Removal of Pb from Water: The Effectiveness of Gypsum and Calcite Mixtures

: Anthropogenic lead pollution is an environmental problem that threatens the quality of soils and waters and endangers living organisms in numerous surface and subsurface habitats. Lead coprecipitation on mineral surfaces through dissolution-recrystallization processes has long-term effects on lead bioavailability. Gypsum and calcite are among the most abundant and reactive rock forming minerals present in numerous geological settings. In this work, we studied the interaction of slightly acidic (pH i = 5.5) Pb-bearing aqueous solutions ([Pb] i = 1 and 10 mM) with crystals of gypsum and/or calcite under atmospheric conditions. This interaction resulted in a reduction of the concentration of lead in the liquid phase due to the precipitation of newly formed Pb-bearing solid phases. The extent of this Pb removal mainly depended on the nature of the primary mineral phase involved in the interaction. Thus, when gypsum was the only solid phase initially present in the system, the Pb-bearing liquid-gypsum interaction resulted in Pb removals in the 98–99.8% range, regardless of [Pb] i . In contrast, when the interaction took place with calcite, Pb removal strongly depended on [Pb] i . It reached 99% when [Pb] i = 1 mM, while it was much more modest (~13%) when [Pb] i = 10 mM. Interestingly, Pb-removal was maximized for both [Pb] i (99.9% for solutions with [Pb] i = 10 mM and 99.7% for solutions with [Pb] i = 1 mM) when Pb-polluted solutions simultaneously interacted with gypsum and calcite crystals. Despite the large Pb removals found in most of the cases studied, the ﬁnal Pb concentration ([Pb] f ) in the liquid phase was always well above the maximum permitted in drinking water (0.01 ppm), with the minimum ([Pb] f = 0.7 ppm) being obtained for solutions with [Pb] i = 1 mM after their interaction with mixtures of gypsum and calcite crystals. This result suggests that integrating the use of mixtures of gypsum-calcite crystals might help to develop more efﬁcient strategies for in-situ decontaminating Pb-polluted waters through mineral coprecipitation processes.


Introduction
The quantity and the quality of available freshwater have been steadily declining over the last century in both industrialized and developing countries. Human activities like farming, mining, or manufacturing, which provoke the continuous rise of salinity and pollutant contents of freshwater in many sites worldwide overworld, are the main factors of this decline [1][2][3][4], as well as the adoption of new lifestyles by large groups of citizens in densely populated cities, which is responsible for the general increase of the drug and pharmaceutic contents in wastewaters. The current scenario of progressive climate warming, which sketches a future of groundwaters with increasing salinity, adds stress on the availability of clean fresh water supplies around the world [5].
Pb is a most hazardous heavy metal whose concentration in the atmosphere, soils, and waters underwent a striking worldwide increase during the 20th century associated with the development of the automobile industry and the generalized use of leaded gasolines until their banning in the 1980s [6,7]. Other industrial activities like the production of ammunition, batteries, pigments, and certain glasses have also contributed anthropogenic Pb inputs to the hydrosphere and the biosphere [8][9][10][11][12]. Mining activities have contributed to the long-term release of Pb as a result of the interaction of large volumes of fine-grained Pb-ores exposed in tails with groundwaters and running waters [13][14][15][16][17]. Furthermore, most catastrophic emissions of Pb environments, which have caused lasting damage in subaquatically and subaerially with great impact on soil fertility and productivity and wildlife diversity, have resulted from punctual wastewaters spills in mine sites and tails [18][19][20]. Pb was the main component of plumbing instalments in worldwide households until the 1970s, after which the use of lead pipes rapidly declined in industrialized countries. Interestingly, Pb-related pollution of drinking water has become a hot topic in recent years as information on the relationship between the presence of lead-tainted water pipes in old buildings and high Pb contents being found in the blood and bones of children in deprived neighborhoods in industrialized countries has reached the news [12,21]. Since Pb is potentially toxic to mammals, water pollution that results in excessive Pb intake often has dramatic impacts on human health, including long-lasting damage to the nervous, skeletal, circulatory, enzymatic, endocrine, and immune systems [17,18,[22][23][24][25][26]. This damage is particularly dramatic in the case of sensitive populations, like children, pregnant women, and elderly people [27][28][29][30].
