Nano-Sheet-like Morphology of Nitrogen-Doped Graphene-Oxide-Grafted Manganese Oxide and Polypyrrole Composite for Chemical Warfare Agent Simulant Detection

Chemical warfare agents (CWAs) have inflicted monumental damage to human lives from World War I to modern warfare in the form of armed conflict, terrorist attacks, and civil wars. Is it possible to detect the CWAs early and prevent the loss of human lives? To answer this research question, we synthesized hybrid composite materials to sense CWAs using hydrothermal and thermal reduction processes. The synthesized hybrid composite materials were evaluated with quartz crystal microbalance (QCM) and surface acoustic wave (SAW) sensors as detectors. The main findings from this study are: (1) For a low dimethyl methyl phosphonate (DMMP) concentration of 25 ppm, manganese dioxide nitrogen-doped graphene oxide (NGO@MnO2) and NGO@MnO2/Polypyrrole (PPy) showed the sensitivities of 7 and 51 Hz for the QCM sensor and 146 and 98 Hz for the SAW sensor. (2) NGO@MnO2 and NGO@MnO2/PPy showed sensitivities of more than 50-fold in the QCM sensor and 100-fold in the SAW sensor between DMMP and potential interferences. (3) NGO@MnO2 and NGO@MnO2/PPy showed coefficients of determination (R2) of 0.992 and 0.975 for the QCM sensor and 0.979 and 0.989 for the SAW sensor. (4) NGO@MnO2 and NGO@MnO2/PPy showed repeatability of 7.00 ± 0.55 and 47.29 ± 2.69 Hz in the QCM sensor and 656.37 ± 73.96 and 665.83 ± 77.50 Hz in the SAW sensor. Based on these unique findings, we propose NGO@MnO2 and NGO@MnO2/PPy as potential candidate materials that could be used to detect CWAs.


Introduction
Chemical warfare agents (CWAs) are considered extremely lethal weapons that have the potential to cause mass destruction [1]. The CWA attacks on the Tokyo subway (1995) [2] and the Syrian civil war (2013) [3] demonstrated to the world the ramifications of the misuse of CWAs. Based on this, we understand that the early detection of CWAs is crucial for preventing the loss of human lives. Figure S1 shows the chemical structure of G-series nerve agents, such as sarin (GB series), soman (GD series), Tabun (GA series), and simulant DMMP. For many years, detection technologies have been suggested and developed to detect CWAs. Some of these include gas chromatography, Raman spectroscopy, Fourier-Transform Infrared Spectroscopy, atomic emission detection, and ion mobility spectroscopy [4,5]. This equipment is used for the detection of both CWAs and their simulants, and it is accurate, precise, and reliable.
However, this equipment has crucial problems when used in field monitoring of CWAs. For practical purposes, it is important to detect CWAs in the field. For the field monitoring and detection of CWAs, devices with less power consumption, high accuracy, portability, reliability, inexpensive manufacturing and maintenance costs, and stability and the requirement of more power [26]. In the case of liquid measurements, the main disadvantage is the reflection of longitudinal waves in the SAW sensor [27].
In the pursuit of discovering highly sensitive functionalized material for the detection of different vapors, carbon nanotubes [28], metal oxides [29], conducting polymers [6,30], and metal organic frameworks [31] have been developed thus far. Manganese dioxide is one of the materials used in the detection of CWAs that has applications in gas sensors [32], energy storage, water treatment, batteries, supercapacitors, and decomposition. PPy is a conducting polymer that has affinity towards gas sensors and that has been used with its hybrid composites to detect DMMP (M.W. = 124.08 g/mol), methanol, ammonia, hydrogen sulfide, nitrogen dioxide, and carbon dioxide [33,34].
It was reported that the incorporation of ZnO with manganese dioxide effectively absorbed DMMP vapor, producing significant data for CWA simulant detection [35]. Another approach for CWA simulant detection is the use of hybrid composite materials containing PPy that showed excellent sensor responses [36]. Another study [37] on graphene oxide demonstrated that it generated remarkable sensitivity, repeatability, and stability for CWA simulant detection. Based on the Tanimoto coefficient and Euclidean distance [38], DMMP is one of the preferred simulants of the GB nerve agents. Figure S2 illustrates the chemical structure of the synthesized hybrid composite materials.
In this study, we present the characterization of hybrid composite materials and study their morphology using Fourier-Transform Infrared Spectroscopy (FTIR), X-ray Diffraction (XRD), X-ray Photoelectron Spectroscopy (XPS), Scanning Electron Microscopy (SEM), and Transmission Electron Microscopy (TEM). NGO@MnO 2 and NGO@MnO 2 /PPy were deposited in the QCM and SAW sensors to test their frequency shifts, selectivity, polar plots, linearity, repeatability, and response/recovery times. The sensing performance of the fabricated composite materials was analyzed and compared for both the sensors. The effect of R.H. on the sensing performance was also investigated. Note that the used acronyms in this paper are listed in Table S1.

Synthesis of NGO
An Improved Hummers method was used to synthesize GO materials. This method is reported in the literature [39]. The calculated amount of graphite flakes (30 g) and NaNO 3 (7 g) were placed together with H 2 SO 4 (300 mL) at a temperature of 0 • C in an ice bath. Next, 6 g of potassium permanganate (KMnO 4 ) was added slowly to the mix kept in the bath, and the temperature of the bath solution was increased to 50 • C with 12 h of continuous stirring. At the end of the process, the color of the resultant product turned greenish brown. Subsequently, filtration of the resultant product was completed, and the product was rinsed with water and ethanol several times. Hence, the obtained GO was heated for 24 h at 50 • C in the vacuum oven.
In order to prepare NGO, the prepared GO (3 g) was dissolved into water (300 mL) and sonicated for 4 h to allow for uniform dispersion of the GO. After proper dispersion, the solution was filtered using a Millipore filter. The filtered solution was then dried for 12 h at 95 • C in an oven. The filtered GO solution was mixed with ammonia and urea (10 mL) together with continuous stirring for 12 h at 95 • C to acquire a homogenous solution. The solution obtained out of this process thus far was dried in an oven for 12 h at 180 • C and then purified further using an ethanol solvent. As the last step, the resultant product was calcined for 12 h at 200 • C.

NGO@MnO 2 Composites
In order to prepare the NGO@MnO 2 composites, a calculated amount 0.6 g of asprepared NGO (0.01 M), manganese acetate and KMnO 4 (0.01 M) were dissolved in doubledistilled (DD) water in a glass vessel of 500 mL. The ammonia solution (25 mL) was introduced into the solution in the glass vessel and stirred continuously for 12 h at 95 • C. The solution obtained after stirring was then transferred to a Teflon-lined autoclave (300 mL) and roasted for 12 h at 180 • C. After the roasting was completed, the resulting product was purified using Millipore water and dried for 12 h at 95 • C. At the end of the drying process, the collected composite sample was calcinated for 10 h at 450 • C in a vacuum furnace.

NGO@MnO 2 /PPy Composite Synthesis
The hybrid composite was prepared using the following procedure. First, the asprepared NGO (0.6 g), manganese acetate (0.01 M), KMnO 4 (0.01 M), pyrrole (6 g), and FeCl 3 (1.5 g) and H 2 O 2 (10 mL) were blended into 100 mL of DD water in a glass vessel. The solution obtained after blending was continuously stirred for 4 h at room temperature to acquire a homogenous dispersion. Subsequent to this stirring, the solution was allowed to undergo in situ oxidative polymerization for 2 h at 90 • C. After the polymerization was finished, the solution was transferred to a Teflon-lined autoclave and heated there for 10 h at 450 • C. After heating, the resulting product was collected and used in our experiments without further processing. Figure 1 shows the schematic diagram illustrating the fabrication process of NGO@MnO 2 /PPy.   Figure 2a illustrates the bubbler flask used in the presented study. The 500 mL bubbler flasks were cleaned using ethanol for several times and dried in an oven for one hour at 60 °C. After the chemicals were received, they were poured into the pre-cleaned bubbler flask with the help of a pre-cleaned glass funnel. Then, two steel tubes attached with Teflon cork were inserted into the inlet and outlet of the bubbler. The steel tubes allowed nitrogen and air to flow through the inlet to outlet of the bubbler. The Teflon cork made the bubbler system airtight to prevent the leakage of generated target vapor. After the bubbler flask was used in the experiments, it was stored in a Chemsafe systems Storage  Figure 2a illustrates the bubbler flask used in the presented study. The 500 mL bubbler flasks were cleaned using ethanol for several times and dried in an oven for one hour at 60 • C. After the chemicals were received, they were poured into the pre-cleaned bubbler flask with the help of a pre-cleaned glass funnel. Then, two steel tubes attached with Teflon Nanomaterials 2022, 12, 2965 5 of 23 cork were inserted into the inlet and outlet of the bubbler. The steel tubes allowed nitrogen and air to flow through the inlet to outlet of the bubbler. The Teflon cork made the bubbler system airtight to prevent the leakage of generated target vapor. After the bubbler flask was used in the experiments, it was stored in a Chemsafe systems Storage Cabinet (Samillab, AL-D 1002).

Sensor Measurement System
The gas response measurement system is divided into three major parts-namely, vapor-generating bubbler, analyte delivery system, and detection chamber and data ac quisition. In our previous works [6,40], we briefly introduced the gas response measure ment system. In this presented study, each part will be explained in detail.

Vapor-Generating Bubbler
A two-necked bubbler was used to vaporize the liquid simulant and VOCs for th gas response measurements. Dry N2 gas for QCM (dry air for SAW) acting as a carrier ga enters into the base of the target vapor containing bubbler through the left inlet steel tube Then, the carrier gas is distributed all over the target liquid sample, which takes the ana lyte vapor upward and finally through the right outlet steel tube. Figure 2b shows th schematic of the vapor generating process in the presented work. The output flow rate o analyte vapor, Fs, can be formulated using a bubbler equation [41]: where Fc is the flow rate of carrier gas in sccm (standard cubic centimeters per minute), P is the outlet pressure in the bubbler headspace above the target liquid, and Pth is the ther modynamic vapor pressure of the target analyte.
The vapor pickup efficiency (α) is defined as the relative ratio of the output targe vapor flow rate by the carrier gas flow rate. These parameters were acquired under th standard conditions of T = 298 K and P = 760 mmHg (=1 atm) in the presented study. P converts to Ps (saturated vapor pressure) when the carrier gas is totally saturated with th

Sensor Measurement System
The gas response measurement system is divided into three major parts-namely, a vapor-generating bubbler, analyte delivery system, and detection chamber and data acquisition. In our previous works [6,40], we briefly introduced the gas response measurement system. In this presented study, each part will be explained in detail.

Vapor-Generating Bubbler
A two-necked bubbler was used to vaporize the liquid simulant and VOCs for the gas response measurements. Dry N 2 gas for QCM (dry air for SAW) acting as a carrier gas enters into the base of the target vapor containing bubbler through the left inlet steel tube. Then, the carrier gas is distributed all over the target liquid sample, which takes the analyte vapor upward and finally through the right outlet steel tube. Figure 2b shows the schematic of the vapor generating process in the presented work. The output flow rate of analyte vapor, F s , can be formulated using a bubbler equation [41]: where F c is the flow rate of carrier gas in sccm (standard cubic centimeters per minute), P 0 is the outlet pressure in the bubbler headspace above the target liquid, and P th is the thermodynamic vapor pressure of the target analyte. The vapor pickup efficiency (α) is defined as the relative ratio of the output target vapor flow rate by the carrier gas flow rate. These parameters were acquired under the standard conditions of T = 298 K and P = 760 mmHg (=1 atm) in the presented study. P th Nanomaterials 2022, 12, 2965 6 of 23 converts to P s (saturated vapor pressure) when the carrier gas is totally saturated with the target analyte vapor. P s of the target vapor can be calculated from the Antoine Equations (2) and (3) expressed as below [42]: where T is the temperature of the bubbler controlled by a digital water bath (LabTech, LWB-106D). The value of P s is greatly dependent on the temperature of target liquid containing bubbler controlled by the digital water bath. The main reason is that the output flow rate was calculated under the fixed bubbler temperature and constant carrier gas flow rate. A, B, and C are the coefficients used in the Antoine equation for various vapors, which is given in Table S2.

Analyte Delivery System
The generated target vapor was transported from the bubbler headspace to the detection chamber. The gas feeding system consisted of nitrogen cylinders connected to the mass flow controllers (MFCs) (KOFLOG, 3660), one-way valves, a vapor bubbler system, steel pipes, and one-touch connectors as shown in Figures S3 and S4. The carrier gas flow was controlled by setting the MFC 2 flow rate. The generated target vapor was mixed with the dilution gas flow, F d (2000 sccm in QCM and 1000 sccm in the SAW sensor), by setting the MFC 1. The resultant target vapor concentration, C vapor , in ppm (parts per million) was enumerated from the predetermined output flow rate and the dilution ratio [43]: In the present study, four VOCs (ethanol, methanol, n-hexane, and Toluene) and water were detected using a fixed carrier flow rate (200 sccm in QCM and 100 sccm in SAW) and a bubbler temperature of 25 • C. As the objective of this study was to detect DMMP, the carrier flow rate was varied from 48 to 320 sccm in QCM and 24 to 160 sccm in the SAW sensor with a fixed bubbler temperature of 25 • C. The dilution flow rate was kept constant most of the time (2000 sccm in QCM and 1000 sccm in the SAW sensor). In both the QCM and SAW sensors, the target vapor was fed to the detection chamber alternately for 5 min. The dilution flow was continuously fed to the detection chamber at the fixed flow rate for enough time to reach the sensor steady baseline.

Detection Chamber and Data Acquisition
In the presented work, a QCM controller system (QCM200, Stanford Research System, Sunnyvale, CA, USA) was used to measure the sensing performance of the hybrid composites (See Figure S5a). Regarding the detection chamber or flow cell, after the installation of axial flow cell adapter, a stagnation point flow cell with a small volume was created. The volume of flow chamber was around 0.15 mL. Figure S5b shows the QCM flow cell used in the presented study. The flow cell is built using the chemical resistant Kynar.
It consists of two ports-inlet and outlet-with a 0.040 inch through-hole equipped with barbed hose adapters with 0.062 inch tubing. The flow cell utilizes the axial flow for the adsorption and desorption processes. In the axial flow cell, the gas flows inwards perpendicularly and radially outward from the stagnation point to the exit at the edge of the cell in a volume of 150 µL. The stagnation point is positioned at the center of the QCM crystal electrode where there is zero hydrodynamic flow or absence of surface shear forces [44]. Figure S5d illustrates the flow pattern in an axial flow cell. A QCM200 Digital controller displays the measured resonant frequency of the QCM crystal at the interval of 1 s, which is transferred to a personal computer. The data acquisition was done by software called SRS QCM200 (see Figure S5c).
The SAW sensor setup used in the presented study was developed in the laboratory clean booth. It consists of a detection chamber, test board, vector network analyzer (VNA-MS46122A, Anritsu, Richardson, TX, USA), and SMA connectors (see Figure S6a). Regarding the detection chamber, it has a volume of 420 mL. Figures S6b and S5c show the top and front views of the detection chamber without and with the top cover, respectively. The SAW sensor detection chamber applies a similar principle with the QCM flow cell.
The gas flows perpendicular towards the test board, which holds the SAW sensor, and exits radially outward from the edge of the chamber. Figure S6d shows an illustration of the gas flow in the SAW detection chamber. VNA was used to transfer the raw sensor signal to the personal computer. The data acquisition was done at intervals of 10 s with the help of software called Shockline. Figure S6e shows the SAW sensor signal monitoring program.
Gas response measurements were performed by positioning the hybrid-compositecoated QCM and SAW sensors inside the flow cell or detection chamber and blowing the diluted target vapor over them while concurrently monitoring the frequency changes of the sensors. All the experiments were conducted inside the clean booth in which the temperature and humidity was regulated with the air-conditioned environment maintained at 20 • C (LG Whisen-LPNW1451VJ, LG, Seoul, Korea).

QCM and SAW Sensor
The QCM200 system consists of a built-in frequency counter and resistance meter. It includes a digital controller, crystal holder, crystal oscillator, flow cell adapter, and quartz crystals. QCM crystals are AT-cut quartz crystals embedded between two conducting chrome/gold electrodes operating at 5 MHz with a diameter of 1 inch. Figure S7a shows the QCM crystal used in the presented study. In 1880, the Curie brothers introduced the piezoelectric effect in piezoelectric crystals for the first time [45].
In 1959, Sauerbrey [46] discovered an expression relating to the measurable changes in the crystal's resonant frequency (∆f ) with the mass deposited (∆m) in gaseous/vacuum media, i.e., ∆f = −C f × ∆m. The sensitivity factor (C f ) is 56.6 Hz/(µg·cm −2 ) for AT-cut quartz crystal of 5 MHz at the room temperature [44]. The AT-cut crystals are well known for their zero-temperature coefficient at room temperature. Hence, the inbuilt temperature dependence of QCM crystals is insignificant at room temperature (~1-3 Hz/ • C). The operating temperature of the QCM system is 0-40 • C. However, all the QCM experiments in the present study were conducted at 20 • C.
In 1885, Lord Rayleigh introduced surface acoustic waves propagated along the plane surface of an elastic solid [47]. R. M. White discovered the direct piezoelectric surface wave transduction by using the interdigital transducers (IDTs) [48]. Figure S7b shows a schematic of the 250 MHz SAW sensor used in the presented study. SAW sensors were forged out of a substrate made of ST-cut quartz. The sensor also consists of an aluminum alloy with 1% copper based interdigital transducer (IDT) and a 50 nm Ti adhesive layer deposited using an e-beam evaporator.
The sensitivity of the SAW sensor is influenced by several factors, including the piezoelectric materials, center frequency, wavelength, and SAW velocity. The central frequency (f c ) of the SAW sensor is directly proportional to the SAW velocity (v) and inversely proportional to the wavelength (λ), i.e., f c = v/λ [49]. The center frequency of the SAW sensor was designed to be 250 MHz, the velocity of SAW on the quartz substrate was 3158 m/s, and the wavelength of the aluminum electrode was 12.632 µm [50].

Deposition of Sensing Material on the QCM Sensor
The synthesized NGO@MnO 2 and PPy grafted NGO@MnO 2 were intermixed with IPA in a 10 mg: 1 mL ratio. The solutions were then ultrasonicated for 3 h. After this, 15 µL of this mixture was drop-coated on the QCM surface. The sensing material deposited QCM sensors was then kept in an oven for one hour at 60 • C. Once removed from the oven, the QCM sensors were allowed to cool at room temperature of 20 • C and were used in measurements.

Deposition of Sensing Materials on the SAW Sensor
The sensing material was deposited on the SAW sensor using the following procedure. First, the prepared composite materials were intermixed with IPA in 1 mg:7 mL ratio. Since SAW sensors are highly sensitive to the weight of deposited sensing materials, this ratio was changed from the one previously used for the QCM sensor. The mix was then ultrasonicated for 3 h. After the ultrasonication, 2.5 µL of the solution was drop-coated onto the top surface of the SAW sensor. The SAW sensors deposited with thin sensing films were then heated in an oven at 60 • C for one hour. Once the heating was complete, the obtained SAW sensors were allowed to cool to room temperature at 20 • C.

Material Characterization Apparatus
The hybrid composites were characterized using Vertex 80v FT-IR Spectrometers (Bruker, Billerica, USA) that provide excellent sensitivity and stability. The XRD patterns of the hybrid composites were studied using high resolution X-ray diffractometry-X'pert PRO MRD (Philips/Panalytical, Malvern, UK). FE-SEM-S-4300SE (Hitachi, Tokyo, Japan) was used to study the morphological analysis of the fabricated hybrid composite materials. FE-TEM-JEM-2100F (Jeol, Akishima, Japan) was used to characterize the hybrid composites to study its structure, thickness, and diameter. The elemental analysis of the composite was performed using X-ray photoelectron spectroscopy-K-Alpha (Thermo Fisher Scientific, Waltham, MA, USA).

Gas-Response System Characterization Apparatus
The weight of the sensing materials was measured by analytical balance-AS200.R2 PLUS (Radwag, Radom, Poland). The ultrasonication processes were done using the sonicator-POWERSONIC 410 (Hwashin Technology, Seoul, Korea). An oven-LDO-150F (Daihan LabTech, Namyangju, Korea) was used to heat the bubbler, QCM crystal, and SAW sensor at 60 • C. After the QCM and SAW sensors were used, Petri dishes with the sensors were kept in the vacuum desiccator-VL-B (MadeLab, Hannam, Korea). The outlet of the QCM and SAW chambers were connected to the Fumehood-SFH series 1000/2011 (Shinsaeng, Yonggin, Korea). The R.H. test was conducted using a temperature and humidity chamber-TH-ME-100 (Lab companion, Daejeon, Korea). The temperature and humidity inside the clean booth were measured using a temperature humidity meter-Fluke 971 (Fluke, Everett, D.C., USA) and ETP101 (All-sun, Zhangzhou, China). Figure S8a,b shows the FT-IR spectrum analysis of the hybrid composites. In the FT-IR spectrum, the peak position at 2918-3447 cm −1 represents the adsorption bands of N-H, C-H, and O-H, respectively [51,52]. The intense peak at 1630-1715 cm −1 corresponds to the C=O stretching and C=C stretching adsorption bands of the carboxylic group. C-N in-plane vibration is observed at the intense peak of 1364-1385 cm −1 [51]. It is important to note that these adsorption bands primarily arise from the fabrication of NGO.

FT-IR and XRD Analysis
In addition, the lower wave number from 623-415 cm −1 represents the stretching vibration of the Mn-O-Mn and Mn-O bonds, which, in the present work, indicates the formation of MnO 2 [52]. Finally, PPy shows the peak position from 3447-415 cm −1 , which represents C-H stretching, C-H in-plane deformation, C-H out-of-plane vibration, C=C stretching of pyrrole ring, C-N stretching, and N-H stretching [53].
This also indicates the formation of PPy in the composite materials via a hydrothermal process. Hence, the FT-IR spectrum analysis confirmed the presence of MnO 2 , PPy, and NGO in the samples prepared as part of the presented work.
The XRD analysis of the NGO@MnO 2 composite showed peaks at 12.  Figure S8c). The resulting peaks and planes depict the polycrystalline structure as well as the tetragonal phase of MnO 2 (JCPDS: 44-0141) [54] on the NGO surface. The peak position of 2θ = 25.4 • corresponds to the heap of NGO sheets.
Similarly, the observed peaks of NGO@MnO 2 /PPy match with the lattice plane confirming the crystalline structure and the tetragonal phase of α-MnO 2 (JCPDS: 240734 and 44-0141). We note that PPy might have arisen from the XRD peak between 2θ = 25.4 • and 28.4-33.1 • (see Figure S8d) [54,55]. In Figure S8d, a low intensity peak of NGO was observed between 2θ = 25.4 • and 20.0-25.0 • . From all these observations, we can conclude that the XRD results confirm the presence of NGO and PPy in NGO@MnO 2 /PPy hybrid composite materials.

X-ray Photoelectron Spectroscopy (XPS)
To study the elemental composition of the hybrid composites, we used X-ray photoelectron spectroscopy. The XPS analysis showed the presence of elements, such as Mn, O, C, and N, and this is illustrated in Figure S9. In Figure S9a, 2p 1/2 (654 eV) and 2p 3/2 (642 eV) are shown as the characteristic peaks of Mn 2p. The energy difference between the 2p 1/2 and 2p 3/2 is 12 eV [55]. In Figure S9b, an O 1s characteristic peak is shown at 530 eV, which is mainly due to the bonding of metal and oxygen during the synthesis.
In Figure S9c, C 1s characteristic peaks are shown in 295, 293, and 285 eV corresponding to O-C=O, C-C sp 3 , and C-C sp 2 , respectively, of the NGO@MnO 2 composite indicating oxidation at a high degree [56]. In Figure S9d, the N 1s characteristic peak is shown at 407 eV, which is primarily due to graphitic N in NGO. This also indicates that nitrogen is successfully infused into the hybrid composites. Finally, Figure S9e shows the elemental spectrum of the NGO@MnO 2 composite, which confirmed the presence of Mn, O, C, and N.

SEM and TEM
SEM and TEM analysis were used to study the morphological structures of the hybrid composite high-resolution SEM images. The energy dispersive X-ray analysis (EDX) profiles of the hybrid composite are illustrated in Figure 3. Figure 3a-f illustrates that the smaller particles are stacked haphazardly and formed a cauliflower-like morphology resulting from an aggregation of minute particles. The primary reason for this agglomeration is the incorporation of NGO onto the MnO 2 [54].
The size of the fabricated MnO 2 nanoparticles ranges from~20 to 100 nm in the composites prepared. The observed size of the grains within the composites was~20-30 nm. Figure 3g-i illustrates the EDX profile, which confirms the presence of Mn, O, N, and C. We also found that the elemental percentages by weight of Mn, O, N, and C were 85.5%, 9.0%, 1.0%, and 4.5%, respectively. This is convincing evidence for the formation of MnO 2 and NGO in the synthesis of NGO@MnO 2 . Figure 4g-i illustrates the EDX profile of NGO@MnO2/PPy depicting the elements composed, such as Mn, O, N, C, and Fe. The main reason for the appearance of Fe is because FeCl3 was used as a catalyst in the synthesis of NGO@MnO2/PPy, which left traces of Fe. The elemental composition by the weight of Mn, O, N, C, and Fe are 65.1%, 18.1%, 1.7%, 6.7%, and 8.4%, respectively. From the above observations, we can say that the morphological structure provides convincing evidence for the introduction of PPy, MnO2, and NGO in the hybrid composite.  Similarly, Figure 4a-f shows the high-resolution SEM images of NGO@MnO 2 /PPy. They depict the formation of a cluster of small grain-like structure mainly due to the formation of NGO on the MnO 2 structure. The superficial side of the composites has a spherical grain-like structure (see Figure 4d,e) primarily due to the infusion of PPy onto the NGO@MnO 2 to eventually form the NGO@MnO 2 /PPy composite. The inclusion of PPy is expected to improve the surface area [58] and the conductivity that results from π-electron conjugation [59], which is highly preferable for adsorption/desorption kinetics. .1%, 1.7%, 6.7%, and 8.4%, respectively. From the above observations, we can say that the morphological structure provides convincing evidence for the introduction of PPy, MnO 2 , and NGO in the hybrid composite.
To study the synthesized composites further and to verify the presence of nanosized grains, FE-TEM analysis was conducted. Figure S11a-f shows the structure of the NGO@MnO 2 composite at different magnifications. It is evident from these figures that NGO@MnO 2 consists of nanosized grains in the form of sheets and rods (see Figure S11e,f). The stacked sheets of NGO@MnO 2 have a thickness and diameter ranging from 10 to 20 nm and 30 to 40 nm, respectively.
Figure S12a-f shows the structure of NGO@MnO 2 /PPy at different magnifications. It can be seen from these figures that the surface of NGO@MnO 2 /PPy is modified into spherical or disc-like structures with the introduction of PPy (see Figure S12d-f). These figures also depict the presence of a hollow structure (see Figure S12c-f) in the composite, which is favorable for the reversible gas adsorption/desorption [60]. Hence, with all these findings, the FE-TEM analysis proves the PPy insertion onto the NGO@MnO 2 while forming the NGO@MnO 2 /PPy composite.  To study the synthesized composites further and to verify the presence of nanosized grains, FE-TEM analysis was conducted. Figure S11a-f shows the structure of the NGO@MnO2 composite at different magnifications. It is evident from these figures that NGO@MnO2 consists of nanosized grains in the form of sheets and rods (see Figure  S11e,f). The stacked sheets of NGO@MnO2 have a thickness and diameter ranging from 10 to 20 nm and 30 to 40 nm, respectively.
Figure S12a-f shows the structure of NGO@MnO2/PPy at different magnifications. It can be seen from these figures that the surface of NGO@MnO2/PPy is modified into spherical or disc-like structures with the introduction of PPy (see Figure S12d-f). These figures also depict the presence of a hollow structure (see Figure S12c-f) in the composite, which is favorable for the reversible gas adsorption/desorption [60]. Hence, with all these findings, the FE-TEM analysis proves the PPy insertion onto the NGO@MnO2 while forming the NGO@MnO2/PPy composite.

Frequency Shift towards Different Concentrations of DMMP Vapor
The real time response of the QCM sensor while detecting DMMP is illustrated in Figure 5a,b. During the detection process, the instantaneous change in the frequency shift

Frequency Shift towards Different Concentrations of DMMP Vapor
The real time response of the QCM sensor while detecting DMMP is illustrated in Figure 5a,b. During the detection process, the instantaneous change in the frequency shift was perceived when the DMMP flowed inside the QCM chamber. It can be inferred from Figure 5 that NGO@MnO 2 is sensitive to DMMP. We also notice that the addition of conducting polymer PPy increased the frequency shift of NGO@MnO 2 dramatically due to the formation of a hydrogen bond between the target DMMP and PPy.
ing the transport of charge carriers. The resultant frequency shift is about six-times better at a DMMP concentration of 150 ppm. The literature also reports that the hydrogen atom containing the carboxyl group of PPy can couple non-covalently with the oxygen atom containing a phosphoryl group of DMMP (Figure 5c) through hydrogen bonding [61]. A process similar to this (Figure 5d) occurs when PPy reacts with the sarin. We also observed a fast recovery of the sensing material when nitrogen flowed inside the QCM chamber.

Relationship between the Frequency Shift and Thickness of the Sensing Film
In order to analyze the effect of the thickness of the sensing material to the frequency shift of the QCM sensor, we first define the mass shift. The mass shift, ∆m, is defined as the mass of DMMP molecules adsorbed onto the surface of the sensing film during the adsorption process [62]. Figure S13 shows the effect of mass accumulation of DMMP adsorbed onto the surface of the hybrid composites.
At a concentration of 25 ppm, ∆m in NGO@MnO2 and NGO@MnO2/PPy were 0.140 and 0.463 µg/cm 2 , respectively. At 125 ppm, the ∆m values were 0.367 and 1.315 µg/cm 2 , respectively. From this, we understand that the PPy-incorporated NGO@MnO2 adsorbed 3.58-times more DMMP than NGO@MnO2 at a concentration of 125 ppm. From Figures 5 and S13, it appears that the frequency shift and mass accumulation are correlated [62]. The hydrogen bond changes the charge distribution along the PPy backbone, affecting the transport of charge carriers. The resultant frequency shift is about six-times better at a DMMP concentration of 150 ppm. The literature also reports that the hydrogen atom containing the carboxyl group of PPy can couple non-covalently with the oxygen atom containing a phosphoryl group of DMMP (Figure 5c) through hydrogen bonding [61]. A process similar to this (Figure 5d) occurs when PPy reacts with the sarin. We also observed a fast recovery of the sensing material when nitrogen flowed inside the QCM chamber.

Relationship between the Frequency Shift and Thickness of the Sensing Film
In order to analyze the effect of the thickness of the sensing material to the frequency shift of the QCM sensor, we first define the mass shift. The mass shift, ∆m, is defined as the mass of DMMP molecules adsorbed onto the surface of the sensing film during the adsorption process [62]. Figure S13 shows the effect of mass accumulation of DMMP adsorbed onto the surface of the hybrid composites.
At a concentration of 25 ppm, ∆m in NGO@MnO 2 and NGO@MnO 2 /PPy were 0.140 and 0.463 µg/cm 2 , respectively. At 125 ppm, the ∆m values were 0.367 and 1.315 µg/cm 2 , respectively. From this, we understand that the PPy-incorporated NGO@MnO 2 adsorbed 3.58-times more DMMP than NGO@MnO 2 at a concentration of 125 ppm. From Figures 5 and S13, it appears that the frequency shift and mass accumulation are correlated [62].

Selectivity and Polar Plot
In order to perform further studies on synthesized composite materials, the hybrid composites were subjected to DMMP, water and VOCs, such as ethanol, toluene, n-hexane, and methanol. From the results of this study given in Figure 6a,b, it can be observed that the sensing materials are highly sensitive to DMMP vapor rather than any potential vapors. In order to support this claim, we exposed the composites to toluene vapor with a concentration of 3522 ppm, which was more than 35-fold higher than that of DMMP.

Selectivity and Polar Plot
In order to perform further studies on synthesized composite materials, the hybrid composites were subjected to DMMP, water and VOCs, such as ethanol, toluene, n-hexane, and methanol. From the results of this study given in Figure 6a,b, it can be observed that the sensing materials are highly sensitive to DMMP vapor rather than any potential vapors. In order to support this claim, we exposed the composites to toluene vapor with a concentration of 3522 ppm, which was more than 35-fold higher than that of DMMP.
Despite the elevated concentration of this potential interference, the frequency response of NGO@MnO2 and NGO@MnO2/PPy were more than 24-fold and 186-fold higher in DMMP than in toluene [63]. Hence, from this analysis, we conclude that the composite materials prepared possess an excellent frequency response towards CWAs-in particular, the nerve agent sarin. We also draw the corresponding polar plots for studies conducted for the selectivity of the composites prepared. Figure 6c shows the polar plot response ratios (R1/R2 with R1 and R2 representing the responses of the NGO@MnO2/PPy-coated and NGO@MnO2coated QCMs, respectively) when the composites are exposed to other vapors [64]. It is inferred from the study that the addition of PPy in NGO@MnO2 caused the minimum and maximum response ratios in toluene and ethanol, which are 1.06-times and 32.67-times greater, respectively.
In the case of water vapor and methanol, the response ratios were greater than 1.58 times and 2.68 times, respectively. These response ratios were greater than 14.60 times and 7.11 times, respectively, for N-hexane and DMMP. The addition of PPy greatly facilitates the hydrogen bond formation between DMMP molecules and the sensing film such that more DMMP is absorbed onto the surface enhancing the response of the sensing material [61]. Despite the elevated concentration of this potential interference, the frequency response of NGO@MnO 2 and NGO@MnO 2 /PPy were more than 24-fold and 186-fold higher in DMMP than in toluene [63]. Hence, from this analysis, we conclude that the composite materials prepared possess an excellent frequency response towards CWAs-in particular, the nerve agent sarin.
We also draw the corresponding polar plots for studies conducted for the selectivity of the composites prepared. Figure 6c shows the polar plot response ratios (R1/R2 with R1 and R2 representing the responses of the NGO@MnO 2 /PPy-coated and NGO@MnO 2coated QCMs, respectively) when the composites are exposed to other vapors [64]. It is inferred from the study that the addition of PPy in NGO@MnO 2 caused the minimum and maximum response ratios in toluene and ethanol, which are 1.06-times and 32.67-times greater, respectively.
In the case of water vapor and methanol, the response ratios were greater than 1.58 times and 2.68 times, respectively. These response ratios were greater than 14.60 times and 7.11 times, respectively, for N-hexane and DMMP. The addition of PPy greatly facilitates the hydrogen bond formation between DMMP molecules and the sensing film such that more DMMP is absorbed onto the surface enhancing the response of the sensing material [61].

Linear Relationship between the Frequency Shift and Concentration of DMMP Vapor
We now move on to studying the linear relationship between the frequency shift and concentration of DMMP. Figure 7 shows the linearity of the QCM sensor for DMMP concentrations of 25 to 150 ppm. NGO@MnO 2 and NGO@MnO 2 /PPy present in the QCM sensor have R 2 values of 0.992 and 0.975, respectively. This is primarily due to good cohesive contact between the sensing materials and the electrodes as well as due to the large surface area [28] for DMMP gas detection.

Linear Relationship between the Frequency Shift and Concentration of DMMP Vapor
We now move on to studying the linear relationship between the frequency shift and concentration of DMMP. Figure 7 shows the linearity of the QCM sensor for DMMP concentrations of 25 to 150 ppm. NGO@MnO2 and NGO@MnO2/PPy present in the QCM sensor have R 2 values of 0.992 and 0.975, respectively. This is primarily due to good cohesive contact between the sensing materials and the electrodes as well as due to the large surface area [28] for DMMP gas detection.
The results also suggest that the composite materials developed can be promising candidates for practical applications [63]. Table 1 depicts the observed data of the calibration curve for the QCM sensor.

Repeatability
To investigate the repeatability of the measurements, hybrid-composite-coated QCM sensors were exposed to a constant concentration of 25 ppm for four cycles of exposure and recovery. Figure S14 shows the response of the QCM sensor for four successive cycles. The response curves show similar behaviors under the same conditions, although we noticed small drifts. These small drifts were presumed to be caused by differences in the concentration since the concentration of vapor cannot be controlled by the dilution process highly accurately [65].
The repeatability of each sensing materials in the QCM sensor was evaluated by calculating the coefficient of variation (D) of the repeated responses (n = 4) of the given hybrid composites when exposed to 25 ppm DMMP [66]. The D for the responses of the NGO@MnO2 and NGO@MnO2/PPy in the detection of 25 ppm DMMP was evaluated with the following equation: The results also suggest that the composite materials developed can be promising candidates for practical applications [63]. Table 1 depicts the observed data of the calibration curve for the QCM sensor.

Repeatability
To investigate the repeatability of the measurements, hybrid-composite-coated QCM sensors were exposed to a constant concentration of 25 ppm for four cycles of exposure and recovery. Figure S14 shows the response of the QCM sensor for four successive cycles. The response curves show similar behaviors under the same conditions, although we noticed small drifts. These small drifts were presumed to be caused by differences in the concentration since the concentration of vapor cannot be controlled by the dilution process highly accurately [65].
The repeatability of each sensing materials in the QCM sensor was evaluated by calculating the coefficient of variation (D) of the repeated responses (n = 4) of the given hybrid composites when exposed to 25 ppm DMMP [66]. The D for the responses of the NGO@MnO 2 and NGO@MnO 2 /PPy in the detection of 25 ppm DMMP was evaluated with the following equation: where δ is the standard deviation of a set of responses, and k is the average response of the QCM sensor coated with the hybrid composites when exposed to a DMMP vapor concentration of 25 ppm. The calculated values of δ, k, and D for NGO@MnO 2 and NGO@MnO 2 /PPy are summarized in Table 2. In this table, a lower value of D indicates a higher repeatability [66]. We observe that NGO@MnO 2 and NGO@MnO 2 /PPy show D values of 7.890% and 5.698%, respectively.

Response and Recovery Times
The response times (T 90 ) can be explained as the time taken to attain 90% of the equilibrium, whereas the recovery time is the time taken to reach 10% of the equilibrium via purging with the dry nitrogen [67]. Figure S15 illustrates the response and recovery time of the hybrid composites. NGO@MnO 2 exhibited response and recovery times of 96 and 116 s, respectively, in the detection of 75 ppm DMMP. On the other hand, NGO@MnO 2 /PPy showed response and recovery times of 103 and 117 s, respectively, in the detection of the same 75 ppm DMMP.
For both the composites, the recovery times were longer than the response times. The reversible response also indicates that the adsorption of DMMP onto the sensing materials consists of weak hydrogen bonding or Van der Waals forces [67,68]. The performance comparison was conducted with a composite-material-coated QCM sensor for the detection of DMMP as shown in Table S3. This revealed that sensing materials provided excellent frequency response and shorter response times and recovery times in comparison with the other composite materials.

Experimental Results with a SAW Sensor
The performance of the SAW sensor was investigated at room temperature. Hence, we need to study the effect of oxygen in the sensing performance of the SAW sensor. We also note that atmospheric air contains 78.09% nitrogen, 20.95% oxygen, and 0.96% other gases, and hence studying the effect of oxygen is important. In our experiments, we used synthetic air to analyze the performance of the SAW sensor at a room temperature of 20 • C and 25-30% R.H.

Frequency Shifts towards Different Concentrations of DMMP Vapor
We now study the frequency responses of the SAW sensor in this section. The real-time response of SAW sensors coated with the hybrid composites is shown in Figure 8. The SAW sensors studied were exposed to a DMMP concentration ranging from 25 to 150 ppm. In our experiments, one test cycle comprises a DMMP exposure of 5 min followed by a purging for 5 min using synthetic air.

Selectivity and Polar Plot
To investigate the selectivity of SAW sensors coated with hybrid composites, they were exposed to DMMP, water, and VOCs at a fixed flow rate of 100 sccm. Figure 9a,b shows the logarithmic frequency shifts of SAW sensors when they are exposed to different vapors. From this response, it can be clearly seen that the hybrid composite materials synthesized possess higher frequency shift towards DMMP than the one it has for the inter- At a concentration 25 ppm, the measured frequency shifts of NGO@MnO 2 and NGO@MnO 2 /PPy were 145 and 98 Hz, respectively. Similarly, frequency shifts of 1114 and 1490 Hz were measured at 150 ppm of DMMP for the NGO@MnO 2 and NGO@MnO 2 /PPy composites, respectively. The excellent measured response is attributed to the excellent sensitivities of manganese oxide, graphene oxide, and PPy to detect DMMP [37,69,70]. We also noticed a slight drift on the frequency shift that is due to the low desorption rate of DMMP molecules from the sensing layer surface [71].

Selectivity and Polar Plot
To investigate the selectivity of SAW sensors coated with hybrid composites, they were exposed to DMMP, water, and VOCs at a fixed flow rate of 100 sccm. Figure 9a,b shows the logarithmic frequency shifts of SAW sensors when they are exposed to different vapors. From this response, it can be clearly seen that the hybrid composite materials synthesized possess higher frequency shift towards DMMP than the one it has for the interferences.

Linear Relationship between the Frequency Shift and Concentration of DMMP Vapor
Similar to the test conducted for the QCM sensor, we conducted a linearity test for the SAW sensor. Figure 10 illustrates the linearity of SAW sensors when exposed to a DMMP with a concentration ranging from 25 to 150 ppm. NGO@MnO2 and NGO@MnO2/PPy show R 2 values of 0.979 and 0.989, respectively. From this figure, we can say that hybrid composite materials possess good linearity in the SAW sensors and reinforce the idea of being used for gas sensing [72]. Table 3 represents the observed data of the calibration curve for the SAW sensor. In order to understand this better, we look at a water vapor concentration that was 2892 ppm and 28-fold higher than that of the DMMP. Although the concentration of this potential interference was more, the frequency response of NGO@MnO 2 and NGO@MnO 2 /PPy towards the DMMP was more than 91-fold and 141-fold greater, respectively, than the frequency response of water vapor [63]. Therefore, we conclude that the composite-material-deposited SAW sensor possessed excellent frequency responses towards the target CWA and simulant. Figure 9c shows the polar plot response ratios (R3/R4) of the NGO@MnO 2 /PPy-and NGO@MnO 2 -coated SAW sensor when exposed to various vapors [64]. From this plot, we observe that in the SAW sensor incorporation of PPy into NGO@MnO 2 increased the affinity to ethanol, which was 5.23-times greater than that of NGO@MnO 2 . In the case of DMMP, the R3/R4 of NGO@MnO 2 /PPy was 1.27-times greater than that of NGO@MnO 2 . The mechanism behind this increase is that, during the adsorption process, DMMP associates with the conductive polymer (PPy) to increase the number of free as well as mobile-hole charge carriers [70]. The R3/R4 values of various vapors is given in the following relation: R3/R4 of H 2 O > R3/R4 of Toluene > R3/R4 of N-hexane > R3/R4 of methanol. However, the response ratios of these vapors were less than 1.

Linear Relationship between the Frequency Shift and Concentration of DMMP Vapor
Similar to the test conducted for the QCM sensor, we conducted a linearity test for the SAW sensor. Figure 10 illustrates the linearity of SAW sensors when exposed to a DMMP with a concentration ranging from 25 to 150 ppm. NGO@MnO 2 and NGO@MnO 2 /PPy show R 2 values of 0.979 and 0.989, respectively. From this figure, we can say that hybrid composite materials possess good linearity in the SAW sensors and reinforce the idea of being used for gas sensing [72]. Table 3 represents the observed data of the calibration curve for the SAW sensor.

Linear Relationship between the Frequency Shift and Concentration of DMMP Vapor
Similar to the test conducted for the QCM sensor, we conducted a linearity test for the SAW sensor. Figure 10 illustrates the linearity of SAW sensors when exposed to a DMMP with a concentration ranging from 25 to 150 ppm. NGO@MnO2 and NGO@MnO2/PPy show R 2 values of 0.979 and 0.989, respectively. From this figure, we can say that hybrid composite materials possess good linearity in the SAW sensors and reinforce the idea of being used for gas sensing [72]. Table 3 represents the observed data of the calibration curve for the SAW sensor.

Repeatability
To investigate the repeatability of the sensing process of SAW sensors, the hybridcomposite-deposited SAW sensors were exposed to the same concentration of 75 ppm DMMP, and the corresponding responses are shown in Figure S16. In this study, one test cycle comprised 5 min of DMMP exposure followed by 5 min of purging with synthetic air.
Similar to the one in the QCM sensor, the repeatability of each sensing material in the SAW sensor was evaluated by calculating the D value of the repeated responses (n = 5) of the given hybrid composites when they are exposed to 75 ppm DMMP [66]. D for the responses of the hybrid composites was evaluated using Equation (5). The calculated values of δ, k, and D of the given hybrid-composite-coated SAW sensor are summarized in Table 4. In this table, a lower value of D indicates higher repeatability [66]. From our calculations, we found that NGO@MnO 2 and NGO@MnO 2 /PPy had D values of 11.3% and 11.6%, respectively. Since there were no significant changes in the response curve [73], we concluded that the composite materials developed possess excellent repeatability when they are used to detect DMMP. This repeatability also ensures that the fabricated composite materials can be used in practical applications that require a high level of repeatability in the sensing process [63].

Response and Recovery Times
We now move on to studying the response and recovery times of the developed composite materials. The results are given in Figure S17, which illustrates the response and recovery times of the hybrid-composite-coated SAW sensor for an exposure of 75 ppm of DMMP. The response and recovery times of NGO@MnO 2 were found to be 121 and 208 s, respectively, while those of NGO@MnO 2 /PPy were 120 and 197 s, respectively.
Similar to the QCM sensor, the recovery time was longer than the response time for the SAW sensors. This difference is attributed to the low desorption rate of DMMP molecules, which were initially adsorbed onto the sensing film surface [71]. We also conducted a performance comparison of the composite-material-coated SAW sensor for the detection of DMMP with other composite materials as shown in Table S4. It can be inferred that the composite materials showed shorter response and recovery times compared with other composite materials.

Effect of R.H. on the Sensing Performance
To investigate the environmental effects on the performance of the sensor, we exposed the SAW sensor coated with NGO@MnO 2 /PPy to 25-30%, 60-65%, and 85-90% R.H. conditions at a constant temperature of 20 • C. The results of this analysis are shown in Figure S18. We previously, as part of our research, tested the QCM sensor for the detection of DMMP under 50% and 90% R.H. conditions at 20 • C but focus now on the SAW sensor in the present work.
From the results of this study, the frequency shifts of the NGO@MnO 2 /PPy-coated SAW sensor under 25-30%, 60-65%, and 85-90% R.H. conditions and 75 ppm were found to be 868, 828, and 543 Hz, respectively. The results for an exposure of 150 ppm were 1546, 1210, and 1190 Hz, respectively. For an exposure of 75 ppm, the dampings in frequency shift were about 4.60% and 34.42%, respectively, when the R.H. was changed from 25-30% to 60-65% and from 60-65% to 85-90%. Similarly, for an exposure of 150 ppm, and when the R.H. was changed from 25-30% to 60-65% and from 60-65% to 85-90%, an attenuation in the frequency shift occurred with values of 21.73% and 1.65%, respectively. This reduction in the frequency shift resulted from the adsorption of water molecules onto the thin sensing films, which failed to function properly as the R.H. increased to 100% [74].

Stability of Composite Materials
The stability test of NGO@MnO 2 /PPy-coated QCM sensor was conducted for the detection of 150 ppm DMMP for five consecutive days, which is illustrated in Figure S19. NGO@MnO 2 /PPy revealed satisfying stability results when comparing with the initial frequency shift for the detection of DMMP. Under normal room temperature, it was reported that graphene oxide exhibited only a small declination of 1.34% in the response for the detection of DMMP, which was monitored for 30 days [63].
In addition, manganese oxide-based composites were investigated for the long-term stability of 6 months. The composites revealed satisfying repeatability and stability with a slight declination in frequency shift of~18 Hz [75]. Furthermore, it was reported that the polypyrrole-based composites showed more than 95% retention of the initial sensitivity during a monitoring period of 21 days [61]. Figure 11 shows the proposed adsorption mechanism of the DMMP onto the NGO@MnO 2 /PPy surface. DMMP may interact with the COOH group connected with the NGO, which indicates a strong interaction. Hydrogen bonds are formed due to the adsorption of DMMP between the P=O and hydroxyl groups (O-H-P=O) [76]. The surface selection rule [77] is also likely to occur in which hydroxyl groups are reoriented during the adsorption, leading to additional favorable perpendicular orientation of the OH bond with respect to the surface [76].

Conclusions
As part of the present work, we developed and tested two hybrid composite materials that can act as a potential candidate for the detection of CWAs. Hydrothermal and thermal reduction processes were used to synthesize these composite materials. The hybrid composites were characterized using FT-IR, XRD, XPS, SEM, and TEM to study the elemental composition, intensities, energy differences, and morphological structures. The frequency shifts, selectivity, polar plots, linearity, repeatability, and response/recovery times for various concentrations of DMMP for both the QCM and SAW sensors were measured, plotted, and compared.
All these experiments were performed at a temperature of 20 °C and R.H. of 25-30%. The FTIR analysis confirmed the presence of elements used in the synthesis of the composite materials. The XRD results confirmed the tetragonal nanostructure of the composite materials. To add to this, a study using XPS exhibited the intensities and energy differences between two states of an element. The SEM and TEM results for the two composite materials also showed cauliflower-shaped morphology primarily due to the agglomeration of smaller particles.
In both the QCM and SAW sensors, NGO@MnO2/PPy showed enhanced frequency shifts when compared to the NGO@MnO2 composite. A selectivity test also showed that Figure 11. The proposed mechanism of DMMP adsorption on the NGO@MnO 2 /PPy composites with a Bronsted acid site, Lewis acid site, and hydrogen bonding or van der Waals force. The figure is not to scale. DMMP may be attached to the MnO 2 surface by binding with a Bronsted acid site, and a Lewis acid site, which generates a bidendate structure to the nearest hydroxyl group and with the metal site [75]. In the case of the PPy, the phosphoryl group of DMMP binds with the carboxyl group, thereby, resulting in the formation of hydrogen bonds [61]. It is reported that the polymeric structure increases the hydrogen bonding strength in alcohols [78]. Thus, the NGO@MnO 2 /PPy composite showed increased adsorption to the DMMP.

Conclusions
As part of the present work, we developed and tested two hybrid composite materials that can act as a potential candidate for the detection of CWAs. Hydrothermal and thermal reduction processes were used to synthesize these composite materials. The hybrid composites were characterized using FT-IR, XRD, XPS, SEM, and TEM to study the elemental composition, intensities, energy differences, and morphological structures. The frequency shifts, selectivity, polar plots, linearity, repeatability, and response/recovery times for vari-ous concentrations of DMMP for both the QCM and SAW sensors were measured, plotted, and compared.
All these experiments were performed at a temperature of 20 • C and R.H. of 25-30%. The FTIR analysis confirmed the presence of elements used in the synthesis of the composite materials. The XRD results confirmed the tetragonal nanostructure of the composite materials. To add to this, a study using XPS exhibited the intensities and energy differences between two states of an element. The SEM and TEM results for the two composite materials also showed cauliflower-shaped morphology primarily due to the agglomeration of smaller particles.
In both the QCM and SAW sensors, NGO@MnO 2 /PPy showed enhanced frequency shifts when compared to the NGO@MnO 2 composite. A selectivity test also showed that the composite materials were sensitive to DMMP rather than to any other potential interferences. The polar plots presented in this article show that the addition of PPy in NGO@MnO 2 resulted in an enhanced frequency shift to DMMP and ethanol. Both the hybrid composite materials also exhibited excellent R 2 when used in QCM and SAW sensors. Low values of D for both the composite materials confirmed their high repeatability in both sensors. We also noticed that the recovery times were longer than the response times in both the QCM and SAW sensors, which was mainly due to the low desorption rate.
Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nano12172965/s1, The supporting information includes information about the list of acronyms (Table S1), the chemical structure of chemical warfare agents (a) sarin, (b) soman, (c) Tabun, and its simulant (d) DMMP ( Figure S1), the proposed chemical structure of the synthesized materials ( Figure S2), the coefficients used to calculate the concentration of vapor used (Table S2), a schematic diagram of the gas-sensing system in the QCM and SAW sensors ( Figures S3 and S4), the QCM and SAW sensor experimental setup ( Figures S5 and S6), the QCM and SAW sensors used in the experiment ( Figure S7), the FTIR and XRD analysis of composite materials ( Figure S8), the XPS results of the fabricated composite materials ( Figures S9 and S10), FE-TEM images of composite materials (Figures S11 and S12), the effects of mass accumulation of composite materials ( Figure S13), the repeatability of composite materials in the QCM sensor ( Figure S14), the response and recovery times of composite materials in the QCM sensor ( Figure S15), a summary of the sensing materials and performance in the QCM sensor (Table S3), the repeatability of the composite materials in the SAW sensor ( Figure S16), the response and recovery times of composite materials in the SAW sensor ( Figure S17), a summary of the sensing materials and performance in the SAW sensor (Table S4), the effects of the relative humidity conditions on the SAW sensor coated with sensing materials (Figure S18), and the stability test of the composite materials for the detection of DMMP ( Figure S19).