Water was purified with an Elix UV 3 system (Millipore Co., Billerica, MA). All reagents used were commercially available and of analytical grade. Bis(2,4,6-trichlorophenyl) oxalate (TCPO), Eosin Y, hydrogen peroxide (H₂O₂), acetonitrile, ethyl acetate, and twenty types of α-amino acids were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan).

The CL profiles and CL counts were examined with a batch-type detection system (CLD-100FC; Tohoku Electronic Industrial Co. Ltd., Sendai, Japan). Here, the CL meant the total CL counts obtained during 480 seconds of measurement for CL. We used four kinds of solution as follows: (A) water only, (B) water-acetonitrile mixed solution (1:1, volume ratio), (C) the water-acetonitrile-ethyl acetate ternary mixed solution (15:3:2, volume ratio; water-rich), and (D) the water-acetonitrile-ethyl acetate mixed ternary solution (3:8:4, volume ratio; organic solvent-rich). In this study, the water was prepared as 1.0 mM carbonate buffer, pH 9.0. The component ratios of the solutions are plotted in the phase diagram of the ternary mixed solution shown in Figure 1.

The preparation of CL regent solution and the mixing procedure of the reagent solution and the analyte solution containing fluorescent compounds, such as eosin Y, were carried out as a reference of our previous paper [18], as follows:

the CL reagent solution (2.0 mM TCPO-200 mM H₂O₂, 0.05 mL; acetonitrile was used as a solvent) was added to each of the above solutions (2.5 mL) containing Eosin Y in the petri dish, through a micro-syringe with constant stirring to obtain the CL profile. The CL was measured through a detection window of 3.1 cm². Similarly, we examined CL profiles by using the solutions (A) and (D) containing α-amino acids instead of Eosin Y, and measured the CL through a detection window of 15.9 cm².

The phase diagram of water-acetonitrile-ethyl acetate is shown in Figure 1 together with the component ratios of the solutions (A)-(D). The dotted line indicates the boundary or solubility curve between homogeneous and heterogeneous solution. The CL profiles in the presence of Eosin Y were examined by using the solutions (A)-(D), by the CL measurement system. Figure 2 shows the CL profiles of 1.0 mM Eosin Y with the solutions (A)-(D). As can be seen in the figure, the CL increased in the following order: (A), (B), (C), and (D). We also examined the CL counts of various concentrations of Eosin Y with the solutions (A)-(D) as shown in Figure 3. The CL counts increased in the order of (A), (B), (C), and (D). However, CL counts with (A) and (B) were very small and showed little change at the examined concentrations of Eosin Y. From the data obtained in Figure 2 and Figure 3, it was confirmed that the ternary mixed solutions of (C) and (D), especially solution (D), indicated a large response for Eosin Y based on peroxyoxalate CL. The tendency of the ternary mixed solutions, especially organic solvent-rich solution, to show large CL based on the peroxyoxalate

reaction system was similar to that reported in our previous paper [12], where we examined the influence of the ternary mixed solution on a luminal CL reaction system. The effect of the ternary mixed solution on peroxyoxalate CL is discussed in the next section.

We examined the CL profiles by using various compositions of the ternary mixed solution, the water rich and organic solvent-rich, as shown in the phase diagram of Figure 4. The compositions of the water-rich mixed solutions in Figure 4 were (C) (the water-acetonitrile-ethyl acetate volume ratio, 15:3:2), ① (15:3.2:1.8 volume ratio), ② (15:3.4:1.6 volume ratio), and ③ (15:3.6:1.4 volume ratio). The composition of the organic solvent-rich solutions in Figure 4 were (D) (water-acetonitrile-ethyl acetate volume ratio, 3:8:4), ①’ (3.8:11.8:5.2 volume ratio), ②’ (3.6: 11.4:5.0 volume ratio), and ③’ (3.4:11.8:4.8 volume ratio). The CL profiles of 1.0 mM Eosin Y by using the water-rich mixed solutions of (C) and ①-③ as well as the organic solvent-rich mixed solutions of (D) and ①’-③’ are shown in Figure 5(a) and Figure 5(b), respectively. The CL with the water-rich solutions decreased in this order: (C), ①, ②, and ③, and the CL with the organic solvent-rich solutions decreased in the order: (D), ①’, ②’, and ③’, as shown in Figure 5. The CL gradually decreased as the compositions of the ternary mixed solution moved away from the boundary curve.

There was little information for CL in the ternary water-hydrophilic/hydrophobic organic solvent mixed solution, and the reason why the ternary mixed solution enhanced CL is not yet clear. The ternary homogeneous solution near the boundary curve may cause formation of small clusters (micro-clusters) in the

solvent solution. The interface of the clusters, with features of hydrophilic and hydrophobic interfaces, may be effective as a CL reaction domain. The organic solvent-rich solution of (D) provided a larger CL than the water-rich solution of (C). Although the reason is not yet clear, the organic solvent-rich solution may inhibit the hydrolysis reaction of TCPO. Similar CL behavior based on the luminol reaction was reported and discussed in our previous paper [12].

The CL profiles of Eosin Y and phenylalanine (1 mM each) were examined by using the organic solvent-rich ternary solution of (D). The obtained data are shown in Figure 6. Although the CL profile of phenylalanine was examined with a detection window about 5 times larger than that of Eosin Y, we could observe a sufficient CL signal even for phenylalanine as an α-amino acid by using the specific mixed solution.

The CL counts of twenty types of α-amino acids were examined by using the organic solvent-rich solution of (D), together with solution (A) as a reference. As alanine was not dissolved with solution (A) and (D), we could not examine it.

The obtained data are shown in Figure 7, where the CL counts are the average values of 3 - 5 measurements. The relative standard deviations (n = 5) of phenylalanine, tryptophan, and tyrosine for their CL counts were 1.32%, 1.20%, and 1.91%, respectively. As can be seen in the figure, solution (D) provided larger CL counts for all α-amino acids than solution (A), in particular for phenylalanine, tryptophan, and tyrosine, the fluorescent α-amino acids. An enhanced effect for peroxyoxalate CL could be observed in the presence of fluorescence α-amino acids, similar to co-existing Eosin Y. α-Amino acids other than fluorescent α-amino acids also showed little CL, compared to the blank values that signified CL counts obtained in the absence of α-amino acids. As mentioned above, solution (A) was water only and solution (D) was the water-acetonitrile-ethyl acetate ternary solution (3:8:4, volume ratio; organic solvent-rich); however, solution (A) and (D) could not necessarily have proton concentrations by the effects of

co-existing organic solvents. The difference in the proton concentration or the exact pH value between solution (A) and (D) may cause a little difference in the CL counts for sixteen-amino acids except for the three fluorescent ones between them.