A Novel Surface Acoustic Wave Sensor Array Based on Wireless Communication Network

Article A Novel Surface Acoustic Wave Sensor Array Based on Wireless Communication Network

Yong Pan 1 , Ning Mu 1, Bo Liu 1, Bingqing Cao 1, Wen Wang 2,* and Liu Yang 1,*

1 State Key Laboratory of NBC Protection for Civilian, Beijing 102205, China; [email protected] (Y.P.); [email protected] (N.M.); [email protected] (B.L.); [email protected] (B.C.) 2 State Key Laboratory of Acoustics, Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China * Correspondence: [email protected] (W.W.); [email protected] (L.Y.); Tel.: +86-10-8254-7803 (W.W.); +86-10-6675-8136 (L.Y.) Received: 18 July 2018; Accepted: 4 September 2018; Published: 6 September 2018

Abstract: A novel surface acoustic wave (SAW) sensor array based on wireless communication network is prepared. The array is composed of four SAW sensors, a wireless communication network module, and a global positioning system (GPS) module. The four SAW sensors of the array are coated with triethanolamine, polyepichlorohydrin, fluoroalcoholpolysiloxane, and L-glutamic acid hydrochloride to detect hydrogen sulfide (H2S), 2-chloroethyl ethyl sulfide (CEES), dimethylmethylphosphonate (DMMP), and ammonia (NH3) at film thicknesses of 50–100 nm. The wireless communication network module consists of an acquisition unit, a wireless control unit, and a microcontroller unit. By means of Zigbee and Lora technologies, the module receives and transmits the collected data to a PC work station in real-time; moreover, the module can control the sensor array’s working mode and monitor the working status. Simultaneously, the testing location is determined by the GPS module integrated into the SAW sensor array. H2S, CEES, DMMP, and NH3 are detected in 300 m at different concentrations. Given the practical future application in environment in the future, the low, safe concentrations of 1.08, 0.59, 0.10, and 5.02 ppm for H2S, CEES, DMMP, and NH3, respectively, are detected at the lowest concentration, and the sensitivities of different sensors of the sensor array are 32.4, 14.9, 78.1 and 22.6 Hz/ppm, respectively. With the obtained fingerprints and pattern recognition technology, the detected gases can be recognized.

Keywords: SAW sensor array; wireless communication; network; detection; gases

1. Introduction Environment and public security are becoming increasingly important to human health and safety. With the development of the world, harmful gases or numerous other volatile organic compounds (VOCs), which are widely used in the industries, pose potential risks for serious diseases and environmental and explosion dangers. For example, hydrogen sulﬁde (H2S), a harmful gas, is a colorless toxic gas that can cause adverse effects on the human body at a low concentration of 2 ppm, and can be fatal with prolonged exposure at levels of 100 ppm [1]. Ammonia (NH3) is another harmful gas that is primarily released from combustion in chemical industries, and exposure to high ammonia concentration can result in life-threatening situations [2]. By contrast, chemical warfare agents (CWAs) are powerful weapons and a threat to civil safety because of their extremely high hazard and potential lethality. Particularly in the recent years, numerous terrorism events involving CWAs that resulted in numerous casualties or injuries have been reported. Therefore, the threat of CWAs is crucial to national security and world affairs. Fast and sensitive detection of industrial harmful gases, VOCs, and CWAs is highly essential in protecting humans [3–5]. In the last 20 years, sensor technology has received

Sensors 2018, 18, 2977; doi:10.3390/s18092977 www.mdpi.com/journal/sensors Sensors 2018, 18, 2977 2 of 11 increasing attention for ubiquitous sensing. As important sensing technologies, surface acoustic wave (SAW) devices have exhibited promising characteristics because of their high sensitivity, fast response, excellent specificity, reversibility, battery-powered operation, small size, and low cost, thus offering extensive potential use in the future. Several reviews presented the mechanism for chemical specificity and chemically specific layers of SAW sensors; numerous studies focused on the detection of industrial harmful gases, VOCs, or CWAs [6–8]. A SAW sensor array is typically required to identify the detected gases accurately, and each sensor in the array should be coated with different sensing films. By simultaneous data gathering and analysis of substantial data and pattern recognition methods, the targets can be discriminated; thus, the chemical sensor array can play an important role in the analysis of gas analytes [9]. CWAs have been successfully detected by different SAW arrays [7,10], and several harmful gases, such as H2S, NH3, and NO2, are also detected by SAW technology [1,2,11–13]. However, most currently available SAW sensor arrays for detecting industrial harmful gases, VOCs or CWAs are manually operated, extremely costly, and time consuming [14]; thus, real-time, continuous, long-term monitoring sensors are attractive for in situ detection, possibly reducing the need for manual operation and expensive off-site analyses. Passive wireless SAW resonant sensors can meet the abovementioned requirement because of their application to a harsh environment and simple installation; wiring and penetration into the harsh environment is not needed [15,16]. In recent years, several SAW embodiments have been proposed for wireless, passive SAW sensors [17,18], and these may be practical for portable or wireless sensor networks (WSN). Numerous designers presented the option of integrating general WSN platforms with sensor devices for a variety of applications [19–21], and several wireless and passive SAW-based sensors have been used to detect organophosphorus compounds [22], CO2, NO2 or humidity [23–25]. Several studies on wireless chemical sensors and algorithms for SAW sensors were reviewed [26,27]. Meanwhile, certain design technologies, such as coupling of modes, which is an efficient technique for SAW devices, are used before fabrication to find optimal design parameters of the SAW sensors [28]. All of the abovementioned works provide a good starting point for the development of real practical wireless SAW chemical sensors. In the last decade, with the development of computer systems, particularly the evolution of processor systems and advances in communication networks, WSN is becoming the focus of sensor studies and is regarded as one of the most promising future technologies [29–31]. An increasingly large market is trying to apply WSN to sensor technology, and sensors combined with a wireless communication network are becoming the current trend. Meanwhile, global positioning systems (GPSs) have been used in WSN, enabling the establishment of a wireless sensors communication network to monitor target gases. As the prominent technologies, SAW sensor arrays are tested for use in this monitoring platform [18,21]. In this study, we present a new SAW sensor array based on WSN and established a novel wireless communication platform, which successfully detected harmful gases (H2S, NH3) and warfare agent simulants (DMMP, CEES, simulants of sarin and mustard gas) within 300 m and located the testing position. Although wireless communication is a big challenge in the study of sensors, the WSN that we suggest here may be used in environmental protection or in any other future monitoring application.

