Anal. Chem. 1996, 68, 4015-4019
Lifetime Enhancement of Ultrasmall Fluorescent Liquid Polymeric Film Based Optodes by Diffusion-Induced Self-Recovery after Photobleaching Michael Shortreed, Eric Monson, and Raoul Kopelman*
Department of Chemistry, The University of Michigan, Ann Arbor, Michigan 48109-1055
The major concern with optodes, especially miniaturized ones, has been their photobleaching limited lifetime. Liquid polymer [highly plasticized poly(vinyl chloride)] films are commonly used to prepare fluorescent optical fiber sensors. A major advantage is the ease of their fabrication. It is demonstrated here that, with proper choice of excitation power and illumination time, the sensor will completely recover itself from photobleaching after each measurement. This self-recovery is demonstrated on single-mode optical fibers with 80 µm diameter (3.1 µm active region) and on near-field scanning optical microscope pulled fiber tips with submicrometer diameter (250 nm active region). The single-mode optode can be used for 30 000 measurements with only a 5% signal loss at a signal/noise of >66. This opens the way for prolonged ratiometric application of such optodes. Microelectrodes have been used extensively for intracellular ion measurements.1-5 Though these microelectrodes tend to be somewhat noisier and less selective than standard-sized ionselective electrodes (ISEs), their performance has certainly been adequate. Recently, there has been much interest in miniaturizing optical fiber sensors (optodes)6-12 as an alternative to microelectrodes. Fluorescent optodes are especially attractive in this regard due to their inherent sensitivity. Indeed, a number of fluorescent optical fiber sensors with tip dimensions on the order of 1 µm and smaller have been prepared.6,10,11,13 Fabrication of such miniaturized devices was prompted by advances in the development of light sources14 for near-field scanning optical microscopy (NSOM) and by advances in photopolymerization chemistry.15 The (1) Tsien, R. Y.; Rink, T. J. Biochim. Biophys. Acta 1980, 599, 623-638. (2) Ammann, D. Ion-Selective Micro-electrodes; Springer-Verlag: Berlin, 1986. (3) Steiner, R. A.; Oehme, M.; Ammann, D.; Simon, W. Anal. Chem. 1979, 51, 351-353. (4) Brown, K. T.; Flaming, D. G. Neuroscience 1977, 2, 813-827. (5) Schaller, U.; Bakker, E.; Spichiger, U. E.; Pretsch, E. Talanta 1994, 41, 1001-1005. (6) Rosenzweig, Z.; Kopelman, R. Anal. Chem. 1995, 67, 2650-2654. (7) Shortreed, M. R.; Bakker, E.; Kopelman, R. Anal. Chem., in press. (8) Shortreed, M. R.; Kuhn, M.; Hoyland, B.; Kopelman, R. Anal. Chem. 1996, 68, 1414-1418. (9) Shortreed, M. R.; Dourado, S.; Kopelman, R. submitted to Sens. Actuators B. (10) Tan, W.; Shi, Z. Y.; Kopelman, R. Anal. Chem. 1992, 64, 2985-2990. (11) Tan, W.; Shi, Z. Y.; Smith, S.; Birnbaum, D.; Kopelman, R. Science 1992, 258, 778-781. (12) Tan, W. H.; Shi, Z. Y.; Kopelman, R. Sens. Actuators B 1995, 28, 157-163. (13) Lieberman, K.; Harush, S.; Lewis, A.; Kopelman, R. Science 1990, 247, 5961. (14) Betzig, E.; Trautman, J. K.; Harris, T. D.; Weiner, J. S.; Kostelak, R. L. Science 1991, 251, 1468-1470. S0003-2700(96)00525-2 CCC: $12.00
© 1996 American Chemical Society
fluorescent, analyte-sensitive portion of these sensors is typically a hydrogel where the dye is either covalently attached or simply trapped during the photopolymerization process.6,8,16-18 This process is now well-defined and reproducible, yielding fluorescent polymers immobilized directly in the path of excitation from the fiber. Unfortunately, the procedure still has many steps, thus requiring a 24 h time period for completion. A general problem exists for sensors such as ISEs, ion-selective field-effect transistors (ISFETs), and optodes, namely, a steady decrease in signal (or drift) associated with the loss of analytesensitive components via leaching from the membrane or film phase of the sensor. A number of solutions to this problem have been proposed including covalent attachment of the analytesensitive reagents to the sensor matrix8,10,15,19 and chemical derivatization of the reagents to increase their solubility in the sensor matrix and decrease their solubility in the sample solution.20 Perhaps, however, the most serious limitation specific to fluorescent optodes is the continuous loss in signal due to photobleaching. This is especially significant for microsensors, which contain only a relatively small number of fluorescent molecules within the sensitive polymer, while the photon flux is maintained high. The size of the polymer, concentration of indicator, and intensity of the excitation source incident on the polymer are primary factors that determine the working lifetime of the sensor. Once these sensors are deemed unusable, because the signal/noise has dropped below a specified limit, there is no mechanism for regeneration and a new sensor must be prepared. Until now, no serious attempt to alleviate this problem has been made. There is an important, but as yet unreported, advantage to sensors prepared using liquid polymer films [e.g., plasticized poly(vinyl chloride) (PVC)]. Components dissolved in this organic liquid diffuse randomly throughout the entire volume of the film. One of these components, the chromoionophore, changes its optical properties (absorbance, emission intensity, or emission wavelength) upon binding with a hydrogen ion. These chromoionophore molecules diffuse in and out of the active region of (15) Munkholm, C.; Walt, D. R.; Milanovich, F. P.; Klainer, S. M. Anal. Chem. 1987, 58, 1427-1430. (16) Bronk, K. S.; Michael, K. L.; Pantano, P.; Walt, D. R. Anal. Chem. 1995, 67, 2750-2757. (17) Li, L.; Walt, D. R. Anal. Chem. 1995, 67, 3746-3752. (18) Healey, B. G.; Walt, D. R. Anal. Chem. 1995, 67, 4471-4476. (19) Dumschat, C.; Alazard, S.; Adam, S.; Knoll, M.; Cammann, K. Analyst 1996, 121, 527-529. (20) Dinten, O.; Spichiger, U. E.; Chaniotakis, N.; Gehrig, P.; Rusterholz, B.; Morf, W. E.; Simon, W. Anal. Chem. 1991, 63, 596-603.
Analytical Chemistry, Vol. 68, No. 22, November 15, 1996 4015
the sensor where they are excited by light from an external source which has propagated through the core of the optical fiber. This gives liquid polymer films great potential advantage. Chromoionophore molecules that are photobleached while in the active region of the sensor are free to diffuse laterally away from the excitation zone, and unbleached chromoionophore molecules can diffuse from the bulk of the film into the active region, effectively replacing the photobleached molecules, thereby achieving selfregeneration of the sensor. Another advantage of these fluorescent liquid polymer film optodes is that they can perform activity (rather than concentration) measurements, which is not currently possible with other optodes, including the photopolymerized, hydrogel-type sensors. A number of simple mathematical relationships have recently been shown to be useful in the adaptation of the well-known optode principles and theory of Simon and co-workers21-23 from absorbance to fluorescence,7 thereby greatly enhancing the utility of these optodes for analytical purposes. Two of the three relationships given permit the use of ratiometric methods that eliminate difficulties associated with fluctuations in excitation power, changes in the light collection geometry, and photobleaching. Using a technique developed originally to study the motion of fluorescent molecules and fluorescently tagged proteins and lipids in thin membranes and films, we have demonstrated the efficacy of diffusion in dramatically enhancing the working lifetime of fluorescent optodes. The fluorescence recovery after photobleaching (FRAP) technique24 is based on using a brief, intense pulse of light to photobleach a small circular or Gaussian-shaped region in a thin film, followed by the monitoring of an increase in fluorescence intensity (with a greatly attenuated beam) as unbleached molecules outside of the bleached zone diffuse in. The recovery kinetics are characterized by the diffusion coefficient of the fluorophore, the size of the bleached spot, and the extent or depth of photobleaching. We present here the results of FRAP experiments performed on the ends of single-mode optical fibers and on NSOM light sources, used to study thin, plasticized PVC films doped with fluorescent chromoionophores (later referred to as simulated optodes). These results include an evaluation of the diffusion coefficient of these chromoionophore molecules and a comparison of sensor lifetimes with continuous vs pulsed illumination. We also outline a simple procedure for quick total regeneration of the microsensor while in use on the microscope stage. EXPERIMENTAL SECTION Reagents. PVC, chromoionophore III (ETH 5350), and bis(2-ethylhexyl) sebacate (DOS) were all obtained from Fluka Chemical Corp. (Ronkonkoma, NY). All components were dissolved in freshly distilled tetrahydrofuran (THF). Indicator Selection and Characteristics. The lipophilic Nile blue derivative chromoionophore III (ETH 5350) was chosen as the fluorescent indicator because of its widespread use in chemical sensors. The molar absorptivities25 and fluorescence spectra7 of this and similar molecules were previously presented. Chemical (21) Bakker, E.; Simon, W. Anal. Chem. 1992, 64, 1805-1812. (22) Seiler, K.; Simon, W. Anal. Chim. Acta 1992, 266, 73-87. (23) Bakker, E.; Willer, M.; Pretsch, E. Anal. Chim. Acta 1993, 282, 265-271. (24) Axelrod, D.; Koppel, D. E.; Schlessinger, J.; Elson, E.; Webb, W. W. Biophys. J. 1976, 16, 1055-1069. (25) Bakker, E.; Lerchi, M.; Rosatzin, T.; Rusterholz, B.; Simon, W. Anal. Chim. Acta 1993, 278, 211-225.
4016
Analytical Chemistry, Vol. 68, No. 22, November 15, 1996
stability measurements of similar molecules in plasticized PVC were also previously shown.25 No special emphasis was placed on the photostability of the indictor, rather, the universality as a highly proton selective ionophore molecule was emphasized. The concentration of indicator in the liquid polymer film was 15 mmol/ kg (also a common concentration used in making optodes). Equipment. All silica, single-mode optical fiber with a 3.1 µm mode field diameter and an 80 µm cladding diameter (Thor Labs, Newton, NJ) was used in the simulated optode experiments. The optical path included the following: lasers, Ion Laser Technology (Salt Lake City, UT) air-cooled argon ion laser Model ILT5490AWC and frequency-doubled CW Nd:YAG (ADLAS DPY 305 II); 514.5 nm laser band-pass filter (Newport Corp. Irvine, CA); neutral density filters (Melles Griot); infrared filter; Uniblitz shutter timer controller Model 310B (Rochester, NY); two λ/4, 3/ in. optical flats (Edmond Scientific); optical fiber coupler series 8 F-916 (Newport Corp., Irvine, CA); infinity corrected microscope objectives, Olympus 4× PLAN, 0.10 NA, and Olympus 40× U PLAN Fl, 0.75 NA (Olympus, Lake Success, NY); inverted epifluorescence microscope Model IX70 with wide green dichroic filter cube (Olympus); avalanche photodiode Model SPCM-200 PQ (EG&G PAR); Princeton Applied Research Model 1102 photon counter; and a digitizing oscilloscope Model TDS 420 (Tektronix). Sample Preparation. Optode film cocktails used for the simulated optode experiments were prepared to contain 15 mmol/ kg chromoionophore with 33% PVC and 66% DOS by weight. All components were dissolved in freshly distilled THF. Optodes were prepared by briefly dipping (25 s) signal is the sum of the background noise and the fluorescence signal.
lower than expected. This could be accounted for if the film thickness was much less than that found on the flat single-mode fiber. Optical microscope images confirm that the film thickness is less than 1.0 µm and may be as small as 100 nm (based on comparison of the optical image of the coated tip with the SEM
image of the bare, aluminum-coated fiber tip). Photon counter integration times were consequently limited by this low signal. This result, along with the slow reaction time of our mechanical shutter system and the extremely short recovery times associated with a bleached spot of only 250 nm, eliminated the possibility of observing the short-term recovery. For this reason, long (0.5 s) bleach times were used which likely form a large depletion zone of unbleached fluorophore molecules about the active region. Recovery of the original fluorescence intensity then requires several seconds (Figure 4). In a real experiment, we would expect the recovery time to be on the order of 1 s. Comparison of Pulsed vs Constant Illumination. To further demonstrate the effectiveness of diffusion in regenerating the active region and extending the working lifetime of the sensor, the effect of using brief pulses with a short recovery period between pulses was compared with the effect of continuous illumination. Continuous illumination of the sensor for ∼0.6 s (marked position A in Figure 5) resulted in a 70% loss in the signal intensity. In contrast, 30 consecutive 0.02 s pulses (0.6 s total) at the same laser power with a 1 s recovery period between each consecutive pulse resulted in only a 20% loss in the signal intensity (Figure 5). Continuous illumination results in a small depletion zone local to the excitation spot. These data indicate that intermittent optode measurements followed by reasonable recovery periods is dramatically effective in improving the usable lifetime of the optode. Film Lifetime. The intensity used to produce the constant illumination curve of Figure 5B, would not typically be used. The
probe beam intensity is adjusted by the signal/noise requirements of the user for a real experiment. To demonstrate the lifetime under typical experimental conditions, the film is illuminated continuously with enough laser light to give a S/N of 66 for an 0.1 s integration time (Figure 3 inset). After 30 000 consecutive measurements and a total illumination time of 50 min, only 5% loss in intensity occurred. Furthermore, the decay is not exponential, as would be expected if the fluorophores were immobilized. Operation of such an optode in a ratiometric mode would further mitigate the effect of photobleaching. The mechanical stability of our experimental setup is not sufficient (there is a slight long-term drift associated with the fiber translation and the microscope stages) to attempt this experiment with a 1 s resting period between excitations. On the basis of our results above, we expect to see little or no decay in the signal over the course of the entire experiment. ACKNOWLEDGMENT We gratefully acknowledge the advice received from Dan Axelrod. We acknowledge NIH Grant 1 RO 1 GM50300 01 for financial support.
Received for review May 29, 1996. Accepted August 15, 1996.X AC9605253 X
Abstract published in Advance ACS Abstracts, October 1, 1996.
Analytical Chemistry, Vol. 68, No. 22, November 15, 1996
4019
