Relationship between the Microscopic and Macroscopic World in

Aug 28, 2009 - An investigation based on confocal fluorescence lifetime imaging microscopy (FLIM) of silica-loaded silicone films doped with a molecul...
0 downloads 12 Views 830KB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Relationship between the Microscopic and Macroscopic World in Optical Oxygen Sensing: A Luminescence Lifetime Microscopy Study Juan Lopez-Gejo, David Haigh, and Guillermo Orellana* Department of Organic Chemistry, Faculty of Chemistry, Complutense University of Madrid, 28040 Madrid, Spain Received July 13, 2009. Revised Manuscript Received July 28, 2009 An investigation based on confocal fluorescence lifetime imaging microscopy (FLIM) of silica-loaded silicone films doped with a molecular oxygen-sensitive ruthenium(II) polyazaheterocyclic complex is presented. The effect of the silica type (hydrophilic/hydrophobic), particle size and amount of silica filler on the luminescence decay of the immobilized indicator dye has thoroughly been studied. A higher amount of hydrophilic silica leads to both a higher solubility of molecular oxygen into the silicone film and to higher levels of the metal indicator dye. Thus, incorporation of 10% (by wt) pyrogenic silica into silicone shortens the mean luminescence lifetime from 1.4 to 0.9 μs. However, an excess of filler may lead to overloading of the dye into the film producing new phenomena such as triplet-triplet annihilation and excitation energy homotransfer, as observed from their influence on the emission lifetime of the metal complex. Those phenomena do not take place when trimethylated silica (hydrophobic filler) is used. In this case, no increase on the oxygen or dye concentration is observed after addition of the filler and no significant reduction of the luminescence lifetime is measured. Both the addition of silica and the possible precipitation of dye crystals lead to the appearance of microdomains where the molecular probe exhibits widely different excited state lifetimes. For the first time, such microdomains within the oxygen sensing layer are visualized and analyzed by means of FLIM, showing the potential of this technique and the usefulness of our conclusions to the future design and development of novel luminescent oxygen sensor films for environmental and process analysis.

Introduction Molecular oxygen (O2) concentration is a critical parameter in several fields such as environmental monitoring, wastewater treatment, process analysis (food, chemicals, fuels, etc.), medical science, vehicle design, and cell metabolism or in combustion monitoring, to name a few. Luminescence-based O2 sensors offer advantages over electrochemical devices including ease of miniaturization, lack of analyte consumption, faster response, robustness, insensitivity to interfering agents and absence of electrical interference. Moreover, indicator-mediated fiber-optic luminescence sensing of O2 has led the way in the development of current optoelectronic devices for in situ chemical analysis.1 The low maintenance, extended operational lifetime and reliability of fiber-optic oxygen sensors based on luminescent transition metal (Ru, Pd, Pt, Ir) complexes with polyazaheterocyclic chelating ligands (bipyridines and phenanthrolines,2 porphyrins,3 etc.) is such, that every major manufacturer of environmental monitors is currently offering at least one model for in situ optical measurements of dissolved O2 in water,4 rapidly phasing out the *To whom correspondence should be addressed. E-mail: orellana@ quim.ucm.es. Telephone: þ34 913 944 220. Fax: þ34 913 944 103. (1) (a) Demas, J. N.; DeGraff, B. A.; Coleman, P. B. Anal. Chem. 1999, 71, 793A–800A. (b) Wolfbeis, O. S. In Optical Sensors: Industrial, Environmental and Diagnostic Applications; Springer: Berlin and Heidelberg, Germany, 2004; Vol. 1, pp 1-34. (c) Papkovsky, D. B. In Oxygen Sensing; Academic Press Inc.: San Diego, CA, 2004; Vol. 381, pp 715-735. (d) DeGraff, B. A.; Demas, J. N. In Reviews in Fluorescence; Springer: New York, 2005, pp 125-152. (2) Orellana, G.; Garcı´ a-Fresnadillo, D., in Optical Sensors: Industrial, Environmental and Diagnostic Applications, Springer: Berlin and Heidelberg, Germany, 2004; Vol. 1, pp 309-357. (3) (a) Amao, Y. Microchim. Acta 2003, 143, 1–12. (b) Douglas, P; Eaton, K. Sens. Actuat. B: Chem. 2002, 82, 200–208. (c) Gewehr, P. M.; Delpy, T. Med. Biol. Eng. Computat. 1993, 31, 11–21. (4) See, for instance, the commercial fiber-optic O2-sensing devices Hach LDO probe, YSI 6150 sensor, In-Situ Inc. RDO sensor, Environmental Instruments FL3 Fluoroprobe, PreSens Fibox and Microx sensors, Interlab IE Optosen system, etc.

2144 DOI: 10.1021/la902546k

amperometric Clark electrode introduced almost 50 years ago. Moreover, the vast majority of the current luminescent O2 sensors are based on emission lifetime measurements as they are immune to indicator leaching, lamp and detector aging and, very often, also to the dye photobleaching. High sensitivity and linear response in the widest possible range of O2 concentrations are desirable properties of an ideal sensing film. The appropriate selection of a luminescent dye with a long excited state lifetime and a supporting polymer material with a high O2 permeability are key issues in the design of the ideal oxygen optical sensor.5 As far as the O2-sensitive dye is concerned, polyazaheterocyclic complexes of Ru(II) are among the most suitable indicators due to their large Stokes’ shift, μs lifetimes, photostability and the close to diffusion-limited rate of quenching of their (triplet) metal-to-ligand charge transfer excited state with O2 (>109 L mol-1 s-1).2 For the present study, the widely used tris(4,7-diphenyl-1,10phenanthroline)ruthenium(II) dication (abbreviated RD3)6 was selected as the oxygen-sensing dye due to its long emission lifetime (ca. 5.5 μs in solution under argon) and high luminescence quantum yield (ca. 0.4 in deoxygenated solution). With regard to the indicator polymer support, many different materials have been investigated looking for a high O2 permeability or sacrificing this highly desirable feature to improve sturdiness. In this way, fluorinated polymers,7 copolymers8 and (5) Lu, X.; Winnik, M. A. Chem. Mater. 2001, 13, 3449–3469. (6) (a) Xavier, M. P.; Garcı´ a-Fresnadillo, D.; Moreno-Bondi, M. C.; Orellana, G. Anal. Chem. 1998, 70, 5184–5189. (b) Mingoarranz, F. J.; Moreno-Bondi, M. C.; García-Fresnadillo, D.; de Dios, C.; Orellana, G. Mikrochim. Acta 1995, 121, 107–118. (c) Navarro-Villoslada, F.; Orellana, G.; Moreno-Bondi, M. C.; Vick, T.; Driver, M.; Hildebrand, G.; Liefeith, K. Anal. Chem. 2001, 73, 5150–5156. (7) (a) Obter, O.; Ertekin, K.; Dayan, O.; Cetinkaya, B. J. Fluoresc. 2008, 18, 269–276. (b) Amao, Y.; Asai, K.; Miyashita, T.; Okura, I. Polym. Adv. Technol. 2000, 11, 705–709. (c) García-Fresnadillo, D.; Marazuela, M. D.; Moreno-Bondi, M. C.; Orellana, G. Langmuir 1999, 15, 6451–6549. (8) Amao, Y.; Miyashita, T.; Okura, I. J. Porphyrin Phthalocyan. 2001, 5, 433– 438. (b) Amao, Y.; Miyashita, T.; Okura, I. React. Funct. Polym. 2001, 47, 49–54.

