Treatment of Wastewater from Thermal Desorption for Remediation of Oil-Contaminated Soil by the Combination of Multiple Processes

Thermal desorption (TD) is one of the methods commonly used to remediate contaminated soil. However, as water is the liquid adsorbent of the off-gas treatment system in the TD stage, the wastewater generated after multiple cycles in the TD stage has low biodegradability and contains complex organic pollutants. In addition to petroleum hydrocarbon, there are also a lot of ammonia, emulsified oil, phenols, aldehydes, and ketones. In this study, effective removal of contaminants was achieved using a combined process of demulsification and flocculation (DF), ammonia stripping (AS), Fenton oxidation (FO), and reverse osmosis (RO). The combined process was optimized, and the maximum chemical oxygen demand (COD), NH 3 -N, turbidity, and extractable petroleum hydrocarbons (EPH) removal efficiencies reached 93.3%, 79.8%, 97.6%, and 99.9%, respectively. The FO was the key process for the efficient removal of contaminants. Ultraviolet-visible (UV/Vis), excitation-emission matrix (EEM), fluorescence spectroscopy, and gas chromatography-mass spectroscopy (GC-MS) showed that refractory macromolecular organic pollutants in water were removed, especially aromatics, phenols, and conjugated aldehydes or conjugated ketones, and further ring cleavage of benzene rings and carbocycles with carbon double bonds was observed. The cost-benefit analysis of the combined process was also carried out. The operating cost was 8.73 US$/m 3 , indicating that the combined process involved moderate costs for recalcitrant wastewater treatment. No studies have been published on combined processes for the treatment of wastewater from TD for the remediation of oil-contaminated soils. Therefore, this study could provide fundamental information based on experimental results and guidelines for wastewater treatment in engineering applications.


Introduction
In the process of oil eld exploitation, a particularly serious ecological problem is soil pollution by crude oil. ere are large amounts of oil-contaminated soils in major oil elds worldwide, such as those in the Middle East, North America, western Africa, and East Asia [1]. A bulletin issued by the Chinese government showed that the main pollutants in oil-contaminated soils are volatile and semivolatile pollutants represented by petroleum hydrocarbons and polycyclic aromatic hydrocarbons [2]. Because thermal desorption (TD) is suitable for the treatment of most volatile and semi-volatile contaminants, it has been widely used as a common remediation method for oil-contaminated soils.
is method exhibits several advantages, such as the capability to treat di erent types of contaminants, short treatment period, high e ciency, high safety, lack of secondary pollution, and recycling of soil and contaminants [3]. However, contaminants in soil are volatilized during the TD stage and are transported to the o -gas treatment system by carrier gas. e main contaminants in o -gas include gaseous contaminants (such as VOCs and SVOCs) and pyrolysis products of soil organic matter. ese contaminants could go into the liquid phase at the condenser or/and condensing scrubber, and wastewater from thermal desorption for remediation of oilcontaminated soil, i.e., the thermal desorption wastewater (TDW), is produced.
After oil removal by an oil/water separator, TDW still poses a serious threat to the ecological environment and human health. Due to the continuous circulation of the liquid absorbent produced by the spray in the TD system, TDW from the remediation of oil-contaminated soil has a complex mixture of pollutants, such as petroleum hydrocarbons and surfactants, which causes a high degree of toxicity and emulsification. e biodegradability of TDW is very poor, and it is difficult to be treated. erefore, the treatment of site-specific TDW remains a major challenge.
