United States Environmental Protection Agency Risk Reduction Engineering Laboratory Cincinnati, OH 45268 Research and Development EPA/600/SR-95/068 May 1995 vvEPA Project Summary Calcium Carbonate Dissolution Rate in Limestone Contactors Raymond D. Letterman The project summarized here inves- tigated some of the parameters and relationships used to predict the per- formance of limestone contactors. The purpose of the project was to study the effect of limestone composition, espe- cially the dolomite [CaMg(CO3)2] and impurity content of the stone, on the kinetics of carbonate mineral dissolu- tion and to determine the effect of tem- perature on the rate of dissolution. The rate of dissolution was determined by using a rotating disk apparatus and samples of limestone of varied compo- sition. The limestone samples included a white marble and a selection of sedi- mentary stones. The white marble con- tained a significant amount of silica (approximately 35%). The major min- eral constituents of the sedimentary limestones ranged from approximately 100% calcite (CaCO3) to essentially pure dolomite. The approximate iron (Fe) content of the stones ranged from 15 to 377 mg Fe/100g and the approxi- mate aluminum content (Al) from 1 to 134 mg Al/100g. A heterogeneous reaction model for mineral dissolution effectively explained the results of the rotating disk experi- ments for all samples except the two with the highest dolomite content. The magnitude of the dissolution rate con- stant for fresh stone decreased by ap- proximately 60% as the calcite content of the stone decreased from 0.92 to 0.09 g CaCO3/g stone. The rate of dis- solution of stones with a high dolomite content may have been enhanced by the presence of small amounts of cal- cite. The rate of solubilization of mag- nesium (Mg) was negligible in all samples except the two with the high- est dolomite content (93 and 100 mass percent dolomite). The overall dissolution rate constant decreased as the amount of calcium dissolved from the surface of the stone increased. For a given amount of cal- cium dissolved per unit area of stone surface, the magnitude of the percent- age decrease in the dissolution rate constant increased as the iron and alu- minum content of the stone increased. The effect of sample aging on the rate of dissolution was lowest when the weighted sum of the iron and alumi- num content of the stone was less than about 10 mg/g. The weighted sum is equal to the aluminum content in mg Al/g plus 0.30 times the iron content in mg Fe/g. The presence of silica as the principal impurity in the white marble seemed to reduce the effective (cal- cite) surface area of the stone in pro- portion to the mass of silica in the sample, but it did not appear to affect the dissolution rate of the calcite sur- face. The dissolution rate constant for cal- cite increased with increasing tempera- ture, from 0.38x10-3 cm/s at 5°C to 2.80x10'3 cm/s at 25°C. The apparent activation energy was 101+8 kJ/mol for the surface reaction rate constant and ------- 17±0.3 kJ/mol for the mass transfer rate constant in the heterogeneous reac- tion model. This Project Summary was developed by EPA's Risk Reduction Engineering Laboratory, Cincinnati, OH, to announce key findings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at the back). Background A limestone contactor is a treatment device in which water flows through and dissolves carbonate minerals (typically cal- cium carbonate) from a packed bed of crushed limestone. Dissolution of calcium carbonate (under a closed-to-atmospheric- carbon dioxide condition) increases the pH, alkalinity, and dissolved inorganic car- bon concentration of the water and de- pletes the amount of calcium carbonate in the bed. Limestone contactors are simple, low-cost devices, which usually require minimal maintenance and are, therefore, especially suitable for small water sup- plies. In an earlier study (Letterman et al., 1987), limestone contactors effectively re- duced the dissolution of corrosion byproducts, such as lead, copper, and zinc, from piping system surfaces. A math- ematical model related the depth