Storing sodium in carbon: the secret of sodium-ion batteries

Rechargeable lithium-ion batteries are commonly used in a myriad of electronic devices and even vehicles, but lithium itself is a limited and expensive resource. Sodium-ion batteries are a cost-effective alternative to lithium-ion batteries as sodium is so abundant; however, their implementation is severely hindered by the poor electrochemical performance of current electrodes. 'Hard carbon' is one of the most promising materials to use for sodium-ion anodes, but little is known of the way in which sodium is stored within the carbon.

Conventional diffraction techniques give limited information about disordered structures such as hard carbons, so researchers turned to a relatively new beamline at Diamond Light Source: I15-1, the beamline for X-ray Pair Distribution Funciton (XPDF) measurements. Using this beamline, the structure of hard carbon electrodes were studied as sodium was inserted.

XPDF demonstrated that the hard carbon anode was made up of graphene-like fragments, which are curved due to defects in the fragments. Moreover, there were two sodium storage mechanisms at work: at high voltages sodium was inserted between the layers of carbon on the surface of pores close to the defects, but at low voltages clusters of sodium ions were stored within the pores. By understanding how sodium is packaged within hard carbons it might be possible to design carbon materials with advantageous storage features to maximise their performance in batteries.

Lithium-ion batteries have become ubiquitous in our portable electronics in recent years, offering much higher energy densities than other battery chemistries. However, lithium is a mined and limited resource likely to become scarce if lithium-ion batteries are widely implemented in electric vehicles. Sodium, whilst heavier than lithium, is far more abundant. In applications such as large-scale storage for the electricity grid, where low-cost and sustainability are of greater primary importance than size, sodium-ion batteries are a promising complementary technology to lithium-ion batteries1.

The key challenge for developing sodium-ion batteries is the discovery of electrodes which allow sodium ions to shuttle in and out of the structure reversibly over many thousands of cycles. Graphite, which is used as the anode in lithium-ion batteries, cannot store sodium in the same way. Therefore, alternative materials need to be developed for use as anodes. 'Hard carbons'– so called because they remain disordered even at very high temperatures – are amongst the most promising materials; they are cheap and easy to manufacture from sustainable sources2.

Hard carbons contain a range of pores and spaces into which sodium can be inserted, however, relatively little is known about how this happens. This is because the carbons and the sodium inserted are disordered and so are challenging to characterise using conventional crystallographic techniques. In this study, the pair distribution function analysis (PDF) was used. PDF analysis is a total-scattering technique; it uses both Bragg and diffuse scattering in parallel meaning it can also give information about the local atomic ordering and, therefore, is ideal for looking at disordered materials.

EE008
Figure 1: Operando 23Na NMR spectra for an electrochemical cell with sodium metal and hard carbon electrodes, and NaPF6 electrolyte. Strong features corresponding largely to the electrolyte or metal have been truncated for clarity. Spectra are coloured in the region -200 to 1000 ppm according to their intensity. The corresponding electrochemistry is shown on the right-hand side and selected spectra are depicted at the bottom. Reused from reference 4 under CC BY 3.0 - Published by The Royal Society of Chemistry.

Coin-cell batteries containing the samples were discharged (i.e. sodium was inserted) to points of interest on the electrochemical profile (marked by coloured dots in Fig. 1). Samples were removed from the coin cells and loaded into quartz capillaries in an argon atmosphere and data collected on the I15-1 beamline. Background measurements for an empty capillary were used to remove the contribution of the container from the powder pattern. Data were integrated using DAWN3 and standard corrections were applied before the data were Fourier transformed to obtain the G(r) PDF data of the sample.

EE009
Figure 2: (a) Experimental PDF data for pristine hard carbon and a PDF simulated using a turbostratically disordered graphite model (offset, below); (b) difference PDFs of hard carbon anodes at various states of charge. The red line (top) corresponds to a sample discharged to 300 mV, the blue (middle) to 180 mV and the green (bottom) to 5 mV. The red and green asterisk (*) highlights the range over which additional interactions are formed in the high-voltage process, and low-voltage process, respectively. Reused from reference 4 under CC BY 3.0 - Published by The Royal Society of Chemistry.

The PDF is a weighted histogram of atom-atom distances. It contains peaks at distances corresponding to the separation between pairs of atoms in the structure. At short distances (r < 4 Å), this corresponds to bond distances. The area of a peak is dependent on the number of atom pairs at that separation, and the scattering power of those atoms. Figure 2a shows G(r) for the hard carbon under study. It can be seen that the carbon contains graphene-like fragments which are disordered by defects and sheet curvature, resulting in a range of interlayer spaces and pores. This is in contrast to graphite which is ordered and has a single interlayer spacing of around 3.35 Å.

