Highly Sensitive Nonlinear Temperature Sensor Based ... - IEEE Xplore

3720

IEEE SENSORS JOURNAL, VOL. 17, NO. 12, JUNE 15, 2017

Highly Sensitive Nonlinear Temperature Sensor Based on Modulational Instability Technique in Liquid Infiltrated Photonic Crystal Fiber Nagarajan Nallusamy, R. Vasantha Jayakantha Raja, and G. Joshva Raj

Abstract— A highly sensitive nonlinear temperature sensor, which is based on modulational instability (MI) process, is theoretically demonstrated for the first time using a CS2 -filled photonic crystal fiber (CSPCF). The proposed novel temperature sensor works on the principle of measurement of a temperaturedependent wavelength shift of generated Stokes and anti-Stokes MI Sidebands. Based on the notion of MI dynamics, the performance of the proposed temperature sensor is studied in both anomalous and normal dispersion regimes of an appropriately designed CSPCF. It is found that the sensitivity of the proposed nonlinear temperature sensor is very low when the CSPCF is pumped in the anomalous dispersive region. However, the sensitivity is enhanced by more than 66 times using Stokes line in the normal dispersion regime. The proposed sensor is optimized by varying the structural parameters and pump parameters, such as pitch, air-hole diameter, pump wavelength, and pump power. The proposed nonlinear temperature sensor, which is made up of an appropriate structure of CSPCF having a length of 13 cm exhibits a sensitivity of −82 nm/°C using Stokes line and 435 nm/°C using anti-Stokes line while pumped with a power of 100 W in the normal dispersive region. Index Terms— Modulational instability, photonic crystal fiber, temperature sensor, four-wave mixing.

I. I NTRODUCTION

I

N THE recent years, photonic crystal fiber (PCF) based temperature sensors have been more popular as they find indispensable advantages such as wide detectable range, high sensitivity and simpler configuration. Indeed, PCF based optical sensors are compact and small in size and easily integrable with other devices [1]. The main advantage of PCF based temperature sensors is that the sensitivity can be enhanced not only by an optimized structure of PCF but also by employing alternate materials such as polymer [2], chalcogenide glasses [3], nonlinear liquids [4], etc., for PCF fabrication. A large number of different approaches have already been proposed both theoretically and experimentally using PCFs, such as Fabry-Perot interferometer [5], Mach-Zehender interferometer [6], [7], surface plasmon resonance [8], etc., Manuscript received March 6, 2017; revised April 20, 2017; accepted April 21, 2017. Date of publication April 28, 2017; date of current version May 22, 2017. This work was supported in part by the Fast Track Fellowship under Grant SR/FTP/PS-096/2012 and in part by the CSIR under Grant 03(1360)/16/EMR-II. The associate editor coordinating the review of this paper and approving it for publication was Dr. Daniele Tosi. (Corresponding author: R. Vasantha Jayakantha Raja.) The authors are with the Centre for Nonlinear Science and Engineering, School of Electrical and Electronics Engineering, SASTRA University, Thanjavur 613401, India (e-mail: [email protected]) Digital Object Identifier 10.1109/JSEN.2017.2699186

in order to improve the performance of temperature sensor. For instance, Lin et al., [9] reported a liquid filled PCF based multimodal interferometer temperature sensor with a sensitivity of −24.757 nm/°C and a temperature range of 19.5°C to 22.5°C. Naeem et al., [2] reported a highly sensitive temperature sensor with a sensitivity of 1.595 nm/°C, and a temperature range of 32°C to 38°C using a two core PCF which is selectively filled up with a polymer. A plasmonic temperature sensor based on a PCF which is coated with a nanoscale gold film is reported with a sensitivity of −2.15 nm/°C in Ref. [10]. Qiu et al., fabricated a simple and compact inline non-polarimetric modal interferometer using a liquidsealed PCF, in order to realize a temperature sensor with a sensitivity of −166 pm/°C [11]. In the year of 2012, an optical fiber Sagnac interferometer based temperature sensor was constructed by a selectively filled polarization maintaining PCF with a sensitivity of 2.58 nm/°C [12]. Recently, Ayyanar et al., have proposed a temperature sensing configuration with a sensitivity of 42.99nm/°C using an asymmetric core mode in a choloform filled PCF of length 1.41 cm [4]. In addition to the above sensing mechanisms, nonlinear phenomena such as four-wave mixing (FWM) [13], [14], modulational instability (MI) [15], [16], etc., have also been recently used to realize fiber optic based sensors. The MI can be interpreted as a FWM process, phase matched through Kerr nonlinearity and dispersive effects, generating two sidebands of frequencies around the pump photon which are called Stokes and anti-Stokes lines [17]. The MI based fiber optic sensors are working on the basis of detecting the frequency shift in Stokes or anti-Stokes photons. The main advantages of MI based fiber sensors are that they are too simple to be experimentally realized with easily available components and require simple fabrication techniques. The first nonlinear sensor based on MI technique was proposed by Ott et al., [18] for label-free and selective nonlinear fiber optical biosensor. Using chalcogenide fiber, markos et al., demonstrated an optical fiber sensor for biosensing applications [3]. Then the idea was extended by Frosz et al., [19] to experimentally demonstrate a highly sensitive compact refractive index sensor. Based on the shift of FWM Stokes and anti-Stokes peaks, a strain sensor also is demonstrated [14], [20]. Even though several authors have reported different techniques for the realization of temperature sensor using PCF, MI based temperature sensing mechanism has not been reported yet to the best of our knowledge. In this line, we have opened the way for realizing

