Solvent‐Mediated Control of the Electrochemical Discharge Products of Non‐Aqueous Sodium–Oxygen Electrochemistry

Abstract The reduction of dioxygen in the presence of sodium cations can be tuned to give either sodium superoxide or sodium peroxide discharge products at the electrode surface. Control of the mechanistic direction of these processes may enhance the ability to tailor the energy density of sodium–oxygen batteries (NaO2: 1071 Wh kg−1 and Na2O2: 1505 Wh kg−1). Through spectroelectrochemical analysis of a range of non‐aqueous solvents, we describe the dependence of these processes on the electrolyte solvent and subsequent interactions formed between Na+ and O2 −. The solvents ability to form and remove [Na+‐O2 −]ads based on Gutmann donor number influences the final discharge product and mechanism of the cell. Utilizing surface‐enhanced Raman spectroscopy and electrochemical techniques, we demonstrate an analysis of the response of Na‐O2 cell chemistry with sulfoxide, amide, ether, and nitrile electrolyte solvents.


Contents Methods Section
Voltammetric Analysis Tables   Table S1 -

Methods Section
Solvent purification: All solvents were purified by distillation over CaH2. The distillate of which was then dried over freshly activated molecular sieves (4 Å) reducing the water content to a value of ≤ 5ppm water. This was determined using a coulometric Karl Fischer titrator (Mettler-Toledo).
Tetraethylammonium trifluoromethanesulfonate (TEAOTf) (≥ 99.0 %, Aldrich) and sodium trifluoromethanesulfonate (NaOTf) were dried under vacuum at 120 o C for 16 hours before use. It has previously been noted that the purity of commercial NaOTf contains impurities such as NaOH and hydrates. However, characterisation by FTIR (Fig S9) of the commercial powder found no such impurity and therefore the salt was without further purification. These were washed in Milli-Q water (18.2 M Ω) and sonicated between slurries and before being dried at 120 o C under vacuum overnight. Before the electrochemical cells were setup, all electrodes were rinsed with dried electrolyte containing ≤ 5 ppm water. A platinum coiled wire was used as a counter electrode and a silver wire as a quasi-reference electrode. The quasi-reference electrode was standardised against an internal ferrocene reference which has a potential of + 0.4 V vs. NHE. These were prepared in a similar manner to the electrochemical cell accepting the counter electrode that was additionally flame annealed before use electrochemical techniques were carried out at 25 o C.

