Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 2963807 2017-04-11
70 Description
71 Title
72 Microwave Enabled Portable, Label-free, High-throughput Detection and
Content Sensing System
73 for Lab on a Chip Platforms
74 Field of the Invention
75 This invention lies in the field of microwave and microfluidic
engineering.
76 In recent years, there has been growing interest in droplet-based
microfluidics because of its promise
77 to facilitate a broad range of scientific research and
biological/chemical processes. Potential
78 applications can be found in many areas such as cell ana1ysis,1-4 DNA
hybridization,5detection of
79 bioassays, 6 bio-reactions,7-9 drug screenine and diagnostics. 11,12
Major advantages of droplet-
80 based microfluidics versus traditional bioassays include its capability
to provide highly uniformed,
81 well isolated environment for reactions with orders of magnitude higher
throughput (i.e. kHz). Most
82 droplet-based microfluidic studies rely on high speed imagine-17to
provide details of droplet
83 generation and transport, which usually require expensive and bulky high
speed camera, and
84 exhaustive post imaging analysis. In addition, in order to differentiate
subtle differences in droplet
85 content, fluorescent imaging is often used which, however, tends to
lower down the throughput
86 because longer residence time is needed for the droplet to stay in the
field of view in order to obtain
87 sufficiently high fluorescent intensity. Although this can be improved
by using a pulse solid state
88 laser that is synchronized wiih the camera, which further complicates
the system due to the need for
89 precise alignment and fluorescent labelling.
90 In contrast, electrical techniques allow the miniaturization of multiple
sensor arrays and their
91 integration into one single microfluidic chip with low power
requirement. Of these capacitive,
92 electrochemical and impedance based electrical detection methods are
widely available. In
93 electrochemical detection, the measurements are based on the
interactions between analytes and
94 electrodes or probes that usually occur in an electrolytic cell. They
are not able to distinguish
95 analytes that are not electroactive.18-2 In addition, the detection
electrodes are sensitive to variations
96 in temperature, ionic concentration and pH that affect the shelf life of
the sensor and shift electrodes'
97 response requiring frequent calibration.18,21,22 Conventional capacitive
and impedance detection
98 approaches operate at low frequencies, which causes either low signal-to-
noise ratio or long
99 response time and thus limit their applications to droplet microfluidics
where droplets are generated
100 at high frequencies. For example, the throughput achieved by a
capacitive sensor23 for droplet
CA 2963807 2017-04-11
101 detection was up to 90 Hz with reasonable sensitivity and for an
electrical impedance-based
102 detection' around 10.
103 Microwave technology, as a versatile non-optical method, has the
potential to address the above
104 issues because it eliminates the need for chemical modification or
physical intrusion of the sample
105 and operates at high frequencies (i.e. GHz). It differentiates
materials based on their electrical
106 properties including e1ectric41 conductivity and/or dielectric
constant. Previously we demonstrated a
107 microwave sensor that can be integrated with microfluidic devices to
differentiate single phase
108 fluids in microchannels and detect the presence of droplets at a very
low frequency (i.e. up to 1.25
109 Hz).25 The low detection frequency was mainly restricted by the
response time of the vector network
110 analyzer (VNA). In addition, the sensing of droplet content was not
achieved because insufficient
111 sampling of droplets did not allow the accurate determination of the
time for the droplet to arrive at
112 the capacitive gap, neither differentiation of the content changes.25
Also, in order to get a reliable
113 reading by the microwave sensor, the effect of droplet geometry on
sensing performance must be
114 eliminated, and the sensitivity of the microwave sensor must be
sufficiently high to detect subtle
115 variations.
116 In this study, we present a sensitive, low-cost, portable microwave
circuitry suitable for detection of
117 droplet presence and label-free sensing of individual droplet content
in microfluidic devices. More
118 importantly, for future point-of-care application purposes, we limited
ourselves to the choices of
119 cost-effective off-the-shelf components for developing the circuitry.
Basically the circuitry that
120 consists of surface mount components is able to generate microwave
signal and measure the
121 response of the sensor (reflection coefficient of the sensor) in a very
fast manner. We validated that
122 the system has a detection limit of several kilohertz (kHz) itself, and
in the experiments we reached
123 over 3 kHz. This microfluidic microwave system might potentially be
used as a coulter counter and
124 content analysis in many applications.
125
126
127
128
129
130
6
CA 2963807 2017-04-11
131
132
133
134
135
136
137 REFERENCES
138 1 L. Li, Q. Wang, J. Feng, L. Tong and B. Tang, Highly Sensitive and
Homogeneous Detection of
139 Membrane Protein on a Single Living Cell by Aptamer and Nicking
Enzyme Assisted Signal
140 Amplification Based on Micro fluidic Droplets, Anal. Chem.,
2014,86,5101-5107.
141 2 E. W. M. Kemna, L. I. Segerink, F. Wolbers, I. Vermes and A. van den
Berg, Label-free, high-
142 throughput, electrical detection of cells in droplets, Analyst,
2013,138(16), 4585-4592.
143 3 X. Ding, Z. Peng, S.-C. S. Lin, M. Geri, S. Li, P. Li, Y. Chen, M.
Dao, S. Suresh and T. J. Huang,
144 Cell separation using tilted-angle standing surface acoustic waves,
Proc. Natl. Acad. Sci. U.
145 S. A., 2014,111(36), 12992-12997.
146 4 Y. Chen, P. Li, P.-H. Huang, Y. Xie, J. D. Mai, L. Wang, N.-T. Nguyen
and T. J. Huang, Rare cell
147 isolation and analysis in microfluidics, Lab Chip, 2014,14(4), 626-
645.