1558-1748 © 2017 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

NALLUSAMY et al.: HIGHLY SENSITIVE NONLINEAR TEMPERATURE SENSOR BASED ON MI TECHNIQUE IN LIQUID INFILTRATED PCF

Fig. 1. The schematic diagram of the mechanism of temperature sensor using MI process in a CSPCF.

an MI based temperature sensor by studying the MI process under the influence of temperature [21]. It is found that the MI sidebands make a significant variation under the influence of temperature and such temperature dependent characteristic of MI motivates us to propose an MI based temperature sensor. The main objective of this paper is to theoretically propose a highly sensitive temperature sensor that is based on MI technique, using a liquid infiltrated PCF. In order to accomplish a highly sensitive temperature sensor using liquid core PCF, it is required that the liquid is needed to be highly temperature sensitive. Among the several liquids that have been used to fabricate PCFs, CS2 emerges as an interesting candidate [22], [23]. Further, the refractive index of CS2 is highly temperature dependent. Also, CS2 liquid has a wide transmission window from visible to mid IR regimes which will help to improve the detecting range of temperature. Even though CS2 liquid is highly toxic and volatile, due to the higher nonlinear coefficient and wide transmission window, they have already been experimentally used in the applications such as supercontinuum generation [24], slow light [25], etc. In addition, focused ion beam milled microchannels method has been realized for the fabrication of liquid infiltrate in PCF [26]. The paper is organized as follows: sec II explains the sensing mechanism which we use to realize a temperature sensor by deriving the MI gain using a modified nonlinear Schrödinger equation (MNLSE). In Sec. III, an appropriate structure of CSPCF is proposed. In sec. IV, the suitable region of nonlinear pulse propagation, in order to achieve maximum sensitivity, is identified by investigating anomalous and normal dispersive regions. Sec. V is dedicated to maximize the sensitivity of the proposed sensor by optimizing the parameters, viz. pitch, air-hole size, pump wavelength, and pump power. The conclusions of the present work are drawn in Sec. VI. II. S ENSING M ECHANISM To understand the sensing mechanism by MI technique, the pulse propagation in PCF with fourth order dispersion (FOD) and Kerr nonlinearity may be described by the following MNLSE [27]

3721

Fig. 2. The cross section and the field distribution of the fundamental mode guided by the proposed CSPCF with lattice pitch () of 7.2 μm and diameter of of the airhole (d) is 5.85 μm with the core diameter (dc = d/3).

respectively. βñ is the temperature dependent nt h order coefficient of dispersion parameter and γ˜ is the temperature dependent Kerr parameter. The process of MI leading to the sensing mechanism can be understood by the linear stability analysis. The MI gain corresponding to the dispersion relation is calculated by G()=2 Im(K) as follows: ˜ 2 4 2 β 2 + β˜4 + β˜4 , (2) G() = 2γ˜ P0 + β˜2 2 24 2 24 where P0 is the pump power. We now calculate the peak frequency which is also called optimal modulation frequency opt of MI generated Stokes and anti-Stokes lines at which the gain reaches its maximum value as follows:

(3) opt = 2 3(3β˜22 − 2P0 γ˜ β˜4 )/β˜4 − 6β˜2 /β˜4 , The detection mechanism of nonlinear temperature sensor is working on the basis of temperature dependent shift of the peak wavelength opt of MI gain. As dispersion and nonlinear parameter of the proposed CSPCF are a function of temperature, the peak wavelength of the generated Stokes (ω S = ω0 − ) and anti-Stokes (ω As = ω0 + ) lines make temperature dependent shift. The schematic representation of the MI based sensing mechanism is described in Fig. 1 III. F IBER D ESIGN In order to realize the temperature sensing scheme which is introduced in the previous section, an appropriate structure of CSPCF is designed as a first step.The cross section and the field distribution of the fundamental mode guided by the proposed CSPCF is shown in Fig. 2. The background of the CSPCF material is silica and central airhole is filled with C S2 liquid. The temperature dependent refractive indices of silica are taken as in [10]. The refractive indices of C S2 liquid at 20°C are calculated using the following Sellmeier’s equation for the wavelength λ: n 20 (λ) = 1.580826 + 1.52389 × 10−2 λ−2

