Remediation of NAPL-contaminated porous media using micro-nano ozone bubbles: Bench-scale experiments

Journal Pre-proof Remediation of NAPL-contaminated porous media using micronano ozone bubbles: Bench-scale experiments

Hobin Kwon, Mohamed M. Mohamed, Michael D. Annable, Heonki Kim PII:

S0169-7722(19)30199-8

DOI:

https://doi.org/10.1016/j.jconhyd.2019.103563

Reference:

CONHYD 103563

To appear in:

Journal of Contaminant Hydrology

Received date:

21 June 2019

Revised date:

8 September 2019

Accepted date:

19 October 2019

Please cite this article as: H. Kwon, M.M. Mohamed, M.D. Annable, et al., Remediation of NAPL-contaminated porous media using micro-nano ozone bubbles: Bench-scale experiments, Journal of Contaminant Hydrology(2018), https://doi.org/10.1016/ j.jconhyd.2019.103563

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2018 Published by Elsevier.

Journal Pre-proof

of

Remediation of NAPL-Contaminated Porous Media Using Micro-nano Ozone Bubbles: Bench-scale Experiments a Hobin Kwon , Mohamed M. Mohamedb,c,**,Michael D. Annabled, Heonki Kima,* a Dept. of Environmental Sciences and Biotechnology, Hallym University, Chuncheon, Gangwon-do, 24252, Korea b Civil and Environmental Engineering Dept., United Arab Emirates University, Al Ain, 15551, UAE c National Water Center, United Arab Emirates University, Al Ain 15551, UAE d Dept. of Environmental Engineering Sciences, University of Florida, Gainesville, Florida, 32611, USA

Jo ur

na

lP

re

-p

ro

*Corresponding author at: Dept. of Environmental Sciences and Biotechnology, Hallym University, Chuncheon, Gangwon-do, 24252, Korea E-mail address: [email protected] (H. Kim) ** Corresponding author at: Civil and Environmental Engineering Dept., United Arab Emirates University, Al Ain, 15551, UAE E-mail address: [email protected] (M.M. Mohamed)

Journal Pre-proof

Abstract Aqueous solutions of micro-nano bubbles (MNBs) containing ozone gas were injected through a NAPL-contaminated glass bead column. The glass column (15 cm × 2.5 cm) was packed with glass beads: the first 12 cm was packed with coarse glass beads while much finer glass beads were used to pack the remaining 3.0 cm of the column. Decane was used as the representative NAPL, to which an oil-soluble fluorescence tracer was

of

added. The fluorescence tracer was considered as a constituent of the NAPL that readily

ro

reacts with ozone. Air and ozone-containing oxygen were used to generate MNB

-p

solutions, and injected through the column. In addition, H2O2 was introduced to the O3containing MNB (O3-MNB) solution to investigate the effect of hydroxyl free radicals

re

on the NAPL removal. An ozone gas sparging experiment was also conducted for

lP

comparison. After 72 hours of O3-MNB application, a significant mass of n-decane (27.6 % of the initial mass applied) was removed from the column. H2O2 injection into

na

the column during O3-MNB application was effective in increasing the n-decane mass

Jo ur

removal by 22 %, compared to the O3–MNB experiment. The rate of NAPL removal during O3-MNB flushing was significant, although slower than ozone sparging. During O3-MNB application, fast decay of fluorescence was observed; whereas, during coinjection of H2O2 and O3-MNB solutions, only a slight change in the fluorescence was observed. This indicates that oxidative degradation of NAPL during H2O2 and O3-MNB injection takes place only at the NAPL-water interface due to the reactivity of hydroxyl free radical, whereas ozone diffusion into NAPL induced the decay of the fluorescence tracer in the bulk NAPL. The removal characteristics during MNB application and ozone gas sparging were investigated based on the analysis of NAPL using mass spectrophotometer. When O3-MNB and H2O2 were co-injected, only n-decane was

2

Journal Pre-proof

detected in the NAPL; while when O3-MNB was used for flushing, oxidative products were found in the NAPL. More hydrophilic compounds were found in the NAPL after ozone sparging. This implies different removal mechanisms depending on the kind of oxidation agent, and the state of oxidizing fluid. Based on the findings in this study, the application of O3-MNB could be a feasible option for cleaning up NAPL-contaminated

of

aquifers.

Jo ur

na

lP

re

-p

ro

Key words: aquifer, remediation, micro-nano bubbles, NAPL, ozone, oxidation

3

Journal Pre-proof

1. Introduction The subsurface environment can be contaminated by non-aqueous phase liquids (NAPLs), due to accidental releases from NAPL-processing facilities. Once stabilized in an aquifer (or the vadose zone), they tend to persist at the location for extended periods of time, serving as a contaminant source zones. The fate of NAPLs in the source zones differs depending on their physicochemical properties and the hydrogeological

of

conditions. NAPLs include a wide spectrum of petroleum products with different

ro

aqueous solubility, density, microbial availability, viscosity, interfacial properties,

-p

making it challenging to remove NAPL from the subsurface environment.

re

A number of subsurface remediation technologies have been developed depending on the nature of contaminants and hydrogeological conditions. One practice for in situ

lP

NAPL removal is the application of chemicals to the source zone for oxidative

na

contaminant degradation (Krembs et al., 2010). This approach targets contaminants with low aqueous solubility, low vapor pressure, and being recalcitrant to biodegradation.

Jo ur

Oxidizing agents include chemicals with high oxidizing potential. Oxygen and ozone are common gases that can be used for this purpose (Alcantara-Garduno et al., 2008; Kim et al., 2013; Nimmer et al., 2000). Water dissolved oxidizing agents are also used. Permanganate, percarbonate, and persulfate are examples of aqueous oxidizing agents (Kao et a., 2008). The more powerful oxidizing agent, hydroxyl free radical, is often generated in situ by the mixing of chemicals (e.g., ozone and hydrogen peroxide, zerovalent iron and hydrogen peroxide) to degrade NAPL constituents in the source zone (Watts and Teel, 2005; Tawabini, 2014). The direct injection of gaseous oxidizing agent (e.g., ozone) is effective only in the radius of influence where the ozone-containing gas is in contact with the NAPL. The

4

Journal Pre-proof

size of the radius of influence in the aquifer during gas sparging is very limited to the vicinity of the injection points (McCray and Falta, 1996). Although the radius of influence may be extended using surfactants (Kim et al., 2004), the gas flow is still confined to a limited region. At the same time, the gas flows through higher permeable regions in the aquifer, resulting in significant preferential flow, leaving lower permeable regions un-impacted (Tomlinson et al., 2003; Tsai, 2007). In contrast, aqueous phase

of

oxidizing agents may flow through the aquifer, sweeping both low and high permeable

ro

regions. (Sung et al., 2017). However, water-soluble oxidizing agents often contain

-p

heavy metals (e.g., manganese), which may be toxic to ecosystems or decomposes

re

quickly in the aqueous solution, limiting their potential for degradation of contaminants. Micro-nano bubble (MNB) solutions may be a feasible option for the continuous

lP

introduction of gaseous oxidizing agents (e.g., ozone) to aquifers. Recently, MNB

na

solutions have received attention in the field of groundwater remediation, due to its potential as an effective carrier of gaseous compounds in aqueous solutions. MNBs are

Jo ur

gas bubbles in the diameter range from sub-micrometer to a few micrometers, and are known to be stable in liquids at high internal gas pressures for substantial periods of time (e.g., several days) (Ushikubo et al., 2010). Early studies on MNBs focused on their generation, characterization and physical properties in aqueous solutions (Alheshibri et al., 2016; Kwan and Borden, 2010; Liu et al., 2013). MNB solutions show typical properties of colloidal suspensions scattering light. In aqueous solutions they shrink and disappear over time due to high gas pressures in the bubbles and the associated gas transfer to the bulk phase, before they coagulate and float up to the water surface. Ozonation processes using MNBs containing gaseous ozone were found to significantly increase the degradation rate of a wide range of organic compounds in bulk

5

Journal Pre-proof

solutions (Ikeura et al., 2011; Khuntia et al., 2012; Li et al., 2009; Takahashi et al., 2012; Yasuda and Ban, 2012). The transport study of MNB solutions through porous media confirmed that the solution flows in a plug-flow manner when the bubbles are at micrometer scale (Choi et al., 2008). Other studies detected significant analogies to the transport of colloidal solutions, such as the retention of MNBs in the interstices of porous media particles

of

(Hamamoto et al., 2017, 2018; Wan et al., 2001). Moreover, MNB transport is affected

ro

by chemical properties (i.e., zeta potential) as well as the size of bubbles (Hamamoto et

-p

al., 2018). Introducing tiny air (or oxygen) bubbles into groundwater was also studied

re

for their potential to enhance dissolved oxygen levels which may help to stimulate microbial communities for bioremediation (Li et al., 2014).

lP

MNBs containing air or oxygen could be used to introduce oxygen into the aquifer

na

over extended temporal and spatial scales at high concentrations for bioremediation of a NAPL-contaminated source zone (Li et al., 2014), whereas ozone-containing MNBs

Jo ur

(O3-MNBs) are generated in aqueous solution, and applied for the more aggressive chemical treatment of NAPL (Hu and Xia, 2018; Agarwal et al., 2016; Sung et al., 2017). The overall mass transfer rate of ozone from MNBs to bulk water was increased and faster degradation of organic chemicals (e.g., water soluble dye) was also observed (Hu and Xia, 2018). Only limited applications of MNBs at the field scale have been reported (Scheffer and van de Ven, 2010). Despite its promising potential for degrading organic contaminants, only a few studies have been performed on the effectiveness of O3-MNB solutions, providing limited quantitative information for the remediation of NAPL-contaminated porous media. The primary objective of this study was to evaluate the mass removal of NAPL (n-

6

Journal Pre-proof

decane) from porous media (glass beads) using an O3-MNB solution. A onedimensional bench-scale column packed with glass beads of different sizes was used for the NAPL removal experiments. The mass removal of NAPL from the column under different flushing conditions including air-containing MNB solution, O3-MNB and H2O2 mixture solution, and bulk water, was estimated and compared with that from an O3-MNB solution flushing experiment. The purpose of using a mixture of O3-MNB and

of

H2O2 solutions was to evaluate the effect of hydroxyl free radicals formed by the

ro

reaction of H2O2 and O3 molecules on NAPL removal. The degradation of NAPL due to

-p

ozone gas sparging for the same experimental setting was also evaluated, and compared

re

to that from the MNB solution flushing experiments. The characteristics of the oxidative reactions and the locations of those reactions during the O3-MNB, the mixture of O3-

lP

MNB and H2O2 solution flushing, and ozone gas sparging were investigated based on

na

the GC-MS analysis of the NAPL following treatments.

2.1 Materials

Jo ur

2. Materials and methods

Two different glass bead types (0.75 – 1.0 mm, 0.05 – 0.15 mm diameter) were used as the packing material in the glass column (2.54 cm id, 15.0 cm length). Reagent grade n-decane (> 99.9%, Sigma-Aldrich Co.) was used as the NAPL. An oil-soluble dye (DFSB-K175, Riskreactor Co.) was dissolved in n-decane, and was used as the fluorescence tracer and representative organic contaminant in NAPL. Reagent grade (99 %) 2, 4 – dimethyl – 3 - pentanol (DMP) was provided by Sigma-Aldrich Co., and was used as the NAPL-partitioning tracer. Bromide (> 99.9%), as in potassium bromide,

7

Journal Pre-proof

was provided by Sigma-Aldrich Co., and used as the non-reactive tracer. HPLC grade methylene chloride from J.T. Baker Co. was used as the extraction agent. Hydrogen peroxide solution (35 %) was provided by Junsei Chemicals Co. (Tokyo, Japan) and was diluted using DI water prior to use. Tap water was used throughout this study.

