Aqueous Singlet Oxygen Sensitization of Porphyrin-Embedded Silica Particles with Long-Term Stability

Abstract

Aqueous singlet oxygen (¹O₂) sensitization is of high interest due to its wide application in bio-imaging and photodynamic therapy. For organic photosensitizers like porphyrin derivatives, surfactant-assisted micelles have been intensively explored for dispersing hydrophobic sensitizers in aqueous phase; however, they can suffer from poor long-term stability. In this work, palladium octaethylporphyrin (PdOEP)-embedded silica particles were prepared with assistance from Tween micelles, and their corresponding application in aqueous ¹O₂ sensitization was explored. With assistance from Tween 80 at a >3 mg/mL concentration, superior (>95%) solubilization of PdOEP was observed in aqueous solution, leading to a high ¹O₂ quantum yield (Φ_Δ ≈ 93%). By optimizing the synthesis conditions, >95% of micellar PdOEP was embedded into silica particles, exhibiting comparable Φ_Δ (up to 70%) to micellar systems by effectively suppressing PdOEP aggregation in particles. The PdOEP-embedded silica particles exhibited dramatically enhanced long-term stability (more than one year) compared to corresponding micelles with a half-life of ~38 days. In addition, aqueous ¹O₂ sensitization by PdOEP-embedded silica particles was demonstrated upon two-photon excitation in a near-infrared regime (λ_ex = 1030 nm), highlighting the great potential of this method for future biological applications.

1. Introduction

The efficient generation of singlet oxygen (¹O₂) in aqueous media holds critical importance for biological applications such as photodynamic therapy and bioimaging [1,2,3,4]. Although photosensitizers such as porphyrin derivatives exhibit exceptional ¹O₂ sensitization capabilities, they require surfactant micelles for dispersion in aqueous media due to their inherent hydrophobicity [5,6,7,8], and the poor stability of the micellar structures in solution hinders the long-term storage of dispersed photosensitizers [9,10,11]. To overcome the poor stability of soft surfactant micelles during long-term storage, silica has been deposited on micelles to enhance their thermodynamic stability [12,13,14]. Based on this approach, we previously developed a robust method for incorporating organic dyes and hydrophobic drugs into silica particles [15,16]. Initially, we solubilized hydrophobic molecular guests within surfactant micelles. Subsequently, we deposited a silica network onto the micelle surfaces through hydrolysis–condensation with tetraethyl orthosilicate (TEOS) as a precursor, ultimately yielding silica particles embedded with hydrophobic guests [17,18,19]. This strategy shows promise for achieving long-term, stable aqueous ¹O₂ sensitization in micellar systems containing embedded photosensitizers. During the silica deposition process through the hydrolysis and condensation of siloxanes [20,21], the micellar packing structure within the as-prepared silica particles was strongly affected by the reaction parameters. The packing structure governs the aggregation state of the particles and the optical properties of embedded guests [22,23,24,25]. However, systematic investigations into the colloidal stability of the silica particles and the aggregation behavior of the embedded photosensitizers remain scarce, thereby constraining their performance in ¹O₂ sensitization [26,27,28].

In this study, palladium octaethylporphyrin (PdOEP) was selected as a model photosensitizer, due to its high ¹O₂ quantum yield (Φ_Δ) in organic solvents [29], and embedded into silica particles via a Tween micelle-assisted approach. Furthermore, as a neutral molecule lacking specific interactions with surfactants, PdOEP represents a versatile model for hydrophobic photosensitizers. At surfactant concentrations >3 mg/mL in micellar systems, Tween 80 achieved superior solubilization of PdOEP (>95%) in aqueous solution compared to Tween 20. Although PdOEP demonstrated good solubilization and a high Φ_Δ (~93%) in Tween 80 micelles, it showed limited stability with a half-life of approximately 5 weeks. After silica deposition, over 95% of the solubilized PdOEP was successfully embedded within the silica matrix, resulting in substantially enhanced stability, with the half-life exceeding one year. After optimizing the reaction conditions, PdOEP embedded in silica particles achieved a Φ_Δ > 70%, comparable to that in micellar systems. In addition, we demonstrated ¹O₂ generation under 1030 nm two-photon excitation, highlighting the potential of this system for deep-tissue applications.