Published on Web 08/28/2009

Langmuir 2010, 26(3), 2144–2150

L opez-Gejo et al.

silica sol-gels,9 together with organically modified silicates (ormosils),10 zeolites11 or even more complex nanostructured materials12 have been used, with the intention of finding the ideal matrix for each particular application. Most of the matrices are microheterogeneous materials formed by several components (fillers, cross-linkers, plasticizers, catalyst, etc.) that play different roles (increase the dye or oxygen solubility, enhance flexibility, improve resistance, etc.). Among the solid supports for oxygen indicator dyes, polydimethylsiloxane (silicone) is probably the most widespread material due to its extraordinary permeability to the analyte.13 Thus, oxygen sensors based on luminescent ruthenium complexes immobilized into silicone films have been extensively studied.1,2,14 Together with the large number of materials investigated, the interplay between the support and the O2 indicator dye has been addressed often.14d,15 Lu and Winnik authored a very complete review, where the principal questions concerning that issue were discussed.5 In spite of the fact that in simple systems a monoexponential luminescence decay could be expected, for dyes attached to or dissolved into a pure polymer film such behavior is the exception rather than the rule.16 Of course, in more elaborated systems, e.g., where a filler is added to increase solubility of the luminescent dye and make it compatible with the physicochemical properties of the polymer support, the existence of many microdomains yields multiexponential profiles. The presence of different environments around the indicator dye with different polarity and accessibility of the quencher (O2 in this case) is likely responsible for the nonlinear Stern-Volmer quenching pattern.1c,3b,17 Another possibility is association of the dye molecules or even precipitation of dye salt microcrystals within the polymer film due to the scarce solubility of salts in a hydrophobic macromolecular network. In the latter cases, selfquenching would critically influence the lifetime of the dye adding shorter components to the luminescence decay of the sample.18 To the best of our knowledge, although widely accepted by the scientific community, the existence of such microdomains with different emission lifetimes has never been visualized so far. Bowman et al. demonstrated the existence of microdomains within an O2-sensitive membrane by means of fluorescence (9) (a) Chu, C.-S.; Lo, Y.-L. Sens. Actuat. B: Chem. 2007, 124, 376–382. (b) Bukowski, R. M.; Ciriminna, R.; Pagliaro, M.; Bright, F. V. Anal. Chem. 2005, 77, 2670–2672. (10) (a) Klimant, I.; Ruckruh, F.; Liebsch, G.; Stangelmayer, A.; Wolfbeis, O. S. Mikrochim. Acta 1999, 131, 35–46. (b) Higgins, C.; Wencel, D.; Burke, C. S.; MacCraith, B. D.; McDonagh, C. Analyst 2008, 133, 241–247. (c) Chen, X.; Li, Z.; Jiang, Y. Q.; Zhong, Z. M.; Wang, X. R.; Wong, K. Y. Spectrosc. Spectral. Anal. 2004, 22, 796–799. (d) Bukowski, R. M.; Davenport, M. D.; Titus, A. H.; Bright, F. V. Appl. Spectrosc. 2006, 60, 951–957. (11) (a) Meier, B.; Werner, T.; Klimant, I.; Wolfbeis, O. S. Sens. Actuators, B 1995, 29, 240–245. (b) Payra, P.; Dutta, P. K. Microporous Mesoporous Mater. 2003, 64, 109–118. (c) Schulz-Ekloff, G.; Wohrle, D.; van Duffel, B.; Schoonheydt, R. A. Microporous Mesoporous Mater. 2002, 51, 91–138. (12) Fernandez-Sanchez, J.-F.; Cannas, R.; Spichiger, S.; Steiger, R.; SpichigerKeller, U.-E. Anal. Chim. Acta 2006, 566, 271–282. (13) Robb, W. L. Ann. N.Y. Acad. Sci. 1968, 146, 119–137. (14) (a) Klimant, I.; Wolfbeis, O. S. Anal. Chem. 1995, 67, 3160–3166. (b) Chan, C.-M.; Chan, M.-Y.; Zhang, M.; Lo, W.; Wong, K.-Y. Analyst 1999, 124, 691–694. (c) Mills, A. Sens. Actuators, B 1998, 51, 60–68. (d) Xu, W.; McDonough, R. C.; Langsdorf, B.; Demas, J. N.; DeGraff, B. A. Anal. Chem. 1994, 66, 4133–4141. (15) (a) Orellana, G.; Moreno-Bondi, M. C.; Garcı´ a-Fresnadillo, D.; Marazuela, M. D. In Frontiers in Chemical Sensors: Novel Principles and Techniques; Springer: Berlin and Heidelberg, Germany, 2005; Vol. 3, pp 189-225. (b) Hartmann, P.; Leines, M. J. P.; Lippitsch, M. E. Sens. Actuators, B 1995, 29, 251–257. (16) Carraway, E. R.; Demas, J. N.; DeGraff, B. A. Langmuir 1991, 7, 2991– 2998. (17) (a) Lippitsch, M. E.; Pusterhofer, J.; Leiner, M. J. P.; Wolfbeis, O. S. Anal. Chim. Acta 1988, 205, 1–6. (b) Carraway, E. R.; Demas, J. N.; DeGraff, B. A.; Bacon, J. R. Anal. Chem. 1991, 63, 337–342. (c) Gewehr, P. M.; Delpy, D. T. Med. Biol. Eng. Comput. 1994, 32, 659–664. (18) Draxler, S.; Lippitsch, M. E.; Klimant, I.; Kraus, H.; Wolfbeis, O. S. J. Phys. Chem. 1995, 99, 3162–3167.