Many unit processes for the treatment of oil-containing wastewater have been reported in previous studies, including the following unit treatment processes: pretreatment, secondary treatment, and advanced treatment [4]. e purpose of pretreatment is to remove suspended solids and oil from water to reduce the burden of the subsequent treatment, usually including demulsification-flocculation [5], coagulation-flocculation [6,7], air stripping [8], and dissolved air flotation [9]. However, the effect of pretreatment on the removal of recalcitrant organics from water is limited, and a subsequent secondary and/or advanced treatment is required. erefore, effective treatment requires complete mineralization of these recalcitrant pollutants. is can be achieved using advanced oxidation processes (AOPs), which can effectively oxidize organic pollutants in the secondary treatment [10]. Compared with other AOPs, the Fenton process is the most popular oxidation process because of its exciting advantages such as wide application range, strong anti-interference ability, simple operation and rapid degradation, and mineralization [11]. But Fenton oxidation suffers some limitations that hinder its large-scale applications. In addition to the high cost of treatment, the inability to completely remove micropollutants is also one of the main problems [12,13]. erefore, to reduce the operating cost of the Fenton process and further improve the treatment effect, it needs to be combined with other treatment processes. As a kind of advanced treatment, biological treatment is also widely used as a follow-up process of Fenton treatment [14]. However, it is not suitable for practical thermal desorption engineering because of the long-term acclimation of biological treatment.
e main advantage of membrane technology as an advanced treatment is that it generates stable water without the addition of chemicals, with relatively low energy use, and is an easy and well-arranged process [15]. Membrane technology is widely used in petrochemical, power generation, and steel industries due to its excellent performance, while reverse osmosis (RO) has the best separation performance among all membrane technologies [16]. However, there are some limitations in the operation of RO, such as the need for strict inlet water quality [17]. If membrane technology is used directly for oily wastewater, membrane pollution will seriously reduce the treatment efficiency [18,19]. erefore, proper pretreatment has been stressed repeatedly as the first line of defense in controlling membrane fouling and assuring success for RO operation [20]. e Fenton process can be combined to improve the inlet water quality, reduce membrane pollution, improve the service life of membrane components, and reduce energy consumption and operating costs [21].
During the oil exploitation, some additives such as solvents, surfactants or inorganic salts, and water-soluble and oil-soluble polymers are used, so the wastewater in this study contains phenols, aldehydes, ketones, and ammonia, except for petroleum hydrocarbons. e configured unit process should be established according to the specific conditions and the physicochemical properties of the wastewater. erefore, a combined process of demulsification and flocculation (DF), ammonia stripping (AS), Fenton oxidation (FO), and RO has been developed to remove oil, ammonia, chemical oxygen demand (COD), and small molecular organics. e operating parameters of each process were optimized. e fate and possible removal pathway of the pollutants in wastewater were determined. Also, the engineering applicability was evaluated, and the benefit analysis of the combined process was performed to provide alternative technologies for the effective treatment of wastewater from TD for oil-contaminated soils.

ermal Desorption Wastewater.
e TDW was collected from indirect thermal desorption for remediation of crude oil-contaminated soil in Tahe, Xinjiang, China. e wastewater is the liquid absorbent from the indirect thermal desorption treatment system. After a long recycling operation of water, the water contained high amounts of emulsified oil and was rich in refractory organic matter. e water was light brown, turbid, and had a pungent odor with high concentrations of COD, oil, and NH 3 -N (see Figure S1 for an on-site water sample map). Table 1 shows the main water quality.

Experiments.
DF experiments were performed using a jar test apparatus. Each sample to be coagulated was added with NaClO and placed under a state of rapid stirring (250 rpm). CaCl 2 was slowly added after stirring for 60 s. After adding PAC, the flocculant was stirred slowly for 5 min at a speed of 50 rpm. At the last stage, PAM was added to make the flocculant settle for 30 min. e supernatant obtained after settling was subjected to analysis.
AS experiments were conducted by placing the supernatant after DF in a 500 mL Erlenmeyer flask, and spherical aeration heads were installed inside, through which external air could be supplied. e air required for stripping was controlled at a pressure of <30 Pa using a blower and regulator, and a flow meter was used to control the volume. e change in ammonia concentration in wastewater was observed by changing pH, temperature, time, and aeration rate.
FO experiments were conducted on the AS effluent and putting it into a wide mouth bottle. A fixed volume of 250 mL of wastewater was fed into the reactor and the desired amount of FeSO 4 •7H 2 O was transferred to the reactor after adjustment to the corresponding pH with H 2 SO 4 (2 mol/L). To initiate the reaction, H 2 O 2 was introduced under constant stirring at 150 rpm (to homogenize the mixture). ereafter, the reaction was terminated by spiking the sample with NaOH (2 mol/L), which adjusted the pH of the solution to above 10.0. is resulted in decomposing residual H 2 O 2 and precipitation of the iron as Fe(OH) 3 which was then filtered using a 0.45 μm filter. Lastly, the supernatant was used for COD, oil concentration, and petroleum hydrocarbon analyses.
RO experiment was performed using FlowMem-0021-HP equipped with Dow SW30-2514 RO membrane (maximum operating pressure: 800 (55) psig (ber); seepage discharge: 150 (0.6) gpd (m 3 /d); pH range suitable: 1.0-12.0). By pouring FO effluent into a bucket, the pressure was slowly raised to 6 MPa. e filtrate passes through the RO membrane in the device and is collected in the discharge bottle for subsequent analysis. e orthogonal experimental design tables are shown in the Supplementary Materials (Tables S1-S3). All experiments were carried out using wastewater from the contaminated site in Tahe, Xinjiang, China, and the process flow is shown in Figure S2.

Analysis. COD was measured via a HACH DRB200
coupled with a DR1010 attachment, and a turbidimeter (SGZ-200A) was used to measure turbidity. NH 3 -N was determined by indophenol blue spectrophotometry [22]. e oil concentration of the permeate solutions was recorded using a UV-vis spectrophotometer (UV2600 A, China) based on an absorbance wavelength of 225 nm. Liquid n-hexane was used as the solvent to extract oil from the filtrate [23].
Gas chromatography-flame ionization detection (GC-FID; Agilent GC-8890) was used to analyze extractable petroleum hydrocarbons (EPH) in water samples. e chromatographic column was an HP-5 column, the heating program was as follows: 60°C for 1 min and heating to 320°C for 5.5 min at a rate of 40°C/min; the injection port temperature was 300°C, and the detection temperature of the device was 305°C.
Ultraviolet-visible (UV-vis) spectra were recorded using a UV-2600i UV-vis spectrophotometer, with Milli-Q water as a blank, and a 1-cm quartz cuvette in the wavelength range of 200∼500 nm. e 3D-EEM spectra were performed over excitation wavelengths between 200 and 450 nm in 10 nm intervals and emission wavelengths ranging from 280 to 550 nm in 10 nm intervals.
Gas chromatography-mass spectroscopy (GC-MS; GC-MS2020NX) was used to identify the compounds. e instrument was equipped with a DB-5MS column, the heating program was as follows: 60°C for 1 min and followed by a ramp at 10°C/min to 280°C, and a ramp at 5°C/min to 300°C, which was held for 4 min; the injection port temperature was 290°C, and the detection temperature of the device was 300°C.

Demulsification: Oil Removal.
