ASSESSMENT OF WALL SALINITY IN THE SELECTION OF RENOVATION PLASTER SYSTEMS

This article presents issues related to assessing the degree of wall salinity to select plaster systems for renovations of damp building walls. The most commonly used salt concentration tests pose many difficulties. If used uncritically, they risk failing to select the right system or its incorrect make. The accuracy of the colorimetric method for testing chloride, nitrate, and sulphate content was analysed to exemplify the magnitude of the problem. Both multi-salt solutions of known concentrations and unknown composition extracted from drillings in the walls of a historical facility were examined. A comparative methodology using ion chromatography as a standard was employed in the research. The analytical methods and the selected modules of the „Statistica” software were used to analyse data and present the results. The colorimetric method has been shown to distort salt concentration values, posing a risk of unsuccessful repair work on high-salinity walls. A method for determining the correction reducing the measurement error has been proposed. The factors affecting the error were also mentioned. Attention has also been drawn to the resolution and application of a method with a correct concentration range intended to improve work efficiency and optimize the costs incurred in renovating the salty wall.


Introduction
When in operation, each facility is continuously exposed to adverse outdoor conditions. If the walls are not adequately protected, they may become wet and thus salty, as the transported water is, in fact, always a salt solution. Depending on the magnitude of the phenomenon, the salts cause relatively great damage to building materials due to crystallization. This problem has been known for years. In 420 B.C., Herodotus (HERODOTUS 420BC), an ancient Greek historian, said, "I observed ... that salt exuded from the soil to such an extent as even to injure the pyramids". In modern times, information on this issue appeared already in 1895 (LUQUER 1895). Doehne estimated that in 2002, there were more than 1800 bibliogra phic entries in the scientific literature on the destruction of porous material caused by salts (DOEHNE 2002). Many of them have been cited in literature extensive review by Doehne, Goudie, Viles, and Charola (DOEHNE 2002, GOUDIE, VILES 1997, CHAROLA 2000 . The impact of the three salts (chlorides, nitrates and sulphates) is assumed to be the most significant for the building material (PAVLÍKOVÁ et al. 2011, PEŘINKOVÁ et al. 2021, WÓJCIK 2006. These are substances covered by instructions and recommendations from industry institutions and manufacturers. For example, the EUREKA EU-1270 programme proposed a salinity grade and damage classification based on the salt type and concentration (Table 1) (STILLHAMMEROVÁ 2006). This classification uses five basic salinity degrees and identifies possible hazards and degradation degrees caused by these salts. According to Table 1 ( STILLHAMMEROVÁ 2006), salinity classes are defined as follows: • Class 0 -low salt concentration (trace amounts), no damage to the wall; • Class 1 -very low salinity of the structure. Structures that are continuous moisture-borne engage in water capillary action (or other unfavourable phenomena) and may get damaged; • Class 2 -average salinity of the structure. The life of the plasters and paintings is slightly reduced; • Class 3 -high salinity of the structure. The life of plaster and paintings is reduced considerably. Despite the use of active vertical isolation, the wall is wet due to the hygroscopic nature of the salt; • Class 4 -extra high salinity. The structure gets destroyed in a very short time.
In the context of damage, this classification is rather general because, in reality, the size of the damage is affected by the structure of the material under analysis. A small quantity of salt can cause extensive damage in material with small pore diameters, and conversely, high concentrations will not result in significant changes in materials with large pore diameters (BLÄUER BÖHM 2005   Knowledge of of salt concentrations is of key importance here (PAVLÍKOVÁ et al. 2011, KONCA et al. 2016, GACZEK, FISZER 2014, KNOP 2016, NOCOŃ 2016. The pores of WTA restoration plasters are hydrophobic, which limits the deposition of salt in the plaster coating. If the substrate exhibits a medium to a high degree of salinity, double layer systems should be used, often with a base plaster storing salts (GACZEK, FISZER 2014) (Table 3). If the salinity of the substrate is incorrectly determined, mainly understated, the volume of the storage layer may be too small, which will cause premature saturation of the system and entail the necessity to rework, which involves significant costs. The situation is similar for the opposite. If concentrations get overestimated, the accepted (exaggerated) thickness of the storage layer puts the investor at risk of incurring unreasonable costs. Consequently, the investor questions the work effectiveness or states an abuse of material and thus refers the case to court. The judicial decisions are made based on the commissioned studies with the highest possible accuracy.
Contractors and architects typically use simple strip tests for salinity diagnostics. Some claim that they are estimates and do not give accurate results. There are no studies in the scientific literature describing the accuracy of these methods. Oberta used the colorimetric method (KOL) to research the capillary zone electrophoresis (CZE) method. The results showed deviations from the actual salt concentrations (Table 4 and   Table 5) (OBERTA 2015).  Table 5 Comparison of test results for concentrations of water-soluble salts obtained using the CZE and KOL method

