WO2018039204A1

WO2018039204A1 - Method of removing oil from water

Info

Publication number: WO2018039204A1
Application number: PCT/US2017/047969
Authority: WO
Inventors: Ahmed Kayvani FARD; Muataz A. HUSSEIN; Tarik RHADFI; Marwan K. Khraisheh
Original assignee: Qatar Foundation For Education, Science And Community Development
Priority date: 2016-08-24
Filing date: 2017-08-22
Publication date: 2018-03-01

Abstract

The method for removing oil from water includes contacting modified carbon nanotubes with the water to adsorb oil from the water, and then isolating the modified carbon nanotubes with the oil adsorbed thereon. The carbon nanotubes can be doped with a metal, such as iron, silver, alumi/num, zinc or copper, or an oxide of any of these metals, using thermal treatment. In order to make iron oxide-modified carbon nanotubes for removing oil from water, ferric nitrate and iron oxide are dissolved in ethanol to form a first solution. Carbon nanotubes are dissolved in ethanol to form a second solution. The first and second solutions are then sonicated separately, followed by mixing the first solution and the second solution to form a third solution, which is then sonicated to form a sonicated mixture, and the sonicated mixture is heated to provide a solid residue, which is then calcined.

Description

METHOD OF REMOVING OIL FROM WATER

TECHNICAL FIELD

The disclosure of the present patent application relates to the purification of water, such as industrial wastewater, groundwater and the like, and particularly to a method of removing oil from water using modified carbon nanotubes (CNTs).

BACKGROUND ART

The significance of energy generated from nonrenewable sources such as oil and natural gas in the modern era is well known. While this type of energy production has many advantages, it is associated with huge volumes of liquid waste. For example, industrial wastewater and produced/co-produced water contain a large quantity of emulsified and non- dissolved oil. Discharging such waste can pollute surface and underground water, including soil and seawater. Based on the United States Environmental Protection Agency (USEPA) standards for treated produced water discharge to offshore, the maximum daily limit for oil and grease in groundwater is 42 mg/L, and the monthly average limit is 29 mg/L. Qatar and several other water-stressed countries generate huge volumes of oil-contaminated water due to their oil and gas industries. Thus, these countries are in particular need of efficient and cost-effective treatment methods to remove pollutants, such as oil, from groundwater as a way to supplement their limited fresh water resources.

When oil mixes with the water, it forms an oil- water emulsion (or "floating film") that should be removed before it is discharged into the environment. Due to the generation of produced water, as well as oil-contaminated water, by the oil industries, many countries are focusing on efforts to find cost-effective and efficient treatment technologies to remove emulsified oil particles from water as a way to augment their limited fresh water resources. Such water, if treated to meet the environmental limits and regulations, can be used for aquifer recharge, irrigation, livestock or wildlife watering and habitats, as well as industrial applications (e.g., vehicle washing, power plant cooling water and fire control).

Numerous techniques are known for treating oil-contaminated water. Examples of such techniques include reverse osmosis, ultra-filtration and micro-filtration, a variety of flotation methods (e.g., dissolved air, column flotation, electro- and induced air), adsorption, gravity separation, activated sludge treatment, membrane bioreactors, biological treatment, chemical coagulation, electro-coagulation and coalescence. Adsorption/sorption is believed to be one of the most promising processes for the removal of oil from water due to its low operational and capital cost, in addition to its high removal efficiency. A variety of materials are known to be oil de-emulsifiers for oil-water treatment, such as natural sorbents, organic polymers (synthetic) and mineral materials (inorganic). Thus, a method of removing oil from water solving the aforementioned problems is desired.

DISCLOSURE OF INVENTION

The method of removing oil from water, such as wastewater, groundwater, produced/co-produced water and the like, includes contacting modified carbon nanotubes (CNTs) with the water to adsorb oil from the water, and then isolating the modified carbon nanotubes with the oil adsorbed thereon. The modified carbon nanotubes may include carbon nanotubes doped with a metal, such as iron, silver, aluminum, zinc or copper, or oxides thereof. For example, the carbon nanotubes can be doped with iron oxide (Fe₂0₃), also known as hematite. The modified carbon nanotubes preferably have an outer diameter of about 10 nm to about 20 nm and a length between about 1 μιη and about 10 μιη.

These and other features of the present disclosure will become readily apparent upon further review of the following specification and drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a schematic diagram of the process of loading hematite (a-Fe20s) nanoparticles on a carbon nanotube (CNT) support.

Fig. 2 A is a scanning electron microscope (SEM) micrograph of undoped CNTs.

Figs. 2B, 2C, 2D, and 2E are SEM micrographs of CNTs doped with 1 wt%, 10 wt%, 30 wt%, and 50% Fe2C>3 nanoparticles, respectively.

Fig. 3A is a transmission electron microscope (TEM) image of unmodified CNTs.

Figs. 3B, 3C, 3D, and 3E are TEM micrographs of CNTs doped with 1 wt%, 10 wt%, 30 wt%, and 50% Fe₂0₃ nanoparticles, respectively.

Fig. 4 are diffractograms comparing the powder X-ray diffraction (XRD) patterns of a-Fe₂0₃/CNT composites at 1, 10, 30, and 50 wt% Fe₂0₃ nanoparticle loading, pure a-Fe₂0₃, and pure CNTs.

