Tuning energy transfer between chromophores. Switchable molecular photonic systems Hurenkamp, Johannes Henricus

Transcription

1 University of Groningen Tuning energy transfer between chromophores. Switchable molecular photonic systems Hurenkamp, Johannes Henricus IMPRTANT NTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2008 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Hurenkamp, J. H. (2008). Tuning energy transfer between chromophores. Switchable molecular photonic systems s.n. Copyright ther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 Tuning Energy Transfer Between Chromophores Switchable Molecular Photonic Systems Johannes Henricus Hurenkamp

3 Johannes Henricus Hurenkamp, Groningen, 2008 Cover picture : The work described in this thesis was carried out at the department of rganic and Molecular Inorganic Chemistry, Stratingh Institute, University of Groningen, The Netherlands. Part of the work described in this thesis was financially supported by the Zernike Institute for Advanced Materials. Zernike Institute PhD thesis series ISSN: PhD thesis Groningen University ISBN: (printed version) ISBN: (electronic version)

4 RIJKSUNIVERSITEIT GRNINGEN Tuning Energy Transfer Between Chromophores Switchable Molecular Photonic Systems Proefschrift ter verkrijging van het doctoraat in de Wiskunde en Natuurwetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op vrijdag 20 juni 2008 om uur door Johannes Henricus Hurenkamp geboren op 7 februari 1978 te Voorst

5 Promotores : Prof. dr. B.L. Feringa Prof. dr. J.H. van Esch Beoordelingscommissie : Prof. dr. J.B.F.N. Engberts Prof. dr. L.D.A. Siebbeles Prof. dr. J.G. Vos

6 Contents Chapter 1 Excited states: Theory and Applications Introduction Electronic absorption and emission spectroscopy Unimolecular photochemistry and electrochemistry Energy transfer in multicomponent systems Photo-induced electron transfer Donor Acceptor Systems Multiple Donors, one Acceptor systems using FRET Tuning excited state processes Thesis outline Chapter 2 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties Introduction Results and Discussion Electronic properties Redox properties Solvent dependence of spectroscopic properties Energy transfer Conclusions Experimental section Chapter 3 Tuning Energy Transfer in Switchable Donor-Acceptor Systems Introduction Synthesis Electronic and Photochemical Properties Discussion Conclusions Experimental section Chapter 4 Tuning of Energy Transfer Between Two Photoswitchable Coumarin-Dithienylcyclopentene Systems Through Structural Modification Introduction Synthesis Electronic properties Redox Properties Spectroelectrochemistry Excited state dynamics Discussion Conclusions Experimental section

7 Chapter 5 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches Introduction Results & Discussion Redox properties UV/Vis absorption, pen and PSS state Fluorescence spectroscopy of the pen State and PSS Photochromic Switching and Stability Energy transfer Electron Transfer Quenching Regarding the Energy Transfer and Quenching Mechanism Conclusions Experimental section Chapter 6 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentenes Introduction to computational chemistry Basic methods Molecular mechanics (MM) Ab initio methods Density functional theory (DFT) Experimental Calculations on three perylene bisimide models Partial density of state diagrams Exploring the remarkable different behaviour of two diphenyldithienylcyclopentene switches coupled with differently substituted coumarin donor groups Nederlandse Samenvatting Dankwoord

8 Chapter 1 Excited states: Theory and Applications The harvesting of light energy is perhaps the most important of physical processes to life on this planet. Indeed with notable exceptions, 1 life in one way or another depends on the transformation of light to chemical energy (i.e. photosynthesis). 2 The increasing importance of renewable energy sources over recent years has lead to an ever increasing interest in solar energy conversion, in particular solar cells. 3 This has driven fundamental research towards understanding of the processes involved in the conversion of light to other forms of energy and has led to exciting opportunities in future energy production methods. 4 7

9 Chapter Introduction 5 In this thesis the design and synthesis of multi-component systems, that engage in intramolecular energy transfer upon photo-excitation, is described. The primary goal is to develop approaches in which it is possible to exert control over both the direction and efficiency of energy transfer. In this chapter the underlying issues, i.e. energy transfer, stability and modulation, which must be considered in designing systems that are able to perform the task required will be discussed. In addition, the physical basis for the techniques that are employed to characterize the behaviour of multicomponent photoactive systems, from the initial absorption of light to the processes that follow that event, are discussed. 1.2 Electronic absorption and emission spectroscopy What is colour? Colour is what we see. 6 More specifically, the perception of colour is determined by the wavelengths of the visible light that reaches our retina. The absorption of light by 11-cis-retinal, followed by a cis to trans isomerisation is the first step in the process of vision. 7 The origin of the colour of the object observed lies in the wavelengths that are absorbed by the object, these are wavelengths of light that do not reach our eye, hence the colour is determined by the remaining non-absorbed or reflected wavelengths of light. 8 In order for a photon (i.e. hν) to be absorbed by an atom or molecule the energy of the photon must match an energy gap between two electronic states (e.g., S 0 S 1 ). Absorption of the photon increases the energy of the atom or molecule and promotes it to a photoexcited state (Figure 1.1, left). S 1 S 1 A + hν A* hν A* A + hν hν S 0 S 0 Figure 1.1 Absorption (left) and emission (right) of a photon by an atom. With the molecule now in the photo-excited state it is energetically unstable and quickly dissipates the excess energy either as heat, a chemical reaction or by emitting a photon 8

10 Excited states: Theory and Applications (Figure 1.1, right). In the following sections each of these photophysical events will be discussed in more detail Electronic absorption spectroscopy At its simplest level, electronic absorption occurs when an atom interacts with a photon and where the energy of that photon (i.e. hν) matches an energy gap between two electronic states (e.g., S 0 S 1 ), the atom or molecule can be promoted to an electronically excited state (Figure 1.2). Figure 1.2 Schematic representation of the excitation of molecule A to the first singlet excited state. In a simplified way, this transition is usually described as the transition of an electron from the HM to the LUM. However, in molecules, this description is not correct because it is not a single electron that is promoted from the highest occupied molecular orbital (HM) to the lowest unoccupied molecular orbital (LUM), but in fact the distribution of electron density in the molecule is changed. Also it should be realized that the geometry of a molecule in the ground state is not static, but it is vibrating, which means the bond lengths are changing continuously. Figure 1.3 shows the statistical distribution of the C=C bond length of ethene in the ground electronic state. However, the time scale of an electronic transition is such that the bond lengths and angles of a molecule do not change significantly during the transition (Born-ppenheimer approximation vide infra) and as a result the structure (i.e. bond lengths and angles) of the molecule in the excited state is identical to that in the ground state. The excited state formed, which has the same bond lengths as the ground state, is referred to as a Franck-Condon state (Figure 1.3). 9 There is, however, not just one Franck-Condon state. Both ground and excited states consist of discrete vibrational levels (the higher the number of a vibrational level, the higher in energy it is and the closer in energy it is to neighbouring vibrational levels), in principle a transition from the zeroth vibrational level of the ground state (S 0 ) to any of the vibrational levels of the S 1 is possible (Figure 1.4A), however the probability of most transitions is very low. 9

11 Chapter 1 Quantum mechanically the vibrational levels are interpreted as wave functions (Figure 1.4B). In the case of the molecule in Figure 1.4, the wave function of the S 0,v=0 has the largest overlap with the wavefunction of the vibrational state of S 1,v=2, and therefore at that energy difference the absorption probability is highest and a maximum will appear in the absorption spectrum. The theory that the most efficient vertical transition (vertical meaning no change in molecular geometry) has the largest overlap between the wave functions of the vibrational states involved, and therefore the highest probability of occurring, is called the Franck-Condon principle. 9 The Franck-Condon principle implies that all electronic transitions are assumed to be vertical with respect to nuclear motion, i.e. the Born-ppenheimer approximation. 10 This assumption is based on the fact that the nuclear movement is negligible on the time scale of the electronic absorption (i.e. nuclear kinetic energy is neglected), for this reason the nuclear positions can be treated as coordinates in calculations. Figure 1.3 A) Potential well diagram showing the probability of the C=C bond length of an alkene in the lowest vibrational state of the ground state. B) Potential well diagram showing: a. absorption, b. Franck-Condon state, c. vibrational relaxation, d. fluorescence, e. displacement of the excited state minimum from that of the ground state (cf. the Huang-Rhys factor). 10

12 Excited states: Theory and Applications Figure 1.4 Potential well diagram of the ground and first excited state of a hypothetical molecule showing probable transitions AU Wavenumber (cm -1 ) Figure 1.5 A) Excitation of a molecule from its electronic ground state (S 0 ) to one of the vibrational levels (i.e. v=1, v=2, etc.) of the first excited state (S 1 ) B) Schematic absorption spectrum showing the vibrational fine structure. C) UV/Vis spectrum of anthracene showing its vibrational fine structure. This transition to different vibrational levels of the S 1 electronic state gives rise to the shape of the absorption spectrum (Figure 1.5B). Different transitions (i.e. 0 0, 0 1, etc.) give rise to maxima at different energies. The efficiency of the different transitions (transition probability) determines the eventual intensity of each individual absorption band, the absorption probability decreases on going from the 0 0 to the 0 3 transition in this simplified example (Figure 1.5A). 11

13 Chapter 1 However, the molecular geometry of the lowest excited state S 1,v=0, which is reached after vibrational relaxation (cooling), is different to that in the Franck-Condon state (i.e. excited state displacement, Figure 1.3B(e)), due to changes in the distribution of electron density and hence bond strength and vibrational structure. Therefore the potential well, which represents the first excited state, has an offset nuclear coordinate compared to the ground state (Figure 1.3B). This offset is usually quantified with the Huang-Rhys factor (S), with S < 1 indicating weak coupling between ground and excited state, and S > 1 indicating strong coupling, 11 for a particular coordinate Emission nce a molecule is in an electronically excited state several processes can occur, however, when there are no other molecules present with which it can interact these can be narrowed down to four probable processes: fluorescence, phosphorescence, dissociative bond breaking (See 1.3 ) and non-radiative decay (events that can occur when other interacting molecules are present will be discussed in 1.4 to 1.8 ). The first two involve the emission of a photon (of lower energy than that absorbed), whereas the last option does not and therefore generates heat (which is transferred to the solvent as vibrational heat). The first two processes are of most interest in this thesis. However, as with absorption, the situation for molecules is more complicated than for atoms (e.g., gaseous ions, lanthanide metals) (Figure 1.6 and Figure 1.7). Figure 1.6 Schematic representation of a molecule A* returning back to the ground state, through emission of a photon. 12

14 Excited states: Theory and Applications AU Wavenumber (cm -1 ) Figure 1.7 A) Emission from a molecule by relaxation from its first excited state (S 1 ) to one of the vibrational levels (i.e. v=0, v=1, etc.) of its ground state (S 1 ) B) Schematic emission spectrum showing the vibrational fine structure. C) Emission spectrum of anthracene showing its vibrational fine structure. Figure 1.8 Schematic representation of possible processes after a. excitation: b. vibrational relaxation, c. fluorescence, d. intersystem crossing and e. phosphorescence. 13

15 Chapter 1 nce a molecule is in the excited state (i.e. the Franck-Condon state, the result of a vertical transition, Figure 1.3B(a) and Figure 1.8(a)), in this case S 1,v=3, it will first relax to the lowest vibrational state (S 1,v=0 ), with the rate of vibrational relaxation (also known as vibrational cooling, Figure 1.3B) being fast (< 10 ps). Following relaxation, emission (fluorescence) of a photon will occur by a transition from the lowest vibrational state of the excited state (S 1,v=0 ) to one of the S 0 vibrational states (Kasha s rule). 13 The photon emitted is of lower energy than that which generated the excited state due to the different nuclear coordinates of S 0 and S 1 (Figure 1.8). This difference in energy is known as the Stokes shift and can be observed as the difference between the λ max of absorption and the λ max of emission. The most probable transition (as with absorption) is that in which the wavefunction of the S 1,v=0 has the largest overlap (i.e. largest probability) with the wavefunction of one of the S 1 vibrational states, in this case S 1,v=0 to S 0,v=1 (See wavefunctions in Figure 1.4B). Another possibility is phosphorescence (Figure 1.8), 14 which requires conversion of the excited state from the first singlet excited state (S 1 ) to the first triplet excited state (T 1 ). Since these states have a different spin multiplicity, this is called intersystem crossing (ISC). ISC is a quantum mechanically forbidden process; it is, however, still possible for ISC to take place. This is due to spin-orbit coupling (i.e. an interaction between the spin angular momentum and the orbital angular momentum), which allows the change in spin angular momentum (i.e. S 1 to T 1 ) to coincide with a change in orbital angular momentum, thereby compensating for the change in the former. This spin orbit coupling also explains the heavy atom effect, where significantly higher ISC rates for molecules, which contain heavy atoms (e.g., I, Br or Cl) are observed. 15 This arises from an increased mixing of spin and orbital quantum number with increased atomic number. After ISC, phosphorescence is possible by radiative relaxation to the ground state S 0, this is a relatively slow process (i.e. rates from 10-1 to 10 6 sec -1 ) due to its spin forbidden nature and was not investigated for molecules studied in this thesis (all luminescence lifetimes were < 100 ns). 14

16 Excited states: Theory and Applications 1.3 Unimolecular photochemistry and electrochemistry Unimolecular photochemistry Fluorescence and phosphorescence are only two of the processes that can occur when a molecule is in the photoexcited state. Photochemistry can occur also, in which the molecule undergoes a light induced structural transformation (e.g., bond breaking/making, 16 changes in conformation 17 ). Several mechanisms for a unimolecular photochemical reaction are depicted in Figure 1.9. A diabatic (i.e. non-adiabatic) reaction is a reaction which involves two energy surfaces. The molecule is excited to the first excited state, which has a minimum that lies between the reactant and the product ground state minima. The exact location of the excited state minimum (i.e. the molecular geometry) relative to the ground state energy surface is important, since it will determine whether the final state reached is the original reactant or the product. If the minimum is to the left of the local maximum of the ground state energy surface after relaxation to the S 0 state the molecule will return to the original state, if this happens to the right of the maximum the product will be formed. This type of transition area is called a funnel or conical intersection. Figure 1.9 Photochemical reactions depicted with schematic energy surface diagrams A) diabatic, B) adiabatic and C) hot ground state. Less frequently encountered mechanisms for photochemical reactions include adiabatic reactions, in which the reaction occurs via one potential energy surface, in this case the surface of the excited state (Figure 1.9B). Another mechanism is the so called hot ground state reaction, in which the molecule relaxes from the excited state via internal conversion (Figure 1.9C). The energy which has been transferred to vibrational excited states is used to initiate a thermal reaction on the ground state surface, this is possible because the molecule relaxes to a higher energy vibrational state thereby lowering the activation energy. Products resulting from these hot ground state reactions are identical to those obtained from thermal reactions. 15

17 Chapter Examples of unimolecular photochemical reactions The cis-trans isomerization of an alkene is, perhaps, a paradigm photochemical reaction. This isomerization is a diabatic photochemical reaction (Figure 1.9A); in the excited state the π bond is broken and rotation around the C-C bond is possible. The steric interaction of the R-groups and electronic repulsive effect (of the electrons) favour an orthogonal conformation, which is similar to that which would be observed for the thermal reaction (Figure 1.10). In this case the minimum in the S 1 energy surface is very close to the maximum in the S 0 surface, which is ideal for the formation of a funnel that can allow for the formation of both the cis and the trans states (See Figure 1.9A). Figure 1.10 Photochemical cis-trans isomerization of an olefin Switchable molecules using cis-trans isomerization Stilbene is a well know example of this photochemical cis-trans isomerization and studies into the isomerization of diversely substituted stilbenes have provided a considerable level of understanding regarding the excited state structures of these types of molecules (Figure 1.11). 18,19 cis-stilbene is also able to undergo a photocyclization reaction; this type of photochemical reaction will be treated in more detail later. Figure 1.11 Photoisomerization and photocyclization of stilbene. The isoelectronic azobenzenes are also known to isomerize under the influence of light. Irradiation with λ = 300 nm light gives a photostationary state (PSS) of 80% cis and 20 % trans (Figure 1.12). However, when irradiation is discontinued, the cis form converts back to the more stable trans form slowly, because the cis form is not stable thermally

18 Excited states: Theory and Applications Figure 1.12 Photoisomerization of azobenzene. A photostationary state is reached when the rate of conversion from trans to cis and from cis to trans are equal. The conversion rates depend on the molar absorptivity of each form at the irradiation wavelength and the quantum yield (QY) of the photochemical process (i.e. the fraction of light which is absorbed that is used for the process e.g., isomerization), and the concentrations. At the PSS, this leads to Eq [ trans] ε cφ c = Eq. 1.1 cis ε Φ [ ] t t Where ε is the molar absorptivity at the wavelength of irradiation and Φ the quantum yield of the process. Light-induced cis-trans isomerization under the influence of light has been used as the basis of chiroptical switches based on overcrowded alkenes, which were developed in the Feringa group. 21 The overcrowded (i.e. sterically strained) central alkene allows for isomerization with relatively low energy light. Introducing different substituents on the lower half (i.e. N 2 or NMe 2 ) makes it possible to address two PSS with different ratios: irradiation with λ = 435 nm light lead to a PSS consisting of 10% M-(trans-nitro) and 90% P-(cis-nitro), and using λ = 365 nm light a PSS with a ratio of 70% M-(trans-nitro) and 30% P-(cis-nitro) is obtained. 22 Taking this one step further has led to the unidirectional motors based on overcrowded alkenes. These molecular motors use a four step process to make a 360 o turn; to achieve this, two photochemical steps with high PSS in favour of the so-called unstable state are each followed by a thermal isomerization step. A single chiral centre is used to induce the unidirectional preference in the cycle

19 Chapter 1 S S 435 nm 2 N NMe nm 2 N NMe 2 S S M-(trans-nitro) P-(cis-nitro) Figure 1.13 Chiroptical switch based on an overcrowded alkene. Figure 1.14 A four step unidirectional rotation by a sterically overcrowded alkene motor. 18

20 Excited states: Theory and Applications Photocyclization Another type of photochemical reaction is photocyclization, where rearrangement of double bonds leads to the formation of a cyclic structure. Figure 1.15 Photocyclization of butadiene (left) and hexatriene (right). This reaction can usually take place thermally also, however there is one major difference, the stereochemical outcome (i.e. in the case of substituted dienes) of both reactions, photochemical and thermal, is opposite. These pericyclic reactions (i.e. concerted, with a transition state of cyclic geometry) have to be symmetry-allowed to be able to occur. This means, as is described in the Woodward-Hoffmann rules, 24 that the overlap of molecular orbitals must be constructive for a bonding interaction to develop across the transition state. For a system with 4n+2 π-electrons (e.g., 2,4,6-octatriene) the thermal reaction involves the ground state, and therefore constructive overlap leading to ring-closure can be achieved by a disrotatory movement of the orbitals (i.e. one orbital rotating clockwise and one orbital rotating counter clockwise, Figure 1.16, top). The photochemical reaction involves the first excited state orbitals, and therefore for a system with 4n+2 π-electrons the situation is reversed and conrotatory motion is required to obtain a symmetry-allowed reaction (Figure 1.16, bottom). For systems with 4n π-electrons (e.g., 2,4-hexadiene) the situation is reversed. thermal = Ground state disrotatory Meso photochemical = S 1 state conrotatory RR (or SS) Figure 1.16 Thermal and photochemical cyclization of 2,4,6-octatriene and its stereochemical outcome according to the Woodward-Hoffmann rules. 19

21 Chapter Diarylethenes; switching by photocyclization Photochromic diarylethenes are a well known example, which is based on reversible photocyclization. These molecules are able to undergo a reversible photochemical cyclization (Figure 1.17) and, due to heterocyclic aromatic groups (e.g., thiophene, furan), both states are thermally stable, making them suitable for application in e.g., data storage. Upon ring-closure either the RR or SS form of the ring-closed diarylethene is obtained, which leads to a racemic mixture in solution. Figure 1.17 The stereoselective photocyclization of a diarylethene photochromic switch. The fact that these molecules are synthetically very versatile, their structure and properties are tunable, and that the compounds are also photochromic opens up many interesting possibilities 25 as described in this thesis. For example diarylethenes have been applied as switchable gelators, 26 switchable crystals 27 and were applied in data storage 28 as well as switchable surfaces, 29 but most important, with respect to this thesis, is their use as fluorescence intensity modulators. 30 ther well known photochromic switches that use photochemical cyclization are fulgides 31 and spiropyrans, 32 however since they are not discussed in this thesis they are not considered in detail here Electrochemistry In a redox reaction a molecule can undergo a reaction by either accepting or donating electrons. This can be spontaneous or under the influence of an electric potential. When a neutral molecule is oxidized the HM loses one electron and it becomes positively charged (now a SM, Singly ccupied Molecular rbital), whereas reduction provides an additional electron to the LUM (which becomes a SM also). The reaction can be reversed (chemically reversible) or the oxidised/reduced molecule can undergo a reaction and form new compounds (chemically irreversible). 20

22 Excited states: Theory and Applications Figure 1.18 Schematic representation of the oxidation and reduction of a neutral molecule. By using cyclic voltammetry it is possible to determine the energy levels of the HM and (indirectly) of the LUM. This can provide valuable information about the relative energy of the HM and LUM levels of the components in multicomponent systems, which can be used to determine if inter-component events, such as energy or electron transfer, are favourable Application of theoretical chemistry Molecular properties can be predicted and understood using calculations, e.g., density functional theory (DFT) and molecular dynamics (either based on theory or empirical data or a mixture of both). 33 Due to the increase in computational power it is possible to apply calculations to increasingly larger molecules whilst using a decreasing number of assumptions. The results from these calculations still show improvements in the degree of accuracy when compared to data obtained empirically. It is possible to calculate properties such as molecular orbitals, energy levels, electronic transitions and reaction pathways. A more in depth discussion of methods and the accuracy of calculations can be found in the introduction to Chapter Energy transfer in multicomponent systems Introduction nce a molecule is in an excited state and a suitable energy acceptor is present (i.e. within sufficient range, possibly even attached covalently), it is possible to transfer the energy from the donor to the acceptor (Figure 1.19). 21

23 Chapter 1 Figure 1.19 Schematic representation of energy transfer from a donor (D) to an acceptor (A). ften energy transfer is rationalized using either of two mechanisms: Dexter energy transfer ( through bond ) 34 or Förster resonance energy transfer (FRET, through space ). 35 However, these models are two idealized situations and in practice for small molecular systems relative rates of interaction between states are of more importance. In the following paragraphs trivial processes will be treated briefly followed by a discussion of the two primary models describing radiationless energy transfer, and a more detailed examination of their practical application to systems described in the present work Trivial energy transfer Perhaps the most simple of energy transfer mechanisms, 5 which is known as radiative energy transfer also, is excitation of a donor by a photon, emission of a photon of longer wavelength, which is followed by reabsorption by a suitable acceptor. In optically dilute solutions where the donor and acceptor are not connected (either covalently or through supramolecular bonding) the probability of reabsorption by the acceptor is quite low. Even in donor-acceptor systems which are connected covalently, trivial energy transfer (i.e. photon emission and reabsorption) does not contribute significantly, due to its statistical nature Collisional energy and electron transfer In systems where the excited state lifetime is sufficiently long-lived in order that the probability of collision with an acceptor by diffusion is high then one of several processes can occur; i.e. energy transfer, electron transfer and triplet sensitization (e.g., 3 2, Ar). 5 Under ambient conditions in non-viscous solvents collisional energy transfer and electron transfer are not normally observed for donor systems with excited state lifetimes of less than 10 ns. However, the diffusion rate of dioxygen and argon are sufficiently high under saturated conditions to give rise to significant photochemistry. Possible reactions are in particular intersystem crossing (ISC) of the singlet excited state to a triplet excited state and further oxidation from singlet oxygen which is formed. The rate of this excited state quenching can be determined using a Stern-Volmer plot where either I/I 0 (emission intensity with (I) and without (I 0 ) quencher) or τ /τ 0 (excited state 22

24 Excited states: Theory and Applications lifetime with (τ) and without (τ 0) quencher) is plotted against the quencher concentration [Q]. From the slope of this graph the Stern-Volmer constant (K SV ) and hence the rate of quenching can be determined using either the rate of deactivation without a quencher present (k 0 ) or the excited state lifetime without the quencher present (τ 0). I / I 0 = τ / τ 0 = 1+ K SV [ Q] where K SV = k q / k0 = k q τ 0 Eq Dexter energy transfer Dexter type energy transfer is often referred to as through bond energy transfer, which incorrectly suggests that a covalent (conjugated) connection between donor and acceptor is required. The Dexter excitation transfer mechanism, however, is based on electron exchange, this means that it requires an overlap between the wavefunctions of the donor and acceptor, which is also possible without covalent connection (Figure 1.20). Dexter excitation transfer is possible when the donor is in the excited state and an acceptor is present with a compatible LUM (i.e. of equal or lower energy). In that case the electron from the SM of the donor can be transferred to the LUM of the acceptor and simultaneously an electron is transferred from the HM of the acceptor (which must be equal or higher in energy then the HM of the donor). This way the net result is the transfer of the excited state from the donor to the acceptor. Figure 1.20 Schematic representation of the Dexter electron exchange energy transfer mechanism. The rate of energy transfer for the Dexter mechanism is dependent primarily on the distance between donor and acceptor R DA and on the overlap between the respective orbitals (J) (Eq. 1.3). 2RDA Eq. 1.3 k ET ( Dexter) = KJ exp L K is a collection of constants related to specific orbital interactions. J is the spectral overlap integral, see below (Eq. 1.4). The orbital overlap, which is required for the Dexter mechanism, is apparent as an exponential distance dependence of exp(-2r DA /L), where L is related to the van der Waals radii of the chromophores. 23

25 Chapter 1 J = 0 f D ( σ ) ε ( σ ) dσ A Eq. 1.4 For the Dexter exchange mechanism the spectral overlap integral J (Eq. 1.4) is dependent on the integrated overlap of the donor fluorescence spectrum (f D ) and the acceptor absorption spectrum (ε A ), over wavenumbers σ. Both the absorption and emission spectra are normalized to unity (i.e. complete overlap would give J = 1). As a result J does not depend on the magnitude of ε A, and therefore k ET (Dexter) is independent of the oscillator strength of the D* D and A A* transitions; in contrast to Förster resonance energy transfer Förster resonance energy transfer While Förster resonance energy transfer (FRET), 35 is often referred to as fluorescence resonance energy transfer in the literature, fluorescence is not actually involved. However, the overlap between the donor fluorescence spectrum and the absorption spectrum of the acceptor plays an important role. verlap between wavefunctions is not a requirement and exchange of electrons does not occur as with the Dexter mechanism. The mechanism is based on resonance between two oscillating dipoles (i.e. the electric fields associated with electrons). When the excited state has an energetically matching oscillating electric field to that of a nearby molecule in its ground state [i.e. E(D* D)= E(A A*) ], it is possible for the two dipoles to couple coulombically and the excited state energy can be transferred from the donor to the acceptor without exchange of electrons ( through space ), bringing the donor to the ground state and the acceptor to the excited state (Figure 1.21); in a sense like a molecular equivalent of a transformer. Figure 1.21 Schematic representation of the Förster resonance energy transfer mechanism. The Förster mechanism depends largely on three factors, first: the overlap between the donor fluorescence spectrum and the acceptor absorption spectrum (J), second: the orientation between the dipoles of the chromophores (κ), better orientation gives better coupling and therefore better k ET. When the orientation is not known a value of 2/3 is used, which means the relative orientation averaged over all donor-acceptor pairs is assumed to 24

26 Excited states: Theory and Applications be random. The third is a 1/R DA 6 distance dependence, where R DA is the distance between donor and acceptor (Eq. 1.5). k ET 2 κ J n τ R ( Förster) = 4 6 D 28 DA mol Eq. 1.5 In which n is the refractive index of the medium and τ D is the radiative lifetime of the donor with no acceptor present. In the case of FRET the overlap integral J is dependent on the overlap between the donor fluorescence spectrum and the acceptor absorption spectrum (Eq. 1.6, Figure 1.22). The difference with the Dexter mechanism is, however, that the k ET (Förster) depends on the oscillator strength of the A A* transition, stronger absorption by the acceptor (i.e. a higher molar absorptivity) gives a higher efficiency of energy transfer. Therefore, a normalized value is used for the donor fluorescence spectrum only and not for the acceptor (Eq. 1.7). J = 0 ( σ ) ε ( σ ) f ' D A Eq. 1.6 dσ 4 σ Where f D is the normalized emission of the donor: f D ( σ ) f ' D = f D ( σ ) dσ 0 Eq. 1.7 Figure 1.22 Schematic representation of the overlap integral J. FRET is often used to determine distances in large biomolecules (e.g., proteins) by connecting a known donor and acceptor pair to e.g., specific amino acids. These distances can then be used to determine the structure or change in structure upon addition of a 25

27 Chapter 1 substrate to the biomolecule. To determine the distance between the donor and acceptor the following formula can be used: 1 E = Eq R DA 1 + R0 Where E is the energy transfer efficiency and R 0 is the Förster radius, which is the distance at which energy transfer is 50% efficient and which is a property of an individual donor and acceptor pair of molecules. The efficiency can be determined by either measuring the fluorescence lifetime (τ D ) of the donor with and without the acceptor present or the donor fluorescence intensity (F) with and without the acceptor present (Eq. 1.9) τ D ' FD ' E = 1 = 1 Eq. 1.9 τ F D D If the Förster radius is known then the distance between the chromophores can be determined. The Förster radius can be determined using the following equation (Eq. 1.10). 6 R = Φ D κ J Eq The mechanisms compared: Dexter and Förster It is important to take note of some significant differences between the Dexter and Förster mechanisms: The Dexter mechanism does not depend on the oscillator strength of the D* D and A A* transitions, therefore the overlap integral is of the donor fluorescence and acceptor absorption spectra normalized to unity, thereby eliminating the dependence on intensity. For the Förster mechanism the strength of transitions plays an important role since k ET (Förster) depends mostly on the A A* transition oscillator strength, therefore in the overlap integral (J) for this dipole-dipole mechanism the value for the molar absorptivity for the acceptor (ε A ) is also present. Both mechanisms have a distance dependence, however, for the Dexter exchange mechanism this is a much steeper one (i.e. exp(-2r DA /L) compared to 1/R DA -6 for FRET). This means that for systems that meet the requirements for Dexter type energy transfer the value for k ET becomes negligible once the R DA increases beyond 5-10 Å (> 1 nm). 26

28 Excited states: Theory and Applications Another issue involving distance is the fact that it is difficult to assign in a meaningful way the donor-acceptor distance. For example, what is the distance; in both Förster and Dexter theory the distance is that between two point dipoles and while this holds for lanthanide systems, which have metal-centered excited states, in the case where the excited state involves a larger number of atoms (i.e. more than one) an assumption has to be made as to the location of the point dipole (Figure 1.23A). In biological systems where the donor and acceptor units are separated by a large distance relative to their own size, the uncertainty in the donor-acceptor distance (10 nm) in terms of the calculations is minor (Figure 1.23B). In more closely spaced systems the uncertainty in position of the dipoles is large in comparison to the donor-acceptor distance. A B Figure 1.23 A) A supramolecular system where the location of the point dipole is difficult to pinpoint. 36 B) Cartoon of the enzyme RNase H with the FRET dyes attached to residues 3 and Somewhere between Förster and Dexter Most system are rationalized using either a Dexter or Förster mechanism, however, as will be shown in this thesis, this is not always the most effective approach to rationalize the results obtained. As discussed in the previous paragraph, the Dexter and Förster mechanism are two idealised situations. Between these ideal models there is a large grey area which is better described by considering the rates of interaction and relative levels of excited states than by assigning energy transfer to a specific mechanism

29 Chapter 1 In a simple A-B system as shown in Figure 1.24 there are a limited number of possibilities (i.e. considering an excitation wavelength that excites A only): Excitation and emission from A (k 1 and k -1 ). Excitation of A (k 1 ), ET to B (k 2 ) and emission from B (k 3 ), (If state A*-B and A- B* are sufficiently close in energy k -2 can become important). 39 The rates of the respective processes determine the overall rates observed (k obs ). This does not just have to be emission from A or B only; but rather increased emission from A and decreased emission from B or vice versa is possible also. Figure 1.24 Donor-Acceptor system with indicated rates for ET between states. The advantage of considering systems in this way is that it deals with the system as a whole and looks at the current energy of the system (e.g., that of A-B, A*-B or A-B*) relative to other possible states. Rates (which can be determined by ultrafast spectroscopy) can then be used to determine if the observed emission is what would be expected or that other mechanisms are operating (e.g., electron transfer). This approach also allows for more clear representation of bi- and multi-component systems. 1.5 Photo-induced electron transfer Although the primary focus of this thesis is energy transfer, electron transfer is an important photochemical process. There are two mechanisms for electron transfer, both requiring a compatible acceptor in proximity to a donor. With oxidative electron transfer an electron is transferred from the donor excited state to the empty, (lower energy) LUM of a nearby acceptor creating a charge separated state with a positive donor and negative acceptor (Figure 1.25a). If the energy levels are as depicted in Figure 1.25B the end result is reversed. The donor is in the excited state and the acceptor transfers an electron from its 28

30 Excited states: Theory and Applications HM to the vacancy in the HM of the donor, which gives a negatively-charged donor and positive acceptor. Both mechanisms of electron transfer can be followed by back electron transfer, after which both donor and acceptor will be returned to the ground state. However, the high energy charge separated state can also lead to various types of photochemistry, which can lead to undesired (side)reactions and bleaching, it is also the basis of n- and p-type dye based semi conducting solar cells (i.e. the Graetzel type cells 3 ) and photosystem II, 2 in which the rate of back electron transfer is reduced as much as possible. Figure 1.25 Schematic depiction of oxidative (a) and reductive (b) electron transfer. A An example showing the ability of perylene bisimides to undergo electron transfer was reported by the group of Meijer. An oligo(p-phenylene vinylene) (PV) electron donor and perylenebisimide electron acceptor can interact via hydrogen bonding in a 2 :1 ratio (Figure 1.26A). 40 Upon irradiation of the perylene bisimide component (λ ex = 578 nm) electron transfer from the PV to the perylene bismide occurs and a charge separated state is formed with a lifetime of 300 ps. At higher concentrations the system is prone to selfassembly into helical aggregates (Figure 1.26B), and within these aggregates the lifetime of the charge separated state is reduced to 60 ps. B C 12H 25 C 12H 25 C 12H 25 H N H N N H N N N H H H H N N N H N N H N H C 12H 25 C 12H 25 C 12H 25 Figure 1.26 A) Hydrogen bonded perylene bisimide / PV triad ( 1 : 2 ratio) B) AFM image of the helical structures formed by self-assembly of the perylene bisimide / PV triads upon an increase in concentration (>10-5 M in methylcycohexane). 29

31 Chapter Donor Acceptor Systems Introduction Donor acceptor (D-A) systems are capable of transferring their excited state energy from the donor to the accepter either via Coulombic interaction or some form of electron exchange mechanism. In this paragraph several principles and applications will be discussed using relevant examples from the literature Single Donor Single Acceptor systems using FRET Donor - acceptor pairs that meet the requirements stated in and are in sufficient proximity, are able to transfer energy from donor to acceptor, after which an increase in acceptor fluorescence intensity can be observed compared to that of the acceptor alone (irradiated at the same wavelength with the same light intensity). An example of such a donor acceptor pair are coumarin-4 and coumarin-343 (Figure 1.27). Both coumarins are used as laser dyes and have a high quantum yield of fluorescence. The absorption and emission spectra show that there is a good overlap (J) between donor (coumarin-4) emission and acceptor (coumarin-343) absorption (Figure 1.27), a requirement for efficient energy transfer. This dye combination was used in electrostatically self-assembled multilayer films, composed of polycations labelled with either dye, and an energy transfer efficiency of over 70% was observed. 41 A B H N CH Coumarin-4 Coumarin-343 Figure 1.27 A) Structures of donor: coumarin 4 and acceptor: coumarin 343. B) Left: Absorption and emission spectrum of coumarin-4 ( ), Right: absorption and emission spectrum of coumarin-343 (---). Shaded area: overlap between coumarin-4 fluorescence and coumarin-343 absorption. 30

32 Excited states: Theory and Applications Another well know dye pair are N-(7-nitro-2,1,3-benzoxadiazol-4-yl) (NBD) and Rhodamine-B (Figure 1.28). These dyes were used as a fluorescent sensors for vesicle fusion. Lipids were labelled with either dye and vesicles were produced which contain three types of lipids, both of the labelled lipids and a non-labelled lipid. These vesicles were mixed with non-labelled vesicles and when fusion is induced, the dyes were more dilute (i.e. average distance R DA increases) and as a consequence energy transfer was less efficient (Figure 1.29B). 42 N N N NH 2 N CH Cl N NBD Rhodamine B Figure 1.28 A dye pair used to monitor vesicle fusion, left: N-(7-nitro-2,1,3- benzoxadiazol-4-yl) (NBD) and right: Rhodamine B. A B AU Wavelength (nm) Figure 1.29 A) NBD Absorption ( ) and fluorescence ( ) spectra; Rhodamine B absorption (----) and fluorescence ( ). B) Schematic representation of decrease of energy transfer efficiency (between donor (D) and acceptor (A)) as indication of vesicle fusion. 31

33 Chapter 1 Even the war on terror sees application of FRET. By using an amine-terminated donor and acceptor (both coumarin based) it was possible to detect phosgene. 43 In the presence of phosgene three urea based products are obtained (Figure 1.30), two symmetric and one asymmetric, the asymmetric product contains both donor and acceptor and exhibits FRET. In the absence of phosgene and using λ ex = 343 nm for excitation fluorescence from the acceptor coumarin is not observed. When phosgene is present significant enhancement of the emission of the acceptor at λ em = 464 nm is observed. The system is a nice proof of principle, however, the concentrations required for detection are still too high for practical use (i.e. for detection below the lethal dose). Figure 1.30 Phosgene detection using amine terminated coumarins and FRET (Right: changes observed in the fluorescence spectrum in the presence of phosgene.). 1.7 Multiple Donors, one Acceptor systems using FRET Light Harvesting: Nature and Principle Using multiple donor molecules and only one acceptor offers exciting possibilities in energy transfer. The absorption cross-section of the acceptor is increased enabling the molecule to absorb a larger proportion of the available photons with increasing donor number. Also the amount of favourable donor acceptor dipole-dipole orientations is increased (i.e. κ 2 ). These advantages over a 1 : 1 ratio result, as an overall effect, in the fluorescence intensity of the acceptor being amplified substantially compared to a single donor / single acceptor system. This effect is referred to as light harvesting (LH) and has been used for billions of years in biotic systems (i.e. bacteria, plants) where arrays of chromophores (i.e. different types of chlorophylls) funnel energy to a single reaction centre, where the energy is eventually used to synthesize ATP from ADP (Figure 1.31)

34 Excited states: Theory and Applications A B Figure 1.31 A) Structure of Chlorophyll A (R=Me) and B (R=CH). B) Schematic representation of the LH system in plants with LH1, LH2 : rings of chromophores for energy transfer (ET) and the reaction centre (RC) Light Harvesting: Synthetic systems The Fréchet group have reported several examples of multiple donor single acceptor systems based on their well known aryl ether dendrimers. 45 Dendrimers are very suitable candidates for donor-acceptor arrays, as they offer the possibility of arranging donor and acceptor units spatially and of controlling communication between these chromophores. A combination of coumarin-2 donors and a coumarin-343 acceptor was used to build 4 generations of aryl ether dendrimers, the number of donors doubles with every generation (i.e. 2, 4, 8 and 16 donors) (Figure 1.32). 46 The emission spectra of the four different generations of dendrimer show a clear increase in fluorescence output with increasing generations, almost doubling with every step (Figure 1.33). 46 Starting from the third and higher generations the sensitized core emission starts to be enhanced above the level that could be obtained by direct acceptor excitation with the same intensity of irradiation. This is the antenna effect: more light is emitted by the acceptor in combination with the donors than could be emitted by the acceptor alone by absorption under the same irradiation intensity. 33

