CA2821581A1 - Method of preparing porous carbon - Google Patents

Abstract

The present invention relates to methods for preparing porous carbon material, and in particular to methods designed to produce porous carbon which exhibits selectivity for low molecular weight aldehydes, such as formaldehyde, and for hydrogen cyanide. The method comprises preparing the porous carbon in the presence of a nitrogen-donating agent. The invention also relates to filter elements and smoking articles comprising the porous carbon.

Description

Method of Preparing Porous Carbon Field of the Invention The present invention relates to methods for preparing porous carbon material, and in particular to methods designed to produce porous carbon which exhibits selectivity for low molecular weight aldehydes, such as formaldehyde, and for hydrogen cyanide. The selective porous carbon is particularly useful for smoke filtration in smoking articles, as the porous structure provides improved adsorption of these smoke vapour phase constituents which are generally poorly adsorbed by conventional activated carbon.
Background to the Invention Filtration is used to reduce certain particulates and/or vapour phase constituents of tobacco smoke inhaled during smoking. It is important that this is achieved without removing significant levels of other components, such as organoleptic components, thereby degrading the quality or taste of the product.
Smoking article filters may include porous carbon materials to adsorb certain smoke constituents, typically by physisorption. Such porous carbon materials can be made from the carbonized form of many different organic materials, most commonly plant-based materials such as coconut shell. Alternatively, synthetic carbons are used, such as resins prepared by polycondensation reactions.
Activated carbon materials have become widely used as versatile adsorbents owing to their large surface area, microporous structure, and high degree of surface reactivity. In particular, these materials are especially effective in the adsorption of organic and inorganic pollutants due to the high capacity of organic molecules to bind to carbon.
Activated carbons are commonly produced from materials including coconut shell, wood powder, peat, bone, coal tar, resins and related polymers. Coconut shell is particularly attractive as a raw material for the production of activated carbon because it is cheap and readily available, and is also environmentally sustainable.

Furthermore, it is possible to produce from coconut shell activated carbon material which is highly pure and has a high surface area.
An alternative source of microporous carbon is synthetic carbons, such as those formed by a polymerisation reaction, such as resin-based synthetic carbons.
Such carbons may, for example, be prepared by polycondensation of an aldehyde and a phenol.
These synthetic carbons are attractive because some of their physical properties can be controlled during manufacturing, allowing them to be tailored to provide desired filtration characteristics. However, these materials are significantly more expensive than activated coconut carbon and the like.
The performance and suitability of porous carbon material as an adsorbent in different environments is determined by various physical properties of the material, including the shape and size of the particles, the pore size, the surface area of the material, and so on. These various parameters may be controlled by manipulating the process and conditions by which the porous carbon is produced.
Generally, the larger the surface area of a porous material, the greater is the adsorption capacity of the material. However, as the surface area of the material is increased, the density and the structural integrity are reduced. Furthermore, while the surface area of a material may be increased by increasing the number of pores and making the pores smaller, as the size of the pores approaches the size of the target molecule, it is less likely that the target molecules will enter the pores and adsorb to the material. This is particularly true if the material being filtered has a high flow rate relative to the activated carbon material, as is the case in a smoking article.
The precise method used to manufacture porous carbon material has a strong influence on its properties. It is therefore possible to produce carbon particles having a wide range of shapes, sizes, size distributions, pore sizes, pore volumes, pore size distributions and surface areas, each of which influences their effectiveness as adsorbents. The attrition rate is also an important variable;
low attrition rates are desirable to avoid the generation of dust during high speed filter manufacturing.
As explained in Adsotption (2008) 14: 335-341, conventional coconut carbon is essentially microporous, and increasing the carbon activation time results in an increase in the number of micropores and surface area but produces no real change in pore size or distribution.
In accordance with nomenclature used by those skilled in the art, pores in an adsorbent material that are less than 2 nm in diameter are called "micropores", and pores having diameters of between 2 nm and 50 nm are called "mesopores". Pores are referred to as "macropores" if their diameter exceeds 50 nm. Pores having diameters greater than 500 nm do not usually contribute significantly to the adsorbency of porous materials.
Whilst it is well established that activated carbon material exhibits excellent general filtration of unwanted substances from the vapour phase of tobacco smoke, there are some smoke vapour constituents that are poorly adsorbed and these include low molecular weight aldehydes (such as formaldehyde) and hydrogen cyanide (HCN).
The presence of free groups on the surface of the porous carbon material has been found to also affect the carbon's adsorption properties. It is known that the presence of free nitrogen groups may enhance the selective adsorption of constituents including low molecular weight aldehydes and HCN.
The present invention seeks to provide a method for preparing porous carbon materials which have nitrogen-containing groups on the surface of the carbon, to enhance selective adsorption of low molecular weight aldehydes and HCN.
The present invention seeks to provide porous carbon materials having nitrogen-containing groups on their surfaces.

