CHAPTER 1: Miktoarm Star (µ-Star) Polymers: A Successful Story

Well-defined polymers with narrow molecular, structural, and compositional (in the case of copolymers) dispersity are essential for establishing structure or composition–property relationships and indispensable to accomplish one of the ultimate goals of polymer chemistry: designing macromolecules with predetermined properties/applications.

Among branched polymers, regular or symmetric stars consisting of several identical linear chains linked together at one chain-end initially attracted the attention of scientists since the star structure has the simplest form of branching. The earliest attempt to synthesize star polymers was that by Schaefgen and Flory in 1948,¹  who synthesized the first 4- and 8-arm star homopolymers (polyamides) by polymerizing ε-caprolactam in the presence of either cyclohexanone-tetrapropionic or dicyclohexanone-octacarboxylic acid.

Fourteen years later, Morton and coworkers,²  taking advantage of the living character of anionic polymerization, succeeded to synthesize 4-arm star homopolystyrenes (PS) by ‘terminating’ living polystyryllithium with tetrachlorosilane (linking agent). Although the produced materials were mixtures of 3- and 4-arm PS, this work eventually led to the preparation of star polymers with up to 128 arms.

In 1963, Orofino and Wenger³  were the first to use tri(chloromethyl)benzene in combination with anionic polymerization as a linking agent to prepare 3-arm star PS. Mayer⁴  used 1,2,4,5-tetra-(chloromethyl)benzene to prepare 4-arm star di- and triblock copolymers of styrene and isoprene. It was difficult to extend the functionality (f) of stars beyond f = 6 with chloromethylbenzene derivatives due to the unavailability of chloromethylbenzene-based linking agents.⁵  Other compounds used as linking agents, such as the cyclic trimer of phosphonitrilic chloride,⁶  2,4,6-tri(allyloxy)triazine,⁷  1,1,4,4-tetraphenyl-1,4-bis (diallyloxytriazine)butane,⁸  tin tetrachloride,⁹  and phosphorus trichloride,¹⁰  suffer the same disadvantage. Decker and Rempp¹¹  demonstrated for the first time the validity of divinylbenzene (DVB) as a linking agent by preparing and properly characterizing PS stars with 6 to 15 arms. The DVB method was apparently first alluded to by Milkovich¹²  but, unfortunately, in his patent there was no clear indication that star-branched polymers had been prepared. It should be noted that the DVB method does not allow the accurate control of the number of star arms since the polymerization of DVB (difunctional monomer) with the living chains is not well controlled.¹³  Considering the disadvantages of the aforementioned compounds as linking agents, multifunctional chlorosilane compounds became the reagents of choice for the preparation of well-defined stars. Table 1.1 summarizes the evolution of the synthesis of symmetric star polymers with chlorosilane linking agents.

In 1989, Roovers and collaborators,²⁴  using a multifunctional linking agent designed/prepared by hydrosilylation of a low molecular weight linear or star 1,2-polybutadiene, succeeded to synthesize star polybutadienes (PB) with 200 and 270 arms. The exhaustive studies of the properties of these well-defined stars led to important conclusions concerning the influence of the star architecture on their properties in solution and bulk.²⁵  In addition, these model polymers were used to test the existing related theories.²⁶ 

Many other interesting linking systems using cationic, group transfer, or living ring-opening metathesis polymerization were later developed, leading to symmetric star vinyl ethers,^27–31  isobutylenes,^32–36  methacrylates,³⁷  and norbornenes.^38,39 

Later on, the synthesis of stars with different arms either in molecular weight (molecular weight asymmetry; asymmetric stars) or chemistry (chemical asymmetry; miktoarm stars) was achieved (Scheme 1.1a and 1.1b). The term miktoarm stars (coming from the Greek word µικτός meaning mixed) was adopted by our group for stars with chemical asymmetry. The term heteroarm stars (hetero coming from the Greek word έτερος, meaning the other) is not appropriate for this class of polymers, since it does not convey the concept of a star composed of dissimilar arms. Later, the term was expanded to stars with molecular weight asymmetry, which can be considered miktoarm homopolymers. Stars having arms of similar chemical nature but different end-functional groups also belong to this category. Finally, topologically asymmetric stars are also µ-stars. They consist of diblock copolymer arms attached by different chain-ends to the star center. π-Shaped, H-shaped, super-H or pom-pom copolymers can be considered as double miktoarm stars (Scheme 1.1c).

Pennisi and Fetters⁴⁰  were the first to report the synthesis of 3-arm asymmetric star homopolymers of PB and PS. Mays⁴¹  and our group,⁴²  based on the work of Fetters, were the first to synthesize a 3µ-star copolymer consisting of two polyisoprene (PI) and one PS arm⁴¹  and a 3µ-star terpolymer of PS, PI, and polybutadiene (PBd or PB),⁴²  respectively. Later, different miktoarm stars were synthesized by anionic polymerization and selective chlorosilane-based linking chemistry, as reported mainly by our group,^43–61  representative examples of which are given in Scheme 1.1. The evolution of the synthesis of different miktoarm structures are summarized in Table 1.2.

Synthetic efforts based on anionic polymerization leading to miktoarm stars (µ-stars) are reviewed in this chapter. In addition, a few examples of the striking influence of the star structure on the morphology of block co- and terpolymers are given. The structures synthesized by anionic polymerization guided scientists working with other types of polymerization techniques, such as controlled/living, ring opening, catalytic, ring opening metathesis polymerization reactions, etc., towards the synthesis of star architectures. The tremendous influence of miktoarm stars on polymer science is evidenced by: (i) the need for a book on miktoarm stars, (ii) the high number of references miktoarm stars have produced in the last years (h-index = 70, references = 21 275, from our first paper to August 2016, source: ISI web of science), and (iii) the significant amount of novel nanostructures reported by this class of materials.

Two general strategies have been developed for the synthesis of miktoarm stars via living anionic polymerization. The first one is based on divinyl compounds, either homopolymerizables (e.g., divinylbenzene) or non-homopolymerizables (e.g., double diphenylethylene, DDPE), and the second one on multifunctional linking agents. Several linking agents have been used for the synthesis of star polymers.⁶²  The most commonly used ones are chlorosilanes,²  adopted mostly for non-polar chains, and chloro-/bromomethyl benzenes,^3,4  adopted for polar ones. A few more complex linking agents have also been used in the synthesis of star-like copolymers.^63–65 

The use of DVB for the synthesis of miktoarm stars was first recognized by Eschwey and Burchard⁶⁶  and developed mainly by Rempp and colleagues.^67–70  The general route is given in Scheme 1.2.

The living macroinitiator/precursor ALi polymerizes a small amount of DVB,⁷¹  leading to the formation of a star molecule bearing a number of active sites within its core (microgel nodule of DVB), which is theoretically equal to the number of incorporated A arms of the star polymer. Subsequent addition of another monomer, B, or the same monomer A, yields a µ-star copolymer or asymmetric homopolymer star, respectively. The growing B arms have anionic sites at their outer ends, thus providing the possibility of reacting with electrophilic compounds or other monomers towards the preparation of end-functionalized stars or star-block copolymers. Because of its simplicity, this method can be carried out under inert atmosphere, thereby avoiding the use of highly demanding and time-consuming vacuum techniques. The DVB method has been applied in the synthesis of µ-star copolymers of the A_nB_n type, with the A arm being polystyrene and the B arm poly(tert-butyl methacrylate),⁶⁸  poly(tert-butyl acrylate),^69–71  poly(ethylene oxide),⁷²  or poly(2-vinyl pyridine).⁷³  Usually, n varies between 6 and 20. PS µ-star homopolymers of the type A_nA′_n have also been prepared by this method.⁷⁴ 

The DVB method is characterized by several disadvantages, the foremost being architectural limitations. Only stars with equal number of arms different in chemical nature or molecular weight, A_nB_n, can be prepared. In fact, since the polymerization of DVB by living chains is not controllable, n is an average value influenced by several parameters. Specifically, n increases with the decreasing molecular weight of the precursor A and the molar ratio of DVB to living chains. Another disadvantage is that the B arms cannot be isolated and characterized independently. Finally, reaction of the living ends with the remaining double bond of the DVB nodule can lead to the formation of loops (intramolecular reaction) or networks (intermolecular reaction).

Hocker and Latterman⁷⁵  recognized in 1976 the usefulness of the addition of living chains to non-homopolymerizable divinyl compounds. They were the first to propose that 4µ-stars could be obtained by adding living chains to bis(1-phenylvinyl)benzenes, followed by the subsequent addition/polymerization of another monomer. In 1978, Szwarc and coworkers⁷⁶  studied the kinetics of addition of PSLi to several divinyl compounds in benzene by UV spectroscopy. They found that, in the case of the para-double diphenyl ethylene (PDDPE) 1,4-bis(1-phenylethenyl)benzene, the ratio of the rate constants of the first and second additions was equal to 13. In 1983, Leitz and Hocker⁷⁷  reported that the reaction of two moles of sec-BuLi with the meta-double diphenylethylene (MDDPE) 1,3-bis(1-phenylethenyl)benzene proceeds rapidly and efficiently to produce a dilithium initiator. The ratio of the rate constants of the first and second additions is almost identical in toluene.

