Green Chemistry

View Article Online

PAPER View Journal | View Issue

Biomass-derived binderless fibrous carbon

Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

Cite this: Green Chem., 2016, 18,

electrodes for ultrafast energy storage†
1506
R. Berenguer,a F. J. García-Mateos,a R. Ruiz-Rosas,b D. Cazorla-Amorós,b
E. Morallón,*c J. Rodríguez-Mirasol*a and T. Corderoa

Received 7th October 2015, The possibility of storing energy efficiently and sustainably at little cost is crucial to prevent climate
Accepted 12th October 2015
change and the exhaustion of natural resources. In this work we demonstrate that interconnected and
DOI: 10.1039/c5gc02409a porous carbon fibers easily obtained from lignin exhibit ultrafast charge–discharge and excellent energy
www.rsc.org/greenchem density and cyclability performances, to be used as binderless and flexible electrodes in supercapacitors.

Electrochemical capacitors, usually referred to as supercapaci- Activated carbons (ACs) derived from biomass and poly-
tors, are electrochemical devices for electrical energy storage mers have been identified as the currently most viable
and harvesting applications that can complement or replace materials for supercapacitors, from both economic and sus-
batteries when high power delivery or uptake and/or long tainability points of view.1a To attain the specifications for
cycling stability are required. Additionally, these devices widespread commercialization, various issues need to be
improve efficiencies in supply systems (such as internal com- solved, the most important being to produce ACs with a
bustion engines, renewable energy systems, batteries and fuel highly accessible surface area and a sufficiently high electrical
cells) by storing energy when in excess or not needed. Thus, conductivity. On the one hand, biomass and natural polymers
they are considered as one of the most powerful technologies show a heterogeneous structure and contain impurities that
that will provide a more efficient and sustainable utilization of complicate their processing to produce adequate conductive
energy in the short-term and real scenario of increasing energy carbon electrodes. On the other hand, the derived ACs are
costs and threatening climate change.1 conventionally obtained as powder or granular materials. Both
In spite of such relevance, the widespread utilization of the particle-like morphology and porosity strongly increase
supercapacitors has not been achieved due to a high cost to their inter- and intra-particle electrical resistances,1e respecti-
performance ratio. Among different electrode candidates, the vely, which makes their processing into electrode pastes
last decade has witnessed a very intense research in the capaci- necessary (by using auxiliary binders and conductivity promo-
tance and storage performance of different nanostructured ters). This generates many electric point contacts that
carbons, like graphene, CNTs, fullerenes, CNFs, onions, tem- decrease their mechanical and chemical stability.2 Hence the
plated carbons, etc.1 Nevertheless, apart from the significant development of binderless porous carbons by a direct route
progress in fundamental aspects, the complex and expensive would improve the conductivity and stability of carbon electro-
manufacture as well as the complicated handling and elec- des, and at the same time, it would reduce the time, com-
trode processing of these carbon materials make uncertain plexity and the environmental and economic impacts of the
their feasible application. Accordingly, while novel electrode overall manufacture. In addition, the possibility of using
materials and chemistries are being developed to improve the biomass precursors for the preparation of these electrodes
storage performance, further research on simpler and cheaper could provide a unique opportunity for their valorization into
manufacture and processing is demanded. high added-value products.
In this work, we report a sustainable and inexpensive
approach to process one of the most abundant polymers
a
in nature, lignin,3 into binderless flexible carbon electrodes
Universidad de Málaga, Andalucía Tech, Departamento de Ingeniería Química,
29071 Málaga, Spain. E-mail: [email protected]
with ultrafast response and overall high-performance for
b
Departamento de Química Inorgánica and Instituto Universitario de Materiales, energy storage in supercapacitors. The approach involves the
Universidad de Alicante, Apartado 99, Alicante, Spain preparation of interconnected and porous carbon fibers (CFs),
c
Departamento de Química Física and Instituto Universitario de Materiales, with a submicron diameter and high surface area and con-
Universidad de Alicante, Apartado 99, Alicante, Spain. E-mail: [email protected]
ductivity. Particularly, we demonstrate that these materials
† Electronic supplementary information (ESI) available: Details of material
characterization, the preparation of the cells, and the determination of gravi- can be simply but uniquely accomplished by electrospinning
metric capacitance, and specific energy and power. See DOI: 10.1039/c5gc02409a and subsequent thermal conversion of a suitable lignin into

1506 | Green Chem., 2016, 18, 1506–1515 This journal is © The Royal Society of Chemistry 2016
 View Article Online

Green Chemistry Paper

CFs, with an easy control of their connectivity and porosity

degree.
The preparation of the biomass-derived fibrous carbon elec-
trodes firstly entails the electrospinning of Alcell®-lignin/
ethanol solutions, at ambient temperature and pressure, into
submicron-diameter lignin fibers.4 Electrospinning is a simple
and powerful technique for fiber production,5 whereas the
chemical and structural properties of lignin make it a unique
precursor for the attainment of carbon materials.6 Further-
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

