fiber optic sensors for hydrocarbon detection

Sensors and Actuators B 105 (2005) 430–436

Fiber optic sensors for hydrocarbon detection

R. Falate, R.C. Kamikawachi, M. Muller, H.J. Kalinowski, J.L. Fabris∗

Centro Federal de Educa¸cao Tecnol´ogica do Paran´a, Av. Sete de Setembro, 3165, Curitiba – PR, CEP 80230-901, Brazil

Received 2 April 2004; received in revised form 21 June 2004; accepted 29 June 2004Available online 18 November 2004

Abstract

We report the use of optical sensors based on long period gratings written in standard telecommunication fibers. Applying electricalarc discharges from a fusion splicer produces the optical devices. Results from the hydrocarbon detection in fuel, water and atmosphericenvironment are shown. Wavelength shifts in the long period gratings attenuation peak as longer as 10.3, 2.6, 50.6, and 6.1 nm are obtainedwhen different concentrations of turpentine, naphtha, paint thinner, and anhydrous alcohol, respectively, are added to the commercial Braziliangasoline blend. The exposition of long period gratings to a mixture of butane and propane in air results in a wavelength shift of 0.6 nm inthe attenuation peak. In the experiments carried out in water environment, a 6 nm wavelength shift is measured in the hydrocarbon presence.T bon vaporsi©

K

1

isvcawths

dbgatv

n thehus,sol-gal

theare

dro-to

m be-dente, re-repair

ragg

peri-ber.ngesBGced

0d

hese results point to the prospect of using the long period gratings sensor for fuel quality control and for the detection of hydrocarn the air or petroleum pollutants in water environments.

2004 Elsevier B.V. All rights reserved.

eywords:Optical sensor; Hydrocarbon detection; Long period grating; Refractive index sensor; Water pollutants

. Introduction

Optical fiber sensors present many interesting character-stics, e.g., immunity to electromagnetic interference, highensitivity, spectral multiplexing[1], reduced weight, smallolume, high fusion temperature and low attenuation. Theseharacteristics make the optical fiber sensors safe and suit-ble to remote operation. An additional feature is importanthen the sensors are used in the hydrocarbon detection: as

he environment can be very inflammable, it is interesting toave a sensor device that does not produce sparks. Fiber opticensors fulfill all those requirements.

There is an increasing interest of the governments and in-ustries in the development of methods and equipments toe used in the fuel quality control. In Brazil the mixture ofasoline with solvents has legal validity for an anhydrouslcohol proportion between 20 and 25%, which depends on

he Brazilian alcohol production. Nevertheless, as some sol-ents have lower prices (about 60%) than gasoline, a common

∗ Corresponding author. Tel.: +55 41 3104642; fax: +55 41 3104683.

malpractice is to increase the solvents concentrations icommercial gasoline that will be sold to car owners. Tprocedures to verify the content of alcohol and differentvents in the fuel mixture[2] are required to assure the lelimit and to protect the consumer’s rights.

The optical sensors can also play an important role inenvironment control. Refineries, pipelines and harborsthe main places where accidents take place involving hycarbon petroleum spills[3]. These places are often closerivers, swamps, oceans and dams increasing the problecause of the high water flow. The proportions of the accican be reduced if the leakage is discovered in a short timducing the damage to the environment and the costs toit.

A special class of fiber sensors makes use of fiber Bgrating (FBG)[4] and long period grating (LPG)[5] as thesensing device. Both of them are mainly produced by aodic refractive index change in the core of the optical fiAlthough these two types of gratings are sensitive to chain the refractive index of the surrounding medium, the Fsensitivity is reached only if the fiber cladding is redu

E-mail address:[email protected] (J.L. Fabris). to allow the access to the core’s evanescent field. However,

925-4005/$ – see front matter © 2004 Elsevier B.V. All rights reserved.oi:10.1016/j.snb.2004.06.033

R. Falate et al. / Sensors and Actuators B 105 (2005) 430–436 431

this procedure makes the FBG a very fragile device and thehandling can easily breaks the grating. In the other hand, theLPG couple light from core to cladding modes, which are verysensitive to refractive index changes of the external medium[5,6], allowing its use without etching the cladding. The useof such fiber sensing devices makes possible the insertion ofseveral gratings wavelength coded in a single optical fiberresulting in a quasi-distributed sensing link that can covera wide area of interest (e.g. around a petroleum refinery orbetween several gas stations). Furthermore, optimized LPGcan be used to measure concentrations of solutions with sen-sitivities equal to, or better than, that of conventional Abberefractometry[6,7].