Quirk and coworkers⁷⁸  have further developed this ‘living linking method’ for the synthesis of 3µ- and 4µ-stars. The general reactions are given in Schemes 1.3 and 1.4, respectively. PDDPE is usually employed for the synthesis of 3µ-A₂B and 3µ-ABC, whereas MDDPE is used for A₂B₂ 4µ-stars. More recently, Quirk’s group extended this method to 6µ-stars by using a triple diphenylethylene, 1,3,5-tris(1-phenylethyl)benzene.⁷⁹ 

The key to the living linking procedure is the control of the stoichiometry of the reaction between the living A chains and the DDPE; otherwise, a mixture of star and linear polymers is produced. A major problem is that the rate constants of initiation of the two new active sites differ, resulting in a bimodal distribution. To overcome this problem, polar compounds have to be added. It is well known that they dramatically affect the microstructure of the polydienes formed in later stages. However, addition of lithium sec-butoxide to the living DDPE derivative prior to the addition of the diene monomer was found to produce monomodal well-defined µ-stars with high 1,4 content. Again, the B arms cannot be isolated from the reaction mixture and characterized separately. Nevertheless, Quirk’s method is valuable for the synthesis of ω-functionalized µ-stars.

Regular homopolymer and block copolymer stars can be synthesized by reaction of an excess of living chains, prepared by anionic polymerization, with the appropriate chlorosilane. An excess of the living polymer is needed to force the linking reaction to completion. For the synthesis of µ-stars, each chlorine atom should be replaced stepwise by a different chain. To achieve this goal, the different reactivity of the living chain ends towards the SiCl group must be taken into consideration. The reactivity of the living chain end decreases with the charge delocalization and by increasing the steric hindrance as follows: butadienyl lithium (BdLi) > isoprenyl lithium (IsLi) > styryl lithium (SLi) > diphenyl ethylenyl lithium (DPELi). The reactivity of the living ends is also affected by the chain molecular weight (the lower the molecular weight, the lower the steric hindrance and, consequently, the higher the reactivity), the polarity of the environment (the higher the polarity, the lower the association of the living chains and, consequently, the higher the reactivity), and the temperature (same as the environment).

Chlorosilanes cannot be combined with macroanions of polar monomers, such as (meth)acrylates and 2-vinyl pyridine, since the linking reaction either leads to unstable products [poly(meth)acrylates] or it does not occur at all [poly(2-vinyl pyridine), P2VP]. Instead, linking agents including chloro-/bromomethyl benzene derivatives^3,4  are used although, unfortunately, they display significant drawbacks such as lithium–halogen exchange, leading to linking agents with higher functionalities and, consequently, to a mixture of stars with different functionalities.⁶²  To overcome this problem, potassium (instead of lithium) counter ions and polar solvents at low temperature (−78 °C) should be used.

Trichloromethylsilane (CH₃SiCl₃) and tetrachlorosilane (SiCl₄) are appropriate linking agents for the synthesis of 3µ- and 4µ-stars. The replacement of only one Cl by one chain can be achieved by very fast addition of the living polymer into a large excess of the chlorosilane. Before the addition of the second living chain, the unreacted chlorosilane is removed to avoid contamination of the µ-stars with the homostars with three (CH₃SiCl₃) or four (SiCl₄) arms. The above chlorosilanes can be easily removed on a high vacuum line owing to their low boiling point (CH₃SiCl₃, b.p.: 66 °C and SiCl₄, 57.6 °C). This method was developed by Pennisi and Fetters,⁴⁰  and was used for the synthesis of 3µ-star homopolymers of styrene and butadiene with arm molecular weight asymmetry. The general reactions for the synthesis of 3µ-star homopolymers are given in Scheme 1.5.

The excess A′Li needed for completion of the linking reaction is removed by fractionation, as in the case of regular stars. In the case of asymmetric PS stars, to ensure complete reaction of PSLi with ASi(CH₃)Cl₂, the living PS is end-capped with a few units of butadiene to increase its reactivity towards the chlorosilane.

