View Online / Journal Homepage / Table of Contents for this issue

Green Chemistry
Cite this: Green Chem., 2012, 14, 741 www.rsc.org/greenchem

Dynamic Article Links

PAPER

A one-pot hydrothermal synthesis of tunable dual heteroatom-doped carbon microspheres

Stephanie-Angelika Wohlgemuth,* Filipe Vilela, Maria-Magdalena Titirici and Markus Antonietti
Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

Received 7th November 2011, Accepted 13th December 2011 DOI: 10.1039/c2gc16415a

In terms of broadening the potential application range of carbon based materials, heteroatom doping provides a useful tool to moderate their physical and chemical properties. While nitrogen doped carbons have been extensively researched, the reports for sulfur, and especially nitrogen/sulfur dual-doped materials is far less exhaustive. We present a one-pot synthesis of dual nitrogen/sulfur doped materials which is based on sustainable carbohydrate and amino acid precursors and enables tailoring of the nal materials physicochemical properties. The addition of cysteine gives rise to pending sulfur functionalities, while addition of thienyl-cysteine results in structurally bound sulfur within the carbonaceous framework. Post pyrolysis offers an additional tool for controlling the material stability and results in microporosity as well as superior conductivity relative to undoped carbon microspheres from glucose. Nitrogen doping levels of about 4 wt% and sulfur doping levels of 3 wt% to 12 wt% could be achieved. The samples were, amongst other methods, characterized by XPS, 13 C MAS solid state NMR and both N2 and CO2 sorption isotherms. Since this is the rst report on heteroatom doped, sulfur containing carbon materials obtained via hydrothermal carbonization, we anticipate our work to be a starting point for more sophisticated studies concerning the role of sulfur (and nitrogen) on the physicochemical properties of hydrothermal carbon.

Introduction
With the constantly growing demand for greener technologies, the application of carbon based materials in almost all aspects of daily life has gained much popularity in recent years. In order to access a greater variety of different applications, strategies for the ne-tuning of material properties are required. The introduction of dopants into carbon materials is a useful method of altering their physical and chemical properties. Successful work thus far includes the superior performance of doped carbons in applications such as electrode materials for catalysis13 or energy storage,48 stationary phases,9,10 and chemoselective adsorption.11,12 The majority of research in this eld has focused on nitrogen doping, which is known to induce favorable changes in the carbon material, e.g. increased conductivity13,14 and high activity in the Oxygen Reduction Reaction (ORR) in fuel
Max Planck Institute of Colloids and Interfaces, Am Mhlenberg 1, 14467 Potsdam, Germany. E-mail: stephanie.wohlgemuth@mpikg.mpg.de; Fax: +4-331-56- 9502; Tel: +49-331-567-9562 Electronic supplementary information (ESI) available: Further characterization details, SEM images, elemental analysis data, and XPS data. See DOI: 10.1039/c2gc16415a

cells.13 The synthesis of nitrogen doped carbons has been achieved via a variety of pathways, such as post treatment of carbon with ammonia,15 amines or urea16 and also more direct approaches using acetonitrile,3 pyrrole,17 polyacetonitrile4,13 or polyaniline18 as starting products. Complementing nitrogen as a dopant, sulfur is receiving increasing attention in current carbon materials research. Overall, literature reports suggest that nitrogen is the dopant of preference concerning the tunability of electronic properties of the carbon material, whereas sulfur, due to its large size, has been used more for applications where its easily polarizable lone pairs (and thus chemical reactivity) are of importance. Sulfur doped carbon materials have, for example, shown benecial effects on the selective adsorption of waste metals11 and the desulfurylation of crude oil.12 The synthesis of these sulfur doped materials generally involves the pyrolysis of sulfur containing polymer based carbons,11,12,19 but also arc vaporization in the presence of sulfur containing compounds such as thiophenes.20 Concerning the combined incorporation of sulfur and nitrogen within the same material, only a few reports exist in the literature. Sulfur-assisted growth of carbon nanotubes by chemical vapor deposition of acetonitrile was shown to increase the
Green Chem., 2012, 14, 741749 | 741

