1 Introduction

The sol–gel process is a well-known route for the preparation of hybrid organic–inorganic materials, e.g., coatings. Such coatings can modify or completely change the properties of substrate. Main reason of the sol–gel process popularity is its simplicity and convenience. It enables the preparation of the glassy matter in the laboratory conditions without the need of utilizing special equipment. However, there are still some limitations, such as multistep-synthesis or the need for final high-temperature curing.

To prepare sol–gel coating with various functions, the default precursor functionalization is used often. As functionalization, organic compounds are used. In this case, a hybrid inorganic–organic material is prepared [1,2,3]. Hybrid materials can be classified into three main groups, regarding the type of bonding in their structure [1, 2, 4, 5].

The first type is characterized by the presence of weak bonding interactions (hydrogen bonds, van der Waals bonds, or electrostatic bonds) between inorganic and organic structural units. Materials of this kind can be prepared in various ways. During the polycondensation step of the network forming precursor in sol–gel synthesis, organic substances can be dispersed in sol and then retained in structure within the gelation phase. The hybrid of group 1 can be also prepared in the xerogel phase of a sol–gel system. The porous inorganic structure is immersed into a solution with organic substance, thereby adsorbed into pores. After the evaporation of a solvent, the organic substance is entrapped and a hybrid structure arises. Another possibility is a simple preparation of two independent inorganic and organic structures. Hybrids of the first type are often used for immobilization of biomolecules or dyes. A well-known example is one of the oldest dyes, Maya blue. This dye was prepared by incorporation of indigo molecules into palygorskite mineral [1, 2, 4, 6, 7].

The second type of hybrid materials is distinguished by the covalent bonds. The preparation manner is alike as conventional sol–gel synthesis, although it has specific requirements for modifying organic compounds. Organic precursors have to contain hydrolyzable inorganic functional group, or organic functionalizations can be introduced as substituted alkoxysilanes. Meeting this requirement permits hydrolysis and polycondensation reactions, resulting in a covalently bonded inorganic–organic network. Thanks to organic functionalizations, the possibilities of material properties alterations broaden immensely. Hybrids of the second type are mostly utilized as functional coatings with various properties [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18].

The third type combines covalent bonding and weak bonding interactions, more precisely, the organic part is covalently bonded to the inorganic part on several sites, and concurrently weak bonding interactions significantly influence the organic part within the whole structure. Hybrids of the third type are materials frequently consisting of network-forming precursors and e.g., silylated dyes [4]—significantly big chromophore molecules may contain hydrogen bond acceptors while the silylated part is covalently anchored into (inorganic) network. The significant advantage of preparing dyed material as a third type hybrid is in increased resistance against washing out the dye [1, 4].

As it was stated above, for functional coatings the second type of hybrids is commonly used. Default alkoxysilanes allow the formation of an inorganic silica network (eg. tetraethoxysilane, tetramethoxysilane), functionalizational alkoxysilanes (selected ones are listed in Table 1 [2, 7, 19]) bring organic functional groups, thus modifyng materials properties. Depending on the nature of organic groups, even those precursors can be viewed as network-forming. This classification is permitted only if the precursors contain polymerizable bonds. Polymerization on those bonds allows secondary organic network to form [2, 4, 5, 19].

Table 1 Selected functionalization precursors and their effects in sol–gel coatings
Full size table

Generally, functionalizational precursor structure can be described as (RO)3–Si–Y, where RO represents hydrolyzable group (sol–gel mechanism) and Y represents arbitrary organic group. The group Y can be even chemically reactive (thio-, vinyl-, amino-, etc.) or inert (alkyl groups) [19]. Immense diversity in organic substituents gives still possibilities to produce new materials with extraordinary properties with only limitation, which is appropriate organic derivatives availability and their reactivity [2, 19].