(1)

+ 4.8578 × 10−4 λ−4 − 8.2863 × 10−5 λ−6 + 1.4619 × 10−5 λ−8 , (4)

where U is the slowly varying amplitude of electrical-field envelope, and z, t represent propagation distance and time

where n 20 (λ) represents the refractive index of CS2 liquid at 20◦ C temperature for a given wavelength λ. By using the refractive indices obtained by Eq. 4 for 20◦ C, refractive

i

∂U + ∂z

4 n=2

βñ

i n−1

∂nU

n! ∂t n

+ γ˜ |U |2 U = 0,

3722

Fig. 3. Variation of n e f f and ZDW as a function of temperature at the pumping wavelength 1550 nm and 1420 nm in the anomalous and normal dispersion regime.


Fig. 4. Variation of β2 and β4 as a function of temperature in anomalous and normal dispersion regimes with the pumping wavelength 1550 nm and 1420 nm.

indices for other temperatures are calculated by the following empirical relation, dn , (5) n(λ, T ) = n 20 (λ) + T dT where T denotes the desired temperature for which refractive indices are needed to be calculated and T is the difference between T and 200 C. dn/d T gives the value of the rate of variation of refractive index as a function of temperature and it is measured as −7.91 × 10−4 K−1 for CS2 [21], [28]. Using fully vectorial effective index method [29], effective refractive index of the CSPCF is calculated from which other fiber parameters such as second order dispersion (SOD), FOD, effective area (Ae f f ) and nonlinearity are evaluated. The CSPCF is designed in such a way that it should be almost single mode and FOD possess a negative value for the considered range of pump wavelength in order to induce MI. As CS2 material has a high refractive index than that of silica, the CSPCF bring multi mode propagation for larger d/ values. In order to reduce the V-parameter [29] [30], the core diameter is reduced by a factor 3 of airhole diameter. On numerical simulations, it is found that a CSPCF which is having structural parameters of pitch 7.2 μm, diameter of airhole 5.85 μm, and core diameter 1.95 μm supports anomalous dispersion at the wavelength of 1550 nm. The V parameter of the proposed CSPCF structure is calculated as 2.852 which indicates that the fiber is slightly multi-moded. In order to make the CSPCF as a single mode, a technique which is called adiabatic coupling can be adopted [30], [31]. Fig. 3 depicts the variation of effective index and ZDW as a function of temperature for a range of 26°C to 27°C. It is observed that the ZDW shifts towards the higher wavelength region, from 1423 nm to 1424.05 nm as the temperature is increased from 26°C to 27°C. From Fig. 4, it is found that β2 can be tuned from −0.03598 ps 2 /m to −0.03562 ps 2 /m in the anomalous dispersion regime and 8.13065×10−4 ps 2 /m to 0.00113 ps 2 /m in the normal dispersion regime. Similarly, β4 can be tuned from −2.83368 × 10−8 ps 4 /m to −2.83105 × 10−8 ps 4 /m in the anomalous dispersion regime and −1.75345 × 10−8 ps 4 /m to −1.75319 × 10−8 ps 4 /m in the normal dispersion regime. As an attempt to improve the

Fig. 5. Variation of effective area and nonlinearity as a function of temperature in the anomalous (1550 nm) and the normal dispersion (1420 nm) regimes.

sensitivity, in addition to the linear refractive index, nonlinear refractive index of CS2 which is given below is also considered [32]: n2 =

12π 2 × 103 × χ (3) (T ), (n 0 )2 c

(6)

where c is the velocity of light. The relationship between third order susceptibility χ (3) and temperature (T) is given by the following empirical relation [28]: χ (3) (T ) = C + D[1 + p(T − T1 )u ]e G/T ,

(7)

where T1 = 207.7 K, C = 0.4600, D = −0.4000 × 10−2 , p = 1.548 × 10−4 , u = 1.790 and G = 0.8000 × 103 . By calculating the effective area of CSPCF and using Eq. 6, the value of nonlinearity is calculated and depicted in Fig. 5. The variation of nonlinearity values clearly indicates that one could adjust this value by changing temperature. Thus, all the fiber parameters of the proposed CSPCF are found to be tunable by changing the temperature from 26°C to 27°C. IV. N ONLINEAR T EMPERATURE S ENSOR A. Anomalous Dispersion Regime In order to propose a highly sensitive nonlinear PCF based temperature sensor, as a first step, the temperature sensitivity is calculated in the anomalous dispersive region using Stokes


Fig. 6. The generated MI sidebands in the anomalous dispersion (1550 nm) regime for different temperatures for the design which is characterized by a lattice pitch of = 7.2 μm, diameter of airhole of d = 5.85 μm, and core diameter of 1.95 μm at the pump power of 50 W.

