Analysis of the Self-Cleaning Potential of Glass Fiber Reinforced Concrete (GRC) with TiO₂ Nanoparticles

Abstract

The materials used in civil construction are undergoing significant advances to achieve reduced maintenance and increased durability. This study analyzed the self-cleaning potential of Glass fiber Reinforced Concrete (GRC) with the addition of titanium dioxide (TiO₂) in contents of 3, 5, and 7% with respect to the mass of cement. We evaluated the self-cleaning GRC plates and the compressive and flexural strength of cylindrical and prismatic specimens. Prepared GRC sample plates were stained with dye solution (rhodamine B and methylene blue) and exposed to the four cardinal solar orientations of a building façade (north, south, east, and west) at different inclination angles (0°, 45°, and 90°) with respect to ground level. Results showed that the samples that presented the greatest performance were plates positioned in a north orientation and inclined at 0° in relation to ground level. The inclusion of TiO₂ positively affected the consistency of the mixtures and improved the properties of the GRC in the hardened state. Measured rupture stresses were greater than 100 MPa in compressive strength and 20 MPa in flexure. The results of this study showed that the introduction of TiO₂ in concrete with high strengths did have great relevance for the self-cleaning of white concrete.

1. Introduction

White concrete and glass fiber reinforced concrete (GRC) have been used in building façade elements when aesthetics and sustainability characteristics are desired [1]. This result is achieved due to the reduction in coating materials (e.g., mortar and paint) applied to building façades and maintenance costs of this construction system [2]. Any façade elements are subjected to an elevated concentration of environmental pollution which dirties the surface and generates sustainability concerns [3].

GRC is a versatile and consolidated technology material that can be used in constructing façade elements. It has been used in façade elements since the early 1970’s in more than 50 countries. It consists of a cement mixture reinforced with alkali-resistant fibers with high flexural strength [4,5]. This allows the manufacture of elements of reduced thickness and less specific weight [6]. When applied to façades, GRC allows standardization of the manufacture and assembly of panels which reduces costs with no decrease in performance. As such, GRC can be used in new buildings or in the restoration and rehabilitation of old ones [7].

GRC is considered a cementitious composite with ductile behavior which results in visible deformations and multiple cracking [8]. As reported by Iskender and Karasu [6], the fiber content in GRC improves its performance with respect to load absorption, impact resistance, and flexural and tensile strength. However, building façades are also often subjected to staining due to weathering and urban pollution [9]. This process is not restricted to particles suspended in the atmosphere which accumulate on the surface but also includes the result of chemical reactions. These chemical reactions are induced by solar irradiation or inherent in the manufacture and assembly processes [10]. It was noted by Mansour and Al-Dawery [11] that the dirtying of a façade could be a result of the assembly and attachment technique used or the growth of microorganisms such as fungi and bacteria.

Façade preservation could be achieved with the application of products that decrease the accumulation of dirt or provide self-cleaning. A possible technique is the incorporation of titanium dioxide (TiO₂) in the cement matrix [11,12]. Titanium dioxide prevents the impregnation of dirt when exposed to solar irradiation [13]. Studies have shown that TiO₂ is a hydroxyl with photocatalytic and hygroscopic properties whose chemical reactions, when mixed in GRC, are accelerated under UV radiation [14]. Photocatalysis emulates natural photosynthesis adsorbed and contributes to the separation of electron pairs [15,16]. The activated photocatalyst promotes CO₂ photoreduction and contributes to the elimination of microorganisms [17,18].

Studies of self-cleaning TiO₂-based products have confirmed that photocatalysis occurred under UV light. Treviso [19] tested mortar façades with 5%, 10%, and 15% of TiO₂ content and confirmed self-cleaning performance. Such mineral self-cleaning additives applied to façades and subjected to solar irradiation and water decreased the surface adhesiveness of particles, thus preventing the deposit of fine or organic materials even on rough surfaces. The speed of the self-cleaning oxidation reaction is related to UV intensity, temperature, relative humidity, TiO₂ content, and particle adhesiveness [11,18,20].

