International Workshop

SMART MATERIALS, STRUCTURES & NDT in AEROSPACE

Conference

NDT in Canada 2011 2 - 4 November 2011, Montreal, Quebec, Canada

EFFECTS OF TEMPERATURE AND HUMIDITY ON UHF RFID PERFORMANCE

Li Zhenzhong1,3, Mrad Nezih2*, Xiao George3, Ono Yuu4, Liu Guocheng1,3, Ban Dayan1
1

Department of Electric and computer Engineering, University of Waterloo Air Vehicles Research Section (AVRS), Defence R&D Canada (DRDC), Department of National Defence (DND), National Defence Headquarters Tel: (613) 993-6443; Nezih.Mrad@drdc-rddc.gc.ca

Institute for Microstructural Sciences (IMS), National Research Council of Canada(NRC)

4

Department of Systems and Computer Engineering, Carleton University ABSTRACT

Radio Frequency IDentification (RFID) technology is contributing to increased manufacturing efficiency, improved supply chain operations, and enhanced logistics. The characteristic associated with RFID technology is currently being exploited by several industries for sensing applications as an alternative to current power inefficient wireless approaches in both fields of structural health monitoring and environmental monitoring. Many environmental challenges continue to persist impeding substantial growth and accelerated implementation of the technology in these fields. This work presents an investigation of the performance of a passive Ultra-High-Frequency (UHF) RFID based system under changing environmental conditions to facilitate increased technology exploitation. The investigation addresses the impact of temperature and humidity variations on the system response (e.g. changes in resonant frequency). Experimental results demonstrate a linear relationship between passive UHF RFID tags resonant frequency and temperature at relative humidity (RH) of 50% and 80%. No significant change in the resonant frequency was observed as result of humidity variation at temperature ranging from 20oC to 80oC. Results suggest that the frequency and return loss are independent of each other and are good indicators for temperature and humidity monitoring. Keywords: RFID, radio frequency, UHF, humidity, temperature, sensing

2011 CANSMART CINDE IZFP

INTRODUCTION Radio Frequency technology has served as an identification technology for more than 50 years. After the development of Ultra-High-Frequency (UHF) radio frequency identification (RFID) technology by IBM in 1990s, and after the RFID system was adopted by many large companies and organizations in early 2000s, the RFID technology has attracted much attention from several sectors of the industry. RFID tags can be made in different sizes and shapes and can be adhered or secured to several items with different degree of complexity. Over the years, the cost of each tag has been reduced to about 10-15 cents per tag, for large quantities of labeling tags [1]. The technology further offers indirect and no line-of-sight detection of tagged articles with simultaneous multiple tags detection, hence reducing check-in and check-out time in stores, airports, hospitals and improving asset management efficiency. After the adoption of this technology by several big early adapters, such as United State Department of Defence (DoD) and Wal-mart [2], the technology is now widely used in asset tracking and management among other applications. The simplicity and low cost associated with the implementation of RFID technology extends its potential to other areas, including passive wireless sensing and wireless power transmission. Near field wireless power transmission has been realized and utilized to charge cellular phone batteries [3]. The application of RFID technology has also been extended into new areas beyond the supply chain, from tracking the movement of goods to monitoring their health state while in transit [4-7]. With selective coatings, HighFrequency (HF) RFID tags can be used to detect a variety of contaminates in the environment through the monitoring of changes in the tags resonant frequency [8] or its capacitance [9]. On the other hand, UHF passive tags are still considered as transponders for other sensors for the purpose of wireless communication [10]. This works focus is to assess the potential use of UHF passive RFID tags for humidity and temperature sensing. Our previous work [11-12] showed that position, distance, orientation and power variation has affects on the tags readability and detectability. EXPERIMENT Experimental Setup One thin (paper like) 915 MHz UHF tag (98.2 x 12.3 mm) and one 840~960 MHz flat antenna (220 x 220 x 30 mm) were used [6] (Fig. 1) in this experimental assessment. An ESPEC (ESX-3CA) environmental chamber was used for humidity and temperature control, and an Anristu network analyzer (VNA-MS2026C) was used for data collection and analysis (Fig. 1(a)). Inside the environmental chamber (Fig. 1(b)), the antenna was placed at a fixed 10 cm away from the tag. The antenna was connected to the network analyzer which was placed outside the environmental chamber. The environmental chambers operating conditions are shown in Fig.2.

2011 CANSMART CINDE IZFP

Antenna

VNA Tag Chamber (a) (b)

Fig. 1 : Illustration of experimental setup; (a) environmental chamber and network analyzer; (b) tag and antenna.

Fig. 2: Operating condition of the environmental chamber. Experimental Measurements In this performance assessment work the tag-antenna system was subjected to a varying temperature and humidity environment. In the temperature variation experiment, the relative humidity (RH) was fixed, independently, at 50% and 80%, while the temperature increased from 20oC to 80oC at increments of 10oC. Similarly, at the same humidity levels, the temperature was ramped down at the same increments of 10oC from 80 oC to 20oC. To assess the repeatability of the results the tests were repeated twice. In the humidity variation experiment, the temperature was fixed, independently, at 50 oC and 80 oC, while the RH varied from 20% to 80% and from 80% to 20% at increments of 10%. Similar to the temperature experiments, and to assess the repeatability of the results the tests were repeated twice. In both humidity and temperature tests, the environmental chamber was considered as an enclosure having metal sidewalls. The adverse effects of metal on radio frequency propagation are well know and are presented in the open literature [13-14]. To discern the metal effect and to project a more realistic environment, additional data was taken with the environmental chamber door open, at the above presented conditions. The door was

2011 CANSMART CINDE IZFP

kept open for few seconds at each test point. For example, after taking data at 50 oC and 50% RH, the door of the chamber was opened, for few seconds, to allow for signal propagation and data collection for the particular condition. This process reduces the interaction of the conductive enclosure on the RF signal propagation. RESULTS AND DISCUSSION In this investigation, analysis has mainly focused on the passive UHF RFID tags resonant frequency variation as a result of environmental changes. Selected tags frequency characteristics (Fig. 3) were monitored. These characteristics include resonant frequency, imaginary impedance valley and peak frequency and real impedance peak frequency. The resonant frequency is denoted by fo (Fig. 3(a)), the valley and peak frequencies of the imaginary impedance are denoted by F1 valley and F1 peak (Fig. 3(b)); whereas, the peak frequency of the real impedance measurement is denoted by F2 peak.

(a)

(b)

(c)

Fig. 3: Frequency measurements; (a) tags resonant frequency; (b) valley and peak frequency of the imaginary impedance measurement; (c) peak frequency of the real impedance measurement. Temperature Variation Effects In varying the temperature while holding the RH constant, a change in the resonant frequency of the RFID tag was observed, as shown in Fig. 4. In both cases where the environmental chamber door was open and closed, and at 80% RH, a linear relationship was obtained between the resonant frequency and the temperature. In the experimental setup where the door of the environmental chamber was closed, and at 80% RH, the frequency decreased from 912.727 MHz to 909.425 MHz with a rate (slope of the trend line) of 55.05 kHz/oC while the temperature increased from 20oC to 80 oC (Fig. 4(a)); when the door was open, the decrease of frequency was more pronounced and varied from 912.927 MHz to 907.524 MHz with a rate of 90.05 kHz/oC (Fig. 4(b)). The rate of frequency decrease is higher when the chamber door was open. This is a result of the effect of the metal enclosure on the UHF response. Although the heat transfer from the chamber to its outside, when the door was opened, might have certain effect on the frequency shift, the heat loss would result in lower temperature indicative of lesser resonant frequency shift. Additionally, the rate of decrease between the temperature ramp up and down for both closed and open chamber door, is found to be

2011 CANSMART CINDE IZFP

57.1 kHz/oC and 82.8 kHz/oC respectively. This difference between temperature ramp up and down is believed to be due humidity variation. This difference is practically insignificant in the open door case. It is noted that the rates of increase or decrease are obtained as an averages between the ramp up and ramp down. At 50% RH, similar frequency decreases were also observed (Fig. 4(c)). A rate of decrease of 25.5 kHz/oC for closed chamber door and of 25.8 kHz/oC for open chamber door, were obtained, respectively. Similar to the 80% RH case, the higher rate of decrease is associated with the open door configuration, indicating the similar effects of the metallic enclosure on the RF signal propagation. Additionally, at this lower level of humidity, the rate of change was slower than at higher humidity levels. Such higher frequency variation at higher humidity level is also observed in the humidity experiment (next section). This change can be attributed to the change of antenna properties, including both resistance and capacitance. Larger variation on temperature may result higher change on the antenna resistance and capacitance, which could alter the resonant frequency further away from its nominal value.
913.5 913 912.5 912 911.5 911 910.5 910 909.5 909 10 30 Up Down Trend line (Up) Trend Line (Down)

914 913
Frequency (MHz)

Frequency (MHz)

912 911 910 909 908 907

Up Down Trend line (Up) Trend Line (Down)

50 70 Temperature (oC)

90

10

30

50 70 Temperature (oC)

90

(a)
914 913 912

(b)
RH-50 open RH-50 close RH-80 open RH-80 close open close

Frequency (MHz)

911 910 909 908 907 906 905 10 20 30 40 50 60 Temperature (o C) 70

80

90

(c)

Fig. 4: The resonant frequency shift as a result of temperature variation; (a) chamber door closed (at 80% RH); (b) chamber door open (at 80% RH); (c) comparison between 50% RH and 80% RH.

2011 CANSMART CINDE IZFP

Humidity Variation Effects In varying the humidity while holding the temperature constant, a change in the resonant frequency of the RFID tag was observed, as shown in Fig. 5. In both cases where the environmental chamber door was open and closed, and at 80 oC, a linear relationship was obtained between the resonant frequency and relative humidity. In this experiment, both the humidity ramp up and down are provided at 80oC but only the ramp up was recorded at 50oC, unlike the case of temperature variation where both temperature ramp up and down were provided for both cases in Fig. 4.
910 909 908 907 906 905 904 903 902 901 900 10
910 909.5 909
Frquency (MHz)

Frquency (MHz)

908.5 908 907.5 907 906.5 906 905.5 905 T-80 open T-50 open T-80 open T-50 open 10 20 30 T-80 close T-50 close T-80 close T-50 close 40 50 60 70 80 90

open close Trend line (open) Trend Line (close) 30 50 70 Relative Humidity (%) 90

Relative Humidity (%)

(a)

(b)

Fig. 5: Resonant frequency shift due to humidity variation; (a) comparison between two cases (open and close) at 80oC; (b) comparison between 50 oC and 80oC. In the experimental setup where the door of the environmental chamber was closed, and at 80oC, the frequency increased from 906.529MHz to 908.279MHz, for variations in humidity from 20% RH to 80%RH at intervals of 10% RH (Fig. 5(a)), with a variation rate of 12.1 kHz/RH. Whereas, in the case of open chamber door, the variation of frequency was increased from 907.126MHz to 908.627MHz (Fig. 5(a)) with a variation rate of 0.08 kHz/RH. The rate of frequency increase is higher when the chamber door was closed. This is a result of the effect of the metal enclosure on the UHF response as well as heat transfer. This difference is practically insignificant in the open door case believed to be due to reflection and deflection of water vapor at higher relative humidity level beside the metal effect caused by the chamber itself. The added reflection and deflection generally introduces phase shift to the original signal, and such phase shift can be both positive and negative, which could cause the resonant frequency shift either way. As a result, more frequency variation was observed when the chamber was closed. The gap between the two trend lines may be explained by heat transfer loss, identified earlier. The higher the humidity level, the more humidity and frequency loss would be expected when the door is open. Given the fact, the trend lines R2 values are low for both closed and open door cases, 0.106 and 0.00001 respectively, these trend lines are not conclusive. At 50oC, similar frequency behavior was also observed (Fig. 5(b)). A rate of increase of 12.2 kHz/oC for closed chamber door and a rate of decrease of 9.7 kHz/oC for open chamber door, were obtained, respectively. Similar to the 80oC, the similar rate of increase is associated with the open door configuration, indicating the similar effects of
2011 CANSMART CINDE IZFP

the metallic enclosure and humidity impact on the RF signal propagation and showing the repeatability of the experiment. Additionally, the constant displacement of approximately 0.88 MHz also confirmed the frequency shift due to temperature variation (previous section). Although based on Fig. 4(a), the expectation of the displacement would be about 1.5 MHz, such difference could be caused by humidity variation, which is 2.75MHz. It is observed that, at both 50oC and 80oC, the change in frequency shift between the closed and open door configurations, increased as the humidity increased. This increase took place at similar rates but opposite direction. If the average of the frequency shift for both open and closed configurations, for both 50oC and 80oC, is obtained, no change of the frequency shift would be observed with increase in humidity.
701 700
Frequency (MHz)

F1 Valley F1 Peak F2 Peak

Frequency (MHz)

704 702 700 698 696 694 692 690 688

F1 Valley F1 Peak F2 Peak

699 698 697 696 695 694 10 30 50 Temperature (oC) 70 90

20

40

60

80

100

Temperature (oC)

(a)
702 701 700
Frequency (MHz)

(b)
704 702
Frequency (MHz)

F1 Valley F1 Peak F2 Peak

F1 Valley F1 Peak F2 Peak

700 698 696 694 692 690

699 698 697 696 695 694 0 20 40 60 Temperature (oC) 80 100

10

30

50 Temperature (oC)

70

90

(c)

(d)

Fig. 6: The peak and valley frequency shift as a result of temperature variation; (a) close case at 80% RH; (b) open case at 80% RH; (c) close case at 50% RH; (d) open case at 50% RH In addition to the use of the UHF RFID resonant frequency, for the assessment of the impact of temperature and humidity variations on the tags response, characteristics such as impedance valley and peak frequency were also used. Fig. 6 and Fig. 7 present the valley frequency (F1 valley), peak frequency (F1 peak) of imaginary impedance and peak

2011 CANSMART CINDE IZFP

frequency (F2 peak) for real impedance at different temperature and humidity variations, respectively. These experimental conditions, for obtaining the variations in frequency, are identical to those employed in the above two sections.
702 700

705 700

698 696 694 692 690 10 30 50 Relative Humidity (%) 70 90

Frequency (MHz)

Frequency (MHz)

695 690 F1 Valley 685 680 10 30 50 70 Relative Humidity (%) 90 F1 Peak F2 Peak

F1 Valley F1 Peak Real Peak

(a)

(b)
705
Frequency (MHz)

701 699
Frequency (MHz)

700 695 690 685 680 F1 Valley F1 Peak F2 Peak

697 695 693 691 689 687 685 10 30 50 70 Relative Humidity (%)
(c)

F1 Valley F1 Peak F2 Peak 90

10

30

50 Relative Humidity (%)

70

90

(d)

Fig. 7: The impedance frequency shift as a result of humidity variation; (a) close case at 80oC; (b) open case at 80oC; (c) close case at 50oC; (d) open case at 50 oC Fig. 6 illustrates distinctively the rate of increase of F1 valley, F1 peak and F2 peak frequencies as function of temperature increase in both, open and close door configurations and for both high and low humidity, 80% RH and 50% RH, respectively. It is observed that the rate of increase, with temperature increase was lower for open door configuration for both cases of RH. However, as seen in Fig 4.c, a higher rate of decrease was observed for the open door configuration. At this stage of the research, it is not clear why as the temperature increased the resonant frequency shifts decreased, while the F1 valley, F1 peak and F2 peak frequencies increased. The difference between the frequencies at 80% RH and 50% RH was not as pronounced in Fig 6 as was in Fig 4.c. Additionally, Fig. 7 illustrates an insignificant change in F1 valley, F1 peak and F2 peak frequencies as the relative humidity increased from 20% to 80% for both open and closed door configuration and for both 80oC and 50 oC. If the average of open and closed door
2011 CANSMART CINDE IZFP

