Covalent functionalization of
graphene significantly broadens its application potential via tuning
its electronic and surface properties.[1,2] Therefore,
a wide range of dry and wet chemistry approaches have been developed
for graphene functionalization.[3−7] Despite recent progress in this field, covalent modification of
graphene is still hampered by its low reactivity. Moreover, the reactivity
depends on the type of graphene support[8,9] and number
of graphene layers.[10−12] Consequently, there is a need for new strategies
that permit high yielding graphene functionalization under more controlled
conditions.Very recently, fluorographene (FG), a stoichiometric
(C1F1) and well-established graphene derivative,[13−18] has been shown to be susceptible to reductive defluorination[14] and nucleophilic attack.[19] These findings suggest that FG may be a useful alternative
material to graphene for the preparation of graphene derivatives.
This idea is supported by recent achievements in the field employing
nucleophilic substitution of fluorine atoms by other groups, such
as sulfhydryl,[20] amino,[21−24] alkoxy,[22] dichlorocarbene[25] and urea.[26] However, efficient reaction of FG with Grignard
reagents to allow high yield alkylation and arylation of graphene
has not yet been reported.The Grignard reaction is one of the
most well-established methodologies
for the formation of C–C bonds in organic chemistry. Grignard
reagents bear a nucleophilic carbon atom owing to its bonding to magnesium,
and the in situ formed hydrocarbon anion can attack electrophilic
carbons, such as the carbons of FG.[19,22] The Grignard
reaction has been successfully applied to fluorinated carbon nanotubes,[27] but to date, there is only one report regarding
the covalent modification of chlorinated graphene with Grignard reagents.[28] Very recently, the same group reported that
Grignard reaction on fluorinated epitaxial graphene was not feasible.[29]In the present work, we report the first
successful covalent modification
of FG based on the Grignard reaction, which yielded homogeneous and
high-density (5.5–11.2%) functionalization of the graphene
surface. Three different types of organometallic reagents, containing
alkane (pentyl), alkene (allyl), and aryl (anisolyl or p-methoxyphenyl) moieties, reacted successfully, unlike the reaction
with the ethynyl reagent. This behavior was rationalized in terms
of the nucleophilicities of the reactive centers, assessed by computational
chemistry (Figure ). The new covalently functionalized graphene derivatives were characterized
by complementary techniques: spectroscopic, thermogravimetric and
microscopic. In addition, we used density functional theory (DFT)
calculations to evaluate the thermodynamic stabilities of the chemically
modified graphenes, and delineate the influence of the different covalently
attached groups on the electronic properties of graphene.
Figure 1
Overview of
the reaction of fluorographene with Grignard reagents,
yielding covalently functionalized graphenes. The partial charges
on the nucleophilic carbons in the hydrocarbon anions are shown in
bold.
Overview of
the reaction of fluorographene with Grignard reagents,
yielding covalently functionalized graphenes. The partial charges
on the nucleophilic carbons in the hydrocarbon anions are shown in
bold.Initially, a FG suspension was
prepared by sonication of graphite
fluoride (GF) in dry tetrahydrofuran. Next, the Grignard reagent (Figure ) was added dropwise
to the suspension and the mixture was stirred under nitrogen for 5
h. Afterward, excess Grignard reagent was quenched with a saturated
aqueous solution of ammonium chloride and the material was washed
with a copious amount of water. To remove any residues of magnesium
salts, the product was resuspended in aqueous 5% HCl solution and
washed with water, ethanol and dichloromethane, consecutively.When ethynylmagnesium bromide was used as a precursor compound
for the generation of the nucleophile (ethynyl group), the reaction
with FG was unsuccessful, even after heating at 60 °C. To decipher
this result, DFT calculations were carried out to assess the nucleophilicity
of the reacting hydrocarbon anions in terms of the electrostatic potential-derived
partial charges. The calculations indicated that the electrostatic
potential over the nucleophilic ethynylcarbon was the lowest among
the considered nucleophiles (Figure ). It should be noted that our DFT calculations indicated
that ethynyl was thermodynamically stable on graphene owing to the
formation of a strong covalent bond to graphene with high binding
energy of −2.40 eV. This implies that the unsuccessful outcome
of the reaction was due to kinetic reasons.Successful covalent
functionalization of FG by Grignard reagents
was confirmed by Fourier transform infrared (FT-IR) spectroscopy,
owing to the appearance of new bands, not present in pristine GF (Supporting Information, Figure S1A). Aliphatic
C—H stretching vibrations were found at 2980–2850 cm–1 for the alkyl-, alkenyl- and aryl-functionalized
graphenes (Figure A). In particular, the spectrum of allyl graphene (AlG) displayed
four unique bands at 3074, 1633, 990 and 913 cm–1, which were ascribed to the monosubstituted alkene group. Concerning
the anisolyl graphene (AnG), bands due to the C—O bond and para disubstitution of the benzene ring were observed at
1037 and 812 cm–1, respectively. Moreover, the FT-IR
spectra showed the presence of aromatic rings due to the occurrence
of two characteristic bands at around 1590 and 1460 cm–1. The band at 1570 cm–1 was particularly evident
for the pentyl graphene (PeG), and corresponded to conjugated C=C
double bonds, which were formed owing to the reductive defluorination[14,30] of FG occurring simultaneously with the substitution.
Figure 2
FT-IR (A),
C 1s XPS (B) and Raman (C) spectra of graphene derivatives.
The inset (left panel B) shows the XPS spectrum of pristine GF.
FT-IR (A),
C 1s XPS (B) and Raman (C) spectra of graphene derivatives.
The inset (left panel B) shows the XPS spectrum of pristine GF.All graphene derivatives showed
almost quantitative elimination
of fluorine atoms according to atomic content analyses obtained from
X-ray photoelectron spectroscopy (XPS) survey spectra (Supporting Information Figure S1B, and Table
S1). High-resolution C 1s XPS spectra of the covalently functionalized
graphene derivatives (Figure B) showed that the intensities of the C–F components at binding energies of 288–293
eV were significantly reduced with respect to FG (Figure B inset, and Table S2). Moreover, the appearance of a band at 284.7 eV
for all materials, which was absent in the starting GF material, confirmed
the presence of sp2carbons formed as a result of defluorination.
Regarding AnG, the oxygen content was found to be 8.3% (Supporting Information, Table S2) owing to the
presence of oxygen atoms in the anisolyl (p-methoxyphenyl)
functionality, verifying successful attachment of this group to the
carbon lattice.Formation of conjugated double bonds in all
three graphene derivatives
was also confirmed by Raman spectra. Whereas the starting material
was Raman silent,[25] the Raman spectra of
the graphene derivatives exhibited characteristic D and G bands at
around 1330 and 1580 cm–1, respectively (Figure C). The ratio of ID and IG band intensities
(ID/IG, which
reflects the degree of functionalization) for pentyl, allyl and anisolyl
graphene was 1.05, 0.93 and 0.98, respectively. These values, along
with considerable broadening of the bands, indicate a high degree
of graphene functionalization.[31,32] Similar broadening
occurred for the 2D band (Figure S2), restricting
the evaluation of the number of the layers in the derivatives. All
these observations suggest that the treatment of FG with Grignard
reagents generated covalently and high-density functionalized graphenes,
formed via nucleophilic substitution and accompanied by reductive
defluorination. The concurrent substitution and defluorination resulted
in materials with negligible fluorine content which could be considered
as graphene alkanes and alkenes, or aryl graphenes. Therefore, the
reported strategy, i.e., functionalization of FG by nucleophiles,
represents an efficient and robust alternative approach for the direct
functionalization of graphene.To estimate the thermal stability
and functionalization degree
of the graphene derivatives, thermogravimetric analysis (TGA) was
employed under nitrogen. The starting material is known to be stable
below 400 °C and exhibits weight loss of 75% in the temperature
range 450–650 °C,[13] as also
shown in Figure S3. The TGA profiles indicated
that decomposition of PeG started at 180 °C, whereas AlG and
AnG derivatives started to decompose at higher temperatures (220 °C).
These observations suggest that the alkyl-graphene derivative was
less stable than the alkenyl and aryl derivatives. DFT calculations
corroborated the TGA results because they revealed that the binding
energy of pentyl on graphene was 0.54 eV per molecule, whereas the
binding energies of allyl and anisolyl were higher, i.e., 0.81 and
0.91 eV.According to the theoretical calculations, the grafting
of functional
groups onto FG reduced its band gap (BG) from ∼8 eV[33,34] to ∼0.15 eV (Figures A, and S4), suggesting them as
small-BG semiconductors. The BG was direct and narrowed with decreasing
coverage (Figure S4). The BG occurred due
to sp3 defects in the graphene lattice because the molecular
states lay 3 eV (AlG and AnG) and 5 eV (PeG) below the Fermi level
(Figure A). The optical
BGs were also evaluated experimentally from the Tauc plots (Figure S5). Both experimental and theoretical
BG values displayed the same trend: BGPeG > BGAlG > BGAnG, although BG depended on the functional group
rather weakly. The lengths of the C–C bonds between the sp3 lattice-carbons and the carbons of the functional groups
were 1.62 Å, 1.61 and 1.59 Å for pentyl, allyl and anisolyl,
respectively, and correlated with the binding energies. The lengths
of the new C–C bonds were somewhat longer than a typical single
C(sp3)–C(sp3) bond of 1.54 Å, indicating
that their bond orders were slightly below one, and thus they were
susceptible to detachment from the graphene surface.
Figure 3
(A) Density of states
of AnG. (B) TEM and (C) dark field image
with carbon and oxygen EDS maps of AnG flakes. (D) AFM image with
the height profile of a typical AnG flake.
(A) Density of states
of AnG. (B) TEM and (C) dark field image
with carbon and oxygen EDS maps of AnG flakes. (D) AFM image with
the height profile of a typical AnG flake.In TGA, the PeG, AlG and AnG derivatives lost 45%, 42.5%
and 44%
of their weight, respectively, between 200 and 580 °C. The weight
losses corresponded to detachment of the functional groups, as suggested
by the DFT calculations. This was also confirmed for the PeG derivative
by mass spectroscopy analysis of the evolved gases and, in particular,
by the detection of a fragment at m/z 71 corresponding to the pentyl group (Mr(C5H11) = 71.1, Figure S3). Considering that the weight losses corresponded to detachment
of the functional groups, we assessed the degree of functionalization
(DF, see S.I. for details) of the individual
graphene derivatives to be 8.5, 11.2, and 5.5%, for PeG, AlG and AnG,
respectively. The presence of impurities detected by XPS was also
taken into account when determining the DF. Such DF values are particularly
high when compared to graphene derivatives synthesized by reactions
on graphene. For example, values of 1.2 to 5% DF were obtained after
reaction with diazonium salts.[31,35,36] Importantly, the high DF of FG resulted in high dispersibility of
the adducts in organic solvents (Figure S6).Finally, we examined the morphology of the graphene derivatives
by transmission electron microscopy (TEM), atomic force microscopy
(AFM) and energy dispersive spectroscopy (EDS) elemental mapping.
All prepared materials exhibited similar morphological characteristics
(Figures and S7). A representative TEM image of AnG (Figure B) showed that the
functionalized material consisted of flakes containing very few layers.
AFM measurements indicated that the thickness of AnG was around 1.8
nm (Figure D), which
corresponded well to the thickness of double-side-functionalized graphene
by the anisol group estimated by DFT calculations (1.77 nm). AFM analysis
on AlG verified the few-layer nature of the flakes (Figure S8). EDS elemental mapping verified the presence of
oxygen on the graphene lattice with a homogeneous distribution of
oxygen over the graphene surface (Figure C).In conclusion, a straightforward,
facile and very efficient approach
for the covalent double-sided and high-degree functionalization of
graphene was developed. This was achieved by exfoliation of the commercially
available material graphite fluoride and its reaction with Grignard
reagents. In three cases, nucleophilic substitution on FG was successful
and alkylated, alkenylated and arylated graphene derivatives were
prepared under mild reaction conditions. Concurrent reductive defluorination
allowed the preparation of fluorine-free and well-defined graphene
derivatives, which were highly dispersible in organic solvents. The
reaction resulted in the homogeneous, high-degree (5.5–11.2%)
and double-sided functionalization of graphene. Moreover, we verified
through theoretical calculations that the value of nucleophilicity
governs the success of nucleophilic substitution on FG. The prepared
allyl derivative is particularly interesting because the presence
of the double bond enables further chemical reactions.[37−39] Therefore, a wide range of successive graphene derivatives could
be prepared starting from the allyl graphene.
Authors: Vasilios Georgakilas; Michal Otyepka; Athanasios B Bourlinos; Vimlesh Chandra; Namdong Kim; K Christian Kemp; Pavel Hobza; Radek Zboril; Kwang S Kim Journal: Chem Rev Date: 2012-09-25 Impact factor: 60.622
Authors: Jan M Englert; Christoph Dotzer; Guang Yang; Martin Schmid; Christian Papp; J Michael Gottfried; Hans-Peter Steinrück; Erdmann Spiecker; Frank Hauke; Andreas Hirsch Journal: Nat Chem Date: 2011-03-20 Impact factor: 24.427
Authors: Qing Hua Wang; Zhong Jin; Ki Kang Kim; Andrew J Hilmer; Geraldine L C Paulus; Chih-Jen Shih; Moon-Ho Ham; Javier D Sanchez-Yamagishi; Kenji Watanabe; Takashi Taniguchi; Jing Kong; Pablo Jarillo-Herrero; Michael S Strano Journal: Nat Chem Date: 2012-08-12 Impact factor: 24.427
Authors: Miroslav Medveď; Giorgio Zoppellaro; Juri Ugolotti; Dagmar Matochová; Petr Lazar; Tomáš Pospíšil; Aristides Bakandritsos; Jiří Tuček; Radek Zbořil; Michal Otyepka Journal: Nanoscale Date: 2018-03-08 Impact factor: 7.790