Silicon Nanoparticle Photophysics and Singlet Oxygen Generation

Published on Web 05/21/2010 pubs.acs.org/Langmuir ... Ciudad Universitaria, Pabell´on 2, Piso 3 (C1428EHA) Buenos Aires, Argentina. Received March 10...
0 downloads 0 Views 1MB Size
pubs.acs.org/Langmuir © 2010 American Chemical Society

Silicon Nanoparticle Photophysics and Singlet Oxygen Generation )

)

Manuel J. Llansola Portoles,† Pedro M. David Gara,† Monica L. Kotler,‡ Sonia Bertolotti,§ Enrique San Roman, Hernan B. Rodrı´ guez, and Monica C. Gonzalez*,† INIFTA, Dpto. Quı´mica, FCE, UNLP. CC 16 Suc. 4, (1900) La Plata, Argentina, ‡Departamento Quı´mica Biol ogica, Facultad de Ciencias Exactas y Naturales, UBA. Ciudad Universitaria, Pabell on 2, Piso 4 (C1428EHA) Buenos Aires, Argentina, §Fac Cs. Exactas, Fı´sico-Quı´micas y Naturales, UNRC. Ruta Nac. 36, Km. 601 (X5804BYA) Rı´o Cuarto, Argentina, and INQUIMAE, Facultad de Ciencias Exactas y Naturales, UBA. Ciudad Universitaria, Pabell on 2, Piso 3 (C1428EHA) Buenos Aires, Argentina )



Received March 10, 2010. Revised Manuscript Received May 7, 2010 The effect of molecular oxygen and water on the blue photoluminescence of silicon nanoparticles synthesized by anodic oxidation of silicon wafers and surface functionalized with 2-methyl 2-propenoic acid methyl ester is investigated. The particles of 3 ( 1 nm diameter and a surface composition of Si3O6(C5O2H8) exhibit room-temperature luminescence in the wavelength range 300-600 nm upon excitation with 300-400 nm light. The luminescence shows vibronic resolution and high quantum yields in toluene suspensions, while a vibronically unresolved spectrum and lower emission quantum yields are observed in aqueous suspensions. The luminescence intensity, though not the spectrum features, depends on the presence of dissolved O2. Strikingly, the luminescence decay time on the order of 1 ns does not depend on the solvent or on the presence of O2. To determine the mechanisms involved in these processes, time-resolved and steady-state experiments are performed. These include low-temperature luminescence, heavy atom effect, singlet molecular oxygen (1O2) phosphorescence detection, reaction of specific probes with 1O2, and determination of O2 and N2 adsorption isotherms at 77 K. The results obtained indicate that physisorbed O2 is capable of quenching nondiffusively the particle luminescence at room temperature. The most probable mechanism for 1O2 generation involves the energy transfer from an exciton singlet state to O2 to yield an exciton triplet of low energy (kBT (0.0256 eV at room temperature) or the lowest triplet is a nonluminescent state. The luminescence lifetime of the order of 1 ns is too short to allow O2 molecules to approach the MMASi-NP by diffuson. Under such conditions, only static quenching may be operative.17 A dynamic adsorption equilibrium of O2 on the MMASi-NP surface leads to a linear dependence of I0/I with [O2] as shown in (16) Gross, E.; Kovalev, D.; K€unzner, N.; Diener, J.; Koch, F.; Timoshenko, V. Y.; Fujii, M. Phys. Rev. B 2003, 68, 115405. (17) Valeur, B. Molecular Fluorescence. Principles and Applications; Wiley-VCH: Weinheim, Germany, 2002.

Langmuir 2010, 26(13), 10953–10960

Llansola Portol es et al.

