paper PDF - Washington University in St. Louis - Computer Science

Self-powered Wireless Biosensing based on Integration of Paper-based Microfluidics with Self-assembling RFID Antennas Mingquan Yuan∗ , Evangelyn C. Alocilja† and Shantanu Chakrabartty∗ ∗ Department

of Computer Science and Engineering Washington University in St. Louis, St. Louis, Missouri USA 63130 † Department of Biosystems and Agricultural Engineering Michigan State University, East Lansing, Michigan USA 48910 Email: [email protected]

Abstract—This paper extends our previous work on wireless biosensing by proposing and demonstrating the integration of self-assembling radio-frequency antennas with paper-based microfluidics. The integration substrate is constructed using polyethylene and the patterning of the antenna on the substrate has been achieved using a low-cost ink-jet printing technique. The use of paper-based microfluidics enables self-powered sample acquisition, sample mixing and sample flow to areas on the substrate where antennas can self-assemble only when target analytes are present in the sample. When the integrated substrate is combined with a passive radio-frequency identification (RFID) tagging technology, the resulting sensor-tag can be used for continuous monitoring in a food supply-chain where direct measurement is considered to be impractical and reducing false alarms is a key consideration. We validate the proof-of-concept operation of the proposed sensor-tag using IgG as a model analyte.

I. I NTRODUCTION End-to-end monitoring of a food supply chain is one of the keys towards preventing food-borne disease outbreaks and product recalls like the 2015 Listeria outbreak due to contaminated ice-cream that resulted in 3 fatalities [1]. Fortunately, two converging economic trends have now made the vision of end-to-end supply-chain monitoring a reality. The first trend is that over the last decade the price of passive RFID tags have reduced by orders of magnitude when compared to the cost of packaging or materials [2], [3]. As a result, it is now economically viable to embed or attach a passive tag to every package of food-item (as shown in Fig. 1). The second trend, also illustrated in Fig. 1, is that the new generation of smartphones are equipped with the capability to read RFID tags. Given the rapid penetration of smart-phones in the consumer market, the tags can be interrogated at different segments of the supply-chain, starting from the food-source all the way to the market shelves. Thus, sensors integrated with passive RFID tags provide an attractive technology to continuously and wirelessly monitor the quality of a food-product in a supplychain where direct measurement is not considered practical. Also, practical supply-chain requirements require high degree of reliability in detection when these passive sensor-tags have to operate under real-world operating conditions.

978-1-4799-7234-0/15/$31.00 ©2015 IEEE

RFID writer/reader

L

Milk Sensor RFID tag

Fig. 1. Illustration of end-to-end monitoring of food quality in a supply-chain where smart-phone readers and sensors embedded in RFID tags could be used for inspection.

Unfortunately, most passive RFID based biosensors that have been reported in literature operate on a unifying principle, which is to measure the deterioration in the tag’s RF reflection properties using a remote reader. The concept is illustrated in Fig. 2(a) where the deterioration manifests itself as a shift in the center frequency and the reduction in the quality factor of the tag antenna. This principle has been used for designing RFID based biosensor reported in [4] for monitoring the freshness of milk or fish by detecting changes in milk’s dielectric constant or by changes in odor volatilities of fish. A similar principle was used in [5] to detect biogenic amine putrescine, commonly used as a marker for food spoilage. A molecular imprinted polymer (MIP) based passive RFID sensor reported in [6] used the deterioration concept for detecting histamines in spoiled fish. However, this approach for RFID biosensing is prone to high false-positives, as several environmental factors could degrade and detune the tag antenna as well. In [7] and [8] we had presented a novel approach of RFID biosensing based

f1

f0

(a) “Low entropy” to “high entropy”

S-parameters

S-parameters

f2

gold nanoparticle

f0

electrodes

aIgG

e-

e-

V

+

I

V

-

silver

f1 f2

(b) “High entropy” to “low entropy”

Fig. 2. Two different RFID sensing approaches. (a) Traditional approach based on detuning of the antenna; (b) Proposed approach where the antenna tunes or assembles itself when the target is detected.

