Liquid-Core Optical Fibers

This is a continuation from the previous tutorial - Tapered Fibers and Specialty Fiber Microcomponents

1. INTRODUCTION

It is widely recognized that the need for larger bandwidths for communication systems was the main driving force for the development of optical fibers. The invention of the laser in 1960 triggered great expectations regarding the possibility of increasing the amount of information carried by a modulated wave using an optical signal. Simultaneously, it was acknowledged that a suitable transmission medium was needed so that optical signals could propagate over long distances with minimum losses.

Today several technological barriers have been broken, leading to the development of optical communication systems relying on optical fibers as a transmission medium. Such systems not only have fulfilled the needs foreseen in the early 1960s but also have enabled the development of more sophisticated technologies. Optical communication systems can now manage simultaneously the transmission of video, data, and voice, thereby conveniently exploiting the large bandwidths offered by optical fibers.

Throughout the years, fabrication methods for low-loss silica fibers have improved considerably, and it is now possible to tailor the spectral properties of glass fibers when using the appropriate materials within the core. However, in the early stages, liquids were among the first materials tested as core media for optical fibers. While solid materials fully compatible with silica had not been found, several liquids offered two main features that were attractive enough for fiber fabrication: a low-absorption coefficient and a refractive index higher than that of glass.

Thus, liquids provided two essential requirements for optical transmission, namely, low losses and wave-guiding by means of total internal reflection. Although several limitations for long-distance transmission were found, these were regarded as perfect step-index multimode fibers, with a constant refractive index across the core and a sharp transition at the core–cladding boundary.

The development of liquid-core fibers for long-distance communications was not further pursued once the fabrication methods for low-loss glass fibers were available. Nonetheless, several fundamental concepts and fiber characterization techniques currently in use were developed using liquid-core fibers.

A number of liquids were tested in the early days as core materials and these studies led to a better understanding of scattering effects on these waveguides. Besides having a direct influence on fiber losses, scattering in liquid-core fibers proved later on to be a useful mechanism to enhance nonlinear effects in liquids, mainly because of the long interaction lengths offered by the fiber.

As we will see in the following sections, this and other interesting features are crucial for all the wide variety of applications that benefit from the use of liquid-core fibers.

2.  PROPAGATION OF LIGHT IN LIQUID-CORE FIBERS: MODAL FEATURES, DISPERSION, AND POLARIZATION EFFECTS

Generally speaking, liquid-core fibers are multimode waveguides and mode theory provides suitable theoretical background to understand the propagation of light within these fibers. Operation of these waveguides is based on multimode propagation of light within the liquid core, which is contained by a hollow glass or capillary tube used as fiber cladding.

Because the liquids used as core materials have a higher refractive index than that of the glass tube, the guided beams propagate by multiple total internal reflections at the core–cladding interface (Fig.1). As the first practical realization of a suitable transmission medium for light over relatively long distances, liquid-core fibers offered a convenient testbed for several experimental studies. Furthermore, they were used in the first demonstration of television broadcasting through optical fibers.

Among the fundamental concepts studied with liquid-core fibers, modal effects and dispersion were perhaps the most interesting. Studies regarding mode launching conditions yielded very important information that was extremely useful to understand how light propagates in a cylindrical waveguide. Even though the glass tubes used as cladding in the early days had very high losses, it was demonstrated that total transmission attenuation could be lower than 20 dB/km.

Further research later showed that this was due to the numerical aperture \(\text{(NA)}\) and the multimode nature of the fibers. Because these fibers can have very high \(\text{NA}\), high-order modes can be effectively filtered or coupled to lower order modes, thereby showing low transmission losses and high bandwidths.

Early studies also demonstrated the effects of cladding losses upon varying the angle of incidence of the probe beam. This method allowed for the identification of mode filtering effects at the cladding, that is, an increase in losses after an angle of incidence smaller than the critical angle indicated that the modes propagating at such angles would experience higher attenuation. 

Because communications applications were not further pursued with these fibers, information regarding dispersion is limited to the early papers. Among other key features, pulse rates of \(200\text{Mbits/km}^{-1}/\text{sec}^{-1}\) were shown to be feasible with liquid-core fibers and it was demonstrated that an increase in the fiber core diameter and a reduction in fiber bend radius both increase the dispersion. As confirmed in these early experiments, mode conversion during propagation is responsible for the dispersion effects observed in these fibers.

