Biochip with Integrated Vertical Emitting Light Source - Book Report/Review Example

Biochip with Integrated Vertical Emitting Light Source Abstract: This report discusses the results from integration of microfluidic device into semiconductor laser material. Laser emitted from the active region can increase assay temperature beyond 37O C which can be fatal for any organisms being studied. This temperature must never be allowed to reach such levels and the heat must be dissipated to avoid such a situation. COMSOL Multiphysics simulation environment is used to perform the tasks. The properties investigated are fluid flow in the channel, Vpulse results in electric conductivity in the active region, and dissipation of heat resulting from laser light produced in the course of current passing through semi-conductor. Heat transfer through convection and conduction is also studied. This transfer is investigated distributing from the active region to rest of the structure. Finally, we present results using the principles of fluid dynamics and heat transfer using electromagnetic modules of COMSOL Multiphysics. These modules deal with movement of fluid in biochip microfluidic channel, the convection and conduction in the active region over time, and Vpulse for electric conductivity in active region produced as a result of passing current through it. We also illustrate heat transfer through the structure and show pattern of convection and conduction in the active region. Introduction The fabrication of microfluidic channels is critical in evaluating the contents of fluid inside the channel. In recent years the implementation, integration of fluid, analytical separation and detection techniques on individual microfabricated devices for complete on-chip analysis have been one of the main objectives of microfluidics research. Some of the integrated features applied in microfluidic devices include “measuring the optical transmission, absorbance or fluorescence of the channel contents” (Course handout). Laser technology emitting very narrow light beams is an appropriate technology to be used in microfabrication of microfluidic devices. The beam from these light sources becomes very small using LEDs and VCSELs by virtue of being semi-conductor based. These reduce the averaging effect and also avoid “cross-talk between adjacent channels” (Irawan et al, 2007). LEDs can also be composed of organic compounds and such LEDs are called organic light emitting diodes (OLEDs). The emitting material can come from small organic molecules (Tang & VanSlyke, 1987) or a polymer (Burroughes et al, 1990). Polymer materials can be flexible (Gustafsson, 1992). Known as flexible organic LEDs (FOLED), they have opened up possibilities of many new applications such as backlit displays (Friend et al, 1999). Usually very small in area, measuring less than 1 mm2, the built-in optical components can shape its radiation patterns (Moreno & Sun, 2008). LEDs have many advantages such as “broad range of available emission wavelengths, high efficiency in power, stable output, long lifeline, durability, reliability, low cost and low self-heating” (Zukauskas et al, 2004: p. 128). One of the fabrication methods which “integrates microfluidic channels on the semiconductor laser material itself” was presented by Cran-McGreehin et al. This allows the lasers beams to “couple directly into the flow channel” (2006: p. 7729). Microfabrication methods can be used to allow semiconductor light sources to be repeatedly mounted in close proximity to the edge of the channel. The mounting of the small emitter devices is as follows (Course handout). Step 1: The emitter device is placed aperture downwards on a freshly cured planar PDMS surface along with any additional electrical contact pads. The PDMS forms a temporary seal against the surface of the device and pads. The second device contact is bonded to the back surface of the device. Step 2: The device and pads are encapsulated in an optical epoxy which forms a flat planar surface against the cured PDMS. Step 3: After curing the epoxy, the PDMS is removed from the inverted structure to reveal a flat surface containing the light emitting aperture and the additional contact. The epoxy forms a strong bond around the edges of the device and contact by does not encroach onto the surface of either component because of the PDMS seal in place during curing. Step 4: Next, the epoxy surface is coated with an evaporated metal which is subsequently pattered to expose the light emitting aperture of the device. The metal layer forms an electrical connection between the upper surface of the device