Mays prepared⁴¹  one sample of 3µ-star PS(PI)₂ by this method. The method was further developed by our group to all possible combinations of A₂B µ-stars, where A and B were PS, PI, or PB.⁴⁵  Furthermore, by using SiCl₄, we prepared PS(PI)₃ 4µ-stars.⁴⁶  A more sophisticated high vacuum technique was used to ensure the synthesis of well-defined µ-stars.⁴²  A high degree of molecular and compositional homogeneity was identified by size exclusion chromatography (SEC), with refractive index and UV-detectors, low-angle laser light-scattering (LALLS), membrane (MO) as well as vapor-pressure osmometry (VPO), and NMR spectroscopy. The chlorosilane approach was also adopted for the synthesis of µ-stars of the PS(PI-b-PS)_{2 or 3} type.⁵¹ 

When only one chlorine needs to be replaced by a highly reactive living chain, for example low molecular-weight PBLi or PI, the living end has to be transformed by end-capping into a less reactive carbanion. By decreasing the reactivity, the selectivity is increased and the replacement of only one chlorine is achieved. Using DPE, we prepared model mono- and difunctional 3µ-star PBs⁴⁶  and (d-PB)₂(PI), (d-PB)(PI)₂, where d-PB is deuterated PB,⁸⁰  having arm molecular weights below 10 000 g mol⁻¹. The same goal could be achieved more easily (avoiding the necessity for end-capping) by linking the first arm or arms at a temperature low enough to create selectivity between the successive steps of replacement of chlorines.⁴⁸ 

Using the chlorosilane method, 4µ-stars (PS-b-PB)(PB)₃ were synthesized by Tsiang.⁸¹  Living polybutadiene PB chains were first reacted with SiCl₄ in a molar ratio 3 : 1, followed by addition of the living diblock (PS-b-PB)Li. The key step of this method was the successful synthesis of the (PB)₃SiCl intermediate. The high reactivity of the PBLi chain end posed questions about the purity of this polymer, since several byproducts such as (PB)₂SiCl₂, (PB)₄Si, or PBSiCl₃ can be formed in the first step. SEC analysis was performed to monitor the reaction sequence. It is obvious that the method developed by Tsiang is very demanding with regard to the stoichiometry of the reagents. The byproducts are almost impossible to be separated from the main miktoarm star.

For the synthesis of 3µ-ABC terpolymers (Scheme 1.6), three chlorines have to be consecutively and selectively replaced. The first chlorine can be replaced by PI using an excess of CH₃SiCl₃, the second is replaced by titration with PSLi, and the third by an excess of PBLi.⁴²  The order of linking of the different living chains to the chlorosilane plays an essential role. The most sterically hindered chain, PSLi, has to be added by titration in the second step, the less sterically hindered chain end, PBLi, has to be added at the last step of the synthesis. The whole procedure can be monitored by SEC, as shown in Figure 1.1.

Using the same route, asymmetric AA′B miktoarm stars were also prepared.⁸²  These are stars having two chemically identical A arms but of different molecular weights. The synthesis of the 4µ-star copolymer (PS)₂(PI)₂ was accomplished in a similar way.⁴³  PSLi was added in two separate steps (excess and titration) to achieve maximum control over the polymer architecture.

Using the selective replacement of the chlorines of SiCl₄, a miktoarm star quarterpolymer was prepared consisting of PS, poly(4-methyl styrene) (P4MeS), PI, and PB.⁴³  The reaction sequence for the preparation of this miktoarm star is presented in Scheme 1.7. The main feature of this method is that two of the arms are incorporated by titration. PS was chosen to react first with an excess of SiCl₄, followed by titration with the more sterically hindered P4MeS after evaporation of the excess silane. The moderately hindered PILi anion was then added by titration. The least sterically hindered PBLi anion was used to complete the linking reaction.

Xie and Xia were the first to prepare a 4µ-star A₂B₂.⁸³  Living PS chains were reacted with SiCl₄ in a 2 : 1 molar ratio leading to the formation of a two-arm product. The two remaining SiCl bonds were used for a linking reaction with living poly(ethylene oxide) (PEO) chains. The process is facilitated by the increased steric hindrance of the living PS chain ends. It is, indeed, very difficult to prepare 3- or 4-arm PS stars. From this point of view, control of the stoichiometry is less important. A₂B miktoarm stars have also been prepared using CH₃SiCl₃ instead of SiCl₄.

1,2-Bis(trichlorosilyl)ethane (SiCl₆) has been used for the synthesis of PS(PI)₅ 6µ-stars.⁵³  Stoichiometric addition of PSLi to the SiCl₆ linking agent was chosen in this case, instead of an excess of SiCl₆, since the unreacted solid chlorosilane cannot be removed by distillation. The reaction sequence is outlined in Scheme 1.8.

