**1. Introduction**

Cu(II) is an active producer of oxidative stress for both plants [1–3] and animals [4]. Human uptake of Cu is usually in the range of 0.6–1.6 mg per day [5]. Excess uptake of Cu in human beings is related to cancer and aging [5]. It is also reported to be related to diseases of the nervous system such as Alzheimer's, Menkes, and Wilson diseases [6,7]. Because of its biological e ffects, control of Cu contamination is an important aspect of environmental protection.

The biotic ligand model (BLM) considers the interactions of all parameters in a natural system to predict the bioavailability of metal ions [8,9]. Bioavailable Cu concentrations predicted by the BLM correlate well with measured Cu LC50s. Total Cu does not correlate well with actual toxicity [9]. However, the BLM is based on an indirect measurement of bioavailable Cu(II), that is, it is based on measurements of organic carbon, pH, other metal ions and several other parameters. At present there is no viable method for measuring bioavailable Cu(II) directly.

Several studies report ligands that change fluorescence when they bind Cu(II). These can potentially be used to measure bioavailable Cu(II). There are some fluorogenic ligands that have increased fluorescence when they bind Cu(II) [10–13]. However, some of them can only be applied in organic solvents such as THF [10] and acetonitrile [11,14], which are not appropriate for the detection of bioavailable Cu(II) in water systems. Low sensitivity, long response times, poor selectivity and

ligands with inappropriate Cu(II)-complex formation constants are other problems that render reported ligands unsuitable for Cu(II) monitoring.

Many other fluorescent sensors have decreased or "turn off" fluorescence upon Cu(II) binding due to Cu quenching of the fluorogenic ligands [15–17]. The strategy of developing a fluorogenic ligand that is capable of measuring bioavailable Cu(II) has ye<sup>t</sup> to succeed. Furthermore, even if successful, it would only be applicable to Cu(II).

We prefer to base detection on metal ion induced changes in a water-soluble polymer conformation detected via fluorescence. This approach separates the fluorophore from the metal, rendering it less subject to metal ion quenching, a frequent issue with Cu(II). Furthermore, the selectivity of this approach can be modified by changing the ligand while keeping the rest of the indicator platform. Du et al. synthesized a ratiometric fluorescent Cu(II) indicator platform [18]. Cu(II) binding neutralizes the charge on the ligand, which causes poly(N-isopropylacrylamide) (pNIPAM) to change conformation. This in turn affects the environment of a dansyl comonomer [18]. The indicator developed by Yao et al. [19] is based on fluorescence resonance energy transfer (FRET) [20]. Cu(II) binding introduces positive charge repulsion which separates copolymer strands disrupting FRET. However, neither of these systems has the required sensitivity for environmental Cu(II) measurements. In Du et al.'s indicator, the fluorophore utilized is not that efficient, and for Yao et al.'s indicator, the limit of detection is not low enough. Osambo et al. demonstrated an indicator platform based on changes in FRET accompanying metal ion induced nanoparticle swelling [21]. However, the excitation wavelength is too short to be practical. We also synthesized ratiometric indicators with both donor and acceptor fluorophores on the same polymer chain, but the signal changes with time due to slow polymer untangling. Therefore, our goal is to demonstrate an indicator platform that is both stable and sensitive, and involves wavelengths in the visible spectrum.

The indicator discussed in this paper is based on cross-linked pNIPAM nanoparticles. A negatively charged ligand is used to make the nanoparticle swell in the absence of metal ions. Addition of metal ions neutralizes the negative charge causing the nanoparticle to shrink. This results in a change in fluorescein concentration per unit volume. The fluorescence signal of fluorescein decreases with increasing concentration due to self-quenching when the concentration is above a critical concentration [22]. Our approach is illustrated schematically in Figure 1. However, nanoparticles alone can undergo self-agglomeration, which affects the volume change, and may also block the Cu(II) binding sites. In order to avoid agglomeration, the nanoparticles were embedded in a polyacrylamide gel. The PA gel increases the stability of the single nanoparticles. This approach makes it possible to synthesize particles with a wider range of sizes.

**Figure 1.** Sensing mechanism of self-quenching poly(N-isopropylacrylamide) (pNIPAM) nanoparticles. Because of the negative charges on the ligand, the nanoparticles swell. When Cu(II) bind to the ligand, the charge neutralization results in less swelling of the nanoparticles, hence the fluorescence intensity also decreases.

The data we show here are for Cu(II). However, the approach is general because binding of other metal ions will also change the charge on the polymer backbone leading to swelling or shrinking depending on whether the charge increases or decreases. Thus, the indicator platform we demonstrate here is applicable to other metal ions depending on the particular ligand that is incorporated into the polymer.

## **2. Experimental Materials**

Materials: Sodium dodecyl sulfate (SDS), *N*-isopropyl acrylamide (NIPAM), *<sup>N</sup>*,*N*-Methylenebisacrylamide (BIS), fluorescein *o*-acrylate, potassium persulfate (KPS), acrylamide, ammonium persulfate (APS), *N*,*N*,*N* ,*N*-Tetramethylethylenediamine (TEMED), Copper (II) nitrate trihydrate, and Zinc (II) nitrate hexahydrate were purchased from Sigma-Aldrich. Aqueous solutions were prepared from doubly distilled water from a Corning Mega-Pure distillation apparatus. Dialysis tubing with a molecular weight cut-off (MWCO) of 3.5–5 kDa was purchased from Spectrum Labs.

Equipment: Fluorescence responses were measured using the scan mode on a Varian Cary Eclipse fluorometer equipped with a Peltier thermostatted single cell holder. Scanning Electron Microscopy (SEM) was performed on a Tescan Lyra3 GMU Focused Ion Beam (FIB) SEM. A Branson model 1800 sonicator was used for reagen<sup>t</sup> dissolution and sonication. A Buchi RE111 Rotavapor was used to evaporate solvents. Separation of precipitated polymer from the solution was performed on a Beckman GP centrifuge (8000 rpm) or an Eppendorf centrifuge 5415 C (14,000 rpm). A FreeZone Plus 2.5 Liter Cascade Benchtop Freeze Dry System was used to lyophilize samples.
