Fluorinated Metal Phthalocyanines: Interplay between Fluorination Degree, Films Orientation, and Ammonia Sensing Properties

In this work, the sensor response of MPcFx (M = Cu, Co, Zn; x = 0, 4, 16) films toward gaseous NH3 (10–50 ppm) was studied by a chemiresistive method and compared to that of unsubstituted MPc films to reveal the effects of central metals and F-substituents on the sensing properties. A combination of atomic force microscopy and X-ray diffraction techniques have been used to elucidate the structural features of thin MPcFx films deposited by organic molecular beam deposition. It has been shown that the sensor response of MPcF4 films to ammonia is noticeably higher than that of MPc films, which is in good correlation with the values of binding energy between the metal phthalocyanine and NH3 molecules, as calculated by the density functional theory (DFT) method. At the same time, in contrast to the DFT calculations, MPcF16 demonstrated the lesser sensor response compared with MPcF4, which appeared to be connected with the different structure and morphology of their films. The ZnPcF4 films were shown to exhibit a sensitivity to ammonia up to concentrations as low as 0.1 ppm, and can be used for the selective detection of ammonia in the presence of some reducing gases and volatile organic compounds. Moreover, the ZnPcF4 films can be used for the detection of NH3 in the gas mixture simulating exhaled air (N2 76%, O2 16%, H2O 5%, and CO2 3%).


Introduction
Ammonia is an important commercial chemical used to make fertilizers, household cleaners, and refrigerants, and is used to synthesize other chemicals. Despite its natural origin and wide distribution, ammonia is both a highly toxic and corrosive gas in its concentrated form. It is classified as an extremely hazardous substance, and is subjected to strict monitoring of its concentration in the environment, as well as in the automotive and chemical industry [1]. The detected concentration levels of ammonia depend on the application areas and can be varied in a very wide range, from ppb to hundreds ppm [1].
Recently, interest has been escalating into the study of exhaled breath as a noninvasive method of diagnostics for bronchopulmonary, cardiovascular, gastrointestinal, and other diseases [2]. Inference can be made regarding the changes in the metabolism and about the presence of a particular disease according to changes in the ratios of substances released in human breath. For example, an ammonia concentration of >1 ppm indicates renal failure in nephritis, atherosclerosis of the renal arteries, toxic affections of kidneys, and other diseases [3].
There are several ammonia detection devices described in the literature. Among those, optical gas analyzers [4][5][6], catalytic ammonia sensors [1], metal-oxide gas sensors [7,8], conducting polymer gas detectors [9][10][11], and chemiresistive sensors based on carbon nanomaterials and two-dimensional (2D) transition metal dichalcogenides [12] are used for the detection of gaseous ammonia, with their virtues and shortcomings. Electrolytic devices usually suffer from their low detection limits and limited accuracy, while optical sensors have very good sensitivity, but they are usually suited only for laboratory testing rather than for low cost portable sensors. Conducting polymer-based sensors generally suffer from irreversible sensor response and low selectivity in the presence of other gases [13].
Thin films of metal phthalocyanine (MPc) derivatives and their hybrid materials are of considerable interest as active layers of chemiresistive sensors for ammonia detection [14,15]. The introduction of various substituents into the phthalocyanine macrocycle can significantly alter the films' structure and morphology, and in turn, leads to a change of their electrical and sensing properties [15,16]. Fluorine substituents decrease the electron density of the aromatic ring and increase the oxidation potential of the MPc molecule [17]. As a result, fluorosubstituted phthalocyanines exhibit a higher sensor response to reducing gases, such as ammonia and hydrogen [18]. The better sensor response of the ZnPcF 16 and PdPcF 16 films towards gaseous ammonia compared with their unsubstituted analogues was presented by Schollhorn et al. [19,20] and Klyamer et al. [15], respectively. To the best of our knowledge, only sporadic data on the structural features and sensing behavior of tetrafluorosubstituted metal phthalocyanine (MPcF 4 ) films are available in the literature [15,16,21]. In our previous work [15], we studied the structure of CoPcF 4 films deposited by thermal evaporation and their sensor response to ammonia. It has been shown that, similarly to the case of MPcF 16 , the sensor response to ammonia is noticeably higher compared with unsubstituted CoPc films. To the best of our knowledge, the systematic analysis of the interplay between the fluorination degree, films' orientation, and sensing properties have never been carried out in the literature.
