Inkjet-printing and Performance Evaluation of UHF RFID Tag Antennas on Renewable Materials with Porous Surfaces

Transcription

1 Inkjet-printing nd Performnce Evlution of UHF RFID Tg Antenns on Renewble Mterils with Porous Surfces Toni Björninen, Johnn Virkki, Juh Virtnen, Luri Sydänheimo, Leen Ukkonen Deprtment of Electronics nd Communictions Engineering Tmpere University of Technology Tmpere, Finlnd Mnos M. Tentzeris The School of Electricl nd Computer Engineering Georgi Institute of Technology Atlnt, GA, USA Abstrct Mteril choices in electronic devices hve huge impct on the environment. The use of environmentl-friendly processes nd renewble mterils is growing trend. We present the inkjet-printing of ntenns for pssive UHF RFID tgs on wood, pper, nd crdbord. We printed the prototype ntenns without surfce tretments. We evluted the performnce of the fully ssembled tgs through wireless tg mesurements. The fbricted tgs chieved 6-to-9.5 meters red rnge under the Europen RFID trnsmission regultions. Index Terms Antenn modeling, Inkjet-printing, Renewble mterils, RFID, Wireless tg mesurements, WSN I. INTRODUCTION Internet of things (IoT) is conceptul vision to connect everydy items nd devices in order to crete ubiquitous computing world. Things re expected to become ctive prticipnts in business, informtion, nd socil processes where they re enbled to interct nd communicte mong themselves, with humns, nd with the environment. IoT hs gret potentil e.g. in home utomtion, intelligent trnsporttion, nd helthcre [1]-[2]. Rdio-frequency identifiction (RFID), wireless sensor networks (WSN) nd mbient power hrvesting ply key role in the development towrd IoT, where omnipresent ntenns nd electronics will be semlessly nd unobtrusively integrted with items ll round us. In the future, mteril choices in wireless devices will hve huge effect on the environment. This is why the use of renewble, environmentl-friendly mterils nd dditive mnufcturing methods, such s inkjet-printing, is growing trend. Gret potentil lies in the use of wood nd pper s pltforms for ntenns used in rdio-frequency identifiction (RFID) tgs, wireless sensor networks (WSN) nd mbient power hrvesting [3]-[]. Importntly, the cpbility to print ntenns nd electronics directly on vrious surfces would remove the need for dditionl pltforms, hence improving the integrtion nd reducing the cost of rw mterils. This is importnt for bringing novel smrt products into mss production.integrtion nd reducing the cost of rw mterils. This is importnt for bringing novel smrt products into mss production. Renewble mterils, such s wood, pper, nd crdbord re typicl mterils especilly in construction nd pckging industry. Due to the porosity nd high surfce roughness, these mterils re chllenging pltforms for inkjet-printing. The ink droplets re esily bsorbed in the wood nd pper, preventing the nnoscle metlliztion prticles contined in the ink from forming conductive lyer. However, inkjet-printing of conductive ptterns directly without surfce tretments on these mterils, is mjor prcticl spect, s it will provide svings in the process cost nd time. We present the ntenn design for pssive UHF RFID tgs embedded in wood structures. Antenns were simulted nd then inkjet-printed on wood, pper nd crdbord, nd the performnce of the fully ssembled tgs ws evluted in terms of the mximum red rnge estimted bsed on wireless tg mesurements. II. MODELING OF A TAG ANTENNA FOR WOOD VENEER A full-wve electromgnetic solver, ANSYS HFSS v12, bsed on finite element method, ws used in ntenn modeling. In the simultion, the wood veneer (moisture content 12 %wt) ws modeled with the dielectric properties ε r =2.2 nd tnδ=0.1, which were estimted bsed on [5]-[6]. To chieve semless integrtion nd protection towrd environmentl stresses, the ntenn ws designed to be embedded in between two mm 3 pieces of wood veneer. This removes the need for dditionl structures. Bsed on the mesurements of ink lyer cross-sections nd ink dtsheet [7], the ntenn conductor ws modeled s uniform lyer with the thickness nd conductivity of 5 µm nd 25 MS/m, respectively. During testing, we found tht to chieve the mximl ntenn performnce, it ws best to lign the min rditing current of the ntenn long the grin of the wood. This provided the most relible print outcome. Becuse of the This reserch ws funded by Finnish Funding Agency for Technology nd Innovtion (TEKES), Acdemy of Finlnd, Centennil Foundtion of Finnish Technology Industries, nd Finnish Forest Foundtion / IEEE 1721

