Mechanistic Aspects of Copper-Mediated Living Radical

Chapter 11

Mechanistic Aspects of Copper-Mediated Living Radical Polymerization

Downloaded by CORNELL UNIV on August 8, 2012 | http://pubs.acs.org Publication Date: June 26, 2003 | doi: 10.1021/bk-2003-0854.ch011

Jeetan Lad, Simon Harrisson, and David M. Haddleton Department of Chemistry, University of Warwick, Coventry C V 4 7 A L , United Kingdom

A series of aminomethacrylate ((dimethylamino)ethyl methacrylate, (diethylamino)ethyl methacrylate and t-butylaminoethyl methacrylate) monomers and a series of methoxy[poly(ethylene glycol)] methacrylate macromonomers of different molecular weights have been copolymerized with methyl methacrylate under free radical and transition-metal mediated conditions. Significantly greater levels of comonomer incorporation were observed under transition-metal mediated conditions for all comonomers. This is attributed to complex formation between comonomer and catalyst, which is presumed to affect the reactivity of the monomer double bond.

148

© 2003 American Chemical Society

In Advances in Controlled/Living Radical Polymerization; Matyjaszewski, K.; ACS Symposium Series; American Chemical Society: Washington, DC, 2003.


149 Transition-metal mediated living radical polymerization, reported independently by Sawamoto (1) and Matyjaszewski (2) in 1995, has proved to be a remarkably efficient method of producing polymers with a wide range of functionalities and architectures (J). The polymerization is widely agreed to follow a free-radical mechanism, shown in Scheme 1, and containing as its key step the reversible abstraction of a (pseudo)halogen from the initiator or dormant polymer chain, creating a free radical which propagates via free-radical addition to monomer. Reactivity ratios, tacticities, and insensitivity to many functional groups are in most cases similar to those observed in free-radical polymerizations. Nevertheless, the presence of an additional step (activation/deactivation) and of an additional species (the metal complex) in the reaction may be expected to affect the mechanism to some extent. Certain monomers that contain donor atoms such as Ν or Ο are likely to coordinate to the catalyst, and with many catalysts, π-bonding of the monomer to the catalyst may be significant (4). Furthermore, a number of studies on similar transitionmetal mediated additions (5-7) have implied the involvement of a "caged" radical, which is constrained within the coordination sphere of the catalyst. ^act

R-X+

M "-Y/Ligand« — » t

Kteact

, R * + X-M n

\

n+1 t

-Y/Ligand

>/ \ k

p

k'N

termination Scheme 1. Literature mechanism of transition-metal mediated living radical polymerization. Under appropriate conditions, significant mechanistic differences can be observed between free-radical and transition metal-mediated polymerizations. For example, changing the polarity of the reaction medium has dramatic effects on the rate of polymerization, e.g. in the presence of oxyethylene groups (8), substituted phenols (9) and in aqueous (10) or ethylene carbonate (11) solution. More generally, while it is clear that transition-metal mediated polymerizations resemble conventional free-radical polymerizations, data obtained from freeradical polymerizations cannot be assumed to be applicable to transition-metal mediated polymerizations. In this contribution, we present the results of some recent experiments, which show significant mechanistic differences between the two types of polymerization. In a previous study, the reactivity ratio of M M A with poly(lactic acid) methacrylate was measured and found to be significantly different between conventional free radical and atom transfer radical polymerizations (12). In that work and a previous study (13) the difference in the reactivity ratios between the two types of polymerization was attributed to the large size of the



150 macromonomer: in conventional free radical polymerization, rapid growth of the polymer chain depletes the local concentration of macromonomer, which diffuses to the active chain end more slowly than the small comonomer. The much longer chain lifetimes associated with living polymerizations allow the macromonomer to diffuse to the active site, maintaining equality between local and bulk concentrations. However, monomers such as poly(lactic acid methacrylate) that contain donor atoms such as Ν or Ο are also potentially able to coordinate to the catalyst. Thus, the aim of this work was to determine whether the observed difference in reactivity ratios could be due to monomer coordination to the transition metal used in transition-metal mediated polymerization. This was investigated using low mass monomers such as (dimethylamino)ethyl methacrylate which should diffuse at approximately the same rate as the comonomer. In a further set of experiments, a series of methoxy[poly(ethylene glycol)] methacrylate monomers containing different numbers of repeat units were copolymerized with M M A under transition-metal mediated conditions to investigate the effect of increasing molecular weight on monomer reactivity.

