GB2447412A - Versatile micro-mixing system with chemical and biological applications - Google Patents

Abstract

The present invention provides microfluidic technology for mixing discrete volumes of chemical and biological solutions and suspensions in an accurate, versatile and predetermined manner. In particular, the method of mixing solutions in a microfluidic system comprises: a) moving a carrier solution 8 which is immiscible with the chemical solutions 9 to be mixed in a conduit 1; b) loading chemical solutions into reservoirs; c) operating a device 6 to move a known volume of a first chemical solution through a first opening into the conduit; d) where required, operating a metering mechanism 2 to move the carrier fluid to move the volume of first solution to a position near a second opening; e) operating a second device to dispense a known volume of a second solution into the conduit causing it to coalesce 10 with the first solution droplet; f) repeating steps d) and e) until all the desired solutions have been mixed together; g) operating the metering mechanism to move the carrier fluid, thereby moving the mixture away from the openings. A method of priming and calibrating the device is also provided. The apparatus may further have a sipper device which is able to pick up a plurality of samples by dipping into them in sequence. The apparatus may be particularly useful for crystallisation operations, e.g. protein crystallization.

Description

Versatile micro-mixing system with chemical and biological

applications

TECHMCAL FIELD

This invention relates to microfluidic systems, and in particular to mixing discrete volumes of chemical and biological solutions and suspensions in an accurate, versatile and predetermined manner. The volumes to be mixed range from less than a femtolitre to a few milliliters.

BACKGROUND OF TIlE INVENTION

Microfluidic systems have the advantages of using very small samples, consuming small amounts of samples and reagents, and compact size.

Many applications have been suggested for microfluidic systems. Applications and suggested applications include protein crystallization, DNA analysis, DNA synthesis, genomics, cell based systems, clinical diagnostics, bio-defense sensors, high throughput screening, pharmaceutical lead optimization, high content screening, combinatorial chemistry and genetic fingerprinting.

"Lab-on-a-chip" is a term intended for devices that allow several laboratory functions to be carried out on a chip with an area of a few square centimeters. However, existing microfluidic devices are not really like laboratories because they are limited to a few functions that they were designed for. They are therefore more like miniature chemical plants or reactors, and they are not able to carry out new functions and experiments at will as one could in a real laboratory.

The present invention aims to satisfy the long-felt need for microfluidic devices that are easily adaptable to novel research and testing applications. The invention is able to generate multiple discrete liquid aliquots, where each aliquot contains a mixture that can be unique and is individually controllable. For example, it can set up combinatorial experiments using dozens of ingredients, where every mixture is determined using a computer program for experimental design. These aliquots take the form of discrete volumes or droplets in a carrier fluid such as oil. The droplets take on the role of test-tubes or SBS-format (Society for Biomolecular Scientists) plastic plates in the laboratory.

Some other microfluidic systems use droplets in oil. See, for example, http://www.raindancetechnologies.com/formulate.html. However, other systems do not possess hardware or use methods that allow the contents of individual droplets to be varied at will.

The present invention is based on, and is a further development of ideas disclosed in patent GB 2 097 692B by the same inventor, which was published in 1984. GB 2097 692B disclosed a microfluidic device where chemical solutions or samples were mixed and moved around the conduits of the device as droplets in oil (or other liquid that was immiscible with the chemical solutions). The suggested apparatus of GB 2 097 692B was constructed by forming indentations or grooves in one or more plates and clamping, gluing or welding these plates against each other to form closed conduits.

The present invention discloses methods for dispensing, priming and calibrating a microfluidic device that were not disclosed in GB 2 097 692B. It also discloses novel hardware.

BRIEF SUMMARY OF THE INVENTION

One aspect of the invention is a microfluidic apparatus with reservoirs filled with chemical or biological solutions, emulsions or suspensions (hereinafter referred to as "chemical solutions"). These reservoirs are in communication with a conduit where mixing of the chemical solutions takes place. This mixing conduit initially contains a carrier fluid (which could be oil), and the chemical solutions (which could be aqueous) are moved into the mixing conduit where they combine and mix with each other, forming discrete volumes or droplets in the carrier fluid. By moving the carrier fluid, the droplets can be moved.

In addition to reservoirs for chemical solutions and a mixing conduit, the apparatus may comprise a detector that can detect the presence of droplets or slugs in carrier fluid as they move past the detector.

A second aspect of the invention is a method of priming the dispensing system by loading its reservoirs with chemical solutions.

A third aspect of the invention is a method of using the detector to calibrate the system with regard to accurately dispensing the solutions into the mixing conduit.

The carrier fluid must be immiscible with the reactants and samples that are to be mixed in the apparatus.

