Porous carbon materials and smoking articles and smoke filters therefor incorporating such materials

Abstract

A porous carbon material suitable for incorporation in smoke filters for cigarettes has a BET surface area of at least 800 m.sup.2/g and a pore structure that includes mesopores and micropores. The pore volume (as measured by nitrogen adsorption) is at least 0.9 cm.sup.3/g and from 15 to 65% of the pore volume is in mesopores. The pore structure of the material provides a bulk density generally less than 0.5 g/cc. The material may be produced by carbonizing and activating organic resins and may be in the form of beads for ease of handling.

Claims

1. A smoking article comprising smoking material and a porous carbon material, wherein the porous carbon material has: a BET surface area of about 900 to about 1638 m.sup.2/g, a pore structure that includes mesopores and micropores, a pore volume (as measured by nitrogen adsorption) of at least 0.9 cm.sup.3/g, a bulk density of not more than 0.5 g/cc, 15 to 65% of the pore volume of the porous carbon material (as measured by nitrogen adsorption) is in mesopores, and the porous carbon material is in particulate form having a mean particle size of 50 to 1000 micron.

2. A smoking article according to claim 1 comprising a rod of smoking material and a filter, and the porous carbon material is incorporated in the filter.

3. A smoking article according to claim 1, wherein the pore volume (as measured by nitrogen adsorption) of the porous carbon material is at least 1.0 cm.sup.3/g and from 30 to 65% of the pore volume is in mesopores.

4. A smoking article according to claim 1, wherein less than 20% of the pore volume of the porous carbon material is in pores having diameters in the range 2-10 nm.

5. A smoking article according to claim 1, wherein the porous carbon material has a BET surface area of 900 to 1300 m.sup.2/g.

6. A smoking article according to claim 5, wherein the porous carbon material has a BET surface area of from 1000 to 1250 m.sup.2/g.

7. A smoking article according to claim 1, wherein the pore volume of the porous carbon material in micropores and mesopores is from 1.1 to 2 cm.sup.3/g.

8. A smoking article according to claim 1, wherein from 35 to 55% of the pore volume of the porous carbon material is in mesopores.

9. A smoking article according to claim 1, wherein the porous carbon material is in the form of microbeads.

10. A smoking article according to claim 1, wherein the porous carbon material has a mean particle size of 50 to 700 microns.

11. A smoking article according to claim 10, wherein the porous carbon material has a mean particle size of 150 to 500 microns.

12. A smoking article according to claim 1, wherein the porous carbon material has a D90/D10 particle size distribution of at least 10.

13. A smoking article according to claim 1, wherein the porous carbon material is substantially free of particles smaller than 10 microns.

14. A smoking article according to claim 1, wherein the porous carbon material is composed of a carbonised organic resin.

15. A smoking article according to claim 14, wherein the organic resin contains nitrogen.

16. A smoking article according to claim 15, wherein the resin is produced by condensing a nucleophilic component with an electrophilic cross linking agent in the presence of a pore former.

17. A smoking article according to claim 15, wherein the nucleophilic component or the cross-linking agent is an organic nitrogen compound.

18. A smoking article according to claim 15, wherein the nucleophilic component comprises a novolak resin.

19. A smoking article according to claim 15, wherein the crosslinking agent comprises hexamethylene tetramine.

20. A smoking article according to claim 15, wherein the pore former comprises ethylene glycol.

21. A smoking article filter comprising a porous carbon material having: a BET surface area of about 900 to about 1638 m.sup.2/g, a pore structure that includes mesopores and micropores, a pore volume (as measured by nitrogen adsorption) of at least 0.9 cm.sup.3/g, a bulk density of not more than 0.5 g/cc, 15 to 65% of the pore volume of the porous carbon material (as measured by nitrogen adsorption) is in mesopores, and the porous carbon material is in particulate form having a mean particle size of 50 to 1000 microns.

22. A smoking article filter according to claim 21, wherein the pore volume (as measured by nitrogen adsorption) of the porous carbon material is at least 1.0 cm.sup.3/g and from 30 to 65% of the pore volume is in mesopores.

23. A smoking article filter according to claim 21, wherein less than 20% of the pore volume of the porous carbon material is in pores having diameters in the range 2-10 nm.

24. A smoking article filter according to claim 21, wherein the porous carbon material has a BET surface area of 900 to 1300 m.sup.2/g.

25. A smoking article filter according to claim 24, wherein the porous carbon material has a BET surface area of from 1000 to 1250 m.sup.2/g.

