Anal. Chem. 1997, 69, 2213-2216
Technical Notes
Fast Temporal Response Fiber-Optic Chemical Sensors Based on the Photodeposition of Micrometer-Scale Polymer Arrays Brian G. Healey† and David R. Walt*
The Max Tishler Laboratory for Organic Chemistry, Tufts University, Medford, Massachusetts 02155
Fiber-optic chemical sensor microarrays for the detection of pH and O2 have been developed with subsecond response times. Sensor microarrays are fabricated by the covalent immobilization (pH sensor arrays) or the physical entrapment (O2 sensor arrays) of fluorescent indicators in photodeposited polymer matrices on optical imaging fibers. Polymer microarrays are comprised of thousands of individual elements photodeposited as hemispheres such that each element of the sensor array is coupled directly to a discrete optical element of the imaging fiber and is not in contact with other neighboring elements. Because of the hemispherical shape and the individuality of the array elements, diffusion of analyte to the sensor elements is dominated by radial diffusion, resulting in a rapid response time. pH-sensitive arrays based on fluorescein respond to a 1.5-unit pH change within 300 ms, while the O2-sensitive arrays respond to O2 changes within 200 ms (90% of steady state response). A need exists in the field of chemical sensors for fast response times. For instance, many biological events at the cellular level, such as neurotransmitter release and Ca2+ fluctuations, occur at time scales on the order of tens to hundreds of ms. Few optical sensor techniques can meet this demand. Submicrometer optical chemical sensors have been developed by Kopelman and coworkers with response times on the order of hundreds of ms.1-3 Although the response of these optical sensors is analytically useful, the sensors have limited utility because the analytesensitive fluorescence must be collected by an external meanssa microscope objective under the sample. This geometry precludes the use of this sensor type for in vivo measurements. The faster response of the submicrometer optical chemical sensors is due to the extremely small size and thickness of the sensing layer. Analyte diffusion to the sensing region is dominated by radial diffusion as opposed to lateral diffusion. Radial diffusion of analyte to the sensor surface decreases the sensor response time, analogous to the decreased response times for microelectrodes over bulk electrodes.4 Microelectrode arrays have also been † Present address: Bristol-Myers Squibb Pharmaceutical Research Institute, 5 Research Parkway, Wallingford, CT 06492. (1) Rosenzweig, Z.; Kopelman, R. Anal. Chem. 1995, 67, 2650. (2) Tan, W.; Shi, Z. Y.; Kopelman, R. Anal. Chem. 1992, 64, 2985. (3) Tan, W.; Shi, Z. Y.; Smith, S.; Birnbaum, D.; Kopelman, R. Science 1992, 258, 778. (4) Forster, R. Chem. Soc. Rev. 1994, 289.
S0003-2700(96)01058-X CCC: $14.00
© 1997 American Chemical Society
fabricated with decreased response times compared to those of bulk electrodes due to radial diffusion of analyte and have been utilized to measure the spatial distribution of a chemical species.5 Previous work from our laboratory has reported on sensor arrays produced by spin-coating thin layers of analyte-sensitive polymer on an optical imaging fiber.6,7 These sensors provide spatial distributions of chemical concentrations with approximately 7 µm resolution, with the possibility of in vivo use due to the small size of the imaging fiber (350 µm) and the ability of the fiber to both deliver the excitation light and collect the analyte-sensitive fluorescence. Although these arrays provide information on chemical distribution, they suffer from multisecond response times due to the planarity and thickness of the sensing layer. We now report a simple technique for fabricating arrays of individual micrometer-scale analyte-sensitive polymer matrices with fast temporal responses. The method combines two techniques reported previously: (1) site-selective photodeposition of micrometer-scale polymer patterns on optical imaging fibers8 and (2) immobilization chemistry developed for preparing optical sensors.9-13 Specifically, we use the discrete light pathways in an optical imaging fiber to photopolymerize an array of thousands of individual polymer spots in a variety of sizes and patterns on a 350-µm imaging fiber. A pH-sensitive array has been fabricated by covalently immobilizing acryloylfluorescein in cross-linked polyacrylamide. O2-sensitive arrays have been fabricated by physically entrapping Ru(Ph2phen)32+ in a photopolymerizable siloxane. Both microarray types have response times of less than 300 ms and should have the ability to measure chemical distributions in vivo due to the small size of the microarray sensors (350 µm) and the sensors’ ability to collect the analyte-sensitive fluorescence. EXPERIMENTAL SECTION Materials. (80-85%)Dimethyl(15-20%)(acryloxypropyl)methylsiloxane copolymer (PS802) was purchased from Gelest, Inc. (5) Caudill, W. L.; Howell, J. O.; Wightman, R. M. Anal. Chem. 1982, 54, 2532. (6) Bronk, K. S.; Michael, K. L.; Pantano, P.; Walt, D. R. Anal. Chem. 1995, 67, 2750. (7) Pantano, P.; Walt, D. R. Anal. Chem. 1995, 67, 481A. (8) Healey, B. G.; Walt, D. R. Science 1995, 269, 1078. (9) Munkholm, C.; Walt, D. R. Talanta 1988, 35, 109. (10) Watts, R. J.; Crosby, G. A. J. Am. Chem. Soc. 1971, 93, 3184. (11) Lin, C. T.; Bottcher, W.; Chou, M.; Creutz, C.; Sutin, N. J. Am. Chem. Soc. 1976, 98, 6536. (12) Demas, J. N.; DeGraff, B. A. Anal. Chem. 1991, 63, 829A. (13) Xu, W.; McDonough, R. C., III; Langsdorf, B.; Demas, J. N.; DeGraff, B. A. Anal. Chem. 1994, 66, 4133.
