Abstract<\/h2>\n\nThere is an unmet demand for microfluidics in biomedicine. This paper describes contactless fabrication of microfluidic circuits on standard Petri dishes using just a dispensing needle, syringe pump, three-way traverse, cell-culture media, and an immiscible fluorocarbon (FC40). A submerged microjet of FC40 is projected through FC40 and media onto the bottom of a dish, where it washes media away to leave liquid fluorocarbon walls pinned to the substrate by interfacial forces. Such fluid walls can be built into almost any imaginable 2D circuit in minutes, which is exploited to clone cells in a way that beats the Poisson limit, subculture adherent cells, and feed arrays of cells continuously for a week. This general method should have wide application in biomedicine.<\/p>\n<\/div>\n<\/section>\n<\/div>\n
<\/div>\n\n\n1 Introduction<\/h2>\n
As sensitivities of methods for detecting biomolecules improve, demand for handling ever-smaller volumes increases\u2014and this drives development of microfluidic approaches.[<\/sup>1<\/a>–3<\/a><\/sup>]<\/sup>\u00a0However, few of these are found in biomedical workflows, with the exception of those involving microplates. Why? Reasons given include: devices are expensive and take days\/weeks to make, they are complicated to operate, their contents are inaccessible, and they are not made with cell-friendly materials.[<\/sup>1<\/a><\/sup>]<\/sup><\/p>\nIn the everyday world, gravity is such a dominant force that most objects are made with solids, and one cannot contemplate building them out of liquids, which just collapse into puddles. Consequently, liquids are always contained by solid walls, otherwise they drain away. In the microworld, gravity becomes irrelevant, and interfacial forces dominate (think of water striders skimming over ponds, and dewdrops sticking to blades of grass).\u00a0Consequently, an interface between two fluids can act as a robust wall separating the two. This paradigm enabled the emergence of open microfluidics,[<\/sup>4<\/a><\/sup>]<\/sup>\u00a0where solid walls are replaced by air:water[<\/sup>5<\/a><\/sup>]<\/sup>\u00a0or oil:water interfaces.[<\/sup>6<\/a>–8<\/a><\/sup>]<\/sup><\/p>\nOpen microfluidic devices are generally easier to integrate into biomedical workflows as they offer better optical and physical access to samples, reduced adhesion of reagents to solid surfaces and are resistance to blocking by air bubbles. However, despite the ever-increasing efforts to simplify manufacturing workflows for such devices, most of them still require etching of the substrate,[<\/sup>9<\/a><\/sup>]<\/sup>\u00a0surface treatment[<\/sup>10<\/a><\/sup>]<\/sup>\u00a0or contact,[<\/sup>7<\/a><\/sup>]<\/sup>\u00a0or some combination of these[<\/sup>8<\/a><\/sup>]<\/sup>\u00a0to confine fluidic structures. Such complex manufacturing processes deter many biologists who favor fast flexible prototyping without having to compromise biocompatibility.[<\/sup>11<\/a><\/sup>]<\/sup>\u00a0Recently, microfluidic arrangements were created simply by dragging a Teflon rod\/stylus resting on the surface of a dish to reshape cell-culture media and an immiscible overlay into the wanted pattern.[<\/sup>7<\/a><\/sup>]<\/sup>\u00a0However, motion of the stylus relative to its holder introduced play in the\u00a0x<\/i>\u2013y<\/i>\u00a0plane, reducing accuracy and precision. Moreover, proteins in media aggregate on the stylus to reduce reproducibility and increase risks of cross contamination. Additionally, the stylus needs frequent change due to wear. All these drawbacks stem from the effects of contact.<\/p>\nHere, we describe a contactless method to fabricate microfluidic devices on demand, where the only \u201cbuilding\u201d materials used are those in the biocompatible trio\u2014cell media, the immiscible fluorocarbon FC40, and a polystyrene Petri dish. FC40 is \u201cjetted\u201d from a dispensing needle through bulk FC40 and media on to the untreated bottom of a dish. Complex microfluidic structures with features <50\u00a0\u00b5m in size can be produced reproducibly with high accuracy in minutes. The aqueous phase is confined by fluid walls\u2014media:FC40 interfaces\u2014which are robust yet easily pierced (so liquids can be added\/removed through them at any preselected point) whilst being transparent. The physics underlying flow during such jetting is complex;[<\/sup>12<\/a>–15<\/a><\/sup>]<\/sup>\u00a0therefore, we establish appropriate conditions. We then exploit jetting to \u201cbeat\u201d the Poisson limit to clone single mammalian cells, subculture them (again using a contactless method), and perfuse them steadily with fresh media for a week.<\/p>\n<\/section>\n\n2 Results<\/h2>\n\n2.1 Approach<\/h3>\n
