WO2014080230A1 - Carbon materials and their use - Google Patents

Abstract

Micro-meso or micro-macro porous bead carbons and their use in novel processes for the removal of pesticides or pharmaceutical residues from drinking water are provided. In particular it is applicable to removal of water-soluble compounds that cannot be readily treated using conventional adsorbents. The invention provides a process for treating water contaminated with pesticide or pharmaceutical residues which comprises contacting the contaminated water with a micro-porous, micro-mesoporous or micro-macro-porous carbon with very little heteroatom surface functionality that has been activated in carbon dioxide to minimise the introduction of surface oxygen functionality

Description

CARBON MATERIALS AND THEIR USE

Field of the Invention

This invention relates to micro-meso or micro-macro porous bead carbons and their use in novel processes for the removal of pesticides or pharmaceutical residues from drinking water. In some embodiments it relates to the removal of water soluble compounds that cannot be readily treated using conventional adsorbents. The invention also relates to an improved process for using these carbons. Background to the Invention

Final stage granular activated carbon (GAC) filtration/ Adsorption treatment is now widely practiced for the removal of pharmaceutical and pesticide residues from drinking water reflecting that these compounds are not adequately removed by conventional water treatment processes. However there is a growing problem due to the increasing use of more water soluble compounds which are poorly retained by conventional activated carbons. Metaldehyde is a specific and particularly significant example of this problem.

The typical final stage treatment process also involves the use very large carbon beds which have to be taken offsite for regeneration due to the severity of the conditions required to remove the strongly adsorbed hydrophobic components. This high severity regeneration then results in the destruction of a significant part of the carbon and a reduction in the properties of the regenerated carbon.

Summary of the Invention

We have now found that a micro-meso porous activated carbon based on phenolic resin has a surprisingly high adsorption efficiency for the removal of metaldehyde and other more water soluble hydrophilic contaminants in comparison to conventional activated carbons. Furthermore the unique pore structure of the phenolic resin-derived carbons allows efficient, in situ steam regeneration for both hydrophilic and hydrophobic contaminants. A bead form of the carbon which can exhibit high attrition resistance and mechanical stability also allows the introduction of novel moving bed technologies which can dramatically reduce the size of the adsorption beds and the overall operating costs of the system.

We have shown that a novel finely controlled multi-porous activated carbon material based on the carbon dioxide activation of a phenolic resin derived carbon can be used for the efficient removal of more water soluble pesticides and pharmaceutical residues from water, a separation that cannot be efficiently achieved using conventional activated carbons. The phenolic resin derived carbon can be produced as beads, granules or monoliths and the preferred form will depend on the way in which the separation process is implemented.

The invention provides for a method for treating contaminated water (including, but not limited to, sewage wastes, river water, industrial effluents, hospital effluents etc.) which comprises passing the contaminated water through a bed of the adsorbent material where the bed can comprise a simple packed bed of the beads or granules, a stirred batch reactor, fluidised or expanded bed or moving bed containing the carbon or the carbon produced in a monolithic form where the water passes along the channels of the monolith. The latter structure is less effective in the adsorption of the chemical contaminants due to the hydrodynamic characteristics of the channel structure but satisfactory removal can still be achieved by continuous recycle and this may be beneficial in the treatment of smaller streams.

The invention also provides for the insitu regeneration of the controlled structure carbon. Conventional activated carbons used in final stage GAC filtration/ Adsorption water treatment have to be regenerated ex-situ as the process is effectively re-activation and as such involves very high temperatures. This also results in the partial destruction of the carbon which then has to be continuously replaced. The costs associated with the removal, transportation and re-loading of the carbon during regeneration and the down time whilst it is reactivated (approx. 2 weeks) means that the frequency needs to be minimised and can be between 2 and 6 years depending on the feed water quality and the overall process. This then leads to a requirement for very large beds. To treat around 12 million litre/day typically requires 5 adsorbers, each with a 50m³ capacity equivalent to approximately lOOte of activated carbon.

We have found the phenolic resin derived carbons can be regenerated using steam at ca. 150°C. This allows more frequent in situ regeneration that can dramatically reduce the size of the facility, restricted only by limitations due to bed velocity and residence time. This, combined with the superior mechanical properties of the phenolic resin derived carbons, can also allow the use of different process systems including moving bed and fluid bed systems that reduce the complexity of the plants (shift from multiple to single beds)leading to enhanced overall process efficiency. This can result in a reduction in both capital and operating costs.

In some embodiments the invention provides apparatus for the treatment of water liable to be contaminated with one or more organic contaminants, said apparatus comprising:

a bed of carbon adsorbent configured for through-flow of water;

a supply line configured for delivery of water liable to be contaminated to said bed; and

a delivery line for receiving treated water flowing from the bed;

wherein the carbon comprises a micro-mesoporous or micro-macro-porous carbon that has been activated in carbon dioxide to minimise the introduction of surface oxygen functionality and that has low heteroatom surface functionality.

In embodiments of the apparatus the carbon is from carbonization and activation of a mesoporous or macroporous phenolic resin which may be in the form of beads. In the latter case the bed and the supply and delivery lines may be configured for operation in expanded bed mode, or they may be configured as an inverse (down-flow) liquid- solid fluidised bed system, fiuidized bed systems having the advantage of low pressure drop.. A steam generator may be configured to supply steam to the bed for regeneration thereof e.g. in a plurality of regeneration cycles. The steam generator may be configured to supply dry superheated steam to the bed or it may be configured to supply saturated steam to the bed. A line from the bed may lead to an oily water separator for separating contaminant organics from condensate.

BRIEF DESCRIPTION OF DRAWINGS

How the invention will be put into effect is described by way of example only with reference to the accompanying drawings, in which:

Fig. 1 is a block diagram showing methods for producing carbons of controlled structure derived from phenolic resins; Fig 2 is a plot of adsorbed volume against pressure for a micro-mesoporous and micro-macroporous carbon;

Fig 3 is a plot of DV/DLogR against mean pore size (A) for various carbons;

Fig 4 is a diagrammatic end perspective view of part of a carbon monolith showing channel structure, with an enlarged detail showing macro-particles forming part of a wall structure and with a further enlarged detail showing micro-domains within a macro-particle;

Fig 5 is a cross-section of a simple extrudate, trilobe extrudate and Rachig ring;

Fig. 6 is a diagram showing breakthrough characteristics of a fixed bed adsorber;

Fig 7 shows adsorption isotherms (mg/g) of metaldehyde on activated carbons;

Fig. 8 is a diagram showing a comparison of the pore structures of conventional activated carbon (A) and pyrolysed phenolic resin (B);

Fig 9 is a graph of concentration of metaldehyde (mg/ml) showing the results of a flow test for metaldehyde using a solution of 0.5 mg/1 of metaldehyde flowing at 0.4 ml/minute and Fig 10 is a graph showing the results of a similar flow test for Atrazine;

Fig 11 is a diagram showing the effects of activated carbon in the removal of metaldehyde for various water samples;

Fig 12 is a graph of TOC (total organic carbon, ppm) against volume of effluent comparing the uptake characteristics of the phenolic resin derived carbon and 4 commercial activated carbons for refinery waste water;

Figs 13 and 14 are cyclic response curves for a conventional activated carbon and for a phenolic resin derived activated carbon after steam regeneration cycles at 200°C;

Fig. 15 is a block diagram of a two bed adsorption system and Fig. 16 is a block diagram of a three bed system for enhanced purification; and

Fig 17 is a block diagram of an inverse (Down Flow) liquid solid Fluidised Bed

System DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Figure 1 shows routes to the sintered monolithic structures (micro-macro porous) and solution dispersed bead (micro-meso porous) materials. In both cases the materials are produced from a nucleophilic component which may be a phenolic compound or a phenol condensation pre-polymer e.g. a novolak cross linked with an electrophilic cross-linking component e.g. formaldehyde, furfural or hexamethylene tetramine (Hex).

For the sintered structures a physical mixture of milled resin and hex is dry cured at 150°C. is then milled/classified and formed into the sintered resin structure by for instance pelleting, pressing, extrusion or injection moulding (US 2005/126395).

