Lab-on-a-chip devices for gold nanoparticle synthesis and their role as a catalyst support for continuous flow catalysis

3.1 Tubular or plug flow reactors

Plug flow reactors are tubular reactors that are utilized for continuous chemical processes. These types of reactors are filled with the desired catalyst for chemical conversion to take place. The geometry of a plug flow reactor plays an important role depending on the type of chemical reaction, required selectivity and conversion efficiency. They also have the ability to add multiple reactant mixtures at different points if needed. Usually, a plug flow reactor is called a packed bed reactor when it is packed with a solid material, in most cases, a solid catalyst. The reactant to product conversion within the plug flow reactor is measured with respect to the residence time of the plug, which is a function of its position in the reactor. Plug flow reactors have the ability to run for extended periods and have high volumetric unit conversion. In a plug flow reactor, a stable concentration profile can be obtained at steady state with concentrations varying in space as the reaction happens in the flow path. The heat transfer rate can be controlled with ease when the reactor is packed with thinner tubes or fewer thicker tubes packed parallel. Another advantage of using a plug flow reactor is its low operational cost. However, undesired thermal gradients, poor temperature control, gas exchange limitations in a sealed reactor and expensive maintenance are some of the current challenges [54]. A typical millifluidic tubular reactor is made up of polymer or glass capillaries with a diameter in the order of 1 mm. Tubular millifluidics has also been utilized to produce hierarchically organized multiple emulsions or particles with a good control over sizes and shapes [55].

The other types of reactors such as gas-liquid trickle bed reactor, crossflow packed bed, parallel packed bed with integrated temperature sensors [49] (Figure 3), flow cell reactor [56], drip flow reactor [57], capillary reactor [58], membrane reactors [59], etc. are few examples of tubular or plug flow reactors that can be used for biological, chemical, physical analyses, and studies.

3.2 Lab-on-a-chip systems: microfluidic and millifluidics

A LOC device is a single chip designed to have components with varying dimensions. The size of this device can range from a few centimeters to nanometers and can handle lesser volumes of fluid. Various laboratory functions can be integrated into a LOC device when all the components are precisely positioned in order to have reliable and repeatable reaction conditions [60]. Some of the challenges in tubular reactors such as requirement for larger space, unavoidable zero dead fluid volume, longer response time, higher power consumption, lower reproducibility, higher waste, and so on can be addressed using chip-based LOC devices [61, 62]. A variety of designs of chip-based LOC devices have been reported demonstrating flow principles, mixers, and traps ranging from the simplest form of the T-section systems, where two fluid inputs enter through channels at the bottom and slowly diffuse over the length of the microchannel to tangential microchannels, where the channels can exchange fluid through the shared area of contact [63]. More sophisticated designs include a pillar array PDMS-based microfluidic channel for the SERS detection of hazardous materials to microfluidic traps using optical, mechanical, dielectrophoretic, electrophoretic, acoustic, and magnetic forces.

In a LOC device, scaling up of the reaction conditions is simple and straightforward, which is done either by carrying out the reaction in continuous flow mode or by increasing the throughput through parallel processing. The use of LOC devices can also decrease the time for process transfer from lab to pilot-plant and then to industrial scale, and the products can be monitored online [62]. The kinetics, selectivity, and the product yield can also be improved easily, given that the external parameters like pressure, temperature, and reactant mixing is taken into account [64]. As the volume needed for the LOC experiments is very less, i.e., in the range of microliters to nanoliters, miniaturization of the flow channels allows us to change the experimental conditions along the reaction path in a shorter time (milli- to microseconds). Hence, LOC devices allow systematic investigation for the development of new synthetic strategies for investigating larger parameters if an online readout is available [65].

To overcome the demerits of conventional-based methods for NP preparation, LOCs can be operated at optimized and steady state with control over the reaction conditions like reagent addition, mixing, temperature, and scaling higher throughputs involving parallel operation of multiple reactor units [62, 66]. There have been reports on demonstration of wet chemical synthesis of semiconducting, metallic, dielectric, magnetic, and core shell NPs using the microfluidic methods [67–74]. However, prohibiting the NPs from growing over the walls of the LOCs is a challenge in the continuous flow reactor while preparing colloidal metal dispersions. Significant deposition of NPs and aggregation over the reactor walls has been attributed to the high surface to volume ratios [65]. For reactions happening at laminar flow within a microreactor, by selecting the appropriate choice of reaction conditions and microreactor wall material, polymer additives, and developing concentric laminar flow patterns, the issue over the particle deposition and aggregation can be overcome or minimized [65, 75–78]. It should also be noted that broader particle size distributions within a millifluidic or microfluidic reactor can be achieved with higher residence time of reactants [79].

LOC devices such as the microfluidic systems are constructed to perform chemical reactions, which are designed in such a way to maximize the efficiency of the reaction conditions. The significant aspect of performing a reaction in a microfluidic reactor is that the reaction conditions can be controlled desirably according to flow rate, concentration, pressure, etc. Early microfluidic devices had stainless steel, polymer tubing to provide reactant flow into the reactor. However, with repeated experiments, silica tubing is being used nowadays to achieve proper mixing and preparing hydrodynamic micro- and nanostructures. Other advantages of using microfluidic devices are low fluid volumes, good process control, quicker analysis, compact, safe, and reproducible results [80]. Especially for flow catalysis, an ideal microfluidic reactor should be capable of working at the desired working conditions (pressure, temperature, chemical compatibility, concentration, etc.) without breaking the system. A few examples of such microfluidic reactors are shown in Figure 5 [81].

