Characterization of oxygen containing functional groups on carbon materials with oxygen K-edge X-ray absorption near edge structure spectroscopy

CARBON

4 9 ( 2 0 1 1 ) 1 7 4 5 –1 7 5 1

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/carbon

Characterization of oxygen containing functional groups on carbon materials with oxygen K-edge X-ray absorption near edge structure spectroscopy Kyungsoo Kim a b c

a,b

, Pengyu Zhu

a,b

, Na Li

b,c

, Xiaoliang Ma b, Yongsheng Chen

a,b,*

Department of Energy and Mineral Engineering, The Pennsylvania State University, University Park, PA 16802, USA EMS Energy Institute, The Pennsylvania State University, University Park, PA 16802, USA College of Chemistry and Chemical Engineering, China University of Petroleum, Dongying, Shandong 257061, China

A R T I C L E I N F O

A B S T R A C T

Article history:

Surface functional groups on carbon materials are critical to their surface properties and

Received 13 September 2010

related applications. Many characterization techniques have been used to identify and

Accepted 23 December 2010

quantify the surface functional groups, but none is completely satisfactory especially for

Available online 30 December 2010

quantification. In this work, we used oxygen K-edge X-ray absorption near edge structure (XANES) spectroscopy to identify and quantify the oxygen containing surface functional groups on carbon materials. XANES spectra were collected in fluorescence yield mode to minimize charging effect due to poor sample conductivity which can potentially distort XANES spectra. The surface functional groups are grouped into three types, namely carboxyl-type, carbonyl-type, and hydroxyl-type. XANES spectra of the same type are very similar while spectra of different types are significantly different. Two activated carbon samples were analyzed by XANES. The total oxygen contents of the samples were estimated from the edge step of their XANES spectra, and the identity and abundance of different functional groups were determined by fitting of the sample XANES spectrum to a linear combination of spectra of the reference compounds. It is concluded that oxygen K-edge XANES spectroscopy is a reliable characterization technique for the identification and quantification of surface functional groups on carbon materials. Ó 2010 Elsevier Ltd. All rights reserved.

1.

Introduction

Carbon is a versatile material as it forms different structures exhibiting different properties. For example, carbon nanotubes (CNT) are used for field electron emission, energy storage, and atomic force microscopy due to their cylindrical conformation, exceptional mechanical strength, and thermal and electrical conductivity [1,2]. Carbon nanofibers (CNF) can be used for material reinforcement and hydrogen storage for their strong individual structure the and the availability to produce various structures [3,4]. Having unique graphitic

multi-layer structure with hollow center morphology, carbon nano-onions (CNO) are good solid lubricant and possible protector for air sensitive materials [5,6]. Other than the bulk properties of carbon materials, for some applications, the surface properties play a very important role. For example, carbon materials can be used as adsorbent [7–10] and catalyst support [11–13]. In these cases, oxygen containing surface functional groups strongly influence the properties of the carbon materials. Extensive efforts have been devoted to preparing or modifying surface functional groups for different applications [14–19].

* Corresponding author at: Department of Energy and Mineral Engineering, The Pennsylvania State University, University Park, PA 16802, USA. E-mail address: [email protected] (Y. Chen). 0008-6223/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.carbon.2010.12.060

