Lithium-ion batteries (LIBs) are widely used due to their high capacity, high power, and lightweight characteristics. However, the increasing number of battery fires has become a significant concern as LIB electrolytes contain flammable organic solvents. This study investigated the effect of electrolyte component mixing on the limiting oxygen concentration (LOC) of diffusion flames to understand the combustion characteristics of LIB electrolytes. Propylene carbonate (PC) was mixed with diethyl carbonate (DEC), vinylene carbonate (VC), and fluoroethylene carbonate (FEC) as additives. Laminar diffusion flames were formed using these fuel mixtures, and the LOC was measured using the wick method under multiple ambient air flow velocity conditions (5 and 10 cm/s). The LOC of PC flames was 17-18 % regardless of flow velocity. The LOC of the PC-DEC mixture was 17 %, which was lower than that of pure PC flames. When VC was added to the PC-DEC mixture, the LOC decreased with increasing VC concentration. The LOC of the PC-DEC-5 wt%VC mixture was 16 %. Furthermore, fuel consumption increased and flame height became higher compared to flames without VC addition. When FEC was added to PC, the LOC increased with increasing FEC concentration. The LOC of the PC-3 wt%FEC mixture was 18 %. The addition of FEC to PC decreased fuel consumption, and the flame height was lower than that of pure PC.

In order to further investigate the discrepancy between the numerical and experimental outcomes of ion current trends of stagnated ammonia-air flames, the measurement technique of the ion current was altered from the flame being scanned by a single vertically moving cathode probe, to both the anode and the cathode moving in tandem horizontally through the flame. As with the vertically scanning cathode method, horizontal scanning results show that, as the equivalence ratio increases from 0.89 to 1.19, the peak value of ion current generally decreases as progress is made into the rich zone. The horizontal scan technique results in a peak ion current at φ = 1.01, whereas the vertical scanning method gives a peak at φ = 0.95. Elementary reaction calculations performed using an expanded ionic mechanism and the burner-stabilized stagnation flame model, predicts the dominant positive ion NO⁺ to peak at φ = 0.98. Furthermore, the new expanded mechanism predicts the existence of high concentrations of OH^- ions in the burned gasses peaking at φ = 0.95. These negative ions could possibly explain the variance in equivalence ratios where the peak ion current occurs with different measurement techniques. This paper attempts to create a model to explain the phenomena based on these findings. It is concluded that in stagnated ammonia-air flames, and possibly in ammonia-air flames in general, the technique used to measure the ion current must be taken into consideration when used as a diagnostic tool.

Laser Induced Thermal Grating Spectroscopy (LITGS) is an anticipated technique for the quantitative temperature measurement especially for high pressure environment with high accuracy. In general, higher spatial resolution is required for flame measurement. However, the detailed discussion on the spatial resolution in LITGS has not been sufficient. To understand the spatial resolution in LITGS, quantitative temperature measurements to non-reacting jet with two different configurations, i.e., (a) non-reacting jet with quasi-one-dimensional temperature gradient and (b) non-reacting jet from a micro nozzle, were conducted. Non-reacting acetone/air premixture was employed for the measurement object. For the experiment (a), the direction of the probe volume in LITGS was aligned in the perpendicular and the parallel directions to the temperature gradient, i.e., the former has no temperature gradient in the probe volume, but the latter has temperature gradient in it. In addition, the size of probe volume was adjusted by the adjustment of crossing angle and the beam separation in the pump beams of LITGS. As a result, the measured temperature using LITGS showed good agreement with the temperature measured by thermocouple when there was no temperature gradient in the probe volume. On the other hand, the difference of measured temperature was observed when there was temperature gradient in the probe volume. In the experiment (b), temperature could be measured even though the size of the measurement object was smaller than that of the probe volume although the signal intensity was decreased. However, the signal-to-noise ratio decreased, and the uncertainty of the measured temperature increased compared to the case that the scale of measurement target was sufficiently large compared to the probe volume.