e TDW with a high oil content of 13.8% (w/w) was an oil/water (O/W) emulsion, so various demulsifiers were used in this study. NaClO, CaCl 2 , PAC, and PAM were added successively. NaClO can destroy the close adhesion of oil and water and assist the subsequent flocculation. Ca 2+ collides with the oil-water interface film and destroys the stability of the oil-water mixture. At the same time, Ca 2+ adsorbs the tiny negatively charged oil particles in the emulsion and compresses the electric double layer of these particles. Demulsifiers work by surrounding the interfacial film around the droplets, and in doing so change the rheological property of the film by gradually weakening until it collapses leading to the aggregation and coalescence of the dispersed droplets. As the droplets coalesce, their size increases until the eventual separation of the phases [24]. PAC and PAM play the role of coagulators to further remove suspended pollutants from water. After adding NaClO, CaCl 2 , PAC, and PAM in sequence, large flocs precipitated, and the turbidity decreased significantly from 596.0 NTU to 14.3 NTU ( Figure S3). Table 2 shows the settling velocity, turbidity, layered effect, floc size, and COD removal after adding NaClO, CaCl 2 , PAC, and PAM. e orthogonal test results showed that under the conditions of 20 mL/L NaClO, 6 g/L CaCl 2 , 6 g/L PAC, and 3 g/L PAM, the oil removal efficiency was the highest at 65.3%, the removal efficiencies of turbidity and COD were 97.6% and 15.3%, respectively (Table S1). e results showed that DF could efficiently remove the turbidity of wastewater, which greatly reduces the burden of subsequent treatment. e low COD removal efficiency was mainly because the water contained high amounts of refractory organic matter, which cannot be removed. Yang et al. [25] also found that low doses of coagulants and flocculants were insufficient to destroy the stability of all colloidal particles, so it is necessary to add enough demulsifier and flocculants for wastewater with high COD and oil contents. Yao et al. [26] found that after the addition of 5 g/L CaCl 2 , 4 g/L PAC, and 0.02 g/L APAM (anionic polyacrylamide), COD concentration of oily wastewater decreased from 36,270 mg/L to 29,054 mg/L, indicating that DF has a limited COD removal efficiency. In addition, Das and Somasundaran [27] reported that in addition to electrostatics, other forces (such as hydrogen bond bridging) between polymer flocculant and particles were also responsible for the flocculation. erefore, too large amounts of acid or alkali are not conducive to the treatment of oily wastewater. us, in the experiment, to save costs and ensure the removal effect, the wastewater can be kept neutral, without additional adjustment of pH.

Ammonia Nitrogen Removal.
After DF treatment, the PH of effluent is about 8.5-9.0, and the turbidity and oil content of effluent are greatly reduced, but the ammonia nitrogen and COD concentrations are still very high. Considering that AS needs to be in alkaline condition, and AS can not only effectively remove ammonia nitrogen but also has a certain ability to remove COD, which also reduces the burden for subsequent processing. erefore, the wastewater after DF treatment was treated with AS, and the effects of pH, temperature, aeration rate, and treatment time on NH 3 -N removal were verified ( Figure 1). With the increase in the pH, the NH 3 -N removal efficiency gradually increased. When the pH increased from 9.0 to 10.0, the removal efficiency increased sharply, and when the pH was 12.0, the NH 3 -N removal efficiency reached 80%. As is well known that when the pH is alkaline, NH 3 -N exists mainly as free ammonia (NH 3 ). By changing the equilibrium constant of the two nitrogen forms (NH 3 -N and NH 4 + -N) in water, the percentage of NH 3 gas generation is increased, and NH 3 -N removal is achieved by air stripping [28]. e NH 3 -N removal efficiency gradually increased with an increase in the aeration rate. When the aeration rate was higher than 0.5 L/min, the NH 3 -N removal efficiency reached 99.8% (Figure 1). An increase in aeration increases the diffusion driving force of NH 3 -N in water and strengthens the desorption and mass transfer rate of NH 3 -N from the liquid to the gas phase so that the removal efficiency is improved.
With an increase in the stripping temperature, the NH 3 -N removal efficiency gradually increased and was greater than 90% at 55°C (Figure 1). AS is a mass transfer process, and the concentration of gas in the gas phase and liquid phase conforms to Henry's law, so the increase in temperature under certain pressure conditions can increase the diffusion coefficient of NH 3 -N and reduce the solubility of NH 3 -N in water [28,29]. On the other hand, when the temperature was above 55°C, the rising trend of the NH 3 -N removal efficiency gradually leveled off (Figure 1) because at this temperature the diffusion of NH 3 -N in the gas-liquid two phases reaches equilibrium. us, the removal efficiency of NH 3 -N does not increase anymore when the temperature continues to rise, but the energy consumption of the equipment greatly increases.
During the stripping process, gas is continuously discharged over time, which changes the ammonia concentration in the gas phase, so that the actual ammonia concentration is always lower than its equilibrium concentration, and the NH 3 -N in the wastewater can be continuously removed. After 150 min, the NH 3 -N removal efficiency increased slowly and finally reached 98.0% (Figure 1). At a pH of 10.0, a temperature of 70°C, and a treatment time of 70 min, the NH 3 -N concentration decreased from 204 mg/L to 41 mg/L (Table S2).

