Production and characterization of membranes of recycled waste materials: cellulose acetate, obtained from sugarcane bagasse with polystyrene from plastics cups.
Brazil is the world's major sugarcane producer and the 2006/2007 harvest reached 457 million metric tons. Because of federal incentive, this production tends to increase since many new sugarcane mills will be assembled. Only in the region known as Triangulo Mineiro, 13 new sugarcane mills will be installed until 2010. Two of these will be assembled in Uberlandia, where our group is situated. This increase in the production is driven by the increase in the use of ethanol as a biofuel. One of the residues generated from this activity is sugarcane bagasse (SCB), which makes up for around 30% in weight of sugarcane. In the literature, several papers can be found regarding the utilization of this residue (1-6); one of the alternatives for utilizing it is the production of cellulosic derivatives. A derivative of great industrial importance is cellulose acetate, which can be used, for instance, for producing membranes for separation processes such as reverse osmosis, hemodialysis, and controlled release of drugs (7-10).
The technology of separation by membranes has become important due to operational simplicity, using of compact modules, and low energy demand (10). In chemical and pharmaceutical industries, membranes are used in processes such as ultrafiltration (UF), microfiltration (MF), reverse osmosis (RO), nanofiltration (NF), pervaporation, dialysis, gas separation, and electrodialysis. The main difference among these processes is their separation mechanism (11). For example, UF and MF are driven by difference of pressure, reaching 2 atm in MF and 10 atm in UF. The size of the particles that can be retained is determined by the kind of membrane being used. MF membranes can retain particles with molecular weight higher than 500 kDa, while UF membranes retain particles with molecular weight ranging from 5 to 500 kDa. RO and membranes, which also operate by difference of pressure, retain solutes of low molecular weight (ions) (11). Other separation processes are driven by difference of concentration or potential. However, in each of these processes the membrane has a key role in the efficiency of separation and, depending on its characteristics, limits the choice of a process for a given application (10).
Recently, we have shown the viability of producing membranes of cellulose acetate from SCB (12) and blends of cellulose acetate with polystyrene from plastic cups (13), (14). In these previous studies with the cellulose acetate/polystyrene system, blends were produced at the concentration 6% w/w. These were studied from the point of view of compatibility, interactions at the molecular level between the components of the system, and transport properties: water flux and water diffusion through membranes. On the other hand, it is important to characterize these membranes in the sense of knowing its transport properties to point it for a given process. Thus, in this work, membranes of cellulose acetate from SCB cellulose and blends with polystyrene were produced and characterized regarding ion diffusion by dialysis, pure water rate properties, and polyethylene glycol (PEG) rejection. In addition, these membranes were thermally evaluated by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC), The morphology of the membranes' cross-sections was evaluated by scanning electron microscopy (SEM).
Production of Cellulose Acetate
Bagasse was purified in accordance to the procedure described by Filho et al. (12) Cellulose acetate was produced through homogeneous acetylation of cellulose from SCB cellulose, which uses acetic anhydride as acetylating agent, acetic acid as solvent, and sulfuric acid as catalyst (13). The degree of substitution of the produced material was 2.79, and therefore the material is characterized as cellulose triacetate. The viscometric molar weight ([M.sub.v]) is 48.000 g [mol.sup.-1]. [M.sub.v] was determined using dichloromethane/ethanol (8/2) for which K = 13.9 X [10.sup.-3] mL [g.sub.-1] and a = 0.834 (15). For polystyrene, [M.sub.v] was determined as 96,398 g [mol.sup.-1] and toluene was used as solvent. K = 3.8 X [10.sup.-5] and a = 0.630 (16).
Production of Membranes
Cellulose acetate was dissolved in dichloromethane (12%w/w). This solution was stirred for 72 h, at 20[degrees]C, and then cast on a glass dish with a cast knife set at 300 [mu]m, at 20[degrees]C. The time of solvent evaporation was 150 s at 20[degrees]C, after which, the dish was dipped into a distilled water bath at 10[degrees]C to detach the membrane. The average thickness of this membrane was 39.0 [+ or -] 1.9 [mu]m. Blends were produced utilizing 0, 10, 30, and 50% (w/w) of PS in relation to CA, which from now on will be referred to as CA, PS 10. PS30, and PS50, respectively.
DSC and TGA Experiments
DSC experiments were performed in a DSC-50, Shimadzu. Sealed aluminum crucibles containing 10 mg of the samples were heated and cooling rates of 10[degrees]C [min.sub.-1] and nitrogen flow of 50 [cm.sup.3] [min.sup.-1].
The crystallinity of the samples was determined according to Eq. 1.
