Commercially available kakishibu (an extract naturally fermented from the juice of un ripe persimmon fruit, tannin content, 4.0%) was immobilized as follows. One part of formalin (35% aqueous formaldehyde solution) was mixed with four parts of the above kakishibu solution. After 2 h, a dark red homogeneous gel was obtained. This gel was then crushed into small particles (diameter 150 - 300 μm) and washed thoroughly using deionized water prior to use in the Cr(VI) removal experiments.

An aqueous solution of Cr(VI) in the form of K₂Cr₂O₇ (10 mL, 50 ppm) at pH 2.0 (adjusted using 0.1 M HCl) was passed through a column of persimmon gel (diameter 8 mm; bed volume, 19.3 mL). Desorption was then attempted under both acidic and basic conditions. Under acidic conditions, desorption was carried out by passing a 0.1 M HCl solution (600 mL) followed by a 1 M HCl solution (400 mL) through the column. In contrast under basic conditions, the solutions of 0.1 M Na₂CO₃ (300 mL), 1 M Na₂CO₃ (700 mL), and 0.1 M NaOH (300 mL) were passed through the column. As the persimmon gel was soluble in 1 M NaOH, this solution was not employed during the desorption experiments.

The quantity of Cr(VI) removed by the persimmon gel was then determined using the previously reported diphenyl carbazide method [23] and measuring the absorbance of the solution at 540 nm using a UV-Vis spectrophotometer (U-1800, Hitachi). Finally, the quantities of removed total chromium were determined by atomic absorption spectrometry (AA-6300, Shimadzu, Kyoto, Japan).

A sample of persimmon gel (31.5 mg, dry wt. basis) was added to a solution of potassium dichromate(VI) (100 mL) containing 50.0 ppm Cr(VI) and stirred for 1 h at 30˚C. The gels were then collected by filtration through a membrane filter (pore size 0. 2 μm). For each experiment, the amount of adsorbed Cr(VI) was 86.9% ± 0.2% and the amount of reduced Cr(III) in the solution was 11.8% ± 0.4% based on the original Cr(VI) solution. The recovered gels were then added to a 1 M HCl solution (100 mL) and stirred for 1 h at the desired temperature (i.e., between 30˚C and reflux. Finally, the quantities of removed Cr(VI) and total chromium in the filtrate were quantified using the above-mentioned method.

A sample of the adsorbent (15 mg, dry wt. basis) was suspended in a solution of potassium dichromate(VI) (10 mL) containing 50 ppm (1.0 mM) Cr(VI) at 30˚C for the desired time between 5 min and 24 h. After the appropriate time, the quantities of removed Cr(VI) and total chromium in the filtrate were quantified using the above-mentioned method.

The oxidation state of the chromium species adsorbed on the persimmon gel was then investigated using X-ray absorption fine structure (XAFS) spectroscopy. The X-ray absorption fine spectra were collected at the BL14B2 beamline at the Spring-8 facility in Hyougo, Japan. All measurements were conducted at the Cr K-edge and data were collected in transmission and fluorescence modes for the standard and chromium-adsorbed samples, respectively, between 5660 and 7000 eV. The standard solutions, employed herein were aqueous solutions of Cr(VI) in the form of K₂Cr₂O₇ (2500 ppm) and Cr(III) in the form of Cr(NO₃)₃ (5000 ppm). The chromium-adsorbed gel samples were collected by filtration through a membrane filter (pore size 0.2 μm) as described above for determination of chromium removal from the potassium dichromate(VI) solution by persimmon gel.

A sample of the adsorbent (15 mg, dry wt. basis) was suspended in a solution (pH 2.00) containing the desired Cr(VI) concentration (i.e., 50 - 1000 ppm, 10 mL) for 1 h at 30˚C. The quantities of removed Cr(VI) and total chromium in the filtrate were determined as described previously.

The oxidation state of the chromium adsorbed on the persimmon gel was then determined by X-ray absorption fine structure XAFS. These X-ray absorption fine spectra were collected at the BL9A beam line at the High Energy Accelerator Research Organization (KEK) in Tsukuba, Ibaraki, Japan, and measurements were carried out as described above but with an energy range of 5500 - 7100 eV.