The bioavailability of dissolved Pb in natural environments is largely controlled by sorption processes that develop during the interaction between Pb-polluted waters and the surface of rock-forming minerals like calcite and gypsum, which are especially abundant in sedimentary basins as main constituents of limestones (calcite) and sulfate evaporitic deposits (gypsum) [31][32][33]. Sorption of pollutants by mineral surfaces can take place through different mechanisms, including adsorption, absorption, and surface precipitation [34][35][36][37]. The first mechanism, adsorption, involves the bonding of the pollutant to the mineral surface by chemical bonds that can be relatively strong, or through longer-range coulomb forces and/or hydrogen bonds. The second mechanism, absorption, requires the diffusion of the pollutant into the structure of the mineral. Adsorption contribution to sorption is significant in the case of minerals like zeolites, whose structures contain large, interconnected channels that allow for ion exchange to occur at relevant rates [38]. In contrast, under the temperature conditions reigning in surface and subsurface environments, the kinetics ion diffusion through carbonate and sulfate mineral structures is most sluggish and absorption contribution to sorption can be considered negligible. The third mechanism, surface precipitation, involves the co-precipitation of the pollutant which incorporates in the crystal structure of new mineral phase. This new phase forms through the chemical reaction between the pollutant and ions that are released to the fluid phase due to the dissolution of the primary mineral. This sorption mechanism commonly progresses by the development of coupled dissolution-crystallization reactions. Coupled dissolution-crystallization reactions are commonplace in surface and subsurface geological environments, where they affect primary mineral phases that can be highly soluble, like halite (NaCl), relatively soluble, like gypsum, or sparingly soluble [39][40][41][42], like leucite or wollastonite [43][44][45]. Further evidence of the widespread nature of these reactions in natural environments have also been provided by a variety of macroscopic and microscopic experiments [46,47]. In fact, coupled dissolution reactions that involve different sulfate and carbonate minerals are considered an effective long-term way of removing inorganic pollutants from natural water and wastewaters [48][49][50][51] as well as reducing the increase atmospheric CO 2 [52,53]. Indeed, in situ AFM imaging has unraveled that coupled dissolution-crystallization reactions that involve calcite effectively remove pollutants like manganese [54], cadmium [37], lead [55], phosphate [56,57], and chromate [58] from aqueous solutions. Similarly, coupled dissolution-crystallization reactions that occur during gypsum-aqueous solution interaction can result in the sequestration of dissolved components like barium and strontium [59], lead [60], arsenate [61], phosphate [62,63], and carbonate [64] through their immobilization in the structure of new phase. Different authors have applied microscopic and macroscopic approaches to study the influence of the interaction of Pb-bearing fluids and common calcium-bearing carbonate and sulfate rock-forming minerals (namely, calcite, aragonite, gypsum, and anhydrite) in the fate of dissolved Pb in natural environments. This interaction has been addressed in numerous studies, most of which have been conducted at room temperature, using very fine-sized crystals (crystal size in the micrometer range), and considering circumneutral to slightly basic (pH i = 7-8) pHs as well as initial Pb ([Pb i ]) contents ≤ 1000 ppm [49,[65][66][67][68][69][70]. Some studies have assessed the long-term immobilization of Pb by calcium carbonate mineral surfaces conducting ageing experiments [66,69]. Recently, Di Lorenzo et al. [71] conducted an in-depth analysis of the characteristics of Pb immobilization by calcite and aragonite surfaces.
