A Fast temperature regulator for Cytoskeleton studies
G. Velve Casquillas, J. Costa, F. Carlier Grynkorn, A. Mayeux, P.T. Tran
Introduction about fast temperature regulator for microscopy
Cell cytoskeleton is fundamental for cell function. It plays important roles in cell polarity and mitosis. Better comprehension of these processes is of major interest for biological studies. Controlling cytoskeleton dynamics, in particular microtubules, is essential to fully understand it. The most common ways to manipulate microtubule dynamics are: on one hand the use of drugs such as carbendazim (methyl benzimidazol-2-yl carbamate) or nocodazole, on the other hand changes in the temperature. Both, treatments with specific drugs or changes of temperature, perturb microtubule polymerization. Decreasing temperature causes tubulin polymerization to slow down resulting in shorter microtubules, these will completely disappear when temperature is below5°C. Since microtubules grow or shrink with velocity around 2µm/min, controlling temperature at a time scale lower than minute allow to precisely control the length of microtubule in the cell.
As thermal time constant of a system decrease with its size, miniaturized device allows to perform fast temperature change. Thus, microfluidic systems, which allow manipulating fluids at the microscale, are good candidate to perform fast temperature change. Moreover, with the development of technology based on the molding of PDMS (Poly Di-Methyl Siloxane), microfluidic shows a strong potentiality to fabricate tool dedicated to cell biology experiment.
We will present here a detailed protocol to fabricate and use a temperature control microfluidic device that allows changing the temperature in a range from2°Cto50°Cin less than 10 sec. Thereby this new device gives the ability to perform temperature changes and high resolution imaging simultaneously. The device described here has been optimized for S.Pombe yeast, which is a model organism for cytoskeleton studies but can eventually be adapted to other type of cell or organisms. Depending on the design of the channel containing the cells, that type of device can be coupled with other microfluidic functionality like mechanical deformation or drug injection. The microfluidic devices fabrication uses soft lithography of PDMS and molding methods, so production of these devices is inexpensive[2, 3]. These temperature control devices can be an interesting tool not only to study the microtubule cytoskeleton dynamics but also to determine the role of a specific protein at different stages of the cell cycle. Because essential gene cannot be deleted from the genome, to study it, biologists construct thermosensitive mutant strains. Then allowing protein activation or deactivation by changing the temperature. Fast temperature control could be a powerful tool to study such thermosensitive protein at different precise time point with possibility to deactivate them several times during cell cycle.
Device and temperature regulator for microscopy presentation
The temperature control device presented here is a bilayer PDMS (Poly Di-Methyl Siloxane) device bond on a 150µm thick glass coverslip. The channels containing cells are in contact with the glass and are topped by a channel network dedicated to temperature control (Figure 1A). A thin 15µm PDMS membrane separates both channel layers and avoids direct fluidic contact between the two layers. Flowing water with controlled temperature in the upper channel will, by heat diffusion that occurs through the PDMS membrane, change the temperature of the cell channel layer. Temperature control channels are simply a parallel network of 100µm thick channel for water circulation. In contrast cell channels can be fabricated with various design/shapes and thickness depending of the type of cells involved (yeast or bacteria) and the required applications (cell deformation, drug screening) . The temperature control setup is composed of two Peltier modules, a syringe pump and a microscope (Figure 1B). We use the Peltier module to control water temperature before injecting it in the microfluidic device. For this purpose, one Peltier modules is plugged upstream (inlet) of the device and the other downstream (outlet ). Once Peltier modules are at the desired temperatures, the temperature in the cells channel can be quickly changed from one to the other. This is possible because changing the direction of the water flow allows to changes the Peltier in which the water goes through before entering the microfluidic device. The temperature change of the device is then limited by the time required to reverse the flow and not by Peltier modules time constant.
Mold and device fabrication
Devices are fabricated using soft lithography of PDMS (Poly Di Methyl Siloxane).This method allows fabricating hundreds of microfluidic devices using a single mold fabricated by photolithography. PDMS has several advantages for the fabrication of microfluidic devices dedicated to cell biology. First, this elastomer is transparent, biocompatible  and permeable to gas . Those characteristics allow easy cell culture and microscope observations. Second, from a technological point of view, this material is cheap, easy to mold, has a low young modulus and can be easily covalently stick to a glass slide using plasma treatment. The small young modulus of PDMS is particularly interesting for implementation of fluidic valves . The ability to covalently bond the PDMS replica on a glass slide using plasma oxydation allow to irreversibly seal the PDMS microfluidic device on it. Moreover, two or more PDMS replicas can be bonded together using plasma treatment that allows the fabrication of multilayer microfluidic devices.
