Download PDF
Synlett 2016; 27(03): 355-368
DOI: 10.1055/s-0035-1560725
account
© Georg Thieme Verlag Stuttgart · New York

Cycloadditions of Singlet Oxygen for Responsive Fluorescent Polymers

Samuel W. Thomas III*
Tufts University, Department of Chemistry, 62 Talbot Avenue, Medford, MA 02155, USA   Email: sam.thomas@tufts.edu
,
Esra Altinok
Tufts University, Department of Chemistry, 62 Talbot Avenue, Medford, MA 02155, USA   Email: sam.thomas@tufts.edu
,
Jingjing Zhang
Tufts University, Department of Chemistry, 62 Talbot Avenue, Medford, MA 02155, USA   Email: sam.thomas@tufts.edu
› Author Affiliations
Further Information

Publication History

Received: 30 August 2015

Accepted after revision: 21 September 2015

Publication Date:
09 December 2015 (online)

 
 PDF Download Permissions and Reprints All articles of this category


Abstract

This account describes progress in the author’s laboratory in the area of new fluorescent polymers that respond to the reactive oxygen species singlet oxygen (1O2). Key to the development of these materials are the [4+2] cycloaddition reactions between singlet oxygen and dienes such as acenes and furans. When covalently bound to conjugated polymer backbones, cycloadditions of these dienes with singlet oxygen can yield dramatic changes in the wavelength and intensity of luminescence: three such examples are given here. The account also summarizes our work to understand how changing the chemical structures of acenes affects reactivity with singlet oxygen as well as the cycloreversion of the resulting endoperoxides.

1 Introduction

2 Motivation

3 Organic Soluble Diene-Linked Conjugated Polymers

4 Red-Shifting Woes Lead to Reversibility

5 Building a Database

6 Integrating Other Acenes

7 Conclusion and Looking Forward


#

Key words

cycloadditions - singlet oxygen - fluorescence - polymers - acenes

Biographical Sketches

Zoom Image

Samuel W. Thomas III is an associate professor of chemistry at Tufts University, where his research focuses on new stimuli-responsive materials. He received a B.S. in chemistry from the University of Rochester in 2000, during which time he worked in the laboratories of Joshua Goodman (Rochester) and Louis Hegedus (Colorado State). After one year as a research chemist at Eastman Kodak, he earned his Ph.D. in 2006 under the supervision of Professor Timothy Swager at MIT focusing on amplifying fluorescent polymers. He then spent three years as a postdoctoral fellow with Professor George Whitesides at Harvard. Since starting at Tufts in 2009, his group’s research has been recognized with a DARPA Young Faculty Award (2009), a Thieme Publishers Journal Award (2010), an NSF CAREER award (2012), and a 3M Nontenured Faculty Award (2013). His teaching of undergraduate organic chemistry at Tufts University has also been recognized with a Tufts Teaching with Technology Award (2013).

Zoom Image

Esra Altinok was born in Izmir, Turkey, in 1987. She received her B.Sc. in chemistry from Middle East Technical University (Ankara, Turkey) in 2010. She then moved to Tufts University (Medford, MA), where she obtained her doctoral degree in August 2015 in the group of Professor Sam Thomas. During her time as a graduate student, her research mostly focused on the synthesis and photophysical characterization of singlet-oxygen-responsive small molecules and polymers. She is currently a research chemist at NBD Nanotechnologies in Boston.

Zoom Image

Jingjing Zhang is a postdoctoral research associate at University of Wisconsin-Madison, where his research focuses on organic open-shell materials. Jingjing graduated from Nanjing University, China, in 2007 with a B.S. in chemistry. He joined the group of Professor Samuel Thomas at Tufts University in 2009 and earned his Ph.D. in 2014, during which time his research focused on singlet-oxygen-responsive materials and substituted acenes.

1

Introduction

This account describes work in the author’s laboratory on the development of fluorescent materials that respond to the important reactive oxygen species (ROS) singlet oxygen (1O2). Fluorescent polymers, especially conjugated polymers (CPs), have received significant attention as high-performance, chemically sensitive fluorophores.[1] Much of this attention stems from the properties of amplification that these materials offer, resulting from light harvesting and exciton mobility characteristics.[2] In comparison to a more traditional, small molecule fluorophore–receptor pair, in which each binding event or reaction of analyte yields a change in a single chromophore, the harvesting of photons together with transport of excited states through, for example, energy transfer mechanisms, allows a single excited state to sample many potential reaction sites. The result is a funneling of excited states generated far from the site of reaction or binding to a low-energy trap. These processes of light harvesting and exciton mobility of CPs have proven useful in not only chemical and biological sensing, for which the sensing of nitroaromatic explosives through amplified photoinduced electron transfer quenching has been particularly successful,[3] [4] but are also the initial steps in the operation of bulk heterojunction organic photovoltaics.[5]

