Novel λ3-Iodane-Based Functionalization of Synthetic Carbon Allotropes (SCAs) - Common Concepts and Quantification of the Degree of Addition

Article abstract of DOI:10.1002/chem.201404662

The covalent functionalization of carbon allotropes represents a main topic in the growing field of nano materials. However, the development of functional architectures is impeded by the intrinsic polydispersibility of the respective starting material, th

Full text of DOI:10.1002/chem.201404662

DOI: 10.1002/chem.201404662
Full Paper
&
Synthetic Carbon Allotropes
Novel l3-Iodane-Based Functionalization of Synthetic Carbon
Allotropes (SCAs)—Common Concepts and Quantification of the
Degree of Addition
Ferdinand Hof, Ricarda A. Schꢀfer, Cornelius Weiss, Frank Hauke, and Andreas Hirsch*[a]
Abstract: The covalent functionalization of carbon allotropes
represents a main topic in the growing field of nano materi-
als. However, the development of functional architectures is
impeded by the intrinsic polydispersibility of the respective
starting material, the unequivocal characterization of the in-
troduced functional moieties, and the exact determination
of the degree of functionalization. Based on a novel carbon
allotrope functionalization reaction, utilizing l3-iodanes as
radical precursor systems, we were able to demonstrate the
feasibility to separate and to quantify thermally detached
functional groups, formerly covalently linked to carbon
nanotubes and graphene through thermogravimetric GC-
MS.
Introduction
been addressed or solved so far: 1) the exact quantification of
the degree of addition of a specific functional entity and 2) a
direct comparison of the reactivity of different SCAs under
equivalent reaction conditions. Herein, we provide a profound
answer for these two fundamental challenges. Based on
a novel and versatile reductive arylation of graphite/graphene
and carbon nanotubes—utilizing l3-iodane precursors—the
degree of functionalization has been quantified by a thermog-
ravimetric product analysis, coupled with gas-chromatographic
separation and mass-spectrometric characterization (TG-GC-
MS). With this setup, we are now able to separate the thermal-
ly detached functional entities and to selectively identify and
quantify each individual component attached to the carbon al-
lotrope framework.
Inspired by the rich chemistry of fullerenes, the functionaliza-
tion of other synthetic carbon allotropes (SCAs), in particular
carbon nanotubes and more recently graphene, has steadily
been progressed.[1] In parallel to the development of efficient
reaction protocols, such as the reductive alkylation and aryla-
tion of SCAs,[2] tools for product characterization have continu-
ously been adjusted and improved.[3] Typical techniques are:
thermogravimetric analysis,[4] Raman spectroscopy,[3g] STM/
AFM,[5] and XPS.[6] However, the powerful analytical tool-kit,
routinely used for the structural characterization of organic
molecules, such as NMR spectroscopy, mass spectrometry, and
X-ray crystallography can hardly be applied for the analysis of
covalently functionalized carbon nanotubes and graphene,
due to their intrinsic polydisperse nature. Here, only mixtures
of adducts with a broad distribution of sizes, shapes, and de-
grees of addition are accessible for analysis. Therefore, only
averaging information can be obtained, which requires the de-
velopment/application of techniques for data acquisition
founded on a broad statistical basis. Along these lines, we
have recently introduced scanning Raman spectroscopy (SRS)
and scanning Raman microscopy (SRM) as versatile techniques
for a reliable and reproducible characterization of covalently
functionalized SCA bulk samples.[3f,g] Despite these improve-
ments, there are still fundamental questions that have not
Results and Discussion
For the l3-iodane-based reductive arylation, the respective SCA
was initially reduced with a stoichiometric amount of potassi-
um yielding the corresponding nanotubide and graphenide
salts,[7] which represent highly activated intermediates. We
have recently shown that the stoichiometry of potassium is
a critical factor to fine tune the degree of addition in the final
single-wall carbon nanotube (SWCNT) derivatives.[8] Therefore,
the investigation of the reaction of nanotubides and graphe-
nides with different l3-iodanes was carried out by the variation
of the potassium/carbon ratio (T(1:4), T(1:8), T(1:16), and T(1:24) as
well as G(1:4), G(1:8), G(1:16), and G(1:24) exhibiting a respective
carbon to potassium ration of 1:4, 1:8, 1:16, and 1:24)
(Scheme 1).
[a] F. Hof, R. A. Schꢀfer, C. Weiss, Dr. F. Hauke, Prof. Dr. A. Hirsch
Department of Chemistry and Pharmacy
and Institute of Advanced Materials and Processes (ZMP)
Friedrich-Alexander-Universitꢀt Erlangen–Nꢁrnberg (FAU)
Henkestrasse 42, 91054 Erlangen (Germany)
Fax: (+49)9131-85-26864
Hypervalent iodine compounds easily form radicals and can
be used as functional-group transfer reagents.[9] In the SCA
chemistry they have so far only been applied for the electro-
chemical functionalization of carbon electrodes[10] as well as
other types of surfaces.[11] In line with the radical mechanisms
E-mail: andreas.hirsch@fau.de
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201404662.