In this work, the sensor response of MPcF x (M = Cu, Co, Zn; x = 4, 16) films toward gaseous NH 3 (10-50 ppm) was studied by the chemiresistive method and compared to that of unsubstituted MPc films to reveal the effects of central metals and F-substituents on the sensing properties. A combination of atomic force microscopy and X-ray diffraction techniques have been used to elucidate the structural features and molecular orientation of thin films of MPcF x deposited by organic molecular beam deposition. Density functional theory (DFT) calculations have been performed to estimate the probable structure of MPcF x -analyte complexes and their bond formation energies. The sensor characteristics of ZnPcF 4 films were studied in more detail to demonstrate their application for the selective detection of a low concentration of ammonia (up to 0.1 ppm) in the presence of CO 2 and water vapors, as well as in the gas mixture with the composition close to exhaled air (N 2 76%, O 2 16%, H 2 O 5%, and CO 2 3%).

Preparation and Study of Thin Films
Unsubstituted (MPc, M = Cu, Co, Zn), tetrafluorosubstituted (MPcF 4 , M = Cu, Co, Zn), and hexadecafluorosubstituted (MPcF 16 , M = Cu, Co, Zn) phthalocyanines were synthesized, according to the procedures described elsewhere [15,16], from the corresponding phthalonitrile derivatives and corresponding metal chlorides. MPcF 4 derivatives were prepared as a statistical mixture of four regioisomers because of the various possible positions of the fluorine substituents. The isomers were not separated because of the close parameters of sublimation.
The thin films of all of the investigated phthalocyanines were deposited by an organic molecular beam deposition under a vacuum of 10 −5 Torr, onto platinum interdigitated electrodes (Dropsens, Oviedo, Spain). The electrodes have the following dimensions: the gap between digits is 10 µm; number of digits is 125 × 2 with a digit length of 6760 µm; and cell constant is 0.0118 cm −1 . The nominal thickness of the phthalocyanine films was about 100 nm. XRD studies of the thin film samples were carried out using a Shimadzu XRD-7000 diffractometer (CuKα, λ = 1.54187Å, Bragg-Brentano scheme, θ-θ goniometer, copper anode sealed tube 30 mA@40 kV with a Ni filter and scintillation counter detector). The scan range was from 5 • to 30 • 2θ, with the step of 0.03 • , and the acquisition time of 40 s per step. The atomic force microscopy (AFM) in the tapping mode with a Nanoscope IIIa (Veeco Instruments, Plainview, NY, USA) scanning probe microscope was used for the characterization of the films' morphology.
To test the chemiresistive sensor response the films were put into the flow cell and held for 10 min under air flow until the resistance reached a steady state value. Then NH 3 gas (0.1-50 ppm) was diluted with air and injected. Air was used as the dilution and carrier gas. The electrical resistance of phthalocyanine films was measured with a Keithley 236 electrometer by applying a constant dc voltage (8 V). All gas sensing measurements were carried out at room temperature.
The binding energy (E b ) was calculated according to the Equation (1), as a difference of the total energies of the corresponding adduct and its interacting components, as follows: The effective charge q(NH 3 ) was calculated according to the Equation (2), as follows: where Z n is the nuclear charge of the atom n; and P IJ and S IJ are the elements of the density and overlap the matrixes corresponding to the atomic orbitals I and J. This scheme, realized in ORCA, is based on the widely used Mulliken population analysis [29,30]. A bond order was estimated using Mayer's method [31,32].

Experimental Study of the Dependence of Sensing Response on Phthalocyanine Molecular Structure
The sensor response of MPcF x (M = Cu, Co, Zn; x = 4, 16) films was studied by a chemiresistive method. The choice of phthalocyanines of copper, cobalt, and zinc was determined by their better sensitivity to ammonia, according to the experimental data and DFT calculations performed earlier by Liang et al. [33]. The change of the film resistance during the sequential injection of the gas analyte and air purging was measured. The typical sensor response toward ammonia is shown in Figure 1, using CoPcF 4 and CoPcF 16 (b) films as an example.
