1. Introduction
The production of crude oil (particularly heavy crude oil) involves the coproduction of process waters. In nondomestic wastewater, the petroleum industry generates a significant load of leachate (WW) with diverse compositions, depending on the geographical location of the exploitation, the crude type, and the method for the extraction of the petroleum, among other factors. The petroleum industry generates large amounts of contaminated WW (
Table 1). The United States has official data for the years 2007 and 2012, and Colombia has official data for the year 2015 [
1,
2,
3].
The WW generated during the production of crude oil usually contains dissolved gases (CO
2, H
2S), salts, suspended solids, radioisotopes, hydrocarbons, and metal ions. In accordance with the guidelines in the Standard Methods for the Examination of Water and Wastewater™ [
4] REF, the parameters measured to analyze WW and treated wastewater (TWW) are pH, salinity, electrical conductivity (EC), total alkalinity (TA), total suspended solids (TSSes), total dissolved solids (TDSes), chlorides, total hardness (TH), chemical oxygen demand (COD), biological oxygen demand (BOD), copper, nickel, and other compounds [
4].
Different regulations in oil-producing countries establish the maximum limits of pollutants in WW that is discharged into surface water bodies (
Table 2). The environment ministries and secretariats of Colombia, Mexico, Brazil, Peru, Venezuela, Asturias (Spain), and China and the World Bank Group (WBG) members’ oil-producing countries were consulted to compare differences and similarities between their maximum limits of effluents.
WW can be discharged into surface water bodies and public sewage if it meets the requirements established in the regulatory norms that are in effect in a country, e.g., Spain [
18], China [
19], Colombia [
20], Mexico [
21], Brazil [
22], Peru [
23], Venezuela [
24], and the World Bank Group (WBG) [
25,
26]. As a consequence, the petroleum industry has tested different strategies to regulate pollutant discharge into water bodies, marine water, and water used in households, industry, agriculture, water sports, and in power generation, among other applications.
The results of a detailed review indicated that there are a variety of methods to remove pollutants from wastewater, including (a) physical methods, such as adsorption, cyclones, enhanced flotation [
16,
27], flocculation [
28], and activated carbons [
29]; (b) chemical methods, such as precipitation, electrochemical techniques, [
16,
30], oxidation [
27,
31], photocatalytic techniques [
32], and demulsifiers [
33]; (c) biological methods, such as bioreactors [
16,
30,
34]; (d) membrane techniques, such as polymeric, ceramic, or inorganic membranes, microfiltration, ultrafiltration, reverse osmosis, membrane distillation, and nanofiltration [
16,
27,
30,
35,
36,
37,
38,
39]; and (e) combined or hybrid methods, such as coagulation–flocculation and flotation, biological treatment with activated carbon and reverse osmosis (RO), and bioelectrochemical reactor and coagulation membrane processes [
30,
35,
40,
41] for reducing the COD, total petroleum hydrocarbon (TPH), and other contaminants present in petroleum WW.
Among physical treatments, the flocculation with zero-valent iron–ethylenediaminetetraacetic acid (EDTA) and air (ZEA) process and granular-activated carbon removed 92% of COD and 97% of TPH and 72.7%–88.2% of TPH, respectively [
28,
42,
43]; and shaking extraction recovered about 60% of TPH from oil [
44]. Of the chemical treatments used, electrocoagulation removed 85.81% of COD [
45], and electrochemical removed 85%–96% of COD [
46,
47].
On the other hand, biological treatment studies have shown that rotating biological contactor (RBC) discs removed 78%–97% of COD and 95%–99% of TPH [
48,
49,
50,
51], and a membrane bioreactor (MBR) removed 96% of COD [
11]. Nevertheless, new technologies have been tested, such as membrane in reverse osmosis (RO) and filtration membranes (which removed 82%–99% of COD [
52,
53,
54,
55]) and a hybrid microfiltration (MF)/ultrafiltration (UF) process (which removed 94.4%–98.8% of COD [
56]).
The oil industry uses fractional distillation (FD) or extraction with solvents, and the emphasis of the technique is on obtaining the optimal benefit for the crude oil in refining or recovery operations. However, the use of distillation or extraction as a decontamination treatment of wastewater still requires further research and data. In this respect, catalytic vacuum distillation has been shown to reduce COD by 99% [
57]. In addition, distillation has been investigated for the treatment of seawater. On the other hand, an extraction technique used to treat pond sludge removed 67.5% [
58] and 40%–60% of COD using the solvents methyl ethyl ketone (MEK) and ethyl acetate (EA) [
44].
To date, there have been no publications on the systematic use of continuous liquid–liquid extraction (CLLE) or FD, either individually or in sequence, as preliminary treatments for the decontamination of WW (including decontamination in sedimentation ponds during production).
This work presents and discusses the results of laboratory-scale assays using CLLE with dichloromethane (CLLEDCM) and high-power fractional distillation (HPFD), individually and in sequence, on authentic WW to produce TWW of better quality than the original water to illustrate the potentiality of these techniques for decontaminating WW. CLLEDCM and HPFD were specifically chosen due to preliminary trials showing that extraction with DCM removed more contaminants from WW and that distillation while heating low- or medium-potency WW prolonged the time of experimentation without appreciable improvement in the yield or quality of the TWW obtained.
3. Results and Discussion
3.1. Guidelines on the Permissible Limits of Pollutant Discharge from WW
The maximum permissible limits of COD and TPH in the norms of China, Peru, WBG members, and Colombia are stricter than those in other countries. In Colombia, the norms regulating the quality of water discharged into water bodies include more analytical parameters than those of Venezuela and Asturias (Spain), and they have less rigorous limits for TSSes, TPH, and COD than do WBG members (
Table 2). Regulations in Brazil and Mexico do not require the evaluation of most of the analytical parameters included in the norms of other countries.
3.2. Physicochemical Analysis of WW
Previous studies [
28,
43,
60,
61] have presented results on the analytical characterization of WW but have not specified the type of operation or the production stage in which the samples were collected.
Table 3 shows data from the physicochemical analysis of the WW sampled from the sedimentation pond of the production area of an oil company in Colombia. The values of the physicochemical parameters in the WW were analyzed and found to exceed the permissible limits for effluent discharge into surface water bodies according to any of the norms of the countries mentioned in this study.
The peaks corresponding to aliphatic isoprenoid hydrocarbons (pristane (n-C17) and phytane (n-C18)) in the GC chromatogram of the WW (
Figure 2A) are indicative of TPH of marine origin [
62] (contaminants in the sample). In the GC chromatogram of WW, the unimodal distribution of peak intensities corresponded mainly to n-alkanes of low molecular weight (Cn, n ≤ 25), which ranged from n-C10 to n-C32, from n-C13 to n-C31, and from n-C17 to n-C31 [
63]. These results were evidence of the presence of aliphatic and aromatic contaminants.
In the WW and GC chromatograms of TWW (
Figure 2A–D), pentacyclic triterpenes (hopanes) (n-C27–C32) were identified. They included steranes (n-C27–C29), with a predominance of n-C27 [
64], and an n-C30/n-C29 ratio higher than 1.0 [
65]. All of these species were remnants of pollutants present in the original WW, and they were found in notably minor concentrations in the TWWs obtained by using any of the tested treatments (CLLE
DCM, HPFD, or CLLE
DCM–HPFD).
3.3. Physicochemical Analysis of TWW by CLLEDCM, HTFD, or CLLEDCM–HTFD
Columns 4–6 in
Table 3 show the values of the physicochemical parameters characterizing each TWW obtained by performing CLLE
DCM, HPFD, or CLLE
DCM–HPFD on WW. Except for the pH, the other values of the physicochemical parameters in the analysis of the TWW generally decreased significantly. Judging by the appearance of the TWW obtained and by the decrease in TPH, COD, and chloride, the HPFD and CLLE
DCM–HPFD treatments were more effective than CLLE
DCM in the removal of the pollutant load from WW. Nonetheless, CLLE
DCM removed most of the TPH (>99%) and diminished the COD in the WW by more than 90% (
Table 4).
Similar or higher decreases in COD and TPH from decontamination treatments of wastewater have been reported in the literature by only a few other studies using more expensive and sophisticated treatments that require special equipment and more controlled conditions than those of HTFD or CLLE
DCM [
42,
66,
67,
68].
The pH of the TWW obtained by HTFD or CLLE
DCM did not comply with the requirement of most regulations (
Table 2 and
Table 3). These pH values are explainable as an effect of the treatments changing the concentrations of chemical species in acid–base equilibria that were present as pollutants in the WW [
69,
70]. An example of such a mechanism is considered in
Scheme 1.
With the HPFD treatment, CO2(g) and H2S(g) went from the hot bottom to the steam phase. Cooling these compounds resulted in their redissolution and concentration, and the pH of the distillate decreased (pH 5.22). The solubility of CO2(g) and H2S(g) in the extract decreased as the extract warmed, and these gases were then lost from the gas phase during solvent evaporation–condensation cycles in the CLLEDCM treatment. The ions in equilibrium, as dissolved species, were concentrated in the raffinates, and the pH increased (pH 9.36).
The raffinates of the CLLE
DCM treatment with scarce CO
2(g) and H
2S
(g) products were subjected to HPFD, which produced distillate with an appropriate pH (pH 7.50). In other words, the TWW obtained by CLLE
DCM–HPFD complied with the current legal requirements of the countries examined in this study [
18,
19,
20,
21,
22,
23,
24,
25,
26].
The conductivity, salinity, and alkalinity of the raffinates obtained by CLLE
DCM or CLLE
DCM–HTFD directly related to the high concentrations of Cl
–, Na
+, and Ca
2+ in the original WW [
69,
70]. These characteristics are unfavorable for the quality of the TWW yielded from these treatments.
The peak intensities in the GC chromatograms of the TWW product of CLLE
DCM, HPFD, or CLLE
DCM–HPFD treatments of WW (
Figure 2B–D) evidenced a lower TPH content than in the starting WW (
Figure 2A). The peaks between n-C13 and n-C31 in their GC chromatograms decreased (
Figure 2C,D) or practically disappeared (
Figure 2B). The persistence of the peaks corresponding to pristane and phytane (
tR ≈ 21.7 and 23.2 min) and the increase in the pristane/phytane ratio, particularly in the GC chromatograms of the TWW obtained as raffinates (
Figure 2C,D), indicated the presence of n-alkanes, as well as methyl palmitate and methyl oleate (residuals of biodiesel,
tR ≈ 24.9 and 27.1 min), in the TWW.
Since their peaks were absent in the chromatogram of the extracts, the transfer of those species from the feed to the extracts in the CLLE
DCM process must have been incomplete. The similarities between the GC chromatograms of WW and the raffinates confirmed the common origin of the pollutants found in those samples (
Figure 2A,C,D). The prevalence of a few of such peaks in the GC chromatograms of the distillates obtained by HPFD or CLLE
DCM–HPFD indicated that a tiny fraction of the contaminant load was transferred from the WW to distillates during the process. The absence of the majority of such peaks in the GC chromatogram of the distillate showed the potentiality of HPFD for decontaminating WW (
Figure 2B).
3.4. Operation Time, Energy Consumption, and Performance of CLLEDCM, HPFD, and CLLEDCM–HPFD
Table 5 shows the average duration and the average mass of the TWW relative to the yields obtained. The average energy consumption (kWh) and time (h) required to obtain a unit of mass (kg) of the respective TWW are also shown, which were calculated after being subjected to
n measured durations and masses according to the statistical treatment described in
Section 2.4. Such measures and estimates (
Table 4 and
Table 5) were established at the given laboratory conditions described in
Section 2.3.
The CLLEDCM treatment recovered 98.5% of the initial WW mass, indicating an approximate yield of 50.0 kg/kWh of TWW at a rate of 0.1 kg/h. CLLEDCM produced yellowish and opalescent water with a pH of 9.36, and TPH and COD decreased by 99% and 90.5%, respectively, compared to the starting WW: That is, CLLEDCM did not achieve percentages of removal of TPH, COD, TH, TDSes, and TSSes that would allow the TWW obtained by this method to meet the regulatory requirements of the countries in this study.
The inferential statistical analysis (Shapiro–Wilk test) of the mean weight of solvent-free extracts collected at 20, 40, 60, 90, 120, 180, and 240 min in the four CLLE
DCM replicates (28 values) and six CLLE
DCM–HPFD replicates (42 values) indicated that these data fit a normal distribution (
p > 0.05) (
Table 6A). The results of the χ2 test indicated that the mean weights of the solvent-free extracts collected at different time points in these replicates were independent (
p > 0.05) (
Table 7).
The Welch test (
p < 0.05) was applied in cases of violation of the homoscedasticity assumption in the Levene and Bartlett tests. There were no significant differences in the mean weights of the extracts, and this is a requirement for the ANOVA test. The ANOVA analysis indicated that there were significant differences (
p < 0.05) in the mean weight of the extracts collected at different time points (
Table 7).
The results of Tukey’s test demonstrated that the data on the mean weight of the solvent-free extracts from four CLLE
DCM replicates and five CLLE
DCM–HPFD replicates were significantly different (
p < 0.05) and that the highest percentage of contaminants was removed in the first 40 min using CLLE
DCM and in the first 20 min using CLLE
DCM–HPFD (
Table 6B).
The HPFD treatment recovered 94.1%–97.3% of the initial WW mass, indicating an approximate yield of 2.8 ± 0.7 kg/kWh at a rate of 0.3 ± 0.1 kg/h. HPFD produced colorless, transparent water with a pH of 5.22 (a low value), and TPH and COD decreased by 95.0%–99.9% and 98.6%–99.6%, respectively, compared to the starting WW. The TWW obtained by HPFD presented the lowest residual values of the analytical parameters for TPH, EC, TA, TH, TDSes, TSSes, chlorides, and salinity measured in the original WW analysis (
Table 3 and
Figure 2B).
The CLLEDCM–HPFD treatment recovered 92.4% of the initial WW mass, indicating an approximate yield of 9.1 ± 0.2 kg/kWh of TWW at a rate of 0.05 kg/h. CLLEDCM–HPFD produced colorless and transparent water with a pH of 7.50, and TPH and COD decreased by 100.0% and 99.9%, respectively, compared to the starting WW. The TWW obtained by CLLEDCM–HPFD met the requirements of current regulations in WBG countries, except for Peru and China, but presented appreciable remnants of salinity, TH, TDSes, and TSSes. Among the three treatments, this method also had the highest estimated average consumption of time and energy required.
Nonetheless, energy consumption decreased and performance slightly increased (98.5%) as the operation time of HPFD increased (
Table 5). Therefore, it is still better to produce TWW with the quality achieved by individual HPFD process than wastewater with slightly better quality that meets the regulations of the analyzed countries and WBG memberˈs for effluent discharge (
Table 2 and
Table 3) at a relatively lower yield (92.4%), longer operation time, and higher energy consumption (using CLLE
DCM–HPFD) (
Table 5).
A duration of 100 min for the CLLE
DCM operation, as either an individual treatment or as part of the CLLE
DCM–HTFD sequence, is recommended in order to remove a relatively high fraction of the pollutants from WW with a reasonable minimal consumption of time and energy (
Figure 3).
The recommended total duration of operation for the other treatments (HPFD or CLLEDCM–HPFD) and the yields of TWW depend on the boiling regime, particularly at the end of the process. When the boiling was affected by the concentration or even the crystallization of poorly soluble contaminants in the distillation bottom, material projections could propagate from the bottom to the distillate collector.
Obviously, the yields, and probably the quality, of the TWW obtained by using any of these treatments can be optimized by using better-controlled conditions than those used in this first instance of laboratory-scale experiments. The laboratory conditions in this study that can be optimized include the control of losses due to handling, heating, the use of a vacuum, temperature control, and improvements in the energy transfer and utilization in each treatment system. Such optimization is outside the scope of this work, but it could be of some engineering interest to take advantage of the demonstrated potential of these treatments for decontaminating the WW generated by the petroleum industry or other sources.
4. Conclusions
This work constitutes the first systematic assay on a laboratory scale determining the potential for using CLLEDCM, HPFD, or CLLEDCM–HPFD as treatments to decontaminate wastewater generated in sedimentation ponds in the production stage of the petroleum industry (WW). The results of using these treatments demonstrated significant yields (94.1%–97.3% by HPFD, 98.5% by CLLEDCM, and 92.4% by CLLEDCM–HPFD) of water, in which decreases in TPH ranged from 95.0% (HPFD) to 100% (CLLEDCM–HPFD), decreases in COD ranged from 90.5% (CLLEDCM) to 99.9% (CLLEDCM–HPFD), and most of the other pollutants decreased compared to the starting WW: That is, water treated by HPFD and CLLEDCM–HPFD met the requirements of the permissible limits of TPH and COD in national and international regulations, including Colombia and China.
Among the tested treatments to decontaminate WW, HPFD was a more effective treatment than CLLEDCM–HPFD or CLLEDCM because it produced at least a 92.4% yield of water with an acceptable quality (only slightly less than the quality achieved by CLLEDCM–HPFD and higher than that obtained using CLLEDCM). Furthermore, HPFD neither demands the long process time or large energy consumption typical of CLLEDCM–HPFD nor produces the low water quality obtained by CLLEDCM, and its pH of 5.22 can be easily adjusted to meet all the requirements of the norms.
The demonstrated potential of HPFD could be attractive as an engineering optimization study. If the scarcity of water of good quality in the world becomes critical, humanity will need to meet the strictest regulations on the disposal and environmentally safe use of wastewater to satisfy water demand and thus preserve life on the planet, which will supersede concerns for economic and energy costs. The factors that currently render the use of HPFD prohibitive for the decontamination of WW will be comparatively minor problems among the others that engineers will have to solve in the future.