Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption

CARBON

x x x ( 2 0 1 4 ) x x x –x x x

Available at www.sciencedirect.com

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

Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption Yonghong Zeng a, Luisa Prasetyo a, Van T. Nguyen a, Toshihide Horikawa b, D.D. Do a,*, D. Nicholson a a

School of Chemical Engineering, University of Queensland, St. Lucia, Qld 4072, Australia Department of Advanced Materials, Institute of Technology and Science, The University of Tokushima, 2-1 Minamijosanjima, Tokushima 770-8506, Japan b

A R T I C L E I N F O

A B S T R A C T

Article history:

We propose a new method, using the ambient temperature adsorption of methanol in the

Received 14 July 2014

Henry law region as a molecular probe, to determine the concentration of surface oxygen

Accepted 24 September 2014

functional groups (Ca) on carbon adsorbents. Two adsorbents: porous A5 and non-porous

Available online xxxx

Carbopack F have been chosen here as experimental examples. The theoretical Henry constant (Ka) is the product of the intrinsic interaction between a methanol molecule and one functional group, a, having a concentration (Ca) and is estimated by carrying out a volume integration of the Boltzmann factor for methanol interacting with the functional group. Our results show that Ca determined with methanol for A5 compares well with that determined previously with water as the molecular probe (Nguyen, 2014 [1]); and both are in good agreement with results from Boehm titration. For Carbopack F, we find that Ca, determined by methanol adsorption, gives more realistic concentrations than those obtained by Boehm titration, which are subject to large errors because of the very low concentration of the functional groups. The method proposed here is fast and easy to implement, and serves as an alternative to the Boehm titration technique.  2014 Published by Elsevier Ltd.

1.

Introduction

The adsorption of associating fluids on carbonaceous materials is significantly affected by the presence of oxygen-containing functional groups, which are hydrophilic and act as strong adsorption sites for these fluids [2–7]. A typical example is the adsorption of water, a strong associating fluid, on activated carbon, which contains several oxygenated surface groups (e.g., carboxyl [8], carbonyl [4,8] and hydroxyl groups [2,8]) that are attached at the edge of graphene layers. Water

molecules interact strongly with these groups via electrostatic forces to form a nucleus onto which further adsorption of water occurs as pressure is increased to form clusters. These clusters merge at higher pressure and can fill the volume space of micropores [9]. This illustrates the importance of functional groups in increasing the ‘‘apparent’’ affinity of carbon towards water despite the hydrophobic nature of carbon surfaces. Boehm titration [10,11] is valuable widely used technique to detect and quantify the functional group concentration

* Corresponding author. E-mail address: [email protected] (D.D. Do). http://dx.doi.org/10.1016/j.carbon.2014.09.077 0008-6223/ 2014 Published by Elsevier Ltd.

Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

CARBON

x x x ( 2 0 1 4 ) x x x –x x x

on carbon surfaces. It is the most frequent use stems from its relative simplicity and the fact that it is inexpensive compared to spectroscopy analysis [12]. However, the method is time consuming and may modify the surface chemistry of the adsorbent. More importantly, it is subject to large errors when the functional groups are at low concentrations. There is therefore a need for a new method that circumvents these problems. Recently, we developed a method that meets this need based on the adsorption of water on activated carbon [1]. At low enough pressures the interaction between water and the adsorbent surface is dominated by the functional groups [13], and the strength of this interaction is described by the experimental Henry constant (K). The theoretical Henry constant (Ka) for a specific group, can be determined from a volume integration of the Boltzmann factor over the rotational coordinates and the accessible space around the group. Knowing the experimental Henry constant K and the intrinsic single-site constant Ka, the functional group concentration (Ca) is simply calculated as Ca = K/Ka. This new method is easy to implement and has been found to give good agreement with concentrations measured on an active carbon by Boehm titration [1]. The potential advantages of using the adsorption of associating fluids to determine the concentration of functional groups is extended in this paper by examining methanol as an alternative adsorbate. It has been shown that at low enough pressures methanol interacts primarily with functional groups on the carbon surface [14,15], but because the partial charges are weaker than for water, different Henry law constants are expected. Here we examine methanol adsorption on a porous activated carbon, A5, and a nonporous graphitized thermal carbon black, Carbopack F. On the latter adsorbent, the concentration of the functional groups is very low, and it serves to highlight the potential of our method compared to Boehm titration which is unsatisfactory in this case.