The introduction of ammonia to the gas cell leads to the increase of resistance of the CoPc and CoPcF 4 films. Similar behavior typical of organic semiconductor films possessing p-type conductivity [34] was also observed in the case of ZnPc, CuPc, ZnPcF 4 , and CuPcF 4 films.
The resistance-based sensing mechanism of the semiconducting sensors has been studied in the literature [35,36]. It has been reported that the formation of the charge-transfer complexes by the coordination of O 2 to MPc, at the air/phthalocyanine interface and at the grain boundaries, leads to the formation of oxidized MPc + and O 2− species, and the injection of hole charge carriers into the films' bulk [37,38]. When a p-type semiconductor gas sensor is exposed to the reducing NH 3 gas, the electrons injected into the material through the oxidation reaction between the reducing gas and the O 2− species on the semiconductor surface decrease the concentration of the holes in the layer, which in turn increases the resistance of the MPc films [39]. The introduction of ammonia to the gas cell leads to the increase of resistance of the CoPc and CoPcF4 films. Similar behavior typical of organic semiconductor films possessing p-type conductivity [34] was also observed in the case of ZnPc, CuPc, ZnPcF4, and CuPcF4 films.
The resistance-based sensing mechanism of the semiconducting sensors has been studied in the literature [35,36]. It has been reported that the formation of the charge-transfer complexes by the coordination of O2 to MPc, at the air/phthalocyanine interface and at the grain boundaries, leads to the formation of oxidized MPc + and O 2− species, and the injection of hole charge carriers into the films' bulk [37,38]. When a p-type semiconductor gas sensor is exposed to the reducing NH3 gas, the electrons injected into the material through the oxidation reaction between the reducing gas and the O 2− species on the semiconductor surface decrease the concentration of the holes in the layer, which in turn increases the resistance of the MPc films [39].
On the contrary, the MPcF16 (M = Co, Cu, Zn) films exhibit a decrease of their resistance upon interaction with the electron donor NH3 molecules. It is known that perfluorinated metal phthalocyanines demonstrate the n-conducting behavior because of the effect of the electronwithdrawing fluorine substituents [40,41]. When an n-type semiconductor gas sensor is exposed to the reducing NH3 gas, ionized oxygen anions are used to oxidize the reducing gas, and the released electrons inject into the semiconducting core, which decreases the sensor resistance proportionally to the concentration of the reducing gas-analyte [36].
To study the influence of the phthalocyanine molecular structure on the sensing behavior, the sensor responses of the MPcFx (M = Co, Cu, Zn; n = 0, 4, 16) films toward ammonia were compared in the concentration range from 10 to 50 ppm. Figure 2 shows the dependence of the relative sensor response Rn = |R-Ro|/Ro (where R is the resistance at a certain concentration of the analyte, Ro is the resistance before injection of the analyte vapors) for the MPc, MPcF4, and MPcF16 films. It can be seen that the sensor response decreases in the order of CoPcFx > ZnPcFx > CuPcFx, both in the case of the unsubstituted ( Figure 2a) and fluorinated derivatives (Figure 2b,c). For instance, the sensor response of the CoPc films toward 10 ppm of ammonia is about two times higher compared with the ZnPc films, and eight times higher compared with the CuPc films ( Figure 2a). An even more pronounced difference is observed in the case of the MPcF4 and MPcF16 films, for example, the sensor response of the CoPcFx (x = 4, 8) films toward 10 ppm of ammonia is about four times higher compared with the ZnPcFx films, and 13 times higher compared with the CuPcFx films (Figure 2b,c). On the contrary, the MPcF 16 (M = Co, Cu, Zn) films exhibit a decrease of their resistance upon interaction with the electron donor NH 3 molecules. It is known that perfluorinated metal phthalocyanines demonstrate the n-conducting behavior because of the effect of the electron-withdrawing fluorine substituents [40,41]. When an n-type semiconductor gas sensor is exposed to the reducing NH 3 gas, ionized oxygen anions are used to oxidize the reducing gas, and the released electrons inject into the semiconducting core, which decreases the sensor resistance proportionally to the concentration of the reducing gas-analyte [36].