2.2 Experimental setup

of

Figure 1 shows a schematic of the experimental set–up used in this study. The water

ro

in the MNB solution reservoir was fed into the MNB generator (KET-1, Korea EMB

-p

Technology Co., Ltd., Incheon City, Korea). The MNB solution was circulated while a

re

constant volume (50 L) of MNB solution was maintained. The fraction of MNB solution in the reservoir was replaced with fresh tap water at the rate of 1.0 L/min, while the

lP

solution was drained at the same rate to maintain constant temperature at 25 ℃. Gases

na

were introduced to the MNB generator by control valves (v1, v2, and v3). An ozone generator (Model LAB2B, Ozonia Co.) was installed between the oxygen gas reservoir

Jo ur

and MNB generator to produce the O3-MNB solution. The ozone concentration in the gas effluent from the ozone generator was 43.5 ± 0.2 g/Nm3, measured by the standard iodometric method provided by the International Ozone Association. Two black lamps were installed close to the column. These lamps emit long-wave UV light (UV-A) which activated the fluorescence of the dye added to the n-decane. A reservoir of H2O2 solution was connected to the column. By operating the peristaltic pump (p2) and the valve (v4), the H2O2 solution was introduced to the column. The lower (inlet) part of the column (12.0 cm) was packed with coarse glass beads (0.75 – 1.0 mm), whereas the top 3.0 cm of the column (outlet) was packed with fine glass beads (0.05 - 0.15 mm). The overall porosity of the column was 0.40. The column was connected to a peristaltic

8

Journal Pre-proof

pump (p1) by which a constant flow of aqueous solution through the column was achieved. All experiments were conducted at 25 ± 2 °C.

2.3 Experimental procedure

2.3.1 Gas saturation change during MNB flushing

of

The dry-packed glass bead column, described in the previous section, was flushed

ro

with de-gassed DI water to achieve complete water saturation. The water-saturated

-p

column was then installed in the flow system (Fig. 1), and flushed with air-containing

re

MNB (Air-MNB) solution. The air-MNB solution was generated using air fed to the MNB generator at the rate of 2.0 L/min. The air-MNB solution was injected into the

lP

column using a peristaltic pump (p1) at the rate of 2.0 mL/min. The gas (air) saturation

na

of the column was estimated based on the weight of the column. Each column was weighed at the beginning (water-saturated condition), during MNB flushing, and when

Jo ur

the flushing experiment was completed. Upon completion of the air-MNB flushing experiment, the column was re-saturated with DI water by flushing the column with de-gassed DI water. The re-saturated column was then flushed with oxygen-containing MNB (O2-MNB) solution. The feed oxygen gas bypassed the ozone generator, and was introduced directly to the MNB generator from the compressed oxygen reservoir. The gas content in the column was estimated by the weight change of the column. After the O2-MNB flushing experiment was completed the O3-MNB flushing experiment was conducted. The experimental procedure was the same, except ozone was fed to the MNB generator. To produce ozone, the oxygen reservoir was connected to the ozone generator at the flow rate of 2.0

9

Journal Pre-proof

L/min. All experiments were conducted in triplicate to measure the maximum air saturations.

2.3.2 NAPL removal - MNB flushing experiments (MNB-1~4) The dry-packed glass bead column was saturated by flushing the column with sufficient amount (at least 10 pore volumes) of de-gassed DI water. About 10 mL of n-

of

decane solution (containing 200 mg/L fluorescence dye) was injected into the column

ro

through the inlet at the bottom of the column using a glass syringe, followed by flushing

-p

with DI water at the rate of 10 mL/min for 2 min. The column was then flushed with DI

re

water at the rate of 2 mL/min for 30 min. leaving n-decane at the residual saturation. A series of partitioning tracer tests were conducted to evaluate the n-decane saturation.

lP

About 0.25 pore volume (PV) of the tracer solution containing DMP (830 mg/L) and

na

bromide (30 mg/L) was injected into the column at the flowrate of 1.0 mL/min using a high precision pump (M925, Yonglin Co., Anyang City, Korea). The effluent of the

Jo ur

column was then analyzed for the breakthrough curves (BTCs) of the tracers. An HPLC system (YL9120, Yonglin Co.) was used for the bromide analysis: column - Dionex AS4A, eluent – sodium borate 3.0 mM, eluent flowrate 2.0 mL/min, detection wavelength – 205 nm. A GC system (Agilent 6890 plus) was used to analyze the DMP extract (solvent: methylene chloride): column – DB 624, temperature: injector - 80℃, oven - 160℃, detector (FID) - 250℃. Information regarding data process for tracers [i.e., calculation of temporal moments of the BTCs and resulting retardation factors (Rt)] are available in previous studies (e.g., Jin et al., 1995). The column was then installed in the MNB flushing system (Fig. 1). For the control experiment (MNB-1, Table 1), the tap water in the water reservoir (no MNB applied)

10

Journal Pre-proof

was introduced into the column at the flow rate of 2.0 mL/min. The flushing continued for 72 hours. Photos of the fluorescence emission were taken during flushing with the black lamps on. Upon the completion of water flushing, the column was weighed and flushed using degassed DI water. Another set of tracer tests were conducted to evaluate the NAPL saturation after the flushing process. After the flushing experiment (and the tracer test), the column was dismantled, and the n-decane left in the column was

of

extracted using 100 mL of methylene chloride, followed by GC analysis, as previously

ro

described.

-p

For the air-MNB flushing experiment (MNB–2), compressed air was fed into the

re

MNB generator (2.0 L/min). The air-MNB solution was introduced to the column in the same manner as described previously (section 2.3.1). The MNB solution flushing

lP

continued for 72 hours and photos of the fluorescence emission from the column under

na

the black lamps were taken. These photos were overlapped with a transparent grid film and the area where the fluorescence emission disappeared was estimated. The column

Jo ur

effluent was collected in a flask containing a pre–determined volume of extraction solvent (methylene chloride), to measure the mass of n-decane washed out of the column during MNB flushing. After completion of the flushing process, the weight of the column was measured. The column was flushed using de-gassed DI water for resaturation. Tracer tests were conducted to measure n-decane saturations before and after the air-MNB flushing. The flushing experiment (MNB-3) using O3-MNB followed the same procedure as that of the air-MNB experiment, except the gas fed to the MNB generator was ozone. After MNB-3 and MNB-4 experiments were completed, the column weight changes were measured for the final gas saturations. The H2O2 solution was applied for experiment MNB-4 during the O3-MNB flushing

11

Journal Pre-proof

experiment. H2O2 solution was injected using the pump (p2) and the valve (v4) while the O3-MNB solution was introduced into the column. The injection rates of H2O2 and O3MNB solutions were 0.50 mL/min, and 2.0 mL/min, respectively, resulting in the total flow rate of 2.50 mL/min. The H2O2 concentration in the column was estimated to be 0.81 g/L. Tracer tests were conducted before and after the flushing experiments as previously described. Table 1 lists the experimental conditions used in this study. The

of

final masses of n-decane after the flushing experiments of MNB-2, MNB-3, and MNB-

ro

4 were measured using the same methods as MNB-1. All MNB experiments including

-p

the control (MNB-1) were duplicated.

re

The methylene chloride extracts of the column after experiments MNB-3 and MNB-4 were analyzed using a GC-MS (7890D-5977MSD, Agilent Technologies Co.) to

lP

investigate any oxidative products formed in the NAPL phase during MNB flushing. A

na

capillary column (UA-5, 30 m, Frontier Lab Co., Japan) was used for the GC analysis at

10 – 800 m/z.

Jo ur

the oven temperature of 150 – 250 ℃ (5 ℃/min ramping rate). MS scan range was set to

2.3.3 NAPL removal - ozone sparging experiment (MNB-5) The ozone sparging experiment (MNB–5) was conducted. Ozone-containing gas (same gas used for MNB–3 and MNB–4) generated by the ozone generator was injected directly into the n-decane loaded column at the gas flow rate of 2.0 mL/min. The same experimental procedure was used as for MNB flushing experiments except that ozone gas was introduced to the column instead of MNB solution. The column packing process, tracer tests for the n-decane saturation before and after the sparging experiment, and the analytical process for the methylene chloride extract of the column

12

Journal Pre-proof

after the sparging process were the same as for the MNB flushing experiments described in the previous section.

3. Results and discussion

3.1 Gas saturation change during MNB flushing

of

Figure 2(a) shows the air saturation change during the MNB flushing process for the

ro

column with no NAPL present. The air saturation increased at a constant rate up to

-p

approximately ten pore volumes, and stabilized. Monotonic increase in the air saturation

re

during the first few pore volumes suggests that all the bubbles were caught in the coarse medium, due to the finer glass bead layer installed at the end of the column. The

lP

accumulation of gas bubbles in the porous medium reported by Hamamoto et al. (2017),

na

showed a similar pattern for the increase in the gas saturation (Sg, volume of gas/total void volume) during MNB injection into homogeneous media. When the Sg in the

Jo ur

column reached at a critical level (0.16 – 0.19), the gas started to migrate through the fine glass bead layer, resulting in a stable overall trapped gas saturation in the column. During the steady state condition (no change in Sg), the aqueous solution flow rate decreased slightly from the set flow rate of 2.0 mL/min due to gas bubbles in the column effluent. The slopes of the linear increase (up to 8 pore volumes) are 0.0183, 0.0180, and 0.0151 for O3-, O2- and air-MNB flushing, respectively, which correspond to gas contents (mLgas/mLsolution) of 1.83 %, 1.80 %, and 1.51 % in the O3-, O2- and air-MNB solutions, respectively. Using the mean diameter (4.9 ± 3.3 μm) of the bubbles (O2MNB) estimated from the microscope observation [Fig. S-1 in Supplementary

13

Journal Pre-proof

Material], the concentrations of bubbles for O3-, O2- and air-MNB solutions were 2.8 × 108, 2.7 × 108, and 2.3 × 108 bubbles/mLsolution, respectively, which are similar to those reported by Wan et al. (2001) but higher than for other studies (i.e., Hamamoto et al., 2017). The difference in the bubble generation mechanism is considered to be responsible for the different bubble concentrations. The Sg at steady state was slightly different depending on the gas species in the bubbles [Fig. 2(b)]. It is obvious that the

of

different gas contents in the MNB solutions were responsible for the different Sg in the

ro

column. However, the reason for the higher gas content in the O3-MNB solution than

-p

other MNB solutions is not clear.

re

The Sg values for columns with NAPL (n-decane) present were higher than those measured for columns without NAPL [Fig. 2(b)]. With n-decane blobs occupying pore

lP

space, the size of the cross section available for flow was reduced, resulting in more

na

trapping of gas bubbles in the region packed with coarse glass beads. As measured by Hamamoto et al. (2017) significantly lower zeta potential for O2-MNB (note that the

Jo ur

majority of gas content in O3-MNB is oxygen) causes stronger repulsive force at the surface of the glass beads, which excludes the volume of aqueous phase available for bubble transport. The negative charge of the bubbles likely restricts bubble mobility through negatively charged media. Thus, O3-MNBs, with a lower zeta potential, were trapped to a great extent in the interstices of the porous media than air-MNBs, increasing Sg further in addition to the effect of narrowed flow paths due to n-decane ganglia.

3.2. Changes in the fluorescence emission: Oxidation reaction in the NAPL phase Since the fluorescence agent used in this study is soluble in n-decane, and loses its

14

Journal Pre-proof

fluorescence-emitting capability quickly when ozone is applied, it is a good indicator of the oxidative condition in the oil phase. No changes in the fluorescence emission were observed for the control experiment (MNB-1) [Fig. 3(a)]. Thus the water used in this study had little oxidative capability. The fluorescence emission from the columns during experiments MNB-2 and MNB-4 decreased slightly over time [Fig. 3(b) and Fig. 3(d)]. The rates of fluorescence reduction during MNB-2 and MNB-4 were approximately the

of

same (Fig. 4); about 30 % loss of fluorescence emission at the end of flushing (72 hr).

ro

The mass of n-decane flushed out of the column during Air-MNB flushing (MNB-2)

-p

is considered to be most responsible for the loss of fluorescence emission. About 15 %

re

of the n-decane was estimated to be washed out of the column based on the analysis of column effluent. Also n-decane blobs might be mobilized particularly at the inlet side of

lP

the column, due to the MNB solution flow leaving the n-decane free zone [Fig. 3(b)].

na

Air accumulation in front of the finer glass bead might force the n-decane blobs out of the column. In addition to the mass loss due to mobilization during the MNB-2,

Jo ur

hydroxyl free radicals formed possibly due to the collapse of bubbles might also affect the degradation of the fluorescence agent (Takahashi et al., 2007). Lower changes in the fluorescence emission during O3-MNB/H2O2 flushing (MNB-4) were observed than anticipated [Fig. 3(d)]. With H2O2 solution mixed with the O3-MNB solution, hydroxyl free radicals formed by the reaction between hydrogen peroxide and dissolved ozone (Peroxone process) were expected to react with the fluorescence tracer. Considering the reactivity of hydroxyl free radicals, fast decay of fluorescence was expected, which did not take place in this study. Since the radicals are very short–lived in the aqueous phase, and very reactive and hydrophilic, they may not have been able to diffuse into the n-decane phase. Thus, it is very likely that only organic molecules at the

15

Journal Pre-proof

surface of n-decane blobs reacted with the hydroxyl radicals, while most fluorescence tracer molecules dissolved in the bulk NAPL did not react with the radical molecules. The reduction of fluorescence during MNB-4 is considered to be the result of n-decane mobilization (less than 5 % n-decane flushed as estimated from column effluent analysis) and oxidative reaction occurred primarily at the surface of the n-decane phase. When the column was flushed with the O3-MNB solution (MNB-3), a faster decay of

of

fluorescence emission (loss of fluorescence over time) was observed compared to

ro

MNB-2 and MNB-4 [Fig. 3(c)]. Ozone is more soluble in n-decane than in water (Bin,

-p

2006), and thus it is likely that aqueous ozone molecules partitioned into the n-decane

re

phase reacting with fluorescence tracer. Fluorescence emission during the ozone sparging process (MNB-5) decreased at the highest rate among the experiments

lP

conducted in this study. The direct supply of ozone gas into the column resulted in high

na

Sg (0.54 ± 0.06) and likely more ozone dissolution into the n-decane phase. A rapid accumulation of ozone-containing gas in the medium was observed particularly below

Jo ur

the fine glass bead zone. The gas saturation in the coarse glass bead region below the fine glass bead zone was expected to be even higher than the average Sg (0.54) meaning more ozone mass flux to the n-decane phase resulting in faster decay in the fluorescence emission in this region. Although it was expected that the ozone sparging process would be more efficient for maintaining oxidative conditions in the NAPL than other experiments performed in this study, O3-MNB solution flushing could be a feasible remedial option based on the observation made in this study of diminishing fluorescence in the NAPL.

3.3 Oxidative mass removal of n-decane

16

Journal Pre-proof

Figure 5 shows the n-decane mass removal during the MNB solution flushing experiments (MNB–2 to MNB–4) and ozone sparging (MNB–5). Note that the mass removal was estimated based on the n-decane mass removal only by oxidative degradation, excluding the mass flushed out of the column. About 6 % (or 0.17 g) of the n-decane mass was estimated to be degraded during 72 hours of the air-MNB flushing experiment (MNB-2). Since the mass recovery of the control experiment (MNB-1) was

of

103 ± 1.1 % (not shown in Fig. 5), the mass removal ratio of MNB-2 was considered as

ro

marginal. Hydroxyl free radicals formed during the collapse of MNBs in the bulk water

-p

could be the cause of oxidative degradation during the air-MNB flushing experiment

re

(MNB–2). The generation of hydroxyl free radicals during MNB–2 might be very limited because of the pH range of the MNB solution, which was controlled at 7.0 ±

lP

0.5. Note that acidic conditions are favorable for the generation of hydroxyl free

na

radicals during MNB collapse (Takahashi et al., 2007). The n-decane mass removal percentages, 27.6% and 26.0 %, for experiments MNB–3

Jo ur

and MNB–4, respectively, were nearly identical, whereas the mass (0.86 g) removed during MNB–4 (O3-MNB and H2O2 flushing) was about 20 % higher than MNB–3. This discrepancy between the removal ratio and mass removed comes from the difference in the initial n-decane mass applied to the column. Oxidative degradation due to ozone molecules can be better evaluated by using NAPL with a low vapor pressure and low aqueous solubility (e.g., n-decane). Studies on ozone-containing MNB often used dense–NAPL (e.g., trichloroethylene) of which the high solubility and vapor pressure are mostly responsible for the mass removal during MNB flushing (Sung et al., 2017). The n-decane mass removal achieved in this study (MNB–3 and MNB–4) is considered to be significant, because it is comparable to that of MNB–5 (ozone sparging). As

17

Journal Pre-proof

already mentioned in the previous section, the oxidative reaction of ozone molecules partitioned into the n-decane phase was confirmed by the reduction of fluorescence. Thus it is reasonable to conclude that n-decane molecules reacted with ozone resulting in mass reduction. The mechanism for the n-decane mass removal during MNB-4 might be considerably different from that of MNB-3. Hydrogen peroxide molecules introduced into the inlet of

of

the column during MNB-4 reacted with ozone molecules (Peroxone process) in the

ro

aqueous phase producing hydroxyl free radicals (Mitani et al., 2002; Weiss, 1935):

-p

O3 + OH¯ → O2•¯ + •HO2

re

O2•¯ + O3 + H+ → 2O2 + •HO •HO + O3 → H+ O2•¯ + O2

lP

Due to the extremely high reactivity of hydroxyl free radical (E0 = +2.73 V, Armstrong

na

et al., 2013; Weiss, 1935), they readily react with organic compounds with high rate constants (Nimmer et al., 2000). The short lifetime of hydroxyl free radicals in the

Jo ur

aqueous phase very likely kept them from partitioning into the n-decane phase, limiting the oxidation reaction only to the air–NAPL interface. Since the rate of fluorescence emission change during MNB–4 was much lower than that of MNB–3, it is a reasonable argument that the majority of the oxidation reaction occurred at the NAPL–water interface, leaving the bulk NAPL unoxidized. The difference in the reaction characteristics between MNB-3 and MNB-4 suggests a different remedial strategy for the use of micro-nano bubbles. When NAPL droplets spread out with large specific NAPL-water interfacial area (m2/m3), the addition of hydrogen peroxide during MNB flushing could induce fast oxidative degradation of NAPL, whereas applying O3-MNB might be a better option for aquifers contaminated

18

Journal Pre-proof

with large ganglia of NAPL. Since ozone is known to partition more into the organic liquid phase than water (Bin, 2006), introducing ozone (rather than hydroxyl free radical) might be a better choice for the degradation of bulk NAPL. Based on the mass removal (1.14 g in 72 hours) of n-decane during MNB–5 (ozone sparging), the O3–MNB flushing processes (MNB–3 and MNB–4) were fairly efficient for the oxidative removal of n-decane. Even though a small amount of ozone gas (1.5 –

of

1.8 % by volume) was applied during the MNB solution flushing experiments, the n-

ro

decane removal for MNB-3 and MNB-4 was almost 60 – 75 % of that achieved by the

-p

ozone sparging process (MNB-5), which used 100 % gas flow with no liquid flow. The

re

n-decane mass removal per unit volume of ozone gas (g/mL) for MNB-3 and MNB-4 were 4.5×10-3 g/mL and 4.5×10-3 g/mL, respectively, which are much larger than that

lP

of MNB-5 (1.3×10-5 g/mL) confirming the efficiency of the MNB application.

na

Dissolved ozone in the aqueous phase during MNB flushing supplied additional ozone molecules for the oxidation reaction. Since MNB solutions are supposed to behave like

Jo ur

aqueous solutions applying ozone as an oxidizing agent for the NAPL in MNB solutions could be advantageous under many field conditions.

3.4 Changes in the chemical composition during oxidation reaction The n-decane saturations (Sn , volume of n-decane/total void volume) in the column at the end of 72 hours of MNB flushing (and sparging) experiment were estimated using a partitioning tracer test. The breakthrough curves for tracers before and after the experiments are shown in Fig. S–2 ~ Fig. S–6 (Supplementary Material). The retardation factors (Rt) of the partitioning tracer (DMP) were calculated, and are shown in the Figures. The Sn was estimated based on the Rt values and the partitioning coefficient of

19

Journal Pre-proof

n-decane (11.5, Thal et al., 2007). Figure 6 shows Sn values estimated from the tracer tests and the mass extraction analyses (section 3.3) after the flushing (or sparging) experiments were completed. For MNB–4, Sn from the tracer test is in good agreement with that from the mass extraction data. However, there was a substantial discrepancy between the values of Sn estimated for MNB–3. The mass of n-decane was overestimated by the tracer test, which

of

was due to the changes in the chemical composition in the NAPL phase. As the

ro

oxidation reaction proceeded in the NAPL phase, less hydrophobic reaction products

-p

were formed. The reduction of NAPL hydrophobicity increased the tracer (DMP)

re

partition and retardation factor, resulting in higher Sn. This observation of a changing partitioning coefficient for DMP and NAPL, due to changes in the NAPL chemical

lP

composition, agrees with a previous study (Cho et al., 2003). The effect of increased

na

oxidized chemicals in the NAPL during MNB–5 was even more significant. The Rt after the sparging was larger than the initial value (Fig. S–6 Supplementary Material). This

Jo ur

implies that the reduction of Rt value due to the decreased NAPL volume during ozone sparging was overcome by the increased partitioning due to the increased fraction of more hydrophilic oxidants in the NAPL. Figure 7 shows the GC-MS chromatogram of NAPL left after the MNB flushing experiments (MNB–3 and MNB–4). Few oxidized compounds were found for the sample (methylene chloride extract of NAPL after flushing) from MNB–4, meaning the majority of the oxidative reaction might have occurred at the NAPL-water interface. On the other hand, several oxidized compounds were found for the extract sample from MNB–3 [Fig. 7(a) and Fig. 8(a)]. Also, additional oxidized chemical species were found in the extract sample from MNB–5 [Fig. 8(b)]. It is reasonable to assume that these

20

Journal Pre-proof

hydrophilic compounds (e.g., nonanoic acid) were produced in the NAPL during MNB – 3, and some of them were washed out of the column during flushing. No aqueous flow for MNB–5 left oxidation products in the NAPL, whereas chemicals sufficiently oxidized to small molecular weights might be vaporized out of the NAPL, and from the column. This evidence indicates that the location of the redox reaction was different depending on the kind of active oxidizing agent (hydroxyl free radical, ozone) and how

of

it was applied (aqueous phase, gas phase). Analysis of pathways for the oxidative

-p

ro

reaction was not attempted here, which is considered beyond the scope of this study.

re

4. Summary

Bench-scale experiments were conducted to investigate the effect of O3-MNB

lP

flushing on the removal of n-decane (NAPL) from porous media. It was found that

na

significant removal of n-decane was achieved by applying an aqueous solution of O3MNB. More mass of n-decane was removed when hydrogen peroxide (and thus

Jo ur

hydroxyl free radicals) was applied during the O3–MNB flushing process. Although the oxidative degradation of n-decane was not as fast as that of ozone sparging at the same fluid (MNB solution, ozone gas) flow rate, MNB flushing can be an attractive option for the remediation of contaminated aquifers due to the liquid-like physical characteristics of MNB solution. Depending on the physical state of ozone applied to the media (ozone sparging, O3-MNB flushing), the location of the actual oxidative reaction was different. With hydroxyl free radical (peroxone process) the reaction took place mostly at the NAPL-water interface, whereas ozone molecules partitioned in the NAPL during ozone sparging and O3-MNB flushing reacted with organic chemicals.

21

Journal Pre-proof

Acknowledgements This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education, Science and Technology (NRF-2016R1D1A1B01014415). This work was also supported by the National Water Center at United Arab Emirates University, UAE (grant

Jo ur

na

lP

re

-p

ro

of

No. 31R112).

22

Journal Pre-proof

Literature Cited Agarwal, A., Zhou, Y., Liu, Y., 2016. Remediation of oil-contaminated sand with selfcollapsing air microbubbles. Environ. Sci. Pollut. Res. 23, 23876-23883. Alcantara-Garduno, M.E., Okuda, T., Tsai, T.Y., Nishijima, W., Okada, M., 2008. Experimental and mathematical evaluation of trichloroethylene removal from saturated soil using acetic acid with saturated ozone. Sep. Purif. Technol. 60(3), 299-

of

307.

ro

Alheshibri, M., Qian, J., Jehannin, M., Craig, V.S.J., 2016. A history of nanobubbles.

-p

Langmuir 32, 11086-11100.

re

Armstrong, D.A., Huie, R.E., Lymar, S., Koppenol, W.H., Merenyi, G., Neta, P., Stanbury, D.M., Steenken, S., Wardman, P., 2013. Standard electrode potentials

na

4), 59-61.

lP

involving radicals in aqueous solution: inorganic radicals. Bioinorg. React. Mech. 9(1-

Bin, A.K., 2006. Ozone solubility in liquids. Ozone-Sci. Eng. 28, 67-75.

