Influence of Multiwalled Carbon Nanotubes as Additives in Biomass

Subscriber access provided by University of Winnipeg Library

Energy, Environmental, and Catalysis Applications

Influence of multiwalled carbon nanotubes as additives in biomass-derived carbons for supercapacitor applications Natalia Rey-Raap, Marina Enterria, José Inácio Martins, Manuel Fernando R. Pereira, and José Luís Figueiredo ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b19246 • Publication Date (Web): 17 Jan 2019 Downloaded from http://pubs.acs.org on January 18, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Influence of multiwalled carbon nanotubes as additives in biomass-derived carbons for supercapacitor applications Natalia Rey-Raapa*, Marina Enterríaa, José Inácio Martinsb,c, Manuel Fernando R. Pereiraa, José Luís Figueiredoa a

Associate Laboratory LSRE-LCM, Departamento de Engenharia Química, Faculdade

de Engenharia, Universidade do Porto, R. Dr. Roberto Frias s/n, 4200-465 Porto, Portugal b

Departamento de Engenharia Química, Faculdade de Engenharia, Universidade do

Porto, R. Dr. Roberto Frias s/n, 4200-465 Porto, Portugal c

Lab2PT, Instituto de Ciências Sociais, Universidade do Minho, Portugal

Keywords: Hydrothermal carbons; glucose-derived carbons; electric double-layer capacitor; oxygen functionalities; high porosity. Abstract Glucose-derived carbon/carbon nanotube hybrid materials were prepared by hydrothermal carbonization of glucose in the presence of carbon nanotubes and subsequent carbonization, physical or chemical activation. The proportion of carbon nanotubes added during the hydrothermal polymerization of glucose was varied in order to ascertain the optimum dose to maximize the performance of the carbon hybrids in supercapacitor applications. Both the thermal treatment applied and the addition of carbon nanotubes lead to changes in the textural and chemical properties of the activated carbons. It was observed that samples bearing carbon nanotubes exhibit higher number of nucleation centres for glucose oligomers to polymerize and, consequently, the behaviour of the hydrothermal carbon towards activation differs according to the activating agent employed. Moreover, the initial chemical speciation dominated by acidic groups shifts to more basic functionalities (quinones and carbonyl groups) with the addition of carbon nanotubes. The effect of the different physicochemical properties of the prepared carbons on their electrochemical behaviour was evaluated. The addition of 2 wt.% of carbon nanotubes and subsequent chemical activation leads to electrode materials yielding 206 F g-1 and 78 % of capacitance retention up to 0.8 V and 20 A g-1, and high rate cyclability (97 % after 5000 cycles). The outstanding performance is ascribed to the high surface *Corresponding author: [email protected] (Natalia Rey-Raap) Tel. 00351 220414919 ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 31

area, narrow mesopores, and phenol/carbonyl surface functionalities, which enhance molecular diffusion, the amount of stored energy and electronic transportation, respectively. 1. Introduction The continuous increase in the consumption of fossil fuels during the last decades is causing the gradual depletion of the accessible fuel-deposits of the world. This fact has generated a deeper interest in developing new technologies that allow the incorporation of other sources of energy that may compete with petroleum. Alternative energies, such as wind and solar, are highly intermittent and, consequently, energy storage became as important as energy conversion. The development of energy storage systems providing an efficient supply in peaks of energy demand or shortage is therefore a key feature to displace a petroleum-based economy. Supercapacitors are energy storage devices capable of supplying high power density, i.e., fast energy delivery

1-3,

and their use is of great

interest in a large number of applications such as stand-by power systems, portable electronics, hybrid electric vehicles, etc. 4-5. Unfortunately, these devices suffer from low energy density (capacitance) 1, 6, which will restrict their applicability over the long-term, as the requirements of the applications become more demanding 2. The capacitance of a supercapacitor is directly related to the properties of the electrode material 7-9, and hence, further breakthroughs in the development of advanced materials are required in order to increase the amount of stored energy without compromising the power density, which is the most attractive feature of supercapacitors. Advanced carbon materials have been extensively explored as electrodes for supercapacitors owing to their high surface area, outstanding electrical conductivity and excellent electrochemical stability

10-12.

Carbon nanotubes, carbon nanofibers, carbon-

based composites and graphene-based materials are some examples of advanced carbon materials that have been used as electrodes. The reader is referred to some excellent reviews on this topic

1, 5, 7, 13-16.

Although these materials may provide for high-

performance supercapacitors, they are mostly obtained from chemical compounds or fossil-fuels, or their production process is complex and expensive. In addition, supercapacitors are also of interest in daily applications such as electric vehicles; therefore, a facile, low-cost and environmentally friendly route to produce novel valuable carbon electrode materials is needed.

2

ACS Paragon Plus Environment

Page 3 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In this context, the valorisation of biomass as a source of electrode carbon materials through hydrothermal processes has emerged as an interesting alternative, since: i) biomass is an abundant and low cost resource; and ii) the production of hydrothermal carbons is considered an affordable and environmentally friendly method owing to its low cost and mild synthesis conditions 2, 17-23. However, one of the main limiting factors that hinder the applicability of hydrothermal carbons is that they usually exhibit limited porosity and inadequate chemical properties

24-25.

In order to overcome this drawback,

different strategies have been addressed: i) modification of the duration and temperature of the hydrothermal process 6, 26-27, ii) carbonization and/or activation of the as-prepared hydrothermal carbons

19, 21, 28-29

or iii) incorporation of different heteroatoms (nitrogen,

phosphor, boron, etc.) or advanced carbon materials (graphene oxide or carbon nanotubes) into the initial precursor solution

6, 10, 18-19, 30-31.

Although these strategies

allow to modify the textural and chemical properties of the hydrothermal carbons, sometimes it is not easy to come up with materials that provide large microporosity and high electrical conductivity, two essential features in supercapacitor applications. The incorporation of carbon nanotubes during the hydrothermal carbonization of biomass is an interesting solution to increase the electrical conductivity of the hydrothermal carbons. In fact, the hybridization of CNT with other types of materials such as redox-based electrocatalysts has been shown as an excellent approach to improve the kinetic and cyclic capabilities of supercapacitor electrodes

32.

However, the effect of adding CNT during

the synthesis of carbons with low conductivity is still uncertain and additional work is needed to investigate the effect of the addition of carbon nanotubes into the biomassderived carbon structures, emphasising the relationship between the textural, chemical and conductive properties and their electrochemical performance, thus enabling the rational design of supercapacitor electrodes. In the present study, glucose-derived carbons were prepared by hydrothermal carbonization of glucose in the presence of multiwalled carbon nanotubes (CNT). An attempt has been made to find the optimal concentration of CNT that leads to the best supercapacitor performance without greatly increasing the production cost of the final material. In addition, the as-prepared hybrid materials were subjected to carbonization and physical and chemical activation, with the aim to promote modifications in the textural and chemical properties of the synthesized materials. Both three- and twoelectrode measurements were performed to fully characterize the electrochemical

