Trivalent Cr is one of the heavy metals that are difficult to be removed from soil using electrokinetic study because of its geochemical properties. High buffering capacity soil is expected to reduce the mobility of the trivalent Cr and subsequently reduce the remedial efficiency thereby complicating the remediation process. In this study, geochemical modeling and migration of trivalent Cr in saline-sodic soil (high buffering capacity and alkaline) during integrated electrokinetics-adsorption remediation, called the Lasagna process, were investigated. The remedial efficiency of trivalent Cr in addition to the impacts of the Lasagna process on the physicochemical properties of the soil was studied. Box-Behnken design was used to study the interaction effects of voltage gradient, initial contaminant concentration, and polarity reversal rate on the soil pH, electroosmotic volume, soil electrical conductivity, current, and remedial efficiency of trivalent Cr in saline-sodic soil that was artificially spiked with Cr, Cu, Cd, Pb, Hg, phenol, and kerosene. Overall desirability of 0.715 was attained at the following optimal conditions: voltage gradient 0.36 V/cm; polarity reversal rate 17.63 hr; soil pH 10.0. Under these conditions, the expected trivalent Cr remedial efficiency is 64.75 %.
In early 1992, a discussion took place between the then Monsanto Chief Executive Officer (CEO) and Administrator of the United States Environmental Protection Agency (USEPA) which ultimately led to the invention of the Lasagna process [
First application of electrokinetics took place in India in the 1930s. It was used to remove excess salts from alkali soils in order to restore it to arable condition [
Applications of Lasagna process at bench-scales from inception to date.
Treatment zone material | Contaminant | Soil type | Cell dimensions |
Polarity reversal/ |
Removal efficiency, % | Voltage gradient, (V/cm)/ current (mA) | Power consumption, kWhr/m3 | Run time, days | Electroosmotic conductivity, cm2 V−1 s−1 (×10−5) | Treatment zone spacing, cm | Reference |
---|---|---|---|---|---|---|---|---|---|---|---|
AC* + sand, bacteria + AC + sawdust |
|
Kaolinite | 10 cm ID, 21.6 cm long | Yes/ |
90–99 | 1–7/3 (constant) | 10 | 20 | 2.5 | 6 | [ |
|
|||||||||||
AC (Bamboo charcoal) | Cd | Sandy loam | 24 cm × 10 cm × 8 cm | Yes/ |
79.6 | 1/7–27 | — | 12 | — | 10 | [ |
|
|||||||||||
AC (Bamboo charcoal) | Cd | Kaolin | 24 cm × 10 cm × 8 cm | No/ |
93 | 1/3–23 | — | 8 | — | 10 | [ |
|
|||||||||||
AC (Bamboo charcoal) | 2,4-dichlorophenol and Cd | Sandy loam | 24 cm × 10 cm × 10 cm | Yes/ |
75.97 (Cd); |
1/Variable | 121.91–128.48 | 10.5 | — | 16 | [ |
|
|||||||||||
GAC** | Cr, Cd, Cu, Pb, Hg, Zn, phenol, kerosene | Saline-sodic clay | 24 cm × 10 cm × 12 cm | No/ |
75.9 (Cr); 34.4 (Cd); 41 (Cu); 55.8 (Pb); 92.49 (Hg); 26.8 (Zn); 100 (phenol); 49.8 (kerosene) | 0.6–1/880 | 1777–4273 | 21 | 425 | 6 | [ |
Applications of Lasagna process at pilot- and field-scales from inception to date.
