《复合材料 Composites》课程教学资源（学习资料）第二章 增强体_carbon fiber_Bulk and surface chemical functionalities of type III PAN-based carbon fibres

CARBON PERGAMON Carbon4l(2003)1905-1915 Bulk and surface chemical functionalities of type Ill PAN-based carbon fibres Elzbieta Pamula,. Paul G. Rouxhet AGH University of Science and Technology, Faculty of Materials Science and Ceramics, al. A. Mickiewic=a 30, 30-059 Krako Poland Universite catholique de Louvain, Unite de chimie des interfaces, Croix du Sud 2118, 1348 Louvain-la-Neuve, Belgium Received 15 June 2002; accepted 15 April 2003 Abstract PAN-based carbon fibres carbonised at relatively low temperature, i.e. type Ill carbon fibres, were submitted to heat treatment at 2300C(GR)or oxidation in nitric acid. The samples were characterised by XPs, FtiR, wetting measurements, as adsorption, elemental analysis and acid/ base titration. While oxidation only slightly affects the nitrogen concentration, it produces an appreciable change in the nature of the chemical functions, namely the conversion of pyridine-type nitrogen and quaternary nitrogen into aliphatic functions. Oxidation treatment modifies all the material constituting the fibre, the oxygen concentration being about 1.5 times higher at the fibre external surface compared with the whole material. Three components 531.2, 532.6 and 533.8 ev) are clearly identified in the oxygen XPS peak, allowing a comparison to be made between the whole material and the external surface regarding chemical species. The acidic groups are mainly carboxyl. Fibres submitted to extensive oxidation also show a high basicity, attributed mainly to calcium carboxylate. Although the acidic and basic groups present in the whole material can be titrated with aqueous solutions, the fibres develop only a very small surface area and no microporosity as determined by krypton adsorption. The material may be viewed as a sponge, collapsed when dry but able to swell in water and developing a high cation-exchange capacity e 2003 Elsevier Ltd. All rights reserved Keywords: A. Carbon fibers; B. Chemical treatment; C. X-ray photoelectron spectroscopy, Infrared spectroscopy; D. Functional groups 1. Introduction resin composites, is treatment with nitric acid. It was found to affect the texture and surface chemical composition of Polyacrylonitrile-based carbon fibres have been used both types of fibres, although the susceptibility to oxida- extensively in composites technology for the last two tion of type I carbon fibres is less than that of type ll, due decades. High modulus carbon fibres(type I, HM)are to a more graphitic and less reactive surface [1, 2]. Electro- obtained at high temperature (above 2000 C)and high chemical oxidation not only results in modification of the strength fibres(type Il, HT)are obtained typically at 1500 surface of type Il carbon fibres, but also leads to the C. The surface properties of both types of carbon fibres formation of chemical functionalities in deeper parts of the have been the subject of many papers, reviews and fibres 3] monographs [1, 2]. It is known that specific surface area, On the other hand, there are only a few publications on a surface energy and surface chemical functionalities in- third type of PAN-based carbon fibre, which is obtained at fluence the behaviour of the carbon fibre-matrix interface relatively low carbonisation temperature(below 1300C) nd, as a result, the mechanical performance of compos- [4-6]. These fibres are called low-temperature carbonised, tes. One of the most widely used wet oxidative treatments, or high-strain because of their relatively high elongation effectively improving the shear properties of carbon fibre or simply type Ill [1]. Due to the lower carbonisation temperature they are less crystalline, which makes them orresponding author. Tel. +48-12-617-2503: fax: +48-12- more sensitive to oxidation than type I or ll; however, the 633-4630 effect of oxidation of these fibres in wet conditions has E-imail address: epamula@ uci agh.edu.pl(E. Pamula) received little attention so far. Nitric acid oxidation of 0008-6223/03/S-see front matter 2003 Elsevier Ltd. All rights reserved doi:10.1016/S0008-6223(03)00177-5

Carbon 41 (2003) 1905–1915 B ulk and surface chemical functionalities of type III PAN-based carbon fibres a, b Elzbieta Pamula , Paul G. Rouxhet * a AGH University of Science and Technology, Faculty of Materials Science and Ceramics, al. A. Mickiewicza 30, 30-059 Krakow´ , Poland b Universite catholique de Louvain ´ ´ , Unite de chimie des interfaces, Croix du Sud 2/18, 1348 Louvain-la-Neuve, Belgium Received 15 June 2002; accepted 15 April 2003 Abstract PAN-based carbon fibres carbonised at relatively low temperature, i.e. type III carbon fibres, were submitted to heat treatment at 2300 8C (GR) or oxidation in nitric acid. The samples were characterised by XPS, FTIR, wetting measurements, gas adsorption, elemental analysis and acid/base titration. While oxidation only slightly affects the nitrogen concentration, it produces an appreciable change in the nature of the chemical functions, namely the conversion of pyridine-type nitrogen and quaternary nitrogen into aliphatic functions. Oxidation treatment modifies all the material constituting the fibre, the oxygen concentration being about 1.5 times higher at the fibre external surface compared with the whole material. Three components (531.2, 532.6 and 533.8 eV) are clearly identified in the oxygen XPS peak, allowing a comparison to be made between the whole material and the external surface regarding chemical species. The acidic groups are mainly carboxyl. Fibres submitted to extensive oxidation also show a high basicity, attributed mainly to calcium carboxylate. Although the acidic and basic groups present in the whole material can be titrated with aqueous solutions, the fibres develop only a very small surface area and no microporosity as determined by krypton adsorption. The material may be viewed as a sponge, collapsed when dry but able to swell in water and developing a high cation-exchange capacity.  2003 Elsevier Ltd. All rights reserved. Keywords: A. Carbon fibers; B. Chemical treatment; C. X-ray photoelectron spectroscopy, Infrared spectroscopy; D. Functional groups 1. Introduction resin composites, is treatment with nitric acid. It was found to affect the texture and surface chemical composition of Polyacrylonitrile-based carbon fibres have been used both types of fibres, although the susceptibility to oxida￾extensively in composites technology for the last two tion of type I carbon fibres is less than that of type II, due decades. High modulus carbon fibres (type I, HM) are to a more graphitic and less reactive surface [1,2]. Electro￾obtained at high temperature (above 2000 8C) and high chemical oxidation not only results in modification of the strength fibres (type II, HT) are obtained typically at 1500 surface of type II carbon fibres, but also leads to the 8C. The surface properties of both types of carbon fibres formation of chemical functionalities in deeper parts of the have been the subject of many papers, reviews and fibres [3]. monographs [1,2]. It is known that specific surface area, On the other hand, there are only a few publications on a surface energy and surface chemical functionalities in- third type of PAN-based carbon fibre, which is obtained at fluence the behaviour of the carbon fibre–matrix interface relatively low carbonisation temperature (below 1300 8C) and, as a result, the mechanical performance of compos- [4–6]. These fibres are called low-temperature carbonised, ites. One of the most widely used wet oxidative treatments, or high-strain because of their relatively high elongation, effectively improving the shear properties of carbon fibre– or simply type III [1]. Due to the lower carbonisation temperature they are less crystalline, which makes them *Corresponding author. Tel.: 148-12-617-2503; fax: 148-12- more sensitive to oxidation than type I or II; however, the 633-4630. effect of oxidation of these fibres in wet conditions has E-mail address: epamula@uci.agh.edu.pl (E. Pamula). received little attention so far. Nitric acid oxidation of 0008-6223/03/$ – see front matter  2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0008-6223(03)00177-5

1906 E. Pamua, P G. Rouxhet /Carbon 41(2003)1905-1915 carbon fibres obtained at 1250C was found to influence Millipore)and dried in a nitrogen flow(sample Ox+1+ their mechanical properties. a decrease of Youngs HCD) modulus from 200 to 100 GPa [7 and an increase of their ultimate strain to failure from 1.5 to 2.6%[8 were 2.2 Methods observed, making these fibres appropriate for medical applications, as ligament and tendon prostheses or sutures 2.2/. Textural measurements the chemistry of these fibres has not been reviewed in the optica diameter of the carbon fibres was measured by [7-10]. However, the influence of nitric acid oxidation on scopy (Lanametr). The average diameter and literature up to now confidence interval were calculated from 100 measure- The aim of the present work was to study the influence ments of individual fibres(a of nitric acid oxidation on the properties of type Ill PAN-based carbon fibres and to investigate the nature the chemical functionalities created, distinguishing be- determined by krypton adsorption at 77K with an tween the external fibre surface and the whole material matic Micrometrics 2400 apparatus and calculated The results also provide advances in the assignment of the the Brunauer, Emmet and Teller (BET)equation. KPS O,s peak components and in the characterisation of basic sites. For comparison, a carbon fibre sample heated 2. 2.2. Contact angle at high temperature in an inert atmosphere was also Wetting of the carbon fibres was determined by the Wilhelmy method using a Cahn Dynamic Contact Angle analy Balance(DCA 322). An elementary fibre(10 mm long) was attached to a nichrome wire with a small piece of 2. Experimental adhesive paper and hung on the arm of the microbalance First, the perimeter of the fibre was determined by means 2.1. Materials of hexane, a low surface tension liquid that wets the surface of the fibre (0=0). This was performed by Carbon fibres (in the form of braid consisting of 18 recording one wetting cycle of immersion and emersion owings and 1800 elementary fibres in each roving) were (advancing-receding), which did not show any hysteresis obtained from polyacrylonitrile(PAN) precursor(produced Then the wetting cycle was recorded with water. The rate from 93% acrylonitrile, 5.9% methyl methacrylate and of displacement was 20 um/s. The cos e was calculated 1. 1% itaconic acid; Chemical Fibres Institute, Lodz, Po- using Eq (D) after correction for buoyancy(F,advancing land)by thermal processing, first in air(stabilisation or receding force; n, surface tension of the liquid; P. process) and then in an inert atmosphere(carbonisation perimeter process).Therefore, the samples were heated in air(heat- F=Px cos 8 ing rate 100C/h)to 230C, maintained at this tempera- ture for 8 h and then heated in nitrogen(heating rate 300 The surface tensions of water and hexane were checked °C/h)to1250C[l using the platinum ring method(Kruss, Hamburg) Six different samples of carbon fibres were prepared. All fibres were carbonised at 1250C (untreated, UN). One 2.2.3. XPs analysis sample(GR) was further heat treated to 2300C in an inert The surface chemical compe position of the carbon fibres atmosphere. A batch of uN was oxidised for 1 h with was determined by XPS, using an SSI X-Probe(SSX-100/ boiling 68% HNO,(oxidised, OX)[Il] After oxidation 206)spectrometer from Fisons, interfaced to a Hewlett- fibres were washed in flowing distilled water for 3 h and Packard 9000/310 computer allowing instrument control then dried at 105C. In order to increase the degree of data accumulation and data treatment. The spectrometer oxidation, OX samples, stored for about I year in normal used monochromatised Al Ka X-ray radiation(1486.6 ev). atmosphere, were treated again in boiling 68% HNO, for Since the samples were conductors, the fiood-gun was 5-3 h(samples OX+0.5 OX+I and OX+3, respective turned off. The pressure during the analyses was between ly ). After treatment, carbon fibres were again washed in 4.5x10 and 2.4x10 Torr. The irradiated zone was an distilled water for 3 h and then dried at 105C. The fibres elliptic spot with a shorter axis of 1000 um. The constant were not submitted to more extensive washing in alkaline pass energy was 150 and 50 ev for wide-scan and detailed conditions or organic solutions, the aim was to characterise peak analysis, respectively. The following sequences of the materials as processed, keeping in mind that extraction petra were recorded: wide-scan spectrum, Cis, os,N would influence the nature of the product obtained Ca2p and Cis again to check for the absence of sample In order to confirm XPS spectral assignments, 0. 2 g of degradation. sample OX+I was immersed in 2.5 ml of 0.1 N HCI Data treatment was performed with the ESCA.3 D solution in a plastic vial for 72 h at room temperature, software provided by the spectrometer manufacturer. The washed in water(purified with a Milli-Q plus system from binding energy (Eb)of the main lines(Cis, Ois, nis)was

