HomeAbout usPolicyProfessional DevelopmentQualifications and CoursesEventsMembershipMedianavigationend 

The impact of chemical treatments on the wear, gloss, roughness, maintenance, and slipperiness of glazed ceramic tiles

Volume:09

Issue:02

Year: 2009

Dr François Quirion1, André Massicotte1, Sophie Boudrias1, Patrice Poirier1
1 QI Recherche et Développement Technologique Inc., Montréal, Québec, Canada.

Correspondence : Dr François Quirion, QInc, 10301 avenue Pelletier, Montréal, Québec, Canada, H1H 3R2.

Abstract

Glazed ceramic tiles are known to be slippery when wet and many surface treatments are offered on the market to improve their slip resistance. Surprisingly, there are very few systematic investigations of the impact of these surface treatments on the slip resistance and surface properties of these tiles. Among the most common treatments are those based on hydrofluoric acid and those based on ammonium bifluoride. In this investigation, these two treatments were applied to glazed ceramic tiles and the impact on the mass loss, resistance to abrasive wear, average roughness, gloss, ease of maintenance, dry and wet friction and aquaplaning threshold was evaluated.

Scanning electron microscopy of the treated glaze suggests that the hydrofluoric acid treatment dissolves part of the glaze and leaves holes at the surface while the ammonium bifluoride treatment leaves a layer of submicron particles on top of the glaze. The glaze treated with hydrofluoric acid is more fragile and sensitive to abrasive and maintenance wear than the untreated tiles. The layer of submicron particles does not adhere strongly to the glaze and it wears out quite easily. The microscopic changes at the surface of glazed tiles have little impact on the average roughness but they slightly increase the slip resistance of the treated tiles. However, the treatments also reduce the gloss and make the tiles more difficult to clean.

Key words: Bifluoride; chemical etch; environmental health; fluorhydric acid; friction; glazed ceramic tiles; slips and falls; slippery floors.

Introduction

Smooth glazed ceramic tiles are known to be slippery when wet. For instance, Cholet et al., (2000) investigated the friction of different flooring using various apparatus and the friction of the smooth and glazed ceramic tiles was always among the lowest. Even so, glazed ceramic tiles are still found in bathrooms, kitchens and halls here they are likely to become wet and slippery. When slips and falls occur, the owners usually prefer to apply a treatment on their slippery floor because it is less expensive than replacing it by non slippery tiles. Many options, including chemical treatments based on either hydrofluoric acid, HF, or ammonium bifluoride, ABF, are offered on the market.

It was suggested by Chang et al., (2001) that friction depends not only on the macroscopic texture of a

surface but also on its microscopic texture. In the UK, HSL (Lemon and Worth, 2008) has developed a ‘slip assessment tool’ based on the measurement of surface micro-roughness. Interestingly, most manufacturers of the chemical treatments claim that their product generates roughness at the micro-level. Surprisingly, there are very few investigations of the impact of these chemical treatments on the slip resistance and roughness of glazed ceramic tiles.

Grönqvist et al., (1992, 2003) reported a significant increase of the wet friction of a glazed ceramic tile after a treatment with a solution of ammonium bifluoride at 5% in water. The increase in wet friction was accompanied by an increase of the micro-roughness.

Bowman et al., (2002) report friction measurements with various methods on a polished porcelain tile treated with a “proprietary floor surface etching treatment” but the chemical nature of the treatment and the application procedure are not described. They observed a significant loss of gloss and that “the SATRA STM 103, Pendulum and wet barefoot ramp tests were all able to determine an improvement in the slip resistance of the etched polished porcelain tile. The VIT and oil-wet ramp tests were unable to detect an improvement”. They also mention that these conclusions are contrary to those of Di Pilla (2000) who used the same VIT to emphasise the improvement of the wet friction of glazed ceramic tiles treated with various products.

In Di Pilla’s study (2000), the surface treatment was applied by the vendors and the nature and the application procedure of these treatments are not given. Nevertheless, they find that some treatments drastically improve the wet friction of glazed ceramic tiles while others have little or no effects. However, they also mention that “there appeared to be low correlation between product claims and the efficacy of the product”. Di Pilla (2000) finally concludes that “more work needs to be done in evaluating the efficacy of these products”.

These results, sometimes contradictory, raise some questions on the efficacy of chemical treatments to improve the slip resistance of glazed ceramic tiles. Moreover, they do not answer the questions related to the impact of these treatments on the resistance to wear, the roughness, the gloss and the ease of maintenance of the treated tiles. In this investigation, we present a systematic investigation of the impact of hydrofluoric acid and ammonium bifluoride treatments on the surface properties of glazed ceramic tiles.

Chemical treatments and tiles tested

All the experiments described in this article were performed in our laboratory using glazed ceramic tiles and chemical treatments available commercially. Two types of chemical treatment were tested. The first one contains hydrofluoric acid in an aqueous surfactant solution and the second contains ammonium bifluoride in a surfactant solution. For comparison purposes, some tiles were treated with either water or a neutralised solution of the hydrofluoric acid treatment.