Results and Discussion There have been several reports on reactivity ratios in A T R P and other transition-metal mediated polymerizations (14-16). These have generally concluded that the reactivity ratios are very similar to those observed in conventional free radical polymerization. Any differences are generally small compared to the precision of the data. Recently, simulations have shown that large (order of magnitude) differences in the rates of activation of the two monomers can produce apparent reactivity ratios that differ from those observed in free radical polymerizations, even though the underlying rate coefficients of homo- and cross-propagation are unchanged (17). Once again, however, these differences are generally small except in the very unusual case where both reactivity ratios are significantly greater than 1 (17). In the case of the methyl methacrylate-poly(lactic acid)methacrylate systems (12) the differences between conventional and transition-metal mediated polymerizations are large, and well outside experimental error. This report prompted us to investigate the reactivities of a range of small monomers that would be expected to coordinate with the catalyst. The monomers chosen were (dimethylamino)ethyl methacrylate ( D M A E M A , 1), (diethylamino)ethyl methacrylate ( D E A E M A , 2) and (/-butylamino)ethyl methacrylate ( T B A E M A , 3).


151

DEAEMA, 2


DMAEMA, 1

TBAEMA, 3 These monomers were copolymerized with a large excess of M M A , and the relative rates of consumption of the comonomers were monitored using N M R spectroscopy. Molecular weights of the resulting polymers were not measured, but A T R P polymerizations under similar conditions (100:1 monomencatalyst ratio) give polymers with number-average degrees of polymerization very close to the theoretical value of IOOJC (where χ is conversion). The use of a large excess of M M A allows the reactivity ratio of M M A (>*MMA) towards the comonomer to be evaluated using a simplified form of the copolymer composition equation, which can be readily integrated. The resulting, integrated equation is shown below (eq. 1): [M ]/[M ]o = ([M ]/[M ] ) 1

1

2

2

rI

(1)

0

in which M is the monomer in excess ( M M A in these experiments). This expression is equivalent to the more common, linear expression given in equation 2 (20). {

ln{[Mi]o/[Mi]} =η.1η{[Μ ]ο/[Μ ]} 2

2

(2)

Significant chain length dependence of the rate coefficients of propagation is generally observed in the first few propagation steps (18). This leads to problems in the application of the instantaneous form of copolymer composition equation to living polymerizations as at low conversions the polymer is dominated by oligomeric species. B y using an integrated form of the copolymer composition equation, it is possible to obtain results at conversions of 0.1-0.95 ( M - 1000 - 9500). Within this molecular weight range, the rate coefficients of propagation and hence the reactivity ratios are expected to be approximately constant and equal to their long-chain limits. n



152

1

Conversion(MMA)

1(b) Figure 1. Linear and non-linearfitsto experimental data (MMA-TBAEMA copolymerization by copper-mediated polymerization) using (a) equation 2. (b) equation 1. Open circles are experimental data, solid lines show linearfitto equation 2, dotted lines non-linearfitto equation 1.


153 In free radical polymerizations high molecular weight polymer is produced from the start of the reaction and reactivity ratios should be constant across the entire conversion range. While equation 2 is a convenient form in which to graph results, the transformation of experimental results ([M]/[M] ) into logarithmic expressions produces severe distortions in the experimental error structure, with the effect that data points are effectively given greater weight as the conversion increases. This is illustrated in Figure 1, which shows experimental data from a copolymerization of M M A and D E A E M A under A T R P conditions, with lines of best fit calculated using both linear (eq. 2) and non-linear (eq. 1) methods. It is clear from Figure l b that the linear method provides an excellent fit to the final data points, but at the expense of data at lower conversions. The non-linear method produces a line, which fits all data points equally well. This distortion also affects the estimation of errors in r which will generally be underestimated using the linear method. In this work the problem was avoided by the use of equation 1 in conjunction with non-linear least squares fitting (20). The results obtained are shown in Table I. It can be seen from the table that linear fitting can give erroneous results, particularly when the reactivity ratio to be measured is not close to unity. Figure 2 gives a graphical representation of the results, including 95% confidence intervals (obtained from the non-linear fitting procedure). There is a clear difference between the free-radical and transition-metal mediated copolymerizations. Unlike the MMA-poly(lactic acid)methacrylate case (12% this cannot be explained by differences in hydrodynamic radius between the comonomers, as these differences are minimal. The increased incorporation of the aminoethyl methacrylate ( A E M A ) monomers into the transition-metal mediated copolymers could potentially be explained by A E M A terminated radicals having an increased rate constant of activation, £ , compared to MMA-terminated radicals. It is by no means clear, however, that this would be sufficient to produce the large deviations observed here, even i f k of A E M A monomers were several orders of magnitude greater than that of


0

u

act

&ct

Table I. Reactivities of Coordinating Monomers (I-MMA) i n Free Radical and Transition-Metal Mediated Copolymerizations

M DMAEMA DEAEMA TBAEMA 2

0

Free radical Linear NLLS? 0.96(2) 0.96 0.98(1) 0.99 0.97(1) 0.98 0

Transition-metal mediated Linear NLL^ 0.74(3) 0.77 0.79(3) 0.86 0.69(3) 0.74 0

b

Estimated by linear fitting of eq. 2. Estimated by nonlinear least squares fitting of eq. 1. Figures in parentheses are standard errors in the final digit.


154 1.1 free radical

1 4 I

0.9 -

^

0.8 4 Φ 0.7 4 TM-mediated

Φ


0.6

DMAEMA

DEAEMA

TBAEMA n

ree

Figure 2. Reactivities of coordinating monomers (VMMA) l f radical (filled) and transition-metal mediated polymerizations (open), showing 95% confidence intervals.

M M A , as deviations from free-radical behavior according to this model appear to be small except when both monomer reactivity ratios are substantially greater than 1 (e.g. examples 1C, 3B, 3C and 2D in reference 17). Such a large A would make controlled polymerization of A E M A s by transition-metal mediated polymerization virtually impossible due to the high radical concentrations and termination rates that would ensue. Nevertheless, there are many reports of successful controlled polymerizations of these monomers (21-24). Hence it appears that the two explanations previously advanced to support differences in reactivities between A T R P and free radical copolymerizations are untenable in this case. Many reactivity ratios show solvent effects (25), but the addition of catalytic amounts of copper complex is unlikely to have a significant effect on the solvent polarity. Such solvent effects are most often seen when charge-transfer structures play a significant role in stabilizing the transition state of the cross-propagation reaction as in styrene-MMA or styrene-acrylonitrile this is unlikely to be the case for M M A - D M A E M A where the two double bonds are very similar. N M R spectra of mixtures of aminoethyl methacrylates and copper(I) bromide in ds-tolmnQ show clear shifts in the absorptions of protons close to the nitrogen atom as well as the downfield vinylic proton (cis- to the ester moiety), even in the presence of equimolar amounts of pyridyl methanimine ligand (Figure 3). No change was observed in the position of the IR absorptions of the carbonyl groups of the aminomethyl methacrylates in the presence of copper(I) bromide. act


In Advances in Controlled/Living Radical Polymerization; Matyjaszewski, K.; ACS Symposium Series; American Chemical Society: Washington, DC, 2003. » n

ι 1111 ρ

ι 11111

il H

11

III

a

H

1 1 1 1 1 » 11

ι^ι ι

c

1111

ppm

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5

ι| iιιH

μ

> λ

1 1 1 1 1 11

ill ι^ι

9.0 8.5

I

ί

II i

t

3.0

2.5

t

1

d

11 11 I ! 11 I I

Figure 3. *HNMR spectra ofN-propyl 2-pyridylmethanimine (top), DMAEMA (middle) and a mixture of both ligands with Cu(I)Br.

(2/2/1)

Ligand + D M A E M A + Cu(I)Br

/ Τ

S .

2



156 This is indicative of coordination to copper through the amine and ester groups of the aminoethyl methacrylate, as shown in Scheme 2. The structure shown is not definitive, and we are unable to explain why the vinylic proton cis to the ester group is shifted upfield, indicating increased shielding, rather than dowfield as might be expected from deshielding due to the copper. However, as this shift is only observed in the presence of copper (and not in mixtures of ligand and D M A E M A ) it is logical to attribute it to interactions between the proton and either the copper itself or other ligands. Further evidence of the ability of D M A E M A to form complexes with copper is provided by the ability of this monomer to undergo living radical polymerization in the absence of additional ligand - presumably in this case the monomer itself acts as ligand (26). Coordination may affect the monomer reactivity by altering the electronic structure of the double bond, or simply through a mass effect (the mass of the monomer-copper complex will be much greater than that of the monomer alone, and this should cause an increase in the pre-exponential factor of the rate constant) (27). It is well known that the presence of Lewis acids may affect reactivity ratios, particularly in the case of S T Y - M M A polymerizations (18, 28); it is possible that a similar effect is observed in this case, with copper acting as the Lewis acid.

n-Pr' Scheme 2. Possible structure of a DMAEMA /N-propyl 2-pyridyl methanimine / copper(I) complex. Trends in reactivities within the monomers appear to support this explanation, as the secondary amine, T B A E M A , shows the greatest deviation from free-radical reactivity ratios, while the most sterically hindered tertiary amine, D E A E M A , shows the least, suggesting that the extent of deviation is correlated with decreasing congestion around the nitrogen. This is difficult to explain except in the context of monomer-copper complex formation. The study was extended to a series of methoxy[poly(ethylene glycol)] methacrylates (4, 5) in order to investigate the effect of increasing molecular weight on reactivity. Similar results were seen as for the amino methacrylates, with higher levels of incorporation into the copolymer (lower r ) in transitionmetal mediated polymerizations than in free radical polymerizations for all molecular weights (Figure 3, Table II). MMA


5a: M = 475, η - 6 5b: M = 1100, η - 2 1 5c: M = 2 0 8 0 , n ~ 4 3

4:MW=188

n

n


n

Table II. Reactivities of Methoxy [Polyethylene glycol) Methacrylate Monomers ΟΜΜΑ) in Free Radical and Transition-Metal Mediated Copolymerizations Macromonomer

Free radical 0.98(1) 0.95(3) 0.93(2) 0.95(3)

4 Sa 5b 5c

Transition-metal mediated 0.85(3) 0.60(5) 0.76(4) 0.76(3)

1.1 1 0.9

I

0.8 4 0.7 0.6 0.5 0.4 if

Figure 3. Reactivities of methoxy[poly(ethylene glycol)] methacrylates as a function of number average molecular weight in free radical (closed circles) and transition-metal mediated polymerizations (open circles). Error bars show 95 % confidence intervals.


158 Within the transition-metal mediated results, there is a general trend towards lower Γ ΜΑ as molecular weight increases. This trend is not compatible with a diffusion control explanation, which predicts that increasing chain length will only affect reactivities in free radical polymerization. The results presumably reflect macromonomers' increased ability to coordinate to copper as the number of potential ligating groups are increased. This should be counterbalanced by the increased likelihood that the copper will be coordinated to donor atoms, which are too far away from the monomer double bond to influence its reactivity as chain length increases. The value of r decreases sharply from M = 188 (4) to M„ = 475 (5a) before increasing to what appears to be a stable value (5b and c) at longer chain lengths. This suggests that 5a, containing on average 6 oxyethylene repeat units, is close to the optimum length for complexing the copper catalyst in such a way that it remains sufficiently close to the monomer double bond to affect its reactivity. Μ

M M A


n

Conclusions We have shown that for a series of aminoethyl methacrylate monomers and a series of poly(ethylene glycol) macromonomers of different weights, the monomer reactivities towards methyl methacrylate differ significantly from those observed in conventional free radical polymerizations. This is attributed to complex formation between monomer and catalyst. It is likely that monomer reactivities in transition-metal mediated copolymerizations will differ significantly from those measured in free radical copolymerizations for a wide range of fiinctionalized monomers with the potential to undergo similar monomer-catalyst interactions.

Experimental

Materials. Methyl methacrylate and aminomethacrylate monomers (99%, inhibited with monomethyl ether hydroquinone) were obtained from Aldrich and passed over a column of activated basic alumina before use. Methoxy[poly(ethylene glycol)] methacrylate monomers (Aldrich) were used as received. Copper(I) bromide (Aldrich, 98%) was purified by a modification of the method of Keller and Wycoff (29). N-propyl 2-pyridylmethanimine was synthesized according to the method of Haddleton et al. (30). Azobis(isobutyronitrile) (ACROS Chimica, 98%) was recrystallized from methanol. A l l other chemicals were purchased from Aldrich or A C R O S Chimica and used as received.


159 Analysis. N M R spectra were obtained on a Bruker DPX400 spectrometer using

Mechanistic Aspects of Copper-Mediated Living Radical

Recommend Documents