Although gypsum and calcite often appear spatially and genetically associated in sedimentary basins, in our knowledge, all previous experimental studies of the interaction between Pb-bearing aqueous solutions and gypsum and calcite have been conducted considering these phases separately, forgetting the fact that the dissolution of any of these Ca-bearing minerals can significantly alter physicochemical features of the aqueous solution like alkalinity, pH, or saturation state, thereby influencing the characteristics of the interaction of the fluid phase with the counterpart Ca-bearing mineral. Aiming to fill this gap, in this work we investigate the interaction of Pb-bearing aqueous solutions with gypsum and calcite, first separately and then simultaneously. We assess the degree of Pb removal from the liquid phase that is achieved depending on the solid phases involved in the interaction. Furthermore, in each case, we evaluate the relationship between the initial Pb concentration in the solution and the kinetics of the Pb removal. We focus our study on slightly acidic Pb-bearing aqueous solutions because acidic pHs are commonplace in running waters downstream Pb-bearing ore and mining sites. Our results evidence that more Pb becomes removed from the liquid phase when this simultaneously interacts with both gypsum and calcite than when the interaction only involves one of these phases. Apart from helping to better understand the fate of dissolved Pb in geological settings where sedimentary formations are predominant, taking this result into consideration might help to design more efficient strategies for decontaminating Pb-polluted waters through mineral coprecipitation processes.

Experimental Methods
Interaction experiments between gypsum and/or calcite natural crystals and Pbbearing aqueous solutions (50 mL) equilibrated with atmospheric pCO 2 (10 −3.5 atm.) were conducted under ambient temperature conditions (~20-22 • C). The Pb-aqueous solutions (1 and 10 mM) were prepared by dissolving reagent-grade Pb(NO 3 ) 2 (Panreac) in high purity deionized water (MiliQ) (18 MΩ cm). A set of experiments was carried out with 1 g of gypsum, calcite, or mixtures of gypsum and calcite (1:1) crystal grains with a relatively high size (average 3-4 mm) to avoid a fast dissolution of the primary minerals. To prepare these crystal grains, large natural optically clear selenite-type gypsum crystals (Alameda de la Sagra, Toledo, Spain) and Iceland spar quality calcite crystals (Proaza, Asturias, Spain) were first crushed in an agate mortar, then sieved to the selected size range, and, finally, cleaned in an ethanol bath. Furthermore, the so-obtained crystal grains were inspected under a stereomicroscope Leica MZ75 (Leica microsystems) and impurity fragments were hand-picked and removed by using 50× magnification. Pb-bearing solutions and crystals were placed in a glass vessel. The system was then sealed with parafilm to avoid contact with the atmosphere and minimize evaporation. The aqueous solution was gently stirred (stirring rate < 60 rpm) along the duration of the experiment (from 5 min up to 96 h) by means of a magnetic stirrer. To avoid crystal fragmentation due to stirring-related impacts, a floating magnet was used. Additionally, further experiments carried out with mixtures of gypsum and calcite or with single calcite were made last, up to 7 days or 30 days, respectively, to assess the impact of long interaction times on Pb removal. Independent experimental runs were conducted for each initial physicochemical condition considered. Furthermore, experiments were run triplicate to confirm experimental reliability.
The pH of the Pb-bearing solutions was measured prior to the beginning of the experiments and immediately after the experiments were terminated using a digital pHmeter (pH Meter basic 20-CRISON). Initial pH values were 4.59 ± 0.05 for the solution with The solid phase sample was immediately washed with ethanol and then dried overnight at room temperature. Afterwards, scanning electron microscope SEM, JEOL-6610LV (JEOL, Ltd., Akishima, Japan) imagining was used to identify dissolution features and newly formed phases on the surface of crystal grains. Energy-dispersive X-ray spectroscopy (EDX, INCA Energy 350; Oxford Instruments, Abingdon-on-Thames, UK) semiquantitative analyses on selected areas of the interacted samples provided information on the chemical composition of these newly formed phases. Raman spectra were collected from the surface of interacted crystals. Furthermore, in those experiments where precipitation from the bulk was observed, solutions were filtered after termination time and the recovered solids were also analyzed by Raman spectroscopy. All Raman spectra from 200 to 4000 cm −1 were obtained at room temperature on a microRaman BKTEK using as the excitation source the 532 nm line of an Ar ion laser and excitation times of 10 s. Powder X-ray diffraction (XRD) was used to identify the newly formed phases on the crystal surfaces and the precipitates from the bulk solution. Previously, interacted solid samples were powdered in an agate mortar. A PAN analytical X'Pert Pro diffractometer was used to collect the diffraction data from 5 to 80 • 2θ using CuKα radiation with a step size of 0.02 • . XRD patterns were processed with X'Pert HighScore Plus (PANalytical B.V., Erie, PA, USA) software.