One of the rewarding aspects of working with singlet oxygen is the direct connection to lessons taught in introductory chemistry on molecular orbital theory and electronic structure. There are a number of sources with excellent details on the physical and chemical characteristics of 1O2;[6] [7] [8] [9] [10] [11] only brief descriptions are given here. Two low-energy electronic configurations with singlet multiplicity exist above the ground state triplet of O2, as shown in Figure [1]. The higher energy of the two, in which electrons with opposite spin occupy different π* antibonding orbitals with term symbol 1Σg, has a very short lifetime, and quickly undergoes internal conversion into the lowest energy electronic excited state with paired electrons with term symbol 1Δg. In the remainder of this paper, 1O2 refers to this lowest energy excited state of O2. The excited state energy of 1O2 is relatively low (~1 eV), giving weak phosphorescence at 1270 nm that is an important observable for monitoring excited-state processes of 1O2. The lifetime of 1O2 is also strongly dependent on solvent, ranging from 3 μs in water to 87 ms in carbon tetrachloride (CCl4). A related unique aspect of 1O2 quenching dynamics is the strong dependence of lifetime on solvent deuteration due to decreased coupling to bond vibrations of deuterated solvents:[12] as an example, 1O2 decays one order of magnitude more slowly in deuterium oxide (D2O) than in H2O.

Zoom Image
Figure 1 (a) Electronic configurations of the electronic ground state (3Σ) and the two lowest energy electronic excited states (1Δ and 1Σ) of O2. (b) General scheme for the preparation of 1O2 by photosensitization, and the structures of some generally useful and commercially available photosensitizers (TPP = tetraphenylporphyrin, MB = methylene blue, RB = rose bengal)

As a spin forbidden transition, the direct absorbance of photons to promote the conversion of 3O2 into 1O2 is not efficient. Given the low excited state energy of 1O2, however, its generation via energy transfer from other molecules can, in contrast, be highly efficient.[13] Although it is possible to transfer energy from singlet excited states to 3O2 to produce 1O2, most highly efficient sensitizers proceed through excited triplet states in a triplet annihilation process, as shown in Figure [1]: excitation of a sensitizing chromophore and intersystem crossing yields the triplet state of the sensitizer, which upon energy transfer to O2 regenerates the ground electronic state of the sensitizer and 1O2. Efficiencies of this process have been tabulated for a wide variety of sensitizers.[14] An important consequence of this efficient sensitization process and the ubiquity of O2 is that in those samples that are not deoxygenated, 1O2 is a commonly generated reactive oxygen species: 1O2 is a typical culprit for the photooxidative decomposition of materials.[15] [16] The stoichiometrically consumed reagents are photons, and O2 if the 1O2 goes on to react by some other pathway other than unimolecular or bimolecular physical quenching. This process does not consume the sensitizer, although sensitizer decomposition by other excited-state pathways can of course occur. Therefore, substoichiometric quantities of the sensitizers can be used to generate 1O2 and downstream products.

In addition to its interesting unimolecular behavior and method of production, 1O2 has a rich palette of organic reaction chemistry, which is usually broken into three classes: ene reactions to form hydroperoxides,[17] as well as [2+2] cycloaddition reactions with electron-rich alkenes[18] [19] and [4+2] cycloaddition reactions with dienes,[20] both of which yield peroxides. These reactions, especially [4+2] cycloadditions between 1O2 and furan moieties, have found great utility in the preparation of natural products.[21] For applications in organic electronics, however, this reaction is generally considered a nuisance. Pentacene, one of the most thoroughly studied organic semiconductors, is an excellent exemplar of this concept. Organic thin-film transistors fabricated using pentacene represent a benchmark in terms of performance, with hole mobility values among the highest reported.[22] What is often considered an unfavorable side effect of pentacene and many of its derivatives is its rapid rate of [4+2] cycloaddition with 1O2 to form the 6,13-endoperoxide (Scheme [1]), the bimolecular rate constant of which is approximately 4 × 108 M–1s–1.[23] Shorter acenes, such as tetracene or anthracene, are also highly reactive (k~105–107 M–1s–1),[12] giving the corresponding endoperoxides with high yields. Given the ease with which light, O2 and many chromophores (including acenes themselves) generate 1O2, this reactivity and the resulting destruction of the highly conjugated structures of acenes presents a challenge for their use in electronic devices. As a result, there exist a number of strategies relying on substituent effects to reduce the observed reactivity of acenes with 1O2 by up to several orders of magnitude.[24]

Zoom Image
Scheme 1 Endoperoxide formation through [4+2] cycloadditions of 1O2 and linear acenes

For our group, however, these characteristics: (i) large changes in optical and electronic properties, (ii) the use of reactants that are ubiquitous under ambient conditions, (iii) rapid and clean addition chemistry, and (iv) the chemical amplification derived from very small amounts of a sensitizer to produce large quantities of the reactive intermediate, make acene–1O2 cycloadditions attractive for stimuli-responsive materials. In fact, these types of [4+2] cycloaddition reactions, typically with anthracenes or isobenzofurans, are responsible for the most well-known molecules that show colorimetric or fluorescent responses to 1O2.[25] Fluorescent sensors or dosimeters for 1O2 are important due to its central nature in both a variety of biological processes, and as a key cytotoxic ROS in photodynamic therapy.[26] [27] In addition, due to the fact that the production of 1O2 is chemically amplified relative to the amount of sensitizer, reactions of 1O2 that produce a change in fluorescence can be used as a replacement for enzymatic catalysis of the formation of a luminescent product in bioassays, as was demonstrated in the bead-based luminescent oxygen-channeling immunoassay[28] [29] commercialized as the AlphaLISA platform.