1906 E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 carbon fibres obtained at 1250 8C was found to influence Millipore) and dried in a nitrogen flow (sample OX111 their mechanical properties. A decrease of Young’s HCl). modulus from 200 to 100 GPa [7] and an increase of their ultimate strain to failure from 1.5 to 2.6% [8] were 2 .2. Methods observed, making these fibres appropriate for medical applications, as ligament and tendon prostheses or sutures 2 .2.1. Textural measurements [7–10]. However, the influence of nitric acid oxidation on The diameter of the carbon fibres was measured by the chemistry of these fibres has not been reviewed in the optical microscopy (Lanametr). The average diameter and literature up to now. confidence interval were calculated from 100 measure￾The aim of the present work was to study the influence ments of individual fibres (a 5 95%). of nitric acid oxidation on the properties of type III The specific surface area of the carbon fibres was PAN-based carbon fibres and to investigate the nature of determined by krypton adsorption at 77 K with an auto- the chemical functionalities created, distinguishing be- matic Micrometrics 2400 apparatus and calculated using tween the external fibre surface and the whole material. the Brunauer, Emmet and Teller (BET) equation. The results also provide advances in the assignment of the XPS O peak components and in the characterisation of 1s 2 .2.2. Contact angle basic sites. For comparison, a carbon fibre sample heated Wetting of the carbon fibres was determined by the at high temperature in an inert atmosphere was also Wilhelmy method using a Cahn Dynamic Contact Angle analysed. Balance (DCA 322). An elementary fibre (10 mm long) was attached to a nichrome wire with a small piece of adhesive paper and hung on the arm of the microbalance. 2. Experimental First, the perimeter of the fibre was determined by means of hexane, a low surface tension liquid that wets the 2 .1. Materials surface of the fibre (u 5 08). This was performed by recording one wetting cycle of immersion and emersion Carbon fibres (in the form of braid consisting of 18 (advancing–receding), which did not show any hysteresis. rovings and 1800 elementary fibres in each roving) were Then the wetting cycle was recorded with water. The rate obtained from polyacrylonitrile (PAN) precursor (produced of displacement was 20 mm/s. The cos u was calculated from 93% acrylonitrile, 5.9% methyl methacrylate and using Eq. (1) after correction for buoyancy (F, advancing 1.1% itaconic acid; Chemical Fibres Institute, Lodz, Po- or receding force; g , surface tension of the liquid; P, l land) by thermal processing, first in air (stabilisation perimeter of the fibre): process) and then in an inert atmosphere (carbonisation process). Therefore, the samples were heated in air (heat- F 5 Pg cos u (1) l ing rate 100 8C/h) to 230 8C, maintained at this tempera￾ture for 8 h and then heated in nitrogen (heating rate 300 The surface tensions of water and hexane were checked using the platinum ring method (Kruss, Hamburg). 8C/h) to 1250 8C [11]. Six different samples of carbon fibres were prepared. All fibres were carbonised at 1250 8C (untreated, UN). One 2 .2.3. XPS analysis sample (GR) was further heat treated to 2300 8C in an inert The surface chemical composition of the carbon fibres atmosphere. A batch of UN was oxidised for 1 h with was determined by XPS, using an SSI X-Probe (SSX-100/ boiling 68% HNO (oxidised, OX) [11]. After oxidation, 206) spectrometer from Fisons, interfaced to a Hewlett- 3 fibres were washed in flowing distilled water for 3 h and Packard 9000/310 computer allowing instrument control, then dried at 105 8C. In order to increase the degree of data accumulation and data treatment. The spectrometer oxidation, OX samples, stored for about 1 year in normal used monochromatised Al K X-ray radiation (1486.6 eV). a atmosphere, were treated again in boiling 68% HNO for Since the samples were conductors, the flood-gun was 3 0.5–3 h (samples OX10.5 OX11 and OX13, respective- turned off. The pressure during the analyses was between 29 29 ly). After treatment, carbon fibres were again washed in 4.5310 and 2.4310 Torr. The irradiated zone was an distilled water for 3 h and then dried at 105 8C. The fibres elliptic spot with a shorter axis of 1000 mm. The constant were not submitted to more extensive washing in alkaline pass energy was 150 and 50 eV for wide-scan and detailed conditions or organic solutions; the aim was to characterise peak analysis, respectively. The following sequences of the materials as processed, keeping in mind that extraction spectra were recorded: wide-scan spectrum, C , O , N , 1s 1s 1s would influence the nature of the product obtained. Ca and C again to check for the absence of sample 2p 1s In order to confirm XPS spectral assignments, 0.2 g of degradation. sample OX11 was immersed in 2.5 ml of 0.1 N HCl Data treatment was performed with the ESCA 8.3 D solution in a plastic vial for 72 h at room temperature, software provided by the spectrometer manufacturer. The washed in water (purified with a Milli-Q plus system from binding energy (E ) of the main lines (C , O , N ) was b 1s 1s 1s

E. Pamua, P G. Rouxhet /Carbon 41(2003)1905-1915 determined by setting a value of 284.8 eV for the main C lemental composition component, due to carbon only bound to carbon and tal analysis of carbon, hydrogen, nitrogen and hydrogen. The peak area was determined using Shirley difference)of the analysed fibres was carried type non-linear background subtraction. The spectral en- a Perkin-Elmer ChN Elemental Analyser, Model elopes were decomposed using a non-linear least-square routine, assuming a Gaussian/Lorenzian(85: 15)function OIs and Ni peaks were decomposed without a priori 2. 2.6. Fourier transform infrared spectroscopy constraints regarding the characteristics of the components To prepare the samples for FTIR analysis the fibres were Intensity ratios were converted into molar concentration rushed, powdered and pressed with potassium bromide to atios using the sensitivity factors proposed by the manu- form a pellet containing about 0. 2% carbon fibres. The facturer(Scofield emission cross sections, variation of the absorption spectra were recorded on a Digilab FTS 60v electron mean free path according to the 0.7 power of(Bio-Rad) Fourier transform spectrometer. The spectra kinetic energy, constant transmission function)[12] were obtained by averaging 256 scans at a resolution of 4 cm in the frequency range 400-4000 cm 2.2. 4. Acid and base uptake The concentration of acidic and basic groups 3. Results determined by titration. Distilled water was boiled prepare a NaOH solution. a known portion of carbon 3. I. Texure fibres (2.0 g) was immersed in 25 ml 0. 1 N NaOH solution in a plastic vial for 72 h at room temperature. After that Table I presents the average diameter of the fibres, time an aliquot of the liquid phase was back-titrated with obtained by optical microscopy, and their BET surface N HCI using a CP-315 pH meter (Elmetron). De- area. The surface area of UN carbon fibres was about termination of the concentration of basic groups was m /g and decreased to 0.42 m/g after heat treatment at carried out in the same manner, but carbon fibres were 2300C(GR)and to about 0.35 m /g for fibres submitted immersed in HCI solution and back-titrated with Naoh to post-treatment oxidation(OX+0.5, OX+l, OX+3) The krypton adsorption isotherm(p/po range up to 0.3) Table I Texture, water contact angle, acidity and basicity of analysed fibres BET Water contact angle 0(deg) (m2/g dvancing Receding Acidic 0.42 10.3 82.7 67.9 19.3 (0.1) (38) (5.4) 76.5 358 (2.6) Ox+0.5 72.8 44.2 (0.15) OX+I 8.8 75.0 (0. (2.1) 0.45 74.4 43.3 730 (0.13) (1.8) 7 Confidence interval in parentheses (a=0.95) is the slope of the amount of adsorbed krypton with respect to the amount adsorbed on UN at the same relative pressure Immersion in water impossible Micropore volume 0.03 cm/g From which about 20% is related to compounds which dissolved upon immersion in 0. 1 N NaOH From which about 50% is related to compounds which dissolved upon immersion in 0. I N NaOH

