Time-resolved laser fluorescence spectroscopy of organic ligands by europium: Fluorescence quenching and lifetime properties

Time–resolved Laser Fluorescence Spectroscopy (TRLFS) has proved its usefulness in the fields of biophysics, life science and geochemistry to characterize the fluorescence probe molecule with its chemical environment. The purpose of this study is to demonstrate the applicability of this powerful technique combined with Steady-State (S-S) measurements. A multi–mode factor analysis, in particular CP/PARAFAC, was used to analyze the interaction between Europium (Eu) and Humic substances (HSs) extracted from Saint Lawrence Estuary in Canada. The Saint Lawrence system is a semi–enclosed water stream with connections to the Atlantic Ocean and is an excellent natural laboratory. CP/PARAFAC applied to fluorescence S-S data allows introspecting ligands– metal interactions and the one–site 1:1 modeling gives information about the stability constants. From the spectral signatures and decay lifetimes data given by TRLFS, one can deduce the fluorescence quenching which modifies the fluorescence and discuss its mechanisms. Results indicated a relatively strong binding ability between europium and humic substances samples (Log K value varies from 3.38 to 5.08 at pH 7.00). Using the Stern-Volmer plot, it has been concluded that static and dynamic quenching takes places in the case of salicylic acid and europium interaction while for HSs interaction only a static quenching ∗Corresponding author Email address: nouhi@univ-tln.fr (A. Nouhi ) Preprint submitted to Journal of LTEX Templates November 13, 2017 is observed.

Fluorescence spectroscopy in S-S measurements provides quantitative and qualitative information. However, more data such as fluorescence lifetime measurements could be gathered using TRLFS. In fact, TRLFS is a convenient technique when dealing with different elements in the same sample because it can offer 15 the spectral and temporal resolution together with its high sensitivity. The discriminating capability of TRLFS relies on the fact that the different chemical species between organic matter and a fluorescent metal ion can have different decay lifetimes and spectral shapes which is adequate for the study of such a complex system Collins et al. (2011). The greatest advantages of TRLFS are 20 its relatively low detection limit and good discriminating capability for different chemical species.  from the ligand site, but only for SA Aoyagi et al. (2004). Analysis of the literature has shown that data of humic substances complexes with Eu(III) from the ligand site is still lacking. As a matter of fact, it raises some interesting questions that have not been discussed before. Therefore, with this study, we wanted to investigate how these interactions occur focusing on the HSs fluorescence emis-45 sion. Analysis of fluorescence lifetimes and Stern-Volmer plots Lakowicz (2006) will give reliable information on the quenching process that takes place. posed of a complex mixture of organic and organomineral compounds and can differ depending on its type. Therefore, SA provides a carboxyl and hydroxyl group that mimic the important functionalities of HSs Stevenson (1994) and form strong complexes with Eu(III) Kuke et al. (2010).
Salicylic acid (NormaPur R , Prolabo) was used as received. Its aqueous solu-60 tion was prepared at a concentration of 5 × 10 −4 M with UHQ water. HA and FA were obtained from the St. Lawrence Estuary Tremblay and Gagné (2009) following the IHSS protocol Swift (1996).
Europium standard (1,000 mg.L −1 in 2% HNO 3 , SPEX CertiPrep) was used as a quencher without further purification. Necessary pH adjustments of metal solutions were done with NaOH (1M) and HCl (1 M) and the changes in the volume were taken into account.  Every mixture was allowed to equilibrate for at least 10 min and bubbled by a nitrogen flux to avoid photobleaching. 5

Data Treatment
For S-S data treatment, CP/PARAFAC algorithm was used to determine 115 the contribution of components to the EEMs by separating independent sources after elimination of scattering phenomena Bro (1997 was selected as optimal model. Iterations and number of runs were implemented until the algorithm converged. The optimization between experimental and calculated fluorescence using one-130 site model with one ligand Ryan and Weber (1982) was executed by evaluating the bias errors. The bias is equal to the sum of the square of the subtraction between the experimental fluorescence intensity and the calculated fluorescence intensity.
Data obtained with TRLFS were analyzed using a home-made MATLAB R 135 program. The evaluation of the best fitting of the integrated fluorescence intensity was done by time deconvolution to determine fluorescence amplitudes and fluorescence lifetimes. Supposing a system excited by an infinitely sharp pulse of light (δ-function) and assuming the response to be linear, the measured fluorescence decay is given by the convolution integral : where h(n) is the intrinsic fluorescence decay and IRF is the Instrument Response Function. Experimentally, the IRF is gathered by measuring only the 6 solvent diffusion.
The estimation of h(n) was achieved by time deconvolution in order to obtain the amplitudes of each fluorophore and lifetimes. Assuming a multi-exponential 145 decay as follow : The amplitudes α i and the lifetimes τ i can be considered to be adjustable parameters. They are calculated employing Simplex algorithm minimizing the error of the equation : It is known that quenching can occur by different mechanisms. Dynamic quench-150 ing occurs when the excited-state fluorophore is deactivated upon contact with other molecules in solutions due to physical collision Lakowicz (2006). In this case, the fluorophore returned to the ground-state during a diffusive encounter with the quencher. The molecules are not chemically modified in the process and can be shown by the Stern-Volmer equation 4.
where I 0 and I Q are the fluorescence intensities in the absence and presence of quencher respectively, τ Q is the fluorescence lifetime in the presence of metal, K SV is the Stern-Volmer quenching constant and k q is the bimolecular quenching constant.
If dynamic quenching takes place, the intensity of fluorescence and lifetime fol-160 lows the relationship: In this expression τ 0 is the fluorescence lifetime in the absence of quencher. It is noteworthy that dynamic quenching is a time-dependent process.
Aside from collisional quenching, fluorescence quenching can occur due to a complex formation. Fluorophores can form non-fluorescent complexes with 165 quenchers Lakowicz (2006). This process is referred to as static quenching since it occurs at ground-state and does not rely on diffusion or molecular collisions.