2.

Experimental

2.1.

Materials

Two carbons with widely different concentrations of functional groups and porous structure were chosen: (1) A5 (supplied by Ad’all Co. in Japan) which is a porous activated carbon with high concentrations of functional groups and (2) a highly graphitized thermal carbon black, Carbopack F (supplied by Supelco, USA), which is highly homogeneous and non-porous, and has very low concentrations of functional groups. The BET surface areas, obtained from nitrogen isotherms at 77 K, for A5 and Carbopack F are 510 m2/g and 4.9 m2/g, respectively.

2.2.

Measurement

Water and methanol adsorption were measured at temperatures in the range between 263 K and 298 K using a constant volume adsorption apparatus (BELSORP-max, BEL Japan). At each point on the isotherm, the system was initially allowed

Table 1 – Concentrations of functional groups on A-5 and Carbopack F from Boehm titration. Adsorbent

A5 Carbopack F

Carboxylic

Lactonic

Phenolic

(mmol/g)

(mmol/g)

(mmol/g)

0.103 0.019

0.000 0.000

0.184 0.054

5

MeOH - A5 Amount adsorbed (mmol/g)

2

263 K 283 K 298 K

4

101

3

2 100

1 10-1 10-2

0 0.0

0.2

0.4

0.6

0.8

1.0

P/P0

Fig. 1 – Adsorption isotherms of methanol on A5 at different adsorption temperatures, 263, 283 and 298 K. (A colour version of this figure can be viewed online.)

to equilibrate for 300 s if the pressure change over this time was less than 0.3%, the measurement was accepted as being at (quasi) equilibrium; if a larger change was observed, equilibration was continued for a further 300 s until this criterion was met. The samples were degassed at 473 K for 5 h under vacuum at pressures less than 0.1 mPa to remove any physically adsorbed components before each measurement. The concentrations of surface functional groups for each carbon were measured by Boehm titration [10,11] and are listed in Table 1. Only carboxylic and phenolic groups were detected in A5 and Carbopack F. For A5 carbon, approximately 0.1 g of solid was added to excess (100 cm3) standard base solution (0.01 N NaOH, 0.001 N Na2CO3, and 0.001 N NaHCO3), and the acidic oxides on the surface were determined by back-titration with HCl after allowing the mixture to stand for 48 h at 298 K. For Carbopack F, larger quantity (1 g of solid) was used because of the low concentrations of functional groups.

3.

Theory

3.1.

Henry’s law

Henry’s law expresses the amount adsorbed on the surface of a solid at very low loadings as a linear function of the bulk gas concentration, given in the following equation: C¼

K P Rg T

ð1Þ

Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

CARBON

0.030

MeOH - Carbopack F

(a)

263 K 273 K 283 K 298 K

0.08

Water - Carbopack F

(b)

10-2

0.025 Amount adsorbed (mmol/g)

Amount adsorbed (mmol/g)

0.10

3

xxx (2014) xxx–xxx

0.06 10-2

0.04

10-3

10-4

0.02

10-3

263 K 273 K 283 K 298 K

0.020 10-4

0.015

0.010

10-5 10-4

10-3

10-2

10-1

0.005 10-5 10-3

0.00 0.0

0.2

0.4

10-2

0.6

10-1

0.8

0.000 0.0

1.0

0.2

0.4

0.6

0.8

1.0

P/P0

P/P0

Fig. 2 – Adsorption isotherms of methanol (a) and water (b) on Carbopack F at different adsorption temperatures, 263, 273, 283 and 298 K. (A colour version of this figure can be viewed online.)

Table 2 – K (m3/g) and Ka (1027 m3/functional group) at various temperatures using methanol as molecular probe.

K for methanol on A-5 K for methanol on Carbopack F Ka for methanol (Dihedral angle: 180) Ka for methanol (Dihedral angle: 90)

263 K

273 K

283 K

298 K

0.1510 1.20 · 105 1318.61 10595.00

– 7.85 · 106 833.75 2962.57

0.0397 5.97 · 106 631.86 1999.39

0.0122 2.09 · 106 214.48 887.98

Table 3 – K (m3/g) and Ka (1027 m3/functional group) at various temperatures using water as molecular probe.

