Zong-wei Wu,
Xiao-chen Xu*,
Hong-bin Jiang,
Ruo-yu Zhang,
Shuai-nan Song,
Chuan-qi Zhao and
Feng-lin Yang*
Key Laboratory of Industrial Ecology and Environmental Engineering, China Ministry of Education, School of Environmental Science and Technology, Dalian University of Technology, Linggong Road 2, Dalian 116024, P. R. China. E-mail: xxcep@dlut.edu.cn
First published on 12th September 2017
An integrated process of catalytic ozonation–persulfate oxidation was developed for the pretreatment of dry-spun acrylic fiber (DAF) wastewater on a pilot scale. Box–Behnken design and response surface methodology (RSM) were used for the design and optimization of the integrated process. A second-order polynomial regression equation was established to describe the chemical oxygen demand (COD) and total nitrogen (TN) removal efficiency of the integrated process and was validated by the analysis of variance and residual techniques. The interaction effects of operational parameters were investigated using response surface analysis. Results showed that the maximum COD and TN removal efficiency of 42.36% and 28.51% were achieved for DAF wastewater when the reaction time, reaction temperature, addition of ozone and addition of persulfate were 4.44 h, 61.82 °C, 40 g h−1 and 1.3 kg t−1, respectively. Furthermore, the biodegradability of raw and treated DAF wastewater was compared, and the (BOD5)/COD (B/C) ratio increased from 0.078 to 0.315, indicating a significant biodegradability improvement. The comparison of N composition before and after the integrated process indicated that organonitrogen was converted into ammonia nitrogen, nitrate nitrogen, and nitrogen, which resulted in the removal TN. This result confirmed that the integrated process of catalytic ozonation–persulfate oxidation was effective in improving the biodegradability of DAF wastewater and was promising for pretreatment prior to biological treatment.
Presently, most DAF manufacturing factories have adopted traditional biological treatment or biological treatment followed by physicochemical treatment such as coagulation, neutralization, and adsorption. However, the high concentrations of polymers, inorganic salts, and other toxic compounds in DAF wastewater have strongly negative effects on these biochemical and physical procedures. The effluents after treatment, which still contain high concentrations of COD, NH4+–N, total nitrogen (TN) and toxic compounds, could not meet effluent discharge standards.5 Therefore, it is increasingly urgent to develop an effective and feasible pretreatment for DAF wastewater prior to biological treatment.
Approaches such as Fenton process,6 photo-Fenton process,7 electro-Fenton process,8 microelectrolysis,9 coagulation–flocculation, electro-coagulation,10 and even microbubble-ozonation11 have been used to decrease the amount of toxic contaminants in DAF wastewater, and to improve the biodegradability of the contaminants in subsequent treatment processes. However, there is a serious application bottle-neck due to their higher operating costs.
Catalytic ozonation has recently gained significant attention as an effective process used to improve the biodegradability and reduce the toxicity of refractory wastewater.12,13 This technique has overcome some of the limiting factors of the traditional ozonation process such as low ozone dissolution and low efficiency of ozone. The generation of hydroxyl free radicals (˙OH) from ozone can improve the oxidation effects because ˙OH possesses a higher oxidation potential (2.80 V oxidation–reduction potential, second only to fluorine) than molecular ozone. In catalytic ozonation process, refractory or toxic organic compounds can be oxidized and transformed into biodegradable small molecules or carbon dioxide and water.
As a newest oxidant, persulfate, which has greater efficiency in degrading refractory organic pollutants and fewer adverse impact on water quality than conventional chemical oxidation processes, has gained increasing attentions in recent years.14–16 Stable at room temperature, persulfate can be activated via heat, transition metals, ultraviolet (UV) light, or other means, forming the highly reactive sulfate radical (˙SO4−).17,18 The sulfate radical exhibits a higher standard redox potential ( = 2.5–3.1 V vs. SHE) than the hydroxyl radical at neutral pH, which makes the sulfate radical superior in mineralizing numerous organic compounds.19 The sulfate radical is also more efficient than the hydroxyl radical for degradation of many organic compounds, since it is more selective for oxidation by electron-transfer reaction. Furthermore, the sulfate radical exhibits high oxidation efficiencies in both carbonate and phosphate buffer solutions.20 Recent studies have demonstrated the ability of surface radicals for simultaneous removal of refractory organic contaminants and ammonia in landfill leachate.21 However, there are few studies on the treatment of DAF wastewater using this process.