35 Chapter 1 Figure 1.32 The first and fourth generation of the coumarin-2 donor coumarin-343 acceptor aryl ether dendrimers. Figure 1.33 Emission spectra of the four generations of coumarin-2 donor coumarin-343 acceptor aryl ether dendrimers (numbers 1-4), and of the directly excited coumarin-343 emission (arrow), intensity compensated for irradiation power

36 Excited states: Theory and Applications Fluorescence studies of the donor dyes in the presence of the acceptor show that the fluorescence intensity of the donor dyes diminishes by more than 40-fold in the first three generations, this indicates an energy transfer efficiency of more than 97%. In the fourth generation dendrimer, quenching was less effective and efficiency decreased to 86% with some direct donor emission being observed (Figure 1.33); this could be due to the increase in donor acceptor distance, the rate of emission by the donor itself is now in competition with the rate of energy transfer Different Donors Single Acceptor System : Energy gradient Nanostar-type dendrimers having phenylacetylene dendrons and a perylene core have been reported by the groups of Moore and Kopelman (Figure 1.34). 47 This type of dendrimer was one of the first types to show that the energy transfer efficiency from donor to acceptor could be optimized by adding an energy gradient towards the acceptor core, similar to that used by nature. By a stepwise increase in the conjugation length (i.e. lower energy absorption) in the direction of the core the energy absorbed by the peripheral chromophores can be directed towards the perylene acceptor. The mechanism via which the energy transfer in these systems occurs is not directly clear as both through bond and through space mechanism seem conceivable. The chromophores are connected by conjugated pathways, however, the meta-substitution is supposed to break this conjugation, therefore neither option can be excluded, a priori. A B Figure 1.34 Structure of two the Nanostar-type phenylacetylene dendrimers A) without energy gradient. B) with built in energy gradient. 35

37 Chapter 1 The Fréchet group has synthesized a dendrimer containing two different donors: coumarin- 2 and fluorol-7ga. These were connected to a perylene bisimide acceptor using aryl ether branches (Figure 1.35 A). Experiments with different model compounds showed that energy can follow two routes to the acceptor when irradiated at the coumarin maximum absorption (λ ex = 342 nm), either via the fluorol-7ga or directly to the perylene bisimide acceptor and this results in an overall FRET efficiency of over 95%. 48 The emission spectra show that both irradiation of the coumarin-2 or of the fluorol-7ga results in efficient transfer of the energy to the acceptor, resulting in a 6.9 fold increase in acceptor emission intensity compared with direct coumarin-2 excitation (i.e. compared to direct acceptor excitation at λ ex = 555 nm) (Figure 1.35 B), which is a good demonstration of the antenna effect. A B N N N N N N N N N N N N N N N N N N Figure 1.35 A) D-A dendrimer with two donor types and one perylene bisimide based acceptor. B) Emission spectra of the dendrimer shown on the left, excited into: the coumarin-2 (λ ex = 342 nm), into the fluorol-7ga (λ ex = 418 nm) and into the perylene bisimide (λ ex = 555 nm). The group of Müllen synthesized a large rigid polyphenylenic dendrimer with a central terrylenediimide (TDI) core, which is surrounded by perylenemonoimide (PMI) and naphthalenemonoimide (NMI) chromophores (Figure 1.36A). The combined absorption spectra cover nearly the entire visible spectrum. The absorption and emission spectra overlap in such a way that allows for efficient Förster energy transfer from NMI to PMI to TDI (Figure 1.36B). Whether the energy from NMI is transferred to TDI in a stepwise manner (via PMI) or directly could not be determined; most likely both processes compete

38 Excited states: Theory and Applications A B Figure 1.36 A) Polyphenylenic dendrimer with terrylenediimide (TDI), perylenemonoimide (PMI) and naphthalenemonoimide (NMI) chromophores. B) Absorption and emission spectra of the tree separate chromophores (i.e. NMi, PMI and TDI) in the dendrimer. Even though the rigidity of the branches makes determination of the inter-chromophoric distances easier, the fact that the branches are built from directly connected phenyl groups could compromise these distances due to conjugation or overlap of wavefunctions, which could affect the inter-chromophoric distance, by transfer via through bond, i.e. Dexter type interactions. 1.8 Tuning excited state processes Unimolecular Tuning the excited state processes in a single molecule is often more difficult then in a multicomponent system, since the required properties have to be present in one molecule and cannot be obtained by combining two components with different properties. Molecules that have a π-electron system containing heteroatoms have a large change, directional and/or in strength, in their dipole upon going to the excited state. This is due to internal charge transfer and this redistribution of charge has an influence on the solvent molecules surrounding the chromophore. 55 The reorientation of the dipoles of the solvent molecules after excitation causes a change in energy of the thermally-equilibrated excited 37

39 Chapter 1 (THEXI) state and thereby causes a shift in the emission maximum (Figure 1.37A). Also molecules with a π-electron system that contain both electron donating and withdrawing substituents often show an increased response to a change in solvent polarity (Figure 1.37B). A B N CF 3 Figure 1.37 A) rigin of the shift in the emission maximum due to a change of the energy of the THEXI state. B) Example of a chromophore (i.e. Coumarin 153) with both electron donating and withdrawing substituents that exhibits high sensitivity to solvent polarity. 50 The group of Barton reported that ruthenium complexes with ligands that have the ability to intercalate in DNA (e.g., [Ru(bpy) 2 dppz] 2+ or [Ru(phen) 2 dppz] 2+, Figure 1.38) show a remarkable increase in the QY of emission upon addition of DNA. 51 This effect is explained by the shielding effect the DNA has on the intercalated dppz ligand. The DNA prevents the formation of hydrogen bonds of the solvent water with the nitrogens in the dppz ligand, thereby raising the energy level of the MLCT state to above that of the emissive MLCT state and preventing the fast (~ 3 ps) internal conversion from MLCT to MLCT, and thereby increasing the QY of emission of the complex. 52,53 38

40 Excited states: Theory and Applications A B Figure 1.38 A) Structure of [Ru(phen) 2 dppz] 2+ B) Excited state diagram of [Ru(phen) 2 dppz] 2+ in H 2 without and with DNA present. Another possibility for tuning unimolecular properties are changes that can be obtained by the energetic inversion of two states, e.g., from nπ* (non-emissive in solution) to ππ* (emissive) or vice versa. 5 This also holds for pyrene-1-carboxaldehyde (Figure 1.39A). In apolar solvents, such as n-hexane, there is little emission from the nπ* excited state, which is the LUM in apolar solvents. 54,55 A B Figure 1.39 A) pyrene-3-carboxaldehyde. B) Inversion the nπ* and ππ* states of pyrene-1-carboxaldehyde upon change from an apolar (i.e. n-hexane) to a hydrogen bonding (i.e. methanol) solvent, leading to a 100 fold increase in the quantum yield of emission. However, when the solvent is changed to a solvent that is capable of H-bond donation such as MeH, the energy of the nπ* state is increased to above that of the emissive ππ* state (Figure 1.39B), which now becomes the LUM and an up to one hundred fold increase in the QY of fluorescence was observed Tuning inter-component energy transfer With inter-component energy transfer the distinction is made that the energy transfer takes place between two parts of a molecule which can be seen as separate units (i.e. that can be 39

41 Chapter 1 addressed separately) and are connected by non-conjugated (e.g., aliphatic) spacers. The energy transfer between two components can be tuned by either making the donor or acceptor less or more suitable for energy transfer with the other component or by providing an alternative pathway, e.g., an additional suitable component for energy transfer. A The group of Irie developed a fluorescent switch based on a diarylethene and an anthracene-based fluorophore which are connected by a non-conjugating adamantyl spacer (Figure 1.40). 56,57 In the open form the molecule shows fluorescence at λ max = 503 nm. Upon ring-closing the diarylethene unit changes its absorption spectrum and now absorbs at longer wavelength, λ max = 618 nm. This absorption is lower in energy than that of the anthracene fluorophore and is able to quench the excited state, changing the QY of anthracene fluorescence from 0.73 to By embedding this fluorescent switch in a polymer matrix it was possible to open and close individual molecules. B Figure 1.40 A) The fluorescent switch reported by the group of Irie. B) Energy levels of the two components in the fluorescent switch in the open and closed form. 56 The group of Tian used the same approach by connecting two naphthalimide-based chromophores to a diarylethene via non-conjugated spacers to obtain very efficient fluorescence switching, which is reversible over several cycles also (Figure 1.41). 58 The fluorescent switches were embedded in PMMA and could be used for reversible pattern generation and therefore, combined with their stability, could be used for data storage. 40

42 Excited states: Theory and Applications A B Figure 1.41 A) Fluorescent switch studied by the group of Tian, consisting of a diarylethene unit and two naphthalimide chromophores. B) The switching of fluorescence intensity from the open ( ) state to the closed ( ) state, inset: reversibility of the switching over three cycles. The first literature example of energy transfer modulation between two components by a third unit is by the group of Effenberger. 59 They used a triad consisting of an anthracene donor and a coumarin acceptor combined with a photochromic fulgimide that can undergo a reversible ring closure reaction (Figure 1.42). In the open form the coumarin is the lowest energy acceptor and upon irradiation into the anthracene band at λ irr = 258 nm emission from the coumarin is observed at λ em = 480 nm. Irradiation at λ irr = 366 nm lead to ringclosure of the fulgimide and to the appearance of a new absorption band between λ = 450 and 600 nm. Irradiation into the anthracene absorption band (λ irr = 258 nm) sees a considerable decrease in the fluorescence intensity of the coumarin, due to the lower energy absorption of the ring-closed fulgimide, which acts as an excited state quencher. Figure 1.42 A switchable energy transfer triad consisting of an anthracene donor, a photochromic fulgimide and a coumarin acceptor. The perylene bisimide based dyad shown in Figure 1.43A shows energy transfer under neutral or acidic conditions. However, under basic conditions the phenol group is deprotonated and almost no ET can take place due to rapid deactivation of the excited state 41

43 Chapter 1 A of the hydroxy substituted perylene bisimide. 60 An additional feature of this dyad is, that upon addition of Fe(III), which can coordinate to the succinamide bridge, the fluorescence is fully quenched under neutral conditions. If acid is simultaneously added, the succinamide bridge is protonated and the Fe(III) then coordinates to the hydroxyl-substituted perylene bisimide and only fluorescence from the phenyl substituted perylene bisimide is observed. Different combinations can be made with the three additives (i.e. Fe(III), acid and/or base), each combination giving different emission intensities and quantum yields (Figure 1.43B). B Figure 1.43 A) Dyad of two differently substituted perylene bisimides. B) Fluorescence spectra of the dyad in THF excited at λ = 330 nm with different additives, which can be a THF solution of (a) Fe (Cl 4 ) 3, (b) TFA and (c) Et 3 N or a combination thereof (see graph). ther examples of energy transfer modulation/tuning are discussed in several recent reviews Tuning inter-component electron transfer As with energy transfer, electron transfer can also be tuned, with the principle being similar to energy transfer: the formation or removal of a suitable donor and/or acceptor. Many examples of sensors for either protons and/or ions 61 have been introduced, especially by the group of de Silva. Sensors for ions, protons and even temperature, have been reported. 61 The molecule shown below is an example of a molecular AND logic gate (Figure 1.44B) and fluorescence is being observed only if both inputs are present. 62 Without the inputs present both sensors, the amine and the crown-ether substituted phenyl, are able to donate an electron to the excited cyano-naphthalene fluorophore via photoinduced electron transfer, thereby quenching its fluorescence. This cannot occur if both H + and Na + are present; these ions change the energy level of the protonated amine and the chelating 42

44 Excited states: Theory and Applications crown ether in such a manner that they are no longer able to donate an electron and increased emission from the cyano-naphthalene is observed. A B H + Na + utput Figure 1.44 A) Molecular AND logic gate with two inputs (H + and Na + ) and one output (fluorescence). B) A truth table with the outcome of the AND logic. A diarylethene-porphyrin-fullerene triad has been reported by the group of Gust, 63 that has the ability to switch off an internal photoinduced electron transfer process between the porphyrin and the fullerene unit by ring opening or closing the third component; a diarylethene (Figure 1.45). As observed for the systems of Irie and Tian (vide supra), upon ring closing of the diarylethene an alternative pathway is provided, which quenches the energy of the porphyrin excited state. Figure 1.45 Diarylethene-porphyrin-fullerene triad capable of electron transfer switching. 43

45 Chapter 1 This example and many interesting and complex systems, either shown in previous paragraphs or found elsewhere in literature, show that there are numerous possibilities, however there are still many fundamental issues that need to be addressed. Understanding of the fundamental processes involved in energy transfer in donor acceptor systems and in particular tuning of energy transfer in such systems requires high photochemical stability. Designing systems which fulfil all criteria from scratch and attempting to predict the properties and behaviour of individual components in multicomponent systems is a major challenge on which the research described in this thesis is focussed. 1.9 Thesis outline The overall goal of this thesis is to show that a rational approach towards designing these systems, capable of tuneable/controllable energy transfer, is possible, and that the complexity in photophysical properties which arise from intercomponent interactions can be approached in a rational sense. In Chapter 2 the design and synthesis of a tetra-coumarin-perylene bisimide system is described. In this case the primary aim is to design a multi-donor single acceptor system which is stable under multi-photon excitation conditions and which shows efficient intramolecular energy transfer. In Chapter 3 a first attempt at designing and building a system in which the energy transfer between a donor and acceptor component can be controlled. This builds on the work presented in chapter 2. The results obtained for the controllable energy transfer system highlight several design issues that need to be addressed. In developing more robust systems the components used in Chapter 2 were modified and/or exchanged for more promising alternatives. In Chapter 4 several other components are investigated and lessons are learnt from the unexpected consequences that the variation of a single component can have for the system as a whole. In Chapter 5 the experience gained in the preceding chapters is combined into a triad system that shows many of the properties, including robustness and efficient energy transfer modulation, that were deemed essential to achieve the goal of tuneable energy transfer. 44

46 Excited states: Theory and Applications Theoretical chemistry can be a useful tool in designing systems and in predicting the properties and behaviour of the molecules, which were observed and described, for example, in Chapters 2 and 5. In Chapter 6, the properties of several molecules from these chapters were calculated using DFT and the results compared to the empirical results obtained References 1 C. L. van Dover The Ecology of Deep-Sea Hydrothermal Vents, 2000, Princeton University Press. ISBN (a) X. Hu, A. Damjanovic, T. Ritz and K. Schulten, Proc. Natl. Acad. Sci. U.S.A., 1998, 95, (b) G. McDermott, S.M. Prince, A.A. Freer, A.M. Hawthornthwaite-Lawless, M. Z. Papiz, R. J. Cogdell and N. W. Isaacs, Nature, 1995, 374, 517. (c) R. J. Cogdell, A. T. Gardiner, A. W. Roszak, C. J. Law, J. Southall and N. W. Isaacs, Photosynth. Res. 2004, 81, (d) C. J. Law, A. W. Roszak, J. Southall, A. T. Gardiner, N. W. Isaacs and R. J. Cogdell, Mol. Membr. Biol. 2004, 21, (a) B. Regan and M. Graetzel, Nature, 1991, 353, (b) A. Hagfeldt and M. Graetzel, Acc. Chem. Res., 2000, 33, N. Armaroli and V. Balzani, Angew. Chem. Int. Ed., 2007, 46, (a) N. J. Turro, Modern Molecular Photochemistry, University Science Books 1991, Mill Valley, CA. (b) J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Plenum Press 1983, New York. (c) M. Montalti, A. Credi, L. Prodi and M. T. Gandolfi, Handbook of Photochemistry, 3 rd edition, CRC; The Science of Color. 2 nd Edition. S. Shevell, ed. ptical Society of America / Elsevier, ISBN R. Birge, Biochim. Biophys. Acta 1990, 1016, Emmission can give rise to colour also, however in that case the percieved colour is determined by the wavelength of the emitted photons. 9 (a) J. Franck, Trans. Far. Soc. 1926, 21, (b) E. Condon, Phys. Rev. 1926, 27, 640. (c) E. Condon, Phys. Rev. 1926, 28, (d) E. Condon, Phys. Rev. 1928, 32, M. Born and R.ppenheimer, Ann. Phys. 1927, 84, K. Huang and A. Rhys, Proc. R. Soc. London, Ser. A 1950, 204, e.g. a Rh-Cl vibration, a C-C-H vibration, etc. 13 M. Kasha, Discuss. Faraday Soc. 1950, 9, C. N. Banwell, Fundamentals of Molecular Spectroscopy, McGraw Hill Book Co. Ltd, May E. U. Condon and G. H. Shortley, The Theory of Atomic Spectra, Cambridge University Press, ISBN B. L. Feringa (ed.), Molecular Switches, Wiley-VCH, 2001, Weinheim. 45

47 Chapter 1 17 W. R. Browne, M. M. Pollard, B. de Lange, A. Meetsma, and B. L. Feringa, J. Am. Chem. Soc. 2006, 128, (a) Y. V. II'ichev, W. Ktihnle and K. A. Zachariasse, Chem. Phys. 1996, 211, (b) H. Meier, Angew. Chem., Int. Ed. Engl. 1992, 31, A. Momotake and T. Arai, J. Photoch. Photobio. C 2004, 5, (a) G. S. Hartley, J. Am. Chem. Soc. 1938, (b) I. Hauser Naturwissenschaften 1949, (a) B. L. Feringa, Acc. Chem. Res. 2001, 34 (6), (b) B. L. Feringa, R. A. van Delden, N. Koumura and E. M. Geertsema Chem. Rev. 2000, 100 (5), (a) W. F. Jager, J. C. de Jong, Ben de Lange, N. P. M. Huck, A. Meetsma and B. L. Feringa, Angew. Chem., Int. Ed. Engl. 1995, 34, (b) R. A. van Delden, J. H. Hurenkamp and B. L. Feringa, Chem. Eur. J. 2003, 9, R. A. van Delden, M. K. J. ter Wiel, H. de Jong, A. Meetsma and B. L. Feringa, rg. Biomol. Chem., 2004, 2, R. B. Woodward and R. Hoffmann, J. Am. Chem. Soc. 1965, 87; (a) H. Tian and S. J. Yang, Chem. Soc. Rev. 2004, 33, (b) M. Irie, Chem. Rev. 2000, 100, (a) J. Gan, J. L. Pozzo and F. Vogtle, Mol. Cryst. Liq. Cryst. 2005, 430, (b) J. J. D. de Jong, L. N. Lucas, R. M. Kellogg, J. H. van Esch and B. L. Feringa, Science 2004, 304, (c) L. N. Lucas, J. van Esch, R. M. Kellogg and B. L. Feringa, Chem. Commun. 2001, (a) S. Takami, L. Kuroki and M. Irie, J. Am. Chem. Soc. 2007, 129, (b) K. Uchida, M. Walko, J. J. D. de Jong, S. Sukata, S. Kobatake, A. Meetsma, J. van Esch and B. L. Feringa, rg. Biomol. Chem. 2006, 4, (c) M. Irie, S. Kobatake and M. Horichi, Science 2001, 291, (a) C. C. Corredor, Z.-L. Huang, K. D. Belfield, A. R. Morales and M. V. Bondar, Chem. Mater. 2007, 19, (b) J. Areephong, W. R. Browne, N. Katsonis and B. L. Feringa, Chem. Commun. 2006, (c) G. M. Tsivgoulis and J. M. Lehn, Angew. Chem., Int. Ed. Eng. 1995, 34, (a) T. Kudernac, S. J. van der Molen, B. J. van Wees and B. L. Feringa, Chem. Commun. 2006, (b) N. Katsonis, T. Kudernac, M. Walko, S. J. van der Molen, B. J. van Wees and B. L. Feringa, Adv. Mater. 2006, 18, (a) K. Matsuda and M. Irie, J. Photoch. Photobio. C 2004, 5, (b) F. M. Raymo, and M. Tomasulo, Chem. Soc. Rev. 2005, 34, (c) F. M. Raymo and M. Tomasulo, J. Phys. Chem. A 2005, 109, (d) F. M. Raymo and M. Tomasulo, Chem. Eur. J. 2006, 12, Y. Yokoyama, Chem. Rev. 2000, 100, G. Berkovic, V. Krongauz and V. Weiss, Chem. Rev. 2000, 100,

48 Excited states: Theory and Applications 33 (a) C. J. Cramer, Essentials of Computational Chemistry, Wiley, (b) W. Koch, and M. C. Holthausen, A Chemist's Guide to Density Functional Theory, Wiley-VCH, Weinheim, 2 nd ed., D. L. Dexter, J. Chem. Phys. 1953, 21, (a) T. Förster, Ann. Phys. 1948, 2, (b) T. Förster, Disc. Faraday Soc. 1959, 27, J. M. Haider, R. M. Williams, L. De Cola and Z. Pikramenou, Angew. Chem. Int. Ed. 2003, 42, E. V. Kuzmenkina, C. D. Heyes, and G. U. Nienhaus, Proc. Natl. Acad. Sci. 2005, 102(43), M. R. Wasielewski, J. rg. Chem. 2006, 71, G. J. Wilson, A. Launikonis, W. H. F. Sasse and A. W.-H. Mau, J. Phys. Chem. A 1997, 101, A. P. H. J. Schenning, J. van Herrikhuyzen, P. Jonkheijm, Z. Chen, F. Würthner and E. W. Meijer, J. Am. Chem. Soc. 2002, 124, J. F. Baussard, J. L. Habib-Jiwan and A. Laschewsky, Langmuir 2003, 19, D. K. Struck, D. Hoekstra and R. E. Pagano, Biochemistry 1981, 20, H. Zhang and D. M. Rudkevich, Chem. Commun. 2007, (a) X. C. Hu, A. Damjanovic, T. Ritz and K. Schulten, Proc. Natl. Acad. Sci. U. S. A. 1998, 95, (b) G. Mcdermott, S. M. Prince, A. A. Freer, A. M. Hawthornthwaitelawless, M. Z. Papiz, R. J. Cogdell and N. W. Isaacs, Nature 1995, 374, (c) R. J. Cogdell, A. T. Gardiner, A.W. Roszak, C. J. Law, J. Southall and N. W. Isaacs, Photosynth. Res. 2004, 81, (d) C. J. Law, A. W. Roszak, J. Southall, A. T. Gardiner, N. W. Isaacs and R. J. Cogdell, Mol. Membr. Biol. 2004, 21, S. K. Grayson and J. M. J. Fréchet, Chemical Rev. 2001, 101, (a) A. Adronov, S. L. Gilat, J. M. J. Fréchet, K. hta, F.V.R. Neuwahl and G.R. Fleming J. Am. Chem. Soc. 2000, 122, (b) S. L. Gilat, A. Adronov and J. M. J. Fréchet J. rg. Chem. 1999, 64, (a) M. R. Shortreed, S. F. Swallen, Z. Y. Shi, W. H. Tan, Z. F. Xu, C. Devadoss, J. S. Moore and R. Kopelman, J. Phys. Chem. B 1997, 101, (b) C. Devadoss, P. Bharathi and J. S. Moore, J. Am. Chem. Soc. 1996, 118, J. M. Serin, D. W. Brousmiche and J. M. J. Fréchet, Chem. Commun. 2002, M. Cotlet, T. Vosch, S. Habuchi, Tanja Weil, K. Müllen, J. Hofkens and F. De Schryver, J. Am. Chem. Soc. 2005, 127, G. Jones, W. R. Jackson, S. Kanoktanaporn and W. R. Bergmark, Photochem. Photobiol. 1985, 42, A. E. Friedman, J. C. Chambron, J. P. Sauvage, N. J. Turro and J. K. Barton, J. Am. Chem. Soc. 1990, 112,

49 Chapter 1 52 E. J. C. lson, D. Hu, A. Hörmann, A. M. Jonkman, M. R. Arkin, E. D. A. Stemp, J. K. Barton and P. F. Barbara, J. Am. Chem. Soc. 1997, 119, The model discussed is somewhat simplified and indeed recent studies suggest a dark 3 ππ* state is involved also, see for example: G. Pourtois, D. Beljonne, C. Moucheron, S. Schumm, A. Kirsch-De Mesmaeker, R. Lazzaroni and J.-L. Brédas, J. Am. Chem. Soc. 2004, 126, K. Kalyanasundaram and J. K. Thomas, J. Phys. Chem. 1977, 81, A. P. de Silva, H. Q. N. Gunaratne, T. Gunnlaugsson, A. J. M. Huxley, C. P. McCoy, J. T. Rademacher and T. E. Rice, Chem. Rev. 1997, 97, M. Irie, T. Fukaminato, T. Sasaki, N. Tamai and T. Kawai, Nature 2002, 420, T. Fukaminato, T. Sasaki, T. Kawai, N. Tamai and M. Irie, J. Am. Chem. Soc. 2004, 126, G. Y. Jiang, S. Wang, W. F. Yuan, L. Jiang, Y. L. Song, H. Tian and D. B. Zhu, Chem. Mater. 2006, 18, J. Walz, K. Ulrich, H. Port, H.C. Wolf, J. Wonner and F. Effenberger, Chem. Phys. Lett. 1993, 213, Y. Li, H. Zheng, Y. Li, S. Wang, Z. Wu, P. Liu, Z. Gao, H. Liu, and Daoben Zhu, J. rg. Chem. 2007, 72, (a) A. P. de Silva, S. Uchiyamab, T. P. Vance and B. Wannalerse, Coord. Chem. Rev. 2007, 251, (b) D. C. Magri, T. P. Vance and A. P. de Silva, Inorg. Chim. Acta 2007, 360, (c) J. F. Callan, A. P. de Silva and D. C. Magri, Tetrahedron 2005, 61, (d) A. P. de Silva, D. B. Fox, A. J. M. Huxley and T. S. Moody, Coord. Chem. Rev. 2000, 205, and references therein. 62 A. P. de Silva, H. Q. N. Gunaratne and C. P. McCoy, Nature (London) 1993, 364, P. A. Liddell, G. Kodis, A. L. Moore, T. A. Moore and D. Gust, J. Am. Chem. Soc. 2002, 124,

50 Chapter 2 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties The synthesis and characterization of a novel coumarin donor / perylene bisimide acceptor light harvesting system is reported, in which an energy transfer efficiency of >99% is achieved. Comparison of the excited state properties of the donor-acceptor system with model compounds revealed that although the photophysical properties of the perylene bisimide acceptor unit are affected considerably by the nature of the substituent at the imide positions and the solvent employed, through bond interaction between the donor and acceptor units is negligible. Energy transfer in the present system can be described as occurring via a through space (Förster) energy transfer mechanism. Careful consideration of the redox properties of the donor relative to the acceptor units allows for avoidance of potentially deleterious excited state electron transfer processes. Part of this work has been published in: J.H. Hurenkamp, W.R. Browne, R. Augulis, A. Pugžlys, P.H.M. van Loosdrecht, J.H. van Esch and B.L. Feringa, rg. Biomol. Chem. 2007, 5, R. Augulis, A. Pugžlys, J. H. Hurenkamp, B. L. Feringa, J. H. van Esch and P. H. M. van Loosdrecht, J. Phys. Chem. A, 2007, 111 (50),

51 Chapter Introduction Light-harvesting (LH) systems have attracted considerable attention in recent years due both to the importance of photosynthesis (e.g., plants, bacteria), 1 and to increasing pressure to find alternatives (e.g., solar energy) to hydrocarbon fuels. 2 Nature has served as a source of inspiration in understanding the basic requirements for building efficient LH systems, 1c,d in particular in mimicking aspects of the complex architecture of the light harvesting complexes PS I and PS II. The spectroscopic properties of the individual components must be complimentary to enable efficient energy transfer, which makes selection of the separate components a central design issue. In developing synthetic light harvesting systems it is essential that the components employed are compatible energetically not only for efficient energy transfer but also to avoid potentially deleterious photochemistry, e.g., irreversible photoinduced electron transfer. 3 Dendritic systems, i.e. large regularly branched molecules, 4 are of special interest as candidates in LH systems, as demonstrated by the groups of Balzani, 5 Fréchet 6 and Wasielewski. 7 Dendrimers offer the possibility of arranging donor and acceptor units and of controlling communication between these chromophores, and it is this approach, which is taken in the present study. Recently we reported a tetra-coumarin-porphyrin based dendritic system, which shows efficient intramolecular energy transfer. 8 However, the 7-methoxycoumarin-3-carboxylic acid selected for the tetra-coumarin-porphyrin system showed a decreased quantum yield of fluorescence when coupled to an amine, due to direct conjugation of the amide with the coumarin double bond. Furthermore, the porphyrin acceptor proved to be unstable under the intense irradiation conditions required to saturate the system. Here we report the design and photophysical characterisation of a coumarin donor / perylene bisimide acceptor system, which enables efficient light harvesting (Figure 2.1). The perylene bisimide core is an excellent alternative for the porphyrin acceptor due to its high fluorescence quantum yield, redox properties, stability and the ability of perylene bisimides to engage in both electron and energy transfer processes. 7,9 These properties make them attractive for application in photonic devices and as substitutes for inorganic phosphorescent systems. Indeed, the similarity of both electronic and redox properties of perylene bisimides with the paradigm complex [Ru(bpy) 3 ] 2+ is remarkable. 10 The planarity of perylene bisimides enables the formation of H- and J- aggregates. 11 Such aggregation behaviour is advantageous for supramolecular systems 12 in achieving gelation 13 and in liquid crystalline behaviour

52 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties However, in studying unimolecular processes, such as energy and electron transfer in dendritic systems, aggregation is less desirable. In recent years, substitution in the bay area of the perylene bisimide has proven to be effective in inhibiting aggregation and as a consequence increasing solubility dramatically. Substitution in the bay area (i.e. on the 1, 6, 7 and 12 position of the perylene bisimide) can be employed to connect the perylene bisimide unit to donor units also, 15 thereby combining increased solubility with increased functionality. In the present contribution substitution at the bay area is employed, as mentioned above, to facilitate solubility. Whilst other substituents, specifically energy donor components, are introduced at the bisimide positions to provide the desired functionality. The perylene bisimide core has been employed sucessfully in single- and multistep donor acceptor arrays and has been shown to be a good acceptor with high stability as demonstrated by the groups of Fréchet, 6 Müllen, 16 De Schryver 16 and Würthner. 17 Figure 2.1Schematic representation of the convergent approach to the construction of the coumarin-perylene bisimide donor-acceptor system. The choice of donor unit is critical for the study of energy transfer in antenna systems, specifically Förster resonance energy transfer. Importantly, the redox properties of the donor and the acceptor must not facilitate excited state electron transfer between donor and acceptor. 9c To match these requirements, a highly fluorescent and stable, coumarin based, donor was selected. 1b The 7-methoxycoumarin-3-acetic acid 1 employed in this study was chosen to be compatible with the perylene bisimide core, in terms of both its electronic and 51

53 Chapter 2 redox properties. Furthermore, the carboxylic acid functionality facilitates the use of amide chemistry. Another important consideration lies in the properties of the bridging unit between the donor and acceptor components since, in studying Förster energy transfer, the bridging unit should not allow through bond energy transfer processes to occur. Therefore, piperazine and amides were used as spacer groups to disrupt through bond electronic communication between the donor and acceptor units. The polarity of the amide bonds provides an opportunity to examine the effect of solvent polarity on the conformation of the dendritic structure and also its effect on the energy transfer efficiency. In this paper we present the design and synthesis of an efficient donoracceptor system (Figure 2.1), together with a photophysical investigation of both the donor and acceptor components and of the assembled donor acceptor system. 2.2 Results and Discussion The construction of the donor-acceptor systems employs sequential amide coupling steps in a convergent manner. The coumarin donor units were connected to branching unit 3 via two consecutive amide coupling reactions (Figure 2.2). First the acetic acid functionalized coumarin 1 was connected to a mono-boc piperazine, after which the protecting group was removed to give the amine functionalized coumarin 2. The functionalized coumarin 2 was then coupled to 5-(N-Boc-amino)isophthalic acid using 4-(4,6-dimethoxy-1,3,5-triazin-2- yl)-4-methylmorpholinium chloride (DMTMM) with an overall yield of 20%, after which the protecting group was removed yielding 5, which was used without further purification. Figure 2.2 Synthesis of the dicoumarin branch 5, a) CDI, CH 2 Cl 2, N-Bocpiperazine; b) CH 2 Cl 2, CF 3 CH, 4 h; c) DMTMM, THF 18 h; d) CH 2 Cl 2, CF 3 CH, 4 h. 52

54 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties The core unit was prepared using literature procedures. 18 The perylene bisanhydride 6 was transformed to the n-butyl bisimide 7 by condensation with n-butylamine. The bisimide was tetra-chlorinated using sulfuryl chloride in nitrobenzene to yield 8. Aromatic substitution using 4-tert-butylphenol provided the tetraphenol substituted perylene bisimide 9. The bisimide was saponified with potassium hydroxide to give the tetra(4-tertbutylphenoxy)substituted bisanhydride 10 after acidic workup in 46% yield over 4 steps. R 1 R 1 R 1 N N N a) b) Cl Cl c) R 2 R 2 d), e) Cl Cl R 2 R 2 R 2 R 2 R 2 R 2 N N N R 1 R 1 = n-butyl R 1 R 1 R 2 = Figure 2.3 Synthesis of the core acceptor unit 10 a) n-butylamine, quinoline, 220 o C, 6 h; b) S 2 Cl 2, I 2, PhI, PnN 2, reflux, 80 o C, 20 h; c) 4-tert-butylphenol, K 2 C 3, NMP, 130 o C, 3 d; d) KH, H 2, t-buh, reflux, 24 h.; e) HCl (aq). Condensation of 10 with an excess of 5 provided, after purification, the tetra coumarin perylene bisimide 12c in 16% yield. The model compounds 12a and 12b were synthesized following similar procedures (Figure 2.4). R R + NH 2 a) R R b) + 5, 12c N NH ,11 N 12a 11,12b R=H R= R = N N N R 12a-c R Figure 2.4 Synthesis of 12a-c a) toluene, 120 o C, 5 d; b) DMA, 120 o C, 5 d. The compounds were purified by column chromatography and characterized by 1 H and 13 C NMR spectroscopy and MALDI-TF mass spectroscopy (see experimental section for details). 53

55 Chapter Electronic properties The absorption spectra of the perylene bisimide model compounds and 12c in CH 2 Cl 2 are shown in Figure 2.5 and the data are summarized in Table 2.1. Examination of the spectra of the substituted perylene bisimides (Figure 2.5) reveals a bathochromic shift with decreasing electron donating strength of the substituent. The visible absorption spectrum (λ > 400nm) of 12b is almost identical to that of 12c. This indicates that the electronic influence of the substituent is limited to the bridging unit itself and any influence of the coumarin component observed is unlikely to be through bond in character. Table 2.1 Absorption and emission spectra of 4, 9 and 12a-c. a Absorption λmax / nm (10 3 ε / cm -1 M -1 ) Emission λ max / nm (Φ fl ) Lifetime τ / ns b 1a 322(18.3) 393 d (0.46 d ) 1.4 d 4 324(31.3) 394 d (0.50 d ) 1.5 d 9 266(40.6), 286(49.6), 451(16.7), 608 e (0.66 i ) 6.7 f 539(26.7), 577(43.1) 12a 265(43.0), 290(43.6), 452(16.9), 612 e (0.59 i ) 6.5 f 540(28.1), 580(45.7) 12b 289(41.9), 455(15.7), 544(28.1), 616 e (0.57 i ) 6.3 f 584(44.6) 12c 295(63.6), 322(61.5), 458(14.5), 546(25.7), 587(41.9) 394 d, 618 d (<0.03, d, g 0.47 h, i ) 5.9 f c c 294, 321, 455, 544, d, 619 d (n.d.) n.d. a recorded in CH 2 Cl 2 at RT. b emission lifetime, experimental uncertainty ~ 2.5%. c 2 : 1 mixture. d λ ex = 322 nm. e λ ex = 450 nm. f λ ex = 420 nm. g residual coumarin emission. h direct excitation of the perylene bisimide unit. i λ ex = 539 nm. Emission spectra (λ ex = 450 nm) of the model compounds (Figure 2.6) and 12c show a trend analogous to that observed in the absorption spectra; a slight bathochromic shift is observed with decreasing electron donating strength of the substituent (Table 2.1). As a consequence of the similarity of the redox (vide infra), absorption and emission properties of 12b and 12c, the former compound was chosen as a model for the core acceptor unit in photophysical studies. 54

56 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties abs abs Wavelength (nm) Wavelength (nm) Figure 2.5 Absorption spectra of 9 ( ), 12a ( ), 12b ( ) and 12c ( ) in CH 2 Cl 2 at RT; Inset: expansion of the 560 to 600 nm region. As for the acceptor unit, a suitable model for the donor part must be identified in order to examine the energy transfer processes within 12c. Due to overlap of absorption in 12c of the coumarin and perylene bisimide components confirmation of the suitability of 4 as a model compound was obtained from comparison of the spectrum of a 2 : 1 molar mixture of 4 and 12b with the absorption spectrum of 12c (Figure 2.7). The absorption spectrum of the model mixture of 12b and 4 is almost identical to that of 12c, with no significant shifts in either the red or the blue region, indicating little or no communication between the coumarins and perylene bisimide is present. The location of the maximum at λ = 324 nm and the intensity of the coumarin 4 absorption coincide with the absorption of the coumarin component of 12c. Similarly the emission λ max of the model coumarin and the residual emission in 12c system compare well (vide infra). 55

57 Chapter 2 1.0M 800.0k Intensity 600.0k 400.0k 200.0k Wavelength (nm) Figure 2.6 Emission spectra of 9 ( ), 12a ( ) and 12b ( ) in CH 2 Cl 2 at RT (λ ex = 450 nm), showing the decrease in the relative quantum yield of emission. Spectra were corrected for absorption. 4 abs Wavelength (nm) Figure 2.7 Absorption spectra of 12c ( ), the 1:2 mixture of 12b and 4 ( ), and the spectrum of 4 ( ), spectra were recorded in CH 2 Cl 2 at RT, the spectra of 12c and the mixture of 4 and 12b are normalized to the perylene bisimide absorption maximum ~ 590 nm. 56

58 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties The fluorescence lifetimes of the model perylene bisimide compounds 9, 12a and 12b (λ ex = 420 nm), substituted with butyl, phenyl and bridge model 11, respectively (Figure 2.4), show a modest but significant decrease in excited state lifetimes and are accompanied by a concomitant decrease in the fluorescence quantum yield (Table 2.1). The trend observed is comparable to the bathochromic shift observed in both absorption and emission spectra. The decrease in emission lifetime is not unexpected and can be rationalised on the basis of the energy gap law, 19 which predicts a decrease in the nonradiative emission lifetime with a decrease in ground / excited state energy gap Redox properties The redox potentials of perylene bisimides (9 and 12a-c) were measured by differential pulse and cyclic voltammetry in CH 2 Cl 2 / 0.1 M TBAPF 6 between +1.6 V and 1.2 V vs. SCE. At positive potentials a reversible one-electron redox peak (E ½ V to V vs. SCE) is observed for all perylene bisimide based compounds (Figure 2.8). Table 2.2 Redox potentials of 9 and 12a-c. Compound x. (V) a Red. (V) a V (V) b a b c a differential pulse voltammetry, b V is the separation between E ½ of the first oxidation and the first reduction. (V vs. SCE, in CH 2 Cl 2 / 0.1 M TBAPF 6 ). Similarly at negative potentials two reversible one-electron redox waves are observed, with the separation between the first process (E ½ V to V vs. SCE) and the second process (E ½ V to V vs. SCE) showing only minimal dependence (~ 5 mv) on the substituent employed at the imide position. Comparison of 12c with the model perylene bisimide systems confirms that both the first oxidation and the first and second reduction processes are based on the perylene bisimide core. 9b,18b Within the potential window examined (+1.6 to -1.6 V vs SCE) no redox activity was observed for the coumarin or amide components. 57