of lime- stone needed to reach the desired effluent water chemistry to the influent water chem- istry, the limestone particle size and shape, the limestone bed porosity, and the tem- perature and superficial velocity of the water. Limited field experiments showed that contactor performance decreases as the water temperature decreases. Another study (Haddad, 1986) monitored the long-term operation of a contactor con- taining somewhat impure, high-calcium limestone. Here, the author concluded that as the calcium carbonate dissolved, the rate of dissolution decreased because rela- tively insoluble impurities formed a resi- due layer. As the thickness of this layer increased, the rate of transport of calcium ion from the calcium carbonate surface to the bulk solution decreased, and, thus, contactor performance decreased with time. Field experiments have shown (Letterman et al., 1987) that the tempera- ture of the water flowing through a lime- stone contactor can affect its performance. For a given set of design and operating conditions, contactor performance tends to decrease with decreasing temperature. One of the objectives of this study was to obtain a better understanding of this rela- tionship. Experimental Materials, Apparatus, and Methods Limestones The study was conducted using 13 samples of limestone including a white marble (sample WM) from a quarry in Proctor, VT, a sedimentary limestone (sample SL) from a quarry near Boonville, NY, Black River limestone (sample BR) from a quarry near Watertown, NY, and 10 samples (samples A-J) from a dolo- mite quarry near York, PA. A sample of each stone was powdered and dissolved in concentrated hydrochlo- ric acid. Dilutions of this solution were used to determine the calcium, magne- sium, iron, and aluminum content of the stone with a direct current plasma spec- trometer and an atomic absorption spec- trophotometer. For a number of the samples, some translucent material, prob- ably quartz, remained after 2 days of dis- solution. The measured calcium and magnesium content of the samples was used to estimate the calcite, dolomite, and insoluble residue content of the samples. In these calculations the magnesium was assumed, based on x-ray diffraction and thin-section photomicrography results, to be associated only with dolomite. The results of these calculations and the measured iron and aluminum values are listed in Table 1. In several cases, where the sum of the calcite and dolomite fractions was slightly greater than 100g/ 100g of stone, the insoluble residue con- tent was set equal to zero. Samples WM, SL and BR as well as a number of the samples from the York do- lomite quarry (samples A, B, D, E and F) are high calcium content limestones. Other York samples (samples C, G, H, and I) are predominately dolomite, and sample J is essentially pure dolomite. The WM sample had the highest in- soluble residue content (36 g/100g) but relatively low amounts of iron and alumi- num (34 mg Al/100g and 71 mg Fe/100g of stone). It is very likely that the insoluble residue in this sample is quartz. Sample I, from York, had the highest amount of iron (377 mg Fe/100g) and sample H had the highest amount of aluminum (134 mg Al/ 100g). The stone disks used in the rotating disk apparatus were prepared by cutting 3.10- or2.45-cm-diameter, cylindrical cores from pieces of rock collected at the quar- ries. Each core was cut into a number of 3-mm-thick disks using a rock saw. The disk faces were smoothed and polished on a lapwheel with a silicon carbide abra- sive. The back face and edge of the disks were coated with plastic so that only the polished face was available for dissolu- tion. Each disk was mounted in a Teflon- coated1 brass holder (Figure 1). Between dissolution rate experiments, each stone sample was "aged" by controlled dissolu- tion in dilute acid solution. The cumulative amount of calcium and magnesium dis- solved during aging was determined by measuring their concentrations in the di- lute acid solution. Rotating Disk Apparatus The reactor used in the rotating disk apparatus (Figure 1) was 14 cm in diam- eter and the clearance between