Kinetics of Atom Transfer Radical Polymerization from PVDF
Macroinitiators
Tichelaar, M.; Voet, V. S. D.; Loos, K.; ten Brinke, G.; Department of Polymer Chemistry, Zernike Institute of Advanced Materials,
University of Groningen, 2012
B SC THESIS 2012
Kinetics of Atom Transfer Radical Polymerization from PVDF
Macroinitiators
Tichelaar, M.; Voet, V.S.D.; Loos, K.; ten Brinke, G.; Department of Polymer Chemistry, Zernike Institute of Advanced Materials,
University of Groningen, 2012
ABSTRACT: Strain-mediated magnetoelectric coupling strongly depends on the morphology of ferroelectric and ferromagnetic domains in composite materials. The unique self-assembling properties of block copolymers can be used to control the morphology. For this purpose, we aimed for the synthesis of ferroelectric poly(vinylidene fluoride) (PVDF) containing block copolymers.
Trichloromethyl- and phenyl methylchloride-terminated PVDF macroinitiators were used to initiate the atom transfer radical polymerization (ATRP) of styrene (S), 4-vinylpyridine (4VP) and tert- butylmethacrylate (tBMA). The use of these monomers facilitate the preparation of a nanoporous ferroelectric network, which can be backfilled subsequently by a ferromagnetic material. Furthermore, the kinetics were studied extensively to investigate the controlled behavior of these reactions.
1. Introduction
1.1 Multiferroics (magnetoelectrics)
Although pioneering research already started in the fifties and sixties, only the last decade much attention has been devoted to materials exhibiting both ferroelectric and ferromagnetic properties.1,2 These materials belong to the class of the magnetoelectrics and multiferroics.1
Ferromagnetic materials show spontaneous macroscopic magnetization even in the absence of an externally applied magnetic field.3 This magnetization is caused by the alignment of unpaired spins between different molecules and results from exchange interaction between neighboring atoms. When spin alignment occurs in randomly orientated domains, no bulk magnetization is observed (Fig. 1a).
However, these domains may be orientated along a common direction by applying an external
magnetic field, H. When the applied field is removed, a residual magnetization, Br, is left in the material (Fig. 1b).3 Furthermore, ferromagnets behave hysteretically (Fig. 1b), making them very useful for data storage. As can be retrieved from the hysteresis loop, reversal of the magnetic field induces opposite magnetization in the material, -Br. The two directions can be identified as ‘0’and ‘1’.4 Ferroelectrics behave according to a similar principle, but instead of spins, dipoles are aligned in certain domains. Ferroelectrics have a similar kind of hysteresis loop, where an external electric field orientates the domains resulting in a net polarization even after the electric field is removed.3 In the same way, the direction of polarization can be used for data storage.5
Since both ferroelectric and ferromagnetic materials posses interesting properties for data storage, their combination leads to new possibilities. For example, a four-state memory can be constructed, where both the direction of magnetization and polarization can be switched independently.6 Furthermore, the two properties may be coupled by the magnetoelectric effect, describing the electric (magnetic) response of a material to an applied magnetic (electric) field.1 This allows to write data using electric fields, while reading can be performed using magnetic fields.1
(a) (b)
Figure 1. Schematic representation of randomly orientated magnetized domains (a) and a typical hysteresis curve for a ferromagnetic material (b).3
α - phase
β - phase
γ - phase
Only few single phase multiferroics exist, however, two-phase systems are more common. Since ferroelectric and ferromagnetic materials are piezoelectric and piezomagnetic as well, magnetoelectric coupling is usually mediated by strain.2 For example, in a two-phase system a magnetic field will induce strain in the ferromagnetic domains of the material (piezomagnetism). Since the ferroelectric domains are in direct contact, the strain is therefore transferred to the ferroelectric material, inducing an electric polarization (piezoelectricity).2 Besides the above-mentioned data storage applications, such systems may find use in microwaves and magnetic field sensors.2
1.2 Magnetoelectrics based on Poly(vinylidene fluoride)
A lot of attention has been devoted to oxides possessing ferroelectric behavior, although other materials, including fluorides, contain this property as well.1 Partly due to the highly crystalline nature of poly(vinylidene fluoride) (PVDF, 50 – 70%), it exhibits excellent physical and electrical properties.
Besides ferroelectric behavior, it demonstrates piezoelectric and pyroelectric properties as well.7 Although five distinct crystalline phases (α, β, γ, δ and ε) exist, the aforementioned properties derive from the all-trans configuration (β – phase) (Fig. 2).7 The resulting crystal contains a high polarity, since the chain dipoles have a common direction and therefore do not cancel each other.8 Fortunately, the β-phase is frequently observed together with the less polar γ- and apolar α – phase (Fig. 2).7 Since PVDF has a high Curie temperature* of 195 – 197 °C 9, it has already been used in many applications today, like sonar arrays, electrolytes, crystal oscillators, Li-ion batteries, headphones and mechanical actuators.7 Even more interesting, magnetoelectric effects were reported for composites of PVDF and Terfenol-D.10
* The transition temperature above which the ferroelectric properties disappear.
Figure 2. Chain configurations of the most common crystalline phases of PVDF. Carbon, hydrogen and fluorine are represented by the grey, white and green spheres, respectively.
