Heterostructures Based on Cobalt Phthalocyanine Films Decorated with Gold Nanoparticles for the Detection of Low Concentrations of Ammonia and Nitric Oxide

This work is aimed at the development of new heterostructures based on cobalt phthalocyanines (CoPc) and gold nanoparticles (AuNPs), and the evaluation of the prospects of their use to determine low concentrations of ammonia and nitric oxide. For this purpose, CoPc films were decorated with AuNPs by gas-phase methods (MOCVD and PVD) and drop-casting (DC), and their chemiresistive sensor response to low concentrations of NO (10–50 ppb) and NH3 (1–10 ppm) was investigated. A comparative analysis of the characteristics of heterostructures depending on the preparation methods was carried out. The composition, structure, and morphology of the resulting hybrid films were studied by X-ray photoelectron spectroscopy (XPS) and inductively coupled plasma atomic emission (ICP-AES) spectroscopy, as well as electron microscopy methods to discuss the effect of these parameters on the sensor response of hybrid films to ammonia and nitric oxide. It was shown that regardless of the fabrication method, the response of Au/CoPc heterostructures to NH3 and NO gases increased with an increase in the concentration of gold. The sensor response of Au/CoPc heterostructures to NH3 increased 2–3.3 times compared to CoPc film, whereas in the case of NO it increased up to 16 times. The detection limits of the Au/CoPc heterostructure with a gold content of ca. 2.1 µg/cm2 for NH3 and NO were 0.1 ppm and 4 ppb, respectively. It was shown that Au/CoPc heterostructures can be used for the detection of NH3 in a gas mixture simulating exhaled air (N2—74%, O2—16%, H2O—6%, CO2—4%).


Introduction
Gas sensors have a wide range of applications in various fields, including the chemical industry, food and cosmetics production, and agriculture and environment monitoring. Another important area of application for gas sensors is non-invasive medical diagnostics. The exhaled breath contains a large number of volatile organic compounds (VOCs), sulfurcontaining compounds, halogen-containing compounds, nitrogen-containing compounds, etc. [1]. Among the gases considered as biomarkers for different medical conditions, the most important are ammonia, nitric oxide, hydrogen sulfide, acetone, ethane, and aldehydes. Recently, interest in the study of exhaled breath as a non-invasive method of diagnosing diabetes mellitus, renal failure, bronchopulmonary and cardiovascular disease, and other ailments has increased in the world [2][3][4]. These platforms are based on the analysis of the exhaled air of patients and the determination of biomarkers indicating

Preparation of Au/CoPc Heterostructures
Au/CoPc heterostructures were obtained according to the general scheme shown in Figure 1. Au/CoPc heterostructures were obtained according to the general scheme shown in Figure 1. First, CoPc films were prepared by the PVD method. The deposition conditions were as follows: the total pressure was 5 × 10 −5 Torr, the evaporation temperature was 450 °C, the deposition time was 1 h, and the substrate temperature was 60 °C. It is known that when CoPc films are heated at temperatures above 200 °C, a phase transition from α-modification to β can occur, and the size of crystallites and their packing density can change, affecting the sensor properties. For this reason, before AuNPs deposition, CoPc films were annealed under the conditions of the MOCVD experiment (T = 280 °C, Ar/O2 = 2, 1 h) to obtain CoPc films with the same phase composition in order to exclude the effect of this parameter on the sensor properties.
Second, the surface of annealed CoPc films was decorated with AuNPs using three different methods, viz. MOCVD, PVD, or DC. AuNPs were deposited on annealed CoPc films by MOCVD using (CH3)2Au(thd) as a precursor in a reactor with cold walls, equipped with an Ar lamp. The evaporator temperature was 60 °С (the saturated vapor pressure of (CH3)2Au(thd) was given by the equation ln(p, atm) = 36.61-15,231/T(K) [55] corresponding to the precursor partial pressure 5 × 10 −2 Torr). The deposition temperature was varied from 210 to 280 °С, and the optimal deposition temperature was 280 °С. Other experimental parameters were set as follows: the precursor weight was 3 mg, the ratio of the reactant gaseous mixture (Ar/O2) in the reaction zone was 2; the total reactor pressure was ~2 Torr; and the deposition time was 3 min. The deposition procedure was performed one, two, or three times sequentially, and the samples obtained in this way were designated as Au_CVD1/CoPc, Au_CVD2/CoPc, and Au_CVD3/CoPc, respectively ( Figure 1).
AuNPs were deposited on annealed CoPc films by PVD from the load of metallic Au in vacuum (P = 10 −6 Torr) using a UVM-71 installation. The load of gold was varied from 0.6 to 5 mg, and the load of 0.6 mg was chosen as optimal for obtaining samples. The evaporator temperature was 1532 °C (which corresponds to a partial pressure of Au 10 −2 Torr), and the substrate temperature was 100 °C. The deposition procedures were carried out until complete evaporation of Au (0.6 mg). This procedure was repeated one, two, or three times sequentially, and the samples obtained in this way were designated as Au_PVD1/CoPc, Au_PVD2/CoPc and Au_PVD3/CoPc, respectively ( Figure 1).
AuNPs were deposited on annealed CoPc films by DC of 5 × 10 −4 M AuNPs solution in hexane, prepared according to the method proposed by Eah et al. [56]. In brief, 400 µL of 0.05 M of NaBH4 in an aqueous NaOH was added slowly to 100 µL of 0.05 M of HAuCl4·3H2O in aqueous HCl, which led to the formation of a red solution. The distilled water was added to obtain 5 × 10 −4 M AuNPs solution. After that, 5 mL of acetone was added to 10 mL of 5 × 10 −4 M AuNPs aqueous solution and the mixture was shaken manually for 2 min. Then, 0.1 g of dodecanethiol in 5 mL of hexane was added rapidly First, CoPc films were prepared by the PVD method. The deposition conditions were as follows: the total pressure was 5 × 10 −5 Torr, the evaporation temperature was 450 • C, the deposition time was 1 h, and the substrate temperature was 60 • C. It is known that when CoPc films are heated at temperatures above 200 • C, a phase transition from α-modification to β can occur, and the size of crystallites and their packing density can change, affecting the sensor properties. For this reason, before AuNPs deposition, CoPc films were annealed under the conditions of the MOCVD experiment (T = 280 • C, Ar/O 2 = 2, 1 h) to obtain CoPc films with the same phase composition in order to exclude the effect of this parameter on the sensor properties.
Second, the surface of annealed CoPc films was decorated with AuNPs using three different methods, viz. MOCVD, PVD, or DC. AuNPs were deposited on annealed CoPc films by MOCVD using (CH 3 ) 2 Au(thd) as a precursor in a reactor with cold walls, equipped with an Ar lamp. The evaporator temperature was 60 • C (the saturated vapor pressure of (CH 3 ) 2 Au(thd) was given by the equation ln(p, atm) = 36.61-15,231/T(K) [55] corresponding to the precursor partial pressure 5 × 10 −2 Torr). The deposition temperature was varied from 210 to 280 • C, and the optimal deposition temperature was 280 • C. Other experimental parameters were set as follows: the precursor weight was 3 mg, the ratio of the reactant gaseous mixture (Ar/O 2 ) in the reaction zone was 2; the total reactor pressure was~2 Torr; and the deposition time was 3 min. The deposition procedure was performed one, two, or three times sequentially, and the samples obtained in this way were designated as Au_CVD1/CoPc, Au_CVD2/CoPc, and Au_CVD3/CoPc, respectively ( Figure 1).
AuNPs were deposited on annealed CoPc films by PVD from the load of metallic Au in vacuum (P = 10 −6 Torr) using a UVM-71 installation. The load of gold was varied from 0.6 to 5 mg, and the load of 0.6 mg was chosen as optimal for obtaining samples. The evaporator temperature was 1532 • C (which corresponds to a partial pressure of Au 10 −2 Torr), and the substrate temperature was 100 • C. The deposition procedures were carried out until complete evaporation of Au (0.6 mg). This procedure was repeated one, two, or three times sequentially, and the samples obtained in this way were designated as Au_PVD1/CoPc, Au_PVD2/CoPc and Au_PVD3/CoPc, respectively ( Figure 1).
AuNPs were deposited on annealed CoPc films by DC of 5 × 10 −4 M AuNPs solution in hexane, prepared according to the method proposed by Eah et al. [56]. In brief, 400 µL of 0.05 M of NaBH 4 in an aqueous NaOH was added slowly to 100 µL of 0.05 M of HAuCl 4 ·3H 2 O in aqueous HCl, which led to the formation of a red solution. The distilled water was added to obtain 5 × 10 −4 M AuNPs solution. After that, 5 mL of acetone was added to 10 mL of 5 × 10 −4 M AuNPs aqueous solution and the mixture was shaken manually for 2 min. Then, 0.1 g of dodecanethiol in 5 mL of hexane was added rapidly and the mixture was shaken manually for 2 min. The separation of the water and hexane layers was accompanied by discoloration of the lower aqueous phase and staining of the upper organic phase. The layer of hexane was collected. The new portion of 0.1 g of dodecanethiol in 5 mL of hexane was added to the water layer, and the above procedure was repeated 5 times. After combining the hexane layers, the total mixture was concentrated in a vacuum rotary evaporator. The addition of 5 mL of ethanol to the concentrated hexane solution resulted in the precipitation of dark amorphous solid of AuNPs. To remove dodecanethiol, the precipitate was dissolved again in hexane and re-precipitated by ethanol. Finally, the precipitate was dissolved in 10 mL of hexane, giving~5 × 10 −4 M AuNPs solution. One, two, or three drops (one drop corresponded to 1 µL) were dripped onto the surface of CoPc films and dried in an oven at 150 • C for 2 h to obtain Au_DC1/CoPc, Au_DC2/CoPc, and Au_DC3/CoPc samples, respectively ( Figure 1).

