PRIORITY CROSS REFERENCE

[001 ] The present application claims priority to Australian Provisional Patent Application No. 2015901503 filed 28 April 2015 the contents of which should be understood to be incorporated into the present specification by this reference.

TECHNICAL FIELD

[002] The present invention generally relates to a non-covalent magnetic graphene oxide composite material and method of production thereof. The invention is particularly applicable as an adsorbent in fluid treatment, particularly aqueous solution treatment and it will be convenient to hereinafter disclose the invention in relation to that exemplary application. However, it is to be appreciated that the invention is not limited to that application and could be used as an adsorbent in a number of other applications.

BACKGROUND OF THE INVENTION

[003] The following discussion of the background to the invention is intended to facilitate an understanding of the invention. However, it should be appreciated that the discussion is not an acknowledgement or admission that any of the material referred to was published, known or part of the common general knowledge as at the priority date of the application.

[004] Graphene oxide (GO) has attracted considerable research interest in recent years due to its potential in a vast array of applications including optics, stabilization of interfaces, spectroscopic sensors, and more recently, water treatment. GO readily disperses in water to form stable colloidal dispersions, facilitating its deployment in aqueous systems. This effect stems from the low acidity constant of carboxyl groups at the periphery of GO sheets (pKa = 4.3), meaning they readily dissociate into carboxylate anions. Therefore, GO maintains a negative surface charge down to very low values of pH (<1 ) retaining charge-based stability across a wide pH range. These surface properties, along with the vast surface area to mass ratio provided by the sheets, explain to a large extent why investigations into the use of GO as an adsorbent material for the removal of toxins from aquatic environments are becoming increasingly prevalent.

[005] Previous studies (for example WO2012170086A1 ) have shown that GO is a suitable material for purifying water of many types of pollutants, from antibiotics to heavy metals. Typical adsorption capacities of GO (ca. 100 mg/g for heavy metal ions) are approaching those seen for the zeolites commonly used, with the advantage that GO is readily and cheaply prepared from abundant natural graphite deposits. These discoveries have inspired the development of a wide variety of graphene and graphene oxide composite materials that have shown great potential for many environmental applications concerning pollution. Certain composites have been ideally formulated and proven to be effective in the adsorption of gases, while others are specialized for removing certain compounds from water such as dye molecules.

[006] However, for the use of GO in the capture of pollutants to be feasible, the material would need to be recovered following sorption. This realization has led to the production of GO composites containing magneto-responsive components, providing a simple method for removing the material from solution after it has been deployed. Magnetic composites incorporating GO and reduced GO have been successfully implemented in the removal of many toxic metal elements from aqueous systems including arsenic, chromium, cadmium, selenium, and mercury. In each of these accounts, the magnetic nanoparticles are introduced to the GO through synthetic pathways that covalently bind the particles to the surfaces of the sheets.

[007] For example, international patent publication No. WO2014094130A1 teaches the use of a GO grafted with magnetic particles (iron oxide) using a controlled Shen's method. The magnetic particles are firmly entrapped and immobilized onto GO sheets via strong covalent bonding. The magnetically grafted GO is used to absorb the heavy metal from wastewater. The process uses a magnetic field to separate the heavy metal loaded magnetic graphene oxide (MGO) from the remaining aqueous solution. Once separated, the heavy metal can be stripped from the heavy metal loaded MGO particle so that it can be reused.

[008] The drawback to using conventional GO composite materials, such as taught in WO2014094130A1 , is that their production involves complex, high energy, and multistep procedures that permanently change the structural makeup of the GO. In addition, they have a comparatively narrow range of uses. Therefore, in essence, the utilization of pure, "virgin" GO as an adsorbent for water decontamination would be ideal. Aside from being difficult to remove from solution, GO also possesses in vivo, toxic characteristics associated with its oxygenated functional groups.

[009] It would therefore be desirable to develop a new or improved method of forming a magnetic graphene oxide composite or complex.

SUMMARY OF THE INVENTION

[010] The present invention provides in a first aspect a method of preparing a magnetic graphene oxide composite material, comprising:

forming a dispersion of graphene oxide and the magnetic material within a dispersant solution; and

controlling the pH of the dispersant solution to cause the graphene oxide and magnetic material to have different surface charges,

thereby forming a magnetic graphene oxide composite material comprising a non-covalent complex of graphene oxide and a magnetic material.

[01 1 ] The present invention utilises the unique charging properties of graphene oxide to prepare of a range of noncovalent magnetic graphene oxide composite materials. Here, non-covalent charge interactions are used to bind the magnetic material and graphene oxide together such that the magnetic material forms a charge-based complex with graphene oxide. This non-covalent complex of graphene oxide and magnetic materials is formed by controlling, and more typically changing the pH conditions of the dispersant solution. Here, a decrease in pH (more acidic) is used to cause the graphene oxide and magnetic material to have different surface charges, resulting in attraction and formation of a magnetic graphene oxide composite material. Advantageously, graphene oxide is unaltered by the process of the present invention. Accordingly, there is no compromise to its original properties following adsorption, and, thus, the graphene oxide and magnetic material can be used in various applications, such as adsorption processes or the like.

[012] The process of the present invention utilises cheap and readily processed materials to form the desired complex. The present invention therefore circumvents the need to perform difficult and lengthy syntheses of GO nanocomposites and is a low energy alternative to centrifugation and polymer flocculation.

[013] The surface charge of a species or material in solution is largely determined by the pH of that solution and the isoelectric point of that species or material. The present invention can therefore preferably operate where the pH of the dispersant solution is controlled to have a pH lower than the isoelectric point of the magnetic material and a pH higher than the isoelectric point of graphene oxide.

[014] For this charge interaction to work effectively, it is preferred that the difference between the isoelectric point of the magnetic material and the isoelectric point of graphene oxide is large, preferably at least 3 pH units, more preferably at least 5 pH units. In exemplary embodiments the difference in isoelectric points of graphene oxide and the magnetic material is at least 6 pH units, and more preferably at least 7 pH units.

[015] The step of controlling the pH of the dispersant solution can encompass either controlling the pH of the dispersant solution to fall within the desired pH range or maintaining the pH of the dispersant solution to fall within the desired pH range.

[016] Thus, in some embodiments, adsorption (and desorption - see below) of the magnetic material from graphene oxide can be controlled by modification of pH within a desired pH unit window. Thus, in some embodiment the pH of the dispersant solution is modified to lower the pH by < 4 pH units, preferably < 3 pH units, more preferably < 2 pH units below the isoelectric point of the magnetic material. [017] However, it should be appreciated that in some embodiments the pH of the dispersant solution may already fall within the desired pH unit window which causes the graphene oxide and magnetic material to have different surface charges. For example, the dispersant solution may comprise an acidic solution having the desired or suitable pH. In this embodiment, the pH would not need to be modified, but rather maintained within the pH window to facilitate the formation of the magnetic graphene oxide composite material.

[018] The pH of the dispersant solution can be controlled by any suitable means. In some embodiments, a solution with a desired pH, for example acidic or alkaline can be added to the dispersant solution to provide a desired pH within the dispersant solution. In other embodiments, a gas can be introduced into the dispersant solution to control the pH of the dispersant solution. The gas can be introduced using any suitable technique. In embodiments, the gas is introduced using sparging.

[019] The type of gas introduced depends on the desired pH of the solution. Where it is desired to decrease the pH (i.e. become more acidic) an acid gas can be introduced into the dispersant solution, for example CO ₂ , SO ₂ , H ₂ S or the like. In preferred embodiments, CO ₂ is introduced into the dispersant solution. Where it is desired to increase the pH (i.e. become more alkaline) gases such as argon, nitrogen or air, and ammonia can be introduced into the dispersant solution.

[020] The magnetic material used in the complex can comprise any suitable material having an isoelectric point which is different to the isoelectric point of graphene oxide.

[021 ] In some embodiments, the magnetic material comprises micro or nano particles. Here, the micro or nano particles of magnetic material preferably have a size range of 5 nanometres to 100 micrometres.

[022] In some embodiments, the magnetic material comprises a ferrous material. One preferred magnetic material used in the present invention is iron oxide, Fe ₂ O ₃ . However, it should be appreciated that a number of other magnetic materials could also be used including other oxides such as Fe ₃ O ₄ and metal nanoparticles such as Fe, Co and Hf.

[023] In other embodiments, the magnetic material comprises a magnetic ionic liquid (preferably a surfactant). Suitable examples include 1 - methyl-3- butylimidazolium tetrachloroferrate (mim) or dodecyltrimethylammonium trichloromonobromoferrate (DTA). However, it should be appreciated that other suitable magnetic ionic liquids could equally be used in the present invention.

[024] Where the magnetic material comprises micro or nano particles, the graphene oxide and the micro or nano particles magnetic material could be dispersed together within a dispersant solution. In alternate embodiments, the graphene oxide and micro or nano particles of magnetic material are dispersed within two separate dispersant solutions that are subsequently mixed together.

[025] The dispersant solution can comprise any solution in which the magnetic material and graphene oxide readily disperses therein. Suitable dispersant solution comprises at least one of: water, other high-dielectric solvents such as formamides and acetamides, or other organic solvents with or without molecular stabilisers present. In exemplary embodiments, the dispersant solution comprises an aqueous solution. In this regard, graphene oxide readily disperses in water to form stable colloidal dispersions, facilitating its deployment in aqueous systems.

[026] For composites formed from graphene oxide and magnetic nanoparticles (for example Fe2O3 nanoparticles), a low magnetic material:graphene oxide mass ratio is preferred to favour flocculation of graphene oxide. Higher ratios (>5:1 ) can cause overcharging of the surfaces resulting in restabilisation of the separate dispersed components. Accordingly, in some embodiments the mass ratio of magnetic material to graphene oxide is preferably less than 5:1 , and more preferably less than 4:1 .

[027] The graphene oxides of the present invention can come in the form of variable sized sheets. Such sheets may have lengths or diameters that range from about a few nanometers to a few hundred microns to several centimeters. In more specific embodiments, the graphene oxides may have lengths or diameters that range from about 1 nanometer to about 3 centimeters. The graphene oxide is also preferably provided in a micro or nano-particle form. In preferred embodiments, the graphene oxide comprises flakes, preferably nano- platelets or micro-platelets. Here, the flakes of graphene oxide preferably have an average platelet size of 100 nanometers to 10 micrometers and a thickness of 1 to 10 molecular layers.

[028] The benefit conferred by using charge-based adsorption of the present invention is that the process is reversible, and the GO can be captured and separated from the magnetic nanomaterial, such that both components can be recycled. Capture of graphene oxide via this process is fully reversible as the GO can be redispersed into solution by readjusting the system pH. Where it is desired to separate the graphene oxide-magnetic material composite from the dispersant solution the method can further include the step of:

magnetically separating the graphene oxide-magnetic material composite from the dispersant solution.

