New Contact Probe and Method to Measure ... - Wiley Online Library

DOI: 10.1002/ente.201600127

New Contact Probe and Method to Measure Electrical Resistances in Battery Electrodes Nils Mainusch,*[a, b] Torge Christ,[a] Thammo Siedenburg,[b] Tom OQDonnell,[c] Meylia Lutansieto,[c] Peter-Jochen Brand,[c] Gerhard Papenburg,[d] Nina Harms,[d] Bilal Temel,[e] Georg Garnweitner,[e] and Wolfgang Viçl[a, b] Electrical resistivity is an important measure to qualify electrodes for lithium-ion batteries. A reliable determination of conductivity is of high practical importance with regard to, for example, electrode production improvements and quality control. To complement state-of-the-art measuring techniques, a new method has been developed based on a new “micron-powder probe”. Following a simple measuring procedure, the system allows nondestructive, highly reproduci-

ble, and rapid data acquisition. In this paper, we describe the new concept thoroughly and present experimental results. These results demonstrate that an initial determination of resistance values in battery electrodes is beneficial especially if it is combined with an electrical postmortem analysis of cycled cathode disks. The outcome of our investigation is validated with regard to the electrochemical performance of cathodes in half-cells.

Introduction The amount of energy that can be extracted from lithium-ion batteries (LIB) is related directly to the voltage levels that are spanned in the course of cell operation. Phenomena that cause undesirable voltage losses are electrochemical polarization and internal resistances.[1] Internal resistances are partially attributed to imperfect electron transfer. LIB electrodes usually consist of metal foils (current collectors) that are coated with a particle-based composite (active layer). Active layers are designed typically with a thickness of 50–150 mm, which depends on the targeted application (e.g., high-power vs. high-energy cells). The electronic pathways within the composite are aggravated by insulators (electrode active material) and incoherent components (organic binder, carbon black, and graphite). As the particles are of micron or submicron scale, Ohmic drops occur at multiple interfaces. Moreover, it is proven that electric barriers occur on the surface of current collectors and might hamper conductivity. This can be a limiting factor for the overall cell performance and impact the usable battery capacity, respectively.[2, 3] State-of-the-art options to maximize electron transport comprise the doping and coating of cathode active materials,[4–6] incorporation of conductive additives,[7] and current collector modifications.[8–10] Apart from the use of tailored preproducts, the entire electrode fabrication affects the quality of the conductivity network of the electrode. Optimized manufacturing implies ideal slurry formulation and the perfect dispersion of the constituents as well as an optimal coating and drying routine in addition to adequate compaction (so-called “calendering”). Electrical conductivity is an important criterion to qualify elaborated electrodes, and quality assessment requires electrochemical tests. These are both timeand material-consuming so it has to be ensured that, in terms

of conductivity, only the best electrodes are submitted to extensive testing. In conclusion, it becomes clear that the determination of conductivity from electrode sheets is highly ben-

[a] N. Mainusch, T. Christ, Prof. Dr. W. Viçl Faculty of Natural Sciences and Technology University of Applied Sciences and Arts Hildesheim/Holzminden/Goettingen Von-Ossietzky-Str. 99, 37085 Gçttingen (Germany) E-mail: [email protected] [b] N. Mainusch, T. Siedenburg, Prof. Dr. W. Viçl Application Center for Plasma and Photonics APP Fraunhofer Institute for Surface Engineering and Thin Films IST Von-Ossietzky-Str. 100, 37085 Gçttingen (Germany) [c] T. O’Donnell, M. Lutansieto, Dr. P.-J. Brand Center for Tribological Coatings Fraunhofer Institute for Surface Engineering and Thin Films IST Bienroder Weg 54 E, 38108 Braunschweig (Germany) [d] G. Papenburg, N. Harms Technische Universit-t Braunschweig Institute of Environmental and Sustainable Chemistry Hagenring 30, 38106 Braunschweig (Germany) [e] B. Temel, Prof. Dr. G. Garnweitner Technische Universit-t Braunschweig Institute for Particle Technology Volkmaroder Str. 5, 38104 Braunschweig (Germany) The ORCID identification number(s) for the author(s) of this article can be found under http://dx.doi.org/10.1002/ente.201600127. T 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA. This is an open access article under the terms of the Creative Commons Attribution Non-Commercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited, and is not used for commercial purposes. Part of a Special Issue on “Li-Ion Batteries”. To view the complete issue, visit: http://dx.doi.org/10.1002/ente.v4.12