At the time that we started this research in 2010, the molecule DMAX (9-[2-(3-carboxyl-9,10-dimethyl)anthryl]-6-hydroxy-3H-xanthen-3-one) from the group of Nagano was a benchmark fluorescent dosimeter for 1O2, comprising a 9,10-dialkylanthracene moiety bound to a xanthene dye.[30] Before cycloaddition with 1O2, the anthracene moiety efficiently quenches the xanthene luminescence through photoinduced electron transfer (Scheme [2]) from the anthracene highest occupied molecular orbital (HOMO) to the lower energy singly occupied orbital of the excited xanthene. Anthracene–1O2 cycloaddition, however, lowers the HOMO energy to such an extent as to render this photoinduced electron transfer endergonic, resulting in a ‘turn-on’ of fluo­rescence from the dye in response to 1O2. DMAX was an improvement over an earlier analog (DPAX = 9-[2-(3-carboxyl-9,10-diphenyl)anthryl]-6-hydroxy-3H-xanthen-3-one), with similar mechanism, that used 9,10-diphenylanthracene as the reactive quencher;[31] this anthracene derivative reacts with 1O2 more slowly than the dialkyl analog used in DMAX. This technology has been commercially successful as well, with an analog of DMAX currently available as ‘Singlet Oxygen Sensor Green’ (SOSG).[32]

Zoom Image
Scheme 2 Oxidation of DMAX with 1O2 to give its endoperoxide. This eliminates the driving force for fluorescence quenching by photoinduced electron transfer from the anthracene to the xanthene fluorophores rendering the process endergonic, and increases the observed fluorescence quantum yield from the xanthene

Concurrent with our work described below, several important improvements on the DMAX/SOSG design have been reported, including: (i) ‘Aarhus Sensor Green’ from the group of Ogilby, a DPAX analog that solves the problem of 1O2 sensitization by the endoperoxide product through fluorination of the fluorescein,[33] (ii) ‘Si-DMA’ from the group of Majima, which combines minimal self-oxidation and far-red emission,[34] and (iii) a series of ratiometric probes for 1O2 based on interruption of the conjugation of extended diarylisobenzofurans from the group of Nam.[35]


# 2

Motivation

Keeping in mind the advantages of CPs as amplifying fluorescent sensing materials, it seemed to us that two potential routes for innovation in 1O2-sensitive fluorophores would be, (i) to use a CP backbone instead of a simple fluorescent small molecule dye to enable light-harvesting and exciton mobility to improve performance in response to 1O2, and (ii) expand the structural space of 1O2-reactive acenes used in 1O2-responsive materials beyond anthracenes, as longer acenes generally tend to be more reactive with 1O2 than shorter acenes. In addition to CPs providing a pathway for improving 1O2-responsive materials, we found tantalizing the general challenge of making a CP that responds to a reactive oxygen species through increased fluorescence intensity, without degradation, as fluorescent conjugated polymers, especially the highly fluorescent 9,9-dialkylfluorene-based CPs, do have a reputation for being photooxidatively unstable.[36] Finally, as a side note, it is important to mention in this context that initial inspiration for this work derived from the PI’S unexpected observation as a graduate student in the laboratory of Professor Tim Swager, in which thin films of an anthracene-substituted poly-(phenylene-ethynylene) showed an increase in fluorescence intensity and a hypsochromic shift upon acquisition of emission spectra. Whether that exact observation was due to endoperoxide formation or some other photoreaction such as [4+4] anthracene dimerization remains unknown.


# 3

Organic Soluble Diene-Linked Conjugated Polymers

There are a number of strongly fluorescent polymer backbones from which to choose in designing a CP-based sensing material. Based on some previous experience with sensors for nitroalkanes, we chose a highly fluorescent ­poly(fluorene-alt-phenylene) conjugated polymer backbone.[37] The reasons for this choice were: (i) they are highly fluorescent both in solution and as solids, (ii) 2,7-diboronates of 9,9-dialkylfluorene are commercially available for use in Suzuki polymerization reactions, (iii) the HOMO–LUMO gaps of these CPs are large; the maximum wavelength of emission is ~420 nm, which maximized the probability that modification of the fluorescence of the polymer by electron transfer or energy transfer would be energetically favorable, and (iv) dialkoxyphenylene backbone units are readily functionalized through alkylation of the hydroquinones. We therefore first prepared polymer P2 as shown in Scheme [3].[38] In this design, the conjugated polymer backbone takes the place of the xanthene in DPAX. The synthesis of this polymer was quite straightforward, mainly because of the ease of preparing anthracene derivatives that are unsymmetrically substituted about the long axis. Simple bromination of commercial 9-phenylanthracene and Suzuki coupling with 4-methoxyphenylboronic acid were known procedures at the time, as was the synthesis of the alkylating agent 1. The assembly of M2 and its co-polymerization with a commercially available fluorene diboronate monomer occurred quickly.