In this paper, experiments are carried out with a LPG sen-sor immersed in external media of: gasoline blend and solventmixtures, air and butane–propane mixture and water environ-ment with gasoline. The changes in the refractive index of theexternal medium are obtained increasing the solvent concen-tration in pure and commercial gasoline samples, by addingthe butane–propane mixture to air, or by adding gasoline inwater. The obtained results show that long period gratingscan be applied for fuel quality control[8], detection of gaspresence in pipelines or small places and detection of hydro-carbon in water environments.

2

fo tings( ichh LPGp od-u PGo fibert tingsc tingm ectori alld thel nantw ivenb

λ

w andbn thect berh odes,w icroa t thec pledf leav-

ing several transmission dips in the transmission spectrum,each one corresponding to a specific coupling governed byEq.(1).

LPG can be very sensitive to changes in the refractive in-dex of the external medium. If the parameter to be measuredaffects the fiber surrounding refractive index, this alterationwill also change the matching condition expressed by Eq.(1)and will lead to wavelength shifts of the resonance dips inthe LPG transmission spectrum. Such displacement occursbecause the effective indexes of the cladding modes are de-pendent on the refractive index of the core, cladding, and ex-ternal medium. Depending on the grating periodicity and thecladding mode order, changes in the resonance wavelengthdip can be so high as 300 nm or so low as 13 nm for externalrefractive indexes ranging from 1 to 1.44[9]. For compari-son, in the case of Bragg gratings, the effective index of themode in the core fiber depends on the core and the claddingrefractive indexes. A change in the external medium index isonly perceived when the cladding is almost entirely removedto expose the evanescent field of the core mode to the ex-ternal medium index. LPG can also be designed to presenttemperature sensitivities as high as 100 pm/◦C [10] or as lowas 1.8 pm/◦C [11] depending on the grating pitch and thecladding mode order.

Chiang et al[12] presented analytical expressions to de-s n re-s ngesi , forew

δ

wi ,a hatd ctivei rδ

sen-s ltingi

3

thatd t-i sions ex-t there lledt witht dura-t the

. Principle of operation

In 1996, Vengsarkar et al.[5] reported the principle operation and some characteristics of long period graLPG). In comparison with standard Bragg gratings, whave spatial periods in the order of a micrometer, theresent a longer spatial period for the refractive index mlation, in the order of hundreds of micrometers. The Lperates by coupling the fundamental mode in the core

o co-propagating cladding modes. Standards Bragg graouple the fundamental mode with a counter-propagaode in the core, and the large change in the wave v

mplies short periods for the grating. For a LPG, the smifference in the co-propagating wave vectors requires

onger spatial periods in the index modulation. The resoavelength of coupling to a particular cladding mode is gy [5]:

m = (nco − nmcl )Λ (1)

here λm is the peak wavelength of the resonance between the core mode and the cladding mode,nco andmcl are, respectively, the effective refractive index ofore mode and of them-th order cladding mode, andΛ ishe grating pitch. The power exchange in the optical fiappens between the guided mode and cladding mhich are strongly affected by fiber imperfections, mnd macro bending, and by the boundary condition aladding–external medium interface. Thus, the light courom the core to the cladding modes leaks out the fiber,

cribe the shift at the resonance wavelength of a LPG iponse to the etching of the fiber cladding or to the chan the refractive index of the external environment. Thusxternal changes in the refractive index fromnex0 to nex, theavelength shiftδλ0 is [13]:

λ0 ≈ u2∞λ30Λ

8π3nclρ3

[1

(n2cl − n2

ex0)1/2

− 1

(n2cl − n2

ex)1/2

](2)

hereu∞ is them-th root of the Bessel functionJ0 [14], λ0s the resonance wavelength atnex0, Λ is the grating pitchndρ is the cladding radius. It is important to notice tifferent external media that promotes the same refra

ndex change fromnex0 to nex will lead to the same value foλ0, and cannot be recognized by the LPG.

