Direct Evidence of the Exfoliation Efficiency and Graphene Dispersibility of Green Solvents toward Sustainable Graphene Production

Achieving a sustainable production of pristine high-quality graphene and other layered materials at a low cost is one of the bottlenecks that needs to be overcome for reaching 2D material applications at a large scale. Liquid phase exfoliation in conjunction with N-methyl-2-pyrrolidone (NMP) is recognized as the most efficient method for both the exfoliation and dispersion of graphene. Unfortunately, NMP is neither sustainable nor suitable for up-scaling production due to its adverse impact on the environment. Here, we show the real potential of green solvents by revealing the independent contributions of their exfoliation efficiency and graphene dispersibility to the graphene yield. By experimentally separating these two factors, we demonstrate that the exfoliation efficiency of a given solvent is independent of its dispersibility. Our studies revealed that isopropanol can be used to exfoliate graphite as efficiently as NMP. Our finding is corroborated by the matching ratio between the polar and dispersive energies of graphite and that of the solvent surface tension. This direct evidence of exfoliation efficiency and dispersibility of solvents paves the way to developing a deeper understanding of the real potential of sustainable graphene manufacturing at a large scale.


Shear mixing exfoliation
Sigma-Aldrich graphite (SA, 100 mesh) and Alfa-Aesar graphite (AA, 300 mesh) were exfoliated with a concentration of 25 mg/mL in a range of non-toxic, low boiling point solvents (D.I., EtOH, MeOH, EA, Ace, IPA) and solvent mixtures [(EtOH:D.I. (1:1 vol. %), IPA:Ace (1:1 vol. %)] using a shear mixer at 5000 rpm for 3 hours. The total volume of 150 ml was kept constant during shear mixing by 'topping up' with the relevant solvent. After exfoliation, the samples were centrifuged at 500 rpm for 1 hour to separate the exfoliated graphene from large graphite aggregates for characterisation of the supernatant.

NMP-Redispersion and Green Solvent-Redispersion
NMP-Redispersion: All samples exfoliated in green solvents (see list in 1.2.1) were dried at 100 o C in a vacuum oven to evaporate the solvent after exfoliation. NMP (equal volume as the evaporated amount) was added.

GS-Redispersion:
All samples exfoliated in NMP were washed with ethanol using vacuum filtration with 0.22 micron PTFE membrane before drying them in a vacuum oven at 100 o C to evaporate the remaining solvent. Same volume of green solvents were added to disperse the NMP exfoliation products. For both NMP and GS-R, a vortex mixer was used to homogenise the samples to eliminate possible effects of ultrasonic cavitation. After homogenisation, the samples were centrifuged at 500 rpm for 1 hour and the supernatant was taken for concentration analysis. (*Exfoliation products are the graphene-containing products with graphite aggregates obtained after exfoliation and before centrifugation.)

Characterisation of graphite and graphene
The samples were characterised by means of scanning electron microscopy (SEM), Brunauer-Emmet-Teller (BET), Raman spectroscopy, UV-vis spectroscopy, and atomic force microscopy. Sample preparation and operating conditions are described below:

Scanning Electron Microscopy
The graphite was sparingly sprinkled onto sticky copper tape for SEM analysis of the raw materials. SEM micrographs were taken at 5 kV, using JEOL JSM-840F with cold cathode field emission gun.

Brunauer Emmet Teller
Brunauer Emmet Teller (BET) measurements were performed under nitrogen atmosphere (nitrogen adsorptiondesorption isotherm at 77K) using a Micromeritics' Gemini Vll . The samples were degassed at 225 ˚C overnight under nitrogen prior to the measurement. Sample mass of 0.3228 g was used on all the samples.

Raman spectroscopy
The graphene dispersions were dropcast onto a Si/SiO 2 wafer and dried under ambient air for Raman spectroscopy. For the graphene samples dispersed in NMP, the graphene dispersion was dropcast onto a Si/SiO 2 wafer, then dried at 150 o C on a hotplate. The Raman D, D' and G peaks were fitted using a Lorentzian function. Raman Spectroscopy was performed using the Horiba Scientific LabRAM ARAMIS. An objective lens of x50, laser wavelength of 532 nm and laser intensity of 10 % were used to acquire the Raman spectra.

UV-Vis spectroscopy
The UV-Vis spectra were acquired using Varian Cary 5000. The concentration of graphene dispersion was calculated using the Lambert-Beer Law (Eq. 1) A absorbance, is the molar absorptivity [ml/mg ], = • m is the optical path length (0.01 m) and is the concentration (mg/ml) of the attenuating species (graphene). The molar absorptivity, was determined using the calibration curves (see 4).

Atomic Force Microscopy (AFM)
An Agilent 5400 AFM was used to evaluate the sample morphology and graphene flake thickness. The samples were prepared by drop casting graphene solutions on a Si/SiO 2 wafer placed on a hot plate, heated to 80 o C (for green solvents) and to 150 o C (for NMP). Absorptivity values obtained for graphene dispersed in NMP ( Figure S2a) and D.I. water (Figure S2b). The lower the absorptivity value, the higher the calculated graphene concentration.

Figure S2c
illustrates dispersed graphene (of equal mass) in solvents with different dispersibility. It is visible that graphene is less well dispersed in EtOH and D.I.; i.e., their ability to disperse graphene is lower which leads to low absorbance and thus lower concentration for a given mass of graphene. Figures S3-S4 show the concentration of GR150 and GR50 (produced from the same mass of G150 and G50 respectively) in different solvents.

Fig. S3: (a) and (b) revealing the concentrations of as produced GR150 and GR50 in different solvents. (b)'Close-up' of (a)
to reveal the differences between the solvents with NMP off scale.

Interfacial contact angle measurement (Washburn method) and surface energy analysis
The interfacial contact angle (θ) measurement for graphite flakes (G150) and graphite powder (G50) in selected test liquids (exfoliation solvents used in this work) was done by means of KRUSS Tensiometer K100. Hexane was used to determine the capillary constant. The contact angle was calculated using the following equation: (eq. 2) The viscosity, surface tension, and density were obtained from KRUSS instrument database (Table S1) apart from the solvent mixtures of EtOH:D.I (1:1) and IPA:Ace (1:1) which were taken from the literature as referenced:  [4] 0.7808 [5] 0.308 [6]  The values used for polar and dispersive components of solvent surface tension for graphite surface energy calculation are listed as follows: 10 Raman spectroscopy of IPA-and NMP-exfoliated graphene in comparison to the commercial graphene Fig. S9: Comparison of IPA-and NMP-exfoliated GR150 Raman spectra with commercial graphene. Two sepctra are collected for each sample. The Raman spectra are normalised to the G peak. D and 2D peaks give information on the defect and layer number. There is no obvious difference in the D and 2D peaks between commercial, IPA-and NMP-exfoliated graphene. The type of exfoliation solvents used has no significant effect on the graphene quality.