2 Dimensionl prmeters in millimeters. L w s t u v Figure 1. Tg ntenn for wood veneer. spreding of the ink, nrrow trces nd gps in between djcent trces were to be voided. These ppliction specific requirements were stisfied with dipole ntenn, which is simple single-lyer structure with well-known impednce mtching nd minituriztion techniques [8]-[9]. Compred with slot-type ntenn, which is nother cnonicl singlelyer ntenn, dipole requires less ink nd ws hence considered fit choice lso from the cost perspective. The prototype ntenn is shown in Fig. 1. It is stright dipole equipped with T-mtching loop (prms. u,v,t). The loop trnsforms the cpcitive impednce of the dipole to inductive. This mkes it possible to chieve complex conjugte impednce mtching between the ntenn nd the cpcitive tg IC [8]. The dditionl coupling between the dipole body (prms. L,w,s) nd sections of the mtching loop situted prllel to it llow for brodbnd mtching, similr to wht is chieved with the double T-mtching technique [9]. We used the built-in genetic optimizer in the modeling tool to choose the mtching loop prmeters. The overll gol in the optimiztion ws to chieve brodbnd nd power-efficient impednce mtching in nother words: to chieve high ntenn relized gin over brod rnge of frequencies. Power reflection coefficient is the rtio of the power vilble from the tg ntenn (P ) nd the power reflected bck (P r ) from the ntenn-chip interfce due to impednce mismtch. It is relted to ntenn nd IC impednces through Γ = Pr Z = P Z ic ic Z + Z * 2, (1) where Z nd Z ic re the ntenn nd IC impednces, respectively. Thus, the ntenn relized gin is given by G r ( θ φ ) = ( 1 Γ) D( θ, φ ) e, (2), r where D(θ,φ) is the ntenn directivity s function of the sptil observtion ngle nd e r is the ntenn rdition efficiency. To evlute equtions (1-2) we used the simulted ntenn properties nd the tg IC impednce model bsed on [10]. All the ssocited prmeters re shown in Fig. 2. Here the ntenn gin is defined: G(θ,φ) = D(θ,φ)e r. As summry, our simultions predicted firly constnt relized gin over the frequency rnge of MHz. The simulted rdition efficiency remined lso pproximtely constnt: 0.60 e r The brodside directivity ws pproximtely 2 dbi, which is typicl vlue for dipole ntenn. Importntly, we chieved brod 120 MHz (12%) mtched bndwidth (Γ 0.1). Figure. 2. Simultion results: ntenn-ic impednce mtching (top), rdition properties towrd the positive z-xis shown in Fig. 1 (bottom). III. ANTENNA FABRICATION A. Inkjet printing Inkjet printing is not just single technique, but collection of technologies iming t the sme trget: forming smll ink droplets nd guiding them onto the substrte in controlled wy. In inkjet technology, smll ink droplets (500 fl 2nl) re formed in the printhed nd ejected on the substrte. Inkjet printing llows the use of low viscosity inks, which is extremely importnt, since it llows for the formultion of inks tht only contin n ctive mteril nd solvents without the need for binders. In ddition, inkjet llows chep nd contct-free mnufcturing bsed on digitl imges. Although the digitl imges re redily swpped nd modified, nd no dditionl msks re required in the process, presently inkjet printing tends to be slow nd high throughput is only chieved by using lrge number of printheds in prllel. This hs introduced yield concerns relted to the misfiring of individul heds during printing of pttern. Moreover, severl process conditions must be ccurtely controlled for successful print outcome. Most importntly, the ink droplets must hve n pproprite shpe nd size when lnding on the substrte. The drop form nd velocity re controlled with the printhed prmeters (distnce to substrte, substrte temperture, jetting voltge nd voltge wveform). These prmeters need to be optimized for ech ink nd substrte. 1722

3 Inkjet utilizes liquid mterils, which sets some mteril requirements. The fluid should hve certin viscosity (typiclly vlues between 8 nd 25 P s) nd surfce tension vlues (should remin between N/m). Obviously, the requirements re lso ffected by the desired interction with the chosen substrte. Currently, silver nnoprticles re commonly used s conductive fillers in inkjet ink formultions. In the future, copper nd crbon bsed conductive inks my provide more ffordble lterntives [11]-[12]. B. Key prmeters of the inkjet-printing process We fbricted the ntenns on ll the studied pltforms with Fujifilm Dimtix DMP-2800 mterils printer [13] using with 10 pl printhed nozzles nd Hrim NPS-JL silver nnoprticle ink [7]. The key prmeters of the process re listed in Tble 1. We considered NPS-JL especilly suitble for creting conductive ptterns