2. Materials and Methods

2.1. Reagents and Instruments The SAW sensor array used in this work is fabricated by Institute of Acoustic of Chinese Academy of Science. The array consisted of a five-channel sensing device with two-port resonator configuration: one device is naked as the reference to compensate the influence of temperature and pressure, and the other four devices are coated with different films for sensing target gases. Each device is fabricated lithographically on ST-X quartz, the corresponding working frequency is set to 300 MHz. The proposed SAW resonator is composed of two interdigital transducers (IDTs) and two reflector located in each Sensors 2018, 18, 2977 3 of 11 side of the IDT. Also, the Al-electrodes are used to build the SAW devices. To deposit sensitive film Sensors 2018, 18, x FOR PEER REVIEW 4 of 11 easily, the resonant cavity of 2 mm2 between the IDTs was designed. The developed sensing devices solutionare collected prepared, into the and differential v2 is the volume oscillation removed loop, which(1 μL). is m composed1 is equal ofto them2; phasetherefore, shifter, the amplifier, average thicknessand mixer of described TEA can inbe Figureestimated1. The using mixed Formula oscillation (3) and frequency is approximately signal is collected 70 nm. The to evaluate other three the kindsgases of to films be detected. are also prepared similarly.

(a) (b)

FigureFigure 1. 1. ((aa) )Schematic Schematic of of the the SAW SAW sensor sensor array array (sensor (sensor 1, 1, coated coated with with TEA; TEA; sensor sensor 2, 2, coated coated with with PECH,PECH, sensor 3, coatedcoated withwith SXFA, SXFA sensor, sensor 4, 4, coated coated with with GAH); GAH); (b) The(b) The photo photo of the of SAW the SAW sensor sensor array. array. Triethanolamine (TEA), polyepichlorohydrin (PECH), and L-glutamic acid hydrochloride (GAH) were 2.3.purchased Structure from of the Aladdin Wireless Chemical Communication Reagent Company SAW Sensor (Shanghai, Array China). Fluoroalcoholpolysiloxane (SXFA) was synthesized by our laboratory. The detected gases, H S, and NH were provided by Beijing Haipu The wireless communication network module of the2 SAW sensor3 array system was designed Company (Beijing, China), and DMMP and CEES were generated in our laboratory. and developed by our laboratory. This system consists of an acquisition layer, a terminal layer, and IP-LINK1223-5152 is an embedded wireless module based on IEEE 802.15.4/ZigBee technology an application layer. The three parts are equipped together, forming the total system. A block launched by Helicomm Company (San Diego, CA, USA). E32 series wireless communication module diagram of the detection and wireless communication system of the SAW sensor array is shown in is a wireless serial module that uses the LoRa spread spectrum technology made by Semtech Company Figure 2. (Camarillo, CA, USA) and works in the 410–441 MHz band (the default 433 MHz). RXM-GPS-FM-T is a Linx Technologies (London, UK) FM series GPS module. The accuracy of the horizontal position is 3 m, which supports a maximum of 66 channels and reaches a sensitivity value of −161 dBm.

2.2. Preparation of SAW Sensor Array

As mentioned above, H2S, CEES, DMMP, and NH3 were detected by SAW sensors coated with TEA, PECH, SXFA and GAH. All the films selected in this study are viscoelastic. TEA is a kind of alkaline film, atom S contained in H2S could combine with atom H contained in TEA by weak hydrogen-bond and van der Waals force, CEES could permeate into PECH easily for its good solubility, while the use of SXFA could form conjugated structure with P=O of organophosphorous compounds due to the –CF3 in SXFA, for GAH, NH3 could dissolve easily in GAH because of the rule of similarity. The sensing mechanisms of adsorption between detected gases and films had been discussed before, and the studies were effective. On these bases, TEA, PECH, SXFA and GAH were selected in this study as sensitive membrane materials to detect H2S, CEES, DMMP, and NH3, respectively. The structure and photo image of the SAW sensor array used in this study are shown in Figure1. Five channels are included in the sensor array: the uncoated middle channel is selected as the reference to delete the influence of temperature and pressure. The other four sensors (sensor 1, sensor 2, sensor 3, and sensor 4), which are coated with TEA, PECH, SXFA, and GAH, are used to detect H2S, CEES, DMMP,Figure and 2. NH Detecting3. and wireless communication system of the SAW sensor array. 1—Decision Making User; 2—Display Platform; 3—Database Server; 4—Cloud Computing Platform; 5—Application Server 1; 6—Application Server 2; 7—Control Terminal 1; 8—Database 1; 9—Database 2; 10—Cable/Wireless Network; 11—Control Terminal 2; 12—SAW Sensor 1; 13—SAW Sensor 2; 14—SAW Sensor 3; 15—SAW Sensor 5; 16—SAW Sensor 4; 17—Operating User.

Sensors 2018, 18, x FOR PEER REVIEW 4 of 11

solution prepared, and v2 is the volume removed (1 μL). m1 is equal to m2; therefore, the average Sensorsthickness2018, 18 of, 2977 TEA can be estimated using Formula (3) and is approximately 70 nm. The other4 of three 11 kinds of films are also prepared similarly.

