Elemental iron powders produced by different processes were separated by nitrogen elutriation into fractions based on particle size. The fractions containing fine particles (7–10 μm) of electrolytic iron, hydrogen (H2)- and carbon monoxide (CO)-reduced iron, and the whole sample of carbonyl iron (92% <11 μm) were blended into freshly milled wheat flour and stored at about 21°C. Storing the enriched flour up to 9 months had no effect on the relative biological value (RBV) of iron for rats from each type of iron supplement. Baking bread with flour containing electrolytic, H2-reduced, and carbonyl iron after storage for 10 months also had no effect on the RBV of the iron powders. There was no change in the RBV of iron for rats from flour enriched with electrolytic iron (200 mg/kg) and ascorbic acid (200 and 1665 mg/kg flour) after storage for 1 week and 3 months. The decreases in RBV due to increase in size of particles (7–10 vs. 20–26 μm) were significant for H2-reduced iron and CO-reduced iron but not for electrolytic iron and carbonyl iron. The RBVs of elemental iron powders for rats were similar to the values obtained with humans, based on iron absorption after an overnight fast.

The relative biological value (RBV) for rats of the iron from 7 elemental iron powders (produced by electrolysis, reduction with hydrogen (H2), carbon monoxide (CO), and desiccated ammonia (NH3), and carbonyl process) were compared with the in vitro solubility of the iron powders in 0.2% HC1 (w/v, ca 0.05/V) for periods of 5 to 90 min. The values obtained for per cent solubility in 10 min were within the fiducial limits of the individual RBVs of 6 iron samples; the exception was one preparation of carbonyl iron whose solubility, but not RBV, was as high as that of electrolytic iron. These data indicate that dissolution rates were good predictors of bioavailability of some types of elemental iron powders. There was good agreement between specific surface areas and particle size distributions of discrete fractions containing fine particles of comparable sizes (7—10 μm) of electrolytic, H2- and CO-reduced iron, and whole preparations of carbonyl iron, but these measurements were not satisfactory criteria for predicting bioavailability. SLOWPOKE neutron activation analysis of the iron powders showed that the carbonyl iron samples contained the lowest concentrations of elements producing short- and long-lived radioisotopes (24Na, 27Mg, 49Ca, 38Cl, 28A1, 66Cu, 56Mn, 52V, 101Mo, 51Cr, 122Sb, and 60Co) with the exception that the Mo content of one preparation was ca 2 times higher than in the other iron samples. Electrolytic iron contained higher levels of Na, Ca, Al, Cu, Mn, V, Sb, and Co than carbonyl iron. H2-, CO-, and NH3-reduced iron samples contained the highest levels of impurities, notably Na, Mg, Ca, CI, Al, Cu, Mn, V, and Cr.