Team:TU Delft-Leiden/Project/Microfluidics



From paper microfluidic devices to "lab on a chip" diagnostics, microfluidics is an exciting field with huge development potential. Several characteristics of microfluidics make it an excellent technology to couple with synthetic biology: It creates a closed system, meaning cells can be contained reducing risk both of outside contamination and the spread of GM organisms into the outside world. Microfluidics also allows for the precise assaying of small volumes, with the same amount of analyte guaranteed every time. These qualities of containment and precision make it the perfect technology to employ in the easy-to-use modular biosensor system we envision. Microfluidics then is not only a useful tool for research, but a technology that could help bring synthetic biology into the world outside the lab.

Microfluidics Device for Electron Transfer

Since the final vision for Electrace is for a product which can be used by non-scientists in the field, microfluidics is a central part of this. It was therefore important to prototype a microfluidics device with built in electrodes. A novel proof-of-principle device was designed and tested with voltammetry to determine whether measurable currents could be produced.

Figure 1: Electrodes are made by injecting carbon paste into blocked off side channels, and flowing an etchant through a main channel remove the walls, thus exposing the electrodes. [1]

Incorporating electrodes adds to the complexity of the microfluidics device, especially since both carbon electrodes and silver/silver chloride electrodes need to be built in. Several options were explored, including the injection of carbon paste into the device as describes in this paper [1], and the use of silver/silver chloride wire inserts.

Finally, the use of screen printed electrodes produced by Dropsens was chosen. This is a miniaturized device featuring a carbon working electrode, carbon counter electrode, and a silver/silver chloride reference electrode, screen printed onto a ceramic chip. It consists of To our knowledge the use of such a mass produced electrode in a microfluidics device has never been done. There are a couple of instances of custom-made electrodes incorporated into microfluidics[2][3}, however by using a readily available prefabricated device, fabrication time and cost is greatly reduced. Since a commercial electrode is used, it also means specifications are available, and periphery products (such as the DropSens Potentiostat) optimised for use with these devices can be used. These were to be embedded into the PDMS of the microfluidics device, providing a reservoir where the electrodes would be in contact with the fluid.

Figure 2: 3d images of the microfluidics device design, seen from the top and bottom.

Figure 3: Dimension drawing of the mask design for the microfluidics device. With a total height of 200 microns, this creates a volume of 33 microlitres.

Since the design features comparatively large dimensions, the complex high precision methods such as photo-lithography were not required to make the mold. Instead, we considered several simpler approaches. 3D printing was first considered, but was discarded for its limited resolution and poor surface finish (at this point, 3d printers still create a rough finish which would be problematic at the small scales of microfluidics. Laser-cutting was considered next. Several materials were considered and tested with the laser-cutter, before PVC electrical tape was chosen- as described here. This can be applied to a slide - and layered to vary the height - and cut either by hand for quick prototypes, or with a laser-cutter for greater precision.

Figure 4: Electrical tape applied to a glass slide provides a quick "low-tech" solution for making microfluidic mold, a popular method amongst "DIY-Biologists". Patterns can either be cut by hand, or for more precise results, with a laser cutter (as shown here).

The screen printed electrode was lightly bonded to the mold with a thin film of vacuum grease (a very light adhesive was needed - something to hold the electrode in position for pouring, but would still allow removal from the mold later), before being poured with PDMS. This resulted in the electrode being fully encased for PDMS except for where it was in contact with the tape - namely the fluid reservoir, and the opening for the electrical contacts. Since the introduction of a rigid material in the PDMS made tearing a greater risk, a higher ratio of catalyst than typical (8:1 PDMS:catalyst by weight) was used to create a firmer, more durable device.

Figure 5: The screen printed electrode was pressed onto the mold and held in place with vacuum grease

FIgure 6: PDMS cast with embedded electrode (the final dimensions of the channels were altered in light of early tests).

The device features a 33uL reservoir with an inter-digitated electrode (carbon-carbon-silver/silver chloride). Wires can be soldered onto the three exposed contacts for connection with a potentiostat or other instrumentation. This makes the device perfect for electrochemical analysis of small quantities of analyte.

Figure 7: The completed microfluidics devices


Once the MF device with the embedded electrode was produced, it was crucial to prove its capability to generate reliable and quantifiable results while performing electrochemistry. The test solution chosen for this purpose was 0.25M Ascorbic Acid while demi-water was intended for the corresponding negative control.

Initially 33 μl of demi-water was pumped into the device via a Harvard Apparatus 11 plus syringe pump set-up at the rate of 0.350 ml/min using 7mm ID syringes. The terminals of the DropSens electrodes are directly accessible and were soldered to extendible wires which were in-turn plugged into the DropSens commercial potentiostat. Two tests were carried out by applying voltages of 0.2V and 0.4V respectively. Theoretically, plain water is not capable of hosting redox activity and hence no/minute quantities of steady state current should be observed.

Next, the water was pushed out of the device following which it was filled once again with an equal amount of Ascorbic Acid solution (0.25M). As above, two experiments were performed at 0.2V and 0.4V respectively and then the corresponding results were analyzed. The steady state current values now observed must exceed the values of the previously conducted negative control. The results of the four experiments have been plotted below in figure 1:

Figure 8 :Experimental results of Ascorbic Acid (0.25M) and negative control(Water)

The initial peaks occur in accordance with the capacitor effect in electrochemical cells. Due to the close proximity of the two metallic plates (i.e. the electrodes) a quantifiable amount of capacitance develops in the system. Hence upon application of voltage at the Working Electrode, this pseudo-capacitor gets rapidly charged and then gradually discharges till the current flowing through the cell converges to a steady state value. The effect of this virtual capacitor is unavoidable but can be neglected by analyzing the steady state current samples. This effect is consequently observed in all four cases.

The corresponding results with water at the end of 300 seconds were 0.03 μA and 0.005 μA respectively. These minute values of current would ultimately converge to zero if the experiment is conducted for long enough and sufficient time is provided for the capacitor to completely discharge. This result is in accordance with expectation. This can also be verified from the following two graphs where the water curves have been enlarged (fig 2 and fig 3). Clearly, a visible negative gradient exists even after 300 seconds.

Figure 9:Water at 0.2V

Figure 10:Water at 0.4V

On the contrary, continuously sustained steady state current values of 0.28 μA and 1.35 μA were observed when 0.2V and 0.4V were applied respectively to the ascorbic acid solution (0.25M) (figure 8). Upon application of a positive potential, the ascorbic acid gets oxidized and loses electrons while the Carbon electrode gets reduced. This leads to the formation of the persistent positive current as presented above.

The results of the experiments clearly indicate that the redox reactions responsible for the continual current were proceeding unhindered inside the micro – fluidic well. Hence the validity of the aforementioned device was established. The device can now be further utilized for various kinds of electro-analytical experiments such as Cyclic Voltammetry, Linear Sweep Voltammetry etc. Finally we propose utilizing this device to characterize BioBricks belonging to the ET module by attempting to probe the current emanated by the E. coli.

In the future when the entire ET module is ready a tentative protocol to be followed is as follows:

  • Prepare 10ml of analyte containing the E.coli in a flask.
  • Add the required amounts of TNT and wait for 4 hours
  • Use a syringe to pump 33 μl of the analyte + DDT into the MF device with the embedded electrodes
  • Measure the current at the Counter Electrode while applying the voltage at the Working Electrode.
  • A positive current would indicate that the electrons are being transported out of the bacteria

Figure 11: The general protocol

Mother Machine

The “Mother Machine” is a microfluidic device designed for the study of single E.coli cells via controlled restriction of cellular movement, through mechanical means of the design of the channels.

Figure 12: Three Dimensional schematic of the "Mother Machine" microfluidics device [3].

The device consists of a wide central trench, through which fluid is flowed. Smaller closed channels, ranging from 0.3 to 0.8 um width, branch off perpendicular to the main channel. This confine individual cells of E. coli and thus serve as growth channels - allowing for the study of individual cells, without the need of chemical fixation. Having kindly been given a set of molds for this devices from the TU Delft department of Nanobiotechnology, we fabricates mother machines, and intended to use them for characterization in two modules:

Characterizing Curli - When uninduced, no biofilm should be produced, however when induced with rhamnose a biofilm should be produced. By flowing induced cells into the main channel of the Mother Machine, formed biofilm should then be observable. Due to time restraints this experiment was not carried out, however it was used to test the BBa_k13160016 BioBrick in the main channel, a positive control for this module.

Characterizing landmine detection - The small side channels of the mother machine can harbour single cells along their longitudinal axis. This allows the fluorescence of each cell to be quantified individually - a significant step in creating a truly quantitative and accurate output for a biosensor. Unfortunately we were unable to use the single side channels of the mother machine for this purpose.

Figure 13: Cell culture is flowed into the device with a Harvard Apparatus 11 Plus syringe pump.

For results of the characterization performed with the Mother Machine, please visit Curli Module - Characterization.


We made our microfluidic devices by casting Polydimethylsiloxane (PDMS) onto a positive mold in the pattern of the desired channel layout(made via photolithography or laser cutting) Once set, the PDMS is removed from the mold is bonded to a glass slide, creating a relatively low cost device with extremely precise fluid channels.


Mold (PDMS, Photoresist, lasercut, or similar)


PDMS Catalyst

Microscope Cover Slips

Tygon Tubing


Vacuum Pump

Plasma Preen Machine

Hole Cutter



- Mix PDMS and catalyst with a ratio 9:1 by weight, mix thoroughly (typically, about 10-15g is needed for one device)

- Construct a tray out of foil around the mold

- Pour pdms mixture into mold to a thickness of 5-10mm

- De-gas mixture in a vacuum pump for ~20 minutes or until no bubbles are visible

- Place in an 80C stove for 10 minutes

- Cut out device from the mold with a scalpel

- Peel the device off the mold with a pair of tweezers

- Cut holes for tubing (through the top down)

- Clean a slide with ethanol and water, and dry thoroughly with a nitrogen gun

- Place PDMS plus clean cover slip in a petri dish and plasma activate for ~10s

- Press PDMS and cover-slip together, making sure a full seal is made

- Brush a thin line of PDMS around the seams and place in stove at 80C for 10 minutes

- Insert tubing into device

Paper Microfluidics

Paper microfluidics offer a simple, low cost method for employing some of the advantages of microfluidics for analytical purposes. Disposable analytical devices can be made easily and quickly with no specialised equipment. We explored the possibility combine a paper microfluidic device with our Electrace E. coli, and printed electrodes, to create a “test strip” for our analyte, which can be used to measure the voltage output of our biosensor. There are several methods that can be employed to create a paper microfluidic device. The two we chose to look at due to their simplicity and potential for good results were method utilising Parafilm, by E. M. Dunfield et al[5], and FLASH (Andres W. Martinez et al)[6] - which utilises a form of photolithography.

Parafilm Method

We first tested the Parafilm method, as described here. This involves stacking a piece of filter paper, a heat resistant mask (in the shape of the desired fluid channels) and a layer of Parafilm, and applying heat and pressure to this stack. The principle is that the Parafilm melts, and is forced into the filter paper, creating a hydrophobic area. The mask prevents the Parafilm entering the paper, thus this area remains hydrophilic - providing the fluid channels. A regular clothes iron was used to provide heat, and pressure was provided by simply pressing down on the iron. We tested several types of paper, including tissue paper, coffee filter, filter paper (?) and Whattman grade 1) thicker grades of paper resulted in failure, as the Parafilm would not fully penetrate the paper. Best results were had with the coffee filter paper.

Figure 14: The Parafilm method allows simple devices to be made with no specialist equipment.

With a heat press, such as those used to print images on to t-shirts, more heat and pressure could be applied in a consistent manner, likely improving results even in thicker grades of paper, however such a device was not accessible. Whilst we has some degree of success with this method, it had several shortcomings which made us move onto the second method, namely:

- Didn’t work with thicker grade of paper (with our methods)

-Some Parafilm “leakage” across the edges of the mask, results in a degree of imprecision in the creation of the channels.