Fig. 5 are thermogravimetric analysis (TGA) plots of pure CNTs and a-Fe₂0₃/CNT nanocomposites having 1 wt%, 10 wt%, 30 wt% and 50 wt% Fe loadings. Fig. 6 is a plot of N2 adsorption-desorption isotherms of CNTs and CNT/a-Fe₂0₃ composites having 1 wt%, 10 wt%, 30 wt% and 50 wt% Fe loadings_.

Fig. 7 is a plot of percentage removal of oil as a function of contact time comparing pure CNTs with CNTs loaded with 1, 10, 30, and 50 wt% Fe₂0₃.

Fig. 8 is a plot of percentage removal of oil as a function of adsorbent dosage comparing pure CNTs with CNTs loaded with 1, 10, 30, and 50 wt% Fe₂0₃.

Fig. 9 is a graph showing the Freundlich adsorption isotherms of oil adsorbed by pure CNTs and CNTs loaded with 1, 10, 30, and 50 wt% Fe₂0₃.

Fig. 10 is a plot showing the Temkin adsorption isotherms of oil adsorbed by pure CNTs and CNTs loaded with 1, 10, 30, and 50 wt% Fe₂0₃.

Fig. 11 is a plot showing the pseudo-first-order kinetics for the adsorption of oil by pure CNTs and CNTs loaded with 1, 10, 30, and 50 wt% Fe₂0₃.

Fig. 12 is a plot showing the pseudo-second-order kinetics for the adsorption of oil by pure CNTs and CNTs loaded with 1, 10, 30, and 50 wt% Fe₂0₃.

Fig. 13 is a plot showing the intra-particle diffusion model kinetics for the adsorption of oil by pure CNTs and CNTs loaded with 1, 10, 30, and 50 wt% Fe₂0₃.

Similar reference characters denote corresponding features consistently throughout the attached drawings.

BEST MODES FOR CARRYING OUT THE INVENTION The method for removing oil from water, such as wastewater, groundwater, produced/co-produced water and the like, includes contacting modified carbon nanotubes (CNTs) with the water to adsorb oil from the water, and then isolating the modified carbon nanotubes with the oil adsorbed thereon. The carbon nanotubes are doped with a metal, such as iron, silver, aluminum, zinc or copper, or an oxide of any of these metals, using thermal treatment. The modified carbon nanotubes preferably have an outer diameter of about 10 nm to about 20 nm and a length between about 1 μιη and about 10 μιη.

Carbon nanotubes (CNTs) are molecules of pure carbon that are relatively long and thin, and which are shaped like tubes. The modified carbon nanotubes for removing oil from water include carbon nanotubes doped with a metal or metal oxide. For example, the carbon nanotubes can be doped with iron oxide (Fe₂0₃), i.e., hematite. The method for removing oil from water can include the step of placing the carbon nanotubes impregnated with iron oxide into contact with wastewater to adsorb the oil, followed by isolating the carbon nanotubes with the adsorbed oil.

The iron oxide-modified carbon nanotubes may be fabricated by impregnating carbon nanotubes (CNTs) with iron oxide Fe2C>3 (hematite). The iron oxide-impregnated carbon nanotubes (Fe20₃/CNTs) possess an improved capacity to remove oil when compared against conventional oil adsorbents. In order to make iron oxide-modified carbon nanotubes for removing oil from water, ferric nitrate and iron oxide are dissolved in ethanol to form a first solution. Carbon nanotubes (CNTs) are dissolved in ethanol to form a second solution. The first and second solutions are then sonicated separately, followed by mixing the first solution and the second solution to form a third solution. The third solution is sonicated to form a sonicated mixture, and the sonicated mixture is heated to provide a solid residue, which is then calcined. The sonicated mixture can be heated in a furnace at a temperature ranging from about 60°C to about 90°C for a time period sufficient to evaporate the ethanol. The solid residue can be calcined at about 350°C for about 4 hours to impregnate the iron oxide into the carbon nanotubes (CNTs). The concentration of iron oxide used may be typically between 1% and 50%. The sonication may be ultra-sonication and may be conducted for about forty-five minutes. The Fe20₃-modified carbon nanotubes may include single-walled and/or multi-walled carbon nanotubes. Fig. 1 schematically illustrates the synthetic process used to produce CNTs impregnated with hematite (Fe2C>3).

Experiments using the present modified CNTs resulted in successful removal of almost 100% oil from water with very high oil content. The doped CNTs exhibited significantly enhanced removal efficiency of oil from water when compared against conventional oil adsorbents. The CNTs may be doped with different metals or metal oxides using thermal treatment. The CNTs may be doped with iron oxide or other metals or metal oxides to remove oil and other oil-derivatives from water. Suitable metals that may be used with varying weight percentages include iron (Fe), silver (Ag), aluminum (Al), zinc (Zn), and copper (Cu). In some cases, the removal efficiency of modified CNTs can be at or near 100%. Use of modified carbon nanotubes can provide a very effective and economical solution for removing oil from water.

The CNTs used in the examples described herein were purchased from the Chengdu

Organic Chemicals Co. Ltd. of China, with a purity of >95%, lengths of 1-10 μιη, outer diameters of 10-20 nm, and BET surface areas of 156 m²/ g. Liquid ethanol (98% purity), used as a solvent, and ferric nitrate (Fe(N0₃)3-9H₂0) (99 % purity), used as a precursor of iron nanoparticles, were obtained from the Sigma-Aldrich, Inc. of St. Louis, Missouri. Liquid gasoline, used as an oil source, was purchased in Doha, Qatar, with an octane number of 97. All of the chemicals were used without further purifications.

Example 1

Preparation of Modified Fe₂0₃/CNT Modification of the surface of the CNTs by iron oxide (hematite) nanoparticles was carried out using the wet impregnation method. About 1.443 g of Ferric (III) Nitrate, [Fe(N0₃)3-9H₂0], was dissolved in 500 ml of ethanol for 1% of iron oxide. Separately, 19.88 g of CNTs was also dissolved in 400 ml of absolute ethanol. Both solutions were sonicated for 45 minutes separately and then mixed together to form a mixture. The mixture was once again sonicated for one hour at room temperature. The ethanol was evaporated at about 70°C in an oven. The residue was calcined for four hours at 350°C in a furnace to prepare CNTs impregnated with 1% iron oxide. Similarly, CNTs were doped with different loadings of iron oxides by weight per cent, i.e., 1%, 10%, 30%, 50%, etc. The product obtained had a purity of about 95%. The modified carbon nanotubes had an outer diameter of about 10 nm to about 20 nm and a length of about 1 μιη to about 10 μιη. Fig. 1 further illustrates the process by which CNTs loaded with hematite are produced.

The iron oxide-modified carbon nanotubes (CNTs) were characterized as follows. Powder X-ray diffraction (XRD) patterns were recorded by using a MiniFlex-600 X-Ray diffractometer, manufactured by Rigaku^®/USA, Inc. of Massachusetts, with Cu Ka radiation λ = 1.54A at a rate of 0.4% over Bragg angles ranging from 10-90°. The operating voltage and current were maintained at 40 KV and 15 mA, respectively. The morphologies of the samples were analyzed with a TESCAN MIRA 3 FEG-SEM field emission scanning electron microscope, manufactured by Tescan of the Czech Republic, using an acceleration voltage of 20 kV. The Zeta potentials for a suspension of 50% of CNTs/Fe in deionized water (DI) solution were determined by a Zetasizer Nano ZS, manufactured by Malvern Instruments, Ltd. of the United Kingdom. Elemental analysis (Fe, C and O) was performed with energy- dispersive X-ray spectroscopy (EDX) analysis. The surface areas were measured by N2 adsorption at 77 K using a 15 -point BET technique on the analysis port of the analyzer, manufactured by Micromeritics^® of Georgia, to determine the amount of a-Fe20₃ present on the CNT surface. The thermogravimetric analyses (TGA) were performed with a TGA analyzer at a heating rate of 10°C/min in air. Before conducting any experiment, the TGA analyzer was calibrated for temperature readings and mass changes using nickel reference material. The gasoline concentration was measured using a combustion type total organic carbon (TOC) analyzer (model TOC-L, manufactured by the Shimadzu Corporation of Japan), with a detection range of 4 μg/L to 30,000 mg/L.

Example 2

Batch Experimental Studies The effects of adsorption parameters on the adsorption capacity of oil removal from water by undoped and doped CNTs with different loadings of Fe20₃ nanoparticles were investigated using batch mode experiments. In these experiments, the effects of adsorbent dosage, initial oil concentration, and contact time on the percentage removal and adsorption capacity of emulsified oil from water were studied. The adsorption capacity and removal efficiency were calculated as follows:

(Ci - C_f) x V

Q =— — (1)

(c, - c_f)

RE (%) = ^TJ X 100 , (2) where C (mg/L) is the initial concentration of emulsified oil in the water, Cf (mg/L) is the final concentration of the remaining oil in the water, V (L) is the volume of the water, and W_g is the mass of the CNTs.

All of the batch experiments were performed by adding 20 mL of aqueous solution containing specific concentrations of oil to 20 mg of undoped and doped CNTs. The experiments were performed at room temperature. The effect of the oil concentrations on the adsorption capacity was examined by varying the initial oil concentrations from 425 to 7460 mg/L. The equilibrium studies were carried out for 120 min to determine the equilibrium contact time and capacity. To find the optimum amount of adsorbent dosages for maximum removal of emulsified oil from water, different weights of undoped and doped CNTs were varied from 5 to 35 mg. All of the samples were agitated using a mechanical shaker at 400 rpm. Example 3

Adsorption Kinetic Studies

The aqueous solutions (20 ml) with 20 mg of undoped and doped CNTs were agitated at 400 rpm using a mechanical shaker at 27 °C. The samples (1.2 ml) were taken from the solution at each preset time interval, and the final concentrations of oil were analyzed using a TOC analyzer. In order to find the maximum oil uptakes by CNTs and to interpret the experimental data acquired, three kinetic models (pseudo-first-order, pseudo-second-order and intra-particle diffusion) were used. A linear fitting procedure was performed using software manufactured by the OriginLab Corporation of Massachusetts.

The Lagergren pseudo-first-order model proposes that the rate of sorption is proportional to the number of sites unoccupied by the adsorbate. The pseudo-first-order equation can be written in linearized form as follows:

\n(Q_e - Q_t) = \n Q_e - k₁t (₃) where Q_t is the sorption capacity (mg/g) at any preset time interval (i) and k₁ is the first- order rate constant (min^-1). A graph of ln(Q_e— Q_t) as a function of time is plotted and the constant is found. Additionally, the adsorption data were analyzed using the pseudo-second- order kinetic mode. The pseudo-second-order kinetic model can be written in linearized form as follows:

where k₂ is the second-order rate constant (g/mg- min). By plotting tlQ_t as a function of time, straight lines were obtained and the constants were found.

