A detailed kinetic study of the RAFT polymerization of a bi-substituted acrylamide derivative: influence of experimental parameters
Introduction
RAFT mediated controlled radical polymerization (CRP) is based on an equilibrium between active and dormant species, achieved by a degenerative chain transfer process [1], [2], [3] (Scheme 1). The chain transfer agent (CTA) is typically a thiocarbonylthio compound (Fig. 1) whose efficiency depends on the nature of the R and Z substituents [3] as well as on the type of monomer.
Considering acrylamide and its derivatives, only a few monomers have been polymerized by the RAFT process to date, although this technique seems much more suited for such monomers than other CRP processes. In fact, the use of nitroxide compounds is very unusual [4], [5], [6] and ATRP [7], [8], [9] is not recommended due to some problems related to the inactivation of the catalytic system leading to limited conversion and MW, although some recent articles concerning dimethylacrylamide (DMA) showed some improved results [10], [11].
The first example of RAFT polymerization of an acrylamide derivative [2] dealt with DMA, polymerized in benzene in the presence of benzyl dithiobenzoate. Polydispersity indices (PDI) of 1.2 were obtained for MW exceeding 100,000 g mol−1 with, however, a conversion limited to 26%. Then, DMA was polymerized in the presence of dimethylthiobenzoyl thiopropionamide [12] up to very high conversion (>90%); however long polymerization times were necessary (3 days). This duration could be reduced to 3 h by performing the polymerization in water [13] at 80 °C in the presence of sodium 4-cyanopentanoic acid dithiobenzoate. Moreover, polyDMA (PDMA) brushes of 23,000 g mol−1 (PDI=1.4) were obtained by RAFT polymerization from functionalized silicate substrates [14].
Another acrylamide derivative, monosubstituted, N-isopropylacrylamide (NIPAM), was polymerized by RAFT in benzene, dioxane and methanol/toluene 1/1 mixture in the presence of benzyl [15a,c] and cumyl dithiobenzoates [15a] as well as in the presence of benzyl and cumyl dithiocarbamates [15b] and 1-phenylethyl phenyldithioacetate [15d]. In these cases however, PDI was greater than 1.2 for MW above 20,000 g mol−1. Two other mono-substituted acrylamide derivatives, sodium 2-acrylamido-2-methylpropane sulfonate (AMPS) and sodium 3-acrylamido-3-methyl butanoate (AMBA) [16], as well as a sulfobetaine derivative [17] were polymerized in aqueous media in the presence of 4-cyanopentanoic acid dithiobenzoate. Here, PDI remained lower than 1.3 for MW until 60,000 g mol−1. Concerning acrylamide itself, it was polymerized by RAFT in aqueous media in the presence of either a xanthate [18] or a dithioester [19], until MW values of 30,000 g mol−1 with PDI around 1.2–1.3. These various examples confirm the versatility of the RAFT process to synthesize well controlled water-soluble polyacrylamide-based polymers.
In a previous paper [20], we investigated the RAFT polymerization of a bi-substituted acrylamide derivative, namely N-acryloyl morpholine (NAM) (Fig. 1). This monomer leads to polymer chains soluble in a wide range of solvents and exhibiting several interesting properties [21], [22], [23], [24], [25]. For instance, polyNAM was used as a substitute for PEO in various kinds of biological applications [26], [27], [28], [29]. In addition, we had previously studied the conventional free-radical homo- and copolymerization [30], [31] of NAM and it appeared particularly challenging to evaluate the potential of RAFT to polymerize NAM in order to yield polyacrylamide-based polymers with controlled MW and architecture.
First, the influence of the CTA structure on NAM polymerization was investigated by comparing several dithioesters [20], especially the only commercially available dithioester, carboxymethyl dithiobenzoate (CMDB) which bears a carboxylic function in the R group, and tert-butyl dithiobenzoate (tBDB, Fig. 1). This latter was synthesized by a one-step very convenient new biphasic process—based on a thioacylation reaction— leading to a very high yield. The performances of the dithiobenzoates were compared in terms of kinetics and molecular weights. Better control of NAM polymerization was obtained with tBDB compared to CMDB, with a linear increase of number average molecular weight, Mn, vs. conversion over the whole conversion range and with polydispersity indices below 1.1, as determined by aqueous size exclusion chromatography with on-line light scattering detection. However, at 60 °C, the polymerization duration was rather long (10 h).
Then, to further improve the RAFT polymerization of NAM, two strategies were developed. As we pointed out in our previous paper, the significant rate retardation (compared to a conventional polymerization) is due both to the lifetime of the intermediate radicals (IR) and to the loss of some propagating radicals consecutive to termination reactions onto the IR [20]. It was thus interesting to try to reduce IR lifetime. In one strategy, a phenyl dithioacetate was used as RAFT agent instead of a dithiobenzoate [32]. A significant improvement of both polymerization rate and polydispersity indices was obtained. In another strategy described here, we investigated if, for a given dithioester, changing the experimental conditions would increase the fragmentation reaction efficiency and thus would improve the polymerization control. RAFT polymerization of NAM in the presence of tBDB was studied in details, especially to demonstrate the influence of several experimental parameters on the kinetics and the control of MW, such as temperature, monomer concentration [M], CTA chain-length, dithioester to initiator molar ratio [CTA]/[AIBN] and monomer to dithioester molar ratio [M]/[CTA] [33].
To date, few systematic studies have been carried out with the RAFT technique to compare the influence of such experimental parameters on the polymerization of a given monomer/CTA pair, in terms of polymerization duration, limit of conversion, adequacy between experimental and calculated Mn values and polydispersity index. In addition, when dealing with the influence of [CTA]/[AIBN] ratio, it seems better to vary the initiator concentration while keeping the dithioester concentration constant, in order to avoid the possible side effects linked with a varying targeted Mn value. Finally, we also investigated the success of polyNAM chain-end functionalization, using two independent techniques, MALDI–TOF MS and 1H NMR spectroscopy, and performing a chain extension experiment.
Section snippets
Materials
N-acryloylmorpholine (NAM) (Aldrich, 97%) was distilled under reduced pressure (120 °C; 10 mmHg) to remove inhibitor. 2,2′-Azobis(isobutyronitrile) (AIBN) (Fluka, 98%) was purified by recrystallization from ethanol. 1,4-dioxane (Acros, 99%) was distilled over LiAlH4 (110 °C). Trioxane (Acros, 99%) and other reagents were all used without further purification. tert-butyl dithiobenzoate (tBDB) was synthesized according to a previously published process [20], [34]. After silica gel chromatography,
Influence of temperature
The first investigated parameter was the temperature with the aim of reducing the duration of the polymerization (runs 1 and 2, respectively, in Table 1). At the generally employed temperature, 60 °C, NAM polymerization in the presence of tBDB reaches 90% conversion in 10 h. At 90 °C, the same conversion is obtained in less than 1 h (Fig. 2). This strong increase in the polymerization rate originates from both the higher primary radical concentration at initial stages and the increase of all the
Conclusions
The RAFT polymerization of N-acryloylmorpholine (NAM) has been studied in detail, in the presence of tert-butyl dithiobenzoate (tBDB) as chain transfer agent. This dithioester, easily synthesized with a very high yield from the commercially available carboxymethyl dithiobenzoate, proved to be a good candidate to control NAM polymerization [20]. We have recently confirmed that tBDB is indeed very adequate for RAFT polymerization of hydrophilic [60] or hydrophobic [61] acrylamide derivatives,
Acknowledgements
The authors thank Catherine Ladavière (Unité Mixte CNRS/bioMérieux), Frédéric Delolme and Guy Dessalces (SCA/CNRS, Solaize, France) for the MALDI–TOF MS analysis, Marie-France Llauro (Service RMN de la Fédération des Polyméristes Lyonnais, Solaize, France) for the 1H NMR spectrum of the polymer, Thierry Delair (Unité Mixte CNRS/bioMérieux) for interesting discussions and Maël Bathfield (Unité Mixte CNRS/bioMérieux) for technical assistance with the chromatogram figures.
References (61)
- et al.
Macromol Chem Phys
(2001) - et al.
J Appl Polym Sci
(2003) - et al.
Macromolecules
(2003) - et al.
Macromol Rapid Commun
(2002) - et al.
Macromolecules
(1998) - Le TP, Moad G, Rizzardo E, Thang SH. PCT Int. Appl. WO 9801478 A1 980115. Chem Abstr...
- et al.
Polym Int
(2000)et al.Macromolecules
(2003)et al.Macromolecules
(2003) - et al.
Macromolecules
(1998) - et al.
J Am Chem Soc
(1999) - et al.
J Polym Sci, A: Polym Chem
(2002)
Macromolecules
Macromolecules
Macromol Rapid Commun
Macromolecules
Macromolecules
Macromolecules
Macromolecules
Macromolecules
Macromolecules
Macromolecules
J Am Chem Soc
Macromolecules
Macromolecules
Macromolecules
J Polym Sci, A: Polym Chem
Macromol Rapid Commun
Macromolecules
Macromolecules
Hebd Seances Acad Sci
Chim Med
Tetrahedron Lett
Polymer
J Chromatogr
J Pharm Sci
Cited by (63)
Novel anion exchange membranes based on quaternized diblock copolystyrene containing a fluorinated hydrophobic block
2018, Journal of Membrane ScienceCitation Excerpt :As is well known, the macro-RAFT agent to initiator molar ratio (mR/I) is a key parameter affecting the RAFT polymerization process [35]. Thus, high R/I ratios typically result in slow RAFT polymerizations and polymers with low polydispersity index (PDI) values [35,36]. For comparison, all the polymerizations were carried out at different mR/I ratios and polymerization times.