Recent advances in materials and biomedical sciences have shown that many interesting phenomena are controlled by trace metals in the system1), and therefore not only the identification/determination of metals but also information on the chemical state is significant. When the quantity of the sample is limited, and the sample is only a liquid drop, however, it becomes extremely difficult to analyze its trace components. In the present study, total reflection sample support has been introduced to improve the detection power of the chemical shifts of absorption edges by X-ray fluorescence detection2).
The experiment was carried out using the grazing incidence X-ray spectrometer3) installed at BL-4A. A mirror-polished silicon wafer was employed as a sample support. The advantage is a very high signal to background ratio which is principally due to the extremely small penetration to the substrate and the availability of a big solid angle of a Si(Li) detector4). The measurement procedure is as follows: (i) Tune the energy at the post edge of the metal of interest, (ii) Align a silicon mirror below the critical angle so that X-rays are totally reflected, (iii) Measure an XRF spectrum (blank evaluation), (iv)Drop a solution by a micro pipette, typically 0.5 - 3micro litter, (v) Quickly scan the incident X-ray energy around the absorption edge, before the solution is dried out. The energy range is typically 30 ~ 50 eV, and the measuring time is 3 - 5 min in total. By reading out the chemical shifts of the absorption edges, one can judge the chemical states.
Figure 1 shows the near edge absorption spectrum obtained by X-ray fluorescence from diluted iron solution (5mM, 3micro litter) of the standard materials. The measurement was done for a wet drop. It is confirmed that the edge shift of FeCl3 (Fe(III)) is larger than that of FeSO4 (Fe(II)). This means that trace Fe(II) and Fe(III) in the small quantity sample can be distinguished. The conspicuous pre-edge structure of K3[Fe(CN)6] and K4[Fe(CN)6] would be feasible as well for identification besides the use of edge shift. It has been found that the practical lower limit is around 0.1mM for micro litter solution in our case. Though the scattering background from a usual support or a cell for a solution sample degrades the sensitivity critically, the present procedure has improved it to almost the same standard as a relatively thin case in bulk analysis2).
The present procedure was applied to donor horse serum (ICN Biomedicals, Inc, USA) centrifuged (10,000 g, 30 min at 4 C). It has been found that the protein-bound iron (in the order of sub mM) is in the mixed state of Fe(II) and Fe(III) as shown in Fig.2, but the free iron (in the order of EmicroM1)) could not be detected because of still insufficient sensitivity. When the material of interest is in a dried state and the volume of a residue is extremely small, the perpendicular detector geometry was found to be effective. The sensitivity in the absolute amount was at least 1000 times better than that of a wet drop in the present experiment. Such a big difference is due to the difference in the ratio of the signal to scattering background from the sample itself, and this suggests the significance of the improvement of the detector system and optimization of the geometry.
The author would like to thank Dr. H.Shintani of the National Institute for Health and Dr. N.Usami of the Photon Factory for their helpful discussion on the preparation of biomedical samples.
1)Clinical Chemistry Special issue:"Trace elements in clinical chemistry",21,(1975)No.4.
2)K.Sakarai, A.Iida and Y.Gohshi,Adv.in X-Ray Anal.32,167(1989).
3)K.Sakarai, A Iida,other article in this issue.
4)Y.Yoneda and T.Horiuchi,Rev.Sci.Instrum.42,1069(1971).
Figure 1 Near edge absorption spectrum of trace iron in a liquid drop obtained by X-ray fluorescence detection at the total reflection condition.
