Kinetics of Multicomponent Polymerization Reaction Studied in a

May 16, 2012 - Dan Voicu , Milad Abolhasani , Rachelle Choueiri , Gabriella Lestari , Caroline Seiler , Gabriel Menard , Jesse Greener , Axel Guenther...
0 downloads 9 Views 464KB Size
Download PDF
Article pubs.acs.org/Macromolecules

Kinetics of Multicomponent Polymerization Reaction Studied in a Microfluidic Format Dan Voicu,† Clement Scholl,† Wei Li,† Dinesh Jagadeesan,† Irina Nasimova,‡ Jesse Greener,*,†,# and Eugenia Kumacheva*,†,§,⊥,# †

Department of Chemistry, University of Toronto, Toronto, Ontario M5S 3H6, Canada Department of Physics, Moscow State University, Moscow 119991, Russia § Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario M5S 3E5, Canada ⊥ Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Ontario M5S 3G9, Canada # FlowJEM, Inc., Toronto, Ontario M5S 3H6, Canada ‡

S Supporting Information *

ABSTRACT: We report a high-throughput study of the kinetics of a multicomponent polymerization reaction in a microfluidic reactor integrated with in situ attenuated total reflection Fourier transform infrared spectroscopy. The technique was used to study the kinetics of an exemplary free-radical polymerization reaction of N-isopropylacrylamide, which was initiated by ammonium persulfate in the presence of the accelerator N,N,N′,N′-tetramethylethylenediamine in water. By monitoring the rate of disappearance of the monomer double bonds, we determined the effects of the concentration of the monomer, initiator, and accelerator on the rate of polymerization and the effect of the pH of the reaction system on the reaction kinetics. This work opens the way for the kinetic studies of complex polymer systems in a microfluidic format. working in parallel (the “numbering up” strategy) offers high reproducibility and an increase in productivity.2 Continuous polymerization in flow microreactor systems has been demonstrated for radical polymerization of vinyl monomers,3 coordination polymerization,4 polycondensation,5 anionic,6 and ring-opening polymerization.7 Advantages of polymerization reactions in a MF format included excellent control of reaction conditions, the capability to vary the degree of polymerization by modulating the monomer-to-initiator ratio by changing their relative flow rates, and the ability to generate, in a high-throughput manner, large libraries of polymers for rapid evaluation. The potential applications of the MF platform for studies and optimization of chemical formulations for polymerization reactions would not be fully realized without rapid, on-chip characterization of reagents and products. Implementation of in situ chemical characterization opens the possibility for (i) the rapid screening of the effect of reaction variables, (ii) feedback for reaction control parameters, (iii) detection of transient species, which may not exist upon their removal from the MF reactor, and (iv) kinetic studies with sufficient time resolution, specifically during the initial stage of the reaction.

1. INTRODUCTION Exploring a broad range of experimental variables for multicomponent reactions is time-consuming and costinefficient. A systematic optimization of formulations for such reactions rapidly becomes challenging and time-consuming, especially when the parameter space includes concentrations of multiple reagents and the variations in temperature, pH, or pressure. Numerous reiterations are needed to evaluate and quantify the effect of superposition of multiple reaction parameters. Studies of chemical reactions in a microfluidic (MF) format provide an efficient experimental platform for the exploration of a large parameter space and optimization of formulations due to the ability to vary the concentrations of reagents in a highthroughput manner, especially when this is complemented by rapid on-chip characterization of reaction products. In MF synthesis, the type and the concentrations of multiple reagents introduced in the reaction system are controlled in a highthroughput manner by varying the ratio of their volumetric flow rates.1 Advantages of MFs as a chemical discovery platform also include reduced consumption of reagents, excellent control over heat and mass transfer, improved safety in handling dangerous species, and the ability to carry out multistep reactions without exposure to ambient conditions. Moreover, in comparison with the challenge in scaling up of conventional multireagent reactions, a combination of multiple MF reactors © 2012 American Chemical Society

Received: March 4, 2012 Revised: April 21, 2012 Published: May 16, 2012 4469

dx.doi.org/10.1021/ma300444k | Macromolecules 2012, 45, 4469−4475

Macromolecules

Article

7.26 phosphate buffer solution was prepared by dissolving NaH2PO4 in the predissolved aqueous reagent solution and titrating it with 1 M HCl to pH = 7.26. To prepare pH 9.50 buffer, ethanolamine was mixed with the aqueous reagent solution and titrated with a 10 M NaOH solution to pH = 9.50. The flow rate of each reagent solution was independently controlled using a separate syringe pump (PHD2000, Harvard Apparatus). All reactions were conducted at 21 ± 1 °C. An attenuated total reflection Fourier transform infrared (ATR-FTIR) spectrometer (Vertex 70, Bruker Corp.) was interfaced with the MF reactor using a single reflection diamond ATR crystal (MIRacle, Pike Technologies). Unless otherwise specified, the spectra were generated from 32 scans at 10 kHz scan speed with 4 cm−1 spectral resolution. Infrared absorbance of chemical species of interest was determined by measuring the intensity of an absorption peak and relating it to the molar concentration. Error bars were generated from the standard deviation from three separate measurements. The spectrometer was purged with air supplied from a purge gas generator (model 75-45, Parker Balston) to limit absorption by ambient CO2(g) and H2O(g). Detection was accomplished using a deuterated L-alanine-doped triglycene sulfate (DLATGS/D301, Bruker Inc.). Opus 6.5 software was used for computer-control of data acquisition and analysis. Measurements of pH were conducted using a VWR Symphony SB70P pH meter connected to a probe (MI408C, Microelectrode Inc.) with flow-through reference probe (ME16730, Microelectrode Inc.). Probes and fluidic connections were interfaced with a customized microfluidic reactor (FlowJEM Inc.) using an interface component (PCIC, FlowJEM Inc.).

Various characterization tools have been utilized for MF polymerization reactions. Continuous online size exclusion chromatography enabled monitoring of the molecular weight and molecular weight distribution of polymers synthesized in a tubular reactor by nitroxide-mediated polymerization.8 Multidetection gel permeation chromatography of polymer samples collected at the outlet of a microreactor was used for the characterization of linear and branched polymers by combining a concentration- and a mass-sensitive detection technique.9 In situ Raman spectroscopy was used to characterize with acceptable accuracy the change in composition and degree of conversion of methacrylate-based droplets in a MF reactor.10 Small-angle light scattering was utilized to monitor and characterize the formation and size distribution of multilamellar vesicles of a diblock copolymer in aqueous solutions.11 Here we present the results of the first in situ MF study of the kinetics of a polymerization reaction using infrared spectroscopy. This approach builds upon other FTIR-based analytical approaches to MFs,13 while maintaining the microchannel dimensions within the microprobe region. We conducted a throughput systematic study of the kinetics of a multicomponent reaction at different pH values with high temporal resolution by using a MF reactor interfaced with a Fourier transform infrared (FTIR) spectrometer. Fourier transform infrared spectroscopy is a well-established technique that is applicable to full-spectrum characterization of chemical species, which allows quantitative analysis of multiple reagents and products.12 Infrared spectroscopy can also be used to characterize solvent−analyte interactions, hydrogen bonding, changes to protonation states, and conformation of macromolecules, and effects of temperature or electromagnetic fields. These factors affect vibrational spectra by changing absorbance peak intensity and peak position or by causing subtle changes to spectral line shape, which can be revealed through second derivative spectroscopy.14 We examined an exemplary polymerization reaction: the polymerization of N-isopropylacrylamide in water, which was initiated by ammonium persulfate in the presence of the accelerator N,N,N′,N′-tetramethylethylenediamine. This important reaction is used for the preparation of gels in cell biology,15 gel electrophoresis,16 and the encapsulation of cells.17 We show the capability of MF synthesis integrated with infrared spectroscopy to rapidly examine the effects of varying the concentration of the monomer, initiator, and accelerator and the effect of pH of the solution on the kinetics of this polymerization reaction.

3. EXPERIMENTAL DESIGN A MF reactor was fabricated in polycarbonate, as described elsewhere.18 Fluid reagents were introduced into the MF reactor via a threaded interface system,19 which connected 1.6 mm poly(ether ether ketone) tubing (IDEX Corp.) to inlets on the reactor. Figure 1 shows the design of the MF reactor used in the present work. The reactor had a microchannel width and height of 200 and 50 μm, respectively. Unless specified, the cross-sectional area of the microchannel was 104 μm2. Aqueous solutions of APS, TEMED, and NIPAAm with initial concentrations CAPS,i, CTEMED,i, and CNIPAAm,i, respectively, and water were introduced into four syringes and supplied to the MF reactor through inlets (i)−(iv), respectively (Figure 1). The mixing between the reagent streams was carried out in a stepwise manner: first, by mixing the liquids supplied through inlets (i) + (ii) and (iii) + (iv) and then by combining two streams in the reaction chamber (the serpentine channel (v)), where the reagents mixed and the reaction took place. In the mixing channels the crosssectional area was reduced from 104 to 7.5 × 103 μm2, due to protrusions from the channel walls, introduced to enhance mixing of the reagents (Figure S1, Supporting Information). The changes in the reaction mixture were characterized after it flowed through the reaction chamber and reached the ATR crystal, a temperature probe, and a pH probe (the probes were located at positions P1, P2, and P3, respectively). A detailed description of the experimental setup, including integration of the ATR-FTIR, pH, and temperature probes with the microfluidic reactor, is described in detail elsewhere.20 The solution containing the polymer product and unreacted reagents exited the MF reactor through the outlet (vi).

