Anal. Chem. 2003, 75, 2407-2413
Sol-Gel-Derived Sensor Materials That Yield Linear Calibration Plots, High Sensitivity, and Long-Term Stability Ying Tang, Elizabeth C. Tehan, Zunyu Tao, and Frank V. Bright*
Department of Chemistry, Natural Sciences Complex, University at Buffalo, The State University of New York, Buffalo, New York 14260-3000
Novel O2-sensing materials based on spin-coated noctyltriethoxysilane (Octyl-triEOS)/tetraethylorthosilane (TEOS) composite xerogel films have been synthesized and investigated. These sensors are based on the O2 quenching of tris(4,7-diphenyl-1,10-phenanthroline)ruthenium(II) ([Ru(dpp)3]2+) sequestered within the xerogels. Scanning electron microscopy and luminescence measurements (steady state and time resolved) have been used to investigate the structure of these films and their analytical figures of merit and determine the underlying reasons for their observed performance. The results show that certain [Ru(dpp)3]2+-doped Octyl-triEOS/TEOS composites form uniform, crack-free xerogel films that can be used to construct high-sensitivity O2 sensors that have linear calibration curves and excellent long-term stability. For example, an 11-month-old sensor based on 50 mol % Octyl-triEOS exhibits more than 4-fold greater sensitivity in comparison to an equivalent sensor based on pure TEOS. Over an 11-month time period, the sensitivity of a pure TEOS-based sensor drops by more than 400% whereas a sensor based on 50 mol % Octyl-triEOS remains stable (RSD ) 4%). An ideal chemical sensor will possess adequate selectivity, sensitivity, and detection limits, be easy to calibrate, and be robust and stable over the long term. Real sensors that meet all these ideals are rare. In the chemical sensing literature, O2 is a common target analyte because it is important in biological, environmental, and industrial applications. O2 is traditionally quantified by using a Clark electrode (CE);1 however, the CE consumes O2 and it can be poisoned by sample constituents (e.g., H2S, proteins, certain anesthetics).2 Given these limitations, researchers have expended substantial effort to develop optically based O2 sensors.3 The most common of these sensors exploit the well-known4 effects of O2 quenching on the intensity (I) or excited-state lifetime (τ) of an immobilized luminophore (fluorophore or phosphore). * Corresponding author: (voice) 716-645-6800, ext 2162; (fax) 716-645-6963; (e-mail)
[email protected]. (1) Clark, L. C., Jr. Trans. Am. Artif. Intern. Organs 1956, 2, 41-49. (2) Bacon, J. R.; Demas, J. N. Anal. Chem. 1987, 59, 2780-2785. (3) Demas, J. N.; DeGraff, B. A. In Topics in Fluorescence Spectroscopy; Lakowicz, J. R., Ed.; Plenum: New York, 1994; Vol. 4, Chapter 4. (4) Lakowicz, J. R. Principle of Fluorescence Spectroscopy, 2nd ed.; Kluwer Academic/Plenum Publishers: New York, 1999; Chapters 8 and 9. 10.1021/ac030087h CCC: $25.00 Published on Web 04/12/2003
© 2003 American Chemical Society
In homogeneous solution without complicating factors, luminophore quenching is described by the Stern-Volmer relationship:4
I0 τ0 ) ) 1 + KSV[Q] ) 1 + kqτ0[Q] I τ
(1)
which relates the ratio of the steady-state intensities or lifetimes in the absence of quencher (I0 and τ0) to the intensity or lifetime in the presence of quencher (I and τ) through the dynamic SternVolmer quenching constant, KSV, the bimolecular quenching rate constant that describes the collisional encounter kinetics between the luminophore and quencher, kq, and the quencher concentration, [Q]. For this ideal case, a plot of I0/I or τ0/τ versus [Q] (called the Stern-Volmer plot) will be linear with a slope equal to KSV and an intercept of unity. When luminophores are immobilized, as they are in a chemical sensor, eq 1 is generally not obeyed and the Stern-Volmer plots become nonlinear.2,3,5-10 The degree of nonlinearity depends on a wide variety of factors, and several models have been developed to help explain these results.2-5,11 However,the origin of the nonlinearity in every case is associated with the luminophore (5) (a) Carraway, E. R.; Demas, J. N.; DeGraff, B. A.; Bacon, J. R. Anal. Chem. 1991, 63, 337-342. (b) Xu, W.; McDonough, R. C.; Langsdorf, B.; Demas, J. N.; DeGraff, B. A. Anal. Chem. 1994, 66, 4133-4141. (c) Demas, J. N.; DeGraff, B. A.; Xu, W. Anal. Chem. 1995, 67, 1377. (6) (a) MacCraith, B. D.; McDonagh, C. M.; O’Keeffe, G.; Keyes, E. T.; Vos, Johannes G.; O’Kelly B.; McGilp, J. F. Analyst 1993, 118, 385-388. (b) McEvoy, A. K.; McDonagh, C. M.; MacCraith, B. D. Analyst 1996, 121, 785-788. (c) Mongey, K. F.; Vos, J. G.; MacCraith, B. D.; McDonagh, C. M.; Coates, C.; McGarvey, J. J. J. Mater. Chem. 1997, 7, 1473-1479. (d) McDonagh, C. M.; MacCraith, B. D.; McEvoy, A. K. Anal. Chem. 1998, 70, 45-50. (e) McDonagh, C. M.; Shields, A. M.; McEvoy, A. K.; MacCraith, B. D.; Gouin, J. F. J. Sol-Gel Sci. Technol. 1998, 13, 207-211. (f) Lavin, P.; McDonagh, C. M.; MacCraith, B. D. J. Sol-Gel Sci. Technol. 1998, 13, 641645. (g) Malins, C.; Fanni, S.; Glever, H. G.; Vos, J. G.; MacCraith, B. D. Anal. Commun. 1999, 36, 3-4. (h) Mongey, K. F.; Vos, J. G.; MacCraith, B. D.; McDonagh, C. M. Coord. Chem. Rev. 1999, 185-186, 417-429. (i) McDonagh, C. M.; Kolle, C.; McEvoy, A. K.; Dowling, D. L.; Cafolla, A. A.; Cullen, S. J.; MacCraith, B. D. Sens. Actuators, B 2001, 74, 124-130. (7) (a) Klimant, I.; Ruckruh, F.; Liebsch, G.; Stangelmayer, A.; Wolfbeis, O. S. Mikrochimica Acta 1999, 131, 35-46. (b) Wolfbeis, O. S.; Oehme, I.; Papkovskaya, N.; Klimant, I. Biosens. Bioelectron. 2000, 15, 69-76. (8) (a) Krihak, M.; Murtagh, M.; Shahriari, M. R. Proc. SPIE-Int. Soc. Opt. Eng. 1996, 2836, 105-115. (b) Krihak, M. K.; Shahriari, M. R. Electron. Lett. 1996, 32, 240-242. (c) Shahriari, M. R. In Optical Fiber Sensor Technology; Grattan, K. T. V., Meggitt, B. T., Eds.; Kluwer Academic: London, 1998; Vol. 4. (d) Murtagh, M. T.; Kwon, H. C.; Shahriari, M. R.; Krihak, M.; Ackley, D. E. J. Mater. Res. 1998, 13, 3326-3331. (e) Murtagh, M. T.; Shahriari, M. R.; Krihak, M. Chem. Mater. 1998, 10, 3862-3869.
Analytical Chemistry, Vol. 75, No. 10, May 15, 2003 2407
molecules being distributed simultaneously between two or more sites that exhibit different kq or τ0 values. The upshot of these nonlinear Stern-Volmer plots is that any sensor based on such a platform will require a multipoint calibration strategy. In addition, if one’s sensor response is not stable/reproducible over time, the issue of calibration and drift can completely derail development of a reliable sensor platform. Over the years, a plethora of methods have been described to immobilize sensing chemistries.12 These methods include physisorption, covalent attachment, and entrapment/sequestration. Physisorption methods are the most simple; however, several disadvantages exist. Random orientation of the sensing chemistry on the substrate can lead to target analyte inaccessibility or distributions of accessibilities. In addition, because this method lacks covalent chemical bonds, the “immobilized” sensing chemistry often leaches/desorbs, causing drift problems. Certain lifetime-based sensor strategies6,7 can address some of these issues, but these approaches require more costly and complex instrumentation. Covalent attachment strategies eliminate the leaching problem; however, they involve more complex chemistry, they tend to be more time-consuming and costly, and they do not guarantee full accessibility nor preclude drift due to surface reorganization, denaturation of protein-based sensing chemistries, or both. Sequestration of the recognition chemistry within a porous, three-dimensional network has become an attractive means to immobilize sensing chemistries.13 Sol-gel processing methods13 have been used to alleviate several of the aforementioned immobilization problems. There have been a significant number of luminescence-based O2 sensors reported in the literature based on sol-gel-derived materials.6-10 However, a common feature of these sol-gel-derived O2 sensors is that the Stern-Volmer plots are nonlinear, the sensor response is not stable over the long term, or both. (9) (a) Liu, H.-Y.; Switalski, S. C.; Coltrain, B. K.; Merkel, P. B. Appl. Spectrosc. 1992, 46, 1266-1272. (b) Innocenzi, P.; Kozuka, H.; Yoko, T. J. Phys. Chem. B 1997, 101, 2285-2291. (c) Maruszewski, K.; Jasiorski, M.; Salamon, M.; Strek, W. Chem. Phys. Lett. 1999, 314, 83-90. (d) Choi, M. M. F.; Xiao, D. Analyst 1999, 124, 695-698. (e) Choi, M. M. F.; Xiao, D. Anal. Chim. Acta 2000, 403, 57-65. (f) Jian, Y.-Q.; Li, Z.; Zhong, Z.-M.; Chen, X.; Wang, X.-R. Chem. Res. Chin. Univ. 2001, 17, 374-379. (g) Xu, H.; Aylott, J. W.; Kopelman, R.; Miller, T. J.; Philbert, M. A. Anal. Chem. 2001, 73, 41244133. (h) Chen, X.; Zhong, Z.; Li, Z.; Jiang, Y.; Wang, X.; Wong, K. Sens. Actuators, B 2002, 87, 233-8. (10) (a) Dunbar, R. A.; Jordan, J. D.; Bright, F. V. Anal. Chem. 1996, 68, 8, 604-610. (b) Watkins, A. N.; Wenner, B. R.; Jordan, J. D.; Xu, W. Y.; Demas, J. N.; Bright, F. V. Appl. Spectrosc. 1998, 52, 750-754. (c) Baker, G. A.; Wenner, B. R.; Watkins, A. N.; Bright, F. V. J. Sol-Gel Sci. Technol. 2000, 17, 71-82. (11) Lehrer, S. S. Biochemistry 1970, 10, 3254. (12) (a) Weetall, H. H. Immobilized Enzymes, Antigens, Antibodies, and Peptides: Preparation and Characterization; Marcel Dekker: New York, 1975; Chapter 6, pp 246-292, and Chapter 8, pp 419-497. (b) Taylor, R. F. Protein Immobilization: Fundamentals and Applications; Marcel Dekker: New York, 1991; Chapter 8, pp 263-303. (13) (a) Brinker, C. J.; Scherer, G. W. Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing; Academic Press: New York, 1990. (b) Hench, L. L.; West, J. K. Chem. Rev. 1990, 90, 33-72. (c) Dave, B. C.; Dunn, B.; Valentine, J. S.; Zink, J. I. Anal. Chem. 1994, 66, 1120A-1127A. (d) Avnir, D.; Braun, S.; Lev, O.; Ottolenghi, M. Chem. Mater. 1994, 6, 1605-1614. (e) Avnir, D. Acc. Chem. Res. 1995, 28, 328-341. (f) Ingersoll, C. M.; Bright, F. V. CHEMTECH 1997, 27, 26-35. (g) Chen, Q.; Kenausis, G. L.; Heller, A. J. Am. Chem. Soc. 1998, 120, 4582-4585. (h) Brennan, J. D. Appl. Spectrosc. 1999, 53, 106A-121A. (i) Bradshaw, J. T.; Mendes, S. B.; Saavedra, S. S. Anal. Chem. 2002, 74, 1751-1759. (j) Jin, W.; Brennan, J. D. Anal. Chim. Acta 2002, 461, 1-36.
2408
Analytical Chemistry, Vol. 75, No. 10, May 15, 2003
In this paper, we report the analytical figures of merit for a series of luminophore-doped xerogels that are composed of n-octyltriethoxysilane (Octyl-triEOS) and tetraethyoxysilane (TEOS). The luminophore is tris(4,7-diphenyl-1,10-phenanthroline)ruthenium(II) ([Ru(dpp)3]2+), and the target analyte is O2. Our results show that these composite xerogels can be tailored to produce uniform, uncracked sensing layers that simultaneously exhibit high sensitivity, long-term stability, and linear response curves (i.e., obey eq 1). THEORY SECTION Equation 1 describes the idealized behavior of a luminophore with a single excited-state lifetime in a homogeneous environment undergoing dynamic quenching. In the situation where luminophore molecules are simultaneously distributed in an ensemble of domains, each luminophore/site combination can exhibit its own characteristic quenching behavior depending on its local environment. Under these circumstances, eq 1, for m domains/ sites, becomes2,3,5
I0 I
)
[
m
∑1 + K i)1
]
-1
fi SV,i[O2]
(2)
where fi denotes the fractional contribution to the total emission from domain/site i and KSV,i is the Stern-Volmer quenching constant associated with domain/site i. In this multisite model, the luminophore time-resolved intensity decay kinetics are given by m
I(t) )
∑R e i
-t/τi
(3)
i)1
In this expression, Ri denotes the preexponential amplitude associated with the ith component, τi is the excited-state luminescence lifetime of the ith component, and m is the number of discrete single-exponential components required to describe the time-resolved intensity decay kinetics. Thus, for a luminophore distributed between m sites, one can, in principle, anticipate there being m excited-state luminescence lifetimes and m bimolecular quenching constant. If eq 2 is recast for the special case where m ) 2, one has the familiar Demas “two-site” model.2,3,5 In the case of two primary luminophore sites with only one being accessible/responsive to quencher (i.e., KSV,2 ) 0), eq 2 collapses to the Lehrer expression.11 The Demas and Leher models are nonlinear. EXPERIMENTAL SECTION Chemical Reagents. Tris(4,7′-diphenyl-1,10′-phenanathroline)ruthenium(II) chloride pentahydrate was purchased from GFS Chemicals, Inc. and purified as described in the literature.14 TEOS and Octyl-triEOS were purchased from United Chemical Technologies. HCl was obtained from Fisher Scientific Co. EtOH was a product of Quantum Chemical Corp. All reagents were used as received unless mentioned otherwise. Deionized water was prepared to a specific resistivity of at least 18 MΩ‚cm by using a Barnstead NANOpure II system. (14) Lin, C.-T.; Bottcher, M. C.; Creutz, C.; Sutin, N. J. Am. Chem. Soc. 1976, 98, 6536-6544.
Preparation of [Ru(dpp)3]2+-Doped Octyl-triEOS/TEOS Composite Xerogel Sensing Films. A pure TEOS-derived sol was prepared by mixing TEOS (3.345 mL, 15 mmol), water (0.54 mL, 30 mmol), EtOH (3.4 mL, 60 mmol), and HCl (15 µL of 0.1 M HCl, 0.0015 mmol). This sol solution was then capped and magnetically stirred under ambient conditions for 6 h. This particular formulation was chosen because it is representative of TEOS-based xerogels reported in the literature.6c,d,h,8d,e,9b,10 The Octyl-triEOS/TEOS composite sols were prepared by mixing TEOS and Octyl-triEOS (6.5 mmol in total) together to form solutions that contained 20, 40, 50, 60, or 80 mol % OctyltriEOS. These particular precursors were selected to provide a wide range of physicochemical properties in the final xerogels. To each of these sol solutions we added EtOH (1.25 mL, 22 mmol) and HCl (0.4 mL of 0.1 N HCl, 0.04 mmol). These solutions were then capped and magnetically stirred under ambient conditions for 1 h. These solutions were then diluted 1:1 (v/v) with EtOH. We found that the EtOH dilution step was necessary to lower the overall solution viscosity, slow the onset of gelation, and improve the final spin-coated film quality. The luminophore-doped sol solutions were prepared by mixing 60 µL of 2 mM [Ru(dpp)3]2+ (in EtOH) with 540 µL of the corresponding sol solution (vide supra). Blanks were prepared by omitting the [Ru(dpp)3]2+. These sol mixtures were capped and magnetically stirred under ambient conditions for 10 min prior to spin casting. Xerogel films were formed by spin casting15 onto 2.5 cm × 2.5 cm glass microscope slides. Each slide was first cleaned by soaking in 1 M NaOH for 24 h. All slides were rinsed with copious amounts of deionized water and EtOH and dried under ambient conditions. Films were formed by delivering 100 µL of a given sol solution onto a glass slide placed in the spin coater. The spin coater was then engaged and the rotational velocity adjusted to 3000 rpm. Spinning was continued for 30 s. All films were stored in the dark under ambient conditions for the long-term aging studies. The aging time clock begins immediately after a film is cast. Experiments were conducted over an 11-month period. Profilometry measurements were performed at regular time intervals on films that had aged for between 1 week and 11 months. There was no more than a 10% change in the individual film thickness over 11 months. Samples and blanks were prepared in triplicate on five separate occasions by using fresh reagent batches. The average and standard deviation for all measurements (n ) 15) are reported. The blank contribution was 20% OctyltriEOS. The linearity of the Stern-Volmer plots open a door to simple two-point calibration strategies. Third, the recovered KSV values (Table 1) for the pure TEOS-based xerogel films decrease as the film age increases. The recovered KSV values for the Octyl-
f1b
τ1 (ns)
1.00 0.46 ( 0.01 1.00 0.25 ( 0.01 1.00 1.00 1.00
4773 ( 18 2741 ( 48 5660 ( 11 2983 ( 59 5569 ( 2 5645 ( 4 5453 ( 8
τ2 (ns) 6872 ( 77 6668 ( 32
r2 0.6585 0.9967 0.9872 0.9929 0.9993 0.9982 0.9978
Samples were aged for 3 months. b f1 + f2 ) 1.
triEOS/TEOS composite xerogel films do not suffer this problem; they exhibit excellent long-term stability. This result is entirely consistent with Figure 3. Finally, inspection of Figure 4 shows that the error bars associated with the pure TEOS-based xerogel films are 3-10-fold larger in comparison to the Octyl-triEOS/ TEOS composite xerogel films. This result demonstrates that the film-to-film reproducibility is significantly better for the composite xerogels in comparison to the pure TEOS-based xerogels. In Figure 5, we report the effects of xerogel composition on the average Stern-Volmer quenching constant, . These results show that increases as we increase the mole percent Octyl-triEOS in the xerogel. To explore the origin of the improved sensitivity, we carried out a series of time-resolved intensity decay experiments. Time-Resolved Intensity Decays. Figure 6 presents a representative series of time-resolved intensity decay traces for 3-month-old [Ru(dpp)3]2+-doped xerogels when they are maintained in a pure N2 atmosphere along with fits to single (‚‚‚) and double (s) exponential decay models. The recovered decay terms are collected in Table 2. These results clearly demonstrate that the [Ru(dpp)3]2+ time-resolved intensity decay is multiexponential in the pure TEOS-based xerogel. This is a common scenario for luminophores doped in a wide variety of host matrixes.5-10 As the mole percent of Octyl-triEOS in the composite xerogel increases, the time-resolved intensity decay traces become single exponential. To the best of our knowledge, these represent the first luminophore-doped xerogels that exhibit (those >20 mol % OctyltriEOS) purely single-exponential intensity decay traces and linear Stern-Volmer plots (Figure 4). Taken together, these results Analytical Chemistry, Vol. 75, No. 10, May 15, 2003
2411
Figure 6. Typical time-resolved intensity decay traces and fits (lines) for 3-month-old [Ru(dpp)3]2+-doped Octyl-triEOS/TEOS composite xerogels in a pure N2 environment. The [Ru(dpp)3]2+ intensity decay is clearly biexponential for the pure TEOS and 20 mol % Octyl-triEOS composite xerogels. The [Ru(dpp)3]2+ intensity decay is single exponential in those xerogels that contain >20 mol % Octyl-triEOS.
