2.1 Urea adsorption experiments

The effect of process parameters viz., initial urea concentration in urine (3.375, 6.75, 10.125 and 13.5 g L⁻¹), adsorbent loading (0.5–2.5 g), solution temperature (30–40 °C), pH (5.6–7.0) and incubator shaker speed (100–200 rpm) was investigated using an incubator shaker (Orbitek LT, Scigenics Biotech, India). Solutions of different urea concentrations were prepared by diluting the prepared synthetic urine with deionized water; pH was adjusted using HCl (0.1 M) or NaOH (0.1 M). For every experimental module involving the effect of a process parameter, all others were held constant. In each experimental run, 100 mL synthetic urine of known concentration was mixed with a fixed quantity of fly ash; pH of the mixture was adjusted and kept in the incubator shaker at a constant temperature and shaker speed until the establishment of equilibrium. At predetermined intervals of time, 5 mL aliquots were withdrawn, centrifuged (REMI Instruments, AUCI–5230, India) and filtered through a Ministat filter (0.45 μm; Sartorius Stedim Biotech GmbH, Germany). Spectrophotometric determination (Shimadzu UV–1601, Japan) of urea was done by observing change in absorbance (λ_max = 430 nm) (With et al., 1961). The aliquots were returned to the adsorption system after the spectrophotometric determination. pH of the adsorbate was adjusted using HCl or NaOH. The adsorptive urea recovered or urea removed (q_e; mg g⁻¹) was calculated using Eq. (1); here, C₀ and C_t (mg L⁻¹) represent the amount of urea at t = 0 and at any time ‘t’ in the sorbate, V is the sorbate volume (L) while W is the amount (g) of fly ash used in the experimental run.

To validate the experimental findings for synthetic urine solutions, additional experiments were performed using human urine with an average urea concentration of 12.7 g L⁻¹ and pH of 6.1. The sorption system was set up by using 100 mL human urine and 1.5 g ash with the incubator shaker kept at 30 °C and 150 rpm.

2.2 Isotherm, mass transfer and kinetic analysis

To understand the distribution of urea between the sorbate and the adsorbent (fly ash) as well as to gain insights into the adsorptive performance of the sorption system, experimental data were analyzed against several well–known model equations. With the aim of identifying the best isotherm equation, we analyzed two parameter models (Langmuir, Freundlich, Dubinin–Radushkevich, Temkin and Flory–Huggins) in their linearized form and three parameter models (Redlich–Peterson, Sips, Toth and Koble–Corrigan) through non–linear optimization (Supplementary Information, Eq. (s1)–(s9)). In this regard, a review by Foo and Hameed (2010) is particularly helpful in understanding the various model equations, their fundamental characteristics and significance. In Section 2.3, the methodology followed for identifying and selecting the best model equation has been detailed.

Process kinetics was evaluated by using the Lagergen’s pseudo–first–order model (Acemioğlu, 2004), Ho and McKay’s (1999) pseudo–second–order model as well as Weber and Morris’s (1963) intra–particle diffusion model. Mass transfer of urea was investigated according to procedure detailed elsewhere (McKay et al., 1981; Eq. (2)); the dimensionless parameter K was estimated from the Langmuir isotherm model constants, S_s is the surface to volume ratio (cm⁻¹) and represents the area available for urea transfer (assuming fly ash particles to be spherical) and m is the amount of fly ash per unit volume of urea–free adsorbate.

2.3 Statistical analysis

Spectrophotometric measurements of urea absorbance were performed with 10 scans per measurement with a photometric accuracy of ±0.002 Abs, repeatability of ±0.001 Abs and noise level of 0.002 Abs. The standard deviation over the adsorption experiments carried out in triplicate was <5%, and mean values have been quoted in the illustrations. For the two parameter linearized isotherm models, coefficient of determination (R²), Normalized Deviation (ND; Eq. (3)) and Normalized Standard Deviation (NSD; Eq. (4)) were calculated to evaluate the goodness of fit.

For error analysis of the three parameter isotherm models, the procedure detailed by Ho et al. (2002, p.10) was followed. Five different error functions viz., Err_SS – Sum of Squares of Errors, Err_CF – Composite Fractional Error Function, Err_MPSD – Derivative of Marquardt’s Percent Standard Deviation, Err_AR – Average Relative Error and Err_SA – Sum of Absolute Errors were evaluated (Eq. (5)–(9)) with the model parameters determined for each function and a ‘sum of normalized error’ was calculated for each parameter set. The parameter set that yielded the minimum sum of normalized error was considered optimal. Statistical analysis was performed using Microsoft® Excel 2010/XLSTAT© Pro 11 (v. 7.5, Addinsoft Inc., NY, USA).

