- Email: [email protected]
Thermal treatments of activated carbon fibres using a microwave furnace
Microporous and Mesoporous Materials 47 (2001) 243±252
www.elsevier.com/locate/micromeso
Thermal treatments of activated carbon ®bres using a microwave furnace P.J.M. Carrott a,*, J.M.V. Nabais a, M.M.L. Ribeiro Carrott a, J.A. Menendez b a
Departamento de Quõmica, Universidade de Evora, Col egio Luõs Ant onio Verney, Rua Rom~ ao Ramalho 59, 7000-671 Evora, Portugal b Instituto Nacional del Carb on, CSIC, Apartado 73, 33080 Oviedo, Spain Received 22 February 2001; received in revised form 25 May 2001; accepted 29 May 2001
Abstract Thermal treatment of activated carbon ®bres (ACF) in a ¯ow of N2 gas has been carried out using a microwave device operating at 2450 MHz and with a power input of 1000 W, instead of a conventional furnace, and the samples were analysed by means of low temperature N2 adsorption, elemental analysis and determination of points of zero charge. The results show that microwave treatment for periods between 5 and 30 min aects the porosity of the ACF, causing a reduction in micropore volume and micropore size. More importantly, the results also show that microwave treatment is a very eective method for modifying the surface chemistry of the ACF. During microwave treatment surface groups are completely eliminated, whereas oxygen and nitrogen atoms bonded within the pseudo-graphitic layer planes are retained. On re-exposure to air the surface groups only reform to a very limited extent and as a result very basic carbons, with points of zero charge approximately equal to 11, are readily obtained. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Activated carbon ®bres; Microwave treatment; Surface oxygen complexes; Basic carbons
1. Introduction Carbon materials with low oxygen content, basic properties, highly hydrophobic character and resistance to aging can be obtained by means of appropriate thermal treatments in dierent gaseous environments [1±8]. Generally, these thermal treatments have to be carried out at relatively high temperatures (>800°C) while ¯owing an inert or
* Corresponding author. Tel.: +351-266-745311; fax: +351266-744971. E-mail address: [email protected] (P.J.M. Carrott).
reducing gas (e.g. N2 , H2 ) over a suitable carbon precursor, during at least 1 or 2 h. The high temperatures are invariably achieved by convective and/or conductive heating of the sample, which is placed in a conventional heating system, such as a tubular furnace. However, recent work has indicated that treatment using a microwave device instead of the conventional heating systems, can also be very eective [9,10]. In view of the possible advantages associated with the use of microwave heating systems, it was therefore decided to study the eect of microwave heating in more detail using a carbon precursor (activated carbon ®bre) with a better de®ned microstructure than that used previously.
1387-1811/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 7 - 1 8 1 1 ( 0 1 ) 0 0 3 8 4 - 5
244
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
In conventional heating the heat source is located outside the carbon bed which is heated by conduction and/or convection. With microwave heating, on the other hand, the microwaves supply energy to the carbon particles themselves. Some carbons have free electrons whose displacement is restricted by grain boundaries. When these carbons are subjected to an electromagnetic ®eld, space charge polarisation takes place. Entire macroscopic regions of the material become either positive or negative synchronising their orientation with the ®eld. This mechanism is often called the Maxwell± Wagner eect [11]. At low frequency the polarisation synchronises its orientation with the ®eld, but as the frequency of the waves increases there is a phase lag between the polarisation and the applied ®eld. This leads to an absorption of energy and Joule heating of the carbon particles [11]. Microwaves are now being used in various technological and scienti®c ®elds in order to heat dielectric materials [11,12]. The main advantage of using microwave heating is that the treatment time can be considerably reduced, which in many cases represents a reduction in the energy consumption as well. Microwave energy is derived from electrical energy with a conversion eciency of approximately 50% for 2450 MHz and 85% for 915 MHz [12]. In addition, the consumption of gases used in the treatment can also be reduced. Yet, in the particular case of carbon materials, there are relatively few publications that describe the use of microwaves for producing [13,14] and regenerating [15±17] activated carbons. Surface chemistry modi®cation of active carbons by means of microwave heating was also studied in previous works [9,10]. The aim of the present work was to further explore the eect of microwave heating on the textural and surface chemistry properties of active carbon ®bres. 2. Experimental 2.1. Materials The precursors used for the production of activated carbon ®bres (ACF) were three acrylic textile ®bres provided by Fisipe (Barreiro, Portugal).
According to the manufacturer all of the ®bres had been polymerised from acrylonitrile (90 wt.%) and vinyl acetate (10 wt.%) monomers. Fibres F1 and F1N were bright 3.3 dtex (g/10,000 m) ®laments of varying length produced in 1995 (F1) and 1999 (F1N) with slightly dierent fabrication processes. Fibre F2 was also produced in 1995 but was in the form of matte 3.3 dtex ®laments of 60 mm length. This latter ®bre also contained trace quantities of a titanium dioxide optical brightener. For the production of the ACF a horizontal tubular furnace made by Termolab and with Eurotherm 904 temperature controllers and a 1 m tubular ceramic insert was used. The internal temperature of the furnace was ®rst calibrated and the length and position of the constant temperature hot zone determined. About 12 g of ®bre were placed in a 10 cm stainless steel boat with perforated ends to facilitate gas ¯ow, and this was positioned in the centre of the constant temperature zone. Stabilisation of the ®bres was carried out by heating to 300°C at a rate of 1°C min 1 under a constant N2 ¯ow of 85 cm3 min 1 and maintaining for 2 h. The ®bres were then carbonised by raising the temperature at a rate of 5°C min 1 to 800°C and maintaining at that temperature for 1 h. The carbonisation yield, in relation to the initial mass of unstabilised ®bre, was 50±52 wt.%. Activation was carried out by raising the temperature again by 15°C min 1 to 900°C and then switching to a CO2 ¯ow of 85 cm3 min 1 , maintaining for the appropriate time in order to obtain burn-os within the range of 10±90 wt.% (indicated in the sample designations), switching back to the N2 ¯ow and allowing to cool to below 50°C before removing the ACF from the furnace and storing in a sealed sample ¯ask. The conditions used and the corresponding sample designations are indicated in Table 1. For the preliminary study we also used three samples of commercial activated carbons from NORIT, namely Norit AZO (NAZO), Norit S51 (NS51) and Norit SX (NSXP). 2.2. Microwave treatments Sub-samples weighing 2:00 0:01 g and sized to less than 1 mm of ACF were placed in a quartz
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
245
Table 1 Conditions of activation and elemental analysis of ACF before and after the microwave treatments and yields of microwave treatments Sample Fibre F1 at 900°C F1 F1-0 F1-0m5 F1-0m15 F1-0m30 F1-37 F1-37m15 F1-53 F1-53m15 F1-76 F1-76m15 Fibre F1N at 900°C F1N F1N-0 F1N-0m5 F1N-0m15 F1N-0m30 F1N-12 F1N-12m15 F1N-36 F1N-36m15 F1N-61 F1N-61m15 Fibre F2 at 900°C F2 F2-0 F2-0m5 F2-0m15 F2-0m30 F2-26 F2-26m15 F2-48 F2-48m15 F2-74 F2-74m15
Activation time (h)
Burn-o (%)
o.f. 0
0
2
37
5
53
9
76
o.f. 0
0
1
12
2
36
7
61
o.f. 0
0
2
26
5
48
8
74
N (%)
C (%)
H (%)
O (%)
26.32 16.47 9.87 7.82 6.91 8.44 7.37 5.61 4.49 4.17 3.44
64.21 74.94 84.18 88.78 83.09 80.95 87.61 82.18 89.50 80.04 90.05
5.91 1.01 0.28 0.43 0.43 0.60 0.32 0.47 0.15 0.30 0.18
3.55 9.63 3.96 1.71 3.84 11.54 2.21 12.97 7.67 7.72 6.88
22.92 12.62 11.52 10.14 9.10 6.90 5.38 5.71 6.83 4.84 3.94
64.65 79.25 79.40 84.59 84.62 77.75 87.04 84.13 84.95 83.13 86.86
5.47 0.97 0.52 0.56 0.47 0.87 0.29 0.42 0.41 0.51 0.33
4.31 7.45 6.01 3.03 3.44 13.18 3.79 11.42 5.64 8.70 4.11
23.36 11.78 10.06 7.59 7.75 5.31 6.22 3.97 4.67 3.95 4.52
64.57 77.90 84.36 88.88 87.39 81.99 85.41 82.31 86.72 76.09 88.48
5.77 0.83 0.63 0.45 0.36 0.34 0.39 0.35 0.34 0.25 0.21
4.26 6.93 2.40 1.49 3.06 10.48 4.81 9.24 3.49 11.15 1.59
Yield (%)
85 87 89
81 81 82
82 82 86
(o:f: original acrylic ®bre; -0 carbonised but non-activated ®bre; -X mY is ®bre with burn-o X% after Y min of microwave treatment).
