The original microfluidic fabrication technologies were derived from silicon microelectronics, which were well developed in semiconductor industry. However, these techniques are very expensive, complicated, and time consuming. In addition, silicon is not suitable to be applied in a microfluidic device due to its opacification to the visible and UV light in addition to its high cost. Feasible techniques for the fabrication of microfluidic devices include microelectromechanical systems (MEMS) with its procedures illustrated in Figure 1. During the fabrication, a material, especially a polymer, is deposited on a substrate first, and then a pattern in a master is transferred into the material by lithography. After an etching process (wet or dry etching), either the exposed or unexposed material is removed. Finally a cover is attached on the surface of the chip to enclose the microchannels.

Another widely adopted method to fabricate a microfluidic system is casting, in which a soft polymer elastomer of high optical transparency is used, i.e., poly(dimethylsiloxane) (PDMS) or poly(methyl methacrylate) (PMMA). Figure 2 shows the steps for the fabrication of a microfluidic device by casting. A mold is produced by soft lithography [12,13] or laser fabrication [14–16] in a photoresist layer (SU-8) or a metallic sheet. Hot embossing technique [14,16,17] or poured molding method [15,18–20] is used to duplicate the mold in a polymer sheet. The peeled polymer replica is sealed to a flat surface to enclose the channels. Much complex structures in a microfluidic device can be developed by stacking multiple polymer layers (100 μm in thickness per layer), similar to a sandwich structure [18,21]. The time period is less than two days starting from design to the realization of a functional device. A list of materials used in the fabrication of opto-microfluidic devices with different molding fabrication techniques is presented in Table 2, which shows that PDMS is the most popular material used in poured molding method and PMMA is widely used in hot embossing technique. In some cases, optical components, especially optofluidic waveguides, have been integrated into the chips and their effectiveness in the transmission of light has been demonstrated. Examples include solid core/solid cladding waveguides [22,23], solid core/liquid cladding waveguides, liquid core/solid cladding waveguides [24], liquid core/liquid cladding waveguides, and hybrid waveguides [18,25,26].

For the applications in optofluidics, the drawbacks of using polymer materials such as solubility in many common solvents, damage upon tightly focused laser irradiation, and fluorescence at certain common wavelengths appear to be severe. In contrast, the glass-based devices with high optical transparency to visible light and inertia to chemical solvents are very suitable for these applications. Consequently, new fabrication techniques are required to realize microfabrication in various glasses. In this regard, ultrafast laser microfabrication, which utilizes ultrashort laser pulses of pulse width as narrow as ∼100 femtoseconds (a femtosecond laser), has been revealed as a powerful approach to fabricate opto-microfluidic devices in glasses [27]. Figure 3 shows a schematic illustration of a home-made femtosecond laser microfabrication station established in our laboratory at the Memorial University of Newfoundland [28]. By shining a piece of glass with a focused femtosecond laser of a peak intensity well above the damage threshold of the glass, three-dimensional microstructures could be generated in the glass. Figure 4 shows a femtosecond-laser-microfabricated microchannel in a fused silica observed by an atomic force microscope (AFM), indicating widths at the top and bottom of the channel of ∼5 and 2 μm, respectively.

After hydrofluoric (HF) acid etching for 3–5 hours, microchannels are generated, which can be adopted in opto-microfluidic devices. The procedures for the fabrication of opto-microfluidic devices with a femtosecond laser are shown in Figure 5(a). Examples of opto-microfluidic channels and microstructures fabricated in our group are shown in Figure 5(b,c).

