Electrochemical oxidation of 5-hydroxymethylfurfural to 2,5-furandicarboxylic acid (Fdca) in acidic media enabling spontaneous fdca separation

Article abstract of DOI:10.1002/cssc.201800532

2,5-Furandicarboxylic acid (FDCA) has become an increasingly desirable platform chemical to replace terephthalic acid in the production of a variety of polymeric materials, including polyethylene terephthalate. FDCA can be produced by the oxidation of 5-hydroxymethylfurfural (HMF), which can be derived from cellulosic biomass. Oxidation of HMF to FDCA is typically performed under basic conditions. Separation of FDCA is most easily accomplished by lowering the pH until FDCA is insoluble and filtering it from solution. In a large-scale process, this would lead to a high operating cost to purchase the required acid and base and to dispose of the resulting salt waste. In this study, electrochemical oxidation of HMF was carried out in acidic media by using a manganese oxide (MnO_x) anode to remove the need to vary the pH to separate FDCA. The MnO_x anode afforded a FDCA yield of 53.8 % in a pH 1 H₂SO₄ solution, in which FDCA precipitation occurred spontaneously from the same reaction solution without altering the pH or other aspects of the solution composition. Electrochemical oxidation in acidic media offers a new pathway to convert HMF into maleic acid, which is another desirable biomass-derived platform molecule. The performance of the MnO_x anode was investigated in comparison with that of a Pt anode to identify unique electro-catalytic properties of the MnO_x anode for HMF oxidation.

Full text of DOI:10.1002/cssc.201800532

DOI: 10.1002/cssc.201800532
Full Papers
Electrochemical Oxidation of 5-Hydroxymethylfurfural to
2,5-Furandicarboxylic Acid (FDCA) in Acidic Media
Enabling Spontaneous FDCA Separation
Stephen R. Kubota and Kyoung-Shin Choi*[a]
2,5-Furandicarboxylic acid (FDCA) has become an increasingly
desirable platform chemical to replace terephthalic acid in the
production of a variety of polymeric materials, including poly-
ethylene terephthalate. FDCA can be produced by the oxida-
tion of 5-hydroxymethylfurfural (HMF), which can be derived
from cellulosic biomass. Oxidation of HMF to FDCA is typically
performed under basic conditions. Separation of FDCA is most
easily accomplished by lowering the pH until FDCA is insoluble
and filtering it from solution. In a large-scale process, this
would lead to a high operating cost to purchase the required
acid and base and to dispose of the resulting salt waste. In this
study, electrochemical oxidation of HMF was carried out in
acidic media by using a manganese oxide (MnOx) anode to
remove the need to vary the pH to separate FDCA. The MnOx
anode afforded a FDCA yield of 53.8% in a pH 1 H2SO4 solu-
tion, in which FDCA precipitation occurred spontaneously from
the same reaction solution without altering the pH or other as-
pects of the solution composition. Electrochemical oxidation in
acidic media offers a new pathway to convert HMF into maleic
acid, which is another desirable biomass-derived platform mol-
ecule. The performance of the MnOx anode was investigated in
comparison with that of a Pt anode to identify unique electro-
catalytic properties of the MnOx anode for HMF oxidation.
Introduction
Biomass is a promising sustainable material source for the
manufacture of key building block chemicals as well as fuels to
reduce or eliminate our dependence on fossil fuels.[1–4] 5-Hy-
droxymethylfurfural (HMF), which can be derived from cellulo-
sic biomass, has generated a considerable amount of attention
as a platform molecule to synthesize industrially and commer-
cially desirable products. For example, 2,5-furandicarboxylic
acid (FDCA), one of the oxidation products of HMF, can serve
as a monomer in a variety of polymeric materials. In particular,
it is known to be an excellent replacement for terephthalic
acid in many polyesters such as polyethylene terephthalate
(PET).[5–7] HMF can be oxidized to FDCA through two possible
pathways (Scheme 1). The alcohol can be first oxidized to an
aldehyde resulting in 2,5-diformylfuran (DFF) or the aldehyde
can be oxidized first to form 5-hydroxymethyl-2-furancarboxyl-
ic acid (HMFCA). DFF or HMFCA is then subsequently oxidized
to 5-formyl-2-furancarboxylic acid (FFCA) and then finally to
FDCA.
Scheme 1. Possible HMF oxidation pathways to produce FDCA.
metric chemical oxidants as well as removing the need for the
use of high pressures of O2 as an oxidant. Furthermore, in an
electrochemical system, the oxidation half reaction of HMF to
FDCA can be coupled with a value-added reduction half reac-
tion, such as water to H2, increasing the overall profitability
and utility of the system.[11]
Recent studies have shown that efficient and selective elec-
trochemical oxidation of HMF to FDCA is indeed possible.
