Unexpected reactivity of graphene oxide with DBU and DMF

Article abstract of DOI:10.1039/c8ta03529f

An unusual reaction between GO, DBU and DMF, that typically uses base and solvent, has been thoroughly analyzed providing valuable basic knowledge about the reactivity of GO, which is essential to control functionalization and therefore the properties of

Full text of DOI:10.1039/c8ta03529f

Journal of
Materials Chemistry A
View Article Online
View Journal
PAPER
Unexpected reactivity of graphene oxide with DBU
and DMF†
Cite this: DOI: 10.1039/c8ta03529f
a
a
a
b
Rafael Ramırez-Jimenez, Mario Franco, Eduardo Rodrigo, Raquel Sainz,
´
´
b
cd
e
Rafael Ferritto, Al Mokhtar Lamsabhi,
Jose Luis Acena
*
´
˜
ad
and M. Belen Cid
*
´
An unusual reaction between GO, DBU and DMF, that typically uses base and solvent, has been thoroughly
analyzed providing valuable basic knowledge about the reactivity of GO, which is essential to control
functionalization and therefore the properties of graphene derivatives. A combination of characterization
techniques, experiments and theoretical calculations allowed us to propose a structure for the materials
and a plausible mechanism. The catalytic activity of the new basic materials was also analyzed.
Received 17th April 2018
Accepted 31st May 2018
DOI: 10.1039/c8ta03529f
rsc.li/materials-a
As a continuation of our research, and in order to avoid the
presence of a basic nitrogen that could potentially interfere in
Introduction
Graphene and its derivatives are currently promising materials the catalytic process, we decided to study the anchoring of
for the development of heterogeneous catalytic systems, both different catalysts through the opening of epoxides in GO by
metallic and organic, as they display a large surface area and reaction with suitably functionalized alcohols. To our knowl-
1–6
unique physicochemical characteristics. Most research in this edge, this transformation has not been described thus far. As
eld involves functionalization of the surface or edges of gra- a consequence, we carried out a control experiment consisting
phene oxide (GO) through its multiple oxygen-based groups of the reaction of GO with 2-(methylthio)ethanol in the presence

7–9
(
alcohol, epoxide, ketone and carboxylic acid). Aer func- of 1,8-diazabicycloundec-7-ene (DBU) and using N,N-dime-
tionalization, the reduction of GO to reduced graphene oxide thylformamide (DMF) as solvent (Scheme 2). These conditions
(
rGO) partially restores the electronic properties of were chosen as they are typically used for epoxide-ring-opening
10,11
graphene.
and also because DMF features excellent dispersion proper-
13
We recently described an rGO material decorated with ties.‡ Nevertheless, the elemental analysis of the resulting
piperazine rings (rGO-NH, A), which was assembled through material revealed an unexpected increase of the nitrogen
nitrogen epoxide-opening in GO by treatment with piperazine content (8.5%) instead of the incorporation of the sulfur-
12
followed by reduction with hydrazine (Scheme 1). Material A containing moiety.
presented excellent catalytic activity as a bifunctional hetero-
Although far from our initial objective, but conscious of the
geneous organocatalyst in several processes through iminium importance of understanding the chemistry of graphene deriv-
and basic activation. The Lewis basic role of the electron-rich atives for future applications, we decided to perform a deep
graphene layer served to stabilize the iminium and ammo- study of this unexpected high ratio of nitrogen obtained, which
nium intermediate species, triggering a higher performance of must be attributed to DBU and/or DMF. Therefore, to fully
this amino- and carbocatalyst compared to piperazine alone.
explore the nature of the aforementioned transformation, we
carried out a systematic research by preparing different mate-
rials arising from the reaction of GO with DMF (material B),
DBU (material C), and DBU/DMF (material D). These materials
have been characterized by different techniques, and a reason-
a
Department of Organic Chemistry, Universidad Aut ´o noma de Madrid, Cantoblanco,
2
8049 Madrid, Spain. E-mail: belen.cid@uam.es
b
NanoInnova Technologies SL., Parque Cient ´ı co de Madrid, C/Faraday, 7, 28049 able mechanistic explanation for their formation was estab-
Madrid, Spain
lished from experimental results as well as a detailed DFT
calculation study. The catalytic activity of materials B and D and
a greener method to produce rGO-NH (A) avoiding the unde-
sirable use of hydrazine are also presented herein.
c
Department of Chemistry, Universidad Aut ´o noma de Madrid, Cantoblanco, 28049
Madrid, Spain
d
Institute for Advanced Research in Chemical Sciences (IAdChem), Universidad
Aut ´o noma de Madrid, 28049 Madrid, Spain
e
Departament of Organic and Inorganic Chemistry, Chemical Research Institute
“
Andr ´e s M. del R ´ı o” (IQAR), Universidad de Alcal a´ , Alcal a´ de Henares, 28871
Madrid, Spain. E-mail: jose.acena@uah.es
†
Electronic supplementary information (ESI) available. See DOI:
0.1039/c8ta03529f
‡ It is known that dispersions of GO in DMF and THF last for at least 3 weeks aer
1
sonication for 1 h (see ref. 13).
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
material B showed a signicant increase in the carbon content
(77.4%) compared to the starting GO (49.8%), indicating a high
degree of deoxygenation of GO. On the other hand, the amount
of nitrogen was 2.8%. This result agrees with the literature as
the use of DMF for the solvothermal reduction of GO and
14–16
concomitant introduction of nitrogen has been described,
due to the generation of dimethylamine by cleavage of DMF
17,18
under the reaction conditions.