Fig. 7. The Variation of peak wavelength (OMF) as a function of temperature in the anomalous dispersion regime at the pump wavelength of 1550 nm with a pump power of 50 W.

and anti-Stokes sidebands. For this purpose a pulse is launched into the fiber at 1550 nm with a power of 50 W. From Fig. 6, it is found that the pulse propagation through the proposed CSPCF undergoes spectral modifications and evolute into MI sidebands under a phase matched condition between Kerr nonlinearity and dispersion which is derived in [27]. Thus, MI generates Stokes and anti-Stokes photons around the pump wavelength by obeying the energy conservation law. The variation of MI gain as a function of temperature for a range of temperature from 26°C to 27°C is plotted in Fig. 6. The corresponding variation of peak wavelengths of Stokes and anti-Stokes sidebands is calculated using Eq. 3 and represented in Fig. 7. From Fig. 7, a set of six peak wavelengths are observed at 1450.9 nm, 1450.8 nm, 1450.7 nm, 1450.7 nm, 1450.6 nm, and 1450.5 nm for temperatures 26°C, 26.2°C, 26.4°C, 26.6°C, 26.8°C, and 27°C respectively. Thus, the peak wavelength of the Stokes photon makes blue shift from 1450.9 nm to 1450.5 nm as temperature is increased from 26°C to 27°C with shift of −0.4003 nm. At the same time, the peak wavelengths for the anti-Stokes sideband are observed at 1661.8 nm, 1661.9 nm, 1662.0 nm, 1662.1 nm, 1662.2 nm, and 1662.3 nm for the same temperature range and the total wavelength shift

3723

Fig. 8. The variation of sensitivity as a function of temperature for Stokes and anti-Stokes lines in anomalous dispersion regime.

Fig. 9. The generated MI sidebands in the normal dispersion (1420 nm) region for different temperatures for the design of lattice pitch of () = 7.2 μm and diameter of the airhole (d)= 5.85 μm with core diameter of 1.95 μm at the pump power of 50 W.

is calculated as 0.5255 nm. The results show that the peak wavelength changes linearly with the temperature in the range of 26°C to 27°C. The sensitivity of the MI sensor is calculated through the shift of peak wavelength for the variation in temperature. The sensitivity can be defined as λ p , (8) T where λ p is the peak wavelength. Fig. 8 depicts the calculated wavelength shift of the MI spectrum for each 0.2°C increase in temperature from 26°C. Calculations show a temperature sensitivity of −0.4003/°C for Stokes line and 0.5255 nm/°C for anti-Stokes line in the anomalous dispersive region. S=

B. Normal Dispersion Regime Now, the proposed sensing mechanism is studied in the normal dispersive regime. For this purpose, a pulse of 50 W is launched into the proposed CSPCF at a wavelength of 1420 nm which falls in the normal dispersive region. It is found that the peak wavelengths of Stokes and anti-Stokes photons experience red and blue shift respectively with the increase of temperature which is similar to the case of anomalous dispersive regime. At the same time, it is interestingly observed that the sidebands make larger shift in the normal dispersive region

3724


Fig. 10. The variation of peak wavelength (OMF) as a function of temperature in normal dispersion regime at the pump wavelength of 1420 nm with pump power of 50 W.

than in the anomalous dispersive region. This is due to the interaction between β4 and normal dispersion characteristic of the medium. Fig. 9 shows the temperature variation of MI gain for Stokes line only because of the anti-Stokes line falls far away from the pump wavelength. Even though the generated MI sidebands cover a wide wavelength spectrum which extend to far-IR region, they can be easily detectable by the state-of-the-art IR detector which can measure radiation up to 12μm. Fig. 10 depicts the position of peak wavelengths of the generated Stokes and anti-Stokes lines for the temperature range of 26°C to 27°C. It is found from the Fig. 10 that the peak wavelengths for Stokes line are observed at 847.4 nm, 840.45 nm, 833.56 nm, 826.81 nm, 820.20 nm, and 813.74 nm for temperatures 26°C, 26.2°C, 26.4°C, 26.6°C, 26.8°C, and 27°C, respectively and the total wavelength shift is calculated as −33.75 nm. The peak wavelengths for the anti-Stokes line are observed at 4326.8 nm, 4559.1 nm, 4773.2 nm, 5007.3 nm, 5264.2 nm, and 5547.1 nm respectively for the same temperature range and the total shift is observed as 1184 nm. Fig. 11 portrays the deviation of peak wavelength for each 0.2°C increment in temperature from 26°C. The sensitivity is numerically calculated as −33.75 nm/°C for Stokes line and 1184 nm/°C for anti-Stokes line in the normal dispersive region. It is found that the sensitivity of the proposed scheme can be increased by more than 66 times using the Stokes line of a pulse of power 50 W which is launched at 1420 nm in the normal dispersion regime than in the case of anomalous dispersive region. Similarly, using anti-Stokes line of the same pulse which is launched at normal dispersive regime, a sensitivity of more than 2368 times is achieved. V. O PTIMIZATION In this section, the focus is shifted towards the optimization of the proposed sensing configuration. From Eqs. 2 and 3, it is identified that MI gain and OMF (peak wavelength) are function of Kerr nonlinearity, dispersion, pump power, pump wavelength, and length of the fiber and the generated MI gain is temperature tunable too. Among these dependent parameters of the sensitivity of the proposed sensing mechanism,

Fig. 11. The variation of sensitivity as a function of temperature for Stokes and anti-Stokes lines in the normal dispersion regime.

nonlinearity and dispersion parameters highly depend on the structure of CSPCF. Hence, the design of the CSPCF is first optimized for maximum sensitivity and then pump wavelength and pump power are considered. A. Optimization of Structural Parameters The objective of optimization of the aforementioned parameters is to achieve maximum OMF shift which in turn increase the sensitivity. As identified earlier, MI sidebands in the normal dispersion regime respond more rapidly to the temperature. Hence, we intent to optimize the design parameters of CSPCF in order to get maximum shift by launching the pulse at 1420 nm with a pump power of 50 W. The optimization of design parameters is done with the following condition: the ZDW is fixed at 1423 nm for a temperature of 26°C which is obtained for the design of pitch value of 7.2μm and air hole diameter value of 5.85μm with the core diameter dc = d/3. Fig. 12 depicts the calculated pitch values for various air hole diameter which support ZDW at 1423 nm for the temperature 26°C. It is found that the FOD turns to be positive at and below the pitch value of 7.1 μm. The peak wavelength of Stokes and anti-Stokes lines of the MI gain are calculated for the same temperature range. It is calculated from Fig. 13 that the maximum MI gain shift is achieved as −39.46 nm/°C and 16285 nm/°C respectively for Stokes and anti-Stokes lines for the lattice pitch value of 7.1μm and air-hole diameter value of 5.752μm with the core diameter of 1.91μm. Thus, the optimization of the design parameters shows that the maximum sensitivity is obtained for the pitch value of 7.1μm and air-hole diameter value of 5.752μm at a pump wavelength of 1420 nm. B. Optimization of Pump Wavelength The pump wavelength is optimized by varying it across a wide range of wavelength that encompass both normal and anomalous dispersive regimes in the structurally optimized CSPCF with a pump power of 50W. Fig. 14 portrays the variation of sensitivity under the influence of pumping wavelength for Stokes and anti-Stokes lines. The figure clearly shows that the shift of stokes and anti-Stokes photons move away from


Fig. 12. Variation of airhole diameter (d) and pitch () to keep ZDW of the CSPCF at 1420 nm at a pump power of 50 W.

3725

Fig. 14. Sensitivity plot for the optimized design for which the lattice pitch is () = 7.1 μm and airhole diameter (d) is 5.752 μm with a core diameter of 1.91 μm for different wavelengths at a pump power of 50 W.

Fig. 13. Sensitivity Vs difference designs at a wavelength of λ = 1420 nm with a pump power of 50 W for Stokes and anti-Stokes lines.

the pump wavelength under the influence of temperature when the pulse is launched nearer to ZDW. However, the quantum of frequency shift is suddenly reduced beyond the ZDW as β2 start to dominate β4 . Here we optimized the wavelength where the sensitivity is maximum in stokes line which falls in the easily detectable regime. The numerical simulations show that the sensitivity of the proposed CSPCF based temperature sensor is −53.34 nm/°C using Stokes line and 944.5nm/°C using anti-Stokes line at 1423 nm. C. Power Optimization and Length Finally, the proposed nonlinear temperature sensing scheme is optimized for pump power and length of the fiber. In order to find the exact length and required power, it is better to solve Eq. 1 through direct numerical simulation. As the proposed CSPCF has a large loss when compared to silica PCF, we include the loss term in the mathematical model so that experimental results can be reproduced in our numerical analysis. Hence, the Eq. 1 has been modified as α˜ i n−1 ∂ n U ∂U +i + βn + γ˜ |U |2U = 0, ∂z 2 n! ∂t n 4

i

(9)

n=2

where αˆ stands for temperature dependent loss. As the loss value does not change significantly with respect to temperature, the value 0.45 dB/m corresponding to 26°C is considered

Fig. 15. The evolution of MI as a function of length of the CSPCF at a temperature of 26°C for the pump wavelength of 1423 nm with a pump power of 100 W.

for this simulations. To investigate the power and the length in the proposed PCF, we numerically solve Eq. 9 using split step Fourier method with initial of Gaussian pulse envelope √ t2 given by U(0, t) = P0 exp − 2T 2 . Numerical simulation 0

is carried out for the input pulse by setting the optimized wavelength λ = 1423 nm and the pulse width having full width at half maximum (FWHM) of 6 ps. The minimum power required to generate MI in the lossy fiber can be calculated as [33] Pt h =

5 , 2γ L e f f

(10)

where L e f f = [1 − exp(−αL)]/α. It has been calculated through Eq. 10 that the minimum power value to generate MI at 26°C is 6.5 W for the considered length of the fiber 15 cm. But, this power value is not enough to generate MI at 27°C. To optimize the sensor, the MI should be generate at both 26°C and 27°C at that time. Hence, we considered the Gaussian pulse with a moderate power of 100 W. To implement experimental conditions in our theoretical analysis, we have

3726


of the fiber for the optimized design of pitch value or 7.1 μm and diameter of the air-hole value of 5.752 μm with a core diameter of 1.91 μm at 1423 nm. The sensitivity in the proposed nonlinear based temperature sensor is almost 10 times greater than that of [4]. The sensor can be used for a temperature range of 20°C to boiling point of the CS2 . However, as the ZDW shifts with respect to temperature, one has to change the pumping wavelength according to the variation of ZDW to get the same maximum sensitivity. VI. C ONCLUSION

Fig. 16. The evolution of MI as a function of different length of CSPCF at a temperature of 27°C for the pump wavelength of 1423 nm with the pump power of 100 W.

Fig. 17. Spectral density of the proposed CSPCF for different temperatures at the pump power of 100 W with pump wavelength of 1423 nm.

To conclude, a highly sensitive temperature sensor is theoretically proposed using the wavelength shift of MI sidebands. A sensitivity of −0.4003 nm/°C is achieved by using the Stokes sideband which is generated by pumping the proposed CSPCF in the anomalous regime. However, the sensitivity is increased by 66 times than that of anomalous dispersion regime when the pulse is launched in the normal dispersion region. In order to design an efficient temperature sensor, the sensitivity is optimized by using structural parameters, and pump parameters such as pump wavelength and pump power. The optimized configuration of the proposed sensing mechanism reveals the hitherto record sensitivity of −82nm/°C using Stokes line and 435nm/°C using anti-Stokes lines at 1423 nm with a low pump power of 100 W, using a CSPCF design which is optimized with a pitch value of () = 7.1μm and diameter value of the air-hole (d) = 5.752μm with the core diameter of 1.91μm with length of 13cm for the stable temperature configuration. The proposed CSPCF temperature sensor is mainly suitable for biological and chemical sensing applications. ACKNOWLEDGMENT

calculated spectral density using λ max

S(z, λ) =

S(z, λ)dλ = λmin

c frep |U˜ (z, v)|2 , λ2

(11)

where the repetition rate frep is considered as 80 MHz. The calculated spectral density and the fiber length are depicted in Fig. 15 and Fig. 16. It is observed from these figures that the generated MI sidebands become supercontinuum spectrum if the pulse is allowed to be propagated beyond 20 cm. Also, it has been observed that MI is generated at a shorter distance for 26°C and it becomes supercontinuum at a shorter distance when compared with 27°C. Thus, we fix the length as 13 cm where one can detect the Stokes and anti-Stokes lines for both 26°C and 27°C simultaneously. At this length, it is calculated that the MI gain value by Stokes and anti-Stokes for 26°C and 27°C is grater than 50 dB/m. Fig. 17 depicts the wavelength shift of Stokes and anti-Stokes lines of MI for 26°C and 27°C at a fiber length of 13 cm. From this figure,it is calculated that the maximum shift for Stokes and anti-Stokes lines is obtained from 992 nm to 910 nm and from 2510 nm to 2945 nm respectively. It is calculated numerically that the largest Stokes wavelength shift of −82 nm/°C and 435 nm/°C is obtained with a pump power of 100 W at 13 cm length