on concept of “growing” antennas which is triggered only when target analytes are present in the sample. This approach of sensing can be viewed as a process that evolves from a high-entropy state (disassembled detuned antenna) to a low-entropy state (assembled tuned antenna), as shown in Fig. 2(b) and therefore requires an influx of energy from the environment. Thus, compared to the traditional detuning approach, the self-assembly based approach should be more robust to environmental artifacts and hence produce lower false-positives. However, the challenge for the self-assembly based method lies in controlling the process of antenna growth on the passive RFID tag, given that there are no continuous sources of power. Addressing some of these challenges is one of the goals of this paper. First, we show that paper-based microfluidics can be used for low-cost, self-powered sample acquisition, sample flow and sample mixing. Second, we show the integration of microfluidics on antennas printed on a plastic substrate using a low-cost ink-jet technology. Third we verify and demonstrate the proof-of-concept detection by integrating the substrate with a commercial RFID tag compliant with the Gen-2 ultra-highfrequency (UHF) standard. II. A NTENNA S ELF - ASSEMBLY BASED ON S ILVER - ENHANCEMENT The proposed biosensor and the self-assembly of the RF antenna is based on our previous reported silver enhancement technique [7], [8]. The principle is illustrated in Fig. 3 using a pair of silver electrodes that are separated by a gap. Silver-enhancement requires conjugation of gold nanoparticles (AuNPs) inside a matrix or on bio-receptors (for instance antibodies shown in Fig. 3) which need to be immobilized between the two electrodes. Since the dimensions of the AuNPs are in nanometers (∼10nm in this paper), they are not large enough to electrically bridge the separated electrodes (Fig. 3(a)). When silver-enhancement solution comprising of Ag ions (I) and hydroquinone (photographic developing solution) is applied in between the electrodes silver ions start reducing into metallic silver on the surface of the gold nanoparticles. The procedure is completely self-powered and does not require any external biasing. During this reaction, gold nanoparticle works as a catalyst and facilitates further reduction of silver ions. As more

5um

(a) Open circuit

(b) Subthreshold

(c) Short circuit

Fig. 3. Three stages of the silver-enhancement process: (a) isolation mode: before the silver enhancement solution applied, the electrodes are electrically isolated; (b) subthreshold mode: during the silver enhancement process, the silver ions get reduced to form metallic silver; (c) above-threshold mode: the path between the two electrodes are electrically bridged and more silver ions are reduced.

silver ions are reduced, a chain of gold nanoparticle cored silver micro-monopole antennas self-assemble in between the electrodes as shown in Fig. 3(b). In this situation, electrons can hop between the two electrodes since the growth of the silver enhanced particles provide shorter paths compared to that shown in Fig. 3(a). With the progression of time and in the presence of more analytes, more silver ions get reduced and in the limit the chain of micro-monopole antennas gets completely bridged as shown in Fig. 3(c). Electrons can now freely flow when a potential difference is applied across the two electrodes. Experimental results show that the change in conductance after silver enhancement process is monotonic with respect to the concentration of the target analyte [9], which is omitted for the sake of brevity in this paper. III. S ENSOR I NTEGRATION AND P RINCIPLE The working principle of the proposed integrated RFID biosensor is shown in Fig. 4. A dipole antenna is ink-jet printed on the mesoporous printing substrate media using silver ink, as shown in Fig. 4(a). To avoid confusion, the side of the substrate where antenna is printed is referred to as the “front side” and the other side (which does not contain printed antenna) is referred to as the “rear side”. The use of ink-jet printing allows a more precise control over the gap and aperture size( lengths less than 100μm ) where silver-enhancement could occur. This provides improvement over the previous methods [7], [8] that used razors to manually create gaps. Two small apertures are created on the plastic printing substrate on two sides of the gap (labeled as “SE pass aperture” in Fig. 4(b)). These apertures provide a path for the silver-enhancement solution to flow through from the rear side of the substrate. The antenna gap is then covered by a nitrocellulose (NC) membrane (“sample pad 1”), with the nitrocellulose side facing the gap on the front side. The pores in the NC membrane allows the liquid sample to flow using capillary force. The “sample pad 2” and “adsorption pad”, are attached on the rear side of the substrate along with the nitrocellulose side which faces the “SE pass apertures”, as shown in Fig. 4. These pads act as reservoirs that control the flow of the liquid through the NC channel.

Adsorption Sample pad 2 pad IC chip SE pass aperture

Sample pad 1

substrate

Printed antenna

(a) Schematic design of the microfluidic RFID biosensor

Adsorption pad

SE pass aperture

Sample pad 2

Gap

(b) Enlarged dash box area in Fig. 4(a) Fig. 4. Assembly and integration of microfluidic channels within the RFID biosensor.

