applied sciences Article
Refractive Index Sensing through Surface Plasmon Resonance in Light-Diffusing Fibers Nunzio Cennamo *, Luigi Zeni, Ester Catalano, Francesco Arcadio and Aldo Minardo
ID
Department of Engineering, University of Campania Luigi Vanvitelli, Via Roma 29, 81031 Aversa, Italy;
[email protected] (L.Z.);
[email protected] (E.C.);
[email protected] (F.A.);
[email protected] (A.M.) * Correspondence:
[email protected]; Tel.: +39-0815010367
Received: 28 June 2018; Accepted: 16 July 2018; Published: 18 July 2018
Abstract: In this paper, we show that light-diffusing fibers (LDF) can be efficiently used as host material for surface plasmon resonance (SPR)-based refractive index sensing. This novel platform does not require a chemical procedure to remove the cladding or enhance the evanescent field, which is expected to give better reproducibility of the sensing interface. The SPR sensor has been realized by first removing the cladding with a simple mechanical stripper, and then covering the unclad fiber surface with a thin gold film. The tests have been carried out using water–glycerin mixtures with refractive indices ranging from 1.332 to 1.394. The experimental results reveal a high sensitivity of the SPR wavelength to the outer medium’s refractive index, with values ranging from ~1500 to ~4000 nm/RIU in the analyzed range. The results suggest that the proposed optical fiber sensor platform could be used in biochemical applications. Keywords: light-diffusing fibers (LDF); surface plasmon resonance (SPR); optical sensors
1. Introduction Refractive index (RI) sensing is important for biological and chemical applications since a number of substances can be detected through measurements of the refractive index. Among RI sensors, those based on surface plasmon resonance (SPR) are especially useful as they permit label-free and real-time detection of biological/biochemical binding reactions [1–3]. In particular, fiber-optic SPR probes have the advantages of remote sensing, continuous analysis, and in situ monitoring. Fiber-optic SPR sensors meet the demands of time- and space-saving, low sample volume, low cost, high sensitivity, portability, and miniaturization. Therefore, these biosensors have recently been extensively investigated [4–6]. Numerous versions have been introduced, exploiting several kinds of optical fibers and/or different manufacture procedures, in an attempt to optimize the sensitivity of the plasmonic sensor platforms [7–14]. In general, the excitation of a surface plasma wave at the boundary between a thin metal film and the surrounding medium requires access to the evanescent part of propagating light [13]. In single-mode fibers, this requires a chemical etching of the cladding, which increases the cost of manufacturing and impacts on the robustness of the sensor. In multimode fibers with polymeric cladding (the so-called hard-clad fibers), the cladding may be more simply removed by mechanical tools. However, the multimode SPR sensor exhibits a strong dependence of its response on the angle of incidence of the input light [7,12]. Herein we demonstrate a novel fiber-optic SPR sensor using a light diffusing fiber (LDF) as the host element. LDFs are designed with light-scattering centers in the core, providing very efficient scattering of light through the sides of the optical fiber and along their entire length. Therefore, the SPR response of an LDF is virtually independent of the angle of incidence of excitation light. Note that Appl. Sci. 2018, 8, 1172; doi:10.3390/app8071172
www.mdpi.com/journal/applsci
Appl. Sci. 2018, 8, 1172 Appl. Sci. 2018, 8, 1172
2 of 8 2 of 8
an platform based on aon plastic LDFLDF has has beenbeen recently demonstrated [14]. However, as seen in thatSPR an SPR platform based a plastic recently demonstrated [14]. However, as seen Reference [14], a polishing procedure was still adopted in order to remove the cladding and produce in Reference [14], a polishing procedure was still adopted in order to remove the cladding anda side-polished surface. surface. produce a side-polished In the following, we first characterize, numerically and and experimentally, the LDF in terms of transmission loss loss as asaafunction functionofofthe thesurrounding surrounding refractive index (SRI). Through analysis, refractive index (SRI). Through thisthis analysis, we we estimate optical energy within fibercore coreisiscompletely completelyscrambled scrambledover over all all the the possible estimate thatthat thethe optical energy within thethe fiber angles. Successively, we present the results of an SPR sensor, realized by depositing a thin gold film around TheThe experimental results are finally compared to thosetofound the literature around the theunclad uncladregion. region. experimental results are finally compared thoseinfound in the for a hard-clad silica (HCS) fiber. literature for a hard-clad silica (HCS) fiber. 2. SPR Sensor Sensor Based Based on 2. SPR on an an LDF LDF ® by Corning® , New York, NY, USA) is composed The The LDF LDF chosen chosen as as the the host host platform platform (Fibrance (Fibrance® by Corning®, New York, NY, USA) is composed of a silica core with a diameter of 170 µm, and a low-index of a silica core with a diameter of 170 µm, and a low-index polymeric polymeric cladding cladding with with aa diameter diameter of of 230 µm [15] (see Figure 1a). In the chosen LDF, the guided light is scattered radially outward 230 µm [15] (see Figure 1a). In the chosen LDF, the guided light is scattered radially outward through through aa plurality distributed in in the the core, core, and plurality of of helical helical voids voids randomly randomly distributed and wrapped wrapped around around the the long long axis axis of of the the optical fiber (see Figure 1b). The pitch of the helical voids controls the amount of the side-emitted light, optical fiber (see Figure 1b). The pitch of the helical voids controls the amount of the side-emitted light, with ranges in with smaller smaller pitches pitches scattering scattering more more light light than than larger larger pitches, pitches, while while their their diameter diameter ranges in size size from from 50–500 nm; consequently, they scatter the propagating light almost independently of the wavelength 50–500 nm; consequently, they scatter the propagating light almost independently of the wavelength of of light used [15]. light used [15].
(a)
(b)
Figure 1. (a) Schematic cross-section of the light-diffusing fiber (LDF) used for surface plasmon Figure 1. (a) Schematic cross-section of the light-diffusing fiber (LDF) used for surface plasmon resonance (SPR) sensing; (b) Schematic lateral view of the LDF used for SPR sensing, showing the resonance (SPR) sensing; (b) Schematic lateral view of the LDF used for SPR sensing, showing the portion of unclad fiber. portion of unclad fiber.
For our tests, a 2 cm length of the fiber was unclad by means of a Miller stripping tool. In the first For our tests, a 2 cm length theoffiber was unclad by means of awhite Miller stripping tool. the first experimental configuration, one of end the fiber was coupled to the light emitted by aIntungsten experimental oneOcean end of Optics), the fiberwhile was coupled to the white light by athrough tungstena halogen light configuration, source (HL2000, the transmitted light wasemitted collected halogen light source (HL2000, Ocean Optics), while the transmitted light was collected through a spectrophotometer with a detection range from 350 to 1023 nm (FLAME-S-VIS-NIR-ES, Ocean Optics). spectrophotometer with a detection range from 350 to 1023 nm (FLAME-S-VIS-NIR-ES, Ocean Optics). The outline of this sensor system (intensity-based LDF sensor) is shown in Figure 2a. The outline of sensor this sensor (intensity-based LDF sensor) shown coating in Figure(Bal-Tec 2a. The SPR was system then obtained by depositing, throughissputter SCD 500), a The SPR sensor was then obtained by depositing, through sputter coating SCD 500), thin gold film over the same unclad portion of the LDF. Gold was chosen due to its(Bal-Tec high stability and aease thinofgold film over the same unclad portion of the LDF. Gold was chosen due to its high stability functionalization for biosensing applications. The sputtering process was repeated twice, and of functionalization for in biosensing applications. The sputtering process wasThe repeated twice, uponease rotating the fiber by 180°, order to metalize the whole fiber circumference. so-obtained ◦ , in order to metalize the whole fiber circumference. The so-obtained upon rotating the fiber by 180 gold film had a nominal thickness of 60 nm, and presented a good adhesion to the substrate, tested gold had a nominal thickness of 60 nm, and good adhesion to the substrate, tested by by itsfilm resistance to rinsing in de-ionized water. Aspresented shown inaFigure 2b, the changes in the transmission its resistance to rinsing in de-ionized water. As shown in Figure 2b, the changes in the transmission spectrum induced by changes of the SRI have been recorded by using the same equipment previously spectrum changes of the SRI have been recorded by using the same equipment previously describedinduced (halogenby lamp and spectrometer). described (halogen lamp and spectrometer).
Appl. Appl. Sci. Sci. 2018, 2018, 8, 8, 1172 1172
of 88 33 of
(a)
(b) Figure 2. Outline of the LDF sensor systems: (a) Intensity-based LDF sensor; (b) SPR-LDF sensor. sensor.
3. Results Results and and Discussion Discussion 3. As a first test, test, we have recorded the changes in the transmission spectrum of the fiber, with the water–glycerol mixtures with RI ranging fromfrom 1.3321.332 to 1.443. From uncoated fiber fiberregion regionimmersed immersedinin water–glycerol mixtures with RI ranging to 1.443. the measurements, it resulted that the external medium’s RI only influenced the total power From the measurements, it resulted that the external medium’s RI only influenced the total LDF, while while its its spectral spectral distribution distribution kept kept constant. constant. Therefore, we report in transmitted through the LDF, see that, that, for for RI RI Figure 3a only the changes in the transmitted power, as a function of the SRI. We We see ranging from 1.332 to ~ 1.38 the transmitted power remained constant, while for RI larger than 1.38 ranging from 1.332 to ~1.38 the transmitted power remained constant, while for RI larger than 1.38 the the transmitted power dropped sensibly –4when dB when varying thefrom RI from to 1.443). transmitted power dropped sensibly (∆P~(P −4~dB varying the RI 1.3321.332 to 1.443). This behavior behavioris is consistent a ray-optic description of light transmission through a This consistent withwith a ray-optic description of light transmission through a multimode multimode fiber [16]: as long as the external medium’s RI is lower than the cladding index fiber [16]: as long as the external medium’s RI is lower than the cladding index (about 1.38 (about in our 1.38 inthe ourpower case), the power guided fiber is onlyby affected by evanescent wave absorption (EWA), case), guided in the fiberinisthe only affected evanescent wave absorption (EWA), which is which is negligible in our case due to the short sensing region. Vice versa, when the RI of the external negligible in our case due to the short sensing region. Vice versa, when the RI of the external medium medium situated the between the cladding andRI, thethe core RI, the variation in the transmitted power is situatedisbetween cladding RI and theRIcore variation in the transmitted power is mostly is mostly to the reduction of theofnumber of propagation modes satisfying the totalreflection internal due to thedue reduction of the number propagation modes satisfying the total internal reflection (TIR) condition at the boundary the fiber core and the external medium (see Figure (TIR) condition at the boundary betweenbetween the fiber core and the external medium (see Figure 1b). 1b). These arguments are confirmed a numerical analysis carried using model described These arguments are confirmed by aby numerical analysis carried outout using thethe model described in in Reference [17]: in brief, the normalized power transmitted over a sensing length 𝐿 is calculated by Reference [17]: in brief, the normalized power transmitted over a sensing length L is calculated by −1 ( n 𝜃 (𝜃𝑐 , 𝜋⁄θ2c )=, sin summing the thepower power contributions over the acceptance where summing contributions over all the all acceptance angles (θcangles , π/2), where 𝑐 c0 = )) cl /n −1 (𝑛 ⁄ 𝑠𝑖𝑛 𝑐𝑙 𝑛𝑐0 )) is the critical angle outside the sensing region. Assuming an unpolarized collimated light source, we can express the normalized transmitted power as [17]:
Appl. Sci. 2018, 8, 1172
4 of 8
is the critical angle outside the sensing region. Assuming an unpolarized collimated light source, Appl. Sci. 2018, 8, 1172 4 of 8 we can express the normalized transmitted power as [17]: 𝜋⁄2 𝜋⁄2 ) 𝑁(𝜃 ) N𝑖(θi ) ∑𝜃 =𝜃 𝑃0 (𝜃P𝑖 )(θ∙ 𝑅)·𝑝R(𝜃(𝑖 )θ𝑁(𝜃 π/2 π/2𝑃0 (𝜃𝑖 ) ∙ 𝑅𝑠 (𝜃𝑖 ) N (θ𝑖i ) 1 ∑𝜃𝑖=𝜃 P θ R θ 𝑐 𝑐 ) ( )· ( ) 𝑖 ∑ ∑ p s 0 0 i i i i 𝑃𝐿 P= =(1 θi =𝜋θc⁄2 + + θi =θc𝜋⁄2 ) L2 π/2 π/2 𝑃0 (𝜃𝑖 ) ∑𝜃 ∑ (𝜃 )θ ) ∑∑ 𝑃 2 0 𝑖 =𝜃 𝜃 =𝜃 P P θ ( ( ) 𝑖 θ𝑐=θc 0 𝑖 θ =𝑐θc 0 i i i i
(1) (1)
where 𝑃0 (𝜃𝑖 ) is the initial power propagated by TIR, 𝑁(𝜃𝑖 ) = 𝑐𝑜𝑡(𝜃𝑖 )𝐿⁄𝑑 is the number of where P (θ ) is the initial power propagated by TIR, N (θ ) = cot(θi ) L/d is the number of reflections, reflections,0 𝑑i is the core diameter, 𝑅𝑝 and 𝑅𝑠 are ithe Fresnel reflectance coefficients at the d is the core diameter, R p and Rs are the Fresnel reflectance coefficients at the boundary between the boundary between the core and the sensing medium for p-polarization and s-polarization, core and the sensing medium for p-polarization and s-polarization, respectively. As regards to the respectively. As regards to the distribution of power among the various modes, the typical distribution of power among the various modes, the typical assumption about multimode fibers is assumption about multimode is a Lambertian source,when whichthe is actually the most accurate when a Lambertian source, which fibers is actually the most accurate fiber is illuminated through a thelens fiber is illuminated through (NA) a lenslarger with than a numerical aperture (NA) this larger than that the of the fiber. with a numerical aperture that of the fiber. Under assumption, power Under this assumption, the power distribution can be expressed as [17–19]: distribution can be expressed as [17–19]: 2 𝑛𝑐𝑜 𝑠𝑖𝑛𝜃𝑖 𝑐𝑜𝑠𝜃𝑖 (2) n2co2sinθi cosθ 𝑃0 (𝜃𝑖 ) ∝ (2) P0 (θi ) (1 ∝ − 𝑛𝑐𝑜 𝑐𝑜𝑠 2 𝜃𝑖 )i22 (1 − n2co cos2 θi ) From Figure 3, however, we see that the loss computed under the Lambertian source assumption From Figure however, we seeobserved that the loss computed under the Lambertian source assumption does not match the3,experimentally behavior. Instead, an excellent agreement is found if does not match the experimentally observed behavior. Instead, an excellent agreement is found if is assuming a perfectly scrambled angular distribution (𝑃0 (𝜃𝑖 ) = 𝑐𝑜𝑛𝑠𝑡 in Equation (1)). This finding assumingconsidering a perfectly scrambled angular (P0 (operated θi ) = const Equation This finding is reasonable the scrambling of distribution optical energy byinthe helical (1)). voids running along reasonable considering the scrambling of optical energy operated by the helical voids running along the core. To best illustrate this point, we compare in Figure 3b the angular distribution of power as the core. To best illustrate this point, we compare in Figure 3b the angular distribution of power as expressed by Equation (2) (Lambertian source) with the one speculated for our LDF. It can be seen expressed by Equation (2) (Lambertian source) with the one speculated for our LDF. It can be seen that, that, in the case of the Lambertian source, the power density of rays is higher near the critical angle in the case of the Lambertian source, the power density of rays is higher near the critical angle (θ ≈ 72◦ (𝜃𝑖 ≈ 72° in our case), therefore the loss of TIR for these rays, consequent to the increase ofi the SRI, in our case), therefore the loss of TIR for these rays, consequent to the increase of the SRI, has a greater has a greater impact on the transmitted power. Instead, for the scrambled, non-Lambertian case, impact on the transmitted power. Instead, for the scrambled, non-Lambertian case, power is equally power is equally distributed among the various angles; therefore, the relative weight of rays escaping distributed among the various angles; therefore, the relative weight of rays escaping away due to loss away dueis to loss of TIR is lower. of TIR lower.
Figure transmitted power through the as a offunction of the external Figure3.3.(a) (a) Variation Variation ofofthethe transmitted power through the LDF as LDF a function the external medium’s medium’s RI. The continuous lines are the simulation results. (b) Angular distribution of optical RI. The continuous lines are the simulation results. (b) Angular distribution of optical power. power.
After this preliminary study, we have used the SPR-based LDF sensor configuration. Figure 4a After this preliminary study, we have used the SPR-based LDF sensor configuration. Figure 4a shows the experimental spectral transmission of our sensor, normalized to the spectrum recorded shows experimental spectral transmission of our sensor, normalized spectrum with the air as the surrounding medium, for different water–glycerol solutions.toAthe clear red-shiftrecorded of the with air as the surrounding for different A clear medium. red-shift of resonance wavelength (SPRmedium, dip) is observed whenwater–glycerol increasing the solutions. RI of the sensing Forthe resonance wavelength dip) 4b is we observed when increasing spectra the RI obtained of the sensing medium. comparison purposes,(SPR in Figure show the corresponding by using the sameFor comparison purposes, in Figure 4b we show the corresponding spectra obtained by using the same model previously described (Equation (1) with 𝑃0 (𝜃𝑖 ) = 𝑐𝑜𝑛𝑠𝑡 ), with the only difference that reflectance coefficients are now calculated taking into account the metal film.
Appl. Sci. 2018, 8, 1172
5 of 8
model previously described (Equation (1) with P0 (θi ) = const), with the only difference that reflectance 5 of 8 coefficients are now calculated taking into account the metal film.
Appl. Sci. 2018, 8, 1172
Figure 4. Experimental (a)(a) and spectra obtained obtainedafter afternormalization normalization to the Figure 4. Experimental andnumerical numerical(b) (b) SPR SPR spectra to the transmission spectrum in in air. transmission spectrum air.
In order to to examine the recall some someparameters, parameters,such such sensitivity In order examine theobtained obtainedresults, results, we we recall as as sensitivity andand signal-to-noise ratio (SNR). (S)can canbebe defined as the in resonance wavelength signal-to-noise ratio (SNR).The The sensitivity sensitivity (S) defined as the shiftshift in resonance wavelength per per unit unit change changeininrefractive refractive index [12,19–22]. is usually reported in nanometers of shift peakper shift index [12,19–22]. It isItusually reported in nanometers of peak RI per Unit(nm/RIU): (nm/RIU): RI Unit i h
δλ nm S𝛿𝜆 =SPR SPR (3) nm RIU δn sm 𝑆= [ ] (3) 𝛿𝑛 δn RIU where δλSPR is the SPR wavelength shift, and𝑠m is the variation of refractive index of the sensing sm medium. the SNR of an SPR with interrogation a dimensionless where 𝛿𝜆𝑆𝑃𝑅 Analogously, is the SPR wavelength shift, and sensor 𝛿𝑛𝑠𝑚 is thespectral variation of refractiveisindex of the sensing parameter defined as [19]: medium. Analogously, the SNR of an SPR sensor with spectral interrogation is a dimensionless ∆λSPR parameter defined as [19]: SNR = (4) ∆λ
∆𝜆SPR where ∆λSPR is the SPR wavelength shift induced by a prescribed RI change, and ∆λ is the spectral(4) SNR = ∆𝜆some reference level of the transmitted power [19]. width of the SPR response curve corresponding to From 4 we see that, differently from the previous results, the agreement between experimental where ∆𝜆Figure 𝑆𝑃𝑅 is the SPR wavelength shift induced by a prescribed RI change, and ∆𝜆 is the spectral and numerical data is notcurve excellent: in particular,tothe experimental a deeper and wider[19]. width of the SPR response corresponding some referencespectra level ofexhibit the transmitted power resonance. Furthermore, the RI-induced SPR dip shift is larger in the experiments (173 nm in the From Figure 4 we see that, differently from the previous results, the agreement between experimental analyzed range against a shift of 128 nm derived from simulations). In other words, the experimental and numerical data is not excellent: in particular, the experimental spectra exhibit a deeper and wider sensitivity (S) is better than numerical sensitivity; vice versa, the SNR is worse. This discrepancy may resonance. Furthermore, the RI-induced SPR dip shift is larger in the experiments (173 nm in the be attributed to several reasons: first, some nonuniformity of the gold film thickness may exist along analyzed range against a shift of 128 nm derived from simulations). In other words, the experimental the length and/or circumference of the fiber; second, our model only takes into account meridional sensitivity (S) is better than numerical versa, the is worse. This discrepancy may rays, while some skew rays may exist,sensitivity; especially ifvice considering theSNR influence of the scattering centers; be attributed to severalofreasons: first, some nonuniformity of the gold film thickness exist along third, the scattering light due to the helicoid voids has been taken into account only may partly in our the length and/or circumference of the fiber; second, our model only takes into account meridional model, in particular setting a uniform angular power distribution. However, the presence of the rays,scattering while some skew rays may exist, if of considering the p-polarization influence of the scattering centers; centers could also affect theespecially distribution energy among and s-polarization. third, the scattering light due to the helicoid voids haswithin been the taken partlyonly in our Unfortunately, an of accurate description of light scattering fiberinto coreaccount may be only performed if a statistical distribution the helicalangular voids were available, which was not ourthe case. It is worth model, in particular setting of a uniform power distribution. However, presence of the noting that, for nsm = 1.332, experimental response indicates a resonance wavelength of ≈630 and nm sscattering centers could alsotheaffect the distribution of energy among p-polarization and a minimum transmittivity of ≈ 52%, while the corresponding numerical values are ≈ 650 nm and polarization. Unfortunately, an accurate description of light scattering within the fiber core may be ≈75%, respectively. As discussed in Reference [23], the propagation skew rayswhich leads to a downward, performed only if a statistical distribution of the helical voids wereofavailable, was not our case. as well as a backward, shift in the SPR curve, which is consistent with our observations and suggests a It is worth noting that, for 𝑛𝑠𝑚 = 1.332 , the experimental response indicates a resonance wavelength major influence of skew rays, compared to a conventional multimode fiber.
of ≈630 nm and a minimum transmittivity of ≈52%, while the corresponding numerical values are ≈650 nm and ≈75%, respectively. As discussed in Reference [23], the propagation of skew rays leads to a downward, as well as a backward, shift in the SPR curve, which is consistent with our observations and suggests a major influence of skew rays, compared to a conventional multimode fiber. Finally, Figure 5 shows the experimentally obtained resonance wavelength versus the SRI, together with the results of a quadratic fitting. From Figure 5, we see that the sensitivity is not constant, increasing from ≈1500 to ≈4000 nm/RIU in the analyzed range.
Appl. Sci. 2018, 8, 1172
6 of 8
Finally, Figure 5 shows the experimentally obtained resonance wavelength versus the SRI, together Appl. Sci. 2018, with 8, 1172the results of a quadratic fitting. From Figure 5, we see that the sensitivity is not 6 of 8 constant, increasing from ≈1500 to ≈4000 nm/RIU in the analyzed range.
Figure 5. SPR wavelengthas asaafunction function of areare extracted fromfrom experimental data, data, Figure 5. SPR wavelength of the theSRI. SRI.The Thecircles circles extracted experimental while the continuous line is the quadratic fitting curve. while the continuous line is the quadratic fitting curve.
Several comparative analysesbetween betweendifferent different fiberoptic-based fiber-optic-based SPR Several comparative analyses SPR sensor sensorconfigurations configurations (D(D-shaped tapered fibers, single-modefibers, fibers, U-shaped U-shaped fibers, fibers, heterocore fibers,fibers, shaped fibers,fibers, tapered fibers, single-mode fibers,TFBGs TFBGs fibers, heterocore etc.) can be found in excellent review papers [4,5,24]. Here, we just compare the performance of ourof our etc.) can be found in excellent review papers [4,5,24]. Here, we just compare the performance sensor with those similarSPR SPR sensor sensor configuration onon an an HCS optical fiberfiber (FT300EMT sensor with those ofofa asimilar configurationbased based HCS optical (FT300EMT from Thorlabs) [19]. In that paper, a very similar setup based on a tungsten halogen light source and a from Thorlabs) [19]. In that paper, a very similar setup based on a tungsten halogen light source and spectrophotometer was used to characterize the sensor. The HCS sensor presented in Reference [19] a spectrophotometer was used to characterize the sensor. The HCS sensor presented in Reference [19] had a length of 2 cm, a core and cladding diameter of 300 and 330 µm, respectively, and was coated hadwith a length 2 cm, core and cladding of 300 and 330 µm, respectively, a thinof gold film.aWe compare in Table 1diameter the performance of our LDF-based SPR sensorand withwas thosecoated with thin gold film. compare Table 1 [19] the performance of our LDF-based sensor of athe HCS sensor asWe extracted frominReference for two refractive indices of interestSPR (1.367 and with those of the HCS sensor as extracted from Reference [19] for two refractive indices of interest 1.384). These values have been chosen as they are typical of biosensors based on bioreceptors with(1.367 andanalytes 1.384). in These have beenand chosen as they are typical of thin biosensors onreceptors bioreceptors watervalues solution (≈1.36), of chemical sensors based on films of based polymer with compounds in water solution(≈1.36), (≈1.38).and Furthermore, the ∆λ appearing in Equation with analytes in water solution of chemical sensors based on thin films of polymer SPR parameter (4) has with been chosen in relation to ∆nsolution whileFurthermore, the ∆λ was taken the middle level between sm = 0.015, receptors compounds in water (≈1.38). theat∆𝜆 parameter appearing in 𝑆𝑃𝑅 the maximum and minimum power. Equation (4) has been chosen in relation to ∆𝑛𝑠𝑚 = 0.015, while the ∆𝜆 was taken at the middle level between the maximum and minimum power. Table 1. Comparison of performance of two different SPR multimode optical fiber sensors for two refractive indices (1.367 and 1.384).
Table 1. Comparison of performance of two different SPR multimode optical fiber sensors for two Resonance Sensitivity SNR refractive indices (1.367 and 1.384). Sensor Configuration n sm
SPR
SPR sensor based on multimode HCS optical fiber [19] Sensor Configuration SPR based on LDF SPR sensor based on multimode HCS optical fiber [19] sensor based onSPR multimode HCS optical fiber based on LDF
1.367 1.367 1.384 [19] 1.384
Wavelength [nm]
𝒏𝒔𝒎
[nm/RIU]
(∆nsm = 0.015 )
647Resonance 1592Sensitivity 701 Wavelength [nm]2831 [nm/RIU] 715 2043 647 1592 755 3554
SNR 0.404 0.254 (∆𝒏𝒔𝒎 = 𝟎. 𝟎𝟏𝟓 ) 0.924 0.404 0.276
1.367 SPR based on LDF 1.367 701 2831 0.254 SPR sensor based on multimode HCS optical fiber [19] 1.384 715 2043 0.924 Table 1 shows that on theLDF sensitivity of the LDF 1.384 sensor is better sensor; vice versa, SPR based 755than the HCS3554 0.276
the SNR is worse. We emphasize that the HCS sensor chosen for comparison had a core diameter (300 µm)1 larger (170 µm), general, the performance of SPR sensors Table showsthan thatour theLDF sensitivity of and the that, LDF in sensor is better than the HCS sensor; vicebased versa, the
SNR is worse. We emphasize that the HCS sensor chosen for comparison had a core diameter (300 µm) larger than our LDF (170 µm), and that, in general, the performance of SPR sensors based on multimode fibers improve with the diameter of the fiber [20]. We argue that the LDF sensor has a higher sensitivity compared to the HCS due to a higher number of modes internally excited by the
Appl. Sci. 2018, 8, 1172
7 of 8
on multimode fibers improve with the diameter of the fiber [20]. We argue that the LDF sensor has a higher sensitivity compared to the HCS due to a higher number of modes internally excited by the nanovoids. Similarly, the increased modal content may also explain the higher FWHM of the SPR response. However, a more extensive analysis of the modal distribution inside the LDF must be carried out in order to fully explain the obtained results. 4. Conclusions An LDF has been proposed as host material for SPR-based refractive index sensing. The realized SPR sensor shows a high sensitivity to the external refractive index (about 4000 nm/RIU for an SRI of 1.394). A tentative model has been developed in terms of power distribution within the fiber, which has been shown to accurately describe the changes of transmitted power along an unclad LDF. The proposed sensor does not require any specific processing of the fiber, apart from a metallization of the probe region, and exhibits a better sensitivity than similar SPR sensors based on a multimode optical fiber. In contrast, the SNR is lower than the HCS fiber, and much lower than single-mode fiber SPR sensors. However, we believe that the observed SNR is still sufficient to allow biochemical applications as demonstrated in the literature with other multimode SPR configurations. The sensor is easily repeatable, disposable, free of labeling, and has a potential for real-time detection, so it is ideally suited for biochemical sensing and environmental monitoring. Author Contributions: N.C., L.Z. and A.M. conceived, designed and realized the SPR sensor device. All the authors realized the experimental measurements, analyzed the results and contributed to the writing of paper. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflicts of interest.
References 1. 2. 3. 4. 5.
6. 7. 8.
9. 10. 11. 12.
Homola, J. Present and future of surface plasmon resonance biosensors. Anal. Bioanal. Chem. 2003, 377, 528–539. [CrossRef] [PubMed] Homola, J.; Yee, S.S.; Gauglitz, G. Surface plasmon resonance sensors: Review. Sens. Actuators B Chem. 1999, 54, 3–15. [CrossRef] Liu, Y.; Liu, Q.; Chen, S.; Cheng, F.; Wang, H.; Peng, W. Surface Plasmon Resonance Biosensor Based on Smart Phone Platforms. Sci. Rep. 2015, 5, 12864. [CrossRef] [PubMed] Caucheteur, C.; Guo, T.; Albert, J. Review of plasmonic fiber optic biochemical sensors: Improving the limit of detection. Anal. Bioanal. Chem. 2015, 407, 3883–3897. [CrossRef] [PubMed] Klantsataya, E.; Jia, P.; Ebendorff-Heidepriem, H.; Monro, T.M.; François, A. Plasmonic fiber optic refractometric sensors: From conventional architectures to recent design trends. Sensors 2017, 17, 12. [CrossRef] [PubMed] Anuj, K.; Sharma, R.J.; Gupta, B.D. Fiber-optic sensors based on surface Plasmon resonance: A comprehensive review. IEEE Sens. J. 2007, 7, 1118–1129. Trouillet, A.; Ronot-Trioli, C.; Veillas, C.; Gagnaire, H. Chemical sensing by surface plasmon resonance in a multimode optical fiber. Pure Appl. Opt. 1996, 5, 227–237. [CrossRef] Pollet, J.; Delport, F.; Janssena, K.P.F.; Jans, K.; Maes, G.; Pfeiffer, H.; Wevers, M.; Lammertyn, J. Fiber optic SPR biosensing of DNA hybridization and DNA–protein interactions. Biosens. Bioelectron. 2009, 25, 864–869. [CrossRef] [PubMed] Masson, J.F.; Kim, Y.C.; Obando, L.A.; Peng, W.; Booksh, K.S. Fiber-optic surface plasmon resonance sensors in the near-infrared spectral region. Appl. Spectrosc. 2006, 60, 1241–1246. [CrossRef] [PubMed] Slavik, R.; Homola, J.; Ctyroky, J. Single-mode optical fiber surface plasmon resonance sensor. Sens Actuators B Chem. 1999, 54, 74–79. [CrossRef] Jorgenson, R.C.; Yee, S.S. A fiber-optic chemical sensor based on surface plasmon resonance. Sens. Actuators B Chem. 1993, 12, 213–220. [CrossRef] Cennamo, N.; Massarotti, D.; Conte, L.; Zeni, L. Low Cost Sensors Based on SPR in a Plastic Optical Fiber for Biosensor Implementation. Sensors 2011, 11, 11752–11760. [CrossRef] [PubMed]
Appl. Sci. 2018, 8, 1172
13.
14.
15.
16.
17. 18.
19. 20. 21. 22. 23. 24.
8 of 8
Liang, G.; Luo, Z.; Liu, K.; Wang, Y.; Dai, J.; Duan, Y. Fiber Optic Surface Plasmon Resonance–Based Biosensor Technique: Fabrication, Advancement, and Application. Crit. Rev. Anal. Chem. 2016, 46, 213–223. [CrossRef] [PubMed] Galatus, R.; Farago, P.; Cennamo, N.; Cristea, C.; Feier, B. SPR based hybrid electro-optic biosensor platform: SPR-cell with side emitting plastic optical fiber. In Proceedings of the 2017 IEEE 23rd International Symposium for Design and Technology in Electronic Packaging (SIITME), Constanta, Romania, 26–29 October 2017; pp. 328–331. Logunov, S.; Fewkes, E.; Shustack, P.; Wagner, F. Light Diffusing Optical Fiber for Illumination. In Renewable Energy and the Environment; Paper DT3E.4; OSA Technical Digest (online), Optical Society of America: Washington, DC, USA, 2013. Apriyanto, H.; Ravet, G.; Bernal, O.D.; Cattoen, M.; Seat, H.C.; Chavagnac, V.; Surre, F.; Sharp, J.M. Comprehensive Modeling of Multimode Fiber Sensors for Refractive Index Measurement and Experimental Validation. Sci. Rep. 2018, 8, 5912. [CrossRef] [PubMed] Sharma, A.K.; Gupta, B.D. Absorption-based fiber optic surface plasmon resonance sensor: A theoretical evaluation. Sens. Actuators B Chem. 2004, 100, 423–431. [CrossRef] Sequeira, F.; Duarte, D.; Bilro, L.; Rudnitskaya, A.; Pesavento, M.; Zeni, L.; Cennamo, N. Refractive Index Sensing with D-Shaped Plastic Optical Fibers for Chemical and Biochemical Applications. Sensors 2016, 16, 2119. [CrossRef] [PubMed] Kanso, M.; Cuenot, S.; Louarn, G. Sensitivity of optical fiber sensor based on surface plasmon resonance: Modeling and experiments. Plasmonics 2008, 3, 49. [CrossRef] Dwivedi, Y.S.; Sharma, A.K.; Gupta, B.D. Influence of design parameters on the performance of a surface plasmon sensor based fiber optic sensor. Plasmonics 2008, 3, 79. [CrossRef] Iga, M.; Seki, A.; Watanabe, K. Gold thickness dependence of SPR-based hetero-core structured optical fiber sensor. Sens. Actuators B Chem. 2005, 106, 363. [CrossRef] Sharma, A.K.; Gupta, B.D. On the sensitivity and signal to noise ratio of a step-index fiber optic surface plasmon resonance sensor with bimetallic layers. Opt. Commun. 2005, 245, 159–169. [CrossRef] Yogendra, S.D.; Sharma, A.K.; Gupta, B.D. Influence of skew rays on the sensitivity and signal-to-noise ratio of a fiber-optic surface-plasmon-resonance sensor: A theoretical study. Appl. Opt. 2007, 46, 4563–4569. Gupta, B.D.; Kant, R. Recent advances in surface plasmon resonance based fiber optic chemical and biosensors utilizing bulk and nanostructures. Opt. Laser Technol. 2018, 101, 144–161. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
