Project Topics
Get Complete Project Material File(s) Now! »
Chapter 3 Microstructure Printer – Hardware & Software
Various methods described in the previous chapters clearly indicate the advantages of direct writing of CPs allowing fabrication of controlled patterns and structures directly on the desired substrate. To fabricate 2D patterns and 3D structures at micrometric levels of scale necessitates the following conditions to be met:
A reservoir that can dispense monomer/polymer dispersion
A precisely controlled micrometric nozzle to dispense the dispersion
Means to accurately position and control the polymer dispenser over the substrate
Control and monitoring capabilities over the printed pattern/structure.
Taking the above requirements into consideration, a 3D microstructure printer was constructed with an overall assembly architecture shown in Fig. 3.1, and the sections below provide the details of the hardware and the software used for this purpose.
Hardware
Reservoir & Nozzle – Borosilicate glass pulled micropipette
Dip-pen nanolithography discussed in section 2.2.6.3 clearly explains the disadvantage of depositing CP dispersion with the reservoir located elsewhere from the printing head. A reservoir that continuously dispenses the printer fluid would thus render direct-writing of CP structures a more lucrative option. As the structure of interest (micro-hairs) being printed would require only few picolitres, to print multiples of these structures during a single print, the volume of liquid to be contained would be just a few microliters.
Towards implementing a means of dispensing a small amount of liquid at only required times, the surface tension of the liquid was exploited. Rather than implementing expensive precision micro-pumps or air displacement micropipettes driven by a linear actuator to control when the fluid is dispensed and when it is not, a reservoir with an orifice was implemented. When the surface tension of the liquid meniscus at the orifice of the reservoir is greater than the pressure head of the liquid above it, the fluid flow out of the orifice ceases and when this surface tension reduces when the meniscus comes in contact with a wettable surface, a fluid channel is established as shown in Fig. 3.2.
When the distance between the meniscus and the surface increases, the surface tension increases, narrowing the fluid channel between the orifice and surface until the channel breaks and surface tension reverts to the state as in Fig. 3.2 (b), thereby stopping the liquid from flowing out further. This effect of surface tension was utilized by implementing a reservoir with a micrometric orifice and by moving the orifice and in turn the meniscus to make and break contact with the surface, the liquid flow was initiated and stopped respectively during the CP patterning process.
To fabricate this reservoir capable of holding the CP dispersion, the orifice diameter was determined by the smallest feature dimension of the intended pattern. In this case, the structures under interest were micro-hairs with dimensions depending on the required sensitivity to flow. The most sensitive hair-like structures in the human body, for example, present in the cochlear region of the ears, where bundles of stereo-cilia are present which respond to pressure waves caused from acoustic sources, range in a few nanometres in diameter [183]. These nano-metric scaled hair-like cilia are extremely sensitive but their extreme sensitivities become superfluous to the intended target application. In the same way the surface hairs on humans, which range from 100-250 µm in diameter, are far less sensitive, despite their large aspect-ratio, as evolution has defined their primary role as thermal insulators rather than as a sensing medium.
Inspired from nature, arthropods for instance and Cricket Cercus in particular, have their abdominal cerci covered by thousands of hair-like cilia called filiform hairs [184]. These structures are primarily used for detecting changes in viscous drag of the surrounding air flow (Fig. 3.3). These hair-structures range from 1-9 µm in diameter and are distributed from 0.3 to 1.3 mm in length over the surface of the cerci and have been observed to respond to flow velocities as low as 3 mm/s. Making the structures in micrometric scale thus makes it ideal as it doesn’t become too sensitive to pick up pressure waves arising from audible sources or mechanical vibrations and thus tamper with the actual flow measurement and at the same time ensures that these structures do not droop/sag down under their own weight as high aspect-ratio structures a few hundreds of micrometres in diameter do. Moreover, if the structures were to be made with a large diameter (few hundreds of micrometres) the rate at which the CP solidifies would increase dramatically, thereby slowing down the printing process and primarily limiting it to two dimensional structures. The sensitivities of the hair-like cilia to fluid flow also depend on the stiffness of the material apart from the physical dimensions. Though stiffness can be decreased by increasing the length of the hair-structures, limiting the hair-structures within a micrometric range makes choosing CP materials with low Young’s modulus highly critical towards fabricating flexible structures.
With the orifice diameter to create the micro-hairs optimally set to only several micrometres, the reservoir to hold the CP while printing was to be decided next. With the dimensional extremities of the micro structures capable of being produced with the set orifice dimensions ranging from 1µm in diameter and 300µm in length (essentially micro-dots) up to 9µm in diameter and 1300 µm in length, the volume of the micro-hair ranges from 2.5×10-16 m3, to a maximum of 8.5×10-14 m3. However, when solidifying, CPs, PEDOT:PSS dispersed in H2O for instance, become devoid of their dispersion medium (as essentially CPs are insoluble, but easily dispersible in water). Taking this loss of volume into account, to theoretically print up to 1000 microstructures with one fill of the reservoir at maximum possible dimensions within a micrometric scale, the minimum reservoir volume was set at 5µL.
Commercially available borosilicate glass microcapillaries were an ideal candidate for the reservoir with volumetric capabilities and dimensions that would suit the 3D printer’s requirements. Thick walled capillaries were commercially available from suppliers such as Harvard Apparatus and Sutter and Warner Scientific and these were essentially hollow glass cylinders, whose internal diameter ranged from 580 µm up to 1.62 mm of which the smallest diameter was chosen for this research purpose. Available in varying lengths, a section of desired length can be cut from these microcapillaries. However, the internal diameter was too large for the intended microstructures to be fabricated and at the same time, large orifice diameters imply the surface tension at the orifice would be much less than the pressure head of the liquid within the capillary causing the CP dispersion to flow out until the surface tension becomes higher than the pressure head to contain the liquid within.
To overcome these issues, the diameter of the orifice had to be reduced to a few micrometres in range. Usually the glass capillaries are heated over a focussed flame at the centre and pulled apart by applying force at either ends to create a smaller diameter pipette shape (Fig. 3.4). As a controlled approach for reducing the diameter, manual pulling of this glass over a high temperature flame pipette was not an ideal solution as there is limited to no control over the pulling force. An alternative to this manual pulling, using a programmable laser puller provides a greater and repeatable control of the orifice diameter.
A Sutter laser puller P-2000, was used to create the required glass micropipettes, herein referred to simply as µpipettes, from commercially sourced microcapillaries. The laser puller has a programmable laser scanning pattern, power output of the laser, velocity of soft pull and velocity of hard pull with a programmable delay before the execution of the hard pull. The parameters used for pulling the µpipettes used within the scope of this research are provided in Appendix A1. Using this Sutter P-2000 laser puller, µpipettes with orifice diameters ranging from 200 nm to 10 µm can be fabricated from 75 mm long thick-walled borosilicate glass capillaries with 580 µm ID and 1000 µm OD procured from Harvard Apparatus (GC100-7.5). Though the Sutter P-2000 laser puller was used to fabricate pipettes ranging from 200 nm to 10 µm, µpipettes with 5 µm orifice diameter were fabricated in batches of 60 µpipettes to create the indented micro-hair like structures.
Laser patterns and heating and pulling rates of the P-2000 laser puller were programmed to obtain 5 µm diameters, with various taper lengths (opening angles) towards comparing the effects of these opening angles during the printing process. Though higher taper lengths greater than 4 mm allowed printing of densely packed microstructures, the printing was highly unrepeatable, sometimes unprintable and the µpipettes themselves were prone to tip breakages at the interface due to increased resistance to fluid flow which prevented the target CP dispersion from reaching the orifice and forming a meniscus (Fig. 3.5).
Taper lengths less than 800 µm were also deemed unfit for printing as the pulling process involved implementing thinner scanning patterns with high values of laser power (heat parameter in the program) and a sudden large pulling force (hard pull) to break the pipette resulted in erratic uncontrolled dimensions of the orifice. Though further work into fine tuning the laser puller parameters would have yielded repeatable orifice diameters, these higher opening angles limit the density and height of the microstructures printed as the µpipettes themselves interact with adjacent microstructures, thereby deforming the already printed structures. The µpipettes used within the scope of this research all had taper lengths of 3±0.75 mm (opening angles of 20°±5°) which allowed printing of 1.3 mm long structures as close as 300 µm apart from each other.
As the µpipettes had an ID of 580 µm at the free end, normally available air displacement micropipettes or medical syringes were not suitable as a means of filling the laser pulled µpipette. For this purpose, MicroFil needles from World Precision Instruments were procured which had an OD of 350 µm allowing a clearance of 230 µm from the µpipette during the process of filling, allowing uninterrupted displacement of air. These 28 gauge needles were 97 mm long and demanded extra care during handling as the metallic needle tips, though extremely flexible, were still fragile.
Rough positioning stages – Thorlabs Micrometric Linear Stages
Positioning and moving these µpipettes is of critical importance to ensure controlled fabrication of microstructures and patterns. The following were the ideal requirements of suitable micrometric linear stages:
Maximum resolution 500 nm (for having better control over the position)
Maximum range 1.3 mm (to be able to print the highest intended structure)
Minimum speed 0.5 µm/s (along the x and y-axes to control the µpipette movement during 3d printing but precision down to tens of nanometres along the z-axis) Maximum speed 500 µm/s
With these requirements in consideration and to establish a low cost printing solution allowing end users (medical lab technicians/suppliers/hospitals) to setup the printer at an affordable price, Thorlabs MTS25/M-28 stages (Fig. 3.6) were selected to be ideal as a high specification/cost solution.
With a maximum resolution of 0.5 µm, the MTS25/M-28 stages also exhibited a total travel range of 25 mm which allows a maximum working envelope of 15.625 cm3; much greater than the requirements for the intended print volume. This large working envelope ensures safe approach and return distances of the printer head (µpipette) to and from the substrate on which the printing is to be done, allowing ease of µpipette change and substrate change without damaging either and allowing placement of probes for monitoring the printing process. To establish a 3D printing space, three of these stages were procured along with the mounting brackets (MTS25A-Z8, MTS25B-Z8, and MTS25C-Z8) and servo motor controllers (TDC001) from Thorlabs and were assembled in the configuration as shown in schematic representation (Fig. 3.7).
The stage mounted on the MTS25A-Z8 was fixed to the base, and aligned in such a way that it was set as the x-axis of the printer. On the linear platform of this x-axis stage, the MTS25B-Z8 adapter was installed which rotates the platform to be mounted on this adapter by 90° in the XY plane with respect to the linear platform of the x-axis stage. The stage attached to the MTS25B-Z8 acts as the y-axis stage and holds the MTS25C-Z8 right angle bracket on its linear platform. The right angled platform allows the final linear stage to be mounted at a 90° rotation about the x-axis aligning the linear motion of this stage to the z-axis. By individually or collectively controlling the three linear stages, the position of the linear platform of the z-axis stage can be controlled within a 25 mm3 envelope.
The TDC001 servo motor controller (Fig. 3.6) has a closed-loop digital PID position control with a quadrature encoder and limit switch inputs which accurately read the servo motor encoder and the limit switches present at either end of the controlled MTS25/M-28 linear stage. The closed-loop PID controller ensures repeatable positioning of the MTS25/M-28 linear stage. As the y-axis and the z-axis stages are mounted on the x-axis linear platform, with each of these stages weighing 310 g, the base stage (x-axis stage) is subjected to a total weight of 0.62 kg, which was well below the rated loading limits of 4.5 kg, ensuring optimal performance. The overall cost of implementing these linear stages was less than $4200 (USD) including the cost of the servo motor controllers.
Fine positioning stage – PI NanoCube® 3-axis linear stage
With the Thorlabs MTS25/M-28 stages capable of micrometric scale movements, 3D structures ranging down to few micrometres could be easily printed. However, a recently disassembled SICM [185] allowed easy access to a piezo nano-stage capable of nanometric precision. The P-611.3S NanoCube® (Fig. 3.8a) piezo stage had three axis movement capabilities of up to 100 µm along each axes in a single package, with strain gauge sensors to monitor the exact position of the nano-stage. The PI P-611.3S had a step resolution as high as 0.2 nm with a compact footprint of 44 mm × 44 mm with adjustable mounting points at 5 mm from the edges of the base. The Thorlabs MTS25/M-28 linear stages had a 43 mm platform with four 4-40 tap holes located 38.1 mm apart from each other providing a flush mounting space for the procured NanoCube®. The linear platform of the z-axis stage of the three axes Thorlabs setup was available for mounting this NanoCube® thereby primarily assigning the Thorlabs stages for micrometric precise positioning and large scale printing (structures up to 25 mm in all directions) and the PI NanoCube® acting as fine positioning stage, more importantly along the z-axis direction, with nanometric precision positioning of the µpipette location in the 3D space. The main reason for introducing increased precision along the z-axis direction is to provide precise positioning of the µpipette tip on the printer substrate without any damage around the fragile glass orifice as this orifice determines the dimension of the printed micro-structure.