148 5 H. Zhu, G. Wang, D. Xie, B. Cai, Y. Liu and X. Zhao, Au nanoparticles
enhanced fluorescence
149 detection of DNA hybridization in picoliter microfluidic droplets,
Biomed. Microdevices,
150 2014,16(3), 479-485.
151 6 M. T. Guo, A. Rotem, J. A. Heyman and D. A. Weitz, Droplet
microfluidics for high-throughput
152 biological assays, Lab. Chip, 2012,12(12), 2146-2155.
153 7 N. Wu, F. Courtois, R. Surjadi, J. Oakeshott, T. S. Peat, C. J.
Easton, C. Abell and Y. Zhu,
154 Enzyme synthesis and activity assay in microfluidic droplets on a
chip, Eng. Life Sci., 2011,
155 11(2), 157-164.
156 8 L. Mazutis, J.-C. Baret, P. Treacy, Y. Skhiri, A. F. Araghi, M.
Ryckelynck, V. Taly and A. D.
157 Griffiths, Multi-step microfluidic droplet processing: kinetic
analysis of an in vitro translated
158 enzyme, Lab Chip, 2009,9(20), 2902-2908.
159 9 H. Zhou, G. Li and S. Yao, A droplet-based pH regulator in
microfluidics, Lab Chip, 2014,
160 14(11),1917-1922.
161 10 E. Brouzes, M. Medkova, N. Savenelli, D. Marran, M. Twardowski, J.
B. Hutchison, J. M.
162 Rothberg, D. R. Link, N. Perrimon and M. L. Samuels, Droplet
microfluidic technology for
163 single-cell high-throughput screening, Proc. Natl. Acad. Sci. U. S.
A., 2009,106(34), 14195-
164 14200.
165 11 R. Sista, Z. Hua, P. Thwar, A. Sudarsan, V. Srinivasan, A. Eckhardt,
M. Pollack and V. Pamula,
7
CA 2963807 2017-04-11
166 Development of a digital microfluidic platform for point of care
testing, Lab Chip, 2008,
167 8(12),2091-2104.
168 12 A. M. Foudeh, T. Fatanat Didar, T. Veres and M. Tabrizian,
Microfluidic designs and techniques
169 using lab-on-a-chip devices for pathogen detection for point-of-care
diagnostics, Lab Chip,
170 2012, 12(18),3249-3266.
171 13 M. R. de Saint Vincent, S. Cassagnere, J. Plantard and J.-P.
Delville, Real-time droplet caliper
172 for digital microfluidics, Microfluid. Nanofluid., 2012, 13(2), 261-
271.
173 14 J. Lim, P. Gruner, M. Konrad and J.-C. Baret, Micro-optical lens
array for fluorescence detection
174 in droplet-based microfluidics, Lab Chip, 2013, 13(8), 1472-1475.
175 15 L. Mazutis, J.-C. Baret and A. D. Griffiths, A fast and efficient
microfluidic system for highly
176 selective one-to-one droplet fusion, Lab Chip, 2009, 9(18), 2665-
2672.
177 16 M. Fukuyama, Y. Yoshida, J. C. T. Eijkel, A. van den Berg and A.
Hibara, Time-resolved
178 electrochemical measurement device for microscopic liquid interfaces
during droplet
179 formation, Microfluid. Nanofluid., 2012, 14(6), 943-950.
180 17 G. D. M. Jeffries, R. M. Lorenz and D. T. Chiu, Ultrasensitive and
High-Throughput
181 Fluorescence Analysis of Droplet Contents with Orthogonal Line
Confocal Excitation, Anal.
182 Chem., 2010, 82(23), 9948-9954.
183 18 N. M. M. Pires, T. Dong, U. Hanke and N. Hoivik, Recent developments
in optical detection
184 technologies in lab-on-achip devices for biosensing applications,
Sensors, 2014, 14(8),
185 15458-15479.
186 19 J. Wu and M. Gu, Microfluidic sensing: state of the art fabrication
and detection techniques, J.
187 Biomed. Opt., 2011, 16(8), 080901.
188 20 J. Sochor, J. Dobes, O. Krystofova, B. Ruttkay-nedecky and P.
Babula, Electrochemistry as a
189 Tool for Studying Antioxidant Properties, Int. J. Electrochem. Sci.,
2013, 8, 8464-8489.
190 21 K. Mitton and J. Trevithick, High-performance liquid chromatography-
electrochemical detection
191 of antioxidants in vertebrate lens: glutathione, tocopherol, and
ascorbate, Methods Enzymol.,
192 1994, 233, 523-539.
193 22 N. Wongkaew, P. He, V. Kurth, W. Surareungchai and A. J. Baeumner,
Multi-channel PMMA
194 microfluidic biosensor with integrated IDUAs for electrochemical
detection, Anal. Bioanal.
195 Chem., 2013, 405(18), 5965-5974.
196 23 C. Elbuken, T. Glawdel, D. Chan and C. L. Ren, Detection of
Microdroplet Size and Speed
197 Using Capacitive Sensors, Sens. Actuators, A, 2011, 171, 55-62.
198 24 E. V. Moiseeva, A. A. Fletcher and C. K. Harnetten, Thin-film
electrode based droplet detection
199 for microfluidic systems, Sens. Actuators, B, 2012, 155, 408-414.
200 25 M. S. Boybay, A. Jiao, T. Glawdel and C. L. Ren, Microwave sensing
and heating of individual
201 droplets in microfluidic devices, Lab Chip, 2013, 13(19), 3840-3846.