3



behaviour of the prepared carbons. Promising results concerning the addition of carbon nanotubes during the hydrothermal treatment of glucose-derived carbons are detailed, showing the effectiveness of the proposed synthesis strategy in converting a low-value biomass-derived product into a high-value carbon material with high supercapacitor performance. 2. Experimental 2.1. Materials preparation Glucose derived carbon/carbon nanotube hybrids were synthesized via hydrothermal treatment. To this end, the selected starting materials were D-glucose (Merck), multiwalled carbon nanotubes (NANOCYL™ NC3100 series, with average diameter of 9.5 nm, average length of 1.5 µ, carbon purity higher than 95 wt.% and 2.6 wt.% of oxygen 33) and deionized water (from a Panice reverse osmosis water system). Glucose and CNTs were first dispersed in deionized water in an unsealed glass beaker under sonication for around 30 min at room temperature. Each mixture was then transferred into a 100 mL Teflon-lined autoclave, sealed and heated at 200 ºC for 16 h. In all cases, the precursor mixtures were prepared with a solid content of 15 wt.% and a maximum volume equal to 2/3 that of the autoclave. The solid product was recovered by filtration, washed with abundant deionized water and dried in air at 80 ºC for 24 h. Different amounts of CNTs were added based on the hydrothermal carbonization yield of glucose obtained in the present study (35-40 %). Dried samples (organic resin/CNT hybrids) were placed in a vertical furnace where three different thermal treatments were performed: carbonization, physical activation and chemical activation. Carbonization was performed at 700 ºC under a nitrogen flow of 50 cm3 min-1 and the heating rate and carbonization time were fixed at 5 ºC min-1 and 2 h, respectively. Physical activation was performed at 900 ºC under a carbon dioxide flow of 150 cm3 min-1 and the heating rate and activation time were fixed at 5 ºC min-1 and 9 h, respectively. Dried samples were also chemically activated using potassium hydroxide (Fluka) as activating agent and a KOH/organic resin mass ratio of 3:1. Activation was performed at 700 ºC under a nitrogen flow of 50 cm3 min-1. The heating rate and activation time were fixed at 5 ºC min-1 and 2 h, respectively. After activation with KOH, samples were mixed with HCl (1 M) to neutralize the activating agent and/or its decomposition

4


Page 4 of 31

Page 5 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


products, and then washed with boiling distilled water under reflux until complete chloride removal. Finally, washed samples were air dried at 80 ºC for 24 h. The hydrothermal carbons (HTC) with no further thermal treatment were denoted as G_X%CNT, while carbonized, physically and chemically activated samples were labelled CG_X%CNT, AG_X%CNT and AG_X%CNT_KOH, respectively. In all cases, X is the amount of carbon nanotubes added (0, 2, 4 and 8 wt.%). 2.2. Textural and chemical characterization The textural properties of all materials were characterized by analysis of the nitrogen adsorption-desorption isotherms at -196 ºC using a Quantachrome Autosorb iQ automated gas sorption analyser. Prior to analysis, all samples were degassed under vacuum at 150 ºC overnight. The specific surface area (SBET) was evaluated by means of the BET equation applied to the nitrogen adsorption isotherms in a linear region within the relative pressure range of 0.09-0.2. The micropore volume (VDR) was determined according to the Dubinin–Radushkevich method and the total pore volume (VP) was calculated from the volume of nitrogen adsorbed at saturation point (relative pressure of 0.99). The QSDFT method, which takes into account the effects of surface roughness and heterogeneity, was used to obtain accurate pore size distributions. The morphology of the samples was examined with a Quanta FEG 400 scanning electron microscope (FEI Company). The samples were previously attached to an aluminium pin using conductive double-sided adhesive tape. An accelerating voltage of 15 kV and a secondary electron detector Everhart-Thornley (ETD) were used in all the analyses. The surface chemistry of the organic resins was studied by recording the FTIR spectra using a JASCO FT/IR 6800 spectrometer. An average of 256 scans at a maximum resolution of 4 cm-1 over a range of 4000-400 cm-1 was applied. Raman spectra were recorded with a Jobin-Yvon Xplora apparatus (Horiba Scientific) using a wavelength laser of 532 nm. In addition, the nature and amount of surface oxygenated groups were evaluated by temperature programmed desorption (TPD) carried out in an Altamira Instruments AMI-300 equipment connected to a mass spectrometer (Dycor Dymaxion). The samples (70 mg) were placed in a U-shaped quartz tube located inside an electrical furnace and heated up to 1100 °C. The heating rate and the flow rate of helium were fixed at 5 °C min-1 and 25 cm3 min-1, respectively. Finally, in order to quantify the amount of 5



CO2 and CO released during the TPD experiments, the calibration of these gases was performed at the end of each analysis. The wettability of the samples was measured by means of an Attension Theta contact angle meter. To this end, a water droplet was placed on the surface of the material with a glass micrometric syringe. The contact angle between the surface of the material and the water droplet was measured during 10 s, and registered in Young-Laplace analysis mode with the software One Attension. All measurements were made in triplicate to ensure reproducibility. It should be noted that this technique was applied to the final electrodes (active material with binder), and not to the powder sample, as the binder added to conform the pellets may influence the hydrophobic character of the electrode. 2.3. Electrode and cell preparation Disc-shaped pellets were prepared by mixing 90 wt.% of each sample and 10 wt.% of polytetrafluoroethylene (PTFE, 60 wt.% suspension in H2O, Sigma-Aldrich) as binder. Both compounds were hand-mixed using absolute ethanol as solvent until a homogeneous paste was obtained. The mixture was then rolled out in order to form a thin film from which the disc-shaped pellets were punched. In order to study the electrochemical behaviour and the supercapacitor performance of the prepared carbons, two types of cells were assembled: i) three-electrode and ii) twoelectrode configuration, respectively. Three-electrode configuration involves a beakertype cell with three electrodes (working, counter and reference) immersed in 1 M H2SO4 electrolyte solution. The working electrodes were prepared by sandwiching carbon pellets, with the same area (0.78 cm2) and similar masses (9-10 mg), between two stainless steel meshes and pressing to 5 t for 30 s. The counter electrode was prepared in the same manner as the working electrode but using Norit DLC Supra 50® as active material (pellets of 20 mg). Working and counter electrodes were stacked using filter paper as separator and the supporting meshes were connected to two Pt wires as current collectors. An Ag/AgCl (KCl 3 M) electrode was used as reference electrode. In the case of the two-electrode cell, two carbon pellets with similar masses (4-5 mg), which were prepared by pressing at 5 t in a mould (0.78 cm2) for 30 s, were used directly as selfstanding symmetric working electrodes. Two stainless steel bars (Hastelloy C276®, Goodfellow) were fixed with a PTFE Swagelok® tube fitting and used as current

6


Page 6 of 31

Page 7 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


collectors. The carbon electrodes were stacked between the steel bars and impregnated with 1 M H2SO4 electrolyte solution. A Whatmann® glass microfiber filter (Grade 934AH), soaked in 1 M H2SO4 electrolyte solution, was placed in-between, as separator. Regardless of the cell-configuration, all the electrodes were vacuum-degassed for 4 h at 120 ºC and then immersed into the electrolyte solution for 1 day before assembling the cell. Additional details of the assembling procedure for both cell configurations are described elsewhere 6. 2.4 Electrochemical measurements The electrochemical measurements were performed on a PGZ 402 Voltalab potentiostat (Radiometer Analytical) by using the three-electrode and two-electrode cell configurations described above. For the study of the electrochemical behaviour of the electrodes, i.e. three-electrode cell configuration, nitrogen gas was bubbled through the electrolyte for 30 min prior to the measurements to remove all O2 present in the electrolyte solution. Cyclic voltammetry (CV) experiments were carried out by varying the scan rates from 2 mV s-1 to 500 mV s-1 using a voltage window of 0.8 V (from -0.2 to 0.6 V). The specific gravimetric capacitance (F g-1) of a single electrode was calculated by equation 1. 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑎𝑛𝑐𝑒 =

1 𝑉1 2 𝑚𝑎𝑐𝑡 ∆𝑉 𝑟∫𝑉0𝐼(𝑉)𝑑𝑉

(1)

where mact is the mass of the active material in the working electrode (g), ∆V is the voltage window (V), r is the scan rate (V s-1) and I is the measured current (A). The supercapacitor behaviour of the carbon electrodes was also studied in a two-electrode cell. CV experiments were performed at a voltage sweep rate of 2 mV s-1 using different voltage windows from 0.6 V to 1.2 V. The specific gravimetric capacitance (F g-1) of the cell was calculated by equation 2. 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑎𝑛𝑐𝑒 =

2 𝑉1 𝑚 ∆𝑉 𝑟∫𝑉0𝐼(𝑉)𝑑𝑉

(2)

where m is the sum of the masses of the active material of the two electrodes (g), ∆V is the voltage window (V), r is the scan rate (V s-1) and I is the measured current (A).