These studies utilized homogeneous reactions, in which HMF
oxidation is mediated by an electrochemically active homoge-
neous catalyst,[11] or heterogeneous reactions, in which HMF is
oxidized directly on the surface of catalytic anodes.[12–15] Com-
monly, all of these methods are performed under alkaline con-
ditions. Even non-electrochemical methods to oxidize HMF to
FDCA generally rely on basic pH conditions.[9] However, FDCA
is most easily isolated from basic aqueous solutions by lower-
ing the pH to <2–3, in which FDCA is no longer soluble; the
resulting FDCA precipitate can then be filtered from the solu-
tion. In an industrial application, raising and lowering the pH
before and after every reaction would generate a tremendous
amount of salt as waste, in addition to the cost associated
Traditional HMF oxidation to FDCA typically relies on chemi-
cal oxidants or noble metal catalysts in high temperature and
O2 pressure reactions.[7–10] Aqueous electrochemical oxidation
is a very promising alternative because reactions are driven by
an applied potential, removing the need for wasteful stoichio-
[a] S. R. Kubota, Prof. K.-S. Choi
Department of Chemistry, University of Wisconsin-Madison
1101 University Ave., Madison, WI 53706 (USA)
E-mail: kschoi@chem.wisc.edu
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under:
https://doi.org/10.1002/cssc.201800532.
ChemSusChem 2018, 11, 1 – 9
1
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
with purchasing the necessary acid/base and removing excess
salt from solution.
Some non-electrochemical studies have successfully oxidized
HMF to FDCA under near-neutral conditions.[16–18] Neutral con-
ditions would reduce the amount of acid needed to precipitate
FDCA but would not completely eliminate the need to change
the pH. Furthermore, the scaled-up oxidation processes per-
formed under near-neutral conditions would require the use of
a buffer to maintain the pH. Otherwise, protons generated
during FDCA production would rapidly decrease the pH from
the neutral region, which can affect the performance and sta-
bility of the catalysts if they operate optimally under near-neu-
tral conditions or dissolve under acidic conditions. Therefore,
the development of catalysts and methods that can oxidize
HMF to FDCA under strongly acidic conditions, in which FDCA
can be readily isolated from the reaction solution without re-
quiring extra procedures, will be a breakthrough for FDCA pro-
duction. Furthermore, considering that HMF can be obtained
from cellulosic biomass through acid catalyzed dehydration re-
actions,[1–4] developing a process that can further convert HMF
into FDCA under similar acidic conditions would streamline the
entire conversion process from biomass into FDCA.
Figure 1. SEM images of a MnOx film as deposited (a) and after annealing at
4008C for 2 h (b).
peak separation of the two Mn3s peaks can be used as a relia-
ble measure to evaluate the oxidation state of Mn.[20–24] The
separation of the two Mn3s peaks increases as the oxidation
state of Mn decreases. For example, typical literature-reported
peak separation for MnO2 is 4.5–4.7 eV, whereas that for Mn2O3
is 5.2–5.4 eV.[20–23] Figure 2 shows the XPS of our MnOx films as
well as the reference samples, in which the positions of the
Mn3s main peaks are aligned to be at the same energy so that
the peak separations of all samples can be easily compared.
(The raw XPS results can be found in Figure S1). The peak sep-
arations for purchased MnO2 and Mn2O3 reference compounds
were 4.37 eV and 5.23 eV, respectively, which agreed with
those reported in the literature. The peak separation of the as-
deposited MnOx film was estimated to be 4.41 eV, suggesting
that the as-deposited film contained mostly Mn4+. The peak
separation of the annealed MnOx film was 4.56 eV, suggesting
that a non-negligible amount of Mn3+ ions were formed
during annealing. The decomposition of MnO2 to Mn2O3 by
heat treatment is a well-known phenomenon with complete
The materials that can be used as a catalytic anode for elec-
trochemical HMF oxidation in acidic media are those that are
stable in acid media under anodic bias and also prefer HMF ox-
idation to water oxidation to O2. In this study, we report elec-
trochemical oxidation of HMF to FDCA in a strongly acidic so-
lution (pH 1) using manganese oxide (MnOx) as a catalytic
anode, which satisfies the aforementioned conditions. To our
knowledge, this is the first demonstration of the production of
FDCA in acidic media as the major product, which enables
spontaneous FDCA precipitation from the same reaction solu-
tion.
Results and Discussion
The MnOx anodes used in this study were prepared by electro-
deposition from an aqueous solution containing MnSO4 as the
Mn2+ source. During anodic deposition, Mn2+ ions in the plat-
ing solution are oxidized to Mn4+ ions, which are no longer
soluble and precipitate as a MnOx film on the working elec-
trode (WE).[19] The as-deposited films were annealed at 4008C
for 2 h to increase their chemical stability as well as their adhe-
sion to the conducting substrate. SEM images of the as-depos-
ited and annealed MnOx films showing their surface morpholo-
gies are displayed in Figure 1. Both the annealed film and as-
deposited film were X-ray amorphous.
The oxidation states of Mn in MnOx films were examined by
comparing Mn3s peaks of X-ray photoelectron spectroscopy
(XPS), which appear as a doublet owing to the exchange of
electrons between the 3s and 3d levels.[20] Although the abso-
lute positions of the Mn3s peaks can be used to evaluate the
oxidation states of Mn, the absolute peak positions can shift
considerably from the known literature values when the sam-
ples have low conductivity such as our as-deposited hydrated
MnOx films.[21] When the absolute peak positions appear to be
affected by factors other than the oxidation state of Mn, the
Figure 2. XPS of the Mn3s peak splitting of as-deposited MnOx films, after
annealing, and after electrolysis compared to MnO2 and Mn2O3 standards.
The spectra binding energy is manually offset to clearly show the peak split-
ting.