Moreover, to prevent drops from sticking to the walls of the mixing conduit and other conduits, the walls of these conduits should be made from or coated with materials with properties similar to the carrier fluid, and dissimilar to the droplet phase. For handling aqueous samples and solutions the carrier fluid could be an oil etc. and conduits could be formed from or lined with fluoropolymers, silanized material, silicones etc.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure la shows a schematic view of a simple embodiment of the invention with four liquid channels, comprising four reservoirs for chemical solutions, a mixing conduit, four devices for moving chemical solutions, an accurate metenng mechanism for moving the carrier fluid, a detector and an outlet.

Figure lb shows a configuration where the mixing conduit increases from a relatively small cross-section near the most upstream opening, to a relatively large cross-section at the most downstream opening.

Figures 2a to 2d show a sequence of liquid movements that can be used to load a chemical solution into one or more liquid channels, thereby pnming solution reservoirs and openings.

Figures 3a to 3d show a sequence of liquid movements that can volumetrically calibrate one liquid channel.

Figures 4a to 4d show a sequence of Liquid movements to dispense two chemical solutions from two liquid channels and form a droplet by mixing them.

Figure 5 is an exploded view showing one possible method of constructing the apparatus by forming indentations or grooves on the surfaces of one or more plates, and bringing two or more such plates into face-to-face contact.

Figures 6a to 6d show various methods of forming conduits by bringing plates into face-to-face contact.

Figure 7 shows extra inlets for introducing extra carrier fluid at or near the point where a conduit is connected to a hole that passes through a plate.

Figure 8 shows that a detector could be associated with e.g. a flexible outlet tube that is located downstream of the mixing conduit.

Figure 9 shows a schematic view of an embodimert of the invention where a pressure chamber with valves is used with a time-and-pressure approach to dispense chemical solutions.

Figure 10 shows a schematic view of an embodiment where solutions are dispensed in discrete steps using pairs of valves located in each channel.

Figure 11 shows a schematic view of an embodiment where metering of all liquids is performed by a single metering mechanism that is placed downstream of the mixing conduit.

Figure 12 shows a schematic view of an embodiment that is able to pick up and mix a large number of samples from e.g. one or many 96- well plates.

Figure 13 shows a schematic view of a system for automatic protein crystallization, where chemical solutions are mixed and passed out of a dispensing tip into a multi-well plate.

DETAILED DESCRIPTION OF THE INVENTION

The invention will now be described in detail, with reference to the accompanying drawings.

In the accompanying drawings, like features are denoted by like numerals.

Figure 1 shows a schematic view of a simple embodiment of the invention that is capable of mixing four chemical solutions. It comprises a mixing conduit (1), which is connected at one end to a fluid metering mechanism (2) and at the other end to an outlet (3). Positioned at intervals along the conduit's length is a plurality of openings (4). Each opening is in fluid communication with a reservoir (5) and with a device (6) for moving fluid in this reservoir.

A branch of the microfluidic apparatus comprising an opening, a reservoir and a device for moving fluid is hereinafter referred to as a "liquid channel".

Downstream of the openings there is a detector (7) that can distinguish between various different fluids at a defined location within the lumen of a conduit. (The detector (7) is here shown associated with the mixing conduit, but this need not be the case -see below.) A carrier fluid (8) (a fluid that is immiscible with the chemical solutions to be mixed by the system) can be loaded into the mixing conduit, and a chemical solution (9) can be loaded into each reservoir. By operating one or more of the devices (6) for moving fluid in the reservoirs, droplets (10) containing one or more chemical solutions can be dispensed into the mixing conduit.

The carrier fluid will typically be a liquid, although in certain cases gases, or combinations of gases and liquids, can be used. For example, the carrier phase could comprise paraffin oil with air bubbles arranged periodically along the conduits that contain the carrier phase to prevent droplets from sticking to the conduits etc. Generally speaking, however, carrier fluids that are liquids are preferable for this invention because they are incompressible.

Constrictions can advantageously be placed at the openings (4) to increase the accuracy of dispensing.

Liquid handling may be improved by using a mixing conduit (I) with a circular cross-section. Similarly, if a reservoir (5) has circular cross-sections in the region close to the opening (6), liquid handling may be improved. At points further from the opening, other cross-sections can be used. This point will be dealt with in more detail below.

Figure Ia and several of the figures that follow show an apparatus with four channels, but clearly apparatuses with more channels can be used. A typical number of channels is 2 to 100.

The droplet shown in figure Ia is large enough to become elongated by the conduit that it is in. Droplets that have a diameter that is less than the diameter of the conduit that they are in may be difficult to position accurately. This is due to the Poiseuille flow profile where the flow in a conduit is faster near the center of the conduit. For the same carrier fluid flow a small (relative to the cross section) droplet will move faster than a large (relative to the cross section) droplet. Since a small droplet can move close to the wall of the conduit or near to the middle of the conduit, the speed at which it moves may vary. Large droplets move at more predictable speeds. Also, it may be difficult to add other chemical solutions to small droplets. It may therefore be preferable not to use conduits that have cross-sections that are greater than the diameters of the droplets that it is planned to use.