-The need to cut out the mask limits its ability to be used for complex shapes and small dimensions, particular if access to a laser cutter or similar is not possible.

FLASH Method

The FLASH method employs SU-8 photoresist to create the hydrophobic areas in the paper. Photoresist is an epoxy which polymerises when exposed to light. The PACE method involves soaking the paper in photoresist, placing a mask over the paper (inkjet printed onto transparency film) exposing to UV. The UV polymerises the photoresist not covered by the mask, creating a hydrophobic paper/epoxy composite. The paper can then be rinsed in acetone to remove all un-set photoresist from the masked areas, creating the hydrophilic channels.

First we tested whether undiluted SU-8 could fully penetrate the filter paper. Three types of paper were tested: coffee filters, a thin grade filter paper, and Whattman grade 1. The photoresist was applied to a sample of each with a small paint brush and visual inspection was used to check whether each sample had become fully saturated with the resin. Full saturation was not possible with thick papers such as the Whattman grade 1.

A batch of SU-8 soaked filter paper was made up and dried, and cut into 20x40mm samples strips. Each strip was to be tested for its hydrophobic/hydrophilic properties (and thus determining if the photoresist had worked as intended) by pipetting a small drop of loading dye on to the paper. One strip was soaked in a bath of acetone and rinsed with ethanol prior to any UV Exposure as a negative control. All of the SU-8 appeared to be washed off, as the paper resumed its original hydrophilic properties. A second strip was exposed to sunlight for 15 minutes before being soaked in acetone and rinsed with ethanol. This strip remained hydrophobic. Various UV were tested, however it is important to note that the optimal wavelength for th SU-8 is 365nM, so the UV source should be as close to this as possible - several of the UV sources tested were unsuccessful. Finally a 370nM lamp (name of device?) was used for exposure of 5 minutes.

Figure 15: Masked SU-8 soaked paper being exposed to UV light

After exposure, the sample was placed on a ~90°c hotplate, where an instantaneous change of colour was noted. The sample was left on the hot plate for 5 minutes to cure. This was also carried out with a masked sample, which was equally successful. A clear colour difference could be observed between the masked and unmasked areas.

Figure 16: Proposed concept for a paper microfluidics device for the Electrace cells.

To test the microfluidic strips, a drop of loading dye was applied to the end of the channel on the sample, where it moved up the channel through capillary action. The progress from capillary action was fairly slow however. Paper microfluidics hold great potential for low cost diagnostic devices, and could one day be used with a system such as Electrace. However the need for a way of immobilising cells in the device, and to incorporate electrodes, amongst other challenges, made the use of paper microfluidics infeasible for the scope of the project, so further development was terminated.

Figure 17: Completed paper microfluidic slips, with blue loading dye applied


1. Andrea Pavesi et al.: How to embed three-dimensional flexible electrodes in microfluidic devices for cell culture applications. Cite this:Lab Chip, 2011,11, 1593

2. Sirley V. Pereiraet al.: A microfluidic device based on a screen-printed carbon electrode with electrodeposited gold nanoparticles for the detection of IgG anti-Trypanosoma cruzi antibodies. Analyst, 2011,136, 47453.embedded electrode

3. Yong-Jin Yoon et al.: Microfluidics biosensor chip with integrated screen-printed electrodes for amperometric detection of nerve agent.

4. Moolman et al.: Electron beam fabrication of amicrofluidic device for studying submicron-scale bacteria. Journal of Nanobiotechnology. 2013 11:12.

5. E. M. Dunfieldet al.: Simple and rapid fabrication of paper microfluidic devices utilizing Parafilm® 10 Apr 2012. Available at:

6. Andres W. Martinez, Scott T. Phillips, Benjamin J. Wiley, Malancha Gupta, and George M. Whitesides. “FLASH: A Rapid Method for Prototyping Paper-Based Microfluidic Devices” Lab on a Chip 8.12 (2008): 2146-2150. Available at:

facebook twitter