Results
The processes that take place during the interaction between Pb-bearing aqueous solutions and gypsum, calcite, or mixtures of gypsum and calcite crystals were studied by monitoring both chemical changes in the liquid phase and mineralogical changes in the solid phase in macroscopic batch experiments. Significant differences regarding the kinetics of both the dissolution of the primary phases and the precipitation of newly-formed Pb-bearing secondary phases are observed depending on (i) the initial Pb concentration ([Pb] i ) and (ii) the solid phase(s) originally present in the system.

Chemical Evolution of the Liquid Phase
In all the experiments, the solution pH followed a similar evolution trend, which was characterized by a short initial period of marked pH increase that transitioned to an almost steady pH. This transition took place at different rates depending on the initial conditions, which determined that different final pHs were reached in the different experiments ( Figure 1). Thus, when the Pb-bearing aqueous solution exclusively interacted with gypsum, the pH remained approximately constant (under the unavoidable analytical error), with values that were close to 5 along the whole duration of the experiments. When the interaction took place with calcite, the initial pH increase was marked during the first minutes of the experiment. Afterwards, an asymptotic pH value (pH f = 5.6) was reached almost immediately when the interaction occurred with a solution with [Pb] i = 10 mM. In contrast, in the experiment where calcite crystals interacted with a solution with [Pb] i = 1 mM, the pH continued to slowly grow during the latest stages of the experiments, with the solution showing a pH f = 6.3 at termination time ( Figure 1).
The highest final pHs were found when Pb-bearing solutions simultaneously interacted with both gypsum and calcite crystals. In this case, an initial marked pH increase was also observed. This was followed by a steady slow pH increase to reach an asymptotic pH value~24 h after the beginning of the experiment, with pH f = 6.6 and 6.5 for solutions with [Pb] i = 1 and 10 mM, respectively.

Newly Formed Phases: Nature and Morphology
In all the interaction experiments, the formation of secondary phases was observed, regardless of [Pb] i and the solid phase(s) initially present in the system. When Pb-bearing aqueous solutions interacted with gypsum, a precipitate was observed to form in the bulk solution. Furthermore, SEM imaging of the surface of interacted gypsum crystals also evidenced the formation of a precipitate (Figure 4a,b). Both precipitates were confirmed as anglesite (PbSO 4 ) by XRD and Raman analysis, respectively (Figures 5 and 6). The morphology of anglesite crystals formed on gypsum surfaces during the early stages of the interaction appeared elongated along the b-axis and showed a habit defined by flat surfaces that belonged to the pinacoid {001} and the rhombic prism {210}, as well as strongly stepped surfaces roughly parallel to (100) (Figure 4a). The latter surfaces appeared serrated and consisted of steps belonging to {210}. As the interaction progressed, anglesite crystals become lozenge-like, with their habit dominated by the pinacoid {001} and further bounded by much smaller {210} surfaces (Figure 4b).  All diffraction peaks can be assigned to gypsum (Gp) and anglesite (Ang). The sign "*" indicates that one reflection can be asigned to both gypsum (Gp) or Anglesite (Ang). Different reaction pathways were observed during the interaction between Pb-bearing aqueous solutions and calcite crystals depending on [Pb] i . When [Pb] i = 1 mM, SEM imaging calcite surfaces showed the formation of a precipitate that consisted of elongated prismatic crystals (Figure 7a). All peaks in the XRD pattern of these crystals matched well those of cerussite (PDF 5-417) (Figure 8a). A more complex precipitation behavior was observed when a solution with [Pb] i = 10 mM interacted with calcite. In this case, the precipitate that formed on the calcite surface consisted of two differently shaped types of crystals: Elongated prisms and hexagonal thin plates (Figure 7b). The X-ray diffraction pattern of this precipitate showed the main features characteristic of cerussite together to a small diffraction peak at 34.075 (2θ) whose location matched well that of the most important reflection (110) of hydrocerussite (PDF 13-131) (Figure 8b). Raman spectra of elongated prism-like crystals obtained in surface areas where they were the only newly formed phases showed the typical characteristics of cerussite ( Figure 9). Taking into consideration the result of the X-ray diffraction analysis together with morphological criteria, the thin plate-like crystals were attributed to hydrocerussite.   