Fenton Oxidation.
After the processes of DF and AS, although the contents of oil, turbidity, and NH 3 -N in water effectively decreased, COD concentration was still as high as 6160 mg/L, and a subsequent oxidation treatment process was still needed. As an advanced oxidation process, the Fenton reagent is suitable for the treatment of high-concentration organic wastewater that is difficult to biotreatment [30]. e degradation efficiency of organic pollutants during FO depends on operating parameters, such as wastewater pH, treatment time, and the dosage of H 2 O 2 and Fe 2+ . In Figure 2, the COD removal efficiency increased continuously with an increase in the dosage of H 2 O 2 . When the dosage of H 2 O 2 was 1.0 mol/L, the COD removal efficiency reached 66.7%. When the Fe 2+ dosage increased from 7 to 21 mmol/ L, the COD removal efficiency increased from 58.4% to 65.8%, the COD concentration decreased to 2359 mg/L ( Figure 2). However, when the Fe 2+ dosage continued to increase to 35 mmol/L, the COD removal efficiency decreased slightly. According to the reaction mechanism, it is found that if Fe 2+ is added in excess, a large number of active •OH will be generated very quickly with H 2 O 2 , while the reaction of •OH with the substrate is relatively slow so that unconsumed free radicals accumulate and interact with each other. e reaction generates water, causing a part of the •OH to be consumed, reducing the utilization rate of •OH, and excessive dosage of Fe 2+ increases the color of the water [31,32]. Figure 2 shows the relationship between COD removal efficiency and H 2 O 2 , Fe 2+ , pH, and time. e decomposition rate of H 2 O 2 varied with different pH. Under acidic conditions, the decomposition rate of H 2 O 2 becomes slower, which is beneficial to the oxidation reaction. At the same time, the conversion of Fe 2+ to Fe(OH) 3 can be achieved by  Journal of Chemistry adjusting the pH, and color and COD removal can be improved by coagulation/flocculation [33]. In Figure 2, the organic compounds in water were oxidized by HO• completely within the first 2 hours, and a further increase in the treatment time has little effect on the COD removal. e optimal COD removal efficiency was achieved at the pH of 3.0, the Fe 2+ dosage of 32.4 mmol/L, the H 2 O 2 dosage of 0.9 mol/L, and the treatment time of 5 h. Moreover, the reaction conditions and Fe 2+ /H 2 O 2 ratio in the experiment were within the appropriate range [34]. e COD decreased from 6160 mg/L after AS to 2250 mg/L (Table S3).

Reverse Osmosis.
Because COD concentration in the effluent after FO was still higher than 2000 mg/L, the discharge standard was not met. According to the Fenton oxidation effluent, the organic pollutants in water were mainly small molecular matters such as ketones, aldehydes, carboxylic acids, and esters ( Figure S5 and Table S5). Because RO provides smaller pore sizes (below 1 nm), it can effectively separate organic liquids (such as alkanes, alcohols, aromatics, and cyclic alkanes) [35]. Al-Huwaidi et al. [36] modeled and simulated the hybrid system of the RO process for the removal of phenol from wastewater. Pozderovic. et al [37] also investigated the influence of processing parameters and membrane types on permeate flux when reverse osmosis concentrates different alcohols, esters, and aldehydes and found that the permeate flux is lower in the concentration of alcohol and ester solutions of the higher molecular mass, a decrease of the permeate flux during the concentration is greater with greater alcohol molecular mass. In this study, FlowMem-0021-HP unit equipped with Dow SW30-2514 RO membrane for RO treatment of FO effluent. e pressure was controlled at 6 MPa, the COD concentration in the effluent was reduced to 150 mg/L, and the COD removal efficiency reached 93.3%, which met the second-class standard of Integrated Wastewater Discharge Standard of China (GB 8978-1996). Figure 3 shows EPH removal in TDW at various process stages. After the combined DF, AS, FO, and RO process, EPH concentration decreased from 818 mg/ L to 0.76 mg/L, and the removal efficiency was up to 99.9%. EPH was mainly removed in the DF and FO processes. e DF process effectively removed EPH, with a removal efficiency of 67.2%, and a reduction from 818 mg/L to 218 mg/L. e remaining EPH was completely removed in the FO process, after which the EPH concentration dropped to 1.24 mg/L, and the removal efficiency reached 99.8%. Previous studies on EPH degradation by means of oxidation have also been reported. Hassan [34] used the FO process to degrade Iraqi Petroleum Refinery Wastewater in experiments and determined that the optimal H 2 O 2 /Fe 2+ molar ratio and treatment time were 15 and 60 min, respectively, the COD removal efficiency was 86%, and the petroleum hydrocarbon concentration decreased from 742 mg/L to 23 mg/L, corresponding to a removal efficiency of 97%. e combination of FO and bioremediation of petroleum hydrocarbon-contaminated soil could achieve a 55% petroleum hydrocarbon removal [38].

Pollutant Degradation Analysis.
First, the raw wastewater and the effluents from the DF, AS, and FO processes were characterized by UV-vis spectrophotometry. UV-vis can quickly obtain the characteristics of organic pollutants in wastewater. Figures 4 and 5 show the UV spectra of the effluents from each process and the calculation results of the corresponding parameters, respectively. In Figure 4, the UVvis spectra of TDW did not show many absorption peaks, which indicates that there were many organic functional groups in the raw wastewater and little interference with each other. e raw wastewater, DF effluent, and AS effluent had obvious absorption peaks at 233 nm, indicating that there were many conjugated aldehydes or conjugated ketones in water; the absorption peak at 283 nm indicated that the water contained many phenols as well [39]. ese two absorption peaks were absent in the effluent of FO. Based on the A 200∼240 and A 240∼400 parameters shown in Figure 5(a), the absorption peak and integral area of the FO effluent were greatly reduced, indicating that the organic matter in the water had been effectively removed in the FO process. In addition, the SUV 254 and A 253/203 parameters of the effluents from each process were calculated ( Figure 5(b)). e SUV 254 value has been widely used as an indicator for evaluating the concentration of aromatics in water, and a SUV 254 value < 0.4 represents a poor degradability of organic matter [40]. e SUV 254 value decreased to below 0.4 after FO, indicating that the concentration of aromatics in water decreased, while the organic matter became more difficult to degrade.
e A 253/203 ratio reflects the concentration of carbonyl, carboxyl, hydroxyl, and ester substituents, and its changing trend was similar to that of SUV 254 . e results from SUV 254 , A 253/203 , A 200∼240 , and A 240∼400 showed that the organics in the raw wastewater were in an aged state, that is, the molecular structure of the organics was complex, containing many aromatic compounds and hydroxyls. Substituents, such as carbonyl, carboxyl, and esters indicate that organics in water have high stability and are difficult to degrade [41,42]. After FO treatment, organic compounds were obviously removed, especially aromatic species, phenols, and conjugate aldehydes or conjugate ketones. Comparing FO and RO, it can be found that the organic matter that cannot be completely degraded in FO is effectively removed in the RO stage ( Figure 4). UV-vis detection has certain limitations in terms of sensitivity and detection limit. By contrast, EEM fluorescence spectroscopy can play an important role in the detection and analysis of organic pollutants because it has multiple advantages, such as providing a highly comprehensive characterization and abundant fluorescence information, high sensitivity, fast detection speed, and good selectivity [43]. In Figure 6, EEM fluorescence analysis of the effluents from each process, P1, P2, P3, and P4 correspond to the fluorescence spectra of the raw wastewater, DF effluent, AS effluent, and FO effluent after 100-fold dilution, respectively. P5 is the fluorescence spectrum of FO effluent without dilution. e fluorescence spectrum in Figure 6 could be divided into five regions, in which regions I and II were located in the range of Ex/Em � (200∼250) nm/ (260∼320) nm and Ex/Em � (200∼250) nm/(320∼380) nm; the ranges of regions III and IV were Ex/Em � (200∼250) nm/(>380) nm and Ex/Em � (250∼450) nm/(260∼380) nm respectively, and region V was located at Ex/Em � (250∼450) nm/(>380) nm [44]. Regions I and II are mainly easily degradable organic substances, mainly aromatic protein substances with a benzene ring structure, regions III and IV belong to the fluorescent region of degradable organic matter, and region V is the fluorescent region of refractory organic matter, for which macromolecular humic substances are the main host [45]. Based on Figure 6, P1, P2, and P3 were similar, the main peak appeared in region IV, and all other peaks appeared in regions I, III, and V, while the main peak in P4 was in region V, and there were no obvious peaks in other regions.
ese results indicate that the raw wastewater, DF effluent, and AS effluent contained complex organic compounds, and after FO, the organic compounds represented by regions I, III, and IV were almost completely removed. e refractory organics represented by region V were also effectively removed. Based on the fluorescence integration of the five regions (Figure 7), the integrated area decreased significantly after FO. Related research showed that the oxidation of Fe 2+ /H 2 O 2 mainly destroys the benzene ring structure and then generates unsaturated fatty acids with π * -π conjugated double bonds [46]. Because the raw wastewater contained substances with benzene ring structure, after being oxidized, Fe 2+ /H 2 O 2 could further decompose unsaturated fatty acids, which was one of the reasons for the more thorough oxidation of organic matter. In addition, there were still fluorescent substances in region V of P4 and P5 (Figure 6), indicating that there were still some refractory organic substances that could not be degraded, and these refractory organic compounds were removed in RO (P6 and P7), which is consistent with the COD degradation results and UV-vis spectra. Figure 8 shows the 3D-EEM of the raw wastewater. e characteristic peaks of phenols are in the vicinity of Ex/ Em � 270/300 nm and Ex/Em � 220/300 nm, where C and D peaks in the region I were located [47], while other monocyclic compounds such as dichlorobenzene and benzene may also contribute to the C and D peaks [43]. e A peak located in region II may be related to bicyclic and tricyclic aromatic hydrocarbons. e fluorescence peaks near Ex/Em � 228/340 nm and Ex/Em � 256/363 nm at which the A peak was located are characteristic peaks of naphthalene and phenanthrene series [48]. In addition, the B peak located in region V generally exists in the fluorescence spectrum of various petroleum substances. Except for saturated hydrocarbons that do not fluoresce in petroleum components, other components involve a luminescence phenomenon, mainly polycyclic aromatic hydrocarbons, heterocyclic compounds, and benzene substances [49]. Peaks A and B located in regions IV and V are both  Journal of Chemistry indicative of refractory macromolecular organics [50], which is also consistent with the UV-vis spectral analysis results. Figure 9 shows the GC-MS total ion chromatogram of the raw wastewater and FO effluent. e peak area of organic matter in Figure 9(b) was greatly reduced compared with that in Figure 9(a), and the types of organic matter also changed significantly. Mass spectrometry analysis of the total ion chromatogram showed that phenolic organic compounds such as phenol, 3-cresol, and 2,3-dimethylphenol may have existed in the raw wastewater, in addition to 2-methyl-2-cyclopenten-1-one, 2,5-hexanedione, 5-methylfuran aldehyde, or 2,3-dimethyl-2cyclopentenone. Degraded conjugated aldehydes or conjugated ketones were also present ( Figure S4), which is consistent with the results of UV-vis spectroscopy and EEM fluorescence spectroscopy. However, there may have still been acetoxy-2-acetone, 2,5-hexanedione, 1,1,3-trichloroacetone, and other ketone organic compounds in the effluent of FO, indicating the presence of conjugated ketones  and cyclic ketones. Although the organic matter was degraded in the FO stage, the generated organic matter could not be completely mineralized by FO ( Figure S5). e remaining organic pollutants present in the raw wastewater and FO effluent are shown in Supplementary Tables S4 and S5, respectively.

Project Application Evaluation and Benefit Analysis.
Based on the data collected in the study and related literature, we compared the cost estimation models of wastewater treatment technology from different authors [51] and found that the cost assessment model of previous studies could not be fully applied to TDW from the actual polluted site of Tahe, Xinjiang, China. For the DF process in this study, the costs mainly include reagents, sludge treatment labor, and maintenance costs, of which the labor cost is estimated to account for about 80%, mainly for sludge treatment [52]. e only disadvantage of the AS process was that it required a large amount of electricity to drive the basic operating equipment, which increase the costs. In the FO process, due to the high COD concentration of wastewater, more hydrogen peroxide in the oxidation stage was consumed, and the cost of chemicals and iron sludge disposal obviously increased. In addition, for the RO membrane process, the costs were mainly composed of four aspects: electricity, chemicals, membrane, and maintenance, of which the electricity costs account for approximately 55% of the total costs, and the maintenance and membrane material costs for 44% [53], 40   while the labor costs were very low in the membrane treatment [54]. A preliminary estimate of the electricity and chemical costs for each TDW treatment process was made, and the operating costs were listed in Table 3 (see Tables S6-S7 for  the main structures and equipment and detailed calculation  procedures). Lee et al. [55] in the experiment used the combined process of coagulation-flocculation⟶dissolved air flotation⟶ozonation⟶ultrafiltration⟶reverse osmosis to treat shale-wastewater with a COD of 1250 mg/L up to the standard, and the operation and maintenance costs were 5.91 US$/m 3 . Similarly, Xu [56] reported multi effect evaporation⟶microelectrolysis⟶Fenton⟶VTBR-Fenton combined process to treat refractory pesticide wastewater, and the COD of the pesticide wastewater reached up to tens of thousands, and the standard treatment costs were 11.4 US$/m 3 . It can thus be seen, the treatment costs for recalcitrant wastewater are generally higher.

Conclusions
To solve the potential environmental threat of wastewater from the indirect thermal desorption process, several process schemes were combined. e characteristics and main pollutants of TDW were analyzed, and the removal properties of the unit processes DF, AS, FO, and RO were investigated.
With the DF process, the removal efficiencies of COD, turbidity, and oil content were 15.3%, 97.6%, and 65.3%, respectively. NH 3 -N concentration decreased from 204 mg/L to 41 mg/L in the AS process. COD concentration further decreased to 2250 mg/L through the FO process. After RO, COD was less than 150 mg/L. With the combined process, EPH concentration decreased from 818 mg/L to 0.76 mg/L, removal efficiency reached 99.9%. In conclusion, COD and NH 3 -N have reached the second-level standard of China's integrated wastewater discharge standard, and petroleum pollutants have reached the first-level standard of China's integrated wastewater discharge standard. e in-depth analysis of the degradation process of pollutants by UV-vis spectroscopy and EEM fluorescence spectroscopy showed that the FO process effectively removed organic macromolecules, especially aromatic compounds, phenols, and conjugated aldehydes or conjugated ketones. It was further proven by GC-MS that oxidation had an obvious removal effect on phenol and other phenolic compounds and an obvious ring-opening effect on cycloketone and cycloenone.
e engineering applicability and economic analysis of the combined DF⟶AS⟶FO⟶RO process were performed for TDW. e operation and maintenance costs of the combined process were found to be 8.73 US$/m 3 , which was higher than the general sewage treatment costs but moderate for recalcitrant wastewater.

Data Availability
e data used to support the findings of this study are included within the supplementary information file(s).

Conflicts of Interest
e authors declare that they have no conflicts of interest. Figure S1. On-site pictures in Tahe Oilfield, Xinjiang (in order are oil-contaminated soil, soil thermal desorption device, and condensed wastewater). Figure S2. Experimental setup and working schematic diagram for treating TDW. Figure S3. Effect drawing of wastewater treatment in DF process (from left to right, NaClO, CaCl2, PAC, and PAM are added in order). Figure S4. Mass spectra of characteristic pollutants of TDW. Figure S5. Mass spectra of characteristic pollutants of FO effluent.