Compared with MEMS techniques, the ultrafast laser microfabrication possesses salient advantages of simplicity, high efficiency, and versatility. In addition to the fabrication of microchannels, optical waveguides can also be fabricated by an ultrafast laser [29–31]. When a femtosecond laser beam with pulse energy of several hundred nano-Joules is irradiated on the glass, the refractive index of the region in glass will increase. This type of waveguide is called a Type I waveguide. Another type of waveguide, i.e., Type II waveguide, can be fabricated with higher peak energy. Laser irradiation with peak intensity higher than the damage threshold of the glass causes the material at the focus to vaporize, thus the high pressure pushes the material to both sides of the irradiated region. Consequently, the tracks at the two sides result in a higher refractive index while the region between the tracks guides light. In this case, the distance between the two tracks should be shorter than 50 μm. Figure 6 is a Type II waveguide coupling with a He-Ne laser at 633 nm. The propagation loss of this kind of waveguide is ∼4 dB/cm, which is somewhat higher than the values of waveguides fabricated by other techniques. However, the propagation loss in an optical waveguide fabricated by a femtosecond laser could be reduced after optimization of the fabrication parameters including the laser properties. The relatively higher propagation loss does not have a serious effect for the adoption of these waveguides in opto-microfluidic devices.

Table 3 lists a comparison of the properties of some reported optofluidic waveguides fabricated by different techniques. Among three leading fabrication technologies, i.e., MEMS, casting, and femtosecond laser microfabrication, the resultant optofluidic waveguides have quite different light propagation properties. It is important to keep in mind that the propagation loss of a waveguide is relative to the material property of the core and cladding (refractive index) and the carried wavelength. For the waveguides fabricated by MEMS or casting techniques, their core and cladding materials can be either liquid or solid. The reported propagation losses of the waveguides fabricated by MEMS techniques range from 0.5 to 4 dB/cm while the values for the waveguides fabricated by molding fall in a range of 1.8–8.2 dB/cm. The femtosecond laser microfabrication technique is a powerful method to fabricate opto-microfluidic waveguides, in particularly on glasses. With the Type I or II waveguides fabricated with different peak intensities of the lasers, different propagation losses of the waveguides have been achieved, in which values around ∼1 dB/cm were found for the Type I waveguides and slightly higher propagation losses for the Type II waveguides.

Any of the aforementioned microfabrication techniques is effective to fabricate opto-microfluidic devices with pros and cons. Table 4 compares the advantages and drawbacks of different microfabrication techniques used for fabricating opto-microfluidic sensors. Judicious selection is necessary in order to choose a suitable technique for a specific application. In some cases, special approach may be required to overcome the drawbacks of specific microfabrication technique.

The original motivation for developing opto-microfluidics was for biological analysis, in which immunoassay and DNA separation have been the two main applications. Immunoassay is a biochemical technique that detects the presence and quantities of antibodies or antigens in the samples [32–34]. Competitive immunoassay and sandwich assay are two common methods [28]. When a labeled (such as an enzyme or fluorescent dye) analytical reagent with special antigens or antibodies is mixed with the sample, a specific binding will form between the antigen and the corresponding antibody. The changes in the color and intensity of light from the labeled binding are recorded to identify the immunity of the sample. The traditional technique requires a large amount of reagents and a long incubation time. Microfluidics-based immunoassays have been developed by several research groups. For instance, in a sandwich assay for antigen, antibody is injected into the microchannels by pressure-driven [35], electrokinetic control [36,37] or centrifugal force [38], and deposited onto the wall of the microchannels to form solid-phase antibody, then antigen in the sample is mixed with solid-phase antibody in the microchannel, and finally another labeled antibody is added for signal generation. The color signals from the antibody-antigen-labeled antibody bindings are detected by a fluorescence microscope after removing excess antibodies and antigens through washing steps. All steps can be completed in one hour, and the reagent consumption is only on the order of microliters. Roos et al. [39] and Herrmann et al. [40] used microspheres to support the binding, and the fluorescence signals were collected from these microbeads. These microspheres provided increased surface area to support the immune complex. Nanoparticle-labeled microfluidic immunoassay [41] was reported by Lin et al., in which free labeled antigens were deposited on the nanoscaled gold particles and the scattered light from the particles provided readout to trace the bindings between antibodies and antigens.

