Value-added chemicals from biomass-derived furans: radical functionalisations of 5-chloromethylfurfural (CMF) by metal-free ATRA reactions

Article abstract of DOI:10.1039/d1ob00013f

Biomass-derived 5-chloromethylfurfural (CMF), a congener of the well-known carbohydrate-based platform chemical 5-hydroxymethylfurfural (HMF), can efficiently be functionalised by radical transformations of its benzylic chloromethyl group. We report here the first examples of these radical reactions by way of metal-free, triethylborane/oxygen-induced atom transfer radical addition (ATRA) reactions between CMF and styrenes, which proceed with high yield and selectivity. The key intermediate, the 2-formyl-5-furfuryl radical derived from CMF, and its radical addition reactions were studied with regard to its electronic structure,i.e.spin density distribution and frontier molecular orbitals based on the NBO ansatz and activation barriers of the addition step using DFT and post-HF methods.

Full text of DOI:10.1039/d1ob00013f

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Value-added chemicals from biomass-derived
furans: radical functionalisations of
5-chloromethylfurfural (CMF) by metal-free ATRA
reactions†
Cite this: Org. Biomol. Chem., 2021,
19, 1626
Rajesh Dasi, a Daniel Schmidhuber,b Lisa Marie Gronbach,a Julia Rehbein *b and
a
Malte Brasholz
*
Biomass-derived 5-chloromethylfurfural (CMF), a congener of the well-known carbohydrate-based plat-
form chemical 5-hydroxymethylfurfural (HMF), can efficiently be functionalised by radical transformations
of its benzylic chloromethyl group. We report here the first examples of these radical reactions by way of
metal-free, triethylborane/oxygen-induced atom transfer radical addition (ATRA) reactions between CMF
and styrenes, which proceed with high yield and selectivity. The key intermediate, the 2-formyl-5-furfuryl
radical derived from CMF, and its radical addition reactions were studied with regard to its electronic
structure, i.e. spin density distribution and frontier molecular orbitals based on the NBO ansatz and acti-
vation barriers of the addition step using DFT and post-HF methods.
Received 4th January 2021,
Accepted 28th January 2021
DOI: 10.1039/d1ob00013f
rsc.li/obc
Introduction
Decades ago, 5-hydroxymethylfurfural (HMF, 1) was identified
as an industrially viable renewable platform chemical for the
production of fuel additives, polymers, C-6 diacids and diols,
and numerous other strategically relevant high-volume chemi-
cal products. While HMF is available from the Lewis- or
Brønsted-acid-mediated dehydration of several mono- and dis-
accharides, its efficient retrieval from cellulosic biomass is still
challenging today as the attempt often results in low yields
and selectivities.1 By contrast to HMF, its chlorinated and con-
sequently much more lipophilic congener 5-chloromethyl-
furfural (CMF, 2) can be obtained in a high yield directly from
crude biomass, by its simple treatment with aqueous hydro-
chloric acid in combination with biphasic extraction tech-
niques (Scheme 1a).2
The derivative chemistry of HMF (1) and CMF (2) can be
subdivided into a furanic and a levulinic manifold, and both
have successfully been converted on a large scale into impor-
tant bulk chemicals such as furan-2,5-dicarboxylic acid, 2,5-
dimethyl-furan, levulinic acid, adipic acid and others, by con-
ventional stoichiometric or more elaborate catalytic oxidation,
aUniversity of Rostock, Institute of Chemistry, Albert-Einstein-Str. 3A, 18055 Rostock,
Germany. E-mail: malte.brasholz@uni-rostock.de
bUniversity of Regensburg, Institute of Organic Chemistry, Universitätsstr. 31, 93053
Regensburg, Germany. E-mail: julia.rehbein@chemie.uni-regensburg.de
†Electronic supplementary information (ESI) available: Full experimental details
and compound characterization data. See DOI: 10.1039/d1ob00013f
Scheme 1 (a) Furans HMF (1) and CMF (2) and their retrieval from
biomass. (b) Industrial-scale conversion of 1 and 2 into value-added
downstream bulk intermediates. (c) The idea of radical functionalizations
of CMF (2) via the “wild type” (2-formyl-5-furanyl)methyl radical 3.
