SciELO - Scientific Electronic Library Online

 
vol.39 número2Synthesis and structural characterization of L-(-)-malic acid pillared layered double hydroxidesAn h-adaptive unstructured mesh refinement strategy for unsteady problems índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados

Articulo

Indicadores

  • No hay articulos citadosCitado por SciELO

Links relacionados

  • En proceso de indezaciónCitado por Google
  • No hay articulos similaresSimilares en SciELO
  • En proceso de indezaciónSimilares en Google

Bookmark


Latin American applied research

versión impresa ISSN 0327-0793

Lat. Am. appl. res. vol.39 no.2 Bahía Blanca abr./jun. 2009

 

ARTICLES

Interpenetrating polymer networks based on castor oil polyurethane/cellulose derivatives and polyacrylic acid

M.A. Gómez Jiménez, J.L. Rivera Armenta, A.M. Mendoza Martínez, J.G. Robledo Muñiz, N.A. Rangel Vazquez and E. Terres Rojas

División de Estudios de Posgrado e Investigación, Instituto Tecnológico de Ciudad Madero, Cd. Madero Tams., 89440, México. jlriveraarmenta@yahoo.com, amendo@yahoo.com, alejandrogomezj@yahoo.com
Laboratorio de Microscopía Electrónica, Instituto Mexicano del Petróleo, México D.F., 07730, México. eterres@imp.com.mx

Abstract - Castor oil, 2,4-tolylene diisocyanate (2,4 TDI), cellulose acetate, and hydroxiethyl cellulose, based polyurethane (PU) - polyacrylic acid (PAA) sequential interpenetrating polymer networks (IPN's) were synthesized. PU's were prepared by reaction of hydroxyl groups from castor oil, and cellulose acetate or hydroxiethyl cellulose with 2,4 TDI, using dibutyl tin dilaurate (DBDTL) as catalyst. PU's were swollen in acrylic acid monomer solution and subsequently polymerized by radical polymerization initiated with 4,4-azobis(4-cyanovaleric acid) (ACVA), and N,N'-methylenebis acrylamide (NNMBA) as crosslinking agent. Series of three PU/PAA ratios (75/25, 50/50, 25/75) were prepared. Viscoelastic properties were studied by means of dynamic mechanical analysis (DMA), showing a maximum modulus for IPN with 25% PAA content. Tan δ curve showed two main peaks, and γ relaxation which is due to lateral chains of PU, suggesting phase separation. The existence of two Tg in each IPN was also confirmed by differential scanning calorimeter DSC). Morphology of samples was observed using a scanning electron microscope (SEM), which revealed different fracture surfaces between the compatible and incompatible PU/PAA ratios studied. Fractures were decreasing to turn into roughness surfaces when the PAA was increasing within the IPN's.

Keywords - Interpenetrating Polymer Networks; Castor Oil; Polyurethane; Polyacrylic Acid.

I. INTRODUCTION

An interpenetrating polymer network is a material containing two (or more) polymer networks, which are combined and at least one of them is synthesized and/or crosslinked in the immediate presence of the other. Crosslinked polymer networks are held together by permanent entanglements with only accidental covalent bonds between the polymers. The three conditions for eligibility of an IPN are: (1) the two polymers are synthesized and/or crosslinked in the presence of the other, (2) the two polymers have similar kinetics, and (3) the two polymers are not dramatically phase separated.

IPN's offer the possibility of combining in network form which otherwise are non-compatible polymers with opposite properties (Pissis et al., 2002; Sperling 1981). As long as the reacting components are mixed well during synthesis, thermodynamic incompatibility can be overcome due to permanent interlocking of network segments and IPN's with limited phase separation are obtained (Hsieh et al., 1999).

Polyurethanes comprise a kind of materials that can vary from rubbery to glassy thermoplastics from linear polymers to thermoset. This versatility can be further controlled in terms of processing and composition to fulfill a vast variety of product requirements (Hsieh et al., 1999). The IPN's formation, be either by "sequential" or "simultaneous" network formation, has often brought about remarkable synergism in properties such as mechanical, thermal, and energy absorption (Parizel et al., 1995).

By in situ sequential IPN's, the elastomeric network, PU, is formed first, and then the rigid crosslinked PAA. The existence of an initial network impedes gross phase separation in the final material. For in situ simultaneous IPN's, the synthesis of both networks is initiated at the same time and proceeds to completion more or less simultaneously, thus bigger domains may form, as there is no topological restriction from an already present structure.

