Microgap Multicavity Fabry–Pérot Biosensor

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 25, NO. 7, JULY 2007

1797

Microgap Multicavity Fabry–Pérot Biosensor Yan Zhang, Member, OSA, Xiaopei Chen, Student Member, IEEE, Yongxin Wang, Kristie L. Cooper, Member, IEEE, and Anbo Wang, Senior Member, IEEE

Abstract—This paper presents a microgap multicavity Fabry– Pérot interferometric sensor fabricated by wet etching and fusion splicing of single-mode optical fibers. The temperature dependence of the optical thickness measurement of self-assembled thin films can be compensated by extracting the temperature information from the multiplexed temperature sensor. Experimental results demonstrate that thin-film characteristics under temperature variations can be examined accurately. The high-temperature sensitivity of the temperature sensor also enables biosensing under temperature variations. This greatly improves the flexibility in sample handling and provides the opportunity to investigate temperature effects in biological applications. Index Terms—Fabry–Pérot interferometers, multicavity sensor, optical fiber sensors, temperature compensation.

I. I NTRODUCTION

A

MONG the many types of fiber-optic sensors, fiber-optic Fabry–Pérot interferometric (FFPI) sensors are distinctive because of their high sensitivity, ease of fabrication, localization, and lead insensitivity. FFPI sensors can be classified as extrinsic Fabry–Pérot interferometric (EFPI) sensors and intrinsic Fabry–Pérot interferometric (IFPI) sensors. An EFPI sensor uses an air cavity between two cleaved fiber ends inserted into an alignment ferrule and bonded by laser welding or epoxy adhesive [1], [2]. Although the EFPI sensor is attractive in various applications, it has intrinsic disadvantages such as difficulty in bonding, diameter nonuniformity due to the alignment ferrule, and limitation on the cavity size due to coupling loss. The large mismatch in thermal expansion coefficients of fibers, alignment ferrules, and bonding materials will cause severe stress in sensor construction. The bonding adhesive such as epoxy undergoes viscoelastic creep and cannot survive at high temperature [3]. The geometric discontinuity will create difficulty in protecting and mounting the sensor in the measurement. The large coupling loss reduces the multiplexing capability. In contrast, an IFPI sensor contains the sensing element, i.e., the Fabry–Pérot (FP) cavity, inside the fiber. The fiber both guides light and experiences the perturbation of interest. IFPI sensors reduce the bonding difficulties experienced in EFPI sensor

fabrication and provide miniature size, continuous geometry, robust structure, and versatile installation. In tubing-based multicavity Fabry–Pérot interferometric (MFPI) sensors, cleaved fiber ends separated by air gaps can serve as reflectors for the FP cavity [4]. However, the cavity size is limited by the coupling efficiency, which thus reduces the flexibility in fabrication and multiplexing. The local reflectors or mirrors inside the fiber can also be fabricated with various methods, such as dielectric thin films [5], [6] and fiber Bragg gratings (FBGs) [7]. Single-layer [5] or multilayer [6] dielectric mirrors can be deposited onto the fiber by magnetron sputtering. After splicing with another fiber, internal mirrors with reflectance of greater than 85% can be achieved [6]. Although this technique has shown success, it is still limited by the need for a special coating on the fiber and deterioration in film quality during splicing. Moreover, additional loss could occur due to fiber end surface roughness, cleave angle, or reflection into the cladding of the fiber [8]. Another method of producing high-finesse fiber cavities is to combine two FBGs as mirrors [7]. High-finesse values can be achieved with narrow-spectralwidth FBGs [9]. Recently, chirped Bragg gratings with much wider spectral widths have been examined in FP filters [10]. Theoretical analyses predict that high finesse and wide spectral widths could be achieved with chirped Bragg gratings [11]. This technique remains expensive and complicated because of the need for chirped FBGs. In this paper, microgap FP sensors were fabricated by combining wet etching and fusion splicing. This process is cost effective, easy to multiplex, and suitable for batch production [12]. Multicavity sensors for thin-film applications were constructed by multiplexing individual microgap sensors. This approach not only has similar advantages to tubing-based MFPI sensors [13] but also is notable for its simple fabrication process and flexibility in cavity lengths. Furthermore, the temperature sensitivity for temperature compensation has increased by an order of magnitude from the tubing-based MFPI sensor. Thus, thin-film characteristics under temperature variations can be examined more accurately. II. S ENSOR F ABRICATION