reactor, which in turn was placed inside a multimode resonant microwave cavity. Microwave treatments consisted of subjecting the samples to microwave action for 15 min for the ACF samples, NS51 and NSXP, 5, 15 or 30 min for the carbonised ®bres and 15 or 30 min for NAZO. The time of treatment is indicated in the sample designation (XXXm15, means 15 min of microwave treatment). The yields of the microwave treatments, which varied between 81% and 89%, are
indicated in Table 1. An inert atmosphere was maintained during treatment and cool-down intervals by passing an N2 ¯ow of 100 cm3 min 1 through the sample bed. The input power of the microwave equipment was set at 1000 W, and the microwave frequency used was 2450 MHz. As we operated at a constant input power the amount of energy absorbed by each sample, and in consequence its temperature during the treatment, were dierent depending on the nature of the sample.
246
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
The temperature of the carbon bed during microwave treatment were measured using an infrared optical pyrometer. Details of the microwave device as well as temperature measurements are given elsewhere [9,10]. 2.3. Physical and chemical characterisation of the carbon ®bres A preliminary study (the results of which are not presented here) con®rmed the results of previous work [9] which had shown that the surface properties of microwave-treated carbons gradually alter over a period of up to about 10±14 days after being removed from the microwave furnace, presumably due to reoxidation and readsorption of residual CO2 , which also occurs after conventional heating. Hence, in order that the characterisation information on the carbon samples was obtained with the surfaces in a stable condition, the physical and chemical characterisation of each sample was only carried out four weeks after its microwave treatment. Nitrogen adsorption isotherms at 77 K were determined using a CE Instruments Sorptomatic 1990 after outgassing the samples at 380°C to a residual vacuum of 5 10 6 mbar. Elemental analysis of carbon, hydrogen, sulphur, nitrogen and oxygen was carried out using a Eurovector EuroEA elemental analyser. No evidence for the presence of sulphur was found for the ACF. For the Norit samples sulphur is present in small quantities, which are negligible. The quantities of the other elements are included in Table 1 for the ACF and Table 2 for the Norit samples. The point of zero charge (pzc) of each sample was estimated from the pH of a concentrated Table 2 Elemental analysis of the Norit samples before and after the microwave treatments (mY sample after Y min of microwave treatment) Sample
N (%)
C (%)
H (%)
O (%)
NAZO NAZOm15 NAZOm30 NS51 NS51m15 NSXP NSXPm15
0.12 0.40 0.38 0.10 0.22 0.39 0.28
82.34 90.20 88.52 78.71 80.20 79.89 93.11
0.47 0.18 0.18 0.63 0.24 0.39 0.22
5.45 1.19 1.45 8.68 0.38 8.68 0.56
dispersion. Approximately 0.3 g of the ACF were added to 5 cm3 of a solution 0.1 M in NaNO3 (p.a. grade, >99:5% pure). The suspension was placed in a thermostat bath with stirring (GRANT model SS40-D) for 48 h, then ®ltered with Whatman #1 paper and the equilibrium pH measured with a CRISON microelectrode and model 2002 potentiometer. 3. Results and discussion 3.1. Microwave heating Fig. 1 shows the evolution of the temperature of the carbon bed during the dierent microwave treatments of the ACF. In all cases heating of the samples was quite fast taking, in general, less than a minute to reach a temperature above 800°C. However, the temperature was not constant during the 15 min of the treatment. At the beginning there was a rapid increase in temperature until it reached a maximum around two min after commencement of the treatment. The maximum temperature was in most cases between 900°C and 1000°C, that is, higher than the temperatures used during carbonisation and activation of the ®bres. The exceptions are the two lowest burn-o ACF F1N-12 and F2-26 which only reached temperatures slightly above (F1N-12) or below (F2-26) 800°C. There was then a decrease in temperature which, depending on the sample, was as much as 250°C (for sample F1N-61) or about 75°C (for samples F1-76 or F2-74). At the end of the 15 min of treatment the temperature was in each case approximately stable. From Table 1 it can be seen that the yield of the microwave treatment also varied from sample to sample, over the range 81±89%. The results indicate that the yield increases slightly with increasing burn-o and that the samples giving the highest yields (the three F1 samples and F2-74) are those which reached and maintained the highest temperatures during the treatment. During microwave heating the energy supplied by the microwave furnace is deposited directly in the sample and the temperature reached will therefore depend on the fraction of the supplied
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
247
the degree of microwave absorption. However, it seems that for a given series the higher the degree of activation the higher the temperature reached by the sample and the higher the yield. It is also interesting to note that, with the samples prepared from ®bre F1, higher maximum temperatures were reached during the initial heating period, and these temperatures were more eectively maintained throughout the treatment and, in addition, the yields were higher. Previous work has shown that the microstructure of ®bre F1 is slightly dierent to that of ®bres F1N and F2 [18]. 3.2. Eect of microwave heating on the characteristics of the carbon ®bres
Fig. 1. Temperature±time pro®les during microwave treatment for series (a) F1, (b) F1N and (c) F2.
energy which is actually adsorbed by the sample. The ability of a carbon material to absorb microwave energy depends on several factors such as particle size, amount of carbon, moisture content and nature of the carbon (degree of order of the basal planes, chemical composition, etc.) [11,12]. In the present experiments the amount of carbon, the moisture content and the particle size were all about the same for all samples. Hence, the dierences in the temperatures reached by each sample and its evolution during the treatment, as well as the dierences in yield, may be a consequence of the dierent microstructures of the carbons. With the present data it is dicult to establish any ®rm correlation between the nature of the sample and
3.2.1. Texture Fig. 2 shows the nitrogen adsorption isotherms determined on one of the Norit samples before and after 15 min of microwave treatment. It can be seen that the surface area and porosity of the sample were not signi®cantly changed by the treatment. Similar results were obtained here with the other two Norit samples and in previous work with other non-ACF samples [9]. It is noteworthy that the Norit samples and the samples used in the previous work [9] were mesoporous, whereas the ACF samples used in this work are completely microporous. The nitrogen isotherms determined on the ACF obtained from ®bre F1N before and after microwave treatment are shown in Fig. 3. Similar results were obtained with the other ®bres. It was found that, with one exception, the limiting uptake of N2 at high relative pressure was signi®cantly reduced after treatment. It was also evident that in most cases the knee of the isotherm became less rounded after treatment, suggesting a decrease in the mean pore size. The only exception to this behaviour was the sample F2-26, for which the nitrogen isotherm was not altered. It should be noted that this was the only sample which did not reach the carbonisation temperature of 800°C during the microwave treatment. The nitrogen isotherms were analysed by means of the as and DR methods, and representative plots for the samples before microwave treatment can be found in Ref. [18]. After microwave treatment
248
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
Fig. 2. Low-temperature nitrogen adsorption/desorption isotherms determined on NS51 before and after the microwave treatment. (light symbols ± adsorption; heavy symbols ± desorption).
the plots obtained were similar. From the slope and intercept of the multilayer region of each as plot the values of the external surface area, Sext , and the total micropore volume, Vs , given in Table 3, were calculated. From the slope and intercept of each DR plot the values of the characteristic energy, E0 , and the micropore volume, V0 , also given in Table 3, were calculated. Estimates of the mean pore width, L0 , were calculated from the relationship [19]: L0 10:8= E0
11:4
1
These values are also included in the ®nal column of Table 3. Comparison of the corresponding values of Vs and V0 shows that all samples with a burn-o of less than 40% contained only narrow (primary) micropores both before and after microwave treatment, whereas the samples with a burn-o greater than 40% contained both primary micropores and
secondary micropores. With these higher burn-o samples the dierence between Vs and V0 was somewhat reduced after microwave treatment, especially in the case of sample F1N-61m15. In all cases, except F2-26, as previously noted, the results con®rm that microwave treatment brought about a signi®cant reduction in pore volume. The values of the mean pore width, L0 , were also reduced in most, although not all cases. Much larger dierences were found with the samples that maintained the highest temperatures during the microwave treatment, namely the three F1 samples and F2-74. With these four samples the mean pore size was reduced by between 0.18 and 0.35 nm. With the other samples the variation was always less than 0.1 nm. 3.2.2. Elemental composition The results of elemental analysis given in Tables 1 and 2 show that microwave treatment for periods of up to 15 min results in an increase in the
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
249
Fig. 3. Low-temperature nitrogen adsorption/desorption isotherms determined on F1N before and after the microwave treatment (light symbols ± adsorption; heavy symbols ± desorption).
carbon content for all of the carbonised and activated ®bres as well as for the Norit samples. This indicates an overall reduction in the total quantity of heteroatoms present in the samples. In particular, it should be noted that the oxygen content is always reduced. The reduction is particularly signi®cant with the Norit samples where the oxygen content after microwave treatment is <1:5% for all samples, compared with values between 1.5% and 7.7% for the ®bres. The higher values obtained with the ®bres may indicate that at least part of the oxygen is very strongly bound, possibly within the pseudo-graphitic layer planes. It would seem likely that this oxygen became included in the structure during polymerisation of the vinyl acetate comonomer. With regard to the nitrogen content, the Norit samples contain only low amounts, <0:4%, both before and after the microwave treatment. The carbonised ®bres initially contain relatively high amounts of nitrogen, and although this is signi®cantly reduced by the microwave treatment the
®nal nitrogen contents of about 7±10% are still quite high. Treatment of the activated ®bres results in a small reduction in the nitrogen content for ®ve of the ®bres (the three F1 ®bres and two of the F1N ®bres) but a small increase for the other four ®bres (one of the F1N ®bres and the three F2 ®bres). The ®nal values of between 3.4% and 7.4% are also relatively high. These results suggest that the carbonised ®bres contain nitrogen in dierent forms. Part of the nitrogen is in the form of surface groups which are readily removed by microwave treatment or by activation. The other part, as was also observed in the case of oxygen, is strongly bound and much more resistant to removal either by microwave treatment or by increasing activation. Presumably, this nitrogen is also contained within the pseudo-graphitic layer planes where it was included during the cyclisation of the nitrile groups of the polymer precursor during carbonisation. The carbonised ®bres and Norit AZO were also treated for the longer period of 30 min. It can be
250
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
Table 3 Textural characteristics of the original ACE and treated ®bres (sample designations as in Table 1) Sample
SBET (m2 g 1 )
as method
DR method
Vs (cm g )
Sext (m g )
V0 (cm3 g 1 )
E0 (kJ mol 1 )
L0 (nm)
3
1
2
1
F1-37 F1-37m15 F1-53 F1-53m15 F1-76 F1-76m15
390 219 1239 768 1341 774
0.18 0.09 0.54 0.36 0.60 0.36
6 1 8 2 6 2
0.19 0.09 0.46 0.30 0.49 0.30
26.1 30.7 20.8 23.8 19.9 23.0
0.74 0.56 1.15 0.87 1.28 0.93
F1N-12 F1N-12m15 F1N-36 F1N-36m15 F1N-61 F1N-61m15
370 163 791 318 1284 730
0.17 0.08 0.35 0.15 0.60 0.32
4 4 6 3 1 4
0.16 0.08 0.34 0.15 0.41 0.30
22.1 17.8 22.9 20.6 20.4 20.7
1.01 1.10 0.94 0.95 1.20 1.16
F2-26 F2-26m15 F2-48 F2-48m15 F2-74 F2-74m15
433 453 1014 655 1397 772
0.19 0.20 0.45 0.29 0.62 0.36
11 1 1 1 15 4
0.19 0.19 0.39 0.26 0.50 0.31
24.1 23.6 20.7 21.2 20.0 23.2
0.85 0.88 1.16 1.11 1.26 0.92
SBET is BET surface area; Vs and Sext are total pore volume and external surface area from as plot; V0 , E0 and L0 are micropore volume, characteristic energy and mean micropore width from DR plot.