For the solution dispersed materials the resin and Hex are dissolved in ethylene glycol which is then dispersed in oil at 150°C to form and cure beads, after which the ethylene glycol is removed by either water washing or vacuum drying, see US-A- 7850942. The pore structure of the solution dispersed beads, measured by nitrogen adsorption and converted to a pore size distribution using the BJH method, is shown in Figures 2 and 3. These show the microstructure (<2nm) of the phenolic resin derived carbon and the meso/macro structure introduced by the solvent pore forming.

Contaminated water can be treated using microporous/mesoporous or microporous/macroporous carbon in the form of beads, granules or monoliths. The carbon is preferably produced by the carbonisation and activated of controlled structure phenolic resin materials. The substances that can be removed include a wide range of chemical, pharmaceutical and pesticide materials. However the carbon material of the invention is surprisingly effective at removing more hydrophilic contaminants that are not effectively removed by the conventional carbons used in final stage GAC treatment. This includes for example the common molluscicide metaldehyde which is particularly difficult to process using conventional carbons and other relatively hydrophilic compounds such as atrazine and estradiol which are also poorly retained.

Preparation of porous carbon from phenolic resin

The applicants have developed a number of processes for the production of activated carbon containing micro-, meso- and macropores from porous phenolic resins, the products commonly taking the form of beads, granules or monoliths, and much of this technology is applicable in the present invention.

As used herein, the term "micropore" refers to pores with diameter <2 nm, as measured by nitrogen adsorption and mercury porosimetry methods and as defined by IUPAC.

As used herein, the term "mesopore" refers to pores with diameter from ca. 2 nm to ca. 50 nm, as measured by nitrogen adsorption and mercury porosimetry methods and as defined by IUPAC.

As used herein, the term "macropore" refers to pores with diameters larger than 50 nm, as measured by nitrogen adsorption and mercury porosimetry methods and as defined by IUPAC. In relation to this invention there are two types of macropores. In macroporous beads they are located within beads and formed by pore-formers. Their size is 50-500nm, typically 50 - 200 nm. In simple monoliths there are macropores present that are formed due to the voids between sintered particles. Their size is typically 700 - 2000 nm. These macropores do not provide any adsorption potential and simply provide transport pores into the structures..

The evidence from the trial work applicants have carried out has demonstrated that some mesopores or macropores are needed to provide for the adsorption of the hydrophilic contaminants present in the water streams in a typical water treatment plant. Typically a precursor resin formulation is used which comprises a significant proportion of pore former, e.g. 250 parts ethylene glycol or other pore former to 100 parts of resin-forming components although this high porosity can preferably be achieved through the use of additives such as urea in combination with the ethylene.

US 2008/025907 (Tennison et al.,) the disclosure of which is incorporated herein by reference) discloses making a mesoporous resin by condensing a nucleophilic component which comprises a phenolic compound or a phenol condensation pre- polymer with at least one electrophilic cross-linking agent selected from formaldehyde, paraformaldehyde, furfural and hexamethylene tetramine in the presence of a pore- former selected from the group consisting of a diol (e.g. ethylene glycol), a diol ether, a cyclic ester, a substituted cyclic ester, a substituted linear amide, a substituted cyclic amide, an amino alcohol and a mixture of any of the above with water to form a resin. The pore-former is present in an amount effective to impart meso- or macro porosity to the resin (e.g. at least 120 parts by weight of the pore former being used to dissolve 100 parts by weight of the total resin forming components, i.e. nucleophilic component plus electrophilic component), and it is removed from the porous resin after condensation by cascade washing with water or by vacuum drying.

The resulting resin may be carbonised by heating in an inert atmosphere to a temperature of at least 600°C to give a material having a bimodal distribution of pores, the pore structure as estimated by nitrogen adsorption comprising micropores and mesopores or macropores. The value for the differential of pore volume with respect to the logarithm of pore radius (dV/dlogR) for the mesopores is greater than 0.2 for at least some values of pore size in the range 20-500A. The mesoporous carbon may have a BET surface area of 250-800m²/g without activation. It may be activated by heating it at high temperature in the presence of carbon dioxide, steam or a mixture thereof, e.g. by heating it in carbon dioxide at above 800°C. It may then have surface areas of up to 2000 m2/g and even higher e.g. 1000-2000m2/g. As used herein the term "BET surface area" is determined by the Brunauer, Emmett, and Teller (BET) method according to ASTM D1993-91, see also ASTM D6556-04. For the purposes of the current invention it is preferred to use carbon dioxide. In contrast to other applications for these porous carbons we have found that it is the essentially hydrophobic character of the carbon that is critical for the adsorption of hydrophilic components from water which can only be achieved through carbon dioxide activation.

Phenolic resins - nucleophilic component

Resins for making carbonaceous material can be prepared from any of the starting materials disclosed in US 2008/025907. Nucleophilic components may comprise phenol, bisphenol A, alkyl phenols e.g. cresol, diphenols e.g. resorcinol and hydroquinone and aminophenols e.g. w-amino-phenol.

It is preferred to use as nucleophilic component a phenolic novolac or other similar oligomeric starting material which because it is already partly polymerized makes polymerization to the desired resin a less exothermic and hence more controllable reaction. The preferred novolacs have average molecular weights (AMW) in the range of from 300 to 3000 prior to cross-linking (corresponding to a DP with respect to phenol of about 3-30). Where novolac resins are used, they may be solids with melting points in the region of 100°C. Novolac resins of MW less than 2000 and preferably less than 1500 form resins which on carbonisation tend to produce carbons with desired pore size distributions using lower amounts of pore former. Novolacs are thermally stable in that they can be heated so that they become molten and cooled so that they solidify repeatedly without structural change. They are cured on addition of cross-linking agents and heating. Fully cured resins are infusible and insoluble.

Whilst commercial novolacs are largely produced using phenol and formaldehyde, a variety of modifying reagents can be used at the pre-polymer formation stage to introduce a range of different oxygen and nitrogen functionalities and cross-linking sites. These include but are not limited to: -

(a) Dihydric phenols e.g. resorcinol and hydroquinone. Both are more reactive than phenol and can lead to some cross-linking at the pre-polymer production stage. It is also possible to introduce these compounds at the cross-linking stage to provide different cross-linking paths. These also increase the oxygen functionality of the resins.

(b) Nitrogen containing compounds that are active in polycondensation reactions, such as urea, aromatic (aniline, m-amino phenol) and hetero aromatic (melamine) amines. These allow the introduction of specific types of nitrogen functionality into the initial polymer and final carbon and influence the development of the mesoporous structure of both the resins and the final carbons. Like hydroquinone and resorcinol, all the nitrogen containing nucleophilic modifying reagents which can be used possess two or more active sites and are more reactive in condensation reactions than phenol or novolacs. It means that they are first to react with primary cross-linking agents forming secondary cross-linking agents in situ.

The nucleophilic component may be provided alone or in association with a polymerization catalyst which may be a weak organic acid miscible with the novolac and/or soluble in the pore former e.g. salicylic acid or oxalic acid. Whilst these can be used in the current invention the use of phenol alone is preferred to minimise the concentration of more hydrophilic sites.

The concentration of novolac in the pore former may be such that when combined with the solution of cross-linking agent in the same pore former the overall weight ratio of pore former to (novolac + cross-linking agent) is at least 125:100 by weight. The actual ratios of novolac:pore former and cross-linking agenfcpore former are set according to convenience in operation e.g. in the case of the process disclosed in WO 2008/043983 (Tennison) by the operational requirements of a bead production plant and are controlled by the viscosity of the novolac:pore former solution such that it remains pumpable and by the ratio of cross-linking agent:pore former such that the cross-linking agent remains in solution throughout the plant. For the purposes of this invention the presence of some mesoporosity is beneficial but should be minimised to maximise the density of the final carbon. Cross-linking agents for phenolic resins

The cross-linking agent is normally used in an amount of from 5 to 40 parts by weight (pbw) per 100 parts by weight of the nucleophilic components e.g. novolac,. It may be, for example, an aldehyde e.g. formaldehyde or furfural, it could be hexamethylenetetramine (hexamine), or hydroxymethylated melamine.

Hexamine is preferably used as cross-linking agent. In embodiments requiring a completely cured resin (beads or granular materials), it is preferably used for cross- linking novolac resin at a proportion of 10 to 25 pbw e.g. about 15 to 20 pbw hexamine per 100 pbw of novolac. This ensures formation of the solid resin with maximal cross- linking degree and ensures the stability of the mesopore structure during subsequent removal of the pore former

Pore-formers

The pore former also acts as solvent. Thus, the pore former is preferably used in sufficient quantities to dissolve the components of the resin system, the weight ratio of pore former to the total components of the resin system resin being preferably at least 1.25: 1. Below this level the resulting resins have essentially no mesoporosity.

