Results & Discussion
When the calcite powder is exposed to the polymer solution, the pH changes with time, as shown in Figure 4. The pH stabilizes at a value between 8.3 and 9, depending on the concentration of KOH in the solution. Rapid equilibration is achieved at pH 7.2, and attack on the calcite at that pH is minimal, so this is the pH range used in the testing the salt resistance of the PAA treatment.
As shown in Figure 5, the amount of PAA adsorbed on calcite increases almost linearly with the concentration of the polymer solution up to about 0.19%, after which the amount adsorbed remains nearly constant. Increasing the concentration by an order of magnitude causes only a slight additional deposit, as shown in Figure 6. This is about 5 times the concentration found to produce monolayer coverage in the study by Thompson et al. , but their polymers were not neutralized; moreover, as explained below, we apparently obtained multilayer coverage.
The adsorption measurements were made on a calcite powder with specific surface area SBET = 4465 cm2/g. The weight percent adsorbed, cA , is
where mP is the mass of polymer adsorbed on mass mC of calcite, whose total surface area is S = mC SBET . The mass of adsorbed polymer per unit area of calcite is
where the second equality follows from eq. (2).
The adsorption data can be fitted to a Langmuir isotherm of the form
where mA is the mass adsorbed per unit area in equilibrium with a solution containing concentration cS of polymer, and K is a constant. From the fit in Figure 6, we find that K = 8.36 and the maximum amount of polymer that can be adsorbed is mAmax = 0.058 m2/g = 5.8 x 10-6 cm2/g.
If ρP is the density of the adsorbed layer of polymer, then the layer thickness of adsorbed polymer, δ, is
With ρP in g/cm3 and cS in weight percent, the adsorbed thickness in nanometers is
The density ρP to be used in eq. (5) depends on the configuration of the adsorbed polymer. We do not know that value, but it is certainly bounded by the density of the solid polymer, !Pmax = 1.3
g/cm3 , and the density of the solvent, !Pmin = 1.0 g/cm3, which would be approached if the polymer were highly dilated. These two bounds are used for the calculations in Table 1. At the peak of adsorption, where mA = mAmax , eq. (5) indicates that δ lies between about 45 and 58 nm for the maximum and minimum adsorbed densities, respectively.
The radius of gyration of a PAA molecule with a molecular weight of 5000 is estimated to be about rG ≈ 2.6 nm, based on the data of Reith et al. . Estimating the thickness of an adsorbed monolayer to be δM ≈ 2rG ≈ 5 nm , we conclude that our peak adsorption was about 10 monolayers thick. The deprotonated polymer layers may have been bound together by calcium ions liberated from the calcite substrate, as well as potassium ions from the KOH used to neutralize the solution. In practice, we do not want that much polymer to adsorb, because desorption of those layers would result in free PAA that could act as a nucleation inhibitor for salt. That could be dangerous: if the salt concentration in the pores built to the point that the inhibitor was consumed, then the supersaturation could be so large that the crystals would grow with extremely high chemical potential. For that reason, we want only enough polymer to constitute ≤ 1 monolayer.
To achieve monolayer coverage, we want to choose cS such that δ ≤ δM. According to eq. (5), this means that the solution in equilibrium with the adsorbed layer should have a concentration of cM ≈ 0.011-0.015 wt% (estimated using !Pmin and !Pmax , respectively). Of course, this applies in the case where there is an infinite reservoir of solution at that concentration. In contrast, during treatment of stone, the volume of the solution is limited to the volume of the pores. Our goal is to introduce a solution into the stone such that the concentration drops to cM after the monolayer of polymer has adsorbed.
Suppose that the stone has pore volume Vp (cm3/g) = φ/ρB , porosity φ, bulk density ρB, and specific surface area S (cm2/g). When the stone is saturated, in every gram of stone there is a volume of solution equal to Vp in contact with surface area S, onto which it deposits thickness δ of polymer with density ρP. The mass of the adsorbed polymer per gram of stone is MA = S δ ρP. The specific volume of solution remaining after equilibration is VS = Vp – S δ, and it is intended to contain concentration c∞ = cM. The concentration (wt%) of polymer remaining in solution per gram of stone is found from
where ρS[c∞] is the density of a solution with concentration c∞. The initial concentration (mass fraction) of the solution used for the treatment is given by
The density of the solution as a function of PAA concentration, shown in Figure 7, is approximated by
We are interested in relatively dilute solutions, so we can approximate ρS by ρW, the density of water. Therefore, the concentration of the treatment is
Since the target final concentration is c∞ ≈ 0, this reduces to
This formula differs insignificantly from the exact solution, based on eqs. (8) and (9), which used to calculate the concentrations shown in the last column of Table 1.
Based on these results, PAA solutions were prepared at concentrations of 0.5, 0.75, 1.0, and 1.5 wt%. The predicted thickness of the layer deposited by these solutions is shown in Table
2, for polymer densities of !min and !max . These submonolayer levels were chosen with the PP intention of minimizing free polymer in the pore solution.
In the first capillary rise test, the Cadeby and Indiana limestones behaved similarly. The untreated samples were the first to show signs of damage, followed by the heavily (1.5 wt%) and lightly (0.5 wt%) treated samples. In the case of the heavily treated samples, the damage originated deeper inside the stone than in the lightly treated and untreated samples. For both sets of stone, the moderately (0.75 and 1.0 wt %) treated samples lasted the longest, suggesting that the adsorbed layers provided protection against crystallization pressure without excessive desorption. The earlier failure of the more heavily treated stone (estimated to have about 1/3 of a monolayer adsorbed) suggests that some of the polymer has desorbed and inhibited precipitation, leading eventually to growth at higher supersaturations. If significant desorption occurs from a partial layer, then this treatment will be too difficult to control for field use. On the other hand, it may be that the treatments were too conservative: if submonolayers provide insufficient protection, then significant crystallization stresses would have developed in the stone, in which case the damage reflects random variations in sample strength, rather than a meaningful pattern. The latter interpretation is supported by the second capillary rise test, in which 10 samples of Cordova Cream were tested with a similar range of layer thickness, and the results showed no discernible pattern. Future tests will explore a wider range of PAA concentration.
Use of polyacrylic acid shows promise for protection of limestone against crystallization pressure, but the treatment has not yet been brought under control. In the capillary rise tests, inconsistent results were obtained, which might reflect an excessively conservative application of polymer. Future tests will explore a wider range of compositions. To provide quantitative information, dilatometric tests will be performed on samples of stone during crystallization  to determine the effect of PAA treatment on the stress generated in the stone.