seen from Tables 1 and 2 that, after this prolonged microwave treatment, there was a small decrease in the carbon content, when compared with the 15 min treatment, and an increase in the oxygen content. It should be remembered that the elemental analyses were carried out about four weeks after treating the samples. Although the samples were stored in sealed containers during this time they will have adsorbed some oxygen and carbon dioxide from the atmosphere. It is likely that the residual oxygen content of the Norit samples treated for 15 min and a small part (but not all) of the oxygen content of the ®bres after 15 min treatment are due to this reoxidation. The fact that after 30 min of microwave treatment the oxygen content is higher than that found after 15 min suggests that the longer treatment period resulted in a more reactive surface which was subsequently reoxidised to a greater degree. 3.2.3. Point of zero charge Fig. 4 shows the variation in pzc for the carbonised ®bres and the Norit samples as a function of the time of microwave treatment. The initial
values for the untreated Norit samples NAZO and NSXP are dierent from values previously published [20] indicating that in the intervening ®ve years the nature of the surfaces changed, presumably due to adsorption of atmospheric CO2 in the form of surface groups (unfortunately, elemental analysis was not available to us when the previous work was published). It can be seen from Fig. 4 that for up to 15 min treatment time the pzc increases but that, in the case of the carbonised ®bres, after 30 min the pzc is lower than the corresponding value obtained after 15 min treatment. These variations are consistent with the observed variations in the residual oxygen content if, as already suggested, the residual oxygen is due to reoxidation of the samples after the microwave treatment. In all cases, including the samples which were initially slightly acid, the samples are basic after treatment. The very high pzc value of 11.3 given by NAZO after 15 min treatment, which is similar to the value of 11.8 previously reported, is particularly noteworthy. The pzc values of the ACF are indicated in Fig. 5. It can be seen that in all cases the pzc increases
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
Fig. 4. Evolution of the pzc of the carbonised ®bres and Norit samples with the time of microwave treatments.
after microwave treatment reaching, in all but one case, the almost constant value of 10:8 0:2 independently of the initial pzc of the untreated samples. In order to verify that the change in pzc was not due solely to the temperature reached by the samples during the microwave treatment, the ®bres were also heated for 2 h in a ¯ow of N2 in a conventional furnace and then stored for four weeks under the same conditions used for the microwave treated samples. It was found that the pzc was very little aected by this conventional treat-
251
ment. It should also be pointed out that the fact that the same value of pzc is obtained for almost all of the microwave treated samples cannot be associated with a buering eect of the water, as we have already seen with NAZO that higher pzc values are possible. In a previous paper [20] we have presented a detailed consideration of the principal factors which in¯uence the surface ionisation of carbon type materials containing acid groups and basic groups and clari®ed the eect of changing site concentrations and ionisation constants on the value of the pzc and iep (isoelectric point). Considering the case of a basic carbon containing one type of basic site with ionisation constant pKb and one type of acid site with ionisation constant pKa , then the pzc will not depend on the absolute concentration of the basic sites, but will depend on the value of pKb . Over the range of pH values of approximately pKb 2 the basic sites turn from completely ionised to completely unionised and the pzc will occur between these limits. Over the same range of pH values the acid sites will, in general, be completely ionised and the pzc will not therefore be dependent on the value of pKa . On the other hand, the exact value of the pzc will be dependent on the concentration of the acid sites and will decrease with increasing acid site concentration. On the basis of these observations a plausible explanation for the results given in Figs. 4 and 5 is the following. During microwave treatment most, if not all, of the non-ring surface functional groups, which
Fig. 5. Pzc of the ACF samples before (®lled bars) and after (empty bars) the microwave treatments.
252
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
would appear to be mainly acid in nature, are removed. Most of the oxygen and nitrogen remaining is included in basic ring groups, and the pzc therefore increases. Furthermore, although the absolute concentrations of the basic groups may vary from one sample to another, their chemical nature is probably the same and hence the pzc has the same value in each case. During storage of the treated samples, some reformation of acid surface groups may occur but to a lesser extent than before treatment, and the pzc therefore remains higher than the initial value. Furthermore, as the pzc remains constant the extent of reformation of acid groups must also be similar in all cases. Thirty-minute treatment of the samples causes more profound changes than 15-min treatment, resulting in a more reactive surface on which reformation of acid surface groups can occur to a greater extent. 4. Conclusions The results presented show that microwave heating is a very eective means of modifying the porosity and, of special importance, the surface chemistry of ACF. Comparatively short heating periods are sucient to completely remove the surface groups, but without removing oxygen and nitrogen atoms bonded within the pseudo-graphitic planes. As the surface groups are mainly acidic in nature, whereas ring oxygen and nitrogen atoms impart basic functionality, the resulting carbon materials are very basic, giving in all cases pzc values close to 11. After microwave treatment the surfaces of the carbon materials are left in a reactive state. In the present work no direct measures (apart from storing under an inert atmosphere) were taken to block the surface reactivity and, as a result, some reformation of surface groups must have occurred slowly during storage. In future work, it would be interesting to study the reaction of the microwavetreated surface, while still in the reactive state, with reducing or other gases, with a view to enhancing the basic character or of anchoring speci®c types of functionality onto the surface. In the former case, the resulting materials would be valuable
adsorbents, for example, for the adsorption of acid gases such as SO2 . In the second case, there are evident opportunities for the development of new catalysts. Acknowledgements The authors are grateful to FISIPE ± Fibras Sinteticas de Portugal for the provision of samples and to the Fundacß~ao para a Ci^encia e a Tecnologia (Portugal), the European Regional Development Fund (FEDER) and the program PRAXIS XXI for ®nancial support (project PRAXIS/3/3.1/ MMA/1781/95). References [1] J.A. Menendez, J. Phillips, B. Xia, L.R. Radovic, Langmuir 12 (1996) 4404. [2] J.A. Menendez, B. Xia, J. Phillips, L.R. Radovic, Langmuir 13 (1997) 3414. [3] J.A. Menendez, L.R. Radovic, B. Xia, J. Phillips, J. Phys. Chem. 100 (1996) 17243. [4] C.A. Le on y Le on, J.M. Solar, V. Calemma, L.R. Radovic, Carbon 30 (1992) 797. [5] H.P. Boehm, M. Voll, Carbon 8 (1970) 227. [6] S.K. Verma, P.L. Walker Jr., Carbon 30 (1992) 837. [7] R.C. Bansal, F.J. Vastola, P.L. Walker Jr., Carbon 12 (1974) 355. [8] T. Suzuki, C. Ishii, N. Shindo, K. Kaneko, J. Porous Mater. 4 (1997) 181. [9] J.A. Menendez, E.M. Menendez, M.J. Iglesias, A. Garcõa, J.J. Pis, Carbon 37 (1999) 1115. [10] J.A. Menendez, E.M. Menendez, A. Garcõa, J.B. Parra, J.J. Pis, J. Microwave Power EE 34 (1999) 137. [11] A. Zlotorzynski, Critic. Rev. Analytic. Chem. 25 (1) (1995) 43. [12] K.E. Haque, Int. J. Miner. Process 57 (1999) 1. [13] K.M. Holland Patent GB 9105735 (May 2, 1991). [14] L.M. Norman, C.Y. Cha, Chem. Eng. Commun. 140 (1996) 87. [15] K.E. Haque, P.D. Kondos, R.J.C. MacDonal, D.E. Laforest, Patent CA 2008242 (July 22, 1991). [16] K.M. Holland, Patent GB 9426661 (November 24, 1994). [17] Y. Kong, C.Y. Cha, Energy Fuels 6 (1996) 1245. [18] P.J.M. Carrott, J.M.V. Nabais, M.M.L. Ribeiro Carrott, J.A. Pajares, Carbon 39 (10) (2001) 1543. [19] H.F. Stoeckli, P. Rebstein, L. Ballerini, Carbon 28 (6) (1990) 907. [20] P.J.M. Carrott, M.M.L. Ribeiro Carrott, A.J. Est^ev~ao Candeias, J.P. Prates Ramalho, J. Chem. Soc. Faraday Trans. 91 (4) (1995) 2179.
