Tube Continuous Flow Reactors - Wiley Online Library

Macromolecular Reaction Engineering

Communication

Preparation of Forced Gradient Copolymers Using Tube-in-Tube Continuous Flow Reactors Simon Saubern,* Xuan Nguyen, Van Nguyen, James Gardiner, John Tsanaktsidis, John Chiefari The preparation of forced gradient polymers has received considerable attention using batch reactors, while the preparation of usable quantities of forced gradient copolymers using continuous flow reactors has been hampered by the need to vary the composition of the monomer feedstock continuously during the reaction. A reactor that allows for addition of a monomer feedstock continuously at all points along the length of the reactor tubing allows for the preparation of forced gradient copolymers in continuous flow reactors, allowing for the scale-up and bulk preparation of these polymers. This study reports here the initial investigation of preparing forced gradient copolymers using the reversible addition–fragmentation chain transfer methodology in tube-in-tube continuous flow reactors.

1. Introduction Gradient copolymers comprise a gradual change of mono mer composition over the length of the polymer chain, usually commencing with mostly one monomer, moving to a mixture of two (or more) monomers, before finishing with a different monomer.[1–4] They therefore occupy a structure between the random (or statistical) copolymer, and the sharp change of composition obtained in block copolymers. Gradient copolymers are predicted to have improved physical characteristics, especially those related to surfactants and dispersants, but also for use in more diverse applications such as fiber optics, battery electrolytes, and lenses.[1,4–7] Gradient copolymers have been classified as either “spontaneous” gradient copolymers, where the gradient (change of composition along the chain length) arises Dr. S. Saubern, X. Nguyen, V. Nguyen, Dr. J. Gardiner, Dr. J. Tsanaktsidis, Dr. J. Chiefari CSIRO Manufacturing Bag 10, Clayton, VIC 3169, Australia E-mail: [email protected] Macromol. React. Eng. 2017, DOI: 10.1002/mren.201600065

© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

purely from the different reactivities of the two monomers, or they are classified as “forced” copolymers, where the ratio of monomers available to the propagating polymerization chain is driven externally to the reaction by introducing different monomer ratios over the course of the polymerization reaction.[4,8] In this laboratory and elsewhere, researchers have been investigating the use of continuous flow reactors as a low capital cost, small footprint means of manufacturing and modifying polymers, especially those belonging to the reversible-deactivation radical polymerization process known as radical addition–fragmentation chain transfer (RAFT).[9–17] This work has focussed on the preparation of homopolymers, random- and block-copolymers in a homogenous phase. While the preparation of forced gradient polymers has received considerable attention using batch reactors,[1–4,8,18] the preparation of usable quantities of forced gradient copolymers using continuous flow reactors has been hampered by the need to vary the real-time ratio of the monomer feedstock continuously during the reaction, and to sample the reaction to determine composition. Preparation of these polymers in continuous flow reactors would convey several advantages, such as, reduced

wileyonlinelibrary.com

DOI: 10.1002/mren.201600065

(1 of 6) 1600065


S. Saubern et al.

www.mre-journal.de

branching[9] and the ability to rapidly optimize the propRecently, flow reactor equipment vendor Vapourtec erties of a product by making libraries of that product, an released a reactor designed to control extremely rapid approach much used in the pharmaceutical sector.[19–21] and exothermic reactions by slowly adding reagent over the length of the reactor tubing (see Figure S1, Supporting We report here our preliminary investigation of preInformation). paring forced gradient copolymers using the RAFT methTheir reactor is a liquid–liquid tube-in-tube reactor, odology in continuous flow reactors in order to develop a where the inner stainless steel tubing is laser etched means of preparing libraries of these polymers for future uniformly along its length with 50 µm holes, providing a applications. porous inner tube through which monomer solution can In a batch reaction, the composition of unreacted be added at points downstream from the reactor entry monomer in the reactor can be continuously varied by point (Figure 1). The inner tube begins mixing its conchanging the feedstock composition in order to produce tents into the outer tube ≈6 cm from the start of the outer the gradient copolymer.[1–4,18] tube, and ceases mixing ≈86 cm from the end of the outer However, in a flow reactor, once the monomer feedtube. stock enters the (usually) narrow tubing of the reactor, As such, the mixing of the second reagent solution into that composition is the only composition available to the the reaction stream occurs before the start of the reactor propagating polymer chain until that proportion of the hot zone and therefore before the commencement of the feedstock emerges from the reactor. Any changes to polymerization reaction when a thermal initiator is used the composition of the feedstock at the entry of the (as is here). The second monomer is thus present from reactor do not affect the composition of available mont = 0 in a small but nonzero quantity. omer at points further down the reactor tubing. If the reactor is connected in reverse, the mixing of the Similarly, sampling of the reaction during the course of second reagent solution into the reaction stream comthe polymerization to determine monomer composition mences after a fixed distance into the reactor hot zone and in the forming polymer chains can be done at any time in a batch reactor, but only at the exit of a flow reactor. This then would only determine the monomer composition of the final polymer, without indicating the change of monomer composition as the polymer chain was built up. So for a batch reactor, the final polymer solution will consist of polymer chains all of which will have the same Mn and same gradient. However, in a flow reactor, where the monomer feedstock has been varied over the course of the polymerization, the earliest product emerging from the flow reactor will be a polymer consisting entirely of the first monomer, but then the polymer in the product stream will slowly change in composition to the final monomer mixture. Critically, the total polymer solution collected over the course of the reaction will thus consist of all possible random copolymer compositions, and not a gradient copolymer of uniform composition. What is needed is a reactor that allows for addition of a monomer feedstock continuously at all points along the length of the reactor tubing. In the case of free radical polymerizations, the Figure 1. Diagram showing “forward” flow in the reactor with the two reagent streams reaction must also be protected from mixing before the heated zone, and “reverse” flow where the polymerization can begin quenching by atmospheric oxygen.[10] in the first stream before the second stream is mixed in.

1600065 (2 of 6)

Macromol. React. Eng. 2017, DOI: 10.1002/mren.201600065 © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com


Preparation of Forced Gradient Copolymers Using Tube-in-Tube Continuous Flow Reactors

www.mre-journal.de

therefore after the commencement of the polymerization reaction when using a thermal initiator, which at that point contains only the first monomer. The second monomer is therefore present only from t > 0, forming a block-gradient polymer.[22] These reactor configurations would then allow for the preparation of a subset of forced gradient copolymers in continuous flow reactors, allowing for the scale-up and bulk preparation of these polymers. Herein, we investigate the use of only the first configuration.

2. Results and Discussion We chose acrylic acid (AA) and N,N-dimethylacrylamide (DMA) as an exemplar, as the reactivity ratios[23] of r1 = 0.36 and r2 = 0.35 should give close to a 1:1 consumption of monomers regardless of reaction time when preparing a random copolymer. Deviation from this 1:1 consumption would indicate forced gradient conditions. 2.1. Control As a control, we measured conversion of the two monomers by combining equal volumes of equimolar solutions of AA and DMA through a simple T-piece, and passing the combined reaction solution into a stainless steel reactor coil. This coil is a regular single-channel coil of diameter 1.1 mm, and of similar volume (15 mL) to the tube-in-tube reactor used to prepare gradient polymers. The reactor coil was heated to 85 °C. Each experiment consists of a plug-flow polymerization of 6 mL (3 mL of each monomer stock solution) conducted at different flow rates to mimic the sampling of a batch reaction at different times. Each monomer stock solution was passed through an in-line degasser (Knauer Smartline degasser) to deoxygenate the solutions before being combined in a simple T-piece and pumped through the reactor.

The DMA stock solution contained the initiator (VAZO68) and the RAFT chain-transfer agent (3-((((1-carboxyethyl)thio)carbonothioyl)thio)propanoic acid). Both stock solutions contained equal amounts of 3-(trimethylsilyl)1-propanesulfonic acid sodium salt as an internal standard. Monomer conversion was determined by integration of the cis-vinyl signal for each monomer in the 1H-nmr spectra compared to unreacted stock solution (t = 0). The results are shown in Table 1. The results in Table 1 indicate a near equal consumption of each monomer versus reaction time, as would be expected from two monomers with similar reactivity ratios. 2.2. Gradient Flow Using Tube-in-Tube Reactor (AA and DMA) The single channel reactor coil was then replaced with the tube-in-tube reactor coil. This coil consists of a porous inner tube (o.d. = 1.1 mm) of volume 4 mL inside an outer tube (o.d. = 2.2 mm) of volume 14.7 mL. The inner tube begins mixing its contents into the outer tube ≈6 cm from the start of the outer tube, and ceases mixing ≈86 cm from the end of the outer tube. A plug of 3 mL of each of the reagent solutions was pumped into the reactor (after degassing) giving a combined flow rate of 0.3 mL min−1 (residence time = 49 min) at 85 °C. Conversions measured by 1H-nmr were 68% for AA, 94% for DMA, clearly displaying a now biased monomer consumption. A repeat of the experiments in Table 1 at different flow rates (different residence times) and using the tubein-tube reactor are shown in Table S1 of the Supporting Information with mixed results. While continuing to show a biasing in monomer consumption, the conversion values did not follow the expected smooth consumption pattern with increasing reaction time. We concluded that this was due to the inherent problem of timing two

Table 1. Monomer conversions for single channel reactor coil at various flow rates for poly(AA-ran-DMA).

Entry

Flow ratea) [mL min−1]

t [min]

AA integral

DMA integral

%conv AAb)

%conv DMAb)

1

0.3

50

1.393

2.576

94.6

89.8

2

0.4

37.5

2.294

3.482

91.1

86.2

3

0.5

30

3.614

4.607

86.0

81.8

4

0.6

25

4.581

5.744

82.2

77.3

5

0.7

21.4

6.151

7.390

76.2

70.7

0

25.794

25.262

0

0

a)

Combined flow rate of both reagent feeds; b)Percentage conversions by 1H-nmr for 15 mL stainless steel reactor coil at various flow rates and 85 °C with VAZO-68 as initiator, and 3-((((1-carboxyethyl)thio)carbonothioyl)thio)propanoic acid as RAFT agent.



(3 of 6) 1600065


S. Saubern et al.

www.mre-journal.de

Table 2. Monomer remaining of fractions collected from poly(AA-grad-DMA) using tube-in-tube flow reactor.

ta) [min]

AA integral

DMA integral

%remaining AAb)

%remaining DMAb)

1

32.0

24.3

0

101.0

0.0

2

28.7

26.4

0

109.8

0.0

3

25.3

25.7

1

106.9

4.1

Fraction

4

22.0

22.7

2.25

94.4

9.1

5

18.7

16.1

3.7

66.9

15.0

6

15.3

11.3

6.7

47.0

27.2

7

12.0

7.3

8.6

30.4

34.9

8

8.7

5.8

10.5

24.1

42.6

9

5.3

5

13.4

20.8

54.4

10

2.0

3.4

13.8

14.1

56.0

0

24.048

24.624

a)Residence

time for leading edge of fraction. Fractions collected at 1 mL intervals (3 min 20 s per fraction); b)Percentage of monomer remaining by 1H-nmr for 14.7 mL stainless steel tube-in-tube reactor coil at 85 °C with VAZO-68 as initiator, and 3-((((1-carboxyethyl) thio)carbonothioyl)thio)propanoic acid as RAFT agent. Integrations were repeated several times to minimize the error resulting from overlapping signals and necessary baseline correction, and that a standard deviation of ±4% was obtained for the value of 101%.

different reagent streams to meet in the reactor at the same time.[24] While not a problem with a steady-state continuous flow synthesis, with a plug-flow synthesis this would result in nonoverlapping plugs with the plug not containing initiator passing through the reactor unreacted or only partially reacted. Instead, the two monomer solutions were pumped continuously into the reactor coil, repeating the experiment in Entry 1 of Table S1 in the Supporting Information, but in a true continuous flow mode. After the leading edge of the reagent stream had been heating for 32 min, the heating element was turned off. Because of the efficient thermal transfer in a flow reactor, the reaction temperature dropped quickly below 60 °C after only 2 min, and reached 40 °C after 4 min. The half-life of VAZO-68 at 60 °C is in the order of 1 month,[25] so the polymerization process is effectively halted at that point, as the rate of termination will now exceed the rate of propagation.[26] Pumping was continued, and 1 mL fractions were collected. Conversions were determined by 1H-nmr as above giving an average value for each fraction. The results are shown in Table 2 as remaining monomer as a function of time in the reactor heated zone (lower fraction numbers emerge from the coil first and have therefore been heated for longer periods). Although the reaction times for each fraction are only an approximate measure, these results show that the rapidly reacting DMA monomer is almost completely consumed by the time the reaction medium begins emerging from the reactor coil, and that fresh AA monomer solution is still being added to the polymerization at the far end of the tube.

1600065 (4 of 6)

Differential consumption of the two monomers is observed over the course of the polymerization process, indicating the formation of a gradient copolymer. However, as AA is still being added to the reaction as it emerges from the reactor coil, not all of the AA has had time to react and polymerize. So a second single channel coil needs to be fitted in series after the tube-in-tube reactor in order to allow for all of the monomer to be consumed. This will of course change the pressure profile in the first reactor, giving rise to a slightly different gradient in the copolymers. 2.3. Gradient Flow Using Tube-in-Tube Reactor (Other Monomers) These experiments were then repeated to form a gradient copolymer of DMA and methacrylic acid (MAA) again with a 1:1 molar feedstock ratio, and additionally with a single tube reactor (10 mL) fitted in series after the tube-in-tube reactor (14.7 mL) to allow remaining monomer to react. In the control experiment forming a random copolymer, the conversion for DMA was 100%% and MAA 66%, while the conversions obtained for the tube-in-tube reactor to form a gradient copolymer were 74% and 71%, respectively (Table 3). More extensive variation in monomer consumption could be seen with other monomer pairs. Feeding DMA into polymerizing vinyl acetate afforded conversions of 76% and 61%, respectively, compared to the equivalent random copolymer where conversions of 100% and 24% were observed. With butyl acrylate (BA) and 3-chloro-2-hydroxypropyl methacrylate (CHPMA), we initially saw no consumption




Preparation of Forced Gradient Copolymers Using Tube-in-Tube Continuous Flow Reactors

www.mre-journal.de

Table 3. Gradient copolymers prepared using the tube-in-tube flow reactor. Experiments in Table 3 used a tube-in-tube gradient reactor (14.7 mL) to form the gradient polymer, followed by a single tube reactor (10 mL) heated to the same temperature to act as a finishing reactor. An equimolar mix of the two monomers was used.

t [min]

Restricted M1a)

% conv M1b)

% conv M2b)

CTAc)

Id)

T [°C]

Solvent

Mne)

Đ

AA-DMA

83

AA

59 (94)

97 (90)

BM1429

VAZO-68

85

Water

12 874

1.17

DMA-MAA

83

MAA

66 (71)

100 (74)

BM1429

VAZO-68

85

Water

18 310

1.23

DMA-VA

83

DMA

76 (100)

61 (24)

BM1542

VAZO-88

100

EtOAc

18 555

1.10

BA-CHPMA

83

CHPMA

38 (40)

34 (0)

BM1542

VAZO-68

85

CH3CN

4599

1.57

Monomers

a) M1 = monomer 1 (restricted feed); M2 = monomer 2; b)Conversions for the equivalent random copolymer in brackets; c)BM1429 = 3-((((1-carboxyethyl)thio)carbonothioyl)thio)propanoic acid; BM1542 = 2-cyanobutanyl-2-yl 3,5-dimethyl-1H-pyrazole-1-carbodithioate; d) VAZO68 = 4,4′-Azobis(4-cyanopentanoic acid); VAZO88 = 1,1′-Azobis(cyanocyclohexane); e)Mn measured using PS standard, so absolute values are not corrected.

of BA when preparing the random copolymer under the conditions listed in Table 3, but 34% conversion of BA when feeding the CHPMA solution into the propagating polymer reaction.

3. Conclusions We have shown that the new tube-in-tube liquid–liquid reactor originally developed by Vapourtec to better control reagent addition in extremely exothermic reactions is suitable for preparing gradient copolymers in a flow reactor. With equimolar feed rates, significant variation of the monomer consumption can be achieved compared to the equivalent random copolymer. Better modelling of the reagent flow through the reactor would allow for the generation of a wider variety of gradients by varying flow rates and monomer ratios.

4. Experimental Section See the complete details of the Experimental Section in the Supporting Information.

Supporting Information Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgements: The authors are indebted to Duncan Gutherie of Vapourtec, UK, for early access to the liquid–liquid reactor. Thanks to Almar Postma and Graeme Moad for useful discussions. Received: October 26, 2016; Published online: ; DOI: 10.1002/ mren.201600065


Keywords: flow chemistry; gradient copolymers; radical polymerization; reversible addition–fragmentation chain transfer (RAFT); reversible-deactivation radical polymerization

[1] M. Kryszewski, Polym. Adv. Technol. 1998, 9, 244. [2] K. Matyjaszewski, M. J. Ziegler, S. V. Arehart, D. Greszta, T. Pakula, J. Phys. Org. Chem. 2000, 13, 775. [3] R. Wang, Y. Luo, B. Li, X. Sun, S. Zhu, Macromol. Theory Simul. 2006, 15, 356. [4] U. Beginn, Colloid Polym. Sci. 2008, 286, 1465. [5] Z. Zheng, X. Gao, Y. Luo, S. Zhu, Macromolecules 2016, 49, 2170. [6] Y. Chen, H. Chen, M. Feng, Y. Dong, Eur. Polym. J. 2016, 85, 489. [7] F. Jasinski, V. L. Teo, R. P. Kuchel, M. M. Mballa, S. C. Thickett, R. H. G. Brinkhuis, W. Weaver, P. B. Zetterlund, Polym. Chem. 2016, DOI: 10.1039/C6PY02062C. [8] A. Feldermann, A. Ah Toy, H. Phan, M. H. Stenzel, T. P. Davis, C. Barner-Kowollik, Polymer 2004, 45, 3997. [9] P. Derboven, P. H. M. Van Steenberge, J. Vandenbergh, M.-F. Reyniers, T. Junkers, D. R. D’hooge, G. B. Marin, Macromol. Rapid Commun. 2015, 36, 2149. [10] C. H. Hornung, C. Guerrero-Sanchez, M. Brasholz, S. Saubern, J. Chiefari, G. Moad, E. Rizzardo, S. H. Thang, Org. Process Res. Dev. 2011, 15, 593. [11] J. Chiefari, C. Hornung, S. Saubern, WO2012037596, 2012. [12] C. H. Hornung, A. Postma, S. Saubern, J. Chiefari, Macromol. React. Eng. 2012, 6, 246. [13] C. H. Hornung, X. Nguyen, G. Dumsday, S. Saubern, Macromol. React. Eng. 2012, 6, 458. [14] C. H. Hornung, X. Nguyen, S. Kyi, J. Chiefari, S. Saubern, Aust. J. Chem. 2012, 66, 192. [15] J. Chiefari, C. Hornung, A. Postma, S. Saubern, WO2013086585, 2013. [16] C. H. Hornung, A. Postma, S. Saubern, J. Chiefari, Polymer 2014, 55, 1427. [17] C. H. Hornung, K. von Känel, I. Martinez-Botella, M. Espiritu, X. Nguyen, A. Postma, S. Saubern, J. Chiefari, S. H. Thang, Macromolecules 2014, 47, 8203. [18] X. Li, S. Liang, W.-J. Wang, B.-G. Li, Y. Luo, S. Zhu, Macromol. React. Eng. 2015, 9, 409. [19] B. J. Reizman, K. F. Jensen, Acc. Chem. Res. 2016, 49, 1786. [20] C. M. Thompson, J. L. Poole, J. L. Cross, I. Akritopoulou-Zanze, S. W. Djuric, Molecules 2011, 16, 9161.


(5 of 6) 1600065


S. Saubern et al.

www.mre-journal.de

[21] A. Gioiello, V. Mancino, P. Filipponi, S. Mostarda, B. Cerra, J. Flow Chem. 2016, 6, 167. [22] O. V. Borisova, L. Billon, M. Y. Zaremski, O. V. Borisov, Polym. Sci., Ser. C 2015, 57, 86. [23] J Brandrup, E. H. Immergut, Eric A Grulke, Polymer Handbook, Wiley, New York 1999.

1600065 (6 of 6)

[24] H. Lange, C. F. Carter, M. D. Hopkin, A. Burke, J. G. Goode, I. R. Baxendale, S. V. Ley, Chem. Sci. 2011, 2, 765. [25] S. P. Vernekar, N. D. Ghatge, P. P. Wadgaoknar, J. Polym. Sci., Part A: Polym. Chem. 1988, 26, 953. [26] G. Moad, D. H. Solomon, The Chemistry of Radical Polymerization, 2nd ed., Elsevier Science, Amsterdam 2006.