2. MATERIALS AND METHODS N-Isopropylacrylamide (NIPAAm) was purchased from Sigma-Aldrich Canada and recrystallized prior to use. The initiator ammonium persulfate (APS), the accelerator N,N,N′,N′-tetramethylethylenediamine (TEMED), and ethanolamine were purchased from SigmaAldrich Canada and used as received. Deionized water was supplied by a Milli-Q Plus system (Millipore Corp.). Aqueous reagent solutions were prepared prior to experiments and purged with N2 gas for 15 min. Phosphoric acid was purchased from Caledon Laboratories, Ltd. Hydrochloric acid was purchased from VWR (Radnor, PA). Dihydrogen phosphate monosodium and sodium hydroxide were purchased from ACP Chemicals (Quebec, CA). For pH-dependent polymerization experiments, we used three buffer solutions with pH values of 2.23, 7.26, and 9.50. The pH 2.23 phosphate buffer solution was prepared by mixing H3PO4 with the predissolved aqueous reagent solution (either NIPAAm, TEMED, or APS) and titrating it to pH = 2.23 with a 1 M HCl solution. The pH

4. DATA ACQUISITION AND ANALYSIS 4.1. Reaction Time and Reagent Concentrations. The reaction time was calculated as t = DA /QT

(1)

where D is the distance between the beginning of the reaction chamber (v) and the ATR-FTIR probe (P1) (D = 160 mm); A is the average area of the cross section of the microchannel in the reaction compartment, A = 7.5 × 103 μm2 (7.5 × 10−3 mm2); and QT is the total flow rate of the reaction mixture. The value of QT is 4470

dx.doi.org/10.1021/ma300444k | Macromolecules 2012, 45, 4469−4475

Macromolecules

Article

Figure 2. (a) Characteristic IR spectra for independent reagents acquired for the aqueous solutions of (i) APS, (ii) TEMED, (iii) NIPAAm, and (iv) PNIPAAm, collected at QT = 3 mL h−1. Arrows mark characteristic “fingerprint” peaks chosen for subsequent intensity measurements. The groups responsible for the characteristic vibrations were (i) S2O82− (1270 cm−1), (ii) C−N (1020 cm−1), (iii) CC−H (975 cm−1), and (iv) N−H (1560 cm−1). All spectra were acquired in situ by averaging five spectra, each composed of 32 scans at 10 kHz. (b) Variation in concentration-dependent absorbance of the IR bands marked in (a). The inset shows the absorbance of the corresponding solutions at concentrations below 20 mM.

Figure 1. Schematic of the MF reactor. Solutions of APS, TEMED, and NIPAAm and water were supplied via inlets (i)−(iv), respectively. Four small wavy channels following the inlets were used to increase hydrodynamic resistance in order to stabilize flow and avoid cross-talk between the channels. Mixing of reagents occurred in a stepwise manner: first, between liquids supplied through inlets 1 and 2 and between liquids introduced via inlets 3 and 4, and then, by merging resulting streams in a T-junction just before the serpentine channel (reaction chamber (v)), where mixing was enhanced and the reaction took place. The composition of the reaction mixture was characterized by ATR-FTIR using a probe placed at point P1. A temperature and a pH probe were placed at points P2 and P3, respectively. The solution of the polymer product and unreacted reagents were evacuated from the MF reactor via outlet (vi). The average height and width of the microchannels were 50 and 200 μm, respectively. The scale bar is 1 cm.

QT = Q NIPAAm + QTEMED + Q APS + Q H2O

= 300 mM, which was the maximum on-chip concentration used in the present work (section 5.4). In the range of analyte concentrations studied, the intensity of an absorbance peak A was related to the analyte concentration by the Beer−Lambert law as −1

where C is the analyte concentration (mol L ), l is the path length (cm) of the evanescent infrared light in the solution, and ε is the molar extinction coefficient (M−1 cm−1). Molar extinction coefficients were determined by replacing A/C in eq 4 with the slopes of the calibration graphs in Figure 2b. We used the variation in absorbance of the double bond of NIPAAm vs t to measure the time-dependent concentration of the monomer during the reaction, that is, conversion. We calculated the penetration depth, lNIPAAm, to be 1.54 μm using the incident angle of light at the ATR crystal surface (60°), the indices of refraction of the liquid media and diamond ATR crystal (1.5 and 2.4, respectively), and the excitation wavelength of the band of interest (975 cm−1), the spectral position of NIPAAm characteristic peak.21 Using eq 4, we determined εNIPAAm to be 1.08 × 104 M−1 cm−1. 4.3. Determination of the Rate of Polymerization. We modulated the total flow rate of the reaction mixture to tune the reaction time within the interval 0 ≤ t ≤ 6.75 s, which limited the extent of conversion to less that 30% for all reagent concentrations. A relatively low conversion enabled us to avoid problems associated with increase in viscosity of the solution, due to the formation of high-molecular weight PNIPAAm, as well as polymer adsorption on the ATR crystal. We prepared three stock solutions with CNIPAAm,i = 525 mM, CTEMED,i = 180 mM, and CAPS,i = 180 mM and introduced them in the MF reactor. The manipulation of the flow rate of water (the fourth inlet) enabled controlled dilution of the reagent streams. For example, we used QNIPAAm = 0.235 mL h−1, QTEMED = 0.118 mL h−1, QAPS = 0.235 mL h−1, and QH2O = 0.118 mL h−1 (QT = 0.706 mL h−1) to calculate the reaction time t = 6.75 s using eq 1. Furthermore, by using eq 3, we determined initial on-chip reagent concentrations to be CNIPAAm,d = 175 mM, CTEMED,d = 30 mM, and CAPS,d = 60 mM.

(2)

where QNIPAAm, QTEMED, QAPS, and QH2O are the individual volumetric flow rates of the solutions of NIPAAm, TEMED, APS, and water, respectively. After on-chip mixing, the solutions of reagents were diluted to the initial on-chip concentrations C NIPAAm,d = C NIPAAm,i × Q NIPAAm/QT

(3a)

C TEMED,d = C TEMED,i × QTEMED/QT

(3b)

CAPS,d = CAPS,i × Q APS/QT

(3c)

(4)

A = Clε

After a reaction time t, the concentrations of the reagents changed to CNIPAAm,t, CTEMED,t, and CAPS,t. 4.2. Correlating Absorbance to Concentration. The use of in situ ATR-FTIR characterization enables direct measurements of reagent and product concentrations. Figure 2a shows representative IR spectra collected for the individual solutions of APS, TEMED, NIPAAm, and PNIPAAm. The arrows specify characteristic peaks for each species. Prior to the reaction, we verified that the spectrum of the reaction mixture was a weighted average of the individual reagent spectra. Figure 2b shows the variation in the intensity of the IR peaks, labeled in Figure 2a, for solutions with solute concentrations varying from 1 to 150 mM. Furthermore, we verified that the linear trend of IR peak intensity vs monomer concentration held until CNIPAAm 4471

dx.doi.org/10.1021/ma300444k | Macromolecules 2012, 45, 4469−4475

Macromolecules

Article

CTEMED,d and CNIPAAm,d were maintained constant, eq 7 took the form of eq 8a. Similarly, when CTEMED,d or CNIPAAm,d was changed and the concentrations of two other reagents were maintained constant, eq 7 took the form of eqs 8b and 8c, respectively.

After a stabilization time of 5 min, the FTIR spectra were acquired and the concentration of NIPAAm was determined from the intensity of NIPAAm absorbance peak at 975 cm−1 by using eq 4. Following data acquisition, the value of QT was changed to achieve a new reaction time, without changing the ratio between CNIPAAm,d, CTEMED,d, and CAPS,d, by modulating all reagent flow rates by the same multiplicative factor. We examined the effect of four different reaction times in the range from 2.25 to 6.75 s. The flow rates of all individual reagent streams are summarized in the Supporting Information, Table S1. The fifth data point at t = 0 was acquired by measuring the intensity of the absorption peak of unreacted NIPAAm for the monomer solution at QNIPAAm = 0.235, QTEMED = 0, QAPS = 0, and QH2O = 0.475 (mL h−1), which based on eq 3 yielded CNIPAAm,d = 175 mM. Table S2 (Supporting Information) summarizes the flow rate ratios and the corresponding diluted reagent concentrations for every series of experiments conducted in this work. As NIPAAm was consumed over the course of polymerization, the variation of CNIPAAm,t vs t yielded an exponentially decaying curve (Figure S2, Supporting Information). Such curves were plotted for various CNIPAAm,d, CTEMED,d, and CAPS,d. Each decay curve (CNIAAm,t vs t) was fit to an exponential decay function CNIPAAM,t = CNIPAAM,de−bt, where CNIPAAM,d = 175 mM, and b is the exponential decay constant (determined from fitting). The initial rate of decay at t = 0 was acquired by taking the derivative of the decay function with respect to time and setting t = 0 (that is, dM/dt|0 = −b × 175 mM s−1). We converted the initial rate of decay of the monomer to the initial polymerization rate as −dM/dt|0 = dP/dt|0, where the terms on the left and right side are the rate of change of the monomer concentration and the rate of change of the polymer concentration at t = 0, respectively. 4.4. Determining the Reaction Order of the Reagents. In assuming a steady-state approximation, the rate of free radical polymerization in the absence of an accelerator species is determined as dP dt

0

⎛ k ⎞1/2 = k p⎜f d ⎟ CAPS,d1/2C NIPAAm,d ⎝ kt ⎠

⎛ dP ⎞ ⎟ = x ln(C ln⎜ APS,d) + ln(wa) ⎝ dt 0 ⎠

(8a)

⎛ dP ⎞ ⎟ = y ln((C TEMED,d) + ln(wb) ln⎜ ⎝ dt 0 ⎠

(8b)

⎛ dP ⎞ ⎟ = z ln(C NIPAAm,d) + ln(wc) ln⎜ ⎝ dt 0 ⎠

(8c)

where wa, wb, and wc are constants with the values ln(wa) = ln k′ + y ln(CTEMED,d) + z ln(CNIPAAm,d), ln(wb) = ln k′ + x ln(CAPS,d) + z ln(CNIPAAm,d), and ln(wc) = ln k′ + x ln(CAPS,d) + y ln(CTEMED,d). Plotting ln(dP/dt|0) vs ln(C) yields a linear plot with a slope equal to the reaction order with respect to the reagent with changing concentration.

5. RESULTS AND DISCUSSION 5.1. Effect of Concentration of the Initiator on Polymerization Kinetics. In the first series of experiments, we tested the effect of the concentration of the initiator APS, CAPS,d, on the rate of NIPAAm polymerization. Figure 3a shows representative overlaid spectra of the NIPAAm double bond peak for different t during the polymerization reaction. The arrow shows that in the time interval 0 ≤ t ≤ 6.75 s the characteristic NIPAAm absorbance peak decreased with reaction time, indicating consumption of the monomer.

(5)

where f is the fraction of initiator radicals reacting with the monomer, and kp, kd, and kt are the rate constants for chain propagation, initiator decomposition, and chain termination, respectively.22 In the present work, TEMED accelerated the reaction by enhancing decomposition of APS and by forming its own radical species23 and determined reaction rate as dP dt

= k′CAPS,d xC TEMED,d yC NIPAAm,d z 0

(6) Figure 3. (a) Absorbance spectra acquired in the course of polymerization of NIPAAm at different reaction times. The rate of decay of the peak at 975 cm−1 (CC−H) of NIPAAm corresponds to the rate of consumption of the monomer (the rate of polymerization). The arrow indicates the direction of the peak intensity. CNIPAAm,d = 175 mM, CTEMED,d = 30 mM, and CAPS,d = 60 mM. (b) Decay curves for monomer absorbance for CAPS,d of 15 (◆), 30 (□), 45 (▲), and 60 (○) mM. CTEMED,d = 30 mM; CNIPAAm,d = 175 mM. The slopes, acquired at t = 0 from exponential fits, provide the respective initial rates of polymerization, dP/dt|0. (c) A plot of ln(dP/dt|0) vs ln(CAPS,d) generated based on the data shown in (b). All measurements were conducted three times. Error bars were generated based on the standard deviation between all three measurements.

where k′ = kp[f(kd/kt)]1/2. The equivalent form of eq 6 is ⎛ dP ⎞ ⎟ = x ln(CAPS,d) + y ln(C TEMED,d) + z ln⎜ ⎝ dt 0 ⎠ ln(C NIPAAm,d) + ln k′

(7)

To test the reaction order with respect to a particular reagent, dP/dt|0 was determined for several concentrations of the reagent, while the concentrations of the other two reagents were kept constant. For example, when CAPS,d changed and 4472

dx.doi.org/10.1021/ma300444k | Macromolecules 2012, 45, 4469−4475

Macromolecules

Article

The experiments were conducted at four values of CAPS,d, while other reagents’ concentrations remained unaltered (Table S2, Supporting Information). Figure 3b shows the variation in CNIPAAm,t vs t. The initial rate of monomer consumption (−dM/ dt|0) from each curve was determined by fitting the decay curves and determining the initial slope, as discussed in section 4.3. Table S3 summarizes the fitting results for Figure 3b. By using −dM/dt|0 = dP/dt|0, we plotted ln(dP/dt|0) vs ln(CAPS,d) (Figure 3c) and determined the slope to be 0.50. On the basis of eq 7a, we determined the order of reaction with respect to APS to be ∼0.5, as expected for free radical polymerization reactions.22 5.2. Effect of Concentration of the Accelerator on Polymerization Kinetics. In the next series of experiments, the values of CAPS,d and CNIPAAm,d were maintained at 30 and 175 mM, respectively, and the decay curves for monomer concentration were collected at four values of CTEMED,d. The value of dP/dt| 0 was determined for each monomer concentration decay curve, as described in section 5.1. Figure 4 shows the acquired data and the linear fit for ln(dP/dt|0) vs

which is the equation of a line with the slope equal to x + y, the sum of the orders of reactions with respect to APS and TEMED. (In the previous sections, we found that x = 0.50 and y = 0.38, respectively.) For CNIPAAm,d = 175 mM, we collected decay curves and determined the initial rate of decay for varying concentrations of Cm,d (Figure 5). A linear fit of the data points gave a slope of 0.83, which was within ∼5% of the expected value of 0.88 from eq 9b.

Figure 5. Variation of ln(dP/dt|0) vs Cm,d. In all cases the reaction mixture included CNIPAAm,d = 175 mM.

5.4. Effect of Monomer Concentration on Polymerization Kinetics. To investigate the effect of monomer concentration on the polymerization kinetics, we used CAPS,d = CTEMED,d = 30 mM and modulated CNIPAAm,d in the range of 90−305 mM (Supporting Information, Table S2). The initial rate of decay of the monomer concentration was measured for each CNIPAAm,d, and a plot of ln(dP/dt) vs ln(CNIPAAm,d) was generated (Figure 6). Using eq 8c, we used the slope of the

Figure 4. Variation of ln(dP/dt|0) vs ln(CTEMED,d). CAPS,d = 30 mM; CNIPAAm,d = 175 mM.

ln(CTEMED,d). By using eq 7b, we determined from the slope of the plot the order of the reaction with respect to TEMED to be 0.38, which indicated that TEMED produces radicals that participate in polymerization of NIPAAm. This finding was in agreement with previous work.23 5.3. Effect of the Concentration of (Initiator + Accelerator) Complex on Polymerization Kinetics. We simultaneously changed the molar concentrations of both APS and TEMED, while keeping their ratio CAPS,d:CTEMED,d at 1:1. The concentration of the “initiator + accelerator” complex immediately after dilution (at the entrance of the reaction chamber (v), Figure 1) was denoted as Cm,d. To examine the effect of the change in Cm,d on the polymerization kinetics, we rearranged eq 7 as ⎛ dP ⎞ ⎟ = x ln(Cm,d) + y ln(Cm,d) + ln(wd) ln⎜ ⎝ dt 0 ⎠

Figure 6. Variation in ln(dP/dt|0) vs ln(CNIPAAm,d). The reaction mixture included CAPS,d = CTEMED,d = 30 mM.

graph in Figure 6 to determine the reaction order with respect to NIPAAm. This method gave us the order of 1.09, with respect to the monomer, close to 1.00, the order of free radical polymerization reaction with respect to the monomer.24 5.5. Effect of pH on Polymerization Kinetics. We conducted the polymerization reaction at three different pH values, in order to monitor changes to polymerization kinetics as a function of pH. To maintain a particular value of pH, we used three buffer solutions at pH 2.23, 7.26, and 9.50. First, we ensured that the pH value was maintained at the original value

(9a)

where ln(wd) = ln k′ + y ln(CTEMED,d). Rearranging eq 9a yielded ⎛ dP ⎞ ⎟ = (x + y) ln(Cm,d) + ln(wd) ln⎜ ⎝ dt 0 ⎠

(9b) 4473

dx.doi.org/10.1021/ma300444k | Macromolecules 2012, 45, 4469−4475

Macromolecules

■

ACKNOWLEDGMENTS The authors thank NSERC Canada (I2I program) and MaRS Innovation (Proof of Principle Program) for financial support of this work.

during the polymerization reaction, by monitoring the pH in real-time off-chip (bulk) experiments at CNIPAAm,d = 175 mM, CAPS,d = 30 mM, and CTEMED,d = 30 mM. Next, the buffered reagent solutions were injected into the MF reactor, and their flow rates were adjusted to achieve different reaction times until the reaction mixture reached the IR probe. The absorbance peak of NIPAAm was measured for each reaction time, as discussed in section 5.1, and the corresponding CNIPAAm was calculated using eq 4. Figure 7 shows the change in CNIPAAm in

■

each buffer solution. The initial reaction rates, dP/dt|0, in pH 2.23, 7.26, and 9.50 buffer solutions were determined to be 2.27, 2.97, and 3.50 M L−1 s−1, respectively. This trend was in agreement with previously reported increase in dP/dt|0 with increasing pH of the reaction system.25,26

6. CONCLUSIONS We demonstrated a systematic, high-throughput study of a multicomponent polymerization reaction in a MF reactor integrated with in situ FTIR. The technique was used to study the kinetics of a free-radical polymerization reaction of Nisopropylacrylamide, which was initiated in water by ammonium persulfate in the presence of the accelerator N,N,N′,N′-tetramethylethylenediamine. The MF format of these studies allowed rapid exploration of the reaction parameter space and thus enabled the determination of the reaction kinetics. The MF study was also used to examine the effect of pH of the reaction mixture on polymerization kinetics. This work opens the way for the kinetic measurements of complex polymer systems. With the use of feedback control of programmable syringe pumps this work opens the way for fully automated on-chip characterization of polymerization reactions for optimization of chemical formulations. ASSOCIATED CONTENT

S Supporting Information *

Design of the MF reactor, residence times in the reaction compartment achieved at different flow rates, and an IR spectrum of the reaction mixture. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

(1) Li, W.; Pham, H. H.; Nie, Z.; MacDonald, B.; Guenther, A.; Kumacheva, E. J. Am. Chem. Soc. 2008, 130, 9935−9941. (b) Mason, B. P.; Price, K. E.; Steinbacher, J. L.; Bogdan, A. R.; McQuade, D. T. Chem. Rev. 2007, 107, 2300−2318. (c) Jensen, K. F. Chem. Eng. Sci. 2001, 56, 293−303. (d) Lee, C.; Sui, G.; Elizarov, A.; Shu, C. J.; Shin, Y.; Dooley, A. N.; Huang, J.; Daridon, A.; Wyatt, P.; Stout, D.; Kolb, H. C.; Witte, O. N.; Satyamurthy, N.; Heath, J. R.; Phelps, M. E.; Quake, S. R.; Tseng, H. Science 2005, 310, 1793−1796. (e) Li, W.; Nie, Z.; Zhang, H.; Paquet, C.; Seo, M.; Garstecki, P.; Kumacheva, E. Langmuir 2007, 23, 8010−8014. (2) (a) Li, W.; Greener, J.; Voicu, D.; Kumacheva, E. Lab Chip 2009, 9, 2715−2721. (b) Nisisako, T.; Torii, T. Lab Chip 2008, 8, 287−293. (3) (a) Wu, T.; Mei, Y.; Cabral, J. T.; Xu, C.; Beers, K. L. J. Am. Chem. Soc. 2004, 126, 9880−9881. (b) Iwasaki, T.; Yoshida., J. Macromolecules 2005, 38, 1159−1163. (4) Miyazawa, A.; Kase, T.; Shibuya, T. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 1841−1844. (5) Yoshida, J.; Okamoto, H. In Advanced Micro & Nanosystems; Kockmann, N., Ed.; Wiley: Weinheim, 2006; Vol. 5. (6) (a) Baskaran, D.; Müller, A. H. E.; Kolshorn, H.; Zagala, A. P.; Hogen-Esch, T. E. Macromolecules 1997, 30, 6695−6697. (b) Nagaki, A.; Miyazaki, A.; Tomida, Y.; Yoshida, J. Chem. Eng. J. 2011, 167, 548− 555. (7) (a) Miyazaki, M.; Honda, T.; Nakamura, H.; Maeda, H. Chem. Eng. Technol. 2007, 30, 300−304. (b) Wilms, D.; Nieberle, J.; Klos, J.; Löwe, H.; Frey, H. Chem. Eng. Technol. 2007, 30, 1519−1524. (8) Rosenfeld, C.; Serra, C. A.; O’Donohue, S.; Hadziioannou, G. Macromol. React. Eng. 2007, 1, 547−552. (9) Bally, F.; Serra, C. A.; Brochon, C.; Anton, N.; Vandamme, T.; Hadziioannou, G. Macromol. React. Eng. 2011, 5, 542−547. (10) Barnes, S. E.; Cygan, Z. T.; Yates, J. K.; Beers, K. L.; Amis, E. J. Analyst 2006, 131, 1027−1033. (11) Norman, A. I.; Zhang, Q.; Beers, K. L.; Amis, E. J. J. Colloid Interface Sci. 2006, 299, 580−588. (12) Griffiths, P. R. In Handbook of Vibrational Spectroscopy; Chalmers, J. C., Griffiths, P. R., Eds.; John Wiley & Sons: Chichester, UK, 2002; pp 33−70. (13) Suga, S.; Okajima, M.; Fujiwara, K.; Yoshida, J. J. Am. Chem. Soc. 2001, 123, 7941−7942. (14) (a) Dong, A.; Huang, P.; Caughey, W. S. Biochemistry 1990, 29, 3303−3308. (b) Olsztynska, S.; Komorowska, M.; Vrielynck, L.; Dupuy, N. Appl. Spectrosc. 2001, 55, 901−907. (c) Baujard-Lamotte, L.; Noinville, S.; Goubard, F.; Marque, P.; Pauthe, E. Colloids Surf. 2008, 63, 129−137. (d) Noinville, S.; Revault, M.; Baron, M. H.; Tiss, A.; Yapoudjian, S.; Ivanova, M.; Verger, R. Biophys. J. 2002, 82, 2709− 2719. (e) Ekgasit, S.; Fulleborn, M.; Siesler, H. W. Vib. Spectrosc. 1999, 19, 85−91. (15) Potorac, S.; Popa, M.; Maier, V.; Lisa, G.; Verestiuc, L. J. Anal. Biochem. 2010, 401, 236−241. (16) Laremore, T. N.; Ly, M.; Solakyildirim, K.; Zagorevski, D. V.; Linhardt, R. J. Anal. Biochem. 2010, 401, 236−241. (17) Klouda, L.; Perkins, K. R.; Watson, B. M.; Hacker, M. C.; Bryant, S. J.; Raphael, R. M.; Kasper, F. K.; Mikos, A. G. Acta Biomater. 2011, 7, 1460−1467. (18) Greener, J.; Li, W.; Ren, J.; Voicu, D.; Pakharenko, V.; Tang, T.; Kumacheva, E. Lab Chip 2010, 10, 522−524. (19) Greener, J.; Tumarkin, E.; Debono, M.; Kwan, C. H.; Abolhasani, M.; Guenther, A.; Kumacheva, E. Analyst 2012, 137, 444−450. (20) (a) Greener, J.; Tumarkin, E.; Debono, M.; Dicks, A. P.; Kumacheva, E. Lab Chip 2012, 12, 696−701. (b) Greener, J.; Abbasi, B.; Kumacheva, E. Lab Chip 2010, 10, 1561−1566.

Figure 7. Variation in CNIPAAm vs time at different pH values of the reaction mixture: 2.23 (◆), 7.26 (□), 9.50 (▲). The initial concentrations of monomer, initiator, and accelerator were 175, 30, and 30 mM, respectively, for all pH values.

■

Article

AUTHOR INFORMATION

Notes

The authors declare no competing financial interest. 4474

dx.doi.org/10.1021/ma300444k | Macromolecules 2012, 45, 4469−4475

Macromolecules

Article

(21) Mirabella, F. M. Practical Spectroscopy Series: Internal Reflection Spectroscopy: Theory and Applications; Marcel Dekker Inc.: New York, 1993; pp 17−52. (22) Carraher, C. E. Seymor/Carraher’s Polymer Chemistry, 5th ed.; Marcel Dekker Inc.: New York, 2000. (23) (a) Feng, X. D.; Guo, X. Q.; Qiu, K. Y. Makromol. Chem. 1988, 189, 77−83. (b) Guo, X. Q.; Qiu, K. Y.; Feng, X. D. Makmmol. Chem. 1990, 191, 577−587. (24) Odian, G. Principles of Polymerization; John Wiley & Sons, Inc.: Hoboken, NJ, 2004; pp 214−215. (25) Caglio, S.; Righetti, P. G. Electrophoresis 1993, 14, 554−558. (26) Furman, O. S.; Teel, A. L.; Ahmad, M.; Merker, M. C.; Watts, R. J. J. Environ. Eng. 2011, 137, 241−247.

4475

dx.doi.org/10.1021/ma300444k | Macromolecules 2012, 45, 4469−4475