202
8
CA 2963807 2017-04-11
203
204
205 Description of Prior Art
206 There have been the development of inventions regarding microwave
detection and microfluidics
207 multiphase flows.
208 Patent US 2005/0191708 Al Microwave Microfluidics
209 Patent US 6605454 B2 Microfluidic Devices with Monolithic Microwave
Integrated Circuits
210 Patent US 2009/0236330 Al Microwave Heating of Aqueous Samples on a
Micro-Optical-Electro-
211 Mechanical System
212 Patent US 2008/0277387 Al Use of Microwaves for Thermal and Non-Thermal
Applications in
213 Micro and Nanoscale Devices
214 Patent US 2010/0089907 Al Instantaneous In-Line Heating of Samples on a
Monolithic Microwave
215 Integrated Circuit Microfluidic Device
216
217
218
9
=
CA 2963807 2017-04-11
219 Summary of the Invention
220
221 System overview
222 The system illustrated in Figure 1 consists of a microfluidic chip
integrated with a microwave
223 sensor, a pumping unit which could be a pressure controller (Fluigent
MFCS-8C) or a syringe pump
224 (Pump33, Harvard Apparatus) depending on the particular study case, an
inverted microscope
225 (Eclipse Ti, Nikon) mounted with a high-speed camera (Phantom v210,
Vision Research) and the
226 developed microwave custom circuitry. Fluid reservoirs are connected to
the microfluidic chip via
227 ethyltrifluoroethylene (ETFE) tubing and connectors (Tefzel, Upchurch
Scientific). Two slightly
228 modified configurations (simple flow focusing and double T-junction)
were used for droplet
229 generation. For the detection of droplet presence, the simple flow
focusing geometry was used while
230 for the sensing of droplet content, the double T-junction geometry was
used where droplets with
231 different contents were alternatively generated by the two T-junctions.
Droplet generation and
232 transport were manipulated through the microfluidic channel network
design by adjusting the
233 pressures of the inlets or the pumping flow rate of the syringe pump.
The high speed camera was
234 used to record microscopic images and videos through the image
processing program ImageJ
235 (National Institute of Health, MD, USA) which were also used to
validate the experimental results
236 obtained through the developed circuitry. A data acquisition device and
Labview software (National
237 Instruments) were used to control the system and set off the computer
interface.
238 Materials
239 Fluorinated oil (FC40 from Sigma-Aldrich) with a 2% custom made
surfactant was used as the
240 continuous phase. The surfactant has a chemical structure of PFPE-PEG-
PFPE (or Krytox-
241 Jeffamine¨Krytox, where Krytox has a molecular weight of 7500 and
Jeffamine 900). D-(+)-
242 Glucose (Sigma- Aldrich) arid potassium chloride (EMD Millipore)
solutions were prepared in ultra-
243 pure water. Penicillin¨streptomycin¨ neomycin antibiotic mixture
(containing 5000 units penicillin,
244 5 mg streptomycin and 10 mg neomycin per mL), Fetal Bovine Serum (FBS;
Sigma-Aldrich) and
245 milk (contains 2% MF) were used in the sensing of droplet content
without further purification or
246 dilution. For further demonstration of the developed system for
potential bioapplications,
247 Alzheimer's disease (AD) testing was chosen. Tau-derived hexapeptide
(AcPHF6) which is used to
248 model the tau-protein aggregation related to AD was purchased from
Celtek Peptides. The assay was
249 carried out with orange G, which is a known inhibitor to the tau-
protein aggregation (Sigma-
CA 2963807 2017-04-11
250 Aldrich). AcPHF6 was prepared in ultrapure water at a concentration of
2.5 mg m1-1 as stock
251 solution, and diluted to a final concentration of 0.316 mM. All other
solutions were prepared in
252 morpholinepropanesulfonic acid (MOPS) buffer with 0.01% NaN3 and
adjusted to pH 7.2, and with
253 assay grade DMSO at 1% (v/v).
254
255 Microwave Sensor
256 The designed microwave sensor works essentially as a resonator. The
sensor structure is made of
257 two concentric copper loops similar to the one presented previously25.
Microwave signal is excited
258 by the outer coplanar transmission line loop, which supplies a time-
varying oscillating current
259 circulating around the loop and a magnetic field passing through the
loop. The inner loop with a
260 small gap constructs the resonator and the microchannel where droplets
are passing through is
261 aligned on top of this gap. When materials with different electrical
properties (permittivity,
262 conductivity) pass by the gap region, the capacitance of the gap
changes and the resonance
263 frequency shifts which can be used to characterize the materials. Take
water-in-oil emulsion as an
264 example, water droplets have a much higher dielectric constant (-80)
than the carrier fluid, oil (-2-
265 3). When a water droplet passes by the resonator, the resonance
frequency will be shifted which can
266 be used to detect droplet's presence. Similarly, when droplets with
different materials pass by the
267 resonator, the magnitude of shift in the resonance frequency can be
used to characterize the droplet
268 content. The resonance frequency shift caused by a perturbation in the
permittivity of the medium is
269 described by26.
At"¨ f AEit Eodv
270 (1)
f f(Eg.Eo + pH. H o)dv
271
272 where E0 and E are the electric fields before and after the
perturbation, HO and H are the magnetic
273 fields before and after the perturbation, f is the resonance frequency
before the perturbation, E is the
274 permittivity of the medium and itt is the permeability of the medium.
In this study, a spiral-shaped
275 capacitive region is designed for sensing purposes because it allows
the system to operate at lower
276 frequencies compared to T-shaped designs25, which thus allows
inexpensive off-the-shelf
277 components to be chosen for the circuitry design.
