Influence of the Nature of Aminoalcohol on ZnO Films Formed by Sol-Gel Methods

Here we present comparative studies of: (i) the formation of ZnO thin films via the sol-gel method using zinc acetate dihydrate (ZAD), 2-methoxyethanol (ME) as solvent, and the aminoalcohols (AA): ethanolamine, (S)-(+)-2-amino-1-propanol, (S)-(+)-2-amino-3-methyl-1-butanol, 2-aminophenol, and aminobenzyl alcohol, and (ii) elemental analyses, infrared spectroscopy, X-ray diffraction, scanning electron microscopy, absorption and emission spectra of films obtained after deposition by drop coating on glass surface, and thermal treatments at 300, 400, 500 and 600 °C. The results obtained provide conclusive evidences of the influence of the AA used (aliphatic vs. aromatic) on the ink stability (prior to deposition), and on the composition, structures, morphologies, and properties of films after calcination, in particular, those due to the different substituents, H, Me, or iPr, and to the presence or the absence of a –CH2 unit. Aliphatic films, more stable and purer than aromatic ones, contained the ZnO wurtzite form for all annealing temperatures, while the cubic sphalerite (zinc-blende) form was also detected after using aromatic AAs. Films having frayed fibers or quartered layers or uniform yarns evolved to “neuron-like” patterns. UV and photoluminescence studies revealed that these AAs also affect the optical band gap, the structural defects, and photo-optical properties of the films.


Materials
The AAs (a-e, see Figure 1) were obtained from of Acrös Organics, ZAD from reac, and ME from Aldrich. All the reagents were used as received.

Film Preparation
A 0.5 g amount of ZAD was treated with the equimolar amount of the selected (a-e) and afterwards 5 mL of the solvent (ME) was added. The mixtures were stirre room temperature for 30 min and glass substrates were then drop-coated with the tained inks, and heated to different temperatures (T = 300, 400, 500, and 600 °C) to pro the films 1a-1e. Moreover, recent studies point towards some influence of the stabilizer's molecular flexibility on bond torsion and breaking, and that, therefore, the substituents may have a strong influence on the decomposition processes of the ZAD-stabilizer system [33], and also on the composition and properties of the films obtained after thermal treatment. The final ZnO obtained by using diethanolamine {HN(CH 2 CH 2 OH) 2 , DEA} as stabilizer retained more N than the one obtained by using aminoalcohol a {H 2 NCH 2 CH 2 OH, MEA} under identical conditions, even after calcination at high temperatures. Moreover, peaks due to other intermediate and undesirable species, such as Zn(OH) 2 , were revealed by X-ray diffraction studies in the first case [44], thus indicating that the AA has a crucial role in the sol-gel ZnO formation process and also on the properties of the final material.
In previous reports, we have demonstrated that aminoalcohols (AA), such as (S)-(+)-2amino-1-propanol (b) and (S)-(+)-2-amino-3-methyl-1-butanol (c) (both shown in Figure 1 and closely related to (a), as well as the phenyl containing derivatives: 2-aminophenol and aminobenzyl alcohol (Figure 1, d and e, respectively), are also valuable stabilizers in the sol-gel process [33,34,41,42]. However, there is a lack of information about the role and influence of the nature of these five AAs in the complex and multistep transformation from the precursors to the final ZnO films and in their properties. Since a-e are valuable stabilizers in ZnO sol-gel formation and have different capability to deprotonate-which affects the formation of the alkoxide or alkoxy-Zn(II) complexes-they appeared to be excellent candidates to find out the effects produced by: a) the nature of the AA (aliphatic (a-c) or aromatic (d-e)), b) the presence of the substituents Me (in b) or i Pr (in c) or of the -CH 2 -unit (in e) on the ZnO formation processes as well as on stability, crystallographic, morphological, optical, and spectroscopical properties of the final ZnO thin films (1a-1e). In this work, we present a comparative study of the ZnO thin films obtained by SGM, using ZAD, the AAs (a-e), and ME, after thermal treatments at different temperatures.

Materials
The AAs (a-e, see Figure 1) were obtained from of Acrös Organics, ZAD from Panreac, and ME from Aldrich. All the reagents were used as received.

Film Preparation
A 0.5 g amount of ZAD was treated with the equimolar amount of the selected AA (a-e) and afterwards 5 mL of the solvent (ME) was added. The mixtures were stirred at room temperature for 30 min and glass substrates were then drop-coated with the obtained inks, and heated to different temperatures (T = 300, 400, 500, and 600 • C) to produce the films 1a-1e.

