sensors Article

A Novel Passive Wireless Sensor for Concrete Humidity Monitoring Shuangxi Zhou 1 , Fangming Deng 2,3, *, Lehua Yu 1 , Bing Li 3 , Xiang Wu 2 and Baiqiang Yin 3 1 2 3

*

School of Civil Engineering and Architecture, East China Jiaotong University, Nanchang 330013, China; [email protected] (S.Z.); [email protected] (L.Y.) School of Electrical and Automation Engineering, East China Jiaotong University, Nanchang 330013, China; [email protected] School of Electrical Engineering and Automation, Hefei University of Technology, Hefei 230009, China; [email protected] (B.L.); [email protected] (B.Y.) Correspondence: [email protected]; Tel.: +86-791-8704-6203; Fax: +86-551-6290-4435

Academic Editor: Ferran Martín Received: 20 May 2016; Accepted: 13 September 2016; Published: 20 September 2016

Abstract: This paper presents a passive wireless humidity sensor for concrete monitoring. After discussing the transmission of electromagnetic wave in concrete, a novel architecture of wireless humidity sensor, based on Ultra-High Frequency (UHF) Radio Frequency Identification (RFID) technology, is proposed for low-power application. The humidity sensor utilizes the top metal layer to form the interdigitated electrodes, which were then filled with polyimide as the humidity sensing layer. The sensor interface converts the humidity capacitance into a digital signal in the frequency domain. A two-stage rectifier adopts a dynamic bias-voltage generator to boost the effective gate-source voltage of the switches in differential-drive architecture. The clock generator employs a novel structure to reduce the internal voltage swing. The measurement results show that our proposed wireless humidity can achieve a high linearity with a normalized sensitivity of 0.55% %RH at 20 ◦ C. Despite the high losses of concrete, the proposed wireless humidity sensor achieves reliable communication performances in passive mode. The maximum operating distance is 0.52 m when the proposed wireless sensor is embedded into the concrete at the depth of 8 cm. The measured results are highly consistent with the results measured by traditional methods. Keywords: wireless humidity sensor; RFID technology; concrete humidity measurement; rectifier; regulator; patch antenna

1. Introduction Structural Health Monitoring (SHM) systems are automated tools, aimed at rapidly identifying the onset of structural damage and at tracking the condition of structures during forced or natural excitation [1]. Current SHM systems almost adopt Non-Destructive Testing (NDT) techniques, which can be divided into two types: on-demand type and in situ. The on-demand NDT techniques are based on instrumentation brought to the measurement site on-demand, including X-ray, infrared thermography, laser scanning and microwave radar [2–5]. However, these techniques cannot measure the physical quantities of concrete directly and the measured results are not real-time. The second family is based on the deployment of several local sensors permanently coupled with structural elements. The in situ techniques can also be divided into two groups, according to the use of wired or wireless technology. Among the wired ones [6,7], high measuring precision, great resolution, and real time monitoring can be obtained, however they are quite time-consuming and expensive because of the complicated manufacturing as well as the extensive signal processing.

Sensors 2016, 16, 1535; doi:10.3390/s16091535

www.mdpi.com/journal/sensors

Sensors 2016, 16, 1535

2 of 15

Thanks to fast and easy installation, wireless sensors are gaining a growing interest. Most of the wireless sensors for SHM are based on Wireless Sensor Networks (WSN), which are battery-powered and permit long-range operating [8,9]. However, once these wireless sensors are immersed in concrete, their batteries cannot be replaced again and their lifespan are limited to several years. Furthermore, their cost and complexity are still relatively high. Recently, with the rapid development of Radio Frequency Identification (RFID) technology, the wireless sensors based on passive RFID technology have aroused great interest [10]. The passive tag is able to passively communicate with the interrogator on a zero-powered backscatter mechanism, resulting in simple architecture, low power, and low cost. The passive RFID sensors are excellent choices for long-term and low-cost SHM [11,12]. Humidity is one of the parameters considered as a critical factor in the progress of cement maturity by influencing the stability, the transformation of hydrates, and strength development [13]. The result of this process is an increased strength and decreased permeability. Uncontrolled moisture diffusion during the setting and curing process can cause a number of detrimental effects. For example, fast moisture diffusion resulting from high temperature during curing, may lead to higher shrinkage stresses and lower long-term concrete strength [14,15]. An insufficient water supply due to this diffusion and the evaporation of the internal moisture will hinder the hydration of cement [16]. Besides, during the process of casting reinforced concrete, the corrosion of steel reinforcing bars due to high humidity will decrease the service life of reinforced concrete structures [17]. Therefore, humidity monitoring is significant for concrete. There are various kinds of RFID-based humidity sensors. The simplest kind is the chipless RFID temperature sensor, which requires no Integrated Circuits (ICs) and transmits sensor data by changing the radar cross section of the RFID antenna [18]. The characters of transmitting analog sensor data and no digital blocks lower its performances and constrain its applications mainly in low-cost fields. In situations where there are high accuracy requirements or a large number of sensors in close proximity, chip-based RFID sensors are employed to incorporate digital modulation. A humidity-to-frequency sensor in CMOS technology with wireless readout is presented in [19], however this wireless humidity sensor only achieves a working range of 30 mm because of 13.56 MHz operating frequency. A wireless temperature and humidity sensors based on Ultra-High Frequency (UHF) is proposed in [20]. Because of the complex architecture and off-chip humidity sensor, the wireless humidity sensing range in air is only 2.7 m, which cannot meet the tough requirements of concrete monitoring. This paper presents a novel wireless humidity sensor for SHM. To our knowledge, this is the first work to measure the concrete humidity based on UHF RFID technology. The rest of the paper is organized as follows. Section 2 analyzes electromagnetic wave transmission in concrete and proposes a novel architecture of wireless humidity sensor for ultra-low power application. Section 3 presents the detailed design of key blocks. Section 4 illustrates the measurement results and compared it with the results of traditional measurement. The conclusion is made in Section 5. 2. System Design of Proposed Wireless Humidity Sensor Since concrete has distinct electromagnetic properties at different humidity conditions, the total power loss for electromagnetic waves penetrating concrete are analyzed through the 1 MHz to 1 GHz frequency range for various humidity conditions (at depth of 0.2 m), i.e., 0.5%, 2.5%, 5.5%, and 13%. The matlab simulation results are shown in Figure 1, which are calculated from the ratio of water to the volume of the specimen [21–23]. As expected, due to the reverse variations of the transmission and propagation losses, an optimum frequency range exists, within which there is significantly smaller power loss. For example, the total loss in 20–80 MHz frequency range for wet concrete (13% humidity) is about 5 to 10 dB less than the total loss at either 1 MHz or 1 GHz. The proposed wireless sensor works at 915 MHz, the total loss is about 10 dB at 915 MHz, which is acceptable for concrete monitoring.

Sensors 2016, 16, 1535 Sensors 2016, 16, 1535 Sensors 2016, 16, 1535

3 of 15 3 of 15 3 of 15 15 15

10 10

h= 0.5% h= 0.5% 2.5% h= h= h= 2.5% 5.5% h= 5.5% 13% h= h= 13%

5 5

0 6 10 0 6 10

107 Frequency 107

108

109

(Hz) 108 109 Frequency (Hz) Figure 1. 1. Total loss of of electromagnetic electromagnetic wave wave penetrating penetrating concrete. concrete. Figure Total loss Figure 1. Total loss of electromagnetic wave penetrating concrete.

The aim of the proposed RFID sensor is to perform humidity measurement inside the concrete, The aim aim of the the proposed RFID RFID sensor is is to perform perform humidity humidity measurement measurement inside inside the the concrete, concrete, thus The the typeof of theproposed antenna shouldsensor be takentointo consideration. Figure 2 compares the return loss thus the the type the antenna should be consideration. Figure 22 compares compares the the return return loss thus type of of should be taken taken into into consideration. Figure loss performances of the the antenna same RFID tag respectively equipped with two typical types of antenna (dipole performances of the same RFID tag respectively equipped with two typical types of antenna (dipole performances of the same RFID tag respectively equipped with two typical types of antenna (dipole and patch) in wet concrete and air. The results is based on the extended Debye model of concrete [24] and patch) patch) in in wet wet concrete concrete and and air. air. The results results is is based based on on the extended extended Debye Debye model model of of concrete concrete [24] [24] and and simulated by HFSS. The two The pairs of RFID tags, onethe residing in air and another one residing and simulated by HFSS. The two pairs of RFID tags, one residing in air and another one residing inside and simulated by HFSS. two pairsare of designed RFID tags, residing in MHz. air andFrom another one2aresiding inside the concrete slab atThe 8 cm depth, toone resonate at 915 Figure we can the concrete slab at slab 8 cmatdepth, are designed to resonate at 915 MHz. From Figure 2a we can see can the inside the concrete 8 cm depth, are designed to resonate at 915 MHz. From Figure 2a we see the performance of dipole is dramatically affected when it is embedded in the concrete, the performance of dipoleof is dipole dramatically affected when it is embedded inembedded the concrete, the bandwidth is see the performance is dramatically affected when it isHowever the concrete, bandwidth is wider and the transmission loss increases obviously. seeninfrom Figure 2b, the the wider and the transmission loss increases obviously. However seen from Figure 2b, the performance bandwidth is of wider and the transmission loss increases obviously. However seen from Figure the performance patch antenna is less sensitive to the humidity change of concrete. Hence, the2b, patch of patch antenna is lessantenna sensitiveistoless thesensitive humiditytochange of concrete. Hence, the patchHence, antenna ispatch more performance of patch the humidity change of concrete. the antenna is more suitable for concrete monitoring. suitable for concrete monitoring. antenna is more suitable for concrete monitoring.

Figure 2. Computed S-parameters of two antennas coupling in free space and wet concrete: (a) return Figure 2. S-parameters of and wet Figure 2. Computed Computed S-parameters of two two antennas antennas coupling coupling in in free free space space and wet concrete: concrete: (a) (a) return return loss of dipoles; (b) return loss of patches. loss of dipoles; (b) return loss of patches. loss of dipoles; (b) return loss of patches.

The theoretical practicable operating power of an RFID tag Pt is calculated from the Friis The theoretical operating power of an RFID tag Pt is calculated from the Friis transmission equationpracticable [25]: The theoretical practicable operating power of an RFID tag Pt is calculated from the Friis transmission equation [25]: transmission equation [25]:  2  λ 2 2  P E G     (1) (1) Pt =t Er ·r Ga a· ηrr ·     d  Pt  Er  Ga  r   44πd (1)  4 d  where where EErr is is the the effective effective isotropic isotropic radiation radiation power power of of aa reader, reader, G Gaa is is the the tag tag antenna antenna gain, gain, ηηrr is is the the where E r is the effective isotropic radiation power of a reader, G a is the tag of antenna gain, η r is the RF-to-DC power conversion efficiency of the rectifier, λ is the wavelength the electromagnetic RF-to-DC power conversion efficiency of the rectifier, λ is the wavelength of the electromagnetic RF-to-DC efficiency of the rectifier, λ is(1), thethe wavelength of thedistance electromagnetic wave, dd is communication distance. From Equation communication dd can wave, and andpower is the theconversion communication distance. From Equation (1), the communication distance can be be wave, and d is the communication distance. From Equation (1), the communication distance d can be expressed as: expressed as: expressed as:

Sensors 2016, 16, 1535

4 of 15

s 4 of 15 λ Er · Ga · ηr d= (2) 4π E  GPt  r a r d (2) 4 Pdistance, Hence, in order to achieve longer communication lower Pt and higher η r are critical for t the design of UHF RFID tag, for Er is limited by regional regulations (4 W is the maximum transmitted Hence, in order to achieve longer communication distance, lower Pt and higher ηr are critical for power) and Ga is roughly determined by the allowable antenna area (1.64 for the λ/2 dipole antenna). the design of UHF RFID tag, for Er is limited by regional regulations (4 W is the maximum transmitted Because is a determined high-loss material for electromagnetic wave transmission, this paper power) andconcrete Ga is roughly by the allowable antenna area (1.64 for the λ/2 dipole antenna). proposesBecause a novelconcrete architecture to lower the overall power dissipation of wireless sensor longer is a high-loss material for electromagnetic wave transmission, this for paper operating distance. 3 shows architecture the proposed wireless humidity sensor based on proposes a novelFigure architecture to the lower the overallofpower dissipation of wireless sensor for longer UHFoperating RFID technology. The blocks, for the antenna and matching network, aresensor integrated distance. Figure 3 showsexcept the architecture of the proposed wireless humidity basedon a on chip. UHF The RFIDsensor technology. Thewhich blocks,isexcept for the antenna and matching network, are integrated single antenna, matched with the sensor chip through the matching network, on a single chip. The sensor antenna, which is matched with the sensor chip through the matching receives the electromagnetic waves from the RFID reader. The rectifier multiplies and transfers the network, receivestothe electromagnetic waves fromthe thesubsequent RFID reader. The rectifier and received RF signal a DC supply voltage VR for circuitry. Oncemultiplies the output of the transfers the received RF signal to a DC supply voltage V R for the subsequent circuitry. Once the rectifier reaches the operating voltage, the Power-On-Reset (POR) block generates a reset signal Rst for output of the rectifier reaches the operating voltage, the Power-On-Reset (POR) block generates a the temperature sensor. The clock generator provides a reference clock CLK for the sensor interface. reset signal Rst for the temperature sensor. The clock generator provides a reference clock CLK for the Because the sensor interface and the clock generator should operate under a stable supply voltage sensor interface. Because the sensor interface and the clock generator should operate under a stable VDD , a regulator block is employed to stabilize the output voltage of the rectifier. This architecture supply voltage VDD, a regulator block is employed to stabilize the output voltage of the rectifier. This doesarchitecture not includedoes the not demodulator the baseband blocks in theblocks traditional tag, which means include the and demodulator and the baseband in the RFID traditional RFID tag, thatwhich this wireless humidity sensor will operate without any addressing as long as the sensor receives means that this wireless humidity sensor will operate without any addressing as long as the enough energy from the RFID reader sensor receives enough energy from [26]. the RFID reader [26]. Sensors 2016, 16, 1535

Figure 3. Architecture of the proposed wireless humidity sensor.

Figure 3. Architecture of the proposed wireless humidity sensor.

3. Key Blocks Design

3. Key Blocks Design

Inspired by the reported CMOS humidity sensor designs [27–30], interdigitated top metal Inspired bypolyimide the reported CMOS humidity sensor designs [27–30], interdigitated top metal fingers fingers with filled into the finger gaps, can be utilized for capacitive humidity sensing. withFigure polyimide filled into finger gaps, canproposed be utilized for capacitive humidity Figure 4a 4a illustrates thethestructure of the capacitive humidity sensor. sensing. The proposed illustrates the structure of the proposed capacitive humidity sensor. The proposed humidity sensor humidity sensor was fabricated in the TSMC 0.18 μm 1P6M CMOS process. The top metal layer was (Metal fabricated in the TSMCand 0.18patterned µm 1P6Mwith CMOS process. The lithography top metal layer 6) was deposited 6) was deposited standard optical and (Metal wet etching over the layer to standard form the interdigitated structure. capacitor then covered a the and isolation patterned with optical lithography andThe wetsensing etching over thewas isolation layer towith form humidity-sensitive polyimide layer.capacitor The sensor is then fabricated in awith standard CMOS process without interdigitated structure. The sensing was covered a humidity-sensitive polyimide any post-processing. As seeninfrom Figure 4b, L is the lengthwithout of metal any electrodes, S is the width eachfrom layer. The sensor is fabricated a standard CMOS process post-processing. Asofseen electrode, and W is the distance between adjacent electrodes. The thickness of the polyimide layer is H, Figure 4b, L is the length of metal electrodes, S is the width of each electrode, and W is the distance which is generally larger than metal thickness h. For a N finger array sensor, the total sensor capacitance between adjacent electrodes. The thickness of the polyimide layer is H, which is generally larger than Ch can be expressed as [31]:

metal thickness h. For a N finger array sensor, the total sensor capacitance Ch can be expressed as [31]: Lh Ch  Nε wet Lh W C = Nε h

wet

W

(3)

(3)

where L is the length of metal electrodes, h is the thickness of metal layer, W is the distance between

where L is the length ofand metal h isdielectric the thickness of metal W isfilm the with distance between adjacent electrodes, εwet electrodes, represents the constant of the layer, polyimide absorbed water.electrodes, Considering theεwet factors including capacitance, area sensitivity,film etc.,with this absorbed work adjacent and represents thesensor dielectric constantchip of the polyimide chooses N = 40, L =the 200factors μm, W =including 2.5 μm, S =sensor 2.5 μm,capacitance, h = 1 μm, andchip H = 2area μm. sensitivity, etc., this work water. Considering chooses N = 40, L = 200 µm, W = 2.5 µm, S = 2.5 µm, h = 1 µm, and H = 2 µm.