Figure lb shows a configuration where the mixing conduit (1) increases from a relatively small cross-section near the most upstream opening (20), to a relatively large cross-section at the most downstream opening (21).

As the droplet passes down the mixing conduit, it tends to increase as more chemical solutions are added to it. Therefore the conduit can be larger further downstream. It is helpful that the droplets are large enough to be elongated by the conduits, but if a droplet becomes excessively elongated it may break up into two or more drops, and it may show a tendency to stick to the walls of the conduit.

Figures 2a to 2d show a sequence of liquid movements that can be used to load a chemical solution into one liquid channel, thereby priming solution reservoirs and openings.

The start of the sequence is shown in figure 2a. The mixing conduit (1) and all of the reservoirs (5) have been filled with the carrier fluid (8).

In figure 2b, the fluid metering mechanism (2) and the device (6) for moving liquid in one reservoir are simultaneously operated. The reservoir is in liquid communication with a source of a chemical solution (not shown). Therefore the camer fluid and the chemical solution move as indicated by the arrows. Once the reservoir is fill, chemical solution will come out of the opening into the mixing conduit (I) where it will form droplets of undefined volume (11), which are broken off and move downstream towards the detector (7). The volumes of the droplets formed are determined by the flow rates of fluids and the geometry and physical properties of the channel.

In figure 2c, shows the situation when the first of the droplets of undefined volume has been detected by the detector. At this point the fluid metering mechanism (2) is inactivated while the device (6) for moving liquid continues to operate until it is certain that enough chemical solution has been dispensed to occlude the mixing condutt. This is done to make sure that no small volume of chemical solution that is insufficiently large to be moved by the carrier fluid remains within the mixing conduit, attached to the outlet by surface tension.

In figure 2d, the device (6) for moving liquid is inactivated while the fluid metering mechanism (2) continues to operate, thereby removing all droplets from the mixing conduit (1).

Solution has now been loaded into one reservoir. By repeating the sequence of figures 2b to 2d for each chosen liquid channel in turn, solution can be loaded into other reservoirs as desired.

Figures 3a to 3d show a sequence of liquid movements that can volumetrically calibrate one liquid channel. In figure 3a, after loading, one device (6) for moving liquid in a reservoir is operated for a suitable period, then inactivated. The chemical liquid moves as indicated by the arrow. This results in a droplet or slug (14) of chemical liquid moving into the mixing conduit (1).

Next, the fluid metering mechanism (2) is activated until the slug (14) is moved along the conduit (1) to a position where it is detected by the detector (7), as shown in figure 3b. Liquid movements are again indicated by arrows.

Jn figure 3c, the metenng mechanism (2) continues to be activated until the detector (7) detects the absence of the slug (14), and the presence of carrier fluid.

By using the information gained, especially the timing of the changes detected by the detector and the volume of camer fluid moved along the conduit to make the transition from the situation rn figure 3b to the situation in figure 3c, an estimate can be made of the volume of chemical solution dispensed into the slug. This volume information can in turn be used to calibrate the chosen liquid channel with its chemical solution.

For example, if the device (6) for moving solution in the reservoir comprised a source of constant pressure and a valve, an estimate could be made of the volume dispensed per unit time that the valve was open.

If the device (6) compnsed a pump, an estimate could be made of the volume dispensed per cycle of the pump.

Clearly the procedure of figures 3a to 3c can be repeated several times for one liquid channel for greater accuracy. It can also be performed with slugs of different volumes to see if departures from linearity are present. If necessary, a look-up table or mathematical function could be used to compensate for non-lineanty. Such a table can be used to specify the timing or movements of a device (6) that will move any desired volume of chemical solution.

Clearly also the procedure of figures 3a to 3c can be repeated for the other liquid channels. Note that the procedure calibrates a particular liquid channel in combination with a particular chemical solution. Any irregularities in the physical apparatus are automatically taken into account. In principle, it is not necessary to know anything about the viscosity, density etc. of the chemical solutions in advance.

Moreover, accuracy can be maintained even with highly viscous chemical solutions such as concentrated solutions of polyethylene glycol, DNA, polymers, glycerol, emulsions and suspensions. These viscous liquids will flow slowly into the mixing conduit, so longer periods of dispensing or more energetic dispensing is required to dispense a particular volume compared to non-viscous solutions.

Now we come to dispensing and mixing defined predetermined volumes of chemical solutions, shown in figures 4a to 4d.

Figures 4a to 4d show a sequence of liquid movements to dispense two chemical solutions from two liquid channels and form a droplet by mixing them.