26. A smoking article filter according to claim 21, wherein the pore volume of the porous carbon material in micropores and mesopores is from 1.1 to 2 cm.sup.3/g.

27. A smoking article filter according to claim 21, wherein from 35 to 55% of the pore volume of the porous carbon material is in mesopores.

28. A smoking article filter according to claim 21, wherein the porous carbon material is in the form of microbeads.

29. A smoking article filter according to claim 21, wherein the porous carbon material has a mean particle size of 50 to 700 microns.

30. A smoking article filter according to claim 29, wherein the porous carbon material has a mean particle size of 150 to 500 microns.

31. A smoking article filter according to claim 21, wherein the porous carbon material has a D90/D10 particle size distribution of at least 10.

32. A smoking article filter according to claim 21, wherein the porous carbon material is substantially free of particles smaller than 10 microns.

33. A smoking article filter according to claim 21, wherein the porous carbon material is composed of a carbonised organic resin.

34. A smoking article filter according to claim 33, wherein the organic resin contains nitrogen.

35. A smoking article filter according to claim 33, wherein the resin is produced by condensing a nucleophilic component with an electrophilic cross linking agent in the presence of a pore former.

36. A smoking article filter according to claim 33, wherein the nucleophilic component or the cross-linking agent is an organic nitrogen compound.

37. A smoking article filter according to claim 33, wherein the nucleophilic component comprises a novolak resin.

38. A smoking article filter according to claim 33, wherein the crosslinking agent comprises hexamethylene tetramine.

39. A smoking article filter according to claim 33, wherein the pore former comprises ethylene glycol.

Description

(1) In order that the invention may be better understood, preferred embodiments thereof will now be described, by way of example onlyin which reference will be made to the following Figures:

(2) FIG. 1which is a graph.

(3) FIG. 2which is a graph.

(4) FIG. 3which is a graph.

(5) FIG. 4awhich is a graph.

(6) FIG. 4bwhich is a graph.

(7) FIG. 4cwhich is a graph

(8) FIG. 4dwhich is a graph.

(9) FIG. 4ewhich is a graph.

(10) FIG. 4fwhich is a graph.

(11) FIG. 4gwhich is a graph.

(12) FIG. 4hwhich is a graph.

(13) FIG. 4iwhich is a graph.

(14) FIG. 4jwhich is a graph.

(15) FIG. 4kwhich is a graph.

(16) FIG. 5which is a graph.

(17) FIG. 6which is a graph.

(18) FIG. 7which is a diagrammatic representation of a smoking article.

(19) FIG. 8which is a diagrammatic representation of a smoking article.

(20) Referring to Table 1, samples of organic resins were prepared by mixing 100 parts by weight of the commercially available novolak phenol-formaldehyde resins specified in Table 1 with ethylene glycol pore former in the proportions indicated in Table 1 at elevated temperature and with stirring to enhance the formation of a clear solution, the temperature of which was then stabilised at 65-70 C. Hexamethylene tetramine (hexamine) cross-lining agent was then added in the proportions indicated in Table 1. The resulting stirred mixture was then heated to the temperature and for the specified reaction time.

(21) The commercial grades of novolak resins used were J1058F available from hexion Specialty Chemicals Inc (formerly Borden Chemical inc), with Mw of about 2400 and containing 5% by weight hexamethylene tetramine, TPR210, with Mw of about 1030, containing salicylic acid to catalyse cross-linking, and J1089F, with Mw of about 1110.

(22) In each case, the resulting viscous solution was poured as a stream with stirring into 2 to 4 times its volume of a preheated (115-120 C.) mineral oil containing 0.5% by volume of a drying oil (known commercially as Danish oil) to retard coalescence. The temperature of resulting emulsion initially dropped to 105-110 C., but on further heating cross-linking occurred at about 115-120 C. Further heating at the rate about 0.5 C. per minute up to 150 C. was applied to complete the reaction. After cooling, the resulting beads of resin were filtered off from the oil and washed several times with hot water to remove the majority of the ethylene glycol and a small amount (less than 5% of total) of low molecular weight polymer. The resulting porous spherical resin, containing water, residual oil, residual pore former and low molecular weight fraction was carbonised by heating at 800 C. to produce spherical porous carbon material. The carbon material was then activated with superheated steam, or carbon dioxide to achieve the weight reduction or burn-off indicated in Table 1.