Methods of Characterization of Heterostructures
The chemical state of gold in the heterostructures with AuNPs obtained by MOCVD and DC methods was investigated using X-ray photoelectron spectroscopy (XPS, SPECS Spectrometer, Berlin, Germany, PHOIBOS-150-MCD-9 analyzer, FOCUS-500 monochromator, Al Kα radiation, hv = 1486.74 eV, 200 W). The binding energy scale (E b ) was calibrated using the positions of the peak energy levels of the Au 4f 7/2 (E b = 84.0 eV) and Cu 2p 3/2 (E b = 932.67 eV). The spectra were processed in the CASA program (Japan) using the Voigt function. The background was taken into account using the Shirley method.
The elemental composition and surface morphology of heterostructures based on CoPc films and AuNPs obtained by CVD, PVD, and DC methods were investigated using scanning electron microscopes, JEOL-JSM 6700 F (connected with EDX-analizator EX-2300BU) and HI-TACHI UHR FE-SEM SU8200. The particle size distribution in the samples were determined using image analysis software (ImageJ, National Institutes of Health, 9000 Rockville Pike, Bethesda, Rockville, MA, USA). The composition and microstructure of Au_CVD1/CoPc and Au_DC1/CoPc samples deposited on a freshly cleaved NaCl crystal or a Cu grid were analyzed using transmission electron microscopy (TEM), electron diffraction, and energydispersive spectroscopy on a JEM-2100 transmission electron microscope (TEM, JEOL). The diffraction patterns were identified using the program DigitalMicrograph (Gatan, Pleasanton, CA, USA) and crystal structure database ICDD PDF 4+ (2020).

Study of the Sensor Properties
The study of the sensor properties of active layers based on CoPc and its heterostructures with AuNPs was carried out by measuring their chemiresistive sensor response to NH 3 and NO. CoPc films and heterostructures were deposited onto a glass substrate with interdigitated Pt electrodes. In this method, the measured value is the resistance of the film, which reversibly changes during the interaction of the layer with the gas to be determined and subsequent purging. The change in the resistance of hybrid film structures was measured using a Keithley 236 universal electrometer by applying a constant DC voltage (10 V). The sensor response (S resp ) was estimated as S resp = (R − R o )/R o , where R o and R are the initial sensor resistances (base resistance) and the resistances in the presence of analyzed gas, respectively. The standard deviation of the response values on each line was calculated using the data for three different samples.
The scheme of the system for investigation of the chemiresistive sensor response with a photo of the gas flow cell and the sensing layer deposited onto IDE is shown in Figure 2.
response to NH3 and NO. CoPc films and heterostructures were deposited onto a glass substrate with interdigitated Pt electrodes. In this method, the measured value is the resistance of the film, which reversibly changes during the interaction of the layer with the gas to be determined and subsequent purging. The change in the resistance of hybrid film structures was measured using a Keithley 236 universal electrometer by applying a constant DC voltage (10 V). The sensor response (Sresp) was estimated as Sresp = (R − Ro)/Ro, where Ro and R are the initial sensor resistances (base resistance) and the resistances in the presence of analyzed gas, respectively. The standard deviation of the response values on each line was calculated using the data for three different samples.
The scheme of the system for investigation of the chemiresistive sensor response with a photo of the gas flow cell and the sensing layer deposited onto IDE is shown in Figure 2. Pure commercial NH3 and NO gases (Company "Chistye Gasy", Novosibirsk, Russia) were used as analyte sources. Air was used as a carrier and diluent gas in the case of investigation of the sensor response to ammonia. Argon (Company "Chistye Gasy", Novosibirks, Russia) was used as a carrier gas and diluent gas in the case of investigation of the sensor response to NO because of its instability in air. The analyzed gases were diluted with the corresponding diluent gases to the required concentration and passed through the cell at the constant flow rate of 300 mL·min -1 . The required gas flow was regulated using mass flow regulators. After the input of the analyzed gas, the cell was purged with an air (or argon) flow until the initial resistance was restored. To achieve complete reversibility of the sensor response to NO, measurements were carried out at the temperature of 40-50 °C. For a correct comparison of the sensor properties of the layers, studies of the sensor response to ammonia were also carried out when the gas-flow chamber was heated at the same temperature.

