mixture was stir-

red for 24 h at room temperature. The solvent was evaporated and the re-

sulting red oil was dissolved in CH₂Cl₂(250 mL), washed with H₂O (3

200 mL), and dried (MgSO₄). Removal of the solvent gave a red oil,

which was purified by column chromatography (SiO₂: CH₂Cl₂/cyclohex-

ane 4:1). The second yellow band (R_f=0.35) was collected and concen-

trated, affording a yellow oil, which was repeatedly dissolved in CH₂Cl₂

(250 mL) and concentrated to give compound 20 (5.14 g, 92%) as a red

oil, which solidified upon standing to give a yellow solid. M.p. 49.5–

50.58C; ¹H NMR (300 MHz, CDCl₃): d=1.36 (t, ³J(H,H)=7.4 Hz, 3H),

2.74 (t, ³J(H,H)=7.1 Hz, 2H), 2.95 (q, ³J(H,H)=7.4 Hz, 2H), 3.08 ppm

(t, ³J(H,H)=7.1 Hz, 2H); ¹³C NMR (75 MHz, CDCl₃): d=14.8, 18.7,

30.8, 31.8, 117.1, 129.3, 142.4, 210.2 ppm; IR (KBr): n˜ =2247 cm^À1(CꢁN);

MS(EI): m/z (%): 279 (100) [M⁺], 88 (84); elemental analysis calcd (%)

for C₈H₉NS₅: C 34.38, H 3.25, N 5.01, S 57.36; found: C 34.60, H 3.22, N

5.07, S 57.48.

¹H NMR (300 MHz, CD₃COCD₃): d=1.20 (t, J(H,H)=7.5 Hz, 3H), 1.33

(s, 18H), 2.36 (s, 3H), 2.60 (q, ³J(H,H)=7.5 Hz, 2H), 3.77–3.83 (m, 6H),

3.86–3.96 (m, 6H), 4.07–4.15 (m, 6H), 4.20–4.25 (m, 4H), 4.29–4.32 (m,

2H), 6.75 (s, 2H), 6.77 (s, 2H), 6.83 (d, ³J(H,H)=9.0 Hz, 2H), 6.94–6.97

(m, 2H), 7.09–7.15 (m, 10H), 7.30–7.35 (m, 7H), 7.39 (t, ³J(H,H)=

8.5 Hz, 1H), 7.77–7.83 ppm (m, 4H); MS(FT-MALDI): m/z (%): 1300

(100) [M⁺].

3

Compound 16: The tosylate 15 (0.22 g, 0.17 mmol) was dissolved in anhy-

drous Me₂CO (50 mL) and KSCN (0.49 g, 5.04 mmol) was added in one

portion. The yellow reaction mixture was heated under reflux for 1 d,

whereupon additional KSCN (0.49 g, 5.04 mmol) was added. The reaction

mixture was heated under reflux for further 1 d before being cooled to

room temperature. After removal of the solvent, the yellow residue was

dissolved in CH₂Cl₂(100 mL), washed with H₂O (250 mL) and dried

(MgSO₄). Concentration in vacuo gave 0.20 g (97%) of the compound 16

as

a

yellow foam. ¹H NMR (500 MHz, CD₃COCD₃): d=1.20 (t,

³J(H,H)=7.6 Hz, 3H), 1.29 (s, 18H), 2.60 (q, ³J(H,H)=7.6 Hz, 2H), 3.37

(t, ³J(H,H)=5.7 Hz, 2H), 3.78–3.81 (m, 4H), 3.89 (t, ³J(H,H)=4.6 Hz,

2H), 3.93–3.95 (m, 2H), 3.99 (t, ³J(H,H)=4.6 Hz, 2H), 4.04–4.08 (m,

2H), 4.09–4.13 (m, 6H), 4.29–4.31 (m, 2H), 4.33–4.35 (m, 2H), 6.76 (s,

2H), 6.77 (s, 2H), 6.84 (d, ³J(H,H)=8.8 Hz, 2H), 6.95–6.98 (m, 2H),

7.10–7.15 (m, 10H), 7.30–7.32 (m, 4H), 7.37 (t, ³J(H,H)=8.5 Hz, 1H),

7.42 (t, ³J(H,H)=8.5 Hz, 1H), 7.81 (d, ³J(H,H)=8.5 Hz, 1H), 7.85 (d,

2-[4-(2-Cyanoethylthio)-5-ethylthio-1,3-dithiole-2-yliden]-5-tosyl-1,3-

dithiolo[4,5-c]-pyrrole (22): Ketone 21 (1.87 g, 6.01 mmol) and thione 20

(1.68 g, 6.01 mmol) were suspended in distilled (EtO)₃P (50 mL) and

heated to 1358C (during heating the two solids dissolved leaving a red so-

lution and after 10–15 min a yellow orange precipitate was formed). Two

additional portions of 20 (each containing 0.84 g, 3.01 mmol) were added

after 15 and 30 min, respectively. The red reaction mixture was stirred for

another 3 h at 1358C, cooled to room temperature and addition of

MeOH (150 mL) yielded a yellow solid, which was filtered and washed

with MeOH (350 mL). The yellow solid was subjected to column chro-

matography (SiO₂: CH₂Cl₂) and the yellow band (R_f=0.4) was collected

and the solvent evaporated to give a yellow solid, which was dissolved in

CH₂Cl₂/MeOH (1:1 v/v, 500 mL) and concentrated to approximately half

of its volume to precipitate the product. The yellow crystals were collect-

ed by filtration, washed with MeOH (50 mL), and dried in vacuo to give

³J(H,H)=8.5 Hz, 1H); IR (KBr): n˜ =2154 cm^À1(S CꢁN); MS(FT-

À

MALDI): m/z (%): 1226 (15) [M⁺+K], 1210 (15) [M⁺+Na], 1187 (100)

[M⁺].

Dumbbell 1: Compound 16 (0.19 g, 0.16 mmol) and the chloride 17

(0.14 g, 0.18 mmol) were dissolved in anhydrous THF/EtOH (2:1 v/v,

50 mL), after which powdered NaBH₄(0.060 g, 1.6 mmol) was added in

one portion. The reaction mixture was stirred for 1 d at room tempera-

ture, whereupon additional NaBH₄(0.060 g, 1.6 mmol) was added and

the reaction mixture was stirred for further 3 d at room temperature.

Thereafter, it was poured into an ice-cooled saturated aqueous NH₄Cl so-

lution (50 mL) and extracted with CH₂Cl₂(250 mL). The combined or-

ganic extracts were dried (MgSO₄) and concentration in vacuo gave a

yellow oil, which was purified by column chromatography (deactivated

SiO₂: CH₂Cl₂/EtOAc 3:2). The yellow band (R_f=0.4) was collected and

the solvent evaporated affording compound 1 (0.21 g, 68%) as a yellow

foam. ¹H NMR (500 MHz, CD₃COCD₃): d=1.20 (t, ³J(H,H)=7.6 Hz,

3H), 1.29 (s, 18H), 2.60 (m, 4H), 3.29 (s, 9H), 3.48–3.50 (m, 6H), 3.62–

3.64 (m, 6H), 3.75–3.82 (m, 14H), 3.84–3.86 (m, 2H), 3.92–3.94 (m, 4H),

4.07–4.12 (m, 12H), 4.24–4.27 (m, 2H), 4.30–4.32 (m ,2H), 4.87 (s, 2H),

4.96 (s, 4H), 6.74 (s, 2H), 6.74 (s, 2H), 6.75 (s, 2H), 6.80–6.98 (m, 10H),

7.10–7.15 (m, 10H), 7.28–7.41 (m, 12H), 7.79 (d, ³J(H,H)=8.6 Hz, 1H),

7.83 ppm (d, ³J(H,H)=8.6 Hz, 1H); MS(MALDI-TOF): m/z (%): 1925

(100) [M⁺]; elemental analysis calcd (%) for C₁₁₀H₁₂₈N₂O₁₈S₅: C 68.58, H

6.70, N 1.45; found: C 68.41, H 6.75, N 1.29.

