The hazards associated with the thermal decomposition of chemically incompatible sodium hydride solvent matrices are known, with reports from the 1960s detailing the inherent instability of NaH/dimethyl sulfoxide, NaH/N,N-dimethylformamide, and NaH/N,N-dimethylacetamide mixtures. However, these hazards remain underappreciated and undercommunicated, likely as a consequence of the widespread use of these NaH/solvent matrices in synthetic chemistry. We report herein detailed investigations into the thermal stability of these mixtures and studies of the formation of gaseous products from their thermal decomposition. We expect this contribution to promote awareness of these hazards within the wider scientific community, encourage scientists to identify and pursue safer alternatives, and most importantly, help to prevent incidents associated with these reactive mixtures.

Dimethyl sulfoxide (DMSO) is widely used as a solvent for chemical reactions, as a cosolvent for crop protection formulations, and in medicines for topical administration of drugs. The potential explosion hazards associated with thermal decomposition of DMSO have been well-documented, with early reports dating back to the late 1950s. However, these explosion hazards are still underappreciated and inadequately communicated, as indicated by the fact that numerous severe accidents have occurred on both laboratory and industrial scales over the years. Differential scanning calorimetry studies show that decomposition of pure DMSO is detected at ca. 278 °C, while accelerating rate calorimetry analysis indicates that thermal decomposition of DMSO occurs at temperatures around its boiling point of 189 °C. Studies also show that the presence of certain substances can significantly lower the onset temperature of DMSO decomposition and also potentially increase the severity of the decomposition reaction through autocatalytic behavior. Further analysis of literature information indicates that there is a wide range of substances that exacerbate the thermal decomposition of DMSO, including acids, bases, halides, metals, electrophiles, oxidants, and reductants. This comprehensive review of explosion hazards associated with the thermal decomposition of DMSO and its mixtures will serve as an educational resource to alert researchers about the need to mitigate these hazards and to incentivize research toward its replacement with safer and greener solvents in the broader chemistry community.

Differential scanning calorimetry (DSC) is a common industry tool used in the assessment of thermal stability of materials. Despite widespread use of DSC for thermal stability hazard evaluation, mistakes in testing methodology or interpretations of data are common. To avoid these issues, a standard operating procedure and list of common practices utilized within our Corteva Agriscience Reactive Chemicals (RC) group is presented in this manuscript. Emphasis within our RC program is placed on device calibration and maintenance, selection of the appropriate sample container, and a unique sample preparation methodology. The use of glass capillary and glass ampoule sample containers for DSC testing is outlined, along with the unique flame-sealing procedure utilized to protect the sample. The results of the glass sample containers using di-tert-butyl-peroxide in toluene compared to gold pan are presented showing the effects that sample containers can have on results. Additionally, glass ampoule sample containers, containing ethylene glycol, are used in DSC testing to show their effectiveness for examining a sample’s oxidative nature. A discussion of the issues and shortcomings of the commonly used aluminum pans, for use with organic samples particularly, is also presented. All this DSC testing information provides insight into our group’s ability to work on a diverse array of samples and generate quality data for understanding the thermal stability hazards present within our company.

Accelerating rate calorimetry (ARC) is a common tool used in thermal stability evaluation of hazardous materials to provide self-heat rate and pressure data that are used to model the kinetics of a reaction and even calculate relief vent sizing. The ARC accomplishes a conservative approach by maintaining adiabatic conditions. However, an issue with ARC instrument heating occurs when a reaction with a high self-heat rate is analyzed that results in nonadiabatic conditions and a "thermal lag". Testing was conducted to determine the self-heat rate threshold for the ARC instrument and to better understand telltale features that occur during analysis of samples with high self-heat rates. The results indicate that while the oven cannot maintain adiabatic conditions, the convective heat loss from the sample container to the ARC oven does not have a significant effect on either the measured self-heat rate or the calculated overall heat of reaction. However, the conductive heat loss from the sample container to the connection fitting does, causing the total heat measured in ARC analysis to be lower by 16% for an adiabatic ARC test, 33% for a nonadiabatic ARC test, and ∼20% for a typical ARC test in comparison with DSC measurements. A corrected phi estimation is proposed with a correction factor to take into consideration the conductive heat loss to address this issue. In addition, the thermal lag caused by the ARC temperature measurement does create a significant difference between the measured peak pressure rate and peak self-heat rate occurrences. Analysis of the results showed that a simple linear correction can be applied to high self-heat rate results for liquid samples, while the pressure rate is still increasing, to correct for these peak rate differences. The pressure rate after correction matches the self-heat rate and better agrees with the prediction from kinetic modeling. Such corrections can be applied to solid samples, where there is an additional thermal lag due to resistance to heat transfer in the solid sample itself. Accounting for the thermal lag in reactions with high self-heat rates is important when modeling of the data is used for kinetic parameters and especially for relief vent sizing.

4597The chemical reactivity of the compounds is much greater than the corresponding chlorine analogues. Both N-bromo derivatives react with AgCl, forming the respective silver salts and BrCl, and with chloroform, liberating BrCl. As expected, their reactivity does not approach that of BrOSOzF with covalent chlorides: BrOSOZF reacts rapidly with CFCl, at 22 OC, whereas solutions of the new N-bromo compounds appear to be stable a t 22 O C .Characterization of the new compounds is given in the experimental section, with the IR, N M R , Raman, and mass spectra being similar t o those of t h e respective a n a l o g u e s (CF3S0z)zNCI'1~1Z and (FS02)zNC1.'3.'7*18 Assignment of frequencies to vN-x in this series is not obvious.