The feasibility of using organosilanes as effective liquid dispersing media for graphene was investigated. In particular, PhTES and GPTMS were tested with the aim of preparing graphene dispersions that could be directly utilized for further uses without recovering graphene in the solid state, thus avoiding any possible restacking of it to pristine graphite and compromising any previous successful exfoliation process.
As described in ‘Methods’ section, the procedure used was as simple as possible and envisaged the direct sonication of graphite without any chemical manipulation. A first indication of the nanometric dimensions of the dispersed graphene particles was provided by the occurrence of the Tyndall effect
]. Both GPTMS (Figure
) and PhTES dispersions exhibit graphene light scattering, thus confirming the colloidal nature of these systems.
Tyndall effect exhibited by a graphene dispersion in GPTMS. When the red laser light passes through the dispersion, it is scattered and becomes visible.
With the aim of determine the best initial graphite/liquid medium ratio that allows obtaining the highest graphene concentration, a series of dispersions with different amounts of graphite was prepared for each of the two organosilanes. In order to find the absorption coefficient α and set-up a reliable method for the determination of graphene concentration in the above media, UV and gravimetric analysis were carried out (see ‘Methods’ section). Figure
shows the Lambert-Beer behavior and the different slopes of the two suspensions, thus indicating that the two media have different dispersibility. Namely, for GPTMS and PhTES, an absorption coefficient of 2,415 and 4,710 ml·mg−1·m−1 was respectively found.
Figure 2 Absorption coefficient α determination. Optical absorbance (660 nm) as a function of graphene concentration in PhTES and GPTMS. A Lambert-Beer behavior is shown, with an absorption coefficient α of 4,710 and 2,415 ml·mg−1 (more ...)
The concentrations of graphene in GPTMS and PhTES as functions of the initial graphite concentration are reported in Figures
, respectively. As far as the GPTMS dispersion is concerned, a direct proportionality between initial graphite and graphene seems to exist up to 5 wt.% of the initial graphite (Figure
). After this value, a decrease of graphene concentration was found. This is probably due to the following observed phenomenon: When high concentrations of graphite are added to the used plastic reactor, it tends to precipitate thus making the subsequent sonication process less effective. It should be highlighted that the maximum concentration of graphene here obtained is one of the highest reported so far by any method
]. An analogous trend was observed also in the case of PhTES (Figure
), for which the maximum graphene concentration (0.66 mg/ml) was found when the initial graphite concentration was 2.5 wt.%. Besides, the maximum amount of graphene that can be dispersed in PhTES is much lower than in the case of GPTMS, 0.66 mg/ml instead of 8.0 mg/ml, thus indicating that this latter liquid medium is more effective than PhTES in the dispersing graphene.
Graphene concentration in GPTMS as a function of initial graphite concentration. Sonication time = 24 h.
Graphene concentration in PhTES as a function of initial graphite concentration. Sonication time = 24 h.
TEM was used to investigate the state of the graphene particles dispersed in organosilanes; indeed, this technique is usually employed for the investigation of graphene dispersions
]. As shown in Figure
, the images revealed a large quantity of flakes of different types. A larger proportion of flakes were few-layer graphene of various dimensions: in particular, very large flakes (lateral size approximately 1 μm) and smaller flakes with an average lateral size of 100 to 200 nm. It should be underlined that, in all cases, we did not observe graphite aggregates. Despite of what was reported in other applications, a relatively large size distribution is generally not considered a drawback in polymer nanocomposites in which the nanometric dimensions of the filler is the predominant factor influencing the properties of the resulting material
TEM micrographies. Graphene dispersed in GPTMS (a,b) and in PhTES (c,d).
A statistical analysis on TEM data was performed in order to verify the exfoliation, thus analyzing carefully the edge of the graphene flakes and measuring the number of layers presented in each flake
]. At this regards about fifty different images have been observed in order to obtain a significant number of flakes for the statistical analysis. The results are reported in Figure
Statistical analysis. Histograms showing the number of layers per flake measured for graphene obtained by exfoliation of graphite in GPTMS (top) and PhTES (bottom).
The flakes present good exfoliation degree with an average number of layers of 2.7 for GPTMS and 2.4 for PhTES; standard deviation was about 0.4 and 0.5, respectively. In both cases, only 14% of the flakes were present more than three layers. Moreover, no more than five layers have been counted in very few flakes thus indicating a narrow dispersion. These results confirm that the exfoliation process was very effective.
Raman spectroscopy is essential for the characterization of graphene. Indeed, it is considered one of the best characterization techniques for discriminating between graphite and graphene
]. As shown in Figure
, typical Raman signals of graphene recovered from the dispersions of GPTMS and PhTES are very similar, both exhibiting the characteristic graphene peaks. In particular, as far as PhTES is concerned, the G band at 1,577 cm−1
, the 2D band at 2,696 cm−1
, and the disorder-related D peak at approximately 1,346 cm−1
are evident. Similarly, graphene obtained from GPTMS shows the G band at 1,574 cm−1
, the 2D band at 2,701 cm−1
, and the disorder-related D peak at approximately 1,345 cm−1
. The shape and position of the 2D peaks (Figure
) is typical of bilayer graphene (four components with the main peak at approximately 2,701 cm−1
, as confirmed by a deconvolution process
]). As a comparison, the 2D peak of graphite consists of two components and the main peak is upshifted to 2,713 cm−1
Raman analysis. Spectra of graphene obtained by sonication in PhTES and GPTMS from 5 wt.% of the initial graphite compared with graphite (a). 2D peaks evaluation for this systems (b).
The disorder-related D peak is present also in the initial graphite powder, but its intensity is higher for graphene. This finding was already reported in which graphene was produced by sonication of graphite and can be attributed to the new edges produced during the sonication process: Ultrasonic treatment causes the decrease in size of the flakes compared to the original graphite, with a consequent increase of the total edge length
]. Comparing the intensity of the peaks and the D
ratio found in the case of graphene obtained in GPTMS and PhTES, only little differences can be found, thus indicating that the disorder induced by the exfoliation process is very similar. Namely, the D
ratio is 0.47 for graphene dispersed in GPTMS and 0.65 for that dispersed in PhTES, while the reference value for graphite powder is 0.14.
On the basis of the above concentration results, some considerations about the use of the Hildebrand solubility parameters δ
should be done. Indeed, Hernandez et al.
] stated that these parameters could be the key for envisaging the best graphene solvent media. In particular, they calculated a δ
value for graphene equal to ca
. 23 MPa1/2
, this value being the same of N
]. However, GPTMS, the best solvent medium reported here, is characterized by δ
= 14.5 MPa1/2
]. This value suggests that graphene solubility parameters should be revised and/or that they are not adequate for any reliable solubility prediction on this respect.