1

compound 22 (2.40 g, 74%) as yellow needles. M.p. 200–2018C; H NM R

(300 MHz, CD₃SOCD₃): d=1.25 (t, ³J(H,H)=7.3 Hz, 3H), 2.38 (s, 3H),

2.84 (t, ³J(H,H)=6.6 Hz, 2H), 2.89 (q, ³J(H,H)=7.3 Hz, 2H), 3.11 (t,

³J(H,H)=6.6 Hz, 2H), 7.39 (s, 2H), 7.46 (d, ³J(H,H)=8.3 Hz, 2H),

7.82 ppm (d, ³J(H,H)=8.3 Hz, 2H); ¹³C NMR (75 MHz, CD₃SOCD₃):

d=15.0, 18.1, 21.1, 29.9, 30.9, 112.3, 112.8 (2 signals), 117.8, 118.9, 124.0,

125.9, 126.0, 126.8, 129.8, 130.4, 134.4, 145.9 ppm; IR (KBr): n˜ =

2250 cm^À1(CꢁN); MS(EI): m/z (%): 542 (11) [M⁺], 387 (28) [M⁺ÀTs],

184 (55), 105 (100), 91 (65); elemental analysis calcd (%) for

C₂₀H₁₈N₂O₂S₇: C 44.25, H 3.34, N 5.16, S 41.35; found: C 44.40, H 3.34, N

5.23, S 41.42.

Compound 23: A solution of the iodide 18 (0.95 g, 1.05 mmol) and 22

[2]Rotaxane RBPTTF·4PF₆:

A

solution of the dumbbell

1

(0.20 g,

(0.55 g, 1.01 mmol) in anhydrous THF (70 mL) was degassed (N₂, 10 min)

0.10 mmol), 2·2PF₆(0.22 g, 0.31 mmol), and the dibromide 3 (0.082 g,

0.31 mmol) in anhydrous DMF (8 mL) was transferred to a teflon tube

and subjected to 10 kbar of pressure at room temperature for 3 d. The

greenish-brown solution was directly subjected to column chromatogra-

phy (deactivated SiO₂) and unreacted dumbbell was eluted with Me₂CO,

whereupon the eluent was changed to Me₂CO/NH₄PF₆(1.0 g NH₄PF₆in

100 mL Me₂CO) and the greenish brown band was collected. Most of the

solvent was removed in vacuo (T<308C), followed by addition of H₂O

(100 mL). The resulting precipitate was collected by filtration, washed

with H₂O (220 mL) and Et₂O (230 mL) and dried in vacuo over P₂O₅,

affording compound RBPTTF·4PF₆(0.15 g, 47%) as a brown solid. The

data given below are for the 1:1 mixture of the two translational isomers;

¹⁹F NMR (376 MHz, CD₃COCD₃): d=À72.4 ppm (d, J=709 Hz);

³¹P NMR (161 MHz, CD₃COCD₃): d=À144.2 (septet, J=709 Hz);

MS(ES): m/z (%): 1369 (15) [M²⁺À2PF₆], 864 (80) [M³⁺À3PF₆], 612

(100) [M⁴⁺À4PF₆]; elemental analysis calcd (%) for C₁₄₆H₁₆₀F₂₄N₆O₁₈P₄

before a solution of CsOH·H₂O (0.174 g, 1.04 mmol) in anhydrous

MeOH (5.0 mL) was added dropwise by means of a syringe over a

period of 75 min at room temperature. Subsequently, the reaction mix-

ture was stirred for 2 d at room temperature, whereupon the yellow reac-

tion mixture was diluted with CH₂Cl₂(150 mL), washed with brine

(150 mL), H₂O (2150 mL), and dried (MgSO₄). Removal of the solvent

gave a yellow foam, which was purified by column chromatography

(SiO₂: CH₂Cl₂/cyclohexane 9:1). The broad yellow band (R_f=0.35) was

collected and concentrated, affording a yellow foam, which was repeated-

ly dissolved in CH₂Cl₂(230 mL) and concentrated to give compound 23

(0.99 g, 77%) as a yellow foam. ¹H NMR (300 MHz, CD₃COCD₃): d=

1.20 (t, ³J(H,H)=7.6 Hz, 3H), 1.23 (t, ³J(H,H)=7.4 Hz, 3H), 1.29 (s,

18H), 2.38 (s, 3H), 2.61 (q, ³J(H,H)=7.6 Hz, 2H), 2.84 (q, ³J(H,H)=

7.4 Hz, 2H), 3.10 (t, ³J(H,H)=6.3 Hz, 2H), 3.84 (t, ³J(H,H)=6.3 Hz,

2H), 3.93–3.99 (m, 4H), 4.02–4.06 (m, 2H), 4.14–4.19 (m, 2H), 4.27–4.33

3

(m, 4H), 6.84 (d, J(H,H)=9.0 Hz, 2H), 6.89–6.97 (m, 2H), 7.06–7.14 (m,

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

275

J. F. Stoddart, J. R. Heath, J. O. Jeppesen et al.

10H), 7.24 and 7.27 (AB q, J=2.1 Hz, 2H), 7.28–7.43 (m, 8H), 7.80–

m/z calcd for C₆₇H₇₆INO₆S₆⁺: 1309.3036; found: 1309.3035; elemental

analysis calcd (%) for C₆₇H₇₆INO₆S₆: C 61.40, H 5.84, N 1.07, S 14.68;

found: C 61.78, H 5.83, N 1.11, S 14.50.

7.85 ppm (m, 4H); MS(MALDI-TOF): m/z (%): 1265 (22) [M⁺], 1111

+

(100) [M⁺+HÀTs]; HiRes-FT-MALDI-MS: m/z calcd for C₇₀H₇₅NO₇S₇

:

1265.3583; found: 1265.3580; elemental analysis calcd (%) for

C₇₀H₇₅NO₇S₇: C 66.37, H 5.97, N 1.11, S 17.72; found: C 65.88, H 5.94, N

1.30, S 17.75.

Compound 28: The iodide 27 (0.48 g, 0.37 mmol) was dissolved in anhy-

drous Me₂CO (50 mL) and KSCN (1.78 g, 18.3 mmol) was added in one

portion. The yellow reaction mixture was heated under reflux for 3 d,

whereupon the reaction mixture was cooled to room temperature. After

removal of the solvent, the yellow residue was dissolved in CH₂Cl₂

(100 mL), washed with H₂O (375 mL), and dried (MgSO₄). Concentra-

Compound 24: Compound 23 (0.85 g, 0.67 mmol) was dissolved in anhy-

drous THF/MeOH (1:1 v/v, 70 mL) and degassed (N₂, 10 min) before

NaOMe (25% solution in MeOH, 2.3 mL, 0.54 g, 10.1 mmol) was added

in one portion. The yellow solution was heated under reflux for 15 min

before being cooled to room temperature, whereupon the solvent was

evaporated. The yellow residue was dissolved in CH₂Cl₂(100 mL),

washed with H₂O (3100 mL) and dried (MgSO₄). Concentration gave a

yellow foam, which was subjected to column chromatography (SiO₂:

CH₂Cl₂). The yellow band (R_f=0.5) was collected and concentrated to

tion in vacuo gave compound 28 (0.45 g, 99%) as

a yellow foam.