Manuscript received July 20, 2006; revised March 25, 2007. This work was supported in part by the U.S. Army Medical Research and Materiel Command under Award DAMD17-03-1-0008. Y. Zhang is with the Fitzpatrick Institute for Photonics, Duke University, Durham, NC 27708 USA (e-mail: [email protected]). X. Chen, Y. Wang, K. L. Cooper, and A. Wang are with the Center for Photonics Technology, Bradley Department of Electrical and Computer Engineering, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061 USA (e-mail: [email protected]; [email protected]; [email protected]; [email protected]). Digital Object Identifier 10.1109/JLT.2007.899169

Chemical etching is widely used in fiber-optic probe fabrication. Conical cores [14] and microwells [15], as well as nanotips [16], can be achieved with wet etching techniques. Chemical etching offers a simple and cost-effective fabrication technique in which the optical fibers are dipped into a balanced solution of hydrofluoric acid (HF) and into an ammonium fluoride (NH4 F) buffer. The process depends on a differential etch rate between the pure silica of the fiber cladding and the germanium-doped

0733-8724/$25.00 © 2007 IEEE

1798

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 25, NO. 7, JULY 2007

TABLE I CONTROL OF THE ARC CONDITIONS (SUMITOMO TYPE-36 FUSION SPLICER) FOR MULTICAVITY SENSOR FABRICATION AND NORMAL SM FIBER SPLICING

Fig. 1.

Fiber end face after 15 min etching in BHF (Zeiss Axiovert 25).

silica of the fiber core. NH4 F is used to reduce the dissolution rate of the germanium-doped core. The cone angle of the etched tip is inversely proportional to the volume ratio of NH4 F in buffered hydrogen fluoride (BHF) [17]. Various cone angles can be obtained by controlling the volume ratio of NH4 F. Etched fiber probes with cone angles larger than 180◦ mean that the etching rate of the core is larger than that of the cladding. The negative cone angle corresponds to a dip in the core area. After splicing with another fiber, the resulting microgap works as a reflector in the optical-fiber FP cavity. For this paper, the etching solution was composed of hydrofluoric acid (HF 50% weight concentration), ammonium fluoride (NH4 F 40% weight concentration), and deionized water. When the volume ratio is NH4 F : HF : H2 O = X : 1 : 1 (X < 1.7 [18]), the etching rate of the core will be faster than that of the cladding. While there is no obvious dependence of the surface quality of the tip on the etching temperature, the etching rate increases significantly with temperature [16]. To simplify the procedure, all the experiments were carried out at room temperature (25 ◦ C). A single-mode (SM) fiber (Corning SMF-28) was first cleaved using a cleaver (Fujikura CT-04B). The reflection intensity of the cleaved fiber was examined by a high-speed spectrometer (Micron Optics SI720 Component Testing System, CTS) to ensure the quality of the end face. Then, the fiber was etched in a BHF solution (NH4 F : HF : H2 O = 0.5 : 1 : 1) for 15 min. After rinsing in deionized water and drying, the end face of the fiber tip was examined under a microscope (Zeiss Axiovert 25). Fig. 1 shows a picture of the etched end face. By comparing the diameter of the fiber tip before and after etching, the etching rate of the cladding can be calculated as 1.2 µm/min. The etching rate of the core is a little bit higher than that of the cladding, so there is a dip in the core area, as shown in Fig. 1. The thickness of the dip is about 1–2 µm as estimated under the microscope. The morphology of the cladding and the dipped core is not purely flat, which may be due to the composition profile during preform fabrication [18]. Generally, this could be eliminated by carefully controlling the preform fabrication conditions. The etched fiber was then spliced with another SM fiber using a conventional fiber splicer (Sumitomo Type-36). A micro-

Fig. 2.

Microgap FP sensor (Zeiss Axiovert 25).

