Project funded by the European Commission under Grant Agreement n°696656 Graphene Core 1 Graphene-Based Disruptive Technologies Horizon 2020 RIA WP13 Functional Foams and Coatings Deliverable 13.3 “Understanding the exfoliation processes of GRM” Main Author(s): Andrea Liscio, CNR Martin Lohe, TUD Artur Ciesielski, UdS Andrea Carlo Ferrari, UCAM Heiner Friedrich, TUD Boaz Pokroy, TECHNION Due date of deliverable: M12 Actual submission date: M12 Dissemination level: Public
Transcript
Project funded by the European Commission under Grant Agreement n°696656
1. Electrochemical exfoliation of GRM ............................................................... 51.1. In-situ monitoring ...................................................................................................... 51.2. Exfoliation process under anodic conditions ........................................................ 6
2. Mechanical exfoliation ..................................................................................... 91.3. Microfluidization processes ..................................................................................... 91.4. GRMs and other 2D related materials (Ball milling exfoliation) ............................ 9
3. Beyond solution monitoring .......................................................................... 101.5. Molecule-assisted exfoliation ................................................................................ 101.6. Geometrical and statistical analysis of GRMs: a dedicated metrology ............. 10
Any other sections .................................................................................................. 121.7. Graphene-based membrane characterization ...................................................... 12
Summary The understanding the exfoliation processes of graphene related materials (GRMs) is a key-
point for the upscaling production of such materials. Focussing on Liquid Phase Exfoliation
approaches, in this report we visualised and characterised the different processes known so
far for exfoliation and intercalation of 2D-layered materials. Aiming at the fundamental
understanding of these processes and possibilities for optimisations in terms of composite
production (such as GRM-host interactions).
We report the advances in electrochemical exfoliation performed by the partners involved in
the WP13.3 activities within the sub-Task 13.1.3 “Electrochemical processing of GRM”, 13.4.1
“Characterization of intercalation and exfoliation process” and 13.4.2 “Characterization and
visualization of electrochemical exfoliation mechanisms”.
Main results:
• New scalable protocol of electrochemical exfoliation: new routes to produce GRMs
compatible with large-scale industrial application. Tuneable properties of the produced
materials. Step-by-step in-situ and ex-situ monitoring to minimise the defects without
significant loss of processability or properties.
• Beyond solution monitoring. Small molecules as surfactants to stabilise GRMs in low-
boiling point solvents. Direct understanding of solution properties. Development of
experimental protocols and mathematical/statistical tool dedicated to the quantitative
analysis of large scale GRMs.
A comprehensive understanding of such issues provided to the production of GRMs suitable
for different applications, as required by other partners and within different WPs, such as
dedicated GRMs for graphene-based inks (WP9 “Flexible electronics”) and processeable
GRMs with good affinity with polymers (WP14 “Composites” and WP15 “Production”).
Introduction Liquid phase exfoliation (LPE) of graphite is one of the most promising and potentially scalable
routes to prepare stable suspension of 2D nanosheets with tuneable properties (electronic,
electrical, mechanical, etc.) and their processing into useful materials. One of the main
advantage for applications of GRMs is that they can be produced in large scales by liquid
phase exfoliation. The exfoliation process shall be considered as a particular fragmentation
process, where the 2D character of the exfoliated objects will influence significantly
fragmentation dynamics as compared to standard materials. As a scaleable method succeeds
and it is indeed upscaled for industrial production, new characterisation protocols and metrics
have to be devised to enable efficient on line quality control of the produced materials.
This deliverable is dedicated to develop new strategies to overcome two main challenges:
Graphene Core 1 D13.3 31 March 2017 5/ 14
• the development of procedures to process graphene in a way that is compatible with
actual production standards (compounding with polymers, coating on metals and
surfaces, integration with silicon-based electronics, etc.);
• the improvement of quality and reliability of the produced materials (larger sheet size,
lower average thickness, lower defect density, etc.).
1. Electrochemical exfoliation of GRM
1.1. In-situ monitoring
In cooperation with Leibniz-Institute for Solid State and Materials Research, TUD employed in-
situ Raman spectroscopy to understand to mechanism of graphite exfoliation. In general, the
intercalation of guest ions started from the boundary to the central part of graphite foils. The
exfoliation was very fast, completed within 1 min (Fig. 1a).