on renewble mterils, since it requires low sintering tempertures, 120 C minimum. Using lower sintering tempertures will result in slightly higher resistivity of the printed conductor, rnging between -6 µω cm. This is pproximtely three times higher thn the resistivity of bulk silver. In our work, we chose to use the temperture of 150 C to mximize the conductivity of the ink lyer [7]. TABLE I. Prmeter Jetting voltge Jetting frequency Crtridge temperture Plten temperture Drop volume Printing resolution KEY PARAMETERS OF THE INKJET-PRINTING PROCESS. Vlue 2 V 9 khz 0 C 60 C 10 pl 23 dpi (60 µm drop spcing) C. Printing nd curing of the tgs The substrte mterils used in this reserch were birch veneer, norml pckging crdbord, nd silk surfce rt pper. During testing, we found tht to chieve the mximl ntenn performnce on wood, it ws best to lign the min rditing current of the ntenn long the grin of the wood. Thus, we ligned the dipole rms long the wood grin (Fig. 3b). Moreover, we found tht to sturte the wood grin with ink, it ws best to first print the five lyers of ink with the printhed moving in the direction of the wood grin. After this, five more ink lyers were dded with the printhed moving ginst the wood grin. Following the ink dtsheet instructions, we sintered the smples t 150 C for to mximize the conductivity of the printed lyer [7]. After sintering, we repeted the bove described process second time, to ensure sufficient conductor thickness for the ntenns. Hence, we formed conductor with totl 20 ink lyers with every five subsequent lyers deposited with the printhed moving either long or ginst the grin of the wood. Unlike the wood veneer, the pper nd crdbord pltforms did not hve visible orienttion dependent surfce fetures. Thus, on these pltforms, we printed the ntenns with the printhed moving longitudinlly long the dipole. To gin insight on the bsorption of the ink into the crdbord, we Figure 3. Inkjet-printed ntenns with ICs on A) pper, B) wood, nd C) crdbord. printed the tgs on this pltform by using two different pproches. In the first pproch, 20 lyers of ink were deposited before sintering t 150 C for. In the second pproch, the sme sintering conditions were used, but printing ws done in two stges. First, 10 lyers of ink were deposited before the first sintering, fter which the process ws repeted second time. On the pper pltform, we printed 15 lyers of ink with the printhed moving longitudinlly long the dipole followed by sintering t 150 C for. After printing nd curing, Alien Higgs-3 RFID ICs (tgs printed on wood) nd NXP RFID ICs (tgs printed on pper nd crdbord) were ttched to the smples using conductive silver epoxy resin. Both chips hve the wke-up power of 18 dbm. Fig. 3 shows the fully ssembled prototype tgs. TABLE II. KEY PARAMETERS OF THE INKJET-PRINTING PROCESS. Substrte Printing Sintering Crdbord Crdbord Pper Wood IV. 10 lyers in the sme direction, sintering (x2) 20 lyers in the sme direction, sintering (x1) 15 lyers in the sme direction, sintering (x1) 5 lyers in the direction of the grin, 5 lyers ginst the grin, sintering (x2) TAG PERFORMANCE EVALUATION We conducted wireless mesurement on the prototype tgs to evlute their performnce in terms of the theoreticl red rnge. In the mesurement, we used n RFID mesurement system [1] which llowed us to monitor the reply from the tg under test to ISO C query commnd while sweeping the trnsmitted crrier power nd the trnsmission frequency. Using the clibrtion tg included in the system hrdwre, we 1723

4 Threshold power [dbm] Forwrd link ttenution [db] Pper (15x, NXP) Crdbord (20x, NXP) Crdbord (10x+10x, NXP) Wood (10x+10x, H3) -25 Theoreticl red rnge [m] Figure. Mesurement results. lso mesured the pth loss in the wireless mesurement chnnel. The pth loss is defined s the power loss fctor (L iso ) from the trnsmitter s output port to the input port of hypothetic polriztion-mtched isotropic ntenn plced t the test loction. To chrcterize the fbricted tgs, we recorded the threshold power (P th ), the minimum reder output power which enbled vlid response from ech tg. Bsed on the mesured dt, we computed theoreticl red rnge from eqution [9] d tg λ EIRP =, (3) π P L th iso where EIRP is the regulted isotropiclly rdited power following the Europen regultion: EIRP = 3.28 W. In prctice, the theoreticl red rnge is ttined in n nechoic environment in the min bem of reder ntenn which is polriztion-mtched with the tg ntenn. Reporting the red rnge in these reference conditions provides site-independent benchmrk for reserchers worldwide. As seen from