Anal. Chem. 2001, 73, 2097-2103
Dual Lifetime Referencing as Applied to a Chloride Optical Sensor Christian Huber, Ingo Klimant,*,† Christian Krause, and Otto S. Wolfbeis
Institute of Analytical Chemistry, Chemo- and Biosensors, University of Regensburg, 93040 Regensburg, Germany
A membrane with an optical response to chloride has been developed that contains two luminophores that display two largely different decay times. The first luminophore (the “reference”) is a chloride-insensitive ruthenium metal-ligand complex possessing a decay time in the microsecond range. The second luminophore is the shortlived chloride-quenchable fluorescent probe lucigenin. Both are contained in a hydrogel matrix and are excited by a blue LED emitting sinusoidally modulated light. Under these conditions, the chloride-dependent fluorescence intensity of lucigenin can be converted in an analytedependent fluorescence phase shift that depends on the ratio of the two luminescence intensities and can be measured at modulation frequencies of typically 45 kHz. The dynamic range of this sensor can be adjusted by either varying the ratio of the two luminophores or selecting a particular optical filter combination. The determination of chloride as the major extracellular anion is routinely performed in clinical chemistry, and several approaches, including potentiometric and optical methods, are known.1-9 Optical chloride sensing can be based on the coextraction of chloride along with protons into a lipophilic phase.1-5 A major disadvantage of these kinds of sensors is their intrinsic strong pH dependence, which limits applicability to samples of unknown pH. This limitation has recently been overcome by coextracting a cationic solvatochromic dye (rather than a proton) along with chloride into a lipophilic phase.6 Another way to eliminate cross-sensitivity to pH is to use fluorescent chloridesensitive probes that undergo dynamic quenching in matrixes of hydrophilic polymers;7,8 however, all of these sensors use fluorescence intensity as the analytical information. The measurement of intensity is simple in terms of instrumentation, but its accuracy † New affiliation: Institute of Analytical Chemistry, University of Technology, 8010 Graz, Austria. (1) Rothmaier, M.; Simon, W. Anal. Chim. Acta 1993, 271, 135-141. (2) Xiao, K. P.; Bu ¨ hlmann, P.; Nishizawa, S.; Umezawa, Y. Anal. Chem. 1997, 69, 1038-1044. (3) Wang, E.; Meyerhoff, M. E. Anal. Chim. Acta 1993, 283, 673-682. (4) Tan, S. S. S.; Hauser, P. C.; Chaniotakis, N.; Suter, G.; Simon, W. Chimia 1989, 43, 257-251. (5) Tan, S. S. S.; Hauser, P. C.; Wang, K.; Fluri, K.; Seiler, K.; Rusterholz, B.; Suter, G.; Kru ¨ ttli, M.; Spichinger, U. E.; Simon, W. Anal. Chim. Acta 1991, 255, 35-44. (6) Huber, Ch.; Werner, T.; Krause, Ch.; Leiner, M. J. P.; Wolfbeis, O. S. Anal. Chim. Acta 1999, 398, 137-143. (7) Urbano, E.; Offenbacher, H.; Wolfbeis, O. S. Anal. Chem. 1984, 56, 427429. (8) Habib Jiwan, J.-L.; Soumillion, J.-Ph. J. Non-Cryst. Solids 1997, 220, 316322. (9) Huber, Ch. Ph.D. Thesis, University of Regensburg, 1999.