The piezo controller was also readily available after the SICM teardown capable of controlling the three axes P-611.3S NanoCube®. The E-664 NanoCube® Piezo controller (Fig. 3.8b) had options to control the NanoCube® in both manual modes by adjusting the potentiometer knobs on the control panel of the controller and through electric voltage signals applied via shielded BNCs located at the rear panel of the controller. A 0-10V signal was needed for this controller to produce a corresponding displacement of 100 µm. The voltage signal to control this NanoCube® must have a maximum resolution of at least 200 µV to be able to accurately control the stage motion by its maximum resolution of 2 nm. The rear panel of the controller also houses output connectors for monitoring the axes positions of the NanoCube® stages as measured by strain gauge sensors. This voltage level output from the strain gauge sensor is a direct indication of the current position of the NanoCube® sage connected to the controller with 0.2 mV corresponding to every 2 nm movement of the corresponding axis. Using this voltage output as a feedback, the position of the NanoCube® was controlled by using PID control implemented in the control software.
Microelectrode Holders & Mounting fixture
With the Thorlabs linear stages allowing a printing envelope of 15.625 cm3, and the PI NanoCube® increasing the accuracy up to 2 nm along the z-axis, when the µpipette is mounted on to these three axes stages, the position of the µpipette tip can be moved and precisely controlled accurately within the printing envelope. However, as discussed earlier, these µpipettes are essentially made of borosilicate glass (and Quartz in case of nano-pipettes) and are fragile for rigid mounting options such as metal clamps and required a soft padded environment with a rigid and robust mounting option with minimal play as any play introduced during the printing process would amplify any surrounding vibrations and tamper with the shape of the structure being printed. Moreover, as mentioned before, glass capillaries come in various diameters and to allow compatibility with any custom made µpipette OD, a more standard mounting solution was sought.
Table of Contents
Abstract
Acknowledgements
Table of Contents
List of Figures
List of Tables
Glossary
Co-Authorship Forms
1. Introduction
1.1. Motivation
1.2. Research objectives and scope
1.3. Thesis Synopsis
2. State of the Art
2.1. Conducting Polymers
2.2. Conducting Polymer Fabrication & Processing
2.3. CP micro structures applications
2.4. Flow sensors
2.5. CP material selection
2.6. Summary
3. Microstructure Printer – Hardware & Software
3.1. Hardware
3.2 Software
4. 3D printed Micro-hair flow sensor – Design and prototyping
4.1. 3D printing of micro-hairs
4.2. Flow Sensor – Design and prototype development
4.3. Micro-Hair Flow Sensor – Characterization and Testing
4.4. Summary
5. Ultra-low flow rate measurement with CP micro-hair flow sensor
5.1. Transition to Ultra-low flow rates
5.3. Summary
6. Fluid-Structure Interaction simulation using Lattice Boltzmann Method & large deflection beam model
6.1. Mathematical model of fluid-structure interaction
6.2. Lattice Boltzmann Method
6.3. Structural Analysis
6.4. Verification of implemented MATLAB FSI mathematical model
6.5. FSI analysis using LBM based CFD and large deflection model
6.6. Summary
7. Portable Ultra-low Velocity Flow Sensor with Integrated Electronics
7.1. Portable flow sensor design
7.2. Sensor electronics
7.3. Portable and Tuneable Micro-hair Flow Sensor
7.4. Summary
8. Conclusion & Outlook
8.1 Research Outcomes
8.2. Research summary
8.3. Future Work
Publications List
References
GET THE COMPLETE PROJECT
3D Printed Conducting Polymer Microstructures as an Ultra-low Velocity Flow Sensor