Characterization Methods
Quantitative determinations of nitrogen, carbon, and hydrogen content in the samples were carried out by elemental organic analyses (EOA) with Thermo EA Flash 2000 equipment working in standard conditions (helium flow of 140 mL/min; combustion furnace at 950 • C; chromatographic column oven at 65 • C). Infrared (IR) spectra were registered at room temperature with a Nicolet 400FTIR instrument using KBr discs.
X-ray diffraction (XRD) data were obtained with a PANalytical X'Pert PRO MPD alpha1 Bragg-Brentano powder diffractometer using a Cu tube operating at 45 kV and 40 mA, a Johansson type Ge (111) primary focalizing monochromator, and a solid-state strip 1D PIXcel 1D detector. High-resolution, high statistics, full angular range, Cu Kα 1 (λ = 1.5406 Å) data were obtained with the following parameters: 2θ/θ scans from 5 to 140 degrees; step size 0.0263 deg.; measuring time per step 200 s (PIXcel 1D active length 3.347 deg.); total measuring time per sample 1.2 h. An automatic divergence slit system and a mask enabled a constant irradiated surface (10 × 12 mm 2 ) over the analyzed samples. The diffracting volume was also constant regarding the small and finite thickness (bellow 5 µm) of the characterized layers. Full profile analysis was performed, applying Rietveld refinement [47] for the identified crystalline phases. The refinements have been performed through the TOPAS v6 software [48]. The peak width of each phase was modelled with the double-Voigt approach by considering both the Lorentzian contribution of the crystallite size effect and the Gaussian contribution of the microstrain to the peak width [49]. Preferential orientation corrections were applied by spherical harmonics [50]. The background was modelled with a 15th order Chebyschev polynomial. The instrumental contribution to the diffraction profile was calculated with the fundamental parameters approach [51].
Film morphology was observed by scanning electron microscopy (SEM) with a JEOL JSM-7100F and a JEOL JSM-6510 in planar and cross-section views. The UV-VIS spectra were measured with a Lambda-950 of PerkinElmer, with wavelengths between 190 and 1100 nm for which the precision UV-VIS is 0.05-5 nm and precision near-IR is 0.2-20 nm. The photoluminescence (PL) emission spectra of the films in a range between 370 and 850 nm were obtained at room temperature by exciting the samples (with the 325 nm line of a He-Cd laser, using a power density of 8 kWcm −2 ) and collecting the emitted light with the same experimental setup; the PL intensity of all spectra was corrected by the spectral response of the system.

Results and Discussion
The first difference detected between the different inks arises from the appearance and stability of the samples obtained immediately after preparation, even before deposition. The aliphatic AAs (a-c) samples were white and no significant change was observed after several weeks of storage at room temperature. In contrast, the aromatic sample d turned black in a few seconds; while the freshly-prepared sample with e was initially whitish but it turned dark brown, nearly black, in a few minutes under identical experimental conditions. These findings suggest that at room temperature the stability of the freshly prepared inks increases according to the sequence d < e < a-c. The lower stability of the aromatic systems could be attributed to their stronger photo-reactivity.
Comparative studies of the appearance, IR spectra, composition, and X-ray diffractograms of the films 1a-1e obtained at 400, 500, and 600 • C were carried out in order to achieve additional information on the effect produced by the different AAs on the ZnO-film formation process, their stability, and their properties.