Energy Technol. 2016, 4, 1550 – 1557 T 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

1550

eficial to refine electrode production and enable quality control.1 However, there is little work published on viable methods to conduct such measurements. Both simple two-point and extended four-point probe techniques are well known.[11] The first generally suffer from “parasitical” resistances, which are Ohmic losses that arise because of an imperfect coupling of a contact pin to an electrode surface (contact resistances, Rcontact ; Figure 1 C). Four-point probe techniques permit the elimination of parasitical resistances. Moreover, they can allow the separate determination of electrode composite conductivities (or composite resistance, Rcomposite) and interfacial resistances, Rinterface, that occur between the current collector and the composite.

Figure 1. A–C) Experimental setup to measure electrical conductivities in electrode sheets. A) The powder probe. Its particles (red spheres) are attached to a magnetic mount and they contact the surface of a composite electrode that consists predominately of dielectric LMO particles (black spheres). The particle size distribution of the powder probe is similar to that of the LMO particles. The red dots on top of the probe particles represent electrically conductive silver deposits (Figures 2 and 3). Different from this, the dashed red lines in (A) and (B) indicate electron pathways that penetrate the matrix of the composite electrode. B) An indented contact stamp and the resulting plastic deformation of the electrode are demonstrated. The schematic of the proposed technique in (A) versus the conventional stamp (B) shows the reduction of parasitical contact resistances in the case of a powder probe application. One explanation for this is that spherical particles shift easily. Therefore, the powder collective conforms to uneven surfaces and will inherently fill in surface gaps or macroscopic roughness. Beyond this, the particular structure of the Ag coating (Figure 2) is believed to promote the conductivity by bridging microscopic voids and efficiently coupling the carbon black conductive network of the electrode, respectively. In contrast, any inflexible stamp will be interstitial even if it is indented. C) Different resistance contributions. Electric flux lines are also plotted.

1

Notably, the electronic properties will change as dry electrode sheets are saturated with liquid electrolyte because of matrix swelling.

Energy Technol. 2016, 4, 1550 – 1557

Wang and co-workers[12] presented a method that allows a layerwise extraction of the electronic resistivity in electrode sheets. To do so, they impose four contacts in a linear arrangement on top of the electrode. As the electrodes might be horizontally stratified, such a depth-profile analysis can provide information on electrical in-plane characteristics. It can also give information about functional film properties in the case of a previous functionalization of a current collector surface, for example. In addition to the determination of an electrode composite resistivity, they specify current collector sheet resistances, Rsq. [W m2]. Here, a key finding is that Rcomposite is in general much lower than Rinterface.2 A problematic aspect of this method is that all results feature high standard deviations (case-dependent up to & 80 %), and it thus possesses little reproducibility. Furthermore, the arrangement requires the use of elongated electrode samples ( & 10 cm in length), and conductivity analysis is done in a horizontal orientation, which does not coincide with the predominant direction of electron transfer in cells in operation, which is basically perpendicular to the composite layer. Ender et al.[13] presented an alternative method. Their setup makes use of a disk-shaped electrode sample, and a constant current is induced at its center by means of a tip. Surface potentials are evaluated as a function of radial distances, and the determination of composite resistivity as well as collector sheet resistances is done by numerical simulations that are fitted to experimental surface potential data for specific radial resistances. The input for the calculation is the electrode geometry. Again, an important outcome of the proof-of-concept study is the collector sheet resistances that dominate composite resistivity. The comparison of different types of cathodes shows that Rinterface is much higher than Rcomposite. The multiplier of Rinterface to equal Rcomposite ranged between 105 and 733, which depends on the investigated cathode sheet. A drawback of this method is the scattering of the measured values. Standard deviations of repeated measurements on identical samples are predominantly 20– 30 %. This is explained by possible misalignment of the tips of the measurement unit. Another problem occurs on closer inspection of the given results that specify LiMn2O4 (LMO) cathodes: A first sample with a thickness of 144 mm is assigned to have Rcomposite & 90 mW and Rinterface & 14.9 W, thus a total electrode resistance of & 15 W. However, a thicker specimen of 170 mm shows Rcomposite & 60 mW, Rinterface & 6.2 W, and in total & 6.2 W. This appears implausible. The methodology makes use of surface potential simulations that are based on the model of a hypothetical, ideal current field distribution. Here, the presupposition is a homogeneous electrode matrix. This excludes the probable existence of zones with discrete electric conduction that are localized in-plane within the matrix. The inconsistency could be explained because such zones cannot be considered in the underlying model. With the intention to address deficiencies in electrical twopoint probe measurements (parasitical contact resistances) as 2

Resistances were calculated from the given layer dimensions in accordance with the published composite resistivity or sheet resistance, respectively.

T 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

1551

well as how to overcome the discussed disadvantages in fourpoint methods (high standard deviations, inconsistent results) a new approach has been developed. The concept involves essentially two elements: First of all, a specific “powder probe” was implemented in an extended two-point measuring setup. With the powder probe, we target an initial reduction of unwanted contact resistances. Second, a particular measuring procedure was established. This aims at the identification of those contact forces that are needed to compensate for contact resistances.

Results and Discussion Given below is a description of the powder probe, its setup, and its methodological characteristics. Afterwards contact resistances for conventional stamp versus probe measurements were evaluated. A validation of conductivity measurements on LMO cathodes that include the electrochemical performance is given, and finally some examples for advanced probe applications are presented. Method It is well known that parasitical electrical losses virtually always occur if a rigid contact medium is applied to a (microscopically) rough surface. One reason for this is an imperfect adaption. Coupling can be optimized, for example, by indenting a stamp into the surface of an electrode. By doing so, the matrix will be deformed plastically, and the electrical connection between stamp and the conductive network will be somewhat improved, however, this is associated with possible changes in the electrical properties of the composite matrix as well. This situation is depicted in Figure 1 B. Different from a solid stamp, the system with the new powder probe basically consists of a collective of particles that serve as contact media. This collective features high plasticity, thus it conforms inherently to the microtopography of the surface as it is driven towards an object.[14] In contrast to inflexible devices, it is the contact media that is deformed and it couples in a nondestructive manner to the conductivity network of an electrode. This setting is illustrated in Figure 1 A. The schematic shown in Figure 1 C enlists the various resistance contributions that can be summed up to an overall resistance. In the middle of the schematic Relectrode indicates all domains that in total represent the electrode resistance (with the exclusion of Rcontact). Furthermore, wiring and contacting that is used to induce the measuring current (typically 1–100 mA) and the separated circuit for the potential detection is displayed. The powder probe is composed of functional micron-sized particles with a mass-weighted median diameter d50 of approximately 15 mm. The particles are spherical and are made from magnetic stainless steel. By means of physical vapor deposition (PVD), the particle surfaces were covered with a highly conductive silver coating. The reference material, functionalized particles, the particle collective, and a complete powder probe are shown in Figures 2 and 3. Energy Technol. 2016, 4, 1550 – 1557

Figure 2. Documentation of spherical stainless-steel particles used for the design of the powder probe. Scanning electron microscopy and transmission electron microscopy (insets) images of pristine particles (top) and Ag-coated particles (bottom). In both cases, particles with a diameter of approximately 10 mm were chosen for the analysis. As can be seen for the Ag-coated particles (bottom), the plasma sputter process permits an all-around covering of the starting material. The silver coating exhibits a granular structure with a particular submicron roughness. The TEM image (bottom inset) indicates a roundish morphology of an isolated Ag deposit on the powder probe material. However, untreated particles possess a comparatively smooth surface with characteristic small fissures and unspecific surface debris (top inset).

Figure 3. Left image: Light microscopy image of stainless-steel particles attached to a magnet. The particles are spread because of the magnetic field. They adopt an antenna-like and pointed structure that constitutes a 3 D shape. The inset documents how the particles are strung together. Right image: The entire powder probe with a copper mount, an inserted magnet, and the attached particle collective.

The actual laboratory setup is displayed in Figure 4. The middle of the photograph shows the force sensor and the appended powder probe. A magnet is inserted into the copper mount, which fixes the particle collective. The probe micro-


1552

Figure 4. Laboratory setup. The photograph documents the force sensor (1), the copper mount with attached powder probe (2), a polymer foil with 8 mm diameter apertures (3), and a LFP laminate (4).

particles are highly magnetizable, which prevents powder losses. Nonetheless, after powder probe application residual powder material lodged in the electrode could sometimes be observed, and these traces can be dusted off and removed thoroughly. Therefore, we do not expect to contaminate electrode sheets with unwanted material (such as iron) and preserve examined electrode sheets for electrochemical followup investigations. A force sensor and a powder probe were adjusted by using a micrometer gauge. Modification from a powder probe to a stamp configuration is feasible. In powder probe applications, a mask borders the contact area. The measuring object shown in Figure 4 is a LiFePO4 (LFP) stand-alone sheet (i.e., an electrode without a subjacent current collector, hereafter referred to as “laminate”). The laminate is placed in between a mask with 8 mm apertures arranged linearly and a dielectric backplate.

A typical result of resistance measurements is shown in Figure 5. In this case the object is a LMO cathode. For each measurement the resistance (left ordinate) and the applied force (right ordinate, red dashed curve) were recorded over a period of 100 s. The comparison of contact stamp (8 mm diameter, gold coated surface) and powder probe (8 mm mask) measurements involves an initial base line acquisition (gray calibration curves) to control the setup performance. This is done by contacting the grounded copper backplate directly, and a slightly lower level is obtained in the case of the contact stamp. In the examination of the cathode sheet, the probe renders significantly lower values ( & 4.7 W) than the stamp (> 65 W). Furthermore, a second measurement reveals excellent data conformity for the probe, whereas the stamp produces an offset. A decrease of parasitical resistances and superior reproducibility for probe measurements can be stated. Resistances in cases of a discrete force increase from 2– 10 N in the aforementioned LFP laminate, which again compare stamp and powder probe measurements, are shown in Figure 6. The effective measuring track is 50 mm, and the measurements were repeated five times. The dashed lines represent fit curves. An algorithm based on a power function was chosen because this corresponds to the physics of contact resistances under external loads.[15] It is clear that in the case of a stamp, the resistance gradient is pronounced and, correspondingly, the calculated exponent (& @0.04) is comparatively high (probe & @0.005). For the probe, only a marginal influence of the contact force on the resistance occurs, and saturation at low forces can be seen. Again, reproducibility for the powder probe is excellent (standard deviation < 0.15 W, not plotted here), and the resistance disparity between the two devices ranges from 519 to 362 W at 2 and 10 N, respectively. A simple extrapolation of the fitting curve of the stamp leads to a force of > 45 N that would have to be imposed to decrease the resistance of the contact stamp to the level of the powder probe. This, in

Figure 5. Survey of resistance measurements on a LMO cathode sheet. Comparison between contact stamp and powder probe over 100 s measuring time.

Energy Technol. 2016, 4, 1550 – 1557


1553

Figure 6. Electrical resistances in a LFP stand-alone electrode sheet as a function of contact forces of 2–10 N. Comparison between stamp and powder probe measurements with a measuring distance of 50 mm.

turn, would provoke sample deformation. Consequently, an undesired structural change of the measuring object itself would take place.

Analysis of contact resistances To determine the quantity of unwanted contact resistances that occur in two-point measurements, investigations on LFP laminates were performed. The use of a laminate allows an in-plane investigation of Ohmic drops as a function of the distance between, for example, one fixed contact and its variable counterpart that is applied to the test structure. Any length of this test structure defines a specific “measuring distance”, and the resulting values correspond to the sum of distance-related contact resistances plus a specific Rcontact. Thus, the approximation d!0 mm gives solely Rcontact. Based on a linear regression of the data (least-square fit) the determination of y axis intercepts and the assignment of Rcontact becomes possible, and one outcome is shown in Figure 7.

The extrapolation of the probe data (gray dotted line) features a y axis intercept that is close to zero. This indicates that probe measurements on the LFP laminate hardly implicate contact resistances or that those must be extremely small compared with the overall resistance. In contrast, the y axis intercept of the stamp is > 300 W. Apart from the evaluation of y axis intercepts, the contact resistance can be deduced by “multiple contacting” and cumulating segment-related resistances. This idea means that first the resistance over a fixed distance is determined in a single measurement. The obtained value is afterwards subtracted from a cumulative resistance value. This value is achieved from multiple measurements performed at two (or more) segments that in total equal the originally chosen distance (e.g., two measurements over 2 cm tracks minus one measurement over a distance of 4 cm). A third possibility to assess the contact resistance of the stamp is given by simply compensating results with values from corresponding probe measurements according to Rstamp@Rprobe. Contact resistances that account for the three methods are shown in Table 1. Good overall data consistency exists and, in particular, the calculated y-axis intercepts deliver figures that are in good agreement with those of the multiple contacting method.

Table 1. Contact resistance values in case of the application of a stamp in contact with a LFP laminate. Force [N]

Rcontact (y-axis intercept) [W]

Rcontact (multiple contact) [W]

Rcontact (probe compensation) [W]

2 4 6 8 10

477 399 350 337 321

484 400 349 336 323

503 429 425 414 367

The overall resistance scales with the elongation of an electrode laminate and, in principle, invariant contact areas in each measurement exist and Rcontact remains constant. At the same time it decreases proportionally with the increasing measuring distances and vice versa. The corresponding R/d ratio for a stamp and a probe are plotted in Figure 8. The curve of the stamp increases steeply at short distances. Conversely, the curves stay fairly constant as the probe is applied. This depiction illustrates how parasitical resistances might dominate the measurement. According to this, for short measuring distances (e.g., across a 100 mm thin electrode layer), Relectrode would not allow analysis by using a stamp because the obtained value basically represented parasitic contact resistances. Validation of LMO cathodes

Figure 7. Electrical resistances in a LFP stand-alone electrode sheet as a function of measuring distances from 20–100 mm. Comparison between stamp and powder probe with applied measuring forces of 2 and 10 N.

Energy Technol. 2016, 4, 1550 – 1557

As stated initially, interfacial resistances can hamper electron transfer between the electrode layer and a current collector. For LMO cathodes, Ender et al. found that Rinterface exceeds


1554

Figure 8. Ratio resistance/distance plotted for all screened electrode distances. Comparison between stamp and powder probe measurements.

Rcomposite by at least 100 times.[13] Thus, it can be assumed that the electrochemical performance of a cathode is improved by enhancing interfacial conductivity. To confirm this and validate the practicability of the powder probe, carboncoated Al foil was implemented as a current collector for LMO cathodes. The carbon coating aims to substitute the dielectric Al2O3 thin film that always occurs on the surface of Al foil. Functionalized current collectors were compared to pristine material by using electrical measurements and current rate (C rate) tests in a half-cell configuration.

Contact resistance compensation As a preparatory investigation before the determination of electrical conductivity, data were acquired in a test series as the contact force was increased successively. By doing so, the resistance approximates to a constant limiting value (red curves, see Figures 9 and 10). This number can be regarded as the genuine electrode resistance because parasitical contact resistances are minimized or even completely eliminated. Moreover, this value can be subtracted from every single overall resistance at each applied force. As a result of this compensation, it is possible to quantify contact resistances. This in turn enables us to identify an appropriate contact

Figure 9. Electrical resistances of a LMO cathode sheet. Blade-coated on bare Al foil. Powder probe with applied measuring forces of 2–10 N.

Energy Technol. 2016, 4, 1550 – 1557

Figure 10. Electrical resistances of a LMO cathode sheet. Blade-coated on Al foil with carbon thin film functionalization. Powder probe with applied measuring forces of 2–10 N.

force to be used and warrants the suitability of the method. The measuring results for both types of LMO cathodes are depicted in Figures 9 and 10. A comparison of the cathodes proves that a coating of carbon on the Al foil enhances electron transfer. The conductivity increases by approximately a factor of 7. Contact resistances diminish with increasing force, and in the case of the functionalized current collector, they become inferior to electrode resistances above approximately 6 N. Apparently the two electrodes possess different parasitical contact resistances (2 W, Figure 9 vs. 0.45 W, Figure 10, both at 2 N). This seems surprising as an identical LMO slurry was cast in a homogeneous manner on both the pristine and carbon-coated Al foil. One possible explanation might be a distinct electrode surface structure or a different surface roughness caused by the underlying carbon coating. As such, the carbon functional film influences the wetting and drying behavior of the LMO slurry explicitly. Further investigation will have to be performed to reveal the influence of the carbon coating on the electrode morphology.

Half-cell tests The benefit of improved electron transfer in terms of available battery capacity is evident from the results of the halfcell tests (Figure 11). LMO cathode sheets with carboncoated collectors (Al/C, red symbols) clearly outperform cells with a pristine collector foil at high discharge rates (AlRef., black). In addition to the initial conductivity enhancement achieved by carbon coating, postmortem analysis of the cycled cathodes showed an even higher resistance disparity (AlRef.: 10–16 W vs. Al/C: < 1 W, see legend insets). The reason for this is Al surface corrosion and deterioration. Explicit corrosion phenomena such as staining and efflorescence were observed on the surface of uncovered pristine current collectors. In contrast, carbon-coated collectors hardly exhibited any defects. Analysis was performed by using light microscopy.