Preliminary step : Microfluidic mold fabrication
The first step required for microfluidic device fabrication is the mold fabrication. Microfluidic molds are fabricated using photolithography of SU-8 on silicon substrate. Since this method is extensively documented and well described in Microchem SU8 datasheet we will not focus on it in this publication. Since our device is composed of two layers of channel then two different molds are necessary for its fabrication. Before spin coating the photoresist, silicon substrate is coated with omnicoat to promote adhesion between resist and silicon. For “temperature control channel mold”, we used a 100µm thick SU8-2050 (or 3050) photoresist channel pattern on silicon wafer. Similarly, for “cell channel mold”, we used a 5µm thick SU8-2005 photoresist channel pattern on silicon wafer. To allow high flow rate of water with small pressure in temperature control channel we use thicker mold. The cell channels are thinner, around 5µm height, once they are designed to confine small cells.
Once molds are fabricated, it is then necessary to coat them with anti-adhesive treatment to avoid wrenching of the photo resist patterns during PDMS molding. For this purpose, we place it on a petri dish in a presence of a 10µL droplet of TMCS (chlorotrimethylsilane from Sigma) during three minutes. Natural evaporation of TMCS on the closed Petri dish allows deposition of a silane monolayer on the whole wafer. Since TMCS is harmful and volatile, this operation has to be performed under a fume hood. Once molds are treated with silane they can be use for PDMS device fabrication several times (10-100) before requiring a newer anti-adhesive treatment.
Photolithography of pattern down to 2µm does not necessary requires clean room facilities. Like in this study, photolithography equipment can be then installed on a classical fume hood. To fabricate our molds we used OAI UV lamp, Laurell spincoater and two Barnstead hotplates. At the exception of isopropanol, used for mold rinse, all chemical reagents have been bought from Microchem.
Mask design for temperature control channels is an array of twelve 1 cm long, 200µm width channels separated by 100µm width wall. Mask design for cells channels can be changed (serpentine, straight channel, feeding system) depending of the desired biological experiment.
The temperature control device fabrication procedure is described on figure2A. The two layers of channel are fabricated independently and are then covalently bond using plasma treatment. The full fabrication process is described underneath.
Step1: PDMS preparation
This step allows production of a homogenized mixture of PDMS and reticulating agent without air bubbles.
- Pour into a cup PDMS and reticulating agent with a ratio 10:1 as follow:
- In first, put into the cup 9 volumes of PDMS, and add in second 1 volumes of reticulating agent.
- Stir vigorously during 3 minutes (with a spoon) to homogenize the mixture. Normally at this step, since the stirring generate bubbles, the PDMS should become white.
- Put the cup in a vacuum chamber for 30 minutes for degassing and remove air bubbles.
- Changing the ratio between reticulating agent and PDMS will change the PDMS stiffness, which increases with the ratio of reticulating agent. The1:10ratio is often considerate as optimal but small change (1:7 to1:13) of this ratio will not be critical for our application.
- Pouring the PDMS before the reticulating agent in the cup allows easier homogenization of the mixture during stirring.
Step2.a : Fabrication of temperature control channels
This step describes the way to fabricate a PDMS replica of the microchannels pattern present on the mold.
- Once PDMS is degassed pour it on temperature control channel mold until PDMS get a thickness comprised between3 mmand6 mm. Wait 3 minutes to allow air bubble reaching the surface, blow gently on the PDMS surface to remove remaining bubbles. Once no air bubbles remain on PDMS put the mold into oven at65°Cduring 2 hours to reticulate PDMS.
- Once PDMS is reticulated, do a rectangular cutting with scalpel around the device and release it from the mold using a scalpel blade.
- Drill inlet and outlet using 20G needle.
- You can store the PDMS block on a closed Petri dish with microchannels side up.
- The recommended two hours of reticulation are not critical for our application and can be eventually extended overnight. Nevertheless, a too-long reticulation could lead to PDMS aging.
- During cutting around the device one should take a minimum 2-3 mmmargin to allow better water tightness of the future device. Moreover, do not forget that the shape of the PDMS block will have to completely cover the cell channel design.
- Avoid touching the PDMS channel side with your finger since the surface properties of the material is crucial for further plasma bonding.