and the embedded electrical contact so allowing both upper and lower contacts of the device to be accesses from the rear of the device. Step 5: Finally, a precise layer of epoxy, typically 2m thick, is spin coated over the device to provide electrical insulation and mechanical robustness. Microfluidic channels are subsequently fabricated on the upper surface of this epoxy layer. In doing so, the lower surface of the channel is within 2m of the light emitting aperture so reducing divergence effects and optimally delivering light to the channel. However, small semiconductor devices have the potential to heat up quickly when light is generated from current and can potentially “stress” or “damage” (Urban et al, 2005: p. 2239) the biological material in a microfluidic device. It is, therefore, necessary that this heat must be dissipated and the cells and proteins be kept at a temperature maintained below 37oC. In this paper, we will use COMSOL Multiphysics to calculate the change in temperature at the bottom surface (top of the thin epoxy layer) of the microfluidic channel over time. Also, to reduce the temperature rise, the device will be driven from a pulsed voltage source where the duty cycle and frequency are chosen to allow heat generated in the device ‘on-time’ to be dissipated before the next pulse. We will use the model in Figure (1) below to find an appropriate pulse width for a 10% on time duty cycle to limit the temperature rise at the lower surface of the channel above the emitter aperture to 0.5oC. Figure (1). View of the complete model geometry The geometry used in this assignment represents the semiconductor (R2), encapsulating epoxy (R6), thin epoxy layer (R1), fluidic channel (R4), channel upper wall (R5) and the active region of the semiconductor (R3). This geometry will be modelled in 2D with the overall length of the model being 5mm and the total height being 1.5mm. The top of the channel is 100m above the surface of the semiconductor with a 2m epoxy layer separating the channel from the device. The device is 2mm long and 300m high. The active region is located in the centre of the upper surface of the device and is 40m wide and 2m high and positioned to be 1.5mm from the start of the channel. It is assumed that all the current flows from the upper surface of the active region to the lower surface of this region so limiting the heat generation to the active region. We will combine fluid flow, thermal and electrical physics regimes to calculate the effect of Ohmic heating on a sample undergoing laminar flow in the channel. It is assumed that the channel contains water and that fluid flow only occurs within the channel subdomain. Current can be assumed to only flow through the active region which has a conductivity of 17Sm-1. Heat will be dissipated through all parts of the model. Figure (2) below shows expanded view of the active region of the model geometry. Figure 2. Expanded view of the active region of the model geometry. Theoretical This section presents the theories and physical equations that will be used when we deal with Fluid flow in the biochip. The flow of fluids is described using the Navier-Stokes equation which can be used “for most phenomena observed in fluid mechanics” (Lagerstrom, 1996: p. 3). However, fluid flow can be turbulent or laminar. To characterise the flow between turbulent or laminar, the Reynolds number (Re) is used (Lagerstrom, 1996: p. 116). Re = If Re < 1,400 the flow is laminar and if Re < 2,100 the flow is turbulent. Generally, fluid is incompressible under the condition meaning that the density of fluid is constant. However, Navier-Stokes equation in this case is: 2 u + (u.)u + Where is the rate of change in momentum or net acceleration, is convection force, n2 u is viscous force, is pressure force and F is the external body force. Figure 3: Velocity of laminar flow between parallel plates. The velocity of laminar flow between parallel plates varies from zero at the walls to a maximum along the centre line of the vessel as shown in Figure 3. The conductivity of water depends on concentration of dissolved salts. Electric conductivity of water sample is used to calculate how salt-free, ion-free or impurity-free the sample is. Pure water has low conductivity. We can get the electric conductivity from the following equation: Where is the current density and E is the electric field. Due to thermistor action, electric conductivity in semiconductor rises with an increase in heat (Seeger, 2004: p. 2). Convection is defined as the exchange of heat between a surface and moving fluid, both being at different temperatures. Convective