Characterization of Heterostructures with AuNPs Obtained by MOCVD
To obtain AuNPs on СoPc films, the experimental parameters of the MOCVD process were chosen in accordance with our previous results [57] on the deposition of AuNPs from (CH3)2Au(thd) onto Si plates. It was experimentally established that AuNPs were grown on the СoPc surface at a deposition temperature of at least 280 °С, while the subsequent increase in the temperature was accompanied by the degradation of the CoPc film. The time and the precursor load were chosen experimentally to ensure the Pure commercial NH 3 and NO gases (Company "Chistye Gasy", Novosibirsk, Russia) were used as analyte sources. Air was used as a carrier and diluent gas in the case of investigation of the sensor response to ammonia. Argon (Company "Chistye Gasy", Novosibirks, Russia) was used as a carrier gas and diluent gas in the case of investigation of the sensor response to NO because of its instability in air. The analyzed gases were diluted with the corresponding diluent gases to the required concentration and passed through the cell at the constant flow rate of 300 mL·min −1 . The required gas flow was regulated using mass flow regulators. After the input of the analyzed gas, the cell was purged with an air (or argon) flow until the initial resistance was restored. To achieve complete reversibility of the sensor response to NO, measurements were carried out at the temperature of 40-50 • C. For a correct comparison of the sensor properties of the layers, studies of the sensor response to ammonia were also carried out when the gas-flow chamber was heated at the same temperature.

Characterization of Heterostructures with AuNPs Obtained by MOCVD
To obtain AuNPs on CoPc films, the experimental parameters of the MOCVD process were chosen in accordance with our previous results [57] on the deposition of AuNPs from (CH 3 ) 2 Au(thd) onto Si plates. It was experimentally established that AuNPs were grown on the CoPc surface at a deposition temperature of at least 280 • C, while the subsequent increase in the temperature was accompanied by the degradation of the CoPc film. The time and the precursor load were chosen experimentally to ensure the formation of AuNPs in each of the Au_CVD1/CoPc, Au_CVD2/CoPc, and Au_CVD3/CoPc samples.
The presence of gold in Au_CVD1/CoPc, Au_CVD2/CoPc, and Au_CVD3/CoPc was confirmed by ICP-AES data ( Table 1). The concentration of gold in the sample Au_CVD1/CoPc was 0.34 ± 0.05 µg/cm 2 . Each subsequent deposition was accompanied by a uniform increase in the concentration of gold in the samples: 0.64 ± 0.08 and 1.1 ± 0.1 µg/cm 2 for Au_CVD2/CoPc and Au_CVD3/CoPc, respectively. Due to the low concentration of gold in Au_CVD1/CoPc, characteristic Au4f lines were absent in its XPS spectrum, but these lines were visualized in the XPS spectra of Au_CVD2/CoPc and Au_CVD3/CoPc. The XPS spectrum of Au_CVD3/CoPc is shown in Figure 3a. Along with Au4f lines, the intense Co2p, C1s, and N1s lines related to CoPc films, as well as low-intensity O1s and Si2p lines from the substrate, were observed in the XPS spectra of Au_CVD2/CoPc and Au_CVD3/CoPc samples. The positions of the main Au4f peaks in the XPS spectrum of Au_CVD3/CoPc indicated the presence of only metallic gold (Au4f 7/2 84.1 eV) in this sample (Figure 3b). This confirms the successful deposition of metallic gold on CoPc by the MOCVD method using [(CH 3 ) 2 Au(thd)] as a precursor. The Co2p can be resolved into two peaks, 2p3/2 and 2p1/2, at 780.3 (its satellite at 782.8 eV) and 796 eV, respectively (Figure 3c). The positions of these peaks are in good agreement with those for the annealed CoPc film described by Kumar et al. [58]. The XPS spectrum of C1s (Figure 3d) exhibited three peaks at about 284.6, 286, and 287.8 eV, assigned to C-C, C-N, and C=N, respectively. The N1s binding energy of the Au_CVD3/CoPc sample showed three types of nitrogen, and peaks at 398.5 and 400.5 eV The Co2p can be resolved into two peaks, 2p3/2 and 2p1/2, at 780.3 (its satellite at 782.8 eV) and 796 eV, respectively (Figure 3c). The positions of these peaks are in good agreement with those for the annealed CoPc film described by Kumar et al. [58]. The XPS spectrum of C1s (Figure 3d) exhibited three peaks at about 284.6, 286, and 287.8 eV, assigned to C-C, C-N, and C=N, respectively. The N1s binding energy of the Au_CVD3/CoPc sample showed three types of nitrogen, and peaks at 398.5 and 400.5 eV were assigned to the nitrogen of the phthalocyanine ring [59], while the peak at 399.3 eV was ascribed to the coordination of cobalt and nitrogen (Figure 3e). Thus, there was no degradation of the CoPc film even after three consecutive MOCVD steps (Au_CVD3/CoPc sample). In the O1s spectrum (Figure 3f), oxygen was in a single state (532.9 eV), which could be attributed to Si-O [60].
The TEM method was used to study the composition and morphology of the heterostructures. Figure 4a shows a TEM image of the Au_CVD1/CoPc sample as an example. The diffraction pattern of Au_CVD1/CoPc contained sets of ring-type reflections from fcc-Au (JCPDS Card No. 04-0784) with polycrystalline structure and from the β-CoPc phase. The surface of the Au_CVD1/CoPc sample was formed by spherical AuNPs evenly distributed over the surface of the CoPc film. This sample was characterized by a unimodal size distribution of AuNPs with a predominant size of 5 nm (Figure 4b). According to the EDX spectrum, the gold content in Au_CVD1/CoPc was 2.8 at.% ( Figure 4c).  During the second deposition (the Au_CVD2/CoPc sample, Figure 5a,b), AuNPs coalesced, which was accompanied by a two-fold increase in particle size compared to the Au_CVD1/CoPc sample. The Au content in the Au_CVD2/CoPc sample was 4.9 at% ( Figure S1, Supporting Information) in accordance with the EDX spectrum. During the third deposition (the Au_CVD3/CoPc sample, Figure 5c,d), Au nanoclusters consisting of 2-4 particles with sizes from 8 to 26 nm, with predominant sizes of 12-20 nm, were formed on the surface of CoPc films. The Au content in the Au_CVD3/CoPc sample was 7.6 at% ( Figure S1, Supporting Information). The average sizes of Au nanoclusters in the heterostructures are summarized in Table 1. Thus, during MOCVD, the growth of AuNPs on CoPc films was accompanied by the nucleation and subsequent fusion of AuNPs with the formation of Au nanoclusters. During the second deposition (the Au_CVD2/CoPc sample, Figure 5a,b), AuNPs coalesced, which was accompanied by a two-fold increase in particle size compared to the Au_CVD1/CoPc sample. The Au content in the Au_CVD2/CoPc sample was 4.9 at% ( Figure S1, Supporting Information) in accordance with the EDX spectrum. During the third deposition (the Au_CVD3/CoPc sample, Figure 5c,d), Au nanoclusters consisting of 2-4 particles with sizes from 8 to 26 nm, with predominant sizes of 12-20 nm, were formed on the surface of CoPc films. The Au content in the Au_CVD3/CoPc sample was 7.6 at% ( Figure S1, Supporting Information). The average sizes of Au nanoclusters in the heterostructures are summarized in Table 1. Thus, during MOCVD, the growth of AuNPs on CoPc films was accompanied by the nucleation and subsequent fusion of AuNPs with the formation of Au nanoclusters. third deposition (the Au_CVD3/CoPc sample, Figure 5c,d), Au nanoclusters consisting of 2-4 particles with sizes from 8 to 26 nm, with predominant sizes of 12-20 nm, were formed on the surface of CoPc films. The Au content in the Au_CVD3/CoPc sample was 7.6 at% ( Figure S1, Supporting Information). The average sizes of Au nanoclusters in the heterostructures are summarized in Table 1. Thus, during MOCVD, the growth of AuNPs on CoPc films was accompanied by the nucleation and subsequent fusion of AuNPs with the formation of Au nanoclusters.