Nevertheless, some discrep-
ancies on the structure of the resulting materials as well as the
possible mechanism will be revisited below.
To study the reaction of GO with DBU in the absence of DMF,
we performed the reaction using THF as solvent in a sealed
tube. Aer the same time (4 days) material C, which contained
around 4% of nitrogen, was obtained (entry 2). As the most
reasonable hypothesis could be an acid–base equilibrium
between DBU and the acids of GO, material C was treated with
Scheme 1 Preparation of rGO-NH (A) and types of catalytic activation.
0
basic media and the new material C subjected again to
elemental analysis (entry 3). It showed a decrease in the amount
+
of C, H and N which agrees with an exchange of DBUH by the
+
not detectable Na and a low incorporation of N on the surface
(
N/C ratio ¼ 0.012) (see ESI† for more details). Conversely, the
simultaneous use of DBU and DMF incorporated more
nitrogen-containing moieties in the resulting material D (N/C ¼
Scheme 2 Unexpected result with introduction of nitrogen in the
material.
0
1
.117, entry 4) than the sum of both reagents separately (entries
–3), thus indicating a probable synergistic effect between DBU
and DMF.
It is well-known that DBU could undergo hydrolysis of the
imidamide moiety in reuxing water to provide amido amine
Results and discussion
19
1. Moreover, the reaction of GO with DBU in water as solvent
Table 1 shows the elemental analysis of the obtained materials
aer submitting GO to different conditions. Material B was
prepared by heating a suspension of GO in DMF at 120 C for 4
days in a sealed tube, followed by ltration and successive
was previously described, and the incorporation of amine 1 was
20
proposed. Although we did not run the reaction in water, we
ꢀ
reasoned that DBU could be partially hydrolyzed under the
reaction conditions and the resulting amino group of 1 would
attack the epoxides of GO. To demonstrate this hypothesis, we
washings with water and acetone (entry 1). Elemental analysis of
Table 1 Reaction of GO with different reagents and evaluation of the nitrogen content
Elem. anal. (%wt)
Ratio
H/C
a
Entry
Conditions
Mat.
C
H
N
N/C
ꢀ
1
2
3
4
DMF, 120 C, 4 days
B
C
C
77.4
60.2
40.0
68.1
1.5
4.5
2.2
3.9
2.8
4.5
0.5
8.0
0.019
0.074
0.055
0.057
0.036
0.075
0.012
0.117
ꢀ
DBU, THF, 80 C, 4 days
b
0
ꢀ
c
DBU, DMF, 120 C, 4 days
D
ꢀ
5
DMF, 120 C, 1 day
E
72.9
3.9
7.2
0.053
0.116
ꢀ
6
EtOH, H
2
O 80 C, 1 day
F
65.0
2.1
5.1
0.032
0.078
a
b
c
ꢀ
Reactions performed in sealed tubes. Material C was washed with a 2 M NaOH solution. When the reaction was performed at 80 C (see ESI),
a lower content of nitrogen was observed (N, 6.3%; N/C, 0.09).
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
reacted 1 with GO during 1 day, which is considered as enough
21
time for the amines to attack to the epoxides. Both DMF and
EtOH were used as solvents as the comparison of the subse-
quent materials E and F (entries 5 and 6, respectively) would
give information about the possible introduction of nitrogen
from DMF. It was found that material E presented slightly lower
percentage of nitrogen than material D (compare entries 4 and
5
6
) but material F that was formed in the absence of DMF (entry
) had a signicantly lower nitrogen content than E and D.
The evolution of the reactions and the structural differences
among these materials were inferred from their characteriza-
tion data employing several analytical techniques. Fig. 1 gathers
TGA plots for materials B, D, E and F. We could observe that the
mass loss in sample B, presumably due to NMe
constant. According to the mass loss in D, E and F (200–550 C)
2
groups is quite
ꢀ
13
Fig. 2 Solid C NMR of (a) GO; (b) material B (GO + DMF); (c) material
their degree of functionalization should be quite similar. The D (GO + DBU + DMF); (d) material E (GO + 1 in DMF); (e) material F (GO
+
2
1 in EtOH/H O).
maximum of the derivative displaced towards higher tempera-
tures in the order F, E and D indicates that the chains anchored
to the graphene are more thermally stable in E and D than in F.
This could t with the assumption that materials D and E are corresponding to C]O and C]C stretches shied to lower
ꢁ
1
functionalized with both chain 1 and NMe
2
1
groups.
wavenumbers, 1713 and 1563 cm , respectively (Fig. 3).
3
ꢁ1
Fig. 2 shows a comparison between the C NMR spectra in Conversely, intense peaks around 1085–1250 cm that could
2
solid state of GO and of materials B, D, E, and F, which turned correspond to C–N stretching from the NMe groups were
out very diagnostic and certainly illustrates the reduction of the observed in materials B, D and E. These data are reinforced by
graphitic structure. The disappearance of broad peaks at 72 and DFT calculations (see ESI†). These peaks are wider for materials
6
0 ppm was a clear indication of the removal of oxygen func- B and D but narrower for material E, probably due to the smaller
tionalities (hydroxyls and epoxides, respectively), whereas the
2
peak at 132 ppm corresponding to sp carbons was displaced
towards higher eld due to the increased level of conjugation.
In material B, the presence of a new peak at 32 ppm could be
attributed to the introduction of Me N groups. The presence of
2
the hydroxyl groups in material E and not in material D could be
a consequence of the longer reaction times in the second case
that would favor water elimination. Despite this difference, the
similar features displayed by D and E led us to suspect the
introduction of compound 1 in material D.