In order to gain insight into the mechanisms and rate controlling steps affecting the kinetics of adsorption, the kinetic experimental results were fitted to the Weber-Morris intra- particle diffusion (IPD) model, which is commonly expressed using the following equation:

Q_t = k_ttl + C , (5) where k is the intra-particle diffusion rate constant (mg/g- min) and C (mg/g) is the intercept. The larger the intercept, the greater the contribution of the surface sorption in the rate- controlling step. If the graph of Q_t vs. t¹^² is linear and passes through the origin, intra- particle diffusion is the sole rate-limiting step for the adsorption process.

The maximum adsorption capacity and the sorption energy of oil on CNTs were analyzed using the Freundlich, Langmuir, and Temkin isotherm models. The Freundlich isotherm can be expressed as:

1

Qe = K_FCg , (6) where C_e is the adsorbate concentration at equilibrium (mg/L), Q_e is the adsorption capacity at equilibrium, and n and K_F are Freundlich empirical constants that indicate the sorption intensity and the sorption capacity of the adsorbent, respectively. The above equation can be linearized and expressed as:

1

ln <?_e = ln i¾r + - ln C_e . (7)

n

The Freundlich model does not consider the sorption saturation, as it assumes a heterogeneous adsorbent surface and an energy distribution for the different sites. On the other hand, the Langmuir isotherm model assumes that the adsorption takes place at defined homogeneous sites on the surface of the adsorbent. The Langmuir isotherm is expressed by:

X_mKC_e

^{Qe =} (1 + KC_e) ' ⁽⁸⁾ where C_e is the equilibrium concentration of adsorbate in aqueous solution (mg/g), Q_e is the adsorption density at the equilibrium solution concentration C_e, K is the Langmuir constant, and X_m is the maximum oil adsorption capacity (mg/g).

The Langmuir isotherm model assumes that when adsorption takes place at defined homogeneous sites on the surface of the adsorbent, suggesting that once an oil molecule occupies a site, no further adsorption can take place at that occupied site. The above equation can be rearranged to give the linearized equation expressed as:

— =— +— . (9)

The Langmuir constants K and X_m (which are related to the constant free energy of sorption) can be determined by representing a linear plot of CJQ_e vs. C_e from the intercept and slope of the plot.

The Tempkin isotherm assumes that the sorption energy during the sorption process decreases linearly with increasing sorption site saturation rather than decreasing exponentially, as implied by the Freundlich isotherm. The Temkin isotherm is given as:

The Temkin isotherm can be linearized as follows: Q_e = B \n A + B \n C_e , (11) where B = RT/b, b is the Temkin constant related to the sorption energy, A is the equilibrium binding constant (L/mg) and B is a constant related to the heat of sorption. The linearized isotherm coefficients were estimated using graphical methods by plotting Q_e vs. In C_e and are reported below.

Example 4

Characterization of CNT and Fe-doped CNT

The undoped and doped CNTs were characterized using field effect scanning electron microscopy (FE-SEM), high resolution Transmission Electron Microscopy (HR-TEM), Thermogravimetry (TGA) techniques, XRD, BET surface area and Zeta potential. The surface morphologies of the undoped and doped CNTs adsorbents were observed using FE- SEM. Fig. 2A is a SEM micrograph of undoped CNTs, which is used as a control. Fig. 2B is a SEM micrograph of CNTs doped with 1 wt% Fe₂C>3 nanoparticles. Fig. 2C is a SEM micrograph of CNTs doped with 10 wt% Fe₂03 nanoparticles. Fig. 2D is a SEM micrograph of CNTs doped with 30 wt% Fe₂03 nanoparticles. Fig. 2E is a SEM micrograph of CNTs doped with 50 wt% Fe₂C>3 nanoparticles. The diameter of the CNTs varied from 20 nm to 40 nm, the average diameter being 24 nm. It can be observed that there are no changes on the surfaces of doped CNTs after the doping, i.e., they are agglomerated and untangled, resembling a cotton-like structure.

High Resolution Transmission Electron Microscopy (HR-TEM) was performed to characterize the structures, sizes and the purity of undoped and doped carbon nanotubes with iron oxide nanoparticles. The TEM micrograph of the unmodified nanotubes is shown in Fig. 3A. Fig. 3B is a TEM micrograph of CNTs doped with 1 wt% Fe₂C>3 nanoparticles. Fig. 3C is a TEM micrograph of CNTs doped with 10 wt% Fe₂C>3 nanoparticles. Fig. 3D is a TEM micrograph of CNTs doped with 30 wt% Fe₂C>3 nanoparticles. Fig. 3E is a TEM micrograph of CNTs doped with 50 wt% Fe₂C>3 nanoparticles. The TEM micrographs show that a highly ordered crystalline structure of CNTs exists. The clear fringes of graphitic sheets are well separated by 0.34 nm and aligned with a tilted angle of about 2° toward the tube axis. The TEM images of CNTs doped with Fe₂C>3 nanoparticles were taken in order to verify the presence of nanoparticle ions on the surfaces of the CNTs (as shown in Figs. 3B and 3C). The distribution and agglomeration of Fe₂C>3 nanoparticles were also investigated. It was observed that there are formations of white crystal structures of Fe₂03 nanoparticles with small sizes and irregular shapes. It can be seen that the Fe20₃ nanoparticles are spread widely on the surfaces of the carbon nanotubes, forming very small crystal particles with diameters varying from 1 nm to 5 nm. When the ratio is increased to 30% and 50%, the size of the Fe₂03 nanoparticles is about 50 nm and agglomerate intensively, as seen in Figs. 3D and 3E, respectively. The doping of Fe20₃ nanoparticles on the surfaces of CNTs was also confirmed by energy dispersive X-ray spectroscopy (EDX) and X-ray diffraction (XRD).

The EDX analysis of the undoped and doped CNTs, which represents the atomic weight percentage (%) of the elements, such as Fe, O and C, with different percentage of a- Fe₂C>3, is shown in Table 1 below. The Fe/C ratios extracted from the EDX are close to the calculated values of the prepared samples.

Table 1: EDX Analysis of CNTs and Modified CNTs

CNT Raw CNTs CNT- CNT- CNT- CNT-

Sample (wt%) Fe₂0₃ Fe₂0₃ Fe₂0₃ Fe₂0₃

(theo) (1%) (10%) (30%) (50%)

Element (wt%) (wt%) (wt%) (wt%)

C 98.50 94.47 84.53 45.9 37.66

O 1.50 3.65 4.5 23.53 18.60

Fe 0 1.50 (1) 11.03 (10) 28.08 (30) 43.74 (50)

Total % 100 100 100 100 100

Fig. 4 shows the X-ray diffraction patterns of undoped and doped CNTs and pure a- Fe2C>3. It has been observed that the XRD diffraction pattern of pure a-Fe20₃ is similar to doped Fe2C>3 nanoparticles, confirming the presence of a-Fe20₃ crystal nanoparticles on the surfaces of the CNTs. There is one characteristic peak of CNTs, which can be seen at 2Θ of 27, while other characteristic peaks were found at 2Θ values of 34.36, 42, 50, 54, 63, 65, 72 and 75, which correspond to a-Fe₂03. These results revealed that the a-Fe₂03 particles were successfully attached to the CNTs.

The thermal oxidation and degradation of the materials is an important factor, as it determines the upper temperature limit. Thermogravimetric analysis (TGA) was conducted to study the oxidation/combustion profile of CNTs and Fe2C>3 with different theoretical mole ratios of Fe on the CNTs (0%, 1%, 10 %, 20%, 30 % and 50%). The TGA profiles of CNTs and modified CNTs with a-Fe20₃ at a heating rate of 10°C/min in air at temperatures ranging from 40°C to 900°C are presented in Fig. 5. The graphs illustrate that CNTs completely oxidize in air at temperatures above 650°C, but the thermal stability of a-Fe20₃/CNTs weakened and they began to decompose at 550°C. These results showed that the initial oxidation temperature of undoped CNTs under air starts approximately at 580°C and then reaches a complete oxidation at 670°C. The initial temperature of doped CNTs with 1 wt%, 10 wt%, 30 wt% and 50 wt% reduces to 520°C, 480°C, 460°C, and 450°C, respectively, and the final oxidation temperatures reduce to 620°C, 600°C, 550°C and 500°C, respectively. It has been found that the loading of Fe20₃ nanoparticles doped on CNTs acts as a heating accelerator agent, which accelerated the heat transfer to the body of the CNTs, as can be seen by the faster combustion of the doped CNTs (i.e., oxidization) compared to undoped CNTs. In addition to the measurement of the thermal stability of CNTs, the TGA provides an accurate estimate of the loading of Fe2C>3 nanoparticles doped on CNTs by comparing the residue of the complete oxidation of doped and undoped CNTs. The final remaining residual of undoped CNTs is 0.99 wt%, and the final remaining residuals of CNTs doped with 1 wt%, 10 wt%, 30 wt% and 50 wt% Fe₂0₃ are 2.05 wt%, 8.8 wt%, 32 wt%, and 48 wt%, respectively.

BET surface area analysis was conducted to measure the surface area of undoped and doped CNTs. The interpretation of the BET results was based on the adsorption-desorption of liquid N2 at 77 K, as shown in Fig. 6. The BET surface area values obtained for the undoped and doped CNTs with 1 wt%, 10 wt%, 30 wt% and 50 wt% Fe2C>3 nanoparticles were 137.7 m²/g, 226.6 m²/g, 295.4 m²/g, 128 m²/g, and 74.86 m²/g, respectively, as shown in Fig. 6. This indicates that the iron oxide nanoparticles at 1 wt% and 10 wt% on surfaces of CNTs enhanced the surface area, thus increasing the number of sites for adsorption, while increasing the doping to 30 wt% and 50 wt% decreases their surface areas due to the aggregation and agglomeration of iron nanoparticles and the formation of a large cluster of nanoparticles that blocks the available surfaces on the CNTs.

The loading of Fe2C>3 nanoparticles on the negative surfaces of carbon nanotubes have a great impact on the stability of the oil emulsion breaking process, which substantially improves the adsorption capacity of the oil on the surface of the carbon nanotubes. The existence of Fe³⁺ on the surfaces of the carbon nanotubes will modify the liquid/liquid and liquid/air surface properties. For example, Fe³⁺ serves to decrease the interfacial tension between the dispersed oil phase and the water, and then increases the interfacial tension between the air bubble and the oil phase. Accordingly, loading Fe2C>3 nanoparticles on the surface of carbon nanotubes will increase oil droplet coalescence. Enhancing this coalescence will also facilitate the adsorption mechanism. This phenomenon can be explained through the zeta potential measurements. Table 2, below, shows that the undoped carbon nanotubes have a negative charge of - 42.6 mV in the oil-in-water emulsion. Loading Fe2C>3 nanoparticles on the negative surface of carbon nanotubes decreases the negative sign of the zeta potential by overcoming the repulsive effects of the electrical double layers to allow the finely sized oil droplets to form larger droplets through coalescence. In these tests, the zeta potential of oil droplets was not measured. However, the literature indicates that oil droplets have a large negative zeta potential. This implies that electrostatic repulsion would make attachment between oil droplets highly unlikely. Thus, it is important to decrease the electrostatic repulsion barrier in the oil emulsion systems under investigation to further improve the adsorption mechanism. In agreement with this, it is noted that increasing the percentage of Fe2C>3 loaded on the negative surface of carbon nanotubes reduced the magnitude of the zeta potential of oil droplets, and this appears to be a crucial factor in the adsorption performance of such emulsion systems. This finding is in a good agreement with the present adsorption measurements, as the percentage of the a-Fe20₃ on the surface of the carbon nanotubes increased the removal of the emulsified oil from the water. It should also be noted that the similarities in the adsorption capacity for 30 wt% and 50 wt% of a-Fe20₃ loaded on the surface of carbon is a result of the similarity in the zeta potential value. Table 2 below shows the Zeta potential of CNTs and a-Fe20₃/CNT nanocomposites with 1 wt%, 10 wt%, 30 wt%

Table 2: Zeta Potential of CNTs and Modified CNTs

CNT/aFe₂0₃ Zeta Potential (mV)

0% -42.6

1% -20.5

10% -17.1

30% -8.9

50% -8.3

Example 5

Effect of contact time

In order to find the optimum equilibrium and contact time for maximum uptake of emulsified oil from water by undoped and doped CNTs, experiments were performed over different contact times ranging from 0 h to 2 h. All other parameters, including shaking speed, adsorbent dosage, initial concentration and pH, were kept constant at 400 rpm, 20 mg, 841 ppm and pH 7, respectively. The samples were taken at different preset time intervals and the concentration of remaining oil was measured using a TOC analyzer. The dynamic sorption of oil with both undoped and doped CNTs with different concentrations of iron oxide nanoparticles is shown in Fig. 7.

Oil removal from produced water with both undoped and doped CNTs increases with an increase in contact time and reaches a maximum adsorption capacity after 20 min. The presence of iron nanoparticle-doped CNTs enhances the removal efficiency and adsorption capacity compared to undoped CNTs. The maximum removal efficiency of undoped and doped CNTs with 1 wt%, 10 wt%, 30 wt% and 50 wt% of Fe₂0₃ nanoparticles was 87.0%, 96.09%, 96.37%, 96.62% and 98.52%, respectively. The amount of sorbent oil particles increases rapidly at the initial stage and then progressively reaches 90% equilibrium capacity in 10-15 min adsorption time. The high removal efficiency rate at the beginning of the contact time was due to the large number of vacant binding sites available for the adsorption of oil. These rapid uptakes coupled with a high sorption capacity were two of the most significant properties for oil sorption by CNTs. As the outside surfaces of CNTs become exhausted and saturated with oil particles, the rate of oil uptake starts to decrease and reaches equilibrium. The doped CNTs reach equilibrium faster than undoped CNTs by almost a factor of two. Compared against prior art magnetic CNTs, the present modified CNTs reached equilibrium three times faster.

Example 6

Effect of adsorbent dosage

Batch adsorption experiments with different doses of adsorbents ranging from 5 mg to 35 mg were designed to investigate the effect of adsorbent dosage on removal efficiency of oil from water using undoped and doped CNTs. For this experiment, all other experimental parameters, including agitation speed, initial oil concentration and contact time, were kept constant at 400 rpm, 841 mg/L and 2 h, respectively. Fig. 8 shows that the final oil concentration decreases with increasing adsorbent dosage. The maximum removal of oil using undoped and doped CNTs was obtained at 30 mg and 25 mg, respectively. It can be seen that the percentage of oil removal for all Fe₂C>₃ nanoparticle loadings was similar at high CNTs dosages. However, at low adsorbent doses, oil removal was more sensitive to the Fe₂C>₃ nanoparticle loadings. The higher removal percentage by doped CNTs was due to the large available active adsorption sites and an increase in oil droplet coalescence. Example 7

Adsorption isotherms for oil removal

The equilibrium adsorption is important in the design of adsorption systems. Equilibrium studies in adsorption indicate the maximum capacity of the adsorbent during the treatment process. The effect of initial concentration on oil adsorption was investigated by varying the initial concentration of oil (from 400 mg/L up to 7500 mg/L) at optimum experimental conditions (adsorbent dosage: 20 mg and contact time: 2 h). Equilibrium adsorption data were used to determine the maximum adsorption capacity of the undoped and doped CNTs. The Langmuir, Freundlich, and Temkin isotherm models were employed to demonstrate the adsorption data.

The Langmuir, Freundlich, and Temkin equations were used to describe the data derived from the adsorption of oil by the different adsorbents over the entire parameter range studied. Based on Figs. 7 and 8, the adsorption capacities (Q_e) and adsorption intensities were determined from the slope and the intercept of each adsorbent graph, respectively. By comparison of Langmuir and Freundlich isotherm data in Table 3, the Freundlich isotherm shows a better fitting model with higher correlation coefficients for both undoped and doped CNTs.