278
11
CA 2963807 2017-04-11
279 Fabrication
280 The microfluidic chip consists of two main components, a glass base
with the microwave
281 components and a polydimethylsiloxane (PDMS) mold with the designed
microchannels for droplet
282 generation and transport, which are fabricated separately and then
bonded together. Thus, the device
283 fabrication consists of two stages: microchannel and microwave
component fabrication. Microwave
284 component. The electrical traces for the microwave components are
fabricated using a combination
285 of photolithography and electroplating. Briefly, the positive
photoresist, S1813 (Rohm-Haas), is
286 spin-coated at 1500 rpm for 60 s onto a 50 nm thick copper film (EMF
Corporation) that is pre-
287 deposited on a glass slide and then baked at 95 C for 120 s. The
design is patterned into the
288 photoresist via UV lithography and subsequently developed with MF-319
(Rohm- Haas) for 2 min.
289 The patterned slide is then immersed in an acidic copper electroplating
solution (0.2 M CuSO4, 0.1
290 M H3B03, and 0.1 M H2504) and electroplated at 2 mA for 4 min and then
7 mA for 20 min. After
291 electroplating, the photoresist is removed with acetone leaving an
electroplated copper film
292 approximately 5 jim thick. Next, the base layer of predeposited copper
is removed by etching with
293 dilute ferric chloride (5%) (MG Chemicals). A passivation layer of a
mixture of PDMS (Sylgard
294 184, Dow Corning) and toluene (1 : 1 (w/w) PDMS/toluene) is spin-coated
at 4000 rpm for 60 s
295 followed by 1 h curing at 95 C to protect the electrical traces. A
subminiature version A (SMA)
296 connector (Tab Contact, Johnson Components) is then soldered to the
electrodes of the microwave
297 components to provide an external connection to the microwave
circuitry. Microchannel.
298 Microchannels are fabricated from PDMS using standard soft-lithography
techniques. The PDMS is
299 mixed in a 10: 1 ratio of base to curing agent, degassed and molded
against SU-8/silicon masters
300 which are fabricated using the same procedure developed previously23
and then cured at 95 C for 2
301 h. The molds are then peeled off from the masters and fluidic access
holes are made using a 1.5 mm
302 biopsy punch. Both the finished microwave components and the PDMS mold
are then treated with
303 oxygen plasma at 29.7 W, 500 mTorr for 30 s. The plasma treatment
process renders PDMS
304 hydrophilic; however, for water in oil droplets, the PDMS channels
should be hydrophobic to form
305 droplets. For this purpose, the walls of the microfluidic channels are
coated with Aquapel (PPG
306 Industries) to ensure that they are preferentially wet by the
fluorinated oil.
307
308
309
12
CA 2963807 2017-04-11
310 Microwave custom circuitry
311 Vector Network Analyzers (VNAs) are widespread tools for microwave
characterization due to their
312 accuracy and user-friendly interface. However, VNAs are expensive
normally which has driven the
313 development of inexpensive alternatives27-29. Regular VNAs have
limitations on the data sampling
314 rate and thus throughput which only allowed a very low throughput (i.e.
up to 1.25 Hz for droplet
315 detection). Another major disadvantage of such bulky benchtop setups is
their size which makes it
316 difficult to be widely applied for point-of-care applications. In this
regard, it is necessary to develop
317 portable yet affordable microwave circuitries that have comparable
accuracy and sensitivity as
318 commercially available VNAs. In this study, such a microwave circuitry
for label-free detection and
319 content sensing of droplets in microfluidic devices is developed.
Considering the microwave
320 structure used in detecting and sensing droplets, a microwave circuit
that measures the reflection
321 coefficient from a one port network is designed since the change in the
resonance frequency can be
322 monitored in the reflection cocefficient. The microwave circuitry is
mainly composed of three sub-
323 systems: i) signal generator, ii) power coupling unit, and iii) gain
detector as shown in Figure 2.
324
325 MICROWAVE SIGNAL GENERATOR
326 This subsystem consists of a voltage controlled oscillator (VCO) (Mini-
Circuits, ROS-2350-519+), a
327 voltage regulator (Rohm Semiconductor, BA17805FP-E2), a data
acquisition system (DAQ)
328 (National Instruments), an op-amp (Texas Instruments, LM358DR), and a
power supply (24V
329 battery) that supplies voltage to the op-amp and voltage regulator. The
VCO provides the required
330 microwave frequency by converting the input analog voltage, which
consists of two components:
331 one voltage source provided by the 24V power supply but regulated by
the voltage regulator to the
332 maximum of 5V and the other by the DAQ (0 - 10V) for tuning purposes.
Tuning voltages are
333 amplified by the op-amp with a gain factor of 2. The total amplified
voltage ranging between 0 to
334 20V controls the tuning voltage of the VCO which is measured by the DAQ
and LabView program
335 and characterized by a spectr.um analyzer (Agilent, E4440A). Serial
capacitors are used in order to
336 reduce the parasitic effects and filter the signal for the op-amp and
VCO input. The microwave
337 signal generator subsystem facilitates the sweeping over the desired
frequency range (1.9 GHz to 2.6
338 GHz). We designed the sensor operating below 3 GHz at which the
corresponding microwave
339 components are widely available and inexpensive that allow the total
cost of the electronic
13
CA 2963807 2017-04-11
340 components below $200. Wider frequency ranges can be achieved by
adjusting the tuning voltage of
341 the VCO.