Figure<\/b>\u00a01a<\/a>\u00a0illustrates the approach. The bottom of a standard tissue-culture dish is covered with a film of cell-growth media, and an FC40 overlay added to prevent evaporation. A dispensing needle filled with FC40, connected to a syringe pump, and held by a three-way traverse is now lowered below the surface of the fluorocarbon; starting the pump jets FC40 on to the dish to push media aside. As FC40 has a low equilibrium contact angle (CA) on polystyrene (<10\u00b0), it wets it better than media (equilibrium CA \u224850\u00b0),[<\/sup>6<\/a><\/sup>]<\/sup>\u00a0so it adheres to the bottom. Moving the microjet sideways then creates a line of FC40 on the dish, and drawing more lines creates a grid with 256 chambers in <2\u00a0min (Figure\u00a01b<\/a>; Movie S1, Supporting Information). Each chamber is isolated from others by liquid walls of FC40 pinned to polystyrene. Interfacial forces dictate chamber geometry\u2014a spherical cap sitting on a square footprint (height \u224875\u00a0\u00b5m; volume \u2248100 nL). Up to \u2248900 nL more media can be pipetted into chambers as fluid walls morph above unchanging footprints. Chambers are then used like wells in microplates: liquids are added\/removed to\/from them by pipetting through FC40 instead of air. The maximum and minimum volumes that can be held in chambers without altering footprints are determined by advancing and receding contact angles; addition of too much media inevitably merges adjacent chambers. Even so, chambers accept a manyfold wider range of volume than equally spaced wells in a microplate, whilst containing \u22481000th the volume.[<\/sup>7<\/a><\/sup>]<\/sup>\u00a0Consequently, if chambers contain cells, the volume ratio of intra- to extracellular fluid more closely resembles that in vivo. Importantly, this method is contactless: the nozzle touches neither dish nor media. Moreover, one pipet tip can add\/extract reagents to\/from many chambers without detectable cross-contamination (shown\u2014for example\u2014by seeding bacteria in every other chamber, adding media to all through one tip, and finding that bacteria grow only in inoculated chambers as others remain sterile).[<\/sup>7<\/a><\/sup>]<\/sup>\u00a0In other words, a tip is washed effectively by passage through FC40 between chambers, and\u2014when using cells\u2014we make doubly sure by additionally washing in 70% ethanol.<\/p>\n\n![\"image\"](\"https:\/\/onlinelibrary.wiley.com\/cms\/asset\/06ada756-69a1-4911-96e0-ddb78b7ee08d\/advs2034-fig-0001-m.png\" "\\"image\\"")
<\/picture><\/a><\/figure>\n<\/section>\n
There is an unmet demand for microfluidics in biomedicine. This paper describes contactless fabrication of microfluidic circuits on standard Petri dishes using just a dispensing needle, syringe pump, three-way traverse, cell-culture media, and an immiscible fluorocarbon (FC40). A submerged microjet of FC40 is projected through FC40 and media onto the bottom of a dish, where it washes media away to leave liquid fluorocarbon walls pinned to the substrate by interfacial forces. Such fluid walls can be built into almost any imaginable 2D circuit in minutes, which is exploited to clone cells in a way that beats the Poisson limit, subculture adherent cells, and feed arrays of cells continuously for a week. This general method should have wide application in biomedicine.<\/p>\n<\/div>\n<\/section>\n<\/div>\n
1 Introduction<\/h2>\n
As sensitivities of methods for detecting biomolecules improve, demand for handling ever-smaller volumes increases\u2014and this drives development of microfluidic approaches.[<\/sup>1<\/a>–3<\/a><\/sup>]<\/sup>\u00a0However, few of these are found in biomedical workflows, with the exception of those involving microplates. Why? Reasons given include: devices are expensive and take days\/weeks to make, they are complicated to operate, their contents are inaccessible, and they are not made with cell-friendly materials.[<\/sup>1<\/a><\/sup>]<\/sup><\/p>\n In the everyday world, gravity is such a dominant force that most objects are made with solids, and one cannot contemplate building them out of liquids, which just collapse into puddles. Consequently, liquids are always contained by solid walls, otherwise they drain away. In the microworld, gravity becomes irrelevant, and interfacial forces dominate (think of water striders skimming over ponds, and dewdrops sticking to blades of grass).\u00a0Consequently, an interface between two fluids can act as a robust wall separating the two. This paradigm enabled the emergence of open microfluidics,[<\/sup>4<\/a><\/sup>]<\/sup>\u00a0where solid walls are replaced by air:water[<\/sup>5<\/a><\/sup>]<\/sup>\u00a0or oil:water interfaces.