Chemical Safety: Dimethyl Sulfoxide Overpressurization Hazard

We would like to report an overpressurization that occurred in our laboratory while conducting a synthesis protocol for a metal organic framework structure (MOF) using the solvothermal method outlined by Shengquian Ma and coworkers (Angew. Chem. Int. Ed. 2008, 47, 4130). In this paper, DMSO (dimethyl sulfoxide) is decomposed to form sulfate in the presence of H2O2 as the bridging ligand between two metal centers.

The described reaction was conducted at 145 °C for 72 hours in a sealed Pyrex tube. This paper inappropriately references Pandey (J. Org. Chem. 2007, 72, 369) as the methodology. In their paper, Pandey and coworkers describe a derivative of the Swern reaction that is conducted at -78 °C due to the exothermic nature of this oxidation.

Referencing Ma, a researcher at Los Alamos National Laboratory scaled up this reaction from 1.5 mL to 12 mL DMSO and conducted the reaction in a Teflon-lined Parr vessel (model number 4745 ADB) that was placed on the bottom of a convection oven set to 150 °C and left for the weekend. The following Monday, the researcher saw that the bottom flange lip of the Parr vessel had sheared off as intended when overpressured. This shearing event was energetic enough to dent the bottom of the oven, causing the metal plate of the oven bottom to contact the heating elements, which resulted in an electrical shortage.

We were initially concerned that perchlorate salts formed during the chemical synthesis and may have detonated with the DMSO, but we have found no evidence supporting this hypothesis. Additional experiments were conducted with appropriate engineered controls to protect against a recurrence. Although nearly identical reaction conditions were used, the Parr vessel did not rupture. However, a thermocouple placed on the Parr vessel recorded an 11.4 °C temperature increase lasting approximately one hour after approximately 27 hours of heating.

Although the exact nature of this energetic release is uncertain, we have concluded that several events may have contributed. First and foremost, the exothermic nature of the oxidation of DMSO was not communicated in the publication, and the experiment was conducted close to the boiling point of DMSO (189 °C). Second, an old Parr vessel (manufactured between 1969 and 1973) with an uncertain history of use was used for experimental chemistry; the vessel was not equipped with a burst disk, which is now the preferred design. Finally, the Parr vessel was placed on the bottom plate of the oven, which was later found to have a temperature that may have been on the order of 
35 °C higher than the oven set temperature of 150 °C. We believe these were critical factors that resulted in the energetic release during the experiment.

Although complete oxidation of DMSO to CO₂, SO₂, and H₂O would generate significant volumes of gas that would exceed the pressure rating of the Parr vessel, there is not enough oxygen present for this to have occurred. A thermal ignition related to the exothermic nature of the reaction and the fact that the vessel was likely to have been much hotter than the 150 °C oven set point may have generated significant moles of gas that would have exceeded the pressure capacity of the Parr vessel.

Best practices would include checking critical published references prior to proceeding with an experiment, using appropriately designed experimental vessels, not placing reaction vessels on the bottom of ovens, and not scaling up reactions until proven safe. Researchers who plan to use DMSO for chemical experiments are directed to the review by T. T. Lam and coworkers that outlines many hazards associated with the energetic decomposition of DMSO at temperatures below its boiling point in the pure phase (J. Thermal Anal. Calor. 2006, 85, 25).

Toti E. Larson and Ruqiang Zou
Los Alamos National Laboratory