7



Page 8 of 31

Electrochemical impedance spectroscopy (EIS) was also applied to the fully discharged cell at 0 V in the frequency region of 10 kHz to 10 mHz with AC amplitude of 10 mV. Charge-discharge (CD) experiments were conducted up to 0.8 V using different current intensities from 0.1 to 20 A g-1. The cycling stability was studied by means of consecutive charge/discharge cycles of the cell up to 1 A g-1 during 10,000 cycles. The capacitance (F g-1) value of one electrode was calculated from the galvanostatic charge-discharge plots according to equation 3. 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑎𝑛𝑐𝑒 =

2𝐼 𝑚𝑎𝑐𝑡 𝑑𝑉 𝑑𝑡

(3)

where I is fixed current (A), mact is the mass of the active material in the working electrode (g) and dV/dt is the slope of the discharge profile (V s-1). Cyclic voltammetry and charge-discharge experiments were also performed using a commercial activated carbon (Norit DLC Supra 50) as working electrode and under the same experimental conditions for comparison. 3. Results and discussion 3.1. Porosity, morphology and surface chemistry The nitrogen adsorption-desorption isotherms of the carbonized and activated carbons prepared using different amounts of carbon nanotubes, and the pore size distributions obtained by applying the QSDFT method to the adsorption branches of the isotherms, are plotted in Figure 1. Sample CG_0%CNT displays a type I isotherm according to the International Union of Pure and Applied Chemistry (IUPAC) classification, which is commonly given by microporous materials

34.

For carbonized samples, the volume of

nitrogen adsorbed at low relative pressures is practically unchanged by the addition of CNT, resulting in materials with similar microporosity. However, carbonized hybrids exhibit a hysteresis loop and an abruptly rise at high relative pressure (Figure 1a), which is characteristic of materials that bear mesopores in their interparticle spaces, like carbon nanotubes. The activation with CO2 of the HTC carbon produces a 5-fold increase of the surface area (Figure 1b, Table 1). In fact, the isotherm of the physically activated HTC

8


Page 9 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(AG_0%CNT) presents a large volume of nitrogen adsorbed at low relative pressures, which indicates the presence of a large volume of micropores (Table 1).

Figure 1. N2 adsorption/desorption isotherms at -196 ºC for (a) carbonized , (b) physically and (c) chemically activated carbons prepared using different amounts of carbon nanotubes and (d, e and f) their corresponding PSDs obtained from the adsorption branches of the isotherms by the QSDFT method . The microporous character of the carbonized samples is further confirmed by the PSD curves (Figure 1d). In the case of the CO2-activated HTC/CNT hybrids, the adsorption at low relative pressures decreases by increasing the amount of CNT, which results in materials with lower microporosity (Figure 1b). The surface area decreases from 2330 to 947 m2 g-1 by increasing the CNT content from 0 to 8 wt.% (Table 1). Similar to carbonized samples, the addition of CNT in the physically activated hybrids leads to capillary condensation at medium-high pressures, which is attributed to the presence of carbon nanotubes. The activation of the organic resins with KOH also increases the surface to a great extent (456 vs 2021 m2 g-1, Table 1). Thus, the adsorption of AG_0%CNT_KOH sample at low relative pressures is comparable to that of AG_0%CNT, which indicates also the presence of a large volume of micropores (Figure 1c, Table 1). However, the activation behaviour of the HTC/carbon nanotube hybrids when using KOH as activating agent greatly differs from that observed in the activation

9



under CO2. The physical activation causes a gradual decrease of the global porosity of the materials, while the chemical activation promotes the generation of mesopores in detriment of the micropore volume. The isotherms of the chemically activated hybrids are concave to the relative pressure axis (P/P0) and the amount of nitrogen adsorbed increases steeply at low pressures, indicating that the volume of micropores is still large enough. Table 1. Textural properties of carbonized and activated carbons Sample

SBET (m2 g-1)

VDR (cm3 g-1)

Vp (cm g-1)

CG_0%CNT

456

0.18

0.19

CG_2%CNT

444

0.18

0.28

CG_4%CNT

461

0.18

0.33

CG_8%CNT

448

0.17

0.55

AG_0%CNT

2330

0.99

1.03

AG_2%CNT

1951

0.85

1.00

AG_4%CNT

1078

0.43

0.64

AG_8%CNT

947

0.39

1.00

AG_0%CNT_KOH

2021

0.88

0.90

AG_2%CNT_KOH

1718

0.59

1.33

AG_4%CNT_KOH

1310

0.48

1.41

AG_8%CNT_KOH

1390

0.44

2.18

In addition, there are remarkable differences in the textural properties of the samples activated with KOH due to the addition of CNT (Figure 1c). Chemically activated carbons evolve from exclusively microporous materials to micro-mesoporous materials due to the increase in the amount of CNT from 0 wt.% to 8 wt.%. The volume of nitrogen adsorbed in samples containing CNT increases almost linearly for relative pressures from 0.1 to 0.5, and then abruptly rises at high relative pressure, which is characteristic of macroporous adsorbents 34. Even if N2 adsorption-desorption is not a precise technique for analysing samples containing macropores, it can be seen that the volume adsorbed at the saturation point increases with the presence of CNT (Table 1). Moreover, it is worth mentioning that the addition of amounts of CNT as low as 2 wt.% (AG_2%CNT_KOH) into the precursor solution causes a huge modification of the textural properties in the final activated carbon. Thus, a defined mesoporous system is readily developed while notably preserving the micropore volume (Figures 1c and 1f). As the content of CNT 10


Page 10 of 31

Page 11 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


increases from 2 to 8 wt.%, the hysteresis loop becomes narrower and shifts to higher relative pressures, indicating an increase in the pore size, which is confirmed by the pore size distributions shown in Figure 1f. However, differences in the hysteresis loop shapes are also observed. Sample AG_2%CNT_KOH exhibits a hysteresis loop similar to type H4, while the hysteresis loops of samples AG_4%CNT_KOH and AG_8%CNT_KOH seem more like a H3 type loop. In the latter, the adsorption branch looks like a type II isotherm, but a hysteresis loop appears. This pseudo-type II isotherm may be given by i) macroporous materials with pores not completely filled, or ii) structures composed of non-rigid aggregates of plate-like particles with slit-shaped pores 34. The observed evolution in the textural properties of both the physically and chemically activated samples can be explained by two alternative hypotheses: i) the shape of the isotherms evolves gradually from activated carbon to pure nanotube-type since the CNT ratio increases in a physical mixture or ii) the porous properties are modified as a consequence of changes either in the polymerization or the activation of glucose due to the presence of the CNTs. In order to delve deeper into these modifications produced by the addition of CNT and/or the different thermal treatments applied, the morphology of the samples was examined by scanning electron microscopy (SEM). Figure 2 shows representative SEM images for the carbonized and activated carbons.

11



Page 12 of 31

Figure 2. SEM images of carbonized HTC (a-d), physically activated HTC (e-h) and chemically activated HTC (i-l) prepared using different amounts of carbon nanotubes.

All the studied materials, with the exception of samples AG_4%CNT_KOH and AG_8%CNT_KOH, are mainly composed of spheres, which is a typical morphology of hydrothermal carbons

17, 26.

The exact growth mechanism of the carbon spheres is still

unclear, as a complex set of simultaneous chemical reactions, such as isomerization, dehydration, condensation or addition, occurs during the hydrothermal process

17, 24-25.

Nevertheless, it has been proposed that, under the appropriate hydrothermal conditions, dehydration of glucose is promoted by water at high temperature 27. Anions formed react with each other via polymerization reactions. Finally, aromatization reactions lead to condensed furan rings bridged by aliphatic regions, which probably results in materials with hydroxyl and carbonyl functional groups

25.

When the concentration of the

polymerized intermediates exceeds a critical concentration, the nucleation of the particles occurs 20. These molecules act as nucleation points from which the growth of the particles takes place, resulting in a spherical-shaped structure

24, 26.

Based on this mechanism, is

not surprising that the pure HTC carbon displays a homogeneous structure composed of 12


Page 13 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


spheres with an average size of 750 nm (Figure 2a). Nonetheless, the addition of CNT leads to inhomogeneous average size distributions of the carbon spheres. Smaller spheres of about 400 nm of diameter appear next to those of 750 nm in the carbonized hybrid materials (Figures 2b-d), due to the presence of CNT. Moreover, the number of the smallsized spheres increases by increasing the percentage of CNT from 0 wt.% to 8 wt.%. As well, the presence of CNT in the carbonized samples also modifies the defined spherical shape of the particles to more elongated plane-shapes. These changes in the morphology suggest that the incorporation of the CNT during the polymerization of glucose may increase the number of molecules that act as nucleation points, leading to the formation of large number of small spherical particles. A further examination of Figures 2a, 2e and 2i, also reveals differences in the size and shape of the spheres due to the activation processes. The physically activated sample (AG_0%CNT, Figure e) presents spheres with size similar to that of carbonized sample (~ 750 nm), while smaller spheres (~ 450 nm) are observed for the chemically activated sample (AG_0%CNT_KOH, Figure i). Moreover, the presence of CNT also results in different morphologies due to the activation process. Physical activation of HTC hybrids results in a more homogeneous structure in comparison with their carbonized counterparts. The activation yield increases with the proportion of carbon nanotubes (yield values are shown in Table S1 in the supporting information), suggesting that CO2 reacts preferentially with the glucose-derived particles, resulting in structures with spheres of smaller size. This effect is more evident for samples AG_4%CNT and AG_8%CNT, which show smooth morphologies (corresponding to CNT) supporting the small carbon spheres (~ 400 nm of diameter). Likewise, interesting changes in the morphology of the chemically activated samples bearing higher proportion of CNT were observed. The morphology of the samples evolved from sphere-shaped (Figure 2i-j), to platelet-like particles with slit-shaped pores (~ 400 nm, Figures 2k and 2l), which demonstrates that the hysteresis loops arising from isotherms plotted in Figure 1c are originated by structures composed of non-rigid aggregates. This phenomenon has already been observed in previous studies in which this type of voids was correlated to the fusion of KOH drops on the surface

19, 30

or by the formation of a liquid-phase intermediate

during activation 21. This effect is much more noticeable when the concentration of CNT is higher than 2 wt.%, suggesting that the effect of the chemical activation differs due to the presence of carbon nanotubes. This phenomenon may be related to the highly reactive

13



nature of the organic resins and the poor mechanical stability produced by the presence of the CNT during the hydrothermal treatment that may modify the colloidal stability of the pre-polymers formed in the solution or provide additional nucleation points. In order to bring some insight about these aspects, the surface of the hydrothermal carbons prior to carbonization or activation (set of samples G_X%CNT), was studied by FTIR spectrometry. The spectra shown in Figure 3a indicate that carbon nanotubes may somehow interact with the polymeric structure. Sample G_0%CNT presents a more intense band at around 3400 cm-1, which is attributed to O—H stretching vibration. The intensity of the peak decreases by increasing the percentage of CNT from 0 wt.% to 8 wt.%, suggesting that the incorporation of the CNT into the polymeric structure results in materials with lower concentration of -OH groups. Moreover, the peak at 1700 cm-1, which is attributed to carbonyl groups and the peaks in the absorption region of 1500 and 800 cm-1, assigned to C–O–C stretching and C–O–H bonds, respectively, become less intense by the addition of CNT. From these results, it can be inferred that sample G_0%CNT will probably have less amount of aromatic rings which may be linked by aliphatic chains.

Figure 3. FTIR spectra of organic HTC glucose-derived materials prior to carbonization or activation (a) and Raman spectra of carbonized glucose-derived materials (b) prepared using different amount of carbon nanotubes. To further correlate this statement with the observed activation trend, carbonized samples were analysed by Raman spectroscopy (Figure 3b). All samples exhibit two main bands at 1350 cm-1 and 1580 cm-1 attributed to D and G band. The D band is assigned to

14


Page 14 of 31

Page 15 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


disordered structures of turbostratic carbon layers or very small graphitic domains while the G band corresponds to the symmetric vibrational mode of graphitic structures. The ID/IG ratio calculated from the integration of the peaks of the different spectra increases by increasing the amount of CNT from CG_0%CNT to CG_8%CNT (the values are shown in Table S2 in the supporting information). As well, both bands shift to lower values of wave number as the percentage of CNT increases. This evolution of the peak parameters suggests that the addition of carbon nanotubes during the hydrothermal treatment of glucose leads to more disordered carbon materials. From the FTIR and Raman studies we can assume that carbon nanotubes play a role on the condensation behaviour of the furan rings generated during the HTC of glucose. The lower proportion of aliphatic regions verified in the organic hybrid materials leads to a more reactive structure towards CO2 and KOH, resulting in notable changes promoted by the different activating agents, in particular by chemical activation. Previous works have demonstrated that a meltdown of carbonaceous materials leads to a morphological rearrangement during hydroxide activations, especially for carbonaceous precursors with a low degree of crystallinity due to the formation of a liquid-phase intermediate

35.

This fact is

consistent with the results in Figure 3b, as carbons yielding low structural order (like those yielding larger proportion of carbon nanotubes), could be more prone to react with KOH 21, 35. The presence of CNT and the method of activation also modifies the final chemical properties of the carbonized and activated materials. As hydrothermal carbons are commonly composed by carbon and oxygen, the nature of the oxygen-containing functionalities was analysed by temperature programmed desorption experiments (TPD). The total amounts of CO (anhydrides, phenols, carbonyls) and CO2 (carboxylic acids, anhydrides, lactones) released were calculated from the corresponding TPD profiles and the values obtained are compiled in Table 2, along with the percentage of oxygen also calculated from TPD curves.

Table 2. Amounts of CO and CO2 released, obtained by integration of the areas under the TPD profiles and the percentage of oxygen content.

15



Page 16 of 31

Sample

CO

CO2

CO/CO2

O (wt. %)

CG_0%CNT

1391

286

4.86

3.1

CG_2%CNT

1116

233

4.79

2.5

CG_4%CNT

1077

214

5.02

2.4

CG_8%CNT

543

61

8.93

1.1

AG_0%CNT

720

261

2.75

2.0

AG_2%CNT

633

229

2.77

1.7

AG_4%CNT

625

183

3.41

1.6

AG_8%CNT

559

56

9.98

1.1

AG_0%CNT_KOH

3460

1263

2.74

9.6

AG_2%CNT_KOH

1943

1199

1.62

6.9

AG_4%CNT_KOH

1444

906

1.59

5.2

AG_8%CNT_KOH

1118

906

1.23

4.7

Regardless of the thermal treatment applied, carbonization or activation, the amount of oxygen determined by TPD decreases with the addition of carbon nanotubes. This trend is in agreement with the FTIR results shown in Figure 3 and the data of the elemental analysis shown in Table S3 in the supporting information. Nevertheless, large amounts of CO and CO2 were still released, as shown in Table 2. The highest degree of surface oxidation is observed in carbon materials after KOH chemical activation, followed by the carbonized samples and physically activated carbons (Table 2). All samples release larger amounts of CO-desorbing groups than those of CO2-desorbing groups. However, the trend of the CO/CO2 ratio differs as a function of the thermal treatment method applied. Carbonization and physical activation originate an increase of the CO/CO2 ratio due to the addition of CNT, while the ratio decreases by performing chemical activation. This suggests possible differences in the basic or acidic nature of the oxygen surface functionalities, which are further studied from the CO2 and CO desorption profiles shown in Figure 4. From these TPD profiles it can be inferred that the chemical surface functionalities of the HTC are modified by both the thermal treatment applied and the amount of CNT added.

16


Page 17 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. CO2 (a and b) and CO (c and d) TPD profiles of carbonized and chemically activated samples. Carbonized samples present similar CO2 profiles with three main peaks at temperature ranges of 200-400 ºC, 400-700 ºC and 500-800 ºC (Figure 4a), which can be attributed to carboxylic, anhydride and lactones groups, respectively

36.