However, many types of functionalized coatings are frequently synthesized in a multi-step process and such preparations often require treatment with high temperatures. Those aspects considerably contribute to economic and technological demandingness, which makes the sol–gel coatings utilization more difficult. With these assumptions, the main objective of this research was to prepare sol–gel derived coating on glass, which enhances the substrate glass corrosion resistance. To allow practical use of the newly developed coating, it must also exert sufficient mechanical resistance and must not impair substrate’s optical properties. To meet the aforementioned assumptions, the sol–gel synthesis was designed as simple as possible, without the need of high-temperature processing.

2 Methods

The coatings were intended for being applied on glass. For that purpose, the conventional microscopic glass slides were used. Chemical composition of the substrate glass was determined via the X-ray fluorescence analysis and the results are listed in Table 2.

Table 2 Composition of substrate glass (XRF), main components [wt. %]
Full size table

2.1 Sample preparation

Every sol was prepared in the polyethylene flask with the screw cap. The preparation of sols consisted of dissolving the precursors (TEOS—tetraethoxysilane and MEMO—3-(trimethoxysilyl)propyl methacrylate or IBTMS—isobutyl(trimethoxy)silane) successively in isopropyl alcohol under continuous stirring at ~230 RPM on a magnetic stirrer. All used chemicals are listed in Table 3. After short homogenization, the demineralized water in composition-relevant quantity was added and the mixture was left to stir for 10 min. The homogenized mixture was acidified with 2–3 drops of hydrochloric acid to pH ~2. The pH value was verified with Litmus paper. After that, the polyethylene flask was closed with the screw cap and sealed with a strip of Parafilm. Prepared sols were left to polycondensate at laboratory temperature for 4 h under continuous stirring. Thereafter, the sol was ready to be applied on glass. Compositions of all tested sols are summarized in Table 4.

Table 3 List of used chemicals
Full size table
Table 4 Composition of sols
Full size table

Prior to the sol application, the microscopic slides were cleaned by means of the polyurethane sponge and conventional household detergent. The detergent residue was rinsed with demineralized water thoroughly. Finally, to remove any remaining surface contamination, the microscopic slides were rinsed with acetone. Cleaned glass substrate was left to air dry afterwards.

Once ready, the sols were applied on cleaned substrate glass via dip-coating method using withdrawal speeds ranging from 5 cm/min to 15 cm/min. For this purpose, a custom-made dip-coating device was utilized. The dip-coating device is equipped with a precise drive, which enables accurate control of the dipping and withdrawal speed. Between the dipping and withdrawal steps, the 30-second delay was set. Coated samples were dried in a drying oven, at 200 °C for 1 h.

2.2 Evaluation methods

The dynamic viscosity of sols was evaluated by the means of vibrational viscometry. Measurement was carried out with A&D Vibro Viscometer SV-10 at 30 Hz.

Appearance of the coatings was assessed visually and with optical microscopy. For this purpose, optical microscope Olympus BX 51was used. The images were captured with digital camera PROMICRA PROMICAM 3-5CP. Subsequent image processing was done in QuickPHOTO CAMERA 3.2 software.

To investigate adhesion of the coating to the glass substrate, cross-cut tape test, defined by standard ISO 2409, was done. The cross-cut tape test consists of cutting the coating in a grid pattern and subsequent sticking of special tape on the grid. After certain time, the tape is peeled off in a defined manner. Results of the test are evaluated with the optical microscope. Rate of the coating’s delamination is observed.

To obtain the light transmittance, the UV–Vis spectrophotometer Shimadzu UV-2450 was used. The transmittance spectra were obtained from a wavelength range of 400–800 nm with sampling interval 0.5 nm. Results were processed with the UVProbe 2.43 software.

Chemical resistance of coated samples was observed with accelerated weathering test (80% RH, 60 °C, 110 days) and by interaction with solutions (pH 2–13, ambient temperature, 7 days). Results were evaluated visually, by means of optical microscopy and UV-Vis spectrophotometry.

Hardness of prepared coatings was assessed qualitatively, according to standard ASTM D3363. This method consists of scratching the coating with graphite pencils with decreasing hardness—the hardness of the coating is attributed to the first pencil, which leaves the coating unaffected.