Analytical Chemistry, Vol. 69, No. 11, June 1, 1997 2213
Remaining reagents were purchased from Aldrich Chemical Co. All reagents were used as received without further purification. All solutions were prepared from deionized water. Imaging fibers, purchased from Sumitomo Electric, had a 350µm diameter and were comprised of 6000, 3-4-µm individual elements. Two-foot lengths of fiber were polished using a polishing bushing and a polishing kit from General Fiberoptics, Inc. Deposition System for Photopatterning. The photodeposition system for the patterning of microspot arrays is a modification of the system described previously.8 Photodeposition was performed using an ultraviolet spot-cure system (EFOS, Ontario, Canada) utilizing a 100-W short-arc mercury lamp and a liquid light guide for light delivery. The spot-cure system has a built-in radiometer for controlling the light intensity and a shutter for controlling the polymerization time. The excitation light exiting the light guide was collimated with a 50-mm-focal length lens, passed through a mask holder, and imaged onto the proximal end of the fiber with a 15× reflecting microscope objective. The fiber was held in a fiber chuck mounted in an x,y positioner with 360° rotation, allowing for precise positioning of the imaged mask on the fiber. The fiber was inspected with a horizontally mounted microscope with an attached CCD camera. Detection System. Fluorescence measurements were made with a modified Olympus fluorescence microscope coupled to a Photometrics PXL frame transfer CCD camera. The instrument has been described in detail previously.6,7 Analyte-Sensitive Microarray Fabrication. pH-Sensitive Polymerization Solution. A working solution was prepared by dissolving 150 mg of acrylamide, 1.5 mg of methylenebisacrylamide (MBA), 30 mg of benzoin ethyl ether (BEE), and 50 µL of an acryloylfluorescein solution (2 mg/mL in n-propanol) in 950 µL of THF. O2-Sensitive Polymerization Solutions. A working solution was prepared by mixing 200 µL of (80-85%)dimethyl(15-20%)(acryloxypropyl)methylsiloxane copolymer with 545 µL of dichloromethane and 12 mg of BEE. A 1 mg/mL Ru(Ph2phen)32+ solution in dichloromethane was prepared. The distal end of the imaging fiber was first functionalized by treatment with 3-(trimethoxysilyl)propyl methacrylate (10% v/v in acetone) to attach a photopolymerizable acrylate group to the glass surface.9 The fiber was then placed on the photodeposition system. The initiation light passed through the mask holder and was imaged onto the proximal end of the fiber. The light flux for deposition was measured to be 14.5 mW/cm2. The distal end of the fiber was coated with a thin film of the polymerization solution by dipping it in a small volume of the solution and then removing it from the solution. The fiber was then illuminated for 1.0 s using an electronic shutter, and excess polymerization solution was removed by rinsing with ethanol. For the O2-sensitive microarray, the fiber was immediately placed in dichloromethane containing 1 mg/mL Ru(Ph2phen)32+ for 1 h and then rinsed with deionized water. Microarrays were then characterized by either scanning electron microscopy (SEM) or atomic force microscopy (AFM). Scanning Electron Micrography (SEM). The distal tip of the fiber was removed and mounted on an aluminum stub with colloidal graphite tape and rf sputter-coated with 150-nm thickness of gold. SEM images were acquired with a JEOL Model H40 SEM. 2214 Analytical Chemistry, Vol. 69, No. 11, June 1, 1997
Figure 1. AFM images of pH- (A) and O2-sensitive microarrays (B). The z-axis in panel A represents 3 µm, and that in panel B represents 7 µm.
Atomic Force Microscopy (AFM). Measurements were obtained using atomic force microscopy (AFM) using a Digital Instruments Dimension 3000 scanning probe microscope. Polymer microstructures were analyzed in the contact mode. Fibers were held for scanning in a custom-built holder. pH-Sensitive Array Testing. Two 0.1 M phosphate buffers of pH 5.0 and 7.0 were prepared containing 140 mM NaCl. The fiber was placed on the imaging system with the distal end placed in 50 µL of pH 7.0 buffer in a 1.5-mL Eppendorf tube. A sequence of images was acquired at 490-nm excitation and 530-nm emission wavelengths. The sequence was acquired for 8 s using a 25-ms exposure with 150 ms between frames. After approximately 4 s of acquisition, 450 µL of pH 5.0 buffer was injected into the tube. After acquisition, the stored images were processed using the image processing software IPLab Spectrum (Signal Analytics). A circular region of interest was drawn around the array, and the average fluorescent intensity of the array was calculated versus time. O2-Sensitive Array Testing. Three test solutions of pH 7.0, 0.1 M phosphate buffer with 140 mM NaCl were bubbled with compressed gas containing 0, 10, and 100% O2, balance N2. A 21% O2 test solution of the same buffer was prepared by storing the solution on the bench and allowing the solution to equilibrate with air. The fiber was placed on the imaging system with the distal end placed in 50 µL of the 21% O2 test solution in a 1.5-mL Eppendorf tube. A sequence of images was acquired at 470-nm excitation, and the fluorescence was collected at emission wavelengths >600 nm. The sequence was acquired for 6 s using a 50-ms exposure with 86 ms between frames. After approximately 3 s of acquisition, 450 µL of a test solution was injected into the tube. The images were then analyzed as described above.
Figure 2. Response of the pH- and O2-sensitive microarrays, A and B, respectively. The pH-sensitive microarray response (A) was tested by diluting pH 7 phosphate buffer with pH 5 phosphate buffer at a dilution of 10-fold. The pH-sensitive microarray reached 90% response in less than 300 ms. The O2-sensitive microarray response (B) was tested by diluting a 21% O2-saturated solution with a 10% (b) and a 100% (O) O2-saturated solution 10-fold. The O2-sensitive microarray reached 90% response in less than 200 ms. Fluorescence was measured at 490-nm excitation and 530-nm emission for the pHsensitive microarray (A) and at 470-nm excitation and >600-nm emission for the O2-sensitive microarray (B).
RESULTS AND DISCUSSION O2- and pH-Sensitive Microarray Fabrication. The present technique is based on the fabrication of micrometer-scale polymer patterns previously developed in our laboratory.8 The reported procedure produced thousands of individual polymer microspots on an optical imaging fiber, with each element of the array polymerized only on the individual cores of the imaging bundle. Polymer deposits as hemispheres on the fiber and only forms on the cores because the glass claddings between the fiber elements do not propagate light. Because of the hemispherical shape of the polymer microspots, radial diffusion of analyte to the microarray elements should occur. By immobilizing analyte-sensitive indicators in the polymer, optical chemical sensor microarrays with fast response times have been fabricated. The pH-sensitive microarray is created by covalently immobilizing a pH indicator, acryloylfluorescein, in polyacrylamide. This indicator exhibits an enhancement in fluorescence intensity upon deprotonation.9 The present immobilization protocol is new and involves photopolymerization of the indicator in the acrylamide. Previously, fiber-optic chemical sensors utilizing polyacrylamide were fabricated by aqueous thermal polymerization or photopolymerization using potassium persulfate or riboflavin,
Figure 3. Calibration curves of pH- and O2-sensitive microarrays, A and B, respectively. The pH-sensitive microarray is sensitive between pH 5.5 and 7.5, with a pKa of 6.8. The calibration of the O2-sensitive microarray is based on a Stern-Volmer analysis. The O2-sensitive microarray is sensitive over a wide dynamic range and follows a single-site quenching mechanism.
respectively.9,14 These procedures could not be used to fabricate polyacrylamide microarrays because of the difficulty in controlling the polymerization rate. By switching the polymerization solvent to THF, benzoin ethyl ether could be used as the initiator, which provides excellent control of the polymerization rate.9,14 The O2sensitive microarray is based on fluorescence quenching of Ru(Ph2phen)32+ entrapped in a photopolymerized gas-permeable siloxane. At high O2 concentrations, the fluorescence of the complex is quenched due to dynamic quenching by molecular oxygen.10-13 A polymer microarray of methacrylate-functionalized polysiloxane was first deposited on an imaging fiber and subsequently immersed in a dichloromethane solution containing Ru(Ph2phen)32+. The organometallic complex diffuses into the polymer microspots and remains trapped after solvent evaporation. Analyte-sensitive microarrays are formed when polymerization is carried out for 1.0 s at 14.5 mW/cm2. Once an analyte-sensitive microarray was fabricated, it was characterized with atomic force microscopy (AFM). Figure 1 shows AFM images of pH-sensitive and O2-sensitive microarrays, A and B, respectively. The microspots were found to be hemispherical, 2.8 µm in diameter, with 3.6-µm center-to-center spacing and heights of 0.6 and 1.2 µm for the pH-sensitive and O2-sensitive microarrays, respectively. (14) Kulp, T. J.; Camins, I.; Angel, S. M.; Munkholm, C.; Walt, D. R. Anal. Chem. 1987, 59, 2849.
Analytical Chemistry, Vol. 69, No. 11, June 1, 1997
2215
Sensor Measurements. After AFM characterization, microarrays were placed in pH 7.0 phosphate buffer. Microarrays were then placed on the fluorescence detection system, and the response was tested by the procedures described in the Experimental Section. Briefly, the sensors were placed in a small volume of buffer at one analyte concentration, and then a large excess of a second analyte concentration was added. Data were analyzed by processing the images to give the mean intensity of the array versus time. Figure 2A shows the response of the pH-sensitive microarray to a 1.5-pH unit decrease. The pH-sensitive microarray responds to the pH change in approximately 300 ms, with response defined as the time to reach 90% change. Figure 2B shows the response of an O2-sensitive microarray to both an increase and a decrease in oxygen tension. The microarray responds to both O2 changes in 200 ms. In both microarrays, the decreased response time compared to that of bulk sensors is due to the small size of the individual polymer array elements and to the increase in radial diffusion of H+ or O2 to the individual sensor elements within the microarrays. Figure 2 also demonstrates that neither the pH- nor the O2-sensitive microarrays show any photobleaching during the continuous excitation throughout the response testing, as demonstrated by the absense of any signal decrease either before or after injection of a test solution. Figure 3 shows the calibration curves of the pH- and the O2sensitive microarrays, A and B, respectively. The pH-sensitive microarray is sensitive in the pH 5.5-7.5 range, with pKa of 6.8. Calibration of the O2-sensitive microarrays is based on the SternVolmer analysis of the microarray response. Data were analyzed using the Stern-Volmer equation, I0/I ) 1 + Ksv[O2], where I0 is the intensity at zero O2 tension, I is the intensity at a given O2
2216
Analytical Chemistry, Vol. 69, No. 11, June 1, 1997
tension, and Ksv is the Stern-Volmer quenching constant. Figure 3B demonstrates that the sensor response is in agreement with theory and that the sensor has a large dynamic range. The linearity of the response compared to those of other O2 sensors suggests that the microspots are homogeneous in composition.11 The sensitivity factor of the O2-sensitive microarray, defined as the intensity in the absense of O2 divided by the intensity at O2saturated conditions (I0%/I100%), is 5.8, in good agreement with other published optical O2 sensors.1,10,11 These microarray sensors for pH and O2 offer several advantages over existing optical chemical sensors. First, the sensors have a faster response time, making them more suitable for measuring biological events and kinetic studies. The microarray sensors also can be used for remote sensing because the imaging fiber carries both the excitation and emission light to and from the analyte-sensitive microarray. Finally, because the microsensors form an array, they can be used to measure spatial distributions of chemical concentrations.6,7 ACKNOWLEDGMENT The authors thank the National Institutes of Health for financial support (GM-48142) and ONR (DURIP award) for the atomic force microscope.
Received for review October 10, 1996. Accepted February 27, 1997.X AC961058S X
Abstract published in Advance ACS Abstracts, April 15, 1997.
