Full Paper
Poly[ethylene-co-(vinyl alcohol)]-graft-poly(e-caprolactone) Synthesis by Reactive Extrusion, 1 - Structural and Kinetic Studya Fre´de´ric Becquart,* Yvan Chalamet, Jianding Chen, Yanchao Zhao, Mohamed Taha
Chemical modification of EVOH in the molten state at 185 8C by a grafting from process of poly(e-caprolactone) in batch was studied. 1H NMR was used to characterize the structure evolutions of PCL grafts. In addition to grafting reactions, dynamic covalent transesterification reactions between EVOH residual alcohols and the polyester grafts led to a redistribution of the PCL grafts length. Dpn up to 27 and SR up to 80% were obtained. Experiments made in a corotating mini twin-screw extruder also confirmed these results. The effect of the alcohol to caprolactone ratio and catalyst concentration (SnOct2) on kinetic evolution showed that few minutes were necessary to complete the polymerization. A kinetic model was proposed and adequate conditions for the synthesis by reactive extrusion were defined. Introduction
F. Becquart, Y. Chalamet, Y. Zhao, M. Taha Universite´ de Lyon, F-69003, Lyon, France Universite´ de Saint Etienne, F-42023, Saint Etienne, France CNRS, UMR5223, Inge´nierie des Mate´riaux Polyme`res, France IMP/LRMP 23, rue du docteur Paul Michelon F-42023 Saint Etienne cedex 2 (France) Fax: 00334 77 48 51 26; E-mail:
[email protected] J. Chen, Y. Zhao School of Materials Science and Engineering, East China University of Science and Technology, 200237 Shanghai, China a Part 2: cf. (in this issue) Macromol. Mater. Eng. 2009, 294, 643–650 ß 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
e-Caprolactone (CL) ring opening polymerization has been well studied during the last fifteen years. The main reasons for this interest are the possibility of controlling the ring opening polymerization with suitable organometallic catalysts or initiators[1] and because polycaprolactone (PCL) is biodegradable. Among catalysts and initiators, tin carboxylates or titanium and aluminium alkoxydes[2–4] were often used. To improve their reactivity, organic molecules or polymers containing one or more alcohol functions were added as co-initiators. These organometallic catalysts and initiators often limit, in ‘‘soft’’ conditions, the main possible side reaction, which is the backbiting reaction. Mechanistic studies with SnOct2 first proposed insertion-coordination mechanisms.[5–7] When tin carboxylate initiators are chosen, it is now admitted[8] that the alkoxyde pattern of the alcohol, used as a co-initiator,
DOI: 10.1002/mame.200900134
643
F. Becquart, Y. Chalamet, J. Chen, Y. Zhao, M. Taha
first undergoes an exchange reaction with the carboxylate groups leading to the acid formation. The formed tin alkoxyde finally acts as the polymerization initiator. In all cases, the polymerization degree is described as varying with the CL/co-initiator ratio and not with the CL/initiator ratio when the co-initiator is in excess.[7,9] ‘‘New’’ initiator systems are now being studied, since the ‘‘oldest’’ are thoroughly understood and relatively well controlled. Recently, low toxicity bismuth (III) triacetate was used at 150 8C for bulk polymerization with polyfunctional alcohol coinitiators.[10] A low transesterification activity[11] was described but at least 4 h are needed to reach 80% monomer conversion. Trinuclear zinc alkoxyde has been presented as a highly efficient initiator when used with alcohols as coinitiators.[12] Low molar mass alcohol co-initiators have generally been chosen to obtain PCL with specific end chains, such as butan1,4-diol, ethylene glycol, neopentylglycol or benzyl alcohol. Synthetic hydroxylated polymers have also been used[13–19] with SnOct2. In this case, the objective was always to modify the polymer with PCL grafts using a ‘‘grafting from’’ process. Transesterification reactions are always described as low, even for several hours of reaction time. Reactive extrusion (REX) is an interesting route for cost effective one-step preparation of polymer materials, and also prepolymers, by polymerization, copolymerization and grafting reactions. Reactions that previously required heavy equipment can be completed in a more efficient continuous way using twin-screw extruders.[20–24] This study concerns a one step preparation of poly[ethyleneco-(vinyl alcohol)]-graft-polycaprolactone (EVOH-g-PCL) by reactive extrusion. Because of the extreme pressure, shearing and temperature used in this reactive process, kinetic and structural studies need to be realized prior to extrusion in order to find the correct conditions to lead to the required conversion and also properties of the prepared products.[25–27] These studies are reported in the first part of this research.
operating with a RW 28 W IKA motor at 80 rpm, a condenser and a T-type thermocouple probe were fixed to the cover. A nitrogen flow, previously dried in a silica column, allowed air elimination from the reactor. The reactor was heated using an IKA HBR4 bath with silicon oil. The hydroxylated polymer, EVOH, was introduced into the reactor with the silicon oil bath temperature at 200 8C. The lactone was first heated in a glass bottle at 120 8C and introduced into the reactor just after the polymer. 20 min were necessary to obtain a homogeneous viscous aspect around 155 8C. When the reactive system became an homogeneous viscous liquid, the silicon oil bath temperature was decreased from 200 8C to 185 8C. The reactant temperature was finally stabilized at 185 8C. The initiator, previously diluted with twice its weight of monomer, was introduced into the reactor when the temperature in the reactor reached 185 8C.
Experiment Monitoring During the reaction, the reactor anchor torque was recorded. The torque was representative of the viscoelastic evolution and consequently allowed the polymerization reaction to be followed through the viscosity increase.
Reactive Extrusion Reactive extrusion was realized using a MiniLab II, Haake rheomix CTW5 corotating mini twin-screw extruder. The extruder screws are conical (5/14 mm diameters) and 109.5 mm in length. Reactants (typically 5 g) were introduced through the extruder hopper. A bypass, positioned at the screw end, can force the material through a feedback channel equipped with two pressure sensors. It can also force the material out of the barrel through the extruder die. The barrel temperature was set at 185 8C and the screw rotation rate was 50 rpm. The reaction was monitored by the screw’s torque.
Characterization
Experimental Part Reagents The EVOH used was E105B from EVAL Europe (melting point ¼ 165 8C, glass transition temperature, Tg ¼ 55 8C (EVAL data)) with an ethylene content of 44 mol-%, given by EVAL data and determined by 1H NMR spectroscopy in DMSO. Tin (II) bis(2-ethylhexanoate), also generally called stannous octanoate or stannous octoate, was purchased from Aldrich and used as received. The e-caprolactone was purchased from Solvay and distilled before use.
Grafting Reactions Batch syntheses were made in a 250 mL glass reactor (90 mm diameter) with a three necked steel cover. A steel anchor stirrer
644
Macromol. Mater. Eng. 2009, 294, 643–650 ß 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
NMR Analysis The 1H NMR and 13C NMR analyses were performed with a Bruker DRX250 spectrometer operating at 250 MHz for proton acquisition and 62.5 MHz for carbon acquisition. The modified EVOH were analyzed in DMSO-d6 at 90 8C. Chemical shift values (d) were in ppm with reference to internal tetramethylsilane (TMS).
Maldi-TOF Spectroscopy A Biosystems Voyager DE-STR was used with wavelength at 337 nm (nitrogen laser). The matrix was composed of dithranol at 10 g L1 in DMSO or HABA (2-(4’-hydroxybenzeneazo) benzoic acid) and DHB (2,5-dihydroxybenzoic acid) at 10 g L1 in DMSO in a linear positive mode. The samples were prepared with a 10/1/1 matrix/sample/NaI ratio.
DOI: 10.1002/mame.200900134
Poly[ethylene-co-(vinyl alcohol)]-graft-poly(e-caprolactone) Synthesis by Reactive . . .
from the resonances integrals (Equation (1)):
Dpn ¼
Figure 1. Structure and nomenclature of EVOH-g-PCL.
I ða þ a0 Þ I ð"0 Þ
The a and a0 protons of PCL are both located in a clear domain allowing the calculation of integral values with a better accuracy than the e and e0 methylene protons, classically selected, which are overlapped by EVOH resonances. For the same reason, instead of using end chain e0 protons for calculation, the methine protons located in the 5.2–4.8 ppm domain, corresponding to the start chain, were preferred. Considering that a methine integral is half a methylene integral, one obtains:
Dpn ¼ m ¼ 1
(1)
Iða þ a0 Þ 2 IðmethineðCHOCOLÞ
(2)
H NMR Characterization of EVOH-g-PCL
Average Polymerization Degree: Dpn The 1H NMR characterization of EVOH-g-PCL required the previous Fraction of the EVOH Alcohol Functions Bearing a PCL assignments of both PCL and EVOH spectra under the same Graft: Substitution Rate (SR) conditions (temperature and solvent). The PCL and EVO chemical [4,18,28] shifts are well known, and are composed of two well resolved The substitution rate (SR) is defined as the fraction of the EVOH resonance regions, between 1–2 ppm for methylene protons and alcohol functions bearing a PCL graft. To complete the grafted EVOH between 2.8–4.6 ppm for both methine and alcohol protons. The description, it is necessary to determine the SR. The substitution addition of trifluoroacetic acid (TFA) shifts the labile hydrogen rate is theoretically expressed by the following expression (OH) to low field >10 ppm and shows consequently the three methine dyads. The structure of such modified EVOH is described in Figure 1, with x, y and z being, respectively, the ratio of ethylene units, unreacted alcohol units and grafted alcohol units, esterified by ring opening polymerization. The sum (x þ y þ z) is always normalized to 1. m is the apparent average polymerization degree of lactone grafts (DP). For the initial EVOH: x ¼ 0.44, y ¼ 0.56 and z ¼ 0. Grafting onto EVOH is fully observed by the appearance of EVOH backbone methines linked to the PCL graft at 4.7 to 5.2 ppm (Figure 2). Two preliminary assumptions have been made: each PCL chain is grafted on the EVOH chain from an alcohol group and each PCL graft is terminated by a CH2OH end-group. These assumptions are fully justified considering the known polymerization mechanisms. Since the 1H NMR analysis allows us to distinguish, in the grafted copolymer (Figure 2), the e0 methylene at the chain end (CH2OH) from the (m–1) other methylenes (CH2COO) noted a and a0 between 2.0 and Figure 2. 1H NMR spectrum of the EVOH-g-PCL in DMSO-d6 at 90 8C (s: protonated residue 2.4 ppm, the average DP can be calculated of solvent). Macromol. Mater. Eng. 2009, 294, 643–650 ß 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.mme-journal.de
645
F. Becquart, Y. Chalamet, J. Chen, Y. Zhao, M. Taha
(c.f. Figure 1): SR ¼
z z ¼ y0 ðy þ zÞ
(3)
Each PCL chain corresponds to an alcohol substitution on the EVOH polymer backbone. Then the substitution ratio (SR in %) may be calculated from the 1H NMR spectrum since the attribution of all the spectrum resonances is easily realized A classic approach consists of determining at least as much as separate domains, with attributions of the different protons in the different domains, as a number of unknown variables (here, x, y, z and m) and to finally solve an equation system to determine the variables describing the structure. Such a domain here could be: Domain 1 (S1): 0.8 to 1.9 ppm, Domain 2 (S2): 2.7 to 4.6 ppm, Domain 3 (S3): 4.6 to 5.2 ppm and (S4): 2.0 to 2.4 ppm. Table 1 presents the attributions of the different protons in the different domains (alcohol protons are not mentioned) with a drop of TFA added for the NMR analysis, shifting them to low field. The following equation system derives from these attributions, where x, y and z values can be established from the integrations (S1, S2, S3 and S4): S1 ¼ 4 x þ 2 y þ 3ð2m þ 1Þ z
(4)
S2 ¼ y þ ð2m þ 1Þz
(5)
S4 ¼ 2m z
(6)
errors allowing the calculation of x and y with the different domains. The obtained substitution rates are not satisfying while the productsTS DPn , which are equal to the polymerized and grafted lactone quantity, overestimate the real used lactone quantity compared to the total hydroxyl site number chosen as reference. Taking account of this remark, a SR can be simply calculated knowing the initial OH/CL ratio and calculating, each time, easily the Dpn . Using both these values, it becomes possible to calculate SR, using: Total polymerized lactone units ¼ Number of PCL chains Average polymerization degree (8) Then: Number of PCL chains ¼ Total polymerized lactone=Average polymerization degree (9) where the number of PCL chains is equal to the number of OH from EVOH, substituted by PCL grafts. Knowing the initial CL/OH ratio, designated by R, it becomes possible to make the link between both these terms and to express the substitution rate SR: substituted OH by PCL grafts OH0 Number of > CH OPCL methine protons ¼ total number of hydroxyl sites in the formulation
SR ¼
(10) Finally:
S3 ¼ z
(7)
The resolution of this system gives, in many cases, erroneous results. For example, the ethylene content of the EVOH backbone is retrieved as between 0.46 and 0.56, which is always superior to the original value (0.44). This discrepancy is attributed to integration
SR ¼
S3 S3 R 2 S3 R ¼ ¼ ðOHÞ0 ðCLÞ0 S4
(11)
while twice the initial lactone quantity is equal to all the polymerized a and a0 protons after a total conversion, well observed each time, in the NMR spectrum. Equation (11) was finally used.
Table 1. Resonances attribution of protons in the EVOH-g-PCL.
Pattern
Proton
Ethylene
Methylene Methine Methylene Grafted methine
Vinyl Alcohol CL
Domain
Number
CH2
S1
4x
>CH
S2
y
CH2
S1
2y
>CH
S3
z
0
CH2
S2
2(mz)
a a0
CH2
S4
2(mz)
CH2
S1
6(mz)
OH
S2
z
ee 0
0
b bd d g g Hydroxyl
646
Macromol. Mater. Eng. 2009, 294, 643–650 ß 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0
DOI: 10.1002/mame.200900134
Poly[ethylene-co-(vinyl alcohol)]-graft-poly(e-caprolactone) Synthesis by Reactive . . .
Results and Discussion
The tin catalyst concentration can be neglected compared to alcohol concentration.
Homopolymer Formation To apply the previous calculation methods, it is necessary to confirm that all the monomer is copolymerized with the EVOH. The obtained reaction products solubilities in different solvents were examined. They were soluble in methanol or acetone, a non-solvent of PCL homopolymer. This is a first indication of homopolycaprolactone absence. In addition, samples were analyzed by MALDI-TOF spectroscopy. No detectable PCL oligomers were observed. From these results it is reasonable to consider the absence of PCL homopolymer.
Structure Determination of EVOH-g-PCL Experiments were first realized at 185 8C with different caprolactone to alcohol ratios; reactions were stopped after 20 min of reaction. In all cases, the monomer had totally reacted after fifteen minutes or less, confirmed on the proton NMR spectrum by the disappearance of the monomer a-methylene linked to the carbonyl). The catalyst to alcohol ratio was fixed as inferior to 1/150. The products were readily cooled in liquid nitrogen and analyzed by 1 H NMR. The DP and the SR (Figure 3) were calculated using respectively Equation (2) and Equation (11). When the ring opening polymerization of lactone occurs in the presence of an alcohol as co-initiator, the average polymerization degree was theoretically calculated from Equation (12). Dpn ¼
½"Cl0 2½SnOct2 þ ½ROH 0
(12)
Figure 3. Structure evolution: apparent average polymerization degree (*) and substitution rate (*) from the hydroxyl function, for EVOH-grafted-PCL by ring opening polymerization in molten state at 185 8C ([SnOct2]0/[OH]0)