(23) TABLE-US-00001 TABLE 1 Resin Ethylene Precursor Glycol Hexamine Novolak Ex. No (PBW) (PBW) (100 PBW) Activation Conditions 1 200 15 J1058F Steam at 850 C. 2 200 11 34% (CO.sub.2 + air modification) 3 200 11 34% (CO.sub.2) 4 200 11 38 5 200 11 34% (CO.sub.2) 6 300 11 33% (CO.sub.2) 7 200 15 37% (CO.sub.2) 8 400 11 36% (CO.sub.2 + air modification) 9 400 11 36% (CO.sub.2) 10 600 11 36% (CO.sub.2) 11 200 11 53% (CO.sub.2) 12 400 11 37% (CO.sub.2) 13 400 11 37% (CO.sub.2 + air modification)

(24) The resulting beads exhibited high durability and very low attrition rate as compared with carbon derived from coconut shell. In particular, when physically handled, the beads had little or no soiling effect on the hands when rubbed, and when physically agitated formed very little dust. The beads also had excellent flow characteristics, the spherical shape of the beads causing the material to flow easily and to form much flatter heaps, i.e. conical piles with a much lower slump angle, or angle of repose, than natural carbon.

(25) For the purposes of comparison, two further samples of carbon material (comparative Examples B and C) were prepared by a technique similar to that described above, using the ingredients and activation conditions referred to in Table 2. A sample of commercially available coconut charcoal, grade 208C was also used for the purposes of comparison (comparative Example A).

(26) TABLE-US-00002 TABLE 2 Resin Precursor Pore Cross-linking Activation Ex. No former agent Resin Conditions B 100 pbw 11 pbw hexamine 100 pbw Novolak 36% (CO.sub.2) ethylene J1058F glycol C water m-amino-phenol- 27% (CO.sub.2) formaldeyde A Activated coconut charcoal grade 208C

(27) FIGS. 1 to 3 show the distribution of pore size as determined by mercury porosimetry for the porous carbon materials of Examples 3 and 9, and comparative Example A (coconut charcoal). In each graph, the left-hand ordinate indicates the logarithm (log) of the differential intrusion of mercury into the sample in ml/g, the right hand ordinate indicates the cumulative intrusion of mercury in ml/g and the abscissa indicates, on a logarithmic scale, the pore size diameter in nanometers over the range 5 nm-110.sup.6 nm. The large peak to the left hand side of each graph is caused by the intrusion of mercury into the gaps between individual particles within the sample. The peaks towards the right hand side of the graph are caused by the intrusion of mercury into the micro, meso and macropores.

(28) FIGS. 4a-k show the distribution of pore size for the samples of the material as determined by nitrogen adsorption. In these drawings, the mean pore size in Angstrom units is plotted on a logarithmic scale on the abscissa against a value indicative of the number of pores of a particular size obtained from nitrogen adsorption studies, which is the differential of the pore volume with respect to the logarithm of the pore size (dV/d log R).

(29) The BET surface area and porosity of the activated porous carbon materials described in Tables 1 and 2 are shown in Table 3, and the corresponding properties of comparative samples A, B and C are shown in Table 4. The BET surface areas were calculated using the BET method over a range of partial pressure for nitrogen (P/P.sub.oN.sub.2) of from 0.07-0.3. The figures shown for nitrogen adsorption is the quantity of nitrogen in milliliters adsorbed under ambient conditions per gram of carbon sample at a relative nitrogen pressure (P/Po) of 0.98, but normalised by the specific gravity of liquid nitrogen at corresponding temperature.

(30) TABLE-US-00003 TABLE 3 Total BET Pore Mesopore Micropore Bead Surface Nitrogen Bulk Volume volume volume % % Ex size area sq absorption Density cc/g cc/g cc/g Micropore Mesopore No (m) m/g cc/g g/cc (by N.sub.2) (by N.sub.2) (by N.sub.2) (by N.sub.2) (by N.sub.2) 1 250-500 918 1.05 1.26 0.66 0.60 47.62 53.38 2 250-500 1077 1.04 0.41 1.30 0.58 0.72 55.38 44.62 3 250-500 1094 1.09 0.41 1.35 0.62 0.73 54.07 45.93 4 250-500 1155 1.16 0.59 1.41 0.66 0.75 53.19 46.81 5 150-250 1057 1.14 0.39 1.42 0.68 0.74 52.11 47.89 6 250-500 1165 1.25 0.27 1.51 0.73 0.78 51.66 48.34 7 250-500 1057 1.23 0.36 1.51 0.79 0.72 47.68 52.32 8 250-500 1203 1.27 0.23 1.53 0.72 0.81 52.94 47.06 9 250-500 1230 1.36 0.23 1.62 0.79 0.83 51.23 48.77 10 250-500 1139 1.37 0.25 1.62 0.87 0.75 46.30 53.70 11 250-500 1466 1.48 0.33 1.68 0.84 0.84 50.00 50.00 12 500-1000 1085 1.44 0.22 1.73 0.96 0.77 44.51 55.49 13 500-1000 1186 1.60 0.20 1.86 1.08 0.78 41.94 58.06