Eq.(2) shows that the LPG operation and the gratingitivity are dependent on the cladding mode order, resun a higher sensitivity asu∞ increases.

. Experimental and results

The LPG are produced using a technique similar toescribed by[15]. A bare fiber without its protective coa

ng is inserted between the electrodes of a commercial fuplicer. A small weight is suspended in one of the fiber’sremities to keep a constant longitudinal tension. The oxtremity of the fiber is mounted on a computer controranslation stage. An electrical arc discharge is appliedhe fusion splicer, using an adequate current and timeion settings. After the discharge, the fiber is moved by

432 R. Falate et al. / Sensors and Actuators B 105 (2005) 430–436

required period of the grating, before another arc is applied.After a suitable number of point-to-point discharges, a peri-odic pattern is engraved in the refractive index profile of thefiber, because of heating activated processes. An optical set-up, which uses a halogen lamp, a monochromator, objectivelens, detector, and a personal computer, is employed dur-ing the writing process to monitor the transmission spectrumthrough the fiber. When the measured spectrum shows suit-able characteristics to the intended application, the processis terminated. Usually 40–70 points are necessary to form agrating for the described sensor. The advantage of using theelectrical arc is that no special fiber (hydrogenated or pre-sensitized) is required. The experiments are performed withthree different LPG fabricated in a fiber with a cladding radiusof 62.5�m and a core radius of 4.5�m. The fusion splicerarc parameters used in the writing process are a current of12 mA and fusing time of 0.5 s.

3.1. Solvent detection in commercial gasoline

A LPG with a pitch of 649�m and 53 points was employedfor gasoline blend and solvent mixture characterizations. Forthis grating the strongest attenuation peak, in air, is locatedat 1586.4 nm.

To measure the solvent sensitivity of the LPG in gasoline,t ecipi-e ticalfi theg ient,t rfer-e lled toa whichi isa Lu-m andh zerA ilityo df ctivei glassr

hy-d verale e andp s al-c holf

g arges ainttas indi-c entsa r-p wer

Fig. 1. Grating response when extra amounts of four different solvents areadded to commercial Brazilian gasoline. Alcohol and naphtha shift the at-tenuation peak towards longer wavelengths, whereas turpentine and paintthinner shift it towards shorter wavelengths. The lines through the data pointsare solely a visual aid. Uncertainty in measured points is typically less thansymbol size.

wavelengths. It can also be seen that the LPG is more sensitiveto alcohol addition in commercial gasoline than to naphthaaddition.

Fig. 2shows the evolution of the LPG transmission spec-trum when the refractive index of the external mediumchanges. The observed changes in the transmission spectraare obtained increasing the paint thinner proportion in com-mercial gasoline. For the sake of comparison, the spectra arealso shown when the external medium is air (n= 1.0000), puregasoline (n = 1.4470), and pure paint thinner (n = 1.4652).From the graphs inFigs. 1 and 2, it can be seen that forpaint thinner concentration above 70%, small concentrationchanges lead to larger resonance wavelength shifts. For suchconcentrations, the refractive index of the external medium

F iron-m ergiesa of thea nner).A

he sensor is inserted into a specially designed glass rnt with four openings, two of them used to insert the opber with the LPG and the two others to insert and to drainasoline samples. With the LPG inserted into the recip

he fiber ends are immobilized to avoid fiber-bending intence on the sensor response. Another parameter controvoid errors on the sensor response is the temperature,

s maintained between 20± 1.5◦C. The glass recipientlso immobilized to avoid its movement. A LED source (inent MREDSP5003, central wavelength at 1530.3 nmalf bandwidth of 52 nm) and an Optical Spectrum Analynritsu MS9710B (0.1 nm resolution, wavelength stabf ±5 pm, wavelength accuracy of±0.05 nm) are employe

or the transmission spectrum measurements. The refrandexes of the samples, after being drained from theecipient, are measured with an Abbe refractometer.

To study the LPG response when different petroleumrocarbon solvents are mixed to commercial gasoline, sextra amounts of anhydrous alcohol, naphtha, turpentinaint thinner are added to it. As we only used anhydrouohol in the experiments, it will be referred just as alcorom now on.

Pure paint thinner has a refractive index (n = 1.4652)reater than the fiber cladding, which accounts for the lhift of the resonant wavelength when the proportion of phinner in commercial gasoline is increased (seeFig. 1).Fig. 1lso shows that alcohol (n= 1.3665) and naphtha (n= 1.4059)hift the attenuation peak towards higher wavelengths,ating a reduction of the refractive index when these solvre mixed in commercial gasoline (n= 1.4240), whereas tuentine (n= 1.4411) shifts the attenuation peak towards lo

ig. 2. LPG response for different refractive indexes of the external envent. These spectra show the shift of the attenuation peak to higher ens the refractive index increases, and the decrease of the amplitudettenuation peak for the refractive index close to 1.4614 (87% paint thifter this value, the amplitude of attenuation peaks is very low.


Fig. 3. Grating sensitivity to changes in the external medium refractive indexand theoretical fitting given by Eq.(2). Uncertainty in measured points istypically less than symbol size.

is close to the cladding one (n = 1.458), and the best LPGsensitivity is reached[6].

Fig. 3shows the curve of the grating sensitivity to changesin the external medium refractive index. It also shows the bestfit of the analytical curve given by Eq.(2) for nex < ncl, whichagrees with the experimental data. The parameters used to thefitting are: (a) fixed parameters:Λ = 649�m, nex0 = 1.0000(air), ρ = 62.5�m; (b) ncl = 1.4640± 0.0007,u∞ = 14.46± 0.38, andλ0 = (1585.8± 0.5) nm, as variable parameters.These results allow to find the value of the correspondingcladding index (nm

cl ) and the order of them-th cladding modecoupled by the grating. TheJ0 root closest to theu∞ calcu-lated value is 14.93, corresponding to the fifth order mode[14].

The main product added to commercial Brazilian gasolineis anhydrous alcohol with an allowed proportion of 20–25%.Fig. 4shows the LPG response for a mixture of alcohol in bothcommercial (open circles) and pure gasoline (solid circles).

F olines ,( tt datap

Fig. 5. LPG temporal response when the butane concentration is changed.

The data from pure gasoline can be fitted by the empiricalequationy = a − bcx, which corresponds to the solid line.When the alcohol proportion in commercial gasoline is in-creased, the grating response agrees with the characteristic (orempirical) curve. When other solvents are added to the sam-ple of commercial gasoline, the total sample volume changesand consequently the alcohol proportion in this mixture de-creases, resulting in the other curves presented onFig. 4. Asit can be seen from that figure, only the legal commercialgasoline blend matches the characteristic curve when the al-cohol proportion is changed in the sample. This fact permitsa single, previously calibrated, LPG based sensor to monitorthe quality of the fuel[16].

Measurements were carried out both with a liquid flowand in a static condition. Additional shifts in the LPG reso-nance wavelength dip because of the liquid stream becomeinsignificant by increasing the longitudinal strain that holdsthe fiber.

3.2. Butane–propane detection

The experiments with butane–propane mixture (20%propane, 80% butane) are performed using a LPG with apitch of 604�m, 60 points of interaction, and two resonancepeaks centered, in air, at 1559 nm and 1573 nm.

on iss entsc lon-g ctedi per-l 2 nmh tioni

ngtha ed tot ong as thei ards

ig. 4. LPG attenuation peak wavelength vs. alcohol proportion in gasample when: alcohol is added to pure gasoline ((�) experimental points—) empirical curve); alcohol (©), naphtha (�), turpentine ( ) and painhinner (�) are added to commercial gasoline. The lines through theoints are solely a visual aid.

The experimental set-up for butane–propane detectiimilar as the one used for the liquid gasoline and solvharacterizations. The grating is fixed under constantitudinal strain, and the butane–propane sample is inje

nside the glass recipient. A superluminescent LED (Suum model 761, 1546.68 nm central wavelength and 53.1alf bandwidth) is used as light source. The OSA resolu

s set to 0.07 nm.Fig. 5 shows the changes in the LPG higher wavele

ttenuation peak when a butane–propane mixture is addhe atmospheric environment. The detection is made alpan of 600 s, with two gas insertions in this period. Asnjection process evolves, the peak attenuation shifts tow


shorter wavelengths, with a maximum shift of 0.06 nm in thefirst gas insertion. The same behavior is noticed when thegas is kept in contact with the LPG in a closed environmentalong time intervals up to 5 min. After this interval, the boxis opened and the LPG recovers the air wavelength in a shorttime interval. This result allows attributing the wavelengthshift to the refractive index change and not to temperaturechanges resulting from the gas injection.

Apparently the smaller wavelength shift on the second gasinsertion comes from a thick film of heavier hydrocarbonsresidues, which are present in the commercial gas mixture,and that condenses on the LPG surface.

3.3. Water pollutants detection

The water pollutants detection is carried out using a LPGwith a pitch of 614�m and 60 points of interaction. The inter-rogation set-up uses a super luminescent LED as light sourceand the optical spectrum analyzer. A cylindrical gutter madeof PVC is used as the monitored environment, with 10 cm ofdiameter by 34 cm length. A water stream is inserted in one ofthe channel sides, and in the opposite side there is a hole thatserves as fluid exit. The optical fiber with the sensing device(LPG) runs throughout the length of the gutter, being insertedthrough two small holes in the opposite end walls and keptu ho-s wingt car-b aturei na olinep moret

mon-i flowo e-t is toi terfl icityd urei

me.D dis-p ersedi LPGr closet is1 sioni n beb y thefi fiberp ct thes an rem y thef8

Fig. 6. LPG temporal response when the gasoline is inserted in the waterflow.

4. Conclusion

We report the use of a LPG to detect the presence of hy-drocarbons in water and air environments, and to evaluatethe quality of the automotive gasoline. Characterization ofthe LPG for samples with refractive indexes ranging from 1to 1.4614 resulted in an attenuation peak wavelength shift of65.5 nm. For the LPG operation in the best sensitivity range,corresponding to refractive indexes of the surrounding me-dia between 1.44 and 1.46, the average sensitivity is approx-imately 5× 10−4 nm−1. This sensitivity relates to a smallestrefractive index variation of 5× 10−5 that can be measured,for an equipment with wavelength accuracy of±0.05 nm.

Additional amounts of alcohol, naphtha, turpentine andthinner to the commercial gasoline can be detected by theLPG. Even low concentrations (about 8%) results in measur-able wavelength shifts (from 0.08 to 0.86 nm for alcohol andthinner, respectively).

Measurements of ethanol proportion when mixed withgasoline are reported using a long period grating-based sen-sor. Calibration of the process has been checked by the useof liquids with known refractive index. For the anhydrousethanol measurements the average sensitivity for the wave-length shift of the resonance band is 0.15 nm/%, for the regionwhere the ethanol proportion changes from 20 to 50% in them , thiss om0 redt singe ssi-b ltingi

ectedi t toa volv-i ectrals sure-m the

nder constant longitudinal strain. The fiber height is cen so that the LPG is close to the water surface, alloo the sensing device to enter in contact with the hydroon sample. During the whole experiment, the temper

s controlled to be constant within±1◦C. In this detectiorrangement the wavelength shift resulting from the gasresence can be up to 25 times the obtained shift for the

emperature sensitive gratings[10].The procedure used for the measurements consists in

toring the LPG transmission spectrum, under a waterf 18 ml/s at 18◦C, during the first 2 min. The interval b

ween successive measurements is 3 s. The next stepnsert 30 ml of commercial Brazilian gasoline in the waow and the spectra are acquired with the same perioduring the next 36 min. After that time, the whole proced

s repeated.Fig. 6shows the LPG wavelength evolution along the ti

uring the first 2 min, the dip in the transmission spectralays the value corresponding to the LPG response imm

n water (1555.58 nm). For both gasoline insertions, theesponse shifts the wavelength towards lower values,o 1549.73 or 1549.12 nm. Fort = 4 min the wavelength555.09 nm, a value that is close to the one for immer

n water. The small hysteresis shown here (0.49 nm) caecause of the adsorption of hydrocarbon molecules bber surface, since both the hydrocarbon and the silicaresent a polar behavior. Such hysteresis does not affeensor performance, as the subsequent water stream cove the adsorbed film. This hypothesis is supported b

act that the wavelength returns to its initial value fort =min.

-

ixture with gasoline. For the naphtha measurementsensitivity is 0.03 nm/% for naphtha proportion ranging frto 33%. In spite of this lower sensitivity when compa

o the ethanol sensitivity, the process can be improved uquipment with a better spectral resolution. Another poility is to use a narrow line bandwidth LPG sensor, resu

n better resolution accuracy.The hydrocarbon presence in the water media is det

n a very short time (about 1 min), what is very importanvoid damages that could happen to the environment in

ng petroleum hydrocarbon leakage. The observed sphifts (around 6 nm) are easy to detect because of meaents in the optical domain. For leakage detection in


environment, where changes in the temperature can also re-sult in LPG wavelength shifts, a fiber Bragg grating can beused in the same optical link to compensate the effect of thetemperature.

For the butane–propane detection, one can observe that thecentral wavelength of attenuation peak doesn’t return to itsinitial value. This effect is also observed with lower intensitywhen gasoline is added to water. In this case, the water flowcleans the fiber surface, removing gasoline residues that couldchange the grating sensitivity. The absence of such a cleaningflow for the gas detection, besides gas adsorption process bythe fiber and/or the presence of residual amount of gas in therecipient could account for the observed difference.

For all the experiments, the sensor response time showedto be influenced mainly by the electronics in the measure-ments systems. Response times of less than 1 min and around10 s are found for liquid and gas hydrocarbons detection, re-spectively. The times to the LPG recovers the initial condi-tions after the liquid or gas draining are less than 1 min andaround 10 s, respectively.

Acknowledgements

This work was partially done under the auspiceso H-A T-P thes efin-e

R

l. 15

ho-hem.

tionpillednol.

Ap-Nor-

rdo-rs, J.

e ofave

lu-7.-the

and

ong266.

[10] V. Bhatia, Applications of long period gratings to single and multi-parameter sensing, Opt. Express. 4 (1999) 457–466.

[11] V. Bhatia, D.K. Campbell, D. Sherr, T.G. D’Alberto, N.A. Zaba-ronick, G.A. Ten Eyck, K.A. Murphy, R.O. Claus, Temperature in-sensitive and strain insensitive long period grating sensors for smartstructures, Opt. Eng. 36 (1997) 1872–1876.

[12] K.S. Chiang, Y. Liu, M.N. Ng, X. Dong, Analysis of etched long-period fibre grating and its response to external refractive index,Electron. Lett. 36 (11) (2000) 966–967.

[13] M. Miyagi, S. Nishida, An approximate formula for describing dis-persion properties of optical dielectric slab and fiber waveguides, J.Opt. Soc. Am. 69 (1979) 291–293.

[14] M.R. Spiegel, Mathematical Handbook of Formulas and Tables,McGraw-Hill Inc., USA, 1968.

[15] G. Rego, O. Okhotnikov, E. Dianov, V. Sulimov, High-temperaturestability of long-period fiber gratings produced using an electricalarc, J. Lightwave Technol. 19 (10) (2001) 1574–1579.

[16] Medidor otico de qualidade de combustıvel para mediçao local emedicao remota”, Patent PI0203712-2, Brazil, 2002.

Biographies

Rosane Falatewas born in Curitiba (Brazil) in 1976. She received thedegree in Electric Engineering and obtained M.Sc. degree from CentroFederal de Educaçao Tecnologica (CEFET-PR) in 2000 and 2002, respec-tively. CEFET-PR is also where she is obtaining Ph.D. degree. Her currentresearch interests are the production and application of photonics devices,such as Bragg gratings and long-period gratings, to communication ands

R er no( re hei uctiona long-p

M ntroF ego razil)a .H efectsc ory atC . Herc pticalfi

H ni-v theP ver-s CSELT( entroF chi pticalfi versityo hass s andO andO

f the Brazilian National Agency of Petroleum (PRNP/MME/MCT 10 CEFET/PR), CNPq, CAPES, CETRO and LITs/TECPAR.The authors acknowledgeupport provided by Pres. Getulio Vargas petroleum rry (REPAR).

eferences

[1] A.D. Kersey, et al., Fiber gratings sensors, J. Lightwave Techno(8) (1997) 1442–1463.

[2] J.A.P. Lima, M.S.O. Massunaga, H. Vargas, L.C.M. Miranda, Ptothermal detection of adulterants in automotive fuels, Anal. C76 (2004) 114–119.

[3] S. Halmemies, S. Grondahl, K. Nenonens, T. Tuhkanen, Estimaof the time periods and processes for penetration of selected soils and fuels in different soils in the laboratory, Spill Sci. TechBull. 8 (5–6) (2003) 451–465.

[4] A. Othonos, K. Kalli, Fiber Bragg gratings: Fundamentals andplications in Telecommunications and Sensing, Artech. House,wood, 1999.

[5] A.M. Vengsarkar, P.J. Lemaire, J.B. Judkins, V. Bhatia, T. Egan, J.E. Sipe, Long-period fiber gratings as band-rejection filteLightwave Technol. 14 (1) (1996) 58–65.

[6] H.J. Patrick, A.D. Kersey, F. Bucholtz, Analysis of the responslong period fiber gratings to external index of refraction, J. LightwTechnol. 16 (9) (1998) 1606–1612.

[7] R. Falciai, A.G. Mignani, A. Vannini, Long period gratings as sotion concentration sensors, Sens. Actuators, B 74 (2001) 74–7

[8] R. Falate, H.J. Kalinowski, J.L. Fabris, M. Muller, Long period grating application for fuel quality control, in: In Proceedings ofSixth International Conference on Optoelectronics, Fiber OpticsPhotonics, Mumbai, 2002, pp. 351–354.

[9] X. Shu, L. Zhang, I. Bennion, Sensitivity characteristics of lperiod fiber gratings, J. Lightwave Technol. 20 (2) (2002) 255–

ensor systems.

icardo Canute Kamikawachi was born in Jales (Brazil) in 1978. Heceived the degree in Physics from Universidade Federal do Paraa andbtained M.Sc. degree from Centro Federal de Educaçao TecnologicaCEFET-PR) in 2001 and 2003, respectively. CEFET-PR is also whes obtaining Ph.D. degree. His current research interests are the prodnd application of photonics devices, such as Bragg gratings anderiod gratings, to communication and sensor systems.

arcia Muller is currently an Associate Professor of Physics at Ceederal de Educaçao Tecnologica do Parana (CEFET-PR), Brazil. Shraduated in Physics from the Universidade Federal do Parana (Brazil),btained M.Sc. degree from the Universidade Federal Fluminense (Bnd Ph.D. degree from the University of Sao Paulo, Brazil, in 1994er research focused on energy transfer in rare earth/molecular do-doped doped crystals. She helped to found the Laser LaboratEFET-PR in 1996, where she is a laboratory co-Director nowadaysurrent main research area is photonics, with special interest in ober grating based sensors and optical cammunications.

ypolito Jose Kalinowski obtained the B.Sc. in Physics from the Uersidade Federal do Parana (Curitiba), M.Sc. and Ph.D. degrees fromontifical Catholic University of Rio de Janeiro, Brazil. He joined Uniidade Federal Fluminense and was a post doctoral researcher inTorino, Italy). From 1991 onward he is an Associate Professor at Cederal de Educaçao Tecnologica do Parana, where his current resear

nterests are photonics devices for optical communications and ober sensors. In 2001, he also spent a sabbatical leave at the Unif Aveiro (Telecommunications Institute), Portugal. Dr. Kalinowskierved as the past president of the Brazilian Society for Microwaveptoelectronics (SBMO) and he is also a member of IEEE, SPIESA.


Jose Luıs Fabris is currently an Associate Professor of Physics at CentroFederal de Educac¸ao Tecnologica do Parana (CEFET-PR), Brazil. Heobtained the B.Sc. in Physics from the Universidade Federal do Parana(Brazil), M.Sc. degree from the Universidade Federal Fluminense (Brazil)and Ph.D. degree from the University of Sao Paulo, Brazil, in 1994.

His research focused on color center lasers and laser spectroscopy. Hehelped to found the Laser Laboratory at CEFET-PR in 1996, where heis a laboratory co-Director nowadays. His current main research area isphotonics, with special interest in optical fiber grating based sensors andoptical cammunications.

fiber optic sensors for hydrocarbon detection

Documents