In Situ Study of Road Marking Durability Using Glass Microbeads and Antiskid Aggregates as Drop-On Materials

: Road markings play an important role in road safety because they provide signiﬁcant information to drivers about the road. For that reason, they must be replaced when they are not correctly perceived by road users. To analyse which are the main factors that affect road marking perception over time, a test section was designed in a two-lane rural highway, running actual trafﬁc over 18 different types of markings fabricated with different combinations of drop-on materials. Chromatic coordinates, luminance, and retroreﬂectivity of each sample were measured during 18 months in order to study their evolution over time. The results obtained show different behaviours depending on the aggregates and application method used. An increment of the durability has been observed with the use of different layers and mixtures of glass microbeads with different sizes.


Introduction
Until now, four main engineering treatments have generally been carried out in order to improve road safety: (i) highway updating; (ii) road signing refurbishment; (iii) repainting of road markings; and (iv) pavement resurfacing. Updating the horizontal or vertical signage is the most affordable option and, in addition to this, road safety increases up to 35% [1].
Road markings provide continuous information to road users about roadway alignment and vehicle positioning [2]. Whether lines or pictograms are used, both should transmit an understandable message for any user. In rural highways, where the traffic average speed is high, it is especially important that this kind of signs is clear and understandable [3].
Road marking should meet certain standards of durability, resistance, and visibility [4]. The service threshold is established according to drivers' safety needs and traffic requirements, and for that reason it is requested [5]: • Defining road marking's essential features for that they achieve their function, from the users' point of view. • Setting the minimum acceptable value for these features, according to traffic circumstances and the intended objectives.
To ensure a good road marking quality, it is very important to make a proper selection of the materials to be used, for which we should consider economic, environmental, application, and use

Test Section
For this piece of research, an in situ test section was set on the CV-904 rural two-lane highway from Crevillente to Catral, located at the existing intersection on 4 + 400 kilometric point (KP), in the municipality of Catral, province of Alicante, Spain (see Figure 1). There is virtually no traffic diversion in this junction because this section is only connected with minor rural paths, so it guarantees a continuous traffic flow along the testing samples (transversal stripes). In addition, speed reduction of passing vehicles is negligible, preventing skid marks that could pollute the samples, altering the wearing process. Furthermore, placing the test section in a crossing makes easier to create a third virtual traffic lane to divert the traffic while making or measuring the samples. The tests were performed under actual traffic and weather conditions, which provide more accurate results than those obtained in laboratory simulations. In the technical datasheet of the traffic count station ID 904020, located at the CV-904 highway, covering traffic flow from 1 + 300 KP to 7 + 000 KP [12], we can observe that traffic volume is balanced between the two lanes, approximately 50% in direction to Catral (ascending) and 50% in direction to Crevillente (descending). Table 1 shows that that traffic flow is regular over time, with Annual Average Daily Traffic (AADT) values around 4000 vehicles/day, and a decreasing percentage of trucks (PT). We also notice (see Table 2) that the average traffic flow rates during working days are greater than on weekends, which means that it is a rural road with a predominant mobility function. Annual Average Daily Truck Traffic (AADTT) remains stable throughout the year, being less intense on weekends as well.

Materials and Application Systems
We have analysed different compositions of road marking materials and application systems in order to study their damage over time.

Base Material
Base material used in all testing samples is white styrene acrylic painting. Technical specifications are described in Table 3.  In the technical datasheet of the traffic count station ID 904020, located at the CV-904 highway, covering traffic flow from 1 + 300 KP to 7 + 000 KP [12], we can observe that traffic volume is balanced between the two lanes, approximately 50% in direction to Catral (ascending) and 50% in direction to Crevillente (descending). Table 1 shows that that traffic flow is regular over time, with Annual Average Daily Traffic (AADT) values around 4000 vehicles/day, and a decreasing percentage of trucks (P T ). We also notice (see Table 2) that the average traffic flow rates during working days are greater than on weekends, which means that it is a rural road with a predominant mobility function. Annual Average Daily Truck Traffic (AADTT) remains stable throughout the year, being less intense on weekends as well.

