The thermal decomposition of finely divided Mg(OH)2 does not occur smoothly. Several maxima of the gas evolution rate are observed, when microgram quantities of Mg(OH)2 are heated under ultrahigh vacuum conditions. This phenomenon is attributed to the formation of an overlayer of partially decomposed hydroxide Mg(OH)2-x Ox,2 [[]x,~ with x--+ 2 ([--] : anion-sized neutral vacancy left by H20 molecules in the hydroxide lattice). This defect layer represents a diffusion barrier for further H20 molecules. When a critical thickness is reached, it spontaneously recrystallizes to cubic MgO exposing fresh Mg(OH)2 surface.The differential thermal analysis of magnesium hydroxide, Mg(OH)2, in the usual quantities, namely a few milligrams, yields a smooth endothermic reaction just as one would expect for the simple dehydration reaction Mg(OH)2 ~ MgO + + H20. However, we have repeatedly observed [1,2] that the dehydration of microgram quantities of magnesium hydroxide under ultrahigh vacuum conditions, as measured mass-spectroscopically by the gas evolution rate, does not result in a smooth curve. ExperimentalThe samples consisted of very high purity MgO-powder (total cation impurity content <5 ppm) either superficially hydroxylated or bulk hydroxylated and deuteroxylated to Mg(OH)2 and Mg(OD)2 respectively. The diameter of the individual hexagonal platelets was well below 1 pro, and their thickness in the 0.1 pm range.The samples were heated in a fused-silica capillary tube attached to an oil-free, all-glass ultrahigh vacuum system operating at a base pressure of 10 -~~ mbar. In case A a step mode with 25 ~ intervals and 7 min. isothermal heating periods was used [1 ] whilst in case B a linear heating rate of 2~[2] was used. The gas evolution from the sample was monitored by a memory-free, all platinum Omegatron mass spectrometer [3]. In case A the gases evolved were pumped away rapidly through a wide tube of large conductance, whilst in case B pumping was carried out more slowly through a capillary of known conductance for quantitative evaluation.

The formation of compound (3) from salt ( 1 ) could be explained by a slow protonation and ring opening of the kinetically favored aziridine (2) followed by base catalyzed ring closure to nickel and loss of Ni-N bond. We demonstrated that ( 2 ) , in CHZCIZ/CH~NOZ (1: 1) solution, was quantitatively transformed into (3) (15 minutes, 25"C), confirming the required acid-base catalysis.The complex (3) was stable in inert solvents (T < 1W0C) but was quantitatively demetalated by acids into base ( 4 ) , as its corresponding salt (CF3COOH < 5 min; AcOH: 24 h).Reaction with ethanolic KOH gave tetraphenylporphinatonickel in loo"/, yield. When heated in solution (0.5% in xylene, 18 h under argon), compound (3) rearranged to the known homoporphinatonickel derivatives ( 5 ) and (6) (89 %; ratio 46: 54, as well as 10 % tetraphenylporphinatonickel).This reaction may occur via aziridine ( 2 ) , which was already known to yield (6) on heating at 80°C in solution['! The higher temperature (140°C) required to transform (3) into ( 5 ) and (6) explained the occurrence of both products, whose interconversion by ring inversion is rapid above 11O'C.-Reactions of the type described here could warrant attention for understanding the substrate binding in porphyrin and corrin coenzymes. N -Ethoxycarbonylmethylene-meso-tetraphenylporphinatonickel( 11) ( 3 )Triethylamine (35ml) and Na2S04 (3g) were added to a stirred solution of (I) (735mg) in CHzClz (40ml). After 3 days' stirring at 25°C the solution was filtered, the solvent evaporated and the residue chromatographed on silica gel. Toluene eluted the product which was crystallized from CH2C12/CH30H. Crystals suitable for the X-ray studies were obtained from CHC13/CH,0H. All physica data for (3) are compatible with the structure (IR, visible absorption, MS,

in Kollidin und arbeitete danach chromatographisch auf (Kieselgel; Dichlormethan); es resultierte gelbes (6) vom Fp = 133-134°C (Ausb. 64 %). In der Erwartung, (8) als eines der Endprodukte zu erhalten, wurde (6) in Dichlormethan rnit Diazomethan unter (CuCl)2-Katalyse umgesetzt, das rohe Homologisierungsprodukt darauf bei 0°C mit Triphenylmethylhexafluorophosphat reagieren gelassen und das dabei gebildete Carbeniumionen-Gemisch [(7) und Isomere] schlieBlich rnit Sproz. Kaliumhydroxid hydrolysiert. Bei der Chromatographie (Aluminiumoxid, Woelm neutral; Dichlormethan) des auf diesem Wege erhaltenen oligen Gemischs trennte man eine gelborange Fraktion ab, aus der durch praparative Dunnschichtchromatographie (Kieselgel; Ather/Hexan) (8) gewonnen werden konnte; gelborange Nadeln (aus Diisopropylather) vom Fp= 103-104°C (Ausb. 5 %). Erwarmen von (8) in 2~

Durch neue massenspektroskopische Untersuchungen ist es jetzt gelungen, die D2‐Abspaltung aus deuterioxyliertem MgO zweifelsfrei zu bestätigen.