10.1021/ac9914364 CCC: $20.00 Published on Web 03/30/2001
© 2001 American Chemical Society
often is compromised by drifts in the optoelectronical setup, loss of light in the optical path, and is difficult in the case of turbid samples. Therefore, efficient referencing methods are sought for precise measurement of intensity. Among those, ratiometry, that is, the measurement of fluorescence intensity at two or more wavelengths of a single fluorophore or a fluorophore plus an added standard, is common.10 Alternatively, decay time can be measured, which is advantageous, because it is virtually independent of the overall signal intensity. Fluorescence decay times can be measured down to picoseconds, but longer lifetimes are preferred for the sake of instrumental simplicity and resolution.11 Decay-time sensing of chloride based on dynamic quenching shows decay times in the lower nanosecond range. Lucigenin, for example, displays a decay time of 20 ns in its unquenched state, which decreases to about 1 ns on exposure to 25 mmol‚L-1 of chloride.12 Hence, modulation frequencies in the MHz range are needed. In previous work, a chloride sensor was obtained by dissolving a chloride carrier along with a pH-sensitive absorber and an analyte-insensitive luminescent ruthenium metal complex with a luminescence lifetime of 1.1 µs in a poly(vinyl chloride) membrane. The sensing scheme exploits the radiative energy transfer that occurs, to a chloride-dependent extent, from an absorber to the ruthenium complex (the donor).13 Again, this sensor suffers from a strong pH dependence, a significant cross-sensitivity to oxygen and a rather poor photostability of the acceptor. Recently, the combination of an analyte-insensitive µs-lifetime luminophore with an analyte-sensitive nanosecond-lifetime fluorophore has been presented as a general logic to reference fluorescence intensity signals.14-16 The fluorescence intensity information can be converted into either a phase shift or a timedependent parameter. In this contribution, we describe the application of this scheme to a chloride sensor. Specifically, lucigenin-doped hydrogel beads are used along with polyacrylonitrile beads containing the ruthenium luminophore Ru(dpp), both contained in a polyurethane hydrogel. (10) Parker, J. W.; Laksin, O.; Yu, C.; Lau, M.-L.; Klima, S.; Fisher, R.; Scott, I.; Atwater, B. W.; Anal. Chem. 1993, 65, 2329-2334. (11) Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Plenum Press: New York/London, 1999. (12) Huber, Ch.; Fa¨hnrich K.; Krause Ch.; Werner, T. J. Photochem. Photobiol. 1999, 128, 111-120. (13) Huber, Ch.; Werner, T.; Krause, Ch.; Klimant, I.; Wolfbeis, O. S. Anal. Chim. Acta 1998, 364, 143-151. (14) Klimant, I.; Wolfbeis, O. S. 4th European Conference on Optical Sensors & Biosensors (Europt(r)ode IV), book of abstracts: Mu ¨ nster, Germany 1998, 125-126. (15) Klimant, I. (inv.) German Patent Application DE 198.29.657, Aug. 1, 1997. (16) Lakowicz, J. R.; Castellano, F. N.; Dattelbaum, J. D.; Tolosa, L.; Rao, G.; Gryczynski, I. Anal. Chem. 1998, 70, 5115-5121.
Analytical Chemistry, Vol. 73, No. 9, May 1, 2001 2097
Table 1. Composition of Sensor Membranes M1-M5.a membrane
lucigenin/Hypan conjugate [mg]
Ru(dpp)/PAN beads [mg]
M1 M2 M3 M4 M5
500 500 500 500 0
0 14 22 32 14
Figure 1. Schematic presentation of the membrane. The polyester film serves as an inert and optically transparent mechanical support.
a The quantities indicated were added to 3.7 g of a 10% hydrogel solution in ethanol/water (9:1, w/w).
EXPERIMENTAL SECTION Chemicals. Lucigenin (bis-N-methylacridinium dinitrate) was from Molecular Probes (Eugene, OR), and ruthenium(II)tris(4,7diphenyl-1,10-phenanthroline) [referred to as Ru(dpp)] was synthesized as described.17 Hypan hydrogel (HN80) was from Kingston Technology Inc. (Dayton, NJ), polyacrylonitrile (PAN) from Hoechst (Kelheim, Germany), the polyurethane hydrogel D4 from CardioTech (Ringo, NJ). The inorganic salts, ethanol, and dimethylformamide (DMF), all of analytical grade, were obtained from Merck (Darmstadt, Germany). The pH of the sample solutions was adjusted to 7.1 using phosphate buffer (cH2PO4- + cHPO42- ) 13 mmol‚L-1). Aqueous solutions were prepared from air-saturated, double-distilled water. The polyester support (prod. no. LS 146585) was obtained from Goodfellow (Cambridge, U.K.). Membrane Preparation. (A) Preparation of Ru(dpp)/PAN Beads. 1 g of PAN and 20 mg of Ru(dpp) were dissolved in 100 mL of DMF. Water (200 mL) was added dropwise under vigorous stirring. 20% aqueous sodium chloride solution (10 mL) was then added to the clear suspension to precipitate PAN particles containing Ru(dpp). The particles were washed three times with 100 mL of water, suspended in 100 mL of water, and heated to 70 °C to remove DMF. Afterward, the particles were washed again with water. (B) Preparation of Lucigenin/HN80 Beads. Hypan (5 g, granular size < 100 µm) was dispersed in 100 mL of an aqueous 500 µM lucigenin solution. Lucigenin was immobilized to Hypan beads as described.9 The beads were washed with a 200 mmol‚L-1 sodium fluoride solution until the filtrate showed no yellow coloration. This was followed by washing first with water to remove the sodium salt, then with dry ethanol and dry ether. The powder was air-dried and sieved to -1.5 0.6
Figure 9. Calibration plot of membrane M3 for selected anions.
