Colloid Polym Sci (2013) 291:1903–1912 DOI 10.1007/s00396-013-2926-9

ORIGINAL CONTRIBUTION

Preparation of epoxy monoliths via chemically induced phase separation Yu-Shun Luo & Kuo-Chung Cheng & Ching-Lin Wu & Chiu-Ya Wang & Teh-Hua Tsai & Da-Ming Wang

Received: 22 August 2012 / Revised: 4 February 2013 / Accepted: 11 February 2013 / Published online: 5 March 2013 # Springer-Verlag Berlin Heidelberg 2013

Abstract Epoxy porous monoliths were prepared from a commercial epoxy resin, D.E.R. 331, that cured with a tertiary amine, 2,4,6-tris-(dimethylaminomethyl) phenol, in the presence of a solvent, diisobutyl ketone (DIBK). During the curing process, polymers were formed and a decrease in its solubility in DIBK; the solution thus phase-separated, usually referred to as chemically induced phase separation. The phase separation formed interconnected polymer-poor phase that then became interconnected pores after the removal of DIBK. By varying the content of DIBK from 32 to 40 vol.%, epoxy monoliths with interconnected pores were prepared, with surface pore size ranging from 0.20 to 2.33 μm, overall porosity from 0.41 to 0.60, and ethanol permeability from 10 to 4,717 L/(m2 h−1 bar−1). The glass transition temperatures of the epoxy monoliths, measured with differential scanning calorimetry, were all higher than 100 °C, and temperatures of 5 % weight loss, analyzed by thermal gravimetry, were higher than 350 °C, evidencing the monoliths’ high thermal stability. Also, the monolith morphology was found to be strongly related to the reaction mechanism of polymerization. The results indicate that the mechanism of chain initiation and propagation associated with the tertiary amine can effectively form monoliths with interconnected pores, which cannot be easily prepared with a stepwise polymerization mechanism associated with using primary amine as the curing agent. Electronic supplementary material The online version of this article (doi:10.1007/s00396-013-2926-9) contains supplementary material, which is available to authorized users. Y.-S. Luo : K.-C. Cheng (*) : C.-L. Wu : C.-Y. Wang : T.-H. Tsai Department of Chemical Engineering and Biotechnology, National Taipei University of Technology, Taipei 106, Taiwan e-mail: [email protected] D.-M. Wang Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan

Keywords Epoxy monolith . Chemically induced phase separation . Tertiary amine . Thermal stability

Introduction Polymers with porous structure have varieties of applications; for example, serving as monoliths for separation processes, beads for chromatography columns, or scaffolds for tissue engineering [1–12]. For thermoplastic polymers, which can usually melt at temperatures above their melting points or dissolve in appropriate solvents, the porous structure can be generated by using thermally induced phase separation or nonsolvent-induced phase separation. However, these two methods cannot be used to generate porous structure in thermosetting polymers, which do not melt or dissolve after gelation. Porous structure needs to be generated before gelation. Chemically induced phase separation (CIPS) is a method that can generate porous structure in thermosetting polymers. Monomers or prepolymers are first dissolved in suitable solvents to form homogeneous solutions and then are polymerized during when their molecular weight increases and their solubility in the solvent decreases. With a certain degree of polymerization, the solution phase separates to form polymer rich and poor phases. The continuing polymerization would make the polymer-rich domains gel, which also fixes the morphology induced by the phase separation. By heating the gelled solution at a temperature above the glass transition temperature of the polymer-rich domains, the contained solvent in the gelled solution can be removed; porous thermosetting polymers can thus be obtained [13–18]. For example, macroporous epoxy networks with a narrow pore size distribution and a closed-cell morphology were prepared from an epoxy resin cured with a primary amine in the presence of hexane or cyclohexane [19, 20]. Epoxy thermosetting polymers have several advantages over thermoplastic polymers, such as good chemical resistance and high thermal stability. Therefore, some limit of the

1904

application of porous thermoplastic polymers in harsh environment, such as high temperature, low or high pH, could be lifted by using porous epoxy thermosetting polymers [21–23]. The present work focuses on preparation of porous epoxy monoliths for separation. As discussed above, porous epoxy with closed pores can be prepared by curing an epoxy resin with a primary amine in the presence of hexane or cyclohexane. However, monoliths with closed pores are not suitable for filtration applications because of the low permeation flux. To have permeation flux high enough for practical applications, epoxy monoliths with interconnected pores are needed. It has been shown that porous epoxy networks with interconnected pores can also be obtained by curing an epoxy resin with a primary amine, but replacing the solvent, such as hexane or cyclohexane, with thermoplastic polymers, such as poly(vinyl methyl ether) (PVME) or polyethylene glycol (PEG) [21, 23]. For such an approach, the removal of PVME and PEG after CIPS cannot be accomplished by evaporation. The removal of PVME was suggested to be performed by oxidative thermal decomposition and that for PEG by leaching with water. Complete removal of PVME or PEG would not be easy. We propose here to use another approach: by using tertiary amine, instead of primary amine, as the curing agent. The epoxy resin can be also cured by using a small amount of a Lewis base, such as a tertiary amine, that acts as an initiator of the anionic polymerization [24–27]. In this study, 2,4,6-tris-(dimethylaminomethyl) phenol (DMP-30) was chosen as a curing agent, and porous epoxy networks were synthesized by the chemically induced phase separation. It will be shown that the pores are interconnected in the epoxy monoliths prepared by using the proposed method. The influence of the fraction of solvent on the morphology, liquid permeability, and thermal properties of the cured samples has been investigated. The dependence of the porous structure of the epoxy networks on the curing temperature and the amount of DMP-30 is discussed in this work. Also, it will be demonstrated that that the monolith morphology is strongly related to the reaction mechanism of polymerization.

Colloid Polym Sci (2013) 291:1903–1912

samples prepared by different lots of DIBK. However, the trend of the results is consistent. Furthermore, the samples were synthesized by the same lot (A019839601) of DIBK in a short time to avoid the effects of the impurities of DIBK. Preparation of epoxy monolith D.E.R. 331 was mixed with DIBK, in different ratios, as shown in Table 1, and the mixture was stirred to form a homogeneous solution. Then, the curing agent DMP-30 was added to the mixture. The solution was transferred into a plate mold, which was then covered and sealed with contacting sheets that were coated with Teflon. The solution was cured at 40 °C for 24 h. The gelled sample was taken out of the mold then put in a vacuum oven at 170 °C for 24 h to post-cure and remove the solvent DIBK. Characterization of monolith morphology The cured epoxy monoliths were examined with a scanning electron microscope (SEM; Hitachi S-3000H). The monoliths were fractured in liquid nitrogen to prepare specimens for SEM examinations. All the specimens were sputtering coated with gold. Evaluation of the thermal properties of monoliths The glass transition temperature (Tg) of the epoxy monolith was determined by differential scanning calorimetry (DSC; TA 2910) at a heating rate of 10 °C/min. In addition, the thermal stability, such as decomposition temperature, of the monolith was measured by thermal gravimetric analysis (TGA; TA 2950) under a nitrogen purge. Measurement of ethanol permeability

Experimental

The cured monolith with a thickness of approximately 0.25 mm was mounted between two steel circular rings with a 13-mm inner diameter. The ethanol flow through the monolith was measured at room temperature and with a constant transmonolith pressure drop, 0.8 bar. The permeability J′ was calculated by the following equation:

Materials

J0 ¼

D.E.R. 331 (EEW=186), a diglycidyl ether of bisphenol A type epoxy resin from Dow Chemical Company, was used to prepare epoxy monoliths. Two kinds of curing agent were used to cure D.E.R. 331: DMP-30 from Aldrich and diethylenetriamine (DETA; 98 %) from Panreac. Diisobutyl ketone (DIBK; 149440010) was purchased from Acros. All the chemicals were of reagent grade and used as received without further purification. It was found that there were few variable results of the

where Q is the steady volumetric flow rate of ethanol (99 %), A is the cross-sectional area of the monolith, and ΔP is the transmonolith pressure drop (0.8 bar in this study).

Q A  Δp

ð1Þ

Determination of monolith porosity Monoliths were immersed in ethanol to determine their porosities. The weight differences between dry and wet

Colloid Polym Sci (2013) 291:1903–1912

1905

Table 1 Pore structures and ethanol permeabilities of the epoxy monoliths prepared from D.E.R. 331 and cured with DMP-30 (10 phr) and different amounts of DIBK, at 40 °C for first stage Samplea

Fraction of DIBK (vol.%)

Surface mean pore diameter (μm)

Surface porosity

Overall porosity

J′ (Lm−2 h−1 bar−1)

Monolith thickness (mm)

32 34 36 38 40

0.20 0.25 0.74 1.10 2.33

0.07 0.12 0.29 0.43 0.59

0.41 0.48 0.51 0.55 0.60

10 67 251 864 4717

0.26 0.26 0.28 0.25 0.25

M-32 M-34 M-36 M-38 M-40 a

The sample code, M-x, denotes the monolith made from the solution with x in volume percent of DIBK

monolith samples (before and after the immersion) were measured, and the monolith porosities can then be determined by the following equation [28]: "v ¼

Ww  Wd ρe  V

ð2Þ

where Ww is the weight of the monoliths wetted by the ethanol, Wd is the weight of dry monoliths, ρe is the density of ethanol at room temperature, and V is the volume of the fully wetted monolith. Monitoring of phase separation process The solution of D.E.R. 331, DIBK, and DMP-30 was sealed into a thin glass mold on a heating/freezing stage (THMS600, Linkam). The light transmission (around 500 nm) through the sample was monitored by an UV/visible spectrometer (TP300, Ocean Optics) coupled with an optical fiber. The phase separation process was further observed by a phase-contrast microscope (MDS-3600 PRO, Mediscope) and a cone-and-plate rheometer (ARES, Rheometric Scientific) at frequency of 1 Hz, and strain of 1 %.

Results and discussion Morphologies of epoxy monoliths The epoxy monoliths discussed in the present work were cured with DMP-30, if the curing agent is not specifically mentioned. Figure 1 shows the SEM micrographs of the cross-sectional morphologies of the cured epoxy monoliths prepared with various DIBK contents. A transparent monolith was formed at 28 vol.% of DIBK. No apparent macroporous structure was found as indicated in Figs. 1a and 2a. It implies that the cured sample is a dense monolith. By increasing the volume fraction of DIBK to 32 vol.%, the sample became opaque and the pores could be seen clearly by the

SEM micrograph. As shown in Figs. 1 and 2, with increasing solvent, the pore size increases. The process at 40 % of DIBK was further monitored by the phasecontrast microscope. It was found that, at the very beginning of time, a few particles appeared in the solution, and the quantity of the particles increased with the reaction time. The CIPS process is very different from that cured by a primary amine (Figs. A4 and A5 in the Electronic supplementary material (ESM)), which will be discussed in the later section. The surface morphology of the epoxy thermosets prepared with various fractions of solvent is illustrated in Fig. 2. A dense structure was formed with 28 vol.% of DIBK, but, at a volume fraction higher than approximately 30 vol.%, some pores developed on the surface. Furthermore, the pore diameter was measured by the SEM image and the mean pore diameter was calculated as [28]: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi P ni di 2 P dm ¼ ni

