Electrochemical synthesis of biobased polymers and polymer building blocks from vanillin

Article abstract of DOI:10.1039/d1ra00649e

Vanillin, one of the few biobased aromatic compounds available on an industrial level, is an attractive candidate for the synthesis of biobased polymers and polymer building blocks. This study presents a detailed investigation of the reductive electrochemical coupling process by pinacolization of vanillin and divanillin in an electrochemical H-type cell setup to the polymer building block hydrovanilloin and to polyvanillin, respectively. Therein, different cathode materials are screened by linear sweep voltammetry for their capability and activity of hydrodimerization of phenolic aromatic aldehydes in alkaline aqueous media. Product distributions and faradaic efficiencies of the electrochemical vanillin reduction are investigated in bulk electrolysis experiments. Dependencies on electrochemical parameters such as current densities, applied charges and cathode materials are studied. Furthermore, the polyvanillin synthesis from divanillin is also investigated by bulk electrolysis experiments. The effects of selected electrochemical parameters (current density, applied charge and electrode material) on yield and structural features (weight-average molecular weight (M_W), number-average molecular weight (M_N), polydispersity (M_W/M_N)) measured by size exclusion chromatography of the obtained polyvanillin were evaluated. Structural features of isolated polyvanillin were determined by 2D-NMR (HSQC,¹³C/¹H) analyses and by³¹P-NMR analyses afterin situlabeling with Cl-TMDP and possible pathways for their generation are discussed. These two promising electro-synthetic processes studied are free of hazardous materials and reagents and highlight the contributions of preparative electrochemistry to green chemistry and further pave the way toward the application of electrochemistry in the synthesis of biobased building blocks and polymers.

Full text of DOI:10.1039/d1ra00649e

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Electrochemical synthesis of biobased polymers
and polymer building blocks from vanillin†
Cite this: RSC Adv., 2021, 11, 8970
a
Volkmar M. Schmidt,b Carsten Cremers,a Dominik Muller,a
*
Robin Kunkel,
¨
Detlef Schmiedlc and Jens Tubkea
¨
Vanillin, one of the few biobased aromatic compounds available on an industrial level, is an attractive
candidate for the synthesis of biobased polymers and polymer building blocks. This study presents
a detailed investigation of the reductive electrochemical coupling process by pinacolization of vanillin
and divanillin in an electrochemical H-type cell setup to the polymer building block hydrovanilloin and to
polyvanillin, respectively. Therein, different cathode materials are screened by linear sweep voltammetry
for their capability and activity of hydrodimerization of phenolic aromatic aldehydes in alkaline aqueous
media. Product distributions and faradaic efficiencies of the electrochemical vanillin reduction are
investigated in bulk electrolysis experiments. Dependencies on electrochemical parameters such as
current densities, applied charges and cathode materials are studied. Furthermore, the polyvanillin
synthesis from divanillin is also investigated by bulk electrolysis experiments. The effects of selected
electrochemical parameters (current density, applied charge and electrode material) on yield and
structural features (weight-average molecular weight (MW), number-average molecular weight (MN),
polydispersity (MW/MN)) measured by size exclusion chromatography of the obtained polyvanillin were
evaluated. Structural features of isolated polyvanillin were determined by 2D-NMR (HSQC, 13C/1H)
analyses and by 31P-NMR analyses after in situ labeling with Cl-TMDP and possible pathways for their
generation are discussed. These two promising electro-synthetic processes studied are free of
hazardous materials and reagents and highlight the contributions of preparative electrochemistry to
green chemistry and further pave the way toward the application of electrochemistry in the synthesis of
biobased building blocks and polymers.
Received 25th January 2021
Accepted 20th February 2021
DOI: 10.1039/d1ra00649e
rsc.li/rsc-advances
production of the ne chemical vanillin, which is generated on
an industrial scale with appropriate yields up to 7% and high
1 Introduction
selectivity via classical thermal oxidation by oxygen at copper
catalysts8 or more recently by sustainable electrochemical
oxidation.9–11 Due to vanillin's aromatic character, multi-
functionality and abundance, many attempts have been made
to utilize vanillin. For example, over the last years vanillin has
been intensively studied as key-intermediate in high perfor-
mance biobased polymer synthesis, which was reviewed in 2015
by Fache et al.12 Vanillin has been exploited as a building block
in, e.g. phenolic polymers, epoxy and benzoxazine resins, poly-
esters as well as acrylate and methacrylate polymers.
Electrochemical conversion of vanillin would be an attractive
route enabling an overall green process, since electrochemistry
is intrinsically fullling several important criteria of green
chemistry, such as prevention of waste and use of toxic or
hazardous reagents as only electrons serve as reactants. More-
over, this leads to a better atom efficiency and an inherently safe
process. Also, electrochemical conversions oen can be con-
ducted under mild conditions, whereas for comparable
conventional syntheses high temperatures and pressures are
needed.13,14 Despite the mentioned advantages electrochemistry
The development of sustainable processes exploiting renewable
resources for the production of new green monomers, drop-in
chemicals replacing petrochemicals, and biobased polymers is
of urgent matter.1–4 Therefore, several approaches have been
recently published valorizing lignin, which is obtained on
a million-ton scale from lignocellulose by either the Kra
process, the sulte process or the organosolv process in the
paper and pulp industry. Currently lignin is mostly used ener-
getically to provide process heat and for the recovery of the
pulping chemicals being the least utilized lignocellulosic
biopolymer.5–7 One promising valorization route is the
aFraunhofer Institute for Chemical Technology ICT, Department of Applied
Electrochemistry, Joseph-von-Fraunhofer-Str. 7, D-76327 Pnztal, Germany. E-mail:
robin.kunkel@ict.fraunhofer.de
bMannheim University of Applied Sciences, Institute of Chemical Process Engineering,
Paul-Wittsack-Str. 10, D-68163 Mannheim, Germany
cFraunhofer Institute for Chemical Technology ICT, Department of Environmental
Engineering, Joseph-von-Fraunhofer-Str. 7, D-76327 Pnztal, Germany
† Electronic supplementary information (ESI) available. See DOI:
10.1039/d1ra00649e
8970 | RSC Adv., 2021, 11, 8970–8985
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
Another electrochemical approach exploiting the same
reaction type of pinacolization is the synthesis of polyvanillin by
reductive polymerization of 6,60-dihydroxy-5,50-dimethoxy-(1,10-
biphenyl)-3,30-dicarboxaldehyde, commonly called divanillin or
5,50-bis-vanillin (Scheme 2). The feasibility of this synthesis was
shown by Amarasekara et al. in 2012 at a lead cathode in 1 M
sodium hydroxide solution leading to a total vanillin-based
polymer.23 The oxidative phenol coupling of vanillin to diva-
nillin can be classically performed by means of Fe(II)-based
catalysts with yields of about 50%.24–28 The sustainable synthesis
is performed enzymatically employing either laccase from Tra-
metes versicolor in oxygenated solution29 or horse-radish
peroxidase type I and hydrogen peroxide30–33 with yields of
about 80% enabling a green alternative for this C–C coupling
step. Besides for the electrochemical synthesis of polyvanillin
divanillin was e.g. used in the preparation of Schiff-base poly-
mers, which were investigated as chelating agents with metal
ions.34
Fig. 1 summarizes the biobased value-added chain from
lignocellulose to the biobased polymer polyvanillin and the
polymer building block hydrovanilloin. As these product lines
only employ sustainable process steps and the use of the side
stream product lignin as starting material, a total green process
would be realized.
The electrochemical reduction of the aldehyde groups in the
pinacolization steps compete in aqueous alkaline media with
the hydrogen evolution reaction (HER):
Scheme 1 Electrochemical vanillin reduction in alkaline aqueous
sodium hydroxide solution. The covalent bond O–Na of the metal
ketyl is only used for convenience.18
is nowadays still sparsely used as synthetic process step in the
preparative organic chemistry.15,16
A promising rst reaction step for the synthesis of biobased
thermoplastic, such as polycarbonates and thermosetting
resins, is the electrochemical conversion of vanillin to its
hydrodimer hydrovanilloin by pinacolization of the aldehyde
group. Electrochemical pinacolization of vanillin to hydro-
vanilloin was rst described in 1952 by I. A. Pearl at lead elec-
trodes in aqueous sodium hydroxide solution.17 Jow et al. later
investigated product distributions and the kinetics of the
cathodic reduction of vanillin at a mercury pool cathode.18
Vanillyl alcohol and hydrovanilloin were found as the only
products, in which the latter being the dominant product at
lower current densities, higher pH values and high Na+
concentrations. On the basis of their results and published
literature of the electroreduction of aromatic aldehydes they
proposed the following mechanism, where vanillin is reduced
via an adsorbed metal ketyl intermediate (Scheme 1).19,20 Due to
the bisphenolic structure of hydrovanilloin the molecule could
be used as a renewable drop-in for exemplary bisphenol A. The
feasibility of the usage of hydrovanilloin as building block was
shown e.g. in the synthesis of a hydrovanilloin-formaldehyde
2H2O + 2eꢀ / 2OHꢀ + H2
(1)
Therefore, cathode materials showing sufficient high over-
potentials for the HER are required to achieve adequate faradaic
efficiencies (FEs) for the pinacol product. These are oen
hazardous materials such as lead, mercury and cadmium,
which can suffer from cathodic corrosion due to dissolution or
can form highly toxic organometallic compounds.35,36 The use of
these heavy metals electrode materials, as in the above
mentioned earlier studies, should therefore be avoided in
future electro-synthetic processes.
copolymer21 or
resin.22
a hydrovanilloin-diglycidyl ether phenoxy
To further establish reliable electrochemical processes for
the pinacolization C–C coupling steps of vanillin to
Scheme 2 Synthesis route of vanillin to polyvanillin.