heat transfer can be categorized into two groups, free convection and forced convection. Any fluid flow which depends only on density gradients caused by temperature differences is termed “natural” or “free”. Whereas, any motion which is caused by external factors, is termed as “forced” (Muneer et al, 2003: p. 162). For example, if a fluid is heated on a surface such as on a hot plate, this would be free or natural convection. However, if hot fluid is passed though a pipe, and heat transfer takes place, this would be described as forced convection. Sometimes, both types of convection may occur together, and in this case, are termed as “mixed convection” (Muneer et al, 2003: p. 162). This is in contrast to conductive heat transfer which depends solely on vibrational energy transferred between molecules. It differs also from radiation energy which transfers heat through electromagnetic waves (Coulson, 1999: p. 381). In the COMSOL Users Conference (2006), Curet et al noted the governing equations for heat transfer basing them on thermophysical properties. The method for effective heat capacity was considered appropriate to analyse thawing in materials that melt over a range of temperature. Using the generalised heat equation, where Q is the internal heat generation source and measures the power dissipated by dielectric losses. Furthermore, in this modelling, Heaviside step function is used. The Heaviside step function returns a value of zero for -∞ ≤ x < 0 and a value of 1 for all other positive values where x is the input value (Chen et al, 2003: p. 49). The function is described as: The COMSOL formula used for Heaviside function is: flc1hs (ton1, tris) - flc1hs (ton2, tris) Methodology In order to calculate the effect of Ohmic heating on a sample undergoing laminar flow in the channel, the model geometry was dealt with in the three following stages. 1. Microfluidic flow: The fluid flow was simulated using incompressible form (Transient analysis) of Navier-Stokes equation in 2D from model Navigator in COMSOL Multiphysics. We used options in COMSOL to create table with following values in constants: rho_water =1000 which is density of water eta_water = 0.001 kg/ (m*s) which is viscosity of water Umax =0.01 m/s which is peak fluid inlet velocity Then, Subdomain Settings was selected from the physics menu, only the channel was active in this section, we defined the physical properties of the fluid, in this case viscosity and density, according to the values above. The Navier-Stokes equation for the subdomain was: MISSING Where ρ is density, n is viscosity. After that, boundary setting was selected from physics menu and the inlet was defined with laminar flow profile using ‘s’ variable where the equation for the boundary condition velocity inlet was: U= -u0n Where U0 was set to (Umax*4*s*(1-s). We did that because the flow was laminar in the parallel plate channel. The boundary condition for outlet was defined as Normal stress of zero as the fluid will flow out from it and the pressure would be zero in the fluid flow: P0=0 This was to set the outlet pressure to 0 Pa gauge. The equation in this case for the outlet boundary condition was: n (0 MISSING Then, we kept the top and bottom surface channel boundary as wall and set the boundary condition “no slip” to simulate friction. The solver parameter was selected, and the analysis type as stationary. The solver manger was set to use the stored solution for unsolved variables. After that, the solve button was selected. 2. Electric conductivity: Staying with the same model geometry, we simulated conductivity using model navigator, selected COMSOL Multiphysics. We chose conductive media from the electromagnetic option. We used the options in COMSOL to add the following value to constant in the previous table: Vmax = 2.3 V which is the voltage applied to active region Then the boundary expression was selected from options to put the value: Vpulse = Vmax * (flc1hs (t-1e-5, 1e-8) – flc1hs (t – 1.1e-5, 1e-8)) which is the definition of the pulse applied to the active region. Then subdomain setting was selected from physics menu. In this case the active region was active because it is the area for ‘light source’ laser. The conductivity was set to (σ = 17 s/m), where the equation for active region subdomain was: Then, we moved to boundary setting. The top of active region was set to electric potential with a value of Vpulse. The equation in this case was V=V0 V0 = Vpulse After that, the bottom was set to Ground. The equation in this case was V = 0. The left side and the right side were set to electric insulation, where the equation was: n.J = 0 Then, the option “Time Dependent” was selected from the solver parameter and the solve time was defined to be range (. 