Summary of the Invention Accordingly, in a first aspect of the invention there is provided a method of preparing porous carbon with adsorbent properties for use in smoke filtration, the method comprising preparing the porous carbon in the presence of a nitrogen-donating agent.
According to a second aspect of the invention, a porous carbon is provided which is obtained or obtainable by a method according to the first aspect of the invention.
According to a third aspect of the invention, a filter element for a smoking article is provided, comprising a porous carbon according to the second aspect of the invention.
According to a fourth aspect of the invention, a smoking article is provided, comprising a porous carbon according to the second aspect of the invention.
Detailed Description of the Invention The present invention relates to a method involving the addition of nitrogen-containing groups to the surface of porous carbon by preparing the carbon in the presence of a nitrogen-donating agent. Preferably the porous surface structure of the carbon is formed in the presence of the nitrogen-donating agent, so that the nitrogen groups are present within the porous structure In one embodiment of the invention, the porous carbon is a resin-based synthetic carbon, such as the carbon prepared by polycondensation of an aldehyde and a phenol. If available, commercially available polycondensates may be used.
To produce the polycondensate, the starting material may be a phenolic compound such as phenol, resorcinol, catechin, hydrochinon and phloroglucinol, and an aldehyde such as formaldehyde, glyoxal, glutaraldehyde or furfural. A commonly used and preferred reaction mixture comprises resorcinol (1,3-dihydroxyben2ol) and formaldehyde, which react with one another under alkaline conditions to form a gel-like polycondensate. The polycondensation process will usually be conducted under aqueous conditions.
To produce the polycondensate, the reaction mixture may be warmed. Usually, the polycondensation reaction will be carried out at a temperature above room temperature and preferably between 40 and 90 C. According to the present invention, the polycondensation reaction is carried out in the presence of a nitrogen-donating agent so that the resultant resin has an increased nitrogen content and increased presence of nitrogen-containing groups on its surface. In preferred embodiments, for example, an aqueous solution containing the nitrogen-donating agent is added to an aqueous solution of resorcinol and formaldehyde under vigorous stirring to yield a homogeneous solution. This solution is then incubated to provide the polycondensate. The incubation period may be between 5 minutes and 24 hours.
The rate of the polycondensation reaction as well as the degree of crosslinking of the resultant gel can, for example, be influenced by the relative amounts of the alcohol and catalyst. The skilled person would know how to adjust the amounts of these components used to achieve the desired outcome.
In order to produce particles of a desired size, it has been shown to be advantageous to reduce the size of the polycondensate before further processing.
The size reduction of the polycondensate may be carried out using conventional mechanical size reduction techniques or grinding. It is preferred that the size reduction step results in the formation of granules with the desired size distribution, whereby the formation of a powder portion is substantially avoided.
In one embodiment of the present invention, the polycondensate (which has optionally been reduced in particle size) then undergoes pyrolysis. The pyrolysis may also be described as carbonisation.
According to an embodiment of the invention, the surface properties of the resultant carbon are changed by treating the polycondensate before, during or after pyrolysis with a nitrogen-donating agent, optionally as well as with other more conventional means, such as steam, air, COõ oxygen or a mixture of gases, which may be diluted with nitrogen or another inert gas. It is particularly preferred to use a mixture of nitrogen and steam.
The activation stage preferably takes place in a gaseous atmosphere comprising nitrogen, water and/or carbon dioxide. In other words, the dried gels used in the present invention may be non-activated or, in some embodiments, activated, for example steam activated or activated with carbon dioxide. Activation is preferred in order to provide an improved pore structure.
Where the starting material is a carbon precursor, the carbon precursor is preferably pyrolysed before then being activated. Conventional methods of pyrolysis may be used. The nitrogen-donating agent may be added to the material before or after the pyrolysis step, but it is preferably added before any pyrolysis step and before the activation step.
Pyrolysis (or carbonisation) is a chemical process of incomplete combustion of a solid when subjected to high heat. By the action of heat, pyrolysis removes hydrogen and oxygen from the solid, so that the remaining product, the char, is composed primarily of carbon. Suitable pyrolysis or carbonisation methods that may be used include those that will be familiar to the skilled person, such as the pit method, the drum method, and destructive distillation. The incubation temperature and time may be between 300 C and 1000 C, and between 30 minutes and 4 hours, respectively. For example, the pyrolysis step may involve heating the pre-treated carbon to a temperature of at least 500 C and maintaining the carbon at that temperature for a number of hours. In one embodiment, the pyrolysis step involves heating the pre-treated carbon at a rate of 5-10 C/minute to 600 C under N, flowing at a rate of 10-200 cm3/min. In one embodiment, the pyrolysis step is carried out at a temperature of no more than 600 C, more preferably at a temperature of no more than 550 C, or of about 500 C. Pyrolysis at these temperatures is preferred as they provide a high nitrogen content and good surface area. Pyrolysis at higher temperatures may lead to a reduction in nitrogen content and can result in a lower surface area due to structural shrinkage.
After pyrolysis, the carbon is cooled and the carbon surface is preferably deactivated, for example by exposure to a humid N, flow. This deactivation is necessary because of the high risk of exothermic 02 adsorption causing red-heat.
Subsequently, to increase the surface area, the pyrolysed carbon is activated.
This may be done by either physical or chemical means, and conventional activation techniques can be used. Preferably the material is activated by physical means, and most preferably the material is activated using nitrogen and steam, or alternatively, CO,.
In one embodiment of the invention, the material is activated by reaction with steam under controlled nitrogen atmosphere in a kiln such as a rotary kiln.
The temperature is important during the activation process. If the temperature is too low, the reaction becomes slow and is uneconomical. On the other hand, if the temperature is too high, the reaction becomes diffusion controlled and results in loss of the material.
Activation of the material using nitrogen and steam may be performed at a temperature of between 600 C and 1100 C, and preferably activation is performed at a temperature of between 700 C and 900 C. Most preferably, the material is activated at about 850 C. The activation process is preferably carried out for between 30 minutes and 4 hours. Most preferably, the material is activated for hour. As the temperature is increased, the nitrogen content is decreased.
In an alternative embodiment, the material is activated by reaction with carbon dioxide. In this case, activation of the material may be performed at a temperature of between 400 C and 1000 C, and preferably activation is performed at a temperature of between 600 C and 800 C. The activation process is preferably carried out for between 30 minutes and 4 hours.