Finally, an influence of [Pb] i in the nature of newly formed phases was also found during the interaction of Pb-bearing aqueous solutions with gypsum-calcite mixtures. This influence only regarded secondary phases that grew on the calcite surface, since anglesite was the only phase precipitated on gypsum surfaces regardless of [Pb] i (Figure 10a). Thus, when [Pb] i = 1 mM (Figure 10b), hydrocerussite formed on calcite surfaces. In contrast, the precipitate that formed when calcite interacted with a solution containing [Pb] i = 10 mM consisted of anglesite, hydrocerussite, and cerussite ( Figure 10c). Furthermore, in the latter case, the relative abundance of these three phases changed as the reaction progressed: The precipitate mainly consisted of hydrocerussite, with a minor content of anglesite and a very minor content of cerussite during the first 5 min of the interaction. Afterwards, anglesite and hydrocerussite contents progressively decreased, and cerussite became the main component of the precipitate after 24 h. A detailed inspection of the distribution of the precipitate on calcite surfaces evidenced a close spatial relationship between cerussite and hydrocerussite crystals, with most crystals of the former phase growing on the surface of hydrocerussite plates (Figure 10c). No morphological differences were observed between anglesite, cerussite, or hydrocerussite precipitated during the interaction of Pb-bearing solutions with gypsum-calcite mixtures or with gypsum or calcite crystals independently. It is worth noting that only anglesite crystals were observed to precipitate from the bulk solution when gypsum-calcite mixtures were used, regardless of [Pb] i .    [49,68] reported similar [Pb] drops in solutions that initially contained 1000 mg/L Pb after they interacted gypsum or anhydrite, respectively. According to these authors, [Pb] dropped due to the development of a dissolution-crystallization process. Dissolution-crystallization reactions have been invoked by different authors as the main mechanism responsible for the removal of different dissolved components from aqueous solutions interacting with gypsum crystals [61,68].
This process started as soon as a Pb-bearing aqueous solution that was undersaturated with respect to either gypsum or anhydrite entered in contact with crystals of any of these phases. The dissolution of a calcium sulfate mineral, either gypsum, anhydrite, or the hemihydrated bassanite, released Ca 2+ and SO 4 −2 ions to the Pb-bearing liquid phase, whose supersaturation with respect to anglesite (PbSO 4 ) progressively increased. Eventually, the supersaturation threshold that allowed for the formation of anglesite nuclei was overcome and anglesite precipitated after reaction times shorter than 5 min regardless of the [Pb] i concentration [68]. The precipitation of anglesite removed both SO 4 −2 and Pb 2+ ions from the aqueous solution and promoted further dissolution of the calcium sulfate phase. Further release of Ca 2+ and SO 4 −2 ions to the fluid promoted further anglesite precipitation, defining of a dissolution-precipitation loop that was operative to remove Pb from the aqueous phase as long as the balance between calcium sulfate crystal dissolution and anglesite precipitation kept the system undersaturated with respect to the former phase and supersaturated with respect to the latter. The progress of the dissolution-precipitation reactions did not induce significant changes in the aqueous solutions pH. The final [Pb] f = 3 mg/L detected in the experiments by Astilleros et al. and Morales et al. [49,68] as well as in this work derived from the solution being simultaneously equilibrated with anglesite and the initially present calcium sulfate phase. This value was directly related to anglesite solubility and, consequently, it represented the minimum [Pb] that could be achieved when anglesite precipitation was the main mechanism scavenging Pb from a Pb-bearing solution. This minimum [Pb] value (3 mg/L) was still 300 times too high to meet standard potability requirements for drinking water (10 µg/L), according to the World Health Organization and the European Community directive [73]. Therefore, although effective to rapidly remove large Pb amounts from aqueous solutions, the dissolution-crystallization reaction that operated during the interaction of Pb-bearing solutions with gypsum (and other calcium sulfate rock-forming minerals) could only be considered as an initial step in a strategy for the decontamination of highly polluted waters, which would need to be implemented together with complementary Pb-uptake mechanisms if aimed at the production of drinking water, or even for treating mildly Pb-contaminated waters.