2
The IR analysis also conrmed a restoration of the sp
carbon network (or removal of oxidative debris) when
22
comparing with GO. For example, in material B the peaks
Fig. 1 Thermograms of material B (GO + DMF); material D (GO + DBU
+
H
DMF); material E (GO + 1 in DMF) and material F (GO + 1 in EtOH/
O).
2
Fig. 3 FTIR spectra of GO and materials B–F.
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
amount of NMe in material E. For E and F the peak is narrow
Paper
2
This hypothesis is reinforced by the fact that when the non-
and shied to higher wavenumbers, this could correspond to volatile 4-phenylbutylamine was treated with DMF under the
0
the C–N stretching from compound 1. Spectra of samples C , D, conditions of entry 1 (Table 3) about only 67% of the mass was
E and F showed a reduction in intensity for the C]O stretching recovered but the formylated amine 3 was isolated in 64% yield
ꢁ
1
and a new peak around 1640 cm that could correspond to the from a very clean mixture. In the absence of DBU the yield was
formation of the carboxylate salt. Assignment of this peak as somewhat lower (entry 2), suggesting a role of DBU in the
possible C]O stretching of amides was excluded since this reaction mechanism. Nevertheless, the latter is an important
ꢁ
1
would appear as a more intense peak around 1634 cm (amide point to consider as DMF could interfere when used as solvent
ꢁ1
23
I) together with a second peak around 1585 cm (amide II).
in the amine functionalization of GO via ring opening of the
Finally, the spectrum of material C displayed an intense peak epoxides by lowering the yield and incorporating NMe groups
2
ꢁ
1
around 1645 cm corresponding to C]N stretch of DBU, but to GO. This transamidation would also explain that the reaction
aer washing with a NaOH solution (material C ) the peaks of 1 with DMF aer only 1 day also incorporated NMe2 to
0
corresponding to DBU disappeared, in agreement with an provide material E (Table 1, entry 5). The low yield obtained in
+
+
exchange of DBUH by Na as previously observed by elemental the experiment shown in entry 3, that uses K
analysis. DBU, put forward that DBU is not acting as a simple base.
Next, we performed additional experiments in order to
2 3
CO instead of
ꢀ
When material F was treated with DMF at 120 C for 72 h, the
understand the process by which materials D and E incorporate ratio of nitrogen did not increase (C, 71.5; H, 3.1; N, 4.8; S, 0.03;
NMe2 groups. First, we analyzed the nature of the resulting N/C, 0.067). This experiment supports the idea that a previous
supernatant solution when GO was treated with DMF or N,N- reaction between DMF and DBU enables the introduction of
dimethylacetamide (DMA) (Table 2). We could isolate some different nitrogen-containing groups on GO. Moreover, when
ꢀ
unaltered DBU along with compounds 2. When using DMF, the GO was treated with DBU and DMF at 120 C for 4 days
a mixture 2a/DBU in 62 : 38 ratio was obtained (entry 1). Due to (conditions of entry 4, Table 1) in the presence of BHT as
the anchoring process, only around 65% of the mass of the a radical inhibitor, the material also contained a high ratio of
initial DBU was recovered. This result is consistent with the nitrogen (C, 73.7; H, 3.7; N, 6.0; S, 0.06; N/C, 0.081), thus sup-
breakdown of the amide, the anchoring of dimethylamine porting a non-radical mechanism.
(blue) to the GO surface and a concurrent transamidation of the
At this point, we evaluated the basicity of the differently
acyl group (red) to the amine residue. The formation of functionalized materials (Fig. 4). The similarity in pH values of
compounds 2 also conrms that DBU was partially hydrolysed. materials A and B in aqueous dispersions suggests that most of
We reasoned that the low conversion and slow evolution the amino groups must be located on the surface of GO, instead
towards 2a (aer two days the ratio 2a/DBU was 44 : 56, entry 2) of forming amides by reaction with the acid groups sited at the
15
could be a consequence of the relatively low amount of water in edges as it has been proposed. IR analyses (see above) and
24
the reaction media, presumably produced from the elimination other reports also support the absence of amides. Moreover,
of the hydroxyl groups of GO. Indeed, we observed a faster
reaction using one equivalent of water, and aer 2 days only 2a
was detected (entry 3). When using DMA as solvent, acetamide
Table 3 Formylations of 4-phenylbutylamine
2b was isolated in 36% yield (entry 4).
Table 2 Reaction of DBU with DMF and DMA in the presence of GO
Entry
Reagents/conditions
Yield (%)
ꢀ
1
2
3
GO, DBU, 120 C, 4 days
64
54
<20
ꢀ
GO, 120 C, 4 days
ꢀ
GO, K CO
2 3
, 120 C, 4 days
t
a
Entry
Solvent
(days)
2/DBU ratio
Product
Yield (%)
b
1
2
3
DMF
4
2
2
4
62 : 38
44 : 56
>95 : 5
2a
2a
2a
2b
38
DMF
DMF
DMA
ND
ND
36
c
d
4
ND
a
1
b
ꢀ
Determined by H NMR. When the reaction was performed at 80 C 2
c
d
was not detected. 1 equiv. of H
2
O was added. Ratio not determined
due to overlapping of signals.