Langmuir 2010, 26(3), 2144–2150

Article

microscopy and two-photon absorption microscopy, but no emission lifetime data of the O2 indicator dye in such microdomains was reported.19 Since in oxygen-sensing applications the average emission intensity or luminesce lifetime is measured across the entire polymer indicator layer, no information concerning the microdomains can be obtained. New techniques such as confocal fluorescence lifetime imaging microscopy (FLIM) afford the opportunity of measuring emission lifetimes in local areas of indicator films with high spatial resolution.20 Therefore, FLIM displays particular potential for the study of the aforementioned domains in microheterogeneous samples such as optical O2 sensing films, thus providing valuable information toward the engineering of novel lifetime-based luminescent oxygen sensors. For this study a series of commercial room-temperaturevulcanization (RTV) silicones were chosen as supporting materials of the RD3 indicator dye. The addition of silica or glass powder (commonly referred to as filler) into the silicone rubber has the purpose of increasing the sturdiness of the polymer film.21 Incorporation of the filler has an effect on the permeability, solubility and diffusivity of oxygen within the matrix13 and, therefore, is expected to change the gas sensing properties of the luminescent polymer films. Moreover, silica has several other side effects and the overall result is more complex than a simple increase of the oxygen concentration into the matrix. For instance, in the case of ruthenium-based optical sensors, the concentration of the metal complex in the polymer film is dependent on the silica concentration.5 In this work, the effect of the silicone type, silica concentration and particle size on the emission lifetime-based oxygen sensing behavior of RD3-doped layers has been rationalized thanks to FLIM measurements.

Experimental Section Materials and Preparations. Sensors films were prepared using three different commercial RTV silicones, namely Dow Corning 732 (Wiesbaden, Germany; composition by weight: >60% dimethylsiloxane, 7-13% amorphous silica, 5-10% ethyltriacetoxysilane, 5-10% methyltriacetoxysilane), Dow Corning 3140 (Wiesbaden, Germany; composition by weight: >60% dimethylsiloxane, 10-30% “trimethylated” silica, 5-10% methyltrimethoxysilane)22 and Krafft 62833 (Guip uzcoa, Spain; exact composition not disclosed). RD3 was prepared as chloride salt according to the method previously published for similar homoleptic Ru(II) complexes.23 Silica was added in some cases as a filler. Three silica gel with different particle sizes (dp) were used namely flash column chromatography grade (Silica Gel 60, Merck, dp=40-63 μm),24 Partisil-5 (Whatman, dp=5 μm)25 and pyrogenic silica (fumed silica, Sigma 99.8%, dp = 0.007 μm).26 A thorough mixture of silicone and silica (0, 5 or 10% by weight) was prepared and applied by knife-coating onto a polyester sheet (85  25  0.1 mm, Goodfellow, Huntingdon, U.K.) to obtain a film after curing the mixture for 15 days at room temperature. Once prepared, the silicone layer is removed from (19) (a) Bowman, R. D.; Kneas, K. A.; Demas, J. N.; Periasamy, A. J. Microsc. 2003, 211, 112–120. (b) Bedlek-Anslow, J. M.; Hubner, J. P.; Carroll, B. F.; Schanze, K. S. Langmuir 2000, 16, 9137–9141. (20) (a) Gadella, T. W. J. FRET and FLIM techniques, 1st ed.; Elsevier: Amsterdam, 2009.(b) Rietdorf, J Microscopy Techniques 1st ed.; Springer: Heidelberg, Germany, 2005. (21) (a) Warrick, E. L.; Lauterbur, P. C. Ind. Eng. Chem. 1955, 47, 486–491. (b) Meyer, L. H.; Cherney, E. A.; Jayaram, S. H. IEEE Electr. Insul. Mag. 2004, 20, 13–21. (22) Data sheet given by manufacturer, www.dowcorning.com. (23) Garcia-Fresnadillo, D.; Georgiadou, Y.; Orellana, G.; Braun, A. M.; Oliveros, E. Helv. Chim. Acta 1996, 79, 1222–1238. (24) Data sheet given by the manufacturer: www.merck.de. (25) Data sheet given by the manufacturer: www.whatman.com. (26) Data sheet given by the manufacturer: www.sigmaaldrich.com.

DOI: 10.1021/la902546k

2145

Article

L opez-Gejo et al.

Table 1. Absorption and Emission Data from the Different Membranes Studied under Air as a Function of the Amount of the Pyrogenic Silica Added silicone

% wt added silica

a λabs max [nm]

Amaxb

[Ru] [mM]c

a λem max [nm]

IL max [au]b

0 454 0.016 0.022 608 275d 5 450 0.063 0.085 610 425d 10 448 0.44 0.60 611 862d DC732 0 446 0.46 0.62 611 240e 5 448 1.24 1.7 619 518e 10 449 2.91 3.9 636 443e a b c -1 -1 23 Uncertainty: ( 1 nm. Uncertainty: 5%. Approximate concentration of RD3 using εmax = 29500 L mol cm and no scattering contribution at the visible maxima (see text). d Exc/em slits: 10/7. e Exc/em slits: 10/5. DC3140

