Correlation between Internal Structure and Electrochemical

Article pubs.acs.org/ac

Correlation between Internal Structure and Electrochemical Impedance Spectroscopy of Multiphase Slurry Systems Seoung Jai Bai† and Young Seok Song*,‡ †

Department of Mechanical Engineering, Dankook University, 126 Jukjeon-dong, Suji-gu, Yongin-si, Gyeonggi-do, Korea Department of Fiber System Engineering, Dankook University, 126 Jukjeon-dong, Suji-gu, Yongin-si, Gyeonggi-do, Korea

‡

S Supporting Information *

ABSTRACT: Slurry for lithium-ion batteries is prepared from an active material, a carbon conductive additive, and a polymeric binder in a solvent, and its morphological change is evaluated over time using electrochemical impedance spectroscopy. A schematic model of the internal structure and dispersion states of the slurry components during 7 days of storage is proposed on the basis of the electrochemical impedance spectroscopy (EIS) measurement. The EIS results reveal that the conductive path constructed by the network structure of the slurry components breaks over time, which can be worsened by mechanical agitation. In order to confirm the morphological change, the slurry is freeze-dried and then prepared to fixate the locations of the slurry components. The existence of a network structure (or flocculation) is verified by morphological observations. In addition, the dispersity index and Micro-CT are introduced as new methods representing the dispersion state of the slurry components.

T

capacitance, and mass-transfer resistance.16−18 In the EIS experiment, the response analysis to the periodic small amplitude AC signal applied to a target system provides electrochemical information about the system structure, electrode interface, and relevant reactions.19,20 Compared with other electrochemical methods such as galvanostatic measurement, cyclic voltammetry, and potentiostatic intermittent titration technique, the EIS can play a significant role in quantitatively characterizing battery system-related transport properties (e.g., electronic conductive) and electrochemical reactions (e.g., rate constant), which are a key consideration to design and fabricate a material system for battery electrodes.21−24 The goal of this study is to understand the morphological change of the LIB components using electrochemical impedance spectroscopy over time. In this study, we investigated the degradation of the slurry using EIS as well as morphological observation and image processing. Changes in the physicochemical features of the slurry over time have been measured using EIS and estimated using equivalent circuit modeling.

he lithium-ion battery (LIB), one of the most important power sources in various energy-storage technologies, has been widely used for portable electronic devices such as cellular phones, laptops, cameras, etc., due to its high energy density, relatively good safety, and good cycling stability.1−4 The manufacturing processes of the LIB are composed of electrode preparation, cell assembly, grading and formation, and safety evaluation.5,6 For preparation of the electrode, active materials, conductive additives, and polymeric binders are generally blended into a solvent with the use of a mixer, and then, the resulting slurry is transported to a storage tank. Therefore, in order to avoid degradation of the slurry, it is necessary to disperse the slurry constituents homogeneously in the solvent and maintain its mixing quality before the following electrode coating step. Furthermore, for rigorous quality control, one needs to monitor and evaluate the change in the internal structure of the slurry during the processing stages.7−10 Proper electrode preparation is critical to achieve the high power performance requirements for LIB applications. Even if battery performance depends on the electrode material itself and preparation methods, the focus of most LIB studies has been on the synthesis, structure, and property of electrode materials.11−15 To the best of our knowledge, this is the first report that evaluates the internal structure of battery components in a slurry state with the help of electrochemisty. Recently, electrochemical impedance spectroscopy (EIS) is being adopted to exploit complex fluid−solid interfaces and electrode processes as well as fundamental electrochemical characteristics, such as charge-transfer resistance, double layer © 2013 American Chemical Society

■

EXPERIMENTAL SECTION Slurry Preparation. The slurries prepared in this study encompassed LiCoO2 powder (JesEChem, Korea), carbon Received: November 20, 2012 Accepted: March 12, 2013 Published: March 12, 2013 3918

dx.doi.org/10.1021/ac303187r | Anal. Chem. 2013, 85, 3918−3925

Analytical Chemistry

Article

Figure 1. Schematic illustration of the change in the internal structure of slurry constituents.

microtomography system (Micro-CT, 1172, Skyscan, Belgium). A 100 kV and 10 Mp X-ray source was employed for 2D image analysis. Dispersion states of the slurry constituents were quantified via entropy-based image processing.26

black (Ketjen Black International Company, Japan), and polyvinylidenefluoride (PVdF, Kureha America, USA) known as an active material, a conductive additive, and a binder, respectively. The used LiCoO2 and carbon black particles had average diameters of 10 μm and 30 nm, respectively. The solid components were mixed in N-methylpyrrolidone (NMP, Aldrich, USA) at 70 wt % with a composition ratio of 95:2:3. The slurries were produced as follows: the active material and the conductive additive were first compounded for 1 h with an agitating machine prior to the addition of the binder and the solvent. Thereafter, the mixture was blended for an additional 2 h with a homogenizer at 3000 rpm.25 In order to identify the time effect on the internal structure of the slurry, the fresh slurry sample was kept for seven days. EIS Instrumentation. EIS data were recorded using a potentiostat (CompactStat, Ivium Technologies, The Netherlands). All EIS measurements were conducted in a twoelectrode system. Two Pt meshes were used as a working electrode and a counter/reference electrode. The Pt working electrode was polarized at 1 V with respect to the counter/ reference electrode. A 10 mV sinusoidal voltage with frequencies ranging from 1 Hz to 100 kHz was applied while the corresponding current was recorded. Freshly prepared slurry was used for an initial impedance measurement. After that, the slurry was stored in a storage tank with or without agitation. Another impedance measurement was conducted using the same configuration as the initial experiment. Equivalent Circuit Modeling. In order to understand the morphological change of slurry over time, an equivalent circuit composed of simple electrical components was introduced. The equivalent circuit design, relevant data processing, and impedance estimation were conducted using a complex nonlinear least-squares (CNLS) method with the aid of commercial software (ZView, Scribner Associates Inc., USA). Morphological Observation and Dispersion Quantification. Slurry samples for morphological measurement were freeze-dried (Freeze-dryer, 74200−40, Labconco, USA), and their cross sections were imaged with a field emission scanning electron microscope (FE-SEM, JSM-6700F, JEOL, Japan). Additionally, the energy dispersive spectroscopy (EDS) measurement was performed to identify the dispersion states of the slurry constituents. Computed tomographic analysis of the slurry was carried out using a high-resolution desktop X-ray

■

RESULTS AND DISCUSSION EIS and Internal Structure Change of Slurry and Its Constituents over Time. Figure 1 illustrates the conforma-

Figure 2. (a) Bode plot and (b) Nyquist plot of the slurry under agitation and no agitation over time.

tional changes of slurry over time. When freshly prepared, the active materials maintain a uniform network surrounded by conductive additives and binders. However, as time elapses, the network breaks down due to aggregation of the conductive additives and binders driven by their internal interactions. In 3919



Article

Figure 3. Impedance changes of not agitated specimen over 3 days: (a) Bode plot of slurry, (b) Nyquist plot of slurry, (c) Bode plot of active material (66.5 wt % in NMP), and (d) Bode plot of carbon conductive additive (1.4 wt % in NMP).

Figure 4. Impedance changes of agitated specimen: (a) Bode plot of slurry, (b) Nyquist plot of slurry, (c) Bode plot of active material (66.5 wt % in NMP), and (d) Bode plot of carbon conductive additive (1.4 wt % in NMP).

the LIB preparation process, slurry is mixed and then coated on the electrode. However, for some reasons, the prepared slurry is likely to be kept in a tank for a period of time prior to the following processes such as coating and compressing. In this sense, it is critical to understand the change in the internal structure of slurries over time considering the real manufacturing processes of LIB. The three slurry components are presumed to keep experiencing configurational change over time as follows: (i) The polymeric binders slightly decompose and aggregate. (ii) The conductive additives agglomerate. (iii) The active materials undergo the breakdown of the initial

network structure. These assumptions are verified experimentally in the following section. Figure 2 shows the impedance changes of the slurry with and without agitation over time. As shown in Figure 2a, both slurries with and without agitation exhibited the increase of impedance after 7 days in a storage tank. However, agitation exacerbates the impedance increase slightly. Figure 2b depicts the Nyquist plot of the results in Figure 2a. In each measurement, a depressed semicircle in the frequency range of 100 kHz to 100 Hz was observed, which was caused by a constant phase element. In the low frequency region below 100 Hz, the Warburg impedance was detected. 3920



Article

Figure 5. (a) Equivalent circuit of the system for impedance fitting: EIS and fitting results showing the change of impedance after storage for 7 days, (b) magnitude of Bode plot, and (c) phase of Bode plot.

Figure 6. Morphological results of day one slurry: (a) FE-SEM images, (b) EDS mapping of carbon, (c) EDS mapping of fluorine, and (d) EDS mapping of cobalt.

In order to understand the cause regarding the impedance change of the slurry, the impedances of active materials and

conductive additives in the solvent were measured with two different slurry samples: first, active materials without 3921



Article

the equivalent circuit, RCT, CPE, and CDL represent a chargetransfer resistance, a constant phase element, and a double layer capacitance of the working electrode, respectively. The constant phase element was used to model the Warburg impedance, which is represented below:

Table 1. Summary of the Values and Errors of the Electrochemical Components Used for Data Fitting calculated impedance data working electrode

slurry

RCT (Ω) CPEE-C (F) CPEE-P CDL (F) RSO (Ω) RA&C (Ω) CPEB-C (F) CPEB-P

day 0 value

day 7 value

8971 1.35 × 10−6 0.75 1.56 × 10−8 2.5 × 107 1.01 × 105 1.4 × 10−5 0.18

17171 8.3 × 10−7 0.834 1.77 × 10−8 1.2 × 108 7.6 × 106 8.45 × 10−6 0.2

ZCPE =

1 C(iw)P

(1)

where C is the capacitance and ω is the angular frequency. When P = 0.5, a CPE represents a Warburg impedance. Equivalent circuit components in the slurry were chosen on the basis of the electrical path proposed in Figure 1 and the SEM image of freeze-dried slurry shown in Figure 6. First, the main component of the equivalent circuit is a resistance of conductive path by the network of the conductive additives and the active materials. The constituents are tightly bound and form electrical paths. Most of the current flows through conductive additives as the impedance of the conductive additives is lower than that of the active materials. Therefore, the impedance of the conductive additives is represented as an ohmic resistance (RA&C) in the equivalent circuit. The second component is CPEB which indicates a capacitance of the binder in the slurry. Since the binder is inherently nonconductive and forms a thin layer between the active materials/conductive additives and the solvent, it can be modeled as a capacitor. However, in this work, a constant phase element was used to attain a better fitting result to the experimental data. The last component is the solvent in the slurry. Due to ions in the solvent, it can be modeled as a resistor (RSO). Figure 5b,c shows typical impedance results of the slurry right after the preparation and storage for 7 days. Right after the preparation, the result exhibits relatively low impedance due to the tight mechanical integrity of the network between the active material, binder, and conductive additive. However, after 7 days of storage, the network of the slurry changes, and the overall impedance increased in the frequency region between 1 Hz and 100 kHz. The experiment was repeated several times. The EIS results from the slurry were fitted, as shown in Figure 5b,c. Data fitting was conducted in a frequency range between 1 Hz and 100 kHz, where the impedance change was detected. First, the resistance, capacitance, and constant phase elements of the working electrode were estimated on the basis of its geometry and the slurry solvent. After that, the values of the components in the equivalent circuit were obtained using data fitting on day 0. Using the values as initial points, another

conductive additives were mixed in the solvent, and the slurry was used for impedance measurements. Then, the same procedures were applied for conductive additives. Figure 3 demonstrates the impedance change of the slurry, active materials, and conductive additives in the solvent (NMP) after 1 and 2 days of storage without agitation. As shown in Figure 3c, the active materials in the solvent exhibited a slight impedance change in the entire frequency region, whereas as shown in Figure 3d, the conductive additives in the solvent exhibited the impedance change mainly in the low frequency region. Since the impedance in the low frequency region corresponds to a resistor, the ohmic conductive path of the conductive additives may have collapsed as time elapses. One possible reason is the aggregation of the conductive additives. As shown in Figure 3c,d, the impedance from the conductive additives is lower by 2 orders of magnitude than that from the active materials. Therefore, the impedance change induced by the conductive additives mainly contributed to the impedance change in the slurry presented in Figure 3a,b. Figure 4 illustrates the impedance change of the slurry, active materials, and conductive additives in the solvent after 1 and 2 days of storage under agitation. Both the active materials and the conductive additives under agitation showed larger impedance increases than those without agitation. As proposed earlier, agitation may accelerate the aggregation of the conductive additives, which results in the network breakdown of active material in the slurry. Equivalent Circuit Modeling for Slurry. In order to understand the morphological change of slurry over time, an equivalent circuit composed of simple electrical components was constructed. Figure 5a shows an equivalent circuit corresponding to a system comprising a working electrode and slurry to fit the impedance results on day 0 and day 7. In

Figure 7. Morphological results of not agitated day seven slurry: (a) FE-SEM image and (b) EDS mapping of carbon. 3922



Article

Figure 8. Morphological results of agitated day seven slurry: (a) FE-SEM image and (b) EDS mapping of carbon.

Figure 9. Results of image processing: (a) day one slurry, (b) not agitated day seven slurry, (c) agitated day seven slurry, and (d) dispersity index.

data fitting was conducted for the impedance result on day 7. The result is summarized in Table 1. Morphological Analyses of Slurry. The morphological analyses of freeze-dried slurries were performed to investigate the in situ internal structures of slurry components. The FESEM images of the slurries are presented in Figures 6a and 7a. The day one slurry shows a network structure (or flocculation) of active materials, which differs from the day seven slurry. The

interaction between binder and active materials is a major determinant of whether the network structure exists or not. Figures 6b and 7b show the EDS mapping images of the carbon conductive additives in the day one and day seven slurries. It is observed that the carbon particles of the day seven sample occupy a larger domain than those of the day one specimen. This is due to the aggregation of conductive additives over time. The result of the polymer binder reveals a similar 3923



Article

Figure 10. Micro-CT image: (a) day one slurry, (b) not agitated day seven slurry, and (c) agitated day seven slurry.

proposed. The equivalent circuit modeling indicates that the impedances for the active material, conductive additive, and binder increase, which was caused by the collapse of the network structure regarding the slurry over time. Further investigations by the electron microscope images, EDS images, and Micro-CT images of the slurry and constituting components confirm the schematic model of the LIB slurry. The findings show that EIS can act as a powerful tool to identify the internal structure of multiphase slurry systems compared with other analysis methods of slurries such as rheology and morphology. Furthermore, for the real application, EIS is expected to help control the mixing quality of slurries before the following coating process.

phenomenon (Figures 6c and S3a in Supporting Information). This binder behavior is presumed to lead to the carbon agglomeration as a result of the binder−carbon particle interaction. Figure 6d (also Figure S3b in Supporting Information) displays the EDS mapping images of cobalt, indicating LiCoO2. This result coincides with the FE-SEM image result. Figure 8 demonstrates the effect of agitation on the morphology of the slurry. Compared with the images shown in Figure 7, the breakdown of the network structure seems to accelerate due to mechanical stress. Also, the aggregation of carbon conductive additives and the degradation of polymeric binders are deteriorated by the agitation. The distribution state of slurry elements can be quantitatively analyzed to comprehend the internal structure of the slurry. Figure 9a−c presents the image processing steps necessary for computing the dispersity indices of conductive additives. The calculation result is given in Figure 9d. As the carbon conductives are aggregated, the dispersity index decreases. The dispersity index also provides information on the population density of particles as well as the dispersion state. In order to look into the network structure formed by active materials in a more in-depth morphological manner, Micro-CT images of the slurry are presented in Figure 10. Similar to the FE-SEM images, the presence and collapse of the network structure are identified. To sum up, the slurry material for a lithium-ion battery is a highly complicated system, thereby requiring comprehensive understanding of internal structures in regards to slurry constituents.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel.: +82-31-8005-3567. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Yoo, M.; Frank, C. W.; Mori, S. Chem. Mater. 2003, 15, 850−861. (2) Kuo, H. S.; Tsai, W. T. J. Electrochem. Soc. 2000, 147, 149−154. (3) Aurbach, D.; Markovsky, B.; Salitra, G.; Markevich, E.; Talyossef, Y.; Koltypin, M.; Nazar, L.; Ellis, B.; Kovacheva, D. J. Power Sources 2007, 165, 491−499. (4) Facci, T.; Huguenin, F. Langmuir 2010, 26, 4489−4496. (5) Yoo, M.; Frank, C. W.; Mori, S. Chem. Mater. 2004, 16, 1945− 1953. (6) Ligneel, E.; Lestriez, B.; Richard, O.; Guyomard, D. J. Phys. Chem. Solids 2006, 67, 1275−1280. (7) Dominko, R.; Gaberscek, M.; Drofenik, J.; Bele, M.; Pejovnik, S. Electrochem. Solid-State Lett. 2001, 4, A187−190.

■

CONCLUSIONS EIS, electron microscopy, EDS, Micro-CT, and dispersity analyses were conducted to understand the morphological change of the LIB slurry over time. The EIS measurement of the slurry with or without agitation showed that the electrochemical impedance increases as time elapses and becomes larger with agitation. On the basis of the EIS results, an equivalent circuit and a schematic model of the slurry were 3924



Article

(8) Lin, B.; Wen, Z.; Gu, Z.; Huang, S. J. Power Sources 2008, 175, 564−569. (9) Guan, H.; Wang, X.; Li, H.; Zhi, C.; Zhai, T.; Bando, Y.; Golberg, D. Chem. Commun. 2012, 48, 4878−4880. (10) Gupta, A. K.; Bajpai, R.; Keller, J. M. J. Polym. Res. 2008, 15, 275−283. (11) Li, C. C.; Lee, J. T.; Tung, Y. L.; Yang, C. R. J. Mater. Sci. 2007, 42, 5773−5777. (12) Li, C. C.; Wang, Y. W. J. Electrochem. Soc. 2011, 158, A1361− 1370. (13) Fongy, C.; Jouanneau, S.; Guyomard, D.; Lestriez, B. J. Power Sources 2011, 196, 8494−8499. (14) Song, J. Y.; Lee, H. H.; Wang, Y. Y.; Wan, C. C. J. Power Sources 2002, 111, 255−267. (15) Ariel, N.; Ceder, G.; Sadoway, D. R.; Fitzgerald, E. A. J. Appl. Phys. 2005, 98, 023516. (16) Qiu, X. Y.; Zhuang, Q. C.; Zhang, Q. Q.; Cao, R.; Ying, P. Z.; Qiang, Y. H.; Sun, S. G. Phys. Chem. Chem. Phys. 2012, 14, 2617− 2630. (17) Kim, H. S.; Periasamy, P.; Moon, S. I. J. Power Sources 2005, 141, 293−297. (18) Mantia, F. L.; Vetter, J.; Novák, P. Electrochem. Acta 2008, 53, 4109−4121. (19) Wu, X.; Chang, S. H.; Park, Y. J.; Ryu, K. S. J. Power Sources 2004, 137, 105−110. (20) Nagaura, T.; Tozawa, K. Prog. Batteries Solar Cells 1990, 9, 209− 217. (21) Gaberscek, M.; Dominko, R.; Jamnik, J. J. Power Sources 2007, 174, 944−948. (22) Lim, S. G.; Jin, E. M.; Zhao, X. G.; Park, K. H.; Kim, N. I.; Gu, H. B.; Park, B. K. J. Power Sources 2011, 196, 6631−6636. (23) Sawai, K.; Yamato, R.; Ohzuku, T. Electrochim. Acta 2006, 51, 1651−1655. (24) Patey, T. J.; Hintennach, A.; La Mantia, F.; Novák, P. J. Power Sources 2009, 189, 590−593. (25) Lee, G. W.; Ryu, J. H.; Han, W.; Ahn, K. H.; Oh, S. M. J. Power Sources 2010, 195, 6049−6054. (26) Sul, I. H.; Youn, J. R.; Song, Y. S. Carbon 2011, 49, 1473−1478.

3925


Correlation between Internal Structure and Electrochemical

Recommend Documents