%C = [[DELTA][H.sub.m]/[DELTA][H.sub.m.sup.o]]x100 (1)
where %C is the crystalline content, [DELTA][H sub.m] is the enthalpy of fusion of the sample, [DELTA][H sub.m.sup.o] is the enthalpy of fusion of a perfect crystal of cellulose acetate, which according to Cerqueira et al. (17) is 58.8 J [gsup.-1].
TGA experiments were performed in a TGA-50, Shimadzu. Ten milligram of the samples were heated from room temperature to 600[degrees]C at a rate of 10[degrees]C [min.sup.-1] under nitrogen atmosphere.
Diffusion of Ions
For these experiments, it was used a two-compartment system separated by the studied membrane. One of the compartments was filled with deionized water, and the other with KC1 solution ([10.sup.-1] [mol [L.sup.-1]). A calibration curve was built, by measuring the conductivity of several concentrations of KCI solutions. This curve was posteriorly used to calculate the concentrations of KC1 during the diffusion experiments. From the slope of the concentration in function of time, the flow (J) through the membrane was calculated. The permeability coefficient (P) was calculated from Eq. 2 (18).
P = [J/[DELTA]C] (2)
Where [DELTA]C is the concentration difference between the two compartments.
The diffusion coefficients (D) through the membranes were calculated using the Eq. 3.
D = P x d (3)
where d is the membrane thickness.
Measurement of Pure Water Permeation Rate and Rejection to PEG
Measurements of pure water permeation rate through the membranes were carried out to test their resistance to pressure. The flow was calculated using Eq. 4.
[J.sub.w] = [Q/A[delta]T] (4)
Where [J.sub.w] is the water flow ([Lm.sup.-2] [h.sup.-1]), Q is the permeate volume (L), A is the area of the membrane ([m.sup.2]). and [DELTA]T is the time of permeation (h) (10).
Measurements of rejection to PEG were carried out to know the size of particles that can be rejected by the membranes, i.e., estimate its molecular weight cut-off. The test was performed using PEG solutions at the concentration of 1% w/v, using PEG of different molecular weights (45 and 80 kDa). During the rejection tests the concentration of PEG in the permeate was determined by measurements of refraction index, using a calibration curve. Rejection was calculated by Eq. 5.
%R = 1 - [[C.sub.p]/C.sub.0]] x 100 (5)
in which [C.sub.p] and [C.sub.0] are the concentrations of the permeate and initial solutions, respectively (10).
Films were fractured in liquid nitrogen and mounted on a metal stub with a double-sided adhesive tape. The cross-section and surface of the membrane were initially gold coated in a Bal-Tec SCD-050 and the microscopies were obtained in a Jeol/Scaning Eletron Microscope, JSM 6060, operated at 10 kV.
RESULTS AND DISCUSSION
Figures 1 and 2 show TGA plots and the DSC thermogram, respectively, for CA membrane and CA/PS blends, in distinct compositions of PS.
[FIGURE 1 OMITTED]
In Fig. 1, a loss of mass in the range between 300 and 400[degrees]C is observed, which is attributed to the degradation of cellulose acetate (14), (19) and other loss of mass in the range from 390 to 490[degrees]C due to rupture of chain of PS (14), (20). The results show that as PS is added to the membranes, the temperature in which degradation starts shifts for higher temperatures, thus, improving the thermal stability of the system.
The DSC thermogram, Fig. 2, shows that the enthalpies of fusion are 12.12, 20.19, 15.35. and 11.02 J [g.sub.-1] for CA, PS 10, PS30, and PS50, respectively. These values of enthalpy of fusion correspond to 20.61, 34.34, 26.10, and 18.74% of crystallinity for CA, PS10, PS30, and PS50 of PS, respectively.
[FIGURE 2 OMITTED]
PS10 shows the highest increase in crystallinity, which corresponds to about 70%. Previous results shows that PS is disperse in the matrix, cellulose acetate, fundamentally creating microregions of compatibility where reordering effects take place. Such reordering effects are fundamentally produced by van der Waals forces. However, such forces are strong enough to have influence in some of the hydrogen bonds present in cellulose acetate, as demonstrated in our previous paper (14). On the other hand, PS30 shows a decrease in crystallinity, what may be explained by the beginning of the phase separation process in the blend, what would affect the compatibility microregions and thus result in an increase of the amorphous character of the material (14) since with the addition of PS to the system there is a phase inversion in the blend. That turns PS in the dispersing phase, and thus less crystalline. Such characteristics have influence on the transport phenomena through the blend, as discussed in the next sections, but not in a linear way since both the crystallinity effects and the increase in the hydrophobic character of the system with the increase of PS content must be evaluated.
Diffusion of Ions
The ion diffusion coefficient calculated for the membranes are presented in Table 1. The ion diffusion coefficient of CA membrane is comparable to that found in the literature for membranes of commercial cellulose triacetate, 8.47 X [10.sup.-8] c[m.sup.2] [S.sup.-1] (18). The ion diffusion coefficients decrease with the increase of PS content in the blends. The behavior observed for the diffusion of ions through these membranes is related to the presence of PS in the membranes.
TABLE 1. Ion diffusion for CA membrane and CA/PS blends. Membrane D([cm.sup.2][s.sup.-1]) CA 3.12 X [10.sup.-8] CA90/PS10 2.42 X [10.sup.-9] CA70/PS30 1.65 X [10.sup.-9] CA50/PS50 1.51 X [10.sup.-9]
CA, which is 20.61% crystalline, presents the highest ion diffusion. As 10% PS is added (PS10), the ion diffusion. coefficient decreases in relation to CA. That can be attributed to the increase crystallinity to 34.34% which hampers the ion diffusion through the membrane. Membranes PS30 and PS50 also present a decrease in the ion diffusion coefficient due to the increase in PS content. However, the crystallinity of these membranes are 26.10 and 18.74%, respectively. In this case, the crystallinity effect may be being counterbalanced by the increase in PS content, which makes the membranes less permeable due to the increase in their hydrophobicity.
The cross section SEM microcopies of the membranes are shown in Fig. 3. CA membrane shows the presence of flake like structures. On the other hand, as PS is added, the structure becomes more compact, and thus, the diffusion of ions through these membranes would be more difficult, corroborating the data obtained in the ion diffusion experiment.
[FIGURE 3 OMITTED]
Pure Water Permeation Rate and Rejection to PEG
In a previous paper (13), the use of PS for producing blends with cellulose acetate was pointed as a viability of recycling both materials, since the experiments of water vapor transport, which included the comparison with nanofiltration and reverse osmosis commercial membranes, were enough to indicate the necessity of further steps in which data regarding water permeation and solute rejection were important to determine the properties for the use of the membranes in separation processes.
Therefore, from the experiments carried out in the present article, it was possible to obtain useful results for classifying the membranes according to the operational pressure and with the molar weight of the solutes.
During permeation experiments the blend membranes broke when the lowest operational pressure of the pump was utilized, 0.5 atm. Membranes of pure cellulose acetate resisted up to 1.5 atm and presented a good performance at 1 atm. The breaking of the membranes containing PS must be explained by the fact that although the addition of 10% PS to the system improved some properties, such as crystallinity, this improvement occurs fundamentally in microregions of compatibility, not being enough to produce adequate mechanical resistance to resist to the used pressures. That happens because PS, even in low concentration, will weaken some of the hydrogen bonds of cellulose acetate.
The pure water permeation rate through the membrane CA was calculated during 1 h of operation, and a diminishing flow was observed, as shown in Fig. 4. This diminishing is related to a compaction of the membrane caused by the pressure during operation (10), (21). This compaction is also observed on the micrographies shown in Fig. 5, which shows the fracture of this membrane before (a) and after (b) being subjected to pressure.
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
The rejection of determined solutes by a membrane indicates its cut-off, which is defined as the molecular weight in which 90% of solute is rejected (10). In the experiment of PEG rejection, CA presented a rejection of 15% to PEG 45 kDa and 27%. to PEG 80 kDa. From these data it may only be estimated that the cut-off of this membrane will be higher than 100 kDa. Thus, CA could be indicated for use in processes of UF and MF, since these processes are used to retain particles ranging from 5 to 500 kDa (UF) or higher (MF) (11). A higher rejection to solutes of lower molecular weight could be reached by producing asymmetric membranes, which will be the next step of this work.
As PS content was increased in the blends with CA, it was possible to observe changes in the thermal properties of the membranes, as well as changes in their coefficients of ion diffusion, which may be useful for several applications.
Regarding transport driven by pressure, rejection to PEG of CA membrane increased considerably with the increase in molar weight of PEG from 45 to 80 kDa. These results indicate that CA membrane could be used in a range of applications comprehending the processes of UF or MF.
The authors thank FAPEMIG (CEX140-05/CEX-733/07), CAPES for "Portal Periodicos," PROAP. Meireles and Cerqueira thank CAPES for their master's and doctor scholarships, respectively; Mello thanks CNPq for her DTI's scholarchip; Lorenzi thanks CNPq for her scholarship.
(1.) C.F. Liu, R.C. Sun, A.P. Zhang, and J.L. Ren, Carbohydr. Polym., 68, 17 (2006).
(2.) D. Pasquini, M.N. Belgacem, A. Gandini, and A.A.S. Curvelo, J. Colloid Interface Sci., 295, 79 (2006).
(3.) S. Gamez, J.J.G. Cabriales, J.A. Ramiurez, G. Garrote, and M. Vasquez, J. Food Eng., 74, 78 (2006).
(4.) M.A. Khan, S.M. Ashraf, and V.P. Malhotra, Int. J. Adhes. Adhes., 24, 485 (2004).
(5.) C.F. Liu, R.C. Sun, and J. Ye, Polym. Degrad. Stab., 91, 280 (2006).
(6.) G. Rodrigues Filho, R.G.V. Portugues, C.S. Meireles, R.M.N. Assuncao, and J.G. Vieira, Carbohydr. Polym., 67, 182 (2007).
(7.) T.K. Dey and B.M. Misra, J. Polym. Mater., 16, 13 (1999).
(8.) K.C. Khulbe, F. Hamad, C. Feng, T. Matsuura, and M. Khayetb, Desalination, 161, 259 (2004).
(9.) Z. Liao, E. Klein, C.K. Poh, Z. Huang, J. Lu, P.A. Hardy, and D. Gao, J. Membr. Sci., 256, 176 (2005).
(10.) G. Arthanareewaran, P. Thanikaivelan, K. Srinivasn, D. Mohan, and M.S. Rajendran, Eur. Polym, J., 40, 2153 (2004).
(11.) M. Mulder, Basic Principles of Membranes Technology, Kluwer, Netherlands (1997).
(12.) G.R. Filho, S.F. Cruz, D. Pasquini, D.A. Cerqueira, V.S. Prado, and R.M.N. Assuncao, J. Membr. Sci., 177. 255 (2000).
(13.) G. Rodrigues Filho, R.C. Silva, C.S. Meireles, R.M.N. Assuncao, and H. Otaguro, J. Appl. Polym. Sci., 96, 522 (2005).
(14.) G. Rodrigues Filho, C.S. Meireles, R.M.N. Assuncao, K. Mello, and M. Zeni, J. Appl. Polym. Sci., 104, 909 (2007).
(15.) S. Knaus and B. Bauer-Heim, Carbohydr. Polym., 53, 383 (2003).
(16.) G. Rodrigues Filho, R.M.N. Assuncao, B. Royer, J.S. Oliveira, and L.A.C. Motta, J. Appl. Polym. Sci., 96, 1534 (2005).
(17.) D.A. Cerqueira, G. Rodrigues Filho, and R.M.N. Assuncao, Polym. Bull., 56, 475 (2006).
(18.) T. Murata, M.-S. Lee, and A. Tanioka, J. Colloid Interface Sci., 220, 250 (1999).
(19.) A.A. Hanna, A.H. Basta, H. El-Saied, and I.F. Abadir, Polym. Degrad. Stab., 63, 293 (1999).
(20.) G.A. Silva, Compatibilizacao de Blendas de acetato de celulose e poliestireno com viniltriacetoxissilano: Comportamento morfologico e biodegradacao, Instituto de Quimica da UNICAMP, PhD Thesis (2002).
(21.) R.B. Carvalho, C.P. Borges, and R. Nobrega, Polimeros: Ciencia e Tecnologia. 11, 65 (2001).
Correspondence to: Guimes Rodrigues Filho; e-mail: email@example.com
Published online in Wiley InterScience (www.interscience.wiley.com).
[c] 2008 Society of Plastics Engineers
Carla daSilva Meireles, (1) GuimesRodrigues Filho, (1) Rosana MariaNascimento de Assuncao, (1) Daniel Alves Cerqueira, (1) MaraZeni, (2) Katia Mello, (2) SuellenLorenzi (2)
(1) Instituto de Quimica da Universidade Federal de Uberlandia, Av. Joao Naves de Avila 2121, CEP 38400-902, Cx. P. 593, Uberlandia-Minas Gerais, Brasil
(2) Departamento de Fisica e Quimica da Universidade de Caxias do Sul-Caxias do Sul, Rio Grande do Sul, Brasil
|Printer friendly Cite/link Email Feedback|
|Author:||Meireles, Carla da Silva; Filho, Guimes Rodrigues; Assuncao, Rosana Maria Nascimento de; Cerqueira,|
|Publication:||Polymer Engineering and Science|
|Article Type:||Technical report|
|Date:||Aug 1, 2008|
|Next Article:||Effect of intercalating agents on clay dispersion and thermal properties in polyethylene/montmorillonite nanocomposites.|