Analysis of the Degradation Process of Sand-Lime Plasters Under the Impact of Crystallization Pressure

The article investigates internal basement wall plasters, especially the effects of their improper use. For this purpose, the results of author's research in the form of selected technical expertise accounts involving various building structures were used. The said expert opinions were developed in the years 1990÷2020 [4].

The article presents only some of the characteristic cases of damage to internal plasters observed in the examined buildings [4]. All the buildings are constructed in brick traditional technology. The following three buildings were examined: a school building constructed in traditional technology before 1945 (Poland), a two-family residential building completed at the beginning of the 20^th century (Germany) and single-family residential building constructed in 1934 (Poland).

The damaged roof drainage pipe in the school building was the source of significant dampness, which caused numerous discolouration of the plaster as well as the development of mold fungi. The consequences of the damage were found both in the basement of the examined building and in its aboveground part. Intense dampness in the basement was caused as well by a failure of the water supply system in a two-family residential building.

Degradation of basement plasters occurred as well in the basement of a single-family building (Fig. 1). Its construction was completed in 1934. No proper vertical waterproofing was provided at that time. Such insulation was not installed until 2004. Therefore, for many years the basement walls had been exposed to groundwater infiltration. As a result of these phenomena, many fragments of walls or plasters were affected by moistness, observed up to the height of the adjacent ground. The interiors of the examined buildings were in different technical conditions. The main factors that contributed to the observed damage include a lack of proper drainage system in the buildings, lack of proper waterproofing protection, lack of proper care for the technical condition of buildings and failures and leaks in water installations.

Figure 1.

Interior plasters are very important building elements, widely used in general construction. The most commonly used ones comprise cement-based plasters with the use of other materials such as lime and gypsum or lime-gypsum plasters. Progressively more plasters are improved with various additives, even polymeric and organic ones [5, 6].

The classical technology of cement materials describes their functional characteristics in terms of mixture composition, he proportion of components and technological parameters.

By examining the composition of mortar cement paste we can distinguish the following elements of the structure [7]:

hydration products C-S-H gels with a layered structure, constituting over 80% of all hydration products,

The mentioned elements make up a structure in which solid grains (of the filler, cement) are surrounded by hydration products in the form of consolidated gels having the nature of solid particles. The existing pores are filled up with water (pore liquid) and air [7]. Plaster mortar as a cement-based material, as a multi-component material, is a very complex system. During its hardening, the transformation of the individual phases takes place, whereof role and meaning are changing. In fresh material, the basic role in terms of workability is played by the particles with which capillary phenomena are related, i.e. grains of sand, cement or ash, as well as water. In the hardened material, the most significant are the particles which are active in terms of physicochemical properties, i.e. cement grains, as well as the liquid phase (pore liquid) and gas phase separated by liquid meniscuses that control capillary phenomena [1, 7, 8, 9]. The porosity of the paste depends primarily on the w/c ratio. As we know, full hydration of cement is obtained with about 20% of water, which corresponds to the w/c ratio = 0.25, and usually it is much higher [1, 7, 9]. The excess of water has an impact on the share of large capillary pores. The share of these pores - initially very large, decreases as the hydration process is progressing. The pores are filled up with hydration products, and the proportion of molecular (gel) pores, typical for the C-S-H phase, increases [10]. A special role in the obtained structure is played by gel pores and capillary pores [1, 7], which differ in size. Gel pores constitute about 28% of the total gel volume [1,9]. As the hydration process progresses, the volume of gel pores increases, while the volume of capillary pores decreases.

Porous materials are mainly characterized by their effective porosity, defined as the quotient of the volume of connected pores and the volume of the material [11]. Pores vary in shape and size. As determined by IUPAC (International Union of Pure and Applied Chemistry) [10], pores are conventionally divided by their diameter into: micropores (d < 2 nm), mesopores (2 nm 50 nm), macropores (d > 50 nm). Macropores in materials with well-developed surfaces make up a small fraction of their volume, but they play a decisive role in the transfer of moisture to mesopores and micropores. Mesopores are accountable for transport, and on their surface the adsorption of moisture particles takes place. Micropores are the basic carrier of sorption properties [11].