&
ChemSusChem 2018, 11, 1 – 9
www.chemsuschem.org
2
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
conversion occurring at ꢀ5008C.[25–27] Our intention for anneal-
ing was not to convert MnO2 into Mn2O3 but to increase the
chemical stability and adhesion of the as-deposited film. We
identified that annealing at 4008C for 2 h was optimal to ach-
ieve our goal, but this condition also resulted in the formation
of a small amount of Mn3+ ions. We expected that the pres-
ence of a small amount of Mn3+ ions should not affect the cat-
alytic performance of MnOx as they would readily be oxidized
to Mn4+ under the anodic bias used for HMF oxidation.
Indeed, XPS of the MnOx film after HMF oxidation showed that
the peak separation decreased to 4.46 eV, suggesting that the
majority of Mn remained as Mn4+ ions once an anodic bias
was applied.
tion of water, which can be observed in the LSV obtained with-
out HMF or its oxidation intermediates. After the addition of
HMF, DFF, HMFCA, or FFCA, the onset potential for the oxida-
tion current commonly shifted to a less positive potential. This
indicated that the oxidation of HMF, DFF, HMFCA, and FFCA
was more favorable than water oxidation on the MnOx surface.
This also indicated that the oxidation of HMF to FDCA could
be achieved without inducing water oxidation by selecting an
oxidation potential that does not allow for water oxidation.
We performed the same LSV experiments using Pt as a con-
trol anode to compare the performance of MnOx. Previously,
Strasser and co-worker studied non-electrochemical HMF oxi-
dation on a carbon-supported Pt catalyst at 508C and 10 bar
O2 at pH 1, 6, 10, and 13.[13] They reported that after a 4 h reac-
tion time, HMF conversion and FDCA yield were 100% and
75%, respectively, at pH 13 and 52% and approximately 5%,
respectively, at pH 1. The major oxidation product of HMF at
pH 1 was DFF (35%). This result suggested that oxidation of
HMF to FDCA on Pt was not favorable under acidic conditions,
particularly owing to the difficulty in the oxidation of aldehyde
groups to carboxylic groups. This result made us wonder
whether similar results would be obtained with Pt for electro-
chemical oxidation of HMF.
Annealing at temperatures below 4008C was not sufficient
to achieve the desired chemical stability and adhesion of MnOx
films. Also, annealing temperatures above 4008C were not fa-
vorable because when bulk crystalline Mn2O3 was formed, it
was not as catalytically active for HMF as the MnOx formed at
4008C. Although the XPS results showed that the MnOx films
were mostly composed of Mn4+ ions, because the exact Mn3+
/Mn4+ ratios in the as-deposited MnOx films, after annealing,
and after HMF oxidation were slightly different and could be
accurately determined, we use the notation MnOx throughout
this study.
The LSVs of Pt measured under the same conditions used
for MnOx are shown in Figure 3b. The onset potential of Pt for
water oxidation was more positive than that of MnOx for water
oxidation, indicating that Pt was less catalytic for water oxida-
tion in a pH 1 solution. When HMF or its oxidation intermedi-
ates were added, a significant decrease in the current density
and a shift of the anodic current onset potential in the positive
direction were observed. In this case, it was not possible to de-
termine whether the observed oxidation current was owing to
water oxidation or the oxidation of HMF, DFF, HMFCA, or FFCA
solely based on LSVs. However, because the onset potentials
of HMF and its oxidation intermediates do not appear before
that of water oxidation suggested that Pt was not particularly
catalytic for the oxidation of HMF and its intermediates.
The catalytic activity of MnOx for HMF oxidation to FDCA
was first examined using linear sweep voltammetry (LSV) in a
H2SO4 (pH 1) solution with and without 20 mm HMF and also
with 20 mm HMF oxidation intermediates (DFF, HMFCA, and
FFCA) (Figure 3a). The major competing reaction with HMF ox-
idation under aqueous conditions is the electrochemical oxida-
To analyze the oxidation products and their yields, we per-
formed constant potential oxidation of HMF using a cell divid-
ed with a glass frit. The WE compartment (anolyte) contained
15 mL of a 20 mm HMF solution in H2SO4 (pH 1), whereas the
counter electrode (CE) compartment (catholyte) contained
15 mL of a H2SO4 solution (pH 1) solution. The anode, cathode,
and the overall reactions are summarized in Equations (1)–(3).
Anode reaction : HMF þ 2 H2O ! FDCA þ 6 Hþ þ 6 eꢁ
Cathode reaction : 6 Hþ þ 6 eꢁ ! 3 H2
ð1Þ
ð2Þ
ð3Þ
Overall reaction : HMF þ 2 H2O ! FDCA þ 3 H2
HMF oxidation was first performed with MnOx as the anode
at 1.6 V versus reversible hydrogen electrode (RHE; 1.34 V
versus Ag/AgCl) at 608C. An elevated temperature (608C) was
used to improve the HMF oxidation kinetics of MnOx and
shorten the time required for the completion of the reaction.
The effect of the temperature on the enhancement of the oxi-
dation kinetics of HMF and its oxidation intermediates was evi-
Figure 3. LSVs of a MnOx anode (a) and a Pt anode (b) obtained in a pH 1
H2SO4 solution without (black) and with 20 mm HMF (red), 20 mm DFF
(blue), 20 mm HMFCA (purple), and 20 mm FFCA (green) at room tempera-
ture at a scan rate of 5 mVsꢁ1
.
