Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
THE FIRST STEP TOWARD SUZUKI POLYCONDENSATION (SPC): A PERSONAL VIEW
In 1986, when I joined the newly founded Max Planck Institute for Polymer Science in Mainz as an experienced synthetic organic chemist, I was a real greenhorn in all matters regarding polymers. Gerhard Wegner, one of the directors, and Jim Feast (Durham University, England), an advisor to the institute in polymer chemistry-related issues at that time, convinced me in several discussions of the importance of an innovative polymer synthesis for organic materials science. So I decided to devote myself to this area of synthesis. They foresaw the considerable impact structurally new polymers would have if their structures were sufficiently novel and they could be obtained with structural integrity even under high analytical standards. They suggested as an important synthesis target straight-connected poly(arylene)s such as poly(para-phenylene)s (PPPs) with their all-carbon, conjugated, and conformationally rigid backbones. These structural characteristics make PPPs attractive candidates for liquid-crystalline, optoelectronic, and other applications but, at the same time, render their syntheses complicated. Unlike the established aromatic polyesters or polyamides, where the aromatic units are connected via carboxylic esters or amides and require CO-bond and CN-bond formations to synthesize, the synthesis of poly(arylene)s involves the formation of CC bonds, which is more difficult to achieve. Consequently, despite attempts by Kovacic and Jones,1 Yamamoto et al.,2 and ICI researchers,3 there were still no syntheses that would furnish structurally defined and high molar mass material.
Two aspects needed to be addressed to develop a passable trail into this class of polymers: high coupling conversion and solubility. On the basis of a retrosynthetic analysis of PPP, it was decided to connect the aromatic units directly to one another. Other possibilities for generating aromatic units sometime during the sequence3 were considered disadvantageous. Consequently, instead of chain growth, a step-growth procedure had to be developed for which it was known that extremely high conversions per individual bond-formation step were a strict necessity if high molar mass polymer was to be obtained. The CC-bond-formation reaction, therefore, had to be chosen with the greatest care. The second aspect was solubility. From short, linearly (1,4-) connected oligophenylenes, it is known that the solubility already drops to negligibly small values after a few connected benzene rings. The solubility of all-para-linked nonaphenylene, for example, which is just a very short model for PPP, is less than 10−8 g/L in toluene at room temperature.4 Any synthetic protocol that did not account for this problem was, therefore, prone to fail.
Beginning in 1987, I started on the PPP project together with Matthias Rehahn, a very able, synthetically oriented doctoral student. During many brainstorming sessions with chalk and blackboard, we analyzed the aforementioned synthesis attempts for their respective advantages and disadvantages and discussed the solubility aspect. Of the three routes, Yamamoto's was considered most interesting because it was the only one that guaranteed the required straight, 1,4-connection of benzene rings [Scheme 1(a)]. The accessible molar masses of approximately 1000 g/mol, however, and the intractability and infusability of the obtained powders were unacceptable. From the work of Kern, Heitz, and others, we knew that the solubility of rigid molecules increases drastically upon substitution with flexible side chains.5, 6 They render the dissolution of the molecules more attractive, mostly for entropic reasons. These considerations led us to believe that a simple decoration of 1,4-dibromobenzene, the Yamamoto monomer, with flexible alkyl chains may open a generally applicable route into PPPs [Scheme 1(b)]. Matthias synthesized 1,4-dibromo-2,5-dihexyl benzene (I),7 which turned out to be an extremely valuable compound for other purposes as well,8 and applied the Yamamoto conditions. The result, however, was quite disappointing.9 Regardless of how he did the reaction and which catalyst precursors he used, exclusively oligomeric products were obtained. The only improvement was the excellent solubility of the oligomers, which enabled him to accurately determine their molar masses and chemical structures. We determined that the steric hindrance imposed by the alkyl groups might have been responsible for termination at this early stage of growth, a view that was supported some time later by the successful Yamamoto-type synthesis of a sterically nonhindered polyarylene.10
In a parallel attempt to synthesize monodisperse oligophenylenes as model compounds, Matthias made an observation that, in retrospect, was essential for what followed. He found that under the conditions applied, magnesium/bromo-exchange reactions between bromo aromatics and Grignard aromatics could not be prevented, which made it impossible to obtain monodisperse compounds this way. He then used boronic acids as Grignard analogues and Suzuki cross-coupling (SCC) conditions for these model reactions and quickly obtained the desired compounds. The benefit from this was not obtaining these models, because they could have been made somehow anyway. More important was the personal experience of working with boronic acids and finding out that they were easier to handle than Grignard compounds and that the initially suspected complications with the known self-condensation of boronic acids could be managed. With this valuable experience in mind and actively supported by Gerhard Wegner and Jim Feast, Matthias then did the essential step and applied the Suzuki protocol to the bifunctional compound II and thus to polymer synthesis [Scheme 1(c)]. We compared the aromatic regions of the 13C NMR spectra of PPP 2 obtained according to Yamamoto and Suzuki and became excited. Although in the first spectrum [Fig. 1(a)] there were still end groups of considerable intensity, in the second nothing like that could be seen [Fig. 1(b)]. Applying other methods such as gel permeation chromatography (GPC) and viscometry, we quickly perceived that the polymerization went much better with 3 than with 1 and that we were on the verge of a breakthrough.