Fig., the mesured threshold power of the tg optimized for embedding in wood structures is the lowest round 900 MHz. This is the frequency where the fbricted tg ttined its pek relized gin nd the corresponding tg red rnge of 8.5 meters. Moreover, the threshold power remined firly constnt within the frequency rnge of MHz with the corresponding tg red over seven meters. This confirmed the brodbnd response predicted by the simultions, but reveled frequency shift of pproximtely 0 MHz towrd higher frequencies. This is likely due to the fct tht to fcilitte the testing, we did not embed the fbricted ntenn in between veneer lyers s in the simultion. The link ttenution is pproximtely the sme for ll tgs, since the mesurement chnnel ws the sme for ll mesured tgs. The ttenution loss incresed fter 950 MHz due to the limited bndwidth of the trnsmit ntenn. We optimized the printed ntenn for tg embedded inside wood structures. However, we lso fbricted the ntenn on crdbord nd pper for the purpose of further experimentl performnce evlution. These dditionl tgs were lso equipped with different tg IC. The chip hs, however, similr impednce nd the sme wke-up power s the IC on the wood tg. Therefore, no mjor performnce degrdtions were expected. As seen from Fig., the pper tg ctully chieved the highest red rnge (9.5 meters) mong the studied tgs. Beyond 970 MHz, the ntenn printed on crdbord with sintering fter 10 nd 20 ink lyers provided the sme performnce s the wood tg. Hence, we believe tht with further pltform- nd IC-specific optimiztion it is possible to chieve eqully good performnce on crdbord nd pper s well. 172

5 V. CONCLUSION Fbriction of ntenns nd electronics directly on renewble, environmentl-friendly mterils, such s wood nd pper-bsed mterils is impertive for the development of future wireless pltforms. We presented pssive UHF RFID tgs with ntenns inkjet-printed on wood, pper, nd crdbord without ny surfce tretments. We provided guidelines for chieving relible print outcome on the rough nd porous surfces. Our mesurement results showed tht the fbricted prototype tgs on ll the studied pltforms chieved 6-to-8 meter red rnges under the Europen RFID trnsmission regultions. This performnce is without doubt sufficient for mny future pplictions. Future work will be the optimiztion the mount of the ink used s well s sintering times. We will lso investigte the inkjet printing of copper inks s n lterntive to silver inks for potentil cost reduction. REFERENCES [1] Commission of the Europen Communities, Internet of Things An Action Pln for Europe, policy/rfid/documents/commiot2009.pdf [2] G. Kortuem, F. Kwsr, D. Fitton, V Sundrmoorthy, Smrt objects s building blocks for the Internet of things, IEEE Internet Computing, vol. 1, no. 1, pp. -51, Jn.-Feb [3] G. Shker, S. Sfvi-Neini, M. M. Tentzeris, Inkjet printing of ultrwidebnd (UWB) ntenns on pper-bsed substrtes, IEEE Antenns Wireless Propg. Lett., vol. 10, no. 1., pp , Dec [] L. Yng, A. Rid, R. Vys, M. M. Tentzeris RFID tg nd RF structures on pper substrte using inkjet-printing technology, IEEE Trns. Microw. Theory Techn., vol. 55, no. 12, pp , Dec [5] A. Koub, P. Perr, R. M. Hutcheon, J. Lessrd, Complex dielectric properties of the spwoods of spen, White Birch,Yellow Birch nd Sugr Mple, Drying Technology, vol. 26, no. 5, pp , My 1, [6] M. Tbssum, B. Colpitts, K. Glik, Dielectric Properties of Softwood Species Mesured with n Open-ended Coxil Probe, 8 th Intl.IUFRO Wood Drying Conf., pp , 2-29 Aug. 2003, Brsov, Romni. [7] Hrim Chemicls Group, Inc., Jpn: NPS-JL nnopste: [8] G. Mrrocco, The rt of UHF RFID ntenn design: impedncemtching nd size-reduction techniques, IEEE Antenns Propg. Mg. vol. 50, no. 1, pp , Feb [9] T. Björninen, A. Z. Elsherbeni, L. Ukkonen, Performnce of single nd double T-mtched short dipole tg ntenns for UHF RFID systems, J. Appl. Computtionl Electromgn. Soc., vol. 26, no. 12, pp , Dec [10] T. Björninen, M. Luri, L. Ukkonen, R. Ritl, A. Z. Elsherbeni, L. Sydänheimo, Wireless mesurement of RFID IC impednce, IEEE Trns. Instrum. Mes., vol. 60, no. 9, pp , Sep [11] K. Woo, Y. Kim, B. Lee, J. Kim, J. Moon Effect of crboxylic cid on sintering of inkjet-printed copper nnoprticulte films, ACS Appl. Mter. Interfces, vol. 3, no. 7, pp , My [12] L. Hung, Y. Hung, J. Ling, X. Wn, Y. Chen, Grphene-bsed conducting inks for direct inkjet printing of flexible conductive ptterns nd their pplictions in electric circuits nd chemicl sensors, Nno Reserch, vol., no. 7, pp , Jul [13] Fujifilm Dimtix, Inc., Snt Clr, CA, USA: [1] Voyntic, Ltd., Espoo, Finlnd: