Manuscript for JOURNAL OF ALLOYS AND COMPOUNDS Lifetime and Fluorescence Quantum Yield of Uranium(VI) Species in Hydrolyzed Solutions Günther Meinrath1,3, Stefan Lis1, Zdzislaw Stryla1 and Chicgua Noubactep3 RER Consultants Passau Adam-Mickiewicz-University Poznan TU Bergakademie Freiberg Corresponding address: Dr. G. Meinrath RER Consultants Schießstattweg 3a D-94032 Passau Germany Tel.: +49 851 70372 FAX: +49 851 70372 e-mail: meinrath@passau.baynet.de meinra01@fsrz1.rz.uni-passau.de meinrath@geo.tu-freiberg.de 1 Lifetime and Fluorescence Quantum Yield of Uranium(VI) Species in Hydrolyzed Solutions by Günther Meinrath*1,2,5), Stefan Lis3), Zdzislaw Stryla4) and Chicgua Noubactep5) 1* ) RER Consultants, Schießstattweg 3a, 94032 Passau, Germany 2) Institute of Inorganic Chemistry, TU Bergakademie Freiberg, 09596 Freiberg, Germany 3) Department of Rare Earths, Faculty of Chemistry, A.-Mickiewizc-University, 60780 Poznan, Poland 4) Quantum Electronics Laboratory, Institute of Physics, A.-Mickiewicz-University, 60780, Poznan, Poland 5) Institute of Geology, TU Bergakademie Freiberg, 09596 Freiberg, Germany Key Words: uranyl(VI) hydrolysis, time-resolved laser-induced fluorescence spectroscopy, speciation, relative fluorescence yield Abstract : Fluorescence lifetimes and emission intensities of uranium(VI) species UO22+, (UO2)2(OH)22+ and (UO2)3O(OH)3+ (=(UO2)3(OH)5+) are determined after excitation by a dye laser at 420 nm. Despite the widely varying species concentrations in the different solutions, single-exponential decay with emission lifetimes between 2 µs and 15.5 µs has been observed, continuously varying with solution composition. Solution composition has been assessed by interpreting UV-Vis absorption spectra in the range 340 nm to 580 nm by available single component spectra. Factor analysis is used to assess the number of absorbing species in solution. Relative fluorescence quantum yields 12 ± 3 and 10 ± 3 times higher than for UO22+ have been found for the (UO2)2(OH)22+ and (UO2)3O(OH)3+ species. Introduction Uranium in its hexavalent state is the only fluorescent actinide whose luminescence can be studied without often prohibitive precautions against radiation effects. Abundance of ura- nium in natural waters is reported at concentration levels between 0.1 ppb to 40 ppb [1], mostly in its hexavalent redox state. These concentrations are principially sufficient for 2 direct fluorescence detection of uranium(VI) in natural aqueous systems. Despite recent progresses of understanding of some fundamental aspects of U(VI) fluorescence [2-5], it is far away from being a developped field. Chemistry of uranium has been -and probably still is- seen almost exclusively under its strategic aspects, thus neglecting many aspects of general interest. Among those aspects, the fluorescence lifetime of U(VI) in aqueous solutions has to be mentioned. Fluorescence lifetimes depend on temperature [6], presence of quenchers [4,7], speciation [8-10] and ionic strength [11,12]. These dependencies, however, remain to be understood in detail [12]. Fluorescence lifetime of U(VI) e.g. depends also on the concentration of acids, e.g. sulfuric and perchloric acid. The effect of perchloric acid, however, seems to be as pronounced as is the effect of sulfuric acid despite of the fact that perchloric acid certainly will not coordinate to U(VI) to an extent comparable to sulfuric acid [11,12]. A more profund understanding of these points requires availability of a wide range of experimental observations. Concerning U(VI) fluorescence, often only a very limited amount of experimental facts are available and the range of conditions studied is limited, too. In this work, 13 well-characterized hydrolyzed U(VI) solutions at elevated concentra- tions (2.10-4 M to 2.1 . 