Further fundamental research on fiber dispersion focused on the effects of bending radius and launching conditions. Because core homogeneity and very low scattering for the wavelength of interest, no bandwidth limit was apparent with these fibers. Remarkably, attenuation did not seem to increase dramatically with bend radius, which is interesting because glass core fibers are susceptible to bend losses due to their weakly guiding structure. Dispersion effects are, thus, related to the multimode nature of the fiber and to the \(\text{NA}\).

Other experimental techniques were used to evaluate modal effects in liquidcore fibers and proved very useful to validate mode-coupling theory. As an example, experiments based on far-field observation verified a theoretical model describing coupling between modes supported by optical fibers.

Results showed that lateral stress applied to the fiber can generate mode coupling and mode scrambling. Coupling between modes was expressed in terms of a normalized mode conversion coefficient, which turned out to be two orders of magnitude smaller for liquid-core fibers compared to glass core fibers. Liquid-core fibers are, thus, less susceptible to stress effects and the excited modes for a given launching condition will propagate with minimum coupling.

In fact, most of the early reports on liquid-core fibers show that single-mode propagation is maintained in these fibers once the fundamental mode is excited.

Further understanding of the multimode propagation of light in fibers was also obtained upon using the high-temperature dependence of the liquid core. A detailed experimental study was carried out in which the mode cutoff frequencies were observed in transmitted and scattered light when varying the temperature of a short segment of fiber.

Selective excitation of the modes supported by the fiber was achieved through adjustments on the temperature and the launching conditions. Thus, mode cutoffs were readily observed as a function of temperature. These results were shown to compare well with theoretical predictions obtained with weakly guiding mode theory.

Another interesting feature of the experimental results was an oscillating behavior attributed to modal interference. This modal effect can be related to core size and ellipticity, thereby suggesting a characterization method for fiber geometry.

Polarization properties of liquid-core fibers have also been reported. In agreement with the aforementioned studies on stress effects, liquid-core fibers do not exhibit stress birefringence, which is commonly observed in glass core fibers. This yields polarization maintaining behavior for straight fibers as long as single-mode operation is sustained.

Nonetheless, bends in the fiber can create axes of birefringence, which are orthogonal to each other, so polarization of the guided beam can be adjusted mechanically. The phase shift induced by these axes can be compensated by rotating the plane of curvature, as is commonly done with fiber polarization controllers with rotatable paddles. Besides bending and stress, no other birefringence sources were reported in the early papers.

For most liquid-core fibers, uniformity and purity of the liquid core is most likely to minimize the effects of other sources of birefringence. However, as we will see, liquid crystals can be used in the core to develop polarization-sensitive fiber devices.

3. FABRICATION AND CHARACTERIZATION METHODS 

The simplest description of a liquid-core fiber is a capillary tube filled with a liquid that has suitable optical properties for a given application. There are, however, some variations on the geometry and even on the cladding material that widens the types of waveguides using a liquid core.

Regarding geometry, several studies describe the use of liquid-core planar waveguides. Devices and applications based on this geometry are not within the scope of this chapter. Instead, we focus on fibers with glass cladding. Other designs and materials are further described in APPLICATIONS. 

Liquid-core fibers are fabricated using a hollow glass tube that serves as a cladding. Because of the high \(\text{NA}\) and the modal effects observed in these fibers, there are no stringent requirements regarding losses in the glass.

In fact, some of the early papers note the low losses obtained with these fibers in spite of the high bulk losses of the glass tubes. Several types of glasses have been reported to be useful for fabrication of these waveguides and even capillaries have been reported to yield adequate features in the transmission properties of the fiber. The glass tube is pulled following the same procedure as that used for pulling glass fibers. As can be seen in early reports on fabrication of liquid-core fibers, the predecessors of the modern draw towers were first developed for pulling hollow glass tubes.

As we will see in APPLICATIONS, reports have demonstrated that ‘‘holey’’ fibers can also be filled with liquids and, thus, serve as liquid-core fibers. In this particular case, the fabrication process is more complicated than simply pulling a single glass tube, but the pulling method is similar in both cases.

Evidently, the goal is to avoid collapsing the glass tube so a properly sized core can be created. Once the tube is pulled to the required dimensions, the cladding is ready to host the liquid core, although some applications may require an extra coating on the inner wall.