This journal is The Royal Society of Chemistry 2012

View Online

nitrogen doping levels as well as the magnetic properties of the nanotubes.21 Choi et al. synthesized heteroatom doped carbon materials by the pyrolysis of amino acid/metal chloride composites. They obtained sulfur doping levels of 2.74 wt% using cysteine and were able to show that materials containing both nitrogen and sulfur increased the materials ORR activity in acidic media, relative to undoped or purely nitrogen doped carbons.22 These results lead us to the conclusion that dual heteroatom-doped carbon materials may be a highly desirable class of materials. Concerning sustainability, most of the aforementioned synthesis methods for heteroatom doped carbon materials show drawbacks in the, often harsh, reaction conditions used. In order to avoid these aspects a hydrothermal carbonization route was chosen. Within the eld of energy efcient and sustainable synthesis of carbon materials, hydrothermal carbonization (HTC) is a meanwhile well-established process. Briey, this technique involves heating ubiquitously available starting materials (carbohydrates, waste biomass) in water in a closed system at mild temperatures (130 C to 200 C) and pressures (ca. 10 bar) and results in their conversion into functional, nanostructured, carbonaceous products. The simplicity of the HTC approach presents an elegant addition to the toolbox of carbon material synthesis with the possibility of tuning the product morphology, structural stability and surface functionality.23,24 While HTC has previously been employed for the synthesis of nitrogen doped materials, to the best of our knowledge, no reports on sulfur doped carbons exist so far. Nitrogen doped carbon microspheres were obtained by hydrothermally treating naturally occurring and nitrogen containing compounds, such as glycine, N-acetyl glucosamine or chitosan.2528 Albumin, a glycoprotein, was used as structure directing additive in the HTC of glucose to produce nitrogen doped carbon aerogels.29 The aim of this work is to extend the sustainable HTC approach towards sulfur and nitrogen multi doped carbon materials with emphasis on the use of cheap and bio-based precursors. Glucose is hydrothermally treated in the presence of cysteine or its derivative thienyl- cysteine. The amino acids serve as nitrogen and sulfur source. Materials with nitrogen contents of about 5 wt% and sulfur contents of 312 wt% can thereby be obtained. We show that the binding state of sulfur and nitrogen can be tuned (from pending groups such as thiols, sulfonates or amides to more structural thiophenic or pyridinic units) by varying the sulfur and nitrogen source. Post pyrolysis of the carbonaceous HTC products allows additional control over material stability and functionality. The pyrolised materials exhibit superior conductivity compared to their undoped counterpart. In addition to exhibiting micropores with CO2 sorption derived surface areas of over 700 m2 g1, a remarkably high sulfur content of more than 7 wt% can be retained even after heat treatment at 900 C. It is expected that the presented synthetic strategy can be extended to various amino acids or proteins, yielding all-sustainable bio-based materials with tunable structural, mechanical, thermal and electronic properties.
742 | Green Chem., 2012, 14, 741749

Experimental methods and sample characterization

Hydrothermal synthesis of the sulfur and nitrogen doped carbon materials
D-(+)-Glucose

was purchased from Roth chemicals. L-Cysteine and S-(2-thienyl)-L-cysteine were purchased from Sigma Aldrich. All chemicals were used as received.

Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

In a typical experiment 1.5 g of glucose and X g (X = 0.1, 0.2, 0.3 or 0.4) of amino acid or its derivative were dissolved in 13.5 g of double distilled H2O. The solution was lled into a glass inlet which was then placed in a Teon lined, stainless steel autoclave (45 ml volume, purchased from Parr Instruments). The autoclave was placed in a pre-heated furnace at 180 C for 5.5 h after which it was allowed to cool down to room temperature. The brown/black solid product was collected by ltration and washed three times with approx. 150 ml H2O and EtOH (until the solution ran clear) before drying in vacuo at 70 C overnight. The samples were labeled according to the following scheme: AAX where AA = Cys for cysteine or TCys for thienyl-cysteine and X denotes the mass (g) of amino acid added. Cys0.2 hence corresponds to an experiment in which 0.2 g of cysteine was added to 1.5 g of glucose. TCys0.2 is the thienyl-cysteine equivalent. The mass yield was between 20 wt% and 30 wt% with respect to the total mass of the starting products, which corresponds to a higher carbon yield (the starting materials contain heavy heteroatoms which are partially lost during carbonization, e.g. as water) and is in the typical range for hydrothermal carbon.41 Note: In order to exclude any contributions from physisorbed molecules on the surface of our materials, Soxhlet extraction was performed on both Cys0.2 and TCys0.2 and the values for elemental analysis compared to those obtained from Cys0.2 and TCys0.2 by washing with EtOH until the ltrate ran clear. Since the results were almost identical, the time consuming Soxhlet extraction step was no longer carried out.
Pyrolysis of Cys0.2 and TCys0.2

The dried sample was placed in a ceramic crucible and covered with a ceramic lid. The crucible was placed in a nitrogen atmosphere furnace and ushed for 30 min before heating to the nal temperature at a heating rate of 10 K min1. The nal temperature was kept for 4 h after which the furnace was allowed to cool down to room temperature. The obtained black powders were characterized without further purication. The sufx _550 or _900 was added to the labels for samples treated at 550 C and 900 C, respectively.
Physical and chemical characterization

Extensive experimental characterization details are provided with the ESI.

This journal is The Royal Society of Chemistry 2012

View Online

Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

Fig. 1 SEM images of solid product after HTC of a) pure glucose, b) glucose with cysteine (Cys0.2) and c) glucose with thienyl-cysteine (TCys0.2). The pure glucose sample was treated at 180 C for 16 h. Both amino acid containing samples were treated at 180 C for 5.5 h. The amino acids catalyze the sphere formation process and give rise to discrete particles with a heterogeneous size distribution. Pure glucose shows homogeneous, interconnected particles.

Results and discussion

Morphology and surface area of the heteroatom doped carbon materials

HTC of glucose at 180200 C typically yields amorphous, nonporous carbon microspheres with a smooth surface which consist of a largely polyfuranic framework and have abundant surface oxygen functional groups (hydroxyl, carboxylic acid).23,24,3032 A simplied reaction mechanism involves the dehydration and subsequent polymerization of the carbohydrate precursors. More detailed explanations of the mechanism have been provided by the groups of Baccile and Titirici et al.32,33 The SEM images in Fig. 1 show, that with the addition of amino acids, a spherical morphology is also obtained. It is interesting to note that, while pure glucose gives rise to interconnected spheres of roughly 1 m in diameter, the amino acid addition results in discrete microspheres which are considerably larger and exhibit a broader size distribution (1 m up to 15 m). This can be attributed Maillard type reactions. Both amines and thiols can react with the aldehyde group of the glucose molecule and give rise to a variety of products which result in the early formation of nucleation seeds.34,35 The complexity of the many reaction pathways that Maillard reactions can undertake implies that nucleation seeds are formed at different points in time. Particle growth according to the la Mer model nally results in larger particles with a heterogeneous size distribution.36 This reaction mechanism also catalyses particle formation which is reected in the signicantly lower reaction times required for the amino acid containing samples (5.5 h) compared to pure glucose (16 h). Pyrolysis does not affect the particle morphology (ESI Fig. S1). In terms of surface area, our materials show no porosity after HTC at 180 C, but exhibit some microporosity after post pyrolysis. At a pyrolysis temperature (Tp) of 550 C, nitrogen adsorption isotherms yield BET surface areas of 440.4 m2 g1 and 281.4 m2 g1 for Cys0.2_550 and TCys0.2_550, respectively (isotherms not shown). This is due to the loss of strongly micropore-bound decomposition products during the heat
This journal is The Royal Society of Chemistry 2012

