Fique fibers obtained from the leaves of Furcraea spp., a highly abundant plant in the mountains of South America, may offer an alternative as biosorbents in desalination processes as they exhibit high removal capacities (13.26 meq/g for chloride ions and 15.52 meq/g for sodium ions) up to four times higher than exchange capacities commonly observed in synthetic resins. The ion removal capacity of the fibers was also found to be a function of the pH of the solution with the maximum removal of ions obtained at pH 8. Unlike most commercial ion exchange resins, our results suggest that fique fibers allow simultaneous removal of chloride and sodium ions.
The concentration of sodium ions in seawater is typically around 11 g/L while chloride concentration nears 20 g/L (Benjamin 2002). High salt concentrations cause corrosion in distribution pipes, damage soil and crops, and affect the ecology of water bodies. Therefore, desalination of waste streams with high salt concentration is required prior to their disposal into surface water bodies or their use in irrigation.
Desalination can be achieved using adsorption and ion exchange processes (Miller 2003). During adsorption, surface complexation reactions, including surface hydrolysis and the formation of coordinative bonds with ligands and metals, take place. Several materials capable of removing sodium and chloride by ion exchange or sorption have been recently developed. For example, ZnAl-NO3 layered double hydroxides have uptake capacities of up to 1.3 meq Cl−/g (Lv et al. 2009). The chloride uptake of Amberlite IRA-420, a commercial resin, can reach values of up to 4 meq/g (Carmona et al. 2008). Other commercial synthetic anion exchange resins exhibit typical uptake capacities ranging between 1 and 4 meq/g. Similar uptake capacities are found in commercial cation exchange resins (Zheng et al. 1996; Zhao et al. 2014).
Low-cost sorbents of biological origin, aka biosorbents, can also be used to remove ions and other contaminants from water (Veglio & Beolchini 1997; Bailey et al. 1999; Annadurai et al. 2003; Anirudhan & Radhakrishnan 2007). Anion exchange capacities of about 4 meq/g have been obtained using rice hull, sugarcane bagasse, pure cellulose, and pure alkaline lignin (Orlando et al. 2002). All biosorbents contain cellulose, hemicellulose, and/or lignin (Orlando et al. 2003). Several ion-binding mechanisms are known to occur during biosorption including ion exchange, complexation, chelation, physical adsorption, and microprecipitation among others (Volesky 2001; Gadd 2009).
Fique fibers are obtained from the leaves of Furcraea spp., a plant indigenous to the Andean region of Colombia in South America (Chacón-Patiño et al. 2013). Fique fibers are widely used to manufacture rope and crafts, and they are composed of cellulose (63%), hemicellulose (17.5%), lignin (14.5%), and other minor components including pectin and waxes (5%) (Castellanos et al. 2012). Due to the widespread availability, their low cost, high surface area, and high cellulose content, we decided to explore the feasibility of fique fibers for the removal of sodium and chloride ions from water. Desalination is a novel application for fique fibers.
Fiber preparation and characterization
Raw fibers, obtained from a local market, were cut in pieces ranging between 0.2 and 0.8 mm in length, and soaked in ultrapure water and ethanol (90%) (Gurgel & Gil 2009). The fibers were oven-dried (103 °C, 12 h) and separated using a 850 μm sieve. Surface area, average pore size, and pore volumes of the fibers were determined with an Autosorb-iQ Automated Gas Sorption Analyzer (Quantachrome Instruments). The surface chemistry of the fibers was analyzed using a Fourier transform infrared spectroscope (FTIR) (Shimadzu IR Prestige-21, Japan). Field emission scanning electron microscopy (FESEM) analyses were carried on a FEI QUANTA FEG 650 instrument equipped with a Large Field Detector; samples were coated with a thin layer of carbon before analysis and the micrographs were taken at 10 kV.