Hydrofluoric acid treatment, HF

Titration with sodium hydroxide confirms the presence of 17.2 % (w/w) of HF in the commercial product. As indicated on the label, the treatment was performed with a solution consisting of one part of the product and one part of tap water, resulting in a solution containing 8.6% HF. That solution was applied on the surface of the glazed ceramic tiles (~ 400 ml/m2) and scrubbed gently with a piece of red floor pad to maintain the surface wet for the duration of the treatment (20 minutes). After the treatment, the tiles were rinsed thoroughly under running water, without any neutralisation, and dried at room temperature. The tiles treated in that manner are referred to as HF tiles.

Ammonium bifluoride, ABF

The glazed ceramic tiles were treated with two commercial products containing ammonium bifluoride. The results were very similar and they are combined without distinction in this investigation. Assuming that 2 moles of sodium hydroxide (NaOH) are required to neutralise ABF (NH4HF2), titration of both products led to a concentration of 1.25 % (w/w) of ABF. As recommended by the manufacturers, approximately 400 ml/m2 of the product was applied directly on the glazed ceramic tiles. The samples were scrubbed gently with a piece of red floor pad to maintain the surface wet for the duration of the treatment (40 minutes). The samples were then rinsed thoroughly under running water, without any neutralisation, and dried at room temperature. The tiles treated in this manner are referred to as ABF tiles.

Blank treatment, BLANK

The blank treatment is essentially a treatment that is performed in the same manner as the HF or ABF treatments but with water or a neutralised solution of the HF treatment. The blank solution was applied on the glazed ceramic tiles at a surface concentration of ~400 ml/m2 and scrubbed gently with a piece of red floor pad to maintain the surface wet for the duration of the treatment (20 minutes). The samples were then rinsed thoroughly under running water and dried at room temperature. These tiles are refered to as BLANK tiles.

Glazed ceramic tiles

All the results presented in this investigation were obtained with glazed ceramic tiles (Cecrisa, White Basic Matte, PEI = 4, 20 cm x 20 cm or Portobello, ARQ NEVE, PEI = 3, 7.2 cm x 7.2 cm). The Cecrisa tiles were cut to the required size while the Portobello tiles were used as received. Six surface conditions were tested. The first three correspond to the tiles treated with the BLANK, HF or ABF treatment. The other three correspond to the tiles treated and worn mechanically.

Methodology

The methods used to evaluate the impact of the chemical treatments on the surface properties and slipperiness of glazed ceramic tiles are described below.

Wear

In this investigation, wear corresponds to the mass loss caused by an action performed on the surface of the glazed ceramic tiles and it is expressed in g/m2.

Abrasive wear

Abrasive wear simulates the wear caused by traffic. Experimentally, the sample tiles were sanded twice for 20 seconds each time with an orbital sander (Al2O3, grit 220) with the pressure from the weight of the sander (1.2 kg). The abrasive wear is determined as the mass difference before and after sanding. The test is rather soft but it allows us to identify very fragile surfaces.

Chemical wear

Chemical wear corresponds to the mass loss, g/m2, caused by treating the surface of the tiles with either the chemical or the blank treatments (see section above for details). The mass is measured before the treatments and after the tiles were treated and dried at 24 ± 1°C for at least 18 hours.

Maintenance wear

Maintenance wear corresponds to the mass loss through daily floor cleaning of the initially clean tiles. The maintenance wear was determined for damp mopping of the treated tiles with either water or a neutral floor cleaner (0.15% nonylphenol ethoxylate in tap water, pH ~10, T ~23°C) and machine scrubbing at 175 RPM with the neutral floor cleaner and a red floor pad (diameter of 43 cm). The maintenance wear is 

Tiles fig 1


Figure 1.0 Example of the analysis of data generated by the Falling Plate
Method.

The left side presents the sliding distance of a Neolite plate on an HF tile as a function of the apparent water thickness.The right side shows the friction ratio, μR (calculated with the sliding distances and equation 1) and the best fit obtained with equation 2 (Fitting parameters are μR, = 0.12, a = 1.5 and t* = 24 μm).

 

 

expressed as the mass loss, g/m2, following the equivalent of six months of daily floor cleaning.

Gloss

Gloss is a measure of the reflectivity of light on a surface. Experimentally, a red light beam (630 nm) was projected at an angle of 45° on the surface of the sample tiles and the intensity of the light reflected at 45° was measured using a photo resistive cell. In this investigation, the reflectivity of a given tile is expressed as the percentage of the reflectivity of the BLANK tiles. If the reflectivity of a tile, R, is lower than 100%, then its gloss is lower than the gloss of the BLANK tiles. The reflectivity is always the average of five measurements at five different locations on a tile and the standard deviation from tile to tile was typically ± 1 %, except for the tiles treated with HF where it is around ± 2 %.

Ease of maintenance of the tiles (EM)

To determine the ease of maintenance, a given amount of vegetable oil was spread homogeneously on the Portobello tiles, which were then cleaned using a procedure developed in our laboratory to simulate damp mopping (Quirion, 2004). The cleaned tiles were then dried at room temperature and weighed to determine the residual amount of oil left after cleaning. The ease of maintenance, EM, of a given tile is defined as the ratio of the residual amount of oil on that tile to the residual amount of oil on the BLANK tiles. EM increases from 0 to 1 with 1 being the ease of maintenance of the BLANK tiles. If EM <1, then the tiles are more difficult to clean than the BLANK tiles. For these experiments, the floor cleaner was an anionic degreaser containing sodium dodecylbenzene sulfonate and 2-butoxyethanol diluted at 0.4% in tap water with a pH ~ 11.

