Development of SiO2-coumarin fluorescent nanohybrid and its application for Cu(II) sensing in aqueous extracts of roadside soil

A SiO2-coumarin nanohybrid was investigated for its Cu(II) sensing performance in aqueous media, and in comparison with the Cu(II)-selective coumarin used alone. Fluorescence of both coumarin itself and the nanohybrid, λex/λem 435/481 nm, was selectively quenched by Cu(II) when tested against a range of multivalent cations. The nanohybrid had enhanced Cu(II) sensing properties when compared to the coumarin including (i) improved limit of detection from μM-level (0.48 μM) of Cu(II) using coumarin alone to nM-level (0.033 μM) and (ii) an extended linear detection range of 0.033–260 μM (0.0005–4.1 mg/mL) Cu(II) compared to 0.48–55 μM for the coumarin itself. The lower limit of detection and extended range were achieved with a smaller amount of coumarin and no traces of organic solvents used to help coumarin dissolution. Characterization suggested that under applied test conditions at pH = 5, SiO2 nanoparticles with negative surface charges adsorbed coumarin and then (when present) Cu(II) ions. The SiO2-coumarin nanohybrid was then applied for the determination of Cu(II) levels in aqueous soil extracts reaching over 94% recovery rates when used against the standard soil analysis method by inductively coupled plasma mass spectrometry (ICP-MS).


Introduction
Copper is an essential trace element that plays a significant role in biological and environmental applications. In the human body, Cu(II)-and Cu(II)dependent enzymes are involved in critical biological processes to ensure normal cellular function (Grasso et al. 2013;Domaille et al. 2010). However, excessive copper intake can lead to imbalance and harmful effects in living organisms (Hao et al. 2015;Zhang et al. 2014). In animals and humans, Cu(II) has been shown to have a role in the development of neurodegenerative disorders such as Menkes, Wilson's, and Alzheimer's diseases, and influence the development of certain cancers (Ge et al. 2021). Copper exposure can occur via contamination of soil and drinking water originating mostly from human activities, such as corrosion (Jakeria et al. 2021;Tang et al. 2021), mining, or lackadaisical waste management practices. Thus, easy-to-use, affordable, and sustainable sensing methods for Cu(II) are highly sought after in clinical, food/water, and environmental industries (Ji et al. 2017;Zhu et al. 2012). In addition, the allowable Cu(II) is highly regulated in many environmental media, such as in soil (e.g. 100 mg/kg in Australia) (He et al. 2015) and drinking water (e.g. 2.0 mg/mL in Australia) (Tang et al. 2021;Seeley et al. 2013).
Among existing sensing techniques, optical sensing using colorimetric or fluorescent molecular probes is readily available and non-expensive tool for the analysis of liquid samples and dissolved Cu(II) levels. An often employed sub-class of fluorescent chemosensors for Cu(II) (Zhang et al. 2014;Ramdass et al. 2017;Sivaraman et al. 2018) sensing are coumarin derivatives (Kraljević et al. 2016;Seo et al. 2011) due to their sufficient photostability, tuneable sensitivity, and selectivity (Geißler et al. 2005;Li et al. 2019). For example, hydrazine derivative 1 was developed as a "turnon" fluorescent probe for the detection of Cu(II) in aqueous acetonitrile with a limit of detection (LOD) of 0.058 μM (Table 1) (Zhang et al. 2019). In comparison, the coumarin-N-acylhydrazone derivative 2 was utilized for both colorimetric and fluorometric detection of Cu(II) in CH 3 CN/H 2 O (v/v, 9/1) with a LOD of 8 μM (Table 1) (Li et al. 2019). While there are a large number of fluorescent probes reported for the sensing of particular metal ions in solution, and despite their satisfactory selectivity, they are often limited by insufficient water solubility reducing their ease of use and applicability for environmentally friendly, on-site sensing (Huang et al. 2014;Wang et al. 2016) Another limiting factor is their inability to selectively distinguish the copper cations from a mixture containing multiple metal cations with similar electronic properties, such as in the case of soil extracts.
One way to overcome mentioned shortfalls is to use nanomaterials that can be surface modified with molecular probes, and that facilitates the dispersion of analyte-selective dyes in aqueous solutions while preserving the sensing performance, especially selectivity (Lin et al. 2006). Such approaches often result in improvements in the limit of detection (LOD) due to the changes in surface to mass ratio of the hybrid sensing materials (Meiling and Guowen 2017) Among others, SiO 2 nanoparticles (SiO 2 NPs) are water dispersible and non-toxic with well-established, facile synthesis methods via, for example, the the Stöber method (Stöber et al. 1968). A previously reported combination using negatively charged SiO 2 NPs and Table 1 Examples of fluorescent Cu(II) sensors and their reported excitation and emission wavelengths, and limits of detection in given solvents Page 3 of 12 114 Vol.: (0123456789) positively charged molecular probe 4 (Table 1) was demonstrated for its use as a metal ion sensor in aqueous media (Peng et al. 2018). The authors discussed that electrostatic interactions between the dye 4 and the SiO 2 NP surface resulted in a locally increased concentration of 4 while an enhancement of the fluorescence intensity was recorded. For Cu(II) in tap water, use of 0.25 mg/mL SiO 2 -4 was delivering a sensing range of 11.2-100 nM and an LOD of 0.2 nM, an improvement from the LOD of 10 μM when 4 was used alone (Peng et al. 2018). A fluorescent, covalently appended SiO 2 -naphthalin sensor was also reported for Hg(II) detection with a sensing range of 0.1-1 μM and LOD of 6.8 nM in water between pH 4.5 and 8 (He et al. 2009).
The current study aimed to improve the aqueous dispersibility and Cu(II) sensing performance of previously reported coumarin 3 (Qian et al. 2019) by coupling to non-toxic SiO 2 NPs to form SiO 2 -3 nanohybrid. The aim of this work is for on-site, facile, low-cost measurements of water and soil quality that can be conducted by non-trained personnel.
Synthesis of coumarin 3, SiO 2 NPs, and the SiO 2 -3 nanohybrid is described. Then relevant optical, physical, and chemical properties are discussed and compared for each material in aqueous media and in the absence and presence of Cu(II) and other metal ions. While the interaction of Cu(II) with coumarin 3 is well studied, here the focus is given to its interaction with SiO 2 NPs and Cu(II) to understand how these affect the observed sensing properties. Finally, the SiO 2 -3 nanohybrid was tested on aqueous extracts of soil samples and its Cu(II) sensing performance was validated against the industry standard, inductively coupled plasma mass spectrometry (ICP-MS) method.
Infrared spectra were recorded on Nicolet 6700 FTIR (Thermo Scientific, USA). UV-vis and fluorescence spectra were measured on a Cary Eclipse 5G UV-Vis-NIR spectrometer and a Cary Eclipse fluorescence spectrophotometer (USA), respectively. Nuclear magnetic resonance and mass spectra were recorded on Bruker Ascend 400 Hz and Thermo Q Exactive in a high-resolution electrospray mode mass spectrometer, respectively. The SiO 2 NP separation process was conducted using an Eppendorf Mini Spin centrifuge. A Jeol 1010 TEM instrument was used for the imaging of SiO 2 NPs and SiO 2 -3 complex structure. Size distribution and surface charge were studied using a Malvern ZEN3600 Zetasizer instrument. Elemental analyses of the SiO 2 NP surface were characterized by X-ray photoelectron spectroscopy (Thermo Scientific K-Alpha XPS). Thermogravimetric analysis (TGA) was conducted on Mettler Toledo micro balances. ICP-MS (Agilent with laser ablation capability) was used for the quantification of metal ion concentration in soil extracts and ICP-OES (Varian Vista-Pro) was applied for the Cu(II) concentration before and after the interaction with SiO 2 NPs.

Synthesis of coumarin 3 and SiO 2 nanoparticles
Coumarin 3 (Table 1) was synthesized via coumaric acid precursors as previously published (Nielsen and Houlihan 2004) (and detailed in the Supporting Information), followed by amide bond formation using PyBOP as a coupling agent (Scheme SI 1). The moderate yield was counter balanced by a simple work-up procedure by the removal of water-soluble by-products (SI Scheme 1) (Qian et al. 2019). After the preparation of SiO 2 NPs via Stöber synthesis (Fors et al. 2013), the SiO 2 -3 nanohybrid was assembled as described in the "Experiment" section. The color of the asreceived nanohybrid was yellow compared to the white SiO 2 NPs due to the presence of coumarin 3. 114 Page 4 of 12 Vol:. (1234567890)

Preparation of SiO 2 -3 nanohybrid
The preparation of coumarin 3 and SiO 2 NPs can be found in the supporting information (SI 1.1-1.3). To assemble SiO 2 -3, SiO 2 NP (50 mg) was mixed into an aqueous 250-μM suspension of 3 (100 mL). The pH was adjusted to 5 using 0.01 M hydrochloric acid, and the resulting mixture was incubated at 37 ℃ for 2 h while gently stirred. The SiO 2 -3 nanohybrid was separated from the mixture by centrifuging the solution at 13,000 rpm and decanting the liquid phase and retaining the supernatant. The resulting particles were washed with Milli-Q water (3 × 1.5 mL) and pH = 5 was adjusted using hydrochloric acid, to remove the non-adsorbed coumarins. The obtained nanohybrid solution was centrifuged at 13,000 rpm and dried in a vacuum oven at 50 °C.

Concentration dependence of fluorescence intensities
Fluorescence spectra of 3 (40 μM) was recorded at pH 5 at concentrations of 5-80 μM in DMSO/ HEPES buffer (1:9, v/v, 20 mM, pH = 5) at λ ex = 435 nm with slit width of 5 nm. Meanwhile, for the fluorescence intensity of SiO 2 -3 nanohybrid, the stock solution was diluted into the concentrations from 5 to 50 mg/L using HEPES buffer (20 mM, pH = 5) and tested using the same instrument settings.
pH dependence of fluorescence intensities Fluorescence intensity of 3 (40 μM) and SiO 2 -3 nanohybrid (7.5 mg/L) at λ ex = 435 nm with slit widths of 5 nm was recorded, respectively, across the pH range of 3-13, which was adjusted by 1 mmol NaOH and 1 mmol HCl solution.

Quantitation of surface-adsorbed Cu(II) on SiO 2 NPs and SiO 2 -3 nanohybrid by ICP-OES
A 30 mg/L stock solution of SiO 2 NPs (10 mL) was mixed with 300 μM CuCl 2 solution (10 mL) and then incubated at 37 ℃ for 2 h. The final concentration of SiO 2 NPs was 15 mg/L and Cu(II) concentration was 9.4 mg/L. Then, the SiO 2 NPs were separated from the mixture by centrifuging the solution at 13,000 rpm. The resulting solution (top layer) was sent to ICP-OES for the measurement of residual Cu(II) concentration. The same process was applied for the SiO 2 -3 nanohybrid.
Interference studies Cu(II) sensing in real soil samples Surface soil samples were taken from roadsides in the west of Melbourne, Victoria, Australia. Each soil sample (2.5 g) was dispersed in water (25 mL) and shaken for 1 week to extract water-soluble constituents, including copper. The extracts were filtered through a nylon filter membrane with a pore size of 0.45 μm and analyzed for Cu using ICP-MS. The extracts were also measured for their Cu(II) concentration via fluorescence as described previously.
Comparison of the thermogravimetric analysis (TGA) curves recorded for SiO 2 -3 nanohybrid (black line) and SiO 2 NPs (red line) at a heating rate of 5 °C/min (Fig. 3) provided approximation of the quantity of surface-adsorbed coumarins as 3 has a lower decomposition temperature than SiO 2 NPs. For the SiO 2 -3 nanohybrid, weight loss of 18% (100 to 82%) was observed between room temperature to about 200 °C, attributed to the evaporation of physically bound water (Tham et al. 2020). Beyond, the SiO 2 -3 nanohybrid displayed 7.0% weight loss (82 to 75%) in the range from 200 to 450 °C which may be correlated to the combination of the SiO 2 NP surface functional groups degrading (mainly -OC 2 H 5 and -OH (Zhuravlev 2000)) amounting to 3% (based on the SiO 2 NP TGA, 88 to 85%) and the amount of the surface bound coumarin 3 of 4% (7 to 3%). This allows for a coarse comparison between the loading capacity of the SiO 2 -3 nanohybrid and their fluorescence versus the use of the standalone coumarin 3. This means when 7.5 mg/L SiO 2 -3 nanohybrid is used in the experiments, the concentration of 3 will be about 0.8 μM but instead of an even distribution of coumarin 3 in the whole solution, now it would be adsorbed and concentrated on the surface of SiO 2 NPs. This localized 3 concentration increase would then expectedly lead to changes in optical properties. Zeta potential of the SiO 2 NPs was − 30.5 mV (7.5 mg/L in HEPES buffer, 20 mM, pH 5) due to the abundant and deprotonated surface silanol groups at testing pH of 5 when prepared using TEOS (Table 2) (Qiao et al. 2018). Upon surface adsorption of the coumarins to form SiO 2 -3 nanohybrid, the hydrodynamic size of the particles increased from 114.2 to 122.1 nm, and the zeta potential changed from -30.5 to -28.4 mV, a slight increase. Upon the addition of Cu(II) (at the final concentration of 40 μM) to the SiO 2 -3 nanohybrid solution, the zeta potential dropped to − 22.8 mV. As for comparison, when SiO 2 NPs were treated with Cu(II) at the same concentration, the zeta potential displayed an even more obvious change to − 13.2 mV, suggesting significant surface adsorption of Cu(II) on the surface of SiO 2 as expected due to the presence of negatively charged hydroxylate as well as -OH surface groups (Chen et al. 2021). Cu(II) adsorption by SiO 2 surface as a reason behind improved sensing performance as it is another evidence of Cu(II) adsorbing onto SiO 2 .

Fluorescence study of coumarin 3 in the absence and presence of Cu(II)
The pH and concentration dependence of the fluorescence properties of the coumarin 3 were studied in aqueous solution to find the optimum conditions for their application for Cu(II) sensing. Fluorescence intensity of the coumarin 3 was studied across the pH range from 3 to 11 (Fig. S11 (a)) showing that, in the range of pH 5-8, the molecule exhibited the highest intensity. However, when fluorescence intensities at concentrations of 0-80 µM dissolved coumarin 3 were recorded (Fig. S11(b)), no significant differences were observed above 40 µM.
Fluorescence study of SiO 2 -3 nanohybrid in the absence and presence of Cu(II) Fluorescence of the SiO 2 NPs was negligible in HEPES buffer (20 mM, pH = 5, λ ex = 435 nm) (Fig. S12). To find the optimal working parameters, the fluorescence intensity of the SiO 2 -3 nanohybrid was recorded at the excitation wavelength of 3, 435 nm, and across the pH range of 3-11 ( Fig. S13(a)). The highest fluorescence emission intensity occurred at pH 4. In addition, various concentrations of the SiO 2 -3 nanohybrid (50 to 0.5 mg/L) ( Fig. S13(b) and (c)) in HEPES buffer were considered. The fluorescence emission intensity of the SiO 2 -3 nanohybrid was highest at 7.5 mg/L.
Although the greatest fluorescence intensity was observed at pH 4, at this pH, SiO 2 -3 nanohybrid does not respond to Cu(II) (Fig. S14). However, at pH 4.5 or 5, the fluorescence quenching of SiO 2 -3 nanohybrid by Cu(II) was more significant. Thus, pH 5 was selected as the working pH in the following studies. The pH dependence of the fluorescence quenching of coumarin 3 by Cu(II) was also assessed (Fig. S15), and similarly, the extent of fluorescence quenching was greater with increasing basicity of the solution in the presence of equal amounts of Cu(II) added.
Fluorescence titration of both 3 (40 μM) and SiO 2 -3 nanohybrid (7.5 mg/L) with CuCl 2 (0-260 µM) in HEPES buffer was conducted (Fig. 5). The SiO 2 -3 nanohybrid displayed linear correlation between fluorescence quenching and Cu(II) concentration over the 0.033-55 and 55-260 μM ranges (R 2 = 0.98), though the sensitivities (slope) (inset of Fig. 5) were different for each range. This is a valuable extension of the sensing range when compared to coumarin 3, especially as the regulatory limit for Cu(II) in drinking water is specified by the Australian Drinking Water Table 2 Hydrodynamic size (nm) and zeta potential (mV) of SiO 2 NPs (7.5 mg/L) before and after surface interaction with coumarin 3 (7.5 mg/L) in HEPES buffer (20 mM, pH = 5), and in the absence or presence of Cu(II)(40 μM)

Samples
Size ( Guidelines at 30 µM (NHMRC N 2011) and 10 mg/ kg in soil. The LOD was 0.033 μM, tenfold lower than that of coumarin 3 at 0.48 μM.
Interference study of SiO 2 -3 nanohybrid To study the influence of other metal ions on the Cu(II)/SiO 2 -3 nanohybrid interaction, interference study was conducted with a variety of metal ions applied at an excess concentration (200 μM each) to SiO 2 -3 nanohybrid (7.5 mg/L) and Cu(II) (80 μM). Such additional ions were Al(III), Ca(II), Cd(II), Co(II), Fe(III), Mn(II), Pb(II), Hg(II), Ni(II), Zn(II), Mg(II), Cu(I), and Fe(II). Fluorescence emission spectra were examined with and without the addition of Cu(II) (Fig. 6). The measurements indicated no significant interference apart from Fe(III), which means the selectivity of 3 was successfully preserved. Cu(II) adsorption by SiO 2 surface as a reason behind improved sensing performance ICP-OES was employed to determine the amount of Cu(II) adsorbed on both SiO 2 NPs and SiO 2 -3 nanohybrid (both presenting (-OH/-O − surface groups) using the back-titration procedure described in the "Experiment" section (Table S1). SiO 2 NPs and SiO 2 -3 nanohybrid were used at 15 mg/L, respectively, with Cu(II) added at 9.4 mg/L. ICP-OES results showed the residual Cu(II) in the mother liquor as 7.0 mg/L in both cases. This means the SiO 2 NP has a 2.4 mg/L Cu(II) adsorption. This is slightly lowered when coumarin 3 is adsorbed in SiO 2 -3 nanohybrid at a ~ 4% (weight ratio). Hence, interaction of Cu(II) with both the coumarin 3 and the SiO 2 NP surface results in the less steep slope/increased sensitivity of the sensing observed upon addition of Cu(II) when compared to that of the coumarin 3 itself (Fig. 5). This results in an extended linear detection range of SiO 2 -3 nanohybrid for Cu(II) sensing in the range of 0.033-260 μM. Moreover, approximately tenfold less amount of coumarin 3 was present in the SiO 2 -3 nanohybrid versus when coumarin 3 used alone. This allows for less organic dye to be used while achieving improved applicable sensing range.
Application of SiO 2 -3 nanohybrid on real soil samples Roadside surface soil samples were collected in the western parts of Melbourne, Victoria, Australia (Li et al. 2016). The soil samples were dispersed in water and digested while shaken for 7 days to ensure maximal dissolution of all the soluble metals. The soil samples were analyzed for their Cu content by the industry standard quantification: inductively coupled plasma mass spectrometry (ICP-MS). Next, fluorescence titration using 15 mg/mL SiO 2 -3 nanohybrid (1.5 mL) was conducted at pH 5. Then the soil extracts (1.5 mL) were then merged with the SiO 2 -3 nanohybrid  solution resulting in a concentration of 7.5 mg/ mL. The background fluorescence of the soil extracts is shown on Fig. S16. Fluorescence intensities were converted into Cu(II) concentrations based on the titration curves (Fig. 5). The results obtained were then compared against the ICP-MS results. The recovery of Cu(II), directly from the water extract without further sample preparation, was between 105 and 94% for all three soil samples tested (Table 3) (Qian et al. 2019).

Conclusions
SiO 2 -3 nanohybrid was designed, synthesized, characterized, and evaluated for their use for Cu(II) sensing in aqueous media while omitting any organic solvent carrier (required for 3 used alone). SiO 2 -3 nanohybrid exhibited equal selectivity towards Cu(II) in the presence of other multivalent metal ions compared to coumarin 3. The sensing signal of SiO 2 -3 nanohybrid versus Cu(II) concentration was linear in the range of 0.033-260 μM, which is in good correspondence with the relevant WHO and Australia standards for Cu(II) content in soil and water, and even intra-cellular imaging. This linear sensing range of the nanohybrid was significantly extended when compared to that of 3 (0-20 μM) from μM to nM level. The limit of detection in aqueous media was 0.033 μM, a more than tenfold improvement from LOD 0.48 μM of 3. SiO 2 -3 nanohybrid also outperforms some of the previously reported surface-modified SiO 2 NP systems used for Cu(II) sensing (Table S2). The impact of this work is the demonstrated (i) improvement of Cu(II) sensing performance; (ii) use of tenfold reduction amounts of organic molecular probe used when loaded onto the SiO 2 nanoparticles while achieving superior sensing performance; and (iii) omission of organic solvents, when SiO 2 -3 nanohybrid was used and compared to 3 alone. This approach offers reduced cost of synthesis and potentially the amount of (in some cases toxic) organic dye used. SiO 2 -3 nanohybrid was validated on real soil samples against the industry standard ICP-MS technique and showed satisfactory accuracy. Therefore, the concept of SiO 2 -3 nanohybrid offers a convenient and customizable strategy for the use of organic sensing probes in fully aqueous media with the added benefit of improved optical sensing properties, such as linear sensing range and detection limit.