95e-5:1e-7:5e-5). In the Solver Manager the incompressible Navier-Stokes physics was disabled and only the conductive media physics was solved. 3. Heat transfer: We continued with the same model geometry as in the preceding two stages. The objectives in this stage were to: 1. Calculate the change in temperature at the bottom surface (top of the thin epoxy layer) of the microfluidic channel over time. 2. To reduce temperature rise. 3. Calculate the effect of Ohmic heating on a sample undergoing laminar flow in the channel. We simulated convection and conductive (Transient analysis) using model navigator, selected COMSOL Multiphysics and then heat transfer. Using the options again, we added the value of ambient temperature to the constants: Ta = 289 k Then the Subdomain setting was selected from physics menu. The material library in subdomain was used to select the materials as given below: Subdomain Material Thin Epoxy Layer MEMS_Polymer_polyimide Device MEMS_semiconductor_GaAs Active Region MEMS_semiconductor_Gaas Heat Source: Q_dc Channel Thermal conductivity 0.6 Channel Density : rho_water Heat capacity: 4200 Velocity: u, v Channel Upper Wall MEMS_Polymers_polyimide Encapsulating Epoxy MEMS_Polymers_polyimide The Velocity was set to u, v in order to combine the convection and conduction to the incompressible Navier-Stokes fluid flow to apply convection in the channel and then conduction throughout the rest of the device. The equation for subdomain was as following: where T is the temperature. The initial temperature for all subdomains was then set to a value of Ta. Subsequently, moving to the boundary setting, all external boundaries were selected to have a temperature of Ta. Only the channel outlet was set to a convective Flux, where the boundary equation here was: n. (-k Then, we selected transient analysis in the solver parameter for convection and conductive. Also, in solver manager, the incompressible Navier-Stokes physics was disabled and only the convection, conduction and the conductive Media physics was solved. This was done in order to occurring resistive heating and the heat generated passes through the rest of the device. Finally, we added the following values to the post processing parameters as shown in the table: Type Parameter Surface T- 298 Arrow X component: u Y component: v X points: 100 Y points: 350 Arrow: cone, scale = 0.125, colour = white, length = proportional Time Show at end of pulse and times after end of pulse Result and discussion The results are presented below in the same sequence of stages as we followed in the Methodology section, that is, Microfluidic Flow, Electric Conductivity and Heat Transfer. 1. Microfluidic flow: Figure (4): shows the result of simulation fluid flow in the channel The result from the simulation of fluid flow in channel is shown in figure (4). The model shows that the water flowed from the inlet towards the outlet. It is also clear that the velocity varies in the channel from zero at the wall to its maximum in the centre of the channel confirming what was mentioned earlier in the Theoretical section above. The maximum velocity for laminar flow was 1e-3 in the centre of the channel and the minimum velocity was 0 in the walls of the channel. 2. Electric conductivity (Conductive Media DC): Figure (5): shows the output for Vpulse in active region The output from Vpulse in the active region is as shown in figure (5). Pulse source (Vpulse) was used to test the transient response of the active region. This provided with out time-variant input source. It is clear that when the pulse is not "on", the value is zero. This can be zero or negative. We can see that when the pulse is fully turned “on”, the value becomes 0.92e-4, where 1*10-5 is the rise time in seconds and same 1.01*10-5 is the fall time in seconds of the pulse. The period between 1*10-5 and 1.01*10-5 is the pulse width. This is the duration in seconds that the pulse stays fully on. Vpulse arises in the beginning because of the need to perform DC testing of laser diodes prior to mounting on a thermal management device. With short duration pulses, the laser diode's average power dissipation has minimal thermal effects. 3. Heat transfer (Convection and Conduction): Figure (6): shows the simulation of convection and conduction through the structure Figure 6 shows the temperature conducting through the structure in the active region (dark red area). The range of temperature is shown between the maximum temperature 298.094K and the minimum temperature 298 K (dark blue) furthest from the active region. The following results are clear from this simulation. 