treatment.29,37 Interestingly, the surface area obtained for TCys0.2_550 is much lower than for Cys0.2_550, indicating that there are inherently less unstable, volatile moieties present. After pyrolysis at 900 C, no nitrogen uptake could be observed. A similar effect was already reported by White et al. who attributed this to micropore closure upon further carbon network condensation.29 However, sorption isotherms obtained with CO2 show that pore shrinkage, rather than complete closure, has occurred. Ultramicropores (i.e. smaller than 1 nm in diameter) giving rise to surface areas of around 630 m2 g1 in the case of Cys0.2_900, and around 730 m2 g1 in the case of TCys0.2_900 were observed (Fig. 2).
Elemental composition

The presence of sulfur and nitrogen was veried by elemental analysis. Table 1 summarizes the elemental composition (in wt %, shown in black) of varying concentrations of a) cysteine and b) thienyl-cysteine added to a constant amount of glucose prior to hydrothermal treatment. As a rst observation, the deviation of the S/N ratio for the amino acid monomers versus the nal carbonaceous products immediately conrms that the amino acids have reacted in an asymmetric fashion. The values imply that not the entire amino acid molecule is incorporated, but occasionally only the sulfur or nitrogen groups. Overall, the trend shows decreasing S/N values with increasing addition of amino acids, indicating the loss of nitrogen containing moieties. Concerning the correlation of the nal sulfur content with the acids initial concentration, a notable difference is seen for the cysteine versus the thienyl-cysteine series. For cysteine, there is a direct correlation, i.e. the more cysteine is added the higher the nal sulfur content. Doping concentrations of 5.66 wt% up to 11.92 wt% of sulfur were obtained. For thienyl-cysteine, on the other hand, the correlation is less pronounced. Over the same concentration variation there is a less than 2 wt% increase in sulfur content, while at the lowest concentration the sulfur doping is already at 10.70 wt% and remains roughly constant throughout the series of experiments. This is a rst indication
Green Chem., 2012, 14, 741749 | 743

View Online

Table 1 Elemental analysis data (wt%) for a) the glucose + cysteine experiment series and b) the glucose + thienyl-cysteine experiment series. The surface elemental composition for Cys0.2, TCys0.2 and the corresponding pyrolised samples at 550 C and 900 C are shown in italic font. The quantities were converted from at % to wt% in order for direct comparison with results from elemental analysis Elemental analysis (wt%) / XPS (wt%) a) Cysteine Sample Cys0.1 Cys0.2 Cys0.3 Cys0.4 Cys0.2_550
Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

Elemental analysis (wt%) / XPS (wt%) S/N 1.71 2.61 2.29 2.50 2.78 1.03 1.44 0.65 1.69 2.29 b) Thienyl-cysteine Sample TCys0.1 TCys0.2 TCys0.3 TCys0.4 TCys0.2_550 TCys0.2_900 Thienyl-cysteinea C 61.61 61.81 70.66 61.65 61.10 76.91 86.84 82.24 88.18 41.32 H 4.32 4.46 4.50 4.45 2.75 0.94 4.43 N 2.64 2.61 2.35 3.52 3.62 3.65 2.23 3.66 6.60 6.89 S 10.70 12.00 3.42 11.41 12.61 11.22 4.08 7.16 3.42 31.55 S/N 4.05 4.60 1.46 3.24 3.48 3.07 1.83 1.96 3.31 4.58

C 62.43 62.94 69.28 60.89 60.73 79.24 84.88 89.01 90.48 29.71

H 4.77 4.81 5.00 4.79 2.93 0.48 5.78

N 3.31 3.73 2.64 3.94 4.28 4.62 3.00 4.53 2.14 11.55

S 5.66 9.72 6.05 9.85 11.92 4.74 4.32 2.95 3.61 26.47

Cys0.2_900
L-Cysteine
a

Labeling scheme: AAX where AA = Cys for cysteine or TCys for thienyl-cysteine and X denotes the mass (g) of amino acid added. The sufx _550 or _900 corresponds to samples pyrolized at 550 C and 900 C, respectively.a Calculated values.