R. Vasantha Jayakantha Raja acknowledges DST for providing financial assistantship through Fast track fellowship (SR/FTP/PS-096/2012) and CSIR: 03(1360)/16/EMR-II. R EFERENCES [1] L. Rindorf and O. Bang, “Sensitivity of photonic crystal fiber grating sensors: Biosensing, refractive index, strain, and temperature sensing,” J. Opt. Soc. Amer. B, vol. 25, no. 3, pp. 310–324, 2008. [2] K. Naeem, B. H. Kim, B. Kim, and Y. Chung, “High-sensitivity temperature sensor based on a selectively-polymer-filled two-core photonic crystal fiber in-line interferometer,” IEEE Sensors J., vol. 15, no. 7, pp. 3998–4003, Jul. 2015. [3] C. Markos and O. Bang, “Nonlinear label-free biosensing with high sensitivity using As2 S3 chalcogenide tapered fiber,” IEEE J. Lightw. Technol., vol. 33, no. 13, pp. 2892–2898, Jul. 1, 2015. [4] N. Ayyanar, R. B. V. J. Raja, D. Vigneswaran, B. Lakshmi, M. Sumathi, and K. Porsezian, “Cover image Highly efficient compact temperature sensor using liquid infiltrated asymmetric dual elliptical core photonic crystal fiber,” Opt. Mater., vol. 64, pp. 574–582, Feb. 2017. [5] R. Jha, J. Villatoro, G. Badenes, and V. Pruneri, “Refractometry based on a photonic crystal fiber interferometer,” Opt. Lett., vol. 34, no. 5, pp. 617–619, Mar. 2009. [6] X. B. Zheng, Y. G. Liu, Z. Wang, T. T. Han, C. L. Wei, and J. Chen, “Transmission and temperature sensing characteristics of a selectively liquid-filled photonic-bandgap-fiber-based Sagnac interferometer,” Appl. Phys. Lett., vol. 100, no. 14, p. 141104, 2012. [7] H. Liang et al., “Fiber in-line Mach–Zehnder interferometer based on near-elliptical core photonic crystal fiber for temperature and strain sensing,” Opt. Lett., vol. 38, no. 20, pp. 4019–4022, Oct. 2013.


[8] C. Liu et al., “A highly temperature-sensitive photonic crystal fiber based on surface plasmon resonance,” Opt. Commun., vol. 359, pp. 378–382, Jan. 2016. [9] W. Lin et al., “Liquid-filled photonic-crystal-fiber-based multimodal interferometer for simultaneous measurement of temperature and force,” Appl. Opt., vol. 54, no. 6, pp. 1309–1313, 2015. [10] Q. Liu, S. Li, H. Chen, J. Li, and Z. Fan, “High-sensitivity plasmonic temperature sensor based on photonic crystal fiber coated with nanoscale gold film,” Appl. Phys. Exp., vol. 8, no. 4, p. 046701, 2015. [11] S.-J. Qiu, Y. Chen, F. Xu, and Y.-Q. Lu, “Temperature sensor based on an isopropanol-sealed photonic crystal fiber in-line interferometer with enhanced refractive index sensitivity,” Opt. Lett., vol. 37, no. 5, pp. 863–865, 2012. [12] Y. Cui, P. P. Shum, D. J. J. Hu, G. Wang, G. Humbert, and X.-Q. Dinh, “Temperature sensor by using selectively filled photonic crystal fiber Sagnac interferometer,” IEEE Photon. J., vol. 4, no. 5, pp. 1801–1808, Oct. 2012. [13] H. M. Frosz, T. Sørensen, and O. Bang, “Nanoengineering of photonic crystal fibers for supercontinuum spectral shaping,” J. Opt. Soc. Amer. B, vol. 23, no. 8, pp. 1692–1699, 2006. [14] B. Gu, W. Yuan, M. H. Frosz, A. P. Zhang, S. He, and O. Bang, “Nonlinear fiber-optic strain sensor based on four-wave mixing in microstructured optical fiber,” Opt. Exp., vol. 37, no. 5, pp. 794–796, 2012. [15] A. Hasegawa, “Generation of a train of soliton pulses by induced modulational instability in optical fibers,” Opt. Lett., vol. 9, no. 7, pp. 288–290, 1984. [16] J. D. Harvey et al., “Scalar modulation instability in the normal dispersion regime by use of a photonic crystal fiber,” Opt. Lett., vol. 28, no. 22, pp. 2225–2227, 2003. [17] G. P. Agarwal, Nonlinear Fiber Optics. 4th ed. San Diego, CA, USA: Academic, 2007. [18] J. R. Ott, M. Heuck, C. Agger, P. D. Rasmussen, and O. Bang, “Label-free and selective nonlinear fiber-optical biosensing,” Opt. Exp., vol. 16, no. 25, pp. 20834–20847, Dec. 2008. [19] M. H. Frosz, A. Stefani, and O. Bang, “Highly sensitive and simple method for refractive index sensing of liquids in microstructured optical fibers using four-wave mixing,” Opt. Exp., vol. 19, no. 11, pp. 10471–10484, 2011. [20] S.-H. Tang, Z.-X. Wu, F. Xu, and Y.-Q. Lu, “Simulation of optical microfiber strain sensors based on four-wave mixing,” IEEE Sensors J., vol. 16, no. 9, pp. 3068–3074, May 2016. [21] G. J. Raj, R. V. J. Raja, N. Nagarajan, and R. Ganapathy, “Tunable broadband spectrum under the influence of temperature in IR region using CS2 core photonic crystal fiber,” J. Lightw. Technol., vol. 34, no. 15, pp. 3503–3509, Aug. 1, 2016. [22] R. V. J. Raja, A. Husakou, J. Hermann, and K. Porsezian, “Supercontinuum generation in liquid-filled photonic crystal fiber with slow nonlinear response,” J. Opt. Soc. Amer. B, vol. 27, no. 9, pp. 1763–1768, 2010. [23] R. V. J. Raja, K. Porsezian, and K. Nithyanandan, “Modulationalinstability-induced supercontinuum generation with saturable nonlinear response,” Phys. Rev. A, Gen. Phys., vol. 82, no. 013825, p. 013825, 2010. [24] D. Churin, T. N. Nguyen, K. Kieu, R. A. Norwood, and N. Peyghambarian, “Mid-IR supercontinuum generation in an integrated liquid-core optical fiber filled with CS2 ,” Opt. Mat. Exp., vol. 3, no. 9, pp. 1358–1364, 2013. [25] D. O. Herrera, L. Schneebeli, K. Kieu, R. A. Norwood, and N. Peyghambarian, “Slow light based on stimulated Raman scattering in an integrated liquid-core optical fiber filled with CS2 ,” Opt. Exp., vol. 21, no. 7, pp. 8821–8830, 2013. [26] F. Wang, W. Yuan, O. Hansen, and O. Ban, “Selective filling of photonic crystal fibers using focused ion beam milled microchannels,” Opt. Exp., vol. 19, no. 18, pp. 17585–17590, 2011.

3727

[27] S. Pitois and G. Millot, “Experimental observation of a new modulational instability spectral window induced by fourth -order dispersion in a normally dispersive single-mode optical fiber,” Opt. Commun., vol. 226, nos. 1–6, pp. 415–422, 2003. [28] D. Fu-xi and W. Bai-shi, “The temperature-dependence of the 3rd-order nonlinear susceptibility chi((3)) of CS2 and dodecylbenzene,” Appl. Opt., vol. 33, no. 24, pp. 5495–5497, 1994. [29] R. V. J. Raja and K. Porsezian, “A fully vectorial effective index method to analyse the propagation properties of microstructured fiber,” Photon. Nanostruct.-Fundam. Appl., vol. 5, no. 4, pp. 171–177, 2007. [30] R. Zhang, J. Teipel, and H. Giessen, “Theoretical design of a liquidcore photonic crystal fiber for supercontinuum generation,” Opt. Exp., vol. 14, no. 15, pp. 6800–6812, 2006. [31] J. D. Love, W. M. Henry, W. J. Stewart, R. J. Black, S. Lacroix, and F. Gonthier, “Tapered single-mode fibres and devices. Part 1: Adiabaticity criteria,” IEEE Proc.-J., vol. 138, no. 5, pp. 343–354, Oct. 1991. [32] W. R. Boyd, Nonlinear Optics, 3rd ed. Amsterdam, The Netherlands: Elsevier, 2008. [33] C. Wei, M. Zhou, Z. Hui-Juan, and L. Hong, “Efects of input spectra on the threshold of modulation instability in a single-mode fiber,” Can. Phys. Lett., vol. 29, no. 4, p. 1–044213, 2012.

Nagarajan Nallusamy received the B.Sc. degree from the Bishop Heber College, Tiruchirappalli, in 2009, the M.Sc. degree from St. Josephs College in 2012, and the M.Phil. degree from the Loyola College, Chennai, in 2014, all in physics. He is currently pursuing the Ph.D. degree in nonlinear fiber optics with SASTRA University, Thanjavur. His research interest is in the field of supercontinuum generation and modulation instability for nonlinear sensor applications.

R. Vasantha Jayakantha Raja received the M.Sc. degree in physics from the University of Madras, the M.Phil. degree in physics from Pondicherry University, and the Ph.D. degree from Pondicherry University. He is currently an Assistant Professor at SASTRA University, Thanjavur, where he is actively working in the research field of nonlinear fiber optics and ultrafast optics. He has mainly focused his research on numerical modeling of nonlinear pulse propagation through solid and liquid core photonic crystal fiber for various nonlinear applications, including supercontinuum generation, soliton, switching dynamics, modulational instability. Recently, he focused his topic on fiber laser and nonlinear optical fiber sensor based on four wave mixing process.

G. Joshva Raj, photograph and biography not available at the time of publication.