Characterization of Heterostructures with AuNPs Obtained by PVD
In the PVD processes, the Au loads were varied from 0.6 to 5 mg in order to obtain samples with concentrations comparable to Au_CVD1/CoPc, Au_CVD2/CoPc, and Au_CVD3/CoPc samples. According to the ICP-AES data, the concentration of gold in Au_PVD1/CoPc was 0.47 ± 0.07 µg/cm 2 at the Au load of 0.6 mg. The concentrations of gold in Au_PVD2/CoPc and Au_PVD3/CoPc were 0.99 ± 0.09 and 2.1 ± 0.2 µg/cm 2 , respectively ( Table 1). As a result, Au_CVD3/CoPc and Au_PVD2/CoPc samples had similar concentrations of gold.
According to SEM data, ultradispersed AuNPs (sizes up to 6.5 nm) with uniform distribution were formed on the surface of CoPc films after the first deposition (Au_PVD1/CoPc sample, Figure 6a,b).

Characterization of Heterostructures with AuNPs Obtained by PVD
In the PVD processes, the Au loads were varied from 0.6 to 5 mg in order to obtain samples with concentrations comparable to Au_CVD1/CoPc, Au_CVD2/CoPc, and Au_CVD3/CoPc samples. According to the ICP-AES data, the concentration of gold in Au_PVD1/CoPc was 0.47 ± 0.07 µg/cm 2 at the Au load of 0.6 mg. The concentrations of gold in Au_PVD2/CoPc and Au_PVD3/CoPc were 0.99 ± 0.09 and 2.1 ± 0.2 µg/cm 2 , respectively ( Table 1). As a result, Au_CVD3/CoPc and Au_PVD2/CoPc samples had similar concentrations of gold.
According to SEM data, ultradispersed AuNPs (sizes up to 6.5 nm) with uniform distribution were formed on the surface of CoPc films after the first deposition (Au_PVD1/CoPc sample, Figure 6a,b). After the second deposition (Au_PVD2/CoPc sample, Figure 6c,d), the unimodal distribution of AuNPs on the surface of CoPc film remained, while the average size of AuNPs increased by no more than 1.3-1.4 times compared to that for the Au_PVD1/CoPc sample (Table 1). After the third deposition (sample Au_PVD3/CoPc, Figure 6e,f), Au After the second deposition (Au_PVD2/CoPc sample, Figure 6c,d), the unimodal distribution of AuNPs on the surface of CoPc film remained, while the average size of AuNPs increased by no more than 1.3-1.4 times compared to that for the Au_PVD1/CoPc sample (Table 1). After the third deposition (sample Au_PVD3/CoPc, Figure 6e,f), Au nanoclusters consisting of 2-3 particles with sizes from 7 to 14 nm, with a predominance of sizes 8-12 nm, formed on the surface of CoPc films. Au content in the Au_PVD1/CoPc, Au_PVD2/CoPc, and Au_PVD3/CoPc samples according to their EDX spectra was 3.7, 6.1 and 9.1 at%, respectively ( Figure S2). Thus, the PVD growth of AuNPs on CoPc films was accompanied by the formation of particles with a narrower size distribution compared to the samples obtained by MOCVD.

Characterization of Heterostructures with AuNPs Obtained by DC
During deposition by drop casting, Au concentrations on heterostructures were set experimentally to obtain Au_DC1/CoPc, Au_DC2/CoPc, and Au_DC3/CoPc samples with Au content of 0.33, 0.66, and 0.99 µg/cm 2 , respectively, taking into account the geometric dimensions of the substrates (IDE, working zone S = 0.35 cm 2 ) ( Table 1). The chemical state of gold in the samples was studied by the XPS method. The XPS spectra of Au_DC3/CoPc are shown in Figure 7 as an example. In the XPS spectrum of Au_DC3/CoPc, along with Au4f lines, intense C1s, N1s, and Co2p lines related to CoPc films, as well as O1s and Si2p lines from the substrate, were observed ( Figure 7a).    (Figure 7d,e) confirms the presence of CoPc in the Au_DC3/CoPc sample. In the O1s spectrum (Figure 7f), the peak at 532.9 eV corresponds to Si-O [60].
Ultradispersed AuNPs are observed on the TEM image of Au_DC1/CoPc (Figure 8a,b). Their sizes (5.5-6 nm) correspond to those in an AuNP solution studied by Eah et al. [56]. In addition to ultradispersed AuNPs, the oval Au nanoclusters with sizes up to 14 nm were also observed on the surface of Au_DC1/CoPc. Unlike Au_CVD1/CoPc and Au_PVD1/CoPc samples, in the case of Au_DC1/CoPc, single AuNPs were scattered on the surface of the CoPc film (Figure 8b). Such a distribution may be due to the diffusion of AuNPs solution into the depth of the CoPc film. The gold content in the Au_CVD1/CoPc sample was 1.9 at.% according to its EDX spectrum (Figure 8c). After the second DC deposition (sample Au_DC2/CoPc, Figure 9a,b), AuNPs with a bimodal size distribution were observed on the CoPc surface. The size of AuNPs in Au_DC2/CoPc varied from 6 to 24 nm (Table 1), while their distribution became more dense. From the EDX spectrum, the gold content in Au_DC2/CoPc was 3.1 at.% ( Figure  S3). After the third DC deposition of the AuNPs (sample Au_DC3/CoPc, Figure 9c,d), the surface of the CoPc was covered with Au nanoclusters with a wide size distribution (the predominant sizes were 17-20 nm). The gold content in Au_DC3/CoPc was 5.1 at.% ( Figure S3). Thus, the distribution of AuNPs obtained by DC on CoPc films was not homogeneous and was accompanied by their fusion with the formation of Au nanoclusters with sizes up to 45 nm. After the second DC deposition (sample Au_DC2/CoPc, Figure 9a,b), AuNPs with a bimodal size distribution were observed on the CoPc surface. The size of AuNPs in Au_DC2/CoPc varied from 6 to 24 nm (Table 1), while their distribution became more dense. From the EDX spectrum, the gold content in Au_DC2/CoPc was 3.1 at.% ( Figure S3). After the third DC deposition of the AuNPs (sample Au_DC3/CoPc, Figure 9c,d), the surface of the CoPc was covered with Au nanoclusters with a wide size distribution (the predominant sizes were 17-20 nm). The gold content in Au_DC3/CoPc was 5.1 at.% ( Figure S3). Thus, the distribution of AuNPs obtained by DC on CoPc films was not homogeneous and was accompanied by their fusion with the formation of Au nanoclusters with sizes up to 45 nm. dense. From the EDX spectrum, the gold content in Au_DC2/CoPc was 3.1 at.% ( Figure  S3). After the third DC deposition of the AuNPs (sample Au_DC3/CoPc, Figure 9c,d), the surface of the CoPc was covered with Au nanoclusters with a wide size distribution (the predominant sizes were 17-20 nm). The gold content in Au_DC3/CoPc was 5.1 at.% ( Figure S3). Thus, the distribution of AuNPs obtained by DC on CoPc films was not homogeneous and was accompanied by their fusion with the formation of Au nanoclusters with sizes up to 45 nm.

