A Hybrid Nanomaterial Based on Single Walled Carbon Nanotubes Cross-Linked via Axially Substituted Silicon (IV) Phthalocyanine for Chemiresistive Sensors

In this work, the novel hybrid nanomaterial SWCNT/SiPc made of single walled carbon nanotubes (SWCNT) cross-linked via axially substituted silicon (IV) phthalocyanine (SiPc) was studied as the active layer of chemiresistive layers for the detection of ammonia and hydrogen. SWCNT/SiPc is the first example of a carbon-based nanomaterial in which an axially substituted phthalocyanine derivative is used as a linker. The prepared hybrid material was characterized by spectroscopic methods, thermogravimetry, scanning and transmission electron microscopies. The layers of the prepared hybrid were tested as sensors toward ammonia and hydrogen by a chemiresistive method at different temperatures and relative humidity as well as in the presence of interfering gases like carbon dioxide, hydrogen sulfide and volatile organic vapors. The hybrid layers exhibited the completely reversible sensor response to both gases at room temperature; the recovery time was 100–200 s for NH3 and 50–120 s in the case of H2 depending on the gas concentrations. At the relative humidity (RH) of 20%, the sensor response was almost the same as that measured at RH 5%, whereas the further increase of RH led to its 2–3 fold decrease. It was demonstrated that the SWCNT/SiPc layers can be successfully used for the detection of both NH3 and H2 in the presence of CO2. On the contrary, H2S was found to be an interfering gas for the NH3 detection.


Synthesis and Characterization
For the creation of SWCNT/SiPc hybrid, the novel axially bis(propynoxy)substituted silicon (IV) phthalocyanine (SiPc) was synthesized by the reaction of dichlorosilicon (IV) phthalocyanine with propargyl alcohol in pyridine. The SWCNT/SiPc hybrid material is the first example of the three-dimensional carbon nanotubes cross-linked via axially substituted phthalocyanines. After purification, the structure of this novel phthalocyanine was confirmed by spectroscopic methods such as UV-Vis, 1 H-NMR, FT-IR, as well as by mass spectrometry and elemental analysis. In the FT-IR spectrum of the SiPc, the vibration band belonging to Si-O-C was observed at 1081 cm −1 . The characteristic vibration bands of C≡C-H and C≡C were observed at 3261 and 2110 cm −1 , respectively [54]. The other main vibrations of aromatic-CH and aliphatic-CH groups were monitored at 3010 and 2957-2852 cm −1 , respectively. In the 1 H-NMR spectrum of SiPc, the aromatic Hα and Hβ protons of phthalocyanine ring were observed between 9.68-9.66 and 8.37-8.36 ppm, respectively [55]. The aliphatic protons in Si-O-CH2 and C≡C-H groups were shifted to the negative region due to the magnetic anisotropy behavior of silicon (IV) phthalocyanine ring. These protons were observed at −0.99 ppm for CH2 and 0.63 ppm for C≡C-H as singlets. In the MALDI-TOFF mass spectrum of this phthalocyanine, the molecular ion peak was observed at 650.66 m/z as [M] + and 595.43 m/z as the [M-OCH2CCH] + fragment, and this result was compatible with the theoretical molecular weight of this compound [56].
FT-IR spectra of the silicon (IV) phthalocyanine substituted with terminal propynoxy groups in axial positions (SiPc), SWCNTs bearing azido groups (SWCNT-N3) and their hybrid nanomaterial