Chem. Eur. J. 2014, 20, 16644 – 16651
16644
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 1. a) Diaryliodonium triflates A, B and C. b) Reductive activation and dispersion of SWCNTs/graphite with variation of the potassium concentration.
c) Covalent functionalization of negatively charged carbon intermediates with l3-iodanes.
stated for reductive arylation/alkylation reactions and diazoni-
Table 1. Statistical Raman spectroscopic data of the reaction products
T(1:n)(A–C) and G(1:n)(A–C)—I(D/G) values with standard deviation D.
um-based SCA functionalizations,[12] an equivalent pathway can
be anticipated for the reaction of l3-iodanes. In order to screen
SWCNT I(D/G)
(l=633 nm)
D
Graphite/graphene I(D/G)
(l=532 nm)
D
their reactivity as SCA trapping electrophiles, three different
para-functionalized l3-iodanes, that is, bis-(4-(tert-butyl)phenyl)
iodonium trifluoromethane-sulfonate (A), (4-bromophenyl)(4-
(tert-butyl)phenyl iodonium trifluoro-methanesulfonate (B), and
bis-(4-bromophenyl) iodonium trifluoro-methanesulfonate (C),
have been synthesized by adjusting literature-known proce-
dures (for details see the Supporting Information).[13] For the
synthesis of the arylated SCA derivatives, the corresponding
potassium nanotubide/graphenide salt was dispersed in THFabs
(glovebox: Ar) and subsequently 0.5 equivalents per carbon of
the respective l3-iodane A–C were added. The isolation of the
reaction products T(A–C) and G(A–C) was accomplished by
aqueous workup and filtration. In order to gain statistically sig-
nificant bulk information about the success of this covalent
functionalization pathway scanning Raman spectroscopy was
carried out. Here, the I(D/G) (intensity ratio of the D- and G-
bands) values of the individual Raman spectra (at least 1.250
single-point spectra measured) are plotted with respect to
their frequency and the respective distribution function was
determined (Figure 1).
T
(1:4)A
0.95
0.15 G(1:4)
0.11 G(1:4)
0.17 G(1:4)
0.17 G(1:8)
0.24 G(1:16)
0.09 G(1:24)
A
B
C
A
A
A
1.00
0.30
0.30
0.30
0.11
0.30
0.47
T(1:4)
T(1:4)
T(1:8)
T(1:16)
T(1:24)
B
C
A
A
A
0.90
0.93
0.68
0.42
0.26
1.05
0.95
1.25
1.05
0.81
rivatives (lexc =532 nm, Figure 1, bottom) (see also Table 1). For
further Raman data see Figures S1 and S2 in the Supporting In-
formation.
The potential of this novel synthetic route becomes appar-
ent, when the respective I(D/G) values of the carbon nanotube
samples T(1:4)(A–C) are compared with reference samples pre-
pared by literature-known synthetic pathways obtained by re-
ductive hexylation T(1:4)Hex-I or reductive arylation T(1:4)Ph-I of
carbon nanotubes[15] (see Figure 2).
For arylated SWCNT derivatives accessible by our novel l3-
iodane-based functionalization sequence, the degrees of func-
tional group addition, represented by the respective RDIs, dras-
tically exceed the values obtained for systems synthesized by
classical pathways:[15] for example, RDI(T(1:4)A)=1.83 versus
RDI(T(1:4)Ph-I)=0.83. In accordance with our latest results[8] we
also observe a pronounced dependence of the Raman derived
In principle, the D-band intensity is a measure for the
amount of lattice sp3 centers introduced by covalent addend
binding. The successful reaction of the reduced SCAs with the
l3-iodanes A–C can exemplarily be demonstrated (K/C=1:4)
on the basis of the Raman data of the functionalized SWCNTs
(lexc =633 nm, Figure 1, top) and the graphite/graphene[14] de-
Chem. Eur. J. 2014, 20, 16644 – 16651
www.chemeurj.org
16645
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
final degree of functionalization
increases with the respective
amount of potassium used in
the reductive SCA activation
step.
Furthermore, we were able to
show that these covalently
bound aryl moieties can reversi-
bly be detached from the
carbon allotrope framework
either by laser-induced de-func-
tionalization (exemplarily shown
for
T(4:1)A; lexc =785 nm, E=
10 mW, Figure 3, left, Figure S7
in the Supporting Information)
or thermally. The latter approach
opens the possibility to monitor
the detachment progress of the
functional moieties directly by
Raman spectroscopy (Figure 2,
right, Figure S8 in the Support-
ing Information).
Here, for both functionalized
SCA
derivatives—T(1:4)
A
and
Figure 1. Statistical Raman spectroscopic analysis of the functionalized carbon allotrope T(1:4)A (lexc =633 nm) and
G(1:4)A—a continuous decrease of
the D-band intensity is observed
starting from around 1508C
up to 5008C. At temperatures
G
(1:4)A (lexc =532 nm). a) Histogram and b) mean spectrum of T(1:4)A. c) Histogram and d) mean spectrum of G(1:4)A.
around 5008C, the characteristic Raman spectrum of the re-
spective starting material is restored exhibiting an I(D/G) value of
less than 0.1 in the bulk (Figure S8 in the Supporting Informa-
tion). However, neither by SRS nor by temperature-depending
Raman spectroscopy any conclusion about the chemical
nature of the attached moieties can be extracted. Thus, ther-
mogravimetric analysis coupled to mass spectrometry (TG-MS)
was carried out (20–7008C, constant flow of N2 (70 mLminÀ1
,
Figures S9–S14 in the Supporting Information). In the region
between 20 and 1308C, mainly physisorbed THF and water is
released from the samples T(1:4)(A–C). According to the respec-
tive mass traces, sample mass loss beyond 5008C mainly can
be traced back to the degradation of the carbon framework.
The predominant mass loss of 23.2 (T(1:4)A), 17.9 (T(1:4)B), and
28.0% (T(1:4)C) is detected between 130 and 5008C and nicely
correlates with the respective flux of the detected ion currents,
which are the characteristic fragments originating from the de-
tached tert-butylbenzene and/or bromobenzene moieties (Fig-
ures S9–S11 in the Supporting Information). The same consis-
tent data set was obtained for the functionalized graphite/gra-
phene derivatives G(1:4)(A–C). However, in this case the ob-
served overall mass loss is considerably lower: 5.4 (G(1:4)A), 10.5
(G(1:4)B), and 3.2% (G(1:4)C) (see Figures S12–S14 in the Support-
ing Information). The TG-MS-based data nicely corroborates
the Raman results regarding an efficient covalent addend bind-
ing in both carbon allotropes. Nevertheless, up to now it was
not possible to determine an exact value of the degree of the
functionalization of covalently modified SCAs as Raman spec-
troscopy only allows to draw information about the amount of
Figure 2. 2D Raman index plot based on the statistical Raman analysis of
T
(1:4)A, T(1:4)B, T(1:4)C, and the alkylated and arylated reference samples
T(1:4)Hex-I and T(1:4)Ph-I. The respective values for the Raman defect index
(RDI), the Raman homogeneity index (RHI), and the Raman selectivity index
(RSI) are summarized in Figure S3 in the Supporting Information.
I(D/G) value on the amount of potassium used for the reduction
for the SWCNT samples and for the graphite/graphene samples
(Figures S4–S6 in the Supporting Information; Table 1). The
Chem. Eur. J. 2014, 20, 16644 – 16651
www.chemeurj.org
16646
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
be identified, exhibiting a peak
area ratio of almost 1:1 (Figur-
es S16 and S17 in the Support-
ing Information).
This result clearly highlights
the potential and versatility of
our novel functionalization se-
quence. Starting from an easily
accessible l3-iodane precursor
system with two different aryl
moieties, carbon-allotrope-based
architectures with complementa-
ry functional groups—like for in-
Figure 3. Left: Laser-induced thermal de-functionalization of the functionalized bulk material T(1:4)A demonstrating
the reversibility of addend binding. Regions in blue color denote areas with lower degrees of functionalization
based on a thermal de-functionalization by “writing” with high laser power. Right: Temperature-dependent Raman
spectroscopy of tert-butylphenyl-functionalized SWCNTs (lexc =633 nm).
stance
donor/acceptor
sys-
tems—can be built up in an
one-pot synthesis.
symmetry breaking defects in the sp2 carbon lattice and ther-
mogravimetric measurements provide only information about
the overall sample mass loss attributed to a thermal detach-
ment of all physisorbed and chemisorbed species. To circum-
vent these obstacles we employed thermogravimetric-based,
gas chromatography coupled to mass spectrometry (TG-GC-
MS, see Figure 4). This allowed us to determine the mass-loss
region, where the covalently bound addend is detached—
based on this data an exact quantification of the amount of
bound functional groups is possible.
Next to the identification of the main components by TG-
GC-MS, formation of byproducts can also be investigated by
this setup. Here, in all three functionalized SWCNT samples
(Figure S16 in the Supporting Information), additional compo-
nents (below 3% total peak area) are detected by GC (reten-
tion time of 1.93, 1.95, and 2.04 min), which can be attributed
to THF (solvent) and degradation products thereof (furane, di-
hydrofurane). Remarkably, in the case of the graphite/graphene
adducts
G(1:4)(A–C) the peak for THF (retention time of
2.04 min) is observed as main component (ꢀ80%) (Figure S17
in the Supporting Information). Moreover, in direct compari-
son, the peak area of the detached tert-butylphenyl moieties
for the SWCNT derivatives T(1:4)A and T(1:4)B is by two orders of
magnitude higher than that detected for the corresponding
graphite/graphene adducts G(1:4)A and G(1:4)B. Apparently, the
degree of covalent functionalization is considerably higher for
carbon nanotubes. These results can be attributed to the fact
that in the case of graphite, the reduction and dispersing steps
do not exclusively yield completely exfoliated and individual-
ized graphene sheets. The detection of THF as major compo-
nent is likely to be traced back to intercalated solvent mole-
cules, which are trapped between the carbon sheets. This also
reduces the accessible carbon surface area for the hypervalent
For this purpose, 1 mL of the carrier gas was collected at
a sample temperature of 2508C (highest mass flux) and the in-
dividual components were separated by a high-performance
GC column (Elite-5MS capillary column)—temperature gradi-
ent: 10 KminÀ1, T=40–2408C. Subsequently, the respective
sample fractions were analyzed by mass spectrometry. The cor-
responding retention times are depicted in Figure 4, right.
Here, the bromophenyl moiety of T(1:4)B and T(1:4)C is detected
at a retention time of 4.93 min, whereas the tert-butylphenyl
addend in T(1:4)A and T(1:4)B is detected at a retention time of
5.84 min (MS fragmentation pattern: see Figure S15 in the Sup-
porting Information). In the mixed functionalized systems
T(1:4)B/G(1:4)B both thermally detached functional moieties can
Figure 4. TG-GC-MS analysis of the functionalized SWCNT samples T(1:4)A, T(1:4)B, and T(1:4)C. Left: Plot of mass loss versus the temperature. Right: GC-MS analysis
of 1 mL gaseous addends detached at 2508C.