Details of suitable pore formers are given in US 2008/025907 (Tennison). The pore former may be, for example, a diol, a diol-ether, a cyclic ester, a substituted cyclic or linear amide or an amino alcohol e.g. ethylene glycol, 1,4-butylene glycol, diethylene glycol, triethylene glycol, γ-butyrolactone, propylene carbonate, dimethylformamide, N-methyl-2-pyrrolidinone and monoethanolamine, ethylene glycol being preferred, and where the selection is also limited by the thermal properties of the solvent as it should not boil or have an excessive vapour pressure at the temperatures used in the curing process.

It is thought that the mechanism of meso- and macropore generation is due to a phase separation process that occurs during the cross-linking reaction. In the absence of a pore former, as the linear chains of pre-polymer undergo cross-linking, their molecular weight initially increases. Residual low molecular weight components become insoluble in the higher molecular weight regions causing a phase separation into cross-linked high molecular weight domains within the lower molecular weight continuous phase. Further condensation of light components to the outside of the growing domains occurs until the cross-linked phase becomes essentially continuous with residual lower molecular weight pre-polymer trapped between the domains. In the presence of a low level of pore former the pore former is compatible with, and remains within, the cross-linked resin domains, (e.g., <120 parts/100 parts Novolac for the Novolac-Hexamine-Ethylene Glycol reaction system), whilst the remainder forms a solution with the partially cross-linked polymer between the domains. In the presence of higher levels of pore former, which exceed the capacity of the cross-linked resin, the pore former adds to the low MW polymer fraction increasing the volume of material in the voids between the domains that gives rise to the mesoporosity and/or macro porosity. In general, the higher the pore former content, the wider the mesopores, up to macropores, and the higher the pore volume.

This phase separation mechanism provides a variety of ways of controlling the pore development in the cross-linked resin structures. These include chemical composition and concentration of the pore former; chemical composition and quantity of the cross-linking electrophilic agents, presence, chemical nature and concentration of modifying nucleophilic agents, chemical composition of phenolic nucleophilic components (phenol, novolac), the presence of water within the solvent and concentration of any curing catalyst if present.

Production of resin precursor and carbon in bead form

In US 2008/025907, production of the resin in both powder and bead form is disclosed. Production of the bead form may be by pouring a solution of a partially cross-linked pre-polymer into a hot liquid such as mineral oil containing a dispersing agent and stirring the mixture. The pre-polymer solution forms into beads which are initially liquid and then, as curing proceeds, become solid. The average bead particle size is controlled by several process parameters including the stirrer type and speed, the oil temperature and viscosity, the pre-polymer solution viscosity and volume ratio of the solution to the oil and the mean size can be adjusted between 5 and 2000μηι, although in practice the larger bead sizes are difficult to achieve owing to problems with the beads settling in the stirred dispersion vessel. The beads can then be filtered off from the oil. In a preparative example, industrial novolac resin is mixed with ethylene glycol at an elevated temperature, mixed with hexamine and heated to give a viscous solution which is poured into mineral oil containing a drying oil, after which the mixture is further heated to effect curing. On completion of curing, the reaction mixture is cooled, after which the resulting porous resin is filtered off, and washed with hot water to remove pore former and a small amount of low molecular weight polymer. The cured beads are carbonized to porous carbon beads which have a pore structure as indicated above, and may be activated as indicated above. The beads can be produced with a narrow particle size distribution e.g. with a D90:D10 of better than 10 and preferably better than 5. However, the bead size distribution that can be achieved in practice in stirred tank reactors is relatively wide, and the more the process is scaled up the worse the homogeneity of the mixing regime and hence the particle size distribution becomes wider.

US2010/0086469 (Tennison; see also US-A-8383703), the disclosure of which is incorporated herein by reference describes and claims a process for producing discrete solid beads of polymeric material e.g. phenolic resin having a porous structure, which process may produce resin beads on an industrial scale without aggregates of resin building up speedily and interrupting production. The process comprises the steps of:

(a) combining a stream of a polymerisable liquid precursor e.g. a novolac and hexamine as cross-linking agent dissolved in a first polar organic liquid e.g. ethylene glycol with a stream of a liquid suspension medium which is a second non-polar organic liquid with which the liquid precursor is substantially or completely immiscible e.g. transformer oil containing a drying oil; (b) mixing the combined stream to disperse the polymerisable liquid precursor as droplets in the suspension medium e.g. using an in-line static mixer;

(c) allowing the droplets to polymerise in a laminar flow of the suspension medium so as to form discrete solid beads that cannot agglomerate; and

(d) recovering the beads from the suspension medium.

Dispersion medium

For bead production, the pore former comprises a polar organic liquid e.g. ethylene glycol chosen in combination with dispersion medium which is a non-polar organic liquid so as to form a mainly or wholly immiscible combination, the greater the incompatibility between the pore former which forms the dispersed phase and the dispersion medium, the less pore former becomes extracted into the dispersion medium. The pore former desirably has a greater density than the dispersion medium with which it is intended to be used so that droplets of the pore former containing dissolved resin- forming components will pass down a column more rapidly than a descending flow of dispersion medium therein. Both protic and aprotic solvents of different classes of organic compounds match these requirements and can be used as pore formers, both individually and in mixtures. In addition to dissolving the reactive components and any catalyst, the pore former should also, in the case of phenolic resins, be compatible with water and/or other minor condensation products (e.g. ammonia) which are formed by elimination as polymerization proceeds, and the pore former is preferably highly miscible with water so that it can be readily removed from the polymerized resin beads by washing.

The dispersion medium is a liquid which can be heated to the temperature at which curing is carried out e.g. to 160°C without boiling at ambient pressure and without decomposition and which is immiscible with ethylene glycol and with the dissolved components therein. It may be hydrocarbon-based transformer oil which is a refined mineral oil and is a by-product of the distillation of petroleum. It may be composed principally of C15-C40 alkanes and cycloalkanes, have a density of 0.8-0.9 depending upon grade and have a boiling point at ambient pressure of 260-330°C, also depending upon grade. Transformer oil has a viscosity of about 0.5 poise at 150°C which is a typical cure temperature. Transformer oil or other dispersion medium may be used in volumes 3-10 times the volume of the combined streams of nucleophilic precursor and crosslinking agent e.g. about 5 times.

Dispersing agents

Preferred dispersing agents which are dissolved in the dispersion medium before that medium is contacted with the reaction mixture to be dispersed therein to retard droplet coalescence are either sold as drying oils e.g. Danish oil or are produced by partially oxidizing naturally occurring precursors such as tung oil, linseed oil etc. The dispersing agents are consumed as the process proceeds, so that if the dispersion medium is recycled, dispersing agent in the recycled oil stream should be replenished. The dispersing agent is conveniently supplied as a stream in solution in the dispersion medium e.g. transformer oil and e.g. in an amount of 5-10%v/v where Danish oil is used which contains a low concentration of the active component to give final concentration of the dispersant in the dispersion medium 0.2 - 1% v/v. Higher dispersant concentrations would be used in the case of oxidised vegetable oils.

Production of the Resin Precursor in Granular Form

The micro-mesoporous resin can also be produced in granular form. In this instance the novolak resin-hexamine-ethylene glycol mixture is poured into trays which are then cured by placing the trays in an oven, raising the temperature to 140-150C and holding this temperature for at least 1 hour. For this process the curing agent concentration must be at least 10% and preferably 20%. At this stage the solid block of resin is crushed to around 1 -5mm particle size prior to post treatment. The higher levels of curing agent are preferred, at least 15%, to prevent loss of the meso-macro pore structure in the resin during milling, washing/vacuum drying and carbonisation. The higher level of curing agent also prevents the resin granules from sintering during subsequent processing. This curing process could also be carried out in a continuous manner using for instance a mesh belt tunnel kiln or a rotary kiln. These processes are well known to those skilled in the art and are used for instance in the production of cured resin used in automotive brake components. Solvent Removal from resin beads and granular materials

The resin beads or granules formed as described above must first be treated to remove the pore former after which they may additionally be formed into monoliths and subsequently carbonised and optionally activated. The pore former can be removed either by water washing or vacuum drying. The beads can be treated directly but In the case of the granular material the washing or vac drying is carried out after crushing to around l-2mm size granules to facilitate extraction. After washing and drying this can then be milled to the required particle size although it is preferred that the material is carbonised and activated prior to fine milling If water washing is used this preferably uses at least a two stage process using hot water at ~80C. This is preferably carried out using a cascade washing process where the water from the second stage, which contains a relatively low level of the pore former, is recycled to the first washing stage. The waste water from the first stage, which contains a high level of the pore former can either be disposed of or the pore former can be recovered by distillation. Vacuum drying can be carried out using any commercially available vacuum dryers although it is preferred that this should use a stirred or moving bed rather than a static tray system.