Sensors 2016, 16, 1535

5 of 15

Sensors 2016, 16, 1535

5 of 15

Sensors 2016, 16, 1535

5 of 15

Figure 4. 4. Proposed of the the Figure Proposed humidity humidity sensor sensor structure structure (a) (a) humidity humidity sensor sensor structure; structure; (b) (b) top top view view of Figure 4. Proposed humidity sensor structure (a) humidity sensor structure; (b) top view of the humidity sensor humidity sensor humidity sensor

The capacitive humidity sensor acts as a capacitor when it works with the sensor interface. The The capacitive sensor acts actsasasa capacitor a capacitor when it works the sensor interface. The capacitivehumidity humidity sensor when it works with with the sensor traditional interface starts with a capacitance-to-voltage converter, which is then interface. followedThe by a The traditional interface starts with a capacitance-to-voltage converter, which is then followed traditional interface starts with a capacitance-to-voltage converter, which is then followed by aby voltage-to-digital converter [32,33]. This technique can achieve high speed and high resolution a voltage-to-digital converter[32,33]. [32,33].This Thistechnique technique canachieve achieve high speed and highresolution resolution voltage-to-digital converter can high speed and high performances. However, due to the use of an operational amplifier, this technique consumes too performances. However, due to the of operational an operational amplifier, technique consumes too performances. However, due to the useuse of an amplifier, thisthis technique consumes too much much power. Figure 5a shows the architecture of proposed sensor interface, which converts the muchFigure power. the architecture of sensor proposed sensorwhich interface, whichthe converts the power. 5aFigure shows 5a theshows architecture of proposed interface, converts sensor signal sensor signal in frequency domain for low-power application. The N-stage ring oscillator is employed sensor signal in frequency domain for low-power application. The N-stage ring oscillator is employed in frequency domain for low-power application. The N-stage ring oscillator is employed to generate to to generate frequency fosc is then countered countered by byaa10-bit 10-bitcounter counter tooutput output generateafrequency acontrolled controlled fosc which which is then thethe a controlled fosc frequency which is then countered by a 10-bit counter to output thetocorresponding corresponding digital output BBo.o.For a standard oscillator,assuming assuming anequal equal rise and fall time corresponding For standard ring ring oscillator, and forfor digital output Bo digital . For a output standard ringaoscillator, assuming an equal risean and fallrise time forfall thetime different thethe different stages respectively, the oscillating frequency of the ring oscillator f osc can be expressed as stagesthe respectively, oscillating of the ring fosc can be expressed stagesdifferent respectively, oscillatingthe frequency of frequency the ring oscillator foscoscillator can be expressed as follows:as follows: follows: I 1 (4) f oscf = 1 = IIll l fosc  (4)(4) ttd VVmCCl osc V C m ll d m where td is the delay time ofofthe loop, is the current current flowingthrough through theinverter, inverter,ClCisl is the equivalent the delaytime timeof theloop, loop,IIIlllis isthe the the equivalent where td is the delay the current flowing flowing throughthe the inverter, Cl is the equivalent where td is load capacitance ofof the is the swing range range ofthe the outputvoltage voltage which mostly equals the load capacitance theloop, loop,and andV isthe the swing swing mostly equals thethe load capacitance of the loop, and VVm mmis range of of theoutput output voltagewhich which mostly equals supply voltage V . Because C is mainly determined by the capacitance C of humidity sensor, f is osc DD l h supply voltage VDD . BecauseCCl lisismainly mainly determined determined by h of humidity sensor, fosc fis supply voltage VDD . Because by the thecapacitance capacitanceCC h of humidity sensor, osc is a sensor-controlled frequency and B is the corresponding digital output of humidity sensor. a sensor-controlledfrequency frequencyand andBBoooisisthe the corresponding corresponding digital sensor. a sensor-controlled digitaloutput outputofofhumidity humidity sensor.

Figure Proposedsensor sensorinterface: interface:(a) (a) architecture architecture of ofof foscfvs. vs. Figure 5. 5. Proposed of sensor sensor interface; interface;(b) (b)simulation simulationresults results osc Figure 5. Proposed sensor interface: (a) architecture of sensor interface; (b) simulation results of fosc vs. temperature on different process corners. temperature on different process corners. temperature on different process corners.

The simulation results of the proposed frequency fosc versus temperature on different process The results proposed frequency ffosc versus temperature on“fast” different process corners are shown in Figureof 5b.the The TT, FF, and SS corners correspond to “nominal”, and “slow” The simulation simulation results of the proposed frequency osc versus temperature on different process corners are shown in Figure 5b. The TT, FF, and SS corners correspond to “nominal”, “fast” and MOSFET devices,inrespectively. TheTT, fosc FF, achieves linearity with temperature and“fast” the simulated corners are shown Figure 5b. The and SSgood corners correspond to “nominal”, and“slow” “slow” MOSFET devices, respectively. The ffosc achieves good linearity linearity with with temperature temperature and worst case variation across corners ±9%. MOSFET devices, respectively. The is oscaround achieves good and the the simulated simulated worst case variation across corners is around ±9%.

worst case variation across corners is around ±9%.

Sensors 2016, 16, 1535 Sensors 2016, 16, 1535

6 of 15 6 of 15

A popular performance metric of a rectifier is its power conversion efficiency which is A popularperformance performance metric a rectifier is its conversion power conversion which as: is A popular metric of aofrectifier is its power efficiencyefficiency which is defined defined as: defined as: Pout P out ηr r= Pout (5) (5) r  Piin (5) n Pin where Pout is the the average average DC DC output output power power generated generated at the output of the rectifier and P out is Pin in is the where Pout is the average DC output power generated at the output of the rectifier and Pin is the RFpower poweravailable availableatat input of the rectifier. a circuit-level of ηview, η r is average RF thethe input of the rectifier. FromFrom a circuit-level point point of view, r is mainly average RF power available at the input of the rectifier. From a circuit-level point of view, ηr is mainly mainly degraded due to the drop forward drop of the switch diodes or [34]. transistors [34]. Differed from degraded due to the forward of the switch diodes or transistors Differed from the reported degraded due to the forward drop of the switch diodes or transistors [34]. Differed from the reported the reported rectifier structure [35–37], Figure shows theofschematic of therectifier, proposed rectifier, which rectifier structure [35–37], Figure 6 shows the 6schematic the proposed which consists of rectifier structure [35–37], Figure 6 shows the schematic of the proposed rectifier, which consists of consists of two identical stages. transistors The NMOSMtransistors MN11–22 and PMOS transistors two identical stages. The NMOS N11–22 and PMOS transistors MP11–22 form theMdifferentialP11–22 form two identical stages. The NMOS transistors MN11–22 and PMOS transistors MP11–22 form the differentialthe differential-drive Inaorder achieve a flat η r curve, bias-voltage VN1,2 and VP1,2 drive switch. In order switch. to achieve flat ηrto curve, the bias-voltage VN1,2the and VP1,2 are employed to boost drive switch. In order to achieve a flat ηr curve, the bias-voltage VN1,2 and VP1,2 are employed to boost are employed to boost the gate-source voltage of NMOS switch and PMOS switch, respectively the gate-source voltage of NMOS switch and PMOS switch, respectively [31]. The large resistor R[31]. S is the gate-source voltage of NMOS switch and PMOS switch, respectively [31]. The large resistor RS is The large resistor is added to block the of VN1,2 and VR . In CMOS process, added to block theRSAC component of V N1,2AC andcomponent VP1,2. In CMOS process, S could be replaced byRaS P1,2 added to block the AC component of VN1,2 and VP1,2. In CMOS process, RS could be replaced by a could be replacedthat by aoperates PMOS transistor that operates inavoid the cut-off region to avoid PMOS transistor in the cut-off region to the large silicon area.the large silicon area. PMOS transistor that operates in the cut-off region to avoid the large silicon area.

Figure 6. Schematic of the proposed two-stage rectifier. Figure 6. 6. Schematic Schematic of of the the proposed proposed two-stage two-stage rectifier. rectifier. Figure

Figure 7 depicts in more details how the proposed scheme sets the DC level of PMOS and NMOS Figure 77depicts details how the the proposed scheme sets the PMOS NMOS depictsininmore more details how proposed scheme setsDC thelevel DCoflevel of and PMOS and gate voltages. The RF intermediate voltage (MU,L) is superimposed on the bias voltages BiasN2 and gate voltages. The RF intermediate voltage (M U , L ) is superimposed on the bias voltages Bias N2 and NMOS gate voltages. The RF intermediate voltage (M , ) is superimposed on the bias voltages BiasP2 to provide the shifted voltages GateN2 and GateP2 UtoLthe gate of NMOS and PMOS switches, BiasP2 provide shifted the voltages Gate N2 andGate GateN2 P2 and to the gate of the NMOS PMOSand switches, and BiasP2 the to provide shifted voltages Gate gate and of NMOS PMOS N2to P2 to respectively. respectively. switches, respectively.

Figure 7. Second stage bias and gate voltages of the proposed gate boosting scheme. Figure Figure 7. 7. Second Second stage stage bias bias and and gate gate voltages voltages of of the the proposed proposed gate gate boosting boosting scheme. scheme.

Due to the large variation of the output voltage of rectifier, a reference circuit with small supply Due to the large variation of the output voltage of rectifier, a reference circuit with small supply Duecoefficient to the large variation of the output voltage of rectifier, a reference circuit with small supply voltage is necessary to generate a stable reference voltage. The traditional way to generate voltage coefficient is necessary to generate a stable reference voltage. The traditional way to generate voltage coefficient is necessary to using generate a stable reference reference circuit voltage. TheHowever, traditional way to meet generate stable 1.25 V reference voltage is a bandgap [38]. it cannot the stable 1.25 V reference voltage is using a bandgap reference circuit [38]. However, it cannot meet the stable 1.25 V requirement reference voltage is using a application. bandgap reference [38]. cannot meet the low-voltage of low-power Figure circuit 8 shows theHowever, schematicit of the proposed low-voltage requirement of low-power application. Figure 8 shows the schematic of the proposed low-voltage requirement of low-power Figure 8 shows the schematic of reference the proposed voltage regulator. It is implemented in application. standard CMOS process without a bandgap and voltage regulator. It is implemented in standard CMOS process without a bandgap reference and voltage It is implemented in M standard CMOS withoutaastart-up bandgapblock, reference and consists regulator. of three parts. The transisitors P0–MP1 and MN0process –MN2 compose which is consists of three parts. The transisitors MP0–MP1 and MN0–MN2 compose a start-up block, which is

Sensors 2016, 16, 1535

7 of 15

Sensors 2016, 16, 1535 Sensors 2016, 1535parts. consists of16, three

7 of 15

7 of 15 The transisitors MP0 –MP1 and MN0 –MN2 compose a start-up block, which is added –M and added as as aaprecautionary precautionarymeasure measureto toensure ensurebias biasininthe thedesired desiredstate. state.The Thetransistors transistorsMM P2 –M P4 and P2 P4 added as a precautionary measure to ensure bias in the desired state. The transistors MP2–MP4 and M composeaa reference reference generator block to provide MN3 N3–MN6 provide aa reference referencevoltage voltageVVREF REF of 0.26 V which which is is N6compose MN3–MN6 compose a reference generator block to provide a reference voltage VREF of 0.26 V which is compensated and M MN7 –M compose compensated by bytemperature temperatureand andsupply supplyvoltage. voltage.The Thetransistors transistorsM MP5P5–M –MP10 P10 and –M N9 compose a N7 N9 compensated by temperature and supply voltage. The transistors MP5–MP10 and MN7–MN9 compose a aregulator regulatortotogenerate generate aa stable voltage V of 1.0 V for other circuits. Figure 9 shows the V − V VDD DD V for other circuits. Figure 9 shows the DD VDD − Vout out regulator to generate a stable voltage VDD◦ of 1.0 V for other circuits. Figure 9 shows the VDD − Vout characteristic 25 °C. C.The Thecircuit circuitstarts startsworking workingproperly properlywith withVV = 0.47 V. characteristic at room temperature of 25 DDDD = 0.47 V. In characteristic at room temperature of 25 °C. The circuit starts working properly with VDD = 0.47 V. In In supply voltage range from0.47 0.47VVtoto2 2V,V,an anaverage averagereference referencevoltage voltageof of 266.5 266.5 mV mV is generated. thethe supply voltage range from the supply voltage range from 0.47 V to 2 V, an average reference voltage of 266.5 mV is generated. In range, the output voltage voltage changes changes at most by 1.8 mV, thus leading In this this measured measured VVDD DD range, leading to to aa line line In this measured VDD range, the output voltage changes at most by 1.8 mV, thus leading to a line sensitivity sensitivity of of 0.441%/V. 0.441%/V. sensitivity of 0.441%/V.

Figure 8. 8. Proposed Proposed voltage voltage regulator. regulator. Figure Figure 8. Proposed voltage regulator.

Figure 9. Simulated supply voltage variation of the proposed rectifier. Figure Figure 9. 9. Simulated Simulated supply supply voltage voltage variation variation of of the the proposed proposed rectifier. rectifier.

The clock generator always employs ring oscillator architecture based on current-starved The clock generator always employs ring oscillator architecture based on current-starved inverters. The generator output voltage traditional inverters is on close to the supplyinverters. voltage The clock alwaysofemploys ring current-starved oscillator architecture based current-starved inverters. The output voltage of traditional current-starved inverters is close to the supply voltage and ground, which improves the output range while increasing the power consumption. As is shown The output voltage of traditional current-starved inverters is close to the supply voltage and ground, and ground, which improves the output range while increasing the power consumption. As is shown in Figure 10, in the order to achieve the balance between output range and power consumption, which improves output range while increasing the power consumption. As is shown in Figurethis 10, in Figure 10, in order to achieve the balance between output range and power consumption, this paper novel structure of ring oscillator for power the clock generator’sthis design. PTAT in orderproposed to achievea the balance between output range and consumption, paperThe proposed paper proposed a novel structure of ring oscillator for the clock generator’s design. The PTAT discussed in Figurefor 3 also the same architecture. The transistors M1–M 6 consist aoscillator novel structure of above ring oscillator theadopts clock generator’s design. The PTAT oscillator discussed oscillator discussed above in Figure 3 also adopts the same architecture. The transistors M1–M6 consist of an Ninstage current-starved transistor MThe 7–Mtransistors 9 and M10–M consist the current mirrors above Figure 3 also adoptsinverter, the samethe architecture. M121 –M consist of an N stage 6 of an N stage current-starved inverter, the transistor M7–M9 and M10–M12 consist the current mirrors of inverter. Compared with the conventional structure, the extra transistors M H1 , M H2 , M L1 , and ML2 current-starved inverter, the transistor M7 –M9 and M10 –M12 consist the current mirrors of inverter. of inverter. Compared with the conventional structure, the extra transistors MH1, MH2, ML1, and ML2 are employed reduce the voltage swingthe [39]. Compared withtothe conventional structure, extra transistors MH1 , MH2 , ML1 , and ML2 are employed are employed to reduce the voltage swing [39]. to reduce the voltage swing [39].

Sensors 2016, 16, 1535 Sensors 2016, 16, 1535 Sensors 2016, 16, 1535

8 of 15 8 of 15 8 of 15

Figure 10. Schematic Schematic of the the proposedclock clock generator. Figure Figure 10. 10. Schematic of of the proposed proposed clock generator. generator.

Microstrip antenna is designed by including a ground plane structure, which has a natural antiMicrostrip antenna is a ground plane structure, which has ahas natural antiMicrostrip isdesigned designedby byincluding including a ground plane structure, which a natural metal advantage compared to the dipole antenna. The length of rectangular microstrip patch antenna metal advantage compared to theto dipole antenna. The length rectangular microstrip patch antenna anti-metal advantage compared the dipole antenna. The of length of rectangular microstrip patch length is approximately equal to half of its electrical length, which is not suitable for the length islength approximately equal equal to halfto of length, which is isnot the antenna is approximately halfitsof electrical its electrical length, which notsuitable suitable for for the miniaturization of tags. As shown in Figure 11a, we presented a miniaturized microstrip antenna miniaturization of of tags. tags. As shown in Figure Figure 11a, 11a, we we presented presented aa miniaturized miniaturized microstrip microstrip antenna antenna miniaturization using the embedded short stub and U-type slot [40]. The impedance of chip was 19-j172 Ω at 915 usingthe theembedded embeddedshort short stub and U-type impedance of chip was 19-j172 Ω at 915 using stub and U-type slotslot [40].[40]. TheThe impedance of chip was 19-j172 Ω at 915 MHz. MHz. The copper traces with the thickness of 0.035 mm were printed on a FR4 substrate, whose MHz. The traces copperwith traces with the thickness of were 0.035printed mm were on a FR4 substrate, whose The copper the thickness of 0.035 mm on a printed FR4 substrate, whose permittivity is permittivity is 4.4 and thickness is 2 mm. The current in the short circuited stub of this antenna is the permittivity is 4.4isand thickness is 2 mm. The current in thestub short stubisof this antennaso is the the 4.4 and thickness 2 mm. The current in the short circuited of circuited this antenna the strongest, strongest, so the impedance of the antenna can be adjusted easily by adjusting the value of Lf1 and strongest, so antenna can by be adjusting adjusted easily by adjusting value Lf1stub and impedance of the the impedance antenna canofbethe adjusted easily the value of Lf1 andthe Lf2. The of short Lf2. The short stub embedded in radiation patch internal can lead to the reduction of the antenna Lf2. The short stub embedded in radiation patch internal can of lead the reduction the antenna embedded in radiation patch internal can lead to the reduction thetoantenna volume.ofFurthermore, volume. Furthermore, because the current can flow around the slot, it results in the increasing of the volume.the Furthermore, the current canit flow around slot, it results the increasing the because current canbecause flow around the slot, results in thethe increasing of the in length of currentof path length of current path and further achieving the miniaturization purpose. The return-loss plot of this length of current path the andminiaturization further achieving the miniaturization purpose. of this and further achieving purpose. The return-loss plot ofThe thisreturn-loss antenna isplot shown in antenna is shown in Figure 11b. The return-loss at 915 MHz frequency is −29 dB and the bandwidth antenna is shown in Figureat11b. frequency is −29 dB and − the Figure 11b. The return-loss 915 The MHzreturn-loss frequency at is 915 −29MHz dB and the bandwidth below 10bandwidth dB is from below −10 dB is from 889 MHz to 923 MHz. The gain of the proposed patch antenna is −3.2 dB and below −10todB from The 889 gain MHzoftothe 923proposed MHz. The gainantenna of the proposed antenna is −3.2 dB and 889 MHz 923isMHz. patch is −3.2 dBpatch and the maximum radiation the maximum radiation direction is right above the antenna. The detailed parameters of the proposed the maximum radiation direction is right above the antenna. The detailed parameters of the proposed direction is right above the antenna. The detailed parameters of the proposed patch antenna is shown patch antenna is shown in Table 1. patch antenna is shown in Table 1. in Table 1.

Figure 11. Proposed patch antenna: (a) antenna design; (b) return-loss plot. Figure 11. 11. Proposed Proposed patch patch antenna: antenna: (a) (a) antenna antenna design; design; (b) (b) return-loss return-loss plot. plot. Figure

Wp Lp Wp Lp 42 mm 53 mm Wp Lp 42 mm 53 mm 42 mm 53 mm

Table 1. Design parameters of the patch antenna. Table1.1. Design Designparameters parametersof ofthe thepatch patchantenna. antenna. Table Ls1 Ls2 Ws1 Ws2 Ws3 Lf1 Lf2 Ls1 Ls2 Ws1 Ws2 Ws3 Lf1 Lf2 23 mm 35 mm 32 mm 29.5 mm 1.7 Ws3 mm 15 mm 3 mm Ls1 Ls2 Ws1 Ws2 Lf1 Lf2 23 mm 35 mm 32 mm 29.5 mm 1.7 mm 15 mm 3 mm 23 mm 35 mm 32 mm 29.5 mm 1.7 mm 15 mm 3 mm

S S 2.5 mm S 2.5 mm 2.5 mm

A protection device is necessary before the sensor tag is embedded into concrete, so that the A protection device is necessary before the sensor tag is embedded into concrete, so that the sensor tag is not impaired by hydration heat arising from grouted concrete and not soaked directly sensor tag is not impaired by hydration heat arising from grouted concrete and not soaked directly in water of concrete. The transmission signal of sensor tag is sent out by electromagnetic wave, hence in water of concrete. The transmission signal of sensor tag is sent out by electromagnetic wave, hence it is important to pay attention to wireless transmission impedance matching. Based on civil it is important to pay attention to wireless transmission impedance matching. Based on civil

Sensors 2016, 16, 1535

9 of 15

6.5 6.5cm cm

A protection device is necessary before the sensor tag is embedded into concrete, so that the sensor tag 16, is not Sensors 2016, 1535impaired by hydration heat arising from grouted concrete and not soaked directly 9 of in 15 Sensors 2016, 16, 1535 9 of 15 water of concrete. The transmission signal of sensor tag is sent out by electromagnetic wave, hence it is important totechniques, pay attention to study wireless transmission impedance Based on civil engineering engineering this proposes aa protective high permeability acrylic package box. engineering techniques, this study proposes protective high matching. permeability acrylic package box. In In techniques, this study proposes a protective high permeability acrylic package box. In order to keep order to keep the sensor in contact with the environment fully, 16 holes were punched on the side order to keep the sensor in contact with the environment fully, 16 holes were punched on the side the sensor in were contact with the 16 holes were punched on the side and 20 holes were and 20 punched at the the package box. 12 the top and 20 holes holes were punched atenvironment the bottom bottom of offully, the proposed proposed package box. Figure Figure 12 shows shows the top view view 2the punched at the bottom of the proposed package box. Figure 12 shows top view of protection device. of protection device. The package box covers an area of 4 × 6.5 mm and the diameter of the hole of protection device. The package box covers an area of 4 × 6.5 mm2 and the diameter of the hole is is 2 and the diameter of the hole is 0.2 cm. The package box covers an area of 4 × 6.5 mm 0.2 cm. 0.2 cm.

Figure Top view Figure 12. 12. Top Top view of of the the proposed proposed package package box. box.

4. 4. Experimental ExperimentalCharacterization Characterization 4. Experimental Characterization Figure wireless humidity sensor, which was in TSMC Figure 13a 13a shows showsthe theproposed proposed wireless humidity sensor, which was fabricated the Figure 13a shows the proposed wireless humidity sensor, which was fabricated fabricated in the the in TSMC 2 and2was equipped with this 0.18 μm CMOS process. The sensor tag chip covers around 8 × 3.5 mm TSMC 0.18 µm CMOS process. The sensor tag covers chip covers around 8 ×mm 3.52 mm and equipped was equipped and was withwith this 0.18 μm CMOS process. The sensor tag chip around 8 × 3.5 antenna on FR4 substrate by using flip chip. Then the sensor tag was packaged with the package this antenna on FR4 substrate by using flip chip. Then the sensor tag was packaged with the package antenna on FR4 substrate by using flip chip. Then the sensor tag was packaged with the package box box for monitoring as in 13b. boxconcrete for concrete monitoring as shown in Figure for concrete monitoring as shown shown in Figure Figure 13b.13b.

Figure Photo of the proposed tag chip; (b) Concrete measuring environment. Figure 13. 13. (a) (a) (a) Photo Photo of of the proposed tag chip; (b) Concrete measuring environment.

Figure 14 shows the wireless testing environment for humidity The VISNFigure thethe wireless testing environment for the the proposed proposed humidity sensor. sensor. The sensor. VISNFigure 14 14shows shows wireless testing environment for the proposed humidity R1200 is a special RFID tester from VI Service Network, which can process, analyze, and display the R1200 is a special RFID tester from VI Service Network, which can process, analyze, and display the The VISN-R1200 is a special RFID tester from VI Service Network, which can process, analyze, and testing signals simultaneously. The sensor performance was measured in a temperature and testing signals simultaneously. The sensor performance was measured in a temperature and display the testing signals simultaneously. The sensor performance was measured in a temperature humidity chamber of VCL4003. The temperature range and resolution of VCL4003 are humidity chamber of Votsch Votsch VCL4003. TheThe temperature range and humidity chamber of Votsch VCL4003. temperature rangeand andresolution resolutionof of VCL4003 VCL4003 are are −40–180 °C and 0.1 °C respectively, the humidity range and resolution of VCL4003 are 10%–98% RH −40–180 °C and 0.1 °C respectively, the humidity range and resolution of VCL4003 are 10%–98% RH and and 1% 1% RH, RH, respectively. respectively.

Sensors 2016, 16, 1535

10 of 15

−40–180 ◦ C and 0.1 ◦ C respectively, the humidity range and resolution of VCL4003 are 10%–98% RH and 1% RH, Sensors 2016, 16,respectively. 1535 10 of 15 Sensors 2016, 16, 1535

10 of 15

Figure environment. Figure 14. 14. Wireless Wireless measurement measurement environment. Figure 14. Wireless measurement environment.

The measured ηr curve of the proposed rectifier is shown in Figure 15a. As compared to the The measured measured ηηrr curve of the proposed rectifier is is shown shown in in Figure Figure 15a. 15a. As compared to the the conventional differential-drive rectifier [36], the proposed rectifier shows a flat ηr curve. When input conventional differential-drive rectifier [36], the proposed rectifier shows a flat η curve. When input conventional differential-drive rectifier [36], the proposed rectifier shows a flat ηrr curve. When input power is 13 dBm, the two curves both reach the optimal point 69%. As for the input power range power power is is 13 13 dBm, dBm, the the two two curves curves both both reach reach the the optimal optimal point point 69%. 69%. As for for the the input input power power range range whose ηr is above 60%, the proposed rectifier and the conventional rectifier achieve −19 dBm and whose 19 dBm and whose ηηrr is above above 60%, 60%, the the proposed proposed rectifier rectifier and and the theconventional conventional rectifier rectifierachieve achieve−−19 and −7 dBm range, respectively. Figure 15b illustrates that the output node of the clock generator swings − 7 dBm range, respectively. Figure 15b illustrates that the output node of the clock generator swings −7 dBm range, respectively. Figure swings between the supply voltage 1.0 V and the ground, while the internal nodes of the clock generator between and thethe ground, while the the internal nodes of the generator vary between the thesupply supplyvoltage voltage1.0 1.0V V and ground, while internal nodes ofclock the clock generator vary from 0.3 V to 0.7 V. The reduction of the voltage swing to nearly 50% of its nominal value lowers from 0.3 V 0.3 to 0.7 The of theofvoltage swing to nearly 50% 50% of itsofnominal value lowers the vary from V toV.0.7 V.reduction The reduction the voltage swing to nearly its nominal value lowers the dynamic power consumption of the internal nodes by 75% and the overall power consumption dynamic power consumption of the internal nodes by 75% andand the the overall power consumption by the dynamic power consumption of the internal nodes by 75% overall power consumption by much more than 25%. much more thanthan 25%. by much more 25%.

Figure 15. (a) Measured power conversion efficiency of the rectifier; (b) output waveforms of the Figure 15. 15. (a) (a) Measured Measured power power conversion conversion efficiency efficiency of of the the rectifier; rectifier; (b) (b) output output waveforms waveforms of of the the Figure proposed ring oscillator at different nodes. proposed ring oscillator at different nodes. proposed ring oscillator at different nodes.

Figure 16a shows the twice measured results of the sensor tag at 20 °C, 35 °C, and 60 °C Figure 16a shows the twice measured results of the sensor tag at 20 ◦°C, 35 ◦°C, and 60 ◦°C respectively within the relative humidity (RH)results range of from to 90%. sensor tag C, achieves Figure 16a shows the twice measured the10% sensor tag The at 20 C, 35 and 60high C respectively within the relative humidity (RH) range from 10% to 90%. The sensor tag achieves high linearity and within shows the a maximum error of 12% forrange 40 °Cfrom offset. Thetotemperature dependence of the respectively relative humidity (RH) 10% 90%. The sensor tag achieves linearity and shows a maximum error of 12% for 40 °C offset. The temperature dependence of the ◦ C error. dielectric constant of the apolyimide εwet 16b showsdependence the measured high linearity and shows maximumfilm error ofresults 12% forin40this offset. Figure The temperature of dielectric constant of the polyimide film εwet results in this error. Figure 16b shows the measured results of six test chips at 20 °C, 35 °C, and 60 °C within the range from 10% RH to 90% RH. Due to the dielectric constant of the polyimide film εwet results in this error. Figure 16b shows the measured results of six test chips at 20◦ °C, 35◦ °C, and 60 °C within the range from 10% RH to 90% RH. Due to the process chips different forfrom a given butRH. all the results of sixvariation, test chipsthe at 20 C, exhibit 35 C, and 60 ◦ Cdigital withinoutputs the range 10%RH RHvalue to 90% Dueresults to the the process variation, the chips exhibit different digital outputs for a given RH value but all the results show anvariation, excellent the linearity. The average sensitivity is 0.55% for RH.a The hysteresis of the process chips exhibit different digital outputs given RH valueperformance but all the results show an excellent linearity. The average sensitivity is 0.55% RH. The hysteresis performance of the sensor at excellent 20 °C is shown in The Figure 16c. The maximum difference between the moisture adsorption show an linearity. average sensitivity is 0.55% RH. The hysteresis performance of the sensor at 20 °C is shown in Figure 16c. The maximum difference between the moisture adsorption and desorption at the point 80% RH does not exceed 10%. and desorption at the point 80% RH does not exceed 10%.

Sensors 2016, 16, 1535

11 of 15

sensor at 20 ◦ C is shown in Figure 16c. The maximum difference between the moisture adsorption and desorption at1535 the point 80% RH does not exceed 10%. Sensors 2016, 16, 11 of 15

◦ C, Figure 16. Measured sensor: (a) digital outputs at 20 ◦°C, Measured humidity performances of the wireless sensor: C, 35 °C, ◦ ◦ ◦ ◦ and 60 °C; C; (b) digital outputs of five test chips at 20 °C, C, 35 °C, C, and 60 °C; C; (c) hysteresis performances ◦ C. at 20 °C.

Table compares the the performances performances of of the the proposed proposed RFID RFID humidity humidity sensor sensor with with previous previous RFID RFID Table 22 compares humidity sensor. The chipless RFID humidity sensor [18] is specially designed for ultra-low cost humidity sensor. The chipless RFID humidity sensor [18] is specially designed for ultra-low cost application, however it has the lowest accuracy because there is no digital block in it. The wireless application, however it has the lowest accuracy because there is no digital block in it. The wireless humidity obtain aa high high power power conversion conversion humidity sensor sensor with with 13.56 13.56 MHz MHz operating operating frequency frequency [19] [19] can can obtain efficiency, however its operating distance is very limited. A RFID-based temperature and humidity efficiency, however its operating distance is very limited. A RFID-based temperature and humidity sensor however its its operating operating distance distance is is only only 2.7 2.7 m. sensor [20] [20] can can achieve achieve complex complex function, function, however m. The The proposed proposed wireless humidity sensor achieves a minimum power dissipation of 5.7 μW, resulting in wireless humidity sensor achieves a minimum power dissipation of 5.7 µW, resulting in aa maximum maximum operating forfor the tough sensing application in operating distance distance of of 17 17m min infree freespace. space.Our Ourdesign designisisespecially especially the tough sensing application concrete. in concrete. Table 2. Performances type RFID RFID humidity humidity sensor. sensor. Table 2. Performances comparison comparison of of various various type Design Design [18] [18] [19] [19] [20] [20] This work This work

Frequency Frequency 900 MHz 900 MHz 13.56 MHz 13.56 MHz 900 MHz 900 MHz 915 MHz 915 MHz

Normalized Inaccuracy Distance Normalized Sensitivity Inaccuracy Distance Sensitivity 1.1% 9.2% No 1.1% 9.2% No0.03 m 0.8% 4.7% 0.8% 4.7% 0.03 2.7 m m 0.7% 4.5% 0.7% 4.5% 2.7 m 0.55% 3.8% 17 m 0.55% 3.8% 17 m

Humidity Cost Humidity Range Cost Range 20%–70% Ultra low 20%–70% low 15%–85% Ultra Low 15%–85% Low 20%–80% Low 20%–80% Low 10%–90% Low 10%–90% Low

Figure 17a shows the experimental scheme of the proposed wireless sensor in concrete. The shows humidity the experimental scheme of the proposed sensor concrete. Theplaced depth depthFigure of the17a wireless embedded in concrete is d, thewireless distance of theinRFID reader of the wireless humidity embedded in angle concrete is d, the of the reader above the above the concrete is H, the incidence between thedistance reader and theRFID sensor is θ. placed The photo of the test site is shown in Figure 17b.

Sensors 2016, 16, 1535

12 of 15

concrete is H, the incidence angle between the reader and the sensor is θ. The photo of the test site is shown in Figure 17b. Sensors 2016, 16, 1535 12 of 15 Sensors 2016, 16, 1535

12 of 15

Figure Figure 17. 17. (a) (a)Experimental Experimentalscheme; scheme; (b) (b) Test Test site. site. Figure 17. (a) Experimental scheme; (b) Test site.

We embedded the sensor into the concrete at aa depth of 8 cm and then investigated the influence We We embedded embedded the the sensor sensor into into the the concrete concrete at at depth of 8 cm and then investigated the influence of angle on power transmission. The relationship between the incidence angle θ and power of incidence incidence angle on power transmission. The relationship between the incidence angle θ and power transmission transmission is is shown shown in in Figure Figure 18a. 18a. From From that, that, we we can can see see the the minimum minimum power power loss loss is is achieved achieved From when θ is 90°, the power transmission efficiency is 85%. When θ is larger than 75°, the power ◦ ◦ when θθ isis 90 When θθ is larger than 75 90°,, the the power power transmission transmission efficiency efficiency is is 85%. 85%. When 75°,, the power transmission efficiency is nearly 0%. Hence, in order to achieve the best test results, the incidence transmission efficiency is nearly 0%. Hence, Hence, in in order to achieve the best test results, the incidence angle should be as vertical as possible. When we set θ = 90°◦ and d = 8 cm, we measure the maximum angle should be as vertical as possible. possible. When we set θθ == 90 90° and d = 8 cm, we measure the maximum communication communication distance distance of of RFID RFID reader. reader. A A calibrated calibrated humidity humidity sensor sensor SHT75 SHT75 is is embedded embedded together together with the proposed wireless humidity sensor to monitor the concrete humidity. For each measurement, with the proposed wireless wireless humidity humidity sensor sensor to monitor the concrete humidity. humidity. For For each measurement, measurement, the reader was instructed to perform 1000 attempts to read the sensor data from the wireless humidity the reader was instructed to perform 1000 attempts to read the sensor data from from the the wireless wireless humidity humidity sensor. When the success ratio remains above 80%, the communication distance is considered to be sensor. When the success ratio remains above 80%, the communication distance is considered to be reliable. As shown in Figure 18b, the measured maximum communication distance decreases when reliable. maximum communication distance decreases when the reliable. As As shown shownin inFigure Figure18b, 18b,the themeasured measured maximum communication distance decreases when the concrete humidity increases. The maximum communication distance is 0.52 m or 0.11 m concrete humidity increases. The maximum communication distance distance is 0.52 m is or 0.11 the concrete humidity increases. The maximum communication 0.52 m mrespectively or 0.11 m respectively when the concrete humidity is 0% or 30%. when the concrete is 0% or 30%.is 0% or 30%. respectively when humidity the concrete humidity

Figure 18. (a) Influence of incidence angle on power transmission; (b) Influence of concrete humidity Figure 18. 18. (a) Figure (a) Influence Influence of of incidence incidence angle angle on on power power transmission; transmission; (b) (b) Influence Influence of of concrete concrete humidity humidity on power transmission. on power transmission. on power transmission.