10-3 M U(VI)) in perchlorate medium were studied by quantitative UV-Vis absorption spectroscopy and lifetime measurements. These data extend previous studies of U(VI) fluorescence at lower concentrations [8,11]. Experimental Uranium(VI) solutions were prepared from a stock solutions. The stock solution was prepared by dissolution of UO3 . 2 H2O in 0.1 M perchloric acid. Changes in pH were made by addition of 0.1 M NaOH stock solution. A pH range from pH 3 to pH 4.5 and an uranium concentration range 2 . 10-4 M to 2.1 . 10-3 M are studied. In this region, U(VI) oligomers form but precipitation of a solid phase is prevented. The solutions were allowed 3 * corresponding address to equilibrate for one week in equilibrium with air. UV-Vis absorption spectra were collected in digital form by double beam UV-Vis spectrometer (Shimadzu UV-2401PC) and processed by programs written by the authors. Noise reduction was achieved by averaging multiple scans. Interpretation of these spectra by factor analysis [13] suggested three species, in agreement with previous results [13,14]. Single component spectra of UO22+, (UO2)2(OH)22+ and (UO2)3O(OH)3+ were taken from previous work [13,15] and used without further processing. Statistical parameters were obtained from Spendley's quadratic approximation approach [16]. Within the validity of Beer's Law, the model function to interpret the observed absorption values is linear and, hence, confidence regions for solution compositions are exact (in the statistical sense). However, high correlation coefficients c of the single component spectra with c((UO2)2(OH)22+, (UO2)3O(OH)3+) as large as -0.951 have been observed. This high correlation is a property inherent in the system under study and can not be resolved. Relevance of the numerical results have been estimated by the consistency of experimentally and analytically obtained total U(VI) concentrations. Lifetimes are measured after excitation with a dye-laser pulse at 420 nm. Laser pulse energy was about 100 µJ. The laser system, the miniaturized dye laser and the detection system are home-made by one of the auhors (Z.S.). Temperature has been (22 ± 1) °C. Results and Discussion The spectra included into the study are given in Fig. 1. The properties of the 13 spectra are investigated by abstract factor analysis (AFA) [13-15,17] with some results summarized in Table 1. AFA is a model-free technique that allows inference into UV-Vis absorption data without a priori assumptions and is recently applied to hydrolyzed U(VI) solutions [13- 15,18,19]. Main advantage of factor analysis techniques is the model-free approach and strong commitment to statistical criteria. In the present context, AFA is used to assess the number of factors (species) contributing significantly to the observed absorption spectra. 4 Included into the first column of Table 1 are the eigenvalues of the data matrix A consisting of 13 column vectors, each holding an experimental spectrum of a one of the 13 U(VI) solutions. Eigenvectors and eigenvalues are extracted from that data matrix by the Singular Value Decomposition. More details on the application of factor analysis techni- ques to U(VI) spectral analysis are given in [13-15] and will not be repeated here. The second column gives the result of a χ2 test together with the critical values in the third column. A factor is considered significant if the χ2 value is larger than its critical value. Hence, the χ2 test suggests three factors. The fourth column holds the so-called real error in the data matrix, RE. The real error accepts a factor if the corresponding RE value is larger than the average noise and background effect in the spectral data. For two factors, RE is 7.7 . 