Study of the Sensor Response of AuNP/CoPc Heterostructures to Ammonia and Nitric Oxide
The study of the sensor properties of active layers based on CoPc and its heterostructures with AuNPs was carried out by measuring their chemiresistive sensor response to NH 3 and NO. In this method, the measured value is the resistance (R, Figure 10), which reversibly changes during the interaction of the layer with the gaseous analyte and the subsequent purging. The typical sensor response of an Au_PVD3/CoPc sample to NO (0.1-0.7 ppb) and NH 3 (1-7 ppm) is presented in Figure 10 as an example. When gaseous NO is introduced into the measuring cell, a sharp decrease in resistance is observed (Figure 10a). The resistance returns to its original value with subsequent purging with argon. In contrast, the introduction of ammonia leads to an increase in the film resistance ( Figure 10b).

Study of the Sensor Response of AuNP/CoPc Heterostructures to Ammonia and Nitric
The study of the sensor properties of active layers based on CoPc an heterostructures with AuNPs was carried out by measuring their chemiresistive s response to NH3 and NO. In this method, the measured value is the resistance (R, F 10), which reversibly changes during the interaction of the layer with the gaseous an and the subsequent purging. The typical sensor response of an Au_PVD3/CoPc sam NO (0.1-0.7 ppb) and NH3 (1-7 ppm) is presented in Figure 10 as an example. W gaseous NO is introduced into the measuring cell, a sharp decrease in resistan observed (Figure 10a). The resistance returns to its original value with subseq purging with argon. In contrast, the introduction of ammonia leads to an increase i film resistance (Figure 10b). This behavior was observed for both bare CoPc films and all investi heterostructures. Similar behavior was also observed in other works [61,62], w unsubstituted phthalocyanines and their hybrids with carbon nanomaterials, whic p-type semiconductors, were used as active layers. When the adsorption of NH3 o surface occurs, the electron density transfers from the NH3 molecule to the CoPc l This reduces the concentration of the major charge carriers and increases the resist The opposite effect is observed in the case of NO.
To reveal the effect of the deposition of Au nanoparticles on the sensor propert CoPc films, the sensor response of CoPc films to NH3 and NO was compared with th This behavior was observed for both bare CoPc films and all investigated heterostructures. Similar behavior was also observed in other works [61,62], where unsubstituted phthalocyanines and their hybrids with carbon nanomaterials, which are p-type semiconductors, were used as active layers. When the adsorption of NH 3 on the surface occurs, the electron density transfers from the NH 3 molecule to the CoPc layer. This reduces the concentration of the major charge carriers and increases the resistance. The opposite effect is observed in the case of NO.
To reveal the effect of the deposition of Au nanoparticles on the sensor properties of CoPc films, the sensor response of CoPc films to NH 3 and NO was compared with that of the heterostructures in which AuNPs were deposited by three different techniques. Figure 11a,b shows the dependencies of the sensor response of CoPc films and heterostructures, in which Au nanoparticles were deposited by the PVD technique, on the concentration of NO and NH 3 , respectively. It is seen that the deposition of AuNPs on the surface of CoPc films leads to a substantial increase in the sensor response to both gases. For example, the sensor response of Au_PVD1/CoPc to NO (10 ppb) is four times higher compared to a CoPc film. The further increase in the number of nanoparticles in the samples Au_PVD2/CoPc and Au_PVD3/CoPc leads to the 7-fold and 16-fold growth of the sensor response to NO, respectively (Figure 11a). In the case of ammonia detection, the sensor response also increases in the order of CoPc < Au_PVD1/CoPc < Au_PVD2/CoPc < Au_PVD3/CoPc. The value of the response of Au_PVD3/CoPc to NH 3 (at 2 ppm) is 3.3 times higher than in the case of a CoPc film (Figure 11b). It is important to note that a further increase in the concentration of AuNPs by increasing the number of PVD cycles to four and five does not result in significant growth in the response to both NO and NH 3 . The dependence of the sensor response to NO and NH 3 on the number of PVD cycles is shown in Figure S4. This means that the number of nanoparticles on the surface of CoPc films has a significant effect on the sensor sensitivity. The Au_PVD3/CoPc heterostructure with the Au content of 2.1 µg/cm 2 and the average size of Au nanoparticles of 7-14 nm has an optimal composition to ensure the maximum response to both NO and NH 3 . Similar to the data presented by Korotcenkov et al. [63], the growth of the sensor response is observed until a certain concentration of Au nanoparticles is reached. A further increase in the number of nanoparticles can lead to the formation of large clusters and the overgrowth of the film surface, thereby reducing the number of active centers.   Figure 12a,b shows a comparison of the sensor response of CoPc films to NO and NH3 with that of heterostructures in which Au nanoparticles were deposited by MOCVD. Similar to heterostructures with AuNPs deposited by PVD, the sensor response of the heterostructure is higher than in the case of CoPc films and increases when the number of nanoparticles grows. The Au_CVD3/CoPc heterostructure demonstrates a 10-fold increase in the sensor response toward NO and a 2-fold increase in the response to NH3 compared to the original CoPc film, which is noticeably lower than in the case of similar heterostructures obtained by PVD (see above). Similarly to the case of heterostructures prepared by PVD, the further increase in the number of CVD cycles does not lead to an increase in the sensor response ( Figure S5).  Figure 12a,b shows a comparison of the sensor response of CoPc films to NO and NH 3 with that of heterostructures in which Au nanoparticles were deposited by MOCVD. Similar to heterostructures with AuNPs deposited by PVD, the sensor response of the heterostructure is higher than in the case of CoPc films and increases when the number of nanoparticles grows. The Au_CVD3/CoPc heterostructure demonstrates a 10-fold increase in the sensor response toward NO and a 2-fold increase in the response to NH 3 compared to the original CoPc film, which is noticeably lower than in the case of similar heterostructures obtained by PVD (see above). Similarly to the case of heterostructures prepared by PVD, the further increase in the number of CVD cycles does not lead to an increase in the sensor response ( Figure S5). heterostructure is higher than in the case of CoPc films and increases when the number of nanoparticles grows. The Au_CVD3/CoPc heterostructure demonstrates a 10-fold increase in the sensor response toward NO and a 2-fold increase in the response to NH3 compared to the original CoPc film, which is noticeably lower than in the case of similar heterostructures obtained by PVD (see above). Similarly to the case of heterostructures prepared by PVD, the further increase in the number of CVD cycles does not lead to an increase in the sensor response ( Figure S5). The sensor response of heterostructures in which different concentrations of AuNPs were deposited by drop casting was also compared. Figure 13a,b shows the same trend as in the case of heterostructures prepared by PVD and CVD methods: the sensor responses The sensor response of heterostructures in which different concentrations of AuNPs were deposited by drop casting was also compared. Figure 13a,b shows the same trend as in the case of heterostructures prepared by PVD and CVD methods: the sensor responses toward both NO and NH 3 increase in the order of CoPc < Au_DC1/CoPc < Au_DC2/CoPc < Au_DC3/CoPc. In the case of NO (10 ppb) detection, the response of Au_DC3/CoPc is about six times higher than that of the CoPc film, while the response toward NH 3 (2 ppm) shows an increase of only two times.