2. Results

2.1. Preparation and Characterization of Particles

PdOEP-embedded silica particles were prepared through a typical two-step approach [30], as illustrated in Figure 1a. Initially, hydrophobic PdOEP was solubilized in aqueous solution through co-dissolution with Tween surfactants in toluene, followed by solvent removal and aqueous redispersion. This process enabled the spontaneous formation of Tween micelles, which solubilized hydrophobic PdOEP molecules through hydrophobic interactions.

The solubilization capacity of surfactant micelles facilitates the embedding of hydrophobic PdOEP into silica particles. Tween surfactants are a class of nonionic surfactants chemically classified as polysorbates (their chemical structure is shown in Supplementary Materials, Figure S1). Tween surfactants form micelles containing hydrophobic cores that predominantly comprise alkyl chains from various fatty acids [31,32]. This structural variation leads to differential solubilization capacities among Tween micelles, with the aliphatic chain length strongly affecting the solubilization capacity [33]. Thus, two Tween surfactants with distinct alkyl chain configurations, Tween 80 (C16) and Tween 20 (C11), were selected for comparative investigation in this study.

As expected, the Tween 80 and Tween 20 surfactants demonstrated distinct solubilization capacities. Ultraviolet–visible (UV-Vis) spectroscopy at 546 nm revealed the superior performance of Tween 80, which achieved a solubilization capacity (c/c₀) exceeding 95% at a concentration of 3 mg/mL. In contrast, Tween 20 required a concentration of 10 mg/mL to achieve similar efficiency (Figure 1b). Subsequent silica coating via TEOS hydrolysis–condensation under NaF catalysis enabled structural stabilization. Following centrifugal separation, PdOEP-embedded silica particles were obtained. Centrifugal separation yielded PdOEP-embedded silica particles. The embedding efficiency, denoted as ω and defined as the proportion of PdOEP solubilized within the silica particles, was determined spectrophotometrically by measuring the absorption of the supernate.

The UV-Vis spectra (Figure S2) show a concentration-dependent decrease in ω for both Tween systems (Figure 1c). At 10 mg/mL, the ω values reached 70% and 60% for Tween 80 and Tween 20, respectively. The decrease in ω is attributed to the increased residual surfactant concentration in the supernate at higher initial surfactant concentrations [33]. This excess surfactant enhances the solubilization capacity of the supernate, resulting in a higher concentration of solubilized PdOEP and consequently a lower ω value [34]. As the surfactant concentration increased, the micellar solubilization capacity increased, while the embedding efficiency of silica particles decreased. To clearly demonstrate the relationship between the amount of PdOEP embedded in the silica particles and the surfactant concentration, we introduced the embedding capacity (C), defined by Equation (1):

$$C={\displaystyle \frac{c}{c_0}}\times \mathsf{\omega }.$$

(1)

As demonstrated in Figure 1d, embedding capacity (C) exhibited a parabolic dependence on surfactant concentration. Tween 80 achieved a maximum C of 92% at 3 mg/mL, surpassing the maximum capacity attained by Tween 20 at 8 mg/mL (80%). This substantial performance difference confirms the structural advantages of the longer-chain Tween 80 for effective PdOEP embedding within silica particles.

As shown in Figure 2a, the silica particles maintained consistent hydrodynamic diameters of 109–117 nm across all surfactant concentrations (the detailed distribution of diameters measured by dynamic light scattering is shown in Figure S3). Compared to micelle-solubilized PdOEP, PdOEP embedded in silica particles demonstrated superior colloidal stability in aqueous media (Figure 2b). The micelle-solubilized PdOEP maintained stability for two weeks, after which a gradual decline in absorbance was observed. The degradation rate increased progressively over time. This stepwise degradation process was fitted using the Weibull model, revealing a half-life of approximately five weeks. By the sixth week, the absorbance at 546 nm had decreased to 30% of the initial value (A/A₀) [34,35]. In contrast, the absorbance of silica-embedded PdOEP remained constant over six weeks, with a fitted half-life exceeding one year [18]. High-resolution transmission electron microscopy (TEM) images (Figure 2c and Figure S4) revealed particles with an approximate size of 91 nm (smaller than the hydrated diameter of 109–117 nm) and mesoporous structures featuring pore diameters of 4–5 nm [36]. The mesoporous structures were formed using Tween 80 micelles containing embedded PdOEP as templates.

