Europium binding to humic substances extracted from deep underground sedimentary groundwater studied by time-resolved laser fluorescence spectroscopy

ABSTRACT

Humic substances (HSs) are ubiquitous in various environments including deep underground and play an important role in the speciation and mobility of radionuclides. The binding of Eu³⁺, a chemical homologue of trivalent actinide ions, to HSs isolated from sedimentary groundwater at −250 m below the surface was studied by time-resolved laser fluorescence spectroscopy combined with parallel factor analysis (PARAFAC) as a function of pH and salt concentration. PARAFAC modeling reveals the presence of multiple factors that correspond to different Eu³⁺ species. These factors resemble those observed for Eu³⁺ binding to HSs from surface environments; however, detailed comparison shows that there are some particularities in Eu³⁺ binding to the deep groundwater HSs. The distribution coefficients (K_d) of Eu³⁺ binding to the HSs calculated from the PARAFAC modeling exhibits a rather strong salt effect. At 0.01 M NaClO₄, the K_d values are relatively large and comparable to those to the surface HSs; they are decreaed at 0.1 M NaClO₄ by more than an order of the magnitude. The K_d values are larger for the humic acid fraction of the deep underground HSs than the fulvic acid fraction over the entire range of pH and salt concentration investigated in this study.

KEYWORDS:

• Humic substances
• deep underground
• europium
• time-resolved laser fluorescence spectroscopy
• parallel factor analysis

1. Introduction

Humic substances (HSs) are natural organic matters resulting from degradation and condensation of animal, plant, and microbial residues, and play an important role in the fate of radionuclides in environments [1–3]. Positively charged radionuclide ions bind to the functional groups of HSs, which changes their aqueous speciation, sorption to minerals and rocks, and migration behaviors [3–5].

HS is a group of organic (macro)molecules with certain physicochemical properties in common [6]. There is an operational classification, based on the solubility to water at different pH; humic acid (HA) is insoluble at acidic pH (pH 1 or 2); fulvic acid (FA) is soluble regardless of pH; humin is an insoluble fraction [1]. HSs are ubiquitous in various environments: surface and shallow underground water, soil, and atmosphere [2,7]. Deep underground environment is not an exception. There are increasing evidences of the presence of HSs in deep underground environments [8–16], although their physicochemical and ion binding properties still remain largely unknown. According to several studies, deep underground HSs are often distinctive from their counterparts in surface environments [12,15,16]. This is no wonder as deep underground HSs, although they may originate from the surface by groundwater recharge or deposited organic matters in rocks [17], experience alternation by prolonged reactions with rocks and microbial activities, or they may even originate from remains of indigenous microorganisms. The authors have demonstrated that HA and FA extracted from groundwater at −250 m below ground level (bgl) at the Horonobe underground research laboratory (URL) of the Japan Atomic Energy Agency (JAEA) have distinctive physicochemical properties and exhibit unique proton (H⁺) and copper (Cu²⁺) binding behaviors [15].

Binding of metal ions to HSs have been an active topic of research over decades [3,18–21]. For divalent metal ions, numerous binding data have been compiled for HSs mostly from surface and shallow underground environments. Mechanistic models have been proposed to describe ion binding to HS over a wide range of conditions [18,22]. Trivalent metal ions such as Fe³⁺ and Al³⁺ are also important for various environmentally relevant processes [3]. Their geochemistry can be greatly modulated by HSs since these ions tend to strongly bind to HSs. In the safety assessment of geological disposal of high-level nuclear wastes, it is of great importance to understand the binding of trivalent and even tetravalent radionuclides to deep underground HSs. Nevertheless, the binding of such metal ions to HSs has been less studied [3,23], compared with divalent metal ions. This is partly because of a lack of reliable analytical techniques such as potentiometric titration with an ion-selective electrode that is extensively used for divalent ions such as Ca²⁺ and Cu²⁺ and that allows one to quantify the binding of the metal ions without separating HSs from the solution [24].