Target specific antibodies (in this case anti-IgG) are then immobilized with the NC membrane gap. Similar to the operation of an lateral-flow immunoassay [10], the target analyte (in this case IgG) first conjugates with the gold nanoparticle labeled anti-IgG (aIgG) to form a partial sandwich (IgG-aIgGAuNP) structure. When the conjugate is applied to the sample pad 1, it flows through the NC membrane due to a capillary force. Due to antibody-antigen hybridization, a complete sandwich structure (aIgG-IgG-aIgG-AuNP) is formed when the partial sandwich structures (IgG-aIgG-AuNP) flow into the spot where aIgG is immobilized. Any unbound aIgGAuNP are washed away due to the capillary force in the NC membrane. The state of the antenna gap region after the sandwich formation is similar to that shown in Fig. 3(a), where the gap remains electrically insulated because the size of the gold nanoparticles is not sufficient to electrically bridge the separated segments of the antenna. When silver enhancement solution is applied to the sample pad 2, it moves through the NC channel to the adsorption pad, through SE pass aperture and sample pad 1. Once the sandwich structure is exposed to a silver enhancement solution (initiator and enhancer mixed with volume ratio of 1:1), silver enhancement process occurs. As time progresses, more silver ions get reduced on the gold nanoparticles, as a result of which the size of the shell grows and ultimately electrically bridges the antenna gap, as shown in Fig. 3(c). Assembly of the antenna within the gap tunes the antenna which in turn enhances its reflection properties. IV. M ATERIALS AND M ETHODS Silver Enhancement Kit was obtained from Sigma-Aldrich (St. Louis, MO, USA). Anti-rabbit IgG (whole molecule) conjugated with gold nanoparticles, anti-rabbit IgG and IgG were all purchased from Sigma-Aldrich (St. Louis, MO, USA). Glutaraldehyde and Methanol were also obtained from

Sigma-Aldrich (St. Louis, MO, USA). Nitrocellulose (NC) membrane with flow rate of 135 sec/4cm was purchased from Millipore (Billerica, MA, USA). Deionized (DI) water used in the experiment was obtained through Millipore water purification systems (Billerica, MA, USA). EPC Gen 2 ALN9640 tags were from Alien Technology (San Jose, CA, USA). An EPSON stylus C88+ ink-jet printer was used to print the antenna. The printing substrate and the JS-B25P silver ink were purchased from Novacentrix (Austin, TX, USA). Scanfob Ultra-BB2 Wireless GEN2 UHF RFID Reader/Writer (Cedar Park, TX, USA) was used for remote interrogation and measurements. All the experiments were carried out in a certified Biological Safety Level II laboratory. Layout of the dipole antenna was optimized using finiteelement simulation to ensure that the feed to the antenna is impedance matched to the ALN-9640 RFID tag chip. The tag was printed using EPSON C88+ printer and JS-B25P silver conductive ink on a mesoporous printing substrate. After antennas have been printed the substrate is annealed in the oven at the temperature of 75◦ C for 1 hour. The annealing procedure creates a uniform conductance across the antenna structure. The ALN-9640 RFID tag chip was removed from a commercial tag and was attached to the ink-jet printed antenna using a tape, as shown in Fig. 5(a). The NC membrane was cut into different shapes and sizes to form the sample pad 1, sample pad 2 and adsorption pad. Using a procedure described in [8] anti-IgG was immobilized on the nitrocellulose surface of sample pad 1 facing the antenna. The sample pad 1 was then attached to the antenna using a tape. Two small SE pass apertures (dimension around 1mm×2mm) were created on the mesoporous substrate on two sides of the gap as shown in Fig. 5(a). These two apertures are created to provide a path for the silver enhancement solution to flow from the sample pad 2 which was attached on the rear side of the substrate. The sample pad 2 and the adsorption pad were attached to antenna such that their end-points covered the two SE pass apertures as shown in Fig. 5(b). The sequential operation of sample processing can be controlled by the adjusting the length of the flow channel and the pore size of the NC membrane. For instance, silver-enhancement procedure requires mixing of the initiator and the enhancer solution in the volume ratio of 1:1. The silver-enhancement solution then flows to the end-point of the sample pad 2 (beneath the SE pass aperture) which is then drawn by the sample pad 1 through SE pass aperture and then by the adsorption pad if sufficient reagent is provided. During this process, the aIgG-IgG-aIgG-AuNP sandwich structure is exposed to the silver-enhancement solution. Post silver-enhancement, the response of the tag was measured using the 915MHz Scanfob Ultra-BB2 reader. A schematic of a complete RFID based detection and measurement system which includes a bluetooth compatible 915MHz RFID writer/reader (Scanfob Ultra-BB2) that interfaces with a laptop through cable or a smartphone through bluetooth for data analysis and display is shown in Fig. 1.

sample pad 1

SE pass aperture spot

IC chip Printed antenna (a) Front side of the sensor prototype

rear

gap spot

0.7

Normalized detection distance

front

0.6 0.5 0.4 0.3 0.2 0.1 0.0

sample pad 2 adsorption pad (b) Rear side of the sensor prototype Fig. 5. Photos of the front and rear side of the RFID biosensor prototype.