Fig. 2 a) CO2 adsorption and desorption isotherms for Cys0.2_900 and TCys0.2_900, and b) the corresponding pore size distributions. In both cases, ultramicropores of less than 1 nm in diameter are obtained, which is attributed to pore shrinkage upon condensation of the carbonaceous framework during heat treatment.

that the incorporation mechanisms of the two amino acids are inherently different. Samples Cys0.2 and TCys0.2 were chosen as representatives of the cysteine and thienyl-cysteine experiment series, respectively. They were pyrolized at both 550 C and 900 C. The benets of post treating hydrothermal carbon in this way are twofold: Firstly, thermal treatment of the carbonaceous material causes loss of weakly bound groups and rearrangement of the framework into a more organized, condensed structure, resulting in a higher degree of aromaticity. Both material stability (thermal and chemical) and specic conductivity can hence be increased, extending its applicability towards electrochemical or energy storage applications. This directly implies that pyrolysis provides a tool for the tuning of the chemical composition and binding state by controlling the structural rearrangements.
744 | Green Chem., 2012, 14, 741749

Secondly, loss of small organic molecules at high temperatures may result in the formation of micropores, rendering the initially low surface area product microporous, in our case with surface areas of about 280 m2 g1 to 730 m2 g1. For both Cys and TCys, the nitrogen content increases by about 1 wt% during pyrolysis (compared to Cys0.2 and TCys0.2). This is caused by the loss of other atoms such as oxygen, hydrogen and sulfur which increase the relative wt% of nitrogen without additional incorporation. An interesting difference between the two acids is seen regarding the sulfur content. Whilst for cysteine the sulfur content is already halved at Tp = 550 C, the material containing the thienyl derivative shows a sulfur loss of less than 1 wt%. At Tp = 900 C the sulfur content in Cys0.2 is now merely a third of that before pyrolysis while for TCys it is approximately half. These ndings are visualized in the ESI Fig. S2. The material obtained with thienyl-cysteine
This journal is The Royal Society of Chemistry 2012

View Online

Table 2 Peak assignments for the the N 1(s) and S 2(p)3/2 photoelectron envelopes for Cys0.2 and TCys0.2 samples after HTC at 180 C and subsequent pyrolysis at 550 C and 900 C Binding energy (eV) Fraction of species (%) Peak N 1s Cys0.2 399.98 79.8 401.24 20.2 S2p3/2 163.77 81.2 165.95 18.8 Cys0.2_550 398.57 23.9 400.69 62 404.10 14.1 164.26 94.4 168.60 5.6 Cys0.2_900 398.54 27.5 401.25 72.9 164.27 92.6 166.65 7.4 TCys0.2 400.10 77.4 402.29 22.6 164.22 93.6 167.44 6.3 TCys0.2_550 398.85 41.3 400.96 58.7 164.35 100 TCys0.2_900 398.86 26.9 401.29 73.1 164.26 85.3 165.30 14.4 Assignment Pyridinic25,26 Amine25 Pyrrolic26 Quaternary22,26 Pyridinic N-oxide22 Aliphatic -SH. disulde or thioether19,22,38 Thiophenic (aromatic) CSC19,38 Oxidized sulfur19,39

Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

clearly exhibits higher thermal more stability with respect to sulfur loss than the one obtained with cysteine, presumably caused by a difference in the sulfur binding state. Note that our sulfur content for Cys0.2_900 is very similar to that obtained by Choi et al. (2.74 wt%) by the pyrolysis of cysteine in the presence of metal salts.22 For TCys0.2_900 however, more than 7 wt % sulfur can be retained even after heat treatment at 900 C, showing the importance of the sulfur binding state in the as-synthesized carbon material. In order to further investigate the atom binding states of our carbons, XPS was carried out on Cys0.2, TCys0.2 and the corresponding post-treated samples at 550 C and 900 C. The surface elemental composition (converted from at % to wt% for direct comparison with elemental analysis) is shown in grey in Table 1. In contrast to the results from elemental analysis, the values for the sulfur content are similar for Cys and TCys. Since XPS is a surface sensitive technique and elemental analysis applies to the bulk, this indicates that most of the sulfur atoms in TCys are bound within the particle core as structural sulfur. The deconvoluted C1s scans (ESI Table S1) show several peaks with varying contributions at different pyrolysis temperatures. In general, the binding energy value of 285 eV can be assigned to sp2 hybridized carbon in an aromatic environment.19,22,25 The contribution of this peak increases with increasing temperatures. The peaks at 286.00286.38 eV show the second highest contribution and correspond to carbon atoms bound to sulfur, nitrogen or oxygen in the form of thiols, sulfonates, amines or alcohols.22 The binding energy values at 287.30288.35 eV (carbonyl or amide groups),22,37 289.00289.60 eV (ester or carboxylic groups)37 and 191.1 eV (* shake up satellite seen in aromatic systems)12,25 are present in minor contributions while the latter is seen only for the samples treated under the highest pyrolysis temperature of 900 C. Overall, the C1s binding energies conrm that carbonization and aromatization increase after pyrolysis, and that a signicant amount of surface carbon atoms are bound to either sulfur, nitrogen or oxygen.
This journal is The Royal Society of Chemistry 2012

Only a small fraction of the heteroatoms bound to carbon are present in an oxidized state. The peak assignments for the deconvoluted spectra of the N1s (ESI Fig. S3) and the S2p3/2 (ESI Fig. S4) peaks are summarized in Table 2. In the case of nitrogen, a trend of formation of increasingly stable binding states with increasing pyrolysis temperature can be seen, i.e. from amine (399.98 eV),25 pyridinic oxide (402.29404.10 eV)22 and pyrrolic (400.10 eV400.96 eV)26 towards quaternary graphitic (401.24 eV401.29 eV)22,26 and pyridinic (398.54398.86 eV).25,26 The samples treated at 900 C contain only nitrogen bound within the carbon framework (quaternary) or at the edges of the framework ( pyridinic). After HTC at 180 C, Cys0.2 contains mainly surface amine groups (79.8% of nitrogen species), whereas mainly pyrrolic functionalities (77.4% of nitrogen species) were observed. After pyrolysis at both 550 C and 900 C, the nitrogen binding states behave similarly for both amino acids. In the case of sulfur, the trend is not as straightforward. For Cys0.2, the majority of sulfur species are present as aliphatic thiols, disuldes or thioethers (163.77 eV),19,22,38 with the remainder being oxidized sulfur species (165.30 eV168.60 eV).19,39 No thiophenic sulfur (164.26 eV164.35 eV)19,38 could be found. This dramatically changes upon pyrolysis at 550 C and 900 C, whereafter all aliphatic sulfur species disappear and more than 90% of sulfur is bound in a thiophenic fashion. For TCys0.2, most of the surface sulfur is aromatically bound already after HTC at 180 C. After pyrolysis at 550 C, 100% of surface sulfur is bound in a thiophenic state. After pyrolysis at 900 C, some oxidized species appear (14.4% of sulfur species). This may be the result of the reaction of surface sulfur with adjacent oxygen molecules. Overall, XPS analysis suggests, that using cysteine as additive results in pending functional groups (amines, thiols), whereas the addition of thienyl-cysteine results in the incorporation of sulfur into the carbon framework already at early stages of the reaction process. For further insight into the chemical environment of the bulk, solid state 13C MAS NMR spectra were recorded (Fig. 3).
Green Chem., 2012, 14, 741749 | 745

View Online

A spectrum of pure glucose hydrothermally treated at 180 C is shown for reference (Glu_180). General regions of interest can be assigned as follows: carbonyl and carboxylic acid groups at 210 ppm to 175 ppm, furanic groups at 175 ppm to 100 ppm and nally aliphatic groups at <75 ppm.25,32,40 Sp2 hybridized carbon atoms in aromatic rings generally resonate at around 128 ppm130 ppm.32,33,41 More specically, in the Glu_180 spectrum, the peaks at 150 ppm and 110118 ppm correspond to the carbon atoms of furan, labeled C1 and C2, respectively (Fig. 3).41 While the carbonyl regions are similar for Glu_180 and the doped samples, a clear difference in spectra is found in the regions highlighted by dashed grey boxes. The furanic peaks are still present in the doped samples, although the relative intensities are greatly shifted. The resonance at 110118 ppm, which is strong for Glu_180, only appears as a weak shoulder in Cys0.2 and TCys0.2, which directly shows that the contribution of furanic sites has decreased signicantly. The peak at 150 ppm, on the other hand, which in furans should have a similar intensity as the one at 110118 ppm, appears as a very strong, sharp signal in the doped samples. The high intensity can be attributed to both the presence of mobile heteroatom substituents on aromatic rings.4244 This would give rise to the observed peak sharpness due to high mobility. Contributions from thiophene rings and pyrrolic or pyridinic units in the region of 142157 ppm are seen as weak shoulders.4547 An additional strong signal at approximately 130 ppm is found both for Cys0.2 and TCys0.2 which is entirely absent for Glu_180. This all carbon-bound aromatic peak is typical for pyrolized hydrothermal carbon, i.e. after decarboxylation and aromatization have taken place.25,33 However, for the doped materials, this peak is already present before post treatment, indicating that a signicant amount of structural aromatization/pseudo graphitization takes place throughout hydrothermal treatment. This effect has previously been observed for nitrogen-doped hydrothermal carbon obtained from glycine and N-acetyl-glucosamine.25,28 Comparing Cys0.2 and TCys0.2 with each other, it can be seen that aromatization is more pronounced for TCys0.2. In the aliphatic region, a signicant broadening, as well as an upeld shift of the peaks is observed for the amino acid containing samples. The peak at around 25 ppm may refer to aliphatic thiol groups,48 but a simple shift of the methyl/methylene peaks due to a different chemical environment within the doped framework cannot be excluded. The peak observed at 9 ppm seems extremely low and an assignment cannot currently be provided. Qualitatively however, it can be said that the intensity of the aliphatic versus the aromatic region is smaller for the doped samples, implying an overall higher fraction of atoms being present in a condensed structure, rather than bound in aliphates.

Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

Fig. 3 13C MAS solid state NMR spectra for Cys0.2 and TCys0.2 after hydrothermal treatment at 180 C (black). A spectrum obtained from pure glucose hydrothermal carbon is shown for reference (grey). While the carbonyl region is similar for the doped and undoped carbon materials, there is a clear difference in both the aromatic and aliphatic regions, highlighted by dashed grey boxes.

Table 3 Summary of the d-spacing for Cys0.2 and TCys0.2 after HTC at 180 C and post pyrolysis at 550 C and 900 C (from bottom to top) Cys0.2 Tp (C) 550 900 2() 19.9 22.8 23.9 d ()a 4.46 3.90 3.72 TCys0.2 2 () 18.2 21.8 22.8 d ()a 4.87 4.08 3.90

a The Bragg equation, n = 2d sin was used for the calculation, where n is an integer, is the wavelength of the incident wave (Cu radiation, = 1.542 ), d is the spacing between the planes in the atomic lattice, and is the angle between the incident ray and the scattering planes.

Structural analysis

The results obtained from XRD measurements are summarized in Fig. 4 and Table 3. For the samples before post treatment, a broad peak at around 19 is seen, indicative of amorphous materials. After pyrolysis at 550 C, two peaks at roughly 22 and 41 appear which shift to around 23 and 44 after pyrolysis at 900 C, respectively. While these values are in a similar range as the values for graphitically ordered carbons (26 and 43,
746 | Green Chem., 2012, 14, 741749

corresponding to the 002 and 100 lattice spacing),49 they also indicate that the interlayer spacing of the sulfur and nitrogen doped materials are much larger than in graphite. We have observed a similar increase in the interlayer spacing for undoped hydrothermal carbons, showing that the material obtained contains many structural defects even after post carbonization. The point of interest here, however, is the considerable increase in the d-spacing of the TCys materials compared to the Cys materials. This effect is already known for nitrogen doped carbon.6 For sulfur doping, the interlayer distance is expected to increase even more due to the large size of the sulfur atom. Glenis et al. observed an interlayer expansion of 0.048 after S doping into graphite (from 3.357 to 3.405 ).20 In their case, the sulfur atoms were located between the graphite layers, whereas in our material, sulfur is incorporated into the carbon framework. This is expected to cause buckling of the carbon layers and give rise to an even larger increase in the d-spacing, as is indeed observed. Also note the increased intensity at lower angles for the samples after pyrolysis at 900 C. This is
This journal is The Royal Society of Chemistry 2012

View Online

Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

Fig. 4 XRD diffractograms for Cys0.2, Cys0.2_550, Cys0.2_900 (solid black line) and overlay plots of TCys0.2, TCys0.2_550 and TCys0.2_900 (dashed grey lines). At all three temperatures, an increased d-spacing is observed for the TCys0.2 sample relative to Cys0.2, indicating higher levels of structural sulfur incorporation.

Table 4 Specic conductivity for pure glucose, Cys0.2 and TCys0.2 after pyrolysis at 550 C and 900 C. Post pyrolysis of the HTC materials results in condensation of the carbonaceous framework, increasing conductivity. At Tp = 900 C, the doped samples show a superior conductivity compared to the pure glucose reference Specic conductivity (S m1) Tp (C) 550 900 Glu 5.7 832.0 Cys0.2 6.7 10 2656.7
4

Fig. 5 TGA curves for Cys0.2 (solid) and TCys0.2 (dashed) after HTC at 180 C (black) and post pyrolysis at 550 C (dark grey) and 900 C (light grey). After HTC at 180 C and post pyrolysis at 550 C, TCys0.2 contains more stable binding motifs compared to Cys0.2. Note the reversal of stability at Tp = 900 C which is attributed to a more nitrogen doped graphitic framework in Cys0.2_900 with little sulfur groups. This condensed backbone is now more stable than that of TCys0.2_900, which contains more than 7 wt% sulfur, most likely bound in a thiophenic fashion.

TCys0.2 2.0 105 2018.6

indicative of the presence of pseudo-periodically ordered micropores which cause X-ray diffraction in a similar manner as crystal lattices and has been observed for pyrolized hydrothermal carbon before.25,26 These ndings are hence coherent with the sorption data shown earlier.

Thermal stability and conductivity of the doped materials

The structural ndings presented earlier also reect in altered material properties of the doped carbons. TGA curves for Cys0.2 and TCys0.2 after HTC at 180 C, as well as after pyrolysis at 550 C and 900 C are shown in Fig. 5. While TCys0.2 and TCys0.2_550 shows a higher thermal stability than Cys0.2 and Cys0.2_550, the relationship is reversed for Cys0.2_900 and TCys0.2_900. XPS analysis showed that at lower temperatures, TCys0.2 contains more stable, aromatic structures such as thiophenic and pyridinic/pyrollic structures, as opposed to amine/ thiol functionalities in Cys0.2. At 900 C, the carbonaceous framework is well developed in Cys0.2_900, whereas TCys0.2_900 still contains many thiophenic units which are now less stable compared to benzene units. Since the nitrogen content is similar for both Cys0.2_900 and TCys0.2_900, the lower thermal
This journal is The Royal Society of Chemistry 2012

stability of the latter is ascribed to the high structural sulfur content, determined by elemental analysis. This trend is also nicely reected in the conductivity values for the obtained materials (Table 4). Up to calcination temperatures of 550 C, the conductivity is still low, as the occurring rearrangements were insufcient to create a cohesive aromatic system (which is required for good electrical conductance). This clearly points to the fact that the observed aromatization throughout HTC in the presence of amino acids is an isolated, tectonic process. Once pyrolized at 900 C however, the conductivity of the doped materials is more than three times higher for Cys0.2_900 than for Glu_900. We attribute the slightly improved conductivity of Cys0.2_900 over TCys0.2_900 to the even increased sulfur content of TCys0.2_900. It has been reported that sulfur can act as n-type dopant in graphite and diamond semiconductors and Kumari et al. showed that sulfur doping increases the conductivity of amorphous carbon lms at optimum concentrations, but causes a decreased conductivity if present in excess.50,51 Similar effects have been reported for nitrogen doping. The conductivity dependence on the dopant concentration can be explained by a competition of two processes taking place. On the one hand, the incorporation of dopants results in a disordered graphitic lattice while, on the other hand, additional charge carriers are introduced into the system.52 A maximum in conductivity is therefore observed when a compromise between the two effects is reached. We therefore postulate that, while the nitrogen content in our materials is in the correct range to increase conductivity, the sulfur content in the TCys0.2_900 is too high, hence causing a slight decrease in charge carrier availability. However, the specic conductivity values are still superior compared to undoped carbon, rendering electronic or electrochemical applications feasible.
Green Chem., 2012, 14, 741749 | 747

View Online

Conclusion
The incorporation of sulfur and nitrogen in carbon microspheres via a sustainable hydrothermal synthesis using bio-based molecules, as well as the tunability of the heteroatom binding states was successfully demonstrated. An amino acid and its derivative (cysteine and thienyl-cysteine, respectively) were chosen as sulfur and nitrogen source. They could very effectively be incorporated into the carbon framework by nucleophilic addition within so-called Maillard schemes. Addition of cysteine gave rise to a material with more pending functional groups such as thiols, amides and sulfonates. Addition of thienyl-cysteine, on the other hand, resulted in more structurally bound sulfur, such as in thiophenes. In both cases, the nitrogen binding motifs moved from less stable, pending groups towards more stable structural motifs ( pyridines, pyrroles, and graphitic nitrogen) with increasing pyrolysis temperatures. After pyrolysis at 900 C, the resulting materials showed almost three times higher specic conductivity than a corresponding undoped sample made from pure glucose, as well as a highly increased interlayer distance of, presumably, buckled carbon sheets. The sulfur binding motif is relevant with respect to chemical, mechanical and electrical properties arising from different heteroatom binding states. Pending surface thiol groups could, for example, be useful for the adsorption of metal nanoparticles in a catalyst support material, but are expected to be dominant for the low temperature cysteine series, whereas structurally bound sulfur and nitrogen are known to alter electronic properties such as conductivity.20 Bandosz et al. highlighted the importance of distinguishing between structural (CSC type) and pending (thiol or sulfonate) sulfur in the selective arsine adsorption capacities of sulfur containing carbons.11 Since in situ monitoring of our closed reaction system at 180 C and 10 bar is difcult to carry out, we can only postulate on the pathway of heteroatom incorporation. Both amino acids contain a free amine group which can react in a similar fashion, and the difference in sulfur binding states of CysX versus TCysX must be attributed to the reactivity of the corresponding sulfur functional group. Unlike the thiol group in cysteine, the thienyl unit rather reacts via a cycloaddition mechanism with various intermediates in the HTC mechanism, creating structural sulfur already at early reaction stages. This explains why nitrogen shows a similar trend for both CysX and TCysX materials whereas sulfur does not. It should be noted that thiophene itself is not a good DielsAlder substrate due to its high level of aromaticity. We expect that the synthesized materials have potential for a variety of applications where the mechanical, chemical and electronic properties of the carbon material are important. More specically, the testing of the herein presented materials for heavy and noble metal adsorption, as well as introduction of surface area and subsequent application for electrode materials in the oxygen reduction reaction, are matter of on-going investigation.

University of Vigo. Dr Niki Baccile from the University Pierre et Marie-Curie in Paris is acknowledged for assistance with the solid state NMR measurements. We are also grateful to Dr Hiromitsu Urakami for his contributions to both scientic discussions, and life.

Notes and references

1 K. P. Gong, F. Du, Z. H. Xia, M. Durstock and L. M. Dai, Science, 2009, 323, 760764. 2 K. A. Kurak and A. B. Anderson, J. Phys. Chem. C, 2009, 113, 6730 6734. 3 P. H. Matter, E. Wang, M. Arias, E. J. Biddinger and U. S. Ozkan, J. Phys. Chem. B, 2006, 110, 1837418384. 4 T. Iijima, K. Suzuki and Y. Matsuda, Synth. Met., 1995, 73, 920. 5 Y. P. Wu, S. B. Fang and Y. Y. Jiang, J. Mater. Chem., 1998, 8, 2223 2227. 6 Y. P. Wu, S. B. Fang, Y. Y. Jiang and R. Holze, J. Power Sources, 2002, 108, 245249. 7 E. Frackowiak, Phys. Chem. Chem. Phys., 2007, 9, 17741785. 8 G. Lota, B. Grzyb, H. Machnikowska, J. Machnikowski and E. Frackowiak, Chem. Phys. Lett., 2005, 404, 5358. 9 C. West, C. Elfakir and M. Lafosse, J. Chromatogr., A, 2010, 1217, 32013216. 10 G. W. Muna, V. M. Swope, G. M. Swain and M. D. Porter, J. Chromatogr., A, 2008, 1210, 154159. 11 C. Petit, G. W. Peterson, J. Mahle and T. J. Bandosz, Carbon, 2010, 48, 17791787. 12 M. Seredych, M. Khine and T. J. Bandosz, ChemSusChem, 2011, 4, 139 147. 13 S. H. Lim, H. I. Elim, X. Y. Gao, A. T. S. Wee, W. Ji, J. Y. Lee and J. Lin, Phys. Rev. B, 2006, 73. 14 D. P. Kim, C. L. Lin, T. Mihalisin, P. Heiney and M. M. Labes, Chem. Mater., 1991, 3, 686692. 15 F. Jaouen, M. Lefevre, J. P. Dodelet and M. Cai, J. Phys. Chem. B, 2006, 110, 55535558. 16 R. Pietrzak, H. Wachowska and P. Nowicki, Energy Fuels, 2006, 20, 12751280. 17 S. Glenis, A. J. Nelson and M. M. Labes, J. Appl. Phys., 1996, 80, 5404 5407. 18 L. Li, E. Liu, Y. Yang, H. Shen, Z. Huang and X. Xiang, Mater. Lett., 2010, 64, 21152117. 19 J. P. Paraknowitsch, A. Thomas and J. Schmidt, Chem. Commun., 2011, 47, 82838285. 20 S. Glenis, A. J. Nelson and M. M. Labes, J. Appl. Phys., 1999, 86, 4464 4466. 21 T. X. Cui, R. Lv, Z. H. Huang, F. Y. Kang, K. L. Wang and D. H. Wu, Nanoscale Res. Lett., 2011, 6, 77. 22 C. H. Choi, S. H. Park and S. I. Woo, Green Chem., 2011, 13, 406412. 23 B. Hu, K. Wang, L. H. Wu, S. H. Yu, M. Antonietti and M. M. Titirici, Adv. Mater., 2010, 22, 813828. 24 M. M. Titirici and M. Antonietti, Chem. Soc. Rev., 2010, 39, 103116. 25 Z. Li, N. Baccile, S. Gross, Z. Yuanjian, W. Wei, S. Yuhan, M. Antonietti and M. M. Titirici, Carbon, 2010, 48, 377837873787. 26 R. J. White, M. Antonietti and M. M. Titirici, J. Mater. Chem., 2009, 19, 86458650. 27 N. Baccile, M. Antonietti and M.-M. Titirici, ChemSusChem, 2010, 3, 246253. 28 N. Baccile, G. Laurent, C. Coelho, F. Babonneau, L. Zhao and M.-M. Titirici, J. Phys. Chem. C, 2011, 115, 89768982. 29 R. J. White, N. Yoshizawa, M. Antonietti and M.-M. Titirici, Green Chem., 2011, 13, 24282434. 30 B. Hu, S. H. Yu, K. Wang, L. Liu and X. W. Xu, Dalton Trans., 2008, 54145423. 31 M. M. Titirici, A. Thomas and M. Antonietti, New J. Chem., 2007, 31, 787789. 32 M. M. Titirici, M. Antonietti and N. Baccile, Green Chem., 2008, 10, 12041212. 33 N. Baccile, G. Laurent, F. Babonneau, F. Fayon, M.-M. Titirici and M. Antonietti, J. Phys. Chem. C, 2009, 113, 96449654. 34 C. Billaud, C. Maraschin, M. N. Peyrat-Maillard and J. Nicolas, in Maillard Reaction: Chemistry at the Interface of Nutrition, Aging, and

Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

Acknowledgements
We would like to acknowledge Sylvia Pirok for elemental analysis, and Rona Pitschke and Heike Runge for SEM measurements. XPS analysis was carried out by Carmen Serra at the
748 | Green Chem., 2012, 14, 741749

This journal is The Royal Society of Chemistry 2012

View Online

35 36 37 38 39 40 41

Downloaded by Max Planck Institut fur Kolloid on 07 March 2012 Published on 18 January 2012 on http://pubs.rsc.org | doi:10.1039/C2GC16415A

42 43

Disease, ed. J. W. Baynes, V. M. Monnier, J. M. Ames and S. R. Thorpe, New York Acad Sciences, New York, 2005, vol. 1043, pp. 876885. C. Billaud, C. Maraschin, Y. N. Chow, S. Cheriot, M. N. Peyrat-Maillard and J. Nicolas, Mol. Nutr. Food Res., 2005, 49, 656662. V. K. Lamer and R. H. Dinegar, J. Am. Chem. Soc., 1950, 72, 4847 4854. M. Sevilla and A. B. Fuertes, Chem.Eur. J., 2009, 15, 41954203. B. J. Lindberg, K. Hamrin, G. Johansson, U. Gelius, A. Fahlman, C. Nordling and K. Siegbahn, Phys. Scr., 1970, 1, 286298298. U. Gelius, C. J. Allan, G. Johansso, H. Siegbahn, D. A. Allison and K. Siegbahn, Phys. Scr., 1971, 3, 237. R. Demir-Cakan, N. Baccile, M. Antonietti and M.-M. Titirici, Chem. Mater., 2009, 21, 484490. C. Falco, N. Baccile and M.-M. Titirici, Green Chem., 2011, 13, 3273 3281. J. D. Mao, B. Xing and K. Schmidt-Rohr, Environ. Sci. Technol., 2001, 35, 19281934. B. R. Sinha, F. D. Blum and D. Oconnor, J. Appl. Polym. Sci., 1989, 38, 163171.

44 G. Erbatur, O. Erbatur, M. F. Davis and G. E. Maciel, Fuel, 1986, 65, 12651272. 45 T. Narasimhaswamy, N. Somanathan, D. K. Lee and A. Ramamoorthy, Chem. Mater., 2005, 17, 20132018. 46 L. Frydman, A. C. Olivieri, L. E. Diaz, A. Valasinas and B. Frydman, J. Am. Chem. Soc., 1988, 110, 56515661. 47 B. Nowacki, E. R. deAzevedo and L. Akcelrud, Polym. Test., 2011, 30, 342347. 48 A. Abraham, E. Mihaliuk, B. Kumar, J. Legleiter and T. Gullion, J. Phys. Chem. C, 2010, 114, 1810918114. 49 Z. Q. Li, C. J. Lu, Z. P. Xia, Y. Zhou and Z. Luo, Carbon, 2007, 45, 16861695. 50 I. Sakaguchi, M. N. Gamo, Y. Kikuchi, E. Yasu, H. Haneda, T. Suzuki and T. Ando, Phys. Rev. B, 1999, 60, R2139R2141. 51 L. Kumari and S. V. Subramanyam, Appl. Phys. A: Mater. Sci. Process., 2009, 95, 343349. 52 Z. R. Ismagilov, A. E. Shalagina, O. Y. Podyacheva, A. V. Ischenko, L. S. Kibis, A. I. Boronin, Y. A. Chesalov, D. I. Kochubey, A. I. Romanenko, O. B. Anikeeva, T. I. Buryakov and E. N. Tkachev, Carbon, 2009, 47, 19221929.

This journal is The Royal Society of Chemistry 2012

Green Chem., 2012, 14, 741749 | 749