Living PS chains were reacted with 1,2-bis(trichlorosilyl)ethane in a 5 : 1 ratio. Dropwise addition of the living polymer solution into the vigorously stirred solution of the linking agent was performed to minimize multiple coupling products. Under these conditions, 15% of the dimeric product was formed. The pentachlorosilane-capped PS was then reacted with an excess of PILi, followed by fractionation to isolate the desired polymer.

Model 16µ-star copolymers with eight PS and eight PI arms, PS₈PI₈, were synthesized using a dendritic carbosilane Si{[CH₂CH₂Si(CH₃)][CH₂-CH₂Si(CH₃)Cl₂]₂}₄ or SiCl₁₆ with 16 SiCl bonds.⁵⁰  This compound was prepared using tetravinylsilane as the initial core molecule, methyldichlorosilane as the propagating units, and vinylmagnesium bromide for the transformation of silicon chloride to the silicon vinyl group. It was found that PS must be linked first, followed by addition of an excess of PI, in order to successfully synthesize these miktoarm stars due to two basic constrains: (a) the inability of the sterically hindered polystyryl anions to react with both chlorines at each peripheral silicon atom and (b) the ability of the less sterically hindered polyisoprenyl anions to react completely with the remaining chlorosilane bonds. Living PS chains were reacted with the linking agent in an 8 : 1 molar ratio for the preparation of the PS 8-arm star. Even a slight excess of PSLi (∼5%) could be used without the incorporation of more than eight arms due to the steric hindrance of the already attached PS chains on each Si atom. A small excess of PILi to SiCl was finally added to prepare the final product.

The term Vergina star copolymers was coined by our group for A₈B₈ miktoarm stars due to their similarity with the 16-rayed star emblem discovered by the late Professor M. Andronikos of the University of Thessaloniki in the ancient tomb of King Philippos of Macedonia in Vergina, a place close to Thessaloniki (Figure 1.2).

By using the same methodology and a chlorosilane with 64 peripheral SiCl groups, Roovers⁸⁴  prepared a 64µ-star copolymer of the A₃₂A′₃₂ type, where A and A′ are PB arms with different molecular weights.

Double µ-stars have been synthesized by selective coupling bifunctional living chains instead of monofunctional ones. The structures obtained are shown in Scheme 1.1c. As an example, the synthetic procedure for the super H-shaped block copolymer⁴⁴  is outlined in Scheme 1.9.

More complex miktoarm stars (Figure 1.3) have been synthesized by our group.⁸⁵  Styrenic single and double star-tailed macromonomers were first obtained by reaction of living homo/miktoarm stars with the SiCl groups of 4-(chlorodimethylsilyl)- and 4-(dichloromethylsilyl)styrene, respectively. The in situ anionic homopolymerization of macromonomers with sec-BuLi or copolymerization with butadiene or styrene, led to single/double homo/miktoarm star-tailed molecular brushes and combs, as well as to a block copolymer consisting of a linear polystyrene chain and a double miktoarm (PBd/PS) star-tailed brush-like polymer.

Miktoarm stars involving diblock copolymers as arms have also been reported by our group.^49,51,86,87  Typical cases are (a) inverse 4-arm star copolymers, (PS-b-PI)₂-junction-(PS-b-PI)₂, where each arm is a block copolymer of styrene and isoprene but two of the arms are connected to the junction point with the PS block, whereas the other two are so with the PI block, and (b) A(BA)_{n = 2, 3} and super-H shaped stars of the (AB)₃A(BA)₃ type, where A = PS and B = PI. Recently, similar but asymmetric non-linear copolymers A(BA′)_{n = 2, 3} (Scheme 1.10) have been reported.⁸⁸  These complex miktoarm-based copolymers have tremendous impact on the microdomain morphology of block copolymers.^49,86 

By applying the chlorosilane approach, our group was able to prepare model 1,4-poly(butadienes) of the A₂A, AA′A″, A₂A′A₂, A₃A′A₃, and A₅A′A₅ types and, by hydrogenation, the corresponding branched polyethylenes.⁸⁹ 

A₂B stars with two PS arms and one poly(2-vinyl pyridine) (P2VP) arm, (PS)₂(P2VP), were prepared by Khan et al.⁹⁰  Living PS chains were linked to dichloromethylsilane, CH₃SiCl₂H, to produce the two arms of the star. In another reactor, living P2VP was terminated with allyl bromide. The 3µ-star was produced through hydrosilylation addition of SiH to the allyl group. Due to incomplete hydrosilylation, the µ-stars were found to have high polydispersity (1.33 to 1.50).

Anionic polymerization techniques and naphthalene chemistry were used by Teyssie et al. to prepare A₂B miktoarm stars, where A was PEO and B was PS, PI, poly(α-methyl styrene), or poly(t-butyl styrene).⁹¹ 

A special technique was employed by Naka et al. for the preparation of A₂B stars, A being PEO and B polyoxazoline (POX), by forming Ru(iii) complexes with bipyridyl terminated polymers.⁹² 

Star polymers with several PS arms and only one poly(2-vinyl naphthalene) (PVN) arm were prepared by Takano et al. by anionic polymerization techniques.⁹³  Sequential anionic block copolymerization of (4-vinylphenyl)dimethylvinylsilane (VS) and VN was employed. The double bonds attached to silicon have to remain unaffected during the polymerization of VS. This was accomplished with K counterions in THF and short polymerization times. The PVS block with the unreacted double bonds was used as a multifunctional linking agent. Subsequent addition of living PS chains produced miktoarm stars of the type (PS)_nPVN. Characterization studies revealed that n = 13.

Similar structures of the AB_n type were prepared by Wang and Roovers, where A = PS and B = PB or P2VP.⁹⁴  Due to the much higher molecular weight of the PS arm, these µ-stars were called umbrella copolymers (Scheme 1.11). The reaction sequence for the preparation of the PS(PB)_n copolymers is given in Scheme 1.11.

Butadiene was polymerized anionically in the presence of dipiperidinoethane (dipip), followed by addition of styrene. Hydrosilylation chemistry was employed to add the Si(CH₃)Cl₂ or Si(CH₃)₂Cl groups to the 1,2-PBd double bonds. Subsequent addition of 1,4-PBdLi or P2VPK led to the formation of umbrella copolymers. The limited control exercised over the hydrosilylation reaction is the main reason why the number of arms cannot be accurately controlled.

Roovers and collaborators⁹⁵  also succeeded in preparing umbrella-star copolymers [(PS-uPI)_n]_m (Scheme 1.12). The synthesis is based on the reaction of (1,2-PB-b-PS)Li with a chlorosilane having 32 peripheral SiCl bonds, followed by hydrosilylation of the PB double bonds and reaction with PILi.

Ishizu and Kuwahara⁹⁶  have developed a macromonomer technique for the synthesis of miktoarm stars of the A_nB_m type. PS and PI macromonomers with vinyl end-groups were anionically copolymerized in benzene solutions with n-BuLi. The products can be considered miktoarm stars of the type A_nB_m, as evidenced by their solution and solid-state properties.

Diblock macromonomers with central vinyl groups were used for the synthesis of (PS)_n(PtBuMA)_n µ-stars.^97,98  The macromonomers were prepared by sequential anionic polymerization of styrene, 1,4-divinylbenzene (DVB), and t-butyl methacrylate (tBuMA). The DVB monomer was left to react with the living PS chains for short times (5 min) so that only a few DVB units were incorporated at the end of the PS chains, and the formation of PS stars was avoided. Free radical polymerization in solution and in bulk using AIBN as the initiator, tetramethylthiuram as the photosensitizer, and ethylene glycol dimethacrylate as the crosslinking agent was carried out for the synthesis of the µ-stars. A similar experiment was performed using PS-b-P2VP diblocks having central isoprene units.

Finally, a cyclophosphazene derivative was used as a linking agent to produce µ-stars consisting of PS and Nylon-6 arms.⁹⁹ 

Fujimoto et al. prepared (PS)(PDMS)(PtBuMA) µ-stars,¹⁰⁰  as described in Scheme 1.13. The lithium salt of p-(dimethylhydroxy)silyl-diphenylethylene was used as the initiator for the polymerization of hexamethylcyclotrisiloxane (D₃). Living PS chains were reacted with the end double bond of the diphenylethylene, followed by the anionic polymerization of tBuMA.

PDMS has a rather high molecular weight distribution (I ∼ 1.4) and fractionation needs to be performed before the following synthetic steps. The polymethacrylate branch cannot be isolated and studied independently.

A similar synthetic route was adopted by Stadler et al. for the synthesis of (PS)(PB)(PMMA) stars,¹⁰¹  as shown in Scheme 1.14.

Living PS chains were end-capped with p-bromomethyldiphenylethylene. The next step involved the linking of living PB chains, prepared in THF at −70 °C, to the end double bond. This reaction produces a new active center used to initiate the polymerization of MMA.