Figure 1. (a) morphological changes on graphite foil during exfoliation; (b) in-situ Raman spectra as function of working bias on graphite electrode; (c) scheme of graphite exfoliation with external bubbling; (d) Raman spectra of exfoliated graphene produced with (without) bubbling process.
The measurements of in-situ Raman spectroscopy were focussed on a fixed position of
graphite. With increasing working bias (from 0 to 4 V), the intensity of D band was notably
enhanced (Fig. 1b), which resulted from countless oxygen radicals by water splitting. In a real
experiment, a high voltage of 10 V was used. Exfoliation could be easily achieved when electric
repulsion (F1) is bigger than van der Waals force (F2). Since massive bubbles were generated
in the exfoliation process, the attached bubbles isolate graphite foil from electrolyte, protecting
graphene flakes from excess oxidation. To verify this hypothesis, external bubbling was
Graphene Core 1 D13.3 31 March 2017 6/ 14
applied in a typical exfoliation procedure in solution of sulphate salt. However, the intensity
ratio of D band to G band (Id/Ig=0.4) is the same as that without bubbling. The actual
mechanism behind electrochemical exfoliation is still not clear.
Figure 2. (a) scheme of graphite exfoliation with alternating current; (b) the curve of working bias on anode; (c) proposed mechanism for exfoliation; (d) possible chemical reactions at graphite anode and cathode.
In contrast to conventional graphite exfoliation by direct current, a novel scalable protocol
based on alternating current was developed. An aqueous solution containing organic sulphate
was used as electrolyte and an alternating current with tuneable frequency (e.g. 0.1 Hz) was
applied as driving force for the intercalation of guest ions. For the first time, graphite exfoliation
has been realised at anode and cathode, affording ultrahigh production rate exceeding
20 gram per hour in laboratory trials, which is appealing for upscaling production. Thanks to
fast and efficient exfoliation as well as in-situ electrochemical reduction, exfoliated graphene
sheets exhibit outstanding quality, with respect to low defect density (Id/Ig<0.2 in Raman
spectra, a C/O=21.2). The exfoliated graphene also holds superior processability in various
solvents, which make it ideal for a wide range of graphene-related applications.
1.2. Exfoliation process under anodic conditions
The properties of GRMs strictly depend on their processing and production methods. Though
monocrystalline sheets of GRMs can be easily prepared via mechanical exfoliation (i.e. scotch-
Graphene Core 1 D13.3 31 March 2017 7/ 14
tape method), such an approach is not suitable for mass production, which is a key requirement
towards practical GRM-based applications. In this context, UdS investigated the use of
electrochemical exfoliation (EE) to produce large quantities of graphene flakes with low degree
of oxidation. These could be conveniently employed for the preparation of large-area graphene
films on arbitrary substrates by means of low-cost solution processing methods. More
specifically, they explored the influence of the EE process under anodic conditions on the
electrical properties of graphene both at the single-flake level and in thin films. By making use
of a combination of experimental techniques ─ including X-ray photoelectron spectroscopy
(XPS), atomic force microscopy (AFM), transmission electron microscopy (TEM), as well as
electrical measurements in field-effect transistors (FETs) ─ the UdS team carried out a
thorough characterisation of electrochemically-exfoliated graphene (EEG) sheets and
unravelled their structure-property relationship. he electrical properties of EEG ─ prepared
using common processing conditions (e.g. ref.[1]) ─ are mostly limited by mechanical defects,
such as cracks, holes and tears within individual sheets, rather than by oxygen functionalities,
which could be significantly reduced by making use of microwave irradiation, as recently
reported.[2] Furthermore, the performance of FETs based on EEG thin films was found to be
comparable to that of devices based on individual flakes, revealing that the film-assembly
process itself was not responsible for degrading the charge-transport properties.
Figure 3 illustrates the methodology used for the electrical characterization of single flakes of
EEG, which were deposited on oxidized silicon substrates (ρSi≈0.001 Ω·cm, tox=290 nm) by
spin-coating from dimethylformamide (DMF) solutions. Multiterminal back-gated FETs were
fabricated using standard e-beam lithography, metal deposition (3/40 nm of Ti/Au) and lift off.
The four-probe measurement configuration was employed to remove the contribution of the
contact resistance and access the intrinsic sheet resistivity, which was found to span within
the range 15-30 kΩ/sq. To minimize the influence of environmental adsorbates, such as O2
and H2O, all the measurements were carried under inert atmosphere (N2-filled glovebox).
