The morphologies of the samples were characterized by field emission scanning electron microscopy (FESEM; Carl Zeiss Supra 55).

The sensor used was of a “delay-line” type with a center frequency of 69.4 MHz. Fifty pairs of interdigital transducers (IDTs) with a diameter of 2,500 μm and a period of 40 μm were fabricated on piezoelectric substrates. Gold IDTs with a thickness of 150 nm were deposited on 10-nm chromium IDTs using standard photolithographic techniques. The effective area of the quartz substrate was 8 mm², and its geometric structure was a parallelogram so as to reduce the acoustic reflection on the edges of the substrate.

A schematic diagram of this experimental setup is shown in Figure 2. The sensor measurement was carried out in a sealed chamber under static conditions. The temperature in the chamber could be set to a specific value and monitored by using a thermometer. By setting the voltage and current of the heating stage, the surface temperature of the sensor acted as the working temperature. The corresponding liquid solution containing a certain amount of formaldehyde was injected into a small heating plate that was fixed in the mixing chamber. After formaldehyde gas was mixed with indoor air evenly, the indoor valve was opened so the gas could come into contact with the sensor. The concentration of formaldehyde can be obtained by controlling the number of droplets. After each measurement, nitrogen was pumped into the chamber to purge the sensor environment.

FESEM investigations confirmed the deposition of a continuous graphene/PMMA film on the surface of the sensor devices (Figure 3A). As can be seen from Figure 3A, the film smoothly and compactly covers the entire substrate. A high-magnification SEM image (Figure 3B) further confirms the uniform deposition of the graphene/PMMA film by the multi-source spray deposition method. The graphene sheets (gray-white bulk in the image) were uniformly dispersed in the PMMA film, and thus, the film formed a three-dimensional framework structure. The detailed side view of the as-prepared film in Figure 3C shows that the typical thickness is about 2.2 μm. The film thickness can be increased by prolonging the deposition time, while the surface morphology can be tuned by changing the spraying parameters and solution composition. However, increasing the amount of the sprayed solution and the carrier gas flow rate and decreasing the gap between the nozzle and the substrate changes the morphology of the film from a uniform dispersion to a dense structure.

The response of sensors with different graphene/PMMA film thicknesses to 4.0 ppm formaldehyde at 120°C is shown in Figure 4. All the samples exhibited responses and could be recovered after formaldehyde was extracted. The response amplitude varied with the coating thickness. The results show that the sensor with 2.2-µm film thickness had the highest sensitivity and performance compared to the other sensors. When the graphene/PMMA film is extremely thin or thick, its sensitivity to formaldehyde drastically reduces. As mentioned previously, thickness of films is controlled by controlling the spraying time. Since the structure of all the films was the same in all factors other than thickness, it can be concluded that the difference in sensor response was mainly due to the difference in film thickness. The increase in film thickness affects the normal direction of the surface, interferes with the propagation of the wave, increases the insertion loss, and reduces the center frequency of the sensor sample. Therefore, sensitivity can be adjusted by adjusting the thickness of graphene/PMMA sensitive films.

The effect of temperature on the formaldehyde vapor sensor was also investigated. The sensor responses of the same sample with the graphene/PMMA film thickness of 2.2 μm at temperatures ranging from 30°C to 180°C are depicted in Figure 5. The response performance underwent a sharp frequency shift when the temperature reaches 120°C, following which the response decreased with the increasing temperature.

Figure 6 shows the response of the graphene/PMMA film sensor sample with a thickness of 2.2 μm to formaldehyde gas of different concentrations at an operating temperature of 120°C. The sensor was exposed to varying concentrations of formaldehyde via the synthetic air that was introduced into the room for calibration, and the dynamic responses of the sensor were measured; the results showed good consistency at different times. The relationship between the frequency shift and gas concentration is linear. After purging formaldehyde with dry synthetic air, the working frequency of the SAW sensor increased and returned to baseline. In these experiments, the graphene/PMMA sensor samples exhibited immediate response upon both exposure to and recovery of formaldehyde, indicating the good response and recovery features of the sensor.

Figure 7 depicts the simultaneous interpretation of the response curves of sensors when coated with different sensing films (pure PMMA, pure graphene, and the graphene/PMMA composite). Compared with sensors based on pure graphene and PMMA films, the graphene/PMMA multilayer film sensor exhibited excellent performance. The frequency shift exhibited by graphene/PMMA was approximately 6.9 kHz, while that of pure graphene and PMMA sensors was 2.0 and 0.5 kHz, respectively. Furthermore, the results show that the sensitivity of the acoustic electric effect to the gas film interaction is far greater than that of the mass loading effect, indicating that the acoustic electric effect is the dominant effect of graphene/PMMA-laminated film sensors.

Good selectivity is generally considered an important function of high-performance gas sensors. Figure 8 shows the response of the graphene/PMMA sensor to four gases, ethanol, NO, NH₃, and acetone, where PMMA is insoluble in NO, NH₃, and ethanol and soluble in (acetone). The gas sensitivity measurement was carried out at 120°C. The frequency shifts of 5 ppm formaldehyde, 100 ppm ethanol, 100 ppm NO, 100 ppm NH₃, and 100 ppm acetone were 7.9, 1.1, 0.4, 0.2, and 1.4, respectively. The results show that the graphene/PMMA sensor had a significant response to formaldehyde and a comparatively weak response to the other gases, revealing that the developed sensor has a high selectivity to formaldehyde.

Two response mechanisms are proposed in terms of graphene/PMMA chemical resistance to formaldehyde. First, due to the tunneling effect of quantum mechanics, a conductive path is formed in the graphene/polymer composite; the distance between the conductive elements allows for electron hopping (Goh et al., 2003; Valentini et al., 2003; Wang et al., 2017). Due to the absorption of the organic vapor, the polymer matrix expands and increases in size, therefore leading to an increase in the distance between graphene sheets and, consequently, a further increase in the contact resistance. The degree of distension and the resulting electrical response depends on the solubility of the polymer in steam. Second, the electronic properties of the conductive elements in the composite film undergo changes, owing to the charge transfer caused by the adsorption of organic molecules (Fan et al., 2002; Someya et al., 2003).

As shown in Figure 8, the graphene/PMMA film has a high sensitivity to formaldehyde. In spite of acetone being an ideal solvent for PMMA, the response of the composite membrane to other solvents, including acetone, is significantly less compared with that to formaldehyde. In addition, solvents that are insoluble or insoluble in PMMA, such as ethanol, would cause a recognizable response as well, which indicates that the main mechanism of the reaction to the formaldehyde vapor cannot be attributed to volume expansion. Thus, it can be concluded that the interaction between the composite surface and formaldehyde was responsible for the observed response. As mentioned previously (Jeremy et al., 2008), formaldehyde seems to adsorb to the graphene sheet through interactions with polar functional groups (such as hydroxyl groups), the existence of which in the graphene structure is a long established fact. This could affect the conductivity of graphene materials, owing to the occurrence of the charge exchange between the p-type semiconductor graphene (Lu et al., 2009) and the adsorbed formaldehyde molecules. Graphene nanoflakes dispersed uniformly in the three-dimensional framework structure of the polymer matrix provide a high aspect ratio and a high surface-area-to-volume ratio for the nanocomposites; due to this interaction, the number of holes in the semiconductor would be reduced, and the electric resistance of the device would thus increase.