Yang & Volesky

Biosorption and Elution of Uranium with Seaweed Biomass (Presented to Conference IBS'99-Spain)

Jinbai Yang and Bohumil Volesky¹ Department of Chemical Engineering , McGill University, 3610 University Street, Montreal, Canada H3A 2B2

1. Corresponding author, e-mail: boya@chemeng.Lan.mcgill.ca

Abstract
While only weakly radioactive, traces of natural uranium in surface waters represent a danger because of high toxicity. Mine drainage brings enough uranium into the environment as to cause concern. Acid washed, protonated, non-living biomass of Sargassum seaweed sequestered uranyl ions from solution extremely effectively. At pH 4.0, pH 3.2 and pH 2.6, the maximum uranium uptake values were 560 mg/g, 330 mg/g and 150 mg/g, respectively. The uranium biosorption mechanism was affected by the solution pH through the hydrolysis of uranyl ions in aqueous solution. At low pH value, the uranium is present in the solution mainly in the form of free UO₂²⁺ ions, and it was competing with protons for the binding sites on the biomass. The high uranium biosorption at higher pH was attributed to the ion exchange between the hydrolyzed uranyl ions, UO₂OH⁺, (UO₂)₃(OH)⁵⁺, (UO₂)₂(OH)₂²⁺ and protons. Experimental sorption isotherms could be reliably fitted by conventional Langmuir and/or Freundlich models. A flow-through biosorption column demonstrated a high overall column uranium sorption capacity of 105 mgU/g at pH 2.5. About 36 bed volumes of 238 mg/L uranium solution was purified before the breakthrough (at 1 mgU/L in the column effluent). The regeneration of the column by elution with 0.1 N HCl resulted in a very narrow peak of the elution curve reflecting a high efficiency in uranium recovery.

1. Introduction

Uranium is one of the most seriously threatening heavy metals mainly because of its high toxicity, not so much radioactivity. Uranium contamination poses a threat in some surface and groundwaters [1, 2]. Activities associated with the nuclear industry have brought excessive amounts of uranium into the environment [3]. Biosorption has been proposed as a potential alternative for removing toxic heavy metals. Various biosorbents based on non-living biomass have demonstrated an excellent uranium adsorption performance. For example, filamentous fungi, yeast, bacteria, actinomycetes and fresh-water algae, such as Chlorella, have been reported binding uranium in excess of 150 mg/g of dry biomass [4, 5, 6, 7, 8, 9, 10]. Marine algae represent an important biosorbent type. They proliferate ubiquitously and abundantly in the litoral zones of world oceans often posing environmental nuisance. Metal biosorption capacities of non-living seaweed biomass were summarized by Kuyucak and Volesky [11]. The brown alga Sargassum fluitans has been is particularly effective in binding heavy metal ions of gold, cadmium, copper, zinc, etc. [12]. The high sorption capacity, easy regeneration and low-costs make this biomass of special interest for purification of large volumes of wastewater with lower concentration levels of metal toxicity to be removed [12, 13, 14, 15]. This is either difficult and/or expensive to accomplish by conventional metal-removal processes.

Although the fact that marine algae are capable of biologically concentrating radionuclides such as radium, thorium and uranium has been known for a long time [16], the biosorption of uranium by non-living marine algae has not been reported. The present work extended investigations of metal biosorption by the protonated Sargassum alga biomass to the uranium removal and recovery. In this work, basic parameters of equilibrium uranium biosorption were determined and the biosorption-desorption in a flow-through packed column was examined.

2. Materials and Methods

2.1. Preparation of sorbent. Beach-dried Sargassum fluitans, collected in Naples, FL, was chopped up in a homogenizer and sieved to different fraction sizes. The batch of dry biomass with particle size (1.0 - 1.4) mm was selected for subsequent protonation pretreatment aimed at standardizing the biomass by eliminating the light metals Ca²⁺, Mg²⁺ etc. The protonation wash using 0.1 N HCl (10 g biomass / L ) resulted in some biomass weight loss. After 3 hours of contacting with acid, the biomass was rinsed with deionized water in the same volume many times until a stable wash solution pH 4.0 was reached. The biomass was then dried in an oven at 40-60 ^OC overnight. So prepared biomass was stored for later use.

Metal concentration analysis.

Dissolved uranium and lithium concentrations in filtered or centrifuged sample supernatant solutions were assessed simultaneously by an inductively coupled plasma atomic emission spectrophotometer (ICP-AES, Thermo Jarrel Ash, Model TraceScan). The ICP analyses were conducted at wavelengths of 409.014 nm and 367.8 nm for uranium and lithium, respectively. Trial tests confirmed that no sorption of dissolved metals occurred during processing of samples.