For a liquid, properties such as wettability and capillarity are associated with surface tension. Wettability is understood as the ability of a liquid to cover the surface of a solid. The measure of wettability is defined as the contact angle of the solid phase with the liquid one [1, 9, 11]. Capillarity is defined as the behaviour of a liquid in thin tubes with a diameter d< 10⁻⁷m (capillaries), in which we observe the rise (or fall) of the liquid column. The capillary action results in water soaking into porous materials, moisture rising in walls and migration of water in the sub-soil. Liquid water state in comparison to sorption moisture is able to dissolve mineral salts and contributes to the movement of them. There is no such ability in the case of sorption moisture.

Different pore sizes are connected with a flow of capillary water [9]. In that sense pores can be as: micropores (d < 10⁻⁷m, no capillary action), capillary pores (10⁻⁷–10⁻⁴m, capillary action), air pores (d > 10⁻⁴ m, capillary breaking). Water is absorbed in porous materials in either liquid or vapour form. All types of pores are able to absorb water vapour. The other factor governing the pores, in which condensation will occur, is the relative air humidity. Small pores, i.e. microspores or gel pores, fill up at very low relative humidity through condensation in the capillaries. The larger the pores, the higher the prevailing relative air humidity must be for condensation to occur. Capillary porous or air pores, can only be filled when the dew point is reached, i.e., the relative humidity must be nearly 100%. The uptake of liquid water is also related to the pore size. Whereas small pores, such as micropores and gel pores, stop absorbing water after a while, capillary pores will continue to absorb it as long as there is a water supply. Large air pores can only absorb water under pressure. This means that the level of saturation in the porous material is due to capillary water absorption [12].

Plaster mortar coatings, as porous materials, are exposed to the action of water, which is connected with the expansion of the material filling up the pores. This expansion is related to the crystallization of certain chemical compounds in the capillaries of the porous material. The phenomenon of salt crystallization has a very negative impact on the durability of the material, and therefore they have been investigated by many researchers [7, 13,14,15,16,17,18,19]. Correns [15] formulated an equation that allows calculating of the crystallization pressure: (1) Pcryst=RT/Vm⋅ln s {{\rm{P}}_{{\rm{cryst}}}} = \left( {{\rm{RT}}/{{\rm{V}}_{\rm{m}}}} \right) \cdot {\rm{\;ln\;s}} where: P_cryst – crystallization pressure [MPa], R – gas constant (8.3145 J/mol·K), T – absolute temperature [K], V_m – molar volume [cm³/mol], s – supersaturation/degree of a solution [−].

As can be seen from equation (4), the supersaturation degree of a solution significantly influences the value of crystallization pressure.

The authors of the work [17] found that ettringite in the first phase of sulfate corrosion seals the pores of the material, which is manifested by the increase in strength. However, with further growth of the crystals, very high internal stresses develop. For example, the crystallization pressure at the transition of CaSO₄ into CaSO₄·2H₂O is approx. 110 MPa [17]. Salt crystallization processes have been studied by Kurdowski [7], Scherer [16], Szeląg [17], as well as by the authors of the study [5]. Studies on modelling of salt solution migration or crystallization phenomena in plasters or renders can be observed in the works of Falchi et al. [20], Huinink et al. [23], Petkovic et al. [24, 25].

Analyzed models in these works refer especially to physical properties, not mechanical ones.

The problems of salt formation in renovation plasters on salt-laden substrates and the problem of plaster resistance depending on its type and structure were analyzed in the work [21].

The study [22] discusses the most important factors influencing the risk of damage caused by the action of salt, taking into account both the service life of the repaired plaster and the the conservation requirements of the existing building. Properties and utility parameters of various plaster categories were defined and reviewed. In all these studies, similar conclusions can be formulated regarding the importance of the properties of wall substrate and plaster for the transport of water and salt, and in particular: moisture diffusion, distribution and volume of pores.