gap was formed in the core area between the fibers and behaves as a reflector in the FP cavity. Shorter arc time and lower arc power are preferred during splicing. Otherwise, the reflectance from the microgap will be reduced or even eliminated. At the same time, in order to maintain the strength of the splicing point, the arc power should not be too low. There is similar requirement on the arc duration. Table I compares the different arc conditions for multicavity sensor fabrication and normal SM fiber splicing. The two surfaces of the microgap differ from each other. One is an etched dip, and the other is a well-cleaved fiber end face. Due to the curvature of the etched dip, the reflection at that surface is much lower than a well-cleaved fiber end face. Normally, a 5-dB decrease in reflection intensity is observed at the etched fiber tip compared with a well-cleaved fiber end face. Furthermore, the air gap in the reflector is very thin, so the reflection strength does not vary much over the optical range of our interest. Therefore, the microgap can be estimated as a single reflector. By cleaving the etched fiber, another reflector of the FP cavity was formed by the end face. Fig. 2 shows a picture of a microgap sensor under the microscope. A single-cavity microgap FP sensor itself can work as a thin-film sensor in chemical sensing or biosensing by monitoring the changes in the reflection spectrum during adsorption. Furthermore, the microgap sensor is distinguished for its multiplexing capability. By multiplexing with another microgap sensor, temperature and thin-film adsorption can be examined simultaneously. This not only reduces the temperature-induced error but also provides the opportunity to investigate the film under temperature variations. By splicing a microgap sensor with another etched fiber, a second microgap was formed between the sensor and the etched fiber. A multicavity sensor was fabricated by cleaving the second etched fiber. A picture of a multicavity sensor is shown in Fig. 3. The two FP cavities are designed with different purposes. The first FP cavity (FP1 in Fig. 3) provides the

ZHANG et al.: MICROGAP MULTICAVITY FABRY–PÉROT BIOSENSOR

Fig. 3.

1799

Microgap multicavity FP sensor (Zeiss Axiovert 25).

Fig. 5. Fourier transformation of the reflection spectrum.

Fig. 4.

Reflection spectrum of a microgap FP sensor. Fig. 6. Fourier transformation of the bandpass-filtered spectrum.

temperature-sensing capability, and the second cavity (FP2 in Fig. 3) offers the thin-film sensing functionality.

I = A + B cos(2KL + φ)

III. S IGNAL A NALYSIS A. Single-Cavity Sensor As discussed in the preceding section, the microgap sensor in Fig. 2 can be simplified as a single-cavity sensor. Fig. 4 shows a sample spectrum of a microgap FP sensor in the wavenumber domain. The sinusoidal waveform validates our assumption. The Fourier transformation of the reflection spectrum is given in Fig. 5. The low-frequency component is mainly due to the direct current offset in the reflection spectrum. A bandpass filter is able to efficiently separate the fiber cavity signal from the lower-frequency signal and higher-frequency harmonics [19]. Finite-impulse response filters were selected due to their linear phase response to frequency. Fig. 6 shows the Fourier transformation of the bandpass-filtered spectrum. The singlefrequency signal has been separated by the filter. The phase delay of an N th-order filter can be expressed as [20] n = (N − 1)/2.

The reflection spectrum of a low-finesse single-cavity FP sensor can be described by

(1)

The delay can be easily compensated by shifting the filtered signal backward by n samples.

(2)

where A and B are coupling coefficients and independent of wavelength, K = 2π/λ is the wavenumber, L = nd is the optical length of the FP cavity, and φ is the coupling phase shift. The combination of single-peak (single-valley) tracing and multiple-peak (multiple-valley) demodulation provides large dynamic range and high resolution simultaneously [21]. The first step is to determine the specific order number of a certain peak. From (2), the adjacent peaks satisfy 2KL + φ = 2πm 2K  L + φ = 2π(m + 1)

(3)

where m is the order number of a peak and should be a positive integer. The order number can be expressed as m = [K/∆K + φ/2π]

(4)

where the square brackets mean to round toward the nearest integer, and ∆K = K  − K. Here, the phase shift φ is from the coupling of light in the microgap. It can be calculated from the estimated size of the microgap, which is similar to the analysis

1800

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 25, NO. 7, JULY 2007

Fig. 7. Coupling phase shift versus the cavity length. Fig. 8.

Reflection spectrum of a microgap multicavity FP sensor.

of splice loss of two SM fibers [22]. When the lateral and angular misalignments are minimized during fabrication, the phase shift is mainly due to the longitudinal misalignment φ = −atan(G/2)

(5)

where G = Z/(KS 2 ) is the longitudinal displacement factors. Z = 2d/a is the normalized longitudinal displacement, where d is the size of the microgap, and a is the fiber core radius. The effective displacement is two times the cavity length because the incident light reflects from the reflection fiber and couples back into the lead-in √ fiber. K = ka is the normalized wavenumber, S = s/a = w/( 2a) is the normalized intensitybased spot size radius, and w is the amplitude-based spot size radius and can be approximated by the empirical formula [23] w/a = 0.65 + 1.619/V 3/2 + 2.879/V 6 .

(6)

Here, a = 4.1 µm, V = 2.15, and S = 0.84 (Corning SMF-28, λ = 1550 nm). Fig. 7 shows the relationship between the coupling phase shift φ and the cavity length d. The size of the microgap is only about 1–2 µm, so the phase shift is much less than 0.1 rad and does not affect the rounding result in the calculation of the order number (4). Then, the optical cavity length (OCL) can be calculated as L = (2πm − φ)/2K.