59 Chapter 2 Current (A) 70.0µ 60.0µ 50.0µ 40.0µ 30.0µ 20.0µ 10.0µ 0.0 * * * * 9 12a 12b 12c -10.0µ Voltage (V) Figure 2.8 Cyclic voltammetry of 9 and 12a-c in CH 2 Cl 2 / 0.1 M TBAPF 6 vs. SCE. The current is offset for clarity.* indicates the open circuit potential. Table 2.3 Redox potentials a of 9 and 12a-b in dichloromethane and chloroform. Dichloromethane Chloroform E 1/2 V (V) E 1/2 V , -0.77, , -0.87, a 1.28, -0.72, , -0.80, b 1.31, -0.66, , -0.72, c 1.33, -0.64, a determined by differential pulse voltammetry. (V vs. SCE, in respective solvent / 0.1 M TBAPF 6 ). The correlation between the HM-LUM energy gap determined electrochemically and spectroscopically is well established, 21 providing both the oxidation and reduction involve the same chromophoric unit. The separation ( V) between first oxidation and first reduction decreases with decreasing electron donating ability of the substituent. This indicates a decrease in the HM-LUM energy gap. For 12b and 12c, the similarity in the separation ( V) indicates a comparable influence on the perylene bisimide core by both imide substituents. Comparison of 9 and 12b indicates that electron withdrawing groups destabilize the LUM to a greater extent than the HM, and thereby show a decrease of the HM-LUM gap. 58

60 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties 2.5 Solvent dependence of spectroscopic properties Absorption and emission spectra of all compounds were obtained in acetone, dichloromethane and chloroform. For the coumarin model 4 both absorption and emission spectra were identical in dichloromethane and chloroform (aggregation was observed in acetone) A.U Wavelength (nm) Figure 2.9 Absorption and emission spectra of 12b in acetone ( ), CH 2 Cl 2 ( ) and chloroform ( ) at RT (λ ex = 450 nm, spectra normalized for clarity). For the perylene bisimide containing compounds, however, considerable solvent dependence in both the emission and absorption spectra was observed (Figure 2.9 and Table 2.4). The spectra show a red-shift in both absorption and emission between acetone, dichloromethane and chloroform. The origin of this solvent dependence can be assigned to the interaction of the imide carbonyls of the perylene bisimide unit with the solvent and is similar to the effect observed upon substitution at the imide nitrogen (vide supra). This solvent effect highlights the influence of the imide carbonyls on the HM-LUM levels of the perylene bisimide core. The nature of the interaction, vis à vis HM vs. LUM stabilisation, can be estimated from the redox potentials of the compounds in these solvents. 59

61 Chapter 2 The redox properties of the perylene bisimide compounds in dichloromethane and chloroform are shown in Table 2.3. It is apparent from the shift in first reduction potential (vs. SCE) going from dichloromethane to chloroform for each separate compound (9, 12a-c) that the LUM level is affected by the solvent to a much greater extent than the HM level. A key aspect of energy transfer is the relative importance of through bond and through space contributions to the overall interaction between the donor and acceptor units. Typically the through bond contribution is estimated by comparison of the spectroscopic properties of the separate donor and acceptor component with those of 12c. It is clear that although minor differences in the absorption and emission spectra between the 2 : 1 mixture (of 4 and 12b) and 12c are observed, these differences are comparable to the effect of local solvent environment. Table 2.4 Solvent dependence of the absorption and emission maxima of the perylene bisimide emission (λ ~ 580 nm) of 9 and 12a-c. Absorption a λ max / nm) Emission a,b λ max / nm) Acetone CH 2 Cl 2 CHCl 3 Acetone CH 2 Cl 2 CHCl a b c a Measurements taken at RT, b λ ex = 450 nm. The molecular geometries and the molecular orbitals diagrams of 9, 12a and 12b were calculated using the hybrid Hartree Fock/density functional method (B3LYP/6-31G(d)). 22 The tert-butyl groups on the phenol bay substituent were replaced by methyl groups and for 12b the piperidine groups were replaced by dimethylamine substituents, this was done to increase symmetry and to reduce calculation cost, all replacement substituents were selected to have minimal impact on the electronic structure. Molecular orbital diagrams of 9, 12a and 12b show a considerable contribution of the imide carbonyls to both the HM and LUM of the model compounds (for details see Chapter 6). It is clear that the imide substituents are not involved in the frontier orbitals of the perylene compounds examined despite holding a strong influence over the electronic character of the carbonyl bonds. This 60

62 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties suggests that the effect of both the solvent and the imide substituents on the electronic properties is due to the perturbation of electron density on the carbonyl groups. 2.6 Energy transfer The absorption spectrum of the coumarin substituted branch 4 in CH 2 Cl 2 at room temperature (RT) shows an absorption maximum at λ = 324 nm (Figure 2.7). At this wavelength the absorption of the perylene bisimide model shows low absorption (Figure 2.5), allowing for direct excitation of the donor unit with minimal direct excitation of the perylene bisimide. In order for energy transfer to occur via a Förster energy transfer mechanism, overlap of the donor emission spectrum and acceptor absorption spectrum is required. The emission of 4 in CH 2 Cl 2 at RT shows a maximum at λ = 394 nm and overlaps with the absorption of the perylene bisimide acceptor unit (Figure 2.10) A.U Wavelength (nm) Figure 2.10 Absorption spectrum of 12c (abs: ) and absorption and emission spectra of 4 (abs:, fl:, λ ex = 322 nm) normalized to 12c absorption maximum at λ = 587 nm in CH 2 Cl 2 at RT. 61

63 Chapter 2 The emission spectra of a 2 : 1 mixture of 4 and 12b, and of 12c were measured in CH 2 Cl 2 at RT (Figure 2.11). Excitation at λ ex = 450 nm results in perylene bisimide emission of similar intensity at λ em = 616 nm for both systems (not shown). This confirms that the spectroscopic properties of the perylene bisimide core are unaffected by the covalent attachment of the coumarin. Similarly, upon excitation at λ ex = 322 nm, corresponding to the λ max of the coumarin absorption, both the mixture and 12c show emission at λ max = 394 nm (coumarin emission) and 616 nm (perylene bisimide emission). The intensity of the coumarin emission in both systems is very different, however. The reduced intensity of the coumarin emission, observed for the 12c, indicates quenching of the coumarin emission is taking place (the coumarin Φ fl decreases from 0.5 to < 0.03, Table 2.1). By contrast the intensity of the perylene bisimide emission in the 12c system is increased compared to that in the model system. Hence intramolecular energy transfer from the coumarin to the perylene bisimide core is taking place in 12c. 3.0x x x10 6 Intensity 1.5x x x Wavelength (nm) Figure 2.11 Emission spectra corrected for the absorption at λ ex = 322 nm of the model mixture (a 2:1 mixture of 4 (3*10-5 M) and 12b (1.5*10-5 M), λ ex = 322 nm, ), and 12c (0.5*10-5 M λ ex = 322 nm, ) in CH 2 Cl 2 at RT. The absence of intermolecular energy transfer was confirmed through excitation spectroscopy (Figure 2.12). The excitation spectrum of 12c matches the corresponding absorption spectrum closely and, importantly, shows the contribution of the coumarin absorption to the perylene bisimide emission in 12c. Comparison of the excitation spectra 62

64 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties of 12b and the 2 : 1 mixture of 4 and 12b shows that the free coumarin does not contribute to emission of the perylene bisimide model compound. Hence, it is unlikely that intermolecular energy transfer occurs in solution for 12c A.U Wavelength (nm) Figure 2.12 Excitation spectra of 12b ( ), 12c ( ), and a 2 : 1 mixture of respectively 4 and 12b ( ) in CH 2 Cl 2 at RT, normalized at the perylene bisimide absorption maximum at λ = 586 nm, the emission was monitored at λ = 610 nm. Preliminary time-resolved emission spectroscopy confirms that energy transfer from the coumarin donor units to the perylene bisimide acceptor core is fast (< 15 ps, Fig. 13). The presence of four donor coumarin units and the non-zero absorption of the perylene bisimide unit itself in the near UV region open the possibility that several components of the system can be pumped optically to an electronically excited state within the lifetime of the perylene bisimide acceptor excited state, resulting in potentially complex excited state behaviour. Under the low excitation intensity conditions employed here, however, statistically only one unit in the array is excited at any one time (i.e. either one of the four coumarin donor units or the perylene acceptor core itself). Hence the rapid decay of the coumarin emission and the concomitant rise time of the perylene bisimide component observed by time resolved spectroscopy provides a strong indication that energy transfer form the coumarin to the perylene bisimide is efficient. 63

65 Chapter 2 Figure 2.13 Time resolved emission spectroscopy of 12c in CHCl 3. irf instrument response function, 385 nm λ em (coumarin unit), 620 nm - λ em (perylene bisimide unit). Excitation wavelength λ ex = 325 nm. 2.7 Conclusions In understanding and characterising energy transfer processes it is essential that the models used for comparison with the dendritic system are suitable. The imide substituent is expected to have a negligible effect on the properties of the perylene bisimide core; however, it is clear from comparison of 9 with 12b that a significant shift in the absorption spectrum results from a change in the imide substituents. Differences between the absorption spectra of 12c and model compounds 9, 12a and 12b become less for the imidesubstituted models, which are structurally most similar to branch unit 4 used in 12c. This is quite evident when considering the trends observed in the absorption spectra, but can also be seen when comparing the redox properties. Hence 12b is the most suitable model compound for the acceptor unit of the donor-acceptor system 12c. The remaining differences (~2 nm shift) can be rationalized by considering the small differences in local environment, most likely caused by the branches, and are much smaller than differences observed with different solvents (i.e. changing from dichloromethane to chloroform results in a shift of about 10 nm). 64

66 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties These small differences suggest that through-bond orbital interaction is minimal if indeed it is present and hence through-bond energy transfer is unlikely. The model mixture of 4 and 12b also shows that the energy transfer is not due to trivial (intermolecular) energy transfer, since irradiation at the λ max of 4 shows no increase of the perylene bisimide emission compared to a solution of 12b irradiated at the same wavelength. Irradiation of 12c at the λ max of 4 shows a clear increase in fluorescence of the perylene bisimide, indicating transfer of energy from the coumarin donor to the perylene bisimide acceptor. This is confirmed by comparison of the excitation spectra of 12c and model mixture of 4 and 12b which show that the energy originates from the coumarin donor and not only from the residual perylene bisimide absorption at that excitation wavelength (the coumarin λ max ). In this chapter the synthesis and characterisation of a tetra coumarin donor - perylene bisimide acceptor array is described. Energy transfer from the donor units to the acceptor was found to be very efficient (>99%) and to take place via a through-space mechanism. In addition to the high efficieny of energy transfer, which is comparable to several related systems reported previously, 6,16,17 the four donor - one acceptor array shows high stability and its redox properties indicate that no undesired, and possibly deleterious, photoinduced electron transfer between the donor and acceptor takes place. Importantly recognition of the sensitivity of the acceptor to both solvent properties and the nature of the imide substituents is critical for excluding a through bond contribution to the energy transfer mechanism. An investigation of the picosecond energy transfer dynamics and the behaviour of the donoracceptor system under multiphoton excitation conditions has been reported elsewhere Experimental section Uvasol-grade solvents (Merck) were employed for all spectroscopic measurements. All reagents employed in synthetic procedures were of reagent grade or better, and used as received unless stated otherwise. N-Boc-piperazine, 24 5-(N-Boc-amino)isophthalic acid, 25 7, 18b 8, 18c and 9 18d,e were prepared according to literature procedures. 1 H NMR spectra were recorded at 400 MHz; 13 C NMR spectra at 50.3 or MHz. All spectra were recorded at ambient temperature, with the residual proton signals of the solvent as an internal reference. Chemical shifts are reported in ppm relative to TMS. CI and EI mass spectra were recorded on a Jeol JMS-600 mass spectrometer in the scan range of m/z with an acquisition time between 300 and 900 ms and a potential 65

67 Chapter 2 between 30 and 70 V. MALDI-TF spectra were recorded on an Applied Biosystems Voyager-DE Pro. UV/Vis absorption spectra (accuracy ±2 nm) were recorded on a Hewlett-Packard UV/Vis 8453 spectrometer. The fluorescence measurements were performed on a SPF-500C spectrofluorometer (SLM Aminco), and a Jobin-Yvon Fluorolog 3-22, the sharp features between λ = 450 and 500 nm in the excitation spectra are instrumental artefacts, the excitation and emission spectra are uncorrected for variations in lamp intensity and detector response. Sample concentration typically 10-5 M, spectra were recorded in 10 mm pathlength quartz fluorescence cuvettes. Quantum yields where determined using 9,10-diphenylanthracene (Φ fl = 0.88, EtH) 26 for 1a, 4 and for the residual coumarin emission of 12c (CH 2 Cl 2, λ ex = 322 nm). Rhodamine 101 (Φ fl = 1.00, EtH) 27 was used for compounds 9 and 12a-c (CH 2 Cl 2, λex = 539 nm). The absorbance of all measured solutions was kept < 0.1. Electrochemical measurements were carried out on a Model 630B Electrochemical Workstation (CHInstruments). Analyte concentrations were typically mm in anhydrous CH 2 Cl 2 containing 0.1 M TBAPF 6 (except where stated otherwise in the text). Unless otherwise stated, a Teflon-shrouded glassy carbon working electrode (CHInstruments), a Pt wire auxiliary electrode and SCE or non-aqueous Ag/Ag + ion reference electrode were employed. Reference electrodes were calibrated with 0.1 mm solutions of ferrocene (0.38 V versus SCE in 0.1 M TBAPF 6 / CH 3 CN). Solutions for reduction measurements were deoxygenated by purging with dry N 2 gas (presaturated with solvent) prior to the measurement. Cyclic voltammograms were obtained at sweep rates of between 10 mvs -1 and 50 Vs -1 ; differential pulse voltammetry (DPV) experiments were performed with a scan rate of 20 mvs -1, a pulse height of 75 mv and a duration of 40 ms. For reversible processes the half-wave potential values are reported; identical values were obtained from DPV and CV measurements. Redox potentials are ± 10 mv. Luminescence lifetime measurements were obtained using an Edinburgh Analytical Instruments (EAI) time-correlated single-photon counting apparatus (TCSPC) comprised of two model J-yA monochromators (emission and excitation), a single photon photomultiplier detection system model 5300, and a F900 nanosecond flashlamp (N 2 filled at 1.1 atm pressure, 40 khz) interfaced with a personal computer via a Norland MCA card. A 400 nm cut off filter was used in emission to attenuate scatter of the excitation light (337 nm). Excitation at λ = 322 and 420 nm corresponding to emission lines of the dinitrogen 66

68 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties flash lamp. The instrument response has a FWHM of 0.89 ns, 4096 channels were used for acquisition with a peak count of 1800 in the maximum channel. Data correlation and manipulation was carried out using EAI F900 software version Emission lifetimes were calculated using a single-exponential fitting function, Levenberg-Marquardt algorithm with iterative deconvolution (Edinburgh instruments F900 software). The reduced χ 2 (less then 1.1) and residual plots were used to judge the quality of the fits. Lifetimes are ± 5%. Picosecond time resolved spectra were recorded as described elsewhere. 23 (1) 7-Methoxycoumarin-3-acetic acid 2-Hydroxy-4-methoxybenzaldehyde (25 g, 0.16 mol) and succinic anhydride (50 g, 0.50 mol) were placed in a 250 ml three-necked round-bottom flask fitted with a reflux condenser. The solid mixture was heated on a metal bath to 90 o C and stirred for 30 min. The melt was then heated to 190 o C. Anhydrous disodium succinate (38 g, 0.23 mol) was added in small portions over 4 h. The hot melt was poured into 10% HCl(aq) and left overnight. The yellowish precipitate was filtered and washed with water till neutral. The residue was dissolved in 5% NaHC 3 (aq) and filtered. The filtrate was added to cold 15% HCl(aq) and left overnight. The precipitate was filtered and the residue washed with H 2, dried, and crystallised from H 2 /ethanol yielding brownish crystals (8.28 g, 35.8 mmol) in 21.8% yield. m.p o C. 1 H NMR (400 MHz, DMS-d6) δ = (s, 1H), 7.90 (s, 1H), 7.60 (d, J = 8.6 Hz, 1H), 7.01 (d, J = 2.4 Hz, 1H), 6.95 (dd, J = 8.6, 2.5 Hz, 1H), 3.85 (s, 3H), 3.45 (s, 2H) ppm. 13 C NMR (50 MHz, DMS-d6) δ = (s), (s), (s), (s), (d), (d), (s), (d), (s), (d), (q), (t) ppm. MS(EI) for C 12 H 10 5 m/z 234 [M + ], HRMS calcd for C 12 H 10 5 : , found: (1a) 7-Methoxy-3-(2-oxo-2-(N-Boc-piperazine)-1-yl-ethyl)coumarin N,N'-Carbonyldiimidazole (CDI) (1.4 g, 8.5 mmol) was added to a suspension of 7-methoxycoumarin-3-acetic acid (2.0 g, 8.5 mmol) in dry CH 2 Cl 2. The reaction mixture was stirred at RT under N 2 until C 2 evolution was complete and stirring was continued for another 30 min. N-Boc-piperazine (1.57 g, 8.5 mmol) was added and the reaction mixture was stirred under N 2 at RT overnight. The reaction mixture was extracted twice with 1M HCl(aq), twice with 5% NaHC 3 (aq) and once with brine. The solution was dried over Na 2 S 4 and the solvent evaporated. The crude mixture was purified using column chromatography (2% MeH in CH 2 Cl 2, Si 2, R f = 0.4) giving a yellow solid ( 3.13 g, 7.76 mmol) in 91.8% yield. m.p o C. 1 H NMR (400 MHz, CDCl 3 ) δ = 7.65 (s, 1H), 7.32 (d, J = 8.6 Hz, 1H), 6.80 (dd, J = 8.5, 2.4 Hz, 1H), 6.78 (d, J = 2.3 Hz, 1H), 3.83 (s, 3H), 3.58 (dd, J = 9.0, 5.0 Hz, 4H), 3.56 (s, 2H), (m, 2H), (m, 2H),

69 Chapter 2 (s, 9H) ppm. 13 C NMR (100.6 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (d), (d), (s), (s), (d), (d), 80.2 (s), 55.7 (q), 46.0 (t), 41.8 (t), 34.0 (t), 28.3 (q) ppm. MS(EI) for C 21 H 26 N 2 6 m/z 402 [M + ], HRMS calcd for C 21 H 26 N 2 6 : , found: General deprotection method for BC-protected amines The BC protected amine was stirred in a mixture of 1:1 CH 2 Cl 2 : CF 3 CH for 4 h. An equal amount of water was added and the mixture was neutralized by addition of solid NaHC 3, after which the aqueous layer was separated and the organic layer was washed with saturated NaHC 3 solution. Subsequently the organic layer was dried over Na 2 S 4 and the solvent evaporated. The resulting product was used in subsequent steps without further purification. (4) (3,5-Bis-{4-[2-(7-methoxy-2-oxo-2H-chromen-3-yl)-acetyl]-piperazine-1-carbonyl}- phenyl)-carbamic acid tert-butyl ester BC protected 7-methoxy-3-(2-oxo-2-(N-Boc-piperazine)-1-yl-ethyl)coumarin (1a) was deprotected using the general method described above and (2.0 g, 6.6 mmol) of the deprotected compound was suspended in THF with 5-(N-Boc-amino)isophthalic acid (0.66 g, 2.4 mmol). 4-(4,6-dimethoxy-1,3,5-triazin-2-yI)-4-methylmorpholinium chloride (DMTMM) (2.0 g, 7.2 mmol) was added and the suspension was stirred overnight. The solvent was evaporated and the crude mixture was purified using column chromatography (4% MeH in CH 2 Cl 2, Si 2, R f = 0.4), yielding a light yellow solid (0.87 g, 1.02 mmol, 43%) m.p o C. 1 H NMR (400 MHz, CDCl 3 ) δ = 7.66 (s, 2H), 7.53 (s, 2H), 7.34 (d, J = 8.6 Hz, 2H), 7.08 (s, 2H), 6.97 (s, 1H), 6.82 (dd, J = 8.6, 2.4 Hz, 2H), 6.79 (s, 2H), 3.84 (s, 6H), (m, 20H), 1.49 (s, 9H) ppm. 13 C NMR (50 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (s), (d), (s), (s), (d), (d), (s), (d), (s), (d), (d), 80.9 (s), 55.7 (q), 47.5 (t), 46.0 (t), 41.9 (t), 34.1 (t), 28.2 (q) ppm. MALDI-TF MS (MW =849.32) m/z = [M + Na + ]. (10) 1,6,7,12-tetrakis-[4'-tert-butylphenoxy]-3,4:9,10-perylenetetracarboxylic dianhydride 9 (3.77 g, 3.4 mmol) was suspended in tert-butyl alcohol (270 ml) in a 500 ml roundbottom flask fitted with a reflux condenser, followed by addition of water (13 ml ) and KH (26.8 g, 478 mmol). Under N 2 atmosphere the mixture was stirred and heated at reflux for 24 h. Subsequently the mixture was cooled down to RT and poured into 10% HCl(aq) (800 ml) and left overnight. The precipitate was filtered off, washed with water 68

70 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties and dried in a vacuum oven (40 o C, 200 mbar). The dark red compound was used without further purification (3.03 g, 3.1 mmol, 91.2 %). 1 H NMR (400 MHz, CDCl 3 ) δ = 8.2 (s, 4H), 7.2 (d, J = 8.8 Hz, 8H), 6.8 (d, J = 8.7 Hz, 8H), 1.3 (s, 36H) ppm. 13 C NMR (100.6 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (d), (s), (d), (s), (d), (s), (s), (q) ppm. MALDI-TF MS (MW = ) m/z = [M + ]. (12a) N,N'-Bisphenyl-1,6,7,12-tetrakis-[4'-tert-butylphenoxy]-3,4:9,10- perylenetetracarboxylic diimide. In a 100 ml roundbottom flask fitted with a reflux condenser 10 (100 mg, 0.10 mmol), aniline (1.0 g, 11 mmol) and toluene (10 ml) were heated at 120 o C under a N 2 atmosphere for 5 d, after which the solvent was evaporated. The crude compound was dissolved in CH 2 Cl 2 and extracted twice with 1M HCl (aq) and once with brine. The organic layer was dried and evaporated in vacuo. The crude product was purified using column chromatography (CH 2 Cl 2 /cyclohexane 3:1, Si 2 ) giving the pure compound 12a (67 mg, mmol, 59 %). 1 H NMR (400 MHz, CDCl 3 ) δ = 8.19 (s, 4H), (m, 10H), 7.18 (d, J = 8.8 Hz, 8H), 6.80 (d, J = 8.8 Hz, 8H), 1.22 (s, 36H) ppm. 13 C NMR (100.6 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (s), (d), (d), (d), (d), (s), (s), (d), (s), (d), 34.3 (s), 31.4 (t) ppm. MALDI-TF MS (MW = ) m/z = [M + ]. (12b) N,N'-(3,5-(piperidine-1-carbonyl)-phenyl)-1,6,7,12-tetrakis-[4'-tertbutylphenoxy]-3,4:9,10-perylenetetracarboxylic diimide. In a 100 ml round-bottom flask fitted with a reflux condenser 10 (100 mg, 0.10 mmol), 11 (after deprotection of 14 via the general method described above) (126.2 mg, 0.4 mmol) and dimethylacetamide (20 ml) were heated at 130 o C under a dinitrogen atmosphere. After 5 d. the reaction mixture was poured into 1M HCl(aq) and left overnight. The precipitate was filtered off and the residue washed with 1M HCl(aq) and water. The residue was taken up into MeH/CH 2 Cl 2 1:1, dried over Na 2 S 4 and the solvent evaporated. The crude compound was purified by column chromatography (3% MeH in CH 2 Cl 2, Si 2 ) giving the solid purple product (60 mg, mmol, 38 %) 1 H NMR (400 MHz, CDCl 3 ) δ = 8.25 (s, 4H), 7.51 (s, 2H), 7.35 (s, 4H), 7.22 (d, J = 8.8 Hz, 8H), 6.82 (d, J = 8.7 Hz, 8H), 3.66 (s, 10H), 3.38 (s, 10H), (m, 20H), 1.27 (s, 36H) ppm. 13 C NMR (50 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (s), (s), (s), (d), (d), (d), (s), (s), (d), (s), (d), 48.9 (t), 43.4 (t), 34.4 (t), 31.4 (q), 26.4 (t), 25.6 (t), 24.5 (t) ppm. MALDI-TF MS (MW = ) m/z = [M + ]. 69

71 Chapter 2 (12c) N,N'-(3,5-Bis-{4-[2-(7-methoxy-2-oxo-2H-chromen-3-yl)-acetyl]-piperazine-1- carbonyl}-phenyl)-1,6,7,12-tetrakis-[4'-tert-butylphenoxy]-3,4:9,10- perylenetetracarboxylic diimide. In a 100 ml roundbottom flask fitted with a reflux condenser 10 (100 mg, 0.10 mmol), 5 (225 mg, 0.30 mmol) (obtained after deprotection of 4 via the standard procedure) and dimethylacetamide (20 ml) were heated and stirred at 120 o C under dinitrogen. After 5 days the reaction was judged complete (by TLC, 10% MeH in CH 2 Cl 2, Si 2 ). The reaction mixture was poured into 50 ml 1M HCl(aq) and left to stand overnight. The precipitate was filtered off and the residue washed with 1M HCl(aq) and water. The residue was taken up into MeH/CH 2 Cl 2 1:1, dried over Na 2 S 4 and the solvent evaporated. The crude compound was purified by column chromatography (4% MeH in CH 2 Cl 2, Si 2 ) yielding the purple product 12c (40 mg, mmol, 16 %). 1 H NMR (400 MHz, CDCl 3 ) δ = 8.26 (s, 4H), 7.65 (s, 4H), 7.63 (t, J = 1.5 Hz, 2H), 7.43 (d, J = 1.3 Hz, 4H), 7.35 (d, J = 8.6 Hz, 4H), 7.23 (d, J = 8.9 Hz, 8H), (m, 16H), 3.86 (s, 12H), (m, 40H), 1.26 (s, 36H) ppm. 13 C NMR (100.6 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (s), (s), (s), (s), (d), (s), (s), (s), (d), (d), (d), (s), (s), (d), (s), (s), (d), (d), (d), (d), 56.8 (q), 48.7 (d), 47.1 (d), 43.3 (d), 35.4 (d), 35.2 (d), 32.5 (q), 30.7 (s) ppm. MALDI-TF MS (MW = ) m/z = [M + Na + ]. (13) 5-tert-Butoxycarbonylamino-isophthalic acid bis-(2,5-dioxo-pyrrolidin-1-yl) ester A suspension of 5-(N-Boc-amino)isophthalic acid (1.0 g, 3.5 mmol) and N- hydroxysuccinimide (886 mg, 7.7 mmol) in dry THF under a dinitrogen atmosphere was stirred at 0 o C. N,N'-dicyclohexylcarbodiimide (1.6 g, 7.7 mmol) was added and the solution was stirred overnight at RT. The suspension was filtered and the filtrate collected. The solvent was evaporated and the resulting solid was crystallised from 2-propanol yielding a white powder (853 mg, 1.77 mmol, 51%). m.p H NMR (400 MHz, CDCl 3 ) δ = 8.50 (s, 1H), 8.44 (s, 1H), 6.99 (s, 2H), 2.90 (s, 8H), 1.52 (s, 9H) ppm. 13 C NMR (50.3 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (d), (d), 82.2 (s), 28.4 (q), 25.9 (t) ppm. MS(EI) for C 23 H 33 N 3 4 m/z [M + ]. (14) [3,5-Bis-(piperidine-1-carbonyl)-phenyl]-carbamic acid tert-butyl ester A suspension of 13 (2.0 g, 4.2 mmol), piperidine (1.1 g, 12,6 mmol) and triethylamine (1.3 g, 21.6 mmol) in CH 2 Cl 2 (100ml) was stirred overnight under a dinitrogen atmosphere. The suspension was washed subsequently with a 1M HCl(aq) (2x), a saturated aqueous NaHC 3 solution (2x), water and finally with brine. The organic layer was dried with 70

72 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties Na 2 S 4 and the solvent evaporated. The solid residue was dissolved in MeH, after which the product was precipitated by dropwise addition to 1M HCl(aq). The white precipitate is filtered, washed with water and dried in a vacuum oven at 40 o C, yielding 14 as a white powder (1.18 g, mmol, 57 %). 1 H NMR (400 MHz, CDCl 3 ) δ = 9.66 (s, 1H), 7.51 (s, 2H), 6.87 (s, 1H), 3.56 (s, 4H), 3.25 (s, 4H), 1.48 (s, 9H), (m, 12H) ppm. 13 C NMR (50.3 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (d), (d), 80.9 (s), 49.0 (t), 43.4 (t), 28.5 (q), 26.7 (t), 25.8 (t), 24.7 (t) ppm. MS(EI) for C 23 H 33 N 3 4 m/z 415 [M + ], HRMS calcd for C 23 H 33 N 3 4 : , found: References 1 (a) X. Hu, A. Damjanović, T. Ritz and K. Schulten Proc. Natl. Acad. Sci. U. S. A. 1998, 95, (b) G. McDermott, S. M. Prince, A. A. Freer, A. M. Hawthornthwaite- Lawless, M. Z. Papiz, R. J. Cogdell and N. W. Isaacs Nature 1995, 374, (c) R. J. Cogdell, A. T. Gardiner, A. W. Roszak, C. J. Law, J. Southall and N. W. Isaacs Photosynth. Res. 2004, 81, (d) C. J. Law, A. W. Roszak, J. Southall, A.T. Gardiner, N. W. Isaacs and R. J. Cogdell Mol. Membr. Biol. 2004, 21, (a) N. Armaroli and V. Balzani Angew. Chem. Int. Ed. 2007, 46, (b) T. L. Benanti and D. Venkataraman Photosynth. Res. 2006, 87, (c) W. M. Campbell, A. K. Burrell, D. L. fficer and K. W. Jolley Coord. Chem. Rev. 2004, 248, (d) A. F. Nogueira, C. Longo and M.-A. De Paoli Coord. Chem. Rev. 2004, 248, (e) T. Markvart Prog. Quant. Electron. 2000, 24, F. Chaignon, M. Falkenström, S. Karlsson, E. Blart, F. dobel and L. Hammarström, Chem. Commun. 2007, (a) S. K. Grayson and J. M. J. Fréchet, Chem. Rev. 2001, 101, (b) A. Adronov and J.M.J. Fréchet, Chem. Commun. 2000, (c) F. Zeng and S.C. Zimmerman, Chem. Rev. 1997, 97, (a) V. Balzani, P. Ceroni, C. Giansante, V. Vicinelli, F.G. Klärner, C. Verhaelen, F. Vögtle and U. Hahn, Angew. Chem. Int. Ed. 2005, 44, (b) V. Balzani, P. Ceroni, M. Maestri and V. Vicinelli, Curr. pin. Chem. Biol. 2003, 7, (c) V. Balzani, P. Ceroni, A. Juris, M. Venturi, S. Campagna, F. Puntoriero and S. Serroni, Coord. Chem. Rev. 2001, 219, (a) W. R. Dichtel, S. Hecht and J. M. J. Fréchet, rg. Lett., 2005, 7, (b) P. Furuta, J. Brooks, M. E. Thompson and J. M. J. Fréchet, J. Am. Chem. Soc. 2003, 125, (c) J. M. Serin, D. W. Brousmiche and J. M. J. Fréchet, Chem. Commun. 2002, (a) M. R. Wasielewski, J. rg. Chem. 2006, 71, (b) L. E. Sinks, B. Rybtchinski, M. Iimura, B. A. Jones, A. J. Goshe, X. Zuo, D. M. Tiede, X. Li and M. R. Wasielewski, Chem. Mater. 2005, 17, (c) B. Rybtchinski, L. E. Sinks and M. R. Wasielewski, J. Am. Chem. Soc. 2004, 126, (d) M. J. Ahrens, 71

73 Chapter 2 L.E. Sinks, B. Rybtchinski, W. Liu, B. A. Jones, J. M. Giaimo, A. V. Gusev, A. J. Goshe, D. M. Tiede and M. R. Wasielewski, J. Am. Chem. Soc. 2004, 126, P. R. Hania, D. J. Heijs, T. Bowden, A. Pugžlys, J. van Esch, J. Knoester and K. Duppen, J. Phys. Chem. B 2004, 108, (a) A. Sautter, B. K. Kaletaş, D. G. Schmid, R. Dobrawa, M. Zimine, G. Jung, I. H. M. van Stokkum, L. De Cola, R. M. Williams and F. Würthner, J. Am. Chem. Soc. 2005, 127, (b) F. Würthner, Chem. Commun. 2004, (c) C. C. You and F. Würthner, J. Am. Chem. Soc. 2003, 125, (d) F. Würthner, A. Sautter, D. Schmid and P. J. A. Weber, Chem. Eur. J. 2001, 7, (a) V. Balzani, A. Juris, M. Venturi, S. Campagna and S. Serroni, Chem. Rev. 1996, 96, (b) A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser and A. Von Zelewsky, Coord. Chem. Rev. 1988, 84, (a) W. Wang, L. S. Li, G. Helms, H. H. Zhou and A. D. Q. Li, J. Am. Chem. Soc. 2003, 125, (b) A. D. Q. Li, W. Wang and L. Q. Wang, Chem. Eur. J. 2003, 9, A. P. H. J. Schenning, J. von Herrikhuyzen, P. Jonkheijm, Z. Chen, F. Würthner and E. W. Meijer, J. Am. Chem. Soc. 2002, 124, K. Sugiyasu, N. Fujita and S. Shinkai, Angew. Chem. Int. Ed. 2004, 43, C. Göltner, D. Pressner, K. Müllen and H.W. Spiess, Angew. Chem., Int. Ed. Engl. 1993, 32, S. L. Gilat, A. Adronov and J. M. J. Fréchet, J. rg. Chem. 1999, 64, (a) S. M. Melnikov, E. K. L. Yeow, H. Uji-i, M. Cotlet, K. Müllen, F. C. De Schryver, J. Enderlein and J. Hofkens, J. Phys. Chem. B 2007, 111, (b) C. Flors, I. esterling, T. Schnitzler, E. Fron, G. Schweitzer, M. Sliwa, A. Herrmann, M. van der Auweraer, F. C. De Schryver, K. Müllen and J. Hofkens, J. Phys. Chem. C 2007, 111, (c) M. Cotlet, T. Vosch, S. Habuchi, T. Weil, K. Müllen, J. Hofkens and F. De Schryver, J. Am. Chem. Soc. 2005, 127, (a) L. Flamigni, B. Ventura, C. C. You, C. Hippius and F. Würthner, J. Phys. Chem. C 2007, 111, (b) C. C. You, C. Hippius, M. Grüne and F. Würthner, Chem. Eur. J. 2006, 12, (a) D. Dotcheva, M. Klapper and K. Müllen, Macromol. Chem. Phys. 1994, 195, (b) F. Würthner, C. Thalacker, S. Diele and C. Tschierske, Chem. Eur. J. 2001, 7, (c) R. Iden and G. Seybold, (BASF AG), Ger. Pat. Appl., DE A1, 1985 (Chem. Abstr. 1985, 103, 38696q). (d) G. Seybold and G. Wagenblast, Dyes Pigm. 1989, 11, (e) G. Seybold and A. Stange, (BASF AG), Ger. Pat., DE , 1987 (Chem. Abstr. 1988, 108, 77134(c)). 19 (a) J. V. Caspar and T. J. Meyer, Inorg. Chem. 1983, 22, (b) J.V. Caspar and T.J. Meyer, J. Am. Chem. Soc. 1983, 105,

74 Intramolecular Energy Transfer in a Tetra-Coumarin Perylene System: Influence of Solvent and Bridging Unit on Electronic Properties 20. An alternative explanation for the decrease in emission lifetime is that it is due to a decrease in the radiative lifetime of the excited state. However, this would imply that the oscillator strength of the lowest energy absorption would decrease concomitantly also, c.f. the Strickler-Berg expression (S. J. Strickler and R. A. Berg, J. Chem. Phys. 1962, 37, ), which is not observed in the present case. 21 (a) M. J. Cook, A. P. Lewis, G. S. G. McAuliffe, V. Skarda, A. J. Thomson, J. L. Glasper and D. J. Robbins, J. Chem. Soc., Perkin Trans , (b) M. J. Cook, A. P. Lewis, G. S. G. McAuliffe, V. Skarda, A. J. Thomson, J. L. Glasper and D. J. Robbins, J. Chem. Soc., Perkin Trans , (c) G. Roelfes, V. Vrajmasu, K. Chen, R. Y. N. Ho, J. U. Rohde, C. Zondervan, R. M. la Crois, E. P. Schudde, M. Lutz, A. L. Spek, R. Hage, B. L. Feringa, E. Münck and L. Que, Inorg. Chem. 2003, 42, Gaussian 03, Revision B.04, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. chterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain,. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. rtiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al- Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez and J. A. Pople, Gaussian, Inc., Wallingford CT, R. Augulis, A. Pugžlys, J. H. Hurenkamp, B. L. Feringa, J. H. van Esch and P. H. M. van Loosdrecht, J. Phys. Chem. A, 2007, 111, (a) E. A. A. Wallén, J. A. M. Christiaans, E. M. Jarho, M. M. Forsberg, J. I. Venäläinen, P. T. Männisto and J. Gynther, J. Med. Chem. 2003, 46, (b) L. A. Carpino, E. M. E. Mansour, C. H. Cheng, J. R. Williams, R. Macdonald, J. Knapczyk, M. Carman and A. Łopusiński, J. rg. Chem. 1983, 48, J. Bitta and S. Kubik, rg. Lett. 2001, 3, M. Mardelli and J. lmsted III, J. Photochem. 1977, 7, T. Kartstens and K. Kobs, J. Phys. Chem. 1980, 84,

75 74

76 Chapter 3 Tuning Energy Transfer in Switchable Donor-Acceptor Systems The synthesis and characterisation of a coumarin - dithienylcyclopentene - coumarin symmetric triad (CSC) and a perylene bisimide - dithienylcyclopentene - coumarin asymmetric triad (PSC) is described. In both triads the switching function of the photochromic dithienylcyclopentene unit is retained. For CSC an overall 50% quenching of the coumarin fluorescence is observed upon ring-closure of the dithienylcyclopentene component, which, taken together with the low PSS (< 70%), indicates that energy transfer quenching of the coumarin component by the dithienylcyclopentene in the closed state is efficient. Upon ring opening of the dithienylcyclopentene unit the coumarin emission is restored fully. The PSC triad shows efficient energy transfer from the coumarin to the perylene bisimide unit when the dithienylcyclopentene unit is in the open state. When the dithienylcyclopentene is in the closed (PSS) state a 60% decrease in sensitized perylene bisimide emission intensity is observed due to competitive quenching of the coumarin excited state and partial quenching of the perylene excited state by the closed dithienylcyclopentene unit. This modulation of energy transfer is reversible over several cycles for both the symmetric and asymmetric tri-component systems. Part of this work was published in: J.H. Hurenkamp. J.J.D. de Jong, W.R. Browne, J.H. van Esch and B.L. Feringa rg. Biomol. Chem. 2008, 6,

77 Chapter Introduction Energy (ET) and electron (E n T) transfer between molecular entities is of continuing interest in the development of molecular based photonic systems, including photovoltaics, 1 molecular electronics 2 and sensor technologies. 3 The excellent efficiency in energy and electron transfer between donor and acceptor components in the photosynthetic apparatuses of plants and bacteria is achieved through the optimal supramolecular spatial arrangement and energetic matching of donor and acceptor units. 4 Achieving such control represents a considerable challenge in synthetic systems, where the tight organization exerted by nature though membrane and protein structures is not present a priori. 4,5 Efficient energy transfer in synthetic donor acceptor systems can be achieved either by through bond (superexchange) interactions or by through space energy transfer. Depending on the mechanism of through space energy transfer (e.g., Dexter 6 or Förster 7 energy transfer) between donor and acceptor chromophoric units, the efficiency is dependent on the absorption cross-section of the energy acceptor and its overlap with the fluorescence spectrum of the donor unit and also on their spatial arrangements. The electronic properties of the donor and acceptor units can be tuned synthetically to achieve an optimal energetic overlap. However, control over the spatial and orientational arrangement between components is often more difficult to achieve. verall the approaches taken to control this latter aspect can be divided into two groups i.e. the covalent and non-covalent (supramolecular) arrangement of donor-acceptor units. Both approaches have seen considerable success in achieving efficient energy transfer and in furthering our understanding of the physical basis of, e.g., Förster resonance energy transfer (FRET). 8 Previously, we have shown that efficient energy transfer can be achieved in a coumarin donor perylene bisimide acceptor based system. 9,10 In this tetra-coumarin-perylene bisimide system, we demonstrated that energy transfer with an efficiency of > 95% could be achieved with good stability even under conditions of high near-uv irradiation flux, without requiring through-bond interaction between the donor and acceptor components. Furthermore, careful matching of the energetics of the donor and acceptor units avoided potentially deleterious competing electron transfer processes. However, energy transfer in this molecule, although efficient, is not subject to post-synthetic control, i.e. the energy transfer efficiency from the donor coumarin units to the perylene bisimide acceptor cannot be altered or modulated reversibly after synthesis. 76

78 Tuning Energy Transfer in Switchable Donor-Acceptor Systems Controlling energy transfer post-synthetically, e.g., by changing the direction and efficiency of the process, represents an interesting, albeit considerable, challenge. The incorporation of an addressable component into supramolecular systems capable of attenuating energy transfer from an energy donor to an energy acceptor would allow for modulation of the emission output of the donor-acceptor system. Control over fluorescence intensity has been demonstrated by several groups, and, typically, this is achieved through quenching of the fluorescence of the chromophore by an additional unit, 11 whose electronic structure can be changed upon external perturbation, e.g., by photo- 12,13 or electrochemical 14 switching, ph change, 3,15 or by disconnection of a quenching unit. 16 Dithienylcyclopentene switches, which belong to a class of photochromic switches that show potential as photoswitchable fluorescence quenching units, are suitable candidates to impart functionality in these donor-acceptor systems. These photochromic switches undergo a photochemical cyclization reaction upon irradiation with UV light, which is reversible upon irradiation with visible light (Figure 3.1). Dithienylcyclopentenes show sufficient stability, good to very good photostationary states (PSS) and the two states have very different absorption spectra, which are both thermally stable. 17 The synthetic routes, which are available, 18 facilitate attachment of substituents (in this case chromophores). UV R 1 S R 2 S Vis R 1 S S pen Closed R 2 Figure 3.1 Schematic representation of the ring-open and ring-closed state of a dithienylcyclopentene switch. Indeed, dithienylcyclopentenes have been employed successfully already as fluorescence quenchers, examples of which have been reported by the groups of Lehn, 19 Irie, 20 Tian 21 and Branda. 22 However, control of energy transfer efficiency between an energy donor and acceptor pair by a third, addressable, component allows for more versatile control of excited state properties, such as the triad system reported by Walz et al. 23 This approach to switchable triad systems has been demonstrated in systems involving electron transfer also, 24 however the present work focuses on through-space energy transfer. In this contribution we report a covalently linked donor-switch-acceptor triad based on a coumarin (donor), a dithienylethene (switch) and a perylene bisimide (acceptor) unit. The energy transfer between the donor and acceptor units can be redirected by photochemical 77

79 Chapter 3 isomerization of the central switching unit. In the open form, the dithienylcyclopentene acts as a photophysically innocent bridging unit. In the closed form it acts to quench the emission of the coumarin donor and to a lesser extent of the perylene bisimide acceptor, thereby modulating the luminescence output of the perylene bisimide unit (Figure 3.2), by reducing the efficiency of energy transfer from the coumarin donor to the perylene acceptor. Figure 3.2 Schematic representation of a Donor Switch Acceptor triad in two different states: a) The switch is in the open State 1: excitation of the donor (D), energy transfer to the acceptor (A), followed by sensitized emission. b) The switch is in the closed State 2: when D is excited the energy is quenched by the switch and sensitized emission is prevented. 3.2 Synthesis The symmetric triad, consisting of two coumarin donors and a central dithienylcyclopentene unit, was prepared to demonstrate the quenching concept for the selected donor and switchable-acceptor chromophores (Scheme 3.1). It was decided to use amides in combination with piperazines as a molecular resistor (i.e. similar to the use of adamantanes 20a or ethers 25 in other multicomponent systems) to construct a symmetric triad, in which the coumarin chromophore is electronically decoupled from the dithienylcyclopentene. This approach was used in building a tetra coumarin-perylene bisimide system (Chapter 2). 10 The amide coupling method used here to construct the substituted dithienylcyclopentene photochromic switching units has been employed successfully in the synthesis of switchable gelator systems also Methoxycoumarin-3-acetic acid 1 was coupled to mono Boc-protected piperazine using the amide coupling reagent 1,1'-carbonyldiimidazole (CDI), 27 which allows for straightforward work-up and subsequent purification by column chromatography. The coumarin N-Boc-piperazine 2 was deprotected subsequently with trifluoroacetic acid (TFA). Two equivalents of the free amine coumarin piperazine 3 were coupled to the dicarboxylic acid switch 4 using 2-chloro-4,6-dimethoxy-1,3,5-triazine (CDMT), a reagent, which has been found to give good yields in combination with the dicarboxylic acid switch, 26 and N-methyl-morpholine (NMM) in CH 2 Cl 2 to yield the pure coumarin-switch- 78

80 Tuning Energy Transfer in Switchable Donor-Acceptor Systems coumarin triad (CSC) 5 in 16% yield (non-optimized) after purification by column chromatography (Scheme 3.1). + HN NBoc A H N 1 2, R = Boc N R B 3, R = H H S 4 S H C N N S 5 S N N Scheme 3.1 Synthesis of CSC 5: A) CDI, CH 2 Cl 2, RT (92%); B) CF 3 CH, CH 2 Cl 2 (quant.); C) 1. CDMT, NMM, CH 2 Cl 2, 0 o C 2. NMM, 3, (16%). A switch-acceptor unit was constructed, using an amine substituted perylene bisimide 9 (Scheme 3.2), which can be coupled to the dithienylcyclopentene diacid 4. Perylene bis-n-butylimide 6 28 was saponified partially to provide a mixture of the perylene-bisanhydride and the perylene-mono-imide-monoanhydride (~ 2 : 1, respectively, as determined by 1 H-NMR spectroscopy). 29 The mixture was condensed with 4-amino-1-Bocpiperidine, 30 to yield a mixture of substituted perylene bisimides with either two piperidines or one n-butyl and one piperidine group at the imide positions. The mixture was separated chromatographically providing the mono N-Boc-piperidine mono n-butyl perylene bisimide 8 in 8.6% yield from the perylene bis-n-butylimide 6 (Scheme 3.2). Ar = R N N N Ar Ar Ar Ar A Ar Ar Ar Ar B Ar Ar Ar Ar C 9, R = H N N N 6 7 8, R = Boc Scheme 3.2 Synthesis of mono piperidine perylene bisimide 9: A) KH, H 2, i-prh, 15 h (yield n.d.); B) 4-Amino-1-boc-piperidine, toluene, 120 o C, 24 h (8.6% from 6); C) CF 3 CH, CH 2 Cl 2 (quant.). 79

81 Chapter 3 Attempts to monosubstitute the diacid dithienylcyclopentene 4, with the coumarin piperazine 3 followed by coupling to the perylene bisimide 9 were unsuccessful due to difficulties in purification of the mono acid after the first coupling step. Therefore, it was decided to use a one step procedure with equimolar amounts of the chromophores followed by isolation of the desired compound by column chromatography. First one equivalent of the perylene-mono-piperidine-mono-n-butyl 8 was deprotected using TFA and the product 9 (1 equivalent) was coupled, together with the coumarin piperazine 3 (1 equivalent) to the diacid dithienylcyclopentene photochromic switch 4 (1 equivalent) to give, in addition to the homo-coupling products, the target Perylene-Switch-Coumarin triad (PSC) 10 in 4.4% yield (non-optimized, Scheme 3.3). H S 4 S H A,B N N S 10 S N N N Scheme 3.3 Synthesis of PSC 10: A) CDMT, NMM, CH 2 Cl 2, 0 o C; B) NMM, 1 eq 3, 1 eq 9, 24 h. The model compound piperidine dithienylcyclopentene piperidine (pipspip 11, Figure 3.4, right) was prepared by similar methods as for CSC 5. All compounds were purified by column chromatography and characterized with 1 H and 13 C NMR spectroscopy and (MALDI-TF) mass spectrometry (see experimental section for details). 3.3 Electronic and Photochemical Properties Coumarin-Switch-Coumarin triad (CSC) 5 The absorption spectrum of CSC 5 in the open and closed (i.e. at the photostationary state (PSS) obtained by irradiation at λ exc = 312 nm, PSS 312 nm ) form are shown in Figure 3.3. The spectra show features of both coumarin and open or closed switch and the maxima correspond closely to those of the model compounds indicating that no direct, or through- 80

82 Tuning Energy Transfer in Switchable Donor-Acceptor Systems bond, electronic communication between the switch and the coumarin is present (Figure 3.3 and Table 1). Although at 298 K irradiation at λ exc = 312 nm resulted in 10-20% decomposition per cycle, at 220 K photochromic switching was fully reversible ε (cm -1 M -1 ) Wavelength (nm) Figure 3.3 Absorption spectra of CSC 5 open ( ), CSC PSS 312 nm ( ) irradiated at 220 K with λ = 312 nm, and coumarin model 2 are shown ( ). The spectra were recorded at RT in CH 2 Cl 2. Table 3.1 Absorption and emission spectra of CSC 5 and PSC 10 open and PSS 312 nm Compound Absorption a λ max / nm (10 3 ε / cm -1 M -1 ) Emission a λ max /nm 2 Coumarin pip Boc 322(18.3) Perylene Bisimide butyl 266(40.6), 286(49.6), 451(16.7), 539(26.7), 577(43.1) pipspip open 239(18.0) 11 b pipspip PSS 312 nm 236(13.6), 487(29.2) 5 CSC open 321 (31.7) CSC PSS b 312 nm 323 (33.3), 493 (4.0) PSC open 267 (50.2), 286 (54.2), 452 (13.4), 540 (22.4), 579 (36.2) 391, PSC PSS b 312 nm 267 (47.5), 286 (52.7), 453 (14.4), 540 (23.2), 579 (36.7) 391, 613 a measurements taken in CH 2 Cl 2 at RT. b at RT after irradiation with λ = 312 nm light at 220 K to PSS. 81

83 Chapter 3 Comparison of the difference spectrum obtained by subtraction of the spectra of 5 in the open form and in the PSS 312 nm with the difference spectrum of the closed form of the model compound 11 shows the similarity between the two systems, confirming that the change in absorption observed upon irradiation at 220 K 31 is due to photochemical ring closure and that the photochromism of the dithienylcyclopentene is retained in the triad (Figure 3.4). CSC 5 can undergo several ring opening/closing cycles with a near complete recovery in the absorption spectrum of the open form after each cycle (Figure 3.5) Abs Wavelength (nm) Abs Wavelength (nm) N S S 11 pipspip N Figure 3.4 Left: The UV/Vis difference spectrum obtained by subtraction of the spectrum of the CSC 5 PSS 312 nm state from the spectrum of the CSC 5 open form recorded at 298 K; the PSS was obtained by irradiation with λ = 312 nm at 220 K. Right: The difference spectrum obtained by subtraction of the spectrum of the pipspip 11 PSS 312 nm state from the spectrum of the pipspip 11 open form recorded at 298 K; the PSS was obtained by irradiation with λ = 312 nm at 220 K Abs pen 0 Closed 1 pen 1 Closed 2 pen 2 Closed 3 pen 3 Closed 4 pen 4 Abs pen 0 Closed 1 pen 1 Closed 2 pen 2 Closed 3 pen 3 Closed 4 pen 4 Wavelength (nm) Switching # Figure 3.5 Left: Absorption spectrum of CSC 5 before and after photochemical ring closure (λ exc > 312 nm) and opening (λ exc > 400 nm). Irradiation was carried out at 220 K in CH 2 Cl 2. Right: Absorption at λ = 493 nm plotted against number of times switched. 82

84 Tuning Energy Transfer in Switchable Donor-Acceptor Systems The fluorescence spectrum of the open form of 5 (λ ~ 390 nm), is identical to that of the free coumarin (Figure 3.6). As for absorption spectroscopy, photochemical ring closure results in changes to the luminescence properties of 5. Irradiation at λ 312 nm results in a decrease in the intensity of the characteristic coumarin emission by 50%, which is reversed by irradiation at λ > 400 nm. This change is not observed when opening and closing a 2 : 1 mixture of the coumarin model 2 and 11, thus excluding trivial or radiative energy transfer being responsible for the changes seen in the symmetric triad 5. The absence of a change in the absorption of the coumarin band in the spectrum of 5 upon ring closure of the dithienylcyclopentene switching unit indicates that the ring closing does not perturb the electronic structure of the coumarin moieties. Hence the decrease in emission intensity can be assigned to energy transfer quenching by the dithienylcyclopentene unit in the closed state. Indeed the absorption spectrum of the closed dithienylethene unit shows good spectral overlap with the emission spectrum of the coumarin; a prerequisite for energy transfer (Figure 3.6) M M 1.2M M A.U. 0.4 CPS 1.0M 900.0k k 700.0k k pen 0 Closed 1 pen 1 Closed 2 pen 2 Closed 3 pen 3 Closed 4 pen 4 Wavelength (nm) Switching # Figure 3.6 Left: Absorption spectra of 5 open ( ) and CSC 5 PSS 312 nm ( ), and fluorescence spectra of 5 open ( ) and 5 PSS 312 nm ( ). Right: Effect of ring closing and subsequent opening on the fluorescence at λ = 394 nm (ring closing with λ exc = 312 nm at 220 K and opening with λ > 400 nm at 298 K). Spectra recorded in CH 2 Cl 2 at 298 K. In the open state, the emission of 5 (λ exc = 337 nm, λ em = 420 nm) decays monoexponentially with a fluorescence lifetime of 1.1 ns. At the PSS 312nm in which a mixture of the open and closed form of the switch are present, it is no longer possible to fit the emission decay with a monoexponential function. A biexponential fit provided the expected decay of the coumarin emission 1.1 ns and a second crosscorrelated component, i.e. a component with a decay lifetime less than the resolution of the instrument (500 ps). 33 This latter component can be attributed to emission from the coumarin in the closed form of 83

85 Chapter 3 5 in which efficient energy transfer results in the coumarin emission lifetime being limited by the rate of energy transfer from the excited coumarin to the dithienylcyclopentene unit in the closed state. 10 The degree of quenching of coumarin fluorescence observed (in this case 50% for 5 at the PSS) is dependent on the open closed ratio at the PSS state. Separation of a mixture of the open and closed form of 5 by HPLC was unsuccessful. However, for the model compound 11 separation was achieved and a ratio of 30% open 70% closed was determined for PSS 312nm. 34 This value can be taken as the upper limit for CSC 5 (and for PSC 10, vide infra) since the PSS is dependent on both the quantum yield for ring opening and closing, and the relative absorption cross-section of the open and closed forms at λ = 312 nm. Since the dithienylcyclopentene switch is electronically decoupled from the coumarin in 5 the quantum yield of ring opening and of ring closing are not likely to be significantly different between 5 and 11. However, the coumarin components of 5 absorb at λ = 312 nm and energy transfer from the coumarin to the closed dithienylcyclopentene component will increase the effective absorptivity of the closed dithienylcyclopentene at this wavelength in comparison to 11. This will serve to change the photostationary state of 5 at λ = 312 nm in favour of the open state in comparison to Abs Wavelength (nm) Figure 3.7 Absorption spectra of PSC 10 ( ) and a 1:1:1 mixture of the individual components 2, 6 and 11 ( ) recorded in CH 2 Cl 2 at 298 K. 84

86 Tuning Energy Transfer in Switchable Donor-Acceptor Systems Perylene-switch-coumarin triad (PSC) 10 The absorption and emission spectra of the PSC triad are shown in Figure 3.7 and Figure 3.8, respectively. The absorption spectrum of PSC 10 correlates closely with the spectrum of a 1:1:1 mixture of the perylene bisimide butyl 6, pipspip 11 and coumarin-pip-boc 2. The near perfect overlap of the λ max of the perylene bisimide and coumarin components indicates that the amide-based bridging units do not allow for significant through-bond electronic communication between the three units or perturbation of the electronic structure of the individual components. In contrast to the absorption spectra, the emission spectra of the PSC triad 10 and of a 1:1:1 mixture of the separate components show considerable differences. When the 1:1:1 mixture is excited at λ = 322 nm, the emission spectrum shows the characteristic emissions of the coumarin (λ em = 393 nm) and perylene bisimide (λ em = 609 nm) components. The excitation spectra recorded at the maxima of the coumarin and perylene bisimide emissions show the characteristic shapes of the coumarin and perylene bisimide absorption spectra, respectively. CPS 1.2M 1.1M 1.0M 900.0k 800.0k 700.0k 600.0k 500.0k 400.0k 300.0k 200.0k 100.0k 0.0 CPS 800k 700k 600k 500k 400k 300k 200k 100k Wavelength (nm) Wavelength (nm) Figure 3.8 Left: Emission spectra of 10 ( ) and of the model mixture 1:1:1 of 2, 6 and 11 ( ) irradiated at λ = 322 nm at RT in CH 2 Cl 2, spectra were corrected for absorption at the excitation wavelength. Right: Excitation spectra of PSC 10 ( ) and of the model mixture (a 1:1:1 ratio of 2, 6 and 11) ( ) monitored at λ = 620 nm at RT in CH 2 Cl For the PSC triad 10, excitation at λ = 322 nm shows emission characteristic of the coumarin and perylene bisimide components also, however, the coumarin emission is very weak when compared with the 1:1:1 mixture and the perylene bisimide emission is more 85

87 Chapter 3 intense (Figure 3.8, left).that this is due to energy transfer from the coumarin to the perylene bisimide components is apparent from the excitation spectra recorded at the emission maximum of the perylene bisimide component. The excitation spectrum of PSC 10 shows that a large contribution to the emission at λ = 620 nm originates from an absorption with a maximum at λ ~ 325 nm, the λ max of the coumarin absorption. This contribution is not seen in the excitation spectrum of the 1:1:1 mixture. This confirms that in 10 in the open state, efficient intramolecular energy transfer from the coumarin to the perylene bisimide takes place (Figure 3.8, right) Absorption and emission spectroscopy of 10 at the PSS 312nm The changes in the absorption spectrum on irradiation to the PSS at λ = 312 nm are minor, due to the relatively low molar absorptivity of the switching unit compared with those of the coumarin and perylene bisimide components (Figure 3.9). The changes are more apparent in the difference spectra and are similar to those observed for the model switching unit 11 (Figure 3.4 right), 36 and confirms that the dithienylethene unit retains its photochromic behaviour in the PSC triad 10. As for 5, ring opening and closing of the switching unit of 10 can be performed over several cycles (Figure 3.10) Abs 0.2 Abs Wavelength (nm) Wavelength (nm) Figure 3.9 Left: PSC 10 open ( ) and PSS 312 nm ( ) irradiated at 220 K 37 with λ = 312 nm in CH 2 Cl 2, spectra recorded at RT. Right: Difference spectrum for PSC 10 at t = 0 min and PSS after t = 10 min irradiation at λ = 312 nm at 220K, (see Figure 3.4 for comparison with 5 and 11). 86

88 Tuning Energy Transfer in Switchable Donor-Acceptor Systems 0,6 0,4 pen 0 Closed 1 pen 1 Closed 2 pen 2 0,232 0,230 0,228 0,226 Abs 0,2 Abs 0,224 0,222 0,220 0,218 0, Wavelength (nm) 0,216 pen 0 Closed 1 pen 1 Closed 2 pen2 Wavelength (nm) Figure 3.10 Left: Absorption spectra of PSC 10 before and after switching from PSC open to PSC PSS, irradiated with λ = 312 nm at 220K and λ > 400 nm light at RT in CH 2 Cl 2 respectively. Right: Absorption at 493 nm plotted over three switching cycles of PSC using λ = 312 nm light at 220K to close and λ > 400 nm at RT to ring-open the dithienylcyclopentene unit. Measurements were performed in CH 2 Cl 2 and spectra were recorded at RT M irradiation at 312 nm irradiation at > 450 nm 2.0M M CPS 1500 CPS 1.0M k Wavelength (nm) Wavelength (nm) Figure 3.11 Left: Emission spectra of PSC 10 in the open state (solid line), after irradiation at λ = 312 nm to form the PSS 312 nm state (dashed line) and after visible (> 450 nm) irradiation to reform the open state (dotted line). All traces were recorded by excitation at λ = 450 nm in CH 2 Cl 2 at RT. Right: Emission spectra of PSC 10 open ( ) and PSS 312 nm ( ) irradiated at 220K with λ = 312 nm in CH 2 Cl 2. All traces were recorded by excitation at λ = 322 nm in CH 2 Cl 2 at RT. The absence of a large overlap of the absorption of the dithienylethene component in the closed state and the emission spectrum of the perylene unit would suggest that energy transfer between the two units would be inefficient. However, comparison of the emission spectra of the perylene unit in the open and PSS 312nm states (under direct excitation i.e. 87

89 Chapter 3 λ exc = 450 nm, Figure 3.11, Left) shows that the perylene bisimide emission is quenched significantly (21%) upon ring closing. Taking the low photostationary state (i.e. < 70% closed form) into account, in the closed state the efficiency of quenching of the perylene excited state by the dithienylethene unit is ca %. A decrease of 30% in the intensity of the perylene bisimide fluorescence of 10 in its PSS 312 nm is observed (Figure 3.11, Right) when excited at the λ max of the absorption of the coumarin (λ = 322 nm). This could indicate that energy transfer from the coumarin unit is less efficient in the closed state. This decrease, however, is not accompanied by a concomitant increase in the fluorescence of the coumarin component. Hence, the decrease in perylene bisimide emission intensity is due to the introduction of an alternative quenching pathway for the coumarin component, and not a decrease in the overall efficiency of quenching of the coumarin excited state in the triad. CPS 2.5M pen 0 Closed 1 pen 1 2.0M Closed 2 pen 2 1.5M Closed 3 pen 3 1.0M 500.0k Wavelength (nm) CPS 2.4M 2.3M 2.2M 2.1M 2.0M 1.9M 1.8M 1.7M 1.6M 1.5M pen 0 Closed 1 pen 1 Closed 2 pen2 Closed 3 pen 3 Switching # Figure 3.12 Left: Emission spectra of switching cycles for PSC 10 from PSC open to PSC 10 closed PSS, irradiated with, respectively, λ = 312 nm at 220 K and λ > 400 nm light at RT in CH 2 Cl 2, compensated for absorption. Right: Fluorescence intensity at λ = 614 nm plotted for three switching cycles of PSC 10 using λ = 312 nm light at 220 K to close and λ > 400 nm at RT to open, measurements were performed in CH 2 Cl 2 and spectra were recorded at RT. The reduction in the intensity of the perylene bisimide component is reversed upon irradiation at λ > 400 nm and the closing/opening cycle can be repeated several times with limited degradation (< 8% per cycle, Figure 3.12). The reversibility of the changes in emission spectrum of 10 confirms that the effect is due to the opening and closing of the dithienylethene switch component of the triad. A steady increase of the coumarin fluorescence (λ max = 391 nm) is observed with each cycle, which is assigned to photodegradation of the perylene bisimide unit. Indeed, when irradiating at λ = 254 nm 88

90 Tuning Energy Transfer in Switchable Donor-Acceptor Systems rapid decomposition of the perylene bisimide with a near complete recovery of the expected emission intensity of the coumarin component is observed (Figure 3.13) M M 1.2M M Abs 0.2 CPS 800.0k 600.0k k 200.0k Wavelength (nm) Wavelength (nm) Figure 3.13 Left: The change in the absorption spectrum of PSC 10 upon irradiation over 20 min with λ = 254 nm light in CH 2 Cl 2 at RT. Right: The change in emission spectra (λ ex = 322 nm) of PSC 10 by irradiation over 18 min with λ = 254 nm light in CH 2 Cl 2 at RT. The fluorescence decay kinetics for the triad show considerable differences, when comparing the open state and PSS 312 nm. In the open form, the emission decay of 10 (recorded at λ em = 420 nm) is biexponential, with a slow component of ~ 2.0 ns, assigned to the fluorescence decay of free coumarin, and a short (cross-correlated) component, which is attributed to energy transfer (between the coumarin and perylene bisimide unit) rate limited fluorescence decay of the coumarin component of the triad. This decay time is comparable with that observed in the tetra coumarin-perylene bisimide system reported earlier. 9,10 Compound 10, at the PSS 312 nm, shows biexponential decay kinetics with similar lifetimes and contributions as in the open state. The emission decay kinetics of 10 in the open state, recorded at λ em = 615 nm (i.e. the λ max of the emission of the perylene bisimide component), show monoexponential decay kinetics with a decay lifetime of ~7 ns, which corresponds closely to the lifetime of the perylene bisimide model compound 6. The same value was observed for the PSC 10 PSS 312 nm, indicating that the closed form of the switch does not perturb the energy of the emissive excited state of the perylene, i.e. the radiative and non-radiative decay rates are unaffected. 89

91 Chapter PSC pen PSC pen >400nm 1500 PSC Closed PSS 312nm Counts Time (ns) Figure 3.14 TCSPC spectra of PSC 10 PSS 312 nm irradiated with λ = 322 nm light and fluorescence decay counts measured at λ = 615 nm. All traces were recorded over the same acquisition time to enable comparison of the signal intensity, top (black): PSC 10 open, bottom (dark grey): PSC PSS 312 nm by irradiation with λ = 312 nm light for 4 min, and middle (light grey): PSC 10 open after irradiating the PSS 312 nm form with λ > 400 nm light for 20 min, all traces recorded in CH 2 Cl 2 at RT. Fluorescence decay traces were recorded for 10 in the open, PSS 312 nm and reopened state. (Figure 3.14). Irradiation of the open state of 10 to the PSS 312 nm (closed) state results in a decrease in emission intensity, however the emission decay lifetimes measured in either state are unaffected. This indicates that even though there is less energy being transferred to the perylene bisimide acceptor from the coumarin donor, this is not caused by a change in the electronic structure of the perylene bisimide. Upon irradiation with λ > 400 nm to reopen the dithienylcyclopentene switch component, the intensity of the fluorescence increases again, recovering to nearly its original intensity, as observed by emission spectroscopy (Figure 3.14). 90

92 Tuning Energy Transfer in Switchable Donor-Acceptor Systems 3.4 Discussion The symmetric CSC triad 5 In the symmetric triad CSC 5, the ability of the dithienylcyclopentene unit to switch between an open and closed state is apparent from the appearance of the characteristic absorption band of the closed state in the visible region (λ max ~ 500 nm) upon UV irradiation and its subsequent disappearance upon irradiation with visible light. The decrease in the fluorescence intensity of the coumarin components observed upon ring closing of the dithienylcyclopentene unit can be assigned to energy transfer quenching of the excited coumarin unit by the closed dithienylcyclopentene on the basis of an absence of such an effect in the 2:1 mixture of 2 and 11 and the biexponential nature of the emission decay upon ring closure. The incomplete quenching (~50%) of the fluorescence of the coumarin components at the PSS 312 nm is not indicative of inefficient quenching, but reflects the low photostationary state achievable for the dithienylcyclopentene unit, which is less than 70% in favour of the closed form. This is supported by the fluorescence lifetime decay traces where, for the open form, strictly monoexponential behaviour is observed with a lifetime similar to the free coumarin component, whereas at the PSS 312 nm the fluorescence decay is no longer monoexponential but shows two distinct contributions a component identical to that observed in the open state and a cross-correlated component with a lifetime considerably less than the instrument resolution (i.e. < 500 ps) (Figure 3.15). Figure 3.15 Energy level diagram of the spectroscopic processes observed in the open and closed state of 5 (C = coumarin, S = dithienylcyclopentene). The biexponential decay at the PSS 312 nm reflects the presence of both the open form and closed form of the symmetric triad in solution. Nevertheless, it is clear that the dithienylcyclopentene component can act as a switchable efficient energy sink for the coumarin components. 91

93 Chapter The asymmetric PSC triad 10 The ability to quench the fluorescence of the coumarin by the dithienylcyclopentene component in the closed state but not the open state can be used to modulate the emission output of coumarin/perylene bisimide based donor-acceptor systems. 9,10 In the present study the dithienylcyclopentene unit was incorporated between the energy donating coumarin unit and the energy accepting perylene bisimide unit, i.e. in the triad 10. The absorption spectrum of PSC 10 is almost identical to that of the 1 : 1 : 1 mixture of 2, 6 and 11 indicating that the amide units employed to link the three components together do not facilitate ground state electronic communication and that the covalent tethering of the individual components does not result in a perturbation of their electronic properties. Indeed the photochemistry of 11, with respect to ring opening and closing of the dithienylcyclopentene unit, is retained in 10. With respect to luminescence, for 10 the covalent attachment of the chromophores does not affect the spectral shape compared with the emission spectra of the separate components. However, the relative intensities of each emission component show substantial changes. When the emission spectrum of the open form of PSC 10 is compared to that of the model mixture, two aspects are notable. First it is evident that the intensity of the coumarin fluorescence in the model mixture is much higher then that in the emission spectrum of 10. Secondly the intensity of the perylene bisimide acceptor fluorescence is increased significantly in the emission spectrum of 10 compared with 6. This shows that there is intramolecular energy transfer from the coumarin-donor to the perylene bisimide-acceptor only in the triad system and not in the solution of the mixture of the separate component units. Comparison of the emission spectra of the open and PSS 312 nm state of PSC 10 shows an intensity decrease of the perylene bisimide emission, which is not accompanied by a proportional increase in emission from the coumarin component. Hence, energy transfer from the coumarin donor is still taking place, but is directed elsewhere. Considering 5, it is most likely that energy transfer from the coumarin to the closed form of the dithienylcyclopentene unit is taking place. The decrease in emission intensity of 10 upon irradiation to the PSS 312 nm state is ~ 30% (Figure 3.11), however this does not take into account the direct excitation of the perylene bisimide. Surprisingly, in the closed state, not only the coumarin emission is quenched but the perylene bisimide excited state is also partly quenched by the dithienylethene. This is 92

94 Tuning Energy Transfer in Switchable Donor-Acceptor Systems unexpected since the overlap of the dithienylethene absorption spectrum and the emission spectrum of the perylene bisimide is negligible. verall the modulation of the fluorescence by photochromic switching can be attributed to several possible effects. The ring-closing of the switching unit has two effects, one is an increase in the rigidity of the triad, thereby increasing the average distance between coumarin-donor and perylene bisimide-acceptor. Secondly, by ring closing, a new, low energy chromophore is formed, which allows for competitive (with the perylene bisimide) energy transfer quenching of the coumarin emission, as observed for 5. That the latter mechanism is most likely to be the major effect is confirmed by the absence of a proportional increase in emission intensity of the coumarin component in the PSS 312 nm state. In the open state of 10 it is not possible to quench the coumarin excited state via the dithienylcyclopentene (open switch) since its lowest excited state lies higher in energy than that of the coumarin (vide supra), and the rate of fluorescence decay of the coumarin is limited by the rate of energy transfer to the perylene bisimide unit (Figure 3.16). Figure 3.16 Energy level diagram of the spectroscopic processes observed in the open and closed state of PSC 10 (P = perylene bisimide, C = coumarin, S = dithienylcyclopentene switch). In the closed state of 10 there are additional competing energy transfer and decay processes (Figure 3.16). From previous results it is known that the energy transfer from the coumarin donor to the perylene bisimide acceptor (k 1 ) is a fast process (i.e. ~ 11 ps, see ref. 10). For k 2, energy transfer from the coumarin to the closed dithienylcyclopentene, to be competitive it must be of the same order of magnitude or faster, which in this case is probable since a ~ 60% decrease in fluorescence intensity is observed upon irradiation to the PSS 312nm of 10. nce energy has been transferred to the closed dithienylcyclopentene it can dissipate through a fast non-radiative decay path (k 3 ) or be transferred to the perylene 93

95 Chapter 3 bisimide (k 4 ). Energy transfer to the perylene bisimide through the Förster mechanism is unlikely since the Φ fl for the closed dithienylcyclopentene is low 26a and a high quantum yield is a prerequisite for efficient FRET. Nevertheless it is clear from Figure 3.11 (left) that the dithienylethene unit in the closed state can itself quench the emission of the perylene bisimide component (k -4 ), albeit with low efficiency (55-65%), implying that the proximity of the perylene and closed dithienylethene components is sufficient to allow for energy transfer to take place. This means that energy transferred to the closed dithienylcyclopentene dissipates through a fast non-radiative decay pathway. Thus in the closed state the dithienylcyclopentene component provides a fast and efficient route to quench the coumarin emission as was observed for 5 and as an inefficient route able to quench the perylene emission only partially. The present system is comparable to the system of Walz et al., who have used energy transfer quenching to influence intramolecular energy transfer in a triad molecule. 23 In that system and in contrast to 10, the photochromic switch (a fulgimide) is closed and provides the lowest energy state of the system. It is, therefore, able to quench both of the chromophores excited states (i.e. anthracene and coumarin) and acting as an energy sink for all intramolecular processes. For the present system (i.e. 10) the lowest excited state of the photochromic switch in the closed state lies between the donor coumarin and the perylene bisimide acceptor excited states (Figure 3.16). Hence it is able to quench the excited state of the coumarin efficiently, however, the perylene bisimide excited state is quenched only partially. 3.5 Conclusions In this chapter two energy transfer donor-acceptor systems are reported. In the first system the function of the dithienylcyclopentene based photochromic switching unit to act as a molecular switch, i.e. turn coumarin emission on and off, is demonstrated. In the asymmetric PSC triad system 10 we have demonstrated that energy transfer efficiency in a donor-acceptor system can be addressed through modulation of the energy transfer quenching abilities of a photoactive unit. The low PSS achievable (< 70%) and the poor photostability at room temperature in the present system, both related to intrinsic properties of the dithienylcyclopentene unit will be addressed in further studies. Nevertheless, the present synthetic approach enables connection of the photoactive units covalently without loss of molecular function. These combined observations show that it is possible to build a molecular triad which allows for modulation of the energy transfer in a donor-acceptor system by introducing a switchable selective quencher of the donor unit and thereby control the emission output. 94

96 Tuning Energy Transfer in Switchable Donor-Acceptor Systems 3.6 Experimental section Uvasol-grade solvents (Merck) were employed for all spectroscopic measurements. All reagents employed in synthetic procedures were of reagent grade or better, and used as received unless stated otherwise. N-Boc-piperazine, 38 2, 9 3, and 6 39 were prepared according to literature. 1 H NMR spectra were recorded at 400 MHz; 13 C NMR spectra at 101 MHz. All spectra were recorded at ambient temperature, with the residual proton signals of the solvent as an internal reference. Chemical shifts are reported relative to TMS. CI and EI mass spectra were recorded on a JEL JMS-600 mass spectrometer in the scan range of m/z with an acquisition time between 300 and 900 ms and a potential between 30 and 70 V. MALDI-TF spectra were recorded on an Applied Biosystems Voyager-DE Pro. UV/Vis absorption spectra (accuracy ± 2 nm) were recorded on a Hewlett-Packard UV/Vis 8453 spectrometer. Fluorescence measurements were performed on a SPF-500C (SLM Aminco) or a Jobin-Yvon Fluorolog 3-22 spectrofluorimeter, the sharp features between λ = 450 and 500 nm in the excitation spectra are instrumental artefacts, the excitation and emission spectra are uncorrected for variations in lamp intensity and detector response. Sample concentration typically 10-5 M, spectra were recorded in 10 mm pathlength quartz fluorescence cuvettes. Luminescence lifetime measurements were obtained using an Edinburgh Analytical Instruments (EAI) timecorrelated single-photon counting apparatus (TCSPC) comprised of two model J-yA monochromators (emission and excitation), a single photon photomultiplier detection system model 5300, and a F900 nanosecond flashlamp (N 2 filled at 1.1 atm pressure, 40 khz) interfaced with a personal computer via a Norland MCA card. A 400 nm cut off filter was used in emission to attenuate scatter of the excitation light (337 nm). Data correlation and manipulation was carried out using EAI F900 software version Emission lifetimes were calculated using a single-exponential fitting function, Levenberg-Marquardt algorithm with iterative deconvolution (Edinburgh instruments F900 software). The reduced χ 2 and residual plots were used to judge the quality of the fits. Lifetimes are ± 5%. (5) Coumarin-Switch-Coumarin (CSC) triad Diacid 4 (200 mg, 0.58 mmol) was suspended in CH 2 Cl 2 (20 ml) and placed in an ice bath. Subsequently N-methylmorpholine (0.13 ml, 1.21 mmol) was added whereupon the solid dissolved. 2-Chloro-4,6-dimethoxytriazine (192 mg, 1.16 mmol) was added and the reaction mixture was stirred for 4 h at 0 C, after which another two equivalents of N- methylmorpholine (0.13 ml, 1.21 mmol) were added followed by 3 ( 350 mg, 1.16 mmol). Stirring was continued for 1 h at 0 C, and overnight at room temperature. CH 2 Cl 2 (50 ml) was added and the solution was washed with, respectively, 1M aq. HCl (2 x 20 ml), brine 95

97 Chapter 3 (1x 20 ml), saturated aqueous bicarbonate solution (1 x 20 ml) and H 2 (1 x 20 ml). The organic phase was dried over Na 2 S 4 and the solvent was removed in vacuo. The resulting solid crude product was purified using column chromatography (2% MeH in CH 2 Cl 2, Si 2 ), yielding the light yellow solid (85 mg, mmol, 16 %) 1 H NMR (400 MHz, CDCl 3 ) δ = 7.63 (s, 2H), 7.32 (d, J = 8.6 Hz, 2H), (m, 6H), 3.82 (s, 6H), 3.62 (s, 16H), 3.58 (s, 4H), 2.75 (t, J = 7.4 Hz, 4H), 2.10 (s, 6H), (m, 2H) ppm. 13 C NMR (101 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (d), (s), (s), (s), (s), (d), (d), (s), (s), (d), (d), 55.6 (q), 45.8 (t), 41.7 (t), 37.6 (t), 34.1 (t), 23.0 (t), 14.3 (q) ppm. MALDI-TF MS (MW =916.28) m/z = [M + ]. (8) N-(n-butyl)-N -(1 -Boc-piperid-4 -yl)-1,6,7,12-tetra(4-tert-butylphenoxy)perylene- 3,4,9,10-tetracarboxylic acid bisimide Partial saponification of N,N -di-n-butyl-1,6,7,12-tetra(4-tert-butylphenoxy)perylene- 3,4,9,10-tetracarboxylic acid bisimide 6 (3.3 g, 3.0 mmol) was carried out with KH (75 g, 134 mmol) in a mixture of isopropyl alcohol (500 ml) and H 2 (50 ml) under a dinitrogen atmosphere by stirring at reflux for 13 h, followed by separation of the basic aqueous layer. The organic layer was poured onto an aqueous 10% HCL (1 l) solution and left overnight, during which the color changed from green to orange to dark red. Filtration and thorough washing and drying, yielded a mixture of perylene bisanhydride and perylene mono butylimide in a ratio of ~ 2:1 (3.3 g) as determined by 1 H NMR spectroscopy. This mixture was heated at reflux under a dinitrogen atmosphere in dry toluene (330 ml) with 4-amino-1- Boc-piperidine (2 g, 10 mmol) for 3 d. The solvent was evaporated and the remaining crude product was purified by column chromatography (0.5 % MeH in CH 2 Cl 2, Si 2 ) providing the mono butyl mono N-Boc-piperidine perylene bisimide as a red solid (315 mg, 0.26 mmol, 8.6 %). 1 H NMR (400 MHz, CDCl 3 ) δ = 8.22 (s, 2H), 8.20 (s, 2H), (m, 8H), (m, 8H), (m, 1H), (m, 3H), (m, 1H), 2.66 (dq, J = 11.8, 3.5 Hz, 2H), (m, 4H), 1.45 (s, 9H), 1.39 (dd, J = 15.1, 7.5 Hz, 2H), 1.29 (s, 36H), 0.94 (t, J = 7.3, 7.3 Hz, 3H) ppm. 13 C NMR (101 MHz, CDCl 3 ) δ =163.7 (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (d), (s), (s), (s), (s), (d), (d), (s), (s), (d), (d), 79.4 (s), 52.0 (d), 44.3 (t), 43.5 (t), 40.3 (t), 34.3 (t), 31.4 (q), 30.1 (t), 28.4 (q), 20.3 (t), 13.7 (q) ppm. MALDI-TF MS (MW = ) m/z = [M + ]. General deprotection method for BC protected amines 2 and 8: The Boc protected amine was stirred in a mixture of 1:1 CH 2 Cl 2 : CF 3 CH for 4h. An equal volume of water was added and the mixture was neutralized by addition of solid NaHC 3, after which the aqueous layer was separated and the organic layer washed with a 96

98 Tuning Energy Transfer in Switchable Donor-Acceptor Systems saturated NaHC 3 solution (aq). The organic layer was dried over Na 2 S 4 and solvent removed in vacuo. The product was used in subsequent steps without further purification. (10) Perylene-Switch-Coumarin (PSC) triad Diacid 4 (93 mg, 0.27 mmol) was suspended in CH 2 Cl 2 (20 ml) and placed in an ice bath. Subsequently N-methylmorpholine (0.06 ml, 0.56 mmol) was added whereupon the solid dissolved. 2-Chloro-4,6-dimethoxytriazine (98 mg, 0.56 mmol) was added and the reaction mixture stirred for 4h at 0 C, after which another two equivalents of N-methylmorpholine (0.06 ml, 0.56 mmol) were added followed by the deprotected mono butyl mono N-Bocpiperidine perylene bisimide 9 (300 mg, 0.27 mmol) and 3 (81 mg, 0.27 mmol). Stirring was continued for 1h at 0 C, and overnight at room temperature. CH 2 Cl 2 (50 ml) was added and the solution was washed with, respectively, 1M aq. HCl (2 x 20 ml), brine (1x 20 ml), saturated aqueous bicarbonate solution (1 x 20 ml) and H 2 (1 x 20 ml). The organic phase was dried on Na 2 S 4 and the solvent was evaporated. The resulting solid crude product was purified using column chromatography (2 % MeH in CH 2 Cl 2, Si 2 ), providing a dark red solid (20 mg, mmol, 4.4 %) 1 H NMR (400 MHz, CDCl 3 ) δ = 8.20 (s, 2H), 8.19 (s, 2H), 7.61 (s, 1H), 7.29 (d, J = 9.2 Hz, 1H), 7.23 (d, J = 8.9 Hz, 8H), 6.94 (s, 1H), (m, 11H), (m, 1H), 4.48 (s, 2H), (m, 2H), 3.82 (s, 3H), (m, 8H), 2.94 (s, 2H), (m, 6H), 2.15 (s, 2H), 2.02 (s, 2H), (m, 1H), 1.61 (s, 3H), (m, 4H), (m, 2H), 1.29 (s, 18H), 1.82 (s, 18H), 0.93 (t, J = 7.3 Hz, 3H) ppm 13 C NMR (101 MHz, CDCl 3 ) δ = 168.8, 164.0, 163.6, 163.6, 163.3, 162.5, 162.1, 156.3, 156.1, 155.4, 153.1, 153.0, 147.6, 142.0, 139.6, 138.7, 135.6, 135.6, 135.5, 135.5, 133.5, 133.1, 133.0, 132.6, 131.2, 130.2, 128.7, 126.9, 122.8, , 120.9, 120.5, 120.3, 120.0, 119.9, 119.7, 119.6, 119.6, 119.5, 113.2, 112.7, 100.7, 55.9, 51.8, 46.1, 42.1, 40.6, 38.2, 37.9, 34.6, 34.4, 31.7, 30.4, 28.7, 23.3, 20.6, 14.7, 14.4, 14.0 ppm. MALDI-TF MS (MW = ) m/z = [M + ]. (11) PipSpip 11 was synthesized using a procedure similar to that for CSC 5 using piperidine as amine and starting from 4 (200 mg, 0.58 mmol). Purification provided pipspip as a cream solid (42 mg, mmol, 15 %) 1 H NMR (400 MHz, CDCl 3 ) δ = 6.84 (s, 2H), (m, 8H), 2.76 (t, J = 7.4 Hz, 4H), 2.05 (s, 6H), (m, 2H), (m, 4H), (m, 8H) ppm. 13 C NMR (101 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (d), 37.9 (t), 26.0 (t), 24.6 (t), 22.9 (t), 14.3 (q) ppm. MS(EI) for C 27 H 34 N 2 2 S 2 m/z 482 [M + ], HRMS calcd for C 27 H 34 N 2 2 S 2 : , found:

99 Chapter References 1 (a) B. Regan, M. Graetzel, Nature 1991, 353, (b) A. Hagfeldt and M. Graetzel, Acc. Chem. Res. 2000, 33, (a) M. R. Wasielewski, J. rg. Chem. 2006, 71, (b) F. D. Lewis, R. L. Letsinger and M. R. Wasielewski, Acc. Chem. Res. 2001, 34, (c) M. R. Wasielewski, Chem. Rev. 1992, 92, (a) A. P. de Silva, H. Q. N. Gunaratne, T. Gunnlaugsson, A. J. M. Huxley, C. P. McCoy, J. T. Rademacher and T. E. Rice, Chem. Rev. 1997, 97, (b) J. F. Callan, A. P. de Silva and D. C. Magri, Tetrahedron 2005, 61, (a) X. Hu, A. Damjanovic, T. Ritz and K. Schulten, Proc. Natl. Acad. Sci. U.S.A. 1998, 95, (b) G. McDermott, S. M. Prince, A. A. Freer, A. M. Hawthornthwaite- Lawless, M. Z. Papiz, R. J. Cogdell and N. W. Isaacs, Nature 1995, 374, (c) R. J. Cogdell, A. T. Gardiner, A. W. Roszak, C. J. Law, J. Southall and N. W. Isaacs, Photosynth. Res. 2004, 81, (d) C. J. Law, A. W. Roszak, J. Southall, A. T. Gardiner, N. W. Isaacs and R. J. Cogdell, Mol. Membr. Biol. 2004, 21, D. Gust, T. A. Moore, and A. L. Moore, Chem. Commun. 2006, N.J. Turro, Modern Molecular Photochemistry (University Science Books, Sausalito, 1991 ) 7 B. W. Van Der Meer, G. Coker III and S.-Y. S. Chen, Resonance Energy Transfer, Theory and Data, VCH, Weinheim, G. D. Scholes, Annu. Rev. Phys. Chem. 2003, 54, J. H. Hurenkamp, W. R. Browne, R. Augulis, A. Pugzlys, P. H. M. van Loosdrecht, J. H. van Esch and B. L. Feringa, rg. Biomol. Chem. 2007, 5, R. Augulis, A. Pugžlys, J. H. Hurenkamp, B. L. Feringa, J. H. van Esch and P. H. M. van Loosdrecht, J. Phys. Chem. 2007, 111, (a) K. Matsuda and M. Irie, J. Photochem. Photobiol. C, 2004, 5, (b) F. M. Raymo and M. Tomasulo, Chem. Soc. Rev. 2005, 34, (c) F. M. Raymo and M. Tomasulo, J. Phys. Chem. A 2005, 109, (d) F. M. Raymo and M. Tomasulo, Chem. Eur. J. 2006, 12, P. Belser, L. De Cola, F. Hartl, V. Adamo, B. Bozic, Y. Chriqui, V. M. Iyer, R. T. F. Jukes, J. Kuhni, M. Querol, S. Roma and N. Salluce, Adv. Funct. Mat. 2006, 16, (a) R. A. van Delden, N. P. M. Huck, J. J. Piet, J. M. Warman, S. C. J. Meskers, H. P. J. M. Dekkers, and B. L. Feringa, J. Am. Chem. Soc. 2003, 125, (b) N. P. M. Huck and B. L. Feringa, J. Chem. Soc., Chem. Commun. 1995,

100 Tuning Energy Transfer in Switchable Donor-Acceptor Systems 14 W. R. Browne, M. M. Pollard, B. de Lange, A. Meetsma and B.L. Feringa, J. Am. Chem. Soc. 2006, 128, S. Abad, M. Kluciar, M. A. Miranda and U. Pischel, J. rg. Chem. 2005, 70, T. Komatsu, K. Kikuchi, H. Takakusa, K. Hanaoka, T. Ueno, M. Kamiya, Y. Urano and T. Nagano, J. Am. Chem. Soc. 2006, 128, (a) H. Tian and S. J. Yang, Chem. Soc. Rev. 2004, 33, (b) M. Irie, Chem. Rev. 2000, 100, L. N. Lucas, J. J. D. de Jong, J. H. van Esch, R. M. Kellogg and B. L. Feringa, Eur. J. rg. Chem., 2003, 1, (a) A. Fernandez-Acebes and J. M. Lehn, Chem. Eur. J. 1999, 5, (b) A. Fernandez-Acebes and J. M. Lehn, Adv. Mater. 1998, 10, (a) M. Irie, T. Fukaminato, T. Sasaki, N. Tamai and T. Kawai, Nature 2002, 420, (b) T. Fukaminato, T. Sasaki, T. Kawai, N. Tamai and M. Irie, J. Am. Chem. Soc. 2004, 126, (c) L. Giordano, T. M. Jovin, M. Irie and E. A. Jares-Erijman, J. Am. Chem. Soc, 2002, 124, L. Sun, S. Wang and H. Tian, Chem. Lett. 2007, 36, T. B. Norsten and N.R. Branda, J. Am. Chem. Soc. 2001, 123, J. Walz, K. Ulrich, H. Port, H. C. Wolf, J. Wonner and F. Effenberger, Chem. Phys. Lett. 1993, 213, (a) A. Hartschuh, I. B. Ramsteiner, H. Port, J. M. Endtner and F. Effenberger, J. Lumin. 2004, 108, (b) J. M. Endtner, F. Effenberger, A. Hartschuh and H. Port, J. Am. Chem. Soc, 2000, 122, (c) P. A. Liddell, G. Kodis, A. L. Moore, T. A. Moore and D. Gust, J. Am. Chem. Soc. 2002, 124, A. Adronov, S. L. Gilat, J. M. J. Fréchet, K. hta, F. V. R. Neuwahl and G. R. Fleming, J. Am. Chem. Soc. 2000, 122, (a) J. J. D. de Jong, L. N. Lucas, R. M. Kellogg, J. H. van Esch and B. L. Feringa, Science 2004, 304, (b) Ph.D. Thesis L. N. Lucas, Towards Photoresponsive Supramolecular Materials 2001, University of Groningen, ISBN G. Anderson and R. Paul, J. Am. Chem. Soc. 1958, 80, (a) F. Würthner, C. Thalacker, S. Diele and C. Tschierske, Chem. Eur. J. 2001, 7, (b) R. Iden and G. Seybold, (BASF AG), Ger. Pat. Appl., DE A1, 1985, (Chem. Abstr. 1985, 103, 38696q). (c) D. Dotcheva, M. Klapper and K. Müllen, Macromol. Chem. Phys. 1994, 195, F. Würthner, A. Sautter, D. Schmid and P. Weber, Chem. Eur. J. 2001, 7, F. Würthner, B. Hanke, M. Lysetska, G. Lambright and G. Harms, rg. Lett. 2005, 7,

101 Chapter 3 31 Irradiation was carried out at 220 K to suppress bimolecular reactions, in particular interference by traces of water. 32 Some degradation is visible per cycle, with the most significant decrease in intensity of the absorption of the closed state at λ = 493 nm after the first cycle, after which the intensity is seen to stabilize. 33 Cross-correlated: the lifetime of the process is less than the FWHM of the excitation pulse. 34 HPLC separation of pipspip 11 was performed on an Alltech Econosphere Silica 10 µm column using n-heptane : 2-propanol 95:5 and a flow of 1.0 ml/min. The retention times for the open and closed form were 29.4 and 32.3 min, repectively. The PSS was determined at λ = 304 nm, an isoabsorptive point for both forms. 35 The sharp features between λ = 450 and 500 nm are instrumental artifacts, the spectra are uncorrected for variations in lamp output intensity. 36 The feature at λ ~ 600 nm is caused by sensitivity of the perylene unit to solvent polarity. Cooling freezes out residual water present in the CH 2 Cl 2, thereby decreasing solvent polarity and causing a small shift in the perylene absorption after the second measurement, which as a result causes this spectral distortion. 37 Irradiation of PSC 10 at room temperature resulted in significant photodegradation of the PSC triad 10 (see also Figure 3.13). However, at 220 K, degradation is suppressed and switching of the dithienylcyclopentene component is observed. 38 (a) E. A. A. Wallén, J. A. M. Christiaans, E. M. Jarho, M. M. Forsberg, J. I. Venäläinen, P. T. Männisto and J. Gynther, J. Med. Chem. 2003, 46, (b) L. A. Carpino, E. M. E. Mansour, C. H. Cheng, J. R. Williams, R. Macdonald, J. Knapczyk, M. Carman and A. Łopusiński, J. rg. Chem. 1983, 48, (a) G. Seybold and G. Wagenblast, Dyes Pigm. 1989, 11, (b) G. Seybold and A. Stange, (BASF AG), Ger. Pat., DE , 1987, (Chem. Abstr. 1988, 108, 77134c). 100

102 Chapter 4 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification The synthesis and characterisation of dithienylcyclopentene switches, with two different coumarin substituents, is described. Both functionalized switches show the reversible ring opening and closing reaction characteristic for dithienylcyclopentenes. However, only one shows reversible quenching of the coumarin fluorescence, which is unexpected considering the small differences in coumarin structure and electronic properties of the system. The reason for this large difference in energy transfer efficiency was investigated using several spectroscopic techniques as well as cyclic voltammetry and the results are rationalised in terms of the Förster and Dexter energy transfer mechanism. 101

103 Chapter Introduction The demand for portable storage capacity, with storage density as high as possible, is rapidly approaching the limits of the top down approach of data storage (e.g. photolithography, 1 magnetic 2 / optical data storage 3 ). To achieve even higher storage densities a bottom up approach may be more suitable. 4 The information density that could, potentially, be reached by using addressable individual molecules provides a strong incentive for this molecular approach. However, it also poses significant challenges with respect to addressability and fatigue resistance. ver the past decades, a structurally diverse range of switchable molecular systems have been reported. 5 These molecules have at least two states that can be addressed separately and, preferably, have a read-out method, which does not influence the molecular state (i.e. non-destructive). Photochromic molecular switches (e.g. fulgides, 6 spiroindolizines, 7 spiropyrans 7 and dithienylcyclopentenes 8 ) have shown considerable potential. The ability to change the absorption spectrum through a structural change, induced by an external stimulus (e.g., ph, 9, redox, 10 irradiation 5 or temperature 11 ), makes it possible to control optical properties in a reversible manner to provide functionality in a system. For example, the combination of efficient fluorophores (e.g., laser dyes) with photochromic switches, 12 thereby having a unit with a high quantum yield of fluorescence (Φ fl 1) for read out purposes, combined with an efficient switch to influence emission output, provides a highly promising approach (Figure 4.1). The properties of both parts can be optimised separately to get efficient quenching and good results have been obtained already by combining efficient chromophores (i.e. porphyrin, coumarin, and anthracene) with photochromic switches. 12 Figure 4.1 Schematic representation of a fluorophore photochromic switch dyad. A) When the photochromic unit is in State 1, the fluorophore (F) is excited by λ1 and emits light at longer wavelengths (λ2). B) When the photochromic unit is in State 2, F is still excited by λ1, however this excited state is quenched by the photochromic unit and emission is not observed. 102

104 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification Dithienylcyclopentenes belong to a class of photochromic switches which show potential as photoswitchable fluorescence quenching units (Figure 4.2). The synthesis routes, which have been developed earlier, facilitate attachment of substituents (in this case chromophores). Dithienylcyclopentenes show good stability, good to very good photostationary state (PSS) and two states, with very distinct absorption spectra, which are both thermally stable. 8 Dithienylcyclopentenes already have been used successfully as switchable fluorescence quenchers, examples of which are to be found in the work of the groups of Lehn, 13 Irie, 14 Tian, 15 Branda 16 and others. 17 Figure 4.2 Schematic representation of the ring open and ring closed state of a dithienylcyclopentene switch. In Chapter 3 we demonstrated that a coumarin dithienylcyclopentene coumarin triad system, with amide connecting units, shows up to 50% quenching of the coumarin emission in the closed (or rather PSS) state. However, this system has some disadvantages that render it unsuited for practical applications. The system shows considerable degradation at room temperature (not at 220 K) and the low PSS increases the probability of readout errors (i.e. when used for data storage). Adding phenyl spacers between the dithienylcyclopentene unit and the amide connecting units improves the stability considerably as well as the PSS that can be reached at room temperature. 18 Previous synthetic experience within our group has provided us with suitable starting components, both chromophores and phenyl substituted dithienylcyclopentenes, 19,20 which can be combined to build molecules that meet the requirements needed for efficient fluorescence N/FF switching. However, it will become clear that even if components seem to meet the basic photophysical requirements, variation in one component can hold consequences for the operation of the second component. In this chapter two coumarin-dithienylcyclopentene systems are described, in which a minor change in the structural design of one of the components results in a considerable difference in spectroscopic properties (Figure 4.3). In the first system efficient quenching of the coumarin excited state energy in both the open and the closed state is observed. In the second system a structural change in the coumarin component, returns the high fluorescence output in the open state, which is combined with efficient quenching (>98%) 103

105 Chapter 4 in the closed state, thereby showing the properties of a very efficient fluorescence N / FF switch. Both systems were investigated using a range of spectroscopic techniques (i.e. UV/Vis and fluorescence spectroscopy, cyclic voltammetry and time correlated single photon counting) to understand why such an apparently minor structural change leads to very different photophysical behaviour and to gain insight into the energy transfer processes taking place in these systems. Previously, it has been shown that ring closing of the dithienylcyclopentene switch can be achieved by electrochemical oxidation in solution and both ring-opening and closing can be achieved when the switch is immobilised on an IT surface. 21 Taken together this opens the possibility of electrochemical switching of fluorescence output. Figure 4.3 Ring-open and ring-closed state of the coumarin-dithienylcyclopentene systems (3 and 8) and the respective effects on the fluorescence. Figure 4.4 Coumarin-substituted dithienylcyclopentene switches and switch/coumarin models investigated in this chapter. 104

106 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification 4.2 Synthesis Amide coupling chemistry with acid functionalized dithienylcyclopentene switches has proven to be a versatile strategy in building multicomponent systems because it is a robust approach to covalent connectivity and precludes through bond intercomponent interactions. 20,22 In the synthesis of the symmetric donor-acceptor systems described in Chapters 2 and 3, amide coupling chemistry was employed successfully to connect donor fluorophores to carboxylic acid functionalized dithienylcyclopentenes and perylenes. In this chapter, the amide coupling approach is employed again to connect the coumarin donor molecules to mono acid functionalised diphenyldithienylcyclopentenes. The carboxylic acid functionalized diphenyl switch 1 was synthesized as reported previously. 19 The synthesis of the piperazine functionalized 7-methoxy-coumarin-3-acetic acid 2 is described in Chapter 2. The mono-methoxy coumarin functionalized diphenyldithienylcyclopentene 3 was synthesized via coupling of the mono acid 1 to the 7-methoxy-coumarin-3-acetyl piperazine 2 using 2-chloro-4,6-dimethoxy-1,3,5-triazine (CDMT) and N-methyl-morpholine (NMM) in CH 2 Cl 2, providing the monomethoxy coumarin switch (MMCS) 3 in 26% yield (Figure 4.5). 22 S + S H N HN 1 2 A S 3 S Figure 4.5 Synthesis of the mono-methoxy coumarin functionalized diphenyl switch 3 : A) i) 1 CDMT, NMM, CH 2 Cl 2, 0 o C. ii) NMM, 2 (26%). As will be discussed below, MMCS 3 did not meet the expectations of a fluorescence switching unit (Figure 4.3, vide infra). It was suspected that this was due to unfavourable energetics of the mono methoxy coumarin relative to the diphenyl switch. To solve this problem the coumarin absorption maximum needed to be shifted towards lower energy (vide infra: data and discussion). N N 105

107 Chapter 4 A literature search revealed that a minor modification to the coumarin, i.e. the addition of an extra methoxy group, might have the desired effect on the coumarin absorption maximum. 23 Therefore 6,7-dimethoxy-coumarin-4-acetic acid 4 was synthesized 23 and coupled to diphenyl switch 1 using a method similar to that employed for 3. The position of attachment for the acetic acid also changes, however, for the mono methoxy coumarin with an acetic acid on the four position spectroscopic characteristics are the same as for the mono methoxy coumarin with the attachment to the three position, so this change is expected to have little or no influence. The 6,7-dimethoxy-coumarin-4-acetic acid 4 was coupled to N-Boc-piperazine 5 using 1,1'-carbonyldiimidazole (CDI). This coupling reagent facilitates purification, as was observed with the related reaction with 7-methoxy-coumarin-3-acetic acid in Chapter After deprotection, the piperazine functionalized dimethoxy coumarin was coupled to the mono acid diphenyl dithienylcyclopentene 1 using CDMT and NMM in CH 2 Cl 2 providing the dimethoxy coumarin switch (DMCS) 8 in a moderate yield (46%, Figure 4.6). Figure 4.6 Synthesis of the di-methoxy coumarin functionalized diphenyl switch 8: A) CDI, CH 2 Cl 2, RT; B) CF 3 CH, CH 2 Cl 2 ; C) i) 1, CDMT, NMM, CH 2 Cl 2, 0 o C. ii) NMM, 7 (46%). All compounds were purified by column chromatography and characterized by 1 H and 13 C NMR spectroscopy and (MALDI-TF) mass spectroscopy (see experimental section for details). 106

108 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification 4.3 Electronic properties Monomethoxycoumarin-dithienylcyclopentene switch 3 (MMCS) UV/Vis spectra of MMCS 3 in the open and closed form (at the photostationary state (PSS) λ irr = 365 nm) are shown in Figure 4.7. The spectra show features of both coumarin and model switch (open or PSS 365 nm ) and the maxima closely match those of the model compounds MMCpipboc 9 (N-Boc protected 2) and PrPhSPh 10 (Figure 4.7 and Table 4.1), which would indicate little or no electronic communication between coumarin and dithienylcyclopentene unit. The changes observed in the spectra of MMCS 3 and in the spectra of the model compound, PrPhSPh 10, upon photochemical ring closure from the open state to the closed state (PSS 365 nm ) are nearly identical. 25 The similarities are more apparent by comparison of the difference spectra shown in Figure 4.8, confirming that the change seen in the absorption spectrum of 3 is due to the switching of the dithienylcyclopentene unit. The change in the absorption spectra show that the photochromism of the diphenyldithienylcyclopentene unit is retained in MMCS 3 and an excellent PSS (>95% open : closed ratio) is being achieved. Furthermore, the absorption of the coumarin moiety in 3 is unaffected by ring closure indicating that no direct electronic communication between coumarin and dithienylcyclopentene unit is present ε (cm -1 M -1 ) ε (cm -1 M -1 ) Wavelength (nm) Wavelength (nm) Figure 4.7 Left: UV/Vis spectra of MMCS 3 in the open state ( ), PSS 365 nm ( ) and MMCpipboc 9 ( ), spectra recorded in CH 2 Cl 2 at RT. Right: UV/Vis spectra of PrPhSPh 10 in the open state ( ) and PSS 365 nm state ( ), spectra recorded in CH 2 Cl 2 at RT. 107

109 Chapter 4 Table 4.1 Absorption and emission data of the dithienylcyclopentene switches in the open state and at the PSS 365 nm and of the appropriate model compounds. Compound Absorbance a λ max / nm (10 3 ε / cm -1 M -1 ) Emission a λ max / nm MMCpipboc 9-323(18.4) DMCpipboc 6 293(4.9) 346(11.4) PrPhSPh (25.5) PrPhSPh PSS 365 nm 10 b 287(22.9) - 372(9.3) 547(12.4) 394 MMCS pen 3 294(34.1) 318(35.9) MMCS PSS 365 nm 3 b 294(30.4) - 370(10.6) 541(15.6) 394 DMCS pen 8 287(29.3) 321(24.6) DMCS PSS 365 nm 8 b 285(25.9) - 353(17.0) 541(14.9) 425 a Recorded in CH 2 Cl 2 at RT. b After irradiation with λ = 365 nm light at RT till PSS was reached Abs Wavelength (nm) Figure 4.8 The UV/Vis difference spectra of MMCS 3 ( ) and PrPhSPh 10 (, normalized) obtained by subtraction of the spectrum of the open state from the spectrum of the PSS 365 nm state, spectra recorded in CH 2 Cl 2 at RT. 108

110 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification Surprisingly in the open form the fluorescence of MMCS 3 is weak ( < 5%) compared to that of free MMCpipboc 9. At the PSS 365 nm state the position of the emission band shifts towards the blue and the intensity decreases further (Figure 4.9, left). 35.0k 40.0k 1.0M 30.0k 25.0k 35.0k 30.0k 800.0k CPS 20.0k 15.0k 10.0k 5.0k ε (cm -1 M -1 ) 25.0k 20.0k 15.0k 10.0k 5.0k 600.0k 400.0k 200.0k CPS Wavelength (nm) Wavelength (nm) 0.0 Figure 4.9 Left: Fluorescence spectra of MMCS 3 in the open state ( ) and PSS 365 nm state ( ), spectra recorded in CH 2 Cl 2 at RT. Right: UV/Vis spectra (left side) and fluorescence spectra (right side) of MMCS 3 in the open state ( ) and free MMCpipboc 9 ( ). Fluorescence spectra were corrected for absorption at the excitation wavelength. The spectra were recorded in CH 2 Cl 2 at RT. When comparing the fluorescence intensity of the free MMCpipboc 9 with the fluorescence intensity of MMCS 3 in the open state there is almost no fluorescence visible for the open form of MMCS 3 (Figure 4.9, right). This shows that already in the open form of MMCS 3 the fluorescence of the coumarin is almost fully quenched by the dithienylcyclopentene unit, making it unsuitable for fluorescence switching (i.e. only FF/FF states are available) Dimethoxycoumarin dithienylcyclopentene switch 8 (DMCS) The UV/Vis absorption spectra of DMCS 8 in the open and closed form (at the photo stationary state (PSS 365 nm )) are shown in Figure 4.10 left. As for MMCS 3, the spectra show features of both the dimethoxy coumarin component and of the open or PSS 365 nm model switch, respectively, and the absorption maxima match those of the spectrum obtained by adding the individual spectra of model compounds 6 and 10 (Table 4.1, Figure 4.23). 109

111 Chapter ε (cm -1 M -1 ) Abs Wavelength (nm) -0.2 Abs Wavelength (nm) Figure 4.10 Left: UV/Vis absorption spectra of DMCS 8 in the open state ( ),DMCS 8 PSS 365 nm ( ) and DMCpipboc 6 ( ), spectra recorded in CH 2 Cl 2 at RT. Right: The UV/Vis difference spectrum for DMCS at the open and closed PSS 365 nm state. The spectra were recorded in CH 2 Cl 2 at RT. The difference spectra (between the open and PSS 365 nm states) of DMCS 8, of PrPhSPh 10 and of MMCS 3 are similar, (Figure 4.8) confirming that the change seen in the absorption spectrum is due to the switching of the dithienylcyclopentene unit (Figure 4.10, Right) and, importantly, that the photochromic properties of the dithienylcyclopentene unit are unaffected by the coumarin unit in this dyad also pen 0 Closed 1 pen 1 Closed 2 pen 2 Closed 3 pen 3 Closed 4 pen 4 Abs Wavelength (nm) 0.0 pen 0 Closed 1pen 1 Closed 2 pen 2 Closed 3 pen 3Closed 4 pen 4 Cycle # Figure 4.11 Left: UV/Vis absorption spectra of DMCS 8 recorded over several switching cycles. DMCS 8 in the open state is converted to the DMCS closed state (PSS 365 nm ) by irradiation at λ = 365 nm, the closed isomer is returned to the open state by irradiation with λ > 400 nm light. The spectra were recorded in CH 2 Cl 2 at RT. Right: Absorbance at λ = 541 nm for the DMCS 8 open and PSS 365 nm state during four switching cycles. The spectra were recorded in CH 2 Cl 2 at RT. 110

112 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification The photochemical switching can be performed with DMSC 8 over several open/closing cycles (Figure 4.11) and shows very efficient switching at RT with a high PSS of up to 95%, which is comparable to the switching efficiency of MMCS 3 and PrPhSPh 10. The emission spectra of the open form of DMCS 8 match closely that of a 1:1 mixture of the individual components 6 and 10, showing, in contrast to MMCS 3, that the fluorescence of the coumarin component is not quenched by the open switch. At the PSS 365 nm of DMCS 8, i.e. in the ring closed form, a large decrease in coumarin fluorescence intensity is observed, due to quenching by the closed switch (Figure 4.12, left). This decrease in fluorescence is > 98%, i.e. when closed to the photostationary state using λ = 365 nm light, indicating that energy transfer to the switch is very efficient. The energy transfer quenching observed is typical for Förster resonance energy transfer, 26 which is probable considering the considerable spectral overlap of the emission spectrum of the coumarin and the absorption spectrum of the closed switch unit. 2.0M 2.5M 2.0M 1.5M 1.5M CPS 1.0M CPS 1.0M 500.0k 500.0k Wavelength (nm) pen 0 Closed 1 pen 1 Closed 2 pen 2 Closed 3 pen 3 Closed 4 pen 4 Cycle # Figure 4.12 Left: fluorescence spectra (λ exc = 380 nm) of DMCS 8 in the open state ( ) and PSS 365 nm state ( ). The spectra were compensated for absorption and recorded in CH 2 Cl 2 at RT. Right: Fluorescence intensity (λ exc = 380 nm) at λ = 420 nm over several switching cycles of DMCS 8 using λ = 365 nm for ring-closing and λ > 400 nm at RT for ring-opening. The spectra were recorded in CH 2 Cl 2 at RT. The switching of fluorescence was followed over four switching cycles and is found to be reversible, with high efficiency (Figure 4.12, right). However, some decrease in the maximum fluorescence for the open form is observed, probably due to an irreversible rearrangement of the closed switch, which generates a species with absorption at slightly higher energy (λ max ~ 525 nm, Figure 4.12, right). The remaining absorption at shorter wavelength (i.e. after opening the switch with λ > 400 nm) supports the latter conclusion (see Figure 4.11). A type of rearrangement which results in this type of absorption shift was observed before by the groups of Branda and Irie

113 Chapter 4 S Closed S R hv S S Rearranged R Figure 4.13 Probable side-product obtained when irradiating Redox Properties MMCS 3 The redox chemistry of the open form of MMCS 3 is characterized by an irreversible oxidation at E p,a = 1.22 V [vs SCE], as obtained in related dithienylcyclopentene switches, 21 and an irreversible oxidation at 1.78 V [vs SCE] (not shown), the edge of the solvent window, as for MMCpipboc 9. In the return cycle two new reduction processes are observed at potentials coincident with those of the closed form. This indicates that oxidative ring closure to MMCS(closed) 2+ is occurring. MMCS(closed) 2+ can then be reduced, first to MMCS(closed) + at 0.73 V and finally to MMCS(closed) at 0.40 V (Figure 4.14). Repetitive cycling results in a significant build-up of the closed form (MMCS open closed) within the diffusion layer of the electrode, as seen in the second cycle where two new redox processes (0.48 V and 0.81 V), originating from the closed form of MMCS 3, are observed. The electrochemical characteristics together with the absence of redox process between -2.0 V and 1.4 V for the coumarin model MMCpipboc 9, provide compelling evidence that the oxidative processes observed are centred on the dithienylcyclopentene unit. Table 4.2 Redox properties of the open and closed form of MMCS 3 and DMCS 8. MMCpipboc 9 DMCpipboc 6 E 1/2, V vs SCE (E p,a where irr or qr) a,b 1.78 (irr) c 1.48 (qr) MMCS 3 pen 1.78 (irr) c, 1.22(irr) Closed 1.78(irr) c, 0.77, 0.43, -1.74(irr) DMCS 8 pen 1.50(qr), 1.18(irr) Closed 1.50(qr), 0.78, 0.43, -1.72(irr) a in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE b irr=irreversible qr = quasi reversible 28 c edge of solvent window. 112

114 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification -40.0µ Mo (Mo 2+ ) Mc µ Current (A) -20.0µ -10.0µ µ Mc Mc + Mc 2+ Mc Mc + Mc µ Potential (V vs SCE) Initial scan Figure 4.14 Cyclic voltammetry of the open form of MMCS 3 (Mo = MMCS open, Mc = MMCS closed) in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE at 1 V s µ -10.0µ Mc Mc + Mc µ Current (A) µ 10.0µ Mc Mc + Mc µ Mc 2- Mc 20.0µ Potential (V vs SCE) Figure 4.15 Cyclic voltammetry of the closed form of MMCS 3 (Mc) in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE at 0.2 V s

115 Chapter 4 In the closed state, MMCS 3 shows less anodic redox potentials than in the open state, which indicates destabilisation of the HM in the closed state. For the closed form of MMCS two fully reversible oxidation processes are observed between 0.25 and 1.0 V [vs SCE], assigned to two one-electron oxidation steps (Figure 4.15). These two oxidation processes are in good agreement with the two oxidation processes that appear when oxidizing the open form and were assigned to the closed form. Similarly, two irreversible reduction steps are observed between -1.5 and -2.2 V [vs SCE] (the second reduction is not shown) DMCS 8 The redox chemistry of DMCS 8 is similar to that of MMCS 3 and is characterized by an irreversible oxidation at E p,a = 1.18 V [vs SCE] leading to oxidative ring closure to DMCS(closed) 2+. This component can then be reduced, first to DMCS(closed) + at 0.74 V and finally to DMCS(closed) at 0.41 V (Figure 4.16). Also a quasi-reversible oxidation at 1.50 V [vs SCE] is observed (Figure 4.18), at the same potential as for DMCpipboc 6. Repetitive cycling results in a significant build-up of the closed form (DMCS open closed) in the diffusion layer of the electrode, shown in the second cycle by two new redox processes (0.46 V and 0.79 V), originating from oxidation of the closed form of DMCS µ -15.0µ Do (Do 2+ ) Dc 2+ Current (A) -10.0µ -5.0µ 0.0 Dc Dc + Dc µ 10.0µ Dc Dc + Dc 2+ Initial scan Potential (V vs SCE) Figure 4.16 Cyclic voltammetry of the open form of DMCS 8 (DMCS open (Do), DMCS closed (Dc)) in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE at 0.2 V s -1. In the closed state, DMCS 8 shows less anodic redox potentials than in the open state and hence destabilisation of the HM. For the closed form of DMCS two fully reversible 114

116 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification oxidation processes are observed between 0.25 and 1.0 V (vs SCE), assigned to two oneelectron oxidation steps (Figure 4.17). These two oxidation processes are in good agreement with the two oxidation processes that appear when oxidizing the open form and were assigned to the closed form, identical as observed for MMCS 3. Similarly, two irreversible reduction steps are observed between -1.5 and -2.2 V [vs SCE] (not shown) µ Dc Dc + Dc µ Current (A) µ Dc Dc + Dc µ Dc 15.0µ Dc 2- Potential (V vs SCE) Figure 4.17 Cyclic voltammetry of the closed form of DMCS 8 (Dc) in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE at 0.1 V s -1. When the potential window is extended to 1.5V [vs SCE] another quasireversible redox wave is observed for DMCS(closed) 2+ DMCS(closed) 3+, due to oxidation of the coumarin moiety (also observed for the coumarin model DMCpipboc 6, not shown). This third oxidation makes the cation insoluble in CH 2 Cl 2 and results in its precipitation on the electrode surface. As a result on the subsequent cathodic sweep a surface confined (nondiffusion dependent) process takes place, observed as a desorption spike at 0.68 V, where the precipitated polycationic molecules are reduced to the better soluble DMCS(closed) 1+ state, allowing it to desorb from the surface of the electrode and return to solution (Figure 4.18). This shows that it is important to stay away from the coumarin oxidation when closing the switch by cyclic voltammetry, since adsorption on the surface of the electrode might hinder the closing process for the diarylcyclopentene switches which are still in solution. 115

117 Chapter µ Dc 2+ Dc µ Dc Dc + Dc µ Current (A) µ 10.0µ Dc 3+ Dc 2+ (qr) 15.0µ Dc 2- Dc Desorption spike 20.0µ Dc Dc + Dc Potential (V vs SCE) Figure 4.18 Cyclic voltammetry of the closed form of DMCS 8 (Dc) in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE at 0.1 V s Spectroelectrochemistry In order to further investigate the possibility of opening and closing the switch by cyclic voltammetry and thereby to be able to use is as an electrochemical fluorescence switch, the system was investigated using spectroelectrochemistry. The investigation showed similar behaviour as for phenyl and p-methoxyphenyl substituted dithienylcyclopentene, which were investigated previously within our group. 21 When a solution of DMCS 3 in CH 2 Cl 2 is subjected to a constant potential of 1.2 V [vs Ag/Ag + ] the absorption spectrum shows similar changes as observed for the phenyl substituted dithienylcyclopentene (Figure 4.19). During the first phase (1) the formation of DMCS 2+ closed (Dc 2+ ) from DMCS open (λ max at 432 and 566 nm) is accompanied by some formation of the monocation (Dc +, λ max at 758 nm). After oxidation the slow formation of an unidentified species Dx is observed (peak at 646 nm). nce all DMCS open has been transformed into Dc 2+ and the electrolysis is stopped, the amount of Dc 2+ starts to decrease while the amount of Dc + remains constant. During this period the amount of Dx steadily increases. After 20 min a steady state is reached were the amount of Dc 2+ and Dc + reaches a minimum and for Dx a maximum amount is observed (Figure 4.19, state 3). The formation of Dx, which occurs when electrochemically closing the switch in solution, is problematic since it is, therefore, not 116

118 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification possible to reach a fully closed state (i.e. only Dc 2+ ) and thereby electrochemically switch from the open to the closed form. However, this behaviour is fully consistent with a previous study in solution of the switching unit. 21 It was already shown by Areephong et al. that it is possible to electrochemically switch these dithienylcyclopentenes when they are connected to a surface. 19 This results show that if we want an electrochemically controllable fluorescence switch (i.e. that can be reversibly opened and closed) the next step would need to be towards a functionalised surface Dc Do -2 e - Do 2+ Dc 2+ + e - - e - Dc + Dx Abs Dx 3 Dc Wavelength (nm) Figure 4.19 Spectroelectrochemistry of DMCS 3 in the open state during phase (1) ( ), phase (2), no electrolysis (- - -) and the final state (3) ( ). During 20 min at 1.2 V vs Ag/Ag + in CH 2 Cl 2 / 0.1 M TBAPF 6 at RT. 117

119 Chapter Excited state dynamics MMCS 3 Time-correlated single photon counting (TCSPC) measurements were performed on the open form and closed form (PSS 365 nm ) of MMCS 3. In the open state, the emission decay of the MMCS 3 (λ exc = 337 nm, λ em = 420 nm) is cross-correlated with the excitation pulse and the process has a fluorescence lifetime of lower than ~ 200 ps (Figure 4.20, left). This decay lifetime is due to the quenching of the coumarin excited state by energy transfer, most likely by a Dexter type mechanism (vide infra). At the PSS 365 nm state in which a mixture of the open and closed form of the switch is present, even though the PSS equilibrium of MMCS 3 is largely towards the closed form (>95%), the emission decay shows a long 2.0 ns and a short < 200 ps contribution to the decay (Figure 4.20, right). Figure 4.20 Left: TCSPC spectrum of MMCS 3 (open state) with: the decay trace of MMCS open, the fit to the fluorescence decay and IRF: the instrument response function. Right: TCSPC spectrum of MMCS (PSS 365 nm ) with: the decay trace of MMCS PSS 365 nm, the fit to the fluorescence decay and IRF: the instrument response function. The short contribution can be attributed to the residual coumarin emission during the lifetime of energy transfer from the coumarin to the closed dithienylcyclopentene unit. The long lifetime, which only has a small overall contribution to the trace, is most likely due to free coumarin, which is formed upon irradiation by degradation of the dithienylcyclopentene unit itself, thereby eliminating the quenching effect and allowing emission from a small amount of the coumarin. This conclusion is supported by a similar fluorescence decay lifetime for MMCpipboc 9 (1.4 ns). verall the bulk of the energy 118