Article

Table 1. Emission Quantum Yields (Φem), Luminescence Decay Times (τ) Obtained by 375 nm Excitation, and 1O2 Quantum Yields (ΦΔ) by 355 nm Excitation, Measured for MMASi-NP under Different O2 Concentrations in Toluene and One-Month-Aged Aqueous Suspensions of pH 7 at Room Temperature aqueous suspensions

toluene suspensions

% O2 v/v

Φem

τ (% emitted light)a

ΦΔ

Φem

τ (% emitted light)a

ΦΔ

0 22 100

0.20 ( 0.02 0.18 ( 0.02 0.16 ( 0.02

1.5 ns (85) and 8.1 ns (15) 1.2 ns (92) and 7.6 ns (8) 1.3 ns (83) and 6.9 ns (17)

----------0.02 ( 0.01

0.76 ( 0.07 ----0.44 ( 0.05

1.4 ns (87.5) and 9.5 ns (12.5) 1.3 ns (89.4) and 7.5 ns (10.6) ------

----------0.15 ( 0.04

ε(λ) a

0.28 dm2 g-1 (340 nm) or 2169 M-1 cm-1 In all cases, χ values range from 1.00 to 1.05.

0.76 dm2 g-1 (360 nm) or 5887 M-1 cm-1

2

Figure 3. Normalized emission, λex = 340 nm (full line), and exci-

tation, λem = 420 nm (dotted line), spectra of MMASi-NP in airsaturated aqueous suspensions at pH 7 and room temperature retrieved from the bilinear regression analysis of the excitationemission matrix. The dashed line represents the emission spectrum of the same suspensions obtained at 77 K upon 340 nm excitation. Inset: Plot of the ratio of the total emission intensity in the absence and presence of molecular oxygen (I0/I) vs the partial pressure of oxygen, P(O2), for 0.500 g/L MMASi-NP aqueous suspensions of pH 3 (O), 7 (3), 12 (0), and in toluene suspensions (b).

eq 1, where K is the adsorption equilibrium constant. In the latter situation, the excited-state lifetime of the uncomplexed fluorophore is unaffected though the luminescence intensity of the suspension decreases upon addition of the quencher (O2), as observed here. Considering the abundant evidence in the literature on the adsorption of molecular oxygen on silicon surfaces, this situation will be further discussed in the following sections. Io ¼ 1 þ K½O2  I

ð1Þ

MMASi-NP in Aqueous Suspensions: Effect of pH and the Presence of Molecular Oxygen. A bilinear analysis of the excitation-emission matrix in aqueous suspensions indicates the existence of one well-defined population of particles emitting in the blue, as the emission spectrum does not depend on the excitation wavelength. The excitation and emission spectra of MMASiNP in aqueous suspensions at pH 7 are shown in Figure 3. The luminescence of MMASi-NP in aqueous suspensions strongly differs from that in organic solvents, both in emission intensity and in band shape.9 Despite the fact that the emission maximum remains in the 430-440 nm wavelength range, the excitation maximum in aqueous suspension is shifted 40 nm to the UV, and both the excitation and emission spectra lose their peak resolution. The emission intensity, but not the spectrum shape, strongly depends on dissolved oxygen concentration and aging of the suspension. The luminescence intensity increases with aging at rates strongly depending on pH. MMASi-NP suspended in aqueous solutions at pH 7 needed to be aged over a week to achieve the same luminescence increase obtained after several minutes at pH 12. Langmuir 2010, 26(13), 10953–10960

Figure 4. FTIR spectra of MMASi-NP supported on KBr pellets before (top spectrum) and after (bottom spectrum) suspension in an aqueous solution of pH 7 for a week.

