This Article Was Published in an Elsevier Journal. the Attached Copy

Home , Dodecanol

This article was published in an Elsevier journal. The attached copy is furnished to the author for non-commercial research and education use, including for instruction at the author’s institution, sharing with colleagues and providing to institution administration. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit:

http://www.elsevier.com/copyright Author's personal copy

Journal of Chromatography A, 1163 (2007) 212–218

Fabrication of high-permeability and high-capacity monolith for protein chromatography Kai-Feng Du, Dong Yang, Yan Sun ∗ Department of Biochemical Engineering, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China Received 15 March 2007; received in revised form 14 June 2007; accepted 19 June 2007 Available online 26 June 2007

Abstract A novel approach for the fabrication of macroporous poly(glycidyl methacrylate-ethylene glycol dimethacrylate) monolith is presented. The method involved the use of sodium sulfate granules and organic solvents as co-porogens. Compared with the conventional monoliths [ML-(1-3)] using organic solvents only as a porogen, the improved monoliths [MLS-(1-3)] showed not only higher column efﬁciency and dynamic binding capacity (DBC) for protein (bovine serum albumin, BSA), but also higher column permeability and lower back pressure. It is considered that the superpores introduced by the solid granules played an important role for the improvement of the monolith performance. Moreover, poly(glycidyl methacrylate-diethylamine) tentacles were grafted onto the pore surface of MLS-3 monolith. This has further increased the DBC of BSA to 74.7 mg/ml, about three times higher than that of the monoliths without the grafted tentacles. This grafting does not obviously decrease the column permeability, so a new monolith of high column permeability and binding capacity has been produced for high-performance preparative protein chromatography. © 2007 Elsevier B.V. All rights reserved.

Keywords: LC; Monolithic column; Solid porogen; Permeability; Dynamic binding capacity; Graft polymerization

1. Introduction monomers and some porogen in a column tube by in-situ polymerization [10]. However, the polymerization system can hardly Porous monolithic polymer materials for use in chro- be changed because each variation of porogenic solvent com- matography have been extensively studied over the last position has significant effect on the structure of the resulting decade [1] due to their potential applications in the sepa- materials [11,12]. Therefore, the pore size distribution of mono- ration of macromolecules. The monolithic stationary phases lithic columns is difficult to control due to the close interrelation have evolved from solvent-swollen hydrophilic acrylates [2] between concomitant porosity and reaction conditions. How to to methacrylate-based polymers [3], and then to more rigid tune the pore sizes concerning these materials thus becomes one polystyrene–divinylbenzene porous materials [4]. The inter- of the most challenging issues [13]. connected macroporous channels in monolithic columns can Recently, a novel porogenic method, cooperation of solid facilitate mass transport, while the micropores in the mono- granules and solvents as porogen, was developed in our lab- lithic skeleton provide the binding sites for the macromolecules oratory. The method can efficiently improve the permeability [5–7]. Therefore, monolithic columns with these unique struc- of beaded media without obvious loss of the dynamic binding tures enable high flow rates at low back pressure with less loss capacity (DBC) of proteins [14,15]. In general, this approach is in column efficiency, resulting in fast separation [8]. based on a templating strategy in combination with phase sepa- Polymer-based monolithic columns were first prepared in ration, which can easily control the microstructure of materials. early 1990s [9]. It was synthesized with cross-linking agent, To date, many templates, such as solid granules [16] and emul- sion [17] have been employed to prepare bimodal pore structure, in which macropores (namely, convective pores) allow fast mass ∗ Corresponding author. Tel.: +86 22 27404981; fax: +86 22 27406590. transfer while the micropores (namely, diffusive pores) give rise E-mail address: [email protected] (Y. Sun). to high surface area that assists the contact of solutes to the solid

