Polythiophenes (PTs) result from the polymerization of thiophenes, a sulfur heterocycle, that can become conducting when electrons are added or removed from the conjugated π-orbitals via doping.
The study of polythiophenes has intensified over the last three decades. The maturation of the field of conducting polymers was confirmed by the awarding of the 2000 Nobel Prize in Chemistry to Alan Heeger, Alan MacDiarmid, and Hideki Shirakawa “for the discovery and development of conductive polymers." The most notable property of these materials, electrical conductivity, results from the delocalization of electrons along the polymer backbone – hence the term “synthetic metals”. However, conductivity is not the only interesting property resulting from electron delocalization. The optical properties of these materials respond to environmental stimuli, with dramatic color shifts in response to changes in solvent, temperature, applied potential, and binding to other molecules. Both color changes and conductivity changes are induced by the same mechanism—twisting of the polymer backbone, disrupting conjugation—making conjugated polymers attractive as sensors that can provide a range of optical and electronic responses.
A number of comprehensive reviews have been published on PTs, the earliest dating from 1981.[1] Schopf and Koßmehl published a comprehensive review of the literature published between 1990 and 1994.[2] Roncali surveyed electrochemical synthesis in 1992,[3] and the electronic properties of substituted PTs in 1997.[4] McCullough’s 1998 review focussed on chemical synthesis of conducting PTs.[5] A general review of conjugated polymers from the 1990s was conducted by Reddinger and Reynolds in 1999.[6] Finally, Swager et al. examined conjugated-polymer-based chemical sensors in 2000.[7] These reviews are an excellent guide to the highlights of the primary PT literature from the last two decades.
Contents |
Electrons are delocalized along the conjugated backbones of conducting polymers, usually through overlap of π-orbitals, resulting in an extended π-system with a filled valence band. By removing electrons from the π-system (“p-doping”), or adding electrons into the π-system (“n-doping”), a charged unit called a bipolaron is formed (see Figure 1). Doping is performed at much higher levels (20–40%) in conducting polymers than in semiconductors (<1%). The bipolaron moves as a unit up and down the polymer chain, and is responsible for the macroscopically observed conductivity of the polymer. For some samples of poly(3-dodecylthiophene) doped with iodine, the conductivity can approach 1000 S/cm.[8] (In comparison, the conductivity of copper is approximately 5×105 S/cm.) Generally, the conductivity of PTs is lower than 1000 S/cm, but high conductivity is not necessary for many applications of conducting polymers (see below for examples).
Simultaneous oxidation of the conducting polymer and introduction of counterions, p-doping, can be accomplished electrochemically or chemically. During the electrochemical synthesis of a PT, counterions dissolved in the solvent can associate with the polymer as it is deposited onto the electrode in its oxidized form. By doping the polymer as it is synthesized, a thick film can build up on an electrode—the polymer conducts electrons from the substrate to the surface of the film. Alternatively, a neutral conducting polymer film or solution can be doped post-synthesis.
Reduction of the conducting polymer, n-doping, is much less common than p-doping. An early study of electrochemical n-doping of poly(bithiophene) found that the n-doping levels are less than those of p-doping, the n-doping cycles were less efficient, the number of cycles required to reach maximum doping was higher, and the n-doping process appeared to be kinetically limited, possibly due to counterion diffusion in the polymer.[9]
A variety of reagents have been used to dope PTs. Iodine and bromine produce high conductivities[8] but are unstable and slowly evaporate from the material.[10] Organic acids, including trifluoroacetic acid, propionic acid, and sulfonic acids produce PTs with lower conductivities than iodine, but with higher environmental stabilities.[10][11] Oxidative polymerization with ferric chloride can result in doping by residual catalyst,[12] although matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) studies have shown that poly(3-hexylthiophene)s are also partially halogenated by the residual oxidizing agent.[13] Poly(3-octylthiophene) dissolved in toluene can be doped by solutions of ferric chloride hexahydrate dissolved in acetonitrile, and can be cast into films with conductivities reaching 1 S/cm.[14] Other, less common p-dopants include gold trichloride[15] and trifluoromethanesulfonic acid.[16]
The extended π-systems of conjugated PTs produce some of the most interesting properties of these materials—their optical properties. As an approximation, the conjugated backbone can be considered as a real-world example of the “electron-in-a-box” solution to the Schrödinger equation; however, the development of refined models to accurately predict absorption and fluorescence spectra of well-defined oligo(thiophene) systems is ongoing.[17] Conjugation relies upon overlap of the π-orbitals of the aromatic rings, which, in turn, requires the thiophene rings to be coplanar (see Figure 2, top). The number of coplanar rings determines the conjugation length—the longer the conjugation length, the lower the separation between adjacent energy levels, and the longer the absorption wavelength. Deviation from coplanarity may be permanent, resulting from mislinkages during synthesis or especially bulky side chains; or temporary, resulting from changes in the environment or binding. This twist in the backbone reduces the conjugation length (see Figure 2, bottom), and the separation between energy levels is increased. This results in a shorter absorption wavelength.