Brunauer–Emmett–Teller analysis demonstrated a surfactant-concentration-dependent evolution in pore architecture (Figures S5 and S6) [37]. As shown in Figure S6, increasing the Tween 80 concentration reduced the average pore size from ~4.7 nm (1 mg/mL) to ~4.0 nm (2–10 mg/mL), which corresponded to a significant enhancement in solubilization capacity (c/c₀). This phenomenon may be attributed to the less dense molecular packing within micelles at low surfactant concentrations, where weak intermolecular van der Waals forces led to the inefficient dispersion of PdOEP [38]. Concurrently, insufficient micelle formation at the Tween 80 concentration of 1 mg/mL limited the solubilization capacity and produced silica particles with a pore volume of only 0.45 cm³/g (Figure 2d); this was attributed to the reduced number of micelles in the reaction solution [39]. Increasing the surfactant concentration to 3 mg/mL increased the pore volume to 0.75 cm³/g, indicating enhanced micelle formation and expanded silica porosity. However, further increasing the concentration to 5 mg/mL reduced the pore volume to approximately 0.6 cm³/g.

According to the literature [40], increasing surfactant concentration decreases silica pore volume, attributed to structural changes in surfactant micelle molecular packing. As concentration rises, enhanced hydrophobic interactions between molecules promote tighter packing of surfactant molecules within the micelles acting as template agents. This denser packing reduces micellar volume, consequently decreasing the pore volume. As Tween 80 concentrations exceed 3 mg/mL, the reduction in the volume of silica-embedded micelles explains the observed decrease in C values. However, this structural transition of micelles within silica particles also reduces their internal free volume, potentially decreasing their capacity to accommodate PdOEP [16]. To further investigate the micellar accommodation of PdOEP within the mesopores, the relationship between intra-pore PdOEP concentration and surfactant concentration was determined using pore volume analysis combined with quantitative C determination (Equation (S1)), as shown in Figure 2d. The PdOEP concentration peaked at approximately 1.3 × 10⁻⁴ M between 5 and 8 mg/mL rather than at 3 mg/mL. A significant reduction in intra-pore concentration occurred only at 10 mg/mL (declining to 1.0 × 10⁻⁴ M). These results suggest an increased molecular density of PdOEP per unit pore volume, despite only a marginal reduction in C. The limited micellar free volume inevitably promoted the intermolecular aggregation of PdOEP, significantly reducing ¹O₂ generation efficiency. Subsequent UV-Vis spectroscopic analysis will elucidate how the increased aliphatic chain packing density affects PdOEP’s aggregation and photosensitizing properties.

2.2. Effect of Micellar Structures on PdOEP Aggregation

As mentioned above, UV-Vis absorption spectroscopy revealed surfactant-concentration-dependent aggregation (Figure 3a). The area ratio of the peak at 546 nm to the shoulder peak at 556 nm in the Q-band of PdOEP was determined via peak fitting (Figure 3b). This ratio increased significantly as the Tween 80 concentration increased from 5 to 10 mg/mL, even though the intra-pore PdOEP concentration decreased from 1.3 × 10⁻⁴ to 1.0 × 10⁻⁴ M (Figure 2d) [40,41,42]. This disparity between aggregation degree and intra-pore concentration indicates that aggregation is likely governed by aliphatic chain packing density rather than the PdOEP concentration in the mesopores of the silica particles. Aggregation abruptly occurred in the surfactant concentration range of 3–5 mg/mL, corresponding to the previously indicated critical packing transition of aliphatic chains (Figure 3c). The packing transition of the micelles within the silica particles reduces their internal free volume, thereby inducing aggregation of the embedded PdOEP molecules.