the disk and the walls of the vessel was greater than 4 cm. The disk was centered about 3 cm above the bottom of the vessel, and its rotational speed was varied over the range 200 to 1200 rpm. The reactor was constructed with double-glass walls. A water bath was used to circulate water between the walls to maintain the reactor contents at preselected temperatures in the range 4° Mention of trade names or commercial products does not constitute endorsement or recommendation for use. Table 1. Estimated Major and Minor Constituents of the Stone Samples (g/100g) Stone ID Calcite Dolomite Insoluble Fe Al WM SL BR A B C D E F G H 1 J 64 92 99 89 92 17 79 71 89 23 38 9 0 1 4 2 16 4 68 18 29 9 59 53 93 100 35 4 0 0 4 15 3 0 2 18 9 0 0 0.071 0.101 0.019 0.024 0.029 0.189 0.040 0.041 0.015 0.294 0.154 0.377 0.189 0.034 0.114 0.044 0.012 0.001 0.093 0.037 0.025 0.005 0.129 0.134 0.032 0.010 ------- Plexiglass Drive shaft Drive shaft support r \ Limestone disk Water jacketec Reactor support reactor Figure 1. Schematic diagram of the rotating disk apparatus. to 25+0.2°C. The plexi-glass cover on the reactor had holes for inserting the rotating shaft, pH electrode, and wetted nitrogen inlet tube. Additional holes were provided for measuring the temperature and pipetting samples for the calcium mea- surement. The bulb of the pH electrode was located 1.5 cm from the rotating disk and 3 cm above the bottom of the vessel. All solutions used in the rotating disk experiments were made with distilled and deionized water that had been boiled for a few minutes, several hours before use, to remove carbon dioxide. Fisher analytical grade (ACS Certified) chemicals were used (KCI, N/10 HCI). The background electro- lyte was 0.079 M KCI. Methods A free-drift method, in which the pH was allowed to increase as the carbonate minerals dissolved from the stone, was used in all experiments. Experimental so- lutions (600 mL) were prepared as needed by adding potassium chloride and the re- quired volume of acid to boiled water and then transferring this to the reactor. Each experiment began by raising the vessel and solution into place beneath the rotating disk and against the plexi-glass cover. Samples of solution (either 2- or 5- ml volume) were withdrawn from the ves- sel at 6 or 9 min intervals for a period of 1.5 hrwith the use of an automatic pipette (1 to 5 ml). The samples were stored in polyurethane disposable test tubes at 4°C for no longer than 2 days before the ion concentrations were measured by atomic absorption spectrophotometry. Experimental Results The calcium concentration and pH change with time in a typical rotating disk experimental run are shown in Figures 2 and 3. In this example, the WM stone sample was used, the rotational speed was 600 rpm, and the initial acidity was 0.01 meq/L. At the end of the experiment the pH was 9.04 and the calcium concen- tration was about 2 mg/L. For an initial acidity of 0.01 meq/L and with no calcium in the solution at t = 0, the calculated equilibrium calcium concentration is 11.6 mg/L and the calculated equilibrium pH is 10.02. The overall dissolution rate constant, ko, was determined for each experimental run by using the measured calcium con- centrations and, in some cases, the mea- sured magnesium concentrations. For the stones that released negligible amounts of magnesium, the calcium concentrations (Ct) were substituted in the relationship, M = In {(Ceq - C,)/(Ceq - Co)}(V/A) (1) where Ceq and Co are the equilibrium and initial calcium concentrations, respec- tively, and A is the surface area of the stone sample disk exposed to the solu- tion. Co was zero in all experiments. Vt is the volume of the solution in the rotating disk apparatus. For sample WM and samples A-J, the limestone disk was 3.6 cm in diameter and, therefore, A was 10.17 cm2. For the 3.1-cm diameter SL sample and the 2.5-cm-diameter BR sample, A was 7.91 cm2 and 4.71 cm2, respectively. ------- 2.5 I 1.5 1.0 I 0.0, 0.0 0.2 0.4 0.6 0.8 1.0 Time (hours) 1.2 1.4 1.6 Figure 2. Calcium concentration in the rotating disk apparatus as a function of time: WM sample; w = 600 rpm and initial