Fig. 4 pH measurements of the suspension of materials in water.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
Scheme 3 Plausible mechanism that explain formation of materials D and E though a synergistic effect of DBU and DMF and transamidation
reaction.
the higher pH of material D is consistent with a higher amount below the entrance channel. The following step (transamidation
of basic nitrogens.
reaction, in red) starts with a new hydrogen bond interaction
With all this information we are in position to propose two between DMF and 1 leading to complex minA-3. In this case, the
ꢁ
1
possible pathways that explain the formation of materials D and stabilization energy is around ꢁ17.8 kJ mol . Finally, the latter
ꢁ
1
E. Under the non-anhydrous conditions and the continuous complex needs to overpass an energy barrier of ꢂ201 kJ mol
formation of water by heating GO, DBU could be easily hydro- to get the transamidated product 2 and HNMe , thus
2
lyzed to afford 1, which could evolve by attacking both the accounting for the lack of compound 2 if the reaction is per-
ꢀ
epoxide groups of GO as well as DMF to provide 2 as a result of formed at 80 C (see Table 2 and ESI†).
the transamidation reaction (Scheme 3, route A).
On the other hand, if DMF interacts rst with DBU (route B,
According to the observed cooperative effect of both DBU and Scheme 3) the complex formed is again a hydrogen bond-based
DMF in this transformation, we reckon that route B must be also structure (minB-1 in Fig. 5, red pathway). In this case, the
operative to some extent. In this route B, DBU could react with stabilization energy is less than the previous one observed in
ꢁ
1
DMF to evolve though intermediate I (le bottom side in Scheme route A (ꢂꢁ18.5 kJ mol ). The process towards compound 2
3
) towards hydrolysis and provide compound 2. Intermediate I and HNMe would begin by a reaction between DBU with DMF
2
has been suggested in several examples that proposed that DBU that, in the rst step, presents a high activation barrier
2
5–29
ꢁ1
could promote the transfer of acyl groups.§
Therefore inter- (ꢂ280 kJ mol ).
mediate I (route B) could promote other transamidation reactions
Since DFT calculations were performed in gas phase, the
of DMF if a suitable nucleophile is present (see Table 3, entry 1). formation of the cationic intermediate I is not probable due to
Nevertheless, as it will be shown below, some considerations have the high energy that should be needed to break DMF homo-
to be made regarding this intermediate I.
litically. Nevertheless, DFT calculations by hydrogen transfer
As a proof for the feasibility of these mechanisms, a density between both entities led us to compound 4. This could be an
functional theory (DFT) calculation was performed at B3LYP/6- intermediate to take into account when studying other related
3
0,31
3
11G(d,p) level of theory.
the different steps that could take place and the intermediates under the basic conditions used (DBU).
formed in the reaction between DBU and DMF in the presence The second step is the hydrolysis followed by a double
of water to explain the formation of HNMe , compound 2, and hydrogen transfer between the involved species to obtain 2 and
materials D and E. Therefore, if the reaction between water and HNMe . Although according to DFT calculations, route A is
Our main target was to highlight processes, as its formation should be very favored in solution
2
2
DBU takes place rst (route A, Scheme 3), the initial interme- more favored than route B, both pathways could be possible
diate is a hydrogen bond complex between both species (minA-1 under the experimental conditions applied to the systems under
ꢀ
in Fig. 5, green pathway). This interaction is exergonic by study (T ¼ 120 C).
ꢁ
1
a release of ꢂꢁ41 kJ mol in energy. This complex is trans-
2
In both routes the HNMe liberated to the reaction media is
formed into compound 1 in two steps, namely the addition of easily trapped by the GO shiing the equilibrium. In fact, the
the hydroxyl group which occurs as a consequence of hydrogen interaction of HNMe with a GO model (containing an epoxide,
transfer from a water molecule to DBU (TsA-12) involving an a hydroxyl and a carboxylic group) shows a great electrostatic
2
ꢁ
1
energy activation of ꢂ174.7 kJ mol , and a second intra- linkage between both species (Fig. 6). The estimation of the
32
molecular hydrogen transfer in minA-2 from the OH group to binding energy by means of ONIOM model reveals an energy
ꢁ
1
the nearest nitrogen atom (TsA-21) with an estimated activation value about ꢁ57 kJ mol , which is in the range of strong
ꢁ
1
33
barrier of ꢂ149.7 kJ mol . The resulting compound 1 is highly hydrogen bonds. This has been ratied by a non-covalent
ꢁ
1
stable in the potential energy surface, about ꢁ63.6 kJ mol
interaction (NCI) analysis which has shown a great attractive
character in the region where HNMe is attached to the GO
epoxide group (see structure (c) in Fig. 6).
2
§
DBU is considered non nucleophilic but it can also act as a nucleophile.
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
Fig. 5 Energy profile of the DBU and DMF reaction in presence of water to get HNMe
Scheme 3 (all values are in kJ mol ).
2
. The colours code follow-up the same reasoning as in
ꢁ1
the reaction between aromatic aldehydes and nucleophiles such
as malononitrile or ethyl 2-cyanoacetate to furnish Knoevenagel
adducts 5a, 5b in high yields and reasonable reaction times
(Table 4, entries 1–3). Furthermore, the analogous condensa-
tion employing ketones was also successful when longer reac-
tion times were applied, albeit material D proved much more
effective than material B (entries 4 and 5). Leaching of func-
tional groups from the catalyst was ruled out as the elemental
analysis of material D aer the Knoevenagel reaction (entry 2)
2
Fig. 6 Structures of the complex between a GO-model and HNMe .
(
a) On the plane interaction (b) on the side Interaction. (c) NCI-plot showed comparable results (C, 68.4; H, 3.7; N, 8.2; N/C, 0.120) as
analysis of structure. The values presented are the binding energies at those included in Table 1.
ꢁ
1
B3LYP//6-311G* level of theory in kJ mol
.