Treatment zone material | Contaminant | Soil type | Site dimensions |
Polarity reversal/ |
Removal efficiency, % | Voltage gradient, (V/cm)/Current (A) | Power consumption, kwh/m3 | Run time, month | Electroosmotic conductivity, cm2 V−1 s−1 (×10−5) | Treatment zone spacing, cm | Reference |
---|---|---|---|---|---|---|---|---|---|---|---|
AC + sand |
|
Kaolin/ |
1.22 m × 0.61 m × 0.61 m | Yes/ |
98 | 1 (constant)/96.2 (based on current density) | 51 | 3 | 0.56–1.7 | 35.56 | [ |
|
|||||||||||
GAC1 | TCE2 | Clay loam | 4.6 m × 3 m |
Yes/ |
99 | 0.35–0.45/40 (constant) | — | 4 | 1.2 | 60 | [ |
|
|||||||||||
Iron filings + kaolin | TCE | Clay loam | 6.4 m × 9.2 m |
Yes/3-week | 95–99 | 0.23–0.31 (constant)/ |
— | 12 | 1.2 | 60 & 150 | [ |
|
|||||||||||
Iron filings + kaolin | TCE | Clay loam | 27.4 m × 22 m |
No/pulse mode | 99 | 0.15–0.26 (constant)/ |
— | 24 | — | 150 | [ |
|
|||||||||||
Iron filings + kaolin | TCE | Clay loam | 33 m × 24 m |
— | 60 (after 1 year) | 0.16 (constant)/ |
— | 24 | — | 150 | [ |
The geochemical properties of the most stable forms of Cr, that is, trivalent and hexavalent Cr under electrokinetic remediation, have been extensively studied in different types of soils (kaolin, glacial till, etc.) by Reddy and his coworkers and other investigators [
Redox potential (Eh)-pH diagram for Cr–O–H system [
Empirical modeling using response surface methodology (RSM) offers great and numerous advantages which include large amount of information from a small number of experiments, evaluation of simultaneous interaction effects of the independent parameters on the responses, and simultaneous optimization of multiple factors and responses for obtaining optimal conditions [
Some microbially-driven biotransformation processes may affect the soil physicochemical properties after electrokinetic remediation because of the passage of electric current and development of pH gradients [
Natural saline-sodic clay, obtained from Al-Hassa Oasis, Saudi Arabia, was used in this study. The soil has the following characteristics: pH (8.3), moisture content (3.91%), soil organic matter (2.59%), electrical conductivity (15.24 dS/m), specific surface area (9.07 m2/g), pore volume (0.014 cm3/g), pore size (62.55 Å)—mineralogy from X-ray diffraction (XRD), quartz (SiO2) (87.4%), calcite (CaCO3) (5.2%), and dolomite (CaMg(CO3)2) (7.4%). X-ray fluorescence spectroscopy (XRF) revealed that the soil consists of the following elements: Ca (37.64%), Si (34.73%), Fe (10.41%), Al (7.6%), K (3.42%), Mg (2.48%), Pd (2.85%), and Ti (0.86%). These properties were determined using methods of the American Society of Testing and Materials (ASTM) standards and were reported elsewhere [
Single and competitive adsorption of five heavy metals (Cr, Cd, Cu, Zn, and Pb) were performed to determine the selectivity sequence and to understand the adsorption behavior of these metals under different pH conditions. This is particularly important to this study, because soil mineralogy affects heavy metal adsorption behavior and selectivity sequence under different pH conditions. Lukman et al. [
Fifteen (15) bench-scale experiments, each having a 21-day run time, were designed and performed to investigate the migration and distribution of trivalent Cr in a contaminant mixture using the coupled electrokinetics-adsorption technique and to understand the operating variables’ effects on saline-sodic soil.
The Plexiglas reactor total volume was about 2268 cm3, made of seven chambers. The overall reactor dimensions are 24 cm (long) × 10 cm (width) × 12 cm (depth). Approximately 1 kg of local KSA soil was artificially spiked with kerosene, heavy metals (Cu, Cr, Cd, Pb, Zn, and Hg), and phenol at predetermined concentrations. Thorough mixing was done using mechanical mixer (Gilson Company Inc.) so as to achieve a homogeneous distribution of the contaminants around the soil matrix. The mixed spiked soil was placed in a fume-hood for drying over a period of time necessary to evaporate the solvents (hexane and distilled water). Distilled water was added to adjust the final moisture content of the soil to about 33–70%. The initial conditions of the soil pH, moisture content, organic matter, and electrical conductivity were measured as well as the actual initial concentrations of the contaminants. Then, the uniformly mixed contaminated soil was placed into the cell layer by layer. Each layer was compacted with stainless steel spatula so that the amount of void space was minimized. The reactor used for the experiments consists of the cell, two graphite electrodes serving as anode and cathode, DC power supply (LG, GP-505), processing fluid reservoirs, heavy duty recirculation pump (BVP Instratec), portable data logger (TDS-303, Tokyo Sokki Kenkyujo Co., Ltd) for real-time monitoring of temperature, electric current, and voltage across the system (Figure
Coupled electrokinetics-adsorption experimental setup.
To evaluate reliability of the analytical procedures, duplicate samples were analyzed for each sample. Quality control (QC) protocols spelt out in EPA method 7000B [
Box-Behnken design (BBD) was chosen for the experimental design because of its advantages over central composite design (CCD) and 3-level factorial design when dealing with only three factors. In BBD, the experimental points are hyperspherically arranged, equidistant from the central point [
Codification and ranges of factors.
Variable | Designation | Units | Coded variable levels | ||
---|---|---|---|---|---|
−1 | 0 | +1 | |||
Polarity reversal |
|
hr | 0 | 24 | 48 |
Voltage gradient |
|
V/cm | 0.2 | 0.6 | 1 |
Concentration |
|
mg/kg | 20 | 60 | 100 |
Design of experimental runs using the Box-Behnken design.
Run order | Polarity reversal, |
Voltage gradient, |
Concentration, |
Remedial efficiency, % |
---|---|---|---|---|
1 | 0 | 0.6 | 20 | 0.00 |
2 | 48 | 0.6 | 20 | 0.00 |
3 | 24 | 1 | 20 | 0.00 |
4 | 24 | 1 | 100 | 0.00 |
5 | 24 | 0.6 | 60 | 79.97 |
6 | 0 | 1 | 60 | 72.73 |
7 | 24 | 0.2 | 20 | 0.00 |
8 | 0 | 0.2 | 60 | 36.93 |
9 | 48 | 1 | 60 | 65.66 |
10 | 48 | 0.6 | 100 | 0.00 |
11 | 0 | 0.6 | 100 | 25.50 |
12 | 24 | 0.2 | 100 | 0.00 |
13 | 48 | 0.2 | 60 | 34.88 |
This experimental design was preliminarily evaluated using variance inflation factor (VIF) to check for orthogonality (independence of factors) and leverage which quantitatively measures the potential of a design point to have significant influence on model fit [
Following design evaluation, the responses were fitted to a quadratic model which was fine-tuned by removing any insignificant term. This will maximize
The developed models were evaluated using the rich diagnostic tools provided in Design-Expert which include normal plot of residuals (to test the assumption of normality of residuals), predicted versus actual plot (to test the assumption of constant variance), Box-Cox plot (to check the need for data transformation), and externally studentized residuals (to check the presence of any outlier in the data). The effects of factors were compared at a particular point in the design space using the perturbation plot. Response surface and contour plots were then generated.
Desirability function, being one of the mathematical methods for computation of critical values (maximum or minimum) and measuring overall success of optimizing multiple responses using geometric mean, was employed for the optimization of trivalent Cr remedial efficiency. A search for a combination of factor levels which simultaneously satisfies the goals imposed on factors and responses is first carried out, followed by combining these goals into an overall desirability function that ranges from 0 (outside of the optimization limits) to 1 (at the goal). Combining all responses into overall desirability eliminates favoring one response over another. The aim is not to clinch a desirability value of 1 but to find a good set of conditions that will meet all the set goals for each factor and response [
Discussion of the monitored results obtained after performing thirteen (13) tests with 3 centre points will be focused on geochemical processes affecting sorption/desorption and migration/removal mechanisms such as the development of acid/base fronts, migration and reactions, dissolution/precipitation, oxidation/reduction reactions, complexation, and metallic ion speciation. In addition, presentation of the developed mathematical models and discussion on how the factors affect the respective responses will follow.
Lukman et al. [
The saline-sodic nature of the soil necessitates the use of processing fluids (2N NaOH and 1N HNO3) to continuously neutralize the rapidly generated H+ and OH− ions at the anode and cathode, respectively. These fluids were monitored every 8 hours and replaced as they degraded. HNO3 and NaOH are strong acid and base, respectively, and dissociate completely according to the following reactions:
Because of the electrochemical decomposition of water, OH− and H+ ions are produced at the cathode and anode, respectively, as shown in (
The electrochemically generated H+ and OH− ions due to water electrolysis at the anode and cathode, respectively, are neutralized to form water molecules (
Weekly pH variation.
Weekly soil electrical conductivity variation.
Cumulative electroosmotic volume for each test.
It was observed from Figure
Perturbation plots showing the relative significance of factors on soil pH (a) and electrical conductivity (c) (left). 3D response surface and contour plots showing how the influential factors affect soil pH (b) and electrical conductivity (d) (right).
Anderson and Whitcomb [
pH profile with two GAC treatment zones for investigating bipolar effects (R11).
Sparks [
These parameters make the measured electroosmotic volume for all the tests to vary temporally. The reduction of the thickness of the diffuse double layer resulting from higher ionic concentration with subsequent higher ionic strength causes reduction in the electroosmotic flow [
(a) Perturbation plot showing the relative significance of factors on electroosmotic volume. (b) 3D response surface and contour plots showing the influence of voltage gradient on cumulative electroosmotic volume.
Comparing electrical current with voltage gradient and soil pH for all tests.
Run | Current, A | Voltage gradient, V/cm | pH |
---|---|---|---|
R6 | 3.02 | 1 | 12.9 |
R9 | 2.65 | 1 | 12.6 |
R3 | 2.25 | 1 | 12.6 |
R4 | 2.04 | 1 | 12.7 |
R2 | 1.32 | 0.6 | 10.9 |
R1 | 1.17 | 0.6 | 10.2 |
R10 | 1.12 | 0.6 | 11.9 |
R5 | 1.03 | 0.6 | 11.2 |
R11 | 0.61 | 0.6 | 12.0 |
R8 | 0.21 | 0.2 | 9.8 |
R12 | 0.15 | 0.2 | 8.3 |
R7 | 0.14 | 0.2 | 8.1 |
R13 | 0.13 | 0.2 | 10.4 |
Comparing variations of electric current with soil temperature: (a) current; (b) temperature.
(a) Perturbation plot showing the relative significance of factors on average electric current. (b) 3D response surface and contour plots showing the influence of voltage gradient on average electric current.
Figure
Trivalent Cr distribution and migration from the contaminated chamber to the GAC chambers after 13 tests.
Speciation diagram for trivalent Cr species at different weekly pH values.
Mass balance analyses of Cr were performed for Runs 8, 11, and 13. From Table
A sample mass balance analysis of trivalent Cr for Runs 8, 11, and 13.
Runs | Run 8 | Run 11 | Run 13 |
---|---|---|---|
Initial concentration, mg/kg | 37.20 | 77.95 | 37.20 |
Residual concentration, mg/kg | 23.46 | 58.08 | 24.23 |
|
|||
1st GAC chamber, F | |||
Initial concentration, mg/kg | 6.90 | 6.90 | 6.90 |
Residual concentration, mg/kg | 21.15 | 0.00 | 19.70 |
|
|||
2nd GAC chamber, G | |||
Initial concentration, mg/kg | 6.90 | 6.90 | 6.90 |
Residual concentration, mg/kg | 14.48 | 0.00 | 25.30 |
|
|||
Mass balance, % | 121.75 | 74.51 | 148.99 |
The tests were sorted in decreasing order of remedial efficiency (Table
Comparing trivalent Cr remedial efficiency with factors and some responses.
Runs | Remedial | Current, | Residual, | Electroosmotic volume, | Polarity reversal | Voltage gradient, | Initial Cr |
---|---|---|---|---|---|---|---|
efficiency, % | A | pH | mL | rate, hr | V/cm | concentration, mg/kg | |
R5 | 79.97 | 1.03 | 11.2 | 2344.50 | 24 | 0.6 | 60 |
R6 | 72.73 | 3.02 | 12.9 | 1201.50 | 0 | 1 | 60 |
R9 | 65.66 | 2.65 | 12.6 | 1399.50 | 48 | 1 | 60 |
R8 | 36.93 | 0.21 | 9.8 | 81.00 | 0 | 0.2 | 60 |
R13 | 34.88 | 0.13 | 10.4 | 63.00 | 48 | 0.2 | 60 |
R11 | 25.50 | 0.61 | 12.0 | 1387.84 | 0 | 0.6 | 100 |
R1 | 0.00 | 1.17 | 10.2 | 1728.00 | 0 | 0.6 | 20 |
R2 | 0.00 | 1.32 | 10.9 | 2542.50 | 48 | 0.6 | 20 |
R3 | 0.00 | 2.25 | 12.6 | 814.50 | 24 | 1 | 20 |
R4 | 0.00 | 2.04 | 12.7 | 2272.50 | 24 | 1 | 100 |
R7 | 0.00 | 0.14 | 8.1 | 396.00 | 24 | 0.2 | 20 |
R10 | 0.00 | 1.12 | 11.9 | 2236.50 | 48 | 0.6 | 100 |
R12 | 0.00 | 0.15 | 8.3 | 324.00 | 24 | 0.2 | 100 |
Weekly percentage removal of trivalent Cr for 13 tests.
Equation (
(a) Perturbation plot showing the relative significance of factors on trivalent Cr remedial efficiency. (b) 3D response surface and contour plots showing the influence of initial contaminant concentration on trivalent Cr remedial efficiency.
Experimental validation of trivalent Cr remedial efficiency and soil pH using voltage gradient = 1 V/cm; average concentration = 44.15 mg/kg; and polarity reversal rate = 0 hr.
Response | Experimental result | Model prediction | Prediction error, % | 90% CI* | 90% CI | 90% PI** | 90% PI |
---|---|---|---|---|---|---|---|
low | high | low | high | ||||
Cr, remedial efficiency | 75.88 | 51.11 | 32.64 | 31.17 | 75.95 | 18.36 | 100.00 |
Residual soil, pH | 12.3 | 12.6 | 2.35 | 11.7 | 13.5 | 10.8 | 14.0 |
**Prediction interval.
Optimal factor levels required to maximize remedial efficiency of trivalent Cr.
Item | Value |
---|---|
Polarity reversal, hours | 17.63 |
Voltage gradient, V/cm | 0.36 |
Concentration, mg/kg | 60.00 |
Expected remedial efficiency of trivalent Cr | 64.75 |
Expected residual soil pH | 10.00 |
Desirability | 0.715 |
Combined and individual response desirability values for all responses and factors.
3D surface plot of the overall desirability variation relative to influential factors.
Preceding sections have elaborately discussed and modeled the impacts of the proposed remediation technique on the soil pH and electrical conductivity. Additionally, the passage of electric current and soil pH gradients will result in the following physicochemical interactions:
Values of soil surface area and pore volume and size, before and after treatment.
Description | BET* surface area, | Pore volume, | Pore size, |
---|---|---|---|
m2/g | cm3/g |
| |
Before | 9.07 | 0.014 | 62.55 |
After | 11.21 | 0.045 | 163.24 |
Soil mineralogical transformations before and after treatment.
Phase name | Before, % | After, % |
---|---|---|
Quartz, SiO2 | 87.4 | 55.3 |
Calcite, CaCO3 | 5.2 | 44.7 |
Dolomite, CaMg(CO3)2 | 7.4 | — |
Values of constituent soil elements, before and after treatment.
Element | Before, % | After, % |
---|---|---|
Ca | 37.64 | 42.06 |
Si | 34.73 | 23.42 |
Fe | 10.41 | 15.06 |
Al | 7.6 | 9.55 |
K | 3.42 | 4.61 |
Mg | 2.48 | 2.49 |
Pd | 2.85 | 1.46 |
Ti | 0.86 | 1.35 |
The study reported herein investigated the migration of trivalent Cr ions from a multiple contaminated natural saline-sodic soil. The soil salinity and sodicity, which provided large amount of dissolved salts and minerals (carbonates) in the pore fluid for sustained high electrical conduction, were responsible for the extremely high electric current flow. This led to excessive soil heating, high energy and process fluid consumption, high electroosmotic volume, and in some cases higher percentage removal of trivalent Cr. Significant migration of Cr from the contaminated chamber to the granular activated carbon chamber was recorded which led to highest remedial efficiencies (79.97–34.88%) for tests involving 60 mg/kg initial trivalent Cr concentration, whereas no removal was recorded for all tests involving 20 mg/kg. Even under low electric current, electroosmotic flow, and voltage gradient (0.2 V/cm), up to 36.93% of the trivalent Cr was removed from the contaminated chamber. It has been shown that high voltage gradient (1 V/cm) or passage of high electric current does not necessarily translate into high remedial efficiency. Bipolar effects did not manifest due to the presence of carbonate minerals that impact high acid buffering capacity. For test without polarity reversal, trivalent Cr moved toward the anode due to the formation of high amount of anionic
The authors declare that there is no conflict of interests regarding the publication of this paper.
The authors would like to acknowledge the support provided by King Abdul-Aziz City for Science and Technology (KACST) through the Science & Technology Unit at King Fahd University of Petroleum & Minerals (KFUPM) for funding this work through Project no. 11-Env1669-04, as part of the National Science, Technology and Innovation Plan.