The Falling Plate Method: Aquaplaning threshold, t*, and wet friction, μwet

The Falling Plate Method is detailed in an earlier paper (Quirion and Poirier, 2007). In summary, a flat and thin Neolite plate standing perpendicular to a wet tile is tilted until it falls freely on the tile. After the plate hits the surface, it slides over a given distance, d. The Neolite plate is 64 mm high, 41 mm wide and it weighs 76 g. The average roughness of the Neolite surface was kept around Ra = 0.6 ± 0.2 μm over the period of the investigation.

The sliding distance was measured on the dry tiles and as a function of the amount of water on the surface of the tiles. The amount of water is expressed as a liquid thickness, t, (1 ml/m2 = 10-6 m3/m2 = 10-6 m = 1 μm). It is well known (Rabinowicz, 1995) that the sliding distance is correlated with the reciprocal of the friction coefficient, μ. Thus, the ratio of the dry, ddry, and wet, dwet, sliding distances should be correlated with the ratio, μR, of the wet, μwet, and dry, μdry friction coefficients (equation 1).

Tiles table 1

Table 1.0 Chemical wear (g/m2) of Cecrisa and Portobello ceramic tiles following different treatments. The Portobello tiles were treated by two different operators using the sameprocedure.

1 Negative values correspond to an increase in the mass of the test sample

equation 1

equation 2

Figure 1.0 shows a typical d vs. t data set obtained on a HF tile. As t increases, water fills the valleys and the sliding distance increases. Eventually, the surface of the tile becomes saturated with water and the sliding distance reaches a plateau. Figure 1.0 shows that the relative friction (μR) remains constant and low at high water thickness, in accordance with the concept of aquaplaning. The thickness of liquid required to reach that plateau is thus called the aquaplaning threshold, t*, and it was obtained by fitting each independent μR vs. t data set to equation 2 (μR, is the friction ratio at t =  and a is an exponent that accounts for the rate of friction drop). Note that t* can also be obtained by fitting equation 2 to μwet vs. t data sets, provided that one knows the value of μdry to get μwet from equation 1.

Typically, one μR vs t data set consisted of 10 to 15 sliding distances obtained in the range t = 8 to 100 μm of water. The values of t* reported in this paper are the average of two to six independent data sets.

Dry friction and wet friction

The dry friction was determined by pulling the Neolite slider on the dry tiles at a velocity around 22 mm/sec. The dynamic coefficient of friction corresponds to the ratio of the pulling force to the weight of the slider. The friction coefficient of one tile corresponds to the average of five determinations and the values of μdry reported in this investigation are the average of at least two tiles. The wet friction, μwet, was evaluated from the Falling Plate Method results using the dry friction of the Neolite slider and equation 1.

The dry and wet frictions were also obtained at a higher load using the Brungraber Mark II apparatus equipped  with the Neolite slider used for the Falling Plate Method. Initially, the foot of the Mark II was tilted so that its rear end was in contact with tile. As the weight was released, the slider fell on the tile in a manner very similar to the Falling Plate Method, but this time with a much higher load. That procedure was used to determine the friction under dry and wet conditions.

Average roughness (Ra)

The average roughness, Ra, was determined with a DekTak 3030 equipped with a diamond stylus having a tip radius of 12.5 μm. The measurement proceeded at low speed over a length of 5 mm with a 0.05 mN (5 mg) force applied on the stylus. The value of Ra was determined on five locations of a tile and the values reported are the average of 2 to 6 sample tiles.

Results

This section compares the wear, gloss, roughness, slipperiness and ease of maintenance of glazed ceramic tiles treated with a solution of hydrofluoric acid or ammonium bifluoride or with a blank solution (no active ingredients).

Wear of glazed ceramic tiles

The wear is associated with the mass loss caused by the chemical treatments, the abrasion from traffic and the maintenance of glazed ceramic tiles.

Chemical wear

The chemical wear for BLANK, HF and ABF tiles was determined for Cecrisa and Portobello tiles and the results are compared in Table 1.0. The Portobello tiles were treated by two different operators using the same procedure and the results are fairly reproducible from one operator to the other.

The first observation is that the chemical wear caused by 

Tiles fig 2


Figure 2.0 Scanning electron microscope (SEM) images of the treated Cecrisa tiles. The black horizontal bar corresponds to 10 μm except for the bottom right image (1 μm) that corresponds to the dashed rectangle of the bottom left image.

the HF solution is much more important than the chemical wear caused by the BLANK treatment. This is true for the Cecrisa and the Portobello tiles, suggesting that the effect is not limited to one brand of tile. The relatively high mass loss strongly suggests that hydrofluoric acid solubilises part of the glaze which is then washed away during the rinsing step. This was observed by Fang et al., (1997) for ceramics containing SiO2 and Al2O3 and by Lee et al., (2001) during the treatment of alumino-silicate fibres with HF. On the contrary, the treatment with ABF results in a slight increase of the mass, suggesting that a small amount of material is deposited on the tiles.

To better understand the difference between the action of HF and ABF, the surface of the treated tiles was analysed with scanning electron microscopy, SEM. As seen in Figure 2.0, the surface of the BLANK tile is rather smooth with some isolated peaks. After the HF treatment, the glaze presents a lot of holes, differing in size and depth, in accordance with the dissolution of part of the glaze by HF. The ABF treatment results in a thin layer of submicron particles deposited on the smooth glaze. In order to get some information on the chemical  composition of the treated surfaces, the Auger emissions were analysed and the scans are shown in Figure 3.0.

The peaks correspond to the atoms bombarded by the electron beam during the SEM experiments. For the SEM experiments, the surface of the samples is made electrically conductive with a very thin carbon film, which explains the origin of the carbon peak for all the samples investigated.

The peaks of the BLANK tile confirm that the surface of the glaze is mainly composed of silicon and oxygen (SiO2) and also sodium (Na), magnesium (Mg) and aluminum (Al) which are common atoms in glazes.

After the HF treatment, the glaze still has the same overall composition. This suggests that the glaze exposed by the formation of the holes has the same composition as the original glaze. This indicates that the HF treatment performed in our laboratory does not remove completely the glaze from the surface of the tiles.

As seen in Figure 2.0, the ABF treatment leads to submicron particles deposited on a smooth glaze.

Tiles fig. 3

Figure 3.0 Auger emissions of the treated Cecrisa tiles during the SEM experiments

The composition of the smooth glaze is almost the same as that of the BLANK tile. However, the composition of the submicron particles is quite different. They are mainly composed of F, Na, Mg, Al with very little SiO2. Since there is almost no fluorine atom on the original glaze, these results strongly suggest that the submicron particles originate from ammonium bifluoride which contains fluorine (NH4HF2). This explains the mass increase following the ABF treatment.

Abrasive wear

The resistance to wear caused by traffic was simulated by sanding the glazed tiles for a brief period of time and the results are expressed in terms of abrasive wear in Table 2.0 for Cecrisa and Portobello tiles. The abrasive wear is significantly higher for the HF tiles compared to the BLANK tiles suggesting that the HF treatment makes the glaze more fragile. These results are in agreement with Fang et al., (1997) who also observed a faster erosion rate for ceramics containing SiO2 and Al2O3 after they were treated with a solution containing HF. The mass loss caused by the abrasion of the ABF tiles corresponds fairly well with the mass gained during the BF treatment, suggesting that the layer of submicron particles is removed from the surface of the glaze during the abrasion test. The abrasion test is rather gentle indicating that the layer of submicron particles does not adhere strongly to the glaze.

Maintenance wear

Table 3.0 compares the impact of six month of daily floor cleaning by damp mopping with water or a neutral floor cleaner or by machine scrubbing with a neutral floor cleaner on BLANK and HF Portobello tiles.

Damp mopping with either a neutral floor cleaner or water resulted in a mass loss 19 and 30 times more important for the HF tiles relative to the BLANK tiles. Machine scrubbing with a neutral floor cleaner also removed 19 times more glaze on the HF tiles relative to the BLANK tiles. Once again, this suggests that the surface of the HF tiles is more fragile and sensitive to the action of floor cleaning than the BLANK tiles.

The chemicals present in the floor cleaner do not seem to be responsible for the mass loss because the maintenance with water also results in a significant mass loss. Not too surprisingly, the more aggressive machine scrubbing results in a higher maintenance wear for both the BLANK and the HF treated tiles.

Tiles table 2

Table 2.0 Abrasive wear (g/m2) of treated Cecrisa and Portobello tiles

Tiles table 3

Table 3.0 Maintenance wear (g/m2) of the treated Portobello tiles

Maintenance wear (g/m2) of the treated Portobello tiles caused by the equivalent of six months of daily floor cleaning by damp mopping with a neutral floor cleaner (NN) or water and machine scrubbing at 175 RPM with a red floor pad (43 cm) and a neutral floor cleaner.

Gloss and roughness of glazed ceramic tiles

Glazed ceramic tiles are often chosen because of their high gloss which is associated with smooth and glazed surfaces. This section looks at the impact of the chemical treatments on the gloss and roughness of the glaze.

Gloss

The gloss is expressed as the percentage of the reflectivity, R (%), of the BLANK tiles. Hence, in Table 4.0 the BLANK tiles have a reflectivity of 100 % and a loss of gloss corresponds to a value lower than 100 %.

As expected, the surface changes caused by the HF and ABF treatments reduce the gloss to about 65% of the Cecrisa tiles and 74% of the Portobello tiles. Surprisingly, the magnitude of the effect seems to be similar for the holes (HF tiles) and the submicron particles (ABF tiles). The effect seems to be greater for the smoother Cecrisa  tiles (Ra = 1.4 μm for Cecrisa and 4.4 μm for Portobello). The loss of gloss was also observed by Bowman et al., (2002) after etching glazed porcelain tiles. However, the comparison with our results is difficult since they did not give the nature of the chemical treatment.

Table 4.0 also indicates that abrasive wear has little impact on the gloss of BLANK tiles while it increases the gloss of the HF and ABF tiles. The increase is rather small for the worn HF tiles (+ 4% and + 2%) suggesting that the holes generated by the treatment are deeper than the action of the soft abrasion. The situation is different for the worn ABF tiles for which the gloss increases significantly (+ 20% and + 8%) following the soft abrasion. This corroborates that the layer of submicron particles does not adhere strongly and is easily removed from the smooth glaze. This also suggests that the ABF treatment has a reversible impact on the gloss while the HF treatment would decrease the gloss irreversibly.