E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 1907 determined by setting a value of 284.8 eV for the main C1s 2 .2.5. Elemental composition component, due to carbon only bound to carbon and Elemental analysis of carbon, hydrogen, nitrogen and hydrogen. The peak area was determined using Shirley- oxygen (by difference) of the analysed fibres was carried type non-linear background subtraction. The spectral en- out using a Perkin-Elmer CHN Elemental Analyser, Model velopes were decomposed using a non-linear least-square 2400. routine, assuming a Gaussian/Lorenzian (85:15) function. O and N peaks were decomposed without a priori 2 .2.6. Fourier transform infrared spectroscopy 1s 1s constraints regarding the characteristics of the components. To prepare the samples for FTIR analysis the fibres were Intensity ratios were converted into molar concentration crushed, powdered and pressed with potassium bromide to ratios using the sensitivity factors proposed by the manu- form a pellet containing about 0.2% carbon fibres. The facturer (Scoffield emission cross sections, variation of the absorption spectra were recorded on a Digilab FTS 60v electron mean free path according to the 0.7 power of (Bio-Rad) Fourier transform spectrometer. The spectra kinetic energy, constant transmission function) [12]. were obtained by averaging 256 scans at a resolution of 4 21 21 cm in the frequency range 400–4000 cm . 2 .2.4. Acid and base uptake The concentration of acidic and basic groups was 3. Results determined by titration. Distilled water was boiled to prepare a NaOH solution. A known portion of carbon 3 .1. Texture fibres (2.0 g) was immersed in 25 ml 0.1 N NaOH solution in a plastic vial for 72 h at room temperature. After that Table 1 presents the average diameter of the fibres, time an aliquot of the liquid phase was back-titrated with obtained by optical microscopy, and their BET surface 0.1 N HCl using a CP-315 pH meter (Elmetron). De- area. The surface area of UN carbon fibres was about 1 2 2 termination of the concentration of basic groups was m /g and decreased to 0.42 m /g after heat treatment at 2 carried out in the same manner, but carbon fibres were 2300 8C (GR) and to about 0.35 m /g for fibres submitted immersed in HCl solution and back-titrated with NaOH to post-treatment oxidation (OX10.5, OX11, OX13). solution. The krypton adsorption isotherm ( p/p range up to 0.3) 0 T able 1 Texture, water contact angle, acidity and basicity of analysed fibres a bc Sample Diameter BET a Water contact angle , Chemical UN (mm) surface u (deg) functions a area (meg/g) 2 Advancing Receding (m /g) Acidic Basic d d GR 7.6 0.42 0.48 – – 1.1 10.3 (0.1) (0.03) UN 8.1 1.06 1 82.7 67.9 9.9 19.3 (0.1) (0.02) (3.8) (5.4) e e f OX 8.5 – – 76.5 53.0 358 225 (0.1) (2.5) (2.6) OX10.5 8.7 0.35 0.39 72.8 44.2 936 392 (0.15) (0.04) (1.7) (2.7) OX11 8.8 0.35 0.40 75.0 45.2 1200 508 (0.19) (0.04) (2.6) (2.1) g OX13 7.6 0.39 0.45 74.4 43.3 1430 730 (0.13) (0.04) (1.8) (1.7) a Confidence interval in parentheses (a 5 0.95). b a is the slope of the amount of adsorbed krypton with respect to the amount adsorbed on UN at the same relative pressure. UN c Standard deviation in parentheses (n53). d Immersion in water impossible. e 3 Micropore volume 0.03 cm /g. f From which about 20% is related to compounds which dissolved upon immersion in 0.1 N NaOH. g From which about 50% is related to compounds which dissolved upon immersion in 0.1 N NaOH

E. Pamua, P G. Rouxhet /Carbon 41(2003)1905-1915 of sample OX was different from those of other samples that the total peak area is proportional to the mole fraction Therefore, all isotherms were compared with that of un by of oxygen in each sample. In the literature the oxygen peak a method analogous to the I plot (de Boer) or a plot of carbon materials is frequently decomposed into two (Singh)[13]: the adsorbed amount for a given sample was components: one in the range 531. 2-532.6 eV, attributed to plotted as a function of the amount adsorbed by un at the oxygen doubly bound to carbon, and the second in the same relative pressure. All plots, except that of oX, passed range 532.8-533. 1 ev, attributed to oxygen singly bound to through the origin; the slopes obtained are given in Table 1 carbon [15, 18, 19]. Fig. 2 shows that three peaks must be and their variation fits the variation of the BET surface taken into consideration for the set of spectra of the area. For OX, the intercept was about 0.17 mmol Kr/g, materials investigated here. These three components are revealing a microporosity of about 0.03 cm/g(taking into found at ( 1)531. 2,(11)532.6 and ( iii)533.8 ev and may be account the density of Kr at 77K)[14]. It thus appears that assigned, respectively, on the basis of data for carbon oxidation of the carbon fibres created microporosity, which materials, polymers and organic salts, as follows: (i) could no longer be observed after oxidation post-treatment. carboxylate(Co0)[12 and oxygen doubly bound to carbon( o=C-o)in carboxylic acids and esters [12, 20, 21] 3.2. Wetting properties 1i)oxygen of phenol, alcohol, aliphatic ether [21-23] and ketone functions [24];(iii) oxygen of aryl ethers p-o-p The results of dynamic contact angle measurements are [12, 21 and oxygen singly bound to carbon in carboxylic also presented in Table 1. Advancing and receding contact acids and esters (o-C-o)[12, 20, 21]. The additional angles decreased according to the sequence untreated- oxidation considerably changed the shape of the O,s peak oxidised-submitted to oxidation post-treatment, indicating by increasing the component at 531 2 eV, subsequent acid that the carbon fibre surface became more hydrophilic incubation(sample OX+I soaked in HCD)restored ap- upon oxidation. The wetting properties were not signifi proximate equal intensity for the components at 531.2 and cantly influenced by the duration of the second oxidation. 533. 8 ev(Table 2) On the other hand. the 2300 oc treatment caused an Fig. 2 shows representative Nis peaks, the total peak increase in surface hydrophobicity; this was such that the area is also proportional to the mole fraction of nitrogen GR fibre could not be immersed in water, while it was well for each sample. The Nis peak is decomposed into three netted by hexar nponents: a peak at 398.4 eV attributed moieties [25, 26], the main peak at 400-401 eV containing 3.3. XPS analysis a contribution from quaternary or protonated nitrogen [12, 26, 27] and a peak at 4055 ev due to oxidised forms of e. Table 2 presents the carbon, oxygen and nitrogen nitrogen(NO, )[25-28] ernal surface concentrations of the analysed fibres. For After moderate oxidation in nitric acid(OX) the shape UN, GR and oX these were the only elements detected by of the O,s peak did not change significantly, a new peak XPS. The post-treated fibres(OX+0.5, OX+1, oX+3) due to-NO2 species appeared at 405.5 ev and the main gave a Ca/(C+O+N) molar ratio of 0.016-0.019 nitrogen peak shifted from 401.0 to 400.6 eV. After an After oxidation, the CIs peak shows the presence of additional oxidation(samples OX+0.5, OX+ I and OX+ components assigned to oxygenated chemical functions, 3)the component at 398.4 ev, attributed to pyridine already described in the literature [3, 15-17. Fig. I shows groups, disappeared, the main nitrogen peak further shifted epresentative O,s peaks, the ordinate scale is adjusted to 400.0 ev without variation of the associated nitrogen Surface composition determined by XPS [mol/100 mol(C+O+N)J Binding energy (ev): 531.2 532.6 533.8 398.4 400-401 405.0 Csot O,et Not Ca DC-OH 0=C-OH NO, ①O BDL 97.3 2.7 BDLBDL 224 BDL 80.915.6 3.5 BDL 0x+0.5 4.9 BDL 2.8 OX+I 11.3 BDL 2.7 0.6 1.6 72.823.9 OX+1+HCI 7.8 7.9 6.1 BDL 2.8 0.6 74.821.834 BDL Pyridine-type nitrogen See discussion