ChemSusChem 2018, 11, 1 – 9
www.chemsuschem.org
3
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
dent when the LSVs for the oxidation of HMF, DFF, HMFCA,
and FFCA at room temperature and 608C (Figure S2). Another
advantage of using the elevated temperature was that the sol-
ubility of FDCA was considerably increased compared to that
at room temperature. As a result, FDCA precipitation did not
occur during electrochemical oxidation until the reaction was
completed and the solution was cooled down to room tem-
perature. This was highly favorable as the precipitation of
FDCA on the electrode surface during electrolysis could hinder
electrochemical oxidation. By using an acidic medium, the sep-
aration of FDCA was enabled by altering the temperature in-
stead of altering the pH, which significantly simplified the sep-
aration process.
A very small amount of HMFCA was observed throughout
the reaction, whereas an accumulation in DFF concentration
was evident in the early stages of the electrolysis, indicating
that HMF oxidation followed the DFF pathway shown in
Scheme 1, in which the alcohol is the first group was oxidized.
This result was consistent with the LSV data (Figure 3a), which
showed that the HMF and HMFCA oxidation rates were high-
est, indicating that the oxidation of the alcohol groups was
faster on MnOx than the oxidation of the aldehyde groups.
Over the course of the reaction, the accumulation of DFF and
FFCA was observed, before the final production of FDCA as
the major product.
The stoichiometric amount of charge required to completely
convert 15 mL of 20 mm HMF solution into FDCA is 173.7 C.
However, because a significant amount of FFCA (a yield of
31.6%) was observed even at 200 C, the oxidation was contin-
ued to 250 C. The product analysis obtained at 250 C showed
that the HMF conversion was 99.9% and the FDCA yield was
53.8%, which corresponded to a FE of 33.8%. At this charge
passed, FFCA still remained in the solution at a yield of 12.6%.
However, passing more charge than 250 C did not increase the
FDCA yield because the side reactions discussed below were
dominant beyond 250 C.
The concentration changes of HMF and its oxidation prod-
ucts in the anolyte were monitored during the HMF oxidation
by using HPLC. The HMF conversion, yields of the oxidation
products, and faradaic efficiency (FE) for the production of
FDCA were calculated by using Equations (4)–(6) where F is the
Faraday constant, and the results are displayed in Figure 4a.
mol of HMF consumed
mol of initial HMF
ð4Þ
ð5Þ
ð6Þ
HMF Conversion ð%Þ ¼
ꢂ 100 %
mol of product formed
mol of initial HMF
The major byproduct observed in HMF oxidation with the
MnOx anode was maleic acid (21.9% at 250 C), which was ob-
tained through the reaction in Scheme 2a. Interestingly, to our
knowledge, maleic acid has not been reported as an oxidation
product of HMF in any previous electrochemical oxidations of
HMF. However, we have previously observed that a lower pH
during electrochemical HMF reduction reactions increases the
favorability for electrochemical ring-opening reactions.[28]
Product yield ð%Þ ¼
FE ð%Þ ¼
ꢂ 100 %
mol of FDCA formed
ꢂ 100 %
total charge passed=ðF ꢂ 6Þ
Scheme 2. a) Oxidation of HMF to maleic acid. b) Reduction of maleic acid
to succinic acid.
The formation of maleic acid was not favorable in terms of
the FDCA yield; however, maleic acid is also a desirable prod-
uct.[29] For example, it can serve as an intermediate to produce
succinic acid (Scheme 2b). Succinic acid, along with FDCA, is
one of the top value-added chemicals from biomass identified
by a recent US Department of Energy report.[30] Electrochemical
means to reduce maleic acid to succinic acid have been previ-
ously reported.[31–34] Therefore, our future studies will identify
factors that can suppress or enhance maleic acid formation so
that we can obtain both FDCA and maleic acid as pure prod-
ucts. The presence of maleic acid in the reaction mixture did
Figure 4. Conversion of HMF (%) and yield (%) of oxidation products ob-
tained by using MnOx anodes at 1.6 V versus RHE (a) and Pt anodes at 2.0 V
versus RHE (b) during the course of electrochemical oxidation of HMF at
608C in a pH 1 H2SO4 solution containing 20 mm HMF at various amounts of
charge passed.
&
ChemSusChem 2018, 11, 1 – 9
www.chemsuschem.org
4
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
not affect the facile separation of FDCA in acidic solution be-
cause maleic acid is very soluble.
for Pt because it showed approximately the same current den-
sity as that of MnOx at 1.6 V versus RHE (Figure 4b). (Pt did not
generate any appreciable amount of current at 1.6 V versus
RHE.) Although the conversion profile of HMF by Pt was similar
to that with MnOx, there was a clear difference in the product
yield. DFF was the major oxidation product of HMF that still
contains the furan ring, and the formation of HMFCA, FFCA,
and FDCA was negligible. For example, at 250 C, the yields of
FFCA and FDCA were 1.8% and 0.1%, respectively. This sug-
gested that although Pt can oxidize the alcohol group at-
tached to the furan ring to an aldehyde to form DFF, it was
not catalytic for the oxidation of the aldehyde groups attached
to the furan ring to form FDCA. Maleic acid production was
also observed with Pt (13.0%), further indicating that low-pH
conditions facilitated maleic acid formation.
Figure 5 shows the precipitation of FDCA in the reaction
mixture after 250 C was passed and the temperature was
cooled down to room temperature. For this reaction, because
the complete conversion of FFCA into FDCA was not achieved,
During HMF oxidation by a Pt anode, the sum of the prod-
uct yields was significantly lower than HMF conversion, indicat-
ing the formation of undetectable side-products (Table 1). This
difference, referred to as “HMF missing”, continuously in-
creased during the course of oxidation. For example, the HMF
missing at 25 C and 250 C was 28.6% and 79.8%, respectively.
We also noticed that during oxidation by a Pt anode, the color
of the reaction solution changed to yellow, which was not ob-
served with MnOx, although we were unable to identify any
side-products that could be responsible for the color change
by HPLC and NMR. The decomposition of HMF into humins
under acidic conditions through cross-polymerization reactions
is a known phenomenon,[35–37] and the presence of humins in
solution is reported to cause color changes.[35] Humins are gen-
erally only characterized if they are produced as solids in suffi-
cient quantities by solid state techniques or after the solids are
redissolved in other solvents.[36–39] Therefore, the majority of
the missing HMF during the oxidation with Pt could be owing
to humin formation.
Figure 5. Photographs of the anode compartment of a split cell after the
electrochemical conversion of 20 mm HMF at 608C (left) and after the tem-
perature was cooled down to room temperature (right, solution agitated to
clearly show precipitate).
the precipitate contained FFCA as well as FDCA. However, this
demonstration indicated that if the complete conversion of
FFCA into FDCA can be achieved by further optimizations, the
use of a gentle temperature change (between 608C and room
temperature) to control the solubility of FDCA in a pH 1 solu-
tion would offer the most facile and cost-effective way to
couple FDCA production and FDCA separation.
The SEM images of the MnOx anode were re-examined after
HMF oxidation and no changes in the surface morphology and
no visual indications of film loss were observed (Figure S3),
confirming the stability of the MnOx anode used in this study
in a pH 1 solution under anodic bias.
Constant potential oxidation of HMF was also performed
using a Pt anode. A potential of 2.0 V versus RHE was chosen
The HMF conversion and product yields by the MnOx anode
summarized in Table 2 show that HMF missing also occurred
Table 1. Results of HMF oxidation by Pt at 608C and 2.0 V versus RHE.
Charge passed [C] HMF conv. [%] DFF yield [%] HMFCA yield [%] FFCA yield [%] FDCA yield [%] Maleic acid yield [%] HMF missing yield [%]
25
50
100
150
200
250
32.3
57.7
88.3
>99.9
>99.9
>99.9
3.4
14.3
13.1
16.3
9.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.4
0.8
1.5
1.8
0.0
0.0
0.0
0.0
0.0
0.1
0.2
0.0
9.9
6.8
7.1
28.6
43.3
64.8
76.0
81.7
79.8
5.3
13.0
Table 2. Results of HMF oxidation by MnOx at 608C and 2.0 V vs. RHE.
Charge passed [C] HMF conv. [%] DFF yield [%] HMFCA yield [%] FFCA yield [%] FDCA yield [%] Maleic acid yield [%] HMF missing yield [%]
25
50
100
150
200
250
33.5
60.9
95.8
>99.9
>99.9
>99.9
24.3
38.2
41.9
12.7
1.0
0.3
0.5
0.2
0.0
0.1
0.1
3.1
9.7
34.7
48.0
31.6
12.6
0.1
0.5
4.2
17.0
39.3
53.8
1.0
2.7
5.3
10.4
16.1
21.9
4.7
9.3
9.6
11.8
11.8
11.4
0.1
ChemSusChem 2018, 11, 1 – 9
www.chemsuschem.org
5
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
with the MnOx anode, but it was considerably lower and did
not exceed 12%. Because a very dilute HMF solution (20 mm)
was used in this study, we expect that 3–5% HMF missing can
easily occur owing to the adsorption of HMF and its products
to the various components of the electrochemical cell. Then,
the loss of HMF owing to the formation of humins or other un-
detectable side-products during the oxidation with the MnOx
anode can be considered to not be significant. Although the
formation of insoluble humins can affect the purity of precipi-
tated FDCA, acid-soluble humins that require neutralization to
precipitate should not affect FDCA precipitation when an
acidic solution is used for HMF oxidation.[39]
Conclusions
The electrochemical oxidation of HMF to FDCA at pH 1 was in-
vestigated by using MnOx as a catalytic anode to enable FDCA
precipitation in the same reaction solution without altering the
pH. The use of elevated temperature (608C) during oxidation
increased the oxidation kinetics and effectively suppressed
FDCA precipitation on the electrode surface during HMF oxida-
tion. Cooling the temperature of the reaction solution to RT
after HMF oxidation led to FDCA precipitation without addi-
tional processes. The MnOx anode could oxidize both the alco-
hol and the aldehyde groups of HMF to carboxylic acids and
form FDCA, whereas Pt, used as a control electrode, could oxi-
dize the alcohol groups, forming DFF but was unable to fur-
ther oxidize DFF to FDCA. Pt also promoted the formation of a
significant amount of undesirable humins, preventing high-
yield conversions of HMF into more valuable products. The
performance comparison of MnOx at 1.6 and 2.0 V versus RHE
suggested that electrochemical humin formation was facilitat-
ed at a more positive potential.
Comparing HMF missing with MnOx and Pt anodes, it ap-
pears that humin formation was particularly facilitated by Pt.