Chemically, castor oil is a natural triglyceride (ester) of fatty acids. Approximately 90% of the fatty acid content is ricinoleic acid, an 18-carbon acid having a double bond in the 9-10 position and a hydroxyl group on the 12th carbon (Klempner and Frisch, 1989).

Cellulose has been re-evaluated as a functional material to meet diverse needs of today's society as a result of the unique reactivities and the nontoxic, biodegradable, biocompatible, hydrophilic, and chiral characteristics (Zhang et al., 1999). Abundant hydroxyl groups in cellulose derivatives facilitate the formation of polyurethane groups with isocyanate groups from 2, 4 TDI. PU based on cellulose derivatives were prepared and tested for biodegradability (Gao et al., 2001, Rivera et al., 2004). Morphology and properties of IPN's depend upon the degree of phase morphology, which is related mainly to polymer miscibility. The goal is to interpenetrate networks when IPN's are made of immiscible components (Vlad et al., 2002). Some studies on the use of castor oil as a component in PU with IPN's have been published (Lu and Zhang, 2002; Lung et al., 2003). Therefore the addition of natural components like castor oil and cellulose derivatives in IPN systems is essential for the successful utilization of renewable resources and improvement of biodegradability.

The present study is concerned with IPN's synthesis made of PU based on castor oil/cellulose derivatives and PAA. IPN's were prepared by sequential method. Microstructure, thermal and mechanical properties of three PU/PAA ratios (75/25, 50/50, 25/75) are reported here. The existence of two glass temperatures (Tg) of the components was found by DSC and DMA. Phase separation which was formed during synthesis was observed by SEM.

II. METHODS

A. Materials

Castor oil with density 0.961g/mL at 25 C,  4-tolylene diisocyanate (2,4-TDI), 95% purity, 74.16g/mol molecular weight and 1.214g/mL density at 25 C, and dibutyltin dilaurate (DBTL), 95% purity, density 1.066g/mL at 25 C were used for the synthesis, all were from Aldrich. Cellulose acetate (CA) Mn ∼30,000, 1.3 g/mL density at 25°C and 2-hydroxyethyl cellulose (2-HEC), Mw ∼250,000, 0.6 g/mL density at 25°C, also from Aldrich. The acrylic acid (AA), purity =99.0% (GC), 1.051 g/mL density at 25 C were also from Aldrich. 4,4-azobis(4-cyanovaleric acid) (ACVA), 98% purity, 280.28g/mol molecular weight was from Fluka. The N,N'-methylene bisacrylamide (NNMBA), 98% purity, 154.17g/mol molecular weight from Aldrich was used as crosslinker.  99.9 % purity acetone from Fermont was used as solvent. All reagents were used as they were received from suppliers.

B. Synthesis

The IPN's were synthesized by sequential method. PU synthesis was carried out adding castor oil, cellulose derivatives and 2,4-TDI to the reactor previously mixed at room temperature. Acrylic acid was heated at 60°C, followed by crosslinnking agent addition. Afterwards, the acrylic acid solution was added to PU-prepolymer and mixed for 20 minutes. Stechiometric amount of initiator to acrylic acid was added at this stage and continuing heating to initiate polymerization. Finally PU catalyst was added to the reaction mix. After 20 minutes the prepolymers were moulded and heated for 24h at 60ºC. IPN's were demoulding and postcurated for the next 24h at 60ºC in a vacuum atmosphere (Hourston et al., 1998, Elmas et al., 2005, Siddaramaiah et al., 1999, Cai et al., 2000).

Experimental parameters are described in Table 1, which describes the isocynate/hydroxyl group relation to PU and three PU/PAA ratios. Table 2 shows crosslinker, initiator levels, and catalyst used to prepare IPN's. The OH number of the castor oil was about 147mgKOH/g.