1908 E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 of sample OX was different from those of other samples. that the total peak area is proportional to the mole fraction Therefore, all isotherms were compared with that of UN by of oxygen in each sample. In the literature the oxygen peak a method analogous to the t plot (de Boer) or a plot of carbon materials is frequently decomposed into two (Singh) [13]: the adsorbed amount for a given sample was components: one in the range 531.2–532.6 eV, attributed to plotted as a function of the amount adsorbed by UN at the oxygen doubly bound to carbon, and the second in the same relative pressure. All plots, except that of OX, passed range 532.8–533.1 eV, attributed to oxygen singly bound to through the origin; the slopes obtained are given in Table 1 carbon [15,18,19]. Fig. 2 shows that three peaks must be and their variation fits the variation of the BET surface taken into consideration for the set of spectra of the area. For OX, the intercept was about 0.17 mmol Kr/g, materials investigated here. These three components are 3 revealing a microporosity of about 0.03 cm /g (taking into found at (i) 531.2, (ii) 532.6 and (iii) 533.8 eV and may be account the density of Kr at 77 K) [14]. It thus appears that assigned, respectively, on the basis of data for carbon oxidation of the carbon fibres created microporosity, which materials, polymers and organic salts, as follows: (i) 2 could no longer be observed after oxidation post-treatment. carboxylate (COO ) [12] and oxygen doubly bound to ] carbon (O=C–O) in carboxylic acids and esters [12,20,21]; ] 3 .2. Wetting properties (ii) oxygen of phenol, alcohol, aliphatic ether [21–23] and ketone functions [24]; (iii) oxygen of aryl ethers F–O–F ] The results of dynamic contact angle measurements are [12,21] and oxygen singly bound to carbon in carboxylic also presented in Table 1. Advancing and receding contact acids and esters (O=C–O) [12,20,21]. The additional ] angles decreased according to the sequence untreated– oxidation considerably changed the shape of the O peak 1s oxidised–submitted to oxidation post-treatment, indicating by increasing the component at 531.2 eV; subsequent acid that the carbon fibre surface became more hydrophilic incubation (sample OX11 soaked in HCl) restored ap￾upon oxidation. The wetting properties were not signifi- proximate equal intensity for the components at 531.2 and cantly influenced by the duration of the second oxidation. 533.8 eV (Table 2). On the other hand, the 2300 8C treatment caused an Fig. 2 shows representative N peaks; the total peak 1s increase in surface hydrophobicity; this was such that the area is also proportional to the mole fraction of nitrogen GR fibre could not be immersed in water, while it was well for each sample. The N peak is decomposed into three 1s wetted by hexane. components: a peak at 398.4 eV attributed to pyridine moieties [25,26], the main peak at 400–401 eV containing 3 .3. XPS analysis a contribution from quaternary or protonated nitrogen [12,26,27] and a peak at 405.5 eV due to oxidised forms of Table 2 presents the carbon, oxygen and nitrogen nitrogen (–NO ) [25–28]. ] 2 external surface concentrations of the analysed fibres. For After moderate oxidation in nitric acid (OX) the shape UN, GR and OX these were the only elements detected by of the O peak did not change significantly; a new peak 1s XPS. The post-treated fibres (OX10.5, OX11, OX13) due to –NO species appeared at 405.5 eV and the main 2 gave a Ca/(C1O1N) molar ratio of 0.016–0.019. nitrogen peak shifted from 401.0 to 400.6 eV. After an After oxidation, the C peak shows the presence of additional oxidation (samples OX 1s 10.5, OX11 and OX1 components assigned to oxygenated chemical functions, as 3) the component at 398.4 eV, attributed to pyridine already described in the literature [3,15–17]. Fig. 1 shows groups, disappeared, the main nitrogen peak further shifted representative O peaks; the ordinate scale is adjusted so to 400.0 eV without variation of the associated nitrogen 1s T able 2 Surface composition determined by XPS [mol/100 mol (C1O1N)] Binding energy (eV): 531.2 532.6 533.8 398.4 400–401 405.0 C O N Ca tot tot tot a b Assignment: O] ]] ] =C–OH C–OH O=C–OH N N –NO2 2 COO O=C F–O–F ] ] ] GR BDL 1.8 0.9 BDL BDL BDL 97.3 2.7 BDL BDL UN 2.0 5.1 2.1 0.6 2.6 BDL 87.6 9.2 3.2 BDL OX 3.9 9.6 2.1 0.5 2.7 0.3 80.9 15.6 3.5 BDL OX10.5 11.0 4.9 4.7 BDL 2.8 0.9 75.7 20.6 3.7 1.6 OX11 11.3 5.9 4.4 BDL 2.7 0.6 75.1 21.6 3.3 1.8 OX13 12.7 5.7 5.5 BDL 2.7 0.6 72.8 23.9 3.3 1.9 OX111HCl 7.8 7.9 6.1 BDL 2.8 0.6 74.8 21.8 3.4 BDL BDL, below detection limit. a Pyridine-type nitrogen. b See Discussion

E. Pamua, P G. Rouxhet /Carbon 41(2003)1905-1915 GR GR A OX 408406404402400398396 Binding Fig. 2. Nis XPS spectra of carbon fibres: untreated (UN), oxidised with nitric acid(OX), submitted to post-treatment with nitric acid 532 adjusted so that the peak area is proportional to the nitrogen mole ng energy(e fraction Fig. 1. O XPS spectra of carbon fibres: untreated (UN), oxidised th nitric acid(OX), submitted to post-treatment with nitric acid in an increase of acidic groups up to about 1. 4 meq/g, and for I h(OX+1), and heat treated to 2300C(GR). Ordinate scale of basic groups up to about 0.7 meq/g. The immersion of djusted so that the peak area is proportional to the oxygen mole OX and post-treated fibres in Naoh led to a yellow and fraction. black solution colour, respectively. Back-titration of the latter with HCI clearly shows two steps: the first step at pH 7-8 allowed computation of the total acidity of the starting concentration and the intensity of the peak assigne material, while the second step, near pH 4, was attributed to back-titration of dissolved species. The latter represent about 20 and 50% of the total acidity of OX and OX+3, 3.4. Acidic and basic groups Table I presents the concentration of acidic and basic 3.5. Elemental composition groups determined by NaoH and HCl uptake, respectively expressed in ueq/g. The concentration of acidic groups for Table 3 presents the elemental composition of the UN carbon fibres was of the order of 10 ueq/g and that of analysed samples expressed in wt% and in mole fraction of basic groups was about twice that. Heating UN fibres at C,O and n with respect to C+O+N(for comparison with 2300C(GR) produced a decrease of both basic and acidic XPS data, Table 2). Due to treatment with nitric acid, the groups, the amount of the latter being close to the concentration of oxygen and hydrogen increased, while quantification limit. Prolonged chemical treatment resulted only a slight increase of nitrogen was observed. On the

E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 1909 Fig. 2. N XPS spectra of carbon fibres: untreated (UN), oxidised 1s with nitric acid (OX), submitted to post-treatment with nitric acid for 1 h (OX11), and heat treated to 2300 8C (GR). Ordinate scale adjusted so that the peak area is proportional to the nitrogen mole fraction. Fig. 1. O XPS spectra of carbon fibres: untreated (UN), oxidised 1s in an increase of acidic groups up to about 1.4 meq/g, and with nitric acid (OX), submitted to post-treatment with nitric acid of basic groups up to about 0.7 meq/g. The immersion of for 1 h (OX11), and heat treated to 2300 8C (GR). Ordinate scale adjusted so that the peak area is proportional to the oxygen mole OX and post-treated fibres in NaOH led to a yellow and fraction. black solution colour, respectively. Back-titration of the latter with HCl clearly shows two steps: the first step at pH 7–8 allowed computation of the total acidity of the starting concentration and the intensity of the peak assigned to material, while the second step, near pH 4, was attributed –NO species increased. to back-titration of dissolved species. The latter represents 2 about 20 and 50% of the total acidity of OX and OX13, 3 .4. Acidic and basic groups respectively. Table 1 presents the concentration of acidic and basic 3 .5. Elemental composition groups determined by NaOH and HCl uptake, respectively, expressed in meq/g. The concentration of acidic groups for Table 3 presents the elemental composition of the UN carbon fibres was of the order of 10 meq/g and that of analysed samples expressed in wt% and in mole fraction of basic groups was about twice that. Heating UN fibres at C, O and N with respect to C1O1N (for comparison with 2300 8C (GR) produced a decrease of both basic and acidic XPS data, Table 2). Due to treatment with nitric acid, the groups, the amount of the latter being close to the concentration of oxygen and hydrogen increased, while quantification limit. Prolonged chemical treatment resulted only a slight increase of nitrogen was observed. On the

E. Pamula, P.G. Rouxhet Carbon 41(2003)1905-1915 Table 3 Elemental composition of carbon fibres omposition given in mole fraction with respect to C+N+O(mol%) sh (wt%) (wt%) (wt%) (C+O+N) (C+O+N) (C+O+N) UN 1 0.56 8.81 5.8 OX+I OX 18 12.3 6.1 BDL, below detection limit; o, obtained by difference tration of oxygen hydrogen and nitrogen providing more spectra (ox 1-ox and (ox+3)=ox not shown had than 99% carbon exactly the same sh ith a higher intens latter 3. 6. Infrared analysis The FTIR spectra of the carbon fibres presented in Fig 3 4. Discussion show a variation similar to that observed for high surface area carbons submitted to oxidation [29], with an increase 4.1. Chemical functions of the bands at 3100-3600 cm(OH stretching vibrations 30,31]), 1720 cm(C=O stretching vibrations of car- XPS analysis(Table 2, Fig. 1) shows that the oxidation bonyl and carboxy l),1600 cm(characteristic of carbons of carbon fibres increases the concentration of oxygen- carbonised at low temperature and assigned to the stretch- containing functions at the external surface and modifies ing of C=C and C=N bonds, with a possible contribution of their distribution in terms of function type. In the case of asymmetric stretching vibrations of -Coo groups [32] UN fibres the concentrations of oxygen responsible for the and 1250-1400 cm(symmetric stretching vibrations of two side-components(531 2 eV attributable to COOH and COo groups, OH deformation vibrations and C-o CO0 and 533.8 ev attributable to COOH and aryl ether) Heat treatment at 2300 C removes all carboxyl and carboxylate functions [36], as shown by the disappearance OX+3 of the 531.2 eV component in the O,s spectrum of GR. After oxidation, particularly after application of a post 0X+1 treatment, the component at 531.2 ev becomes much more intense than the component at 533.8 eV, indicating the presence of carboxylate For OX+0.5, OX+I and OX+3 the difference of molar ratio between the two types of oxygen is about 0.07 oxygen atom per atom of (C+o+N) (Table 2). On the other hand, XPS reveals a Ca/(C+o+ N)ratio of 0.016-0.019. This indicates that about 0.035 carboxy late groups per atom of(C+o+N) are present at the fibre external surface in the form of calcium car- boxylate. The latter is attributed to carboxyl deprotonation nd calcium pick up during extensive rinsing subsequent to the chemical treatment; this is in agreement with the increase of ash content detected by chemical analysis (Table 3 ). This assignment is confirmed by the restoration of a similar intensity for the 531.2 and 533.8 ev com- 350030002500 ponents of sample OX+I soaked in HCI. This is in accordance with the recent observation that the surfac Fig. 3. FTIR spectra of carbon fibres: untreated (UN), oxidised concentration of na and ca increase as a result of the th nitric acid(OX), and submitted to post-treatment with nitric electrochemical oxidation of carbon fibers [37 acid for 1 and 3 h(OX+ I and Ox+3) Wet oxidation of carbon fibres also affects the nature of

1910 E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 T able 3 Elemental composition of carbon fibres (wt%) and composition given in mole fraction with respect to C1N1O (mol%) diff Sample C O N H Ash C/ O/ N/ (wt%) (wt%) (wt%) (wt%) (wt%) (C1O1N) (C1O1N) (C1O1N) (mol%) (mol%) (mol%) GR 99.3 0.07 BDL 0.13 0.50 99.95 0.05 BDL UN 91.5 1.71 5.55 0.61 0.63 93.8 1.3 4.9 OX 85.7 7.36 5.97 0.56 1.11 89.0 5.7 5.3 OX10.5 83.1 8.81 6.48 0.74 0.87 87.3 6.9 5.8 OX11 82.3 9.51 6.61 0.65 0.93 86.5 7.5 6.0 OX13 76.0 15.28 6.63 0.91 1.18 81.6 12.3 6.1 BDL diff , below detection limit; O , obtained by difference. other hand, the 2300 8C treatment reduced the concen- stretching vibrations, respectively [31–35]). Differential tration of oxygen, hydrogen and nitrogen, providing more spectra (OX11)2OX and (OX13)2OX (not shown) had than 99% carbon. exactly the same shape, with a higher intensity for the latter. 3 .6. Infrared analysis The FTIR spectra of the carbon fibres presented in Fig. 3 4. Discussion show a variation similar to that observed for high surface area carbons submitted to oxidation [29], with an increase 4 .1. Chemical functions 21 of the bands at 3100–3600 cm (OH stretching vibrations 21 [30,31]), 1720 cm (C=O stretching vibrations of car- XPS analysis (Table 2, Fig. 1) shows that the oxidation 21 bonyl and carboxyl), 1600 cm (characteristic of carbons of carbon fibres increases the concentration of oxygen￾carbonised at low temperature and assigned to the stretch- containing functions at the external surface and modifies ing of C=C and C=N bonds, with a possible contribution of their distribution in terms of function type. In the case of 2 asymmetric stretching vibrations of –COO groups [32]) UN fibres the concentrations of oxygen responsible for the 21 and 1250–1400 cm (symmetric stretching vibrations of two side-components (531.2 eV attributable to COOH and 2 2 ] COO groups, OH deformation vibrations and C–O COO and 533.8 eV attributable to COOH and aryl ether) ] ] are similar, so they may be entirely assigned to COOH. Heat treatment at 2300 8C removes all carboxyl and carboxylate functions [36], as shown by the disappearance of the 531.2 eV component in the O spectrum of GR. 1s After oxidation, particularly after application of a post￾treatment, the component at 531.2 eV becomes much more intense than the component at 533.8 eV, indicating the presence of carboxylate. For OX10.5, OX11 and OX13, the difference of molar ratio between the two types of oxygen is about 0.07 oxygen atom per atom of (C1O1N) (Table 2). On the other hand, XPS reveals a Ca/(C1O1 N) ratio of 0.016–0.019. This indicates that about 0.035 carboxylate groups per atom of (C1O1N) are present at the fibre external surface in the form of calcium car￾boxylate. The latter is attributed to carboxyl deprotonation and calcium pick up during extensive rinsing subsequent to the chemical treatment; this is in agreement with the increase of ash content detected by chemical analysis (Table 3). This assignment is confirmed by the restoration of a similar intensity for the 531.2 and 533.8 eV com￾ponents of sample OX11 soaked in HCl. This is in accordance with the recent observation that the surface 1 21 Fig. 3. FTIR spectra of carbon fibres: untreated (UN), oxidised concentration of Na and Ca increase as a result of the with nitric acid (OX), and submitted to post-treatment with nitric electrochemical oxidation of carbon fibers [37]. acid for 1 and 3 h (OX11 and OX13). Wet oxidation of carbon fibres also affects the nature of

E. Pamula, P.G. Rouxhet Carbon 41(2003)1905-1915 the nitrogen-containing chemical functionalities. The chemical state of nitrogen in PAN-based oxidised carbon vestigations of Pels et al. [26] and Stanczyk et al. 12/22 8P 20 model compound carbonisation indicated that, above 723 K, pyrrole-type nitrogen(N-5) disappears gradually and is converted into pyridine-type nitrogen(N-6 )and quaternary 8 15 nitrogen(N-Q). Amides and lactams are less stable and are ully converted or decomposed above 400 and 800C espectively [38]. Our results clearly show the presence of N-6(3984 ev)and N-Q(401 ev)incorporated in graphene layers at the external surface of UN carbon fibers. Heating UN carbon fibres to 2300C(sample GR) totally removes nitrogen from the graphene layers. Oxida- tion with nitric acid creates nitroxy groups at the carbon fibre external surface and causes the disappearance of the component attributed to pyridine-type nitrogen(N-6). The shift of the main Ns component from 401.0 to 400.6 ev, ON(C+O+N), whole(mole%) and, upon further oxidation, to 400.0 eV, suggests the Fig. 4. Plot of the surface concentration of oxygen, obtained br conversion of quaternary nitrogen into aliphatic structures XPS, vS. the oxygen concentration in the whole material: (O)UN; like amide (399.9 ev), amine (399.7 ev), nitroso(400. (O)GR,(O)OX,()post-treated fibres(OX+0.5, OX+I ev)and lactam(399.7 eV)[23, 26- Ox+3) moles of (C+O+N), it was considered that the con- 4.2. Bulk vs surface reactivity centrations of other constituents(H, ash) were negligible Fig. 5a shows that the whole fibre material develops an <s The increase of the total oxygen concentration(Table 3) acidity of the order of 17 eq per 100 oxygen atoms. Fig 5b much too high to be attributed exclusively to a modi- reveals that, for 100 oxygen atoms present near the fibre fication of the external surface. Relating the acidity and extermal surface, about 20 are in the form of OH from basicity of fibres submitted to oxidation post-treatment to carboxyl or in the form of aryl ether. The formation of aryl the BEt surface area would give results three orders of ether is expected to be much less than the formation of magnitude higher than reasonably possible. This has also carboxyl upon oxidation. It thus appears from Figs. 4 and 5 been observed for type Il fibres oxidised by electrochemi that the concentrations of oxygen and acid functions, in cal treatment [3, 38]. This shows that oxidation took place particular carboxyl functions, are only slightly higher(less within the fibres, as confirmed by the appearance of than 1.5 times) at the fibre external surface compared with carboxyl, carboxylate and hydroxyl bands in the FTir the whole material petra. While the whole material is accessible to chemical Fig. 6a shows that the whole fibre material reagents in aqueous solution, this is due, as discussed basicity of about 8 eq per 100 oxygen atoms The basic elow, to swelling and not to the presence of defined properties of carbons are attributed to oxygen function- pores, which should be quantified by gas adsorption. In alities such as pyrone-type structures, quinones and order to avoid any ambiguity, the termwhole material is chromene- type groups as well nitrogen-containing groups thus preferred to ' internal surface for the designated bulk [1, 2, 37]. Our results, however, reveal the presence of analytical data, while 'external surface'refers to XPS data. carboxylate groups, which may reprotonate and thus Fig 4 presents a plot of the oxygen mole fraction at the possess a basic character. Fig. 6b reveals that, at the external surface (XPS, Table 2) as a function of the external surface of post-treated fibres, about 30 of 10 ygen mole fraction in the whole material(chemical oxygen atoms are in the form of carboxylate, this would fibres heated at 2300C, the oxygen concentration at the calcium carboxylate formed upon rinsing in water ove as analysis, Table 3). It shows that, for untreated fibres or represent 15 carboxylate functions, identified ab external surface is much higher than the whole oxygen In contrast to oxy taining functions, nitrogen concentration. In contrast, the increase of the oxygen less concentrated at the external surface(Table 2)com concentration due to oxidation is only about 1.5 times pared with the whole material (Table 3). The available higher at the external surface compared with the whole data do not allow an evaluation of the respective amounts material of nitrogen present in the form of acidic, basic or neutral Figs. 5a and 6a present plots of acidity and basicity as a(amide) functions. However, the possible concentration of function of the oxygen concentration, using comparable any of these functions at the extemal surface is appreciably concentration units. In order to convert 1 g of fibre into less than the concentration of carboxyl and carboxylate

