The samples of SCG in this study were collected from coffee machine at convenience store in Shobara City, Hiroshima, Japan. First, SCG washed using distilled water and dried in the oven for 1 - 2 days at 60˚C. After, it was sifted using test sieve by Tokyo Screen Co., Ltd. with sieving size 425 µm. Then, SCG immersed in 0.04 M calcium hydroxide solution with ratio of 2 g of SCG in 160 mL of calcium hydroxide solution during 24 hours and dried in the oven at 60˚C. The modified SCG washed using distilled water subsequently until pH become 7.5. All chemicals were either reagent on analytical grade and purchased from Kanto Chemical Co. Inc. and Wako pure chemical corporation.

Figure 1 shows SEM images of original and modified SCG. SEM was used to observe the morphology of the adsorbent. The porosity of the modified SCG was relatively more formed than the original SCG. The addition of calcium hydroxide gives advantage to SCG for adsorb more PO 4 3 − and NO 3 − .

FTIR was used to analyze the functional groups on the adsorbent before and after modified. The spectra are shown in Figure 2, both of original and modified SCG have peak between 3600 and 3200 cm⁻¹ is related to the hydroxyl groups of O-H stretching vibration. The hydroxyl groups causes an increase in the

electrostatic adsorption of anions, such as nitrate [10]. The two sharp bands at 2923 and 2852 cm⁻¹ are attributed to C-H stretching vibration. The original SCG has peak between 1700 and 1600 cm⁻¹ is highly associated with chlorogenic acids and caffeine [11]. However, in modified SCG the peak has shifted become 1567 cm⁻¹ due to carboxylate groups [12].

Based on Figure 3 and Figure 4, the figures show that the adsorption capacity of the modified SCG increases than original SCG. The adsorption capacity for PO 4 3 − is 36.74 mg/g and for NO 3 − is 20.21 mg/g, respectively. The pH of the solution is an important variable in the adsorption process because it affects the surface charge of the adsorbent and also the chemical speciation of the adsorbate [13]. The effect of pH is showed in Figure 5, in the case of PO 4 3 − the optimum adsorption occurs at pH 3 and pH 1 - 3 for NO 3 − . The modified SCG surface would be positively charged when the solution pH was lower, which enhanced

the adsorption possibility for the negatively charged PO 4 3 − and NO 3 − .

Figure 6 shows that the increasing mass of adsorbent doesn’t increase the adsorption capacity of PO 4 3 − and NO 3 − . Furthermore, the addition of calcium hydroxide has an important role in the adsorption process of PO 4 3 − and NO 3 − The less amount of adsorbent, the more active sites are formed.

Adsorption isotherm model provide information about the capacity of the adsorbent and the solute-sorbent interaction. In the present work, the Langmuir and Freundlich isotherms were used to analyze the experimental equilibrium data. Table 1 shows the data of (a) Langmuir and (b) Freundlich isotherm models. Figure 7 shows the best fit of adsorption PO 4 3 − and NO 3 − .

The Langmuir model is indicating that there is monolayer coverage of adsorbate on a homogeneous adsorbent surface. The Langmuir isotherm equation is [14]:

C e / q e = 1 / q m K L + C e / q m (4)

whereq_m (mg/g) is the maximum of the adsorption capacity in monolayer condition, q_e(mg/g) is equilibrium capacity, K_L (L/mg) and C_e(mg/L) is equilibrium concentration. The applicable of isotherm equation is compared on the basis of

correlation coefficients (R²). However, R² of Langmuir isotherm model was lower than that of the Freundlich isotherm.

The Freundlich isotherm model is based in assumption that the heterogeneity of the adsorbent material and multi-layer coverage of the adsorbate. The Freundlich isotherm equation is:

log q = log K L + 1 / n log C e (5)

where K_f(L/mg) is the Freundlich isotherm, n is the adsorption intensity, C_e(mg/L) is equilibrium concentration, and q_e (mg/g) is the equilibrium capacity.

The best fit of adsorption PO 4 3 − and NO 3 − were obtained by the Freundlich with R² values was 0.98 for PO 4 3 − and 0.97 for NO 3 − . The data was indicating that adsorption of PO 4 3 − and NO 3 − onto the adsorption are not monolayer adsorption. The addition of calcium hydroxide played important role in adsorption PO 4 3 − and NO 3 − , which is indicating not only physical adsorption but also chemical adsorption.

Figure 8 shows the equilibrium time for PO 4 3 − adsorption is 240 minutes and the optimum time for NO 3 − is 90 minutes. The kinetic model is used to determine the speed of the adsorption process and the stages that control the adsorption process. The kinetic data obtained is in the form of adsorption capacity, which can be obtained through modeling using pseudo-first order and pseudo-second order models. To find out the appropriate kinetic model for the adsorption system of PO 4 3 − and NO 3 − , it is necessary to describe the relationship between the concentration of PO 4 3 − and NO 3 − adsorbed and equilibrium with time. The kinetics of PO 4 3 − and NO 3 − adsorption by modified SCG was investigated by pseudo-first order and pseudo-second order models [15], were determined using the following equations:

ln ( q e − q t ) = ln q t − k 1 t (6)

t / q t = ( 1 / k 2 q e 2 ) + t / q e (7)

where k₁ is the pseudo-first order adsorption rate constant, k₂is the pseudo- second order rate constant, q_e(mg/g) is the adsorption capacity at equilibrium, and q_t(mg/g) is the adsorption capacity at t (minute). Figure 9 shows that both adsorption data is more linear using the pseudo-second order adsorption

kinetic model, evidenced by the R² value of 0.9791 for PO 4 3 − and 0.9347 for NO 3 − .

Adsorption by chemical bonding or ion exchange or combination of both, then desorption can be effected by stronger desorbent like acid or alkali solution. If the adsorption is by physical bonding then the loosely bound metal ion can be easily desorbed with distilled water in most of the cases. The result of PO 4 3 − desorption rate using sodium hydroxide solution showed results reaching 24.35%, while the optimum value for desorption of NO 3 − reached 39.39%. These results indicate that the adsorption on the adsorbent modified SCG is not completely reversible and bonding between it and adsorbed PO 4 3 − and NO 3 − is likely to be be strong. Thus, it is relatively difficult to desorb PO 4 3 − and NO 3 − from the modified SCG [16].

The application of modified SCG after enrichment with PO 4 3 − and NO 3 − for plant growth (Raphanussativus) can be seen in Figure 10 and Figure 11. The dose of 0.3% was a significant growth in the plant height and leaf width. However, seeds can grow on the third day. It is estimated because of the chemical bonds between PO 4 3 − and NO 3 − in the active site of modified SCG are quite strong. Therefore, nutrients need time to release into the soil. It is linear with a desorption

rate of only 26% - 30%.

The original SCG is characterized by a strong phytotoxic impact, such as caffeine content [17]. Caffeine (1.8 mg/g) present in SCG may serve as a chemical defense mechanism in some plants [5]. FTIR data on the original SCG indicated that there was a peak associated with caffeine content. However, the addition of sodium hydroxide has the effect of lowering caffeine levels in SCG. The effect of toxicity can also be seen from the germination rate test. The result of germination rates in Table 2 shows that even though the small dose of fertilizer can contribution to plant growth significantly.