www.elsevier.com/locate/micromeso
Thermal treatments of activated carbon ®bres using a microwave furnace P.J.M. Carrott a,*, J.M.V. Nabais a, M.M.L. Ribeiro Carrott a, J.A. Menendez b a
Departamento de Quõmica, Universidade de Evora, Col egio Luõs Ant onio Verney, Rua Rom~ ao Ramalho 59, 7000-671 Evora, Portugal b Instituto Nacional del Carb on, CSIC, Apartado 73, 33080 Oviedo, Spain Received 22 February 2001; received in revised form 25 May 2001; accepted 29 May 2001
Abstract Thermal treatment of activated carbon ®bres (ACF) in a ¯ow of N2 gas has been carried out using a microwave device operating at 2450 MHz and with a power input of 1000 W, instead of a conventional furnace, and the samples were analysed by means of low temperature N2 adsorption, elemental analysis and determination of points of zero charge. The results show that microwave treatment for periods between 5 and 30 min aects the porosity of the ACF, causing a reduction in micropore volume and micropore size. More importantly, the results also show that microwave treatment is a very eective method for modifying the surface chemistry of the ACF. During microwave treatment surface groups are completely eliminated, whereas oxygen and nitrogen atoms bonded within the pseudo-graphitic layer planes are retained. On re-exposure to air the surface groups only reform to a very limited extent and as a result very basic carbons, with points of zero charge approximately equal to 11, are readily obtained. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: Activated carbon ®bres; Microwave treatment; Surface oxygen complexes; Basic carbons
1. Introduction Carbon materials with low oxygen content, basic properties, highly hydrophobic character and resistance to aging can be obtained by means of appropriate thermal treatments in dierent gaseous environments [1±8]. Generally, these thermal treatments have to be carried out at relatively high temperatures (>800°C) while ¯owing an inert or
* Corresponding author. Tel.: +351-266-745311; fax: +351266-744971. E-mail address: [email protected] (P.J.M. Carrott).
reducing gas (e.g. N2 , H2 ) over a suitable carbon precursor, during at least 1 or 2 h. The high temperatures are invariably achieved by convective and/or conductive heating of the sample, which is placed in a conventional heating system, such as a tubular furnace. However, recent work has indicated that treatment using a microwave device instead of the conventional heating systems, can also be very eective [9,10]. In view of the possible advantages associated with the use of microwave heating systems, it was therefore decided to study the eect of microwave heating in more detail using a carbon precursor (activated carbon ®bre) with a better de®ned microstructure than that used previously.
1387-1811/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 7 - 1 8 1 1 ( 0 1 ) 0 0 3 8 4 - 5
244
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
In conventional heating the heat source is located outside the carbon bed which is heated by conduction and/or convection. With microwave heating, on the other hand, the microwaves supply energy to the carbon particles themselves. Some carbons have free electrons whose displacement is restricted by grain boundaries. When these carbons are subjected to an electromagnetic ®eld, space charge polarisation takes place. Entire macroscopic regions of the material become either positive or negative synchronising their orientation with the ®eld. This mechanism is often called the Maxwell± Wagner eect [11]. At low frequency the polarisation synchronises its orientation with the ®eld, but as the frequency of the waves increases there is a phase lag between the polarisation and the applied ®eld. This leads to an absorption of energy and Joule heating of the carbon particles [11]. Microwaves are now being used in various technological and scienti®c ®elds in order to heat dielectric materials [11,12]. The main advantage of using microwave heating is that the treatment time can be considerably reduced, which in many cases represents a reduction in the energy consumption as well. Microwave energy is derived from electrical energy with a conversion eciency of approximately 50% for 2450 MHz and 85% for 915 MHz [12]. In addition, the consumption of gases used in the treatment can also be reduced. Yet, in the particular case of carbon materials, there are relatively few publications that describe the use of microwaves for producing [13,14] and regenerating [15±17] activated carbons. Surface chemistry modi®cation of active carbons by means of microwave heating was also studied in previous works [9,10]. The aim of the present work was to further explore the eect of microwave heating on the textural and surface chemistry properties of active carbon ®bres. 2. Experimental 2.1. Materials The precursors used for the production of activated carbon ®bres (ACF) were three acrylic textile ®bres provided by Fisipe (Barreiro, Portugal).
According to the manufacturer all of the ®bres had been polymerised from acrylonitrile (90 wt.%) and vinyl acetate (10 wt.%) monomers. Fibres F1 and F1N were bright 3.3 dtex (g/10,000 m) ®laments of varying length produced in 1995 (F1) and 1999 (F1N) with slightly dierent fabrication processes. Fibre F2 was also produced in 1995 but was in the form of matte 3.3 dtex ®laments of 60 mm length. This latter ®bre also contained trace quantities of a titanium dioxide optical brightener. For the production of the ACF a horizontal tubular furnace made by Termolab and with Eurotherm 904 temperature controllers and a 1 m tubular ceramic insert was used. The internal temperature of the furnace was ®rst calibrated and the length and position of the constant temperature hot zone determined. About 12 g of ®bre were placed in a 10 cm stainless steel boat with perforated ends to facilitate gas ¯ow, and this was positioned in the centre of the constant temperature zone. Stabilisation of the ®bres was carried out by heating to 300°C at a rate of 1°C min 1 under a constant N2 ¯ow of 85 cm3 min 1 and maintaining for 2 h. The ®bres were then carbonised by raising the temperature at a rate of 5°C min 1 to 800°C and maintaining at that temperature for 1 h. The carbonisation yield, in relation to the initial mass of unstabilised ®bre, was 50±52 wt.%. Activation was carried out by raising the temperature again by 15°C min 1 to 900°C and then switching to a CO2 ¯ow of 85 cm3 min 1 , maintaining for the appropriate time in order to obtain burn-os within the range of 10±90 wt.% (indicated in the sample designations), switching back to the N2 ¯ow and allowing to cool to below 50°C before removing the ACF from the furnace and storing in a sealed sample ¯ask. The conditions used and the corresponding sample designations are indicated in Table 1. For the preliminary study we also used three samples of commercial activated carbons from NORIT, namely Norit AZO (NAZO), Norit S51 (NS51) and Norit SX (NSXP). 2.2. Microwave treatments Sub-samples weighing 2:00 0:01 g and sized to less than 1 mm of ACF were placed in a quartz
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
245
Table 1 Conditions of activation and elemental analysis of ACF before and after the microwave treatments and yields of microwave treatments Sample Fibre F1 at 900°C F1 F1-0 F1-0m5 F1-0m15 F1-0m30 F1-37 F1-37m15 F1-53 F1-53m15 F1-76 F1-76m15 Fibre F1N at 900°C F1N F1N-0 F1N-0m5 F1N-0m15 F1N-0m30 F1N-12 F1N-12m15 F1N-36 F1N-36m15 F1N-61 F1N-61m15 Fibre F2 at 900°C F2 F2-0 F2-0m5 F2-0m15 F2-0m30 F2-26 F2-26m15 F2-48 F2-48m15 F2-74 F2-74m15
Activation time (h)
Burn-o (%)
o.f. 0
0
2
37
5
53
9
76
o.f. 0
0
1
12
2
36
7
61
o.f. 0
0
2
26
5
48
8
74
N (%)
C (%)
H (%)
O (%)
26.32 16.47 9.87 7.82 6.91 8.44 7.37 5.61 4.49 4.17 3.44
64.21 74.94 84.18 88.78 83.09 80.95 87.61 82.18 89.50 80.04 90.05
5.91 1.01 0.28 0.43 0.43 0.60 0.32 0.47 0.15 0.30 0.18
3.55 9.63 3.96 1.71 3.84 11.54 2.21 12.97 7.67 7.72 6.88
22.92 12.62 11.52 10.14 9.10 6.90 5.38 5.71 6.83 4.84 3.94
64.65 79.25 79.40 84.59 84.62 77.75 87.04 84.13 84.95 83.13 86.86
5.47 0.97 0.52 0.56 0.47 0.87 0.29 0.42 0.41 0.51 0.33
4.31 7.45 6.01 3.03 3.44 13.18 3.79 11.42 5.64 8.70 4.11
23.36 11.78 10.06 7.59 7.75 5.31 6.22 3.97 4.67 3.95 4.52
64.57 77.90 84.36 88.88 87.39 81.99 85.41 82.31 86.72 76.09 88.48
5.77 0.83 0.63 0.45 0.36 0.34 0.39 0.35 0.34 0.25 0.21
4.26 6.93 2.40 1.49 3.06 10.48 4.81 9.24 3.49 11.15 1.59
Yield (%)
85 87 89
81 81 82
82 82 86
(o:f: original acrylic ®bre; -0 carbonised but non-activated ®bre; -X mY is ®bre with burn-o X% after Y min of microwave treatment).
reactor, which in turn was placed inside a multimode resonant microwave cavity. Microwave treatments consisted of subjecting the samples to microwave action for 15 min for the ACF samples, NS51 and NSXP, 5, 15 or 30 min for the carbonised ®bres and 15 or 30 min for NAZO. The time of treatment is indicated in the sample designation (XXXm15, means 15 min of microwave treatment). The yields of the microwave treatments, which varied between 81% and 89%, are
indicated in Table 1. An inert atmosphere was maintained during treatment and cool-down intervals by passing an N2 ¯ow of 100 cm3 min 1 through the sample bed. The input power of the microwave equipment was set at 1000 W, and the microwave frequency used was 2450 MHz. As we operated at a constant input power the amount of energy absorbed by each sample, and in consequence its temperature during the treatment, were dierent depending on the nature of the sample.
246
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
The temperature of the carbon bed during microwave treatment were measured using an infrared optical pyrometer. Details of the microwave device as well as temperature measurements are given elsewhere [9,10]. 2.3. Physical and chemical characterisation of the carbon ®bres A preliminary study (the results of which are not presented here) con®rmed the results of previous work [9] which had shown that the surface properties of microwave-treated carbons gradually alter over a period of up to about 10±14 days after being removed from the microwave furnace, presumably due to reoxidation and readsorption of residual CO2 , which also occurs after conventional heating. Hence, in order that the characterisation information on the carbon samples was obtained with the surfaces in a stable condition, the physical and chemical characterisation of each sample was only carried out four weeks after its microwave treatment. Nitrogen adsorption isotherms at 77 K were determined using a CE Instruments Sorptomatic 1990 after outgassing the samples at 380°C to a residual vacuum of 5 10 6 mbar. Elemental analysis of carbon, hydrogen, sulphur, nitrogen and oxygen was carried out using a Eurovector EuroEA elemental analyser. No evidence for the presence of sulphur was found for the ACF. For the Norit samples sulphur is present in small quantities, which are negligible. The quantities of the other elements are included in Table 1 for the ACF and Table 2 for the Norit samples. The point of zero charge (pzc) of each sample was estimated from the pH of a concentrated Table 2 Elemental analysis of the Norit samples before and after the microwave treatments (mY sample after Y min of microwave treatment) Sample
N (%)
C (%)
H (%)
O (%)
NAZO NAZOm15 NAZOm30 NS51 NS51m15 NSXP NSXPm15
0.12 0.40 0.38 0.10 0.22 0.39 0.28
82.34 90.20 88.52 78.71 80.20 79.89 93.11
0.47 0.18 0.18 0.63 0.24 0.39 0.22
5.45 1.19 1.45 8.68 0.38 8.68 0.56
dispersion. Approximately 0.3 g of the ACF were added to 5 cm3 of a solution 0.1 M in NaNO3 (p.a. grade, >99:5% pure). The suspension was placed in a thermostat bath with stirring (GRANT model SS40-D) for 48 h, then ®ltered with Whatman #1 paper and the equilibrium pH measured with a CRISON microelectrode and model 2002 potentiometer. 3. Results and discussion 3.1. Microwave heating Fig. 1 shows the evolution of the temperature of the carbon bed during the dierent microwave treatments of the ACF. In all cases heating of the samples was quite fast taking, in general, less than a minute to reach a temperature above 800°C. However, the temperature was not constant during the 15 min of the treatment. At the beginning there was a rapid increase in temperature until it reached a maximum around two min after commencement of the treatment. The maximum temperature was in most cases between 900°C and 1000°C, that is, higher than the temperatures used during carbonisation and activation of the ®bres. The exceptions are the two lowest burn-o ACF F1N-12 and F2-26 which only reached temperatures slightly above (F1N-12) or below (F2-26) 800°C. There was then a decrease in temperature which, depending on the sample, was as much as 250°C (for sample F1N-61) or about 75°C (for samples F1-76 or F2-74). At the end of the 15 min of treatment the temperature was in each case approximately stable. From Table 1 it can be seen that the yield of the microwave treatment also varied from sample to sample, over the range 81±89%. The results indicate that the yield increases slightly with increasing burn-o and that the samples giving the highest yields (the three F1 samples and F2-74) are those which reached and maintained the highest temperatures during the treatment. During microwave heating the energy supplied by the microwave furnace is deposited directly in the sample and the temperature reached will therefore depend on the fraction of the supplied
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
247
the degree of microwave absorption. However, it seems that for a given series the higher the degree of activation the higher the temperature reached by the sample and the higher the yield. It is also interesting to note that, with the samples prepared from ®bre F1, higher maximum temperatures were reached during the initial heating period, and these temperatures were more eectively maintained throughout the treatment and, in addition, the yields were higher. Previous work has shown that the microstructure of ®bre F1 is slightly dierent to that of ®bres F1N and F2 [18]. 3.2. Eect of microwave heating on the characteristics of the carbon ®bres
Fig. 1. Temperature±time pro®les during microwave treatment for series (a) F1, (b) F1N and (c) F2.