(31) TABLE-US-00004 TABLE 4 BET Total Bead Surface Nitrogen Bulk Pore Mesopore Micropore Ex size area sq absorption Density Volume volume volume % % No (m) m/g cc/g g/cc cc/g cc/g cc/g Micropore Mesopore A 996 0.55 0.5 0.78 0.04 0.74 94.87 5.13 B 250-500 1040 0.6 0.61 0.87 0.09 0.78 89.66 10.34 C 250-500 726 0.65 0.59 0.9 0.31 0.59 65.56 34.44

(32) Table 5 gives further details of the pore size distribution of Examples 2, 3, 4, 7, 8, 9, 12 and 13, and of comparative examples A and C.

(33) TABLE-US-00005 TABLE 5 Pore Size Distribution (N.sub.2 adsorption). (% of total pore volume by pore size) Example No >10 nm 5-10 nm 2-5 nm <2 nm 2 34.62 5.38 7.69 52.31 3 36.30 6.67 6.67 50.37 4 34.04 8.51 7.80 49.65 7 46.00 4.00 5.33 44.67 8 39.87 3.92 7.84 48.37 9 42.59 3.09 5.56 48.77 12 51.45 2.31 4.05 42.20 13 53.76 2.15 4.84 39.25 A 1.28 1.28 7.69 89.74 C 31.11 1.11 5.56 62.22

(34) It can be seen from the above Tables and FIGS. 1-4a-k that the carbon materials of the invention have distribution of pore sizes that extends over the micro-, meso- and, sometimes, macropore ranges. Although nitrogen adsorption cannot be used to estimate macropore volumes, the presence of significant macropore volumes is indicated by from the positive slope of the pore size distribution curve towards the upper end of the measurement range of nitrogen adsorption, as seen for Examples 8 (FIG. 4f), 9 (FIG. 4g), 10 (FIG. 4e) 12 (FIG. 4g) and 13 (FIG. 4i). The presence of macropores can be confirmed by mercury porosimetry studies, as shown in FIGS. 2 and 3.

(35) It can also be seen from the nitrogen adsorption measurements that there is a clear minimum in the pore size distribution of the carbon materials of the Examples of the invention, in each case in the range 2-10 m. Within this range the mesopores account for less than 20% of the combined meso and micropore volumes, usually less than 15% and more often less than 10% of the combined volume.

(36) The effect of the carbon materials of the examples upon tobacco smoke was tested by preparing standard cigarettes comprising a paper-wrapped rod of US blended style tobacco and a smoke filters 27 mm in length connected to the rod by a tipping paper. Each filter was composed of two cellulose acetate plugs separated by a 3-5 mm cavity containing a 60 mg sample of the material, the length of the cavity being adjusted to accommodate the sample snugly. The cigarettes were smoked to within 3 mm of the end of the tipping paper in a conventional cigarette smoking engine according to an ISO standard smoking regime. The levels of volatile carbonyl compounds were estimated by trapping the whole mainstream smoke for each cigarette in 2,4-dinitrophenyl hydrazine stabilised with a buffer, and analysed for aldehydic components by liquid chromatography (HPLC) with an ultra-violet detection system. Levels of vapour phase components of the smoke were estimated by passing the mainstream smoke through a 44 mm Cambridge filter pad to remove particulate material, collecting the vapour phase of the smoke in a 31 Tedlar bag, and analysing the vapour by GCMS. Levels of hydrogen cyanide (HCN) in the smoke were estimated by trapping the whole mainstream smoke in an aqueous solution of sodium hydroxide and subjecting the solution to continuous flow analysis. Each test was repeated on four samples, and the averages calculated in each case. For each sample, comparative tests were performed using a control sample, comprising a cigarette with an identical filter having an empty cavity 4 mm in length, and a sample with a cavity containing 60 mg of coconut charcoal.