more, Alcell lignin can be dissolved in ethanol7 and it is the

only variety of this polymer that, up to now, has been electro-
spun at room temperature, directly without any modification
treatment and without any added binder polymer.8 Interest-
ingly, its pulping (organosolv) process uses ethanol as both
the delignifying agent and the solvent, and exhibits higher
pulp yields and it is more environmentally friendly than the
production of kraft lignin, the other most-produced lignin
today.3,4a Since lignin is a worldwide by-product of the pulp
and paper industries and is generated together with ethanol,
in future it may lead to wood-to-ethanol bio-refineries,3 and
the hereby reported preparation procedure, the resulting
electrode materials and their use for energy storage are con-
sidered a model example of biomass valorisation, green chem-
istry and sustainability. In fact, the valorization of lignin into
new and highly valuable products is essential for the replace-
ment of its petroleum-derived counterparts in multiple
applications.3–5
After electrospinning, the lignin submicrofibers were ther-
mally air-stabilized and carbonized under various conditions
to generate carbon submicrofibers with different connectivities
and porosities. The connectivity of fibers, i.e. the production
of individual (loose) or interconnected CFs, was controlled by
using a suitable heating rate during stabilization. At 0.08 °C
min−1, the heating rate is slow enough to stabilize lignin and
prevent its softening during subsequent carbonization,9 ren-
dering loose CFs (sample LCF-L in Fig. 1a and b). By contrast,
thermostabilization of lignin fibers at 0.16 °C min−1 resulted
Fig. 1 SEM images, covering different areas, of loose (a, b), inter-
in their partial softening and interconnection at contact points connected (c, d) and interconnected + partially gasified (e, f) lignin-based
between fibers during carbonization (sample LCF-I in Fig. 1c CFs; (g) histogram of fiber diameters from SEM images; (h) photograph
and d). of a binderless LCF-IA electrode; (i) SEM of a pasted LCF-IA electrode.
Both the loose and interconnected CFs exhibited quite
similar diameters, mainly in the range of 400 nm to 1.5 μm
(Fig. 1g) and are essentially microporous, as deduced from On the other hand, the microporosity of the interconnected
their characteristic type-I N2 isotherms (Fig. 2a). From these CFs was further developed (Fig. 2a), again without additional
data, the specific surface area (Table 1) and the pore size distri- steps, by partial gasification during carbonization in the pres-
bution (Fig. 2b) were calculated (see the Experimental section). ence of 1 vol% O2 (5% burn-off ) (LCF-IA sample). This was
These materials display a high surface area of ca. 700 m2 g−1 possible because of the high oxidation resistance of CFs
and pores in the range 0.4–1.3 nm, with those between obtained from lignin,4 again a unique feature of this precur-
0.6–0.8 nm prevailing. Such microporosity develops directly sor. Such a high oxidation resistance was confirmed by
during carbonization, without any additional activation thermogravimetric analysis (Fig. 2c), illustrating that these CFs
process, and seems to be related to the high oxygen content in start to react in air at very high temperatures (from 550 °C).
lignin.4,10 These pores of less than 2 nm size have been found The resulting interconnected CFs (sample LCF-IA in Fig. 1e
to be optimum for double layer charging, and the porous and f ) displayed similar diameters (Fig. 1g), but a higher
structure of CFs favors the accessibility and diffusion of ions;11 surface area of 1000 m2 g−1 (Table 1), the presence of some
these are well-known properties beneficial for energy storage wider micropores (d = 1.2–1.7 nm) (Fig. 2b) and a larger con-
in supercapacitors.1 centration of oxygen surface groups (Table 1). Particularly,

This journal is © The Royal Society of Chemistry 2016 Green Chem., 2016, 18, 1506–1515 | 1507
 View Article Online

Paper Green Chemistry

these groups mainly consisted of phenol- and quinone-like

functionalities, as evidenced by their evolution as CO
between 600–900 °C during TPD experiments12 or their clear
contributions at 286.1 and 289.7 eV, respectively, in XPS
spectra (Fig. 2d), which can actively contribute to charge
storage by pseudocapacitive reactions.1,12 Thus, while the
high carbonization temperatures applied for the obtainment
of sufficiently conducting carbons usually restrain their
functionalization (≤1.0 mmol CO per g for LCF-L and LCF-I),
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

the hereby reported O2-activation allows obtaining not only a

higher surface area but also an electro-active surface chem-
istry (3.5 mmol CO per g for LCF-IA), directly without any
post-treatment. In addition, by interconnection the materials
became flexible (Fig. 1h), the prerequisite to achieve stretch-
able and bendable energy storage devices.13 This is in agree-
ment with the enhanced mechanical behavior caused by
inter-fiber bonding of lignin-based carbon fibers prepared by
electrospinning.14
The structural features of the obtained lignin-based CFs
were also analyzed. These properties have been widely related
to their electron conductivity, but they were found practically
independent of the different connectivities and porosities of
the studied fibers. For instance, the XRD, Raman and TEM
characterization of the most porous CFs (LCF-IA) is shown in
Fig. 2e–i. The X-ray diffractogram and the Raman spectrum
reflect some common features characteristic of disordered
porous carbons. The observed broad (002) and (10) diffraction
peaks (Fig. 2e), centered at around 2θ = 25° and 43°, respecti-
vely, point out the lower crystallinity of the CFs compared to
that of graphite.15 By using the Scherrer equation,15a the mean
size of the ordered (crystalline) domains in the perpendicular
(Lc) and parallel (La) directions of graphene layers was esti-
mated to be 18 and 33 Å, respectively. On the other hand, the
strong D band (at 1355 cm−1) in the Raman spectrum is
indicative of the presence of structural defects, and the weaker
ones at 1170 and 1515 cm−1 have been ascribed to oxygen
surface groups and interstitial defects, respectively (Fig. 2f ).4,16
However, TEM reveals that the studied lignin-based CFs
exhibit an apparently smooth surface with no macro-/micro-
Fig. 2 (a) N2 adsorption isotherms, (b) pore-size distributions and scopic defects (Fig. 2g), even at the connecting points (Fig. 2h)
(c) thermogravimetric analysis of the studied lignin-based CFs. (d) C(1s)
or ending regions (Fig. 2i). Moreover, although they cannot be
XPS deconvoluted spectrum, (e) X-ray diffractogram, (f ) deconvoluted
first-order Raman spectrum, and (g–i) TEM images (scale bars are compared with those of graphitic or graphitized carbons, the
0.1 μm (g) 1 μm (h) and 1 μm (i)) of LCF-IA fibers. calculated mean crystal dimensions are among the largest

Table 1 Specific surface area (SBET); volume of micropores (Vt); CO and CO2 evolved from TPD experiments; mass surface composition; and yields
of thermostabilization and carbonization treatments and that of the overall production

TPD (mmol g−1) XPS (wt%) Yields (wt%)

SBET Vt
Sample (m2 g−1) (cm3 g−1) CO CO2 C(1s) O(1s) Stab. Carb. Overall

LCF-L 690 0.33 0.6 0.5 95.9 4.1 77.0 39.1 30.1
LCF-I 640 0.27 1.0 0.4 95.8 4.2 73.2 38.0 27.8
LCF-IA 1005 0.44 3.5 0.5 95.4 4.6 74.1 34.5 25.6

1508 | Green Chem., 2016, 18, 1506–1515 This journal is © The Royal Society of Chemistry 2016
 View Article Online

Green Chemistry Paper

for porous carbons prepared at temperatures in the range yield when carbonization and activation are both considered)
700–1000 °C.17 These values and the position of the (002) peak of the raw material. The overall yields (involving stabilization
(2θ = 25.08°), with an interlayer spacing of d002 = 3.548 Å which and carbonization treatments) for the production of the
is close to that obtained after carbonization of kraft lignin at reported lignin-based CFs range between 26–30%, similar to
temperatures as high as 2000 °C,16 suggest the existence of an those achieved in biomass based activated carbons20 and are
incipient graphitic structure. This can also be deduced from in agreement also with those of CFs obtained from other
the position of the G band in the Raman spectrum organosolv-based lignins.7,21 Hence, not only the described
(1595 cm−1) which approaches that of graphite.15,16a production of lignin fibers by electrospinning of ethanol
It is well known that the thermostabilization step and, solutions, but also their thermal processing into inter-
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

therefore, the fabrication of lignin-based CFs are a very slow connected CFs can be considered an environmentally improved
process.10 This causes a large energy consumption. In route to these types of materials.
addition, many exhausts evolve during the stabilization and The electrochemical behavior of the as-prepared lignin-
carbonization of lignin precursors.4b,10 These specifications based CFs was studied in a three-electrode cell in the sulfuric
are inherent to the low glass transition temperature of any acid electrolyte (Fig. 3). Cyclic voltammetry (CV) revealed the
lignin source and their high oxygen content. In fact, the slow ideal rectangular shape of a pure electric double-layer capaci-
thermostabilization step is also common for most CF precur- tor (EDLC),1 no matter what the connectivity of the CFs was
sors, like PAN, pitches, cellulose, etc.,9 however, the evolution
of H2, H2O, CO, CO2, light hydrocarbons and waste products
(tar and other condensable volatiles) is unavoidable when
biomass precursors are used. Although the acceleration of
thermostabilization of the lignin fibers is the aim of intense
research, most approaches involve the modification or purifi-
cation of lignin, the use of thermal pre-treatments or the utiliz-
ation of polymers and/or chemicals.18 As a result, additional
steps, reagents and energy consumption are involved and by-
products could be generated. In this sense, the hereby pro-
posed method for the preparation of interconnected lignin-
based CFs reduces the duration of the conventional air thermo-
stabilization process to half by raising the heating rate, just
without any modification of the conventional procedure or any
added compound or step. Thanks to it working slightly over
the glass transition temperature of the material, interconnec-
tion of fibers is also achieved without additives and the acti-
vation and functionalization with oxygen surface groups occur
simultaneously to thermal processing, without post-treat-
ments. Table 1 collects the yields of thermostabilization and
carbonization treatments. Stabilization yields not only account
for the condensation and decarboxylation reactions that result
in weight losses, they also consider the removal of adsorbed
ethanol on the surface of the lignin fibers, which represents
around 15% of their initial weight, thus leaving the weight loss
caused by lignin oxidation to be only ∼10%. The carbonization
yields were found to be slightly lower than those for the direct
carbonization of the same Alcell lignin under similar con-
ditions (40.3 wt%, not shown for brevity’s sake). This can be
explained considering that the high amount of oxygen groups
introduced during the stabilization treatment will be later
released as CO and CO2, thus removing part of the carbon that
otherwise would remain in the carbonized product. Neverthe-
less, the preparation yield is high considering the development
of surface area achieved after the carbonization step (which
could be seen as a combined carbonization/activation treat- Fig. 3 (a) Steady-state CVs of LCF-L, LCF-I and LCF-IA binderless elec-
trodes; (b) 4th scan CVs of LCF-L electrodes upon the step-wise opening
ment). In fact, for obtaining a similar textural development in
of the potential window on the positive and negative sides; (c) CV of the
the activation of non-porous carbon fibers from biomass pre- LCF-L electrode before and after repetitive cycling up to an expanded
cursors or even from phenolic resins, a burn-off higher than potential window of 1.0 V. Note: All data were obtained in a three-elec-
20% would be necessary,19 resulting in a worse use (lower trode cell; scan rate: 10 mV s−1.

This journal is © The Royal Society of Chemistry 2016 Green Chem., 2016, 18, 1506–1515 | 1509
 View Article Online

Paper Green Chemistry

(Fig. 3a). In addition, the different CFs displayed a pseudo- spatial arrangement of the graphitic crystallites and undefec-
capacitive contribution at around 0.4 V, related to the electro- tive surfaces (see Fig. 2e–i and the corresponding discussion);
chemical reaction of quinone-like oxygen functionalities.12 The as well as the binderless conformation of the electrodes, avoid-
significance of this contribution was found to be proportional ing the stability problems conventionally derived from the use
to the concentration of these oxygen groups, which can be of additives.
related to the amount of CO evolved during TPD experiments The practical application of the lignin-based CFs as electro-
(Table 1). The formation of such kinds of CO-evolving groups des for supercapacitors was investigated in symmetric two-elec-
is known to occur upon oxygen (ambient conditions) exposure trode cells, which were constructed by using the as-prepared
of a carbon surface prepared at high temperatures.22 From the materials as binderless electrodes, without using any kind of
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

