Procedures for Initiation of an ATRP Reaction
Selecting the correct conditions for initiation of an ATRP reaction is the obvious first step in any well controlled ATRP. Control over initiation is critical for the preparation of materials where exploitation of other aspects of controlled material synthesis is desired. As noted in the bullets above six different procedures have been developed for initiation of an ATRP together with two procedures discussing conditions for improving cross-propagation kinetics when chain extending from a macroinitiator and a discussion on conditions for conducting an ATRP from surfaces. The initiation procedures are listed in the order in which they were developed and do not indicate a preference for a standard procedure for initiation of an ATRP.
The procedure termed "halogen exchange" provides a tool that allows one to alter the order of addition of monomers to a sequential block copolymerization from that dictated by reactivity thereby allowing increased freedom in designing polymer architecture. This is a unique advantage of ATRP and cannot be used in SFRP or RAFT. However, with the development of procedures that use low levels of catalyst, halogen exchange cannot work and therefore another procedure developed to overcome this limitation: preparation of a copolymer in the second block.
Initiation should be fast and complete at low monomer conversion since this provides control over molecular weight, PDI, structure, functionality. When applied to large scale production initiation procedures should be robust, inexpensive, and provide a clean ATRP.
The procedure for "normal" initiation of an ATRP reaction employs a molecule, either a small molecule, macromolecule, or a functionalized surface, with one or more transferable atoms or groups, most often a (pseudo)halide that undergoes a one electron redox reaction with a transition metal catalyst in a lower oxidation state forming the reactive species. (1-7) In all of the published literature on ATRP this R-X molecule has been called the initiator, even though in contrast to a standard free radical polymerization initiator, this molecule is an inherently thermally stable entity and is incorporated in its entirety into the final polymer.
The added initiator R-X can be a mono functional initiator, a multifunctional initiator, i.e. it can either possess more than one initiating functionality or it can be used to introduce additional functionality into the alpha-chain end, it can be a macroinitiator (a polymer containing initiator site(s)), or initiators attached to a surface; either a particle, flat surface, fiber, or porous material.
Plot taken from CRP Consortium Meeting October 2006
In small scale reactions care should be taken to reduce the amount of dissolved oxygen in the system, since oxygen will oxidize the catalyst complex forming the redox conjugate, or persistent radical, and reduce the rate of reaction. (8, 9) Note, however, that the addition of a low level of the redox conjugate does provide complete control from the start of the polymerization, which increases the amount of added initiator incorporated into the final product, by reducing the fraction of termination reactions between low molecular weight species normally required to form the persistent radical. (2) This is important in commercial processes where cost of initiator may be a major concern. The selection of the appropriate radically transferable atom or group on the initiator and transition metal complex is important to provide control over the efficiency of the initiation reaction. (10, 11)
In a reverse ATRP transition metal complexes in the higher oxidation state (e.g. CuII complex) are added to the reaction and the ATRP initiator (R-X) and activator are generated in situ by reactions triggered by decomposition of conventional free radical initiators. (7, 9, 12-14) The components of the initial system are less sensitive to air, therefore the catalyst precursors are easier to handle, and the procedure is compatible with commercial processes. The initiation step does not proceed by activation of an alkyl halide with a Mtn/L catalyst, but rather by thermal decomposition of a conventional free radical initiator, such as AIBN. Once radicals are generated, either they react immediately with the higher oxidation state transition-metal complex to form the reduced transition-metal species and a dormant species (I* + X-Mtn+1/L forming I-X), or they react with monomer to form a propagating radical, I-P1*, which is then quickly deactivated by reaction with X-Mtn+1/L to form Mtn/L and a dormant species (I-P1-X).
In subsequent steps, the reduced transition-metal species, Mtn/L, reacts with the newly formed halogen terminated chains, as in a normal ATRP initiation/propagation process.
As noted, above the advantages of reverse ATRP include starting with the more stable transition metal complex, (which is particularly useful when one wants to use more active catalyst complexes that are easily oxidized, as in miniemulsion systems) (15) for the preparation of a range of linear copolymers with good "-chain end functionality.