Because of its expected importance, we publicized this finding quickly, first by presenting it in the early summer of 1988 at a conference in Como, Italy, and shortly thereafter in lectures at the University of Massachusetts (Amherst, MA), Cornell University (Ithaca, NY), Eastman Kodak Chemical Co. (Rochester, NY), and Bayer AG (Leverkusen, Germany). The first publication was submitted to Polymer in August 1988.11 As is sometimes the case, important discoveries are made not only in one laboratory. Kim and Webster from DuPont (Wilmington, DE) also applied SCC, not to a bifunctional monomer such as in our case but to a trifunctional one. The subjection of 3,5-dibromobenzene boronic acid furnished a hyperbranched polyphenylene. This work was presented at the fall 1988 American Chemical Society meeting,12 was published in 1990, and has not been further followed since then.13
These findings have been further developed by us and other laboratories all over the world into a powerful and versatile tool of modern polymer chemistry, SPC.14 The following is a short overview on what has been accomplished during somewhat more than a decade since the discovery. Some important polymer chemical aspects of SPC are discussed, and a brief description of the impressive wealth of synthesized polyarylenes is given. This report closes up with a brief description of related procedures for which considerable progress has also been achieved.
SPC is a step-growth polymerization of bifunctional aromatic monomers to poly(arylene)s and related polymers (see the general outline in Scheme 2).14, 15 The required functional groups, boronic acid or esters on the one side and bromide, iodide, and so forth on the other, may be present in different monomers (AA/BB approach) or combined in the same monomer (AB approach). Both approaches have been successfully applied. The majority of the publicized work used the former despite the advantages of the latter for achieving high molar mass material. AB-type monomers intrinsically have the stoichiometric balance between the two different functional groups that, according to Carother's equation,16 is a strict necessity in step-growth polymerizations when high molar mass polymer is concerned. There is a simple synthetic reason the AA/BB approach is nevertheless favored. Normally, it is easier to synthesize aromatic monomers with two identical substituents in opposite positions (for benzene, 1,4) than those with different ones. An additional factor is that once an aromatic dibromide is obtained, its conversion into the corresponding diboronic acid or ester can often be achieved in one simple step and on a large scale. The price to be paid for this, however, is the necessity of applying the AA and BB monomers in strictly equal molar amounts. This requirement sounds almost trivial on paper. In reality, however, it may turn into a real experimental challenge. Purities, methods of how to completely transfer monomers into the polymerization vessel, and losses of some of the functional groups during polymerization become important and, all of a sudden, even critical aspects when the molar mass difference between two monomers is very large.17
The matter of purity is of real importance for SPC and should, therefore, be briefly addressed. Free boronic acid or one of the many cyclic boronic esters are used as boron-based functional groups. During polymerization, these esters may hydrolyze to the acids that then enter the normal cross-coupling or follow an independent mechanism.18 Boronic acids always contain some water. Otherwise, they are partially or completely condensed to cyclic boroxines. This water content has to be precisely determined for the reasons mentioned previously.19 Boronic esters, which do not have the problem with additional water, tend to partially hydrolyze on the column upon attempted purification. This renders weighing and, thus, stoichiometry control also somewhat problematical. Nowadays, easily applied solutions are available for both problems. They comprise procedures to prevent self-condensation,19 the use of pinacole esters to reduce ester hydrolysis,20 the rigorous application of high-field NMR, and, most importantly, series of test polymerizations. High molar mass polymer can be obtained only if the impurities are correctly quantified. Figure 2 shows a 500-MHz 1H NMR spectrum of diboronic acid (III), which is one of SPC's drosophilas. The small signal at δ = 3.4 ppm stems from water, whose content was determined as 2.7% (w/w) by NMR integration. Monomer III of this purity can be obtained on a scale of several tens of grams. The boronic monomer counterparts in SPC are aromatic bromides, iodides, or triflates. The bromo group is by far the most often encountered coupling partner in SPC. Iodides21 and triflates22 were only seldom used, although iodo compounds may gain increasing attraction because they were recently found to furnish higher molar mass products than their bromo analogues (discussed later). Chloro aromatics, although successfully used in organic chemistry SCC,23 have not been transferred to polymer chemistry yet. The purity of the dibromo monomers also has to be addressed with great care. Recrystallization for simple compounds and column chromatography for the more complex ones need to be repeatedly applied until a purity greater than 99% is achieved. Figure 2 shows an enlarged part of a 1H NMR spectrum of a pure sample of diiodo monomer IV. The main signal at δ = 3.9 ppm stems from the methoxy group. The enlargement make its 13C satellites visible (marked), which can be used as an internal NMR integration standard. Besides the spinning side bands, small signals also appear, which obviously stem from impurities. Because the intensity of the satellites is 0.5% of the main signal, the degree of impurities seems to be rather small. A quantification would, however, require knowledge of the impurities' structure.