Tiles table 4

Table 4.0 Reflectivity, R (%), and average roughness, Ra (μm), of the Cecrisa and Portobello tiles after different treatments Values of R (%) are ± 1 % except ± 2 % for the HF tiles.

Average roughness

There are many parameters used to express the roughness of a surface (Chang et al., 2004) and most of them depend on the experimental conditions (scan length, tip force, tip radius, horizontal speed, cut-off length, etc.). In this investigation, we used a DekTak profilometer that generates only the average roughness, Ra, and the values are reported in Table 4.0. Some may argue that the average roughness is not the best parameter to correlate the roughness of a surface with its wet friction. However, it seems that in certain conditions, Ra may be correlated with wet friction. For instance, Loo-Morrey (2007) has reported values of Ra and Rz for different floor types. We analysed these values and we found a very good correlation between Ra and Rz. The author mentions that “the value of waterwet PTV (Pendulum Test Value) measurements increases as the Rz surface roughness of the natural or man-made stones increases”. Since Ra is directly correlated with Rz, then one could also say that the water-wet PTV increases with Ra. In another report, Loo- Morrey (2006) compared the wet PTV with the wet friction coefficient obtained using the ramp test and it is reported that “the ramp CoF (coefficient of friction) increases as the pendulum CoF increases”. Hence, considering that Ra is correlated with the wet PTV and that the wet PTV is correlated with the wet ramp coefficient of friction, then one can assume that Ra is also correlated with the wet ramp coefficient of friction, at least for the flooring materials and the experimental conditions described by Loo-Morrey (2006, 2007).

The treatment of Cecrisa tiles with either HF or ABF increases only slightly the value of Ra. The results for the Portobello tiles even suggest a slight decrease of Ra upon treating the tiles with either HF or ABF. However, the uncertainty on Ra for these tiles is quite high so that the effect remains rather insignificant.

Grönqvist et al., (1992) reported an increase of Ra from 0.5 to 1.0 μm after the application of an undefined antislip treatment on a glazed ceramic tile. This variation is similar to what we observed (+ 0.2 and + 0.4 μm) for the HF and ABF treatments of the Cecrisa tiles but in disagreement with the decrease (-0.3 and -0.7 μm) observed for the rougher Portobello tiles. In a later study, they reported (Grönqvist et al., 2003) a significant increase of Rz from 2.6 ± 0.6 to 5.9 ± 0.9 μm following the treatment of a glazed ceramic tile with a solution of ammonium bifluoride (5 % w/w). Maybe the application of a more concentrated solution of ABF (5 % vs. 1.2 % in our case) leads to a higher increase of the roughness. Or maybe Rz is more sensitive to roughness changes than Ra.

Using atomic force microscopy Luo et al., (2001) obtained the average roughness, Ra, of a dental ceramic rich in SiO2 following the treatment with hydrofluoric acid (9.6%). For a 20 μm x 20 μm area, the value of Ra increased from 6 nm to about 278 nm, suggesting that Ra increases by 0.3 μm following the HF treatment. That value is similar to what was observed for the Cecrisa tiles (+ 0.2 μm). But once again, the experimental conditions are too different for any quantitative comparison.


Finally, abrasive wear (treated and worn tiles) appears to reduce the average roughness of the tiles. The effect remains small but it is observed for all the tiles tested, suggesting a definite trend. This decrease of the surface roughness following abrasive wear is in agreement with the increase of the gloss noted in the previous section.

Slipperiness of glazed ceramic tiles

The friction coefficient of a smooth Neolite sample with the glazed ceramic tiles was measured under dry (horizontal pull and Brungraber Mark II) and wet (Brungraber Mark II) conditions. The Falling Plate Method (Quirion and Poirier, 2007) was also used to evaluate the aquaplaning threshold and the wet friction.

Tiles table 5

Table 5.0 Dry friction, μdry, of Neolite on Cecrisa tiles after different treatments. Results were obtained with the Horizontal Pull method and the Brungraber Mark II

Tiles fig 4

Figure 4.0 Apparent friction, μwet, obtained using the Falling Plate Method with a Neolite slider as a function of the apparent water thickness, t, on the tiles. Tile numbers only distinguish independent data sets. Results obtained on the treated tiles are compared with those obtained on worn tiles

Dry friction

Table 5.0 summarises the dry friction coefficient obtained with the horizontal pull method and the Brungraber Mark II. The absolute values of the friction coefficient obtained with the horizontal pull method are lower than those obtained with the Mark II apparatus.

However, both methods suggest a slight increase of the dry friction caused by the HF treatment and little impact for the ABF treatment. These results are in agreement with the observations of Di Pilla (2000) who noted only a slight increase of the dry friction of glazed ceramic tiles following different surface treatments.

Once the HF and ABF tiles become worn, their dry friction drops to the same level and sometimes lower than tha of the BLANK tiles.

Aquaplaning threshold

Table 6.0 reports the average values of the aquaplaning threshold, t*, obtained from the analysis of the μwet vs. t data sets presented in Figure 4.0. In a previous investigation (Quirion and Poirier, 2005) we reported preliminary results of the aquaplaning threshold and wet friction on BLANK and HF Cecrisa tiles. At that time, only one data set for each tile was obtained with water and with a detergent solution (sodium lauryl sulphate at 0.15% in water). It was concluded that treating the Cecrisa tiles with HF had little impact on the average roughness and the aquaplaning threshold of the glazed ceramic tiles.