We propose to use magnetoelectrics derived from PVDF, that can be prepared using VDF containing block copolymers. The proposed route towards these novel materials is outlined in Fig. 3. Block copolymers are good candidates, since they arrange themselves into highly ordered structures (see next section). After self assembly, the second block (the sacrificial block) may be removed and a ferroelectric nanoporous polymer network can be obtained. Backfilling with a ferromagnetic material will result in the desired magnetoelectric composite.
As a requirement, the second block has to be easily removed or modified in order to create a nanoporous network. Two methods are well known: First, block copolymers containing poly(4- vinylpyridine) (P4VP) can be complexated in advance with low molecular weight amphiphiles, such as pentadecylphenol (PDP), which can be washed away after self-assembly, rendering a nanoporous network.11 Second, Poly(tert-butyl methacrylate) (PtBMA) containing block copolymers can be hydrolyzed, generating nanopores.12
1.3 Block copolymer self assembly
Block copolymers are macromolecules consisting of two or more constitutionally different blocks.13 Generally, two chemically distinct polymer segments are incompatible with each other and tend to phase separate. Since the two blocks are covalently linked, segregation can only occur on nanoscale.
This process is known as microphase separation. As a result, highly ordered structures may be obtained with feature sizes ranging from 10-6 to 10-9 m.14
Figure 3. Proposed route towards novel magnetoelectrics, using VDF containing block copolymers as a template.
The driving force for microphase separation is the reduction of contacts between dissimilar blocks (mainly enthalpic), although this is counterbalanced by entropic forces.15 A block copolymer at the interface contains less translational ‘freedom’ and has to adopt extended chain configurations, reducing the configurational entropy (Fig. 4). Since mixing entropy is inversely proportional to the length, N, increasing the length will favor phase separation.15
The incompatibility between blocks of distinct monomers (A and B) is expressed in the Flory-Huggins interaction parameter, χ 14:
χ depends on the number of nearest neighbor contacts (Z), and the interaction energies between similar (εAA and εBB) and dissimilar monomers (εAB). A positive value indicates a unfavorable interaction between monomer A and B and χ tends to be positive even when the chemical difference is very small.
Furthermore, χ varies inversely with temperature and hence, lowering the temperature will favor ordered structures.
Summarized, microphase separation occurs for block copolymers with a positive χ and sufficiently large N. Therefore, the order-disorder transition (ODT) is mainly dependent on the product of the two, χ·N. For symmetric diblock copolymers, the mean field theory predicts the critical point at χ·N =10.5.
Below this value, the system is under entropic control, therefore no phase separation occurs and a disordered state is obtained. Above the ODT, several morphologies may be found depending on N and the volume fraction, f, of the components in the block copolymer.
Fig. 5 presents a theoretical phase diagram for diblock copolymers17, along with the most commonly observed morphologies: Lamellar (L), Cylindrical (C), Spheres (S) and Gyroid (G). A more realistic phase diagram for polyisoprene-polystyrene (PI – PS) diblock
copolymers
18is also depicted in Fig.
5. Most notably are the differences in symmetry and the value for the critical point of χ·N. The non- equilibrium (metastabile) Hexagonally Perforated Layer (HPL) morphology was revealed as well.
Figure 4: Driving force for phase separation is the reduction of contacts between dissimilar blocks, however this is counterbalanced by entropic forces.16
These morphologies can be prepared from diblock copolymer systems, but a whole variety of morphologies may be found for block copolymers with more complex architectures, which can be obtained in an almost infinite number of ways by altering the number of blocks19, monomer type and topology (linear vs. branched)20.
1.4 Atom Transfer Radical Polymerization (ATRP)
Controlled Radical Polymerizations (CPR) are of great interest for the synthesis of well-defined polymers, due to the relatively low sensitivity to impurities and reactive groups. Furthermore, living polymerization techniques like anionic polymerization is much more demanding and not feasible to polymerize VDF. Atom Transfer Radical Polymerization (ATRP)21,22 is now one of the most employed CPR techniques in polymer chemistry, which can be used to prepare block copolymers by the macroinitiator approach.The general scheme for the reaction is depicted below, illustrated with the PVDF macroinitiator used in this study.
Scheme 1. Atom Transfer Radical Polymerization using a chloride terminated PVDF macroinitiator.
DMF, 110°C
Figure 5. Mean-field phase diagram for diblock copolymers17 (left) and a phase diagram for polyisoprene-polystyrene (PI – PS) diblock copolymers18 (right). Besides the Lamellar (L), Cylindrical (C), Spherical (S) and Gyroid (G) morphologies, a metastabile Hexagonally Perforated Layer (HPL) morphology is found for the PI – PS diblock copolymer.
The macroinitiator is involved in a redox equilibrium together with a transition metal species in its lower oxidation state.23 The transition metal species homolytically abstracts a chlorine atom from the PVDF macroinitiator, leading to an oxidized metal species and an active radical species.22 In the active state, monomer addition results in the formation of the block copolymer.
To achieve a controlled polymerization, it is desired to diminish termination reactions. Since termination and propagation are, respectively, second and first order in concentration of active radical species, lowering this concentration strongly affects the termination.