The circles in Scheme 2 represent aromatic units, which are substituted benzenes in practically all cases but also include naphthalines, thiophenes, pyridines, and pyrroles (with an acceptor on nitrogen). When substituted with boronic acids, electron-rich aromatics tend to undergo deboronification reactions,24 which lead to stoichiometric misbalance with its detrimental impact on the achievable molar mass. This is why, for example, thiophenes in SPC are always used as dibromides and not as diboronic acids. These aromatic units are connected to one another to linear poly(arylene)s (for benzenes, PPPs) in more than 95% of all publicized cases. In a few examples, regularly kinked poly(arylene)s or related conjugated polymers containing additional olefinic or acetylenic units or other functional groups as part of the main chain are generated. Linear poly(arylene)s, whose chemical constitution in principle allows the attainment of a totally straight conformation, are considered rigid-rod-type polymers.6 Although they certainly have bent backbones and attain coiled conformations in solution, they have less conformational degrees of freedom, which are available at low energetic cost than, for example, saturated polymers such as polystyrene (PS). As a result, these poly(arylene)s show poor solubility because they have little driving force to dissolve molecularly dispers.
This is why in most cases when SPC comes into play, it is applied to monomers that carry flexible chains of some sort. These chains help keep the growing (and final) polymer in solution and accessible to further growth until growth reaches its system's intrinsic limits.6 These limits comprise termination through reduction of the bromo group or phosphorous incorporation through ligand scrambling channels (discussed later) or the removal of catalytically active Pd complexes through the precipitation of Pd(0) intermediates such as Pd black.
The substituents on the poly(arylene)s are not only important for solubility (and processability) reasons. They can also be used to incorporate function, a feature that has been increasingly and astoundingly successfully used in recent years. As far as the electronic properties of the backbones are concerned, substituents may, however, be disadvantageous, too. They normally lead to an increase of the dihedral angle of consecutive aromatic units, which reduces electronic conjugation and thus further increases the polymers' already quite large highest occupied molecular orbital (HOMO)/lowest unoccupied molecular orbital (LUMO) gap.
As for SCC, SPC involves only the carbon atoms that carry the functional groups. Polymerizations proceed regiospecifically. This is important because some of the properties of poly(arylene)s depend on their backbone's ability to attain straight conformations without kinks. Also, the functional group compatibility of SPC is the same as for SCC. Aldehydes, nitro and cyano groups, sulfonic esters, ethers, various protected alcohols and amines, amides, and so forth can be present. Even free hydroxy and free amines have been reported, although they do not seem to work too well. The reaction conditions are like the ones Suzuki reported in his famous, original article of 1981.25 Other solvent systems were also applied whenever required by the solubility of the polymer. For example, SPC has even been done in water with both water-soluble monomers and catalyst precursors.26 The mechanism of SPC is supposed to involve the same steps of oxidative addition, transmetallation, and reductive elimination as for SCC. The standard catalyst precursor is Pd(PPh3)4. Although SPC has not yet been developed into a reaction that is catalytic in an industrial sense, 0.5 mol % Pd complex is sufficient in many cases. Pd complexes with other phospine ligands have also been employed. For example, ortho- and para-tolyl ligands proved successful.19, 21 Although the choice of the best catalyst precursor is still a matter of intuition, it is accepted knowledge that the complex used should be as pure as possible. Thus, the commercially available Pd(PPh3)4 should not be used as obtained but rather should be recrystallized and used directly thereafter (under nitrogen). Best results are normally obtained when the Pd complexes are self-prepared and used freshly. Even if kept in a Teflon-sealed tube in a high quality glove box (<1 ppm oxygen), aging cannot be prevented, which inevitably leads to a reduction of molar mass even after a few days of storage.
Types of Polymers Prepared and Some of Their Applicational Aspects
Since the discovery of SPC in 1988, approximately 70 structurally different polymers with conjugated, fully or partially aromatic backbones have been prepared. From an inspection of the overview provided in Chart 1, one can see that the vast majority of these polymers are PPPs, which just differ in their substitution pattern. Whenever possible, the structures were drawn so that the mode of synthesis can be extracted: only one aromatic in the repeat unit indicates an AB-type approach, whereas two indicate an AA/BB-type approach. In the latter case, the left aromatic of the repeat unit stems from a diboronic acid or ester monomer, respectively, and the right one stems from a dibromide [or diiodide (14), di(trifluoromethylsulfonate) (23)]. For some of the more complex systems such as 37 or 56–58, this rule cannot be applied, and the literature needs to be consulted. Not all of the polymers shown were prepared in one step from monomers that already carry the final substitution pattern. Polymers 43–46, for example, were obtained by the chemical modification of PPP 13, and the double-stranded polymers 48–50, 52, 54, and 55 were obtained from single-stranded SPC precursor polymers (discussed later). The polymers in Chart 1 are divided into eight groups (a–h) according to their structural features, which correlate with certain properties in many but not all cases: (a) polymers with alkyl or alkoxy chains, (b) amphiphilic PPPs, (c) polyelectrolytes, (d) PPP precursors for ladder polymers, (e) polymers with main-chain chirality, (f) dendronized PPPs, (g) poly(arylene vinylene)s and poly(arylene ethinylene)s, and (h) others. These classes are discussed in the following, with a focus on certain especially typical or successful representatives.
Group a Polymers
This class comprises polymers 1,112,263,274,285,296,307,31, 328–11,33, 3412,3513,3514,3615,3616–18,3719,3820,3921,40, 4122,42 and 23–25.22 Polymer 1 is an important representative of this class, although it has not gained much importance in materials science. Its successful synthesis in 1988 marks the birth of SPC. For the first time, a structurally defined, all-para-linked, and processable PPP derivative of reasonable molar mass was available. The opportunity for a systematic and rational tailoring of suitable polymers for certain applications also came within reach for this kind of conjugated polymer.
PPPs have been investigated for applications in light-emitting diodes (LEDs).43, 44 They have a relatively large HOMO/LUMO gap that should give rise to a blue emission. Leising et al.45 was able to show this for the parent material. The development of a commercializable polymer-based LED is a complex enterprise, and many variables have to be considered and adjusted. A couple of Chart 1 PPPs have been investigated under this aspect. Polymer 21 (x = 0)41 seems to have won the race. Dow Chemical Corp. has announced the development of a blue-emitting LED based on this polymer and stated its imminent commercialization. Polymer 21 has just the right emission in the deep blue at λ = 465–470 nm, which cannot be easily reached by PPPs without partially planarized benzene rings. It has the additional advantage that it can be easily modified by copolymerization to give, for example, 21 (x = 1),40 for which, besides the right color, relatively high photoluminescence (PL) and electroluminescence (EL) efficiencies were demonstrated. The optoelectronic properties of polymers 24 and 25 for their use in multilayer LEDs were also investigated.22 The interested reader is referred to the literature as this whole matter cannot be treated here further.43, 44
A long-standing problem of surface sciences, the interaction of organic compounds with metal surfaces, was addressed with polymers 12 and 13.35 As observed earlier for rigid-rod poly(imide)s with terphenylene units (polymer 7046, 47), these polymers adsorb onto gold surfaces in ultrathin layers (15–25 Å) from solution. On copper, the thickness of these layers depends on the presence or absence of oxidizing agents. Thicknesses on copper of up to 900 Å were found. The metal surfaces can be completely hydrophobized by these layers as indicated by contact angles of approximately 90°. Upon the addition of alkylthiols, the adsorbed polymers on gold are pushed aside, and a nonhomogeneous mixture of domains of ordered, chemisorbed alkylthiols and thicker polymer layers are formed. Finally, polymers 14 and 1536 gained some importance as one-dimensional matrices with (protected) hydroxy anchor groups (one-dimensional matrices) to which other large substituents such as dendrons can be attached. Very recently, copolymers based on fluorenylidene-linked conjugated oligo(p-phenylene)s were reported.48
Group b Polymers
All polymers of this section, 26,49, 5027–33,5134,5235,52 and 36,53 carry oligoethyleneoxide chains that were incorporated to mediate some compatibility of the unpolar PPP backbone and polar media such as salts or water. Research is directed toward either the potential application of PPPs as separators in rechargeable solid-state lithium cells (26)49, 50 or the study of the aggregation behavior of these structurally novel rigid-rod amphiphiles in colloidal solutions or on surfaces (27–36).51–53 For 26, a structural model was also developed according to which the rods are ordered in layered structures separated by a matrix of the side chains. This matrix can be systematically swollen by blending with Li salts that are preferentially incorporated into this matrix. First investigations into the aggregation behavior of polymers 27–33 show that they form reversible aggregates in a tetrahydrofuran solution. Amphiphile 36 forms transferable Langmuir monolayers at the air/water interface with the backbones arranged parallel to the surface and the oligoethyleneoxide (OEO) chains dipping into the water phase. This finding was interpreted as evidence for the unusual ability of amphiphilically equipped PPPs to segregate lengthwise (referring to the backbone) in polar and unpolar domains.53
Group c Polymers
PPP polyelectrolytes 37,5438,55–5739,58–6040,6141,6242,63, 6443–45,65 and 4665, 66 were prepared mainly for three reasons: to show that SPC can be run in water, which broadens the applicability of this method considerably; to construct blue-emitting EL devices using the supramolecular ordering effects of charged macromolecules; and to investigate the formation of hierarchical structures in both solution and solid state. Wallow and Novak54 did the decisive experiment to prove the compatibility of SPC with an entirely aqueous medium. The usage of dibromobiphenyl monomer bearing two carboxylic acid functions and a water soluble Pd catalyst precursor with monosulfonated triphenylphosphine ligands together with a biphenyl bisboronic acid ester indeed gave polymer 37,54 which, because of its decoration with the two acid functions, turned out to be soluble in a diluted aqueous base. Similarly to some of the group 1 polymers, PPPs 38,55–5741,62 and 4263, 64 were investigated for their EL behavior. EL devices were prepared by the spin-coating of a solution of saponificated 38 onto an indium tin oxide (ITO) -covered glass substrate. ITO and aluminum were used as hole-injecting and electron-injecting electrodes, respectively. Intensive blue EL emission could be observed already at about 4 V if sodium counterions were used. Unfortunately, the lifetimes of the devices under operation were between 12 and 24 h, which is significantly too short for any commercial application. Blue emissions were also observed for polymers 4162 and 42.63, 64 LEDs constructed with them by both layer-by-layer electrostatic deposition and hybrid ink-jet printing methods also showed some decline in the initial properties.