Whilst the precise mechanism by which the nitrogen-containing groups are transferred to the surface of the carbon is not known, it is believed that the donating agent is able to add nitrogen-containing groups to the carbon's surface and these groups act as preferential sites for aldehyde and HCN adsorption. This hypothesis is supported by the experimental data provided and discussed below, and in particular by the XPS (X-ray photoelectron spectroscopy) results.
It has also been demonstrated that the porous carbon produced according to the methods of the present invention exhibit significantly increased selectivity for formaldehyde and HCN.
According to the present invention, the nitrogen-donating agent may be an amino acid or amino acid derivative, an amine, including an aromatic amine, or an imidazole, an imidazole derivative or a compound including the pyridine-like nitrogen of imidazole, such as 1-methylimidazole. In a preferred embodiment, the nitrogen-donating agent is lysine, L-hydroxylysine, L-arginine, L-histidine, L-aspartic acid, 1-methylimida2ole (MIM), ethylenediamine (EDA), propylamine, dimethyamine, 2-propylamine, trimethylamine or aniline. Particularly preferred agents are lysine, MIM and EDA.
The nitrogen-donating agent is preferably added to the constituent reagents of the polycondensate in the form of an aqueous solution. This solution may be added to the mixture of phenolic compound and aldehyde prior to polymerisation. The amount of nitrogen-donating agent used as a molar ratio to the amount of the phenolic compound may be between 30:1 and 3:1 (phenolic compound: nitrogen donating agent). The molar ratio of phenolic compound to water is preferably in the range of between 1:3 and 1:50 (phenolic compound: water).
The surface areas of activated carbon materials are estimated by measuring the variation of the volume of nitrogen adsorbed by the material in relation to the partial pressure of nitrogen at a constant temperature. Analysis of the results by mathematical models originated by Brunauer, Emmett and Teller results in a value known as the BET surface area.