acidity of 0.01 meq/L. 9.0 6.0 0.0 0.2 0.4 0.6 0.8 1.0 Time (hours) Figure 3. pH versus time for the rotating disk experiment of Figure 2. The magnitude of Ceq was determined for each experimental run by using a chemical equilibrium model and effec- tive solubility products determined for a number of the stone samples (Table 2). For the stone samples that were not in- cluded in the solubility product experi- ments, i.e., samples B, D, E, G, H and J, the average value of the effective solubil- ity products (pKsp = 8.81) for the samples from the same quarry was used. pKsp = 8.35 was used for sample BR because of its similarity to sample SL. Table 2. Effective Solubility Products for Calcium Carbonate and Calcium-Magnesium Carbonate in Selected Limestone Samples.' Stone sample ID WM SL A C F 1 Negative log of the effective solubility product 8.20 ±0.07 8.35 ±0.06 8. 76 ±0.09 8.72 ±0.07 8.88 ±0.05 8.89 ±0.04 'Values are for 25°C and infinite dilution. As samples were withdrawn during an experiment, the magnitude of Vt de- creased. A value of Vt was calculated for each value of Ct using the relationship, V, = V0 - nv (2) where Vo is the volume of the solution in the reactor at the start of the experi- ment, v is the volume of each sample withdrawn for the calcium and magne- sium measurements, and n is the total number of samples withdrawn from the reactor up to that sample. In the dissolu- tion rate experiments, Vo was 600 ml_ and v was either 2 or 5 mL A straight line was fitted to the M ver- sus time points using the method of least squares (M is given by Equation 1). The slope of this line is equal to the overall dissolution rate constant. Figure 4 is an M versus time plot for a fresh sample of WM stone. In this experi- ment, the disk rotational speed was 600 rpm, the water temperature was 25°C, and the initial acidity was 0.01 meq/L. The slope of the least squares line in Figure 4 yields an overall dissolution rate constant of 3.3x10'3 cm/s. Effect of Insoluble Residue on the Dissolution Rate The overall dissolution rate constant for fresh calcitic stones (stones with low dolo- mite content) tended to decrease as the ------- -10 -12 0.0 0.2 0.4 0.6 0.8 1.0 Time (hours) 1.2 1.6 Figure 4. Determination of the overall dissolution rate constant (ko) for the experiments of Figures 2 and 3 (ko = o.oo33 cm/s). estimated amount of insoluble residue in the stone increased. We concluded that the insoluble impurities reduced the area of calcite exposed to the solution. To test this hypothesis, we assumed that the area of exposed calcite is proportional to the mass percent of calcite in the stone. The rate constants were "corrected" for the residue content by dividing them by the mass percent of calcite in the stone. The results of this calculation, listed in the right-hand column of Table 3, show that this correction reduces the effect of the residue content on the overall dissolution rate constant. The corrected overall dissolution rate constant for the coarse-grained WM sample (3.12x10'3 cm/s) is somewhat less than the values of 3.51x10'3 and 3.75x10'3 cm/s for the fine-grained SL and BR samples. Effect of Aluminum and Iron Content on the Dissolution Rate We observed that the extent to which the dissolution of calcium from the stone surface reduced the overall dissolution rate constant depended on the aluminum and iron content of the stone. Figure 5 shows the normalized overall dissolution rate constant (i.e., the mea- sured value divided by the initial, fresh stone, value, ko/koi) plotted versus the amount of calcium dissolved from the sur- face of the stone, Cad, for stones A through H. (The results for stones I and J were Table 3. Comparison of Experimental and Corrected Overall Dissolution Rate Constants (kjfor Essentially Fresh Limestone Disks.' Stone Mass % Experimental kox103 Corrected kox103 Calcite (cm/s) (cm/s) Calcium dissolved in "aging" = 0.2 mg Ca/cm2 WM 64 SL 92 BR 100 Calcium dissolved in "aging" ~ 0 mg Ca/cm2 B 92 F 89 1.99 3.26 3.75 4.39 3.46 3.12 3.51 3.75 4.77 3.89 'Small amounts of calcium had been dissolved from samples WM, SL, and BR before the first rate constants