Method Development for Quality Control of Suspensions for Lithium

---

Article pubs.acs.org/IECR

Method Development for Quality Control of Suspensions for Lithium-Ion Battery Electrodes Henning Dreger,*,†,‡ Marlene Huelsebrock,§ Linus Froboese,†,‡ and Arno Kwade†,‡ †

Institute of Particle Technology, Technische Universität Braunschweig, Volkmaroder Straße 5, 38104 Braunschweig, Germany Battery LabFactory Braunschweig (BLB), Technische Universität Braunschweig, Langer Kamp 19, 38106 Braunschweig, Germany § Feddem GmbH & Co. KG, Mosaikweg 19, 53489 Sinzig, Germany ‡

ABSTRACT: A method to determine the agglomerate and aggregate sizes of carbon black (CB), commonly used in anode and cathode suspensions for lithium-ion battery electrodes, is presented. An analysis via light diffraction and scattering was evaluated, and measuring parameters and the development of sample preparation are described in detail. Within this work, different dispersing additives were tested with regard to their ability to stabilize the CB agglomerates and aggregates after dispersing. Furthermore, a sample preparation routine was set up which enables the determination of CB particle sizes in about 10 min. This includes the separation of active material particles and the particle size analysis itself. Furthermore, the method was tested with discontinuously and continuously processed suspensions using a laboratory dissolver and a pilot-scale extruder. In these experiments, the progress of CB deagglomeration in the dispersing step could be proven. For this reason, the method represents a suitable instrument for a quality check in an early production stage. with regard to carbon black deagglomeration. While Kil et al.2 and Chang et al.4 present results from CB particle size analysis via dynamic light scattering (DLS) using dispersing additives, the analyzing process of manufactured slurries, measurement parameters, or any sample preparation for the analysis is not or only insufficiently described. The reasons for this lack of scientific literature regarding the particle size analysis of CB in electrode suspensions are the challenges in the sample preparation in combination with the measurement method. Because of the composition of the high-concentrated slurry with its multiple components, mainly large active material particles and agglomerates of CB, and the necessity for reliable sample preparation, several actions have to be taken for a reliable analysis. In this paper, we present the step-by-step development of an adapted analysis method for quality control in the dispersing step of electrode manufacturing. In this method, a direct measurement of CB agglomerates and aggregates sizes via light scattering and diffraction is used. A detailed insight into the analysis and sample preparation is also provided. In the first step, the refractive index of CB and solvent N-methyl-2-pyrrolidone (NMP) was determined using a software algorithm implemented in the analysis device. Next, a screening

1. INTRODUCTION In the near future, the production capacity of lithium-ion batteries will increase significantly. Capacities of several thousand megawatt-hours will be installed in Asia and North America,1 indicating a market ramp-up in the fields of consumer electronics, grid, and automotive applications. This market ramp-up is promoted by reduced production costs due to economy of scale at constant product quality. Product quality has to be ensured by a high degree of automation and a well-developed quality management at each step of the process chain. A reliable quality check in the dispersing step for the preparation of the anode and cathode slurries, usually consisting of active material, conductive agent such as carbon black (CB), and a binder diluted in a solvent, is of particular interest. This process step has a distinct influence on important cell parameters (e.g., electrochemical performance, calendric lifetime). Especially on the cathode side, the CB distribution plays an important role for enhanced electrode conductivity because typical active materials such as lithium nickel cobalt manganese oxide (NCM) are low conductors. Several publications focus on the homogeneous distribution and deagglomeration of CB in the slurry and subsequently in the electrode, partially using dispersing additives to enhance the battery performance.2−6 Usually, rheological characteristics like the viscosity or yield point of the suspension are determined for quality control reasons.2,4,6−8 These rheological characteristics are affected by different parameters like suspension formulation, especially the solids content, and represent only an indirect quality parameter © XXXX American Chemical Society

Received: Revised: Accepted: Published: A

June 1, 2016 January 5, 2017 February 13, 2017 February 13, 2017 DOI: 10.1021/acs.iecr.6b02103 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research for a suitable dispersing additive was carried out. The additives purpose is the stabilization of CB particles in processed slurry. Thus, the actual dispersing degree of CB, meaning the size of CB agglomerates and aggregates, is maintained, and no or minimal reagglomeration occurs. Furthermore, the influence of active material particles on the particle size analysis (PSA) of CB was investigated. The PSA of samples with or without active material particles was compared and the necessity for separation of active material and CB was proven. For method completion, the sample preparation sequence routine was evaluated. As a result, an analysis method for the determination of stressed CB agglomerates and aggregate sizes was developed. Finally, the method was tested using a laboratory scale dissolver and a continuously working extruder for the production of anode and cathode slurries. These experiments prove that the developed method is able to measure the progress of CB deagglomeration in the dispersing step. Thus, this method could be applied for further understanding of influence of CB morphology on cell performance and moreover used as an economic quality check in the large-scale production of electrode slurries for lithium-ion batteries.

Table 2. Dispersing Additves from Evonik Corp. product active name ingredient [%]

PVDF

CB

NCM

6 4

4 6

90 90

amount of solids content [% w/w] anode

dissolver

CB

SMG-SG1

extruder

dissolver

6

3

91

50

40

Tego 685 Tego 690

100 100

dibasic ester:methoxypropylacetate 1:1 methoxypropylacetate:butyl acetate 1:1.5 n/a n/a

120 500 85 000 35 000

nonvolatile ingredients [%]

solvent

density [g/mL]

52

methoxypropylacetate

1.01

100

n/a

1.06

n/a

n/a

1.05

homogenizer (HD 2200, Bandelin Corp.) with an effective high-frequency power of 200 W was used to disperse CB in NMP. Thus, 40 g of a NMP-based suspension with 5% w/w of CB and different stabilizers was produced. The amount of stabilizer originates from the average value of CB-related concentration ranges taken from datasheets of different stabilizers. These suspensions were dispersed for 8 min, and the CB particle size was measured after 1, 2, 4, and 8 min. Furthermore, a long-term analysis after a 120 h storage time at ambient temperature was carried out. 2.2.2. Dissolver. The discontinuous dispersing process in a dissolver represents a standard laboratory process. In this process, a fast rotating geared dissolver disc disperses the solids. The particles are mainly stressed by turbulent flow inside the mixing tank. First, all solids were dry-mixed in a rotary drum mixer (Turbula T2F, Willy A. Bachofen Corp.) with a rotational speed of 49 min−1 for 15 min, resulting in a coarse premix of all solids. Afterward, the mixed solids were added to the solvent in the dissolver (Dispermat CA, VMA-Getzmann Corp.) over a period of 5 min at low circumferential speed of 1.30 m s−1, resulting in a suspension volume of 150 mL. In the next step, the speed was raised to 9 m s−1 and maintained for 60 min. The suspension was cooled at 15 °C and degassed during dispersing. 2.2.3. Extruder. This continuous process uses a corotating twin-screw extruder (ZSK 18, Coperion Corp.) which are very common in, for example, the plastics industry, to continuously produce high viscous suspensions or molten masses. In addition, first experiments were carried out using the extrusion process for lithium-ion battery applications.11 Compared with the extrusion process presented in this paper, a standard process for the production of electrode suspensions also uses a dry mixing step where solids are homogeneously mixed. For the large-scale production of suspensions, planetary mixers, with a volume of several hundreds of liters, are used to mix all solids with a suitable solvent (e.g., NMP, water). After a dispersing time of several hours, the suspension is then continuously coated on the current collector. The dry mixing and dispersing step may vary with respect to used active materials, conductive additives, binders, or formulations.12 As mentioned before, all solids were dry-mixed in a rotary drum mixer with rotational speed of 49 min−1 for 15 min. Next, the mixed solids were transferred to a gravimetric twin-screw

60 − − 50 solids in suspension [% w/w]

PVDF

40

Disperbyk 2150 Disperbyk 2155 Byk 9077

solids in suspension [% w/w] extruder

Tego 672

product name

Table 1. Recipes for Cathode and Anode Suspensions

cathode

40

viscosity [mPas]

Table 3. Dispersing Additves from BYK Additives & Instruments Corp.

2. EXPERIMENTAL SECTION 2.1. Materials. The cathode suspensions were produced using commercial lithium nickel cobalt manganese (HED NCM-111, BASF Corp.; NCM) with an average particle size of x50 = 12 μm. The anode suspension was produced using commercial graphite (SMG-SG1, Hitachi Corp.; SMG-SG1) with an average particle size of x50 = 20.9 μm as anode active material. Carbon black (C-NERGY Super C 65, Imerys Graphite & Carbon Corp.; CB) conducting agent was added to the active materials. It consists of large agglomerates a few micrometers in size, while the CB aggregates have a particle size of about 100−500 nm consisting of sintered primary particles of about 32 nm.9 A polyvinylidene fluoride binder (PVDF Solef 5130, Solvay Solexis S.A.S.; PVDF) was used for solvent-based suspensions. N-Methyl-2-pyrrolidone (standard grade, BASF; NMP) was used as solvent, and suspensions were prepared using the compositions from Table 1.

amount of solids content [% w/w]

Tego 670

solvent

The screening of suitable stabilizers was carried out using seven different NMP-soluble dispersing additives from two different companies (Evonik Industries Corp., BYK Additives & Instruments Corp.). Additive properties are listed in Tables 2 and 3. All stabilizers have a steric stabilization mechanism using high-molecular polymers in different solvents. 2.2. Dispersing Methods. 2.2.1. Ultrasonic Homogenizer. For the first experiments, an ultrasonic disperser was used representing a standard laboratory dispersing device for small suspension amounts up to about 1 L which uses acoustic waves resulting in cavitation forces.10 In the first step, an ultrasonic B

DOI: 10.1021/acs.iecr.6b02103 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

for ni using a software algorithm included in the measuring device. To be able to define an appropriate CRI, the measured and the calculated light intensity distributions are compared to each other using the algorithm mentioned before. For this purpose, the R-parameter is defined as the deviation of the measured and calculated scattering light of each detector:19

feeder which is connected to the extruder via a funnel. The solvent was conveyed to the extruder by a peristaltic pump. The extruder (screw diameter: 18 mm) was operated at a temperature of 35 °C for all housing elements, with different rotational speeds up to n = 1200 min−1 and different conveyed volumes (cathode: 0.5 l/h, anode: 0.75 l/h). Average residence time in the extruder was about 5−10 min with maximum shear rates of up to 11.000 s−1. A schematic overview of the twinscrew device used for this experiment is shown in Figure 1.

1 R= N

2.3. Particle Size Analysis. For particle size analysis, measurements were executed with an optical device with 87 photodiode detector channels (LA-960, Horiba Corp.). A laser diffraction and scattering method was used based on Fraunhofer diffraction and Mie scattering theory.13 Devices using this theory measure the light intensity distribution generated by particles to calculate a size distribution. This distribution is compared to a calculation of a particle size distribution based on Fraunhofer and Mie theory for particles with a known diameter assuming a complete sphere. For small particles (1 μm). This tendency also remains beyond a dispersing time of 8 min. Long-term stability investigations (Figure 4) confirm the dispersing results from Figure 3. While stabilizers Tego 672 and 690 and Byk 9077 show a quite larger median particle size, the other stabilizers result in much lower and almost constant particle size. For this reason, further screening experiments are carried out only with the four following stabilizers: Disperbyk

Figure 4. Median particle size as function of long-term stability of CB.

2150 and 2155 and Tego 670 and 685. The stabilizer Tego 690, which shows almost comparable results in dispersing degree and long-term stability, was rejected to reduce further experimental trial. Beyond this, based on the fact that Tego 690 has a composition similar to the other tested Tego additives, an analysis of stabilizing additives with a different chemical composition has an additional scientific value due to changes in the formulation of suspensions adding active material and binder. In the next step, the impact of stabilizer concentration on the median particle size was investigated. In Figures 5−8 the median particle size of CB dispersed with Disperbyk 2150 and 2155 and Tego 670 and 685 is plotted over the dispersing time using the ultrasonic (US) homogenizer. While there is no significant concentration dependency on particle size for Disperbyk 2150 and 2155, a tendency for smaller particle sizes with higher concentrations of Tego 670 and 685 is revealed taking the dispersing time of 1 min into account. Except for Tego 670 (t = 4 min) and Tego 685 (15% concentration curve and t = 4 min), all other samples for both additives exhibit the same particle size of about 0.25 μm, which is characteristic for CB aggregates. Summarizing these results, the four different dispersing additives stabilize the carbon black aggregates and agglomerates efficiently. For Tego 670 and 685, slight concentration D

DOI: 10.1021/acs.iecr.6b02103 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

Figure 5. Median particle size of CB over dispersing time. Additive: Disperbyk 2150.

Figure 8. Median particle size of CB over dispersing time. Additive: Tego 685.

3.3. Preparation of Cathode Suspensions for Particle Size Analysis. 3.3.1. Separation of Active Material Compared to Unprepared Suspension. For the following experiments, only two stabilizers (Disperbyk 2150 and Tego 670) were used. Although both Byk stabilizers were slightly better in stabilizing the CB, Tego 670 was chosen because of a different chemical composition representing an alternate additive compared to additives from Byk. The different chemical composition becomes important because in the following experiments active material and binder were added, changing the slurry formulation; thus, they are able to influence the stabilizing process. Therefore, differences in chemical composition of stabilizers are beneficial. In this step, NMP-based cathode suspensions consisting of active material, binder, and CB (see Table 1) were produced in a dissolver as described in section 2.2.2 to investigate whether a separation of active material is necessary for the correct CB particle size measurement. In the following experiments, the concentration in the suspension was set to 80% w/w of all solids for both additives. This ensures an excess of stabilizer, which is available for CB particle surface even if the stabilizer adsorbs on the active material particle surface. Thus, two sample preparations were realized: In the first, suspensions were diluted for the PSA without stabilizer (sample size, ∼12 mL). In the second, one sample of each suspension was mixed with one of the two stabilizers. In the next step, these mixtures were diluted and centrifuged (Biofuge Primo, Heraeus Corp.) for 5, 10, and 15 at 500 min−1 before particle size analysis. The centrifugation step separates the larger and heavier active material particles from the CB still floating in the solvent. Based on the following theory,22 the distance, tE, a spherical particle travels in a centrifugal field is described as

Figure 6. Median particle size of CB over dispersing time. Additive: Disperbyk 2155.

tE =

⎛r⎞ 18η ln 2 2 ⎜ ⎟ (ρP − ρF )d ω ⎝ r0 ⎠

(5)

where η is the dynamic viscosity of solvent; ρP and ρF are the particle and fluid density, respectively; d is the particle diameter; ω is the angular speed; r0 (6 cm) and r (12.4 cm) are the orbit starting and end radius, respectively. Therefore, the distance covered in the centrifuge is at maximum 6.4 cm. While active material particles are spherical (NCM) or flakelike (SMG-SG1), CB generates nonspherical, fractal, porous

Figure 7. Median particle size of CB over dispersing time. Additive: Tego 670.

dependencies were observed in the investigated concentration ranges. As a result, all stabilizers seem to be suitable for the stabilization of CB. E

DOI: 10.1021/acs.iecr.6b02103 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

This large light transmission indicates an insufficient particle amount in the measuring cell and hence in the supernatant sample volume, resulting in an error message from the PSA device. Because of their low concentration in the supernatant and an error message from the PSA device, the authors assume that active material particles have only an inferior impact on the scattering light emission of centrifuged samples, although a certain influence is possible. In addition, sample centrifugation emphasizes a bimodal particle size distribution (see Figures 9 and 10), with one peak at about 1 μm, representing CB agglomerates, and the other at about 200 nm, representing CB aggregates. In this case, differences become more apparent because of a higher resolution. In contrast, without centrifugation, a large peak at active material particle sizes (>3 μm) arises, which reduces the CB peak resolution. 3.3.2. Preparation Routine. To quantify the mixing sequence influence for the PSA preparation, three different preparation routines were tested, where the mixing sequence was varied as in Table 6.

structures due to the branched structure of aggregates forming agglomerates. For this reason, the agglomerate density ρP of CB is significantly lower as a spherical and solid particle of the same size. While all other parameters (see eq 5) are considered constant, the particle density of CB agglomerates differs from the theoretical density of a spherical, solid particle of the same size. This results in a different behavior in a centrifugal field in terms of distance traveled. As a result, a good separation of CB active material particles in diluted suspension using centrifugation was found. Solid active material particles are separated after 5 min of centrifugation (Figures 9 and 10), as indicated by the disappearing

Table 6. Tested Sample Preparation Routines routine

sequence of addition

1 2 3

suspension → stabilizer → NMP suspension → NMP → stabilizer NMP → stabilizer → suspension

The NMP-based cathode suspensions were produced with a dissolver (see section 2.2.2), and samples were prepared using the following two stabilizers with a stabilizer concentration of 80% w/w of all solids each: Disperbyk 2150 (sample A) and Tego 670 (sample B). As expected, the sample without centrifugation in Figure 11 shows the largest particle size caused by the active material particle influence in diluted suspension. Looking at the two different stabilizers, we can identify that the particle size distribution using both stabilizers is in a range comparable with a d90 of 2−2.5 μm. With regard to the sample preparation, the first routine (Figure 11; A1, B1) reveals the smallest median particle sizes for both stabilizers, meaning less CB reagglomeration. Therefore, routine A1 was used in further experiments. Overall, the developed sample preparation and particle size measurement requires about 10 min. Furthermore, it could be reduced by shortening the centrifugation time or by sample measurement automation. 3.3.3. Dispersing Progress of Cathode Suspensions. In the next step, the dispersing progress of a standard cathode suspension in a dissolver experiment and its reproducibility with respect to the PSA were investigated with a NMP-based cathode suspension (see Table 1). For this purpose, the sample preparation was carried out three times using Tego 670. Reproducibility is characterized by standard deviation, which exhibits no significant difference to minimum or maximum particle size values, although only small sample numbers were taken. In Figure 12, a smaller particle size is observed with increasing dispersing time. This observation agrees with the fact of more frequent stress input deagglomerating the CB over time. The largest change in particle size (from 2.75 to 1.75 μm) is obtained after a short dispersing time of 5 min, where large and unstable agglomerates are separated into smaller

Figure 9. Particle size distribution of un/prepared samples with varying centrifugation time. Additive: Disperbyk 2150.

Figure 10. Particle size distribution of un/prepared samples with varying centrifugation time. Additive: Tego 670.

peak at 10 μm. On the other hand, there are still particles in a range of a few micrometers (10) is needed to ensure well-founded statements. In Figure 13 the volume density distribution of the suspension prepared by dissolver is plotted as a function of the particle size. Here, unmodified samples with a dispersing time of 0 and 15 min are compared to samples which have been centrifuged. Obviously, the unmodified samples, comprising active material particles, show a large peak at 10 μm. This peak almost absorbs the whole CB agglomerates peak at ∼2 μm because of the active material particle influence with their dominant light scattering. In addition, the differences in dispersing progress are only slightly visible in the unmodified samples. In contrast, the centrifuged samples show notably larger CB peaks with a peak shifting from 3 to 1 μm between 0 and 15 min dispersing time, caused by deagglomeration of CB. G

DOI: 10.1021/acs.iecr.6b02103 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research

standard deviation. These findings contradict the results of the PSA in section 3.4, where the standard deviation is much smaller. At this point, the reasons for these divergences are not completely understood, but the authors suggest again a higher sample number (e.g., >3 samples) to reduce this phenomenon and achieve better results.

4. CONCLUSIONS We developed a method to analyze the CB dispersing behavior in suspensions for lithium-ion battery electrodes based on laser diffraction and scattering. Furthermore, a sample preparation was established, in which the CB is stabilized using different additives and separated from active material. In the first step, the measuring device parameters were evaluated and a standard measuring routine was developed using a model suspension consisting of CB and solvent. Next, seven different dispersing additives were tested with respect to their ability to stabilize the CB dispersing degree. Within these experiments, the short- and long-term stability was tested. On the basis of these results, four stabilizers with different concentrations were used for further experiments. Results indicate a negligible influence on the CB particle size over a wide concentration range. In the next step, we proved the necessity of separation in sample preparation using a multicomponent electrode suspension: PSA carried out with unprepared suspensions shows significant influence by active material particles. The high concentration of these particles suppresses and sophisticates the share of CB particles, emphasizing the importance of active material separation. Additionally, a sample preparation sequence was developed and used for different dispersing processes demonstrating the ability to determine the CB dispersing progress in electrode suspensions. Experiments carried out with a discontinuously working dissolver led to reliable results with cathode and anode suspensions, showing the CB dispersing progress with increasing dispersing time. The method was then applied to suspensions which were processed continuously using an extruder. Again, a progress in CB dispersing was observed with small standard deviations on the part of the cathode suspensions. On the anode side, larger standard deviations reduce the reliability of the method, although the progress of the dispersing degree is still obvious. Summarizing all results, we are able to describe the dispersing progress by the degree of CB deagglomeration in suspensions, which is one important aspect for the quality of electrodes because of the significant influence on structure and electrochemical performance.11 Furthermore, this method has the potential for a fast standard quality check (∼10 min) in continuous mass production, where parameters like the CB dispersing degree must be kept constant for ensuring a high quality level and traceability of products. Another advantage besides the short preparation time is the direct measurement of a quality parameter compared to the indirect analysis of CB dispersing degree and deagglomeration using rheological methods.

Figure 14. Median particle size as a function of dispersing time. Additive: Tego 670.

3.5. PSA Method Application on Continuously Processed Suspensions. In the next step, the method was applied to continuously processed suspensions manufactured with the twin-screw extruder (see section 2.2.3). Therefore, cathode and anode suspensions were produced (see Table 1) using different circumferential screw speeds between 120 and 1200 min−1. Concentration of stabilizer used for the sample preparation was maintained at 80% w/w for cathode suspensions respectively 100% w/w of solids for the anode suspension. In Figure 15, the median CB particle size (d50) of

Figure 15. Median particle size, d50 (CB), of continuously processed anode and cathode suspensions. Additive: Tego 670.

one anode and two cathode experiments is plotted as a function of different screw speeds. Examining the cathode experiments, it becomes apparent that the median particle size decreases from ∼1.35 to 0.4 μm with higher tip speeds or higher stress number, respectively. Additionally, there are only minor differences between both cathode experiments, and standard deviations are moderate, leading to reasonable results. On the other side, the anode experiment confirms the results of cathode experiments with a decreasing CB particle size from ∼1.5 to 0.5 μm with higher stress intensities. In contrast to the cathode experiments, the anode experiment reveals a higher

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: +49-531-391 94655. Fax: +49-531-391 9633. ORCID

Henning Dreger: 0000-0001-9613-0325 H

DOI: 10.1021/acs.iecr.6b02103 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Article

Industrial & Engineering Chemistry Research Notes

(19) Horiba. Understanding the Chi Square and Residual R Parameter 2008. http://www.horiba.com/fileadmin/uploads/ Scientific/Documents/PSA/TN153.pdf. (20) Won, Y.-Y.; Meeker, S. P.; Trappe, V.; Weitz, D. A.; Diggs, N. Z.; Emert, J. I. Effect of Temperature on Carbon-Black Agglomeration in Hydrocarbon Liquid with Adsorbed Dispersant. Langmuir 2005, 21 (3), 924−932. (21) Probst, N. In Carbon Black - Science and Technology; Donnet, J.B., Bansai, R.-C., Wang, M.-J., Eds.; CRC Taylor & Francis/Marcel Dekker, Inc.: New York, 1993; pp 271−288. (22) Stieß, M. Mechanische Verfahrenstechnik - Partikeltechnologie 1. Springer: Berlin, 2008. (23) Schilde, C.; Kampen, I.; Kwade, A. Dispersion kinetics of nanosized particles for different dispersing machines. Chem. Eng. Sci. 2010, 65 (11), 3518−3527.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We kindly thank the Federal Ministry for Economic Affairs and Energy for funding of the projects S-ProTrak and DaLion (Project Grant Number: 01MX13003B, 03ET6089) whereby this work could be realized.

■

REFERENCES

(1) Chung, D.; Elgqvist, E.; Santhanagopalan, S. Automotive Lithiumion Battery (LIB) Supply Chain and U.S. Competitiveness Considerations; NREL/PR-6A50-63354; CEMAC, National Renewable Energy Laboratory: 2015. (2) Kil, K. C.; Kim, G. Y.; Cho, C.-W.; Lim, M. D.; Kim, K.; Jeong, K.-M.; Lee, J.; Paik, U. Acid−base interaction between carbon black and polyurethane molecules with different amine values: Dispersion stability of carbon black suspension for use in lithium ion battery cathodes. Electrochim. Acta 2013, 111 (0), 946−951. (3) Dominko, R.; Gaberscek, M.; Drofenik, J.; Bele, M.; Pejovnik, S.; Jamnik, J. The role of carbon black distribution in cathodes for Li ion batteries. J. Power Sources 2003, 119−121 (0), 770−773. (4) Chang, C.-C.; Her, L.-J.; Su, H.-K.; Hsu, S.-H.; Yen, Y. T. Effects of Dispersant on the Conductive Carbon for LiFePO4 Cathode. J. Electrochem. Soc. 2011, 158 (5), A481−A486. (5) Patey, T. J.; Hintennach, A.; La Mantia, F.; Novák, P. Electrode engineering of nanoparticles for lithium-ion batteriesRole of dispersion technique. J. Power Sources 2009, 189 (1), 590−593. (6) Lee, J.-H.; Kim, J.-S.; Kim, Y. C.; Zang, D. S.; Paik, U. Dispersion properties of aqueous-based LiFePO4 pastes and their electrochemical performance for lithium batteries. Ultramicroscopy 2008, 108 (10), 1256−1259. (7) Lee, G.-W.; Ryu, J. H.; Han, W.; Ahn, K. H.; Oh, S. M. Effect of slurry preparation process on electrochemical performances of LiCoO2 composite electrode. J. Power Sources 2010, 195 (18), 6049−6054. (8) Bitsch, B.; Dittmann, J.; Schmitt, M.; Scharfer, P.; Schabel, W.; Willenbacher, N. A novel slurry concept for the fabrication of lithiumion battery electrodes with beneficial properties. J. Power Sources 2014, 265, 81−90. (9) Spahr, M. E.; Goers, D.; Leone, A.; Stallone, S.; Grivei, E. Development of carbon conductive additives for advanced lithium ion batteries. J. Power Sources 2011, 196 (7), 3404−3413. (10) Bittmann, B. A. Ultraschalldispergierung von anorganischen Nanopartikeln in Epoxidharz und Charakterisierung der resultierenden Werkstoffe; IVW: Kaiserslautern, Germany, 2009. (11) Dreger, H.; Bockholt, H.; Haselrieder, W.; Kwade, A. Discontinuous and Continuous Processing of Low-Solvent Battery Slurries for Lithium Nickel Cobalt Manganese Oxide Electrodes. J. Electron. Mater. 2015, 44 (11), 4434−4443. (12) Yoshio, M.; Brodd, R. J.; Kozawa, A. Lithium-Ion Batteries: Science and Technologies; Springer: New York, 2010. (13) Mie, G. Beiträge zur Optik trüber Medien, speziell kolloidaler Metallösungen. Ann. Phys. 1908, 330 (3), 377−445. (14) Pohl, M. C. Light Scattering. In Powder Metal Technologies and Applications; ASM Handbook; ASM International: Materials Park, Ohio, 1990; Vol. 7, pp 250−255. (15) Trofimowa, T. J. Physik. In Physik; Vieweg: Verlag, 1997; p 227. (16) Fox, M. Optische Eigenschaften von Festkörpern; Oldenbourg Wissenschaftsverlag: München, Germany, 2012. (17) Kinoshita, T. The method to determine the optimum refractive index parameter in the laser diffraction and scattering method. Adv. Powder Technol. 2001, 12 (4), 589−602. (18) Bond, T. C.; Bergstrom, R. W. Light Absorption by Carbonaceous Particles: An Investigative Review. Aerosol Sci. Technol. 2006, 40 (1), 27−67. I

DOI: 10.1021/acs.iecr.6b02103 Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX