Graphite functionalization by ball milling with sulfur
- 137 Downloads
The present study reports on ball-milling of high-surface-area graphite in the presence of sulfur, mainly considering high graphite/sulfur (G/S) weight ratios. Graphitic crystallites after co-milling with sulfur maintain essentially unaltered crystalline order while exhibit heavily functionalized lateral edges, mainly by O=S=O and S=O groups. Solvent extraction procedures and thermogravimetric experiments show that co-milled graphite/sulfur samples contain two different kinds of sulfur: an extractable one being only physically adsorbed on graphite (exhibiting a higher thermal stability) and a non-extractable fraction chemically bonded to defective graphitic sites (exhibiting a lower thermal stability). The amount of sulfur covalently bonded to graphite crystallites increases with the co-milling time up to a limit value, which is in the range 3.0–3.5 wt% and is poorly dependent on the starting G/S ratio.
KeywordsX-ray diffraction Edge functionalization Thermogravimetric analysis Solvent extraction
For many graphite and graphene applications, mainly for those needing high electrical conductivity, the presence of ordered graphitic structure with functional groups being confined to the graphitic edges, is often preferred [1, 2, 3, 4, 5, 6].
Graphite functionalization by ball-milling of graphite with gases (CO2, N2, F2 and Cl2), liquids (SO3, Br2) and solids (I2, maleic anhydride and elemental sulfur) have been described by many reports. These low-cost procedures are environmentally sustainable and can also lead to selective functionalization of graphite edges by different functional groups [2, 7, 8, 9, 10, 11, 12, 13].
In particular, by ball-milling of graphite in the presence of sulfur it was demonstrated that sulfur could edge-selectively functionalize graphene nanoplatelets with a concomitant homogeneous adsorption on and within the graphitic crystallites [14, 15].These literature studies have been conducted by using excess of sulfur with respect to graphite, (maximum G/S ratio equal to 1) and showed that only a low fraction of sulfur is covalently bonded to graphite while most of sulfur is simply adsorbed by graphite, even for long term milling [14, 15].
Graphene nanoplatelets edge-functionalized with sulfur have been proposed as highly efficient lithium-sulfur batteries cathode materials [14, 15] and could be in principle suitable as reinforcing and reactive agents for polymer composites.
The aim of the present study is to explore the possibility to achieve graphite functionalization by using much smaller amounts of sulfur and to establish a simple method to discriminate between bound and adsorbed sulfur.
2 Experimental section
High surface area graphite, exhibiting a surface area of 330 m2/g, was obtained from Asbury Graphite Mills Inc.
Ball-milling was conducted by a “FRITSCH Pulverisette 7 Premium Line” with spinning rate of 400 rpm.
Sulfur extraction was conducted by treatment with liquid chloroform followed by powder desiccation at 60 °C.
Thermogravimetric analyses (TGA) were conducted by a TA-Instruments Q500, under nitrogen flow of 50 mL/min, with a heating rate of 10 °C/min, up to 800 °C.
FTIR spectra were obtained with a FTIR (BRUKER Vertex70) spectrometer equipped with deuterated triglycine sulfate (DTGS) detector and a KBr beam splitter, at a resolution of 2.0 cm−1. The frequency scale was internally calibrated to 0.01 cm−1 using a He–Ne laser. 32 scans were signal averaged to reduce the noise. Spectra of powder samples were collected by using potassium bromide (KBr) pellets.
Differential Scanning Calorimetry (DSC) scans were collected by a DSC Q2000 provided by (TA instruments).
Surface areas of carbon and oxidized carbon samples were measured by nitrogen adsorption at liquid nitrogen temperature (77 K) with a Nova Quantachrome 4200e instrument. Before adsorption measurements, samples were degassed at 60 °C under vacuum for 24 h. The surface area (SABET) values were determined by using 11-point BET analysis.