Sorption dynamics experiments.

In order to determine the contact time required for the sorption equilibrium experiments, the sorption dynamics experiments were conducted first. 0.1 g of biomass was mixed with 50 mL of 200 mg/L UO₂(NO₃)₂ solution in a magnetically stirred vessel with standard baffles. The pH value of the solution was controlled by a computer-driven autotitrator assembly (PHM82 pH meter, TTT80 Titrator and ABU80 AutoBurette, Radiometer, Copenhagen, Denmark). The autotitrator was set in the end-point titration mode, maintaining the pH value of the reacting solution at the level of the designed end-point. The 0.05 N LiOH solution was added into the titrated system by the internal high-speed pump and a burette. The pH value and the volume of the alkali vs. contact time were recorded by the controlling computer. A series of 0.2 mL samples of solution were removed from the vessel at pre-defined time intervals. After appropriate dilution, the samples were analyzed by the ICP-AES for metal concentrations.

Sorption equilibrium experiments.

With control samples containing no biomass, a series of uranium nitrate solution concentrations (50 mL) were mixed with 0.1 g biomass in 100 mL Erlenmeyer flasks which were shaken on a rotary shaker at 3Hz and room temperature. The pH was adjusted with 0.05 N LiOH or 0.05 N HNO₃ during the sorption process and the volume of LiOH added was recorded. After 3 hours of contact (according to the preliminary sorption dynamics tests), the sorption equilibrium was reached and the solution was filtered or centrifuged. The biomass was then cleaned by soaking and rinsing it with deionized water several times (no biosorbed metal loss occurred) before drying it at 40-60^oC in an oven overnight. The dried metal-loaded biomass was used in desorption experiments later. The supernatant was diluted with D-H₂O for uranium and lithium concentration analyses by the ICP-AES.

The initial uranium concentrations C_icorrespond to the control samples, and the final uranium concentrations C_f were from the supernatant solution. The uranium uptake was calculated by the concentration difference method that is based on the mass balance as follows:
q = (C_i - C_f) V / W (1)

with V being the solution volume, and W being the mass of biosorbent.

In the desorption experiments, 0.1 g of metal-loaded biomass was mixed with 50 mL 0.1 N HCl in a 100 mL Erlenmeyer flask. The remaining procedure was the same as that in the sorption equilibrium experiments except that no pH adjustment was required. The eluted biomass metal content could be calculated directly from the amount of metals desorbed into the HCl solution as follows:

q_des.= C_des.V / W (2)

with q_des. being eluted metal content per gram of biomass, and C_des. being the metal concentration in the HCl eluent solution.

Sorption and desorption column experiments.

The column (Diameter/Length: 3/45 cm) was uniformly packed with 22.64 g (dry basis) of protonated biomass. During the column sorption operation, an aqueous solution containing 238 mg/L (1 mM) uranium at pH 2.5 was pumped upward through the column at a constant flowrate (340 mL/h) continuously. The samples, collected from the outlet of the column by a fraction collector (Gilson 205) at pre-set time intervals, was analyzed for the uranium concentration by the ICP-AES. The pH value of the outlet solution was recorded by the controlling computer. After the biomass in the column became saturated, the column was washed at the same flowrate by D-H2O for several hours, before a subsequent uranium elution with 0.1 N HCl acid. The outlet sample collection and analysis was the same as that used in the biosorption uptake run.

3. Results
3.1. Sorption dynamics and isotherms at different pH values in the batch system. In order to determine the minimum contact time for the equilibrium experiments, the sorption dynamics was examined first. The profiles of dimensionless uranium concentration vs. contact time are plotted in Figure 1.

Figure 1: The uranium concentration decreases with the contact time for different pH values,
                   C₀ = 190.5 mg/L.

    The uranium biosorption rate was strongly influenced by the sorption system pH value, the uranium solution concentration in the solution decreased with contact time faster at higher pH values. At various pH values, approximately 70 - 80% of the uranium present originally in the solution was sorbed onto the biomass in about 15 minutes after the start of biosorption and the equilibrium could be reached within 3 hours. This provided a guide for the biosorption contact time to be used in the following equilibrium experiments.

The equilibrium of the uranium biosorption on Sargassum biomass was expressed in resulting biosorption isotherms for pH 2.6, pH 3.2 and pH 4.0 (Figure 2).