Figure 5

Examples of (A) metal, (B) glass/glass, and (C) silicon/Pyrex and microreactors. (Reproduced with permission from Ref. [81], Copyright Elsevier, 2012.)

Millifluidic devices can be considered as low-cost LOC devices that have been utilized for synthesis of a variety of micro- and nanomaterials. They offer more advantages over the microfluidic devices. For example, larger quantities of NPs can be produced with good control over their size-distribution and shape. They do not require expensive lithography for their fabrication. They can also be used as flow reactors with potential opportunities to manipulate and functionalize the products at different stages of synthesis [82].

The flow rates of the reactants that are fed into the millifluidic reactor can be increased, thereby, reaching easy scale up by integrating several reactors on a single platform. Owing to the continuous reaction conditions and low-cost methodology, millifluidic devices are used as a potential production tool in industries. The millifluidic devices can be easily assembled and disassembled without any expensive techniques such as lithography or etching for fabrication. As the channel dimensions are increased when compared to the microfluidic reactor, channels are less prone to clogging in a millifluidic reactor, thereby, resulting in lesser residence time of the reactants. The millifluidic devices can also be used to prepare atomically precise nanomaterials and study their formation in situ in order to obtain time-resolved kinetic details of the nanomaterials formed. A chip-based LOC is a device that is made up of polymer or glass as a substrate. These devices can be fabricated into any desirable shape, which can be fully integrated to analytical equipment for characterization, identification, and separation processes. Moreover, the reaction channel in a chip-based LOC can be designed into various shapes (like zigzag, spiral-shaped, serpentine-shaped channels) according to the experimental conditions. To summarize, a simple millifluidic device bridges the gap between a microfluidic device and bulk reactor. A traditional microfluidic device has a channel dimension in micrometer scale and can hold a fluid volume of about nanoliters to a few microliters. In contrast, a millifluidic device has a channel dimension in millimeter scale, which can carry fluid volume up to milliliters and offers high throughput without compromising on the flow properties. In addition, it offers superior in situ monitoring capabilities due to higher signal-to noise ratio than traditional microfluidic devices.

4.1 Overview of LOC synthesis of metal nanocatalysts

Many inorganic and organic nanocatalysts have been synthesized in a continuous manner with more control over the particle size distribution, shape, and quality of the nanomaterial. LOC devices for synthesis of inorganic nanomaterials have recently been reviewed [23]. A variety of metal NPs, oxide NPs, semiconductor NPs, quantum dot (QD) core-shell-structured micro-, and nanostructures are few types of nanomaterials synthesized using the LOC devices. Most often, reducing agents such as sodium borohydride, lithium hydrotriethyl borate, 3-(N,N-dimethyldodecylammonia) propane sulfonate, and sodium citrate are used as the reducing salts for synthesizing metal NPs. Ligands, though required to stabilize NPs from agglomeration, have a deleterious effect on the catalytic activity of NPs. For example, the most commonly used ligands like thiols, phosphines, amines, etc., for gold NP synthesis hinder the reactant to reach the metal surface when they are in heterogeneous phase, thereby, reducing their catalytic activity. A well-known example for heterogeneous-phase gold catalysis is the gas-phase CO oxidation reaction [83]. However, in solution-based homogeneous catalysis, the ligands tend to disperse in the solvent, thereby, allowing the reactants to reach the metal surface and, hence, resulting in increased catalytic activity, for example, styrene oxidation in toluene [84]. In general, these ligands can be removed from the NP surface by several methods like calcination [85], oxygen treatment [86], and ozone treatment [87]. The synthesis of metal nanocatalysts within the LOC devices is carried out in two ways. The first method is where the metal precursors are fed into the reactor followed by the addition of the second reactant into it as shown in Figure 6 [88]. To reduce the risk of agglomeration, ligands and surfactants are passed through the reactor in the final step. The disadvantage of using this approach toward the nanocatalyst synthesis is that the reactants are fed into the LOC one after the other leading to a large residence time distribution. Owing to the continuous single-phase reaction condition, the possibility of reactant clogging within the channel is high, thereby, resulting in loss of optical and absorption information. The second approach toward the synthesis of metal nanocatalyst using the LOC devices is that the reaction solutions are fed simultaneously within the reactor through separate inlets as shown in Figure 7 [79]. These reactants mix together within the LOC device, where the nucleation and the growth of the nanocatalyst take place. The second approach for the synthesis of the metal nanocatalyst is preferred due to its short residence time distribution and less chances of clogging [89].

Figure 6

Schematic illustration of reactant feed into the LOC device. Modular microreactor arrangement for flow-through process. (Reproduced with permission from Ref. [88], Copyright Elsevier, 2008.)