To study the influence of the phthalocyanine molecular structure on the sensing behavior, the sensor responses of the MPcF x (M = Co, Cu, Zn; x = 0, 4, 16) films toward ammonia were compared in the concentration range from 10 to 50 ppm. Figure 2 shows the dependence of the relative sensor response R n = |R-R o |/R o (where R is the resistance at a certain concentration of the analyte, R o is the resistance before injection of the analyte vapors) for the MPc, MPcF 4 , and MPcF 16 films. It can be seen that the sensor response decreases in the order of CoPcF x > ZnPcF x > CuPcF x , both in the case of the unsubstituted ( Figure 2a Figure 3 demonstrates the effect of the F-substitution in the phthalocyanine ring on the sensing response to ammonia, using the ZnPcFx (x = 0, 4, 16) films as an example. The sensor response decreases in the order of ZnPcF4 > ZnPcF16 > ZnPc. The same order is also observed for the CuPcFx and CoPcFx films. The MPcF4 films exhibit the maximal sensor response to ammonia among all of the investigated phthalocyanines, for example, the sensor response of the MPcF4 (M = Zn, Co, Cu) films is 3-10 times higher than that of the MPcF16 films, and 30-70 times higher than that of the MPc films.  Figure 3 demonstrates the effect of the F-substitution in the phthalocyanine ring on the sensing response to ammonia, using the ZnPcF x (x = 0, 4, 16) films as an example. The sensor response decreases in the order of ZnPcF 4 > ZnPcF 16 > ZnPc. The same order is also observed for the CuPcF x and CoPcF x films. The MPcF 4 films exhibit the maximal sensor response to ammonia among all of the investigated phthalocyanines, for example, the sensor response of the MPcF 4 (M = Zn, Co, Cu) films is 3-10 times higher than that of the MPcF 16 films, and 30-70 times higher than that of the MPc films. Therefore, the introduction of the F-substituents to the phthalocyanine macrocycle leads to a substantial increase of their sensitivity to ammonia.  Figure 3 demonstrates the effect of the F-substitution in the phthalocyanine ring on the sensing response to ammonia, using the ZnPcFx (x = 0, 4, 16) films as an example. The sensor response decreases in the order of ZnPcF4 > ZnPcF16 > ZnPc. The same order is also observed for the CuPcFx and CoPcFx films. The MPcF4 films exhibit the maximal sensor response to ammonia among all of the investigated phthalocyanines, for example, the sensor response of the MPcF4 (M = Zn, Co, Cu) films is 3-10 times higher than that of the MPcF16 films, and 30-70 times higher than that of the MPc films. Therefore, the introduction of the F-substituents to the phthalocyanine macrocycle leads to a substantial increase of their sensitivity to ammonia.      The sensor response of the sensing layers depends on several factors, among them are the molecular structure of sensing material that governs the nature and strength of its interaction with an analyte, and the sensing layer structure and morphology that determines the number of active sites, and the rate of adsorption-desorption process.

Theoretical Study of the Dependence of Sensor Response on the Phthalocyanine Molecular Structure
The DFT calculations have been performed to study the interaction of the NH 3 molecules with MPcF x , and to elucidate the different sensor responses of MPcF x with different x and central metals.
To check the validity of the theoretical model, the calculated vibrational spectra of MPcF x were compared with the experimental ones, as it has already been described elsewhere [15].
The most favorable structure of the MPc····NH 3 aggregates simulated by the DFT calculations was that with the NH 3 molecule binding with phthalocyanine, via its central metal. The binding of MPcF x with the NH 3 molecule increases the out-of-plane distortion of the Pc ring (e.g., the out-of-plane displacement of the Zn atom in ZnPcF 4 leads to an increase in the Zn-N α bond length from 2.007 Å to 2.043 Å, on average). The binding parameters for NH 3 with MPcF x are presented in Table 2 for comparison. It has already been shown elsewhere [15,42] that the ammonia and MPcs form complexes with a charge transfer from the NH 3 to phthalocyanine molecule, via the interaction of NH 3 with the central metal ion inside the phthalocyanine macrocycle. The formation of this bond is based on the displacement of electron density from NH 3 molecule to MPc, through the central metal atom and, as a result, NH 3 acquires a positive effective charge increasing in the order of CuPc < ZnPc < CoPc, both for the unsubstituted and fluorinated derivatives ( Table 2). At the same time, the M-NH 3 bond order increases, and the respective distance d between the metal atom and the ammonia nitrogen atom decreases in the same order. The obtained theoretical data are in a good correlation with the experimental investigations of the sensor response of MPc (M = Cu, Zn, Co), which is higher in the case of cobalt phthalocyanines.
As for the effect of the F-substituents, the binding energy between NH 3 and MPcF x and the positive effective charge of NH 3 increases in the order of MPc····NH 3 < MPcF 4 ····NH 3 < MPcF 16 ····NH 3 ( Table 2). The experimental investigations of the sensor response of the unsubstituted and fluorinated phthalocyanines showed that its value is higher in the case of MPcF 4 . However, it is necessary to mention that, in contrast to the theoretical calculations, the experimental sensor response of the MPcF 4 films is higher than that of MPcF 16 films. It is conceivable that such behavior can be associated with  16 films. It has already been mentioned above that the MPcF 4 films possess the p-type conductivity, whereas the MPcF 16 films demonstrate the n-conducting behavior. One more important factor governing the sensing properties is the structure and morphology of the sensing layers.

Characterization of Thin Films
To study the effect of fluorination, the structure and morphology of the MPcF x films were investigated by XRD and AFM methods. X-ray diffraction patterns of thin films of all nine phthalocyanine derivatives are shown in Figure 5. The diffraction patterns contain a single strong diffraction peak in the range from 5 • to 7 • 2θ and several barely visible peaks with the corresponding interplanar distances d, which are the natural fractions of the d 0 of the strong peak. This type of diffraction patterns is a typical feature of thin films with a strong preferred orientation. Comparing the interplanar distances with the calculated ones known from the single crystal data [43,44], the CoPc and CuPc thin films were identified as metastable α-polymorphs. There are no known structural data for α-ZnPc, however, some works show that α-ZnPc is isostructural to α-CuPc and α-CoPc, and it forms when deposited onto the substrate surface at temperatures lower than 100 • C [45].  CuPcF4, CoPcF4, and ZnPcF4 are isostructural with PdPcF4 [16], and crystallize only in one triclinic (P-1 space group) phase. There are two known polymorphs for CuPcF16, that is, α-CuPcF16 (P-1 space group, Z = 1) [46] and triclinic β-CuPcF16 (P-1 space group, Z = 2) [47], and both of them have very similar values of interplanar distances for the first peak on the calculated diffraction pattern. α-CuPcF16 grows on the substrate surface at room temperature, while the β-CuPcF16 are obtained at 360 °C. As, in this work, the substrate temperature was about 20 °C, it is reasonable to assume that the CuPcF16 thin films consists of a α-phase. No crystal structure data are known for the α-polymorphs of CoPcF16 and ZnPcF16, but as α-CuPc/α-CoPc, CuPcF4/CoPcF4/ZnPcF4, and β-CuPcF16/β-CoPcF16/β-ZnPcF16 are isostructural to each other, we assumed that the CoPcF16 and ZnPcF16 thin films are also α-polymorphs, with the same structure as α-CuPcF16. Figure 6 shows the AFM images of the surface of the ZnPc (a), ZnPcF4 (b), and ZnPcF16 (c) films. As can be clearly seen, the ZnPc films surface consists of roundish grains (Figure 6a) and has the rms roughness value of 14.2 nm. The ZnPcF4 film, having an rms roughness of 6.7 nm, is formed by azimuthally disordered elongated grains (Figure 6b). The ZnPcF16 films exhibit a high density of azimuthally disordered roundish grains, with the size noticeably smaller than those of the ZnPc films and the minimal rms roughness values (4.2 nm) among the investigated films (Figure 6c). The more rough and inhomogeneous surface of the ZnPcF4 films can also be responsible for their higher sensor response to ammonia, compared with ZnPcF16 films. CuPcF 4 , CoPcF 4 , and ZnPcF 4 are isostructural with PdPcF 4 [16], and crystallize only in one triclinic (P-1 space group) phase. There are two known polymorphs for CuPcF 16 , that is, α-CuPcF 16 (P-1 space group, Z = 1) [46] and triclinic β-CuPcF 16 (P-1 space group, Z = 2) [47], and both of them have very similar values of interplanar distances for the first peak on the calculated diffraction pattern. α-CuPcF 16 grows on the substrate surface at room temperature, while the β-CuPcF 16 are obtained at 360 • C. As, in this work, the substrate temperature was about 20 • C, it is reasonable to assume that the CuPcF 16 thin films consists of a α-phase. No crystal structure data are known for the α-polymorphs of CoPcF 16 and ZnPcF 16 , but as α-CuPc/α-CoPc, CuPcF 4 /CoPcF 4 /ZnPcF 4 , and β-CuPcF 16 /β-CoPcF 16 /β-ZnPcF 16 are isostructural to each other, we assumed that the CoPcF 16 and ZnPcF 16 thin films are also α-polymorphs, with the same structure as α-CuPcF 16 . Figure 6 shows the AFM images of the surface of the ZnPc (a), ZnPcF 4 (b), and ZnPcF 16 (c) films. As can be clearly seen, the ZnPc films surface consists of roundish grains (Figure 6a) and has the rms roughness value of 14.2 nm. The ZnPcF 4 film, having an rms roughness of 6.7 nm, is formed by azimuthally disordered elongated grains (Figure 6b). The ZnPcF 16 films exhibit a high density of azimuthally disordered roundish grains, with the size noticeably smaller than those of the ZnPc films and the minimal rms roughness values (4.2 nm) among the investigated films (Figure 6c). The more rough and inhomogeneous surface of the ZnPcF 4 films can also be responsible for their higher sensor response to ammonia, compared with ZnPcF 16 films.
α-polymorphs of CoPcF16 and ZnPcF16, but as α-CuPc/α-CoPc, CuPcF4/CoPcF4/ZnPcF4, and β-CuPcF16/β-CoPcF16/β-ZnPcF16 are isostructural to each other, we assumed that the CoPcF16 and ZnPcF16 thin films are also α-polymorphs, with the same structure as α-CuPcF16. Figure 6 shows the AFM images of the surface of the ZnPc (a), ZnPcF4 (b), and ZnPcF16 (c) films. As can be clearly seen, the ZnPc films surface consists of roundish grains (Figure 6a) and has the rms roughness value of 14.2 nm. The ZnPcF4 film, having an rms roughness of 6.7 nm, is formed by azimuthally disordered elongated grains (Figure 6b). The ZnPcF16 films exhibit a high density of azimuthally disordered roundish grains, with the size noticeably smaller than those of the ZnPc films and the minimal rms roughness values (4.2 nm) among the investigated films (Figure 6c). The more rough and inhomogeneous surface of the ZnPcF4 films can also be responsible for their higher sensor response to ammonia, compared with ZnPcF16 films.

Sensor Characteristics of Phthalocyanine Films
The sensor characteristics of the MPcF 4 films demonstrating the best sensitivity to ammonia among the investigated samples were studied in more detail to demonstrate their applicability for the detection of NH 3 at lower concentrations, down to 0.1 ppm, in the presence of other gases. A typical sensor response of a ZnPcF 4 layer toward ammonia, in the concentration range from 1-4 ppm, is shown in Figure 7a. To demonstrate the possible application of ZnPcF 4 films for the detection of gases-biomarkers in exhaled air, the sensor response of ZnPcF 4 films to ammonia was also tested in a mixture of gases, with the composition close to the exhaled air of healthy people. For this purpose, small amounts of ammonia (1-4 v.%) were added to the preliminarily prepared gas mixture (N 2 -76%, O 2 -16%, H 2 O-5%, CO 2 -3%). The sensor response of ZnPcF 4 films to ammonia (1-4 ppm) diluted with the mixture of gases N 2 76%, O 2 16%, H 2 O 5%, and CO 2 3% is shown in Figure 7b.

Sensor Characteristics of Phthalocyanine Films
The sensor characteristics of the MPcF4 films demonstrating the best sensitivity to ammonia among the investigated samples were studied in more detail to demonstrate their applicability for the detection of NH3 at lower concentrations, down to 0.1 ppm, in the presence of other gases. A typical sensor response of a ZnPcF4 layer toward ammonia, in the concentration range from 1-4 ppm, is shown in Figure 7a. To demonstrate the possible application of ZnPcF4 films for the detection of gases-biomarkers in exhaled air, the sensor response of ZnPcF4 films to ammonia was also tested in a mixture of gases, with the composition close to the exhaled air of healthy people. For this purpose, small amounts of ammonia (1-4 v.%) were added to the preliminarily prepared gas mixture (N2-76%, O2-16%, H2O-5%, CO2-3%). The sensor response of ZnPcF4 films to ammonia (1-4 ppm) diluted with the mixture of gases N2 76%, O2 16%, H2O 5%, and CO2 3% is shown in Figure 7b. The ZnPcF4 films demonstrate a reversible sensor response in the investigated concentration range, with a quite good response and recovery time; the response time varied from 15 s to 30 s, depending on the NH3 concentration, while the recovery time increased from 28 s to 90 s when the NH3 concentration changes from 1 to 4 ppm. The dependence of the sensor response on the NH3 concentration is given in Figure 8. The minimum detected concentration of NH3 in the case of ZnPcF4 films was found to be 0.1 ppm. depending on the NH 3 concentration, while the recovery time increased from 28 s to 90 s when the NH 3 concentration changes from 1 to 4 ppm. The dependence of the sensor response on the NH 3 concentration is given in Figure 8. The minimum detected concentration of NH 3 in the case of ZnPcF 4 films was found to be 0.1 ppm. Figure 7. Sensor response of a ZnPcF4 layer toward ammonia in the concentration range from 1 to 4 ppm, in air (a) and in a mixture of gases with the composition close to exhaled air of healthy people (N2-76%, O2-16%, H2O-5%, and CO2-3%) (b).
The ZnPcF4 films demonstrate a reversible sensor response in the investigated concentration range, with a quite good response and recovery time; the response time varied from 15 s to 30 s, depending on the NH3 concentration, while the recovery time increased from 28 s to 90 s when the NH3 concentration changes from 1 to 4 ppm. The dependence of the sensor response on the NH3 concentration is given in Figure 8. The minimum detected concentration of NH3 in the case of ZnPcF4 films was found to be 0.1 ppm.  To study the selectivity of ZnPcF 4 -based sensors, their response was tested against ammonia (10 ppm), acetone (1000 ppm), dichloromethane (10 4 ppm), carbon dioxide (10 4 ppm), and ethanol (10 4 ppm). Figure 9a shows that the sensor exhibited a significantly higher response to ammonia in comparison with that toward the other investigated analytes. This obviously indicates the viability of this type of sensors to detect ammonia selectively in the presence of other gases, such as those tested in this work. Note that the investigated interfering gases were taken at a much higher concentration compared with the ammonia. To study the selectivity of ZnPcF4-based sensors, their response was tested against ammonia (10 ppm), acetone (1000 ppm), dichloromethane (10 4 ppm), carbon dioxide (10 4 ppm), and ethanol (10 4 ppm). Figure 9a shows that the sensor exhibited a significantly higher response to ammonia in comparison with that toward the other investigated analytes. This obviously indicates the viability of this type of sensors to detect ammonia selectively in the presence of other gases, such as those tested in this work. Note that the investigated interfering gases were taken at a much higher concentration compared with the ammonia. The dependence of the sensor response on the relative humidity (RH) was also examined and the results are presented in Figure 9b, which show that the initial resistance of the ZnPcF4 films decreases with the increase of RH from 5% to 70%. The value of the sensor response to NH3 at RH 5% and 30% is almost the same, however, it is found to decrease noticeably when increasing the RH to 70%. The main reason for such behavior appears to be a competitive sorption of the NH3 and H2O molecules on the surface of the ZnPcF4 film.