Initially, no radicals are present and consequently, considering the equilibrium, the rate of formation of the transient radical species equals the rate of formation of the oxidized metal species. Since bimolecular termination of the transient radicals is inevitable, their concentration will decrease, and consequently oxidized metal species are formed irreversibly. Usually, the termination at the beginning of the reaction does not exceed 5% of the total amount of growing chains.24 When the concentration of active radical species has been decreased to a certain extend (and that of the oxidized metal species has been increased), the regeneration of the dormant species dominates.25 This mechanism is known as the Persistent Radical Effect (PRE). The fast exchange between dormant and active chains is essential to obtain uniform growth. In addition, the initiation rate must be fast compared to the propagation rate, and a fast equilibrium needs to be reached. Under these conditions, provided that chain transfer does not occur, the polymerization will demonstrate a pseudo-living character.
1.4.1 Kinetics of ATRP
The rate law for propagation is given by:
where [P·] is the concentration of active radical species and [M] is the monomer concentration. When the concentration of active radical species is sufficiently low, the termination becomes negligible and we may assume that [P·] remains constant. The rate law rearranges to a pseudo-first-order rate equation (3) with a general solution as given by equation (4).
where and [M0] is the initial monomer concentration. [M] is determined from the conversion, p, using the following relation:
Considering a controlled reaction, a linear relationship is obtained indicating no termination occurs.
Furthermore, when no chain transfer occurs, a linear relationship between the molecular weight and
conversion should be obtained. Therefore, the molecular weight can be predetermined by choosing an appropriate ratio of initial concentrations of the monomer and initiator species ([M]0 and [I]0):
where represents the initial number average molecular weight, in this case the weight of the macroinitiator.
1.5 Aim of the research
Block copolymer self-assembly provides an excellent way to control the morphology and therefore the properties of the final magnetoelectric composite. This research is concerned with the synthesis of new VDF containing block copolymers using Atom Transfer Radical Polymerization (ATRP). It has been reported that tricholoromethyl-terminated copolymers of VDF and hexafluoropropylene (HFP) initiate the ATRP of styrene.26 Furthermore, PS-b-P(VDF-co-HFP)-b-PS and PS-b-PVDF-b-PS copolymers have been prepared from phenyl methylchloride-terminated macroinitiators.27
A similar strategy is adopted to prepare new VDF containing block copolymers. Both trichloromethyl- terminated (MI 1) and phenyl methylchloride-terminated (MI 2) fluorpolymers are used to initiate the ATRP of S, 4VP and tBMA. Both macroinitiators as well as the monomers are depicted in scheme 2.
Due to the chemical structure, the use of MI 1 leads to AB diblock copolymers, while the use of MI 2 results in ABA triblock copolymers. Furthermore, the kinetics of these reaction are investigated to determine whether the reaction is controlled and low polydispersities might be expected.
As mentioned before, 4VP and tBMA are good candidates to act as sacrificial blocks. However, also S can be used, since the amorphous PS domains can be selectively removed by a nitric acid etch.28
Scheme 2. PVDF macroinitiators and monomers used to prepare novel block copolymers.
Trichloromethyl-terminated PVDF macroinitiator (MI 1)
Phenyl methylchloride-terminated PVDF macroinitiator (MI 2)
Styrene (S)
tert-Butylmethacrylate (tBMA)
4-Vinylpyridine (4VP)
2. Experimental
2.1 Materials
Styrene (S, Acros, 99%), 4-vinylpyridine (4VP, Aldrich, 95%) and tert-butylmethacrylate (tBMA, Aldrich, 98%) were distilled from calcium hydride under reduced pressure, followed by a second distillation from dibutylmagnesium.
Copper chloride (CuCl, Acros, 99.99%), N,N,N’,N”,N”–pentamethyldiethylenetriamine (PMDETA, Acros, 99+%) and dimethylformamide (DMF, extra dry, Acros, 99.8%) were used as received.
Both macroinitiators were synthesized by Vincent S.D. Voet. MI 1 was synthesized in a pressure vessel by emulsion polymerization of vinylidene fluoride (VDF) using potassium persulfate (KPS) as initiator and chloroform (CHCl3) as transfer agent. The macroinitiator has a number average molecular weight of = 5.7 kg/mol and a polydispersity index (PDI) of 1.4. MI 2 was synthesized in a pressure vessel using 4-chloromethyl benzoyl peroxide as initiator. MI 2 has a number average molecular weight of = 16.1 kg/mol and a PDI of 1.3.
2.2 Characterization
The molecular weight and molecular weight distribution of the PVDF macroinitiators was determined by Gel Permeation Chromatography (GPC) in DMF with 0.01 M LiBr on a Viscotek GPCMAX equipped with model 302 TDA detectors using two columns (Pl – gel 5µ, 30 cm mixed – C, Polymer Laboratories). Calibration was performed using narrow disperse PS standards. GPC was not suitable to determine the molecular weight of the VDF containing block copolymers. Therefore, the monomer conversion and molecular weight was determined by 1H-NMR, recorded on a 400 MHz Varian VXR operating at room temperature. DSC was performed on a TA Instruments Q1000 with a heating rate of 10.0 °C/min ranging from -80 to 200°C. TGA was performed on a Perkin Elmer Thermogravimetric Analyzer TGA7 with a heating rate of 10.0 °C/min ranging from 20 to 900°C.
2.3 ATRP from a trichloromethyl-terminated PVDF macroinitiator
The ATRP of S, 4VP and tBMA was initiated by a trichloromethyl-terminated PVDF macroinitiator.