As expected, similar peaks

attributed to the thermal decomposition of anhydride groups can be observed in the CO profiles (deconvolution of the CO2 and CO profiles for the carbonized samples are plotted in Figure S1). In addition to anhydrides, two main peaks are also observed (Figure 4c). The first one is centred at a 620 ºC and is ascribed to phenols, while the second peak appears at a temperature between 750 and 780 ºC, which is attributed to carbonyl/quinone groups. Although the TPD profiles for all carbonized samples are quite similar, some differences due to the addition of CNT can be highlighted: i) the temperature at which carbonyl/quinone groups are released increases with the amount of CNT and ii) the ratio between phenols and carbonyl/quinones decreases by increasing the percentage of CNT

17



from 0 to 8 wt.%. These results suggest an increase in the basicity of the surface due to the presence of carbon nanotubes. The nature of the oxygen functionalities of the samples physically activated is similar to those of carbonized samples (CO2 and CO desorption profiles are shown in Figure S2). However, some differences in the contribution of the functional groups are observed. The CO2 profiles of the activated samples exhibit less carboxylic acids and more lactones than carbonized samples, due to the higher temperature used during the activation (Figure S2a). Like in carbonized samples, the temperature at which carbonyl/quinone groups are released as CO increases with the addition of CNT, while the contribution of such functional groups decreases (Figure S2b). On the contrary, samples activated with KOH present CO2 profiles with only two broad peaks at temperature ranges of 200-400 ºC and 400-700 ºC which could be attributed to carboxylic and anhydride groups, respectively. In contrast, noticeable differences are found in the CO profiles (Figure 4d), suggesting that the variation in CO/CO2 ratio shown in Table 2 is probably due to the functional groups released as CO. As expected, similar peaks attributed to the thermal decomposition of anhydride groups can be observed at low temperatures in the CO profiles. However, at higher temperatures the most intensive evolution of CO was observed for sample AG_0%CNT_KOH, which shows a broad peak centred at 630 ºC commonly attributed to phenols. The CO profile for sample AG_2%CNT_KOH shows two peaks centred at 650 ºC and 820 ºC, attributed to phenols and carbonyls, respectively. As observed for samples AG_4%CNT_KOH and AG_8%CNT_KOH, the addition of higher percentage of carbon nanotubes not only reduces the amount of oxygenated groups, but also shifts the CO evolution to higher temperature. Once again, the decrease of the phenol/carbonyl ratio in the latter samples suggests a lower acidity of the surface by addition of CNT and a preferential activation of HTC carbon rather than the carbon nanotubes. The above commented differences in the surface chemistry of the samples are also reflected in their hydrophobic character. The hydrophilic/hydrophobic character of the different electrodes prepared from the carbonized and activated carbons was estimated by contact angle measurements. The contact angle, and hence the hydrophobic character of the sample, increases with the concentration of carbon nanotubes from 132º to 150º in carbonized samples, from 136º to 154º for physically activated carbons and from 122º to 147º in the case of chemically activated materials (more information regarding hydrophobicity of the samples can be found in Table S4 of the supporting information). 18


Page 18 of 31

Page 19 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


It should be noted that the values of the contact angle are lower for those samples activated with KOH, as the presence of surface oxygen-containing groups enhances the wettability of the carbons 37. However, it is important to highlight that the electronegative character of the oxygen functionalities may also decrease their electrical conductivity. It is therefore essential to find a compromise between the percentage of carbon nanotubes added and the thermal treatment applied (carbonization or activation) to obtain samples with high microporosity, an adequate wettability and a high electrical conductivity, as such features will result in materials with better electrochemical performance. 3.2. Electrochemical measurements The electrochemical behaviour of the prepared materials was investigated in a three electrode cell. Figure 5 shows the cyclic voltammograms for carbonized and activated samples bearing different proportions of CNT at a scan rate of 2 mV s-1. The global specific capacitances calculated from the cyclic voltammograms plotted in Figure 5 are listed in Table 3. Carbonized samples display highly resistive voltammograms with noticeable H2 evolution at low potential values (Figure 5a), which is indicative of aqueous electrolyte decomposition. A slight increase of the capacitance is observed for the sample bearing the highest proportion of carbon nanotubes (CG_8%CNT), although a capacitance of only 58 F g-1 is obtained. In contrast, activated samples (Figure 5b and 5c), yield almost rectangular curves, suggesting a capacitive and reversible behaviour (electric double-layer capacitor, EDLC) as well as current peaks deviating from linearity in the cathodic/anodic sweeps. The presence of those humps is attributed to fast faradaic redox processes that also contribute to the electrochemical charge storage in those materials (pseudocapacitance)

5, 12.

Redox peaks are related to electrochemically-active surface

oxygen functionalities which, in carbon materials, are generally associated with the quinone-hydroquinone redox couple 8. Previous reports have pointed out that both phenol and pyrone groups can also present faradaic activity at slightly lower potentials than quinones 9. The relationship between faradaic redox processes (Figure 5b) and the surface oxygen functionalities observed in TPD measurements (Figure S2) is quite evident for physically activated samples. AG_2%CNT displays well-resolved and intense redox humps ~ 0.27 V vs Ag/AgCl (KCl, 3 M) due to its large proportion of carbonyl-quinones. The fainter peaks shifting to lower potentials (~ 0.22V) observed for the AG_0%CNT electrode are ascribed to its larger proportion of phenol-like groups.

19



Figure 5. Cyclic voltammograms measured in a three-electrode cell at room temperature in 1 M H2SO4 at a scan rate of 2 mV s-1 for carbonized (a), physically (b) and chemically (c) activated carbon materials. Chemically activated samples deliver the highest capacitance values among all the studied materials, probably because of their rich surface chemistry (Figure 4d) and high surface area (Table 1). Like CO2 activation, the samples bearing 0 and 2 wt.% of CNT outperform those having larger proportions. The AG_2%CNT_KOH electrode exhibits anodic and cathodic peaks at around 0.30 V while the faradaic activity for sample AG_0%CNT_KOH shifts to lower potentials (0.24 V). This fact is in agreement with the TPD profile (Figure 4e), which showed a large release of CO-desorbing groups at higher temperature for AG_2%CNT_KOH (carbonyl-quinones) and large release of phenols for AG_0%CNT_KOH. Samples AG_4%CNT_KOH and AG_8%CNT_KOH do not seem to induce such pseudocapacitive charge storage due to their lower surface functionalization. From these results, it can be concluded that the incorporation of carbon nanotubes into the structure shifts the redox peaks to higher potentials, confirming a higher proportion of quinones as the proportion of carbon nanotubes increases in both physically and chemically activated samples.

20


Page 20 of 31

Page 21 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 3. Electrochemical parameters for the activated carbons extracted from cyclic voltammograms with three- and two-electrode configuration. Capacitance (F g-1) Sample

CVa (3E)

CVb (2E)

CG_0%CNT

29

27

CG_2%CNT

32

29

CG_4%CNT

45

41

CG_8%CNT

58

57

AG_0%CNT

149

139

AG_2%CNT

196

190

AG_4%CNT

105

100

AG_8%CNT

88

84

AG_0%CNT_KOH

239

218

AG_2%CNT_KOH

230

220

AG_4%CNT_KOH

129

121

AG_8%CNT_KOH

124

117

a

Calculated from equation 1.

b

Calculated from equation 2.