the polyester support mechanically or by dipping the ensemble into methylene chloride for a couple of seconds. The thickness of the membrane was determined with a digital micrometer giving a value of 250 ( 25 μm. The luminescent RD3 indicator dye is loaded into the silicone films by immersing them in a 0.25 mM solution of the dye in HPLC-grade dichloromethane for a few minutes. During this process the membrane swells and releases some unreacted polydimethylsiloxane. After drying in air for three days at room temperature, the membrane recovers its original dimensions. Silica-free silicone films were produced using the single component conformal coating 1-2577 from Dow Corning (composition by weigh: >60% dimethyl methylphenylmethoxysiloxane, 15-40% toluene, 3-7% methyltrimethoxysilane).22 A solution of 0.4 mg of RD3 in 2 mL of toluene or CH2Cl2 was prepared and 1.0 mL of this solution was mixed with 0.5 mL of the conformal coating. Some drops were placed over a glass slide and dried for 1 day under air at room temperature (200 ( 20 μm). Photochemical Characterization Techniques. Absorption spectra were recorded with a Cary 3-Bio (Varian Inc., Hansen Way, Palo Alto, CA) spectrophotometer using for each sample the corresponding Ru-free silicone film as reference. Luminescence spectra (λexc 450 nm) were recorded at room temperature with a LS50 spectrofluorometer (Perkin-Elmer, Uberlingen, Germany). The Horiba-Jobin-Yvon (Piscataway, NJ) DynaMic FLIM used to characterize the O2-sensing films includes an epifluorescence confocal microscope (Olympus BX51, NY) equipped with two objectives (10x and 40x) and a CCD camera (uEye UI-1450C, IDS, Germany) with 1600  1200 pixels for bright field image recording. A laser diode (H-J-Y NanoLED-470LH) was used as excitation source (463-nm peak wavelength; 900-ps pulse width; 100-, 50- or 10-kHz repetition rate). A 470-nm interference filter (Chroma HQ470/20x, Rockingham, VT) was placed in the excitation path and a 490-nm dichroic mirror (Olympus Q490DCXR) was placed in the 6-position cube turret. For highly emitting films, a neutral density filter (Thermo-Oriel, Stratford, CT; optical density 1.00) was used to attenuate the excitation beam. Since the emission maximum of the investigated membranes was around 630 nm, a 590 nm cutoff filter (CG-OG590-1.00-3, CVI Technical Optics, UK) was used to minimize the detection of scattered light in the emission beam. The H-J-Y FluoroHub controller for sequential single photon timing detection was interfaced to a H-J-Y TBX picosecond photon detection module equipped with a fast red-sensitive PMT detector with thermoelectric cooling for lower noise operation. Variable delay times with respect to the trigger pulse could be adjusted with an external gate and delay generator (Ortec 416A, TN). Luminescence decays were measured with a 10-, 20-, or 50-μs window by accumulating at least 5000 counts in the peak channel and emission lifetimes were obtained from triexponential curve fittings using the proprietary H-J-Y hybrid grid-search minimization algorithm (without deconvolution) for stable chi-squared minimization.27 Measurements were carried out under atmospheric pressure (711 ( 5 Torr) at least on three different spots over the sample surface (200-μm pinhole, 10 objective; approximately (27) Johnson, M. L. Anal. Biochem. 1992, 206, 215–225.

2146 DOI: 10.1021/la902546k

area of analysis = 300 μm2). Mean values for the three emission lifetime components and the corresponding pre-exponentially weighted emission lifetimes (τm) are reported throughout. To obtain the Stern-Volmer oxygen sensing plots, luminescence lifetime measurements under a controlled atmosphere were performed by introducing a small film square into a 0.4  1.0-mm optical paths thickened walls Suprasil emission cell (Hellma, M€ ulheim, Germany) fitted with a rubber septum, which allows N2/O2 mixtures from an electronic mass flow controller (ICP, Cantoblanco, Madrid) to be flowed in before and during the emission lifetime measurements. The cell was positioned horizontally under the objective of the confocal microscope.

Results and Discussion Incorporation of siliceous fillers such as glass powder, pyrogenic silica, kaolin, bentonite, sericite, etc. into silicone membranes is the most common way to increase the consistency of silicone materials to make them usable as films (sealants, adhesives, medical implants, etc.). State-of-the-art luminescent sensors for O2 measurements rely also on silicone thin films doped with long emission lifetime indicator dyes. The photophysical and photochemical (including O2 quenching) properties of the latter will be dependent on their immediate environment around the probe. Increased O2 and dye solubility, self-aggregation of the dye, gas diffusivity changes, restricted O2 accession and variations of the medium polarity, among others, are typical effects observed upon the addition of the filler. Nevertheless, some uncertainties appear in the literature concerning the location of dye and quencher in the film after addition of the filler, or the reason behind the shift of absorption and emission bands of the dye embedded in a silicone/silica polymer matrix.28 In order to shed some light onto these discrepancies, we have investigated first how the silica filler concentration, particle size and type influence the oxygen response of the sensor silicone films. Effect of Silica Concentration. Table 1 collects the absorption and emission data for the different RD3-dyed films prepared with selected commercial silicones. In all cases, as the (pyrogenic) silica filler content increases, a rise in the absorption at ca. 450 nm is observed. This band corresponds to the metal-to-ligand charge transfer (MLCT) absorption of the ruthenium complex.23 However, while the absorption maximum is located at 454 nm for the unreinforced DC3140 silicone, it shifts up to 8 nm to the blue after addition of silica. Such a hypsochromic shift must be attributed to a change in the polarity around the metal complex due to incorporation of the silica within the hydrophobic silicone. Similar blue shifts with a decrease of the solvent polarity have been observed previously for ruthenium polypyridyls due to the higher polarity of the MLCT excited state compared to the ground state.29 Lu et al. have measured shifts of the excitation spectrum of platinum octaethylporphine embedded in silicone films with the (28) Lu, X.; Manner, I.; Winnik, M. A. Macromolecules 2001, 34, 1917–1927. (29) Castro, A. M.; Delgado, J.; Orellana, G. J. Mater. Chem. 2005, 15, 2952– 2958.