The filling of the hollow fiber is generally carried out through hydrostatic pressure and the dimensions of the fiber will determine the time required for this process. A wide range of applications involve the use of small lengths of fibers, so capillary forces are enough to fill short lengths of a hollow fiber. However, for long lengths, the filling process requires an increase in pressure to achieve convenient filling times.

A cell can be specially designed for controlling the hydrostatic pressure and to host the hollow fiber, the liquid, and a window so that light can be launched into the fiber core. As an example, early papers on liquid-core fibers report the use of a Monel cell (Fig. 2) with a Teflon plunger that allowed filling a 50-m long fiber in half an hour.

Evidently, the core size is important in evaluating the time required to fill the fiber. This can be readily seen from formal analysis of laminar flow in small-bore pipes. As an example, the time \((T)\) taken to fill a fiber of length \(L\) is given.

\[\tag{1}T=\frac{16\mu}{P}\left(\frac{L}{d}\right)^2,\]

where \(P\) is the applied pressure, m is the coefficient of viscosity of the liquid, and \(d\) is the core diameter. It is clear from this expression that for a fixed pressure, long lengths of fiber will require longer filling times, and conversely, fibers with larger core diameters require shorter filling times.

The maximum pressure that can be applied to a glass fiber before rupture is related to the tensile strength of the glass \((S)\), the outer diameter of the tube \((D)\), and the core diameter. Explicitly, this can be evaluated approximately as

\[\tag{2}P_\text{max}=\frac{S(D-d)}{D}.\] 

Both of these equations can be used to determine the optimum core size \((d_\text{opt})\) required to fill a fiber of length \(L\) in a minimum time. An increase in pressure to reduce filling times is, thus, limited to the mechanical properties of the glass tube. Other methods that may be used focus on reducing the capillary forces upon reducing the viscosity of the liquid, provided that the optical properties remain unaltered. 

Several liquids have been used as fiber cores ranging from bromobenzene and o-dichlorobenzene (the first reported in the literature) to water and ethanol. The selection of the liquid for the fiber depends on the required optical features for a given application.

Some of the desired characteristics in a liquid are evident; for instance, if a simple capillary or hollow fiber is used as cladding, the liquid has to have an index of refraction higher than that of the glass. However, this condition is not necessary if the liquid is to be enclosed inside a microstructured fiber.

For transmission over long lengths of fiber, the liquid must have low loss at the wavelength of interest and, in general, the liquid has to be stable and nonvolatile and must have low viscosity.

Also, scattering effects are important in applications involving nonlinear effects; thus, a suitable liquid with proper scattering coefficient must be chosen. The \(NA\) of liquid-core fibers is generally high compared to that of glass fibers. Evidently, the value for this parameter depends on the refractive indices of the liquid core and the cladding.

Values ranging from 0.2 to 0.6 have been reported, and once again, the optimum value is determined by the requirements of a given application. Finally, the optical properties of liquid-core fibers are determined using what are today the standard characterization methods for optical fibers.

4. APPLICATIONS

As mentioned earlier, liquid-core fibers were first envisioned as a suitable transmission medium for optical communications systems. Practical limitations for the fabrication of these waveguides and, more importantly, the development of highly transparent glass core fibers proved that other materials offered better features for communications applications. However, the possibility of selecting a liquid core with very specific optical properties has remained highly attractive for a wide range of applications.

Among other interesting features, liquid-core fibers offer the possibility for transmission of ultraviolet \((UV)\) light, enhancement of nonlinear effects in liquids, and even the use of liquid crystals for developing polarization-sensitive fiber devices.

The development of microstructured fibers has also opened new possibilities for the fabrication of waveguides with unique features that can be tailored upon selecting a suitable liquid. Development of light sources based on nonlinear effects and sensors using liquids are two broad categories in which applications of liquid-core fibers can be classified.

Some specific examples that fall into these categories are reviewed in the following sections. 

Waveguides for Special Spectral Regions and Optical Chemical Analysis 

Early studies on the spectral features of liquid-core fibers focused on the visible and infrared (IR) regions, mainly because of the availability of laser sources at these wavelengths. Naturally, upon selecting the appropriate liquid for the core, the fiber also offered the possibility of guiding light at other nonstandard wavelengths. As an example, the first reports on transmission losses at the \(IR\) region \((3.39\mu m)\) used a fiber with a tetrachloroethylene \((C_2Cl_4)\) liquid core. Losses at this particular wavelength region are high for this fiber \((10^4-10^5 \text{dB/km)}\), but these reports suggested that other liquids with suitable optical properties were likely to offer better results.