K.-F. Du et al. / J. Chromatogr. A 1163 (2007) 212–218 213 surface. However, with the increase of through-pores formed by of dodecanol and cyclohexanol, and then the solution was added solid templates, the specific surface area decreases and then lim- slowly to the mixture of GMA and EDMA. After the polymer- its the active sites available for adsorption [17]. It would result ization system was purged with nitrogen for 15 min to remove in lower detection limit in protein microarrays [18] and lower oxygen, it was poured into a stainless-steel tube which was output in the preparation and separation of desired proteins [19]. sealed with rubber stopper at both ends. The polymerization One way to alleviate this problem is to graft tentacle-type poly- was performed in a water bath at 55 ◦C for 24 h. After the reac- mers on the surface of the internal channel of monoliths [20–24]. tion, the column was connected to an HPLC system for washing For example, Frechet´ and coworkers [23] and Muller¨ [24] have with ethanol/water (1/1, v/v) for 3 h and then with water for 1 h demonstrated that grafting tentacles onto the inner surface of to remove the porogenic agents and other soluble compounds chromatographic media using ammonium ceric nitrate (CAN) present in the polymer monolith. as initiator could afford a high DBC. They thought that DBC of the grafted monoliths depended not only on the surface area 2.2.2. Preparation of high permeable monolith of monoliths, but also on the structure of their surface. These Three high permeable monoliths, noted as MLS-(1-3), were tentacles attached on the surface of monoliths would provide prepared by introducing Na2SO4 granules in the polymerization multiple functionalities emanating from each individual surface system. The conditions and procedure for MLS preparation were site and dramatically increase the ligand density by forming the same as those for ML, except that the porogenic agents were three-dimensional surfaces [21]. composed of both organic solvents and Na2SO4 granules. In In this paper, we report a novel approach for the fabrica- addition, the column filled with the polymerization mixture was tion of macroporous poly(glycidyl methacrylate-ethylene glycol kept in a thermostatic bath at 55 ◦C and rotated inversely at dimethacrylate) (GMA-EDMA) monolith using Na2SO4 gran- 30 cycles per minute to keep the Na2SO4 granules suspending ules and organic solvents as co-porogens. In the process, we homogeneously. The rotation was driven by a mechanical motor, faced the problem of the decrease of protein binding capacity and continued till the polymerization was completed (24 h). with the introduction of the solid granules. To overcome this problem, polymer tentacles were grafted by CAN-induced graft 2.3. Preparation of anion-exchange monolith polymerization using GMA as the monomer. The monoliths were extensively characterized for protein chromatography. The preparation of weak anion-exchange monolith was based on the ring-opening reaction of the epoxy groups on the pore 2. Experimental surface of the monoliths as described in literature [7]. Herein, we converted the monoliths with epoxy groups to the weak 2.1. Materials anion-exchange monoliths with diethylamine by two different approaches, and compared their chromatographic behaviors. Glycidyl methacrylate (GMA) and ethylene glycol dimethacrylate (EDMA) were purchased from Yuanji (Shanghai, 2.3.1. Direct modification of monolith with diethylamine China). Before use, GMA and EDMA were extracted with For the derivation of diethylaminohydropropyl groups, a mix- 10% aqueous sodium hydroxide solution and distilled water, ture of diethylamine and tetrahydrofuran (THF) (1/1, v/v) was respectively, dried over anhydrous magnesium sulfate, and then continuously pumped through the monoliths (ML or MLS) for distilled under vacuum. Benzoyl peroxide (BPO) (95%) was 7 h at 60 ◦C, 0.2 ml/min. Thereafter, the column was washed from Damao (Tianjin, China) and recrystallized before use. routinely with THF, distilled water and 10 mM Tris–HCl buffer Bovine serum albumin (BSA) was purchased from Sigma (St. pH 7.6 (buffer A). Because the reaction of epoxy groups with Louis, MO, USA). Na2SO4 granules were prepared by addition diethylamine was not complete, the residual epoxy groups were ◦ of hot ethanol (50 C) to saturated aqueous Na2SO4 solution at hydrolyzed in 1 M sulfuric acid. 50 ◦C. The fine crystals were recovered by filtration and dried in air. Particle size distribution of the Na2SO4 granules was 2.3.2. Graft polymerization and modification with measured to be 0.7–1.5 ␮m with a Mastersizer 2000 particle diethylamine size analyzer (Malvern, Malvern, UK). Other reagents (such as The grafting and modification of MLS-3 are described in dodecanol, cyclohexanol, ammonium ceric nitrate and diethyl Fig. 1. In the reaction, GMA molecules polymerize to form long amine) were all of analytical grade and used without further poly(GMA) tentacles from the starting radical sites on the sur- purification. face of monolith MLS-3. The reactions included four steps. First, the epoxy groups on the surface of MLS-3 were hydrolyzed to 2.2. Preparation of poly(GMA-EDMA) monolith hydroxyl groups for 5 h at 60 ◦C using 0.5 M sulfuric acid. Sec- ond, the monolith with hydroxyl groups was activated with 0.1 M ◦ 2.2.1. Preparation of conventional monolith CAN in 1 M HNO3 for 4 h at 60 C. Third, the solution for the General monoliths (noted as ML-(1-3)) were synthesized by activation mentioned above was displaced with degassed GMA, a free-radical polymerization of monomer, GMA and a cross- and the graft polymerization proceeded for 20 min at 40 ◦C. In linking agent, EDMA, in the presence of organic porogen in a each step described above, the reactant solution was pumped into stainless-steel tube of 50 mm × 4.6 mm I.D. At first, BPO (1% the column through a syringe with a 5 ml loop, and then the col- (w/w) of GMA) was dissolved in porogenic solvent composed umn was sealed and placed in water bath. The polymerization Author's personal copy