1746

CARBON

4 9 ( 2 0 1 1 ) 1 7 4 5 –1 7 5 1

To better understand how surface functional groups on carbon materials affect their performance, proper identification and quantification of these functional groups are essential. Many characterization methods have been used for this purpose, among them, X-ray photoelectron spectroscopy (XPS), infrared spectroscopy (IR), temperature-programmed desorption (TPD), and titration are the most common techniques. Boehm compared the advantages and disadvantages of these techniques in great details in a review paper [20]. Although all these techniques give useful information about carbon materials and their surface functional groups, each technique has its own limitations, and proper identification and reliable quantification of surface functional groups are still not completely satisfactory. With XPS, quantification usually involves deconvolution of overlapping peaks, which is somewhat arbitrary as peak position, peak shape, and peak width are assumed for different species and the assumptions are difficult to verify [20]. If XPS analysis is performed based on C 1s spectra, a strong interference from the carbon substrate is expected due to dominant C–C species in carbon materials. Moreover, charging of carbon materials, which is not uncommon, may cause peak distortion and add more uncertainties in the analysis. By combining XPS with chemical derivatization, the uncertainties may be reduced [21]. TPD depends on evolved gases such as CO and CO2 from surface functional groups for identification and quantification purposes. However, same functional groups may result in multiple gases and there is also possibility of secondary reactions, which brings uncertainties to the TPD analysis [20]. In addition, TPD analysis also involves deconvolution of overlapping peaks, which is the same problem faced by XPS. X-ray absorption near edge structure (XANES) spectroscopy is a technique that can potentially overcome the limitations faced by the commonly used techniques. XANES is an element specific technique that probes local electronic structure around a specific element of interest. By performing oxygen K-edge XANES spectroscopy, the interference from carbon in the substrate can be avoided [22,23]. In addition, intense synchrotron X-ray radiation provides high intensity which allows the detection of very low level of surface functional groups [24]. Moreover, surface functional groups may have unique XANES spectra, by which qualitative analysis is possible by fingerprinting [25]. Both carbon and oxygen K-edge XANES works have been reported in the literature for characterization of surface functional groups [25–30]. Carbon K-edge XANES spectroscopy has been used to characterize chemical structure of carbon, and the surface functional groups are often identified as minor components of the material. Methods have been developed to quantify different carbon species including deconvolution of overlapping peaks. Uncertainties result from the fact that the surface functional groups in most cases are minor components compared to the dominant contribution from other species in the substrate that do not contain oxygen such as C–C and C–H bonds. To overcome these problems, it seems that oxygen K-edge XANES spectroscopy is a convenient choice since only oxygen containing surface functional groups will be detected and no interference from other species in the substrate is expected. In addition, reference compounds with only one type of functional groups are widely available, thus obtaining reference

spectra of different functional groups would not be an issue. Surprisingly, there are only few reports in the literature on oxygen K-edge XANES works, and they are limited to qualitative analysis [25,30]. The main reason is believed to be the charging effect that distorts the XANES spectra of poorly conducting carbon materials when the spectra are collected in total electron yield (TEY) mode. In the present work, fluorescence yield (FY) mode is used to collect oxygen K-edge XANES spectra, which effectively overcomes the charging effect (see Fig. S1 in the Supplementary data). We obtain XANES spectra of reference compounds for individual functional groups and use linear combination fitting method to identify and quantify the functional groups present in two activated carbon samples. In addition, the difference in absorption before and after the edge (or edge step) is used to determine the total oxygen contents in the activated carbon samples. It is shown that oxygen K-edge XANES spectroscopy is a reliable tool for identifying and quantifying different types of surface functional groups on carbon materials.

2.

Experimental

2.1.

Reference compounds for oxygen functional groups

Oxygen K-edge XANES spectra were collected for carbonyl, aldehyde, ketone, carboxyl, anhydride, ester, ether, and hydroxyl groups. Table 1 lists the reference compounds that were used in this work. The compounds were commercially available aromatic compounds and polymers with functional groups of one type. All the compounds were used as received without any further treatment. The reference compounds of same functional group on different substrates, such as 9-acethylanthracene and PVMK (ketone), 1-naphthoic acid and PAA (carboxylic acid), and poly(ethylene-alt-maleic anhydride) and phthalic anhydride (anhydride) are chosen to investigate the influence of substrate structure on XANES spectrum.

2.2.