The supercapacitor performance of the electrodes was studied in a two-electrode cell and the global specific capacitances calculated from the cyclic voltammograms are listed in Table 3. Since chemically activated samples deliver the highest capacitance values among all the studied materials, only the most relevant results obtained for these samples in a two-electrode cell are shown below. The supercapacitor performance of carbonized and physically activated samples can be found in the supporting information. The cyclic voltammograms obtained for chemically activated samples are plotted in Figure 6a. As previously

observed

in

the

three-electrode

measurements

(Figure

5),

the

AG_0%CNT_KOH and AG_2%CNT_KOH electrodes were those showing the highest performance. This is attributed to their larger surface areas (Table 1) and pseudocapacitive activities (Figure 5c). It should be noted that the addition of a small quantity of CNT originates a noticeable change in the resistance, either molecular or electronic. Hence, the profile of the cyclic voltammetry curves changes readily from AG_0%CNT_KOH to AG_2%CNT_KOH, which displays highly rectangular shape at

21



the limit potentials. Figure 6a also shows the cyclic voltammogram obtained for the commercial activated carbon (Supra 50), which is widely used in supercapacitor devices. As can be seen, the commercial material exhibits a capacitance value (122 F g-1) lower than those of samples AG_0%CNT_KOH to AG_2%CNT_KOH (239 and 230 F g-1) and similar to those of samples AG_4%CNT_KOH and AG_8%CNT_KOH (121 and 117 F g-1). However, the commercial carbon yields the highest resistive behaviour, pointing out the positive effect of the CNT in the as-synthesized electrode hybrid materials in terms of high electrical conductivity by itself. Hence, the addition of conductive additives in the electrode formulation, a common strategy used in a large number of publications, is not necessary to decrease the resistive behaviour of biomass-derived carbons.

Figure 6. Cyclic voltammograms measured in a two-electrode cell at room temperature in 1M H2SO4 at a scan rate of 2 mV s-1 (a), galvanostatic charge/discharge profiles at a current density of 0.2 A g-1 (b) for chemically activated carbon materials. Similar tendencies are observed in the galvanostatic charge/discharge profiles shown in Figure 6b. The chemically activated sample (AG_0%CNT_KOH) presents similar capacitance as AG_2%CNT_KOH. In addition, a further increment above 2 wt. % in the proportion of CNT causes a dramatic drop of the capacitance. This effect was also observed for physically activated samples (Figure S4). However, it is important to note that the IR drop decreases gradually with the incorporation of CNT both for physically and activated samples, and hence, samples bearing the highest capacitance are also those displaying the largest resistance. In order to further understand the differences in the resistance and capacitive behaviour due to the addition of CNT, electrochemical impedance spectroscopy (EIS) measurements were also performed (Figure 7).

22


Page 22 of 31

Page 23 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The cell resistance (Rs) of the sample CG_0%CNT decreases from 0.66 Ω to 0.25 Ω due to the addition of just 2 wt.% of CNT. A further increase in the proportion of CNT, results in a slightly increase of the cell resistance, which could be explained by the hydrophobic character of the samples. Moreover, samples CG_0%CNT and CG_2%CNT do not exhibit the common semicircle observed at medium-high frequencies suggesting an excessive internal resistance (Rp) that prevents the transport of electrons through the structure. The addition of amounts of carbon nanotubes larger than 2 wt.% increases drastically the electronic and molecular conduction, and hence, higher capacitances are obtained (Figure 7a).

Figure 7. Nyquist plots obtained from electrochemical impedance spectroscopy measurements in a two-electrode cell for carbonized (a), physically activated (b) and chemically activated (c) carbons. Among the physically activated samples, the AG_0%CNT and AG_2%CNT electrodes exhibit the highest resistive behaviour (Figure 7b), probably caused by their high surface area (Table 1). The presence of pores in high surface area activated carbons can prevent the transport of electrons through the aromatic structure due to the presence of structural defects. Moreover, at medium frequencies both samples display diffusion limitations due to electron-trapping functional groups on their surface (Table 2). This phenomenon can be confirmed by the behaviour of the sample AG_8%CNT, which shows the lower diffusional resistance and the lower oxygen content. However, increasing the amount of CNT from 2 wt.% to 8 wt.% results in a larger electronic resistance (Rp) due to the higher hydrophobic character of the samples. These results confirm the need to find the optimum concentration of CNT to be added in the precursor solution. In this context, it is interesting to highlight the capacitive behaviour of sample AG_2%CNT at low frequencies, which is comparable to that of the sample with the highest proportion of CNT, suggesting that the addition of a small percentage of CNT is enough to improve the electrochemical 23



performance of physically activated HTC. This phenomenon is also observed with the KOH activated samples. Indeed, sample AG_2%CNT_KOH exhibits a similar galvanostatic charge/discharge profile as that of sample AG_2%CNT (capacitances of 214 and 227 F g-1 are obtained for sample AG_2%CNT and AG_2%CNT_KOH, respectively, calculated at 0.2 A g-1). However, comparing the IR drop of both activated samples bearing 2 wt.% of carbon nanotubes, significant differences are observed as a result of their textural and chemical properties, especially due to the contribution of the carbonyl/quinone functional groups. Nyquist plots in Figure 7c show that chemically activated samples present similar cell resistance (Rs) but different internal resistance (Rp) at medium-high frequencies. Sample AG_0%CNT_KOH yields the highest resistive behaviour, probably caused by its high surface area (Table 1) and the large proportion of electron-trapping functional groups on its surface (Table 2). The microporous nature of samples AG_0%CNT_KOH and AG_2%CNT_KOH implies diffusion limitations that are not observed in the electrodes having higher proportions of carbon nanotubes, which present a well-developed meso-macroporous system acting as diffusion channels, as shown in Figure 1f. At low frequencies, samples bearing CNT display a nearly vertical line indicating preferential formation of the EDCL as compared to their counterpart without CNT, which could be explained by a better compromise between adequate porosity and chemical surface speciation. It can be therefore concluded that the addition of CNT during the hydrothermal treatment of glucose increases both the electronic (less structural defects) and the molecular (creation of mesopores) transportation in the final activated carbon materials. As a consequence, the slope of the charge/discharge profile increases linearly with time for samples AG_4%CNT_KOH and AG_8%CNT_KOH, while a slight deviation from the triangular form is observed for samples AG_0%CNT_KOH and AG_2%CNT_KOH at potentials lower than 0.3 V. This fact is consistent with the observation of faradaic current intensities in the cyclic voltammetry curves (Figure 5c). It can be therefore concluded that, as observed in the three-electrode characterization, non-faradaic processes are the main contribution to the total charge storage when adding a concentration of carbon nanotubes higher than 2 wt.%, while samples bearing no or low proportion of nanotubes integrated in their structure present both electrostatic and faradaic charge storage. As a result, AG_0%CNT_KOH and AG_2%CNT_KOH yield higher capacitances at low operation rates (sweep rate and current intensity). However, not only the maximum capacity provided for each electrode material is a key factor, but also its capacitance retention with increasing the current. This 24


Page 24 of 31

Page 25 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


point is of major relevance for real implementation of materials as supercapacitor electrodes, since the most remarkable feature of these devices is the fast release of energy. In addition, increasing the energy density by tuning the textural and chemical properties is only desirable if the power density is maintained. In this context, charge-discharge tests up to 0.8 V using current densities from 0.1 to 20 A g-1 were also carried out. The variation of the specific capacitances of the chemically activated carbons calculated from the charge/discharge cycles versus the working current densities (from 0.1 to 20 A g-1) is plotted in Figure 8a. It should be noted that the commercial carbon Supra 50 is also plotted but just up to a current density of 2 A g-1, as a further increase in the current density led to a significant increase of the IR drop and a great loss of capacitance (Figure S5). The variation of the specific capacitances of the carbonized and physically activated carbons calculated from the charge/discharge cycles versus the working current densities (from 0.1 to 20 A g-1) were also calculated and are plotted in Figure S6.