Figure 2: Uranium sorption isotherms experimental data and Langmuir model regression,
(s ) pH 2.6; (u ) pH 3.2; (n ) pH 4.0; (—) Langmuir model; (---) Freundlich model

The lines in Figure 2 refer to model-calculated values and the points are for experimental uranium uptakes. At pH 2.5 and pH 3.2, the isotherms could be represented well by the Langmuir sorption model, q = q_m*C_f /(K + C_f), while the Freundlich model, q = k*(C_f)^n, could represent a better regression than the Langmuir model at pH 4.0.

The model parameters, the maximum uptake capacities q_m and the equilibrium constants K in the Langmuir model as well as k and n in the Freundlich model, were regressed from the experimental data at various pH values and are listed in Table 1:

Table 1: Regressed Langmuir and Freundlich sorption isotherm model parameters

Parameters	pH 2.6	pH 3.2	pH 4.0
K (mmol/L)^-1	0.233	0.084	0.1695
q_m (mmol/g)	0.701	1.215	2.40
k (Freundlich Model)	NA	NA	1.756
n (Freundlich Model)	NA	NA	0.249

Table 1 shows that q_m and K were largely dependent on the final solution pH values. The fact that q_m increased while the K decreased for higher solution pH values indicated that the sorption affinity of uranium for the biomass was enhanced at higher solution pH values. It is worth noting that the q_m value at pH 4.0 was close to the amount of the biomass binding sites, 2.25 mmol/g, as determined from titrations [17]. The control samples demonstrated that the very high uranium uptake could not be attributed to micro-precipitation.

3.2. Desorption of uranium by HCl. The uranium-loaded biomass was eluted by various elutants, NaHCO₃, (NH4)₂SO₄ and by mineral acids. It was established that the diluted mineral acids, such as H₂SO₄, HNO₃ and HCl, were effective in uranium desorption and that no significant biomass damage resulted after several sorption-desorption cycles. The experimental results for the elution of the uranium-loaded biomass with various initial uranium loading by 0.1 N HCl are presented in Figure 3.

Figure 3: Comparison of the initial uranium loading and the uranium elution with 0.1 N HCl

The Y axis in Figure 3 stands for the percentage of the eluted uranium metal over the initial uranium loading on the biomass. It could be noticed that the elution percentage values are close to unity within 4% error range, indicating that the elution with 0.1 N HCl was complete. The biomass weight loss during the acidic desorption process was less than 5%. The biomass was also protonated at the same time and it was thus ready for the next cycle of uranium biosorption uptake.

3.3. Uranium sorption and elution in a packed biomass column. In the biosorption of uranium in the column packed with Sagarsum biomass, approximately 10 L of 238 mg/L uranium solution was processed before the column breakthrough point occurred which was arbitrarily established at 1.0 mg/L of uranium in the column outlet. In this case, the column residence time was approximately 49 minutes and the total amount of 2,380 mg of uranium was accumulated on 22.64 g (dry) biomass. That gives the column overall uranium biosorption capacity of 105 mg U/g (dry biosorbent), including the only partially saturated portion of the dynamic sorption zone still inside at the column breakthrough (outlet 1 mgU/L). The results are illustrated in Figure 4 where the uranium concentration in the column outlet is plotted vs. the number of the column volumes that passed through the column. The column breakthrough took place at 36.5 bed volumes passing through.

Figure 4: Biosorption column breakthrough curve. Feeding uranium concentration C₀ = 238mg/L, flowrate F = 340 mL/h, volume of empty column V_bed = 280 mL, biomass weight W = 22.64 g (dry), pH 2.5.

The elution curve for the column acid wash and recovery of uranium is shown in Figure 5. The uranium concentration in the elution acid (0.1N) at the outlet of the column was plotted against the volume of the elutant passing through the column. The narrow peak, about 6000 mg/L average concentration in 400 mL volume, and the low residual uranium concentration (< 1 ppm) indicated a highly efficient and complete column elution. After one month of continuous sorption-regeneration operation (5 cycles), no significant damage in biomass structures was observed.

