The qualification file holds two certificates for the same natural hydroxyapatite. One reports a crystal size near 60 nanometres. The other, from an equally reputable lab, reports over a micron. Neither lab made an error, and both worked from the same X-ray diffraction data. The difference is a line of arithmetic.<\/p>\n
A 2020 study in Nanomaterials<\/em> ran the comparison explicitly on natural hydroxyapatite from three sources. Using the classic Scherrer equation across all available peaks, \u201cthe sizes of crystallites obtained by the XRD patterns in each method for hydroxyapatite from cow, pig, and chicken were 1371, 457, and 196 nm.\u201d The same patterns, run through the Halder\u2013Wagner method, returned \u201c4 nm for all samples.\u201d And a corrected Scherrer model \u2014 a straight line forced through the origin \u2014 \u201cgave 60, 60, and 53 nm, which shows much improvement.\u201d<\/p>\n From 4 nm to 1,371 nm, on identical data. The authors are blunt about the worst offender: the all-peaks Scherrer fit \u201cled to unreasonable values.\u201d Their corrected model, by contrast, landed within a few nanometres of an independent BET surface-area cross-check (56, 52 and 49 nm) \u2014 which is what validation looks like.<\/p>\n The honest answer: none of them, without the method attached. A \u201ccrystallite size\u201d figure is not a property of the powder alone; it is a property of the powder plus<\/em> a peak set, a fitting model, and a set of assumptions about strain. Quote the figure without those, and a buyer comparing two suppliers may be comparing arithmetic, not materials.<\/p>\n There is a second trap stacked on the first. Crystallite size is not particle size. As the paper puts it, \u201ccalculation of nanoparticle size extracted by XRD patterns is not possible because a particle has several nanoscale or microscale crystals.\u201d A datasheet that slides from an XRD crystallite number to a \u201cparticle size\u201d claim has changed subject mid-sentence \u2014 particle size belongs to laser diffraction, DLS or microscopy, each with its own disagreements.<\/p>\n None of this is exotic knowledge inside a good analytical lab \u2014 the methods are decades old and their disagreements are documented. The problem is what survives the journey from instrument to certificate: a single number, shorn of its peak set and fitting model, rounded into false precision. By the time it reaches a purchasing spreadsheet it looks as objective as a weight. Two such numbers from two suppliers get compared as if they measured the same thing, and a sourcing decision quietly inherits an assumption nobody made on purpose.<\/p>\nWhich number belongs on a certificate?<\/h2>\n
The question that sorts suppliers<\/h2>\n