Materials and Application Systems
We have analysed different compositions of road marking materials and application systems in order to study their damage over time.

Base Material
Base material used in all testing samples is white styrene acrylic painting. Technical specifications are described in Table 3.      We have used white marble sand (Macael 400) as non-transparent antiskid aggregate and sodium-calcium glass particles (Glass 600) as transparent antiskid aggregate. Macael 400 aggregate properties are described in Table 9. Transparent antiskid Glass 600 calcium and sodium granules have irregular shapes and sharp edges to be mixed with glass microbeads, in order to generate roughness. Technical specifications are shown in Table 10.

Application Systems
Two different types of application systems were used in this study: • "Monolayer" or single layer system (AS1): it is the traditional system of application. First, the base material is sprayed and immediately after, mixture of drop-on material is projected (it can be just glass microbeads or a mixture of microbeads with transparent or non-transparent antiskid aggregates). • "Bilayer" or double layer system (AS2): the first step is to create a complete AS1 layer, as explained before; once it is dry, another identical layer is applied over the previous one.

Test Section Design and Sample Fabrication
Following the indications of the Spanish regulations on road surface marking design, included in the Ministerial Order FOM/3053/2008 [26], we sketched the placement of the testing samples on the pavement with chalk and adhesive tape. As shown in Figure 2, stripes were painted in pairs (except the first one, that is just one stripe), spaced 15 m between groups, except for the single stripe, that is 20 m apart. The last stripe is 50 m away from the intersection, considered as a conflict point.

Application Systems
Two different types of application systems were used in this study:  "Monolayer" or single layer system (AS1): it is the traditional system of application. First, the base material is sprayed and immediately after, mixture of drop-on material is projected (it can be just glass microbeads or a mixture of microbeads with transparent or non-transparent antiskid aggregates).  "Bilayer" or double layer system (AS2): the first step is to create a complete AS1 layer, as explained before; once it is dry, another identical layer is applied over the previous one.

Test Section Design and Sample Fabrication
Following the indications of the Spanish regulations on road surface marking design, included in the Ministerial Order FOM/3053/2008 [26], we sketched the placement of the testing samples on the pavement with chalk and adhesive tape. As shown in Figure 2, stripes were painted in pairs (except the first one, that is just one stripe), spaced 15 m between groups, except for the single stripe, that is 20 m apart. The last stripe is 50 m away from the intersection, considered as a conflict point. There are nine stripes per traffic direction. In addition, every stripe is divided into two testing samples covering half of the lane, so that both are treaded by a tyre at the same time, totalling 18 testing samples with different combinations of road marking materials (see Table 11). The scheme was repeated in the opposite direction, obtaining a total of 36 testing samples, labelled as shown in  There are nine stripes per traffic direction. In addition, every stripe is divided into two testing samples covering half of the lane, so that both are treaded by a tyre at the same time, totalling 18 testing samples with different combinations of road marking materials (see Table 11). The scheme was repeated in the opposite direction, obtaining a total of 36 testing samples, labelled as shown in Figure 3. As AADT is similar in both directions, it is possible to compare and contrast the results obtained from each testing sample.  While fabricating each of in situ testing samples, we obtained specimens in a 15 cm × 30 cm metal plate placed on the pavement while painting to examine them under a laboratory microscope. Figure  4 shows the sample fabrication process. While fabricating each of in situ testing samples, we obtained specimens in a 15 cm × 30 cm metal plate placed on the pavement while painting to examine them under a laboratory microscope. Figure 4 shows the sample fabrication process. While fabricating each of in situ testing samples, we obtained specimens in a 15 cm × 30 cm metal plate placed on the pavement while painting to examine them under a laboratory microscope. Figure  4 shows the sample fabrication process.