3.1 Effect of fly ash loading and initial concentration

For a fixed amount of the target compound that needs to be adsorbed, there exists an optimal dosage of the adsorbent that maximizes its uptake capacity at equilibrium. To identify the ideal amount of fly ash to be utilized in the sorption system, a series of experiments were conducted with varied fly ash loading (0.5–2.5 g) (Fig. 1). It was evident that the urea sorption was high (>380 mg g⁻¹) at ash loadings ⩽1.5 g which is possibly due to the availability of sorption sites on the ash surface. When dosage exceeds 1.5 g, the equilibrium adsorption capacity decreases and this is presumably because the number of sorption sites is far more than the amount of urea present in the solution. With increase in ash loading from 0.5 to 2.5 g causes the removal of urea to also increase from 27 to 91%. However, the increase in urea recovered is insignificant (<5%) for ash dosage greater than 1.5 g; hence, 1.5 g ash loading was considered to be the optimal dosage for urea recovery from 100 mL of synthetic urine.

Close

Fig. 1

Urea recovery from synthetic urine investigated at various fly ash dosages and [CO(NH₂)₂] = 13.5 g L⁻¹, 30 °C, 125 rpm and pH = 6.5.

Export to PPT

Using a fixed adsorbent loading of 1.5 g, the initial urea concentration in the synthetic urine was varied by dilution with deionized water to investigate the effect of concentration on the sorption (Fig. 2). The strong influence of concentration is clearly evident with urea uptake capacity of the ash increasing from 176.2 to 409.3 mg g⁻¹ as concentration varies from 25 to 100%. Further, urea recovery was seen to be >80% for all concentrations suggesting that initial urea concentration affects only the quantity of the urea recovered at equilibrium. Furthermore, at fixed adsorbate concentration of 3.375 g L⁻¹ and adsorbent loading of 1.5 g, the effect of shaker speed (100–200 rpm) was observed to be insignificant and hence, an average value of 150 rpm was selected for use in further experiments (Fig. S3).

Close

Fig. 2

The adsorption profile of urea investigated at various initial concentrations of urea with 1.5 g ash loading, 30 °C, 100 rpm and pH = 6.0.

Export to PPT

3.2 Effect of sorption temperature

Experiments were performed at three different incubator temperatures while fly ash loading, initial concentration, shaker speed and pH were held constant at 1.5 g, 13.5 g L⁻¹, 150 rpm and 7.0 respectively (Fig. 3). Results indicated that urea removal from urine over ash could be exothermic in nature as adsorption capacity decreased in the order, 30 °C > 35 °C > 40 °C; urea uptake capacity decreased from 409.3 to 282.4 mg g⁻¹ as temperature increased from 30 to 40 °C. At low temperatures, the average molecular energy is low; hence, the interaction of urea and fly ash is higher which resulted in enhanced adsorption. An observation contrary to this trend has been made in the case of industrial dye sorption where higher temperatures are seen to be favorable (Sarkar and Acharya, 2006); however, the exothermic nature of sorption of organic compounds has been noted by Gupta and Ali (2001) who investigated the removal of pesticides using bagasse fly ash. In any case, small variations in sorption temperature should not significantly affect the adsorptive separation of solute from the solution.

Close

Fig. 3

Effect of sorption temperature on the uptake of urea by coal fly ash investigated at [CO(NH₂)₂] = 13.5 g L⁻¹, 1.5 g ash dosage, 150 rpm and pH = 6.5.

Export to PPT

3.3 Effect of urine pH

The variation in urea uptake capacity of fly ash was also investigated under different synthetic urine pH (Fig. 4). In these experiments the pH was varied between 5.5 and 7.5 considering the variability of pH of human urine from person–to–person as well as different dietary preferences. In any case, the effect of pH was not considerably significant as urea was the dominant molecule in the solution across the range of pH examined. Maximum urea adsorption was observed to occur at a synthetic urine pH of 7. It could be quite possible that urea adsorption onto fly ash is governed by non–electrostatic interactions and thus, to increase its removal from urine, either the pore volume of the ash has to be augmented or the molecular size of the solute has to be increased through agglomeration. Alternatively, increasing ionic strength has also been advocated as a suitable option for improving adsorption capacity at a given pH (Newcombe and Drikas, 1997). Radovic et al. (1997) recorded similar observations in their study involving nitrobenzene adsorption onto activated carbon.

Close

Fig. 4

Influence of urine pH on the urea adsorption capacity of coal fly ash studied at 1.5 g ash loading, [CO(NH₂)₂] = 6.75 g L⁻¹, 30 °C and 150 rpm.

Export to PPT

3.4 Experiments with human urine

The average maximum adsorption capacity of coal fly ash for adsorption of urea from human urine studied in triplicate was found to be 388.3 mg g⁻¹ which agreed well (δ = 5.12) with that observed in experiments with synthetic urine (409.3 mg g⁻¹). In all experiments, ⩾90% removal of urea was recorded. Although not tested in this study, in order to facilitate urea desorption from the fly ash and allow fly ash re-use, the methodology for regeneration suggested by Ganesapillai et al. (2016) can be followed; deionized water of a pre-determined volume based on the fly ash adsorption capacity could allow recovery of urea as an aqueous concentrated solution for subsequent use in agriculture as a crop fertilizer. Alternatively, fly ash adsorbed with urea can be used as such for crop fertilization especially for the amendment and conditioning of acidic soils (Yadav et al., 2016) given the alkaline nature of the fly ash (pH_1:5 ∼ 8).

3.5 Contact time and adsorption kinetics

It is evident from Figs. 1–4 that the sorption was rapid as the time required for establishment of equilibrium was observed to be ∼3600 s. Majority (>95%) of the sorption was seen to occur in t < 70 min across all the experimental conditions investigated. Besides, the adsorption profile revealed three distinct sections: swift urea removal from urine was observed for t < 500 s, gradual removal for 500 s < t < 2500 s and slow uptake at t > 2500 s. Such behavior has been observed elsewhere in studies involving the adsorption of organics (Brás et al., 2005).

Tables 2–4 summarize the results of the kinetic experiments. When experimentally derived sorption data were tested against the three kinetic models it was observed that, using the coefficient of determination as the selection criteria for goodness of fit yielded the pseudo–second–order equation as the best fit (Ho and McKay, 1998). This was a consistent observation (R₂² > 0.99) throughout all the investigated experimental modules – concentration (C; g L⁻¹), temperature (T; °C) as well as fly ash loading (W; g). Likewise, the predicted urea adsorption capacity at equilibrium (q₂) was very close to observed values. In contrast, log (q_e − q_t) versus t plots (not shown) for the pseudo–first–order model indicated the deviation between predicted and experimental data, especially in the initial portion of the adsorption (t < 500 s). The pseudo–second order rate constant (k₂) was found to increase with concentration and ash dosage (⩽1.5 g) and decreased with sorption temperature. An Arrhenius type relationship (R² > 0.9) was seen between k_i (k₁, k₂ or k_id) and temperature as seen through Eq. (10) with the activation energy (E_A) calculated to be <22 kJ mol⁻¹; R (8.314 J mol⁻¹ K⁻¹) is the gas constant, T (K) is the temperature and k₀ is the Arrhenius factor. This is consistent with values reported in literature where Nollet et al. (2003) provided similar observations for fly ash in experiments involving the adsorptive separation of PCBs. The low energy of activation also suggested that diffusion processes featured strongly in determining the rate and that, the sorption was physical in nature (Doğan et al., 2009). Besides, it can also be advocated that Van der Waal’s and electrostatic interactions occur between urea (polar compound) and the surface of fly ash particles. The initial rate of urea adsorption displayed an inverse relationship with temperature; such experimental findings have been reported by others where the suggestive explanation for such behavior is that, adsorbate–adsorbent forces of attraction tend to be weak for exothermic processes (Ofomaja and Unuabonah, 2013).

3.6 Mass transfer and rate determination

Linear plots based on Eq. (2) were obtained and used for the estimation of the mass transfer coefficient using the method of slopes (Fig. S5). As seen in Table 5, the mass transfer analysis again confirmed the significant contribution of fly ash loading and initial solute concentration in determining the extent and rate of urea adsorption. To determine and distinguish the mechanism governing the transfer and/or diffusion of urea molecules, the procedure detailed by Boyd et al. (1947), Reichenberg (1953) and Helfferich (1962) was followed (Eqs. (11) and (12)). F is calculated as the ratio of q_t to q_∞, i.e. the fraction of urea adsorbed (mg g⁻¹) with respect to urea adsorption at infinite time (in this study, 5 h). Bt versus time plots were linear (R² > 0.86) for t < 300 s but did not pass through the origin (Fig. 5). Ignoring the surface adsorption of urea that can be assumed to be spontaneous and thus, non–rate–controlling, it can be suggested here that film diffusion of urea features stronger than its pore diffusion and thus, is rate determining. The estimated coefficient of diffusion values are provided in Table 5 and were found to vary between 7.07 × 10⁻¹³ and 1.77 × 10⁻¹¹ cm² s⁻¹.

Close

Fig. 5

Bt versus t plots for urea adsorption with 1.5 g ash loading, 30 °C, 100 rpm and pH = 6.0 at various initial concentrations: 3.375 g L⁻¹ (■), 6.75 g L⁻¹ (□), 10.125 g L⁻¹ (▲) and 13.5 g L⁻¹ (○).

Export to PPT

3.7 Linear and non–linear isotherm analysis

Isotherm analysis was conducted by testing sorption data from experiments involving the change in concentration (ΔC), ash dosage (ΔW) and temperature (ΔT) against 9 isotherm models. Initially, regression of the data against linearized two parameter equations was performed and results are summarized in Table 6. The favorability of the sorption process investigated was confirmed by the estimated separation factor (0 < R_L < 1) as well as the Freundlich adsorption intensity (n_FR > 2). Thermodynamic parameters were evaluated using the Eyring equation (Doğan et al., 2009) and found to be as follows – ΔH^∗: –75.4 kJ mol⁻¹, ΔS^∗: +0.19 kJ mol⁻¹ and ΔG^∗ varied as −129 to −132 kJ mol⁻¹. In addition, the average free energy (E_S) calculated using the Dubinin–Radushkevich model constant was found to be <8 kJ mol⁻¹ affirming that the process was indeed physical. Utilizing a criterion of high R², low ND and low NSD indicated that, the Flory–Huggins isotherm was seen as an appropriate choice of linearized isotherm model equation. This indicated that that urea adsorption onto coal fly ash was a spontaneous process and that, the removal of urea from urine followed the mechanism of physisorption.

Results of the non–linear regression and error analysis performed as per the guiding notes of Ho et al. (2002, p.10) are presented in Tables 7–9. For every model equation, the parameter set that resulted in the least Err_SN was considered the optimal solution and was used to further evaluate the correlation between experimental and predicted adsorption capacity of fly ash as investigated under various process conditions. For example, as seen in Table 7 for the Toth model, minimum Err_SN was 4.8 suggesting the model parameters Q_t, K_t and t_t were 401, 2.5 and 3, respectively. Considering the overall analysis (four isotherms and three modules) where the optimization objective was to minimize Err_SN, Err_SS and Err_MPSD provided the best parameters sets across eight systems. Moreover, the Redlich–Peterson model, using parameters derived from Err_MPSD and Err_SS, was able to adequately capture the sorption process across all the systems (Table S6). Since the Redlich-Peterson model is derived from a combination of the Freundlich and Langmuir isotherm model relations, the fit of the experimental data is expectedly better than either of them; further, this is also indicative of surface heterogeneity of the fly ash in providing the active sites for urea adsorption.

3.8 System design

The design of a batch adsorption system can be carried out using the results of the isotherm analysis which indicated that as far as the three parameter models are concerned, the relationship between q_e and C_e can be adequately described by the Redlich–Peterson equation (Eq. (13)). Assuming that the fly ash does not adsorb any water from the solute, the adsorbate balance can be expressed as in Eq. (14). Simultaneously solving Eqs. (13) and (14) yields the amount of fly ash required (Eq. (15)) to result in a predetermined extent of reduction/removal of the target compound.

Since a multistage adsorber reduces the amount of adsorbent required to effect the same reduction in the target compound vis–à–vis single stage adsorption, let us consider a two stage adsorber wherein the amount of fly ash required in each stage can be estimated by Eqs. (16) and (17). With the objective of minimization of the total (M₁ + M₂) amount of fly ash required (Eq. (18)), an expression for the intermediate concentration (C₁ in this case) as a function of initial and equilibrium adsorbate concentration can be found (Eq. (19)). By simultaneously considering Eqs. (16), (17), and (19), an expression unique to the investigated system was found for multistage adsorption involving N stages (Eq. (20)).

Under the experimental conditions of pH = 6.0, sorption temperature of 30 °C, shaker speed of 150 rpm and initial urea concentration as 13.5 g L⁻¹, the minimum amount of fly ash required for 90% removal of urea from different volumes of influent synthetic urine solutions was calculated and is enlisted in Table 10. It is evident that increasing the number of stages decreases the total fly ash required while increasing the volume of urine necessitates higher ash dosages (Fig. 6).

Close

Fig. 6

Estimation of amount of fly ash (g) required to treat a given volume of wastewater (synthetic urine in this study) under various removal (%) of urea scenarios.

Export to PPT