The purpose of this study is to investigate europium (Eu³⁺) binding, a known chemical homologue of trivalent actinide ions such as Cm³⁺ and Am³⁺, to Horonobe HA and FA, which are collectively denoted as the HHSs (Horonobe HSs) hereafter, by time-resolved laser fluorescence spectroscopy (TRLFS). Europium is a trivalent fluorescent metal ion, whose emission spectrum and decay lifetime depend on the local chemical environment around the Eu³⁺ ion [25–27]. TRLFS enables one to directly estimate the binding amount of a fluorescent metal ion to simple ligands, mineral surfaces, and HSs and even differentiate its binding environments [28–32]. This is especially advantageous for the HHSs, as they have relatively small sizes [15], for which classical separation techniques such as ultrafiltration may fail. The binding of Eu³⁺ to the HHSs are estimated from a series of the TRLFS data obtained as a function of pH and ionic strength with the help of multi-mode factor analysis called parallel factor analysis (PARAFAC) [28,30,31,33], and the obtained results are compared with those reported for HSs from surface environments to reveal the particularity of the deep underground HSs for the binding of trivalent actinide ions.

2. Material and methods

2.1. Materials

For the entire experiments, milli-Q grade pure water (18.2 MΩ) and analytical-grade chemicals purchased from Wako Pure Chemical Industries were used, unless otherwise noted.

HA and FA were extracted from groundwater collected at the experiment stages at −250 m bgl of the Horonobe URL [15] and that from a surface borehole (HDB-10) at −500 m bgl. The geology and geochemistry of the site are described elsewhere [34,35]. The −250 m stage is located at the boundary of the Pleiocene Koetoi and the Miocene Wakkanai formations, both of which are composed of siliceous mudstones. The groundwater is weakly alkaline Na⁺/HCO₃⁻ type water with relatively high TOC levels (12 mg/L). The extraction and purification of HA and FA fractions from this groundwater are reported elsewhere [15]. Groundwater from the HDB-10 was pumped by a packer-type groundwater sampling system from −496 to 550 m bgl, where the Wakkanai formation lies, and had circumneutral pH and 22 mg/L of TOC. The same procedure used in the extraction of HSs from the −250 m groundwater was used for this groundwater [36]. By processing 1743 L of the groundwater, 615 and 12 mg of HA and FA were obtained, which corresponds to 1.6% and 0.03% of TOC. Hereafter, the Horonobe HA and FA collected at the two locations will be denoted as HHA250, HFA250, and HFA500. As the yield of HHA500 was too small, it was not used for the further study. The physicochemical properties of HHA250 and HFA250 as well as the binding of H⁺ and Cu²⁺ to them are reported elsewhere [15].

The samples for TRLFS measurements were prepared according to [30]. Briefly, 10 mL of 50 mg/L HS samples with 70 μM Eu³⁺ were prepared in 0.01 or 0.1 M NaClO₄ as a function of pH. The Eu³⁺ concentration was set to achieve a reasonable signal to noise ratio in an acceptable measurement time for single TRLFS measurement, according to the previous research [30], and may not well correspond to the ones expected for the trivalent actinide elements in real disposal situations. Sample preparation was carried out at room temperature (298 K) in an Ar-filled glove box. The pH of the samples was measured by a pH meter (ORION 5-star, Thermo Fisher Scientific) with a ROSS combined electrode (Thermo Fisher Scientific Inc., 1 M NaCl as the inner filling solution) and adjusted from 3 to 7.6 by 0.1 or 0.01M of HClO₄ and NaOH. During the TRLFS measurements, we did not monitor the pH values of the samples. All samples were allowed to reach equilibrium for at least 24 hours before TRLFS and UV/Vis measurements.

2.2. TRLFS measurements

The samples were separated into two portions for TRLFS and UV-Vis absorption measurements. The absorption spectra of the samples were measured by UV/Vis spectrometer (UV-3100, Shimadzu). The obtained spectra were used to correct TRLFS data for inner filter effects [37]. The TRLFS measurements were performed at constant room temperature (294 K), using the same setup reported elsewhere [28,30,33]. The details of the setup are found in the Supplemental material and only briefly mentioned here. The excitation source was 130-fs Ti:sapphaire laser operating at 394 nm laser beam with the power of 0.1 mJ/pulse and the repetition rate of 1 kHz. The fluorescence from a sample was collected at 90°, energy-dispersed by a spectrograph, and measured by a time-dated ICCD camera. Fifty temporal scans were collected with the gate width and step of both 19 μs and initial delay of 10 μs. Two hundred measurements were accumulated for each sample.