However, because the HMF oxidation performed with Pt used
a more positive potential than MnOx, we could not exclude the
possibility that the formation of humins observed with Pt
could be owing to the more positive potential applied to Pt
and not owing to the intrinsic nature of Pt. Therefore, we ex-
amined HMF oxidation by MnOx at 2.0 V versus RHE (Table S1).
Indeed, when the applied potential to MnOx was increased
from 1.6 V versus RHE to 2.0 V versus RHE, as with Pt, a yellow-
colored solution was observed after electrolysis indicating the
formation of humins, and the amount of missing HMF at 250 C
increased from 11.4% to 35.1%. This result clearly indicated
that electrochemical humin formation was facilitated at a more
positive potential. However, HMF missing observed with MnOx
at 2.0 V versus RHE was significantly lower than that observed
with Pt at the sample potential. Also, MnOx still produced DFF
(17.6%), FFCA (20.3%), and FDCA (5.9%) at 250 C at 2.0 V
versus RHE. Therefore, it is reasonable to conclude that Pt had
a stronger tendency for undesired oxidation of HMF (i.e.,
humin formation) while it was not capable of FDCA produc-
tion.
We also discovered that electrochemical oxidation of HMF in
acidic media offers a new pathway to form maleic acid, which
is a valuable building-block compound with industrial applica-
tions. Although the formation of maleic acid limited the maxi-
mum FDCA yield achievable in this study, further studies on
identifying factors that can enhance and suppress maleic acid
formation will allow us to optimize the reaction conditions to
maximize FDCA or maleic acid production. This study, which
shows the possibility of electrochemical production of FDCA in
acidic media and discusses the possible side reactions that
need to be suppressed for more efficient FDCA production,
will serve as a new foundation to facilitate further studies on
the production of FDCA in an acidic medium.
When 2.0 V versus RHE was applied to the MnOx anode,
water oxidation became a significant contributor to the cur-
rent, which was supported by the fact that even at 250 C of
charge passed, a small amount of HMF and a sizeable amount
of HMF oxidation intermediates (i.e., DFF, FFCA) remained
(Table S1). This result was expected from the LSV results shown
in Figure 3a, in which water oxidation became considerable on
MnOx at a more positive potential than 1.6 V versus RHE. Inter-
estingly, the yield for maleic acid was similar for the MnOx
anode when applying 1.6 V and 2.0 V versus RHE (21.9% and
19.4%, respectively), despite the lower overall conversion of
HMF and its oxidation intermediates related to FDCA produc-
tion as well as a significant decrease in FDCA yield obtained at
2.0 V versus RHE. This indicated that the conversion of HMF
into maleic acid was more favorable than the conversion of
HMF into FDCA at a more positive potential. This also suggest-
ed that a MnOx anode with a higher surface area, which can
generate sufficient currents for the completion of HMF oxida-
tion at a less positive potential than 1.6 V versus RHE, could
decrease the yield of maleic acid while minimizing humin for-
mation, allowing for improved FDCA production.
Experimental Section
Materials
MnSO4 monohydrate (ꢀ98%, Sigma–Aldrich), Na2SO4 (ACS Grade,
DOT Scientific Inc.), HMF (97%, Alfa Aesar), DFF (97%, TCI America),
HMFCA (98%, Oxchem), FFCA (>98%, TCI America), FDCA (>98%,
AstaTech Inc.), sodium maleate dibasic anhydrous (100.3%, Chem-
Impex Int’l Inc.), NaOH (ꢀ98%, Sigma–Aldrich), H2SO4 (95.0–98.0%,
Sigma–Aldrich), MnO2 (ꢀ99%, Sigma–Aldrich), and Mn2O3 (98%,
Aldrich) were all commercially purchased and used without further
purification.
Electrode preparation
Fluorine-doped tin oxide (FTO) glass was purchased from Hartford
Glass Company Inc. and cut into 1ꢁ2.5 cm or 2ꢁ2.5 cm pieces and
then sonicated sequentially in acetone, isopropyl alcohol, and
water for 15 min. Cu tape was attached to the top of the conduct-
ing side of the FTO to provide a contact for the potentiostat.
Teflon tape was also applied to the top of the electrode to prevent
direct contact between the Cu tape and the electrolyte, which
yielded a working area of either 2 cm2 (for linear sweep voltamme-
try) or 4 cm2 (for constant potential HMF oxidation).
&
ChemSusChem 2018, 11, 1 – 9
www.chemsuschem.org
6
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
MnOx was electrochemically deposited onto the FTO in an aqueous
solution composed of 50 mm MnSO4 and 100 mm Na2SO4. An undi-
vided three-electrode cell was used. FTO and Pt were used as the
WE and CE, respectively. A Ag/AgCl (4m KCl) electrode was used as
the reference electrode (RE). MnOx was anodically deposited by ap-
plying 0.9 V to the WE in the plating solution kept at 608C while
passing 0.5 Ccmꢁ2. The anodic bias oxidized Mn2+ ions in the plat-
ing solution to insoluble Mn4+ ions, which precipitated as a MnOx
film on the WE.[19] The as-deposited films were washed with deion-
ized (DI) water, dried in a stream of air, and then annealed in air at
samples from the WE compartment were diluted 8ꢁ, samples from
the CE were examined undiluted, and the pH of both were raised
using NaOH to ensure FDCA solubility for HPLC measurements.