For saponificated polymer 39,60 a model was developed to describe how rigid-rod polyelectrolytes aggregate in water. An analysis of the small-angle X-ray patterns of aqueous solutions of the fully anionic 39 reveals the formation of cylindrical micelles with a diameter of d = 3.1 nm and a radial aggregation number of 11 independent of concentration. In these micelles, all PPP backbones lie parallel to one another on the surface of a cylinder, with the sulfonate groups pointing outward and the unpolar dodecyl chains inward. In salt-free solutions, these micelles do not deaggregate into single polymer molecules even at concentrations as low as c = 0.001 g/L. From static and dynamic light scattering, it was concluded that above a critical concentration, the micelles form lyotropic objects that consist of approximately 130 micelles and are oblate spheroids with dimensions of 400 and 600 nm. Cylindrical aggregates in solution were also found for polymer 46, which in the solid state aggregates in columnar mesomorphous structures. Polymer 43,65 although not a polyelectrolyte, is mentioned here because it is an important intermediate on the way from 13 to PPPs 44–46.65, 66 Polymer 44 is exceptional because it carries four charges at every repeat unit.
Group d Polymers
Ladder polymers have been of long-standing interest for their thermal, electrical, and optical properties.67, 68 One of two basically possible routes into this class of polymers starts from a single-stranded polymer that is designed to allow a second independent strand to be made by some subsequent chemical modification. Polymers 47,6951,70, 71 and 5372, 73 are examples for this. They carry functional groups with which ring closures between two consecutive benzene rings were achieved to give ladders 48, 52, and 54, respectively. The conversions of these reactions can be driven to approximately 90–95%. Polymers 49, 50, and 5574 were obtained accordingly. Polymers 48 and 4975 have excellent EL characteristics but have not yet reached the state of technical production.
Group e and f Polymers
The development of enantioselective polymer-based catalysts is important in the efficient production of optically pure, chiral organic compounds, including drug molecules. The major advantage of polymer-based catalysts is the ease of recovery and reusability. Applications in flow reactors may also be feasible. Polymers 56–5876 are good examples here. Complexed with AlMe3 or Me2AlCl, they were used to catalyze hetero Diels–Alder reactions in a highly enantioselective manner. Dendronized polymers, which are polymers with dendritic side chains, have gained some importance in research directed toward molecular objects on the nanometer scale. For polymers such as 5977, 78 and 60,17, 79 it was proven that, because of the enormous steric congestion at each repeat unit, they are exceptionally rigid and attain a cylindrical shape in solution and when adsorbed on surfaces. This matter has been treated comprehensively in the literature.80 From a synthetic point of view specifically, polymer 60 is rather amazing. It was obtained in high yields and very high molar masses, which proves that SPC proceeds with conversions greater than 98% even in cases where monomers are sterically enormously loaded. The complex mechanistic cycle of SPC does not seem to be detrimentally influenced by this.
Group g Polymers
Although poly(arylene vinylene)s and poly(arylene ethinylene)s are normally prepared by other reactions (discussed later), SPC has also been applied to a few representatives of these classes of polymers. Besides polymers 61–63,31, 81 the already mentioned polymer 57 also formally belongs to this group. 62 was used to compare some of its EL and PL characteristics with those of the formally identical polymer prepared by the Heck polycondensation (see the Related Procedures section). Differences were actually observed and attributed to some defective connections occurring during the Heck procedure.
Group h Polymers
Most polymers of this group did not gain importance as materials for an application but should be noticed for synthetic considerations. Polymers 64–6682 show that SPC can be applied to monomers containing functional groups between two aromatic units. The dendritic macromolecule 6713, 83 proves that besides the conventional bifunctional SPC monomers, the trifunctional 3,5-dibromo benzeneboronic acid can also be employed. Although unprotected pyrrole cannot be subjected to SPC, the tert-butyloxycarbonyl-protected pyrrole can. The corresponding 2,5-dibromopyrrole derivative together with 2,5-didodecylbenzene bisboronic acid furnishes polymer 68,84 whose thermal treatment gives the polyarylene copolymer 69 with one or three pyrrolic units per repeat unit, which is otherwise not accessible. Polymer 7046, 47 is a novel rigid-rod type polyimide; polymer 7185 can be used for retro Diels–Alder chemistry to access, for example, unusual poly(arylene ethinylene)s; and the reduction of polymer 7286 yields the radical anion 73, which was investigated as a model for one-dimensional polaronic ferromagnetism.