2.3. PARAFAC modeling

Single TRLFS data that forms a matrix with wavelength × time coordinates was first corrected for its constant background, inner-filter effect and laser power variation. Finally, a series of the data collected as a function of pH were stored in 10 × 724 × 50 3D array (pH × wavelength × time) for a given HS and ionic strength and processed by PARAFAC, using the N-way toolbox for MATLAB® [38]. Detailed description of PARAFAC is beyond the scope of this article and described in the literatures [39–41] and references therein. The application of PARAFAC for the speciation of Eu³⁺ with surface HSs from various origins is found in [30]. Non-negativity constraint was applied on the coordinates of pH and wavelength and exponential constraint on the temporal coordinate. Estimation of appropriate number of factors in PARAFAC was performed with the aid of sum of squared residual (SSR) and core consistency (CC) [40]. Errors associated with the peak ratios and the fluorescence lifetimes of the extracted factors were performed by cross validation proposed by Riu and Bro [41]. The spectrum of a factor was normalized by the peak intensity of the ⁵D₀ → ⁷F₁ transition at 593 nm and the fluorescence decay curve by the intensity at the first point of the temporal series. The resulting normalization constants were used to calculate the fluorescence concentration profile of the factor. Note that the obtained profiles are approximations of the actual concentration profiles of relevant Eu³⁺ species since the quantum efficiencies are not necessarily same among them [42].

3. Results and discussion

3.1. TRLFS data overview

The fluorescence spectra of Eu³⁺ with the HHSs vary with pH and with time after laser irradiation. Figure 1 shows the normalized spectra of Eu³⁺ at t = 10 μs (the first time step) as a function of pH (a) and the decay profile of the normalized intensity at 592 nm (b) for HHA250 at 0.1 M NaClO₄. Similar results are obtained for the other HSs and salt levels (data not shown). The peak at 615 nm originates from the transition from the ⁵D₀ to ⁷F₂ states, which is an electrical dipole allowed and hypersensitive transition, meaning that its transition probability strongly depends on the symmetry around Eu³⁺ [43]. An increase of the intensity at 615 nm with pH indicates that the symmetry decreases due to displacement of hydration water of Eu³⁺ by HHA250. The decay profiles (Figure 1(b)) show that the lifetime tends to decrease with pH. It is also apparent that the decay profiles do not follow simple mono-exponential decay, which suggests the presence of several species with different local environments around Eu³⁺ at a given pH.

Figure 1. (a) Normalized fluorescence spectra and (b) decay profiles of Eu³⁺ in the presence of HHA250 as a function of pH at 0.1 M NaClO₄. In (b), the profiles only at selected pH (3.0, 4.5, 6.0, and 7.6) are plotted for clarity.

3.2. PARAFAC modeling

PARAFAC modeling was applied for the series TRLFS data measured as a function of pH for HHA250 and HFA250 at 0.01 and 0.1 M NaClO₄. For HFA500, only the TRLFS data at 0.1 M NaClO₄ were measured and processed by PARAFAC due to limited availability of the sample. Determination of the number of factors is the first and an important step in PARAFAC modeling. The SSR and CC were used as indicators for this purpose [40]. Three factors are selected to describe the TRLFS data of HHA250 and HFA250 at 0.01 M NaClO₄ and that of HFA500 at 0.1 M NaClO₄, and two factors are turned out to be necessary for the data of HHA250 and HFA250 at 0.1 M NaClO₄. The SSR and CC at the chosen numbers of factors are given in Table 1. Note that the factor 3 in HFA500 was considered to be a so-called ‘garbage’ factor [30] and is not discussed hereafter, as it lacks meaningful features in its spectrum (Figure 2(e)) and its decay lifetime and concentration profile show no physicochemically sound trends (data not shown). It is likely that this factor results from systematic errors in the measurements. It is also noteworthy to mention that the presence of three factors at 0.01 M NaClO₄ agree with the results of the previous study with surface HSs at the same salt concentration [30].

Table 1. Peak ratios and lifetimes of the factors extracted by the PARAFAC decomposition.