In figure 4a, the apparatus starts with carrier fluid (8) in the mixing conduit (I), and with the chemical solutions (9) to be mixed in the reservoirs. In this example, two chemical solutions are present. First, the device (6) for moving liquid in a first reservoir is activated to a degree and duration that is known to dispense a droplet of the first chemical solution with the desired volume (15) into the mixing conduit (1) as shown.

In figure 4b, the metering mechanism (2) is activated to move the droplet (15) near to the opening of a second reservoir containing a second chemical solution.

In figure 4c, the device (6) for moving liquid in the second reservoir is activated to dispense the desired volume of the second chemical solution into the mixing conduit (1). As it comes into the conduit, this second solution coalesces with the previously-dispensed droplet to form a larger droplet (10).

To ensure coalescence, it may be necessary to dispense the second fluid slowly. This would be the case if the camer fluid was a viscous liquid. In other cases, such as using non-viscous carrier fluids, the solutions coalesce very rapidly. Surfactants can also be used to facilitate coalescence, see below.

In figure 4d, the metenng mechanism (2) is activated again, to move the final droplet (10) away from all openings of reservoirs. It can be moved past the detector (7) to confirm its size and position, and, possibly, its composition.

Thereafter, the droplet can be moved out of the outlet (3). It can then be stored within the microfluidic apparatus, or passed out of the apparatus and into e.g. a multi-well plate.

For clarity, figures 4a to 4d show the method applied to mixing only two chemical solutions. Clearly more chemical solutions can be mixed simply by repeating the steps of figure 4b and 4c for other liquid channels.

Note that the method of figures 4a to 4d can be used irrespective of the methods used to load and calibrate the system. In particular, the methods of figures 2a to 2c, and figures 3a to 3c, need not be used. For example, chemical solutions could be loaded into the reservoirs from the mixing conduit. Again, calibration may not be necessary. For example, accurate pumps that do not require calibration could be used, such as positive displacement pumps, piston pumps or gear pumps.

If the reactants and samples are aqueous, a carrier that is a paraffin oil, liquid silicone or liquid fluorocarbon could be used, with conduits made from or lined with a fluoropolymer such as PTFE, PES, PA!, PPS, PEEK, PFA, FEP, PCTFE, ETFE, ECTFE, polycarbonate, polypropylene, COC, polystyrene, nylon, acetal, silanized glass, or a silicone such as PDMS. Similarly, if the reactants and samples are hydrophobic, an aqueous or polar carrier fluid could be used with conduits with polar surfaces such as untreated glass. Fluorinated oils and surfaces are particularly suitable for conducting experiments with biological materials since they do not promote solution wetting by denaturing biological molecules.

It is very helpful, though not essential, for the conduits to be made from a transparent material so that operation of the system can be followed visually. FEP, PCTFE, polycarbonate, polypropylene, COC, polystyrene, glass and PDMS are examples of suitable transparent materials.

A variety of surfactants can be using in the system. Some surfactants will encourage chemical solutions to coalesce with each other. Others will discourage the droplets from sticking to the walls of conduits.

Still others prevent protein, cells or bacteria from moving to the surfaces of drops. Protein at drop surfaces may become denatured. Surfactants can be added to both the chemical solutions and to the carrier fluid. Examples of useful surfactants are I -s-Heptyl-B-D-thioglucoside, 1 -s-Nonyl-f3-D-thioglucoside, 1 -s-Octyl-B-D-thioglucoside, ABIL EM9O, ABIL WEO9, Anapoe 20, Anapoe 35, Anapoe 58, Anapoe 80, Anapoe CIOE6, Anapoe C1OE9, Anapoe CI2EIO, Anapoe C13E8, Anapoe X-l 14, Anapoe X-305, Anapoe X-405, BAM, C12E8, C12E9, C8E5, CHAPS, CHAPSO, C-HEGA-l0, C-HEGA-1 1, C-HEGA-9, CFAB, Cymal -l, Cymal -2, CYMAL -3, Cymal -4, CYMAL -5, CYMAL -6, CYPFOS-3, DDAO, DDMAB, Deoxy BigChap, FOS-Choline -lO, FOS- Choline -12, FOS-Choline -8, FOS-Choline -9, HECAMEG, HEGA-lO, HEGA-8, HEGA-9, Heptyl- B-D-thioglucoside, IPTG, LDAO, MEGA-8, MEGA-9, n-Decanoylsucrose, n-Decyl-B-D-maltoside, n- Decyl-B-D-thiomaltoside, n-Dodecyl-N,N-dimethylglycine, n-Dodecyl-B-D-maltoside, n-Dodecyl-B-D-maltotrioside, n-Hexadecyl-B-D-maltoside, n-Hexyl-13-D-glucoside, n-Nonyl-8-D-maltoside, n-Nonyl-B- D-maltoside, n-Octanoylsucrose, n-Octyl-B-D-glucoside, n-Octyl-B-D-thiomaltoside, Nonyl-B-D-glucoside, n-Tetradecyl-B-D-maltoside, n-Tridecyl-B-D-maltoside, n-Undecyl-B-D-maltoside, Pluronic F-68, SPAN 80, Sucrose monolaurate, Thesit , TRITON X-100, Tween 80, Zonyl FSN, ZWITTERGENT 3-10, ZWITTERGENT 3-12, Zwittergent 3-14,, Zwittergent -3-08. In addition, useful detergents can be made by substituting hydrogen atoms with fluorine atoms or vice versa in the detergents listed above. Also, adding polymers such as polyethyleneglycol or polydirnethylsiloxane can lead to useful properties of the detergents.