Langmuir 2010, 26(3), 2144–2150

L opez-Gejo et al. Table 2. Emission Lifetime of RD3 in Different Silicone Films under Nitrogen as a Function of the Amount of Pyrogenic Silica Addeda silicone % wt added silica τ1 [μs] (%) τ2 [μs] (%) τ3 [μs] (%) τm [μs]b DC3140 0 0.37 (25) 2.77 (34) 7.40 (41) 4.04 5 0.42 (13) 3.40 (36) 7.62 (51) 5.19 10 0.39 (10) 3.72 (43) 7.85 (47) 5.30 DC732 0 0.51 (11) 3.16 (35) 7.47 (54) 5.17 5 0.54 (6) 3.82 (53) 7.87 (41) 5.26 10 0.51 (12) 2.98 (47) 6.21 (41) 4.02 a Maximum uncertainty: 15% for τ1, 5% for τ2, 3% for τ3 and 2% for τm; χ2 < 1.1 in every fit. b τm = Σt%tτt/100

addition of silica, but the authors attributed such band shifts just to scattering of the excitation light.28 In our case, scattering effects must be ruled out because we use a blank silicone film in the reference optical path of the double beam spectrophotometer. At the same time, by increasing the silica gel content a higher amount of ruthenium complex enters the film as evidenced from the increase in the absorption and emission intensities (Table 1). According to the absorption data, much more ruthenium dye is loaded into the DC732 than into to DC3140 silicone. This finding might be surprising if we take into account that the latter contains up to 30% of “trimethylated silica” as thickening agent (see Experimental Section) compared to the 7-13% of silica particles within the former one. Trimethylated silica is a unique form of fumed silica with alkoxy groups and trimethylsilyl groups at the individual particle surface made to improve the poor affinity of conventional fumed silica particles for a silicone oil and to avoid their aggregation and precipitation at the time of mixing or removal of the solvent.30 Therefore, even at high levels of filler, the hydrophobic trimethylated silica particles do not adsorb as much dicationic dye as the conventional (hydrophilic) silica does. Commercial DC732 silicone already contains ca. 10% of silica by weight so that further loading is added on top of this amount. This fact would also explain the behavior of the absorption maximum discussed above for both silicones. An increase in the intensity of the absorption band at 448 nm is observed for the DC732 films as more silica is introduced but, in this case, the luminescence intensity values do not show a monotonous increase along the series (Table 1): silicone with 10% of added silica shows a lower emission intensity and a 17 nm bathochromic shift compared to the silicone films with 5% added silica. The very high amount of ruthenium dye loaded into the thin film produces an overlap between the absorption and emission bands (see Figure S1 in the Supporting Information) with the consequence of self-absorption or an energy transfer (homotransfer) process. As a result, the luminescence intensity decreases in the region of spectral overlap and an apparent emission red-shift can be observed.31 With the aim of investigating the effect of silica on the emission lifetime of RD3 in the silicone films without interference of the O2 quenching, luminescence decays of the different films were recorded by SPT-FLIM in the absence of this gas. Table 2 shows the emission lifetime data for two silicone films with different pyrogenic silica content. Samples were measured three times on different spots in order to obtain information on the homogeneity of the sensing film. Multiexponential decays were always recorded revealing the presence of several microdomains where the sensing dye dwells within the film. Such decays were successfully fitted to (30) Satoshi Kuwata, A; Takahiro Goi, K.; Takaaki, S.; Tsutomu, O. US Pat. 4,983,388. (31) Valeur, B. Molecular Fluorescence, Principes and Applications, 1st ed.; Wiley-VCH: Weinheim, Germany, 2002; pp 110-112.

Langmuir 2010, 26(3), 2144–2150

Article Table 3. Emission Lifetime of RD3 in Different Silicone Films under Air (Patm = 911 Torr) as a Function of the Amount of Added Pyrogenic Silicaa silicone % wt added silica τ1 [μs] (%) τ2 [μs] (%) τ3 [μs] (%) τm [μs]b DC3140

0 0.20 (29) 1.09 (47) 3.30 (24) 1.36 5 0.25 (29) 0.88 (58) 2.93 (13) 0.96 10 0.28 (29) 0.84 (59) 2.58 (12) 0.88 DC732 0 0.37 (37) 0.95 (54) 2.68 (9) 0.88 5 0.33 (39) 0.73 (52) 1.96 (9) 0.68 10 0.34 (29) 0.92 (58) 2.39 (14) 0.95 KF62833 0 0.33 (29) 1.07 (52) 3.42 (19) 1.30 5 0.30 (38) 0.93 (51) 2.86 (11) 0.90 10 0.35 (41) 0.86 (51) 2.53 (8) 0.77 a Maximum uncertainty: 15% for τ1, 10% for τ2, 5% for τ3 and 5% for τm. b τm = Σt%tτt/100

a sum of three exponentials but indeed many more than just three microenvironments should exist in the samples. The shortest lifetime (τ1) might be associated with RD3 molecules in close proximity to other RD3 molecules so that strong luminescence self-quenching occurs (see below). The other two lifetimes might correspond to the centers of distributions from molecules which are in hydrophobic silicone-rich (τ2) or hydrophilic silica-rich environments (τ3). In fact, most of the samples show similar values for the mean emission lifetime (τm =5.3 ( 0.1 μs, Table 2). Just two samples fall well outside this range. The DC3140 film, which does not contain any added (highly polar) pyrogenic silica but (hydrophobic) trimethylated silica, shows a τm of 4.0 μs. The blue shift of its absorption maximum was attributed to the hydrophobicity of this film compared to the high polarity of the added pyrogenic silica surface. According to the Lippert-Mataga model of the solvent polarity effect on the emission lifetime,32 the less polar the environment, the shorter the lifetime. In fact, such a polarity-sensitive behavior of ruthenium complexes has already been investigated in solution and used to develop hydrocarbon-in-water luminescent sensors.29 In the case of the DC3140 with no added pyrogenic silica (Table 2), the contribution of the longer component attributed to luminescence emission of the dye in a polar environment, can only be explained by the presence of some amount of regular silica within the trimethylated silica added by the silicone manufacturer as filler.33 The other sample with a shorter lifetime value (4.0 μs) is the DC732 film with the maximum load of silica. As mentioned above, the addition of silica significantly increases the amount of the dye loaded into the film. In this case, the reduction of the lifetime from the average value may indicate the presence of additional excited state deactivation mechanisms at high concentrations of RD3. Deactivation of photoexcited ruthenium complexes bound to nanocrystalline TiO234 or adsorbed onto porous Vycor glass (PVG)35 and cellulose36 has been studied previously. Using a laser source it was found that a triplet-triplet annihilation process can take place when the ruthenium complex is bound to TiO2 nanoparticles. This is due to an efficient energy transfer from the excited state of the dye to spatially adjacent ground state dye molecules.34 A similar deactivation process has been observed in [Ru(bpy)3](PF6) crystals excited with a ca. 0.01-mJ cm-2 laser (32) (a) Lippert, E. Z. Electrochem. 1957, 61, 962–975. (b) Mataga, N.; Kaifu, Y.; Koizumi, M. Bull. Chem. Soc. Jpn. 1956, 29, 465–470. (33) The commercial “trimethylated silica” contains actually a minimum of 60% (by weight) of trimethylated silica (see www.dowcorning.com). (34) Kelly, C. A.; Farzad, F.; Thompson, D. W.; Meyer, G. J. Langmuir 1999, 15, 731–737. (35) Gafney, H. D. Coord. Chem. Rev. 1990, 104, 113–141. (36) Milosavljevic, B. H.; Thomas, J. K. J. Phys. Chem. 1983, 87, 616–621.