Chem. Eur. J. 2014, 20, 16644 – 16651
www.chemeurj.org
16647
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
iodine compounds, which decreases the overall degree of
functionalization and yields an explanation for the high
amount of THF in the TG-GC-MS experiments. This rationale
could be corroborated by a reference experiment where
a graphite starting material with initial smaller flake size was
chosen. Here, only a negligible amount of THF in relation to
the tert-butylphenyl signal is observed by GC (Figure S18 in
the Supporting Information).
of the peak area of the corresponding tert-butylbenzene de-
tached from the respective SCA (0.95 for T(1:4)A and 0.16 for
G(1:4)A) (Figure S22 in the Supporting Information). For each
thermogravimetric experiment the corresponding amounts of
tert-butylbenzene were extracted (Figure S23 in the Supporting
Information) and were correlated to the peak area of the GC-
MS elugram by using different weight portions of samples
during the thermogravimetric analysis of the specified samples
versus the integrated peak area of the GC-MS trace. The linear
fit (Figure 6) yields the pre-factor (3.24Æ0.14)ꢂ10À11 (standard
deviation: 4%) and provides the basis for the direct quantifica-
tion of the functional groups attached to carbon nanotubes as
well as to graphene/graphite. On the basis of the experimen-
In order to confirm that the obtained GC-MS traces originate
from covalently bound functional groups and not from interca-
lated l3-iodanes, we carried out an additional reference experi-
ment where the hypervalent iodine compound B was blended
with graphite (Figure 5). The GC elugram of the reference
tally
determined calibration
factor, the amount of tert-butyl-
benzene in the mixed-functional-
ized samples T(1:4)B and G(1:4)
B
can be calculated as 1.98 and
0.01 mmol, respectively. Accord-
ing to the detected 1:1 ratio of
the peak areas (tert-butylben-
zene/bromobenzene) it is fur-
thermore possible to quantify
the amount of the detached
bromobenzene
groups
in
T(1:4) (2.03 mmol) and T(1:4)
B
C
Figure 5. Reference experiment—graphite/l3-iodane blend. Left: Elugram of the pristine l3-iodane B
(TTG=2308C) thermal fragmentation products. The mass-spectrometric-identified compounds are attributed to
the respective retention times of 4.93, 5.84, 9.51, 10.62, 10.85, and 11.86 min. Right: For comparison—elugram of
the covalently functionalized SWCNT derivative T(1:4)B.
(3.25 mmol) (Figure S16 in the
Supporting Information). An
amount of 0.010 mmol was
sample differs distinctively from the elugrams of the covalently
functionalized SCA derivatives as it contains not only the char-
acteristic fragmentation pattern of the tert-butyl- and bromo-
phenyl groups, but also several other compounds originating
from the trifluoromethyl sulfonate counter ion of the l3-iodane
B (not found for covalently functionalized samples).
The separation of thermally detached functional moieties
and the identification of their chemical identity is a fundamen-
tal advantage provided by the TG-GC-MS setup. This fact is
also nicely illustrated by the characterization of an n-hexyl-
functionalized SWCNT reference sample (Figure S19 in the Sup-
porting Information). Here, besides n-hexane as main compo-
nent, characteristic byproducts[8] like hexanol, hexanal, and
hexyliodide are also individually separated by GC and identi-
fied by MS. These results nicely confirm our published data
concerning the multifunctional product distribution of reduc-
Figure 6. TG-GC-MS calibration curve by using different sample weights
tively functionalized SWCNT samples and highlight the
demand of powerful analytical techniques for their investiga-
tion. Nevertheless, for an exact determination of the amount
of covalently attached addends a reliable calibration for the
TG-GC-MS-based analysis is a fundamental prerequisite. For
this purpose, we systematically varied the amount of T(1:4)A
and G(1:4)A (Figures S20 and S21 in the Supporting Information).
As the detachment of the addends takes place mainly in the
temperature regime between 130 and 5008C, the detected
overall mass loss in this region was weighed by the percentage
&
(0.5–8.3 mg) of T(1:4)A and G(1:4)B ( , c=linear fit) during the thermogravi-
metric analysis. R2 =0.9835, Dx=(3.24Æ0.14)ꢂ10À11
.
found for G(1:4)B and of 0.026 mmol for G(1:4)C (Figure S17 in the
Supporting Information).
Furthermore, the amount of potassium used in the SCA acti-
vation step (Figures S24 and S25 in the Supporting Informa-
tion) can be correlated with the final amount of attached func-
Chem. Eur. J. 2014, 20, 16644 – 16651
www.chemeurj.org
16648
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tional moieties. The respective data is collected in Table ST2 in
the Supporting Information.
trapping electrophile is covalently bound. As mentioned
above, hypervalent iodine compounds can easily form radicals
and therefore a similar radical mechanism can be anticipated
for our novel l3-iodane-based functionalization route. First evi-
dence for this assumption was obtained from a GC-MS-based
byproduct analysis. Here, the organic components in the sol-
vent phase, obtained after the aqueous workup of the primary
reaction mixture—l3-iodane C, were separated by GC and
identified by mass spectrometry. Besides 1-bromo-4-iodoben-
zene as main component, biphenyl derivatives as characteristic
radical recombination products were detected (Figure S30 in
the Supporting Information).
For the first time it is now possible to exactly quantify the
degree of functionalization (f-addends in relation to SCA lattice
carbon atoms, Table 2) of a given functionalization sequence,
as the region of mass loss attributed to the functional-group
detachment is identified by TG-MS and the amount of a specific
Table 2. Correlation of the amount of attached functional addends with
the amount of lattice carbon atoms in the respective functionalized SCA
derivatives.