E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 1911 the nitrogen-containing chemical functionalities. The chemical state of nitrogen in PAN-based oxidised carbon fibres is not clearly defined in the literature. The in￾vestigations of Pels et al. [26] and Stanczyk et al. [27] on model compound carbonisation indicated that, above 723 K, pyrrole-type nitrogen (N-5) disappears gradually and is converted into pyridine-type nitrogen (N-6) and quaternary nitrogen (N-Q). Amides and lactams are less stable and are fully converted or decomposed above 400 and 800 8C, respectively [38]. Our results clearly show the presence of N-6 (398.4 eV) and N-Q (401 eV) incorporated in graphene layers at the external surface of UN carbon fibers. Heating UN carbon fibres to 2300 8C (sample GR) totally removes nitrogen from the graphene layers. Oxida￾tion with nitric acid creates nitroxy groups at the carbon fibre external surface and causes the disappearance of the component attributed to pyridine-type nitrogen (N-6). The shift of the main N component from 401.0 to 400.6 eV, 1s Fig. 4. Plot of the surface concentration of oxygen, obtained by and, upon further oxidation, to 400.0 eV, suggests the XPS, vs. the oxygen concentration in the whole material: (d) UN; conversion of quaternary nitrogen into aliphatic structures (s) GR; (h) OX; (j) post-treated fibres (OX10.5, OX11, like amide (399.9 eV), amine (399.7 eV), nitroso (400.1 OX13). eV) and lactam (399.7 eV) [23,26]. moles of (C1O1N), it was considered that the con- 4 .2. Bulk vs. surface reactivity centrations of other constituents (H, ash) were negligible. Fig. 5a shows that the whole fibre material develops an The increase of the total oxygen concentration (Table 3) acidity of the order of 17 eq per 100 oxygen atoms. Fig. 5b is much too high to be attributed exclusively to a modi- reveals that, for 100 oxygen atoms present near the fibre fication of the external surface. Relating the acidity and external surface, about 20 are in the form of OH from basicity of fibres submitted to oxidation post-treatment to carboxyl or in the form of aryl ether. The formation of aryl the BET surface area would give results three orders of ether is expected to be much less than the formation of magnitude higher than reasonably possible. This has also carboxyl upon oxidation. It thus appears from Figs. 4 and 5 been observed for type II fibres oxidised by electrochemi- that the concentrations of oxygen and acid functions, in cal treatment [3,38]. This shows that oxidation took place particular carboxyl functions, are only slightly higher (less within the fibres, as confirmed by the appearance of than 1.5 times) at the fibre external surface compared with carboxyl, carboxylate and hydroxyl bands in the FTIR the whole material. spectra. While the whole material is accessible to chemical Fig. 6a shows that the whole fibre material develops a reagents in aqueous solution, this is due, as discussed basicity of about 8 eq per 100 oxygen atoms. The basic below, to swelling and not to the presence of defined properties of carbons are attributed to oxygen function￾pores, which should be quantified by gas adsorption. In alities such as pyrone-type structures, quinones and order to avoid any ambiguity, the term ‘whole material’ is chromene-type groups as well nitrogen-containing groups thus preferred to ‘internal surface’ for the designated bulk [1,2,37]. Our results, however, reveal the presence of analytical data, while ‘external surface’ refers to XPS data. carboxylate groups, which may reprotonate and thus Fig. 4 presents a plot of the oxygen mole fraction at the possess a basic character. Fig. 6b reveals that, at the external surface (XPS, Table 2) as a function of the external surface of post-treated fibres, about 30 of 100 oxygen mole fraction in the whole material (chemical oxygen atoms are in the form of carboxylate; this would analysis, Table 3). It shows that, for untreated fibres or represent 15 carboxylate functions, identified above as fibres heated at 2300 8C, the oxygen concentration at the calcium carboxylate formed upon rinsing in water. external surface is much higher than the whole oxygen In contrast to oxygen-containing functions, nitrogen is concentration. In contrast, the increase of the oxygen less concentrated at the external surface (Table 2) com￾concentration due to oxidation is only about 1.5 times pared with the whole material (Table 3). The available higher at the external surface compared with the whole data do not allow an evaluation of the respective amounts material. of nitrogen present in the form of acidic, basic or neutral Figs. 5a and 6a present plots of acidity and basicity as a (amide) functions. However, the possible concentration of function of the oxygen concentration, using comparable any of these functions at the external surface is appreciably concentration units. In order to convert 1 g of fibre into less than the concentration of carboxyl and carboxylate

1912 E. Pamula, P.G. Rouxhet Carbon 41(2003)1905-1915 a 8×2 0.5 8101214 O/(C+O+N), whole(mole % O/C+O+N), whole(mole % 5 to+8o 0 0 0 O/(C+O+N), surface(mole% ON(C+O+N), surface(mole%) Fig. 5.(a) Total acidity measured by titration vs. oxygen con- Fig. 6.(a) Total basicity measured by titration vs. oxygen centration in the whole material. (b)Concentration of oxygen concentration in the whole material.(b)Concentration of oxygen the form of OH of carboxyl or aryl ether(Os33. s)at the surface vs in the form of carboxylate (Os312-Os33s)at the surface vs. the the total oxygen concentration at the surface, as determined by total oxygen concentration at the surface, as determined by XPS XPS:(O)UN;(O)GR;(O)OX;() post-treated fibres(OX+ (●)UN,(o)GR;(口)Ox;(■) post-treated fibres(Ox+0.5, 0.5,Ox+1,OX+3) Ox+1,Ox+3) 43. Fibre texture and reactivity to slit-shaped pores of I or 2 nm thickness, their surface area would be 60 or 30 m /g, respectively. The micro- The BET surface area of GR and post-treated samples pores collapsed as a result of oxidation post-treatment; this (Table 1)is in good agreement with the external surface may be attributed to the greater hydrophilic character of area(0.25-0.30 m /g) computed for model cylinders the micropore walls, developing greater capillary forces characterised by the fibre diameter and a density of 1.78 upon drying. g/cm[40]. The larger surface area measured for UN The variation of the fibre diameter according to the indicates a greater fibre roughness or the presence of pores, sample is small, but seems to be significant in view of the which may be partially removed by heat treatment; a small confidence interval. The smaller value for GR omparison of the adsorption isotherms shows that these compared with UN is attributed to the loss of non-carbon re not micropores. The first oxidation caused the creation elements, as confirmed by a decrease in the oxygen of micropores, presumably due to dissolution of oxidised concentration, along with the removal of pores or extermal compounds. Dissolution of aromatic compounds is indeed roughness. The larger fibre diameter of OX, OX+0.5 and confirmed by the yellow colour of solutions obtained after OX+I compared with UN results from the balance of oxidation [41, 42]. If the pore volume of 0.03 cm/g is due oxygen incorporation, dissolution of oxidation products

1912 E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 Fig. 6. (a) Total basicity measured by titration vs. oxygen Fig. 5. (a) Total acidity measured by titration vs. oxygen con- concentration in the whole material. (b) Concentration of oxygen centration in the whole material. (b) Concentration of oxygen in in the form of carboxylate (O 2O ) at the surface vs. the the form of OH of carboxyl or aryl ether (O ) at the surface vs. 531.2 533.8 533.8 total oxygen concentration at the surface, as determined by XPS: the total oxygen concentration at the surface, as determined by (d) UN; (s) GR; (h) OX; (j) post-treated fibres (OX10.5, XPS: (d) UN; (s) GR; (h) OX; (j) post-treated fibres (OX1 OX11, OX13). 0.5, OX11, OX13). 4 .3. Fibre texture and reactivity to slit-shaped pores of 1 or 2 nm thickness, their surface 2 area would be 60 or 30 m /g, respectively. The micro￾The BET surface area of GR and post-treated samples pores collapsed as a result of oxidation post-treatment; this (Table 1) is in good agreement with the external surface may be attributed to the greater hydrophilic character of 2 area (0.25–0.30 m /g) computed for model cylinders the micropore walls, developing greater capillary forces characterised by the fibre diameter and a density of 1.78 upon drying. 3 g/cm [40]. The larger surface area measured for UN The variation of the fibre diameter according to the indicates a greater fibre roughness or the presence of pores, sample is small, but seems to be significant in view of the which may be partially removed by heat treatment; a small confidence interval. The smaller value for GR comparison of the adsorption isotherms shows that these compared with UN is attributed to the loss of non-carbon are not micropores. The first oxidation caused the creation elements, as confirmed by a decrease in the oxygen of micropores, presumably due to dissolution of oxidised concentration, along with the removal of pores or external compounds. Dissolution of aromatic compounds is indeed roughness. The larger fibre diameter of OX, OX10.5 and confirmed by the yellow colour of solutions obtained after OX11 compared with UN results from the balance of 3 oxidation [41,42]. If the pore volume of 0.03 cm /g is due oxygen incorporation, dissolution of oxidation products

E. Pamua, P G. Rouxhet /Carbon 41(2003)1905-1915 913 and pore collapse. The dissolution of degradation products compared with the whole fibre Oxidation does not mark would dominate in the case of ox+3. which shows a fibre edly change the nitrogen concentration, but modifies the diameter similar to that of the fibres treated at 2300 oC ature of the nitrogen-containing functions, as evidenced at the fibre surface by XPS analysis: pyridine-type nitroge and quaternary nitrogen are converted into nitrogen in the surface more hydrophilic due to the creation of function- form of aliphatic functions(e.g. amide or amine). The alities containing oxygen. Oxidation increases the cosine small concentration of nitro groups results from oxidation. of both advanced and receding angles by a factor of about Three kinds of oxygen species are revealed by the com 2. The observed hysteresis is in agreement with recent ponents of the O,s peak(531.2, 532.6 and 533. 8 eV). The investigations of AKZo Tenax carbon fibres [37] and may acidic and basic groups seem to be mainly carboxyl and be tentatively explained in terms of surface roughness calcium carboxylate, respectively. For the whole set of or/and surface heterogeneity [43]. According to Johnson fibres, the fraction of oxygen present in acidic groups and Dettre's theory [44 the advancing angle is controlled(carboxyl) is only slightly higher at the external surface by the low-energy(hydrophobic) part of the surface and compared with the whole fibre. For highly oxidised fibres. the receding contact angle is determined by the high- most basic sites are produced by the conversion of energy(hydrophilic)part carboxyl to calcium carboxy late; the fraction of oxygen is Penetration of oxidising reagents into the fibres may be about twice higher at the external surface compared with due to the presence of micropores or to the formation of the whole fibre; these basic groups are thought to be micropores by progressive dissolution of oxidised com- formed by carboxyl deprotonation and calcium pick-up unds. Some micropores were indeed detected in the ox from rinsing water. ample, but not in fibres submitted to more extensive While the whole material constituting extensively oxidation. This indicates that micropores created by oxida- dised fibres is accessible to acids and bases in tion close up upon drying, however, the internal material solutions, the fibres develop a small surface area and remains accessible to the acidic and basic solutions used not microporous, as shown by krypton adsorption. They for titration and is thus swollen by water. Note that the may thus be considered as sponges which are able to swell access of micropores of type Il Thornel T-300 carbon in contact with water, the internal void volume collapsing fibres was shown to depend on the nature of the gas used upon drying [39]. They may thus find interesting applica- for adsorption [45], which may be attributed to the tions as ion exchangers to be used as fibres, braids, mats, swelling power of the adsorbate: the apparent surface area threads or tissues. While type Ill carbon fibres can be measured by CO, adsorption increased from 1 to 132 transformed into such a material with nitric acid treatment. m/g with increasing extent of electrochemical oxidation type Il fibres seem to require more aggressive treatment of the fibres, whereas the surface area measured by N such as electrochemical oxidation dsorption remained close to 1 m'/ Similar behaviour--small surface area, titration of inter nal acid groups via swellinghas already been reported Acknowledgements for electrochemically oxidised type II fibres 39]: it was suggested that the material might be represented as a The authors thank Prof. P. Grange for the texture Interestingly, oxidation of these fibres with 70% nitric acid measurements,Y. Adriansen and S. Derclaye for their help and prof. s. blazewicz and m.. genet for fruitful discus- at 115"C produced a similar increase in oxygen con- sions. The support of the Belgian Foundation for Scientific entration at the fibre external surface, but did not lead to Research and the Belgian Office for Scientific, Technical such an extensive increase of acidic function concentration in the whole material [40] and Cultural Affairs (Interuniversity Poles of Attraction Programme: Programme of Collaboration with Central and Eastern Europe) is gratefully acknowledged 5. Conclusions The above results show that, upon treatment nitrIc References d, PAN-based fibres carbonised at 1250C are subject to textural changes and profound chemical modifications at [1 Bonnet nsal RC. editors. Carbon fibers. 2nd ed.. New the external surface(XPS data), and also in the whole Dekker,1990,pp.145-260 material (infrared spectra, elemental analysis, acid/base 2 Carnet ang TK. Rebouillat S. Peng JCM. editors fibers, 3rd ed, New York: Marcel Dekker; 1998, pp The oxygen concentration is higher at the fibre external 3] Pittman Jr. CU, Jiang W, Gardner S, Wang L, Toghiani H, urface compared with the whole fibre. In contrast, the Leon y Leon CA. Surface properties of electrochemically nitrogen concentration is lower at the external surface oxidized carbon fibres. Carbon 1999 37: 1797-807

E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 1913 and pore collapse. The dissolution of degradation products compared with the whole fibre. Oxidation does not mark￾would dominate in the case of OX13, which shows a fibre edly change the nitrogen concentration, but modifies the diameter similar to that of the fibres treated at 2300 8C nature of the nitrogen-containing functions, as evidenced at (GR). the fibre surface by XPS analysis: pyridine-type nitrogen The oxidation of carbon fibres makes the external and quaternary nitrogen are converted into nitrogen in the surface more hydrophilic due to the creation of function- form of aliphatic functions (e.g. amide or amine). The alities containing oxygen. Oxidation increases the cosine small concentration of nitro groups results from oxidation. of both advanced and receding angles by a factor of about Three kinds of oxygen species are revealed by the com- 2. The observed hysteresis is in agreement with recent ponents of the O peak (531.2, 532.6 and 533.8 eV). The 1s investigations of AKZO Tenax carbon fibres [37] and may acidic and basic groups seem to be mainly carboxyl and be tentatively explained in terms of surface roughness calcium carboxylate, respectively. For the whole set of or/and surface heterogeneity [43]. According to Johnson fibres, the fraction of oxygen present in acidic groups and Dettre’s theory [44] the advancing angle is controlled (carboxyl) is only slightly higher at the external surface by the low-energy (hydrophobic) part of the surface and compared with the whole fibre. For highly oxidised fibres, the receding contact angle is determined by the high- most basic sites are produced by the conversion of energy (hydrophilic) part. carboxyl to calcium carboxylate; the fraction of oxygen is Penetration of oxidising reagents into the fibres may be about twice higher at the external surface compared with due to the presence of micropores or to the formation of the whole fibre; these basic groups are thought to be micropores by progressive dissolution of oxidised com- formed by carboxyl deprotonation and calcium pick-up pounds. Some micropores were indeed detected in the OX from rinsing water. sample, but not in fibres submitted to more extensive While the whole material constituting extensively oxi￾oxidation. This indicates that micropores created by oxida- dised fibres is accessible to acids and bases in aqueous tion close up upon drying; however, the internal material solutions, the fibres develop a small surface area and are remains accessible to the acidic and basic solutions used not microporous, as shown by krypton adsorption. They for titration and is thus swollen by water. Note that the may thus be considered as sponges which are able to swell access of micropores of type II Thornel T-300 carbon in contact with water, the internal void volume collapsing fibres was shown to depend on the nature of the gas used upon drying [39]. They may thus find interesting applica￾for adsorption [45], which may be attributed to the tions as ion exchangers to be used as fibres, braids, mats, swelling power of the adsorbate: the apparent surface area threads or tissues. While type III carbon fibres can be measured by CO adsorption increased from 1 to 132 transformed into such a material with nitric acid treatment, 2 2 m /g with increasing extent of electrochemical oxidation type II fibres seem to require more aggressive treatment of the fibres, whereas the surface area measured by N such as electrochemical oxidation. 2 2 adsorption remained close to 1 m /g. Similar behaviour—small surface area, titration of inter￾nal acid groups via swelling—has already been reported Acknowledgements for electrochemically oxidised type II fibres [39]; it was suggested that the material might be represented as a The authors thank Prof. P. Grange for the texture sponge, collapsed in the dry state, but able to swell. measurements, Y. Adriansen and S. Derclaye for their help Interestingly, oxidation of these fibres with 70% nitric acid and Prof. S. Blazewicz and M.J. Genet for fruitful discus￾at 115 8C produced a similar increase in oxygen con- sions. The support of the Belgian Foundation for Scientific centration at the fibre external surface, but did not lead to Research and the Belgian Office for Scientific, Technical such an extensive increase of acidic function concentration and Cultural Affairs (Interuniversity Poles of Attraction in the whole material [40]. Programme; Programme of Collaboration with Central and Eastern Europe) is gratefully acknowledged. 5. Conclusions References The above results show that, upon treatment with nitric acid, PAN-based fibres carbonised at 1250 8C are subject to [1] D onnet JB, Bansal RC, editors, Carbon fibers, 2nd ed., New textural changes and profound chemical modifications at York: Marcel Dekker; 1990, pp. 145–260. the external surface (XPS data), and also in the whole [2] D onnet JB, Wang TK, Rebouillat S, Peng JCM, editors, material (infrared spectra, elemental analysis, acid/base Carbon fibers, 3rd ed., New York: Marcel Dekker; 1998, pp. titrations). 161–229. The oxygen concentration is higher at the fibre external [3] P ittman Jr. CU, Jiang W, Gardner S, Wang L, Toghiani H, surface compared with the whole fibre. In contrast, the Leon y Leon CA. Surface properties of electrochemically nitrogen concentration is lower at the external surface oxidized carbon fibres. Carbon 1999;37:1797–807

1914 E. Pamua, P G. Rouxhet /Carbon 41(2003)1905-1915 [4] Alexander MR, Jones FR. Effect of electrolytic oxidation on characterization of activated carbons with oxygen and nitro- the surface chemistry of type A carbon fibres-Part I. X-ray gen surface groups. Carbon 1997; 35(12): 1799-810 hotoelectron spectroscopy. Carbon 1994; 32(5): 785-94. 24] Beamson G, Briggs D, editors, High resolution XPS of [5] Alexander MR, Jones FR. Effect of electrolytic oxidation rganic polymers: the Scienta ESCA A300 database, Chi- Opon the surface chemistry of type A carbon fibres. Ill chester. Wiley: 1992, pp. 98-136 Chemical state, source and location of surface nitrogen. [25] Nakajima T,Koh M. Synthesis of high crystalline carbon- Carbon1996;34(9):1093-102 nitrogen layered compounds by CVD using nickel and cobalt [6]Dryzek J, Pamula E, Blazewicz S. Positron annihilation in catalysts. Carbon 1997: 35(2 ): 203-8 carbon fibers. Phys Status Solidi (a)1995: 151: 39-46 26] Pels JR, Kaptein F, Moulin JA, Zhu Q, Thomas KM. [7 Blazewicz S, Chlopek J, Blazewicz M, Wailer C. Carbon Evolution of nitrogen functionalities in carbonaceous materi- materials. In: Middleton J, Pande gn, Williams KR, als during pyrolysis. Carbon 1995: 33(11): 1641-53 omputer methods in biome 27] Stanczyk K, Dziembaj R, Piwowarska Z, Witkowski S anics and biomedical engi tional; 1992, pp. 364-2neering, Books Journal Intema- Transformation of nitrogen structures in carbonization of model compounds determined by XPS. Carbon 1995;33(10:1383-92. implant behaviour of chemically-modified carbon braids. [28]Koh M, Nakajima T. Adsorption of aromatic compounds Ceram Int1994202)99-103 C N-coated activated carbon. Carbon 2000: 38: 1947-54 9】 Blazewicz S, Wailer C, Chlopek J. Static and dynamic 29OReilly JM, Mosher RA. Functional groups in carbon black of carbon ligament prosthesis. J Biomed by FTIR spectroscopy. Carbon 1983; 21(1): 47-51 Mater Res1996;32(2:2l5-9 30 Zawadzki J IR spectroscopic investigations of mechanism of [10] Blazewicz M, Blazewicz S, Wailer C. Carbon materials in oxidation of carbonaceous films with HNO, solution. Carbon medicine. In: Vincenzini P. editor. Innovative materials 1980,18(4):281-5. Faenza: Techna Srl; 1993, pp. 595-602. B1Hontoria-Lucas C, Lopez-Peinado A, Lopez-Gonzalez JD, [11] Polish Patent P-282577, Sposob otrzymywania weglowych Rojas-Cervantes ML, Marti-Aranda RM. Study of oxygen- aterialow protetycznych i szewnych, 1993 graphite oxides: phy [12] Gerin PA, Dengis PB, Rouxhet PG. Performance of XPS chemical characterization. Carbon 1995: 33(11): 1585-92 sis of model biochemical compounds. J Chim Phys B2]Zawadzki J IR spectroscopy investigations of acidic charac. 1995:92:1043-65 ter of carbonaceous films oxidized with HNo, solution. [13 Gregg SJ, Sing ksw. Adsor rface area and porosity, Carbon I981;191):19-25 2nd ed.. London: Academic 33] Starsinic M, Taylor RL, Walker PL, Painter PC. FTIR studies [14 Gas Encyclopaedia. L'Air Amsterdam: Elsevier of Saran chars. Carbon 1983; 21(1): 69-74 1976:1005-1009 34]Zawadzki J. An infrared study of acid sites on carbons by [15] Kozlowski C, Sherwood PMA. X-ray photoelectron spectro- pyridine adsorption. Carbon 1988: 26(5): 627-33 scopic studies of carbon fiber surfaces. Vill. A comparison of B5] Rouxhet PG, Robin PL, Nicaise G. Characterization of type I and lI fibers and their interaction with thin resin films kerogens and their evolution by infrared spectroscopy. In: Carbon1987;25(6:75l-60. Durand B, editor, Paris: Kerogen. Technip, 1980, pp. 163- 0, chapter 6 Jr. CU. Surface characterization of carbon fibres using 36 Otake Y, Jenkins RG. Characterization of oxygen-containing resolved XPS and ISS. Carbon 1995: 33(5): 587-95 surface complexes created on a microporous carbon by [17 Polovina M, Babic B, Kaluderowic B, Dekanski A Surface and nitric acid treatment. Carbon 1993: 31(1): 109-21 naracterization of oxidised activated carbon cloth. Carbon 37] Vickers PE, Watts JF, Perruchot C, Chehimi MM. The surface chemistry and acid-base properties of a PAN-based [18] Bradley RH, Ling X, Sutherland J. An investigation of carbon fibre Carbon 2000- 38: 675-89 bon fibre surface chemistry 38 Jansen RJ, van Bekkum H. XPS of nitrogen-containing energy functional groups on activated carbon. Carbon [19xie Y, Wang TJ, Franklin O A. X-ray photo- 1995;33(8):1021-7. 9]Waseem SF, Gardner SD, Guoren He, Jiang w, Pittman Jr U Spectrosc199246(4)645-51 dhesion and surface analysis of carbon fibres electro- chemically oxidized in aqueous potassium nitrate. J Mater Sci1998;33(12):3151-62. Carbon [40] Pittman Jr. CI r wu b, gardner SD. Chemical modification of carbon fiber surfaces by nitric acid oxidation i2 caree, bertine. k chenmicman modif followed by reaction with tetraethylenepentamine. Carbon 199735(3):317-31 [41]Ehrburger P, Donnet J-B. Study of the oxidation of carbons [22] Huttinger wK, Hohmann-Wien S, Seiferling M. A to graphitic oxides. Carbon 1973; 11(4)309-16 essile drop technique for the determination of the tions at carbon-liquid interfaces. Carbon 1991; 29( [23 Biniak S, Szymanski G, Siedlewski J, Swiatkowski A. The [43] Andrade JD, Smith LM, Gregonis dE. The contact angle and