(7)

It is difficult to derive the exact cavity length from the above equation because of the unknown phase shift from the microgap. The OCL can be estimated as L ≈ πm/K = mλ/2

(8)

since φ  2πm. The error of this estimation depends on the phase shift. In a rough estimation, a 0.1-rad phase shift will cause an OCL error of 0.01 µm (λ = 1.55 µm). However, we care more about the relative changes of the OCL than the absolute cavity length itself. It is possible to calculate the changes in the OCL as ∆L = (2πm − φ)∆λ/4π ≈ m∆λ/2

(9)

Fig. 9. Fourier transformation of the reflection spectrum of a microgap multicavity FP sensor.

where ∆λ is the wavelength shift of the mth-order peak. The relative error is δ(∆L)/∆L = φ/2πm  1.

(10)

For an FP cavity with OCL of 200 µm, the order number of a peak at λ = 1550 nm is m = 2L/λ ≈ 258. If the phase shift is 0.1 rad, the relative error of the changes in OCL is only 6 × 10−5 . In this paper, the change in OCL is always less than 1 µm, so the error will be smaller than 0.1 nm. B. Multicavity Sensor A multicavity sensor with structure shown in Fig. 3 is constructed by two FP cavities. Fig. 8 shows a reflection spectrum of that type of sensor. The Fourier transformation in Fig. 9 clearly illustrates the spatial frequency components from each FP cavity and their combinations. Following a similar method as for the single-cavity sensor, the frequency component of each cavity can be filtered out by specially designed bandpass filters. Then, the cavity lengths and their changes can be calculated as described in the previous section.

ZHANG et al.: MICROGAP MULTICAVITY FABRY–PÉROT BIOSENSOR

1801

Fig. 10. Schematic of microgap multicavity FP sensor system for thin-film applications.

IV. T EMPERATURE C OMPENSATION The microgap sensor has been used in temperature, strain, and thin-film measurement [12]. The sensor is distinguished by its simplicity, low cost, and high sensitivity to monitoring various changes. It is capable of measuring ultrathin films in chemical or biological applications by comparing the changes in optical thickness of the cavity. However, the thermal-induced errors in thin-film measurement cannot be neglected. Therefore, it is necessary to introduce temperature compensation in highaccuracy measurements at varying temperatures. Compared with the tubing-based MFPI sensors whose temperature resolution is only about 2 ◦ C because of the low thermal expansion coefficient [13], the microgap multicavity sensor increases the temperature sensitivity by an order of magnitude, simplifies the fabrication procedures, and provides flexibility in cavity lengths. The experimental setup of the microgap sensor system is shown in Fig. 10. A low-noise fiber ring laser from a CTS (Micron Optics SI720) was coupled into the sensor. The reflection spectrum of the sensor was monitored by the detector in the CTS and analyzed by a personal computer. The CTS offers a measurement range of 1520–1570 nm with 1-pm accuracy. A multicavity sensor was constructed with a temperature sensor and a thin-film sensor as in Fig. 10. The optical thickness of the thin-film sensing cavity is highly dependent on the environment temperature due to the thermo-optic effect and thermal expansion. The change in optical thickness of the cavity with temperature can be expressed as ∆(nd)f = ∆nf · df + ∆df · nf = αn ∆T · df + αd df ∆T · nf = αf nf df ∆T

(11)

where αn is the thermo-optic coefficient, αd is the thermal expansion coefficient, and αf = αn /nf + αd is the effective temperature coefficient. The effective temperature coefficient is about 7.4 × 10−6 /◦ C (λ = 0.633 µm) [13]. The temperature sensitivity can be derived as ∆T = ∆L/αf L.