Jo ur

Cho, J., Annable, M.D., Rao, P.S.C., 2003. Residual alcohol influence on NAPL saturation estimates based on partitioning tracers. Environ. Sci. Technol. 37, 16391644.

Choi, Y.J., Park, J.Y., Kim, Y.-J., Nam, K., 2008. Flow characteristics of microbubbles suspensions in porous media as an oxygen carrier. Clean-Soil Air Water 36(1), 59-65. Hamamoto, S., Nihei, N., Ueda, Y., Moldrup, P., Nishimura, T., 2017. Effects of flow rate and gas species on microbubble and nanobubble transport in porous media. J. Environ. Eng. 143(7), 04017023-1~04017023-7. Hamamoto, S., Takemura, T., Suzuki, K., Nishimura, T., 2018. Effects of pH on nanobubble stability and transport in saturated porous media. J. Contam. Hydrol. 208, 61-

23

Journal Pre-proof

67. Hu. L., Xia, Z., 2018. Application of ozone micro-nano-bubbles to groundwater remediation. J. Hazard. Mater. 342, 446-453. Ikeura, H., Kobayashi, F., Tamaki, M., 2011. Removal of residual pesticides in vegetables using ozone microbubbles. J. Hazard. Mater. 186(1), 956-959. Jin M., Delshad, M., Dwarakanath, V., McKinney, D.C., Pope, G.A., Sepehrnoori, K.,

of

Tilburg, C.E., 1995. Partitioning tracer test for detection, estimation, and remediation

ro

performance assessment of subsurface nonaqueous phase liquids. Water Resour. Res.

-p

31(5), 1201-1211.

re

Kao, C.M., Huang, K.D., Wang, J.Y., Chen, T.Y., Chien, H.Y., 2008. Application of potassium permanganate as an oxidant for in situ oxidation of trichloroethylene-

lP

contaminated groundwater: A laboratory and kinetic study. J. Hazard. Mater. 153(3),

na

919-927.

Khuntia, S., Majumder, S.K., Ghosh, P., 2012. Microbubbles-aided water and

Jo ur

wastewater purification: a review. Rev. Chem. Eng. 28, 191-221. Kim, H., Soh, H.–E., Annable, M.D., Kim, D.-J., 2004. Surfactant-enhanced air sparging in saturated sand. Environ. Sci. Technol. 38(4), 1170-1175. Kim, H., Yang, S., Yang, H., 2013. Surfactant-enhanced ozone sparging for removal of organic compounds from sand. J. Environ. Sci. Heal. Part A. 48(5), 526-533. Krembs, F.J., Siegrist, R.L., Crimi, M.L., Furrer, R. F., Petri, B.G., 2010. ISCO for groundwater remediation: Analysis of field applications and performance. Ground Water Monit. Remediat. 30(4), 42-53. Kwan, J.J., Borden, M.A., 2010. Microbubbles dissolution in a multigas environment. Langmuir 26(9), 6542-6548.

24

Journal Pre-proof

Li, H., Hu, L., Song, D., Al-Tabbaa, A., 2014. Subsurface transport behavior of micronano bubbles and potential applications for groundwater remediation, Int. J. Environ. Res. Public Health 11, 473-486. Li, P., Tsuge, H., Itoh, K., 2009. Oxidation of dimethyl sulfoxide in aqueous solutioin using microbubbles. Ind. Eng. Chem. Res. 48(17), 8048-8053. Liu, S., Kawagoe, Y., Makino, Y., Oshita, S., 2013. Effects of nanobubbles on the

of

physicaochemical properties of water: The basis for peculiar properties of water

ro

containing nanobubbles. Chem. Eng. Sci. 93, 250-256.

-p

McCray, J.E., Falta, R.L., 1996. Defining the air sparging radius of influence for

re

groundwater remediation. J. Contam. Hydrol. 24(1), 25-52. Mitani, M.M., Keller, A.A., Bunton, C.A., Rinker, R.G., Sandall, O.C., 2002. Kinetics

lP

and products of reactions of MTBE with ozone and ozone/hydrogen peroxide in

na

water. J. Hazard. Mater. B89, 197-212.

Nimmer, M.A., Wayner, B.D., Morr, A.A., 2000. In-situ ozonation of contaminated

Jo ur

groundwater. Environ. Prog. 19(3), 183-196. Scheffer, B., van de Ven, E. 2010. Experiences of perozone and c-sparge at two former dry cleaner sites in the Netherlands. Ozone-Sci. Eng., 32(2), 130-136. Sung, M., Teng, C.-H., Yang, T.-H., 2017. Dissolution enhancement and mathematical modeling of removal of residual trichloroethene in sands by ozonation during flushing with micro-nano-bubble solution. J. Contam. Hydrol. 202, 1-10. Takahashi, M., Chiba, K., Li, P., 2007. Free-radical generation from collapsing microbubbles in the absence of a dynamic stimulus. J. Phys. Chem. B 111(6), 13431347. Takahashi, M., Ishikawa, H., Asano, T., Horibe, H., 2012. Effects of microbubbles on

25

Journal Pre-proof

ozonized water for photoresist removal. J. Phys. Chem. C 116(23), 12578-12583. Tawabini, B.S., 2014. Simultaneous removal of MTBE and benzene from contaminated groundwater using ultraviolet-based ozone and hydrogen peroxide. Int. J. Photoenergy 2014, 1-7. Thal, Jr., A.E., Knox, R.C., Sabatini, D.A., 2007. Estimating partitioning coefficients of tracers. Ground Water Monit. Remediat. 27(4), 135-142.

of

Tomlinson, D.W., Thomson, N.R., Johnson, R.L., Redman, J.D., 2003. Air distribution

ro

in the Borden sand aquifer during in situ air sparging. J. Contam. Hydrol. 67(1-4),

-p

113-132.

re

Tsai, Y.J., 2007. Air flow paths and the porosity/permeability change in a saturated zone during in situ air sparging. J. Hazard. Mater. 142, 315-323.

lP

Ushikubo, F.Y., Furukawa, T., Nakagawa, R., Enari, M., Makino, Y., Kawagoe, Y.,

na

Shiina, T., Oshita, S., 2010. Evidence of the existence and the stability of nanobubbles in water. Coll. Surf. A 361, 31-37.

