Nanocomposite composed of multiwall carbon nanotubes covered by hematite nanoparticles as anode material for Li-ion batteries

Electrochimica Acta 228 (2017) 82–90

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Nanocomposite composed of multiwall carbon nanotubes covered by hematite nanoparticles as anode material for Li-ion batteries Marcin Krajewskia,* , Po-Han Leeb , She-Huang Wub , Katarzyna Brzozkac , Artur Malolepszyd , Leszek Stobinskid,e , Mateusz Tokarczykf , Grzegorz Kowalskif , Dariusz Wasikf a

Institute of Fundamental Technological Research, Polish Academy of Sciences, Pawinskiego St. 5B, 02-106 Warsaw, Poland Department of Materials Engineering, Tatung University, Taipei, 104 Taiwan, ROC c Faculty of Mechanical Engineering, Department of Physics, University of Technology and Humanities in Radom, Krasickiego St. 54, 26-600 Radom, Poland d Faculty of Chemical and Process Engineering, Warsaw University of Technology, Warynskiego St. 1, 00-645 Warsaw, Poland e Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka St. 44/52, 01-224 Warsaw, Poland f Faculty of Physics, Institute of Experimental Physics, University of Warsaw, Pasteura St. 5, 02-093 Warsaw, Poland b

A R T I C L E I N F O

Article history: Received 19 August 2016 Received in revised form 15 December 2016 Accepted 9 January 2017 Available online 11 January 2017 Keywords: anode material hematite Li-ion battery multiwall carbon nanotube nanocomposite

A B S T R A C T

This work describes the detailed studies performed on the nanocomposite composed of chemicallymodified multiwall carbon nanotubes covered by hematite nanoparticles which diameters vary from 10 nm to 70 nm. This nanomaterial was fabricated in two-steps facile chemical synthesis and was characterized with the use of several experimental techniques, such as: thermogravimetric analysis, differential thermal analysis, Raman spectroscopy, X-ray diffraction, and transmission Mössbauer spectroscopy in order to determine its structure precisely. Moreover, the investigated nanocomposite was tested as an anode material of Li-ion batteries. Its cycling performance was stable during 40 cycles, while its capacity was retained at the level of 330 and 230 mAh/g at the discharge/charge rate of 25 and 200 mA/g, respectively. © 2017 Elsevier Ltd. All rights reserved.

1. Introduction At present, a development of portable technologies is very fast and there is no sign of slowdown in this field. However, all sorts of mobile devices need to have an adequate source of energy to work properly. Therefore, an essential issue is the energy storage, usually in a form of electricity. This can be achieved by the electricity conversion into the chemical energy which is also easy to accumulate [1]. Thereby, such conversion process is commonly used in the case of different energy storage systems, including supercapacitors, fuel cells as well as Li-ion batteries. Nevertheless, so far the last mentioned energy storage device has only been successfully applied in many portable technologies, such as: laptops, smartphones, tablets, pacemakers for artificial hearts, etc., but can be also used as an electricity storage devices for electric

* Corresponding author. E-mail addresses: [email protected], [email protected] (M. Krajewski). http://dx.doi.org/10.1016/j.electacta.2017.01.051 0013-4686/© 2017 Elsevier Ltd. All rights reserved.

vehicles (EVs) [2]. Among all of applications, the Li-ion batteries require high specific energy, high specific power, broad working temperature range, very low self-discharge and long life [3–5]. In general, Li-ions batteries are composed of the three main components: negative electrode (anode), positive electrode (cathode) and an electrolyte [2,6]. To improve the electrochemical performance of this type of batteries, the studies on each of these elements are important. However, up till now the most crucial issue in the case of Li-ion batteries improvement seems to be a search of new and high-performance anode materials which could replace the commercially applied graphite electrodes [4–7]. This is mainly caused by the fact that the graphitic anode cannot reach a capacity higher than 372 mAh/g and this capacity limit is almost achieved in some currently used commercial graphite electrodes [8]. However, it is not common phenomenon. Therefore, there is a vast amount of publications where the new potential anode materials are shown as the alternative anodes for Li-ion batteries. Usually, they are classified into three groups: the insertion-type materials (e.g. graphite and other carbon-based materials,

M. Krajewski et al. / Electrochimica Acta 228 (2017) 82–90

83

Li4Ti5O12, TiO2), the conversion-type materials (e.g. cobalt oxides, iron oxides, nickel oxides), and the alloying-type materials (e.g. Sn, Si, Ge, Al) [2,5,9,10]. Their simplified lithium storage mechanisms are presented below:
Insertion-type material MOy + x Li+ @ LixMOy

(1)

where MOy  TiO2 or carbon nanotubes
Conversion-type material MO + 2 Li+ @ M + Li2O

(2)

where MO  transition metal oxide
Alloying-type material M + Li+ @ LiM

(3)

where LiM  alloy As usual, each of anode materials has its own advantages and drawbacks. For instance, the lithium storage in the case of insertion-type materials is based on insertion (intercalation) of Li+ ions into the electrode structure during the battery charging and their deinsertion during the battery discharging [2,5,6]. Hence, this type of materials exhibits a low expansion coefficient during the process of lithium storage. At the same time, the insertion-type materials usually display much lower values of specific capacities than other groups of anode materials. On the other hand, most of conversion- and alloying-types of materials suffer for a low electric conductivity and they reveal the high expansion coefficients due to formation of metals covered by lithium oxides or alloys (see presented mechanisms). This is a main reason of a physical crumbling of electrode which is often called pulverization [2,6]. This process aims to a progressive irreversible degradation of electrode even during the first few discharge/charge cycles [11,12]. Hence, an idea of nanocomposite materials preparation, which can merge the advantages of different anode materials, has been proposed by the research community. Although in recent years plenty of publications have appeared showing different possible anode materials, the new carbon-based nanocomposites are still one of the most frequently studied in the case of Li-ion batteries [5–8]. Undoubtedly, this is related to several issues, including: the enhancement of electronic conductivity of anode, a better diffusion of Li+ ions into electrode and the preservation of a morphological stability during battery cycling. Moreover, it is worth noting that the carbon-based nanocomposite materials usually combine the insertion mechanisms of lithium storage with conversion or alloying mechanisms [6]. All of these