[029] Where it is desired to separate the graphene oxide and magnetic material in the complex, the method can further include the step of:

controlling the pH of a solution containing the magnetic graphene oxide composite material to cause the graphene oxide and magnetic material to have the same surface charge,

thereby separating the graphene oxide and magnetic material, such that one or both of the graphene oxide and magnetic material can be recycled.

[030] In this step, the pH is controlled, typically modified or increased to cause the graphene oxide and magnetic material to have the same sign of surface charge, resulting in repulsion which causes separation of the graphene oxide and magnetic material. Advantageously, this separation enables the graphene oxide and magnetic material to be recycled, and be preferably reused to form another magnetic graphene oxide composite material.

[031 ] Once again, the surface charge of a species or material in solution is largely determined by the pH of that solution and the isoelectric point of that species or material. Hence, separating the graphene oxide and magnetic material preferably requires the pH of the solution to be modified to have a pH of at least or greater than the isoelectric point of the magnetic material. In some embodiments, the pH of the solution is modified to have a pH of at least 1 pH unit greater than the isoelectric point of the magnetic material.

[032] Again, the pH of the solution can be controlled by any suitable means. In some embodiments, a solution with a desired pH, for example acidic or alkaline can be added to the solution to provide a desired pH within the solution. In other embodiments, a gas can be introduced into the solution to control the pH of the solution. The gas can be introduced using any suitable technique. In embodiments, the gas is introduced using sparging.

[033] The type of gas introduced depends on the desired pH of the solution. Where it is desired to decrease the pH (i.e. become more acidic) an acid gas can be introduced into the solution, for example CO ₂ , SO ₂ , H ₂ S or the like. In preferred embodiments, CO ₂ is introduced into the solution. Where it is desired to increase the pH (i.e. become more alkaline) gases such as argon, nitrogen or air, and ammonia can be introduced into the solution.

[034] The present invention also relates to the magnetic graphene oxide composite material formed from a process according to the first aspect of the present invention. A second aspect of the present invention provides a magnetic graphene oxide composite material comprising a non-covalent complex of graphene oxide and a magnetic material dispersed in a dispersant solution having a pH that is lower than the isoelectric point of the magnetic material and higher than the isoelectric point of graphene oxide.

[035] Like the first aspect, adsorption and desorption of the magnetic material from graphene oxide can be controlled by modification of pH within a desired pH unit window. Thus in some embodiments, the pH of the dispersant solution is controlled to lower the pH by < 4 pH units, preferably < 3 pH units, more preferably < 2 pH units below the isoelectric point of the magnetic material. [036] Again, the magnetic material used in the complex can comprise any suitable material have an isoelectric point which is different to the isoelectric point of graphene oxide. In some embodiments, the magnetic material comprises micro or nano particles. In some embodiments, the magnetic material comprises a ferrous material. One preferred magnetic material used in the present invention is iron oxide, Fe ₂ O ₃ . In other embodiments, the magnetic material comprises a magnetic ionic liquid (preferably a surfactant). Suitable examples include 1 - methyl-3-butylimidazolium tetrachloroferrate (mim) or dodecyltrimethylammonium trichloromonobromoferrate (DTA). However, it should be appreciated that other suitable magnetic ionic liquids could equally be used in the present invention.

[037] Similarly, suitable dispersant solutions are the same as described in relation to the first aspect of the present invention.

[038] A third aspect of the present invention provides a method of treatment of water using a magnetic graphene oxide composite material, comprising:

incorporating a magnetic graphene oxide composite material formed using the method of the first aspect of the present invention or according to the second aspect of the present invention in a treatment solution including at least one capture material capable of being adsorbed by graphene oxide;

allowing the magnetic graphene oxide composite material to adsorb the capture material or materials; and

magnetically separating the magnetic graphene oxide composite material from the treatment solution.

[039] This third aspect of the present invention advantageously uses the fully recyclable nature of the assembly and capture process, and vast adsorption capacity of GO (for example 100 mg/g for heavy metal ions), for decontamination and fluid treatment applications. The GO content of the composition can be used as an adsorbent, capturing material from solution, and then separated in the form of the graphene oxide-magnetic material composite using an applied magnetic field. [040] Once separated, the graphene oxide-magnetic material composite can be separated into the two components once again by increasing the pH of the dispersant solution, such that both the graphene oxide and magnetic material components can be recycled. In such embodiments, the method further includes the step of:

adding the magnetic graphene oxide composite material to a solution; and

controlling the pH of the solution to cause the graphene oxide and magnetic material to have the same surface charge,

thereby separating the graphene oxide and magnetic material, such that one or both of the graphene oxide and magnetic material can be recycled.

[041 ] Again, separating the graphene oxide and magnetic material preferably requires the pH of the solution to be controlled to have a pH of at least or greater than the isoelectric point of the magnetic material. In some embodiments, the pH of the solution is controlled to have a pH of at least 1 pH unit greater than the isoelectric point of the magnetic material.

[042] Again, the pH of the solution can be controlled by any suitable means. In some embodiments, a solution with a desired pH, for example acidic or alkaline can be added to the solution to provide a desired pH within the solution. In other embodiments, a gas can be introduced into the solution to control the pH of the solution. The gas can be introduced using any suitable technique. In embodiments, the gas is introduced using sparging.

[043] The type of gas introduced depends on the desired pH of the solution. Where it is desired to decrease the pH (i.e. become more acidic) an acid gas can be introduced into the solution, for example CO ₂ , SO ₂ , H ₂ S or the like. In preferred embodiments, CO ₂ is introduced into the solution. Where it is desired to increase the pH (i.e. become more alkaline) gases such as argon, nitrogen or air, and ammonia can be introduced into the solution.

[044] Typically after separation of the graphene oxide and the magnetic material, the capture material can be stripped or otherwise separated from the loaded graphene oxide and the graphene oxide can be recycled. A variety of stripping or separation processes well known in the art can be used to separate the graphene oxide and the capture material. Selection of a suitable process depends on the nature and properties of the capture material.

[045] The magnetic graphene oxide composite can be formed in situ in the treatment solution or separately formed and then added to the treatment solution. In some embodiments, the magnetic graphene oxide composite is formed in the treatment solution using a method according to the first aspect of the present invention, in which the dispersant solution is the treatment solution. In alternate embodiments, the magnetic graphene oxide composite is added to the treatment solution after being formed using a method according to the first aspect of the present invention.

[046] Again, it should be appreciated that the step of controlling the pH of the dispersant solution (or the treatment solution in those embodiments where the treatment solution is the dispersant solution) can encompass either controlling (typically modifying) the pH of the dispersant solution to fall within the desired pH range or maintaining the pH of the dispersant solution to fall within the desired pH range. Again, in some embodiments the pH of the dispersant solution/ treatment solution may already fall within the desired pH unit window which causes the graphene oxide and magnetic material to have different surface charges. For example, the treatment solution may comprise an acidic solution having the desired or suitable pH. In this embodiment, the pH would not need to be modified, but rather maintained within the pH window to facilitate the formation of the magnetic graphene oxide composite material.

[047] As noted above, graphene oxide is a suitable material for purifying water of many types of pollutants, from antibiotics to heavy metals. The magnetic graphene oxide composite of the present invention can therefore adsorb or otherwise capture a large variety of capture materials.

[048] In some embodiments, the capture material comprises a metal or metal ion, such as at least one of at least one of gold, silver, platinum, chromium, cadmium, mercury, selenium, lead, copper, zinc, chrome and arsenic. In some embodiments, the capture material comprises at least one heavy metal, heavy metalloid or ionic equivalent selected from chromium, cobalt, nickel, copper, zinc, arsenic, selenium, silver, cadmium, antimony, mercury, thallium and lead, or a combination thereof. In some embodiments, the capture material comprises a metal ion, for example a heavy metal ion or other forms of charged metals. In some embodiments, the capture material comprises a metal micro or nanoparticle.

[049] In other embodiments, the capture material comprises at least one radioactive element, chlorate, perchlorate, organohalogen, or combinations thereof. In more specific embodiments, the materials to be purified include, without limitation, polycyclic aromatics, chlorinated and brominated dibenzodioxins and dibenzofurans, chlorinated biphenyls, lindane, dichlorodiphenyltrichloroethane (DDT) and other similar hydrophobic xenobiotics.

[050] In some embodiments, the capture material or materials include radioactive elements. In some embodiments, the radioactive elements comprise at least one of metals, salts, metal salts, radionuclides, actinides, lanthanides, and combinations thereof. In more specific embodiments, the radioactive elements in the environment include radionuclides, such as thallium, iridium, fluorine, americium, neptunium, gadolinium, bismuth, uranium, thorium, plutonium, niobium, barium, cadmium, cobalt, europium, manganese, sodium, zinc, technetium, strontium, carbon, polonium, cesium, potassium, radium, lead, actinides, lanthanides and combinations thereof. In some embodiments, the radionuclide is actinide.

[051 ] In more specific embodiments, the radioactive elements include, without limitation, americium(lll), actinide(lll), actinide(IV), thallium(IV), plutonium(IV), neptunium(V), uranium(VI), strontium(ll), technetium(VII), and combinations thereof. In more specific embodiments, the radioactive elements include, without limitation, thallium-201 , iridium- 192, fluorine- 18, americium-241 , americium-243, neptunium-237, Gd- 153, niobium-93, barium-133, cadmium- 109, cobalt-57, cobalt-60, europium-152, manganese- 54,sodium-22, zinc-65, technetium-99, strontium-90, thallium-204, carbon- 14, polonium 210, cesium- 137, and combinations thereof.

[052] In some embodiments, the capture material or materials include chlorates, such as ammonium chlorate, barium chlorate, cesium chlorate, fluorine chlorate, lithium chlorate, magnesium chlorate, potassium chlorate, rubidium chlorate, silver chlorate, sodium chlorate, and combinations thereof. In some embodiments, the capture material or materials include perchlorates, such as ammonium perchlorate, barium perchlorate, cesium perchlorate, fluorine perchlorate, lithium perchlorate, magnesium perchlorate, perchloric acid, potassium perchlorate, rubidium perchlorate, silver perchlorate, sodium perchlorate, and combinations thereof. In additional embodiments, the capture material or materials include organohalogens, such as polychlorinated biphenyls (PCB) and halogenated flame retardants.

[053] It is also contemplated that the magnetic graphene oxide composite of the present invention can be used to remove other materials, compounds and ions from aqueous solutions including proteins, surfactants, dyes and pharmaceutical actives and excipients.

[054] One capture material may be within the treatment solution, or two, three or multiple capture materials may be within the treatment solution. The magnetic graphene oxide composite material can be used to capture/adsorb each of the capture materials within the treatment solution.

[055] Adsorption of the capture material onto graphene oxide may occur variously by charge-based attraction, i.e. following the same attraction principles as adsorption of the magnetic material thereon, or by other colloidal attractions including van der Waals forces, Lewis acid-base interactions, specific molecular attractions, chemisorption or physisorption.