An aqueous solution (1 L) containing 10 ppm Cr(VI) (190 μM) at pH 2.00 was passed through a column (bed volume, 3.8 mL) of persimmon gel (202 mg, dry wt. basis) at a space velocity of 15.8 h⁻¹. The chromium-adsorbed gels sample was then heated in 1 M HCl (100 mL) at reflux for 1 h using a batch system. Following the successful desorption of chromium from the gel, this adsorption-desorption cycles was repeated 8 times (9 cycles in total) to confirm the possibility of recycling of this system. The quantities of Cr(VI) and total chromium in the solution were then determined quantified using the above-mentioned method.

We initially investigated the extent of chromium adsorption/desorption from a Cr(VI) solution using persimmon gel. As previously reported, the amount of chromium removed from solution using persimmon gel was strongly affected by the pH of the solution [21] . More specifically, we found that a pH of 1 - 2 was optimal for our studies, as this pH resulted in the quantitative removal of Cr(VI) from solution, and any further increase in pH resulted in decreased adsorption [21] . In addition, the quantity of total chromium removed was maximised at pH 2 [21] . We therefore examined the adsorption of chromium at pH 2, while the desorption of chromium from the persimmon gel column was examined using both acidic and basic conditions. As indicated in Figure 1 and Figure 2, the majority of Cr(VI) present in the solution was successfully adsorbed on the persimmon gel column (86.8% for a 49.1 ppm Cr(VI) solution and 86.5% for 54.3 ppm Cr(VI) solution), and a percentage of this Cr(VI) was reduced to Cr(III) in solution (i.e., 13.2% and 13.5% respectively). Furthermore, Cr(III) was not adsorbed under these conditions [21] . We also found that relatively small quantities of chromium were desorbed under acidic conditions (16.0%) using 0.1 M

and 1 M HCl solutions (Figure 1), and under basic conditions (12.5%) using 0.1 M, 1 M Na₂CO₃ solutions or 0.1 M NaOH solution (Figure 2) at 30˚C. Moreover, we found that 100% of the desorbed chromium had been reduced to Cr(III).

As described above, the quantity of desorbed chromium was rather low at 30˚C and slightly greater quantities of chromium were desorbed under acidic conditions compared to under basic conditions. Thus, we moved on to examine the effect of temperature on chromium desorption under acidic conditions. As shown in Figure 3(a), the degree of chromium removal increased significantly with increasing temperature, and the majority of adsorbed chromium (i.e., 99.1%) was successfully desorbed under reflux conditions. Furthermore, all adsorbed Cr(VI) was reduced to Cr(III).

The relationship between the reciprocal of temperature and the natural logarithm of the equilibrium constant is shown in Figure 3(b). As indicated from this plot, the desorption of adsorbed chromium from persimmon gel using hydrochloric acid obeys the Van’t Hoff equation,

ln K = − Δ H 0 / R T + Δ S 0 / R (1)

where K is the equilibrium constant (the ratio of desorbed/residual chromium), DH⁰ is the enthalpy change (J∙mol⁻¹), R is the gas constant (8.314 J∙K⁻¹∙mol⁻¹), T is the temperature (K), and DS⁰ is the entropy change (kJ∙K⁻¹∙mol⁻¹) [24] . The enthalpy change was therefore estimated from the slope of the line to be

45.6 kJ∙mol⁻¹, which indicates that the desorption of chromium using dilute hydrochloric acid was an endothermic reaction. Furthermore, the entropy change was estimated to be 0.137 kJ∙K⁻¹∙mol⁻¹. Moreover, as all desorbed chromium was in the form of Cr(III), it was clear that reduction had taken place on the persimmon gel prior to desorption.

The effect of time on both the chromium oxidation state and on Cr(VI) removal from solution using persimmon gel were then examined. As shown in Figure 4(a), the removal of Cr(VI) was rapid with equilibrium being reached within 10 min when a sample of the adsorbent was suspended in the potassium dichromate (VI) solution at 30˚C. Indeed, ~90% of Cr(VI) was absorbed by the persimmon gel, with 10% of that being reduced to Cr(III) in solution.