Preparation of activated carbons

An ultra fine petroleum coke-based activated carbon with a product name AC4, was purchased from Kansai Coke and Chemicals Company Ltd. To demonstrate that surface functional groups can be created by an oxidizer, AC4 activated carbon was treated by a mixture of 1.5 M ammonium persulfate ((NH4)2S2O8) and 1 M sulfuric acid (H2SO4). First, 1 g of the AC4 activated carbon was mixed with 20 ml of the oxidizer, and heated at 60 °C in a flask for 3 h. The reacted sample was filtered and washed with distilled water, and then dried at 110 °C in a vacuum oven overnight. The two activated carbon samples are referred to as AC4 (untreated activated carbon) and OAC4 (oxidized activated carbon).

2.3.

Oxygen K-edge XANES spectroscopy

Oxygen K-edge (543.1 eV) XANES measurements were performed at Beamline U4B of the National Synchrotron Light Source (NSLS) at Brookhaven National Laboratory, Upton, NY, USA. Detailed information about the experimental setup can be found elsewhere [31]. The storage ring was operated

CARBON

4 9 ( 20 1 1 ) 1 7 4 5–17 5 1

1747

Table 1 – Reference compounds for different functional groups. Compound Anthrone 9-Anthraldehyde 9-Acethylanthracene Poly(vinyl methyl ketone) (PVMK) 1-Naphthoic acid Poly(acrylic acid) (PAA) Poly(vinyl chloride-co-vinyl acetate)(PVCVA) Poly(ethylene-alt-maleic anhydride) Phthalic anhydride Poly(vinyl alcohol) (PVA) Poly(4-vinylphenol) (PVP) Methyl cellulose

Functional group Carbonyl Aldehyde Ketone Ketone Carboxyl Carboxyl Ester Anhydride Anhydride Aliphatic hydroxyl Phenolic hydroxyl Ether

Supplier Alfa Aesar TCI America Alfa Aesar Sigma Aldrich Alfa Aesar Sigma Aldrich Sigma Aldrich Sigma Aldrich Alfa Aesar Sigma Aldrich Sigma Aldrich Sigma Aldrich

with an electron beam energy of 800 MeV and an average current of 600 mA. XANES spectra were collected in both fluorescence yield (FY) and total electron yield (TEY) modes simultaneously. Reference spectrum of a material containing V, O, and Cr was also collected in every scan for energy alignment. The energy range was from 510 to 590 eV. The energy interval was 0.1 eV in the critical region (from 520 to 560 eV) and 0.5 eV in the other region. Samples were pressed onto a copper tape uniformly, and no cracks were observed. The samples were about tenths of a millimeter thick so no interference from the adhesive on the copper tape was expected. The samples were placed in an ultra high vacuum chamber maintained at a pressure lower than 108 Pa. The XANES spectra were processed, and quantitative analysis was performed by linear combination fitting using Athena [32].

3.

Results and discussion

3.1.

XANES spectra of individual functional groups

XANES spectra of the reference compounds containing different functional groups are shown in Figs. 1–3. All spectra were collected in fluorescence yield (FY) mode, and were rescaled

Fig. 1 – Oxygen K-edge XANES spectra of carboxyl-type functional groups: ester (PVCVA), carboxylic acid (a) (PAA), carboxylic acid (b) (naphthoic acid), anhydride (a) (phthalic anhydride), and anhydride (b) (poly(ethylene-alt-maleic anhydride)).

Fig. 2 – Oxygen K-edge XANES spectra of carbonyl-type functional groups: carbonyl (anthrone), aldehyde (anthraldehyde), ketone (a) (9-acethylanthracene), and ketone (b) (PVMK).

for better comparison. As a general observation, oxygen K-edge spectra consist of pre-edge feature (sharp peak around 530 eV) and broad whiteline (first intense post-edge peak, near 540 eV in this case). Some functional groups have very similar XANES spectra, and are very difficult to discriminate, while some are very different and can be much easier to distinguish. For this reason, functional groups are grouped into three types. Carboxyl-type functional groups include carboxylic acid (PAA and naphthoic acid), anhydride (phthalic anhydride and poly(ethylene-alt-maleic anhydride)), and ester (PVCVA). As shown in Fig. 1, the carboxyl-type functional groups show a strong pre-edge peak at about 531 eV and a broad whiteline at 539 eV. They also have a small second pre-edge peak at about 533.5 eV. Carbonyl-type functional group includes carbonyl (anthrone), aldehyde (anthraldehyde) and ketone (9-acethylanthracene and PVMK). Similar to carboxyl-type functional groups, carbonyl-type functional groups have a strong pre-edge peak below 530 eV and a broad whiteline at 538–544 eV, as shown in Fig. 2. They do not have a small second pre-edge peak at about 533.5 eV, however. Hydroxyl-type functional groups include phenolic hydroxyl (PVP), aliphatic hydroxyl (PVA), and ether (methyl cellulose). In contrast to carboxyl-type and carbonyl-type functional