Comparison of the Sensor Properties of Heterostructures with AuNPs Deposited in Different Ways
The comparative analysis described above showed that in the case of heterostructures in which nanoparticles were deposited in three different ways, a similar trend is observed, namely, when the concentration of nanoparticles increases to a certain value, the sensor response to both ammonia and nitric oxide increases. To understand which method is the most preferable in terms of increasing the sensitivity of the active layer, the sensor characteristics of heterostructures containing a similar amount of AuNPs were compared. As shown above, the heterostructures Au_PVD2/CoPc, Au_CVD3/CoPc, and Au_DC3/CoPc have almost the same concentration of gold. The dependencies of the sensor response for Au_PVD2/CoPc, Au_CVD3/CoPc, and Au_DC3/CoPc samples on the concentrations of NO and NH3 are shown in Figure 14. Figure 14 shows that Au_CVD3/CoPc demonstrates the highest sensitivity to NO in the concentration range from 10 to 30 ppb. At 40-50 ppm NO, the difference between the sensor responses of Au_CVD3/CoPc and Au_PVD2/CoPc is within the measurement error. Thus, Au_PVD2/CoPc and Au_CVD3/CoPc samples, which are characterized by a narrow distribution of small sizes of AuNPs or Au nanoclusters, have the highest sensor responses to NO. The heterostructure Au-PVD2/CoPc with a particle size of 8 nm is

Comparison of the Sensor Properties of Heterostructures with AuNPs Deposited in Different Ways
The comparative analysis described above showed that in the case of heterostructures in which nanoparticles were deposited in three different ways, a similar trend is observed, namely, when the concentration of nanoparticles increases to a certain value, the sensor response to both ammonia and nitric oxide increases. To understand which method is the most preferable in terms of increasing the sensitivity of the active layer, the sensor characteristics of heterostructures containing a similar amount of AuNPs were compared. As shown above, the heterostructures Au_PVD2/CoPc, Au_CVD3/CoPc, and Au_DC3/CoPc have almost the same concentration of gold. The dependencies of the sensor response for Au_PVD2/CoPc, Au_CVD3/CoPc, and Au_DC3/CoPc samples on the concentrations of NO and NH 3 are shown in Figure 14. Figure 14 shows that Au_CVD3/CoPc demonstrates the highest sensitivity to NO in the concentration range from 10 to 30 ppb. At 40-50 ppm NO, the difference between the sensor responses of Au_CVD3/CoPc and Au_PVD2/CoPc is within the measurement error. Thus, Au_PVD2/CoPc and Au_CVD3/CoPc samples, which are characterized by a narrow distribution of small sizes of AuNPs or Au nanoclusters, have the highest sensor responses to NO. The heterostructure Au-PVD2/CoPc with a particle size of 8 nm is characterized by the most rapid increase in response with an increase in NO concentration. According to [63], the smaller the Au nanoparticles, the greater their activity in charge transfer processes, as well as in the adsorption and desorption reactions of molecular gases. Interestingly, unlike NO, the sensor response to NH3 is practically the same for heterostructures in which nanoparticles were obtained by vapor deposition and wet methods. As has already been mentioned above, these heterostructures had almost the same concentration of gold but different sizes and distribution of AuNPs (Table 1). This behavior may be related to the different mechanisms of interaction of NO and NH3 with heterostructures based on phthalocyanine and gold nanoparticles, which are discussed below.
The improvement in the sensitivity of the Au/CoPc heterostructures to both gases with an increase in the concentration of AuNPs may be associated with the sensitization mechanism. Based on the literature analysis [63][64][65][66][67], we believe that in Au/CoPc heterostructures the charge transfers from CoPc to Au and makes AuNPs negatively charged, increasing the number of hole carriers on the CoPc surface. Note that the concentration of hole carriers decreases with the adsorption of donor gas, but increases with the adsorption of acceptor gas [65,68]. Comparing the two sensitization mechanisms described in the literature [65], we can assume that in the case of Au/CoPc heterostructures, the most likely mechanism is electronic sensitization, in which the main role of Au is to increase the number of hole carriers. The mechanism of chemical sensitization based on the formation of O2 − (ad.) species that would react with NH3(ad.), giving electrons back to CoPc, seems unlikely here due to the low operating temperatures of the sensor (50 °C).
The difference in the behavior of NO and NH3 gases may be related to the stronger binding of NO with AuNPs. According to the work of Chinh et al. [69], in Au/ZnO heterostructures, NO molecules can adsorb on Au cores, forming NO − (ad.) species, according to Equation (1): (1) Figure 14. Dependence of the sensor response of Au_PVD2/CoPc, Au_CVD3/CoPc, and Au_DC3/CoPc on NO (blue lines) and NH 3 (red lines) concentrations.
In the case of nanoparticles obtained by the DC method, the sensor response to nitric oxide is 2-3 times lower compared to the heterostructures obtained by the other two methods. This may be due to the possible presence of an organic shell of AuNPs necessary for their stabilization when they are obtained by the wet method. A similar trend is also observed when comparing the sensor response of heterostructures with gold nanoparticles of the same size.
Interestingly, unlike NO, the sensor response to NH 3 is practically the same for heterostructures in which nanoparticles were obtained by vapor deposition and wet methods. As has already been mentioned above, these heterostructures had almost the same concentration of gold but different sizes and distribution of AuNPs (Table 1). This behavior may be related to the different mechanisms of interaction of NO and NH 3 with heterostructures based on phthalocyanine and gold nanoparticles, which are discussed below.
The improvement in the sensitivity of the Au/CoPc heterostructures to both gases with an increase in the concentration of AuNPs may be associated with the sensitization mechanism. Based on the literature analysis [63][64][65][66][67], we believe that in Au/CoPc heterostructures the charge transfers from CoPc to Au and makes AuNPs negatively charged, increasing the number of hole carriers on the CoPc surface. Note that the concentration of hole carriers decreases with the adsorption of donor gas, but increases with the adsorption of acceptor gas [65,68]. Comparing the two sensitization mechanisms described in the literature [65], we can assume that in the case of Au/CoPc heterostructures, the most likely mechanism is electronic sensitization, in which the main role of Au is to increase the number of hole carriers. The mechanism of chemical sensitization based on the formation of O 2 − ( ad. ) species that would react with NH 3 ( ad. ), giving electrons back to CoPc, seems unlikely here due to the low operating temperatures of the sensor (50 • C).
The difference in the behavior of NO and NH 3 gases may be related to the stronger binding of NO with AuNPs. According to the work of Chinh et al. [69], in Au/ZnO heterostructures, NO molecules can adsorb on Au cores, forming NO − (ad.) species, according to Equation (1): In this case, the sensor response is the result of a dynamic equilibrium between NO adsorption and NO − (ad) desorption. Thus, the role of Au nanoparticles is to reduce the energy barrier of adsorption and desorption for NO and generate NO − (ad.) species. Recently, Korotcenkov et al. [63,64] described the mechanism of chemical sensitization of gold-containing heterostructures based on the generation of O − (ad.) species near Au clusters, which then further react with gases. Note that the mechanism of chemical sensitization depends entirely on the efficiency of generation and the availability of these species [63,65,66,70]. Apparently, in the case of the studied heterostructures, mixed sensitization occurs, which is based on the generation of NO − (ad.) species that can migrate to hole carriers in the CoPc film. In this case, the adsorption/desorption processes that determine gas sensitivity have an activation character, and the parameters describing these processes depend on the grain size and distribution [66]. Thus, the sensor response of Au/CoPc heterostructures to NO depends not only on the concentration of gold but also on the size of the AuNPs and distribution parameters, which are responsible for the efficiency of the generation and migration of NO − (ad.) species.

Investigation of the Sensor Characteristics of AuNP/CoPc Heterostructures
Au_PVD3/CoPc was shown above to exhibit the best sensor response to both ammonia and nitric oxide among the investigated heterostructures. For this reason, it were chosen for a more detailed investigation of its sensor characteristics and to assess the possibility of its use for detecting low concentrations of ammonia and nitric oxide in exhaled air. The sensitivity of the Au_PVD3/CoPc sensor to NO and NH 3 was tested in a wider range of concentrations. The dependences of the sensor response on the concentration of NO (10-1000 ppb) and NH 3 (0.5-100 ppm) are shown in Figure 15a Recently, Korotcenkov et al. [63,64] described the mechanism of chemical sensitization of gold-containing heterostructures based on the generation of O -(ad.) species near Au clusters, which then further react with gases. Note that the mechanism of chemical sensitization depends entirely on the efficiency of generation and the availability of these species [63,65,66,70]. Apparently, in the case of the studied heterostructures, mixed sensitization occurs, which is based on the generation of NO − (ad.) species that can migrate to hole carriers in the CoPc film. In this case, the adsorption/desorption processes that determine gas sensitivity have an activation character, and the parameters describing these processes depend on the grain size and distribution [66]. Thus, the sensor response of Au/CoPc heterostructures to NO depends not only on the concentration of gold but also on the size of the AuNPs and distribution parameters, which are responsible for the efficiency of the generation and migration of NO − (ad.) species.

Investigation of the Sensor Characteristics of AuNP/CoPc Heterostructures
Au_PVD3/CoPc was shown above to exhibit the best sensor response to both ammonia and nitric oxide among the investigated heterostructures. For this reason, it were chosen for a more detailed investigation of its sensor characteristics and to assess the possibility of its use for detecting low concentrations of ammonia and nitric oxide in exhaled air. The sensitivity of the Au_PVD3/CoPc sensor to NO and NH3 was tested in a wider range of concentrations. The dependences of the sensor response on the concentration of NO (10-1000 ppb) and NH3 (0.5-100 ppm) are shown in Figure 15a,b. The dependence of the response on the concentration of NO was linear in the range from 10 to 100 ppb, while on the concentration of NH3 the dependence was linear from 0.5 to 10 ppm. Both of these intervals cover the concentration range important for determining these gases in the exhaled air [10]. The limit of detection (LOD) of NO, defined as 3σ/m, where σ is the standard deviation of the sensor response to 10 ppm NH3 and m is the slope of the corresponding calibration plot (Figure 15) in the linear region, was 4 ppb, while the response and recovery times of the sensor, determined at 10 ppb of NO, were 40 s and 100 s, respectively. The LOD of NH3 was 0.1 ppm; the response and recovery times determined at 1 ppm of NH3 were 20 and 85 s, respectively.
The reproducibility of the response was studied by repeated exposure to different The dependence of the response on the concentration of NO was linear in the range from 10 to 100 ppb, while on the concentration of NH 3 the dependence was linear from 0.5 to 10 ppm. Both of these intervals cover the concentration range important for determining these gases in the exhaled air [10]. The limit of detection (LOD) of NO, defined as 3σ/m, where σ is the standard deviation of the sensor response to 10 ppm NH 3 and m is the slope of the corresponding calibration plot (Figure 15) in the linear region, was 4 ppb, while the response and recovery times of the sensor, determined at 10 ppb of NO, were 40 s and 100 s, respectively. The LOD of NH 3 was 0.1 ppm; the response and recovery times determined at 1 ppm of NH 3 were 20 and 85 s, respectively.
The reproducibility of the response was studied by repeated exposure to different concentrations of the investigated analytes, as shown in Figure 16. Au_PVD3/CoPc sensing layers showed very good repeatability of the sensor response to both ammonia and nitric oxide. Because the operating temperature is a unique characteristic of each sensor, the sensor response of a Au_PVD3/CoPc layer toward NH 3 and NO was tested at temperatures from 40 to 100 • C ( Figure S6). An increase in the temperature to 70 • C practically does not affect the sensor response to both gases, while its further growth to 100 • C leads to an increase in the response. This behavior appears to be due to the growth of the concentration of charge carriers in p-type semiconductors with temperature. practically does not affect the sensor response to both gases, while its further growth to 100 °C leads to an increase in the response. This behavior appears to be due to the growth of the concentration of charge carriers in p-type semiconductors with temperature. Selectivity is another parameter important for the practical application of chemical sensors. To investigate the selectivity, the response of the Au_PVD3/CoPc layers to various gaseous analytes was tested (Figure 17a). It is obvious that the response of sensing layers to both NH3 and NO gases was much higher than to interfering analytes such as carbon dioxide, acetone, ethanol, hydrogen, and a mixture of aldehydes (butanal, heptanal, and octanal). It is necessary to mention that the concentration of ammonia and nitric oxide given in the diagram is much less than that of the other volatile organic analytes. These results make the investigated sensing layers attractive for detecting ammonia and nitric oxide in gas mixtures.
To demonstrate the possible application of the investigated heterostructures for the detection of ammonia in exhaled air, the sensor response of Au_PVD3/CoPc was tested in a mixture of gases with a composition close to the exhaled air of healthy people (N2-74%, O2-16%, H2O-6%, and CO2-4%). For this purpose, small amounts of ammonia (1-4 v.%) were added to the preliminarily prepared gas mixture (N2-74%, O2-16%, H2O-6%, and CO2-4%). The sensor response of Au_PVD3/CoPc to ammonia (1-4 ppm) in the gas mixture was compared with that to ammonia in air, as shown in Figure 17b. It was shown that the determination of ammonia in a mixture similar in composition to exhaled air did not lead to a significant change (within the measurement error) in the sensor response to ammonia measured in air. Thus, it was shown that the investigated heterostructures are promising materials for use as the active layers of sensors for determining low concentrations of ammonia in exhaled air. Selectivity is another parameter important for the practical application of chemical sensors. To investigate the selectivity, the response of the Au_PVD3/CoPc layers to various gaseous analytes was tested (Figure 17a). It is obvious that the response of sensing layers to both NH 3 and NO gases was much higher than to interfering analytes such as carbon dioxide, acetone, ethanol, hydrogen, and a mixture of aldehydes (butanal, heptanal, and octanal). It is necessary to mention that the concentration of ammonia and nitric oxide given in the diagram is much less than that of the other volatile organic analytes. These results make the investigated sensing layers attractive for detecting ammonia and nitric oxide in gas mixtures.
practically does not affect the sensor response to both gases, while its further growth to 100 °C leads to an increase in the response. This behavior appears to be due to the growth of the concentration of charge carriers in p-type semiconductors with temperature. Selectivity is another parameter important for the practical application of chemical sensors. To investigate the selectivity, the response of the Au_PVD3/CoPc layers to various gaseous analytes was tested (Figure 17a). It is obvious that the response of sensing layers to both NH3 and NO gases was much higher than to interfering analytes such as carbon dioxide, acetone, ethanol, hydrogen, and a mixture of aldehydes (butanal, heptanal, and octanal). It is necessary to mention that the concentration of ammonia and nitric oxide given in the diagram is much less than that of the other volatile organic analytes. These results make the investigated sensing layers attractive for detecting ammonia and nitric oxide in gas mixtures.
To demonstrate the possible application of the investigated heterostructures for the detection of ammonia in exhaled air, the sensor response of Au_PVD3/CoPc was tested in a mixture of gases with a composition close to the exhaled air of healthy people (N2-74%, O2-16%, H2O-6%, and CO2-4%). For this purpose, small amounts of ammonia (1-4 v.%) were added to the preliminarily prepared gas mixture (N2-74%, O2-16%, H2O-6%, and CO2-4%). The sensor response of Au_PVD3/CoPc to ammonia (1-4 ppm) in the gas mixture was compared with that to ammonia in air, as shown in Figure 17b. It was shown that the determination of ammonia in a mixture similar in composition to exhaled air did not lead to a significant change (within the measurement error) in the sensor response to ammonia measured in air. Thus, it was shown that the investigated heterostructures are promising materials for use as the active layers of sensors for determining low concentrations of ammonia in exhaled air. To demonstrate the possible application of the investigated heterostructures for the detection of ammonia in exhaled air, the sensor response of Au_PVD3/CoPc was tested in a mixture of gases with a composition close to the exhaled air of healthy people (N 2 -74%, O 2 -16%, H 2 O-6%, and CO 2 -4%). For this purpose, small amounts of ammonia (1-4 v.%) were added to the preliminarily prepared gas mixture (N 2 -74%, O 2 -16%, H 2 O-6%, and CO 2 -4%). The sensor response of Au_PVD3/CoPc to ammonia (1-4 ppm) in the gas mixture was compared with that to ammonia in air, as shown in Figure 17b. It was shown that the determination of ammonia in a mixture similar in composition to exhaled air did not lead to a significant change (within the measurement error) in the sensor response to ammonia measured in air. Thus, it was shown that the investigated heterostructures are promising materials for use as the active layers of sensors for determining low concentrations of ammonia in exhaled air.
Due to the low stability of NO in air, its determination in exhaled air samples is possible only with strict control of the inert atmosphere. For this reason, we did not determine gaseous NO in the gas mixture close to exhaled air. It has been shown that Au/CoPc hybrid materials are very promising for the determination of nitric oxide. The properties of the sensors studied in this work are superior in some characteristics to the sensors described in recent years ( Table 2). They have high sensitivity to NO, good selectivity, relatively short response and recovery times, and, importantly, they operate at significantly lower temperatures with a low detection limit. To detect NO in oxygen-and water-containing gas mixtures, some researchers first quantitatively oxidized NO to NO 2 , and then conducted a study of the sensor response to NO 2 . The use of this approach will be the subject of a separate publication.