S5
In situ SERS measurements were setup similarly with a glass, multi-necked gas-tight cell fitted with a sapphire window. A millimetre behind this window an electrochemically roughened Au working electrode was placed. The Au working electrode was roughened using an oxidation/reduction cycle (ORC) described previously 1 (Fig S10) shows the effect of applying the ORC (oxidation-reduction cycle) treatment to a Au electrode with pyridine drop cast on the electrode. Beyond 25 cycles a reduction in signal intensity is observed. Spectra were recorded using a 50x objective on a Raman spectrometer (Renishaw In via) with a 633 nm laser (2 mW) calibrated against a silicon wafer. 23 Na NMR studies were carried out within an in house sealed NMR tube on a Bruker Ultrashield 400 MHz spectrometer. The setup contained a 2M NaCl standard solution in a sealed capillary tube which was dried overnight before use and put in the glovebox a week before using. Along with this 0.2M NaClO4 in the corresponding solvent was added to the NMR tube for testing. Both the spectrum from 23 Na and 1 H NMR was taken as a control measure and the shift standardised to that of the NaCl peak within the spectra.
Fourier Transformed Infrared (FTIR) spectroscopy was carried out upon a Nicolet iS50 FT-IR.
The transmission spectrum was taken using a pelletised mix of caesium bromide and the compound in question.
Voltammetric peak to peak separation offers little in terms of reversibility trend between electrolytes (Table S1), however there is a clear visual decrease in the ratio of charge and peak current as you switch the non-aqueous solvent with the trend of reversibility DMSO > DMA > DEGDME > MeCN (see S6   Tables S1 and S2 on Au working electrode and Tables S3 and S4 using GC working electrode). As a comparative electrolyte property Gutmann donor number (G.D.N. Fig. S11) is useful to ascertain the differences between the chemistries of these electrolytes. Although this is an important consideration other factors may be pivotal in the underlying fundamentals of this mechanism. This may include the solubility and diffusion of NaO2 in these solvents as well as the viscosity of solvent. 2  The observable difference between the two surfaces is a lesser change in the Qa/Qc ratio of the process from 75% to 71%. Another difference is the formation of a secondary peak in the reduction response on Au what is broader if not present on GC surfaces. This is similar to the Li-O2 voltammetric response, whereby there is the formation of LiO2 in the initial voltammetric peak and subsequent formation of Li2O2 in secondary peak. Here due to the absence of any signals for Na2O2 detailed in out SERS analysis this secondary peak is feasibly caused the formation of NaO2 precipitated on the surface.
The same process is occurring to a lesser extent on the GC electrode, but with a lesser tendency for adsorption on the surface of the discharge products of this reaction this feature is broad and diminished.
The surface dependent kinetics of dioxygen redox chemistry perhaps hinders the reaction, at this scan rate, affecting the charge ratio of the peak/peak current ratio more greatly on Au electrode surfaces. The greater propensity for surface adsorption of reactants and electrolyte components at this interface is a secondary influence on the calculated parameters.
The proposed explanation is that the Lewis acidity of DMSO solvated NaO2 induces an ion pair interaction between Na + and O2resembling the quasi reversible electrochemical response of [TEA + --O2 -] ion pairs within the electrolytes (0.1 M Na OTf: 0.18 V TEA OTf: 0.17 V peak separations). DMA S7 based electrolytes with Na + observe a decreased current density on discharge and a positively shifted doublet upon charge (2.64 V & and 2.81 V). The reduction peak decreasing in current density by a factor of 2 from -2.7 mA cm -2 to -1.38 mA cm -2 . The unchanging shift of which concludes that the process, at least on reduction, is most likely a similar [Na + --O2 -] interaction as in DMSO. Due to the slight decrease in Lewis acidity, and subsequent decrease in the favourability of the interaction between the Na + and O2 -, it suggests a lesser solubility of NaO2 and therefore enhanced tendency for precipitation on the surface. This explains the decreased current density but non-shift of the oxidation peak current. The supporting SERS conclusions from this data suggest a lone discharge product of NaO2, which is not supported by the electrochemistry's complex charging characteristics, with peaks arising at 2.64 V and 2.81 V. More investigation is needed to ascertain the mechanism on charge however at present it is presumed these features are due to NaO2 and [Na + --O2 -] species..
Upon discharge as solvent donor number decreases there is a further decrease in peak current density observed. For MeCN this is a decrease of 84% which shows almost total passivation of the electrode surface upon reduction. Similar characteristics for DEGDME are seen when comparing Na OTf electrolytes with TEA OTf electrolytes. However, the reduction in peak current density is only 36% on Au surface displaying the gradual increase in surface passivation upon decreasing solubility or solvent stabilisation of NaO2. However this ratio is masking a reduced peak current for oxygen reduction in TEA OTf electrolytes. The passivation is occurring within this electrolyte may be of similar behaviour to that discussed by Johnson et al 3 in there study of Li-O2 battery electrolytes. Here they discuss competing surface and solution based mechanisms within DEGDME through the limited ability of the solvent to stabilise and solvate LiO2 removal from the surface. This subsequently increases discharge capacity before passivation occurs through surface formed and solvent deposited Li2O2.
If applied to this study with sodium electrolytes interestingly the surface passivation of GC is much more enhanced in this electrolyte observing 72% reduction in peak current comparing Na OTf and TEA OTf electrolytes within this media. Charging characteristics observe little response compared to high S8 donor number solvents, but a similar positive peak shift in oxidative processes with a number of peaks observed. It is speculated these peaks may arise from oxidation of NaO2 and Na2O2, but further experiments are needed to confirm the identity of these species.     Figure S1.  Figure S11. 23 Na NMR correlated to Gutmann donor number for various non-aqueous solvents and water in 0.2 M NaClO4 internally referenced against 2 M NaCl