Electrospun nanofiber mats for evanescent optical fiber sensors

Accepted Manuscript Title: Electrospun nanofiber mats for evanescent optical fiber sensors Author: Aitor Urrutia Javier Goicoechea Pedro J. Rivero Ignacio R. Mat´ıas Francisco J. Arregui PII: DOI: Reference:

S0925-4005(12)01033-7 doi:10.1016/j.snb.2012.10.009 SNB 14642

To appear in:

Sensors and Actuators B

Received date: Revised date: Accepted date:

2-8-2012 1-10-2012 2-10-2012

Please cite this article as: A. Urrutia, J. Goicoechea, P.J. Rivero, I.R. Mat´ias, F.J. Arregui, Electrospun nanofiber mats for evanescent optical fiber sensors, Sensors and Actuators B: Chemical (2010), doi:10.1016/j.snb.2012.10.009 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Electrospun nanofiber mats for evanescent optical fiber

ip t

sensors

cr

Aitor Urrutia*, Javier Goicoechea, Pedro J. Rivero, Ignacio R. Matías, and Francisco J. Arregui

us

Nanostructured Optical Devices Laboratory, Electric and Electronic Engineering Department, Public University of Navarra, Edif. Los Tejos, Campus Arrosadía S/N, 31006, Pamplona, Spain.

an

*[email protected]

M

ABSTRACT

d

In this work, a study about the optical response of Electrospun Nanofiber (ENF) coatings for their use

te

in evanescent optical fiber sensors is presented. Several types of ENF mats composed of poly(acrylic acid) (PAA) were developed with different ENF diameters and densities. These ENF mats were

Ac ce p

deposited onto an optical fiber core in order to fabricate humidity evanescent optical fiber sensors. The devices were exposed to Relative Humidity (RH) variations from 30% RH to 95%RH. The transfer functions of the devices (transmitted optical power versus relative humidity) presented two welldifferenced behaviors depending on the ENF diameter and the ENF mat density. The devices with lower ENF diameters and higher mat density showed an increase in the transmitted optical power when RH increased. On the contrary, the devices with higher ENF diameters and lower mat density showed a decrease in the transmitted optical power when RH increased. In addition to this, sensors with thinner ENF overlays, showed a higher sensitivity. In order to study the response time of these devices, the ENFs sensors were submitted to human breathing cycles and presented a response time around 340 ms 1 Page 1 of 23

(exhalation). In spite of the high RH conditions of this experiment, the devices showed a recovery time around 210 ms and a negligible hysteresis or drift with respect to the initial condition (inhalation).

ip t

KEYWORDS

cr

Electrospinning, swelling, humidity sensing, evanescent optical fiber sensors, optical response

us

1. INTRODUCTION

Recently, new nanostructured materials with extremely high specific surface area have been

an

developed thanks to the use of nanotechnology. With respect to the bulk materials, these new materials

M

can enjoy better characteristics for sensing such as a higher sensitivity or a shorter response time. The combination of different techniques such as Layer-by-Layer [4], PVD [5], sol-gel [6] or sputtering [7]

te

and morphology of the films.

d

make possible the tuning of the overall properties of the coating by properly designing the composition

A good alternative to fabricate materials with a high specific surface area is the electrospinning

Ac ce p

method. By means of this technique, continuous polymeric nanofibers can be fabricated by applying a high voltage field to a polymeric solution. Schematically, when such high voltage is applied to the polymeric solution with respect to a grounded electrode, the polymeric solution becomes charged and, due to electrostatic attraction, a stream of liquid erupts from the surface and is launched towards the grounded electrode where the substrate to be coated has been previously placed. If the molecular cohesion of the solution is high enough, a continuous thin fiber is formed and elongated following the electrical field direction. Simultaneously, the solvent is evaporated during the flight of the fibers and, finally, this electrospun nanofibers (ENF) are deposited on the grounded screen (substrate). The thinning of the fiber leads to the fabrication of filaments up to several tens of nanometers in diameter that superpose forming a fiber web [8]. These fibrous membranes obtained using the electrospinning 2 Page 2 of 23

technique have a strong potential application in sensing and other fields such as in the synthesis of biocompatible or antimicrobial surfaces [9]. Such fiber mats may have a specific surface area of approximately one to two orders of magnitude larger than the specific surface area of continuous flat films [10], boosting the efficiency of the desired device. In fact, several recent sensors have been

ip t

reported creating polymeric electrospun membranes to sense RH [11-13], pH [14], ethanol [15], ammonia [16] and other gases [17, 18]. These sensors are based on quartz crystal microbalances,

cr

interdigital electrodes, a three-electrode system or a spectrofluorometer.

us

Very recently, a new approach based on the ENF coating of a hollow core optical fiber has been reported for RH sensing [19]. The sensing characteristics of these optical fiber sensors depend

an

dramatically on the ENF diameter and density of the ENF mat. For this reason, a preliminary experimental study of the optical response of the ENFs for evanescent optical fiber sensors has a

M

relevant importance to improve the final characteristics of the new sensor devices. In this work, the formation of ENF mats with different ENF diameters and mat densities were

d

performed based on poly(acrylic acid (PAA). The aim of this work is the experimental demonstration

te

for the first time that ENF membranes can be used as sensing coatings in optical fiber sensors. The

Ac ce p

hydrogel-nature of the PAA ENFs, and therefore, their swelling behavior with the humidity makes possible a notable response to RH variations. According to the ENFs characteristics, these variations were described and analyzed. In addition, this ENF-optical fiber configuration has been successfully used for human breathing monitoring. To our knowledge this is the first work in which PAA ENF mats have been experimentally studied for their application in optical fiber humidity sensors.

2. MATERIAL AND METHODS 2.1.- Materials The polymer poly(acrylic acid), PAA (Mw≈100.000 sol. 35%wt. in water), was provided from Sigma Aldrich and was used without any further purification. Plastic clad-silica (PCS) fiber with core and 3 Page 3 of 23

cladding diameters of 200 and 230 μm respectively was purchased by OFS (Specialty Photonics Division).

2.2.- Electrospinning of the nanofibers and sensor fabrication Firstly, PCS fiber was cleaved (with a Fujimura cleaver) obtaining a 3 cm length section. Then, the

ip t

cladding was chemically removed from this section. This optical fiber core fragment is then used as the substrate in the electrospinning deposition process. The polymeric solution was composed of only PAA,

cr

35 wt. % in water. This solution was immediately used for the ENFs deposition. The solution flux was

us

fixed to 1 mL/hour using a syringe pump. The needle acted as the anode and an aluminum screen as the cathode. A distance of 22 cm was fixed between them, and a voltage of 18.5 kV was applied. The fiber

an

fragment was placed 2 cm above the aluminum cathode perpendicularly to the syringe direction, in order to collect the ENFs. A more detailed description of the electrospinning technique can be found

M

elsewhere [20, 21]. In this particular arrangement, the optical fiber fragment was spun at 30 revolutions

d

per minute during the electrospinning process to achieve a uniform ENF-coating throughout all the

Ac ce p

te

sensing section of the optical fiber. An illustrative scheme of this set-up is shown in Figure 1.

4 Page 4 of 23

Figure 1. Scheme of the electrospinning method. The fiber fragment is placed 2 cm above the aluminum cathode perpendicularly to the syringe direction, in order to collect the ENFs. Once the segment was coated with the ENFs, it was submitted to a thermal treatment of 150°C for 4 hours. Finally, the segment was spliced to optical fiber pigtails at both extremes so as to monitor the

ip t

transmission spectra of the device.

This fabrication process was repeated to create several devices with different types of ENFs: with

cr

different ENFs diameters (varying the PAA viscosity) and with different mat densities (varying the

us

deposition time). The rest of all fabrication process parameters were fixed to achieve a good

an

repeatability in each device.

M

2.3.- ENFs characterization

The ENFs were morphologically characterized using optical microscopy (Leica DM2500P). Optical

d

properties of the fibrous coatings were measured by UV-VIS spectrometry (Jasco V-630).

te

The experimental setup used for testing the optical fiber devices is shown in Figure 2. It consists of a

Ac ce p

white light source connected to one of the optical fiber pigtails. The other one is connected to a UV-VIS spectrometer (USB2000-FLG from Ocean Optics). The use of this spectrometer makes possible to cover a wavelength range from 380 to 1000 nm.

5 Page 5 of 23

ip t cr us

M

3. RESULTS AND DISCUSSION

an

Figure 2. Experimental setup to monitor the optical sensor response.

3.1.- Electrospun fiber mat characterization

d

Once the deposition process is finished, the ENFs were characterized using optical microscopy. In

te

Figure 3, it is possible to see the morphology of ENF coating onto the optical fiber core. As expected,

Ac ce p

the ENFs are mainly placed perpendicularly to the guided light fiber direction. This ENF arrangement on the optical fiber core can be explained by the spinning movement of the optical fiber segment during the electrospinning process, which make the ENFs wrap the optical fiber core.

6 Page 6 of 23

an

us

cr

ip t

100 μm

Figure 3. Optical fiber core coated with ENFs.

M

In order to characterize and analyze the ENFs diameter and density, several deposition processes were performed with the conditions indicated in the Experimental Section using glass slides as substrates.

d

Two batches of PAA with different viscosities (47 and 49 cps. respectively) were used so as to obtain

te

different ENF diameters. As PAA solutions were used without further manipulation it is possible to

Ac ce p

keep constant the two PAA viscosities, thus achieving repeatability in the ENFs. Samples fabricated with the first batch were denoted as samples A, and samples composed of more viscous PAA were denoted as samples B. Furthermore three deposition times were applied to achieve different ENF mat densities: 5 minutes (denoted as samples 1), 15 minutes (samples 2) and 30 minutes (samples 3). Experimentally, the ENF density has been estimated using the absorbance of the ENF coatings in the visible region. All samples showed a white visual aspect, and their transparency were inversely proportional to the density and thickness of the electrospun polymeric webs. The optical micrographs of the fabricated samples of the ENF depositions are shown in Figure 4, and ENF diameter values and absorbance values are presented in Figure 5. The average fiber diameter results in samples A were 1.61 ± 0.28 µm, 1.54 ± 0.32 µm, and 1.51 ± 0.43 µm respectively, taking the 7 Page 7 of 23

error as the standard deviation. On the other hand, the average fiber diameter in samples 1B, 2B and 3B were 0.59 ± 0.15 µm, 0.48 ± 0.09 µm and 0.53 ± 0.18 µm respectively. This demonstrates that a tiny viscosity variation (from 47 to 49 cps) in the electrospinning solution affects dramatically to the ENF

te

d

M

an

us

cr

ip t

diameter.

Figure 4. Optical micrographs of the fabricated samples. Thicker ENFs correspond to samples A and

Ac ce p

thinner ENFs correspond to samples B.

8 Page 8 of 23

ip t cr us an M

d

Figure 5. UV-VIS absorption values at 450 nm of the ENFs in all samples. Inset: Summary table with

te

the sample codes and the diameter measurements.

Ac ce p

Figure 5 also shows the absorbance values of all samples obtained by UV-VIS spectroscopy at λ = 450 nm. ENF diameters are also presented in the inset table of Figure 5. It is clearly appreciable how ENF deposition time determines the electrospun mat thickness, giving higher UV-VIS absorbance values. It is important to remark than an increment in time of the deposition process increases the ENF mat thickness, while the ENF diameter is kept constant. On the other hand, it possible to see how thinner ENF diameters also yield higher absorbance values because the polymeric fibers form a denser mat.

3.2.- ENF coated optical fiber devices In order to study the optical response of the ENF coated optical fiber devices, a programmable climatic chamber was used to expose the sensors to different variations and cycles of RH at a constant 9 Page 9 of 23

temperature of 25ºC. Cycles consisted on RH ramps with a constant slope from 30% to 95%. Then, the absorbance spectra were captured taking as a reference the optical signal from each sensor at 30%RH. In Figure 6 the relative absorbance spectra of the sensor 1B are presented (device with a denser ENF mat). The results showed that the relative absorbance spectrum is decreased as the RH increased. This

ip t

means that a higher amount of optical power is transmitted through the optical fiber at higher RH environments, giving a lower absorbance value. Contrary to that, the response of the sensor 1A, device

cr

with a lower ENF mat density, showed an increment of the absorbance when the RH grew, see Figure 7.

us

In this case, this means that in high RH environments, the optical losses of the ENF coated region are increased and therefore it contributes to a reduction of the overall transmitted optical power. The more

an

pronounced variation in optical power transmission is observed around the 700 – 800 nm range in both cases.

M

251658240

60%RH

70%RH

te Ac ce p

Absorbance (a.u.)

d

50%RH

80%RH 90%RH

Wavelength (nm) Figure 6. Relative absorbance spectra variations at different RH values (optical response, sensor 1B).

10 Page 10 of 23

90%RH

60%RH

70%RH

an

50%RH

us

cr

ip t

Absorbance (a.u.)

80%RH

Wavelength (nm)

M

Figure 7. Relative absorbance spectra variations at different RH values (optical response, sensor 1A).

d

3.3.- Influence of the ENF mat density

te

Relative humidity tests were performed for the other sensor types (2A, 3A, 2B and 3B). These results

Ac ce p

were coherent with the optical characteristics shown in figures 6 and 7. The swelling/deswelling behaviour of the PAA ENFs in presence of RH causes important changes in the evanescent field and therefore in the transmitted optical power. However the optical responses from sensors 3A and 3B showed significantly higher hysteresis and drift values than the 1A, 1B, 2A and 2B sensors. It is known that higher thickness ENF mats presents more difficulties to trap and release the environmental humidity [22] yielding a slow and hysteretic optical response. In addition, it has been observed by optical microscopy that sensors 3A and 3B had a significant amount of loose electrospun polymeric fibers which are not well fixed to the mat. These loose fibers could induce irreversible morphological alterations in the ENF optical fiber overlay, giving high optical drift values. In consequence, the optical response from sensors 3A and 3B was not taken in 11 Page 11 of 23

consideration for the analysis. In Figure 8 it is shown the transfer functions calculated at 750 nm of type 1 and 2 sensors. In all cases, the sensors have two well different operating regions. For humidity values from 30% to 50% the sensor response in absorbance was almost negligible (around 0.01dB/%RH for the 1B). In contrast, for the RH

ip t

range between 70% and 95%, the optical response is significantly more sensitive to RH (up to 0.11dB/%RH for the 1A).

Ac ce p

te

d

M

an

us

cr

251658240

Figure 8. Transfer function of the devices 1A, 2A, 1B and 2B for a 50%-95% RH range at a wavelength of 750 nm. The first remarkable aspect of the above mentioned transfer functions, is the difference in the slope of the curve, depending on the sensor type (A and B). As it has been experimentally demonstrated with sensors 1A and 1B, the slope sign of the optical response to relative humidity variations depend on the 12 Page 12 of 23

density of the ENF (Figures 6 and 7). Type A sensors have thicker fibers but lower mat densities, and their optical losses are directly related with RH increments. Oppositely, type B sensors have higher mat density, and their optical losses are inversely related with RH. The classical evanescent optical fiber sensors typically show an increment of the optical absorbance as

ip t

the effective refractive index of the sensing coating is increased [23, 24]. This happens when the sensing coating has a lower refractive index than the optical fiber core [23] and it is consistent with the

cr

behaviour observed in type A sensors, where the transmitted optical power decreases as the RH is

us

increased. Nevertheless the type B sensors show the opposite optical response as the RH is increased. This apparently anomalous behaviour is discussed in the following paragraphs.

an

CoreRI has been estimated in 1.463 at ! =450nm, and 1.448 at ! =1000nm using the well-known Sellmeier equation [25], and PAA solution RI was measured by a refractometer (model 320 ABBE,

M

Zuzi), thus estimating for pure PAA a RI of 1.50, which is consistent with other values measured by ellipsometry found in literature [26]. Therefore the RI of the pure polymer is higher than the optical

d

fiber core one. Nevertheless it is very important to remark that the overall RI of the ENF coating does

te

not depend only on the RI of the electrospun fiber material, but also in the fiber diameter, distribution,

Ac ce p

packaging, etc [27]. For this reason, in this particular optical configuration of the optical fiber sensor, the External Effective Refractive Index (EERI) is composed by the ENF mat, water and air gap contributions.

According to the evanescent field theory [23] when EERI is lower than the optical fiber core refractive index (coreRI), small variations of the EERI may cause an increment of the evanescent optical losses, that depend on the difference between EERI and the coreRI. For high RH values, ENFs swell and increase their volume and thus this ENF swelling causes the suppression of air gaps and leads to increase the EERI. As it has been previously mentioned, the evanescent losses observed in type A sensors fit well with the evanescent field theory when EERI < coreRI. Nevertheless Figure 8 shows how type B sensors 13 Page 13 of 23

had the opposite behaviour; they show lower absorbance values (more transmitted light) with higher RH values. The optical losses dependence with EERI variations is summarized in Table 1, for the cases in which EERI > coreRI and EERI < coreRI.

higher than the optical fiber’s core and the other with a lower one [28].

ip t

Table 1. Schematic dependence of the optical losses with the EERI variation for two different refractive index coatings, one

Optical losses

EERI lower than optical fiber core

Increase

Increase

EERI higher than optical fiber core

Increase

cr

External Effective Refractive Index (EERI)

us

Decrease

The experimental results showed in Figure 8 demonstrated that the sign of the slope of the sensor

an

transfer function depend on the ENF density rather than the ENF thickness, as far as type A sensors show positive optical loss slopes with RH increasing, while type B sensors have negative optical loss

M

slopes with RH. This optical response slope dependence is summarized in Table 2.

d

Table 2. Experimental dependence found in the assayed sensors with relative humidity variations for the two ENF densities

te

(type A and B sensors). These trends have been extracted from Figure 8. Relative Humidity (%)

Absorbance

Lower (Type A sensors)

Increase

Increase

Higher (Type B sensors)

Increase

Decrease

Ac ce p

Fiber mat density

According to the theoretical and experimental information displayed in Tables 1 and 2, it is possible to relate the A and B type sensors with their respective EERI (lower or higher than the coreRI). In type A sensors (thicker diameter ENF), the polymeric fibers are less compacted onto the optical fiber core and the overall overlay morphology show more air gaps. These air gaps contribute to lower the EERI, and as a result it is slightly lower than coreRI. In these cases, the transmitted optical power decreases with a small EERI increase caused by the RH [28, 29]. The authors’ hypothesis is that in type B sensors the higher density of the ENF confers an EERI slightly higher than the coreRI. In such type B mats, the diameter of the ENFs is thinner, and hence more densely packaged. These ENF overlays have less air

14

Page 14 of 23

gaps and, consequently the EERI is slightly higher than coreRI. Therefore when RH raises the difference between EERI and coreRI is higher, and consequently the transmitted optical power increases [28].

3.4.- Influence of the ENF mat thickness

ip t

Other significant aspect shown in Figure 8 is that the thickness of the overlays dramatically affects to the sensitivity of the sensors. Thicker overlays devices (2A and 2B) lead to worse sensitivities than

cr

thinner mat sensors (1A and 1B). In Figure 8, it can be seen as the devices 1A and 1B have a higher

us

sensitivity and also a higher dynamic range than 2A and 2B respectively. This fact confirms that depending on the initial ENF thickness the RH-induced ENF variations affects differently to the

an

evanescent losses, and consequently to the optical response of the device.

As far as the evancescent field around the optical fiber core has a well defined penetration depth [23],

M

the EERI within this region is what define the evanescent losses of the optical sensor. In this way, the

d

RH-induced optical variations of the ENF must be confined in this region to impact in the sensor optical

sensor response.

te

signal. All optical medium variations outside this evanescent penetration region will not affect to the

Ac ce p

In sensors type 1 (thinner ENF mat), the fibrous mats experiment important RH-induced thickness variations within the evanescent field sensitive region. Consequently these ENF morphology changes induce higher impacts in the evanescent optical losses for the same RH range as in sensors 2 and 3. In these thicker coatings, although the ENF morphology change could be greater than in type 1 sensors, it may have a lower relative impact in the EERI, and thus in their evanescent losses.

3.3.- Dynamic response Regarding to the dynamic response, all the sensors were submitted to the same climatic chamber assays. Sensors 1A, 2A, 1B and 2B followed almost the same dynamic response than the climatic chamber electronic RH sensor, with low hysteresis and drift. As it has been previously mentioned, 15 Page 15 of 23

sensors 3A and 3B showed a very hysteretic behavior, with high drift values, and therefore they were not studied in the present section. One of the most exigent dynamic applications for RH sensors is a real human breathing test. The ENF sensors were submitted to several exhalation and inhalation cycles in order to demonstrate their

ip t

potential in high performance such sensing application. All sensors exhibited optical responses when

cr

they were exposed to several human breathing cycles.

us

2.5 2.0 1.5

an M

95%

1.5

338ms

d

1.0

1.0 0.5 0.0

0

5

10

15

20

25

30

7.0

7.5

Time (s)

209ms

0.0

0.5

1.0

1.5

2.0

5%

te

0.5

0.0

Transmitted Power (dB)

2.0

Ac ce p

Transmitted Optical Power (dB)

2.5

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Time (s)

Figure 9. Monitoring of the absorbance spectra to several human breathings (sensor 1B). The best results were obtained for 1B sensor. These 1B ENFs had the thinner diameter fibers and they also had the higher fiber density. These characteristics probably confers to them a higher specific surface to interact with the surrounding medium, and their thin diameter minimizes the environmental water trap/release times, giving a very quick response. The results showed a fast average time response of 300 ms, drift values close to zero and a high repeatability (see Fig. 9). More in detail, the response 16 Page 16 of 23

times was around 340 ms (exhalation) and 210 ms (recovery time). This fast behavior of the sensors against RH makes them excellent for human breathing monitoring. The measurements showed a difference of 2 dB on average between the ambient condition and the exhalation periods. Figure 9 represents the breathing test for sensor 1B in the 730-780 nm range.

ip t

4. CONCLUSIONS

cr

In this work, a novel study of the influence of ENF characteristics in the optical response of new

us

evanescent fiber-optic sensors is presented. PAA was electrospun to develop several types of ENFs with different densities (types A and B) and thicknesses (types 1, 2 and 3). As more viscose is the PAA

an

solution, ENF diameter is lower, and the resultant ENF mat density is higher. ENFs were characterized and used to create several humidity fiber-optic sensors, and their optical response was analyzed when

M

they were exposed to RH variations between 30%RH and 95%RH. It was found that ENF density was a key factor to determine the optical sensor behavior. Low density ENF sensors showed increments in the

d

optical losses as the RH grows, while higher density ENF sensors decreases their losses in the same

te

conditions. It has been demonstrated that this effect is related with the value of effective refractive index

Ac ce p

of the surrounding sensitive medium respect to the refractive index of the optical fiber core. The impact of the ENF thickness was also studied and it was found that devices with low ENF mat thickness obtained higher sensitivities and lower hysteresis values. The ENF sensors showed fast responses (300ms) thanks to the large specific surface area of the ENFs making possible their use in applications such as human breathing monitoring.

The combination of the ENF swelling property and the characteristic EERI of the devices explain the different optical responses and their transfer functions reported in this work, clearly showing the great importance of the ENF diameter and mat density in the evanescent field response for the optical fiber sensors created by electrospinning. The technical aspects enclosed in this work make possible the optimization of ENF membranes for further optical sensing approaches by simply tuning the ENF 17 Page 17 of 23

thickness and density, opening new possibilities for the fabrication of diverse sensors, not only for RH sensing but for many other variables such as pH, gases or other target substances. To our knowledge this is the first time that PAA ENF mats have been experimentally studied for its application in optical fiber sensors.

ip t

ACKNOWLEDGEMENTS

cr

This work was supported in part by the Spanish Ministry of Education and Science- CICYT-FEDER

us

TEC2010-17805 Research Grant, and a UPNA pre-doctoral research grant.

an

REFERENCES

[1] B. Culshaw, A. Kersey, Fiber-optic sensing: A historical perspective, J. Lightwave Technol. 26

M

(2008) 1064-1078.

d

[2] R. Jindal, S. Tao, J.P. Singh, P.S. Gaikwad, High dynamic range fiber optic relative humidity

te

sensor, Optical Engineering. 41 (2002) 1093-1096.

Ac ce p

[3] F.J. Arregui, Y. Liu, I.R. Matias, R.O. Claus, Optical fiber humidity sensor using a nano FabryPerot cavity formed by the ionic self-assembly method, Sensors and Actuators, B: Chemical. 59 (1999) 54-59.

[4] X. Zhang, H. Chen, H. Zhang, Layer-by-layer assembly: From conventional to unconventional methods, Chemical Communications. (2007) 1395-1405. [5] D.M. Mattox, Physical vapor deposition processes, Prod Finish (Cincinnati). 65 (2001) 72. [6] L.L. Hench, J.K. West, The Sol-Gel process, Chem. Rev. 90 (1990) 33-72. [7] M.Y. Ali, W. Hung, F. Yongqi, A review of focused ion beam sputtering, International Journal of 18 Page 18 of 23

Precision Engineering and Manufacturing. 11 (2010) 157-170. [8] J. Doshi, D.H. Reneker, Electrospinning process and applications of electrospun fibers, J. Electrostatics. 35 (1995) 151-160.

ip t

[9] P.J. Rivero, A. Urrutia, J. Goicoechea, Y. Rodríguez, J.M. Corres, F.J. Arregui, I.R. Matías, An antibacterial submicron fiber mat with in situ synthesized silver nanoparticles, J Appl Polym Sci.

cr

(2012).

us

[10] X. Wang, C. Drew, S.-. Lee, K.J. Senecal, J. Kumar, L.A. Samuelson, Electrospun Nanofibrous Membranes for Highly Sensitive Optical Sensors, Nano Letters. 2 (2002) 1273-1275.

an

[11] X. Wang, B. Ding, J. Yu, M. Wang, F. Pan, A highly sensitive humidity sensor based on a

M

nanofibrous membrane coated quartz crystal microbalance, Nanotechnology. 21 (2010). [12] W. Wang, Z. Li, L. Liu, H. Zhang, W. Zheng, Y. Wang, H. Huang, Z. Wang, C. Wang, Humidity

te

(2009) 404-409.

d

sensor based on LiCl-doped ZnO electrospun nanofibers, Sensors and Actuators, B: Chemical. 141

Ac ce p

[13] D. Aussawasathien, J.-. Dong, L. Dai, Electrospun polymer nanofiber sensors, Synth. Met. 154 (2005) 37-40.

[14] C.-. Kuo, Y.-. Tung, W.-. Chen, Morphology and pH sensing characteristics of new luminescent electrospun fibers prepared from poly(phenylquinoline)-block-polystyrene/ polystyrene blends, Macromolecular Rapid Communications. 31 (2010) 65-70. [15] X. Song, Z. Wang, Y. Liu, C. Wang, L. Li, A highly sensitive ethanol sensor based on mesoporous ZnO-SnO2 nanofibers, Nanotechnology. 20 (2009). [16] B. Ding, J. Kim, Y. Miyazaki, S. Shiratori, Electrospun nanofibrous membranes coated quartz 19 Page 19 of 23

crystal microbalance as gas sensor for NH3 detection, Sensors and Actuators, B: Chemical. 101 (2004) 373-380. [17] W. Zheng, X. Lu, W. Wang, Z. Li, H. Zhang, Z. Wang, X. Xu, S. Li, C. Wang, Assembly of Pt nanoparticles on electrospun In2O3 nanofibers for H2S detection, J. Colloid Interface Sci. 338 (2009)

ip t

366-370.

cr

[18] J.-. Park, J. Moon, S.-. Lee, S.H. Kim, H.Y. Chu, T. Zyung, SnO2-ZnO hybrid nanofibers-based

us

highly sensitive nitrogen dioxides sensor, Sensors and Actuators, B: Chemical. 145 (2010) 592-595. [19] J.M. Corres, Y.R. Garcia, F.J. Arregui, I.R. Matias, Optical fiber humidity sensors using PVdF

an

electrospun nanowebs, IEEE Sensors Journal. 11 (2011) 2383-2387.

M

[20] Z.-. Huang, Y.-. Zhang, M. Kotaki, S. Ramakrishna, A review on polymer nanofibers by electrospinning and their applications in nanocomposites, Composites Sci. Technol. 63 (2003) 2223-

d

2253.

te

[21] D.H. Reneker, I. Chun, Nanometre diameter fibres of polymer, produced by electrospinning,

Ac ce p

Nanotechnology. 7 (1996) 216-223.

[22] K.E. Secrist, A.J. Nolte, Humidity Swelling/Deswelling Hysteresis in a Polyelectrolyte Multilayer Film, Macromolecules. 44 (2011) 2859-2865. [23] V. Ruddy, B.D. MacCraith, J.A. Murphy, Evanescent wave absorption spectroscopy using multimode fibers, J. Appl. Phys. 67 (1990) 6070-6074. [24] L. Xu, J.C. Fanguy, K. Soni, S. Tao, Optical fiber humidity sensor based on evanescent-wave scattering, Opt. Lett. 29 (2004) 1191-1193. [25] C.R. Zamarreño, M. Hernáez, I. Del Villar, I.R. Matías, F.J. Arregui, Optical fiber pH sensor 20 Page 20 of 23

based on lossy-mode resonances by means of thin polymeric coatings, Sensors and Actuators, B: Chemical. 155 (2011) 290-297. [26] H.-. Su, W.-. Chen, Photosensitive high-refractive-index poly(acrylic acid)-graft-poly(ethylene glycol methacrylate) Nanocrystalline Titania hybrid films, Macromolecular Chemistry and Physics. 209

ip t

(2008) 1778-1786.

cr

[27] J.L. Davis, H.J. Walls, K.C. Mills, K. Guzan, R. Yaga, D.S. Ensor, Optical properties of electrospun nanofiber substrates, Materials Research Society Symposium Proceedings. 1240 (2010) 49-

us

54.

an

[28] Charles F. Cole, Jr., Alois J. Adams and Robert A. Sims, Optical fiber refractometer for

M

measuring the dispersion of turbid fluids, Proc. SPIE. 1796 (1993) 383-392. [29] A. Arie, R. Karoubi, Y.S. Gur, M. Tur, Measurement and analysis of light transmission through a

Ac ce p

FIGURE CAPTIONS

te

d

modified cladding optical fiber with applications to sensors, Appl. Opt. 25 (1986) 1754-1758.

Figure 1. Scheme of the electrospinning method. The fiber fragment is placed 2 cm above the aluminum cathode perpendicularly to the syringe direction, in order to collect the ENFs. Figure 2. Experimental setup to monitor the optical sensor response. Figure 3. Optical fiber core coated with ENFs. Figure 4. Optical micrographs of the fabricated samples. Thicker ENFs correspond to samples A and thinner ENFs correspond to samples B.

21 Page 21 of 23

Figure 5. UV-VIS absorption values at 450 nm of the ENFs in all samples. Inset: Summary table with the sample codes and the diameter measurements. Figure 6. Relative absorbance spectra variations at different RH values (optical response, sensor 1B).

ip t

Figure 7. Relative absorbance spectra variations at different RH values (optical response, sensor 1A). Figure 8. Transfer function of the devices 1A, 2A, 1B and 2B for a 50%-95% RH range at a wavelength

cr

of 750 nm.

an

us

Figure 9. Monitoring of the absorbance spectra to several human breathings (sensor 1B).

VITAE

M

Aitor Urrutia received the M.S. degree in telecommunications engineering and the M.S. degree in communications from the Public University of Navarra (UPNA), in 2009 and 2011, respectively.

d

Currently, he is working towards the PhD degree in the Communications Doctoral Program, Department

nanostructured materials.

te

of Electrical and Electronic Engineering, UPNA. His research interests include optical fiber sensors and

Ac ce p

Javier Goicoechea received his MS degree in electrical engineering and the PhD degree in 2003 and 2008 respectively from the Public University of Navarra (UPNA), Pamplona, Spain. In 2004 he obtained a predoctoral scholarship from the Spanish Ministry of Culture and Science. In 2010 he became reader at the Public University of Navarra. His research interest includes organic light emitting diodes, organic photovoltaic cells and fiber optic sensors. He has authored more than 21 scientific international publications and 4 book chapters (international edition). Pedro J. Rivero obtained his MS degree in Chemistry and Biochemistry in 2006 and 2007 respectively from the Catholic University of Navarra, Pamplona, Spain. In 2010, he graduated in Materials Engineering Postgraduate Master from the Public University of Navarra (UPNA). Nowadays, he is working towards his PhD degree in the Electrical and Electronic Engineering Department at the 22 Page 22 of 23

Public University of Navarra. His main areas of interest are the research and development of nanostructured functional thin films, optical fiber devices and their engineering applications. Ignacio R. Matías received the M.S. degree in electrical and electronic engineering and the Ph.D. degree in optical fiber sensors from the Polytechnic University of Madrid, Madrid, Spain, in 1992 and

ip t

1996, respectively. He became a Lecturer at the Public University of Navarra in 1996, where presently he is a Permanent Professor. He has coauthored more than 200 chapter books, journal and conference

cr

papers related to optical fiber sensors and passive optical devices and systems. He is an Associate Editor

us

of IEEE Sensors Journal.

Francisco J. Arregui (M’04) is a Full Professor at the Public University of Navarra, Pamplona,

an

Spain. He was part of the team that fabricated the first optical fiber sensor by means of the Layer-byLayer self-assembly method at Virginia Tech, Blacksburg, VA, in 1998. He is the author or coauthor of

M

around 250 scientific journal and conference publications, most of them related to optical fiber sensors based on nanostructured coatings. He is the editor of the book Sensors Based on Nanostructured

d

Materials, which has been recently published. He is an Associate Editor of IEEE Sensors Journal and

Ac ce p

of Sensors.

te

the International Journal on Smart Sensing and Intelligent Systems, as well as Editor-in-Chief of Journal

23 Page 23 of 23