**1. Introduction**

Electrochemiluminescence (ECL), also called as electrogenerated chemiluminescence, is light emission from the excited species generated through multiple redox reactions on an electrode [1,2]. ECL has many advantages as a platform of sensing materials. Owing to unnecessary light excitation, background noise in the luminescence detection can be drastically reduced by removing scattering of incident light. Compared to chemiluminescence which is produced through chemical reactions, it is relatively easy to achieve temporal and spatial control for emission generation. Therefore, the ECL techniques are recognized as a powerful analytical tool and are used in various areas, such as immunoassay, genosensors, and for monitoring small molecules [1–6]. Although a variety of substances can potentially work as a source of ECL, ruthenium(II) (Ru(II)) complexes, especially tris(2,2-bipyridyl)ruthenium(II) (Ru(bpy)3<sup>2</sup>+), are conventionally used because of many superior properties, such as high chemical stability, recyclability, redox properties, and long lifetime in the excited state [2–4]. Meanwhile, oxygen is well known to be a crucial quencher to the photo-excited state of Ru(II) complexes [7–9], followed by the reduction of the ECL intensity [10–12]. Thereby, pretreatments for deoxygenation with the analytical samples are essential before conventional ECL detection. Hence, improvement of oxygen resistance of the ECL system is strongly needed not only for improving accuracy and sensitivity in detection but also for extending applicability to future sensing technologies such as real-time monitoring with the native samples.

Polyhedral oligomeric silsesquioxane (POSS) attracted attention as an "element-block" which is the minimum unit containing heteroatoms for constructing functional materials [13,14]. We report unique and useful properties of POSS-containing hydrophilic materials [15–19]. In aqueous media, it was shown that the water-soluble POSS-containing materials can e fficiently capture various types of gues<sup>t</sup> molecules in the distinct hydrophobic spaces created around the POSS units and are applicable for sensors [20–24]. It is proposed that the cubic core is favorable for adsorbing various gues<sup>t</sup> molecules with hydrophobic interaction. In particular, it is shown that the captured molecules in the POSS-core dendrimer are protected from photobleaching. It should be noted that stimuli responsiveness is compatibly observed from the captured gues<sup>t</sup> molecules, although these molecules are isolated by POSS. Because of the steric structure of POSS, accessibility of reactive oxygen species which are promised to induce degradation can be lowered. Meanwhile, energy and electron transferring is acceptable through the POSS sca ffold. Owing to these characters, environment-responsive photon upconversion including multiple energy transferring steps [23] and photo-triggered redox reactions [24] are accomplished in the POSS-containing materials.

Inspired by the enhancement of environmental resistance of POSS without losses of stimuli responsiveness, we presume that oxygen resistance in ECL systems might be reinforced. On the basis of this assumption, we synthesize the modified POSS tethered to the tris(2,2-bipyridyl)ruthenium(II) complex (Ru-POSS) and evaluate the influence of dissolved oxygen on the ECL properties in the aqueous mixture system containing tripropylamine (TPrA, coreactant). Initially it is shown that ECL intensity and electric current can be enlarged by the connection to POSS. Moreover, from the comparison studies, it is clearly indicated that ECL quenching by oxygen is suppressed in the Ru-POSS/TPrA system. We elucidate how this phenomenon could originate from the improvement of TPrA oxidation by Ru-POSS adsorption onto the ITO electrode. Finally, we also accomplish the ECL-based detection for the water pollutant without the degassing pretreatment. The oxygen-resistance enhancement and target detection which tend to show the trade-o ff relationship are compatible by employing the POSS element-block.

#### **2. Results and Discussion**

#### *2.1. Characteristics of Ru-POSS*

The chemical structures and preparation of the Ru(II)-containing materials involving Ru-POSS and the model compound, Ru-Model are shown in Scheme 1. As is often the case with luminescent dyes, concentration quenching is occasionally induced through non-specific intermolecular interaction in the accumulation of the chromophores. To avoid reduction of sensitivity, we designed Ru-POSS to have the single Ru(II) complex. To obtain the target material, we used octa-substituted ammonium POSS (Amino-POSS) as a sca ffold and connected it to the bipyridine ligand for complexation with the Ru(II) ion [24]. After complexation with the commodity reaction condition, the introduction ratio of the Ru(II) complex to the POSS unit in Ru-POSS was estimated from the integration ratios of the signal peaks from the side chains in the 1H NMR spectrum. Purification with Ru-POSS was only precipitation and Ru-POSS was obtained as mixtures containing variable numbers of the Ru complex. The structures of synthesized compounds were confirmed by 1H, 13C and 29Si NMR spectroscopies and high-resolution mass spectrometry (HRMS). The detailed procedures are shown in the Supporting Information. The structures of the compounds used, such as Ru(bpy)3 2+, Ru(bpy)2(dmbpy)<sup>2</sup>+ (dmbpy = dimethylbipyridine) and tripropyl amine (TPrA), are listed in Figure S1.