In cement matrices based on Portland Cement, TiO₂ is an inert material that does not participate in hydration reactions or react with any of the phases of cementitious composite. However, it has been determined by Ma et al. [21] to act as a filler that reduced pores and produced corresponding improvements in the compressive and flexural strength of the GRC. Sikora, Horszczaruk, and Rucinska [22] obtained similar improvements in these mechanical properties with a 3% TiO₂ content when compared to a reference sample. Application of TiO₂ could occur as a surface layer or incorporated into the concrete matrix. In this condition, it was confirmed as a filler material that reduced porosity but also reduced initial strength [7,14].

Additionally, if used in proportions of more than 1%, it negatively affected the workability of the concrete [19,20,23]. Stained samples subjected to UV radiation presented some color recovery, but this effect decreased over time and with higher proportions of TiO₂.

Considering the presented scenario, it is possible to note a lack of studies evaluating the potential of self-cleaning GRC in authentic and natural exposure conditions and with different dirt elements. This study evaluates different percentages of TiO₂ in GRC samples containing glass fiber, considering the exposure in a roof slab with rhodamine B and methylene blue.

2. Materials and Methods

The experimental methodology consisted of tests on fresh and hardened GRC composite plates as well as visual analyses of the samples and spectrophotometry. The GRC composites incorporated increasing TiO₂ contents with respect to the mass of cement and were identified accordingly: GRC3 with 3%, GRC5 with 5%, and GRC7 with 7%.

Materials used for the manufacture of GRC were white Portland Cement type II, quartz sand, titanium dioxide, and AR glass fiber added as 0.5% with respect to total volume. The glass fibers used had a length of 10 mm and a diameter of 12 µm (shape factor of 833). The Polycarboxylate Ethers Superplasticizer (PCE-SP) was added at 1.5% with respect to cement.

Table 1. Mix ratio of samples manufactured for this study (in mass).

Composite	Cement	Quartz Sand	TiO2 * %	Glass Fiber by % Volume **	w/c	SP * %
REF	1	1	0	0.5	0.2	1.5
GRC3	1	1	3	0.5	0.2	1.5
GRC5	1	1	5	0.5	0.2	1.5
GRC7	1	1	7	0.5	0.2	1.5

* addition proportion with respect to cement, ** proportion with respect to total volume of the mixture.

presents the mix ratio, in mass, of the prepared samples.

Slump tests have been conducted in the past on fresh samples following the procedures of standard EN 1170-1 [24]. Mechanical properties were measured through compressive and flexural strength tests on a universal testing machine (INSTRON Machine testing). For the flexural tests, the ASTM C348:2021 [25] method was used. Six cylindrical specimens were produced for each composite by age. That is, in total, 48 cylindrical tests were produced. For the prismatic samples, the same sampling was followed.

Color recoveries of the different TiO₂ contents were evaluated by saturating the plates with rhodamine B dye diluted in water in a solution of 0.5 g/L as recommended by Treviso [19]. Plates were immersed in the dye solution for 24h, followed by exposition to natural sunlight for 21 days followed by visual analyses. The procedure was repeated on the same plates with methylene blue in a 5.0 g/L solution as performed by Balbino [26], followed by exposition to natural sunlight for 21 days followed by spectrophotometry measurements. The plates were repurposed from the rhodamine B testing since façade elements are subjected to repeat dirtying. The color recovery testing procedure is shown in the flowchart in Figure 1.

Natural sunlight exposure was conducted in three distinct inclinations: 0°, 45°, and 90°. As shown in Figure 2, the plates were mounted on wooden frames which secured each inclination.

The frames were placed along the four cardinal orientations (north, south, east, and west) each inclination made use of four plates (coordinates: −29.79193720757752, −51.14931743845786) In total, 96 plates were used to analyze the self-cleaning capacity of the mixtures. Of these, 48 were stained with rhodamine B and 48 with methylene blue. The frames were mounted on the rooftop of an academic building at the Unisinos University campus as seen in Figure 3. The incident UV radiation was in the microclimate area of the city of São Leopoldo. This process was conducted for each type of dye for 21 days each.

As reported by Santos [27], for a surface to receive the greatest radiation possible, it must be placed at the same inclination as the local latitude and at 0° (north) azimuth in the southern hemisphere as is the case of São Leopoldo. In addition to geographical considerations, Macphee and Folli [28] noted that TiO₂ activation was affected both by geographical and seasonal conditions. Since the exposure part of this study occurred between March and May, weather conditions during this period such as solar irradiation, precipitation, temperature, and relative humidity were also obtained. These variations are shown in Figure 4.

Spectrophotometry measures the light absorbed or transmitted through a solid medium. This equipment is used to measure the difference in the tone of the plates with specific spectrophotometer coordinates denoted as L*, a*, and b*. For example, a pure white color tone has reference coordinates L* = 100, a* = 0, and b* = 0. As the surface becomes covered in dirt, the L* coordinate decreases, and the a* and b* coordinates vary up or down depending on the tonality of the dirt (see Figure 5).

With the values of L*, a*, and b* it is possible to calculate the color variation (∆E), which is determined by Equation (1).

$$\mathsf{\Delta }E=\sqrt{\left(\mathsf{\Delta }{L}^2\right)+\left(\mathsf{\Delta }{a}^2\right)+\left(\mathsf{\Delta }{b}^2\right)}$$

(1)

For this study, these measurements accounted for TiO₂ content, orientation, and inclination of the plates. As TiO₂ activated and induced self-cleaning, spectrophotometer coordinates were expected to return to their initial values. It should be noted that the test plates of this study were not perfectly white and consequently had non-zero values for the initial a* and b* coordinate.

Readings were taken at 4 locations on each plate with a Konica Minolta portable spectrophotometer model CM-2500d. Readings were aided by a guide surface with cutouts 1 cm in diameter, with the equipment placed at the center of each cutout. Three sets of measurements were taken prior to the saturation with methylene blue (day 1), after saturation (day 2), and after 21 days of exposure to natural sunlight (day 3).

For self-cleaning and color recovery tests under ambient sunlight, plates measuring 120 mm × 120 mm × 20 mm (width × length × thickness) were produced with a single layer of each mixture into molds. This layer was mechanically densified using a vibration table. From the results, the dirtiness reduction index (∆S) of each GRC was calculated, using Equation (2).

$$\mathsf{\Delta }S(\%)=\left(\frac{\mathsf{\Delta }{E}_{inicial }-\mathsf{\Delta }{E}_{j,days}}{\mathsf{\Delta }{E}_{\mathrm{inicial}}}\right)\cdot 100$$

(2)

Folli [29] noted that TiO₂ is inert. It did not react chemically with the other components of the cementitious matrix to increase strength but could contribute to filler effects. Figure 6 shows the granulometric curve of the materials in the mixture. As seen in the figure, TiO₂ had the smallest diameters among the other materials and, as such, was also designated as nano-titanium dioxide.

3. Results

3.1. Properties in the Fresh State

Slump test results are shown in

Table 2. Slump test results.

Composite	Ø (mm)	Image	Composite	Ø (mm)	Image
REF	65		GRC5	90
GRC3	75		GRC7	95

. Table 2 displays a noticeable effect of TiO₂: the slump diameter increases on average 10 mm proportionally to content.

The studies of Meng et al. [30] and Sorathiya et al. [31] reported that TiO₂ addition was proportionally detrimental to the fluidity of the cementitious matrix with increasing content. Decreased workability was also noted and attributed to the increase in specific surface area due to the presence of nano-TiO₂ [32]. Other studies such as Zapata et al. [33] and Mukharjee et al. [34] confirmed that the change in the characteristic of the concrete was related to the fineness modulus and surface area of TiO₂ particles. Reduced workability was also measured by Jalal et al. [35] as decreases in slump diameter of 10, 20, 40, 60, and 70 mm for TiO₂ contents of 1, 2, 3, 4, and 5%, respectively, in the mass of the mixture.

On the other hand, other studies have obtained results opposite to the previous trends. Noorvand et al. [36] evaluated the effect of TiO₂ added to black rice husk mortar. Results showed increased fluidity proportional to TiO₂ content with 30% rice husk substitution. This effect was believed to be associated with the filler effect of TiO₂ in the pores of the matrix. Mohseni et al. [37] also demonstrated improved workability with higher TiO₂ content and the slump diameter was increased by as much as 5.3% with respect to a reference mixture. This result was attributed to the sphericity of TiO₂ particles. This behavior may also have occurred in this study, increasing the fine content and thus, with filler effect.

3.2. Compressive and Flexural Strength Results

Average strengths and standard deviation of axial compressive and flexural strength tests of the mixtures at 7 days and 28 days of curing are shown in

Table 3. Compressive and flexural strength results.

Composite	Compressive Strength (MPa)		Flexural Strength (MPa)
	7 Days	28 Days	7 Days	28 Days
REF	93.1 ± 5.0	118.1 ± 3.0	17.9 ± 3.0	19.7 ± 3.0
GRC3	99.5 ± 1.1	118.3 ± 2.6	20.2 ± 4.0	20.8 ± 4.0
GRC5	105.2 ± 6.0	120.0 ± 3.0	18.2 ± 1.0	20.3 ± 1.0
GRC7	96.3 ± 7.0	121.8 ± 2.0	17.8 ± 3.0	21.4 ± 2.0

.

At 7 days, the compressive strength increased with increasing TiO₂ content up to 5%, which matched the results of Linsebigler et al. [38]. At this age, GRC5% obtained the highest compressive strength among the composites and the difference between them was up to 13%. This positive effect from small additions of TiO₂ was also observed in Ma et al. [21]. Although it did not follow a linear increasing trend, sample GRC7 still presented a higher compressive strength than the reference, as with all the other samples. A similar trend is observed at 28 days of curing but with the overall values closer to the reference sample strength. Still, in this case, sample GRC7 presented the highest compressive strength.

An understanding of the effects of TiO₂ on compressive strength could be gained by examining grain packing in the concrete mixtures. Yunsheng et al. [39] reported that increased strength and decreased porosity could be traced to the packing particle curve since finer particles increased density by reducing void spaces. This effect would be assisted by the inert nature of TiO₂, which did not react with any of the concrete phases or take part in cement hydration reactions.

Table 3 shows that TiO₂ did not induce substantial gains in flexural strength. Overall, results followed Fernandes [40] and Austria [41], who determined that a 10% TiO₂ content yielded the most increase in flexural strength, albeit a small one. Austria [41] noted that this is possibly a consequence of the fineness of the material resulting in a filler effect. Fiber reinforcement was known to improve matrix post-cracking performance due to the reduced spacing between reinforcement elements when compared to a conventional material [42,43].

3.3. Plate Color (before Staining)

Differences in surface tonality were observed immediately after demolding the plates. As shown in Figure 7, samples were gradually lighter in tone as TiO₂ content increased.

This was expected since TiO₂ is considered a pigment and used as a whitening and opacity agent in other products such as coatings, plastics, paints, glasses, and others [44] and this effect was known to be proportional to TiO₂ concentration. This way, the staining of the plates with titanium dioxide was more apparent in the larger plates due to the difference in the white tone. However, evaluations related to the staining reduction index were carried out among the specimens of each sample in different periods.

3.4. Visual Analysis with Rhodamine B

This chapter will present the results obtained with plates impregnated with rhodamine B. As mentioned in the introduction section, the UV incidence reacts with TiO₂ allowing the plates to self-clean, and thus, the angulation and orientation play an important role in the mentioned effect.

Figure 8 presents visual results of plate samples impregnated with rhodamine B with respect to preset inclinations of 0°, 45°, and 90° and a north orientation.

As seen in Figure 8, sample GRC7 obtained the best results with lighter tonality in all inclinations. It should be noted that most rhodamine B staining was removed after 7 days and the remaining staining along the edges disappeared after 14 days. The north orientation resulted in the most UV incidence and, coupled with precipitation between day 7 and day 21, induced sufficient rhodamine B decomposition chemical reactions. This result partially matched Treviso [19], which obtained the most change in color with a north orientation and 45° inclination, with 0° and 90° not too far behind. In the case of this study, an inclination of 0º yielded the best lighter color results at 14 days and at 21 days. This was likely a result of water accumulation due to the horizontal inclination of the plate. In contrast, Burger and Rüther [45] concluded that, at high latitudes, a vertical position (as in 90°) was the most favorable for solar irradiation. As for cleaning time, Sikora et al. [22] claimed that TiO₂ cleaning effects could be triggered in as short as two hours of exposition.

Figure 9 presents visual results of the plates with respect to preset inclinations of 0°, 45°, and 90° and an east orientation.

Results of Figure 9 show that east-oriented plates did not present the same self-cleaning evolution as the north orientation of Figure 8. The reference and sample GRC3 inclined at 0° and 90° still displayed traces of color at 14 days. The same traces could be observed for sample GRC5 at a 45° inclination. Sample GRC7 inclined at 45° showed staining due to the oxidation of the molds immersed in rhodamine B, which was not removed by the self-cleaning process since it was not organic in nature. This was previously verified by Diamanti et al. [46] in that inorganic iron oxide reduced the self-cleaning ability of GRC. Beeldens [47] also noted that TiO₂ was only able to degrade organic dirtying and not iron oxide staining. Figure 9 also demonstrates that there was little visible difference between inclinations of 0° and 90° although the plates at 0º still presented some advantage probably due to deformations caused by the accumulation of water.

Figure 10 presents visual results of the plates with respect to preset inclinations of 0°, 45°, and 90° and a west orientation.

Figure 10 shows that, for all preset inclinations, results were similar for all samples but GRC7 presented the best performance. Overall, most staining had degraded by day 7 with some coloration remaining along the edges of the reference sample, which did not last past day 14. For a west orientation, it should be noted that the best result was obtained at 0°. Considerable apparent deformations were observed at 90°, probably due to water accumulation on the rough surface which was unable to runoff due to the inclination.

Figure 11 presents visual results of the plates with respect to preset inclinations of 0°, 45°, and 90° and a south orientation.

Figure 11 shows that sample GRC7 at 0° inclination presented the best result with all staining degraded by day 7. On the other hand, GRC7 plates with a 90° inclination presented inorganic oxidation stains which were not degraded as expected. Overall, the south-facing plates presented the most difference between inclinations compared with the other orientations with 0° being lighter and 90° more stained.

3.5. Visual and Spectrophotometric Analysis of Methylene Blue Staining

The GRC plates stained with methylene blue were also exposed to 21 days of natural sunlight. Throughout this period, no precipitation occurred from day 0 to day 7 but was present between day 7 and day 21. Consequently, the reductions in dirtiness presented in the first 7 days of exposure were the consequence of UV radiation only. Results are presented in Figure 12 for plate inclinations of 0°, 45°, and 90°.

Results of Figure 12 show that the dirtiness reduction index (∆S) was nearly identical for all GRC samples in the first 7 days of exposure. This was not only observed across all GRC samples but also regardless of orientation or inclination. It should be stressed that no precipitation occurred during this period and thus stressed the role of water as a trigger for TiO₂ reactions. Diamanti et al. [46] pointed out that water was a necessary component in TiO₂-based self-cleaning compounds since it contributed to the activation and release of chemical substances that decreased the dirtiness of surfaces.

In addition to the absence of water, the low ∆S index observed in the first 7 days was also a result of the effect of the contaminant on the surface of the plates, which reduced the amount of UV radiation passing through to the GRC matrix. The layer of dirt and contaminants on the sample’s surface may have hampered for UV rays to penetrate until they reached the TiO₂ contained in the GRC. Since TiO₂ was effectively blocked from the UV rays, chemical reactions were retarded and so was the self-cleaning effect. Despite the lack of water and blocked UV rays, the GRC matrices with TiO₂ addition still presented a higher ∆S than the reference sample. Past day 7, dirtiness decreased both visually and on the ∆S index curve from precipitation and degradation of the contaminant layer, which progressively allowed more UV rays to penetrate the matrix. These external parameters were also cited by Diamanti et al. [46] as relevant to the effectiveness of self-cleaning.

Overall, increases in TiO₂ content in the cementitious matrix resulted in greater reductions in staining. Consequently, sample GRC7 with 7% TiO₂ addition presented the most promising results with an ∆S of up to 93%. The effects of concentration became more evident depending on UV exposure conditions. Thus, a 0^o inclination coupled with a north or west orientation had the most intense and longest UV exposure, which resulted in more active TiO₂ reactions and a self-cleaning effect as reported by Treviso [19].

The reference sample also presented reductions in staining, but these were related to degradation from natural cycles of wetting and drying. In this process, dirtiness indices (∆S) were over 50% especially in plates with a north orientation. However, there was visible residual staining which was much more noticeable than in the plates with TiO₂ content.

Evaluating the effect of inclination, fewer ∆S indices corresponded to 90°. At this inclination, UV incidence was unfavorable, and, at certain periods of the day, shades were observed on the surface of the plates. For the plates at 90° and with a south orientation, ∆S varied from 57% to 93% which demonstrated the severe impact of shading on the effectiveness of the self-cleaning action of TiO₂.

Surface temperature also played a role as significant as the presence of water and UV incidence. Plate temperature can intensify TiO₂ chemical reactions and accelerate the self-cleaning effect. In theory, that would result in decreased performance in plates facing south and east since those should have lower temperatures. In practice, results showed little variation in ∆S between south and east-facing samples regardless of TiO₂ content. However, plates with the same inclination and facing north presented increasing differences in ∆S as TiO₂ content increased, with GRC7 samples having the best performance.

4. Conclusions

This study evaluated the effect of TiO₂ addition on the self-cleaning and compressive and flexural strength of the GRC plates. Sample GRC7, with 7% of TiO₂ in the mass of cement reached a higher dirtiness reduction index (∆S) after 21 days of exposition. In relation to ∆S, it was already expected that the results would be better than the other mixtures due to the physicochemical functioning of the compound. The following conclusions were drawn from this study:

Higher TiO₂ content resulted in higher self-cleaning as there were visible improvements in self-cleaning effects when compared to the reference sample. Due to the location of São Leopoldo in the Southern hemisphere, the best results were obtained with the plates facing north. While variations in inclination were not as effective, a horizontal 0° inclination had the best result from the incidence of solar irradiation throughout the day.

TIO₂ did not produce a self-cleaning effect on inorganic materials during the relatively short term of this study. Further, long-term studies should be performed to confirm this phenomenon.

The layer of dirt and contaminants on the surface of the samples may have made it difficult for UV rays to penetrate until they reached the TiO₂ contained in the GRC, resulting in variations in the self-cleaning results. Thus, depending on the contaminant and exposure conditions to which the plates are exposed, self-cleaning is not guaranteed only by increasing the concentration of TiO₂.

The solar orientation directly influenced the ∆S results of the samples, since it is related to the exposure to UV rays and the time of insolation. However, the inclination of the plates was also decisive, since the best results were obtained at 45°. On this slope, good levels of insolation and the presence of water/moisture on the face after exposure to rain were observed. Water was an important precursor agent for TiO₂ activation, in addition to favoring the runoff of dirt through “washing”.