[056] The treatment solution can comprise various solutions containing a capture material. In preferred embodiments, the treatment solution is an aqueous solution, for example wastewater or other contaminated or polluted water. [057] In some embodiments, the dispersion formed with the magnetic material within a dispersant solution comprises a dispersion of carbon-based nanomaterials selected from graphene oxide (GO), reduced graphene oxide (rGO) or carbon nanotubes (CNTs) with the magnetic material. In these embodiments, the aqueous dispersibility of the carbon-based nanomaterials in the dispersion solution can be controlled by light via the photoisomerisation of a photoswitchable surfactant molecule adsorbed to the surface of these materials. One photoswitchable surfactant molecule comprises a cationic azobenzene photosurfactant. The incorporation of a photoswitchable surfactant molecule into a GO, rGO or CNT dispersions allows these dispersions to be separated and redispersed on command utilizing UV radiation, preferably at 365 nm.

BRIEF DESCRIPTION OF THE DRAWINGS

[058] The present invention will now be described with reference to the figures of the accompanying drawings, which illustrate particular preferred embodiments of the present invention, wherein:

[059] Figure 1 illustrates (a & b) Conceptual schematic of the experiment: pH adjustment is used to effect charge attraction or repulsion between the GO sheets and Fe2O3 particles, (c) Zeta potentials of graphene oxide and Fe2O3, demonstrating the pH ranges at which electrical double-layer (EDL) attraction or repulsion would be expected.

[060] Figure 2 illustrates GO adsorption and recovery using Fe ₂ O ₃ microparticles including (a) AC (intermittent contact) mode atomic force microscopy (AFM) height image showing GO adsorbed onto Fe ₂ O ₃ microparticles. (b) The corresponding AFM amplitude image highlighting edge features. In (a) and (b), the white scale bars represent 200 nm. (c) Micrograph of Fe ₂ O ₃ microparticles at 100χ magnification. The horizontal dimension of the image is 120 μιη. (d) A histogram showing the size distribution (projected area) of the particles in (c). (e and f) The effect of an external magnetic field on GO (0.2 img/mL) without (left vials) and with (right vials) added Fe ₂ O ₃ microparticles (0.05 g) at high (e) and low (f) pH, demonstrating the separation of Fe ₂ O ₃ from GO in the nonadsorbed state and capture of the GO-Fe ₂ O ₃ complex when adsorption occurs at low pH.

[061 ] Figure 3 illustrates flocculation of GO using Fe ₂ O ₃ microparticles including: (a) Samples containing 20 mg of Fe ₂ O ₃ with varying concentrations of aqueous GO dispersion at pH 2.8. (b) Equivalent samples to those in (a) but at pH 1 1 .5. (c) UV-visible spectrophotometry data showing the concentration of GO remaining in the supernatant layer at high and low pH as a function of initial GO concentration. The dashed, black line shows a y = x trend for comparison (i.e., the limit of no adsorption).

[062] Figure 4 illustrates the effect of pH on flocculation of GO with Fe ₂ O ₃ , including: (a) Samples containing GO at 1 .5 mg/mL with 20 mg of Fe ₂ O ₃ microparticles present at different pH values, (b) The concentration of GO remaining in the supernatant layer as a function of pH for the fixed starting concentrations specified in (a) determined by UV-visible spectrophotometry.

[063] Figure 5 illustrates (a) Apparent hydrodynamic diameters of synthesized Fe ₂ O ₃ nanoparticles as a function of pH as determined using dynamic light scattering (DLS). Error bars correspond to the standard error for each point, and the dashed line is a guide to the eye. (b) AFM height image of Fe ₂ O ₃ nanoparticles dried onto mica. The white scale bar represents 1 μιη, and the dashed, blue line shows the position of the cross-sectional height profile that is presented in (c). (d) The magnetic extraction of GO in water using Fe ₂ O ₃ nanoparticles. Both vials contain 0.2 mg/mL of GO; however, the right-hand vial also includes a 0.2 mg/mL concentration of nanoparticles.

[064] Figure 6 illustrates (a) AFM height image of Fe ₂ O ₃ nanoparticles adsorbed onto sheets of graphene oxide at pH 3 and dried onto mica. The scale bar represents 200 nm. (b) Samples containing 0.15 mg/mL of aqueous GO and 0.15 mg/mL Fe ₂ O ₃ nanoparticles at different pH values, (c) Samples containing 0.1 mg/mL of GO in water (pH 3) and increasing concentrations of Fe ₂ O ₃ nanoparticles.

[065] Figure 7 illustrates (a and b) Magnetic response of magnetic surfactant-GO systems: all samples contain 0.2 mg/mL GO in water, but only the right-hand vials contain additionally 1 imM surfactant, of which the corresponding molecular structure is shown directly beneath the vials, (c) Adsorption isotherms for each surfactant on GO. Due to difficulties encountered in obtaining spectra for the higher concentrations of mim, approximate points (marked with crosses) have been added to demonstrate the approximate adsorption saturation concentration. The dashed trendlines have been added as guides to the eye.

[066] Figure 8 illustrates the magnetic compression of GO-Fe ₂ O ₃ materials: (a) The change in volume of a GO-Fe ₂ O ₃ microparticle matrix as a function of time when one is subjected to a magnetic field and the other is not. Measuring began after the initial settling due to gravitational forces having ceased, (b and c) Images of the two identical samples when first placed on the magnet/cardboard magnet (b) and 2 days later (c).

[067] Figure 9 illustrates (a and b) Demonstrations of how system pH can be used to reversibly capture and redisperse GO in the presence of Fe ₂ O ₃ microparticles (a) and Fe ₂ O ₃ nanoparticles (b). (c) At pH 3, Fe ₂ O ₃ microparticles alone are unable to capture gold nanoparticles from dispersion. Both vials contain the same concentration of gold nanoparticles; however, the right-hand vial also includes 10 mg of Fe ₂ O ₃ . (d) A mixture of Fe ₂ O ₃ microparticles and GO can effectively capture and extract gold nanoparticles from dispersion. Both vials contain the same concentration of gold nanoparticles; however, the right- hand vial also includes 0.2 img/mL of GO and 10 mg of Fe ₂ O ₃ . (e) UV-visible spectra showing the effect of magnetic collection of gold nanoparticles.

[068] Figure 10 shows a partial adsorption isotherm of caffeine on a GO-Fe ₂ O ₃ composite material according to one embodiment of the present invention.

[069] Figure 1 1 provides a partial adsorption isotherm of methylene blue dye on a GO-Fe ₂ O ₃ composite material according to one embodiment of the present invention.

[070] Figure 12 provides a photograph showing magnetic recovery of methylene blue dye using magnetic GO composite material. [071 ] Figure 13 provides an image demonstrating the drop in pH (A) and separation of GO-iron oxide complex from the supernatant (B) in the sample sparged with CO ₂ (S) that is not seen in the control (C), with both samples having 0.3 mg/ml GO and 0.15 mg/ml magnetite nanoparticles.

[072] Figure 14 provides (a, b & c) AFM height images of GO (a), rGO (b) and CNTs (c) dried onto mica. The dashed lines correspond to the height profiles in d, e and f and the scale bars represent 1 μιη, 500 nm and 200 nm respectively, (g) Structural schematic showing the reversible, photo-induced trans-cis isomerisation of azoTAB. (h) UV-visible spectra of trans- and cis-azoTAB.

[073] Figure 15 shows (a) Samples containing 0.1 mg/mL rGO and the specified concentrations of azoTAB 30 minutes after preparation. Measured zeta potentials are displayed for certain samples, (b) Identical samples in which the azoTAB solution was irradiated at 365 nm for 10 minutes before the rGO was added, (c) Adsorption isotherm of azoTAB in trans and cis dominated states onto rGO. All samples are at pH 1 1 . (d) An image series of the same 0.1 mg/mL rGO dispersion, demonstrating the effects of photoisomeration of the surfactant (1 .2 mM). (e) Samples containing 0.05 mg/mL of CNTs and azoTAB in trans-dominated (left) and cis-dominated (right) states. These samples were sonicated for approximately 2 minutes and images were taken after 1 hour.

[074] Figure 16 shows (a) Samples comprising 0.2 mg/mL GO and the specified concentrations of azoTAB 30 mins (0-0.5 mM) or 12 h (1 mM) after preparation, (b) Identical samples in which the azoTAB was irradiated at 365 nm for 10 minutes before the GO was added. All samples are at approximately pH 10. (c) Adsorption isotherm of azoTAB photosurfactant onto GO at high (open symbols) and low (solid symbols) pH. The green data series corresponds to samples that were irradiated with UV light for 10 minutes after GO was incorporated into the system, (d) Phase analysis light scattering data showing the change in zeta potential of GO for increasing concentrations of azoTAB in the trans and cis state. The dashed line is a guide to the eye.

[075] Figure 17 shows small-angle neutron scattering data of GO (0.15 mg/mL) or rGO (0.1 mg/mL ) in 0.25 mM (a), 0.5 mM (b) and 1 mM (c) solutions of azoTAB. Symbols represent the experimental SANS data and solid lines are the corresponding theoretical fits. The insets are digital images of each measured sample in quartz 'banjo' cells, (c) A schematic showing the proposed formation of fractal aggregates as azoTAB surfactant loadings are increased in the trans and cis state.

DETAILED DESCRIPTION

[076] The invention concerns the preparation of non-covalent complexes of graphene oxide (GO) and magnetic materials (microparticles, nanoparticles or magnetic ionic liquids), which can be used for use as water treatment adsorbents and capture aids for soluble chemical species. The process of the present invention can utilise cheap and readily processed materials to form the desired complex. The present invention therefore circumvents the need to perform difficult and lengthy syntheses of covalently grafted magnetic GO nanocomposites. The present invention also provides a low energy alternative to non-magnetic GO separation processes such as centrifugation and polymer flocculation, in which the magnetic properties of the magnetic GO particles of the present invention are separable from solution using an applied magnetic field.

[077] The non-covalent complexes of GO and magnetic materials according to the present invention are created by forming a dispersion of GO and the magnetic material within a dispersant solution, typically an aqueous solution. The pH of that dispersant solution is then controlled to cause the GO and magnetic material to have different surface charges, resulting in attraction and formation of a magnetic graphene oxide composite material. Here, non- covalent charge interactions are used to bind the magnetic material and GO together such that the magnetic material forms a charge-based complex with graphene oxide.

[078] The pH of the dispersant solution can be controlled by any suitable means. In embodiments, a gas can be introduced into the dispersant solution to control the pH of the dispersant solution, for example through sparging or another technique. The type of gas introduced depends on the desired pH of the solution. Where it is desired to decrease the pH (i.e. become more acidic) an acid gas can be introduced into the dispersant solution, for example CO ₂ , SO2, H ₂ S or the like. Where it is desired to increase the pH (i.e. become more alkaline) gases such as argon, nitrogen or air, and ammonia can be introduced into the dispersant solution.