were determined. also plotted but are not shown in Figure 5.) For stones C, I, G, and H, the overall dissolution rate constant decreased by more than 60% as the amount of calcium dissolved increased from 0 to 4 mg Ca/ cm2. For stones B, F, and J, the decrease was less than 30%. Values of ko/koi were interpolated from Figure 5 at Cad = 2 mg Ca/cm2 and then listed in Table 4 in rank order, from the highest (ko/koi = 0.90 for stone F) to the lowest (kjkj" = 0.23 for stone G). The stones with the highest aluminum content (> 25 mg Al/100g of stone) had the great- est decrease in the overall dissolution rate constant for this amount of calcium dis- solved. For several stones, especially stone I with ko/koi = 0.36, the iron content seemed to be an additional factor. Since both the iron and aluminum con- tent of the stone seem to determine how sample aging affects the overall dissolu- tion rate constant, a composite parameter that includes a weighted combination of the iron and aluminum concentrations (aC^ + bCFe) was derived, where C^ is the aluminum concentration in mg AI/100 g and CFe is the iron concentration in mg Fe/100 g. The highest linear correlation between ko/koi and the Fe+AI parameter (r2 = 0.92) was obtained with weighting factors a=1 and b=0.3, i.e., (C^ + 0.30 CFe). The quantity ko/koi and correspond- ing values of (C^ + 0.30 CF) are listed in Table 5. According to the results in Table 5, the effect of iron and aluminum on the overall dissolution rate constant will be minimized if the quantity C^ + 0.30 CFe for the stone is less than about 10 mg/100g. In a special experiment, the brownish residue layer that formed on the SL disk was scraped into concentrated nitric acid, ultrasonicated and the solution was ana- lyzed for total soluble aluminum. The soluble aluminum expressed as the amount per area of disk was 0.97 |omoles/ cm2 (26 |ig/cm2). The scraped residue did not dissolve completely in acid which sug- gested the presence of alumino-silicates. The overall dissolution rate constant for the SL stone increased to 90% of its origi- nal value when the residue layer was scraped from the disk surface. Conclusions and Recommendations A heterogeneous reaction model for min- eral dissolution, in which the rate of disso- lution is controlled by a surface reaction and a cation mass transfer resistance act- ing in series, effectively explained the re- sults of the rotating disk experiments for all samples except the two with the high- ------- 6 8 Ca dissolved per unit area of disk (mg/sq cm) Figure 5. Effect of amount of calcium dissolved from disk surface on the fractional decrease in the overall dissolution rate constant. Table 4. Effect of the Iron and Aluminum Content on the Fractional Decrease in the Overall Dissolution Rate Constant (kj at Cad = 2 mg Calcium Dissolved/sq cm of Limestone Surface. Stone ID F A B J E D C 1 H G Calcite 88.9 88.9 92.1 0 70.7 78.9 16.5 8.8 38.3 22.9 /ro*703' (cm/s) 3.5 4.7 4.4 0.9 3.3 4.2 2.7 2.8 3.3 3.2 W 0.90 0.74 0.73 0.70 0.65 0.67 0.43 0.36 0.35 0.23 Fe (mgFe/100g) 15 24 29 189 41 40 189 377 154 294 Al (mgAI/100g) 5 12 1 10 25 37 93 32 134 129 ' Interpolated from Figure 5 at Cad = 2 mg Ca/cm2.* kd = value when negligible Ca dissolved from the stone. Table 5. Effect of the Weighted Sum of Iron and Aluminum in the Limestone on the Fractional Decrease in the Dissolution Rate Constant at 2 mg Calcium/sq cm of Limestone Surface. Stone ID ko/ko: (mg/100g) F A B J E D C 1 H J 0.90 0.74 0.73 0.70 0.65 0.61 0.43 0.36 0.35 0.23 10 19 10 67 37 49 149 145 180 217 'kd = value when negligible Ca dissolved from the stone. est dolomite content. For calcite and the experimental conditions of this study, the surface reaction rate was relatively large and the rate of dissolution was essentially mass transfer controlled. The results show that a calcium ion diffusivity of 0.8 x 10"5 cm2/s (at 25°C, can be used in predicting the mass transfer resistance. The stone samples with the highest cal- cite content and lowest dolomite content had the highest initial rates of dissolution. The magnitude of the overall dissolution rate constant for fresh stone decreased