Moreover, a vacuum-annealing step (p~5×10-8 mbar, T~60 °C) was performed to desorb
solvent traces, as well as O2 and H2O ─ which are known to be detrimental electron-acceptor
traps ─ and thus reduce the level of hole doping within the material. Upon annealing the
behaviour of the EEG FETs changes from unipolar (p-type) to ambipolar, as in the case of
mechanically-exfoliated or CVD-grown graphene devices. A well-defined charge-neutrality
point could be identified in the transfer curves acquired after annealing (Fig. 3c, inset) at Vg
values lower than +5 V. To the best of our knowledge, this is the first observation of ambipolar
transport in EEG, revealing the presence of a significantly lower level of oxidation as compared
to graphene oxide (GO). At this stage, the charge-carrier mobility (Fig. 3d) is still remarkably
lower than in the case of pristine graphene sheets, due to the presence of mechanical defects
Graphene Core 1 D13.3 31 March 2017 8/ 14
generated during the various steps of the electrolysis process (e.g. strong interlayer repulsion,
ion intercalation and evolution of gas bubbles among graphitic layers).
Figure 3. Characterisation of EEG single flakes. Electrical transport: schematics (a) and optical micrograph (b) of the multi-terminal back-gated FETs used for electrical characterization. (c) Ids- Vg transfer characteristics of an EEG FET acquired before (blue) and after (red) high-vacuum annealing of ~60 °C. The plot inset shows the occurrence of the charge neutrality point VCNP at ~4V. (d) µFE values for holes (h+) and electrons (e-), as measured in the two- and four-terminal measurement configuration. Morphological characterisation: (e) AFM image of EEG deposited on SiO2/Si substrates by spin-coating from DMF. (g) Statistical distribution of the number of layers per sheet determined by AFM (blue) and TEM (red). (f) AFM topographic (top) and phase-contrast (bottom) images show mechanical defects; (h) Surface density of mechanical defects as a function of the number of EEG flakes. Chemical analysis: XPS spectra of EEG (i) and GO (j) in the C1s survey.
A morphological characterisation of the EEG sheets (Fig. 3e-h) revealed the presence of
nanoscale holes, cracks and tears within the nanosheets (mostly 1-3 layers thick). A statistical
study based on high-magnification AFM images showed that ~30% of EEG flakes was studded
with mechanical defects, whose average density was found to be approximately 20-35% of the
flake area. Moreover, XPS measurements proved that the oxidation degree of EEG sheets
obtained after 1’ of electrolysis (Fig. 3i) was significantly lower than in GO powder (Fig. 3j).
Finally, large-area films of EEG were produced and investigating their electrical properties
through the fabrication and characterization of back-gated top-contact FETs (Fig. 4a). Large
area EEG films were produced by self-assembly of graphene sheets dispersed in DMF at the
interface with water, leading to thin graphene films with thickness of ca 3 nm and surface
coverage over 85%, as deduced from a set of AFM images as the one reported in Fig. 4b.
After evaporation of Au contacts (40 nm thick, interdigitated geometry) through a shadow
mask. The latter were found to display a p-type field-effect mobility of ~3-4 cm2V-1s-1, similar to
that of FETs based on individual EEG flakes, proving the achievement of uniform coverage
and high-quality overlap among the flakes.
Graphene Core 1 D13.3 31 March 2017 9/ 14
Figure 4. Optical (a) and AFM (b) images of EEG films deposited on SiO2/Si substrates. Transfer curve of an FET based on EEG films.
2. Mechanical exfoliation
1.3. Microfluidisation processes
In order to understand the mechanisms of exfoliation during the microfluidisation process,
UCAM studied its fluid dynamics. During microfluidisation a mixture of graphite particles and
sodium deoxycholate surfactant in water pass through a microchannel under high pressure
(207 MPa). The fluid flow in the microchannel is 300 m/s2 and the Reynolds number 2.7×104,
indicating a fully developed turbulent flow inside the microchannel. We then calculated the
energy dissipation rate (~8.5×109 m2/s3) and the turbulent shear rate ~108 s-1. This is four
orders of magnitude higher than the one required to initiate graphite exfoliation. Thus, the
exfoliation in the microfluidiser is primarily due to shear stress generated by the turbulent flow.
Key advantage over other methods such as sonication and shear-mixing is that the high shear
rate is applied to the whole fluid volume and not just locally. The Kolmogorov length (Kl) is
~100nm. This is the length-scale above which the system is in the inertial subrange (IS) and
below which it is in the viscous subrange (VS), where turbulence energy is dissipated by heat.