Wear resistance of prepared coatings was proved qualitatively, by a model test of wiping the coatings with loaded abrasive cloth. In works [20, 21], the fabric made of a cellulose-polyester mixture, loaded with 20 g weight was used for this test. On the other hand, authors in [16] used more abrasive material, 3 M Scotch-Brite™. The testing cycle consisted in drawing the loaded fabric (in straight direction) towards and backwards. By this manner, abrasive wear was generated. The rate of abrasive wear correlated with number of cycles was observed via the optical microscopy.

Lastly, the approximate coating thickness was assessed by the analysis of scratches in the coating via atomic force microscopy.

3 Results and discussion

3.1 Viscosity of sols

The dynamic viscosity (η) was measured at laboratory temperature (t), immediately after the sol preparation and after the ageing (5 and 14 days), the measured values are presented in Table 5. Apparently, the values oscillate around 2 mPa s, even with increasing ageing time the values remain comparable, differing just in decimals. Main reason for this behavior could lie in sol compositions. Composition is designed for dip-coating processing, therefore the sols must be flowing enough.

Table 5 Viscosity of sols after various ageing times.
Full size table

3.2 Appearance, adhesion, and mechanical properties

All the prepared coatings appeared visually transparent and homogenous. Observation by optical microscopy divulged that the coatings were faultless and that they appeared all the same, independent of composition or withdrawal speed used for their production.

Adhesion of the coating to the glass substrate, evaluated in accordance with ISO 2409, was classified as class 0, the highest adhesion obtainable. This result was independent of composition, ageing time of sols and withdrawal speed during dip-coating method. High adhesion in the system sol–gel coating-glass is quite common. It is most likely caused by nature of used chemicals and the fact, that coating is applied on glass. The glass surface contain high amount of –OH groups, which can react with –OH groups of sol–gel siloxane stuctures or with hydrolyzable bonds, resulting in covalent bonding between coating and glass.

Pencil hardness test of prepared coatings was carried out with graphite pencils of hardness scaling from HB (soft) to 6B (hard). The results are summarized in Table 6. Pencils, which left coating scratched, are marked with check symbol (✔), the unmarked ones left coating unaffected. First unmarked pencil hardness (from the right) is assigned as the coating hardness. Highest hardness was observed at composition ME1 (4H), a bit softer was ME1-IB1 (2H). Moreover, the test proved that there was not any effect of sol ageing on hardness of coating. Composition ME1-IB2 was the softest of the tested samples. A trend is clearly visible—by increasing IBTMS content the hardness of produced coatings grows smaller. The source of hardness in produced hybrid inorganic–organic coatings is presumably functionalizing agent MEMO. The initial inorganic structure, emerging from subsequent hydrolysis and polycondensation reactions occurring between molecules of default precursor TEOS, is altered upon functionalization with MEMO. The MEMO molecule contains hydrolysable sites, which enables it to participate in hydrolyses and polycondensations, forming a siloxane network enriched with methacrylate organic groups in statistical deployment. Although methacrylate groups cannot be subjected to hydrolysis, they can undergo radical polymerization, producing a crosslinked, “secondary” organic structure. Radical polymerization of methacrylate moieties is often catalyzed by light irradiation or by radical-providing compound addition. The formation of such organic structure can lead to reinforcement of the inorganic ones [22]. With the increasing addition of IBTMS, the organic structure can be increasingly modified with isobutyl groups, which can act as a spacer molecule and thus mitigate the reinforcing effect of polymerized methacrylate.

Table 6 Results of pencil hardness test
Full size table

Wear resistance test was realized in modified manner, otherwise described in [20, 21]. As abrasive material, the coarse part of a dishwashing sponge was used (3M Scotch-Brite™-like material). The abrasive material was loaded with 500 g weigh. The wear resistance was evaluated as the number of wiping cycles required to visually damage the samples. Results revealed, that the coating of composition ME1 can withstand more than 100 wiping cycles (damage was only microscopical), the coating ME1-IB1 was visually affected after 80 wiping cycles, the coating ME1-IB2 was damaged just after 10 cycles. The results were overall similar for coatings prepared from sols aged 5 and 14 days. Wear resistance test results imply a link between hardness and wear resistance of the coatings. The sufficiently hard coatings can withstand abrasive wear and vice versa.