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 8970–8985 | 8971

View Article Online
RSC Advances
Paper
Fig. 1 Sustainable value-added chain of lignocellulose to the biobased
polymer polyvanillin and the polymer building block hydrovanilloin.
hydrovanilloin and divanillin to polyvanillin, a more detailed
understanding of the underlying reaction mechanisms and the
effects of electrochemical parameters, such as current densities,
applied charges, on product distributions, yields and structural
features are required. The present study addresses these
subjects extending the feasibility study of polyvanillin and the
state of the art of the electrochemical synthesis of hydro-
vanilloin as well as further screens for green, sustainable and
stable electrode materials as alternatives for the commonly
used heavy metals.
Fig. 2 Electrochemical batch cell set-up (H-type).
(Fig. 2). Anode and cathode compartments were separated by
a Naon® N324 membrane (Ion Power GmbH), which was
brought to the Na+-form by soaking it in 1 M NaOH for at least
24 h. A Bio-Logic SAS SP-150 potentiostat and a VMP3 10 A
booster were used in all electrochemical experiments. The
catholyte was stirred by a magnetic stirrer at 250 rpm
throughout all bulk electrolysis experiments. The H-type cell
was immersed in a water bath and cell temperatures were
ꢁ
conrmed to stay between 22–25 C during all measurements.
2 Experimental
2.1 Chemicals
2.2.2 Electrodes. Three different working electrodes con-
sisting of Pb, Zn and glassy carbon (GC) sheets with dimensions
of 2 ꢂ 6 cm2 were used. The geometric area immersed in the
electrolyte facing the counter electrode was 5 cm2, which was
used as reference area for the calculations of the current
densities. To ensure no limitation by the parallel running
reaction at the anode side, which is the oxygen evolution reac-
tion (OER):
All aqueous solutions were prepared with ultrapure water (0.055
mS cmꢀ1). Vanillin (Sigma Aldrich, reagent plus, 99%), vanillyl
alcohol (Sigma Aldrich, supelco, $99%), vanillic acid (Sigma
Aldrich, purum, $98%), sodium hydroxide (Carl Roth, p.a.,
$98%), 1 M hydrochloric acid (Carl Roth, standard solution/
p.a.), concentrated hydrochloric acid (VWR Chemicals,
supelco), methanol (Carl Roth, LC-MS grade), ethanol absolute
(VWR Chemicals, LC-MS grade), acetic acid (Carl Roth, Ph. Eur.,
100%), H2O2 solution (Sigma Aldrich, Supelco, $30%), horse-
radish peroxidase (Sigma Aldrich, Type I, $50 units per mg),
acetonitrile (Sigma Aldrich, HPLC grade), ortho-phosphoric acid
(Carl Roth, Ph. Eur., 85%), DMSO-d6 (Sigma Aldrich, 99.9
atom% D), pyridine-d5 (Sigma Aldrich, for NMR
spectroscopy, min. 99.8% D), CDCl3 (Sigma Aldrich, for NMR
spectroscopy, min. 99.9% D), 2-chloro-4,4,5,5-tetramethyl-1,3,2-
dioxaphospholane also called Cl-TMDP (Sigma Aldrich, 95%,
for in situ labelling in 31P-NMR-analysis), cyclohexanol (Sigma
Aldrich, 99%, internal standard) were used as received for all
experiments. As electrode materials Pb sheets (ChemPur,
99.99%), Zn sheets (ChemPur, 99.99%), glassy carbon (GC –
HTW Germany, SIGRADUR®) and Ni-foam (AQUA-TITAN, Ger-
many) were used.
4OHꢀ / O2 + 2H2O + 4eꢀ
(2)
a 2 ꢂ 2.5 cm2 Ni-foam anode was used in all experiments due to
its fast kinetics of the OER and high surface area.37 Potentials
were measured against a reversible hydrogen reference elec-
trode (RHE – HydroFlex®, Gaskatel, Germany) placed as close as
possible to the working electrode. The Ni-foam anode was used
without preparation. Prior to each experiment in case of Pb and
Zn the cathodes were polished with SiC papers with decreasing
roughness (Struers GmbH, FEPA #P180/#P500/#P1000 – grade),
rinsed with ultrapure water, dipped in absolute ethanol to
remove organic impurities, held 5 s in 1 M HCl for Zn or
concentrated HNO3 for Pb, respectively, and rinsed with ultra-
pure water again. GC electrodes were obtained mirror nished.
The electrodes were roughly rinsed with water and absolute
ethanol. When reusing them, the mirror nish was refreshed by
polishing with diamond suspension with decreasing particle
size (3 mm, 0.25 mm, 0.05 mm, Buehler, Germany). Mass changes
of working electrodes were determined by weighing before and
aer the electrolysis experiments.
2.2 Electrochemical experiments
2.2.1 Set-up. A conventional electrochemical cell (H-type)
with a standard three electrode setup was built for linear vol-
tammetry studies as well as for bulk electrolysis experiments
8972 | RSC Adv., 2021, 11, 8970–8985
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
2.2.3 Electrochemical procedures. For the vanillin reduc- 2.3 Analytical instrumentation
RSC Advances
tion experiments the catholyte was prepared as 0.2 M vanillin
solution by dissolving 1.522 g of vanillin in 50 mL of 1 M
sodium hydroxide solution. For the divanillin reduction to
polyvanillin a 0.1 M divanillin solution was prepared by dis-
solving 1.511 g of divanillin in 50 mL of 1 M sodium hydroxide.
The anolyte consisted of 50 mL of 1 M sodium hydroxide
solution in both cases. Immediately aer the electrodes were
immersed into the electrolyte and the setup was ready, the
experiment's current was applied to the cell to minimize open-
circuit corrosion of the cathode's surface in aqueous alkaline
media. Linear sweep voltammograms (LSV) were recorded by
sweeping the potential with a potential sweep rate of 10 mV sꢀ1
from ꢀ0.5 V vs. RHE to ꢀ1.6 V vs. RHE. For measuring back-
ground currents, the same experiment was conducted without
adding vanillin to the catholyte compartment. Electrolysis was
performed in galvanostatic mode until a given charge had
passed. Values of charges are presented in F per mole vanillin or
divanillin, respectively, where F is the Faraday unit of charge,
i.e. the charge of one mole of electrons (96 485 C). Fresh elec-
trolyte solutions were prepared for each experiment. Recorded
potentials were corrected by the solution's resistance between
the working electrode and the reference electrode, which was
measured by galvano electrochemical impedance spectroscopy
(GEIS) prior to and aer each experiment. Frequencies were
measured from 100 kHz to 100 mHz at the experiment's elec-
trolysis current. The alternating current amplitude was set to
10% of the electrolysis current. The vanillin electrolysis was
monitored throughout electrolysis by withdrawing aliquots of
10 mL from the catholyte solution every 15–30 minutes
depending on the overall reaction time. The aliquots were
diluted with 100 mL methanol and 10 mL internal standard
(vanillic acid) and then quantitively analyzed by capillary elec-
trophoresis (CE) coupled with UV detection. FEs for the vanillin
reduction products were calculated as follows:
2.3.1 Capillary electrophoresis (CE). A CE system with UV
detection (Prince Technologies, PrinCE-C770 and PrinCE
Next875 DAD) was used to analyze the withdrawn aliquots in the
vanillin electrolysis. A fused silica tubing of 50 mm inner
diameter and an effective length of 8.5 cm (short capillary
application) was used as capillary (total length of capillary was
60 cm). The capillary was conditioned before every run for 2 min
(2000 mbar) with background electrolyte. The wavelength of the
UV detector was selected to be 283 nm according to the litera-
ture.18 For the background electrolyte 0.8 g of phosphoric acid
were dissolved in 20 mL 1 M NaOH, 30 mL Water and 50 mL
methanol. The conductivity of the buffer solution was 6.1
mS cmꢀ1 and the pH value 12.2 at 20 ꢁC. Linear calibration
curves were obtained from peak areas for vanillin and vanillyl
alcohol from pure substances. Hydrovanilloin was obtained
aer a series of treatments of the catholyte solution, which is
described elsewhere.23 10 mL of vanillic acid were added to each
sample as internal standard resulting in an end concentration
of 1.7 mM. The electrophoretic separation was carried out at
ꢀ20 kV and 20 ꢁC (corresponding current in the range of ꢀ60 to
ꢀ70 mA). The injection parameters were 60 mbar for 0.1 min at
ꢁ
20 C.
2.3.2 HPLC-DAD-MSD. HPLC-DAD-MSD analysis of
selected isolated products of the vanillin reduction experiments
and of divanillin was realized on a series 1200 HPLC-System
from Agilent Technologies (binary pump SL 600 bars, auto
sampler SL, DAD SL, LC MS Ion Trap XCT plus Series 6330 with
multi-mode ion source MMAPCI/APESI). The products were
dissolved in methanol and analyzed on a Kinetex C8 HPLC-
˚
column (2.6 mm, 100 A, 3.0 mm ꢂ 100 mm). The DAD wave
length was set to 280 nm ꢃ 5 nm with a ref. wave length of
550 nm ꢃ 50 nm. The mobile phase consisted of A: acetic acid
(0.05%ical) and B: acetonitrile (in case of isolated product from
vanillin) or methanol (in case of divanillin) in gradient program.
The ow rate was set to 0.5 mL minꢀ1 and the oven temperature
z ꢂ F ꢂ cPðtÞ ꢂ VR
FEðtÞ ¼
ꢂ 100%
(3)
ꢁ
to 45 C.
I ꢂ t
2.3.3 Size exclusion chromatography (SEC) analysis. The
structural features weight-average molecular weight (MW) and
number-average molecular weight (MN) of the isolated polymer
samples from the divanillin reduction experiments were deter-
mined by SEC using a HPLC-system with a refractive index
detector (Agilent Technologies, LC 1200, RID cell temperature:
35 ꢁC) with three columns (PSS MCX 10 mm, guard column,
where z is the number of transferred electrons for a given
product, F the Faraday constant (96 485 C molꢀ1), cP(t) the
concentration of a product at a given time t, VR the catholyte's
volume and I the circuit current. The number of transferred
electrons z is equal to 2 for vanillyl alcohol as well as for
hydrovanilloin (see Scheme 1). The change of volume of the
catholyte due to the withdrawing of the aliquots for the CE
analysis was neglectable.
2.2.4 Product isolation. The products were isolated aer
the electrolysis from the catholytes for both reaction as
described in the literature.23 Therefore, the catholyte was
transferred to a beaker, cooled in ice and acidied to pH ¼ 2.
For the titration step 1 M HCl was used in case of the vanillin
reduction experiments and concentrated HCl in case of the
divanillin reduction experiments. The precipitates were ltered,
thoroughly washed with water and dried with silica gel in
a desiccator under vacuum overnight. Isolated yields were re-
ported by weighing the dried samples.
˚
˚
analytical columns 10 m, 100 A and 10 m, 1000 A) according to
the published literature.38 2–3 mg of the samples were dissolved
in 1 mL of 0.1 M sodium hydroxide solution. The mobile phase
was 0.1 M sodium hydroxide solution with a ow rate of 1
mL minꢀ1 and an oven temperature of 35 ꢁC. Conventional
pullulan standards (342 g molꢀ1 to 805 000 g molꢀ1, PSS Mainz,
Germany) were used for calibration. Therefore, molecular
weights of samples are presented versus pullulan.
2.3.4 NMR spectroscopy. 1H-, 13C- and 2D-NMR spectra
(HSQC, 13C/1H) of isolated products were recorded on a Bruker
AVANCE spectrometer operating at 500 MHz. In case of the
vanillin reduction experiments DMSO-d6 was used as solvent
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 8970–8985 | 8973

View Article Online
RSC Advances
Paper
and in case of the divanillin reduction experiments pyridine-d5. adjusted to a value of 4 by adding 250 mL of glacial acetic acid.
Before dissolution the samples were additionally dried under An amount of 1000 units of activity of horse radish peroxidase
1
vacuum. Chemical shis for H-NMR and 13C-NMR are given were added during stirring and 7.5 mL of 30 wt% hydrogen
in ppm downeld from TMS (d ¼ 0.00). The spectra were peroxide solution were added to the solution dropwise. The
adjusted to the C–H cross coupling signals of pyridine-d5 (dH in brown precipitate was ltered aer 15 minutes, washed with hot
ppm/dC in ppm: 7.220/123.87; 7.580/135.91; 8.740/150.35) and water and dried under vacuum (10 mbar, at 45 ꢁC, two days) (8 g,
of DMSO-d6 (dH ¼ 2.5 ppm/dC ¼ 39.51 ppm), respectively. 80%). The product structure was conrmed by HPLC-DAD-MSD,
Generated 2D-NMR (HSQC) spectra were processed on a Mes- elemental analysis (Thermo Scientic, Thermo Electron Flash
tReNova soware.
EA 1112) and Fourier Transform Infrared analysis (Thermo
Further, 31P-NMR spectra were recorded of selected isolated Scientic, NICOLET 6700) (ESI, Fig. S1 and S2†). The remaining
polyvanillin samples of the divanillin reduction experiments fraction of vanillin in the isolated divanillin batch was deter-
aer in situ labeling for quantication of aliphatic and aromatic mined by GC-FID/MSD (Agilent Technologies, HP 6890) on
OH-groups. Three solvents were used for the preparation of the a DB-5-ms column aer calibration with vanillin standard
polyvanillin samples. Solvent I was a homogenous mixture of solutions. The remaining fraction of vanillin was 4.9 wt% in the
16 mL pyridine (water free) and 10 mL CDCl3 (water free). For rst batch and under limit of detection in the second.
the preparation of solvent II 10.0 mg of Cr(III)acetyl acetonate
were weighed into a HPLC vial and 2 mL of solvent I were added.
The vial was caped and the Cr(III) acetyl acetonate was dissolved
and homogenized at room temperature by vortex mixing. For
the preparation of solvent III 21.92 mg of internal standard
3 Results and discussion
3.1 Screening of cathode materials for vanillin reduction
On repeating I. A. Pearls original synthesis17 of hydrovanilloin
by electrochemical pinacolization of vanillin at lead cathodes,
Pb dissolution was found in electrolysis under given conditions.
As Zn and GC exhibit high overpotentials for the competing
HER (see eqn (1)) and are cheap as well as harmless materials,
they are screened for the electrochemical reduction of vanillin
in alkaline aqueous media. Polarization curves of background
currents resulting from the competing HER and of the actual
vanillin reduction measured by LSV are shown in Fig. 3.
Background currents are neglectable for all three electrode
materials at lower overpotentials up to ꢀ0.8 V vs. RHE. With
increasing overpotential a signicant increase of the HER
currents are observed for the three investigated cathode mate-
rials. Background currents are then in order Zn followed by Pb
and GC. For example, at a potential of ꢀ1 V vs. RHE currents of
ꢀ44.3 mA cmꢀ2 for Zn, ꢀ19.2 mA cmꢀ2 for Pb and ꢀ12.2 mA
cmꢀ2 for GC are measured for the HER. At even more negative
values background currents of Pb and GC are approaching each
other, whereas these negative potentials are hardly achieved at
Zn due the higher HER activity depolarizing the Zn cathode. All
three cathode materials show activity for the electrochemical
cyclohexanol (99%ical) was accurately weighed in a HPLC-vial
and exactly 2 mL of solvent I was added. The vial was caped
and the solution was homogenized at room temperature by
vortex mixing. The freshly prepared solvents were allowed to
stay at room temperature. The polyvanillin samples were dried
in a vacuum oven at 10 mbar and at a temperature of 50 ꢁC for
48 h. Exactly 20.00 mg of the dried polyvanillin samples were
weighed in HPLC vials. Then 800 mL of freshly prepared solvent
I, 100 mL of freshly prepared solvent II and 100 mL of freshly
prepared solvent III were added to the samples and the vials
were caped. The samples were mixed by vortex mixing at room
temperature. 10 min before the 31P-NMR-analysis 160 mL of in
situ
labeling
reagent
2-chloro-4,4,5,5-tetramethyl-1,3,2-
dioxaphospholane (Cl-TMDP) were added to the vials and
homogenized again by vortex mixing at room temperature. Only
aer the addition of Cl-TMDP were the samples completely
dissolved and no precipitation was observed. The nal sample
solution was transferred to an NMR tube. The 31P-NMR analysis
was done on a Bruker AVANCE 300 MHz. Chemical shis are
reported relative to the sharp signal (132.2 ppm) originating
from the reaction between water and Cl-TMDP. Following NMR
parameters were used: scans ¼ 1024, pulse delay ¼ 5 s, 90ꢁ pulse
and line broadening ¼ 2 and default baseline correction. 31P-
NMR measurement is based on the developed method by
Granata and Argyropoulos.39 31P-NMR-spectra were processed
on ACD Lab 12.0, 1D-processor soware. Integration for quan-
titative calculation was done according to Pu et al.40 Integration
limits for aliphatic OH-groups were 150.0–145.40 ppm, for
phenolic OH groups 144.50–137.50 ppm and for the internal
standard (cyclohexanol) 145.39–144.75 ppm (ESI, Fig. S12†).
2.4 Synthesis of divanillin
Divanillin was synthesized using horse radish peroxidase and
hydrogen peroxide according to the procedure described in the
Fig. 3 Cathodic LSVs for vanillin reduction at Pb, Zn and GC in static
literature.41 10 g of vanillin were dissolved in 400 mL deionized
solution. Potential sweep rate is 10 mV sꢀ1. Background currents of
ꢁ
water at a temperature of 40 C by stirring. The pH value was neat electrolyte are shown as dashed lines.
8974 | RSC Adv., 2021, 11, 8970–8985
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
Fig. 4 Potential–charge plots for vanillin reduction at Pb, Zn and GC at a current density of 30 mA cmꢀ2 (A) and at current densities of 60 and 100
mA cmꢀ2 (B). Exemplary concentration course of the electrochemical vanillin reduction. Analysis of the catholyte was carried out by CE-UV (C).
Calculated FEs for hydrovanilloin and vanillyl alcohol (D). Parameters: vanillin concentration 0.2 M in 1 M sodium hydroxide solution, catholyte
stirred at 250 rpm, RT.
reduction of vanillin. Onset potential at Zn and Pb differ in the electrochemical vanillin reduction under the three screened
millivolt range for the electrochemical vanillin reduction at electrode materials but exhibits the lowest activity for the
zꢀ0.575 V vs. RHE. At GC the onset potential is shied by competing HER. Almost similar catalytic properties to the
50 mV in negative direction and cathodic currents are inferior to commonly used Pb cathodes are observed for the electro-
Pb and Zn. At a potential of ꢀ0.800 V vs. RHE reductive currents chemical vanillin reduction at Zn cathodes, but also showing
of ꢀ24.7 mA cmꢀ2 for Pb, ꢀ21.4 mA cmꢀ2 for Zn and ꢀ16.1 mA a higher activity of the competing HER in the higher current
cmꢀ2 for GC are observed in static solution. Therefore, GC density region. To evaluate further the impact of the competing
shows comparable the worst catalytic properties for the
Table 1 Product yields, mass differenceꢀs1oaf cathodes and potentials at the start and end of electrolysis of the electrochemical vanillin reduction
after an applied charge of 2 F molVanillin
Current density Mass difference Vanillin conversion Yield hydrovanilloin Yield vanillyl Potential at
Potential at
Cathode material (mA cmꢀ2
)
cathode (mg)
(%)
(%)
alcohol (%) start (mV vs. RHE) end (mV vs. RHE)
Zn
Pb
GC
30
60
100
30
60
100
30
+0.21
+0.00
99.0
90.7
80.1
99.9
95.0
93.7
97.9
92.9
88.0
97.7
88.9
78.6
99.1
93.4
87.1
95.9
78.6
74.9
1.3
1.8
1.5
0.9
1.6
6.6
2.0
14.3
13.1
ꢀ720
ꢀ824
ꢀ1018
ꢀ1112
ꢀ1134
ꢀ1040
ꢀ1083
ꢀ1195
ꢀ1326
ꢀ1640
ꢀ1696
ꢀ0.09
ꢀ2.86
ꢀ5.90
ꢀ19.30
+0.05
ꢀ1073
ꢀ686
ꢀ803
ꢀ1266
ꢀ750
60
100
+0.12
+0.13
ꢀ1012
ꢀ1791
a
Electrolysis was performed under following conditions – catholyte: 0.2 M vanillin in 1 M sodium hydroxide, anolyte: 1 M sodium hydroxide, room
temperature, catholyte stirred at 250 rpm, potentials were corrected by electrolyte resistance measured by GEIS. Catholyte was analyzed by CE-UV.
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 8970–8985 | 8975

View Article Online
RSC Advances
Paper
HER quantitatively as well as determine product distributions
bulk electrolysis experiments were conducted.
Almost no mass difference for the Zn and GC electrodes are
found aer electrolysis for all investigated current densities
deducing stable behavior. Mass differences range from ꢀ0.09 to
+ 0.21 mg and + 0.05 to + 0.13 mg for Zn and GC, respectively.
Small positive changes of the cathode's mass may be due to
small deposits of vanillin or reaction products on the electrode
surface and measurement errors. However, for reproducible
3.2 Product distributions and FEs of vanillin bulk
electrolysis experiments
Bulk electrolysis of vanillin in 1 M NaOH was performed at three potential courses at GC intensive polishing of the electrode
different current densities 30, 60 and 100 mA cmꢀ2 at a xed surface is necessary showing some change of the morphology
applied charge of 2 F molꢀ1. Potential–charge plots in galva- under given conditions. This agrees well with observations by
nostatic electrolysis mode at Pb, Zn and GC cathodes are shown Chandrasekaran et al., who found a sensitive behavior to the
in Fig. 4A and B. Product yields, mass differences of cathodes in pre-treatment by means of polishing of GC cathodes for the
the electrolysis and potentials at start and end of electrolysis are reduction of vanillin in aprotic and aqueous media.42 For Pb
summarized in Table 1.
signicant loss of electrode material was observed. Mass
Throughout the electrolysis at 30 and 60 mA cmꢀ2 the differences were in the range between ꢀ2.86 to ꢀ19.30 mg per
cathode potential drops to more negative values due to electrolysis run. Similar cathodic corrosion of lead was observed
¨
decreasing vanillin concentration and increasing conversion e.g. by Gutz et al., who used leaded bronzes as alternative
resulting in higher overpotentials for the electrochemical cathode material for electro-reduction such as deoxygenation,
reduction of vanillin. For example, at 30 mA cmꢀ2 at an dehalogenations and hydrodimerization.36
applied charge around 0.60 F molꢀ1 for GC and 0.75 F molꢀ1
Nearly full conversion of vanillin was achieved aer an
for Pb and Zn the potential shis to a new more negative applied charge of 2 F molꢀ1 at all three electrode materials at 30
plateau, where preferably the HER is taking place. At this stage mA cmꢀ2. Hydrovanilloin is the main product in all cases and
the vanillin electrolysis is under mass transfer control. An vanillyl alcohol formation is neglectable with a maximum yield
exemplary progression of the electrochemical vanillin reduc- of 2% at GC. With increasing current density vanillin conver-
tion for a Zn cathode at 30 mA cmꢀ2 is shown Fig. 4C and D. A sion is decreasing at the xed applied charge of 2 F molꢀ1
,
change of the slope of the vanillin conversion is observed because the current is consumed by the competing HER. At 100
transitioning from a linear (kinetically controlled) to an mA cmꢀ2 the lowest vanillin conversion of 80.1% is observed for
exponential (mass transport controlled) concentration Zn. At 100 mA cmꢀ2, however, the highest selectivity for the
decrease. At 100 mA cmꢀ2 no drop of the cathode's potential is pinacolization product hydrovanilloin is achieved for Zn among
observed suggesting that the reaction is under full mass the three investigated cathode materials with neglectable yields
transport control from start on and fastest conversion rates are of vanillyl alcohol of 1.5%. In comparison signicant vanillyl
achieved under given conditions. Potentials at the start and at alcohol yields of 13–14% are observed for GC at 60 and 100 mA
the end of the electrolysis decrease with increasing current cmꢀ2 as well as 6.6% for Pb at 100 mA cmꢀ2
density due to a stronger cathodic polarization of the elec-
Analogously to Jow et al. FEs of the reduction products
.
trodes. Similar current–potential–charge behavior was hydrovanilloin and vanillyl alcohol aer an applied charge of
observed by Jow et al. at a mercury pool electrode.18 For 0.6 F molꢀ1 were calculated.18 At this stage the reaction is still
example, starting potentials for a 0.033 M vanillin solution in kinetically controlled at low current densities and FEs stay
1 M NaOH decreased from ꢀ1.72 V vs. SCE at 30 mA cmꢀ2 to
ꢀ2.05 V vs. SCE at 100 mA cmꢀ2. Potential shis to the HER
plateau were slightly earlier at applied charges of 0.1–0.5 F
molꢀ1, which may be due to slower mixing of the catholyte and
different cell and electrode design. Here, for 30 and 60 mA
cmꢀ2 the cathode's potentials for the three materials Zn, GC
and Pb at the start of the electrolysis are approximately similar
and are in the range between ꢀ0.68 and ꢀ0.824 V vs. RHE
except for GC at 60 mA cmꢀ2, where a potential at the start of
ꢀ1.01 V vs. RHE is observed. At the highest current density of
100 mA cmꢀ2 and in the mass transport limited region of the
electrolysis the potential is dominated by the HER. Due to high
HER overpotentials very negative potentials up to ꢀ1.79 V vs.
RHE are reached either at GC at 30 and 60 mA cmꢀ2 in the
mass transport limited regions or at 100 mA cmꢀ2 at Pb and
GC throughout the whole electrolysis. Minor HER over-
potentials at Zn compared to the other two investigated
cathode materials hinder the potential from reaching very
Fig. 5 Faradaic efficiencies for hydrovanilloin and vanillyl alcohol
formation in electrochemical bulk electrolysis of vanillin after 0.6 F
molꢀ1 (vanillin concentration 0.2 M in 1 M sodium hydroxide solution,
catholyte stirred at 250 rpm, RT). Error bars indicate highest and lowest
values of two electrolysis runs.
negative values being at maximum ꢀ1.13 V vs. RHE at 100 mA
cmꢀ2
.
8976 | RSC Adv., 2021, 11, 8970–8985
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
constant (see Fig. 4D).43 For proving reproducibility, the exper- potential of ꢀ0.575 V vs. RHE at Zn seems to t perfectly in the
iments for the FE evaluation were conducted two times and the viable potential window.
resulting mean values are presented in Fig. 5. At the lowest
current density of 30 mA cmꢀ2 FEs for hydrovanilloin formation
are in the range of 85–88% for Pb and Zn, whereas for GC a FE of
z98% is obtained due to its higher overpotential for the
competing HER. FEs for vanillyl alcohol are below 0.5% for Pb
and GC. A slightly higher FE for vanillyl alcohol formation at Zn
cathodes of 2.3% is observed. Comparatively, no vanillyl alcohol
3.3 Product isolation and identication of the
electrochemical vanillin reduction
Isolated product yields obtained by precipitation aer acidi-
cation of the catholyte differ from the analyzed hydrovanilloin
amount in the catholyte solution. Fig. 6 shows the isolated
product yields at 30 mA cmꢀ2 for the three investigated cathode
formation was reported by Jow et al. for a Hg cathode in the
materials versus the applied charge. Isolated yields at 60 and 100
current density range of 15–100 mA cmꢀ2. FEs for hydro-
mA cmꢀ2 at an applied charge of 2 F molꢀ1 are show in the (ESI,
vanilloin decreased from 80% to 45% when increasing the
Table S1†). A maximum isolated yield of 72% is achieved aer
current density from 15 mA cmꢀ2 to 100 mA cmꢀ2. But it should
an applied charge of 2 F molꢀ1 for Zn and Pb cathodes at 30 mA
be noted that the starting vanillin concentration was lower
cmꢀ2, although almost full vanillin conversion to hydro-
vanilloin is observed in the directly analyzed catholyte (see
being 33 mM of vanillin in 1 M NaOH.18 Here, total FEs of
vanillin reduction products decrease gradually to 71–75% at 100
Table 1). Similar isolated yields are reported in the literature
mA cmꢀ2 for Pb or GC and to 55% for Zn. However, comparable
ranging between 69 and 76% for applied charges of 1.1 F molꢀ1
FEs for hydrovanilloin formation of 55–62% are achieved
among the three investigated cathode materials, since the FE
and 1.56 F molꢀ1, respectively.17,44 Published yields of 86% by
Amarasekara et al. could not be reproduced, even though
for vanillyl alcohol formation stays at low percentages for Zn,
isolation was performed similar titrating the catholyte to pH ¼ 2
whereas FEs of 12–13% for vanillyl alcohol are obtained for GC
with HCl and having almost full vanillin conversion. Neither
and Pb.
current densities nor applied charges were mentioned.23
According to the published literature the dissociated
Maximum isolated yields of 69% for the electroreduction of
hydroxyl group in basic aqueous media leads to a decrease of
vanillin at GC are slightly lower compared to Pb and Zn due to
the stability of the metal ketyl intermediate due to electron
slightly higher formation of vanillyl alcohol, which remains in
release to the benzene ring and therefore to a preferably
the ltrate. The isolated product yield generally decreases with
dimerization reaction pathway (see Scheme 1). Signicant
increasing current density from 30 to 100 mA cmꢀ2 at a xed
vanillyl alcohol formation is only observed at very negative
applied charge of 2 F molꢀ1 due to the increasing charge
potentials, where kinetics for the second electron transfer to
consumption of the competing HER. An isolated yield of 49% is
the metal ketyl intermediate would be fast enough compared
reported for the Zn cathode at 100 mA cmꢀ2. The isolated
to the dimerization pathway.18 These negative potentials are
product was identied by HPLC-DAD-MSD, FT-IR and 1H-, 13C-
either reached herein at Pb and GC electrodes at higher
and 2D-NMR (HSQC, 13C/1H) as pure hydrovanilloin (ESI,
current densities or in mass transport controlled region of the
Fig. S4–S6†), which agrees with the published literature.17,23
This shows that the product isolation by the commonly used
titration step is sub optimal giving space for further improve-
ment, e.g. by membrane technology. Interestingly to note is that
electrolysis. For example, signicant vanillyl alcohol forma-
tion at GC at 60 mA cmꢀ2 is only observed aer an applied
charge of 0.5 F molꢀ1, when the potential drops from ꢀ1 V vs.
RHE to the HER plateau of zꢀ1.6 V vs. RHE (ESI, Fig. S3†).
the pinacolization at Zn electrode exhibits the same
Therefore, FEs at 0.6 F molꢀ1 for GC at 60 mA cmꢀ2 are still
low, although a high vanillyl alcohol yields is observed aer
an applied charge of 2 F molꢀ1. It is assumed that the
comparable lower HER overpotential depolarizes the Zn
electrode hindering the potential of reaching very negative
values and therefore inhibiting the second electron being
transferred to the metal ketyl intermediate. As a result, the
highest selectivity for hydrovanilloin at higher current
densities is obtained for Zn and side product formation of
vanillyl alcohol is neglectable. The negative impact of the
minor HER overpotential at Zn on the FE of hydrovanilloin is
low showing comparable FEs of z55% to Pb and GC up to
high current densities. Furthermore, it is interestingly to note
that the viable potential window of Zn for electroreductions in
aqueous alkaline media is very limited. At potentials higher
and equal to the open circuit potential of ꢀ0.450 V vs. RHE
corrosion occurs and at potentials more negative than
Fig. 6 Isolated product yields versus applied charge for the three
investigated cathode materials Zn, Pb and GC. Products were isolated
ꢀ0.800 V vs. RHE a signicant increase of the competing HER
is observed. The electroreduction of vanillin with an onset
by titrating the catholyte to pH ¼ 2 with 1 M HCl followed by filtrating,
washing and drying.
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 8970–8985 | 8977

View Article Online
RSC Advances
Paper
stereoselectivity as it was previous reported for Pb electrode. charge on the carboxyl group due to the electron transfer may be
The pinacolization can lead two different diastereomeric forms delocalized by the aromatic rings leading to a more difficult
of hydrovanilloin, namely the DL-form and the meso-form. Their reduction of the second aldehyde group. Moreover, the two
ratio can be determined from the 1H-NMR methoxy group phenolic groups of divanillin are dissociated in the aqueous
signals or the Ca–Ha/Ca0–Ha0 signals in the aliphatic region of alkaline electrolyte increasing the total negative charge of the
hydrovanilloin. Here, a meso- to DL-form ratio of 78% is calcu- molecule. Indication of a second electroreduction peak from
lated, which agrees well with the literature showing a ratio of the second aldehyde can just be suspected at a potential of
z80% at 30 mA cmꢀ2 synthesized at a Pb cathode.21 However, it zꢀ850 mV vs. RHE, since at these negative potentials reduction
should be mentioned that the meso- to DL-form ratio is depen- currents of the competing HER strongly overlap. As for the
dent on the applied current density and e.g. a maximum ratio of electroreduction of vanillin current densities up to 100 mA
94% was achieved by Amarasekara et al. at 64 mA cmꢀ2 at Pb cmꢀ2 were achievable, maximum applied currents for the
cathodes.21
divanillin reduction have to be reduced in the following bulk
electrolysis experiments. A maximum applied current density of
60 mA cmꢀ2 is chosen.
3.4 Linear sweep voltammetry of divanillin
The molecular weight build-up of divanillin to polyvanillin is
carried out by electroreduction of the two aldehyde groups of
divanillin leading to the pinacol coupling product. Fig. 7 shows
the LSV of divanillin in aqueous alkaline solution at a Zn
cathode and for comparability the LSV of vanillin. It was ex-
pected that for the same number of aldehyde groups of vanillin
and divanillin onset potentials and reduction currents would be
equal due to their similar molecular structure and the far
distance between the aldehyde groups of divanillin. For
example, the electrochemical reduction behavior of iso-
phthalaldehyde bearing two aldehyde groups is similar to the
electroreduction of benzaldehyde in acidic and alkaline media
at twice the concentration.45 However, the electroreduction
product of benzaldehyde in acidic and alkaline media is the
corresponding alcohol.46 For the electroreduction of vanillin the
main product in alkaline media is the pinacol product hydro-
vanilloin due to the induced instability of the dissociated
phenolic group in para-position of the aldehyde leading pref-
erably to dimerization.18 The onset potential of divanillin is
shied 50 mV in negative direction and cathodic currents are
inferior compared to the vanillin reduction at twice the
concentration. It is therefore assumed that the electroreduction
of one aldehyde group of divanillin is affecting the electro-
reduction of the second aldehyde group. The resulting negative
3.5 Bulk electrolysis of divanillin to polyvanillin
Since in the vanillin reduction experiments Zn was found as
excellent replacement for Pb as cathode material for hydro-
dimerization, bulk electrolysis experiments of divanillin are
primarily conducted with Zn as cathode. Bulk electrolysis
experiments of divanillin at the Zn cathode are summarized in
Table 2 and Fig. 8A–D.
At the lowest applied current density of 15 mA cmꢀ2 the
reaction is kinetically controlled showing a drop of the cathode
potential from ꢀ800 mV vs. RHE to ꢀ900 mV vs. RHE at an
applied charge of z2 F molꢀ1 due to decreasing divanillin
concentration (Fig. 8A). In the latter part of the reaction the
divanillin reduction is under mass transport control and, as in
the electrochemical vanillin reduction, the competing HER is
the dominating reaction. No potential drop at a current density
of 60 mA cmꢀ2 corresponding to a cathode potential of zꢀ1 V
vs. RHE is observed indicating that the reaction is under mass
transport control respectively highest possible conversion rate
from start on. The slight increase of the cathodic potential with
increasing applied charge in the mass transport-controlled
region is explained by the decreasing resistance of the electro-
lyte in the cathodic chamber, which was typically 0.5 U at an
applied charge of 8 F molꢀ1. On the basis of charge neutrality
Na+-ions migrate through the ion selective membrane from the
anodic to the cathodic chamber leading to a lower resistance
between the working and the reference electrode on the cathode
side and therefore to a lower recorded potential with increasing
reaction time.
An exemplary molecular weight build-up from vanillin to
polyvanillin at the Zn cathode for a current density of 30 mA
cmꢀ2 in the reductive polymerization step of divanillin is shown
in Fig. 8B. Since in the rst batch of the enzymatic conversion of
vanillin to divanillin a remaining vanillin fraction of 4.9 wt%
was found, a small pre-peak in the SEC spectra of divanillin is
observed. It should be noted that vanillin and divanillin are no
polymeric compounds and have an exactly weight of 152 g
molꢀ1 for vanillin and 302 g molꢀ1 for divanillin. Due to their
low molecular weight such substances are usually not appro-
priate for SEC analysis and the broad molecular weight distri-
butions (MWDs) do not correspond to their actual molecular
weight. However, these compounds are shown in Fig. 8B for
Fig. 7 Cathodic LSV for divanillin reduction in alkaline aqueous media
at a Zn cathode in static solution. Potential sweep rate is 10 mV sꢀ1
.
Background current in pure electrolyte is shown as dashed line. For
comparability the LSV of a 0.2 M vanillin aqueous alkaline solution at
a Zn cathode is shown in red.
8978 | RSC Adv., 2021, 11, 8970–8985
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
Table 2 Impact of current density, applied charge and divanillin concentration on isolated yields and molecular weights of isolated polyvanillina
b
b
Current density
Applied charge
Divanillin concentration
Isolated yield
(%)
MW
MN
Run
(mA cmꢀ2
)
(F molꢀ1
)
(mol Lꢀ1
)
(g molꢀ1
)
(g molꢀ1
)
MW/MN
1
2
3
4
5
6
7
8
15
15
15
30
30
30
60
60
60
60
60
2
4
8
2
4
8
2
4
8
8
8
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.2c
0.3c
57.9
66.2
64.1
74.6
64.2
69.9
71.2
64.8
69.7
72.8
81.1
1771
3132
3171
2095
3103
3023
1687
2864
3016
3082
3208
1360
2321
2362
1464
2174
2186
1239
1748
2201
2260
2344
1.30
1.35
1.34
1.43
1.42
1.38
1.36
1.64
1.37
1.36
1.37
9
10
11
a
Electrolysis was performed under following conditions – cathode: Zn (5 cm2), anode: Ni-foam, catholyte: 0.1 M divanillin in 1 M sodium hydroxide
(rst batch of divanillin containing 4.9 wt% vanillin, if not stated otherwise), anolyte: 1 M sodium hydroxide, room temperature, catholyte stirred at
b
250 rpm, isolated products were analyzed by SEC. Calibrated versus pullulan standard. c Second batch of divanillin containing no vanillin.
Fig. 8 (A) Potential–charge plots of divanillin reduction. (B) Exemplary molecular weight build-up from vanillin to divanillin followed by cathodic
polymerization to polyvanillin at 30 mA cmꢀ2 up to an applied charge of 8 F molꢀ1. (C) Impact of current density on the molecular weight and
shape of polyvanillin. (D) Impact of concentration on the molecular weight and shape of polyvanillin (0.2 M and 0.3 M divanillin reduction was
performed with second batch of divanillin). Electrolysis was performed under following conditions – cathode: Zn (5 cm2), anode: Ni-foam,
catholyte: 0.1 M divanillin in 1 M sodium hydroxide (carried out with first batch of divanillin containing 4.9 wt% vanillin, if not stated otherwise),
anolyte: 1 M sodium hydroxide, room temperature, catholyte stirred at 250 rpm, isolated products were analyzed by SEC. SEC was calibrated
versus pullulan standard.
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 8970–8985 | 8979

View Article Online
RSC Advances
Paper
Table 3 Molecular weight data of synthesized polyvanillin at Pb and
a better visualization of the process. The actual polymerization
of divanillin to polyvanillin shows a bimodal frequency distri-
bution of molar mass under given electrochemical process
conditions. A rst local maximum is observed at a molecular
weight of z1000 g molꢀ1 and a second local maximum at
z3200 g molꢀ1. With increasing reaction time or applied
charge the rst local maximum decreases and the second local
maximum increases until a nal state of the MWD shape is
reached. The rst local maximum becomes more and more
a shoulder. This result is observed for all three investigated
current densities of 15, 30 and 60 mA cmꢀ2 (ESI, Fig. S7 and
S8†). Similar bimodal molecular weight distributions were e.g.
observed for the emulsion polymerization of ethylene47 or
isotactic polystyrene.48 No further change of the shape of the
MWD is observed aer an applied charge of 4 F molꢀ1 (double
theoretical charge for full divanillin pinacolization) for 15 and
30 mA cmꢀ2, whereas this shape is achieved aer 8 F molꢀ1
GC cathodesa
Current
density
Applied
charge
b
b
Cathode
material
MW
MN
(mA cmꢀ2
)
(F molꢀ1
)
(g molꢀ1) (g molꢀ1) MW/MN
Pb
Pb
Pb
GC
GC
GC
15
30
60
15
30
60
8
8
8
8
8
8
2444
2958
1898
2986
3382
1264
2023
2285
1501
2384
2631
1017
1.21
1.29
1.26
1.25
1.29
1.24
a
Electrolysis was performed under following conditions – anode: Ni-
foam, catholyte: 0.1 M divanillin in 1 M sodium hydroxide, anolyte:
1
M sodium hydroxide, room temperature, catholyte stirred at
b
250 rpm, isolated products were analyzed by SEC. Calibrated versus
pullulan standard.
Fig. 9 2D-NMR (HSQC, 13C/1H) spectra of selected polyvanillin samples: (A) Zn cathode, 30 mA cmꢀ2, 2 F molꢀ1 (partly precipitated). (B) Zn
cathode, 30 mA cmꢀ2, 8 F molꢀ1. (C) Zn cathode, 15 mA cmꢀ2, 8 F molꢀ1. (D) Zn cathode, 60 mA cmꢀ2, 8 F molꢀ1. (E) GC cathode, 15 mA cmꢀ2, 8
F molꢀ1. (F) GC cathode, 60 mA cmꢀ2, 8 F molꢀ1 (mostly precipitated).
8980 | RSC Adv., 2021, 11, 8970–8985
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
(quadruple theoretical charge for full divanillin pinacolization)
at the highest current density of 60 mA cmꢀ2. The most energy
efficient polymerization is therefore conducted at low current
densities, where the competing HER is still slow, whereas fast-
est conversion is conducted at 60 mA cmꢀ2 under mass trans-
port control for the given conditions. The formation of the
bimodal frequency distribution of molar mass was at rst sug-
gested due to remaining vanillin in the divanillin, since vanillin
showed a higher activity compared to divanillin. This would
lead to fast formation of small pre-polymers followed by the
actually main polymerization of divanillin. However, poly-
vanillin synthesis experiments with a second divanillin batch
containing no remaining vanillin resulted in a similar shape
and similar polymerization degrees (see Fig. 8D, run 10–11).
Therefore, the inuence of the remaining vanillin fraction of
4.9 wt% in the rst batch can be neglected and the bimodal
frequency distribution of molar mass still remains subject of
further studies.
Further it is found that the maximum degree of polymeri-
zation is independent of the applied current density at a Zn
cathode up to potentials of zꢀ1 V vs. RHE, which corresponds
to a current density of 60 mA cmꢀ2 (Fig. 8C). Weight average
molecular weights MW up to z3200 g molꢀ1 are achieved. The
polydispersity MW/MN is relatively low showing a narrow
distribution. Values for the polydispersity lie in the range
between 1.30 and 1.43 excluding run 8, where a polydispersity of
1.64 is determined. This higher value, however, is calculated for
a partly polymerized sample at a current density of 60 mA cmꢀ2
and an applied charge of 4 F molꢀ1 showing almost two equally
high peaks in the bimodal frequency distribution of molar mass
(ESI, Fig. S8†).
Fig. 10 Proposed structural features of polyvanillin based on the
results of 2D-NMR (HSQC) analysis.
thiophene lead to the formation of electronic conducting poly-
mer lms on the anode.49,50 No deposition of polymers on Zn
cathodes was observed throughout all polyvanillin electrolysis
experiments.
Since higher concentrations of divanillin might lead to an
increase of the polymerization degree, polyvanillin synthesis
was performed for higher concentration of divanillin up to
0.3 M (run 10–11). A higher divanillin concentration would lead
to an increased radical intermediate concentration at the elec-
trode's surface, thus, might increase pinacol formation (radical
combination) instead of alcohol formation terminating polymer
chain growth. Similar observation were shown for the electro-
reduction of vanillin to hydrovanilloin.18 However, no increase
of the polymerization degree was found when tripling the
concentration from 0.1 M to 0.3 M divanillin (Fig. 8D).
Isolated yields ranged between 58% and 74% at a divanillin
concentration of 0.1 M showing no discernible trend when
varying the current density and the applied charge. The results
suggest that reaction products are in solubility equilibrium
when titrating the alkaline catholyte to pH ¼ 2 for product
precipitation also showing potential for improvement in the
isolation step as in the hydrovanilloin synthesis. However, an
increase of the isolated yields is observed for higher concen-
trations reaching values of 81% at a divanillin concentration of
0.3 M. Isolated polyvanillin yields by Amarasekara et al. for
As Pb cathodes were used in the original synthesis of poly-
vanillin by Amarasekara et al.23 and GC cathodes also showed
reasonable activity in the electrochemical vanillin reduction
experiments, the impact of different cathode materials were
further investigated. The applied charge was selected to be 8 F
molꢀ1, since full polymerization was achieved at Zn cathodes at
this amount of applied charge. The molecular weight data of the
polyvanillin samples synthesized at GC and Pb cathodes are
summarized in Table 3. Potential–charge plots and MWDs are
shown in the (ESI, Fig. S9–S11†).
Bimodal frequency distributions of molar mass are also
found for GC and Pb cathodes in all cases. Weight-average
molecular weights MW of z3000 g molꢀ1 are approximately
similar to polyvanillin synthesized at Zn cathodes for the
current densities of 15 and 30 mA cmꢀ2. A signicant drop of
the weight-average molecular weight is observed at Pb and GC at
the highest applied current density of 60 mA cmꢀ2 to 1898 g
molꢀ1 and 1264 g molꢀ1, respectively. A high amount of a low
molecular weight fraction still exists in the polyvanillin sample
synthesized at the GC cathode showing a high rst local
maximum in the bimodal frequency distributions of molar
mass at z800 g molꢀ1. Low polydispersities MW/MN ranging
between 1.21 and 1.29 are found for synthesized polyvanillin at
GC and Pb cathodes also showing very narrow MWDs. Cathode
potentials are signicant more negative compared to Zn
a 0.175 M divanillin solution were slightly higher being 91%
aer an applied charge of 35.2 F molꢀ1
.
23
Finally, it is interestingly to note, that the reductive pinaco-
lization forms soluble polymers. Since on the other hand, the
oxidative polymerization for example of pyrrole, aniline or
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 8970–8985 | 8981

View Article Online
RSC Advances
Paper
reaching values of ꢀ1.30 V vs. RHE and ꢀ1.75 V vs. RHE for Pb from the methoxy group are split into a high sharp peak and
and GC at a current density of 60 mA cmꢀ2, respectively. This a lower broader peak verifying the existence of a very low
agrees well with obtained results in the electrochemical vanillin molecular fraction. Moreover, very sharp peaks are also
reduction experiments. A decrease of the cathode's potential is observed in the aromatic region in the 1H-spectra between
observed at 60 mA cmꢀ2 for Pb and GC at the start of the elec- 6.5 ppm and 8.5 ppm further conrming the existence of a very
trolysis indicating that the reaction is not under full mass low molecular fraction. A cross-coupling signal observed at dH
¼
transport control from begin on, which is due to higher over- 10.1 ppm/dC ¼ 191.4 ppm was assigned to the unreacted alde-
potentials of the competing HER.
hyde end groups (see Fig. 10, green motif). Only very weak to
none signals at all from the remaining unreacted aldehyde end
groups can be seen in the spectra of the high molecular weight
samples (Fig. 9B–E) conrming full conversion divanillin for the
Zn cathode aer an applied charge of 8 F molꢀ1 at the three
current densities 15, 30 and 60 mA cmꢀ2 and for the GC cathode
at the current density of 15 mA cmꢀ2. As expected a strong
signal of unreacted aldehyde groups is found in the 1H-NMR
spectra for the party polymerized sample at the Zn cathode at
30 mA cmꢀ2 when only applying a charge of 2 F molꢀ1 (Fig. 9A).
Surprisingly, unreacted aldehyde groups are also found in the
polyvanillin sample synthesized at the GC cathode at 60 mA
cmꢀ2 aer an applied charge of 8 F molꢀ1 (Fig. 9F). This
explains the low molecular weight analyzed in the SEC analysis.
Since a drop of the cathode's potential at the start of the elec-
trolysis was observed indicating some conversion of divanillin,
it is assumed that the main reaction in the latter period of the
reaction is the HER leading to no further divanillin reduction.
At the moment this phenomenon cannot be explained and is
subject of further studies. Besides undergoing pinacolization
the aldehyde end groups can be reduced by a 2eꢀ reduction to
the corresponding alcohol terminating the polymer chain
growth (see Fig. 10, red motif). As shown in the electrochemical
vanillin reduction experiments and by Jow et al. the ratio
between alcohol formation and pinacolization increases with
decreasing potential respectively under more severe conditions
in favor of alcohol formation.18 The cross-coupling signal of the
primary, terminal alcohol group was assigned to the peak at dH
¼ 5.0 ppm/dC z 64.9 ppm. For the polyvanillin samples
3.6 NMR analyses of polyvanillin
To gather more insight in structural features of the synthesized
polyvanillin 2D-NMR (HSQC, 13C/1H) spectra of selected
sampled were recorded combining the sensitivity of 1H-NMR
with the higher resolution of 13C-NMR (Fig. 9, 10 and Table
4).51 The polyvanillin samples were fully soluble in pyridine-d5
except for the sample synthesized at a Zn cathode at a current
density of 30 mA cmꢀ2 and a low applied charge of 2 F molꢀ1
(Fig. 9A), which partly precipitated and for the sample synthe-
sized at a GC cathode at a current density of 60 mA cmꢀ2 and an
applied charge of 8 F molꢀ1 (Fig. 9F), which mostly precipitated.
In both precipitated cases low molecular weights were observed
in the SEC analysis.
The existents of structural features of polyvanillin were
conrmed in all cases. Cross-coupling signals of the methoxy
groups coupled to the aromatic rings at dH ¼ 3.7 ppm/dC
¼
55.60 ppm were found in every sample. The vicinal hydroxy
groups in the aliphatic region generated by the pinacol forma-
tion (see Fig. 10, brown motif) were assigned to the cross-
coupling signals at dH ¼ 5.0 ppm/dC ¼ 79.5 ppm and dH ¼ 5.4
ppm/dC ¼ 78.9 ppm and cross-coupling signals from the
aromatic ring of the C5-substituted guaiacol to dH ¼ 7.7 ppm/dC
¼ 109.0 ppm, dH ¼ 7.7 ppm/dC ¼ 123.6 ppm and dH ¼ 7.9 ppm/
dC ¼ 130.0 ppm (Table 4). Very broad peaks of the methoxy
groups signals in the 1H-spectra conrm the high molecular
weight of the polyvanillin samples (Fig. 9B–E). In the two
samples with lower molecular weight (Fig. 9A and F) 1H-signals
Table 4 Identified cross-coupling signals in the 2D-NMR (HSQC, 13C/1H) spectra of polyvanillin samples, their assessment and some additional
notes
1
Cross-coupling signals d 13C/d H in ppm
Assessment
Note
z55.60/3.7
64.9/5.0
(CH3–O-Ar)
R–CH2-OH
Methoxyl group
Linked to primary aliphatic OH-group,
sometimes weak signal
79.5/5.0 & 78.9/5.4
109.0/7.7
Ca0–Ha0, Ca00–Ha00
C2–H2
Vicinal OH groups in aliphatic region
C5-substituted guaiacyl, cross-coupling d 13C/
1
d H aromatic ring
123.6/7.7
126.6–125.6/6.7
C60–H60
Cross-coupling d 13C/d 1H aromatic ring
Stilbene like double bond instead of vicinal OH
groups
Ca0–Ha0 & Ca00–Ha00
130.0/7.9
C6–H6
C5-substituted guaiacyl, cross-coupling d 13C/
1
d H aromatic ring
191.4/10.1
Ca–Ha
Aldehyde end group, sometimes weak signal
Pyridine-d5, 99%ical
123.87/7.220; 135.91/7.580 & 150.35/8.740
111.5/7.5; 116.4/7.2; 121.5/7.3
C–H cross-coupling in pyridine
Unknown
Could be cross-coupling signals of 13C/1H in
aromatic rings, differing in chemical
environment
8982 | RSC Adv., 2021, 11, 8970–8985
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
Table 5 Selected results of 31P-NMR-analyses on different polyvanillin samples, after in situ labelling with 2-chloro-4,4,5,5-tetramethyl-1,3,2-
dioxaphospholane (Cl-TMDP) in solvent pyridine/CDCl3 ¼ 1.6/1 (v/v), ISTD: cyclohexanol, Bruker Avance 300 MHz
Current density
Applied charge
Phenolic OH
Aliphatic OH
Ratio aliph.
OH/phen. OH
Cathode material
(mA cmꢀ2
)
(F molꢀ1
)
(mmol gꢀ1
)
(mmol gꢀ1
)
Zn
Zn
Zn
GC
30
30
60
60
2
8
8
8
6.630
6.236
8.775
8.234
3.401
2.730
4.362
0.513
0.513
0.438
0.497
0.067
synthesized at Zn cathodes the amount of terminal alcohol yet, further studies will get to the bottom of this preliminary
groups slightly increases with increasing current density result. However, the low observed degree of polymerization in
respectively decreasing cathode potential and increasing the SEC analysis for the polyvanillin sample synthesized at GC
applied charge (Fig. 9A–D). At GC comparably, a higher amount at a current density of 60 mA cmꢀ2 is conrmed by the 31P-NMR
of these terminal alcohol groups is already observed at 15 mA results showing almost no generated aliphatic hydroxy groups
cmꢀ2 due to more negative potentials. In the sample synthe- by pinacolization.
sized at a current density of 60 mA cmꢀ2 the absolute peak
Finally, it should be noted that in comparison higher weight-
height of the terminal alcohol groups is lower than at a current average molecular weights MW up to 17 000 g molꢀ1 of poly-
density of 15 mA cmꢀ2. However, in this partly polymerized vanillin synthesized at a Pb cathode with a divanillin concen-
sample the 1H-NMR signal of the primary, terminal alcohol tration of 0.175 M were reported in the preliminary feasibility
group is still relatively high compared to the 1H-NMR signal of study by Amarasekara et al., which could not be reproduced.23
the pinacol groups. Moreover, an increasing signal at dH ¼ 6.7 However, exact current densities, SEC calibration standards and
ppm/dC z 126.0 ppm is observed with decreasing cathode columns were not stated and DMF was used as solvent for SEC
potential respectively under more severe conditions. This signal analysis. Therefore, the SEC molecular weight data of poly-
was assigned to stilbene-like double bond systems, since the vanillin is hardly comparable. However, the published 1H-NMR
signal is located in the unsaturated and aliphatic region of the spectra of the polyvanillin sample showed unreacted aldehyde
1H/13C-NMR spectra (see Fig. 10, blue motif). Unfortunately, it end groups and a sharp methoxy group peak indicating not full
cannot be distinguished whether these stilbene-like double conversion of divanillin and existence of a high amount of a very
bond systems are created during the isolation process by low molecular fraction. Therefore, relatively higher molecular
titrating with concentrated HCl or in the actual electrolysis. weights should have been achieved in this presented study.
However, a clear trend of increasing double bond systems with
harsher electrolysis conditions is observed favoring the case of
stilbene formation in the electrolysis. It is suspected that the
generated vicinal alcohol groups in the aliphatic region gener-
4 Conclusions
In conclusion, vanillin was successfully exploited for the elec-
trochemical production of biobased polymers (polyvanillin) and
polymer building blocks (hydrovanilloin). The production of
these compounds was realized in an H-type cell setup by
reductive coupling via pinacolization of vanillin and divanillin
to hydrovanilloin and polyvanillin, respectively, and depen-
dencies of cathode material, current density and applied charge
on their formation were systematically investigated. Selected
viable cathode materials for the electrochemical coupling steps
exhibiting high HER overpotentials (Pb, GC and Zn) were
screened by LSV for their activity. The screening was followed by
bulk electrolysis experiments of vanillin gaining insights into
conversion, product distributions and FEs under various reac-
tion conditions. Thereby, Zn was revealed as stable, sustainable
and excellent replacement for the commonly used toxic cathode
material Pb enhancing the selectivity for hydrovanilloin
formation to almost 100% up to high current densities of 100
mA cmꢀ2 at reasonable FEs of 55–85%. Further, impact
parameters on structural features of polyvanillin, generated in
the divanillin bulk electrolysis experiments, were examined by
SEC, 2D-NMR (HSQC, 13C/1H) and 31P-NMR aer in situ labeling
with Cl-TMDP. SEC analyses revealed weight-average molecular
ated by pinacolization are converted to these stilbene-like
double bond systems by elimination of water followed by
a further reduction to the corresponding double bond. Similar
hetero-catalyzed reductions have been shown for the conversion
of acetylated hydrovanilloin, which is the pinacol product of
vanillin, to the stilbene and stilbene epoxides to alkenes.44,52,53
Furthermore, 31P-NMR analyses of selected polyvanillin
samples were conducted for quantication of phenolic and
aliphatic OH groups. In theory, the polyvanillin molecule
should be characterized by having more or less the same molar
number of phenolic and aliphatic OH groups. By means of 31P-
NMR analysis, aer in situ labeling Cl-TMDP, this should be
illustratable. An exemplary 31P-NMR spectrum is shown in the
(ESI, Fig. S12†) and quantitative calculated results are summa-
rized in Table 5. Contrary to theoretical considerations, the
polyvanillin samples prepared in this study oen exhibit
a molar ratio of aliphatic to phenolic OH groups of approxi-
mately 0.5. This nding can of course be discussed speculatively
with respect to a steric hindrance of the in situ labeling process
or an uncontrolled water separation during the isolation from
the electrolyte. As no conclusive answer could be given for this
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 8970–8985 | 8983

View Article Online
RSC Advances
Paper
weights MW up to 3400 g molꢀ1 (vs. pullulan standard) at low
polydispersities. At Zn cathodes the molecular weights and their
distributions of polyvanillin were observed to be independent
on current density (15–60 mA cmꢀ2) and divanillin concentra-
tion (0.1–0.3 M), whereas for Pb and GC cathodes molecular
weights decreased at higher current densities. Structural
features such as unreacted aldehyde groups, pinacol groups,
terminal alcohol groups and stilbene-like double bound
systems were identied by NMR and their dependency on
reaction conditions discussed.
4 R. Chinthapalli, P. Skoczinski, M. Carus, W. Baltus, D. de
¨
Guzman, H. Kab, A. Raschka and J. Ravenstijn, Ind.
Biotechnol., 2019, 15(4), 237–241.
5 Z. Sun, B. Fridrich, A. de Santi, S. Elangovan and K. Barta,
Chem. Rev., 2018, 118, 614.
6 C. Heitner, D. Dimmel and J. Schmidt, Lignin and Lignans,
CRC Press, 2016.
7 R. Rinaldi, R. Jastrzebski, M. T. Clough, J. Ralph,
M. Kennema, P. C. A. Bruijnincx and B. M. Weckhuysen,
Angew. Chem., Int. Ed., 2016, 55, 8164.
For further scalability of these promising electrochemical
processes a transfer to an electrochemical ow reactor is
required, as batch cells are limited in production rate due to low
electrode area to electrolyte ratios and may suffer from poor
8 H.-R. Bjørsvik and F. Minisci, Org. Process Res. Dev., 1999, 3,
330.
9 M. Zirbes and S. R. Waldvogel, Curr. Opin. Green Sustain.
Chem., 2018, 14, 19.
undened hydrodynamics.54 Thus, the transfer to an electro- 10 D. Schmitt, C. Regenbrecht, M. Hartmer, F. Stecker and
chemical ow reactor of these processes including a detailed
S. R. Waldvogel, Beilstein J. Org. Chem., 2015, 11, 473.
investigation of the mass transport behavior will be addressed 11 C. Z. Smith, J. H. P. Utley and J. K. Hammond, J. Appl.
in future investigations overcoming the mentioned limitations
of batch cells.
Electrochem., 2011, 41, 363.
12 M. Fache, B. Boutevin and S. Caillol, Eur. Polym. J., 2015, 68,
488.
13 B. A. Frontana-Uribe, R. D. Little, J. G. Ibanez, A. Palma and
R. Vasquez-Medrano, Green Chem., 2010, 12, 2099.
Author contributions
¨
14 A. Wiebe, T. Gieshoff, S. Mohle, E. Rodrigo, M. Zirbes and
Conceptualization, R. Kunkel, V. M. Schmidt; data curation, R.
Kunkel and D. Schmiedl; formal analysis, R. Kunkel and D.
Schmiedl; funding acquisition, D. Schmiedl, C. Cremers and J.
S. R. Waldvogel, Angew. Chem., Int. Ed., 2018, 57, 5594.
15 C. Schotten, T. P. Nicholls, R. A. Bourne, N. Kapur,
B. N. Nguyen and C. E. Willans, Green Chem., 2020, 22, 3358.
16 D. Pletcher, Electrochem. Commun., 2018, 88, 1.
17 I. A. Pearl, J. Am. Chem. Soc., 1952, 74, 4260.
18 J.-J. Jow and T. C. Choi, Electrochim. Acta, 1987, 32, 311.
19 R. M. Powers and R. A. Day Jr, J. Am. Chem. Soc., 1958, 80,
808.
¨
¨
Tubke; investigation, R. Kunkel, D. Muller and D. Schmiedl;
methodology, R. Kunkel and D. Schmiedl; project administra-
tion, C. Cremers, D. Schmiedl and J. Tubke; supervision, V. M.
¨
¨
Schmidt and J. Tubke; visualization, R. Kunkel; writing – orig-
inal dra, R. Kunkel; writing – review & editing, R. Kunkel, V. M.
Schmidt, C. Cremers and D. Schmiedl.
´
´
20 P. Zuman, D. Barnes and A. Ryvolova-Kejharova, Discuss.
Faraday Soc., 1968, 45, 202.
21 A. S. Amarasekara and M. A. Hasan, Polym. Sci. B, 2016, 58,
307.
Conflicts of interest
There are no conicts to declare.
22 A. S. Amarasekara, R. Garcia-Obergon and A. K. Thompson,
J. Appl. Polym. Sci., 2019, 136, 47000.
23 A. S. Amarasekara, B. Wiredu and A. Razzaq, Green Chem.,
2012, 14, 2395.
24 H. Yamamoto, T. Hoshino and T. Uchiyama, Biosci.
Biotechnol. Biochem., 1999, 63, 390.
25 F. Tiemann, Ber. Dtsch. Chem. Ges., 1885, 18, 3493.
26 K. Elbs and H. Lerch, J. Prakt. Chem., 1916, 93, 1.
27 D. R. Kelly, S. C. Baker, D. S. King, D. S. de Silva, G. Lord and
J. P. Taylor, Org. Biomol. Chem., 2008, 6, 787.
Acknowledgements
This work was performed within the scope of the joint project
“IntEleK-to” nancially supported by the Federal Ministry of
Food and Agriculture (Germany) via the Fachagentur Nach-
wachsende Rohstoffe (FNR), funding code 22409617. Further-
more, R. Kunkel acknowledges generous support by the
bioeconomy graduate program BBW ForWerts, supported by
the Ministry of Science, Research and the Arts of Baden-
`
28 M. Delomenede, F. Bedos-Belval, H. Duran, C. Vindis,
`
M. Baltas and A. Negre-Salvayre, J. Med. Chem., 2008, 51,
¨
Wurttemberg (MWK) and thanks the Fonds der Chemischen
3171.
´
Industrie (FCI) for a Kekule fellowship.
29 A. Llevot, E. Grau, S. Carlotti, S. Grelier and H. Cramail,
Polym. Chem., 2015, 6, 6058.
30 S. Antoniotti, L. Santhanam, D. Ahuja, M. G. Hogg and
J. S. Dordick, Org. Lett., 2004, 6, 1975.
References
¨
¨
¨
1 M. D. Karkas, B. S. Matsuura, T. M. Monos, G. Magallanes 31 E. Anklam, S. Gaglione and A. Muller, Food Chem., 1997, 60,
and C. R. J. Stephenson, Org. Biomol. Chem., 2016, 14, 1853.
2 T. Iwata, Angew. Chem., 2015, 127, 3254.
43.
32 W. R. Russell, L. Scobbie and A. Chesson, Bioorg. Med.
Chem., 2005, 13, 2537.
´
3 S. Molina-Gutierrez, V. Ladmiral, R. Bongiovanni, S. Caillol
and P. Lacroix-Desmazes, Green Chem., 2019, 21, 36.
8984 | RSC Adv., 2021, 11, 8970–8985
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Paper
RSC Advances
33 L. M. M. Mouterde, A. L. Flourat, M. M. M. Cannet, 43 T. R. Ralph, M. L. Hitchman, J. P. Millington and
P.-H. Ducrot and F. Allais, Eur. J. Org. Chem., 2013, 173. F. C. Walsh, J. Electroanal. Chem., 1999, 462, 97.
34 A. S. Amarasekara and A. Razzaq, ISRN Polymer Sci., 2012, 44 B. G. Harvey, A. J. Guenthner, H. A. Meylemans,
2012, 1.
S. R. L. Haines, K. R. Lamison, T. J. Groshens,
L. R. Cambrea, M. C. Davis and W. W. Lai, Green Chem.,
2015, 17, 1249.
35 A. Roessler, D. Crettenand, O. Dossenbach, W. Marte and
P. Rys, Electrochim. Acta, 2002, 47, 1989.
¨
36 C. Gutz, V. Grimaudo, M. Holtkamp, M. Hartmer, J. Werra, 45 W. J. Bover, D. Johnson, M. S. Baymak and P. Zuman, Indian
L. Frensemeier, A. Kehl, U. Karst, P. Broekmann and
S. R. Waldvogel, ChemElectroChem, 2018, 5, 247.
J. Chem., Sect. A: Inorg., Bio-inorg., Phys., Theor. Anal. Chem.,
2003, 42, 744.
37 M. R. Gennero de Chialvo and A. C. Chialvo, Electrochim. 46 P. Zuman, Electroanalysis, 2006, 18, 131.
Acta, 1988, 33, 825.
47 H. Tobita, J. Polym. Sci., Part A: Polym. Chem., 2002, 40, 3426.
38 S. Baumberger, A. Abaecherli, M. Fasching, G. Gellerstedt, 48 M. Vielhauer, S. Bodendorfer, P. J. Lutz, C. Friedrich and
¨
R. Gosselink, B. Hortling, J. Li, B. Saake and E. de Jong,
Holzforschung, 2007, 61, 459.
39 A. Granata and D. S. Argyropoulos, J. Agric. Food Chem., 1995,
43, 1538.
R. Mulhaupt, Macromolecules, 2013, 46, 8129.
49 V. M. Schmidt and J. Heitbaum, Electrochim. Acta, 1993, 38,
349.
50 C. Li, H. Bai and G. Shi, Chem. Soc. Rev., 2009, 38, 2397.
40 Y. Pu, S. Cao and A. J. Ragauskas, Energy Environ. Sci., 2011, 51 M. Sette, R. Wechselberger and C. Crestini, Chem.–Eur. J.,
4, 3154. 2011, 17, 9529.
41 R. T. Nishimura, C. H. Giammanco and D. A. Vosburg, J. 52 B. He, H. Tian, Y. Geng, F. Wang and K. Mullen, Org. Lett.,
Chem. Educ., 2010, 87, 526. 2008, 10, 773.
42 M. Chandrasekaran, M. Noel and V. Krishnan, J. Appl. 53 J.-M. Huang, Z.-Q. Lin and D.-S. Chen, Org. Lett., 2012, 14, 22.
¨
Electrochem., 1992, 22, 1072.
54 D. Pletcher, R. A. Green and R. C. D. Brown, Chem. Rev., 2018,
118, 4573.
© 2021 The Author(s). Published by the Royal Society of Chemistry
RSC Adv., 2021, 11, 8970–8985 | 8985