Microdroplets-based SERS-microfluidic System

1. Background
   1. Introduction

Online quantified detection of samples containing a complex mixture of compounds in picoliter volumes is an emerging research topic in the food, environmental and biomedical detection areas. The analytes dispersed in various phases, such as gas, organic and aqueous, raise a more challenging task for online multiplex detection. This problem is amplified when trace analytes detection is required, as signals from background molecules can swamp the signals of the analytes.

Raman scattering spectroscopy has a great promise to solve these problems, which is a unique technique to obtain information on the chemical structure of molecules. Discovered in 1928 by the Indian Physicist C. V. Raman, Raman Scattering can offer unique signal information for every particular chemical bond, which offered great promise in the detection and analysis field. However, even though general Raman Scattering can provide rich information, it was not as lucky as Fourier-transform infrared or UV/Vis spectroscopy, to have been widely developed into practical applications, because of its inherent limitation “feeble signal”. The weak signal is due to the extremely small Raman scattering cross section for chemical bonds of most molecules. For example, the benzene molecule, which has relatively strong Raman scattering signal, has 12 to 14 orders of magnitude lower signal intensity than typical fluorescence signal intensity. This weakness had long restricted the applications of Raman Scattering, until the phenomenon of surface enhanced Raman scattering(SERS) was discovered.

50 years later after Raman Scattering was discovered, surface-enhanced Raman scattering(SERS) was discovered by scientists, opening a new gate for Raman spectroscopy technology. By using noble-metal (Ag or Au) nanoparticles and their intense electromagnetic field upon incident excitation, SERS is an extremely sensitive technique that can be tailored to provide the detection of specific analytes through their unique vibrational fingerprints of molecules, even at ultra-trace concentration. Moreover, the narrow linewidth of SERS spectra allows for multiple-analyte detection within complex mixtures, including detection down to the single molecule level. Due to the above advantages, SERS gradually drew attention of optical and material scientists all around the world, especially after 1990s, when nanomaterial science started to experience a dramatic expansion. This spark is due to varieties of factors, such as development of molecule detection, nanotechnology, instrumentation capabilities and the demanding needs for health care, safety, and anti-terrorist threats.

At the same time, microfluidic technology, which is famous for its high controllability on emulsions, has potential to provide steady platforms for SERS substrates generation and stable online analytes detection, helping broaden the applications of Raman scattering technology. Recently, much progress in microfluidic systems has been achieved, gaining numerous advantages such as high efficiency, low dosage, high safety, and high sensitivity compared with conventional macroscopic systems. As high-throughput analytical microscale devices, microfluidic systems show great potential for the high-sensitivity detection of various chemical and biological molecules. The combination of microfluidic and SERS detection has many advantages over conventional SERS detection. This is because the measurement under flowing conditions prevents problems of variable mixing times, scattering geometry, localized heating, and photo dissociation. In addition, due to the advantages of microfluidics, some special research targets could be achieved, such as high throughput single cell screening.

This review mainly focuses on microfluidic-SERS combination technology. Firstly, the basic theory of SERS (part 2) and the brief introduction of microfluidics (part 3) will be presented. And two main methods to combine microfluidics and SERS technology (part 4), as well as some applications of SERS-microfluidic combination (part 5) will be roughly reviewed. Finally, based on my current project and some perspectives, the outlook of the microdroplets-based SERS-microfluidic system will be discussed (part 6).

2.  SERS
   1.    Theories of SERS

Though the mechanisms for surface enhanced Raman scattering by metal nanostructure have long been under debate, there were two main theories standing out and being widely accepted, which are electromagnetic enhancement theory and chemical enhancement theory¹.  In order to understand the enhancement mechanisms, it is instructive to refer to the following equation²:

P = αE

where P is the induced dipole moment,

αis the molecular polarizability, and E is the incident electric field. The Raman scattering intensity I is proportional to the square of the induced dipole moment

I ∝P2

Raman enhancement can take place either by increasing the electric field experienced by the molecule (electromagnetic enhancement) or by changing the molecular polarizability of the adsorbate (chemical enhancement). The electromagnetic enhancement mechanism is explained by a phenomenon known as surface plasmon resonance³. Surface plasmons are oscillations of conduction band electrons at a metal surface. At the surface plasmon resonance frequency, conduction band electrons move easily, producing a large oscillation in the local electric field intensity. The surface plasmon frequency strongly depends on the surface morphology (size and shape of particles), the dielectric properties of the metal, and the wavelength of the incident light. Electromagnetic effects are known to decrease as a function of

1/r3from the surface. Chemical enhancement involves bond formation between the analyte and the metal surface. This bond makes it possible to transfer charge from the metal surface to the adsorbate molecule, and this effect increases the molecular polarizability of the molecule. There have been many experimental demonstrations that both mechanisms play key roles in SERS effects. However, it is generally believed that electromagnetic enhancement may have a greater part to play than chemical enhancement⁴. The main analytical advantages of SERS are enhanced sensitivity, surface specificity, and fluorescence quenching by metal nanoparticles. Furthermore, it is possible to detect multiple analytes simultaneously using SERS because its signals are much narrower than those of fluorescence bands.