Isotherms and sorption kinetics
Solutions of sodium chloride (1 × 10−3 to 0.9 M) were used to determine sorption isotherms. The initial pH of the solutions was adjusted, using nitric acid or potassium hydroxide, to 4, 6, and 8. Approximately 0.5 g of fibers were placed in 50 mL of the solutions and shaken for 2 h at 25 °C. Fibers were separated from the solution prior to measuring sodium and chloride concentrations by filtration (0.45 μm). The concentration of Na+ was measured according to method 3500-Na C (AWWA 1998) via atomic absorption spectrometry (Perkin Elmer Analyst 300 Spectrometer). The concentration of Cl− was measured by using the selective ion method (ME 4500-Cl−) with a 187 pH/IonMeter (Metrohm, Switzerland). Final pH values were measured for all equilibrium experiments. Point of zero charge (PZC) was measured using the pH drift method as previously described (Bouatay et al. 2014). Values for the kinetic constants were obtained by measuring the concentration of dissolved ions at regular intervals during 120 minutes starting at initial concentrations of 0.9 and 0.5 M. All experiments were conducted in triplicate.
RESULTS AND DISCUSSION
The removal of negatively charged ions by lignocellulosic materials can be attributed to the electrostatic interactions between the anions and the protonated carboxyl groups present on the lignin surface (Demirbas 2005; Albadarin et al. 2011). When functional groups such as carboxyl, phenolic, hydroxyl and carbonyl are protonated, the surface is surrounded by hydronium ions, and thus promotes the approach of negatively charged ions.
The sorption capacity (Freundlich constant, Kf) is an order of magnitude higher for sodium than for chloride. This observation supports the conclusion that fique fibers exhibit higher affinity for cations over anions. Higher removal of cations is expected due to the abundance of –OH and C–O–C bonds in the fique fibers as illustrated by the FTIR spectra (Figure 2, right) (Gañán & Mondragon 2002).
The Lagergren equation is commonly used to describe removal kinetics according to a pseudo-first-order kinetic model (Ünlü & Ersoz 2006), where qe is the equilibrium concentration, qt is the amount of solute removed at time t, and Kads is the adsorption constant (Figure 3). The resulting Lagergren fits render: chloride: log (14.53 − qt) = 14.53 − (0.0493/2.303)t, R2 = 0.979; sodium: log(15.82 − qt) = 15.82 − (0.0386/2.303)t, R2 = 0.989) at initial pH of 8 and initial NaCl concentration of 0.9 M. The good quantitative fit of the experimental data to the Lagergren model is an indication that the process does indeed follow pseudo-first-order kinetics. The pseudo-second-order kinetic model was also used to evaluate our experimental data, but R2 values obtained were slightly lower than those obtained with the Lagergren model.
Our results suggest that fique fibers can simultaneously be used for the removal of chloride and sodium ions from water. However, results presented here are only from batch experiments, and further experiments need to be conducted to evaluate the efficiency of the removal process under continuous conditions. Also, it is unknown if the cutting, sieving, washing, and drying of the fibers is required prior to the removal process and it is necessary to conduct further experiments to optimize pretreatments for cost and quality.
Fique fibers were studied as potential biosorbents for the removal of sodium and chloride ions from water. Experimental results indicated that both ions were removed at high rates, reaching saturation within 60 minutes. The removal capacities were significantly higher than most removal capacity values reported in the literature (Orlando et al. 2002; Carmona et al. 2008; Lv et al. 2009), namely up to 15 meq/g for sodium and 13 meq/g for chloride. Fique fibers present an interesting alternative for water desalination, featuring high removal capacity in a material of agricultural origin. Furthermore, unlike most commercial ion exchange resins, our batch test results suggest that fique fibers allow simultaneous removal of chloride and sodium ions.
Funding for this research was provided by Universidad de los Andes and Universidad Libre. We would like to thank Jorge R. Silva-Larrotta for allowing us to use the laboratory facilities at Universidad Libre. This work made use of the Cornell Center for Materials Research Shared Facilities, which are supported through the NSF MRSEC program (DMR-1120296). Juan P. Hinestroza acknowledges financial support from USDA National Institute of Food and Agriculture (Hatch Multistate NC-170).