In the present study, the aquaplaning threshold was determined with a slightly different Neolite slider and the results presented in Table 6.0 are the average of four and six independent data sets (see Figure 4.0) obtained with BLANK and HF tiles covered with water. This time however, the aquaplaning threshold of the HF tiles increases from 16 ± 4 μm to 26 ± 4 μm. Although this improvement is significant, the values of t* remain small. For instance, 1 μm = 1 ml/m2 so that 26 μm of water is smaller than half a teaspoon of water spread on a 30 cm x 30 cm tile. Note also that in a previous investigation (Quirion and Poirier, 2006), values of t* = 58 μm were obtained for new and sealed quarry tiles (Ra = 5.5 ± 0.2μm) which are not considered as slip resistant tiles. The ABF treatment also increases the value of t* (from 16 to 20 μm) but the improvement is not significant within the experimental uncertainty.

The abrasive wear decreases the aquaplaning threshold of the ABF tiles to a value very close to the worn BLANK tiles, suggesting once again that the submicron particles at the surface of the ABF tiles are easily worn out leaving behind a surface similar to that of the worn BLANK tiles. Abrasive wear also decreases the aquaplaning threshold of the HF tiles. However, the worn HF tiles still have a higher aquaplaning threshold than that of the worn BLANK tiles. These results suggest that the HF treatment acts deeper into the glaze and is more resistant to abrasive wear than the ABF treatment. This observation is in accordance with the gloss results (see Table 4.0).

Wet friction

The wet friction coefficient of the Cecrisa tiles was determined using the μwet vs. t data sets generated with the Falling Plate Method (Figure 4.0) and with the Brungraber Mark II (Figure 5.0). The main difference between the two methods is the much higher load at impact for the Mark II (5 kg) compared to the Falling Plate Method (0.076kg).

For both methods, the wet friction decreases as the apparent water thickness increases until it reaches a plateau at low friction where the risk of aquaplaning is high. A quick comparison of Figures 4.0 and 5.0 indicates that the μwet vs. t data sets obtained with the Mark II

Tiles fig 5

Figure 5.0 Mark II friction coefficient of Neolite on treated Cecrisa tiles as a function of the apparent water thickness on the tiles

Tile numbers only distinguish independent data sets. Results obtained on the treated tiles (left) are compared with those obtained on the worn tiles (right). Results for the HF and HF-W were obtained with a solution of sodium lauryl sulphate (SLS) at 0.15% in water.

Tiles table 6

Table 6.0 Aquaplaning threshold, t*, and wet friction at 50 μm of water, μwet,50, of Neolite on Cecrisa tiles after different treatments

apparatus are shifted to lower apparent water thickness, suggesting that the aquaplaning threshold depends on the load at impact and that it occurs at lower water thickness for higher loads.

These observations indicate that, when comparing wet frictions, it is important to specify the apparent water thickness (amount of water per area of tile) at which the frictions are obtained. Table 6.0 compares the wet friction results at t = 50 μm (50 ml of water per m2). At that water thickness, all the tiles tested in this investigation are past the aquaplaning threshold and their wet friction is fairly constant.

The wet frictions at 50 μm obtained with the Falling Plate Method and the Mark II are in fair agreement. Considering that the two experimental methods are completely independent, the overall agreement between the two methods suggests that the Falling Plate Method can provide a fair estimate of the wet friction.

The wet friction obtained with the MARK II for the HF and the ABF treated tiles is essentially the same and slightly higher than that of the BLANK tiles. The Falling Plate Method generates similar results except that the wet friction of the BLANK tiles is similar to that of the HF and ABF treated tiles. Thus, the Mark II results suggest an improvement of the slip resistance while the Falling Plate Method indicates no changes. But in both cases, the wet friction of the treated tiles remains quite low (μwet,50 < 0.10) so that the improvement remains marginal.

These results do not agree with those of Di Pilla (2000) and Grönqvist et al., (2003) who both observed a drastic increase of the wet friction following chemical treatments. For example, Grönqvist reports an increase of the wet friction from 0.08 to 0.65 after treating a glazed ceramic tile with an ABF solution. The value for the untreated tile is in fair agreement with our results, but the wet friction for the treated tile is much higher than what we observed using the Falling Plate Method and the Mark II. This discrepancy corroborates the difficulty to compare friction results originating from different test methods.

Table 6.0 also indicates that abrasive wear has little impact on the wet friction of the BLANK and the ABF tiles. According to the Falling Plate results, the abrasive wear of the HF tiles has little effect on the wet friction while the Mark II suggests a significant decrease.

Tiles table 7

Table 7.0 Ease of maintenance, EM, of Portobello tiles by damp mopping with an anionic degreaser.

Maintenance of glazed ceramic tiles

Floor cleaning experiments on BLANK, HF and ABF tiles were performed in order to determine how the HF and ABF treatments affect the ease of maintenance of glazed ceramic tiles. The cleaning tests were conducted on Portobello tiles covered with vegetable oil and cleaned using damp mopping with an anionic degreaser. The results are reported in Table 7.0.

In this investigation, the ease of maintenance, EM, increases from 0 to 1 with one being the ease of maintenance of a BLANK tile. The first observation is that all values of EM are smaller than one, suggesting that the unworn BLANK tiles, used as the reference tiles, were the easiest to clean.

The HF and ABF treatments drastically reduce the value of EM to 0.12 and 0.14, respectively. In other words, the same cleaning procedure leaves around eight times more oil on the treated tiles relative to the unworn BLANK tiles. As the HF treatment wears out, the ease of maintenance re-increases but only to a value similar to the worn BLANK tiles. In other words, abrasive wear makes the glazed ceramic tiles more difficult to clean but that effect is less pronounced than the decrease of EM following the HF and ABF treatments.

Hupa et al., (2005) also noted that once a smooth glaze is corroded, it becomes more difficult to clean. In the case of HF and ABF tiles, the decrease of EM is probably due to the entrapment of oil into the holes left by the HF treatment or within the layer of submicron particles left
by the ABF treatment at the surface of the glaze. So, in order to properly remove the trapped oil, the treated floors have to be cleaned with a more aggressive procedure. Because of the fragility of the HF treated glaze and the small adherence of the submicron particle layer left by the ABF treatment, making floor cleaning more aggressive should also result in a faster deterioration of the glaze treated with HF and a faster disappearance of the submicron particles deposited through the ABF treatment.

Discussion

This investigation deals with the impact of chemical treatments based on aqueous solutions of hydrofluoric acid, HF, or ammonium bifluoride, ABF, on various properties of glazed ceramic tiles. The impact of these surface changes on the properties of the glazed ceramic tiles is discussed for each treatment separately.

Chemical treatments based on hydrofluoric acid

The HF treatment solubilises part of the glaze and leaves micrometer holes at the surface of the glaze. The chemical composition of the treated glaze does not change but the presence of the holes makes it more fragile and more difficult to clean. Hence, daily traffic and routine floor maintenance should degrade the glaze of HF tiles at a faster rate than the untreated tiles. The holes are fairly deep so that the glaze does not completely recover its gloss as the treatment wears out.

The HF treatment increases only slightly the average roughness and dry friction of the glazed tiles. The treated tiles have a higher resistance to aquaplaning than the untreated tiles but the magnitude of the effect remains relatively small. When the amount of water on the tile is
over the aquaplaning threshold, the wet friction of the HF tiles is a little higher than that of the untreated tiles but still quite low (µwet ~ 0.1). Finally, the slight increase of the average roughness, dry and wet frictions and aquaplaning threshold disappears as the treated tiles wear out.

Chemical treatments based on ammonium bifluoride

The submicron particles deposited on the glazed tiles during the ABF treatment are rich in fluorine and they do not adhere very strongly to the glaze. The treated tiles are more difficult to clean, probably because oils and fats can be trapped into the layer of submicron particles. To counteract that effect, more aggressive floor cleaning methods and detergents are needed, thus accelerating the maintenance wear of the ABF tiles.

The ABF treatment does not increase significantly the average roughness and the dry friction and it slightly increases the resistance to aquaplaning. However, the wet friction remains quite low.

The elimination of the layer of submicron particles exposes a glazed surface very similar to that of the BLANK tiles. Thus, one can expect that the tiles will recover their surface properties as the ABF treatment wears out.

General discussion

Most manufacturers of chemical treatments require that the floor is properly cleaned with a degreaser before the application of the treatment. Interestingly, it was demonstrated (Quirion et al., 2007) that improving the floor cleaning procedure often results in the improvement of the slip resistance. Thus, trying to improve the floor cleaning procedure should always be the first step. And if that does not improve the slip resistance of the floor, then one could think of applying a chemical treatment.

The ABF treatment is less aggressive because it does not alter significantly the glaze. However, its impact on the slip resistance is marginal and it wears out rapidly. That observation is in accordance with the frequent usage recommended by the manufacturers. In that sense, the
ABF solutions resemble more to cleaning solutions than to chemical treatments.

That is not the case for the HF treatments. First, the products contain HF, which is a hazardous chemical that should be used with caution. The improvement of the slip resistance is better than that observed with ABF but it remains fairly small and it disappears as the treated tiles wear out.

Because of its negative impact on the resistance to wear and ease of maintenance and its rather small impact on the slip resistance, treating glazed ceramic tiles with HF solutions should be considered as a short term option for old and worn ceramic floorings that have to be replaced anyway.

It is important to stress that the above observations are based on the results obtained using the experimental procedures and the glazed tiles described in this article. It is possible that changes in the application procedure of the chemical treatments or the use of different glazed tiles would lead to different results.

Conclusions

This investigation deals with the impact of one chemical treatment based on an aqueous solution of hydrofluoric acid, HF, or ammonium bifluoride, ABF, on various properties of glazed ceramic tiles.

The HF treatment:

  • solubilises part of the glaze resulting in micrometric
    holes at the surface but has little impact on the
    average roughness of the tiles,
  • does not affect significantly the chemical
    composition of glaze,
  • reduces significantly the gloss of the glaze,
  • makes the glaze more fragile so that it becomes
    more sensitive to abrasive and maintenance wear,
  • makes the tiles more difficult to clean,
  • increases slightly the aquaplaning threshold and the
    wet friction.

The ABF treatment:

  • leaves a layer of submicron particles rich in fluorine
    that is easily removed from the surface of the glaze
    by abrasive wear,
  • reduces significantly the gloss of the glaze but the
    gloss increases again as the layer of submicron
    particles wears out,
  • has little impact on the average roughness of the tiles,
  • does not seem to make the remaining glaze more
    fragile,
  • makes the tiles more difficult to clean,
  • has little impact on the aquaplaning threshold and
    the wet friction.

Overall, the slight increase in slip resistance observed after one treatment with an HF solution is not worth the deterioration of the glaze that makes it more fragile and more difficult to clean. These treatments should be applied only as a temporary solution on old glazed tiles.
The ABF treatments seem to have little impact on the slip resistance and the sensitivity to wear of the glaze ceramic tiles.

The results obtained are for a given application procedure of the treatments on given glazed ceramic tiles. Different results leading to different conclusions could probably be observed with different application procedures and different tiles.

Acknowledgements

This investigation was supported by the Institut de Recherche Robert-Sauvé en santé et en sécurité au travail under grant 099-179.

References

Bowman R, Strautins C J, Westgate P and Quick G W (2002). Implications for the development of slipresistance standards arising from rank comparisons of friction-test results obtained using different walkwaysafety tribometers under various conditions. In ASTM STP 1424, Metrology of pedestrian locomotion and slip resistance, 112-136.

Chang W R, Kim I J, Manning D P and Bunterngchit Y (2001). The role of surface roughness in the measurement of slipperiness. Ergonomics, 44(13), 1200-1216.

Chang W R, Hirvonen M and Grönqvist R (2004). The effect of cut-off length on surface roughness parameters and their correlation with transition friction. Safety Science, 42, 755-769.

Cholet C, Salimbeni E and Vetter F (2000). Glissance des revêtements de sol: Étude expérimentale. Cahier du CSTB No. 3234, Livraison 411.

Di Pilla S (2000). Slip resistance treatment: Study 2000, ESIS risk control services, Report available online at URL: www.esis.com/InTheNews/ESISSRT2000-0600.pdf [Accessed 18/07/09].

Fang Q, Sidky P S and Hocking M G (1997). The effect of corrosion and erosion on ceramic materials. Corrosion Science 39, 511-527.

Grönqvist R, Hirvonen M and Skyttä E (1992). Countermeasures against floor slipperiness in the food industry. In Advances in Industrial Ergonomics and Safety IV, S. Kumar (ed), Taylor and Francis, 989-996.

Grönqvist R, Matz S and Hirvonen M (2003). Dynamic friction in the shoe-floor interface with respect to floor surface roughness and contamination. In Proceedings of the XV Congress of the International Ergonomics Association, Seoul Korea.

Hupa L, Bergman R, Fröberg L, Vane-Tempest S, Hupa M, Kronberg, Pesonen Leinonen, E, Sjöberg, A M (2005). Chemical resistance and cleanability of glazed surfaces. Surface Science, 584, 113-118.

Lee J C, Song H J, Park M J and Shin H I (2001). Pore development in alumino-silicate fibres treated in hydrofluoric acid solutions, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 187-188, 399-403.

Lemon P and Worth J (2008). Assessing slipperiness: How inspectors do it! Contract Flooring Journal (CFJ), July 2008, 32.

Loo-Morrey M (2007). A study of the slip characteristics of natural and manmade stone flooring materials, Health and Safety Executive, HSE Research Report RR529.

Loo-Morrey M (2006). Ramp testing natural and man made stone floors – Final Report, Health and Safety laboratory, HSL Report No HSL/2006/47.

Luo X P, Silikas N, Allaf M, Wilson N H F and Watts D C (2001). AFM and SEM study of the effects of etching on IPS-Empress 2TM dental ceramic. Surface Science, 491, 388-394.

Quirion F (2004). Optimal cleaning for safer floors, Contemporary Ergonomics 2004, CRC Press, Ed. Paul T. McCabe, p : 28-32. For a detailed description of the cleaning methods used in our laboratory please consult the following report: Massicotte A, Boudrias S and

Quirion F (2000) Conditions optimales d’utilisation des nettoyants à planchers : une approche globale, Rapport de recherche de l’IRSST No R-258. Available online at www.irsst.qc.ca/fr/_publicationirsst_783.html [Accessed 18/07/9].

Quirion F and Poirier P (2005). The impact of heel material, Detergent and Acid Etch on the Risk of Slipping and Sliding on Wet Floorings. In Proceedings of the XIX Annual ISOES Conference in Las Vegas, Lockhart T R S and Fernandez J E (Eds), 501-507.

Quirion F and Poirier P (2006). Aquaplaning threshold and wet friction of various floorings. Proceedings IEA2006 Congress, Eds. Pikaar R N, Koningseld E A P and Settels P J M, Elsevier Ltd.

Quirion F and Poirier P (2007). The Falling Plate Method: Evaluating the risk of aquaplaning on wet floor. Journal of Environmental Health Research. 6(2), 71-80.

Quirion F, Poirier P and Lehane P (2007). Optimal cleaning to prevent slippery floors in restaurants. Journal of Environmental Health Research, 6(1), 25-33.

Rabinowicz E (1995). Friction and Wear of Materials. 2nd edition, John Wiley & Sons.