Z.R. Yue et al. /Carbon 37(1999)1785-1796 each of which contained chemical oxidation was defined in terms of coulombs approximately 3000 1 vere cut from the carbon fiber (AXn) per gram(C/g) tow and positioned of a custom stainless steel sample holder. The vere held firmly in place by a gold foil mask secured to the sample holder with screws 3.1.1. Weight loss of carbon fibers The gold foil contained a machined oval opening in its There was a continual loss of weight of the carbon fibers center that exposed a 1.5 cm by 0.8 cm area of the as the extent of oxidation increased. The weight loss was underlying carbon fibers KPS experiments were performed on a Physical Elec- oxidation(Fig. 1)from the onset of oxidation up to 4000 tronics PHI Model 1600 surface analysis system equipped C/g. At this point 17-18% of the initial fiber weight had with a PHI 10-360 spherical capacitor energy analyzer been lost. a slower loss of weight occurred as the extent of (SCA)fitted with an Omni Focus Ill small-area lens(800 electrochemical oxidation increased from 4000 to 8000 um diameter analysis area)and a high-performance multi- C/g At 8000 C/g 20-21% of the mass was gone. Then a channel detector. Samples were oriented such that the axial arp increase in the extent of weight loss occurred with direction of carbon fibers was in the plane of the X-ray continued oxidation over 8000 C/g. A 30% weight loss angle was at 30 Progressive weight loss occurs with CO, evolution. KPS spectra were obtained using an achromatic Mg k Active site atoms on the fiber surface were oxidized to (1253.6 eV)X-ray source operated at 200 W. Survey scans form such oxygen-containing surface groups as C-Oh were collected from 0-1100 ev with a pass energy equal to C=O, COOH and finally CO,. The types of oxygen 46.95 eV. High-resolution scans were performed with the functions and the simplified step-wise progression mecha- ass energy adjusted to 23.50 ev. The vacuum system nism for carbon surface oxidation in Eq (1) have been pressure was maintained at approximately 10 Torr widely studied [54-59 during all XPS experiments A non-linear least squares curve fitting program XPSPEAK95 software, Version 2.0) with a Gaussian- Lorentzian mix function and Shirley background subtrac- tion was used to deconvolve the xPs peaks. The Lorentz Gaussian mix was 60%. The carbon Is electron binding energy corresponding to graphitic carbon was referenced at 284.6 ev for calibration [52]. The peak constraints for fitting were used. All the higher energy C Is peaks fitted were shifted to higher binding energies by about 1.55, 3.0, 4.0 and 6. 1 ev, respectively. Atomic ratios were calculated from the XPs spectra after correcting the relative peak areas by sensitivity factors based on the transmissio characteristics of the Physical Electronics SCA [53 25. Fourier transform infrared spectroscopy 20 FT-IR spectroscopy was used for analyzing functional groups formed on the electrochemically oxidized carbon fibers. Treated fibers were cut and mixed with KBr. the mixture was analyzed with a Bruker Instruments Inc odel IFS 25 FT-IR spectrometer. 3. Results and discussion 3.I. Influence of the extent of electrochemical oxidation 20004000600080001000012000 High strength PAN-based carbon fibers were continuous- ly electrochemically oxidized by applying current (A)for Extent of electro-oxidation(C/g) specific residence times(n)to the fibers which served as an Fig. 1. eight loss of carbon fiber as a function of the extent of anode in 1% wt KNo solution. The extent of electro- electrochemical oxidation.Z.R. Yue et al. / Carbon 37 (1999) 1785 –1796 1787 Several 2.5 cm sections (each of which contained chemical oxidation was defined in terms of Coulombs approximately 3000 fibers) were cut from the carbon fiber (A3t) per gram (C/g). tow and positioned on top of a custom stainless steel sample holder. The fibers were held firmly in place by a 3.1.1. Weight loss of carbon fibers gold foil mask secured to the sample holder with screws. There was a continual loss of weight of the carbon fibers The gold foil contained a machined oval opening in its as the extent of oxidation increased. The weight loss was center that exposed a 1.5 cm by 0.8 cm area of the approximately proportional to the extent of electrochemical underlying carbon fibers. XPS experiments were performed on a Physical Elec- oxidation (Fig. 1) from the onset of oxidation up to 4000 C/g. At this point 17–18% of the initial fiber weight had tronics PHI Model 1600 surface analysis system equipped with a PHI 10-360 spherical capacitor energy analyzer been lost. A slower loss of weight occurred as the extent of electrochemical oxidation increased from 4000 to 8000 (SCA) fitted with an Omni Focus III small-area lens (800 mm diameter analysis area) and a high-performance multi- C/g. At 8000 C/g 20–21% of the mass was gone. Then a sharp increase in the extent of weight loss occurred with channel detector. Samples were oriented such that the axial continued oxidation over 8000 C/g. A 30% weight loss direction of carbon fibers was in the plane of the X-ray had occurred at about 10 600 C/g. source and the analyzer detection slit. The electron take-off Progressive weight loss occurs with CO evolution. angle was at 308. 2 XPS spectra were obtained using an achromatic Mg K Active site atoms on the fiber surface were oxidized to a form such oxygen-containing surface groups as C–OH, (1253.6 eV) X-ray source operated at 200 W. Survey scans were collected from 0–1100 eV with a pass energy equal to C5O, COOH and finally CO . The types of oxygen 2 functions and the simplified step-wise progression mecha- 46.95 eV. High-resolution scans were performed with the nism for carbon surface oxidation in Eq. (1) have been pass energy adjusted to 23.50 eV. The vacuum system 29 widely studied [54–59]. pressure was maintained at approximately 10 Torr during all XPS experiments. A non-linear least squares curve fitting program (XPSPEAK95 software, Version 2.0) with a Gaussian￾Lorentzian mix function and Shirley background subtrac￾tion was used to deconvolve the XPS peaks. The Lorentz/ Gaussian mix was 60%. The carbon 1s electron binding (1) energy corresponding to graphitic carbon was referenced at 284.6 eV for calibration [52]. The peak constraints for fitting were used. All the higher energy C 1s peaks fitted were shifted to higher binding energies by about 1.55, 3.0, 4.0 and 6.1 eV, respectively. Atomic ratios were calculated from the XPS spectra after correcting the relative peak areas by sensitivity factors based on the transmission characteristics of the Physical Electronics SCA [53]. 2.5. Fourier transform infrared spectroscopy FT-IR spectroscopy was used for analyzing functional groups formed on the electrochemically oxidized carbon fibers. Treated fibers were cut and mixed with KBr. The mixture was analyzed with a Bruker Instruments Inc. Model IFS 25 FT-IR spectrometer. 3. Results and discussion 3.1. Influence of the extent of electrochemical oxidation High strength PAN-based carbon fibers were continuous￾ly electrochemically oxidized by applying current (A) for specific residence times (t) to the fibers which served as an Fig. 1. Weight loss of carbon fiber as a function of the extent of anode in 1% wt KNO solution. The extent of electro- electrochemical oxidation. 3