120 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification absorbed by the coumarin is quenched by another component, most likely by the dithienylcyclopentene, in both the open and the closed state of MMCS DMCS 8 Time correlated single photon counting (TCSPC) measurements were performed on the open form and closed form (PSS 365 nm ) state of DMCS. In the open state, the time profile of the emission of the DMCS (λ exc = 337 nm, λ em = 420 nm) shows a long decay component of 2.3 ns, which is in good agreement with the fluorescence lifetime of DMCpipboc 6 (2.7 ns). Therefore, this long decay component is assigned to the dimethoxy coumarin connected to the open form of the dithienylcyclopentene switch (Figure 4.21, left). A short decay component, which is cross-correlated with the excitation pulse, can be observed and this process has a fluorescence lifetime of lower than ~ 200 ps. This fast decay, which is a minor component, is assigned to energy transfer of the coumarin to the small amount of closed form of the dithienylcyclopentene present in solution. 29 Figure 4.21 Left: TCSPC spectrum of DMCS 8 open with: the decay trace of DMCS open, the fit to the fluorescence decay recorded at λ em = 420 nm for 10 min and IRF: the instrument response function green. Right: TCSPC spectrum of DMCS closed (PSS 365 nm ) with: the decay trace of DMCS PSS 365 nm, green the fit to the fluorescence decay recorded at λ em = 420 nm for 10 min blue, and the instrument response function. At the PSS 365 nm state, in which a mixture of the open and closed form (>95% of closed form) of the switch 8 is present, the emission decay lifetimes were assigned as in the open form, however, the overall intensity has decreased (both spectra were obtained over 10 min accumulation time) (Figure 4.21, right). This indicates that the fluorescence lifetime of the dimethoxy coumarin remains the same (i.e. in the remaining open form) and the rate of energy transfer has not changed (i.e. the short component still has the same lifetime, ~ 200 ps), but also that part of the energy is being transferred elsewhere (since there are less 119

121 Chapter 4 counts coming from the coumarin than in the open form), in this case being quenched by the closed form of the switch (short component). The reversibility of the switching process fluorescence quenching can be followed using TCSPC by recording the decay traces over a set period of time, in this case 10 min. The low intensity irradiation used during the TCSPC assures that the state of the DMCS 8 solution (i.e. open or PSS 365nm ) is not changed during data acquisition (Figure 4.22). The open form of DMCS 8 can be irradiated to the closed form (PSS 365 nm ), then reopened using λ > 400 nm and finally irradiated again to the closed form of 8 (PSS 365 nm ). During this process the observed fluorescence output (e.g. counts per 10 min) is modulating, indicating that in the PSS 365 nm state of 8 the energy of the dimethoxy coumarin is being quenched by the closed form of the dithienylcyclopentene. The change in the extent of fluorescence quenching is reversible, 30 confirming that it is not due to photodegradation but due to the appearance and disappearance of the low energy absorption band of the closed dithienylcyclopentene pen 1500 Reopened λ > 400nm Counts Closed 2nd time λ = 365nm Closed 1st time λ = 365nm Time (ns) Figure 4.22 TCSPC spectra of DMCS 8, open and closed forms (PSS 365 nm ). PSS obtained by irradiating with λ = 365 nm light and the open form of DMCS obtained after irradiating the closed form (PSS 365 nm ) form with λ > 400 nm light for 20 min. Fluorescence decay counts measured at λ em = 420 nm for 10 min; all traces recorded in CH 2 Cl 2 at RT. 120

122 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification 4.7 Discussion Two structurally similar coumarin donor dithienylcyclopentene acceptor diads have been prepared and characterized by NMR and mass spectroscopy, i.e. MMCS 3 and DMCS 8, which differ only in the substitution pattern of the coumarin unit. This change in structure causes a bathochromic shift in the λ max of absorption from the monomethoxy coumarin to dimethoxy coumarin. A corresponding shift is observed in the emission properties of both coumarins and as a consequence the photophysical behaviour of both dyads (i.e. MMCS 3 and DMCS 8) is remarkably different. The absorption spectrum of the monomethoxy switch shows the spectral features of both MMCpipboc 9 and dithienylcyclopentene (Figure 4.9, right). When irradiated at λ = 365 nm the dithienylcyclopentene unit of MMCS 3 undergoes photochemical ring closure, which results in the appearance of a lower energy absorption at λ max = 540 nm. This photochemical ring closure reaches a very high PSS comprising > 95% of the closed form of 3 and also shows good stability at RT over several switching cycles. The initial concept was to use the appearance of this lower energy absorption band to quench the coumarin excited state. However, when considering the emission spectra of the open state and PSS of the MMCS 3 the fluorescence spectrum observed is of very low intensity compared with an equimolar coumarin model MMCpipboc 9 solution (Figure 4.9, Right). Surprisingly, in both the open state and the PSS, the excited state energy of the coumarin is quenched. This is indicated also by the large contribution of the short lifetime component (< 200 ps) observed with TCSPC for both the open and the closed state of 3 (PSS 365 nm ). In order to understand why this coumarin was not energetically compatible with the diphenyl dithienylcyclopentene, the absorption spectrum of the coumarin was compared with that of the model switch, PrPhSPh 10 (Figure 4.23). The absorption spectrum of PrPhSPh 10 overlaps fully with that of MMCpipboc and the lowest energy absorption is attributable to the dithienylcyclopentene unit and not to the coumarin. This situation makes it likely that the energy levels of the coumarin and switch unit are in a range compatible with Dexter type energy transfer, which requires overlap in wavefunction for the energy donor (i.e. the monomethoxy coumarin) and the energy acceptor (i.e. the dithienyl cylopentene). When an alternative coumarin is employed (i.e. DMCpipboc 6) the absorption of the open form of the dithienylcyclopentene is no longer the lowest energy absorption. Instead the dimethoxy coumarin is lowest in energy, thereby allowing the observation of coumarin emission in the open form. 121

123 Chapter 4 ε (cm -1 M -1 ) 25.0k 20.0k 15.0k 10.0k 1.2M 1.0M 800.0k 600.0k 400.0k CPS 5.0k 200.0k Wavelength (nm) Figure 4.23 Absorption spectra of PrPhSPh 10 ( ), MMCpipboc 9 ( ) and DMCpipboc 6 ( ). n the right side the normalized fluorescence spectra for MMCpipboc 9 ( ) and DMCpipboc 6 ( ) are also shown. The spectra were recorded in CH 2 Cl 2 at RT. The diphenyl dithienylcyclopentene substituted with the dimethoxy coumarin (DMCS 8) shows identical switching properties compared to the MMCS 3, with good stability and high switching efficiency (i.e. high PSS). In this compound, however, the open form of the DMCS 8 shows a strong emission at the same wavelength as the coumarin (i.e. DMCpipboc 6) emission (Figure 4.12) and the fluorescence decay lifetime observed for this emission is in good agreement to that of the free coumarin (2.7 ns). Irradiation of DMCS 8 to the closed form using λ = 365 nm light, results in the emission at λ = 420 nm decreasing in intensity by more then 98%. Due to the overlap of the emission band of the coumarin with the low energy absorption band (λ ~ 540 nm) of the closed form of the dithienylcyclopentene unit, the quenching is assigned to a Förster type mechanism. Fluorescence decay lifetime measurements at the PSS state still show a minor contribution of the longer lifetime (i.e. DMCpipboc 6) component, but the largest contribution is due to a < 200 ps component, which can be attributed to the quenched emission of the coumarin by the closed dithienylcyclopentene, as was observed for the MMCS also. pening the closed dithienylcyclopentene using λ > 400 nm light results in almost full recovery of the fluorescence output and this can be repeated over several switching cycles (Figure 4.12, right). This shows that the electronic properties of the coumarin donor have been 122

124 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification sufficiently altered to inhibit quenching by the dithienylcyclopentene open form and to allow for quenching only in the closed state by Förster resonance energy transfer. The redox properties of the individual units of MMCS 3 and DMCS 8, determined by cyclic voltammetry, give a good indication of the differences in HM and LUM levels in the two compounds. For the diphenyl dithienylcyclopentene unit the properties are the same for the three compounds (i.e. PrPhSPh 10, MMCS 3 and DMCS 8), as both first oxidation and first reduction take place at identical potentials (vs SCE). For the mono and dimethoxy coumarin (quasi)irreversible oxidations are observed at 1.6 V and 1.8 V (vs SCE) respectively. This shows that the different substitution pattern (i.e. the extra methoxy group) destabilises the HM level of the coumarin and brings it closer to the LUM, thereby decreasing the bandgap and causing a bathochromic shift in the absorption spectrum for the DMCpipboc Conclusions Investigation of MMCS 3 and DMCS 8 shows that by small changes to the relative energy levels of two components it is possible to change from a Dexter energy transfer mechanism to a Förster mechanism. In the present case, the introduction of an additional methoxy substituent, transforms a barely fluorescent photochromic switch (MMCS 3) into a molecule with up to 98% switchable fluorescence (DMCS 8). Using cyclic voltammetry it is also possible to close the dithienylcyclopentene switch in solution showing that the electrochemical properties of the switching unit are unaffected by the attachment of the fluorophore. n a surface it is has been shown earlier that it is possible to both open and close the dithienylcyclopentene unit electrochemically. 19 Combining the fluorescence switching properties of DMCS 8 in solution with the reversible switching of dithienylcyclopentenes on an IT surface, opens the possibility of electro- and photochemical control of fluorescence output. 4.9 Experimental section For all spectroscopic measurements Uvasol-grade solvents (Merck) were employed. All reagents employed in synthetic procedures were of reagent grade or better, and used as received unless stated otherwise. Jetsuda Areephong is acknowledged for providing compounds 1 and Compounds 2 20, 4 23 and 5 31 were prepared according to literature. 1 H NMR spectra were recorded at 200, 300, or 400 MHz; 13 C NMR spectra at 50.3, 75.4 or MHz. All spectra were recorded at ambient temperature, with the residual proton signals of the solvent as an internal reference. Chemical shifts are reported in ppm relative to TMS. CI and EI mass spectra were recorded on a Jeol JMS-600 mass spectrometer in the scan range of m/z with an acquisition time between 300 and 900 ms and a 123

125 Chapter 4 potential between 30 and 70 V. MALDI-TF spectra were recorded on an Applied Biosystems Voyager-DE Pro. UV/Vis absorption spectra (accuracy ±2 nm) were recorded on a Hewlett-Packard UV/Vis 8453 spectrometer. The fluorescence measurements were performed on a SPF-500C spectrofluorometer manufactured by SLM Aminco, and a Jobin- Yvon Fluorolog 3-22, the sharp features between λ = 450 and 500 nm in the excitation spectra are instrumental artefacts, the excitation and emission spectra are uncorrected for variations in lamp intensity and detector response. Sample concentration typically 10-5 M, spectra were recorded in 10 mm pathlength quartz fluorescence cuvettes. Electrochemical measurements were carried out on a Model 630B Electrochemical Workstation (CHInstruments). Analyte concentrations were typically mm in anhydrous dichloromethane containing 0.1 M TBAP. Unless otherwise stated, a Teflonshrouded glassy carbon working electrode (CHInstruments), a Pt wire auxiliary electrode and SCE or nonaqueous Ag/Ag + ion reference electrode were employed. Reference electrodes were calibrated with 0.1 mm solutions of ferrocene (0.38 V versus SCE in 0.1 M TBAP/ CH 3 CN). Solutions for reduction measurements were deoxygenated by purging with dry N 2 gas (presaturated with solvent) prior to the measurement. Cyclic voltammograms were obtained at sweep rates of between 10 mvs -1 and 50 Vs -1 ; differential pulse voltammetry (DPV) experiments were performed with a scan rate of 20 mvs -1, a pulse height of 75 mv and a duration of 40 ms. For reversible processes the half-wave potential values are reported; identical values were obtained from DPV and CV measurements. Redox potentials are given with an accuracy of ±10 mv. Luminescence lifetime measurements were obtained using an Edinburgh Analytical Instruments (EAI) time-correlated single-photon counting apparatus (TCSPC) comprised of two model J-yA monochromators (emission and excitation), a single photon photomultiplier detection system model 5300, and a F900 nanosecond flashlamp (N 2 filled at 1.1 atm pressure, 40 khz) interfaced with a personal computer via a Norland MCA card. A 400 nm cut off filter was used in emission to attenuate scatter of the excitation light (337 nm). Data correlation and manipulation was carried out using EAI F900 software version Emission lifetimes were calculated using a single-exponential fitting function, Levenberg-Marquardt algorithm with iterative deconvolution (Edinburgh instruments F900 software). The reduced χ 2 and residual plots were used to judge the quality of the fits. Lifetimes are given with an accuracy of ± 5%. 124

126 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification (3) 7-Methoxy-3-{2-[4-(4-{5-methyl-4-[2-(2-methyl-5-phenyl-thiophen-3-yl)-cyclopent- 1-enyl]-thiophen-2-yl}-benzoyl)-piperazin-1-yl]-2-oxo-ethyl}-chromen-2-one (MMCS) Mono carboxylic acid 1 (100 mg, 0.22 mmol) was suspended in CH 2 Cl 2 (20 ml) and placed in an ice bath. Subsequently N-methylmorpholine (25 mg, 0.24 mmol) was added. After the compounds dissolved 2-chloro-4,6-dimethoxytriazine (42 mg, 0.24 mmol) was added. The reaction mixture was stirred for 4h at 0 C, after which one equivalent of N- methylmorpholine (25 ml, 0.24 mmol) was added followed by 2 ( 73 mg, 0.24 mmol). Stirring was continued for 1h at 0 C, followed by stirring overnight at room temperature. CH 2 Cl 2 (50 ml) was added and the solution was washed with, respectively, 1 M aq. HCl (2 x 20 ml), brine (1x 20 ml), saturated aqueous bicarbonate solution (1 x 20 ml) and H 2 (1 x 20 ml). The organic phase was dried on Na 2 S 4, filtered and the solvent evaporated. The resulting solid was purified using column chromatography (2% MeH in CH 2 Cl 2, Si 2 ), yielding 3 as a white solid ( 44 mg, mmol, 26.8 %). m.p o C. 1 H NMR (400 MHz, CDCl 3 ) δ = 7.69 (s, 1H), (m, 4H), (m, 5H), (m, 1H), 7.08 (s, 1H), 7.04 (s, 1H), (m, 2H), 3.87 (s, 3H), 3.69 (bs, 8H), 3.60 (bs, 2H), 2.85 (t, J = 7.4 Hz, 4H), (m, 2H), 2.02 (s, 3H), 1.99 (s, 3H) ppm. 13 C NMR (101 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (d), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (d), (d), (d), (d), (d), (d), (d), (d), (s), (s), (d), (d), (q), (t), (t), (t), (t), (t), (t), (q), (q) ppm. MALDI-TF MS (MW = ) m/z = [M + ]. (6) 4-[2-(6,7-Dimethoxy-2-oxo-2H-chromen-4-yl)-acetyl]-piperazine-1-carboxylic acid tert-butyl ester (DMCpipBoc) 6,7-Dimethoxycoumarin-4-acetic acid 4 (1.0 g, 3.8 mmol) was suspended in 50 ml of CH 2 Cl 2 and stirred under nitrogen atmosphere. N,N'-Carbonyldiimidazole (CDI) (665 mg, 4.1 mmol) was added and the reaction mixture was stirred under N 2 till all C 2 had evolved and stirring continued for another 30 min. N-Boc-piperazine 5 (775 mg, 4.1 mmol) was added and the reaction mixture was stirred under N 2 at RT overnight. The reaction mixture was then extracted twice with 1M HCl (aq), once with water and twice with 5% NaHC 3 (aq). The organic phase was then dried over Na 2 S 4, filtered and the solvent evaporated. The crude mixture was purified using column chromatography (2% MeH in CH 2 Cl 2 ) giving 6 as a yellow solid ( 1.1 g, 2.4 mmol, 63.2%). m.p o C. 1 H NMR (400 MHz, CDCl 3 ) δ = 7.01 (s, 1H), 6.76 (s, 1H), 6.10 (s, 1H), 3.87 (s, 3H), 3.85 (s, 3H), 3.79 (s, 2H), (m, 2H), (m, 2H), (m, 4H), 1.40 (s, 9H) ppm. 13 C NMR (101 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (s),

127 Chapter 4 (s), (d), (s), (d), 99.9 (d), 80.3 (s), 56.3 (q), 56.2 (q), 45.9 (t), 43.4 (t), 41.7 (t), 37.9 (t), 28.2 (q) ppm. MS(EI) for C 22 H 28 N 2 7 m/z 432 [M + ], HRMS calcd for C 22 H 28 N 2 7 : , found: General deprotection method for N-BC protected amines: The Boc protected amine was stirred in a mixture of 1:1 CH 2 Cl 2 : CF 3 CH for 4h. An equal amount of water was added and the mixture was neutralized by adding solid NaHC 3, after which the aqueous layer was separated and the organic layer was washed with saturated NaHC 3 solution. Subsequently, the organic layer was dried over Na 2 S 4 and the solvent evaporated. The resulting product was used in subsequent steps without further purification. (8) 6,7-Dimethoxy-4-{2-[4-(4-{5-methyl-4-[2-(2-methyl-5-phenyl-thiophen-3-yl)- cyclopent-1-enyl]-thiophen-2-yl}-benzoyl)-piperazin-1-yl]-2-oxo-ethyl}-chromen-2-one (DMCS) This compound was synthesized using a procedure similar to the one used for MMCS 3 using 1 (200 mg 0.44 mmol) and 7 (from 6 deprotected using the general method) (162 mg, 0.48 mmol). Purification gave 8 (158 mg, 0.20 mmol, 45.5%). m.p o C. 1 H NMR (400 MHz, CDCl 3 ) δ = (m, 4H), (m, 4H), (m, 1H), 7.07 (s, 2H), 7.03 (s, 1H), 6.84 (s, 1H), 6.16 (s, 1H), 3.94 (s, 3H), 3.91 (s, 3H), 3.84 (s, 2H), 3.68 (s, 4H), 3.55 (s, 4H), 2.84 (t, J = 7.43 Hz, 4H), (m, 2H), 2.02 (s, 3H), 1.98 (s, 3H) ppm. 13 C NMR (101 MHz, CDCl 3 ) δ = (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (s), (d), (d), (d), (d), (d), (d), (d), (d), (s), (d), (d), (q), (q), (t), (t), (t), (t), (t), (t), (q), (q). ppm. MALDI-TF MS (MW = 770.3) m/z = [M + ]. 126

128 Tuning of Energy Transfer Between Two Photoswitchable Coumarin- Dithienylcyclopentene Systems Through Structural Modification 4.10 References 1 W. M. Moreau, Semiconductor Lithography, Plenum Press, New York, (a) J. E. Green, J. W. Choi, A. Boukai, Y. Bunimovich, E. Johnston-Halperin, E. Delonno, Y. Luo, B. A. Sheriff, K. Xu, Y. S. Shin, H. R. Tseng, J. F. Stoddart and J. R. Heath, Nature 2007, 445, (b) A. P. de Silva, M. R. James, B.. F. McKinney, D. A. Pears and S. M. Weir, Nat. Mater. 2006, 5, B. L. Feringa, Ed. Molecular Switches (Wiley VCH, Weinheim, 2001). 6 Y. Yokoyama, Chem. Rev. 2000, 100 (5), G. Berkovic, V. Krongauz and V. Weiss, Chem. Rev. 2000, 100, (a) H. Tian and S. J. Yang, Chem. Soc. Rev. 2004, 33, (b) M. Irie, Chem. Rev. 2000, 100, (a) S. Abad, M. Kluciar, M. A. Miranda and U. Pischel, J. rg. Chem. 2005, 70, (b) S.A. de Silva, K. C. Loo, B. Amorelli, S. L. Pathirana, M. Nyakirang'ani, M. Dharmasena, S. Demarais, B. Dorcley, P. Pullay and Y. A. Salih, J. Mater. Chem. 2005, 15, A. Peters and N. R. Branda, Chem. Commun., 2003, A. Momotake and T. Arai, J. Photoch. Photobio. C 2004, 5, (a) K. Matsuda and M. Irie, J. Photoch. Photobio. C 2004, 5, (b) F. M. Raymo, and M. Tomasulo, Chem. Soc. Rev. 2005, 34, (c) F. M. Raymo and M. Tomasulo, J. Phys. Chem. A 2005, 109, (d) F. M. Raymo and M. Tomasulo, Chem. Eur. J. 2006, 12, A. Fernandez-Acebes and J. M. Lehn, Chem. Eur. J. 1999, 5, (a) M. Irie, T. Fukaminato, T. Sasaki, N. Tamai and T. Kawai, Nature 2002, 420, (b) T. Fukaminato, T. Sasaki, T. Kawai, N. Tamai and M. Irie, J. Am. Chem. Soc. 2004, 126, (c) L. Giordano, T. M. Jovin, M. Irie and E. A. Jares-Erijman, J. Am. Chem. Soc. 2002, 124, (a) S. Wang, W. Shen, Y. L. Feng and H. Tian, Chem. Commun. 2006, (b) G. Y. Jiang, S. Wang, W. F. Yuan, L. Jiang, Y. L. Song, H. Tian and D.B. Zhu, Chem. Mater. 2006, 18, T. B. Norsten and N. R. Branda, J. Am. Chem. Soc. 2001, 123, (a) R. T. F. Jukes, V. Adamo, F. Hartl, P. Belser, L. De Cola, Inorg. Chem. 2004, 43, (b) T. A. Golovkova, D. V. Kozlov and D. C. Neckers, J. rg. Chem. 2005, 70, J. J. D. de Jong, L. N. Lucas, R. Hania, A. Pugzlys, R. M. Kellogg, B. L. Feringa, K. Duppen and J. H. van Esch, Eur. J. rg. Chem. 2003,

129 Chapter 4 19 J. Areephong, W. R. Browne, N. Katsonis and B. L. Feringa, Chem. Commun. 2006, J.H. Hurenkamp, W.R. Browne, R. Augulis, A. Pugžlys, P.H.M. van Loosdrecht, J.H. van Esch and B.L. Feringa, rg. Biomol. Chem. 2007, 5, (a) W. R. Browne, J. J. D. de Jong, T. Kudernac, M. Walko, L.N. Lucas, K. Uchida, J.H. van Esch and B.L. Feringa, Chem. Eur. J. 2005, 11, (b) W. R. Browne, J. J. D. de Jong, T. Kudernac, M. Walko, L.N. Lucas, K. Uchida, J.H. van Esch and B.L. Feringa, Chem. Eur. J. 2005, 11, (a) J. J. D. de Jong, L. N. Lucas, R. M. Kellogg, J.H. van Esch and B.L. Feringa, Science 2004, 304, (b) Ph.D. Thesis L. N. Lucas, Towards Photoresponsive Supramolecular Materials, University of Groningen, 2001, ISBN Y. Ma, W. Luo, P. J. Quinn, Z. Liu and R. C. Hider, J. Med. Chem. 2004, 47, G. Anderson and R. Paul, J. Am. Chem. Soc., 1958, 80, λ = 365 nm was chosen as irradiation wavelength since it lies at the edge of the open switch 10 absorption and UV irradiation of that wavelength gives little degradation. 26 N. J. Turro, Modern Molecular Photochemistry (University Science Books, Sausalito, 1991 ) 27 (a) A. Peters and N. R. Branda, Adv. Mater. pt. Electron. 2000, 10 (6), (b) M. Irie, T. Lifka, K. Uchida, S. Kobatake and Y. Shindo, Chem. Commun. 1999, (8), Here, quasireversibility and irreversibility refer to the chemical stability of the oxidised or reduced species. In all cases electrochemical reversibility (E p,a -E p,c < 80 mv) is observed between 0.01 and 2.0 Vs -1, where the chemical stability of the oxidised/reduced species is sufficient to observe the return wave. Where the I p,a peak is not significantly different to the I p,c peak, the term reversible is applied; similarly, quasireversiblity refers to situations in which chemical reversibility is dependent on scan rate (i.e. reversibility is observed only at higher scan rates). 29 The diphenyldithienylcyclopentene switches are very sensitive to irradiation, it is close to impossible to exclude all light while performing these experiments. 30 A UV/Vis spectrometer was not available at the time of the measurements, hence the lack of full reversibility is due to the fact that it was not possible to confirm the full re-opening/closing of the switch, 31 (a) E. A. A. Wallén, J. A. M. Christiaans, E. M. Jarho, M. M. Forsberg, J. I. Venäläinen, P. T. Mannisto and J. Gynther, J. Med. Chem. 2003, 46, (b) L. A. Carpino, E. M. E. Mansour, C. H. Cheng, J. R. Williams, R. Macdonald, J. Knapczyk, M. Carman and A. Lopusinski, J. rg. Chem. 1983, 48,

130 Chapter 5 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches In the previous chapters several donor acceptor systems of increasing complexity were prepared and characterised spectroscopically. The design approach, which was taken, emphasized the role of bridging units, which in those cases were insulating piperazine moieties. The aim in all these systems was to connect an energy donor unit (i.e. a coumarin) with an energy acceptor unit (i.e. a perylene bisimide or a photoswitchable dithienylcyclopentene). In this chapter the synthesis and spectroscopic characterization of two donor-acceptor systems is described; a symmetric perylene-switch-perylene (PerSPer) and an asymmetric dimethoxycoumarin-switch-perylene (DMCSPer). In the symmetric system the PerSPer shows efficient photoswitchable quenching of emission intensity of the perylene bisimide unit. With DMCSPer, for the coumarin unit both the anticipated energy transfer to the perylene bisimide and the expected quenching by the closed dithienylcyclopentene were observed. The closed dithienylcyclopentene quenches the excited state of both chromophores. This was surprising for the perylene as the overlap between donor and acceptor is poor and the lowest absorption band of the perylene and the closed switch are essentially at the same energy. This makes the efficient quenching observed difficult to rationalize using the Förster model for energy transfer, a more general excited state model was required, which takes other factors such as excited state decay rates into account. 129

131 Chapter Introduction In Chapter 2 the combination of four coumarin donors units with a perylene imide acceptor unit resulted in a system in which efficient and fast energy transfer took place. This system proved to be photochemically very stable. However, it was not possible to control energy transfer within this system after synthesis, i.e. the rate as well as efficiency of the energy transfer was preset. A step further was taken in Chapter 3 through the use of a dithienylcyclopentene photochromic switch as a modulator of energy transfer. However, the system used had several disadvantages. The quenching of the coumarin donor excited state was inefficient, with a maximum of 30% of the total emission being quenched upon ring-closure. The stability of the system was an issue with the triad showing 10% degradation per switching cycle, even at a temperature of 220 K, making it unsuitable for use as a component in molecular photonic devices. In addressing these issues, first the structures of the dithienylcyclopentene photochromic switches and the coumarins were varied to develop a system which showed effective modulation of the energy transfer and high stability during the switching steps (see Chapter 4). Diphenyl dithienylcyclopentene switches were found to show substantially higher stability then the unsubstituted dithienylcyclopentenes. A mono carboxylic acid diphenyl dithienylcyclopentene had already been developed in our group 1 and this switch was used, together with the coumarin employed previously, to determine the ability of the closed form of the dithienylcyclopentene unit to quench the coumarin excited state. However, as the results obtained were unexpected (see Chapter 4) another donor unit, a dimethoxy coumarin, was employed to achieve fluorescence switching. This dyad showed high stability together with efficient switching of the emission intensity (up to 98% quenching in the closed form), which made this dyad a good candidate to use together with a compatible chromophore to build an efficient energy transfer modulating triad. In this chapter the experience gained in previous systems is drawn on to design a system which is an efficient energy transfer modulator, preferably even an energy transfer n / ff switch. The spectroscopic properties of the perylene bisimide used in the earlier triad system showed that it had good potential as an energy acceptor. Especially when combined with the dimethoxy coumarin donor unit, since the overlap of the coumarin emission with the perylene bisimide absorption is optimal for through space energy transfer. The diphenyl 130

132 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches dithienylcyclopentene will be used as a photoswitchable quencher of the coumarin excited state. 5.2 Results & Discussion To combine the separate components a proven method, i.e. that used in Chapter 3, was applied. The dicarboxylic acid diphenyl dithienylcyclopentene was available from known methods. 2 This molecule was used in a one step reaction with the amine-functionalized donor and acceptor to prepare the Donor-Switch-Acceptor triad molecule. As will be shown in the following sections, the triad system shows high stability and efficient switching and quenching combined within one molecule Synthesis Amide coupling chemistry with acid functionalized dithienylcyclopentene switches has proven to be a versatile strategy in building multicomponent systems through the combination of providing a robust approach to covalent connectivity and precluding through-bond intercomponent interactions. 3,4 In the synthesis of the donor-acceptor systems described in Chapters 2, 3 and 4, amide coupling chemistry was employed successfully to connect donor fluorophores to carboxylic acid functionalized dithienylcyclopentenes and perylenes. In this chapter, the amide coupling approach is employed again to connect the coumarin donor and perylene acceptor molecules to diacid functionalized diphenyldithienylcyclopentene 5 (Figure 5.1), which is known to show greater stability and better photostationary states compared to dithienyl cyclopentene switches without the phenyl groups present. 2 The syntheses of the piperazine functionalized 6,7-methoxy-coumarin-4-acetic acids 1 and 2 are described in Chapter 3. The syntheses of the mono piperazine functionalized perylene 3 and 4 are described in Chapter 2. The carboxylic acid functionalized diphenyl switch 5 was synthesized as reported previously. 2 Compounds 6, 7 and 8 were synthesized in a one step procedure with equimolar amounts of the chromophores followed by isolation of the target compounds by column chromatography. ne equivalent of the N-(piperid-4-yl)-N -(n-butyl) perylene bisimide 4 (1 equiv.) was coupled, together with the coumarin piperazine 2 (1 equiv.) to the diacid diphenyldithienylcyclopentene photochromic switch 5 (1 equiv.) using 2-chloro-4,6- dimethoxy-1,3,5-triazine (CDMT) and N-methyl-morpholine (NMM) in CH 2 Cl 2 to give, in addition to the homo-coupling products (6 (not isolated) and 7 (4.8%)), the target Coumarin-Switch-Perylene triad (DMCSPer) 8 in 7.4% yield (non-optimized, Figure 5.1). 131

133 Chapter 5 All compounds were purified by column chromatography and characterized with 1 H and 13 C NMR spectroscopy and (MALDI-TF) mass spectrometry (see experimental section for details). 1 R 1 =Boc 2 R 1 =H N NR 1 R 2 N N N H S 5 S A,B H 3 R 2 =Boc 4 R 2 =H N N S S N N 6 N N N N N S S 7 PerSPer N N N S S 8 DMCSPer N N N Figure 5.1 Synthesis of donor-acceptor molecules 6, 7 and 8. A) CDMT, NMM, CH 2 Cl 2, 0 o C; B) NMM, 1 eq 2, 1 eq 4, 24 h. 132

134 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches Figure 5.2 The model compounds for the perylene bisimide 9, the dithienyl cyclopentene switch 10 and the dimethoxycoumarin switch (DMCS) dyad Redox properties The redox properties of the open and closed form of PerSPer 7 and DMCSPer 8 were investigated using both cyclic voltammetry (CV) and differential pulse voltammetry (DPV). 5 Triads 7 and 8 demonstrate rich electrochemistry, with both molecules showing multiple oxidative and reductive processes (Table 5.1). Table 5.1 Redox properties of the open and closed form of PerSPer 7 and DMCSPer 8 and model compounds. Compound E 1/2, V vs SCE (E p,a where irr or qr) a,b DMCpipboc (qr) Perylene Bis butyl , -0.77, 1.25 DMCS 11 pen 1.18(irr), 1.50(qr) DMCS 11 Closed -1.72(irr), 0.43, 0.78, 1.50(qr) PerSPer 7 pen -0.90, -0.77, 1.23, 1.57(irr) PerSPer 7 Closed -1.67, -0.91, -0.77, 0.45, 0.79, 1.22, 1.56(irr) DMCSPer 8 pen -1.84(irr), -0.87, -0.72, 1.23, 1.59(irr) DMCSPer 8 Closed -1.59, -0.87, -0.73, 0.45, 0.79, 1.23, 1.59(irr) a In CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE. b irr=irreversible, qr = quasi reversible

135 Chapter 5 Current (A) -10.0µ PerSPer 7 For the open form of PerSPer 7 three reversible redox processes can be observed (-0.90, and 1.23 V vs SCE), which have identical shape and potential as observed for 9 (Figure 5.3, left). However, the oxidative process at 1.23 V [vs SCE] has a higher intensity then expected for a one electron process. This is due to the fact that an irreversible process is taking place at slightly higher potential, i.e. that of the open form of the switch (Po 1+ (Po 2+ ) Pc 2+ ). This process is the electrochemical ring closing of the dithienylcyclopentene and is known for diphenyl dithienylcyclopentene switches (see chapter 4 and ref 7). n the return cycle the two characteristic reductions for the closed form dication are observed also (Pc 2+ Pc + Pc). The cyclic voltammogram of the closed form of 7 shows the redox processes already observed for the open form (-0.90, -0.77, 1.23 V vs SCE) and three additional reversible peaks (Figure 5.3, right), which are assigned to the closed form of the dithienylcyclopentene (-1.67, 0.45, 0.79 V vs SCE). -5.0µ µ Pc 2- Pc - Pc Po Pc Po 1+ Pc + (Po 3+ ) Pc 2+ Pc 2+ Pc 3+ Pc 3+ Current (A) -10.0µ -5.0µ µ Pc 3- Pc 2- Pc 2- Pc - Pc Pc Pc Pc + Pc + Pc 2+ Pc 2+ Pc 2+ Pc 3+ Pc 3+ Pc 2- Pc - Pc 10.0µ Potential (V vs SCE) Pc 3- Pc 2- Pc 2- Pc - Pc 10.0µ Potential (V vs SCE) Figure 5.3 Cyclic voltammetry of the open (left) and closed (right) form of PerSPer 7 (PerSPer open (Po), PerSPer closed (Pc)) in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE at 0.1 V s -1. The redox processes are clearer when observed using differential pulse voltammetry (DPV). The open form of 7 now shows four processes, of which three are from the perylene bisimide part of the molecule (Figure 5.4, left). The irreversible process at 1.57 V [vs SCE] now lies just within the solvent window. For the closed form of 7 the three extra redox processes due to the closed form of the dithienylcyclopentene are also observed (Figure 5.4, right). 134

136 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches -16.0µ -8.0µ -12.0µ -8.0µ -4.0µ Current (A) -4.0µ µ 8.0µ Current (A) µ 12.0µ 16.0µ Potential (V vs SCE) 8.0µ Potential (V vs SCE) Figure 5.4 Differential pulse voltammetry of the open (left) and closed (right) form of PerSPer 7 in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE DMCSPer 8 As for the open form of PerSPer 7 the open form of DMCSPer 8 shows three reversible redox processes (-0.90, and 1.23 V [vs SCE]) (Figure 5.5, left), which have identical shape and potential as observed for 9. For 8 the oxidative process at 1.23 V [vs SCE] also obscures the irreversible ring closing process of the dithienylcyclopentene part, once more in the return wave the two peaks for the closed form are also visible (Dc 2+ Dc + Dc). At 1.48 V [vs SCE] the irreversible wave is even more pronounced for DMCSPer 8, this is due to a quasi reversible redox process of the coumarin unit at slightly lower potential than the process visible for PerSPer 7. As with 7 the cyclic voltammogram of the closed form of 8 shows the peaks already observed for the open form (-0.90, -0.77, 1.23 V [vs SCE]) and three additional reversible peaks (Figure 5.5, right), which are known to be of the closed form of the dithienylcyclopentene (-1.59, 0.45, 0.79 V [vs SCE]). For DPV, the open form of 8 now shows four processes, of which three are from the perylene bisimide part of the molecule (Figure 5.6). Also one irreversible process is observed at V [vs SCE] which is probably due to the coumarin unit, since it was not visible for 7. As with 7 for the closed form of DMCSPer 8 the three extra redox processes due to the closed form of the dithienylcyclopentene are also observed. It is interesting to note here the difference in intensity between the peaks of the closed dithienylcyclopentene (i.e. at 0.45 and 0.79 V [vs SCE]) and the peak for the perylene 135

137 Chapter 5 bisimide at 1.23 V [vs SCE] for, respectively, PerSPer 7 (Figure 5.4) and DMCSPer 8 (Figure 5.6). For 7 the area of the peak at 1.23 V [vs SCE] is nearly double that of the peak at either 0.45 or 0.79 V [vs SCE], whereas for 8 the area is roughly the same. This is due to the process in 7 involving two oxidation steps (i.e. both perylene bisimides are oxidized simultaneously) and in 8 a single oxidation takes place (i.e. one electron, which is equivalent to the processes at 0.45 and 0.79 V [vs SCE]). -8.0µ -8.0µ Current (A) -6.0µ -4.0µ -2.0µ µ 4.0µ 6.0µ Dc 2- Dc 2- Dc - Dc - Dc Dc Do Do 1+ (Do 3+ ) Dc 3+ Dc Dc + Dc 2+ Dc 3+ Current (A) -6.0µ -4.0µ -2.0µ µ 4.0µ 6.0µ Dc 2- Dc 3- Dc 2- Dc 3- Dc 2- Dc Dc + Dc - Dc Dc 2+ Dc 3+ Dc Dc + Dc 2+ Dc 2- Dc - Dc 3+ Dc Potential (V vs SCE) Potential (V vs SCE) Figure 5.5 Cyclic voltammetry of the open (left) and closed (right) form of DMCSPer 8 (DMCSPer open (Do), DMCSPer closed (Dc)) in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE at 0.1 and 0.5 V s -1, respectively µ -6.0µ Current (A) -8.0µ -4.0µ µ 8.0µ Potential (V vs SCE) Current (A) -4.0µ -2.0µ µ 4.0µ 6.0µ Potential (V vs SCE) Figure 5.6 Differential pulse voltammetry of the open (left) and closed (right) form of DMCSPer 8 in CH 2 Cl 2 / 0.1 M TBAPF 6 vs SCE. 136

138 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches 5.4 UV/Vis absorption, pen and PSS state UV/Vis spectra of PerSPer 7 and DMCSPer 8 in the open and the photostationary state (PSS, λ irr = 365 nm) are shown in Figure 5.7. The spectrum of the open form of PerSPer 7 clearly shows the three absorption maxima in the visible region of the perylene bisimide spectrum and the maxima (i.e. λ max = 452, 541 and 579 nm) are essentially the same as those of the perylene bisimide model compound 9 (Table 5.2). Table 5.2 Absorption and emission spectra of fluorescent switches in the open and PSS 365nm state and of the model compounds. Compound Absorption a λ max / nm (10 3 ε / cm -1 M -1 ) Emission a λ max / nm DMCpipboc 1 293(4.9) 346(11.4) Perylene Bis butyl 9 266(40.6) 286(49.6) 451(16.7) 539(26.7) 577(43.1) 608 PrPhSPh (25.5) PrPhSPh 10 PSS 365nm b 287(22.9) - 372(9.3) 547(12.4) DMCS pen (29.3) 321(24.6) DMCS 11 PSS 365nm b 285(25.9) - 353(17.0) 541(14.9) PerSPer 7 287(129.1) - 452(33.8) 541(56.2) 579(90.4) 609 PerSPer 7 PSS 365 nm b 287(126.9) 370(19.5) 453(37.3) 542(72.2) 579(107.3) 611 DMCSPer 8 287(75.2) 324(37.7) c 452(15.0) 540(24.8) 579(40.1) 416, 608 DMCSPer 8 PSS 365nm b 287(73.4) 349(22.2) 456(18.2) 542(40.2) 579(55.2) 420, 609 a Recorded in CH 2 Cl 2 at RT. b After irradiation with λ = 365 nm light at RT till PSS was reached. c Shoulder. This indicates that there is little or no through bond communication between the perylene bisimide unit and the dithienylcyclopentene unit. In the PSS 365nm state, obtained by irradiating with λ = 365 nm light, an increase in absorption in the red and a decrease of absorption in the blue can be observed. The difference spectrum between the open and PSS state of 7 shows that the change is identical to that observed for the model switch compound 10 (Figure 5.8, Left). Upon irradiation with λ > 400 nm light the absorption spectrum of 7 recovers to the original shape (i.e. to the open form). 137