342 POWER COUPLING
343 The primary function of the power coupling unit is to provide proper
microwave signals to the
344 sensor and the gain detector which would require careful isolation of
signals without sacrificing the
345 useful power. The high operating frequency at the GHz range brings in
more challenges in the
346 design and fabrication of the microwave components. First, the
impedance of the transmission lines
347 in the printed circuit board must match to the characteristic impedance
of 50 1/ because any
348 mismatch between the transmission line traces causes reflections and
reduces the performance27. For
349 this purpose, 0.79 mm thick FR-4 PCB material (c=4.34) is used and
microstrip impedances are
350 carefully calculated considering the trace width, thickness and
substrate height. The copper ground
351 (at the bottom of the PCB) and ground planes (on the top of the PCB)
are connected with vias closer
352 to 1/20th X where X is the wavelength of the signal flowing through it
to reduce noise'. In order to
353 minimize parasitic coupling to the transmission lines, separations
between the ground planes and all
354 other traces are designed to be at least three times larger than the
substrate thickness. Second,
355 reflections between different components must be controlled well which
include the reflection from
356 the attenuator, VCO, and directional coupler, to the microwave sensor
which is connected to the
357 circuitry through coaxial cables, and that from the microwave sensor
back to the VCO which have
358 disturbing effects on robust frequency generation. The above concerns
are taken into consideration
359 in the design of the power coupling and isolator subsystem.
Specifically, a high directivity bi-
360 directional coupler (Mini-Circuits, SYBD-16-272HP+) with 16 dB coupling
is used to regulate
361 microwave power to the resonator. Additionally, a 20 dB resistive
attenuator is used as an isolator
362 network which is chosen to isolate the reflected signal because of the
mismatch of the sensor while
363 not reducing the useful power significantly.
364 GAIN DETECTOR
365 An integrated circuit (Analog Devices, AD8302) is employed in the gain
detection subsystem,
366 which communicates with the microwave sensor and the power-coupling
unit. The signal traveling
367 from the signal generator is coupled by the bi-directional coupler to
the gain detector as a reference
368 signal and to the resonator, which is aligned with microchannel. Then
the gain detector enables the
369 amplitude and phase difference between the signal reflected from the
sensor and the reference signal
370 to be measured which is described by the reflection coefficient.
14
CA 2963807 2017-04-11
=
371
372 RL = ___________
Reflected Voltage (ZR ¨ Zo)
(2)
Incident Voltage (ZR + Zo)
373
374 where ZR is the frequency dependent input impedance of the device
presented in Figure 1(b) that
375 includes the resonator and the excitation structure and Zo is the
characteristic impedance of the
376 transmission line used for feeding the structure. The gain detector
converts the microwave signals to
377 DC signal, and this magnitude ratio of electronic signal is post-
processed and used to relate
378 nanoliter-sized droplet detection and sensing of its content. For
example, if different materials pass
379 by the sensor, the reflected signal would be different even though the
incident voltage would be the
380 same which can be used for detection and sensing of materials. The
system is able to detect ac-
381 coupled input signals from -60 dBm to 0 dBm. The output reflection
coefficient range can be
382 accurately measured between -30 dB to +30 dB which is scaled to
30mV/dB. The system is also
383 able to measure the phase over a range of 0 -180 . The minimum and
maximum levels of the
384 detection limits are characterized by the limit that each individual
log amp can detect as well as the
385 finite directivity of the coupler.
386 The LabVIEW program is used as an interface to collect and convert the
measurement data, and
387 control the system real time. Meanwhile, a calibration algorithm is
used to correlate the measured
388 data readings of the gain detector to the reflection coefficient of the
microwave sensor which carry
389 the information of the physical droplet system.
390
391 Results and Discussion
392 Prior to the droplet detection and sensing using the developed
microwave circuitry, its sensitivity
393 and accuracy is first evaluated by comparing its measurement results
with that obtained using a
394 commercial VNA (MS2028C, Anritsu). Table 1 below shows the comparison
of the measured
395 resonance frequencies for FC-40 (c=1.9), air (8=1), and water (c=78.54)
between the custom-made
396 circuitry and VNA. The developed microwave circuitry has very similar
performance to the
397 commercial VNA with the maximum difference of 1.283% found for water.
398
399
CA 2963807 2017-04-11
400 Table 1 Comparison of the resonance frequencies between custom
microwave design and the VNA.
Material (Liquid) f@Sii min using f@Sii min using Variation
Percentage
VNA (MHz) Custom Design (%)
(MHz)
Air 2588 2580 0.309
FC-40 2582 2573 0.349
Water 2417 2386 1.283
401
402 Then the detection and counting function of the developed circuitry was
thoroughly evaluated by
403 measuring the reflection coefficient of the resonator for various
fluids in microchannels as a
404 function of frequency. In order to prevent potential contamination
caused by the residual of the
405 previous sample; the microchannels were primed with the solution to be
tested for at least 15 min
406 prior to each test, and flushed with oil for 10 min. The tubing was
cleaned twice before
407 measurements by purging air and then with isopropanol.
408 As shown in Figure 3, the circuitry is able to differentiate between
fluids with permittivity effects
409 dominant (Figure 3a) and conductivity effects dominant (Figure 3b).
Different concentrations of D-
410 (+)-Glucose and potassium chloride solutions (KCI) were prepared with
ultra-pure water. The
411 minimum assessed concentration is 0.001 g/ml for KCI and 0.01 g/ml for
glucose. The frequency
412 step size was 0.1 MHz in the analysis. The lower detection limit was
achieved for the KCI solutions
413 because of the combined effects of permittivity and ionic conductivity
(dominant effect). For
414 example, the conductivity of potassium chloride increase from 17.84
mS/cm to 60.8 mS/cm when its
415 concentration increases from 0.01g/m1 to 0.05 g/ml, which were
experimentally measured using a
416 conductivity meter kit (Thermo Scientific, Orion 5-Star) after
calibration of the probe with three
417 different calibration solutions. While increasing KCI concentration
causes a decrease in the
418 resonance frequency, increase in the glucose concentration results in
higher resonance frequencies.
419 As well, concentration changes cause sharp decline in the reflection
coefficient (S11) so that the
420 change in resonance frequency can be monitored in the reflection
coefficient. The differentiation of
421 fluids with small differences in electrical properties validated the
dynamic performance of the
422 customized microwave circuitry along with the microwave sensor
integrated with the
423 microchannels.
16
CA 2963807 2017-04-11
424 High-Throughput Droplet Detection
425 The performance of the microwave circuitry for droplet detection and
counting is performed using a
426 flow focusing generator as shown in Figure 4. The channel height is 50
pm and assumed to be
427 uniform across the entire chip. The channel width smoothly narrows down
from 300 pm to 80 m at
428 the intersection which is the same as that of the dispersed phase for
generating droplet stably. The
429 wider channels were designed to lower down the hydrodynamic resistance
for easy pumping while
430 the uniform channel widths near the generator is the most stable design
for droplet generation31.
431 Initially, the droplet generation and transport were visualized and
characterized with the optical
432 microscope (Eclipse Ti, Nikon) integrated with a high speed camera
which captured images at a
433 frame rate of 9000 fps. Figure 4(a) shows the image of the generator
and generated droplets which
434 are around Inl considering the droplets are fully confined by the
channel (50um high, 80um wide)
435 with a length varying from 1-3 channel widths (80um 240um). Figure 4(b)
compares the droplet
436 generation frequencies measured by the optical imaging and microwave
sensor. When a droplet
437 passes the capacitive region (gap) of the resonator, the
electromagnetic field is disturbed by the
438 presence of the droplet and the dielectric change (from the oil phase
to the aqueous droplet phase)
439 causes a peak in the collected signal. The perturbation in the EM field
can be used to determine the
440 droplet generation frequency by counting the number of the
perturbations over a specific time
441 period. As shown in Figure 4(c), the signal peaks correspond to droplet
presence while the valleys
442 correspond to the carrier fluid (oil). The resonator was operated at
2.59 GHz which was the
443 resonance frequency for FC40 oil, and at this frequency the signal
change in reflection coefficient is
444 used to determine the droplet presence. It is worthy to mention that
the sinusoidal look of peaks at
445 very high droplet formation frequencies is caused by shorter and
unstable droplet spacing which are
446 likely due to the use of syringe pump which cause unpredictable
uncertainties31. Due to the
447 limitation of the maximum pressure that the pressure controller can
provide which limits the
448 throughput of droplet generation, a syringe pump was used in order to
evaluate the detection
449 performance of the microwave system. With carefully cleaning and
preparation of microfluidic
450 chips we reached as high flow rates as 4000 M1/hr for water and 4750
1/hr for the continuous phase
451 (i.e. oil). Correspondingly, we were able to generate droplets at the
maximum rate of 3.33 kHz
452 which can be detected with the microwave system successfully. Further
high frequencies can be
453 achieved by increasing the flow rate, which however tends to break chip
made of PDMS32.
454 Ideally, in order to detect 'a droplet, at least one signal level needs
to be sampled from carrier
455 fluid and one from dispersed (droplet) phase. This will result to give
a minimum and a maximum
17
CA 2963807 2017-04-11
456 value, and maximum detection limit can be estimated to be the half of
the signal generation rate
457 provided that the data sampling rate of the system is equal or larger
than signal generation rate. In
458 our system, since the signal generation frequency is at microwave range
(i.e., GHz), which is
459 extremely higher than data sampling that the data sampling rate
basically determines the maximum
460 detectable limit. However, since the droplet spacing is low in our very
high droplet generation
461 frequencies because of the droplet generation limitations explained
above, the collected data gives a
462 sinusoidal look. For a clearly resolved droplet detection data for very
high-throughput scenarios, the
463 droplet spacing should be at least one droplet length or higher. Here,
with a data sampling rate of 10
464 i_ts and well-spaced droplets, the theoretical droplet detection limit
of the developed microwave
465 system is 50 kHz.
466 Over a period of ten minutes, two million droplets were counted without
missing any droplet. In
467 order to assess the minimum sensible dielectric variation for
detectable droplets, it is important to
468 evaluate the resolution of the circuitry. An RMS noise level of 0.78 mV
has been calculated over an
469 interrogation time of 20 s. With this noise level, a resolution
threshold of 0.026 dB in reflection
470 coefficient was obtained. Since a resonant microwave sensor is designed
and used in the study,
471 electromagnetic energy into small region is accumulated, which is
extremely sensitive to small
472 changes. Likewise, utilizing the characteristic feature of microwaves,
i.e., operation at GHz
473 frequencies, allows working at shorter time scales. This gives a great
opportunity and advantages
474 over other detection techniques such as capacitive and electrochemical
which operate at lower
475 frequencies.
476
477 Droplet Content Sensing
478 Microwave sensing of droplet content was also carried out with the
spiral resonator design. The
479 spiral resonator was placed 8 mm away from the generator intersection.