The works [23,24,25] discuss research studies on the transport of salts and their accumulation in plaster, depending on pore distribution and the structure of porosity in plaster and substrate. Following the analysis of the results of the studies, we can conclude that the transport and accumulation of salts in the plaster as well as the drying of the system plaster/substrate depend on pore size distribution both in the plaster and in the substrate.

The research on the salinity of plasters with the use of a test simulating salinity conditions was carried out in the work [26]. They designed ageing tests simulating water movement, ice formation and crystallization of salts in lime mortars applied to plaster historic buildings with the aim to determine their strength properties. It was found that the response of mortar blends differs depending on the mechanism and attack factor, choice of an appropriate methodology for determining durability, and it depends on various factors, such as binder content, aggregate ratio, and binder to sand ratio. The research on the impact of the porosity structure of renovation plasters on salt crystallization as well as on the durability of these plasters was investigated in the works [27, 28]. The findings demonstrated that the type of pore size and pores distribution in renovation plasters significantly influenced the speed of their drying and the crystallization of salts, and thus their durability. This confirms the essence and effectiveness of renovation plasters, which are very useful, especially in historical buildings. It is worth noting that the studies on lime mortars for plastering historic buildings were also carried out in the work [26].

In the tables above, using the Correns formula [15] (equation 1), the values of the crystallization pressure of ettringite, portlandite and brucite are presented. Following the analysis of crystallization pressures for the compounds (Tabl. 1A, 1B, 1C) we can conclude that the degree of supersaturation “s” according to the equation (1) has a very large impact on their values. And temperature has a minor influence on the values of these pressures. Therefore in the article, crystallization pressures associated with the water-ice phase transformations will not be analyzed, since the observations were limited to indoor rooms, where the temperature was not lower than 10°C.

Not crumbled hard plaster samples about 1.5 cm thick were collected using a drill bit and then using a special head. Yet, the pull-off measurement [29] in the rupture process of plaster from the substrate could not be applied due to the fact that the plaster samples had very little adhesion to the masonry wall.

Nine plaster samples of diameter 50 mm were drilled off. Therefore, the tensile strength of the plaster was measured using a modified pull-off device (Fig. 2). The collected plaster samples were glued to the working rings of that device. The destruction of all 9 samples always took place in the plaster layer, where fracture planes of the sample could be seen.

Figure 2.

The analyzed plasters have a porous structure. It was assumed that a plaster model can be shaped as a sphere which is hollow inside. This hollow cavity corresponds to the volume of pores. As a result, we obtain a sphere loaded (Fig. 3), where “a” and “b” respectively denote the internal and external radius of the sphere, and pa and pb the internal and external pressure. As a result, we are faced with the problem of polar-symmetric deformation of a thick-walled spherical tank, as in [30].

Figure 3.

It was assumed that the sphere is mentally cut out in the vicinity of the pore (the inner sphere in the model) (Fig. 4). Therefore, there is no external pressure, i.e. p_b = 0. We will then obtain the following formulas for stresses occurring in the adopted model [30]: (2) σt=paa32r3+b32r3b3−a3 {\sigma _t } = {p_a}{{{a^3}\left( {2{r^3} + {b^3}} \right)} \over {2{r^3}\left( {{b^3} - {a^3}} \right)}} (3) σr=paa3r3−b3r3b3−a3 {\sigma _r} = {p_a}{{{a^3}\left( {{r^3} - {b^3}} \right)} \over {{r^3}\left( {{b^3} - {a^3}} \right)}}

Figure 4.

