By Topic

• Abstract

SECTION 1

## INTRODUCTION

Micro- and nanostructures fabricated on fiber end facet can form versatile optical devices with miniature size and flexibility [1], [2], [3]. Many works have been done to promote the combination of nanostructures and optical fibers. For instance, fabricating gold ring grating on multimode fiber's end facet to realize refractive index sensing by a pronounced peak caused by Rayleigh anomaly of the ring grating [4], [5]; proposing and demonstrating a flexible optical probe for photonic integrated circuits, which is based on a single-mode optical fiber with a subwavelength-period gold grating on its facet [6]; and fabricating periodic metal dots or nanostructures on fiber facet to form a surface-enhanced Raman spectroscopy probe or a refractive index sensor [7], [8], [9], [10]. All of these designs are interesting and attractive, but most of them are based on expensive and low productivity nanofabrication techniques, which make these designs far from practical applications.

Photonic crystals (PCs) are one kind of sensing materials with great potential due to their periodic and interconnected structures [11], [12], [13], [14]. Recently, many applications based on three dimensional PCs have been reported, such as relative humidity (RH) sensing based on $\hbox{T}_{\rm i}\hbox{O}_{2}$ inverse opals fabricated by glancing angle deposition [15]; chemical vapor detection by opals [16], [17], [18], [19]; independent multifunctional detection by wettability controlled inverse opal hydrogels [20]; PH value sensor [21], [22] and $\hbox{NH}_{3}\hbox{-HCl}$ detection by polyaniline-infiltrated $\hbox{T}_{\rm i}\hbox{O}_{2}$ inverse opal film [23]. Three dimensional PCs used in these applications were fabricated by several different methods. Among these methods, self-assembly method is the most popular and effective way to produce three dimensional periodic structures with controlled layers and periodicity in sub-micrometer scale. Even two dimensional periodic structures can be made by using one layer of periodical colloidal sphere arrays as deposition mask [24]. However, self-assembled PC films will crack to areas in diameter of $\sim\! 50\ \mu\hbox{m}$, and optical response of the PC films will be nonuniform in different areas when the number of layers of the film is above $\sim$15. Cracks will affect optical properties of the films, and nonuniformity will bring some deviations when the measurements are not taken at the same point in sensing applications. However, it is important to note that recent advances in rapid convective deposition method had resulted in uniform monolayer hexagonal sphere arrays, which had been implemented in both light-emitting diodes [25], [26] and organic LEDs [27].

In this paper, we proposed to fabricate opal films and inverse opal films on single mode optical fiber bundle end facets (FBEFs) (shown in Fig. 1) to overcome some disadvantages of self-assembled films and to combine the advantages of optical fiber probes, which are flexible, reliable, and can be used in harsh environments and remote sensing. Firstly, incident angle of light projected on PC films keeps constant and probe area in the film is fixed. Because in this design, probe light was guided by the fiber core and projected on the PC film which is tightly deposited on fiber end facet. Hence, in sensing applications, repeat detection of optical response of PC films in a same area will be practicable by this structure. Secondly, large area high quality PC films are not required. When the fiber end facet was covered with PC films, only in the core area of fiber the PC film will be probed, which is about 9 $\mu\hbox{m}$ in diameter. Thus, this structure has some tolerance to shortcomings of self-assembled PC films such as cracks and nonuniformities, and optical properties of the PC films on FBEFs will not be affected by those defects. For reasons above, this structure will make applications of PC films more achievable, reliable and adaptable. Because area of the end facet of a single fiber is very small and the facet is a circle plane, it is very hard to fabricate high quality multilayer PC films on it by self-assembly methods. Additionally, when the PC film is more than $\sim$10 layers, it will easily drop from the single fiber tip. Therefore, we use a fiber bundle rather than a single fiber to construct this structure. Optical reflection spectra of this structure were investigated and one kind of RH sensors base on this structure were presented in this article.