10-4 which is larger than the noise of the spectrometer, while the value of 3 . 10-4 for the third factor would be a rather optimistic estimate for the spectrometer's noise level. The SCREE test, giving the residual percent variance in the data matrix as a function of the number of significant factors is given in Fig. 2. As observed previously, the first factor carries about 99% of the data variance [15] due to the strong overlap of the spectra (cf. Fig. 1) as is also indicated by the high off-diagonal values of the correlation matrix. The SCREE test levels out at three factors, again the third factor is uncertain. A typical example for a resolved UV-Vis spectrum is given in fig. 3. In order to get information on the fluorescence intensity contribution of a species, an estimate of the absorption value of the species in a multicomponent mixture is required [20]: I I cf f, .θ φ ε= ⋅ dλ⋅ ⋅ ⋅ ⋅0 2 302 (1) with If,θ : fluorescence intensity at wavelength θ Io : intensity of the excitation source φf : fluorescence quantum yield 5 ελ : decadic molar absorption at excitation wavelength λ d : path length in cm From the absorption values of the individual species at 420 nm, the relative absorption values can be estimated. Assuming the source intensity to be constant in the time average and the fluorescence quantum yield equal for each species, relative fluorescence intensities for each species can be obtained within the limits of the precision in each step of the total analysis. In principal, the fluorescence contributions of each species may be estimated from the lifetime curves, if the lifetimes of the species are sufficiently different and the lifetime of a species is larger than that of the electronically excited state. This behaviour has been observed previously [8,11] in solutions taken from U(VI) solid-aqueous phase equilibria. In the present case, however, only a single-exponential decay has been observed with lifetimes varying with the absorption ratio of the individual species. This observation can be understood that the electronically excited state has a longer lifetime than the species itself. Hence, an excited U(VI) electron may be transform into a different species before electronic relaxation occurs. In Fig. 4, decay curves for two solutions are given. The analytical data of these solutions are given in Table 2. In all 13 cases, single-exponential decay has been observed. Fig. 5a gives the dependence of observed fluorescence lifetime as a function of the absorption ratio r1 with r1 defined as r1 = ε10/ε22 (2) ε10 : decadic molar absorption of UO22+ at 420 nm ε22 : decadic molar absorption of (UO2)2(OH)22+ at 420 nm. From Fig.5a the overwhelming fluorescence intensity of (UO2)2(OH)22+ compared to UO22+ is evident. The characteristic lifetime of the UO22+ species of about 2 µs increases 6 already on the presence of 10% (UO2)2(OH)22+ species. This is a relative concentration at the edge of detectability by the UV-Vis absorption method. In Fig. 5b, the dependence of observed fluorescence lifetimes is given as a function of absorption ratio r2 defined as r2 = ε22 / ε35 (3) ε35 : decadic molar absorption of (UO2)3O(OH)3+ at 420 nm. The relative fluorescence intensity ratio between (UO2)2(OH)22+ and (UO2)3O(OH)3+ is more balanced and the phenomenological lifetime of the (UO2)2(OH)22+ species of about 7- 9 µs increases at absorption ratios r2 of about 1. These findings are discrepant with previous observations where a double-exponential decay has been observed between the UO22+/(UO2)2(OH)22+ species, (UO2)2(OH)22+/(UO2)3O(OH)3+ species and (UO2)3O(OH)3+/UO2CO3° species [9,11]. From the previous work, single component emission spectra of the respective species are evaluated and validated by spectral resolution of multicomponent emission spectra and the Stokes relationship between absorption and emission spectra of each species [5]. The lifetimes given there for UO22+, (UO2)2(OH)22+ and (UO2)3O(OH)3+ are (0.9±0.3) µs, (2.9±0.4) µs and (7 ± 1) µs [11]. The previous observation is also in agreement with 17O NMR studies [8] for the UO22+/(UO2)2(OH)22+ equilibrium. In the range, where (UO2)2(OH)22+ seems to govern the fluorescence emission, however, phenomenological fluorescence lifetimes in the range 7 - 9 µs have been found, much larger than about 2 µs given previously [9,11,21]. The difference in the excitation wavelength (266 nm [11] compared to 420 nm) is unlikely to be a possible reason for the observed differences. The lifetimes of uranium(VI) in aqueous solutions do not lend itself to an easy interpretation and several aspects are not yet understood [5,12]. It seems, however, that the kinetics of the species formation and dissociation depends on solution composition and/or some 7 catalysis of the exchange rate that has not been realized previously due to the very limited amount of experimental studies in the U(VI)-H2O-CO2 fluorescence system. The time traces of the electronic decay carry the information on the number of excited states generated by the laser pulse. For the present system, pulse duration is about 5 ns, much shorter than the decay rate. Hence, the excitation pulse may be considered as instantaneous and all electronically excited states may be considered to be created at a time to. The relative number of excited states may be estimated from the time traces by extrapolating the fitted decay curve to t = to. By adapting eq. 1 to a three-component system and defining the quantity kxy = (I0 2.303 εxy cxy d) -where xy is either 11, 22 or 35- a quantity Q is obtained from numerical interpretation of the UV-Vis absorption spectra in eq. 4: Q = ∫Iφ ,π = k10 * φf,10 + k22 * φf,22 + k35 * φf,35 (4) Q gives a measure of the relative fluorescence quantum yield of UO22+, (UO2)2(OH)22+ and (UO2)3O(OH)3+ species by collecting the emission over all emission wavelengths. Sub- script π indicates that the photons are collected at the photomultiplier tube under a very limited range of angles, φf,xy represents the relative fluorescence quantum yield and xy is representing the respective absorbing species. Only the relative magnitude of φf,xy is meaningful. The numerical interpretation of the data resulted in relative fluorescence yields φf,22 = 12 ± 3 of (UO2)2(OH)22+ and φf,35 = 10 ± 3, while the fluorescence quantum yield of UO22+ is φf,10, = 1. The uncertainties are obtained from a Jackknife analysis that give a 68% confidence region, corrected for the small sample size by appropriate Student's t. The result shows that UO22+ is a poorly fluorescent ion. Formation of oligomeric hydrolysis species strongly increases the fluorescence quantum yield. In agreement with the interpretation of Figs. 5, small amounts of (UO2)2(OH)22+ species are able to govern the 8 fluorescence emission even if UO22+ is the major solution species. The situation is different for (UO2)2(OH)22+ and (UO2)3O(OH)3+ where the difference in the relative quantum yield is hardly significant and the increase in the fluorescence lifetime occurs roughly at r2 ≈ 1. There is, at present, only one other study reporting relative fluorescence efficiencies [22]. The both studies however can not be compared. The solution compositions in the other study are not assessed