GB2355567A

GB2355567A - Fluidic multiplexer

Info

Publication number: GB2355567A
Application number: GB0019769A
Authority: GB
Inventors: Ray Allen; John Russell Tippetts; Vaclav Tesar
Original assignee: University of Sheffield
Current assignee: University of Sheffield
Priority date: 1999-10-20
Filing date: 2000-08-11
Publication date: 2001-04-25
Anticipated expiration: 2020-08-11
Also published as: GB0019767D0; GB9924776D0; GB2355567B; GB0019770D0; GB2355543A; GB0019769D0; WO2001029435A1; GB2355414A; AU7808400A; GB2355566A

Abstract

A fluidic multiplexer for selectably sampling a number of fluid flows 102 to 110 uses fluidic valves 114 to 122 to control the sampling of, for example, the reaction product flows by selectably turning on and off those valves. The valves do not comprise moving mechanical parts. Therefore, advantageously, the valves do not fail as a consequence of mechanical fatigue. One of the fluid flows 102 to 110 is connected to an output 112.

Description

2355567 FLUID MULTIPLEXER The present invention relates to a fluidic

multiplexer for sampling a number of fluid flows. The fluid flows may be taken from, for example, a number of chemical microreactors.

Modern chemical engineering has a tendency to use very small reactors. These reactors are often of submillimeter size and are sometimes constructed on silicon chips using the same technology used to- fabricate microelectronic or VLSI circuits. It can be appreciated that the output rate of reaction products from a single microreactor is very small. Therefore, to achieve a required productivity level, a large number of such microreactors may be operated simultaneously.

one of the advantages of the small size of the microreactors is the capability to control precisely the process conditions under which the reaction takes place. This capability follows as a consequence of the effective and fast response of the microreactors to control actions. It is often possible to have a controller resident on the same chip as the microreactor. The 2S typical parameters that can be controlled in such a microreactor include the temperature or pressure in the reactor and the product composition.

A product composition analyser can be used to monitor the operation of a reactor. When operating or testing, for example, a catalytic process, a product composition analyser is typically connected to the output of the microreactor. It can be appreciated that the 2 analyser must be appropriate to or conf igured for the reaction product produced by the catalytic process under test. It is clearly not a cost effective option, in circumstances where more than one chemical reactor is 5 being monitored, to have one analyser per reactor. Whenever the variations in the processes are sufficiently slow, that is, the processes do not vary significantly during a sampling cycle, a single analyser may be used to monitor a number of reactors.

Ehrfeld W. (Ed) "Microreaction Technology Industrial Prospects", Springer, Berlin 2000, ISBN 3-54066964-7 discloses multiplexers that are suitable for taking fluid samples from mini- reactors and supplying is them to a product composition analyser. These known multiplexers operate using moving components. Therefore, they are large, that is, cannot be fabricated on a chip, require external mechanical drives, are prone to malfunctions, that is, breakage and/or seizure of the moved or flexed parts, and are not suitable for high temperature applicaions.

It is an object of the present invention to at least mitigate some of the problems of the prior art.

2S Accordingly, a first aspect of the present invention provides a fluidic multiplexer for supplying to a common outlet channel one fluid selected from at least a first fluid channel and a second fluid channel each for carrying respective first and second fluid flows; the multiplexer comprising a first fluidic valve to prevent to flow of the first fluid from the first fluid channel to the common outlet channel in response to flow of first 3 control fluid from a first control inlet; and a second fluidic valve to prevent the flow of the second fluid from the second fluid channel to the common outlet channel in response to flow of a second control fluid 5 from a second control inlet.

Advantageously, the first and second fluids represent, in a preferred embodiment, the reaction products of respective chemical reactors. These reaction products -can be selectably directed to a common outlet channel under the control of the fluidic valves. The output of the common outlet channel will be either the first fluid or the second fluid according to whether the first or second fluidic valve has been actuated to is prevent the flow of the first and second fluids respectively.

It can be appreciated that a single analyser can be used to sample reaction products of the two chemical processes occurring in the microreactors that are used to feed the first fluid channel and the second fluid channel respectively. Further, the use of a single analyser finds particular application in chemical processes that vary slowly as compared to the time required for the analyser to cycle through all chemical processes.

Fluidic valves that operate by preventing or restricting the flow of a reaction product out of a corresponding supply nozzle may adversely affect the process conditions in a reactor by, for example, reducing the flow rate of reactants or reaction product through the reactor or by varying the temperature or pressure in the reactor. It can be appreciated that such changes in 4 a catalytic process can be disadvantageous.

Accordingly, an embodiment of the present invention provides a fluidic multiplexer in which the first fluidic valve comprises a first vent arranged to allow flow of the first fluid through the first vent in response to the first control fluid preventing the flow of the first fluid to the common outlet channel.

A further embodiment of the present invention provides a fluidic multiplexer in which the second fluidic valve comprises a second vent arranged to allow flow of the second fluid through the second vent in response to the second control fluid preventing the flow of the second fluid into the common outlet channel.

Advantageously, the fluid flow in the first and second fluid channels can be arranged to be continuous even though the first and/or second fluidic valves has/have been actuated to prevent corresponding fluid flow to the common outlet channel. This ensures that the reactor conditions or chemical process conditions remain substantially constant, that is, the flow rate, amongst other things, through a reactor may be maintained at a 2S given value rather than there being a temporary cessation during sampling of another reaction product carried in another fluid channel.