Production of Sintered Resin Forms

By "monolithic" is meant that the porous carbon is in a single piece i.e. not granular or not composed of granular carbons bound together by a binder etc. The monolithic carbon contains large transport channels. For a symmetrical monolith (4) a continuous channel structure is defined by a channel dimension, W, and a wall thickness, t, or for an asymmetric monolith by channel length and width or other relevant dimensions as well as wall thickness t. These fix the ratio of open to closed area and therefore the flow velocity along the channels of the monolith. The walls of the monolithic carbon have a macroporous structure providing continuous voids or pores whilst the micro structure is contained within the primary particles.

Known methods for the production of complex shaped controlled porosity adsorbent material are discussed in US 2005/126395 (Blackburn and Tennison, see also US-A-7160366 the disclosure of which is incorporated herein by reference). The inventors explain that there are very few viable routes for the production of complex shaped controlled porosity adsorbent materials with good mechanical properties. For instance, they explain that activated carbon is traditionally produced by taking a char, made by pyrolysing an organic precursor or coal, grinding the char to a fine powder, mixing this with a binder, typically pitch, and extruding or pressing to give a "green" body. The green body is then further fired to pyrolyse the binder and this is then typically further activated in steam, carbon dioxide or mixtures of these gases to give the high surface activated carbon product. The drawback to this route is that the binder, which is usually a thermoplastic material, goes through a melting transition prior to pyrolytic decomposition. At this point the material is weak and unable to support a complex form. This, combined with the problems of activating the fired body, limits the size and shape of the products to typically simple extrudates.

An alternative route is to take an activated carbon powder and form this directly into the final shape. In this instance a range of polymeric binders have been used that remain in the final product. The main drawback to this route is that high levels of binders are required and these then tend to both fill the pores of the activated carbon powder and encapsulate the powder leading to a marked reduction in adsorption capacity and deterioration in the adsorption kinetics. The presence of the polymeric phase also degrades the physical and chemical stability of the formed material, severely limiting the range of applicability. A further alternative is to take a formed ceramic material, such as a multichannel monolith, and to coat this with a carbon forming precursor such as a phenolic resin; this can then be fired and activated to produce a ceramic-carbon composite. The main limitations of this route are the cost associated with the ceramic substrate and the relatively low volume loading of carbon.

In the current embodiments carbonised and activated sintered carbons are now formed from phenolic resin precursors. Sintered porous carbon can be made by partially curing a phenolic resin to a solid, comminuting the partially cured resin, extruding the comminuted resin, sintering the extruded resin so as to produce a form-stable sintered product and carbonising and activating the form-stable sintered product. EP 0 254 551 Satchell et al. gives details of methods of production the porous resins suitable for forming the porous carbon used in the present invention and its contents are included herein by reference. US 2004/045438A1 (Place et al, se also US-A-6964695, the disclosure of which is incorporated herein by reference) gives details of producing monolithic structures using the sintered resin structures. In the standard process, the resin cure is controlled so that it is sufficient to prevent the resin melting during subsequent carbonisation but low enough that the particles produced during the milling step can sinter during subsequent processing. The temperature and duration of the partial curing step are selected as to give a degree of cure sufficient to give a sinterable product, and such that a sample of the partially cured solid when ground to produce particles in the size range 106-250μηι and tableted in a tableting machine gives a pellet with a crush strength which is not less than 1 N/mm. Preferably the pellet after carbonisation has a crush strength of not less than 8 N/mm.

By "sintering" we mean a step which causes the individual particles of phenolic resin to adhere together without the need for a separately introduced binder, while retaining their individual identity to a substantial extent on heating to carbonisation temperatures. Thus the particles must not melt after forming so as to produce a molten mass of resin, as this would eliminate the internal open porosity of the article. The open porosity (as opposed to the closed cells found in certain types of polymer foams) is believed to be important in enabling formed particles to retain their shape on carbonisation.

In one embodiment the comminuted resin particles have a particle size of 1- 250μιη, more preferably 10-70μηι. Preferably the resin powder size is 20-50μιη which provides for a macropore size of 4-10μηι with a macropore volume of around 40%. The size of the particles is selected to provide a balance between diffusivity through the interparticle voids and within the particles.

As disclosed in US 2004/045438 the milled powder can then be extruded to produce polymeric structures with a wide range of physical forms and cell structures, limited only by the ability to produce the required extrusion die. These can range from relatively simple "spaghetti" forms up to and including trilobe and quadralobe structures along with for instance Rachig rings. In a further level of complexity the resin can be extruded to form square channel monoliths. At this stage the monolith has a bimodal structure - the visible channel structure with either the central channel in a simple tube or the open cells in a square channel monolith of 100-1000μη cell dimension and cell walls with thickness 100- ΙΟΟΟμηι and the macropore structure within the walls generated by the sintered resin particles. The carbonisation steps take place preferably by heating above 600°C for the requisite time e.g. 1 to 48 hours and takes place under an inert atmosphere to prevent oxidation of the carbon. On carbonisation the material loses about 50% by weight and shrinks by about 65% by volume but, provided the resin cure stage was correctly carried out, this shrinkage is accommodated with no distortion of the monolith leading to a physical structure identical to that of the resin precursor but with dimensions reduced by -30%. The macropore size is also reduced by ~30% although the macropore volume (ml/ml) remains unaltered. During carbonisation the microporosity of the porous carbon develops. After carbonisation there may be partial blocking of the micropores by the decomposition products from the carbonisation process. These blockages may be removed by activation to provide rapid access to the internal structure of the carbon that is essential for the operation of the monoliths as adsorption devices.

Sintered carbon produced from phenolic resins by existing processes have a microporous/macroporous structure and introduction of mesoporosity is not intended. Forming monoliths having mesoporosity intentionally introduced into their structure gives rise to a number of difficulties. As previously described, an embodiment of a standard monolith production process comprises the steps of (i) pouring a mixture of novolak, cross-linking agent (hexamine) and pore former (ethylene glycol) into a tray, (ii) partially curing e.g. at 150°C in an oven, (iii) crushing or hammer milling the cured resin to reduce its particle size, (iv) removing residual pore former by water washing or by vacuum drying, (v) jet milling the washed and dried material, (vi) extruding the material as a dough to form a resin monolith which is stabilised by sintering, and (vii) subjecting the sintered monolith to carbonization and activation.

It is necessary for the above process that the partially cured resin should be in a sinterable state, and that requirement limits the amount of cross-linking agent that can be used. The standard process used by the applicants for making monolithic carbon from phenolic resins uses 5 parts by weight of hexamine as cross-linking agent, but if the same amount is used in the production sequence indicated above the induced mesoporosity collapses during pore former removal. It is therefore desirable to increase the proportion of cross-linking agent to an amount sufficient to stabilise the mesoporous structure but less than an amount that prevents the partially cured resin from sintering. Increase of the amount of cross-linking agent to e.g. 1 1 pbw per 100 pbw of novolac has been found effective, although the resulting monoliths are of reduced mechanical strength.