The BET surface area of the activated carbon materials produced by the method is still important for the adsorption of smoke constituents other than low molecular weight aldehydes and HCN. In particular, the activation step may be controlled to ensure that the resultant product contains the desired volume of micropores.
The porous carbon materials produced according to the present invention preferably have a BET surface area of at least 400, 450, 500, 550, 600, 650, 700, 750, 800, 900, 1000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800 or at least 1900 m2/g.
Porous carbon materials with BET surface areas of between 500 m2/g and 1300 m2/g are preferred, and material with surface areas of between 600 m2/g and 1200 m2/g are most preferred.
The porous carbon materials of the invention preferably have a pore volume (as estimated by nitrogen adsorption) of at least 0.3 cm9g, and desirably at least 0.5 cm9g. Carbon materials with pore volumes of at least 0.5 cm9g are particularly useful as an adsorbent for tobacco smoke. Carbon materials according to the invention with pore volumes significantly higher than 1 cm9g are low in density and are therefore less easy to handle in cigarette production equipment. Such carbon materials are less favourable for use in cigarettes or smoke filters for that reason.
The activated carbon produced by the methods of the present invention may be provided in monolithic or particulate form. Particles will preferably have a particle size in the range of between 10 p.m and 1500 ,m. Preferably the mean particle size is between 100 ,m and 1000 p.m, and more preferably between 150 p.m and 800 p.m.
Most preferably, the particles of activated carbon material have a mean size of between 250 ,m and 750 p.m. The smaller the particles are, the larger is the combined surface area, however, if the size of the particles used is too small, the particles can interfere with manufacturing processes, especially high speed processes as used to manufacture cigarette filters.
Experiments Lysine catalyzed carbon monolith Two carbon samples, 08-12-05 and 09-05-04 (in the form of carbon monoliths) were prepared by polycondensation of resorcinol and formaldehyde in the presence of lysine.
Sample 08-12-05 The mass percentage of (resorcinol+formaldehyde) and lysine in solution was approximately 30 weight A and molar ratio of resorcinol:lysine was 6.6 with thermal curing (1 day at 50 C and 1 day at 90 C). The obtained bulk polymer was dried at 50 C for 1 day and pyrolysed at 700 C for 2 hours under nitrogen atmosphere.
The resultant carbon had a BET surface area of 460 m2/g, a total pore volume of 0.23 cm3/g, and a micropore volume of 0.22 cm3/g. The surface nitrogen was detected using IR but the surface groups were not identified.
Sample 09-05-04 The process used to synthesise sample 08-12-05 was repeated, except that the surface area and porosity were slightly increased, so that the sample had a BET
surface area of 580 m2/g, a total pore volume of 0.27 cm3/g and a micropore volume of 0.24 cm3/g.
Performance of both of these carbon samples in a cigarette was measured by placing 60 mg into the cavity filter of a reference cigarette. The filter construction was a triple filter in the form of cellulose acetate ¨ carbon granules ¨ cellulose acetate.
Filters having an empty cavity, or an equal weight of sorbite (coal based carbon) were used as controls. Sorbite was chosen as a control carbon because it has similar physical properties to sample 08-12-05 (460 m2/g surface area, 0.26 cm3/g total pore volume and 0.25 cm3/g micropore volume).
Once prepared, cigarettes were aged at 22 C and 60% Relative Humidity for approximately 3 weeks prior to smoking.
All cigarettes were smoked under ISO conditions, i.e. a 35m1 volume puff of 2 second duration was taken every minute. The smoke chemistry results are shown below in Table 1, which provides the percentage reductions of various smoke constituents.
Table 1 Carbon Sorbite 08-12-05 09-05-04 Acetaldehyde 10 14 16 Acetone 25 9 10 Acrolein 22 16 26 Butyraldehyde 30 13 20 Crotonaldehyde 42 24 48 Formaldehyde 16 32 67 Methyl ethyl ketone 36 6 15 Propionaldehyde 21 9 12 1,3-butadiene 12 12 11 Acrylonitrile 22 28 8 Benzene 19 5 39 Isoprene 24 10 20 Toluene 13 1 28 A 67% reduction in formaldehyde (as observed using sample 09-05-04) by carbon is outstanding. The reduction in HCN is also significantly improved by using both of the samples produced according to the present invention in comparison to conventional activated carbon.
Further samples of materials produced by polycondensation of resorcinol and formaldehyde in the presence of lysine were pyrolysed at different temperatures (400, 500, 600, 700 and 800 C). The samples were not subsequently activated.
The nitrogen content was measured by element analysis and XPS measurements (Axis Ultra, Kratos Analytical) and the results are set out in Table 2 below.