- During inlet drilling, when the needle goes through the device a little PDMS cylinder remains at the end of the needle and should be taken off before needle removal. Needles used for inlet drilling should be smaller than steel tube coupler to allow tight seal between PDMS and steel tube coupler used during injection procedure. Moreover, to avoid PDMS crack during drilling, needles edge should be previously smoothed using sandpaper.
Step2.b: Fabrication of cell channels layer
To fabricate a bilayer PDMS device, in which the temperature control channels are separated from the cells channels by just few microns, the bottom layer has to be a very thin PDMS layer. This step describes the way to fabricate a 15µm thick PDMS membrane on the mold containing cell channel pattern.
- Once PDMS is degassed place the mold on a spincoater (Figure 2B) and then pour PDMS on it in order to cover around 30% of the surface. Launch spincoater at 500 rpm during 10 seconds (acceleration 100 rot2/min ) followed by a spincoat at 6000 rpm during 30 seconds (acceleration 500 rot2/min ) this will spread the PDMS on the mold as a thin layer ( around 15µm thick) .
- Put the mold on a hotplate at95°Cfor 30 minutes.
- You can store the mold with the PDMS membrane on a closed Petri dish with PDMS side up.
- Membrane thickness is important in our case, since its play an important role in heat transfer occurring between cell channel and temperature control channel. The thickness of this membrane depends mainly of two parameters one is the viscosity of the PDMS, the second is the rotation speed during spincoat. After mixing PDMS and reticulating agent you should not exceed 1 hour to do the spincoat step, since PDMS will slowly reticulate at ambient temperature and change its viscosity.
- As for temperature control the PDMS surface containing the channels should not be touched with finger since surface properties of PDMS are crucial for plasma sticking.
- The recommended 30 minute of reticulation at95°Care not critical for our application and can be eventually extended or shortened (15min=>3hours).
Step3 : Plasma bonding of both PDMS layers (temperature control and cells chanels)
To fabricate the PDMS bilayer assembly it is necessary to covalently stick both PDMS layers. For this purpose the most common technique is plasma treatment. In our laboratory we use a Plasma Cleaner (figure 2C), with a plasma flow module to control the pressure inside the plasma chamber.
- Once both PDMS layers are reticulated insert them with microchannels facing up on the plasma chamber. Once air pressure is stabilized between 500 mT and 1000mT launch RF power at high position for 30 sec. Then, put the temperature control channel replica on the top of cell channel immediately after plasma treatment.
- Put down the bilayer assembly on a hotplate at95°Cfor 30 minutes.
- Cut the PDMS membrane around the assembly with a scalpel and pull up the PDMS block to release it from the mold. Since both PDMS part are covalently bond by plasma, the two layers will be tear off simultaneously from the silicon mold. If the mold have been correctly treated with antiadhesive silane the release of the PDMS device should be easy.
- Drill inlet and outlet of the cell channel using 20G needle.
- You can store the PDMS device on a closed Petri dish with side containing microchannels upside.
- To avoid presence of dust on the surface before plasma sticking, one can blow device with an air gun immediately before placing them on the plasma chamber. Another alternative is to stick and release removable tape (magic 3M scotch) on the channel surface just before plasma step. (The tape designed here does not interfere with the plasma treatment).
- Plasma step is critical for microfluidic fabrication, mistake in time, pressure or presence of impurity in the air can lead to inefficient plasma treatment. In normal condition the plasma should be purple, white/pink plasma indicates too high pressure and evanescent plasma indicates too low pressure. Particularly the presence of refluxing oil from the vacuum pump in the plasma tube leads to inefficient plasma sticking. If the plasma treatment time is too long (more than 1 min) sticking efficiency between PDMS and glass will be lower.
- Contact between the two parts has to be done during 1 or 2 minutes after the plasma treatment. Larger waiting time will lead to less efficient sticking. Moreover, once the two parts are in contact do not try to reposition them.
- Cell channel should be positioned under the center of the temperature control channel network to reach better temperature uniformity. Inlets and outlets of the cell channel must not be under the temperature control channel network since it will be necessary to drill injection hole on it later.
- After PDMS device release from the mold, a PDMS membrane remains on the zone of the mold which was uncovered by the device. One can clean the mold using tweezer or doing PDMS roll with gloved finger.
Step4 : Plasma bonding of the bilayer assembly on a glass coverslip
To close the cell channel layer it is necessary to bond it on a glass slide using plasma treatment.
- Place glass coverslip and PDMS assembly with microchannels facing up on the plasma chamber. Once air pressure is stabilized between 500 mT and 1000mT, launch RF power at high position for 30 seconds. Immediately after plasma treatment, put down the PDMS assembly up to the glass slide.
- Put down the microfluidic device on a hotplate at95°Cfor 30 minutes.
- If the plasma treatment worked well the contact area between PDMS device and glass should spread in seconds all over the surface without help. If it not the case and some non conformal contact remained (white area) one can push gently the PDMS on glass with tweezers (do not apply to much force or you take the risk of collapsing the channels).
- To facilitate device handling and imaging, one can use an homemade chip holder as show in figure 2D.
Temperature regulation setup : first installation
The experimental setup is composed of an inverted microscope ( Nikon confocal with 100x immersion objective) two Peltier modules a water tank and a syringe pump. Beginning from this basic setup it is necessary to add two Peltier temperature controllers and two peristaltic pump to control and maintain Peltier module temperature (Figure 3 A-B).
Step1 : Peltier module microfluidic connection
In this step we will prepare Peltier module fluidic inlet/outlet to allow them to fit with microfluidic device. Each Peltier have two independent couple of inlet/outlet. In this setup we wil use only one of each. In the following paragraphs, Peltier bottom tubing will designate the Peltier tubing on the side connected to the microfluidic device.
Only one of the bottom metallic tubes of the Peltier will be connected to the microfluidic device. To ensure good fitting, between microfluidic device and Peltier bottom metallic tubing plug a 2cm (1.14 ID) PE tubing (furnished with Peltier) to it and then plug a 4cm «Microline» tubing (0.5mm ID) downstream. This tubing assembly is terminated by a stainless steel tube coupler that will fit with inlets of the microfluidic devices. Eventually one can insert a second stainless steel tube coupler inside at the interconnection between «Microline» and PE tubing to facilitate fitting between both tubing. The 6cm long assembly (figure 3C) is long enough to easily handle and plug it on the device and short enough to limit heat transfer between water and environment.
The other side of the Peltier module (upper side) will get the same kind of assembly but with less constrains on the length of the tubing since the water temperature is not critical at this stage. One of the Peltier upper tubing will be plug on a syringe pump and the end of the second Peltier upper tubing will be immerged in a water tank. In our case both «Microline» tubing plugged to Peltier upper side was50 cmlong to be able to reach the syringe pump (or the tank). Since the tubes are deformable, increasing the length of the tube has the disadvantage of increasing the fluidic time constant of the system.
- To avoid accidental lifting of tubing or device movement when touching tubing, one should fix the “Microline» tubing at different strategic points and particularly around the microscope with tape.
- At low temperature, condensation could appear on the Peltier module leading to eventual water droplet falling on the microscope. To eliminate this problem it is possible to add a small sweatband at the bottom of the Peltier module.
- To minimize heat loss at the Peltier bottom metallic tube one can cover metallic part with adhesive tape.
- To allow positioning of Peltier module directly up to microfluidic device inlet we used a common gallows with mobile tweezers.
- To our knowledge, this microfluidic temperature control system can be use with all type of inverted microscope.
Step2 : Connection of peristaltic pump to Peltier module
Each Peltier module is plugged to two independent water cooling closed loops drove by peristaltic pumps with a water reservoir of2 liters. Each Peltier module is connected to a “Harvard 66 peristaltic pump” and a water bottle using “Tygon R-1000”silicon tubing.
- Since they do not influence the system thermal time constant, peristaltic cooling pump can be placed far from the hearth of the system, under the microscope table for example.
- When using Peltier at low temperature like1°C, Peltier modules may not be able to maintain their temperature. To solve this problem one can increase peristaltic pump flow rate of water or decrease water temperature using bottle of water buried in ice. For low temperature experiment during several hours, the water reservoir should be at least 1-2 liters, immersed in a gasket full of ice with a water flow rate of 6 liter/hours. To avoid this kind of problem and the use of costly peristaltic pump one can use the thermal cooling module TCM1 from Warner Instrument or use a thermostated chiller.
- To avoid accidental lifting of tubing or device movement when touching tubing, one should fix tubing at different strategic points and particularly around the microscope.
- Because of the high water flow rate involved for peltier cooling, operator should take care of the tightness of the peristaltic pump tubing fitting on Peltier module. Moreover user should take care of ageing of the silicon tubing in contact with the rotary part of the pump. For more security in case of leakage, peristaltic pumps should be place inside a plastic bowl.
Step3 : Positioning Peltier devices
In this setup we used two SC-20 in-line Warner Peltier devices. The two Peltier devices are sustained4 cmabove the objective area using gallows and tweezers. Both of them are electrically plugged to Peltier temperature controllers with the cable furnished with SC-20 Peltier modules.
Biological validation of temperature regulation setup
Step1 : Device preparation and cell injection
- Plug a 24G needle on a 2-3 mL syringe and insert a «Microline» tube ended with stainless steel tubing.
- Fill syringe and tubing with your cells in renewed medium. For experiment convenience, one should avoid the presence of air bubbles in the syringe/tubing.
- Inject gently the cells in the device. Be sure that no air bubbles remain in the device at the end of the operation.
- If the cell channel is not designed to allow medium renewal and if the biological experiment will run for more than 2/3 hours the cell will begin to miss nutriment and the device could dry due to evaporation through PDMS. To limit this phenomena one should plug a 1cm long «Microline» tube at the device outlet and inject medium from the device inlet to fill outlet tubing without bubble. Once the outlet tube is full of medium then inject the cell from device inlet and cut the inlet tube at the same length than outlet one. Using this trick, medium lost in the device by evaporation will be renewed by medium from the tubes plug at inlet/outlet and allow experiments up to 9 hours depending of the cell concentrations.
- While injecting, one could put down the device on a black paper to facilitate visualization of air bubbles. Since contrary to water, air has refraction index far from the one of PDMS, air bubbles appear in white in the device since channel filled with water become almost impossible to see.
Step2 : Installing device on the setup
- Install a syringe filled with water on the syringe pump
- Plug both Peltier bottom tubing on the device inlet and outlet of temperature control channels
- Launch the syringe pump until all tubing became filled with water. This operation allows to test the device plugging and to ensure that the entire capillary are filled, which is an essential condition for fast temperature change. At this step Peltier module should be at ambient temperature to avoid temperature change in the device during capillary filling.
- Wait at least 1 minute to allow water completely stopping flowing on the tubing
- Launch peristaltic pump at 100 ml/min
- Set Peltier device at the desired temperature
Step3 : Performing temperature change
Once device is plugged, capillary filled and Peltier set at desired temperature the setup is ready to perform fast temperature change. To change device temperature it is necessary to push or refill the syringe (figure 4A). Here cold Peltier module on the syringe pump side will be call Peltier 1 , and the hot Peltier module on the water tank side Peltier 2.
- Push syringe at 2.5mL/min flow rate to thermalise the microfluidic device with water passing through the upstream Peltier module 1. There is a fluidic delay of 10-15 seconds between the action of the syringe pump and the beginning of the temperature change, this delay should be take in consideration for precisely timed experiment. Once the temperature begin to change, ten additional seconds will be necessary to reach the desired temperature value (with a precision <1°C).
- Refill syringe at 2.5mL/min flow rate will change the upstream Peltier module and will thermalise the microfluidic device with water passing through the Peltier module2. Inthis case fluidic and thermal time constant remain the same (10-15 sec of waiting and 10 sec to perform the temperature change).
- To prepare further temperature change, one can change the temperature of the downstream Peltier module without influencing microfluidic device temperature. Peltier modules used here generally require 1-3 min to reach desired temperature. Reaching Peltier temperature close to0°Ccould be longer than 3 min depending of the peristaltic pump flow rate and temperature.
Figure 4E show S.Pombe microtubule depolymerization using this kind of procedure with Peltier set at1°Cand microtubule re-polymerization when coming back at ambient temperature. (Using Immersion objective)
- Water flow rate is an essential parameter; decreasing flow rate could lead to improper microfluidic device temperature. Figure 4B-C shows dependence of microfluidic device temperature as a function of water flow rate with a Peltier module set at1°Cand39°C.
- The temperature measured in the device will be different from the Peltier device temperature. For example with a Peltier module set at1°Cand a room temperature of24°Cthe microfluidic device temperature will be 2.7°C. (Figure 4D give the correspondence between Peltier and device temperature)
- The presence of immersion objective is a critical point since it acts as a heat sink and shifts the microfluidic device temperature. With a Peltier module set at1°C, if immersion objective is in contact with the microfluidic device the temperature shift almost reach 4.5°COne can limit this problem by moving immersion objective far from the experiment area during the time-lapse between two photos (by putting it on the side of the experiment area, downstream or out of contact).