1748

CARBON

4 9 ( 2 0 1 1 ) 1 7 4 5 –1 7 5 1

Fig. 3 – Oxygen K-edge XANES spectra of hydroxyl-type functional groups: phenolic hydroxyl (PVP), aliphatic hydroxyl (PVA), and ether (methyl cellulose). groups, the hydroxyl-type functional groups possess no or very weak pre-edge peak as shown in Fig. 3. Aliphatic hydroxyl and ether have a whiteline at 537 eV while phenolic hydroxyl has a whiteline at 538.5 eV and a shoulder at 534.5 eV. The difference among different types of functional groups is significant. Fig. 4 presents representative normalized XANES spectra of carboxyl-type, hydroxyl-type, and carbonyl-type functional groups, which manifests the differences among them. The main difference lies in the intensity and position of the pre-edge peak. Both carboxyl and carbonyl groups possess intense pre-edge peak with big separation in peak position of greater than 1 eV. In contrast, hydroxyl has negligible pre-edge peak and probably the most intense whiteline. The whiteline appears in the lowest energy among all the examined functional groups. Evidently, these three types of functional groups can be easily distinguished.

3.2. Effect of substrate on XANES spectrum and fitting strategy Functional groups of a same type have very similar XANES spectra as shown in Figs. 1–3. Oxygen K-edge XANES probes

Fig. 4 – Normalized oxygen K-edge XANES spectra of representative carboxyl-type (PAA), hydroxyl-type (PVA), and carbonyl-type (anthrone) functional groups.

local electronic structure around oxygen, which is a direct result of the coordination structure around oxygen. The oxygen in the functional groups binds to carbon through single or double bonds and may also bind to hydrogen. The hydrogen has very similar effect to the oxygen electronic structure when compared to single-bonded carbon, as shown in the spectra of ester and carboxyl groups (Fig. 1), aldehyde and ketone (Fig. 2), or ether and aliphatic hydroxyl (Fig. 3). The nature of the carbon (aliphatic or aromatic) that binds directly to oxygen may also affect the XANES spectrum to some extent, for example, phenolic and aliphatic hydroxyls as shown in Fig. 3. Moreover, the functional groups bind to the carbon substrate through C– C bond. The nature of the carbon from the substrate (or chemical environment) can also have secondary effect on the XANES spectrum. For instance, carboxyl groups are present in both poly(acrylic acid) and naphthoic acid, and the substrates are aliphatic and aromatic carbons, respectively. Although their spectra are very similar, some small differences are discernible, such as slight difference in pre-edge peak intensity and peak position, as shown in Fig. 1. The similarities are also observed for anhydride and ketone (see Figs. 1 and 2). Since XANES spectra of different types of functional groups are significantly different, it is possible to quantify them by principle component analysis or linear combination fitting. As discussed before, functional groups of a same type have similar spectra with slightly different pre-edge peak position, pre-edge peak height, whiteline position and whiteline intensity. One potential problem is the selection of fitting base sets for different types of functional groups, i.e., how many reference spectra should be used in the fitting. Including all possible reference spectra is not feasible since for any sample spectrum to be fitted there is only limited data available. Using too many principle components would inevitably face convergence problem. The problem can be overcome effectively and efficiently by using two reference spectra having the most difference for each of the three types of functional groups and inclusion of a small energy shift in the fitting process. We have found that using two reference compounds with the smallest and the biggest pre-edge peak heights for carboxyl-type and carbonyl-type functional groups and introducing a small energy shift to account for the chemical environment, any other reference compound of the same type can be reasonably well fitted (total abundance automatically comes very close to 1). The reference compounds used in the fitting are anthrone and anthraldehyde for carbonyl-type, naphthoic acid and phthalic anhydride for carboxyl-type functional groups. The energy shift is limited to about 0.6 eV or less to ensure no crossover between these two types. For hydroxyl-type functional groups, aliphatic hydroxyl and phenolic hydroxyl are used, energy shift is limited to 2 eV or less to account for the whiteline shift due to different chemical environment. Before performing the fitting, all spectra, including those of the samples and the reference compounds, are normalized to a unity edge step to compensate for the different oxygen contents in the materials. Fitting range includes both the pre-edge and whiteline regions. One criterion to determine fitting goodness is an R-factor generated in Athena [32]. The R-factor is defined as follows,

