Languages

XRD for Battery Research

Battery Pouch Cells

Characterization of an NMC pouch cell with the EIGER2 R 500K

Pouch cells have become an industry standard battery design due to their efficient shape
and lightweight construction. In operando measurements allow simultaneous monitoring
of the cathode and anode for cycling effects which effect energy storage performance.
This lab report describes in operando characterization of an NMC pouch cell using
XRD with the D8 ADVANCE equipped with Mo radiation and the EIGER2 R 500K detector.

The pouch cell used in this experiment was made of a single NMC (LiNixMnyCozO2) layer (67 μm) coated on Al foil (15 μm), a separator (40 μm), and graphite (82 μm) coated on Cu foil (9 μm). The electrodes were immersed in a LiPF6 electrolyte solution, and packed in a polymer-Al composite bag.
The pouch cell was positioned in the center of the diffractometer, held by two clamps. The total thickness of the pouch cell was about 920 μm. Mo radiation rather than the more common Cu radiation was used for the transmission measurements to reduce the effects of X-ray absorption by the pouch cell. The thick Si sensor of the EIGER2 R 500K is well suited for wavelengths ranging from Cr to Mo, producing high signal while reducing the background by minimizing the effects of charge sharing.

Two charge/discharge cycles were done at C/5 rate (5 hr charge, 5 hr discharge), leading to a total cycling time of 20 hours. The pouch cell was charged at constant current, the charge/discharge profile is shown in figure 1.

Diffraction data were collected during charge/discharge cycling. For this particular experiment, diffraction patterns were collected in transmission geometry from 7-32° 2q. This angular range was covered in a single shot with the EIGER2 R 500K detector. This allows for fast data collection with only 3 min per diffraction pattern, resulting in 400 diffraction patterns collected over the 20 hr experiment. The fine time slicing provides detailed insight on the structural changes happening during the cycling process. The excellent data quality and high intensity of the individual diffraction patterns suggest that data collection time could be reduced below 1 minute.

Figure 2: Iso-intensity plot of the two charge/discharge cycles with DIFFRAC.EVA.

During charging Li+ ions migrate from the positive electrode to the negative electrode, where they intercalate in the graphite layers. This process is reversed upon discharge. The iso-intensity plot in figure 2 displays that the phase composition changes reversibly upon cycling.

Figure 4a shows a qualitative phase analysis on a diffraction pattern taken at 2.7 V (discharged state), figure 4b at 4.3 V (charged state). As expected, LiC6 and other Li/C phases resulting from the Li intercalation are present in the charged state (figure 4a) but absent in the discharged state (figure 4b).

Beyond this qualitative information, more detailed information can be extracted from a Rietveld refinement. A typical in operando experiment will consist of several hundreds or even thousands charge/discharge cycles, hence automated batch mode evaluation is essential to efficiently analyze the large number of diffraction patterns. In this case, all 400 patterns were analyzed in DIFFRAC.TOPAS using batch mode. For a proper Rietveld refinement, the sandwiched layer pouch cell design has to be taken into account including peak intensity corrections, profile and position corrections1.

D8 ADVANCE configuration

Mo radiation (0.71 Å, 17.5 keV)

Focusing Goebel mirror, 1 mm slit, 2.5° primary axial Soller

Pouch cell – 2 charge/discharge cycles at C/5 rate

2.5° panoramic axial Soller
EIGER2 R 500K detector, 2q-optimized mode,
180 mm sample-to-detector distance

Figure 5 shows the NMC and graphite lattice parameters during the two charge/discharge cycles. The intercalation of Li into the layered graphite structure during charge initially leads to a quick expansion of graphite c-lattice parameter and LiCX (X = 24,12,6) were formed and identified as charging continues.

Interestingly, the NMC a- and c-lattice parameters show opposite behavior during cycling. The a-lattice parameter decreases during charge, as intuitively expected since Li transfers from NMC to graphite. The c-lattice parameter shows a strong initial increase during charge, followed by a decrease just before reaching its maximum capacity. This behavior is linked to the layered structure of NMC.

Figure 6 shows that the metal atoms (Ni, Mn and Co) are arranged in layers of MO6 octahedra stacked along the c-axis with Li atoms in-between. During charge, Li leaves the structure leading to strong O–O electrostatic repulsion. As a consequence, the c-lattice parameter expands. The then following reduction of the c-axis at higher charge levels (at > 4V) can possibly be linked to a charge transfer from oxygen to the transition metals, which would reduce again the O-O electrostatic repulsion. At the same time, the transition metal oxidation state increases, reducing the radii of the metal ions and resulting in a stronger M-O attraction. This primarily affects the MO6 layers in the ab-plane and explains the decrease of the a-lattice parameter.

This lab report demonstrates that a wealth of structural information can be obtained from in operando studies on battery materials using a D8 ADVANCE diffractometer equipped with Mo radiation and the EIGER2 R 500K detector. Efficient data collection strategies allow for fine time slicing, providing detailed insight on the structural changes happening during the cycling process.

References:
[1] Rowles, M.R. & Buckley, C.E. (2017), J. Appl. Crys. 50, 240-251

Lithium Batteries using XRD

In Operando Characterization of Lithium Batteries using XRD

Since their commercial introduction beginning of the ‘90ties, small-size lithium-ion batteries have found wide-spread use in consumer electronics due to their high specific energy and energy density. Recently lithium-ion batteries have also started replacing nickel-metal hydride batteries in hybrid and electric vehicles. These automotive applications require large-size batteries with high energy density and superb cycle life time.

 

In particular the cathode material has a major impact on the battery capacity. Therefore improving batteries for automotive and power grid applications is very much focused on researching cathode materials and their characteristics, in particular under in operando conditions.

The battery cell used for the in operando study was developed by the Laboratoire de Réactivité et de Chimie des Solides at the Université de Picardie Jules Verne (Amiens, France), and is exclusively commercialized by Bruker AXS. This user friendly, Swagelock-type battery cell is used by numerous other research groups worldwide and has proven to work efficiently. The basic design is shown in Fig. 1.

The positive electrode, the separator (soaked with electrolyte) and the negative electrode are stacked layer by layer in between a Be window and a stainless steel plunger. Further technical details are described by J.B. Leriche et al. in the Journal of The Electrochemical Society 157 (5) A606-A610 (2010). The battery cell fits straight to the rotating sample holder of the D8 ADVANCE diffractometer. Besides for lithium-based electrode materials, the battery cell can also be used for in operando characterization of sodium-based electrode materials.

For the current experiment, the battery cell was prepared in a glove box using Li+Fe2+PO4 (LFP) as positive electrode material. 15 charge/discharge cycles were run at C/2,5 rate using a potentiostat SP-50 from BioLogic (Claix, France).

Because structural changes occur during cycling it is of utmost importance to collect several diffraction patterns while charging/discharging. The D8 ADVANCE features Dynamic Beam Optimization (DBO), which enables collection of highest quality data over a large angular range in shortest measurement time.

For this particular experiment diffraction patterns were collected from 16°(2q) up to 38°(2q) in only 5 minutes. This time slicing results in about 30 diffraction patterns collected during charge and discharge respectively, about 900 diffraction patterns in total in 75 hours.

Upon charging Li+ ions transfer from the positive electrode to the negative electrode. As a consequence LFP is reduced to the Li-free phase (Fe3+PO4, FP). Upon discharging the opposite reaction occurs. Fig. 2 shows a qualitative phase analysis on a diffraction pattern taken in the course of the charge cycle. Both LFP and FP phases are easily identified.

The iso-intensity plot in Fig. 3a nicely demonstrates that the phase composition changes upon cycling, the waterfall plot in Fig. 3b shows this more in detail for the first charge-discharge cycle.

Fig. 4 shows a zoomed view on the LFP(200) and FP(200) reflections during charge. As expected for this rather slow charge rate the LFP(200) gradually diminishes and the FP(200) gradually increases, indicative of a first-order phase transition.

Beyond this qualitative information, also quantitative and (micro)structural information can be extracted from a Rietveld refinement. This was done in an automated way on all diffraction patterns using DIFFRAC.TOPAS in batch modus. A typical in operando experiment will consist of several hundreds or even thousands charge/discharge cycles, hence automated batch mode evaluation is essential to efficiently analyze the large number of diffraction patterns.

Fig. 5 shows a Rietveld refinement of a typical diffraction pattern. Rietveld refinement enables quantification of the respective crystalline phases, as well as obtaining structural information such as the lattice parameters, and microstructural information such as crystallite size parameters.

The quantitative phase amount as obtained from the Rietveld refinement (Fig. 6) is plotted for the first charge-discharge cycle. The data indicate that upon charging there is a conversion from LFP to FP, which is completly reversible upon discharging. Li+ extraction and insertion also has an effect on the crystal structure of the respective phases and clearly follows the charge/discharge cycle.

Fig. 7 shows the evolution of the c-lattice parameter during cycling. Also the crystallite size changes upon cycling. The effect is rather small for LFP, but the FP crystallite size decreases significantly upon approaching full discharge (Fig. 8). This application note demonstrates that a wealth of information can be obtained from in operando studies on battery materials using a home-lab diffractometer.

Acknowledgement:
The sample for the in operando study was kindly provided by the Laboratoire de Réactivité et de Chimie des Solides (Université de Picardie Jules Verne, Amiens, France).