- When using immersion objective, temperature change generate materials dilatation/contraction in the objective leading to focus drifting. The resulting drift depends of the objective and is around 0.5 µm/°C. Although temperature change in the microfluidic device is very fast, a transient temperature gradient remain in the objective during 2-3 minutes before reaching permanent state leading to focus drifting during this time-lapse. Nevertheless temperature change obtained with this setup is reproducible leading to predicable focus drift.
- During syringe refilling, bubbles could appear in the tubing and the syringe leading to higher fluidic time constant. Moreover, presence of air in the tubing could lead to microchannels corking with air bubble leading to temperature non uniformity.
We described here a protocol which allows fabrication and use of fast microfluidic temperature control device and setup. That kind of system allows fine control of tubulin polymerization and thus a full control of polymerization state of the cytoskeleton. The ability to couple this temperature control with other microfluidic functionality like cell deformation will open new opportunity of biological experiments.
Mold fabrication :
- Spincoater, Laurell, CZ-650 series
- Hotplate, Barnstead int, model HP131720-33
- UV lamp, OAI, model 30 with OAI intensity controller model 2105C2
- Photoresist, microchem ,SU8 2005
- Photoresist, microchem ,SU8 2050
- Photoresist, microchem ,SU8 developer
- Photoresist, microchem ,omnicoat
- Isopropanol, AnalaR, normapur
Device fabrication :
- Hotplate, Barnstead int, model HP131720-33
- Oven, MEMMERT, 14L UNB100
- Plasma cleaner, Harrick plasma, “Extended plasma” cleaner with “plasmaflo” pressure controller
- Inlet outlet drilling: 20G needle smoothed with sandpaper.
- PDMS, Sylgard, 184
- Coverglass, dow corning, 24×40 mm ref 2940-244
Temperature control setup :
- Peltier module, Warner, SC-20 Dual In-line Solution Heater/Cooler
- Peltier Bipolar Temperature Controller, Warner, “CL-100”
- Syringe pump, Harvard apparatus, phd push pull prog high force
- Peristaltic pump, Harvard apparatus, High accuracy peristaltic pump model 66
- Peristaltic pump tubing: Tygon, R-1000 1/8in. ID * 1/4in. OD
- Syringe pump/peltier tubing, Harvard apparatus, «Microline» tubing 0.5mm ID*1.5 mm OD
- Peltier metal tube/«Microline» tubing interface, Warner, Polyethylene tubing 1.57OD*1.14ID (furnished with Peltier module).
- microfluidic device/” «Microline» tubing” interface, Harvard apparatus , Stainless steel tubing coupler, 23 Ga,8 mm
- Syringe for water injection, monoject, 140cc syringe with luer lock tip
Cell injection :
- 2 mL syringe, Terumo
- 24G needle, Terumo
- «Microline» tubing 0.5mm ID*1.5 mm OD, Harvard apparatus
- Microfluidic device/” «Microline» tubing” interface, Harvard apparatus, Stainless steel tubing coupler, 23 Ga,8 mm
4 : Changing downstream temperature does not affect device temperature and then allow to adjust the Peltier temperature for next temperature change.
(B)(C) Device temperature as a function of water flow rate with objective not in contact and Peltier set at1°C(B) and39°C(C).
(D) Calculation of device temperature as a function of Peltier module temperature with 30µL/sec water flow rate with ambient temperature of25°C. Device temperature is given by the following equation:
A=0.071 when objective is in contact with the device coverslip
A= 0.187 when objective is not in contact with the device coverslip
(E) S.Pombe cytoskeleton dynamic control performed in our device. At6°Cmicrotubule depolymerize and finally depolymerize when device coming back to ambient temperature.
1. G. Velve-Casquillas, M. Le Berre, M. Piel and P. T. Tran, Nano Today. 5 (2010) 28-47
2. D. C. Duffy, J. C. McDonald, O. J. A. Schueller and G. M. Whitesides, ANALYTICAL CHEMISTRY-WASHINGTON DC-. 70 (1998) 4974-4984
3. J. C. McDonald and G. M. Whitesides, Acc. Chem. Res. 35 (2002) 491-499
4. M. C. Belanger and Y. Marois, J Biomed Mater Res. 58 (2001) 467-77
5. S. G. Charati and S. A. Stern, Macromolecules. 31 (1998) 5529-5535
6. M. A. Unger, H. P. Chou, T. Thorsen, A. Scherer and S. R. Quake, Science. 288 (2000) 113-6