2.    CB and SERS…

1.3 Microfluidics

Microfluidics is becoming increasingly attractive in analysis field because microfluidic devices require only small amounts of sample, but offer highly precise controlled manipulation of samples within microscale channels⁵. The small size of these devices provides the potential of creating portable lab-on-a-chip devices that facilitate the low detection limits. The field of microfluidics emerged in the early 1950s⁶ with the development of ink-jet technology⁷. Later, silicon wafers and Si–Pyrex technology were successfully used for the miniaturization of gas chromatograph and high-pressure liquid chromatography⁸⁹ by considering the advantage of the fluid propulsion inside microchannels. Nowadays, the field of microfluidics has evolved from simple micro-structures into complex devices applying various technologies simultaneously, such as mixing fluids¹⁰, pumping liquids ¹¹, and individually culturing cells¹² . However, in practical cases, some limitations still exist in the variety of micro-scale components, such as the slow processing operation and expensive manufacturing costs. An alternative way is to evolve microfluidics by combining with other technologies to create a new technology platform to have more advantages.

Microfluidic platforms that integrate optical and spectroscopic analysis, exploiting both absorption and scattering techniques, have been widely reported, and the functionality of microfluidics is significantly expanded when they are combined with those optical techniques. For example, Huang¹³ presented a review for optofluidics-based techniques focusing on single cell manipulation, treatment, and detection, respectively. Among them, SERS has become an emerging technology of choice for the detection of analytes at the cellular and molecular levels. The integration of microfluidics with SERS provides not only a new technology platform but also a transformation to a new paradigm of bio-sys-on-a-chip.

Even though SERS can be used as highly sensitive sensor, it has always been a challenge to obtain repeatable and reproducible SERS signal results under complicated experimental conditions. The degree of metallic particle aggregation, the different size of metal colloids, and the inhomogeneous distributions of molecules on the metal surface all affect the SERS signal reproducibility. Lately it has been found that the combination of SERS and microfluidics could overcome these major drawbacks for SERS-based techniques. More reproducible SERS signals can be obtained in a microfluidic channel, because of the more consistent geometries and the favourable heat dissipation inside microfluidic devices. Furthermore, the integration of microfluidic devices and SERS minimizes the sample volumes, the recording time, and damage to the biological analytes. In addition, the microfluidic-SERS systems consolidate sample preparation, target manipulation and separation, and in situ detection into one working unit, and can significantly improve the effectiveness of these sensing platforms²⁴.

The most common microfluidic devices are made from polydimethylsiloxane (PDMS) and other polymers. They can be made at low manufacturing costs by moulding techniques, such as photolithography¹⁴, and soft lithography¹⁵ with the same layout prototype master. However, PDMS is a Raman active material, which has its own Raman signal. To overcome this problem, confocal Raman microscopy is always used. In confocal Raman microscopy, a confocal slit is designed to collect only the laser-induced signal from a distinct focal volume within the diffraction limit¹⁶. With confocal Raman microscopy, the Raman signals from the PDMS channel itself are effectively removed, and then the target spectra from a small volume inside the microfluidic channel can be obtained¹⁷.

4.  Microfluidic-SERS Combination

There are mainly two methods to combine microfluidic technology and SERS detection. The first method is to mix the analytes and SERS active nanostructures (colloidal nanoparticles) in microfluidic devices. The second is mainly to synthesis or implement SERS-active substrates inside microfluidic devices.

1.4.1 To mix the analytes and SERS active nanostructures in microfluidic devices

For this method, a reproducible mixing performance is a critical need for a successful microfluidic-SERS platform. Microfluidic devices are usually designed to work with liquid samples. Hence, most microfluidic-SERS systems utilize colloidal nanoparticles as SERS-active substrates. Generally, the nano probes are injected through the microfluidic channels where they encounter analytes at a certain location for SERS detection. A successful SERS application to microfluidics requires several key steps, such as an instrumental setup of the Raman microscope, a special design for the microfluidic channel, a synthesis of stable metal nano colloids, and an optimal flow rate control.