Determining the maximum effective conjugation length requires the synthesis of regioregular PTs of defined length. The absorption band in the visible region is increasingly red-shifted as the conjugation length increases, and the maximum effective conjugation length is calculated as the saturation point of the red-shift. Early studies by ten Hoeve et al. estimated that the effective conjugation extended over 11 repeat units,[18] while later studies increased this estimate to 20 units.[19] More recently, Otsubo et al. synthesized 48-[20] and 96-mer[21] oligothiophenes, and found that the red-shift, while small (a difference of 0.1 nm between the 72- and the 96-mer), does not saturate, meaning that the effective conjugation length may be even longer than 96 units.[21]
A variety of environmental factors can cause the conjugated backbone to twist, reducing the conjugation length and causing an absorption band shift, including solvent, temperature, application of an electric field, and dissolved ions. The absorption band of poly (3-thiophene acetic acid) in aqueous solutions of poly(vinyl alcohol) (PVA) shifts from 480 nm at pH 7 to 415 nm at pH 4. This is attributed to formation of a compact coil structure which can form hydrogen bonds with PVA upon partial deprotonation of the acetic acid group.[22] Chiral PTs showed no induced circular dichroism (ICD) in chloroform, but displayed intense, but opposite, ICDs in chloroform–acetonitrile mixtures versus chloroform–acetone mixtures.[23] Also, a PT with a chiral amino acid side chain[24] displayed moderate absorption band shifts and ICDs, depending upon the pH and the concentration of buffer.[25]
Shifts in PT absorption bands due to changes in temperature result from a conformational transition from a coplanar, rodlike structure at lower temperatures to a nonplanar, coiled structure at elevated temperatures. For example, poly(3-(octyloxy)-4-methylthiophene) undergoes a color change from red–violet at 25 °C to pale yellow at 150 °C. An isosbestic point (a point where the absorbance curves at all temperatures overlap) indicates coexistence between two phases, which may exist on the same chain or on different chains.[26] Not all thermochromic PTs exhibit an isosbestic point: highly regioregular poly(3-alkylthiophene)s (PATs) show a continuous blue-shift with increasing temperature if the side chains are short enough so that they do not melt and interconvert between crystalline and disordered phases at low temperatures.
Finally, PTs can exhibit absorption shifts due to application of electric potentials (electrochromism),[27] or to introduction of alkali ions (ionochromism).[28] These effects will be discussed in the context of applications of PTs below.
The asymmetry of 3-substituted thiophenes results in three possible couplings when two monomers are linked between the 2- and the 5-positions. These couplings are:
These three diads can be combined into four distinct triads, shown in Figure 3. The triads are distinguishable by NMR spectroscopy, and the degree of regioregularity can be estimated by integration.[29][30]
Elsenbaumer et al. first noticed the effect of regioregularity on the properties of PTs. A regiorandom copolymer of 3-methylthiophene and 3-butylthiophene possessed a conductivity of 50 S/cm, while a more regioregular copolymer with a 2:1 ratio of HT to HH couplings had a higher conductivity of 140 S/cm.[31] Films of regioregular poly(3-(4-octylphenyl)thiophene) (POPT) with greater than 94% HT content possessed conductivities of 4 S/cm, compared with 0.4 S/cm for regioirregular POPT.[32] PATs prepared using Rieke zinc formed “crystalline, flexible, and bronze-colored films with a metallic luster." On the other hand, the corresponding regiorandom polymers produced “amorphous and orange-colored films.”[33] Comparison of the thermochromic properties of the Rieke PATs showed that, while the regioregular polymers showed strong thermochromic effects, the absorbance spectra of the regioirregular polymers did not change significantly at elevated temperatures. This was likely due to the formation of only weak and localized conformational defects. Finally, Xu and Holdcroft demonstrated that the fluorescence absorption and emission maxima of poly(3-hexylthiophene)s occur at increasingly lower wavelengths (higher energy) with increasing HH dyad content. The difference between absorption and emission maxima, the Stokes shift, also increases with HH dyad content, which they attributed to greater relief from conformational strain in the first excited state.[34]
Unsubstituted PTs are conductive after doping, and have excellent environmental stability compared with some other conducting polymers such as polyacetylene, but are intractable and soluble only in solutions like mixtures of arsenic trifluoride and arsenic pentafluoride.[35] However, in 1987 examples of organic-soluble PTs were reported. Elsenbaumer et al., using a nickel-catalyzed Grignard cross-coupling, synthesized two soluble PTs, poly(3-butylthiophene) and poly(3-methylthiophene-'co'-3’-octylthiophene), which could be cast into films and doped with iodine to reach conductivities of 4 to 6 S/cm.[36] Hotta et al. synthesized poly(3-butylthiophene) and poly(3-hexylthiophene) electrochemically[37] (and later chemically[38]), and characterized the polymers in solution[39] and cast into films.[40] The soluble PATs demonstrated both thermochromism and solvatochromism (see above) in chloroform and 2,5-dimethyltetrahydrofuran.[41]
Also in 1987, Wudl et al. reported the syntheses of water-soluble sodium poly(3-thiophenealkanesulfonate)s.[42] In addition to conferring water solubility, the pendant sulfonate groups act as counterions, producing self-doped conducting polymers. Substituted PTs with tethered carboxylic acids,[43] acetic acids,[44] amino acids,[24] and urethanes[45] are also water-soluble.