(12)

The relative change in optical thickness can be demodulated with up to 10−6 resolution limited by the CTS, which corresponds to around 0.1 ◦ C temperature resolution. The thermal

Fig. 11. Changes in optical cavity length of temperature sensor as the temperature was cycled from 0 to 100 ◦ C.

effect can be reduced by shortening the cavity length. An OCL as small as 20 µm can be achieved by cleaving under a microscope. However, it is increasingly difficult to obtain high-visibility fringes as the size of the cavity becomes shorter. Furthermore, a light source with wider wavelength range is required in the demodulation of the reflection spectrum. From (3), the wavelength range can be expressed as ∆λ ≈ λ2 /2L

(13)

where λ is the central wavelength of the light source. For an FP cavity with OCL as small as 10 µm, the wavelength range of the light source (λ = 1550 nm) should be larger than 120 nm for multiple-peak demodulation. Due to the difficulties existing in both fabrication and signal processing, a singlecavity sensor is not appropriate for thin-film measurement under temperature variations. A microgap multicavity sensor with a built-in temperature sensor and thin-film sensor not only compensates the temperature-induced error but also simplifies the fabrication and signal analysis. To effectively demodulate the reflection spectrum of the multicavity sensor, two cavities are fabricated with different lengths. The temperature-sensing cavity is longer as the temperature sensitivity depends on the relative changes in the optical length of the cavity. The thin-film sensing cavity is shorter for higher resolution of the absolute changes in optical thickness during adsorption. Here, the OCLs of the temperature sensor and thin-film sensor are around 1000 and 200 µm, respectively. The lengths of each cavity are quite flexible as long as their corresponding frequency components can be effectively differentiated from each other. The multicavity sensor was calibrated in an environmental chamber (TestEquity 1020C). The temperature response of each cavity is illustrated in Figs. 11 and 12. The relationship between the OCL and temperature is almost linear, as expressed in (11). The weak nonlinearity may come from the thermal expansion of silica [24]. The changes in OCL can be derived as 
(14) ∆L = (αn /n + αd )∆T + βd ∆T 2 L where αn is the thermo-optic coefficient, and αd and βd are the thermal expansion coefficients. The temperature sensitivity

1802

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 25, NO. 7, JULY 2007

Fig. 12. Changes in optical cavity length of thin-film sensor as the temperature was cycled from 0 to 100 ◦ C.

Fig. 14. Effect of temperature compensation on the optical thickness error of the thin-film sensor. The temperature was cycled twice in steps of 10 ◦ C in the range from 0 to 100 ◦ C. The upper curve shows the optical thickness error before compensation, and the lower curve shows the error after compensation.

Fig. 13. Comparison of changes in optical cavity lengths of temperature sensor with respect to temperature from 0 to 100 ◦ C.

of the microgap sensor is related to the thermo-optic effect and thermal expansion of the FP cavity, while the temperature sensitivity of the tubing-based sensor is based on the thermal expansion of the silica tubing [13]. The thermo-optic coefficient of silica is more than ten times larger than the thermal expansion coefficient [25]. So the temperature sensitivity of the microgap sensor is higher than the tubing-based FP sensor by an order of magnitude. A temperature resolution of 0.1 ◦ C can be achieved in our test. Thus, the thin-film characteristics under temperature variations can be examined more accurately. The relative changes in the OCLs of the temperature sensor and thin-film sensor can be written as ∆Lf /∆LT = Lf /LT

(15)

where Lf and LT are the OCLs of each cavity. Fig. 13 shows the linear relationship between the changes in cavities. Therefore, the thermal-induced error in thin-film sensing can be easily compensated by comparing the changes in OCLs of the two sensors. Fig. 14 shows the effect of temperature compensation on the optical thickness error of the thin-film sensor before adsorption. In the temperature range from 0 ◦ C to 100 ◦ C, the temperature dependence is about ±70 nm before

Fig. 15. Change in optical thickness of the [PAH/PSS]10 film under temperature variations.

compensation. The thermal-induced error greatly affects the thin-film measurement under temperature variations. Temperature compensation effectively reduces the thermal-induced error to less than ±0.2 nm over the entire temperature range. The pattern after compensation may come from the fitting error due to the nonlinear thermal optical or thermal expansion effect of the cavities. V. T HIN -F ILM S ENSING To demonstrate the effect of temperature compensation in thin-film sensing, the multicavity sensor was coated with a self-assembled polyelectrolyte multilayer. The sensor was alternately immersed in a polycationic (PAH) and a polyanionic (PSS) solution. A multilayer thin film with structure of [PAH/PSS]10 was self-assembled onto the sensor end. The film thickness was measured to be 86 nm (n = 1.54) at room temperature (25 ◦ C) by calculating the changes in OCLs of the thin-film sensor. The thin film was examined in the chamber under temperature range of 0 ◦ C–100 ◦ C. Fig. 15 illustrates

ZHANG et al.: MICROGAP MULTICAVITY FABRY–PÉROT BIOSENSOR

1803

the changes in optical thickness of the thin film. The film was heated in 10 ◦ C increments and equilibrated for 30 min at each step. From 0 ◦ C to 20 ◦ C, the optical thickness decreased slowly. From 20 ◦ C to 100 ◦ C, the decrease became more rapid. Then, the film was cooled to 0 ◦ C in the same step and equilibration time. The optical thickness slowly increased, but apparent hysteresis was observed. After the film was kept at room temperature for 24 h, the optical thickness recovered to its initial value. The decrease in optical thickness and hysteresis may be due to the release of bound water molecules from the film and reentry from the air due to temperature and humidity changes [26]. This is a common effect in this type of selfassembled thin film. VI. B IOSENSING In biosensing, thermal-induced errors cannot be neglected due to the significant influence on measurement accuracy. However, the low-temperature resolution of the tubing-based MFPI sensor limits its application in temperature-compensated biosensing [13]. The microgap multicavity sensor composed of a temperature sensor and a thin-film sensor increases the temperature sensitivity by an order of magnitude and has demonstrated its capability in thin-film sensing. Therefore, we applied the microgap MFPI sensor in biosensing under temperature variations. The experimental setup of the microgap sensor system is similar to the description in thin-film sensing (Fig. 10). The environmental temperature of the sensor was controlled by a water bath. A small vial (0.5 mL) containing the solution was immersed in water. The water temperature was controlled by a hot plate. A microgap multicavity FP sensor coated with selfassembled polymer films ([PAH/PSS]2 ) was first exposed to pig IgG solution (20 µg/ml in 0.05 M HEPES). After immobilization of IgG, the sensor was immersed in anti-pig IgG solution (20 µg/ml in 0.05 M HEPES). The kinetics of the antiIgG binding to the immobilized IgG layer were determined by changes in optical thickness, as shown in Fig. 16. The top curve shows the temperature change during the test, which was measured by the built-in temperature sensor. In order to increase the temperature range to be investigated, the water temperature was first decreased to around 7 ◦ C in a refrigerator. Then, during immunosensing, the temperature was slowly increased to around 37 ◦ C. The middle curve shows the change in optical thickness of the thin film without compensation. A total change of 36.9 nm was observed, which is due to both the immunoreactions and thermal effects of the thin-film sensor. The optical thickness slowly increased, and no equilibrium state was achieved within 60 min because of the thermal effects. The microgap sensor had been calibrated in the environmental chamber before coating. According to (15), the changes in the OCLs of the temperature sensor and thin-film sensor have a linear relationship. Therefore, the thermal-induced error in thin-film sensing can be easily compensated by comparing the changes in OCLs of the two sensors. Temperature compensation revealed that the equilibrium state of the immunoreactions was actually attained within 20 min, as can be seen from

Fig. 16. Effect of temperature compensation on the change in optical thickness during immunosensing. The top curve shows the temperature change during the test. The middle curve shows the change in optical thickness before compensation, and the lower curve shows the change after compensation.

the bottom curve in Fig. 16. The change in optical thickness reduced to about 7.1 nm, which corresponded to a thickness change of about 5.0 nm (n = 1.432). The thickness change is similar to the result under room temperature [4], which demonstrates the capability of the microgap MFPI sensor in biosensing with temperature variations. VII. C ONCLUSION Microgap FP sensors have been developed by wet etching and splicing techniques. Small dips in fiber cores can be fabricated by differential etching. By splicing the etched fiber with another fiber, a microgap was generated inside the fiber that functioned as a reflector. Low-finesse FP cavities were formed between the microgap reflectors and cleaved fiber ends. This process provides a simple and cost-effective method for IFPI sensor fabrication. The multiplexing of microgap sensors with other sensors provides an effective method of temperaturecompensated measurement. Multicavity sensors for thin-film measurement were constructed by multiplexing two microgap sensors. This configuration not only has advantages similar to tubing-based MFPI sensors but is also distinguished by its simple fabrication process and flexible cavity lengths. Furthermore, the temperature sensitivity for temperature compensation has increased by an order of magnitude from the tubing-based MFPI sensor. A temperature resolution of 0.1 ◦ C has been achieved with the microgap sensor. Thus, thin-film characteristics under temperature variations can be examined more accurately. The high-temperature sensitivity of microgap MFPI sensors enables biosensing under temperature variations. This greatly improves the flexibility in sample handling and provides the opportunity to investigate temperature effects in biological applications.

1804

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 25, NO. 7, JULY 2007

R EFERENCES [1] B. Qi, G. Pickrell, P. Zhang, Y. Duan, W. Peng, J. Xu, Z. Huang, J. Deng, H. Xiao, Z. Wang, W. Huo, R. G. May, and A. Wang, “Fiber optic pressure and temperature sensors for oil down hole application,” presented at the Fiber Optic Sensor Technol. Applications Conf., Newton, MA, 2001. [2] V. Bhatia, K. A. Murphy, R. O. Claus, M. E. Jones, J. L. Grace, T. A. Tran, and J. A. Greene, “Multiple strain state measurements using conventional and absolute optical fiber-based extrinsic Fabry–Pérot interferometric strain sensors,” Smart Mater. Struct., vol. 4, no. 4, pp. 240–245, Dec. 1995. [3] B. Yu, D. W. Kim, J. D. Deng, H. Xiao, and A. Wang, “Fiber Fabry–Pérot sensors for detection of partial discharges in power transformers,” Appl. Opt., vol. 42, no. 16, pp. 3241–3250, Jun. 2003. [4] Y. Zhang, H. Shibru, K. L. Cooper, and A. Wang, “Miniature fiber-optic multicavity Fabry–Pérot interferometric biosensor,” Opt. Lett., vol. 30, no. 9, pp. 1021–1023, May 2005. [5] T. Bae, R. A. Atkins, H. F. Taylor, and W. N. Gibler, “Interferometric fiber-optic sensor embedded in a spark plug for in-cylinder pressure measurement in engines,” Appl. Opt., vol. 42, no. 6, pp. 1003–1007, Feb. 2003. [6] C. E. Lee, W. N. Gibler, R. A. Atkins, and H. F. Taylor, “In-line fiber Fabry–Pérot-interferometer with high-reflectance internal mirrors,” J. Lightw. Technol., vol. 10, no. 10, pp. 1376–1379, Oct. 1992. [7] P. Betts and J. A. Davis, “Bragg grating Fabry–Pérot interferometer with variable finesse,” Opt. Eng., vol. 43, no. 5, pp. 1258–1259, May 2004. [8] F. Shen, W. Peng, K. Cooper, G. Pickrell, and A. Wang, “UV-induced intrinsic Fabry–Pérot interferometric fiber sensors,” in Proc. Sens. Harsh Environ., Philadelphia, PA, 2004, pp. 47–56. [9] M. Gupta, H. Jiao, and A. O’Keefe, “Cavity-enhanced spectroscopy in optical fibers,” Opt. Lett., vol. 27, no. 21, pp. 1878–1880, Nov. 2002. [10] R. Slavik, S. Doucet, and S. LaRochelle, “High-performance all-fiber Fabry–Pérot filters with superimposed chirped Bragg gratings,” J. Lightw. Technol., vol. 21, no. 4, pp. 1059–1065, Apr. 2003. [11] X. D. Peng and C. Roychoudhuri, “Design of high finesse, wideband Fabry–Pérot filter based on chirped fiber Bragg grating by numerical method,” Opt. Eng., vol. 39, no. 7, pp. 1858–1862, Jul. 2000. [12] X. Chen, F. Shen, A. Wang, Z. Wang, and Y. Zhang, “Novel Fabry–Pérot fiber optic sensor with multiple applications,” in Proc. Sens. Harsh Environ., Philadelphia, PA, 2004, pp. 111–121. [13] Y. Zhang, K. L. Cooper, and A. Wang, “Multicavity Fabry–Pérot interferometric thin-film sensor with built-in temperature compensation,” IEEE Photon. Technol. Lett., vol. 17, no. 12, pp. 2712–2714, Dec. 2005. [14] K. Kurihara, H. Ohkawa, Y. Iwasaki, O. Niwa, T. Tobita, and K. Suzuki, “Fiber-optic conical microsensors for surface plasmon resonance using chemically etched single-mode fiber,” Anal. Chim. Acta, vol. 523, no. 2, pp. 165–170, 2004. [15] J. R. Epstein, M. Lee, and D. R. Walt, “High-density fiber-optic genosensor microsphere array capable of zeptomole detection limits,” Anal. Chem., vol. 74, no. 8, pp. 1836–1840, Apr. 2002. [16] R. Stockle, C. Fokas, V. Deckert, R. Zenobi, B. Sick, B. Hecht, and U. P. Wild, “High-quality near-field optical probes by tube etching,” Appl. Phys. Lett., vol. 75, no. 2, pp. 160–162, Jul. 1999. [17] T. Pangaribuan, K. Yamada, S. D. Jiang, H. Ohsawa, and M. Ohtsu, “Reproducible fabrication technique of nanometric tip diameter fiber probe for photon scanning tunneling microscope,” Jpn. J. Appl. Phys. 2, Lett., vol. 31, no. 9A, pp. L1302–L1304, 1992. [18] A. Klini, T. David, E. Bourillot, S. Emonin, P. Papadopoulos, J. P. Goudonnet, and G. Kotrotsios, “Reproducible optical fiber tips for photon scanning tunneling microscopy with very small (< 5◦ ) cone angle,” J. Lightw. Technol., vol. 16, no. 7, pp. 1220–1227, Jul. 1998. [19] F. B. Shen and A. B. Wang, “Frequency-estimation-based signalprocessing algorithm for white-light optical fiber Fabry–Pérot interferometers,” Appl. Opt., vol. 44, no. 25, pp. 5206–5214, Sep. 2005. [20] T. J. Cavicchi, Digital Signal Processing. New York: Wiley, 2000, pp. 630–631. [21] B. Qi, G. R. Pickrell, J. C. Xu, P. Zhang, Y. H. Duan, W. Peng, Z. Y. Huang, W. Huo, H. Xiao, R. G. May, and A. Wang, “Novel data processing techniques for dispersive white light interferometer,” Opt. Eng., vol. 42, no. 11, pp. 3165–3171, Nov. 2003. [22] S. Nemoto and T. Makimoto, “Analysis of splice loss in single-mode fibers using a Gaussian field approximation,” Opt. Quantum Electron., vol. 11, no. 5, pp. 447–457, 1979. [23] D. Marcuse, “Loss analysis of single-mode fiber splices,” Bell Syst. Tech. J., vol. 56, no. 5, pp. 703–718, May/Jun. 1977. [24] Y. S. Touloukian, Thermal Expansion-Nonmetallic Solids. New York: Plenum, 1977, p. 358.