The walls of the sintered carbon have a macroporous structure. By "macroporous" is meant that the carbon has continuous voids or pores. The macropore structure in the walls is controlled by the particles used to form the structure. When the structure is made from macro-particles with a mean particle size of DP the macro pore size is typically 20% of the size of the precursor resin particles. In the square channel monoliths the particle size can be varied over a wide range from a maximum particle size of approximately 10% of the wall thickness, t, to a minimum practical particle size of about ΙΟμηι. This gives a macropore size of 2-20μηι within the wall for a 1mm wall thickness. For the simpler "spaghetti" structures a wider range of particle sizes is possible. The pore size fixes the diffusivity of the adsorbate molecules within the matrix. In the current embodiments the monoliths are square channel monoliths with a cell structure (cells per square cm) where the channel size is between 100 and 2000μηι and the wall thickness is between 100 and 2000μηα and with an open area of between 30 and 60% to give a good carbon packing density per unit volume and acceptable mass transfer characteristics. Carbonisation and Activation of Resin Structures

In US 2010/098615 (Tennison, see also US-A-8501142, the disclosure of which is incorporated herein by reference) there is provided a process for carbonizing and activating bead or granular polymeric material and especially the solid beads of polymeric material resulting from the process of US 2010/086469, which comprises supplying the material to an externally fired rotary kiln maintained at carbonizing and activating temperatures, the kiln having a downward slope to progress the material as it rotates, the kiln having an atmosphere free of oxygen provided by a counter-current of carbon dioxide or steam, and annular weirs being provided at intervals along the kiln to control progress of the material.

Alternatively the resin beads can be carbonised and activated on a smaller scale using a batch furnace. Here the carbonisation and activation may be carried out as separate steps where the carbonisation takes place in carbon dioxide at ~800°C and the activation in carbon dioxide at between 850 and 950°C or in steam at between 700 and 850°C.

For the purposes of this invention it is preferred to use carbon dioxide as the activating medium.

Post Treatment of Carbon Beads or Granules

The activated carbon beads as produced by this method have a very low concentration of hetero atoms at the surfaces after the activation in high temperature carbon dioxide and are essentially pH neutral. This is in marked contrast to commercial activated carbons, where the production process generally involves activation in a steam/carbon dioxide/air mixture, and which leads to a significant concentration of oxygen groups at the surface. These carbons are significantly more hydrophilic and tend to exhibit an acidic character. The surface chemistry of the phenolic resin can however be extensively modified to produce either acidic or basic surface groups and to significantly change the hydrophobic hydrophilic character.

Activated carbon materials for water treatment in the present patent application have been prepared by the generic methods described below though they may be prepared also by numerous variations of this method. Alternatively mesoporous or macroporous resin-precursors for carbons may be prepared in blocks, then crushed, washed with water or vacuum-dried from ethylene glycol and further processed into monoliths. For the purposes of this invention the introduction of hetero atoms which can increase the hydrophilic character is to be avoided.

Adsorption experiments

Batch adsorption studies were carried out incubating aqueous solution of different challenge molecules such as metaldehyde (40ml, 5% ethanol) with the required mass (mg) of carbon beads or GAC (granular activated carbon) for 48h at 25°C in an orbital shaker model SI500 (Stuart Scientific, UK ) at 90 rpm. The kinetics of the process could be followed by sampling the supernatant liquid at time intervals throughout the 48hour period. Flow studies were carried out at equivalent conditions although in this case the carbon was loaded into a column and the contaminated water stream was pumped through the bed. . In the batch experiments, aliquots (0.5ml) were taken for analysis at definite periods of time. A maxirnmn of 6 aliquot s were taken per sample. The adsorptive capacity has been obtained by the mass of the challenge molecules removed from solution, which was difference between the amount of the challenge at the initial conditions and amount after the incubation with the sorbent, per mass of sorbent. The batch adsorption studies were performed in triplicate.

Process Applications

In conventional water treatment processes final stage GAC Filtration AdsorptiDn only tends to be used where the water will be used for human consumption. As it follows the conventional process the level of contaminants in the feed is very low, typically less than 200ppb, and_^J jg, permits the dsoi ftrs_, to be used for long periods without regeneration, usually from 2-6 years. However this still requires very large beds. A typical smaller plant, processing around 10-20miILion litre/day requires 5 §¾8.ϊ¾4ϊ§. each containing approximately 50m⁵ of activated carbon. In operation there would typically be one djgQjfogr on back flush and one being regenerated. During regeneration approximately 10-12 "A of the carbon is destroyed and is replaced with virgin carbon whilst the performance of the remaining carbon will also deteriorate over time. This entire process system is fixed by the requirement for external regeneration and the need to minimise the frequency of regeneration.

If the carbon can be efficiently regenerated in-situ it becomes possible to consider much shorter operating periods. In principle then the bed volume reduces by the ratio of the time on stream. If for instance each bed is regenerated after 1 week compared to 2 years the operating bed VOlume reduces by approximately xlOO so that for the plant above the total carbon requirement could be reduced to 2.5m^J. The possible reduction would however then be limited by adsorption kinetics. In the process described above, if only 3 beds were operational this corresponds to a liquid hourly space vebcity (LHSV)

The results shown in Figure 4 and Figure 5 were obtained at a space vebcity of 35v/v h but using sigriificantly smaller particle sizes than would be used in a full size reactor. To reduce the bed v lumes to the minimum possible would Κ>£ϋ£3§& the space vebcity to -300 and reduce the EBCT to approx. 0.15 minutes. However the flow test results achieved with atrazine (Table 1) using the n^gsjjpg j^ beads were achieved at a space velocity of 465vfvfh showing that operation at such high velocities is possible. This will however also be a function of the pressure drop through the reactor. With the phenolic derived materials this can be reduced through the use of a tritebfi. or other structured

where the bed yjojdagg, can be adjusted through the physical form of the material (5). This then has the further benefit of erLhancing the adsorption kinetics as the diffusrvity is related to the effective radius. As this is reduced by ~3 the diffusivity should be enhanced by ~9. As in the case of the existing coniLmercial plants this fixed bed plant would operate with multiple beds with one regenerating, one on back flush and 1 or 2 adsorbing.

The cost and viability of the process will then be dominated by the regeneration cost. For a 2.5m³ bed the results suggest that approximately 15m³ of water as steam at 200°C would be required if a fixed bed was used. The frequency would be fixed by the nature of the stream to be treated but a 6 hour regeneration cycle at 1-2 week intervals is likely. The high linear velocities in these systems require careful design of the adsorbent to mir-imise pressure drop and maximise the adsorption kinetics. The tfllobji and j &hjg ring structures are preferred.

This is not however the most efficient process option, TJje.. breakthrough curve in a fixed bed is shown in 6. At any point in time the bed is characterised by three zones - the filled zone, the mass transfer zone (MTZ) and the length of the unused bed (LUB). Only the MTZ is actually in use and this zone moves along the bed until the breakthrough level is exceeded at the bed outlet. The filled zone and LUB can be considered as completely inactive for the entire process cycle.

The more efficient process uses a system where only the saturated part of the bed is continuously withdrawn, regenerated and returned to the bed. Possibilities include moving ¾£α½_Λ liquid phase fluidised beds and expanded beds. In the latter two cases the back flushing step required in fixed beds is no longer an issue and all of the continuous processes could operate with a single bed. In both the tluidized and expanded bed processes control of the adsorbent particle size and density along with good kinetics will make the use of the bead materials essential. For the moving bed granular or bead materials could be used.

Fluidized beds have been proposed for use in waste water treatment but until now only in the earlier stages of the process stream. US 200SJD05897

use of a biological fluidized bed system for biological nutrient removal using immobilised bacteria. US 6716344 proposed the use of a circulating fluid bed system using ion exchange beads for metal ion recovery or protein recovery and the possibility of using this in waste water treatment is mentioned but not exemplified. There is no reference to 5 them being used for the final stage GAC adsorption step. US 2012/0094364 refers to the use of a composite fluidised expanded liquid -solid system for the recovery, purification or reaction of single or multiple components but does not show the use of this in final stage waste water processing. One of the drawbacks of the fluid bed system is the extent of ¾a£kn¾ j g in the bed. This can be overcome by the use of an expanded I D as against fluidised bed system. There are numerous patents relating to the use of this process in biochemical separations (e.g. CN101972558) and the biological stages in waste water treatment although there are no known commercial applications, due in part to the cost of the special adsorbents.

The invention will now be illustrated in the following examples.