139 Chapter k 80.0k 120.0k 100.0k 60.0k ε (cm -1 M -1 ) 80.0k 60.0k 40.0k ε (cm -1 M -1 ) 40.0k 20.0k 20.0k Wavelength (nm) Wavelength (nm) Figure 5.7 Left: UV/Vis spectra of PerSPer 7 in the open state ( ),PSS 365nm ( ) and reopened state (λ irr > 400 nm, ), spectra recorded in CH 2 Cl 2 at RT. Right: UV/Vis spectra of DMCSPer 8 in the open state ( ), PSS 365nm ( ) and reopened state (λ irr > 400 nm, ), spectra recorded in CH 2 Cl 2 at RT. The absorption spectrum of the open form of DMCSPer 8 is similar to that of PerSPer 7, however, it shows several notable differences. The molar absorptivity of the maxima in the open form of 8 are half that of the maxima of PerSPer 7, as expected given the fact that this triad contains only one perylene bisimide. Also the shoulder at ca. λ = 350 nm is more pronounced for 8 due to the increased absorption of the switching unit in this region (i.e. relative to the perylene bisimide unit) and due to the absorption of the coumarin unit. 20.0k 20.0k 15.0k 15.0k 10.0k 10.0k ε (cm -1 M -1 ) 5.0k k -10.0k Wavelength (nm) ε (cm -1 M -1 ) 5.0k k -10.0k Wavelength (nm) -15.0k -15.0k -20.0k -20.0k Figure 5.8 Left: UV/Vis difference spectrum of PerSPer 7 between the open state and the PSS 365 nm state ( ), spectrum recorded in CH 2 Cl 2 at RT. Right: UV/Vis difference spectrum of DMCSPer 8 between the open state and the PSS 365nm state ( ), spectrum recorded in CH 2 Cl 2 at RT. 138

140 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches The PSS 365nm state of DMCSPer 8 shows a similar increase (in the red) and decrease (in the blue) as observed for PerSPer 7, however, the change appears larger than observed for 7. This is not surprising as with 7 the changes are obscured by two perylene units and with 8 only by one perylene unit and therefore the absolute change is the same, as is apparent from the difference spectra for 7 and 8 (Figure 5.8). For 8 the change in the absorption spectrum is reversible upon irradiation with λ > 400 nm light also. 5.5 Fluorescence spectroscopy of the pen State and PSS. The emission spectrum (λ irr = 450 nm) of the open state of PerSPer 7 from 550 to 750 nm has a shape identical to the emission of the perylene bis-n-butyl imide 9, indicating that the emission is based on the perylene bisimide unit (Figure 5.9). Upon photo-conversion to the PSS 365 nm state the emission intensity of 7 is reduced by 83%. This is due to quenching by the closed form of the dithienylcyclopentene unit, the only photochemically responsive component in the system. This change is reversible, upon irradiating with λ > 400 nm light the emission spectrum of PerSper 7 recovers almost completely to the original intensity. 3.0M 1.6M CPS 2.5M 2.0M 1.5M 1.0M 500.0k Wavelength (nm) CPS 1.4M 1.2M 1.0M 800.0k 600.0k 400.0k 200.0k Wavelength (nm) Figure 5.9 Left: Emission spectra (λ ex = 450 nm) of PerSPer 7 in the open state ( ),PSS 365 nm ( ) and reopened state (λ irr > 400 nm, ). The spectra were recorded in CH 2 Cl 2 at RT. Right: Emission spectra (λ ex = 450 nm) of DMCSPer 8 in the open state ( ), PSS 365 nm ( ) and reopened state (λ irr > 400 nm, ). The spectra were recorded in CH 2 Cl 2 at RT. The emission behaviour (λ irr = 450 nm) of DMCSPer 8 is similar to that of 7. In the open form the shape of the emission (λ irr = 450 nm) is identical to that of perylene bisimide model 9. When the triad 8 is irradiated to the PSS 365nm the emission intensity is reduced by 80% with emission recovering to its original intensity upon irradiation with λ > 400 nm light, as was observed for 7. The only difference observed is that the emission of 8 is about 139

141 Chapter 5 half the intensity of that of 7, however this is expected given the presence of two perylene bisimides in 7 and only one in 8. The effective quenching of the perylene bisimide excited state in the closed form of triad 7 and 8 is somewhat unexpected. Förster type energy transfer requires overlap of the fluorescence spectrum of the donor (perylene bisimide unit) with the absorption spectrum of the acceptor (closed dithienylcyclopentene unit) and that the absorption of the acceptor is at longer wavelength (i.e. lower energy) than that of the emission of the donor. This does not seem to be the case in the present system, however, upon closer inspection these requirements are met to a limited degree. From Figure 5.10 it can be seen that the absorption of the closed dithienylcyclopentene (represented by the difference spectrum) extends beyond that of the perylene bisimide towards the red of the spectrum and overlaps with the fluorescence spectrum of the perylene bisimide unit indicating that energetically energy transfer is possible. However, the donor acceptor overlap is poor, especially compared to systems described in previous chapters M Abs M 2.0M 1.5M 1.0M CPS k Wavelength (nm) 0.0 Figure 5.10 n the left side absorption spectra of PerSPer 7 open ( ), PSS 365 nm ( ) and the difference spectrum ( ) are shown. n the right side the fluorescence spectrum of the open form of 7 ( ) is shown. The spectra were recorded in CH 2 Cl 2 at RT. 140

142 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches Photochromic Switching and Stability. Both molecules 7 and 8 show reasonable stability and switching can be performed over several cycles, however, for PerSPer 7 significant degradation is observed over several cycles (Figure 5.11). With increasing numbers of switching cycles the spectrum of the open form of 7 remains unchanged, except for the region between 300 to 350 nm, where recovery is incomplete. For the closed form of 7 the spectrum shows a clear decrease in absorption in the region between 350 to 400 nm and from 450 to 600 nm during consecutive cycles (See also Figure 5.14). The region in which the decrease is observed corresponds to that of the absorption of the closed dithienylcyclopentene switch Abs Abs Wavelength (nm) Wavelength (nm) Figure 5.11 Left: UV/Vis absorption spectra of PerSPer 7 recorded over several switching cycles. Right: UV/Vis absorption spectra of DMCSPer 8 recorded over several switching cycles. (For both: the open state is converted to the closed state (PSS 365 nm ) by irradiation with λ = 365 nm light, the closed isomer is returned to the open state by irradiation with λ > 400 nm light. The spectra were recorded in CH 2 Cl 2 at RT.) Since these absorptions decrease (i.e. from 350 to 400 nm and from 450 to 600 nm), despite constant irradiation times, it is apparent that the decrease is due to reduced formation of the ring closed state. This indicates that the dithienylcyclopentene undergoes an irreversible photochemical reaction. This is apparent in the spectrum of the open form of 7, in which the shape and intensity of the perylene absorption in the visible remain unchanged. For DMCSPer 8 the shape of the spectrum in the open and closed form is stable over several cycles and only minor changes are observed, indicating good fatigue resistance (Figure 5.11, right). Why photochemical degradation is more apparent with PerSPer 7 than DMCSPer 8 is not immediately clear. 141

143 Chapter 5 For the emission spectra the change in emission intensity, as with UV/Vis, is reversible over several switching cycles (Figure 5.12). However, as with the absorption spectra, the reversibility for PerSPer 7 is lower than for DMCSPer 8. For 7 the quenching is less pronounced with increasing number of switching cyles, indicating that the quenching unit, which is most likely the closed dithienylcyclopentene unit, is undergoing an irreversible photoreaction, since the maximum intensity of the open form is unaffected. By contrast, only minor changes are observed for DMCSPer 8 (Figure 5.12, right). Closer inspection of the intensity of the absorption spectra of PerSPer 7 at λ = 550 nm during several switching cycles (Figure 5.13, left) shows that the absorption intensity in the open form remains constant. The absorption in the PSS, however, shows a steady decrease per cycle. For DMCSPer 8 the intensity of the open state and the PSS remains nearly constant. 3.0M 1.4M CPS 2.5M 2.0M 1.5M 1.0M 500.0k Wavelength (nm) CPS 1.2M 1.0M 800.0k 600.0k 400.0k 200.0k Wavelength (nm) Figure 5.12 Left: Emission spectra (λ ex = 450 nm) of PerSPer 7 recorded over several switching cycles. Right: Emission spectra of DMCSPer 8 recorded over several switching cycles. (For both: the open state is converted to the closed state (PSS 365 nm ) by irradiation with λ = 365 nm light, the closed isomer is returned to the open state by irradiation with λ > 400 nm light. The spectra were recorded in CH 2 Cl 2 at RT.) For the emission spectra of PerSPer 7 during the cycling a clear increase of the emission intensity of the PSS is observed (Figure 5.13, right), the open form remains at nearly the same intensity, showing only a minor decrease in intensity. As with the absorption spectra the emission intensity in the open state and the PSS of DMCSPer 8 stays at nearly the same level, showing efficient and reversible modulation of the fluorescence. 142

144 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches 80k 3.0M 70k 2.5M 60k 2.0M ε (cm -1 M -1 ) 50k 40k CPS 1.5M 1.0M 30k 500.0k 20k open0 closed1 open1 closed2 open2 closed3 open3 closed4 open4 Cycle # 0.0 open0 closed1 open1 closed2 open2 closed3 open3 closed4 open4 Cycle # Figure 5.13 Left: UV/Vis absorption at λ = 550 nm during four switching cycles of 7 ( ) and 8 ( ). The spectra were recorded in CH 2 Cl 2 at RT. Right: Emission intensity during four switching cycles for 7 ( ) and 8 ( ), measured at λ = 610 nm (λ ex = 450 nm). The spectra were recorded in CH 2 Cl 2 at RT. For DMCSPer 8 the overall changes are minor and little degradation is observed. For PerSPer 7 some further remarks should be made regarding its degradation. The difference spectra of the open and PSS state relative to the spectrum of the open form before switching give some insight into the source of the photochemical instability which is observed upon irradiation of PerSPer 7 (Figure 5.14) Wavelength (nm) Abs Abs Wavelength (nm) Figure 5.14 Left: Difference spectra of PerSPer 7 of the open form during four switching cycles of 7 relative to the open form at the start of the cycling (i.e. before any irradiation) Right: Difference spectra of PerSPer 7 of the PSS during four switching cycles of 7 relative to the open form at the start of the cycling. The spectra were recorded in CH 2 Cl 2 at RT. 143

145 Chapter 5 Both the absorption spectrum of the open state and PSS show considerable changes in the region where the dithienylcyclopentene unit absorbs, i.e. between 250 and 350 nm for the open form and 350 and 700 nm for the PSS state. Another important observation is the increase of absorption at 250 and 400 nm, which is observed for both the open and PSS difference spectra. This indicates the irreversible formation of another species, indicating a permanent change to the dithienylcyclopentene unit, possibly the rearrangement which has been reported elsewhere (see Chapter 4). 8 The redox chemistry of 7 and 8 in the open and closed states indicate that electron transfer photochemistry is unlikely to play a significant role in the observed degradation (vide supra). 5.7 Energy transfer For DMCSPer 8 the coumarin unit of the asymmetric triad makes energy transfer between the coumarin donor and both perylene bisimide unit and closed dithienylcyclopentene possible (Figure 5.15, right). The emission maximum of the coumarin model DMCpipboc 1 is at λ max = 417 nm. For DMCSPer 8 an emission is observed with a maximum at approximately the same position (λ max = 420 nm). 700k 2.0M 1.5M 1 600k 500k 400k CPS 1.0M 500.0k 8 open 8 PSS CPS 300k 200k 100k Wavelength (nm) Wavelength (nm) Figure 5.15 Left: Emission spectra (λ ex = 450 nm) of DMCSPer 8 in the open state ( ), PSS 365 nm ( ), reopened state (λ irr > 400 nm, ) and of DMCpipboc 1 (, compensated for absorption). Right: Emission spectra of DMCSPer 8 recorded over several switching cycles. The spectra were recorded in CH 2 Cl 2 at RT. However, between the model 1 and DMCSPer 8 the intensity of the emission at this wavelength differs considerably. In equimolar solutions the intensity of the emission of 8 at λ = 420 nm is 60 times less than that of 1. From the dimethoxy coumarin dithienylcyclopentene dyad 11 it was found (see Chapter 4) that a combination of just the switch and the coumarin shows normal (i.e. the same intensity as DMCpipboc 1) intensity 144

146 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches of fluorescence when excited at λ ex = 380 nm. 9 This indicates that in the open form of 8 the energy absorbed by the coumarin unit is being transferred elsewhere. Since the open form of the dithienylcyclopentene has no influence on the coumarin emission in dyad 11, the only remaining candidate is the perylene bisimide chromophore, which, due to the overlap of the coumarin fluorescence with the perylene bisimide absorption, is an obvious acceptor for Förster energy transfer (see Chapter 2 and 3). At the PSS 365 nm of 8 the fluorescence intensity at both λ em = 420 and 605 nm decreases dramatically (with λ ex = 380 nm 9 to 31% and 10% of the open form, respectively). This decrease in intensity can be reversed by ring-opening of the dithienylcyclopentene unit using λ > 400 nm light. This modulation can be performed over several cycles, with minimal degradation (Figure 5.15, right). ver the switching cycles the emission intensity of the coumarin (λ em = 420 nm) is modulated reversibly also (a 66% decrease in intensity upon switching from the open state to the PSS is observed). This indicates that the emission observed is from the coumarin unit attached to the dithienylcyclopentene and not from unattached coumarin in solution. Such modulation of the coumarin emission was observed for DMCS 11 (Chapter 4) also k 600.0k CPS 400.0k 200.0k Wavelength (nm) Figure 5.16 Excitation spectra (λ em = 605 nm) of DMCSPer 8 in the open state ( ) and PSS 365 nm ( ) and excitation spectra of DMCpipboc 1 (λ em = 420 nm, ) and 9 (λ em = 605 nm, ). The spectra were recorded in CH 2 Cl 2 at RT. The excitation spectrum of 8 (λ em = 605 nm) shows features of both DMCpipboc 1 and bis-n-butyl perylene bisimide model 9 (Figure 5.16). When comparing the excitation spectrum of the perylene bisimide model 9 with the excitation spectrum of DMCSPer 8 there is an additional band at ca. λ = 420 nm. This band coincides with the excitation spectrum of the coumarin, which shows that the energy that is absorbed by the coumarin 145

147 Chapter 5 unit is transferred to the perylene bisimide unit and then emitted, thereby demonstrating energy transfer from the coumarin unit to the perylene bisimide in the open form. At the PSS 365 nm state of 8 the overall intensity of the entire excitation spectrum decreases by the same factor (i.e. overall ~ 0.25) with only minor differences (i.e. bands corresponding to perylene absorption appear to have a greater decrease). This shows that at the PSS overall the energy absorbed and subsequently emitted by the perylene bisimide, has decreased and the emission is being quenched by a third component, in this case most likely the closed switch Electron Transfer Quenching. The relative positions of the HM and LUM levels of the individual components in PerSPer 7 and DMCSPer 8 (i.e. coumarin, switch and perylene bisimide, Figure 5.17) can give a good indication whether electron transfer is possible within these triads. The first oxidation gives the energy (V vs SCE) of the HM level. The energy of the λ max of absorption and the first reduction give, respectively, the upper and lower limit of the LUM level relative to the HM Potential (V vs SCE) Coumarin pen Switch Perylene Potential (V vs SCE) Coumarin Closed Switch Perylene Figure 5.17 Depiction of the relative positions of the HM (bottom) and LUM (top) levels of the separate components in PerSPer 7 and DMCSPer 8 in the open (left) and closed (right) state. Values obtained from redox and electronic data. For PerSPer 7 the HM and LUM levels of the perylene bisimide and the switch units are important. In the open state there is no driving force for reductive electron transfer since the HM of both units is at nearly the same level and therefore electron transfer from the switch to the perylene bisimide is not favoured. The LUM of the switch is higher in energy than the LUM level of the perylene bisimide, and hence oxidative electron transfer would be up hill energetically. For the closed state of the switch the situation is different; reductive electron transfer from the excited perylene to the closed switch unit is not 146

148 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches possible, since the LUM level of the switch is not lowered sufficiently. However, the HM level of the closed switching unit is now positioned between the HM and LUM level of the perylene bisimide, this makes reductive electron transfer a possibility (i.e. with the perylene in the excited state an electron is transferred from the switch HM level to the perylene bisimide to fill the HM orbital, after which back electron transfer can take place, where the electron from the perylene LUM level is transferred to the switch SM level and the triad has returned to the ground state. For DMCSPer 8 an additional unit is added to the molecule which can give rise to electron transfer. In the open form reductive electron transfer from the coumarin excited state to the switch is not possible due to the higher energy of the LUM level of the switch. xidative electron transfer from the switch to the excited coumarin is possible; however, due to the small driving force because of the small energy difference it is unlikely. In the closed form reductive electron transfer between the switch and the coumarin is possible, however, the small energy difference makes this unlikely. xidative electron transfer from the closed switch to the excited coumarin seems to be a possibility due to the favorable positioning of the closed switch HM level. Importantly, the HM and LUM levels of the coumarin and the perylene bisimide have the classic positioning required for energy transfer. For the switch and the perylene bisimide unit the same situation holds as for 7, which means that reductive electron transfer quenching is possible energetically. 5.9 Regarding the Energy Transfer and Quenching Mechanism Combining the experimental data obtained from DMCS 11 and PerSPer 7, which shows that the excited state of both chromophores (i.e. coumarin and perylene bisimide) can be quenched by the closed form of the dithienylcyclopentene switch. Therefore, it is likely that the mechanism of energy transfer and quenching which is observed for DMCSPer 8 occurs as depicted in Figure In the open form of 8 the bulk of the excited state energy of the coumarin is transferred (k 2 ) to the perylene bisimide and emitted (k 3 ). It cannot be transferred to the open form of the dithienylcyclopentene, since the excited state of the switch in the open form is too high in energy. In the closed form of DMCSPer 8 the dithienylcyclopentene provides a lower energy sink, to which the energy of both the coumarin (k 4, direct, and k 2, via the perylene bisimide) and the energy of the perylene bisimide (k 5 ) can be transferred. The absorption spectra indicate that the excited state energy levels of the perylene bisimide and dithienylcyclopentene units are very close, which makes the reverse process of k 5 a 147

149 Chapter 5 possibility, however since the dithienylcyclopentene has a fast radiationless decay of the excited state (k 6 >> reverse k 5 ), 10 this reverse process is not significant and consequently a large overall decrease in the perylene bisimide emission is observed. Figure 5.18 Schematic representation of the energy levels and energy transfer within the open (left) and closed (right) form of DMCSPer Conclusions The goal of this chapter was to design a system that is an efficient energy transfer modulator and preferably an energy transfer n / ff switch. PerSPer 7 shows efficient photoswitchable quenching of emission intensity of the perylene bisimide unit analogous to the coumarin-dithienylcyclopentene described in Chapter 4. This was surprising because in case of the coumarin-switch systems energy transfer from the coumarin to the closed dithienylcyclopentene is expected to be efficient due to the very good overlap of the donor emission and acceptor absorption. In contrast, for PerSPer 7 the overlap between donor emission and acceptor absorption is poor and the lowest absorption band of the perylene and the closed switch are essentially at the same energy, which makes the observed efficient quenching difficult to rationalize. A similar situation was observed for the monomethoxy coumarin-switch dyad in the open state (see Chapter 4). As a result of these observations it is apparent that a simple Förster approach is insufficient to describe these systems as a whole and a more general excited state model such as that shown in Figure 5.18 is needed; a model that takes the rates of other photophysical processes such as excited state decay rates into account. 148

150 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches Nevertheless many of the design rule lessons learned in earlier chapters with regard to the choice of molecular components were incorporated successfully. Especially the use of the phenyl-substituted dithienylcyclopentene, which show much better photostationary states than that achievable with the earlier switchable components. With DMCSPer 8, for the coumarin unit both the anticipated energy transfer to the perylene bisimide and the expected quenching by the closed dithienylcyclopentene were observed. The closed dithienylcyclopentene quenches the excited state of both chromophores. In the open state the emission of the coumarin unit is quenched by energy transfer to the perylene; similarly in the closed state the coumarin is quenched by the dithienylcyclopentene unit. In addition, as for PerSPer 7, the closed state of the dithienylcyclopentene unit quenches the perylene bisimide excited state. Remarkably, despite the rich photophysics of this system, it proved to be photochemically more robust than the simpler symmetric system 7. The relatively straightforward synthetic accessibility to photophysically complex systems is demonstrated in compounds such as 8. The strategy to employ both structurally and photophysically well-defined functional building blocks (i.e. chromophores, switching units) connected by insulating bridging units opens fascinating opportunities to explore fundamental photophysical processes. These systems will form the subject of femto- and picosecond spectroscopic investigations in the near future, to elucidate their complex photophysical behaviour Experimental section For all spectroscopic measurements Uvasol-grade solvents (Merck) were employed. All reagents employed in synthetic procedures were of reagent grade or better, and used as received unless stated otherwise. Compounds 1, 2, 3, , 10 and 11 were prepared according to literature (See Chapters 3 and 4 ). Jetsuda Areephong is acknowledged gratefully for providing compound 5, which was prepared according to literature. 2 1 H NMR spectra were recorded at 200, 300, or 400 MHz; 13 C NMR spectra at 50.3, 75.4 or MHz. All spectra were recorded at ambient temperature, with the residual proton signals of the solvent as an internal reference. Chemical shifts are reported in ppm relative to TMS. MALDI-TF spectra were recorded on an Applied Biosystems Voyager-DE Pro. UV/Vis absorption spectra (accuracy ±2 nm) were recorded on a Hewlett-Packard UV/Vis 8453 spectrometer. The fluorescence measurements were performed on a SPF-500C spectrofluorometer manufactured by SLM Aminco, and a Jobin-Yvon Fluorolog 3-22, the sharp features between λ = 450 and 500 nm in the excitation spectra are instrumental artefacts, the excitation and emission spectra are uncorrected for variations in lamp 149

151 Chapter 5 intensity and detector response. Sample concentration typically 10-5 recorded in 10 mm pathlength quartz fluorescence cuvettes. M, spectra were Electrochemical measurements were carried out on a Model 660B Electrochemical Workstation (CHInstruments). Analyte concentrations were typically mm in anhydrous dichloromethane containing 0.1 M TBAP. Unless otherwise stated, a Teflonshrouded glassy carbon working electrode (CHInstruments), a Pt wire auxiliary electrode and SCE or nonaqueous Ag/Ag + ion reference electrode were employed. Reference electrodes were calibrated with 0.1 mm solutions of ferrocene (0.38 V versus SCE in 0.1m TBAP/ CH 3 CN). Solutions for reduction measurements were deoxygenated by purging with dry N 2 gas (presaturated with solvent) prior to the measurement. Cyclic voltammograms were obtained at sweep rates of between 10 mvs -1 and 50 Vs -1 ; differential pulse voltammetry (DPV) experiments were performed with a scan rate of 20 mvs -1, a pulse height of 50 mv and a duration of 0.2 s. For reversible processes the half-wave potential values are reported; identical values were obtained from DPV and CV measurements. Redox potentials are given with an accuracy of ±10 mv. PerSPer 7 and DMCSPer 8 Diacid 5 (135 mg, 0.27 mmol) was suspended in CH 2 Cl 2 (20 ml) and placed in an ice bath. Subsequently N-methylmorpholine (0.06 ml, 0.54 mmol) was added whereupon the solid dissolved. 2-Chloro-4,6-dimethoxytriazine (95 mg, 0.54 mmol) was added and the reaction mixture stirred for 4h at 0 C, after which another two equivalents of N-methylmorpholine (0.06 ml, 0.54 mmol) were added followed by the mono-piperidine-mono-n-butyl perylene bisimide 4 (300 mg, 0.27 mmol) and coumarin 2 (90 mg, 0.27 mmol). Stirring was continued for 1 h at 0 C, and overnight at room temperature. CH 2 Cl 2 (50 ml) was added and the solution was washed with, respectively, 1M aq. HCl (2 x 20ml), brine (1x 20 ml), saturated aqueous bicarbonate solution (1 x 20 ml) and H 2 (1 x 20ml). The organic phase was dried on Na 2 S 4 and the solvent was evaporated. The resulting solid crude product was purified using column chromatography (for 7: 0.5 % MeH in CH 2 Cl 2, Si 2 and 8: 2 % MeH in CH 2 Cl 2, Si 2 ), providing the dark red solids 7 (34 mg, mmol, 4.8 %) and 8 (38 mg, mmol, 7.4 %) PerSPer 7 1 H NMR (400 MHz, CDCl 3 ) δ = 8.20 (s, 4H), 8.19 (s, 4H), 7.48 (d, J = 8.3 Hz, 4H), 7.38 (d, J = 8.3 Hz, 4H), 7.22 (dd, J = 2.9, 8.9 Hz, 16H,), 7.05 (s, 2H), 6.81 (m, 16H), 5.21 (m, 2H), 4.83 (m, 1H), 4.09 (m, 4H), 3.90 (m, 1H), 3.03 (m, 2H), 2.82 (m, 4H), 2.72 (m, 6H), 2.07 (m, 2H), 1.95 (s, 6H), 1.64 (m, 10H), 1.36 (m, 4H), 1.28 (s, 36H), 1.27 (s, 36H), 0.92 (t, J = 7.3 Hz, 6H) ppm. 13 C NMR (101 MHz, CDCl 3 ) δ = 170.0, 163.7, 163.4, 156.0, 155.8, 150

152 Energy Transfer Within Two Spectroscopically Distinct Perylene Bisimide Substituted Switches 152.9, 152.8, 147.3, 138.7, 136.8, 135.7, 135.3, 134.6, 134.2, 132.9, 132.8, 127.7, 126.6, 125.1, 124.6, 124.6, 122.5, 122.5, 120.7, 120.3, 120.0, 119.8, 119.5, 119.4, 119.3, 119.2, 51.7, 51.6, 47.9, 42.5, 40.4, 38.5, 34.3, 31.4, 30.1, 29.7, 28.7, 22.9, 20.3, 14.5, 13.7 ppm. MALDI-TF MS (MW = ) m/z = [M + ]. DMCSPer 8 1 H NMR (400 MHz, CDCl 3 ) δ = 8.19 (s, 4H), 7.48 (m, 4H), 7.36 (m, J = 8.1, 12.5 Hz, 4H), 7.21 (dd, J = 3.0, 8.7 Hz, 8H), 7.06 (s, 1H), 7.04 (s, 2H), 6.80 (t, J = 8.3 Hz, 8H), 6.14 (s, 1H), 5.22 (s, 1H), 4.82 (s, 1H), 4.09 (m, 2H), 3.92 (s, 3H), 3.89 (s, 3H), 3.82 (s, 2H), 3.60 (m, 6H), 3.02 (m, 1H), 2.80 (m, 7H), 2.07 (m, 2H), 2.00 (s, 3H), 1.97 (s, 3H), 1.63 (m, 2H), 1.38 (m, 2H), 1.27 (s, 18H), 1.27 (s, 18H), 0.91 (t, 3H, J = 7.3 Hz) ppm. 13 C NMR (101 MHz, CDCl 3 ) δ = , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , 56.53, 56.50, 56.38, 56.34, 51.62, 46.28, 42.13, 40.35, 38.36, 38.11, 34.33, 31.42, 30.12, 28.64, 22.98, 20.29, 14.50, 14.43, 13.72, ppm. MALDI-TF MS (MW = ) m/z = [M + ]. 151

153 Chapter References 1 J. Areephong, W. R. Browne, N. Katsonis and B. L. Feringa, Chem. Commun. 2006, A. Mulder, A. Jukovic, J. Huskens and D.N. Reinhoudt, rg. Biomol. Chem. 2004, 2, J. J. D. de Jong, L. N. Lucas, R. Hania, A. Pugzlys, R. M. Kellogg, B. L. Feringa, K. Duppen and J. H. van Esch, Eur. J. rg. Chem. 2003, (a) J. J. D. de Jong, L. N. Lucas, R. M. Kellogg, J.H. van Esch and B.L. Feringa, Science 2004, 304, (b) Ph.D. Thesis L. N. Lucas, Towards Photoresponsive Supramolecular Materials, University of Groningen, 2001, ISBN J. steryoung, J. Chem. Educ. 1983, 60, Here, reversibility, quasireversibility and irreversibility refer to the chemical stability of the oxidised or reduced species. In all cases electrochemical reversibility (E p,a -E p,c < 80 mv) is observed between 0.01 and 2.0 Vs -1, where the chemical stability of the oxidised/reduced species is sufficient to observe the return wave. Where the I p,a peak is not significantly different to the I p,c peak, the term reversible is applied; similarly, quasireversiblity refers to situations in which chemical reversibility is dependent on scan rate (i.e. reversibility is observed only at higher scan rates). 7 (a) W. R. Browne, J. J. D. de Jong, T. Kudernac, M. Walko, L.N. Lucas, K. Uchida, J.H. van Esch and B.L. Feringa, Chem. Eur. J. 2005, 11, (b) W. R. Browne, J. J. D. de Jong, T. Kudernac, M. Walko, L.N. Lucas, K. Uchida, J.H. van Esch and B.L. Feringa, Chem. Eur. J. 2005, 11, (a) A. Peters and N. R. Branda, Adv. Mater. pt. Electron. 2000, 10, (b) M. Irie, T. Lifka, K. Uchida, S. Kobatake and Y. Shindo, Chem. Commun. 1999, This wavelength lies, according to the absorption spectra, just inside the absorption spectrum of the coumarin and just outside that of the open form of the dithienylcyclopentene switch, thereby avoiding unwanted ring closure. 10 C. kabe, T. Nakabayashi, N. Nishi, T. Fukaminato, T. Kawai, M. Irie and H. Sekiya, J. Phys. Chem. A 2003, 107, (a) E. A. A. Wallén, J. A. M. Christiaans, E. M. Jarho, M. M. Forsberg, J. I. Venäläinen, P. T. Mannisto and J. Gynther, J. Med. Chem. 2003, 46, (b) L. A. Carpino, E. M. E. Mansour, C. H. Cheng, J. R. Williams, R. Macdonald, J. Knapczyk, M. Carman and A. Lopusinski, J. rg. Chem. 1983, 48,

154 Chapter 6 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentenes In recent years molecular modelling and theoretical calculations have become an invaluable tool in helping to understand and explain experimental observations. In this chapter calculations, primarily using density functional theory, will be applied to find an additional theoretical basis and explanation for several of the observations (vide infra) made in Chapters 2 and

155 Chapter Introduction to computational chemistry 1 Computational chemistry uses theoretical models as a basis for calculating the physical properties of molecules. These properties can range from energy optimized structures to properties including vibrational structure (IR), absorption spectra (UV/Vis), charges, dipoles and reaction pathways. The validity of the results depends on a multitude of factors and is mostly dependent on the assumptions and simplifications that are used. The results should always be validated (when possible i.e. when the molecule exists) by comparison with experimental data. The limiting factor in obtaining results that are in agreement with experimental results lies, in the end, in the calculating power available, which translates as computational time, the more accurate the method, the more time that is required to complete the calculation. For small molecules, consisting of only a few atoms, the most accurate methods can be used. However, for larger molecules less accurate methods are needed to be able to enable completion of calculations within an acceptable time period. In the following sections several common methods and their (dis)advantages will be discussed. 6.2 Basic methods These methods use a combination of Newtonian mechanics and values obtained from experimental data to perform calculations, which are fast, but provide results that often deviate significantly from experimental data. 6.3 Molecular mechanics (MM) This method employs Newtonian mechanics to calculate molecular properties and is usually used to find the energy minimized structure of a molecule. Atoms within the molecule are treated as single particles of a defined radius (typically the van der Waals radius is used). These particles are connected by springs which have an equilibrium distance that was obtained from either experiment or more exhaustive calculations. For a certain conformation the potential energy of the system is calculated by separating the total energy in to covalent and non-covalent contributions (Eq. 6.1). E = E Eq. 6.1 cov alent + E non cov alent These contributions are then separated further into smaller individual interactions (Eq. 6.2 and Eq. 6.3). From the covalent interactions the bond and angle components can be treated as harmonic oscillators of a certain equilibrium value. For the dihedral component this is 154

156 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene more difficult, since several minima exist, and other methods (i.e. functionals) should be used. E = E + E + E Eq. 6.2 cov alent bond angle dihedral The noncovalent components are divided into a van der Waals and an electrostatic component. The van der Waals term takes into account repulsion and van der Waals attraction forces. These forces are usually treated in the context of a 6-12 Lennard-Jones potential, 2 which means that attractive forces decrease with distance by r -6 and repulsive forces by r -12, with r being the inter atomic distance. Electrostatic interactions are more complicated and have considerable effect over longer ranges, which cannot be ignored and can be important in large molecules (e.g., proteins). The common method of modelling the electrostatic interactions is to use the Coulomb potential as the basis functional, which has distance dependence of r -1. E + non alent = E vanderwaal s E Eq. 6.3 cov electrosta tic All these parameters and terms, combined with constants and values from experiments and theoretical calculations are known as the force field. Molecular mechanics is used for fast structure optimizations (also for quick optimization prior to using more advanced methods) and many of the fundamentals behind this method are applied in molecular dynamics simulations. However, the assumptions made and the fact that it is based on Newtonian mechanics make it unsuitable for use for prediction of electronic properties Semi-empirical methods The semi-empirical methods are built on Hartree-Fock theory (vide infra), however several approximations, constants and parameters obtained from empirical data are employed. These changes overall lead to considerable reduction in calculation time compared to ab initio based methods, which permits it to be used for large molecules. However, if empirical parameters are used which were obtained from molecules which are not similar enough to the target molecule, then the results obtained can be unreliable. Well known methods used for semi-empirical calculations are AM1 3 and PM

157 Chapter Ab initio methods Molecular orbital theory Ab initio ( from the beginning ) methods are built on molecular orbital theory. These methods use quantum mechanics and quantum field theory to solve the time-independent molecular Schrödinger equation that is coupled with the molecular Hamiltonian (Eq. 6.4). 5 H Ψ = EΨ Eq. 6.4 Ψ is the wave function, which is a mathematical function used in quantum mechanics to describe the wave characteristics of a particle or a system of particles (i.e. an atom or molecule). Its solution has no direct physical meaning, however, Ψ 2 gives the probability of finding a particle at a certain position at a certain time. The operator H, that gives the systems energy E, is called the Hamiltonian operator. The common form of this operator has five contributions to the total energy of the system (Eq. 6.5). H = k k i k + + i 2me k 2mk i k rik i< j rij k< l rkl h h e Z e e Z Z Eq. 6.5 l Where i and j are electrons, k and l are nuclei, m e is the mass of an electron, m k the mass of nucleus k, 2 the Laplacian operator, e the charge of an electron, Z the atomic number and r ik the distance between particles i and k. From left to right these five terms represent: the kinetic energy of the electrons, the kinetic energy of the nuclei, the attraction between electrons and nucleus, the interelectronic repulsion and the internuclear repulsion. Methods that do not use empirical or semi-empirical parameters in their equations are considered to be ab initio (i.e. only based on theory or results from theoretical calculations). The solution from these calculations, however, is still not an exact one; it is an approximation built up from first principles (current quantum theory). The total energy of the system under investigation is determined by solving the time-independent Schrödinger equation, using the Born-ppenheimer approximation and typically used without relativistic terms. 6 The B approximation states that since the movement of electrons is much faster than nuclear motion nuclear positions can be treated as fixed and therefore the electrons feel an average field generated by the nuclei. This allows for significant simplification of the calculations and the Hamiltonian in Eq. 6.5 can be broken up into a nuclear and an electronic part. The last term, internuclear repulsion becomes a constant for a certain geometry (and is often omitted from the Hamiltonian) and as the nuclei do not 156

158 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene move the nuclear kinetic energy term becomes zero, thereby yielding the so-called electronic Hamiltonian (H elec ) (Eq. 6.6). H elec = i h e Z k e Eq i + 2m r r e i k ik i< j ij This leads to the following electronic Schrödinger equation, combining the electronic Hamiltonian with the electronic wave function, which depends on the electron coordinates: H elec Ψ = E Ψ Eq. 6.7 elec elec elec Finally the total energy of the system (E tot ) is the sum of the E elec and the constant nuclear repulsion term: E E + E tot = Eq. 6.8 elec nucrep It is now feasible to use operators to describe properties of the system, however it is still not possible to construct a wave function that describes the system. Nevertheless, an educated guess can be made. Combining this guess with the variational principle, which states that when you have a system, which has a lower boundary, it is possible to find the best solution for a problem be systematically varying the guess until a minimum value has been reached. This minimum is the best possible solution since the boundary condition, in this case originating in quantum mechanics, makes sure a solution which is lower in energy then the exact description of the system can never be reached. This combination gives the inequality shown in Eq. 6.9 and allows for the fitting of an approximate wave function. ΦHΦdr E 2 0 Φ dr Eq. 6.9 If Φ is the exact wave function describing the system, the energy of that system will be E 0, every other approximate wave function will be higher in energy. By systematically trying guessed wave functions in principle a minimum integral can be obtained, which represents the lowest energy wave function for that level of theory. 157

159 Chapter Hartree Fock Theory In addition to the approximations described in the previous section Hartree-Fock (HF) theory makes further approximations: The wave function Φ of the system (used in combination with the variational principle) is assumed to be a linear combination of the atomic wave functions (ψ) (which is also known as the linear combination of atomic orbitals (LCA)) of the atoms present in the system. The interelectronic repulsion term from Eq. 6.6 is approximated by letting each electron feel an average repulsion effect of all electrons, instead of each separately, thereby neglecting electron correlation. This last approximation has the most significant effects. Neglecting the electron correlation can cause large deviations from experimental results. It also prevents the calculated energy from ever reaching the exact energy (E 0 ) of a system. The minimum energy that is approached by increasing the basis set size (i.e. a set of functions used to describe the molecular orbitals, see also paragraph 6.5.4) of a Hartree-Fock calculation is called the Hartree-Fock limit (Figure 6.1). Figure 6.1 Schematic representation of the Hartree-Fock limit. Several methods have been developed to compensate for this weakness. These methods are collectively called post-hartree-fock methods (e.g., Møller-Plesset perturbation theory) and use different approaches to include electron correlation into the multi-electron wave function. 158

160 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene 6.5 Density functional theory (DFT) Early work, the Thomas-Fermi model. The basis for this theory was developed independently by Thomas and Fermi in Named the Thomas-Fermi model, it uses the electron density (ρ) to calculate the potential energy of a many-electron system, instead of using the wave functions as in Hartree-Fock theory. As discussed above, the electronic Hamiltonian depends only on the position and atomic number of the nuclei and the total number of electrons. This dependency on the total number of electrons makes it possible to describe the energy of a system by using the electron density: integrating the electron density over all space gives the total number of electrons (N) (Eq. 6.10) N = ρ ()dr r Eq The location of the nuclei will correspond to local minima in the electron density, leaving only the atomic number (Z) as an unknown. The atomic number (or nuclear charge) can be derived from the properties of the maximum in electron density at the position (r k ) of atom k by: ρ ( r ) r k k r = 0 k = 2Z ρ k ( r ) k Eq where r k is the radial distance from atom k and ρ is the spherically averaged density. Since all parameters required can be described using the electron density it is possible to construct the Hamiltonian and solve the Schrödinger equation. The energy (E) can be separated into its potential (V) and kinetic (T) components. The most straightforward way is to treat the system in a classical manner; this gives the following components for the potential energy for the attraction between nuclei and the electron density: V ne Nulcei k [ ρ() r ] = ρ()dr r k r Z r k Eq and for the potential energy for the self-repulsion of a classical charge distribution: 159