(37) Table 6 summarises the analytical results for 1,3-butadiene and hydrogen cyanide (HCN). In order to compare the performances conveniently with respect to coconut charcoal, the results for each sample were normalised with respect to the results for the coconut charcoal. The normalised data is plotted in FIGS. 5 and 6, which are scatter charts plotting the percentage reduction in 1,3-butadiene and HCN, normalised with respect to the coconut charcoal of comparative example A, against the total combined volume of meso- and micropores, and the % micropore volume respectively.

(38) TABLE-US-00006 TABLE 6 % 1,3- % Butadiene % HCN 1,3-Butadiene Reduction % HCN Reduction Example Reduction v. Normalised to Reduction Normalised to No control Example A v. control Example A 2 46.09 1.70 54.09 1.05 3 57.34 2.12 58.55 1.14 4 70.68 2.61 60.77 1.18 5 64.25 2.38 78.56 1.53 6 53.74 1.99 85.22 1.66 7 73.94 2.73 79.38 1.54 8 60.03 2.22 78.83 1.53 9 71.48 2.64 84.62 1.65 10 73.94 2.73 79.38 1.54 11 57.50 2.13 73.18 1.42 12 56.58 2.09 75.01 1.46 13 50.72 1.88 64.82 1.26 A 27.05 1.00 51.41 1.00 B 20.33 0.75 36.22 0.70 C 12.88 0.48 22.24 0.43

(39) As can be seen from the data and FIGS. 5 and 6, the carbon materials of the invention, with higher total pore volumes and higher proportions of mesopore volumes than coconut charcoal performed significantly better in relation to the adsorption of HCN and, especially 1,3-butadiene, from tobacco smoke.

(40) The materials tested showed similar adsorption characteristics relative to Examples A, B and C for acreolin, propionaldehyde, crotonaldehyde, methyl-ethyl ketone and butyraldehyde.

(41) Table 7 summarises the properties of five further examples of carbon materials according to the invention in the form of microbeads (Examples 14-18), together with two comparative examples, also in the form of microbeads in the same particle size range (Examples D and E). All the beads had a particle size in the range 250-500 microns.

(42) The carbon material of Example 14 is similar to that of comparative example C, and was prepared from a resin (MAP) produced by polymerising m-amino-phenol and formaldehyde in the presence of water as pore former, but the beads were subjected to more extensive activation in carbon dioxide to achieve a higher surface area Whilst mesopores form a relatively low proportion of the combined meso- and micropore volume of this sample, its bulk density is also low, indicating that a significant pore volume lies in small macropores, not detected by nitrogen adsorption.

(43) The carbon material of example 15 was prepared by carbonising a commercially-available polymer of styrene and divinyl pyrolidone (SDP) and activating in carbon dioxide.

(44) The carbon material of Example 16 was prepared from a phenol-formaldehyde resin (PF) obtained by polymerising 100 pbw phenol and formaldehyde in the presence of 200 pbw ethylene glycol as pore former, without any additional cross-linking agent. The resulting polymer was washed, carbonised and then activated in carbon dioxide to achieve 40% burn-off.

(45) The carbon material of Example 17 was prepared from a phenol formaldehyde (PF) resin obtained in a similar manner to that of Example 16, but using 175 pbw ethylene glycol, with additional washing and in carbon dioxide to achieve a burn off of 36%.

(46) The carbon material of Example 18 was prepared from a phenol formaldehyde (PF) resin in a similar manner to that of Example 17, using 150 pbw ethylene glycol. Like the material of Example 14, mesopores form a relatively low proportion of the combined meso- and micropore volume of this sample, yet its bulk density is also low, indicating that a significant pore volume lies in small macropores, not detected by nitrogen adsorption.

(47) The carbon material of comparative Example D was prepared using a styrene vinyl pyrolidene polymer similar to that used in Example 15. The resulting material had a lower combined micro and mesopore volume and a relatively high density.

(48) The carbon material of comparative Example E was prepared using a phenol-formaldehyde resin obtained in a manner similar to that of Example 17. The resulting material had a lower proportion of mesopore volume, and a higher density.

(49) The performances of the carbon materials in reducing formaldehyde, acetaldehyde, 1,3-butadiene and HCN components of tobacco smoke was tested, using the same test procedures as described above. The results are also set forth in Table 7. The performances are evaluated in terms of the % reductions of the analytes in the tobacco smoke, normalised with respect to the corresponding reductions measured using coconut shell carbon.