Figure 5 is an exploded view showing one possible method of constructing the apparatus by forming indentations or grooves (16) on the surfaces of one or more plates (17), and bringing two or more plates into face-to-face contact. At least one indentation of the appropriate shape and size is formed on the surface of a plate. Indentations can be formed by moulding, injection moulding, machining, blasting, etching, laser cutting, eroding, pressing, stamping etc. If this construction approach using plates is adopted, it is helpful for at least one plate to be made from a transparent material such those noted above.

Plates can be mated in face-to-face contact by gluing, bonding, vulcanizing, welding etc., or simply by clamping or pressing together.

If a face with one or more indentations is mated with a face with no indentations, then conduits with at least one flat side will be formed.

If indentations are formed on two faces that can be aligned with each other, then conduits with round cross-sections etc. can be formed.

Figures 6a to 6d show various methods of forming conduits by bringing plates into face-to-face contact.

The dotted lines in figures 6a to 6d indicate lines where the surfaces of plates have been brought into contact, welded etc. Figure 6a shows a conduit with a square cross-section, such as would formed by bringing a plate with grooves with vertical walls (18) into face-to-face contact with a plate with a flat surface (19). This configuration is easy to construct and works well in many cases. Square cross-sections may be preferable to rectangular cross-sections.

Figure 6b shows a conduit with a square cross-section, such as would formed by bringing three plates into face-to-face contact in a sandwich formation, where the middle plate (30) has elongated holes passing right through it in the form of the desired conduits. The sides of the holes form the sides of the conduits.

The middle plate is sandwiched between two plates with flat surfaces (19). Such a middle "plate" may be very thin -only a few microns, more like a film, and the pattern corresponding to the conduits can be cut by e.g. laser cutting.

Figure 6c shows a conduit with a D-shaped Cross-section, such as would formed by bringing a plate with a U-shaped groove (31) into face-to-face contact with a plate with a flat surface (19). This configuration is in some cases preferable to the square or rectangular cross-section of figures 6a and 6b.

Figure 6d shows the circular cross-section of a conduit such as would be formed by bringing together into face-to-face contact two plates with U-shaped grooves on them (31) such that the grooves are aligned to form one or more conduits. This configuration may give very good results, and it minimizes the tendency for the surfaces of droplets to be pressed against the walls of conduits. It also encourages the maintenance of a film of carrier fluid between the droplets and the walls.

To give reproducible and accurate dispensing, round cross-sections can be used for reservoirs (e.g. 5 on figure Ia), particularly in the region close to the openings (4). Round cross- sections here can reduce the tendency for carrier fluid to move into or out of the reservoir, which can give rise to Inaccuracy. At other points further away from the openings, other cross-sections, such as square or rectangular cross-sections, can often be used with impunity.

In some cases, the use of a particular chemical solution in combination with the chosen carrier liquid and solid surfaces of conduits can give rise to sticking of drops to conduit surfaces. One solution to this problem is to change the geometry of conduits. Conduits with round cross-sections may be particularly helpful. Other methods of avoiding sticking of drops include surface treatments of conduits and the use of surfactants in chemical solutions and in the carrier liquid.

Returning to figure 5, a simple way to make a connection with the conduits is shown, namely to make a cylindrical hole (32) that passes through a plate at approximately right angles to the mated faces. Tubing (33) can be inserted into these holes to make connections with external devices. It is sometimes helpful to flare the ends of the tubing, i.e. to increase inner diameter near the end until it approaches the outer diameter. This avoids a sudden "step" which could cause droplets to stick at the mouth of the tubing Where this method of connection is used, droplets may tend to come into contact with the walls of conduits as they move into the holes. This could give rise to sticking and associated contamination of droplets. Figure 7 shows a method of avoiding this problem at the outlet (3) of the mixing conduit (1).

Figure 7 shows extra inlets for introducing extra carrier fluid at or near the point where a conduit is connected to a hole that passes through a plate. At least one inlet (34) is provided where extra carrier fluid can be injected. This causes droplets to move rapidly through the outlet (3), and can also keep them away from conduit walls.

Figure 8 shows that a detector could be associated with e.g. a flexible tube (33) downstream of the mixing conduit (1).