configurations is taken in the case of Fig. 5, similar conclusions can be drawn about the insignificance change in the resonant frequency shifts. CONCLUSION An investigation on the impact of temperature and humidity variations on a passive UHF RFID system response was conducted. UHF RFID characteristics such as resonant frequency and impedance frequency were used to assess such impact at different temperatures (20 oC to 80oC) and relative humidity (20% to 80%) in closed and open environments. Experimental results demonstrate a linear relationships between passive UHF RFID tags frequency (resonant and impedance) and temperature at different level of humidity. Additionally, as the relative humidity increased, at different levels of temperature, an insignificant change in frequency (resonant and impedance) is observed, indicating the lack of influence of humidity on the tags response. The linear relationship between frequency and temperature could be used as a temperature indicator. Further investigations will focus on understanding the mechanism behind the decrease of resonant frequency and the increase of impedance frequency as function of temperature. REFERENCE [1] RFID Tag Pricing GuideTM 2010, ODIN Technologies Inc, August 10 th, 2010, report. [2] Ron Weinstein, "RFID: A Technical Overview and Its Application to the Enterprise" IT Professional, vol. 7, no. 3, pp. 27-33, May/June 2005, doi:10.1109/MITP.2005.69 [3] X. Z. Jian and H. Z. Yu "A novel wireless charging system for movable telephone with printed-circuit-board windings of different structure and shape respectively", Proc. 2007 Int. Conf. Electrical Machines and Systems, 2007, p.1283. [4] A. Vergara, E. Llobet, J.L. Ramirez, P. Ivanov, L. Fonseca, S. Zampolli, A. Scorzoni, T. Becker, S. Marco and J. Wllenstein, An RFID reader with onboard sensing capability for monitoring fruit quality, Sensors and Actuators B: Chemical 127, 2007, pp. 143149. [5] L. Ho, M. Moh, Z. Walker, T. Hamada, and C. Su, "A prototype on RFID and sensor networks for elder healthcare: progress report", in Proc. of ACM SIGCOMM workshop on Experimental approaches to wireless network design and analysis, 2005. [6] A. Vaz et al., Full passive UHF tag with a temperature sensor suitable for human body temperature monitoring, TCAS II 57 (2), Feb. 2010. [7] G. Marrocco "RFID antennas for the UHF remote monitoring of human subjects", IEEE Trans. Antennas Propag., vol. 55, p.1862 , 2007

2011 CANSMART CINDE IZFP

[8] R.A. Potyrallo and W.G. Morris, Multianalyte chemical identification and quantitation using a single radio frequency identification sensor, Anal. Chem. 79, 2007, pp. 4551 [9] A. Oprea, N. Brsan, U. Weimar, J. Courbat, D. Briand and N.F. de Rooij, Integrated temperature, humidity and gas sensors on flexible substrates for low-power applications, IEEE Sensors 13 (2007), pp. 158161 art. no. 4388360 [10] D. D. Arumugam, and D. W. Engels, Environmental and Performance Analysis of SAW-based RFID Systems, International Journal of Radio Frequency Identification Technology and Applications, 2007. [11] Li Zhenzhong, Mrad Nezih, Xiao George, Guocheng, Liu, Yuu Ono, Ban Dayan, Effects of Orientation and Obstacles on the RFID Performance, Proceedings of the Flyby-Wireless (FBW) 2011 Workshop, Montreal, QC, Canada June 14-17, 2011, the CANEUS Fly-by-Wireless Consortium [12] Guocheng Liu, Mrad Nezih, Xiao George, Li Zhenzhong, Ban Dayan, "RF-based power transmission," Proceedings of the Fly-by-Wireless (FBW) 2011 Workshop, Montreal, QC, Canada - June 14-17, 2011, the CANEUS Fly-by-Wireless Consortium [13] D. M. Dobkin and S. M. Weigand, "Environmental effects on RFID tag antennas", Proc. IEEE Int. Microw. Symp. , 2005 [14] L. Mo and H. Zhang "RFID antenna near the surface of metal", Proc. Int. Symp. Microw., Antenna, Propag. EMC Technol. Wireless Commun., p. 803 , 2007.

2011 CANSMART CINDE IZFP