This work employs concrete made of normal aggregate (1 h pre-wetting, moisture content of This work employs concrete made of normal aggregate (1 h pre-wetting, moisture content of work employs concrete made of normal aggregate h pre-wetting, moisture of 6.6%)This as the sample. The proposed wireless humidity sensor is(1embedded at a depth of 8 content cm inside 6.6%) as the sample. The proposed wireless humidity sensor is embedded at a depth of 8 cm inside 6.6%) as the sample. The proposed wireless humidity sensor is embedded at a depth of 8 cm inside the concrete sample with the calibrated wired humidity sensor SHT75. The results measured by these the concrete sample with the calibrated wired humidity sensor SHT75. The results measured by these the concrete sample with calibrated wired humidity sensor SHT75. The19. results measured by two sensors compared withthe results of the moisture probe are plotted in Figure According to [41], two sensors compared with results of the moisture probe are plotted in Figure 19. According to [41], these two compared with results of the probe are plotted in sensor Figure 19. during thesensors early age of concrete maturing, themoisture traditional wired humidity canAccording get more during the early age of concrete maturing, the traditional wired humidity sensor can get more to [41], during early concrete maturing, traditional wired sensor can get accurate results the in the firstage 14 of days resulting from thethe exchange of air whenhumidity measuring by moisture accurate results in the first 14 days resulting from the exchange of air when measuring by moisture more accurate results theembedded first 14 days resulting from thedrift exchange of air when measuring by probe. Due to the long in time in wet concrete, zero of the traditional wired humidity probe. Due to the long time embedded in wet concrete, zero drift of the traditional wired humidity sensor will decrease the accuracy of the measurement results. Thus, after the 14th day the moisture sensor will decrease the accuracy of the measurement results. Thus, after the 14th day the moisture probe achieves higher measurement accuracy. From Figure 18 we can see in the first 14 days the probe achieves higher measurement accuracy. From Figure 18 we can see in the first 14 days the proposed wireless sensor achieves nearly the same results with the traditional humidity sensor. After proposed wireless sensor achieves nearly the same results with the traditional humidity sensor. After that, the measurement results of the proposed sensor are close to the results measured by the that, the measurement results of the proposed sensor are close to the results measured by the

Sensors 2016, 16, 1535

13 of 15

moisture probe. Due to the long time embedded in wet concrete, zero drift of the traditional wired humidity sensor will decrease the accuracy of the measurement results. Thus, after the 14th day the moisture probe achieves higher measurement accuracy. From Figure 18 we can see in the first 14 days the proposed wireless sensor achieves nearly the same results with the traditional humidity sensor. Sensors 2016, 16, 1535 13 of 15 After that, the measurement results of the proposed sensor are close to the results measured by the moisture probe. Hence the proposed wireless humidity sensor achieves better performances during measurement. the entire measurement.

different sensors. sensors. Figure 19. Performances comparison measured by different

5. Conclusions 5. Conclusions Humidity monitoring monitoring plays plays an an important important role role in in concrete concrete measurement, measurement, and and thus thus this this work work Humidity presents a wireless humidity sensor based on UHF RFID technology. Considering the high losses of presents a wireless humidity sensor based on UHF RFID technology. Considering the high losses of electromagneticwaves wavesininconcrete, concrete, a patch antenna is proposed to ensure the sensor tag caninside work electromagnetic a patch antenna is proposed to ensure the sensor tag can work inside of the concrete. The wireless humidity sensor employs a novel architecture and is carefully of the concrete. The wireless humidity sensor employs a novel architecture and is carefully designed designed for low power application. Theresults measured show the proposed sensorhigh taghumidity achieves for low power application. The measured showresults the proposed sensor tag achieves high humidity linearity with a normalized sensitivity of◦ C. 0.55% at 20operating °C. The distance maximum linearity with a normalized sensitivity of 0.55% %RH at 20 The %RH maximum is operating distance is 0.52 m when the proposed wireless sensor is embedded into the concrete at 0.52 m when the proposed wireless sensor is embedded into the concrete at a depth of 8 cm. Thea depth of 8results cm. The results are highly consistent with the results measured by The traditional measured aremeasured highly consistent with the results measured by traditional methods. whole methods. The whole experiment demonstrates that the proposed wireless humidity sensor can experiment demonstrates that the proposed wireless humidity sensor can provide reliable performance. provide reliable performance. Acknowledgments: This work was supported by National Key Research and Development Program (2016YFC0700802), Natural Foundation (61501162), Postdoctoral Science Foundation Acknowledgments:National This work wasScience supported by National Key China Research and Development Program (2015M571926), and Science and Technology Support Project of Jiangxi Province (20161BBE50076). (2016YFC0700802), National Natural Science Foundation (61501162), China Postdoctoral Science Foundation Author Contributions: Shuangxi Zhou is in charge of Project this work and Fangming provides the instructions (2015M571926), and Science and Technology Support of Jiangxi ProvinceDeng (20161BBE50076). for the design of wireless humidity sensors. Xiang Wu and Baiqiang Yin provide the helps in designing the RFID Authortag. Contributions: Shuangxi Zhou is the in charge ofmeasurement. this work and Fangming Deng provides the instructions sensor Bing Li and Lehua Yu provide helps in for the design of wireless humidity sensors. Xiang Wu and Baiqiang Yin provide the helps in designing the RFID Conflicts of Interest: The authors declare no conflict of interest. sensor tag. Bing Li and Lehua Yu provide the helps in measurement.

References Conflicts of Interest: The authors declare no conflict of interest. 1.

Chang, P.C.; Flatau, A.; Liu, S.C. Review Paper: Health Monitoring of Civil Infrastructure. Struct. Health Monit. 2003, 2, 257–267. [CrossRef] Chang, Flatau, A.; Liu, S.C. Review Paper: Health Monitoring of Civil Infrastructure. Sitek, L.P.C.; Analysis of inner structure changes of concretes exposed to high temperatures usingStruct. micro Health X-ray computed tomography. Monit. 2003, 2, 257–267. Acta Geodyn. Geomater. 2015, 12, 79–89. [CrossRef] Kamoi, Okamoto, Y.; Vavilov, V. Study onofDetection of Buried in Concrete by Sitek, L.A.; Analysis of inner structure changes concretesLimit exposed to highDefects temperatures usingStructures micro X-ray Using Infrared Thermography. Key Eng. Mater. 2004, computed tomography. Acta Geodyn. Geomater. 2015, 270–273, 12, 79–89.1549–1555. [CrossRef] González-orge, H.; Gonzalez-Aguilera, D.; on Rodriguez-Gonzalvez, P. Monitoring in civil Kamoi, A.; Okamoto, Y.; Vavilov, V. Study Detection Limit of Buried Defects inbiological Concrete crusts Structures by engineering structures using intensity from terrestrial laser scanners. Constr. Build. Mater. 2012, 31, Using Infrared Thermography. Key Eng.data Mater. 2004, 270–273, 1549–1555. 119–128. [CrossRef] González-orge, H.; Gonzalez-Aguilera, D.; Rodriguez-Gonzalvez, P. Monitoring biological crusts in civil engineering structures using intensity data from terrestrial laser scanners. Constr. Build. Mater. 2012, 31, 119–128. Pieraccini, M.; Luzi, G.; Mecatti, D.; Noferini, L.; Atzeni, C. A microwave radar technique for dynamic testing of large structures. IEEE Trans. Microw. Theory Tech. 2003, 51, 1603–1609. Ravet, F.; Briffod, F.; Glisic, B.; Nikle, M.; Inaudi, D. Submillimeter crack detection with Brillouin-based fiber-optic sensors. IEEE Sens. J. 2009, 9, 1391–1396.

References

1. 2. 3. 2. 4. 3.

4.

5. 6.

Sensors 2016, 16, 1535

5. 6. 7.

8. 9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22. 23. 24.

25.

26.

27. 28.