Effectiveness of Calcite as Pb Scavenger
Pb removal during the interaction of Pb-bearing aqueous solutions with calcite crystals occurred at a very slow rate compared to that observed when the interaction took place with gypsum. Thus, while in the latter case  (Figure 7a,b) supported a dissolution-precipitation process as the main mechanism responsible for Pb removal during this interaction. As explained in the Section 1, dissolution-precipitation was the main operating mechanism that led to effective removal of a variety of dissolved pollutants from aqueous solutions interacting with calcite [37,55,58]. The dissolution-precipitation process that operated during the interaction of Pb-bearing solutions and calcite had similar characteristics to those of the gypsum dissolution-anglesite precipitation one described in Section 4.1. In contact with Pb-bearing aqueous solutions, calcite crystals underwent dissolution that resulted in the release of Ca 2+ and CO 3 −2 ions to the liquid. As the concentration of the released CO 3 −2 ions mounted, the pH of the solution increased. Both factors determined that the threshold supersaturation for the nucleation of lead carbonate phases was rapidly overcome in the solution. At this point, a calcite dissolution-lead carbonate(s) precipitation feedback loop was established. This loop-operated pH evolved differently in solutions with different [Pb] i , reflecting differences in the balance between the amount of CO 3 −2 ions that were released to the liquid phase through calcite dissolution and the amount of both CO 3 −2 and Pb 2+ ions that became removed from it by precipitation of lead carbonates. In the case of calcite interaction with a solution with [Pb] i = 10 mM, higher supersaturation with respect to lead carbonates was reached, which resulted in a larger amount of lead carbonate precipitation and, consequently, larger depletion of CO 3 2− from the solution than would occur during the interaction of calcite with a solution with [Pb] i = 1 mM. This larger CO 3 2− depletion from solutions with [Pb] i = 10 mM could explain that after an initial increase, pH rapidly reached a steady state value around 5.5. In contrast, the smaller CO 3 2− removal during the interaction of calcite with solutions with [Pb] i = 1 mM explained that the pH continued to slowly grow to approach values around 6.3 at the later stages of the interaction. The removal from the solution of CO 3 2− ions was associated with the removal of the acidic Pb 2+ , which added up to the explain the pH evolution in the latter. The impact of Pb 2+ removal on the pH evolution was much less significant in solutions ([Pb] i = 10 mM) because in this case, the lead concentration remained very high during the whole duration of the experiments. Di Lorenzo et al. [71] pointed out that a small increase in the pH of a solution bearing both lead and carbonate ions translated into a much higher increase of the solution supersaturation with respect to hydrocerussite than with respect to cerussite. According to this, one would expect that hydrocerussite was formed in experiments conducted with an aqueous solution with [Pb] i = 1 mM. However, this was not the case. The formation of hydrocerussite together with cerussite was observed when calcite interacted with solutions with [Pb] i = 10 mM. However, despite the higher pH increase undergone by solutions with [Pb] i = 1 mM during the interaction, only cerussite formed in this case. A plausible explanation for this was that most Pb removal from solutions with [Pb] i = 1 mM took place during the first 24 h of interaction, while pH solutions values remained below 6. Moreover, the different stoichiometry of cerussite and hydrocerussite might also be a key factor that determines the influence of pH in the precipitation of these two lead-bearing carbonates, as pointed by Di Lorenzo et al. [71].
The kinetics of Pb removal is defined by the balance between the rate of the two reactions involved in the calcite dissolution-lead carbonate(s) precipitation loop. Both cerussite and hydrocerussite are very sparingly soluble phases (pK cerussite = −13.13; pK hydrocerussite = −17.46) [72], which means that higher departures from equilibrium can result in the system during the interaction with calcite than during the interaction with gypsum [74,75]. In other words, the release of the same amount of Pb 2+ ions to the solution led to far larger supersaturation changes with respect to lead carbonates than with respect to anglesite. However, in the time scale of the experiments in this work, precipitation of anglesite more efficiently removed Pb from a Pb-bearing solution interacting with gypsum than the formation of cerussite and hydrocerussite did when this solution interacted with calcite. Reaching slightly larger Pb removals through lead carbonate precipitation requires far longer interactions of Pb-bearing solutions with calcite. The comparison of the variation of [Ca] in Pb-bearing solutions during their interaction with gypsum or calcite illustrates the overwhelming difference between gypsum and calcite dissolution rates (Figures 2a and 3a), which has been contrasted by numerous macroscopic and microscopic studies [65,66,68,71,76]. Calcite dissolves much more slowly than gypsum. Consequently, the release of carbonate ions to the liquid phase that resulted from calcite dissolution took place much more slowly than the release of sulfate ions due to the dissolution of gypsum, which completely overbalanced the faster Pb-removal that could be expected from the precipitation of more insoluble phases like cerussite and hydrocerussite compared to the precipitation of anglesite. Nevertheless, it is interesting to note that during the experiments of interaction between Pb-solutions and calcite crystals in the work, more Pb 2+ was removed from the solution than Ca 2+ was released to it. Since calcite dissolution was congruent, this unbalance could imply that there was some CO 2 atmospheric input leading to the formation of lead carbonate phase(s) and/or other Pb-uptake mechanisms than coprecipitation, such as Pb absorption and adsorption on calcite surfaces, were also operating and contributing to the net Pb removal.
From the results in this work, a scarce effectiveness of the interaction with calcite crystals to remove Pb from Pb-bearing aqueous solutions can be concluded. This contrasts with the conclusions reported by Godelitsas et al. [65]. These authors reported a drop of the Pb concentration of solutions bearing 10 to 1000 mg/L Pb to values below the experimental detection limit after 16 h of interaction with calcite crystals. The size of the calcite crystals used in both studies might be in the basis of the very different conclusions reached in both works. Godelitsas et al. [65] conducted interaction experiments with 100-200 micrometer-sized calcite crystals. Calcite crystals in this work were 3-4 mm-sized. The larger crystal surface area exposed to the interaction with the Pb-bearing aqueous solution of the former may have resulted in a significantly faster release of calcium and carbonate ions to the liquid phase, which would have allowed for a faster mounting of the supersaturation of the Pb-bearing solution with respect to lead carbonate phases. This faster supersaturation rate and reached supersaturation actual value could have facilitated the formation of significantly larger amounts of cerussite precipitates and, thereby, larger Pb-removals than observed in this work. Moreover, the larger surface area of smaller-sized crystals could also have provided more sites for other sorption mechanisms to operate. These other sorption mechanisms could have further contributed to the larger balance of scavenged Pb reported by Godelitsas et al. [65].