		Peak ratio*			Lifetime (μs)*
HS name	CNaClO4 (M)	Factor 1	Factor 2	Factor 3	Factor 1	Factor 2	Factor 3	SSR**	CC***
HHA250	0.01	0.43 ± 0.01	3.08 ± 0.01	3.69 ± 0.08	91.0 ± 0.3	23.8 ± 0.1	99.5 ± 0.1	0.00001	85.8
	0.1	0.42 ± 0.01	3.2 ± 0.1	–	97 ± 1	52.3 ± 0.2	–	0.00410	89.7
HFA250	0.01	0.39 ± 0.01	1.76 ± 0.02	3.2 ± 0.3	91.0 ± 0.5	49.1 ± 0.5	131.8 ± 1.5	0.00089	77.8
	0.1	0.50 ± 0.02	3.8 ± 0.2	–	102 ± 1	37 ± 1	–	0.00350	89.3
HFA500	0.1	0.36 ± 0.02	2.2 ± 0.1	–	106.7 ± 0.7	54 ± 2	–	0.00300	70.9

*The associated errors are estimated by the cross validation method in [41].

**Sum of squared residuals normalized by the sum of the total intensity.

***Core consistency.

Figure 2. Normalized fluorescence spectra of the factors extracted by PARFAC for (a) HHA250 at 0.01 M NaClO₄, (b) HHA250 at 0.1 M NaClO₄, (c) HFA250 at 0.01 M NaClO₄, (d) HFA250 at 0.1 M NaClO₄, and (e) HFA500 at 0.1 M NaClO₄.

The spectra and relative concentration profiles of the extracted factors are shown in Figures 2 and 3, respectively. The fluorescence decay lifetime obtained by exponential fitting of the decay profile of each factor is presented in Table 1 together with the peak ratio, the ratio of the intensity at 615 nm (the ⁵D₀ → ⁷F₂ transition) to that at 593 nm (the ⁵D₀ → ⁷F₁ transition) [30]. Without energy transfer from excited Eu³⁺ to HS, the lifetime and peak ratio would increase with the progress of binding to HS. The factors are named Factors 1, 2, and 3 according to their similarities. Factor 1 is characterized by the smallest peak ratios and intermediate lifetimes. Its concentration decreases with pH. This factor is thought to be free Eu³⁺. Somewhat shorter lifetimes and larger peak ratios compared with those of Eu³⁺ aquo ion (113 μs and 0.3, respectively) indicate the presence of energy transfer via diffusional collision between Eu³⁺ and HS molecules (i.e. dynamic quenching), as is observed for surface HSs [30]. In the previous study with surface HSs, it is reported that the lifetime of this factor is positively and negatively correlated with the sizes and aromaticities of the surface HSs, respectively, reflecting frequency of collision and ease of energy transfer [30]. Considering relatively small sizes of the HHSs (<1 nm), the lifetimes of this factor would be much shorter. This indicates the existence of hindered energy transfer from excited Eu³⁺ to the HHSs likely due to the small aromaticity of the HHSs, which is in line with the elemental compositions of the HHSs and the carbon distribution by ¹³C-NMR [15,36].

Figure 3. Relative concentration profiles of the factors extracted by PARFAC for (a) HHA250 at 0.01 M NaClO₄, (b) HHA250 at 0.1 M NaClO₄, (c) HFA250 at 0.01 M NaClO₄, (d) HFA250 at 0.1 M NaClO₄, and (e) HFA500 at 0.1 M NaClO₄.