Tests Performed
We have tested the fundamental parameters of the in situ testing samples: (i) luminance factor β and colour (chromatic coordinates X,Y) for daytime visibility, using a Spectro-Guide 45/0 gloss spectrophotometer (BYK GmbH, Wesel, Germany); (ii) retroreflective luminance coefficient (retroreflectivity) in dry (RL) and wet conditions (RW) for nighttime visibility, using a ZRM 6013 retroreflectometer (Zehntner GmbH, Sissach, Switzerland); and (iii) skid resistance coefficient (SRT) using the TRRL pendulum (Controls SpA, Milan, Italy). All the tests were performed inside the wheel path area. As standards require, each factor was measured a minimum of three times in every testing sample per driving direction, and subsequently the average value is calculated. Therefore, for each kind of mixture tested (S1-S18), at least six values for each parameter and period of time assessed were determined.
For the durability study, we have measured those parameters at different times: just produced, after 1, 6, 12, and 18 months, for a quantitative and qualitative analysis supported by data charts, pictures of in situ testing samples and laboratory specimens using a Leica EZ4D microscope (Leica Microsystems GmbH, Wetzlar, Germany). The weather conditions during data collection are shown in Figure 5.

Tests Performed
We have tested the fundamental parameters of the in situ testing samples: (i) luminance factor β and colour (chromatic coordinates X,Y) for daytime visibility, using a Spectro-Guide 45/0 gloss spectrophotometer (BYK GmbH, Wesel, Germany); (ii) retroreflective luminance coefficient (retroreflectivity) in dry (R L ) and wet conditions (R W ) for nighttime visibility, using a ZRM 6013 retroreflectometer (Zehntner GmbH, Sissach, Switzerland); and (iii) skid resistance coefficient (SRT) using the TRRL pendulum (Controls SpA, Milan, Italy). All the tests were performed inside the wheel path area. As standards require, each factor was measured a minimum of three times in every testing sample per driving direction, and subsequently the average value is calculated. Therefore, for each kind of mixture tested (S1-S18), at least six values for each parameter and period of time assessed were determined.
For the durability study, we have measured those parameters at different times: just produced, after 1, 6, 12, and 18 months, for a quantitative and qualitative analysis supported by data charts, pictures of in situ testing samples and laboratory specimens using a Leica EZ4D microscope (Leica Microsystems GmbH, Wetzlar, Germany). The weather conditions during data collection are shown in Figure 5.  Figure 6 shows the average results of the chromatic coordinates (X,Y) measured on the 18 in situ testing samples over time. The results obtained are all inside the colour polygon defined by its vertices for the white permanent colour, hence this parameter is not a limiting condition according to the aspects studied in this paper.    Figure 6 shows the average results of the chromatic coordinates (X,Y) measured on the 18 in situ testing samples over time. The results obtained are all inside the colour polygon defined by its vertices for the white permanent colour, hence this parameter is not a limiting condition according to the aspects studied in this paper.  Figure 6 shows the average results of the chromatic coordinates (X,Y) measured on the 18 in situ testing samples over time. The results obtained are all inside the colour polygon defined by its vertices for the white permanent colour, hence this parameter is not a limiting condition according to the aspects studied in this paper.   The luminance factor evolution over time is shown in the Figure 7. An analysis of the collected data shows that between the sample groups with AS1 application system (monolayer), the best performance is obtained by the group G1 samples, that only use glass microbeads as drop-on material, followed by group G3, whose samples contain glass microbeads (80%) and glass 600 (20%), and the last one is G2, that has a mixture of glass beads and non-transparent antiskid aggregates as drop-on materials. We should underline that group G4 (bilayer) results follow a similar trend, but getting better outcomes than the equivalent AS1 group with the same kind of drop-on aggregates. The luminance factor evolution over time is shown in the Figure 7. An analysis of the collected data shows that between the sample groups with AS1 application system (monolayer), the best performance is obtained by the group G1 samples, that only use glass microbeads as drop-on material, followed by group G3, whose samples contain glass microbeads (80%) and glass 600 (20%), and the last one is G2, that has a mixture of glass beads and non-transparent antiskid aggregates as drop-on materials. We should underline that group G4 (bilayer) results follow a similar trend, but getting better outcomes than the equivalent AS1 group with the same kind of drop-on aggregates. From the qualitative analysis of the in situ pictures taken from every testing sample over time, we notice that larger glass microbeads become detached first, leaving craters which are filled by dust, dirt, and rubber from tyres, making the road markings darker and losing its luminance factor, as observed in Figure 8a. However, we realize that glass microbeads with an adhesive coating (as in sample S4) are detached with more difficulty, even if they are in bigger sizes. Therefore, a lesser number of craters are created, and the luminance stays better over time, as seen in Figure 8b. 0.75 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 From the qualitative analysis of the in situ pictures taken from every testing sample over time, we notice that larger glass microbeads become detached first, leaving craters which are filled by dust, dirt, and rubber from tyres, making the road markings darker and losing its luminance factor, as observed in Figure 8a. However, we realize that glass microbeads with an adhesive coating (as in sample S4) are detached with more difficulty, even if they are in bigger sizes. Therefore, a lesser number of craters are created, and the luminance stays better over time, as seen in Figure 8b. The luminance factor evolution over time is shown in the Figure 7. An analysis of the collected data shows that between the sample groups with AS1 application system (monolayer), the best performance is obtained by the group G1 samples, that only use glass microbeads as drop-on material, followed by group G3, whose samples contain glass microbeads (80%) and glass 600 (20%), and the last one is G2, that has a mixture of glass beads and non-transparent antiskid aggregates as drop-on materials. We should underline that group G4 (bilayer) results follow a similar trend, but getting better outcomes than the equivalent AS1 group with the same kind of drop-on aggregates. From the qualitative analysis of the in situ pictures taken from every testing sample over time, we notice that larger glass microbeads become detached first, leaving craters which are filled by dust, dirt, and rubber from tyres, making the road markings darker and losing its luminance factor, as observed in Figure 8a. However, we realize that glass microbeads with an adhesive coating (as in sample S4) are detached with more difficulty, even if they are in bigger sizes. Therefore, a lesser number of craters are created, and the luminance stays better over time, as seen in Figure 8b. 0.75 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 Regarding G2 and G3, apart from the craters caused by the detached microbeads, the antiskid aggregates retain dust and rubber particles, which also contributes to reduce luminance ( Figure 9). Regarding G2 and G3, apart from the craters caused by the detached microbeads, the antiskid aggregates retain dust and rubber particles, which also contributes to reduce luminance ( Figure 9).
From the global analysis of the luminance factor in all the 18 testing samples, we notice that they fulfil the minimum acceptance standard of β = 0.30 established by the Spanish regulations [27,28], being that some testing samples slightly under this value in at 18 months age, as in S12 and S17, but could be still acceptable. That is why we consider that the luminance factor β is not the main limiting factor for durability. Figure 10 shows the evolution of the retroreflectivity in dry conditions, RL. Samples from groups G2 and G3 reached lower RL values than those from G1, which only has glass microbeads as drop-on material. From the qualitative analysis of the pictures taken in laboratory, we realized that in G2 the nontransparent antiskid particles were taking up spaces that could be covered with glass microbeads. 700 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18
From the global analysis of the luminance factor in all the 18 testing samples, we notice that they fulfil the minimum acceptance standard of β = 0.30 established by the Spanish regulations [27,28], being that some testing samples slightly under this value in at 18 months age, as in S12 and S17, but could be still acceptable. That is why we consider that the luminance factor β is not the main limiting factor for durability. Figure 10 shows the evolution of the retroreflectivity in dry conditions, R L . Samples from groups G2 and G3 reached lower R L values than those from G1, which only has glass microbeads as drop-on material. Regarding G2 and G3, apart from the craters caused by the detached microbeads, the antiskid aggregates retain dust and rubber particles, which also contributes to reduce luminance (Figure 9).
From the global analysis of the luminance factor in all the 18 testing samples, we notice that they fulfil the minimum acceptance standard of β = 0.30 established by the Spanish regulations [27,28], being that some testing samples slightly under this value in at 18 months age, as in S12 and S17, but could be still acceptable. That is why we consider that the luminance factor β is not the main limiting factor for durability. Figure 10 shows the evolution of the retroreflectivity in dry conditions, RL. Samples from groups G2 and G3 reached lower RL values than those from G1, which only has glass microbeads as drop-on material. From the qualitative analysis of the pictures taken in laboratory, we realized that in G2 the nontransparent antiskid particles were taking up spaces that could be covered with glass microbeads. 700 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 From the qualitative analysis of the pictures taken in laboratory, we realized that in G2 the non-transparent antiskid particles were taking up spaces that could be covered with glass microbeads. Consequently, these opaque particles do not let the light go through them. Moreover, they cast shadows to the glass microbeads placed just behind them ( Figure 11). That is the reason why this sample group has a considerable loss of retroreflectivity compared to G1. Consequently, these opaque particles do not let the light go through them. Moreover, they cast shadows to the glass microbeads placed just behind them ( Figure 11). That is the reason why this sample group has a considerable loss of retroreflectivity compared to G1. Something similar happens with G3 samples. As they have mixed glass beads plus transparent antiskid aggregates, the last ones let part of the light go through them and be retroreflected ( Figure  12). That is why G3 samples obtain better results than in G2, but not as good as in G1. Following the acceptance criteria [27,28], we find that from the eighteenth month, some G2 testing samples have a RL value below Spanish standard requirements (100 mcd/m 2 /lx). For G3, testing samples S13 and S14 are those whose perform best along the different tests, achieving the best results during service life.

Nighttime Visibility
From both qualitative and quantitative analysis, we can conclude that initial retroreflectivity results improve when microbeads used have a greater size. On the other hand, the smaller the glass beads are, the more they embed in the base material, getting less retroreflectivity at first. However, as they detach with more difficulty, retroreflectivity is preserved better along time, extending road marking service life. The samples fabricated with medium-sized glass microbeads (425-850 µ m) obtained good results initially and over time. Another important aspect to consider is that glass microbeads with an adhesive coating are detached from the paint with more difficulty, extending road marking performance.
In the G4 testing samples group, a retroreflection increase is observed after twelve months of service. This behaviour could be explained because the upper paint layer is progressively worn out by traffic, revealing the microbeads embedded in the lower layer (Figure 13), improving retroreflectivity. This process starts from the sixth month of service and extends the service life of bilayer road markings, reducing retroreflectivity decay over time. Something similar happens with G3 samples. As they have mixed glass beads plus transparent antiskid aggregates, the last ones let part of the light go through them and be retroreflected ( Figure 12). That is why G3 samples obtain better results than in G2, but not as good as in G1. Consequently, these opaque particles do not let the light go through them. Moreover, they cast shadows to the glass microbeads placed just behind them ( Figure 11). That is the reason why this sample group has a considerable loss of retroreflectivity compared to G1. Something similar happens with G3 samples. As they have mixed glass beads plus transparent antiskid aggregates, the last ones let part of the light go through them and be retroreflected ( Figure  12). That is why G3 samples obtain better results than in G2, but not as good as in G1. Following the acceptance criteria [27,28], we find that from the eighteenth month, some G2 testing samples have a RL value below Spanish standard requirements (100 mcd/m 2 /lx). For G3, testing samples S13 and S14 are those whose perform best along the different tests, achieving the best results during service life.
From both qualitative and quantitative analysis, we can conclude that initial retroreflectivity results improve when microbeads used have a greater size. On the other hand, the smaller the glass beads are, the more they embed in the base material, getting less retroreflectivity at first. However, as they detach with more difficulty, retroreflectivity is preserved better along time, extending road marking service life. The samples fabricated with medium-sized glass microbeads (425-850 µ m) obtained good results initially and over time. Another important aspect to consider is that glass microbeads with an adhesive coating are detached from the paint with more difficulty, extending road marking performance.
In the G4 testing samples group, a retroreflection increase is observed after twelve months of service. This behaviour could be explained because the upper paint layer is progressively worn out by traffic, revealing the microbeads embedded in the lower layer (Figure 13), improving retroreflectivity. This process starts from the sixth month of service and extends the service life of bilayer road markings, reducing retroreflectivity decay over time. Following the acceptance criteria [27,28], we find that from the eighteenth month, some G2 testing samples have a R L value below Spanish standard requirements (100 mcd/m 2 /lx). For G3, testing samples S13 and S14 are those whose perform best along the different tests, achieving the best results during service life.
From both qualitative and quantitative analysis, we can conclude that initial retroreflectivity results improve when microbeads used have a greater size. On the other hand, the smaller the glass beads are, the more they embed in the base material, getting less retroreflectivity at first. However, as they detach with more difficulty, retroreflectivity is preserved better along time, extending road marking service life. The samples fabricated with medium-sized glass microbeads (425-850 µm) obtained good results initially and over time. Another important aspect to consider is that glass microbeads with an adhesive coating are detached from the paint with more difficulty, extending road marking performance.
In the G4 testing samples group, a retroreflection increase is observed after twelve months of service. This behaviour could be explained because the upper paint layer is progressively worn out by traffic, revealing the microbeads embedded in the lower layer (Figure 13), improving retroreflectivity. This process starts from the sixth month of service and extends the service life of bilayer road markings, reducing retroreflectivity decay over time.  Figure 13. S18 sample in situ detail after 12 months of service. Emergence of glass microbeads from the bottom layer is observed. Figure 14 shows retroreflectivity evolution over time in wet conditions. From this figure, it can be confirmed that from six months old, none of the G1, G2, and G3 testing samples reach the minimum value required by Spanish regulations [27,28]. We also notice that samples with bigger microbeads get better RW results than those with smaller ones, because they protrude over the water surface. Furthermore, as bigger microbeads are detached easier, RW drops quickly over time, as seen on Figure 14. As happened with RL values, RW increase from the sixth to the twelfth month is observed in G4 samples, due to the same reason explained before. This effect makes that these three samples achieve the minimum RW values established by the Spanish regulations [27,28]. Figure 15 shows skid resistance results of the 18 testing samples over time. Results were the opposite of those obtained for β, RL, and RW. So, the best group for SRT value is G2, followed by G3 150 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18

G1
G2 G3 G4 Figure 13. S18 sample in situ detail after 12 months of service. Emergence of glass microbeads from the bottom layer is observed. Figure 14 shows retroreflectivity evolution over time in wet conditions. From this figure, it can be confirmed that from six months old, none of the G1, G2, and G3 testing samples reach the minimum value required by Spanish regulations [27,28]. We also notice that samples with bigger microbeads get better R W results than those with smaller ones, because they protrude over the water surface. Furthermore, as bigger microbeads are detached easier, R W drops quickly over time, as seen on Figure 14.  Figure 14 shows retroreflectivity evolution over time in wet conditions. From this figure, it can be confirmed that from six months old, none of the G1, G2, and G3 testing samples reach the minimum value required by Spanish regulations [27,28]. We also notice that samples with bigger microbeads get better RW results than those with smaller ones, because they protrude over the water surface. Furthermore, as bigger microbeads are detached easier, RW drops quickly over time, as seen on Figure 14. As happened with RL values, RW increase from the sixth to the twelfth month is observed in G4 samples, due to the same reason explained before. This effect makes that these three samples achieve the minimum RW values established by the Spanish regulations [27,28]. Figure 15 shows skid resistance results of the 18 testing samples over time. Results were the opposite of those obtained for β, RL, and RW. So, the best group for SRT value is G2, followed by G3 150 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 As happened with R L values, R W increase from the sixth to the twelfth month is observed in G4 samples, due to the same reason explained before. This effect makes that these three samples achieve the minimum R W values established by the Spanish regulations [27,28]. Figure 15 shows skid resistance results of the 18 testing samples over time. Results were the opposite of those obtained for β, R L , and R W . So, the best group for SRT value is G2, followed by G3 and finally G1. Therefore, non-transparent aggregates have better roughness results than transparent ones.  For G4, testing samples follow same the trend that those testing samples that have the same drop-on materials but are applied using AS1 system. Sample S17 shows the best results in this group, followed by S18 and S16.