Effectiveness of Gypsum + Calcite as Pb Scavenger
Regardless of [Pb] i , the interaction of Pb-bearing aqueous solutions with gypsum and calcite mixtures rapidly resulted to drops in [Pb] that closely paralleled those observed when the interaction took place only with gypsum (Figures 2b and 3b) However, the interaction with calcite-gypsum mixtures resulted in larger Pb removals than those attained in experiments using only gypsum (and, obviously, using only calcite) when longer interactions were considered. The minimum [Pb] reached through the interaction with gypsum and calcite mixtures in the time set of experiments in this work was 1.9 mg/L in a solution with [Pb] i = 10 mM, 0.7 mg/L in a solution with [Pb] i = 1 mM, and after 7 days' interaction. The acidic conditions (pH < 5.6) at the beginning of all the experiments promoted the dissolution of the pristine minerals. Gypsum dissolution was a transport-controlled reaction [77] which occurred at a significantly faster rate than the dissolution of calcite. This fact together with the relatively high solubility of gypsum explained that anglesite precipitation not only took place on gypsum surfaces but also occurred on calcite surfaces and in the bulk solution. Thus, the presence of anglesite crystals on the surface of both interacted gypsum and calcite crystals, together with the similarity between the [Pb] evolution curves in these experiments and those conducted with only gypsum crystals, supported the interpretation that the precipitation of anglesite was main reaction responsible for the fast [Pb] drop to approach [Pb]~3 mg/L, during the first 48 and 96 h in solutions with [Pb] i = 1 and 10 mM, respectively. However, as these Pb-bearing solutions became equilibrated with gypsum and anglesite, anglesite precipitation could no longer occur. Further Pb removal mainly took place due to the precipitation of lead carbonates, most likely in the boundary layer around calcite crystals since these phases were only observed to form on calcite surfaces. During the interaction with gypsum and calcite mixtures, a similar pH evolution was observed in both Pb-bearing aqueous solutions; that with [Pb] i = 1 mM and that with [Pb] i = 10 mM ( Figure 1). Furthermore, in both cases pH values above 6 were reached after relatively short reactions times. This explained the formation of hydrocerussite on calcite surfaces regardless of [Pb] i . This phase was not observed to form from solutions with [Pb] i = 1 mM that interacted only with calcite, whose pH remained below 6 at the beginning of the experiment, as discussed in Section 4.2. Once lead carbonate precipitation became the main Pb removal operating mechanism, Pb removal slowed down. This slowdown was observed at latter stages of the interaction, when the slow calcite dissolution became the rate limiting process defining the kinetics of the calcite dissolution-lead carbonate precipitation feedback loop. This slowdown that occurred was more marked and occurred earlier when gypsum and calcite mixtures interacted with solutions with [Pb] i = 10 mM. A possible explanation for this slowdown could relate to significant Pb adsorption taking place on calcite surfaces that resulted in calcite dissolution inhibition.

Conclusions
Under the mildly acidic conditions in this work, the interaction of aqueous solutions bearing 10 mM ≥ [Pb] ≥ 1mM with gypsum, calcite, or mixtures of crystals or 50:50 mixtures of gypsum and calcite crystals resulted in a Pb removal that could be fast and extensive when gypsum was present in the system, but progressed at a comparably extremely slow rate in the absence of this phase. Regardless of its kinetics, in all the systems studied coprecipitation of Pb-bearing mineral phases was the main mechanism of Pb removal. These Pb-bearing phases formed as a result of the reaction between dissolved Pb and SO 4 2− and/or CO 3 2− ions released during the dissolution of the primary phases, gypsum, and/or calcite. The early formation of anglesite (PbSO 4 ) both in the bulk and on the surface of crystals of the primary phase(s) led to a rapid drop of [Pb] to~3 mg/L. This value is well above the accepted maximum [Pb] for drinking water. No further Pb removal was achieved through interaction of Pb-bearing solutions only with gypsum, no matter the duration of the interaction. Interaction of Pb-bearing solutions with calcite resulted in the formation of cerussite (PbCO 3 ) and hydrocerussite (Pb 3 (CO 3 ) 2 (OH) 2 ). The sluggish kinetics of calcite dissolution made Pb removal through precipitation of lead carbonates a very slow process, which resulted very inefficiently compared to the precipitation of anglesite, at least within the time set of experiments in this study. Interestingly, the simultaneous interaction of Pb-bearing aqueous solutions with crystals of both phases and gypsum-calcite crystals resulted in Pb removals that, depending on [Pb] i , were from as much as up to 5 times larger than those attained through interaction only with gypsum, and up to 500 times larger than those attained when solutions only interact with calcite. Further exploration of Pb removal from aqueous solutions simultaneously in contact with gypsum and calcite could help to shed light on the fate of lead in sedimentary basins as well as help to design more efficient strategies for decontaminating Pb-polluted waters through the integration of sequences of co-precipitation processes.