[079] The structure and properties of GO used in the complex is unaltered by the process of the present invention. Accordingly, there is no compromise to its original properties following adsorption, and, thus, the GO and magnetic material can be used in various applications, such as adsorption processes or the like.

[080] Furthermore, the charge-based adsorption of the present invention is reversible, with the complex between GO and the magnetic material can be redispersed into solution by readjusting the system pH to cause the GO and magnetic material to have the same surface charge and resulting in repulsion between these materials.

[081 ] Without wishing to be bound by any one theory, the inventors have found that the adsorption of GO at a pH lower (moving to more acidic) than the isoelectric point of the magnetic material but higher than the isoelectric point of GO is attributed to attractive electrical double-layer forces between the GO and Fe ₂ O ₃ or surfactants. Magnetic material microparticles or nanoparticles form a changed-based complex with graphene oxide. Conversely in pH conditions proximate to, at or higher (moving to more basic) than the isoelectric point of the magnetic material (and thus also higher than the isoelectric point of GO), the dispersions remain stable due to like-charge repulsions. This provides a mechanism for the redispersion and separation of GO after magnetic capture.

[082] Suitable magnetic materials include ferrous materials, preferably in micro or nanoparticle form, and magnetic ionic liquids such as 1 - methyl-3- butylimidazolium tetrachloroferrate (mim) or dodecyltrimethylammonium trichloromonobromoferrate (DTA). [083] One exemplary magnetic material used in the present invention is iron oxide, Fe ₂ O ₃ . Iron oxide in the form of hematite or maghemite (a- or γ- Fe ₂ O ₃ ) is a material that naturally occurs in large abundance, with the γ form showing strong ferromagnetism and the a form showing weak but size-dependent magnetism due to spin-orbit coupling. The surface chemistry of Fe ₂ O ₃ is dependent on pH in aqueous media, with an isoelectric point (IEP) of 7 to 8. Therefore, in a solution at pH values below its IEP, Fe ₂ O ₃ particles would be expected to experience strong, charge-based attractions with GO sheets, arising from the opposing surface charges. Fe ₂ O ₃ is an attractive material to use in the present invention as while it is less magnetically responsive than other iron oxides (or other magnetic materials) it is by far the cheapest and most abundant, making large-scale wastewater treatment not only plausible but financially viable.

[084] A conceptual schematic of this charge dependent complex formation of the process of the present invention is shown schematically in Figure 1 which shows in Figure 1 (a) a conceptual schematic of pH adjustment is used to effect charge attraction or repulsion between the GO sheets and Fe ₂ O ₃ particles. Figure 1 (b) provides a more detailed schematic showing the adsorbed and non- adsorbed states via pH adjustment, and the effect of application of a magnetic field, which is schematically shown as being able to separate the Fe ₂ O ₃ from solution or dispersion whether when adsorbed to GO or not. Figure 1 (b) therefore schematic demonstrates when GO sheets and Fe ₂ O ₃ particles are attracted, forming the magnetic GO complex of the present invention, a magnetic field can be used to separate that complex and any other material adsorbed thereon (GO) from the surrounding dispersant solution.

[085] Figure 1 (c) provides a plot of the zeta potentials of graphene oxide and Fe ₂ O ₃ , demonstrating the pH ranges at which electrical double-layer (EDL) attraction or repulsion would be expected. Data for hematite are from Palomino, D.; Stoll, S. Fulvic Acids Concentration and pH Influence on the Stability of Hematite Nanoparticles in Aquatic Systems. J. Nanopart. Res. 2013, 15, 1428/1 -8. (solid symbols) or measured in this work using Fe ₂ O ₃ nanoparticles (open symbols) and for GO are from Chen, J.-T.; Fu, Y.-J.; An, Q.-F.; Lo, S.-C; Huang, S.-H.; Hung, W.-S.; Hu, C.-C; Lee, K.-R.; Lai, J.-Y. Tuning Nanostructure of Graphene Oxide/Polyelectrolyte LbL Assemblies by Controlling pH of GO Suspension to Fabricate Transparent and Super Gas Barrier Films. Nanoscale 2013, 5, 1 -7. The plot confirms that at pH 8, there is a cross-over between EDL attraction and EDL repulsion for graphene oxide and Fe ₂ O ₃ that can be exploited by pH adjustment of the dispersant solution in order to form or breakup a complex formed between these two materials.

[086] The reversible charge-based adsorption of the present invention shown in Figure 1 (b) and the adsorbent capacity of GO can be utilised for solution treatment and/or purification applications. For example, GO is a suitable material for purifying aqueous solution of many types of pollutants, from antibiotics to heavy metals. The magnetic GO complex can be formed insitu within the wastewater or aqueous solution to be treated or added after formation in a separate solution. The GO content of the composition can be used as an adsorbent, capturing material from solution, loading the magnetic GO complex with that capture material. The loaded magnetic GO complex can then be separated from the treated aqueous solution using an applied magnetic field/ magnet as shown in Figure 1 (b). In the presence of a magnetic field, each loaded magnetic GO complex move towards the source of the magnetic field. The loaded magnetic GO complex can then be removed and separated from the wastewater. Once magnetically separated, the GO and the magnetic material of the loaded magnetic GO complex can be separated. Once magnetically separated, the GO and the magnetic material of the loaded magnetic GO complex can be separated using by pH adjustment of the dispersant solution in order to break-up the complex formed between these two materials, such that both components can be recycled.

[087] The treatment and/or purification applications are typically used to treat an aqueous solution. Non-limiting examples of aqueous solutions include water, such as radioactive water, contaminated water, and waste water. In some embodiments, the solution includes nuclear fission products. In some embodiments, the solution includes waste or pollution products from industrial operations, mining operations, chemical processing operations or the like. In some embodiments, the solution to be purified is a non-aqueous solution, such as a solution containing benzenes, toluenes, dichloromethane, and other nonaqueous solvents.

[088] A variety of wastewater applications are envisaged, including antibiotics, radioactive elements, chlorates, perchlorates, organohalogens, metals, caffeine, proteins, surfactants, dyes and pharmaceutical actives and excipients and combinations thereof.

[089] Particular examples of metals include, without limitation, gold, platinum and heavy metals such as chromium, cobalt, nickel, copper, zinc, arsenic, selenium, silver, cadmium, antimony, mercury, thallium and lead or combinations thereof.

[090] Particular examples of other materials to be purified include, without limitation, polycyclic aromatics, chlorinated and brominated dibenzodioxins and dibenzofurans, chlorinated biphenyls, lindane, dichlorodiphenyltrichloroethane (DDT) and other similar hydrophobic xenobiotics.

[091 ] Particular examples of radioactive elements to be purified are metals, salts, metal salts, radionuclides, actinides, lanthanides, and combinations thereof. In more specific embodiments, the radioactive elements in the solutions include radionuclides, such as thallium, iridium, fluorine, americium, neptunium, gadolinium, bismuth, uranium, thorium, plutonium, niobium, barium, cadmium, cobalt, europium, manganese, sodium, zinc, technetium, strontium, carbon, polonium, cesium, potassium, radium, lead, actinides, lanthanides and combinations thereof. In more specific embodiments, the radioactive elements to be purified from various environments include, without limitation, americium(lll), actinide(lll), actinide(IV), thallium(IV), plutonium(IV), neptunium(V), uranium(VI), strontium(ll), technetium(VII), and combinations thereof.

[092] In more specific embodiments, the radioactive elements include, without limitation, thallium-201 , iridium- 192, fluorine- 18, americium-241 , americium- 243, neptunium-237, Gd- 153, niobium-93, barium-133, cadmium-109, cobalt- 57, cobalt-60, europium-152, manganese- 54,sodium-22, zinc-65, technetium- 99, strontium-90, thallium-204, carbon- 14, polonium 210, cesium- 137, and combinations thereof.

[093] Particular non-limiting examples of chlorates include ammonium chlorate, barium chlorate, cesium chlorate, fluorine chlorate, lithium chlorate, magnesium chlorate, potassium chlorate, rubidium chlorate, silver chlorate, sodium chlorate, and combinations thereof. Non-limiting examples of perchlorates include ammonium perchlorate, barium perchlorate, cesium perchlorate, fluorine perchlorate, lithium perchlorate, magnesium perchlorate, perchloric acid, potassium perchlorate, rubidium perchlorate, silver perchlorate, sodium perchlorate, and combinations thereof. In some embodiments, perchlorates may have similar sorption profiles to various radioactive elements, such as pertechnetate (Tc(VII)).

[094] The present invention may also be utilized to capture various organohalogens from solution. Organohalogens generally refer to organic compounds that include one or more halogen groups. In some embodiments, the organohalogen is an organochloride. In some embodiments, the organohalogen is polychlorinated biphenyl (PCB). In some embodiments, the organohalogen is a halogenated flame retardant. In further embodiments, the organohalogens include, without limitation, chloromethanes, dichloromethanes, trichloromethanes, tetrachloromethanes, bromomethanes, bromoalkanes, bromochloromethanes, iodoalkanes, iodomethanes, organofluorines, organochlorines, acyclic organohalogens, cyclic organohalogens, and combinations thereof.

[095] In some embodiments, the dispersion formed with the magnetic material within a dispersant solution comprises a dispersion of carbon-based nanomaterials selected from graphene oxide (GO), reduced graphene oxide (rGO) or carbon nanotubes (CNTs). The dispersibility of such CNMs is intimately related to their dimensions, charge and hydrophobicity. In embodiments, the aqueous dispersibility of the carbon-based nanomaterials in the dispersion solution can be controlled by light via the photoisomerisation of a photoswitchable surfactant molecule adsorbed to the surface of these materials. One photoswitchable surfactant molecule comprises a cationic azobenzene photosurfactant. The incorporation of a photoswitchable surfactant molecule into a GO, rGO or CNT dispersions allows these dispersions to be separated and redispersed on command utilizing UV radiation, preferably at 365 nm. Here, the surfactant molecules change from the trans to the cis isomer increasing aqueous solubility and in turn, alters their adsorption affinity for the GO and rGO sheets such that the ratio of free to adsorbed surfactant molecules changes significantly, allowing for reversible phase separation of the colloids. These effects present a unique method for controlling the dispersion behaviour of two-dimensional nanomaterials using light as a clean and low energy external stimulus.

EXAMPLES

[096] The formation of noncovalent, magnetic graphene oxide (GO) materials was explored using three magnetic materials: Fe2O3 microparticles, Fe2O3 nanoparticles, and magnetic surfactants.

MATERIALS

[097] Fe ₂ O ₃ (iron(lll) oxide), iron(ll) sulfate heptahydrate, iron(ll) dichloride tetrahydrate, phosphoric, sulfuric, and hydrochloric acids, potassium hydroxide, and potassium permanganate (all 99% or greater) were obtained from Sigma and used without further purification. Fe2O3 microparticles were prepared by grinding sintered Fe ₂ O ₃ pieces (2-10 mm) in a pestle and mortar. The resulting powder was characterized using light microscopy and powder X-ray diffraction measurements. Hydrogen peroxide solution and aqueous ammonia (30% and 28% w/w in water, respectively) were from ChemSupply (Australia) and used as received. Sodium tetrachloroaurate and sodium citrate (>99%, reagent grade) were from Sigma. The magnetic surfactants 1 -decyl-3-methyl imidazolium tetrachloroferrate (mim) and dodecyltrimethylammonium tnchloromonobromoferrate (DTA) were synthesized and purified as described previously. Mica disks used as substrates for atomic force microscopy (AFM) imaging were from ProSciTech (Thuringowa, OLD, Australia) and were freshly cleaved before use. [098] GO was synthesized from graphite flakes (Sigma, +100 mesh) using a variation on the Hummers method described in Marcano et al. The graphite powder (1 g) was dispersed in 1 13 mL of a 9:1 ratio of concentrated sulfuric and phosphoric acids. This mixture was then stirred while potassium permanganate (6 g) was added gradually. The temperature was elevated to 50 ^< C, and the reaction was left to stir overnight. The resultant orange/brown mixture was then left to cool to room temperature and poured over ice (ca. 300 mL) with approximately 1 mL of 30% H2O2. Large particles were removed from the crude reaction mixture by filtration, and the filtrate was then centrifuged at 6000 rpm for 1 h; the supernatant liquid was discarded and replaced with distilled water. This process was repeated several times, and the clean GO obtained was then dried at 45 ^« C.

[099] Fe ₂ O ₃ nanoparticles were prepared from iron(ll) sulfate heptahydrate (green vitriol) following the procedure of Chen et al.45 Iron(ll) sulfate heptahydrate (8.36 g) was dissolved in 300 mL of water to create a 0.1 M solution. The addition of 30% (w/w) hydrogen peroxide solution (10 mL) rendered the mixture an intense orange colour, and this was then heated to 80 ^< C. In a separate vessel, 50 mL of aqueous ammonia (2.8% w/w) was heated to 60 and then mixed rapidly with the orange iron s olution. The mixture was allowed to stir for 20 min before the reddish precipitate was collected, washed five times by centrifugation, and redispersed in ultrapure water. A small volume of iron(ll) dichloride solution (0.6 mL, 0.1 M) was added to the washed suspension as a catalyst, and the solution was adjusted to pH 4 using hydrochloric acid and heated to boiling for 5 h under reflux and gentle stirring. Upon completion of the reaction, gentle centrifugation (1000 rpm, 2 min) was used to remove any large particles, leaving a dark red solution of Fe2O3 nanoparticles.

[0100] Gold nanoparticles were synthesized using the method of McFarland, A. D.; Haynes, C. L.; Mirkin, C. A.; Van Duyne, R. P.; Godwin, H. A. Color My Nanoworld. J. Chem. Educ. 2004, 81 , 544A-544B, the contents of which are to be understood to be incorporated into this specification by this reference. Briefly, a 1 imM solution of sodium tetrachloroaurate was heated to boiling with vigorous stirring, and to this, sodium citrate solution (38 imM) was added. The mixture was left to boil under stirring until the solution turned a deep red colour, indicating the presence of nanoparticles, after which the solution was allowed to cool.

METHODS

[0101 ] Samples shown throughout were made up to a standard volume of 1 .5 mL. Adjustments to solution pH were made with either hydrochloric acid or potassium hydroxide and measured with a calibrated pH meter. Prior to analysis, all samples were equilibrated for at least 24 h unless otherwise stated.

[0102] Magnetic response was assessed by placing a strong, permanent magnet beside or underneath the vials containing the samples. The magnets used were composed of sintered NdFeB (M35) in a 100 μιη thick nickel casing (Jaycar Electronics, Springvale, VIC, Australia). The magnets were cylindrical in shape with diameter 19 mm and length 28.2 mm, and the magnetic field intensity at the surface was ca. 1 .2 T.

[0103] Magnetization data were collected for dried samples, which were placed in sealed polypropylene tubes and mounted inside a plastic straw for measurements in a magnetometer equipped with a superconducting quantum interference device (SQUID, MPMS XL, Quantum Design, San Diego, CA, U.S.A.).

[0104] UV-visible spectrophotometry measurements were carried out using a Cary 60 instrument from Agilent Technologies. The supernatant of each sample was analyzed across a 200-800 nm wavelength range in clean quartz cuvettes. For samples in which the concentration of surfactant was the point of interest, the GO was centrifuged down to ensure that the spectra obtained were representative of free surfactant only. Correspondence of the measured absorbance values to prepared calibration curves for GO and both magnetic surfactants were performed to obtain the post-adsorption concentrations of each material.

[0105] Dynamic light scattering measurements of Fe2O3 nanoparticles were made using a Brookhaven ZetaPlus instrument. The autocorrelation function of light scattered by the sample from a 30 mW red diode laser was fitted to obtain particle diffusion coefficients, D, and translated into apparent particle hydrodynamic radii, Rh, using the Stokes- Einstein equation. Zeta potential measurements were made using a Brookhaven NanoBrook Omni instrument.

[0106] Atomic force microscopy (AFM) imaging was performed using a JPK Nanowizard 3 AFM in AC (intermittent contact) mode using Bruker NCHV model cantilevers, which had nominal resonant frequencies of 340 kHz and spring constants of 20-80 N/m. Images were obtained with a set-point force of <1 nN and "flattened" only by subtraction of a straight line from each scan line.

CHARACTERIZATION OF MATERIALS

[0107] Graphene oxide (GO) was synthesized from graphite particles using an improved Hummer's method. The shape and size of the resulting particles were then characterized by atomic force microscopy (AFM) and were found to be almost entirely 1 nm in thickness, indicative of monolayer GO. The lateral dimensions were typically several micrometers, with the average flake sizes characterized previously by dynamic light scattering. Optical properties were characterized by UV-visible spectrophotometry, which was also used to quantify GO concentrations throughout, having been initially calibrated gravimetrically. The spectra display the characteristic features expected for GO: a distinct absorption maximum at approximately 230 nm corresponding to π→ TT* transitions and a shoulder at around 300 nm, believed to be n → ττ ^* transitions.

[0108] Iron oxide (Fe2O3) particles were produced by milling sintered Fe2O3 pieces that had been produced by a thermal treatment of Fe ₂ O ₃ precipitated powder. These particles were characterized using atomic force and optical microscopies, superconducting quantum interference device (SQUID) magnetometry, and powder X-ray diffraction (PXRD), demonstrating their morphology, size, magnetization, and apparent dominant crystal structure, respectively.

EXAMPLE 1 - Graphene Oxide and Fe ₂ 0 ₃ Microparticle Systems. [0109] Figure 1 b indicates a large difference in isoelectric points for GO (pH ~ 0) and Fe ₂ O ₃ (pH ~ 7). This large difference provides a wide window in which the two materials exhibit opposite surface charges and thus could be expected to experience Coulombic attraction. It is noteworthy that the pH-dependent zeta potentials and isoelectric points for hematite and maghemite are essentially identical. Previous work at gold-water interfaces has indicated that the approach of assembly via charge attraction works well for GO, as have various layer-by-layer assembly studies.

[01 10] To assess the feasibility of magnetically extracting the GO using charge- assembled magnetic materials, the behaviour of GO/Fe ₂ O ₃ microparticle mixtures were studied as a function of pH (Figure 2). AFM imaging of GO on the particle surfaces (Figure 2a,b) indicated that it adsorbs conformally (i.e., lying flat over the particle surfaces). The particles used had an average radius ca. 2 μιη, calculated from the projected area (Figure 2c,d). These Fe ₂ O ₃ particles were only kinetically stable in water and settled over time.

[01 1 1 ] In order to exemplify the charge-mediated adsorption of GO onto the Fe ₂ O ₃ particles, Figure 2e,f shows the same set of vials at different pH conditions. In all cases, the two vials both contain GO (0.2 img/mL), but in each frame the right-hand vial also contains 50 mg of Fe ₂ O ₃ microparticles. At high pH (Figure 2e) no change in the colour of the dispersion is seen when Fe ₂ O ₃ was added and the sample placed by a magnet, indicating that all of the GO remained dispersed while the Fe ₂ O ₃ is captured by the magnet. However, once the dispersion had been marginally acidified (pH < 5), the GO and Fe ₂ O ₃ coflocculated and were both captured by the magnet (Figure 2f). This serves as a clear indication as to the importance of pH and charge in controlling the surface chemistry in these systems. Although iron oxides become more soluble at low pH, Fe ₂ O ₃ dissolution is vanishingly slow at the pH values dealt with here. Over the course of several months, no change in the particle size of the iron oxide was noted when maintained at pH 2.5.

[01 12] To explore this effect quantitatively, a range of pH and concentration ratios were explored. At an acidic pH of 2.8, the GO and Fe ₂ O ₃ coprecipitate to form a magnetically responsive "network" that settles out quickly, leaving clear water as the supernatant (Figure 3a). It is notable that this coflocculated material appears to be a gelled particle network, as the volume occupied is many times greater than the dry volume occupied by the constituent Fe2O3 and GO. The higher the concentration of GO, the greater the volume of the "network".

[01 13] Conversely, at high pH, no precipitation of the GO was seen and it remained dispersed in the water, whereas the Fe ₂ O ₃ settled (Figure 3b). This effect is readily explained by strong charge attraction between the positive Fe2O3 surfaces and the negative periphery of the GO sheets, due to the dissociated carboxylate groups. The inventors therefore posit a structure for this network whereby GO sheets link Fe ₂ O ₃ particles, acting as a bridging flocculant, in much the same way that many polyelectrolytes are used as particle flocculants.

[01 14] At pH 1 1 .5, the GO and Fe ₂ O ₃ are both strongly negatively charged and hence experience a mutual repulsive electrical double-layer force. At pH 2.8, the GO maintains its negative surface charge; however, the Fe ₂ O3 particles are strongly positively charged from protonation of their surface oxide groups. It can be seen that effectively all of the GO has been removed from the bulk solution in the low pH samples, while for the high pH samples, the majority of the GO is still present (Figure 3c). Some removal appears to have occurred for the higher concentrations of GO at high pH, though this may be due to a self-salting effect where the high volume fraction causes overlap of the electrical double-layers of the GO sheets, resulting in a locally increased counterion concentration that acts to "salt out" the sheets via charge screening. From Figure 3 the adsorption capacity of the microparticles is estimated to be around 0.23 mg of GO per mg of Fe ₂ O ₃ . This capacity is surprisingly high considering the density and surface area difference between the two materials.

[01 15] Exploring the effects of pH indicates a relatively narrow window in which the transition between complete flocculation and full dispersion is seen (Figure 4). The results show that at pH values of 6 and below, close to pure water can be retrieved. At pH 6.96, marginally below the isoelectric point of Fe ₂ O3, the majority of the GO had adsorbed, leaving a low concentration in dispersion and a slightly smaller GO/Fe ₂ O ₃ network. Just above the isoelectric point of Fe ₂ O ₃ (pH = 7.48), some settling and adsorption is apparent; however, most of the GO remains dispersed. At pH values of 8 and above, the GO stays almost fully dispersed. The apparent decrease in dispersed GO seen at pH 12.00 is most likely a result of the GO starting to become chemically reduced, which is known to occur at high pH.

EXAMPLE 2 - Graphene Oxide and Fe2C>3 Nanoparticle Systems.

[01 16] Although microparticles are appealing for reversible capture and dispersion of GO due to the large magnetic force they experience, they suffer from the problem of settling, which means that energy is required to effectively disperse them for capture of GO. At small scales and in stirred, flowing, or otherwise agitated systems (such as wastewater settling tanks, pipe flow, etc.) this does not represent a significant problem. However, deployment at larger scales requires a magnetic capture agent that is colloidally stable - i.e. a material that will not settle over time but will remain dispersed due to Brownian collisions with solvent molecules.

[01 17] One approach is to explore the deployment of small Fe ₂ O ₃ nanoparticles that form thermodynamically stable dispersions. These nanoparticles were synthesized from iron(ll) sulfate heptahydrate using the procedure described by Chen, X.-Q.; Wu, S.-B.; Cao, R.-B.; Tao, J.-S. Preparation and Characterization of Nanosized Hematite Colloids Using Green Vitriol as Ferrum Source. J. Nanomater. 2014, 749562/1 -8, the contents of which should be understood to be incorporated into the specification by this reference. Dynamic light scattering (DLS) results indicate that the nanoparticles are most stable at low pH (2-4) where their surface charge is strongly positive. Their apparent effective diameter at this point is ca. 40 nm, whereas the particles readily attract one another and form larger floes at pH 7-9 (Figure 5a), close to the isoelectric point of the particles where the surface potential drops below the 30 mV required for colloidal stability.

[01 18] AFM imaging indicated that the particles were mostly below 20 nm in size, with surprisingly low polydispersity (Figure 5b, c). A mixture of Fe ₂ O ₃ nanoparticles with aqueous GO dispersion at pH 3 resulted in a flocculated material that was highly responsive to an external magnet (Figure 5d).

[01 19] To confirm specific adsorption of the Fe ₂ O ₃ nanoparticles to the GO sheets, the GO-nanoparticle composite material formed at pH 3 was imaged using AFM (Figure 6a). It is clear that the Fe2O3 nanoparticles show high and specific affinity for the GO, forming large clusters on the surfaces of the sheets. This suggests that, again, electrical double-layer interactions drive adsorption and flocculation processes, as strikingly few nanoparticles were seen on the mica substrate. Moreover, it is clear from Figure 6b that there is a highly specific pH range within which optimal interaction between the nanoparticles and GO occurs, with no adsorption being evident in the high pH sample. This is again in line with expectation from the surface charging behaviour of Fe ₂ O ₃ and GO and echoes the characteristics of the Fe2O3 microparticle dispersions.

[0120] From Figure 6, the adsorption/flocculation capacity of the Fe ₂ O ₃ nanoparticles is estimated as around 1 .5 mg of GO per mg of Fe ₂ O ₃ at pH 3. As expected due to the significantly greater surface area of the nanoparticles, they are significantly more effective flocculants per unit mass than microparticles.

[0121 ] Despite the significant difference in lateral dimensions between the Fe ₂ O ₃ nanoparticles and the microparticles studied above, the mechanism of coflocculation is likely similar, whereby the nanoparticles act to bridge the GO sheets, causing a 3D network to form. The number of effective cross-links between sheets would therefore be related to the ratio of nanoparticles to GO, with higher nanoparticle loadings favouring a denser network structure due to a larger number of "linkages". This appears to be realized (Figure 6c) where increasing nanoparticle loading for the same amount of GO results in a smaller, denser flocculated layer. By nanoparticle:GO ratios of >3:1 , free (unflocculated) nanoparticles are evident in the supernatant above the flocculated GO.

[0122] It is found therefore that an equal mass ratio of Fe2O3 nanoparticles to GO is ideal for recovery. However, a particularly interesting effect occurs at higher nanoparticle loadings, whereby restabilization of the GO was observed. Dispersions thus formed were indefinitely stable, indicating a surface potential above the 30 mV generally required for colloidal stability. The Inventors posit that this effect occurs due to overcharging of the particle surfaces, whereby the Fe2O3 nanoparticles act as a dispersant for the GO, increasing the values. A similar effect has been seen whereby small, highly charged nanoparticles stabilize larger colloids with low, opposite surface charges by adsorption, with the effect explained as a balance of van der Waals attractions and electrical double-layer effects. Similarly, the "supercharging" of anionic silica by adsorption of anionic surfactants has been noted, suggested to be entropic in origin. The concentration dependent flocculation/redispersion effects for nanoparticle adsorption on GO seen here are to the inventor's knowledge the first such example for such a nanomaterial and thus may be advantageous in designed systems for bulk solution deployment of GO.

EXAMPLE 3 - Graphene Oxide and Magnetic Surfactant Systems.

[0123] There has been a recent surge of interest surrounding the use of magnetic surfactants as stabilizers, due to their ability to form micelles, microemulsions, and (macro)emulsion droplets as soft colloids with magnetic response. They have also been employed in the field-induced control of biomacromolecules and, more recently, silica particles. However, as yet, their application for the magnetic recovery of nanomaterials remains unexplored.

[0124] The potential of two such magnetic ionic liquids, 1 -decyl-3- methylimidazolium tetrachloroferrate (mim) and dodecyltrimethylammonium tnchloromonobromoferrate (DTA), were investigated as molecular alternatives to the microparticle and nanoparticle systems examined in the previous examples.

[0125] Adsorption experiments were carried out at pH 5.5 to ensure that the GO had a significant negative surface charge. Interestingly, the behaviour of the magnetic surfactant/GO systems is similar to that of the Fe ₂ O ₃ systems (Figure 7a,b), and the surfactant appears to serve two roles: first as a magnetic material for field-induced recovery of the GO but also as a flocculant. The surfactants were chosen for their positively charged head-groups, as these should experience a strong electrostatic attraction to the dissociated carboxylate groups on GO. To explore these characteristics further, adsorption isotherms for each surfactant on GO were determined (Figure 7c). It is seen that mim has a relatively low affinity for GO, adsorbing at less than 5 mmol of surfactant per gram of GO even at high surfactant loadings. DTA on the other hand adsorbs with moderate affinity at high surfactant concentrations, though we were unable to observe adsorption saturation (asymptotic flattening of the isotherm) due to the experimental uncertainties associated with measuring small changes in absorbance at large concentrations.

[0126] The large difference in affinity aids interpretation of the magnetic response of the two systems (Figure 7a,b). The DTA-GO system shows significant flocculation, forming a loosely aggregated GO matrix, which is moderately attracted to the magnet. Conversely, the mim-GO system forms a much denser flocculated material of which only a small amount responds to the magnet. Thus, it is clear that the more strongly adsorbed DTA surfactant is the more effective of the two for magnetic recovery of GO.

[0127] The inventors theorize that there are two possible reasons for the flocculation behaviour encountered for the surfactants here: (a) the surfactant adsorbs strongly to the anionic carboxylate groups around the GO periphery, thus neutralizing the charge and causing colloidal instability, or (b) the GO sheets become coated with a large amount of surfactant and are therefore rendered hydrophobic and attract one another. Given the high concentrations of surfactant required for significant adsorption, the second explanation seems much more likely. Similar flocculation behaviour is seen for GO when reduction due to increased pH occurs, which also supports this hypothesis.

[0128] As it is the counterion that is paramagnetic, rather than the surfactant ion itself, this raises an important question about the nature of the adsorption and magnetic response. Clearly the fractional level of counterion dissociation is a key parameter, and this was previously determined using small-angle neutron scattering as 0.73 and 0.81 for mim and DTA, respectively. These values are high compared to the same surfactant ions with "conventional" bromide or chloride counterions, indicating an increased hydrophilicity of the iron- containing, magnetic counterions. Thus, it is suspected that bound surfactants contribute to the magnetic response of the surfactant-GO materials in two ways. The first is that undissociated, bound surfactants (adsorbed via hydrophobic interactions with the GO sheets or polar interactions) will directly respond to the magnetic field. The second is that dissociated, bound surfactant ions still retain their counterions in a diffuse layer near the surfactant-GO interface and that, by magnetic movement of the dissociated counterions, the surfactant-GO complex is osmotically "dragged" with them.

[0129] The relative contributions of these two effects would depend on binding strength and concentration, and modeling studies are underway to understand this further.

EXAMPLE 4 - Magnetic Compression and Dewatering of GO.

[0130] For each of the magnetic materials employed, pH-dependent GO flocculation was observed, and so it is pertinent to more systematically study the magnetic response of the coflocculated materials generated. In particular, the magnetic compressibility of the flocculated material was explored by placing the materials on a strong permanent magnet and recording the volume of the flocculated network as a function of time (Figure 8).

[0131 ] Fe ₂ O ₃ matrix was almost halved over a two day period (Figure 8a-c), indicating significant compression and dewatering. The inventors posit that the GO and iron oxide microparticles initially form a loose matrix, whereby charge interactions link the particles and GO sheets together. This is conceptually similar in nature to particle floes induced by addition of polyelectrolyte but clearly different from depletion induced flocculation. This matrix appears to be composed of loose floes that entrain a large amount of water. There is a moderate initial settling of the floes under gravity to provide a matrix, the volume of which scales with the amount of GO present. Under the additional downward force provided by the magnet, this matrix is densified, excluding a supernatant water phase. The precise nature of this flocculated material could be explored further by, for example, confocal microscopy or freeze-fracture electron microscopy to better understand the internal structure of the floes. [0132] When the same magnetic compression test was performed on systems containing Fe2O3 nanoparticles or the magnetic surfactants, effectively no compaction occurred. It is possible that, in the cases with the nanoparticles and surfactant molecules, the matrix is denser or more strongly bonded and therefore unable to compact further regardless of the magnet being present. However, an examination of the energy experienced per particle due to the magnetic field is also illuminating: a 2 μιη radius Fe ₂ O ₃ particle, typical of the microparticle material, would experience an energy due to the magnet of around 7 x 10 ^"15 J (assuming a typical magnetization of 200 A/m for finely divided Fe2O3). A 20 nm radius particle with the same characteristics however would gain an energy of only 7 ^χ 10 ^"21 J, around the same magnitude at thermal energy, kBT at 298 K, 4.1 ^χ 10 ^"21 J. It is difficult to precisely estimate the energy for the surfactant system, but it is expected to be of a similar magnitude per unit mass as Fe2O3, given the susceptibilities of the surfactants themselves.

[0133] It should also be noted that the mass ratio for magnetic surfactant and nanoparticle experiments (around 1 :1 magnetic material:GO) was significantly less than for the microparticle case (2.6:1 magnetic material:GO), also contributing to the increased effectiveness of microparticles. This does however indicate that by transitioning to nanomaterials where surface area-to-volume ratios are significantly higher, lower loadings can be used to capture GO from dispersion. It is clear that the Fe2O3 microparticles are able to exert a significantly larger force on the GO matrix, and although the nanoparticles and surfactants apply sufficient force in a magnetic field to act against Brownian motion and enable collection of the material with a magnet, they cannot effectively compress the matrix to dewater it. In fact, the compressional strength of the GO-Fe ₂ O ₃ microparticle matrix could be estimated from the applied force due to the magnetic field, although this would require a more carefully controlled magnetic field than was employed in the proof-of-principle measurements shown here. However, it is clear that the "gel" matrix strength is on the same order as the force applied by the microparticles in the field and that the nanoparticles and magnetic surfactants are insufficient to overcome the compressional strength of the GO matrix. [0134] In contrast to previous studies where magnetic particles were found to have no significant compaction effect on dewatering of suspensions, the fairly significant compression seen for the GO-microparticle case indicates that, for the loosely flocculated "gels" produced, magnetic compression is a viable and effective method for dewatering.

EXAMPLE 5 - Magnetic Compression and Dewatering of GO.

[0135] Having demonstrated that low pH conditions facilitate strong adsorption of Fe ₂ O ₃ onto GO, it is pertinent to explore whether this process is reversible. The motivation is that if the captured material can be redispersed effectively, then GO can be used as a recyclable adsorbent for (waste) water treatment. The ability to reuse the adsorbent multiple times offers clear energy, cost, and environmental benefits.

[0136] Figure 9a and b indicates that GO can indeed be reversibly captured with Fe ₂ O ₃ microparticles or nanoparticles by changing the pH of the system. Acidic conditions facilitate the adsorption and capture of the GO, whereas readjusting the pH of the samples to moderately basic conditions serves to release the GO, restabilising the dispersion and allowing selective separation of the magnetic particles from the GO.

[0137] These observations also serve to again indicate that surface charge is the overriding, driving force of the adsorption and dispersion processes for these materials, and van der Waals forces (expected to be comparatively weak for the single-layer GO) are of secondary importance. The approach of using noncovalent, charge based attractions in forming magnetic GO materials has significant advantages over existing magnetic GO composites, because in our case, the GO and the magnetic material are able to be recovered without any compromises to their original properties and, hence, can be recycled and reused multiple times.

EXAMPLE 6 - Magnetic GO colloids capture of gold nanoparticles [0138] Magnetic GO colloids/composite material formed following the procedure and materials set out in Example 1 (Fe ₂ O ₃ microparticles) were used to capture gold nanoparticles of diameter 10 nm from dispersion (Figure 9d).

[0139] By adjusting the pH to 3, the gold surface charge is moderately positive and thus experiences an attractive electrical double-layer force for the GO surfaces which are negatively charged at this pH. The gold is expected to stick selectively to the GO surfaces, as the Fe2O3 is strongly positively charged at this pH.

[0140] Figure 9e demonstrates conclusively from the loss of the characteristic plasmon signature for the gold particles that, after application of the GO-Fe ₂ O ₃ material and magnetic recovery of the matrix, the gold has been entirely removed from dispersion (at least within the detection limits of our experiment). The "blank" experiment where only Fe2O3 particles are used, shown in Figure 9c, demonstrates that GO is essential to this process, as the Fe ₂ O ₃ and gold nanoparticles are both positively charged at this pH.

[0141 ] This serves as a demonstration that noncovalent GO-Fe ₂ O ₃ composite materials can be effectively deployed for adsorption and subsequent removal of nanomaterials from dispersion, with obvious applications in water treatment and nanomaterials processing.

EXAMPLE 7 - Magnetic GO colloids capture of Caffeine

[0142] Magnetic GO colloids/composite material formed following the procedure and materials set out in Example 1 (Fe ₂ O ₃ microparticles) were used to capture from an aqueous solution containing caffeine. Figure 10 shows a partial adsorption isotherm of caffeine on GO, demonstrating adsorption of caffeine to the noncovalent GO-Fe ₂ O3 composite material.

EXAMPLE 8 - Magnetic GO colloids capture of methylene blue dye

[0143] Magnetic GO colloids/composite material formed following the procedure and materials set out in Example 1 (Fe ₂ O3 microparticles) were used to capture methylene blue dye. Figure 1 1 provides a partial adsorption isotherm of methylene blue dye the magnetic GO colloids, demonstrating significant adsorption of methylene blue, even at low concentrations.

[0144] Figure 12 provides a photograph showing magnetic recovery of methylene blue dye using magnetic GO colloids formed following the procedure and materials set out in Example 1 . The illustrated vials 100 and 102 contain the same concentration of methylene blue dye (50 micromolar), but the vial 100 on the left also contains 0.2 img/mL graphene oxide and 0.1 img/mL iron oxide microparticles. The graphene oxide/iron oxide composite material adsorbs the methylene blue in the left vial 100, and the graphene oxide/iron oxide composite material is attracted to the magnet 104.

EXAMPLE 8 - Effect of sparging with C0 ₂ on GO and iron oxide for magnetic capture

[0145] CO ₂ was introduced into a solution of magnetite and graphene oxide using sparging in order to determine the effect on the pH of the solution. The pH of a control solution and a sparged solution were tested after 15 minutes of sparging and 15 minutes of settling.

[0146] After 15 minutes of sparging and 15 minutes of settling, there was clear separation in the sparged sample that was not observed in the control (Figure 13A). This separated sample also showed a strong magnetic response that the control did not (Figure 13B). When the pH of the both solutions were tested, the control sample had a pH of 10.2, while the sparged sample had a pH of 5.0, due to acidification through formation of carbonic acid. The sparged sample flocculating and the control sample remaining dispersed is therefore expected and consistent with the zeta potential data, as at pH 5.0 electrostatic attraction between the positive magnetite and the negative GO should cause adsorption and flocculation, while at pH 10.2, both materials are negatively charged and repel each other, therefore being unable to flocculate and remaining disperse.

[0147] This experiment showed that it was possible to control the flocculation of magnetite and graphene oxide through sparging with CO ₂ . Sparging with CO ₂ drops pH sufficiently to heteroflocculate GO and iron oxide for magnetic capture [0148] The use of CO ₂ to raise the pH of the solution removes the need for the large scale use of chemical reagents for that purpose that may further pollute the water or cause salt interferences as well as overcoming the need for energetically expensive mixing processes, as sparging naturally mixes the wastewater as part of the process.

[0149] It should be appreciated that the solution pH can be increased by sparging with another gas including argon, nitrogen or air, and ammonia can be added to further increase the pH to -10.

EXAMPLE 9 - Use of a photoswitchable surfactant to reversibly control the stability of carbon nanomaterials

[0150] A photoswitchable surfactant (azo-TAB) was used to reversibly control the stability of carbon nanomaterials (CNMs), specifically demonstrating the effect for graphene oxide (GO), reduced graphene oxide (rGO) and carbon nanotubes (CNTs) (see Figure 14a to f).

[0151 ] Firstly, the photo-induced phase separation of rGO was investigated. When GO is chemically or physically reduced to rGO, its innate hydrophobicity increases due to the loss of hydrophilic functional groups, and rGO is therefore generally only metastable in aqueous conditions. When adding a cationic surfactant that would be expected to adsorb readily to the negative charge sites on rGO thereby reducing its effective surface potential and increasing its hydrophobicity, it is thus unsurprising that the dispersion fully flocculates, even at very low concentrations of azoTAB in both trans and cis isomerisation states (Figure 15a and b). As surfactant concentration is increased to the CMC of azoTAB (1 imM in the trans state and 2 imM in the cis state) and beyond, the rGO remains fully dispersed. It can also be observed quantitatively in the azoTAB/rGO adsorption isotherm (Figure 15c) that azoTAB adsorption drops off significantly at and past these concentrations. This unexpected result indicates that steric stabilisation by further surfactant adsorption is unlikely to be the effect causing stability of the rGO dispersions at high surfactant loadings. Instead, the highly positive zeta potentials (>30 mV) obtained (Figure 15a) indicate re-stabilisation by charging of the colloids, presumably via surfactant adsorption through van der Waals or pi-stacking interactions. The reduction in adsorbed amount is curious, and may reflect a reconfiguration of the partitioning within the system when moving to a 3-state equilibrium (adsorbed, monomeric and micellised surfactant).

[0152] The difference in surfactant concentration of the two isomers required to induce flocculation of the rGO provides an accessible window in which the dispersibility of the material can be reversibly controlled by exploiting the photo- induced switch of the azobenzene group. By choosing a concentration between 1 and 2 imM (e.g. 1 .2 imM), it can be seen that it is possible to reversibly flocculate and redisperse the rGO by using light (Figure 15d). It is noteworthy that significantly longer illumination times were required to destabilise the colloid when the rGO was present compared to Figure 15b, where the surfactant solutions were irradiated before the rGO was added, due to added optical density of the rGO dispersion. Irradiation with blue light (ca 450 nm) then served to restabilise the rGO, and although not required, this process could be accelerated by short bursts of ultrasonication. The retained stability of the dispersion after being left overnight shows that the restabilisation is a direct result of the isomerisation state of the surfactant.

[0153] Another hydrophobic CNM in the form of carbon nanotubes (CNTs) can be dispersed and recovered using the same method as for rGO above. CNTs do not typically disperse in water, but have been found to do so with the aid of surfactant molecules and sonication. In the case of azoTAB, the CNTs are found to disperse well when the surfactant is in the more surface active trans configuration, with dispersions achieved at sub-CMC concentrations (0.5 imM), whereas it can be seen that the cis isomer does not effectively disperse the CNTs at concentrations below 1 .5 mM and they settle out after only 1 hour (Figure 15e). However, the highly polydisperse nature of the multiwalled CNTs used here meant that the separation was not as 'clean' (i.e. complete) as for rGO, and therefore further studies were not performed. [0154] A very different behaviour is seen for graphene oxide, a hydrophilic and water-dispersible CNM. Here, the dispersion stability is innately much higher and it is therefore more challenging to effect destabilisation through surfactant adsorption. Although significant adsorption was seen in the isotherm when mixing GO with azoTAB at pH 3 (Figure 16c), at the surfactant concentrations where rGO flocculated, GO is instead stable. Notably, a significant difference can be observed in the level of adsorption between not only the trans and cis forms of azo- TAB, but also when the photosurfactant solution is irradiated pre- adsorption GO when compared to UV irradiation of the already-adsorbed trans dominated azoTAB. This result is in line with expectation from the larger dipole moment and increased solubility of the cis isomer, which would cause a shift in the dynamic equilibrium of the system such that the surfactant partitions more into the bulk solution. The difference in adsorption from isomerisation pre- versus postadsorption could indicate that adsorption of the surfactant to the GO in the trans form means that molecules are somewhat stabilised towards desorption.

[0155] When examining the same systems in basic conditions (pH 10), it was found that higher concentrations of azoTAB (> 0.3 imM) were required to flocculate the GO (Figure 16a,b). It can also be seen that flocculation occurred less readily when the surfactant was in the cis state (Figure 16b), reinforcing the notion that light can be used to control the aggregation state of these systems. Surprisingly, when the same isotherm was determined at pH 10 (Figure 16c, hollow symbols) there was no significant change in the adsorbed amount of surfactant, suggesting that adsorption is not driven solely by electrostatics, but likely also interactions such as van der Waals or ττ-stacking. Investigation of the surface charge on the GO sheets showed that increasing the concentration of azoTAB resulted in a gradual increase in the zeta potential of the system (Figure 16d). This accounts for the destabilisation of the dispersion due to insufficient electrical double-layer interactions between sheets. However, given that the zeta potential of the GO increases only marginally with added azoTAB, it is clear that charge-based interactions are not the predominating mode of adsorption in these systems, and that the interactions due to ττ-stacking, van der Waals and hydrophobicity are more significant. [0156] To explore the aggregation mechanism and morphology induced by interaction of azoTAB with GO and rGO, small angle neutron scattering (SANS) was employed, whereby the effects of surfactant concentration and irradiation on assembly were analysed. The scattering seen (Figure 17a,b) is characteristic of fractal aggregation, where intensity increases rapidly at low scattering vector q (an inverse length scale), indicating the formation of large structures with poorly defined morphology; thus a mass fractal model was used to fit these data. It can be seen that scattering intensity is greater for GO than rGO systems, and also greater for trans than cis (Figure 17a,b). Given the previous results of the isotherms this is not surprising, as the surfactant molecules were found to have a higher affinity for GO than rGO, and are responsible for the majority of the scattered intensity due to their greater contrast than the CNMs.

[0157] By exploring systems where the photosurfactant concentration is changed as well as its isomerisation state (Figure 17a,b), it becomes clear that the level of surfactant adsorption is the key process driving flocculation in these systems. Differences in the fitting parameters from the mass fractal model used to quantify these data indicate that the spatially inhomogeneous floes become more compacted with increased surfactant loading. Fractal aggregates are known to be metastable, as at first they possess a large surface area to volume ratio and then undergo a relaxation towards a more stable configuration. By observing these GO/rGO surfactant systems over time, it can be seen that this is indeed the case (see insets to (Figure 17a,b), as the floes eventually settle into a more condensed network at the bottom of their vessels; similar compaction over time of fractal aggregates has been observed recently in yttrium aluminium garnet systems also analysed by SANS.

[0158] In summary, by incorporating a photosensitive surfactant molecule with a cationic head-group into aqueous carbon nanomaterial systems, it is possible to control the dispersion state of these materials using only light as a clean and low energy external stimulus. Using photoisomerisation to subtly shift the equilibrium between free and adsorbed surfactant, the carbon-based materials can be reversibly dispersed and flocculated. Crucially, by exploring a wide range of conditions including concentration and pH effects, it becomes clear that charge is not the only factor at play, and that surfactant adsorption occurs also via van der Waals type interactions. The photo-modulated flocculation of CNMs provides a facile means of recovering these materials from solution, enhancing opportunities for their application, processing and deployment in aqueous systems.

FURTHER CAPTURE MATERIALS

[0159] While the examples described herein refer to the removal of gold caffeine and methylene blue dye from water, magnetic GO colloid particles of the present invention are contemplated to be able to capture (adsorb) other materials, such as heavy metal ions including selenium, mercury, chromium, lead, cadmium, copper, zinc and arsenic. Adsorption of this type is taught in prior publications such as International patent publication No. WO2014/094130 for GO type particles, albeit having a different binding nature to the complex of the present invention.

[0160] Furthermore, international patent publication No. WO2012170086A1 teaches that graphene oxide is suitable for sorption of various materials from an environment, including radioactive elements, chlorates, perchlorates, organohalogens, and combinations thereof. In particular, WO2012170086A1 teaches that graphene oxide and functionalised graphene oxides are suitable for sorption of:

• radioactive elements comprising radionuclides selected from the group consisting of thallium, iridium, fluorine, americium, neptunium, gadolinium, bismuth, uranium, thorium, plutonium, niobium, barium, cadmium, cobalt, europium, manganese, sodium, zinc, technetium, strontium, carbon, polonium, cesium, potassium, radium, lead, actinides, lanthanides and combinations thereof. In more specific embodiments, the radioactive elements include, without limitation, americium(lll), actinide(lll), actinide(IV), thallium(IV), plutonium(IV), neptunium(V), uranium(VI), strontium(ll), technetium(VII), and combinations thereof. In more specific embodiments, the radioactive elements include, without limitation, thallium-201 , iridium- 192, fluorine- 18, americium-241 , americium-243, neptunium-237, Gd- 153, niobium-93, barium-133, cadmium-109, cobalt-57, cobalt-60, europium-152, manganese- 54,sodium-22, zinc-65, technetium-99, strontium-90, thallium- 204, carbon- 14, polonium 210, cesium- 137, and combinations thereof; or

• chlorates selected from the group consisting of ammonium chlorate, barium chlorate, cesium chlorate, fluorine chlorate, lithium chlorate, magnesium chlorate, potassium chlorate, rubidium chlorate, silver chlorate, sodium chlorate, and combinations thereof; or

• perchlorates selected from the group consisting of ammonium perchlorate, barium perchlorate, cesium perchlorate, fluorine perchlorate, lithium perchlorate, magnesium perchlorate, perchloric acid, potassium perchlorate, rubidium perchlorate, silver perchlorate, sodium perchlorate, and combinations thereof; or

• organohalogens comprising at least one of polychlorinated biphenyls (PCB) and halogenated flame retardants.

[0161 ] Yet further, the following publication demonstrates that GO is suitable for adsorption and/or capture of the following ions, compounds and materials:

• Heavy metal and metalloid ions including:

o cadmium and cobalt [Zhao, et al. Few-Layered Graphene Oxide

Nanosheets As Super Sorbents for Heavy Metal Ion Pollution

Management. Environ. Sci. Technol. 201 1 , 45, 10454-10462];

o Cu ²⁺ , Zn ²⁺ , Fe ³⁺ , Pb ²⁺ and Cr ³⁺ [Yuan, et al. Poly(amidoamine)

Controlled Graphene Oxide as an Efficient Adsorbent for Heavy Metal

Ions. Polym. Chem. 2013, 4, 2164-2167];

o Arsenic [ Zhang et al. Graphene Oxide/Ferric Hydroxide Composites for

Efficient Arsenate Removal from Drinking Water. J. Hazard. Mater.

2010, 182, 162-168; and Chandra et al. S. Water-Dispersible

Magnetite-Reduced Graphene Oxide Composites for Arsenic Removal.

ACS Nano 2010, 4, 3979-3986.];

o Cr(VI) [Jabeen et al. B. Enhanced Cr(VI) Removal using Iron

Nanoparticle Decorated Graphene. Nanoscale 201 1 , 3, 3583-3585.]; o Selenium [ Fu et al. Water-Dispersible Magnetic Nanoparticle-Graphene

Oxide Composites for Selenium Removal. Carbon 2014, 77, 710-721 ]; o Mercury [Chandra et al. Highly Selective Adsorption of Hg ²⁺ by a Polypyrrole-Reduced Graphene Oxide Composite. Chem. Commun. 201 1 , 47, 3942-3944; and Sreeprasad et al. Reduced Graphene Oxide- Metal/Metal Oxide Composites: Facile Synthesis and Application in Water Purification. J. Hazard. Mater. 201 1 , 186, 921 -931 .];

• Ionic dye molecules [Deng et al. Liang, J. Simultaneous Removal of Cd(ll) and Ionic Dyes from Aqueous Solution using Magnetic Graphene Oxide Nanocomposite as an Adsorbent. Chem. Eng. J. 2013, 226, 189-200.]; and

• Antibiotics [Gao, et al. Adsorption and Removal of Tetracycline Antibiotics from Aqueous Solution by Graphene Oxide. J. Colloid Interface Sci. 2012, 368, 540- 546];

[0162] It should be understood that the contents of each of these publications is incorporated into this specification by their reference. It is to be understood that the magnetic GO colloid particles of the present invention, and the present invention is intended to encompass all such applications.

CONCULSION

[0163] Each of the three magnetic materials studied: Fe2O3 microparticles, Fe ₂ O ₃ nanoparticles, and magnetic surfactants, was found to coflocculate GO at acidic pH, resulting in materials that could be captured using an external magnetic field.

[0164] An interesting effect was found with Fe ₂ O ₃ nanoparticles, whereby low concentrations resulted in flocculation of GO and higher concentrations caused restabilisation, most likely by an effective overcharging of the GO surfaces. Such behaviour is not without precedent but has not been noted for carbon nanomaterials before, and it provides a unique route to dispersions with enhanced stability and properties.

[0165] As well as magnetic capture of GO itself, these systems were shown to be effective for the removal of a model nanomaterial, gold nanoparticles, from water. These results demonstrate that the unique surface charging behaviour of aqueous GO systems can be readily exploited and manipulated to reversibly control the assembly of GO with various magnetic materials. Previous studies of the sorption properties of GO and functionalise GO provide evidence that the magnetic GO particle of the present invention can be used to treat and purify a wide range of pollutants or other capture materials desirable to be removed from an aqueous solution.

[0166] Such noncovalent materials show obvious cost and energy benefits compared to bespoke syntheses of controlled magnetic GO. The use of the magnetic surfactants and Fe2O3 nanoparticles minimizes the amount of adsorbate required for recovering GO, but they are much less magneto- responsive when compared to Fe ₂ O ₃ microparticles, which form a network with the GO that is not only captured but also readily compressed by exposure to an external magnetic field. This magnetic dewatering result is more effective than for mineral systems.

[0167] Furthermore, the present invention provides a viable and recyclable colloidal technique for deploying and removing GO from water, making its use in large scale water treatment more cost-effective and practical compared to prior methods which utilised high energy and chemically complex covalent magnetic GO composites for this application.

[0168] Those skilled in the art will appreciate that the invention described herein is susceptible to variations and modifications other than those specifically described. It is understood that the invention includes all such variations and modifications which fall within the spirit and scope of the present invention.

[0169] Where the terms "comprise", "comprises", "comprised" or "comprising" are used in this specification (including the claims) they are to be interpreted as specifying the presence of the stated features, integers, steps or components, but not precluding the presence of one or more other feature, integer, step, component or group thereof.