by direct spectroscopic speciation but inferred from numerical calculation. In that calculations, formation constants from a data base [23] have been used. It has been shown previously that these formation constants are not in agreement with similar other assessments of U(VI) hydrolytic behaviour and also discrepant with direct speciation by using UV-Vis and laser-induced fluorescence spectroscopy [13-15,18,19,26]. Conclusions The lifetime behaviour of U(VI) species UO22+, (UO2)2(OH)22+ and (UO2)3O(OH)3+ has been investigated by quantitative methods. In disagreement with previous results, concentration-dependent lifetimes have been observed. At the present state of our under- standing of the U(VI) fluorescence decay, no explanation for these discrepancies is avai- lable. The observed lifetimes are λ10 = (2 ± 0.5) µs, λ22 = (8 ± 1) µs and λ35 ≥ 15 µs. From a quantitative interpretation of the absorption spectra of the solutions studied, absorption values of the respective species at the excitation wavelength are obtained. Thus, the relative fluorescence yields of (UO2)2(OH)22+ and (UO2)3O(OH)3+ are determined as φf,22 = 12 ± 3 and φf,35 = 10 ± 3, compared to φf,10 = 1. Here is, however, a note of caution necessary: The relative quantum yields depend strongly on the quantum yield obtained for the UO22+ species to which the relative values φf,22 and φf,35 are normalized. This value is small and has a comparatively high uncertainty. All experimental quantities are affected by a series of experimental uncertainties, starting with the spectral deconvolution of the multicomponent UV-Vis spectra, stability of the laser system, temporal stability of U(VI) 9 solutions, e.g. absence of quenchers including organic material etc. The magnitudes of relative quantum yields are, however, in agreement with other experimental observations, as already mentioned above. Direct speciation of sample solutions by suitable techniques like UV-Vis absorption spectroscopy is a necessary prerequisite for interpreting fluorescence emission spectra on a quantitative level. The data forwarded from present study seems to be internally consistent. Nevertheless, the error margins in the relative fluorescence yields are considerable. These margins are to a large part due to the rather imbalanced properties of the UO22+/(UO2)2(OH)22+/(UO2)3O(OH)3+ system. The both oligomeric species have high molar absorption values compared to UO22+ and show a strong fluorescence contribution at relative low concentrations. References 1 Rogers, J.J.W., Adams, J.A.S.; "Uranium" in: Handbook of Geochemistry, Part II-2, Chpt. 92 BO, (1972) 2 Moriyasu, M., Yokoyama, Y., Ikeda, S.; J. Inorg. Nucl. Chem., 39 (1977), 2199-2203; Matsushima, R., Fujimori, H., Sakuraba, S.; J. Chem. Soc. Faraday Trans., 70 (1974), 1702 3 Moulin, Ch., Reiller, P., Beaucaire, C., Lemordant, D.; Appl. Spect., 47 (1993), 2172- 2174 4 Moulin, Ch., Beaucaire, C., Decambox, P., Mauchien, P.; Anal. Chim. Acta, 238 (1990), 291 5 Meinrath, G., Kato, Y., Kimura, T., Yoshida, Z.; Radiochim. Acta, 82 (1998), 115-120 6 Formosinho, S.J., M.D.Graca, M.; J. Chem. Soc. Faraday Trans., 80 (1984), 1745-1756 7 Matsushima, R.; J. Am. Chem. Soc., 94 (1972), 6010-6016 8 Park, Y.Y., Sakai, Y., Abe, R., Ishii, T., Harada, M., Kojima, T., Tomiyasu, H.; J. Chem. Soc. Faraday Trans., 86 (1990), 55 10 9 Meinrath, G., Kato, Y., Yoshida, Z.; J. Radioanal. Nucl. Chem., 174 (1993), 299-314 10 Moulin, C., Decambox, P., Moulin, V., Decaillon, J.G.; Anal. Chem., 67 (1995), 348 - 353 11 Kato, Y., Meinrath, G., Kimura, T., Yoshida, Z.; Radiochim. Acta, 64 (1994), 107-111 12 Bouby, M., Billard, I., Bonnenfant, A., Klein, G.; Chem. Phys, 240 (1999), 353-370 13 Meinrath, G.; Radiochim. Acta, 77 (1997), 221-234 14 Meinrath, G.; J. Alloy Comp., 275-277 (1998), 777-781 15 Meinrath, G., Schweinberger, M.; Radiochim. Acta, 75 (1996), 205-210 16 Spendley, W.; in: Optimization (R. Fletcher, ed.), Institute of Mathematics and Its Application, Academic Press, London/UK (1969), 259 17 Otto, M.; Chemometrics, Wiley-VCH, Weinheim/F.R.G. (1998) 18 Dai, S., Burleigh, M.C., Simonson, J.M., Mesmer, R.E., Xue, X.