214 K.-F. Du et al. / J. Chromatogr. A 1163 (2007) 212–218

was calculated as the pressure drop across the monolith. Then, hydraulic permeability of the monolith was calculated using the hydrodynamic data from the Darcy’s law [25]: p uμ = (1) L B0 where p stands for back pressure (Pa), L for column length (m), u for superﬁcial velocity (m/s), μ for mobile phase viscosity 2 (Pa s), B0 for hydraulic permeability (m ). The value of B0 can be used as an index for evaluating the quality of monoliths, helping improve the monolith preparation.

2.5.2. Dynamic binding capacity (DBC) of monoliths The DBC of anion-exchange monolith for BSA was determined by frontal analysis [7]. In the experiments, the column was equilibrated by loading buffer A until the absorbance of the outlet stream had reached that of the inlet stream. After the column equilibration, BSA solution in buffer A was pumped to the monolithic column at a defined flow rate and the outlet absorbance at 280 nm was measured. This led to the determina- tion of BSA breakthrough curve. The column was then eluted with buffer A plus 1 M NaCl (buffer B). The DBC of a monolithic column was calculated at 50% of the final absorbance value of Fig. 1. Graft polymerization of glycidyl methacrylate and modification with the breakthrough curve [7]: diethylamine on the surface of poly (GMA-EDMA) monolith. C × V − V q = 0 ( 0) V (2) reaction was terminated by rapid cooling the column in ice- c ◦ box at 4 C followed by extensive washing with pure methanol where C0 represents feed BSA concentration (mg/ml); V the and water connected in an HPLC system. Finally, the mono- volume of BSA solution pumped into the column at 50% break- lith with poly(GMA) tentacles was modified by further reaction through (ml); V0 the dead volume of the HPLC system (ml); Vc with diethylamine to prepare a weak anion-exchange mono- the total column volume (ml); q for the DBC (mg/ml). lith. The modification procedure was the same as that described above. The anion-exchange monolith thus prepared was noted 2.5.3. Column efficiency of monoliths as MLS-3T. The plate theory was used to investigate the homogeneity and efficiency of the monoliths, and the height of equivalent 2.4. Characterization of monoliths theoretical plate (HETP) was determined from the following equation [14]: An XL 30 environmental scanning electron microscope L × σ2 (SEM, Philips, Eindhoven, The Netherlands) was used to exam- HETP = (3) μ 2 ine the microscopic morphology of the monoliths after sputter 1 2 coating gold to the dried monolith samples. The back pressures where L is the column length (mm); σ the peak variance; μ1 of monolithic columns were measured on AKTA¨ Explorer 100 the first absolute moment. BSA (2 mg/ml in buffer B) was used system (GE Healthcare, Uppsala, Sweden). Other experiments as a tracer in the impulse experiment under unretained con- were performed with Waters HPLC system (Waters, Milford, dition. The column was equilibrated with buffer B, and 20 ␮l MA, USA) with a model 600E multi-solvent delivery system, pulse of the tracer solution was injected into the column. Then, a Rheodyne 7725i injector valve (Rheodyne, Cotati, CA, USA) the column was washed with buffer B and the eluted peak was and a 2748 UV detector. The data were acquired and processed recorded at the column exit. The dead volume of the system was with the PC 800 software (Waters, Milford, MA, USA). measured by injecting acetone solution via blank experiments. 2 From the experimental chromatogram, σ and μ1 are calculated, 2.5. Chromatography respectively, from t2c(t)dt 2.5.1. Hydrodynamic properties of monoliths σ2 = − μ2 (4) c(t)dt 1 The flow hydrodynamic behavior of the monoliths was described by the pressure drop of monolithic column at different and flow rates using pure water as mobile phase. The values of the tc(t)dt system pressure without and with the monolith were measured μ = (5) 1 c(t)dt at each flow rate, and the difference between the two values Author's personal copy

K.-F. Du et al. / J. Chromatogr. A 1163 (2007) 212–218 215

where t stands for time (s) and c(t) for BSA concentration at the column exit (mg/ml) as a function of time.

3. Results and discussion

3.1. Porous morphology and permeability of monolithic columns

The permeability of monolithic columns is one of the most practical factors in designing a novel type of monolithic column. Therefore, to increase the permeability, solid granules were introduced into the polymerization system as co-porogen. In the experiment, six monolithic columns (ML-(1-3) and MLS-(1- 3)) were synthesized under different composition of porogenic Fig. 2. Effect of porogen composition on the back pressure as a function of flow agents as listed in Table 1. Fig. 2 represents the effect of flow rate rate for monoliths ML and MLS. Conditions: column 50 mm × 4.6 mm I.D.; on the back pressures of the monolithic columns prepared using mobile phase, distilled water. different porogenic agents. It can be seen that the back pressures of all monolithic columns were linearly proportional to Typical SEM images of the two monoliths ML-3 and MLS-3 the flow rate, which confirms that the porous materials were not are shown in Fig. 3a and b, respectively. From Fig. 3a, it is found compressed even at high flow rate up to over 2000 cm/h, indicat- that ML-3 contains many interconnected channels composed ing their good mechanical property. Moreover, the composition of neighboring nuclei and clusters, and the pore distribution of porogenic agents significantly affected the back pressures of is in the range of 1–3 ␮m. It is known that nuclei can form monolithic columns. As for the ML columns, the back pressures micro-mesopores and further congregate into clusters, which decreased with increasing the percentage of dodecanol. This can form through-pores according to the phase separation mech- is because the dodecanol is a kind of poor solvent for GMA anism [11,12]. When Na2SO4 granules were added into the monomer, so it can result in the increase of large pores, lead- polymerization system, the pore structure of MLS-3 exhibited a ing to the lower back pressure [11,12]. When Na2SO4 granules significant change. As given in Fig. 3b, there are some less uni- were introduced into the polymerization system (MLS), the back form superpores in MLS-3, and the pore size is up to about 6 ␮m. pressures of the MLS columns decreased obviously as compared It is considered that the superporous structure was left by the to the corresponding ML monoliths. Moreover, their back pres- Na2SO4 granules that were removed by solubilization in water. sures changed little with the increase of dodecanol, indicating Considering the particle size distribution of Na2SO4 granules ␮ that the pores formed by Na2SO4 granules contributed much (0.7–1.5 m), some superpores should have been formed by the more than those formed by the organic solvents. To further inves- aggregates of the granules. Obviously, the superpores in MLS- tigate the effect of Na2SO4 granules on the formation of large 3 have contributed to the lower back pressure of the monolith pores, the permeability values (B0) of the six monoliths were (Fig. 2). calculated from the Darcy’s law [Eq. (1)], and the results are listed in Table 1. It can be seen that the values of B0 increased 3.2. Dynamic binding capacity and dynamic adsorption evidently when introducing the solid granules into the polymer- isotherms ization system. So, it is confirmed that the use of solid granules was very favorable for improving the permeability of monolithic To evaluate the mass transfer and dynamic binding properties columns. of the anion-exchange monoliths, breakthrough curves of BSA

Table 1 a Effect of porogen composition on the permeability B0 and DBC of the poly(GMA-EDMA) monoliths prepared at different condition b −14 2 c Monolith DoOH/CyOH (v/v) Na2SO4 granules (g) B0 (10 m ) DBC (mg/ml) Graft ML-1 1/9 0 1.8 33.4 No ML-2 6/19 0 4.2 29.3 No ML-3 1/2 0 5.0 27.5 No MLS-1 1/9 0.25 7.8 26.4 No MLS-2 6/19 0.25 8.1 26.0 No MLS-3 1/2 0.25 8.8 23.9 No MLS-3T 1/2 0.25 7.2 74.7 Yes a Reaction conditions: glycidyl methacrylate (GMA), 26% (v/v); ethylene glycol dimethacrylate (EDTA), 14% (v/v); organic porogenic solvent, 60% (v/v); benzoyl peroxide (BPO), 1% (w/w) with respect to GMA monomer; temperature, 55 ◦C; polymerization time, 24 h. b Volume ratio of dodecanol (DoOH) and cyclohexanol (CyOH)) in the polymerization mixture. c Flow rate: 360 cm/h. Author's personal copy

