Temperature cross-sensitivity characteristics of singlemode–multimode–singlemode fiber structure Rongxiang Zhang, Tiegen Liu, Qun Han, Yaofei Chen, Lin Li, and X. Steve Yao Citation: Review of Scientific Instruments 86, 013108 (2015); doi: 10.1063/1.4905723 View online: http://dx.doi.org/10.1063/1.4905723 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/86/1?ver=pdfcov Published by the AIP Publishing Articles you may be interested in CO2 sensing at room temperature using carbon nanotubes coated core fiber Bragg grating Rev. Sci. Instrum. 84, 065002 (2013); 10.1063/1.4810016 Pressure effects on the temperature sensitivity of fiber Bragg gratings AIP Conf. Proc. 1511, 1570 (2013); 10.1063/1.4789229 Lithium niobate nanoparticulate clad on the core of single mode optical fiber for temperature and magnetic field sensing Appl. Phys. Lett. 101, 043102 (2012); 10.1063/1.4738884 A thin metal sheath lifts the EH to HE degeneracy in the cladding mode refractometric sensitivity of optical fiber sensors Appl. Phys. Lett. 99, 041118 (2011); 10.1063/1.3615712 Simulation of interferometric sensors for pressure and temperature measurements Rev. Sci. Instrum. 71, 1608 (2000); 10.1063/1.1150505
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
REVIEW OF SCIENTIFIC INSTRUMENTS 86, 013108 (2015)
Temperature cross-sensitivity characteristics of singlemode–multimode–singlemode fiber structure Rongxiang Zhang,1,2 Tiegen Liu,1 Qun Han,1,a) Yaofei Chen,1 Lin Li,1 and X. Steve Yao1
1
College of Precision Instrument and Opto-Electronics Engineering, Tianjin University, Tianjin 300072, China and Key Laboratory of Opto-Electronics Information Technical, Tianjin University, Ministry of Education, Tianjin 300072, China 2 College of Physics Science and Technology, Hebei University, Baoding 071002, China
(Received 2 December 2014; accepted 29 December 2014; published online 16 January 2015) The temperature cross-sensitivity characteristics of a singlemode–multimode–singlemode (SMS) fiber structure packaged by a shell are studied both theoretically and experimentally. By theoretical investigation, we found that the temperature sensitivity of a SMS structure is mainly determined by the thermo-optic effect (TOE) of the cladding of the multimode fiber (MMF). Meanwhile, the TOE of the MMF core, thermal expansion effects (TEEs) of the MMF core, and the packaging material also influence the ultimate sensitivity, and the magnitude of their effects depends on the refractive index of the MMF cladding. Among them, the TEE of the packaging material, inducing an axial strain, is considered to be the second main factor. A temperature sensor based on a packaged SMS structure is designed and investigated to experimentally verify the theoretical findings. The experimentally measured temperature sensitivity of the sensor is −453.4 pm/◦C, which agrees well with the theoretical prediction. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4905723]
I. INTRODUCTION
Singlemode–multimode–singlemode (SMS) fiber structure, as one of the most promising structures utilizing the multimode interference, has attracted great research interests and been successfully used to measure many parameters, including refractive index (RI),1,2 magnetic field or current,3–5 displacement,6 stain,7 curvature,8 and temperature.9–12 Accurate and reliable measurement of temperature is of significance for many chemical, physical, biological, and environmental applications. Compared with the fiber optic temperature sensors using fiber Bragg gratings,13,14 long-period fiber gratings,15 Fabry–Perot interferometer,16 Mach–Zehnder interferometer,17 microstructured optical fiber taper,18 and photonic crystal fiber,19 the SMS fiber optic structure based temperature transducers, taking advantage of the thermo-optic effect (TOE) and thermal expansion effect (TEE) of the employed materials, has superiorities of simple structure, easy fabrication, and big sensitivity with proper materials.11 Recently, there have been some reports about fiber optic temperature sensor based on the SMS structure.9–12 Li et al.9 proposed a reflective SMS-based temperature sensor. However, the measured sensitivity was only 15 pm/◦C. It is well known that for a SMS structure, when temperature changes, two effects will influence the multimode interference within the multimode fiber (MMF) and hence the temperature sensitivity.9 One is the TOE of the MMF, which influences the RIs of the MMF core and cladding and thus effects on the multimode interference, the other is the TEEs of the employed materials, which affect the multimode interference through the physical expansion or contraction of the MMF dimensions.
a)Author to whom correspondence should be addressed. Electronic mail:
[email protected].
Accordingly, the sensitivity of a SMS-based temperature sensor is determined by the thermo-optic coefficient (TOC) and the thermal expansion coefficient (TEC) of the materials constructing the SMS structure. In the above mentioned study, the core and cladding of the MMF, which were both mainly made of silica, had nearly the same thermal properties. Since both the TOC and TEC of silica are low, the temperature sensitivity of the sensor is limited. To improve the sensitivity, several attempts have been made. Wu et al.10 proposed a bent SMS fiber structure and the temperature sensitivity was improved several times to 44.26 pm/◦C or 31.97 pm/◦C. Silva et al.11 used RI standard liquid with a larger RI and TOC as the cladding of the MMF, and a temperature sensitivity of −360 pm/◦C in the temperature range from 25 to 80 ◦C was achieved. Sun et al.12 fabricated a temperature sensor by encapsulating a SMS device into a deionized water-filled aluminum alloy shell, the attained sensitivity as temperature varying from 25 to 80 ◦C was 358 pm/◦C. It can be seen from the aforementioned works that the temperature sensitivity of a SMS structure can be dramatically improved by replacing the cladding of a conventional MMF with materials having a high RI and TOC. In addition, a shell, used to package the sensor, can also influence the temperature sensing characteristics of the SMS structure due to the axial strain introduced by the different TECs of the MMF core and packaging shell. References 11 and 12 mainly investigated the temperature features of a SMS structure with a substituted but fixed MMF cladding or a fixed packaging material. However, according to the principle of the multimode interference, the output transmission spectrum of a SMS structure is simultaneously associated with the RIs of the MMF core and cladding, diameter, and length of the MMF core.1,4 So when the cladding of the MMF is substituted by different materials or the SMS structure is
0034-6748/2015/86(1)/013108/7/$30.00 86, 013108-1 © 2015 AIP Publishing LLC This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-2
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
packaged by different shells, the thermal responses of the RIs and dimensions of the elements in the SMS structure will be different, resulting to the different temperature sensitivities of the constructed sensor. In previous studies, to the best of our knowledge, there is no work about the effects of the thermal properties of the MMF core and packaging shell on the temperature sensitivity of a SMS structure with different MMF claddings. In other words, the temperature cross-sensitivity characteristics, resulting from intrinsic MMF core, MMF cladding, and packaging shell, have not been investigated. The work of this paper is to carry out a comprehensive study about the temperature sensing properties of a SMS structure by exploring the MMF claddings with different RIs and TOCs, and packaging shells with different TECs, with the aim to ascertain the respective impact of each element on the temperature sensitivity of a SMS structure. The study will be of importance for the optimal design and development of SMS-based sensors. According to the obtained results, the proper materials of the MMF cladding and packaging shell can be selected to improve the sensing performance for temperature measurement or eliminate the temperature crosssensitivity in non-temperature sensors. Based on our analyses, a temperature sensor composed with glycerol–water mixture as the MMF cladding and packaged by a glass capillary is proposed and experimentally demonstrated in this paper. The experimental results confirm the theoretical analyses and with the temperature varying from 25 to 80 ◦C, a high sensitivity of −453.4 pm/◦C is obtained. II. TEMPERATURE SENSING PRINCIPLE OF SMS FIBER STRUCTURE
The schematic of a SMS fiber structure is shown in Fig. 1. Two identical single mode fibers (SMFs) are axially spliced to the two ends of a MMF, respectively. Since the diameter of the MMF core is larger than that of the SMF, many high-order eigenmodes {LPnm} are excited, while the beam comes to the MMF section. Suppose that the axes of the SMF and the MMF are perfectly aligned, only the symmetric modes {LP0m} can be excited and the input field in the MMF can be represented as1 E(r,0) =
M
cm Fm (r),
(1)
m=1
where M is the total number of excited modes, Fm (r) is the field profile of the LP0m mode, r is the radial coordinate in the cross section of the fiber, cm is the excitation coefficient of the LP0m mode, which can be expressed by the overlap integral
between E(r,0) and Fm (r) as follows:1 ∞ E(r,0) × Fm (r) ×r dr cm =
0
.
∞
(2)
Fm (r) × Fm (r) ×r dr 0
The field profile of the LP0m mode at a propagation distance z can be then calculated by1 E(r,z) =
M
cm Fm (r)exp(i β m z),
(3)
m=1
where β m is the propagating constant of the LP0m mode. When light is coupled back to the output SMF, the transmittance of SMS structure can be thus calculated by2 T(λ) =
M
2 cm × cn2 × cos[( β m − β n )L],
(4)
m, n=1
where L is the length of the MMF. When the RI of the MMF core or cladding changes, the eigenmodes Fm (r) excited in the MMF vary accordingly, leading to the changes of cm and β m of each mode. The length and diameter of the MMF core also influence the distribution of eigenmodes Fm (r) and the interference between the modes. Therefore, the final output to the SMF as shown in Eq. (4) is determined by the RIs of the MMF core and cladding, and the physical dimensions of the MMF core. Since all these parameters are sensitive to temperature, the output spectrum changes with temperature. This is the underlying principle of a SMS-based temperature sensor. In order to make a comprehensive study of the temperature characteristics of a SMS structure, we consider a model as shown in Fig. 1. The cladding of the MMF is substituted by certain thermal sensitive material (such as some kind of liquid) that has known thermal-optical features and is easy to be prepared. The section of the MMF is packaged by a shell, and both ends are sealed with UV glue. For such a SMS structure, when temperature changes ∆T, the RI changes of the MMF core and cladding induced by the TOE of the MMF can be expressed as ∆nco_1 = ξco × ∆T,
(5)
∆ncl = ξcl × ∆T,
(6)
where ξco and ξcl are the TOCs of the MMF core and cladding, respectively. The changes of the diameter and length of the MMF core induced by thermal expansion then can be expressed as ∆d 1 = d 0 × αco × ∆T,
(7)
∆l 1 = l 0 × αco × ∆T,
(8)
where d 0 and l 0 are the diameter and length of the MMF core at 25 ◦C, respectively, αco is the TEC of the MMF core. In addition, if the TEC of the packaging material (αp) is different from that of the MMF core (αco), an axial strain will be caused and can be given by ε = (αp − αco) × ∆T. (9) FIG. 1. Schematic of SMS fiber optic structure. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-3
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
The changes of the RI, diameter, and length of the MMF core, induced by the axial strain, can be, respectively, described by7,20 3 nco_0
[p12 − ν(p11 + p12)] × αp − αco × ∆T, 2 ∆d 2 = −ν × d × αp − αco × ∆T, ∆l 2 = l × αp − αco × ∆T, ∆nco_2 = −
(10) (11) (12)
where nco_0 is the RI of the MMF core at 25 ◦C, p11 and p12 are elastic-optic coefficients, ν is the Poisson ratio. For silica fibers, p11 = 0.12, p12 = 0.27, and ν = 0.17.20 Therefore, when temperature changes, the RIs of the MMF core and cladding, diameter, and length of the MMF core can be given, respectively, by nco = nco_0 + ∆nco_1 + ∆nco_2,
(13)
ncl = ncl_0 + ∆ncl,
(14)
d = d 0 + ∆d 1 + ∆d 2,
(15)
l = l 0 + ∆l 1 + ∆l 2,
(16) ◦
where ncl_0 is the RI of the MMF cladding at 25 C. For ease of description, we will refer to the TOEs of the MMF core and cladding as the TOEco and the TOEcl, the TOCs of the MMF core and cladding as the TOCco and the TOCcl, the TEEs of the MMF core and packaging material as the TEEco and the TEEp, the TECs of the MMF core and packaging material as the TECco and the TECp, and the RIs of the MMF core and cladding as the RIco and the RIcl, respectively. According to the above discussion, when the temperature changes, due to the TOEco, TOEcl, TEEco, and the axial strain caused by the different values of the TECco and TECp, the RIco, RIcl, diameter, and length of the MMF core change correspondingly, leading to the change of the transmission spectrum of the SMS structure. Hence, the temperature sensitivity of a SMS fiber structure depends on the combined effects of the TOE, TEE, and the induced axial strain of SMS structure. In order to make the temperature cross-sensitivity characteristics clear and provide the basis for choosing appropriate cladding and shell materials for improving the sensor’s performance or meeting various practical requirements, the impact of each factor on the temperature sensitivity of a SMS structure has been analyzed in the rest of the paper.
core is assumed to be made of pure silica with a diameter of 61.5 µm and its RI is calculated by the Sellmeier equation.20 Since the sensitivity demodulated from the wavelength shift of a SMS structure is independent of the MMF length and proportional to the dip wavelength,2 the lengths of the MMF core are selected to obtain an identical dip wavelength at 25 ◦C for various SMS structures with different MMF claddings. Then, based on Eqs. (1)–(4), (6), and (14), the numerical simulation can be carried out by utilizing the mode propagation analysis (MPA) method.1 Considering that the RI sensitivity is changing with the RI range,1 the temperature features of SMS structures with different RIcls and TOCcls are simulated. Fig. 2 shows the relationship between a dip wavelength (1569.16 nm) in the transmission spectrum and the temperature for various SMS structures. The inset shows the corresponding transmission spectra at 25 ◦C, where an interference dip with better visibility appears at 1569.16 nm for both spectra. It can be seen from Fig. 2 that the dip wavelength decreases linearly with the increase of temperature in the range from 25 to 80 ◦C for the SMS structures with different MMF claddings. The respective sensitivity can be obtained from the linear fit. As shown in Fig. 2, for the same TOCcl (ξcl = −1.5 × 10−4/◦C), when the RIcl (ncl) is 1.33 and 1.42, the temperature sensitivity is −31.8 pm/◦C and −278 pm/◦C, respectively. So the temperature sensitivity of the SMS structure significantly increases with the increase of the RIcl. Further, when the RIcl is kept at 1.42, but the TOCcl is changed to −2 × 10−4/◦C, the temperature sensitivity will be −343.6 pm/◦C. This indicates that a higher TOCcl will also be helpful to enhance the temperature sensitivity. So if a higher temperature sensitivity is desirable, we need to choose a material with a higher RI and/or higher TOC to serve as the cladding of the MMF section in a SMS structure. B. Temperature characteristics due to TOE of the MMF core in a SMS structure
In the study of the contribution of the TOEco to the temperature sensitivity of a SMS structure, we assume that the
III. ANALYSIS OF THE TEMPERATURE CHARACTERISTICS OF SMS FIBER OPTIC STRUCTURE A. Temperature characteristics due to TOE of the MMF cladding in a SMS structure
Because of the TOEcl, the RIcl will change with temperature. To numerically analyze the thermal sensing characteristics of a SMS fiber structure induced solely by the TOEcl, we first suppose that the thermal responses of the MMF core and packing shell are not considered, that is, nco, d, and l are unchanged with the changing temperature. In the simulations, the SMF is assumed to be the Corning SMF-28 with a core diameter of 8.2 µm and numerical aperture of 0.14. The MMF
FIG. 2. Dip wavelength as a function of the temperature for SMS structures fabricated with MMF claddings having different RIs and TOCs, the inset shows the corresponding transmission spectra of various SMS structures at 25 ◦C. Only the TOEcl is considered in the simulations. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-4
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
FIG. 3. Dip wavelength versus temperature for SMS structures fabricated with various MMF claddings having different RIs. Only the TOEco is considered in the simulations.
TOEcl, TEEco, and TEEp are not taken into account, i.e., the ncl, d, and l are not varying with temperature, and ∆nco_2 in Eq. (13) is zero. The other parameters are identical to those in Sec. III A. Based on Eqs. (1)–(5) and (13), the dip wavelengths in the transmission spectra as functions of temperature for the SMS structures fabricated with different MMF claddings are depicted in Fig. 3. For the standard silica fibers, ξco = 1.06 × 10−5/◦C.20 From Fig. 3, we can see that the resonant dip wavelengths change linearly with the increase of temperature, but the sensitivities are different. This indicates that the thermal sensing characteristic induced by the TOEco is influenced by the RIcl. The sensitivities of the SMS structures can be obtained by the linear fits. As shown in Fig. 3, when the RIcl is 1 (such as atmosphere), the dip wavelength shifts toward longer wavelength as the temperature increases, and the sensitivity is 11.1 pm/◦C. This sensitivity is close to the previously reported experimental results (15 pm/◦C and 12.9 pm/◦C),9,21 where the core and cladding of the MMF are mainly made of silica, having nearly the same RI and TOC, and hence can be equivalent to the MMF core in our simulations. Fig. 3 shows that with the increasing of the RIcl, the positive temperature sensitivity decreases, while when the RIcl reaches a certain value (between 1.40 and 1.41), the dip wavelength is almost insensitive to the temperature and then it shifts to shorter wavelengths with the rise of temperature. The temperature sensitivity reaches −13 pm/◦C when the RIcl is 1.42. Thus, the effect of the TOEco on the final temperature sensitivity of a SMS sensor depends on the RIcl.
FIG. 4. Dip wavelength versus temperature for SMS structures fabricated with various MMF claddings having different RIs. Only the TEEco is considered in the simulations.
different claddings are depicted in Fig. 4. For the standard silica fibers, αco = 5 × 10−7/◦C.20 As shown in Fig. 4, the temperature responses of the dip wavelengths of the SMS structures with different RIcls are nearly the same. The dip wavelength increases slowly with the increasing of temperature in the range from 25 to 80 ◦C. The small temperature sensitivity of 0.8 pm/◦C is caused by the low TEC of silicon. So we can conclude that the TEEco has less influence on the thermal sensitivity of a SMS structure than the TOEco and TOEcl. Besides the TEEco, the TEEp also influences the sensitivity of a SMS structure by introducing an axial strain, as described in Sec. II. Now, we analyze the effects of both of the TEEco and TEEp. The ∆nco_2 in Eq. (13), ∆d 2 in Eq. (15), and ∆l 2 in Eq. (16) are no longer null. With Eqs. (1)–(4),(7)–(13), (15), and (16), the dip wavelengths changing with temperature for SMS structures with different packing materials are simulated. The RIcl is assumed to be in the range from 1 to 1.42. The simulation results are shown in Fig. 5. Similar to Fig. 4, we can see from Fig. 5 that the dip wavelength does not change with the RIcl if the TECp is fixed. How-
C. Temperature characteristics due to TEEs in a SMS structure
In the evaluation of the contribution of the TEEco to the temperature sensitivity of a SMS structure, it is assumed that nco and ncl are independent of temperature, and ∆d 2 in Eq. (15) and ∆l 2 in Eq. (16) are zero. By using Eqs. (1)–(4), (7), (8), (15), and (16), the simulated relationships between the dip wavelength and temperature for the SMS structures with
FIG. 5. Dip wavelength versus temperature for SMS structures fabricated with various MMF claddings having different RIs and packaging shells with different TECs. Only the TEEco and TEEp are considered in the simulations. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-5
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
ever, the TECp does have an impact on the temperature sensitivity. As shown in Fig. 5, when the TECps are 5×10−6/◦C and 5 × 10−5/◦C, the sensitivities are −9 pm/◦C and −107 pm/◦C, respectively. This means that the higher the TECp is, the larger the negative sensitivity will be. So for SMS-based temperature sensors with positive or negative sensitivities, proper packaging materials with a smaller or bigger TEC should be selected to further optimize the sensing performance. D. Temperature characteristics based on TOE of MMF core and TEEs in SMS fiber optic structure
Finally, by comparing the respective temperature sensitivity induced by the TOEcl, TOEco, TEEco, and both of the TEEco and TEEp, we can conclude that the TOEcl plays a major role in determining the temperature sensitivity of a SMS structure. The thermal properties of the MMF core and the packaging materials also have influences on the sensitivity. In order to optimize the temperature performance of a SMS structure, we consider the combined effect of the TOEco, TEEco, and TEEp on the sensitivity of the sensors with several MMF claddings having different RIs. Simulations are based on Eqs. (1)–(5) and (7)–(16). The simulation results are shown in Fig. 6. Fig. 6 depicts the resonance dip wavelength versus temperature for the packaged SMS structures with various MMF claddings and packaging materials. It can be seen that the results depicted in Fig. 6(a) are similar to those in Fig. 3, that displays the effect of the TOEco, i.e., the change trend and extent of the dip wavelength with temperature are varying with the changing of the RIcl. This is attributed to the small and unchanged temperature sensitivity induced by the smaller TECco and TECp (see Fig. 5). It can be seen from Fig. 6(a) that when the RIcl is 1 and 1.42, the sensitivity is 2.1 pm/◦C and −21.9 pm/◦C, respectively. Whereas when the TECp is larger, Fig. 6(b) indicates that an increase in the temperature from 25 to 80 ◦C causes monotonic decreases in the dip wavelengths without exception, and the negative sensitivity increases with the increasing of the RIcl. A sensitivity of −118.7 pm/◦C has been achieved when the RIcl is 1.42. This result of the all negative sensitivities is due to the larger axial strain induced by the packaging shell (also see Fig. 5). So, the TEEp is the second important factor that influences the temperature sensitivity of a SMS structure. To maximize the temperature sensitivity of a packaged SMS structure, the proper packaging material should be selected according to the TOCcl. For example, if the MMF cladding has positive TOC, a packaging material with a smaller TEC is recommended, otherwise a packaging material with a larger TEC is better. Additionally, from Fig. 6(a), it should be noted that the temperature responses resulting from the TOEco, TEEco, and TEEp may counteract to each other under certain condition, which is useful for some temperature-insensitivepreferable sensors based on a SMS structure. IV. SMS FIBER STRUCTURE BASED TEMPERATURE SENSOR WITH ENHANCED SENSITIVITY A. Experiments
In order to experimentally demonstrate the improved temperature sensing property of a packaged SMS fiber optic
FIG. 6. Dip wavelength versus temperature for SMS structures fabricated with various MMF claddings having different RIs and packaging shell with different TECs. Only the TOEco, TEEco, and TEEp are considered in the simulations.
structure by using proper materials, as well as verify the theoretical findings based on numerical simulations, we designed, fabricated, and experimentally studied a SMS-based temperature sensor. The sensor head, as shown in Fig. 1, is fabricated by first splicing two ends of the MMF to two SMFs, which is accomplished by a commercial fusion splicer in manual operation mode. The SMF and MMF employed in our experiments are the standard SMF-28 (Corning, Inc.) and commercialized no-core fiber (NCF, Prime Optical Fiber Co.), respectively. The NCF is made of pure silica. The diameter and length of the NCF are 61.5 µm and 8.2 cm, respectively. Then, the NCF section is fed into a glass capillary with an approximate TEC of 5 × 10−6/◦C,22 an inner diameter of 0.5 mm, and a length of 10 cm. Finally, the capillary is filled with glycerol–water mixture (3:1 volume mixture), and both ends of the capillary are sealed with UV glue. The glycerol–water mixture serves as the cladding of the NCF. Glycerol–water mixture (3:1 volume mixture) is chosen because it is a conventional and easily accessible liquid with a larger RI around 1.43 and a TOC of −1.827 × 10−4/◦C.23 The packaging capillary not only makes the senor convenient to be
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-6
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
FIG. 7. Experimental setup for investigating the temperature properties of the SMS-based sensor.
used but also enhances the temperature sensitivity of sensor due to the larger TEE of the glass. The experimental setup is shown in Fig. 7, light from a broadband amplified spontaneous emission (ASE, LightComm Technology, Ltd.) source passes through the SMSbased sensor, and the transmission spectrum is recorded by an optical spectrum analyzer (OSA, Yokogawa Electric Co.). In characterization of the temperature response of the sensor, the sensing head is placed into a temperature-controlled oven. A thermometer with an accuracy of ±0.5 ◦C is used to measure the temperature in real time. The sensor head of the thermometer is commercial platinum resistance which is placed near the SMS structure in the oven through a small hole. The sensor is first temperature-cycled between 25 ◦C and 80 ◦C for several times to relax strain. Then, the transmission spectra are recorded with the OSA with a temperature step of 5 ◦C.
It can be seen from Fig. 8 that interference patterns are clearly obtained in the experimental transmission spectrum. We chose the dip at 1592.78 nm as the spectral indicator for the temperature sensing, because it has good visibility and large dynamic range in wavelength domain in our experiment. The measured spectral responses under different temperatures are shown in Fig. 9(a), as can be seen; with the temperature rising, the dip wavelength of the resonant spectrum shifts toward shorter wavelength due to the decreasing RI of the glycerol–water mixture. In order to obtain the temperature sensitivity of the sensor, based on Fig. 9(a), the dependence of the dip wavelength on temperature is explicitly plotted in Fig. 9(b). It is shown that the dip wavelength decreases linearly with the increase of temperature. The experimental results can be fitted with a good linearity, as shown in Fig. 9(b). The slope of the linear fit is −453.4 pm/◦C. In the corresponding range of temperature, this sensitivity is not only much higher than that of SMS-based sensor using conventional MMF with silica as cladding (12–45 pm/◦C)9,10,21 but also higher than that of the SMS-based sensor with commercial RI liquid (a nominal RI value of 1.44 (20 ◦C)) as MMF cladding but without the packaging shell (−360 pm/◦C)11 and also higher
B. Results and discussions
Fig. 8 shows the experimental and simulated transmission spectra of the sensor at 25 ◦C. The simulation is carried out by using the MPA method based on Eqs. (1)–(4), and the parameters are taken as the employed materials described in Sec. IV A. From Fig. 8, one can see that the theoretical spectrum has a reasonable agreement with the measured result. There is a dip wavelength (1592.78 nm) in both of the simulated and experimental spectra. The discrepancy between the two transmission spectra could be mainly due to the imperfect arrangement of the axes of the SMF and NCF.
FIG. 9. (a) Transmission spectra under different temperatures and (b) variation of dip wavelength of the spectrum with the temperature for the SMSFIG. 8. Transmission spectra of the SMS-based sensor at 25 ◦C. based sensor. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-7
Zhang et al.
than that of the SMS device employing deionized water as the MMF cladding and aluminum alloy shell as packing material (358 pm/◦C).12 So here, according to the numerical analyses, the glycerol–water mixture with a high RI and a negative TOC, and the packaging glass capillary having a higher TEC compared with the NCF are both contributors to the enhanced temperature sensitivity. Simulation results with Eqs. (1)–(16) are also shown in Fig. 9(b). In the simulations, all of the effects of the TOEco, TOEcl, TEEco, and TEEp are taken into account. The parameters are still taken as the experimental materials described in Sec. IV A. The sensitivity from the simulation results is −463.6 pm/◦C, which is close to the experimental value. This confirms the validity of our theoretical analyses.
V. CONCLUSIONS
In summary, we carry out a comprehensive analysis of the temperature cross-sensitivity characteristics of a SMS structure. Based on the MPA method, it is found that the TOEcl and TEEp are the first and second important contributors to the temperature sensitivity of a SMS structure, respectively. Further, the combined effects of TOEco, TEEco, and TEEp indicate that, for the MMF cladding with a positive/negative TOC, the packaging material with a smaller/larger TEC is more suitable for improving the thermal sensitivity of the SMS structure. In addition, the RIcl also influences the effects of various factors. Based on the results of numerical analysis, we fabricate a SMS-based temperature sensor with a NCF cladding of glycerol–water mixture and a packing shell made of glass. The measured temperature sensitivity is −453.4 pm/◦C. The experimental data match our simulation very well. Although the sensor, featuring a simple structure and fabrication process, performs very well, its sensing performance can be further improved by using appropriate materials, such as a NCF cladding with a larger RI and/or larger negative TOC, and packaging shell with a bigger TEC. Additionally, the conclusions derived in this paper can also provide a reference for eliminating cross-sensitivity to temperature for the non-temperature sensors based on SMS structure.
Rev. Sci. Instrum. 86, 013108 (2015)
ACKNOWLEDGMENTS
This work was supported by the Natural Science Foundation of Tianjin (Grant No. 13JCYBJC16100), the National Natural Science Foundation of China (Grant No. 61107035), the National Key Scientific Instrument and Equipment Development Project of China (Grant No. 2013YQ03091502), and the National Basic Research Program of China (Grant No. 2014CB340104). 1Q.
Wu, Y. Semenova, P. Wang, and G. Farrell, Opt. Exp. 19, 7937 (2011). Chen, Q. Han, T. Liu, and H. Xiao, IEEE Photon. J. 6, 6801807 (2014). 3H. Wang, S. Pu, N. Wang, S. Dong, and J. Huang, Opt. Lett. 38, 3765 (2013). 4Y. Chen, Q. Han, T. Liu, X. Lan, and H. Xiao, Opt. Lett. 38, 3999 (2013). 5L. Li, Q. Han, T. Liu, Y. Chen, and R. Zhang, Rev. Sci. Instrum. 85, 083107 (2014). 6Q. Wu, Y. Semenova, P. Wang, A. M. Hatta, and G. Farrell, Meas. Sci. Technol. 22, 025203 (2011). 7E. Li, IEEE Photon. Technol. Lett. 19, 1266 (2007). 8Y. Gong, T. Zhao, Y.-J. Rao, and Y. Wu, IEEE Photon. Technol. Lett. 23, 679 (2011). 9E. Li, X. Wang, and C. Zhang, Appl. Phys. Lett. 89, 091119 (2006). 10Q. Wu, Y. Semenova, A. M. Hatta, P. Wang, and G. Farrell, Electron. Lett. 46, 1129 (2010). 11S. Silva, E. G. P. Pachon, M. A. R. Franco, J. G. Hayashi, F. X. Malcata, O. Frazão, P. Jorge, and C. M. B. Cordeiro, Appl. Opt. 51, 3236 (2012). 12H. Sun, M. Hu, Q. Rong, Y. Du, H. Yang, and X. Qiao, Opt. Commun. 323, 28 (2014). 13Y. Liu, Z. Zhou, E. Zhang, J. Zhang, Y. Tan, and M. Liu, Rev. Sci. Instrum. 85, 064905 (2014). 14S. K. Mondal, U. Tiwari, G. C. Poddar, V. Mishra, N. Singh, S. C. Jain, S. N. Sarkar, K. D. Chattoypadhya, and P. Kapur, Rev. Sci. Instrum. 80, 103106 (2009). 15A. P. Zhang, L.-Y. Shao, J.-F. Ding, and S. He, IEEE Photon. Technol. Lett. 17, 2397 (2005). 16J. Wang, B. Dong, E. Lally, J. Gong, M. Han, and A. Wang, Opt. Lett. 35, 619 (2010). 17Z. Cao, Z. Zhang, X. Ji, T. Shui, R. Wang, C. Yin, S. Zhen, L. Lu, and B. Yu, Opt. Fiber Technol. 20, 24 (2014). 18D. Monzón-Hernández, V. P. Minkovich, and J. Villatoro, IEEE Photon. Technol. Lett. 18, 511 (2006). 19W. Qian, C.-L. Zhao, S. He, X. Dong, S. Zhang, Z. Zhang, S. Jin, J. Guo, and H. Wei, Opt. Lett. 36, 1548 (2011). 20S. M. Tripathi, A. Kumar, R. K. Varshney, Y. B. P. Kumar, E. Marin, and J.-P. Meunier, J. Lightwave Technol. 27, 2348 (2009). 21J. G. Aguilar-Soto, J. E. Antonio-Lopez, J. J. Sanchez-Mondragon, and D. A. May-Arrioja, J. Phys.: Conf. Ser. 274, 012011 (2011). 22J. Arndt and F. Häberle, Contrib. Mineral. Petrol. 39, 175 (1973). 23J. Popp, M. Lankers, K. Schaschek, W. Kiefer, and J. T. Hodges, Appl. Opt. 34, 2380 (1995). 2Y.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
REVIEW OF SCIENTIFIC INSTRUMENTS 86, 013108 (2015)
Temperature cross-sensitivity characteristics of singlemode–multimode–singlemode fiber structure Rongxiang Zhang,1,2 Tiegen Liu,1 Qun Han,1,a) Yaofei Chen,1 Lin Li,1 and X. Steve Yao1
1
College of Precision Instrument and Opto-Electronics Engineering, Tianjin University, Tianjin 300072, China and Key Laboratory of Opto-Electronics Information Technical, Tianjin University, Ministry of Education, Tianjin 300072, China 2 College of Physics Science and Technology, Hebei University, Baoding 071002, China
(Received 2 December 2014; accepted 29 December 2014; published online 16 January 2015) The temperature cross-sensitivity characteristics of a singlemode–multimode–singlemode (SMS) fiber structure packaged by a shell are studied both theoretically and experimentally. By theoretical investigation, we found that the temperature sensitivity of a SMS structure is mainly determined by the thermo-optic effect (TOE) of the cladding of the multimode fiber (MMF). Meanwhile, the TOE of the MMF core, thermal expansion effects (TEEs) of the MMF core, and the packaging material also influence the ultimate sensitivity, and the magnitude of their effects depends on the refractive index of the MMF cladding. Among them, the TEE of the packaging material, inducing an axial strain, is considered to be the second main factor. A temperature sensor based on a packaged SMS structure is designed and investigated to experimentally verify the theoretical findings. The experimentally measured temperature sensitivity of the sensor is −453.4 pm/◦C, which agrees well with the theoretical prediction. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4905723]
I. INTRODUCTION
Singlemode–multimode–singlemode (SMS) fiber structure, as one of the most promising structures utilizing the multimode interference, has attracted great research interests and been successfully used to measure many parameters, including refractive index (RI),1,2 magnetic field or current,3–5 displacement,6 stain,7 curvature,8 and temperature.9–12 Accurate and reliable measurement of temperature is of significance for many chemical, physical, biological, and environmental applications. Compared with the fiber optic temperature sensors using fiber Bragg gratings,13,14 long-period fiber gratings,15 Fabry–Perot interferometer,16 Mach–Zehnder interferometer,17 microstructured optical fiber taper,18 and photonic crystal fiber,19 the SMS fiber optic structure based temperature transducers, taking advantage of the thermo-optic effect (TOE) and thermal expansion effect (TEE) of the employed materials, has superiorities of simple structure, easy fabrication, and big sensitivity with proper materials.11 Recently, there have been some reports about fiber optic temperature sensor based on the SMS structure.9–12 Li et al.9 proposed a reflective SMS-based temperature sensor. However, the measured sensitivity was only 15 pm/◦C. It is well known that for a SMS structure, when temperature changes, two effects will influence the multimode interference within the multimode fiber (MMF) and hence the temperature sensitivity.9 One is the TOE of the MMF, which influences the RIs of the MMF core and cladding and thus effects on the multimode interference, the other is the TEEs of the employed materials, which affect the multimode interference through the physical expansion or contraction of the MMF dimensions.
a)Author to whom correspondence should be addressed. Electronic mail:
[email protected].
Accordingly, the sensitivity of a SMS-based temperature sensor is determined by the thermo-optic coefficient (TOC) and the thermal expansion coefficient (TEC) of the materials constructing the SMS structure. In the above mentioned study, the core and cladding of the MMF, which were both mainly made of silica, had nearly the same thermal properties. Since both the TOC and TEC of silica are low, the temperature sensitivity of the sensor is limited. To improve the sensitivity, several attempts have been made. Wu et al.10 proposed a bent SMS fiber structure and the temperature sensitivity was improved several times to 44.26 pm/◦C or 31.97 pm/◦C. Silva et al.11 used RI standard liquid with a larger RI and TOC as the cladding of the MMF, and a temperature sensitivity of −360 pm/◦C in the temperature range from 25 to 80 ◦C was achieved. Sun et al.12 fabricated a temperature sensor by encapsulating a SMS device into a deionized water-filled aluminum alloy shell, the attained sensitivity as temperature varying from 25 to 80 ◦C was 358 pm/◦C. It can be seen from the aforementioned works that the temperature sensitivity of a SMS structure can be dramatically improved by replacing the cladding of a conventional MMF with materials having a high RI and TOC. In addition, a shell, used to package the sensor, can also influence the temperature sensing characteristics of the SMS structure due to the axial strain introduced by the different TECs of the MMF core and packaging shell. References 11 and 12 mainly investigated the temperature features of a SMS structure with a substituted but fixed MMF cladding or a fixed packaging material. However, according to the principle of the multimode interference, the output transmission spectrum of a SMS structure is simultaneously associated with the RIs of the MMF core and cladding, diameter, and length of the MMF core.1,4 So when the cladding of the MMF is substituted by different materials or the SMS structure is
0034-6748/2015/86(1)/013108/7/$30.00 86, 013108-1 © 2015 AIP Publishing LLC This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-2
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
packaged by different shells, the thermal responses of the RIs and dimensions of the elements in the SMS structure will be different, resulting to the different temperature sensitivities of the constructed sensor. In previous studies, to the best of our knowledge, there is no work about the effects of the thermal properties of the MMF core and packaging shell on the temperature sensitivity of a SMS structure with different MMF claddings. In other words, the temperature cross-sensitivity characteristics, resulting from intrinsic MMF core, MMF cladding, and packaging shell, have not been investigated. The work of this paper is to carry out a comprehensive study about the temperature sensing properties of a SMS structure by exploring the MMF claddings with different RIs and TOCs, and packaging shells with different TECs, with the aim to ascertain the respective impact of each element on the temperature sensitivity of a SMS structure. The study will be of importance for the optimal design and development of SMS-based sensors. According to the obtained results, the proper materials of the MMF cladding and packaging shell can be selected to improve the sensing performance for temperature measurement or eliminate the temperature crosssensitivity in non-temperature sensors. Based on our analyses, a temperature sensor composed with glycerol–water mixture as the MMF cladding and packaged by a glass capillary is proposed and experimentally demonstrated in this paper. The experimental results confirm the theoretical analyses and with the temperature varying from 25 to 80 ◦C, a high sensitivity of −453.4 pm/◦C is obtained. II. TEMPERATURE SENSING PRINCIPLE OF SMS FIBER STRUCTURE
The schematic of a SMS fiber structure is shown in Fig. 1. Two identical single mode fibers (SMFs) are axially spliced to the two ends of a MMF, respectively. Since the diameter of the MMF core is larger than that of the SMF, many high-order eigenmodes {LPnm} are excited, while the beam comes to the MMF section. Suppose that the axes of the SMF and the MMF are perfectly aligned, only the symmetric modes {LP0m} can be excited and the input field in the MMF can be represented as1 E(r,0) =
M
cm Fm (r),
(1)
m=1
where M is the total number of excited modes, Fm (r) is the field profile of the LP0m mode, r is the radial coordinate in the cross section of the fiber, cm is the excitation coefficient of the LP0m mode, which can be expressed by the overlap integral
between E(r,0) and Fm (r) as follows:1 ∞ E(r,0) × Fm (r) ×r dr cm =
0
.
∞
(2)
Fm (r) × Fm (r) ×r dr 0
The field profile of the LP0m mode at a propagation distance z can be then calculated by1 E(r,z) =
M
cm Fm (r)exp(i β m z),
(3)
m=1
where β m is the propagating constant of the LP0m mode. When light is coupled back to the output SMF, the transmittance of SMS structure can be thus calculated by2 T(λ) =
M
2 cm × cn2 × cos[( β m − β n )L],
(4)
m, n=1
where L is the length of the MMF. When the RI of the MMF core or cladding changes, the eigenmodes Fm (r) excited in the MMF vary accordingly, leading to the changes of cm and β m of each mode. The length and diameter of the MMF core also influence the distribution of eigenmodes Fm (r) and the interference between the modes. Therefore, the final output to the SMF as shown in Eq. (4) is determined by the RIs of the MMF core and cladding, and the physical dimensions of the MMF core. Since all these parameters are sensitive to temperature, the output spectrum changes with temperature. This is the underlying principle of a SMS-based temperature sensor. In order to make a comprehensive study of the temperature characteristics of a SMS structure, we consider a model as shown in Fig. 1. The cladding of the MMF is substituted by certain thermal sensitive material (such as some kind of liquid) that has known thermal-optical features and is easy to be prepared. The section of the MMF is packaged by a shell, and both ends are sealed with UV glue. For such a SMS structure, when temperature changes ∆T, the RI changes of the MMF core and cladding induced by the TOE of the MMF can be expressed as ∆nco_1 = ξco × ∆T,
(5)
∆ncl = ξcl × ∆T,
(6)
where ξco and ξcl are the TOCs of the MMF core and cladding, respectively. The changes of the diameter and length of the MMF core induced by thermal expansion then can be expressed as ∆d 1 = d 0 × αco × ∆T,
(7)
∆l 1 = l 0 × αco × ∆T,
(8)
where d 0 and l 0 are the diameter and length of the MMF core at 25 ◦C, respectively, αco is the TEC of the MMF core. In addition, if the TEC of the packaging material (αp) is different from that of the MMF core (αco), an axial strain will be caused and can be given by ε = (αp − αco) × ∆T. (9) FIG. 1. Schematic of SMS fiber optic structure. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-3
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
The changes of the RI, diameter, and length of the MMF core, induced by the axial strain, can be, respectively, described by7,20 3 nco_0
[p12 − ν(p11 + p12)] × αp − αco × ∆T, 2 ∆d 2 = −ν × d × αp − αco × ∆T, ∆l 2 = l × αp − αco × ∆T, ∆nco_2 = −
(10) (11) (12)
where nco_0 is the RI of the MMF core at 25 ◦C, p11 and p12 are elastic-optic coefficients, ν is the Poisson ratio. For silica fibers, p11 = 0.12, p12 = 0.27, and ν = 0.17.20 Therefore, when temperature changes, the RIs of the MMF core and cladding, diameter, and length of the MMF core can be given, respectively, by nco = nco_0 + ∆nco_1 + ∆nco_2,
(13)
ncl = ncl_0 + ∆ncl,
(14)
d = d 0 + ∆d 1 + ∆d 2,
(15)
l = l 0 + ∆l 1 + ∆l 2,
(16) ◦
where ncl_0 is the RI of the MMF cladding at 25 C. For ease of description, we will refer to the TOEs of the MMF core and cladding as the TOEco and the TOEcl, the TOCs of the MMF core and cladding as the TOCco and the TOCcl, the TEEs of the MMF core and packaging material as the TEEco and the TEEp, the TECs of the MMF core and packaging material as the TECco and the TECp, and the RIs of the MMF core and cladding as the RIco and the RIcl, respectively. According to the above discussion, when the temperature changes, due to the TOEco, TOEcl, TEEco, and the axial strain caused by the different values of the TECco and TECp, the RIco, RIcl, diameter, and length of the MMF core change correspondingly, leading to the change of the transmission spectrum of the SMS structure. Hence, the temperature sensitivity of a SMS fiber structure depends on the combined effects of the TOE, TEE, and the induced axial strain of SMS structure. In order to make the temperature cross-sensitivity characteristics clear and provide the basis for choosing appropriate cladding and shell materials for improving the sensor’s performance or meeting various practical requirements, the impact of each factor on the temperature sensitivity of a SMS structure has been analyzed in the rest of the paper.
core is assumed to be made of pure silica with a diameter of 61.5 µm and its RI is calculated by the Sellmeier equation.20 Since the sensitivity demodulated from the wavelength shift of a SMS structure is independent of the MMF length and proportional to the dip wavelength,2 the lengths of the MMF core are selected to obtain an identical dip wavelength at 25 ◦C for various SMS structures with different MMF claddings. Then, based on Eqs. (1)–(4), (6), and (14), the numerical simulation can be carried out by utilizing the mode propagation analysis (MPA) method.1 Considering that the RI sensitivity is changing with the RI range,1 the temperature features of SMS structures with different RIcls and TOCcls are simulated. Fig. 2 shows the relationship between a dip wavelength (1569.16 nm) in the transmission spectrum and the temperature for various SMS structures. The inset shows the corresponding transmission spectra at 25 ◦C, where an interference dip with better visibility appears at 1569.16 nm for both spectra. It can be seen from Fig. 2 that the dip wavelength decreases linearly with the increase of temperature in the range from 25 to 80 ◦C for the SMS structures with different MMF claddings. The respective sensitivity can be obtained from the linear fit. As shown in Fig. 2, for the same TOCcl (ξcl = −1.5 × 10−4/◦C), when the RIcl (ncl) is 1.33 and 1.42, the temperature sensitivity is −31.8 pm/◦C and −278 pm/◦C, respectively. So the temperature sensitivity of the SMS structure significantly increases with the increase of the RIcl. Further, when the RIcl is kept at 1.42, but the TOCcl is changed to −2 × 10−4/◦C, the temperature sensitivity will be −343.6 pm/◦C. This indicates that a higher TOCcl will also be helpful to enhance the temperature sensitivity. So if a higher temperature sensitivity is desirable, we need to choose a material with a higher RI and/or higher TOC to serve as the cladding of the MMF section in a SMS structure. B. Temperature characteristics due to TOE of the MMF core in a SMS structure
In the study of the contribution of the TOEco to the temperature sensitivity of a SMS structure, we assume that the
III. ANALYSIS OF THE TEMPERATURE CHARACTERISTICS OF SMS FIBER OPTIC STRUCTURE A. Temperature characteristics due to TOE of the MMF cladding in a SMS structure
Because of the TOEcl, the RIcl will change with temperature. To numerically analyze the thermal sensing characteristics of a SMS fiber structure induced solely by the TOEcl, we first suppose that the thermal responses of the MMF core and packing shell are not considered, that is, nco, d, and l are unchanged with the changing temperature. In the simulations, the SMF is assumed to be the Corning SMF-28 with a core diameter of 8.2 µm and numerical aperture of 0.14. The MMF
FIG. 2. Dip wavelength as a function of the temperature for SMS structures fabricated with MMF claddings having different RIs and TOCs, the inset shows the corresponding transmission spectra of various SMS structures at 25 ◦C. Only the TOEcl is considered in the simulations. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-4
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
FIG. 3. Dip wavelength versus temperature for SMS structures fabricated with various MMF claddings having different RIs. Only the TOEco is considered in the simulations.
TOEcl, TEEco, and TEEp are not taken into account, i.e., the ncl, d, and l are not varying with temperature, and ∆nco_2 in Eq. (13) is zero. The other parameters are identical to those in Sec. III A. Based on Eqs. (1)–(5) and (13), the dip wavelengths in the transmission spectra as functions of temperature for the SMS structures fabricated with different MMF claddings are depicted in Fig. 3. For the standard silica fibers, ξco = 1.06 × 10−5/◦C.20 From Fig. 3, we can see that the resonant dip wavelengths change linearly with the increase of temperature, but the sensitivities are different. This indicates that the thermal sensing characteristic induced by the TOEco is influenced by the RIcl. The sensitivities of the SMS structures can be obtained by the linear fits. As shown in Fig. 3, when the RIcl is 1 (such as atmosphere), the dip wavelength shifts toward longer wavelength as the temperature increases, and the sensitivity is 11.1 pm/◦C. This sensitivity is close to the previously reported experimental results (15 pm/◦C and 12.9 pm/◦C),9,21 where the core and cladding of the MMF are mainly made of silica, having nearly the same RI and TOC, and hence can be equivalent to the MMF core in our simulations. Fig. 3 shows that with the increasing of the RIcl, the positive temperature sensitivity decreases, while when the RIcl reaches a certain value (between 1.40 and 1.41), the dip wavelength is almost insensitive to the temperature and then it shifts to shorter wavelengths with the rise of temperature. The temperature sensitivity reaches −13 pm/◦C when the RIcl is 1.42. Thus, the effect of the TOEco on the final temperature sensitivity of a SMS sensor depends on the RIcl.
FIG. 4. Dip wavelength versus temperature for SMS structures fabricated with various MMF claddings having different RIs. Only the TEEco is considered in the simulations.
different claddings are depicted in Fig. 4. For the standard silica fibers, αco = 5 × 10−7/◦C.20 As shown in Fig. 4, the temperature responses of the dip wavelengths of the SMS structures with different RIcls are nearly the same. The dip wavelength increases slowly with the increasing of temperature in the range from 25 to 80 ◦C. The small temperature sensitivity of 0.8 pm/◦C is caused by the low TEC of silicon. So we can conclude that the TEEco has less influence on the thermal sensitivity of a SMS structure than the TOEco and TOEcl. Besides the TEEco, the TEEp also influences the sensitivity of a SMS structure by introducing an axial strain, as described in Sec. II. Now, we analyze the effects of both of the TEEco and TEEp. The ∆nco_2 in Eq. (13), ∆d 2 in Eq. (15), and ∆l 2 in Eq. (16) are no longer null. With Eqs. (1)–(4),(7)–(13), (15), and (16), the dip wavelengths changing with temperature for SMS structures with different packing materials are simulated. The RIcl is assumed to be in the range from 1 to 1.42. The simulation results are shown in Fig. 5. Similar to Fig. 4, we can see from Fig. 5 that the dip wavelength does not change with the RIcl if the TECp is fixed. How-
C. Temperature characteristics due to TEEs in a SMS structure
In the evaluation of the contribution of the TEEco to the temperature sensitivity of a SMS structure, it is assumed that nco and ncl are independent of temperature, and ∆d 2 in Eq. (15) and ∆l 2 in Eq. (16) are zero. By using Eqs. (1)–(4), (7), (8), (15), and (16), the simulated relationships between the dip wavelength and temperature for the SMS structures with
FIG. 5. Dip wavelength versus temperature for SMS structures fabricated with various MMF claddings having different RIs and packaging shells with different TECs. Only the TEEco and TEEp are considered in the simulations. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-5
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
ever, the TECp does have an impact on the temperature sensitivity. As shown in Fig. 5, when the TECps are 5×10−6/◦C and 5 × 10−5/◦C, the sensitivities are −9 pm/◦C and −107 pm/◦C, respectively. This means that the higher the TECp is, the larger the negative sensitivity will be. So for SMS-based temperature sensors with positive or negative sensitivities, proper packaging materials with a smaller or bigger TEC should be selected to further optimize the sensing performance. D. Temperature characteristics based on TOE of MMF core and TEEs in SMS fiber optic structure
Finally, by comparing the respective temperature sensitivity induced by the TOEcl, TOEco, TEEco, and both of the TEEco and TEEp, we can conclude that the TOEcl plays a major role in determining the temperature sensitivity of a SMS structure. The thermal properties of the MMF core and the packaging materials also have influences on the sensitivity. In order to optimize the temperature performance of a SMS structure, we consider the combined effect of the TOEco, TEEco, and TEEp on the sensitivity of the sensors with several MMF claddings having different RIs. Simulations are based on Eqs. (1)–(5) and (7)–(16). The simulation results are shown in Fig. 6. Fig. 6 depicts the resonance dip wavelength versus temperature for the packaged SMS structures with various MMF claddings and packaging materials. It can be seen that the results depicted in Fig. 6(a) are similar to those in Fig. 3, that displays the effect of the TOEco, i.e., the change trend and extent of the dip wavelength with temperature are varying with the changing of the RIcl. This is attributed to the small and unchanged temperature sensitivity induced by the smaller TECco and TECp (see Fig. 5). It can be seen from Fig. 6(a) that when the RIcl is 1 and 1.42, the sensitivity is 2.1 pm/◦C and −21.9 pm/◦C, respectively. Whereas when the TECp is larger, Fig. 6(b) indicates that an increase in the temperature from 25 to 80 ◦C causes monotonic decreases in the dip wavelengths without exception, and the negative sensitivity increases with the increasing of the RIcl. A sensitivity of −118.7 pm/◦C has been achieved when the RIcl is 1.42. This result of the all negative sensitivities is due to the larger axial strain induced by the packaging shell (also see Fig. 5). So, the TEEp is the second important factor that influences the temperature sensitivity of a SMS structure. To maximize the temperature sensitivity of a packaged SMS structure, the proper packaging material should be selected according to the TOCcl. For example, if the MMF cladding has positive TOC, a packaging material with a smaller TEC is recommended, otherwise a packaging material with a larger TEC is better. Additionally, from Fig. 6(a), it should be noted that the temperature responses resulting from the TOEco, TEEco, and TEEp may counteract to each other under certain condition, which is useful for some temperature-insensitivepreferable sensors based on a SMS structure. IV. SMS FIBER STRUCTURE BASED TEMPERATURE SENSOR WITH ENHANCED SENSITIVITY A. Experiments
In order to experimentally demonstrate the improved temperature sensing property of a packaged SMS fiber optic
FIG. 6. Dip wavelength versus temperature for SMS structures fabricated with various MMF claddings having different RIs and packaging shell with different TECs. Only the TOEco, TEEco, and TEEp are considered in the simulations.
structure by using proper materials, as well as verify the theoretical findings based on numerical simulations, we designed, fabricated, and experimentally studied a SMS-based temperature sensor. The sensor head, as shown in Fig. 1, is fabricated by first splicing two ends of the MMF to two SMFs, which is accomplished by a commercial fusion splicer in manual operation mode. The SMF and MMF employed in our experiments are the standard SMF-28 (Corning, Inc.) and commercialized no-core fiber (NCF, Prime Optical Fiber Co.), respectively. The NCF is made of pure silica. The diameter and length of the NCF are 61.5 µm and 8.2 cm, respectively. Then, the NCF section is fed into a glass capillary with an approximate TEC of 5 × 10−6/◦C,22 an inner diameter of 0.5 mm, and a length of 10 cm. Finally, the capillary is filled with glycerol–water mixture (3:1 volume mixture), and both ends of the capillary are sealed with UV glue. The glycerol–water mixture serves as the cladding of the NCF. Glycerol–water mixture (3:1 volume mixture) is chosen because it is a conventional and easily accessible liquid with a larger RI around 1.43 and a TOC of −1.827 × 10−4/◦C.23 The packaging capillary not only makes the senor convenient to be
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-6
Zhang et al.
Rev. Sci. Instrum. 86, 013108 (2015)
FIG. 7. Experimental setup for investigating the temperature properties of the SMS-based sensor.
used but also enhances the temperature sensitivity of sensor due to the larger TEE of the glass. The experimental setup is shown in Fig. 7, light from a broadband amplified spontaneous emission (ASE, LightComm Technology, Ltd.) source passes through the SMSbased sensor, and the transmission spectrum is recorded by an optical spectrum analyzer (OSA, Yokogawa Electric Co.). In characterization of the temperature response of the sensor, the sensing head is placed into a temperature-controlled oven. A thermometer with an accuracy of ±0.5 ◦C is used to measure the temperature in real time. The sensor head of the thermometer is commercial platinum resistance which is placed near the SMS structure in the oven through a small hole. The sensor is first temperature-cycled between 25 ◦C and 80 ◦C for several times to relax strain. Then, the transmission spectra are recorded with the OSA with a temperature step of 5 ◦C.
It can be seen from Fig. 8 that interference patterns are clearly obtained in the experimental transmission spectrum. We chose the dip at 1592.78 nm as the spectral indicator for the temperature sensing, because it has good visibility and large dynamic range in wavelength domain in our experiment. The measured spectral responses under different temperatures are shown in Fig. 9(a), as can be seen; with the temperature rising, the dip wavelength of the resonant spectrum shifts toward shorter wavelength due to the decreasing RI of the glycerol–water mixture. In order to obtain the temperature sensitivity of the sensor, based on Fig. 9(a), the dependence of the dip wavelength on temperature is explicitly plotted in Fig. 9(b). It is shown that the dip wavelength decreases linearly with the increase of temperature. The experimental results can be fitted with a good linearity, as shown in Fig. 9(b). The slope of the linear fit is −453.4 pm/◦C. In the corresponding range of temperature, this sensitivity is not only much higher than that of SMS-based sensor using conventional MMF with silica as cladding (12–45 pm/◦C)9,10,21 but also higher than that of the SMS-based sensor with commercial RI liquid (a nominal RI value of 1.44 (20 ◦C)) as MMF cladding but without the packaging shell (−360 pm/◦C)11 and also higher
B. Results and discussions
Fig. 8 shows the experimental and simulated transmission spectra of the sensor at 25 ◦C. The simulation is carried out by using the MPA method based on Eqs. (1)–(4), and the parameters are taken as the employed materials described in Sec. IV A. From Fig. 8, one can see that the theoretical spectrum has a reasonable agreement with the measured result. There is a dip wavelength (1592.78 nm) in both of the simulated and experimental spectra. The discrepancy between the two transmission spectra could be mainly due to the imperfect arrangement of the axes of the SMF and NCF.
FIG. 9. (a) Transmission spectra under different temperatures and (b) variation of dip wavelength of the spectrum with the temperature for the SMSFIG. 8. Transmission spectra of the SMS-based sensor at 25 ◦C. based sensor. This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:
169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
013108-7
Zhang et al.
than that of the SMS device employing deionized water as the MMF cladding and aluminum alloy shell as packing material (358 pm/◦C).12 So here, according to the numerical analyses, the glycerol–water mixture with a high RI and a negative TOC, and the packaging glass capillary having a higher TEC compared with the NCF are both contributors to the enhanced temperature sensitivity. Simulation results with Eqs. (1)–(16) are also shown in Fig. 9(b). In the simulations, all of the effects of the TOEco, TOEcl, TEEco, and TEEp are taken into account. The parameters are still taken as the experimental materials described in Sec. IV A. The sensitivity from the simulation results is −463.6 pm/◦C, which is close to the experimental value. This confirms the validity of our theoretical analyses.
V. CONCLUSIONS
In summary, we carry out a comprehensive analysis of the temperature cross-sensitivity characteristics of a SMS structure. Based on the MPA method, it is found that the TOEcl and TEEp are the first and second important contributors to the temperature sensitivity of a SMS structure, respectively. Further, the combined effects of TOEco, TEEco, and TEEp indicate that, for the MMF cladding with a positive/negative TOC, the packaging material with a smaller/larger TEC is more suitable for improving the thermal sensitivity of the SMS structure. In addition, the RIcl also influences the effects of various factors. Based on the results of numerical analysis, we fabricate a SMS-based temperature sensor with a NCF cladding of glycerol–water mixture and a packing shell made of glass. The measured temperature sensitivity is −453.4 pm/◦C. The experimental data match our simulation very well. Although the sensor, featuring a simple structure and fabrication process, performs very well, its sensing performance can be further improved by using appropriate materials, such as a NCF cladding with a larger RI and/or larger negative TOC, and packaging shell with a bigger TEC. Additionally, the conclusions derived in this paper can also provide a reference for eliminating cross-sensitivity to temperature for the non-temperature sensors based on SMS structure.
Rev. Sci. Instrum. 86, 013108 (2015)
ACKNOWLEDGMENTS
This work was supported by the Natural Science Foundation of Tianjin (Grant No. 13JCYBJC16100), the National Natural Science Foundation of China (Grant No. 61107035), the National Key Scientific Instrument and Equipment Development Project of China (Grant No. 2013YQ03091502), and the National Basic Research Program of China (Grant No. 2014CB340104). 1Q.
Wu, Y. Semenova, P. Wang, and G. Farrell, Opt. Exp. 19, 7937 (2011). Chen, Q. Han, T. Liu, and H. Xiao, IEEE Photon. J. 6, 6801807 (2014). 3H. Wang, S. Pu, N. Wang, S. Dong, and J. Huang, Opt. Lett. 38, 3765 (2013). 4Y. Chen, Q. Han, T. Liu, X. Lan, and H. Xiao, Opt. Lett. 38, 3999 (2013). 5L. Li, Q. Han, T. Liu, Y. Chen, and R. Zhang, Rev. Sci. Instrum. 85, 083107 (2014). 6Q. Wu, Y. Semenova, P. Wang, A. M. Hatta, and G. Farrell, Meas. Sci. Technol. 22, 025203 (2011). 7E. Li, IEEE Photon. Technol. Lett. 19, 1266 (2007). 8Y. Gong, T. Zhao, Y.-J. Rao, and Y. Wu, IEEE Photon. Technol. Lett. 23, 679 (2011). 9E. Li, X. Wang, and C. Zhang, Appl. Phys. Lett. 89, 091119 (2006). 10Q. Wu, Y. Semenova, A. M. Hatta, P. Wang, and G. Farrell, Electron. Lett. 46, 1129 (2010). 11S. Silva, E. G. P. Pachon, M. A. R. Franco, J. G. Hayashi, F. X. Malcata, O. Frazão, P. Jorge, and C. M. B. Cordeiro, Appl. Opt. 51, 3236 (2012). 12H. Sun, M. Hu, Q. Rong, Y. Du, H. Yang, and X. Qiao, Opt. Commun. 323, 28 (2014). 13Y. Liu, Z. Zhou, E. Zhang, J. Zhang, Y. Tan, and M. Liu, Rev. Sci. Instrum. 85, 064905 (2014). 14S. K. Mondal, U. Tiwari, G. C. Poddar, V. Mishra, N. Singh, S. C. Jain, S. N. Sarkar, K. D. Chattoypadhya, and P. Kapur, Rev. Sci. Instrum. 80, 103106 (2009). 15A. P. Zhang, L.-Y. Shao, J.-F. Ding, and S. He, IEEE Photon. Technol. Lett. 17, 2397 (2005). 16J. Wang, B. Dong, E. Lally, J. Gong, M. Han, and A. Wang, Opt. Lett. 35, 619 (2010). 17Z. Cao, Z. Zhang, X. Ji, T. Shui, R. Wang, C. Yin, S. Zhen, L. Lu, and B. Yu, Opt. Fiber Technol. 20, 24 (2014). 18D. Monzón-Hernández, V. P. Minkovich, and J. Villatoro, IEEE Photon. Technol. Lett. 18, 511 (2006). 19W. Qian, C.-L. Zhao, S. He, X. Dong, S. Zhang, Z. Zhang, S. Jin, J. Guo, and H. Wei, Opt. Lett. 36, 1548 (2011). 20S. M. Tripathi, A. Kumar, R. K. Varshney, Y. B. P. Kumar, E. Marin, and J.-P. Meunier, J. Lightwave Technol. 27, 2348 (2009). 21J. G. Aguilar-Soto, J. E. Antonio-Lopez, J. J. Sanchez-Mondragon, and D. A. May-Arrioja, J. Phys.: Conf. Ser. 274, 012011 (2011). 22J. Arndt and F. Häberle, Contrib. Mineral. Petrol. 39, 175 (1973). 23J. Popp, M. Lankers, K. Schaschek, W. Kiefer, and J. T. Hodges, Appl. Opt. 34, 2380 (1995). 2Y.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 169.230.243.252 On: Tue, 31 Mar 2015 12:52:17