Figure 8. Capacitance retention of the chemically activated carbons measured within a voltage window of 0.8 V (a) and cyclability test for sample AG_2%CNT_KOH at 1 A g-1 for 5000 and 10000 cycles (b). The specific capacitance of samples AG_0%CNT_KOH and AG_2%CNT_KOH is higher than that of samples AG_4%CNT_KOH and AG_8%CNT_KOH at lower current density, probably due to their higher surface area (Table 1). However, even for current densities as low as 5 A g-1, the capacitance of sample AG_0%CNT_KOH decays more than 50 %. In contrast, samples bearing CNT proportions from 4 to 8 wt.% present capacitance retentions up to 85 % in the studied intensity range. Indeed, the higher the amount of CNT incorporated in the porous structure, the lower the decay of capacitance 25



at high current densities (22 %, 13 % and 10 % at 20 A g-1 for samples AG_2%CNT_KOH, AG_4%CNT_KOH and AG_8%CNT_KOH, respectively). These data demonstrate that to maintain the capacitance of the activated carbons obtained from glucose it is necessary to incorporate CNT into the structure. The capacitance retention for materials presenting pseudocapacitive charge storage is lower than that of electrode materials performing pure double layer charge storage. However, in the case of samples chemically activated, it should be noted that sample AG_2%CNT_KOH presents a remarkable capacitance retention (78%) in spite of exhibiting pseudocapacitive storage. These results demonstrate the importance of achieving a compromise regarding the amount of carbon nanotubes that should be added in order to obtain materials with high surface area, well-developed mesoporosity, appropriate surface oxygen functionalities and high electrochemical capacitance and stability. Among all the samples studied, sample AG_2%CNT_KOH seems to be highly promising as electrode material for supercapacitors. Accordingly, the electrochemical performance of this sample was evaluated by performing the cyclability test at 1 A g-1. The galvanostatic charge/discharge profiles obtained after 5000 and 10000 cycles are shown in Figure 8b. A slight increase in the IR drop and a less triangular form is observed after 5000 cycles, although the specific capacitance only fades by 3 % and 25 %, after 5000 (205 F g-1) and 10000 (159 F g-1) cycles, respectively. These results confirm that the addition of just 2 wt.% of carbon nanotubes is enough to improve the electrochemical performance of low-cost biomass derived carbons. Conclusions Activated glucose derived carbon/carbon nanotube hybrids were prepared by hydrothermal synthesis and subsequent carbonization and CO2 and KOH activation. The effect of the addition of nanotubes as conductive additive was investigated by varying their concentration in the precursor solution during the hydrothermal polymerization of glucose. The effect of this parameter on the physicochemical properties and the supercapacitor performance of the final activated materials was thoroughly investigated. It was concluded that the presence of CNT leads to significant changes in the textural, chemical, morphological and electrochemical properties of the materials. Interestingly, it was observed that chemically activated samples with proportions of carbon nanotubes above 2 wt.% developed mesoporous systems and did not present the sphere-like

26


Page 26 of 31

Page 27 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


morphology commonly observed for hydrothermal carbons. In contrast, they display solid and continuous structures yielding lower surface areas and oxygen contents. Regardless of the thermal treatment applied, carbonization or activation, an increase in the proportion of carbon nanotubes also causes changes in the surface oxygen speciation, which becomes enriched in more basic functional groups. Regarding their application as electrode materials in supercapacitors, activated carbons without carbon nanotubes present noticeable pseudofaradaic charge storage, but yield very poor capacitance retention as the current density is increased. The presence of an amount of carbon nanotubes as low as 2 wt.% constitutes an optimum balance between high specific capacitances and high capacitance retentions at high current densities, especially for those samples chemically activated. Samples having larger proportions of nanotubes, i.e. 4 and 8 wt.%, present higher electronic transportation and good capacitance retention, but lower capacitances. Therefore, it can be concluded that the addition of small quantities of nanotubes greatly enhances the electrochemical properties of low-cost biomass derived carbon electrodes, largely exceeding the supercapacitor performance of the commercial activated carbon Supra 50. Supporting Information Supporting Information available: Yields of the successive steps involved in the preparation of the samples; integrated areas of the peaks D and G obtained by Raman spectroscopy; Elemental and proximate analyses of carbonized and activated carbons; deconvolution of the CO2 and CO profiles for the carbonized samples; TPD profiles of physically activated samples; the value of the contact angle of carbonized and activated samples, the cyclic voltammograms measured in a two-electrode cell for the carbonized and physically activated samples; galvanostatic charge/discharge profiles at a current density of 0.2 A g-1 for the carbonized and physically activated samples and at 3 A/g for sample AG_2%CNT_KOH; and capacitance retention for the carbonized and physically activated samples.

Acknowledgements This work was financed by Project “AIProcMat@N2020 - Advanced Industrial Processes and Materials for a Sustainable Northern Region of Portugal 2020”, with the reference

27



NORTE-01-0145-FEDER-000006, supported by Norte Portugal Regional Operational Programme (NORTE 2020), under the PORTUGAL 2020 Partnership Agreement, through the European Regional Development Fund (ERDF), and by Project POCI-010145-FEDER-006984 - Associate Laboratory LSRE-LCM funded by FEDER through COMPETE2020-Programa Operacional Competitividade e Internacionalização (POCI) and by national funds through FCT - Fundação para a Ciência e a Tecnologia. The authors are indebted to Rui Rocha (CEMUP) for assistance with SEM analyses and Dr. A. Guedes (Departamento Geociências, Ambiente e Ordenamento do Território da FCUP), Dr. C. Pereira (Departamento de Química e Bioquímica da FCUP) and Dr. A. M. Pereira (Departamento de Física e Astronomia da FCUP) for assistance with electrical conductivity and Raman analyses.

References 1. Chen, X.; Paul, R.; Dai, L., Carbon-based Supercapacitors for Efficient Energy Storage. Natl. Sci. Rev. 2017, 4 (3), 453-489. 2. Gao, Z.; Zhang, Y.; Song, N.; Li, X., Biomass-derived Renewable Carbon Materials for Electrochemical Energy Storage. Mater. Res. Lett. 2017, 5 (2), 69-88. 3. Canal-Rodríguez, M.; Arenillas, A.; Rey-Raap, N.; Ramos-Fernández, G.; MartínGullón, I.; Menéndez, J. A., Graphene-doped Carbon Xerogel Combining High Electrical Conductivity and Surface Area for Optimized AqueousSupercapacitors. Carbon 2017, 118, 291-298. 4. Tooming, T.; Thomberg, T.; Kurig, H.; Jänes, A.; Lust, E., High Power Density Supercapacitors Based on the Carbon Dioxide Activated d-glucose Derived Carbon Electrodes and 1-ethyl-3-methylimidazolium Tetrafluoroborate Ionic Liquid. J. Power Sources 2015, 280 (Supplement C), 667-677. 5. Zhang, L. L.; Zhao, X. S., Carbon-based Materials as Supercapacitor Electrodes. Chem. Soc. Rev. 2009, 38 (9), 2520-2531. 6. Enterría, M.; Pereira, M. F. R.; Martins, J. I.; Figueiredo, J. L., Hydrothermal Functionalization of Ordered Mesoporous Carbons: The effect of Boron on Supercapacitor Performance. Carbon 2015, 95, 72-83. 7. Song, Y.; Liu, T.; Qian, F.; Zhu, C.; Yao, B.; Duoss, E.; Spadaccini, C.; Worsley, M.; Li, Y., Three-dimensional Carbon Architectures for Electrochemical Capacitors. J. Colloid Interface Sci. 2018, 509 (Supplement C), 529-545. 8. Hulicova-Jurcakova, D.; Seredych, M.; Lu, G. Q.; Bandosz, T. J., Combined Effect of Nitrogen- and Oxygen-Containing Functional Groups of Microporous Activated Carbon on its Electrochemical Performance in Supercapacitors. Adv. Funct. Mater. 2009, 19 (3), 438-447. 9. Raymundo-Piñero, E.; Leroux, F.; Béguin, F., A High-Performance Carbon for Supercapacitors Obtained by Carbonization of a Seaweed Biopolymer. Adv Mater 2006, 18 (14), 1877-1882. 10. Fan, X.; Yu, C.; Ling, Z.; Yang, J.; Qiu, J., Hydrothermal Synthesis of PhosphateFunctionalized Carbon Nanotube-Containing Carbon Composites for Supercapacitors with Highly Stable Performance. ACS Appl. Mater. Interfaces 2013, 5 (6), 2104-2110 28