SWCNT f-Addend per lattice
Graphite/gra-
phene
f-Addend per lattice
carbon [%]
carbon [%]
T(1:4)
T(1:4)
T(1:4)
T(1:8)
T(1:16)
T(1:24)
A
B
C
A
A
A
3.42
2.51
2.02
2.45
0.80
0.25
G(1:4)
G(1:4)
G(1:4)
G(1:8)
G1:16)
G(1:24)
A
B
C
A
A
A
0.04
0.02
0.01
0.13
0.04
0.02
Conclusion
In conclusion, we have introduced a new l3-iodane-based reac-
tion sequence as a versatile and efficient tool for the function-
alization of reductively exfoliated and activated SCAs. The co-
valent aryl binding can be confirmed by SRS and SRM as well
as by TG-MS studies. By the aid of temperature-dependent
Raman investigations it can be shown that the introduced
functional moieties can reversibly be detached in a temperature
region up to 4808C. This information provided the basis for
TG-GC-MS coupling experiments where it was shown that the
thermally detached entities can be identified by MS as one
major component after GC separation. These findings enabled
us to quantify the amount of covalently bound aryl functionali-
ties, allowing us to exactly calculate the percentage of func-
tionalized carbon atoms for each individual component.
covalently bound addend is determined by TG-GC-MS. This is
the advantage of this setup as contributions of side products,
solvent residues, and sample impurities can be identified and
the detected sample mass loss, used for the calculation of the
degree of functionalization, can exactly be assigned to the spe-
cific functional entity of interest.
Apparently, the TG-GC-MS-determined degrees of addition
in the respective SCA derivatives (Figures S24 and S25 in the
Supporting Information) nicely correlate with the development
of the I(D/G) values, discussed above. To demonstrate that the
reductive activation with potassium is indeed crucial for the
covalent functionalization with l3-iodanes, reference experi-
ments with pristine carbon nanotubes and graphite starting
materials have been carried out under various experimental
conditions: 1) stirring at RT (4 h), 2) stirring under light irradia-
tion (4 h), and 3) stirring under heating to reflux (4 h). The re-
spective samples have been investigated by Raman spectros-
copy and TG-MS (Figures S26–S29 in the Supporting Informa-
tion). In contrast to the measured I(D/G) values for the reductive-
ly activated SCAs (0.95 for T(1:4)A, 1.00 for G(1:4)A) only a negligi-
ble increase of the I(D/G) value (Figures S26 and S28 in the
Supporting Information) in relation to the pristine starting allo-
trope was detected for all three reference experiments. These
findings are also corroborated by the corresponding TG-MS
data (Figures S27 and S29 in the Supporting Information).
These results nicely illustrate the important role of reductive-
ly activated SCA intermediates for the covalent functionaliza-
tion of carbon allotropes. Together with a suitable electro-
phile—here l3-iodanes—functional entities can very efficiently
be grafted onto the SCA framework. Dyke et al.[12a] as well as
Chattopadhyay et al.[12b] have shown that these types of reduc-
tive functionalization sequences are based on single-electron
transfer processes from the charged carbon allotrope inter-
mediates towards the trapping electrophile. This leads directly
to the formation of highly reactive radical species in close
proximity to the SCA carbon framework and subsequently, the
Experimental Section
Materials
Purified HiPco SWCNTs (grade: pure; lot number P0261, TGA resi-
due 13.3% wt) were purchased from Unidym Inc. (Sunnyvale, CA)
and were used without further treatment. Synthetic spherical
graphite (SGN18, 99.99% C, TGA residue 0.01% wt—Future
Carbon, Germany) with a mean grain size of 18 mm and a specific
surface area of 6.2 m2 gÀ1 was used after annealing under vacuum
(3008C).
Chemicals and solvents were purchased from Sigma Aldrich Co.
(Germany) and were used as-received if not stated otherwise.
THF was distilled three times under an argon inert gas atmosphere
in order to remove residual water: 1) over CaH2, 2) over sodium,
and 3) over sodium–potassium alloy. Residual traces of oxygen
were removed by pump freeze treatment (three iterative steps).
THFabs was used for all reactions.
Glovebox
Sample functionalization was carried out in an argon-filled Labmas-
ter sp glovebox (MBraun), equipped with a gas filter to remove sol-
vents and an argon cooling systems, with an oxygen content
<0.1 ppm and a water content <0.1 ppm.
Chem. Eur. J. 2014, 20, 16644 – 16651
www.chemeurj.org
16649
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Raman spectroscopy
Synthesis of the l3-iodanes
Raman spectroscopic characterization was carried out on a Horiba-
LabRAM Aramis confocal Raman microscope (lexc =532, 633,
785 nm) with a laser spot size of about 1 mm (Olympus LMPlanFl
100ꢂ, NA 0.80). The incident laser power was kept as low as possi-
ble to avoid structural sample damage: 127 mW (532 nm) for
SWCNT samples or 1.35 mW (532 nm) for graphite samples, 36 mW
(633 nm) and 100 mW (785 nm). Spectra were obtained with a CCD
array at À708C—grating: 600 grooves per mm (532 and 633 nm)
and a 300 groove per mm (785 nm). Spectra were obtained from
a 50 mmꢂ50 mm area for SWCNTs or 100ꢂ100 mm area for graphite
with 2 mm step size in the SWIFT mode for low integration times.