As a prerequisite step to allow opto-microfluidic measurement, capillary electrophoresis (CE) is a rapidly growing separation technique, which has been applied in bioanalysis, environmental pollutant analysis and food analysis [42,43]. By using high electric fields (larger than 500 V/cm), buffer solution and all ions, positive or negative particles in the sample generate electroosmotic and electrophoretic flow respectively. Electroosmotic flow pulls the analytes in the buffer solution through the capillary toward the cathode. However, the electrophoretic flow reduces the flow rate of the negatively charged analytes and increases the flow rate of positively charged analytes in the capillary. Therefore, the analytes separate due to different mobilities. Ultraviolet (UV) absorption detectors are equipped to detect the ingredients of the sample near the outlet of the capillary. The data from the detector is displayed as an electropherogram, in which peaks at different times are shown for separated chemical compounds.

A CE instrument is formed after assembling the microchannels and reservoirs in a chip with extra high potential. Figure 7 illustrates two types of CE chips and their operations. The CE chips are extensively used in medical research such as DNA analysis [44], infectious disease diagnostics [45] and sample purification [46]. Now some companies such as Micronit Microfluidics BV in The Netherlands, microLIQUID in USA, and Micralyne Inc. in Canada supply CE chips and kits for clinical diagnostics. Meanwhile, several research groups have integrated different components into the CE chips to satisfy diverse needs of analysis. For example, combination of polymerase chain reaction (PCR) with CE chip (PCR-CE chip) is a focus of research, in which one or multiple PCR chambers are integrated into the CE chip to incubate and analyze DNA simultaneously [47–52]. Prakash et al. fabricated PCR arrays on the CE chip with two PCR valves and ports to control the flow of samples in each PCR chamber, so that the CE chip can be reusable without contamination [53]. CE chips integrated with multichannels for mixing, reaction and separation were also reported by some Canadian research groups [54–57]. Munce et al. used CE with optical tweezers to achieve single-cell sorting and analysis in one chip [58]. A CE chip with an acoustic wave sensor successfully trapped one live myocyte in the acoustic wave sensor region [59], in which the contraction and relaxation of the cell were clearly detected without the influence from other cells.

The commonly adopted detection method in a CE chip is a confocal detective system. Bliss et al. integrated optical waveguides into a CE chip [60]. However, the electropherogram demonstrated that the waveguide detection possesses an equal or greater sensitivity than that of the confocal system. A good technique to accurately identify the ingredient of a sample is to connect the outlet of the capillary to an electrospray mass spectrometry (ESMS) [52], as shown in Figure 8.

Polymer particles are highly sought in biology, cosmetics, food process, medicine, and pesticide. However, the incomplete control and high cost of traditional chemical synthesis techniques limit their applications. One research group at the University of Toronto proposed a novel approach to realize continuous and scalable production of core-shell droplets and polymer capsules in microfluidic devices in 2005 [61]. Two or more different kinds of immiscible liquids were infused into a microfluidic device and the mixed liquid flew through a thin “focused” orifice. Different core-shell droplets were synthesized by changing the liquid flow rates, which were controlled by push pumps. UV lamp irradiation made the photopolymerizable droplets solidification when droplets flew along a long wavy microchannel. Polymer particles were then obtained at the output of the device. This technique could precisely control the sizes of the particles, liquid cores, shells, and the number of core droplets. During the subsequent years, further research had been carried out by the same group to address additional topics, such as biopolymer microcapsules [62], Janus and ternary particles [63], microgels of biological polymers [64], monodisperse particles with microspheres, rods, disks and ellipsoids [65], dynamic lattices [66,67], and colloidal particles at gas–liquid interfaces [68]. This kind of chemical sensors could better explore the dynamics and mechanisms of physical and chemical reactions which are difficult to be investigated in conventional experiments.

Centrifugal microfluidic devices were reported by a research group at McGill University, in which multiple chambers and capillary valves were fabricated in a disc. The flow of the solvent was controlled to achieve dissolution [69], mixing [70], extraction [21,71] and reaction [70] between reagents by adjusting the rotational rate of the disc. The components and concentrations of analytes were measured at the detection unit by fluorescence and absorption spectra. In some cases, magnets were utilized in the chamber and base to serve as an agitator to increase the efficiency of chemical reaction. The centrifugal microfluidic devices could be applied in environmental monitoring and soil analysis for their portability and rapidity.

Optofluidic platform for chemical component analysis by laser-induced breakdown spectroscopy (LIBS) was achieved by a research group at the University of Alberta [72,73]. A small thermal or piezoelectric actuator was deposited on the bottom of a microchannel while an orifice of a few microns was opened on the opposite wall of the actuator. During the operation, a bubble was first excited at the orifice by the actuator, and then a fast intense laser pulse broke down the bubble, and thus plasma with the components of reagent was generated. As the plasma was cooled down, electrons and ions were recombined, accompanied by emitting electromagnetic radiation with fingerprint wavelengths of the elements in the reagent, which was detected by its spectroscopy. The most notable advantage of this technique is the non-contact measurement, which completely avoids sample contamination.

Surface plasmon polaritons (SPP) are electromagnetic waves that propagate along a metal and dielectric interface [74,75]. SPP can be evanescent, which are excited by both electrons and photons. The excitation of SPP by photons is usually referred as surface plasmon resonance (SPR). SPP cannot be directly excited by photons due to momentum mismatch. Special arrangements, such as Otto configuration [76], Kretschmann configuration [77], or a diffraction grating [78], are well-known techniques to couple photons into SPP in order to match the wave vectors of the photon and the surface plasmon, as illustrated in Figure 9. The Kretschmann configuration is the most common approach, in which the thickness of the metal layer is usually a few tens of nanometers to ensure the evanescent wave to travel through the metal and couple to a surface plasmon mode, which is shown in Figure 9(d).

Since the wave travels on the interface of the metal and the external dielectric layer, SPR is very sensitive to any change at this interface, such as the density and temperature of the dielectric layer and the structure of the metal surface. Opto-microfluidic SPR sensors for immunoassay and refractive index measurement have been demonstrated. Chien et al. integrated SPR into a microfluidic platform [79], in which diffractive mirrors were attached on the prism to focus the incident light in the direction of wave vectors while liquid solution as the dielectric material was infused into a gold-coated microchannel. When light with different incident angles is irradiated on the surface of the prism and metal, part of the light, which matches with the SPR angle θ_SP, is attenuated into the metal to generate SPR and the rest of the light is reflected. A photodetector can be placed at the end of the device to capture and monitor the intensity of the reflected wave. The reflected light of reduced intensity at the SPR angle would be detected due to the optical absorption by the metal. The sensor has been proved to be effective in measuring refractive indices of liquids by monitoring changes in the SPR angles with NaCl solution as a testing material.

Nanohole based SPR sensors have received considerable attention in recent years. One of the most important characteristics of a nanohole array is that the intensities of the transmitted light enhance at certain wavelengths due to the SPP coupling when the incident light irradiates on the surface of the metal film. By monitoring the wavelengths of the intensity peaks, a nanohole-based SPR sensor is realized. Compared with a standard reflective mode SPR sensor, a nanohole SPR sensor operates at the normal incidence without the necessity to consider the SPP angle. Sinton group at the University of Victoria realized nanohole array SPR sensors [80–85]. For example, a device with flow-through nanohole arrays, instead of dead-ended nanohole arrays in a SPR sensor, could make the regents to travel through the nanoholes rapidly, thus reducing the response time remarkably [85].

With the improvement of SPR techniques, surface plasmon resonance imaging (SPRI) has been proposed as a new kind of detection technique in biology [86]. Kanda et al. at the University of Alberta patterned a surface of metal (gold) with free-labeled protein antigens arrays [87]. When the sample flew through the microchannel located above the patterned gold film, the antigen-antibody bindings were generated at the arrays. A high contrast SPRI based on the adsorbed proteins was realized to evaluate the quantitative and qualitative properties of the antibodies in the sample. Tabrizian group at the McGill University replaced the plain gold film in the SPR with periodic gold nanoposts to detect DNA hybridization. The optimal result showed a fivefold SPRI enhancement compared with the common SPRI [88]. This group also combined SPRI technique with digital microfluidics (DMF) to detect biological samples [89,90]. DMF is a fluid manipulation technique, in which a patterned array of electrodes is etched on the substrate with MEMS techniques and then coated with a waterproof material like Teflon to form the DMF. Droplets can dispense and merge in the DMF when an electrical potential is applied on sequential electrodes in the array [91–95]. Therefore, this combination realized real-time monitoring and detection of reaction occurring as required.

Integrating micro-optical components into a chip could effectively shape beams to enhance the sensitivity of a sensor and increase its portability. Besides optical waveguides discussed in the previous sections, Table 5 lists several micro-optical components integrated in opto-microfluidic devices, which include lenses [96,97], gratings [22,98], mirrors [79,99], and light sources [100,101]. There is continuing effort to realize integration of multiple components to acquire novel optical functionalities.

Optical tweezers are important instruments to trap and sort particles like cells, proteins, and microspheres. A tightly-focused laser beam (Gaussian beam) provides an attractive or repulsive force on the particles due to the changes in the momentum of light upon reflection or refraction. Although the force is on the order of piconewtons, it is large enough to hold and move small objects with sizes of several tens of microns. Taking advantages of the merits of opto-microfluidic devices with narrow channel widths, fiber-microfluidic tweezers have been particularly attractive in studying a variety of biological systems. Sinton group realized trapping microsphere array by dual-beam with a focus on the study of optohydrodynamics and optical interactions between particles, which could be applied for contact-free storage of biological cells as well [102–105]. In addition, Guo et al. at the University of Ottawa fabricated a microfluidic chip with on-chip lens structures to reduce the beam waist of the light and achieved a higher efficiency in optical operation and optofluidic transportation [96].

Microcytometer is another important application of opto-microfluidic sensors. Azmayesh-Fard et al. [18] coupled excitation lasers of different wavelengths into the microfluidic devices by solid-core waveguides and different dyed particles could be resolved from different peaks after a windowed Fourier transform to output signals. Multiple parallel waveguides were integrated into an opto-microfluidic sensor to collect the fluorescent and scattered signals from the labeled cells by Xu group in McMaster University [106]. The same group also fabricated micro-lenses in the chips to shape the excitation beam and improve the signal-to-noise ratio of cytometers [97]. A wide-angle microfluidic cytometer was reported by a research group at the University of Alberta [107–109]. A conventional cytometer can only collect signals through small-angle forward scattering (5°) and side scattering (10°) in which errors are apparent when cells gather or irregular cells and other organelles exist. In the wide-angle microfluidic cytometer, the goniometric measurement and finite-difference time-domain (FDTD) method were adopted to overcome the disadvantages of the conventional cytometers.

Refractive index (RI) sensing is another important application of opto-microfluidic sensors, which have been widely applied in environmental monitoring and optical measurement. Opto-microfluidic RI sensors based on Mach-Zehnder interferometer (MZI) [23,110,111], grating [98,112], and refractometer [99] have been reported by several research groups in Canada. All these reported methods adopted similar principles in which two identical beams were first coupled into different paths such as a solid core waveguide or a liquid core waveguide, then interference occurred due to the difference in the optical paths when the two beams were combined together. The measurement of the refractive index of the liquid can be achieved by monitoring the change of the interference intensity.