Table 2 - Element analysis of the carbon monoliths Sample Pyrolysis N 0 temperature ( C) (wt.%) (wt.%) (wt.%) (wt.%) 1 400 1.71 24.67 3.07 70.55 2 500 1.92 10.15 1.68 86.25 3 600 1.84 9.11 1.66 87.40 4 700 1.43 4.03 0.61 93.93 800 1.28 3.44 0.34 94.89 XPS measurements (Axis Ultra, Kratos Analytical) on N ls signal are enabled to reveal the changes occurring in nitrogen species present on carbon surfaces after 5 pyrolysis. The fitting of the N Is peaks gave the following binding energies: 399.8 0.3 eV (amide and/or pyrrolic nitrogen), 401.4 0.3 eV (quaternary nitrogen), and 402.8 eV (pyridine-N-oxide). Upon pyrolysis, the nitrogen species within the material change significantly with the temperature applied. The nitrogen content of samples are accordingly of 1.3, 1.9, 1.3, 0.8 and 0.5 wt%.
The XPS results clearly show that nitrogen has been incorporated into the carbon structure. They also show that the pyrolysis temperature is significant and that the nitrogen content begins to drop significantly when the carbon is pyrolysed at a temperature above 600 C. The optimal temperature for nitrogen incorporation appears to be around 500 C.
Alternative nitrogen-donating agents Three carbon samples were evaluated against a conventional reference cigarette for filter additive studies. The samples were based on a resorcinol-formaldehyde resin with increased nitrogen content via synthesis with either 1-methylimida2ole (MIM) or ethylenediamine (EDA). This work was conducted to establish whether other nitrogen-containing molecules could be used to increase the activated carbon nitrogen content which would then also show similar reductions in formaldehyde and HCN.

Table 3 sets out the properties of the carbon samples that were prepared. RF
polymer means resorcinol formaldehyde polymer, and the catalyst used was either 1-methylimida2ole (MIM) or ethylenediamine (EDA). The polycondensation of resorcinol with formaldehyde in the presence of MIM was conducted at 50 C for hours and an ambient pressure drying followed by incubation at 800 C. For the polycondensation in the presence of EDA, thermal curing was carried out at 90 C
for 4 hours and, after drying, the sample was thermally treated at 800 C under N2.
Table 3 Sample Carbon Type Surface area Micropore Total pore code (m2/g) volume volume (cm3/g) (cm3/g) 09-12-01 RF polymer + MIM 670 0.27 0.37 (R:MIM=26:1) 09-12-02 RF polymer + MIM 580 0.23 0.33 (R:MIM=13:1) 09-12-03 RF polymer + EDA 520 0.22 0.25 The micropore volumes of the samples were lower than that of coconut carbon currently typically used in cigarette filters (which is generally 0.4-0.5 cm3/g). This was expected to have an effect on the adsorption characteristics of the samples compared to the reference cigarette.
Infra red analysis showed that a number of different nitrogen-containing groups were present on the surfaces of the samples, including NH and CN groups.
60 mg of the carbon samples was weighed into the cavity filter of a reference cigarette. The filter construction was a triple filter in the form of cellulose acetate ¨
carbon granules ¨ cellulose acetate. A filter having an empty cavity of similar dimensions was used as a control.
Once prepared, the cigarettes were aged at 22 C and 60% Relative Humidity for approximately 3 weeks prior to smoking.

All cigarettes were smoked under ISO conditions, i.e. a 35m1 volume puff of 2 second duration was taken every minute. The smoke chemistry results are shown below in Table 4.
Table 4 - Smoke analysis results Smoke Yields ( /0 Reductions) Carbon Empty 09-12-01 09-12-02 09-12-Puff No 6.8 6.9 6.7 6.6 NFDPM (mg/cig) 11.2 9.7 8.4 7.1 Nicotine 0.93 0.83 0.72 0.59 Water 2.5 1.8 1.5 0.8 CO 11 11.2 10.3 10.6 Acetaldehyde 550.1 299.7 (46) 223 (59) 470 (15) ( g/cig) Acetone 285.1 217.6 (24) 116.9 (59) 267.5 (6) Acrolein 64.3 14.7 (77) 4.7 (93) 48.9 (24) Butyraldehyde 37.4 29.3 (22) 19.3 (48) 31.6 (16) Crotonaldehyde 21.4 3.8 (82) 1.5 (93) 12.4 (42) Formaldehyde 34.6 13.5 (61) 11.7 (66) 14.5 (58) Methyl ethyl ketone 68.3 52.5 (23) 26.9 (61) 62.3 (9) Propionaldehyde 48.3 35.6 (26) 19.6 (59) 45.6 (6) HCN 122.1 15.7 (87) 16 (87) 45 (63) 1,3-butadiene 72.8 47.7 (34) 35.5 (51) 57.6 (21) Acrylonitrile 15.2 3.5 (77) 2.3 (85) 8.3 (45) Benzene 53.5 40.9 (24) 27.3 (49) 41.4 (23) Isoprene 644 588 (9) 428 (34) 515 (20) Toluene 73.4 51.5 (30) 28.6 (61) 54.4 (26) All of the samples show much higher reductions in formaldehyde and HCN than is usually observed with activated carbon (compare to the reductions observed using Sorbite provided in Table 1).

The granule strength of the samples was weak and it is likely that some 'dust' was generated causing an increase in pressure drop and thus resulting in the lower yields of NFDPM and nicotine observed (i.e. the smoke particulate phase).
Differences in pressure drop will not affect vapour phase analyte yields and, although some caution should be exercised, percentage reductions are shown without normalising to tar.
It is clear that all of the agents used to incorporate nitrogen into the activated carbon structure have resulted in enhanced adsorption towards formaldehyde and HCN. The greater the nitrogen addition, the greater the reductions, as shown using MIM, where despite a lower surface area and pore volume, sample 09-12-02 (having R:MIM of 13:1) outperformed sample 09-12-01 (in which R:MIM was 26:1).
Sample 09-12-03 gave small reductions in the majority of smoke analytes measured (as expected using a physisorption mechanism with a surface area of only about 500m2/g). However enhanced selectivity (via a chemisorption mechanism) was shown towards formaldehyde and HCN.
From the results of the experiments described above, it appears that the precise nitrogen source is not critical. The experimental data show that the use of activated carbons produced using lysine, MIM, or EDA all provide excellent reductions in smoke formaldehyde and HCN.

Claims (15)

1. A method of preparing a smoking article filter element comprising a porous carbon with adsorbent properties, the method comprising preparing the porous carbon in the presence of a nitrogen-donating agent, wherein the nitrogen-donating agent is an amino acid Or amino acid derivative, or an imidazole, an imidazole derivative or a compound including the pyridine-like nitrogen of imidazole, such as 1-methylimidazole.

2, A method as claimed in claim 1, wherein the porous carbon is a resin-based synthetic carbon.

3. A method as claimed in claim 2, wherein the porous carbon is prepared by polycondensation of an aldehyde and a phenolic compound.

4. A method as claimed in claim 3, wherein the nitrogen-donating agent is added to the aldehyde and the phenolic compound prior to the polycondensation.

5. A method as claimed in either of claims 3 or 4, wherein the molar ratio of the amount of phenolic compound to the amount of nitrogen-donating agent used is between 30:1 and 3:1,

6. A method as claimed in any of claims 3 to 5, wherein the phenolic compound is resorcinol.

7. A method as claimed in claim 1, wherein the porous carbon is prepared by activating a carbon precursor.

8. A method as claimed in claim 7, wherein the carbon precursor is pyrolysed prior to activation.

9. A method as claimed in claim 7 or claim 8, wherein the nitrogen-donating agent is added to the carbon precursor prior to activation.

10. A method as claimed in claim 9, wherein the nitrogen-donating agent is added to the carbon precursor prior to any pyrolysis step.

11. A method as claimed in claim 10, wherein the pyrolysis step is carried out at a temperature of no higher than 600°C.

12. A method as claimed in any one of the preceding claims, wherein the nitrogen-donating agent is lysine, 1-methylimidazole (MIM) or ethylenediamine (EDA),

13. A smoking article filter element obtained or obtainable by a method as claimed in any one of the preceding claims.

14. A smoking article comprising a smoking article filter element as claimed in claim 13.

15. Use of a smoking article filter element as claimed in claim 13 or a smoking article as claimed in claim 14 for the filtration of tobacco smoke.