Transmission of \(UV\) light has also proven to be feasible with specially designed liquid-core fibers. Absorption spectrometry, in which fibers are regularly used as a means to increase the interaction length of the light with the analyte, has increased its spectral range of operation from the visible to the \(UV\) upon the use of aqueous ethanol solutions as fiber core.

Essentially, the fibers are used as cuvettes and the transmission spectrum is registered to detect absorption bands for specific analytes. Because light can travel several meters along the fiber, an enhancement in sensitivity is effectively achieved with these waveguides.

Aqueous solutions require adding solvents to increase the refractive index of the liquid, although approaches such as adding an inner coating to the glass tube have allowed for the use of simple liquids such as water (Fig. 3). The use of Teflon coatings in glass and plastic tubes has yielded good results in highly sensitive detection of pesticides and water pollutants using liquid-core fibers together with chromatographic and spectroscopic techniques.

Theoretical analysis has shown that Teflon layers as small as 5mm are sufficient to confine the light within the liquid core and, thus, avoid environmental effects and scattering by the capillary material. Applications such as \(\text{pH}\) monitoring have also reported enhancement in detection sensitivity using a simple transmission monitoring scheme.

In this case, transmission through the liquid-core fiber is simply monitored as a function of \(\text{pH}\), so the absorption spectra can be used to determine the \(\text{pH}\) value.

Chemical analysis has benefited from the use of liquid-core waveguides in planar and cylindrical geometries. Besides liquid chromatography and absorption spectroscopy, other analytical techniques such as fluorescence and Raman measurements have enhanced their detection limits due to Teflon-coated capillaries.

The Teflon \(AF\) family of fluoropolymers is perhaps the most widely used for inner coatings of waveguides, because the range of refractive indices available (1.29–1.31) is adequate to generate total internal reflection when aqueous solutions are used as a fiber core. Generally speaking, the liquid-core fiber acts as a flow cell and as a reaction chamber in which light can be generated either by a chemical reaction or by the Raman effect. Light is then guided by the

fiber to a detection system to monitor the spectral and/or the intensity features of the guided beam (Fig. 4). AF-polymer tubing is now commercially available and liquid-core fibers are nowadays almost a standard tool in analytical chemistry. Furthermore, availability of miniaturized \(\text{LEDs}\) and detector arrays has also opened the possibility of developing compact and portable chemical analysis systems using liquid-core fibers. 

Fiber Sensors 

The first report on a liquid-core fiber sensor was for voltage monitoring through the Kerr effect. However, detailed analysis of the performance of these fibers as sensors was first carried out in distributed temperature sensing using \(\text{OTDR}\). The main advantage of using liquid-core fibers in this configuration is that Rayleigh scattering and the \(\text{NA}\) are highly dependent on temperature for these types of waveguides. Furthermore, as mentioned earlier, the modal features vary with temperature as well.

The fibers used in these first experiments were pulled from silica tubes; other special preparation included deposition of a layer of high-purity silica inside the tubes, and the addition of an outer polyimide coating.

The core was filled with hexachlorobuta-1,3-diene using a high-pressure syringe system. With this filling system, the authors could fill 150 m of fibers with 150-mm core in 30 minutes. The fibers had an \(\text{NA}\) of 0.2 and 0.54 at 900 and 589 nm, respectively, and losses were measured to be 13 dB/km at 900 nm. 

Results from the first experiments on distributed temperature sensing were very useful to determine the temperature sensitivity of the fibers. Two major

temperature effects are understood to play an important role: changes in \(\text{NA}\) and in scattering loss. Although other means such as transmission losses monitoring can be used to register temperature changes, the backscattering loss allows for the use of \(\text{OTDR}\).

As the temperature increases, the scattering loss increases due to the thermal agitation, which is strongly dependent on temperature. Analysis of the setup used by Hartog showed that a resolution of 1 m over fiber lengths of more than 100 m was attainable, with a temperature accuracy of \(1^\circ\text{C}\).

The sensitivity obtained experimentally was \(23.3\times10^{-3}\text{dB}/^\circ\text{C}\) (0.54%/\(^\circ\text{C})\) over a temperature range of 5–110 \(^\circ\text{C}\).

A simpler approach that does not require an \(\text{OTDR}\) system was later demonstrated based on transmission measurements as a function of temperature. As shown in this report, liquid-core fibers also offer the possibility of creating multiplexed arrays of sensors using different liquids to extend the temperature range of operation. 