Factor 2 is characterized by the relatively large peak ratio and the shortest lifetime among the three (or two) factors. The concentration of this factor tends to increase with pH, suggesting that this factor corresponds to Eu³⁺ bound to the functional groups of the HHSs. Factor 3 observed only at 0.01 M NaClO₄, on the other hand, has the largest peak ratio and the longest lifetime. These factors look similar to Factors C and B, respectively, in the previous study with the surface HSs [30]. In this previous study Factor C is explained as Eu³⁺ bound to the sites of the surface HSs, being characterized by the large possibility of energy transfer and strong influences of the HS carbon backbone structures on the chemical environment around Eu³⁺; meanwhile, Factor B is thought to be Eu³⁺ bound to the functional groups of the HSs, in which detailed composition of the types of the functional groups strongly affects the local symmetry around Eu³⁺, that is, the peak ratio, and the abundance of this factor. Regardless of the apparent similarity, close comparison can reveal differences in these factors between the two HS groups (surface vs. deep underground). The lifetimes of Factor 2 do not follow the trends observed in the surface HSs, and its peak ratios are too small, considering the larger aliphaticities of the HHSs than those of the surface HSs, as is shown in Figures S2 and S3 in the Supplemental material. On the other hand, the peak ratios of Factor 3 of the HHSs follow the observed correlation between the compositions of the functional group types and the peak ratios of Factor B (Figure 4). It is likely that underlying mechanisms characterizing these factors are similar, although the lifetimes of Factor 3 of the HHSs seem to be too short compared with those of the Factor B of the surface HSs, considering the small ash and nitrogen contents of the former. In addition, disappearance of Factor 3 at the higher salt concentration may also contradict with the similarity of this factor with Factor B in the previous study and the description on Factor B above.

Figure 4. Relationship between the peak ratios of Factor 3 (r₃) of HHA250 and HFA250 or Factor B (r_B) of the surface HSs in [30] at 0.01 M NaClO₄ and the compositions of their functional groups. The ordinate corresponds to the ratios of the amounts of the carboxylic groups of the HSs (q₁) and the sum of the amounts of the phenolic (q₂) and amine-type (q₃) groups determined by acid-base titration. The dashed line represents the regression line between the peak ratios of Factor B and q₁/(q₂ + q₃) established in [30].

Overall, the application of PARAFAC for the TRLFS data of Eu³⁺ binding to the HHSs reveals similar results to those of Eu³⁺ binding to the surface HSs; nevertheless, chemical environments surrounding Eu³⁺ and chemical properties of HSs that govern the fluorescence properties of bound Eu³⁺ are different among the two cases. This may relate to the observation that HHSs are different from surface HSs in various ways, consisting of rather small largely aliphatic organic acids that predominantly complex with metal ions with mono-dentate configuration [15]. Local structures of Eu³⁺ bound to the HHS should be investigated, for instance, by X-ray absorption spectroscopy [15,44,45] and compared with those of Eu³⁺ to surface HSs for further understanding of the difference among them. It is also of interest that Factor 3 only appears at 0.01 M NaClO₄. Increased salt concentration leads to stronger screening of the electrostatic potential around HS particles and thus diminishes the binding of cations. Thus, this result suggests that the screening does not only decrease the overall binding amount of Eu³⁺ to the HHSs, but hinders particular types of the interactions.

3.3. Comparison of the K_d values

Once the underlying factors (i.e. species) are revealed from the TRLFS data by PARAFAC, it is easy to calculate the binding amounts or relevant parameters of Eu³⁺ with the HHSs by the following equation: (1) $${\displaystyle K_{\mathrm{d}}=\frac{1}{r_1{c}_{\mathrm{HS}}}\sum _{i\ge 2}{r}_i,}$$(1) where r_i and c_HA stand for the relative concentrations of the factors (i = 1, 2, and 3) and the concentration of HS in Kg/L. Figure 5 compares the distribution coefficients, K_d, of Eu³⁺ to the HHSs as a function of pH at 0.1 and 0.01 M NaClO₄. The K_d increases with pH, as is the case for cation binding to HSs under the conditions, where the effects of any side reactions such as complexation with ligands and hydrolysis, which also increases in their magnitude with pH, are neglected. The K_d strongly depends on the salt concentration; it is decreased by more than an order of the magnitude for HHA250 and more than two orders of the magnitude for HFA250 when increasing the salt concentration from 0.01 to 0.1 M. As mentioned above, binding amounts of a metals ion to HS decreases with an increase of salt concentration; nevertheless, it seems that this effect is rather strong for Eu³⁺ binding to the HHSs, compared with that to surface HSs [23]. Among the HHSs, HHA250 shows the larger K_d values than HFA250 at the both salt concentrations, and the K_d values of HFA500 are similar to those of the HFA250.