1914 E. Pamula, P.G. Rouxhet / Carbon 41 (2003) 1905–1915 [4] A lexander MR, Jones FR. Effect of electrolytic oxidation on characterization of activated carbons with oxygen and nitro￾the surface chemistry of type A carbon fibres—Part I. X-ray gen surface groups. Carbon 1997;35(12):1799–810. photoelectron spectroscopy. Carbon 1994;32(5):785–94. [24] B eamson G, Briggs D, editors, High resolution XPS of [5] A lexander MR, Jones FR. Effect of electrolytic oxidation organic polymers: the Scienta ESCA A300 database, Chi￾upon the surface chemistry of type A carbon fibres. III. chester: Wiley; 1992, pp. 98–136. Chemical state, source and location of surface nitrogen. [25] N akajima T, Koh M. Synthesis of high crystalline carbon– Carbon 1996;34(9):1093–102. nitrogen layered compounds by CVD using nickel and cobalt [6] D catalysts. Carbon 1997;35(2):203–8. ryzek J, Pamula E, Blazewicz S. Positron annihilation in carbon fibers. Phys Status Solidi (a) 1995;151:39–46. [26] P els JR, Kapteijn F, Moulijn JA, Zhu Q, Thomas KM. Evolution of nitrogen functionalities in carbonaceous materi- [7] B lazewicz S, Chlopek J, Blazewicz M, Wajler C. Carbon biomaterials. In: Middleton J, Pande GN, Williams KR, als during pyrolysis. Carbon 1995;33(11):1641–53. editors, Recent advances in computer methods in biomech- [27] S tanczyk K, Dziembaj R, Piwowarska Z, Witkowski S. Transformation of nitrogen structures in carbonization of anics and biomedical engineering, Books & Journal Interna￾model compounds determined by XPS. Carbon tional; 1992, pp. 364–72. 1995;33(10):1383–92. [8] B lazewicz M, Blazewicz S, Wajler C. Mechanical and [28] K oh M, Nakajima T. Adsorption of aromatic compounds on implant behaviour of chemically-modified carbon braids. C N-coated activated carbon. Carbon 2000;38:1947–54. Ceram Int 1994;20(2):99–103. x [29] O ’Reilly JM, Mosher RA. Functional groups in carbon black [9] B lazewicz S, Wajler C, Chlopek J. Static and dynamic by FTIR spectroscopy. Carbon 1983;21(1):47–51. fatigue properties of carbon ligament prosthesis. J Biomed Mater Res 1996;32(2):215–9. [30] Z awadzki J. IR spectroscopic investigations of mechanism of [10] B oxidation of carbonaceous films with HNO solution. Carbon 3 lazewicz M, Blazewicz S, Wajler C. Carbon materials in 1980;18(4):281–5. medicine. In: Vincenzini P, editor, Innovative materials, [31] H ontoria-Lucas C, Lopez-Peinado AJ, Lopez-Gonzalez JD, Faenza: Techna Srl; 1993, pp. 595–602. Rojas-Cervantes ML, Marti-Aranda RM. Study of oxygen- [11] P olish Patent P-282577, Sposob otrzymywania weglowych materialow protetycznych i szewnych, 1993. containing groups in a series of graphite oxides: physical and chemical characterization. Carbon 1995;33(11):1585–92. [12] G erin PA, Dengis PB, Rouxhet PG. Performance of XPS [32] Z awadzki J. IR spectroscopy investigations of acidic charac- analysis of model biochemical compounds. J Chim Phys ter of carbonaceous films oxidized with HNO solution. 1995;92:1043–65. 3 Carbon 1981;19(1):19–25. [13] G regg SJ, Sing KSW. Adsorption, surface area and porosity, [33] S tarsinic M, Taylor RL, Walker PL, Painter PC. FTIR studies 2nd ed.. London: Academic Press; 1982. of Saran chars. Carbon 1983;21(1):69–74. [14] G as Encyclopaedia. L’Air Liquide. Amsterdam: Elsevier, [34] Z awadzki J. An infrared study of acid sites on carbons by 1976:1005–1009. pyridine adsorption. Carbon 1988;26(5):627–33. [15] K ozlowski C, Sherwood PMA. X-ray photoelectron spectro- [35] R ouxhet PG, Robin PL, Nicaise G. Characterization of scopic studies of carbon fiber surfaces.VIII. A comparison of kerogens and their evolution by infrared spectroscopy. In: type I and II fibers and their interaction with thin resin films. Carbon 1987;25(6):751–60. Durand B, editor, Paris: Kerogen. Technip; 1980, pp. 163– 90, chapter 6. [16] G ardner SD, Singamsetty CSK, Booth GL, He G-R, Pittman Jr. CU. Surface characterization of carbon fibres using angle- [36] O take Y, Jenkins RG. Characterization of oxygen-containing surface complexes created on a microporous carbon by air resolved XPS and ISS. Carbon 1995;33(5):587–95. and nitric acid treatment. Carbon 1993;31(1):109–21. [17] P olovina M, Babic B, Kaluderowic B, Dekanski A. Surface characterization of oxidised activated carbon cloth. Carbon [37] V ickers PE, Watts JF, Perruchot C, Chehimi MM. The 1997;35(8):1047–52. surface chemistry and acid–base properties of a PAN-based carbon fibre. Carbon 2000;38:675–89. [18] B radley RH, Ling X, Sutherland J. An investigation of carbon fibre surface chemistry and reactivity based on XPS [38] J ansen RJJ, van Bekkum H. XPS of nitrogen-containing and surface free energy. Carbon 1993;31(7):1115–20. functional groups on activated carbon. Carbon [19] X 1995;33(8):1021–7. ie Y, Wang TJ, Franklin O, Sherwood PMA. X-ray photo￾electron spectroscopic studies of carbon-fibre surfaces. Appl [39] W aseem SF, Gardner SD, Guoren He, Jiang W, Pittman Jr. U. Spectrosc 1992;46(4):645–51. Adhesion and surface analysis of carbon fibres electro- [20] Y oshida A, Tanehashi I, Nishino A. Effect of concentration chemically oxidized in aqueous potassium nitrate. J Mater of surface acidic functional groups on electric double-layer Sci 1998;33(12):3151–62. properties of activated carbon fibers. Carbon [40] P ittman Jr. CU, He G-R, Wu B, Gardner SD. Chemical 1990;28(5):611–5. modification of carbon fiber surfaces by nitric acid oxidation [21] Z ielke U, Huttinger KJ, Hoffman WP. Surface oxidized followed by reaction with tetraethylenepentamine. Carbon carbon fibers. 2. Chemical modification. Carbon 1997;35(3):317–31. 1996;34(8):999–1005. [41] E hrburger P, Donnet J-B. Study of the oxidation of carbons [22] H uttinger WKJ, Hohmann-Wien S, Seiferling M. A modified to graphitic oxides. Carbon 1973;11(4):309–16. sessile drop technique for the determination of the interac- [42] S osa RC, Masy D, Rouxhet PG. Influence of surface tions at carbon–liquid interfaces. Carbon 1991;29(3):449– properties of carbon black on the activity of adsorbed 55. catalase. Carbon 1994;32(7):1369–75. [23] B iniak S, Szymanski G, Siedlewski J, Swiatkowski A. The [43] A ndrade JD, Smith LM, Gregonis DE. The contact angle and