Fig. 1. Schematic diagram of (a) fiber bundle, (b) bundle with opal film and (c) with inverse opal film.
SECTION 2

## FABRICATION AND DISCUSSIONS

### 1. Fiber Bundle Preparation

Fiber bundles used in the experiments are composed of single mode optical fibers, 9/125 $\mu\hbox{m}$, $\sim$50 cm in length, and a copper ferrule, 9 mm in outer diameter, 3 or 5 mm in inner diameter and $\sim$10 mm in length. One end of the fiber bundle was ferruled while the other end is free. The ferruled end was polished to form a facet (FBEF) which is orthogonal to fiber axis $(90 \pm 3^{\circ})$ to assemble PC films on it, and the free end was used to send and receive light. In the ferruled end, interstitials of fibers in the ferrule were filled with epoxy resin.

Surface treatment of FBEFs is vital to deposit high quality opal films and inverse opal films on it by self-assembly method. The surface of FBEFs must be hydrophilic and have been activated before deposition, and roughness of the surface should be less than 0.1 $\mu\hbox{m}$. Oxygen plasma treatment is a feasible way to realize these acquirements.

In the following experiments, the FBEFs were cleaned by ultrasonic bath with deionized water for 5 minutes, and rinsed by deionized water for 2 minutes, then dried in nitrogen flow and followed by $\hbox{O}_{2}$ plasma treatment (Suman) for 3 minutes.

### 2. Fabrication of Polystyrene Opal Films, Composite PC Films, and Inverse Opal Films on FBEFs

Vertical self-assembly method was used to fabricate polystyrene opal films on FBEFs. Several drops of polystyrene microsphere suspension, which was bought from Bangs Laboratories Incorporated and was used as received, were diluted by deionized water to concentration of 0.1 wt%–0.5 wt%, and then, a cleaned end facet of a fiber bundle was immersed into the diluted suspension and was fixed to keep it vertical to the liquid surface. While the suspension is evaporating slowly, polystyrene microspheres are being assembled into ordered arrays at the interface of liquid, bundle facet and air by capillary force. Good quality polystyrene microsphere arrays in long range (millimeter scale) can be formed on the FBEF by precision control of the evaporation parameters, such as temperature, RH, air fluctuate and so on. For this experiment, the suspension was evaporated in an electric oven at temperature of $50\ ^{\circ}\hbox{C} \pm 1\ ^{\circ}\hbox{C}$. A layer of polystyrene opal film is formed on the FBEF when the suspension is dry, as shown in Fig. 2(a) and (b). They are assembled from polystyrene spheres in diameter ${\rm D}_{\rm sphere} = 390\ \hbox{nm}$ and 490 nm, respectively.

Fig. 2. (a)–(d) Pictures of photonic crystal films on fiber bundle end facets: (a) and (b) polystyrene opal films, (c) composite photonic crystal film, (d) silica inverse opal film. (e) Optical microscope image of silica inverse opal film on a fiber bundle end facet. (f)–(h) Scanning electron microscope images of photonic crystal films: (f) polystyrene opal film, (g) composite photonic crystal film and (h) silica inverse opal film.