A typical polymerization procedure is as follows: to a dry 50 ml three-necked, round-bottom flask sealed with a rubber septum, desired amounts of MI 1 and CuCl were added. The specific reaction conditions are pointed out in the results and discussion section. After three vacuum/N2 cycles, dry DMF was added to dissolve the initiator. The light brown/green suspension turned immediately dark blue after addition of PMDETA, and finally dark green due to complexation between the metal and the ligand. Subsequently, the monomer was added, which resulted in a suspension in the case of 4VP (due to coordination of 4VP to the metal). The mixture was degassed by nitrogen with at least 4 freeze-pump-thaw cycles, and the flask was placed in an oil bath thermostated at 110°C. The mixture soon turned dark brown and became increasingly more viscous. At specific time intervals, a sample
(0.1-0.2 ml) was withdrawn from the reaction mixture using a degassed syringe. About half of the sample was used to determine the composition and hence the conversion using 1H-NMR. The remainder was precipitated in H2O/MeOH 1:1†, filtered, washed and dried overnight at 60°C in vacuo.
The molecular weight of the precipitate was determined by 1H-NMR. At the end, the aforementioned procedure was repeated, and brown products were obtained. PVDF-b-P4VP was colored significantly more dark, probably due to a higher copper contamination.
2.4 ATRP from a phenyl methylchloride-terminated PVDF macroinitiator
The ATRP of S, 4VP and tBMA was initiated by a phenyl methylchloride-terminated PVDF macroinitiator. A typical polymerization procedure is as follows: to a dry 50 ml three-necked, round- bottom flask sealed with a rubber septum, desired amounts of MI 2 and CuCl were added. The specific reaction conditions are pointed out in the results and discussion section. After three vacuum/N2 cycles, dry DMF was added to dissolve the initiator. The light yellow/green suspension turned immediately dark green after addition of PMDETA. Subsequently, the monomer was added, which resulted in a suspension for 4VP (due to coordination of 4VP to the metal). The mixture was degassed by nitrogen with at least 4 freeze-pump-thaw cycles, and the flask was placed in an oil bath thermostated at 110°C.‡ At specific time intervals, a sample (0.1-0.2 ml) was withdrawn from the reaction mixture using a degassed syringe. About half of the sample was used to determine the composition and hence the conversion using 1H-NMR. The remainder was precipitated in H2O/MeOH 1:1 (for 4VP ether was used), filtered, washed and dried overnight at 60°C in vacuo. The molecular weight of the precipitate was determined by 1H-NMR. At the end, the aforementioned procedure was repeated, and off-white products were obtained. Again, the 4VP containing block copolymer product contained a darker color compared to the other two polymers.
† The 4VP containing block copolymers did not precipitate in H2O/MeOH 1:1. Considering the first four samples, the H2O/MeOH was evaporated and the polymer was dissolved in DMF and reprecipitated in ether. The remaining samples were directly precipitated in ether.
‡ For S, the color initially remained green, becoming darker after 1 day and finally light brown and grey after 5 days. For 4VP, the color almost instantly turned dark brown, similar to the polymerization with MI 1. For tBMA, the color initially remained green, becoming darker over time.
DMSO H2O H – T
H – H T – T
3. Results and Discussion
Unfortunately, GPC cannot be used to determine the molecular weight and therefore information regarding the PDI cannot be obtained. However, kinetic analysis can be employed to determine whether the reaction is controlled. In the upcoming sections, the characterization of the macroinitiators is discussed. Subsequently, the kinetic data is presented for the ATRP of S, 4VP and tBMA, initiated by two different PVDF macroinitiators. Finally, the thermal properties of the ABA triblock copolymers are assessed.
3.1 Spectroscopic analysis of the macroinitiators
Both macroinitiators were analyzed by 1H- and 19F-NMR. Fig. 6 presents the 1H-NMR spectrum of MI 1. Peaks at 2.6 – 3.1 ppm and 2.2 – 2.4 ppm can be assigned to head-to-tail and head-to-head (tail- to-tail) structures in the VDF sequence. The presence of chlorine end groups is clearly identified by the multiplet at 3.5 – 3.78 ppm, which is due to –CF2 – CH2 – CF2 – CCl3 and –CF2 – CH2 – CCl3
structures.26 The presence of these end groups is vital for the controlled synthesis of block copolymers using ATRP. The peaks centered around 6.3 and 1.78 ppm correspond to H-terminated chain ends (– CH2 – CF2H and –CF2 – CF2 – CH3, respectively).26,29
An excess of CHCl3 relative to the monomer is used to initiate the polymerization through ·CCl3
radicals. However, the peak at 5.68 ppm indicates that some chains are initiated directlyby KPS.
Figure 6. 1H-NMR spectrum of MI 1 ( = 5.7 kg/mol; PDI = 1.4).
A A
B B C
C
The 1H-NMR spectrum of MI 2 is depicted in Fig. 7. Again, the signals at 2.6 – 3.2 ppm and 2.2 – 2.4 ppm correspond to head-to-tail and head-to-head (tail-to-tail) structures, respectively. The presence of halogen containing end groups are confirmed by the singlet and two doublets at 4.85, 7.62 and 8.02 ppm, respectively.27
In order to obtain well-defined self-assembled structures, it is desired that the synthetic route leads to the ABA triblock copolymer only, that is, the macroinitiator must contain two halogen end groups.