14 of 15

Pieraccini, M.; Luzi, G.; Mecatti, D.; Noferini, L.; Atzeni, C. A microwave radar technique for dynamic testing of large structures. IEEE Trans. Microw. Theory Tech. 2003, 51, 1603–1609. [CrossRef] Ravet, F.; Briffod, F.; Glisic, B.; Nikle, M.; Inaudi, D. Submillimeter crack detection with Brillouin-based fiber-optic sensors. IEEE Sens. J. 2009, 9, 1391–1396. [CrossRef] Ramakrishnan, M.; Rajan, G.; Semenova, Y.; Farrell, G. Hybrid Fiber Optic Sensor System for Measuring the Strain, Temperature, and Thermal Strain of Composite Materials. IEEE Sens. J. 2014, 14, 2571–2578. [CrossRef] Li, P.; Olmi, C.; Song, G. Energy efficient wireless sensor network for structural health monitoring using distributed embedded piezoelectric transducers. Proc. SPIE 2010, 7647, 2157–2164. Lédeczi, Á.; Hay, T.; Völgyesi, P.; Robert, H.D.; Nadas, A.; Jayaraman, S. Wireless acoustic emission sensor network for structural monitoring. IEEE Sens. J. 2009, 9, 1370–1377. [CrossRef] Li, B.; He, Y.G.; Zuo, L.; Long, Y. Metric of the Application Environment Impact to the Passive UHF RFID System. IEEE Trans. Instrum. Meas. 2014, 63, 2387–2395. [CrossRef] Caizzone, S.; Giampaolo, E.D. Wireless Passive RFID Crack Width Sensor for Structural Health Monitoring. IEEE Sens. J. 2015, 15, 6767–6774. [CrossRef] Leon-Salas, W.; Halmen, C. A RFID Sensor for Corrosion Monitoring in Concrete. IEEE Sens. J. 2015, 16, 32–42. [CrossRef] Delatte, N.; Williamson, M.S.; Fowler, D.W. Bond strength development with maturity of high-early-strength bonded concrete overlays. ACI Mater. J. 2000, 97, 201–207. Shen, D.; Wang, T.; Chen, Y.; Wang, M.; Jiang, G. Effect of internal curing with super absorbent polymers on the relative humidity of early-age concrete. Constr. Build. Mater. 2015, 99, 246–253. [CrossRef] De Medeiros-Junior, R.A.; De Lima, M.G.; De Medeiros, M.H.F. Service life of concrete structures considering the effects of temperature and relative humidity on chloride transport. Environ. Dev. Sustain. 2015, 17, 1103–1119. [CrossRef] Lothenbach, B.; Matschei, T.; Möschner, G.; Glasser, F.P. Thermodynamic modelling of the effect of temperature on the hydration and porosity of Portland cement. Cem. Concr. Res. 2008, 38, 1–18. [CrossRef] Escalante-Garcia, I.; Sharp, H. Effect of temperature on the hydration of the main clinker phases in Portland cements: Part I, neat cements. Cem. Concr. Res. 1998, 28, 1245–1257. [CrossRef] Feng, Y.; Xie, L.; Chen, Q.; Zheng, L.R. Low-Cost Printed Chipless RFID Humidity Sensor Tag for Intelligent Packaging. IEEE Sens. J. 2015, 15, 3201–3208. [CrossRef] Cirmirakis, D.; Demosthenous, A.; Saeidi, N.; Donaldson, N. Humidity-to-Frequency Sensor in CMOS Technology with Wireless Readout. IEEE Sens. J. 2013, 13, 900–908. [CrossRef] Yu, S.M.; Feng, P.; Wu, N.J. Passive and Semi-Passive Wireless Temperature and Humidity Sensors Based on EPC Generation-2 UHF Protocol. IEEE Sens. J. 2015, 15, 2403–2411. [CrossRef] Pena, D.; Feick, R.; Hristov, H.D.; Grote, W. Measurement and modeling of propagation losses in brick and concrete walls for the 900-MHz band. IEEE Trans. Antennas Propag. 2003, 51, 31–39. [CrossRef] Savov, S.V.; Herben, M.H.A. Modal transmission-line modeling of propagation of plane radiowaves through multilayer periodic building structures. IEEE Trans. Antennas Propag. 2003, 51, 2244–2251. [CrossRef] Sandrolini, L.; Reggiani, U.; Ogunsola, A. Modelling the electrical properties of concrete for shielding effectiveness prediction. J. Phys. D Appl. Phys. 2007, 40, 5366–5372. [CrossRef] Jiang, S.; Georgakopoulos, S.V. Optimum wireless power transmission through reinforced concrete structure. In Proceedings of the 2011 IEEE International Conference on RFID (RFID), Orlando, FL, USA, 12–14 April 2011; pp. 50–56. Cartasegna, D.; Conso, F.; Donida, A.; Grassi, M.; Picolli, L.; Rescio, G.; Regnicoli, G.F. Integrated microsystem with humidity, temperature and light sensors for monitoring the preservation conditions of food. In Proceedings of the 2011 IEEE Sensor, Limerick, Ireland, 28–31 October 2011; pp. 1859–1862. Kapucu, K.; Dehollain, C. A passive UHF RFID system with a low-power capacitive sensor interface. In Proceedings of the 2014 IEEE RFID Technology and Applications Conference (RFID-TA), Tampere, Finland, 8–9 September 2014; pp. 301–305. Gu, L.; Huang, Q.A.; Qin, M. A novel capacitive-type humidity sensor using CMOS fabrication technology. Sens. Actuators B Chem. 2004, 99, 491–498. [CrossRef] Zhao, C.L.; Qin, M.; Huang, Q.A. A fully packaged CMOS interdigital capacitive humidity sensor with polysilicon heaters. IEEE Sens. J. 2011, 11, 2986–2992. [CrossRef]

Sensors 2016, 16, 1535

29.

30.

31.

32. 33.

34.

35. 36. 37.

38. 39. 40. 41.

15 of 15

Kim, J.-H.; Hong, S.-M.; Lee, J.-S.; Moon, B.-M.; Kim, K. High sensitivity capacitive humidity sensor with a novel polyimide design fabricated by MEMS technology. In Proceeding of the 4th IEEE International Conference on Nano/Micro Engineered and Molecular Systems, Shenzhen, China, 5–8 January 2009; pp. 703–706. Dai, C.L.; Lu, D.H. Fabrication of a micro humidity sensor with polypyrrole using the CMOS process. In Proceedings of the 5th IEEE International Conference on Nano/Micro Engineered and Molecular Systems, Nice, France, 20–23 January 2010; pp. 110–113. Paavola, M.; Kamarainen, M.; Laulainen, E.; Saukoski, M.; Koskinen, L.; Kosunen, M.; Halonen, K.A. A micropower-based interface ASIC for a capacitive 3-axis micro-accelerometer. IEEE. J. Solid-State Circuits 2009, 44, 3193–3210. [CrossRef] Shin, D.Y.; Lee, H.; Kim, S. A delta–sigma interface circuit for capacitive sensors with an automatically calibrated zero point. IEEE Trans. Circuits Syst. II 2011, 58, 90–94. [CrossRef] Kamalinejad, P.; Keikhosravy, K.; Mirabbasi, S.; Leung, V. An efficiency enhancement technique for CMOS rectifiers with low start-up voltage for UHF RFID tags. In Proceedings of the International Green Computing Conference (IGCC), Dallas, TX, USA, 3–5 November 2013; pp. 1–6. Theilmann, P.T.; Presti, C.D.; Kelly, D.; Asbeck, P.M. Near zero turn-on voltage high-efficiency UHF RFID rectifier in silicon-on-sapphire CMOS. In Proceedings of the IEEE Radio Frequency Integrated Circuits Symposium (RFIC), Anaheim, CA, USA, 23–25 May 2010; pp. 105–108. Mandal, S.; Sarpeshkar, R. Low-power CMOS rectifier design for RFID applications. IEEE Trans. Circuits Syst. I 2007, 54, 1177–1188. [CrossRef] Kotani, K.; Sasaki, A.; Ito, T. High-efficiency differential-drive CMOS rectifier for UHF RFIDs. IEEE J. Solid-State Circuits 2009, 44, 3011–3018. [CrossRef] Kamalinejad, P.; Keikhosravy, K.; Mirabbasi, S. A CMOS rectifier with an extended high-efficiency region of operation. In Proceedings of the 2014 IEEE RFID Technology and Applications Conference (RFID-TA), Johor Bahru, Malaysia, 4–5 September 2013; pp. 1–6. Leung, N.N. A sub-1-V 15-ppm/◦ C CMOS bandgap reference without requiring low threshold voltage device. IEEE J. Solid-State Circuits 2002, 37, 526–529. [CrossRef] Azarmehr, M.; Rashidzadeh, R.; Ahmadi, M. Low-power oscillator for passive radio frequency identification transponders. IET Circuits Devices Syst. 2012, 6, 79–84. [CrossRef] Eldek, A.A. Miniaturized patch antenna for RFID tags on metallic surfaces. Microw. Opt. Technol. Lett. 2011, 53, 2170–2174. [CrossRef] Grasley, Z.C.; Lange, D.A. Thermal dilation and internal relative humidity of hardened cement paste. Mater. Struct. 2007, 40, 311–317. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).