DOI: 10.1021/la902546k

2147

Article

pulse thanks to an energy migration through the crystal.37 In our case, triplet-triplet annihilation by efficient energy migration between chromophores adsorbed on the silica surfaces is proposed to be an important excited state deactivation pathway for high RD3 concentration in the absence of oxygen. In fact, we have found that the mean luminescence lifetime (τm) of a Ru(4,7diphenyl-1,10-phen)3Cl2 microcrystal under N2 is 0.65 ( 0.03 μs, which is in good agreement with values given by McGee and Mann for microcrystals of similar ruthenium complexes,38 evidencing the contribution of triplet-triplet annihilation processes to the mean excited state lifetime under high dye-loading experimental conditions. Luminescence lifetime values for the different silicone sensing films in the presence of molecular oxygen are collected in Table 3. As in the case of a nitrogen atmosphere, the emission decays required fit to a triexponential function. The existence of such a complex decay profile may obey to several reasons. Ruthenium complexes are located at different positions where O2 might have different accessibility. This possibility cannot be the only cause since it would imply an exponential decay in the absence of O2 instead of the complex pattern observed (Table 2). Therefore, the multiexponential emission decay profile must be also a consequence of the intrinsic heterogeneity of the dye distribution within the film. As far as the luminescence lifetime changes with the silica content is concerned, DC3140 and KF display a τm decrease upon addition of pyrogenic silica filler into the membrane. This fact might indicate a concomitant increase of the O2 concentration into the silicone since, in the absence of O2, no change of τm with the silica loading occurs (Table 2). Nevertheless, the addition of silica to silicone films has been shown to have a small effect on the oxygen solubility. For instance, while O2 has a solubility of 0.18 ( 0.01 cm3(STP)/cm3atm in pure poly(dimethylsiloxane) (PMDS),39 its solubility increases to 0.31 for PMDS filled with as much as 33% by weight silica.13 This raise on the O2 solubility does not explain itself the significant decrease of the dye emission lifetime upon incorporation of just 10% silica (1.36 to 0.88 μs, Table 3). Loading silica into the silicone film has two major effects with regard to molecular oxygen.5 First, it increases the concentration of the latter due to adsorption of molecular oxygen onto the surface of the microporous silica particles.40 At the same time, the silica particles embedded into the film perturb the diffusion of oxygen acting as an obstacle and increasing the diffusion path length of the gas. For instance, Cox and Dunn reported a decrease of the oxygen diffusion coefficient in polydimethylsiloxane from 3.56  10-5 to 2.60  10-5 cm2s-1 by addition of 5 wt % (pyrogenic) silica.41 The overall effect is a higher and longer presence of molecular oxygen into the film, leading to a higher probability of an encounter with the excited dye that reduces its lifetime. Moreover, the ruthenium complex is preferentially located into the pyrogenic silica particle surface (vide infra) leading to an enhancement of the O2 quenching due to the local concentration and reduced dimensionality effects.42 Concomitantly with the lifetime decrease, the luminescence intensity (37) Ikeda, N.; Yoshimura, A.; Tsushima, M.; Ohno, T. J. Phys. Chem. 2000, 104, 6158–6164. (38) McGee, K. A.; Mann, K. R. J. Am. Chem. Soc. 2009, 131, 1896–1902. (39) Merkel, T. C.; Bondar, V. I.; Nagai, K.; Freeman, B. D.; Pinnau, I. J. Polym. Sci., Part B: Polym. Phys. 2000, 38, 415–434. (40) (a) Lambert, B.; Peel, D. H. P. Proc. R. Soc. London, Ser. A 1934, 144, 205– 225. (b) Gavrilov, V. Y. Kinet. Catal. 2005, 46, 403–406. (41) Cox, M. E.; Dunn, B. Appl. Opt. 1985, 24, 2114–2120. (42) Klafter, J.; Drake, J. M. Molecular Dynamics in Restricted Geometrics; Wiley: New York, 1989.

2148 DOI: 10.1021/la902546k

L opez-Gejo et al. Table 4. Luminescence Lifetime Data under Air and Nitrogen of RD3Doped DC3140 Films Loaded with 10 wt % of Different Particle Sizes of Silicaa silica type b

% O2

τ1 [μs] (%)

τ2 [μs] (%)

τ3 [μs] (%)

τm [μs]c

flash 21 0.20 (27) 0.70 (61) 2.05 (12) 0.72 Partisil 21 0.27 (30) 0.75 (61) 2.33 (10) 0.75 pyrogenic 21 0.28 (29) 0.84 (59) 2.58 (12) 0.88 flash 0 0.33 (22) 2.98 (37) 7.57 (42) 4.31 Partisil 0 0.18 (23) 2.39 (25) 7.08 (52) 4.33 pyrogenic 0 0.39 (10) 3.72 (43) 7.85 (47) 5.30 a Maximum uncertainty: 15% for τ1, 10% for τ2, 5% for τ3 and 5% for τm. b See Experimental Section and Table 1 for details. c τm = Σt%tτt/100.

increases upon addition of silica (Table 1). Such an (apparently) contradictory result may be explained by the effect of the silica particles that increase at the same time the concentration of the luminescent molecules and the quenching efficiency. It should be stressed that the initial emission lifetime of RD3 in the DC732 silicone series (film with no added silica, 0.88 μs) is the same than that obtained in the DC3140 silicone with 10% pyrogenic silica load (Table 3). This result clearly indicates that the DC732 silicone has ca. 10% pyrogenic silica or the like within its original ingredients. In the later case, the reduction of lifetime is further observed after adding 5% of silica following the tendency of DC3140 and KF62833 silicone films. This result shows that the RD3 emission lifetime may be a valuable probe of the composition of silicone materials. Nevertheless, incorporation of 10% silica leads to 20% increase in the luminescence lifetime compared to the sample with just 5% of additional silica. As discussed above for this specific sample, an energy homotransfer process was observed by steadystate luminescence experiments (Table 1). These multiple reabsorption and re-emission processes are known to slow down the luminescence decay yielding longer-lived excited states.31 Although energy homotransfer is always present at this level of filler regardless the O2 concentration, such increase in the RD3 lifetime was not observed under nitrogen (Table 2) because triplet-triplet annihilation was the dominant process in the absence of a luminescence quencher. Effect of the Silica Particle Size. Three different types of hydrophilic silica were used in this study (Table 2) namely flash column chromatography grade (40-63 μm), HPLC Partisil (5 μm), and pyrogenic silica (0.007 μm). Table 4 gathers the emission lifetime data of silicone-embedded RD3 under air and nitrogen. As it can be observed, the mean lifetime (τm) of the RD3 indicator dye loaded into the silicone films increases with the filler particle size either in absence or in the presence of oxygen. No significative differences do exist between the 5- and 50-μm silica particles while an appreciable increase of τm is measured when silica nanoparticles are used. Since the same trend is observed under nitrogen and oxygen, the origin of those differences should correspond to an inherent property of the added silicone filler. The small size of the pyrogenic silica grains determines that many RD3 molecules are in contact with the silica surface producing a very polar environment and, consequently, a higher lifetime (see above). Larger silica particles possess less surface so that less indicator particles are in a “silica” environment and the contribution of RD3 molecules surrounded by silicone increases the shorter lifetime component. Effect of the Silicone Preparation. Comparing the emission lifetime data for films fabricated with 5% added silica to the three selected commercial prepolymerization mixtures (Table 3), it can be observed that the indicator lifetime in DC3140 and K62833 are Langmuir 2010, 26(3), 2144–2150

L opez-Gejo et al.

very similar, indicating that the amount of pyrogenic silica filler incorporated in the latter is very small (DC3140 contains no pyrogenic silica according to the manufacturer). As it has been discussed above, the high amount of pyrogenic silica in the DC732 composition is responsible for the lower emission lifetime of the RD3 in the presence of O2. In addition, important information on the silicone fillers can be obtained from the emission lifetime of RD3 acting as a molecular probe into the film. Solubility of oxygen into silica-loaded silicone may be described with a two-phase system where the silica and rubber act independently as sorbents.43 This is the case of (hydrophilic) pyrogenic silica suspended in the silicone (hydrophobic). In this case, O2 is simply dissolved in the polymer matrix according to Henry’s law, while it is adsorbed by the dispersed sieve phase following a Langmuir isotherm.44 However, when the filler has a hydrophobic character, such as the trimethylated silica included into the DC3140 silicone, the growing polydimethylsiloxane (PDMS) interacts extremely well with the filler during the manufacture stage, filling the pores of the silica particles and forming a unique phase. In this case, the solubility of oxygen into the film is described with a model in which the filler is regarded as completely wetted by the PDMS polymer and is thus a nonsorbent. The result is that, after addition of hydrophobic silica (trimethylated silica), the increase in oxygen concentration into the film is negligible while for hydrophilic (pyrogenic silica) the presence of air-filled pores into the particles yields a significant increase of the O2 contents of the film. These effects are clearly demonstrated on the long luminescence lifetime values observed for the neat DC3140 (30% hydrophobic silica) compared to the short values for the DC732 (10% pyrogenic silica). Addition of any type of silica particles to the silicone film should lead to similar decrease of oxygen diffusion coefficient. Since no lowering of the RD3 luminescence lifetime is observed in the presence of ca. 30% trimethylated silica for two silicones containing 10% regular silica (DC3140-10% vs DC732-0%, Table 3), it can be concluded that hindering of the diffusion has no significant effect on the quenching rate. Therefore, only the increase of the O2 solubility and quenching in a confined space (both due to the gas adsorption onto the pyrogenic silica) seem to be responsible for the enhancement of the oxygen sensitivity of the luminescent silicone films in the presence of pyrogenic silica fillers. Stern-Volmer O2 Quenching Plots. Once the effects of the silica content on the luminescence lifetime decay of the films have been discussed and the existence of microdomains has been evidenced, the following step is to determine to what extent the microheterogeneity influences the sensitivity and linearity of the sensor response to O2. To that end, rigorously silica-free silicone films were prepared from a commercial conformal coating (DC1-2577) using either toluene or dichloromethane as solvents for RD3. Films with high transparency and microscopic homogeneity are obtained with the latter solvent, while films prepared from toluene present microheterogeneity due to the low solubility of RD3 (see below). The Stern-Volmer plots for a film prepared with the commercial RTV silicone DC3140, a similar one with 10% additional pyrogenic silica and a silica-free film prepared with DC1-2577 and RD3 in dichloromethane are shown in Figure 1. As it has often been reported,5,14d incorporation of regular silica significantly increases the oxygen sensitivity compared to the RTV silicone and the conformal coating, even taking into account that the former contains up to 30% trimethylated silica. At the same (43) Barrer, R. M.; Barrie, J. A.; Raman, N. K. Polymer 1962, 3, 605–614. (44) Kemp, D. R.; Paul, D. R. J. Polym. Sci. 1974, 12, 485–500.