A further embodiment of the present invention provides a fluidic multiplexer arranged to control selectably the flow of first and second fluids into a common outlet channel using first and second fluidic valves, the first fluidic valve having a first inlet to supply the first fluid to a first outlet channel and a first control inlet to prevent flow of the first fluid to the first outlet channel; the second fluidic valve having a second inlet to supply the second f luid to a second outlet channel and a second control inlet to prevent f low of the second f luid to the second outlet channel; the first and second outlet channels being arranged to feed the common outlet channel.

Often the process conditions within a reactor and in relation to a reaction product are unsuitable for an analyser, which can be a relatively expensive and delicate item of equipment. For example, the pressure of a reaction product may be incompatible with the operating conditions of an analyser.

Accordingly, an embodiment of the present invention provides a fluidic multiplexer comprising a pressure regulator to establish, in use, a predeterminable pressure in the common outlet channel.

Advantageously, the predeterminable pressure can be set to a pressure value that is acceptable to the analyser notwithstanding the pressures required by the chemical process under test.

A still further embodiment of the present invention provides a fluidic multiplexer further comprising a pressure regulator for changing the pressure of at least one of either the first and second fluids from respective first and second pressures to a selectable pressure prior to feeding the common outlet channel.

6 A further advantage of the embodiments of the present invention is that fluid sampling can be undertaken without using any moving components. Therefore, the fluid multiplexer is capable of operating under adverse conditions such as high temperature and/or in an aggressive fluid chemical environment under which corresponding electromechanical fluid valves would fail.

Still further the operation of the valve is not adversely effected by mechanical acceleration, and/or vibration.

Embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings in which: figure 1 shows schematically a fluidic multiplexer; figure 2 illustrates an embodiment of a fluidic valve for use in the multiplexer shown in figure 1; figure 3 illustrates a further embodiment of a fluidic valve that can be used in the multiplexer shown in figure 1; figure 4 illustrates a still further embodiment of a fluidic valve that can be used in the multiplexer shown in figure 1; 25 figure 5 illustrates an embodiment of a fluidic multiplexer that uses a fluidic valve as shown in figure 3; figure 6 illustrates an embodiment of a multiplexer that uses fluidic valves as shown in figure 4; and 30 figure 7 illustrates a further embodiment of a multiplexer that uses fluidic valves as shown in figure 4.

7 Referring to figure 1 there is shown schematically a fluidic multiplexer 100 for directing a plurality of f luids 102 to 110 to a common outlet channel 112 using a plurality of fluidic valves 114 to 122 which can be selectively actuated to prevent or allow flow of a corresponding one of the plurality of fluids 102 to 110. optionally, each channel for carrying the plurality of fluids 102 to 110 is interrupted by a gap which forms 10 a corresponding vent 124 to 132 that is arranged to maintain the fluid flow 102 to 110 of the plurality of fluids 102 to 110 even upon actuation of at least one of the fluidic valves 114 to 122. In the absence of such vents, the fluid flows 102 to 110 would have to be is terminated.

In the schematic embodiment shown in figure 1, there are only five inlets 134 to 142 for receiving, for example, corresponding reaction products, that is, the plurality of fluid flows 102 to 110. However it will be appreciated that the present invention is not limited thereto and, in practice, a significantly greater number of reaction product flows can be provided.

Each of the fluidic valves 114 to 122 is actuated by a corresponding flow of fluid, that is, a corresponding flow of control fluid 144 to 152. Each fluidic valve 114 to 122 comprises a respective control nozzle 154 to 162.

The dimensions of each of the channels and control nozzles 154 to 162 are arranged, taking into account the velocities of the control fluids 144 to 152 and the reaction products 102 to 110 and the viscosities of the control fluids 144 to 152 and the reaction products 102 8 to 110, to result inextremely low respective Reynolds numbers such that the presence of a control fluid 144 to 152 in the path of a reaction product f low 102 to 110 can be, in some embodiments, sufficient to stop or at least reduce that reaction product flow 102 to 110.

Preferably, the Reynolds numbers are less than 100 and more preferably less than 40.

An embodiment is provided in which such a blocking effect is achieved by ensuring the viscosities of the control fluids 144 to 152 are greater -than the viscosities of the reaction products 102 to 110. Alternatively, the control fluids 144 to 152 and the reaction products 102 to 110 can be arranged to be mutually immiscible. Still further, an embodiment is provided in which the control fluid is a liquid and the reaction product is a gas. In such a situation the control fluid is strongly held in position by the action of surface tension and viscous forces.

It can be appreciated from figure 1 that the fourth fluid flow 108 is the fluid that is coupled to the common outlet channel 112 while the remaining reaction product flows 102, 104, 106 and 110 are vented via the corresponding vents 124, 126, 128 and 132 to a common vent 164 in response to actuation of fluidic valves 114, 116, 118 and 122. Accordingly, only the reaction product of the fourth channel is coupled to the common outlet channel for sampling by an analyser.

Referring to figure 2, there is shown, according to an embodiment, a fluidic valve 200 for controlling the flow of reaction product 102 to the common outlet channel 9 112. It can be appreciated from figure 2 that the fluidic valve 200 is shown in a closed state. The embodiment shown in figure 2 can be manufactured by etching substantially planar material to an appropriate depth. It can be appreciated from figure 2 that since the fluidic valve 200 is in a closed state the reaction product 102 will continue to flow via the channels 202 and 204 that form part of the vent rather than flowing into the outlet channel.