Compounds 5a, 5b served as excellent Michael acceptors in
the subsequent addition of nucleophiles. This transformation
could be also conducted in a one-pot Knoevenagel–Michael
sequence and hence nitroderivatives 6a, 6b were readily ob-
tained in the presence of excess of nitromethane in moderate
2
Once the generation of HNMe under the experimental
conditions has been established by the mechanism depicted
above (Fig. 5), and considering the ability of amines to get
attached to the GO surface, we can next propose a pathway for
this insertion. This can occur through a hydrogen transfer from
21
HNMe to GO-epoxide oxygen which is possible by overpassing
2
ꢁ
1
an activation barrier of about 177 kJ mol (Fig. 7). It is worth
noting that other possible pathways could take place under the
same conditions as deprotonation by DBU or the reduction of
34
2
rGO by DMF to produce HNMe as pointed out by Kim et al.
The usefulness of the newly prepared GO-derived materials
was next tested on some C–C bond-forming processes, such as
the Knoevenagel condensation that has also been performed
recently in the presence of other graphene-derived cata-
2
Fig. 7 Energy profile of a possible HNMe insertion in graphene oxide
12,35–39
lysts.
Thus, both materials B and D were able to promote achieved with ONIOM model where the high layer is optimized at
B3LYP/6-311G(d,p) and low layer at B3LYP/STO-3G levels of theory.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
Table 4 Synthetic processes catalyzed by GO-derived materials
0
a
b
Entry
X
R
R
Catalyst
5 (reaction time, yield)
6 (reaction time, yield )
1
2
3
4
5
Cl
Cl
Cl
Br
Br
H
H
H
Me
Me
CN
CN
CO
CN
CN
B
5a (6 h, 90%)
5a (2.5 h, 99%)
5b (18 h, 97%)
5c (96 h, 38% )
5c (96 h, 90%)
6a (36 h, 67%)
c
D
D
B
6a (20 h, 66% )
c
2
Et
6b (20 h, 36% )
d
D
a
b
c
Reaction time of the Michael addition step. Overall yield of the Knoevenagel–Michael sequence. Aer 72 h, 20–30% of compound 7 was also
d
formed. Conversion.
overall yields (Table 4, entries 1–3). It should be noted that which could occur via different mechanisms and could provide
variable amounts of dinitro compound 7 were produced, arising valuable information about the nature of both catalysts. The
from a retro-Michael reaction followed by the addition of Knoevenagel reaction could work either via iminium or basic
a second molecule of nitromethane.
activation and the subsequent Michael addition must occur via
0
These excellent results prompted us to reevaluate the prep- basic activation. In this case, both catalysts A and A behaved in
aration method of our former material A, but avoiding the use a very similar way to provide compound 6d (Scheme 5, path a).
of hydrazine as reducing agent. Thus, GO was rst reacted with On the contrary, the reaction of cinnamaldehyde and nitro-
12
piperazine as in Scheme 1, and further treatment with DMF methane via iminium ion activation produced a mixture of 1,2-
0
produced a rearomatization to provide reduced material A
Scheme 4).
and 1,4-addition products 8 and 9, respectively, in different
ratios depending on the catalyst employed (Scheme 5, path b). It
(
Interestingly, the elemental analysis showed an almost is expected that compound 9 was formed via iminium activation
identical ratio of nitrogen for both materials A (C, 66.5; H, whereas compound 8 could be formed via both mechanisms,
0
2
.8; N, 5.2; N/C, 0.078) and A (C, 66.1; H, 2.1; N, 5.1; N/C, 0.077). namely iminium and basic activation. The higher ratio of 8
0
0
Similarly to material B, FTIR analysis for material A also shows when using catalyst A corroborated the presence of Me N
2
the shiing of C]O and C]C stretches peaks when comparing moieties as additional basic centers that modulate the catalytic
2
with GO (restoration of the sp carbon network or removal of activity of the material, showing a decrease of the iminium
0
oxidative debris). Material A also showed the appearance of an pathway only possible through the piperazine moiety.
ꢁ1
intense and wide peak around 1200 cm that could correspond
to C–N stretching from both NMe
2
groups and piperazine
0
(Fig. 8). Comparison of TGA plots for materials A and A shows
different proles and weight losses that could be assigned to the
NMe groups that have been incorporated during DMF treat-
2
ment (see ESI†).
The catalytic performances of A and the newly prepared
0
material A were compared by performing several reactions
0
0
Fig. 8 FTIR spectra of GO and materials A and A .
Scheme 4 Preparation of material A .
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
Preparation of material C. A suspension of 600 mg of GO in
THF (100 mL) was sonicated for 30 min, whereupon DBU (900
mL, 916 mg, 6 mmol) was added to the suspension. The reaction
ꢀ
mixture was stirred at 70 C in a sealed tube for 96 h. Aer
cooling to room temperature, the material was ltered through
a lter plate, washing with deionized water (4 ꢃ 15 mL) until
neutral pH. Then, the material was washed with methanol (20
ꢀ
mL), acetone (20 mL) and le to stir in a vial tube at 70 C for
24 h.
Preparation of material D. A suspension of 1.5 g of GO in
DMF (110 mL) was sonicated for 30 min, whereupon DBU (2.25
mL, 2.29 g, 15 mmol) was added to the suspension. The reaction
mixture was stirred at 120 C in a sealed tube for 96 h. Aer
0
Scheme 5 Comparison between catalysts A and A in: (a) Knoevena-
gel–Michael sequence and (b) Michael versus Henry and elimination.
ꢀ
cooling to room temperature, the material was ltered through
a lter plate, washing with deionized water (5 ꢃ 20 mL) until
neutral pH. Then, the material was washed with methanol (20
These results demonstrate that DMF is not an innocent
solvent when using GO. In this context, it has been published
that DMF played a relevant role in the catalytic activity of CuO/
GO composites. Although in this case the metal has been
proposed as responsible of the cleavage of DMF, according to
our results it cannot be discarded an effect of the GO.
ꢀ
mL), acetone (20 mL) and le to stir in a vial tube at 70 C for
2
4 h.
Preparation of material E. A suspension of 500 mg of GO in
DMF (60 mL) was sonicated for 30 min, whereupon amine 1 (ref.
40
1
9) (600 mg, 3.53 mmol) was added to the suspension. The
ꢀ
reaction mixture was stirred at 120 C in a sealed tube for 24 h.
Aer cooling to room temperature, the material was ltered
through a lter plate, washing with deionized water (4 ꢃ 15 mL)
until neutral pH. Then, the material was washed with acetone (2
Conclusions
In conclusion, we have found that the reaction between DBU
and DMF triggers the formation of amine 1 and HNMe , and the
2
further covalent bonding of these amines to GO. Some experi-
ments and DFT calculations allowed us to propose a plausible
mechanism and reaffirm the characterization of the obtained
materials. In addition, our estimations showed a great
ꢀ
ꢃ 15 mL) and le to stir in a vial tube at 70 C for 24 h.
Preparation of material F. A suspension of 200 mg of GO in
2
deionized H O (5 mL) was sonicated for 30 min, whereupon
a solution of amine 1 (ref. 19) (342 mg, 2 mmol) in EtOH (5 mL)
was added to the suspension. The reaction mixture was stirred
ꢀ
at 100 C in a sealed tube for 24 h. Aer cooling to room
2
hydrogen bond interaction between HNMe and GO. Under-
temperature, the material was ltered through a lter plate,
washing with deionized water (3 ꢃ 15 mL) until neutral pH.
Then, the material was washed with acetone (2 ꢃ 15 mL) and
standing this type of processes is essential to control graphene
functionalization, which in turn is fundamental to control the
properties and applications of graphene derivatives. We have
demonstrated the catalytic potential of the resulting materials
and proved that the use of DMF for the solvothermal reduction
of functionalized graphene derivatives could be a greener
option but it may also change the properties of the material
compared to other reduction methods. Moreover, this study
reveals that the use of DMF as solvent to introduce amines into
GO via epoxide opening would undergo transamidation reac-
ꢀ
le to stir in a vial tube at 70 C for 24 h.
0
Preparation of material A . A suspension of 500 mg of GO-
12
NH in DMF (50 mL) was sonicated for 30 min. The reaction
ꢀ
mixture was stirred at 120 C in a sealed tube for 72 h. Aer
cooling to room temperature, the material was ltered through
a lter plate, washing with deionized water (5 ꢃ 20 mL) until
neutral pH. Then, the material was washed with acetone (2 ꢃ 20
ꢀ
mL) and le to stir in a vial tube at 70 C for 24 h.
2
tions and introduction of NMe groups. Therefore, these results
must be considered when DMF and/or DBU are used in reac-
tions promoted by GO as the introduction of basic groups might
change the properties of the desired materials.
Knoevenagel reactions
Reactions with aldehydes. A suspension of 20 mg of the
corresponding material in EtOH (1 mL) was sonicated for
Experimental and methods
Preparation of materials B–F and A
10 min. Then, benzaldehyde or p-chlorobenzaldehyde (0.123
0
mmol) and the corresponding nucleophile (malononitrile or
Preparation of material B. A suspension of 100 mg of GO in ethyl 2-cyanoacetate, 0.135 mmol) were added. The mixture was
ꢀ
DMF (5 mL) was sonicated for 30 min. The reaction mixture was stirred at 70 C until completion, whereupon the crude was
ꢀ
stirred at 120 C in a sealed tube for 96 h. Aer cooling to room ltered to remove the material and the solvent was evaporated
temperature, the material was ltered through a lter plate, at reduced pressure, to afford the corresponding product 5a, 5b,
washing with deionized water (4 ꢃ 15 mL) until neutral pH. 5d, in the yields and reaction times indicated in Table 4 and
Then, the material was washed with methanol (20 mL), acetone Scheme 5. Characterization data of 5a, 5b, 5d were in agreement
ꢀ
41–43
(
20 mL) and le to stir in a vial tube at 70 C for 24 h.
with those previously reported.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
Journal of Materials Chemistry A
Reactions with ketones. A suspension of 30 mg of the cor- by comparing the experimental infrared spectrum and theo-
responding material in EtOH (1 mL) was sonicated for 10 min. retical ones. The nice coincidence between the major vibrations
Then, p-bromoacetophenone (19.9 mg, 0.1 mmol) and malo- was the decisive factor for our choice (see ESI†). As a comple-
nonitrile (7.2 mg, 0.1 mmol) were added. The mixture was ment of our exploration a non covalent interaction analysis was
ꢀ
52,53
stirred at 70 C for 96 h, whereupon the crude was ltered to done by means of NCI-plot proposed by Yang et al.
remove the material and the solvent was evaporated at reduced
pressure to afford the corresponding product 5c in the yields
Conflicts of interest
indicated in Table 4. Characterization data of 5c were in
44
agreement with those previously reported.
There are no conicts to declare.
Knoevenagel–Michael one-pot sequence
Acknowledgements
Reaction with nitromethane. A suspension of 20 mg of the
corresponding material in EtOH (1 mL) was sonicated for
0 min. Then, benzaldehyde or p-chlorobenzaldehyde (0.123
We acknowledge the Spanish Government (CTQ-2012-35957,
CTQ2016-78779-R, CTQ2015-63997-C2) for nancial support,
Beatriz Garc ´ı a Alcubilla (UAM), Jorge Legan ´e s (UAM) and
Miguel Mart ´ı n Arroyo (UAM) for some initial experiments; Dr
J. L. Garc ´ı a Fierro (ICP-CSIC) for performing XPS analyses and
Dr M. Jos ´e de la Mata (SIDI-UAM) for useful discussions. A
generous allocation of computing time at the Centro de
Computaci ´o n Cient ´ı ca of the UAM is also acknowledged.
1
mmol) and the corresponding nucleophile (malononitrile or
ethyl 2-cyanoacetate, 0.135 mmol) were added. The mixture was
ꢀ
stirred at 70 C until completion (TLC analysis), and nitro-
methane (131 mL, 2.45 mmol, 20 equiv.) was then added. The
ꢀ
mixture was stirred at 70 C until completion of the reaction,
whereupon the crude was ltered to remove the material and
the solvent evaporated at reduced pressure. The crude was
puried by ash column chromatography (hexane/EtOAc, 5 : 1)
to afford products 6a, 6b, 6d and 7 in the yields and reaction
times indicated in Table 4 and Scheme 5. Characterization data
of 6a, 6b, 6d and 7 were in agreement with those previously
Notes and references
S. Naval ´o n, A. Dhakshinamoorthy, M. Alvaro and H. Garcıa,
´
´
1
2
3
4
5
6
7
8
9
Chem. Rev., 2014, 114, 6179–6212.
C. K. Chua and M. Pumera, Chem.–Eur. J., 2015, 21, 12550–
45–48
reported.
12562.
X. Fan, G. Zhang and F. Zhang, Chem. Soc. Rev., 2015, 44,
3023–3035.
Michael/Henry addition of nitromethane to enals
D. Deng, K. S. Novoselov, Q. Fu, N. Zheng, Z. Tian and X. Bao,
Nat. Nanotechnol., 2016, 11, 218–230.
Reaction with cinnamaldehyde. A suspension of 50 mg of the
2
corresponding material in MeNO (2 mL) was sonicated for
P. Tang, G. Hu, M. Li and D. Ma, ACS Catal., 2016, 6, 6948–
6958.
1
0 min. Then, cinnamaldehyde (38 mL, 0.3 mmol) was added.
ꢀ
The mixture was stirred at 60 C for 72 h, whereupon the crude
was ltered to remove the material and the solvent evaporated
at reduced pressure. The crude was puried by ash column
chromatography (cyclohexane/EtOAc, 10 : 1 to 5 : 1) to afford
products 8 and 9 in the yields and reaction times indicated in
Scheme 5. Characterization data of 8 and 9 were in agreement
M. Hu, Z. Yao and X. Wang, Ind. Eng. Chem. Res., 2017, 56,
3
477–3502.
X.-K. Kong, C.-L. Chen and Q.-W. Chen, Chem. Soc. Rev.,
014, 43, 2841–2857.
2
D. R. Dreyer, A. D. Todd and C. W. Bielawski, Chem. Soc. Rev.,
2014, 43, 5288–5301.
49,50
with those previously reported.
S. Eigler, Chem.–Eur. J., 2016, 22, 7012–7027.
1
0 C. K. Chua and M. Pumera, Chem. Soc. Rev., 2014, 43, 291–
Computational details
312.
The structure optimization together with the frequencies esti- 11 C. K. Chua and M. Pumera, Chem. Commun., 2016, 52, 72–75.
mation along the potential energy surface were done at B3LYP/ 12 E. Rodrigo, B. Garc ´ı a Alcubilla, R. Sainz, J. L. Garc ´ı a Fierro,
30,31
6-311G(d,p) level of theory.
Starting from an asymmetric
R. Ferritto and M. B. Cid, Chem. Commun., 2014, 50, 6270–
distribution of the functional groups in the graphene oxide
6273.
32
which lead to a non periodic system the Oniom model was 13 D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, Chem.
51
choose as implemented in Gaussian 09 series of programs.
The optimization of the high layer was achieved at B3LYP/6- 14 K. Ai, Y. Liu, L. Lu, X. Cheng and L. Huo, J. Mater. Chem.,
11G(d,p) while for the low layer the level of theory was 2011, 21, 3365–3370.
B3LYP/STO-3G. The election of graphene model was condi- 15 C. Bosch-Navarro, E. Coronado and C. Mart ´ı -Gastaldo,
tioned by the active functional groups (epoxide, hydroxyl and Carbon, 2013, 54, 201–207.
carboxylic groups) and the size of the active region (high layer of 16 Y. Shen, H.-B. Zhang, H. Zhang, W. Ren, A. Dasari, G.-S. Tang
Oniom calculation) where the interaction could take place. The and Z.-Z. Yu, Carbon, 2013, 56, 132–138.
Soc. Rev., 2010, 39, 228–240.
3
objective was to reduce the computational cost achieving 17 J. Muzart, Tetrahedron, 2009, 65, 8313–8323.
a reasonable quantum calculation quality. So, we tried different 18 S. Ding and N. Jiao, Angew. Chem., Int. Ed., 2012, 51, 9226–
sizes of graphene and different positions of functional groups
9237.
This journal is © The Royal Society of Chemistry 2018
J. Mater. Chem. A

View Article Online
Journal of Materials Chemistry A
Paper
19 R. S. Gonçalves, P. V. Abdelnur, V. G. Santos, R. C. Simas, 37 S. M. Islam, A. S. Roy, R. C. Dey and S. Paul, J. Mol. Catal. A:
M. N. Eberlin, A. Magalh ˜a es and E. R. P ´e rez Gonz ´a lez,
Amino Acids, 2010, 40, 197–204.
0 X. Zhang, C. Yu, C. Wang, Z. Wang and J. Qiu, Mater. Res.
Bull., 2015, 67, 77–82.
1 A. B. Bourlinos, D. Gournis, D. Petridis, T. Szab ´o , A. Szeri and
I. D ´e k ´a ny, Langmuir, 2003, 19, 6050–6055.
2 J. P. Rourke, P. A. Pandey, J. J. Moore, M. Bates, I. A. Kinloch,
Chem., 2014, 394, 66–73.
38 B. Xue, J. Zhu, N. Liu and Y. Li, Catal. Commun., 2015, 64,
105–109.
39 S. Rana and S. B. Jonnalagadda, Catal. Commun., 2017, 92,
31–34.
40 S. Zhang, W. Gao, J. Li, X. Zhou and Y. Qu, ACS Appl. Mater.
Interfaces, 2014, 6, 22174–22182.
2
2
2
R. J. Young and N. R. Wilson, Angew. Chem., Int. Ed., 2011, 41 K. Yamashita, T. Tanaka and M. Hayashi, Tetrahedron, 2005,
0, 3173–3177. 61, 7981–7985.
3 S. Rani, M. Kumar, R. Garg, S. Sharma and D. Kumar, IEEE 42 J. S. Yadav, B. V. Subba Reddy, A. K. Basak, B. Visali,
5
2
2
2
2
2
Sens. J., 2016, 16, 2929–2934.
4 I. A. Vacchi, C. Spinato, J. Raya, A. Bianco and C. M ´e nard-
Moyon, Nanoscale, 2016, 8, 13714–13721.
A. Venkat Narsaiah and K. Nagaiah, Eur. J. Org. Chem.,
2004, 546–551.
43 J.-Y. Liu, Y.-J. Jang, W.-W. Lin, J.-T. Liu and C.-F. Yao, J. Org.
Chem., 2003, 68, 4030–4038.
5 F. Aric `o , S. Evaristo and P. Tundo, Green Chem., 2015, 17,
1
176–1185.
6 M. Carafa, E. Mesto and E. Quaranta, Eur. J. Org. Chem.,
011, 2458–2465.
44 D. M. Barnes, A. R. Haight, T. Hameury, M. A. McLaughlin,
J. Mei, J. S. Tedrow and J. Dalla Riva Toma, Tetrahedron,
2006, 62, 11311–11319.
2
7 C. Larriv ´e e-Aboussafy, B. P. Jones, K. E. Price, M. A. Hardink, 45 H.-M. Guo, J.-G. Li, G.-R. Qu, X.-M. Zhang and W.-C. Yuan,
R. W. McLaughlin, B. M. Lillie, J. M. Hawkins and
R. Vaidyanathan, Org. Lett., 2010, 12, 324–327.
8 W.-C. Shieh, S. Dell and O. Repi ˇc , J. Org. Chem., 2002, 67,
Chirality, 2011, 23, 514–518.
46 M. Bera, T. Kanti Ghosh, B. Akhuli and P. Ghosh, J. Mol.
Catal. A: Chem., 2015, 408, 287–295.
2
2
3
2
188–2191.
9 V. K. Aggarwal and A. Mereu, Chem. Commun., 1999, 2311–
312.
47 P. Kotrusz, S. Toma, H.-G. Schmalz and A. Adler, Eur. J. Org.
Chem., 2004, 1577–1583.
48 N. R ´ı os-Lombarda, E. Busto, V. Gotor-Fern ´a ndez and
V. Gotor, Eur. J. Org. Chem., 2010, 484–493.
49 C. Dockendorff, S. Sahli, M. Olsen, L. Milhau and
M. Lautens, J. Am. Chem. Soc., 2005, 127, 15028–15029.
2
0 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter
Mater. Phys., 1988, 37, 785–789.
31 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
3
2 S. Dapprich, I. Kom ´a romi, K. S. Byun, K. Morokuma and 50 I. Mager and K. Zeitler, Org. Lett., 2010, 12, 1480–1483.
M. J. Frisch, J. Mol. Struct.: THEOCHEM, 1999, 461–462, 1–21. 51 J. M. Frisch, et al., Gaussian 09, Revision E01, Gaussian, Inc.,
33 I. Alkorta, I. Rozas and J. Elguero, Chem. Soc. Rev., 1998, 27,
Wallingford CT, 2010.
1
63–170.
52 E. R. Johnson, S. Keinan, P. Mori-S ´a nchez, J. Contreras-
Garc ´ı a, A. J. Cohen and W. Yang, J. Am. Chem. Soc., 2010,
132, 6498–6506.
53 J. Contreras-Garc ´ı a, E. R. Johnson, S. Keinan, R. Chaudret,
J.-P. Piquemal, D. N. Beratan and W. Yang, J. Chem. Theory
Comput., 2011, 7, 625–632.
3
3
4 S. Kim, K. Choi and S. Park, Solid State Sci., 2016, 61, 40–43.
5 T. Wu, X. Wang, H. Qiu, J. Gao, W. Wang and Y. Liu, J. Mater.
Chem., 2012, 22, 4772–4779.
3
6 F. Zhang, H. Jiang, X. Li, X. Wu and H. Li, ACS Catal., 2014, 4,
394–401.
J. Mater. Chem. A
This journal is © The Royal Society of Chemistry 2018