charge enclosed by the CVs (Fig. 3a), the gravimetric capaci- additives or conditioning treatments. The advantages of using
tance values of 134, 153 and 201 F g−1 were obtained for interconnected CFs as binderless electrodes can be observed
LCF-L, LCF-I and LCF-IA electrodes, respectively. These values in Fig. 4a, where the response of binderless LCF-I is compared
seem to be proportional to the combination of their specific with that of binderless LCF-L and also with that obtained for
surface areas and the concentrations of electroactive CO-evol- LCF-I processed in the conventional form of an electrode paste
ving groups (Table 1). The stability of the carbon electrodes (sample LCF-Ip in Fig. 1i), adding Teflon and carbon black as
was studied by a step-wise opening of the voltammetric poten- the binder and the conductivity promoter, respectively. At a
tial window, from the open circuit potential (around 0.3 V) to low scan rate (100 mV s−1), the three materials render a quite
both negative and positive sides (Fig. 3b). The large cathodic similar electrochemical (CV) response. However, at faster scan
current observed below −0.4 V, corresponding to proton rates, for example at 1 V s−1, the carbon paste LCF-Ip shows a
electro-reduction, was probably indicative of the negative stabi- sluggish response, while the binderless LCF-I electrode is able
lity potential of LCF-L. In the case of LCF-I and LCF-IA this to keep the characteristic rectangular-shape CV of EDLCs
current was shifted towards lower (more negative) potentials. better than LCF-L. This difference must be undoubtedly
On the positive side, LCF-L also started to show anodic related to differences in their fiber connectivity (loose vs. inter-
currents at lower overpotentials than LCF-I and LCF-IA. This connected), a feature that is correlated to the conductivity of
material was able to reach 1.0 V without apparent volta- the electrode.14 In particular, when the scan rate is raised from
mmetric changes (Fig. 3b). However, after continuous cycling up 10 to 1000 mV s−1 the capacitance retention for the binderless
to this potential a considerable development of pseudocapaci- LCF-I electrode (62%) is higher than that of LCF-L (36%) and
tance was observed when compared to a previous CV registered three times higher than that for the pasted one (17% for
in a narrower potential window (see Fig. 3c). This increase in LCF-Ip), highlighting the beneficial utilization of LCF-I directly
pseudocapacitance may be attributed to the electrochemical as binderless electrodes.
oxidation of the carbon surface upon positive polarization in Capacitance retention for the LCF-I supercapacitor cell is
acid media, which produces the generation of electroactive CO- higher than that found for seamless carbon forms like acti-
type evolving groups.23 As a consequence, we concluded that vated carbon monoliths.24 Only advanced porous carbon
the potential safety limits determined for the three studied materials, such as hierarchical porous carbons25 or electro-
materials were at least −0.4 and 0.9 V (vs. Ag/AgCl/Cl-sat.), and spun PAN-based CO2-activated carbon binderless electrodes,26
that the supercapacitor cells constructed using them as binder- can match the outstanding performance of LCF-I binderless
less electrodes could be operated safely up to 1.3 V. Such a electrodes, though they require the use of more expensive and
high electrochemical stability is in agreement with the high less environmentally friendly carbon precursors or their prepa-
resistance of the lignin-based CFs to air oxidation (Fig. 2c), ration is burdened by complicated and energy-demanding
and may be related to their characteristic structure and the procedures.

Fig. 4 (a) Steady-state CVs from symmetric two-electrode cells of binderless LCF-L and binderless and pasted LCF-I (composition of the carbon
paste: 90/5/5 = LCF/Teflon/acetylene black). (b) Rate performance for the binderless LCF-I and LCF-IA cells. (c) GCD profiles of the binderless
LCF-IA cell at 32 A g−1 (cut-off voltage: 1.3 V). (d) Nyquist plots of freshly constructed two-electrode cells with different electrodes (applied voltage:
50 mV, sine amplitude: 10 mV; frequency ranges: 100 kHz–0.1 Hz).

1510 | Green Chem., 2016, 18, 1506–1515 This journal is © The Royal Society of Chemistry 2016
 View Article Online

Green Chemistry Paper

Under galvanostatic charge–discharge (GCD) conditions, specific currents (Fig. 4b). Considering the maximum cell
the different electrodes display a triangular-shaped response capacitances and the overall high-enough coulombic and
characteristic of the charge and discharge processes in EDLCs energy efficiencies obtained when working at 1.3 V (Table 2),
(Fig. S1†). Table 2 compiles a collection of properties for the this was selected as the operational voltage for both cells.
analysed supercapacitor cells. First, their cell capacitance was Because of their higher surface area and content of electro-
found to be practically independent of the cut-off voltage, and active oxygen groups, the interconnected and more activated
slightly increasing up to a maximum value of 1.3 V (Fig. S1† CFs (LCF-IA) exhibited the highest capacitance, while those
and Table 2). As expected, the cell capacitance values at low without additional activation (LCF-I) showed the best rate per-
specific currents are similar for the LCF-L and LCF-I electro- formance. For LCF-IA, a cell capacitance of 48.0 F g−1 is
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