by approximately 60% as the calcite con- tent of the stone decreased from 0.92 to 0.09 g CaCO3/g stone. The rate of disso- lution of stones with high dolomite content may be enhanced by the presence of small amounts of calcite. For example, the stone that was essentially pure dolomite had a dissolution rate constant that was 66% less than the constant for another dolo- mitic stone with approximately 9% calcite. When the high dolomite content samples were fresh, it appeared that the calcium carbonate component of the dolomite dis- solved faster than the magnesium car- bonate component. The rate of dissolution of magnesium was negligible in all samples except the high dolomite content samples (93 and 100 mass percent dolomite). The overall dissolution rate constant decreased as the amount of calcium dis- solved from the surface of the stone in- creased. Analysis of several stone surfaces, by scanning electron microscopy and x-ray energy spectroscopy, indicated that the density of calcium atoms on the surface of the stone decreased and the density of aluminum, silicon, and iron in- creased as calcium dissolved. For a given amount of calcium dissolved per unit area of stone surface, the magnitude of the decrease in the overall dissolution rate constant increased as the iron and alumi- num content of the stone increased. The results suggest that the effect of sample aging on the rate of dissolution is a mini- mum if the weighted sum of the iron and aluminum content of the stone is less than about 10 mg/g. The weighted sum is equal to the aluminum content in mg Al/g plus 0.30 times the iron content in mg Fe/g. To minimize the negative effect of mineral dissolution and residue-layer buildup on the performance of a limestone contactor during long-term operation, the iron and aluminum content should be less than this weighted sum. The presence of silica as the principal impurity in the white marble reduced the effective surface area of the calcite in proportion to the mass of silica in the sample but did not appear to cause a ------- reduction in the dissolution rate of the calcite surface. The dissolution rate of calcite increased with increasing temperature, from 0.38x10-3 cm/s at 5°C to 2.80x10'3 cm/s at 25°C. The apparent activation energy de- termined for the surface reaction rate con- stant in the heterogeneous reaction model was 101+8 kJ/mol, a value that is signifi- cantly larger than literature values (46 to 63 kJ/mol). The apparent activation en- ergy for the mass transfer rate constant was 17+0.3 kJ/mol, which is consistent with values in the literature for mass trans- fer controlled kinetics. References Haddad, M., 1986. Modeling of Lime- stone Dissolution in Packed Bed Contactors Treating Dilute Acidic Water. Ph.D. Dissertation, Department of Civil Engineering, Syracuse University. Letterman, R. D., C. T. Driscoll, Jr., M. Haddad and H. A. Hsu, 1987. Limestone Bed Contactors for Control of Corrosion at Small Water Utilities. A Report for the Water Engineering Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Cincin- nati, OH (EPA/600/S2-86/099). Letterman, R D., M. Haddad and C. T. Driscoll, 1991. Limestone Contactors: Steady-State Design Relationships, Jour- nal of Environ. Eng.. Am. Soc. of Civil Engineers, 117:339-358. The full report was submitted in fulfill- ment of CR 814926 by Syracuse Univer- sity under the sponsorship of the U.S. Environmental Protection Agency. ------- Raymond D. Letterman is with the Department of Civil and Environmental Engineering, Syracuse University, Syracuse, NY 13244-1190. Jeffrey Q. Adams is the EPA Project Officer (see below). The complete report, entitled "Calcium Carbonate Dissolution Rate in Lime- stone Contactors,"(Order No. PB95-222733; Cost: $27.00, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Risk Reduction Engineering Laboratory U.S. Environmental Protection Agency Cincinnati, OH 45268 United States Environmental Protection Agency Center for Environmental Research Information Cincinnati, OH 45268 Official Business Penalty for Private Use $300 BULK RATE POSTAGE & FEES PAID EPA PERMIT No. G-35 EPA/600/SR-95/068 ------- |