The detector (7) is an important element of the invention. It can be placed in any position downstream of the last opening where chemical solutions are dispensed. For simplicity, in all previous drawings, it was shown associated with the mixing conduit. A more practical position might be on a flexible tube (33) downstream of the outlet (3). However, it should be noted that droplets may move at disparate speeds in a conduit or tube. For example, more viscous droplets will tend to move at different speeds than less viscous ones, and smaller droplets at different speeds than larger ones. It is therefore helpful to place the detector reasonably close to the mixing conduit.

It may be helpful to have two or more detectors. For example, one detector can be placed close to the mixing conduit to allow loading and calibration of the system as described above, while another detector can be placed close to a dispensing tip e.g. to allow the dispensing of droplets together with the minimum of carrier fluid.

Figure 9 shows a schematic view of an embodiment where a pressure chamber is used with valves in a time-and-pressure approach to dispensing chemical solutions. The apparatus comprises a pressure chamber (35) which has a means of establishing or maintaining constant pressure (36) such asa pump or a regulated gas cylinder or supply. The pressure chamber is connected to a plurality of sealed vessels (37) that supply chemical solutions. These vessels could be e.g. the wells of a 96-well plate that have been sealed. However they could also be dedicated inlets on the device, where solutions can be directly deposited. The vessels are connected via valves (38) to the reservoirs (5), described above.

When a valve opens, the chemical solution in the connected reservoir moves into the mixing conduit (1) to form a droplet (15). The volume of the droplet corresponds to the time that the valve remains open.

To slow up dispensing, at least some of the connectors or reservoirs should have a narrow cross-section so that viscous drag can reduce the flow of chemical solutions, thus allowing small volumes to be dispensed. This is important because in practice valves cannot reliably open for very brief periods.

Moreover, if low pressures are used to slow up dispensing, surface tension effects may reduce the accuracy of dispensing. A region with narrow cross-sections (39) is shown on the conduits that connect the vessels to the valves (38).

Figure 10 shows a schematic view of an embodiment where solutions are dispensed in discrete steps using pairs of valves located in each channel. Like the embodiment disclosed in figure 9, this comprises a pressure chamber (35), a means of establishing or maintaining constant pressure (36) and a plurality of sealed vessels (37) that supply chemical solutions.

Two variations of the embodiment exist: in the first, the two valves (51), (52), or the conduit (39) or reservoir that connects the two valves, are formed or partly formed from an elastic matenal. In the second variation, a side-channel containing gas (50) is provided. Such a side-channel is shown only on channel 1 disclosed in figure 10, but would be present on all channels if the channels were constructed from a hard material.

Each liquid channel has an upstream valve (51) and a downstream valve (52).

Using both variations disclosed above, fluid is metered by actuating the valves in various defined sequences. An example of a 4-step sequence of activating the valves follows. In the following table, the first (left) word indicates the state of the upstream valve (51), the second (right) word the stale of the downstream valve (52): Step 1: closed, closed step 2: open, closed step 3: closed, closed step 4: closed, open.

During step 1 no liquids flow. During step 2, pressure supplied by the pressure chamber (35) causes the chemical solution to flow into the reservoir or conduit between the valves (39). (This happens in variation I because the valves, reservoir or conduit expands, and m variation 2 because the gas in the side-channel (50) is compressed.) During step 3 no liquids flow. In step 4, a quantum of chemical solution passes through the downstream valve (52), so that an equivalent volume of chemical solution moves into the mixing conduit forming, or adding to, the droplet (15).

By multiple repetitions of the cycle the desired volume can be dispensed. The final volume of each chemical solution that can be dispensed will have a "graininess" that is equal to the quantum volume dispensed in each cycle.

Increasing or decreasing the pressure in the pressure chamber (35) increases or decreases the quantum volume of each cycle.

Many other configurations exist that use three or more valves. For example, a well-known variation of a peristaltic pump called a finger-pump can be used, which uses three valves.

Figure ha shows a schematic view of an embodiment where metering of all liquids is performed by a single metering mechanism (2) that is placed downstream of the mixing conduit. A valve (38) is provided on each liquid channel and a valve (40) is also provided on the supply of carrier fluid. A pressure chamber (35) and a means of establishing or maintaining constant pressure (36) are provided. A plurality of sealed vessels (not shown in figure 1 Ia) can be provided for loading chemical solutions as in the embodiments disclosed above. A detector (7) is also provided. A valve (42) is provided downstream of the mixing conduit (I), which allows droplets to flow out of the mixing conduit. In this example, a flexible outlet tube (33) is shown, although this tube is not essential.

To dispense a chemical solution from a reservoir (5) into the mixing conduit (I) to form a droplet (15), the valve (38) on that liquid channel opens, and, at the same time, the metering mechanism (2) withdraws carrier fluid into a side arm (41).