Fig. 2. TGA-DTA curves of MWCNTs-COONH4 (MWCNTs) and MWCNTs-COONH4 coated by iron oxide nanoparticles (FexOy-MWCNTs) collected in the artificial air.

features lead to the significant improvement of battery capacity as well as cycleability. Among the carbon allotropes, which can be applied in Li-ion batteries as a nanocomposite anode material, carbon nanotubes (CNTs) seem to be one of the most suitable. They exhibit a high tensile strength (50  200 GPa [13]) as well as a high electronic conductivity (102  106 S/cm [5]), and at the same time they are relatively chemically inactive. These properties provide a good enough conductivity and cause that carbon nanotubes can act as a damper which reduce greatly the electrode expansion during the lithium storage in the form of metal or alloy [5,6]. Therefore, CNTs are the appropriate candidate for being the support matrixes in the case of nanocomposite anodes in Li-ion batteries. It is known that the nanostructured transition metal oxides, in particular iron oxides, satisfy the conversion-type anode material criteria. Moreover, they can be used in order to form the carbonbased nanocomposites through various chemical procedures. Among all of them, hematite (a-Fe2O3) is the most thermodynamically stable one and at the same time this oxide is easily accessible, low-cost, and environmentally-friendly [8,26]. Another benefit of a-Fe2O3 is its high theoretical capacity which equals 1007 mAh/g [7,8,26]. These features allow identifying this iron oxide as a promising material for the application in Li-ion batteries. Considering all of above discussed issues, in this work a preparation route and an application of nanocomposite is described, that is composed of chemically-modified multiwall carbon nanotubes (MWCNTs) covered by randomly-deposited nanoparticles of hematite (a-Fe2O3) which fulfils all requirements

Fig. 1. STEM images of (a) MWCNTs-COONH4, (b) TEM mode and corresponding (c) SEM mode of MWCNTs-COONH4 coated by iron oxide nanoparticles.

84

M. Krajewski et al. / Electrochimica Acta 228 (2017) 82–90

to become an alternative anode material for Li-ion batteries. This nanocomposite links the properties of insertion-type with conversion-type materials, which prevent the possible electrode pulverization. Moreover, a remarkable advantage of the applied fabrication pathway of investigated nanocomposite is its possible usage for a large scale industrial production. Therefore, it is believed that the studies shown in this work constitute a new idea regarding to the preparation of anode materials for Li-ion batteries. 2. Experimental The investigated nanocomposite has been fabricated in twosteps facile chemical synthesis which is introduced in details in our previous report [14]. Therefore, in this work the synthesis procedure is only briefly described below. The initial raw multiwall carbon nanotubes (raw-MWCNTs; 93%wt. of purity) were supplied by the CNT CO., Ltd. from South Korea. Such CNTs were oxidized to form MWCNTs-COOH with heating at 120  C in the concentrated nitric acid (68% HNO3; Sigma Aldrich) for 50 h under a reflux [15–19]. The oxidized CNTs were further modified using a 25% aqueous solution of ammonia. This led to obtain the ammonium salt in form of MWCNTs-COONH4 [18,19] which act as a precursor for the second step of synthesis. The MWCNTs-COONH4 (1 gram) were dispersed in 500 ml of water and then it was cooled down. After the temperature reached 4  C, two aqueous solutions of 0.215 g iron dichloride tetrahydrate (FeCl2  4H2O; 98% of purity; Sigma-Aldrich) and 0.588 g iron trichloride hexahydrate (FeCl3  6H2O; 97% of purity; SigmaAldrich) were added by droplets into the previously prepared carbon nanotubes dispersion. The reacting system was mixed continuously with a magnetic stirrer. After adding of the last drops of iron chlorides solutions, the pH of dispersion was brought up to 10 with 1 M NaOH. Then, it was progressively heated for 30 min up to 100  C under a constant magnetic stirring. Finally, the obtained product was centrifuged several times with distilled water until the value of pH reached 7. In order to determine the structure and morphology of obtained MWCNTs-COONH4 and nanocomposite, the following equipment was used: a Hitachi S-5500 scanning-transmission electron microscope (STEM), a Phillips X’Pert X-ray diffractometer (XRD) equipped with a copper-Ka1 X-ray source, a T64000 Series II Raman spectrometer equipped with Nd-YAG laser (l = 532 nm) of continuous wave excitation, a standard Mössbauer spectrometer

Fig. 4. XRD patterns of (a) MWCNTS-COONH4 and (b) investigated nanocomposite (a-Fe2O3/MWCNTs). Miller indices based on JCPDS cards with no. 41-1487 (graphite) and 84-0311 (a-Fe2O3).

(POLON) and 57Co/Rh source of g-radiation placed on a vibrator working in a constant acceleration mode, and a TA Instruments SDT 2960 Simultaneous DTA-TGA (TGA-DTA). The electrochemical performance and cycleability of MWCNTsCOONH4 and obtained nanocomposite as anode materials were investigated using the coin-type cells. At the beginning, both studied materials were dried in a vacuum oven at 100  C before their further use. This ensured the evaporation of residual moisture. After that, the electrode containing one of the studied materials was fabricated by mixing 85%wt. of this material with 5%wt. of acetylene black (AC; Strem Chemicals Inc., USA) and 10%wt. of polyvinylidene fluoride (PVDF; Arkema Inc., USA) in an adequate amount of N-methyl-2-pyrrolidone (NMP, Mitsubishi Chemical Corporation, Japan). This was performed to produce slurry which was stirred overnight to obtain its good quality. It is worth adding that the slurries of both studied nanomaterials were prepared in exactly the same manner using the same chemicals as well as their proportions. Then, such prepared slurries were deposited on a copper foil (UACJ Corporation, Japan) with a doctor blade coater. This allowed formation of the tapes with the studied nanostructures as the active materials, which then were dried in the vacuum oven at 110  C to evaporate NMP solvent during about 1 day. Afterwards, the circular electrodes with diameters of 10 mm were cut from the previously obtained tapes and pressed under a pressure of about 50 MPa. Such circular anodes were placed in argon-filled glove box (82-2 Spez, MecaPlex, Switzerland) where the concentrations of oxygen and moisture were below 1 ppm. Finally, the cells were assembled using the previously prepared electrodes as a working electrodes, the Hipore membranes supplied by Asahi Kasei Corp. soaked with an electrolyte containing a 1 M solution of lithium hexafluorophosphate (LiPF6) dissolved in 1: 1 (v/v) ethylene carbonate (EC) and diethyl carbonate (DEC) (Mitsubishi Corp., Japan) and the metallic lithium foils (99%, FMC, USA, diameter of 12 mm) acted as counter electrode. Such assembled cells were galvanostatically cycled for two different current densities by means of a homemade battery tester in the voltage range between 0.02 and 2.0 V vs. Li+/Li. 3. Results and discussion