Concentration of the solution determined with CZE [%]
Sandstone 0.20 (-5) 1 0.12 (-43) 1 0.14 (-26) 1 0.21 (5)  The value in brackets refers to the difference between the results obtained using the KOL and CZE methods 2 The value in brackets refers to the difference between the results obtained using the CZE and KOL methods The CZE method overstated the results by 36-79%. On the contrary, tests using the CZE and KOL methods (Table 5) gave similar values for chlorides. The difference in the results obtained using the KOL and CZE methods was -5-0%. It can thus be assumed that the deviation found in Table 4 for the CZE method is also applicable to the KOL method. For other salts, the differences in results obtained using CZE and KOL were more pronounced. For nitrates, they were -43-11% and -4-26% for sulphates. In both cases, the largest deviations applied to sandstone. Unfortunately, concerning nitrates and sulphates, it is difficult to make similar assumptions as for chlorides, as no comparison was made for CZE with the actual concentrations of these salts. Thus, we do not know what discrepancies (even hypothetical) can be expected when determining nitrate and sulphate concentrations using colorimetry. On the other hand, these parameters are very important for renovation works in salty facilities and should be carefully examined. In this work, the accuracy of colorimetric methods used in repair technology using renovation plasters was analysed for all of the above-mentioned salts.
Borelli described it as highly accurate (BORRELLI 1999). The authors' own studies confirmed this opinion.
The accuracy of the chromatograph used proved very high. For the tests, a chromatograph equipped with a conductometric detector and UV-vis detector with a photodiode array and an ion-exchange column dedicated to separating anions such as chlorides, nitrates and sulphates was used. An external standard was used for determining the so-called calibration curve. It showed a correlation between the value measured by the apparatus and the test substance concentration. It was based on an analysis of several samples of known concentrations and integrals, the so-called peaks on the chromatogram (Figure 1). The results of the studies exhibited a very high degree of correlation. Model fit to the trial through the coefficient of determination R 2 (the closer it is to unity, the better the fit).
When determining calibration curves for chlorides, nitrates and sulphates, the R-square (R2) approximation value was always above 0.9998 (Table 6 and Figure 2). However, for comparing several populations of result, c b a a coefficient of variation was used (RSD -relative statistical deviation) (if the resulting RSD is less than 25%, the results are assumed to be only slightly variable). For the calibration curves, the coefficient of variation of the RF response factor (RF %RSD) were 1-6% (Table 6 and Figure 2), where the RF response factor was defined as the quotient of the peak surface area and the concentration (or vice versa) of the standard solution. Therefore, the curve presented a very good fit to the results in the analysed range. Statistically, all values calculated from the correlation function obtained were subject to a minor error.
It was planned to run tests on two types of samples to determine the accuracy of the colorimetric method.
Tests on samples of unknown concentrations were then performed. The salt solutions were extracted from wall drillings of building facilities. 10 g of material was weighed, poured into 100 mL of demineralized water, stirred for 30 s, and filtered through quantitative hard filter papers (type 390, weight 84 g/m 2 ). Individual samples (drillings) differed by colours, so they were grouped according to them. In this way, seven groups with three samples each were obtained, on average. An attempt was made to determine the discrepancy between the colorimetric method and ion chromatography results with such a population of results. To this end, the results obtained by the IC method were taken as a reference against which the results obtained in the KOL methods were compared. Deviations from the IC method were obtained.

Solutions of Known Concentrations
When testing nitrate concentration, the colorimetric method determined the salt concentration without any error. The colours on the strips matched the standard colours for the individual concentrations ( Figure   3). However, when measuring chloride concentrations with the colorimetric method using titration, errors a. b. c.
1 Concentration above the indicated value, slightly coloured towards the next concentration range (strip field) have occurred. A so-called blind trial (solution free from the test substance) yielded a positive result. In a sample that did not contain any chlorides, the test detected a concentration of 0.0006 %. This is a row of magnitude equal to the test resolution, as concentration is determined by adding drops with a volume equal to 6 mL of reagent (this corresponds to a concentration of 0.0006%). The other deviations resulted from the limitation resulting from constant volume (multiples of 6 mL) (Figure 3). The greatest difficulties were encountered in the colorimetric sulphate test. Even though we prepared concentrations corresponding to the ranges indicated on the strips, the individual parts acquired imprecise colours. They overstated the values, which were difficult to determine precisely because the subsequent colour (indicative of higher concentrations) was only partially coloured ( Figure 3). To the contrary, the ion chromatography method determined the solutions with a high degree of precision within limits given to prepare the calibration solutions ( Figure 3).
The differences (deviations) in the KOL and IC results were in the range -65%÷50%. The discrepancies ranged from -100% to 100% for nitrates, and -60% to 2122% for sulphates ( Figure 4 and Table A1). The  Table 7.

Discussion
In the case under analysis, high accuracy in nitrate and chloride measurements was achieved in the range of known concentrations (assigned to scale for colorimetric method tests). The situation significantly changed when solutions with a natural distribution of concentrations obtained from the facility were taken into account, and a significant deviation in the KOL-IC results was observed. A diminishing trend in changes in deviation values as a function of concentrations was observed for all tests. The sulphate test was particularly distinguishing, whereas the variation in deviations was similar for nitrates and chlorides. The scatter of the deviation values appeared similar. As the concentrations increased, the repeatability of the error value increased as well.
At salinity thresholds (according to WTA), i.e. 0.2% for chlorides and 0.3%, it presented a high level. Less repeatability in error values was obtained at a salinity level of 0.1% for chlorides and 0.5% for sulphates.
Thus, in Table 7, we precisely defined the value of deviations for specific salinity degrees in a way corresponding to error repeatability.
The results obtained using colorimetric methods varied from the conclusions drawn based on the Oberta study. The latter suggested positive deviations exceeding the actual chloride concentrations by more than 36%.
However, in a direct study, negative deviations were obtained for WTA salinity thresholds. In the case of nitrates, the deviations were similar to the chloride test. The largest errors were obtained for the sulphate test.
For this test, taking measurements proved problematic. The measurement was the most ambiguous because the fields in the strips were partially coloured, and the resolution was the smallest of all methods.
Underestimated values obtained with the colorimetric method under investigation may be detrimental to the success of repair work on salty walls. A correction may be applied to the colorimetric test to reduce the measurement error, either using the deviations from Table 7 or by determining salt concentration using more accurate methods. Ion chromatography is a well-proven method. It requires a small amount of extract (approximately 1 cm 3 ). For tests on historical sites (which require repairs most often), this becomes a major advantage since other methods require more material to be collected. The weighing method requires 50 cm 3 of extract obtained by pouring 1 to 5 g of the sample per 100 cm 3 (DOMASŁOWSKI et al. 2011). Machine price is the major disadvantage of the IC method. However, it is increasingly applicable to testing wall salinity (FRANZONI et al. 2011, FRANZONI et al. 2014, FRANZONI, BANDINI 2012, LUBELLI et al. 2018, LUBELLI 2006, GONÇALVES et al. 2006, CAMUFFO 2018, MARAVELAKI-KALAITZAKI et al. 2003, GONÇALVES 2007. However, these colorimetric tests may be a good complement to ion chromatography. Their advantage is that they are dedicated to a specific type of salt. An on-site exam can also be performed. However, when using such tests, attention should be paid to concentration ranges. The upper limit of the Chloride Test in question is 0.2% (200 mg/L). This is merely where mid-range concentrations start, according to WTA. Therefore, it will be impossible to distinguish between average and high salinity and select appropriate components of the renovation system for salty walls.
To be on the safe side, if we roughly assume a very high degree of salinity and it is average, then the costs of the materials used would be unjustifiably high. If, in turn, we attempt to determine salinity when the concentration is out of scale, we would need to keep repeating the test, iteratively approaching the scale through subsequent tests. Samples obtained from drillings taken from the same site must be diluted so that the test covers its range. This can consume large amounts of test material, time and accumulate measurement error due to successive dilutions of the solution. The same is true for matching the scale on the test strips against the salinity classification used by the WTA. In the case analysed, the scale did not match this classification (Table 8). This causes difficulties in accurately determining the salinity degree. Table 8 Comparison of salt concentration ranges in colorimetric tests with the wall salinity classification by WTA no. 2-9-04

Conclusions
The use of colorimetric methods for measuring salt concentrations for repairing salty walls using renovation What is important for the application of restoration plaster technology are the deviations of the result obtained by the colorimetric method, which is at the level of the WTA salinity degree thresholds, i.e. concentrations of 0.1; 0.2; 0.3; 0.5% (in mass). They may serve as a correction for the result obtained with a given colorimetric method provided that such deviation shows there is repeatability for successive concentrations.
The accuracy of the correction in a given concentration range is obtained in direct proportion to the repeatability of the deviation.
Ion exchange chromatography may also be used as the main measurement method and the colorimetric method as a complementary method. It is especially applicable when dealing with a multi-anion solution with peaks overlapping on the chromatogram. Assigning specific anions to the individual tests of the colorimetric method may help resolve doubts concerning the assignment of specific anions to peaks on the chromatogram and accelerate analytical work.

Acknowledgments
I acknowledge any support given by Professor Robert Wójcik, especially for the donation of research materials (powder samples drilled from building wall)