The mechanism of PdOEP aggregation was validated through the addition of cholesterol (10⁻⁶ M) at Tween 80 concentrations of 3–5 mg/mL to enhance the packing density of aliphatic chains [43,44]. The enhanced packing density increased aggregation at both 3 and 5 mg/mL (Figure 3d,e), confirming the correlation between packing density and aggregation degree. Furthermore, the silica particles prepared at the Tween 80 concentration of 3 mg/mL, which showed a lower degree of PdOEP aggregation, displayed greater sensitivity to cholesterol addition, with aggregation doubling. Conversely, the particles prepared at 5 mg/mL, which had a higher initial degree of aggregation, showed a weaker response to cholesterol (<20% increase in aggregation).

These results demonstrate that 3 mg/mL is the critical Tween 80 concentration for preparing particles with low aliphatic chain packing density. Under this condition, adding cholesterol (Figure 3f) or increasing the surfactant concentration significantly increases the aliphatic chain packing density within the micellar structure of the silica particles. This enhanced packing promotes PdOEP aggregation, which impairs aqueous ¹O₂ sensitization.

2.3. ¹O₂ Sensitization Activity

To quantify the ¹O₂ generation capacity, we measured the quantum yields (Φ_Δ) of PdOEP-embedded micelles and silica particles using Singlet Oxygen Sensor Green (SOSG) as a fluorescent indicator. As shown in Figure 4a, upon reacting with ¹O₂, the reaction product of SOSG emitted bright green fluorescence. The fluorescence intensity exhibits a linear relationship with the concentration of ¹O₂ [45], allowing the ¹O₂ quantum yield of PdOEP to be measured according to Equation (S2). The Φ_Δ values were determined from the linear slope of the plot of F_t/(F₀·A_PdOEP) vs. time (Figure S10).

At a Tween 80 concentration of 1 mg/mL, the insufficient embedding of PdOEP prevented reliable Φ_Δ quantification due to the negligible slope of the plot of F_t/F₀ vs. time. Consequently, Φ_Δ measurements were performed exclusively at Tween 80 concentrations ≥ 2 mg/mL.

As shown in Figure 4b and Figure S10, PdOEP in micelles maintained a constant Φ_Δ of approximately 93% across surfactant concentrations (2–10 mg/mL), consistent with its value in toluene solution [29]. However, for the PdOEP-embedded silica particles, Φ_Δ decreased sharply from >70% at 2–3 mg/mL Tween 80 to <60% at >5 mg/mL, exhibiting an inverse correlation with PdOEP aggregation degree (Figure 4b). At the maximum tested Tween 80 concentration (10 mg/mL), Φ_Δ decreased further to 44%. These results demonstrate that the Φ_Δ values of PdOEP-embedded silica particles were only comparable to those of micelle-dispersed PdOEP at Tween 80 concentrations ≤ 3 mg/mL. Higher concentrations of Tween 80 in the reaction solution caused severe intra-particle aggregation, significantly reducing Φ_Δ.

The ¹O₂ quantum yield merely assesses the ¹O₂ generation capacity per unit concentration of PdOEP. Thus, given that the PdOEP concentration (C) increased with Tween 80 concentration (up to <3 mg/mL), a comprehensive evaluation of ¹O₂ generation efficiency must also account for the mass of PdOEP embedded within the silica particles and the effects of aggregation-induced spectral changes. Therefore, we further defined the ¹O₂ generation efficiency as denoted by Equation (2):

$$C_{∆}={\displaystyle \frac{A_{\mathrm{m}}}{d}}\times {\Phi }_{\Delta },$$

(2)

where A_m denotes the absorbance of PdOEP at 552 nm per unit mass concentration of silica particles, and d is the optical path length (1 cm).

Due to PdOEP aggregation within the silica particles, C_Δ did not exhibit a simple positive or negative correlation with either C or Φ_Δ. As shown in Figure 4c, increasing the Tween 80 concentration from 2 to 3 mg/mL caused C_Δ to increase from 4.7 × 10⁻³ to 5.5 × 10⁻³ mg⁻¹·L·cm⁻¹. This enhancement primarily resulted from the significant increase in C (from 62% to 92%), which caused A_m to increase substantially from 0.029 to 0.044, while Φ_Δ remained above 70%. However, further increasing the Tween 80 concentration to 5 mg/mL caused C_Δ to decrease to 3.0 × 10⁻³ mg⁻¹·L·cm⁻¹. This decline can be partly attributed to the slight reduction in C (by 11%) and the subsequent decrease in A_m. More significantly, the increased degree of PdOEP aggregation led to a decrease in Φ_Δ to 57%. At a Tween 80 concentration of 10 mg/mL, C_Δ further decreased to 2.3 × 10⁻³ mg⁻¹·L·cm⁻¹. Here, C decreased only to 75%; the primary factor was severe PdOEP aggregation, which caused Φ_Δ to decrease further from 57% to 44%. Therefore, the Tween 80 concentration of 3 mg/mL was optimal for maximizing C_Δ during the synthesis of PdOEP-embedded silica particles.

At a Tween 80 concentration of 3 mg/mL, the as-prepared PdOEP-embedded silica particles exhibited excellent stability during storage while maintaining high ¹O₂ generation efficiency. Building on these properties, we further evaluated their potential for two-photon excitation using a 1030 nm laser (high-repetition-rate femtosecond laser PHAROS, manufactured by Light Conversion, in Vilnius, Lithuania) [46]. The linear relationship of F_t/F₀ with time (Figure 4d and Figure S13) confirms the effectiveness of PdOEP activation via two-photon excitation for generating ¹O₂. This near-infrared excitation capability demonstrates the potential of our PdOEP-embedded silica particles for photodynamic therapy targeting deep subcutaneous tissues.

3. Materials and Methods

3.1. Materials

Palladium octaethylporphyrin (PdOEP) was procured from Aldrich (St. Louis, MO, USA). Tween 80 and Tween 20 (cell culture grade) were purchased from Aladdin (Shanghai, China). Na₂HPO₄ and NaH₂PO₄ were acquired from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Cholesterol was sourced from Xianheng Chemical Co., Ltd. (Hangzhou, China), and toluene was obtained from Tianjin Bohai Chemical Co., Ltd. (Tianjin, China). Rose Bengal was provided by Merck (Darmstadt, Germany), and Singlet Oxygen Sensor Green (SOSG) was purchased from Dalian Meilun Biotech Co., Ltd. (Dalian, China). TEOS was purchased from Macklin (Shanghai, China), and NaF was sourced from Tianjin Yuanli Chemical Co., Ltd. (Tianjin, China). Ultrapure water was prepared using a laboratory ultrapure water system.

3.2. Preparation of PdOEP-Embedded Silica Particles

PdOEP was mixed with different concentrations of Tween 80 and Tween 20 in 10 mL of toluene, with PdOEP concentration set at 10⁻⁵ M and surfactant concentrations varying at 1, 2, 3, 5, 8, and 10 mg/mL. The solvent was removed via rotary evaporation and vacuum drying, followed by redispersion in an equal volume of ultrapure water, forming micelles solubilizing PdOEP within Tween 80 and Tween 20. UV–visible absorption spectroscopy was used to analyze the micellar samples at varying surfactant concentrations. Subsequently, 80 μL of 0.5 M NaF aqueous solution and 100 μL of TEOS were added to each sample, followed by stirring for 24 h to achieve silica coating [47]. After silica coating, the samples were centrifuged, and the supernate was subjected to UV–visible absorption spectroscopy to calculate the particle embedding efficiency (ω) based on the Lambert–Beer law:

$$\mathsf{\omega }=\left(1-{\displaystyle \frac{A_{\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{n}\mathrm{a}\mathrm{t}\mathrm{e}}}{A_{\mathrm{m}\mathrm{i}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{e}}}}\right)\times 100\%$$

(3)

where ω represents embedding efficiency, and A represents absorbance of the supernate and micelles. To minimize scattering effects in the colloidal system, this study employed absorbance at the Q-band (546 nm) in the visible region for quantitative analysis, rather than the Soret band in the ultraviolet region (at 394 nm). The precipitate was then washed three times with ultrapure water and centrifuged. The dried precipitate was re-dispersed, yielding PdOEP-embedded silica particles.

3.3. Characterization of Silica Particles

UV–visible absorption spectra and fluorescence emission spectra were recorded at room temperature using UV-1900 and RF-6000 spectrophotometers (Shimadzu Corporation, Kyoto, Japan), respectively. The particle size of micelles and particles was measured via dynamic light scattering (DLS) using a Zetasizer Nano ZS 90 (Malvern Instruments Ltd., Malvern, UK). The DLS data are detailed in Figure S3. The silica particle samples were calcined in a muffle furnace at 300 °C and 600 °C for six hours each and degassed at 250 °C for six hours [48]. Nitrogen adsorption–desorption isotherms and the pore size distribution of the silica particles were determined using a Micromeritics ASAP 2010 Accelerated Surface Area and Porosimetry Analyzer (Micromeritics, Norcross, GA, USA). The mesoporous structure and morphological features of the samples were observed via transmission electron microscopy (TEM).

3.4. Stability of Silica Particles

To explore the stability of Tween micelles and silica-coated particles, the aqueous dispersions of micelles and silica particles were stored at room temperature, and the UV–visible absorption spectra were recorded weekly to monitor A_t/A₀, where A_t represents absorbance of silica particles or micelles after weeks, and A₀ represents absorbance of silica particles or micelles at the beginning.

3.5. Measurement of ¹O₂ Quantum Yield

The ¹O₂ quantum yield of PdOEP micelles and silica-coated particles was measured at room temperature, following the method which has been outlined in the literature [49]. Singlet Oxygen Sensor Green (SOSG) was employed as the ¹O₂ capture agent. The specific testing procedure involved dissolving 100 μg of SOSG solid in 330 μL of methanol and diluting it with 587 μL of ultrapure water to prepare the SOSG stock solution. A mixture of 2.9 mL of the test sample solution and 0.1 mL of the SOSG stock solution was prepared, ensuring the full dissolution of SOSG, with micelles and silica particles dispersed in 0.5 M PBS buffer. To prevent self-absorption, a 552 nm light source was used to excite the samples to generate triplet states. After capturing ¹O₂, SOSG produced green fluorescence emission at 530 nm, and the emission spectrum was recorded at 30 s intervals. A linear fit was applied to the peak intensity over time, using Rose Bengal (RB) as the reference standard (¹O₂ quantum yield Φ_ΔRB = 75% in water). The absorption spectra and SOSG response data of Rose Bengal are shown in Figure S11. The ¹O₂ quantum yield of the samples was calculated using Equation (S2).

3.6. Two-Photon Excitation Experiment

The SOSG solution prepared in the previous step was mixed with PBS-dispersed samples and stirred continuously. The samples were subjected to two-photon excitation using a 1030 nm laser (the irradiation power density of the laser was 3 W/cm²), with a single excitation duration of 60 s, followed by recording of the SOSG emission spectrum at each interval to qualitatively analyze the ¹O₂ generation capability of the two-photon excited samples.

4. Conclusions

This work demonstrates that PdOEP-embedded silica particles are stable aqueous platforms for ¹O₂ sensitization using a Tween micelle-assisted synthetic approach. In micellar systems, Tween 80 efficiently solubilized PdOEP in aqueous solution. Although PdOEP demonstrated effective dissolution with a high ¹O₂ quantum yield (Φ_Δ ≈ 93%), its stability was limited, with a half-life of approximately 5 weeks. Following silica deposition, over 95% of the solubilized PdOEP was successfully embedded within the silica particles. This embedding substantially enhanced the stability, increasing the half-life to over one year. Under the optimized synthetic conditions, PdOEP embedded in silica particles achieved a Φ_Δ exceeding 70%, comparable to that observed in micellar systems. In addition, we observed efficient ¹O₂ generation under two-photon excitation at 1030 nm, highlighting significant potential for deep-tissue biomedical applications.