The stresses max σ_t occur for r = a. After simplifying the formula (2), we obtain: (4) maxσt=pa2a3+b32b3−a3 {{max}}{\sigma _t } = {p_a}{{2{a^3} + {b^3}} \over {2\left( {{b^3} - {a^3}} \right)}}

The stress σ_r for the radius r = a has the following value, and thus compression is taking place. (5) σr=a=paa3a3−b3a3b3−a3<0 {\sigma _{r = a}} = {p_a}{{{a^3}\left( {{a^3} - {b^3}} \right)} \over {{a^3}\left( {{b^3} - {a^3}} \right)}} < 0

The porosity of the adopted model, defined as the quotient of the internal volume of the sphere with the radius “a” and the sphere with radius “b”, can be calculated from the formula: (6) p=a3b3⋅43π43π=a3b3 p = {{{a^3}} \over {{b^3}}} \cdot {{{4 \over 3}\pi } \over {{4 \over 3}\pi }} = {{{a^3}} \over {{b^3}}}

The calculations demonstrate that the fact of ignoring the external pressure does not cause major differences in stress values. For example, for the pressure p_b = 0.10 MPa, the values of circumferential stresses “σ_t” are lower by approx. 4% as compared to the case when p_b = 0.0 MPa.

Preliminary model calibration:

Using the model equations provided above, the calculations of stresses “σ_t” for different dimensions of radii “a” and “b” were made, with the internal pressure p_a = 1.0 MPa. The calculation results are presented in the tables below. For each combination of the dimensions of the radii, the porosity “P” of the model was calculated. The calculations show that the same values of porosity P correspond to identical values of the circumferential stresses “σ_t”, regardless of the value of the radii of the analyzed model. In Tables 2 and 3, the corresponding values are marked with the same colours. For porosity P = 0.001 – yellow, for porosity P = 0.008 – grey, for porosity P = 0.027 – blue, and for porosity P = 0.064 – green.

The supplementary tests of plaster involved checking the following: a)

temperature and humidity in the basement in an exemplary period from June 5, 2021, to July 7, 2021, just before the collection of samples for testing.

In order to analyze the porosity structure of the plaster, tests were carried out using mercury porosimetry. Sample 2 and sample 3 were chosen from all samples and subjected to testing. Sample 2 was collected from the area of the crumbled plaster, in the damp zone and damage caused by salt crystallization, below the ground level. Sample 3 was collected in the area of hard plaster, above the crumbled plaster. One piece of each sample were prepared for the MIP analysis. To ensure the representativeness of the samples for the porosimetry test, two or three pieces of each sample were prepared so that the MIP operator could choose randomly one small piece of each sample for the MIP analysis.

No micropores were found in the tested plasters. There were only mesopores below 0.05 μm and macropores above diameters of 0.05 μm. Pore size distribution of samples 2 and 3 are similar with more amount of macropores 10–60 μm than of sample 2 (Fig. 7, Tab. 9). Additionally it can be seen that sample 2 had lower open porosity (26%) as compared to sample 3 (27.7%). Sample 2 also had lower permeability as compared for sample 3. The average pore diameter for sample 2 was also lower than of sample 3. As for other parameters, such as tortuosity, apparent density, actual density and total surface area of pores, higher values were recorded in sample 2 as compared to sample 3 (Tab. 9).

Figure 7.

To recognize the changes taking place in the damp plaster thoroughly, chemical tests were also carried out on samples 1, 2 and 3. It should be noted here that the results for samples 1 and 2 were very similar. It was found that samples 1 and 2 contained from 1 to 3% of sulfate ions and sample 3 contained more than 7% in relation to the mass of the part of plaster which was soluble in hydrochloric acid (binder, water and corrosion products), with practically the same content of aggregate equal to about 20% by weight in both samples. The conductivity of the suspensions made of samples 1, 2 and 3 was 6.1, 5.7 and 8.9 mS/cm, respectively. The reaction of samples 1 and 2 was slightly lower than that of sample 3, with the value of 8.5 and 8.9, respectively. The mineralogical analysis with the X-ray diffraction method of samples 1 and 2 showed the presence of the main phases: calcite and quartz, and in smaller amounts: muscovite, albite and microcline. The content of gypsum was questionable. Based on the above studies, it was determined that it was sand-lime plaster since no phases coming from the damaged cement paste, in particular ettringite, were detected.