Binding of Calcium and Magnesium Ions to Chromophoric Dissolved Organic Matter (CDOM): A Combination of Steady-State and Time-Resolved Fluoresce Study

Revealing the binding properties of calcium ion (Ca2+) and magnesium ion (Mg2+) to chromophoric dissolved organic matter (CDOM) facilities understanding the effect of natural water composition on the photophysics of dissolved organic matter. Steady-state and time-resolved fluorescence spectrometry, and dynamic light scattering were applied to investigate the fluorescence quenching process of CODM by Ca2+ and Mg2+. The binding of Ca2+ and Mg2+ preferred terrestrial CDOM to aquatic CDOM. The fluorescence quenching of CDOM by cations mainly occurred in a static process, which was based on the fact that the decrease of steady-state fluorescence intensity was greater than fluorescence lifetime. The fluorescence quenching was profound under longer excitation and emission wavelength. The binding constant (K, L/mol) for Ca2+ to CDOM from terrestrial source ranged from 4.29 to 5.09 (lgK), which was approximately one order of magnitude higher than that of Mg2+ to CDOM (3.86 to 4.56). Fluorescence decay became faster in the presence of Ca2+ and Mg2+. Lifetime distribution of CDOM excited states shifted to small value side in the presence of metal ions, particularly for Ca2+, indicating fluorescence quenching of CDOM mainly through the interaction of Ca2+/Mg2+ with relatively long-lived fluorophores.


Introduction
Natural organic matter (NOM) is ubiquitous in natural water. NOM plays a key role in biological and chemical processes because of the direct competition for solar irradiance and the subsequently generated reactive species [1][2][3][4][5]. Metal ions tend to bind to NOM mainly through the phenolic and carboxylic functional groups. [6] The binding of metal ions to NOM not only altered the bioavailability of metal ions but also affected the photophysical and photochemical properties of NOM [5,7,8].
Among various metal ions, copper ion (Cu 2+ ) [9][10][11] and mercury ion (Hg 2+ ) [12,13], were the most frequently investigated species. Previous studies demonstrated that binding constants between Cu 2+ , Hg 2+ , and NOM could be up to 10 6 L/mol [14,15], leading to a decreased fluorescence intensity of NOM. The binding constants were dependent on NOM source (structure) and experimental conditions such as solution pH and salinity [9,16,17]. The competition of co-existing cations for binding sites in NOM was used to explain the inhibition effect of salinity on the binding of conducted metal ions to NOM [9,[18][19][20]. For example, the presence of up to 11×10 -6 mol/L Ca 2+ broke Hg-NOM complex and increased the fluorescence intensity, whereas Ca 2+ exhibited no apparent effect on NOM fluorescence intensity [19]. However, the binding properties of Ca 2+ and Mg 2+ to NOM have not been fully documented.
Notably, binding properties of metal ions to NOM (components) were determined by Excitation-Emission matrix (EEM) followed with parallel factor analysis (PARAFAC). The knowledge about the effects of metal ions on NOM's fluorescence lifetime is still quite limited [21]. In our previous study [22], we found that fluorophores with relatively short lifetime (< 1 ns) dominated CDOM (the major composition of NOM) excited populations, while these short-lived excited states could not be well documented by steady-state fluorescence because of their minor contribution. Static quenching through the formation of metal-NOM complex was proposed to explain the fluorescence quenching. Nevertheless, certain fluorophores had relatively long lifetime (approximately 10 ns), collision quenching would not be excluded in the presence of millimolar level metal ions.
In this study, interactions between Ca 2+ and Mg 2+ and CDOM (aquatic and terrestrial source) were systematically investigated by steady-state and time-resolved fluorescence spectrometry. The first objective is to quantify fluorescence quenching extents and determine binding constant. The second objective is to reveal the mechanism for fluorescence quenching based on the effects of Ca 2+ and Mg 2+ on the quenching of steady-state fluorescence intensity and fluorescence lifetime, and the change of CDOM dispersity.

Chemicals
Analytical grade CaCl2 and MgSO4 were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China) and used as received. CDOM reference materials, Suwannee River natural organic matter (SRNOM, Lot 2R101N), Suwannee River Humic Acid Standard III (SRHA, Lot 3S101H), Elliott Soil HA Standard IV (ESHA, Lot 5S102H), and Pahokee Peat Humic Acid Standard (PPHA, Lot 1S103H), were purchased from the International Humic Substance Society. CDOM was added to pH 10.0 water and stirred for two hours. Then, pH of the CDOM solution was adjusted to 7.0 and filtered by a 0.22 m nylon membrane (Shanghai Xingya Purification material Factory, China). All the CDOM stock solutions were stored in dark at 4°C and used within one month. The concentrations of CDOM stock solutions were analyzed on an Elementa total organic carbon analyzer (Vario TOC; Elementar, Langenselbold, Germany).

Experimental procedure
Working solutions of CDOM (10 mg C/L), Ca 2+ (0-30 mg/L), and Mg 2+ (0-70 mg/L) were prepared by dilution. The resulting solutions were shaken at 25 ± 0.2°C for at least 12 hours to reach binding equilibrium. The pH values of the working solutions were adjusted to 7.5 by phosphate buffer. The ionic strength was adjusted to 20 mM for all working solutions.

Spectra recording method
Absorption spectra: The absorption spectra were recorded with an ultraviolet (UV)visible spectrometer (UV 2550; Shimadzu, Tokyo, Japan) at 25 ± 0.5°C. The optical path length of the cuvette was 1 cm.
Steady-state fluorescence spectra of CDOM solutions: Steady-state fluorescence spectra of CDOM were recorded with either an Edinburgh FS-5 or a Horiba FluoroLog-3 fluorimeter. The excitation (Ex) and emission (Em) wavelength range for the excitation emission matrix (EEM) fluorescence spectra of CDOM in the absence and presence of Ca 2+ and Mg 2+ were 250−600 nm and 270−750 nm, respectively. The stepsize was 5 nm and 2 nm for excitation and emission, respectively. The slit size was 5 nm. The dwell time was 0.100 s. Fluorescence intensity was corrected using factory-measured correction files. Ultrapure water was used for background and scatter subtraction, and inner-filter effect was also corrected.
To get a more precise fluorescence quenching efficiency, the single emission profile (400−800 nm, increments of 1 nm, dwell time was 0.100 s) was recorded at excitation wavelengths of 375, 440, and 550 nm, and the intensity correction followed the same protocol for EEM fluorescence spectra.
Time-resolved fluorescence spectra of CDOM solutions: Time-resolved fluorescence spectra of CDOM were measured with a FluoTime 200 equipment (PicoQuant GmbH, Berlin, Germany) at 25 ± 1 °C. The excitation source was a picosecond laser diode (LDH-P-C-375B for 375 nm, LDH-P-C-440B for 440 nm). The fluorescence decay of CDOM was detected using a PMA hybrid 07 detector and recorded in a Picoharp 300 time-correlated single-photon counter. The maximum intensity of CDOM was 10, 000 counts, and the counting rate was less than 1% of the excitation rate to avoid pile-up problem. The instrument response function (IRF) was recorded using a Ludox solution by recording the scattered light at 375 nm or 440 nm. Lifetime distribution analysis was applied to model emission decays of CDOM with FAST program (Version 3.5, Edinburgh). Detailed procedures for data analysis and fitting criteria were reported in a previous study. [22] Hydrodynamic diameters of CDOM solutions: Hydrodynamic diameters of CDOM solutions (ESHA: 60 mg C/L; PPHA: 100 mg C/L) in the presence and in the absence of Ca 2+ and Mg 2+ were determined using a Malvern Zen 3700 Zetasizer (Worcestershire, UK).

Changes of CDOM's absorption spectra with Ca 2+ and Mg 2+
As shown in Figure S1, the absorption spectra of ESHA and PPHA did not change much in the presence of Ca 2+ or Mg 2+ , particularly in the wavelength longer than 400 nm. An inconspicuous but discernible change of absorption spectra was reported for the binding of Ca 2+ to SRHA [23]. Notably, light scattering became notable for ESHA in the presence of high Ca 2+ (such as 20 mg/L). The increase of light scattering was in line with the change of hydrodynamic size with Ca 2+ and Mg 2+ ( Figure S2). In the presence of Ca 2+ or Mg 2+ , the hydrodynamic size of ESHA and PPHA became larger, particularly for ESHA. Similar results were also reported for fulvic acid after binding Cu 2+ , Pb 2+ , or Cd 2+ [24], which resulted from cation bridge effects [25].

Changes of CDOM's emission spectra with Ca 2+ and Mg 2+
Compared with the slight change of absorption spectra, the steady-state fluorescence intensity of ESHA and PPHA significantly decreased with an increase of Ca 2+ or Mg 2+ concentration in the whole conducted spectra range ( Figure 1). The inhibition effect of Ca 2+ was more evident than Mg 2+ , while the decrease intensity of ESHA was more profound than PPHA ( Figure S3). As demonstrated in Figure 2, under the excitation of 375, 440, and 550 nm, with an increase of Ca 2+ (0−30 mg/L) or Mg 2+ (0−70 mg/L), the fluorescence intensity of ESHA decreased gradually, and remained almost unchanged at high Ca 2+ concentration (>20 mg/L) or high Mg 2+ concentration (>50 mg/L). The change of PPHA fluorescence intensity with Ca 2+ and Mg 2+ was analogous to ESHA, but to a less extent ( Figure S4). Notably, Ca 2+ and Mg 2+ (20 mg/L) did not quench the fluorescence of either SRNOM or SRHA (data not shown).   (Figure 3) [26]. As listed in Table 1, lgK was excitation and emission wavelength dependent, which was larger at longer excitation and emission wavelength. For instance, under the excitation of 375 nm, K for binding of Mg 2+ and Ca 2+ to ESHA at emission 650 nm was 1.36 and 1.92 times that for emission at 550 nm, respectively. With excitation wavelength increased from 375 nm to 550 nm (Em=650 nm), K for binding of Mg 2+ to ESHA and Ca 2+ to ESHA increased from 1.39×10 4 to 1.72×10 4 L/mol and from 9.

Changes of CDOM's emission decays with Ca 2+ and Mg 2+
Emission decays of CDOM were only slightly affected by Ca 2+ and Mg 2+ . For example, emission decays of ESHA were not affected by either 5 mg/L Ca 2+ or 10 mg/L Mg 2+ ( Figure  4). Analogous results were found in the case of PPHA ( Figure S5). Since it was proved that results obtained from multi-exponential analysis of fluorescence decays had no physical meaning [22], lifetime distribution analysis was applied to quantify the effects of Ca 2+ and Mg 2+ on the deactivation of excited states in CDOM.

4.1.Binding of Ca 2+ and Mg 2+ to CDOM
Consistent with previous reports for other heavy metal ions [27], the binding affinity of Ca 2+ and Mg 2+ to terrestrial source CDOM was larger than that to aquatic source CDOM. Similar results were also reported for fluorescence quenching of CDOM by other organic quenchers [28], in which fluorescence of CDOM or its fractions with larger molecular size were more evidently quenched.
The binding constant of Ca 2+ to CDOM was slightly smaller than that of Cu 2+ reported [9,[18][19][20]. Notably, CDOM used in this study differed from previous studies. Quenching of ESHA by Cu 2+ was further investigated to exclude the effect of CDOM structure on the binding affinity. As shown in Figure S8, the fluorescence quenching efficiency for ESHA at emission wavelength longer than 600 nm by 10 mg/L Ca 2+ was approximately 1.5 times that by 0.3 mg/L Cu 2+ , while the quenching efficiency for ESHA by 0.3 mg/L Cu 2+ was comparable to that by 50 mg/L Mg 2+ . These results clearly supported the finding of a smaller lgK for Cu 2+ to organic matter in a higher salinity water [9]. Moreover, due to diverse water hardness or other chemical compositions, the accuracy of quantifying CDOM or NOM concentration only by in-situ fluorescence spectrometry is doubtful.

4.2.Process for fluoroscence quenching of CDOM by Ca 2+ and Mg 2+
Fluorescence decays of fluorophore would not change in a typical static quenching process. Under the scenario of CDOM, the situation could be complex. Fluorescence decays became faster in the presence of cations, particularly in the case of Ca 2+ , leading to a decreased amplitude-weighted average lifetime (a). The decrease of a could result from several possibilities. 1) Ca 2+ or Mg 2+ quenched excited state of fluorophores in CDOM. 2) Ca 2+ or Mg 2+ formed complex with fluorophores with relatively long lifetime relative to fluorophores with relatively short lifetime. 3) Larger CDOM aggregates formed in the presence of Ca 2+ or Mg 2+ , which could increase light scattering to some extent.
Under the first scenario, fluorophores with the longest lifetime were more likely involved in collisional quenching. However, as demonstrated in Figure 5 and lifetime distribution on the large value side has not shifted except for a decrease in their amplitude. For the component with a lifetime smaller than 1 ns, collisional quenching was negligible when taking a millimolar Ca 2+ or Mg 2+ and a diffusion-controlled rate constant into account. Therefore, collision quenching was not the cause for the decrease of steadystate fluorescence intensity or acceleration of emission decay. The second scenario requires fluorophores within CDOM were not equally quenched, which was supported by the dependence of quenching efficiency on excitation and emission wavelength ( Figure S8). Moreover, as shown in inset Figure 6, amplitude for fluorophores with a lifetime longer than approximately 1 ns decreased with Ca 2+ larger than 5 mg/L. Therefore, the overall decay of CDOM became faster in the presence of Ca 2+ or Mg 2+ . Moreover, the distribution of fluorophore lifetimes was more isolated in the presence of cations than in the absence of cations, indicating the complete quenching of some fluorophores or the interactions between fluorophores was weakened in the presence of Ca 2+ or Mg 2+ .
The third scenario was supported by the different response of steady-state fluorescence intensity and emission decay to Ca 2+ or Mg 2+ concentration. For instance, steadystate fluorescence intensity of ESHA barely changed when Ca 2+ or Mg 2+ was higher than 20 mg/L or 40 mg/L (Figure 3), respectively. This result suggested that the composition or concentration of light-emitting fluorophores did not change, which conflicted with the further acceleration of emission decays. As shown in Figures S1 and S2, larger aggregates or even precipitates formed in the presence of high concentration of Ca 2+ or Mg 2+ , which would increase light scattering that contributed to the fast decay component ( Figure 6). Notably, formation of aggregates could lead to self-fluorescence quenching.

4.3.The residues of CDOM fluorescence in the presence of Ca 2+ and Mg 2+
Results obtained in this study show that CDOM maintained at least 10% percentage of its original steady-state fluorescence intensity among the conducted wavelength range. There are two candidates for the residual fluorescence. The first source is the unquenched fluorophore, and the second source is the metal-CDOM complex.
The first resource requires that some fluorophores are not accessible to metal ions or other quenchers. Recently, the lack of classical solvatochromism phenomena in CDOM optical spectra challenged the validation of charge-transfer model in interpreting the photophysical properties of CDOM [29], which put forward the argument that whether fluorophore responding for locally-excited states or charge transfer complex was freely diffusing and fully accessible [30,31]. As shown in Figures 5 and 6, fluorophores with long lifetime still presented in CDOM solution at high Ca 2+ or Mg 2+ concentration, which clearly showed that certain fluorophores were not accessible to these cations.
Notably, the distribution of ESHA's fluorescence lifetime shifted to the small value side in the presence of 5 mg/L Ca 2+ . However, the shortest lifetime was still longer than 100 ps, which might not result from stray light. Therefore, metal-CDOM complex was also likely to contribute to residual fluorescence. Nevertheless, their contribution was minor because of their large deactivation rate constant.

Conclusions
Based on the change of optical spectra and hydrodynamic size of CDOM with Ca 2+ and Mg 2+ , the current findings showed that both Ca 2+ and Mg 2+ had high binding constant to terrestrial source CDOM. The formation of metal-CDOM complex quenched CDOM fluorescence, whereas collisional quenching could not occur. Moreover, static quenching was more evident for fluorophores with relatively long lifetime. The excitation and emission wavelength dependent quenching efficiency suggested that fluorophores were quenched to different extents. Fluorophores with long lifetime were not fully quenched by Ca 2+ or Mg 2+ , which suggested that metal-CDOM complex was not the only contributor to the unquenched fluorescence.
Supplementary Materials: Figure S1: Absorption spectra of ESHA and PPHA (10 mg C/L) in the presence of Mg 2+ and Ca 2+ , Figure S2: Effects of Ca 2+ or Mg 2+ on the hydrodynamic size of ESHA (60 mg C/L) and PPHA (100 mg C/L) aqueous solution(pH 7.5), Figure S3: Effects of Mg 2+ and Ca 2+ on the EEM spectra of PPHA (10 mg C/L) aqueous solution, Figure S4: Effects of Mg 2+ and Ca 2+ on the steady-state fluorescence spectra of PPHA (10 mg C/L) aqueous solution at excitation of 375, 440, and 550 nm, Figure S5: Effects of Mg 2+ (10 mg/L) or Ca 2+ (10 mg/L) on the time-resolved fluorescence spectra of PPHA (10 mg C/L) aqueous solution at excitation of 375 and 440 nm, Figure S6

Conflicts of Interest:
The authors declare no conflict of interest.