Elemental analysis was performed with a Thermo FlashEA 1112 Series CHNS-O analyzer, after pre-treating samples in an oven at 100 °C for 12 h, by using BBOT as standard.
3 Results and discussion
The first two sections of this paper are devoted to ball-milling procedures on as received graphite and sulfur samples. The subsequent section deals with co-milling of graphite and sulfur, by considering mixtures with prevailing content of graphite.
3.1 Milling of graphite
After short-term milling (e.g., after 10 min, blue curves in Fig. 1A, B) 002 and 004 reflections, corresponding to crystal planes parallel to the graphitic planes, become definitely more intense than for the starting graphite (black curves in Fig. 1A, B). This result can be rationalized by considering that short term milling induces an orientation of graphite platelets that is preferentially parallel to the plane of the sample holder of the diffractometer. This orientation leads to an apparent increase of intensities of 00 l reflections, as collected by an automatic powder diffractometer.
The FWHM of 002 (Fig. 1A) and 004 (Fig. 1B) reflections increase with milling time and correspondingly a remarkable reduction of the correlation length perpendicular to the graphitic planes, down to 4 nm, is evaluated (Fig. 1C). This indicates that long term milling leads to a significant graphite delamination.
Surface area and elemental analysis of graphite after different milling times
Milling times (h)
Informative on the milling process are also surface area values, as evaluated by BET measurements (2nd column of Table 1). After 1 h of milling there is only a small reduction of SA from 330 m2/g down to nearly 300 m2/g. For long term milling (20 h), the surface area becomes lower than 10 m2/g. This phenomenon is possibly due to agglomeration of nanostructured particles, as already known for ball milling of graphite .
The formation of a significant amount of oxygenated groups definitely reduces thermal stability of graphite. This is shown for instance by TGA scans, like those shown in Fig. 2B. It is apparent that 10 min of milling (blue scan) do not change the thermal stability of graphite (black scan) while the decrease of thermal stability is remarkable after 20 h of milling. Moreover, the increase of weight loss below 100 °C with long term ball-milling (Fig. 2B) clearly indicates increase of hydrophilicity.
3.2 Milling of sulfur
DSC scans of the untreated sulfur sample (black curve in Fig. 3B) show two endothermic phenomena: a lower one at 107 °C with ΔHt ≈ 12 J/g, corresponding to the orthorhombic → monoclinic transition [23, 24] and a higher one at 120 °C with ΔHm ≈ 47 J/g, corresponding to the melting of the monoclinic phase. A DSC scan of the sulfur sample after 20 h of milling (green curve in Fig. 3B) again shows two well-separated endothermic phenomena, but the lower peak becomes much broader and centered at lower temperature (100 °C), with only slightly reduced transition enthalpy (ΔHt ≈ 10 J/g).
This confirms that long term ball milling reduces the crystalline order of the orthorhombic phase thus reducing the temperature of the orthorhombic → monoclinic transition. The unaltered melting enthalpy (ΔHm ≈ 47 J/g) clearly indicates that the long-term ball milling does not produce chemical modification of sulfur.
3.3 Co-milling of graphite and sulfur
WAXD patterns of samples milled for shorter times and G/S = 10/1 are shown in Fig. 4B. It is apparent that already after 10 min of co-milling crystalline peaks of sulfur have nearly completely disappeared (only a weak 222 reflection, at 2θ ≈ 23°, is still detectable, purple pattern in Fig. 4B). Also, this latter reflection disappears after 1 h of co-milling (red pattern in Fig. 4B).
The effect of co-milling with sulfur on the crystalline structure of graphite is nearly negligible for G/S ≥ 10/1 (upper patterns of Fig. 4A). As the sulfur content increases, the 002 peak corresponding to the interlayer graphitic distance, becomes broader and broader. For G/S = 1/1, after removal of the crystalline sulfur by extraction by chloroform, the correlation length perpendicular to the graphitic planes, as evaluated by the Sherrer method decreases from ≈ 10 nm to ≈ 3 nm.