CARBON

R-factor ¼

P ðdata  fitÞ2 ; P ðdataÞ2

4 9 ( 20 1 1 ) 1 7 4 5–17 5 1

1749

ð1Þ

where the summations are over the data points in the fitting range. Thus, a smaller R-factor usually suggests a better fit.

3.3.

Functional groups in the activated carbons

The XANES spectra of AC4 and OAC4 before normalization are shown in Fig. 5. The absorption of oxygen in OAC4 is significantly higher than that of AC4, indicating an increase in total oxygen content after oxidation treatment of the activated carbon. As shown in Fig. 5, the XANES spectra of the activated carbon samples have no fine structures beyond 555 eV, and the X-ray absorption does not change noticeably. Theoretical derivation (see the Supplementary data) shows that the difference in absorption before pre-edge features and post whiteline (or the edge step) is a good measurement of absolute oxygen content in carbon materials. It is determined that if the weight fraction of oxygen is less than 43%, the X-ray absorption edge step measured at 560 eV is proportional to the oxygen weight fraction. This condition should be automatically satisfied for most activated carbons. The edge steps measured at 560 eV for AC4 and OAC4 are 0.0155 and 0.0311, respectively; thus, the total oxygen content in OAC4 is twice that of AC4. As shown in Fig. 5, the XANES spectra of both AC4 and OAC4 have a pre-edge peak at 530 eV, which indicates the presence of carbonyl-type and/or carboxyl-type functional groups. More reliable identification and quantitative information are obtained by performing linear combination fitting, in which sample spectra are fitted to a linear combination of the spectra of individual functional groups collected from reference compounds. The original and fitted spectra of AC4 and OAC4 are shown in Figs. 6 and 7 (same plots together with component spectra are available in the Supplementary data, see Figs. S2 and S3). A residual line is also included. Excellent fits are achieved, as confirmed by the small R-factors reported in Table 2 together with other fitted parameters. All three types of functional groups, namely carbonyl-type, carboxyltype, and hydroxyl-type, are observed in both AC4 and OAC4. The fitted energy shifts are small, which assures that

Fig. 5 – Oxygen K-edge XANES spectra of AC4 and OAC4.

Fig. 6 – Linear combination fitting of XANES spectrum of AC4 (untreated activated carbon).

Fig. 7 – Linear combination fitting of XANES spectrum of OAC4 (oxidized activated carbon). chemical environment is taken into account and there is no misinterpretation of functional groups. The absolute uncertainties in the abundance of functional groups are quite small, about 3% or less. The relative uncertainties (uncertainty divided by fitted abundance) are also small for the major components such as carbonyl-type and carboxyl-type functional groups while it is rather large for the minor component, hydroxyl-type functional groups. Note that the summation of all species is very close to but not exactly 1. We do not force the total to be 1, instead, good fits are expected to converge at 1 as we have observed. The small deviation from 1 can easily result from the errors associated with normalization. The abundance or percentage of each type of functional groups do not change significantly before and after oxidation. In AC4 the most abundant functional groups are carboxyltype, about 58% (phthalic anhydride + naphthoic acid), followed by carbonyl-type functional groups, about 34% (anthrone + anthraldehyde). Hydroxyl-type groups are minor, only about 15%. In OAC4 the most abundant functional groups are still carboxyl-type, about 58%, followed by carbonyl-type functional groups, about 31%. Hydroxyl-type groups are about 17%. Please note, however, the fractions for the two reference

1750

CARBON

4 9 ( 2 0 1 1 ) 1 7 4 5 –1 7 5 1

Table 2 – Results from linear combination fitting of AC4 and OAC4 XANES spectra. Reference spectrum

AC4

Anthrone Anthraldehyde Phthalic anhydride Naphthoic acid Phenolic hydroxyl Aliphatic hydroxyl Sum of abundance R-factor a

OAC4

Abundance (%)

Energy shift (eV)

Abundance (%)

Energy shift (eV)

10.3 ± 3.2 24.0 ± 1.8 12.1 ± 2.6 46.0 ± 1.9 14.8 ± 1.7 0 107.2 ± 5.2a 0.00216

0.08 0.58 0.66 0.26 0.05 0

13.9 ± 2.5 16.7 ± 1.4 32.8 ± 1.6 24.8 ± 1.4 12.4 ± 1.4 4.3 ± 0.9 104.9 ± 3.9a 0.00114

0.04 0.57 0.33 0.58 0.25 1.75

The uncertainty of the sum is calculated by taking square root of the sum of the square of the uncertainties of all components.

spectra for each type of functional groups do change, and in some cases quite significantly. For example, naphthoic acid decreases from 46% in AC4 to 25% in OAC4 while phthalic anhydride increases from 12% in AC4 to 33% in OAC4. From the previous discussions (see Section 3.2), the fitting results suggest that there are probably transformation among functional groups of the same type and changes in the carbon substrate or chemical environment to which the functional groups bind after oxidation of the activated carbon. Fine discrimination of the functional groups of the same type by XANES may be possible through analysis of the subtle differences among the functional groups, and additional characterization provided by other techniques such as titration can also be useful. With the total oxygen contents determined from the X-ray absorption edge step and the relative abundance determined by linear combination fitting for each type of surface functional groups, the absolute amount of different types of functional groups in the samples are quantified. The results are presented in Table 3. Note that the absolute amount is expressed in terms of the total oxygen content in AC4, which is yet unknown. The absolute amount of each type of functional groups in the activated carbons almost doubles after oxidation. This shows an example that by using oxygen K-edge XANES spectroscopy it is possible to monitor the change of a certain type of functional groups in a series of samples. In summary, the reliability of the analysis lies in the fact that XANES spectroscopy measures the electronic structure around the element of interest, which is dominantly determined by the configuration in the local coordination struc-

Table 3 – Absolute content of each type of functional groups in activated carbon samples. Samples

AC4 OAC4 a

Total oxygen content

1a 2.01

Amount of functional groups

Carbonyltype

Carboxyltype

Hydroxyltype

0.32 0.59

0.54 1.1

0.14 0.32

Total oxygen content in AC4 is assumed to be 1.

ture. The unique XANES spectra for different types of functional groups affirm the reliable identification and quantification of the functional groups. Compared to the techniques that rely on deconvolution for quantification such as XPS and TPD in which peak position, peak shape and width are assumed, and the assumptions are difficult to verify, the reliability in XANES analysis is greatly enhanced.

4.

Conclusions

We have successfully demonstrated that oxygen K-edge XANES spectroscopy is a reliable characterization technique for identification and quantification of oxygen functional groups on carbon materials. The key to the success is to collect the spectra in fluorescence yield mode, which does not suffer from the charging effect due to poor conductivity of some carbon materials; as a result, spectrum distortion is minimized and quantification is feasible. The total oxygen contents in a series of samples can be determined from the X-ray absorption edge step since it is proportional to the edge step if the oxygen content is within a certain limit, which can be satisfied in most cases. The unique spectra of carbonyltype, carboxyl-type, and hydroxyl-type functional groups allow reliable determination of the existence and abundance of these types of functional groups by linear combination fitting. Another advantage of oxygen K-edge XANES analysis is the interference from the carbon substrate is minimized because XANES is element specific. Reliability in quantitative XANES analysis is assured when compared to other techniques that rely on deconvolution since the assumptions used in the XANES analysis are strongly supported by the experimental observations.

5.

Disclaimer

Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the Defense Advanced Research Projects Agency.

Acknowledgements This material is based upon work supported by the Defense Advanced Research Projects Agency (DARPA) under Award

CARBON

4 9 ( 20 1 1 ) 1 7 4 5–17 5 1

No. N66001-10-1-4020. The authors thank Dr. Ezana Negusse for the help with carbon XANES measurements. Carbon XANES work on U4B Beamline at the National Synchrotron Light Source (Brookhaven National Laboratory) was supported by the U.S. Department of Energy, Office of Basic Energy Sciences, under Contract DE-AC02-98CH10886.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.carbon.2010.12.060.

R E F E R E N C E S

[1] Baibarac M, Gomez-Romero P. Nanocomposites based on conducting polymers and carbon nanotubes: from fancy materials to functional applications. J Nanosci Nanotechnol 2006;6(2):289–302. [2] Singh R, Pantarotto D, Lacerda L, Pastorin G, Klumpp C, Prato M, et al. Tissue biodistribution and blood clearance rates of intravenously administered carbon nanotube radiotracers. Proc Natl Acad Sci USA 2006;103(9):3357–62. [3] Klein KL, Melechko AV, McKnight TE, Retterer ST, Rack PD, Fowlkes JD, et al. Surface characterization and functionalization of carbon nanofibers. J Appl Phys 2008;103(6):1–26. [4] Yue ZR, Jiang W, Wang L, Gardner SD, Pittman Jr CU. Surface characterization of electrochemically oxidized carbon fibers. Carbon 1999;37(11):1785–96. [5] Liu Y, Vander Wal RL, Khabashesku VN. Functionalization of carbon nano-onions by direct fluorination. Chem Mater 2007;19(4):778–86. [6] Subramoney S. Novel nanocarbons-structure, properties, and potential applications. Adv Mater 1998;10(15):1157–71. [7] Al-Ghouti MA, Yousef I, Ahmad R, Ghrair AM, Al-Maaitah AA. Characterization of diethyl ether adsorption on activated carbon using a novel adsorption refrigerator. Chem Eng J 2010;162(1):234–41. [8] Li KQ, Zheng Z, Li Y. Characterization and lead adsorption properties of activated carbons prepared from cotton stalk by one-step H3PO4 activation. J Hazard Mater 2010;181(1–3): 440–7. [9] Wang XL, French J, Kandadai S, Chua HT. Adsorption measurements of methane on activated carbon in the temperature range (281 to 343) K and pressures to 1.2 MPa. J Chem Eng Data 2010;55(8):2700–6. [10] Zhang SJ, Shao T, Kose HS, Karanfil T. Adsorption of aromatic compounds by carbonaceous adsorbents: a comparative study on granular activated carbon, activated carbon fiber, and carbon nanotubes. Environ Sci Technol 2010;44(16):6377–83. [11] Calvino-Casilda V, Lopez-Peinado AJ, Duran-Valle CJ, MartinAranda RM. Last decade of research on activated carbons as catalytic support in chemical processes. Cat Rev – Sci Eng 2010;52(3):325–80. [12] Eguchi M, Okubo A, Yamamoto S, Kikuchi M, Uno K, Kobayashi Y, et al. Preparation of catalyst for a polymer electrolyte fuel cell using a novel spherical carbon support. J Power Sources 2010;195(18):5862–7. [13] Soares O, Orfao JJM, Pereira MFR. Pd–Cu and Pt–Cu catalysts supported on carbon nanotubes for nitrate reduction in water. Ind Eng Chem Res 2010;49(16):7183–92.

1751

[14] Chen JP, Wu S. Acid/base-treated activated carbons: characterization of functional groups and metal adsorptive properties. Langmuir 2004;20(6):2233–42. [15] Jiang Z, Liu Y, Sun X, Tian F, Sun F, Liang C, et al. Activated carbons chemically modified by concentrated H2SO4 for the adsorption of the pollutants from wastewater and the dibenzothiophene from fuel oils. Langmuir 2003;19(3):731–6. [16] Lakshminarayanan PV, Toghiani H, Pittman CUJ. Nitric acid oxidation of vapor grown carbon nanofibers. Carbon 2004;42(12–13):2433–42. [17] Li B, Shi Z, Lian Y, Gu Z. Aqueous soluble single-wall carbon nanotube. Chem Lett 2001;30(7):598–9. [18] Zhao B, Hu H, Yu A, Perea D, Haddon RC. Synthesis and characterization of water soluble single-walled carbon nanotube graft copolymers. J Am Chem Soc 2005;127(22):8197–203. [19] Zhao W, Song C, Pehrsson PE. Water-soluble and optically pHsensitive single-walled carbon nanotubes from surface modification. J Am Chem Soc 2002;124(42):12418–9. [20] Boehm HP. Surface oxides on carbon and their analysis: a critical assessment. Carbon 2002;40(2):145–9. [21] Langley LA, Villanueva DE, Fairbrother DH. Quantification of surface oxides on carbonaceous materials. Chem Mater 2006;18(1):169–78. [22] Bajt S, Sutton SR, Delaney JS. X-ray microprobe analysis of iron oxidation states in silicates and oxides using X-ray absorption near edge structure (XANES). Geochim Cosmochim Acta 1994;58(23):5209. [23] Lamberti C, Bordiga S, Bonino F, Prestipino C, Berlier G, Capello L, et al. Determination of the oxidation and coordination state of copper on different Cu-based catalysts by XANES spectroscopy in situ or in operando conditions. Phys Chem Chem Phys 2003;5(20):4502–9. [24] Chen Y, Xie C, Li Y, Song C, Bolin TB. Sulfur poisoning mechanism of steam reforming catalysts: an X-ray absorption near edge structure (XANES) spectroscopic study. Phys Chem Chem Phys 2010;12(21):5707–11. [25] Turner JA, Thomas KM, Russell AE. Identification of oxygen functional groups in carbonaceous materials by oxygen Kedge XANES. Carbon 1997;35(7):983–92. [26] Braun A. Comment on ‘‘effects of native organic material and water on sorption properties of reference diesel soot’’. Environ Sci Technol 2009;43(13):5158–9. [27] Braun A, Huggins FE, Shah N, Chen Y, Wirick S, Mun SB, et al. Advantages of soft X-ray absorption over TEM-EELS for solid carbon studies – a comparative study on diesel soot with EELS and NEXAFS. Carbon 2005;43(1):117–24. [28] Braun A, Mun BS, Huggins FE, Huffman GP. Carbon speciation of diesel exhaust and urban particulate matter NIST standard reference materials with C(1s) NEXAFS spectroscopy. Environ Sci Technol 2007;41(1):173–8. [29] Braun A, Shah N, Huggins FE, Kelly KE, Sarofim A, Jacobsen C, et al. X-ray scattering and spectroscopy studies on diesel soot from oxygenated fuel under various engine load conditions. Carbon 2005;43(12):2588–99. [30] Lee DW, De Los Santos L, Seo JW, Felix LL, Bustamante A, Cole JM, et al. The structure of graphite oxide: investigation of its surface chemical groups. J Phys Chem B 2010;114(17):5723–8. [31] Chen CT, Sette F. Performance of the dragon soft-X-ray beamline. Rev Sci Instrum 1989;60(7):1616–21. [32] Ravel B, Newville M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J Synchrotron Radiat 2005;12:537–41.