Page 28 of 31

Page 29 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


11. Liang, J.-Y.; Wang, C.-C.; Lu, S.-Y., Glucose-derived Nitrogen-doped Hierarchical Hollow Nest-like Carbon Nanostructures from a Novel Template-free Method as an Outstanding Electrode Material for Supercapacitors. J. Mater. Chem. A 2015, 3 (48), 24453-24462. 12. Calvo, E. G.; Rey-Raap, N.; Arenillas, A.; Menéndez, J. A., The Effect of the Carbon Surface Chemistry and Electrolyte pH on the Energy Storage of Supercapacitors. RSC Adv. 2014, 4 (61), 32398. 13. Shi, X.; Zheng, S.; Wu, Z.-S.; Bao, X., Recent Advances of Graphene-based Materials for High-performance and New-concept Supercapacitors. J. Energy Chem. 2018, 27 (1), 25-42. 14. Sun, L.; Wang, X.; Wang, Y.; Zhang, Q., Roles of Carbon Nanotubes in Novel Energy Storage Devices. Carbon 2017, 122 (Supplement C), 462-474. 15. Zheng, S.; Wu, Z.-S.; Wang, S.; Xiao, H.; Zhou, F.; Sun, C.; Bao, X.; Cheng, H.-M., Graphene-based Materials for High-voltage and High-energy Asymmetric Supercapacitors. Energy Storage Mater. 2017, 6 (Supplement C), 70-97. 16. Li, W.; Liu, J.; Zhao, D., Mesoporous Materials for Energy Conversion and Storage Devices. Nat. Rev. Mater. 2016, 1 (6). 17. Sevilla, M.; Fuertes, A. B., Chemical and Structural Properties of Carbonaceous Products Obtained by Hydrothermal Carbonization of Saccharides. Chemistry - A European Journal 2009, 15 (16), 4195-4203. 18. Zhao, L.; Fan, L. Z.; Zhou, M. Q.; Guan, H.; Qiao, S.; Antonietti, M.; Titirici, M. M., Nitrogen-containing Hydrothermal Carbons with Superior Performance in Supercapacitors. Adv Mater 2010, 22 (45), 5202-5206. 19. Martín-Jimeno, F. J.; Suárez-García, F.; Paredes, J. I.; Martínez-Alonso, A.; Tascón, J. M. D., Activated Carbon Xerogels with a Cellular Morphology Derived from Hydrothermally Carbonized Glucose-graphene Oxide Hybrids and their Performance Towards CO2 and Dye Adsorption. Carbon 2015, 81 (1), 137-147. 20. Li, M.; Li, W.; Liu, S., Hydrothermal Synthesis, Characterization, and KOH Activation of Carbon Spheres from Glucose. Carbohydr. Res. 2011, 346 (8), 999-1004. 21. Falco, C.; Sieben, J. M.; Brun, N.; Sevilla, M.; Van Der Mauelen, T.; Morallón, E.; Cazorla-Amorós, D.; Titirici, M. M., Hydrothermal Carbons from HemicelluloseDerived Aqueous Hydrolysis Products as Electrode Materials for Supercapacitors. ChemSusChem 2013, 6 (2), 374-382 22. Park, S. K.; Kwon, S. H.; Lee, S. G.; Choi, M. S.; Suh, D. H.; Nakhanivej, P.; Lee, H.; Park, H. S., 105 Cyclable Pseudocapacitive Na-Ion Storage of Hierarchically Structured Phosphorus-Incorporating Nanoporous Carbons in Organic Electrolytes. ACS Energy Letters 2018, 3 (3), 724-732. 23. Park, S. K.; Lee, H.; Choi, M. S.; Suh, D. H.; Nakhanivej, P.; Park, H. S., Straightforward and Controllable Synthesis of Heteroatom-doped Carbon Dots and Nanoporous Carbons for Surface-confined Energy and Chemical Storage. Energy Storage Mater. 2018, 12, 331-340. 24. Titirici, M. M., Hydrothermal Carbons: Synthesis, Characterization, and Applications. In Novel Carbon Adsorbents, Elsevier Ltd: 2012; pp 351-399. 25. Li, R.; Shahbazi, A., A Review of Hydrothermal Carbonization of Carbohydrates for Carbon Spheres Preparation. Trends Renew. Energy 2015, 1 (1), 43-56. 26. Falco, C.; Marco-Lozar, J. P.; Salinas-Torres, D.; Morallón, E.; Cazorla-Amorós, D.; Titirici, M. M.; Lozano-Castelló, D., Tailoring the Porosity of Chemically Activated Hydrothermal Carbons: Influence of the Precursor and Hydrothermal Carbonization Temperature. Carbon 2013, 62, 346-355.

29



27. Fang, Z.; Smith Jr, R. L.; Kozinski, J. A.; Minowa, T.; Arai, K., Reaction of d-Glucose in Water at High Temperatures (410 °C) and Pressures (180 MPa) for the Production of Dyes and Nano-particles. J. Supercrit. Fluids 2011, 56 (1), 41-47. 28. Lee, K. K.; Hao, W.; Gustafsson, M.; Tai, C.-W.; Morin, D.; Bjorkman, E.; Lilliestrale, M.; Bjorefors, F.; Andersson, A. M.; Hedin, N., Tailored Activated Carbons for Supercapacitors Derived from Hydrothermally Carbonized Sugars by Chemical Activation. RSC Adv. 2016, 6 (112), 110629-110641. 29. Sevilla, M.; Yu, L.; Ania, C. O.; Titirici, M.-M., Supercapacitive Behavior of Two Glucose-Derived Microporous Carbons: Direct Pyrolysis versus Hydrothermal Carbonization. ChemElectroChem 2014, 1 (12), 2138-2145. 30. Enterría, M.; Martín-Jimeno, F. J.; Suárez-García, F.; Paredes, J. I.; Pereira, M. F. R.; Martins, J. I.; Martínez-Alonso, A.; Tascón, J. M. D.; Figueiredo, J. L., Effect of Nanostructure on the Supercapacitor Performance of Activated Carbon Xerogels Obtained from Hydrothermally Carbonized Glucose-graphene OxideHybrids. Carbon 2016, 105, 474-483. 31. Zhang, C.; Tjiu, W. W.; Liu, T., One-pot Hydrothermal Synthesis and Reusable Oiladsorbing Properties of Porous Carbonaceous Monoliths Using Multi-walled Carbon Nanotubes as Templates. RSC Adv. 2013, 3 (35), 14938-14941. 32. Mahmood, Q.; Yun, H. J.; Kim, W. S.; Park, H. S., Highly Uniform Deposition of MoO3 Nanodots on Multiwalled Carbon Nanotubes for Improved Performance of Supercapacitors. J. Power Sources 2013, 235, 187-192. 33. Enterría, M.; Gonçalves, A. G.; Pereira, M. F. R.; Martins, J. I.; Figueiredo, J. L., Electrochemical Storage Mechanisms in Non-stoichiometric Cerium Oxide/multiwalled Carbon Nanotube Composites. Electrochim. Acta 2016, 209, 25-35. 34. Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouquerol, J.; Siemieniewska, T., Reporting Physisorption Data for gas/solid Systems with special Reference to the Determination of Surface Area and Porosity (Recommendations 1984). In Pure Appl. Chem., 1985; Vol. 57, p 603. 35. Lillo-Ródenas, M. A.; Juan-Juan, J.; Cazorla-Amorós, D.; Linares-Solano, A., About Reactions Occurring During Chemical Activation with Hydroxides. Carbon 2004, 42 (7), 1371-1375. 36. Figueiredo, J. L.; Pereira, M. F. R.; Freitas, M. M. A.; Órfão, J. J. M., Modification of the Surface Chemistry of Activated Carbons. Carbon 1999, 37 (9), 1379-1389. 37. Canal-Rodríguez, M.; Menéndez, J. A.; Arenillas, A., Performance of Carbon Xerogel-graphene Hybrids as Electrodes in Aqueous Supercapacitors. Electrochim. Acta 2018, 276, 28-36.

30


Page 30 of 31

Page 31 of 31 1 2 3 4 5 6 7



Influence of Multiwalled Carbon Nanotubes as Additives in Biomass

Recommend Documents