Nuclear radiation detectors can also be developed using liquid-core fibers. The idea behind these sensors is to fill capillaries with liquid scintillators so that light generated by luminescence is guided by the fiber. Studies on radiation resistance of liquid scintillators and capillaries filled with these liquids have been carried out demonstrating that high scintillating and trapping efficiencies can be achieved.

Such arrangements are capable of yielding a track hit density higher than that of detectors based on plastic fibers or semiconductors. Because of the large attenuation lengths (which, in turn, depend on the \(\text{ID}\) of the capillaries), it is possible to construct detectors with lengths more than 2 m and a spatial resolution of less than 20 microns/hit. Several glass tubes of different grades were used as capillaries for the fibers.

The influence of radiation in quartz was investigated upon comparing the attenuation for scintillating liquid-core fibers with quartz tubes before and after exposing it to radiation. Although glass darkening was observed in low-grade quartz, it was found that radiation resistance of the arrangements was limited by the liquid rather than the capillary glass.

A comparison between plastic scintillating fibers and scintillating liquid-core fibers showed that radiation resistance is much better for the latter arrangements.

From this study, it was suggested that plastic fiber will work well for doses less than 5 Mrad, but that liquid-core fibers can be used for doses higher than 60 Mrad without changing the liquid scintillator.

The use of liquid scintillators as fiber cores has also allowed for the development of novel sensing techniques capable of yielding high-quality imaging of ionizing particle tracks with very high spatial and time resolution. Using a \(\text{CCD}\) as readout, the fibers act simultaneously as target, detector, and light guides.

As in other applications, the liquid composition can be optimized to maximize light output and attenuation length. A passing charged particle creates scintillating light in the liquid core and a fraction of the light is guided by the fiber due to total internal reflection. The amount of light trapped within the core depends on the \(\text{NA}\) and the \(\text{ID}\) of the capillaries.

The use of \(\text{CCDs}\) as readout has also allowed for the development of a sensor array that can function as a vertex detector.

These high-resolution systems include a set of image-intensifier tubes followed by a \(\text{CCD}\) or an electron bombardment \(\text{CCD}\) (EBCCD) camera.

Detectors with as much as \(10^6\) capillaries have been reported to yield resolutions of 20-\(40\mu m\) and are capable of withstanding radiation levels at least an order of magnitude higher than those of other tracking devices of comparable performance.

Other relevant achievements obtained with these vertex liquid-core sensors include the recording of high-quality images of neutrino interactions.

Nonlinear Optical Effects

Scattering properties of liquids have always been attractive for observing nonlinear optical effects. However, most of them require long interaction lengths and high power to yield useful features. Because liquid-core hollow fiber systems allow for higher local intensity and longer gain length, they have been used in several applications involving nonlinear processes.

Stokes-shifted, super-broadening, stimulated scattering was among the first effects demonstrated with a liquid-core fiber. A laser beam with intensities from \(10^6\) to \(5\times10^8 W/\text{cm}^2\) was fed into a 250-cm long fiber using \(\text{CS}_2\) as a liquid core.

Among other features, the spectral range of the stimulated scattering radiation was very large \((>700\text{cm}^{-1})\).

Further research on this topic demonstrated that a nonlinear material used as core in a hollow waveguide was useful for generating ultrafast broadband radiation. In both cases, stable and spectrally broadened radiation was effectively generated because of the extended interaction length and the high power excitation attainable with the liquid-core fiber. 

The flexibility offered by liquid-core fibers for nonlinear interactions is not only limited to the capability of selecting a suitable liquid, but it is also possible to choose a proper core size to manage high powers. This is particularly important for avoiding problems related to laser-induced breakdown of the liquid.

In this sense, the use of these waveguides allowed for detailed experimental analysis of effects such as stimulated Raman scattering \(\text{(SRS)}\), stimulated Rayleigh wing scattering \(\text{(SWS)}\), stimulated Raman–Kerr scattering \(\text{(SRKS)}\), and parametric generation of radiation.

Several liquid samples have been used, yielding superbroadened radiation even with tuning capabilities. The results for the liquids used in these examples can be summarized as follows: (1) \(\text{CS}_2\) and toluene show superbroadening on both the pump line and the \(\text{SRS}\) lines, (2) benzene exhibits superbroadening predominantly on the \(\text{SRS}\) lines, and (3) carbon tetrachloride shows no appreciable broadening on any of the lines.

These experiments, together with a theoretical model, yielded enough information to determine that the threshold and the specific spectral distribution of the superbroadening effect depend on the molecular structure of the liquid.

The Raman-Kerr scattering process is perhaps the most widely studied optical effect with liquid-core fibers. Upon theoretical modeling, it has been possible to identify the requirements to observe this process: (1) The molecules of the liquid Kerr medium must be anisotropic, (2) liquid samples must be transparent or have a small loss factor, (3) pump intensity must be sufficiently high, and (4) the gain length must be long enough. All these requirements can be fulfilled with a liquid-filled hollow fiber.

Kerr liquid-filled hollow-fiber system can, thus, be used as a broadband, multiwavelength coherent light source with spectral and temporal feature that can be tailored for specific requirements.

A clear example of this is shown in He et al. in which the liquid and the hollow fiber are selected to enhance the multiorder \(\text{SRS}\) and SKS that yields the superbroadening effect. In this particular case, \(\text{CS}_2\) was recognized as the most efficient liquid for generating \(\text{SKS}\) and \(\text{SRS}\), whereas the optimized fiber parameters were \(\text{ID}\) 0.1–0.25 mm, length 3–4 m, and focusing length for the input coupling of 10–15 cm.

The total output spectral range was measured to be up to 4000 \(\text{cm}^{-1}\)(i.e., more than six orders larger that the typical \(\text{SRS}\)). The location of the output spectrum depends on the frequency of the pump source and can be in the near \(\text{UV}\) (300–400 nm) to the near \(\text{IR}\) (0.7–2.0 microns). Spectroscopy is clearly one of the main applications that can benefit from further development of light sources based on these optical effects.

Raman spectra of liquids dissolved in \(\text{CS}_2\) can also be obtained with liquid-core fibers, although a wider variety of liquids can also be analyzed upon using a Teflon-coated hollow fiber. It has also been demonstrated that when compared to conventional measurements using cuvettes, the use of a fiber geometry effectively enhances the Raman bands obtained from liquid samples.

Liquid-core fibers have been further used for other applications involving nonlinear optical effects. These include the amplification of amplified spontaneous emission \(\text{(ASE)}\) signals, studies of spectral narrowing of stimulated scattering, and optical limiting of laser pulses. Whereas the first two applications are based on the Raman–Kerr scattering effect, limiting of laser pulses is based on the nonlinear absorption processes that occur in the high-index liquid used as fiber core.

The choice of liquid depends on the pulse duration and an opaque cladding and organic liquid are generally used for this kind of application. Optical limiting action has been attributed to thermal density effects such as self-defocusing, wide-angle nonlinear scattering, and input coupling and propagation mode losses.

Medical Applications

Optical fibers have been used extensively in medical application for diagnostics and laser delivery. Among other features, medical diagnostics require the development of reliable sensors with in situ monitoring capabilities, which in turn implies that noninvasive measurements are also required.

Several fiber optic sensors have been proved to be useful for this application, and although a wide variety of materials have been used, most of the reports involve the use of glass fibers.

As seen in the previous, liquid-core waveguides have been proved to be useful for analytical chemistry, so the realization of liquid-core fiber sensors for in situ measurements and medical diagnostics should be considered feasible.

The realization of an instrument based on capillary optrodes has been already reported. An inexpensive setup was shown to be useful for analyzing small liquid samples and the instrument was sought to be useful for emergency medicine. The capillary optrodes were constructed using the commonly used glass capillaries for blood sampling from a pierced fingertip.

The sensors were sensitized with polymers and fluorophores and each of them could, therefore, be targeted to detect a different analyte (Fig. 5). The principle of operation of these waveguide optrodes is the fluorescence generated in the polymer coating trapped and guided by the wall of the capillary; thus, it can be seen as the equivalent to an evanescent waveguide sensor.

Up to three different analytes were detected per sensor because of the capability of coating the inner surface with three different polymers. The realization of compact systems based on other analytical techniques should also be possible with compact light sources and detector arrays.

Laser delivery waveguides have been extensively studied for laser tissue engineering and therapy applications. The wide range of wavelengths required for the broad variety of medical applications have proven to be one of the main challenges for the design and fabrication of suitable optical fibers.

As a general rule, wavelengths comprised within the visible and the near \(\text{IR}\) use silica fibers as waveguides, whereas the \(\text{UV}\) and mid-\(\text{IR}\) portions of the spectrum require specialty fibers. There is, however, a report on the use of a liquid-core waveguide for laser tissue ablation using a visible laser, showing that high-peak

power transmission is feasible with this arrangement. Although this ‘‘optical catheter’’ was demonstrated in in vivo experiments for laser angioplasty, no further improvements on these types of devices have been reported and silica fibers for this spectral region seem to remain the preferred choice.

Regarding the \(\text{UV}\) wavelengths, the use of liquid-core fibers for laser delivery in medical applications has not been reported. Nonetheless, as seen before, these waveguides have proven to operate in this spectral region. 

Wavelengths in the \(\text{IR}\) spectrum can be guided by several types of optical fibers that have proven to be useful for laser delivery. These include hollow waveguides, \(\text{IR}\)-transmitting glasses, and crystalline fibers. 

However, complicated fabrication processes, high sensitivity to bending losses, and a low damage threshold are, until now, the principal factors limiting these fibers. As an alternative, several studies have focused on the use of liquid-core fibers for laser delivery systems. Besides being an inexpensive option, liquid-core fibers offer attractive features such as variability in diameter, high flexibility, and mechanical stability.

\(\text{IR}\) absorption effects of water in the core and permeation of atmospheric water and of the solvent through the cladding have been reported, leading to fiber designs suitable to operate at \(2.94\mu m\). For this wavelength, carbontetrachloride \(\text{(CCl}_4)\) has been used as core with plastic tube and quartz capillary as cladding. Bending radii below 10 mm are possible and a minimum transmission loss of 2 dB/m can be reportedly achieved.

Further studies have also shown that because of an overlap of the refractive indices of \(\text{CCl}_4\) and fused silica between 500 nm and \(1\mu m\), laser wavelengths in this spectral range (such as those obtained with \(\text{Nd}\)/\(\text{YAG}\) and \(\text{HeNe}\) lasers) cannot be transmitted in this fiber.

However, upon using a mixture of \(\text{CCl}_4\) and tetrachloroethylene \(\text{(C}_2\text{Cl}_4)\), the fiber becomes transparent from the near-\(\text{UV}\) (380 nm) up to the near IR (3 mm), and consequently, it is suitable also for the \(\text{Nd/YAG}\) laser.

Distal energy densities up to 30J/\(\text{cm}^2\) have been achieved, thereby exceeding the ablation threshold of soft tissue. Thus, minimally invasive surgery can potentially be carried out with the aid of these fibers. 

Special Waveguide Structures and Devices with Liquid Cores 

Thus far, we have reviewed applications involving fibers with simple or diluted liquids and capillary tubes. However, there are other materials and waveguide structures that have increased the applications and driven further research involving liquid-core fibers.

As an example, the use of liquid crystals has proven to be useful for fabrication of polarization-sensitive fiber devices. Experimental studies have shown that liquid-crystal core fibers with elliptical geometry yield adequate polarization properties for sensing and communications applications.

Fabrication of fiber devices such as long-period gratings based on liquid-crystal core fibers has also been demonstrated, showing that band rejection filters can be fabricated using this type of waveguide.

Adequate control of the optical properties of liquids is, thus, extremely important for dynamically adjustable fiber devices. Advances in physics of fluids, and in particular in the development of electrical and magnetically controlled liquids, will continue to spur future developments of devices based on liquid-core fibers.

As we have seen in the previous sections, the waveguide structure also plays an important role in the guiding properties of liquid-core fibers. In this sense, the development of complex waveguide structures has renewed the interest of using liquids as core materials.

Theoretical studies have shown that a micro-structured silica–air cladding provides excellent confinement for light guided in a liquid core, provided that the average cladding index is sufficiently below the index of water. Realization of such fiber is, thus, limited by the fabrication of the micro-structured cladding, an engineering problem that most likely will soon be addressed successfully.

Similar structures have already been successfully demonstrated with water and ethanol, so the development of compact biosensor, pollutant monitors, and chemical sensors based on liquid-core micro-structured fibers is successfully underway.

Moreover, dynamic control of other arrangements based on liquid-core and liquid-cladding waveguide structures has been successfully demonstrated. Reconfigurable optical switches, modulators, and optical couplers should, therefore, be possible with all-liquid optical waveguides, thereby increasing the usefulness of light guides for sensing and communications applications.