More recently, poly(3-(perfluorooctyl)thiophene)s soluble in supercritical carbon dioxide[46] were electrochemically and chemically synthesized by Collard et al.[47] Finally, unsubstituted oligothiophenes capped at both ends with thermally-labile alkyl esters were cast as films from solution, and then heated to remove the solublizing end groups. Atomic force microscopy (AFM) images showed a significant increase in long-range order after heating.[48]
PTs can be synthesized electrochemically, by applying a potential across a solution of the monomer to be polymerized, or chemically, using oxidants or cross-coupling catalysts. Both methods have their advantages and disadvantages.
In an electrochemical polymerization, a potential is applied across a solution containing thiophene and an electrolyte, producing a conductive PT film on the anode. Electrochemical polymerization is convenient, since the polymer does not need to be isolated and purified, but it can produce polymers with undesirable alpha-beta linkages and varying degrees of regioregularity.
As shown in Figure 4, oxidation of a monomer produces a radical cation, which can then couple with a second radical cation to form a dication dimer, or with another monomer to produce a radical cation dimer. A number of techniques including in situ video microscopy,[49] cyclic spectrovoltammetry,[50] photocurrent spectroscopy,[51] and electrochemical quartz crystal microbalance measurements,[52] have been used to elucidate the nucleation and growth mechanism leading to deposition of polymer onto the anode. Deposition of long, well-ordered chains onto the electrode surface is followed by growth of either long, flexible chains, or shorter, more crosslinked chains, depending upon the polymerization conditions.
The quality of an electrochemically prepared PT film is affected by a number of factors. These include the electrode material, current density, temperature, solvent, electrolyte, presence of water, and monomer concentration.[2] Two other important but interrelated factors are the structure of the monomer and the applied potential. The potential required to oxidize the monomer depends upon the electron density in the thiophene ring π-system. Electron-donating groups lower the oxidation potential, while electron-withdrawing groups increase the oxidation potential. Thus, 3-methylthiophene polymerizes in acetonitrile and tetrabutylammonium tetrafluoroborate at a potential of about 1.5 V vs. SCE (saturated calomel electrode), while unsubstituted thiophene polymerizes at about 1.7 V vs. SCE. Steric hindrance resulting from branching at the α-carbon of a 3-substituted thiophene inhibits polymerization.[53] This observation leads to the so-called “polythiophene paradox”: the oxidation potential of many thiophene monomers is higher than the oxidation potential of the resulting polymer. In other words, the polymer can be irreversibly oxidized and decompose at a rate comparable to the polymerization of the corresponding monomer. This remains one of the major disadvantages of electrochemical polymerization, and limits its application for many thiophene monomers with complex side groups.
Chemical synthesis offers two advantages compared with electrochemical synthesis of PTs: a greater selection of monomers, and, using the proper catalysts, the ability to synthesize perfectly regioregular substituted PTs. While PTs may have been chemically synthesized by accident more than a century ago,[54] the first planned chemical syntheses using metal-catalyzed polymerization of 2,5-dibromothiophene were reported by two groups independently in 1980. Yamamoto et al. used magnesium in tetrahydrofuran (THF) and nickel(bipyridine) dichloride, analogous to the Kumada coupling of Grignard reagents to aryl halides.[55] Lin and Dudek also used magnesium in THF, but with a series of acetylacetonate catalysts (Pd(acac)2, Ni(acac)2, Co(acac)2, and Fe(acac)3).[56]
Later developments produced higher molecular weight PTs than those initial efforts, and can be grouped into two categories based on their structure. Regioregular PTs can be synthesized by catalytic cross-coupling reactions of bromothiophenes, while polymers with varying degrees of regioregularity can be simply synthesized by oxidative polymerization.
The first synthesis of perfectly regioregular PATs was described by McCullough et al. in 1992.[57] As shown in Figure 5 (top), selective bromination produces 2-bromo-3-alkylthiophene, which is followed by lithiation, transmetallation and then Kumada cross-coupling in the presence of a nickel catalyst. This method produces approximately 100% HT–HT couplings, according to NMR spectroscopy analysis of the diads. In the method subsequently described by Rieke et al. in 1993,[58] 2,5-dibromo-3-alkylthiophene is treated with highly reactive “Rieke zinc"[59] to form a mixture of organometallic isomers (Figure 5, bottom). Addition of a catalytic amount of Pd(PPh3)4 produces a regiorandom polymer, but treatment with Ni(dppe)Cl2 yields regioregular PAT in quantitative yield.[60]
While the McCullough and Rieke methods produce structurally homogenous PATs, they require low temperatures, the careful exclusion of water and oxygen, and brominated monomers. In contrast, the oxidative polymerization of thiophenes using ferric chloride described by Sugimoto in 1986 can be performed at room temperature under less demanding conditions.[61] This method has proven to be extremely popular; H.C. Stark's antistatic coating Clevios P is prepared on a commercial scale using ferric chloride (see below).[62]
A number of studies have been conducted in attempts to improve the yield and quality of the product obtained using the oxidative polymerization technique. In addition to ferric chloride, other oxidizing agents, including ferric chloride hydrate, copper perchlorate, and iron perchlorate have also been used successfully to polymerize 2,2’-bithiophene.[63] Slow addition of ferric chloride to the monomer solution produced poly(3-(4-octylphenyl)thiophene)s with approximately 94% H–T content.[32] Precipitation of ferric chloride in situ (in order to maximize the surface area of the catalyst) produced significantly higher yields and monomer conversions than adding monomer directly to crystalline catalyst.[64][65] Higher molecular weights were reported when dry air was bubbled through the reaction mixture during polymerization.[66] Exhaustive Soxhlet extraction after polymerization with polar solvents was found to effectively fractionate the polymer and remove residual catalyst before NMR spectroscopy.[29] Using a lower ratio of catalyst to monomer (2:1, rather than 4:1) may increase the regioregularity of poly(3-dodecylthiophene)s.[67] Andreani et al. reported higher yields of soluble poly(dialkylterthiophene)s in carbon tetrachloride rather than chloroform, which they attributed to the stability of the radical species in carbon tetrachloride.[68] Higher-quality catalyst, added at a slower rate and at reduced temperature, was shown to produce high molecular weight PATs with no insoluble polymer residue.[69] Laakso et al. used a factorial design to determine that increasing the ratio of catalyst to monomer increased the yield of poly(3-octylthiophene), and claimed that a longer polymerization time also increased the yield.[70]
The mechanism of the oxidative polymerization using ferric chloride has been controversial. Sugimoto et al. did not speculate on a mechanism in their 1986 report.[61] In 1992, Niemi et al. proposed a radical mechanism, shown in Figure 6(top). They based their mechanism on two assumptions. First, since they observed polymerization only in solvents where the catalyst was either partially or completely insoluble (chloroform, toluene, carbon tetrachloride, pentane, and hexane, and not diethyl ether, xylene, acetone, or formic acid), they concluded that the active sites of the polymerization must be at the surface of solid ferric chloride. Therefore, they discounted the possibilities of either two radical cations reacting with each other, or two radicals reacting with each other, “because the chloride ions at the surface of the crystal would prevent the radical cations or radicals from assuming positions suitable for dimerization.”[71] Second, using 3-methylthiophene as a prototypical monomer, they performed quantum mechanical calculations to determine the energies and the total atomic charges on the carbon atoms of the four possible polymerization species (neutral 3-methylthiophene, the radical cation, the radical on carbon 2, and the radical on carbon 5).
Since the most negative carbon of the neutral 3-methylthiophene is also carbon 2, and the carbon with the highest odd electron population of the radical cation is carbon 2, they concluded that a radical cation mechanism would lead to mostly 2–2, H–H links. They then calculated the total energies of the species with the radicals at the 2 and the 5 carbons, and found that the latter was more stable by 1.5 kJ/mol. Therefore, the more stable radical could react with the neutral species, forming head-to-tail couplings as shown in Figure 6 (top).
Andersson et al. offered an alternative mechanism in the course of their studies of the polymerization of 3-(4-octylphenyl)thiophene with ferric chloride, where they found a high degree of regioregularity when the catalyst was added to the monomer mixture slowly. They concluded that, given the selectivity of the couplings, and the strong oxidizing conditions, the reaction could proceed via a carbocation mechanism (Figure 6, middle).[32]
The radical mechanism was directly challenged in a short communication in 1995, when Olinga and François noted that thiophene could be polymerized by ferric chloride in acetonitrile, a solvent in which the catalyst is completely soluble. Their analysis of the kinetics of thiophene polymerization also seemed to contradict the predictions of the radical polymerization mechanism.[72] Barbarella et al. studied the oligomerization of 3-(alkylsulfanyl)thiophenes, and concluded from their quantum mechanical calculations, and considerations of the enhanced stability of the radical cation when delocalized over a planar conjugated oligomer, that a radical cation mechanism analogous to that generally accepted for electrochemical polymerization was more likely (Figure 6, bottom).[73] Given the difficulties of studying a system with a heterogeneous, strongly oxidizing catalyst that produces difficult to characterize rigid-rod polymers, the mechanism of oxidative polymerization is by no means decided. However, the radical cation mechanism shown in Figure 6 is generally accepted as the most likely route for PT synthesis.
A number of applications have been proposed for conducting PTs, but none has been commercialized. Potential applications include field-effect transistors,[74] electroluminescent devices, solar cells, photochemical resists, nonlinear optic devices,[75] batteries, diodes, and chemical sensors.[76] In general, there are two categories of applications for conducting polymers. Static applications rely upon the intrinsic conductivity of the materials, combined with their ease of processing and material properties common to polymeric materials. Dynamic applications utilize changes in the conductive and optical properties, resulting either from application of electric potentials or from environmental stimuli.
As an example of a static application, poly(3,4-ethylenedioxythiophene)-poly(styrene sulfonate) (PEDOT-PSS) product Clevios P (Figure 7) from Heraeus has been extensively used as an antistatic coating (as packaging materials for electronic components, for example). AGFA coats 200 m × 10 m of photographic film per year with PEDOT:PSS because of its antistatic properties. The thin layer of PEDOT:PSS is virtually transparent and colorless, prevents electrostatic discharges during film rewinding, and reduces dust buildup on the negatives after processing.
PEDOT can also be used in dynamic applications where a potential is applied to a polymer film. The electrochromic properties of PEDOT are used to manufacture windows and mirrors which can become opaque or reflective upon the application of an electric potential.[27] Widespread adoption of electrochromic windows could save billions of dollars per year in air conditioning costs.[77] Finally, Phillips has commercialized a mobile phone with an electrically switchable PEDOT mirror.
The use of PTs as sensors responding to an analyte has also been the subject of intense research. In addition to biosensor applications, PTs can also be functionalized with synthetic receptors for detecting metal ions or chiral molecules as well. PTs with pendant[78] and main-chain[28] crown ether functionalities were reported in 1993 by the research groups of Bäuerle and Swager, respectively (Figure 8). Electrochemically polymerized thin films of the Bäuerle pendant crown ether PT were exposed to millimolar concentrations of alkali cations (Li, Na, and K). The current which passed through the film at a fixed potential dropped dramatically in lithium ion solutions, less so for sodium ion solutions, and only slightly for potassium ion solutions. The Swager main chain crown ether PTs were prepared by chemical coupling and characterized by absorbance spectroscopy. Addition of the same alkali cations resulted in absorbance shifts of 46 nm (Li), 91 nm (Na), and 22 nm (K). The size of the shifts corresponds to the ion-binding preferences of the corresponding crown ether, resulting from a twist in the conjugated polymer backbone induced by ion binding.
In the course of their studies of the optical properties of chiral PTs,[79][80][81][82] Yashima and Goto found that a PT with a chiral primary amine (Figure 9) was sensitive to chiral amino alcohols, producing mirror-image-split ICD responses in the π–transition region.[83] This was the first example of chiral recognition by PTs using a chiral detection method (CD spectroscopy). This distinguished it from earlier work by Lemaire et al. who used an achiral detection method (cyclic voltammetry) to detect incorporation of chiral dopant anions into an electrochemically polymerized chiral PT.[84]
A fluorine substituted polythiophene was shown to yield 7% efficiency in polymer-fullerene solar cells.[85]