The microfluidic channel
480 network and droplet generators are shown in Figure 5(a). Fluid pumping
and droplet generation
481 were controlled using a microfluidic pressure controller system
(Fluigent MFCS-8C) which can
482 provide more stable droplet generation29. For this set of experiments,
a double T-junction generator
483 was used to alternatively generate droplet pairs with different
materials encapsulated33 such as type
484 A and type B. The alternating generation works as follows. When one
droplet (i.e. with content type
485 A) is being generated in one of the T-junctions, it obstructs the main
channel as it is growing and
486 thus restricts the flow of the continuous phase, which causes a
dramatic increase in the pressure
18
CA 2963807 2017-04-11
487 upstream of the T-junction intersection. When the pressure increases to
a critical value, it drives the
488 continuous phase to neck and then pinch off the droplet34-36. After
pinch off the remaining interface
489 recoils back to the T-junction inlet. While this process is taking
place, the other T-junction generator
490 repeats similar droplet (type .13) formation process. By well-tuned
applied pressures, two alternating
491 droplets can be formed sequentially. During the formation of droplets,
although two pairs come to
492 close proximity, they do not coalesce or cross-contaminate at certain
operating regimes33. This
493 configuration has advantages over a simple Y-channel design in terms of
operation of the two
494 droplet generators and robustness. In addition, with this configuration
there is no need to add a
495 dilution stream in order to increase droplet spacing.
496 To demonstrate the sensitivity of the sensor and its potential to be
applied in the area of biosensing
497 with appealing features of no chemical and physical intrusion to the
sample, some materials were
498 strategically chosen. In particular, aqueous based solutions with
slight differences in their
499 concentration such as the potassium chloride solutions and glucose
solutions used here, which result
500 in similar dielectric constant and/or electric conductivity values,
were chosen to demonstrate the
501 sensitivity of the sensor. Two biochemical materials, fetal bovine
serum that is a widely used serum-
502 supplement for in vitro cell culture of eukaryotic cells and penicillin-
streptomycin-neomycin
503 antibiotic mixture (contain 5,000 units penicillin, 5 mg streptomycin
and 10 mg neomycin/mL), that
504 is widely used to prevent bacterial contamination of cell cultures due
to their effective combined
505 action against gram-positive and gram-negative bacteria, were chosen to
demonstrate its potential
506 for biosensing. Thawing fetal bovine serum and penicillin solution
started in the fridge at 8 C, then
507 completed at room temperature while the bottles was swirled gently to
mix the solution during the
508 thawing process. D-(+)-Glucose and potassium chloride solutions were
prepared in ultra-distilled
509 water, and 2% fat content of milk was used.
510 In order to ensure that the microwave system can differentiate droplets
with small difference in
511 dielectric properties, the experiments were carefully designed to
eliminate the droplet size effect. As
512 can be seen in Eq. (1), the frequency shift is a function of
permittivity difference and the relative
513 size of the droplet over the resonator. Considering that the
electromagnetic field is accumulated in
514 the sensing region, and the droplet width and height is confined with
the channel, droplet size has no
515 effect on the reflection coefficient as long as its length is longer
than the sensor region. This
516 consideration ensures that the response of the sensor to different
droplets is caused by dielectric
517 property variation, namely by the specific droplet content.
19
CA 2963807 2017-04-11
518 In order to verify the sensing performance of the microwave system,
ultra-pure water droplets were
519 generated from both T-junction generators with FC40 oil as the
continuous phase. The same sized
520 droplets and different sized droplets were sensed with the same signal
magnitude and the longer
521 droplets resulted in wider signals due to their longer residence time
in the sensing region.
522 Subsequently, a set of droplet pairs of the same size were sensed which
include a pair of Fetal
523 bovine serum and penicillin-strep.-neomycin, a pair of D-(+)-
glucose(0.2g/m1) and milk(2%mf), and
524 a pair of potassium chloride(0.03 g/m1) and water droplets. Figure 5(a)
shows the coordinated
525 optical imaging and microwave sensing results while Figure 5(b) and (c)
shows the microwave
526 sensing results.
527 The reflection coefficient difference between the fetal bovine serum
droplets and penicillin droplets
528 is -1.61 dB, which is 1.16 times lower, while -9.01 dB difference with
the baseline of carrier oil
529 FC40. As well, the difference between glucose and milk droplets is -
4.02 dB, and between KCI and
530 water droplets -3.45 dB. It is worthwhile that very low (-5 dBm) output
excitation power was used
531 in order to avoid any heating effect on droplets. These results show
that our microwave module is
532 very sensitive to nanoliter droplet permittivity contrast and can
easily distinguish various droplet
533 contents. Very high reproducibility is accomplished. This microwave
system can also be used with
534 other bio-materials for content analysis or for synthesis and reaction
monitoring. It should be noted
535 that the demonstrated throughput of sensing is not high; however, it is
limited by the throughput of
536 droplet generation for the particular scenarios considered here rather
the sensor which has been
537 demonstrated for high throughput sensing as shown in Figure 4.
538 To demonstrate that this platform has the potential to be used as a
tool for pharmaceutical
539 applications, it is applied to perform a similar assay developed to
screen inhibitors for tau-
540 aggregation that is linked with neurodegenerative disorders such as the
Alzheimer's disease (AD)37-
541 38. The tau-derived hexapeptide (AcPHF6) which is normally considered
as a model for tau-protein
542 aggregation in many assays was used as the peptide and orange G which
is one of the common
543 inhibitors used in the traditional assay was chosen for this
preliminary testing.
544 Figure 6 shows that the microwave sensor is able to differentiate the
droplets with and without the
545 mixture of peptide and inhibitor and the droplets with different
concentrations of the inhibitor
546 (orange G), which are 0.665 mM and 0.332 mM respectively (inhibitor I
and II respectively in the
547 figure). The peptide concentration was kept at 0.316 mM and all
droplets contain Thioflavin S (0.05
548 mg m1-1), which is a fluorescent indicator dye normally used in tau-
aggregation assays. The
CA 2963807 2017-04-11
549 samples were prepared in 4-morpholinepropanesulfonic acid (MOPS) buffer
of 20 mM with a pH of
550 7.2. There is only one set of droplets containing no mixture of orange
G and AcPHF6, which is used
551 as a base similar to the negative control in the traditional assay37.