216 K.-F. Du et al. / J. Chromatogr. A 1163 (2007) 212–218

Fig. 3. SEM images of (a) ML-3, (b) MLS-3, and (c) MLS-3T. in MLS-3 were measured at different flow rates, and DBC values less than 1 ␮m) between the through-pores and reach adsorption were calculated from Eq. (2) (Fig. 4). It is found that with the equilibrium in short time [26]. The favorable property afforded increase of flow rate, the DBC of MLS-3 decreased slightly at the the monolithic column high resolution and DBC regardless of beginning, and then changed little when the flow rate exceeded flow rate. The inset in Fig. 4 displays the breakthrough curves in 500 cm/h, suggesting that the DBC was independent of flow rate. a wide range of flow rate. The sharp increase of the breakthrough It is deduced that the interconnected through-pores in the mono- frontals also confirmed the excellent mass transfer behavior of lith led to fast mass transfer, and the accessible microporous the monolithic column. channels on the surface of the monolithic skeleton provided To investigate the effect of porogenic agents on the DBC of the active sites for protein adsorption [12]. Protein molecules monoliths, the breakthrough curves of the six monoliths were could easily diffuse into the short microporous channels (mostly determined at the same flow velocity (360 cm/h) and their DBC values thus obtained were listed in Table 1. It can be seen that the DBC of ML monoliths decreased from 33.4 mg/ml (ML-1) to 29.3 mg/ml (ML-2) and 27.5 mg/ml (ML-3) along with the reduction of cyclohexanol. This is because the cyclohexanol is a kind of good solvent for the monomers, whose increase led to the decrease of pore size, resulting in the low column permeability. With Na2SO4 granules in the monolith preparation, the DBC of the MLS columns decreased slightly. At the same time, the column permeability B0 increased drastically to about 8 × 10−14 m2. It is considered that the superpores introduced by Na2SO4 granules improved the convective transport for protein solution and increased the permeability at the expense of reduc- ing the accessible active sites due to the decrease of surface area. In order to further improve the DBC of MLS without decreasing their permeability, the MLS-3 was grafted with GMA and Fig. 4. Effect of flow rate on the DBC of MLS-3. The inset represents the breakthrough curves obtained at different flow rates. C is the outlet BSA concentration further modified with DEA (Fig. 1) to produce MLS-3T. It and C0 is the feed protein concentration (2 mg/ml). Conditions: 50 mm × 4.6 mm was confirmed that the graft polymerization should be termi- I.D.; feed, 2 mg/ml BSA in buffer A; mobile phase, buffer A, eluent, buffer B. nated within 20 min, because longer reaction time would jam Author's personal copy

K.-F. Du et al. / J. Chromatogr. A 1163 (2007) 212–218 217

Fig. 5. Breakthrough curves of BSA in MLS-3 and MLS-3T at flow rate of Fig. 7. Effect of flow rate on the back pressures of ML-3, MLS-3 and MLS-3T. 1 ml/min (360 cm/h). Conditions: column, 50 mm × 4.6 mm I.D.; feed, 2 mg/ml Conditions: column 50 mm × 4.6 mm I.D.; mobile phase, distilled water. BSA in buffer A; mobile phase, buffer A, eluent, buffer B. at a constant flow rate (1 ml/min) with different BSA load- the channels of monolith due to the excessive grafting reaction. ing concentrations. The dynamic adsorption isotherms can be In addition, the grafting polymerization could result in the for- described by the Langmuir equation [7]: mation of thin film, which grew from the surface of pores in the monolith, as observed in Fig. 3c. Because MLS-3 has a macrop- qmc q = (6) orous structure, the grafted tentacles could expand the adsorption Kd + c zone from the limited surface area to the internal pore volume where q stands for the DBC, q for the dynamic adsorption and provided the possibility of multilayer protein adsorption, m saturation capacity, c for the BSA concentration in buffer A, which is favorable for high adsorption capacity [20]. and K for the apparent dissociation constant. Calculated from To evaluate the effect of the graft polymerization on protein d Eq. (6), the dynamic capacities (q ) of ML-3, MLS-3 and adsorption and back pressure of the monolith, BSA breakthrough m MLS-3T monolith were estimated at 28.3, 25.1 and 78.7 mg/ml, curves and flow hydrodynamics were measured on the monolith. respectively. Compared with ML-3, the dynamic capacity of Fig. 5 shows the breakthrough curves of the MLS-3T and MLS-3 MLS-3 decreased approximately 11% because of the propor- at 1 ml/min. From the breakthrough curves, it was calculated that tional increase of superpores created by Na SO granules. This the DBC of MLS-3T was high as 74.7 mg/ml, over three times 2 4 confirms that the template method can greatly enhance mono- higher than that of MLS-3 (23.9 mg/ml, Table 1). However, the lith permeability without significant loss of adsorption ability. slope of the breakthrough curve from the MLS-3T was less steep The grafted monolith (MLS-3T) had a dynamic capacity as high than that from the MLS-3. These results suggest that the grafted as 78.7 mg/ml, over three times higher than that of MLS-3. tentacles have enhanced the multipoint adsorption of the protein, This suggests that the tentacles contributed greatly to the high leading to the increase of protein binding capacity. At the same capacity of MLS-3T. time, the grafting also resulted in the increase of mass transfer To check the recovery of protein from the columns, BSA resistance. was eluted with buffer B from the monoliths MLS-3 and MLS- Fig. 6 shows the dynamic adsorption isotherms of ML-3, 3T after breakthrough experiments. As a result, 91% and 96% MLS-3 and MLS-3T, which were determined by frontal analysis of BSA were recovered from MLS-3 and MLS-3T, respectively. The high recovery yield of BSA from MLS-3T column indicates the absence of non-specific protein absorption by the grafted tentacle chains. The back pressure versus flow rate curves of ML-3, MLS-3 and MLS-3T are shown in Fig. 7. It is obvious that the back pressure of MLS-3T was only slightly higher than that of MLS- 3, but much lower than that of ML-3. It is deduced that the MLS-3T can uptake more water than MLS-3 because of the hydrophilicity of the tentacles, leading to higher flow resistance. This is in accordance with previous observation [21].

3.3. Column efﬁciency

The column efﬁciencies of the monoliths were investigated by Fig. 6. Dynamic adsorption isotherms of BSA on ML-3, MLS-3 and MLS-3T. impulse injection experiments using BSA as tracer under unre- Conditions: column 50 mm × 4.6 mm I.D.; ﬂow rate 1 ml/min. strained condition. As shown in Fig. 8, the values of HETP for all Author's personal copy

218 K.-F. Du et al. / J. Chromatogr. A 1163 (2007) 212–218

in the preparation of MLS-3T with similar back pressures and DBC values.