15

Example 1

Preparation of heads of porous phenolic resins and corresponding carbons

A solution of 100 parts by weight of industrial M.(3Kftk¾. resin with an average molecular weight 700-800 D (Hexjo.n Specialty Chemicals) in ethylene glycol was

20 heated to 90-95*C and thoroughly mixed for 2-5 minutes with a solution of 15 - 20 parts by weight of hexamethyknetetra dne (hexamine) in ethylene glycol heated to the same temperature. The resulting clear solution was poured in stream into 2.5-6 fold volume of stirred hot (1 0-155°C) bw viscosity mineral oil (insulating oil or transformer oil) containing 0.2 - 1% (yfv) of a dispersing agent which was an industrial

25 drying oil (Danish oil), a major component being polyunsaturated (oxidised) vegetable oils. The temperature of the mixture fell to 135-140°C, and the mixture was reheated to 150-155°C in 15-20 minutes. Typically curing occurred within 1-2 minutes at around 140°C followed by substantial evolution of gas, predominantly ammonia. Further heating to 150-155°C for 15-20 minutes ensures the completion of curing. The mixture

3 D was cooled and the resulting beads were separated from the oil by filtration or centrifugation Ethylene glycol was removed from the resin either by multiple hot water extraction or by drying in vacuum (120°C at 50 mm Hg . The above procedure shows the difference between "flash" cure with higher hexamine content (15-20 j¾g) as compared to "slow" cure described in previously published patents (bwer hexamine content - 11 p¾¾; per 100 pbjj; of

Water-washed wet, dried or vacuum-dried resin beads were heat treated to produce carbon materials. A typical procedure comprised but is not restricted to carbonisation in a flow of carbon dioxide with temperature ramping from ambient to 800°C at 3°C min, classification by particle size and further "physical" activation of selected fraction in carbon dioxide flow at 900°C. Many variations of this routine known in the art may also be applied.

Pore size distribution in the resulting carbons is pre-determinedby the porosity of the resin-precursor, which is controlled by the content of the solvent/pore former (typically but not restricted to ethylene glycol) in the resin composition. Table 1 bebw gives details of three resins compositions that are precursors to micro-, mgsj^- and macro-porous carbons, as illustrated by nitrogen adsorption tests of activated materials used in adsorption studies (up to -40% of activation burn-off in carbon dioxide) (Figures 2 and 3).

The particle size distribution of resulting resin beads depends on various parameters including but not restricted to the type of stirring tool, stirring rate, viscosity of the resin solution, concentration of the dispersing agent, resin solution to oil ratio and temperature of the dispersion. Though the distribution is typically broad the size of the predommant fraction could effectively be shifted between -10 micron and ~1 mm.

Table 1 Examples of Porous Resin Produce Using Ethylene Glycol as the Solvent

Figure 2 shows nitrogen adsorption isotherms (a) and calculated pore size distributions (BJH model) (b) of activated carbons derived from the resins of examples 1,2 and 3 respectively (compositions from Table 1):

KHS|0JPJ9¾ (black squares);

me^ojjoj3^(blank squares) and

mSJ$ ^i&*y squares).

Example 2

Control of Porosity Through Pore Bw^X jQfl^

Additional materials with different pore structures can be produced by further adjusting the ratio of total solids (rj£¾¾l&k + hexamine) to ethylene glycol (Table 2). The table also shows the resulting structure of the carbons :-

Table 2 Porous resinfomiulations

Example 3

Preparation of carbon monoliths

A hot solution of 100 parts by weight of M93?pj££ resin in 100 _^p., of ethylene glycol was thoroughly mixed with a hot solution of 16 parts by weight of hexa tiine in 190 parts by weight of ethylene glycol. The resulting solution was transferred into a stainless steel tray, covered with a lid and placed into flameproof oven. Raising the temperature to 150°C and mamtaining it for 1-4 ¾u¾ ensured formation of a solid cross- linked resin cake from a resin solution. After crushing the cake into ~ <lcm pieces of resin it was either dried in vacuum at 110-130°C or washed repeatedly with hot (90- 95°C) water to remove ethylene glycol and then dried until water-free, milled and used for the preparation of a monolith. The XM$Q.$gXQM. resin was milled to <100microns. The powder was then formed into ajjpjigh._. y mixing with water and standard extrusion additives including n¾thoj¾¾ and polyethylene oxide according to the previously defined methods. This was extruded to provide a 10mm resin monolith with a 600 cells per square inch geometry. The pore size distribution analysis of the sample was performed by mercury pojpjgjrjagiiy. The pore size distribution plot indicates that the carbon monolith had pores in the mesjjpQre. range of 200-500nm in size and also a larger population of

in the 10000-20000nm range. Pores above 100,000 nm (100 microns) in size most likely represent channels within the monolith. The importance of the irjggQpoj¾_,s.. and macmpareg within the monolith for removal of representative contanunants is considered in Example 5.

Example 4

Adsorption of Met^dehyde on Various Carbons

The adsorption of je jd jysk according to the test method described above is shown in Table 3 for a range of carbons. Within this data the commercial carbons to N were all high surface area powder grade commercial activated carbons. In general these tend to be chemically activated (phosphoric acid) and although their performance is better than the granular activated carbon they could not be used in a fixed, or probably even in a fluid bed due to their particle size, and are included simply for comparison.

Table 3 Metaldehyde adsor tion on various carbons

In this table the data must be compared at the same solution concentration. It can be seen from the GAC results in (F) and (I) and the results for the TE3/34C material in (B) and (H) that the uptake drops with the solution concentration. Taking account of this the data demonstrates the good adsorption performance of the activated porous beads TE7/52C (A) and TE3/34C (B) where the degree of activation in carbon dioxide is given by the 52C (52% bum off) and 34C(34% burn off) as compared to the conventional activated carbon (GAC) (F). It also shows that there is little difference between the TE3 (slightly ffigsgpg QusJ and TE7 (highly rnacn¾.g.rfiusj. The non beads (E1)(C) appear to have a bwer performance but the solution concentration is also bwer.

It also demonstrates that any further treatment of the bead carbons to change the surface chemistry resulted in a decrease in performance. The samples identified by "Ox" were post treated in air at 300C to increase the concentration of surface acidic functions. It can be seen that in both cases the rrje ajdaj¾yd& capacity has been significantly bwered. The sample labelled ΒΞΑ was treated with djass ju j benzene S!ii -hjsnik acid to further increase the surface acidity and it can be seen that this has also reduced the n tajdj^fdg adsorption although not by as much as the oxidation treatment.

Example 5

S Effect of Activation Extent on Micro /Macro porous Beads

Table 4 shows the adsorption of r JdeJjysk™ activated

beads (TES) as a function of the degree of activation as shown by the codes from 00C to 65C representing the level of activation in carbon dioxide. The carbon beads were activated to the %bum off shown in the table in carbon dioxide at 900°C. It can be 10 seen that on the

the adsorption is very poor but is largely independent of the degree of activation between a bum off of 29% and 65%. This is very surprising as adsorption normally correlates with available surface area.

Example 6

Effect of Surface Chemistry on Point of Zero Charge PZC) and Metaldehvdi: Adsorption

20 In this example the micro/macro porous carbon was subjected to a sequence of treatments as shown in Table 5. The initial carbonised material was first activated in carbon dioxide according to the normal preparation to a bum off of 52% weight (A), a sample was also activated to a higher degree (70% bum off (B) which had no effect on the rigjajdejjyde. adsorption. Sample A was then treated with urea at 300C to increase

25 the surface nitrogen content (D). This reduced the PZC marginally and reduced the BOSlftysfcsfe uptake slightly. Sample A was also oxidised in air at 30G°C to increase the surface acidity. This significantly lowered the PZC and nj jajd^j^^

Table 5 ¾fctaXdeh de Adsorption on TET Bead Derived Carbons Post treated to increase the h tero atom content of the surface

Gte¾8dS33SjKS, of f¼ ssdKH&based Q.n TE7 precursor Ad^ar ii5¾. sapacjj¾

mg mg

ΞΒΕΤ ossi metaldehyde/ metaldehyde/

PZC (m²# (cm¾ ΞΒΕΤ m² ± s g carbon ± s

0.032 ±

2.297 C (Gafeed) 1443 1.4 0.22 0.008 38.8 ± 2.7

D Im fed H3t¾ 0.036 ±

8.743 urea 1651 1.85 0.2 0.002 60.3 ± 3.9

B (activated to higher 0.034 ±

9.615 degree) 2045 1.52 0.15 0.0007 69.0 ± 1.4

0.040 ±

10.058 A gajjjjpjj 1649 1.25 0.19 0.001 66.5 ± 2.2

Example 7

EHect of ¾I¾ri¾orosit and Particle Size

In this example the purely mjgjojgpjojjg. beads (El) were evaluated as a function of particle size. It was expected that the absence of the nj¾s.o./macro pores might have a sigrrificant impact on access to the carbon and therefore an impact on the adsorption properties. These are compared with the

(TE7) beads at the largest particle size (250-500microns) at the same solution concentration and carbon loading in Table 6. It can be seen that the for the El beads there is some reduction in performance with particle sise at the largest sise. Comparison of this with the TE7 beads shows that the nucroporyijf;._, eads do have a bwer performance, even for the smallest El ¾gads£0- 45microns). The reduced performance might be expected to be greater in a dynamic experiment as against this static adsorption test. Table 6 Effect of Particle Size on ftU rop ro w Bead Performance

It can be seen that the performance is highest for the smallest beads El beads showing a dependence on the diffusion properties of the material. This is also apparent from the comparison with, the roacj¾.gr<¾j£. bead (TE7) where the performance of the 250- SOOmicron beads exceeds the performance of the smallest mjcj ^pjojjg beds

10

Examples 8

Adsorption of Atrazine and EstadiQl

The adsorption of the pharmaceuticals, atrazine and EsladjoJ, are shown in Table 7. It can be seen that the adsorption of Ej?.tradjpi,and Atrazine on the bead carbons 15 are both dramatically higher than on the conventional G AC .