1626 | Org. Biomol. Chem., 2021, 19, 1626–1631
This journal is © The Royal Society of Chemistry 2021

View Article Online
Organic & Biomolecular Chemistry
Paper
reduction and hydrolytic protocols (Scheme 1b).3 On the other and to correlate it with its behavior in addition reactions with
hand, to the best of our knowledge, the radical chemistry of alkenes.
HMF (1) or CMF (2) has not been explored thus far, despite
The spin density distribution in radical 3 was compared to
CMF (2), with its benzylic chloromethyl group, obviously those of the closely related 2-furylmethyl radical 4, 4-formylto-
offering ample opportunities for transformations in the lyl radical 5 and tolyl radical 6 (NBO, uMP2/6-311+G**). In all
radical manifold (Scheme 1c).
radicals, the highest spin density is in fact located at the
We report here the first radical functionalisations of CMF benzylic methylene carbon, but in radical 3, to a considerable
(2) by way of highly efficient atom transfer radical addition extent also at C-2 (Scheme 2b). The SOMO energies of species
(ATRA)4 reactions with styrenes. Furthermore, we studied the 3–6 were also calculated (NBO6, uMP2/6-311+G**). As shown
properties and radical addition reactions of the key “wild-type” in Table 1 (top row), the SOMO energies ε of the radicals can
(2-formyl-5-furanyl)methyl radical 3 derived from CMF (2) by be ranked as follows: 5 < 3–6 < 4, ranging from −9.168 eV to
NBO population analysis in combination with post-HF −8.167 eV. A correlation of the SOMO energies of radicals 3–6
methods (MP2) and density functional theory (DFT) with HOMO and LUMO energies of diverse alkenes as poten-
calculations.
tial reaction partners showed that, in particular, styrene deriva-
tives 7 possess HOMO energies close to the SOMO energies of
3–6 and therefore should be highly suitable radical acceptors.
Based on the observed differences in Δε (FMO), one may (a)
expect slight modulation in reactivity and (b) a change in the
radicals’ nucleophilic or electrophilic role in the addition reac-
tion depending on the electron density of the styrene. As
shown in Table 1, the combination of (2-formyl-5-furanyl)
methyl radical 3 with electron-poor styrenes (R = CN, F) is
expected to have the fastest reaction based on the low Δε
(FMO) of 0.169 eV. For the much alike 4-formyltolyl radical 5,
the smallest energy gap is again that with the HOMO of the
acceptor-substituted 4-cyanostyrene (Δε 0.171 eV), while the
largest value for each was obtained for donor-substituted
4-methoxystyrene (Δε = 0.894 and 1.234 eV). The correlation of
Δε (FMO) with activation barriers was later checked by calcu-
lating the actual addition barrier.
Results and discussion
During the reaction design, we anticipated that several factors
would potentially impede the desired radical functionalisation
of CMF (2). Irrespective of the chosen induction method, e.g.,
a photoredox-catalysed reductive cleavage of the C–Cl bond or
direct halogen atom transfer (XAT), complications could arise
from the degree of spin delocalization in (2-formyl-5-furanyl)
methyl radical 3. In addition, 3 being a benzylic-like radical, a
comparably slow rate of addition to alkenes would be
expected.5 Finally, 2-furylmethyl radicals are known to
undergo ring-opening, leading to multiple secondary fragmen-
tation products (Scheme 2a).6 Therefore, we initiated a compu-
tational study to elucidate the electronic structure of radical 3
In our synthetic experiments, we initially aimed at develop-
ing photoredox-induced atom transfer radical addition (ATRA)
reactions7 between CMF (2) and styrene (7a). For this purpose,
CMF (2) was first characterized by cyclic voltammetry, and a
range of potential photocatalysts were subsequently evaluated
in the reaction, both under oxidative and reductive quenching
conditions. However, these experiments were only of limited
success. After extensive experimentation, we found that only
+
when using Ir(dtbbpy)(ppy)2 as the photocatalyst, i-Pr2NEt as
the sacrificial electron donor, in combination with tetrabutyl-
ammonium iodide (TBAI), to convert CMF (2) in situ into the
more easily reduced benzylic iodide, the ATRA product 8a was
Table 1 SOMO energies of radicals 3–6, HOMO energies of 4-substi-
tuted styrenes 7 and energy gaps Δε (in eV) calculated using NBOs at the
uMP2/6-311+G** level of theory
3
4
5
6
Radicals
SOMO (eV)
HOMO (eV)
–8.828
–8.167
–9.168
–8.771
Δε (eV)
Δε (eV)
Δε (eV)
Δε (eV)
Scheme 2 (a) Methods for the generation of radical 3 and subsequent
ring opening. (b) Comparison of spin densities (based on NBO analysis,
uMP2/6-311+G**, the plot at an isovalue of 0.04) of radical 3 with those
of the related radicals 4, 5 and 6. (c) Spin density plot of radical 3: blue
shades designate high spin density.
R = H
–8.399
–8.136
–7.934
–8.550
–8.997
0.429
0.692
0.894
0.278
0.169
−0.232
0.031
0.233
−0.383
−0.83
0.769
1.032
1.234
0.618
0.171
0.372
0.635
0.837
0.221
−0.226
R = Me
R = OMe
R = F
R = CN
This journal is © The Royal Society of Chemistry 2021
Org. Biomol. Chem., 2021, 19, 1626–1631 | 1627

View Article Online
Paper
Organic & Biomolecular Chemistry
produced in 32% yield, and this result could not be further m-methoxy derivative 8k was accessible in 73% yield, the
improved. o-methoxy and p-ethoxy derivatives 8l and 8m were obtained
As an alternative to photoredox-induction of the radical exclusively as eliminated products, in moderate yields of 41%
reaction, CMF (2) could also be activated by direct halogen and 30%, respectively.
atom abstraction, and as we attempted triethylborane/O2 as a
Small changes in the structure of the alkene partner exerted
reagent system,8 we could indeed realise the desired ATRA pronounced effects on the reaction efficiency. Utilizing
transformations in synthetically useful yields. The Et3B/O2- α-methylstyrenes 10a–d, the conversion of CMF (2) remained
induced reaction was carefully optimized as shown in Table 2 partially incomplete even after 24 h of reaction time, and iso-
(see also Table 1 of the ESI†). When CMF (2) and styrene (7a) meric C2–C3 and C3–C4 alkenes 11 were isolated in moderate
were treated with 2 equivalents of Et3B in MeCN, under air and yields of 25–33% (Scheme 3b). Consistently, when employing
at 60 °C, the conversion of 2 was 29% within 24 hours, yet, a 1,1-diphenylethylene (12), trisubstituted alkene 13 was
mixture of undefined products was formed (entry 1). However, obtained in 29% yield. A number of alternative electron-rich
upon addition of 1 equivalent of TBAI, the conversion and electron-deficient radical acceptors were evaluated
improved to 87%, and 23% of secondary benzylic chloride 8a (Scheme 3c); however, they were unsuitable reaction partners
was obtained, along with 10% of furanic dimer 9, as deter- for (2-formyl-5-furanyl)methyl radical 3, which mostly under-
mined by quantitative 1H-NMR analysis (entry 2). went unimolecular decomposition in their presence.
Subsequently, the influence of the molar equivalents of TBAI,
The experimental results in Scheme 3 reflect the influence
Et3B and styrene (7a), the oxygen concentration and the of the SOMO/HOMO energy gaps of the reaction partners.
solvent was systematically studied (entries 3–16). Ultimately, While compound 8j from the reaction between CMF (2) and
using 0.5 equiv. of Et3B, 4 equiv. of TBAI and a 10-fold excess 4-cyanostyrene was isolated in 59% yield, the eliminated
of styrene (7a) in acetone, under air and at 60 °C for 24 h, the 1-methoxy and 4-ethoxy derivatives 8l and 8m were obtained in
secondary benzylic chloride 8a was provided in an excellent significantly reduced yields of 41% and 30%, respectively,
78% isolated yield after chromatography, and the undesired which correlates well with the calculated respective FMO
dimer 9 was formed in only 3% yield. Shortening of the reac- energy gaps Δε (Table 1). A DFT calculation (uM06-2X/6-
tion time to 12 h led to largely incomplete conversion of CMF 311+G** basis set) of the energy profiles for the addition reac-
(2). Notably, in none of the reactions could any trace of the sec- tions of (2-formyl-5-furanyl)methyl radical 3 with 4-cyano- and
ondary benzylic iodide corresponding to 8a be detected.
4-methoxystyrene revealed that in both cases, the radical
Under the optimized conditions, a series of ATRA products addition occurs with equally very low energy barriers of
8a–8k could readily be synthesized (Scheme 3a). The yields around 11 kcal mol−1, the energy barrier of the latter reaction
ranged from 40 to 84%, while the 2,6-dichlorophenyl derivative being marginally higher (Scheme 4). Both reactions are
8h formed in just 35% yield, probably as a result of steric hin- exothermic, while the adduct radical 14 from 4-cyanostyrene is
drance of the double bond in parent styrene 7h. While the ca. 2 kcal mol−1 lower in energy than the adduct formed by the
Table 2 Reaction optimization (summary)
Entry
Et3B (eq.)
TBAI (eq.)
7a (eq.)
Solvent
Conv. 2 a [%]
Yield 8a a [%]
Yield 9 a [%]
1
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
0.25
0.5
1.0
0.5
0.5
0.5
0
1
1
0.5
1
2
4
4
4
4
4
4
4
4
4
4
15
15
15
8
8
8
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
THF
29
87
0
0
23
0
0
10
0
2
2
3
3
2
3
2
3
3
5
2
2
3
2
3b
4
100
100
100
100
95
100
100
91
96
98
74
63
16
30
31
40
30
42
30
38
74
51
50
28
75 (78)c
5
6
7
8
8
4
9
10
20
10
10
10
10
10
10
10
11
12
13
14
15
16
DMF
Acetone
97
Reactions were performed on a 0.30 mmol scale of 2. a Determined by 1H-NMR spectroscopy against CH2Br2 as an internal standard. b Reaction
in the absence of O2. c Yield of the isolated product after chromatography.
1628 | Org. Biomol. Chem., 2021, 19, 1626–1631
This journal is © The Royal Society of Chemistry 2021

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 4 Calculated energy profiles and the transition state structure
for radical 3 and 4-methoxystyrene.
Scheme 3 Reactions were performed on a 0.30 mmol scale of 2, under
the conditions of entry 16 in Table 2.
Scheme 5 Proposed reaction mechanism.
reaction with 4-methoxystyrene. The fully optimized transition
state structures show that in both cases, the aromatic rings of give the primary ATRA products, reactive secondary benzylic
the reactants converge in a stacked orientation, and the length iodides 16. Nucleophilic substitution of 16 with the chloride
of the formed bond is very similar for R = CN (2.23 Å) and anion finally delivers the secondary benzylic chlorides 8.
OMe of 2.22 Å (also see the ESI†).
The overall ATRA process occurs efficiently only in air; thus,
The proposed mechanism for the Et3B/O2-induced ATRA when ample O2 is available to continuously supply ethyl rad-
reactions is shown in Scheme 5. Aerobic autoxidation of Et3B icals for initiation and to sustain the short-lived radical chain,
rapidly liberates reactive ethyl radicals9 to abstract atomic 14• + 15 → 3• + 16.10 In the absence of iodide anions, direct
iodine from iodomethylfurfural (IMF, 15), which forms in situ chlorine atom abstraction from CMF (2) is way less efficient (k1
by the reaction between CMF (2) and the iodide anion. Radical ≪ k3). However, once IMF (15) is generated in situ, atomic
addition of the so-produced (2-formyl-5-furanyl)methyl radical iodine is easily abstracted. The radical addition of benzylic-like
3 with styrenes 7 leads to the more stabilized radical adducts radical 3 with styrenes is expected to be slow,5b yet it is fol-
14, which react with IMF (15) in a chain propagation step, to lowed by a fast propagation step (kprop > kadd). The formation
This journal is © The Royal Society of Chemistry 2021
Org. Biomol. Chem., 2021, 19, 1626–1631 | 1629

View Article Online
Paper
Organic & Biomolecular Chemistry
of chloro compounds 8 by chain propagation of adduct rad-
icals 14 with CMF (2) is energetically much less likely.
Acknowledgements
Furthermore, a possible polymerization of the intermediate The authors acknowledge the financial support of this project
radical adducts 14 with excess styrene molecules (atom trans- by the strategic network funding programme of the Leibniz
fer radical polymerization, ATRP)11 is apparently of minor sig- Association,
within
the
project
“Leibniz-
nificance (kpol ≪ kprop) based on the observed reaction yields WissenschaftsCampus–ComBioCat–Rostock”. R. D. thanks this
of ATRA products 8, which are in the range of up to ca. 85%. programme for a Ph.D. fellowship.
Notably, both key radical intermediates 3 and 14 could suc-
cessfully be trapped by TEMPO, and their adducts 17 and 18
were unambiguously identified by ESI-TOF mass spectrometry
(see the ESI†). Moreover, when CMF (2) alone was reacted with
Notes and references
TEMPO and Et3B/O2, adduct 17 was isolated in 44% yield and
1 Selected reviews: (a) C. Thoma, J. Konnerth, W. Sailer-
could be fully characterized.
Kronlachner, P. Solt, T. Rosenau and H. W. G. van
Herwijnen,
ChemSusChem,
2020,
13,
3544–3564;
(b) R.-J. van Putten, J. C. van der Waal, E. de Jong,
C. B. Rasrendra, H. J. Heeres and J. G. de Vries, Chem. Rev.,
2013, 113, 1499–1597.
Conclusions
The derivative chemistry of the biomass-derived furan
5-chloromethylfurfural (2) has successfully been extended into
the radical manifold. As a first demonstration, we developed
efficient ATRA reactions between CMF (2) and styrenes, driven
by the metal-free and environmentally benign initiation
system of Et3B and O2. The key “wild type” 2-formyl-5-furfuryl
radical 3 and its behavior in radical addition reactions were
studied by computational chemistry approaches, and the cal-
culated molecular and reaction parameters correlate well with
the experimental outcomes of our synthetic experiments.
Future approaches to utilize the radical chemistry of the
furanic platform chemicals HMF (1) and CMF (2) are currently
under development in our laboratories.
2 (a) M. Mascal and E. B. Nikitin, ChemSusChem, 2009, 2, 423–
426; (b) M. Brasholz, K. von Känel, C. H. Hornung,
S. Saubern and J. Tsanaktsidis, Green Chem., 2011, 13, 1114–
1117; (c) M. Mascal, ChemSusChem, 2015, 8, 3391–3395.
3 Recent overviews: (a) Q.-S. Kong, X.-L. Li, H.-J. Xu and
Y. Fu, Fuel Process. Technol., 2020, 209, DOI: 10.1016/j.
fuproc.2020.106528; (b) D. Zhao, T. Su, Y. Wang,
R. S. Varma and C. Len, Mol. Catal., 2020, 495, DOI:
10.1016/j.mcat.2020.111133; (c) C. Thoma, J. Konnerth,
W. Sailer-Kronlachner, T. Rosenau, A. Potthast, P. Solt and
H. W. G. van Herwijnen, ChemSusChem, 2020, 13, 5408–
5422; (d) B. Wozniak, S. Tin and J. G. de Vries, Chem. Sci.,
2019, 10, 6024–6034; (e) J. G. de Vries, Adv. Heterocycl.
Chem., 2017, 121, 247–293; (f) M. Mascal and S. Dutta, Top.
Curr. Chem., 2014, 353, 41–83.
4 Reviews: (a) J. M. Muçoz-Molina and T. R. Belderrain, in
Science of Synthesis, C-1 Building Blocks in Organic Synthesis
2, ed. P. W. N. M. Van Leeuwen, Thieme, Stuttgart, 2014,
pp. 459–473; (b) T. Pintauer, Eur. J. Inorg. Chem., 2010, 17,
2449–2460; (c) T. Pintauer and K. Matyjaszewski, Chem. Soc.
Rev., 2008, 37, 1087–1097; (d) K. Severin, Curr. Org. Synth.,
2006, 10, 217–224.
5 Review of absolute rate constants of free-radical additions to
alkenes: (a) H. Fischer and L. Radom, Angew. Chem., Int. Ed.,
2001, 40, 1340–1371. For rate constant data of the less-reac-
tive benzylic radicals, see: (b) K. Héberger, M. Walbiner and
H. Fischer, Angew. Chem., Int. Ed. Engl., 1992, 31, 635–636.
6 (a) Y. Li, Y. Gan and C. Cao, J. Comput. Chem., 2019, 40,
1057–1065; (b) C. Togbé, L.-S. Tran, D. Liu, D. Felsmann,
P. Oßwald, P.-A. Glaude, B. Sirjean, R. Fournet, F. Battin-
Leclerc and K. Kohse-Höinghaus, Combust. Flame, 2014,
161, 780–797; (c) L. D. Kispert, R. C. Quijano and
C. U. Pittman Jr., J. Org. Chem., 1971, 36, 3837–3838.
7 Reviews: (a) D. Bag, H. Kour and S. D. Sawant, Org. Biomol.
Chem., 2020, 18, 8278–8293; (b) T. M. Williams and
C. R. J. Stephenson, in Visible Light Photocatalysis in
Organic Chemistry, ed. C. R. J. Stephenson, D. W. C.
MacMillan and T. P. Yoon, Wiley-VCH, Weinheim, 2018;
pp. 73–92; (c) T. Corant and G. Masson, J. Org. Chem., 2016,
Experimental
General procedure for the synthesis of compounds 8
5-Chloromethylfurfural 2 (0.30 mmol, 1.0 equiv.) and tetra-
butylammonium iodide (TBAI, 1.20 mmol, 4.0 equiv.) were
combined in a 10 mL crimp cap vial and the vial was sealed.
Dry acetone (2.00 mL) and styrene derivative 7 (3.00 mmol, 10
equiv., freshly distilled) were added and the mixture was
heated to 60 °C. After 30 min, triethyl borane (Et3B, 1.0 M solu-
tion in hexanes, 0.15 mmol, 0.5 equiv.) was added. Then, the
septum was pierced with a cannula to allow the capture of
oxygen from the ambient air. After 24 h of reaction time, the
solvent was evaporated. The crude mixture was diluted with
DCM and washed with water and brine (3 × 20 mL). The
organic layer was collected and dried over anhydrous MgSO4
and filtered, followed by concentration under reduced
pressure. The crude mixture was subjected to column chrom-
atography (silica gel, 230–400 mesh) to afford the desired com-
pounds 8.
Conflicts of interest
There are no conflicts to declare.
1630 | Org. Biomol. Chem., 2021, 19, 1626–1631
This journal is © The Royal Society of Chemistry 2021

View Article Online
Organic & Biomolecular Chemistry
Paper
81, 6945–6952; (d) O. Reiser, Acc. Chem. Res., 2016, 49,
1990–1996.
8 Selected overviews: (a) P. Renaud, in Encyclopedia of
Radicals in Chemistry, Biology and Materials, ed. C.
9 For a recent study of the mechanism of Et3B autoxidation,
see: R. Uematsu, C. Saka, Y. Sumiya, T. Ichino,
T. Taketsugu and S. Maeda, Chem. Commun., 2017, 53,
7302–7305.
Chatgilialoglu and A. Studer, Wiley, New York, 2012, Vol. 1, 10 For a detailed study on the effect of the O2 concentration in
pp. 601–628; (b) V. Darmency and P. Renaud, Top. Curr.
Chem., 2006, 263, 71–106; (c) H. Yorimitsu and K. Oshima,
Et3B-initiated radical reactions, see: D. P. Curran and
T. R. McFadden, J. Am. Chem. Soc., 2016, 138, 7741–7752.
in Radicals in Organic Synthesis, ed. P. Renaud and 11 K. Matyjaszewski and J. Xia, Chem. Rev., 2001, 101, 2921–
M. P. Sibi, Wiley-VCH, Weinheim, 2001, vol. 1, pp. 11–27.
2990.
This journal is © The Royal Society of Chemistry 2021
Org. Biomol. Chem., 2021, 19, 1626–1631 | 1631