1555

ly, explicit discrimination is possible if the powder probe is used. Carbon-coated current collector Powder probe measurements are nondestructive. To emphasize this statement we performed an investigation on an Al foil that had been coated previously with carbon films of various thicknesses, and the outcome can be seen in Figure 13. Both left bars in each series represent results from the direct contact with the copper backplate. In the case of the stamp and for Al/C20 nm and Al/C40 nm (red), the values are almost as low as the lowest base value. Only the 60 nm thick carbon coating is detectable and leads to an increased resistance. Figure 11. Electrochemical results from a total of twelve half-cell tests using LMO-electrode disks (specific capacity: 0.4 mAh cm@2) that were cycled in three-electrode cells with Li foil as a counter and reference electrode. The electrolyte is 1 m LiPF6 in ethylene carbonate/dimethyl ethylene carbonate with a ratio of 3:7.

Advanced powder probe applications It is expected that the conductivity measurements by using the powder probe will be superior to that by using a contact stamp to optimize electrode manufacture, refine the design process, and ascertain electrode quality. Two examples underline this assumption.

Calendered LMO cathodes The results from a comparative study that includes compressed LMO cathodes are shown in Figure 12. Repeated resistance measurements were performed, and the stamp is compared to the powder probe. The standard deviations in the context of stamp measurements are such that differentiation between the two specimens becomes difficult. Converse-

Figure 13. Electrical resistances of Al foil sputter-coated with a carbon film. The carbon film thickness is 20, 40, and 60 nm. Comparison between stamp and powder probe measurements with an applied measuring force of 2 N.

In contrast, powder probe measurements provide successively decreasing resistances for all three foil modifications. This is a conclusive outcome because the carbon coating is applied on thoroughly cleaned and sputter-etched foil and the thicker the carbon the better its barrier effect against dielectric oxide films (reoxidation). Thus, conductivity is improved. This experiment confirms the nondestructive character of the powder probe, whereas effects that arise from the partial destruction of the dielectric oxide film on the surface are visible for the stamp measurement.

Conclusions

Figure 12. Electrical resistances of a LMO cathode on an Al current collector with a 20 nm carbon film. Comparison between stamp and powder probe measurements for a pristine cathode and a compacted cathode. The applied measuring force is 2 N.

Energy Technol. 2016, 4, 1550 – 1557

Conventional two-point probe techniques applied in conductivity measurements on thin battery electrodes suffer naturally from parasitical contact resistances. This complicates comparative studies and makes quantitative analysis (resistivity determination) virtually impossible. Advanced four-point methods deliver explicit information on sheet resistances and composite resistivity but high variances restrict their practicability.


1556

The new powder probe technique is based on an initial minimization of contact resistances and requires a specific, but simple measuring procedure. This method provides viable resistance data for lithium-ion battery cathodes. This enables us to specify and to qualify electrodes. Both aspects are proven to be significant with regard to their electrochemical performance. At present, extended test series are underway to determine how long one probe can continue to be used. So far, a decrease in conductivity caused by the loss or degradation of the powder probe material has not been observed. Further experimental work will be conducted to ascertain the applicability of this test in the case of graphite battery anodes, capacitor electrodes, and thin primer coatings. Investigations will also be performed to analyze electrical anisotropy in electrodes. A fully automated measuring system is also under development.

Experimental Section Powder probe particle preparation A dc sputter device was used to produce Ag-coated stainlesssteel particles. Before sputtering, the system was evacuated and flushed with Ar and finally brought to a pressure of 5.0 X 10@4 Pa. Deposition was performed under an Ar atmosphere with a working pressure of 4.9 X 10@1 Pa. A highly magnetizable stainlesssteel powder (15 g; obtained from Eckart GmbH) with a d50 of 15 mm was stirred mechanically during the coating process aimed at a homogenous deposition on all particles. Power density on the target was 0.75 W cm@2, the target-to-substrate distance was 130 mm, and the coating time was 2 h.

Al current collector Various carbon-coated Al foils (alloy AA 1085, 20 mm thickness) were produced. Foils were initially sputter-etched to remove native oxide structures and potential contaminations. All carbon films were deposited by using a dc sputtering system (base pressure 5.0 X 10@4 Pa) with a target power density of 3.7 W cm@2 and a target-to-substrate distance of 60 mm. The deposition rate of carbon was 3.5 nm min@1 at a working pressure of 4.9 X 10@1 Pa. Film thicknesses were 20, 40, and 60 nm, respectively.

LMO cathode production LMO cathodes for electrical and electrochemical measurements consist of 90 wt % active material LiMn2O4 (TODA Kogyo Corp.) with a nominal capacity of 108 mAh g@1 (first charge capacity at 0.1 C). Further slurry components are 5 wt % carbon black (C 65) and 5 wt % binder (PVDF) dissolved in 17 g DMSO. Doctor-bladed electrode layers yield a specific loading of 0.4 mAh cm@2.

Energy Technol. 2016, 4, 1550 – 1557

Electrochemical tests on a total number of 12 LMO-electrode disks (diameter 18 mm) were performed by using three-electrode test cells (from EL-cell GmbH) with Li foil as counter and reference electrode. The electrolyte was 1 m LiPF6 in ethylene carbonate/dimethyl ethylene carbonate with a ratio of 3:7. Formation was done with a total number of three cycles (charge rate 0.05 C up to 4.3 V, 15 min pause, discharge 0.05 C down to 3 V, 15 min pause). Charging in current rate tests was performed with 0.1 C (until 4.3 V) discharge stepwise (0.1, 0.2, 0.5, 1, 2, 5 C) until 3 V.

Acknowledgements The authors would like to thank the Nds. Ministerium fgr Wissenschaft und Kultur of the State of Lower Saxony for the financial support of this work with Graduiertenkolleg Energiespeicher und Elektromobilit-t Niedersachsen (GEENI). Material analysis by TEM was carried out at the Gottfried Wilhelm Leibniz University Hannover, Laboratory of Nano and Quantum Engineering.

Keywords: battery characterization · contact resistance · energy storage · lithium-ion batteries · scanning probe [1] M. Park, X. Zhang, M. Chung, G. B. Less, A. M. Sastry, J. Power Sources 2010, 195, 7904 – 7929. [2] H.-L. Pan, Y.-S. Hu, H. Li, L.-Q. Chen, Chin. Phys. B 2011, 20, 118202. [3] M. Gaberscek, J. Moskon, B. Erjavec, R. Dominko, J. Jamnik, Electrochem. Solid-State Lett. 2008, 11, A170 – A174. [4] H. Tukamoto, A. R. West, J. Electrochem. Soc. 1997, 144, 3164 – 3168. [5] S. Y. Chung, J. T. Bloking, Y. M. Chiang, Nat. Mater. 2002, 1, 123 – 128. [6] H. Li, H. Zhou, Chem. Commun. 2012, 48, 1201 – 1217. [7] Y. H. Chen, C. W. Wang, G. Liu, X. Y. Song, V. S. Battaglia, A. M. Sastry, J. Electrochem. Soc. 2007, 154, A978 – A986. [8] H.-C. Wu, H.-C. Wu, E. Lee, N.-L. Wu, Electrochem. Commun. 2010, 12, 488 – 491. [9] H.-C. Wu, E. Lee, N.-L. Wu, T. R. Jow, J. Power Sources 2012, 197, 301 – 304. [10] S.-K. Chen, K.-F. Chiu, S.-H. Su, S.-H. Liu, K. H. Hou, H.-J. Leu, C.C. Hsiao, Thin Solid Films 2014, 572, 56 – 60. [11] S. L. Bewlay, K. Konstantinov, G. X. Wang, H. K. Liu, Mater. Lett. 2004, 58, 1788. [12] C.-W. Wang, A. M. Sastry, K. A. Striebel, K. Zaghib, J. Electrochem. Soc. 2005, 152, A1001 – A1010. [13] M. Ender, A. Weber, E. Ivers-Tiff8e, Electrochem. Commun. 2013, 34, 130 – 133. [14] N. Mainusch, T. Siedenburg, T. Christ, E. Flade, J. Paulus, W. Viçl, pending patent: DE 10 2015 212 565.3, 2015. [15] R. Holm, Electric Contacts: Theory and Application, Springer, Berlin, Heidelberg, New York, 2000. Received: March 4, 2016 Revised: April 19, 2016 Published online on August 25, 2016


1557