Molecular Weight Determinations, Degrees of Polymerization (DPs), Mechanistic Considerations, and Diiodo Versus Dibromo Monomers
PPPs, polyarylenes in general, and other fully conjugated polymers such as poly(arylene vinylene)s and poly(arylene ethinylene)s are considered rigid rods. Rod-type macromolecules have, for a given molar mass, a larger hydrodynamic volume than flexible ones. In practically all cases, the molar masses and distributions of rigid-rod polymers are determined by GPC, a method that separates according to hydrodynamic volume. Narrow samples of different molar mass PS are used for the calibration. A GPC molar mass of a PPP derivative of, for example, 20,000 means nothing more than that this sample has the same hydrodynamic volume as a sample of PS with the actual molar mass of 20,000. With this argumentation, the actual molar mass of the PPP will be lower. The number 20,000 is just an upper limit. With this in mind, all publicized GPC molecular weights of the PPP derivatives presented here should be considered with care.
Fortunately, there is a thorough study available in which the actual molar mass of PPP 39 (R1 = methyl, R2 = dodecyl, and R3 = 3,5-di-tert-butylphenyl) was determined by light scattering, osmometry, and size exclusion chromatography with universal calibration and compared with the masses obtained from PS calibration.58 The universal calibration was done on the basis of the Mark–Houwink–Sakurada equation with nine fractions of 39 with weight-average molecular weights (Mw's) ranging between 27 and 189 kg/mol. A calibration based on the wormlike chain model gave very similar results. As can be seen from Figure 3, both these calibrations lead to almost superimposable elution curves at much lower molar masses than PS. A quantification of this difference reveals that PS overestimates the real molar mass of PPP 39 by a factor of almost 2. Although substituents certainly have an effect on the stiffness and thus the hydrodynamic volume of a polymer chain, it is reasonable to assume that this effect will normally be small. As a rule of thumb, GPC molar masses of PPP derivatives should, therefore, generally be corrected by this factor of 2 to lower values to have a more realistic estimate of the actual molar mass. For very large substituents, such as in the dendronized polymers 59 and 60, an additional aspect comes into play. If a given polymer mass per unit length is much larger than a PS′, GPC calibrated with PS can underestimate the real molar mass. For polymers such as 59 and 60, the factors of underestimation can reach 1.5–4, as was proven by light and small-angle neutron-scattering investigations.87 This underestimation by GPC is a clear exemption from the rule.
The reported DPs for polymers 1–73 mostly rest on GPC data referenced to PS. They vary greatly. With the aforementioned rule of thumb applied (a factor of 2), an average DP for SPC polymers lies between 30 and 60. Unfortunately, in many cases not even qualitative information is given on the monomer purity and how the catalyst precursors were treated. This range represents, therefore, a lower threshold for the intrinsically achievable DPs. Two quite different examples of the few available may suffice to prove the much greater potential of SPC for high molar mass polymer. The molar masses of polymers 3958 (mentioned previously) and 5978 were determined by state-of-the-art methods, and their DPs were 107 and 110, respectively. According to Carother's equation, such numbers translate into conversions per CC-bond-formation step of over 99%. Thus, if properly applied, SPC can in fact compete with the best step-growth polymerizations known. Because such very high molar masses were not needed for many of the reported SPC polymers, a special synthetic effort was not required. As far as optoelectronic properties are concerned, for example, typical polymer characteristics have already been obtained for quite short chains.
The mechanism of SPC is believed to basically follow the one developed for SCC.25 There is, however, an additional aspect that only comes into play if bifunctional compounds are subjected to SCC, as is the case for SPC. A few years ago, reports appeared in the literature that shed light on potential side reactions of SCC, an aryl–aryl exchange between aryls at the Pd center and phosphorus of the ligand.88 Consequently, aryl boronic acids (and related compounds) were found to couple not only with the aryl halide provided but also with aryls of the phosphine ligand. In low molar mass chemistry, this scrambling is disadvantageous but may be still acceptable as long as the yields of side products are not too high and they can be separated off. In SPC, however, aryl–aryl scrambling would be devastating. Novak et al.88 pointed out that phosphorus-containing groups could be incorporated not only as terminators but also as integral parts of the backbone. Scheme 3, in which substituents are omitted for clarity, shows some of the possible pathways. Up to compound D, everything is assumed to proceed the normal way. D now has the option to do what is required for SPC, which is to set E free by reductive elimination, or to undergo one or even two ligand scramblings to give F and G. These intermediates are now the source of side products, some of which contain phosphorus. Phosphorus acts in I as a terminus and in K as a kink and an electronic insulator. In further steps, networks may be formed with phosphorus as net points (not shown).