10 mL sample aliquots were injected directly onto a 300ꢁ7.8 mm
ICSep ICE-Coregel 87H3 column (Transgenomic), and the column
oven temperature was maintained at 658C. An aqueous mobile
phase composed of 0.1% sulfuric acid was used with a flow rate of
0.5 mLminꢁ1 in isocratic mode. The identification of the products
and calculation of their concentrations were determined from stan-
dard solutions of known concentration. Retention times of 11, 20,
23, 27, 32, and 40 min correlated to maleic acid, FDCA, HMFCA,
FFCA, HMF, and DFF, respectively. Because each compound exhibits
a different light absorption profile, detection wavelengths of 200,
262, 258, 281, 283, and 288 nm were chosen as lmax for maleic
acid, FDCA, HMFCA, FFCA, HMF, and DFF, respectively.
4008C for 2 h, with a ramp rate of 28Cminꢁ1
.
Characterization
XRD was performed using a Bruker D8 Advanced PXRD, 298 K, Ni-
filtered CuKa radiation, l=1.5418 ꢂ. XPS spectra were obtained
using a Thermo Scientific Inc. K-Alpha XPS equipped with a mono-
chromated AlKa line as the X-ray source. The surface morphology
was examined by SEM using a LEO 1530 microscope with a 5 KV
accelerating voltage.
Acknowledgements
This work was supported by University of Wisconsin-Madison.
Electrochemical HMF oxidation
Conflict of interest
The catalytic activity of the electrodes for HMF oxidation was first
examined using LSV. LSVs were performed in pH 1 H2SO4 solutions
without and with 20 mm of HMF, DFF, HMFCA, or FFCA in an undi-
vided three-electrode cell, as described above, with MnOx or Pt as
the WE. The potential was swept in the positive direction from the
open circuit potential at a rate of 5 mVsꢁ1. Potentials were record-
ed versus Ag/AgCl RE and then converted into those versus RHE
by using Equation (7):
The authors declare no conflict of interest.
Keywords: biomass
conversion
·
electrocatalysis
·
manganese · polymers · separations
[1] B. R. Caes, R. E. Teixeira, K. G. Knapp, R. T. Raines, ACS Sustainable Chem.
Eng. 2015, 3, 2591–2605.
[2] G. W. Huber, J. A. Dumesic, Catal. Today 2006, 111, 119–132.
[3] A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411–2502.
[4] X. Tong, T. Ma, Y. Li, Appl. Catal. A 2010, 385, 1–13.
[5] J. J. Bozell, G. R. Petersen, Green Chem. 2010, 12, 539–554.
[6] A. J. J. E. Eerhart, A. P. C. Faaij, M. K. Patel, Energy Environ. Sci. 2012, 5,
6407–6422.
[7] L. Hu, G. Zhao, W. Hao, X. Tang, Y. Sun, L. Lin, S. Liu, RSC Adv. 2012, 2,
11184–11206.
[8] S. E. Davis, L. R. Houk, E. C. Tamargo, A. K. Datye, R. J. Davis, Catal. Today
2011, 160, 55–60.
[9] Z. Zhang, K. Deng, ACS Catal. 2015, 5, 6529–6544.
[10] A. A. Rosatella, S. P. Simeonov, R. F. M. Frade, C. A. M. Afonso, Green
Chem. 2011, 13, 754–793.
[11] H. G. Cha, K.-S. Choi, Nat. Chem. 2015, 7, 328–333.
[12] D. J. Chadderdon, L. Xin, J. Qi, Y. Qiu, P. Krishna, K. L. More, W. Li, Green
Chem. 2014, 16, 3778–3786.
[13] K. R. Vuyyuru, P. Strasser, Catal. Today 2012, 195, 144–154.
[14] N. Jiang, B. You, R. Boonstra, I. M. T. Rodriguez, Y. Sun, ACS Energy Lett.
2016, 1, 386–390.
[15] B. You, X. Liu, N. Jiang, Y. Sun, J. Am. Chem. Soc. 2016, 138, 13639–
13646.
E ðvs: RHEÞ ¼ E ðvs: Ag=AgClÞ þ EAg=AgCl ðref:Þ þ 0:0591 V ꢂ pH
ð7Þ
where EAg/AgCl (ref.)=0.1976 V vs. NHE at 258C. Constant-po-
tential oxidation of HMF was performed to determine the reac-
tion products, yields, and faradaic efficiency. Electrolyses were
performed in an H-shaped glass cell with two 20 mL compart-
ments connected by a 22 mm long, 13 mm diameter glass
tube. The solutions of the two compartments were separated
by a 3.5 mm thick fine glass frit. The MnOx or Pt WE and the
Ag/AgCl RE were immersed in the anodic compartment con-
taining 15 mL of a pH 1 H2SO4 solution with 20 mm HMF. The
Pt counter electrode was immersed in the cathodic compart-
ment with 15 mL of pH 1 H2SO4. The distance between the WE
and CE was approximately 6 cm. Electrolyses were performed
by applying a constant potential to the WE at 608C while stir-
ring with a magnetic stir bar at 1400 rpm.
[16] D. K. Mishra, H. J. Lee, J. Kim, H.-S. Lee, J. K. Cho, Y.-W. Suh, Y. Yi, Y. J.
Kim, Green Chem. 2017, 19, 1619–1623.