The action of the control fluid -in preventing the flow of the reaction product 102 shown in figure 2 can be accomplished by ensuring the control fluid 114 is a very viscous liquid and, preferably, immiscible with fluids to is be sampled, that is, the reaction product flows. If the above conditions are satisfied, only a small amount of control fluid needs to be introduced into the path of a reaction product flow. However, if the control fluid and the reaction products are miscible and the control fluid cannot be expected to be held in or to maintain a stationary position by the action of viscosity or under surface tension, it will be appreciated that some of the control fluid may be carried into the common outlet channel 112. It will also be appreciated that, under such circumstances, the flow of some of the control fluid into the common outlet channel 112 may also take place during initial closure of the fluidic valve 114. During the initial stages, that is, during partial closure of the fluidic valve 200, a small amount of the control fluid emerging from a corresponding nozzle into the relatively strong full flow of the reaction product 102 will result in a small amount of that control fluid being carried into the common outlet channel 112. it is only after delivery of a sufficient volume of control fluid 148 into the path of the reaction product 102 that the fluidic valve will close and prevent the flow of the reaction product 102 into the common outlet channel.

s While, under some circumstances, the flow of the control fluid into the common outlet channel may be acceptable, under other circumstances such flow may not be acceptable. Therefore, an embodiment is provided in which the control fluid is selected so that an analyser connected to the common outlet channel 112. does not generate a corresponding signal in response to the presence of such a control fluid.

is Still further, under certain circumstances, the flow of the control fluid into the analyser may represent a relatively large volume of the sample taken by the analyser and accordingly decrease, by way of dilution of the sample, the effectiveness of the analysis, that is, the analyser will have a product concentration threshold below which a reaction product cannot be reliably analysed. Therefore, an embodiment provides a fluidic valve 114 in which the flow of control fluid is arranged to have a greater component opposing the direction of flow of the reaction product as compared to the component in the direction of flow of the reaction product. Referring to figure 3 there is shown an embodiment of a fluidic valve for achieving this aim. In the embodiment shown it can be seen that figure 3 provides a first channel 302 for carrying a reaction product 102. The first channel comprises a first nozzle 304 for directing the reaction product 102 towards a first outlet 306 that is arranged to feed a common outlet channel 112 via a pair of symmetrically shaped conduits 308 and 310. There is also provided a control nozzle 154 that is arranged to produce a flow of control fluid 312 which opposes the flow of reaction product through the first outlet channel 306. It can be appreciated that under certain circumstances the control fluid may extend into a vented cavity 124. The control nozzle is contained within the first outlet channel that leads the common outlet channel in the embodiment shown. This arrangement reduces the extent of any backflow 314 of control fluid into the outlet channel during valve actuation or operation.

It can be appreciated that the embodiment shown in figure 3 may be manufactured by CVD and etching as is appropriate.

Referring to figure 4 there is shown a further embodiment of a fluidic valve 400 for achieving a more effective suppression of flow of the control fluid into the first outlet channel 402. It can be appreciated that the direction of flow of the reaction product and the control fluid are mutually opposite, and the control nozzle is inclined relative to the first outlet channel. The first outlet channel 402, as with the above embodiments, is arranged to feed the common outlet channel 112. The embodiment shown in figure 4 has the advantage that there does not result an island of etched material, such as island 316 shown in figure 3, which would lead to manufacturing complications if the valve structures from which the valves are constructed were deposited on a suitable substrate after etching rather than prior to etching. It can be - appreciated from figure 4 that at relatively high Reynolds numbers such as, for 12 example, 200, the jet or relatively fast flowing stream of control fluid would prevent the formation of a flow of the control fluid into the common outlet and, preferably, generate a suction effect which would draw the fluid S contained within the first outlet channel 402 in a direction which opposes the flow of supply fluid towards the common outlet channel 112.

Referring to figure 5 it can be seen there is 10 disclosed- a parallel arrangement of first and second channels 502 and S04 for carrying respectivereaction products. The first and second channel S02 and 504 are directed towards respective outlet channels 506 and 508 that are arranged to feed a common outlet channel 510.

The control nozzles 512 and 514 are arranged to direct respective control fluids such that they oppose the flow of reaction products flowing via the first and second inlet nozzles 516 and 518. It can be appreciated that the first and second inlet nozzles are substantially wider than the corresponding inlets 520 and 522 of the outlet channels and the control nozzles 512 and 514.

There is also provided, in the embodiment shown in figure 5, a common vent 524.

Referring to figure 6 there is shown an embodiment of the present invention in which each of a number of fluidic valves 602 to 608 controls the flow of a respective reaction product 610 to 616 using respective control fluids 618 to 624 fed via control fluid nozzles 626 to 632. The embodiment is arranged to control the flow of reaction products 610 to 616 into corresponding outlet channels 634 to 640. It can be seen from figure 6 that pairs of outlet channels of the fluidic valves, that 13 is, outlet channels 634 and 636 and outlet channels 638 and 640, merge into f irst 642 and second 644 common outlet channels. The f irst 642 and second 644 common outlet channels also merge to form an overall common outlet channel 646. The merging pairs of outlet channels have the advantage that they at least reduce or prevent the loss of a fluid sample currently under investigation via backf low into the channel of a f luid that is not current under investigation. The same applies to corresponding features of figure 7.

It can be seen from figure 6 that each fluidic valve 602 to 608 has a vent 648 to 654. Each vent 648 to 654 has a corresponding opening 656 to 662 which is formed in is an overlaying, vertically disposed, plate or larger having respective holes 656 to 662 etched therein.

It can be appreciated that an alternative embodiment to that shown in f igure 6 is one in which, rather than having separate respective vents 648 to 654, at least two of the vents can be combined into a common vent. A still further embodiment is envisaged in which all of the vents 640 to 654 are combined into a single vent as shown in figure 7 at 700.

Models tested in the laboratory used Syngas" as a supply fluid at a temperature of 4000C and a viscosity of SO x 10-6 m2/s. The velocity of the supply f luid at the nozzle exit was 5 m/s. The laboratory models had the following dimensions: supply nozzle width 0.34 mm, control nozzle width 0. 24 mm, and the nozzle depths were 0.15 mm. The gap between the supply nozzle exit and the output channel entrance was 1.14 mm. However, it will be 14 appreciated in actual microfluidic applications, the dimensions will be smaller. Typically, the dimensions may be three to five times smaller.

It can be appreciated that, due to the very low Reynolds numbers used in some or all of the embodiments of the present invention, the mere presence of a blockage or control fluid can be sufficient to close a valve since, in those embodiments, no use of fluid inertia is made to -close the valve as in the prior art. The extremely low Reynolds numbers are such that the inertial effects are quickly damped or countered by fluid viscosity. Typically, the Reynolds numbers of the fluid flows of the embodiments of the present invention will be is less than 100 and in some instances can be less than 40.

Embodiments of the present invention utilise a 0.34 mm. nozzle width. Each value occupies about a 5 mm x 5mm area. A complete fluidic multiplexer occupies a 80 mm x 15 mm area. The depths of the channels of the above embodiments are 0. 1016 mm.

It can be seen that the above embodiments have been described within a catalysis context, that is, within the context of controlling the flow of reaction products. However, it will be appreciated that the present invention is not limited thereto. Embodiments can be realised in which the fluids flows that are controlled are fluids other than reaction products.

The magnitudes of the fluid flows of conventional fluidic valves are such that the supply fluid flow is significantly greater than the control fluid flow.

However, the embodiments of the present can equally well use fluid flow magnitudes in which the control fluid flow is at least equal to or significantly greater than the supply fluid flow. In some embodiments that control fluid flow may be at least 10 times greater than the supply fluid flow.

In applications such as, for example, high throughput catalysis testing, the range of volumes flows for the above embodiments may be between 10 cubic centimetres per minute and 30 cubic centimetres per minute, which corresponds to 60.10-rkgls to 180.10-'kg/s. The above volume flow rate limits were imposed, in practical realisations, by the operating parameters of is the infrared analyser used to undertake the testing and do not reflect any hydrodynamic limitations of the above embodiments. it will be appreciated that the hydrodynamic limitations are determined from the Reynolds number, Re, by Re. h. v M[kg / s] = V where fluid viscosity v= fluid specific volume; and h= the cavity depths in metres.

2S At one end of the spectrum, the skilled man could envisage the flow of a gas at 0=10.10-6 m2/s, v= O.S m/kg at an Re = 1 in a microchannel f or which the nozzle depth is 5Lm, which give a value of M=100.10-2 kg/s. However, it will be appreciated that the other end of the spectrum may have the following prevailing conditions 0=10.10-6 m 2/S, V= 10.10-3 M3 /kg at an Re = 1000 in a microchannel 16 for which the nozzle depth is 2mm, which give a value of M=20.10-' kg/s.

Furthermore, in the above embodiments it is envisaged that the control flow could be between 0.1 and 5 times that of the supply flow.

It will can seen from the attached appendix that there is disclosed a technique for maintaining required pressure conditions within the valves that are utilised in the above embodiments. The appendix has been included in the application in its current form for convenience only. it is fully intended that the technical information contained in the Appendix forms part of the is disclosure of the present invention and represents possible or preferable operating conditions for the above embodiments. Furthermore, it will be appreciated that the technical information contained within the Appendix can be used to maintain constant pressure conditions within a microfluidic valve.

MICROFLUIDICS THE CHALLENGE OP LOW Re FLOW CONTROL ViclaVMSA PayWXA=N, JohnR_TIPP=S University of Sheffield, United Kingdom

Abstract Although real fluidic valves - and circuits in Microfluidics is a technology of generating and control- which they are used - are often more complex, the simple ling fluid flows - preferably without moving components - in example of the jet-type diverter in Fig. I may be useful for micron-sized structures. The basic problem is the low Reynolds clarifying some basic facts. The valve is here placed be number: inertial effects used in large-scale fluidics are too tween the source (which may be a pump), and the device small relative to viscous dissipation and new approaches are re- in which the flow is to be varied. This device represents a quired In the diverter valves able to operate in the subdynamic load for the valve. The source delivers a constant supply regime, developed by the authors, pressure difference is used in flow 0M. supplied into the nozzle while the load re place of the inertial effects. A new characterisation number was ceives the variable valve output flow C) My captured by the introduced replacing the Reynolds number, which ceases to be collector opposite the nozzle, which re-converts the jet important in this regime. into a closed cavity flow. The constantcy of the supply conditions are commonly maintained by a supply regula Key words. Fluidics, microfluidics, flow control, tor. Because of the usually invariant properties of the sup microvalves, Reynolds- number ply circuit, a pressure regulator is commonly used, Fig.2.

Note that the pressure difference maintained constant is that across the source. Usual applications involve many 1. In trodtiction valves supplied from a common source so that the pres ence of the regulator is not a large complication as just a The last decade saw a fast growth of a new technosingle one usually suffices for the whole fluidic system.

logical field of microdevices, manufactured by methods The output flow variation is achieved by the action originally developed for semiconductor electronics. of a control effect. This may be acoustic, electric - but in a Microchips have been demonstrated containing mechan- typical fluidic valve it is usually the action of another fluid ical components such as gears, electric motors and even flow. Depending upon the intensity of this control flow, turbines and complete miniature combustion engines. The some part of the main flow is diverted into the vent outlet prospects are promising: an expectation of 40 billion US and thus prevented from reaching the output. The relative S microdevice business in year 2002 was recently pub- output flow Ly (defined in Fig. 1) is decreased.

lished [1]. Many microdevices work with fluid flows in Without the control action the valve must provide a micron-sized channels in applications ranging from ad- reasonably high value of 4y (in other words, a reasonable sorber heat pumps (cooling), fluid sample chemical analy- percentage of supplied fluid must reach the load) so that sis (DNA identification for disease diagnostics), to chemical microreactors used for compact fuel processing in fuel cell powered automobiles. Flow control by some sort of valves is essential. Microvalves with moving com- Control =ves om nozle ponents were successfully demonstrated, but much more effect a collector by inertia promising is application of the concepts of no-moving- LOAD part fluidics (e.g. [2]).

0 M 2. Jet-type diverter valves S The fundamental problem is that with small channel cross-sections and small flow velocities, the fundamental characterisation. parameter, the Reynolds number Re, tends to be very small, often by several orders of magni- Divwed tude. This results in completely different operating condi- f low tions than in large-scale fluidics, where flow control without moving components usually depends upon dy- OMY namic effects in accelerated fluids. Instead of simply reLy_ 0 plicating the more conventional ("mm-size") fluidic devices on the "micron size" (in itself no trivial task be- cause of the quite different manufacturing methods), An example of flow control by a no-moving-partfluidic valve. The microfluidics calls for the more challenging task of in- path between the supply inlet (left) and the output outlet (right) is in venting new operating principles. terrupted by a gap in which the fluid flow is exposed to a control ac- tion wtuch diverts some fluid and prevents it from reaching the collector.

18. Neigoo Ix Control 3. Microfluidics - and limits offluid dynamics E ' fact Jet The value of this Reynolds number Re = wb Pressure saw y determines the character of fluid flow in a fluidic valve.

S LOAD It is the basic parameter of the scaling law: if two fluidic devices, mutually similar but of diffffent size are to ex hibit the same properties, their Re must be equal. To get Diverted flow a one tenth scale microfluidic valve by scaling down a successfiil design of a large scale fluidic valve (Fig.4) re- V Presmve semor quires ten times higher velocity in the resultant small de- Pressure vice, using the same working fluid. Apart from the fact regulator that fluid viscosity is actually in many microfluidic ap Fig. 2 plications (viscous liquids in DNA tests, high viscosity To avoid output flow changes not caused by the control action, the hot gas in energetic conversions), the ten times higher fluidic diverter valve is usually used with a supply pressure regulator velocity would reach absurdly high supersonic or even hy keeping constant pressure difference between supply S and vent V personic values - for which there is usually not the re terminals. quired hundred (or more) times higher pressure source anyway. In fact, in many microfluidic applications there is a requirement of very low velocity (due e.g. to the re there is something the control action can decrease. To get quired residence time in a chemical microreactor or a this desirable effect in large-scale fluidic devices we em- composition analyser). As a result, aerodynamically simiploy dynamic, inertial effects. To overcome the viscous lar scaling down is usually impossible and we must accept loss occurring in the-gap between the nozzle and the col- much lower Reynolds numbers.

lector, the flow is accelerated in the nozzle so that it re aches the collector still with considerable kinetic energy to be converted into pressure rise. Selecting proper dimen- S sions of the gap is the key task of the valve designer. A wide gap leads to more effective jet deflection control - but leads to large loss ofjet momentum. The interplay be- W tween inertial and viscous forces acting on the jet is deter MEANING OF THE REYNOLDS NUMBER S W characteristic B Velocity: W time: t - W/ I Jet travel F _Fi 41 Surface: -1 2 3 Volume: Mass: - I Device miniaturisation as the way towards microfluidics. The basic V problem is the scaling lawrelations between the properties of a deceleration: -W/t - Wl microfluidic valve B scaled down from a large device A.

velocity graillient: - W This, of course, means increased viscous retardation shear stress: X W of the jet relative to its inertia and deteriorating valve effi Viscous force: 1 2-l; y wl v I ciency. We may use the magnitude of the relative output V 3 W2 1 2 W 2 flow m, in the no control flow regime as a measure of the Inertial force: mass x deceleration - - Inerlialforce W I v I V hydrodynamic efficiency of a fluidic diverter- type valve.

