J. Phys. Chem. 1994,98, 7599-7601
7599
Influence of Cross-Linking upon the Macroscopic Pore Structure of Cellulose S. Tasker and J.
P.S. Badyal'
Department of Chemistry, Science Laboratories, University of Durham, Durham DH1 3LE. England, U.K. Received: February 8, 1994; In Final Form: May 12, 1994'
Pore size distributions of cellulose have been determined using nitrogen adsorption isotherms for differing levels of cross-linking. Macropores located within the cellulose structure are readily altered during cross-linking, whereas the associated mesopores are found to be much more stable.
Introduction
Experimental Section
Cellulosic materials can be employed as effective, low-cost ion exchange chromatography media for protein separation.14 Chemical derivatizationof cellulose is relatively straightforward, thereby enabling a wide range of functionalities to be grafted directly onto the host polymer matrix, which in turn can impart many new and exciting properties.3 Functionalization is primarily governed by the accessibility of cellulose hydroxyl groups and is dependent on short-range ~rystallinity.~Highly crystalline cellulose consists of a close packed supramolecularmicrofibrillar structure, which hinders the penetration of large and/or poorly charged species into the cellulose matrix; this results in a small number of readily accessible reactive sites and, therefore, ultimately a low protein binding ~ a p a c i t y .The ~ level of matrix functionalization is found not to affect its performance as a chromatography medium6 (except in extreme cases, where derivatization fallsbelow an effective threshold). Related studies have demonstrated that the conformational arrangement of the parent polymer can strongly influence protein binding; for instance, hydrophobic folding of the matrix at a high pH gives rise to pore shrinkagein conjunctionwith a concomitantdecrease in accessibility of the charged centers for protein attachment.2 It is the fibrillar nature of cellulose which gives rise to its p o r ~ s i t y . ~Mesoporosity is due to inefficient packing of the cellulosic microfibrils, yielding intrafibrillar pores (25-75 A). Larger pores (500-1 5 0 0 A) are present in between the hydrogenbonded fibrillar network which constitutes the fibers of cellulose. Porosity is usually introduced during manufacturing, and unbeaten fibrous stock forms a more porous network than a heavily beaten material.* The structural and swelling characteristicsof cellulose are important attributes of this type of chromatographic medium.9-14 Cross-linking, drying,'>'' and chemical treatment of the base material can disrupt the hydrogen bonding in cellulose and thereby reduce interfibrillar cohesion, which causes structural rearrangement and redistribution of pore architecture.18J9 Conventional methods used for probing the porosity of cellulose are selected on the basis of pore sizes of interest. Mercury porosymmetry7-a.21is used to measure cellulose pore apertures between 0.1 and 100 pm. Water?223benzene,17 ethanol,I6 and nitrogen9 sorption have been used to evaluate pores in the 10-100 A regime. The solute exclusion method24enables pores in the 100-1OOO A range to be characterized, but this is subject to only a limited degree of accuracy. In this article, the porosity of highly amorphous cellulose has been studied as a function of bulk crosslinking. Porosity values obtained from nitrogen adsorption isotherms are used to track changes taking place in the cellulose pore structure during cross-linking treatment.
Sample Preparation. Regenerated sponge cellulose (BPS SeparationsLtd.) derived from wood pulp and made via the viscose process2$was used as the base material (sample I). This was found to contain a relatively high amorphous content (approximately 49% crystallinity as determined by XRD). Epichlorohydrin in aqueous base solutions was used as a cross-linking reagent for cellulose.2"29 Cross-linking samples 11-V were prepared by using 0.5%. 1.0%, 1.5%, and 2.0% by volume concentrations of epichlorohydrin, respectively. Each treated sample was subsequently rinsed in water, freeze-dried,and then stored under vacuum at 50 "C for 2 weeks prior to analysis. Freeze drying minimizes pore shrinkage, collapse,and distortion.30 Pore Structure Analysis. Nitrogen adsorption isothermswere measured for each cellulose sample using a PMI Brookhaven Sorption apparatus. The Brunauer, Emmett, and Teller3l.32 (B.E.T.) method for specific surface area determination was employed. Pore size distributions were calculated from the adsorption region of the isotherm using the modified33procedure developed by Pierce.34 This uses the "Kelvin equation" (l), which relates the vapor pressure (p) of a condensed liquid in a cylindrical capillary to the radius of curvature of a hemispherical meniscus
* To whom correspondence should be a d d r d .
Abstract published in Advance ACS Absrracrs, July 1, 1994.
0022-3654/94/2098-1599$04.50/0
(4*
Herepo is the saturation pressure of the adsorbate, y the surface tension, Y the molar volume of the liquid, and 6 the angle of contact between the liquid and the pore walls (cos $t = l).20 Since r increases as p/po increases, it follows that condensation occurs at lower relative pressures in smaller radii pores. However, r is not the radius of the pore, and therefore the thickness of the adsorbed layer ( t ) on the walls of the pore must be taken into account using the "Halsey equation" (2).
where IJ is the thicknessof an adsorbed monolayer (inangstr6ms). This then allows calculation of the critical pore radius (rP,J,
rp,c= r + t = -(2Vy cos d / R T ln(p/p,))
+t
(3)
Incremental gas adsorption on the surface of cellulose pores eventuallyleads to pore closure?3 and differencesin the quantity (Ah') of adsorbed gas at different relative pressures can be expressed as an equivalent volume of liquid AVL; this can then be used to calculate the pore size distribution from the adsorption isotherm using (4). Cross-sectionalelectron microscopy of swollen cellulose fibres suggest a "honeycomb" structure of cylindrical pores.12 Therefore cylindrical pore shapes have been assumed: 0 1994 American Chemical Society
7600 The Journal of Physical Chemistry, Vol. 98, No. 31, 1994
Tasker and Badyal
Here i, 7, r P f ,and At correspond to average values of the Kelvin radius, adsorbed film thickness, critical pore radius, and the change in adsorbed film thickness for the increment in relative pressure, respectively; AVp and tprelate to pore volume and average pore radius within a specific pore size interval. By plotting A V P / A ~ ~ . ~ versus rp,c,one can obtain a pore radius distribution (or diameter).
Results Uncrossed-Linked Cellulose. The pore size distribution for untreated cellulose (sampe I) is shown in Figure 1; this is typical of cellulosic materials.17 Most of the pore volume is located at less than 100 A aperture diameter, while the majority of pores are located around 30 A. In addition, there are some large pores present at approximately 600, 800, and 1100 A. Crcws-LinkedCellulose. Reaction of cellulosewith OS%, 1.O%, 1.5%, and 2.0% by volumeof epichlorohydrin cross-linking solution (samples 11-V, respectively) causes a reduction in macroporosity with increasing cross-linker concentration. Figure 2 shows that as thecross-linking reagent concentration is increased, the number of pores a t 600A are reduced to approximately l / 3 of their original value, beyond which they experience no further perturbation. The pores in the 800 A range virtually disappear at low crosslinker concentrations. The pore volume corresponding to 1100 A apertures is slightly lowered by 0.5% cross-linking reagent; however, there isa rapid attenuation of thesemacroporesat higher concentrations. The meso- and microporosity regions remain virtually unperturbed by the epichlorohydrin solution with respect to the uncross-linked cellulose material (sample I).
50
Pore Diamet er(A) Figure 1. Overall pore size distribution of untreated cellulose material (sample I).
Discussion Cross-linking with epichlorohydrin solution can strongly alter the macroporous nature of cellulose, while the mesopore structure remains intact. In the macroporous range, the 600 A pores partially disappear during cross-linking whereas the 800 and 1100 A pores completely collapse a t higher cross-linker concentrations. The cross-sectional area associated with a cylindrical 1100 A diameter pore is approximately 3 times greater than a pore with a diameter of 600 A. The bonding required to sustain a specific pore structure increases with pore diameter size." Smaller pores tend to be far more robust, since they comprise microfibrillar voids and are stabilized by short-range hydrogen bonding. Macropores are formed by fibrillar packing; this is a more loosely bound typeofstructure, which leads to weaker fibrillar association. The latter are usually associated with swelling in various solutions~4 and can give rise to large pore collapse during drying.16J7 Cross-links within the larger pores may be expected to have a much more detrimental effect on the pore stability compared to a similar number of cross-links introduced into a microfibrillar environment. As well as restricting the swelling of cellulose fibres in the wet state, the cross-linking process can also be expected to influence the mechanical26 and polyelectrolyte (e.g. protein) adsorption characteristics of the cellulose matrix.3s For instance, Martin and Rowland36 used partial separation of pairs of sugars with differing molecular weights as a measure of the effect of cross-linking upon cellulose. After cross-linking with formaldehyde, the amorphous cellulose showed reduced permeability to larger molecules but increased permeability to molecules with molecular weights less than 1000.
500 1000
100
I
600
"
1
900
1
'
1
'
'
1200
Pore Diameter(A) Figure 2. Macropore size distributions of cellulose having undergone differing degrees of cross-linking obtained by using O.O%, 0.596, 1.096, 1.5%, and 2.0% by volume concentrations of epichlorohydrin (corresponding to samples I-V, respectively).
causes a preferential disruption of the macropores rather than the mesopores contained within the host matrix.