To achieve efficient mixing performance between metal nano colloids and analytes in the microfluidic channel, passive mixing channels are most widely used in laminar flow because of their simplicity and operational applicability. Figure1 shows several different passive channel designs for the fast and efficient mixing of two confluent streams¹⁸. Figure 1a is a simple diffusive mixing across fluid interfaces¹⁹. Efficient chaotic advections can be achieved by designing different channel structures. Figure 1b shows a zigzag-shaped channel for chaotic mixing at high Reynolds numbers²⁰. Figures 1c and d show 3-D serpentine channels for chaotic flows at low to intermediate Reynolds numbers¹⁰²¹. Figure 2e shows a groove channel for chaotic mixing at low Reynolds numbers²².

Figure 1. Microfluidic channels for mixing continuous flows. (a) Mixing of two miscible fluid streams under laminar flow conditions. The component streams mix only by diffusion. (b) Zigzag-shaped channel for chaotic mixing. (c) 3-D L-shaped channel. (d) 3-D connected out-of-plane channel. (e) Staggered-herringbone grooves for chaotic mixing.

The alligator teeth-shaped microfluidic channel shows a high mixing efficiency compared with the other chaotic channels, because a strong chaotic advection is developed by the simultaneous vertical and transverse dispersion of the confluent streams²³. Figure2 shows a schematic illustration of an alligator teeth-shaped microfluidic channel for efficient laminar flow mixing. Analyte molecules are effectively

adsorbed on the surface of silver nanoparticles by traveling a zigzag-shaped microfluidic channel. As mentioned above, the on-chip Raman detection in a microfluidic channel yields more reproducible results than that of a static condition. In the microfluidic channel, the nano colloids and analytes are introduced into a channel where they are mixed with each other. Then, the Raman signal is accumulated from the flowing stream when the mixed samples pass through a laser beam. At the laser detection point, SERS signals from different aggregates are accumulated and averaged. As a result, the detection reproducibility is greatly improved. Quantitative analysis of analytes based on intensity changes of the SERS signal is possible with an efficient mixing channel design and optimal flow velocity.

Figure 2. Schematic illustration of the alligator teeth-shaped microfluidic channel. The con-fluent streams of silver colloids and trace analytes were effectively mixed in the channel through the triangular structures, which are located on the upper and lower surfaces of the channel in a zigzag manner.

More recently, a microfluidic channel has been used for serial high-throughput microanalysis²⁴. In this work, the SERS microscopic technique was applied to the detection and quantization of analytes in a liquid/liquid segmented microfluidic system. Using this technique, the “memory effects” caused by the photo deposition of analyte/ colloid conjugates to the channel walls could be overcome.

1.4.2 Integration of SERS active substrates in microfluidic devices

In the method stated in 1.4.1, controlling nanoparticle aggregation due to the interaction between nanostructures and channel surfaces will always be a big challenge to obtain high and consistent SERS enhancement signals. An alternative strategy, instead of injecting nano colloids into devices, is that SERS-active substrates be fabricated, aligned, and bound to microfluidic devices directly²⁵²⁶ . Nowadays, more and more on-chip SERS applications integrating SERS active substrates inside microchannels are being reported. Connatser²⁷ integrated nanostructures in to microfluidic devices by applying electron beam lithography on a polymer layer. Another approach for integrated substrates in microfluidics is presented by Banerjee²⁸ by introducing a nanoporous anodic aluminum oxide (AAO) layer of hexagonally packed holes on top of an aluminum substrate (figure 3). A microfluidic channel was placed on the substrate to provide the analyte solution and then achieve a high throughput. The choice of AAO as substrate matrix is due to its high biocompatibility as this platform is used to monitor protein binding to a lipid bilayer.

Figure 3. The schematics of the microfluidic channel on AAO (yellow). The latter is resting on an Al support (blue). Arrow points to the fluid channel and the direction of the flow.

Despite the fact that SERS substrates have been successfully integrated into microfluidic systems, there are still a lot of problems in compatibility and controllability in their implementation. In order to further improve the flexibility and stability of the integrated SERS substrates in the microfluidics, a number of novel techniques have been implemented for in situ functionalization of lab-on-a-chip platforms. Xu²⁹ applied a femtosecond laser technique to produce silver SERS substrates at the desired position of the microchannel and obtained the signal enhancement of ~108 for p-aminothiophenol (p-ATP) and flavin adenine dinucleotide (figure 4).

Figure 4. (a)Sketch of femtosecond laser fabrication of the silver SERS substrate inside a microfluidic channel (b) The application of this SERS monitor for target molecule detection at visible light excitation

Oh and Jeong³⁰ fabricated plasmonic nanoprobes with rich EM hot spots as SERS active substrates in microfluidic channels. The silver film substrate with an initial thickness of 30 nm was etched into nanotips and nanodots with oxygen plasma treatment, resulting in excellent SERS performance (figure 5). This integrated optofluidic device enables the label free and solution phase SERS detection of small molecules with low Raman activity such as dopamine at micromolar level.