Fig. 3. Raman spectra of (a) MWCNTS-COONH4 and (b) investigated nanocomposite (a-Fe2O3/MWCNTs); ID – intensity of D band, IG – intensity of G band, ID’ – intensity of D’ band.

The morphologies and structures of the functionalized multiwall carbon nanotubes (MWCNTs-COONH4) as well as the nanocomposite (MWCNTs covered by iron oxides) have been observed

M. Krajewski et al. / Electrochimica Acta 228 (2017) 82–90

85

Fig. 5. Room temperature Mössbauer spectra of (a) MWCNTs-COONH4 and (b) investigated nanocomposite; black point  experimental data, solid black line K main spectrum, solid magenta line – crystalline a-Fe, solid red line – nonstoichiometric FexC, solid orange line – iron impurity in beryllium, solid blue line – ferrihydrites or Fex(COO)y, dashed blue line – superparamagnetic a-Fe2O3, and solid green line – crystalline a-Fe2O3.

with STEM and they are shown in Fig. 1. The microscopy studies indicate clearly that the chemically-modified MWCNTs are very disordered and they tend to form the bundles. STEM investigations also confirm that CNTs have been successfully coated by the randomly-dispersed iron oxide nanoparticles well-seen in Fig. 1b. The CNTs diameter varies from 10 to 40 nm, while the average diameter of deposited iron oxide nanoparticles is approximately 50 nm. Nevertheless, their dimensions fluctuate between 10 nm and 70 nm. While discussing the microscopy results, it is unreasonable to avoid the fact that MWCNTs-COONH4 contain inside their structures the catalyst residues originating from the fabrication method which was a catalytic chemical vapour deposition (CCVD). The quantity of deposited iron oxide has been determined with use of thermogravimetric analysis (TGA). In turn, the differential thermal analysis (DTA) shows that the deposited iron oxide does not oxidize during the heating in the artificial air (75% of nitrogen and 25% of oxygen). The obtained TGA-DTA curves of MWCNTsCOONH4 and studied nanocomposite are shown in Fig. 2. One can see that the shapes of corresponding curves for both nanomaterials

are very similar. This suggests that the obtained curves are related to the changes occurring rather in carbon nanotubes than in the deposited iron oxides. It is known that the oxidation of carbon nanotubes can lead to the enhancement of their hydrophilic properties [20–22]. This causes that the functionalized MWCNTs can absorb easily the moisture from environment. Therefore, the low and broad endothermic peak appears in the DTA spectra of MWCNTsCOONH4 as well as of studied nanocomposite and it is related to the dehydration process of both nanomaterials. Moreover, all carbon materials tend to be oxidized in a presence of oxygen during heating. This process is often called combustion [23] and leads to formation of gaseous products  carbon dioxide (CO2) or carbon monoxide (CO), which are removed from the experimental system. Thereby, the high exothermic peak is visible in DTA curves and it describes the oxidation of carbon. Additionally, the combustion of carbon originating from CNTs influences the TGA curves of MWCNTs-COONH4 and studied nanocomposite which are marked as the solid lines in Fig. 2. This process causes the substantial mass loss for both materials. The chemically-modified

Table 1 Hyperfine parameters fitted to the Mössbauer spectra for MWCNTs-COONH4 and investigated nanocomposite (A – relative contribution represented by the area under subspectrum, IS – isomer shift, QS – quadrupole splitting and Bhf – hyperfine magnetic field). Iron-containing phase

Parameter

MWCNTs-COONH4

Investigated nanocomposite

a-Fe

A (%) IS (mm/s) QS (mm/s) Bhf (T)

11 0.00 0.00 32.8

– – – –

FexC

A (%) IS (mm/s) QS (mm/s) Bhf (T)

79 0.17 0.02 20.7

7 0.20 – 21.7  0.5

Ferrihydrites or Fex(COO)y

A (%) IS (mm/s) QS (mm/s) Bhf (T)

10 0.42 0.63 –

54 0.34 0.66 –

a-Fe2O3

A (%) IS (mm/s) QS (mm/s) Bhf (T)

