Volume 87 Issue 28 | p. 2 | Letters
Issue Date: July 13, 2009

Chemical Safety: Mg3N2 Hazard

Department: Letters | Collection: Safety Letters

Both Steven Ley (C&EN, June 8, page 4) and Sheldon Crane (C&EN, April 13, page 2) have now reported that explosions can occur when using magnesium nitride to convert esters to amides following the procedure reported by Ley (Org. Lett. 2008, 10, 3623) or one similar to it.

Because I have worked for many years as part of a group whose purpose was to prevent such reactive chemical events from occurring on the laboratory scale all the way to commercial manufacture, four questions always come to mind when I read about these incidents. How much heat is released in the desired reactions? What potential temperatures and pressures could result? How fast are the reactions (that is, how fast is the heat released)? Can the heat released be removed quickly enough to prevent undesired consequences?

A quick Internet search reveals that magnesium nitride reacts rapidly and very exothermically with water to form magnesium hydroxide and ammonia. Thus, this chemistry presumably occurs in two steps. First, the magnesium nitride reacts with the methanol to form magnesium methoxide and ammonia. (Mg3N2 + 6 CH3OH → 3Mg(OCH3)2 + 2 NH3). Second, the ammonia reacts with the ester to form the amide and an alcohol. Using standard heats of formation, the reaction of magnesium nitride with water has a ΔH = –165 Kcal/mol of magnesium nitride.

Because the same types of bonds are broken and formed in the reaction with methanol, this is a good estimate for the heat of reaction of magnesium nitride with methanol. On the scale of 1.3 g of magnesium nitride used by Crane, that translates to –2,123 cal. Just over half of the methanol used will be consumed in the reaction to make magnesium methoxide. Assuming the reaction mixture has a heat capacity of 1 cal/g-ºC, and ignoring the heat capacity of the container, the vaporization of methanol, and the heat from the ester to amide reaction, one calculates an adiabatic temperature rise of 310 ºC. At just 25 ºC, assuming no solubility of the ammonia in the reaction mixture, and using the ideal gas law, one calculates a resultant pressure of 31.5 atmospheres in the closed 20-mL container. Ammonia will liquefy at about 9.9 atm and 25 ºC, but at 68 ºC, its vapor pressure is 31.2 atm. At 310 ºC, the ammonia will generate 61.6 atm of pressure. If the reaction of the magnesium nitride with methanol is rapid and the reaction of the resultant ammonia with the ester is slow, the potential temperatures and pressures are enough to burst a glass vial.

When exploring new or unfamiliar chemistry, one should always try to estimate the heats of reaction (using model reactions if needed), the maximum possible adiabatic temperature rise, and potential pressures generated using simple assumptions as needed. The results can be quite revealing about possible consequences.

I would suggest that this procedure is always a race between rates of consecutive reactions controlled by the time spent in an ice bath or water bath. A safer way to try running the reaction would be to add the magnesium nitride in small portions, letting the ammonia react between portions, or to dissolve/slurry the ester and magnesium nitride in an inert solvent such as toluene, diglyme, etc., and slowly add the methanol at a rate to control the reaction.

Gary Buske
Midland, Mich.

Chemical & Engineering News
ISSN 0009-2347
Copyright © American Chemical Society