K for water on Carbopack F Ka for water (Dihedral angle: 180) Ka for water (Dihedral angle: 90)

263 K

273 K

283 K

298 K

3.55 · 105 2741.70 16354.94

2.54 · 105 1412.55 7781.09

1.99 · 105 1006.99 4126.57

1.30 · 105 458.12 1721.60

where C is the amount adsorbed (mol/g), P is the absolute pressure, Rg is the gas constant, T is the temperature of the system and K is the experimental Henry constant (m3/g).

3.2.

Experimental Henry constant

For water and methanol adsorption on carbons, the oxygen containing functional groups are much stronger sites than the basal plane of the graphene layers. At low concentration the adsorbate will be localised in the region of these groups and the weaker interaction with the basal carbon atoms can be neglected. Therefore, to an excellent approximation, the Henry constant K is a measure of the interaction between the adsorbate molecule and all the functional groups. The adsorption isotherms of methanol on A5 are shown in Fig. 1 and those for methanol and water on Carbopack F are shown in Fig. 2. The insets in these figures depict isotherms plotted on logarithmic scales to confirm that data at low pressures obey Henry’s law. The experimental Henry constants K (m3/g), determined from these plots, are listed in Tables 2 and 3.

3.3.

Theoretical Henry constant

The Henry constant between a molecule and one functional group, Ka, can be calculated by integrating the Boltzmann factor over the volume space around the functional group and over all possible orientations of the molecule [13]: Z Z Z Z Ka ¼ exp½buðr; xÞdrdx  H½uðr; xÞdrdx ð2Þ X

X

where X is domain accessible to adsorbate molecule, b ¼ kB1T, u is the potential energy of interaction between an adsorbate molecule at the position r and orientation x with the functional group, and H is a Heaviside function. Ka has the dimensions of a volume. The intermolecular interaction energy between a fluid molecule i and a functional group j,ui,j is given by the summation of the 12-6 Lennard Jones interaction and the Coulombic interaction: 2 !12 !6 3 A X B C X D qc qd X ra;b ra;b 1 X i j i;j i;j a;b 4 ui;j ¼ 4 i;j  a;b 5 þ ð3Þ a;b c;d 4p r r 0 r a¼1 b¼1 c¼1 d¼1 i;j i;j i;j

Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

4

CARBON

x x x ( 2 0 1 4 ) x x x –x x x

the permittivity of a vacuum. The parameter rc;d i;j is the distance between partial charges. The intermolecular potential of water is modelled with the SPC/E model [16] having one LJ site and three partial charges, and methanol is modelled with the TraPPE-UA model having two LJ sites and three partial charges [17]. The three functional groups considered in

where parameter va;b i;j is associated with a site a on i and a site a;b b on j. The parameters ra;b i;j and i;j are the cross collision diameter and the cross well depth of interaction energy, respectively, and are calculated from the Lorentz–Berthelot mixing rule. Parameter ra;b i;j is the distance between the LJ sites. The parameter qki is the partial charge on the site k of i and 0 is

105

105

3

104

103

102

101 250

260

270

280

290

103

102

101 250

300

carbonyl hydroxyl carboxyl

104

Henry constant, nm

3

(a) Henry constant, nm

(b)

carbonyl hydroxyl carboxyl

260

270

T, K

280

290

300

T, K

Fig. 3 – Theoretical Henry constants (K) of water (a) and methanol (b) as a function of temperature for carboxyl group (circles), hydroxyl group (squares) and carbonyl group (triangles) (readers are referred to Appendix B for the full version of these figures). (A colour version of this figure can be viewed online.)

106

104 Carboxyl

water methanol

3

Henry constant, nm

Henry constant, nm

3

105

water methanol

104

103

102

101

10

0

103

102

101

(b)

(a) Carboxyl

240

10

260

280

300

320

340

360

380

0

Carbonyl

240

260

280

300

320

340

360

380

T, K

T, K

103

Henry constant, nm

3

water methanol

102

101

(c) Hydroxyl 10

0

240

260

280

300

320

340

360

380

T, K

Fig. 4 – Theoretical Henry constants (Ka) as a function of temperature of water (triangles) and methanol (squares) contributed by carboxyl group (a), carbonyl group (b) and hydroxyl group (c) (readers are referred to Appendix B for the full version of these figures). (A colour version of this figure can be viewed online.) Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

CARBON

5

xxx (2014) xxx–xxx

this work are carbonyl, hydroxyl and carboxyl groups; their molecular parameters are taken from the OPLS set [18] and are given in the Appendix (Table A1). In the Monte Carlo integration of Eq. (2), the simulation box consists of eleven stacked finite graphene surfaces with one functional group attached at the edge of the graphite. Details of the simulation box can be found in our previous work [1]. The interaction energy between fluid molecule and the semi-infinite surface layer was determined using the Bojan–Steele potential model [19–21]. The theoretical values of Ka for water and methanol with various functional groups are presented in Fig. 3a and b, respectively. The same figures for an extended temperature range can be found in Appendix B (Figs. B1 and B2). It should be noted that the initial adsorption occurs at those sites whose affinities are highest. As was found earlier for water, the carboxyl group has the highest affinity in the range of temperatures investigated in this work (263–298 K); its Henry constant is two orders of magnitude higher than those for the carbonyl and hydroxyl groups. Therefore, unless the carboxyl group is absent from the surface or its concentration is at least two orders of magnitude lower than the other groups the initial adsorption is dominated by the carboxyl group. It is interesting, but not surprising, that the affinities of water and methanol towards the oxygenated groups attached at the edge of graphite are comparable as both models have two positive partial charges (q+) and a single negative charge (q) of comparable magnitude as shown graphically in Fig. 4a–c.

3.4.

Concentration of surface oxygen-containing groups

Knowing the experimental Henry constant (K) which measures the interaction of an adsorbate molecule with all adsorption sites on the surface and the theoretical Henry constant (Ka) which measures the interaction by one functional group with the adsorbate molecule, the functional group concentration Ca (mol/g) can be calculated from the following equation: K ¼ Ca Ka NA

ð4Þ

where NA is the Avogadro number. In a previous study by Morimoto and Miura [22,23], no carboxyl groups were detected on a graphite sample of 99.5% purity with 0.5% ashes, and BET surface area of 8.63 m2/g after heat treatment at temperatures above 1000 C; although there was a measurable amount of phenolic group. By contrast, our results from Boehm titration on Carbopack F show a detectable concentration of carboxyl group, possibly because the highly ordered structure of Carbopack F stabilizes these groups.

4.

Results and discussion

The experimental K (m3/g) and theoretical Ka (m3) for carboxyl group at ambient temperatures obtained from (1) and (2), respectively, are listed in Tables 2 and 3 for methanol and water, respectively. The concentrations Ca (mol/g) for A5 and

Table 4 – Estimated concentrations of carboxyl groups (mmol/g) of A-5 at various temperatures.

Methanol on A-5 (180) Methanol on A-5 (90)

263 K

283 K

298 K

Average

Boehm

190.19 23.67

104.35 32.98

94.75 22.89

129.76 26.51

0.103

Table 5 – Estimated concentrations of carboxyl groups (mmol/g) of A-5 with CACAOAH dihedral angle of 90 and F = 1.15, F = 1.2 and F = 1.25.

Methanol on A-5 (F = 1.15) Methanol on A-5 (F = 1.2) Methanol on A-5 (F = 1.25)

263 K

283 K

298 K

Average

Boehm

1.07 0.33 0.10

1.47 0.45 0.15

1.23 0.44 0.15

1.26 0.41 0.13

0.103

Table 6 – Estimated concentrations of carboxyl groups (mmol/g) of Carbopack F with optimized CACAOAH dihedral angle of 90 (for methanol) and 100 (for water) and F = 1.2.

Methanol on Carbopack F (F = 1.2) Water on Carbopack F (F = 1.2)

263 K

273 K

283 K

298 K

Average

Boehm

0.000029 0.000021

0.000045 0.000036

0.000067 0.000078

0.000081 0.000148

0.000055 0.000071

0.019

Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

6

CARBON

x x x ( 2 0 1 4 ) x x x –x x x

interaction energy between one adsorbate molecule and one functional group: Ui;j ¼ F  ui;j

104

103

Kα ,10

27

3

m /functional group

105

102

101 0

20

40

60

80

100

120

140

160

180

C-C-O-H dihedral angle,degree

Fig. 5 – Theoretical Henry constant for methanol-carboxyl interaction, as a function of the CACAOAH dihedral angle at 263 K.

Carbopack F are determined using (4) and are given in Tables 4–6.

4.1.

Effect of dihedral angle

In the Monte Carlo integration of Eq. (2), the theoretical Henry constant Ka was initially obtained by positioning the atoms in the carboxyl group with the hydrogen atom in the same plane as the other atoms, giving a CACAOAH dihedral angle of 180. With this arrangement, our result shows that the calculated carboxyl concentrations for A5 using methanol as the molecular probe overestimated those measured with Boehm titration (see Table 4), indicating underestimation of the Ka. value. Clearly it is important to account for the variation in the orientation of the CACAOAH group, and the one that gives the highest Ka is the optimum dihedral angle that maximises the interaction between molecular probe and the carboxyl group. In our previous work, the optimum CACAOAH dihedral angle was determined by rotating the OAH bond about the CAO bond. There it was found that the optimum CACAOAH dihedral angle for water was 100, which gives a Ka value five times greater than that for a dihedral angle of 180. For methanol, the optimum CACAOAH dihedral angle was found to be 90 (See Fig. 5). This gives a significantly higher value of Ka which translates to lower concentrations of carboxyl groups for A5 as seen in Table 4.

4.2.

Comparison with experiment

In our proposed method, the graphite was modelled as perfect graphene layers whereas the partial charges on the carboxyl group were taken from the potential optimized for a simulation (OPLS) of liquid acetic acid [18]. Hence, it is very likely that there are discrepancies between the simulation results and the Boehm titration results as the calculated interaction energy may not correspond to that in the actual physical system because the partial charges on the carboxylic group are modified by the sea of electrons in the graphene layers. To account for this as well as other structural and energetic factors we introduced a factor F in the calculation of the

ð5Þ

where ui;j is the interaction energy between a fluid molecule i and a carboxyl group j. In our previous work using water as the molecular probe, this factor was found to be F = 1.15 (i.e., a 15% increase of interaction energy) and using this factor we find the carboxyl group concentration for A5 of about 0.11 mmol/g, which is in good agreement with the Boehm titration result of 0.103 mmol/g. Using methanol as the molecular probe, a factor F = 1.25 was found to give a good agreement between the calculated concentration (0.1 mmol/ g) and Boehm titration results (0.103 mmol/g) as tabulated in Table 5. We attribute the discrepancy between the F factors for methanol and water to the differences in configuration of oxygenated functional groups on the carbon surfaces. This is under investigation and will be reported in a future correspondence. To further demonstrate the applicability of the new method, we determined the carboxyl group concentrations of non-porous Carbopack F using water and methanol as molecular probes. A factor F = 1.2 was employed in the Monte Carlo integration to estimate Ka for the aforementioned reasons. This factor was chosen as the average of the factors F = 1.15 and F = 1.25, resulting from our analysis of carboxyl groups on A5 using water and methanol and bearing in mind that A5 has a high enough Ca to be able to obtain viable results using the Boehm titration technique. We see in Table 6 that measurements using either water or methanol give Ca values of the same order of magnitude, but lower than those predicted by Boehm titration by about three orders of magnitude. This raises the question ‘‘which method gives credible results?’’

4.3.

BET surface area of Carbopack F

Here we show that the total concentration of functional groups (carboxylic and phenolic) determined by the Boehm titration on Carbopack F (0.073 mmol/g) is incorrect: If this concentration is spread uniformly on a surface then, assuming a molecular projection area of 0.15 nm2, the functional groups would occupy an area of 6.6 m2/g, which is larger than the BET surface area of 4.9 m2/g. Since highly graphitized thermal carbon black consists predominantly of graphene basal planes, nitrogen adsorption on this material will give a statistical monolayer coverage corresponding exactly to the BET surface area. We can conclude that the Boehm titration is seriously in error and is not reliable for carbons with low concentrations of functional groups.

4.4.

Spacing of carboxylic groups on Carbopack F

To add further support to the above assertion, we calculated the spacing between adjacent functional groups on Carbopack F. The electron micrographs of Carbopack F, show that the particles have a polyhedron shape, typical of a highly graphitized thermal carbon black. To a first approximation, we can treat the particles as cubic then, given a BET surface area of 4.9 m2/g and assuming a carbon density of 2000 kg/

Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

CARBON

7

xxx (2014) xxx–xxx

m3, the linear dimension of the cubes is 600 nm. Taking the concentration of the carboxylic functional groups to be that determined by the Boehm titration (0.019 mmol/g) gives a figure of 220 molecules along every 0.3 nm of edge; which is impossible! If, on the other hand, we use the value of Ca determined by methanol adsorption we find 0.6 molecules for every 0.3 nm along the edge of the cubic crystal. This is a reasonable value, and it clearly shows the superiority of our method over the Boehm titration method for determining low surface oxygenated functional groups.

4.5. Comparative plot to estimate the concentration of oxygen functional group The concentration of functional groups containing oxygen can also be estimated by a comparative plot. In this section we present further evidence to support our method for determining the functional group concentrations (Ca), in the form of the comparative plot shown in Fig. 6, where the amount of water adsorbed in porous A5 is plotted against the amount adsorbed in the non-porous Carbopack F. The comparative plot serves to show that Ca for porous A5 is greater than that for Carbopack F. We can identify various stages of adsorption shown as dashed lines:

Amount of water adsorbed on A5 (ml(STP)/g)

(1) The first stage, AB, corresponds to water growth around each cluster and this mechanism is similar for both porous A5 and non-porous Carbopack F. (2) The second stage, BC, occurs when clusters at the edges of the same walls of A5 merge to cover the walls. The upward slope of the graph here reflects the higher concentration of functional groups in A5. (3) In the third stage CD there is a secondary merging, linking clusters across the pore walls. (4) In the fourth stage, DE, the micropores of A5 are filled, followed by the complete filling.

Fig. 7 – Critical cluster diameter dc. (A colour version of this figure can be viewed online.)

The schematic of these adsorption stages is illustrated in Appendix C (Figs. C1–C5). In Fig. 6, D corresponds to the point at which the clusters reach a critical size, dc, and water starts to penetrate into the micropore. We may assume that dc is approximately equal to the width of 1.25 nm of the micropores in A5 (see Fig. 7). Knowing the size of the cluster occupying one functional group, we are able to estimate the amount of water in one cluster. Knowing the total amount of water adsorbed on Carbopack F at point D (= a0(mmol/g)), we can estimate Ca (mmol/g) from the following equation: ! 3 pdc  qm  NA  Ca ¼ a0 ð6Þ 6 where qm is molar density of water (mol/m3) at 298 K. The calculated Ca (0.000193 mmol/g) is approximately 100 times lower than Boehm titration result and is comparable to our previously determined Ca (See Table 6).

Table A1 – Molecular parameters of the models of oxygen functional groups, methanol and water.

200

150

E

100 D

50

Model

Interacting site

˚) r (A

e/Kb (K)

q (e)

Methanol

C O H

3.75 3.02 0.0

98 93 0.0

0.265 0.7 0.435

Water

O H

3.166 0.0

78.2 0.0

0.8476 0.4238

Carbonyl

Ca O

0.0 0.296

0.0 105.8

0.5 0.5

Hydroxyl

Ca O H

0.0 0.307 0.0

0.0 78.2 0.0

0.2 0.64 0.44

Carboxyl

Ca C O(@C) O(AH) H(AO)

0.0 0.375 0.296 0.3 0.0

0.0 52 105.7 85.6 0.0

0.08 0.55 0.5 0.58 0.45

C

0

A

0.0

B

0.1 <0.25>

a0

0.2

0.3

<0.72>

0.4 <0.94>

Amount of water adsorbed on Carbopack F (ml(STP)/g)

Fig. 6 – Comparative plot of water adsorption on A5 and Carbopack F. (A colour version of this figure can be viewed online.)

a

Carbon located in plane of graphene sheet.

Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

8

5.

CARBON

x x x ( 2 0 1 4 ) x x x –x x x

accurately employed, and to a graphitized thermal carbon black, Carbopack F, where Ca is too low to obtain valid results by this method. For the A5 activated carbon, we found that Ca, determined by methanol adsorption is comparable to that determined using water as the molecular probe, reported in our previous work, and these values are in good agreement with those measured by the Boehm titration method. When our method was applied to determine the concentration of

Conclusion

This paper explores the use of methanol as a promising molecular probe to determine the concentrations of functional groups (Ca) by carrying out adsorption at ambient temperatures in the Henry’s law region. To illustrate the potential of our method we applied it to a sample of activated carbon, A5, whose Ca is high enough for the Boehm technique to be

105

105

carbonyl hydroxyl carboxyl

104

Henry constant, nm

3

104

Henry constant, nm

3

carbonyl hydroxyl carboxyl

16 15

103

14 13 12

102

11 10 600

(a)

700

800

900

16 15

103

14 13 12

102

11 10 600

(b)

1000

700

800

900

1000

101

101 0

200

400

600

800

1000

0

200

400

T, K

600

800

1000

T, K

Fig. B1 – Theoretical Henry constants (K) of water (a) and methanol (b) as a function of temperature for carboxyl group (circles), hydroxyl group (squares) and carbonyl group (triangles). (A colour version of this figure can be viewed online.)

105

105

methanol water

16

104

15

14

103

13

12

10

11

2

(a)

10 600

15

14

103

800

900

13

12

10 700

16

104

Henry constant, nm3

Henry constant, nm3

methanol water

11

2

10 600

(b)

1000

700

800

900

1000

Carbonyl

Carboxyl

101

101 200

400

600

800

1000

200

400

T,K

600

800

1000

T,K

105

Henry constant, nm3

methanol water 16

104

15

14

103

13

12

11

102

(c)

10 600

700

800

900

1000

Hydroxyl

10

1

200

400

600

800

1000

T,K

Fig. B2 – Theoretical Henry constants (Ka) as a function of temperature of water (triangles) and methanol (squares) contributed by carboxyl group (a), carbonyl group (b) and hydroxyl group (c). (A colour version of this figure can be viewed online.) Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

CARBON

xxx (2014) xxx–xxx

functional groups in Carbopack F with methanol as the molecular probe, we found that Ca is very low as expected for a highly graphitized sample. On the other hand, however, the Boehm titration gave a high value of 0.07 mmol/g, which corresponds to an area of 6.6 m2/g covered by the functional groups. This is clearly not realistic because: (1) the BET

9

surface area of Carbopack F is only 4.9 m2/g, (2) this sample is highly homogeneous and (3) its surface area consists mainly of basal planes of graphene which are unlikely to have attached functional groups. This example shows the superiority of our proposed method to determine functional group concentration for samples with a low concentration of

Fig. C1 – Illustration of water cluster growth (stage A and B). (A colour version of this figure can be viewed online.)

Fig. C2 – Illustration of the merging of water clusters of the same wall (stage B and C). (A colour version of this figure can be viewed online.)

Fig. C3 – Illustration of the secondary merging of water clusters (stage C and D). (A colour version of this figure can be viewed online.) Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

10

CARBON

x x x ( 2 0 1 4 ) x x x –x x x

Fig. C4 – Illustration of water penetration into the micropore (stage D and E). (A colour version of this figure can be viewed online.)

Fig. C5 – Illustration of water filling the micropore (after stage E). (A colour version of this figure can be viewed online.)

functional groups. Further work on investigations with methanol will be applied to more samples of carbons, and reported in a future correspondence.

Acknowledgement This work was supported by the Australian Research Council, and by the Japan Society for the Promotion of Science, Grantin-Aid for Young Scientists (B), 24750146.

Appendix A. Molecular parameters Table A1

Appendix B. Figs. B1 and B2

Appendix C. Figs. C1–C5

R E F E R E N C E S

[1] Nguyen VT et al. Water as a potential molecular probe for functional groups on carbon surfaces. Carbon 2014;67:72–8. [2] McCallum CL et al. A Molecular Model for Adsorption of Water on Activated Carbon: Comparison of Simulation and Experiment. Langmuir 1998;15(2):533–44. [3] Lodewyckx P et al. Water isotherms of activated carbons with small amounts of surface oxygen groups: fitting the mesopore region. Carbon 2001;39(2):309–10. [4] Brennan JK, Thomson KT, Gubbins KE. Adsorption of water in activated carbons: effects of pore blocking and connectivity. Langmuir 2002;18(14):5438–47.

Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077

CARBON

xxx (2014) xxx–xxx

[5] Liu JC, Monson PA. Monte Carlo simulation study of water adsorption in activated carbon. Ind Eng Chem Res 2006;45(16):5649–56. [6] Birkett G, Do DD. Simulation study of water adsorption on carbon black: the effect of graphite water interaction strength. J Phys Chem 2007;111(15):5735–42. [7] Ye X et al. Prediction of adsorption equilibrium using a modified D-R equation: pure organic compounds on BPL carbon. Carbon 2003;41(4):681–6. [8] Jorge M, Schumacher C, Seaton NA. Simulation study of the effect of the chemical heterogeneity of activated carbon on water adsorption. Langmuir 2002;18(24):9296–306. [9] Ajna To´th KLaszlo. Water adsorption by carbons. Hydrophobicity and hydrophilicity. In: Tasco´n JMD, editor. Novel Carbon Adsorbents. Oxford: Elsevier; 2012. p. 147–71. [10] Boehm HP. Chemical Identification of Surface Groups, in Advances. In: Eley HPDD, Paul BW, editors. Catalysis. Academic Press; 1966. p. 179–274. [11] Boehm HP. Some aspects of the surface chemistry of carbon blacks and other carbons. Carbon 1994;32(5):759–69. [12] Bandosz TJ. Molecular models of porous carbons. In: Radovic LR, editor. Molecular models of porous carbons, in Chemistry and physics of carbon. New York: Marcel Dekker Inc; 2003. p. 41–228. [13] Do DD, Nicholson D, Do HD. On the Henry constant and isosteric heat at zero loading in gas phase adsorption. J Colloid Interface Sci 2008;324(1–2):15–24. [14] Salame II, Bandosz TJ. Adsorption of water and methanol on micro- and mesoporous wood-based activated carbons. Langmuir 2000;16(12):5435–40.

11

[15] Tarasevich YI, Aksenenko EV. Interaction of water, methanol and benzene molecules with hydrophilic centres at a partially oxidised model graphite surface. Colloids Surf, A 2003;215(1–3):285–91. [16] Berendsen HJC, Grigera JR, Straatsma TP. The missing term in effective pair potentials. J. Phys. Chem. B 1987;91(24):6269–71. [17] Chen B, Potoff JJ, Siepmann JI. Monte Carlo calculations for alcohols and their mixtures with alkanes. Transferable potentials for phase equilibria. 5. United-atom description of primary, secondary, and tertiary alcohols. J Phys Chem B 2001;105(15):3093–104. [18] Jorgensen WL, Maxwell DS, Tirado-Rives J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J Am Chem Soc 1996;118(45):11225–36. [19] Bojan MJ, Steele WA. Computer simulation of physisorption on a heterogeneous surface. Surf Sci 1988;199(3):L395–402. [20] Bojan MJ, Steele WA. Computer simulation of physisorbed Kr on a heterogeneous surface. Langmuir 1989;5:625–33. [21] Bojan MJ, Steele WA. Computer simulations of the adsorption of xenon on stepped surfaces. Mol Phys 1998;95(3):431–7. [22] Morimoto T, Miura K. Adsorption sites for water on graphite. 1. Effect of high-temperature treatment of sample. Langmuir 1985;1(6):658–62. [23] Morimoto T, Miura K. Adsorption sites for water on graphite. 2. Effect of autoclave treatment of sample. Langmuir 1986;2(1):43–6.

Please cite this article in press as: Zeng Y et al. Characterization of oxygen functional groups on carbon surfaces with water and methanol adsorption. Carbon (2014), http://dx.doi.org/10.1016/j.carbon.2014.09.077