Figure 5: Elution of uranium with 0.1 N HCl from the biosorption column. Flowrate F = 340 mL/h, volume of empty column V_bed = 280 mL, biomass weight W = 22.64 g (dry)

4. Discussion

4.1. The effect of pH on uranium biosorption mechanisms and the maximum sorption capacity. For algal biomass, ion exchange has been considered as a main mechanism responsible for metal sequestering [18, 19, 20]. The ion exchange mechanism for uranyl ion binding to the biomass is complicated by the fact that the uranium cation UO₂²⁺ is hydrolyzed in aqueous solutions within the pH range of the sorption system studied here. Partioning of the hydrolysed uranium species depends on the solution pH and on the total uranium concentration in the solution. In the range of acidic to near neutral pH values, four major hydrolysed complex ions, UO₂²⁺, (UO₂)₂(OH)₂²⁺, UO₂OH⁺, (UO₂)₃(OH)₅⁺ and a dissolved solid schoepite exist in the solution [21]. The hydrolysis equilibrium constants are pK = 5.8 for UO₂OH⁺, pK = 5.62 for (UO₂)₂(OH)₂²⁺ and pK = 15.63 for (UO₂)₃(OH)₅⁺[22]. The equilibrium coposition calculations could be carried out by the computer program MINEQL+ [21].

At pH 4.0, all hydrolyzed ions UO₂OH⁺, (UO₂)₃(OH)₅⁺ and (UO₂)₂(OH)₂²⁺existed in the solution within all the present experimental concentration ranges of uranium. The percentage of the free ion UO₂²⁺ decreased while that of (UO₂)₂(OH)₂²⁺ increased with the increase in the total uranium concentration. The two monovalent ions took about 10 - 15 % of the total in all concentration ranges. According to Collins and Stotzky [23], the hydrolyzed species can apparently be sorbed better than the free hydrated ions. Particularly the monovalent ions, compared with the divalent hydrolyzed ions, have even higher affinity to the biomass in ion exchange with protons because they could replace single protons on separate binding sites in the biomass. The binding of the hydrolyzed ions onto the biomass would drive the hydrolysis equilibrium toward the formation of hydrolyzed complex ions when the hydrolyzed proton ions, H₃O⁺s, were neutralized by the added LiOH to the system to maintain the constant solution pH 4.0. Eventually, the uranium would be sorbed on the binding sites in the form of hydrolyzed ions. When the hydrolyzed ions exchanged with protons, the ion exchange stoichiometry would be U/H⁺= 1 : 1 for UO₂OH⁺, U/H⁺= 3 : 1 for (UO₂)₃(OH)₅⁺ and U/H⁺= 2 : 2 for (UO₂)₂(OH)₂²⁺, comparing with a U/H⁺= 1 : 2 exchange ratio for the free ion UO₂²⁺. In another word, the hydrolyzed uranyl ions have a higher binding capacity on the biomass than the free ions. When the hydrolysed ions become prodominant in the ion exchange, the maximum molar uranium uptake could thus become close to or even higher than the value for the total binding capacity in the biomass, i.e. 2.25 meq/g. The extremely high experimental q_m value of 2.4 mmol/g at pH 4.0 (Table 1) may be the demonstration of this case.

With decreasing system pH, the percentage of UO₂²⁺in the solution increased accordingly. The lower pH suppressed the enhancement of uranium biosorption occurring normally because of the hydrolyzed ions. When the pH became low enough, for example at pH 2.6, the divalent free UO₂²⁺became the dominant ion form in the solution for a wide uranium concentration range from 0.3 to 1000 mg/L. Apart from a less preferable ion exchange ratio of U / H⁺ = 1 : 2, some binding sites which are far away from other sites may not be available for the divalent UO₂²⁺ which needs to exchange with two protons. This leads to an even lower uranium binding capacity at low pH.

On the other hand, the non-ionic dissolved solid schoepite starts appearing in the solution when the pH is too high. The uranium sorption may be hindered by the decrease in ion concentration in this situation. For example, Guibal et al. [5] observed a decrease in uranium uptake by filamentous fungus biomass at pH 6.0.

In summary, the biosorption of uranium on Sargassum biomass is a ion exchange process between the uranium ions and protons introduced to the biomass binding sites during the acid pre-treatment and biomass regeneration process. The hydrolysis of uranium ions, which is dependent on the solution pH, increased the uranium uptake by forming monovalent hydrolyzed complex ions. The number of available binding sites in the biomass for hydrolyzed ions was twice or more that for the divalent free UO2²⁺ ions. Correspondingly, at pH 4.0, the maximum uranium uptake was as high as 566 mg/g or 2.38 mmol/g, which is quite close to the total amount of biomass binding sites.

