Objectives The ionomic profile of rice grain is usually characterized by techniques such as inductively coupled plasma mass spectrometry (ICP-MS) and inductively coupled plasma emission spectrometry (ICP-OES), which require complicated sample preparation and cannot analyze samples non-destructively. On the contrary, X-ray fluorescence spectrometry (XRF) does not require sample preparation and can detect the elemental content of a sample directly. The purpose of this study was to optimize experimental conditions of XRF for the quantitative analysis of elements in rice grains.

Methods The determining conditions of XRF spectra for individual element in rice grains were compared and then optimized, according to the results. The applicability and accuracy of quantitative results of XRF measurement were evaluated by correlating them with ICP-MS results.

Results Using filter-free scanning at a 50 μm step, with a light source of 30 kV, and a current of 600 μA under vacuum, the signals of P, S, Cl, K, Ca, Mn, Fe were distinctly different from the background noise signals. The quantification of XRF signals for Mn and Zn was highly correlated with their concentrations determined by ICP-MS, the determination coefficients of the linear fitting equations were 0.8415 (P<0.001) and 0.7736 (P<0.001), respectively. For Cd detection, the two spectrum signals, La and Ka, were very close to that of K and Mn, respectively. Cd had two spectrum, La and Ka; the La spectrum was very close to that of K, making it unable to differentiate from K, and the Ka spectrum was close to that of Mn, which was pressed completely by Mn signal and could not be isolated.

Conclusions Under the optimized conditions, XRF scanning can be used to quantitatively determine the concentration of multiple elements, except for Cd, in rice grains simultaneously and non-destructively, particularly for Mn and Zn.