Synthesis and Characterization
For the creation of SWCNT/SiPc hybrid, the novel axially bis(propynoxy)substituted silicon (IV) phthalocyanine (SiPc) was synthesized by the reaction of dichlorosilicon (IV) phthalocyanine with propargyl alcohol in pyridine. The SWCNT/SiPc hybrid material is the first example of the three-dimensional carbon nanotubes cross-linked via axially substituted phthalocyanines. After purification, the structure of this novel phthalocyanine was confirmed by spectroscopic methods such as UV-Vis, 1 H-NMR, FT-IR, as well as by mass spectrometry and elemental analysis. In the FT-IR spectrum of the SiPc, the vibration band belonging to Si-O-C was observed at 1081 cm −1 . The characteristic vibration bands of C≡C-H and C≡C were observed at 3261 and 2110 cm −1 , respectively [54]. The other main vibrations of aromatic-CH and aliphatic-CH groups were monitored at 3010 and 2957-2852 cm −1 , respectively. In the 1 H-NMR spectrum of SiPc, the aromatic H α and H β protons of phthalocyanine ring were observed between 9.68-9.66 and 8.37-8.36 ppm, respectively [55]. The aliphatic protons in Si-O-CH 2 and C≡C-H groups were shifted to the negative region due to the magnetic anisotropy behavior of silicon (IV) phthalocyanine ring. These protons were observed at −0.99 ppm for CH 2 and 0.63 ppm for C≡C-H as singlets. In the MALDI-TOFF mass spectrum of this phthalocyanine, the molecular ion peak was observed at 650.66 m/z as [M] + and 595.43 m/z as the [M-OCH 2 CCH] + fragment, and this result was compatible with the theoretical molecular weight of this compound [56].
FT-IR spectra of the silicon (IV) phthalocyanine substituted with terminal propynoxy groups in axial positions (SiPc), SWCNTs bearing azido groups (SWCNT-N3) and their hybrid nanomaterial (SWCNT/SiPc) are shown in Figure 2. The vibrational peaks corresponding to ethynyl groups in the SiPc derivative (observed at 3261 and 2110 cm −1 ) and azide groups in SWCNT-N3 (observed at 2106 cm −1 ) almost disappear in the FT-IR spectrum of SWCNT/SiPc hybrid which indicates that the "Click" reaction was occurred between SiPc and SWCNT-N3 (Figure 2). At the same time, several peaks Molecules 2020, 25, 2073 4 of 17 belonging to phthalocyanine moiety are observed in the FT-IR spectrum of the hybrid, viz. 2990-2835 (Aliphatic CH), 1608, 1518 (C=C), 1135 (Ar C=C) and 1122 cm −1 (Si-O). A small peak corresponding to azide vibration was still observed in the FT-IR spectrum of SWCNT/SiPc due to the rest of the azide groups in the hybrid nanomaterial.
Molecules 2020, 25, x FOR PEER REVIEW 4 of 18 (SWCNT/SiPc) are shown in Figure 2. The vibrational peaks corresponding to ethynyl groups in the SiPc derivative (observed at 3261 and 2110 cm −1 ) and azide groups in SWCNT-N3 (observed at 2106 cm −1 ) almost disappear in the FT-IR spectrum of SWCNT/SiPc hybrid which indicates that the "Click" reaction was occurred between SiPc and SWCNT-N3 (Figure 2). At the same time, several peaks belonging to phthalocyanine moiety are observed in the FT-IR spectrum of the hybrid, viz. 2990-2835 (Aliphatic CH), 1608, 1518 (C=C), 1135 (Ar C=C) and 1122 cm −1 (Si-O). A small peak corresponding to azide vibration was still observed in the FT-IR spectrum of SWCNT/SiPc due to the rest of the azide groups in the hybrid nanomaterial. The Raman spectrum of SWCNT/SiPc is compared to those of SiPc and SWCNT-N3 in Figure 3. The D and G bands in the spectrum of SWCNT-N3 are located at 1343 and 1590 cm −1 , respectively, while their positions shifted to 1332 and 1582 cm −1 after the formation of the hybrid. The ratio of intensities of D and G bands (ID/IG), which is usually used for the characterization of functionalization of carbon nanomaterials [57], increases more than twice, namely, the ID/IG ratio in the spectrum of SWCNT-N3 is 0.05, while in the spectrum of SWCNT/SiPc hybrid is 0.12. This value is less than that observed in the case of SWCNT hybrids obtained by non-covalent functionalization by phthalocyanine molecules (e.g., the ID/IG ratio was 0.052 in the case of SWCNTs The Raman spectrum of SWCNT/SiPc is compared to those of SiPc and SWCNT-N3 in Figure 3. The D and G bands in the spectrum of SWCNT-N3 are located at 1343 and 1590 cm −1 , respectively, while their positions shifted to 1332 and 1582 cm −1 after the formation of the hybrid. The ratio of intensities of D and G bands (I D /I G ), which is usually used for the characterization of functionalization of carbon nanomaterials [57], increases more than twice, namely, the I D /I G ratio in the spectrum of SWCNT-N3 is 0.05, while in the spectrum of SWCNT/SiPc hybrid is 0.12.
Molecules 2020, 25, x FOR PEER REVIEW 4 of 18 (SWCNT/SiPc) are shown in Figure 2. The vibrational peaks corresponding to ethynyl groups in the SiPc derivative (observed at 3261 and 2110 cm −1 ) and azide groups in SWCNT-N3 (observed at 2106 cm −1 ) almost disappear in the FT-IR spectrum of SWCNT/SiPc hybrid which indicates that the "Click" reaction was occurred between SiPc and SWCNT-N3 (Figure 2). At the same time, several peaks belonging to phthalocyanine moiety are observed in the FT-IR spectrum of the hybrid, viz. 2990-2835 (Aliphatic CH), 1608, 1518 (C=C), 1135 (Ar C=C) and 1122 cm −1 (Si-O). A small peak corresponding to azide vibration was still observed in the FT-IR spectrum of SWCNT/SiPc due to the rest of the azide groups in the hybrid nanomaterial. The Raman spectrum of SWCNT/SiPc is compared to those of SiPc and SWCNT-N3 in Figure 3. The D and G bands in the spectrum of SWCNT-N3 are located at 1343 and 1590 cm −1 , respectively, while their positions shifted to 1332 and 1582 cm −1 after the formation of the hybrid. The ratio of intensities of D and G bands (ID/IG), which is usually used for the characterization of functionalization of carbon nanomaterials [57], increases more than twice, namely, the ID/IG ratio in the spectrum of SWCNT-N3 is 0.05, while in the spectrum of SWCNT/SiPc hybrid is 0.12. This value is less than that observed in the case of SWCNT hybrids obtained by non-covalent functionalization by phthalocyanine molecules (e.g., the ID/IG ratio was 0.052 in the case of SWCNTs This value is less than that observed in the case of SWCNT hybrids obtained by non-covalent functionalization by phthalocyanine molecules (e.g., the I D /I G ratio was 0.052 in the case of SWCNTs non-covalently modified with tetrasubstituted ZnPc [58]), but closer the I D /I G values obtained for SWCNTs covalently modified with other polyaromatic molecules (e.g., 0.092 in the case of 3D hybrids of SWCNT cross-linked via phenylcoumarin molecules [57]). The intensity of radial breathing modes (RBM) observing in the spectrum of SWCNT-N3 in the range from 150 to 230 cm −1 becomes lower in the SWCNT/SiPc spectrum. RBMs correspond to radial expansion-contraction of the nanotube and, their frequencies depend on the nanotube diameter. A decrease of their intensities in the spectrum of SWCNT/SiPc hybrid may relate to the change of electronic structure of SWCNT after the formation of covalent bonds with SiPc molecules and a cross-linked structure. Apart from this, bands corresponding to the vibration of phthalocyanine moiety are clearly observed in the spectrum of SWCNT/SiPc hybrids (see inset to Figure 3). Figure 4 depicts the UV-Visible absorption spectra of the SiPc solution in DMF and the suspension of its hybrid nanomaterial SWCNT/SiPc obtained by sonication in DMF. SiPc shows characteristic absorption bands at 677 and 362 nm in DMF known as Q and B bands, respectively. The vibronic satellite at 610 nm is also observed. The intensities of these bands are lower in the spectrum of SWCNT/SiPc hybrid because of the lower amount of SiPc molecules attached to the surface of carbon nanotubes compared to the SiPc solution.
Molecules 2020, 25, x FOR PEER REVIEW 5 of 18 non-covalently modified with tetrasubstituted ZnPc [58]), but closer the ID/IG values obtained for SWCNTs covalently modified with other polyaromatic molecules (e.g., 0.092 in the case of 3D hybrids of SWCNT cross-linked via phenylcoumarin molecules [57]). The intensity of radial breathing modes (RBM) observing in the spectrum of SWCNT-N3 in the range from 150 to 230 cm −1 becomes lower in the SWCNT/SiPc spectrum. RBMs correspond to radial expansion-contraction of the nanotube and, their frequencies depend on the nanotube diameter. A decrease of their intensities in the spectrum of SWCNT/SiPc hybrid may relate to the change of electronic structure of SWCNT after the formation of covalent bonds with SiPc molecules and a cross-linked structure. Apart from this, bands corresponding to the vibration of phthalocyanine moiety are clearly observed in the spectrum of SWCNT/SiPc hybrids (see inset to Figure 3). Figure 4 depicts the UV-Visible absorption spectra of the SiPc solution in DMF and the suspension of its hybrid nanomaterial SWCNT/SiPc obtained by sonication in DMF. SiPc shows characteristic absorption bands at 677 and 362 nm in DMF known as Q and B bands, respectively. The vibronic satellite at 610 nm is also observed. The intensities of these bands are lower in the spectrum of SWCNT/SiPc hybrid because of the lower amount of SiPc molecules attached to the surface of carbon nanotubes compared to the SiPc solution.
. In the electronic absorption spectrum of SWCNT/SiPc hybrid, the Q band was observed as a broad peak with the absorption maximum at 685 nm, bathochromically shifted compared to that in the electronic absorption spectrum of SiPc. The shift to the higher wavelength might be ascribed to the strong electronic interaction between SWCNTs and SiPc derivative in the SWCNTs/SiPc hybrid (MPc is usually regarded as an electron donor, while SWCNTs as an electron acceptor) [59] , which reduces the aggregation of the hybrids and hence changes the absorption.

Thermogravimetric Analysis
The amount of SiPc molecules covalently bonded to SWCNT was calculated using TGA data ( Figure 5). The weight losses of SiPc, pristine SWCNT, SWCNT-N3 and SWCNT/SiPc at 800 °C were found to be 51.70%, 4.60%, 12.17% and 38.66%, respectively. The weight losses due to functional chemical groups in SWCNT were determined as 7.57% for SWCNTs-N3 and 26.49% for SWCNT/SiPc. The number of azide groups on SWCNTs-N3 was calculated according to the literature [47,60,61] to be 1 per 43 carbon atoms. The weight loss due to elimination of the SiPc fragment from SWCNT/SiPc hybrids was 51.23% (26.49%/51.70%). The amount of SiPc covalently bonded to SWCNT was estimated in the same way to be one SiPc molecule per 57 carbon atoms in the hybrid according to the calculations (51.23% × 650.16)/(48.77% × 12). In the electronic absorption spectrum of SWCNT/SiPc hybrid, the Q band was observed as a broad peak with the absorption maximum at 685 nm, bathochromically shifted compared to that in the electronic absorption spectrum of SiPc. The shift to the higher wavelength might be ascribed to the strong electronic interaction between SWCNTs and SiPc derivative in the SWCNTs/SiPc hybrid (MPc is usually regarded as an electron donor, while SWCNTs as an electron acceptor) [59], which reduces the aggregation of the hybrids and hence changes the absorption.

Thermogravimetric Analysis
The amount of SiPc molecules covalently bonded to SWCNT was calculated using TGA data ( Figure 5). The weight losses of SiPc, pristine SWCNT, SWCNT-N3 and SWCNT/SiPc at 800 • C were found to be 51.70%, 4.60%, 12.17% and 38.66%, respectively. The weight losses due to functional chemical groups in SWCNT were determined as 7.57% for SWCNTs-N3 and 26.49% for SWCNT/SiPc. The number of azide groups on SWCNTs-N3 was calculated according to the literature [47,60,61] to be 1 per 43 carbon atoms. The weight loss due to elimination of the SiPc fragment from SWCNT/SiPc hybrids was 51.23% (26.49%/51.70%). The amount of SiPc covalently bonded to SWCNT was estimated in the same way to be one SiPc molecule per 57 carbon atoms in the hybrid according to the calculations (51.23% × 650.16)/(48.77% × 12). The amount of SiPc molecules in this hybrid is about four times larger than that in SWCNT-based hybrid obtained by covalent functionalization with other phthalocyanine molecules [36,37] and comparable with the amount of other linking molecules, viz. pyrene and coumarin, in similar cross-linked hybrid materials [45,61].

Morphological Characterization of the Hybrid Material
The morphology of the SWCNT/SiPc hybrid was studied by SEM ( Figure 6a) and TEM ( Figure  6b). The SWCNT/SiPc hybrid consists of aggregates of tightly packed cross-linked nanotubes. The TEM image of SWCNT/SiPc ( Figure 6b) demonstrates a structure of cross-linked carbon nanotubes with pores of diameter 1-2.6 nm. Such a porous structure makes the 3D material a good candidate for gas sensor applications.

Sensor Response of SWCNT/SiPc Layers to Ammonia and Hydrogen
The chemiresistive sensor response of the hybrid layers was studied toward reducing gases, viz. NH3 and H2. Figures 7 and 8 show the relative sensor responses of SWCNT/SiPc toward ammonia and hydrogen, respectively. The investigated NH3 concentration varied from 1 to 50 ppm because this is within the range of its maximum permissible concentration, which is important for practical applications. The amount of SiPc molecules in this hybrid is about four times larger than that in SWCNT-based hybrid obtained by covalent functionalization with other phthalocyanine molecules [36,37] and comparable with the amount of other linking molecules, viz. pyrene and coumarin, in similar cross-linked hybrid materials [45,61].

Morphological Characterization of the Hybrid Material
The morphology of the SWCNT/SiPc hybrid was studied by SEM ( Figure 6a) and TEM (Figure 6b). The SWCNT/SiPc hybrid consists of aggregates of tightly packed cross-linked nanotubes. The TEM image of SWCNT/SiPc ( Figure 6b) demonstrates a structure of cross-linked carbon nanotubes with pores of diameter 1-2.6 nm. Such a porous structure makes the 3D material a good candidate for gas sensor applications. The amount of SiPc molecules in this hybrid is about four times larger than that in SWCNT-based hybrid obtained by covalent functionalization with other phthalocyanine molecules [36,37] and comparable with the amount of other linking molecules, viz. pyrene and coumarin, in similar cross-linked hybrid materials [45,61].

Morphological Characterization of the Hybrid Material
The morphology of the SWCNT/SiPc hybrid was studied by SEM ( Figure 6a) and TEM ( Figure  6b). The SWCNT/SiPc hybrid consists of aggregates of tightly packed cross-linked nanotubes. The TEM image of SWCNT/SiPc ( Figure 6b) demonstrates a structure of cross-linked carbon nanotubes with pores of diameter 1-2.6 nm. Such a porous structure makes the 3D material a good candidate for gas sensor applications.

Sensor Response of SWCNT/SiPc Layers to Ammonia and Hydrogen
The chemiresistive sensor response of the hybrid layers was studied toward reducing gases, viz. NH3 and H2. Figures 7 and 8 show the relative sensor responses of SWCNT/SiPc toward ammonia and hydrogen, respectively. The investigated NH3 concentration varied from 1 to 50 ppm because this is within the range of its maximum permissible concentration, which is important for practical applications.

Sensor Response of SWCNT/SiPc Layers to Ammonia and Hydrogen
The chemiresistive sensor response of the hybrid layers was studied toward reducing gases, viz. NH 3 and H 2 . Figures 7 and 8 show the relative sensor responses of SWCNT/SiPc toward ammonia and hydrogen, respectively. The investigated NH 3 concentration varied from 1 to 50 ppm because this is within the range of its maximum permissible concentration, which is important for practical applications.  The concentration of hydrogen was changed from 100 to 60,000 ppm. Sensor response is given as R/Ro where R is the resistance at the detected concentration of analyte and Ro is the initial layer resistance. SWCNT/SiPc layers exhibit a completely reversible sensor response to ammonia and hydrogen in the investigated concentration range at room temperature. The injection of both ammonia and hydrogen into the flow cell leads to an increase of the resistance of SWCNT/SiPc layer. Similar behavior was also observed for SWCNTs and their hybrids with other molecules [62]. The gas detection mechanism is based on a change in the material conductivity, which is the result of the reaction of the gas molecules with the CNT surface. Direct charge transfer between the analyte and CNTs was identified as the main sensing mechanism for polar analytes. Under ambient conditions, nanotubes are p-type semiconductors as a result of physical adsorption of oxygen molecules on their surfaces [63]. Thus, interaction with electron donor molecules such as ammonia leads to the depletion of holes level in the investigated layer [64]. The interactions that occur between surface functional groups and NH3 molecules are weak, such as hydrogen bonds, electrostatic interactions (dipole-dipole interactions), which causes the reversibility of the response [65].
It is necessary to mention that the sensor response to hydrogen is much less than that to ammonia at the same concentration. This may be connected with the better electron donating ability of NH3 in comparison with H2 and/or different nature of interaction of the sensing layer with hydrogen. It is worth mentioning that the sensor response of SWCNT/SiPc to ammonia is 10 times higher than that of pristine SWCNTs studied in our previous works [37,66]. The recovery time was 100-200 s in the case of NH3 and 50-120 s in the case of H2 depending on the gas concentrations.  The concentration of hydrogen was changed from 100 to 60,000 ppm. Sensor response is given as R/Ro where R is the resistance at the detected concentration of analyte and Ro is the initial layer resistance. SWCNT/SiPc layers exhibit a completely reversible sensor response to ammonia and hydrogen in the investigated concentration range at room temperature. The injection of both ammonia and hydrogen into the flow cell leads to an increase of the resistance of SWCNT/SiPc layer. Similar behavior was also observed for SWCNTs and their hybrids with other molecules [62]. The gas detection mechanism is based on a change in the material conductivity, which is the result of the reaction of the gas molecules with the CNT surface. Direct charge transfer between the analyte and CNTs was identified as the main sensing mechanism for polar analytes. Under ambient conditions, nanotubes are p-type semiconductors as a result of physical adsorption of oxygen molecules on their surfaces [63]. Thus, interaction with electron donor molecules such as ammonia leads to the depletion of holes level in the investigated layer [64]. The interactions that occur between surface functional groups and NH3 molecules are weak, such as hydrogen bonds, electrostatic interactions (dipole-dipole interactions), which causes the reversibility of the response [65].
It is necessary to mention that the sensor response to hydrogen is much less than that to ammonia at the same concentration. This may be connected with the better electron donating ability of NH3 in comparison with H2 and/or different nature of interaction of the sensing layer with hydrogen. It is worth mentioning that the sensor response of SWCNT/SiPc to ammonia is 10 times higher than that of pristine SWCNTs studied in our previous works [37,66]. The recovery time was 100-200 s in the case of NH3 and 50-120 s in the case of H2 depending on the gas concentrations. The concentration of hydrogen was changed from 100 to 60,000 ppm. Sensor response is given as R/R o where R is the resistance at the detected concentration of analyte and R o is the initial layer resistance. SWCNT/SiPc layers exhibit a completely reversible sensor response to ammonia and hydrogen in the investigated concentration range at room temperature. The injection of both ammonia and hydrogen into the flow cell leads to an increase of the resistance of SWCNT/SiPc layer. Similar behavior was also observed for SWCNTs and their hybrids with other molecules [62]. The gas detection mechanism is based on a change in the material conductivity, which is the result of the reaction of the gas molecules with the CNT surface. Direct charge transfer between the analyte and CNTs was identified as the main sensing mechanism for polar analytes. Under ambient conditions, nanotubes are p-type semiconductors as a result of physical adsorption of oxygen molecules on their surfaces [63]. Thus, interaction with electron donor molecules such as ammonia leads to the depletion of holes level in the investigated layer [64]. The interactions that occur between surface functional groups and NH 3 molecules are weak, such as hydrogen bonds, electrostatic interactions (dipole-dipole interactions), which causes the reversibility of the response [65].
It is necessary to mention that the sensor response to hydrogen is much less than that to ammonia at the same concentration. This may be connected with the better electron donating ability of NH 3 in comparison with H 2 and/or different nature of interaction of the sensing layer with hydrogen. It is worth mentioning that the sensor response of SWCNT/SiPc to ammonia is 10 times higher than that of pristine SWCNTs studied in our previous works [37,66]. The recovery time was 100-200 s in the case of NH 3 and 50-120 s in the case of H 2 depending on the gas concentrations. The change in the resistance of hybrid films when interacting with H 2 appears to have different mechanism than in the case of ammonia. It was assumed in the literature [67,68] that atmospheric oxygen absorbs at the air/hybrid interface. The formation of charge-transfer complexes at the interface between air and phthalocyanine attached to the SWCNT walls leads to the generation of oxidized MPc + and O 2 − species and injection of hole charge carriers into the layer's bulk. Then the sensor layer is exposed to hydrogen atmosphere, and the hydrogen molecules react with the adsorbed oxygen species. The redox reaction is exothermic and leads to the fast desorption of produced H 2 O molecules from the surface. Due to the released electrons, a change of the resistance of the hybrid layer is observed. The dependence of sensor response of SWCNT/SiPc to ammonia concentration is almost linear in the investigated concentration range (Figure 7b). The sensor response to hydrogen was investigated in the wider concentration range up to 50,000 ppm because hydrogen is a highly flammable gas and it burns at concentrations as low as 4% in air and its detection at comparatively high concentrations in air is also of practical importance. The layer exhibits linear dependence between 100 ppm and 1000 ppm of hydrogen (Figure 8b). For the concentrations above 1000 ppm the response is no more linear and the sensitivity to hydrogen continuously decreases. This means that the sensors reach a saturation state. The limits of detection (LOD) were calculated according to the 3·s/m [69], where m is the slope of the calibration plots (Figures 7b and 8b) and s is the standard deviation of the response value at the NH 3 concentration of 1 ppm and H 2 concentration of 100 ppm. The standard deviation of the response value was calculated using the data for five parallel samples. These values were 0.5 ppm and 70 ppm for ammonia and hydrogen, respectively.
Characteristics of the investigated SWCNT/SiPc-based sensor are compared with those of some sensors described in the literature in Table 1. The sensing layers based on SWCNT/SiPc are quite competitive with the active layers on the basis of such carbon nanomaterials as graphene (G), reduced graphene oxide (rGO), multiwalled carbon nanotubes (MWCNT) functionalized with metal oxides, phthalocyanines, poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT-PSS), described in the literature. The SWCNT/SiPc layers 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. Water vapor often acts as an interfering gas for the detection of electron donor gases. Therefore, the effect of water on the sensor response is of great interest for this study. The sensor responses of SWCNT/SiPc hybrids toward NH 3 (1-50 ppm) measured at RH 5, 25, 45 and 75% are shown in Figure 9 as examples. At RH 20% the sensor response was almost the same as that measured at RH 5%, whereas the further increase of RH to 50 and 70% led to its 2-3 fold decrease. At the same time, the increase of RH does not noticeably influence the response and recovery times. 5%, whereas the further increase of RH to 50 and 70% led to its 2-3 fold decrease. At the same time, the increase of RH does not noticeably influence the response and recovery times.
Like the NH3 molecule, water is also an effective electron donor agent. The presence of water leads to the increase of the base resistance of hybrid layers from 2.4·10 4 to 2.7·10 4 Ohm (Figure 9). Sorption of water molecules on the surface of hybrid material appears to lead to the occupation of active sites and contribute to the saturation of the sensing layer. Consequently, the response to ammonia decreases with the increase of RH [64].

Cross-Sensitivity toward Various Gases
The selectivity and cross-sensitivity of the SWCNT/SiPc hybrid material was tested using H2S, CO2, acetone, dichloromethane, ethanol, and formaldehyde. The diagram of the sensor response of SWCNT/SiPc hybrid layers toward NH3 and these interfering analytes is shown in Figure 10. The sensitivity of SWCNT/SiPc layers to ammonia is noticeably higher compared to that towards other investigated gases and volatile organic vapors (VOC). This makes this hybrid material attractive for the detection of ammonia in the presence of VOCs and CO2.
The detection of ammonia and hydrogen was also carried out in the mixture with H2S and CO2. Figure 11a shows the sensor response of SWCNT/SiPc to NH3 (20 ppm) in presence of CO2 (1·10 4 -5·10 4 ppm). The addition of comparatively high concentrations of CO2 does not lead to any noticeable changes in the sensor response value of SWCNT/SiPc layers to 20 ppm NH3. The value of sensor response to hydrogen is also stable and does not change upon addition of carbon dioxide (Figure 11b). It means that both ammonia and hydrogen can be successfully detected in the presence of CO2. Like the NH 3 molecule, water is also an effective electron donor agent. The presence of water leads to the increase of the base resistance of hybrid layers from 2.4·10 4 to 2.7·10 4 Ohm (Figure 9). Sorption of water molecules on the surface of hybrid material appears to lead to the occupation of active sites and contribute to the saturation of the sensing layer. Consequently, the response to ammonia decreases with the increase of RH [64].

Cross-Sensitivity toward Various Gases
The selectivity and cross-sensitivity of the SWCNT/SiPc hybrid material was tested using H 2 S, CO 2 , acetone, dichloromethane, ethanol, and formaldehyde. The diagram of the sensor response of SWCNT/SiPc hybrid layers toward NH 3 and these interfering analytes is shown in Figure 10. The sensitivity of SWCNT/SiPc layers to ammonia is noticeably higher compared to that towards other investigated gases and volatile organic vapors (VOC). This makes this hybrid material attractive for the detection of ammonia in the presence of VOCs and CO 2 .
Molecules 2020, 25   On the contrary, H2S was found to be an interfering gas for the NH3 detection. Figure 11c shows that while low concentrations of H2S (below 5 ppm) does not affect the sensor response to ammonia (20 ppm), the further growth of H2S concentrations in the investigated gas mixture causes a noticeable increase in the sensor response. The electron donating character of H2S similarly to NH3 appears to cause the additional increase of the layer resistance and, as a consequence, the total sensor response.

Effect of Operating Temperature
Because every sensor has its own preferred operating temperature, the sensor response of SWCNT/SiPc material toward ammonia and hydrogen was tested at 25, 40, 60 and 80 °C (Figure 12).   Figure 10. Sensor response of SWCNT/SiPc hybrid layers to different analytes.
The detection of ammonia and hydrogen was also carried out in the mixture with H 2 S and CO 2 . Figure 11a shows the sensor response of SWCNT/SiPc to NH 3 (20 ppm) in presence of CO 2 (1·10 4 -5·10 4 ppm). The addition of comparatively high concentrations of CO 2 does not lead to any noticeable changes in the sensor response value of SWCNT/SiPc layers to 20 ppm NH 3 . The value of sensor response to hydrogen is also stable and does not change upon addition of carbon dioxide (Figure 11b). It means that both ammonia and hydrogen can be successfully detected in the presence of CO 2 .
Molecules 2020, 25, x FOR PEER REVIEW 10 of 18   On the contrary, H2S was found to be an interfering gas for the NH3 detection. Figure 11c shows that while low concentrations of H2S (below 5 ppm) does not affect the sensor response to ammonia (20 ppm), the further growth of H2S concentrations in the investigated gas mixture causes a noticeable increase in the sensor response. The electron donating character of H2S similarly to NH3 appears to cause the additional increase of the layer resistance and, as a consequence, the total sensor response.  On the contrary, H 2 S was found to be an interfering gas for the NH 3 detection. Figure 11c shows that while low concentrations of H 2 S (below 5 ppm) does not affect the sensor response to ammonia (20 ppm), the further growth of H 2 S concentrations in the investigated gas mixture causes a noticeable increase in the sensor response. The electron donating character of H 2 S similarly to NH 3 appears to cause the additional increase of the layer resistance and, as a consequence, the total sensor response.

Effect of Operating Temperature
Because every sensor has its own preferred operating temperature, the sensor response of SWCNT/SiPc material toward ammonia and hydrogen was tested at 25, 40, 60 and 80 • C ( Figure 12). The value of sensor response to both gases was maximal at 40 • C, however further temperature growth led to its decrease compared to that measured at room temperature. This behavior appears to be due to two competing processes, namely an increase of desorption [74] and growth of the concentration of charge carriers in p-type semiconductors with temperature [75]. Different rate of desorption of the chemisorbed oxygen from the surface of hybrid material at different temperatures may also contribute to these processes. Therefore, the operation temperature is a unique characteristic of each sensor. The maximal sensor response to ammonia is exhibited at 40 • C but the further increase of the temperature to 60-80 • C leads to the noticeable decrease of the response (Figure 12a). Similar tendency is also observed in the case of hydrogen detection (Figure 12b), however, the change is not as significant as in the case of ammonia.
Molecules 2020, 25, x FOR PEER REVIEW 11 of 18 The value of sensor response to both gases was maximal at 40 °C, however further temperature growth led to its decrease compared to that measured at room temperature. This behavior appears to be due to two competing processes, namely an increase of desorption [74] and growth of the concentration of charge carriers in p-type semiconductors with temperature [75]. Different rate of desorption of the chemisorbed oxygen from the surface of hybrid material at different temperatures may also contribute to these processes. Therefore, the operation temperature is a unique characteristic of each sensor. The maximal sensor response to ammonia is exhibited at 40 °C but the further increase of the temperature to 60-80 °C leads to the noticeable decrease of the response (Figure 12a). Similar tendency is also observed in the case of hydrogen detection (Figure 12b), however, the change is not as significant as in the case of ammonia.

Materials
Propargyl alcohol, dichlorosilicon (IV) phthalocyanine, copper(II) sulphate pentahydrate, sodium-L-ascorbate, NaN3, iodine monochloride and all solvents were purchased from Sigma Aldrich (Saint Louis, MO, USA). All reactions were carried out under a dry argon atmosphere unless otherwise noted. Column chromatography for the purification of the crude compounds was performed on silica gel 60. The purity of the products was tested at each step using thin layer chromatography (silica gel F-254 coated TLC plates).

Azido Functionalized SWCNTs (SWCNT-N3)
Azido functionalized SWCNTs (SWCNT-N3) were prepared according to the literature [58]. For this purpose, 250 mg (3.84 mmol) NaN3 and 165 mg (1.01 mmol) iodine monochloride were dissolved in 25 mL of acetonitrile under a nitrogen atmosphere. This mixture was magnetically stirred for 15 min at 0 °C, followed by adding 20 mg SWCNTs to the solution; the reaction was stirred at room temperature for 24 h and filtered off. The solid products were washed with DMF and ethanol to remove unreacted compounds, then dried under vacuum.

Materials
Propargyl alcohol, dichlorosilicon (IV) phthalocyanine, copper(II) sulphate pentahydrate, sodium-L-ascorbate, NaN 3 , iodine monochloride and all solvents were purchased from Sigma Aldrich (Saint Louis, MO, USA). All reactions were carried out under a dry argon atmosphere unless otherwise noted. Column chromatography for the purification of the crude compounds was performed on silica gel 60. The purity of the products was tested at each step using thin layer chromatography (silica gel F-254 coated TLC plates).

Azido Functionalized SWCNTs (SWCNT-N3)
Azido functionalized SWCNTs (SWCNT-N3) were prepared according to the literature [58]. For this purpose, 250 mg (3.84 mmol) NaN 3 and 165 mg (1.01 mmol) iodine monochloride were dissolved in 25 mL of acetonitrile under a nitrogen atmosphere. This mixture was magnetically stirred for 15 min at 0 • C, followed by adding 20 mg SWCNTs to the solution; the reaction was stirred at room temperature for 24 h and filtered off. The solid products were washed with DMF and ethanol to remove unreacted compounds, then dried under vacuum.

Axially-bis(propynoxy)phthalocyaninato Silicon (IV) (SiPc)
Axially-dichlorophthalocyaninato silicon (IV) (200 mg, 0.32 mmol) and propargyl alcohol (0.376 mL, 6.4 mmol) were refluxed in dry pyridine (15 mL) under an argon atmosphere for 48 h. Then the reaction mixture was cooled to room temperature and poured into water. The resulting solid was filtered and washed several times with hot water. The crude product was purified by column chromatography using a mixture of chloroform-EtOH (99:1) solvents as eluent. A green solid product was obtained and dried over phosphorus pentoxide. Yield: 30 mg (31%), Chemical formula: C 38  The synthetic pathways of axially terminal propynoxy substituted silicon (IV) phthalocyanine (SiPc) and its hybrid nanomaterial with SWCNTs are presented in Figure 1.

Characterization Methods
The structure of axially-bis(propynoxy)phthalocyaninato silicon (IV) was characterized by the methods of elemental analysis (Flash 1112 Instrument, Thermo Finnigan, Waltham, MA, USA), MALDI mass-spectrometry (Matrix Assisted Laser Desorption Ionization (MALDI) Microflex LT, Bruker, Bruker Daltonics MS, Bremen, Germany), 1 H-NMR (500 MHz spectrometer, Varian, California, USA, tetramethylsilane was an internal standard), FTIR (Spectrum 100 spectrophotometer, Perkin Elmer, Waltham, MA, USA) and UV-vis spectrophotometry (PG Instruments, Lutterworth, UK or Shimadzu 2101, Kyoto, Japan) in the range of 190-1100 nm. The composition of the novel hybrid material was confirmed by FT-IR and Raman (a LabRAM Horiba single spectrometer equipped with a CCD Symphony detector (Jobin Yvon, HORIBA, Montpellier, France), 488 nm, 20 mW line of Ar-laser) spectroscopies. All vibrational spectra are given without background subtraction. The morphology of the obtained hybrid was investigated via scanning electron microscopy (SEM) using a FEI-Nova system (Hillsboro, OR, USA) with magnification of 160,000× and high voltage of 10 kV and transmission electron microscopy (TEM) (G2 F20S-TWIN instrument, a versatile multipurpose 200 kV STEM/TEM/nano-analysis system, Tecnai, Hillsboro, OR, USA). Thermogravimetric analysis (TGA) was carried out using a STAReThermal Analysis System (Mettler Toledo, Zurich, Switzerland) at the heating rate of 10 • C·min −1 under a nitrogen flow (50 mL·min −1 ).

Sensor Properties
The chemiresistive sensor response of the novel hybrid material toward different gases (NH 3 , CO 2 , H 2 S, H 2 and their mixtures) was studied as described in the literature [37,[76][77][78]. Hybrid suspensions (20 µL, 2 mg·mL −1 ) in dichloromethane subjected to ultrasonic treatment were spun onto a glass substrate with interdigitated Pt electrodes having the 10 µm dimension gap (G-IDEPT10; cell constant is 0.0118 cm −1 DropSens, Oviedo, Spain). Then a substrate connected with an electrometer (mod. 236, Keithley, Beaverton, OR, USA) was placed into the flow cell with the internal volume of 5 cm 3 . The resistance of layer was constantly measured at the dc voltage of 10 V. Analyzed gases diluted by dry air to the required concentration were passed through the cell with the constant flow rate of 300 mL·min −1 during 30 s. Then the gas flow was blocked, and the layer was held until its resistance reached a steady state value. After that, the cell was purged with air. The next portion of gas was supplied after the resistance of the film reached its initial value. The relative sensor response was calculated as R/R o , where R is the current layer resistance at the detected concentration of the analyte and R o is the layer resistance before the analyte injection. For the investigation of humidity effect, the measurements were carried out in the same way but the carrier gas (air) was bubbled through distilled water. The relative humidity (RH) inside the cell was monitored by a humidity meter (MPE-202.013).

Conclusions
The novel hybrid nanomaterial (SWCNT/SiPc) made of of SWCNTs cross-linked by axially substituted silicon (IV) phthalocyanine (SiPc) was obtained by the azide-alkyne cycloaddition (Click) reaction of azido-substituted SWCNTs with axially-bis(propynoxy)phthalocyaninato silicon (IV) (SiPc). The synthesized axially bis(propynoxy) substituted silicon (IV) phthalocyanine and its hybrid were characterized by spectral methods (FT-IR, 1 H-NMR, UV-Vis, mass), thermogravimetry and electron microscopy (SEM and TEM). According TGA data, the amount of linking SiPc molecules covalently bonded to SWCNT was estimated to be one SiPc molecule per 57 carbon atoms. The layers of the prepared hybrid were tested as active layers of chemiresistive sensors toward ammonia (1-50 ppm) and hydrogen (100-60,000 ppm) at different temperatures, relative humidity and in the presence of interfering gases like carbon dioxide, hydrogen sulfide and volatile organic vapors. The hybrid nanomaterial exhibited the completely reversible sensor response to both gases at room temperature; the recovery time was 100-200 s for NH 3 and 50-120 s in the case of H 2 depending on the gas concentrations. At RH 20% the sensor response was almost the same as that measured at RH 5%, whereas the further increase of RH to 50 and 70% led to its 2-3 fold decrease. It was demonstrated that the SWCNT/SiPc layers can be successfully used for the detection of both NH 3 and H 2 in the presence of CO 2 . On the contrary, H 2 S was found to be an interfering gas for the NH 3 detection.