To move the carrier fluid in the mixing conduit (1), all valves close except for the valve (38) on the carrier fluid channel which opens, and, with the valves remaining in this state, the metering mechanism (2) withdraws carrier fluid into a side arm (41).

Using these two operations and combinations of them, all of the liquid handling methods of figures 2a - 2d, figures 3a -3c and figures 4a -4c can be carried out. To move droplets out of the mixing conduit, the valve (38) on the carrier fluid channel opens at the same time as the valve downstream of the mixing conduit so that droplets move out of the mixing conduit e.g. into an outlet tube (33).

To prevent fluids from moving too fast, constrictions can be provided in conduits or reservoirs. As an example, a constriction (43) is shown on the carrier fluid channel which would slow up the movements of camer fluid by increasing viscous drag.

Figure 1 lb shows a schematic view of an embodiment that is similar to the embodiment disclosed in figure 11 a, but it is able to pick up and mix a large number of samples from e.g. one or more 96-well, 384-well or 1536-well etc. plates. This apparatus possesses a "sipper" (53), which is a tube that is able to dip into different stock or sample solutions (54) to pick up many different samples in series. For example, a sample can be picked up from a well, then the sipper can move into a volume of carrier liquid, and carrier fluid can be sucked up, thereby defining a droplet of sample in carrier fluid. Carner fluid can be placed over the samples (60), or placed in one or more separate containers (61).

Figure 12 shows a schematic view of another embodiment that is able to pick up and mix a large number of samples from e.g. one or more 96-well plates. This apparatus also possesses a sipper (53). The sipper is connected to a rotary valve (55) that is situated at the intersection of two liquid channels. Each of these two channels possesses a device for moving fluids (56), (57).

When the valve is in the position indicated by dotted lines, device (56) is in communication with the sipper. When the valve is in the position indicated by solid lines, device (57) is in communication with a microfluidic system (58) similar to that of e.g. figure 1.

By turning the rotary valve (55) to the position indicated by the dotted lines and activating device (56), solution can be drawn up the sipper and through the rotary valve (55). The rotary valve now turns through 90 degrees to the position shown by solid lines, bringing a sample of solution (70) that is held on the bore that passes through the barrel of the rotary valve into the second channel (71). This second channel has previously been filled with carrier fluid. By activating device (57) the sample is moved as a droplet into the main microfluidic system (58).

Obviously, many possibilities exist for analyzing droplets. One example is shown in figure 12. Droplets of mixed solutions (10) are shown in figure 12 coming out of the main microfluidic system (58) in a flexible tube (33), and passing through an analysis station (72). This analysis station could use absorption, scattering, radiolabelling, fluorescence, fluorescence microscopy, confocal microscopy etc, or any other suitable method to analyze the samples in droplets.

In many cases it is necessary to mcubate mixtures at specific temperatures for specific times in order for chemical reactions etc. to take place. Clearly there are many possible approaches to incubation in a microfluidic system. One simple method is to provide a temperature-controlled environment and to place the flexible outlet tubing (33) in such an environment. For example the analysis station (72) of figure 12 could be replaced by a temperature-controlled environment, or a heater. When droplets pass along the flexible tubing their temperature can be controlled by such an environment. to

Figure 13 shows a schematic view of a system for automatic protein crystallization, where chemical solutions including protein are mixed and passed out of a dispensing tip and into a multi-well plate. The apparatus comprises a pressure chamber (35) that is connected to vessels (37) that provide chemical solutions. A suitable number of vessels for this application would be around 3 to 50. These vessels could be the wells of e.g. a 96 or 384-well plate. The vessels are connected to a microfluidic system (58) that works on the same pnnciples of operation as the apparatus of e.g. figure 1. A source of carrier fluid is also provided (74). Since protein may be the solution that has the greatest tendency to stick to and coat the walls of the conduits, it can be loaded into the last (furthest downstream) liquid channel.

The microfluidic system is controlled by a computer program for designing crystallization experiments (75). A suitable program would be able to design all crystallization experiments including the step of initial screening for crystallization conditions; the step of a targeted screen that explores crystallization space around one or more conditions that give crystallization in order to improve the choice of ingredients; the step of setting up multidimensional experiments to optimize expenmental variables especially the concentrations of ingredients; and finally, the step of growing sufficient crystals for protein structure determination by x-ray diffraction analysis. All of these steps can be carried out by factorial experimental designs or by other methods. Examples of factors in protein crystallization experiments are protein concentration, choice of buffer, choice of precipitant, concentration of buffer, concenflation of precipitant, choice of additive etc. Examples of factorial experimental designs include incomplete factonal, fractional factorial, central composite, Box-Behnken, complete factorial, grid or linear designs.