Table 3: Parameters of Langmuir and Freundlich Isotherm Models

Langmuir Freundlich

Parameters £>_m(mg/g) K_L(Umg) R² n K_F R²

CNTs 3.33X10⁴ 7.22E-04 0.51 1.08 53.15 0.99

CNTs-a-Fe₂0₃ 1 wt% 2.00X10⁶ 1.43E-05 0.00 1.01 29.78 1.00

CNTs-a-Fe₂0₃ 10 wt% 5.00X10⁵ 6.23E-05 0.01 1.02 33.25 1.00

CNTs-a-Fe₂0₃ 30 wt% 1.25X10⁵ 2.40E-04 0.33 0.98 27.86 1.00

CNTs-a-Fe₂0₃ 50 wt% 3.33X10⁴ 1.17E-03 0.23 1.16 62.30 0.99 Based on the data analysis, the Langmuir isotherm does not show a good fit for all of the adsorbents, with all of the coefficients of determination (R²) being less than 0.5. On the other hand, both undoped and doped CNTs show very good agreement with the Freundlich model, as shown in Fig. 9. The Langmuir isotherm assumes monolayer coverage on a homogeneous surface with identical adsorption sites. In solution-solid systems, with the hydration forces, mass transport effects, etc., the system is much more dynamic and complex, and obeying the isotherm does not necessarily reflect the validity of the aforementioned assumptions. In such systems, the isotherm adequacy can be seriously affected by the experimental conditions. The Freundlich isotherm is commonly used to describe the adsorption characteristics for heterogeneous surfaces, as it is easier to handle mathematically in more complex calculations (e.g., in modeling the dynamic column behavior), where it may appear quite frequently. Therefore, the Freundlich isotherm model was employed to describe the adsorption of oil on the surface of all the adsorbents. The Freundlich isotherm describes the adsorption process to be reversible and not restricted to the formation of a monolayer. Therefore, the amount of oil adsorbed on the CNTs is the summation of adsorption on all sites, with the stronger energy binding sites being occupied first, until the adsorption energies exponentially decreased upon the completion of the adsorption process.

The Temkin isotherm model analysis results are shown in Fig. 10 and Table 4. The constants in the Temkin isotherm are found by plotting Q_e as a function of In C_e. The correlation coefficient of 0.8 to 0.9 is obtained for the different adsorbents. The A value, which is an indication of binding energy, shows that there is a linear increase in the standard enthalpy of adsorption with surface coverage, and when the surfaces of the CNTs are doped with Fe₂C>3 nanoparticles, the surface binding energy increases. This can be related to the zeta potential and increase in charge density by the introduction of Fe₂C>3 nanoparticles.

Table 4: Parameters of Temkin Isotherm Model

Temkin Isotherm

Samples ^

CNT 4.46E-02 2.20E+03 1.13 0.90

1 wt% Fe 5.48E-02 2.32E+03 1.07 0.87

10 wt% Fe 6.00E-02 2.30E+03 1.08 0.85

30 wt% Fe 5.57E-02 2.43E+03 1.02 0.88

50 wt% Fe 8.69E-02 1.98E+03 1.25 0.80

Example 8

Adsorption kinetics for the removal of oil The modeling of the kinetics for oil adsorption on CNTs was investigated by three common models, namely, the Lagergren pseudo-first order, pseudo-second-order and intra- particle diffusion models. Linearized plots of the three models, which assume an oil concentration of 841 mg/L, are shown in Figs. 11, 12 and 13, respectively. The parameters of the second order equations are tabulated in Table 5. Table 5: Kinetic Parameters for Second Order Models

Second Order Model

Samples

g_e (mg/g) g/mgh) R²

CNT 7.69X10" 2.70X10 " 0.99

1 wt% Fe 8.33X10² 2.30X10^"4 0.99

10 wt% Fe 8.33X10² 2.93X10^"4 0.99

30 wt% Fe 8.33X10² 3.70X10^"4 0.99

50 wt% Fe 8.33X10² 5.54X10^"4 0.99

From Fig. 11, it can be seen that the data do not fit very well to the first order model, as the R² values are less than 0.8 in most cases. The data in Table 5 show that the value of rate constant ¾ is lowest for undoped CNTs and increases as the iron dosage into the CNTs increases. The equilibrium adsorption capacity of undoped CNTs is lower than that of the doped CNTs as well, due to the availability of more active adsorption sites and increasing surface charge density (Fig. 12). The results also imply that different percentage loadings of iron on CNTs do not have a major impact on equilibrium capacity of the CNTs.

Fig. 13 shows the plot of Q_t vs. t¹¹² for the intra-particle diffusion model. It can be clearly seen that the plot does not show a linear trend over the entire time range. There are two almost linear regions, but the sorption time of 20 minutes for the oil uptake is far too short for an intra-particle diffusion mechanism. Nevertheless, the t 1/2 plot does give some visible insight to the mechanism by showing time regions. The primary linear part may be explained as external surface adsorption, in which the oil particles diffuse through the solution to the external surface of the adsorbent. The second and intermediate linear portion refers to a slower adsorption into the pores of the adsorbent with a slower rate than the first portion. During the first few minutes, layers of oil build up on the external surface of the CNTs in the form of a strongly bound layer of oil due to adhesion and then looser bound layers due to cohesion. This represents most of the oil sorption capacity. After several minutes, a slower diffusion of oil into the pores occurs, representing about 5-10% of the capacity. Table 6 summarizes the external mass transfer rate parameter (£,) for all five adsorbents found from the linear fit of the data. The fits are for the values of the intercept C, which offer information regarding the thickness of the boundary layer. Table 6:Kinetic Parameters for Intraparticle Diffusion Models