Acknowledgment. S.T. thanks the SERC and BPS Separations Ltd. for financial support during the course of this work. References and Notes Kennedy, J. F.; Paterson, M. Polym. Int. 1993, 32, 71. Chen, H.-L.; Hon, K. C. React. Polym. 1987, 5, 5 . Callstrom, M. R.; Bednarski. M. D. MRS Bull. 1992, Oct, 54. Kremer, R. D.; Tabb, D. Int. Lub. 1989, July/August, 40. ( 5 ) Rowland, S. P.; Howley, P. S. J . Polym. Sci.: Part A: Polym. Chem. 1985, 23, 183. (6) Woffindin, C.; Hoenich, N. A.; Matthews, J. N . S. Nephrol. Dial. Transplant 1992, 7, 340. (7) Jeffries, R.; Jones, D. M.; Roberts, J. G.; Selby, K.; Simmens, S. C.; Warwicker, J. D. Cellul. Chem. Technol. 1969, 3,255. (8) Dullien, F. A. L. Porous Media: Fluid Transport andPoreStructure; Academic Press: New York, 1979. (9) Porter, B. R.; Rollins, M. L. J. Appl. Polym. Sci. 1972, 16, 217. (10) Gert,N. V.;Torgashov,V. I.;Shishonok,M.V.;Sinyak,S. J.;Kaputsii, F.N. J . Polym. Sei.: Part B Polym. Phys. 1993, 31, 567. (1 1) Lokhande, H. T.; Thakare, A. M. J. Polym. Sci.: Part C: Polym. Lett. 1990..~ 28.,~ 21. (12) Aravindanath, S.; Bhama Lyer, P.; Sreenivasan, S. J. Appl. Polym. Sci. 1992, 46, 2239. (13) Buschle-Diller, G.; Zeronian, S. H. J . Appl. Polym. Sci. 1992, 45, (1) (2) (3) (4)
~
Conclusions Nitrogen adsorption isotherms have shown that chemical crosslinking of regenerated spongecelluloseby epichlorohydrinsolution
9/11
(14) El-Din, N. M. S. Polym. Int. 1993, 32, 13. (15) Weatherwax, R. C.; Caulfield, D. F. Tappi 1971, 54, 985.
Cross-Linking in the Pore Structure of Cellulose (16) Weatherwax, R. C. J. Colloid. Int. Sci. 1977,62,3, 432. (17) Alina, B.; Colloid Polym. Sci. 1975,253,720. (18) Rowland, S.P.; Wade, C. P.; Bertoniere, N. R.J. Appl. Polym. Sci. 1984,29,3349. (19) Rowland, S.P.; Wade, C. P.; Bertoniere, N. R. J. Appl. Polym. Sci. 1986,31, 2769. (20) Gregg, S.J.; Sing, K. S.W. Adsorption,Surface Area and Porosity; Academic Press: Londong 1967. (21) Stbm, G.; Carlsson, G. J. Adhes. Sci. Technol. 1992,6, 745. (22) Zeronian, S.H.; Coole, M. L.; Alger, K.W.; C. Chandler, J. M. J. Appl. Polym. Sci.: Appl. Polym. Symp. 1983,37, 1053. (23) Weatherwax, R. C. J. Colloid Inr. Sci. 1974,49, 40. (24) Stone, J. E.;Trieber, E. E.; Abrahamson, B. Tappi 1%9,52, 108. (25) Trieber, E. E. In Cellulose Chemisrry andlts Applications;NeveII, T. P., Zcronian, S.H., Eds.; Wiley: New York, 1985. (26) Decring, Milliken Research Corp. Bristish Patent No. 8834/57,1960. (27) BPS Separations Ltd. Eur. Pat. Appl. 91909759.2.
The Journal of Physical Chemistry, Vol. 98, No. 31, 1994 7601 (28) Guthrie, J. D.; Bullock,A. L.Ind. Eng. Chem. 1960,52,11. (29) Guthrie, J. D.Ion Exchangers in Organic and Biochemistry;Interscience: New York, 1957;Chapter 30. (30) Hall, J. L., Ed. Elecrron Microscopy and Cytochemistryof Plant Cells; Elsevier: Amsterdam, 1978. (31) Brunauer, S.;Emmet, P. H.; Teller, E. J. Am. Chem. Soc. 1938,60, 309, (32) Brunauer, S.;Dedng, L.S.;Deming, W.S.;Teller, E. J . Am. Chem. Soc. 194,62,1723. (33) Orr, C.; Dalla Valle,J. M. Fine Particle Measurement;Macmillan: London, 1959;p 271. (34) Pierce, C. J. Phys. Chem. 1953,57, 149. (35) Peterson, E. A. Cellulosic Ion Exchangers;Elsevier: Amsterdam, 1980:Chapter 2. (36) Martin, L.F.;Rowland, S.P. J. Polym. Sci.: Part A: Polym. Chem. 1967,5,2563.