The FTIR spectrum of the MMASi-NP obtained after suspension for a week in an aqueous solution at pH 7 followed by solvent evaporation (bottom spectrum in Figure 4) shows important absorption peaks at around 3500 cm-1 absent in the FTIR spectrum of the fresh particles (upper spectrum in Figure 4) which may be assigned to OH stretching in Si-OH. Broader peaks at 1115 and 1055 cm-1 characteristic of Si-O-Si vibrations are also observed. All other peaks due to C-H stretching of the CH3 group (2920 and 2850 cm-1), CdO stretching (1724 cm-1), CH3 deformation modes in alkanes and esters (1450 and 1370 cm-1, respectively), Si;CdC vibration modes (1600 and 960 cm-1), and Si-CH2-CH deformation (1260 cm-1), attributed to the covalent attachment of MMA to the surface,9 are still observed after suspension in water. Hydrolysis of the methyl ester group of surface-attached MMA leading to the formation of terminal COOH groups cannot be discarded since absorption due to CdO and carboxylic HO stretching (3100-2400 cm-1) is observed. However, the small peak at 1724 cm-1 (CdO stretching) does not match the strong OH stretching peak at 3500 cm-1,18,19 and therefore, COOH groups alone cannot account for the strong absorption at 3500 cm-1. Therefore, alkaline solutions catalyze the surface hydroxylation of the particles, in agreement with the reported increased solubilization of silica (SiO2) to alkali silicate in colloidal dispersions at pH>10.20 Hydroxylated silica presents an H-bonding network of water molecules immobilized by the surface hydroxyl groups (Si-OH).21 Considering that surface Si-OH groups are formed in aqueous suspensions of MMASiNP, the observed structureless emission band may be a consequence of fluctuations in the structure of the solvation shell around the particle (inhomogeneous broadening). Once the surface has been modified, the luminescence yield does not show, within the (18) Lambert, J. B.; Shuvell, H. F.; Lightner, D. A. Organic Structural Spectroscopy; Prentice Hall, Inc.: Upper Saddle River, NJ, 1998. (19) Judge, K.; Brown, C. W.; Hamel, L. Anal. Chem. 2008, 80, 4186–4192. (20) Zhuravlev, L. T. Colloids Surf., A 2000, 173, 1–38. (21) Caregnato, P.; Mora, V. C.; Le Roux, G. C.; Martire, D. O.; Gonzalez, M. C. J. Phys. Chem. B 2003, 107, 6131–6138.

DOI: 10.1021/la100980x

10955

Article

experimental error, any variation with aging for several months. These observations suggest that the introduction of surface silanols on the MMASi-NP is related to an increase in the luminescence intensity. The luminescence quantum yield of aged argon-saturated aqueous suspensions of MMASi-NP is around 25% of that observed in toluene suspensions, though the luminescence decay time does not depend on the solvent (see Table 1). Moreover, after water elimination and resuspension in toluene, the particles recover their original spectral band resolution and strong emission. These observations indicate that water affects Φem in two ways: it enhances the luminescence emission by promoting the irreversible hydroxylation of the particle surface and strongly decreases Φem by static quenching favoring radiationless transitions that do not lead to the emitting exciton. As observed for the toluene suspensions, Φem values determined in aqueous suspensions depend on [O2], whereas the luminescence lifetimes are not sensitive to its presence, as shown in Table 1. The luminescence quenching by oxygen does not depend on the excitation wavelength, as expected for a unique emitting chromophore (vide supra). Figure 3 inset shows the linear dependence of Io/I on the partial pressure of molecular oxygen, P(O2). Within experimental error, the data obtained for toluene and aqueous suspensions at different pH values fall within the same line, supporting a molecular oxygen adsorption equilibrium undisturbed by the solvent as the main process capable of quenching the MMASi-NP luminescence. From the slope of the plot in Figure 3, eq 1 yields K = (0.49 ( 0.03) bar-1 at room temperature. A hint on the nature of O2 adsorption (chemical or physical) on the MMASi-NP surface can be obtained from the magnitude of the adsorption enthalpy. From the experiments on O2 quenching in toluene suspensions, a rough estimation yields K=3.3 bar-1 at 77 K (the value of K obtained from the adsorption experiments in Figure 1 could not be taken since large porous silicon particles were also present in these experiments). Therefore, an adsorption enthalpy of -1650 J mol-1 is obtained, which strongly supports a physical adsorption of molecular oxygen on the MMASi-NP, in agreement with the reversibility of the adsorption process observed at low temperature. Luminescence (see Figure 3 dashed spectrum) shows a broad band with maximum at 465 nm at 77 K, 40 nm red-shifted with respect to that observed at room temperature, and coincident with the emission maximum observed in toluene suspensions. This behavior may be a consequence of the impossibility of solvent molecules to reorganize in water at 77 K; therefore, emission arises from a state close to the Franck-Condon state, as in the case of a less polar medium. The latter observations indicate that different dipole moments may be expected for the ground and excited states of the particles, in agreement with the Stokes shift and anisotropy at time zero (ro =0.25) measured for MMASi-NP toluene suspensions at room temperature.9 Singlet Oxygen Generation. The luminescence quenching by O2 may be a sign of formation of 1O2. Therefore, two sets of experiments were performed to probe 1O2 formation and its reactivity toward the particles: (a) time-resolved detection of 1O2 luminescence at 1270 nm and (b) detection of the molecular oxygen consumption in continuous irradiation experiments of particle suspensions in the absence and presence of furfuryl alcohol (FFA) and sodium azide as specific probes for 1O2. Evidence of 1O2 formation in toluene suspensions is evidenced by the time-resolved phosphorescence observed at 1270 nm (Figure 5). From the comparison of the total 1O2 phosphorescence emission generated by the MMASi-NP (IΔ) and that 10956 DOI: 10.1021/la100980x

Llansola Portol es et al.