[25] M. J. Weber, Handbook of Optical Materials. Boca Raton, FL: CRC, 2003, p. 239. [26] Y. Lvov, G. Decher, and H. Mohwald, “Assembly, structural characterization, and thermal-behavior of layer-by-layer deposited ultrathin films of Poly(Vinyl Sulfate) and Poly(Allylamine),” Langmuir, vol. 9, no. 2, pp. 481–486, 1993.

Yan Zhang received the Ph.D. degree in electrical engineering from the Virginia Polytechnic Institute and State University (Virginia Tech), Blacksburg, in 2005. He is currently with the Fitzpatrick Institute for Photonics, Duke University, Durham, NC. His current research interests are in the area of fiber-optic sensors, biomedical sensing, and imaging. Dr. Zhang is a member of the Optical Society of America and of the International Society for Optical Engineers.

Xiaopei Chen (S’03) received the Ph.D. degree in electrical engineering from the Virginia Polytechnic Institute and State University (Virginia Tech), Blacksburg, in 2006. She is currently with the Center for Photonics Technology, Bradley Department of Electrical and Computer Engineering, Virginia Tech. Her research activities focus on the areas of fiber-optic sensors, optical fiber communications, laser characterizations, and optical fiber devices.

Yongxin Wang is currently working toward the Ph.D. degree at the Center for Photonics Technology, Bradley Department of Electrical and Computer Engineering, Virginia Polytechnic Institute and State University (Virginia Tech), Blacksburg. His research interests are in the areas of fiber-optic sensors, signal processing, and circuit design.

Kristie L. Cooper (S’96–M’01) received the B.S., M.S., and Ph.D. degrees in electrical engineering from the Virginia Polytechnic Institute and State University (Virginia Tech), Blacksburg. Since 2002, she has been with the research faculty at Virginia Tech’s Center for Photonics Technology, where her research has focused on nanoscale optical materials and photonic sensors for chemical, physical, and biomedical applications. She has authored or coauthored more than 80 technical papers and has been responsible for over 40 sponsored research programs in excess of $8 million.

Anbo Wang (M’91–SM’92) received the Ph.D. degree in applied optics from Dalian University of Technology, Dalian, China, in 1990. After spending three years as a Research Staff in the Fiber and Electro-Optics Research Center, Department of Electrical and Computer Engineering, Virginia Polytechnic Institute and State University (Virginia Tech), Blacksburg, he then joined the faculty of the same department as an Assistant Professor, where he is currently a Professor and the Founding Director of the Center for Photonics Technology. He is the author/coauthor of about 90 journal articles and 140 conference papers and is the holder of nine patents, all licensed to industry. His research interests include sensors, laser characterization, and tests on the constancy of physical constants. Dr. Wang has chaired/cochaired numerous national or international conferences on photonics and sensors.