4. Conclusions

A novel type of monolithic column has been fabricated using a solid-templating strategy combining with phase separation. By introduction of Na2SO4 granules as a co-porogen with organic solvents, the monolith permeability was greatly improved without obviously decreasing the dynamic binding capacity of protein. Further graft polymerization of the MLS- 3 monolith with GMA created tentacle-type polymer chains on the pore surface of the high-permeability monolith, lead- Fig. 8. Dependence of column efficiency on mobile-phase flow rate for ML-3, ing to a three-fold increase of DBC. Moreover, the monolith MLS-3 and MLS-3T. Conditions: column 50 mm × 4.6 mm I.D.; feed 2 mg/ml of high permeability and high capacity showed acceptable col- BSA in buffer B; mobile phase, buffer B; injection size, 20 ␮l. umn efficiency at high flow rates. So, the approach presented in this article would be a promising way to the fabrication of monolithic columns for high-performance preparative protein the three monoliths decreased with increasing flow rate and kept chromatography. nearly constant at flow rate higher than about 1500 cm/h. This can be explained by the nature of the inner macroporous chan- Acknowledgments nel structures in the monoliths. Generally, the transport of the large molecules was driven by convective flow in the channels, This work was supported by the Natural Science Foundation and they approached the micropores between the channels by of China (No. 20476082 and No. 20636040), and the Program diffusion. Increasing flow rate could increase the back pressure for Changjiang Scholars and Innovative Research Team in Uni- (Fig. 2), leading to the occurrence of convective flow in more versity from the Ministry of Education of China. small-sized channels [14]. So, the column efficiency increased with increasing flow rate at low flow rate range. At higher References flow rates, more convective flow occurred through the monolith. Because the distance between the convective-flow channels [1] B. Gu, J.M. Armenta, M.L. Lee, J. Chromatogr. A 1079 (2005) 382. (that is, the length of diffusion path) was quite short, pore dif- [2] A. Palm, M.V. Novotny, Anal. Chem. 69 (1997) 4499. fusion did not play a significant role in affecting the column [3] M.J. Benes,ˇ D. Horak,´ F. Svec, J. Sep. Sci. 28 (2005) 1855. efficiency [26]. [4] C. Viklund, A. Nordstrom,¨ K. Irgum, Macromolecules 34 (2001) 4361. It is obvious from Fig. 8 that the column efficiency of the [5] T. Umemura, Y. Ueki, K. Tsunoda, A. Katakai, M. Tamada, H. Haraguchi, Anal. Bioanal. Chem. 14 (2006) 566. three monoliths was different, possibly due to the difference in [6] K. Cabrera, D. Lubda, J. High Resolut. Chromatogr. 23 (2000) 93. their pore and surface structures. The pore sizes and propor- [7] M. Zhang, Y. Sun, J. Chromatogr. A 912 (2001) 31. tion of flow-through pores of ML-3 were smaller than those [8] A. Podgornik, M. Barut, A Strancar,ˇ Anal. Chem. 72 (2000) 5693. of MLS-3, resulting in higher over mass transfer resistance in [9] T.B. Tennikova, B.G. Belenkii, F. Svec, J. Liq. Chromatogr. 13 (1990) 63. ML-3 than in MLS-3 [21]. Therefore, the column efficiency [10] F. Svec, J.M.J. Frechet,´ Anal. Chem. 64 (1992) 820. [11] F. Svec, J.M.J. Frechet,´ Chem. Mater. 7 (1995) 707. of ML-3 was lower than that of MLS-3. For MLS-3T, the [12] C. Viklund, F. Svec, J.M.J. Frechet,´ Chem. Mater. 8 (1996) 744. grafted polymer chains on the pore surface would hinder pro- [13] G.S. Chirica, V.T. Remcho, J. Chromatogr. A 924 (2001) 223. tein mass transfer, leading to the lower column efficiency of [14] Q.H. Shi, X. Zhou, Y. Sun, Biotechnol. Bioeng. 92 (2005) 643. MLS-3T. [15] G.Y. Sun, Z. Yang, X.Y. Dong, Y. Sun, J. Appl. Polym. Sci. 103 (2007) 17. We have also checked the reproducibility of the monolith [16] Y. Shi, X.-Y. Dong, Y. Sun, Chromatographia 55 (2002) 405. [17] Y. Shi, Y. Sun, Chromatographia 57 (2003) 29. preparations (MLS-3 and MLS-3T). Nine monoliths of MLS- [18] J. Dai, Z. Bao, G.L. Baker, M.L. Bruening, Langmuir 22 (2006) 4274. 3 were prepared and their chromatographic properties such as [19] I.N. Savina, B. Mattiasson, I.Y. Galaev, Polymer 46 (2005) 9596. flow hydrodynamics and column efficiency were measured. The [20] I.N. Savina, I.Y. Galaev, B. Mattiasson, J. Chromatogr. A 1092 (2005) 199. results demonstrated that the present method could produce [21] K.H. Gebauer, J. Thommes,¨ M.R. Kula, Chem. Eng. Sci. 52 (1997) 405. monoliths with good reproducibility. Moreover, the homogene- [22] E. Muller,¨ Chem. Eng. Technol. 28 (2005) 1295. [23] C. Viklund, F. Svec, J.M.J. Frechet,´ Biotechnol. Prog. 13 (1997) 597. ity of the monolith was confirmed by SEM observations of the [24] E. Muller,¨ US Pat. No. 5 647 987 (1997). monolith at different positions after removing it from the col- [25] C. Martin, J. Coyne, G. Carta, J. Chromatogr. A 1069 (2005) 43. umn. In addition, the graft polymerization could also be repeated [26] F. Svec, J.M.J. Frechet,´ Ind. Eng. Chem. Res. 38 (1999) 34.