General Appearance and Chemical Analysis
Pictures of films obtained after the thermal treatments are presented in Figure 2, showing that after annealing at 300 • C, samples 1d and 1e were clearly darker than 1a-1c, which exhibited a yellowish or brownish tonality. This finding is similar to that observed for the freshly prepared samples at room temperature. After annealing at 500 • C, films 1a-1c turned white; while 1d and 1e still had black (for 1d) or deep-brown (for 1e) regions. This fact suggests that the aliphatic or aromatic nature of the AAs may be important as to modify the ZnO-formation path (as explained in more detail in the following sections). Finally, after annealing at 600 • C, all films showed the typical white color of ZnO, with 1d exhibiting higher transparency. As visible in Figure 2, after storage of these samples for two years in the dark and at room temperature, no significant deterioration was observed. This is relevant in view of their potential applications.
IR spectra of the films obtained at each annealing temperature ( Figure 3) show that the intensity of the bands due to specific organic units or bonds of the aminoalcohols (i.e., C-H, C-C, or C-N, and C-O) decreased gradually as the temperature increased. For T = 400 • C, the IR spectra of 1a-1c have a very weak band at 478 cm −1 that is characteristic of the crystalline ZnO [52]. In Figures 2 and 3 it can be seen that after annealing at 500 • C: (a) films 1a-1c exhibited the typical color of the ZnO and (b) the intensity of the band at 478 cm -1 increased gradually, while those related to the functional groups of the organic units (mainly those between 1200 and 1700 cm −1 ) decreased. Moreover, annealing at 600 • C caused the absorption band related to the ZnO bond to become the most intense one. We also undertook a parallel study with 1d and 1e films (Figure 3), and, as expected, the intensity of the bands observed in the range 1600-1400 cm −1 (due to the carbon-carbon stretching vibrations of the phenyl ring [53]) decreased from 300 to 400 • C. It should be noted that at these temperatures the appearance of 1d-1e samples was markedly different from those of 1a-1c ( Figure 2). The IR spectra of 1d and 1e films after heating at 500 • C showed the ZnO band, but its relative intensity was smaller than in the IR spectra of 1a-1c ones at identical temperature. At 600 • C, this absorption became the most intense one in the studied range. IR spectra of the films obtained at each annealing temperature ( Figure 3) sho the intensity of the bands due to specific organic units or bonds of the aminoalcoho C-H, C-C, or C-N, and C-O) decreased gradually as the temperature increased. 400 °C, the IR spectra of 1a-1c have a very weak band at 478 cm −1 that is character the crystalline ZnO [52]. In Figures 2 and 3 it can be seen that after annealing at 500 films 1a-1c exhibited the typical color of the ZnO and (b) the intensity of the band cm -1 increased gradually, while those related to the functional groups of the organ (mainly those between 1200 and 1700 cm −1 ) decreased. Moreover, annealing at caused the absorption band related to the ZnO bond to become the most intense o also undertook a parallel study with 1d and 1e films (Figure 3), and, as expected, tensity of the bands observed in the range 1600-1400 cm −1 (due to the carbonstretching vibrations of the phenyl ring [53]) decreased from 300 to 400 °C. It sho noted that at these temperatures the appearance of 1d-1e samples was markedly d from those of 1a-1c ( Figure 2). The IR spectra of 1d and 1e films after heating at showed the ZnO band, but its relative intensity was smaller than in the IR spectra 1c ones at identical temperature. At 600 °C, this absorption became the most inten in the studied range. Elemental analyses of the films 1a-1e obtained at each of the temperatures were also carried out and results are presented in Table 1. As expected, as the annealing temperature increased the contents of nitrogen, carbon, and hydrogen (N, C, and H, respectively) decreased gradually. For the aliphatic derivatives (1a-1c) at 500 • C the analytical determinations revealed contents of N and H below 1%, and the percentage of C was relatively low: 0.8% (for 1b), 2.94% (for 1a), and 3.6% (for 1c). Notice that for these samples the appearance as well as the IR studies were consistent with the presence of ZnO.
For the aromatic films 1d-1e, the C, H, and N contents also decrease with increasing temperature, as expected. For each annealing temperature, the amount of organic residues in aromatic films was higher than for the aliphatic analogues 1a-1c (see Table 1). Upon heating at 600 • C, the percentages of C, H, and N decreased in all cases, but the amount of carbon in the aromatic samples was clearly higher (0.71% for 1d and 3.49% for 1e) than in the aliphatic ones (0.24%, 0.27%, and 0.41% for 1a, 1b, and 1c, respectively). Besides, film 1e retained more nitrogen (0.43%) than 1a-1d (0.01%).  Elemental analyses of the films 1a-1e obtained at each of the temperatures were al carried out and results are presented in Table 1. As expected, as the annealing temperatu increased the contents of nitrogen, carbon, and hydrogen (N, C, and H, respectively) d creased gradually. For the aliphatic derivatives (1a-1c) at 500 °C the analytical determin tions revealed contents of N and H below 1%, and the percentage of C was relatively lo 0.8% (for 1b), 2.94% (for 1a), and 3.6% (for 1c). Notice that for these samples the appea ance as well as the IR studies were consistent with the presence of ZnO.
For the aromatic films 1d-1e, the C, H, and N contents also decrease with increasi temperature, as expected. For each annealing temperature, the amount of organic residu in aromatic films was higher than for the aliphatic analogues 1a-1c (see Table 1). Up heating at 600 °C, the percentages of C, H, and N decreased in all cases, but the amou of carbon in the aromatic samples was clearly higher (0.71% for 1d and 3.49% for 1e) th in the aliphatic ones (0.24%, 0.27%, and 0.41% for 1a, 1b, and 1c, respectively). Besid film 1e retained more nitrogen (0.43%) than 1a-1d (0.01%).
Since 1e can be visualized as derived from 1d by replacement the -OH unit (of 1d) the CH2OH one (in 1e), the comparison of the contents of the residual elements indicat that the presence of the pendant -CH2OH arm hinders the "elimination of the organ groups". All these findings suggest that the purity of the final ZnO films (1a-1e) obtain at 600 °C increase according to the sequence 1e < 1d < 1c < 1b ≤ 1a, indicating that und identical experimental conditions the final ZnO films formed from the aromatic AAs a specially from e retain more impurities. This could be related to the higher boiling poin of e and d compared with those of a-c (in the range 275-283 °C as compared to 170-186 °C). The presence of different amounts of residual C (in all cases) and N (in 1e) may important as to modify the number and type of defects and vacancies in the structur and, consequently, the properties and potential applications of these films [1,2,1 17,19,20,29].  Since 1e can be visualized as derived from 1d by replacement the -OH unit (of 1d) by the CH 2 OH one (in 1e), the comparison of the contents of the residual elements indicates that the presence of the pendant -CH 2 OH arm hinders the "elimination of the organic groups". All these findings suggest that the purity of the final ZnO films (1a-1e) obtained at 600 • C increase according to the sequence 1e < 1d < 1c < 1b ≤ 1a, indicating that under identical experimental conditions the final ZnO films formed from the aromatic AAs and specially from e retain more impurities. This could be related to the higher boiling points of e and d compared with those of a-c (in the range 275-283 • C as compared to 170-186.8 • C). The presence of different amounts of residual C (in all cases) and N (in 1e) may be important as to modify the number and type of defects and vacancies in the structures and, consequently, the properties and potential applications of these films [1,2,[15][16][17]19,20,29].

Structural Properties
In order to achieve further information about the formation of 1a-1e and to elucidate whether the nature of the AA (aliphatic vs. aromatic) could modify the path of the ZnO formation, films 1a-1e formed after the thermal treatments were studied by XRD ( Figures S1-S5). For comparison purposes, Figure 4 shows spectra of all films. In this section we will describe first the results obtained for the films derived from the aliphatic AAs (1a-1c) and afterwards those isolated from the aromatic ones (1d and 1c).    All observed peaks in the diffractograms of films 1a-1c after annealing at 400, 500, and 600 • C are indexed with the hexagonal, space group P63mc, ZnO wurtzite type form. ZnO hexagonal wurtzite type is the only crystalline observed phase in these samples. Figures S1-S3 depict the indexed observed diffractograms in the main angular range for 2θ from 25 to 75 degrees. It is well known that this hexagonal form is more stable than the cubic sphalerite (zinc-blende) one, and also than the rock salt type, which can only be obtained at high pressures [1,2].
For films 1a-1c and for each one of the temperatures selected in this study the Rietveld refinements enabled the accurate determination of the lattice parameter a and c and the estimation of the average crystallite size D and lattice strain (ε) (according to the double-Voigt approach [54]).
Data presented in Table 2 for the aliphatic-based films (1a-1c) show that upon heating the lattice parameter a increases, while c follows the opposite trend and the values obtained at 600 • C, as well as the c/a ratio fall within the ranges expected for the hexagonal wurtzite ZnO: 3.2475 Å ≤ a ≤ 3.2507 Å; 5.2042 Å ≤ c ≤ 5.2075 Å, and 1.593 ≤ c/a ≤ 1.603 [1,2]. The estimated crystallite sizes, D, at 600 • C (where ZnO conversion is accomplished) are larger than those at 400 • C while the lattice strain, ε, follows the opposite trend. These findings are indicative of fewer microdefects in the grains [55,56] and better quality [40] of films 1a-1c annealed at 600 • C. Table 2. Summary of the results obtained from Rietveld XRD studies of films 1a-1e after annealing at 400, 500, and 600 • C: lattice parameters (a and c (in Å)) and c/a ratio, ZnO phase present (% in weight), average crystal size (D in nm), and lattice strain (ε ×10 -3 ). When the XRD studies were performed with the films derived from the aromatic AAs 1d and 1e, some interesting and differential features were observed. First, the analysis of data obtained after the treatment at 400 • C revealed the presence of two ZnO forms. One of them was the thermodynamically stable wurtzite form (detected in the XRD of 1a-1c for all temperatures selected). The other one was identified, after indexing the remaining peaks, as the cubic, space group F-43m, sphalerite form ( Figures S4 and S5). The Rietveld refinements enabled the semi-quantification [57] of the ZnO phases, hexagonal wurtzite and cubic sphalerite. The relative abundance of the cubic phase decreased when the annealing temperature increased. In fact, analyses of XRD data obtained for films 1d after the heating treatments at highest temperatures (T = 500 or 600 • C) indicated the presence of the wurtzite form exclusively; while for 1e, the XRD studies revealed the coexistence of both phases and the presence of the cubic form even after the thermal treatment at 500 • C (Table 2 and Figure S6).
For these films 1d and 1e, the Rietveld refinements enabled also the cell parameters determination (a and c for wurtzite and a for sphalerite) and the estimation of the average crystallite sizes, D. The sphalerite type phase shows lower values of D than the wurtzite type phase (Table 2). At 600 • C, where only the wurtzite form remains, D is 30 nm for film 1e (similar to films 1a-1c) whereas for the 1d film is higher, 46 nm (Table 2).
In order to produce additional information on the evolution of the precursor to the final ZnO film, we prepared and analyzed film 1e at 300 • C and, in this case, XRD data indicated the presence of both crystalline forms, but now sphalerite was clearly the predominant phase (80.6% vs. 19.4% wurtzite) ( Figure 5). These findings suggest that for 1d and 1e films, the formation of the final wurtzite ZnO may proceed through sphalerite form.
It is widely accepted that the sphalerite form can be stabilized by growth on cubic substrates [1,2]. Nevertheless, to the best of our knowledge, no evidence for the presence of the cubic form or its coexistence with the hexagonal one (thermodynamically more stable) has been reported so far. XRD studies of films obtained from the aromatic AAs 1d and specially 1e reveal that the ZnO-sphalerite form was the major component at low annealing temperatures (i.e., 400 • C); while for the aliphatic-based ones (1a-1c), only the hexagonal wurtzite form was detected in all the cases, even for films obtained at the lowest annealing temperature, T = 300 • C.

Surface Morphology and Microstructure
In view of the effects produced by the AA nature (aromatic vs. aliphatic) during the annealing process and knowing the importance of the films' surface morphology in technological applications, SEM studies were also undertaken. The morphology of films 1a-1e annealed at 300, 400, 500, and 600 °C was visualized using both secondary and backscattered electrons, as shown in Figure 6. Moreover, the crystallite size in 1d after annealing at 600 • C is larger than that of 1e. These findings indicate that the presence of the -CH 2 -unit between the -OH unit and the phenyl ring (in 1e) is important as to modify the dimension of the grains and the microdefects therein.
XRD spectra of films annealed at 600 • C do not show any clear texture, with minor variation between samples. Only a slight preferential (00l) orientation is found for aliphaticderived films, especially 1b, which is even less visible for aromatic-based samples.

Surface Morphology and Microstructure
In view of the effects produced by the AA nature (aromatic vs. aliphatic) during the annealing process and knowing the importance of the films' surface morphology in technological applications, SEM studies were also undertaken. The morphology of films 1a-1e annealed at 300, 400, 500, and 600 • C was visualized using both secondary and backscattered electrons, as shown in Figure 6.

Surface Morphology and Microstructure
In view of the effects produced by the AA nature (aromatic vs. aliphatic) during the annealing process and knowing the importance of the films' surface morphology in technological applications, SEM studies were also undertaken. The morphology of films 1a-1e annealed at 300, 400, 500, and 600 °C was visualized using both secondary and backscattered electrons, as shown in Figure 6.
(a) Backscattered electrons (Figure 6b) produce small contrast in all the analysed films, indicating continuous layers with chemical uniformity at the microscopic scale for every sample. Only films 1d exhibit a different behaviour, being cracked even at annealing temperatures as low as 300 °C. Although a very thin film remains adhered to the substrate, Backscattered electrons (Figure 6b) produce small contrast in all the analysed films, indicating continuous layers with chemical uniformity at the microscopic scale for every sample. Only films 1d exhibit a different behaviour, being cracked even at annealing temperatures as low as 300 • C. Although a very thin film remains adhered to the substrate, producing a very transparent coverage (see Figure 2), the poor stability induced by AA d produces films difficult to be controlled, unsuitable for most applications.
Secondary electron imaging (Figure 6a,c) indicated that ZnO precursors a-c and e promote films with fractal topography depending on the AA and on the thermal treatment. In particular, annealing at low temperature produce wrinkled layers with morphologies that appear as neural nets, with ridges up to 14 µm thick (sample 1a) or porous structures (sample 1e). Annealing at 400 • C tends to increase the number of large features at the expense of suppressing the smallest ones, while treatment at 500 • C gives rise to films with more regular patterns and smaller ridges ranging from 0.8 (1e) to 6 µm (1a). For higher temperature, this trend seems to be inverted, as annealing at 600 • C promotes the formation of neuron-like prominences that remind us of those observed for low temperatures. A completely different behavior can be observed for sample 1d, which produced cracked films with smooth surfaces at every temperature. Adhesion is lost from the lowest temperatures, and very thin films remain in all cases, allowing excellent transparency (Figure 2).
A closer look allows ascertaining that those films consist on inhomogeneous, dense, agglomerated particles that accumulate preferentially in the ridges. The thickness of the layers differs depending on the region due to these large fractal structures formed by crystallites that group into larger grains, as reported for MEA [44]. The overall thickness of the layers ranges between 1 (continuous bottom layer) and 5 µm (including protrusions). At 600 • C, larger grains are observed, which can be attributed to a coalescence process that change the film topography, producing less uniform layers with thicker protrusions with diameters between 1 µm (1c) and 4 µm (1b). In contrast, for lower temperatures, the prominences seem formed by dark-coloured amorphous and nanocrystalline agglomerations highly contaminated with organic residues.
The images clearly show the different structures promoted by the aliphatic and aromatic AAs used. At low annealing temperatures, at which samples have a high concentration of impurities, the relief observed for films 1e includes rather frayed fibres, very different from the more uniform yarns visible for films 1a to 1c. This difference might be related to the presence of a noticeable amount of sphalerite phase in films 1e annealed at temperatures below 600 • C [58], which disappears at this temperature. The different behaviour is still much more evident for films 1d, where annealing produces quartered layers that at high temperature leave a very thin transparent film. The instability found for films 1d prevents its use for producing good-quality ZnO films, and, therefore, it will be disregarded in the optical characterizations that follow.

UV-Visible (UV-VIS) Spectroscopy
Due to the differences observed in the formation processes, morphologies, and compositions of the ZnO thin films, and the increasing interest in these kinds of materials, a comparative study of the optical properties of the films 1a-1e was undertaken. These properties are related to intrinsic characteristics of the films, such as absorption coefficient (α), optical conductivity, bandgap energy (E g ), extinction coefficient (k), and refractive index (n). Films obtained at 300 • C were not included in this study due to the following reasons: (a) in their IR spectra, the typical band due to ZnO at ca. 478 cm −1 -that is characteristic of the crystalline ZnO [52]-was, when detectable, extremely weak (see Figure 3) and (b) the analytical determinations revealed the presence of large amounts of C, H, and N (especially for the aromatic-based films with a total content around 38% for 1e, or even higher for 1d (ca. 52%)), thus suggesting that films processed at 300 • C contained a small amount of ZnO and large quantities of organic fragments. In all cases, spectra were recorded in triplicate in order to ensure the repeatability of the results.
Films derived from the aliphatic AAs 1a-1c show quite similar UV-VIS behavior (Figure 7), reaching high visible-range transparency as temperature increases. Spectra of films obtained at 400 and 500 • C (which, according to the elemental analyses, still retained small quantities of C, H, and N-residues, see Table 1) exhibited a broad absorption with low resolution below 370 nm. In contrast, spectra of films annealed at 600 • C exhibit an absorption band clearly visible in the range of 355-370 nm (Figure 7). Crystalline ZnO films typically show an absorption band at around 359 nm usually attributed to free excitons [59] that shifts toward a lower wavelength due to quantum confinement.
[59] that shifts toward a lower wavelength due to quantum confinement.
As seen in Figure 7, the UV-VIS spectra of films 1e do not follow the same trend and the absorption band at around 359 nm is not clearly identifiable even after annealing at 600 °C. As explained above, films 1d are quartered by annealing and the very thin films that remain produce very small overall absorption. This is why spectra for films 1d are not included in Figure 7. Since it is well known that the bandgap energy (Eg) is a crucial parameter to predict the electronic behavior of semiconductor films, in the next step we obtained the absorption coefficient dependence to estimate the optical bandgap energy of the films obtained at 400, 500, and 600 °C using Tauc's equation for direct allowed bandgap (1) [60,61] (α hν) 2 where α is the absorption coefficient, hν is the incident photon energy, and B is a constant. The optical band gap was then estimated by plotting (αhν) 2 against the photon energy (hν) and extrapolating the linear portion of the curve to the photon energy axis. Tauc's plots of films (1a-1c and 1e) are presented in Figures S7-S10. The energy bandgap (Eg) value presented in Table 3 is obtained from 3 independent spectra for each film at each annealing temperature, and assuming uniform reflection (around 20%), as usually measured for this kind of polycrystalline films. For illustrative purposes, Tauc's plots of the films annealed at 600 °C and the graphical summary of the obtained bandgaps are presented in Figure 8. As seen in Figure 7, the UV-VIS spectra of films 1e do not follow the same trend and the absorption band at around 359 nm is not clearly identifiable even after annealing at 600 • C. As explained above, films 1d are quartered by annealing and the very thin films that remain produce very small overall absorption. This is why spectra for films 1d are not included in Figure 7.
Since it is well known that the bandgap energy (E g ) is a crucial parameter to predict the electronic behavior of semiconductor films, in the next step we obtained the absorption coefficient dependence to estimate the optical bandgap energy of the films obtained at 400, 500, and 600 • C using Tauc's equation for direct allowed bandgap (1) [60,61] (α hν) 2 where α is the absorption coefficient, hν is the incident photon energy, and B is a constant. The optical band gap was then estimated by plotting (αhν) 2 against the photon energy (hν) and extrapolating the linear portion of the curve to the photon energy axis. Tauc's plots of films (1a-1c and 1e) are presented in Figures S7-S10. The energy bandgap (E g ) value presented in Table 3 is obtained from 3 independent spectra for each film at each annealing temperature, and assuming uniform reflection (around 20%), as usually measured for this kind of polycrystalline films. For illustrative purposes, Tauc's plots of the films annealed at 600 • C and the graphical summary of the obtained bandgaps are presented in Figure 8.  All the aliphatic films present rather similar behaviors, progressing in a relatively reduced bandgap interval (3.17-3.28 eV) and reaching values between 3.22-3.26 eV when the annealing temperature increases to 600 °C. The estimated standard deviation, σ, is ±0.01 eV, and the differences in the Eg values for each sample do not clearly exceed 3σ. Thus, correlation between Eg and temperature for films 1a-1c is not possible. These values are smaller than Eg for pure ZnO reported (around 3.37 eV [1][2][3][4]). According to the literature [30], this could be related to the presence of contaminants that introduce shadow pdoping levels into de bandgap [62] or intrinsic defects, such as Zn or O vacancies or interstitials [30]. Films 1e have Eg clearly smaller than those found for 1a-1c under identical conditions. This could be ascribed to different causes such as the higher content of C, H, and N impurities of films 1e compared with 1a-1c and the presence of the cubic crystalline form. It is well known that bandgap energy for the sphalerite is smaller (3.18 eV) than the one for wurtzite (3.29 eV) phase.
Another useful information that can be extracted from these spectra is related to the structural disorder in thin films. In particular, the tail of the absorption edge indicates the presence of localized states in the bandgap as given by (2) [63,64], where α0 is a constant and EU is Urbach's energy, which is the width of the tail of the All the aliphatic films present rather similar behaviors, progressing in a relatively reduced bandgap interval (3.17-3.28 eV) and reaching values between 3.22-3.26 eV when the annealing temperature increases to 600 • C. The estimated standard deviation, σ, is ±0.01 eV, and the differences in the E g values for each sample do not clearly exceed 3σ. Thus, correlation between E g and temperature for films 1a-1c is not possible. These values are smaller than E g for pure ZnO reported (around 3.37 eV [1][2][3][4]). According to the literature [30], this could be related to the presence of contaminants that introduce shadow p-doping levels into de bandgap [62] or intrinsic defects, such as Zn or O vacancies or interstitials [30]. Films 1e have E g clearly smaller than those found for 1a-1c under identical conditions. This could be ascribed to different causes such as the higher content of C, H, and N impurities of films 1e compared with 1a-1c and the presence of the cubic crystalline form. It is well known that bandgap energy for the sphalerite is smaller (3.18 eV) than the one for wurtzite (3.29 eV) phase.
Another useful information that can be extracted from these spectra is related to the structural disorder in thin films. In particular, the tail of the absorption edge indicates the presence of localized states in the bandgap as given by (2) [63,64], where α 0 is a constant and E U is Urbach's energy, which is the width of the tail of the localized states corresponding to the optical transition between localized states adjacent to the valence band and extended state in the conduction band which is lying above the mobility edge [64]. Although the poor thickness uniformity observed in all the obtained films lowers the precision in its determination, the values of the Urbach's energies have been estimated as the reciprocal of the slopes of the linear portion in the lower photon energy region of ln(α) vs. hν (Table 3). For aliphatic-derived films, their values are between 0.65 (films 1b annealed at low temperature) and 0.22 eV (films 1a annealed at high temperature). E U shows a decreasing trend with increasing T, suggesting less localized states at higher annealing temperatures. The same trend is observed for the aromatic-derived films (1e) but with considerably higher E U values, only approaching to the previous range after annealing at the highest temperature (600 • C).

Photoluminescence (PL)
The photoluminescence spectra obtained for films annealed at 600 • C are presented in Figure 9 (except for sample 1d, owing to its low quality). For comparison purposes, spectra corresponding to the samples annealed at 300 • C are also included.
x FOR PEER REVIEW 16 of 21 films (1e) but with considerably higher EU values, only approaching to the previous range after annealing at the highest temperature (600 °C).

Photoluminescence (PL)
The photoluminescence spectra obtained for films annealed at 600 °C are presented in Figure 9 (except for sample 1d, owing to its low quality). For comparison purposes, spectra corresponding to the samples annealed at 300 °C are also included. For the films annealed at 600 °C, two emission bands can be distinguished, one band in the range 370-420 nm and the other band, with higher intensity, appearing at lower energies extending roughly from 550 to 750 nm. The PL spectra of samples annealed at 300 °C are noisier and clearly different from those at 600 °C, probably owing to the low content of ZnO in the films at this temperature, in good agreement with the literature [65].
The near-band edge (NBE) for UV emission is present in all samples in the low wavelength range, regardless of the annealing temperature. The energy corresponding to this peak can be found in the range 2.95-3.35 eV and seems to increase with the annealing temperature, which matches well with the bandgap energies obtained from UV-VIS spectroscopy. This emission is attributed to excitonic recombination, a process where an elec- For the films annealed at 600 • C, two emission bands can be distinguished, one band in the range 370-420 nm and the other band, with higher intensity, appearing at lower energies extending roughly from 550 to 750 nm. The PL spectra of samples annealed at 300 • C are noisier and clearly different from those at 600 • C, probably owing to the low content of ZnO in the films at this temperature, in good agreement with the literature [65].
The near-band edge (NBE) for UV emission is present in all samples in the low wavelength range, regardless of the annealing temperature. The energy corresponding to this peak can be found in the range 2.95-3.35 eV and seems to increase with the annealing temperature, which matches well with the bandgap energies obtained from UV-VIS spectroscopy. This emission is attributed to excitonic recombination, a process where an electron in the conduction band is recombined with a hole in the valence band. Due to its crystalline structure, wurtzite has three exciton states, at 362, 366, and 367 nm (3.42, 3.39 and 3.38 eV, respectively) [30,31]. The UV emission of our samples at 600 • C (Figure 9 and Table S1) is slightly red shifted (ca. 18 nm), indicating a bandgap narrowing probably due to a high defect level and sub-stoichiometry of materials. Besides, we can see that film 1e has the lowest NBE PL intensity at 600 • C, probably because it retains more C and N than films 1a-1c obtained at identical temperature. Furthermore, we can conclude that the increase in annealing temperature enhances the visible photoluminescence.
The broadening of the band in the visible region might be due to the overlap of several emission peaks close to one another. These deep-level emission bands (DLE) are usually attributed to defects of the ZnO material, such as zinc or oxygen vacancies, interstitials, or anti-sites, among others. Although there is no clear consensus on their origin yet, these emissions are reported to depend on the synthesis procedure, morphology, impurities, vacancies, and surface defects. Typically, green-blue luminescence has been attributed to recombination of simple ionized oxygen vacancies [9], very commonly observed in oxygendeficient ZnO. It is also thought that oxygen-interstitial related defects are responsible for orange-red luminescence. Moreover, some studies have proved that the addition of carbon increases the oxygen vacancies in ZnO by the reduction reaction [10]. Emissions at 400 ≤ λ ≤ 500 nm are usually associated to defects involving zinc [66]. In the PL spectra of our films annealed at 600 • C (Figure 9, left), the main emission occurs at higher wavelengths, suggesting low relevance, if any, of this sort of defect. Therefore, the broad DLE bands observed should be attributed to structural defects of different typology which are relevant in view of their potential interest in specific areas, e.g., as gas sensors [67]. For instance, the presence of neutral, single-, and double-charge oxygen vacancies or interstitials (V O , V O + , V O ++ , and O i ) produce peaks in the ranges 520-570, 570-620, 620-670, and 670-780 nm, respectively, all of them within the 550-750 nm region, where the broad DLE emission bands of our films annealed at 600 • C were observed. These kinds of structural defects are reported to be strongly dependent on the calcination temperature. Since our films were obtained under identical experimental conditions and calcinated at 600 • C, they are well suited to elucidate whether the AA used in the synthesis of the ink could also induce significant variations in the number and type of these defects in the produced film.
In order to clarify the main origin of the DLE bands visible in Figure 9 (left), they were deconvoluted. The calculated wavelet positions and relative contributions of each component are summarized in Table 4. The results obtained show that one (for 1a and 1c), two (for 1e), or even three (for 1b) components have low contributions (<5%), and, therefore, won't be included in the following discussion. Comparison of data shows that transitions from the conduction band to O i levels are more favored in the films prepared using an aromatic AA (1e, with a total weight of ca. 75%) than in the aliphatic-derived films 1a-1c, for which the relative weight was clearly smaller (ca. 37% for 1a and below 50% for 1b and 1c). This finding suggests that in addition to the changes induced by the use of the aromatic AA e on the properties of films 1e described in the previous sections, it also induces the preferential formation of O i defects in the ZnO film; while for the aliphaticbased films 1a and 1c, transitions from the conduction band to deep-level oxygen vacancies (V O , V O + , and V O ++ ) are predominant (total contributions varying from 61% for 1a to 48% for 1c). Table 4. Wavelet position, full width at half maximum (FWHM), area, relative weight, and possible assignment of each DLE component (according to [63]) obtained after Gaussian deconvolution for 1a-1c and 1e films annealed at 600 • C.

Conclusions
In this work we have prepared and compared the main properties of ZnO-thin films (1a-1e) obtained by the sol-gel methodology, using ZAD as the Zn(II) source, the aminoalcohols (a-e) as stabilizers, and 2-methoxyethanol as solvent, identical molar ratios, reaction period, and deposition method, but different annealing temperatures (T = 300, 400, 500, and 600 • C). The results summarized here provide conclusive evidence of the effect produced by the nature of the aminoalcohols (aliphatic (a-c) or the phenyl-containing analogues (d and e)) in the ZnO formation process as well as on the physical appearance, stability, purity, morphology, and properties of the films (1a-1e) at each of the selected temperatures.
In general terms, films (1a-1c) obtained using the aliphatic AAs exhibit a similar behavior, stability, morphology, composition, and properties, but some of them are markedly different from those of their aromatic partners 1d and 1e, processed identically. For instance, films (1a-1c) are clearly more stable than 1e and, especially than 1d, which degrades and crashes quickly even at low annealing temperatures, limiting its potential interest. Moreover, the cubic and hexagonal forms of ZnO coexist, in different molar rations (depending on the temperature), in the films prepared from the aromatic aminoalcohols (d and e); while evidence of the presence of the cubic form was not detected in any of the aliphatic films 1a-1c. This finding suggests that the aromatic AAs induce the preferential formation of the sphalerite form at low annealing temperatures, that transforms gradually into the wurtzite ZnO phase as the calcination temperature increases. Moreover, the energy band gap (E g ) and the Urbach's energy (E U ) of the aliphatic-based films are higher and smaller, respectively, than the value obtained for 1e under identical conditions. Our films exhibit intense emission bands in the visible region, but the analyses of spectral data suggest that the main contribution to the emission spectra is due to transitions from the conduction band to the oxygen-interstitial levels (for 1e) or to the deep level-oxygen vacancies for 1a-1c. This indicates that the characteristics and properties of the AA are also important as to modify the type of structural defects and the main origin of the emissive properties of the ZnO films. It should be noted that the results obtained for film 1e at 600 • C open up a new path to explore other aminoalcohols closely related to aminobenzyl alcohol. The incorporation of substituents with different electron donor/electron acceptor properties on different sites of the phenyl ring may allow to control or tune the bandgap energy and other relevant parameters, as the type and predominance of structural defects present in the final ZnO films, that are important in view of their potential utility to achieve new technological devices.