Ratio: Viscous force Re In good large-scale designs operating at the usual high 1 i Re - typically of the order 10. 10' - the efficiency may be Fig-3 as high as g, =0.80. The example in Fig.5 (evaluated for Derivation of Re by considering inertial and viscous force the valve from Fig. 12) shows fast decrease of achievable acting of a cube-shaped elementary volume of fluid. MY with decreasing Re in the laminar flow region B. As long as Re is not very low, some compensation is possible by decreasing the distance s, between the nozzle and the mined by the value of the Reynolds number Re. In Fig.3 it collector. This calls for much larger control flows, but this is expressed in terms of a characteristic length 1, char- is usually no problem with the small absolute flow mapiacteristic flow velocity, and fluid viscosity. In fluidics, it tudes - some microfluidic valves are known to work sucis common to use Reynolds number evaluated from nozzle cessfially with control flow much larger than the exit width and nozzle exit velocity - Fig.4. controlled flow, which would be an unacceptable paradox inlarge scale fluidics.

19. Aypojw Tlative= flow SdkDYNAMIC FLOWS low Re - w b t 0 zero action Reynolds number Classic 0.1 1 10 100.1000. CREEPING FLOW JjtyF W ICROF MICROFLUIDIC FLOW low velocity W Re large dimensions b - not necessarily low velocity W 14. small dimensions b e.g. Hele-Shaw Re?ion DB small Sh = '-t large Sh = WAt b b J F-F i- -6- J 1he subdynamic, low Re flows need not be -Re L.; j creepmg flows. A subdynamic flow in a rrucrode- Igi I onvice, far from being slow, may cover the small travelled distances extremely fast - the character 6.6, istic times of the order of milliseconds are no U 101 exception.

A t oreticall 4. Pressure driven devices Limit and the rekvant characterisadon number of fluid. Larninar re The remedy to the unacceptably small gy in the dynamics- subdynamic region is to rely not upon the vanishingly Dorrinated by Dominiated by small inertia but to force the fluid towards the load using Viscosity Intertia another effect. It is possible to use the electroosmotic ef fect - the micro flow injection analysis (liFIA) based on F-F71,2-7-5) The typical decrease of a diverter valve efficiency this effect is becoming quite successful in microanalytical L.;." with decreasing Reynolds number Re. systems [3]. The present authors remained in the safe realm of purely mechanical effects, forcing the fluid towards the load by pressure difference. What is needed, as shown in Fig.7, is to provide a second regulator (similar There are, however, microfluidic applications in to the supply regulator from Fig.2) keeping a constant which the operating Reynolds number is extremely low. pressure difference between the vent and output outlets. Fig.5 shows that there is a the theoretical limit Re = 1 Of course, additional regulator is no problem with the below which, in region C, the flow becomes fully domielectronics available an the "intelligent" chip. Note (cf nated by viscosity and inertial effects become negligible. Fig. 1) that this second regulator basically keeps a constant An interesting fact is that the transition into this subdy- pressure across the load - and again a single regulator namic regime (in this case at Re = 6.6) is so distinct and may be used for a large number of parallel loads. With a sharp. Another important fact is that in the subdynamic single valve the regulator would be called to react very regime C below Re - 1 the Reynolds number ceases to fast, but the usual parallel regulation of a large number of be the governing parameter. In this regime the data in Fig.5 show the efficiency to be nearly the same when Re ynolds number decreases by several decimal orders It should be noted that the low Re flow in a micro- Contra[ effect fluidic valve should not be imagined as a slow creeping flow. Since the distances s (proportional to the nozzle LOAD width b, Fig.4) to be travelled by the jet are very short, perhaps of the order of microns, the Strouhal number Sh (Fig.6) may be quite large and the characteristic time S (proportional to the jet travel time) may be very short. In Y Pressure general, flow switching in a microfluidic valve may sensor actually be very fast, especially if the circumstances (and available supply pressure difference) allow working with V high nozzle exit velocity w. Diverted There are not infrequent applications in micro- flow Pressure fluidics which demand operating Re near to and even regulator below the subdynamic regime. As a general rule, 1., due to inertial effects decreases below values of practical in- Fig. 7 terest beneath about Re = 100 and the inertial mode of Nficrofluidic diverter valve operated at very low Re with fluid fluidic valve operation is out of question. driven to the load by the constant pressure difference maintained by an external regulator loads means less de Characterisation numberTe rrands placed upon the ily 1ApYJ b2 regulator frequency Te 177 Te= 2h range, especially if o Re - 64.2 0M opening some of the & Re = 34.2 el valves takes place sim-,.Re=27.2 J4F,,Re 4.2.. e=3 AP [pa] pressure drop ultaneously with clos- 0.5 - Re = 13.8 OM [kgJsl mass flow rate ing other ones. The m Re= 7.0 b Iml nozzle exit width fluid flow through the 0.2 r Re= '4 load is varied due to the c Re= 1.4 I. R 0 v fm'lsl viscosity control flow (or another 0.1 a Re= 0.7 3 1 Tel h [m]... height of cavities control effect) blowing 0.05 U, 5.65 10 away (into the vent) the N,A1 available fluid from the 0.02 Te v 2 5 10 20 50 100 200 MEANING OF THE TESAR NUMBER Velocity: W Elementary volume of fluid _A. e.ple of the ci,dece of the reldve,.tpt flow rate for the mininture fluidic valve shown in Surface: - 1 2 Characteristic Fig.12. Properties in the subdynamic range (Re < 1) Volume: - 13 V dimension are uniquely characterised by Te and aH data points are found on the single universal straight line. At higher W Reynolds numbers this line is just an asymptote.

velocity gradient:

AP I acting shear stress: -C - V W pressure v replaces it in situations where it is pressure force instead difference Pressure force: - -AP 12 of the inertia acting on an elementary fluid volume. In the -AP.. negative, opposite to velocity pressure x area subdynamic range (Fig.5) where the Reynolds number Viscous force I 2'r - V W I difference ceases to be a meaningful characterisation of operating v conditions, Te is the really meaningful parameter com- Pressure force V [API 1 pletely characterising a pressure driven subdynamic Ratio: Viscous force W - Te microfluidic valve. This may be seen in the example in Fig. 10, where the relative output flow was evaluated both experimentally and by numerical flowfield computations

Derivation of Te by considering pressure and viscous force act- for a particular fluidic valve, shown in Fig. 11. The char ing of a cube-shaped elementary volume of fluid - not the acter of the dependence is, in fact, quite generic and prefect analogy to Fig.3. typical for a similar diver-ter devices. The lower is Re ynolds number, the nearer the values are to the asymptotic entrance to the collector. There is usually no problem if line 4Y = K Te this requires considerable input control power the abso- - such linear dependences are, of course, typical for lute magnitudes of powers and flows in microfluidics are laminar flows. In the subdynarnic range, where the prop negligibly small anyway. erties are determined solely by Te, this line becomes a With proper adjustment of the pressure drop, the universal relationship for the valve behaviour.

relative flow rate L, in the "fully open", no control flow state may be made as high as we may wish - even higher 4. An example of a pressure driven diverter than g, = 1.0 (- not achievable, of course, with fluidic ects) - Fig. 11. In- diverter based upon the dynamic eff valve itially as nothing more than just a convenient nondimen- The theoretical predictions based upon the character sional representation of the required pressure drop, useful isation by the introduced Te number were tested on a as an aid for its evaluation and adjustment, a dimension- microfluidic valve [4] developed by the present authors for less number Te, as defined in Fig. 8 was introduced. Sev- an application in high-temperature microreactor technol eral alternative definitions are possible, the one in Fig.8 ogy. It forms the key element of a sampag unit admitting is a practically convenient one, based upon the mean vel- sequentially a gas sample flow from one of 16 microreac ocity tors into a gas composition analyser. The sampling unit w = oMv/bh in the nozzle exit. However, (Fig. 11) is manufactured by etching in stainless steel. The a deeper meaning of the new parameter became soon required performance involves 10 % spillover flow in the obvious. We may note in Fig.9 that it is completely anal- OPEN state to eliminate possible sample contamination ogous to the Reynolds number (as presented in Fig.3) and by mixing %idth fluid from the vent, and a jet pumping applied pressure according to Fig.7 the output flow in.... . :=::.

..

the OPEN state ..........

.........

zero. Fig.13 shows that even with ........

some applied pres- Vory mall:1.111r,.

sure forcing the flowto a sample flow into the output ter- .......... minal, the sample flow still tends to FF-i -g.-I T k...........

............

prefer leaving =..

An example of a microfluidic sampling selector. Only one reac- through the vent - .......

.......

.....

tion product sample from 16 microreactors is admitted at a time In spite of the not........

.....

into the central outlet into a composition analyser. really very low Re here (note the.........

formation of a Re 128 tUy 0.336 jet, which does Te - 33 not take place in the subdynarnic 3.69 b FFig. 13 J regime). With in 2 = 1.47 b creased pressure Computed streamlinesfor the r, - 0.29 b difference AP, as valve in the OPEN state. Only 33.6 shown in Fig.14, % of supplied flow reaches the output :.V.97 b despite the ex- Y, despite an applied (but insuffi r cient pressure difference AP, (note tremely low Re, the nonzero Te) the tendency of 32_ the streamlines to after leav- spread ing the supply nozzle exit and to head into the vent, the pressure action is seen to force them to enter the collector. The diagram Fig.10 was essential in finding the proper Te value and the corresponding proper pressure.

FF-i -g. -1-2- 5. Conclusions

Lw.;_. The basic problem encountered in microfluidics is Geometry of the microfluidic switching valve developed for oper- the low Reynolds number which represents the ratio of ation in the pressure driven mode. Flow of reactants from S to Y dynamic effects to viscous dissipation. As a conseis discontinued and deflected into the vent V by a powerffil. flow quence, the inertial transport of fluid between nozzle and from the control terminal X collector - the basis of operation of jet- type valves as used in the large-scale fluidics - cannot be employed and new approaches are required. The diverter valves, as demonstrated on a practical example, which were developed by effect (backwards output flow L, < 0) in the CLOSED the authors can operate in the extremely low Re subdystate to remove the previous sample from the analyser ca- namic regime. All that is required is an additional presvities. Ile jet-pumping requirement resulted in the un- sure regulator. This creates a favourable pressure usual inclination of the control nozzle - Fig. 12. Another difference across the load which forces the fluid to move unusual feature is the small depth of the cavities, equal to into the collector of the valve and finther downstream into only 0.44 b. The smooth shape of the flow transfer char- the device in which the flow is to be controlled. A new acteristic Fig.15 - assures that the conditions in the characterisation number Te (Fig.8) was introduced reCLOSED state can be always met, even though at a price placing the Reynolds number, which ceases to be of imof using control flow more than 20-times the controlled portance in this regime. The required pressure drop for flow. Of key importance for proper operation is therefore proper operation in the OPEN state can be evaluated the adjustment of the OPEN state conditions. The high from a relationship (Fig.10) between the relative output temperature and very small sample flow rate lead to Re- flow rate and Te. An important feature of this relationynolds number around Re = 30, very near to the transi- ship is the asymptotic straight line, which represents the tion into the subdynamic range, so that without the subdynamic behaviour limit.