Once again, droplets of mixed solutions (10) are moved along a flexible tube (33), past a detector (7), and into a target plate (76). The tip of the flexible tube is moved by an arm (77). A magnified view of a well of the target plate shows a protein crystal (78) growing. The crystal is shown here growing using the microbatch-under-oil method. To crystallize protein using the vapor diffusion method, 96-well plates designed for sitting drop crystallization can be used. The droplets, along with some carrier fluid, can be dispensed into the sample wells. The carrier liquid can be allowed to remain on the droplet. If it is desired that the carrier fluid should be removed, low molecular weight paraffins, silicones or fluorocarbons that are volatile can be used as the carrier fluid, which will evaporate.

II

Claims (10)

1. What I claim is: 1. A microfluidic system for combining chemical solutions comprising an apparatus and a method, where the apparatus comprises: (a) a conduit having a distal and a proximal end, (b) a metering mechanism connected to the distal end of said conduit, (c) an outlet at the proximal end, (d) a plurality of liquid channels, where each liquid channel comprises an opening into said conduit, a reservoir in fluid communication with said opening, and a device for moving fluids in said reservoir, and (e) at least one detector; and the method comprises the steps of: (a) moving a carrier fluid, which is immiscible with the chemical solutions that are to be mixed by the system, into said conduit, (b) loading chemical solutions into said reservoirs, (c) operating one of said devices to move a known volume of a first chemical solution through a first opening into said conduit, (d) where required, operating said metering mechanism to move carrier fluid, thereby moving said known volume along the conduit to a position near a second opening, (e) operating a second device to dispense a known volume of a second chemical solution into the conduit causing it to coalesce with said first chemical solution, (f) repeating steps (d) and (e) until all the desired chemical solutions have been mixed together, (g) operating said metering mechanism to move the carrier fluid, thereby moving the mixture away from said openings.
2. 2. A system as claimed in claim 1 where said chemical solutions are loaded into said reservoirs at step (b) therein by the steps of: (a) operating said metering mechanism and at the same time operating one of said devices, (b) inactivating said metering mechanism when droplets of a chemical solution are detected by at least one detector, (c) continuing to operate said device until it is estimated that enough chemical solution has been dispensed to block the conduit, (d) stopping said device and operating said metering mechanism to break off the last droplet and move all droplets produced out of the conduit, and (e) repeating steps (a) to (d) above until all the desired chemical solutions are loaded into reservoirs.
3. 3. A system as claimed in claims I or 2 where said system is volumetrically calibrated before step (c) therein by the additional steps of: (a) Transitorily operating said device in a first liquid channel in order to move an initially unknown volume of one of said chemical solutions Out of an opening and into said conduit, (b) inactivating said device, (c) operating said metering mechanism to move carrier liquid, thereby moving said unknown volume until at least one detector detects its arrival, (d) continuing to operate said metering mechanism until at least one detector detects the departure of said unknown volume and the arrival of carrier liquid, (e) using the information gained, in particular the volume of carrier liquid moved during step (d), to estimate the volume of said chemical solution that was dispensed at step (a), thereby calibrating the effectiveness of said device in dispensing said chemical solution from said first liquid channel, I2 (f) repeating steps 2 to 5 for further liquid channels to calibrate them;
4. 4. An apparatus as claimed in claim I in which conduits are formed by clamping together two or more plates, one or more of the plates having grooves or indentations on their surfaces, which grooves form conduits when the plates are clamped together, where connections are made to these conduits by forming bores through one or more plates, wherein the improvement compnses the addition of one or more extra conduits that meet the a conduit close to its connection with a bore, allowing extra fluid to be added at or close to this point of connection.
5. 5. An apparatus as claimed in claims 1 to 3 where said device for moving fluids is a valve and a source of constant pressure, liquids being metered by opening said valve for predetermined intervals of time.
6. 6. An apparatus as claimed in claims I to 3 where said device for moving fluids comprises a pair of valves and a source of constant pressure where the valves can be operated in turn to allow a small volume of chemical liquid to pass through the pair of valves on each cycle of operation.
7. 7. An apparatus as claimed in claims I to 3, in which liquids are moved by positive displacement, which comprises a plurality of liquid channels, each one of which possesses at least one valve, where chemical solutions can be moved into a mixing conduit to form or add to droplets by opening a valve on a liquid channel and withdrawing camer liquid from the mixing conduit using a fluid metering mechanism.
8. 8. An apparatus as claimed in claims I to 3 which is in fluid communication with a sipper device which is able to pick up a plurality of samples of chemical liquids by dipping into them in sequence.
9. 9. An apparatus as claimed in any previous claim that possesses a flexible tube along which droplets of mixed chemical solutions can pass.
10. 10. An apparatus as claimed in any previous claim that can carry out macromolecular crystallization by mixing macromolecules with chemical additives in a microfluidic device and then bringing mixtures out of the microfluidic device and dispensing them in an array for incubation and crystallization. .3