In preparing of SAW sensor array, all the films were prepared by dipping method, the operation method is as follows. First, TEA and GAH were prepared in ethanol solvent at a concentration of 0.2 mg/mL. PECH and SXFA were prepared using chloroform and acetone, respectively, at the concentration of 0.1 mg/mL. Second, to get the needed thickness of the films, by removing some volume solution having been prepared, after dipping on the surface of the SAW sensor and calculating, the needed thickness film could be gotten. For example, 1 µL of TEA solution was withdrawn and dipped on the coating area, which was dried with N2 at room temperature. TEA was deposited on the surface of the SAW device, and the thickness of TEA film was calculated using the formula below:

m1 = ρv1 = ρsh (1)

m2 = cv2 (2)

h = cv2/ρs (3) 3 In Formula (1), m1 (mg) is the amount of TEA on coating area, ρ (g/cm ) is the density of TEA, 3 (a) 2 (b) v1 (cm ) is the volume of TEA on the coating surface, s (cm ) is coating area, and h (cm) is the thickness of ﬁlm.Figure In Formula 1. (a) Schematic (2), m2 (mg) of the is theSAW amount sensor ofarray TEA, (sensorc is the 1, coated concentration with TEA; of TEAsensor solution 2, coated prepared, with and v2PECH,is the volumesensor 3, removedcoated with (1 µSXFAL). m, sensor1 is equal 4, coated to m2 ;with therefore, GAH); the(b) The average photo thickness of the SAW of sensor TEA can be estimatedarray. using Formula (3) and is approximately 70 nm. The other three kinds of ﬁlms are also prepared similarly. 2.3. Structure of the Wireless Communication SAW Sensor Array 2.3. Structure of the Wireless Communication SAW Sensor Array The wireless communication network module of the SAW sensor array system was designed andThe developed wireless by communication our laboratory. network This system module consists of the SAWof an sensoracquisition array layer, system a wasterminal designed layer, and and developedan application by our layer. laboratory. The three This parts system are consists equipped of an together, acquisition forming layer, the a terminal total system. layer, andA block an applicationdiagram of layer. the detection The three and parts wire areless equipped communication together, forming system theof the total SAW system. sensor Ablock array diagram is shown of in theFigure detection 2. and wireless communication system of the SAW sensor array is shown in Figure2.

FigureFigure 2. Detecting2. Detecting and and wireless wireless communication communication system system of the SAWof the sensor SAW array. sensor 1—Decision array. 1—Decision Making User;Making 2—Display User; Platform;2—Display 3—Database Platform; Server; 3—Databa 4—Cloudse ComputingServer; 4—Cloud Platform; 5—ApplicationComputing Platform; Server 1;5—Application 6—Application ServerServer 2; 7—Control1; 6—Application Terminal Server 1; 8—Database 2; 7—Control 1; 9—Database Terminal 2; 10—Cable/Wireless 1; 8—Database 1; Network;9—Database 11—Control 2; 10—Cable/Wireless Terminal 2; 12—SAW Network; Sensor 11—Con 1; 13—SAWtrol Terminal Sensor 2;2; 14—SAW12—SAW Sensor Sensor 3; 1; 15—SAW 13—SAW SensorSensor 5; 2; 16—SAW 14—SAW Sensor Sensor 4; 3; 17—Operating 15—SAW Sensor User. 5; 16—SAW Sensor 4; 17—Operating User.

Sensors 2018, 18, x FOR PEER REVIEW 5 of 11

SensorsThe2018 acquisition, 18, 2977 layer is located at the front of the whole system. The layer is composed of 5SAW of 11 sensor array, data acquisition module, GPS module, and Zigbee and Lora transceiver modules. As a wirelessThe transceiver acquisition module, layer is locatedit not only at the collects front ofthe the required whole system.data and The position layer isinformation composed but of SAW also sensortransmits array, data data to acquisitionthe terminal module, layer. GPS The module, terminal and layer, Zigbee which and Loraincludes transceiver the control modules. terminal, As a wirelessdatabase, transceiver and the wireless module, transceiver it not only module, collects is the call requireded the client data andmodule. position Its function information is to butconnect also transmitsthe system data by wired/wireless to the terminal layer.networks, The terminalprovide access layer, which to the includescollection the of controlSAW sensor terminal, data, database, control andeach the sensor’s wireless working transceiver mode, module, and analyze is called and the detec clientt/monitor module. the Its target function gas is types to connect and information the system byon wired/wirelessconcentration in networks, the environment. provide access The human to the collectioncomputer of interaction SAW sensor interface data, control is divided each into sensor’s two workingfunctions, mode, namely, and instrument analyze and control detect/monitor operation the and target target gas gas types information and information display.on concentration in theThe environment. application The layer human is a computerserver module interaction composed interface of the is divided cloud intocomputing two functions, platform, namely, data instrumentserver, application control operationserver, and and display target platform. gas information The layer display. is commonly used to receive information fromThe the applicationcontrol terminal layer isand a serverprovide module information composed to the of thedecision-making cloud computing user platform, to determine data server,target applicationgas information server, in andthe displaydetection/monitoring platform. The enviro layer isnment. commonly Meanwhile, used to the receive database information server stores from theuseful control data, terminaland the application and provide and information data-processing to the servers decision-making analyze the user statistical to determine data. target gas informationThe total in working the detection/monitoring procedure is operated environment. as follows. Meanwhile, When the target the database gas is exposed server stores to the useful SAW data,sensor and array the in application the acquisition and data-processing layer, the detecting servers array analyze data theof frequency statistical data.shifts are outputted and collectedThe totalby the working data acquisition procedure module. is operated Afterward, as follows. the When data theare targettransferred gas is exposedto the main to the control SAW sensorcenter by array GPS in module, the acquisition Zigbee, layer,and Lora the detectingtransceiver array modules. data of After frequency processing shifts of are the outputted original data and collectedby the main by the control data acquisition center, the module. detecting Afterward, array thecurves data are transferreddisplayed toon the computer main control screen center in byreal-time. GPS module, Zigbee, and Lora transceiver modules. After processing of the original data by the main control center, the detecting array curves are displayed on computer screen in real-time. 2.4. Experiments Methods 2.4. ExperimentsThe developed Methods wireless SAW sensor array is evaluated at temperature/humidity of 19 °C and 20% TheRH. developedThe sensor wirelessarray is SAWsealed sensor in a gas array cham is evaluatedber, and the at temperature/humiditydetected gas enters the gas of 19 chamber◦C and 20%and is RH. exposed The sensor to the array sensor is arrays sealed alternately in a gas chamber, via the andgas path. the detected Meanwhile, gas entersthe personal the gas computer chamber andused is to exposed process todetection the sensor data arrays and locate alternately the detec viating the gasposition path. is Meanwhile, 300 m away. the As personal mentioned computer above, usedthe proposed to process SAW detection sensor data system and locate is composed the detecting of the position SAW issensor 300 m array, away. the As mentionedfrequency above,signal theacquisition proposed module SAW sensor (FSAM), system GPS is module, composed Zigbee of the and SAW Lora sensor transceiver array, the modules frequency (wireless signal acquisition module). moduleHence, the (FSAM), sensor GPS signal module, is collected Zigbee by and the LoraFSAM, transceiver and trasmitted modules by the (wireless wireless module). module Hence,to the thecommercial sensor signal received is collected module by in the the FSAM, PC, andand trasmittedthen recorded by the and wireless plotted module by the to thePC commercial using the receivedhomemade module software. in the PC, and then recorded and plotted by the PC using the homemade software.

3. Results

3.1. Electrical CharacterizationCharacterization of the SAW DeviceDevice The SAWSAW sensing sensing devices devices are are electrically electrically characterized characterized by a networkby a network analyzer analyzer (Figure 3(Figure). Figure 3).3 indicatesFigure 3 theindicates frequency the frequency response of response the prepared of the SAW prepared sensing SAW device sensing before device and after before SXFA and (50 after and 200SXFA nm) (50 deposition. and 200 nm) SXFA deposition. deposition notSXFA only deposition induces frequency not only shiftinduces due tofrequency the mass shift loading due but to alsothe evidentmass loading acoustic but attenuation also evident because acoustic of theattenuation viscoelastic because nature; of moreover, the viscoelastic a thick nature; SXFA ﬁlm moreover, produces a largethick attenuation,SXFA film produces which considerably large attenuation, increases which the baseline considerably noise ofincreases the sensor the system.baseline Thus, noise aof thick the polymersensor system. sensitive Thus, ﬁlm a shouldthick polyme be avoided.r sensitive film should be avoided.

Figure 3. Measured frequency response response of of the the prepared prepared SAW SAW sensing sensing device device before before and and after after SXFA SXFA deposition (Device #1—50 nm, Device #2—200 nm).

Sensors 2018, 18, x FOR PEER REVIEW 6 of 11

Sensors 2018, 18, 2977 6 of 11 3.2.Sensors Stability 2018, 18 Test, x FOR and PEER Gas REVIEW Sensor Array Experiment 6 of 11 Replicate experiments were conducted to validate the stability of the SAW sensor array. Under 3.2. Stability Test and Gas SensorSensor ArrayArray ExperimentExperiment room conditions, H2S (7.2 ppm) with carrier gas N2 was selected for detection and passed through the SAWReplicate sensor experiments experiments array sequentially were were conducted for conducted four replicate to validate to validate experiments the stability the stability (Figure of the of4). SAW the sensor SAW sensorarray. Under array. Underroom conditions, room conditions, H2S (7.2 H ppm)2S (7.2 with ppm) carrier with gas carrier N2 was gas selected N2 was selectedfor detection for detection and passed and through passed throughthe SAW the sensor SAW array sensor sequentially array sequentially for four replicate for four replicate experiments experiments (Figure 4). (Figure 4).

Figure 4. Four times responses and repeatability of sensor array to H2S. (1—sensor 1, 2—sensor 2, 3—sensor 3, 4—sensor 4). Figure 4. FourFour times times responses responses and and repeatability repeatability of of sensor sensor array array to toH2 HS.2 (1—sensorS. (1—sensor 1, 2—sensor 1, 2—sensor 2, 2, Notably,3—sensor 3,Figure 4—sensor 4 indicates 4). a typical response profile obtained from four times consecutive testing, and the sensor array shows good reproducible run. Given its TEA coating and sensitivity to Notably, Figure4 indicates a typical response proﬁle obtained from four times consecutive testing, H2S, Notably,sensor 1 yieldedFigure 4frequency indicates data a typical that show response a rapid profile increase obtained upon exposurefrom four to times H2S inconsecutive the initial severalandtesting, the andseconds sensor the arraysensor and showsthe array largest goodshows frequency reproducible good reproducib shift run. uple Givenon run. reaching Given its TEA itsthe coatingTEA adsorption coating and sensitivity andequilibrium. sensitivity to H By 2toS, sensor 1 yielded frequency data that show a rapid increase upon exposure to H2S in the initial several comparison,H2S, sensor 1 the yielded frequency frequency shifts data of the that other show three a rapid sensors increase are lower; upon exposurewhen H2 Sto is H removed2S in the byinitial N2 injection,secondsseveral seconds and the thesensor largestand array the frequency returnslargest tofrequency shift its initial upon shift reachingbaseline. upon theThe reaching adsorption same phenomenon the equilibrium.adsorption is expected equilibrium. By comparison, to occur By the frequency shifts of the other three sensors are lower; when H2S is removed by N2 injection, whencomparison, CEES, DMMP,the frequency and NH shifts3 are ofdetected the other by threethe sensor sensors array. are Theselower; promising when H2S results is removed exhibit by that N2 theinjection, sensor the arrayarray sensor returnsoffers array fast to itsreturnsresponse initial to baseline. and its initial exce Thellent baseline. same repeatability phenomenon The same in detectingphenomenon is expected target tois gases. occurexpected when to CEES,occur DMMP, and NH3 are detected by the sensor array. These promising results exhibit that the sensor whenAfter CEES, the DMMP, above andexperiment, NH3 are detecteddifferent by concentrations the sensor array. of TheseH2S, CEES, promising DMMP, results and exhibit NH3 thatare detectedarraythe sensor offers with array fast the responseoffers fabricated fast and response sensor excellent andarray. repeatability exce Beforellent repeatability experiments, in detecting in targetthe detecting sensor gases. targetarray gases.tends to stabilize After the above experiment, different concentrations of H S, CEES, DMMP, and NH are detected after After20 min, the frequency above experiment, stability can different reach 30concentrations Hz h−1 in the 2of stable H2S, status,CEES, DMMP,and the resultsand3 NH of3 theare withdetectiondetected the fabricated withof the the four fabricated sensor gases array. aresensor Beforeshown array. experiments, in BeforeFigure experiments,5. the As sensor can be arraythe concluded sensor tends toarray using stabilize tends the after tofigure, stabilize 20 min,the −1 responsesfrequencyafter 20 min, stabilityfor frequencythe four can reachgases stability 30 are Hz sensitivecan h reachin the at 30 stable lo Hzw concentrations; status,h−1 in andthe thestable resultsmoreover, status, of theand the detection thesensor results ofarray the of fourcanthe gases are shown in Figure5. As can be concluded using the ﬁgure, the responses for the four gases providedetection theof thesignals four ofgases 165, are 217, shown 87, andin Figure110 Hz 5. Asfor canH2S, be CEES, concluded DMMP, using and the NH figure,3 at the concentrationsareresponses sensitive for at theof low 1.08, four concentrations; 0.59, gases 0.1, are and sensitive moreover, 5.02 ppm, at the relowspectively. sensor concentrations; array These can environmental provide moreover, the signalsthe concentrations sensor of 165, array 217, canare 87, and 110 Hz for H S, CEES, DMMP, and NH at the concentrations of 1.08, 0.59, 0.1, and 5.02 ppm, consideredprovide the safe signals for2 humans.of 165, 217,On the87, basisand 3of110 the Hz results, for H 2weS, CEES,suggest DMMP, that the and sensor NH 3array at the is promisingrespectively.concentrations for These use of 1.08,in environmental gas 0.59, detection. 0.1, andconcentrations 5.02 ppm, respectively. are considered These safe environmental for humans. Onconcentrations the basis of theare results,considered we suggestsafe for that humans. the sensor On arraythe basis is promising of the forresults, use inwe gas suggest detection. that the sensor array is promising for use in gas detection. 450 CEES H S 1000 2 400450 CEES 800 H S 1000 2 350400

600 800 300350

400 600 Frequency Shifts (Hz) Shifts Frequency 250300 Frequency Shifts (Hz) Shifts Frequency

200 400

Frequency Shifts (Hz) Shifts Frequency 200250

Frequency Shifts (Hz) Shifts Frequency 0246810121416 0 5 10 15 20 25 30 Concentration (ppm) 200 Concentration (ppm) 200 0246810121416 0 5 10 15 20 25 30 (a) Curve of frequency shift toward H2S gas (b) Curve of frequencyConcentration shift (ppm)toward CEES gas Concentration (ppm)

Figure 5. Cont. (a) Curve of frequency shift toward H2S gas (b) Curve of frequency shift toward CEES gas

SensorsSensors 20182018, ,1818, ,x 2977 FOR PEER REVIEW 77 of of 11 11

500 1200

DMMP NH3 450 1000 400

350 800

300 600 250

200 400

150 (Hz) Shifts Frequency Frequency Shifts (Hz) 200 100

50 0 012345 0 1020304050 Concentration (ppm) Concentration (ppm)

FigureFigure 5. 5. ResponseResponse of of the the SAW SAW sensor sensor array array to to different different concentration concentration gases. gases.

3.3.3.3. Establishment Establishment of of Sensor Sensor Data Data Transmission Transmission Model Model ReliableReliable transmission modelsmodels are are essential essential to ensureto ensure the correctnessthe correctness and effectiveness and effectiveness of wireless of wirelessdata. Based data. onBased Zigbee on Zigbee technology, technology, a WSN a shouldWSN should consist consist of a base of a station base station and abundant and abundant nodes. nodes.The microprocessor The microprocessor processes processes the raw the data raw and data sends and datasends to data the adjacentto the adjacent nodesby nodes a wireless by a wirelesstransceiver transceiver module. module. Through Through the nodes the of thenodes sensor of the network, sensor data network, are transmitted data are leveltransmitted by level level until byit islevel sent until to the it base is sent station. to the Finally, base datastation. are Finally, transmitted data toare the transmitted host by the to base the station host by through the base the stationserial portthrough to successfully the serial port monitor to successfully the location monitor of the control.the location of the control. TheThe commonly commonly used networknetwork transmissiontransmission models models include include a stara star structure, structure, a mesh a mesh structure, structure, and anda hybrid a hybrid network. network. Each Each network network structure structure exhibits exhi itsbits own its advantages own advantages and defects. and defects. In fact, differentIn fact, differentnetwork network topologies topologies should be should selected be according selected toaccording different applicationto different requirements.application requirements. Star topology Staris a topology single-hop is systema single-hop that supports system point-to-pointthat supports andpoint-to-point point-to-multipoint and point-to-multipoint communication. communication.The structure requires The structure low power requires consumption, low power but its consumption, transmission distancebut its betweentransmission node distance and base betweenstation is node limited. and The base transmission station is distancelimited. ofThe the transmission frequency band distance is usually of the in meters. frequency Star topologyband is usuallyis suitable in meters. for the Star network topology of round is suitable equipment for the near network distance. of round Mesh equipment topology is near a multi-hop distance. system. Mesh topologyBy using multi-hopis a multi-hop routing system. communication, By using multi-hop all wireless routing sensor nodescommunication, perform identical all wireless roles, notsensor only nodescommunicate perform directly identical with roles, each not other only but communicate also conducting directly data transmissionwith each other and but mutual also transmissionconducting datacommands transmission with the and central mutual node. transmission Given that eachcommands sensor nodewith possessesthe central multiple node. pathsGiven to that the centereach sensornode or node other possesses nodes, the multiple network paths exhibits to the strong center robustness node or other and reliability. nodes, the The network mesh topologyexhibits strong system robustnessreaches farther and reliability. transmission The distance mesh topology than the system star-type reaches structure farther and transmission is suitable distance for large than physical the star-typespace-node structure dispersed and systems. is suitable However, for large this physic systemal space-node requires large dispersed power consumption. systems. However, this systemGiven requires the transmissionlarge power consumption. distance requirement in this project, the system adopts mesh-network topology,Given andthe transmission its network architecturedistance requirement is shown in in this Figure project,6. Figure the system6a is the adopts sketch mesh-network map of the topology,established and wireless its network communication architecture module. is shown In Figure in Figure6b, the 6. GPS Figure positioning 6a is the module sketch (yellow map of point the establishedin right) is designedwireless incommunication the node, transmitting module. the In locationFigure and6b, the detection GPS positioning data to the mainmodule control (yellow node point(blue in point right) in left)is designed and meeting in the the node, required transmi communicationtting the location distance and (300 detection m). data to the main control node (blue point in left) and meeting the required communication distance (300 m). 3.4. Actual Wireless Communication of the SAW Sensor Array

Actual wireless communication experiments were conducted within 300 m outdoors. When H2S, CEES, DMMP, and NH3 passed through the SAW sensor array with the concentration of 10.76, 2.94, 2.96 and 21.53 ppm, respectively, response shifts are recorded by the SAW sensor array and transmited to the main control node (PC computer with Zigbee and Lora controllers), and the graphs formed by the detected data and testing position are visible on the screen of the PC computer in real time. The detected results are shown in Figure7.

Sensors 2018, 18, x FOR PEER REVIEW 8 of 11

系统架构图

主控Main Control 中心Center ZigBee

Sensors 2018, 18, 2977 8 of 11

Sensors 2018, 18, x FOR PEER REVIEW 8 of 11

Terminal终端 Terminal系统架构图终端 Terminal终端 Sensor Array Sensor Array Sensor Array 传感器阵列传感器阵列传感器阵列 …… MainGPS GPS GPS 主控GPS GPS GPS Control 中心Center ZigBee (a) (b)

Figure 6. (a) Sketch map of the wireless communication module used in this study; (b) Position located by GPS in practical experiments.

Terminal终端 Terminal终端 Terminal终端 3.4. Actual Wireless Communication of the SAW Sensor Array Sensor Array Sensor Array Sensor Array 传感器阵列传感器阵列传感器阵列 …… Actual wireless communication experiments were conducted within 300 m outdoors. When GPS GPS GPS GPS GPS GPS H2S, CEES, DMMP, and NH3 passed through the SAW sensor array with the concentration of 10.76, 2.94, 2.96 and 21.53 ppm, respectively, response shifts are recorded by the SAW sensor array and (a) transmited to the main control node (PC computer with Zigbee and( bLora) controllers), and the graphsFigure formed 6. ((aa)) SketchbySketch the map mapdetected of of the the wireless datawireless and communication communication testing position module module usedare used invisible this in study; thison study;the (b) Positionscreen (b) Position locatedof the PC computerlocatedby GPS in by in real practicalGPS time. in practical experiments.The detect experiments.ed results are shown in Figure 7.

3.4. Actual Wireless Communication of the SAW Sensor Array Actual wireless communication experiments were conducted within 300 m outdoors. When H2S, CEES, DMMP, and NH3 passed through the SAW sensor array with the concentration of 10.76, 2.94, 2.96 and 21.53 ppm, respectively, response shifts are recorded by the SAW sensor array and transmited to the main control node (PC computer with Zigbee and Lora controllers), and the graphs formed by the detected data and testing position are visible on the screen of the PC computer in real time. The detected results are shown in Figure 7.

(a) Response of SAW sensor array based on TEA (b) Response of SAW sensor array based on PECH sensitive layer toward H2S sensitive layer toward CEES

(c) Response of SAW sensor array based on SXFA (d) Response of SAW sensor array based on GAH sensitive layer toward DMMP sensitive layer toward NH3

Figure 7. Detection results of the SAW sensorsensor array to different gases.gases.

Figure7 indicates the whole processes when the four kinds of gases are detected using the developed SAW sensor array, and the variations are evident when different gases are detected. For example, when H2S gas passes through the sensor array, sensor 1 exhibits the largest signal response and the fastest response speed to H2S, whereas the other three sensors show relatively long response(c) Response times of andSAW small sensor response array based signals. on SXFA This ﬁnding(d) Response is attributed of SAW to thesensor TEA array coating based of on sensor GAH 1, which is speciﬁcsensitive and layer sensitive toward to DMMP H2S; thus, the H2S molecules cansensitive be easily layer absorbed toward NH on3 the surface of

Figure 7. Detection results of the SAW sensor array to different gases.

Sensors 2018, 18, x FOR PEER REVIEW 9 of 11

Figure 7 indicates the whole processes when the four kinds of gases are detected using the developed SAW sensor array, and the variations are evident when different gases are detected. For example, when H2S gas passes through the sensor array, sensor 1 exhibits the largest signal response and the fastest response speed to H2S, whereas the other three sensors show relatively long response times and small response signals. This finding is attributed to the TEA coating of Sensors 2018, 18, 2977 9 of 11 sensor 1, which is specific and sensitive to H2S; thus, the H2S molecules can be easily absorbed on the surface of TEA and rapidly reach equilibrium. Although the other three sensors can respond to

TEAH2S, andthe rapidlyresponse reach signals equilibrium. and speeds Although are highly the otherdifferent; three evidently, sensors can sensor respond 2 (coated to H2S, with the responsePECH to signalsdetect CEES) and speeds of the are array highly shows different; the second evidently, largest sensor signal 2 response, (coated with followed PECH by to sensor detect CEES)3 and sensor of the array4 with shows the third the second and fourth largest largest signal signal response, respon followedse, respectively. by sensor 3 Therefore, and sensor the 4 with difference the third ofand the fourthfour sensors largest in signal detecting response, H2S is respectively. determined, Therefore, and the fingerprint the difference of the of sensory the four array sensors to detect in detecting H2S is Hobtained.2S is determined, Similar findings and the are fingerprint observed of in the the sensory other three array sensors to detect when H2S isNH obtained.3, CEES,Similar and DMMP findings are aredetected, observed and in the the fingerprints other three sensorsare shown when in NHFigure3, CEES, 8. and DMMP are detected, and the fingerprints are shown in Figure8.

350 350 Detection of H2S Detection of CEES 300 300

250 250

200 200 Hz Hz 150 150

100 100

50 50

0 0 Sensor 1 Sensor 2 Sensor 3 Sensor 4 Sensor 1 Sensor 2 Sensor 3 Sensor 4

(a) Fingerprint of H2S (b) Fingerprint of CEES

Deteciton of DMMP 400 Detection of NH3 400 350

300 300 250

Hz 200 Hz 200 150

100 100 50

0 0 Sensor 1 Sensor 2 Sensor 3 Sensor 4 Sensor 1 Sensor 2 Sensor 3 Sensor 4

Figure 8. Fingerprints of the four kinds of gases detecteddetected byby SAWSAW sensorsensor array.array.

WithWith thethe preparationpreparation ofof wirelesswireless communicationcommunication of the SAW sensorsensor array, HH22S, CEES, DMMP, andand NHNH33 can be be detected, detected, and and the the dete detectedcted data data can can be be transferred transferred within within 300 300 m. m.Meanwhile, Meanwhile, the theresponse response signals signals and and response response time time are arehighly highly different different when when different different gases gases are aredetected detected by the by theSAW SAW sensor sensor array, array, and and the the correspon correspondingding fingerprints fingerprints vary vary owing owing to to film film qualities. The detecteddetected gasesgases cancan bebe identifiedidentified byby patternpattern recognitionrecognition algorithm.algorithm.

4.4. ConclusionsConclusions WSNWSN isis widelywidely usedused inin small-scalesmall-scale environmentalenvironmental monitoring.monitoring. WithWith thethe evolutionevolution ofof computercomputer systems,systems, monitoringmonitoring combined combined with with wireless wireless communication communication networks networks is one of theis bestone approachesof the best to evaluateapproaches the changeto evaluate of environment the change inof theenvironment future. As thein the best future. technology As the to best provide technology context awareness,to provide thecontext detecting awareness, platform the based detecting on wireless platform communication based on wireless include numerouscommunication monitor include nodes tonumerous monitor the detected gases [29,32]. Considering the lack of an objective method for determining whether harmful gases leak into large-scale environment, we present a new wireless SAW-based chemical sensor array composed of TEA, PECH, SXFA, and GAH. This approach in detecting H2S, CEES, DMMP, and NH3 by wireless communication technology is also demonstrated in our study. The method indicates numerous special features of the designed sensor array, including the design of an efficient WSN for monitoring, low power consumption, Sensors 2018, 18, 2977 10 of 11 data availability, confidentiality, and easy remote configuration. This monitoring system presented in this paper can easily determine other parameters of interest, such as sensing activity, transmission from sensor nodes to base station, data storage, or visualization. Therefore, the developed approach may be extensively used in determining harmful gases for monitoring a given area. The response of the developed SAW sensor array toward H2S, CEES, DMMP, and NH3, as well as repeatability, is characterized in gas experiments. With the wireless communication network module, H2S, CEES, DMMP, and NH3 are detected within 300 m under the environmentally safe concentrations of 1.08, 0.59, 0.10, and 5.02 ppm, respectively, the sensitivities of different sensors of the sensor array are 32.4, 14.9, 78.1 and 22.6 Hz/ppm, respectively, and the testing location is determined by the GPS module integrated into the sensor array. Although wireless communication is a considerable challenge in sensing and substantial work should be focused on this aspect, the current system that we developed demonstrated the feasibility of the WSN sensor array and its possibility in application. The future work will focus on enhancing the performance parameters of the device and the system, and the database should be established to improve recognition accuracy.

Author Contributions: All authors participated in the work presented here. Y.P. and W.W. proposed the original idea and defined the research topic; N.M. and B.C. conducted most of the experiments; and B.L. participated in the experiment and analyzed the data. Y.P. and L.Y. conceptualized, designed, and performed the method Acknowledgments: This research was funded by National Natural Science Foundation of China. (Grant Nos. 10834010 and 11074268). Conflicts of Interest: The authors declare no conflict of interest.

References

1. Daniel, L.A.F.; Tony, A.R. A new analytical system, based on an acoustic wave sensor for halitosis evaluation. Sens. Actuators B 2009, 136, 73–79. 2. Chi, Y.S.; Shih, Y.L. Surface acoustic wave gas monitor for ppm ammonia detection. Sens. Actuators B 2008, 131, 673–679. 3. Byung, S.J.; Jeung, S.H. Fabrication of polymer SAW sensor array to classify chemical warfare agents. Sens. Actuators B 2007, 121, 47–53. 4. Alizadeh, T.; Zeynali, S. Electronic nose based on the polymer coated SAW sensors array for the warfare agent simulants classification. Sens. Actuators B 2008, 129, 412–423. [CrossRef] 5. Jurs, P.C.; Bakken, G.A. Computational methods for the analysis of chemical sensor array data from volatile analytes. Chem. Rev. 2000, 100, 2649–2678. [CrossRef][PubMed] 6. Tasaltin, C.; Gurol, I.; Harbeck, M. Synthesis and DMMP sensing properties of fluoroalkyloxy and fluoroaryloxy substuted phthalocyanines in acoustic sensors. Sens. Actuators B 2010, 150, 781–787. [CrossRef] 7. Matatagui, D.; Marti, J.; Fernandez, M.J. Chemical warfare agents simulants detection with an optimized SAW sensor array. Sens. Actuators B 2011, 154, 199–205. [CrossRef] 8. Matatagui, D.; Fernandez, M.J.; Forntecha, J. Love-wave sensor array to detect, discriminate and classify chemical warfare agent simulants. Sens. Actuators B 2012, 175, 173–178. [CrossRef] 9. Jagannath, D.; Paul, R.O.; David, W.G. SAW sensors for chemical vapors and gases. Sensors 2017, 17, 801. 10. Horrllo, M.C. Detection of volatile organic compounds using surface acoustic wave sensors with different polymer coatings. Thin Solid Films 2004, 467, 234–238. [CrossRef] 11. Shen, C.Y.; Huang, C.P.; Chuo, H.C. The improved ammonia gas sensors constructed by L-glutamic acid hydrochloride on surface acoustic wave devices. Sens. Actuators B Chem. 2002, 84, 231–236. [CrossRef] 12. Wang, W.; He, S.T.; Li, S.Z. Advances in SXFA-coated SAW chemical sensors for organophosphorous compound detection. Sensors 2011, 11, 1526–1541. [CrossRef][PubMed] 13. Liu, W.W.; Yu, J.H.; Pan, Y. Study on the detection of blister agent mustard by surface acoustic wave technology. FenXi CeShi XueBao 2006, 25, 80–82. 14. Clifford, K.H.; Eric, R.L.; Rawlinson, K.S. Development of a surface acoustic wave for in-situ monitoring of volatile organic compounds. Sensors 2003, 3, 236–247. 15. Boquan, L.; Zhang, C.; Ji, X. An improved performance frequency estimation algorithm for passive wireless SAW resonant sensors. Sensors 2014, 14, 22261–22273. Sensors 2018, 18, 2977 11 of 11

16. David, W.G.; Chin, T.L.; Peng, Z. Surface acoustic wave devices for harsh environment wireless sensing. Sensors 2013, 13, 6910–6935. 17. Ahnsei, S.; Jun, C.; Jiuk, J. An implantable wireless neural interface system for simultaneous recording and stimulation of peripheral nerve with a single cuff electrode. Sensors 2018, 18, 1. [CrossRef] 18. Jegadeesan, R.; Nag, S.; Agarwal, K.; Thakor, N. Enabling wireless powering and telemetry for peripheral nerve implants. IEEE J. Biomed. Health Inform. 2015, 19, 958–970. [CrossRef][PubMed] 19. Lu, S.; Huang, X.; Cui, L.; Zhao, Z.; Li, D. Design and implementation of an ASIC-based sensor device for WSN applications. IEEE Trans. Consum. Electron. 2009, 55, 1959–1967. [CrossRef] 20. Harnett, C.K. Open wireless sensor network telemetry platform for mobile phones. IEEE Sens. J. 2010, 10, 1083–1084. [CrossRef] 21. Tang, K.T.; Li, C.H. An electronic-nose sensor node based on a polymer-coated surface acoustic wave array for wireless sensor network applications. Sensors 2011, 11, 4609–4621. [CrossRef][PubMed] 22. Xu, F.Q.; Wang, W.; Xue, X.F. Development of a wireless and passive SAW-based chemical sensor for organophosphorous compound detection. Sensors 2015, 15, 30187–30198. [CrossRef][PubMed] 23. Lim, C.-B.; Wang, W.; Yang, S.S.; Lee, K.K. Development of a wireless, battery-free SAW volatile organic compounds sensor integrated with temperature sensor. Sens. Lett. 2011, 9, 82–86. [CrossRef] 24. Lim, C.; Wang, W.; Yang, S.S.; Lee, K.K. Development of SAW-based multi-gas sensor for simultaneous

detection of CO2 and NO2. Sens. Actuators B Chem. 2011, 154, 9–16. [CrossRef] 25. Lieberzeit, P.A.; Christian, P.; Dickert, F.L.; Gerhard, F. SAW RFID-tags for mass-sensitive detection of humidity and vapors. Sensors 2009, 9, 9805–9815. [CrossRef][PubMed] 26. Petar, K.; Matthew, D.; Ivana, M.S. Wireless chemical sensors and biosensors: A review. Sens. Actuators B 2018, 266, 228–245. 27. Olivier, H.; Jean-Philippe, P.; Christine, M. A review of algorithms for SAW sensors e-nose based volatilecompound identiﬁcation. Sens. Actuators B 2018, 255, 2472–2482. 28. Wu, T.-T.; Lin, C.-M. Analysis of SFIT SAW ﬁlter using coupling of modes model. J. Chin. Inst. Eng. 2004, 27, 973–979. [CrossRef] 29. Marlon, N.; José, C.; Alberto, B. GTSO: Global trace tynchronization and ordering mechanism for wireless sensor network monitoring platforms. Sensors 2018, 18, 28. 30. Giovanni, P.; Claude, C.; Dixian, Z. Next generation wireless technologies for internet of things. Sensors 2018, 18, 221. 31. He, M.L.; Shen, W.B.; Pan, Y.J. Temporal–spatial surface seasonal mass changes and vertical crustal deformation in south China block from GPS and GRACE measurements. Sensors 2018, 18, 99. [CrossRef] [PubMed] 32. Siddhartha, B.; Neil, B.; Raja, J. Time series analysis for spatial node selection in environment monitoring sensor networks. Sensors 2018, 18, 11.

© 2018 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/).