The chlorosilane method was also employed by our group for the synthesis of µ-stars with PS, PI, and PMMA arms.^52,102  The reaction sequence is presented in Scheme 1.15.

The monofunctional macromolecular linking agent (PS)(PI)(CH₃)SiCl was prepared through a procedure similar to those described above. This linking agent was reacted with a stoichiometric amount of the dilithium initiator prepared from 1,1-diphenylethylene (DPE) and Li. The remaining active center was used to polymerize MMA in THF at −78 °C. In this procedure, the PMMA arm cannot be isolated and cannot be characterized independently.

A technique similar to Stadler’s was employed for the synthesis of miktoarm stars with PS, PEO, poly(ε-caprolactone) (PCL), or PMMA arms.¹⁰³  A PS-b-PMMA diblock copolymer possessing a central DPE derivative bearing a protected hydroxyl function was prepared. After deprotection and transformation of the hydroxyl group into an alkoxide, the anionic ring-opening polymerization of the third monomer (EO or ε-CL) was initiated. Only limited characterization data were given.

Recently, Hirao and coworkers¹⁰⁴  have proposed a promising methodology for the synthesis of 3-, 4-, 5-, and 6-arm regular and µ-stars. Anionic living chains react with specially designed terminators with a predetermined number of methoxy benzyl groups. The methoxy groups are transformed into a chloromethyl group, followed by reaction with another living chain. As an example of the concept, the synthesis of a 6µ-star A₂B₄ is given in Scheme 1.16.

By exploiting the unique features of anionic polymerization to synthesize end- or in-chain amino-functionalized polymers, as well as the living nature of the ring opening polymerization of α-amino acid N-carboxyanhydrites (NCAs) under high vacuum conditions, our group in collaboration with Hirao’s group succeeded to synthesize the following miktoarm star (µ-star) polymer/polypeptide hybrids (macromolecular chimeras): (PS)₂(PBLG or PBLL) and (PS)(PI)(PBLG or PBLL) (3µ-stars), and (PS)₂[P(α-MeS)](PBLG or PBLL) and (PS)₂(PBLG or PBLL)₂ (4µ-stars), where P(α-MeS) is poly(α-methylstyrene), PBLG is poly(γ-benzyl-l-glutamate), and PBLL is poly(ε-tert-butyloxycarbonyl-l-lysine).¹⁰⁵  DPE-functionalized polymers were first prepared and subsequently activated by reaction with a living chain or sec-BuLi, followed by reaction with 1-(3-bromopropyl)-2,2,5,5-tetramethyl-aza-2,5-disilacyclopentane. The silyl-protected group was cleaved under acidic conditions to afford amine-functionalized macroinitiators for the polymerization of BLG and BLL NCAs to afford the desired miktoarm stars. An example is given in Scheme 1.17. Detailed characterization revealed the efficiency of the synthetic strategy and the homogeneity of the synthesized miktoarm chimeras.

The morphology and consequently the properties of a given linear block copolymer are controlled mainly by the volume fraction of its components. Upon changing from linear to a non-linear architecture, the study of miktoarm star copolymer microphase separation becomes very important, since these materials shift the boundaries between different morphologies compared to the corresponding linear copolymers or form new structures (µ-ABC).¹⁰⁶  As an example, the PS-b-PI (PS: 40% by volume) linear copolymer has a lamellar morphology, whereas the non-linear (PI)₂PS and (PI)₃PS block copolymers, with the same overall molecular weight and composition, form hexagonally packed cylinders and spheres in a body centered cubic (BCC) array, respectively.

In the first microphase separation study of A₂B miktoarm stars synthesized by our group,¹⁰⁷  where A and B were PS or PI, respectively, a PS(PI)₂ sample with 37 vol% PS was found by transmission electron microscopy (TEM) to microphase separate, exhibiting PS cylinders in the PI matrix. This observation is in contrast to the alternating lamellar structure expected for a linear diblock copolymer with the same volume fraction. The subsequent studies^47,108,109  with a larger number of samples and covering a wider range of compositions revealed that differences exist in the phase diagram of miktoarm star copolymers in comparison to what is predicted and expected for the corresponding linear diblock copolymers (in composition and molecular characteristics). Analogous shifts were also observed in the case of A₃B (A = PI and B = PS) miktoarm stars.⁴⁷  These findings are in qualitative agreement with the theoretical predictions reported by Milner,¹⁰⁹  which were made just after the first experimental findings.¹⁰⁷ 

Gido’s and our group¹¹⁰  have demonstrated the discrepancies between theoretical predictions and experimental results for A_nB miktoarm stars as the value of n increases. More significant deflections from the theory for (PI)₅PS miktoarm star were reported, even when eliminating the formation of spherical and cylindrical morphologies in such highly asymmetric miktoarm stars. Based on the initial phase diagram introduced by Milner,¹¹⁰  Gido’s and our group proposed the phase diagram shown in Figure 1.4, where the discrepancies with the theoretical predictions are indicated with dark shaded symbols.