Table 7 Pharmaceutical Compound Adsorption

Example 9

on Activated TE7 beads and GAC.

In this instance, rather than carrying out the batch adsorption studies at a single solution concentration, as was the case in the preceding examples, a wide range of solution concentrations and carbon weights were used to provide a range of final solution concentrations after the adsorption had taken place. The weight uptake is then determined from the initial rj¾ d£ jjfe concentration and the

concentration after equilibration with the carbon was complete. The weight uptake is pbtted against the final solution concentration in Figure 7 Ij ca be seen that the GAC exhibits a very shallow isotherm approaching a saturation uptake of lOmg mejajd^h^dg. /gm carbon as a solution concentration in excess of 00mg/L

IT can be seen that at the expected feed concentration in contaminated water of around 1 micro .gflitre the uptake would be negligible

In marked concentration the adsorption on the TE7-20/52C carbon shows a very sharp isotherm, consistent with a much higher energy of adsorption, reaching saturation uptake of 85mg/L carbon at a feed concentration of 15mg L mejajdejgcdg.. This higher energy adsorption would normally be associated with high energy sites on the carbon but the earlier examples have shown that the introduction of such sites on the carbon tends to reduce rather than enhance the adsorption capacity. Based on the cross sectional area of the mejtajdejjydg. molecule it can be shown that the saturation uptake equates to 5-10% of the available carbon surface as determined by nitrogen adsorption.

Without wishing to be bound by the following we believe that this enhanced adsorption performance is the result of the unique pore structure of the phenolic resin derived carbons. In essentially all conventional activated carbons the pore structure can be represented by approximately slit shaped pores bounded by graphitic layer planes (Figure 8A). The adsorption in these pores is predominantly controlled by Van der Waals forces which can be correlated with the relative size of the adsorbing molecule and the slit width between the graphitic layer planes £ the layer spacing is large relative to the adsj¾fca|e. size the energy of adsorption will be reduced In marked contrast, the phenolic resin derived caibon pore structure is primarily formed by the voids between nanoparticles (approx. lOnm diameter) of glassy carbon (Figure SB). These are formed by pyrolysis of the nsn&-domains of highly cured phenolic resin formed during the resin curing process. The pore structure then comprises cusp shaped voids between these lOnm glassy carbon nanoparticles. It can then be seen that there is a continuous change in the effective pore size from effectively zero to the void space between the cbse packed particles. Under these circumstances an ¾3gpjj¾gjg, molecule can then find a position in these pores that maximises the Van der Waals adsorption energy. This would then account for the observed adsorption isotherm with aju h. adsorption energy but for only a very limited part of the total available surface. This would also the account for the enhanced adsorption for other molecules as shown in example 8 and 10 for atrazine and esj.ra^ioj.

On this basis it is reasonable to assume that enhanced adsorption will be seen for a wider range of pharmaceuticals and pesticides where bw retention is observed on conventional G AC. This includes for instance

Makthion

Carbamazepine

Steroidal Oestrogens (Estag& El. ¾&¾¾al E2, 17-ethinylestradiol)

PPCP 's (Personal care products)

Qxesm Example 10

Flow Adsoiption Studies

For these tests ultrapure water was spiked with 0.5mg L of mej£]dej¾£d& or 32mg/L of Atrazine. The flow result for £ 9ldei¾£5te, through a 5mm column. The adsorption at saturation was 11.4 mg

carbon and at breakthrough 3600 bed volumes of solution had been pumped through the column. OD x 70mm column packed with 322mg of TE7/30 carbon beads is shown in Figure 9. The results for the Atrazine solution are shown in Figure 10. The volume flowed through the bed was smaller but the atrazine concentration was much higher (32mgjL) and the resulting uptake on the carbon was much higher at 410mg/g at saturation.

All of the tests reported above were carried out with solutions of the challenge molecules in distilled water. This raised the question as to whether the unique performance of the phenolic resin carbons was due to the presence of specific surface sites that could be bbcked by contaminants in normal water reducing their efficiency for the critical molecules. Batch adsorption tests were therefore carried out using water collected from a variety of natural sources spiked with the n¾t¾j¾Jjyjfe, The waters used were:-

Arjngiy. reservoir (pH 8.1)

B JGGBtfe k&£, (pH 7.8)

Ouse rj¾¾j; (New Haven, pH 8.1 )

SpaiJm lane (OiteMag pH 8.3)

Tap water (Brighton, 22rt)8/12, pH 7.9)

Ultrapure water (pH 6.2)

The river waters were first centrifuged to remove any particulate contamination. The carbon samples were then incubated with the spiked water on an orbital shaker for 2 days to reach equilibrium. It can be seen from Figure 11 that that virtually gompgfe removal of rnejajdjej¾yd& was achieved in all of the natural waters and the residual levels were only very slightly higher than observed with ultrapure water. This is in contrast to the poor adsorption seen for the GAC. Example 11

Steam Regeneration

Steam regeneration trials were carried out using a simple 2mm diameter "spaghetti" micro-macro porous adsorbent produced by particle sintering rather than by the bead dispersion route. In this instance the feed solution for the adsorption trials comprised a simulated refinery effluent stream containing ethanol (15ppm), benzene (17ppm), toluene (14ppm), styrene (3ppm), phenol (Sppm), indene (6^"ppm) and 8¾ltMe_¾e. (7ppm). These concentrations are considerably higher than would normally be found as feed to a final stage water treatment process. The test was carried out at room temperature in flow mode at 36LHSV using a carbon bed comprising 3cm3 of 250-500 μιη carbon prepared by crushing the 2mm § jrjjdaj&. The phenolic resin derived carbon was compared with 4 commercial activated carbons. The regeneration 5 was carried out with the bed held in a fluid sand bath at 200C with water injected into a vaporisation coil at the bed inlet. 5 bed volume of water were pumped into the evaporator to complete the regeneration. The hydrocarbon content of the effluent was determined by TOC (total organic carbon).

The initial adsorption cycles for the carbons are shown in Figure 12. The

I D performance itnprovement with the phenolic resin derived carbon and the coconut shell derived cabon (SS208C) correlates with a high concentration of smaller pores where the adsorption strength is maximised. The converse is also apparent with the highly KKSapiwW A110 showing significantly worse performance. The two best carbons, the phenolic resin derived material and the S320SC were then subjected to multiple

15 adsorption-regeneration cycles. The results for the 208C are shown in Figure 13. It can be seen that the performance deteriorated dramatically after the 1st cycle and after 4 cycles there was immediate breakthrough demonstrating that the regeneration of the carbon had not been completed. Figure 14 shows the result for the phenolic resin derived caibon. It can be seen that whilst there was some loss of capacity after the first

2 D cycle the third and fourth cycles were essentially identical with an initial breakthrough after -2.7 litres of solution, equivalent to 900 bed volumes or approximately 15% weight uptake on the carbon.

Example 12

25 Multiple Fixed -Expanded Bed System with Steam Regeneration

A simplified view of the two bed steam regeneration process utilising an expanded bed of carbon derived from phenolic resin and e.g. in the form of beads as described above is shown in Fig. 1 . The operation is by way of example and is not meant to be an exhaustive description.

