Chapter 25
Copolymerization of n-Butyl Acrylate with Methyl Methacrylate and PMMA Macromonomers by Conventional and Atom Transfer Radical Copolymerization 1
1,2,4
Sebastian G . Roos , Axel H. E . Müller , and Krzysztof Matyjaszewski
3
1
Institut für Physikalische Chemie, Universität Mainz, D-55099 Mainz, Germany MakromolekuIare Chemie II, Universität Bayreuth, D-95440 Bayreuth, Germany Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, PA 15213 2
3
ABSTRACT: The reactivity ratios of n-butyl acrylate (nBuA) with methyl methacrylate (MMA) and ω-methacryloyl-PMMA macromonomers (MM) in conventional and atom transfer radical copolymerization (ATRP) have been determined. For the copolymerization of nBuA with M M A , good agreement of the ratios is observed between conventional and controlled radical copolymerization, indicating that chemoselectivities in both processes are similar. The relative reactivity of the M M (1/r ) in conventional copolymerization is significantly lower than of M M A . It depends on the concentration of the comonomers but is not significantly influenced by the length of the M M . At high concentrations the relative reactivity decreases due to diffusion control of the M M addition. In ATRP the relative reactivity of the M M is much closer to the value of M M A . This is explained by the different time scales of monomer addition in both processes: whereas the frequency for monomer addition is in the range of milliseconds for conventional polymerizations, it is in the range of seconds or minutes in ATRP, thus diffusion control is less important here. This gives the opportunity to copolymerize at much higher concentrations than in conventional radical copolymerization. In addition, two-dimensional chromatography shows that the graft copolymers obtained by ATRP are much more homogeneous in terms of MWD and number of side-chains. nBuA
Graft copolymers offer all properties of block copolymers but are usually easier to synthesize. Moreover, the branched structure leads to decreased melt Corresponding author.
© 2000 American Chemical Society
361
362 viscosities which is an important advantage for processing. Depending on the nature of their backbone and side chains, they can be used for a wide variety of applications, such as impact-resistant plastics, thermoplastic elastomers, adhesives, compatibilizers, and polymeric emulsifiers. The state-of-the-art technique to synthesize graft copolymers is the copolymerization of macro monomers (MM) with the low molecular weight monomers. It allows for the control of the polymer structure which is given by (i) the chain length of sidechains which is controlled by the synthesis of the macromonomer by living polymerization, (ii) by the chain length of backbone which can be controlled in a living copolymerization, and (iii) by the average spacing of the side chains which is determined by the molar ratio of the comonomers and the reactivity ratio of the low-molecular-weight monomer, \ k\\fk\2' However, the distribution of spacings may not be very easy to control due to the incompatibility of the polymer backbone and the macromonomers. In the past, we have used conventional radical copolymerization for the synthesis of poly(n-butyl acrylate)-grq^-PMMA and other acrylic graft copolymers. ' Obviously, a control of backbone chain length is not possible for this mechanism. Thus, we have used ATRP to synthesize the graft copolymers. It should be noted that conventional and controlled polymerizations lead to different copolymer structures. In conventional radical copolymerization the polymers show a chemical heterogeneity of first order. The chemical composition of different polymer molecules is different due to the short period of time needed to form a polymer and the shift of the comonomer ratio during polymerization. In a living polymerization all chains grow simultaneously with the same chemical composition but this changes during the polymerization leading to a heterogeneity of second order, i.e. a compositional shift within all of the chains. In order to control the structure of graft copolymers the reactivity ratios of the copolymerization of η-butyl acrylate (nBuA) and P M M A - M M in the ATRP were investigated and compared to those obtained in conventional radical copolymerization. For comparison, the same was done for the copolymerization of nBuA and M M A . 1