argue that the microenvironment surrounding the [Ru(dpp)3]2+ molecules is homogeneous within those Octyl-triEOS/TEOS xerogel composites that contain more than 20 mol % Octyl-triEOS. The fact that the Stern-Volmer plots appear to be “linear” for the 20 mol % Octyl-triEOS xerogels (Table 1) while the intensity decay traces are clearly multiexponential is a manifestation of the increased information content of the time-resolved measurements in comparison to a simple intensity measurement. The sensitivity of any quenchometric O2 sensor depends on the Stern-Volmer quenching constant, KSV, which in turn depends on two factors (eq 1): τ0 and kq. Figure 7 summarizes the effects of Octyl-triEOS mole percent on the average τ0 and kq (i.e., and ) values for 3-month-old xerogels. These results show that increases from ∼5 µs in the pure TEOS-based xerogel film to ∼6 µs for 60 mol % Octyl-triEOS/TEOS composite, a 20% increase. Thus, the increase in the [Ru(dpp)3]2+ luminescence lifetime alone is not entirely responsible for the 50% increase in O2 sensitivity for the Octyl-triEOS/TEOS composites. The remaining cause of the increased sensitivity (30%) arises from an increase in . Thus, 40% of the total improvement in O2 sensitivity arises from an increase in [Ru(dpp)3]2+’s excited-state lifetime within the Octyl-triEOS/TEOS xerogel composites and 60% comes about from a concomitant increase in (i.e., O2 transport within the Octyl-triEOS/TEOS xerogel composites). 2412 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003
Figure 7. Effects of xerogel composition on the average [Ru(dpp)3]2+ excited-state fluorescence lifetime (〈τ〉 ) ∑(fi τi)) and the bimolecular quenching constant. The films are 3 months old.
CONCLUSIONS We report on the quenchometric behavior of [Ru(dpp)3]2+ sequestered within a series of Octyl-triEOS/TEOS composite xerogel films. Upon adding Octyl-triEOS to TEOS, a number of changes occur. First, the quality of the films improve from being cracked (pure TEOS) to being uniform and crack free (composites that contain 20-50 mol % Octyl-triEOS). Second, the O2 sensor
long-term stability is improved significantly by adding the OctyltriEOS component. Over a period of 11 months, the sensitivity for a pure TEOS-based sensor drops by 400% whereas a 50 mol % Octyl-triEOS/TEOS composite xerogel remains unchanged (RSD ) 4%). Third, the average O2 sensitivity (IN2/IO2) increases as we increase the mole percent Octyl-triTEOS within the composite. In comparison to an 11-month-old TEOS xerogel, a 50 mol % OctyltriEOS/TEOS composite xerogel exhibits a greater than 4-fold improvement in sensitivity. For 3-month-old films, 40% of the observed increase in the observed O2 sensitivity comes about from an increase in the [Ru(dpp)3]2+’s excited-state lifetime and 60% arises from a concomitant increase in O2 transport within the Octyl-triEOS/TEOS xerogel composites (i.e., ). Fourth, based on the Stern-Volmer plots and the time-resolved intensity decay traces, the [Ru(dpp)3]2+ microenvironment changes from
being (1) heterogeneous in a TEOS xerogel (described by the Demas two-state model and double-exponential intensity decay law) to (2) homogeneous in any Octyl-triEOS/TEOS xerogel composite with >20 mol % Octyl-triEOS (described by the SternVolmer model and single-exponential intensity decay law). Overall, the results of our experiments demonstrate a sol-gel-derived sensor platform that is uniform, exhibits excellent long-term stability and improved sensitivity, and offers simplified calibration. We are currently exploring the potential of these formulations with our chemical sensor array platforms.17
(17) (a) Cho, E. J.; Bright, F. V. Anal. Chem. 2001, 73, 3289-3293. (b) Cho, E. J.; Bright, F. V. Anal. Chem. 2002, 74, 1462-1466. (c) Cho, E. J.; Bright, F. V. Anal. Chim. Acta 2002 470, 101-10. (d) Cho, E. J.; Tao, Z.; Tehan, E. C.; Bright, F. V. Anal. Chem. 2002 74, 6177-84.
Received for review March 4, 2003. Accepted March 6, 2003.
ACKNOWLEDGMENT This work was generously supported by the National Science Foundation and Office of Naval Research.
AC030087H
Analytical Chemistry, Vol. 75, No. 10, May 15, 2003
2413