des, being 35% higher when carbonized in the presence of achieved at the lowest tested specific current. Interestingly, the
oxygen (LCF-IA). After a 64-fold increase of the specific cell retains 42% of the capacitance when the applied current is
current, the capacitance severely dropped for the unconnected raised from 0.25 to an ultrahigh specific current density of
electrode (LCF-L). Surprisingly, the interconnected binderless 128 A g−1. A comparable storage rate performance, with a lower
electrode obtained after smooth air activation showed a capacitance but a better capacitance retention (58%), has been
boosted capacitance while preserving a low ohmic drop also found for LCF-I. In particular, under an extremely
(LCF-IA, Fig. S1† and Table 2). It can be seen that, although all demanding specific current of 32 A g−1, the cell with LCF-IA
the electrodes exhibited very low IR drops at 1 A g−1, differ- can be charged and discharged in less than 1.5 s while
ences between the cells arise as the applied current increases, keeping a high capacitance of 34.1 F g−1 and showing only
with LCF-I and LCF-IA (ohmic drops of 0.186 and 0.207 V) out- 103 mV of ohmic drop (Fig. 4c).
performing the results obtained for LCF-L (0.282 V). The cell Moreover, these electrodes show other valuable features for
resistances were calculated from these values, going down their use in symmetric supercapacitors. On the one hand, both
from 3.19 to 2.12 Ohms after the interconnection of the the negative and positive electrodes display quite similar capa-
carbon fibers. As observed in the table, most of the stored citances, causing only a slightly uneven operational potential
charge was effectively recovered during discharge of the cells, window (values of ΔV + and ΔV − of 0.69 and 0.61 V, respectively,
with coulombic efficiencies ranging between 99.0 and 99.5% Fig. S2†). On the other hand, the potential limits that each
in the case of the interconnected electrodes, even when the electrode reached during the charging of the cell were nearly
voltage was increased up to 1.3 V. The energy efficiency of the constant, no matter what the used specific current was
cells, which is usually an underestimated parameter of super- (Fig. S2†). This stability of the potential windows allows the
capacitors, was also measured. It reaches values close to 90% safe operation of this cell at 1.3 V under a wide range of
when working at 1.0 V. When the cut-off voltage was further current loads, a very interesting feature for systems that
increased up to 1.3 V, the larger contribution of redox reac- require or provide variable sources of power.
tions to capacitance led to a small decrease (around 3–5%) in The observed ultrafast response of the lignin-based inter-
the energy efficiency for all the electrodes, being slightly larger connected CFs and their superior rate performance is sup-
for the electrode with the larger amount of surface functional- ported by electrochemical impedance spectroscopy (EIS)
ities (5.6% for LCF-IA). Upon raising the applied current inten- measurements (Fig. 4d). The low and fairly similar Rcell values
sity from 1 to 64 A g−1, the increasing energy losses due to the (the first intercept on the Z′ axis) in Nyquist plots (around
Joule effect made the energy efficiencies decay. Nevertheless, 0.64–0.70 Ω at 100 kHz) observed in all cases reflect the high
the interconnection of fibers seemed to lower these energy conductivity of the whole cell, including that of the electrode
losses when working at higher specific currents, at least for materials,27 which may be associated with the suitable pro-
the 1.3 V operating voltage. perties of lignin to produce ordered carbon structures.16a
Owing to their better response, the rate performance of However, clear differences in the semicircle diameters (and
LCF-I and LCF-IA cells was then analyzed upon a wide range of the Equivalent Series Resistance (ESR)) were found for the

Table 2 Cell gravimetric capacitance (Cg), coulombic (Ceff ) and energy (Eeff ) efficiencies, and resistance (Rcell ) of two-electrode cells constructed
with the lignin-based CFs

1 A g−1 64 A g−1

Voltage Cg Eeff Ceff Cg Eeff Ceff Rcell

Electrode (V) (F g−1) (%) (%) (F g−1) (%) (%) (Ω)

LCF-L 1.0 32 89.1 99.3 19 75.4 99.5 3.19

1.3 34 86.5 98.9 18 69.2 98.0
LCF-I 1.0 32 90.2 99.3 27 75.9 99.4 2.12
1.3 34 86.9 98.9 27 71.1 98.3
LCF-IA 1.0 45 90.6 99.3 35 71.2 99.8 2.36
1.3 47 84.8 98.9 29 69.8 99.9

This journal is © The Royal Society of Chemistry 2016 Green Chem., 2016, 18, 1506–1515 | 1511
 View Article Online

Paper Green Chemistry

different electrode materials, which are indicative of a remark-

able influence of their distinct continuity, connectivity and
porosity, or the amount of functional groups.27 The huge
increase of resistivity observed for the LCF-I electrode when
prepared as a paste highlights the benefits of using binderless
materials to obtain higher conductivities. In addition, ca. 40%
reduction in resistance is achieved when the topology of the
carbon electrode is changed from loose (LCF-L) to a well-inter-
connected network of CFs (LCF-I). This is attributed to the
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

increased network connectivity in the interconnected CFs, pro-

viding a greater number of pathways for charge transport.14
Furthermore, the results also indicate that an extra activation
of the interconnected CFs (LCF-IA) increases their electrical
resistance, probably connected to the presence of a larger
amount of electron-withdrawing surface oxygen groups and
porosity-derived structural defects.4b Nevertheless, in spite of
their narrow porosity, all the electrodes showed a low Equi-
valent Distributed Resistance (EDR),27 reflecting the low electro-
lyte resistance within the highly accessible pore systems of
CFs,28 especially, in those with an ultrafine diameter prepared
by electrospinning.29
The exceptional performance of the lignin-based fibrous
carbon materials was finally highlighted in terms of specific
energy and power of the supercapacitor cells in a Ragone
plot (Fig. 5a). For comparison purposes, the plot also collects
among the best reported performance results of other super-
capacitor devices constructed with flexible30 and electrospun-
based29,31 carbon materials. In terms of energy density, the
most activated CFs (LCF-IA) outperform the other tested
lignin-based materials in a broad range of power densities,
with a maximum energy density close to 10 Wh kg−1 and Fig. 5 (a) Ragone plots for the studied materials. Literature values from
other electrospun-based or flexible supercapacitors are also referred. (b)
near 5 Wh kg−1 when discharged for 1 s (power of 18 kW kg−1).
Durability test of the LCF-I cell carried out at 5 A g−1. A resting time of
Nevertheless, the maximum deliverable power density is 1 h was imposed each 25000 cycles. (a, b) Operating voltage of 1.3 V.
observed for the less porous CFs (LCF-I), showing a
maximum power of 61 kW kg−1 while storing an energy of
2.0 Wh kg−1 that would be delivered in less than 0.2 s under
those conditions. By contrast, the loose CFs (LCF-L) and, to a Another important aspect from a practical point of view is
much greater extent, the pasted discontinuous ones (LCF-Ip), the stability of supercapacitors upon charge–discharge cycling.
show a marked performance decay at highly demanding The cycle life of the lignin-based interconnected CFs was eval-
power densities. Hence, the ultrafast responses of LCF-I and uated upon 100 000 GCD cycles at 5 A g−1 and 1.3 V voltage. As
LCF-IA electrode materials are uniquely attributed to their it can be observed in Fig. 5b, the cell can practically retain
excellent conductivity and the interconnection and long- 100% of its power and around 90% of its total capacitance and
range continuity of the finely electrospun CFs, attained energy, with this slight fading stabilizing after the first 50 000
through the control of their stabilization process and by cycles. An additional test consisting of a holding voltage (the
using them as binderless electrodes, respectively. As it can “floating” accelerated test32) at 1.3 V for 24 h was conducted
also be deduced from the Ragone plot, the energy and power after the cycling one. GCD profiles recorded after the cycling
characteristics of the cells made with both kinds of inter- and holding tests show small changes in the resistivity and the
connected lignin-based CFs are similar to the best results capacitance of the cell (Fig. S3a†), resulting in a reduced
found for supercapacitors based on electrospun electrodes impact in its energy and power (Fig. S3b†).
reported in the literature,31a,c,g but without requiring the use
of porogens or porosity development treatments. Further-
more, the ultrafast energy delivery exhibited by the reported
CFs exceeds by far that of activated carbons and approaches Conclusions
the performance of the top-engineered nanocarbons,1 just by
using an abundant and biorenewable precursor and a simple We have demonstrated that an abundant and renewable pre-
preparation procedure. cursor like lignin can be inexpensively and sustainably pro-