energy which is actually adsorbed by the sample. The ability of a carbon material to absorb microwave energy depends on several factors such as particle size, amount of carbon, moisture content and nature of the carbon (degree of order of the basal planes, chemical composition, etc.) [11,12]. In the present experiments the amount of carbon, the moisture content and the particle size were all about the same for all samples. Hence, the dierences in the temperatures reached by each sample and its evolution during the treatment, as well as the dierences in yield, may be a consequence of the dierent microstructures of the carbons. With the present data it is dicult to establish any ®rm correlation between the nature of the sample and
3.2.1. Texture Fig. 2 shows the nitrogen adsorption isotherms determined on one of the Norit samples before and after 15 min of microwave treatment. It can be seen that the surface area and porosity of the sample were not signi®cantly changed by the treatment. Similar results were obtained here with the other two Norit samples and in previous work with other non-ACF samples [9]. It is noteworthy that the Norit samples and the samples used in the previous work [9] were mesoporous, whereas the ACF samples used in this work are completely microporous. The nitrogen isotherms determined on the ACF obtained from ®bre F1N before and after microwave treatment are shown in Fig. 3. Similar results were obtained with the other ®bres. It was found that, with one exception, the limiting uptake of N2 at high relative pressure was signi®cantly reduced after treatment. It was also evident that in most cases the knee of the isotherm became less rounded after treatment, suggesting a decrease in the mean pore size. The only exception to this behaviour was the sample F2-26, for which the nitrogen isotherm was not altered. It should be noted that this was the only sample which did not reach the carbonisation temperature of 800°C during the microwave treatment. The nitrogen isotherms were analysed by means of the as and DR methods, and representative plots for the samples before microwave treatment can be found in Ref. [18]. After microwave treatment
248
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
Fig. 2. Low-temperature nitrogen adsorption/desorption isotherms determined on NS51 before and after the microwave treatment. (light symbols ± adsorption; heavy symbols ± desorption).
the plots obtained were similar. From the slope and intercept of the multilayer region of each as plot the values of the external surface area, Sext , and the total micropore volume, Vs , given in Table 3, were calculated. From the slope and intercept of each DR plot the values of the characteristic energy, E0 , and the micropore volume, V0 , also given in Table 3, were calculated. Estimates of the mean pore width, L0 , were calculated from the relationship [19]: L0 10:8= E0
11:4
1
These values are also included in the ®nal column of Table 3. Comparison of the corresponding values of Vs and V0 shows that all samples with a burn-o of less than 40% contained only narrow (primary) micropores both before and after microwave treatment, whereas the samples with a burn-o greater than 40% contained both primary micropores and
secondary micropores. With these higher burn-o samples the dierence between Vs and V0 was somewhat reduced after microwave treatment, especially in the case of sample F1N-61m15. In all cases, except F2-26, as previously noted, the results con®rm that microwave treatment brought about a signi®cant reduction in pore volume. The values of the mean pore width, L0 , were also reduced in most, although not all cases. Much larger dierences were found with the samples that maintained the highest temperatures during the microwave treatment, namely the three F1 samples and F2-74. With these four samples the mean pore size was reduced by between 0.18 and 0.35 nm. With the other samples the variation was always less than 0.1 nm. 3.2.2. Elemental composition The results of elemental analysis given in Tables 1 and 2 show that microwave treatment for periods of up to 15 min results in an increase in the
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
249
Fig. 3. Low-temperature nitrogen adsorption/desorption isotherms determined on F1N before and after the microwave treatment (light symbols ± adsorption; heavy symbols ± desorption).
carbon content for all of the carbonised and activated ®bres as well as for the Norit samples. This indicates an overall reduction in the total quantity of heteroatoms present in the samples. In particular, it should be noted that the oxygen content is always reduced. The reduction is particularly signi®cant with the Norit samples where the oxygen content after microwave treatment is <1:5% for all samples, compared with values between 1.5% and 7.7% for the ®bres. The higher values obtained with the ®bres may indicate that at least part of the oxygen is very strongly bound, possibly within the pseudo-graphitic layer planes. It would seem likely that this oxygen became included in the structure during polymerisation of the vinyl acetate comonomer. With regard to the nitrogen content, the Norit samples contain only low amounts, <0:4%, both before and after the microwave treatment. The carbonised ®bres initially contain relatively high amounts of nitrogen, and although this is signi®cantly reduced by the microwave treatment the
®nal nitrogen contents of about 7±10% are still quite high. Treatment of the activated ®bres results in a small reduction in the nitrogen content for ®ve of the ®bres (the three F1 ®bres and two of the F1N ®bres) but a small increase for the other four ®bres (one of the F1N ®bres and the three F2 ®bres). The ®nal values of between 3.4% and 7.4% are also relatively high. These results suggest that the carbonised ®bres contain nitrogen in dierent forms. Part of the nitrogen is in the form of surface groups which are readily removed by microwave treatment or by activation. The other part, as was also observed in the case of oxygen, is strongly bound and much more resistant to removal either by microwave treatment or by increasing activation. Presumably, this nitrogen is also contained within the pseudo-graphitic layer planes where it was included during the cyclisation of the nitrile groups of the polymer precursor during carbonisation. The carbonised ®bres and Norit AZO were also treated for the longer period of 30 min. It can be
250
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
Table 3 Textural characteristics of the original ACE and treated ®bres (sample designations as in Table 1) Sample
SBET (m2 g 1 )
as method
DR method
Vs (cm g )
Sext (m g )
V0 (cm3 g 1 )
E0 (kJ mol 1 )
L0 (nm)
3
1
2
1
F1-37 F1-37m15 F1-53 F1-53m15 F1-76 F1-76m15
390 219 1239 768 1341 774
0.18 0.09 0.54 0.36 0.60 0.36
6 1 8 2 6 2
0.19 0.09 0.46 0.30 0.49 0.30
26.1 30.7 20.8 23.8 19.9 23.0
0.74 0.56 1.15 0.87 1.28 0.93
F1N-12 F1N-12m15 F1N-36 F1N-36m15 F1N-61 F1N-61m15
370 163 791 318 1284 730
0.17 0.08 0.35 0.15 0.60 0.32
4 4 6 3 1 4
0.16 0.08 0.34 0.15 0.41 0.30
22.1 17.8 22.9 20.6 20.4 20.7
1.01 1.10 0.94 0.95 1.20 1.16
F2-26 F2-26m15 F2-48 F2-48m15 F2-74 F2-74m15
433 453 1014 655 1397 772
0.19 0.20 0.45 0.29 0.62 0.36
11 1 1 1 15 4
0.19 0.19 0.39 0.26 0.50 0.31
24.1 23.6 20.7 21.2 20.0 23.2
0.85 0.88 1.16 1.11 1.26 0.92
SBET is BET surface area; Vs and Sext are total pore volume and external surface area from as plot; V0 , E0 and L0 are micropore volume, characteristic energy and mean micropore width from DR plot.
seen from Tables 1 and 2 that, after this prolonged microwave treatment, there was a small decrease in the carbon content, when compared with the 15 min treatment, and an increase in the oxygen content. It should be remembered that the elemental analyses were carried out about four weeks after treating the samples. Although the samples were stored in sealed containers during this time they will have adsorbed some oxygen and carbon dioxide from the atmosphere. It is likely that the residual oxygen content of the Norit samples treated for 15 min and a small part (but not all) of the oxygen content of the ®bres after 15 min treatment are due to this reoxidation. The fact that after 30 min of microwave treatment the oxygen content is higher than that found after 15 min suggests that the longer treatment period resulted in a more reactive surface which was subsequently reoxidised to a greater degree. 3.2.3. Point of zero charge Fig. 4 shows the variation in pzc for the carbonised ®bres and the Norit samples as a function of the time of microwave treatment. The initial
values for the untreated Norit samples NAZO and NSXP are dierent from values previously published [20] indicating that in the intervening ®ve years the nature of the surfaces changed, presumably due to adsorption of atmospheric CO2 in the form of surface groups (unfortunately, elemental analysis was not available to us when the previous work was published). It can be seen from Fig. 4 that for up to 15 min treatment time the pzc increases but that, in the case of the carbonised ®bres, after 30 min the pzc is lower than the corresponding value obtained after 15 min treatment. These variations are consistent with the observed variations in the residual oxygen content if, as already suggested, the residual oxygen is due to reoxidation of the samples after the microwave treatment. In all cases, including the samples which were initially slightly acid, the samples are basic after treatment. The very high pzc value of 11.3 given by NAZO after 15 min treatment, which is similar to the value of 11.8 previously reported, is particularly noteworthy. The pzc values of the ACF are indicated in Fig. 5. It can be seen that in all cases the pzc increases
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
Fig. 4. Evolution of the pzc of the carbonised ®bres and Norit samples with the time of microwave treatments.