[<\/sup>6<\/a>–8<\/a><\/sup>]<\/sup><\/p>\n Open microfluidic devices are generally easier to integrate into biomedical workflows as they offer better optical and physical access to samples, reduced adhesion of reagents to solid surfaces and are resistance to blocking by air bubbles. However, despite the ever-increasing efforts to simplify manufacturing workflows for such devices, most of them still require etching of the substrate,[<\/sup>9<\/a><\/sup>]<\/sup>\u00a0surface treatment[<\/sup>10<\/a><\/sup>]<\/sup>\u00a0or contact,[<\/sup>7<\/a><\/sup>]<\/sup>\u00a0or some combination of these[<\/sup>8<\/a><\/sup>]<\/sup>\u00a0to confine fluidic structures. Such complex manufacturing processes deter many biologists who favor fast flexible prototyping without having to compromise biocompatibility.[<\/sup>11<\/a><\/sup>]<\/sup>\u00a0Recently, microfluidic arrangements were created simply by dragging a Teflon rod\/stylus resting on the surface of a dish to reshape cell-culture media and an immiscible overlay into the wanted pattern.[<\/sup>7<\/a><\/sup>]<\/sup>\u00a0However, motion of the stylus relative to its holder introduced play in the\u00a0x<\/i>\u2013y<\/i>\u00a0plane, reducing accuracy and precision. Moreover, proteins in media aggregate on the stylus to reduce reproducibility and increase risks of cross contamination. Additionally, the stylus needs frequent change due to wear. All these drawbacks stem from the effects of contact.<\/p>\n Here, we describe a contactless method to fabricate microfluidic devices on demand, where the only \u201cbuilding\u201d materials used are those in the biocompatible trio\u2014cell media, the immiscible fluorocarbon FC40, and a polystyrene Petri dish. FC40 is \u201cjetted\u201d from a dispensing needle through bulk FC40 and media on to the untreated bottom of a dish. Complex microfluidic structures with features <50\u00a0\u00b5m in size can be produced reproducibly with high accuracy in minutes. The aqueous phase is confined by fluid walls\u2014media:FC40 interfaces\u2014which are robust yet easily pierced (so liquids can be added\/removed through them at any preselected point) whilst being transparent. The physics underlying flow during such jetting is complex;[<\/sup>12<\/a>–15<\/a><\/sup>]<\/sup>\u00a0therefore, we establish appropriate conditions. We then exploit jetting to \u201cbeat\u201d the Poisson limit to clone single mammalian cells, subculture them (again using a contactless method), and perfuse them steadily with fresh media for a week.<\/p>\n<\/section>\n Figure<\/b>\u00a01a<\/a>\u00a0illustrates the approach. The bottom of a standard tissue-culture dish is covered with a film of cell-growth media, and an FC40 overlay added to prevent evaporation. A dispensing needle filled with FC40, connected to a syringe pump, and held by a three-way traverse is now lowered below the surface of the fluorocarbon; starting the pump jets FC40 on to the dish to push media aside. As FC40 has a low equilibrium contact angle (CA) on polystyrene (<10\u00b0), it wets it better than media (equilibrium CA \u224850\u00b0),[<\/sup>6<\/a><\/sup>]<\/sup>\u00a0so it adheres to the bottom. Moving the microjet sideways then creates a line of FC40 on the dish, and drawing more lines creates a grid with 256 chambers in <2\u00a0min (Figure\u00a01b<\/a>; Movie S1, Supporting Information). Each chamber is isolated from others by liquid walls of FC40 pinned to polystyrene. Interfacial forces dictate chamber geometry\u2014a spherical cap sitting on a square footprint (height \u224875\u00a0\u00b5m; volume \u2248100 nL). Up to \u2248900 nL more media can be pipetted into chambers as fluid walls morph above unchanging footprints. Chambers are then used like wells in microplates: liquids are added\/removed to\/from them by pipetting through FC40 instead of air. The maximum and minimum volumes that can be held in chambers without altering footprints are determined by advancing and receding contact angles; addition of too much media inevitably merges adjacent chambers. Even so, chambers accept a manyfold wider range of volume than equally spaced wells in a microplate, whilst containing \u22481000th the volume.[<\/sup>7<\/a><\/sup>]<\/sup>\u00a0Consequently, if chambers contain cells, the volume ratio of intra- to extracellular fluid more closely resembles that in vivo. Importantly, this method is contactless: the nozzle touches neither dish nor media. Moreover, one pipet tip can add\/extract reagents to\/from many chambers without detectable cross-contamination (shown\u2014for example\u2014by seeding bacteria in every other chamber, adding media to all through one tip, and finding that bacteria grow only in inoculated chambers as others remain sterile).[<\/sup>7<\/a><\/sup>]<\/sup>\u00a0In other words, a tip is washed effectively by passage through FC40 between chambers, and\u2014when using cells\u2014we make doubly sure by additionally washing in 70% ethanol.<\/p>\n2 Results<\/h2>\n
2.1 Approach<\/h3>\n
![\"image\"](\"\/cms\/asset\/1e5e19f2-481c-4a1d-95af-b962ed91cdbd\/advs2034-fig-0001-m.jpg\" "\\"image\\"")
<\/picture><\/a><\/figure>\n<\/section>\n