The sensor response of a ZnPcF4 layer toward ammonia in the air was also compared with that in a mixture of gases, with the composition close to the exhaled air of healthy people. Figures 7b shows that the value of the sensor response to NH3 in the presence of gas mixture (N2 76%, O2 16%, H2O 5%, and CO2 3%) is almost the same as in the mixture with air. This makes the ZnPcF4 films a promising sensing layer for the detection of ammonia in exhaled air, which is used as a gas-biomarker The dependence of the sensor response on the relative humidity (RH) was also examined and the results are presented in Figure 9b, which show that the initial resistance of the ZnPcF 4 films decreases with the increase of RH from 5% to 70%. The value of the sensor response to NH 3 at RH 5% and 30% is almost the same, however, it is found to decrease noticeably when increasing the RH to 70%. The main reason for such behavior appears to be a competitive sorption of the NH 3 and H 2 O molecules on the surface of the ZnPcF 4 film.
The sensor response of a ZnPcF 4 layer toward ammonia in the air was also compared with that in a mixture of gases, with the composition close to the exhaled air of healthy people. Figure 7b shows that the value of the sensor response to NH 3 in the presence of gas mixture (N 2 76%, O 2 16%, H 2 O 5%, and CO 2 3%) is almost the same as in the mixture with air. This makes the ZnPcF 4 films a promising sensing layer for the detection of ammonia in exhaled air, which is used as a gas-biomarker of renal failure in nephritis, atherosclerosis of the renal arteries, and toxic affections of the kidneys [3].
Note that the sensor performance of several sensors towards ammonia has been reported in the literature [48][49][50][51][52][53]. Some examples of sensor characteristics of several sensors, including the data obtained in this work, are summarized in Table 3 for comparison.

Carbon-Containing Nanomaterials and Phthalocyanines
AuNPs/SWNT 0.25-6 0.255 20 (0.4 ppm)/-Room temperature [51] rGO modified with metal tetra-α-iso-pentyloxymetallophthalocyanines (CuPc, NiPc, PbPc) The sensing layers based on ZnPcF 4 are quite competitive with the active layers, based on metal oxides, conducting polymers, and carbon-containing nanomaterials, described in the literature; the ZnPcF 4 films exhibit a reversible sensor response at room temperature, a low detection limit, and low values of response and recovery times, compared with the other sensors.

Conclusions
In this work, unsubstituted metal phthalocyanines (MPc, M = Cu, Co, Zn), tetrafluorosubstituted metal phthalocyanines (MPcF 4 ) and hexadecafluorosubstituted metal phthalocyanines (MPcF 16 ) thin films were deposited by organic molecular beam deposition and studied, to reveal the effects of the central metals and F-substituents on the films' sensor response to ammonia.
It has been shown that the sensor response decreased in the order of CoPcF x > ZnPcF x > CuPcF x , both in the case of the unsubstituted and fluorinated derivatives. The sensor response of the MPcF 4 films to ammonia is noticeably higher than that of the MPc films, which is in good correlation with the values of the binding energy between the metal phthalocyanine and NH 3 molecule, as calculated by the DFT method. At the same time, in contrast to the DFT calculations, MPcF 16 demonstrated the lesser sensor response compared with MPcF 4 , which appeared to be connected with the different structure and morphology of their films.
It has been shown, using ZnPcF 4 films as an example, that they exhibit a sensitivity to ammonia, up to concentrations as low as 0.1 ppm, and can be used for the selective detection of ammonia in the presence of some reducing gases and volatile organic compounds. Moreover, the ZnPcF 4 films can be used for the detection of NH 3 in the gas mixture simulating exhaled air (N 2 76%, O 2 16%, H 2 O 5%, and CO 2 3%). This makes these films promising active layers as chemiresistive sensors for the detection of ammonia in exhaled air, which is a biomarker of some kidney diseases.