3 D The system comprises two adss.rbj¾.beds (1 and 2), a steam generator (3), steam condenser (4) and oil/water separator (5). The general principle of the process is that contaminated water is flowed upwards through one bed with clean water exiting through the top of the bed. This can be operated in an expanded bed mode to prevent blockage although this is not essential. At the same time steam is produced in the generator (3) and passes down flow through the second bed, displacing the adsorbed organics. The steam regeneration can be carried out using either superheated steam if the contaminants are purely organic or if the contaminarLts contain significant inorganic contaminants may be carried out using saturated steam at higher pressures. It has been shown that steam at 150°C (approximately 4 bar) is sufficient. The use of saturated steam is preferred as it is possible to heat the bed to the regeneration temperature more efficiently than with dry, superheated steam. The steam passes from the second bed and through the condenser (4) to the oil/water separator (5). As the components all have a low solubility in water and this will be exceeded in condensate, the organics can be separated from the upper layer of the separator for disposal. The dirty water from the bottom of the separator which will be saturated with the contaminants may then be returned to the feed stream of the 1st adjgpjfee However in the case of mej jde.¾yd£ the majority of the adsorbed njejajdejjydg, is thermally decomposed during the regeneration significantly reducing

of the contamination in the condensate. Provided that the ratio of the dirty water inlet flow to the recycled dirty water is large the addition of the recycled contaminants will not seriously impact on the bed performance. As the steam requirement is typically 6-10 bed volumes, which could be of the order of 100m⁵ for a 10m⁵ bed, and the total water flow in a 2 week period could exceed 200,000m3 this is easily met.

This system can also be extended if additional capacity is required by using additional beds where for a three bed system 2 are adsorbing whilst 1 is regenerating (Fig. 16).

The bead carbons are also ideally suited to use in a fluidised bed system. A possible configuration is shown in Figure 17 where the beads flow down through the column whilst the incorriing fluid passes up the column. This requires that the size and density of the beads be controlled to achieve the down flow. This can be achieved with the materials of this invention by adjusting the micro-porosity by activation and the mggg/macro porosity by control of the pore forming process .

This type of process is disclosed in US 6716344(the disclosure of which is incorporated herein by reference) specifically for use in ion exchange systems but has never been applied to waste water treatment. More specifically with reference to Fig. 17, fluidised bed apparatus comprises a first liquid fluidized bed 10 and a second steam fluidized bed 12 interconnected at their adjacent ends by solid transfer 14 .

The first fluidized bed 10 is a counter-current flow bed wherein solids (carbon beads) as indicated at 18 enter adjacent to the top of the bed 10 as indicated by the line 17 and flow downward and a first fluidizing fluid namely water to be treated 20 enters the bed 10 as indicated schematically at 22 at the bwer end 24 of the bed 10 and flows upward in counter current with the carbon beads 18. Clean water exits the bed at 44

The second fluidized bed 12 on the other hand is a riser fluidized bed wherein the carbon beads 18 transferred from bed 10 via transfer system 14 enter the bed 12 adjacent to the lower end 26 of the bed 12 and flow upward in co-current relation with a steam regeneration stream 28 which enters the bed 12 under pressure in tlue illustrated arrangement via nozzle 30 and inlet 32 both adjacent to the lower end 26 of the bed 12 and flows upward through the bed 12 carrying the particles 18 in its flow.

The steam is produced in the steam generator 62 from clean water introduced at 28. It.JlfiteE?.. into the bottom 26 of the second fluidized bed 12 through a perforated plate inlet 32. The function of the au-dliary stream 62 is to stir up the particles at the bottom of the second fluidized bed 12 to be entrained up the second fluidized bed.

At the top of column 12the steam and carbon beads are separated in the cyclone 16, TJ cleaned dry beads are returned to the main liquid fluidized bed via line 17 whilst the steam, containing the organic material removed by the regeneration is transferred to a condenser 36.

Claims

Claims

1. A method for treating water contaminated with pesticide or pharmaceutical residues which method comprises contacting the contaminated water with a micro- mesoporous or micro-macro-porous carbon that has been activated in carbon dioxide to minimise the introduction of surface oxygen functionality and that has low heteroatom surface functionality.

2. The method of claim 1 , wherein the carbon is a phenolic resin derived carbon.

3. The method of claim 1 or 2, wherein the carbon has a pore size distribution showing a first large population of micropores of size < 2 nm and a second population of meso and/or macropores of size > 20 nm.

4. The method of any preceding claim, in which the carbon has a surface area of 600-2000m²/g.

5. The method of any preceding claim, wherein the carbon is from carbonization and activation of a mesoporous or macroporous phenolic resin.

6. The method of any preceding claim, wherein the activation of the carbon is carried out in carbon dioxide at a temperature greater than 800°C to give a weight loss of between 10 and 70%, preferably between 20 and 50%.

7. The method of any preceding claim, wherein the carbon is in the form of beads.

8. The method of any of the preceding claims where the carbon is in the form of irregular granules

9. The method of any of claims 1-5, wherein the carbon is in the form of an extruded monolith structure.

10. The method of claim 9, wherein the extruded carbon is in the form of trilobe, rachig ring or other such physical form designed to provide adequate bed porosity to allow an acceptable pressure drop and good kinetics as exemplified by an enhanced surface: volume ratio compared to a simple cylindrical extrudate

11. The method of claim 9, wherein the monolithic porous carbon is the result of: i) partially curing a phenolic resin to a solid;

ϋ) comminuting the partially cured resin;

iii) extruding the comminuted resin;

iv) sintering the extruded resin so as to produce a form-stable sintered product; and v) carbonising and activating the form-stable sintered product.

12. The method of any preceding claim, wherein the method is for removal of metaldehjyde.

13. The method of any preceding claim, which is for removal of water from one or more of atrazine, estadiol, propranolol. Malathion, carbamazepine, steroidal oestrogen, PPCP's (Personal care products), mecoprop, clopyralid and quinmerac.

14. The ,method of any preceding claim, wherein the potentially contaminated water comprises water from ground or surface water sources or water in the course of treatment at a water treatment works.

15. The method of any preceding claim, further comprising the step of steam regenerating the carbon.

16. A method for contacting the carbon with the contaminated water in a down flow mode that involves the use of two or more fixed carbon beds whereby at least one of the beds is taken off stream, purged with air to remove excess water, and regenerated in situ with steam.

17. A method for contacting the carbon with the contaminated water that involves the use of a moving carbon bed such that the contaminated carbon is continuously removed from the liquid inlet to the bed for regeneration and is continuously replaced by fresh or regenerated carbon at the clean water exit from the bed.

18. A method for contacting the carbon with the contaminated water that involves the use of a fluidised or expanded bed system whereby the saturated carbon is continuously removed from the bed, regenerated and returned to the bed, whilst the purified water is continuously removed from the top of the fluidized or expanded column.

19. A method for reactivating the carbon once it has been saturated with the chemical residues that involves contacting the carbon with steam at between 120 and 250°C, preferably between 150 and 200°C.

20. Apparatus for the treatment of water liable to be contaminated with one or more organic contaminants, said apparatus comprising:

a bed of carbon adsorbent configured for through-flow of water;

a supply line configured for delivery of water liable to be contaminated to said bed; and

a delivery line for receiving treated water flowing from the bed;

wherein the carbon comprises a micro-mesoporous or micro-macro-porous carbon that has been activated in carbon dioxide to minimise the introduction of surface oxygen functionality and that has low heteroatom surface functionality.

21 The apparatus of claim 20, wherein the carbon is from carbonization and activation of a mesoporous or macroporous phenolic resin.

22. The apparatus of claim 20 or 21 , wherein the carbon is in the form of beads.

23 The apparatus of claim 22, wherein the bed and the supply and delivery lines are configured for operation in expanded bed mode.

24. The apparatus of claim 22, wherein the bed is configured as a fluidised bed.

25. The apparatus of claim 22, wherein the bed is configured as an inverse (down- flow) liquid-solid fluidised bed system

26. The apparatus of any of claims 20-25, further comprising a steam generator configured to supply steam to the bed for regeneration thereof.

27. The apparatus of claim 26, wherein the steam generator is configured to supply dry superheated steam to the bed.

28.. The apparatus of claim 26, wherein the steam generator is configured to supply saturated steam to the bed.

29. The apparatus of any of claims 26-28, wherein a line from the bed leads to an oily water separator for separating contaminant organics from condensate.