Figure 5. Comparison of the distribution coefficients, K_d, of Eu³⁺ binding to the HHSs as a function of pH at 0.1 and 0.01 M NaClO₄.

The K_d values of Eu³⁺ to HHA and HFA250 at 0.01 M NaClO₄ are compared with those to the surface HSs obtained at the same conditions in Figure 6. The latter is calculated, based on the PARAFAC decomposition of the TRLFS data of Eu³⁺ binding to the surface HSs [30]. The surface HSs in this figure include both FA and HA from terrestrial and aquatic origins. Surprisingly, the K_d values to the HHSs are similar to those to the surface HSs at this salt concentration. This is contrast to the case of Cu²⁺ binding to HHA250, which exhibits smaller Cu²⁺ binding to HHA250 than those to surface HSs [15]. Terashima et al. [36] investigated the binding of Eu³⁺ to HHA250 by fluorescence quenching and reported weak binding of Eu³⁺ at 0.1 M salt concentration. Such discrepancy can be explained by strong salt effects on Eu³⁺ binding to the HHSs. At a low salt concentration, Eu³⁺ binding to the HHSs is relatively large and similar to that to surface HSs due to strong electrostatic interaction. This seems to be largely diminished by screening of the electrostatic potential at a larger salt concentration.

Figure 6. Comparison of the distribution coefficients, K_d, of Eu³⁺ binding to the HHSs and the surface HSs at 50 mg/L HS, 70 μM Eu³⁺, and 0.01 M NaClO₄. The surface HSs are the standard HSs from the international humic substances society (IHSS) and Japan humic substances society (JHSS). PAHA: purified Aldrich HA (commercial), BFA: Biwako FA (lake), SRHA: Suwannee river HA (river), SRFA: Suwannee rive FA (river), IHA: Inogashira HA (soil), IFA: Inogashira FA (soil), NHA: Nordic HA (lake), NFA: Nordic lake FA (lake), DHA: Dando HA (soil), EHA: Eliot HA (soil), and PHA: Pahokee HA (peat). The K_d values of the surface HSs are calculated, based on the results of the PARAFAC modeling in [30] in a similar way to those of the HHSs.

4. Conclusion

Binding of Eu³⁺ to HA and FA isolated from sedimentary deep underground water was studied by TRLFS as a function of pH and salt concentration. Multivariate analysis known as PARAFAC was used to distinguish different Eu³⁺ in the presence of the HSs and quantify its binding amounts. PARAFAC successfully separated the contributions of different factors (i.e. different Eu³⁺ species) from the series of TRLFS data. Factor 1 is considered to be free Eu³⁺, whose lifetime is modulated by collisional quenching with the HS particles. It is likely that smaller aromaticity of the HSs hinders energy transfer from Eu³⁺ to the HSs, resulting in somewhat longer lifetimes of this factor, compared with those of the HSs from surface environments. Factors 2 and 3 correspond to Eu³⁺ bound to the functional groups of the HHSs. Close comparison of the spectra and lifetimes of these factors to those of the corresponding factors in Eu³⁺ binding to surface HSs reveals particularities of the former HSs. Furthermore, it turns out that Factor 3 is the Eu³⁺ species that is strongly affected by a change of salt concentration.

The results of the PARAFAC decomposition were used to calculate the K_d values as function of pH at 0.1 and 0.01 M NaClO₄. The K_d values of the HA fraction are larger than those of the FA faction. The K_d values are decreased by more than an order of the magnitude when changing the salt concentration from 0.01 to 0.1 M. This suggests that the binding of Eu³⁺ to the deep underground HSs are rather sensitive to the salt concentration; at a relatively small salt concentration the K_d values of the deep underground HSs are comparable to those of surface HSs.

Acknowledgments

Takumi Saito thanks Mr Steven Lukman for his assistance in the TRLFS measurements. This research was partly supported by the Ministry of Economy, Trade and Industry (METI), Japan, and ‘Grant-in-Aid for Young Scientists (B)’ (Grant No. 25820446), the Japan Society for the Promotion of Science.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

Ministry of Economy, Trade and Industry (METI), Japan, and ‘Grant-in-Aid for Young Scientists (B)’ [grant number 25820446], the Japan Society for the Promotion of Science.