An inverse 4-arm miktoarm star block copolymer of styrene and isoprene (PS: 50% by volume) synthesized by our group and studied morphologically by the Thomas group showed an ordered bicontinuous double-diamond morphology⁴⁹  never seen before for the strong segregation limit of a neat copolymer (Figure 1.5).

The self-assembly of three-component ABC miktoarm star terpolymers is governed by two independent composition variables (φ_A, φ_B, φ_C = 1 − φ_A − φ_B) and by three interaction parameters (χ_AB, χ_AC, and χ_BC), whereas the block sequence is eliminated in such architectures (due to the existence of only one junction point through which all blocks are connected) in contrast to that of ABC linear triblock terpolymers. Furthermore, the star junction points are restricted to lie on periodically spaced parallel lines (not necessarily straight) defined by the mutual intersection of the different domains, as theory predicts.

In order to verify this prediction, the morphology of a series of 3-miktoarm star terpolymers comprised of PS, PI, and PMMA (SIM) were reported by Thomas’ and our group.^112,113  These samples were synthesized via anionic polymerization with a specific method of alternating the functionality of (PS)(PI)Si(CH₃)Cl in order to initiate polymerization of MMA,⁵²  as already described previously. In all cases, the materials exhibited distinct three-phase microdomain structures. In Figure 1.6, the resulting morphology schematics are given for these very interesting miktoarm star terpolymers of the (PS)(PI)(PMMA) type.

In the first case,¹¹¹  the PS and PMMA arms showed pronounced incompatibility towards PI, whereas the PS and PMMA arms were organized into a novel annulus-matrix structure. In the second case,¹¹²  the self-assembly results indicated that, in these (PS)(PI)(PMMA) miktoarm star terpolymers, the composition between the three different domains is favorable towards a structure that allows the junction points to be confined within lines where the three different types of microdomains intersect, verifying therefore the predictions by the constrains involved in the star-type architecture. Such an observation was the first one of its kind in the literature.¹¹³ 

Ikkala’s and our group have reported the first hierarchical smectic self-assembly in a miktoarm macromolecular chimera composed of two coil-like arms (PS and PI) and a mesogenic α-helical polypeptide arm [poly(ε-tert-butyloxycarbonyl l-lysine), PBLL].¹⁰⁵  The PBLL α-helices are packed within lamellar nanodomains, which leads to an overall smectic alteration of the rod- and coil-containing layers typically observed in rod-coil block copolymers. Furthermore, the coil-containing lamellae have an inner structure composed of PS and PI rectangular cylinders, leading to what we call a hierarchical smectic phase (Figure 1.7).¹¹⁴  Similar hierarchical self-assembly have been found in other miktoarm chimeras too.¹¹⁵ 

Miktoarm (a term coined by our group in 1992) star polymers prepared by anionic polymerization high vacuum techniques play an important role in polymer science. In polymer chemistry, polymerization methodologies have evolved towards complex architectures, including miktostar structures and, in polymer physics, they have inspired scientists to elaborate/perform theories/experiments on the influence of the architecture on the self-assembly of block copolymers. In addition, miktoarm stars have led to the discovery of a significant number of novel nanostructures, in solution and bulk, as potential candidates for nanotechnology applications, such as nanomedicine, nanolithography, etc.

N. H. wishes to thank his former Ph.D. students, without their enthusiasm and dedication the µ-star project would not have been possible: Prof. Hermis Iatrou, Prof. Marinos Pitsikalis, Dr Stergios Pispas, Dr Yiannis Tselikas, Dr Vasilis Efstratiadis, Professor Apostolos Avgeropoulos, Professor George Sakellariou, Dr Yiannis Poulos, Dr Stella Sioula, Dr Maria Xenidou, Dr Gabriel Velis, Dr Stella Paraskeva, Dr Panagiota Fragouli, Dr Paraskevi Driva. N.H. wishes also to thank his collaborators Dr G. Polymeropoulos and Dr K. Ntetsikas for the design of the structures comprised in the Schemes. All authors wish to thank the corresponding Universities of Athens, of Ioannina and King Abdullah University of Science and Technology (KAUST) for their continuous support.