161 Chapter 6 [ ρ () r ] 1 = ρ ( r ) ρ( r ) 1 V ee 2 r1 r 2 2 dr dr 1 2 Eq For the kinetic energy component Thomas and Fermi used fermion statistical mechanics to derive a formula describing the kinetic energy of a uniform electron gas (an infinite number of electrons moving in an infinite volume filled with a uniformly distributed positive charge): T [ ρ () r ] = ( 3π ) 2 / ρ 5 / ()dr r Eq The previous terms (Eq Eq. 6.14) for potential and kinetic energy are a function of the density, and the density itself is a function of the three-dimensional spatial coordinates. This means they are a function of a function, which is called a functional. Therefore, the expressions above are referred to as density functionals, and hence the name: density functional theory. This theory, however, was incomplete and only a rough approximation of the true kinetic energy. Since it did not describe the exchange and correlation effects, which are predicted by Hartree-Fock theory, in this form the theory is unsuitable to describe real molecules. Independent work by Bloch (1929), Dirac (1951) and later by Slater (1951) added a correction functional for the exchange energy, leading to the Thomas-Fermi-Dirac model. This however did not improve accuracy sufficiently to enable practical usage Practical density functional theory The first steps in the direction of a practical application of DFT were made by Hohenberg and Kohn in They showed with the Hohenberg-Kohn Existence theorem that the non-degenerate ground-state density determines the external potential of a system (i.e. the nuclei), and thereby the Hamiltonian and from there the wave function. Proving in this way that by using only electron density it is possible to determine the energy of a system. They showed also that a variational principle could be applied to DFT (similar to Eq. 6.9), showing that any energy calculated should be greater then or equal to the true ground state energy (E 0 ). This made a systematic approach to finding the best fitting result possible, since step by step minimisation of the energy of the system would lead to a result that is closest to the real ground state of the system under investigation (i.e. within the limits of the approximations used). 160

162 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene Kohn and Sham (1965) introduced an important simplification to the construction of the Hamiltonian operator. By using a system of non-interacting electrons that have the same density as the system as if they were interacting, it is possible to represent the Hamiltonian as a sum of one electron operators. It also allows a significant part of the energy to be determined exactly and therefore only a small part has to be approximated. This allows the energy of the system to be obtained as follows from the electron density: E [ ρ() r ] T [ ρ() r ] + V [ ρ() r ] + V [ ρ( r) ] + T[ ρ( r) ] + Vee[ ρ( r) ] = Eq ni ne ee where T ni is the kinetic energy of non-interacting electrons, V ne the nuclear-electron interaction, V ee the classical electron-electron interaction, T the correction to the kinetic energy due to electron interaction and V ee all the non-classical corrections due to electronelectron interaction. It is important to note that in this functional all terms are determined exactly except for correction terms T and V ee. These two terms are grouped usually as the exchange-correllation energy (E XC ). This functional can then be used to construct the one-electron (Kohn-Sham) operators for the system: ( r' ) nuclei KS Z k ρ δe xc h i = i + dr + r r ' 2 r r' δρ 1 2 Eq k i k of which the terms on the right hand side are kinetic energy, nuclear-electron interaction, classical electron-electron interaction and the exchange-correlation energy. These Kohn- Sham operators are used to construct the secular equation which is used to minimise the energy of the system and thereby obtain the ground state Exchange-correlational functionals As discussed before the exchange-correlation energy (E XC ) is the only part that cannot be solved exactly, because it has not yet been possible to produce a functional that can determine this energy and that is based on only the electron density (i.e. ρ). Several different functionals have been developed to deal with the difference in kinetic energy between the real and non-interacting system and the non-classical electron-electron interactions, of which the most commonly used one, Becke s three parameter exchange functional (B3) combined with the Lee-Yang-Parr correlational functional (LYP) (which makes B3LYP), will be discussed in more detail. This functional can best be described as a combination of the best methods from several theories (it is therefore called a hybrid functional), and is depicted as: i 161

163 Chapter 6 B LYP xc LSDA HF B LSDA LYP ( 1 a) E + ae + b E + ( + c) E ce E = Eq x x The first two terms are a combination of exact Hartree-Fock exchange energy (E HF X), corrected using the local spin density approach (LSDA) which uses uniform-electron gas to determine the value of the energy at any position using electron density. E B X is a functional by Becke which corrects for the fact that the density is not only dependent on the local value but also on the gradient (i.e. the way the density is changing locally). E LSDA C is the local term for electron corelation based on LDSA and E LYP C the gradient correction of the electron correlation based on the functional by Lee, Yang and Par. This hybrid functional uses 3 constants (i.e. a, b and c), which were optimized by fitting to heats of formation of small molecules and are set to 0.20, 0.72 and 0.81 respectively. This functional is the most widely used exchange-correctional function and shows good results 8 (i.e. it compares well with experimental results) for small and relatively large systems in combination with relatively low calculation cost Basis set To solve the Schrödinger equation modern computer programs use a basis set to approximate orbitals. These basis sets use a combination of mathematical functions to describe the atomic orbitals. The larger this basis set, the better the description of a system, an infinitely large basis set would, in theory, give the best results. However, this would be impossible to use in practice since this would require unacceptable computational time. Therefore, a compromise has to be made between the accuracy required and time available. x C C For calculations described in this chapter the 6-31G(d) (also known as 6-31G*) basis set was used. 9 This implies the 6-31G valence double-ζ basis set, (10s4p/4s)[3s2p/2s], which was combined with an additional set of 6 d functions on C and N to allow for polarisation of the p orbitals Advantages and limitations of DFT Currently, comparing results (i.e. best fit to empirical results) and computational cost (i.e. computer power required to calculate using a certain method in a reasonable amount of time), DFT (i.e. in combination with appropriate functionals and basis-set) gives the best results compared to methods requiring similar amounts of computational power (i.e. HF, MP2). In combination with hybrid functionals, such as the popular B3LYP, which are among the most useful and cost efficient at the moment, DFT can be a useful tool in understanding experimental results. 162

164 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene ne of the disadvantages of DFT is that it is not possible to improve the methods systematically (unlike post-hf methods such as Møller-Plesset or coupled cluster theory) to reach the final goal, the exact description of a system. It is therefore not possible to estimate errors without comparison of results to other methods or experimental data. For the calculation of vibrational frequencies DFT categorically shows better results then HF. The calculation of IR spectra using B3LYP/6-31G(d) gives results comparable or better then second order Møller-Plesset theory (MP2) at reduced computational cost. 10,11 Infrared vibrational frequencies are on average overestimated by ~ 10% due to the approximation of the electron correlation and thereby an overestimation of bond strength. The agreement with experimental results can be increased by using a scaling factor Experimental Hartree-Fock (HF) and Density functional theory (DFT) calculations were carried out with the GAUSSIAN 03W (rev. B.04) program package. 12 Gaussian 03W was run on a computer system equipped with an Intel Pentium D 3.2 GHz CPU and two gigabytes of RAM with Windows XP SP2 as the operating system. GaussView 3.0 was used to prepare jobs and to analyse the results. 13 GaussSum 14 was used to monitor jobs, calculate (partial) density of states ((P)DS) spectra, and to convolute IR spectra. All the DFT calculations were performed on systems in the gas phase using the Becke s three-parameter hybrid functional (B3) 15 with the LYP correlation functional 16 (B3LYP) and 6-31G(d) basis set. The geometry optimization was followed by the frequency calculation to prove that the energy minimum was found. btained frequencies were scaled with a factor of , a value which was determined by Scott and Radom, 11 and was obtained by comparing 1066 experimental vibrations of 122 molecules to their calculated frequencies. This scaling factor is required to compensate for the HF character of B3LYP and by using it the B3LYP/6.31G(d) method gives the lowest RMS error of all methods studied in ref. 11. Assignments of vibrations were made using the possibility to visualize vibrations using GaussView FT-IR spectra were recorded (as intimate mixtures in KBr) in reflectance mode using a Nicolet Nexus FT-IR spectrometer. 163

165 Chapter Calculations on three perylene bisimide models Introduction Three differently substituted perylene bisimides (i.e. 1a, 2a and 3a) have been synthesized and their electronic and redox properties been investigated (see Chapter 2). The experimentally observed absorption maxima (λ max ) showed, that when going from more electron donating (i.e. butyl) to a more electron withdrawing group (i.e. isophthalic) there was a decrease in the HM LUM gap. DFT calculations were employed to predict the orbital nature of the frontier molecular orbitals of perylene bisimide 1b, 2b and 3b in order to investigate if the observed effect could be contributed to the imide substituents. The validity of the calculations was assessed by comparison of predicted physical data with FT-IR spectroscopy, cyclic voltammetry and UV/Vis data of synthesized perylene bisimides 1a, 2a and 3a (vide infra). Figure 6.2 The investigated perylene bisimide, both synthetic compounds 1a, 2a and 3a and simplified models for calculations 1b, 2b and 3b. Calculations were performed on suitable models for perylene bisimide 1a, 2a and 3a. These models have structures similar to the synthesized compounds, some alkyl extensions were removed where possible to decrease calculation time and to increase symmetry. For all calculations instead of 4-tert-butylphenol, 4-methylphenol bay substituents were used, which allows for an increase in symmetry for both molecules. Since the tert-butyl groups are spatially quite separated from the rest of the molecule no steric effects are expected 164

166 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene from this and because the pk a values for 4-tert-butylphenol and p-cresol only differ slightly (10.16 vs , respectively) there is not expected to be any significant electronic effect. For 1a the symmetry group C2 could be used, which almost halves the calculation time for 1b. For 2b and 3b a D2 symmetry could be used. For 3b the calculation time could be reduced further by replacing the piperidine groups by dimethylamine moieties, a change which not expected to influence the properties of the amide, but could, however, have some steric influence on the area around the carbonyls. Structures were first optimized using the Hartree-Fock method and a 6-31G(d) basis set. After this first optimization a more strict optimization was performed using DFT with the Becke s three-parameter hybrid functional 15 using the LYP correlation functional 16 and the 6-31G(d) basis set, which has shown a good agreement with experimental data in similar structures. 17 These energy optimized structures were then used to perform a frequency calculation (IR) using the same functional and basis set. 6.8 Partial density of state diagrams Partial density of state diagrams (PDS) can be used to show the contribution of different parts with a molecule to the density of states and energy of the whole molecule. The partial density of state diagrams in Figure 6.3 show the influence of the imide substituent of the perylene bisimide on the frontier orbitals. Going from 1b to 3b, it is clear that this influence increases when going from more electron donating (i.e. butyl) to a more electron withdrawing group (i.e. isophthalic). However, the substituent shows no direct influence on either HM or LUM in any of the molecules examined. In all cases the perylene bisimide dominates the LUM orbitals. This is the case for the HM also, however, the influence of the substituents on the frontier orbitals becomes larger when going to the more electron withdrawing substituents and for the 3b the substituent has a clear contribution to the HM-1. There is no clear trend in the comparison of the HM and LUM difference values ( ev) obtained by DFT with the V values for the separation between first oxidation and first reduction (Table 6.1). The values obtained from cyclic voltammetry show a decrease in the gap between oxidation and reduction (which is related to the HM LUM gap), with decreasing electron withdrawing character of this substituent, in agreement with UV/Vis spectral data (i.e. a decrease in λ max ) (Chapter 2). This trend is not apparent for ev obtained from calculations, where only very small differences are observed. DFT calculations performed by Pichierri on perylene bisimides without bay substitution, 18 however, show the same trend in the calculations when going from an alkyl group to a 165

167 Chapter 6 phenyl group to a phenyl group with electron withdrawing substituents. Both sets of calculations were performed in the gas phase, which may be the origin of the differences in the values that were calculated, compared with the data from cyclic voltammetry and UV/Vis spectra, which originate from measurements which where performed in solution (CH 2 Cl 2 ). Since the differences observed are quite small, it is possible that solvent effects (i.e. polarity, H-bond formation) are dominant. 19 Table 6.1 First oxidation and reduction potentials of 1a, 2a and 3a combined with calculated ev values for HM and LUM. Comp. x. (V) a Red. V b Model HM (V) a (ev) c LUM (ev) c ev c 1a b -5,195-2,888 2,307 2a b -5,194-2,883 2,311 3a b -5,474-3, a Differential pulse voltammetry, b V is the separation between E ½ of the first oxidation and the first reduction. (V vs. SCE, in CH 2 Cl 2 / 0.1 M TBAPF 6 ), c Values obtained from PDS diagrams Molecular orbitals diagrams The orbital diagrams for the three systems (i.e. 1b, 2b and 3b) of LUM + 1, LUM, HM and HM - 1 (Figure 6.4) show no electron density on the imide substituents and therefore they are not likely to have an influence on the HM LUM gap. The only exception is the HM - 1 of 3b, which is also the molecule which shows the largest shift in the UV (i.e. 3a) compared to 1a. This indicates that the influence of the substituents on the HM - LUM gap is unlikely to be exerted through bond. However, another observation can also be made from the orbital diagrams. In all three molecules the perylene bisimide carbonyls are involved in both HM and LUM. Electron density on the carbonyl groups is sensitive to environmental effects (e.g., solvent, hydrogen bonding, and Lewis acids). 19 Even if the imide substituents are not involved directly (i.e. through bond, by influencing HM and/or LUM) in the HM LUM energy gap, they do influence the direct environment of the carbonyls, either sterically (e.g., by hindering solvent molecules around the carbonyls) or electrostatically (e.g., through interaction with the π system of the differently substituted phenyl groups). 166

168 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene 1b LUM HM Perylene Butyl ev b LUM HM Perylene Phenyl ev b LUM HM Perylene IsophthalDiMeAmine ev Figure 6.3 Partial density of states (PDS) diagrams of 1b, 2b and 3b with contribution of R1 groups to the DS (grey filled areas, Figure 6.2). 167

169 Chapter 6 1b 2b 3b LUM + 1 LUM + 1 LUM + 1 LUM LUM LUM HM HM HM HM - 1 HM - 1 HM - 1 Figure 6.4 Frontier molecular orbital diagrams of LUM + 1, LUM, HM and HM - 1 of 1b, 2b and 3b (DFT B3LYP/6-31G(d), isosurface = 0.03). 168

170 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene Frequency calculations Frequency calculations on the optimized structures provided the calculated IR spectra for compounds 1b, 2b and 3b. 20 Comparison of the FT-IR spectra of 1a, 2a and 3a with the calculated IR spectra of 1b, 2b and 3b shows good agreement (Figure 6.5 and Figure 6.6 respectively). This indicates the optimized model structures obtained from the calculations (1b, 2b and 3b) show good resemblance with the molecules synthesized (1a, 2a and 3a). Comparison of the FT-IR spectra of 1a and 2a with the calculated spectra of 1b and 2b, respectively, show that overall a good agreement is obtained. Differences between the calculated and experimental spectra are due to vibrations of the tert-butyl groups on the bay phenyl substituents (R 2, Figure 6.2), which are not observed for the calculated spectra due to the replacement of the tert-butyl by a methyl substituent. Comparing just the experimental (1a, 2a and 3a) or the calculated spectra (1b, 2b and 3b) there are, of course, differences due to different IR absorption bands for the butyl, phenyl and isophthalic imide substituents. Bands for the core perylene and bay phenyl substituents remain at the same position for both groups of molecules, both calculated and experimental (Table 6.2). More important are the two characteristic imide bands (i.e. imide C=, ν s and ν as ) in the FT-IR spectra of 1a, 2a and 3a show a shift towards higher energy going from 1a to 2a to 3a (Table 6.2). This indicates an influence of the imide substituents on the carbonyl bond strength frequency and since the carbonyls have a contribution to the frontier orbitals as shown in the molecular orbital diagrams (Figure 6.4), the substituents, therefore, can still via their influence on the imide vibrations have an effect on the HM-LUM levels of the perylene core. However, when comparing the IR absorption bands of 1b, 2b and 3b, as with the values obtained from the PDS diagrams, the expected trend breaks down. In the FT-IR spectra the trend from 1a to 2a to 3a seems to be towards higher energy for the imide carbonyls, however, this trend is not observed in the calculated spectra (Figure 6.6). From 1b to 2b the increase in energy is present in accordance with the experimental data, however for 3b the energy is lower in energy, as was also observed for the PDS diagrams. As discussed for the PDS diagrams, the calculations were performed with molecules in the gas phase, which might be the cause of the different trend between calculated and measured IR data; the latter being performed in the solid state. 169

171 Chapter 6 Table 6.2 Characteristic bands from the calculated and experimental infrared spectra of 1a,b; 2a,b and 3a,b. Experimental (ν, cm -1 ) Calculated (ν, cm -1 ) a Assignments b 1a, Bu 2a, Ph 3a, Iso 1b, Bu 2b, Ph 3b, Iso Imide C= ν s Imide C= ν as Iso imide C= ν s ν as Perylene ν C-C, δ C-H Perylene and bay phenyls δ C-H Imide methylene δ C-H Perylene ν C-C, δ C-H a DFT B3LYP/6-31G(d) scaled by b btained by visualization with Gaussview 13 and checked with known literature values Conclusions Even though the bisimide substituents do not have a direct contribution to the HM and LUM orbitals, as seen in the PDS and molecular orbital diagrams, the agreement of experimental and calculated IR spectra seems to indicate that the substituents are able to influence the frontier orbitals indirectly by influencing the electron density on the imide carbonyls and thereby make a considerable contribution to the HM and LUM levels of 1b, 2b and 3b. However, the trend observed experimentally was not in full agreement with the calculations, indicating that the model used was not sufficiently realistic, i.e. other factors are involved such as solvent environment, etc. 170

172 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene Wavenumbers (cm -1 ) Figure 6.5 Infrared spectra of 1a (top), 2a (middle) and 3a (bottom) Wavenumbers (cm -1 ) Figure 6.6 Calculated infrared spectra of 1b (top), 2b (middle) and 3b (bottom). (DFT B3LYP/6-31G(d) scaled by ) 171

173 Chapter Exploring the remarkable different behaviour of two diphenyldithienylcyclopentene switches coupled with differently substituted coumarin donor groups Introduction In Chapter 4 two modified diarylethene switches were described that show different behaviour with respect to their fluorescence properties. Both switches 4 and 5 show full quenching of the coumarin fluorescence in the ring-closed state, which was expected because of the overlap in the donor emission and acceptor absorption spectra. However, in the open state different behaviour was observed. The mono-methoxy coumarin functionalized switch 4 (MMCS) showed no coumarin fluorescence in the open state. In contrast, the dimethoxy coumarin functionalized switch 5 (DMCS) shows no quenching of the coumarin fluorescence (i.e. compared to the coumarin model 7a) in the ring-open state. In this paragraph DFT calculations will be used in combination with experimental data to rationalize this difference in behaviour in the ring-open forms of 4 and 5. Special attention will be paid to the position of the relative energy levels of the coumarin models (6b and 7b) compared to each other and compared to the energy levels of the ring open form of the diarylethene switch. Calculations were performed as described in the experimental 6.6 and as described in using suitable models (Figure 6.8) and the results from the frequency calculations were compared to experimental FTIR spectra of the compounds under investigation to validate the calculation method used. 172

174 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene Figure 6.7 Structure of the diarylethene switches MMCS 4 and DMCS 5. 6a 6b MMC R 1 DMC N 7a N R 1 = R 2 = NHBoc R 1 =NMe 2 7b R 2 =NMe 2 R 2 NHBoc S S PhSPhC R 3 8a 8b R 3 = N H R 3 =NMe 2 Figure 6.8 Structure of the compounds prepared and as described in Chapter 4 (6a, 7a and 8a) and models used in the DFT calculations (6b, 7b and 8b). 173

175 Chapter Partial density of state diagrams The PDS diagrams for MMCS 4 and DMCS 5 show the energy levels of the diphenyl dithienylcyclopentene units and coumarin units within the whole molecule and also of the separate model components (Figure 6.9 and Figure 6.10). Three important observations can be made: First the differences in the energy levels of the diphenyl dithienylcyclopentene unit are negligible in both 4 and 5 (Table 6.3), as expected, since this part of the molecule is identical in both systems and is separated from the coumarin by the piperazine bridge. Secondly the HM-LUM gap for both coumarins is found to be of equal size (both in the separate models 6b and 7b as in 4 and 5). This is unexpected since this would imply an identical lowest energy λ max of absorption for both coumarins, which is not the case (λ max = for 6a 323 nm, for 7a 346 nm, E = 23 nm 0.24eV). Thirdly, compared to values obtained with cyclic voltammetry (CV) (Chapter 4), the energy of the HM levels (from the first oxidation, obtained from CV) for the coumarins are reversed compared to what has been calculated using DFT. For 6a the first oxidation is observed at 1.78 V [vs SCE] and for 7a at 1.48 V [vs SCE]. This a considerable difference, both in absolute and relative energy, compared to that obtained from the DFT calculations. However, from CV it is clear that the first oxidations of both 4 and 5 are localised on the diarylethene units in agreement with the PDS diagrams. The differences can be rationalised by consideration of the environment of the molecules (i.e. solvent, solvent polarity, solutes), both coumarins have groups that can interact with the environment (carbonyls, methoxy groups) and DMC has one extra methoxy group, so one would expect more possibilities for interaction, however it is surprising that the DFT calculations indicate that the extra methoxy group has no influence on the HM-LUM gap, a result which seems highly unlikely. The LUM levels of the diarylethene and coumarins (both for the separate models and for 4 and 5) are grouped closely together (Figure 6.9), with DMC having the lowest energy. MMC and the diarylethene switch are close in their LUM energies, if change in e.g., environment could change the position of one of these levels it would be possible that for MMCS 4 the LUM of the whole molecule would be localised on the diarylethene instead of on the coumarin unit, thereby making energy transfer from MMC to the open diarylethene a possibility. This possibility could explain the absence of fluorescence for the open form of MMCS

176 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene Table 6.3 Energy values for the calculated orbitals of MMCS 4, DMCS 5 and individual components (values obtained from PDS diagrams). Compound Part HM (ev) LUM (ev) ev a MMCS 4 MMC PhSPhC DMCS 5 DMC PhSPhC HM-LUM gap (ev) b b MMC c 7b DMC c 8b PhSPhC a Difference between HM and LUM level of the component switch and coumarin subunits. b Difference between the HM and LUM level within the molecule as a whole. c Compared to the HM level of 8b ev LUM PhSPhC MMC DMC -6 HM Figure 6.9 Partial density of states (PDS) diagrams of 6b, 7b and 8b showing the location of the HM and LUM levels for MMC, DMC and PhSPhC. 175

177 Chapter 6 a 0-2 LUM MMC Piperazine PhSPh -4 ev HM b 0-2 LUM DMC Piperazine PhSPh -4 ev -6 HM Figure 6.10 Partial density of states (PDS) diagrams of a) MMCS 4 and b) DMCS 5. Also showing the contribution of the different coumarin sidegroups (MMC and DMC) to the frontier orbitals (DFT B3LYP/6-31G(d)). 176

178 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene Molecular orbital diagrams For both MMCS 4 and DMCS 5 the frontier molecular orbital diagrams show similar features (Figure 6.11). HM-1, HM and LUM+1 are located on the diarylethene moiety for both molecules. The LUM for both molecules is located on the coumarins. Considering the physical properties of the open state of MMCS 4 and DMCS 5 this is unexpected in the case of 4. The similarities of their orbital structure would suggest that they should exhibit similar properties, however, as expected from the PDS diagrams, the calculated frontier orbitals do not seem to correspond to the experimental data in the case of MMCS 4 (Chapter 4). For this compound one would expect a LUM which is also localised on the diarylethene moiety, thereby making the diarylethene HM-LUM transition the lowest energy one and thereby making MMC fluorescence unfavourable. Considering the fact that Dexter energy transfer requires wave function overlap (which is not necessarily through bond in character) and since both LUM (located on the coumarin) and LUM+1 (located on the diaryethene) are proximate it should be possible with the correct orientation of the coumarin and diarylethene for the orbitals to overlap. If this is combined with favourable matching of the donor and acceptor energy levels, it is possible that the coumarin excited state will be quenched by the diarylethene. Since the LUM s of the MMC and the diarylethene are closest in energy, as observed from the PDS diagrams, the probability of energy transfer in that case is expected to be greater. 177

179 Chapter 6 MMCS LUM + 1 DMCS MMCS LUM DMCS MMCS HM DMCS MMCS HM - 1 DMCS Figure 6.11 Frontier molecular orbital diagrams of LUM + 1, LUM, HM and HM - 1 of MMCS 4 and DMCS 5 (DFT B3LYP/6-31G(d) isosurface = 0.03). 178

180 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene Frequency calculations Frequency calculations were performed on the optimized structures of 4, 5, 6b, 7b and 8b. The calculated spectra of the models show good agreement with compounds 4 and 5, i.e. important IR absorptions (e.g., from carbonyls) that are present in the spectra of the models can be seen in the spectra of switches 4 and 5 at identical locations. The spectra of 4 and 5 are almost a summation of the calculated IR spectra of their respective component unit models (i.e. 6b and 8b for 4 and 7b and 8b for 5, Table 6.4). This indicates that for the DFT calculations of the separate units (6b, 7b and 8b) the results are electronically identical to those performed on the full switches 4 and 5, supporting the conclusion made in Chapter 4 that the components are electronically isolated (Figure 6.12b and Figure 6.13b). The same holds for the experimental FT-IR spectra of 4 and 5 compared to those of models 6a, 7a and 8a. The data are in agreement except for the additional absorption of the Boc group in the coumarin models, which shows a strong absorption at ~ 1700 cm -1. However, features are much more difficult to resolve due to broadness of the peaks. Comparison of the FT-IR spectra with the calculated spectra is not trivial in this case, however, using the spectra of the model compounds (some peaks are clearly present only in the coumarin model compounds (6 and 7, respectively) and some only in that of the switch model 8) it is possible to assign several common absorption bands. The area between 1300 and 1500 cm -1 is crowded and therefore it is not possible to assign bands to specific vibrations with reasonable confidence. For MMCS 4 the agreement is quite good between the FT-IR and calculated spectra, the largest differences are observed for groups which are sensitive to their environment (i.e. carbonyls, amides), which is possible due to the difference between relating to the gas phase of the calculations and the FT-IR spectra which were obtained in the solid state. For DMCS 5 a similar comparison can be made as for MMCS 4. There is, however, one exception; the position of carbonyl absorption band of the calculated dimethoxy coumarin (both for 5 and 7b) is shifted considerably towards higher energy compared to that observed in the FT-IR spectrum. There is no clear direct explanation for this observed shift (DFT optimization of the structure shows no abnormalities and no imaginary frequencies were obtained for the calculation, which would indicate a saddle point). Considering the fact that in the PDS diagrams the HM LUM gap for both coumarins is identical, it is likely that for the dimethoxycoumarin the calculations in this form (i.e. level of theory / basis-set) do not provide reliable results

181 Chapter 6 Table 6.4 Comparison between calculated spectra of 4, 5, 6b, 7b and 8. a Vibrations (ν, cm -1 ) MMC 6b MMCS 4 8b DMCS 5 DMC 7b Coumarin C= ν s Coumarin Amide ν s Switch Amide ν s Coumarin rings ν C-C, δ C-H Coumarin rings ν C-C, δ C-H Switch Phenyls ν C-C, δ C-H Coumarins rings ν C-C, δ C-H Switch δ C-H Coumarin δ C-H Switch and Coumarin δ C-H Piperazine and switch γ C-H Switch Amide Phenyl ν C-C, δ C-H Piperazine/amide γ C-H Amide ν C-N, δ C-H DMC ν C-C, δ C-H Coumarin ν C-C, δ C-H a DFT B3LYP/6-31G(d) scaled by Table 6.5 Comparison between experimental and calculated spectra of 4 and 5. MMCS 4 DMCS 5 Vibrations (ν, cm -1 ) Exp Calc a Exp Calc a Coumarin C= ν s Coumarin Amide ν s 1652(sh) b (sh) b 1694 Switch Amide ν s Coumarin rings ν C-C, δ C-H (sh) b 1616 Switch Phenyls ν C-C, δ C-H Coumarins rings ν C-C, δ C-H Switch δ C-H Coumarin ν C-C, δ C-H a DFT B3LYP/6-31G(d) scaled by b Shoulder, but present in respective model coumarins 6a/7a 180

182 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene a Wavenumbers (cm -1) a a 70 * b Wavenumbers (cm -1 ) b 200 8b Figure 6.12 a) FT-IR spectra of 4 (top), 6a (middle, * = Boc group ) and 8a (bottom). b) Calculated infrared spectra of 4 (top), 6b (middle) and 8b (bottom) (DFT B3LYP/6-31G(d) scaled by ). 181

183 Chapter 6 a Wavenumbers (cm -1 ) a * 8a b Wavenumbers (cm -1 ) b 200 8b Figure 6.13 a) FT-IR spectra of 5 (top), 7a (middle, * = Boc group ) and 8a (bottom) b) Calculated infrared spectra of 5 (top), 7b (middle) and 8b (bottom) (DFT B3LYP/6-31G(d) scaled by ). 182

184 Application of Density Functional Theory to the Electronic Structure of Perylene Bisimides, Coumarins and Dithienylcyclopentene Conclusions The results obtained for both switches give some good indications as to why there is such a difference in behaviour between MMCS 4 and DMCS 5 in the open form which was described in chapter 4 (i.e. possibility for Dexter type energy transfer in MMCS 4). However, it should be borne in mind, that to compare the energy differences between the two different coumarins (i.e. DMC and MMC) and the switch (PhSPhC), it would be better to not compare the coumarin LUM with the switch HM, but to compare the coumarin first excited state with the ground state (HM) of the switch. Calculations on the coumarin first excited state, however, were troublesome and, unfortunately, were unsuccessful until now. Furthermore, there appears to be a fundamental disagreement between the vacuum DFT calculations and experiment (identical HM-LUM gap for the coumarins and an unexpected shift of the coumarin carbonyl band in DMCS 5), which makes it difficult to draw conclusive comparative inferences from the results. It also shows that, even though calculations are an increasingly commonplace tool, the gas phase (or vacuum) conditions under which calculations are performed are very different from the solvated environment for which physical data are recorded References 1 Based on the following sources: (a) C.J. Cramer, Essentials of Computational Chemistry, Wiley N.Y., (b) W. Koch and M.C. Holthausen, A Chemist's Guide to Density Functional Theory, Wiley-VCH, Weinheim, sec. ed., J. E. Lennard-Jones, Proc. Phys. Soc. 1931, 43, M.J.S. Dewar, E.G. Zoebisch, E.F. Healy and J.J.. Stewart, J. Am. Chem. Soc. 1985, 107, J.J.P. Stewart, J. Comput. Chem. 1989, 10, (a) E. Schrödinger, Ann. Phys. 1926, 79, (b) E. Schrödinger, Phys. Rev. 1926, 28, M. Born and R.ppenheimer, Ann. Phys. 1927, 84, P. Hohenberg, W. Kohn, Phys. Rev. B 1964, 136, B Recent studies have shown, however, that for some larger molecules considerable deviations can occur: P.R. Schreiner, Angew. Chem. Int. Ed. 2007, 46, (a) R. Ditchfield, W. J. Hehre and J.A. Pople, J. Chem. Phys. 1971, 54, (b) W. J. Hehre, R. Ditchfield and J. A. Pople, J. Chem. Phys. 1972, 56, (c) P. C. Hariharan and J. A. Pople, Mol. Phys. 1974, 27, ((d) M. S. Gordon, Chem. Phys. Lett. 1980, 76, (e) P. C. Hariharan and J. A. Pople, Theo. Chim. Acta 183

185 Chapter , 28, (f) J. P. Blaudeau, M. P. McGrath, L. A. Curtiss and L. Radom, J. Chem. Phys. 1997, 107, (g) M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley, D. J. DeFrees, J. A. Pople and M. S. Gordon, J. Chem. Phys. 1982, 77, (h) R. C. Binning Jr. and L. A. Curtiss, J. Comp. Chem. 1990, 11, (i) V. A. Rassolov, J. A. Pople, M. A. Ratner and T. L. Windus, J. Chem. Phys. 1998, 109, (j) V. A. Rassolov, M. A. Ratner, J. A. Pople, P.C. Redfern and L.A. Curtiss, J. Comput. Chem. 2001, 22, X.F. Zhou, C.J.M. Wheeless and R.F. Liu, Vib. Spectrosc. 1996, 12, A.P. Scott, L. Radom, J. Phys. Chem. 1996, 100, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. chterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain,. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. rtiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, and J. A. Pople, Gaussian 03, Revision B.04, Gaussian, Inc., Wallingford CT, GaussView, Version 3.09, R. Dennington II, T. Keith, J. Millam, K. Eppinnett, W.L. Hovell, and R. Gilliland, Semichem, Inc., Shawnee Mission, KS, N. M. 'Boyle, GaussSum 2.0, Available at 15 A. D. Becke, J. Chem. Phys. 1993, 98, C. Lee, W. Yang and R. G. Parr, Phys. Rev. B 1988, 37, (a) D. Zoon and A.M. Brouwer, ChemPhysChem 2005, 6 (8), (b) N.U. Zhanpeisov, S. Nishio and H. Fukumura, Int. J. Quantum Chem. 2005, 105 (4), F. Pichierri, J. Mol. Struct. (THECHEM) 2004, 686, (a) K. Das, B. Jain and H.S. Patel, J. Phys. Chem. A 2006, 110, (b) I. Georgieva, N. Trendafilova, A. Aquino and H. Lischka, J. Phys. Chem. A 2005, 109 (51), Software limitations unfortunately did not allow calculation of the IR spectrum of 3b via the standard FREQ method and FREQ=NUMER was used. Results obtained using this method are in agreement with results obtained from the PDS diagrams. 21 By using larger basis sets it may be possible to obtain results which are closer to empirical data. This however would take extensive calculating effort, also because this would have to be performed for all molecules, since otherwise comparisons cannot be made and time did not allow for this. 184

186 Samenvatting Het Controleren van Energieoverdracht tussen Chromoforen: Schakelbare Moleculaire Fotonische Systemen Chemie is de tak van de wetenschap die zich bezighoudt met het synthetiseren (het bouwen) en het bestuderen (meten en berekenen) van moleculen. De organische chemie houdt zich specifiek bezig met moleculen die voornamelijk zijn opgebouwd uit koolstof, waterstof en zuurstof, dezelfde bouwstenen waar de levende natuur (en dus wijzelf ook) grotendeels uit bestaan. Door de juiste reacties opeenvolgend te toe te passen kunnen deze bouwstenen zo met elkaar worden verbonden dat de moleculen die ontstaan de gewenste eigenschappen bezitten (Figuur 1). Hierbij kan dan gedacht worden aan bijvoorbeeld medicijnen, plastics of kleurstoffen. Figuur 1 rganische chemie: door de juiste reactiestappen te combineren kan uit simpele bouwstenen een gecompliceerd molecuul worden gemaakt. De moleculen (zogenaamde chromoforen) die in dit proefschrift worden beschreven hebben de specifieke eigenschap dat ze in staat zijn om licht (foton, energie) op te nemen (absorptie) en vervolgens weer uit te zenden (fluorescentie). Wanneer dit soort moleculen worden samengebracht in één groot molecuul dan is het, wanneer de correcte combinatie wordt gekozen, mogelijk dat moleculen de energie aan elkaar overdragen (Figuur 2). Deze energieoverdracht is een principe wat de natuur al miljoenen jaren gebruikt. Het is de manier waarop planten de energie van de zon gebruiken om uiteindelijk de chemische energie te produceren waarvan ze leven. Figuur 2 Wanneer de juiste combinatie van chromoforen word gebruikt kan door de donor (D) met de juiste golflengte (hν) aan te slaan de energie worden overgedragen naar de acceptor (A) die dan een andere golflengte uitstraalt (hν ) 185

187 Samenvatting Een molecuul kan in een aangeslagen toestand worden gebracht door het molecuul te bestralen met licht van een golflengte die wordt geabsorbeerd. Energieoverdracht kan daarna plaats vinden van het molecuul in de aangeslagen toestand (de donor) naar een ander molecuul (de acceptor) waarvan de aangeslagen toestand na de energieoverdracht van gelijke of lagere energie is. Door een molecuul met een hoog energetische aangeslagen toestand te combineren met een molecuul met energetisch lager gelegen aangeslagen toestand kan een systeem worden ontworpen dat in staat is tot energieoverdracht. Het doel van dit proefschrift is te laten zien dat het mogelijk is om op een rationele manier systemen te ontwerpen, die in staat zijn tot energieoverdracht en om te laten zien dat het mogelijk is om die energieoverdracht te sturen en controleren. Hoofdstuk 1 beschrijft de theoretische achtergrond van energieoverdracht en laat aan de hand van een aantal voorbeelden zien wat er al bereikt is op dit gebied. In Hoofdstuk 2 wordt beschreven hoe is begonnen met zoeken naar een goede combinatie van chromoforen om efficiënte energieoverdracht te bewerkstelligen. De combinatie van een donor (coumarin) met een acceptor (peryleen bisimide) geeft de gewenste hoge efficiëntie bij energieoverdracht en ook een goede stabiliteit onder intense bestraling. De twee chromoforen werden gecombineerd tot een systeem met vier donoren en één acceptor. Dit systeem is in staat om UV licht (dat onzichtbaar is voor het menselijk oog) heel efficiënt om te zetten in rood licht (Figuur 3). Figuur 3 De schematische opbouw van het vier donoren één acceptor systeem dat in staat is om UV licht te om te zetten naar rood licht met behulp van energieoverdracht (ET). 186

188 Het Controleren van Energieoverdracht tussen Chromoforen: Schakelbare Moleculaire Fotonische Systemen Het derde hoofdstuk beschrijft het ontwerp en de synthese van een systeem waarin de energieoverdracht tussen de donor en acceptor kan worden gecontroleerd. Door gebruik te maken van het in Hoofdstuk 2 ontworpen donor-acceptor paar werd een systeem ontworpen dat het mogelijk maakt door met behulp van een diaryletheen schakelaar de energieoverdracht tussen beide chromoforen in het donor-schakelaar-acceptor systeem te controleren (Figuur 4). Dit wordt bereikt doordat de schakelaar in de gesloten vorm dient als een energie put die de energie van de coumarin opneemt en niet meer loslaat (behalve als warmte). Figuur 4 Een donor-schakelaar-acceptor systeem: door de schakelaar om te zetten kan de energieoverdracht (ET) tussen het donor en het acceptor gedeelte worden uitgeschakeld en weer aangeschakeld. m een meer robuust systeem te ontwerpen werden de componenten gebruikt in Hoofdstuk 2 en 3 aangepast en/of vervangen door alternatieven met mogelijk verbeterde eigenschappen. Dit leidde uiteindelijk tot een aangepaste donor (een coumarin met betere efficiëntie bij energieoverdracht) en een verbeterde schakelaar (met hogere stabiliteit en efficiënter schakel gedrag). Het werk dat beschreven staat in Hoofdstuk 4 laat op deze manier zien dat kleine variaties in molecuul struktuur grote gevolgen kunnen hebben voor de eigenschappen (Figuur 5). S S pen Vis N UV N Intensiteit 2.0M 1.5M 1.0M 500.0k pen S S Gesloten N N 0.0 Gesloten Golflengte (nm) Figuur 5 Het verbeterde schakelaar-donor systeem (links) is in staan om de fluorescentie van de coumarin bijna tot nul te reduceren (rechts). 187

189 Samenvatting In Hoofdstuk 5 worden de lessen die zijn geleerd in de voorgaande hoofdstukken benut om een systeem te ontwerpen en te synthetiseren dat de beste eigenschappen van alle onderdelen combineert: een efficiënt donor acceptor paar (Figuur 6), hoge stabiliteit en goed schakelgedrag. Het uiteindelijke systeem vertoont zeer efficiënt schakelgedrag en is in staat om de fluorescentie van de acceptor met 80% te verminderen. pen N N S S N N N Vis UV S S Gesloten Figuur 6 De verbeterde donor-schakelaar-acceptor met een hoge mate van controle over het schakelgedrag en ook over de energieoverdracht. Het zesde hoofdstuk laat zien dat theoretische chemie een belangrijk hulpmiddel kan zijn bij het ontwerpen en voorspellen van eigenschappen van systemen. Met behulp van berekeningen die uitgaan van de elektronendichtheid rondom een molecuul (DFT) worden de eigenschappen van een aantal moleculen die zijn behandeld in Hoofdstuk 2 en 4 berekend (Figuur 7) en vervolgens vergeleken met de experimentele data. Figuur 7 De met DFT berekende moleculaire elektronenschillen van de HM en de LUM van het verbeterde schakelaar-donor systeem uit Hoofdstuk 4 (Figuur 5). In zijn geheel laat dit proefschrift zien dat het mogelijk is om met een systematische aanpak moleculen en systemen te ontwikkelen die in staat zijn om energieoverdracht te controleren. Verder onderzoek naar de mechanismes en efficiëntie van de hier ontwikkelde systemen zou een beter begrip kunnen opleveren van fundamentele processen die plaats vinden bij energieoverdracht en kan mogelijk leiden tot toepassingen in moleculaire machines of voor data opslag. 188