Although it was quite easy to synthesize P2, the responsive characteristics were, predictably, poor. Before exposure to 1O2, the fluorescence spectrum of P2 matched almost perfectly with that of 9,10-dianisylanthracene, which indicated that energy transfer from the CP backbone to the anthracene was occurring. Upon exposure to 1O2 in an organic solvent, the change in the fluorescence spectrum was so small that with just a quick glance one could miss it. A slight blue shift in λmax of emission from 422 nm to 417 nm occurred, with almost no change in intensity. Another disadvantage was that readily perceptible responses required ≥20 minutes of irradiation of methylene blue. These spectra observed did indeed match those of the anthracene (before exposure) and the CP backbone (after exposure). We therefore concluded that energy transfer from the CP backbone to the pendant acene occurred before exposure to 1O2. After endoperoxidation of the acene pendant, however, the much larger HOMO–LUMO gap of the endo­peroxide makes this energy transfer (ET) unfavorable, resulting instead in luminescence from the polymer. Therefore, we realized that in order for an acene-linked CP to show a rapid, clear response to 1O2, we needed acene pendants that were both red-shifted from diarylanthracenes and more reactive.

Zoom Image
Scheme 3 Synthesis of acene-linked diiodophenylene monomers for cross-coupling polymerizations (M2 and M3) and the resulting anthracene-linked (P2) and tetracene-linked (P3) poly(fluorene-alt-phenylene)s. MEK = methyl ethyl ketone

The use of longer linear acenes was our solution to this problem, but posed a greater challenge for synthesis. Unlike anthracenes, for which there is a plethora of literature precedent and commercially available starting materials for the simple synthesis of elaborated, unsymmetrically substituted derivatives, there are far fewer easy options for unsymmetrically substituted tetracene derivatives. The lower solubility of tetracene derivatives and their more reactive nature with respect to photooxidation presented additional challenges. The obvious starting material to us was 5,12-tetracenequinone, but we struggled for a while to prepare a tetracene analog of M2 from it. Initial efforts focused on early introduction of an unsymmetrical substitution pattern by adding only one equivalent of 4-methoxyphenyllithium to the quinone, but all such attempts gave only very poor yields of the γ-hydroxyketone. The same was true at that time for attempting to deprotect only one of the methyl ethers with boron tribromide (BBr3), so we settled on isolating a singly alkylated product of the tetracene diol with 1, followed by methylation of the remaining phenol. Upon isolating M3 by column purification the first time, we did not have much material, perhaps only 20–30 mg, and the compound was contaminated with 10% of an unidentified impurity. The temptation to carry this into the polymerization to see if our idea would work was enormous, given the several months of effort it took to develop the synthetic conditions. The demanding nature of stoichiometry control of step-growth polymerization, however, required that we further purify this first batch of tetracene monomer through three painstakingly careful recrystallizations. In the end, we were able to polymerize about 15 mg of M3 with the fluorene co-monomer, which yielded P3 with a number average molecular weight of 13 kg/mol.

Zoom Image
Scheme 4 Ratiometric fluorescence response of tetracene-linked conjugated polymer P3 to photochemically generated 1O2 in CH2Cl2. FRET = Fluorescence resonance energy transfer. Reproduced in part from Ref. 38 with permission from the Royal Society of Chemistry
Zoom Image
Scheme 5 Example syntheses of trifluoromethyl-substituted diarylfuran-linked poly(fluorene-alt-phenylene)s with one furan per repeating unit (P4) or two (P5)

Data of the type in Scheme [4], collected using the newly characterized P3, are among the most gratifying ever recorded in our lab, particularly because it was the very early days of the lab’s existence. It was clear that the diaryltetracene allowed for a fast, clear ratiometric fluorescence response to 1O2 due to acene–1O2 cycloaddition interrupting energy transfer from the CP backbone to the tetracene pendants. We submitted the paper after a flurry of confirmation, control, and characterization experiments that lasted about a month or two.

Not long after, we investigated a related series of diene-linked CPs for 1O2 responsiveness, this time using 2,5-diarylfurans,[39] endoperoxides derived from which are not stable at room temperature and fragment into several electron-poor products.[40] Inspired by the mechanism through which nitroaromatics quench CPs,[4] we predicted that the electron-poor products would be good fluorescence quenchers by accepting excited electrons from the CP backbone. We used the same type of poly(fluorene-alt-phenylene) backbone, but replaced the acene with a diarylfuran as shown in Scheme [5]. Introducing unsymmetric substitution on the diarylfuran is easier than on tetracenes, but the iodofurans are highly unstable and decompose completely within a couple of days. This instability required us to perform cross-coupling reactions with them immediately upon isolation. It is also worth mentioning the alkylation of the phenol-substituted diarylfurans with 1,4-dibromobutane, which is necessary to link the furan pendant to the poly­merizable diiodoarene. Our initial efforts to link these two units with a butyl chain were exactly analogous to the acene pendants, but reactions of phenol-substituted furans with 1 gave yields so low that it was difficult to prepare the actual CP, which was extremely frustrating especially as it was so late in the synthesis. For reasons we do not understand, the approach shown in Scheme [5], that just swaps which unit we alkylate first with 1,4-dibromobutane, gave good yields (58–93%) for each of the two alkylation steps.

These polymers behaved as predicted upon exposure to 1O2 generated photochemically, showing strong fluorescence quenching due to oxidation of furan pendants (Scheme [6]). There was a trade-off in terms of substituent effects, with increasingly electron-poor furans showing more efficient quenching upon oxidation, due to the products being better electron acceptors, but slower rates of reaction with 1O2, as one expects from typical Diels–Alder reactions. Polymers with two furans per repeat unit showed more fluorescence quenching upon complete furan oxidation (P5 showed 93% quenching). Key to these materials was that the rates of reaction between 1O2 and the furans were so fast as to preclude decomposition of the CP backbone from contributing to the observed quenching.

Zoom Image
Scheme 6 Top: Endoperoxides of diarylfurans yield various electron-poor quenchers. Bottom: Fluorescence quenching of P5 due to oxidation of furan pendants with 1O2. Adapted with permission from Ref. 39. Copyright 2013 American Chemical Society

# 4

Red-Shifting Woes Lead to Reversibility

Having investigated two of the most popular classes of dienes for cycloaddition with 1O2, we decided to focus on acenes because of the ratiometric responses we observed and the large degree to which substituents can tune their luminescence and reactivities. Our efforts then turned to the preparation of CPs that incorporate acenes that were (i) more red-shifted, and (ii) more reactive than the diaryltetracene used in P3. Although the obvious candidates were rubrene (5,6,11,12-tetraphenyltetracene) and 6,13-diarylpentacene derivatives, we have been unable to prepare CPs that contain these moieties as 1O2-reactive pendants. Contributing to this has been the relative insolubility and reactivity of the 6,13-diarylpentacene derivatives. Beyond these types of structures, however, not much had been reported quantitatively on the reactivity of highly red-shifted acenes, which we attribute in part to the requirement for electronics applications to prevent this reaction. Therefore, an important phase of our research at this time turned to understanding what other sorts of novel acene structures could fulfill our goals of red-shifted emission, faster reaction with 1O2, or both.

Zoom Image
Scheme 7 Top: Synthesis of tetrasubstituted tetracenes with both phenyl and ethynyl substituents for increased conjugation. Bottom: Photographs of fluorescence from solutions of three substituted tetracene derivatives in solution under illumination with UV light, showing increased red-shifted emission from the tetrasubstituted derivative. Adapted from Ref. 41 with permission from the Royal Society of Chemistry

Our initial targets were related to rubrene, but instead substituted two alkynes for two of the phenyl substituents for increased conjugation and easy synthetic manipulation (Scheme [7]).[41] Access to these novel acene structures was possible through double addition of alkynyllithium reagents to diphenyltetracenequinone, followed by reduction of the diol intermediate with tin(II) chloride (SnCl2) in aqueous hydrochloric acid (HCl). In line with our design principles, these tetrasubstituted tetracene derivatives had HOMO–LUMO gaps that were significantly lower than those of the diaryltetracene from P3; in fact, their onset of absorbance and fluorescence spectra matched almost exactly those of 6,13-dianisylpentacene. Scheme [7] shows visual comparisons of the fluorescence of disubstituted and tetrasubstituted tetracenes. From the point of view of their reactivity with 1O2, however, they were actually less reactive than 5,12-dianisyltetracene (by 2.5–5 fold). It is indeed well known that the reactivity of alkyne-substituted acenes is suppressed relative to unsubstituted analogs; Fudikar and Linker have elucidated many of the reasons for this, which turn out to depend on the length of the acene.[23] Another approach to accessing increasingly red-shifted tetracenes would be substitution with methoxy groups, as reported by Anthony and co-workers.[42]

As a part of this work we had a goal of separating and characterizing the two endoperoxides that could be formed upon oxidation with 1O2. In previous experiments on dianisyltetracene, we were able to do so easily by flash chromatography (see Scheme [8]). In these cases, however, we were unable to do so: although, based on NMR and TLC analysis of the crude oxidation mixture, there was clearly a mixture of endoperoxides formed, we were only able to isolate one by chromatography, along with significant amounts of the starting acene, which formed on the silica gel with a distinctive red color. We then noticed that upon standing neat, in the dark, the mixture of endoperoxides from the tetrasubstituted tetracenes turned red over the span of several days and showed resonances in the NMR spectrum attributed to the starting acene. By studying the evolution of the mixture of endoperoxides from bis(trimethylsilyl­ethynyl)tetracene, we found that the regioisomer oxidized at the ethynylated ring cycloreverts into the starting acene at room temperature, while the regioisomer at the unsubstituted ring did not, suggesting that ethynyl substitution lowers the barrier for cycloreversion of endoperoxides. Around the same time, Fudikar and Linker reported that diethynylanthracene derivatives cyclorevert rapidly at room temperature (much like the endoperoxide of 1,4-dimethylnaphthalene), attributing the behavior to increased stability of propargyl radical intermediates.[43] Clearly, then, ethynylated acene–endoperoxides show promise as substrates for reversible photoreactions. Although we have yet to capitalize on this discovery for 1O2-responsive polymers and other materials, it opens possibilities to move beyond dosimeters to dynamically responsive, reversible sensors based on 1O2-acene cycloaddition chemistry.

Zoom Image
Scheme 8 The regioisomeric mixture of endoperoxides from 5,12-diaryltetracenes is easily separable, while the endoperoxide formed on the ethynylated positions of diethynyltetracenes cycloreverts to the acene at room temperature

# 5

Building a Database

Our group’s efforts in rational tuning of properties of a tetracene derivative by extending conjugation through ethynyl substitution opened our eyes to the challenges inherent in trying to improve the properties of a molecule along multiple axes simultaneously. In the case of the 5,12-diethynyl-6,11-diphenyltetracenes, the desired characteristic of red-shifted emission also yielded a decrease in reactivity, which is unfavorable for maximum sensitivity to 1O2. Therefore, our strategy shifted from the design of individual acene cores to a systematic study of how variations in key structural elements affect the properties important to our applications: absorbance and fluorescence properties, frontier molecular orbital energies, and reactivity with 1O2. Much of this information has been reported for a large number of anthracene derivatives,[12] [44] a variety of symmetrically substituted pentacene derivatives,[45,46] and diethynyl anthradithiophenes.[45]

We decided to examine how these properties depended on two common structural elements of elaborated acenes.[47] One was whether substitution directly on the acene core was diaryl, diethynyl, or a mixture of the two; in general, we expected that unsymmetric substitution of acenes with one aryl and one ethynyl group, examples of which are rare for acenes longer than three rings, would yield spectral properties and reactivities intermediate between corresponding diaryl and diethynyl derivatives. The other was whether a conjugated acene core was extended by either a fused thiophene or benzene ring; in this case, we expected that the smaller resonance energy of thiophene compared to benzene would result in thienoacenes having properties intermediate between an unextended acene and the corresponding acenes extended with a benzene ring.

Such a line of inquiry therefore required the synthesis of a significant number of compounds, several of which constituted the first examples of their subclass of substituted acenes and thienoacenes. The basic strategy for synthesis was highly conserved (see Scheme [9] for an example with ‘TMT’ derivatives – Tetracene-MonoThiophene). All compounds were prepared using nu­cleophilic addition of the appropriate organolithium reagents to quinones, followed by reduction. In these syntheses, we faced two significant challenges. The first was a general approach to preparing unsymmetrically substituted aryl-ethynylacenes; our previous attempts (vide supra) to prepare unsymmetric diaryl acenes from quinones failed. The work of Tykwinski and co-workers, in which monoadducts of 6,13-pentacenequinone and alkynyllithium reagents were prepared in high yield as intermediates in the synthesis of unsymmetric diethynylpentacenes, gave us an easily implemented and practical method for preparing such compounds.[48] [49] We found that this approach was both (i) generally useful for preparing monoethynyl adducts for all the quinones we tested (anthracene, tetracene, pentacene, thienoanthracene, thienotetracene, and anthradithiophene), and (ii) useful for preparing unsymmetrically substituted arylethynyl derivatives by exposing the intermediate γ-hydroxyketones to excess aryllithium salt.

Zoom Image
Scheme 9 Example syntheses of diethynyl, aryl-ethynyl, and diaryl derivatives of the tetracenothiophene backbone. Reproduced with permission from Ref. 47. Copyright 2014 American Chemical Society

Our second challenge involved the somewhat labile C–H bond on the 2-position of terminal fused thiophene rings. Most previous reports of substituted derivatives of the heteroacene cores under investigation here utilized ethynyl substituents on a central ring, installation of which required alkynyllithium salts that are not sufficiently basic to deprotonate the thiophene moieties. Analogous efforts in our lab to prepare aryl-substituted derivatives of, for example, anthradithiophene, yielded only poor yields of butylated products, presumably through deprotonation of terminal thiophene rings and subsequent alkylation with the 1-bromobutane side-product of halogen–metal exchange. We solved this problem through methylation of the 2-position of all terminal thiophene rings early in the synthesis of the quinones,[50] resulting in clean thienoacenes in acceptable yields for our purposes (30–50%). Although substitution with other groups such as fluorine atoms or nitriles has also been reported to yield highly fluorescent thienoacene derivatives,[51] we chose alkylation to prevent the possibility of electron-withdrawing substituents slowing the rate of oxidation by 1O2. This approach of installing alkyl groups on terminal thiophene rings of these heteroacenes also brings the additional possibility of increasing solubility of these derivatives and their synthetic intermediates, which can be a limiting factor in the practical synthesis of large, elaborated acenes.

Trends in the key properties of these acenes for 1O2-responsive fluorophores (optical properties and reaction rates) generally adhered to what we predicted. Lengthening of acenes through addition of a thiophene ring yielded red-shifted spectra, mostly through an increase in HOMO energies. This shift was due to addition of a benzene ring, which both increased HOMO energies and decreased LUMO energies. In addition, increasing the conjugation of these acenes by substituting ethynyl groups for aryl groups also yielded red-shifted spectra, primarily through a lowering of the LUMO energies. In terms of absorbance and fluorescence, these compounds together cover nearly the entire visible spectrum from violet to red emission (see Figure [2]). The rates of reactivity with 1O2 also corresponded to our expectations, with longer acenes generally reacting faster than shorter acenes, and addition of fused thiophenes contributing smaller increases in rate than addition of fused benzenes. Double ethynyl substitution also lowers the rate of reaction of an acene core with 1O2 (by approximately 3–10 fold), with single ethynyl substitution yielding a smaller decrease in rate. The most obvious exception to this is diethynylpentacenes, which show a ~103 decrease in reactivity, which Linker has discovered is due to a competitive quenching of 1O2.[23] Our calculation of the triplet energy (0.74 eV) of this pentacene suggests that this may be due to energy transfer from 1O2 to the pentacene, as is known to occur for β-carotene. Finally, this work uncovered that the diaryl- and aryl-ethynyl derivatives of tetracenothiophene react with 1O2 with observed rate constants and spectral positions close to those of 6,13-diarylpentacene, but with much higher fluorescence quantum yields (0.4–0.5 vs 0.1), highlighting these and related structures for potential use in 1O2-responsive fluorescent materials.

Zoom Image
Figure 2 Trends of emission colors, visualized using photographs of acene solutions under UV illumination, together with relative observed rates of oxidation for each acene. Adapted with permission from Ref. 47. Copyright 2014 American Chemical Society

# 6

Integrating Other Acenes

Having quantitative reactivity data in hand for a broad range of substituted acenes, our goal has been to integrate 1O2-reactive acenes into polymers in new ways. One recently reported effort centered on linking an acene to a CP backbone through a conjugated linker, as such linkers have been reported to increase energy transfer rates in a number of examples of ‘through bond’ energy transfer cassettes.[52] We therefore targeted the poly(phenylene ethynylene) shown in Scheme [10], which uses a biphenyl-bridged oligo(phenylene-ethynylene) as a conjugated linker between the CP backbone and the pendant diethynyltetracenes.[53] We chose tetracenes in this case as they react faster with 1O2 than any other of the diethynylacenes we examined above. The first key intermediate in this synthesis is the 3,4,5-tri(2-ethylhexyl)ethynylbenzene; the long, branched alkyl chains are required not only for solubility of the final CP, but also for solubility of the monomer and other extended, rigid intermediates. Analogous routes that did not include these solubilizing chains resulted in low yields and difficult purifications because of the insolubility of many of the reaction products. Again, critical to our success was the selective mono-addition of alkynyllithium salts to tetracenequinone, which was followed by excess lithiated 4-ethynyl-4′-trimethylsilylethynylbiphenyl, SnCl2–HCl reduction, and deprotection of the trimethylsilyl (TMS) group to yield a monomer precursor. Finally, Sonogashira cross-coupling with 2,5-bis(trimethylsilylethynyl)iodobenzene and TMS deprotection yielded the dialkyne monomer M6 suitable for Sonogashira polymerizations. When conducted at ambient temperature with the 1,4-di(2-ethylhexyl)-2,5-diiodobenzene co-monomer, these reactions worked well, yielding polymers with M n values of 20–70 kDa. At 65 °C, however, these reactions yielded only insoluble, highly colored gelatinous precipitates that were insoluble in every organic solvent investigated.

Zoom Image
Scheme 10 Synthesis of two-dimensional, 1O2-responsive conjugated polymer P6

The 50–50 alternating copolymer (P6) showed what was, for us at the time, peculiar behavior upon exposure to 1O2. The UV/Vis spectrum behaved as one would expect, with decreasing absorbance of bands attributed to the extended tetracene moiety. The fluorescence spectrum, on the other hand, showed not a decrease in fluorescence intensity from the tetracene at 580 nm, but an increase, even though the concentration of tetracene was decreasing with increasing exposure time to 1O2 (see Figure [3]). This increase continued until approximately 75% conversion of tetracene pendants into endoperoxides, at which time it decreased and fluorescence from the polymer backbone started to emerge from the baseline at 470 nm. An observed increase in average fluorescence lifetime of P6 with increasing endoperoxide formation indicated that dynamic self-quenching of the tetracenes occurs with tetracenes on every repeating unit, while decreasing the local tetracene concentration with endoperoxide formation inhibits this self-quenching. This interpretation explains the unexpected fluorescence response of P6 to 1O2. Finally, we conducted a random copolymerization with M6 diluted four-fold with a 1O2-inert dialkyne monomer so that the overall loading of the tetracene pendant was decreased (by a factor of 4) relative to the 1:1 copolymer P6. In an illustration of the amplifying nature of CPs, this polymer (P6-25) with an average of one tetracene for every eight conjugated aromatic rings along the polymer backbone showed nearly complete energy transfer before exposure to 1O2, and a faster ratiometric fluorescent response to 1O2 than P6, as there was not a large excess of tetracenes that had to be oxidized before fluorescence from the CP backbone could compete with energy transfer.


# 7

Conclusion and Looking Forward

Our group has created a general construct comprising conjugated polymers with 1O2-reactive pendants. We have demonstrated that such materials can show ratiometric fluorescent responses to 1O2, and that such responses can be amplified relative to analogous small molecules. Through fundamental studies of the structure–property relationships of acenes and heteroacenes, we have uncovered structural elements that enable reversibility of 1O2-mediated oxidation, created a library of substituted acene derivatives that span a range of luminescence colors and reactivities, and developed guidelines for predicting the effects of additional fused rings and different substituents on frontier molecular orbital energies. We are currently combining these lines of inquiry into next-generation fluorescent materials with the goals of increased sensitivity to photosensitized 1O2 through incorporation of maximally reactive (hetero)acenes and improvement of energy transfer efficiencies. We are also developing 1O2-responsive materials that work in aqueous environments, building upon previous success in CP/acene thin-film composites that respond ratiometrically to 1O2 generated using sensitizers bound to proteins.

As have our achievements to this point, our future success will rely upon the toolbox of synthesis to control molecular structure. Key design elements such as controlling the electronic coupling between donors and acceptors, developing 1O2-reactive units that can also be integrated into polymer structures in new ways, and enhancing performance in water will require the efficient syntheses of new molecules and materials that combine these individual design elements. As practitioners of synthetic chemistry in pursuit of novel functional materials, the ever-increasing scope of structures accessible with modern synthetic methods is both tantalizing and overwhelming, which results in feeling like the proverbial “kid in a candy store”, with our imaginations the largest hurdle to new discoveries.

Zoom Image
Figure 3 Due to the amplifying properties of CPs, the ratiometric response to 1O2 of P6 accelerates when the tetracene content is reduced four-fold. Adapted with permission from Ref. 53. Copyright 2015 American Chemical Society

#
#

Acknowledgment

This work was supported by the National Science Foundation (CHE-1305832).



   Permissions and Reprints All articles of this category
Zoom Image
Zoom Image
Zoom Image
Zoom Image
Figure 1 (a) Electronic configurations of the electronic ground state (3Σ) and the two lowest energy electronic excited states (1Δ and 1Σ) of O2. (b) General scheme for the preparation of 1O2 by photosensitization, and the structures of some generally useful and commercially available photosensitizers (TPP = tetraphenylporphyrin, MB = methylene blue, RB = rose bengal)
Zoom Image
Scheme 1 Endoperoxide formation through [4+2] cycloadditions of 1O2 and linear acenes
Zoom Image
Scheme 2 Oxidation of DMAX with 1O2 to give its endoperoxide. This eliminates the driving force for fluorescence quenching by photoinduced electron transfer from the anthracene to the xanthene fluorophores rendering the process endergonic, and increases the observed fluorescence quantum yield from the xanthene
Zoom Image
Scheme 3 Synthesis of acene-linked diiodophenylene monomers for cross-coupling polymerizations (M2 and M3) and the resulting anthracene-linked (P2) and tetracene-linked (P3) poly(fluorene-alt-phenylene)s. MEK = methyl ethyl ketone
Zoom Image
Scheme 4 Ratiometric fluorescence response of tetracene-linked conjugated polymer P3 to photochemically generated 1O2 in CH2Cl2. FRET = Fluorescence resonance energy transfer. Reproduced in part from Ref. 38 with permission from the Royal Society of Chemistry
Zoom Image
Scheme 5 Example syntheses of trifluoromethyl-substituted diarylfuran-linked poly(fluorene-alt-phenylene)s with one furan per repeating unit (P4) or two (P5)
Zoom Image
Scheme 6 Top: Endoperoxides of diarylfurans yield various electron-poor quenchers. Bottom: Fluorescence quenching of P5 due to oxidation of furan pendants with 1O2. Adapted with permission from Ref. 39. Copyright 2013 American Chemical Society
Zoom Image
Scheme 7 Top: Synthesis of tetrasubstituted tetracenes with both phenyl and ethynyl substituents for increased conjugation. Bottom: Photographs of fluorescence from solutions of three substituted tetracene derivatives in solution under illumination with UV light, showing increased red-shifted emission from the tetrasubstituted derivative. Adapted from Ref. 41 with permission from the Royal Society of Chemistry
Zoom Image
Scheme 8 The regioisomeric mixture of endoperoxides from 5,12-diaryltetracenes is easily separable, while the endoperoxide formed on the ethynylated positions of diethynyltetracenes cycloreverts to the acene at room temperature
Zoom Image
Scheme 9 Example syntheses of diethynyl, aryl-ethynyl, and diaryl derivatives of the tetracenothiophene backbone. Reproduced with permission from Ref. 47. Copyright 2014 American Chemical Society
Zoom Image
Figure 2 Trends of emission colors, visualized using photographs of acene solutions under UV illumination, together with relative observed rates of oxidation for each acene. Adapted with permission from Ref. 47. Copyright 2014 American Chemical Society
Zoom Image
Scheme 10 Synthesis of two-dimensional, 1O2-responsive conjugated polymer P6
Zoom Image
Figure 3 Due to the amplifying properties of CPs, the ratiometric response to 1O2 of P6 accelerates when the tetracene content is reduced four-fold. Adapted with permission from Ref. 53. Copyright 2015 American Chemical Society