Jo ur

Wan, J., Veerapaneni, S., Gadelle, F., Tokunaga, T.K., 2001. Generation of stable microbubbles and their transport through porous media. Water Resour. Res. 37(5), 1173-1182.

Watts, R.J., Teel, A.L., 2005. Chemistry of modified Fenton’s reagent (catalyzed H2O2 propagation-CHP) for in situ soil and groundwater remediation. J. Environ. Eng. 131(4), 612-622. Weiss, J., 1935. Investigation on the radical HO2 in solution. Trans. Faraday Soc. 31, 668-681. Yasuda, K., Ban, N., 2012. Wastewater treatment for bioethanol production system using ozone microbubbles. J. Chem. Eng. Jpn. 45(9), 672-677.

26

Journal Pre-proof

Figures

Figure 1. Schematic diagram of experimental set up

Figure 2. Changes in gas saturation (Sg) during flushing experiments: (a) changes in gas saturations over time, MNBs were generated using air (triangle), pure oxygen

of

(rectangle), and ozone gas (circle), (b) maximum gas saturations achieved for different

ro

MNB solutions for the glass bead column with and without n-decane applied; vertical

re

-p

lines represent standard deviations

Figure 3. Changes in fluorescence emission during flushing experiments: (a) control

lP

experiment (MNB-1), tap water flushing, (b) flushing with air- MNB solution (MNB-2),

na

(c) flushing with ozone-MNB solution (MNB-3), (d) flushing with ozone-MNB and

Jo ur

H2O2 solutions (MNB-4), (e) ozone sparging (MNB-5)

Figure 4. Estimated percent of fluorescence loss in the column during the flushing experiments

Figure 5. Estimated NAPL (n-decane) removal after 72 hours of flushing (sparging) experiments; vertical lines represent the range of data from duplicated experiments

Figure 6. Residual NAPL saturations (Sn) estimated after 72 hours of flushing (sparging) using tracer tests and analysis of methylene chloride extract of the media; vertical lines

27

Journal Pre-proof

represent the range of data from duplicated experiments

Figure 7. GC-MS chromatogram for the methylene chloride extracts of the column media (MNB-3 and MNB-4)

Figure 8. Parts of the GC-MS chromatograms for the methylene chloride extracts of the

Jo ur

na

lP

re

-p

ro

of

column media (MNB-4 and MNB-5)

28

Journal Pre-proof

Tables Table 1. Conditions for the MNB flushing experiments Experiment code MNB-1

Description

Source of MNB

Volume n-decane (mL)

Saturation n-decane

Conc of H2O2 (g/L)

H2O2 solution Flowratea (mL/min)

MNB flowratea (mL/min)

Specific discharge (cm/min)

control (tap water flushing)

-

4.1

0.13

-

-

2.0 ± 0.1b

0.39 ± 0.02

air

4.0

0.12

-

O3/O2

3.8

0.12

O3/O2

4.8

0.15

-

3.3

MNB-2 MNB-3

MNB solution flushing

MNB-4 MNB-5

ozone sparging

a

n r u

l a

including upper and lower limits of flow rates b flow rate of tap water c flow rate of ozone-containing gas d based on total flow rate including H2O2 and MNB solutions e specific discharge of ozone-containing gas

o J

r P

0.11

o r p

e -

f o -

0.39 ± 0.02 2.0 ± 0.1

4.06 ± 0.03

0.50 ± 0.02

-

-

0.49 ± 0.02d 2.0 ± 0.2c

0.39 ± 0.02e

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure 1

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure 2

31

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure 3

32

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure 4

33

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure 5

34

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure 6

35

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure 7

36

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure 8

37

of

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

Supplementary Material

38

Journal Pre-proof

Jo ur

na

lP

re

-p

ro

of

Figure S-1. Picture of micro-nano bubbles generated in this study

39

Journal Pre-proof

Figure S-2. Tracer breakthrough curves (BTCs) for MNB-1; Br (bromide) non-reactive

Jo ur

na

lP

re

-p

ro

of

tracer, dimethylpentanol (DMP) reactive (partitioning) tracer; (a) BTCs before tap-water flushing, (b) BTCs after tap-water flushing

40

Journal Pre-proof

Figure S-3. Tracer breakthrough curves (BTCs) for MNB-2; Br (bromide) non-reactive

Jo ur

na

lP

re

-p

ro

of

tracer, dimethylpentanol (DMP) reactive (partitioning) tracer; (a) BTCs before MNB (air) solution flushing, (b) BTCs after MNB (air) solution flushing

41

Journal Pre-proof

Figure S-4. Tracer breakthrough curves (BTCs) for MNB-3; Br (bromide) non-reactive

Jo ur

na

lP

re

-p

ro

of

tracer, dimethylpentanol (DMP) reactive (partitioning) tracer; (a) BTCs before MNB (O3/O2) solution flushing, (b) BTCs after MNB (O3/O2) solution flushing

42

Journal Pre-proof

Figure S-5. Tracer breakthrough curves (BTCs) for MNB-4; Br (bromide) non-reactive tracer, dimethylpentanol (DMP) reactive (partitioning) tracer; (a) BTCs before MNB (O3/O2) and H2O2 solution flushing, (b) BTCs after MNB (O3/O2) and H2O2 solution

Jo ur

na

lP

re

-p

ro

of

flushing

43

Journal Pre-proof

Figure S-6. Tracer breakthrough curves (BTCs) for MNB-5; Br (bromide) non-reactive

Jo ur

na

lP

re

-p

ro

of

tracer, dimethylpentanol (DMP) reactive (partitioning) tracer; (a) BTCs before ozone sparging, (b) BTCs after ozone sparging

44

Journal Pre-proof

Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be

Jo ur

na

lP

re

-p

ro

of

considered as potential competing interests:

45

Journal Pre-proof

Highlights:  Quantitative evaluation of NAPL removal using O3-containing MNB solutions  Significant mass removal of n-decane using 1-D column packed with glass beads during O3-containing MNB solutions  More effective O3-containing MNB solutions for NAPL removal when H2O2 solution co-injected

of

 Different oxidative reaction sites during MNB flushing confirmed by chemical

Jo ur

na

lP

re

-p

ro

composition

46