Figure 5. NIR 1O2 phosphorescence at 1270 nm obtained after 355 nm irradiation. Plot A: oxygen-saturated toluene suspensions of DPA (upper curve, 0.16 absorbance) and of MMASi-NP (lower curve, 0.13 attenuance). Plot B: oxygen-saturated deuterated aqueous suspension of MMASi-NP at pD 6.4. Inset B: Total 1O2 emission (IΔ) observed upon 355 nm irradiation of a deuterated aqueous suspension of MMASi-NP as a function of the partial pressure of oxygen.

generated by the reference9,10 diphenyl anthracene (DPA) (IΔR)22 under similar solution absorbance (A) at 355 nm and experimental setup, a quantum yield of 1O2 generation ΦΔ =0.15 ( 0.04 is obtained for MMASi-NP taking ΦΔR =0.8 for the reference in toluene23 (see Supporting Information). The singlet oxygen lifetime τΔ = 20 μs obtained from the decay of these traces are in agreement with reported values in toluene solutions.24 The short lifetime of 1O2 emission in water (τΔ =3.5 μs25) does not allow measurements with our experimental setup; for that reason, D2O (τΔ =68 μs25) was used as the solvent in the transient detection studies shown in Figure 5 for MMASi-NP aqueous suspensions. Evidence of formation of 1O2 is provided by the traces at 1270 nm showing τΔ = 60 ( 15 μs in D2O suspensions. The agreement observed for the decay times of 1O2 generated upon irradiation of MMASi-NP in toluene and D2O suspensions with the 1O2 lifetime reported in these solvents supports a negligible reaction with the MMASi-NP surface, as will be discussed further. The quantum yield of 1O2 generation in oxygen-saturated deuterated aqueous suspensions is approximately 14% of the value observed in toluene. The dependence of the total 1O2 phosphorescence emission with the dissolved oxygen concentration shown in Figure 5B inset may be described by eq 2. This equation describes a Langmuir approach for the dynamic adsorption equilibrium of oxygen molecules on a surface, in agreement with the behavior reported for 1O2 production by porous silicon.5 In fact, considering that the total phosphorescence emission is proportional to an instrument constant (A) and to the product of the probability of energy transfer to the adsorbed molecules (P) times the fraction of the surface covered by O2 molecules, eq 2 agrees with the Langmuir approach if b=K and a=A  P  K, with K=(0.6 ( 0.1) bar-1. Within the experimental error, the latter value is in agreement with that found for the quenching of the luminescence by adsorbed (22) Brauer, H. D.; Acs, A.; Drews, W.; Gabriel, R.; Ghaeni, S.; Schmidt, R. J. Photochem. 1984, 25, 475–488. (23) Scaiano, J. C. CRC Handbook of Organic Photochemistry; CRC Press, Inc.: Boca Raton, 1989; Vol. 2, pp 229-251. (24) Okamoto, M.; Tanaka, F. J. Phys. Chem. 1993, 97, 177–180. (25) Wilkinson, F.; Worrall, D. R.; Williams, S. L. J. Phys. Chem. 1995, 99, 6689–6696.

Langmuir 2010, 26(13), 10953–10960

Llansola Portol es et al.

Article

Figure 6. Molecular oxygen consumption during 350 nm irradiation of an air-saturated aqueous suspension of 6.4  10-6 M MMASi-NP contained in a closed reactor containing the following: (O) no added substrates, (3) 10-3 M FFA, and (b) 10-3 M FFA and 10-2 M azide.

oxygen in toluene and aqueous suspensions. Therefore, there is strong evidence for the energy transfer from the excitons confined in the MMASi-NP to oxygen molecules adsorbed on the NP surface. IΔ ¼

a½O2  1 þ b½O2 

ð2Þ

Continuous irradiation with light of 350 nm of an air-saturated aqueous suspension of MMASi-NP shows no O2 depletion with the irradiation time (Figure 6). However, irradiation of the latter suspensions in the presence of 10-3 M FFA, a well-known chemical quencher of 1O2 (reaction ), leads to the consumption of O2, as also observed in Figure 6. These observations indicate that 1O2 is generated by the reaction system and consumed by reaction 3, with the concomitant depletion of molecular oxygen. Addition of excess azide (10-2 M), an efficient physical quencher of 1O2 (reaction 4), to the reaction mixture inhibits O2 consumption as expected from the competition of reactions 3 and 4. In aqueous solutions, the electronic excitation energy of 1O2 is efficiently converted into vibrational energy of terminal bonds of deactivating water molecules (reaction 5) and no oxygen consumption is expected in the absence of efficient 1O2 chemical quenchers, as observed here. A negligible reactivity of MMASiNP with 1O2 is therefore supported by these experiments. Overall, these experiments show the potential capability of the aqueous suspensions of the particles to be used for the specific oxidation of dissolved substrates.

Heavy Atom Effect on MMASi-NP Luminescence. Internal and external heavy-atom perturbations have long been known to effectively enhance spin-orbit coupling, thus favoring singlet-triplet transitions.28,29 To envisage the participation of a triplet state in the production of 1O2, the luminescence spectrum, lifetime, and 1O2 generation of MMASi-NP suspensions were measured in (i) a solvent mixture composed of toluene and iodoethane (26) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M., Environmental Organic Chemistry; John Wiley & Sons: Hoboken, NJ, 2002. (27) Miskoski, S.; Garcı´ a, N. A. Photochem. Photobiol. 1993, 57. (28) Kasha, M. J. Chem. Phys. 1952, 20, 71–74. (29) Kasha, M. Radiat. Res. Suppl. 1960, 243–275.

Langmuir 2010, 26(13), 10953–10960

Figure 7. Plot A: Emission spectra of 1.5  10-6 M MMASi-NP argon-saturated suspensions (360 nm excitation) in a 1:3 iodoethane/ toluene mixture (lower spectrum) and in toluene (upper spectrum). Inset A: Emission decay at 440 nm for MMASi-NP argon-saturated suspensions in a 1:3 iodoethane/toluene mixture (lower curve) and in toluene (upper curve) exciting at 341 nm. Plot B: Emission spectra of 6.4  10-6 M MMASi-NP argon-saturated aqueous suspensions (340 nm excitation) (full line) in the presence of 0.1 M KI (dashed line) and in the presence of 1  10-4 M MV2þ (dotted line). Inset B: Transient formation at 390 nm during conventional flash photolysis experiments with MMASi-NP in aqueous (upper curve) and toluene (lower curve) suspensions containing 1  10-4 M MV2þ. Irradiation was performed at λ > 300 nm.

in a 1:3 proportion and (ii) an aqueous suspension containing 0.1 M KI. The observed spectrum in argon-saturated toluene/iodoethane mixtures (Figure 7A) resembles that in pure toluene suspensions, though with less peak resolution and with a total luminescence 35% of that in the absence of iodoethane. On the other hand, the presence of 0.1 M KI in an argon-saturated aqueous suspension of MMASi-NP reduces the total emission by 20% and shows no effect on the spectrum shape (Figure 7B). The luminescence decay times of 1.3 ( 0.2 ns measured for the iodoethane and KIcontaining suspensions are in agreement with those measured previously in toluene and aqueous suspensions. Such behavior is expected for processes that cannot be controlled by diffusion due to their short excited-state lifetime (see previous discussions). A reduction of less than 10% in the total luminescence intensity is observed upon saturation of the latter suspensions with molecular oxygen. Irradiation experiments with 355 nm light of MMASi-NP in oxygen-saturated toluene/iodoethane and 0.1 M KI D2O suspensions showed no 1O2 phosphorescence at 1270 nm. The whole set of observations indicate that the involvement of a triplet state emitting at wavelengths of