Sample movement was carried out by an automated XY-scanning
table.
The syntheses of the l3-iodanes were adapted from the work of B.
Olofsson and her research group.[13a] m-Chloroperbenzoic acid was
dried under reduced pressure before use. Dichloromethane was
distilled once under reduced pressure before use.
Synthesis of f-SCA
Carbon nanotubide salts with varying potassium/carbon ratios:
T(1:4) (K/C=1:4)], T(1:8) (K/C=1:8), T(1:16) (K/C=1:16), T(1:24) (K/
C=1:24): In an argon-filled glovebox (<0.1 ppm oxygen,
<0.1 ppm H2O), HiPco SWCNTs (12.00 mg, 1.000 mmol) and the re-
spective amount of potassium—T(1:4) (9.775 mg, 0.250 mmol, K/C=
1:4), T(1:8) (4.887 mg, 0.125 mmol, K/C=1:8), T(1:16) (2.444 mg,
0.063 mmol, K/C=1:16), T(1:24) (1.667 mg, 0.042 mmol, K/C=1:24)—
were heated under stirring at 1508C for 8 h. Afterwards, the re-
spective salt was allowed to cool to RT and isolated as a black (K/
C=1:24, 1:16, and 1:8) or beige (K/C=1:4) material.
Temperature-depending Raman measurements were performed in
a Linkam stage THMS 600, equipped with a liquid nitrogen pump
TMS94 for temperature stabilization under a constant flow of nitro-
gen. The measurements were carried out on Si/SiO2 substrates
Graphite intercalation compounds with varying potassium/
carbon ratios: G(1:4) (K/C=1:4), G(1:8) (K/C=1:8), G(1:16) (K/C=1:16),
G(1:24) (K/C=1:24): In an argon-filled glovebox (<0.1 ppm oxygen,
<0.1 ppm H2O), spherical graphite (SGN18) (12.00 mg, 1.000 mmol)
and the respective amount of potassium—G(1:4) (9.775 mg,
0.250 mmol, K/C=1:4), G(1:8) (4.887 mg, 0.125 mmol, K/C=1:8),
G(1:16) (2.444 mg, 0.063 mmol, K/C=1:16), G(1:24) (1.667 mg,
0.042 mmol, K/C=1:24)—were heated under stirring at 1508C for
8 h. Afterwards, the respective salt was allowed to cool to RT and
isolated as a black (K/C=1:24), bronze (K/C=1:16 and 1:8), or
beige (K/C=1:4) material.
SWCNT: l3-iodane variation—preparation of T(1:4)A, T(1:4)B, T(1:4)C:
In an argon-filled glovebox (<0.1 ppm oxygen, <0.1 ppm H2O),
the carbon nanotubide salt T(1:4) (21.7 mg salt, 1 mmol carbon) and
the corresponding l3-iodane—A (293 mg, 0.50 mmol), B (265 mg,
0.50 mmol), or C (238 mg, 0.50 mmol)—were dispersed by the ad-
dition of THFabs (20 mL, three different 50 mL round bottom flasks)
and by the aid of a 5 min tip ultrasonication treatment step (Ban-
delin UW 3200, 600 JminÀ1). The respective dispersion (T(1:4)A,
T(1:4)B, T(1:4)C) was stirred for 14 h. Afterwards, the reaction mixture
was transferred from the glovebox and water (50 mL) and a few
drops of HCl (until pH 4 is reached) were added to the dispersion.
The reaction mixture was transferred to a separation funnel with
cyclohexane (50 mL). The phases were separated and the organic
layer, containing the functionalized SWCNT material, was purged
three times with distilled water. The organic layer was filtered
through a 0.2 mm reinforced cellulose membrane filter (Sartorius)
and washed THF (3ꢂ100 mL). The covalently functionalized
SWCNTs were scraped off the filter paper and the material was
dried in vacuum.
Graphene: l3-iodane variation—preparation of G(1:4)A, G(1:4)B,
G(1:4)C: In an argon-filled glovebox (<0.1 ppm oxygen, <0.1 ppm
H2O), the graphite intercalation compound G(1:4) (21.7 mg, 1 mmol
carbon) and the corresponding l3-iodane—A (293 mg, 0.50 mmol),
B (265 mg, 0.50 mmol), or C (238 mg, 0.50 mmol)—were dispersed
by the addition of THFabs (20 mL, three different 50 mL round
bottom flasks) and by the aid of a 5 min tip ultrasonication treat-
ment step (Bandelin UW 3200, 600 JminÀ1). The respective disper-
sion (G(1:4)A, G(1:4)B, G(1:4)C) was stirred for 14 h. Afterwards, the reac-
tion mixture was transferred from the glovebox and water (50 mL)
and a few drops of HCl (until pH 4 is reached) were added to the
dispersion. The reaction mixture was transferred to a separation
funnel with cyclohexane (50 mL). The phases were separated and
the organic layer, containing the functionalized material, was
purged three times with distilled water. The organic layer was fil-
(300 nm oxide layer) with a heating rate of 10 KminÀ1
.
Thermogravimetric analysis (TG) combined with gas-chro-
matographic separation (GC) coupled with a mass spectrom-
eter (MS)
Thermogravimetric analysis was carried out on a PerkinElmer Pyris
1 TGA instrument. Time-dependent temperature profiles in the
range of 20 and 7008C (20 KminÀ1 gradient) were recorded under
a constant flow of N2 (70 mLminÀ1). About 2.0 mg initial sample
mass were used if not stated otherwise. The evolved gases detach-
ed from the respective sample in combination with the N2 carrier
gas is transferred into the GC system through a TL9000 TG-IR-GC
interface at a constant temperature of 2808C. The gas-chromato-
graphic separation was achieved by a GC-Clarus 680 with a polysil-
oxane-coated Elite-5MS capillary column: 30 m length, 0.25 mm di-
ameter, 0.25 mm film thickness. A GC injection fraction of 1 mL was
collected at the respective TG temperature: SWCNT samples
(2508C), graphite samples (3008C).GC parameters: injector zone=
2808C, detection zone=2508C, split=8.2 mLmin-1, flow rate
helium=10 mLminÀ1, temperature profile=34 min total run time,
dynamic ramp=24 min, 40–2808C with a10 KminÀ1 gradient fol-
lowed by an isothermal step of 10 min at 2808C.
Online MS measurements without GC-separation were carried out
with a GC-Clarus 680 with an Elite-5MS glass capillary column:
30 m length, 0.25 mm diameter. GC parameters: injector zone=
2808C, detection zone=2508C, split=22.6 mLmin-1, flow rate
helium=10 mLminÀ1, isothermal temperature profile=34 min at
2808C. MS measurements were performed on a MS Clarus SQ8C
(multiplier: 1800 V). The obtained data was processed with the Tur-
boMass Software and Bibliograpic searches where performed with
NIST MS Search 2.0.
1H, 13C and 19F NMR spectroscopy
1
NMR spectra were recorded on a Jeol JNM EX 400 (400 MHz for H
and 100 MHz for 13C) and a Bruker Avance 300 (300 MHz for 1H,
75 MHz for 13C and 282 MHz for 19F) spectrometer. Chemical shifts
are reported in [ppm] at room temperature. Abbreviations used for
splitting patterns are: s=singlet, d=doublet, t=triplet, m=mul-
tiplet.
Chem. Eur. J. 2014, 20, 16644 – 16651
www.chemeurj.org
16650
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tered through a 0.2 mm reinforced cellulose membrane filter (Sar-
torius) and washed three times with 100 mL of THF. The covalently
functionalized material was dried in vacuum.
Keywords: analytical chemistry
graphene · hypervalent iodine · nanotubes
·
carbon allotropes
·
SWCNT: variation of the potassium/carbon ratio—preparation of
T(1:4)A, T(1:8)A, T(1:16)A, T(1:24)A: In an argon-filled glovebox
(<0.1 ppm oxygen, <0.1 ppm H2O), the respective carbon nanotu-
bide salt (1 mmol carbon)—T(1:4) (21.7 mg), T(1:8) (16.8 mg), T(1:16)
(14.4 mg), T(1:24) (13.7 mg)—and A (293 mg, 0.50 mmol) were dis-
persed by the addition of THFabs (20 mL, four different 50 mL
round bottom flasks) and by the aid of a 5 min tip ultrasonication
treatment step (Bandelin UW 3200, 600 JminÀ1). The respective dis-
persion (T(1:4)A, T(1:8)A, T(1:16)A, T(1:24)A) was stirred for 14 h. After-
wards, the reaction mixture was transferred from the glovebox and
water (50 mL) and a few drops of HCl (until pH 4 is reached) were
added to the dispersion. The reaction mixture was transferred to
a separation funnel with cyclohexane (50 mL). The phases were
separated and the organic layer, containing the functionalized
SWCNT material, was purged three times with distilled water. The
organic layer was filtered through a 0.2 mm reinforced cellulose
membrane filter (Sartorius) and washed three times with 100 mL of
THF. The covalently functionalized SWCNTs were scraped off the
filter paper and the material was dried in vacuum.
[1] a) S. Eigler, A. Hirsch, Angew. Chem. 2014, 126, 7852–7872; Angew.
Chem. Int. Ed. 2014, 53, 7720–7738; b) C. K. Chua, M. Pumera, Chem.
Soc. Rev. 2013, 42, 3222–3233; c) N. Karousis, N. Tagmatarchis, D. Tasis,
Chem. Rev. 2010, 110, 5366–5397; d) A. Hirsch, J. M. Englert, F. Hauke,
Acc. Chem. Res. 2013, 46, 87–96; e) A. Hirsch, Nat. Mater. 2010, 9, 868–
871.
[2] F. Liang, L. B. Alemany, J. M. Beach, W. E. Billups, J. Am. Chem. Soc. 2005,
127, 13941–13948.
[3] a) G. L. C. Paulus, Q. H. Wang, M. S. Strano, Acc. Chem. Res. 2013, 46,
160–170; b) S. A. Hodge, M. K. Bayazit, K. S. Coleman, M. S. P. Shaffer,
Chem. Soc. Rev. 2012, 41, 4409–4429; c) M. Quintana, E. Vazquez, M.
Prato, Acc. Chem. Res. 2013, 46, 138–148; d) E. Bekyarova, S. Sarkar, F.
Wang, M. E. Itkis, I. Kalinina, X. Tian, R. C. Haddon, Acc. Chem. Res. 2013,
46, 65–76; e) C. Backes, J. M. Englert, N. Bernhard, F. Hauke, A. Hirsch,
Small 2010, 6, 1968–1973; f) F. Hof, S. Bosch, J. M. Englert, F. Hauke, A.
Hirsch, Angew. Chem. 2012, 124, 11897–11900; Angew. Chem. Int. Ed.
2012, 51, 11727–11730; g) J. M. Englert, P. Vecera, K. C. Knirsch, R. A.
Schꢀfer, F. Hauke, A. Hirsch, ACS Nano 2013, 7, 5472–5482.
[4] Z. Syrgiannis, V. La Parola, C. Hadad, M. Lucꢃo, E. Vꢄzquez, F. Giacalone,
M. Prato, Angew. Chem. 2013, 125, 6608–6611; Angew. Chem. Int. Ed.
2013, 52, 6480–6483.
Graphene: variation of the potassium/carbon ratio—preparation
of G(1:4)A, G(1:8)A, G(1:16)A, G(1:24)A: In an argon-filled glovebox
(<0.1 ppm oxygen, <0.1 ppm H2O), the respective graphite inter-
calation compound (1 mmol carbon)—G(1:4) (21.7 mg), G(1:8)
[5] S. Fujii, T. Enoki, ACS Nano 2013, 7, 11190–11199.
[6] S. Kundu, W. Xia, W. Busser, M. Becker, D. A. Schmidt, M. Havenith, M.
Muhler, Phys. Chem. Chem. Phys. 2010, 12, 4351–4359.
[7] a) A. Pꢅnicaud, P. Poulin, A. Derrꢅ, E. Anglaret, P. Petit, J. Am. Chem. Soc.
2004, 127, 8–9; b) C. Vallꢅs, C. Drummond, H. Saadaoui, C. A. Furtado,
M. He, O. Roubeau, L. Ortolani, M. Monthioux, A. Pꢅnicaud, J. Am. Chem.
Soc. 2008, 130, 15802–15804.
[8] F. Hof, S. Bosch, S. Eigler, F. Hauke, A. Hirsch, J. Am. Chem. Soc. 2013,
135, 18385–18395.
[9] a) E. A. Merritt, B. Olofsson, Angew. Chem. 2009, 121, 9214–9234;
Angew. Chem. Int. Ed. 2009, 48, 9052–9070; b) J. Malmgren, S. Santoro,
N. Jalalian, F. Himo, B. Olofsson, Chem. Eur. J. 2013, 19, 10334–10342.
[10] C. K. Chan, T. E. Beechem, T. Ohta, M. T. Brumbach, D. R. Wheeler, K. J.
Stevenson, J. Phys. Chem. C 2013, 117, 12038–12044.
[11] a) C. Fontanesi, C. A. Bortolotti, D. Vanossi, M. Marcaccio, J. Phys. Chem.
A 2011, 115, 11715–11722; b) K. H. Vase, A. H. Holm, K. Norrman, S. U.
Pedersen, K. Daasbjerg, Langmuir 2007, 23, 3786–3793; c) K. H. Vase,
A. H. Holm, S. U. Pedersen, K. Daasbjerg, Langmuir 2005, 21, 8085–
8089.
(16.8 mg), G(1:16) (14.4 mg), G(1:24) (13.7 mg)—and
A (293 mg,
0.50 mmol) were dispersed by the addition of THFabs (20 mL, four
different 50 mL round bottom flasks) and by the aid of a 5 min tip
ultrasonication treatment step (Bandelin UW 3200, 600 JminÀ1).
The respective dispersion (G(1:4)A, G(1:8)A, G(1:16)A, G(1:24)A) was stirred
for 14 h. Afterwards, the samples are removed out of the glovebox
and water (50 mL) and a few drops of HCl (until pH 4 is reached)
were added to the dispersions. The reaction mixture was trans-
ferred to a separation funnel with cyclohexane (50 mL). The phases
were separated and the organic layer, containing the functional-
ized material, was purged three times with distilled water. The or-
ganic layer was filtered through a 0.2 mm reinforced cellulose
membrane filter (Sartorius) and washed three times with THF
(100 mL). The covalently functionalized material was dried in
a vacuum.
[12] a) C. A. Dyke, J. M. Tour, Nano Lett. 2003, 3, 1215–1218; b) J. Chattopad-
hyay, S. Chakraborty, A. Mukherjee, R. Wang, P. S. Engel, W. E. Billups, J.
Phys. Chem. C 2007, 111, 17928–17932.
[13] a) M. Bielawski, B. Olofsson, Chem. Commun. 2007, 2521–2523; b) M.
Bielawski, B. Olofsson in Efficient One-Pot Synthesis of Bis(4-tert-Butyl-
phenyl) Iodonium Triflate, Wiley, New York, 2009.
Acknowledgements
[14] The reductive exfoliation sequences does not exclusively yield monolay-
er material, which is discussed in the course of the manuscript.
[15] a) F. Liang, A. K. Sadana, A. Peera, J. Chattopadhyay, Z. Gu, R. H. Hauge,
W. E. Billups, Nano Lett. 2004, 4, 1257–1260; b) J. Chattopadhyay, A. K.
Sadana, F. Liang, J. M. Beach, Y. Xiao, R. H. Hauge, W. E. Billups, Org. Lett.
2005, 7, 4067–4069.
The authors thank the Deutsche Forschungsgemeinschaft (DFG
- SFB 953, Project A1 “Synthetic Carbon Allotropes”) and the
European Research Council (ERC; grant agreement no.246622
GRAPHENOCHEM) for financial support. The research leading
to these results has received funding from the European Union
Seventh Framework Programme under grant agreement
no.604391 Graphene Flagship.
Received: July 31, 2014
Published online on October 24, 2014
Chem. Eur. J. 2014, 20, 16644 – 16651
www.chemeurj.org
16651
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim