Due to the composition of starch it produces few X-rays, which in this case would be indistinguishable from X-rays generated by the cellulose fibres below it. In order to determine the distribution of starch it must be selectively tagged with a heavier element ( in this case iodine). The results of mapping for iodine will parallel the distribution of starch. Unfortunately, life is rarely so simple. As seen in Figure 2, the I L peak is on the flank of the Ca K_1,3 peak. The presence of CaCO₃ (pcc) in most ink jet papers indicates a potential problem. Figure 4 shows the results of mapping for iodine and pcc on a starch treated ink jet paper (sample C9). It is immediately obvious that the two maps are very similar, the difference being the lower intensity of the iodine map. Due to the position of the iodine peak on the flanks of the Ca peak, the iodine signal is swamped by the Ca contribution to background. In order to determine the actual distribution of the iodine this background must be removed. The simplest way to do this is to measure background on either side of the iodine peak, interpolate the background under the iodine peak and subtract it. Figure 5 shows the results of this operation. For comparison Figure 6 shows the same operation carried out on a blank paper (i.e. not iodine treated). Although there is still some signal, it is mostly random, and does not reflect the Ca distribution.

Figure 4 Maps of Ca (a) and I before background subtraction (b) for the same area (sample C9). Note the similarity in distribution.

Figure 5 Background subtraction process for an iodine map (sample C9). a) Raw image. b) and c) upper and lower background image respectively. d) background subtracted image ( raw - (upper bg + lower bg)).

Figure 6 Iodine map of unstained sample C9 (i.e. no iodine). a) raw iodine image. b) background corrected image. Figure 5d shows a background corrected map for the same paper treated with iodine.