1512 | Green Chem., 2016, 18, 1506–1515 This journal is © The Royal Society of Chemistry 2016
 View Article Online

Green Chemistry Paper

cessed into interconnected and porous CFs, of submicron (SAIEUS, available online at http://www.nldft.com/) Software.
diameters, which exhibit an ultrafast storage performance For thermogravimetric analysis (TG), a CI Electronics MK2
when used as electrode materials for supercapacitor devices. balance and samples of 10 mg were used.
The electrodes are simply prepared by electrospinning of The amount of oxygen surface groups was determined as
Alcell®-lignin/ethanol solutions and subsequent thermal con- the CO and CO2 desorbed from the samples by temperature-
version into CFs, with an easy control of their connectivity and programmed desorption (TPD) analyses. These analyses were
activation degree. Supercapacitor cells can be easily mounted carried out up to 940 °C (10 °C min−1) in He flow (100 cm3
using the CFs as flexible binderless electrodes, showing power STP per min) using a differential scanning calorimetry-thermo-
and energy densities of up to 61 kW kg−1 and 10 Wh kg−1, gravimetric analysis (DSC–TGA) equipment (TA Instruments,
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

respectively, which are comparable or superior to the up-to- SDT 2960 Simultaneous) coupled to a mass spectrometer
now top engineered structures, and can retain 100% and more (Thermostar, Balzers, GSD 300 T3). The amounts of CO and
than 90% of the initial power and energy densities, respecti- CO2 desorbed from the samples were monitored following the
vely, after 100 000 charge–discharge cycles at 5 A g−1. 28 and 44 m/z signals, which were previously calibrated by the
non-isothermal decomposition of calcium oxalate mono-
hydrate (99.999%). XPS data were registered by using a 5700C
Experimental model Physical Electronics apparatus with Mg Kα radiation
Preparation of lignin-based CFs (1253.6 eV). Raman spectra were recorded with a Renishaw
micro-Raman system using an Ar+ laser at 514 nm as the exci-
Lignin fibers were prepared by electrospinning of Alcell®
tation source with a spectral resolution of 2 cm−1. The fitting
lignin–ethanol solutions by using a co-axial configuration.4
of both the XPS and Raman spectra was done by least squares
The spinnable solution had a lignin concentration of 45 wt%
using Gaussian–Lorentzian peak shapes. The X-ray diffracto-
and a viscosity of 370 cP. The flow rates were 1 mL h−1 for the
grams were obtained using a PANalytical diffractometer (X’Pert
lignin solution (inner spinneret) and 0.1 mL h−1 for pure
PRO) using Cu Kα1 radiation (1.5406 Å). The profile intensities
ethanol (outer spinneret). The tip-to-collector distance was
were measured step by step (0.0167° in 2θ) for a total time of
30 cm and the electrical potential difference was 12 kV (the
3000 s. In order to determine the exact angular position of the
collector was at −6 kV and the tips at +6 kV). Lignin was kindly
(002) peak, 10 wt% Si (Aldrich 99.999%) was used as an
provided by Repap Technologies Inc. (Canada), while ethanol
internal standard. For the application of the Scherrer equation,
absolute ( p.a. Merck) was used as the solvent. The compo-
the instrumental broadening contribution was determined by
sition of the used lignin was determined by CHNS/O elemental
using LaB6 as the crystalline standard.
analysis (PerkinElmer® 2400C Instruments) as: 65.9 wt% C;
6.3 wt% H; 0.2 wt% N; 0.00 wt% S; and 27.6 wt% O, while the
ash content was 0.00 wt%. Thermostabilization and carboniz- Electrochemical characterization
ation treatments were carried out in a horizontal tubular
The electrochemical measurements have been performed in
furnace. Thermostabilization was performed under an air
1 M H2SO4 using an Autolab PGSTAT302 Potentiostat equipped
atmosphere (150 cm3(STP) min−1) and at different heating
with the FRA module for the Electrochemical Impedance
rates up to 200 °C, keeping this final temperature for 36 h. The
Spectroscopy (EIS), Cyclic voltammetry (CV) and Galvanostatic
stabilized materials were carbonized at 10 °C min−1 up to
charge–discharge (GCD) measurements in a three electrode
900 °C under inert (N2) or 1 vol% O2 under a N2 atmosphere
cell and an Arbin SCTS Instrument for the galvanostatic
(150 cm3(STP) min−1).
charge–discharge experiments and durability tests in a two
electrode cell. A T-shaped Teflon Swagelock system equipped
Physico-chemical characterization with gold collectors and an Ag/AgCl 3 M KCl reference elec-
SEM and TEM images were obtained by using a JEOL trode has been used for assessing the cell performance while
JSM-6490LV microscope (at 25 kV voltage) and a Philips registering separately the potential of each electrode. Round-
CM200 microscope (at an accelerating voltage of 200 kV), shaped CF pieces were cut into a surface area of 0.196 cm2, the
respectively. The histogram of fiber diameters was constructed working and counter electrodes constructed by stacking three
from SEM images after an analysis of more than 120 cases for round pieces together (total weight between 1.00–1.20 mg) over
each different CF. The textural properties were characterized the gold collectors at each side of the cell, achieving a surface
by N2 adsorption–desorption at −196 °C using a Micromeritics loading in the range of 5–6 mg cm−2. A round piece of nylon
ASAP2020, after outgassing for 8 h at 150 °C under vacuum. membrane ( pore size: 0.45 µm, 7 mm diameter) was used as
From the N2 adsorption/desorption isotherms, the apparent the separator. The electrodes were dried at 80 °C under
surface area (SBET) was calculated by applying the BET vacuum for 2 h before determining their weight, and after-
equation, and the micropore volume (Vt) was calculated by wards soaked in the electrolyte for 1 h prior to cell construc-
using the t-method. Pore size distribution has been calculated tion and the electrochemical measurements. Further
from the N2 adsorption isotherms considering the proposed information about the determination of gravimetric capaci-
2D-NLDFT heterogeneous surface model,33 and by applying tances, energy and power density, cell resistance, efficiencies
the Solution of Adsorption Integral Equation Using Splines and other relevant parameters is compiled in the ESI.†