Conclusions
All sample groups reach the luminance factor acceptance requirements over time. Thus, luminance is not the most determining factor in order to choose the optimum combination of materials for road marking selection amongst all the samples studied. The craters left on the base material by the detached drop-on aggregates are filled with dust and rubber particles, darkening the road markings and making them lose luminance and therefore daytime visibility over time.
Regarding the chromatic coordinates, all the testing samples analysed are inside the polygon that defines permanent white colour during the 18-month testing period. As happens with the luminance factor, colour is not determinant to find the optimum drop-on mixture.
Retroreflectivity of road marking samples made of just microglass beads as drop-on material is greater than those which have a mixture of microbeads and transparent antiskid aggregates, and these ones obtain better results than those made with opaque antiskid aggregates.
Macael 400 particles take the space that could have been covered by glass microbeads and block the incident light, which reduces RL values in G2 samples. Glass 600 particles also fill that space, reducing RL value; however, as they are transparent, they let part of the light go through them, obtaining slightly better RL values in the G3 sample group.
As the size and quality (less impurities and defective beads) of microbeads increase, better initial RL values are obtained. On the other hand, largest microbeads are detached from the painting easier and faster than the smaller ones, because the first ones are poorly embedded in the base material, losing retroreflection capacity over time. Moreover, microcraters left by missing beads are posteriorly filled by dirt, dust, and rubber from tyres, causing also extra luminance loss. Microbeads' surface coating gives them extra adhesion to remain embedded in the base material, getting really good results, even in the largest diameters. The best RW values are obtained with the largest microbead sizes, because their upper part remains above the water and continues providing some retroreflection capacity. 65 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 For G4, testing samples follow same the trend that those testing samples that have the same drop-on materials but are applied using AS1 system. Sample S17 shows the best results in this group, followed by S18 and S16.

Conclusions
All sample groups reach the luminance factor acceptance requirements over time. Thus, luminance is not the most determining factor in order to choose the optimum combination of materials for road marking selection amongst all the samples studied. The craters left on the base material by the detached drop-on aggregates are filled with dust and rubber particles, darkening the road markings and making them lose luminance and therefore daytime visibility over time.
Regarding the chromatic coordinates, all the testing samples analysed are inside the polygon that defines permanent white colour during the 18-month testing period. As happens with the luminance factor, colour is not determinant to find the optimum drop-on mixture.
Retroreflectivity of road marking samples made of just microglass beads as drop-on material is greater than those which have a mixture of microbeads and transparent antiskid aggregates, and these ones obtain better results than those made with opaque antiskid aggregates.
Macael 400 particles take the space that could have been covered by glass microbeads and block the incident light, which reduces R L values in G2 samples. Glass 600 particles also fill that space, reducing R L value; however, as they are transparent, they let part of the light go through them, obtaining slightly better R L values in the G3 sample group.
As the size and quality (less impurities and defective beads) of microbeads increase, better initial R L values are obtained. On the other hand, largest microbeads are detached from the painting easier and faster than the smaller ones, because the first ones are poorly embedded in the base material, losing retroreflection capacity over time. Moreover, microcraters left by missing beads are posteriorly filled by dirt, dust, and rubber from tyres, causing also extra luminance loss. Microbeads' surface coating gives them extra adhesion to remain embedded in the base material, getting really good results, even in the largest diameters. The best R W values are obtained with the largest microbead sizes, because their upper part remains above the water and continues providing some retroreflection capacity.
Bilayer road markings (G4 samples) results follow the trend of their equivalent monolayer sample. It should be pointed out that what we achieve with this application system is to decrease the wearing of road markings from the first month and to keep constant retroreflectivity values from the sixth month to the end of the testing period, even recovering this feature from the sixth to the twelfth month of service because of the emergence of the bottom layer after the wearing of the upper layer.
The relation between skid resistance is opposite to daytime and nighttime visibility: road markings having better SRT results (G2 sample group) obtain lower values of luminance and retroreflectivity. That is because rougher mixtures capture a greater amount of dirt and rubber particles, darkening the markings surface.