Generally, mechanical properties are affected by the composition of MEMO functionalization to other functionalization ratio.

Coatings thickness approximately measured via the AFM measurement seemed to be dependent only on withdrawal speed during the dip-coating process. Coatings prepared with withdrawal speed 5 cm/min attained roughly 100 nm in thickness, coatings prepared with withdrawal speed 15 cm/min attained approximately 200 nm in thickness. According to equations described in literature [23], the coating thickness is withdrawal speed and viscosity dependant. More precisely, for low withdrawal speeds and low viscosities the coating thickness is given by Landau-Levich equation:

$$h = \frac{{0,94\left( {\eta U} \right)^{2/3}}}{{\gamma _{{{{\mathrm{lg}}}}}^{1/6} \cdot \left( {\rho g} \right)^{1/2}}}$$

where η is dynamic viscosity, U is withdrawal speed, γlg is liquid-gas phase interface tension, ρ is density of the sol and g is gravitational acceleration. It can be seen from the equation, that coating thickness of the wet coating is proportional to viscosity and withdrawal speed to the power of 2/3. Additionally, viscosity measurements suggest that viscosity of prepared sols did not change with composition or ageing time a lot, so thickness could be mainly affected only just by withdrawal speed. It must be taken into account the final coating thickness could be affected with many variables during subsequent processing [23].

3.3 Transmittance measurements

Transmittance of samples was measured on differently prepared coatings to investigate the effect of withdrawal speed, composition, and 5-day ageing of the sol.

3.3.1 Effect of sol composition

The graph summarizing the effect of sol composition on transmittance (Fig. 1) shows that composition ME1, ME1-IB1, and ME1-IB2 exhibit the best results in terms of transmittance — their transmittance is visibly increased through the whole analyzed spectrum. Generally, the highest attained transmittance belongs to the composition ME1-IB2 (up to 95% at 550 nm). All samples used in this measurement were prepared with withdrawal speed 5 cm/min.

Fig. 1
figure 1

Transmittance of samples with differing composition

Full size image

3.3.2 Effect of withdrawal speed

Transmittance of sample ME1 (Fig. 2) showed its maxima around 94% for both withdrawal speeds. On the contrary, higher withdrawal speed produced coating which transmittance proceeds also through local minimum in the analyzed spectrum (about 92% at the spectrum margin). Sample ME3 (Fig. 3) exhibits similar behavior. Coatings prepared with withdrawal speed 5 cm/min have better transmittance in the major part of the spectrum (maximum around 94% at 600 nm) compared to coatings prepared with withdrawal speed 15 cm/min.

Fig. 2
figure 2

Transmittance of samples ME1, prepared using various withdrawal speeds

Full size image
Fig. 3
figure 3

Transmittance of samples ME3, prepared using various withdrawal speeds

Full size image

The highest transmittance value of sample ME1-IB1 (5 cm/min) was around 94% at 500 nm (Fig. 4). Overall transmittance did not lower under 93%. Increasing withdrawal speed (15 cm/min) led to the production of coating with a similar maximum (93.8% around 750 nm), but towards smaller wavelengths transmittance decreased considerably. Samples with higher content of IBTMS (ME1-IB2) performed similarly to ME1-IB1 (Fig. 5): lower withdrawal speed produced coating with peak transmittance of 95% (550 nm), coating prepared with higher withdrawal speed had local transmittance maxima on the spectrum borders (max. 94.5%), while in the middle of the spectrum its transmittance proceeds through minimum (around 92% at 550 nm).

Fig. 4
figure 4

Transmittance of samples ME1-IB1, prepared using various withdrawal speeds

Full size image
Fig. 5
figure 5

Transmittance of samples ME1-IB2, prepared using various withdrawal speeds

Full size image

Another increase in IBTMS content (ME1-IB4) caused further transmittance changes (Fig. 6). No significant increase in transmittance maximum at the sample prepared with withdrawal speed 5 cm/min was observed. Moreover, transmittance at the edge of the spectrum grew smaller to 92%. Maximal transmittance of sample prepared at 15 cm/min increased to 95% (around 500 nm), followed by decline to roughly 92%.

Fig. 6
figure 6

Transmittance of samples ME1-IB4, prepared using various withdrawal speeds

Full size image

Increased amount of MEMO functionalization together with lower IBTMS content (ME3-IB1) gives comparable results as composition ME1-IB4 (Fig. 7). Progress of transmittance through the spectrum remains similar, only maxima decreased slightly — in both cases to around 94%.

Fig. 7
figure 7

Transmittance of samples ME3-IB1, prepared using various withdrawal speeds

Full size image

As can be seen from the graphs (Figs. 2–7), there is a visible trend in transmittance performance: using lower withdrawal speeds leads to the production of thinner coatings. This conclusion is in good agreement with theoretical relations and with AFM observations. Since transmittance depends on the thickness of an optical environment, this is a crucial parameter for coatings fabrication. In terms of transmittance maximization, thickness of the coating with refractive index n should be equal to one-fourth of incident monochromatic light’s wavelength (\(d = \frac{\lambda }{{4n}}\)) [24]. By performing an analysis in the spectrum of wavelengths, changes in transmittance values are observed — as the wavelength changes, the coating thickness may or may not comply “the one-fourth condition”, shifting the resulting transmittance significantly. Thus, coatings prepared with higher withdrawal speeds (thicker coatings) exhibit different wavelength dependence, often with local minima in parts of the analyzed spectrum, than coatings prepared with lower withdrawal speeds. Among other influences, major impact on the transmittance performance may be founded in uneven coating thickness, which is a common dip-coating derived coatings defect. For all tested compositions, in terms of transmittance, 5 cm/min withdrawal speed seemed to be the most promising. Samples prepared in this manner exhibited transmittance without such inflections, as it was at samples prepared with 15 cm/min withdrawal speeds, and attained higher transmittance peaks.

3.3.3 Effect of sol ageing

Ageing of sols for 5 days caused significant transmittance increase only at the composition ME1-IB1 (Fig. 8). The difference between samples prepared from non-aged and aged sol exceeded 0.5% (maximal transmittance was around 94.5%). Other samples did not exhibit any significant improvement in transmittance. All samples used in this measurement were prepared with withdrawal speed 5 cm/min.

Fig. 8
figure 8

Transmittance of samples prepared from 5 days aged sols

Full size image

3.4 Chemical resistance test — accelerated weathering

After 110 days of exposing the samples to humid and hot atmosphere (80% RH, 60 °C), the vast majority of samples was still visually transparent. Only observable visual changes were on the edges of certain samples, especially coatings ME3 and ME3-IB1 exhibited increased visual dullness. Optical microscopy of those samples revealed noticeable signs of corrosion. On the contrary, untreated glass substrate showed distinguishable signs of corrosion just after 40 days, after 110 days of accelerated weathering was its surface strongly corroded. The reason for such pronounced appearance changes in samples with higher MEMO content (ME3; ME3-IB1) can be probably found in slightly different coating structure. Due to higher MEMO content could be the resulting structure less crosslinked, or it could exert other faults leading to acceleration of corrosion.

The visual appearance of coatings ME1 and ME1-IB2 before and after exposure to damp-heat for 110 days is compared in the figures (Figs. 9, 10). In both figures, it is noticeable strong corrosion occurring on the uncoated sites of the samples. On the contrary, the coated sites remained clearer, without stronger corrosion damage.

Fig. 9
figure 9

Visual appearance of sample ME1 after accelerated weathering test, (from the left) untested sample; ME1 (5 and 15 cm/min) after 80 days; ME1 (5 and 15 cm/min) after 110 days (80% RH, 60 °C)

Full size image
Fig. 10
figure 10

Visual appearance of sample ME1-IB1 after accelerated weathering test, (from the left) untested sample; ME1-IB1 (5 and 15 cm/min) after 80 days; ME1-IB1 (5 and 15 cm/min) after 110 days (80% RH, 60 °C)

Full size image

To investigate how accelerated weathering test affected transmittance of the samples ME1 and ME1-IB1, UV–Vis spectrophotometric measurements were carried out after the test (Fig. 11). The results showed that bare substrate glass exhibited a significant drop in transmittance, below 90%. Samples ME1 and ME1-IB1 manifested slightly higher transmittance in the major part of the analyzed spectrum. Quite different was the result of sample ME1-IB1 prepared from 5-day aged sol. Differences were up to 3% at the spectrum margin (λ = 400 nm). Apparently, the exposition to damp heat accelerates not only glass corrosion, but it promotes some further structural and chemical changes, such as surface morphology changes or changes in sol–gel coating thickness, resulting in transmittance changes. Nevertheless, to prove the mentioned hypothesis, subsequent experiments are needed.

Fig. 11
figure 11

Transmittance of samples before and after the accelerated weathering test (110 days, 80% RH, 60 °C)

Full size image

3.5 Chemical resistance test — interaction with solutions

Chemical resistance of the coatings was verified by immersing the samples into the solutions with pH varying from 2 to 13 for 7 days, at laboratory temperature. The pH value of solutions was adjusted by adding the concentrated hydrochloric acid or sodium hydroxide saturated solution into demineralized water. Prepared solutions were transferred into polypropylene flasks with grooves for positioning glass slides. The samples were placed into the flasks and left for 7 days at laboratory temperature. After the test, samples were rinsed with demineralized water and dried for 1 h at 60 °C.

It was found that the samples were not visually affected by interaction with solutions at the conditions described above. During rinsing the samples, changes in wettability were observable — all samples immersed in a solution of pH 13 exhibited higher wettability in comparison with samples immersed in solutions of pH 2 and pH 5. Microscopic analysis was in good agreement with visual observation: coatings remained transparent and seemed to be pristine.

UV–Vis spectroscopy measurements (Figs. 12, 13) showed that interaction with solutions of pH 2 and pH 5 does not affect transmittance at all, moreover, composition ME1-IB1 (Fig. 13) exhibits slight increase in transmittance after the test (this increase does not exceed 1%). Transmittance of samples made from other compositions remained similar after immersion in solutions with pH 2 and 5. Immersion in strong basic solution (pH = 13) led to a decrease in transmittance for all compositions, up to transmittance level of bare, uncoated glass. This could be explained by the full dissolution of coating in basic media. Transmittance attenuation can be accounted for decrease of coating mass as a result of dissolution, moreover the chemical changes can cause uneven, damaged surfaces with increased light scattering. On the contrary, justification of resistance against acid solution can be found in alkali ions deficit in the sol–gel coating. Because of that, alkali ions can not be leached out of the surface layers easily and thus corrosively debase coatings surface.

Fig. 12
figure 12

Transmittance of sample ME1 before and after the interaction with solutions

Full size image
Fig. 13
figure 13

Transmittance of sample ME1-IB1 before and after the interaction with solutions

Full size image

4 Conclusions

This article experimentally verified possibilities in the functionalization method of sol–gel coating preparation. It has been proven that it is possible to prepare coatings which are durable enough to withstand utilization in exteriors. These coatings were prepared using only one-step synthesis, without the firing necessity. Such conditions implicate further diminishing of economic demandingness.

From all prepared coatings the best results were attributed to compositions combining functionalization with 1% MEMO and 1% IBTMS, and also compositions comprising only of MEMO functionalization, up to 2%. Those coatings combined sufficient increase of transmittance in the concerned spectrum while being chemically and mechanically durable. It was shown that a strong basic environment damages coatings more than an acid environment. A significant decrease in mechanical performance with increasing content of IBTMS functionalization was observed.