The disadvantages of reverse ATRP are that it limits the terminal functionality remaining on the initiator residue ("-functionality) to that present as an additional functional group on the "I" residue of the standard free radical initiator, and limits the topology of the polymers that can be prepared. Only linear (co)polymers can easily be prepared.
Furthermore the molecular weight of the final copolymer cannot be independently adjusted irrespective of the activity of the transition metal complex, since the transferable atom or group on the growing polymer chain end is introduced as a ligand on the added catalyst.
Consequently a relatively high amount of catalyst (CuII-ligand) is required: [CuII-ligand]0/[I - I]0 = 1 - 1.5
SR&NI, was developed to overcome the problems "reverse" ATRP has with "-functionality, targeted MW, and more complex polymer architecture, which were amplified by a desire to use even more active catalyst systems in the reaction.
In SR&NI, a small amount of an active activating catalyst complex is generated by decomposition of a standard free radical initiator, such as AIBN, while the majority of the polymer chains are initiated from an added alkyl halide via a normal ATRP process. This allows very active catalysts to be added to the reaction in their stable form and the bulk of the polymer to be formed from the added alkyl halide initiator.
The following schematic is a summary of the normal ATRP and reverse ATRP initiation mechanisms, shown above, illustrating how both initiation procedures are used in SR&NI. The reagents shown in red are the reagents that are added to a SR&NI reaction. The first formed radicals drive the reverse ATRP initiation reaction where an active catalyst complex in the higher oxidation state is reduced to the activator state by reaction with the formed radicals (kdeact) but the bulk of the polymer chains are initiated by the normal ATRP initiation mechanism.
The degree of polymerization is predominately controlled by the concentration of initially added alkyl halide, as expressed in the following equation, where f is the initiation efficiency of the added free radical initiator.
SR&NI was initially developed for bulk polymerization using macroinitiators to prepare block copolymers. (16) However SR&NI was quickly adapted to miniemulsion systems where addition of the catalyst precursor as an oxidatively stable salt prior to sonication simplifies the laboratory procedure (17-21) and allows the preparation of block, star, graft and hybrid copolymers in heterogeneous media.
This initiation system is similar to a SR&NI ATRP in that it starts with alkyl halides as initiators and transition metal complexes in their oxidatively stable state (e.g., CuIIBr2/ligand) as catalyst precursors. Therefore the most active catalyst complexes can be added to the reaction in their stable state. However, instead of employing a conventional radical initiator to activate the catalyst complex as in "reverse" ATRP and SR&NI, a non-radical forming reducing agent is employed to generate the activator. Reducing agents such as tin 2-ethylhexanoate, (22) ascorbic acid, (23) or MAO (24) react with the oxidatively stable Cu(II) complex and generate the activator, i.e. the transition metal complex in its lower oxidation state.
The Activators are Generated by Electron Transfer (AGET) without involvement of organic radicals capable of initiating a radical reaction. Therefore an additional requirement for a successful AGET ATRP is that the reducing agents should be selected so that the reduction occurs without formation of intermediates or products that could form new initiators for an ATRP. Some reducing agents can also react directly with alkyl halides but exchange reactions are too slow to provide polymerization control.
AGET ATRP has a significant advantage over "reverse" ATRP and SR&NI ATRP described above because it provides a route for synthesizing pure tele-functional polymeric materials of any desired architecture. (22-25) All reagents that are directly added to the reaction are shown in red italics in the following scheme. These reagents are stable in the presence of oxygen.
Furthermore, since AGET ATRP can be also successfully carried out in miniemulsion (23) this is of potential commercial importance. All agents initially added to the reaction mix prior to dispersion by high shear forces, or sonication, are stable in the presence of air. The reducing agent can be added, at a controlled rate, after a stable miniemulsion is formed to provide a controlled rate of initiation, i.e. provide highly efficient initiation, and subsequently control the rate of propagation of the reaction, by continuously controlling the ratio of CuI to CuII.
An additional advantage of AGET initiation is that the reducing agent can be used to remove dissolved oxygen from the system and hence the reaction can be conducted in a limited amount of air. (25, 26) Furthermore it was the development of AGET ATRP that allowed extension of an ATRP micromulsion procedure to a true ab initio emulsion system, (27) and extension to inverse miniemulsion. (28)
Pure telechelic homopolymers and pure block copolymers were successfully synthesized via AGET ATRP in bulk and miniemulsion copolymerization reactions.
2-D chromatographic analysis of samples of some of the copolymers prepared by AGET ATRP (see second set of images below) clearly illustrate that we resolved the incentive for this development in initiating systems, i.e. the formation of homopolymer due to the action of a standard radical initiator in SR&NI systems. 2-D chromatographic analysis are particularly useful at showing the presence of minor components in a material prepared even in a well controlled CRP, note the presence of 1% of the coupled product in the following analysis as the reaction was driven to high conversion
2D chromatograms of a linear PMA macroinitiator (A) and the final block copolymer PMA-b-PS (B) synthesized by AGET ATRP in miniemulsion. The 1st dimension is HPLC under the critical condition for PS, and the 2nd dimension is GPC with PS standard as calibration. 
2D chromatogram of star block copolymer PMA3-b-PS3 synthesized using 3-arm trifunctional PMA macroinitiator a) by SR&NI ATRP in miniemulsion; b) by AGET ATRP in miniemulsion. 
These chromatograms show that AGET ATRP can produce the same spectrum of pure copolymers as a normal ATRP while additionally allowing the precursors of very active catalysts to be added to the reaction in their stable oxidative state.
In many ways Activators ReGenerated by Electron transfer is not just another way to initiate an ATRP but is a new way to run an ATRP with much lower concentrations of catalyst present in the system. When we considered the implications of the convenient AGET procedure for initiating an ATRP system, described above, where the activators are generated by electron transfer we realized that it should be possible to use the reducing agents to constantly regenerate the ATRP activator, the Cu(I) species, from Cu(II) species formed during termination processes, without directly or indirectly producing initiating species that generate new chains. (29)
The amount of Cu-based catalysts in atom transfer radical polymerization (ATRP) of styrene could therefore be reduced to a few ppm in the presence of the appropriate reducing agents such as FDA approved tin(II) 2-ethylhexanoate (Sn(EH)2) and glucose (29, 30), ascorbic acid, (31) hydrazine and phenyl hydrazine. (32) Furthermore, since the reducing agents allow starting an ATRP with the oxidatively stable Cu(II) species and the reducing/reactivating cycle can be employed to eliminate air or radical traps in the system it is included in the procedures for initiating an ATRP.
For example, styrene was polymerized by the addition of 5 ppm of CuCl2/Me6TREN and 500 ppm of Sn(EH)2 to the reaction resulting in a polystyrene with Mn=12,500 (Mn,th = 12,600) and Mw/Mn = 1.28. (33) ARGET ATRP has also been applied to polymerization from surfaces even in the presence of limited amounts of air. (34)
Generally, in an ARGET system it is desirable to add an excess of the ligand compared to the amount required to form the transition metal complex in order to compensate for competitive complexation by monomer/solvent/reducing agent that are all present in significant molar excess when compared to the transition metal.
Another advantage of ARGET ATRP is that catalyst induced side reactions are also reduced to a significant degree and it is now possible to drive an ATRP reaction to much higher conversion and prepare copolymers with much higher molecular weight (34, 35) while retaining chain end functionality. This has been confirmed by successful chain extension. (36)
The concept of Initiators for Continuous Activator Regeneration (ICAR) could simplistically be considered a "reverse" ARGET ATRP. In ICAR ATRP a constant source of organic free radicals works to regenerate the CuI activator which is otherwise consumed in termination reactions when catalysts are used at very low concentrations. With this technique, controlled synthesis of polystyrene and poly(meth)acrylates (Mw/Mn < 1.2) can be conducted with catalyst concentrations between 10-50 ppm, where its removal or recycling would be unwarranted for many applications, and the reaction driven to completion with low concentrations of a standard free radical initiator. (35)
Four ATRP catalysts with a broad range of KATRP values were examined in ICAR ATRP of styrene. They included the CuCl2 complexes of tris[2-(dimethylamino)ethyl]amine (Me6TREN), tris[(2-pyridyl)methyl]amine (TPMA), N,N,N",N",N"-pentamethyldiethylenetriamine (PMDETA), and 4,4'-di-(5-nonyl)-2,2'-bipyridine (dNbpy). ICAR ATRP of styrene was first conducted at low temperature (60 C) where organic radicals were produced by the slow decomposition of azobisisobutyronitrile (AIBN) (0.1 eq vs. ethyl 2-bromoisobutyrate (EtBrIB) initiator) in the presence of 50 ppm of CuCl2/L complexes. Interestingly, rates of polymerization differed by less than a factor of two amongst the reactions. This was initially surprising, given that values of KATRP, which govern radical concentration and the rate of polymerization under normal and SR&NI ATRP conditions, differ by more than four orders of magnitude among these four complexes. Additional experiments and kinetic simulations explored the possibility that (1) rates of polymerization and radical concentration under ICAR ATRP conditions are actually controlled by the rate of free radical initiator decomposition and that (2) relative CuI and CuII concentrations conform accordingly as dictated by the KATRP value. Since such a small amount of Cu catalyst is employed in ICAR ATRP, catalysts with large values of KATRP (high concentration of CuII) and fast deactivation rate constants will minimize this ratio, allowing for more even polymer chain growth and ultimately better control. Cu complexes with TPMA have a large value of KATRP with the model polystyrene chain end 1-(bromoethyl)benzene (~7.9*10-6 at 60 C). While the analogous KATRP value of the Cu/PMDETA complex is much lower (~ 5.9*10-8), the deactivation rate constant (kda) for Cu/PMDETA is larger than that of TPMA, which can compensate for the product of kda[CuII].
Therefore Me6TREN and TPMA are more suitable ligands than PMDETA and dNbpy in ICAR ATRP at low Cu catalyst concentrations. Simulations confirmed that the rate of polymerization in ICAR is governed by the rate of free radical initiator decomposition (as in RAFT) while control is ultimately determined by KATRP and the rate of deactivation (as in ATRP). (36)
The need for halogen exchange arises due to a mismatch in reactivity when crossing over from terminal polymeric secondary alkyl halides to add monomers that form tertiary alky halides. Therefore halogen exchange is recommended for the preparation of block copolymers when one is moving from a macroinitiator of lower activity, such as a styrene or an acrylate, to continue the polymerization with a monomer that forms a more reactive dormant species, such as a methacrylate or acrylonitrile. (37-45)
In a typical halogen exchange experiment, an A-Br macroinitiator is used but the catalyst is formed from CuCl rather than CuBr. Once the radical A* is formed in the first activation step, it can add to the double bond of the monomer B yielding the radical B*, as in the previous case. The newly formed radical can be deactivated by the CuII halide complex forming either a B-Br or a B-Cl type dormant species. If the halogen exchange reaction is efficient the majority of the radicals B* are converted to B-Cl dormant species, while most of the A-type dormant species remain in the A-Br form.
The rational behind "halogen exchange" is that the value of the ATRP equilibrium constant for alkyl chloride-type of (macro)initiators is 1-2 orders of magnitude lower than the alkyl bromides with the same structure. These C-Cl bonds are activated more slowly, and thus the rate of propagation is decreased with respect to the rate of initiation from the added macroinitiator, which effectively leads to increased initiation efficiency from the added macroinitiator and preparation of a second block with narrower polydispersity. A well-defined A-b-B block copolymer will result in the case when KATRPB,Br x kp > KATRPA,Br x ki.
The rational behind halogen exchange is that the value of the ATRP equilibrium constant for alkyl chloride-type of (macro)initiators is 1-2 orders of magnitude lower than the alkyl bromides with the same structure. These C-Cl bonds are activated more slowly, and thus the rate of propagation is decreased with respect to the rate of initiation, which effectively leads to increased initiation efficiency from the added macroinitiator and preparation of a second block with narrower polydispersity.
When the CuBr-based catalyst was used the faster activation of the newly formed polyMMA-Br combined with the more efficient addition of MMA-type radicals to MMA, compared to polySty-Br and Sty-type radicals, respectively, resulted in the preparation of copolymers with varying lengths of the polyMMA block mixed with unreacted macroinitiator. The use of CuCl/dNbpy as the catalyst significantly improved the cross-propagation kinetics and the MWD of the resulting copolymers was symmetrical. Moreover, the macroinitiator was practically completely consumed at relatively low MMA conversion.
This is shown clearly in the following comparison of molecular heterogeneity between a block copolymer prepared with and without halogen exchange using Gradient Elution Chromatography (GPEC) and 2-Dimensional Chromatography. The effect of efficient cross-propagation is clearly shown by the clean shift in molecular weight and narrow molecular weight distribution. The 2D chromatographic analysis show that in the copolymer obtained using halogen exchange no macroinitiator remained at 19% MMA conversion, while ca. 10% unreacted polySty-Br macroinitiator was left in the reaction after 22% MMA conversion when halogen exchange was not employed
The rate of halogen exchange depends strongly upon the nature of the ATRP catalyst and is higher when more active catalysts (catalysts high activation rate constant) are used.
Successful chain extension of a polyacrylate or polystyrene macroinitiators with a methacrylate in an ARGET and ICAR ATRP presents a problem, since halogen exchange obviously can not be used for systems with a low concentration of catalyst. The poor initiation efficiency of polystyrene and poly(butyl acrylate) macroinitiators during chain extension with methacrylates was resolved by employing a small amount of styrene as a comonomer in the second block. (36) Model studies with small molecule acrylate and styrene-based initiators determined that only 10 mol % of styrene was needed to provide sufficient control. Well-defined block copolymers poly(Bu acrylate)-b-poly(Me methacrylate-co-styrene) and polystyrene-b-poly(Me methacrylate-co-styrene) were synthesized using activators regenerated by electron transfer (ARGET) and initiators for continuous activator regeneration (ICAR) ATRP. For example, starting from poly(Bu acrylate) macroinitiator (Mn = 18 000, Mn/Mw = 1.12), a poly(Bu acrylate)-b-poly(Me methacrylate-co-styrene) (Mn = 39 100, Mn/Mw = 1.23) was obtained. Without styrene, a block copolymer with a bimodal mol. wt. distribution was formed: poly(Bu acrylate)-b-poly(Me methacrylate): Mn = 36 400, Mn/Mw = 1.70).
Computational simulations also indicated that the improved initiation efficiency was due to a ten fold higher concentration of CuI species in the copolymerization chain extension caused by lower KATRP of polystyrene chain ends.
ATRP has been conducted from a range of surfaces since the concept was first disclosed. (46) Because of their appearance these materials have been called polymer brushes.
As illustrated above polymer brushes have been formed by "grafting from" and "grafting to" inorganic particles (47-53) and flat surfaces. (54-58) The synthesis of organic/inorganic hybrid materials is an area of growing interest as the useful properties of disparate components can be combined into a single material.
This section will focus on "grafting from" reactions/conditions since "grafting to" (59) requires synthesis of block or functional copolymers comprising functional groups suitable for reaction with the target substrate which will therefore be addressed in the material synthesis sections.
Controlled/"Living" Radical Polymerization (CRP) has been demonstrated to be suitable for the preparation of organic/inorganic hybrid materials with varying structural complexity on nano-, meso- and micro-scopic dimensions. ATRP has been particularly successful, as inorganic particles and substrates can be easily functionalized with either initiating alkyl halides or block copolymers can be synthesized with segments that can attach to inherent surface functionality. Target applications include: surfactants, elastomers, opto/magnetic materials, sensors, reinforced ultra-thin films, bio-responsive materials and patterned surfaces.
Spherical Particles Organic/inorganic hybrid nanoparticles containing an inorganic core and tethered glassy or rubbery homopolymers or copolymers have been prepared by the ATRP of styrene and (meth)acrylates from colloidal initiators. SiO2-g-pSt hybrid nanoparticles have been prepared possessing molar masses of tethered pSt in the range of Mn = 5,000 to 33,000 g/mol and were characterized both in the solid state and in solution using respectively transmission electron microscopy (TEM) and dynamic light scattering (DLS). TEM images of SiO2-g-pSt particles revealed the formation of composite materials with interparticle spacing increasing with the increase in the molar mass of the tethered pSt chains. (51) Comparison of hydrodynamic radii (Rh) for hybrid nanoparticles of varying size determined by DLS in toluene, versus molar masses (Mn) of pSt chains cleaved from colloids determined by SEC, revealed a linear relationship. Such linear dependence of Rh vs. Mn is a strong indication that when the particles are dispersed in toluene, the tethered chains adopt highly extended conformations, presumably due to steric interactions caused by the high grafting density and the fact that toluene is a good solvent for polystyrene. (51)
AFM examination of spherical brushes formed by grafting n-butyl acrylate from silica particles deposited on mica surfaces show that the swollen brush has collapsed and that the hard silica core is surrounded by the soft halo of tethered grafted chains. (51)
Since the polymerization from the surface is being conducted by controlled polymerization processes it is relatively simple to isolate the particles and add them to a fresh monomer solution to form tethered block copolymers.
In order to minimize termination reactions and gel formation during a "grafting from" a particle surface polymerization in a "bulk" system one should add excess monomer and target a slow rate of reaction. In accordance with the persistent radical effect, adding significant levels of the deactivator to the contact solution is a convenient tool for slowing down the rate of polymerization and facilitating exchange between active radical and tethered dormant oligo/polymeric species. Often sacrificial initiator is added to the contacting reaction medium to both provide a means of following the reaction and to assist in controlling initiation from the surface. (55)
However in commercial applications this presents an additional expense since the "waste" polymer has to be removed from the contacting medium as solvent and monomer are recycled. In both ATRP and NMP control can be obtained by addition of the persistent radical to the contacting reaction medium and all polymer formed is tethered to the substrate minimizing waste. (56)
As noted one of the problems associated with use of multifunctional initiators in an ATRP is the impact of termination reactions on the properties of the final material. Termination can occur either inter-molecularly resulting in crosslinking, eventually resulting in gel formation, or intra-molecularly which affects functionality and distribution of tethered chains. Accordingly, reaction conditions should be selected to minimize termination reactions including:
Conducting the reaction in a miniemulsion also reduces the impact of inter-particle coupling and stops the formation of a macroscopic gel. (57, 58)
Flat Surfaces Modification of surfaces with thin polymeric films allows one to tailor surface properties such as wetability, biocompatibility, biocidal activity, adhesion, adsorption, corrosion resistance and friction. Polymers with reactive groups or segments can be prepared for "grafting onto" surfaces or functional groups can be attached to the surface for a more efficient "grafting from" approach. The properties of surfaces are addressed elsewhere. In this section, we primarily address "grafting from" tethered initiators.
The density of the tethered initiator can be varied over a wide range. The conformation of the final tethered polymer depends on the graft density, and hence initial initiator density. The nomenclature currently employed to describe the result of different graft density is:
The graft density obviously affects the morphology of the tethered (co)polymer chains.
As shown above, once the initiator coverage exceeds 20% there is complete coverage of the surface with tethered chains providing a relatively smooth film with occasional defects. At lower levels of initiator coverage the grafts were increasingly patchy but at >1% coverage they formed highly uniform, isolated domains of tethered materials indicating the presence of collapsed tethered chains on the surface.
Initiators for CRP processes have been attached to a variety of surfaces including silicon, gold and carbon black, using a variety of linking groups. (60-66) The conditions noted above for grafting from particles should be applied to grafting from surfaces
In our studies initiator density (Id) was assumed to be the same as the molar ratio of the ratio of the molecule with attached initiator functionality (nATRP) and blank tetherable functional molecules. Id = nATRP/nATRP + nBlank)
Grafting for "Everyone": Historically an ATRP, as in any radical polymerization process, had to be carried out in rigorously deoxygenated systems to prevent trapping of propagating radicals by oxygen. However, with the development of ARGET ATRP it is now possible to conduct an ATRP in the presence of limited amounts of air with a very small (typically ppm) amount of copper catalyst together with an appropriate excess of reducing agent. This technique was successfully applied to the preparation of densely grafted polymer brushes, poly(Bu acrylate) homopolymer, and poly(Bu acrylate)-block-polystyrene copolymer from silicon wafers (0.4 chains/nm2). This simple new method of grafting well-defined polymers does not require any special equipment and can be carried out in vials or jars without deoxygenation. The grafting for "everyone" technique is especially useful for wafers and other large objects and may be also applied to the synthesis of molecular hybrids and bioconjugates. (33)