Example 9

CNT removal from water A qualitative assessment of the performance of CNTs in oil-water adsorption was performed in test tubes. The gasoline-water emulsion was prepared by sonication without surfactants, as the water-oil interface is clearly visible due to the light-scattering property of oil droplets. It was found that the CNTs in water precipitated at the bottom of the test tube due to a lack of adsorption and the hydrophobic nature of the CNTs. After the addition of CNTs to the solution, the CNTs were freely dispersed in the gasoline- water emulsion by CNT aggregation around the oil droplets. The behavior of the CNTs in water only indicated that most of the CNTs were undispersed and settled at the bottom of the tube. After two hours of the adsorption process, most of the CNT particulates floated on the top of the tube due to coverage by the oil droplets. After six hours, the CNTs adsorbed more oil and there were fewer CNTs observed in the tube, with most of them floating on the top. After 24 hours, the oil was fully removed by the CNTs, with the water showing as clear with all CNTs at the top of the tube. The floating CNTs were removed from the clean water by a conventional sand filtration process that was simulated by placing a sand layer in a sand filtration column. The effluent of the sand bed was clear water, with no evidence of CNT particles. This demonstration provided support for the feasibility of creating an industrial treatment process consisting of the CNT adsorption process, sedimentation, and sand filtration to fully remove oil and CNTs from water, and the recycling of fresh water. To remove CNTs from the oil/water emulsion after treatment, ethanol was used to extract the oil particles from the CNTs. The remaining residue was agglomerated CNTs that could be used again in oil removal.

In the development of an engineering treatment design process, several parameters should be taken into account to enhance the present separation technique. First, the reusability and regeneration of modified CNTs with iron oxide should be considered. Second, the presence of different co-adsorbates and ions in the aqueous solution along with oil should be examined. The influence of constituents, such as other hydrocarbons, heavy metal ions, and particulate matter in the emulsion, may create adsorption competition and affect the removal efficiency. On an industrial scale, wastewater does not contain only one constituent and usually is a mixture of many materials. The presence of such materials may interact with the crystallinity of the CNTs, and consequently change the physicochemical properties of modified CNTs, and ultimately, their interaction with oil droplets. Finally, advancements in material synthesis, application and reuse can help improve the economic competitiveness of the treatment process using the present modified CNTs. To do so, the first stage is selection of appropriate CNT material, as described above.

It is to be understood that the method of removing oil from water is not limited to the specific embodiments described above, but encompasses any and all embodiments within the scope of the generic language of the following claims enabled by the embodiments described herein, or otherwise shown in the drawings or described above in terms sufficient to enable one of ordinary skill in the art to make and use the claimed subject matter.

Claims

1. A method of removing oil from water, comprising the steps of:

contacting a mixture of oil and water with modified carbon nanotubes, the modified carbon nanotubes being carbon nanotubes doped with a metal or metal oxide selected from the group consisting of iron, silver, aluminum, zinc, copper, and oxides thereof;

adsorbing oil from the mixture of oil and water onto the modified carbon nanotubes to yield purified water; and

isolating the modified carbon nanotubes from the purified water.

2. The method of removing oil from water as recited in claim 1, wherein the modified carbon nanotubes comprise carbon nanotubes doped with iron oxide (Fe20₃), the modified carbon nanotubes having between 1% and 50% iron oxide (Fe₂0₃) by weight.

3. The method of removing oil from water as recited in claim 1, wherein each said modified carbon nanotube has an outer diameter between 10 nm and 20 nm.

4. The method of removing oil from water as recited in claim 3, wherein each said carbon nanotube has a length between 1 μιη and 10 μιη.

5. A method of making an iron oxide-modified carbon nanotube, comprising the steps of:

dissolving ferric nitrate and iron oxide in ethanol to form a first solution;

dissolving carbon nanotubes in ethanol to form a second solution;

sonicating the first solution and the second solution separately;

mixing the first solution and the second solution to form a third solution;

sonicating the third solution to form a sonicated mixture;

heating the sonicated mixture to the ethanol and provide a solid residue; and calcining the solid residue to impregnate the iron oxide into the carbon nanotubes.

6. The method of making an iron oxide-modified carbon nanotube as recited in claim

5, wherein the step of heating the sonicated mixture comprises heating the sonicated mixture at a temperature of between 60 °C and 90 °C.

7. The method of making an iron oxide-modified carbon nanotube as recited in claim

6, wherein the step of calcining the solid residue comprises calcining the solid residue at a temperature of about 350°C for between two hours and five hours.

8. The method of making an iron oxide-modified carbon nanotube as recited in claim

7, wherein the step of sonicating the third solution comprises sonicating the third mixture for a period of at least thirty minutes.

9. An oil adsorbent, comprising a plurality of carbon nanotubes doped with a metal or metal oxide selected from the group consisting of iron, silver, aluminum, zinc, copper, and oxides thereof.

10. The oil adsorbent as recited in claim 9, wherein the metal or metal oxide comprises iron oxide (Fe₂0₃).

11. The oil adsorbent as recited in claim 10, wherein iron oxide comprises between 1 wt% and 50 wt% of each of the doped carbon nanotubes.