X-Ray Emission Microanalysis of Rock-Forming Minerals. VII. Garnets

X-ray emission microanalyses for magnesium, aluminum, silicon, calcium, titanium, manganese, and iron in twenty-two chemically analyzed and five unanalyzed garnets were made independently by Knowles (K) and Bence (B) and used to check correction formulae of Smith (S) and Bence and Albee (BA). Lower microprobe estimates of titanium in many specimens suggest the presence of Ti-rich impurities: three of the chemically analyzed specimens with serious deviations contain optically anisotropic grains. Chemical zoning complicates interpretation of all analyses, especially for manganese. The KS analyses for aluminum tend to be proportionately 2 percent higher than the chemical analyses if Heinrich's absorption coefficients are used and 1-2 percent lower if Theisen's data are used instead; the KBA and BBA analyses are unbiased. All microprobe analyses for manganese tend to be higher than the chemical values when account is taken of the zoning; the bias is around 2-3 percent. The microprobe analyses for calcium may be slightly lower (0-2 percent) than the chemical values. For other major elements most microprobe data, whatever the standards and whatever the correction formula, agree within 2 percent with the chemical values, with some deviations up to about 5 percent. It seems certain that microprobe analysis of major elements from magnesium to iron can be carried out even for complex silicates with an error not greater than 5 percent of the amount present and often not greater than 2 percent. Further progress will require refinement of correction procedures based on synthetic standards with complex compositions.