Based on this, in this study, an integrated process of catalytic ozonation–persulfate oxidation was investigated for DAF wastewater pretreatment in pilot scale. The reactor parameters, including reaction time, reaction temperature, the addition of ozone and the addition of persulfate, were optimized by Box–Behnken design and response surface methodology (RSM), and then the removal rates of the contaminants were investigated under the optimized treatment conditions. The biodegradability the biodegradability of raw and treated DAF wastewater were also evaluated. Moreover, the composition of N before and after the integrated process was also analyzed to interpret the mechanism of TN removal.
Parameter | COD | BOD5 | TN | NH4+–N | Organonitrogen | pH | Temperature |
Unit | mg L−1 | mg L−1 | mg L−1 | mg L−1 | mg L−1 | °C | |
Range | 1679–2377 | 152–176 | 417–480 | 33.0–76.8 | 380–470 | 6.0–8.0 | 60–70 |
Average | 2056 | 162 | 452 | 57.8 | 394 | 6.9 | 67 |
![]() | (1) |
Variables | Coded | Ranges and levels | ||
---|---|---|---|---|
−1 | 0 | 1 | ||
Reaction time (h) | X1 | 4 | 5 | 6 |
Reaction temperature (°C) | X2 | 50 | 60 | 70 |
The addition of ozone (g h−1) | X3 | 20 | 30 | 40 |
The addition of persulfate (kg t−1) | X4 | 0.5 | 0.9 | 1.3 |
Y1 = −26.23203 + 1.78958X1 + 1.10042X2 + 0.23958X3 + 1.68229X4 − 0.025X1X2 + 2.5 × 10−3X1X3 − 0.31250X1X4 − 5 × 10−4X2X3 + 0.06250X2X4 − 0.08750X3X4 + 0.083333X12 − 8.91667 × 10−3X22 + 1.20833 × 10−3X32 + 0.13021X42 | (2) |
Y1 = −105.20156 + 0.83125X1 + 2.79271X2 + 2.00333X3 + 9.63542X4 − 7.5 × 10−3X1X2 − 0.07X1X3 − 1.31250X1X4 − 0.012250X2X3 + 0.081250X2X4 + 0.10X3X4 + 0.37500X12 − 0.019625X22 − 0.013375X32 − 3.28125X42 | (3) |
Statistical testing of the model was performed with the Fisher's statistical test for analysis of variance (ANOVA). The results of ANOVA for COD and TN removal efficiency are depicted in Tables S2 and S3.† The model F value of 160.74 and 198.08 implied that the models were significant for the COD and TN removal efficiency; indeed, there were only a 0.01% chance that these large model F value could occur due to noise. For the removal of COD, the F values for reaction time, reaction temperature, the addition of ozone and the addition of persulfate were 280.93, 95.10, 1557.44 and 117.04, respectively, which implied that the influence of the addition of ozone on the integrated process of catalytic ozonation–persulfate oxidation of DAF manufacturing wastewater was the most significant, followed by reaction time, the addition of persulfate and reaction temperature, orderly. For the removal of TN, the order was the addition of ozone, the addition of persulfate, reaction temperature and reaction time; therefore, the addition of persulfate played an important role in the removal of TN.
The probability value (P value) was equivalent to the proportion of the area under the curve of the F-distribution that lies beyond the observed F value.25 A smaller P value indicates a higher significance of the corresponding coefficient. A P value > 0.10 indicates that the model term was insignificant.26 As shown in the Table S2,† the P value < 0.0001 for the regression model of the removal of COD indicated that the quadratic model was highly significant and adequate for representing the actual relationship between the response and the variables. Furthermore, the individual variables effects (X1, X2, X3 and X4) and interaction terms (X1X2, X2X4, X3X4 and X22) were significant terms (P < 0.10) and had significant effects on the removal efficiency of COD. For the removal of TN (Table S3†), X1, X2, X3, X4, X1X3, X1X4, X2X3, X2X4, X12, X22, X32 and X42 were significant.
The correlation between the experimental data and the predicted values is evaluated by the coefficient of determination (R2). The experimental results and model predictions (eqn (2) and (3)) are shown in Table 3. The obtained R2 value of 0.9938 and 0.9950 indicated that only 0.62% and 0.5% of the total variations were not explained by the model in the removal of COD and TN. This high R2 value suggested that the regression models could well estimate the COD and TN removal response in the studied range. The adjusted determination coefficient (adj R-squared) was 0.9876 and 0.990, showing the high significance of the model. The “Pre R-squared” of 0.9644 and 0.9711 was close to “Adj R-squared”. Adequate precision was a signal to noise ratio and compared the range of the predicted values at the design points to the average prediction error.
Parameter | COD | BOD5 | TN | NH4+–N | Organonitrogen | NO3−–N | The removal of COD | The removal of TN |
Unit | mg L−1 | mg L−1 | mg L−1 | mg L−1 | mg L−1 | mg L−1 | % | % |
Raw water 1 | 1985 | 157 | 454 | 57.6 | 396.4 | — | 42.25 | 28.60 |
Effluent 1 | 1146 | 368 | 324.2 | 68.8 | 152.8 | 29.8 | ||
Raw water 2 | 2164 | 168 | 471 | 61.2 | 409.8 | — | 42.45 | 28.44 |
Effluent 2 | 1245 | 386 | 337 | 73.2 | 154 | 30.3 | ||
Raw water 3 | 2097 | 160 | 461 | 58.5 | 402.5 | — | 42.38 | 28.49 |
Effluent 3 | 1208 | 382 | 329.7 | 69.2 | 153.1 | 29.9 | ||
Average of raw water | 2082 | 162 | 462 | 59.1 | 402.9 | — | 42.36 | 28.51 |
Average of effluent | 1200 | 379 | 330.3 | 70.4 | 153.3 | 30.0 |
Fig. 2a and b are the plot of the predicted values vs. the experimental values of COD and TN removal response. The predicted values calculated using the second-order model were in good agreement with the experimental values, which indicated that there was a satisfactory correlation between the predicted values and the experimental values.
![]() | ||
Fig. 2 Predicted vs. experimental values of COD (a) and TN (b) removal efficiencies of DAF wastewater. |
The reaction residual was another important indicator for evaluating the adequacy of the model in addition to the regression coefficient. The residuals were essentially elements of the variation unexplained by the fitted model and should conform to a normal distribution.27 Fig. S1a and b† are the normal probability plot of the residuals for the removal of COD and TN, which roughly follow a straight line, reveal reasonably well-behaved residuals from the trends. Fig. S1c and d† show the residuals vs. the predicted values for the removal of COD and TN. According to the plots, the residuals appeared to be randomly scattered around zero. Fig. S1e and f† show the residuals in the order of the corresponding observations. Based on the plots, the residuals fluctuated randomly around the center line. All the results above indicated the second-order model could be used to optimize and estimate the removal COD and TN for DAF production wastewater.
With regard to the TN removal, X1X3, X1X4, X2X3 and X2X4 were significant, and the response surface plots and contour plots of them are given in Fig. 4. According to Fig. 4a–d, prolonged reaction time, or increased the addition of ozone or persulfate all could enhance the TN removal efficiency increased under studied range. The probable reason for this behavior was that the generation of ˙OH and ˙SO4− had been promoted by the increase of the addition of ozone or persulfate, which enhanced the oxidation capability of the integrated process, thereby promoting the TN removal efficiency. As shown in Fig. 4e and f, increased reaction temperature led to a rapid increase in the TN removal efficiency, and then slowed nearer the maximum efficiency at approximately 63 °C. Fig. 4g and h show that 60 °C is the optimum reaction temperature for the persulfate oxidation, with lower or higher temperature having a negative effect on TN removal. It may be the reason that persulfate could not be absolutely activated at lower temperature and decomposed at higher temperature.
To further investigate the biodegradability and gain some insight into the degradation process, the GC-MS was also employed to detect the composition of the initial DAF wastewater and after catalytic ozonation–persulfate oxidation treatment. The GC peaks of compounds in the wastewater were shown in Fig. S2 and S3† and it was clearly observed significant differences before and after catalytic ozonation–persulfate oxidation treatment. About 46 compounds in the initial DAF wastewater and 48 compounds in the treatment wastewater had been detected and listed in the Tables S4 and S5.† In order to give more detail about biodegradability and mechanism of TN removal, the compounds in the wastewater were classified into several categories and summarized according to the peak areas fitting. As can be seen from Fig. 5, it was clearly observed that all the categories of the compounds except the ketone ether decreased more than 30% after treatment. The reason may be that ˙OH and ˙SO4− were not selective during the degradation of organic matters,28,29 the nitrile compounds and nitrogen heterocyclic compounds, which were the most toxic and difficult to biodegrade, decreased significantly after catalytic ozonation–persulfate oxidation treatment. On the contrary, ketone ether used as a source of organic carbon increased after treatment. The results could be further confirmed that the integrated process of catalytic ozonation–persulfate oxidation was effective in improving the biodegradability of DAF wastewater and was promising as pretreatment prior to biological treatment.
In this study, the composition of N before and after the integrated process of catalytic ozonation–persulfate oxidation was also investigated to interpret the mechanism of TN removal, which is shown in Tables S2, S3† and Fig. 6. According to Fig. 6a, there is 87.2% organonitrogen and 12.8% NH4+–N in the raw DAW wastewater. Fig. 6b shown that treated by integrated process, the ratio of organonitrogen was a dramatic decline to 46.4%, and the ratio of NH4+–N had a slight increase to 21.3%. Meanwhile nitrate nitrogen and nitrogen appeared and the ratio was 9.1% and 23.2%, respectively. The change of nitrogen existing forms indicated that organonitrogen was converted into ammonia nitrogen, nitrate nitrogen, and nitrogen. The possible degradation pathway of TN in the DAF wastewater was proposed. The nitrogen heterocyclic compounds which were main source of TH, were oxidized firstly to small molecule amino compounds by oxidant such as ozone, ˙OH and ˙SO4−. With following further degradation and mineralization, the amino would be release into the solution, which leaded to the formation of ammonia nitrogen and nitrate nitrogen. In the next step, ozone and ˙OH played a very important role in the transformation of ammonia nitrogen to nitrogen, which resulted in the removal TN. The results closely agreed with the literatures.30–33
2NH4+ + O3 → N2 + 3H2O+ 3O2 + 2H+ | (4) |
NH4+ + 4O3 → NO3− + H2O + 4O2 + 2H+ | (5) |
NH4+ ↔ NH3(aq) + H+ | (6) |
NH3(aq) + 4OH˙ → NO3− + H2O + 5H+ | (7) |
2NH3(aq) + 6OH˙ → N2 + 6H2O | (8) |
![]() | ||
Fig. 6 The composition of N before (a) and after (b) the integrated process of catalytic ozonation–persulfate oxidation. |
(1) The second-order response surface model was valid and adequate for predicting COD and TN removal efficiency of DAF wastewater with four independent variables: reaction time, reaction temperature, the addition of ozone and the addition of persulfate.
(2) ANOVA yielded a high coefficient of determination, ensuring a satisfactory adjustment of the second-order regression model with the experimental data.
(3) The effect of the experimental parameters on the COD and TN removal efficiency was established by the response surface plots. The optimum values of the variables for the maximum COD removal (42.36%) and TN removal (28.51%) developed by the response surface model were found as follows: 4.44 h reaction time, 61.8 °C reaction temperature, 40 g h−1 ozone and 1.3 kg t−1 persulfate.
(4) After the treatment under the optimum conditions, the B/C ratio increased from 0.078 to 0.315, indicating a significant biodegradability improvement.
(5) The comparison of N composition before and after the integrated process confirmed organonitrogen was converted into ammonia nitrogen, nitrate nitrogen, and nitrogen, which resulted in the removal TN.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ra03287k |
This journal is © The Royal Society of Chemistry 2017 |