¹H NMR (300 MHz, CD₃COCD₃): d=1.21 (t, J(H,H)=7.6 Hz, 3H), 1.26

(t, ³J(H,H)=7.3 Hz, 3H), 1.30 (s, 18H), 2.61 (q, ³J(H,H)=7.6 Hz, 2H),

2.87 (q, ³J(H,H)=7.3 Hz, 2H), 3.11 (t, ³J(H,H)=6.4 Hz, 2H), 3.28 (t,

³J(H,H)=5.7 Hz, 2H), 3.76–3.81 (m, 4H), 3.86 (t, ³J(H,H)=6.4 Hz, 2H),

3.94–4.01 (m, 4H), 4.03–4.06 (m, 2H), 4.11–4.14 (m, 2H), 4.16–4.19 (m,

2H), 4.29–4.34 (m, 4H), 6.77 and 6.80 (AB q, J=2.1 Hz, 2H), 6.85 (d,

³J(H,H)=9.0 Hz, 2H), 6.93 (d, ³J(H,H)=8.0 Hz, 1H), 6.96 (d, ³J(H,H)=

8.0 Hz, 1H), 7.06–7.14 (m, 10H), 7.26–7.38 (m, 6H), 7.83 (d, ³J(H,H)=

3

provide compound 24 (0.64 g, 87%) as

a

yellow foam. ¹H NM R

(300 MHz, CD₃COCD₃): d=1.21 (t, ³J(H,H)=7.6 Hz, 3H), 1.26 (t,

³J(H,H)=7.3 Hz, 3H), 1.30 (s, 18H), 2.61 (q, ³J(H,H)=7.6 Hz, 2H), 2.87

(q, ³J(H,H)=7.3 Hz, 2H), 3.11 (t, ³J(H,H)=6.4 Hz, 2H), 3.86 (t,

³J(H,H)=6.4 Hz, 2H), 3.95–4.01 (m, 4H), 4.02–4.05 (m, 2H), 4.16–4.19

(m, 2H), 4.29–4.33 (m, 4H), 6.79 and 6.80 (AB q, J=1.9 Hz, 2H), 6.85

(d, ³J(H,H)=9.0 Hz, 2H), 6.93 (d, ³J(H,H)=8.0 Hz, 1H), 6.96 (d,

³J(H,H)=8.0 Hz, 1H), 7.06–7.14 (m, 10H), 7.26–7.38 (m, 6H), 7.83 (d,

³J(H,H)=8.0 Hz, 1H), 7.86 (d, ³J(H,H)=8.0 Hz, 1H), 10.36 ppm (brs,

1H); MS(MALDI-TOF): m/z (%): 1112 (100) [M⁺]; HiRes-FT-MALDI-

MS: m/z calcd for C₆₃H₆₉NO₅S₆⁺: 1111.3495; found: 1111.3452; elemental

analysis calcd (%) for C₆₃H₆₉NO₅S₆: C 68.01, H 6.25, N 1.26, S 17.29;

found: C 67.74, H 6.36, N 1.28, S 17.06.

8.0 Hz, 1H), 7.86 ppm (d, ³J(H,H)=8.0 Hz, 1H); IR (KBr): n˜ =

2154 cm^À1(S CꢁN); MS(MALDI-TOF): m/z (%): 1241 (100) [M ];

+

À

HiRes-FT-MALDI-MS: m/z calcd for C₆₈H₇₆N₂O₆S₇⁺: 1240.3743; found:

1240.3743; elemental analysis calcd (%) for C₆₈H₇₆N₂O₆S₇: C 65.77, H

6.17, N 2.26, S 18.08; found: C 65.87, H 6.31, N 2.28, S 17.83.

Dumbbell 29: The chloride 17 (0.19 g, 0.24 mmol) and compound 28

(0.25 g, 0.20 mmol) were dissolved in anhydrous THF/EtOH (2:1 v/v,

50 mL), after which powdered NaBH₄(0.15 g, 3.97 mmol) was added in

one portion. The reaction mixture was stirred for 2 d at room tempera-

ture, whereupon it was poured into a saturated aqueous NH₄Cl solution

(50 mL), and extracted with CH₂Cl₂(275 mL). The combined organic

extracts were washed with brine (100 mL) and dried (MgSO₄). Concen-

tration in vacuo gave a yellow oil, which was purified by column chroma-

tography (SiO₂: CH₂Cl₂/EtOAc 2:1). The yellow band (R_f=0.5) was col-

lected and the solvent evaporated affording a yellow oil, which was re-

peatedly dissolved in CH₂Cl₂(225 mL) and concentrated to give com-

Compound 26: Compounds 24 (0.61 g, 0.55 mmol) and 25 (0.25 g,

0.90 mmol) were dissolved in anhydrous DMF (20 mL) and degassed (N₂,

10 min) before NaH (0.055 g of

a 60% suspension in mineral oil,

1.38 mmol) was added. The reaction mixture was stirred for 3.5 h at

room temperature, causing the initially yellow solution to become more

orange. Brine (80 mL) was added dropwise until no more gas evolution

was observed and the resulting yellow precipitate was filtered, washed

with H₂O (20 mL), and dried. The crude product was purified by column

chromatography (SiO₂: CH₂Cl₂). The yellow band (R_f=0.5) was collected

and the solvent evaporated, providing compound 26 (0.50 g, 75%) as a

yellow foam. ¹H NMR (300 MHz, CD₃COCD₃): d=1.21 (t, ³J(H,H)=

7.6 Hz, 3H), 1.26 (t, ³J(H,H)=7.3 Hz, 3H), 1.30 (s, 18H), 2.61 (q,

³J(H,H)=7.6 Hz, 2H), 2.86 (q, ³J(H,H)=7.3 Hz, 2H), 3.11 (t, ³J(H,H)=

6.4 Hz, 2H), 3.61–3.73 (m, 4H), 3.74–3.78 (m, 2H), 3.85 (t, ³J(H,H)=

6.4 Hz, 2H), 3.94–4.00 (m, 4H), 4.02–4.05 (m, 2H), 4.08–4.12 (m, 2H),

4.15–4.19 (m, 2H), 4.28–4.34 (m, 4H), 6.76 and 6.79 (AB q, J=2.0 Hz,

2H), 6.85 (d, ³J(H,H)=9.0 Hz, 2H), 6.93 (d, ³J(H,H)=8.0 Hz, 1H), 6.96

(d, ³J(H,H)=8.0 Hz, 1H), 7.06–7.13 (m, 10H), 7.26–7.38 (m, 6H), 7.83

(d, ³J(H,H)=8.0 Hz, 1H), 7.86 ppm (d, ³J(H,H)=8.0 Hz, 1H);

MS(MALDI-TOF): m/z (%): 1217 (100) [M⁺]; HiRes-FT-MALDI-MS:

m/z calcd for C₆₇H₇₆ClNO₆S₆⁺: 1217.3680; found: 1217.3675; elemental

analysis calcd (%) for C₆₇H₇₆ClNO₆S₆: C 66.01, H 6.28, N 1.15, S 15.78;

found: C 66.14, H 6.30, N 1.20, S 15.61.