Langmuir 2010, 26(3), 2144–2150

Article

Figure 1. Stern-Volmer plots for the O2 quenching of commercial DC3140 silicone films ((), DC3140 with added 10% pyrogenic silica (9) and silica-free DC1-2577 silicone coating (2).

time, the linearity of the response function is lost, especially at high concentrations of O2. This departure from linearity may be ascribed to the presence of microdomains in the film as a consequence of predominant adsorption of the luminescent indicator dye on the pyrogenic silica surface. The slight curvature of the commercial silicone film mighty be explained by the presence of some amount of regular (i.e., non-methylated) silica in the preparations of “trimethylated” silica used as filler.30,33 Interestingly, it should be pointed out that all films, including the rigorously filler-free silicone one (data not shown), display a multiexponential luminescence decay at any O2 level. The possible reasons behind this behavior have been investigated by luminescence lifetime imaging microscopy. Luminescence Lifetime Imaging Analysis. Emission lifetime images of selected RD3-doped silicone films were recorded with the purpose of visualizing the proposed microdomains. Figure 2 shows the FLIM analysis of a rigorously silica-free film prepared with DC1-2577 silicone and a solution of the O2 indicator dye in toluene. As it can be observed, microcrystals of the ruthenium complex within the film can be detected in the bright field microscopy image (Figure 2a). The red luminescence of the metal complex is observed in the emission intensity image (Figure 2c) extracted from the FLIM experiment, where the crystal has a much higher intensity than the rest of the film. However, as far as the (average) lifetime image is concerned (Figure 2b), the luminescence lifetime measured in the crystal pixels is significantly lower than that measured for RD3 in the bulk silicone. This result is a consequence of the triplet-triplet annihilation phenomena discussed above that leads to emission lifetime quenching for highly concentrated domains. Even the smaller RD3 crystal in the upper part of the image (Figure 2a-c) can be recognized in both the intensity and lifetime images demonstrating the potential of FLIM in the characterization of luminescent sensing films. The striking contrast between the emission intensity and lifetime images underlines the necessity to keep the Ru(II) indicator dye concentration in the silicone film as low as possible for achieving the maximum luminescence lifetime and therefore, the maximum sensitivity to O2. Finally, a chromatography grade silica-containing film has been subject to FLIM (Figure 3). Measurements were performed in this case under dinitrogen to maximize the luminescence intensity and to remove the effect of the differential oxygen accessibility to the indicator dye. The flash chromatography silica DOI: 10.1021/la902546k

2149

Article

L opez-Gejo et al.

Figure 2. (a) Bright field, (b) lifetime (τm), and (c) intensity images of a silicone membrane (DC1-2577, 0.4 mg RD3 in 2 mL toluene) under N2. Acquisition time: 300 s/pixel. λexc = 470 nm; cutoff filter at 590 nm. Pinhole: 0.2 mm. Objective: 40.

Figure 3. (a) Bright field, (b) pre-exponential factor of second component (%2), and (c) intensity image of a silicone membrane (DC3140, 10% Flash SiO2) under nitrogen. Acquisition time: 300 s/pixel. λexc = 470 nm; cutoff filter at 590 nm. Pinhole: 0.2 mm. Objective: 10.

particles were chosen as filler because of their large particle size well above the resolution of the confocal microscope. The luminescence intensity image (Figure 3c) confirms that the sensing dye is preferentially located on the silica particles, since their shape can be easily recognized. The emission decays that conform the FLIM image pixels could be fitted to a biexponential function. The image in Figure 3b has been obtained using the pre-exponential factor of the longest-lived component. The reason behind this particular selection is to avoid distortion of the average luminescence lifetime by the important contribution of the light scattering to the short-lived component. In fact, an image constructed with the average lifetime does not show any recognizable silica particle. An additional advantage of such image is that it enhances the contrast between the emission of the RD3 dye adsorbed on the silica particles compared to that of RD3 embedded in the bulk silicone due to the longer lifetime of the former (see above). These pictures demonstrate the power of FLIM to investigate fine details of the intimacy of the O2 sensor films. In the case of silicone films filled with pyrogenic silica, no microdomains could be evidenced under the fluorescence microscope as particles with a 7-nm diameter lie far below its limit of detection. Nevertheless, the existence of microdomains follows from the nonexponential nature of the emission decays as it has been discussed above.

Summary and Conclusion Microenvironments with different luminescence lifetimes have been visualized for the first time in Ru-doped O2 sensing silicone films using FLIM. Incorporation of pyrogenic silica as filler allows a fine control of both the indicator luminescence kinetics and the O2 sensitivity. Such tuning is mainly related to the increase of the solubility of the luminescent dye and the gas into the silicone film. Moreover, coadsorption of quencher and quenchee onto the silica surface increases the probability of an encounter

2150 DOI: 10.1021/la902546k

between the two species, and therefore the O2 sensitivity, at the expense of a loss of linearity of the Stern-Volmer plot. The use of rigorously filler-free silicone leads to crystallization of the ionic ruthenium complex due to its low solubility in the polymer as it is evidenced by FLIM. In this case, the O2 sensitivity is poor even though the supporting material is highly permeable to the analyte. Interestingly enough, incorporation of hydrophobic silica does not help to increase the sensitivity. On the other hand, an excess of silica filler generates an overload of ruthenium complex that reduces the O2 sensitivity of the sensor film due to energy homotransfer and triplet-triplet annihilation of the photoexcited indicator dye. FLIM is shown to be a technique with high potential for the design and characterization of novel indicator films for the manufacturing of future luminescent sensors as it provides the link between the macroscopic and microscopic worlds. Further work is currently in progress in order to investigate such relationships in other Ru(II)-based sensing films (pH, waterborne hydrocarbons, humidity, etc.). Acknowledgment. This project has been funded by the Madrid Community Government (IV PRICYT ref CM-S-505/AMB/ 0374), the European Regional Development Fund, the European Social Fund, the Spanish Ministry of Science and Innovation (CTQ2006-15610-C02-01-BQU and TRA2007-30965-E) and the UCM-B. Santander (GR58-08-910072). G.O. gratefully acknowledges reception of an I3 Intensification of Research grant from the Madrid Community Government. We thank J. B. Navalon for preparation of the polymer films. Supporting Information Available: Figures showing absorption and emission spectra of low and high RD3-loaded films and luminescence lifetime decays of a typical film under different oxygen concentrations. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2010, 26(3), 2144–2150