V. M EASURED R ESULTS AND D ISCUSSIONS For calibration purposes an unmodified 915MHz dipole antenna (identical structure as the biosensor dipole but without gap) was also integrated with the biosensor. All interrogation measurements from the biosensors (L1 ) were normalized with respect to the measurement obtained from the calibration dipole (L2 ). To minimize the mutual loading the sensor tag and the calibration tag were placed next to each other and a ratiometric readout technique was used [8]. Fig. 6 shows the preliminary results which compares the normalized detection range under two conditions: when the target is absent and silver-enhancement does not occur; when the target is present and silver-enhancement grows the antenna. For the first case, the normalized measurement is only 28.6% with respect to that of the calibration antenna. Whereas as shown in Fig. 6, post silver-enhancement, the normalized maximum interrogation range increases to 61.8% with respect to that of the calibration antenna. It validates the proof-ofconcept biosensing based on integration of paper microfluidics with self-assembled RFID antennas. VI. C ONCLUSION In this paper, we have extended our previous work in the area of RFID biosensing by integrating paper-based microfluidics with self-assembling RFID antennas. The sensing procedure relies on a process that evolves from an unassembled antenna or a high-entropy state to an assembled antenna or a low-entropy state. Because the direction of this evolution is opposite to processes that deteriorate the quality of the antenna, this method could potentially produce lower falsepositives. The integrated RFID biosensor prototype presented in this paper represents just a preliminary prototype for proofof-concept demonstration. Future work in this area will require

before silver enhancement after silver enhancement

before silver enhancement

after silver enhancement

Fig. 6. Measured results validating the proof-of-concept RFID biosensor.

optimization of the reagent flow speed control (can be realized by choosing membranes with proper flow-rate) and pad shape uniformity (can be realized using laser to cut instead of manually cut) to facilitate the sequential processing of sample hybridization, silver-enhancement and antenna growth. R EFERENCES [1] Multistate Outbreak of Listeriosis Linked to Blue Bell Creameries Products. [Online]. Available: http://www.cdc.gov/listeria/outbreaks/icecream-03-15/ [2] X. Zhu, S. K. Mukhopadhyay, and H. Kurata, “A review of RFID technology and its managerial applications in different industries,” Journal of Engineering and Technology Management, vol. 29, no. 1, pp. 152–167, 2012. [3] C. Huang and S. Chakrabartty, “An asynchronous analog self-powered CMOS sensor-data-logger with a 13.56 MHz RF programming interface,” Solid-State Circuits, IEEE Journal of, vol. 47, no. 2, pp. 476–489, 2012. [4] R. A. Potyrailo, N. Nagraj, Z. Tang, F. J. Mondello, C. Surman, and W. Morris, “Battery-free radio frequency identification (RFID) sensors for food quality and safety,” Journal of agricultural and food chemistry, vol. 60, no. 35, pp. 8535–8543, 2012. [5] N. R. Tanguy, L. K. Fiddes, and N. Yan, “Enhanced radio frequency biosensor for food quality detection using functionalized carbon nanofillers,” ACS Applied Materials & Interfaces, vol. 7, no. 22, pp. 11 939–11 947, 2015. [6] D. Croux, T. Vangerven, J. Broeders, J. Boutsen, M. Peeters, S. Duchateau, T. Cleij, W. Deferme, P. Wagner, R. Thoelen et al., “Molecular imprinted polymer films on RFID tags: a first step towards disposable packaging sensors,” physica status solidi (a), vol. 210, no. 5, pp. 938–944, 2013. [7] M. Yuan, E. Alocilja, and S. Chakrabartty, “A novel biosensor based on silver-enhanced self-assembled radio-frequency antennas,” Sensors Journal, IEEE, vol. 14, no. 4, pp. 941–942, April 2014. [8] M. Yuan, P. Chahal, E. Alocilja, and S. Chakrabartty, “Wireless Biosensing Using Silver-Enhancement Based Self-Assembled Antennas in Passive Radio Frequency Identification (RFID) Tags,” Sensors Journal, IEEE, vol. 15, no. 8, pp. 4442–4450, Aug 2015. [9] Y. Liu, D. Zhang, E. C. Alocilja, and S. Chakrabartty, “Biomolecules detection using a silver-enhanced gold nanoparticle-based biochip,” Nanoscale research letters, vol. 5, no. 3, pp. 533–538, 2010. [10] T. C. Tisone and B. OFarrell, Manufacturing the next generation of highly sensitive and reproducible lateral flow immunoassay. Springer, 2009.