Figure 5. A schematic illustration of the optofluidic SERS chip with plasmonic nanoprobes self-aligned along microfluidic channels. The perspective scanning electron microscope (SEM) image shows the plasmonic nanoprobes with nanotips and nanodots

The other widely used method is to implement SERS substrate within microdroplets with high uniformity in shape and size in microfluidic devices. At the same time, microdroplets are good precursors for micro-scale materials. The combination of SERS and microdroplets offer a good method for SERS micro-detector fabrication.  Abalde-Cela³¹ used microdroplets as reactors for the fabrication of agarose beads, densely loaded with silver ions, which were subsequently reduced into nanoparticles

using hydrazine. The resulting nanocomposite beads not only display a high plasmonic activity, but can also trap or concentrate analytes, which can be identified by SERS spectroscopy. Figure6 shows the detailed scheme of the beads fabrication process in microfluidic device. The research shows the SERS detector can be restricted to a single bead, thereby strongly reducing the amount of material needed for detection of the analytes.

Fig. 6 (A) Detailed scheme for the on-chip preparation of homogenous Ag+–agarose microbeads (inset). Optical images of the T-junction and flowfocusing system (B), and the chip channels and outlet containing the prepared microbeads (C). (D) Off-chip reduction of the silver ions into nanoparticles inside the beads.

5. Applications of Microfluidic-SERS Detection Technology

1.5.1 In situ monitoring of chemical reactions in a microfluidic channel

The development of synthetic chemistry using microfluidic devices has been of increasing interest in recent years. In particular, dangerous organic reactions can be carried out in a microfluidic channel in relative safety because of the small reacting volumes involved. The application of microfluidic reactors to industrial areas has been achieved by several researchers³²³³. However, highly sensitive detection schemes are important because the detection volume inside the channel is extremely small.

Fortt³⁴ investigated the formation of azo dyes via diazonium salts in a glass microfluidic channel using on-chip Raman spectroscopy. In this work, anhydrous diazotization and chloro- and hydro-deazotization of various arenes in microfluidic reactor systems could be successfully monitored using on-line Raman spectroscopy. It was shown that Raman monitoring is suitable for on-line reaction feedback and control. The advantages of on-line monitoring over conventional off-line GC monitoring and the advantages of on-chip Raman monitoring over capillary monitoring were successfully elucidated in this work. Lee³⁵ illustrated the applicability of confocal Raman microscopy by monitoring the imine formation reaction in a glass microfluidic channel. To monitor the diffusion process in a microfluidic channel, the Raman spectra were measured at various points along the channel with a constant flow rate. In addition, time-dependent Raman spectra were also measured at the fixed position 5 in figure 7 under a static condition in order to monitor the variation of the Raman peaks to complete conversion. In the imine formation reaction, the disappearance of the C=O stretching peak at 1700 cm^-1 for the reactant, benzaldehyde, and the appearance of the C=N stretching peak at 1628 cm^-1 for the product, an imine, were successfully monitored. Figure 7 shows that the concentrations of the reactants and the product are changed during the course of reaction. In this work, it was shown that Raman microscopy is a sensitive detection tool for the in situ monitoring of organic reactions in a microfluidic device. Lin³⁶ used a microfluidic silicon mixer equipped with a freeze-quenching device for trapping metastable intermediates populated during fast chemical or biochemical reactions. Seven staggered pillars were placed in the mixing channel to induce the turbulent flow necessary for efficient mixing at moderate flow rates. In this work, time-dependent resonance Raman spectroscopy was applied to monitor the ultrafast reaction. Leung³⁷  demonstrated the combination of continuous flow microfluidic reactors and on-line Raman spectroscopic detection for rapid and efficient reaction monitoring. The confocal Raman microscopy was successfully utilized for rapid, on-line analysis and reaction condition screening of the catalytic oxidation of IPA. Barnes³⁸ used fiber optic Raman spectroscopy for the quantitative analysis of monomer composition and degree of conversion of methacrylate-based droplets in a microfluidic device. Raman data allowed accurate measurement of the decrease in the C=C bond conversion as a function of increasing cross-linker concentration. The results show that Raman spectroscopy, combined with a lab-on-a-chip, is an effective on-chip analytical tool for screening polymeric materials on the micro meter scale.

Figure 7. Time-dependent Raman spectra of imine formation in a glass microfluidic channel. (a) Nine laser focusing points along the channel. (b) Corresponding Raman spectra of imine formation at the fixed position 5, showing loss of the 1700 cm^-1 peak and gains of the 1593 and 1628 cm^-1 peaks. Time interval for each spectrum was 4 min. (c) Formation of imine.

1.5.2 environmental analysis

The SERS detection technique, in combination with microfluidic technology, has been developed for rapid and sensitive trace analysis of chemical analytes, including environmental pollutants, toxic chemicals, insecticides, and pharmaceutical products. Compared with other detection methods for trace analysis, the detection sensitivity was enhanced by several orders of magnitude. The first SERS application to microfluidics was carried out by Keir³⁹. In this work, a model analyte, an azo dye derived from trinitrotoluene (TNT) in a glass microfluidic channel, was analysed using the SERRS technique. The SERRS substrate, silver nano colloids, was prepared by borohydride reduction of silver nitrate in the microfluidic channel. Using this on-chip SERRS technique, it was possible to detect 10 fmol of an azo dye analyte. Yea⁴⁰ used the SERS technique for a rapid and sensitive trace analysis of cyanide anions in an alligator teeth-shaped PDMS microfluidic channel. The confluent streams of silver colloids and trace analytes of cyanide solution were effectively mixed in the channel through the triangular structures, which are located on the upper and lower surfaces of the channel in a zigzag manner. Because cyanide ions are negatively charged, the polyamine spermine tetrachloride was used as the agent to neutralize the cyanide anions. As a result, the cyanide anions were effectively adsorbed on silver nanoparticles in the channel. Compared with other methods for the trace analysis of cyanide anions, the detection sensitivity was enhanced by several orders of magnitude. Figure 8 illustrates the SERS spectra for different concentrations of cyanide anions in a microfluidic channel and the corresponding calibration curve. In this work, the LOD was estimated to be 0.5–1.0 ppb. A similar SERS detection technique has been applied to the trace analysis of malachite green (MG). MG is a cationic triphenyl methane dye that has been widely used around the world as a fungicide and antiseptic in the aquaculture industry. Nowadays, the use of MG has been banned in several countries and is not approved by the US Food and Drug Administration. Thus, it is important to develop highly sensitive detection systems for monitoring trace amounts of MG in water. Lee⁴¹ used the SERS detection technique for the trace analysis of MG. Under the optimal condition of flow velocity, MG molecules were effectively adsorbed onto silver nanoparticles while flowing along the upper and lower zigzag-shaped PDMS channel. The LOD, using the SERS microfluidic sensor, was found to be below the 1–2 ppb level, and this low detection limit is comparable with the result of the LC-MS detection method. Methyl parathion is one of the most hazardous insecticides used for the control of sucking and chewing insects in a wide range of crops, such as cereals, fruits, vegetables, and cotton, as well as ornamentals. Lee⁴² also used the confocal SERS technique for the fast and sensitive trace analysis of methyl parathion pesticides.

Figure 8. (a) SERS spectra for increasing concentrations of cyanide ion in the microfluidic channel: (a) 0 ppb, (b) 1 ppb, (c) 5 ppb, (d) 50 ppb, (e) 100 ppb, (f) 200 ppb, and (g) 300 ppb. (b) Variation of the CN stretching peak area as a function of cyanide ion concentration (correlation coefficient, R = 0.991).

1.5.3 Microfluidic-SERS assay of living cells

The structure of a cell is complex due to the various biological constituents in and on the surface of the cell, such as proteins, DNA, lipids, and carbohydrates. SERS is perfectly suited to selectively analysing targets with complex components, while microfluidic systems can be used to keep cells stable in solution.

Zhou⁴³ reported a highly sensitive detection method of bovine serum albumin on a PDMS microfluidic chip using SERS analysis. Yang⁴⁴ reported the detection of the proteins lysozyme and cytochrome c as well as the bacterial species S. oneidensis MR-1 in aqueous solutions with high sensitivity using SERS and optical fibers. Lu⁴⁵ differentiated 21 methicillin-sensitive S. aureus (MSSA) and 37 MRSA using clinical isolates within SERS coupled microfluidic chips. In this optofluidic platform, segmented flow was created by using mineral oil as the segregation medium to remove the memory effects; designed snake channels were used to enhance the mixing efficiency; and PDMS background spectra were minimized by implementing a confocal micro-Raman spectrometer (figure 9).

Figure 9. Scheme of the optofluidic platform. Designed winding channels were used to enhance the mixing efficiency

4.    Microfluidic-SERS Within Microdroplets

As mentioned at the end of part 4.2, microdroplets in microfluidic devices are potential supporter for SERS detection. Since microdroplets are widely used as individual microreactors, in which the chemical reactions and other chemical or physical process are independent from outer environment, they have great promise for the analysis of chemical reaction or bio-process, such as single cell cultivation and sorting, in separate small-scaled spaces. Meanwhile, the droplets provide the interface of two different phases, offering possibility of detections analytes in two phases if the SERS substrates can be formed on the interface. Therefore, developing a SERS detection method within microfluidic-based microdroplets could be an interesting research topic which has potential of wide applications.

To implement SERS detection within microdroplets, analytes molecules should be absorbed on the nanostructure of SERS substrates. Conventionally, there are various methods to generate SERS substrates, such as lithograph, self-assembly, template methods and chemical approaches, among which self-assembly, especially on the interface of gas/liquid and liquid/liquid, is a “bottom-up” approach, which can avoid both hard templates and complex equipment as conventional “top-down” processes do, providing a simple and efficient method for 3D substrates generation. In comparison to lithographic methods, the great advantage of the self-assembly method for SERS substrates generation is the preparation of highly ordered nanostructures and extremely narrow inter gaps (several nano-meters) between nanoparticles, enabling large electromagnetic field enhancement between the noble metal surface and the analytes chemical bonds. Furthermore, the close-packed structures resulting from self-assembly can provide a maximum surface density of hot spots within an illuminated area, which is rather important for the ultra-trace chemical detection.

Within microdroplets, nanoparticle self-assembly could be achieved by either functionalizing nanoparticle or adding electrolyte into the solution containing nanoparticles. Van der Waals force or electrostatic force would drive the nanoparticles to aggregate into mono or multi layers on the interfaces. Hydrophobic surfactants such as alkanethiols are often used to drive the nanoparticles toward self-organization at the aqueous interface⁴⁶. For example, Gia⁴⁷ generated plasmonic colloidosomes as 3D SERS platforms by self-assembly of perfluorodecanethiol functionalized Ag nanoparticles at the interface of water and decanes. Ultrasensitive SERS detection of multiphase toxin sensing was achieved with this platform. NaCl was used by Edel’s group⁴⁸ for increasing the ionic strength of the aqueous phase to drive the Au nanoparticles to self-organize at the water-dichloroethane interface. Previously in 2001, Wei⁴⁹ reported a calixarene-based surfactant containing several hydrocarbon tails, providing strong repulsive forces, which has been successfully employed for the fabrication of highly ordered, close-packed gold nanoparticle arrays by self-assembly of large-diameter gold nanoparticles (up to 170 nm in diameter) at the air/water interface. However, all the above studies were based on emulsions generated by vigorous shaking in a tube, where the generation of interfaces is difficult to control. Therefore, the data reproducibility suffered from the bad emulsion monodispersity and lack of controllability, limiting further applications of emulsion based SERS technology.

Luckily, microfluidic approach has great promise to solve these problems by providing controllable microdroplets generation and the liquid/liquid or gas/liquid interfaces, offering stable spaces for nanoparticles’ self-assembly. With this method, a higher reproducibility of 3D substrates’ structure could be achieved due to the reproducibility of the microdroplet structure. In the detection aspect, microfluidic approaches also have many other advantages. For example firstly, because of the transparent materials which are generally used as microfluidic chips, microfluidic platforms could help achieve in situ optical and spectroscopic analysis, which is beneficial for the processes where constant monitoring is required, such as single cell sorting. Another advantage is that the detection processes with microfluidic only needs small volumes of samples, which is beneficial for detection situations when the samples are limited, rare, toxic or expensive. Moreover, in addition of liquid flows, microfluidic platforms can also be used to generate controllable gas flows and monodispersed gas bubbles, which are difficult to achieve in other large-scale platforms. This controllability of gas flows provides a possibility for online airborne absorption and detection with 3D metal nanoparticle substrate by SERS technology, which might be beneficial for explosive or air pollutes analysis. Based on stable emulsion or double emulsions, it is also possible that nanoparticle layers on the interfaces could detect different analytes in two or three immiscible phases simultaneously by using microfluidic devices, which could provide a possibility to avoid complicated sample taking processes. As stated above, integration of microfluidic platforms and SERS technology might be a powerful tool both for the generation of 3D substrates and the analytes detection, becoming a key enabler for many future applications.

2. References

1. Ko, H., Singamaneni, S. & Tsukruk, V. V. Nanostructured Surfaces and Assemblies as SERS Media. Small 4, 1576–1599 (2008).

2. Chen, L. & Choo, J. Recent advances in surface-enhanced Raman scattering detection technology for microfluidic chips. Electrophoresis 29, 1815–1828 (2008).

3. Gopinath, A., Boriskina, S. V, Reinhard, B. M. & Dal Negro, L. Deterministic aperiodic arrays of metal nanoparticles for surface-enhanced Raman scattering (SERS). Opt. Express 17, 3741–3753 (2009).

4. Kneipp, K. Surface-enhanced raman scattering. Phys. Today 60, 40–46 (2007).

5. Whitesides, G. M. The origins and the future of microfluidics. Nature 442, 368–73 (2006).

6. Haeberle, S., Mark, D., Von Stetten, F. & Zengerle, R. Microfluidic platforms for lab-on-a-chip applications. Microsystems Nanotechnol. 9783642182938, 853–895 (2012).

7. Le, H. P. Progress and trends in ink-jet printing technology. J. Imaging Sci. Technol. 42, 49–62 (1998).

8. Terry, S. C., Herman, J. H. & Angell, J. B. A gas chromatographic air analyzer fabricated on a silicon wafer. IEEE Trans. Electron Devices 26, 1880–1886 (1979).

9. Manz, A. et al. Design of an Open-tubular Column Liquid Cbromatograpb Using Silicon Chip Technology. Sensors Actuators B Chem. 1, 249–255 (1990).

10. Chen, H. & Meiners, J. C. Topologic mixing on a microfluidic chip. Appl. Phys. Lett. 84, 2193–2195 (2004).

11. Beebe, D. J., Mensing, G. a & Walker, G. M. Physics and applications of microfluidics in biology. Annu. Rev. Biomed. Eng. 4, 261–286 (2002).

12. Inoue, I., Wakamoto, Y., Moriguchi, H., Okano, K. & Yasuda, K. On-chip culture system for observation of isolated individual cells. Lab Chip 1, 50–55 (2001).

13. Huang, N.-T., Zhang, H.-L., Chung, M.-T., Seo, J. H. & Kurabayashi, K. Recent advancements in optofluidics-based single-cell analysis: optical on-chip cellular manipulation, treatment, and property detection. Lab Chip 14, 1230–1245 (2014).

14. Chen, W., Lam, R. H. W. & Fu, J. Photolithographic surface micromachining of polydimethylsiloxane (PDMS). Lab Chip 12, 391–5 (2012).

15. Xia, Y. & Whitesides, G. M. Soft Lithography. Angew Chem Int Ed. 37, 550–575 (1998).

16. Yasui, T. et al. Confocal microscopic evaluation of mixing performance for three-dimensional microfluidic mixer. Anal. Sci. 28, 57–9 (2012).

17. Sarrazin, F., Salmon, J. B., Talaga, D. & Servant, L. Chemical reaction imaging within microfluidic devices using confocal raman spectroscopy: The case of water and deuterium oxide as a model system. Anal. Chem. 80, 1689–1695 (2008).

18. DeMello, A. J. Control and detection of chemical reactions in microfluidic systems. Nature 442, 394–402 (2006).

19. Kamholz, A. E., Weigl, B. H., Finlayson, B. A. & Yager, P. Quantitative analysis of molecular interaction in a microfluidic channel: The T-sensor. Anal. Chem. 71, 5340–5347 (1999).

20. Mengeaud, V., Josserand, J. & Girault, H. H. Mixing processes in a zigzag microchannel: Finite element simulations and optical study. Anal. Chem. 74, 4279–4286 (2002).

21. Liu, R. H. et al. Passive mixing in a three-dimensional serpentine microchannel. J. Microelectromechanical Syst. 9, 190–197 (2000).

22. Stroock, A. D. et al. Chaotic mixer for microchannels. Science 295, 647–651 (2002).

23. Kim, D. J., Oh, H. J., Park, T. H., Choo, J. B. & Lee, S. H. An easily integrative and efficient micromixer and its application to the spectroscopic detection of glucose-catalyst reactions. Analyst 130, 293–298 (2005).

24. Strehle, K. R. et al. A reproducible surface-enhanced Raman spectroscopy approach. Online SERS measurements in a segmented microfluidic system. Anal. Chem. 79, 1542–1547 (2007).

25. Huang, K.-W., Wu, Y.-C., Lee, J.-A. & Chiou, P.-Y. Microfluidic integrated optoelectronic tweezers for single-cell preparation and analysis. Lab Chip 13, 3721–7 (2013).

26. Im, H., Bantz, K. C., Lindquist, N. C., Haynes, C. L. & Oh, S. H. Vertically oriented sub-10-nm plasmonic nanogap arrays. Nano Lett. 10, 2231–2236 (2010).

27. Connatser, R. M., Cochran, M., Harrison, R. J. & Sepaniak, M. J. Analytical optimization of nanocomposite surface-enhanced Raman spectroscopy/scattering detection in microfluidic separation devices. Electrophoresis 29, 1441–1450 (2008).

28. Banerjee, A., Perez-Castillejos, R., Hahn, D., Smirnov, A. I. & Grebel, H. Micro-fluidic channels on nanopatterned substrates: Monitoring protein binding to lipid bilayers with surface-enhanced Raman spectroscopy. Chem. Phys. Lett. 489, 121–126 (2010).

29. Xu, B.-B. et al. Localized flexible integration of high-efficiency surface enhanced Raman scattering (SERS) monitors into microfluidic channels. Lab Chip 11, 3347–3351 (2011).

30. Oh, Y.-J. & Jeong, K.-H. Optofluidic SERS chip with plasmonic nanoprobes self-aligned along microfluidic channels. Lab Chip 14, 865–8 (2014).

31. Abramowitz, M., Spring, K. R., Keller, H. E. & Davidson, M. W. Basic principles of microscope objectives. Biotechniques 33, 772–781 (2002).

32. Iles, A., Habgood, M., De Mello, A. J. & Wootton, R. C. R. A Simple technique for microfluidic heterogeneous catalytic hydrogenation reactor fabrication. Catal. Letters 114, 71–74 (2007).

33. Lee, H. L., Boccazzi, P., Ram, R. J. & Sinskey, A. J. Microbioreactor arrays with integrated mixers and fluid injectors for high-throughput experimentation with pH and dissolved oxygen control. Lab Chip 6, 1229–1235 (2006).

34. Fortt, R., Wootton, R. C. R. & De Mello, A. J. Continuous-flow generation of anhydrous diazonium species: Monolithic microfluidic reactors for the chemistry of unstable intermediates. Org. Process Res. Dev. 7, 762–768 (2003).

35. Lee, M. et al. Applicability of laser-induced Raman microscopy for in situ monitoring of imine formation in a glass microfluidic chip. J. Raman Spectrosc. 34, 737–742 (2003).

36. Lin, Y. & Yeh, S. R. Microfluidic mixer and ultra-fast freeze-quench device. Biophys. J. 84, 579A–580A (2003).

37. Leung, S.-A., Winkle, R. F., Wootton, R. C. R. & deMello, A. J. A method for rapid reaction optimisation in continuous-flow microfluidic reactors using online Raman spectroscopic detection. Analyst 130, 46–51 (2005).

38. Barnes, S. E., Cygan, Z. T., Yates, J. K., Beers, K. L. & Amis, E. J. Raman spectroscopic monitoring of droplet polymerization in a microfluidic device. Analyst 131, 1027–33 (2006).

39. Keir, R. et al. SERRS. In situ substrate formation and improved detection using microfluidics. Anal. Chem. 74, 1503–1508 (2002).

40. Yea, K. et al. Ultra-sensitive trace analysis of cyanide water pollutant in a PDMS microfluidic channel using surface-enhanced Raman spectroscopy. Analyst 130, 1009–1011 (2005).

41. Lee, S. et al. Fast and sensitive trace analysis of malachite green using a surface-enhanced Raman microfluidic sensor. Anal. Chim. Acta 590, 139–144 (2007).

42. Lee, D. et al. Quantitative analysis of methyl parathion pesticides in a polydimethylsiloxane microfluidic channel using confocal surface-enhanced Raman spectroscopy. Appl. Spectrosc. 60, 373–377 (2006).

43. Zhou, J., Ren, K., Zhao, Y., Dai, W. & Wu, H. Convenient formation of nanoparticle aggregates on microfluidic chips for highly sensitive SERS detection of biomolecules. Anal. Bioanal. Chem. 402, 1601–1609 (2012).

44. Yang, X., Gu, C., Qian, F., Li, Y. & Zhang, J. Z. Highly sensitive detection of proteins and bacteria in aqueous solution using surface-enhanced raman scattering and optical fibers. Anal. Chem. 83, 5888–5894 (2011).

45. Lu, X. et al. Detecting and Tracking Nosocomial Methicillin-Resistant Staphylococcus aureus Using a Microfluidic SERS Biosensor. Anal. Chem. 85, 2320–2327 (2013).

46. Kim, B., Tripp, S. L. & Wei, A. Self-organization of large gold nanoparticle arrays. J. Am. Chem. Soc. 123, 7955–7956 (2001).

47. Phan-Quang, G. C., Lee, H. K., Phang, I. Y. & Ling, X. Y. Plasmonic Colloidosomes as Three-Dimensional SERS Platforms with Enhanced Surface Area for Multiphase Sub-Microliter Toxin Sensing. Angew. Chemie – Int. Ed. 54, 9691–9695 (2015).

48. Cecchini, M. P., Turek, V. A., Paget, J., Kornyshev, A. A. & Edel, J. B. Self-Assembled Nanoparticle Arrays for Multiphase Trace Analyte Detection. Nat. Mater. 12, 165–171 (2012).

49. Kim, B., Tripp, S. L. & Wei, A. Tuning the Optical Properties of Large Gold Nanoparticle Arrays. 676, 1–7 (2001).