22 FF -ig-. -1 _4 Reference5 LMMMMMMA [11 Ehrfeld W. (ed.): "Microreaction technology: In The improvement dustrial prospects", ISBN 3-540-66964-7, Springer, achieved by propefly Berlin 2000.

adjusted pressure dif- [2] Tesar V.: "Valvole fluidiche senza parti mobili", ference AP, jet is Oleodinamica -pneumatica, Milano, Italy, forced into the col lector, despite ex- Anno 39, ISSN 1 122-5017, p. 216 223, 1998.

tremely low Re. A [3] Fletcher P.D.I, Haswell S.J., and Paunov V.N.:

small spill-over flow "Theoretical considerations of chemical reactions on (here 13 %) in this micro-reactors operating inter electroosmotic and electro OPEN state is wel- phoretic control", Ana#sl 124, p. 1273-1282, 1999 come to guarantee.no [4] Tesar V., Tippetts J.R., Allen R.W.K: "Fluidic Spillover cross contamination Valve", British Patent Application between the samples.

Re =3.50 4a y = 0.87 Te 154 Y I Re 27.71 1.0 Te159-FFg-.15A 0.8 Flow tsfer characteristic of the valve from ran OPEN Fig-12 in relative co-ordinates (control flow 0.6 -state: horizontal co-ordinate, output flovr. vertical = 0.9 co-ordinate - both related to the supply 0.4 Y flow). Note the extremely powerful (>20-times the sample flow) control flow required for clos 0.2 x ing the valve and generating the jet pumping 0 CLOSED state effect.

-0.2 LAY< 0 - 0. 4F- - i 1 1 I 0 5 10 15 20

Claims

1. A fluidic multiplexer for supplying to a common outlet channel (112) one fluid selected from at least a first fluid channel and a second fluid channel each for carrying respective first (102) and second (104) fluid flows; the multiplexer (100) comprising a first fluidic valve (114) to prevent flow - of the first fluid from the first fluid channel (102) to the common outlet channel (112) in response to flow of a first control fluid from a first control inlet (154) and a second fluidic valve (116) to prevent the flow of the second fluid (104) from the second fluid channel to the common outlet channel (112) in response to flow of a second control fluid from a second control inlet (156).

2. A fluidic multiplexer as claimed in claim 1 in which the first fluidic valve (114) comprises a first vent (124) arranged to allow flow of the first fluid (102) through the first vent (124) in response to the first control fluid preventing flow of the first fluid (102) to the common outlet channel (112)

3. A fluidic multiplexer as claimed in either of claims 1 and 2 in which the second fluidic valve (116) comprises a second vent (126) arranged to allow flow of the second fluid (104) through the second vent (126) in response to the second control fluid preventing the flow of the second fluid (104) to the common outlet channel (112).

2A4-

4. A fluidic multiplexer arranged to control selectably the flow of a first fluid (102) and a second fluid (104) into a common outlet channel (112) using first (114) and second (116) fluidic valves, the first fluidic valve (114) having a first inlet to supply the first fluid (102) to a first outlet channel and a first control inlet (154) to prevent flow of the first fluid (102) to the first outlet channel; the second fluidic valve (116) having a second inlet to supply the second fluid (104) to a second outlet channel and a second control inlet to prevent flow of the second fluid (104) to the second outlet channel; the first and second outlet channels being arranged to feed the common outlet channel (112) is

5. A fluidic multiplexer as claimed in any preceding claim further comprising a pressure regulator for establishing the pressure in the outlet channel to be a predeterminable pressure.

6. A fluidic multiplexer as claimed in any preceding claim further comprising a pressure regulator for changing the pressure of at least one of either of the first (102) and second (104) fluids from respective first and second pressures to a selectable pressure prior to feeding the common outlet channel (112).

7. A fluidic multiplexer as claimed in any preceding claim in which the direction of flow of the control fluid (104) is arranged to oppose the direction of flow of the supply fluid (102).

Z's

8. A fluidic multiplexer as claimed in any preceding claim in which the dimensions of the channels and inlets are arranged such that the fluids flows carried therein have associated Reynolds numbers 5 that are less than 100.

9. A fluidic multiplexer as claimed in claim 8 in which the associated Reynolds numbers are less than 40.

10. A fluidic multiplexer as claimed in any preceding claim in which the first fluidic valve (114) operates in anti-phase with the second fluidic valve (116).

is

11. A fluidic multiplexer substantially as described herein with reference to and/or as illustrated in any of the accompanying drawings.

12. A fluidic multiplexing method substantially as described herein with reference to and/or as illustrated in any of the accompanying drawings.