552 It should be noted that the sensing shown in Figure 6 only demonstrates
that the developed
553 microwave and microfluidic platform has the potential to serve as a
tool for drug discovery or
554 pharmaceutical applications:These results are not a quantitative
measure of the effects of the
555 inhibitor on tau-aggregation because it is difficult to judge whether
the signal difference is caused by
556 the concentration of the inhibitor or the degree of peptide aggregation
induced by the different
557 inhibitor concentrations. To perform such an assay to quantitatively
compare with the traditional
558 assay would require systematic design of the microfluidic chip and
microwave sensor and require
559 further improvements on the sensor fabrication protocol as well to
improve its sensitivity, which is
560 beyond the scope of this study. However, with a calibration process of
drug assays, the platform
561 developed herein presents promising and insightful results for
Alzheimer's disease drug screening
562 assays.
563
564 \
565 List of figures
566 Figure 1. (a) A schematic description of the microwave-microfluidics
integrated device, (b)
567 schematic of microwave sensor with a spiral resonator design and an
excitation loop, (c) and (d) a
568 closer view of droplet formation channels and spiral capacitive gap,
respectively.
569 Figure 2. Schematic description of the microwave circuitry.
570 Figure 3. Reflection coefficient of the resonator for a series of
glucose¨water (a) and KC1¨water (b)
571 mixtures for testing the circuitry.
572 Figure 4. (a) Image of the generator and generated droplets, (b)
comparison of the droplet generation
573 frequencies; optical imaging vs. microwave sensor, (c) high-throughput
droplet detection.
574 Figure 5. Label-free content sensing of individual droplets. (a)
FBS¨penicillin, (b) glucose (0.2 g
575 m1-1)¨milk, (c) water¨potassium chloride (0.03 g m1-1) droplets.
576 Figure 6. Demonstration of sensing of droplets involving AcPHF6 and
orange G which are the
577 model peptide and inhibitor respectively used in traditional tau-
aggregation assays that is linked to
578 neurodegenerative disorders such as the Alzheimer's disease.
579
21
CA 2963807 2017-04-11
580 References
581 26 D. M. Pozar, Microwave Engineering, 4th edn, John Wiley & Sons, Ltd,
2012.
582
583 27 N. Suwan, Investigation of RF Direct Detection Architecture Circuits
for Metamaterial Sensor
584 Applications, University of Waterloo, 2011.
585
586 28 G. F. Engen, The Six-Port Reflectometer: An Alternative Network
Analyzer, IEEE Trans.
587 Microwave Theory Tech., 1985,25(12), 1075-1080.
588
589 29 F. M. Ghannouchi and A. Mohammadi, The Six-Port Technique with
Microwave and Wireless
590 Applications, Artech House Publishers, 2009.
591
592 30 J. Ardizzoni and D. Falls, A Practical Guide to High-Speed Printed
Circuit Board Layout, Analog
593 Devices Inc., 2005.
594
595 31 T. Glawdel and C. L. Ren, Global network design for robust operation
of microfluidic droplet
596 generators with pressure driven flow, Microfluid. Nanofluid.,
2012,13(3), 469-480.
597
598 32 J. Kim, A. J. deMello, S.-I. Chang, J. Hong and D. O'Hare, Thermoset
polyester droplet-based
599 microfluidic devices for high frequency generation, Lab Chip,
2011,11(23), 4108-4112.
600
601 33 B. Zheng, J. D. Tice and R. F. Ismagilov, Formation of droplets of
alternating composition in
602 microfluidic channels and applications to indexing of concentrations in
droplet based assays, Anal.
603 Chem., 2004,76(17), 4977-4982.
604
605 34 C. N. Baroud, F. Gallaire and R. Dangla, Dynamics of microfluidic
droplets, Lab Chip, 2010,
606 10(16), 2032-2045.
607
608 35 T. Glawdel, C. Elbuken and C. Ren, Passive droplet trafficking at
microfluidic junctions under
609 geometric and flow asymmetries, Lab Chip, 2011,11(22), 3774-3784.
610
611 36 T. Glawdel, C. Elbuken and C. L. Ren, Droplet formation in
microfluidic T-junction generators
612 operating in the transitional regime. I. Experimental observations,
Phys. Rev. E: Stat., Nonlinear,
613 Soft Matter Phys., 2012,85(1), 016322.
614
615 37 T. Mohamad, T. Hoang, M. Jelokhani-Niaraki and P. P. N. Rao, Tau-
derived-hexapeptide
616 VQIVYK aggregation inhibitors: nitrocatechol moiety as a pharmacophore
in drug design, ACS
617 Chem. Neurosci., 2013,12(4), 1559-1570.
618
619 38 S. N. Haydar, H. Yun, R. G. W. Staal and W. D. Hirst, Smallmolecule
protein-protein interaction
620 inhibitors as therapeutic agents for neurodegenerative diseases: recent
progress and future
621 directions, Annu. Rep. Med. Chem., 2009,44,51-69.
622
623
624
22
CA 2963807 2017-04-11
625 Detailed Description
626
627
628
629
630
631
632
633 .
634 Potential Application
635 The integrated microwave and microfluidics platform has application of
detection of E.Coli bacteria
636 in food and drinking water, detection and sensing in dairy product
industry, pharmaceutical drug
637 discovery and biomedical diagnosis.
638
639
640
641
642
643
644
645
646
23
=