(50) It can be seen that the carbon materials according to the invention perform better with respect to coconut shell carbon in the removal of at least three out of the four the smoke analytes tested, and that the comparative examples performed worse than coconut shell carbon in relation to all four of the analytes.

(51) TABLE-US-00007 TABLE 7 BET Bulk Mesopore Micropore Combined % Normalised % reductions Ex S.A. Density vol. vol pore vol mesopore 1,3- No Resin m.sup.2/g g/cc cc/g cc/g cc/g vol Formaldehyde Acetaldehyde butadiene HCN 14 MAP 1059 0.35 0.16 0.75 0.91 17.6 1.26 1.09 0.79 1.66 15 SVP 1638 0.36 0.50 0.91 1.41 35.5 0.54 1.04 2.06 1.01 16 PF 1055 0.25 0.90 0.57 1.47 61.2 2.00 1.62 2.71 1.54 17 PF 1119 0.33 0.83 0.77 1.60 51.9 1.09 1.07 1.80 0.96 18 PF 1075 0.30 0.05 0.97 1.02 4.90 1.42 1.42 2.26 1.08 D SVP 1048 0.51 0.22 0.64 0.86 25.6 0.52 0.48 0.40 0.69 E PF 1085 0.60 0.13 0.81 0.94 13.8 0.67 0.67 0.60 0.32

(52) Specific embodiments of smoking articles and smoke filters according to the invention will now be described by way of example only with reference to FIGS. 6 and 7 in which FIG. 6 is a side elevation, partly in longitudinal cross-section and partially broken away of a smoking article with a smoke filter according to the invention

(53) FIG. 7 is a similar view to FIG. 6 of a smoking article with an alternative smoke filter according to the invention.

(54) In the drawings, which are not to scale, similar features are given like reference numerals.

(55) Referring to the drawings, FIGS. 7 and 8 illustrate smoking articles in the form of cigarettes having a rod 1 of tobacco encased in a wrapper 2 attached to a smoke filter 3 by means of a tipping paper 4. For clarity, the tipping paper 4 is shown spaced from the wrapper 2, but in fact they will lie in close contact.

(56) In FIG. 7, the smoke filter 3 comprises two cylindrical filter elements 3a and 3b. The first filter element 3a at the mouth end of the filter is 15 mm in length, composed of cellulose acetate tow impregnated with 7% by weight of triacetin plasticiser having a 25 mm water gauge pressure drop over its length. The second filter element 3b, positioned adjacent the rod 1 is 12 mm in length, has a 90 mm water gauge pressure drop over its length, and comprises 80 mg cellulose acetate tow impregnated with 4% by weight of triacetin and has 30 mg of an activated porous carbon material according to the invention distributed evenly throughout its volume in a Dalmatian style.

(57) The cigarette shown in FIG. 8 is similar to that of FIG. 7 except that the smoke filter 3 has three coaxial, cylindrical filter elements 3a, 3b and 3c. The first filter element 3a at the mouth end of the cigarette is 10 mm in length, and composed of cellulose acetate tow impregnated with 7% by weight of triacetin plasticiser. The second filter element 3b, positioned adjacent the first filter element 3a is a cavity 7 mm in length containing 100 mg of an activated porous carbon material according to the invention. The third filter element 3c adjacent the second filter element 3b is 10 mm in length and comprises cellulose acetate tow impregnated with 7% by weight of triacetin. A ring of ventilation holes 5 is formed in the tipping paper 4 in a radial plane A-A which deliver air into the second filter element 3b about 3 mm downstream of the junction with the third filter element 3c when smoke is inhaled through the cigarette.

(58) In summary, the Examples provide a porous carbon material suitable for incorporation in smoke filters for cigarettes that has a BET surface area of at least 800 m.sup.2/g and a pore structure that includes mesopores and micropores. The pore volume (as measured by nitrogen adsorption) is at least 0.9 cm.sup.3/g and from 15 to 65% of the pore volume is in mesopores. The pore structure of the material provides a bulk density generally less than 0.5 g/cc. The material may be produced by carbonising and activating organic resins and may be in the form of beads for ease of handling.

(59) Various modifications and variations of the described aspects of the invention will be apparent to those skilled in the art without departing from the scope and spirit of the invention. Although the invention has been described in connection with specific preferred embodiments, it should be understood that the invention as claimed should not be unduly limited to such specific embodiments. Indeed, various modifications of the described modes of carrying out the invention which are obvious to those skilled in the relevant fields are intended to be within the scope of the following claims.