pound 29 (0.31 g, 78%) as

a

yellow foam. ¹H NMR (300 MHz,

3

CD₃COCD₃): d=1.21 (t, J(H,H)=7.6 Hz, 3H), 1.24 (t, J(H,H)=7.3 Hz,

3H), 1.30 (s, 18H), 2.52 (t, ³J(H,H)=6.4 Hz, 2H), 2.61 (q, ³J(H,H)=

7.6 Hz, 2H), 2.84 (q, ³J(H,H)=7.3 Hz, 2H), 3.08 (t, ³J(H,H)=6.4 Hz,

2H), 3.29 (s, 9H), 3.49 (t, ³J(H,H)=6.4 Hz, 2H), 3.48–3.51 (m, 6H),

3.62–3.67 (m, 10H), 3.77–3.82 (m, 6H), 3.83 (t, ³J(H,H)=6.4 Hz, 2H),

3.93–3.98 (m, 4H), 4.01–4.18 (m, 12H), 4.26–4.32 (m, 4H), 4.91 (s, 2H),

5.03 (s, 4H), 6.73 (s, 2H), 6.76 and 6.78 (AB q, J=2.0 Hz, 2H), 6.82 (d,

³J(H,H)=8.5 Hz, 2H), 6.85 (d, ³J(H,H)=9.1 Hz, 2H), 6.92 (d, ³J(H,H)=

8.0 Hz, 1H), 6.94 (d, ³J(H,H)=8.0 Hz, 1H), 6.95 (d, ³J(H,H)=8.6 Hz,

4H), 7.07–7.14 (m, 10H), 7.26–7.35 (m, 8H), 7.39 (d, ³J(H,H)=8.6 Hz,

4H), 7.83 (d, ³J(H,H)=8.0 Hz, 1H), 7.86 ppm (d, ³J(H,H)=8.0 Hz, 1H);

MS(MALDI-TOF): m/z (%): 1976 (100) [M⁺], 1767 (12); HiRes-FT-

+

MALDI-MS: m/z calcd for

C₁₁₀H₁₃₁NO₁₈S₇

:

1977.7406; found:

1977.7405; elemental analysis calcd (%) for C₁₁₀H₁₃₁NO₁₈S₇: C 66.74, H

6.67, N 0.71, S 11.34; found: C 66.64, H 6.45, N 0.77, S 11.15.

[2]Rotaxane RBLOCK·4PF₆: A solution of the dumbbell 29 (0.25 g,

0.13 mmol), 2·2PF₆(0.27 g, 0.38 mmol), and the dibromide 3 (0.10 g,

0.38 mmol) in anhydrous DMF (12 mL) was transferred to a teflon-tube

and subjected to 10 kbar of pressure at room temperature for 3 d. The

red solution was directly subjected to column chromatography (SiO₂) and

unreacted dumbbell was eluted with Me₂CO, whereupon the eluent was

changed to Me₂CO/NH₄PF₆(1.0 g NH₄PF₆in 100 mL Me₂CO) and the

red band was collected. Most of the solvent was removed in vacuo (T

<308C), followed by addition of H₂O (100 mL). The resulting precipitate

was collected by filtration, washed with H₂O (220 mL) and Et₂O (2

30 mL) and dried in vacuo over P₂O₅, affording 0.16 g (41%) of the com-

pound RBLOCK·4PF₆as a red solid. Data for RBLOCK·4PF₆: mp

Compound 27: The chloride 26 (0.46 g, 0.38 mmol) was dissolved in anhy-

drous Me₂CO (60 mL) and NaI (3.42 g, 22.8 mmol) was added in one por-

tion. The reaction mixture was heated under reflux for 6 d, before being

cooled to room temperature and the solvent removed in vacuo. The

yellow residue was dissolved in CH₂Cl₂(75 mL) and washed with H₂O

(350 mL), before being dried (MgSO₄). Concentration in vacuo gave

compound 27 (0.49 g, 99%) as a yellow foam. ¹H NMR (300 MHz,

3

CD₃COCD₃): d=1.21 (t, J(H,H)=7.6 Hz, 3H), 1.26 (t, J(H,H)=7.3 Hz,

3H), 1.30 (s, 18H), 2.61 (q, ³J(H,H)=7.6 Hz, 2H), 2.86 (q, ³J(H,H)=

7.3 Hz, 2H), 3.11 (t, ³J(H,H)=6.4 Hz, 2H), 3.30 (t, ³J(H,H)=6.5 Hz,

3

2H), 3.69 (t, J(H,H)=6.5 Hz, 2H), 3.74–3.78 (m, 2H), 3.85 (t, J(H,H)=

6.4 Hz, 2H), 3.95–4.01 (m, 4H), 4.02–4.06 (m, 2H), 4.08–4.12 (m, 2H),

4.16–4.19 (m, 2H), 4.29–4.34 (m, 4H), 6.77 and 6.80 (AB q, J=2.1 Hz,

2H), 6.85 (d, ³J(H,H)=9.0 Hz, 2H), 6.93 (d, ³J(H,H)=8.0 Hz, 1H), 6.96

(d, ³J(H,H)=8.0 Hz, 1H), 7.06–7.14 (m, 10H), 7.26–7.38 (m, 6H), 7.83

(d, ³J(H,H)=8.0 Hz, 1H), 7.86 ppm (d, ³J(H,H)=8.0 Hz, 1H);

MS(MALDI-TOF): m/z (%): 1309 (100) [M⁺]; HiRes-FT-MALDI-MS:

1

1708C (decomposed without melting); H NMR (400 MHz, CD₃COCD₃):

3

d=1.18 (t, J(H,H)=7.6 Hz, 3H), 1.27 (s, 18H), 1.30 (t, J(H,H)=7.3 Hz,

3H), 2.56 (t, ³J(H,H)=6.2 Hz, 2H), 2.58 (q, ³J(H,H)=7.6 Hz, 2H), 2.72

(d, ³J(H,H)=8.0 Hz, 1H), 2.74 (d, ³J(H,H)=8.0 Hz, 1H), 2.93 (q,

³J(H,H)=7.3 Hz, 2H), 3.29 (s, 9H), 3.41–3.46 (m, 2H), 3.48–3.52 (m,

6H), 3.55 (t, ³J(H,H)=6.5 Hz, 2H), 3.63–3.69 (m, 8H), 3.71 (s, 2H),

276

www.chemeurj.org

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

3.74–3.82 (m, 6H), 4.02–4.22 (m, 10H), 4.34–4.45 (m, 6H), 4.48–4.62 (m,

6H), 4.90 (s, 2H), 5.04 (s, 4H), 6.01–6.14 (brm, 8H), 6.15 (t, ³J(H,H)=

8.0 Hz, 1H), 6.24 (t, ³J(H,H)=8.0 Hz, 1H), 6.44 (d, ³J(H,H)=8.0 Hz,

1H), 6.45 (d, ³J(H,H)=8.0 Hz, 1H), 6.66 and 6.72 (AB q, J=2.2 Hz,

2H), 6.78 (s, 2H), 6.80–6.96 (m, 8H), 7.03–7.10 (m, 10H), 7.26–7.34 (m,

6H), 7.40 (d, ³J(H,H)=8.4 Hz, 4H), 7.50–7.90 (brm, 8H), 8.10–8.50

(brm, 8H), 9.05–9.45 ppm (brm, 8H); UV/Vis (MeCN, 298 K): l_max(e)=

540 nm (920 mol^À1dm^À3cm^À1); MS (MALDI-TOF): m/z (%): 2644 (8)

[M⁺À3PF₆], 2499 (8) [M⁺À4PF₆], 1977 (2), 665 (16) [CBPQT+PF₆⁺],

561 (100); elemental analysis calcd (%) for C₁₄₆H₁₆₃F₂₄N₅O₁₈P₄S₇·2H₂O:

C 56.27, H 5.40, N 2.25, S 7.20; found: C 56.23, H 5.32, N 2.46, S 7.50.