The present work demonstrates that uranium can be very effectively removed from the uranium-containing solution by the continuous-flow biosorption process. The uranium-laden Sargassum-based biosorbent can be conveniently eluted from the sorption column with a small volume of an HCl (0.1 N) wash which concentrates the metal. The high efficiency of biosorption and elution, low biomass damage and stability over a prolonged operation time make the new biosorption process an effective alternative for uranium pollution control which is coupled with the possibility of a feasible uranium metal recovery.

5. Glossary

C₀ Column feeding concentration (mmol/L)

C_i, C_f Initial and final metal concentration (mmol/L)

C_des. Metal concentration in the eluant solution (mmol/L)

F Flowrate passing through column (mL/hr)

q Metal uptake (mmol/g)

q_des. Eluted metal content per gram of biomass (mg/g)

V Solution volume (L)

V_b Column empty bed volume (mL);

W Biomass weight (g)

References

[1] White, S.K. J. Am. Water Works Assoc. 1983, 75, 374.

[2] Laul, J.C. Radioanal. Nucl. Chem. Articles 1992, 156, 235.

[3] Benedict, B.; Pigford, T.H.; Levi, H.W. Nuclear Chemical Engineering, McGraw-Hill
New York, 1981.

[4] Volesky, B.; Tsezos, M. 1981, US Patent 4 320 093; Canadian Patent 1 143 007
(1983).

[5] Guibal, E.; Roulph, C.; Le Cloirec, P. Water Res. 1992, 26, 1139-45.

[6] Macaskie, L.E.; Empson, R.M.; Cheetham, A.K.; Grey, C.P.; Skarnulis, A.J. Science
1992, 257, 782-784.

[7] Munroe, N.D.H.; Bonner, J.D.; Williams, R.; Pattison, K.F.; Norman, J.M.; Faison, B.D.
In Abstracts, American Society for Microbiology Annual Meeting, 1993.

[8] Hu, M.Z.-C.; Norman, J.M.; Faison, N.B.; Reeves, M. Biotechnol. Bioeng. 1996,
51, 237-47.

[9] Horikoshi, T.; Nakajima, A.; Sakaguchi, T. Agric. Biol. Chem. 1979, 332, 617.

[10] Byerley, J.J.; Scharer, J.M.; Charles, A.M. Chem. Eng. Journal 1987, 36, B49-B59.

[11] Kuyucak, N.; Volesky, B. In Biosorption of Heavy Metals ; Volesky, B., ed.;
CRC Press: Boca Raton, FL, 1990, pp 173-198.

[12] Volesky, B.; Holan, Z.R. Biotechnol. Prog. 1995, 11, 235-250.

[13] Kuyucak, N.; Volesky, B. Biorecovery 1989, 1, 189-204.

[14] Leusch, A.; Holan, Z.R.; Volesky, B. J. Chem. Tech. Biotechnol. 1995, 62, 279-288.

[15] Aldor, I.; Fourest, E.; Volesky, B. Can. J. Chem. Eng. 1995, 73, 516-522.

[16] Edgington, D.N.; Gorden, S.A.; Thommes, M.M.; Almodovar, L.R. Limnol. Ocean.
1970, 15, 945-955.

[17] Fourest, E.; Volesky, B. Environ. Sci. Technol. 1996, 30, 277-282.

[18] Crist, R.H.; Oberholser, K.; Schwartz, D.; Marzoff, J.; Ryder, D.; Crist, D.R. Environ.
Sci. Technol. 1988, 22, 755-760.

[19] Crist, D.R.; Crist, R.H.; Martin, J.R.; Watson, J. In Metals-Microorganisms
       Relationships and Applications, FEMS Symposium Abstracts, Metz, France, May;
       Bauda, P., ed.; Societe Francaise de Microbiologie: Paris, France, 1993, pp 13.

[20] Schiewer, S.; Volesky, B. Environ. Sci. Technol. 1995, 29, 3049-3058.

[21] Schecher, W.D. MINEQL+ : A Chemical Equilibrium Program for Personal
       Computers, Users Manual Version 2.22, Environ. Research Software: Hallowell,
       ME, 1991.

[22] Baes, C.F.J.; Mesmer, R.E. The Hydrolysis of Cations, Wiley-Interscience, John
Wiley & Sons: New York, 1976, pp 197-240.

[23] Collins, Y.E.; Stotzky, G. Appl. Environ. Microbiol. 1992, 58, 1592-1600.

[ Bact to the Beginning ] [ Bact to the Index ] [Biosorption Home] [ Dr. Volesky's Home ]

By Jinbai Yang (E-mail: cyya@musica.mcgill.ca )