Conclusions
In this work, new heterostructures based on cobalt phthalocyanine (CoPc) films decorated with gold nanoparticles (AuNPs) were obtained. The AuNPs were deposited on the surface of CoPc films by gas-phase methods (MOCVD, PVD) and drop casting (DC). Within each method, AuNPs were deposited onto CoPc films one, two, or three times sequentially to provide a proportional increase in the concentration of gold in Au/CoPc heterostructures. Au/CoPc heterostructures with a narrow distribution of small sizes of AuNPs were obtained by gas-phase methods, while the drop casting resulted in the production of Au/CoPc heterostructures with a scattered or inhomogeneous distribution of AuNPs.
The effect of gold concentration and the size of AuNPs on the sensor response of Au/CoPc heterostructures to NH 3 (1-10 ppm) and NO (10-50 ppb) was investigated. Regardless of the fabrication method, the response of Au/CoPc heterostructures to NH 3 and NO gases increased with an increase in the concentration of gold. The sensor response of Au/CoPc heterostructures to NH 3 increased 2-3.3 times compared to CoPc film, whereas in the case of NO it increased up to 16 times. A comparison of Au/CoPc samples with the same concentration of gold (~1 µg/cm 2 ) deposited by different methods revealed the influence of AuNPs morphology on the sensor responses of these heterostructures to NO. In particular, the Au/CoPc structure with the smallest AuNPs size of 8 nm, obtained by the PVD method, was characterized by the most rapid increase in the sensor response with an increase in NO concentration. The detection limits of the Au/CoPc heterostructure with the highest gold content (~2.1 µg/cm 2 ) to NH 3 and NO were 0.1 ppm and 4 ppb, respectively. The recovery times determined at 1 ppm NH 3 and 10 ppb NO were 85 and 100 s, respectively. It was shown that Au/CoPc heterostructures can be used for the detection of NH 3 in a gas mixture simulating exhaled air (N 2 -74%, O 2 -16%, H 2 O-6%, and CO 2 -4%). This makes the studied heterostructures promising materials for further investigation as active layers of sensors or sensor arrays for determining low concentrations of ammonia in exhaled air, which is a biomarker of some kidney diseases. Further research in this area should focus on improving the selectivity of layers and the creation of miniature sensor devices based on the studied heterostructures for the determination of NH 3 and NO in real exhaled air samples.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/bios12070476/s1, Figure S1. EDX spectra of Au-CVD2/CoPc (a) and Au-CVD3/CoPc (b); Figure S2. EDX spectra of Au-PVD1/CoPc (a), Au-PVD2/CoPc (b), and Au-PVD3/CoPc (c). Figure S3. EDX spectra of Au-DC2/CoPc (a) and Au-DC3/CoPc (b). Figure S4. Dependence of the sensor response of heterostructures, in which Au nanoparticles were deposited by the PVD technique, on the number of PVD cycles. Figure S5. Dependence of the sensor response of heterostructures, in which Au nanoparticles were deposited by the CVD technique, on the number of CVD cycles. Figure S6. Dependence of the response of Au_PVD3/CoPc sensor on temperature.