D. Physical Characterization of TPPNi active layer
i. Optical Study of TPPNi:
The optical properties of the TPPNi compound have been investigated using UV–vis absorption spectroscopy (wavelength range ~ 300-800 nm) in solution as well as solid-state, as shown in Figure 4 and Figure 4 (inset), respectively.
Porphyrins have two electronic transitions in the visible domain of the electromagnetic spectrum: a Soret band at 350–500 nm and Q-bands around 500–700 nm with typically one order of magnitude lower intensity [19]. The UV–vis absorption spectrum of the TPPNi solution (in chloroform) displayed the characteristic Soret band between 355 and 465nm, with a maximum absorption peak at 415nm, which is attributed to the π-π* transition from the ground state (S0) to the second-lowest singlet state (S2). On the other hand, the broadband at 500–575 nm with peak absorption at 525 nm is due to π–π* electron transition from the ground state (S0) to the lowest excited singlet state (S1). Quite interestingly, albeit the absorption spectra of TPPNi solution (in chloroform) and thin film (prepared by spin coating its solution in chloroform on pre-cleaned glass substrates) are somewhat similar; however, there is a significant difference in the shape of the Soret band of both spectra. Specifically, in the solution state, TPPNi compound exhibited a narrow Soret band; whereas, the Soret band has become substantially broader in the solid-state spectrum. In addition, the characteristic Soret band peak has been observed to be redshifted significantly. The observed results may be due to the aggregates formation in thin film, that ultimately results in an increased π–π interaction, as reported by some other studies [20-22].
ii. Structural study of TPPNi:
The crystalline structure of the TPPNi semiconducting layer has been analyzed by X-Ray Diffraction (XRD) pattern, which displays diffraction intensity as a function of 2θ (as shown in Figure 5). Typically, the existence of a amorphous solid form can be confirmed by detecting the absence of the distinct XRD peaks, which are envisaged to be the characteristic of crystalline order [23]. The appearance of a general "halo" pattern at 2θ ~ 23.5o may further point towards the occurrence of amorphous, glassy, or disordered material.
iii. Surface morphology of TPPNi thin film:
Field Emission Scanning Electron Microscopy (FESEM) has been used to characterize the surface morphology of a pristine TPPNi thin film. The FESEM micrographs (Fig, 6 (a) and (b)) depict the TPPNi thin film at different magnification scales (500 and 1.3k, respectively). It may be clearly observed that the humidity sensing TPPNi layer comprises essentially of micro-pyramidal shaped structures (decorated with inhomogeneous, irregular shaped sub-micron particles). In addition, the internal structure of the sensing layer contains a fine network of voids/pores resulting in a "sponge-like" structure. In fact, the porous morphological characteristic appears to be intrinsic for porphyrin-sponges which is a general name for a variety of phenyl-meso-exchanged metalloporphyrin analogues [24].
The microporous structure of the active thin film is envisaged to be vital for superior humidity sensing ability since it allows a stronger interaction between the analyte (water molecules) and the sensing layer. Hence, we believe that TPPNi is an ideal template for humidity sensing by virtue of bulk porosity and essential void spaces (between microstructures), which may assist the efficient humidity circulation through the bulk. Further, the irregular-shaped sub-micron particles embedded on the pyramid-shaped structures are also believed to provide a larger specific area for improved humidity adsorption.
To supplement the aforementioned experimental results, the morphology of the pristine TPPNi thin film has also been studied via Atomic Force Microscope (AFM). Figure 7 (a) and (b), depict the two and three-dimensional AFM images of the spin coated pristine TPPNi thin film, respectively, with an examination area of 7.5 x 7.5μm. Typically, for surface morphology investigation of (very) rough surfaces, the 3D surface measurements are preferred over the top-view projections [25]. Fig 7 (c) portrays the section analysis at four distinct randomly selected locations (pre-specified in Figure 7 (a)). It may be clearly observed that the surface of the sensing layer is rough and exhibits positive skewness i.e., the surface exhibits mainly peaks and asperities. The prominent high surface roughness in TPPNi humidity sensing film gives a significant rise to surface-to-volume ratio [26], which is ultimately expected to yield higher sensitivity of the humidity sensing device.
E. Electrical Characterization of Humidity sensor
i. Humidity sensing performance study of TPPNi
Generally, humidity influences a wide variety of physical, chemical and biological processes, and these effects can later be exploited to estimate variation in varied humidity levels [27]. When operated in capacitive mode, the fabricated humidity sensor utilizes the TPPNi sensing layer as a dielectric layer. The sensing layer adsorbs and desorbs the water molecules in proportion to the ambient relative humidity during its capacitive mode of operation. The area of the Aluminium electrodes (A), inter-electrodes gap (d) and the dielectric permittivity constant (εr) of the TPPNi dielectric material influence the capacitance of the fabricated device (represented mathematically in Eq. 1), [28].
Where "C" is the capacitance of the fabricated device and "εo" represents the dielectric permittivity of air.
Dielectric permittivity of the active humidity sensing layer is triggered by polarization in the TPPNi layer (humid and desiccated). Typically, there are four mechanisms i.e., dipolar, ionic, space charge, or electronic, which may contribute towards polarizability of the active layer [29]. The Clausius–Mosotti equation defines the relationship between the dielectric constant (εr) and polarizability (αd) as given in Eq. 2 [30].

Whereas Eq. 3 describes the relationship between dielectric constant and capacitance [31].

Here εdry and εwet are the relative dielectric constants for the desiccated and humid active sensing layer, respectively and "n" is the dielectric morphology related factor. Generally, the dielectric permittivity of desiccated organic semiconductor layer is ~5 which is considerably smaller than that of water ~80 [32]. Naturally, with the continuing adsorption of water molecules by the TPPNi thin layer, the dielectric permittivity of the humid sensing layer varies significantly [33].
Figure 8 depicts the capacitance-relative humidity response of the fabricated humidity sensor for a range of 39 to 85 %RH measured at four distinct frequencies (500 Hz, 1 kHz, 10 kHz and 100 kHz) of the AC test signal. In general, for all test frequencies, the capacitance of the fabricated device displays a monotonous nonlinear increase as a function of %RH. Moreover, this nonlinear response can be correlated to the prolonged relaxation period of the dipole moments of adsorbed water molecules [34]. In comparison to high test frequencies (1, 10, and 100 kHz), the influence of %RH variation on the capacitance was shown to be larger at low operating frequency (500 Hz). The sensor's capacitance has shown an increase by 54.36 times in magnitude at test frequency ~ 500 Hz with an increase in %RH from 39 to 85%, as shown in Figure 8. A decrease in the capacitance change has been observed at higher frequencies, precisely 23.65, 19.58, and 15.77 times, for 1 kHz, 10 kHz, and 100 kHz, respectively. The sensitivity of the fabricated device towards ambient humidity has been measured to be 146.17, 51.94, 42.41, and 32.35 pF/%RH at four distinct frequencies of the AC test signal. This is very well correlated with the formerly established fact by E. Pinottie et al. that based on the low intrinsic mobility of organic semiconductors, in some cases, the charge carrier cannot follow the rapid change in the applied electric field due to applied test signal at higher frequencies [35]. As a result, the polarization mechanism becomes less effective, and the dielectric permittivity of the active layer decreases at higher frequencies [36, 37].
The capacitance variation in the 39 – 58 %RH range is not noticeable, as seen in Figure 8, due to the very well-known fact that the coverage of water molecules on the active sensing layer is not noticeable at low ambient humidity levels. Primarily, water molecules are chemisorbed (in the form of a monolayer) on the sensing thin film by virtue of the electron vacancies on the surface [36]. On the chemisorbed water layer template, several physisorbed water molecular layers continue to accumulate as moisture levels rise [38]. Additional water adsorbed molecules strengthen the polarization and significantly increase the capacitance of the sensor [39]. Thus, in a range of 58 – 85 %RH, a steady increase in capacity is conveniently observed.
Interestingly the OSCs provide a technological attractive charge transport property that is significantly modulated with ambient conditions, in particular humidity. The influence of ambient relative humidity, in the 39 – 85 %RH range, on the impedance of the fabricated sensor for three test frequencies (1 kHz, 10 kHz, and 100 kHz) is depicted in Figure 9. It may be conveniently observed that for all test frequencies, impedance of the sensor exhibits a similar trend, i.e., decrease in magnitude with the upsurge in ambient relative humidity. At 1 kHz test frequency, an electrical impedance change of 28.32 times was detected at 85 percent RH compared to 39 percent RH, resulting in a 48.23 kΩ/ percent RH sensitivity. Similarly, the sensitivity at higher frequencies such as 10 kHz and 100 kHz sensitivity of 32.11 kΩ/%RH and 13.00 kΩ/%RH has been recorded. The aforementioned results prove that TPPNi semiconductor-based humidity sensor can effectively function in the dual (capacitive and conductometric) mode for ambient relative humidity monitoring.
The operating mechanism of impedance-type sensors may be described with the help of the Grothus mechanism. At low %RH range, primarily immobile chemisorbed water molecules layer is formed on the surface of TPPNi thin film, and the conduction of the active layer at this stage is mainly by virtue of intrinsic electrons only [40]. Furthermore, as the %RH level rises, layers of multi-physisorbed water molecules are adsorbed on the active sensing layer. These physisorbed layers exhibit liquid-like behavior and swiftly decompose into hydronium ions (H3O)+ as charge carriers, as described in chemical Eq. (4). Therefore, the conductivity of the semiconductor thin film at higher %RH is now dictated by the ionic conduction [41]. In bulk, hydronium ion releases hydrogen ion (H+) to its neighboring water molecule, and the chain reaction continues. The effective proton hopping between neighboring molecules in physisorbed H2O molecules layers considerably reduces the electrical impedance of the TPPNi sensing layer [42].
When analyzing the sensor's performance, the response time or the recovery/reset time is a critical parameter of interest. It is computed during the humidification/desiccation cycle of the humidity sensor's dynamic curve. [43]. The sensor's temporal capacitive responses to step-change in ambient relative humidity levels are depicted in Figures 10 (a) and (b). As shown in Figure10 (a), the sensor in capacitive mode shows a stable baseline initially when measured at 45%RH, and consequently, with step input in %RH from 45% to 85%, the average response time has been evaluated to be ~130 s. Similarly, the reset time in the capacitive mode of operation has been recorded to be 156 s, as shown in Figure 10 (b). The effective diffusivity of water molecules in the active sensing layer can be securely attributed to the constructed humidity sensor's considerably slow response/recovery time.
Table 1 compares the proposed TPPNi-based capacitive and conductometric humidity sensor to previously reported sensors in terms of critical performance metrics. Although it is a little inefficient in terms of response/reset time, the proposed sensor outperforms others in terms of sensitivity. It is expected that by selecting the right doping material and quantity, the sensitivity will be improved, and the response time will be significantly reduced. The impact of doping and other geometrical parameters is being investigated and will be reported later.
Table 1. Comparison of humidity sensors based on key performance parameters.
Material
|
Mode of operation
|
Sensitivity
|
Bandwidth
|
Response/
Reset time
|
DMBHPET [44]
|
Capacitive
|
0.007 pF / %RH
|
30–80 %RH
|
10, 15 sec
|
Polyimide [45]
|
Capacitive
|
22.29 pF/%RH
|
20-90 %RH
|
25 sec
|
Methyl-red [46]
|
Capacitive
|
16.92 pF/%RH
|
30-95 %RH
|
~10 sec each
|
ZnO-SnO2 composite thin film [47]
|
Conductometric
|
8.6 kΩ/%RH
|
32-92 %RH
|
17, 65 sec
|
polyaniline/PVA[48]
|
Conductometric
|
12.6 kΩ/%RH
|
30-85 %RH
|
-
|
TPPNi
|
Capacitive and Conductometric
|
146.17 pF/%RH @ 500 Hz
48.23 kΩ/%RH @ 1kHz
|
39-85 %RH
|
130, 156 sec
|