August 15, 2006 / Vol. 31, No. 16 / OPTICS LETTERS 2405

Optical fiber long-period grating with solgelcoating for gas sensor

Zhengtian Gu and Yanping XuCollege of Science, University of Shanghai for Science and Technology, P.O. Box 249, 516 Jun Gong Road,

Shanghai 200093, China

Kan GaoShanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, P.O. Box 800-211,

Shanghai, 201800, China

Received February 22, 2006; revised May 20, 2006; accepted May 31, 2006;posted June 5, 2006 (Doc. ID 68242); published July 25, 2006

The novel long-period fiber grating (LPFG) film sensor is composed of the long-period grating coated withsolgel-derived sensitive films. The characteristics of the transmissivity of the LPFG film sensor are studied.By analyzing the relation among the sensitivity Sn, the thin film optical parameters, and the fiber gratingparameters, the optimal design parameters of the LPFG film sensor are obtained. Data simulation showsthat the resolution of the refractive index of this LPFG film sensor is predicted to be 10−8. Experimentally,a LPFG film sensor for detection of C2H5OH was fabricated, and a preliminary gas-sensing test wasperformed. © 2006 Optical Society of America

OCIS codes: 050.2770, 060.2370, 060.2430, 160.6060, 310.1860.

The long-period fiber grating (LPFG) has been usedas a refractive index sensor because of its high refrac-tive index sensitivity.1,2 Now LPFGs are used mainlyin chemical concentration measurement, especiallyin online monitoring of the concentration of danger-ous solutions of materials or of materials inaccessibleto industrial product quality control.3 In these appli-cations, the fiber cladding of these kinds of LPFGsensor is directly immersed in the external medium,and such sensors are sensitive only when the refrac-tive indices of the external media are less than orequal to those of the fiber gratings.

Recently attention has been drawn to coating anoverlay of thin film on the fiber cladding of the grat-ing region. Rees4 has coated an organic thin film onthe cladding of the LPFG by the Lamgmuir–Blodgetttechnique and studied the influence of the thicknessof the overlay material on the LPFG response. Theresearch indicates that this sensor just avoided thelimitation that the refractive indices of the externalmedia must be less than or equal to those of the fibergrating. Further, the Langmuir–Blodgett film hasbeen proved to be chemically sensitive, so the LPFGfilm sensors could be used as species-specific chemi-cal sensors.

The solgel technology has many advantages overthe other methods, such as simple and low-cost pro-cessing, precise control of the doping level, and a po-rous structure very desirable for gas sensor applica-tion. Thus, in this Letter we selected the solgel-derived metal-oxide-semiconductor films as sensinglayers. While the coated LPFG is brought into con-tact with the specific gases, the semiconductor sur-face energy changes, which leads to the change ofconductivity and refractive index. Consequently, itresults in the shift of the resonance position of thetransmission spectrum. To bring the change of reso-nance position of the transmission curve into play tothe utmost, the optical parameters of sensing films

and grating structure parameters must be optimized.

0146-9592/06/162405-3/$15.00 ©

In this Letter, the sensitivity to the refractive in-dex �Sn� of the LPFG film sensor was studied basedon the model of triple-clad long-period fiber instead ofthe approximate model of the multiplayer planarwaveguide introduced in Ref. 4. Experimentally, weprepared three different LPFG sensors coated withsolgel-derived SnO2 thin films, and we observed theirresponses to C2H5OH.

For single-mode LPFG the coupling between thefundamental mode propagating in the fiber and co-propagating cladding modes must satisfy the phase-matching condition

�co − �cl�l,v� = l ·

2�

�, l,v = 1,2,3, . . . , �1�

where � is the period of the fiber grating and �co and�cl

�l,v� are the propagating constant of the fundamentalguided mode and the vth cladding mode with the azi-muthal order l, respectively, which could be obtainedby solving the core mode eigenvalue equation5 andthe cladding eigenvalue equation of four-layeredLPFG.6 For untilted LPFG, only the cladding modeswith the azimuthal order l=1 are taken intoconsideration.7

From the cladding eigenvalue equation, the propa-gating constant �cl

�l,v� is related to thin film refractiveindex n3 and thin film thickness h3. While the LPFGsensors coated with metal-oxide-semiconductor filmsare exposed to the specific gases, the transmissionpeak will shift due to the tiny variation of thin filmrefractive indexes.

Figure 1 shows the transmission spectrum of theHE13 mode of the coated LPFG with a different filmrefractive index n3 as the film thickness h3 and gat-ing parameters (length L, period �, and the indexmodulation �) are held constant. It can be seen thatthe film refractive index has an influence on the mag-

nitude and position of the attenuated peak, and they

2006 Optical Society of America

2406 OPTICS LETTERS / Vol. 31, No. 16 / August 15, 2006

are quite distinct from each other in the response tothe change of refractive index for two different pa-rameters.

As an LPFG film gas sensor, the tiny variation ofthe refractive index of the sensing film when exposedto external gas should cause the change of the trans-mission curve. In this case it appears suitable to de-fine the physical quantity describing the sensor sen-sitivity as follows:

Sn = � �T

�n3

n3

T � . �2�

Sn represents the sensitivity to n3 of the sensor. Ob-viously, Sn is a multivariable function of the film op-tical parameters and grating parameters, and therewould be an extremum in a specific region. Thereforethe LPFG film sensor must be designed with optimi-zation techniques.

In practical applications, the LPFG film sensorwith optimum parameters works around the centralwavelength of the selected cladding mode so as toreach the maximum sensitivity. By observing thechange of T while the sensitive film of the sensor re-acts to the (bio)chemical material, the analyte can bedetected.

Figure 2 shows the dependence of �Sn�max on n3 andh3 for wavelengths ranging from 1360 to 1460 nm

Fig. 1. Transmission spectrum of HE13 mode with differ-ent refractive index (h3=500 nm, �=2�10−4, L=1.8 cm,�=450 �m).

Fig. 2. Dependence of �Sn�max on refractive index n3 for h3ranging from 0–800 nm �n3=1.5–1.9�.

( �Sn�max represents the maximum of Sn in the above-

mentioned range). It shows explicitly that there is agreat difference in maximum sensitivity �Sn�max forvarious combinations of n3 and h3. The magnitudes of�Sn�max are available to more than 104. The contour

line of �Sn�max=103, 104 is shown in Fig. 3, fromwhich we can select suitable materials according tooptimum n3 and coat the film with optimum thick-ness h3 for sensor design.

Three typical optical parameters and their sensi-tivities and resolutions are listed in Table 1, where �nrepresents the minimum resolution of the refractiveindex; �n is given by

�n = Sn−1nf�dT

T �−1

, �3�

where dT /T is taken as 10−3 for the usual measuringinstrument in the calculation. From Table 1, we cansee that the resolution of n3 of the film sensor is pre-dicted to be 10−8, by selecting suitable film optical pa-rameters, together with the central wavelength.Similarly, the grating parameters �L ,� ,�� exert aninfluence on the sensitivity Sn just like the film pa-rameters.

The fiber used in the experiment was CorningSMF-28, which was hydrogen loaded for a week in150 atm hydrogen of high purity to enhance its pho-tosensitivity. The light source was an excimer laserwith a wavelength of 193 nm, single-pulse energy of170 mJ/cm2, and a repetition frequency of 5 Hz. Thefiber grating 1.8 cm in length was made using a cop-per amplitude mask with a period of 450 �m, andwas annealed at 150°C for 24 h to improve the tem-perature stability. The SnO2 solution was obtained bydissolving 45.12 g of SnO2·2H2O in 500 ml absolute

Table 1. Typical Film Optical Parameters andRelevant Sensitivity and Resolution for Same

Grating Parametersa

RefractiveIndex n3

Thicknessh3

Wavelength�c

SensitivitySn

Resolution�n

1.52 740 nm 1477.4 nm 1.25�104 1.21�10−7

1.54 640 nm 1476.7 nm 1.77�104 8.70�10−8

1.76 210 nm 1547.0 nm 1.52�103 1.16�10−6

a −4

Fig. 3. Contour line of �Sn�max=103, 104. The contour of104 is located in the center of the right branch.

�=4�10 , L=1.8 cm, �=450 �m.

August 15, 2006 / Vol. 31, No. 16 / OPTICS LETTERS 2407

ethanol. It was refluxed and stirred at 80°C for 2 hand aged for 24 h at 30°C to get the sol.

To investigate the influence of film optical param-eters on the sensitivity of the LPFG film sensor, threekinds of solgel-derived SnO2 film with different pa-rameters were prepared. For the solgel-derived SnO2film, the refractive index was between 1.74 and 1.76under the technological conditions mentioned above(see Ref. 8). Based on the contour of sensitivity, therewould be an area of higher sensitivity when the filmthickness is close to 200 nm. Because the film thick-ness cannot be available to 200 nm by one-step dip-ping, we introduced double-step dipping. The fiberwas drawn again from the sol after drying the first-run dipped fiber in an oven and was heated in a tube-type resistance furnace for 30 min at 150°C, thenheated from 150°C to 500°C for 120 min and kept atthe temperature of 500 °C for 30 min, and thencooled down unaffectedly. The dip rates for LPFGsamples 1, 2, and 3 were selected to be 30, 60, and120 mm/min to get different optical parameters.

The coated long-period grating is housed in a gasflow chamber, which was controlled at 150°C to im-prove the response of the gas sensor and eliminatethe effect of temperature fluctuation on transmission.The long-period grating is illuminated by a broad-band light source, and the experimental data werecaptured on an optical spectrum analyzer.

For the three kinds of coated LPFG sample withdifferent film optical parameters, there were signifi-cant differences in the rate of change of transmissiv-ity T, which brought about the distinction of gas sen-sitivities accordingly. The attenuated peak observedin the experiment was the fourth at wavelengthabout 1555 nm. Figure 4 shows the transmissionspectrum of sample 3 �vd=120 mm/min� before andafter acting on alcohol vapor with the concentrationof 200 parts per million (ppm). The central wave-length of the attenuated peak shifted about 3 nm to-wards the shorter wavelength from 1557 to 1554 nm.The change rate of T is about 25%. There was no vis-ible shift of the attenuated peak for samples 1 and 2.

Theoretically, Sn of the coated LPFG sample 3 wasof the order of 103 since its film thickness was in thevicinity of 200 nm (see Table 1). The sensitivity to therefractive index was higher, which resulted in an ob-vious shift of the attenuated peak. For samples 1 and2, however, Sn was of the order of 101 because thefilm thicknesses were away from the optimal regions��200 nm�, so the sensitivity of the LPFG sensor wasso low that there was almost no response to C2H5OH.With the experimental design parameters, the detec-

tion limit is available to 1 ppm for C2H5OH in thepresented setup.

In summary, the optimal optical parameters of thethin film layer of a novel LPFG sensor are designedby using the optimization numerical method. Datasimulation shows that the resolution of this schemeto the refractive index of the films is predicted to be10−8. The gas-sensing experimental results obtainedin the scheme accord with the theoretical analysisand demonstrate the necessity for optimization of op-tical parameters. Due to the advantages of high sen-sitivity, simplicity, and compactness, our presentedoptical film sensor will likely be developed for appli-cations in the analytical chemistry process, environ-mental monitoring, and biochemical sensing.

This research was supported by Shu Guang project(02SG32) of the Shanghai Municipal Education Com-mission and the Shanghai Education DevelopmentFoundation, and supported by the Shanghai LeadingAcademic Discipline Project (T0501). Z. Gu’s emailaddress is zhengtiangu@yahoo.com.cn.

References

1. V. Bhatia and A. M. Vengsarkar, Opt. Lett. 21, 692(1996).

2. S. W. James and R. P. Tatam, Meas. Sci. Technol. 15, 49(2003).

3. R. Falciai, A. G. Mignani, and A. Vannini, Sens.Actuators B 74, 74 (2001).

4. N. D. Rees, S. W. James, and R. P. Tatam, Opt. Lett.27, 686 (2002).

5. D. Gloge, Appl. Opt. 10, 2252 (1971).6. C. Tsao, Optical Fibre Waveguide Analysis (Oxford U.

Press, 1992).7. T. Erdogan, J. Opt. Soc. Am. A 14, 1760 (1997).8. Z. T. Gu and P. H. Liang, Opt. Laser Technol. 36, 211

(2004).

Fig. 4. Measured wavelength shift of the fourth-orderresonance peak captured on an optical spectrum analyzer.