r =
2
3
Experimental Section 4
Reagents: The purification of reagents has been published elsewhere. Methacryloyl-terminated PMMA macromonomers (PMMA-MM) of low polydispersity carrying a U V label were prepared by group transfer polymerization using a functionalized initiator. ^ Copolymerizations: All copolymerizations were conducted at a weight ratio nBuA MM · Conventional radical copolymerizations were performed in butyl acetate at 60 °C using AIBN as initiator; initial total monomer concentrations ranged from 4-33 wt-%. ATRP was performed in ethyl acetate at 90 °C using methyl α-bromopropionate as initiator, CuBr as catalyst, and 4,4'di(5-nonyl)-2,2'-bipyridine (dNbpy) as ligand, initial total monomer concentration was 50%. Details have been reported elsewhere. Analysis: Comonomer conversions were determined from the reaction solution by G C using a poly(methylsiloxane) capillary column. Decane was used as an internal standard. The conversion of the macromonomer was determined by m
:w
=
2
4
363 4
GPC analysis of the copolymer. P M M A and PnBuA standards were used for GPC calibration. The graft copolymers were characterized by 2-dimensional chromatography. An HPLC and a SEC apparatus are connected by a dual-loop automatic injector. The HPLC was used under critical conditions of adsorption ( L A C C C ) for PnBuA: eluent, THF:acetonitrile (53:47 by weight); flow, 0.01 ml/min, 35 °C; column set: 25 cm χ 4 mm, RP 18 (YMC S 5 μπι), 120 A and 300 Â (reverse phase). The elute of the HPLC was collected in two sample loops of exactly the same size (100 μΐ) and was immediately injected onto the SEC. Conditions for SEC: THF as eluent at a flow rate of 2 ml/min, RT, column set: 2 χ 30 cm, 5 μπι PSS SDVgel, linear and 100 A. Detectors: TSP UV3000 diode array detector and PL EMD-960 evaporative light scattering detector (ELSD) at 40 °C with a gas flow of 3.5-5 1/min. The PSS 2D SEC software (Polymer Standards Service GmbH, Mainz) was used for collecting and evaluating the raw data. 6
7
Results and Discussion Determination of reactivity ratios of M M A and nBuA in A T R P : Using methyl 2-bromopropionate (MBP) as initiator and CuBr/dNbipy as catalyst, the monomer reactivity ratios were obtained by two different methods: the Jaacks method for high monomer feed ratios ([nBuA] /[MMA] = 8 and 0.1, respectively) and the Kelen-Tiidôs method and non-linear (NLLS) optimization with varying feed monomer composition. No differences were found to the Kelen-Tudôs values. In Table 1 the reactivity ratios obtained are compared with an average of the known literature data of conventional radical copolymerization of the mentioned monomers determined by the Kelen-Tudôs method and nonlinear optimization. The nonlinear optimization produces more accurate values. 8
0
0
9
1 0
Table 1: Reactivity ratios of the copolymerization of the conventional and controlled radical polymerization ofnBuA and MMA * conventional radical polymerization system ATRP NLLS Kelen-Ttidôs method Kelen-Tudôs; Jaacks NLLS 4 11-13 4 9 Ref. 0.30 ± 0.03 0.26 ±0.14 rnBuA 0.36 ±0.12 0.39 ± 0.01 1.79 ±0.18 2.15 ±0.37 2.07 ± 0.09 2.19 ±0.03 TMMA a
mean and standard deviation of six values
It is important to realize that the reactivity ratios for the conventional poly merizations were obtained at different temperatures and solvents. However, for this pair of comonomers the ratios are not expected to significantly depend on these parameters within the limits of experimental error. The ratios show a good agreement for both polymerization techniques. Using NiBr (PPh ) as catalyst, Moineau et a l . recently reported comparable NLLS reactivity ratios (** 1.7; r = 0.34). Thus, we conclude that the different mechanisms do not affect the propagation reaction. This means that the equilibrium between dormant and active species does not affect the selectivity of the growing radicals. Similar observations have been made with other comonomer pairs. 2
3
2
1 4
=
MMA
n B u A
1 8
364 Conventional radical copolymerization of nBuA with P M M A - M M : Figure 1 shows a typical molecular weight distribution of a PnBuA-g-PMMA graft copolymer and that of the P M M A - M M used.
ι—·—ι—«—ι—«—ι—«—ι—>—ι—'—ι—«—ι—«—ι— —I 1
16
18
20
22
24
26
28
30
32
34
V./ml
Figure 1: Apparent molecular weight distribution of PnBuA -g-PMMA graft copolymer (( )M = 158,000, PDI =3.1) and PMMA macromonomer used ((· · ·)Μ = 10,900, PDI = 1.19). n a p p
η
Figure 2 shows a typical time-conversion plot. The macromonomer is con verted faster than nBuA due to the methacryloyl end group of the M M . Figure 3 shows the accompanying Jaacks plot for determination of the relative reactivity, l/r BuAn
ι,ο
Ί
t/h
Figure 2: Time-conversion plot of the conventional radical copolymerization of nBuA (A) and PMMA-MM (M) in butyl acetate at 60 °C. [nBuA] = 0.77 mol/l, [nBuAJ /[MMJ = 170, [AIBN] = 5 JO' mol/l; MJMM) = 10,900. Q
3
(
0
Q
The dependence of the relative reactivity of the P M M A - M M on the total weight concentration of monomers, w BUA w MM> was studied. The reciprocal value of r (as determined by the Jaacks method) is equivalent to the relative b
nBuA
365 reactivity of the M M ( l / r = ^BuA-MiAiBuA-nBuA)- In all copolymerizations with nBuA the reactivity of the M M ( l / r < 1.6) is much lower than that of M M A ( 1 ^nBuA „A«3.310). nBuA
n B u A
r
R
1,5
Ί
-
,
n
( - M 1
X
M
)
Figure 3: Jaacks plot obtainedfromthe time-conversion plot in Figure 2. Figure 4 shows that the apparent reactivity of the M M initially increases with increasing total concentration of monomers (and thus total concentration of polymer). After reaching a maximum the reactivity decreases at very high concentration of monomers. Let us first discuss the apparent reactivity decrease at high concentration. A similar effect was observed by Radke and Muller in the copolymerization of P M M A - M M with M M A . It was attributed to the increased viscosity of the reaction solution leading to diffusion control of the mobility of the M M lowering its apparent reactivity. At very high concentrations (38 wt-% of M M in bulk nBuA) the resulting product was reported to be a polymer blend of nBuA and unreacted P M M A - M M , which was not incorporated into the backbone. The initial increase of the reactivity is in excellent agreement to earlier results on the conventional copolymerization of nBuA and P M M A - M M in toluene. Since the M M and the resulting graft copolymers are incompatible, the polymer coils will try to avoid contacts at low concentration. Thus, the actual concentration of macromonomers near the growing radical will be lower than its stoichiometric concentration, similar to the "bootstrap" model which Harwood used in order to explain deviations from the terminal model in the copolymerization of monomers with different polarity. However, in contrast to reports on the copolymerization of the ω-styryl macromonomer of poly(2,6dimethyl-l,4-phenylene ether) and M M A , no evidence for micelle formation could be found by viscosity measurements. With increasing polymer concentration the polymer coils start to overlap and the "bootstrap effect" will become less important than the viscosity effect. Using two macromonomers of different length it was shown that 1/r does not strongly depend on the M M chain length (see Figure 4). Further experiments showed that the molar ratio of comonomers also does not strongly effect the M M reactivity. 5
3
19
20
2 1
366
3f
2-X-*-MMA
2,82.4
\
H
Figure 4: Dependence of ΡMMA macromonomer reactivity on the total i concentration of monomers. Conventional radical copolymerization wit butyl acetate at 60 °C with [nBuA](/[MM] = 170 and M/MM) = 10900 and M (MM) = 5600 (A). ATRP with nBuA in ethyl acetate at 90 °C wit fnBuA]y/[MM] = 83 and M (MM) = 5600 (Y); (x) copolymerization wi MMA. 0
n
0
ι
n
1
1
26
1
1
1
28
1
1
30
32
1
i
1
1
34
36
1
1
38
V /ml e
Figure 5: GPC eluograms of the crude PnBuA-g-PMMA graft copolyme the pure graft copolymer (M = 47200, PDI =1.66) ( ), and macromonomer used (M = 5600, PDI = 1.16) (· · ·/ n a p p
n
Copolymerization by A T R P : nBuA and P M M A - M M were copolymerized by ATRP in ethyl acetate at 90 °C [nBuA] :[MM] :[MBP] :[CuBr] :[dNbpy] :[Cu] = 500:5.8:1:10:20:10). Metallic copper was added in order to reduce CuBr formed via termination reactions. The M M used had M = 5600. Figure 5 shows the GPC eluogram of the graft copolymer obtained. For the calculation of M and PDI of the pure graft 0
0
0
0
0
0
2
22
n
n
367 copolymer the unreacted UV-labeled macromonomer was numerically subtracted from the crude copolymer. Figure 6 show the conversions of the comonomers versus time. The macromonomer is first converted faster than nBuA due to its more reactive methacryloyl end group. The conversion of the M M stopped at 90 % due to the very high viscosity of the reaction solution (ca. 45% solid content) which reduces the mobility of the M M .
1.0-,
0.0 4 ο
,
, 5
,
,
,
10
,
,
15
τ 20
t/h
Figure 6: Time-conversion plot of the atom transfer radical copolymerization of nBuA (A) and PMMA-MM (M), synthesized with [nBuA] = 2.33 mol/l, [nBuA] ç/[MM] = 83 in ethyl acetate at 90 °C. 0
0
Xp
Figure 7: Apparent number-average molecular weight and polydispersity index versus average conversion for the time conversion plot in Figure 6.
368 Figure 7 shows the apparent molecular weights and polydispersities of the pure copolymers (after subtracting residual MM) versus the average conversion, W
°MM )· ê conversion the molecular weight raises linearly indicating a controlled copolymerization. However, the PDI increases slowly with conversion indicating some side reactions. Both termination and transfer to polymer ^' could contribute to this effect. Figure 8 shows the Jaacks plot obtained from the experiment of Figure 6. The relative reactivity of the M M ( l / r A 2.2) is significantly higher than that found in conventional radical polymerization at comparable concentrations (1/^nBuA « 1-2, Figure 4). It is much closer to that of M M A ( l / r A 3.2, Table 1). ( °nBuA**p.*B"A + w ° M M t x p , M M ) / 0 ° n B u A w
+
W
2
i
t
h
i n c r e a s i n
24
=
n B u
=
n B u
Figure 8: Jaacks plot obtainedfromthe time conversion plot in Figure 6. In conventional radical polymerization the lower reactivity of macromonomers compared to M M A was explained by the diffusion control of the reaction. One polymer molecule is built in less than a few seconds and the time interval between two consecutive monomer additions is in the range of milliseconds. Thus, a M M has only a short time to move to the reactive chain end. Especially at the higher monomer concentrations (up to 33%) leading to high viscosities at higher conversions this effect might become important. In ATRP it takes hours to build the same degree of polymerization due to the reversible deactivation. Thus, the time interval for monomer addition is in the range of seconds to minutes, leaving enough time for the monomer and macromonomer to move to the dormant chain end. Once the dormant chain end is converted to a radical the M M has the same chance to react than the low molecular weight monomer. Consequently, the relative reactivity is closer to the value for the low molecular weight model, M M A .
369 G P C and 2D-Chromatography. A comparison of Figures 1 and 5 reveals that the MWDs of the copolymers obtained in ATRP are significantly narrower (although not extremely narrow) than those obtained in conventional radical copolymerization. Recently, we were able to show by two-dimensional chromatography of the products that the polymers obtained in ATRP are much more homogeneous in structure than those obtained by conventional polymerization. The principle of this method is based on separation according to chemical composition by Liquid Adsorption chromatography ( L A C C C ) followed on-line by separation according to total size by SEC. It has been shown that L A C C C of block copolymers allows for the independent determination of the size of block Β under conditions where block A is not separated according to molecular weight because size exclusion and adsorption cancel each other.2$. In our case column and solvent conditions were such that separation according to the size of the PnBuA backbone does not occur. However, polymers with increasing number of P M M A arms elute at decreasing elution volume (SEC mode). Subjecting the eluate to on-line SEC separates graft copolymers with exactly one arm from residual macromonomers. 6
7
Fig. 9 shows a two-dimensional chromatogram of the graft copolymer obtained by conventional polymerization, indicating that four different species are present in the product. Integration of the peaks allows for quantitative determination of the composition and shows that we only find 63% of the desired product under these conditions.
•-f'fiBiiiA'NoTri ©polymer
5.00-
4.50 4.00-
3.50 3.00 2.50 2.00 11.OO
12.00
13.00
14.00
15.00
16.00
Figure 9: 2D chromatogram of PnBuA -g-PMMA obtained by conventional radical polymerization. The ordinate corresponds to LACCC elution volume, the abscissa to SEC elution volume. Graft copolymer (I; 63%), star copolymer (2; 17%), PMMA MM (3; 8%), PnBuA homopolymer (4; 9%). In contrast, ATRP leads to a >90% yield of graft copolymer (Fig. 10). Less than 1% P M M A M M remain in the reaction product, which promises better mechanical properties (as thermoplastic elastomer) of the product. As the
370 chromatogram is scaled linear, the peak of PMMA M M and in the same way star copolymer and PnBuA homopolymer peaks vanish in the signal noise.
5.00
—τ
PnBuA Homopolymer
4.SO 4.00 Star Copolymer
3.50 PMMA-MM
j
3.00
Graft Copolymer
/
2.50-
1
/
2.00 _
ι
10.00
•»"
11.OO
— r—
12.00
13.00
ι
14.00
ι
15.00
I
16.00
Figure 10: 2D chromatogram of a graft copolymer obtained by ATRP. Graft copolymer (1; 91% M = 72600, PDI = 1.8), star copolymer (2; 6%), PMMA MM (3; 1%, M = 5900, PDI = 2.1), PnBuA homopolymer (4; 1%). f
napp
n
Conclusions Both conventional and controlled radical polymerization result in comparable reactivity ratios for low molecular weight monomers where diffusion control is absent. This indicates that the selectivity of the corresponding radicals is independent of the mechanism. The relative reactivity of macromonomers, however, strongly decreases at high concentrations in the conventional radical copolymerization of nBuA and P M M A - M M , due to the diffusion control. In controlled radical copolymerization a much higher relative reactivity of the M M is observed at these concentrations. This is explained by different time scales for monomer addition leading to the absence or a much later onset of diffusion control in ATRP. Accordingly, copolymerization of macromonomers is facilitated in ATRP. As a result, lower polydispersities of the graft copolymers are observed. Acknowledgement: This work was supported by the Deutsche Forschungsgemeinschaft within the Sonderforschungsbereich 262.
References: 1) 2)
Schulz, G. O.; Milkovich, R.J.Appl. Polym. Sci. 1982, 27, 4773. Radke, W.; Roos, S.; Stein, H. M . ; Müller, Α. Η. E. Macromol. Symp. 1996, 101, 19.
371 3)
Roos, S.; Müllier, A. H . E.; Kaufmann, M . ; Siol, W.; Auschra, C. in: "Applications of Anionic Polymerization Research" ; Quirk, R. P., Ed.; Am. Chem. Soc.: Washington, DC, 1998; Vol. 696, pp 208. 4) Roos, S. G.; Müller, Α. H . E.; Matyjaszewski, K. Macromolecules 1999, 32, 8331. 5) Radke, W.; Müller, Α. H . E. Makromol. Chem., Macromol. Symp. 1992, 54/55, 583. 6) Roos, S. G.; Schmitt, B.; Müller, Α. H . E. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1999, 40 (2), 140. 7) Pasch, H.; Much, H.; Schulz, G.; Gorshkov, Α. V. LC GC international 1992, 5, 38. 8) Jaacks, V. Makromol. Chem. 1972, 161, 161. 9) Kelen, T.; Tüdös, F.; Turcsányi, B.; Kennedy, J. P. J. Polym. Sci., Polym. Chem. Ed. 1977, 15, 3047. 10) Dube, M . Α.; Penlidis, A. Polymer 1995, 36, 587. 11) Bevington, J. C.; Harris, D. O. J. Polym. Sci., Part B 1967, 5, 799. 12) Brosse, J.-C.; Gauthier, J.-M.; Lenain, J.-C. Macromolecular Chem. 1983, 184, 505. 13) Emelie, B.; Pichot, C.; Guillot, J. Makromolecular Chem. 1991, 192, 1629. 14) Moineau, G.; Minet, M.; Dubois, P.; Teyssié, P.; Senninger, T.; Jérôme, R. Macromolecules 1999, 32, 27. 15) Greszta, D.; Matyjaszewski, K. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1996, 37(1), 569-70. 16) Arehart, S. V . ; Greszta, D.; Matyjaszewski, K. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1997, 38(1), 705. 17) Matyjaszewski, K. Macromolecules 1998, 31, 4710. 18) Haddleton, D.; Crossman, M . C.; Hunt, K. H.; Topping, C.; Waterson, C.; Suddaby, K. S. Macromolecules 1997, 30, 3992. 19) Stein, H. M . , Diplomarbeit, Universität Mainz Mainz 1992. 20) Harwood, J. Makromol. Chem., Macromol. Symp. 1987, 10/11, 331. 21) Percec, V.; Wang, J. H. Makromol. Chem., Macromol. Symp. 1992, 54/54, 561. 22) Matyjaszewski, K.; Coca, S.; Gaynor, S. G.; Wei, M . ; Woodworm, Β. E. Macromolecules 1997, 30, 7348. 23) Ahmad, N . M . ; Heatley, F.; Lovell, P. A. Macromolecules 1998, 31, 2822. 24) Roos, S. G.; Müller, A. H. E. Macromol. Rapid Commun. , submitted. 25) Falkenhagen, J.; Much, H.; Stauf, W.; Müller, A. H . E. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1999, 40(2), 984.