Figure 7

LOC device for simultaneous mixing of the reactants. A gold nanoparticle seed suspension (S) and aqueous reagent solutions (R1 and R2) are separately delivered into one arm of a microfluidic T-junction, and silicone oil is delivered into the other arm. Droplets are pinched off at the T-junction. Reagents and seeds are rapidly mixed by chaotic advection. The oil forms a thin lubricating layer around the translating droplets, and prevents contact between growing particles and the microchannel walls. (Reproduced with permission from Ref. [76], Copyright Wiley-VCH Verlag GmbH & Co. KGaA 2009.)

Several metal nanocatalysts have been synthesized using the microfluidic devices. A few common nanocatalysts that were prepared using the microfluidic devices are gold NPs and nanorods [65, 73, 75, 77, 79], silver NPs [89] and nanorods [65, 88, 90], copper NPs [91], palladium NPs [92, 93], CdSe@ZnS [93], CdSe@ZnSe [94], SiO₂@TiO₂ [72]_, γ-Fe₂O₃@SiO₂ [95], [CdSe@ZnS] in PLGA microgels and microcapsules [96], γ-Fe₂O₃ in microhydrogels [97], γ-Fe₂O₃ or QDs in PNIPAM microcapsules [98], and superparamagnetic Janus particles [99]. The syntheses of these nanocatalysts vary accordingly with respect to the temperature and the type of microfluidic reactor used. SU-8-PEEK, glass capillary, PVC and PEEK tubing, silicon Pyrex, silica capillary, PDMS glass, PDMS glass+aluminum reflectors, etc., are few types of microreactors that were used to prepare these nanocatalysts [100].

Although millifluidic devices prove to be an efficient tool in continuous flow catalysis, there are only a few reports on millifluidic nanocatalysts, and these are discussed in this review. Millifluidic systems offer similar reaction conditions for the synthesis of metal nanocatalysts like that of the microfluidic systems. Unlike microfluidic devices, metal nanocatalysts synthesized using millifluidic reactors withstand higher pressures and velocity distribution confined to the width of the interfacial zone [101–103].

4.2 Synthesis of nanostructured gold catalysts using LOC devices

Owing to its intrinsic value and properties, gold has always been considered as a noble metal of all the elements [104]. When gold catalyst is in nanoparticulate form, it possesses several unique properties. At nanoscale, gold can be used to transform carbon monoxide to carbon dioxide. It is used to remove toxins from exhaust gases. Gold NPs have different properties from those of bulk gold [105–107]. A number of reactions have been efficiently catalyzed by gold NPs. For example, carbon monoxide oxidation [2], catalytic combustion of hydrocarbons [108], hydrochlorination of ethyne [109], hydrogen sulfide and sulfur dioxide removal, oxidation of glucose to gluconic acid [110], oxidative decomposition of dioxins, oxidative removal of mercury, ozone decomposition [111], reduction of NOx with propene [112], carbon monoxide or hydrogen, selective oxidation, e.g., epoxidation of olefins, selective hydrogenation [113], e.g., of alkynes and dienes to mono-olefins, vinyl acetate synthesis from ethene, acetic acid, and oxygen, etc. However, atomically precise gold NPs catalysts have not yet been explored for their effectiveness in these reactions [85, 114].

Capping agents such as ligands, surfactants, polymers, and dendrimers are commonly used to confine the growth of the gold NPs in a controlled synthesis [115–122]. A good control over the growth of NP with respect to its size, diameters <150 nm was seen in the reduction of tetrachloroaurate ion in aqueous solution along with a reducing agent. In most cases, sodium borohydride has been used as the reducing agent, which is one of the most preferred methods to synthesize spherical gold NPs. Various other methods that were carried out for synthesis of gold NPs were by using poly(N-vinyl-2-pyrrolidone) as a protective agent in ethylene glycol for shape control [119]. A seed-mediated growth approach was carried out with the surfactant cetyltrimethylammonium bromide (CTAB) as the directing agent in order to prepare gold nanorods in aqueous solution [117, 123]. Another method of preparation involved the use of [Au(SO₃)₂]^3-, which was decomposed and precipitated under acidic conditions to synthesize the quasi monodisperse gold microspheres [124]. Also, gold microspheres with diameters of more than 1 mm were synthesized in a controlled flower-like fashion using a simple electrochemical route [125]. Reactants like poly(sodium 4-styrene sulfonate) was used as an emulsifier, water treatment agents like dispersants, flocculation agents, sulfur exchange resin, etc. have been used along with various gold salts for preparing the gold NPs [126].

Owing to a smaller dimension of the microfluidic devices, characterization of gold NPs in order to study its morphology and structural chemistry becomes tedious. In recent times, this disadvantage is overcome by synthesizing catalytically active gold micro- and nanostructures in a millifluidic platform. Gold catalysts prepared in this method offered a greater scope for spectroscopic probing catalysis of reactions as it happens. A comparative study between millifluidics-based synthesis and traditional flask-based synthesis of gold micro-/nanostructures shows that size and morphology of the gold can be better controlled by varying the flow rates using millifluidic systems [102].

There have been several reports on the synthesis of gold NPs using LOC devices. Influence in the reaction conditions such as flow rate, temperature, reducing agent, concentration, catalyst support, etc., play a vital role in particle synthesis using the LOC devices. Gold NPs of proper size and shapes have been synthesized using stabilizing agents. These agents help in preventing particle agglomeration during the synthesis. Tetraoctylammonium bromide (TOAB), 1-dodecanethiol, 11-mercaptoundecanoic acid, polyvinyl pyrolidone (PVP), etc., are a few reagents that have been used in the preparation of ligand-stabilized gold NPs.

Lohse and coworkers studied high-throughput synthesis and functionalization of gold NPs with different sizes and shapes using a simple bench top reactor system [127]. A flow reactor assembly (Figure 8) was used to operate for the synthesis of gold NPs by modifying previously reported synthetic procedures [128, 129]. Citrate-stabilized gold NPs of size 4 nm were synthesized by mixing tetrachloroauric acid and sodium citrate with a residence time of 3 min to obtain a reddish brown solution. An experiment based on the Brust-Schiffrin procedure was also carried out to synthesize 3 nm of mercaptohexanoic acid-stabilized gold NPs [130–132] and CTAB-stabilized Au NP growth procedures [133]. CTAB-stabilized 2- and 8-nm gold NPs were synthesized by mixing tetrachloroauric acid and CTAB with sodium borohydride solution to give deep brown and vibrant red color solutions, respectively.

Figure 8

The integrated millifluidic reactor used for gold nanoparticle synthesis and functionalization is shown. (A) Diagram and picture of the rector for AuNPs synthesis. The reactor is composed of multiple modular commercially available components, and fluid flow is driven by the peristaltic pump. (B) In this reactor, mixing of the growth solution and the seed/borohydride solution occurs in a simple polyethylene Y-mixer. (C) The reactor also features an integrated flow-based purification system, in which a commercially available tangential flow filtration cartridge can be attached to an additional peristaltic pump in order to integrate high-throughput approach for nanoparticle purification or functionalization. (Reproduced with permission from Ref. [127], Copyright ACS Publications, 2013.)

They used the 8-nm gold NPs as the seeds for the synthesis of the 20-nm gold NPs, and in a similar fashion, 20-nm gold NPs were used as seeds for the synthesis of 40-nm gold NPs. All the gold seeds and NPs were stabilized using CTAB in solution.

Lohse and coworkers [127] have also reported the synthesis of gold nanorods using different synthesis protocols. They synthesized gold nanorods with the flow reactor by the seeded growth approach [132–134] where growth solutions containing tetrachloroauric acid, silver nitrate, and l-ascorbic acid were mixed with aqueous CTAB solution. These solutions that were prepared were fed into the flow reactor at 50 ml/min having a residence time of 3 min where they mix with each other within the reactor before they get deposited into a conical flask. Similarly, gold nanorods were prepared by using borohydride and ascorbic acid method within the flow reactor using the reported procedures [135–137]. Gold nanorods with dog bone structure and nanorods with large transverse diameters were prepared by modifying the isotropic overgrowth with the increased ascorbic acid addition [138–140]. As before, the growth solution was prepared by mixing tetrachloroauric acid with silver nitrate and L-ascorbic acid. CTAB solution was mixed with 4 nm of gold seed dispersion, and the mixture was aged for 2 h to prepare the seed solution. The growth and the seed solutions were then fed within the flow reactor at 50 ml/min and deposited into a conical tube. The nanorods formed through this procedure were 1.2 nm in size.

Tsunoyama and coworkers [141] reported the synthesis of gold clusters of size ∼1 nm stabilized by PVP using a microfluidic reactor. Syringes were used to feed the aqueous HAuCl₄ and PVP solutions into the microfluidic reactor kept in a methanol bath at 0°C. The individual solutions were overlaid in an interdigital arrangement in region I (Figure 9), where the solutions were laminated into 16 substreams, and region II had wide multilamellar flow, which was compressed into a single 0.5-mm-wide stream. An Erlenmeyer flask was used to collect the mixed solution eluted from the outlet in an ice bath and was stirred for 1 h. Au clusters stabilized by PVP were obtained by ultrafiltration after subsequent deionization.

Figure 9

Schematic diagram for the synthesis of PVP-stabilized Au clusters in a micromixer. (Reproduced with permission from Ref. [141], Copyright ACS Publications, 2008.)

Shalom and coworkers [73] also reported the synthesis of thiol-stabilized gold NPs of sizes ranging from 2.9 to 3.7 nm with standard deviations of particles of size range between 0.6 and 0.9 nm respectively, using the microfluidic reactor. The thiol-functionalized gold NPs were often called as monolayer-protected clusters (MPCs). The gold-thiolate polymer was prepared by phase transfer of HAuCl₄ into toluene using TOAB followed by the addition of 1-dodecanetiol. A three-layer micromixer (Figure 10) was used to mix the thiol-stabilized gold NPs with NaBH₄, which acts as the reducing agent [116]. Different mole ratios of gold-thiolate polymer were fed into the microfluidic device with NaBH₄ to produce the monolayer-protected NPs.

Figure 10

(A) Three-dimensional schematic of a radial interdigitated mixer. Each mixer is fabricated in three layers. In the first two layers, input flows are directed to two circular bus channels, which, in turn, split the flow into eight identical fluid laminae and deliver reagent streams toward a central mixing chamber. The final layer acts as a cap to enclose channels and as a guide for input and output capillaries. The output is from the center of the uppermost layer. (B) Photograph of the fabricated mixer. Microchannels are filled with dye solutions to show different shadings for the different channels. (Reproduced with permission from Ref. [73], Copyright Elsevier, 2007.)

Pedro and coworkers [142] reported similarly prepared gold NPs with an average diameter of 2.7 nm stabilized by 11-mercaptoundecanoic acid (MUA) using microfluidics. In their report, the gold colloid was prepared by one phase reaction by reducing gold (III) chloride with NaBH₄ prior to stabilizing it with MUA. Syringes were used to control the flow rate of the reactants within the microfluidic device fabricated by low-temperature co-fired ceramics technology (LTCC). A multilayer design of 3D structures was integrated into these static mixers along with electric, mechanic, and fluidic components using LTCC technology. With this static microfluidic mixer, the overall process of MUA-stabilized gold NP synthesis was split into two steps. The first step involved the preparation of gold NPs, where the respective reactants, gold (III) chloride and NaBH₄ were hydraulically focused in a laminar fashion within the microfluidic device. The gold chloride solution was then fed into the microfluidic device between two streams of NaBH₄ in order to have efficient mixture. In the next step, the as-formed gold NPs were stabilized by feeding MUA into the microfluidic device through a separate inlet. The resulting solution that was collected out from the device was found to be fully protected MUA gold NPs.

Polte and coworkers [143] also reported a different method for growth of gold NPs with an average radius of 0.8 nm to about 2 nm using a microstructured static mixer. These gold NPs were prepared at room temperature by homogeneously mixing aqueous tetrachloroauric acid (HAuCl₄) with NaBH₄ within the static micromixer (Figure 11). Different residence times were applied to synthesize the gold NPs. At an increased residence time, it was observed that there was no apparent effect on the particle size suggesting that the growth of the particles happens at a shorter scale. The experimental setup was used successfully for gold NP synthesis using the sodium borohydride as the reducing agent. It was analyzed that the gold nuclei Au⁰ was formed initially at a rapid complete conversion from the gold precursor when the setup was coupled to X-ray absorption near-edge spectroscopy (XANES). These gold nuclei coalesced into larger particles, and within 100 ms, the gold precursor was completely converted to metallic gold due to rapid mixing within the microstructure mixer.

Figure 11

Experimental setup for particle synthesis in continuous flow mode coupling of a microstructured static mixer directly to SAXS analysis in a flow cell. (Reproduced with permission from Ref. [143], Copyright ACS Publications, 2010.)

A similar time-resolution study on the size evolution of gold NP with an average particle size of 2.53 nm in a millifluidic reactor was reported by Li and coworkers [102]. Different residence times of the reactants HAuCl₄ and NaBH₄ were applied to prepare the gold NPs using a millifluidic device, and the spatially resolved size evolution of the gold NP was investigated using transmission electron microscope. They prepared the gold NPs by mixing aqueous HAuCl₄ and DMSA within the millifluidic device followed by the addition of NaBH₄. The flow of the reactants was controlled using a syringe pump (Figure 12). It was observed that the gold NPs showed a broader size distribution at a residence time of 3.53 s, and the control of particle size was much easier with a milliscale LOC device when compared to the conventional-based synthesis of gold NPs.

Figure 12

Schematic illustration of the millifluidic reactor channel used for the synthesis of gold NPs using HAuCl₄ and DMSA. (Reproduced with permission from Ref. [102], Copyright Wiley-VCH Verlag GmbH & Co. KGaA, 2012.)

Saikrishna et al. [103] have reported on the time-resolved mapping of the growth of gold nanostructures using similar millifluidic chip. The millifluidic chip used in their study had serpentine channel through which HAuCl₄ and DMSA solutions were fed at different flow times (such as 1, 3, 5, 7, and 9 h) to form different gold nano/microstructures (Figure 13). The mixing of the precursors resulted in the formation of gold sulfide clusters of size 1–2 nm, which were later reduced by washing them with NaBH₄ solution. The authors have studied the in situ X-ray absorption spectroscopy (XAS) of the formation process of the nanoclusters (Figure 14) and also their further growth into microstructures. They also demonstrated flow catalysis using the gold-deposited chips for the reduction of 4-nitrophenol and ferricyanide.

Figure 13

Schematic for the formation steps of gold structures within millifluidic reactor.

Figure 14

(A) Millifluidic chip marked with different zones where in situ XAS was performed. (B) In situ XAS analysis at different zones within the millifluidic channel. (C) Reaction scheme of precursors. (Reproduced with permission from Ref. [103], Copyright ACS Publications, 2013.)

Synthesis of gold NPs of size 5.6 nm using multistep microfluidic reaction system was reported by Ishizaka and coworkers [144]. They used facile methods to synthesize highly dispersed gold NPs using three micromixers, which had two inlets each in order to prepare the gold NPs by NaBH₄ reduction. The precursor solutions, HAuCl₄ and N,N-dimethylacetamide (DMAc) solution of the poly-(amic acid) (PAA), were fed into the first Y-shaped micromixer using pumps to prepare a homogeneous mixed solution. This homogenous mixed solution was then fed simultaneously with the NaBH₄ solution to form the homogeneously dispersed gold NP in the PAA solutions. Finally, a microemulsion was formed by feeding the resulting solution along with n-hexane. The microemulsion was then collected and heated in a microheat exchanger, and a dispersion of the liquid containing polyimide NPs confining Au NPs were obtained after several processing procedures.

Kohler and coworkers [77] also reported a three-step mixing synthesis of gold NPs at room temperature using a microflow channel. Unlike the previous case, gold NPs synthesized by these authors used ascorbic acid as the reducing agent. There were three mixing zones for their synthesis process. In the first zone, the reducing agent, i.e., the ascorbic acid (educt solution 1) and Fe (II) sulfate (educt solution 2) were mixed at low flow rates within the micromixer. The Fe (II) was then added to polyvinyl alcohol solution before the first mixing step. The second mixing zone had sodium metasilicate solution (educt solution 3) reacting with the mixture from the first step. Finally, 1 mm of HAuCl₄ (educt solution 4) was introduced to the mixture from the second step to form 5-nm-sized Au NPs.

Wagner and coworkers [75] reported the synthesis of gold NPs with size ranges of about 5 to 50 nm using a hydrophobic continuous flow microreactor channel. The microreactor channel walls were made hydrophobic using trichloro (1H, 1H, 2H, 2H-perfluoro-octyl) silane. In their procedure, 1 mm of aqueous HAuCl₄ containing PVP was mixed with ascorbic acid within a microreactor as shown in the Figure 15. The reactant solutions were then fed into the micromixer by polypropylene syringe pumps, and the gold sol was collected separately through the mixer outlet. The pH of the reactant solution was altered in order to suppress microreactor fouling.

Figure 15

Schematic of the experimental setup showing the connectivity of the microreactor (STATMIX 6, area 22×14 mm). (Reproduced with permission from Ref. [75], Copyright ACS Publications, 2005.)

Similar to the work carried out by Wagner and group [75], Cabeza and coworkers [145] also synthesized gold NPs by controlling its size with a segmented flow microfluidic platform under hydrophobic conditions (Figure 16). The channel of the microfluidic platform was precoated with poly(tetrafluoroethylene) (PTFE) hydrophobic layer in order to study the effect of axial dispersion of the gold NP size distribution. Segmented flow of reagents were achieved by feeding three separate streams of reactants viz. NaBH₄, aqueous mixture of chloroauric acid and tetracyltrimethylammonium bromide (TTABr), and the third stream had either air, toluene, or silicone oil. The process was carried out for 10, 20, and 40 s, and different flow rates were maintained for each reagent to get particle sizes of 3.8±0.3, 4.6±2.1, and 4.9±3.0 nm, respectively.

Figure 16

Synthesis of gold NPs using segmented flow system under hydrophobic conditions. The hydrophobic microchannel (water is the dispersed phase). Fluorescein was added to the aqueous phase to improve the optical resolution. (Reproduced with permission from Ref. [145], Copyright ACS Publications, 2012.)

Yet, another hydrophobic-channeled poly(dimethylsiloxane) (PDMS) microfluidic device was used by Lazarus and coworkers [146] for synthesizing monodisperse gold NPs of size 4.3±0.5 nm. In this process, an interdiffusion between the two reagent streams, i.e., chloroauric acid/1-methylimidazole and NaBH₄ in 1-butyl-3-methylimidazolium tetrafluoroborate (BMIM-BF4) was achieved by injecting a stream of pure BMIM-BF4 using a syringe pump at various flow rates through different inlets. In addition, inert oil (polychlorotrifluoroethylene) was pumped into the microfluidic device to define the flow regimes of the reagent stream, and the final product was collected through the outlet in ethanol.

In a different procedure, Duraiswamy and coworkers [79] reported the synthesis of gold NPs of sizes <5 nm using droplet microfluidics in order to prevent the contact between the solutions and the microreactor channel walls. Their procedure involved the production of zero-valent gold (Au⁰) NPs by reducing trivalent gold (Au³⁺) chloride with ascorbic acid in the presence of CTAB as the surfactant. The seed and the reagent solutions were fed into the microfluidic mixer at equal flow rates. The gold NP seed suspension and the aqueous reagent solutions were mixed rapidly by chaotic advection within the T-junction microfluidic device to form droplets. Silicone oil was fed along with the seed and the growth solutions through a separate arm of the T-junction device (Figure 17), which prevented contact of the growing particles with the microchannel walls by forming a thin lubricating layer around the droplets. The droplets were then collected separately in a sampling reservoir and aged to ensure growth completion.

Figure 17

Schematic of the experimental setup for the production of gold NPs using a T-junction microfluidic device. (Reproduced with permission from Ref. [79], Copyright John Wiley and Sons 2009.)

Boleininger and coworkers [65] used microfluidic mixers to synthesize gold NPs with a high concentration of CTAB, which acts as the surfactant for forming rod-shaped gold particles. Millimolar concentration of the spherical metal seed crystals in aqueous growth solution (HAuCl₄) was mixed with ascorbic acid in the micromixer. These metal seeds were produced by reducing the metal salt with freshly prepared NaBH₄. The flow rates of the growth solution and the seeds were maintained independently using the syringe pumps. The seed crystals were allowed to grow at a fixed temperature by directing the seed and the growth solution mixture from the microfluidic device through a temperature-controlled tubing and collected separately.

Another report by Jun and coworkers [147] describes the synthesis of gold NPs of final size range from 3 to 35 nm [148] through a wet chemical synthesis technique using both millifluidic and microfluidic mixers at room temperature (Figure 18). The precursor solution HAuCl₄ was mixed with ascorbic acid (whose pH was adjusted with a calculated amount of NaOH). Rapid mixing of these solutions was achieved by feeding them into the millifluidic mixer aided by two syringe pumps. A separate glass vessel was used to collect the colloidal gold solution for further analysis. The reactants were mixed at various flow rates in order to investigate its effect on the gold NP synthesis. Ball-Berger mixer was used to achieve very fast mixing conditions for high flow rate reactions [149], whereas a butterfly mixer built with poly(dimethylsiloxane) (PDMS) microchip was connected to the syringe pump for low flow rate reactions [150], and the colloidal gold solution obtained was collected through the microtube outlet.

Figure 18

Illustration of the experimental setup. The mixer can be a pressed Teflon tube (millifluidic mixer) (7 or 22 cm length) (left), a Ball-Berger mixer (not shown), or a microfluidic mixer with a butterfly geometry (right). (Reproduced with permission from Ref. [147], Copyright ACS Publications, 2012.)

Kitson and coworkers [151] reported the synthesis of gold NPs of size 10 nm using a 3D-printed millifluidic and microfluidic devices. These 3D-printed reactors had two and three inlets for introducing the reactant solutions (Figure 19). Gold NPs were synthesized using a premixed solution of aqueous HAuCl₄ and sodium citrate. The solutions were fed into the 3D reactor through the first inlet and mixed with NaBH₄, which was fed simultaneously through the second inlet. The formation of the gold NPs was observed by in-line UV-Vis spectroscopy, which decreased over time due to the gold deposition over the channel walls.

Figure 19

3D-printed millifluidic reactors for synthesis of gold NPs. The devices with three inlets each connected to a pump, and the in-line ATR-IR and/or UV-Vis flow cells connected to the outlet. (Reproduced with permission from Ref. [151], Copyright ACS Publications, 2008.)

Weng and coworkers [152] reported the synthesis of citrate-based gold NPs of size 35±2 nm using PDMS-based microfluidic device. They used a pneumatic rotary micromixer with four peristaltic membrane layouts for the synthesis of hexagonal gold NPs with trisodium citrate as the primary reducing agent. Trisodium citrate and preheated HAuCl₄ solutions were fed within the micromixer and mixed vigorously to obtain a ruby red solution, which was then heated to 115°C for the reaction completion. Different heating times and concentrations of the HAuCl₄ and tri-sodium citrate were also tested within the micromixer to synthesize the hexagonal gold NPs with different sizes.

Similar to the previous report, Yang and coworkers [153] used a microfluidic chip on a thermoelectric device (TE) to synthesize gold NPs with different sizes (19, 28, 37, and 58 nm) by manipulating the volumes of the reactant, i.e., chloroauric acid and trisodium citrate at a constant reaction temperature of 100^oC. A cone-shaped condenser was placed on top of the chamber after the reagents were mixed into it to maintain a constant reactant concentration. A micropump was used to inject the sodium citrate, and a vortex-type micromixer was used to stir the reagents in the mixing chamber in order to form gold NPs by reducing the chloroauric acid solution, which was observed by a color change from light yellow to ruby red. The reaction process was carried out for 10 min before the solution was cooled down to room temperature.

4.3 Gold nanocatalysts supported within LOC systems

In the previous section, various synthetic protocols that have been used for making gold NPs using the LOC systems were summarized. In this section, the focus is on gold NPs that are supported within the LOC devices for catalysis. The gold NPs catalysts were either prepared in situ enabling attachment to the channel walls or synthesized using conventional methods and loaded or immobilized over the channels of the LOC system (Figure 1). Loading the gold NPs into the LOC systems are carried out by various methods. For example, Abahmane and coworkers loaded the gold nanocatalysts into the capillaries and packed it by supporting vibrations [20]. Similarly, dip coating the internal surface of the capillary [154], calcination [155], crosslinking the copolymer to the gold NP [156, 157] spin-coating techniques [158] were used in order to make gold nanocatalyst-supported reaction systems.

Synthesis of polypyridine derivatives using alumina-supported gold NPs of size 2 nm using microcontinuous flow conditions has been reported by Abahmane and coworkers [20]. After supporting the gold NPs on Al₂O₃ through impregnation, the heterogeneous catalysts were used in the microcontinuous flow system (Figure 20) as the packed bed capillary reactor (PBCR). Reactants were fed through the PBCR under optimized continuous flow conditions to retrieve the products. The authors also reported the synthesis of pyridine [159] and propargylamines (Figure 21) [21] using similar continuous microflow system packed with gold NP-impregnated alumina.

Figure 20

Synthesis of polypyridine derivatives using alumina-supported gold NPs using microcontinuous flow setup (feedstock 1: bis-α-H-ketone solution, feedstock 2: propargylamine solution; M, micro mixer; PBCR1, PBCR2, packed bed capillary reactors (ID: 1 mm, 50 cm length); P, pressure sensor; T1, T2, temperature sensors; S1, S2, syringe pumps). Inlay: image of the PBCRs. (Reproduced with permission from Ref. [20], Copyright John Wiley and Sons, 2009.)

Figure 21

Flow chemistry setup schemes for the different reaction regimes for the synthesis of propargylamines. (A–C). P1, P2, pumps; P(1), pressure sensor; T(1), T(2), temperature sensors; PBCR1 (Montmorillonite K-10) and PBCR2 (Au-NP@Al₂O₃), packed-bed capillary reactors; BPR, back-pressure regulator. (Reproduced with permission from Ref. [21], Copyright John Wiley and Sons, 2011.)

Packed bed capillary reactors were also used for liquid phase hydrogenation of citral on Au/TiO₂ thin films within capillary microreactors by Protasova and coworkers [154]. Gold NPs of size 4.5±0.5 nm prepared by NaBH₄ reduction were doped on titania films to form the Au/TiO₂ sol. The coating of Au/TiO₂ sol within the capillary reactor was done after pretreatment and calcination processes and used to study the kinetics of citral hydrogenation. A similar Au/TiO₂ catalyst was reported by Cao and coworkers [160]. They used Au-Pd over TiO₂ for benzyl alcohol oxidation within the microstructured reactors (Figure 22). The Au-Pd/TiO₂ catalyst was prepared separately and loaded into the microreactors and was later used for catalytic studies.

Figure 22

Microreactor, reactor assembly with temperature control and the schematic of the experimental setup for benzyl alcohol oxidation reaction. (Reproduced with permission from Ref. [160], Copyright Elsevier, 2011.)

Juarez and coworkers [155] reported the continuous flow carbamoylation of aniline by dimethyl carbonate using Au/CeO₂ of size 3–4 nm within a microreactor (Figure 23). The Au NPs supported on CeO₂ support was prepared using convention wet impregnation method, coated within the stainless steel microreactor plate and used for carbamyoylation reaction.

Figure 23

Parts of the microreactor showing the microchannel plate coated with a thin film of nanoparticulated ceria. (Reproduced with permission from Ref. [155], Copyright Elsevier, 2011.)

Wang and coworkers [156] reported the oxidation of alcohols with molecular oxygen using the gold-immobilized mirochannel flow reactor. Microencapsulated gold prepared from chlorotriphenylphosphine gold (AuClPPh₃) and copolymer in THF solution, was used as a gold source for the immobilization [157, 161]. Gold was immobilized on the capillary column reactor by reducing the cyanopropyl groups using lithium aluminum hydride to the corresponding amine (Figure 24). The modified capillary column was pumped with a colloidal solution of the microencapsulated gold and heated to 170°C for 5 h in order to have a crosslinking of the copolymer to result in the gold immobilization, which was further used for oxidation studies (Figure 25).

Figure 24

Immobilization of the gold catalyst into capillary column reactor. (A) Reduction of the cyano group to an amine. (B) Preparation of microencapsulated gold. (C) Immobilization of the gold catalyst. (Reproduced with permission from Ref. [156], Copyright John Wiley and Sons, 2009.)

Figure 25

Experimental setup of the gold-catalyzed oxidation of alcohols. (Reproduced with permission from Ref. [156], Copyright John Wiley and Sons, 2009.)

Apart from using immobilized gold nanocatalyst prepared through the bottom-up approaches, gold NPs synthesized using the top-down method were also used within the microfluidic reactors for catalytic applications. For example, Adleman and coworkers [158] carried out heterogeneous catalysis mediated by plasmon heating with gold NPs of estimated average diameter size of 20±5 nm prepared by lithography and immobilized over a glass microscope slide, which was temporarily bonded to PDMS. Figure 26 shows the schematic of gold NPs immobilized on the microfluidic channel attached to the PDMS support.

Figure 26

Schematic of the process of plasmonic heating (side view). A microfluidic channel with gold nanoparticles attached to a glass support; fluid flows from left to right. (Top) A laser at or near the frequency of the plasmon resonance of the gold nanoparticles is focused on the top of the support, and the subsequent heat generated in the nanoparticles is transferred to the surrounding fluid and forms vapor. The vapor phase components react on the catalyst forming gas bubbles, which are carried downstream. (The channel height is 40 μm, and the radius of the nanoparticles is ∼10 nm). (Reproduced with permission from Ref. [158], Copyright ACS Publications, 2009.)

Saikrishna et al. [103] have demonstrated reduction reactions of 4-nitrophenol and ferricyanide using a gold microstructure-coated millifluidic chip. The authors achieved ∼90% conversion of the reactants using gold-coated chip compared to 20% conversion using a chip without gold coated within the channel. Their gold catalysts were also reusable up to 40 catalytic cycles (80 h), which establishes the significant advantage of flow catalytic process.

There are some challenges that need to be addressed when a NP-supported LOC device is used for catalytic conversion reactions. While loading a metal catalyst like gold NP onto a support or a reactor, there are chances that these particles might lose their functional activity and selectivity due to environmental conditions. For example, while immobilizing the NPs onto a substrate, proper pretreatment procedures have to be carried out in advance in order to have an improved catalyst attachment. There are chances that the catalyst might undergo agglomeration due to the thermal pretreatment like calcination, mechanical vibrations, etc.