To understand the origin of the two DTG peaks associated with sulfur loss, co-milled samples have been subjected to different solvent extraction procedures. Gravimetric measurements show that G/S = 20/1 mixtures (S content of 4.8%), co-milled for 20 h and 30 h, present a weight loss after extraction by chloroform of nearly 2% and 1.5%, respectively. Elemental analyses of G/S 20/1 co-milled for 20 h and 30 h, show that the sulfur content in the extracted samples is close to 2.9% and 3.5%, respectively (Table S1 of Supplementary Information), in satisfactory agreement with the weight loss data. Similar contents of bonded sulfur were obtained also for mixtures with much larger sulfur content.
Particularly informative are DTG scans of Fig. 7B, D, showing that the solvent extraction involves essentially only sulfur whose thermal loss occurs at higher temperature (corresponding to the DTG peak centered in the range 325–350 °C). In fact, the lower temperature DTG peak, roughly centered at 230 °C, remains nearly unaltered.
These results can be easily rationalized by assuming that the higher temperature weight losses of Figs. 6 and 7 correspond to sulfur absorbed on graphene surfaces while the lower temperature weight losses correspond to degradation of covalently bound sulfur. The amount of covalently bound sulfur, possibly located on edges of graphitic planes, increases by increasing ball milling time (Fig. 6).
Although these bands remain broad for co-milled G/S samples, for high sulfur content (upper spectrum of Fig. 8A) and for long term ball milling (upper spectra of Fig. 8B) better defined peaks, centered in the 1130–1125 cm−1 and 1085–1080 cm−1 ranges, possibly associated with stretching of O=S=O and S=O bonds, respectively, are apparent. It is also possible to exclude the presence of S–H bonds due to the absence of the typical bands in the 2600–2540 cm−1 range .
A comparison between FTIR spectra of three co-milled samples before and after chloroform extraction is reported in Fig. 8B. It is apparent that extraction of absorbed sulfur does not change significantly the observed FTIR spectra.
For a more precise definition of the sulfur containing species, FTIR identification of species released by thermogravimetric measurements could be included in future studies.
Crystallinity of graphite is poorly affected by ball-milling procedures, with only a reduction below 5 nm of correlation length perpendicular to the graphitic crystallites, for long milling times. This behavior is not changed by co-milling with sulfur. Sulfur, on the contrary, as a consequence of milling with graphite, becomes fully amorphous and time needed to a complete loss of crystal order markedly decreases with the graphite/sulfur weight ratio, becoming of only few minutes for G/S ≥ 20.
However, even for G/S ratios as high as 20, relevant functionalization of graphitic edges is achieved, mainly occurring as O=S=O and S=O groups.
Solvent extraction procedures and thermogravimetric measurements show that co-milled graphite/sulfur samples contain two different kinds of sulfur: (A) an extractable one only physically adsorbed on graphite, exhibiting a higher thermal stability, with DTG peak close to 350 °C; (B) a non-extractable fraction chemically bonded to defective (mainly edge) graphitic sites, exhibiting a lower thermal stability, with DTG peak close to 230 °C.
The amount of sulfur covalently bonded to graphite crystallites increases with the co-milling time up to a limit value, which is in the range 3.0–3.5 wt% with respect to graphitic carbon and is poorly dependent on the used G/S ratio.
Highly crystalline graphite as functionalized by co-milling with sulfur could be considered as reinforcing and reactive filler for polymer composites.
Compliance with ethical standards
Conflict of interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
- 24.Černosek Z, Holubova J, Černoskova E, Ruzicka A (2009) Sulfur—a new information on this seemingly well-known element. J Non Oxide Glasses 1:38–42Google Scholar
- 26.Lin-Vien D, Colthup NB, Fateley WG, Grasselli JG (1991) The handbook of infrared and Raman characteristic frequencies of organic molecules. Elsevier, AmsterdamGoogle Scholar