Silica inverse opal films on FBEFs were fabricated by sol-gel co-assembly method and calcination. Deionized water diluted polystyrene microsphere suspension with concentration of 0.2 wt%–0.6 wt% was added with hydrolyzed tetraethyl orthosilicate (TEOS) solution which consists of a 1 : 1 : 1.5 ratio by weight of TEOS (98% Aldrich), 0.1 M HCl, and ethanol (100%), respectively, and stirred at room temperature for 1 hour prior to use [28]. Volume ratio of added TEOS solution and polystyrene microsphere suspension is $\sim$0.001. The added TEOS solution worked as silica sol-gel precursor solution. And the silica precursor sol in the mixed suspension will be infiltrated in interstitials of the polystyrene spheres while the spheres are self-assembling. Then, the infiltrated sol will aggregate to form silica precursor gel, which can turn to silica by calcination, in gaps of the spheres as the evaporation continues. A cleaned FBEF was immersed vertically into the mixed suspension and fixed. The evaporation temperature for this experiment is $45\ ^{\circ} \hbox{C} \pm 1\ ^{\circ}\hbox{C}$. In about 24 $\sim$ 36 hours, the solvent was evaporated and a layer of composite PC film, which is polystyrene opal film infiltrated with silica precursor gel in gaps of the spheres, was formed on the FBEF, as shown in Fig. 2(c). It was assembled from polystyrene spheres in ${\rm D}_{\rm sphere} = 690\ \hbox{nm}$. The formed film was calcined at $500\ \sim\ 700\ ^{\circ}\hbox{C}$ for $\sim$5 minutes and gradually cooled down to room temperature, and then, the silica inverse opal film was obtained, as shown in Fig. 2(d) and (e).

### 3. Structural Characterization of PC Films on FBEFs and Discussions

With methods mentioned above, monodispersed polystyrene spheres with diameters ranging from 390 nm to 690 nm were assembled on FBEFs in face-centered-cubic structure (which is also called opal), and inverse opal films (three dimensional macro porous materials) were also formed on FBEFs with pore diameters ranging from $\sim$360 nm to $\sim$660 nm. The opal films and inverse opal films on end facet of fibers are mostly uniform in both thickness and colors, as we can observe in Fig. 2(a), (b), and (d). Fig. 2(e) is an optical microscope image which shows details of the inverse opal film in (d). Although cracks can be clearly seen, areas $(\sim\!40\ \mu \hbox{m}\ \times \sim\!400\ \mu\hbox{m})$ with no cracks in the film are large enough to cover the fiber core. The inverse opal films on epoxy resin which was infiltrated in fiber interstitials are mostly damaged during calcination, while the inverse opal films on end facets of the fibers are preserved. Mostly, area of opal films and inverse opal films fabricated on FBEF with high quality is above $40\ \mu\hbox{m} \times 40\ \mu\hbox{m}$ and typically in strip shape.

By scanning electron microscope (SEM), microstructures of these PC films were examined. Typical top view SEM images of polystyrene opal films, composite PC films, and silica inverse opal films on FBEFs are exhibited in Fig. 2(f)(h), respectively. The insets provide larger magnification images of these films from which differences of the three kinds of films can be clearly observed. In the opal film, as shown in (f), mono-sized spheres are hexagonal close packed layer by layer, which forms a face-centered-cubic structure, and it is {111} plane parallel to the FBEF. Air voids in the opal film is $\sim$26% in volume according to face centered cubic structure. Compared to the opal film, it is silica precursor gel rather than air in interstitials of the polystyrene spheres in the composite PC film, as shown in (g), which is the only difference between them. And differences between the inverse opal film, shown in (h), and the composite PC film are: (i) polystyrene spheres were removed (replaced by air) in the inverse opal film; (ii) the gel turns to silica in the inverse opal film. Therefore, volume ratio of air and silica in the inverse opal film is $\sim$74% and $\sim$26%, respectively. If interstitials of polystyrene spheres in opal films were filled with silica and then removed these spheres, the polystyrene opal film will turn to silica inverse opal film. It is important to note that it is inter-connected by air channels and pores for opal and inverse opal films, respectively.

As described above, the three kinds of PC films are the same in lattice structure and can be fabricated by similar methods while they have differences in volume ratio of air in them and materials that constitute them. The complementary characteristics of the three kinds of PCs make the proposed fiber-PC film structure more flexible to different applications.

To further improve the quality of self-assembled PC films, it will be effective to precisely control the optimized evaporation parameters during the period of assembling. It is also very important to decrease polydispersity of microspheres (deviations of diameters of microspheres from their average value) in suspension as small as possible, which is less than 10% in our experiments, because tiny differences in size of spheres will disrupt the long range order in the films and will largely affect the reflection spectra of the films.

Thickness of the vertically self-assembled films which can be expressed as the number of layers of the spheres is controlled mostly by the concentration of polystyrene microspheres in suspension. Number of layers of the assembled films on FBEFs can be controlled from 1 to about 40 by adjusting concentrations of polystyrene spheres in suspensions, which is similar to vertical self-assembly of spheres on homogeneous planar substrates [29]. Increased concentration will result in increased number of layers. Using large volume of suspension in experiments can decrease the influence of continuous increased concentrations of spheres, which were caused by solvent evaporation. If the assembled film is too thick, structure defects in it will increase due to the mechanism of the fabrication method, and the film is easy to drop from the FBEFs. If the film is too thin, its reflection peak intensity is weak and reflection bandwidth is wide, as we will see in Fig. 7(a). Considering the structure quality and optical properties of the PC films, which will be discussed in detail in Section 3.2, the optimized number of layers of PC films on FBEFs is 15–20.

When there is only one layer of polystyrene spheres on FBEFs, this layer can be a deposition mask for producing periodic metal patterns on fiber facet which could form local surface plasmon resonance or surface-enhanced Raman spectroscopy detectors [24]. If the one layer of polystyrene spheres is infiltrated with gel, after removing polystyrene spheres by calcination, two dimensional periodic structures can be obtained on FBEFs. According to the infiltration ratio which is infiltrate height against sphere diameter, resulted two dimensional periodic structures will have different morphologies, as shown in Fig. 3. SEM image Fig. 3(a) is a two dimensional periodic structure formed with high infiltration ratio, and image (b) is formed with low infiltration ratio. Image (a) indicates that each cavity is connected to its six neighbor cavities by pores on the wall.

Fig. 3. SEM images of two morphologies of 2-D periodic structures formed on fiber bundle end facets.
SECTION 3

## OPTICAL CHARACTERIZATION

### 1. Simulation Results

Three dimensional finite difference time domain method was used to simulate both the reflection spectra and inner electric field distribution of PC films on fiber end facet.

The simulation was performed by EastFDTD software (DONGJUN Technology). The simulation model is a single optical fiber with polystyrene opal film on its end facet, and it is {111} plane of the film that parallel to fiber end facet, as shown in Fig. 4(a). Light was incident into the model from the other end of the fiber. Parameters of the fiber model were set as: fiber length 30 $\mu\hbox{m}$, cladding diameter 15.6 $\mu\hbox{m}$, core diameter 9 $\mu\hbox{m}$, cladding refractive index 1.45, and core refractive index 1.46. Parameters of the polystyrene opal film model were set as: diameter of spheres ${\rm D}_{\rm sphere} = 675\ \hbox{nm}$, sphere refractive index 1.56, air refractive index 1, number of layers of the film is 39, and area of the film is $15.6\ \mu\hbox{m} \times 15.6\ \mu\hbox{m}$.

Fig. 4. (a) Shows the simulation model. (b) Simulated reflection spectrum of the opal film on fiber bundle end facets, and (c) simulated images of electric field distribution in longitudinal section of the model at different frequencies when light propagate in this structure.

To avoid divergence and get accurate results in simulations, mesh size should be far small than the minimal wavelength simulated. However, small mesh size will increase the calculate complex dramatically. According to our simulation results, mode field diameter of this single mode fiber model at 1550 nm is $\sim\! 9.6\ \mu\hbox{m}$, and electric field intensity in this mode decreases to $\sim$3% of its peak value at radius of 15.6 $\mu\hbox{m}$. Hence, we chose 15.6 $\mu\hbox{m}$ as cladding diameter approximately to decrease simulation time while the introduced errors are acceptable.

Reflection coefficient of this model in range from 1100 nm to 2000 nm in wavelength was calculated by reflection coefficient calculation module of the software, which is plotted in the form of reflectivity in Fig. 4(b). Results of the simulated electric field distribution in the model when light propagates in it are presented in Fig. 4(c). The three pictures show the electric field distribution in longitudinal section of this model at three frequencies near or in the reflection band of the PC film. Distribution width of electric field of incident light in the opal film is $\sim\!\!10\ \mu\hbox{m}$ for all the three pictures. Based on this result, we consider that interactions of light which is emitted from neighboring fibers are weak in the opal film, and we can infer that when all fibers in the bundle are lighted at the same time, each light spot on the opal film will be spatially distinguishable, though the fibers are closely arranged in the bundle.

### 2. Experimental Results and Discussions

Reflection spectra of opal films and inverse opal films deposited on FBEFs were measured by the free end of the fiber bundle. A single mode 1 × 2 coupler with 50/50 split ratio was used to connect a white light source (Yokogawa AQ4305), an optical spectrum analyzer (Yokogawa AQ6370 or Ocean Optics USB2000), and one fiber in the fiber bundle. The white light source and the optical spectra analyzer are connected to the coupler by port1 and port2, respectively, as shown in Fig. 5. Reflectivity of the PC films was defined as: ${\rm R} = ({\rm P}_{2} \times 2)/{\rm P}_{3}$. ${\rm P}_{3}$ is the power of light projected on the PC film in port3. ${\rm P}_{2}$ is power of reflected light in port2 when port3 connected the structures.

Fig. 5. Schematic diagram of measuring reflection spectra of photonic crystal films on fiber bundle end facets.

Fig. 6(a) and (b) shows the measured optical reflection spectra of polystyrene opal films and silica inverse opal films on FBEFs in different ${\rm D}_{\rm sphere}$. A band of strongly increased reflectivity is observed in every measured spectrum which corresponds to the photonic band gaps of opal films and silica inverse opal films. Reflection peak position $\lambda_{\rm c}$ and peak reflectivity ${\rm R}_{\rm c}$ of the reflection spectra plotted in Fig. 6(a) and (b) are listed in Table 1. Number of layers of these PC films ${\rm N}_{\rm L}$ as well as the effective refractive indices of them can be calculated from the well defined Fabry-Perot interference fringes and reflection peaks in their reflection spectra, which were also given in Table 1. In the calculation, we assume that diameters of spheres in the films are the same, and their values were indicated in Table 1, line 2. However, there is a little shrinkage in diameter of pores in inverse opal films after calcination, which is about 460 nm and 660 nm. Average refractive indices of the fabricated polystyrene opal films and silica inverse opals are 1.350 ± 0.035 and 1.086 ± 0.018.

Fig. 6. Reflection spectra of (a) opal films and (b) inverse opal films on fiber bundle end facets, and (c) reflection peak positions of opal films versus diameter of spheres. (d) Distribution of reflection peak positions of the polystyrene opal film on fiber bundle end facets.
TABLE 1 CHARACTERISTICS OF REFLECTION SPECTRA OF PHOTONIC CRYSTAL FILMS PLOTTED IN FIG. 6(a) AND (b)

Because lattice structure of opal films and silica inverse opal films is the same and optical properties of them are similar, we will only discuss optical characteristics of opal films for example. We plotted the experimentally measured $\lambda_{\rm c}$ of opal films in Fig. 6(c). Deviations of reflection peak positions from linearly fit line are 22.6 nm, 24 nm, 17.8 nm, and 20 nm, respectively. According to Bragg diffraction law, these points should be in a straight line. Factors that may cause these deviations are polydispersity of the polystyrene spheres, structure defeats (layer mismatch), different number of layers of these films, tiny discrepancy in angles between fiber facet and fiber axis among different fibers $(\pm 3^{\circ})$, and the influence of environment conditions (temperatures and relative humidities) when measurements were taken.

To examine the degree of optical uniformity of PC films on the FBEFs, 17 fibers in a bundle with an opal film on its facet were randomly selected to measure their reflection spectra. The opal film was assembled from polystyrene spheres in ${\rm D}_{\rm sphere} = 390\ \hbox{nm}$. Reflection peak positions of these spectra were plotted in Fig. 6(d). All the peak positions are distributed in a range from 831 nm to 866 nm, while 64.7% of them are distributed in the range from 838 nm to 852 nm. Decrease polydispersity of the spheres and improve the film quality by precision control of evaporation parameters in fabrication will effectively increase the optical uniformity of PC films on FBEFs.

The reflection peak position of this structure is adjustable by changing ${\rm D}_{\rm sphere}$, and spheres in diameter ranging from tens of nanometers to hundreds of nanometers are suitable for self-assembly method. Reflection band of this structure can locate in a wide range, which makes it adaptable to different optical communication networks.

Full-width at half-maximum (FWHM) of the reflection band plotted in Fig. 6(a) and (b) are shown in Table 1. As we can see from these data, FWHM will increase as ${\rm D}_{\rm sphere}$ increases except for ${\rm D}_{\rm sphere} = 580\ \hbox{nm}$ and 690 nm. According to previous reports [30], [31], [32], we believe that this is due to the difference of ${\rm N}_{\rm L}$ of the two films, which is 15 ± 1 and 48 ± 1. To investigate the influence of ${\rm N}_{\rm L}$ on optical properties of opal films on FBEFs, we measured optical reflection spectra of opal films in ${\rm D}_{\rm sphere} = 690\ \hbox{nm}$ with different layers, as shown in Fig. 7(a). Number of layers is calculated from the well defined Fabry-Perot interference fringes. Characteristics of these measured reflection spectra as well as simulated results are plotted in Fig. 7(b)(d); the tendency of simulation and experiment results are the same. The value difference between simulations and experiments may be caused by the imperfection of face centered cubic lattice of opal films and losses (absorption and scattering) that not accounted for in the calculation model. ${\rm R}_{\rm c}$, $\lambda_{\rm c}$, and FWHM of the reflection peaks are not invariant when ${\rm N}_{\rm L}$ changes. ${\rm R}_{\rm c}$ will increase [Fig. 7(b)], while $\lambda_{\rm c}$ and FWHM will decrease [Fig. 7(c) and (d)] when ${\rm N}_{\rm L}$ increases. When ${\rm N}_{\rm L}$ is large enough, in this case ${\rm N}_{\rm L}\ >\ 50$, ${\rm R}_{\rm c}$, $\lambda_{\rm c}$, and FWHM will approach a certain value and change very little when N continues to increase.

Fig. 7. (a) Measured reflection spectra of polystyrene opal films in ${\rm D}_{\rm sphere} = 690\ \hbox{nm}$ on fiber bundle end facets with different layers. (b)–(d) Characteristics of reflection spectra of polystyrene opal films versus number of layers of the films: (b) full-width at half-maximum of the reflection peak, (c) reflection peak position and (d) reflection peak reflectivity.
SECTION 4

## RH SENSING

### 1. Principles of PC Films for Sensing

Reflection peak positions of PCs can be predicted by Bragg diffraction equation. It has the same form for the three kinds of PCs which we fabricated and can be expressed as TeX Source $$m \lambda_{\rm c} = 2\times n_{eff}\times (\sqrt{6}/3) \times {\rm D}_{\rm spheres} \times \sin\theta\eqno{\hbox{(1)}}$$ where $m$ is an integer determined by the order given, $n_{eff}$ is effective refractive index of the diffractive media, $(\sqrt{6}/3) \times {\rm D}_{\rm spheres}$ is the inter-layer distance and ${\rm D}_{\rm spheres}$ is diameter of polystyrene spheres or air spheres in inverse opal films, and $\theta$ is incident angle of light. The $n_{eff}$ of PC films can be calculated as TeX Source $$n_{eff} = \sqrt{(\varepsilon_{1} \times f_{1} + \varepsilon_{2} \times f_{2} + \cdots)}\eqno{\hbox{(2)}}$$ where $\varepsilon_{\rm i}\ ({\rm i} = 1, 2, \ldots)$ are dielectric constant of materials that comprise the PC films, and $f_{\rm i}\ ({\rm i} = 1, 2, \ldots)$ are volume ratio of the corresponding materials in PCs. If there is a change in effective refractive index of the PCs, a shift of the reflection peak will be detected, which is the mechanism that most sensing applications of PC films are based on.

### 2. Experimental Method to Measure the Optical Response of PC Films on FBEFs to RH

Since saturated salt solutions in closed container can keep RH of air above them at constant value if the environment temperature is kept constant, several kinds of saturated salt solutions were selected to obtain the desired relative humidities. A FBEF with PC film on it was inserted into the closed container and kept the PC film above the liquid surface. In a couple of minutes, RH in the closed container reached a constant value, which was monitored by a commercial RH detector (TESTO 605) fixed in the container near the FBEF. During the period of measurement, the environment temperature was set at $20 \pm 1 ^{\circ}\hbox{C}$.

### 3. Optical Response of Opal and Inverse Opal Films to RH

Fig. 8(a) and (c) show the changing reflection spectra of PC films, which were measured when environment RH increased from 12% to 97% at 20 °C. Now the most acceptable RH sensing mechanism for PC films is [15]: as the RH increases, water vapor penetrates the connected pores in PCs and condenses in capillary voids in PC films. This progress increases the average refractive index of the PC films. Based on this explanation, hydrophilic surface of pores in PC films is helpful for water vapor condensing on them [13]. Because polystyrene opal films have bad surface hydrophilicity at 20 °C, shifts of spectral features (such as the reflection peak position) are not obvious when RH changes, especially in low RH [see Fig. 8(b)]. For silica inverse opal films, although hydrophilicity of their surface is not good either, there are more capillary voids or cracks in their structures due to their fabrication process. And water vapor can easily condense in these voids to change the effective refractive index of the films. So their spectral features shift obviously to longer wavelength as RH increases, as shown in Fig. 8(d). Compared with opal films and inverse opal films that have been widely researched as RH sensors recently, composite PC films have not been focused on as RH sensing elements, which also have good performance and great potential.

Fig. 8. Reflection spectra of (a) polystyrene opal films and (c) silica inverse opal films on a fiber bundle end facet in different relative humidities at 20 °C. Reflection peaks of (b) polystyrene opal films and (d) silica inverse opal films versus relative humidity.

### 4. Optical Response of Composite PC Films to RH and Discussions

The changing reflection spectra of a composite PC film, which were measured when RH increased from 12% to 86% at 20 °C, are shown in Fig. 9(a). The changing reflection spectra show that the reflection peak shifts to longer wavelength and peak reflectivity decreases as RH increases.

Fig. 9. (a) Measured reflection spectra of composite photonic crystal film on a fiber bundle end facet versus relative humidity. (b)–(d) Features of the reflection spectra versus relative humidity: (b) reflection peak position versus relative humidity, (c) reflection peak intensity versus relative humidity, (d) full-width at half maximum versus relative humidity.

As we have discussed, the composite PC film is composed of polystyrene spheres and silica precursor gel. The gel is in interstitials of the polystyrene spheres, and it consists of a solid three dimensional network that spans the volume of water and ensnares it through surface tension effects. For a certain RH at fixed temperature, equilibrium between water evaporation and water vapor condensation in the gel can be achieved. When RH varies, a new equilibrium in the gel will be formed, but quantities of water in the gel will be changed. And effective refractive index of the gel is sensitive to volume ratio of water in it. Thus, effective refractive index of the composite film will be changed when RH varies, which results in a shift of the reflection peak. Compared with the gel, polystyrene spheres can be seen as solid with no voids, and therefore, they have no contributions to water vapor absorption, and their refractive index can be seen as constant.

Fig. 9(b) shows average $\lambda_{\rm c}$ of three circles of measurement. One measurement circle includes two parts: raise environment RH from low to high and then decrease it from high to low. In Fig. 9(b), an approximate linearity between $\lambda_{\rm c}$ and RH in range from 12% RH to 76% RH can be seen, and $\lambda_{\rm c}$ will increase about 0.325 nm when RH increases 1% in this range. A repeatable hysteresis loop is observed in Fig. 9(b). Hysteresis commonly exists in mesoporous solid and is created by capillary condensation effect [15]. While hysteresis has been examined at $\lambda_{\rm c}$, it affects other features of the reflection spectra. Reflection peak intensity is different in the rising and falling proportion of the circle, as well as FWHM of the reflection spectra, as shown in Fig. 9(c) and (d). Fig. 9(d) is average FWHM of three measurement circle. FWHM is relating to refractive index contrast of the materials in PC films, and FWHM will decrease when the contrast decreases. During the RH rising proportion of the circle, FWHM first increases and reaches a maximum at $\sim$33% RH. After this point, FWHM decreases as RH increased further. When RH changes to decrease in this circle, FWHM decreases to a minimum value at about 76% RH and then increases. According to the behavior of FWHM, we believe that refractive index contrast between polystyrene spheres and the gel has the same changing profile as RH varies.

SECTION 5

## CONCLUSION

In this paper, robust polystyrene opal films and silica inverse opal films with controlled layers are deposited on FBEFs. The optimized number of layers for PC films on FBEFs is 15–20. Characteristics of reflection spectra of PC films on FBEFs were investigated. Reflection bands of the assembled PC films can locate in a wide range by choosing microspheres of proper refractive index and diameters. Number of layers of PC films has influence on reflection peak reflectivity, reflection band width, and reflection peak position of the films. Probe light emitted from different fibers in a bundle has weak interactions with each other in the PC films when the films are not too thick. Monodispersity of microspheres is very important to produce high quality PC films on FBEFs with good optical properties. The composite PC films, composed of silica precursor gel and polystyrene spheres, on FBEFs are sensitive to RH. Relationship between the reflection peak position of the composite PC film and RH is approximately linear in the range from 12% RH to 76% RH at 20 °C. In this range, reflection peak of the film will shift $\sim$0.325 nm for a 1% change in RH. Two dimensional periodic structures with different morphologies formed on FBEFs can be utilized to fabricate metal patterns on fiber tips to form surface-enhanced Raman spectroscopy and local surface plasmon resonance sensors. Further, if the PC films are deposited on the end facet of a fiber-optic imaging bundle, in situ image contrast will make sensing applications of PC films more reliable and effective, especially for a small shift in wavelength.

### ACKNOWLEDGMENT

We would like to thank Prof. J. Li for useful discussions and giving some good suggestions.

## Footnotes

This work was supported by the National Natural Science Foundation of China under Grants 61178044 and 91123015, the Specialized Research Fund for the Doctoral Program of Higher Education under Grant 20113207110004, the Jiangsu Province Prospective Joint Research Project under Grant BY2012005, and the University Postgraduate Research and Innovation Project of Jiangsu Province, China, under Grant CXZZ02-0405. Corresponding author: M. Wang (e-mail: wangming@njnu.edu.cn).

## References

No Data Available

## Cited By

No Data Available

None

## Multimedia

No Data Available
This paper appears in:
No Data Available
Issue Date:
No Data Available
On page(s):
No Data Available
ISSN:
None
INSPEC Accession Number:
None
Digital Object Identifier:
None
Date of Current Version:
No Data Available
Date of Original Publication:
No Data Available