7
The dependence of the fluorescence intensity upon quencher concentration is given by the following expression where K S is the association constant of the complex.

170
Note that the I 0 /I Q on [Q] is linear, which is identical to dynamic quenching.
Hence, the measurement of fluorescence lifetimes is the most ultimate method to distinguish static and dynamic quenching.

175
The CP/PARAFAC decomposition of SA-Eu 3+ quenching experiment shows that two components are obtained in this interaction (Fig. 1).  Table 1 and was estimated as Log K SA = 3.38 based on the 1 : 1 complexation assumption between metal and ligand. This value is higher than Log K=2.08 ± 0.02 obtained by Aoyagi et al. (2004) at pH 4 from the slope of the intensity SternVolmer plot and higher than Log K=2.02±0.05 obtained by Hasegawa et al. (1989) at pH 6 from the glass electrode method.

190
The stability constant of SA-Eu 3+ found in this study is consistent considering that the stability constant should increase when pH increases.  interaction with Eu 3+ suggests the presence of three components (Fig. 1.B). The first component in the UV range is located at (λ ex /λ em )=270/445 nm, and has been designated as peak A. It was attributed to humic substances Coble (1996).
The relative contribution of this component to the EEMs during the quenching experiment is shown in Fig. 2.B. and the corresponding stability constant is For the FA, two components were found by CP/PARAFAC showing an interesting fluorescence quenching (Fig. 2. C.) The first component ( Fig. 1. C) located at (λ ex /λ em )=310/415 nm is generally another marine humic-like component and is designated as peak M Coble (1996). The second component 220 (λ ex /λ em )=265/475 nm is larger and is designated as peak A Coble (1996) with a slight red shift in emission. The corresponding fluorescence intensities are given in Fig. 2

TRLFS results
As mentioned above in section 2.3, the quenching process cannot be assessed using steady-state measurements. Therefore, fluorescence lifetime measurements using TRLFS are the most ultimate method to distinguish whether static, dynamic or both quenching processes occur based on Stern-Volmer plots Lakow- For the interaction between SA-Eu 3+ , the decay was found to be monoexponential. The average lifetime τ a recorded was 2.97±0.2 ns. In Fig. 3.A This discrepancy could be due to the pH value as they worked at pH 4.
For HA-Eu 3+ interaction, CP/PARAFAC decomposition gives three components while the time deconvolution suggests a bi-exponential decay and two lifetimes denoted by exponent indexes τ A and τ B . One very short, indicated by Fig. 3. B and the second one, specified by τ B 0 /τ B Q , longer than the first one (11.33 ns). The two fluorescence lifetimes turned out to be independent on total concentration of Eu 3+ . This suggests a static quenching. The dependence of I 0 /I Q of component 1 and component 2 on [Eu 3+ ] is therefore non-linear.
In the case of static and dynamic quenching, the characteristic feature of the 265 Stern-Volmer plot is an upward curvature, concave towards the y-axis. The concave pattern of the Stern-Volmer plot is interpreted as the manifestation of fluorophores that are not accessible for quenching Lakowicz (2006). Thus, the data cannot be accurately fitted using the linear Stern-Volmer equation 4 in this situation. However, Reiller and Brevet (2010)   The intensity ratios of the fulvic acid with Eu 3+ in Fig. 3. C. seem to be linear except the penultimate point which could be attributed to an error in the measurement. According to Tiseanu et al. (1998), the existence of a tri-modal decay time distribution with time values centered around 0.7, 3 and 11 ns was reported. In this study, we found a bi-exponential time decay. The first average decay time component (0.7ns) refered to as τ A 0 /τ A Q may be assumed as typical for FA of different origins Kumke et al. (1998) and similar to Tiseanu et al. (1998). The second lifetime indicated by τ B 0 /τ B Q in Fig. 3. C. is around 12 ns and is similar too to the literature. In the presence of Eu 3+ , the dependency of the relative amplitudes of the fluorescence decay of FA was almost unchanged quenching. The previous study ) also reported a static quenching in the case of FA-Eu(III) interaction.

Conclusion
In this work, the interactions of salicylic acid, humic acid and fulvic acid with Eu 3+ were studied by steady-state and time-resolved laser-induced fluorescence 295 spectroscopy focusing on the ligand fluorescence emission using a nanosecond pulsed laser. Lifetime analysis using Stern-Volmer plot showed that part of dynamic quenching process or both static and dynamic quenching contributed to the global quenching of fluorescence in the case of salicylic acid-Eu 3+ interaction.
Data obtained is not exactly in agreement with literature Aoyagi et al. (2004) since they found a static quenching in the case of SA-Eu(III) interaction. This discrepancy with SA should be investigated at different pH to define precisely the quenching process dependence on pH. Static quenching process took place for humic acid and fulvic acid extracted from St. Lawrence Estuary. Tiseanu et al. (1998) also found a static quenching in the case of FA-Eu(III) interaction, 305 which is in a good agreement with our study. The stability constants of all the interactions studied were calculated using the one-site 1 : 1 model and are in concordance with literature for SA and HA. For FA, we found constants slightly lower than those in the literature which may be due to the origin of FA. It is noteworthy that TRLFS successfully discriminates the fluorescence quenching 310 process that takes place and gives original lifetimes of HSs and constant of complexation with Eu(III). Further application can be extended to quantitatively analyze other lanthanides and actinides.