[17] T. Gao, T. Gao, W. Fang, Q. Cao, Mol. Catal. 2017, 439, 171–179.
[18] X. Han, L. Geng, Y. Guo, R. Jia, X. Liu, Y. Zhang, Y. Wang, Green Chem.
2016, 18, 1597–1604.
Product analysis
[19] M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solutions,
2nd ed., National Association of Corrosion Engineers, Houston, TX,
1974.
[20] A. J. Nelson, J. G. Reynolds, J. W. Roos, J. Vac. Sci. Technol. A 2000, 18,
1072–1076.
[21] M. Chigane, M. Ishikawa, J. Electrochem. Soc. 2000, 147, 2246–2251.
[22] V. R. Galakhov, M. Demeter, S. Bartkowski, M. Neumann, N. A. Ovechkina,
E. Z. Kurmaev, N. I. Lobachevskaya, Y. M. Mukovskii, J. Mitchell, D. L.
Ederer, Phys. Rev. B 2002, 65, 113102.
After constant-potential oxidation of HMF, products were identified
using Bruker Avance III 400 and 500 MHz NMR spectrometers. HMF
conversions and product yields were determined with a Shimadzu
Prominence-i LC-2030C 3D HPLC system containing a photodiode
array (PDA) detector. Over the course of the electrolysis, some HMF
and HMF oxidation products diffused from the WE compartment
to the CE compartment, so both solutions were examined, and the
results were combined to quantify each product. After electrolysis,
ChemSusChem 2018, 11, 1 – 9
www.chemsuschem.org
7
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[23] J. M. Cerrato, M. F. Hochella, W. R. Knocke, A. M. Dietrich, T. F. Cromer,
Environ. Sci. Technol. 2010, 44, 5881–5886.
[33] F. Zhao, F. Yan, Y. Qian, Y. Chu, C. Ma, Int. J. Electrochem. Sci. 2012, 7,
12931–12940.
[24] M. Huynh, C. Shi, S. J. L. Billinge, D. G. Nocera, J. Am. Chem. Soc. 2015,
137, 14887–14904.
[34] F. Wang, X. Yan, M. Xu, S. Li, W. Fang, Electrochim. Acta 2013, 97, 253–
258.
[25] M.-S. Wu, P.-C. J. Chiang, J.-T. Lee, J.-C. Lin, J. Phys. Chem. B 2005, 109,
23279–23284.
[26] Y. U. Jeong, A. Manthiram, J. Electrochem. Soc. 2002, 149, A1419–A1422.
[27] Y. Omomo, T. Sasaki, M. Watanabe, Solid State Ionics 2002, 151, 243–
250.
[35] B. F. M. Kuster, Starch/Staerke 1990, 42, 314–321.
[36] S. K. R. Patil, C. R. F. Lund, Energy Fuels 2011, 25, 4745–4755.
[37] I. Van Zandvoort, Y. Wang, C. B. Rasrendra, E. R. H. Van Eck, P. C. A. Bruij-
nincx, H. J. Heeres, B. M. Weckhuysen, ChemSusChem 2013, 6, 1745–
1758.
[28] J. J. Roylance, K.-S. Choi, Green Chem. 2016, 18, 2956–2960.
[29] R. Wojcieszak, F. Santarelli, S. Paul, F. Dumeignil, F. Cavani, R. V. Gon-
calves, Sustainable Chem. Process. 2015, 3, 1–11.
[30] T. Werpy, G. Petersen, A. Aden, J. Bozell, J. Holladay, J. White, A. Man-
heim, D. Elliot, L. Lasure, S. Jones, M. Gerber, K. Ibsen, L. Lumberg, S.
Kelley, Top Value-Added Chemicals From Biomass, Volume 1: Results of
Screening for Potential Candidates from Sugars and Synthesis Gas; U. S.
Department of Energy, Washington, DC, 2004.
[38] B. A. Fachri, R. M. Abdilla, H. H. Van de Bovenkamp, C. B. Rasrendra, H. J.
Heeres, ACS Sustainable Chem. Eng. 2015, 3, 3024–3034.
[39] I. V. Sumerskii, S. M. Krutov, M. Y. Zarubin, Russ. J. Appl. Chem. 2010, 83,
320–327.
Manuscript received: March 13, 2018
Revised manuscript received: May 14, 2018
Version of record online: && &&, 0000
[31] D. Vasudevan, J. Appl. Electrochem. 1995, 25, 176–178.
[32] A. V. Muzumdar, S. B. Sawant, V. G. Pangarkar, Org. Process Res. Dev.
2004, 8, 685–688.
&
ChemSusChem 2018, 11, 1 – 9
www.chemsuschem.org
8
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
FDCA approved: Electrochemical con-
version of 5-hydroxymethyfurfural,
which can be derived from biomass, to
2,5-furandicarboxylic acid (FDCA) and
subsequent spontaneous separation of
FDCA from solution was achieved in an
acidic media. The acidic media also pro-
vided a new route for the electrochemi-
cal production of maleic acid.
S. R. Kubota, K.-S. Choi*
&& – &&
Electrochemical Oxidation of 5-
Hydroxymethylfurfural to 2,5-
Furandicarboxylic Acid (FDCA) in
Acidic Media Enabling Spontaneous
FDCA Separation
ChemSusChem 2018, 11, 1 – 9
www.chemsuschem.org
9
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