1. In the fluid, the heat is transferred from one place to another by convection. 2. Heat energy is transferred from the hot area (active region) to the cooler area (the rest of structure) by convection. 3. The hot area (active region) is less dense than the cold areas (the rest of structure). This is because the particles in hot area move faster than the particles in cooler areas. 4. Finally, the heat energy can move through structure (metal) by conduction, transferred from the hot end of the structure to the cold end. In order to calculate the change in temperature at the bottom surface (top of the thin epoxy layer) of the microfluidic channel over time, the temperature probe points was plotted. The location of temperature point is as shown in figure (7) of X 0.00152, Y 3.02e-4 Figure (7): shows the location of temperature point. Figure 8. Graph 1 Figure 8. Graph 2. The graph 1 in Figure (8) shows that the temperature rose at 1 microsecond (the x point of the Heaviside function) up to 298.012 K .The pulse stays on for 0.01 microsecond and then it takes 1ns to rise and fall. The graph 2 in Figure (8) shows that once the maximum temperature is reached, the device gradually cools. This results in Convection and conduction through structure. Conclusion: The experiment in this paper simulated the fluid flow, electric conductivity, and heat transfer in Biochip with Integrated Vertical Emitting Light Source. Temperature change at the bottom surface (top of the thin epoxy layer) of the microfluidic channel over time was calculated. The objective was to reduce temperature rise and this was successfully shown. The result showed that the fluid flow in the channel was laminar, and the maximum velocity was present in the centre of channel reducing to zero at walls. It was also demonstrated that the device could be driven from a suitable pulse source, Vpulse, where the duty cycle and frequency can be chosen to allow heat to be dissipated before the next pulse arrives. During the pulse width, the laser diode's average power dissipation has minimal thermal effects. The results presented the associated change in temperature over time where the change was explained by convection and conduction through the structure. References Burroughes, J.H. et al (1990) Light-emitting diodes based on conjugated polymers. Nature, 347, 539-541. Chen, W., DeKee, D., & Kaloni, P. N. (2003) Advanced mathematics for engineering and science. Hackensack, NJ: World Scientific Publishing Company. Coulson, J.M., Richardson, J.F., Backhurst, J.R., & Harker, J.H. (1999) Chemical engineering: Fluid flow, heat transfer and mass transfer. 6th edition. Oxford, UK: Butterworth-Heinemann. Cran-McGreehin, S., Krauss, T.F., & Dholakia, K. (2006) Monolithic integration of microfluidic channels and semiconductor lasers. Optics Express, 14(17), 7723-7729. Curet, S., Rouaud, O., & Boillereaux, L. (2006) Heat Transfer Models for Microwave Thawing Applications. Excerpt from the Proceedings of the COMSOL Users Conference. COMSOL. [Online] Available from http://cds.comsol.com/access/dl/papers/1656/Curet.pdf [Accessed May 29, 2010]. Friend, R.H. et al (1999) Flexible light-emitting diodes made from soluble conducting polymers. Nature, 397, 121-128. Gustafsson, G. et al (1992) Flexible light-emitting diodes made from soluble conducting polymers. Nature, 347, 477-479. Irawan, R., Tjin, S.C., Fang, X., & Fu, C.Y. (2007) Integration of optical fiber light guide, fluorescence detection system, and multichannel disposable microfluidic chip. Biomedical Microdevices, 9(3), 413-9. Lagerstrom, P.A. (1996) Laminar Flow Theory. Princeton, NJ: Princeton University Press. Muneer, T., Kubie, J., & Grassie. T. (2003) Heat transfer: a problem solving approach, Volume 1. London, UK: Taylor and Francis. Moreno, I. & Sun, C.C. (2008) Modeling the radiation pattern of LEDs. Optics Express, 16(3), 1808-1819. Seeger, K. (2004) Semiconductor physics: an introduction. 9th edition. Vienna, Austria: Springer. Tang, C.W. & VanSlyke, S.A. (1987) Organic electroluminescent diodes. Applied Physics Letters, 51(12), 913. Urban, S. et al (2005) Temperature measurements in microfluidic systems: Heat dissipation of negative dielectrophoresis barriers. Electrophoresis, 26(11), 2239-2246. Zukauskas, A., Shur, M.S., & Gaska, R. (2003) Optical measurements using light emitting diodes. In Shur, M.S. & Zukauskas, A. (eds.) UV Solid-State Light Emitters and Detector: Proceedings of the NATO Advanced Research Workshop. New York, NY: Springer. Read More