As Cr(III) is not adsorbed by persimmon gel below pH 2 [21] , it was clear that all chromium adsorption observed herein was in the form of Cr(VI). Therefore, to confirm the variation in chromium oxidation state with time, the adsorbed chromium was analyzed by XAFS. The Cr K X-ray absorption edges were measured for Cr(VI) in a K₂Cr₂O₇ standard and for Cr(III) in a Cr(NO₃)₃ standard, where the presence of a sharp pre-edge peak at 5992 eV corresponds to Cr(VI) and a broad signal at ~6008 eV corresponds to Cr(III). However, as shown in Figure 4(b), no signal at 5992 eV was observed in the XAFS spectrum after10 min,

with the presence of a broad signal at ~6008 eV indicating that all adsorbed Cr(VI) had been reduced to Cr(III) within this time.

The effect of the Cr(VI) concentration on both the removal of chromium and on the oxidation state of chromium in persimmon gel was then examined. As shown in Figure 5(a), the percentage of Cr(VI) removal from the aqueous solution was quantitative at initial Cr(VI) concentrations <400 ppm (7.7 mM). However, above 400 ppm, the degree of Cr(VI) and total chromium removal decreased at higher initial chromium concentrations. Furthermore, the quantity of Cr(VI) and total chromium removed (μmol chromium/g dry wt. adsorbents) increased linearly upon increasing the concentration of chromium in the aqueous solution up to ~400 ppm (i.e. quantitative adsorption up to a maximaum concentration of 400 ppm), after which only very small increases were observed. For example, when the original chromium concentration was 1000 ppm, ~7000 μmol of Cr(VI) and 6000 μmol of total chromium/g dry wt. adsorbents was removed (1,000 μmol of Cr(VI)/g dry wt. adsorbents was reduced to Cr(III) in solution), which correspond to approximately 55% and 45% of the Cr(VI) and total chromium present in the original solution, respectively.

The oxidation states of the adsorbed chromium in this experiment were also analysed by XAFS. As shown in Figure 5(b), no signal was observed at 5992 eV after 1 h when an initial Cr(VI) concentration of 50 ppm was employed, thereby indicating that the adsorbed Cr(VI) had been reduced quantitatively. However, upon increasing the original Cr(VI) concentration above 200 ppm, the intensity of this signal increased. As such, the ratio of Cr(VI) to Cr(III) could be calculated using their respective peak areas, as shown in Table 1. More specifically, ~60% of Cr(VI) was reduced where an initial Cr(VI) concentration of 800 - 1000 ppm had been employed.

Furthermore, desorption of the adsorbed chromium from the persimmon gel was examined at initial Cr(VI) concentration between 200 and 1000 ppm. As shown in Figure 5(c), the majority of adsorbed chromium was successfully desorbed independent of the initial Cr(VI) concentration. In addition, all desorbed chromium was in the form of Cr(III), thus indicating that any remaining unreduced Cr(VI) adsorbed on the persimmon gel was reduced to Cr(III) during the desorption procedure.

To obtain an indication of the recyclability of this system, Cr(VI) adsorption and its subsequent desorption as Cr(III) was repeated over 9 cycles. As shown in Figure 6, using a 1 M HCl solution at reflux, quantitative Cr(VI) adsorption, reduction to Cr(III), and subsequent, Cr(III) desorption were observed. More specifically, after 8 cycles, Cr(VI) adsorption was quantitative, while reduction to

Cr(III) remained at a relatively constant value of ~80%. However, during the 9^th cycle, the Cr(VI) adsorption decreased to 82%, and reduction to Cr(III) decreased to <60%, although Cr(III) desorption was remained quantitative. We expect that this decrease in Cr(VI) adsorption was likely due to oxidation of the polyhydroxyphenyl group in the persimmon gel to give a polyquinone moiety. We could therefore conclude that under the conditions examined herein, persimmon gel could be successfully recycled 8 times for the adsorption/reduction/desorption of chromium species from aqueous solutions. It need to oxidized persimmon gel to reduce initial form for more recycle times.