The stresses applied on the flakes in the IS are much higher than in VS. Thus, a lower Kl
makes the process more energy efficient. Kl can be tuned, by either increasing the kinematic
viscosity of the dispersion or decreasing the energy dissipation rate, thus extending the viscous
subrange of turbulence realising a milder exfoliation.
1.4. GRMs and other 2D related materials (Ball milling exfoliation)
Ball milling approach previously used extensively to mill expanded graphite. In particular,
TECHNION focussed the activities to study the exfoliation processes of 2D materials mixed
with TiO2 for photocatalytic coatings provided by Italcementi and UNIBO partners. The focus
was to explore the effects of the experimental parameters to develop a standard procedure to
prepare the GRM/TiO2 composites. Exfoliation processes were performed by milling and the
obtained mixed materials were characterised by Hi-resolution X-Ray Diffraction measurements
Graphene Core 1 D13.3 31 March 2017 10/ 14
performed using synchrotron radiation at European Facility ERSF. Others 2D materials such
as WS2 and MoS2 were mixed with melamine to improve the exfoliation and the affinity with
TiO2. TECHNION studied quantitatively the exfoliation degree taking by varying three
parameters the milling parameters. As expected, exfoliation degree increases with time and
rotation speed while the number of milling balls did not affect the final 2D materials. The
developed procedure allowed to optimise the milling parameters such as: 90-120 min at 250
rpm using standard steel balls.
3. Beyond solution monitoring
1.5. Molecule-assisted exfoliation
CNR partner demonstrated that it is possible to combine the use of organic solvents and
organic surfactants such as perylene diimide (PDI) to obtain dispersions of GRM in solvents in
which graphene would not be stable on its own such as IPA, THF and CHCl3. In particular,
CNR showed that few-layer graphene flakes can be stabilised in organic solvents by adding
small amounts of suitable molecules, in a similar but different way to the typical action of soaps
in water.[3] The composite materials thus obtained can be included in a straightforward way in
elastomer or thermoplastic materials, rendering them conductive. It is noteworthy to underline
that the process could be industrially relevant because the employed surfactants are
commercial PDI molecules that are already used as polymer additives. This approach could
be compatible with large-scale industrial application in the polymer industry.
1.6. Geometrical and statistical analysis of GRMs: a dedicated metrology
The production of large quantities of 2D materials in solution with well-controlled morphological
properties of the nanosheets (i.e. area, lateral size and shape) is not only a technological
challenge but also a fundamental one, because several scientific aspects still need to be
clarified for a detailed understanding of the fragmentation mechanisms and the control of the
final products. Furthermore, it is necessary to develop fast and reliable protocols to measure
and analyse a large number of 2D objects and well as to find a set of “robust” parameters to
achieve an accurate multi-scale description of the system. CNR developed a new, fast and
accurate approach[4] positively tested using a standard 2D material as the target system, i.e.
GO completely exfoliated without any aggregation, featuring more than 99% of monoatomic
nanosheets in water.[5] GO in water underwent a standard ultrasonication treatment from 0 to
100 h. GO in water underwent a standard ultrasonication treatment from 0 to 100 h. Combining
the analysis of images acquired with different techniques such as Fluorescence (FM),
Scanning Electron (SEM) and AFM microscopies, the distribution of the 2D fragments in the
suspension was monitored studying the morphological parameters from the millimetre to the
Graphene Core 1 D13.3 31 March 2017 11/ 14
nanometre scale. The quantitative analysis of the fragments distributions on multi-scale
provides to measure the mechanical properties of the GO single sheet such as the Young’s
modulus and the fracture strength.
Figure 5. Evolution of size distribution of the 2D nanosheets for different sonication times. The measured distributions are fitted by: (blue) power law, (red) Gamma and (green) exponential functions. Light green, light yellow and light blue backgrounds indicate different active regimes: pristine fragments, bulk fragmentation and edge fragmentation.
Moreover, the analysis of the fragment distributions gives detailed information on the dynamics
of the 2D fragmentation providing direct evidence of different regimes given by the interplay of
two breaking mechanisms, such as core fragmentation and peripheral erosion. Showing
unambiguously the coexistence of different GO sheets, it was proved that the GO suspensions
is described as a blend of large sheets together with small debris fragments (similar to fulvic
oxides), casting a new light on the results reported in the last years on this topic. The use of
area distributions allowed a new quantitative approach to measure the heterogeneity of 2D
materials in solution. CNR proposed to extend to 2D objects concepts already developed for
1D polymers. In particular, the mass-molar dispersity (ÐM) (a.k.a. polydispersity) has been
generalised, by defining the area dispersity of 2D materials as: Ð"# = 𝐴" 𝐴 ", where A is
the sheet area. Figure 6a plots the time evolution of Ð2D of GO sheets during sonication. In the
first half hour the dispersity increased indicating a higher heterogeneity of the solution due to
the persistence of unbroken pristine material, as discussed before. Afterwards, with continuing
sonication Ð2D reduced reaching a value close to 2. Finally, the validity of this parameter is
Graphene Core 1 D13.3 31 March 2017 12/ 14
tested demonstrating how it can explain some macroscopic property of the material studied.
In analogy with 1D polymers, dispersity variation influences the way that light is scattered by
2D objects in solvent. Static light scattering (SLS) measurements performed on GO
suspension showed that the measured signal is proportional to Ð2D (fig. 6b) suggesting that
the smallest GO sheets produced by long sonication times had no relevant folding in water
suspensions, behaving as quasi-planar objects, in agreement with further Dynamic Light
Scattering DLS measurements.
Figure 6. Experimental measurements of 2D areal dispersity index. (a) Time-dependence of areal dispersity Ð2D measured with different techniques. (b) Experimental evidence of the correlation between SLS signal (black squares) and Ð2D for t > 20 h. The red line is a linear fit of the experimental data.
Any other sections
1.7. Graphene-based membrane characterisation
TU/e partner focussed on the 2D/3D structural characterisation of GOx membranes by
(cryo)TEM and electron tomography in collaboration with CNR within the Subtask 13.4.4. For
multiscale analysis by TEM, membranes need to be either prepared directly on the TEM
support (only successful for membranes >500 nm thick) or transferred to the TEM support
(successful for >100 nm thick membranes). Large area imaging was carried out, showing
significant variations of thickness (±50% of nominal thickness) for directly deposited
membranes (Fig. 1a), while for transferred membranes only small thickness variations (±10%
of nominal thickness) are found (Fig. 7b). Thickness measurements of thin homogeneous
membranes by profilometry (190 nm) and TEM (100 nm) indicate the presence of pores
throughout the structure. With a suitable membrane TEM model system at hand, the next steps
in the task will address 3D characterization by electron tomography (to resolve the pores) and
cryo-tomography to visualise differences between the dry and wet state and including water
soluble ions such as magnesium and chlorine.
Graphene Core 1 D13.3 31 March 2017 13/ 14
Figure 7. GOx membranes: a) prepared by direct filtration on a 700 mesh hexagonal grid (ISOF-CNR) and b) transferred to 100x400 mesh slotted grid (TU/e).
Conclusions We visualised and characterised the processes involved in the electrochemical, mechanical,
and molecule-assisted exfoliation in liquid-phase (i.e. fragmentation and intercalation of 2D
layered materials). The quantitative monitoring performed by combining microscopic,
spectroscopic techniques, and X-ray diffraction allowed us to optimise processes, to propose
new protocols to produce high quality of GRMs by LPE (i.e. stable suspensions in water) and
by ball milling (GRM composites with TiO2). Moreover, a quantitative study of the geometrical
properties of the fragments opened the way to define a proper mathematical/statistical
approach to study the metrology of 2D materials.
Graphene Core 1 D13.3 31 March 2017 14/ 14
References 1 Parvez, K. et al. Exfoliation of Graphite into Graphene in Aqueous Solutions of
Inorganic Salts. J Am Chem Soc 136, 6083-6091, doi:10.1021/ja5017156 (2014). 2 Voiry, D. et al. High-quality graphene via microwave reduction of solution-exfoliated
graphene oxide. Science 353, 1413-1416, doi:10.1126/science.aah3398 (2016). 3 Liscio, A. et al. Exfoliation of Few-Layer Graphene in Volatile Solvents Using
4 Andrea, L. et al. Evolution of the size and shape of 2D nanosheets during ultrasonic fragmentation. 2D Materials 4, 025017 (2017).
5 Treossi, E. et al. High-Contrast Visualization of Graphene Oxide on Dye-Sensitized Glass, Quartz, and Silicon by Fluorescence Quenching. J Am Chem Soc 131, 15576, doi:Doi 10.1021/Ja9055382 (2009).