Although termination of fluoropolymers usually occurs by combination27,30, signals at 1.78 and 6.3 ppm suggest the presence of – CF2CH3 and – CF2H end groups. These may either result from chain transfer processes or derive from 1,5 hydrogen shifts (“backbiting”).31 The former mechanism terminates the main chain, while the latter process introduces small side-chain branching. Process optimization is currently performed to diminish – CF2CH3 and – CF2H end groups.
The multiplet at 4.63 ppm indicates that the decarboxylation of the 4-chloromethyl benzoyl peroxide initiator during initiation was not complete.32
The PVDF backbone structures are also identified on 19F-NMR spectra (Appendix 1). The signal at -92 ppm corresponds to the head-to-tail structure. Multiplets at -94.8, -113.8 and -116.1 result from defect structures in the polymeric chain (–CH2CF2 – CF2CH2 – CH2CF2 – CH2CF2–, –CH2CF2 – CH2CF2 – CF2CH2–, –CH2F2 – CF2CH2 – CH2CF2–, respectively).27
9 8 7 6 5 4 3 2 1 0
8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 1.9 1.7
ppm
A B
C D
F B
C E
E F
D A
H – T
H – H T – T
H2O DMSO
Figure 7. 1H-NMR spectrum of MI 2 ( = 16.1 kg/mol; PDI = 1.3).
3.2 Results for the ATRP of styrene
The progress of the ATRP of styrene initiated by the two different macroinitiators has been followed using 1H-NMR in d6-acetone. Samples withdrawn from the reaction mixture were analyzed to determine the conversion and molecular weight at specific time intervals. No quenching agents were added, since the decrease in temperature will prevent the reaction to continue. The conversion, p, has been calculated using the following relation:
where HP and Hm are the integration values corresponding to a single proton of the polymer-segment (PS) and the monomer (S), respectively. Fig. 8a presents a characteristic 1H-NMR spectrum obtained for the reaction mixture. The doublets at 5.2 and 5.8 ppm belong to single protons of styrene and therefore Hm is given by the average value. The remaining six protons of styrene are visible as narrow peaks in the region between 6.7 and 7.6 ppm. The two broad peaks between 6.0 and 7.6 ppm can be attributed to the five phenyl protons of the PS-block. Therefore, HP is calculated by integrating the area between 6.0 and 7.6 ppm, subsequently subtracting 6 times Hm and dividing it by 5.
Since the molecular weight (and polymerization degree (
)) of the PVDF macroinitiator is known, the relative intensities between PVDF and PS can be used to determine the molecular weight.
Fig. 8b depicts a characteristic spectrum for the precipitated samples and the molecular weight is calculated as follows:
1. Integrate the area around 7 ppm and divide it by 5 (single proton value of the PS segment).
2. Integrate the area around 2.4 and 2.95 ppm and divide the total value by 2 (single proton value of the PVDF segment).
3. Divide the single proton value for PS by the total of the single proton values of the PS- and PVDF-segments (mol fraction of PS, XPS).
4. The polymerization degree of the PS-segment (
) follows from:
5. The molecular weight of the block copolymer (
) is given by:
9 8 7 6 5 ppm 4 3 2 1 0
9 8 7 6 5 4 3 2 1 0
9 8 7 6 5 ppm 4 3 2 1 0
ppm (a)
(b)
Figure 8. 1H-NMR spectra of PVDF-PS block copolymers. S and PS signals are used to determine the conversion (a). The molecular weight of the block copolymer is calculated using the characteristic PVDF and PS peaks together with the determined molecular weight of the macroinitiator (b).
Figure 9. The kinetic plots for the ATRP of styrene initiated by MI 1.
Table 1. Specific reaction conditions for the ATRP of styrene initiated by two different macroinitiators at 110°C.
Initiator
[Initiator]
(M)
*[CuCl]
(M)
*[PMDETA]
(M)
*[Styrene]
(M)
*DMF (ml)
MI 1 0.015 0.30 0.30 4.0 5.0
MI 2 0.005 0.05 0.05 4.0 5.0
* Concentrations are calculated relative to the amount of DMF added
3.2.1 Kinetic analysis of ATRP from MI 1 with styrene
The kinetic plots for the ATRP of styrene initiated by MI 1, with reaction conditions as specified by Table 1, are depicted in Fig. 9. The concentration of propagating species remains constant throughout the polymerization and furthermore first-order kinetics with respect to the monomer is observed as indicated by the linear semi-logarithmic behavior. Note the decrease in initial rate, which is probably due to the persistent radical effect (PRE). Initially, a high concentration of radical species results in termination, thereby irreversibly creating an increased amount of deactivator species (Cu2+) needed to obtain control over the reaction. It is reported that the addition of a small amount of deactivator reduces the termination during the initial stage of the reaction, although the increased steady-state concentration of Cu2+ species affords lower polymerization rates.33 The relatively slow polymerization rate of this reaction (Fig. 9), however, may be caused by a relatively large steady-state concentration of Cu2+ species due to the PRE.
The molecular weight dependence on the conversion is presented in Fig. 10. The linear relationship with conversion indicates that negligible chain transfer occurs. With the absence of significant termination, the polymerization behaves controlled. The molecular weights determined from 1H-NMR agree reasonably well with theoretical values. This implies complete initiation.
0 50 100 150 200 250
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Conversion
Time (h)
0 50 100 150 200 250
0.0 0.2 0.4 0.6 0.8 1.0
ln([M0]/[M])
Time (h) R2=0.991
0.0 0.1 0.2 0.3 0.4 0.5 0.6 4000
6000 8000 10000 12000 14000 16000 18000 20000
Measured Theoretical Mn (g/mol)
Conversion
R2=0.999
Figure 12. The kinetic plots for the ATRP of styrene initiated by MI 2.