**Scheme 1.** Syntheses of the Ru-POSS and the Ru-Model.

#### *2.2. ECL Properties of Ru(II)-Containing Materials*

Optical measurements were performed to evaluate the influence on the electronic structure of the Ru complex moiety by the connection with POSS (Figure S2). ECL and photoluminescence (PL) spectra of the materials are shown in Figure 1 and Table 1. All compounds containing Ru(II) exhibited the emission bands at similar positions, indicating that ECL originated from the metal-to-ligand charge transfer transition in the triplet excited state (3MLCT) in the Ru(II) complex moiety [2–4]. By subtracting TPrA or Ru(bpy)3<sup>2</sup>+, critical reduction of electric currents was obtained (Figure S3), supporting the conclusion that ECL is produced from the redox cycles including the Ru(II) complex and TPrA through the several proposed mechanisms (Scheme S1). The increase in the Ru(bpy)3<sup>2</sup>+ concentration was minimally influenced on the enhancement of ECL intensity (Figure S4). Concentration quenching could occur under condensed conditions. Slight bathochromic shifts were observed in UV−vis absorption and ECL spectra of the solutions containing Ru-POSS and Ru-Model compared to the ECL spectrum of Ru(bpy)3<sup>2</sup>+. It is likely that a substituent effect is responsible for these negligible changes [25]. From these results it was shown that electronic structure at the Ru(II) complex moiety is hardly influenced by connecting to the POSS unit.

**Figure 1.** ECL (jagged line) and PL spectra (smooth line) of the Ru(II)-containing materials. The spectra of Ru(bpy)3<sup>2</sup>+ and Ru-Model are shifted in the *y*-axis direction for clarity.


**Table 1.** Luminescent properties of the Ru(II)-containing materials *a*.

*a* Based on the concentration of the Ru(II) complex unit (1.0 × 10−<sup>5</sup> M). *b* Excited at λabs\_MLCT. *c* Calculated as an absolute value. *d* Measured with the excitation light at 375 nm. *e* Determined at the peak value in CV experiments.

The cyclic voltammograms and corresponding ECL profiles in the TPrA co-reactant system are shown in Figure 2. The POSS unit in Ru-POSS induced significant changes in electric and optical properties. There are two advantageous issues resulting from the connection of the Ru(II) complex to POSS. First, it was found via Ru-POSS that ECL appeared in the lower voltage region. HOMO energy levels were estimated from the onset values (*E*onset) in the oxidative wave of the CV curves (Figure S5). Obviously, the *E*onset of Ru-POSS decreased compared to the Ru-Model, as well as Ru(bpy)3<sup>2</sup>+. Although it was reported that the substituent effect can decrease *E*onset [26], the degree of change was several times larger by the connection to POSS than those in the previous work. In our ECL system, another pathway for decreasing *E*onset could exist. Additionally, even in the absence of TPrA, lower *E*onset values were obtained from the Ru(II)-containing materials except for Ru(bpy)3<sup>2</sup>+ (Figure S6), proposing that the lowering value of *E*onset might be induced by changes in molecular morphology on the electrode. The discussion will be made in a later section. Second, a larger anodic current and ECL intensity were obtained from Ru-POSS than those from Ru(bpy)3<sup>2</sup>+ and Ru-Model. To understand the roles of POSS in these changes, further experiments were executed.

**Figure 2.** Cyclic voltammograms (solid lines) and corresponding ECL profiles (dashed lines) of 0.1 mM Ru(II) complexes and 100 mM TPrA in 0.20 M PBS buffer (pH 8.8) with an ITO electrode at a scan rate of 100 mV/s (*n* = 3). The concentration of Ru-POSS was based on the Ru(II) complex unit.

In the bulk electrolysis (BE) at 1.2 V vs. Ag/AgCl, it was also observed that a larger current was able to be generated from Ru-POSS than those of Ru(bpy)3<sup>2</sup>+ and Ru-Model (Figure S7). In particular, the initial non-faradaic current was one order magnitude larger. To explain these behaviors, including lower *E*onset, we initially investigated the possibility that Ru-POSS could adsorb at the electrode surfaces and accelerate the redox reactions. To evaluate the validity of this speculation, we examined BE with Ru(bpy)3<sup>2</sup>+ in the presence of Amino-POSS under the same condition as above (Figure S8). It was clearly indicated that the connection to POSS was essential for obtaining a larger ECL. Next, the electrodes modified by immobilization of Ru(II) complexes were prepared, and their properties were examined. The ITO electrode was immersed in the ECL reaction solution containing the series of Ru(II) complexes, such as Ru-POSS, Ru(bpy)3<sup>2</sup>+ and Ru-Model, rinsed with deionized water and dried in vacuo. Then, the BE and ECL properties were evaluated in the buffer solution containing only TPrA without adding Ru(II)-containing molecular species in the solution (Figures S9–S11). The reduction in current intensity and disappearance of ECL were observed in the ITO electrodes treated with Ru(bpy)3<sup>2</sup>+ and Ru-Model, while a large current and ECL were still detected in the Ru-POSS-immersed ITO electrode. This data clearly indicated that the adsorption of Ru-POSS occurred on the ITO electrode. It has been reported that ammonium groups can form tight interaction to the ITO electrode from the studies regarding self-assembling monolayers [27–29], poly-l-lysine [30–32], and cytochrome C [33]. It is likely that the multiple ammonium groups in Ru-POSS play a critical role in adsorption to the electrode. Moreover, from the adsorption experiments under variable pH conditions, enhancement of ECL was observed by using the modified ITO electrode immersed with Ru-POSS solution in an acidic condition (pH 5.2), supporting the conclusion that the ammonium-ITO interaction is responsible for anchoring Ru-POSS onto the electrode surface (Figure S12). The assembly of functional groups into the compact space as well as the high symmetry of POSS might be favorable for tight interaction.

Regarding the lowering *E*onset, it was reported that modification of ITO electrodes with CNT [34], poly(amidoamine) (PAMAM) dendrimer [35], Au nanoparticles [36], and thin film consisting of vertically aligned silica mesochannels [37] lowered the onset potentials and enhanced ECL intensities by facilitating the redox cycles at the electrode surfaces. Similar to these modifications, it is proposed that Ru-POSS could promote the redox cycles for the generation of ECL. Bard et al. reported about the influence of molecular morphology on intensity. They claimed that some degree of distance would be needed for obtaining emission from the Ru(II) complex adsorbed to electrodes, otherwise direct quench by the electrode would occur [38]. From our experiments, although POSS could induce adsorption of Ru(II) species on the electrode, larger ECL was obtained. It is implied that steric hindrance of the silica cube might prevent quenching by the electrode.

#### *2.3. Resistance to Oxygen Quenching*

Oxygen resistance of this ECL system was evaluated (Figure 3 and Table 2). To examine the influence of oxygen on the luminescent and electric properties, the dissolved oxygen (DO) levels in the samples were tuned by argon (DO: less than 1 mg/mL) and oxygen bubbling (ca. 20 mg/mL) before monitoring. As readily expected, the aerated sample containing Ru-complex showed critical annihilation of ECL prepared by oxygen bubbling, whereas significant influence were hardly observed in the electric currents (Figure S13). It is well known that DO readily facilitates non-radiation decay of the triplet-excited state, resulting in emission annihilation. In contrast, it was shown that the intensity level from the hypoxic sample containing Ru-POSS was maintained at the higher DO level. These data clearly indicated that the connection of Ru-complex to POSS was responsible for improving oxygen resistance in the ECL system.

**Figure 3.** Time courses of ECL with (**a**) Ru-POSS, (**b**) Ru(bpy)3 2+ and (**c**) Ru-Model in BE with 0.1 mM Ru(II) complexes and 100 mM TPrA in 0.20 M PBS bu ffer (pH 8.8) with Ar (solid line) and O2 bubbling (dashed line) with an ITO electrode at 1.2 V vs. Ag/AgCl. The concentration of Ru-POSS was based on a Ru(II) complex unit.


**Table 2.** Residual rates of maximum ECL intensities and onset increase rates in BE in the aerated solutions compared to the hypoxic ones.

*m*: Slope of onset intensity curve.

To obtain further insight regarding the detailed mechanism and the role of POSS in enhanced oxygen resistance, we performed measurements under various conditions. It was shown that the connection of the Ru(II) complex was necessary for expressing oxygen resistance in the former section (Figure S8 and Table S1). From the PL measurement with the Ru-POSS sample under the aerated condition, it was found that emission annihilation was induced (Figure S14), meaning that oxygen access for the Ru(II) complex moiety could be maintained even in the presence of POSS. Therefore, it was proposed that the rate of TPrA oxidation could be significantly enhanced in the Ru-POSS system. The pH e ffect on ECL of Ru-POSS was initially investigated (Figure S15 and Table S2). The ECL intensities of all Ru(II)-containing materials drastically decreased under acidic conditions. It should be emphasized that even in the Ru-POSS system, which had high resistance against DO, a lower intensity was shown at pH 5.2 than that at pH 8.8. According to previous literature, deprotonation in TPrA followed by oxidation of TPrA [39–41] could be suppressed under acidic conditions. Therefore, the redox cycle for generating ECL should be inhibited. It is already known that there are several pathways in the ECL mechanism of Ru(bpy)3 <sup>2</sup>+/TPrA system (Scheme S1) [12,40–44]. In these mechanisms, the TPrA oxidation should be the critical step before presenting luminescence. Furthermore, it was reported that the excess amount of TPrA radicals (TPrA•) generated by the TPrA oxidation can contribute to reducing of the amount of DO near the ECL reaction layer and subsequently can suppress ECL quenching [12]. In summary, the enhancement of the TPrA oxidation by Ru-POSS might be responsible for improving oxygen resistance in the ECL system as well as ECL intensity.

#### *2.4. Analysis of Antibiotics without Deoxygenation*

Because ECL is usually obtained from the triplet-excited state, most conventional ECL systems always su ffer from emission quenching by DO in the pristine sample. Therefore, degassing for lowering DO levels in the sample was previously needed in practical usage. To demonstrate the advantage of oxygen-resistant ECL based on Ru-POSS, we finally performed the detection without degassing as a pre-treatment. By evaluating the decrease in ECL intensity, the target concentration was determined [45–49]. By adding oxytetracycline (OTC), which is an anti-bacterial agen<sup>t</sup> for domestic animals that often induces water pollution near livestock barns, to the samples, the changes in ECL intensity were monitored. The OTC aqueous solution was added to the bu ffer containing 0.1 mM Ru(II) complex and 0.1 M TPrA with stirring and then ECL measurements in BE (1.2 V vs. Ag/AgCl) of the mixtures were taken (Figure 4). Apparently, ECL intensity from the Ru(bpy)3 <sup>2</sup>+/TPrA system was too small to discriminate the target, meanwhile critical decrease in emission intensity was detected from the Ru-POSS sample, meaning that the existence of OTC was able to be detected even in the aerobic sample. In particular, the linear relationship between Δ*I*ECL and the concentration in the range from 1.3 × 10−<sup>5</sup> to 2.0 × 10−<sup>4</sup> M ( *R*<sup>2</sup> = 0.993) was obtained when using Ru-POSS. The limit of detection was 9.8 × 10−<sup>6</sup> M (S/N = 3). It was suggested that ECL quenching may be caused by oxidized OTC at the electrode, which is a benzoquinone derivative [45]. Benzoquinone groups are well known to have the ability to quench ECL [48,49]. The decay of the excited state of the Ru(II) complex and TPrA intermediate-radical quenching were induced by benzoquinone and the latter pathway was predominant when TPrA direct oxidation occurred [49]. Regarding the detection limit, it has been reported that the OTC concentrations in wastewater are in the range between 10−<sup>6</sup> and 10−<sup>9</sup> M

order [47]. The detection limit of our system was not so high compared with the other ECL system [45], however, our method could be directly applicable for detecting serious contaminated conditions.

**Figure 4.** Time courses of ECL with (**a**) Ru(bpy)3<sup>2</sup>+ and (**b**) Ru-POSS in BE of 0.1 mM Ru(II) complexes and 100 mM TPrA and various concentrations of OTC in 0.20 M PBS buffer (pH 8.8) stirred for 20 min with an ITO electrode at 1.2 V vs. Ag/AgCl. (**c**) Standard curve prepared from the data from (**b**) (*I*ECL and *I*ECL\_0 represent the maximum ECL intensity in the time course of BE with and without OTC, respectively).