– – – –

39 0.38 0.20 48.7

86

M. Krajewski et al. / Electrochimica Acta 228 (2017) 82–90

MWCNTs and nanocomposite have lost about 97.6%wt., and 72.2%wt. of their initial mass, respectively. In other words, the mass of products after the combustion equals about 2.4%wt. and 27.8%wt. for the functionalized multiwall carbon nanotubes and the studied nanocomposite, respectively. Considering these values, it is possible to conclude that the content of deposited iron oxides during applied synthesis is around 25.4%wt. Furthermore, the purity of raw CNTs used in this work was around 93%wt. (see Experimental Section), while the purity of MWCNTs-COONH4 has been determined as about 97.6%wt. Therefore, the application of simple TGA experiment confirms the previous findings [16–19] which have shown that the oxidation of CNTs with use of nitric acid can decrease the residual traces of catalyst coming from the fabrication process. Hence, it is believed that the influence of this contaminant is minimal in the case of application of both materials in Li-ion batteries. Comparing the obtained TGA-DTA curves, it is worth noting that the deposition of iron oxide nanoparticles on MWCNTs causes the decrease of temperature related to combustion of carbon from 550  C to 475  C. This may be explained by the weakening of carbon-carbon bonds in the studied nanocomposite. Raman spectroscopy has been applied to determine the possible structure of investigated MWCNTs-COONH4 and nanocomposite. The Raman spectra of both nanomaterials are shown in Fig. 3. Considering both spectra, they reveal four bands which are characteristic for the carbon structures i.e. D band (1340 cm1), G (1580 cm1), D’(1610 cm1) and 2D (G’; 2690 cm1) [24]. The G band is related to the E2g vibrational mode of graphite representing the stretching C-C vibrations. On the other hand, the D and D’ modes indicate the disordered structure of graphite related to the zone-edge and mid-zone phonons, respectively, while the 2D mode is associated with the first overtone of the D mode [16,24]. All mentioned bands are typical of the carbon nanotubes, which confirms that the chemically-modified MWCNTs applied in the synthesis of nanocomposite do not change drastically under the process of a-Fe2O3 deposition. However, a very small shift of D, G and D’ bands towards low frequencies has been observed in the case of nanocomposite spectrum. Moreover, the ratios between D band and G band intensities (ID/IG) are almost identical for MWCNTs-COONH4 and carbon nanotubes present in nanocomposite and they equal about 1.49. In turn, the value of D’ to G peak intensities (ID’/IG) for the investigated nanocomposite is

slightly lower. This might be associated with the partial suppression of mid-zone phonons caused by the presence of a-Fe2O3 on MWCNTs. The bands of iron oxides deposited on MWCNTs are broad and some of them are probably merged together. Such observation has been already reported for hematite nanoparticles which have decorated the walls of single-wall carbon nanotubes (SWCNTs) [7]. Nevertheless, the positions of iron oxide bands in spectrum of nanocomposite are in an agreement to those of well crystalline a-Fe2O3 [25]. Therefore, the difference between the bands of the typical crystalline a-Fe2O3 and a-Fe2O3 deposited on MWCNTs can be caused, for instance, by the interactions between iron oxide nanoparticles and carbon nanotubes [26]. It is commonly known that the laser irradiation can lead to a significant sample heating during the measurement and it can influence the final results, especially in the case of iron oxides. These side effects may cause the irreversible transformation of some iron oxides (e.g. Fe3O4, g-Fe2O3) into hematite [27–29]. Hence, the powder X-ray diffraction measurements have been also performed on MWCNTs-COONH4 and nanocomposite to verify their structures determined with a use of Raman spectroscopy. The obtained XRD patterns are demonstrated in Fig. 4. The Miller indices for CNTs and a-Fe2O3 have been assigned with regard to JCPDS cards with no. 41-1487 and 84-0311, respectively. According to these results, the chemically-modified carbon nanotubes have been successfully coated by the hematite and no significant changes in the peak positions and their relative intensities for CNTs after the deposition of a-Fe2O3 can be perceived. Moreover, no signal originating from the residual traces of catalyst has been observed. This indicates again that the content of contaminant from the fabrication process is really low (agreement with TGA) so it is below the detection limit of the applied X-ray method. The last applied experimental technique to determine all phases containing iron ions in MWCNTs-COONH4 and in the investigated nanocomposite was the transmission Mössbauer spectroscopy (TMS). In other words, this method allows establishing precisely the structures of catalyst residues and deposited iron oxides. The collected Mössbauer spectra for both nanomaterials measured at room temperature are shown in Fig. 5, while Table 1 includes the information about: (i) the contributions of the subspectra corresponding to different individual phases containing iron and (ii) the hyperfine parameters fitted to all subspectra. According to Mössbauer data, the signals of iron-containing phases

Fig. 6. Galvanostatic discharge/charge profiles recorded at discharge/charge current density of 25 mA/g for (a) MWCNTs-COONH4 and (b) investigated nanocomposite.

M. Krajewski et al. / Electrochimica Acta 228 (2017) 82–90

in the case of chemically-modified CNTs are associated with three components represented by a-Fe, nonstoichiometric iron carbide (FexC) and probably ferrihydrites. Although this spectrum has been collected within 8 weeks, its statistic is not perfect due to very low content of iron-containing phases in this sample. Moreover, one of the doublets in this spectrum was identified as an iron impurity in beryllium present in the window of the proportional counter used for g-ray detection. This effect is a consequence of extremely long time of the experiment. Nevertheless, the values of matched hyperfine parameters (obtained after necessary removal of this extra doublet) agree quite well with the similar Mössbauer studies performed on MWCNTs-COONH4 [19]. The Mössbauer spectrum describing the investigated nanocomposite at room temperature was much more complex to analyse. However, it consists of three main subspectra including a sextet corresponding to nonstoichiometric iron carbide (FexC), a doublet attributed to ferrihydrites or iron ions attached to carboxylic groups (marked as Fex(COO)y in Fig. 5), and also a sextet with high hyperfine magnetic field 48.7 T. Both isomer shift (0.38 mm/s) and quadrupole splitting (0.20 mm/s) fitted for the sextet are characteristic of a-Fe2O3 but hyperfine magnetic field is markedly smaller than the value typically observed for this phase at room temperature (51 T). All this features as well as the broadening of lines in the sextet can be attributed to temperature relaxations of magnetic moments of a-Fe2O3 nanoparticles, which causes reduction of hyperfine magnetic field [30]. At this point, it is worth adding that the intense doublet placed in the middle of spectrum is a combination of the signals originating from ferrihydrites or Fex(COO)y and superparamagnetic hematite. This can be explained taking into account a spread of the sizes of deposited iron oxide nanoparticles. According to the microscopy observations discussed before, their diameters vary between 10 nm to 70 nm. Therefore, the larger nanoparticles show reduced magnetic splitting while the smaller ones exhibit the superparamagnetic behaviour that is possible to observe for the hematite nanoparticles even up to about 40 nm [31]. The Mössbauer spectra gathered for such nanoparticles at room temperature usually display a single doublet in spectrum [32– 34]. Besides that, in the investigated material the hyperfine parameters of the components forming the doublet are quite similar. The fitted values of relative contribution (A), isomer shift

Fig. 7. Cycling performance curves of MWCNTs-COONH4 and investigated nanocomposite at the charge/discharge current density of 25 mA/g and 200 mA/ g; open and filled symbols correspond to charge and discharge current, respectively.

87

(IS) and quadrupole splitting (QS) equal 38%, 0.35 mm/s and 0.56 mm/s for superparamagnetic a-Fe2O3 and 16%, 0.34 mm/s and 0.92 mm/s for ferrihydrites or Fex(COO)y, respectively. This issue causes that the doublet shown in spectrum for nanocomposite should be considered as a combination of above described phases. Moreover, the signal originating from a-Fe could be additionally fitted to the Mössbauer spectra of nanocomposite. However, its intensity is in the same range as the measurement uncertainty so it is omitted in the spectrum. Finally, the analysis of Mössbauer spectrum obtained for the investigated nanocomposite leads to conclusion that the external structural changes of MWCNTsCOONH4 are related to the deposition of hematite as well as to the attachment of few iron ions to carboxylic groups formed during the modification process of CNTs. Reassuming the obtained Mössbauer results, the iron-containing phases like iron carbide, ferrihydrites and a-Fe should be rather considered as the inner parts of MWCNTs, while other registered phases seem to be located on the surface of outermost graphene sheet forming CNTs. Obviously, the applied synthesis method allows coating CNTs by the randomly-dispersed iron oxide nanoparticles visible in Fig. 1 and in fact the Mössbauer signal of those particles can be associated with a-Fe2O3 phase. Besides that, one of the synthesis steps is related to the formation of carboxylate salts which can be written as MWCNTs-(COO)3Fe3+ or MWCNTs(COO)2Fe2+. Part of these salts transforms into oxide form during the heating in NaOH alkaline environment, while some remaining groups may be still present and may act as the ‘side products’. Thereby, none of the previously applied experimental techniques (XRD, Raman spectroscopy) could detect those iron carboxylate salt groups because they do not exhibit a long range order. The galvanostatic discharge/charge curves of two electrode coin cells with MWCNTs-COONH4 and investigated nanocomposite as anodes are demonstrated in Fig. 6. These curves have been recorded at current densities of 25 mA/g in the potential window of 0.02 and 2.0 V vs. Li+/Li. Comparing the obtained profiles for both nanomaterials, it is visible that they are slightly different from each other. This proves that the lithium storage is carried out via two various mechanisms in the case of a-Fe2O3/MWCNTs, including insertion and conversion mechanisms. Nevertheless, it is obvious that the main Li+ ions storage is related to insertion/deinsertion processes because the studied nanocomposite is mainly composed of carbon nanotubes. Principally, it is worth noting that the discharge/charge profiles for the chemically-modified MWCNTs investigated in this work are pretty similar to those found in the literature [35–40]. The characteristic feature of such curves is that they exhibit no voltage plateau [6,35] and this has been explained previously considering several issues. For instance, Maurin et al. [40] suggest that it depends on the degree of CNTs graphitization, while Mi et al. [35] and Lin et al. [37] propose the additional lithium storage inside the microspores and defects. Also, such phenomenon could be referred to the quasi-reversible interactions between lithium ions and carboxylic groups resulting in formation of MWCNTs-COO(Li+) salts [39]. All of above explanations are also a source of discharge capacity enhancement (especially during the 1st discharge) with regard to the typical graphite electrodes. However, a large portion of such capacity is usually associated with the irreversible capacity (cirr) [35–37]. The same situation is observed for the chemicallymodified multiwall carbon nanotubes studied in this work. They exhibit an extremely high magnitude of the initial discharge capacity which reaches almost 2230 mAh/g using a discharge current density of 25 mA/g, whilst the initial charge capacity equals 396 mAh/g. It means that the irreversible capacity is about 1830 mAh/g and it reflects a poor coulombic efficiency of 18%. However, the values of discharge and charge capacities are indeed higher than the theoretical capacity of graphite electrode (372 mAh/g [8]).

88

M. Krajewski et al. / Electrochimica Acta 228 (2017) 82–90

Table 2 Discharge capacities for MWCNTs-COONH4 and investigated nanocomposite (a-Fe2O3/MWCNTs) derived from the cycling performance tests with two various discharge/ charge current densities. Sample

Discharge/ charge current density (mA/g)

1st cycle discharge capacity 10th cycle discharge (mAh/g) capacity (mAh/g)

20th cycle discharge capacity (mAh/g)

40th cycle discharge capacity (mAh/g)

MWCNTsCOONH4

25

2227

338

289

257

200 25

1308 1925

217 395

199 357

184 332

200

1154

302

264

228

a-Fe2O3 /MWCNTs

In addition, the fade of capacity is evident for the first ten cycles, while the coulombic efficiency increases gradually from 68% to 76%, 88%, 96% for 2nd, 3rd 10th and 40th cycle, respectively. Moreover, it is worth underlining that to our knowledge nobody has reported so far such a high initial discharge capacity for the multiwall carbon nanotubes. As it was mentioned before, the discharge/charge curves for studied nanocomposite differs a bit from the curves of MWCNTsCOONH4. For instance, they are not as smooth as the profiles for the chemically-modified CNTs. This is obviously caused by the presence of iron oxide nanoparticles deposited on the outer walls of CNTs. The discharge/charge profiles for a-Fe2O3/MWCNTs can be described considering four regions. They correspond to the insertion of lithium inside a-Fe2O3 (small knee at around 1.7 V), the reduction of Fe3+ to Fe2+ (small knee at around 1.1 V) and the reduction of Fe2+ to Fe0 (plateau at 0.9 V). The forth region (voltage below 0.9 V) is related to the formation of solidelectrolyte interphase (SEI) layer due to the decomposition of the solvent from the electrolyte. This phenomenon is typical for the carbon-based electrodes [35–39] as well as the metal oxide electrodes [8,41–43]. Therefore, this region is clearly visible for the initial discharge curves of MWCNTs-COONH4 and investigated nanocomposite. Furthermore, the first three regions of initial discharge curve recorded for the nanocomposite can be attributed to the chemical reactions with following numbers (4), (5) and (6) [7,43]. Fe2O3 + x Li+ + x e ! LixFe2O3

(4)

LixFe2O3 + (2-x) Li+ + (2-x) e ! Li2Fe2O3

(5)

Li2Fe2O3 + 4 Li+ + 4 e @ 2 Fe + 3 Li2O

(6) +

These reactions show that the consumption of 6 Li ions leads to formation of metallic iron and lithium oxide. However, the first charge process corresponding to the reverse reaction usually causes the oxidation of Fe0 to Fe2+ according to the chemical reaction represented by Eq. (7) [7,43]. 2Fe + 2 Li2O @ FeO + 4 Li+ + 4 e

(7)

This phenomenon is well seen as the short plateaus at voltage of about 1 V in the consecutive discharge curves beginning from 2nd cycle. Moreover, this induces the fade of capacity because only 4 moles out of 6 moles of lithium ions are stored during the subsequent reversible discharge/charge reactions [7,43]. The initial discharge capacity of investigated nanocomposite is lower than the initial discharge capacity of MWCNTs-COONH4 and it reaches about 1930 mAh/g. This can be explained considering the lithium storage in the form of MWCNTsCOO(Li+) discussed before. Briefly, the chemically-modified

CNTs contain much more available carboxylic groups than CNTs present in the nanocomposite. This is caused by the partial consumption of carboxylic groups for the deposition of iron oxide nanoparticles on carbon nanotubes. On the other hand, the initial charge capacity and corresponding coulombic efficiency for 1st cycle are higher for studied nanocomposite and equal 469 mAh/g and 24%, respectively. Then, the coulombic efficiency rises from 74% for 2nd to 90% for 10th cycle and it remains at the level of about 93% up to 40th cycle. The cycling performance curves of MWCNTs-COONH4 and investigated nanocomposite as anodes are shown in Fig. 7. They have been collected for two various discharge/charge current densities: 25 mA/g and 200 mA/g. Moreover, Table 2 contains the precise values of discharge capacities for MWCNTs-COONH4 and investigated nanocomposite for 1st, 10th, 20th and 40th cycle. Comparing the cycling experimental data, it should be noted that the pronounced fade of capacity for all curves is observed up to 5th cycle and then the capacities retain the steady levels. Applying the discharge/charge current density of 200 mA/g during 40 cycles, it is possible to keep the discharge capacity of about 184 mAh/g and 228 mAh/g for MWCNTs-COONH4 and investigated nanocomposite, respectively. These values are not as high as for a-Fe2O3/ carbon nanosprings composite [4] or a hybrid film composed of a-Fe2O3/single-walled carbon nanotube [7] or CNTs filled by Fe2O3 [8]. Nevertheless, the studied a-Fe2O3/MWCNTs exhibit the improved specific capacity in comparison with Sb/CNTs and Bi/ CNTs composites [44], TiO2/CNTs composite [45], TiO2 nanotubes [45,46], Sn-doped TiO2 nanotubes [45] nano-sized carboxylates [47] and the various advanced carbon materials (mesophase soft carbon, mesophase graphite, etc.) [48]. The major drawback of both nanomaterials studied in this work seems to be the high value of irreversible capacity. Hence, it is believed that this inconvenience should be solved to increase the possible application of MWCNTs-COONH4 and investigated nanocomposite as the anode electrode in Li-ion batteries. For instance, the reduction of cirr could be achieved with a partial removal of COONH4 groups attached to the chemically-modified carbon nanotubes as well as with a shortening of the carbon nanotubes lengths [36,49], which decrease effectively the diffusion pathways for Li+ ions [6]. Moreover, in the case of studied nanocomposite a better quality of deposited iron oxide can cause the improvement of reversible capacity. 4. Conclusions This work describes the two-steps facile chemical synthesis which aims to form the nanocomposite composed of the chemically-modified multiwall carbon nanotubes (MWCNTsCOONH4) covered by the randomly-dispersed hematite nanoparticles. According to the thermogravimetric results, the proportions of used chemicals have enabled us to deposit about 25.4%wt.

M. Krajewski et al. / Electrochimica Acta 228 (2017) 82–90

of iron oxide on the outermost wall of CNTs. Such amount of iron oxide influences clearly the cycling performance of the investigated nanocomposite. Its values of specific capacity are stable and they retain at the level of 330 mAh/g and 230 mAh/g after 40 cycles applying the discharge/charge rate of 25 mA/g and 200 mA/g, respectively. In contrary, the corresponding values of capacity for MWCNTs-COONH4 equal 257 mAh/g and 184 mAh/g applying the same discharge/charge currents. Obviously, this indicates that the storage of lithium ions in the investigated nanocomposite is carried out via insertion and conversion mechanisms. The major disadvantage of investigated nanocomposite is the high value of irreversible capacity. It seems that this inconvenience is mainly related to the irreversible lithium storing processes occurring rather in MWCNTs-COONH4 than in iron oxide nanoparticles. Therefore, the future works performed on a-Fe2O3/ MWCNTs will be associated with the shortening of the carbon nanotubes lengths which may increase their reversible capacity. In addition, it is worth noting that, among many advantages of the proposed synthesis procedure, the most important one is a possibility of fabrication a large amount of material during one synthesis. It does not require the application of high temperatures and protective atmospheres or vacuum as in the case of other methods used for an iron oxides deposition on carbon materials [4,7,8,26,42]. This affects the simplicity of synthesis system as well as decreases the cost of synthesis. It should be also added that the developed procedure can be easily fitted to the preparation of other nanocomposites composed of CNTs and different metal oxides for the various applications in the field of energy storage, catalysis, sensors, solar cells etc. Acknowledgements This work was partially supported by the Foundation for Polish Science within International PhD Projects Programme co-financed by the EU European Regional Development Fund. References [1] R.M. Dell, D.A.J. Rand, Energy storage  a key technology for global energy sustainability, J. Power Sources 100 (2001) 2. [2] P.G. Bruce, B. Scrosati, J.M. Tarascon, Nanomaterials for rechargeable lithium batteries, Angew. Chem. Int. Ed. 47 (2008) 2930. [3] B.J. Landi, M.J. Ganter, C.D. Cress, R.A. DiLeo, R.P. Raffaelle, Carbon nanotubes for lithium ion batteries, Energy Environ. Sci. 2 (2009) 638. [4] J. Shao, J.X. Zhang, J.J. Jiang, G.M. Zhou, M.Z. Qu, a-Fe2O3@CNSs nanocomposites as superior anode materials for lithium-ion batteries, Electrochim. Acta 56 (2011) 7005. [5] X.M. Liu, Z.D. Huang, S.W. Oh, B. Zhang, P.C. Ma, M.M.F. Yuen, J.K. Kim, Carbon nanotube (CNT)-based composites as electrode material for rechargeable Liion batteries: A review, Compos. Sci. Technol. 72 (2012) 121. [6] C. de las Casas, W.Z. Li, A review of application of carbon nanotubes for lithium ion battery anode material, J. Power Sources 208 (2012) 74. [7] Z.Y. Cao, B.Q. Wei, a-Fe2O3/single-walled carbon nanotube hybrid films as high-performance anodes for rechargeable lithium-ion batteries, J. Power Sources 241 (2013) 330. [8] W.J. Yu, P.X. Hou, F. Li, C. Liu, Improved electrochemical performance of Fe2O3 nanoparticles confined in carbon nanotubes, J. Mater. Chem. 22 (2012) 13756. [9] A.S. Arico, P.G. Bruce, B. Scrosati, J.M. Tarascon, W.V. Schalkwijk, Nanostructured materials for advanced energy conversion and storage devices, Nat. Mater. 4 (2005) 366. [10] S. Goriparti, E. Miele, F. De Angelis, E. Di Fabrizio, R.P. Zaccaria, C. Capiglia, Review on recent progress of nanostructured anode materials for Li-ion batteries, J. Power Sources 257 (2014) 421. [11] J.H. Ryu, J.W. Kim, Y.E. Sung, S.M. Oh, Failure modes of silicon powder negative electrode in lithium secondary batteries, Electrochem. Solid State Lett. 7 (2004) A306. [12] A. Mukhopadhyay, B.W. Sheldon, Deformation and stress in electrode materials for Li-ion batteries, Prog. Mater. Sci. 63 (2014) 58. [13] D. Qian, G.J. Wagner, W.K. Liu, M.F. Yu, R.S. Ruoff, Mechanics of carbon nanotubes, Appl. Mech. Rev. 55 (2002) 495. [14] M. Krajewski, A. Malolepszy, L. Stobinski, S. Lewinska, A. Slawska-Waniewska, M. Tokarczyk, G. Kowalski, J. Borysiuk, D. Wasik, Preparation and characterization of hematite-multiwall carbon nanotubes nanocomposite, J. Supercond. Nov. Magn. 28 (2015) 901.

89

[15] W. Xia, Y. Wang, R. Bergstrasser, S. Kundu, M. Muhler, Surface characterization of oxygen-functionalized multi-walled carbon nanotubes by high-resolution X-ray photoelectron spectroscopy and temperature-programmed desorption, Appl. Surf. Sci. 254 (2007) 247. [16] L. Stobinski, B. Lesiak, L. Kover, J. Toth, S. Biniak, G. Trykowski, J. Judek, Multiwall carbon nanotubes purification and oxidation by nitric acid studied by the FTIR and electron spectroscopy methods, J. Alloy Compd. 501 (2010) 77. [17] L. Stobinski, B. Lesiak, J. Zemek, P. Jiricek, Time dependent thermal treatment of oxidized MWCNTs studied by the electron and mass spectroscopy methods, Appl. Surf. Sci. 258 (2012) 7912. [18] I. Kuryliszyn-Kudelska, A. Malolepszy, M. Mazurkiewicz, L. Stobinski, K.J. Kurzydlowski, W. Dobrowolski, Enhanced coercivity of as-prepared and chemically modified multiwalled carbon nanotubes, Phys. Status Solidi A 208 (2011) 1787. [19] A. Jamrozik, J. Przewoznik, M. Mazurkiewicz, A. Malolepszy, L. Stobinski, G. Trykowski, I. Habina, K. Matlak, J. Korecki, C. Kapusta, M. Sitarz, K. Burda, Influence of iron contaminations on local and bulk magnetic properties of nonfunctionalized and functionalized multi- wall carbon nanotubes, Phys. Status Solidi A 211 (2014) 661. [20] Z. Liu, X.M. Sun, N. Nakayama-Ratchford, H.J. Dai, Supramolecular chemistry on water-soluble carbon nanotubes for drug loading and delivery, ACS Nano 1 (2007) 50. [21] S. Kundu, Y.M. Wang, W. Xia, M. Muhler, Thermal stability and reducibility of oxygen-containing functional groups on multiwalled carbon nanotube surfaces: a quantitative high-resolution XPS and TPD/TPR Study, J. Phys. Chem. C 112 (2008) 16869. [22] N.K. Mehra, V. Mishra, N.K. Jain, A review of ligand tethered surface engineered carbon nanotubes, Biomaterials 35 (2014) 1267. [23] B.J. Landi, C.D. Cress, C.M. Evans, R.P. Raffaelle, Thermal oxidation profiling of single-walled carbon nanotubes, Chem. Mater. 17 (2005) 6819. [24] T. Belin, F. Epron, Characterization methods of carbon nanotubes: a review, Mater. Sci. Eng. B-Adv. Funct. Solid-State Mater. 119 (2005) 105. [25] S.J. Oh, D.C. Cook, H.E. Townsend, Characterization of iron oxides commonly formed as corrosion products on steel, Hyperfine Interact. 112 (1998) 59. [26] N. Yan, X.H. Zhou, Y. Li, F. Wang, H. Zhong, H. Wang, H.Q. Chen, Fe2O3 nanoparticles wrapped in multi-walled carbon nanotubes with enhanced lithium storage capability, Sci. Rep. 3 (2013) 3392. [27] O.N. Shebanova, P. Lazor, Raman study of magnetite (Fe3O4): laser-induced thermal effects and oxidation, J. Raman Spectrosc. 34 (2003) 845. [28] A.M. Jubb, H.C. Allen, Vibrational spectroscopic characterization of hematite, maghemite, and magnetite thin films produced by vapor deposition, ACS Appl. Mater. Interfaces 2 (2010) 2804. [29] M. Krajewski, K. Brzozka, W.S. Lin, H.M. Lin, M. Tokarczyk, J. Borysiuk, G. Kowalski, D. Wasik, High temperature oxidation of iron-iron oxide core-shell nanowires composed of iron nanoparticles, Phys. Chem. Chem. Phys. 18 (2016) 3900. [30] K. Brzozka, M. Krajewski, A. Malolepszy, L. Stobinski, T. Szumiata, B. Gorka, M. Gawronski, D. Wasik, Phase analysis of magnetic inclusions in nanomaterials based on multiwall carbon nanotubes, Acta Phys. Pol. A (2017) In Press. [31] T.P. Raming, A.J.A. Winnubst, C.M. van Kats, A.P. Philipse, The synthesis and magnetic properties of nanosized hematite (a-Fe2O3) Particles, J. Colloid Interface Sci. 249 (2002) 346. [32] F. Bodker, M.F. Hansen, C.B. Koch, K. Lefmann, S. Morup, Magnetic properties of hematite nanoparticles, Phys. Rev. B 61 (2000) 6826. [33] M.F. Hansen, C.B. Koch, S. Morup, Magnetic dynamics of weakly and strongly interacting hematite nanoparticles, Phys. Rev. B 62 (2000) 1124. [34] R. Zboril, M. Mashlan, D. Petridis, Iron(III) oxides from thermal processess synthesis, structural and magnetic properties, Mössbauer spectroscopy characterization, and applications, Chem. Mater. 14 (2002) 969. [35] C.H. Mi, G.S. Cao, X.B. Zhao, A non-GIC mechanism of lithium storage in chemical etched MWNTs, J. Electroanal. Chem. 562 (2004) 217. [36] J.Y. Eom, D.Y. Kim, H.S. Kwon, Effects of ball-milling on lithium insertion into multi-walled carbon nanotubes synthesized by thermal chemical vapour deposition, J. Power Sources 157 (2006) 507. [37] K.H. Lin, Y.H. Xu, G.R. He, X.L. Wang, The kinetic and thermodynamic analysis of Li ion in multi-walled carbon nanotubes, Mater. Chem. Phys. 99 (2006) 190. [38] S.Y. Chew, S.H. Ng, J.Z. Wang, P. Novak, F. Krumeich, S.L. Chou, J. Chen, H.K. Liu, Flexible free-standing carbon nanotube films for model lithium-ion batteries, Carbon 47 (2009) 2976–2983. [39] E. Frackowiak, F. Beguin, Electrochemical storage of energy in carbon nanotubes and nanostructured carbons, Carbon 40 (2002) 1775. [40] G. Maurin, C. Bousquet, F. Henn, P. Bernier, R. Almairac, B. Simon, Electrochemical lithium intercalation into multiwall carbon nanotubes: a micro-Raman study, Solid State Ionics 136 (2000) 1295. [41] Y.M. Lin, P.R. Abel, A. Heller, C.B. Mullins, a-Fe2O3 nanorods as anode material for lithium ion batteries, J. Phys. Chem. Lett. 2 (2011) 2885. [42] S.L. Yang, C.Y. Cao, G. Li, Y.B. Sun, P.P. Huang, F.F. Wei, W.G. Song, Improving the electrochemical performance of Fe3O4 nanoparticles via a double protection strategy through carbon nanotube decoration and graphene networks, Nano Res. 8 (2015) 1339. [43] M.V. Reddy, T. Yu, C.H. Sow, Z.X. Shen, C.T. Lim, G.V.S. Rao, B.V.R. Chowdari, a-Fe2O3 nanoflakes as an anode material for Li-ion batteries, Adv. Funct. Mater. 17 (2007) 2792. [44] Y. Nuli, J. Yang, M.S. Jiang, Synthesis and characterization of Sb/CNT and Bi/CNT composites as anode materials for lithium-ion batteries, Mater. Lett. 62 (2008) 2092.

90

M. Krajewski et al. / Electrochimica Acta 228 (2017) 82–90

[45] H. Huang, W.K. Zhang, X.P. Gan, C. Wang, L. Zhang, Electrochemical investigation of TiO2/carbon nanotubes nanocomposite as anode materials for lithium-ion batteries, Mater. Lett. 61 (2007) 296. [46] N.A. Kyeremateng, F. Vacandio, M.T. Sougrati, H. Martinez, J.C. Jumas, P. Knauth, T. Djenizian, Effect of Sn-doping on the electrochemical behaviour of TiO2 nanotubes as potential negative electrode materials for 3D Li-ion micro batteries, J. Power Sources 224 (2013) 269.

[47] X.Y. Wu, J. Ma, Y.S. Hu, H. Li, L.Q. Chen, Nano-sized carboxylates as anode materials for rechargeable lithium-ion batteries, J. Energy Chem. 23 (2014) 269. [48] M.D. Fang, T.H. Ho, J.P. Yen, Y.R. Lin, J.L. Hong, S.H. Wu, J.J. Jow, Preparation of advanced carbon anode materials from mesocarbon microbeads for use in high C-rate lithium ion batteries, Materials 8 (2015) 3550. [49] X.X. Wang, J.N. Wang, L.F. Su, Preparation and electrochemical performance of ultra-short carbon nanotubes, J. Power Sources 186 (2009) 194.