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
10000 20000 30000 40000 50000 60000
Measured Theoretical Mn (g/mol)
Conversion
R2=0.998
3.2.2 Kinetic analysis of ATRP from MI 2 with styrene
The molecular weight dependence on the conversion and the kinetic plots for the ATRP of styrene initiated by MI 2 are presented in Figures 11 and 12, respectively. The rate of the reaction is much larger compared to initiation with MI 1, reaching a conversion of 0.7 after 120 hours. Since different reaction conditions were employed (Table 1), further comparison of the results is difficult. However, the equilibrium is shifted towards the active radical side as follows from equation 2. Furthermore, a more pronounced decrease in initial rate is observed (Fig. 12.). The linear time dependence of ln([M0]/[M]) after the initial stage indicates that the contribution of termination reactions was minimal.
As stated above, the termination during the initial stage of the reaction might be reduced by addition of the Cu2+ species, while maintaining acceptable polymerization rates.
The molecular weight is linearly dependent on the conversion and significantly lower than predicted (Fig. 11). For example, at 50% conversion the theoretical weight exceeds the measured molecular weight by 28%. As mentioned before, the molecular weight of the macroinitiator has been determined
0 20 40 60 80 100 120
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Conversion
Time (h)
0 20 40 60 80 100 120
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
ln([M0]/[M])
Time (h)
R2=0.931
Figure 10. The dependence of the molecular weight on the conversion for PVDF-b-PS copolymers initiated by MI 1.
Figure 11. The dependence of the molecular weight on the conversion for PVDF-b-PS copolymers initiated by MI 2.
by GPC, calibrated using narrow disperse PS standards. Deviation between the measured and real molecular weight might explain the discrepancy. However, the linear relationships, as obtained above, indicate controlled behavior.
3.3 Results for the ATRP of 4-vinylpyridine
1H-NMR measurements of 4VP containing block copolymers were performed in d6-DMSO. The conversion was calculated in a similar manner as for the S containing block copolymers. A characteristic spectrum is depicted in Fig. 13a. The doublets at 5.49 and 6.1 ppm both correspond to a single proton of 4VP, therefore Hm (in equation 7) is given by the average value. The multiplet at 6.72 ppm can be assigned to the remaining non-aromatic proton of 4VP. The broad signals at 6.58 and 8.25 ppm both correspond to two aromatic protons in the P4VP segment. For convenience, the peak at 6.58 ppm has been used to determine HP by integrating the area between 6.35 and 6.95 ppm, subsequently subtracting Hm and dividing it by a factor 2.
Fig. 13b presents a 1H-NMR spectrum of a 4VP containing block copolymer used to determine the molecular weight, which was calculated as follows:
1. Integrate the area around 6.58 ppm and divide it by 2 (single proton value of the P4VP segment).
2. Integrate the area around 2.45 and 2.9 ppm and divide the total value by 2 (single proton value of the PVDF segment).
3. Divide the single proton value for P4VP by the total of the single proton values of the P4VP- and PVDF-segments (mol fraction of 4VP, X4VP).
4. The polymerization degree of the P4VP-segment (
) follows from:
5. The molecular weight of the block copolymer (
) is given by:
3.3.1 Kinetic analysis of ATRP from both macroinitiators with 4VP
4VP and P4VP can act as strong coordinating compounds. This may result in the formation of pyridine-coordinated copper complexes during the reaction, especially since the monomer concentration is much higher compared to the concentration of the ligand. The reduced effectiveness of the catalyst may result in lower reaction rates and higher polydispersities.34
9 8 7 6 5 4 3 2 1 0
ppm
9 8 7 6 5 ppm 4 3 2 1 0
P4VP (a)
(b)
Figure 13. 1H-NMR spectra of PVDF-P4VP block copolymers. 4VP and P4VP signals are used to determine the conversion (a). The molecular weight of the block copolymer is calculated using the PVDF and P4VP peaks together with the determined molecular weight of the macroinitiator (b).
For convenience, only the peak at 6.58 ppm was used to determine the molecular weight.
Figure 14. Kinetic plots for ATRP of 4VP initiated by two different macroinitiators;
M1 (a), MI 2 (b).
Table 2. Specific reaction conditions for the ATRP of 4VP initiated by two different macroinitiators at 110°C.
Initiator
[Initiator]
(M)
*[CuCl]
(M)
*[PMDETA]
(M)
*[4VP]
(M)
*DMF (ml)
MI 1 0.020 0.10 0.30 2.0 2.5
MI 2 0.005 0.05 0.05 4.0 5.0
* Concentrations are calculated relative to the amount of DMF added
0 50 100 150 200 250
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Conversion
Time (h)
0 50 100 150 200 250
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
ln([M0]/[M])
Time (h)
0 10 20 30 40 50
0.0 0.2 0.4 0.6 0.8 1.0
ln([M0]/[M])
Time (h)
Fig. 14 presents kinetic plots for the ATRP of 4VP, initiated by MI 1 and MI 2, with reaction conditions as specified by Table 2. Clearly, both reactions are uncontrolled. Especially, the observed behavior for the initiation with MI 1 (Fig. 14a), is unexpected and inexplicable.