ð3Þ

Where ni is the number of pores with diameter di. The monolith surface porosity was determined by calculating the ratio of the area of all pores to the monolith area containing them, taken from the SEM micrographs. As indicated in Table 1, the surface pore size increases with increasing DIBK fraction. The mean diameter of the surface pores was about 0.20 μm with 32 vol.% of DIBK, and it became 2.33 μm as DIBK increased to 40 vol.%. Both surface and overall porosities increased with increasing solvent fraction, as shown in Fig. 3. Ethanol permeation through epoxy monoliths The results of ethanol permeabilities through the epoxy monoliths are reported in Table 1 and Fig. 4. It was found that no ethanol flux was detected with a transmonolith pressure drop of 0.8 bar, for the epoxy monolith prepared with 30 vol.% of DIBK. As the fraction of DIBK increased

1906

Colloid Polym Sci (2013) 291:1903–1912

Fig. 1 SEM images of the cross-sectional morphologies of the epoxy monoliths

(a) M-28 (

from 32 to 40 vol.%, the overall porosity raised from 0.41 to 0.60, and the ethanol permeability increased dramatically from 10 to 4,717 Lm−2 h−1 bar−1.

5µm)

(d) M-36

(b) M-30

(e) M-38

(c) M-32

(f) M-40

shows that the epoxy monoliths we prepared possessed good thermal stability.

Thermal properties of epoxy monoliths

Dependence of monolith morphology on the curing temperature and the content of DMP-30

The Tg of the epoxy monoliths were measured with DSC, and the results are shown in Fig. 5 and Table 2. It was found that all the Tg were high than 100 °C. The Tg of the dense epoxy monolith, M-28, was about 116 °C, and it decreased from 113 to 104 °C when the DIBK used to prepare the monoliths raised from 32 to 40 vol.%. In addition, the thermal decomposition of the cured epoxy networks was analyzed by TGA under a nitrogen purge, and the weight loss and derivative thermogravimetric curves are plotted in Fig. 6. The thermal properties of the epoxy thermosets are summarized in Table 2. At a heating rate of 10 °C/min, a 5 % weight loss of the sample occurred from 351 to 367 °C, and the temperature of the maximum decomposition rate was higher than about 435 °C. This result

The porous structure of the epoxy monolith can be changed by the competition between phase separation and curing rate, both of which are affected by the thermodynamics of phase separation and the kinetics of the polymerization during cure [29]. Since the rate of polymerization strongly depends on the curing temperature and the amount of curing agent, these two factors may play important roles in determining the monolith morphology. The dependence of monolith morphology on the curing temperature is shown in Fig. 7. The monolith pore size decreased with higher curing temperature, implying that the phase-separated domains (polymer rich and poor phases) had less time to coarsen and coalescence. Higher curing temperature increased the polymerization rate, which could reduce the gel time. Therefore, it retarded the phase-separated domains to

Colloid Polym Sci (2013) 291:1903–1912

1907

Fig. 2 SEM images of the surface morphologies of the epoxy monoliths

(a) M-28 (

(b) M-30

(e) M-38

(c) M-32

(f) M-40

grow. In contrast, the domains could grow faster owing to a lower viscosity at the higher temperature. The results indicated that the factor of curing rate dominated the growth of phase-separated domains.

0.7

0.5 0.4 Surface

0.3 0.2 0.1 0.0 M-30

M-32

M-34

M-36

M-38

Membranes Fig. 3 Overall and surface porosities of the epoxy monoliths

M-40

Permeability (L/m2 hr bar)

Overall

0.6

Porosity

(d) M-36

5µm)

10000

4717

864

1000 251 67

100

10

10

1 M-32

M-34

M-36

M-38

Fig. 4 Ethanol permeabilities of the epoxy monoliths

M-40

1908

Colloid Polym Sci (2013) 291:1903–1912 100

M-0

Exo Heat flow (W/g)

M-28

M-28 M-32 M-40

80

Weight (%)

M-30 M-32 M-34 M-36 M-38

DTG %oC-1

Tg

60 40 20

M-40 0

100

60

80

100

120

Temperature

140

160

180

200

300

400

Temperature

500

600

o

C

Fig. 6 Thermogravimetry and derivative thermogravimetry (DTG) curves of the epoxy monoliths

oC

Fig. 5 DSC diagrams of the epoxy monoliths

Figure 8 shows the decrease in pore size with increasing amount of DMP-30. The trend is similar to that shown in Fig. 7. The results could be explained by the higher rate of polymerization caused by higher amount of DMP-30. The faster polymerization rate accelerated the gel time, and the phase-separated domains had not as much of a time to coarsen and coalescence. Moreover, the conversion at the gel point as well as the chemical structure and the interaction parameters of the formed polymers all depend on the ratio of DMP-30, which was probable to change the morphology of epoxy monolith. Effect of curing agents on monolith morphology The results presented above show that, by using DMP30 as the curing agent, epoxy monoliths with interconnected pores can be obtained. To be shown in this section is that the connectivity of the monolith Table 2 The thermal properties of the monoliths determined by DSC and TGA Sample

M-28 M-30 M-32 M-34 M-36 M-38 M-40

Tg (°C)

Tda (°C)

Tmaxb (°C)

Char residue at 600 °C (wt.%)

116 110 113 115 114 110 104

367 363 359 360 361 353 351

435 437 436 437 440 439 437

12 12 11 10 10 10 12

a

The temperature at 5 % weight loss under a nitrogen purge at a heating rate of 10 °C/min

b

The temperature at the maximum decomposition rate

pores is strongly affected by the curing agent used. With the same solvent DIBK, when D.E.R. 331 was cured with DETA, with the equal stoichiometry of amine hydrogen to epoxy groups, the obtained epoxy monolith contained closed pores, as shown in Fig. 9 but not interconnected pores. The reason for such a dramatic change in monolith morphology is discussed below. A major difference between the two curing agents is that they cured the epoxy resin with different reaction mechanisms. The curing of D.E.R. 331 with DETA is via a stepwise polymerization of the epoxy group with either the primary or secondary amine group in DETA [30–33]. For the curing with a tertiary amine, the reaction mechanism is much more complicated, including chain initiation and propagation: a quaternary ammonium alcoholate is first formed via the initiation reaction of the epoxy and tertiary amine groups; then the chain propagation proceeds by the active site via the anionic polymerization [24, 25, 31]. The difference in the reaction mechanism for the two curing agents can be seen from the DSC curves shown in Fig. 10. For the system with DMP-30, there are two main exothermic peaks on the DSC curve. While for the system with DETA, only one exothermic peak was found during the curing process. The polymerizations of epoxy resin and various ratios of DMP-30 with addition of DIBK, or not, were monitored by the DSC as shown in Figs. A1 and A2 (ESM). Similarly, two main exothermic peaks on the DSC curves appeared under the bulk polymerization, and the enthalpy of first peak increased with the increase in DMP30 (Fig. A2 in the ESM). It implies that the two exothermic peaks resulted solely from the reaction mechanisms of epoxy resin with DMP-30 but not the effect of the phase separation or the solvent. Moreover, there was only one exothermic peak found for the curing system of another tertiary amine: benzyldimethylamine [24] that has only

Colloid Polym Sci (2013) 291:1903–1912

1909

Fig. 7 SEM images of the epoxy monoliths prepared at different temperatures. a–c Surface and d–f cross section

(a) 40oC (

one tertiary amino group and no phenol group. Therefore, the two exothermic peaks might be caused by the different reactivity, or substitution effect of the three tertiary amino groups on DMP-30, and the phenol group could also play a role during polymerization. The phase separation behavior of the D.E.R. 331/DIBK solution cured with different curing agents also differed a lot, as indicated in the light transmission curves shown in Fig. 10. We measured the intensity of the light transmitted through the solution in the sealed mold for curing. When the solution phase separated, the solution became turbid or even opaque, reducing the intensity of the transmitted light. The light transmission of a solution was defined as the ratio of the intensity of the transmitted light through it to that of the transmitted light through the initial transparent solution. During the curing of a solution, the change in the light transmittance of the solution can be interpreted as how the solution phase separates. The light transmission curve shown in Fig. 10b indicates that, with DETA as the

5µm)

(d) 40oC

(b) 50 oC

(e) 50oC

(c) 60oC

(f) 60oC

curing agent, the solution remained transparent for a time period during which the molecular weight of the epoxy polymer was increasing but not high enough to induce phase separation. When the degree of polymerization was high enough to induce phase separation, the light transmittance dropped, indicative of formation of polymer rich and poor domains in the solution. On the other hand, the light transmission curve for the solution cured with DMP-30 (Fig. 10a) shows a phase separation behavior much different from that cured with DETA (Fig. 10b). Drop of light transmittance was observed almost immediately after the beginning of curing, but the drop continued only for a short period. The solution remained translucent for a while and a second drop of the light transmission was then observed. The light transmission curve presented in Fig. 10a seemed to correspond well to the two-step curing mechanism of tertiary amine. The change of the phase was further observed by the phase-contrast microscope, which could enhance the interfacial profiles between two phases, as shown in Figs. A4

1910

Colloid Polym Sci (2013) 291:1903–1912

Fig. 8 SEM images of the epoxy monoliths prepared with different amount of DMP-30. a–c Surface and d–f cross section

(a) 5 phr (

5µm)

(d) 5 phr

(b)10 phr

(e) 10 phr

(c)15 phr

(f) 15 phr

and A5 in the ESM. For DMP-30 system, there were many small particles appeared at the very beginning; then, more particles were formed successively at the first stage. The images became opaque and dark around 150 min. The result

10 µm

Fig. 9 Cross-sectional SEM image of the epoxy monolith prepared from the mixture of D.E.R. 331/DETA (11 parts per hundred (phr))/DIBK (40 %)

was consistent well with the light transmission curve shown in Fig. 10, and the mechanisms of polymerization initialized by DMP-30 mentioned above. The CIPS process could be described by a monomer-solvent-polymer pseudo-ternary phase diagram. At first, the system of epoxy resin (monomer) with solvent was one-phase; then the system became two-phase quickly owing to the polymers formed via the chain polymerization while the miscibility gap is shifting to the monomer vertex [22]. The second-stage phase change is believed to be caused by the second reaction mechanism as indicated by the DSC diagram. In contrast, for DETA system, the phase change observed by the phase-contrast microscope occurred around 66 min, and the transmission of light dropped quickly simultaneously (Fig. 10). There were many separated domains formed, coarsened, and coalesced to form larger ones. The domain was solvent-rich phase, which was further confirmed by the SEM pictures after removing the solvent. In contrast to the ternary phase diagram for the chain polymerization, the CIPS process of the stepwise polymerization of epoxy resin with primary amine has been well discussed by a pseudo-binary phase

Colloid Polym Sci (2013) 291:1903–1912

1911

80

a2

60

(a)

20

a3

0

0

100

200

300

400

500

600

700

800

b1

100 80

b2

Heat flow (W/g)

Transmission (%)

40

60 40

(b) b3

20 0

0

100

200

300

400

500

600

700

800

Time (min) Fig. 10 DSC curve and transmission of the back light through the reactive solution at 40 °C. a D.E.R. 331/DMP-30 (10 phr)/DIBK (40 %) and b D.E.R. 331/DETA (11 phr)/DIBK (40 %)

diagram, of the conversion or degree of polymerization versus composition of polymer and solvent [22]. Epoxy resin was first dissolved in solvents to form a homogeneous solution. The molecular weight of polymers increased with the conversion, and their solubility in the solvent decreased. At a certain conversion, the solution phase separates to form polymer rich and poor phases. The CIPS process was further monitored by a cone-andplate rheometer. The shear storage and loss modulus (G′ and G″) as well as the complex viscosity profiles dependent on the reaction time are shown in Fig. 11. At early stage of the polymerization, the G″ was larger than G′ owing to the liquid-like behavior of the solution with lower molecular weight or minor quantity of polymers. Once the G′ crossed over the G″, the elastic portion dominated. The gel point characterizes the transition from liquid to solid state, and it can be estimated from the crossover point of the G′ and G″ [34]. For the DETA system, the G′ and G″ and complex viscosity increased gradually with the growth of molecular weight of the polymers via stepwise polymerization. The gel time was about 180 min, and the phase separation started about 70 min (Fig. 10), which indicated that the separated domains have longer than 100 min to grow and coarsen as larger isolated cells under low viscosity. On the other hand, for the DMP-30 system, after the first-stage phase separation, the G′ and G″ and complex viscosity increased rapidly at about 75 min, and the crossover point of G′ and G″ was at about 90 min. After that, the value of G″ was still very close to that of G′. It implies that there were still a lot of

monomers or small molecules in the reaction system. The second-stage polymerization accompanied with a secondstage transmission light dropped after the crossover point. Meanwhile, there were more small particles further formed observed by the microscope. Because the second-phase separation process occurred under high complex viscosity, the slower domain growing rate could help to retain the nascent structure formed by phase separation, and the final monolith pore size for the DMP-30 system (Fig. 1; Fig. A4 in the ESM) was smaller than that for the DETA system (Fig. 9; Fig. A5 in the ESM). According to the results we obtained, it is clear that the morphology of the epoxy monoliths is strongly related to the polymerization mechanism. The mechanism of chain initiation and propagation associated with the tertiary amine, DMP-30, can effectively form monoliths with interconnected pores, which cannot be easily prepared with a stepwise polymerization mechanism associated with using primary amine as the curing agent. However, the other cause, such as the different interaction parameters between the two curing systems, might also play a role on the CIPS, and it is worthy of being further studied. (a) G' and G'' dyne/cm 2

a1

10

7

10

6

10

5

10

4

10

3

10

2

10

1

10

0

DETA

DMP-30

X G'' O G' 0

50

100

150

200

250

time (min)

(b) 10

complex viscosity (cP)

100

10

10

9

10

8

10

7

10

6

10

5

10

DETA

DMP-30 4

10

3

10

2

0

50

100

150

200

250

time (min) Fig. 11 a Shear storage and loss modulus (G′ and G″) as well as b complex viscosity profiles dependent on the reaction time at 40 °C, in which the initial compositions are the same as those in Fig. 10

1912

Conclusions The epoxy monoliths with interconnected pores have been prepared via the polymerization of epoxy resin cured with a tertiary amine, DMP-30, in the presence of 32 to 40 vol.% of DIBK. The interconnected pore structure of the cured epoxy monolith was further verified by the ethanol permeation test. The morphology and liquid permeability of the epoxy monolith can be controlled by the initial reaction composition, such as the amount of solvent or DMP-30, and the curing temperature. Furthermore, the prepared epoxy monoliths possessed good thermal stability. The temperatures at 5 % weight loss of the epoxy monoliths determined by TGA were higher than 350 °C. As observed in SEM, the morphology of the porous epoxy thermosetting polymers via curing with a tertiary amine is different from that prepared with a primary amine. This result implies that a different polymerization mechanism could cause the special pore structure to form. However, the aim of this work is to demonstrate that an epoxy monolith with interconnected canals and high thermal stabilities can be achieved by chemically induced phase separation. The dependence of the phase separation process on the various chemical reaction mechanisms is worthy of careful study in the future. Acknowledgment We thank the National Science Council of Taiwan for the financial support of this study under Contract NSC 97-2221-E027-013-MY2.

References 1. Sotiropoulou S, Vamvakaki V, Chaniotakis NA (2005) Biosens Bioelectron 20(8 SPEC. ISS):1674 2. Huck CW, Bonn GK (2005) Chem Eng Technol 28(12):1457 3. Barbetta A, Carnachan RJ, Smith KH, Zhao CT, Cameron NR, Kataky R, Hayman M, Przyborski SA, Swan M (2005) Macromol Symp 226(1):203 4. Higuchi A, Shindo Y, Gomei Y, Mori T, Uyama T, Umezawa A (2005) J Biomed Mater Res B Appl Biomater 74(1):511

Colloid Polym Sci (2013) 291:1903–1912 5. Safinia L, Mantalaris A, Bismarck A (2006) Langmuir 22(7):3235 6. Romeo HE, Vílchez A, Esquena J, Hoppe CE, Williams RJJ (2012) Eur Polym J 48(6):1101 7. Svec F, Kurganov AA (2008) J Chromatogr A 1184(1–2):281 8. Varilova T, Madera M, Pacakova V, Stulik K (2006) Curr Proteomics 3(1):55 9. Zilberman M (2005) Acta Biomater 1(6):615 10. Xie S, Svec F, Fréchet JMJ (1998) Chem Mater 10(12):4072 11. Svec F, Fréchet JMJ (1999) Ind Eng Chem Res 38(1):34 12. Peters EC, Svec F, Fréchet JMJ, Viklund C, Irgum K (1999) Macromolecules 32(19):6377 13. Kiefer J, Hedrick JL, Hilborn JG (1999) Macroporous thermosets by chemically induced phase separation. Adv Polym Sci 147:161–247 14. Kiefer J, Hilborn JG, Hedrick JL, Cha HJ, Yoon DY, Hedrick JC (1996) Macromolecules 29(26):8546 15. Della Martina A, Hilborn JG, Mühlebach A (2000) Macromolecules 33(8):2916 16. Garcia Loera A, Dumon M, Pascault JP (2000) Macromol Symp 151:341 17. Ai H, Xu K, Chen W, Liu H, Chen M (2009) Polym Int 58(1):105 18. Luo YS, Cheng KC, Huang ND, Chiang WP, Li SF (2011) J Polym Sci B Polym Phys 49(14):1022 19. Kiefer J, Hilborn JG, Hedrick JL (1996) Polymer 37(25):5715 20. Kiefer J, Hilborn JG, Månson JAE, Leterrier Y, Hedrick JL (1996) Macromolecules 29(11):4158 21. Loera AG, Cara F, Dumon M, Pascault JP (2002) Macromolecules 35(16):6291 22. Williams RJJ, Rozenberg BA, Pascault JP (1997) Adv Polym Sci 128:95 23. Tsujioka N, Ishizuka N, Tanaka N, Kubo T, Hosoya K (2008) J Polym Sci A Polym Chem 46(10):3272 24. Han JL, Hsieh KH, Chiu WY (1993) J Appl Polym Sci 50(6):1099 25. Dzhavadyan EA, Bogdanova LM, Irzhak VI, Rozenberg BA (1997) Polym Sci Ser A 39(4):383 26. Dell’Erba IE, Williams RJJ (2006) Polym Eng Sci 46(3):351 27. Kuzub LI, Irzhak VI (2001) Colloid J 63(1):86 28. Chakrabarty B, Ghoshal AK, Purkait MK (2008) J Colloid Interface Sci 320(1):245 29. Chen JL, Chang FC (2001) Polymer 42(5):2193 30. Vazquez A, Bentaleb D, Williams RJJ (1991) J Appl Polym Sci 43(5):967 31. Rozenberg B (1986) Kinetics, thermodynamics and mechanism of reactions of epoxy oligomers with amines. Adv Polym Sci 75:113–165 32. Shechter L, Wynstra J, Kurkjy RP (1956) Ind Eng Chem 48(1):94 33. Cheng KC, Don TM, Rwei SP, Li YC, Duann YF (2002) J Polym Sci B Polym Phys 40(17):1857 34. Winter HH, Chambon F (1986) J Rheol 30(2):367