In fact, Novak et al.89 found NMR spectroscopic evidence for phosphorus in SPC polymers. Because this would considerably decrease the value of this method for the synthesis of structurally perfect rodlike polymers, as was believed to be the case, this matter was thoroughly investigated with polymers 14 and 15.21 In several independent preparations with different catalyst precursors and reaction conditions, the extent of phosphorus incorporation was quantified. According to 31P NMR spectroscopic results with internal integration standards, one phosphorus atom is found on roughly every 400 repeat units. It has, meanwhile, been accepted that this degree of incorporation of phosphorus in SPC polymers is negligible and need not be considered in all practical cases.
As mentioned at the beginning of this section, chloro aromatics have not been employed in SPC so far, although the variety of dichloro aromatics is much broader than that of dibromo or even diiodo ones (see, however, the Related Procedures section). There are a few reports regarding SPC with diiodo monomers.76, 88, 90, 91 Because SCC of iodoaromatics proceeds at a high rate under especially mild conditions, the potential for SPC for diiodo monomer 74 was investigated and directly compared with the otherwise identical dibromo monomer 76 (Scheme 4). The main focus was on the achievable molecular weight and the degree to which phosphorus was incorporated.21 Scheme 4 shows the reactions performed, and Table I lists the molar masses obtained (referenced to PS and a PPP standard58). The considerable increase in the molar masses of polymer 14 prepared from 76 compared with the masses of 74 needs no further comment. Fortunately, this increase does not need to be paid for with an increased imperfection of the backbone by phosphorus incorporation. Roughly, only every fourth chain contained one phosphorus atom, most likely located at the chain terminus.
Table I. Molecular Weights of Polymer 14 Synthesized from 74 and 76
The costs for the polymers synthesized via SPC reported here are high. Apart from commodity polymers, however, this does not automatically discourage their application. For an LED, for example, only very small amounts are needed, and the price of the device does not really depend on the costs of its organic active component. A good example for this is polymer 21, which is already commercially produced. Other SPC polymers will, however, never make it into any industrial-scale application. They are nevertheless indispensable for all kinds of basic research studies that will provide us further knowledge about rigid-rod polymers specifically and add to the understanding of polymer science in general. Because of the great expectation associated with poly(arylene)s, which goes beyond optoelectronic devices, it is not surprising that scientists in industry and academia already tried very early on to develop some ways to make certain PPPs available at a reasonable cost. The two important keys here were the nature of the leaving group and the catalyst. The question was whether bromine could be replaced with chlorine and Pd with Ni. The Union Carbide researcher Colon made the important observation in 1986 that under certain conditions chlorobenzene can be coupled to biphenylene in the presence of Ni complexes and Zn as a reducing agent [Scheme 5(a)],92 as already found for some poly(hetarylene)s. In a detailed study, which also shed light onto possible side and termination reactions, he found that the yield for this coupling step may exceed 98%. This was a clear indication that this CC-bond-forming reaction may be of use for step-growth polymerizations. It was again Colon and Kwiatkowski93 who made with this methodology the first polymer, the aromatic polyether sulfone 74 [Scheme 5(b)]. Unlike for most SPC cases, only one monomer is needed here, which carries two identical leaving groups (Cl). These groups are reductively cleaved off during polymerization. On the basis of this observation, Percec et al.94 took PPP synthesis a major step forward by applying this protocol to monomers 75a–c, which contain not only the halides chloride and bromide as leaving groups but also trifluoromethylsulfonate (triflate, Tf). They obtained the PPP derivatives 76 with a GPC molar mass of up to 6300 [Scheme 5(c)]. The polymerization of the ditriflate 75c went best and was, therefore, systematically further developed to the point that regioregular and regioirregular PPP copolymers with CF3 and OCF3 substituents with extremely high GPC molar masses became available.95 The regioregular polymer 77, for example, was obtained in a yield of 83% with a number-average molecular weight (Mn) of 55.000, which corresponds to 360 linearly connected repeat units [Scheme 5(d)]. Both triflates and the significantly cheaper methylsulfonates (mesylates, Mf) were used as leaving groups in this study. Because the corresponding monomers are prepared from hydroquinones, this methodology widens the applicability of transition-metal-catalyzed polycondensation considerably. Just 1 year after Percec submitted his first article on this matter (September 9, 1991), Maxdem Inc. launched a press release (November 24, 1992) in which the commercialization of polymer 78 was announced [Scheme 5(e)].96 Even though not all the experimental details were disclosed, the researchers Marrocco and Gagné seem to have used a procedure very similar to Colon's to obtain their PPP, which they named Poly-X™. Presumably because of the interesting properties reported by Maxdem for the thermoplastic polymer 78 (tensile modulus, 0.9–2.6 MSI; tensile strength, 15–35 KSI; dielectric constant, 3.0–3.3; moisture uptake, 0.2–0.3%; moldability, etc.) DeSimone et al.97 and Wang and Quirk98 also entered this research field. They mostly concentrated on the effect that changes in the synthetic protocol, for example, the nature of the chelating ligand in the Ni-catalyzed polymerization of 2,5 dichlorobenzophenone in the presence of excess zinc, have on the microstructure of 78 and, consequently, on properties such as solubility and glass-transition temperature. Wang and Quirk found that, for example, by changing the monodentate ligand triphenylphosphine to the bidentate 2,2′-bipyridine, they produced polymer 78 with a 68 °C higher glass-transition temperature and a λmax value in the ultraviolet–visible spectrum that is bathochromically shifted by 24 nm. Together with other evidence, these striking differences were attributed to more head-to-tail structures in 78 when the bidentate ligand was used.
During the early 90s, the group of Kaeriyama contributed to this field when they reported the synthesis of the ester derivative of PPP, 79,94 which was used as a precursor for parent PPP99 [Scheme 5(f)]. Hydrolysis of the ester functions in 79 afforded the free acids, which were catalytically decarboxylated with copper(II) oxide, basic copper(II) carbonate, and copper(I) oxide. PPP was obtained as an amorphous powder [Scheme 5(e)]. Very similar experiments were more recently reported by Ueda's group.100 BASF chemists finally were also active in this area during the same period but did not reach the commercialization stage. Naarmann and Kallitsis101 made a series of PPPs with the general structure 80 [Scheme 5(g)]. They found that the best results were obtained with 20 mol % catalyst precursor/mol of monomer. Such a high amount of catalyst is of course disadvantageous for many reasons, including financial ones. Recently, it was again Percec et al.102 who showed that PPPs with side-chain liquid-crystalline (SCLC) groups such as 81 can be obtained from dichloro monomers. Although research is still ongoing, first results indicate that this polymer may be the first SCLC polymer to form the NII nematic phase. During the 90s, even more articles emerged from this group on Ni- and Pd-catalyzed polyarylene syntheses.103
In a little over 10 years, the area of poly(arylene)s with their conjugated, all-hydrocarbon backbones has been revolutionized. These previously practically nonexistent polymers were made available through SPC and related procedures with a broad variety of substitution patterns. The decoration with substituents not only made them soluble and processable but also mediated function. This amazing development was driven not only by synthetic chemists, for whom poly(arylene)s had been an embarrassing blank space on the list of wanted polymers, but also by physicists and material scientists. They stressed that the use of polyarylenes with their now available structural integrity would enable one to more reliably correlate the measured properties of, for example, a polymer-based device with the polymer structure. Although there are still some synthetic hurdles to be overcome, the most important being to develop a catalytic process with high turnover numbers104 and to obtain high molar mass polymers from chloro monomers, the future of polymers prepared according to SPC and related procedures looks bright. The ever increasing demand for conjugated and rigid polymers will nourish the chemists' creativity and involvement to overcome these hindrances and design new polymer structures. These polymers are expected to have considerable impact in optoelectronics, as liquid-crystalline materials, in the area of colloidal chemistry, for molecular reinforcement purposes, and for surface coatings. It may especially surprise academic researchers to learn from the Maxdem press release95 that certain Poly-X™ resins are expected to sell for only about $10/lb. This shows the dramatic effect that using the two keys, the leaving group and catalyst, can have. The Ni/Zn-based cross-coupling polycondensation has been used not only for the synthesis of PPP derivatives alone but also for many heteroarylenes and related aromatic polymers. This work could unfortunately not be treated in this overview.105
The author cordially thanks G. Wegner (Mainz) and W. J. Feast (Durham, England) for igniting his interest in polyarylene synthesis and M. Rehahn for doing the first important steps together with him. The author thanks the latter and other coworkers (whose names are mentioned in the references) for their involvement and invaluable input. Financial support over the years from the following sources was essential for this work: Deutsche Forschungsgemeinschaft, Max-Planck-Gesellschaft, Freie Universität Berlin, and the Fonds der Chemischen Industrie. Finally, the author thanks the referees for helpful comments.
A. D. Schlüter
A. Dieter Schlüter received his doctorate in 1984 from the Ludwig-Maximillians-Universität München. After postdoctoral work with Professor K. P. C. Vollhardt (Berkeley) and Professor W. J. Feast (Durham, England), he joined the Max Planck Institute for Polymer Research in Mainz in 1986, where he concentrated on preparative macromolecular chemistry in the department of Professor G. Wegner. In 1991, he received his habilitation in organic chemistry at the University of Mainz. Shortly thereafter, he was awarded the Dozentenstipendium of the Fonds der Chemischen Industrie and was briefly an associate professor at the Polymer Institute of the University of Karlsruhe before accepting a chairmanship for organic and macromolecular chemistry at the Freie Universität of Berlin in 1992, where he has been ever since. He is on the executive board of the Wiley-VCH journals Macromolecular Chemistry and Physics and Macromolecular Rapid Communications and is a member of many committees and councils; moreover, he has held administrative functions, such as in the academic senate of the Freie Universität and in the chemistry department (as chairman). In 1998, he was awarded the Steinhofer-Preis of the Faculty of Chemistry and Pharmacy of the Universität Freiburg.