in this experiment are shown in Figure 4 and the calibration plots of membrane M3 when using emission filters of different cutoff wavelengths in Figure 7. Again, the normalized plots of cot Φ vs chloride concentration are identical to plots in which no reference luminophore was added (Figure 7D). The fact that the emission spectra of lucigenin and the reference luminophore are different is advantageous, because the fractional intensity of the fluorophore and the luminophore can be modified by changing the emission filter; however, it is essential that the fluorophore and the luminophore have at least partial spectral overlap. Selectivity. The fluorescence of lucigenin is quenched by, in increasing order, chloride, bromide, salicylate, thiocyanate, and iodide. On exposure to 100 mM iodide, the fluorescence of lucigenin is completely quenched, which results in a phase angle that is almost the same as that of plain Ru(dpp)/PAN. The phase angle of membrane M3 increases slightly on exposure to phosphate, sulfate, nitrate, and fluoride, thus indicating a small effect on lucigenin fluorescence. Figure 8 shows the cross-sensitivities of the major interfering agents relative to a buffer containing 200 mM fluoride solution; Figure 9, the respective calibration plots. opt Selectivity coefficients log KCl,A relative to chloride were determined by the separate solution method (SSM) and are listed in Table 4, along with the points of inflection and the phase shift of membrane M3 when exposed to 50 mM of the interfering anion. As reported previously,12 oxygen has no effect on the fluorescence of lucigenin; however, ruthenium metal-ligand complexes are known to be dynamically quenched by oxygen.17 Therefore, calibration plots of M3 were measured both in air-saturated and in deoxygenated buffers. The phase angle increases by about 1° 2102 Analytical Chemistry, Vol. 73, No. 9, May 1, 2001
a log KOpt is defined as the log of the ratio of the concentrations Cl,A giving the same signal referred to 50 mM chloride. b On exposure to 50 mM of the respective anion.
on changing from air-saturated to deoxygenated solutions, independent of the respective chloride concentration. Even such a rather small quenching effect results in a decrease of the point of inflection by about 6 mM chloride. A change of the phase angle by 1° at the point of inflection results in a concentration change of (10 mM chloride. We are presently working on the design of Ru(dpp) beads that are not at all quenched by oxygen. CONCLUSION A powerful scheme for internal referencing of luminescence intensities is presented, which we refer to as dual lifetime referencing (DLR). It is based on the use of an indicator fluorophore with a short lifetime and a long-lived reference luminophore. The luminescence of the reference is independent of analyte concentrations in terms of intensity and lifetime and provides a constant background signal. The average phase shift depends on the ratio of the luminescence intensities of the indicator and the reference luminophore and reflects the intensity of the analyte-dependent fluorescent indicator. The modulation frequency is adjusted to the lifetime of the luminophore. Indicator and reference dyes need to have overlapping excitation spectra and emission spectra, respectively. As a result, they can be excited by the same light source and luminescence can be detected by the same photodetector. Drifts caused by the optoelectronic system, the optical parameters of the sample, bending effects of filter optics, and light losses in the optical path are referenced out by this method; however, losses in the
luminescence intensity of the indicator, for example, by leaching or photobleaching, will not be referenced. The turning points of calibration plots depend on the intensity ratio of the fluorophore and the luminophore. The apparent dynamic range and, thus, the point of inflection can be adjusted by varying the concentration of the fluorophore and the luminophore or by varying the optical filters; however, it is important that the relative proportions of the short and long lifetime
fluorophores remain constant during measurement. In addition, the calibration curves for a DLR sensor will change if the sensing and reference fluorophores photobleach at different rates.
Received for review December 15, 1999. Accepted March 29, 2000. AC9914364
Analytical Chemistry, Vol. 73, No. 9, May 1, 2001
2103