Tab. 2 Cellulose derivatives [NCO]/[OH], (AC, HEC)

Tab. 3 Initiator, catalyst and crosslinker relations

C. Characterization

Fourier transform infrared (FTIR) spectroscopy was carried out at room conditions using a Perkin Elmer tral range, with 4cm-1 resolution, using an attenuated total reflactance (ATR) accesory under 16 scans. Differential scanning calorimetry (DSC) was carried out on a TA Instrument 2010 calorimeter, under a nitrogen atmosphere and two scans, first at  20ºC/min heat rate, from -100 to 250ºC and a second scan of 5ºC/min from -100 to 250ºC, using the second one for analysis. Dynamic Mechanical Analysis (DMA) was performed on a TA Instrument 2980 dynamic mechanical analyzer under multifrequency mode at a temperature range of -80 to 200°C (heating rate 5°C/min), and a fixed frequency of 1Hz using a dual cantilever clamp. Scanning electron microscopy (SEM), was performed for sample microstructural analysis on an electron microscope Joel JSV 5800 LV model. The voltage of acceleration was of 15 Kw.

III. RESULTS AND DISCUSION

A. Fourier Transformed Infrared (FTIR)

Figure 1 shows IR spectra for IPN including HEC with 3 PU/PAA ratios. Bands corresponding to urethane
(-CH2 a 2926 - 2907cm-1, -CH3 a 2851 - 2841cm-1 y C=O a 1725 - 1701cm-1) can be observed. Incorporation of acrylic acid was confirmed by the presence of a peak at 1691cm-1 attributed to carbonyl of acrylic acid units (also -CH3 2911cm-1) network. No evidence of NCO groups (peak at 2270cm-1) appears, which indicates that all isocyanate groups reacted. Also a small shoulder at 3300cm-1 attributed to -NH stretching of amide II characteristic of PU, decreased as PU amount decreased within the IPN. There is no clear evidence of HEC on IPN, by the fact that signals attributed to this material apear at 1100cm-1 (C-O- C bond, which is overlapped with signals from PU, and OH groups that appear in the same region of NH groups, 3300cm-1). Furthermore OH groups from HEC and CA also are reacting with isocyanate gruops integrating PU network. HEC and CA amounts used in synthesis are too small so that can be a reason there is no clear evidence. Since spectra do not show appearence of additional bands there by ruling out the possibility of any chemical interaction between the component networks (Athawale and Kolekar, 1998).


Fig. 1 FT-IR spectra of IPN's,  HEC 1w% (3w% ACVA, 2.5w%NNMBA)

IR spectra of PU/PAA IPN including 3% CA is showed in Fig. 2. Typical absortion peaks for PU and PAA networks appear. Polyacrylic acid signal is overlapped at the carbonyl region of 1715 - 1727cm-1 for CA and PU. These bands were decreasing when PAA was increasing; which reflected the acrylic acid polymerization (Hourston et al., 1998), and there is no evidence of residual monomer and secondary reactions, which is indicative of complete polymerization and physical crosslinking. Same behavior was observed for this IPN when HEC was present.


Fig. 2. FT-IR spectra of IPN's with 3w% CA (3w% ACVA, 2.5w% NNMBA)

B. Differential Scanning Calorimetry

DSC PU and PAA standards thermograms and PU/PAA IPNs (75/25, 50/50, 25/75) with CA and HEC are shown in figs 3, 4 and 5 respectively. Figure 3 shows endothermic relaxations for PU and PAA standards (around -32°C and 128°C, respectively).


Fig.  3. DSC thermograms of PAA and PU standards


Fig.  4. DSC thermograms of IPN's with 3w% AC (2.5w% ACVA, 2w% NNMBA)


Fig.  5. DSC thermograms of IPN's with 1w% HEC (3w% ACVA, 1.5w% NNMBA)

Thermograms in Figs 4 and 5 show that all IPN exhibited two Tg, which suggests a heterogeneous phase morphology characteristic of a semi IPN, Tg shifting value inward from their respective homopolymer. The lower Tg value corresponds to PU phase and the higher value to PAA phase. This kind of Tg shift suggests that interpenetration occured at some extent. CA was added to the PU of the IPN's, thus this was able to modify the Tg of the PU phase. However, as soon as cellulose derivative was decreasing with PU phase, Tg did not recover due to the relaxed PU network allows more effective interpenetration and leads to close packing of the PAA chains (Hsieh et al., 1999). Tg values attributed to PAA networks, show a slight diminishing value in some IPN, indicating some miscibility with PU network. When CA was including on the IPN composition, there was no significant effect on Tg value. For IPN with a 75/25 ratio there was no Tg around 120°C, attributed to PU as the dominating phase on IPN.

Figure 5 shows that HEC affected lightly to Tg value for PU network on IPNs due to its bigger structure which increased the intermolecular space between PU networks and PAA. Thus a lower heat flow was needed to detect transition relaxations. Also for IPN with PU/PAA ratio 25/75 and 50/50 a relaxation is detected around 50°C, which can be attributed to soft segments of PU and HEC.

Some authors report a compatibility factor, depending on weight ratio of each individual network and Tg value, using DiBenedetto equation (Athawale and Kolekar, 1998), to determinate interactions between networks.

C. Dynamic Mechanical (DMA) Measurements

DMA was used to study the dynamic mechanical behavior of IPNs. Figures 6 and 7 shows storage modulus for PU/PAA with AC and PU/PAA with HEC as function of temperature respectively.


Fig. 6. Storage modulus, 3% wt CA (2.5% ACVA, 2% NNMBA)


Fig.  7. Storage modulus, 3w% HEC (2.5% ACVA, 2.5% NNMBA)

For PU/PAA with 3%(wt) AC, a decrease in modulus occurred at the PU (around -18°C) and PAA (around 115°C) transitions, reflecting a typical behavior of a two-phase system with incomplete phase separation because of the absence of a pronounced lowering of storage modulus with temperature at the Tg region of the individual networks. With PAA increasing content in the  IPN, the rubbery plateau between the PU's Tg and the PAA's Tg increased to a higher modulus showing the change in composition (Widmaier and Drillières, 1997).

The amorphous structure of PU softens with temperature and heat flow, while PAA structure resisted better under test conditions. In all cases, the higher modulus belongs to 25/75 PU/PAA IPN. IPNs with CA had storage modulus higher than IPN's with HEC, providing better viscoelastic properties. CA has acetyl groups which are less bulky than HEC, so IPNs containing acetyl groups had a better rearrangement within their PU structures which would affect the viscoelastic properties. The storage modulus decreased because the PU is the amorphous phase, which has a diminishing capacity for energy absorption when is heated. The rubber behavior starts around -20ºC for three IPNs ratios.

The Tan δ versus temperature curve is often used to determinate damping behavior of materials, and also to identify transitions. A narrow peak indicates a high miscibility degree, whereas two clearly separated peaks in a Tan δ curve transitions with low inter transition Tan d values are indicative of gross phase separation. Tan δ in Figs. 8 and 9 have two main peaks, a narrow peak at 0°C and a broad peak at 140°C attributed to individual PU and PAA networks for all IPNs, and whose magnitudes decreased when concentrations of the components varied. PAA peaks are not as big as the PU, but it is clear enough thus it can be concluded that PU phase is predominant within the IPN. A γ relaxation appeared at 50/50, and 25/75 IPNs ratios at around 60°C. This γ relaxation is attributed to the fact that lateral chains of PU are relaxed and the segmental motions are typical of the glass-rubber transition, also these chains can be associated to CA structure present in the PU network. These results are in good agreement with the results of other authors (Hsieh et al., 1999; Hourston et al., 1998).


Fig.  8. Tan δ, 3w% CA (2.5% ACVA, 2% NNMBA)


Fig. 9. Tan δ, 3w% AC (2.5% ACVA, 2.5% NNMBA)

Tgs for PU phase were higher for DSC than DMA due to the fact that latter is performed under dynamic conditions whereas DSC applies a constant flow heat over the samples. DMA can recognize small transition regions that are beyond the resolution of DSC. Furthermore the position of the glass transition of a polymer can be influenced by adding low molecular weight materials (cellulose derivatives) which reduce intermolecular forces and essentially "lubricate" the macromolecular chains.

D. Scanning Electron Microscopy Micrographs

The morphology of two phases was studied by SEM. Figure 10 shows IPNs with 3% wt content of AC for three PU/PAA relations (75/25, 50/50, and 25/75).  Micrograph 10a shows a surface like beach lines, originated when sample was cutted and characteristic of highly amorphous or elastomeric material which has process memory to cut (Hsieh et al., 2001). The presence of some big domains is indicative of phase separation (Athawale and Raut, 2000). When PU hard segments content is high, exist a finer dispersion in the IPN, and if PAA content increase (10b and 10c) vitreous matrix indicates a better compatibility than those observed for rubbery matrix (Wang et al., 2001), and that is not the case for a phase separation which was proven by DMA and DSC analysis.


Fig. 10. SEM images of IPN's a) 75/25, b) 50/50, c) 25/75 (3w% CA, 2.5w%ACVA, 1.5w% NNMBA).

Figure 11 shows the IPN's of 3%w HEC for three PU/PAA ratios 75/25, 50/50, and 25/75. In Fig. 11 (a) for a 75/25 relation, the sample reveals layers or fractures, and a phase separation between them. The surface shows increased roughness  as PAA contents grew. Figure 11 (b) shows a rough surface when the PAA content was increased too. A brightness loss and a phase separation appeared for this sample. Micrography (c) shows more roughness when the PAA is 75%w of the material. PU was interpenetrated in the PAA network, while the little clear spots on the surface seem to be the PU phase. The morphology of the sample with higher PU content is different from that with higher PAA content (Bartolotta et al., 2002; Sperling 1992; Chiu et al., 2001; Siddaramaiah, 2003).


Fig.  11. SEM images of IPN's a) 75/25, b) 50/50, c) 25/75, (3w% HEC, 2.5w%, 1.5 w%)

IV. CONCLUSIONS

It was possible to prepare IPNs PU based in castor oil CA and HEC/PAA, varying the crosslinker agent concentration, catalyst and PU/PAA ratio in 3 levels. By FTIR was possible to detect characteristic signals of PU, PAA, but was not clear to prove presence of CA and HEC due to overlapping with other groups in the same typical region and because the cellulose derivatives amount was low. Furthermore, there was no evidence of secondary reactions in PU and PAA network formation. By means of  DSC analysis two Tg's were detected for the samples, corresponding to PU and PAA networks, which suggest phase separation in a macroscopic level, characteristic of semi-IPN. Tg values shifted inward its respective homopolymer as PU/PAA ratio was changed. Using this analytical technique was not possible to identify the effect of CA and HEC inclusion on Tg value.

By DMA was possible to confirm the presence of two different phases in the synthesized materials solely based on Tg. Tan δ presented two main peaks for each individual component and γ relaxations from lateral chains from PU. Their storage modulus values were good when the PAA phase was 75% of the IPN. When the PU content was increased, the rigidity of the system decreased.

Trough SEM, the samples showed some characteristics of phase separation, which are proper of IPN's. Therefore when PU or PAA increased in the materials, these presented a heterogeneous surface, due to the control of one of the individual components.

REFERENCES

1. Athawale, V., and S. Kolekar, "Interpenetrating polymer network based on polyol modified castor oil poluyrethane and polymethyl methacrylate," European Polymer Journal, 34, 1447-1451 (1998).         [ Links ]

2. Athawale, V., and S. Raut, "Transparent semi- and full-interpenetrating polymer networks based on uralkyd-polymethyl methacrylate," Physics chem. Chem. Phys, 2, 1249-1254 (2000).         [ Links ]

3. Bartolotta, A., G. Di Marco, M. Lanza, G. Carini, G. D'Angelo, G. Tripodo, A. Fainleib, I. Danilenko and S. Sergeeva, "Mechanical behavior of polycyanurate-polyurethane sequential full-interpene-trating polymer networks," Journal of Non-Crystalline Solids, 307 - 310, 698-704 (2002).         [ Links ]

4. Cai, Y., P. Liu, X. Hu, D. Wang and D. Xu, "Microstructure - tensile relationships of polyurethane /poly(urethane-modified bismaleimide-bismalei-mide) interpenetrating polymer networks," Polymer, 41, 5653 - 5660 (2000).         [ Links ]

5. Chiu, H.C., A.T. Wu and Y.F. Lin, "Synthesis and characterization of acrylic acid-containing dextran hydrogels," Polymer, 42, 1471 - 1479 (2001).         [ Links ]

6. Elmas, B., M. Tuncel, G. Yalcin, S. Senel and A. Tuncel, "Anti C-Reactive Protein Immobilization Onto Poly(glycidyl Methacrylate) Microspheres and its Use for Diagnosis," Colloids and Surfaces A: Physicochem. Eng. Aspects, 269, 125-134 (2005).         [ Links ]

7. Gao, S., "Effect of the synthesis route on the structure and properties of polyurethane/nitrokonjac glucomannan semi-interpenetrating polymer networks," Macromolecules, 34, 2202 - 2207 (2001).         [ Links ]

8. Hourston, D.J., M. Song, F.U. Schafer, H.M. Pollock and A. Hammiche, "Modulated differential scanning calorimetry: 13. analysis of morphology of poly(ethylmethacrylate)/polyurethane interpenetrating polymer networks," Thermochimica Acta, 324, 109-121 (1998).         [ Links ]

9. Hsieh, T., K. Hsieh, G.P. Simon and C. Tiu, "Interpenetrating polymer networks of 2-hidroxyethyl methacrylate terminated polyurethanes and polyurethanes," Polymer, 40, 3153 - 3163 (1999).         [ Links ]

10. Hsieh, K.H., J.L. Han, C.T. Yu and S.C. Fu, "Graft interpenetrating polymer networks of urethane-modified bismaleimide and epoxy (I): mechanical behavior and morphology," Polymer, 42, 2491 - 2500, (2001).         [ Links ]

11. Klempner, D. and K.C. Frisch (Eds.), Advances in interpenetrating polymer networks I," Technomic Publishing Company, Lancaster, Pensilvania (1989).         [ Links ]

12. Lu, Y., and L. Zhang, "Morphology and mechanical properties of semi-interpenetrating polymer networks from polyurethane and benzyl konjac glucomannan," Polymer, 43, 3979 - 3986 (2002).         [ Links ]

13. Lu, Y., L. Zhang, X. Zhang and Y. Zhou, "Effects of secondary structure on miscibility and properties of semi-IPN from polyurethane and benzyl konjac glocomannan," Polymer, 44, 6689 - 6696 (2003).         [ Links ]

14. Parizel, N., G. Meyer and G. Weill, "Morphologies of semi and full interpenetrating polymer networks by nuclear magnetic resonance relaxation times," Polymer, 36, 2323-2330 (1995).         [ Links ]

15. Pissis, P., G. Georgoussis, V.A. Bershtein, E. Neagu and A.A. Fainleb, "Dielectric studies in homogeneous and heterogeneous polyurethane/polycyanurate interpenetrating polymer networks," Journal of Non-Crystalline Solids, 305, 150-158 (2002).         [ Links ]

16. Rivera, J.L., T. Heinze and A.M. Mendoza, "New polyurethane foams modified with cellulose derivatives," European Polymer Journal, 40, 2803-2812 (2004).         [ Links ]

17. Siddaramaiah, T.J., P. Mallu and A. Varadarajulu, "Interpenetrating polymer networks from castor oil-based polyurethane and polystyrene," Polymer Degradation and Stability, 63, 305-309 (1999).         [ Links ]

18. Siddaramaiah, T.J., "Synthesis and characterization of polyaniline filled PU/PMAA interpenetrating polymer networks," European Polymer Journal, 39, 569-578 (2003).         [ Links ]

19. Sperling, L.H., Interpenetrating polymer networks and related materials, Plenum Press (1981).         [ Links ]

20. Sperling, L.H., Introduction to physical polymer science, John Wiley & Sons Inc., 2nd Edition, Bethlehem, Pennsylvania (1992).         [ Links ]

21. Vlad, S., A. Vlad and S. Oprea, "Interpenetrating polymer networks based on polyurethane and polysiloxane," European Polymer Journal, 38, 829-835 (2002).         [ Links ]

22. Wang, S.H., S. Zawadzki and L. Akcelrud, "Morphology and damping behavior of Polyurethane/PMMA simultaneous interpenetrating networks," Materials research, 4, 27-33 (2001).         [ Links ]

23. Widmaier, J.M., and S. Drillières, "Relationships between polymerization activating systems and viscoelastic properties of subsequent polyurethane/poly(tert-butil acrylate) interpenetrating polymer networks," J. Appl. Polym., 63, 951-958 (1997).         [ Links ]

24. Zhang, L., J. Zhou, J. Huang, P. Gong, Q. Zhou, L. Zheng and Y. Du, "Biodegradability of regenerated cellulose films coated with polyurethane/natural polymers interpenetrating polymer networks," Ind. Eng. Chem. Res., 38, 4284 - 4289 (1999).         [ Links ]

Received: February 22, 2008.
Accepted: July 9, 2008.
Recommended by Subject Editor: Ricardo Gómez.