When sodium is inserted into the hard carbon, two processes were observed in the electrochemical profile (marked as 1 and 2 on Fig. 1). Process 1 has a sloping voltage profile at higher voltages (0.8 – 0.1 V); process 2 takes place close to approximately 5 mV. In order to understand the structural changes associated with the different processes, G(r) for the pristine electrode (i.e. with no sodium inserted) and at various stages of sodium insertion were compared. By carefully subtracting the G(r) for the pristine electrode from those electrodes containing sodium, differential PDFs (dPDF; dG(r)) containing only the interactions of sodium ions with the electrode and with other sodium ions could be extracted (Fig. 2b). These highlight additional interactions present on the insertion of sodium. The most important observation in this data is that there are two sodium storage regimes; at high voltages, the dPDF shows only low-intensity features below 7 Å (red * in Fig. 2(b)), implying that the sodium inserted is disordered over the surface of the carbon pores. At low voltages, the dG(r) shows additional peaks out to about 10 Å (green * in Fig. 2b), suggesting clusters with a correlation length of approximately 10 Å are formed.

Complementary operando 23Na solid-state nuclear-magnetic resonance (NMR) spectroscopy experiments, where a battery cell is cycled within the NMR magnet whilst acquiring spectra, were performed at the University of Cambridge. These experiments give nucleus specific information, i.e. about only the sodium local environments (Fig. 1). Two peaks at −10 ppm and 1135 ppm in the spectra for the pristine electrochemical cell correspond to the electrolyte and the sodium metal counter electrode, respectively. During process 1, an additional signal at approximately -40 ppm emerges and grows in intensity. This peak is assigned to sodium inserted into the carbon. During process 2, this peak continues to grow in intensity but now proceeds to shift to higher frequencies, finally reaching 760 ppm at the end of sodium insertion after around 22 hours. This shift implies that the sodium species become increasingly metallic during this period of the electrochemistry.

Taken together with our PDF results, the data shows that the sodium is initially inserted as ions which are stored between appropriately spaced sheets and on the surface of pores close to defects. At low voltages, the sodium-carbon interaction becomes less ionic as the carbon is reduced and, in areas where the pores are sufficiently large, metallic sodium clusters of approximately 10 Å across are formed.

These results suggest that the presence of defects has a significant effect on the electrochemical properties. Defects will affect how sodium is stored during the high-voltage process, but will also help control exactly how the graphite-like, but disordered, carbon sheets stack and, therefore, the pore architecture of the carbon. In turn, the pore architecture will affect the size of the clusters which can form at low voltages. Different methods of manufacture will affect the type and number of defects present in hard carbons. This implies that by optimising the synthesis conditions, ‘designer’ carbons with enhanced electrochemical properties could be obtained.

References:

  1. Yabuuchi, N. et al. Research Development on Sodium-Ion Batteries. Chem. Rev. 114 (23), 11636–11682, doi:10.1021/cr500192f (2014).
  2. Irisarri, E. et al. Hard Carbon Negative Electrode Materials for Sodium-Ion Batteries. Electrochem. Soc. 162, A2476-A2482, doi:10.1149/2.0091514jes (2015).
  3. Filik, J. et al. Processing 2-dimensional X-ray Diffraction and Small-Angle Scattering Data in DAWN 2, J. Appl. Cryst. 50 doi:10.1107/S1600576717004708 (2017).
  4. Stratford, J.M. at al. Mechanistic insights into sodium storage in hard carbon anodes using local structure probes. Chem. Comm. 52, 12430-12433, doi:10.1039/C6CC06990H (2016).

Funding acknowledgement:

The authors thank Diamond Light Source for the provision of beamtime on XPDF (I15-1) and access to beamline I15 (EE8840, EE13681) that contributed to the results presented here. The authors acknowledge funding from EPSRC and the European Commission under grant agreement no. 696656 (Graphene Flagship), the School of the Physical Sciences of the University of Cambridge, Gonville and Caius College, Cambridge, and the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 655444.

Corresponding author:

Dr Phoebe Allan, University of Cambridge and Diamond Light Source, phoebe.allan@diamond.ac.uk

Related publication:

Stratford JM, Allan PK, Pecher O, Chater PA, Grey CP. Mechanistic insights into sodium storage in hard carbon anodes using local structure probes. Chemical Communications 52, 12430, doi: 10.1039/C6CC06990H (2016).

Publication keywords:

Batteries; Energy; Pair distribution function analysis; Diffraction