Microfluidics toward a Lab-on-a-Chip
History and state-of-the-art
The historical beginnings of microfluidics are difficult to point accurately. Indeed, in the 1940s, the implementation of high-pressure liquid chromatography and gas chromatography have required fluid manipulation at the microscale with high precision , . In 1975, Terry reported the first proof of concept of a miniaturized analytical silicon made device, through a gas chromatographic analyser. This device was efficient enough to separate a simple mixture of model compounds in a few seconds. Although this early proof of concept was very promising, the real development of microfluidics as a specific domain in science, and the name ―microfluidics‖ itself, emerged in the 1990s thanks to the know-how developed by the Microelectronics community. In 2006, microfluidics has been defined by G.M. Whitesides as ―the science and the technology of systems that process or manipulate small (10-9 to 10-8L) amounts of fluids, using channels with dimensions of tens to hundreds of micrometres. ».
To illustrate this definition, Figure I.1shows the characteristic sizes of microfluidic devices.
During the early emerging phase of microfluidics, the community of analytical chemistry took an active part in miniaturized analytical devices, to integrate the different steps of the analytical chain such as the sample pre-treatment, analytes separation and detection within the same device. In 1990, Andreas Manz introduced the concept of miniaturized micro total analysis system (µTAS). Thereafter, microfluidics raised a strong interest among various scientific communities, who proposed various applications. Notably, microfluidics appears as a powerful toolbox with numerous already developed applications in biology and life sciences, and rapidly emerging ones in chemistry, medicine, and environment. Such progress and the increasing level of complexity associated with it requiredin particular new tools for of liquids handling, and therefore the development and the integration of micro pumps and valves[9–11],Thus, thefastdevelopment of miniaturized fluidic systems is directly connected with the ability to make new types of microstructures or to duplicate existing structures at micro-scale levels.
The current microfabrication techniques, such as micromachining, soft or hard lithography, printing, embossing, injection moulding have been developed and optimized for different materials. Silicon, glass and elastomer have beenthe mostly used materials for microchip microfabrication . Nevertheless, in thelast years, silicon tends to be usedless and less because of its laborious microfabrication, its optical opacity and difficulties to integrate other components. G. Withesides was among the early pioneers in this domain. Glass and silicon, mostly used during the early developments of microdevices, were progressively replaced by polymeric materials such as poly(dimethylsiloxane) (PDMS), an elastomer which is optically transparent, easy to mould and to be used by non-experts. At present,PDMS is the most commonly used material to fabricate microfluidic systems at the research laboratory level . This material, however, is not the ideal one for many industrial applications, and new materials based on mouldable thermoplastics such as COC, Polycarbonate, PMMA, are emerging. The accession to new materials and innovative techniques dedicated to the microfluidic systems’ fabrication are continually evolving and already gave rise to many reviews , .
In the present work, we used PDMS microchip fabricated by soft lithography. Such a technique is particularly useful for pattern replication as it enables rapid prototyping of microfluidic devices and with no need of a cleanroom. Herein, are briefly described the properties, Table 1. For more details, please refer to .
Besides these microfabrication progresses, an important contribution towards the integration of complex bioanalytical processes in microdevices lead to the creation of simple methods to fabricate pneumatically activated valves, , mixers[21–23]and pumps . To handle such microfluidic elements it is of high importance to finely control the fluid within the device and thus to understandthe physical phenomena involved in fluid transport in microchannels[26–28]. The manipulation of fluids in microenvironments, such as channels with typical dimensions of tens to hundreds of micrometres, implies some hydrodynamic properties of fluids further discussed in this manuscript.
The use of microfluidic devices offer several advantages for analytical and bioanalytical applications, notably the use of minute amounts of reagents and samples (down to picolitres) and an important time reductiondue to the decrease of diffusion distance. In addition the large surface-to-volume ratio offers an intrinsic compatibility between microfluidic systems and surface-based assays. Microfluidic systems also offer strong potential for portability,low cost, versatility in design for integration and multiplexing development.
The field of micro total analysis systems has rapidly expanded and has especially evolved towards bioanalytical and biochemical applications (Figure I.2). A rather wide account, ranging from basic research in academia to commercial applications was presented in a series of review articles by A.Manz group , , [29–33].
Figure I.2: Timeline of the evolution of microfluidic technology.
Progress in the development of bioanalytical microfluidic devices presents particularly interesting features for the development of miniaturized, portable, user friendly and low cost systems called ―point-of-care‖ (POC) systems. A well-known example of lab-on-a-chip is already commercially available on the market: the strip test for pregnancy . In this device, the sample flows across a membrane, which gathers labelling reagents embedded within it, and flows over an area that contains immobilized molecules, the labelled captured analytesform a visible peak. The other major class of POC test is the blood glucose test. The latter is also performed on membranes and uses signal amplification by redox enzyme. The glucose test has improved diabetic patient’s quality of live and reduced risks of over or under dosage. This is by nature a particularly well-fitted application for POC, since the useful sensitivity is within the mM range, i.e. not very challenging, the needed frequency of the testing frequency can be high, typically several tests a day. Many other POC applications, however, would be more demanding regarding for instance sensitivity, so numerous progress is still needed.
The basis physics of microfluidics
In order to better understand the physics at play in our systems, let’s briefly introduce the basic fluids properties and the characteristic of dimensional parameters.
At the microscale level, the inertial effects are generally negligible but viscosity and surface tension effects are prevalent. The dimensionless Reynolds number ( ) gives a measure of the ratio of inertial forces to viscous forces, it is defined for a microchannel with a circular section as:Where is the fluid density ( . −3), is the characteristic velocity of the fluid ( . −1), is the hydraulic channel’s diameter ( ) and is the dynamic viscosity of the fluid ( . ). Such a dimensionless number allows to describe the flow regime; laminar or turbulent according to the Reynolds number ( )value. We can note that when the characteristic dimension of the channel d decreases, the Reynolds number decreases strongly. It is generally smaller than 1 (typically 10−2or10−3)at the velocities usually encountered in microfluidics < 1 indicates a laminar flow, without any sign of turbulence, while for >> 1, the flow is considered to be turbulent (Figure I.3).Between ~1 and ≫ 1 lies a transition regime, in which turbulence generally does not occur, but inertial effects may be significant.
In microchannels the flow is almost exclusively laminar due to their small cross section and even if two or more streams are used, no mixing is observed except by diffusion (Figure I.4) .
In a laminar flow, mixing between several liquids in microdeviceis governed by advection on larger scales, and diffusionacross flow lines on small scales. This involves that transport of mass, energy or momentum in a direction perpendicular to the flow is essentially diffusive. Depending on the applicationthe laminar flow can be desirable or not. With laminar flows, the sorting and analysis of products is in general easier. On the contrary, for some applications notably those requiring a reaction between species initially contained in different fluids or solids, a mixing between the different reagents injected is required.
The relative importance of advectionand diffusion effects is given by the Péclet number:
Where ν is the velocity (m.s-1), d is a covered distance by the particle during time ( ) and is the diffusion coefficient of the particle ( 2. −1). Péclet number gives an indication on the diffusion process in microfluidic device. One of the principal applications of diffusive mixing is known as ―T-sensor‖ or ―Y-mixer‖. Two flow fluids are injected to flow alongside each other down the channel, and solute molecules in each stream diffuse into the other one, forming an interdiffusion zone(Figure I.5).
As the diffusion process is slow, a long channel is required so that an effective mixing can take place, especially to increase the contact area between species to be mixed. In this aim, several approaches have been considered. They can be characterized as either active, where the sample species is mixed by using an energy input from the exterior, or passive, where particular microchannels configuration increases the contact area between different incoming fluids. Hessel et al. and Lee presented a review on microstructured mixer devices, their mixing principles and their mixing performances.Table 1and Table 3provide a non-exhaustive list of the different categories available to perform a good mixing in microfluidics.
This last decade, many flow control methods have been described in microfluidics. The liquid flow can be controlled using either passive or active pumps. In passive pumping, the forces involved to drive liquids are due to chemical gradients on surfaces, to osmotic pressure, and to permeation through the PDMS or capillary forces.
In active pumping, an external power source is needed; the flow rate can be controlled essentially by electric field, magnetic field or centripetal force, by mechanical displacement(e.g. syringe or peristaltic pump), or by hydrostatic pressure.Until recently, syringe pumps have been the most common in microfluidic laboratories, .
However, the flow rate range that can be applied limited and it depends on the syringe capacity, so that a compromise must be made between accuracy and capacity. The syringe is connected to the microfluidic chip with tubing connections (device shown in Figure I.8.a), lead to large dead volumes and significant flow hysteresis. In 2004 a new generation of pressure controller was made available to overcome the defects encountered with the syringe pump (Futterer et al., 2004) and marketed by Fluigent Company®. Further details are given in the MAESFLO controller appendix.
Whatever the type of flow control applied, a pressure-driven flow leads to a parabolic velocity profile. Such flow profile causes an axial dispersion since the fluid velocity is zero at the tube wall and maximum at the centreline, as depicted in the Figure I.8.b with a schematic representation of Poiseuille flow and Taylor dispersion. Due to this phenomenon, the pressure-driven flow is less attractive for separation as it leads mostly to a decrease of separation performance and consequently to a lower resolution.
Figure I.8: (a) Schematic drawing of syringe pump. (b) Poiseuille flow and Taylor dispersion . Another approach to control the flow is based on the movement of molecules under an electric field. When a liquid comes into contact with the charged surface of a microchannel, the formation of an interfacial charge causes a rearrangement of the local free ions in the liquid so as to produce a thin region of nonzero net charge density, near the interface. This is the electric double layer (EDL). In presence of an electric field, the solvated ions are driven towards the oppositely charged electrode, and by viscous drag transfer the motion to the rest of the liquid which runs uniformly as a plug-like flow (see Figure I.9). This phenomenon is called electroosmosis, or electroendosmosis. Electroosmotic flow control has been first exploited with glass microchips. But others materials can also generate electroosmotic flow. For instance, the oxidation of PDMS by plasma treatment generates silanol groups (Si-OH) on the surface of the microchannels.
The electroosmotic flow (EOF) creates a uniform flow profile, thus allowing the transport of small volumes of samples while avoiding band broadening in contrast with the hydrodynamic dispersion, in the pressure-driven flow. Nevertheless, the EOF control exhibits important drawbacks for bioassays, including buffer incompatibility i.e. the fluids used in such systems are often so highly conductive that they could generate deleterious Joule heating for assay performance.
Table of contents :
Chapter I. Magnetic handling of particles in microfluidic devices
1. Microfluidics toward a Lab-on-a-Chip
a. History and state-of-the-art
b. The basis physics of microfluidics
i. Reynolds number
iii. Flow control
2. Lab-on-a-Chip: magnetic microparticles and immunoassays
a. Magnetic microparticles for bioanalysis
i. Magnetic properties
ii. Superparamagnetic microparticles
iii. Structure of microparticles
b. Immunoassay principle
i. Heterogeneous immunoassay
3. Technical aspects of microparticles manipulation
a. On-chip microparticles manipulation
i. Chemical and mechanical trapping of microparticles
Beads immobilization based on chemical interaction
ii. Dynamic manipulation of microparticles
b. Magnetic microparticles concept for bioassays
i. System using electromagnets
ii. System using permanent magnets
Chapter II. Integrating fluidized beds in microfluidic systems
2. Magnetic microparticle transport in a microfluidic device
a. Magnetic force acting on a particle
b. Viscous drag force
c. Particle mobility
3. Magnetic microparticle motion within a plug
a. Microchip design: 1st generation
b. Microparticle capture principle
c. Structure of the magnetic plug
d. Forces acting on a magnetic bead in a flow
e. Hydrodynamic behaviour of the plug
i. Experimental procedure
ii. Hydrodynamic characterization of the plug
Particle bed behaviour
Influence of microparticle amount
Influence of the gap between the magnets
Influence of the microparticle size
4. Second device generation: towards an integrated fluidized bed
b. Fluidized bed: some basics
c. Influence of the bed porosity on the fluid flow
d. Darcy’s law: pressure drop across the bed
e. Pressure-flow relationship in fluidized bed: Minimum fluidization velocit
f. Magnetically assisted fluidized bed
g. Minimum fluidisation velocity
5. Microchip based fluidized bed
6. Spatial magnetic field distribution
7. Fluid flow and particle distribution in magnetic fluidized beds
a. Microparticle capture
b. Pressure drop in the magnetic bed
c. Pressure vs. flow rate hysteresis
d. Flow distribution in the fluidized bed
e. Bed expansion and porosity
f. Fluidized bed activation: control of the fluid resistance
g. Balance between magnetic and drag forces
Chapter III. On-chip immunoextraction
2. Model compounds Ab/Ag
3. Off-chip Ag capture and elution
a. Off-chip characterisation and optimisation of the protocol
i. Materials and methods
ii. Off-chip experiments
4. On-Chip Immunoextraction
a. Description of the experimental set-up
b. Optimisation of the on-chip immunoextraction
5. Results and discussion
6. Preliminary conclusion