For the MSTJ devices, x_eqcannot be obtained on account of the fact that

N_MSCC/N_Totalcannot be measured directly. As a result, the kinetic data for

the MSTJ were acquired based on the assumption that the current associ-

ated with the pure GSCC (I_GSCC) is much smaller than the current associ-

ated with the pure switch-closed state (I_CLOSED). Given this assumption,

the following approximate equality holds N_MSCC/N_TotalꢂI_OPEN/I_CLOSED, as

elaborated in the next section. The assumption I_GSCC!I_CLOSEDhas been

verified by the theoretical studies.^[36]

Derivation of switching amplitudes in MSTJ devices: Based on the re-

fined hypothesis that, the high-conductance (switch-closed) state of an

MSTJ corresponds to the MSCC, but that the low-conductance (switch-

open) state is related to the MSCC/GSCC ratio at equilibrium, the mea-

sured current (I) can be defined in terms of the intrinsic conductance

properties of each co-conformation and the percentage of the co-confor-

Data analysis: Exponential decay curves were obtained for each temper-

ature by recording the CV over a range of different scan rates utilizing a

modification of previously used methods.^{[25, 55]}The time interval, t, was

determined as the time between formation of the MSCC—reduction to

the TTF neutral redox state—and its oxidation to the monocationic state,

equivalent to the point of measurement of the proportion of MSCC re-

maining. The proportion of MSCC remaining N_MSCC/N_Totalwas obtained

by integrating the area under the peak of the MSCC in the second

anodic scan, N_MSCC, and then normalized to a single electron (N_Total, 1e^À)

by halving the total area (2e^À) of the first or second anodic scan. The in-

tegrated areas were determined from normalized data. The accuracy of

the normalizing procedure was checked by comparing the integrated area

from each of the anodic and cathodic scans and was found to vary by no

more than Æ5%. The integration regions were selected to include the

contributions to the CV from the diffusional tail that follows the main

peak. Two additional points were added to the decay curve that corre-

spond to time zero, t=0 s, N_MSCC/N_Total=1.0 and, at long relaxation times,

t=300 s, at which the value of N_MSCC/N_Totalwas set equal to the steady-

state value determined by the first anodic scan. The resulting decay curve

was fitted to a single exponential curve in order to obtain the time con-

stant, t, and hence the rate constant, k at T. Errors in the kinetics data

(t, k, and DG^ꢀ) were obtained from the absolute errors determined from

the fit of each decay curve (N_MSCC/N_Totalversus t). Errors in the DH^ꢀand

DS^ꢀparameters were obtained from the errors in the fits to the Eyring

plots. Errors in the equilibrium thermodynamics data were estimated

from the vanꢀt Hoff plots.

mations N_MSCC/N_Totaland N_GSCC/N_Total

.

Consequently, I_OPENcorresponds to a thermal equilibrium condition and

is a mixture of the GSCC and MSCC, whereas I_CLOSEDis 100% of the

MSCC. This model influences the meaning of the ratio I_OPEN/I_CLOSED

.

The conductance properties of these systems can be described as follows:

Firstly, the GSCC and MSCC have intrinsic current values I_GSCCand

I_MSCC, which are constants at a certain temperature T. Therefore, at any

time the current measured (I_t) is a summation of these two contributions.

The magnitude of each contribution is scaled by the proportions of the

GSCC (N_GSCC/N_Total) and MSCC (N_MSCC/N_Total) present in the mixture.

This leads to the following general formula [Eq. (7)] for the current I_t:

I_t¼ ðN_MSCC=N_TotalÞ_tI_MSCCþ ðN_GSCC=N_TotalÞ_tI_GSCC

ð7Þ

Therefore, for the trivial situation when I_CLOSEDis measured at t=0, we

assume that N_GSCC=0 and N_MSCC/N_Total=1 confirming that I_CLOSED=I_MSCC

(see Figure 9).

Derivation of the kinetic parameters: The kinetic data—rate constants

and activation barriers—for the relaxation from the MSCC back to the

equilibrium ratio of MSCC:GSCC were acquired based on the following

equilibrium [Eq. (3)]:

k₁

MSCC

GSCC

H

ð3Þ

G

k₂

The relative concentration of the MSCC as a function of time can be de-

noted as x_t=[MSCC]/[Total], where t is the time. Let x₀=x_t=0, and at

equilibrium, let x_eq=x_t!1. The first derivative of x(t) with respect to time

is given by Equation (4).

Figure 9. Schematic representation of a volatility curve defining I_CLOSED

and I_OPEN

.

d½MSCC=dt ¼ Àk₁x½Total þ k₂ð1ÀxÞ½Total

) dx=dt ¼ Àk₁x þ k₂ð1ÀxÞ

ð4Þ

Now consider what happens at thermal equilibrium (t=1), defined as

I_OPEN[Eq. (8)]

Integration of Equation (4) leads to the following equality [Eq. (5)]

I_OPEN¼ ðN_MSCC=N_Total

Þ

₁I_CLOSEDþ ðN_GSCC=N_Total

Þ

₁I_GSCC

ð8Þ

x_t

t

Z

dx

¼

Àðk₁þ k₂Þdt

ðk₁þ k₂ÞxÀk₂

x₀

0

ð5Þ

Consequently, the ratio I_OPEN/I_CLOSED, which happens to be the inverse of

the switching amplitude, can be expressed as Equation (9).

ðk₁þ k₂Þx_tÀk₂

ðk₁þ k₂Þx₀Àk₂

) ln

¼ Àðk₁þ k₂Þt

I

_OPEN=I_CLOSED¼ ðN_MSCC=N_TotalÞ₁þ ðN_GSCC=N_TotalÞ₁ðI_GSCC=I_CLOSED

Þ

ð9Þ

At equilibrium, 0=dx/dt=Àk₁x + k₂(1Àx))k₂=(k₁x_eq)/(1Àx_eq). Putting

this expression for k₂into Equation (5), we obtain Equation (6).

If the intrinsic conductance of the GSCC is very small compared to

I_CLOSED, the term I_GSCC/I_CLOSEDgoes to zero and therefore we can use

Equation (10).

x_t¼ x_eqþ ðx₀Àx_eqÞexp½Àk₁t=ð1Àx_eqÞ ¼ a þ b expðÀt=tÞ

ð6Þ

By substituting the relationship k₁=(1Àx_eq)/t into the expression DG^ꢀ=

ÀRTln(hk₁/k_BT), DG^ꢀcan be calculated.

I

_OPEN=I_CLOSED¼ N_MSCC=N_Total

ð10Þ

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

277

J. F. Stoddart, J. R. Heath, J. O. Jeppesen et al.

For example, in the case of R(A)TTF⁴⁺, N_MSCC/N_Total=1/10 and assuming

an intrinsic conductance of the GSCC that is 100 times smaller than the

MSCC, I_GSCC/I_Closed=1/100 then the ratio at t=1 is given by Equa-

tion (11)

Mantooth, K. F. Kelly, L. A. Bumm, J. D. Monnell, J. J. Stapleton,

D. W. Price, A. M. Rawlett, D. L. Allara, J. M. Tour, P. S. Weiss, Sci-

ence 2001, 292, 2303–2307.

[8] J. F. Smalley, S. B. Sachs, C. E. Chidsey, S. P. Dudek, H. D. Sikes,

S. E. Creager, C. J. Yu, S. W. Feldberg, M. D. Newton, J. Am. Chem.

Soc. 2004, 126, 14620–14630.

[9] a) P. E. Kornilovitch, A. M. Bratkovsky, R. S. Williams, Phys. Rev. B

2002, 66, 165436, 1–11; b) Y. Chen, G. Y. Jung, D. A. Ohlberg, M.

Li, D. R. Stewart, J. O. Jeppesen, K. A. Nielsen, J. F. Stoddart, R. S.

Williams, Nanotechnology 2003, 14, 462–468; c) C. A. Richter, D. R.

Stewart, D. A. Ohlberg, R. S. Williams, Appl. Phys. A 2005, 80,

1355–1362.

[10] a) A. Aviram, M. A. Ratner, Chem. Phys. Lett. 1974, 29, 277–283;

b) J. Jortner, M. Ratner, Molecular Electronics, Blackwell Science,

Oxford, 1997.

[11] R. L. Carroll, C. B. Gorman, Angew. Chem. 2002, 114, 4556–4579;

Angew. Chem. Int. Ed. 2002, 41, 4378–4400.

I

_OPEN=I_CLOSED¼1=10 þ ð9=10 ꢃ 1=100Þ ¼ 1=10 þ 9=1000 ¼

0:1 þ 0:009 ¼ 0:109

ð11Þ

Consider also how RBPTTF⁴⁺behaves at low temperatures (MSCC/

GSCC=3:4) [Eq. (12)]

I

_OPEN=I_CLOSED¼ 3=7 þ ð4=7 ꢃ 1=100Þ ¼ 0:43 þ 0:006 ¼ 0:436

ð12Þ

In other words, the ratio of N_MSCC/N_Totaldominates the I_OPEN/I_CLOSEDmea-

sured ratio at equilibrium and therefore the switching amplitude in the

condition when the intrinsic conductance of the GSCC is small.

Whereas in the condition when the intrinsic conductance of the GSCC

were higher such as if I_GSCC/I_CLOSED=1/10, then for R(A)TTF⁴⁺and

RBPTTF⁴⁺Equations (13) and (14), respectively, are valid.

[12] G. Maruccio, R. Cingolani, R. Rinaldi, J. Mater. Chem. 2004, 14,

542–554.

[13] G. Ho, J. R. Heath, M. Kondratenko, D. Perepichka, K. Arseneault,

M. Pezolet, M. Bryce, Chem. Eur. J. 2005, 11, 2914–2922.

[14] a) G. J. Ashwell, J. R. Sambles, A. S. Martin, W. G. Parker, J. Chem.

Soc. Chem. Commun. 1990, 1374–1376; b) A. S. Martin, J. R. Sam-

bles, G. J. Ashwell, Phys. Rev. Lett. 1993, 70, 218–221; c) G. J. Ash-

well, D. S. Gandolfo, J. Mater. Chem. 2001, 11, 246–248; d) G. J.

Ashwell, D. S. Gandolfo, J. Mater. Chem. 2002, 12, 411–415; e) G. J.

Ashwell, D. S. Gandolfo, R. Hamilton, J. Mater. Chem. 2002, 12,

416–420; f) G. J. Ashwell, W. D. Tyrrell, A. J. Whittam, J. Am.

Chem. Soc. 2004, 126, 7102–7110.

[15] a) R. M. Metzger, B. Chen, U. Hçpfner, M. V. Lakshmikantham, D.

Vuillaume, T. Kawai, X. L. Wu, H. Tachibana, T. V. Hughes, H. Sa-

kurai, J. W. Baldwin, C. Hosch, M. P. Cava, L. Brehmer, G. J. Ash-

well, J. Am. Chem. Soc. 1997, 119, 10455–10466; b) R. M. Metzger,

Acc. Chem. Res. 1999, 32, 950–957; c) T. Xu, I. R. Peterson, M. V.

Lakshmikantham, R. M. Metzger, Angew. Chem. 2001, 113, 1799–

1802; Angew. Chem. Int. Ed. 2001, 40, 1749–1752; d) R. M. Metzger,

T. Xu, I. R. Peterson, J. Phys. Chem. B 2001, 105, 7280–7290;

e) R. M. Metzger, Chem. Rev. 2003, 103, 3803–3834.

[16] a) M.-K. Ng, L. Yu, Angew. Chem. 2002, 114, 3750–3753; Angew.

Chem. Int. Ed. 2002, 41, 3598–3601; b) P. Yiang, G. M. Morales, W.

You, L. Yu, Angew. Chem. 2004, 116, 4571–4575; Angew. Chem. Int.

Ed. 2004, 43, 4471–4475.

[17] F. C. Krebs, H. Spanggaard, N. Rozlosnik, N. B. Larsen, M. Jørgen-

sen, Langmuir 2003, 19, 7873–7880.

[18] a) S. C. Chang, Z. Y. Li, C. N. Lau, B. Larade, R. S. Williams, Appl.

Phys. Lett. 2003, 83, 3198–3200; b) D. R. Stewart, D. A. Ohlberg,

P. A. Beck, Y. Chen, R. S. Williams, J. O. Jeppesen, K. A. Nielsen,

J. F. Stoddart, Nano Lett. 2004, 4, 133–136; c) N. Lau, D. R. Stewart,

R. S. Williams, M. Bockrath, Nano Lett. 2004, 4, 569–572.

[19] A. N. Shipway, I. Willner, Acc. Chem. Res. 2001, 34, 421–432.

[20] C. P. Collier, G. Mattersteig, E. W. Wong, Y. Luo, K. Beverly, J.

Sampaio, F. M. Raymo, J. F. Stoddart, J. R. Heath, Science 2000, 289,

1172–1175.

I

_OPEN=I_CLOSED¼1=10 þ ð9=10 ꢃ 1=10Þ ¼ 1=10 þ 9=100 ¼

0:1 þ 0:09 ¼ 0:19

ð13Þ

ð14Þ

I

_OPEN=I_CLOSED¼ 3=7 þ ð4=7 ꢃ 1=10Þ ¼ 0:43 þ 0:06 ¼ 0:49

Comparison between the two cases, in which I_GSCCis comparatively

smaller (1%) or larger (10%), leads to switching amplitudes for

R(A)TTF⁴⁺of 9 and 5, respectively, whereas for RBPTTF⁴⁺they corre-

spond to 2.3 and 2.0.

Small intrinsic conductances of the GSCC relative to the MSCC are not

so unlikely and have been calculated^[36]for related TTF-containing bista-

ble catenanes, based on the theory of coherent electron transport, to be

approximately 1/10000.

Acknowledgements

This research was funded by the Office of Naval Research (ONR), the

National Science Foundation (NSF), the Molectronics Program of the

Defense Advanced Research Project Agency (DARPA), the Microelec-

tronics Advance Research Corporation (MARCO) and its Focus Center

on Functional Engineered NanoArchitectonics (FENA) and Materials,

Structures and Devices, and the Center for Nanoscale Innovation for De-

fense (CNID) in the US and by the Danish Natural Science Research

Council (SNF, grants #21-03-0014 and #21-03-0317), the Oticon and

MODECS foundation in Denmark. J.W.C. acknowledges the Samsung

Lee Kun Hee Scholarship Foundation for a graduate fellowship. N.N.P.M.

acknowledges The Netherlands Organization for Scientific Research

(NWO) for a TALENT fellowship.

[21] C. P. Collier, J. O. Jeppesen, Y. Luo, J. Perkins, E. W. Wong, J. R.

Heath, J. F. Stoddart, J. Am. Chem. Soc. 2001, 123, 12632–12641.

[22] Y. Luo, C. P. Collier, J. O. Jeppesen, K. A. Nielsen, E. DeIonno, G.

Ho, J. Perkins, H.-R. Tseng, T. Yamamoto, J. F. Stoddart, J. R.

Heath, ChemPhysChem 2002, 3, 519–525.

[23] H. Yu, Y. Luo, K. Beverly, J. F. Stoddart, H.-R. Tseng, J. R. Heath,

Angew. Chem. 2003, 115, 5884–5889; Angew. Chem. Int. Ed. 2003,

42, 5706–5711.

[24] M. R. Diehl, D. W. Steuerman, H.-R. Tseng, S. A. Vignon, A. Star,

P. C. Celestre, J. F. Stoddart, J. R. Heath, ChemPhysChem 2003, 4,

1335–1339.

[25] D. W. Steuerman, H.-R. Tseng, A. J. Peters, A. H. Flood, J. O. Jep-

pesen, K. A. Nielsen, J. F. Stoddart, J. R. Heath, Angew. Chem.

2004, 116, 6648–6653; Angew. Chem. Int. Ed. 2004, 43, 6486–6491.

[26] H.-R. Tseng, D. Wu, N. X. Fang, X. Zhang, J. F. Stoddart, Chem-

PhysChem 2004, 5, 111–116.

[1] R. L. McCreery, Chem. Mater. 2004, 16, 4477–4496.

[2] A. Nitzan, Annu. Rev. Phys. Chem. 2001, 52, 681–750.

[3] J. R. Heath, M. A. Ratner, Phys. Today 2003, 56, 43–49.

[4] H. McConnell, J. Chem. Phys. 1961, 35, 508–515.

[5] a) P. S. Vincett, G. G. Robert, Thin Solid Films 1980, 68, 135–171;

b) K. V. Mikkelsen, M. A. Ratner, Chem. Rev. 1987, 87, 113–153.

[6] a) E. E. Polymeropoulos, J. Sagiv, J. Chem. Phys. 1978, 69, 1836–

1847; b) T. Lee, W. Wang, J. F. Klemic, J. J. Zhang, J. Su, M. A.

Reed, J. Phys. Chem. B 2004, 108, 8742–8750.

[7] a) D. L. Allara, T. D. Dunbar, P. S. Weiss, L. A. Bumm, M. T. Cygan,

J. M. Tour, W. A. Reinerth, Y. Yao, M. Kozaki, L. Jones, Ann. N. Y.

Acad. Sci. 1998, 852, 349–370; b) J. Chen, M. A. Reed, A. M. Raw-

lett, J. M. Tour, Science 1999, 286, 1550–1552; c) T. D. Dunbar,

M. T. Cygan, L. A. Bumm, G. S. McCarty, T. P. Burgin, W. A.

Reinerth, L. Jones, J. J. Jackiw, J. M. Tour, P. S. Weiss, D. L. Allara,

J. Phys. Chem. B 2000, 104, 4880–4893; d) Z. J. Donhauser, B. A.

278

www.chemeurj.org

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Chem. Eur. J. 2006, 12, 261 – 279

Molecular Switches

FULL PAPER

[27] A. H. Flood, A. J. Peters, S. A. Vignon, D. W. Steuerman, H.-R.

Tseng, S. Kang, J. R. Heath, J. F. Stoddart, Chem. Eur. J. 2004, 10,

6558–6564.

[38] M. J. Blandamer, P. M. Cullis, J. B. F. N. Engberts, J. Chem. Soc.

Faraday Trans. 1998, 94, 2261–2267.

[39] a) P. R. Ashton, R. Ballardini, V. Balzani, S. E. Boyd, A. Credi,

M. T. Gandolfi, M. GomezLopez, S. Iqbal, D. Philp, J. A. Preece, L.

Prodi, H. G. Ricketts, J. F. Stoddart, M. S. Tolley, M. Venturi, A. J. P.

White, D. J. Williams, Chem. Eur. J. 1997, 3, 152–170; b) P. R.

Ashton, V. Balzani, J. Becher, A. Credi, M. C. T. Fyfe, G. Matter-

steig, S. Menzer, M. B. Nielsen, F. M. Raymo, J. F. Stoddart, M. Ven-

turi, D. J. Williams, J. Am. Chem. Soc. 1999, 121, 3951–3957.

[40] M. B. Nielsen, J. O. Jeppesen, J. Lau, C. Lomholt, D. Damgaard, J.

Jacomsen, J. Becher, J. F. Stoddart, J. Org. Chem. 2001, 66, 3559–

3563.

[28] a) J. O. Jeppesen, J. Perkins, J. Becher, J. F. Stoddart, Org. Lett.

2000, 2, 2631–2634; b) J. O. Jeppesen, J. Perkins, J. Becher, J. F.

Stoddart, Angew. Chem. 2001, 113, 1256–1261; Angew. Chem. Int.

Ed. 2001, 40, 1216–1221; c) J. O. Jeppesen, K. A. Nielsen, J. Perkins,

S. A. Vignon, A. DiFabio, R. Ballardini, M. T. Gandolfi, M. Venturi,

V. Balzani, J. Becher, J. F. Stoddart, Chem. Eur. J. 2003, 9, 2982–

3007; d) T. Yamamoto, H. R. Tseng, J. F. Stoddart, V. Balzani, A.

Credi, F. Marchioni, M. Venturi, Collect. Czech. Chem. Commun.

2003, 68, 1488–1514; e) H.-R. Tseng, S. A. Vignon, P. C. Celestre, J.

Perkins, J. O. Jeppesen, A. DiFabio, R. Ballardini, M. T. Gandolfi,

M. Venturi, V. Balzani, J. F. Stoddart, Chem. Eur. J. 2004, 10, 155–

172; f) S. Kang, S. A. Vignon, H.-R. Tseng, J. F. Stoddart, Chem.

Eur. J. 2004, 10, 2555–2564; g) I. C. Lee, C. W. Frank, T. Yamamoto,

H.-R. Tseng, A. H. Flood, J. F. Stoddart, J. O. Jeppesen, Langmuir

2004, 20, 5809–5828; h) J. O. Jeppesen, S. Nygaard, S. A. Vignon,

J. F. Stoddart, Eur. J. Org. Chem. 2005, 196–220.

[41] R. Castro, K. R. Nixon, J. D. Evanseck, A. Kaifer, J. Org. Chem.

1996, 61, 7298–7303.

[42] The impact of temperature on equilibrium constants, K and their as-

sociated population ratios, MSCC/GSCC, are related by the vanꢀt

Hoff equation [ln(K₁/K₂)=À(DH/R)(1/T₁À1/T₂)] such that it is the

enthalpic contribution that determines the temperature sensitivity in

K.

[29] a) E. Katz, O. Lioubashevsky, I. Willner, J. Am. Chem. Soc. 2004,

126, 15520–15532; b) D. Ryan, S. N. Rao, H. Rensmo, D. Fitzmaur-

ice, J. A. Preece, S. Wenger, J. F. Stoddart, N. Zaccheroni, J. Am.

Chem. Soc. 2000, 122, 6252–6257; c) B, Long, K. Nikitin, D. Fitz-

maurice, J. Am. Chem. Soc. 2003, 125, 15490–15498.

[30] a) A. M. Brouwer, C. Frochot, F. G. Gatti, D. A. Leigh, L. Mottier,

F. Paolucci, S. Roffia, G. W. H. Wurpel, Science 2001, 291, 2124–

2128; b) A. Altieri, G. Bottari, F. Dehez, D. A. Leigh, J. K. Y. Wong,

F. Zerbetto, Angew. Chem. 2003, 115, 2398–2402; Angew. Chem.

Int. Ed. 2003, 42, 2296–2300; c) A. Alteri, F. G. Gatti, E. R. Kay,

D. A. Leigh, D. Martel, F. Paolucci, A. M. Z. Slawin, J. K. Y. Wong,

J. Am. Chem. Soc. 2003, 125, 8644–8654; d) F. Cecchet, P. Rudolf, S.

Rapino, M. Margotti, F. Paolucci, J. Baggerman, A. M. Brouwer,

E. R. Kay, J. K. Y. Wong, D. A. Leigh, J. Phys. Chem. B 2004, 108,

15192–15199.

[31] a) A. Livoreil, C. O. Dietrich-Buchecker, J. P. Sauvage, J. Am.

Chem. Soc. 1994, 116, 9399–9400; b) J. P. Collin, P. Gavina, J. P.

Sauvage, New J. Chem. 1997, 21, 525–528; c) L. Raehm, J. M. Kern,

J. P. Sauvage, Chem. Eur. J. 1999, 5, 3310–3317; d) I. Poleschak,

J. M. Kern, J. P. Sauvage, Chem. Commun. 2004, 474–476.

[32] Although the Coulombic driven switching movement of the

CBPQT⁴⁺ring has not been measured, we estimate that the barrier

corresponding to the mono- and dication TTF^2+/+would be at least

~3 and ~6 kcalmol^À1, respectively, less than the 16 kcalmol^À1barri-

er observed for the free-energy barrier between the MSCC and

GSCC leading to room temperature time constants of t~500 and

3 ms, respectively.

[43] J. O. Jeppesen, K. Takimiya, F. Jensen, T. Brimert, K. Nielsen, N.

Thorup, J. Becher J. Org. Chem. 2000, 65, 5794–5805.

[44] J. G. Hansen, K. S. Bang, N. Thorup, J. Becher, Eur. J. Org. Chem.

2000, 2135–2144.

[45] P.-L. Anelli, P. R. Ashton, R. Ballardini, V. Balzani, M. T. Gandolfi,

T. T. Goodnow, A. E. Kaifer, D. Philp, M. Pietraszkiewicz, L. Prodi,

M. V. Reddington, A. M. Z. Slawin, N. Spencer, C. Vicent, D. J. Wil-

liams, J. Am. Chem. Soc. 1992, 114, 193–218.

[46] N. Svenstrup, K. M. Rasmussen, T. K. Hansen, J. Becher, Synthesis

1994, 809–812.

[47] C. L. Butler, A. G. Renfrew, L. H. Cretcher, B. L. Souther, J. Am.

Chem. Soc. 1937, 59, 227–229.

[48] D. D. Perrin, W. L. F. Armarego, Purification of Laboratory Chemi-

cals, Pergamon, New York, 1988.

[49] DG^ꢀat each temperature was calculated from the Eyring equation:

DG^ꢀ=ÀRTln(hk/k_BT) in which R, h, k and k_Bare the gas constant,

Planckꢀs constant, the first-order rate constant and Boltzmann con-

stant, respectively. k corresponds to (1Àx_eq)/t, in which x_eqis N_MSCC

/

N_Totalat equilibrium and t is 1/e decay time constant. DH^ꢀand DS^ꢀ

were obtained from the regression analysis of Gibbs–Helmholtz

plot, DG^ꢀ/T vs 1/T.

[50] The activation barrier (E_a) was calculated from the Arrhenius equa-

tion: k=Aexp(ÀE_a/RT) in which A is the activation coefficient. E_a

was obtained from a regression analysis of the Arrhenius plots, lnk

vs 1/T.

[51] The remnant molecular signature is a sequence of voltage pulses of

1 s that are applied to poly-Si electrodes with 100 mV step sizes and,

in between each pulse, is a read voltage of +0.1 V held for 1.5 s to

record the current. The metal top electrodes were connected to

ground through a preamplifier.

[33] K. Nørgaard, J. O. Jeppesen, B. W. Laursen, J. B. Simonsen, M. J.

Weygand, K. Kjaer, J. F. Stoddart, T. Bjørnholm, J. Phys. Chem. B

2005, 109, 1063–1066.

[34] B. W. Laursen, S. Nygaard, J. O. Jeppesen, J. F. Stoddart, Org. Lett.

2004, 6, 4167–4170.

[35] A. H. Flood, J. F. Stoddart, D. W. Steuerman, J. R. Heath, Science

2004, 306, 2055–2056.

[36] a) W. Deng, R. P. Muller, W. A. Goddard III, J. Am. Chem. Soc.

2004, 126, 13562–13563; b) Y. Kim, S. Jang, Y. Jang, W. A. God-

dard III, Phys. Rev. Lett. 2005, 94, 156801, 1–4.

[37] a) K. N. Houk, S. Menzer, S. P. Newton, F. M. Raymo, J. F. Stoddart,

D. J. Williams, J. Am. Chem. Soc. 1999, 121, 1479–1487; b) F. M.

Raymo, M. D. Bartberger, K. N. Houk, J. F. Stoddart, J. Am. Chem.

Soc. 2001, 123, 9264–9267.

[52] V. Daniel, Dielectric Relaxation, Academic Press, 1967.

[53] M. Iwamoto, C. Wu, Phys. Rev. E 1997, 56, 3721–3724.

[54] To the first order, the normalized current at equilibrium defined by

the I_OPEN/I_CLOSEDratio is approximately equal to N_MSCC/N_Totalif the

intrinsic conductance of the GSCC is smaller than the MSCC by

more than two orders of magnitude.

[55] a) R. S. Nicholson, I. Shain, Anal. Chem. 1964, 36, 706–723; b) L.

Raehm, J.-M. Kern, J.-P. Sauvage, Chem. Eur. J. 1999, 5, 3310–3317.

Received: August 1, 2005

Published online: December 1, 2005

Chem. Eur. J. 2006, 12, 261 – 279

ꢁ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chemeurj.org

279

Article Doi

Ground-state equilibrium thermodynamics and switching kinetics of bistable [2]rotaxanes switched in solution, polymer gels, and molecular electronic devices

DOI: 10.1002/chem.200500934

Source and publish data:

Authors:

Article abstract of DOI:10.1002/chem.200500934

Full text of DOI:10.1002/chem.200500934

Products guided by the article

R&D Labs maybe for 846571-55-9

Relevant to this article

Hot Product