after microwave treatment reaching, in all but one case, the almost constant value of 10:8 0:2 independently of the initial pzc of the untreated samples. In order to verify that the change in pzc was not due solely to the temperature reached by the samples during the microwave treatment, the ®bres were also heated for 2 h in a ¯ow of N2 in a conventional furnace and then stored for four weeks under the same conditions used for the microwave treated samples. It was found that the pzc was very little aected by this conventional treat-
251
ment. It should also be pointed out that the fact that the same value of pzc is obtained for almost all of the microwave treated samples cannot be associated with a buering eect of the water, as we have already seen with NAZO that higher pzc values are possible. In a previous paper [20] we have presented a detailed consideration of the principal factors which in¯uence the surface ionisation of carbon type materials containing acid groups and basic groups and clari®ed the eect of changing site concentrations and ionisation constants on the value of the pzc and iep (isoelectric point). Considering the case of a basic carbon containing one type of basic site with ionisation constant pKb and one type of acid site with ionisation constant pKa , then the pzc will not depend on the absolute concentration of the basic sites, but will depend on the value of pKb . Over the range of pH values of approximately pKb 2 the basic sites turn from completely ionised to completely unionised and the pzc will occur between these limits. Over the same range of pH values the acid sites will, in general, be completely ionised and the pzc will not therefore be dependent on the value of pKa . On the other hand, the exact value of the pzc will be dependent on the concentration of the acid sites and will decrease with increasing acid site concentration. On the basis of these observations a plausible explanation for the results given in Figs. 4 and 5 is the following. During microwave treatment most, if not all, of the non-ring surface functional groups, which
Fig. 5. Pzc of the ACF samples before (®lled bars) and after (empty bars) the microwave treatments.
252
P.J.M. Carrott et al. / Microporous and Mesoporous Materials 47 (2001) 243±252
would appear to be mainly acid in nature, are removed. Most of the oxygen and nitrogen remaining is included in basic ring groups, and the pzc therefore increases. Furthermore, although the absolute concentrations of the basic groups may vary from one sample to another, their chemical nature is probably the same and hence the pzc has the same value in each case. During storage of the treated samples, some reformation of acid surface groups may occur but to a lesser extent than before treatment, and the pzc therefore remains higher than the initial value. Furthermore, as the pzc remains constant the extent of reformation of acid groups must also be similar in all cases. Thirty-minute treatment of the samples causes more profound changes than 15-min treatment, resulting in a more reactive surface on which reformation of acid surface groups can occur to a greater extent. 4. Conclusions The results presented show that microwave heating is a very eective means of modifying the porosity and, of special importance, the surface chemistry of ACF. Comparatively short heating periods are sucient to completely remove the surface groups, but without removing oxygen and nitrogen atoms bonded within the pseudo-graphitic planes. As the surface groups are mainly acidic in nature, whereas ring oxygen and nitrogen atoms impart basic functionality, the resulting carbon materials are very basic, giving in all cases pzc values close to 11. After microwave treatment the surfaces of the carbon materials are left in a reactive state. In the present work no direct measures (apart from storing under an inert atmosphere) were taken to block the surface reactivity and, as a result, some reformation of surface groups must have occurred slowly during storage. In future work, it would be interesting to study the reaction of the microwavetreated surface, while still in the reactive state, with reducing or other gases, with a view to enhancing the basic character or of anchoring speci®c types of functionality onto the surface. In the former case, the resulting materials would be valuable
adsorbents, for example, for the adsorption of acid gases such as SO2 . In the second case, there are evident opportunities for the development of new catalysts. Acknowledgements The authors are grateful to FISIPE ± Fibras Sinteticas de Portugal for the provision of samples and to the Fundacß~ao para a Ci^encia e a Tecnologia (Portugal), the European Regional Development Fund (FEDER) and the program PRAXIS XXI for ®nancial support (project PRAXIS/3/3.1/ MMA/1781/95). References [1] J.A. Menendez, J. Phillips, B. Xia, L.R. Radovic, Langmuir 12 (1996) 4404. [2] J.A. Menendez, B. Xia, J. Phillips, L.R. Radovic, Langmuir 13 (1997) 3414. [3] J.A. Menendez, L.R. Radovic, B. Xia, J. Phillips, J. Phys. Chem. 100 (1996) 17243. [4] C.A. Le on y Le on, J.M. Solar, V. Calemma, L.R. Radovic, Carbon 30 (1992) 797. [5] H.P. Boehm, M. Voll, Carbon 8 (1970) 227. [6] S.K. Verma, P.L. Walker Jr., Carbon 30 (1992) 837. [7] R.C. Bansal, F.J. Vastola, P.L. Walker Jr., Carbon 12 (1974) 355. [8] T. Suzuki, C. Ishii, N. Shindo, K. Kaneko, J. Porous Mater. 4 (1997) 181. [9] J.A. Menendez, E.M. Menendez, M.J. Iglesias, A. Garcõa, J.J. Pis, Carbon 37 (1999) 1115. [10] J.A. Menendez, E.M. Menendez, A. Garcõa, J.B. Parra, J.J. Pis, J. Microwave Power EE 34 (1999) 137. [11] A. Zlotorzynski, Critic. Rev. Analytic. Chem. 25 (1) (1995) 43. [12] K.E. Haque, Int. J. Miner. Process 57 (1999) 1. [13] K.M. Holland Patent GB 9105735 (May 2, 1991). [14] L.M. Norman, C.Y. Cha, Chem. Eng. Commun. 140 (1996) 87. [15] K.E. Haque, P.D. Kondos, R.J.C. MacDonal, D.E. Laforest, Patent CA 2008242 (July 22, 1991). [16] K.M. Holland, Patent GB 9426661 (November 24, 1994). [17] Y. Kong, C.Y. Cha, Energy Fuels 6 (1996) 1245. [18] P.J.M. Carrott, J.M.V. Nabais, M.M.L. Ribeiro Carrott, J.A. Pajares, Carbon 39 (10) (2001) 1543. [19] H.F. Stoeckli, P. Rebstein, L. Ballerini, Carbon 28 (6) (1990) 907. [20] P.J.M. Carrott, M.M.L. Ribeiro Carrott, A.J. Est^ev~ao Candeias, J.P. Prates Ramalho, J. Chem. Soc. Faraday Trans. 91 (4) (1995) 2179.