However, the high viscosity might pose a problem to withdraw proper samples and obtain reliable values for the conversion. Compared to the ATRP of styrene (using similar conditions), a decrease of the conversion by a factor 2 was demonstrated, while the molecular weights remained constant (data not shown). Probably, due to an inhomogeneous reaction mixture, a higher concentration of monomer has been present in the sample compared to the monomer concentration in the reaction mitxure.
0 10 20 30 40 50
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Conversion
Time (h)
(a)
(b)
Figure 15. The dependence of the molecular weight on the conversion for PVDF-b-P4VP copolymers initiated by MI 2.
0.0 0.1 0.2 0.3 0.4 0.5 0.6
15000 20000 25000 30000 35000 40000 45000 50000 55000
Measured Theoretical Mn (g/mol)
Conversion
R2=0.996
Nevertheless, it can be concluded that the reaction does not demonstrate controlled behavior.
Furthermore, the reaction rate is relatively slow.
Unfortunately, difficulties arose during the molecular weight determination, due to low intensity signals in the 1H-NMR spectrum corresponding to the P4VP-segment (6.58 ppm). Therefore, the molecular weight is strongly influenced by the signal-to-noise ratio, and as a result no reliable values have been obtained.
For the ATRP of 4VP with MI 2 as initiator, different reaction conditions were selected (Table 2). A high polymerization rate at the initial stage of the reaction was observed, reaching a conversion of 0.41 in less than 5 hours (Fig 14b). The non-linear behavior in the pseudo-first-order approximation is most likely caused by termination reactions, indicating a high concentration of active radical species present throughout the polymerization. The high degree of termination results in a poor control over the reaction, and finally the polymerization rate levels off near 60% conversion. Decreasing the temperature generally results in decreased values for the propagation constant, kp, and therefore lower polymerization rates, however, it also reduces the atoms transfer equilibrium constant.24 The equilibrium will therefore shift more towards the dormant state, reducing the amount of terminations.
Due to the unexpected high rate of the reaction, the first sample has been withdrawn at a relatively high conversion (i.e. 0.41). Consequently, the molecular weights were measured over a short range of conversions (Fig. 15). However, once again the molecular weights are low compared to the theoretical values.
3.4 Results for the ATRP of tert-butylmethacrylate
The progress of the reaction was monitored using 1H-NMR in CDCl3. The conversion has been calculated in a similar manner as for the S and 4VP containing block copolymers. A characteristic spectrum is depicted in Fig. 16a. The doublets at 5.41 and 5.93 ppm both correspond to a single proton of tBMA. Therefore, Hm is given by the average value. The broad peak around 1 ppm can be attributed
to the methyl group in the PtBMA segment. The value for HP can be obtained by integrating this peak and divide the area by a factor 3.
The same methyl group was used to determine the molecular weight. A characteristic spectrum for a precipitated sample of the PVDF-PtBMA diblock copolymer is presented in Fig. 16b. A very low signal for the PVDF block was observed corresponding to the head-to-tail structure of the VDF sequence. As will be discussed in the upcoming sections, considering the triblock copolymerization, the PVDF signal was demonstrated to be absent. The molecular weights were calculated as follows:
1. Integrate the area between 0.5 and 1.2 ppm and divide it by 3 (single proton value of the PtBMA segment).
2. Integrate the signal around 2.73 ppm and divide it by 2 (single proton value of the PVDF segment).
3. Divide the single proton value for PtBMA by the total of the single proton values corresponding to the PtBMA- and PVDF-segments (mol fraction of PtBMA, XPtBMA).
4. The polymerization degree of the PtBMA-segment (
) follows from:
5. The molecular weight of the block copolymer (
) is given by:
3.4.1 Kinetic analysis of ATRP from MI 1 with tBMA
The reaction conditions for the ATRP of tBMA initiated by MI 1 were similar to the ATRP of 4VP (Table 3). The conversion versus time and the first-order kinetic plot are depicted in Fig. 17. The linear semi-logarithmic behavior indicates a constant number of radical species throughout the polymerization. However, a relative large decrease in initial rate is observed.
Table 3. Specific reaction conditions for the ATRP of tBMA initiated by two different macroinitiators at 110°C.
Initiator
[Initiator]
(M)
*[CuCl]
(M)
*[PMDETA]
(M)
*[tBMA]
(M)
*DMF (ml)
MI 1 0.020 0.10 0.30 2.0 2.5
MI 2 0.005 0.05 0.05 4.0 5.0
* Concentrations are calculated relative to the amount of DMF added
9 8 7 6 5 ppm 4 3 2 1 0
2.5 2.0 1.5 ppm 1.0 0.5 0.0
(a)
(b)
Figure 16. 1H-NMR spectra of PVDF-PtBMA block copolymers. tBMA and PtBMA signals are used to determine the conversion (a). For the reaction initiated by MI 1 it was possible to determine the molecular weight, however, only the head-to-tail sequence of the PVDF block was observed (b).
Figure 17. The kinetic plots for the ATRP of tBMA initiated by MI 1.
0 10 20 30 40 50 60 70
0.0 0.2 0.4 0.6 0.8 1.0
Conversion
Time (h)
0 10 20 30 40 50 60 70
0.0 0.5 1.0 1.5 2.0
ln([M0]/[M])
Time (h) R2=0.84
The molecular weight increases linearly with conversion, however, it exceeds the theoretical weight by orders of magnitude (Fig. 18). This indicates a very low initiator efficiency (Ieff = Mn,th/Mn,nmr). Various reasons may account for this behavior, as discussed below.
First of all, a fraction of tBMA might hydrolyze to methyl methacrylic acid (MMA), which may act as an inhibitor for deactivation of active radical species.35,36 It is suspected that metal carboxylates are formed at the expense of deactivator species.36 Consequently, a high concentration of radicals results in excessive termination. Furthermore, quartenization of the amine ligand may influence the coordination.36 Considering this behavior, however, the poisoning should be effective throughout the complete polymerization process, and this is not observed. After the initial stage, the polymerization behaves in a controlled manner, suggesting that the low initiator efficiency is caused by termination or side reactions during the initial stage of the polymerization.
An efficient initiator should demonstrate an observed initiation rate constant (ki
obs = Koeq ki) equal to or larger than the observed propagation rate constant (kp
obs = Keq kp)§.37 Probably, the Keq of the CuCl/PMDETA system with tBMA is relatively large and as a result the above-mentioned condition may not be satisfied. A slow establishment of the Cu2+ equilibrium concentration may result in a significant amount of termination. It has been reported that the rate of initiation relative to the propagation rate can be increased using mixed halogen systems (I – Br/CuCl). The use of the more labile Br end groups provides a faster initiation, while the rate of propagation is still governed by the more stabile Cl end groups. This approach has successfully been applied in the ATRP of tBMA initiated by PS macroinitiators.38 However, this requires the use of Br end-functionalized PVDF macroinitiators.
§ Koeq is the equilibrium constant for the activation of the initiator:
Keq is the equilibrium constant for the activation of the polymer chain:
ki and kp are the respective rate constants for the monomer addition in both reactions.
Figure 18. The dependence of the molecular weight on the conversion for PVDF-b-PtBMA copolymers initiated by MI 1.
Figure 19. The kinetic plots for the ATRP of tBMA initiated by MI 2.
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0 50000 100000 150000 200000 250000 300000
Measured Theoretical Mn (g/mol)
Conversion R2=0.997
As mentioned before, the low initiator efficiency might be caused by side reactions at the initial stage of the polymerization. Generally, side reactions are less pronounced at lower temperatures.24 And indeed, the ATRP of tBMA is commonly carried out at lower temperatures than adopted here.
3.4.2 Kinetic analysis of ATRP from MI 2 with tBMA
Similar results concerning the semi-logarithmic behavior are observed for the ATRP of tBMA initiated by MI 2 (see Table 3 for specific reaction conditions). Again, a decrease in initial rate is observed, followed by a linear time-dependence of ln([M0]/[M]) (Fig. 19).
Unfortunately, it was shown to be impossible to monitor the molecular weight development, since the PVDF signal did not appear in the 1H-NMR spectrum. The absence of the PVDF signal in the spectrum is most likely explained by a very low initiator efficiency, as observed for the MI 1 initiated ATRP of tBMA.
0 5 10 15 20 25 30
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Conversion
Time (h)
0 5 10 15 20 25 30
0.0 0.2 0.4 0.6 0.8 1.0
ln([M0]/[M])
Time (h) R2=0.967
Figure 21. DSC thermograms of PS-b-PVDF-b-PS (a) and PtBMA-b-PVDF-b-PtBMA (b).
Figure 20. TGA curves of PS-b-PVDF-b-PS (––), P4VP-b- PVDF-b-P4VP (––) and PtBMA-b-PVDF-b-PtBMA (––).
100 200 300 400 500 600 700 800 900 1000
0 10 20 30 40 50 60 70 80 90 100
PVDF-b-PtBMA PVDF-b-P4VP
Weight (%)
Temperature (oC)
PVDF-b-PS
3.5 Thermal analysis
The thermal properties of the triblock copolymers were examined by TGA and DSC. The PS-b-PVDF- b-PS copolymer is thermally stabile up to 370 °C, where the PS block starts to degrade (Fig. 20). The thermal decomposition temperature of the PVDF block was demonstrated to be 450°C. This is consistent with previous reported results.27 The P4VP-b-PVDF-b-P4VP copolymer is slightly less stabile as indicated by the lower onset weight-loss temperature of 310 °C. This is expected due to the lower stability of P4VP relative to PS.39 The first weight-loss step at 200 °C for the PtBMA-b-PVDF- b-PtBMA copolymer is due to the loss of the tert-butyl group and concomitant anhydride formation.40 The second weight-loss step corresponds to further decomposition of both PtBMA and PVDF blocks.
-80 -60 -40 -20 0 20 40 60 80 100 120 140 160 180 200 220 1
2 3 4
Heat flow (W/g)
Temperature (oC)
-80 -40 0 40 80 120 160 200
1.0 1.5 2.0 2.5 3.0
Heat flow (W/g)
Temperature (oC)
DSC analysis reveals only one Tg corresponding to the sacrificial blocks of S and tBMA containing triblock copolymers (Fig. 21). The absence of a glass transition for the PVDF-segment suggests a high degree of crystallinity. The melting of the PVDF segments is identified by the endotherm around 160°C. A bimodal peak is observed due to recrystallization from the melt.41 The baseline shifts around 0°C for the PtBMA-b-PVDF-b-PtBMA copolymer might arise from the presence of impurities in the sample.
(a) (b)