This journal is © The Royal Society of Chemistry 2016 Green Chem., 2016, 18, 1506–1515 | 1513
 View Article Online

Paper Green Chemistry

Acknowledgements 14 I. Dallmeyer, L. T. Lin, Y. Li, F. Ko and J. F. Kadla, Macro-

mol. Mater. Eng., 2014, 299, 540.
Financial support by the Spanish Ministerio de Economía y 15 (a) M. A. Short and P. L. Walker Jr., Carbon, 1963, 1, 3;
Competitividad, for the MAT2013-42007-P, P09-FQM-5156R, (b) D. D. L. Chung, J. Mater. Sci., 2002, 37, 1475.
CTQ2012-36408, JCI2011-10566, JCI-2012-12664 and the joint 16 (a) J. Rodríguez-Mirasol, T. Cordero and J. J. Rodríguez,
Spanish-Japanese (PRI-PIBJP-2011-0766) projects, FEDER and Carbon, 1996, 34, 43; (b) A. Sadezky, H. Muckenhuber,
the Junta de Andalucía is gratefully acknowledged. H. Grothe, R. Niessner and U. Poschl, Carbon, 2005, 43,
1731.
17 (a) M. M. Dubinin, Thermal treatment and microporous
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

structure of carbonaceous adsorbents, in Proceedings of the

References fifth Conference on Carbon, ed. Y. Zhou, Pergamon Press
Inc., 1962; (b) Z. Ryu, H. Rong, J. Zheng, M. Wang and
1 (a) Y. Zhai, Y. Dou, D. Zhao, P. F. Fulvio, R. T. Mayes and B. Zhang, Carbon, 2002, 40, 1131; (c) X. Ma, H. Yang, L. Yu,
S. Dai, Adv. Mater., 2011, 23, 4828; (b) S. L. Candelaria, Y. Chen and Y. Li, Materials, 2014, 7, 4431.
Y. Shao, W. Zhou, X. Li, J. Xiao, J.-G. Zhang, Y. Wang, J. Liu, 18 (a) E. Sjöholm, G. Gellerstedt, R. Drougge and I. Brida, WO
J. Li and G. Cao, Nano Energy, 2012, 1, 195; (c) T. Chen and 2012/112108A1, 2012; (b) E. Sjöholm, G. Gellerstedt,
L. Dai, Mater. Today, 2013, 16, 273; (d) H. Nishihara and R. Drougge and I. Norberg, WO 2013/112100A1, 2013;
T. Kyotani, Adv. Mater., 2012, 24, 4473; (e) M. Inagaki, (c) S. Kubo and J. F. Kadla, J. Polym. Environ., 2005, 13, 97;
H. Konno and O. Tanaike, J. Power Sources, 2010, 195, 7880. (d) S. Hu and Y.-L. Hsieh, J. Mater. Chem. A, 2013, 1, 11279;
2 S. Shiraishi, Boletín del Grupo Español del Carbón, ISSN-e (e) I. Dallmeyer, L. T. Lin, Y. Li, F. Ko and J. F. Kadla, Macro-
2172-6094, 2013, 28, 18. mol. Mater. Eng., 2013, 298, 1; (f ) W. Qin and J. F. Kadla,
3 A. J. Ragauskas, G. T. Beckham, M. J. Biddy, R. Chandra, J. Appl. Polym. Sci., 2012, 126, E203; (g) J. Lin, S. Kubo,
F. Chen, M. F. Davis, B. H. Davison, R. A. Dixon, P. Gilna, T. Yamada, K. Koda and Y. Uraki, BioResources, 2010, 7,
M. Keller, P. Langan, A. K. Naskar, J. N. Saddler, 5634.
T. J. Tschaplinski, G. A. Tuskan and C. E. Wyman, Science, 19 (a) A. Oya, S. Yoshida, J. Alcaniz-Monge and A. Linares-
2014, 344, 1246843. Solano, Carbon, 1996, 34, 53; (b) P. T. Williams and
4 (a) M. Lallave, J. Bedia, R. Ruiz-Rosas, J. Rodríguez-Mirasol, A. R. Reed, Biomass Bioenergy, 2006, 30, 144; (c) W. M. Qiao,
T. Cordero, J. C. Otero, M. Márquez, A. Barrero and M. Huda, Y. Song, S.-H. Yoon, Y. Korai, I. Mochida, O. Katou,
I. G. Loscertales, Adv. Mater., 2007, 19, 4292; (b) R. Ruiz- H. Hayashi and K. Kawamoto, Energy Fuels, 2005, 19,
Rosas, J. Bedia, M. Lallave, I. G. Loscertales, A. Barrero, 2576.
J. Rodríguez-Mirasol and T. Cordero, Carbon, 2010, 48, 696. 20 H. Marsh and F. R. Reinoso, Activated Carbon, Elsevier,
5 A. Greiner and J. H. Wendorff, Angew. Chem., Int. Ed., 2007, 2006.
46, 5670. 21 S. Kubo, Y. Uraki and Y. Sano, Carbon, 1998, 36, 1119.
6 (a) J. M. Rosas, R. Berenguer, M. J. Valero-Romero, 22 T. J. Bandosz, in Carbon Materials for Catalysis, ed. P. Serp
J. Rodríguez-Mirasol and T. Cordero, Front. Mater., 2014, and J. L. Figueiredo, John Wiley & Sons, Inc., 2008,
1(29), 1; (b) W.-J. Liu, H. Jiang and H.-Q. Yu, Green Chem., pp. 45–92.
2015, DOI: 10.1039/C5GC01054C. 23 (a) R. Berenguer, J. P. Marco-Lozar, C. Quijada, D. Cazorla-
7 Y. Ni and Q. Hu, J. Appl. Polym. Sci., 1995, 57, 1441. Amorós and E. Morallón, Carbon, 2009, 47, 1018;
8 (a) D. A. Baker and T. G. Rials, J. Appl. Polym. Sci., 2013, (b) R. Berenguer, H. Nishihara, H. Itoi, T. Ishii,
130, 713; (b) E. Frank, L. M. Steudle, D. Ingildeev, E. Morallón, D. Cazorla-Amorós and T. Kyotani, Carbon,
J. M. Spörl and M. R. Buchmeiser, Angew. Chem., Int. Ed., 2013, 54, 94; (c) Z. Tabti, R. Berenguer, R. Ruiz-Rosas,
2014, 53, 5262. C. Quijada, E. Morallón and D. Cazorla-Amorós, Electro-
9 J. L. Braun, K. M. Holtman and J. F. Kadla, Carbon, 2005, chemistry, 2013, 81, 833.
43, 385. 24 V. Ruiz, C. Blanco, R. Santamaría, J. M. Ramos-Fernández,
10 J.-W. Jeon, L. Zhang, J. L. Lutkenhaus, D. D. Laskar, M. Martínez-Escandell, A. Sepúlveda-Escribano and
J. P. Lemmon, D. Choi, M. I. Nandasiri, A. Hashmi, J. Xu, F. Rodríguez-Reinoso, Carbon, 2009, 47, 195.
R. K. Motkuri, C. A. Fernandez, J. Liu, M. P. Tucker, 25 D.-W. Wang, F. Li, M. Liu, G. Q. Lu and H.-M. Cheng,
P. B. McGrail, B. Yang and S. K. Nune, ChemSusChem, Angew. Chem., Int. Ed., 2008, 47, 373.
2015, 8, 428. 26 (a) Y. Gao, V. Presser, L. Zhang, J. J. Niu, J. K. McDonough,
11 A. Linares-Solano and D. Cazorla-Amorós, Activated C. R. Pérez, H. Lin, H. Fong and Y. Gogotsi, J. Power
Carbon Fibers, in Handbook of Advanced Ceramics, ed. S. Sources, 2012, 201, 368; (b) C. Kim, J. Power Sources, 2005,
Somiya, Academic Press, Elsevier Inc., 2013, p. 155. 142, 382; (c) E. J. Ra, E. Raymundo-Piñero, Y. H. Lee and
12 M. J. Bleda-Martínez, J. A. Maciá-Agulló, D. Lozano- F. Béguin, Carbon, 2009, 47, 2984.
Castelló, E. Morallón, D. Cazorla-Amorós and A. Linares- 27 R. Kötz and M. Carlen, Electrochim. Acta, 2000, 45, 2483.
Solano, Carbon, 2005, 43, 2677. 28 M. J. Bleda-Martínez, D. Lozano-Castelló, D. Cazorla-
13 K. Xye and B. Wei, Adv. Mater., 2014, 16, 3592. Amorós and E. Morallón, Energy Fuels, 2010, 24, 3378.

1514 | Green Chem., 2016, 18, 1506–1515 This journal is © The Royal Society of Chemistry 2016
 View Article Online

Green Chemistry Paper

29 M. Inagaki, Y. Yang and F. Kang, Adv. Mater., 2012, 24, 2547. Y. Song and L. Liu, Chem. Eng. J., 2014, 249, 216; (d) C. Lai,
30 (a) W. Liu, X. Yan, J. Lang, C. Peng and Q. Xue, J. Mater. Z. Zhou, L. Zhang, X. Wang, Q. Zhou, Y. Zhao, Y. Wang,
Chem., 2012, 22, 17245; (b) S. H. Aboutalebi, R. Jalili, X.-F. Wu, Z. Zhu and H. Fong, J. Power Sources, 2014, 247,
D. Esrafilzadeh, M. Salari, Z. Gholamvand, S. Aminorroaya 134; (e) T. Le, Y. Yang, Z. Huang and F. Kang, J. Power
Yamini, K. Konstantinov, R. L. Shepherd, J. Chen, Sources, 2015, 278, 683; (f ) T. Wang, D. Song, H. Zhao,
S. E. Moulton, P. C. Innis, A. I. Minett, J. M. Razal and J. Chen, C. Zhao, L. Chen, W. Chen, J. Zhou and E. Xie,
G. G. Wallace, ACS Nano, 2014, 8, 2456. J. Power Sources, 2015, 274, 709; (g) C. Tran and V. Kalra,
31 (a) E. J. Ra, E. Raymundo-Piñero, Y. H. Lee and F. Béguin, J. Power Sources, 2013, 235, 289.
Carbon, 2009, 47, 2984; (b) A. Kurniawan, L. K. Ong, 32 D. Weingarth, A. Foelske-Schmitz and R. Kötz, J. Power
Published on 14 October 2015. Downloaded by Universidad de Alicante on 9/19/2018 3:17:58 PM.

F. Kurniawan, C. X. Lin, F. E. Soetaredjo, X. S. Zhao and Sources, 2013, 225, 84.

S. Ismadji, RSC Adv., 2014, 4, 34739; (c) C. Ma, Y. Li, J. Shi, 33 J. Jagiello and J. P. Olivier, Carbon, 2013, 55, 70.

This journal is © The Royal Society of Chemistry 2016 Green Chem., 2016, 18, 1506–1515 | 1515