-L.; Radiochim. Acta, 81 (1998), 195-199 19 Lubal, P., Havel, J.; Chem. Papers, 51 (1997), 213-220 20 Perkampus, H.-H.; Encyclopedia of Spectroscopy, Wiley-VCH, Weinheim/F.R.G. (1995), 669p. 21 Azenha, M.E.D.G., Burrows, H.D., Formosinho, S.J., M.G.M.Miguel, Daramanyan, A.P., Khudyakov, I.V.; J. Luminescence, 48/49 (1991), 522-526 22 Eliet, V., Bidoglio, G., Omenetto, N., Parma, L., Grenthe, I.; J. Chem. Faraday Trans., 91, (1995), 2275-2285 23 Grenthe, I., Fuger, J., Konings, R., Muller, A.B., Nguyen-Trung, C., Wanner, H.; Chemical Thermodynamics of Uranium, OECD/NEA, Paris/France, (1992) 24 Fuger, J.; Radiochim. Acta, 58/59, (1992), 81-91 25 Fuger, J., Khodakhovskiy, I.L., Sergeyeva, E.J., Medvedev, V.A., Navratil, J.D.; The Chemical Thermodynamics of Actinide Elements and Compounds.-Part 12. The Actinide Inorganic Complexes, IAEA, Wien/Austria, (1992), 13-41 11 26 Meinrath, G.; J. Radioanal. Nucl. Chem., 232 (1998), 179-188 12 Table 1: Results of AFA for the first eight of 13 eigenvalues _________________________________________________________________________ eigenvalue χ2 critical values of χ2 RE 8.2661422 116241 144 3.956 . 10-3 0.0464031 1132 121 7.694 . 10-4 0.0014725 127 100 2.928 . 10-4 0.0000872 10.3 81 2.247 . 10-4 0.0000208 2.5 64 1.314 . 10-4 0.0000059 1.3 49 9.552 . 10-5 0.0000043 0.7 36 8.332 . 10-5 0.0000028 0.43 25 7.253 . 10-5 _________________________________________________________________________ RE : real error in the data matrix 13 Table 2: analytical data of U(VI) solutions given in fig. 4 _________________________________________________________________________ lifetime 7.8 µs 14.8 µs pH 4.27 4.42 [UO22+] 9.9 . 10-4 ± 1.0 . 10-5 3.1 . 10-4 ± 2.31 . 10-5 [(UO2)2(OH)22+] 2.05 . 10-4 ± 1.7 . 10-6 6.2 . 10-5 ± 2.8 . 10-6 [(UO2)3O(OH)3+] 4.8 . 10-6 ± 1.9 . 10-7 2.9 . 10-5 ± 1.0 . 10-6 _________________________________________________________________________ [] : concentration; uncertainty in pH is ±0.03 (68 % confidence limit) Figure Captions: Fig. 1: experimental UV-Vis spectra of hydrolyzed U(VI) solutions at different pH in the range pH 3 to pH 4.5 and uranium concentration range 2 . 10-4 M to 2.1 . 10-3 M. Fig. 2: residual percent variance as a function of factors (SCREE test) Fig. 3: experimental UV-Vis absorption spectrum at pH 4.42 and 5.2 . 10-4 M U(VI). Numerical data is included into Table 2 14 Fig. 4: time traces of fluorescence decay of hydrolyzed U(VI) solutions with composition given in Table 2. Fig. 5a,b: Variation of fluorescence lifetimes as function of the ratio r1 (a) and ratio r2 (b) 15 350 400 450 500 0,00 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 0,09 0,10 1 ] 16 0 1 2 3 4 5 6 7 8 9 10 0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 1,1 1,2 17 0,000 0,005 0,010 0,015 0,020 0,025 0,030 350 400 450 500 18 10 20 30 40 50 0,00 0,05 0,10 0,15 0,20 0,25 : measured : calculated τ = 14.8 ± 0.12 µs figure 4 τ = 7.83 ± 0.02 µs in te ns ity / [re l. un its ] decay time / [µs] 19 -1 0 1 0 2 4 6 8 10 12 14 16 2.2 µs 7.5 µs 15.5 µs figure 5a flu or es ce nc e lif e tim e / [ nm ] log(absorption ratio r1 at 420 nm) 20 -1 0 1 6 8 10 12 14 16 8.2 µs 15.5 µs figure 5ab flu or es ce nc e lif e tim e / [ µs ] log(absorption ratio r2 at 420 nm) 21