[American Journal of Science, Vol. 309, October, 2009, P. 633– 660, DOI 10.2475/08.2009.01] ELEMENTAL RELEASE RATES FROM DISSOLVING BASALT AND GRANITE WITH AND WITHOUT ORGANIC LIGANDS E. M. HAUSRATH*,**, A. NEAMAN***, and S. L. BRANTLEY* ABSTRACT. Bacteria, fungi, lichen and plants all produce organic acids, which can strongly affect weathering by increasing the solubility and mobility of elements. Leaching by organic acids may therefore produce trace element signatures which could record the presence of life in the rock record from early Earth. To elucidate this effect, long term column experiments were performed with powdered granite and basalt with and without 0.01 M citrate at pHⴝ6 for 45 weeks. Both granite (8.44 ⴛ 10ⴚ13, 3.39 ⴛ 10ⴚ13) and basalt (2.94 ⴛ 10ⴚ14, 6.47 ⴛ 10ⴚ14) dissolution rates mol (Ca, Mg) (mⴚ2 sⴚ1 respectively) were enhanced in the presence of citrate relative to the organic-free controls: granite (3.17 ⴛ 10ⴚ14, 4.4 ⴛ 10ⴚ15) and basalt (1.01 ⴛ 10ⴚ14, 1.04 ⴛ 10ⴚ14). Enhanced release of individual elements in the presence of citrate was strongly correlated with the stability constant of the citrate-element complex. Elements which might be useful as biosignatures are those elements that showed a strong enrichment in the presence of citrate: Zr, Sc and Mn (basalt), V and Zn (granite), and Y, La, Ce, Th, Ti, Al, P, Pb, Ni and Fe (both basalt and granite). Release of these elements from the rock material in the columns is consistent with dissolution of apatite ⴙ Fe sulfides ⴙ Fe oxides ⴙ augite in basalt and apatite ⴙ sphene ⴙ hornblende in granite. Similar groups of elements have been reported to be enriched in organic-rich rivers, suggesting leaching of strongly-complexed elements could be useful as biosignatures and may have left mineralogical traces on early Earth. Key words: Basalt, Granite, Weathering, Citrate, Biosignature, Trace elements. introduction Biota affect chemical and physical weathering by many mechanisms. One important mechanism for enhanced chemical weathering in the presence of biota is the secretion of organic acids. Bacteria, fungi, lichen and plants all produce organic acids, which can strongly affect weathering by increasing the solubility and mobility of elements. However, the effects and mechanisms of organic acids remain an area of active research and debate. A significant body of work has focused on mineral dissolution in the presence of organic acids (for example, Drever and Stillings, 1997, and references therein). Some minerals which have been dissolved in the presence of citrate include feldspar, quartz, augite, muscovite, kaolinite, illite, hornblende, apatite and phosphate rocks (Huang and Kiang, 1972; Manley and Evans, 1986; Bennett and others, 1988; Lundstrom and Ohman, 1990; Zhang, ms, 1990; Bennett, 1991; Welch and Ullman, 1993; Kpomblekou-a and Tabatabai, 1994; Stillings and others, 1996; Zhang and others, 1996; Blake and Walter, 1999; Zhang and Bloom, 1999; Welch and others, 2002; Wang and others, 2005). However, the mechanism by which organic ligands enhance dissolution is still not known. The effect of ligands on dissolution may be due to the direct effect of the ligands on the mineral surface, whereby they polarize and weaken the bond between the cation and mineral lattice (Furrer and Stumm, 1983). According to this model, the overall rate of dissolution is therefore the sum of proton-promoted dissolution and the * Department of Geosciences, The Pennsylvania State University, University Park, Pennyslvania 16802, USA ** Present address: Department of Geoscience, University of Nevada, Las Vegas, Nevada 89154, USA; [email protected] *** Facultad de Agronomı́a, Pontificia Universidad Católica de Valparaı́so, Casilla 4-D, Quillota, Chile 633 634 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates ligand-promoted dissolution. The equilibrium constant for sorption of ligands to a mineral surface increases with the equilibrium constant for formation of the ligandmetal complex in solution (Sigg and Stumm, 1981). However, indirect mechanisms have also been proposed, whereby the complexation of the ion in solution indirectly affects the dissolution of the mineral surface (Oelkers and Schott, 1998; Kubicki and others, 1999). Drever and Stillings (1997) in their review of the effect of organic acids on mineral dissolution, particularly feldspars, have found that a concentration of at least 1 ⫻ 10⫺3 M is necessary to cause a significant effect on dissolution. Concentrations of organic acids in soils are difficult to measure (Banfield and others, 1991), but some reported values for bulk soils include 1.5–3.0 ⫻ 10⫺5M (Shen and others, 1996) to ⬃1 ⫻ 10⫺3 to ⬃4 ⫻ 10⫺3 M for aliphatic organic ligands in modern soils (Stevenson, 1991). Values measured near tree roots can be much higher (Jones and Edwards, 1998). If organic acids are affecting mineral dissolution in field environments, such weathering might be a valuable biosignature, or signature of life. Microbial biosignatures can take a variety of forms, as recently summarized by Fisk and others (2006), including body fossils, stromatolites, trace fossils, reduced or oxidized minerals, alteration of geochemical cycles, fractionation of stable isotopes, chirality, metabolic byproducts, and organic molecules. Banfield and others (2001) have extensively reviewed mineral biosignatures. They discuss the interpretation of phases, trace element, major element, and isotopic compositions, surface composition and morphology, crystal morphology, particle size, spatial arrangement, aggregation state, and the presence of organic molecules that may indicate a biogenic origin. Extensive evidence has documented that micron-sized tunneling may be due to microbial action (Fisk and others, 2006) and that mineral pitting (Buss and others, 2003) and element depletion on the surface of a mineral (Kalinowski and others, 2000; Hausrath and others, 2008) can result from biotic impacts. Hausrath and others (2007) also tested the effect of methanogen growth on glass dissolution to determine whether a Ni biosignature might occur. They demonstrated that the observed increase in Ni release in the presence of methanogens was not due to the presence of cell exudates, low molecular weight organic acids, lysates, direct cell-mineral reactions, such as biofilms or pitting, but instead changes in pH. To determine whether the presence of organic acids in paleosols might be a useful biosignature distinguishable from reductive inorganic weathering, previous work has examined the dissolution of basalt and granite in the presence of organic acids with and without oxygen (Neaman and others, 2005a; Neaman and others, 2005b; Neaman and others, 2006). Total elements removed were calculated from the rocks reacted with a variety of organic acids, with and without oxygen present. Neaman and others conclude that enhanced release of Fe, P and Y indicate organic-rich conditions, and that enhanced release of Cu indicates oxygen-rich conditions (Neaman and others, 2005a). Here we expand upon this previous work, to perform column dissolution experiments dissolving Columbia River Basalt and Tuolomne River Series granite under oxic conditions with and without citrate. Column experiments using whole rock instead of separate minerals were chosen because they most closely simulate a natural soil environment. Citrate was chosen for these column dissolution experiments because it is a common secretion product of prokaryotes, fungi, lichens, and plant roots, a common degradation product of biomolecules, and a common constituent in modern soil solutions (Baziramakenga and others, 1995; Krzyszowska and others, 1996; Neaman and others, 2005b). Citrate has also been previously demonstrated to have the greatest effect on basalt dissolution from among aliphatic ligands acetate, formate, fumarate, glutarate, lactate, malonate, oxalate, and succinate, and aromatic ligands from dissolving basalt and granite with and without organic ligands 635 benzoate, gallate, phthalate, and salicylate (Neaman and others, 2005a; Neaman and others, 2005b). The reason that citrate has this large effect is likely two-fold 1) it is able to form a strong tridentate ligand and 2) aliphatic ligands have been previously demonstrated to have a stronger effect than aromatic ligands (Neaman and others, 2005, 2006). Trace elements were measured, to expand upon the work of Neaman and others in determining biosignatures (Neaman and others, 2005a; Neaman and others, 2005b; Neaman and others, 2006). We then compare our trace element results across a range of scales to batch dissolution experiments and modern rivers to test whether laboratoryobserved trace element signatures may in fact yield an important indicator of biological factors in natural weathering environments. methods Material Columbia River basalt, and granite from the Half-Dome Tuolumne River Series were collected and powdered for dissolution experiments in deionized water with and without 0.01 M citrate in flow-through column dissolution experiments. Basalt was obtained from the same quarry as the BCR-1 basalt reference sample of the U.S. Geological Survey (Flanagan, 1967), and the granite was sampled in Yosemite National Park at Olmsted Point along Highway 120 (Neaman and others, 2004). Samples were hammer-broken, cleaned with distilled water, crushed and powdered with a tungsten carbide jaw crusher and disk mill, sieved to 100 to 200 mesh (75-150 m), and cleaned of fine particles by ultrasonication, acetone washing, and drying at 60 °C. The concentrations of major and trace elements in the rocks were determined by X-Ray Fluorescence (XRF) and Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (Neaman and others, 2004). The concentrations in the basalt are very similar to the elemental compositions reported for the BCR-1 basalt reference sample (Govindaraju, 1994). Mineralogy of Basalt and Granite The mineralogy of the Half-Dome granite Tuolumne River Series has been previously determined (Bateman and Chappell, 1979), and the chemical composition of the minerals previously determined by electron microprobe microanalysis (Neaman and others, 2005b) (table 1). A grain mount of the powdered granite was examined optically and using Back-Scattered Electron Microscopy (BSE). Grains were determined to be primarily monomineralic (fig. 1). The powder used in the columns was prepared identically to powders used in previous batch experiments (Neaman and others, 2005a; Neaman and others, 2005b; Neaman and others, 2006), however, both basalt and granite used in the columns have a higher surface area than the powders previously dissolved in the batch experiments. The normative mineralogy of the Columbia River basalt was determined from its elemental composition (Neaman and others, 2005b). BCR-1 has previously been petrographically characterized as an aphanitic, hypocrystalline basalt with an interstitial texture of plagioclase laths, interstitial augite, partially devitrified brown glass and iron oxides (Flanagan, 1967). Previous X-ray diffraction (XRD) and electron probe microanalyses (EPMA) of the Columbia River basalt have documented the presence of the following major phases: plagioclase feldspar, alkali feldspar, quartz, and augite, as well as minor phases ilmenite, and magnetite-ulvospinel solid solution, and trace phases including fluorapatite, Fe/Ti oxides, Cu/Fe sulfide, and Fe-sulfide (Neaman and others, 2005b). The volume percent of the different minerals in the Columbia River basalt was quantified using Energy Dispersive X-ray Spectroscopy (EDS) elemental maps of 636 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates Table 1 Granite and basalt rock mineralogy 1 The chemical composition of minerals was determined by electron microprobe microanalysis (Neaman and others, 2005b). 2 Percentages of minerals determined by (Bateman and Chappell, 1979) by point counting. 3 Percentages determined from EDS images by image analysis as described in text. 4 Percentages determined from basalt normative calculation as described in text. 5 Simplified from (Neaman and others, 2005b). polished and carbon coated sample and the INCA-mapping software (Buss, ms, 2006). All magnesium was attributed to augite, all potassium to glass, and aluminum to plagioclase (where Al maps did not overlap with other minerals). The percentage of each of these volumes was determined, the total coverage of these maps scaled to 100 percent, and then averaged (table 1). Grain mounts of powdered basalt 75 to 150 m grain size identical to that used in the columns were examined optically and with BSE, and grains were determined to be primarily polymineralic (fig. 1). Surface Area Surface area and pore size distribution of samples before and after dissolution were measured by gas adsorption using a Quantachrome Autosorb-1 MP LP. Samples were prepared for surface analysis by degassing at 250°C under vacuum overnight (at least 12 hours or more) until the leak rate (pressure rise) was less than 30 micron/ minute. Specific surface area and pore structure of the minerals were determined by N2 sorptometry (ASAP 2010, Micromeritics). Surface area was calculated using multipoint adsorption data from the linear segment of the N2 adsorption isotherms between relative pressures of 0.05 and 0.3 (generally 5-6 points) using the Brunauer-EmmettTeller (BET) isotherm (Brunauer and others, 1938). If the y-intercept of the BET plot was less than zero, the lowest points were eliminated. Pore size distributions were calculated from desorption branch isotherms using the Barrett-Joyner-Halenda (BJH) from dissolving basalt and granite with and without organic ligands 637 Fig. 1. Backscattered electron micrograph of the unreacted granite (A) and basalt (B) grains identical to those within the dissolution columns. Granite grains appear to be primarily monomineralic while the basalt grains appear to be primarily multimineralic. The higher the atomic number, the brighter the mineral appears—therefore minerals such as magnetite and sphene appear white, hornblende, feldspars and quartz gray, and the foliated grains are biotite. Scale bars are shown in the lower right corner. method (Barrett and others, 1951), assuming the pores to be cylindrical, perpendicular to the mineral surface and closed on one end, and using the Halsey layer thickness equation (Halsey, 1948). Only data from the desorption isotherm between relative pressures of 0.995 and 0.3 were used (see table 2). According to these calculations, 638 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates Table 2 BET surface areas before reaction and after reaction with and without citrate1 Rock Before reaction Without citrate With citrate Basalt 12.03 12.06 9.54 Granite 0.193 0.33 0.3 1 m2/g. about half of the basalt surface area is attributable to pores that are about 3.5 nm in size (data not shown). Examination of Reacted Grains Although the columns had reached steady state by 45 weeks (fig. 2), and this paper examines the water chemistry from the first 45 weeks, solutions continue to flow through the columns. This is to allow future analyses of long term experiments similar to White and others (1999). Therefore, in order to not disturb the water chemistry analyzed in this work (the first 45 weeks), approximately 20 –30g of reacted powdered basalt and granite was removed from inside the columns after 94 weeks. To remove the sample, flow was stopped, the top of the column was removed, and sample scooped out 550 Basalt citrate 2500 500 2000 450 Si (µM) Si (µM) 400 1500 350 1000 300 250 500 200 Basalt no ligand 0 0 10 20 30 40 50 0 10 20 30 4 0 140 Granite citrate 1200 Granite no ligand 120 1000 100 Si (µM) Si (µM) 800 600 400 80 60 40 200 20 0 0 10 20 Time (week) 30 40 50 0 10 20 30 4 0 Time (week) Fig. 2. Si concentrations corrected for background versus time reveal that the columns had reached ⬃steady state by weeks 40 to 45, which is the time period we are focusing on in this paper. Si concentrations show three characteristic behaviors versus time (A) increasing with time (B) constant with time, and (C) and (D) initially decreasing rapidly and then remaining steady. from dissolving basalt and granite with and without organic ligands 639 for analysis. Because the samples were removed after the sampling for this manuscript was finished (45 weeks), the flow rates and elemental concentrations were not disrupted. These samples were rinsed twice with spectrophotometric grade acetone, and once with MilliQ DDI water, frozen for 24 hours, and then freeze-dried. Samples were examined under low vacuum Scanning Electron Microscopy (SEM), uncoated, for areas of rock coating, and etch pit formation. Experimental Setup Six columns were prepared, 3 of which contained 0.01 M citrate, and 3 distilled/ deionized water (MilliQ) (ligand-free). Columns were designed to be similar to columns in White and Brantley (2003). The input into each column was adjusted to pH 6 with concentrated high purity NH4OH, or 1 N HNO3, and was in equilibrium with atmospheric oxygen. Each condition (citrate or ligand-free) consisted of one column containing basalt, one containing granite, and one empty column (control). Basalt (350g) and granite (330 g) were wet packed in the columns: each column was filled with deionized water and the rock poured as a dry powder into the water-filled column. Each of the Pyrex columns had a total volume of 246 cm3, and therefore, assuming a density of 3 g/cm3 basalt and 2.75 g/cm3 granite, the pore volume of the columns is 129 cm3 (basalt) and 126 cm3 (granite). The input solutions flowed from input reservoirs above the columns down through 1/16⬙ (0.16cm) tubing into the bottom of the columns and then up through the columns. At the top of each column was a 0.2 m filter, which was connected by 1/16⬙ (0.16cm) tubing to the outlet which drained into another set of reservoirs, at which point samples were collected. Since the input reservoirs were higher than the columns, the effect of gravity was sufficient to cause the solutions to flow, which allowed there to be no moving parts, an important consideration in a long-term experiment such as this. In addition, since columns were wet-packed and flow occurred upwards, flow through the columns was saturated and development of preferential flow paths was assumed to be minimal. Flow rates were controlled by adjusting the height of the outlet, and clamps on the tubing leading to the column. Filters were changed as needed to prevent a decrease in flow rate due to blocking of the pores in the filters or when visual inspection indicated evidence of precipitation on the filter. For the first 3 months filters on the basalt and granite-containing columns with citrate were changed weekly. All solutions were made up to contain 0.05 percent azide, to prevent microbial growth (Na-azide for 0-14 weeks, Li-azide after 14 weeks). The solutions contained no pH buffer, as such buffers have previously been shown to impact mineral dissolution experiments (Brantley, 2008). Sampling The effluent discharge from the top of each column passed through a 0.2 m filter as previously described and was collected in bottles. A flow rate was determined by weighing the effluent discharged since the last sampling period and dividing it by that time period (1-2 weeks). Because the effluent collected for flow rates flowed into an open bottle with a large fluid surface area, the measured flow rate (mass of water/time between samples) was corrected for evaporation (0.21 ml/hr). Since samples collected for chemical analysis had a small fluid surface area, no correction was deemed necessary or performed. Approximately 10 to 15 ml of effluent was collected for major and trace element analysis by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) and Inductively Coupled Plasma Atomic Emissions Spectrometry (ICP-AES). Elements analyzed include Rb, Sr, Y, Zr, Mo, Ba, La, Ce, W, Pb, Th, Na, Mg, Al, P, S, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, K and Si. Sampling occurred approximately weekly throughout the first 3 months of the experiment, and at least every 2 weeks thereafter. Samples were first analyzed for pH, and then acidified with 640 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates high purity HNO3 to 3%v/v. Input solutions were sampled and analyzed by ICP-MS beginning in the 19th week. Concentrations used for calculations were generally analyzed by ICP-MS except for non-dilute samples for which ICP-AES concentrations were used. In some cases the sample concentrations were below detection, in which case a value of 10⫻ the instrument detection limit (IDL) was used in calculations of cumulative release, and zero for release rate calculations. We considered this to be the most realistic detection limit (the instrument detection limit is very low), and also the most conservative value for calculating the difference between the elemental concentrations released in the citrate-containing columns and the citrate-free columns. Using a smaller detection limit the difference could have been calculated to be larger than it actually was, whereas this is unlikely to occur if we use a conservative detection limit. In some cases results were determined to be anomalous, or were not measured, in which case a value of the mean of the remaining values was used in calculations. results Concentrations Elemental concentrations with time displayed three characteristic behaviors, either increasing, decreasing or remaining steady with time (fig. 2). However, it appeared that concentrations reached approximate steady state by weeks 40 to 45 (fig. 2). Steady state was defined as a constant concentration for a number of residence periods. The residence period of the water in the columns was approximately 3 pore volumes per week, or approximately 2 days (determined by dividing pore volume by flow rate). Since steady state concentrations had been reached, the average concentrations for weeks 40 to 45 (table 3) corrected for input concentrations were used in calculations. Flow Rates and pH Flow rates were variable with time. Flow rates in the granite citrate containing column increased slightly with time, and flow rates in the rest of the columns decreased slightly with time. (See table 4 for average flow rates weeks 40-45). The pH of the effluent from the columns decreased in the order: basalt citratecontaining ⬎ basalt ligand-free ⬎granite ligand-free⬃ control ligand-free ⬎granite citrate-containing ⬎citrate control (table 5). pH values for the basalt citrate column increased and then decreased through time, and the rest of the columns showed an initial drop in pH and then steady state pH values for the remainder of the experiment. Reacted Grains No pitting or coating was observed on reacted samples removed from inside the columns, and no difference was observable between unreacted samples, samples reacted with deionized water, and samples reacted with citrate for either basalt or granite. Measured surface areas of reacted grains (table 2) were used for calculation of release rates. calculations Release Rates Release rates were calculated as described in White and Brantley (1995) using the expression R⫽ 共Cout ⫺ Cin兲Q , Am (1) where Cout is the measured output concentration, Cin is the inlet concentration, Q is the flow, A is the specific surface area, and m is the mass. Here Cin was assumed to equal the 14,000 3,200 5.66 448,000 203,000 141 799 362 11,900 915 38.2 16.8 17.3 2,390 20,000 1,610,000 1,600 323 11 1,060 178,000 2,170,000 130,000 41,900 822 534 608 1,240 3.7 99.8 1 SD is one standard deviation. bd indicates below detection. Na K Rb Mg Ca Sr Ba Sc Ti V Cr Mo95 Mo97 W Mn Fe Co Ni Cu Zn Al Si P S Y Zr La Ce Pb Th 1,300 1,100 0.69 18,000 14,000 12 330 26 330 48 4.3 2.4 2.3 300 1,100 920,000 41 37 17 150 12,000 480,000 3,200 1,400 33 51 26 58 1.1 4.2 Basalt citrate concentration SD1 (x10-3) (µMx10-3) 12,700 3,700 14.1 44,700 112,000 193 0 4.61 2,030 86 2.5 4.5 4.6 50.6 1,500 35,400 29.2 354 92 420 82,000 105,000 86,700 510 30.9 6 156 207 8.7 270 1,900 1,200 2.2 7,300 7,100 16 0 0.6 210 12 1.9 4.7 4.8 7.5 200 4,800 3.5 770 41 560 11,000 38,000 5,600 330 1.5 3 10 12 7.7 39 Granite citrate concentration SD1 (x10-3) (µMx10-3) 11,500 bd 0.57 71,200 69,300 bd bd 2.75 1.55 26.3 bd 5.9 6 900 320 90 42.5 0.017 7.7 180 1,120 430,000 bd 59,000 bd 1.05 bd 0.0071 0.37 0.069 0 0.13 0.075 0.86 15,000 1.4 1.4 250 19 45 2.6 0 9.9 150 160 110,000 0.7 0.68 4.4 0.34 5,800 2,300 990 Basalt no ligand concentration SD1 (x10-3) (µMx10-3) Average concentrations and standard deviations (M) (Weeks 40 – 45) corrected for background Table 3 4,600 5,800 21.3 650 4,800 110 bd 0.21 0.61 5.7 1.4 2.7 2.7 46.6 56 99 22.3 0.61 74 87.3 56.8 38,100 76 273 bd 0.59 bd 0.071 0.55 0.048 0 0.78 0.041 0.44 0.17 0 1.2 1.9 1.5 1.5 4.5 11 0 3.7 0.3 19 0 5.9 3,700 40 18 1,600 1,800 3.2 270 1,300 53 Granite no ligand concentration SD1 (x10-3) -3 (µMx10 ) from dissolving basalt and granite with and without organic ligands 641 642 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates Table 4 Flow rates weeks 40 – 45 corrected for evaporation (0.21 ml/hr) Column Basalt with citrate Granite with citrate Control with citrate Basalt no ligand Granite no ligand Control no ligand Average flow rate (ml/hr) 1.7 2.7 2.5 2.2 2.6 2.7 Standard deviation (ml/hr) 0.2 0.4 0.7 0.6 0.3 0.7 measured concentrations of elements in effluent from the control columns (citrate and ligand-free). This accounts for any leaching from any part of the columns, although such leaching appears to be negligible. Release rates and standard deviations for each element for weeks 40 to 45 are reported in table 6. Cumulative Release The cumulative release (B) of each element from each column was also calculated over the duration of the experiment (30 samples from 45 weeks) using the expression 冘 共C 30 B⫽ n out n ⫺ Cin 兲Qn tn, (2) n⫽1 where Cout is the measured output concentration, Cin is the inlet concentration (again estimated as the output of the control columns), Q is the flow, and t is the time between sampling. The percentage (P) of the total element released in the rock-containing columns is calculated using the expression P⫽ B Cm (3) where B is the cumulative release over the 45 weeks of the experiment, C is the concentration of the element in the basalt or granite rock, and m is the mass of the rock in the column (table 7). Element Mobility To assess elemental mobility in a soil, element concentration in a soil can be normalized by the concentration of an assumed immobile element (often Ti, Zr, and Nb) to calculate the fractional mineral loss or enrichment (Anderson and others, 2002) i,j ⫽ Cj,w Ci,p ⫺ 1. Cj,p Ci,w (4) Table 5 pH values weeks 40 – 45 Citrate Ligand-free Basalt 7.33 7.04 Granite 6.19 6.66 Control 6.07 6.61 from dissolving basalt and granite with and without organic ligands 643 Table 6 Average element release rates and standard deviations week 40 – 45 Basalt citrate Granite citrate Basalt no ligand Granite no ligand RR1x1016 SD2x1016 RR1x1016 SD2x1016 RR1x1016 SD2x1016 RR1x1016 SD2x1016 Rb Sr Y Zr Mo95 Mo97 Ba La Ce W Pb Th Na Mg Al P S Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn K Si 1 2 0.0082 0.205 1.19 0.770 0.0240 0.0250 1.17 0.88 1.80 3.43 0.0044 0.140 20.3 647 256 189 60.6 294 0.520 17.3 1.32 0.0550 29.0 2400 2.32 0.470 0.010 1.55 4.6 3120 0.0013 0.032 0.15 0.042 0.0035 0.0034 0.59 0.12 0.24 0.31 0.0025 0.018 3.1 57 10 21 6.7 34 0.020 1.7 0.05 0.0030 3.7 1400 0.26 0.063 0.022 0.33 1.8 700 1.05 14.6 2.35 0.40 0.32 0.33 0.10 1.9 0.40 0.33 0.30 0.31 0.0083 0.0025 1.42 7.4 0.33 4.1 0.0012 0.0083 0.0084 0.0012 0.0017 0.0018 0.013 0.18 0.18 0.024 0.10 0.10 11.9 15.8 3.8 0.71 20.3 953 3390 6200 6600 21 8440 0.35 153 6.5 0.21 114 2680 2.21 26 7.2 20 280 7900 1.9 2.5 0.6 1.24 0.16 0.72 0.00030 0.00034 3.0 0.000046 0.000061 57 16.9 4.6 670 104 28 1100 1.61 0.37 1200 30 76.8 6.4 940 101 28 0.05 0.0041 0.0018 10 0.0010 0.0013 1.1 0.0376 0.0086 0.17 21 0.47 0.12 460 0.11 0.11 0.36 0.062 0.016 63 4.1 0.007 0.015 33 0.11 0.22 110 2900 604 190 0.0015 3.09 0.018 0.0031 300 44 1.3 1.7 5.9 317 0.012 0.007 0.38 0.031 3.2 1.1 1.48 0.020 4.9 1.0 380 2510 0.0026 0.56 0.037 0.0025 110 19 2.1 2.8 9.1 99 0.012 0.016 0.11 0.076 1.6 2.6 0.30 0.026 1.3 2.4 120 280 Release rate (mol m⫺2 s⫺1). Standard deviation (mol m⫺2 s⫺1). Here i,j is the fraction of mobile element or mineral j lost (i, ⬍ 0) or gained (i, ⬎ 0) assuming that element or mineral i is immobile. C is the concentration of the immobile and mobile elements in the parent and weathered materials (w and p refer to weathered and parent material respectively). Ti was used as an immobile element to calculate for each of the elements in each of the columns. Despite the fact that Ti is much more mobile in the presence of citrate than without, it is still very immobile, with only 0.14 percent removed from the basalt and 0.35 percent from the granite in the presence of citrate (table 7). The cumulative loss from each of the columns (B from equation 2) was subtracted from the total elemental content of the original columns (Cp, also represented as the product Cm in equation 3) to obtain Cw for both mobile and immobile elements (i,j) (table 8). The percentages removed (table 7) can also be 644 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates Table 7 Percentages of element removed and the ratio of percent removed with citrate to percent removed without citrate Na K Rb Mg Ca Sr Ba Sc Ti V Cr Mo95 Mo97 W Mn Fe Co Ni Cu Zn Al Si P S Y Zr La Ce Pb Th Basalt citrate % removed 0.43 0.071 0.46 2.3 0.64 0.36 0.4 1.6 0.14 0.42 0.24 bd bd bd 4.6 6.4 bd 4.6 1.3 2.2 0.29 0.93 8.2 bd 7.5 0.98 13 11 0.37 13 Granite Basalt no Granite no Basalt citrate % ligand % ligand % citrate:no removed removed removed ligand 0.55 2.2 1.5 0.19 0.1 0.2 0.2 0.36 0.23 0.27 0.41 1.7 1.6 0.61 0.13 3.7 1.8 0.31 0.16 2 0.69 0.036 0.084 10 0.091 bd 0.008 NA 0.27 0.028 0.039 55 0.35 4.4E-05 0.00015 3200 0.63 0.052 0.017 8.1 0.17 0.018 0.065 14 bd bd bd NA bd bd bd NA bd bd bd NA 2.1 0.093 0.25 50 0.64 0.0002 0.00069 33000 bd bd bd NA 3 0.056 0.05 82 12.4 0.047 4.7 27 7.2 0.36 0.039 6.2 0.28 0.0032 0.0018 91 0.13 0.3 0.03 3.1 87 0.023 0.058 360 bd bd bd NA 2 0.0038 0.0079 2000 0.065 0.0056 0.016 170 7.9 0.0092 0.0048 1400 4.9 0.0064 0.0036 1800 1.5 0.011 0.011 32 31 0.07 0.01 190 Granite citrate:no ligand 0.38 0.52 0.56 13 11 8.2 11 6.8 2300 36 2.7 NA NA NA 8.3 930 NA 59 2.6 190 160 4.4 1500 NA 250 4.1 1700 1400 150 3200 Calculated percentages of less than 1 ⫻ 10⫺5 are reported as below detection (bd). used to determine which elements might indicate the presence of organic acids over geologic time. Mineral Dissolution Rates Mineral dissolution rates were calculated for the basalt and granite columns by inverse modeling using PHREEQC (Parkhurst and Appelo, 1999). The input solution was assumed to be pure water at pH ⫽ 6 (input pH) in contact with atmospheric CO2, and the output solution was the average concentration of weeks 40 to 45 corrected for the concentrations in the control columns (table 9). An effort was made to obtain the simplest model that fit the data, and therefore the only minerals allowed to precipitate 645 from dissolving basalt and granite with and without organic ligands Table 8 t values less than ⫺0.1 La Ce Th P Cu Basalt Citrate -0.13 -0.11 -0.13 NA NA Basalt no ligand -0.000091 -0.000063 -0.00070 NA NA Granite citrate NA NA -0.31 -0.86 -0.12 Granite no ligand NA NA -0.000095 -0.000095 -0.047 NA indicates that those elements did not have t values less than ⫺0.1 in those rocks. were amorphous silica, gibbsite, amorphous iron oxide, Mg-containing saponite, and kaolinite. Similarly, minerals allowed to dissolve included plagioclase (of the specific composition in each rock), hornblende, quartz, fluorapatite, magnetite, and augite as well as secondary kaolinite (see table 10 for mineral reactions). Since the reacting solutions entered the columns in equilibrium with atmospheric oxygen, and the oxygen participates in mineral reactions, it was necessary to artificially allow Fe2⫹ to precipitate, as if it had first been oxidized to Fe3⫹ and then precipitated (see table 10). In order to allow fluorapatite to dissolve, it was necessary to input fluoride concentrations into PHREEQC for the input and effluent solutions. However, since fluoride had not been measured during the experiment, fluoride concentrations were unknown. Therefore, effluent fluoride concentrations of 10 M were used for each column, but uncertainty on the fluoride concentrations was set to 1000 percent to allow the model to adjust fluoride concentrations during model calculations based on measured elemental concentrations. As an estimate of uncertainty, 1 standard deviation around the average concentration during weeks 40 to 45 (table 3) was assumed. Uncertainty was estimated at 40 percent for Ca (the granite-containing column with no ligand) and Ca and Mg concentrations (the granite-containing column with citrate). The increased uncertainties in Ca and Mg were necessary to allow PHREEQC to successfully produce models for the granite-containing columns, and are likely due to the low concentrations of Mg and Ca in the effluent from the granite. Moles of mineral dissolved per kg of water as calculated by PHREEQC were converted to mineral dissolution rates using the expression V⫽ 共MQ兲 , Am (5) Table 9 Modeled solutions in PHREEQC Condition T(ºC) pH Basalt citrate 20 7.3 Basalt no ligand 20 6.1 Granite citrate 20 6.2 Granite no ligand 20 6.7 1 2 Na2 14 11.1 12.7 4.55 K2 P2 Ca2 3.17 130 203 4.61 0.0051 69.3 3.65 86.7 112 5.79 0.756 4.75 Below detection, therefore the value used was 10⫻ the IDL. All concentrations in M. Si2 2170 427 105 38.1 Al2 178 1.12 81.5 0.0568 Fe2 1610 0.09 35.4 0.0987 Mg2 448 71.2 44.7 0.645 646 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates Table 10 Mineral dissolution and precipitation reactions in PHREEQC Mineral phase labradorite oligoclase hornblende quartz fluorapatite magnetite augite kaolinite amorphous silica gibbsite amorphous amorphous glass Mg-saponite Mineral reaction Ca0.5Na0.5Si2.5Al1.5O8 + 6H+ +2H2O = 0.5Ca2+ + 0.5Na+ +2.5H4SiO4 +1.5Al3+ Ca0.3Na0.7Si2.7Al1.3O8 + 5.2H+ + 2.8H2O = 0.3Ca2+ + 0.7Na+ + 2.7H4SiO4 + 1.3Al3+ Ca2Mg3.5Fe1.5Si8O22(OH)2 + 14H+ + 8H2O = 2Ca2+ + 3.5Mg2+ + 1.5Fe2+ + 8H4SiO4 SiO2 + 2 H2O = H4SiO4 Ca5(PO4)3F + 3H+ = F- + 3HPO42- + 5Ca2+ Fe3O4 + 8H+ = 2Fe3+ + Fe2+ + 4H2O MgFe0.5Ca0.5Si2O6 + 4H+ + 2H2O = Mg2+ + 0.5Fe2+ + 0.5Ca2+ + 2H4SiO4 Al2Si2O5(OH)4 + 6H+ = H2O + 2H4SiO4 + 2Al3+ SiO2 + 2H2O = H4SiO4 Al(OH)3 + 3H+ = Al3+ + 3H2O Fe+3 oxide Fe(OH)3 + 3H+ = Fe3+ + 3H2O Fe2+ oxide Fe(OH)2 + 2H+ = Fe2+ + 2H2O K0.125Na0.125Al0.25SiO2.5 + 1.5H2O + H+ =0.125K+ +0.125Na+ + 0.25Al3+ +H4SiO4 Mg3.165Al0.33Si3.67O10(OH)2 + 2.68H2O + 7.32H+ = 0.33Al3+ + 3.165Mg2+ + 3.67H4SiO4 where V is the dissolution rate in moles m⫺2s⫺1, M is the moles of mineral dissolved per kg of water as calculated by PHREEQC, Q is the flow rate in Ls⫺1, A is the specific surface area of the rock in the columns (m2g⫺1), and m is the mass of mineral in the columns (g). Surface area was assigned to the different minerals based on mass as shown in equation (5). In some cases PHREEQC produced more than one value for moles of mineral dissolved, in which case we report the full range in calculated rates (table 11 and fig. 3). Experimental and Computational Uncertainties These column experiments were designed after White and others (1999) to flow for long periods of time. They were therefore designed with no moving parts (such as peristaltic pumps) that can break under such long collection times. This experimental design, while it allows the collection of data that would not otherwise be possible, results in some uncertainties. Here we discuss some of these uncertainties, including variations in flow rate, secondary mineral precipitation, and oxidation state. Although the flow rates are low and somewhat variable, the columns and the residence times are relatively long (⬃2 days). During the period that we report release rates, flow rates were particularly carefully controlled (table 5). Therefore, the effects of the variations in flow rates on mineral dissolution rates are likely to be minimal. Secondary mineral precipitation was observed on the filters, although not on the grains by SEM, as discussed above. We included this effect in our PHREEQC modeling, by allowing the precipitation of secondary phases. Ideally, this precipitation would have been quantified. However, since the precipitation was primarily observed on the columns containing citrate, it does not change the conclusions: the presence of precipitates merely means that the observed enrichments in the presence of citrate are minima. Entering solutions were in equilibrium with atmospheric O2; however, mineral reactions within the columns (for example, oxidation of Fe2⫹ present in the parent minerals) likely lowered the oxygen concentration within the columns. These columns were designed to be similar to long-term columns used by White and others (1999), who document the oxidation of sulfide throughout their similar granite-filled column. In addition, our PHREEQC modeling requires the oxidation and precipitation of iron. 1 Dissolution rate (mol m⫺2 s⫺1). NP indicates that mineral is not present in that rock. Augite Glass Plagioclase Fluorapatite Magnetite Biotite Quartz K-feldspar Hornblende Basalt citrate Max Min dissolution dissolution rate1 x1012 rate1 x 1012 0.18 0.17 0.016 0.0078 0.012 0.0079 0.014 0.015 0.59 0.17 NP NP NP NP NP NP NP NP Basalt no ligand Min Max dissolution dissolution rate1 x 1012 rate1 x1012 0.077 0.069 0 0 0.0099 0.0083 0 0 0.0033 0.0011 NP NP NP NP NP NP NP NP Granite citrate Max Min dissolution dissolution rate1 x 1012 rate1 x1012 NP NP NP NP 0.38 0.28 183 176 19 7.1 0.89 0 0 0 0.15 0 3.7 2.5 Dissolution rates from PHREEQC Table 11 Granite no ligand Max Min dissolution dissolution rate1 x1012 rate1 x 1012 NP NP NP NP 0.074 0.063 0.23 0.071 0.0012 0.00040 0.73 0 0.78 0 0.12 0 0.36 0.25 from dissolving basalt and granite with and without organic ligands 647 648 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates Fig. 3. Dissolution rates in the presence of (filled symbols) and without citrate (open symbols). All values from this study are plotted versus final pH. In each case, the range in dissolution rates represents the largest and smallest values of the range in models (table 11), and are compared to data from the literature (line art symbols). Dissolution rates for (A) bulk basalt rock and basalt glass (calculated as described in methods), (B) fluorapatite, (C) magnetite, (D)augite, (E)hornblende, and (F) oligoclase. The fact that the magnetite dissolution rates are smaller than other values from the literature suggest that Fe may be precipitating. from dissolving basalt and granite with and without organic ligands 649 Therefore, these columns are assumed to be oxic, although not in equilibrium with oxygen throughout the length of the columns. discussion Implications for Biosignatures If microorganisms, lichens or plants produced organic acids which leached rocks and soil, the differential mobility of trace elements might provide a biosignature of their presence. Previous batch experiments suggest that differential mobility of the elements P and Y in paleosols can be attributed to the presence of organic acids and host mineral phase (Neaman and others, 2005a; Neaman and others, 2005b). Since geochemical behavior is observable at multiple scales from the molecular to the laboratory and field scale, we here compare our results to phenomena observed at the molecular, laboratory and field scales to determine whether observed behaviors are consistent across these scales, and therefore indicative of organomarkers, observed at the field and laboratory scales and predicted by the molecular mechanisms of organic acids. Metal Complexation Elements that are preferentially removed by citrate in these column experiments (elements with a ratio of B with citrate to B with no ligand ⱖ 30) are Zr, Sc and Mn from basalt, Y, La, Ce, Th, Ti, Al, P, Pb, Ni and Fe from both basalt and granite, and V and Zn from granite (fig. 4). Figure 5 indicates a plot of the ratio of the percentage of the element released (P) with citrate to without citrate, versus the NIST stability constant of the metal with citrate (Martell and Smith, 2001). Some have argued that the release rate of elements from minerals in the presence of a metal-complexing ligand may be controlled by the stability constant of the metal with the ligand (Casey and Westrich, 1992; Neaman and others, 2005a; Neaman and others, 2005b; Goyne and others, 2006). Figure 5 suggests a strong correlation between the ratio of the release with citrate to that without citrate with the NIST stability constant of the metal with citrate in solution. The labeled elements (Th and Al in basalt) do not follow the trend as strongly as other elements. This may be due to reprecipitation, or differences in mineralogy between basalt and granite. Other elements not present in the NIST database (Zr, Mo, W, P, S, Sc, Ti, V, Cr, Si) are not compiled in this figure. In addition to the complexing effect of the ligand, the elemental release rate is strongly affected by the host mineral phase. Due to the fine grained nature of the basalt, no attempt was made to analyze mineral grains. Apatite is assumed to host La, Ce, Sr, P, Y, Th based on previous work (Neaman and others, 2005b; Prowatke and Klemme, 2006). Augite is assumed to host Mg, Sc, V, Cr, Mo, W, Mn based on previous work (Hart and Dunn, 1993), as well as knowledge of chemical reactivity. Glass is assumed to host K, Rb, Ba, based on previous work (Ho and Cashman, 1997) as well as assumptions as to the similarity of the elements. The Fe-Ti minerals were assumed to host Ti and Zr based on previous work (Neaman and others, 2005b). The copper iron sulfides were assumed to host Co, Ni, Zn, and Pb, as well as Cu and S, based on the chalcophyle and siderophile nature of these elements. Bateman and Chappel (1979) have extensively characterized the Tuolomne river series granite. They document that Ni, Cr, V, Sc, Co, Mn, and Zn are present in biotite and hornblende, Rb and Pb in both biotite and feldspar, Mn, V and Cr in magnetite, Sr in alkali feldspar, Th in sphene, and La, Ce, and Y in sphene, hornblende, and apatite in that relative order. Zr is primarily present in zircons, and Cu and S in Cu-Fe sulfide inclusions in Fe oxide phases (Neaman and others, 2005b). Based on other work on the Sierra Nevada batholiths, we infer that Ba is present primarily in biotite (Dodge and others, 1982). Molybdenum is assumed to be present in magnetite (Kuroda and Sandell, 1954). 650 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates 10 Ratio citrate: no ligand 10 10 10 10 10 5 Basalt Granite 4 3 2 1 0 -1 10 Na K RbMgCa Sr Ba Sc Ti V CrMnFe Ni CuZn Al Si P Y Zr LaCePb Th Element Fig. 4. Ratio of percent of element released with citrate (Pcitrate) to percent of element released without citrate (Pnoligand) for basalt and granite. High ratios (⬎30) indicate elements that may be potential biosignatures (for example, Y, La, Ce, Th, Al, P, Ti, Fe, Ni for both basalt and granite, Zr, Sc, and Mn for basalt, and Pb, V, and Zn for granite). Line is drawn at a ratio of 30. Elements are organized after the periodic table, with first s block elements, from smallest to largest atomic radius, then d block elements, then p block and f block elements. This allows the comparison of elements with similar properties. Therefore, broadly, the elemental organomarkers documented here may also be consistent with mineralogical biosignatures. In basalt, if P, Y, La, Ce, and Th are present in apatite, then their coupled enhanced release suggests that the dissolution of apatite may be an indicator of organic acids in the soil environment. Apatite also dissolves in inorganic solutions, although more slowly. If Pb and Ni are present in Cu-Fe sulfides, then their coupled release may reflect the oxic dissolution of these phases. The enhanced release of Zr and Ti may reflect the enhanced dissolution of Fe-Ti and Fe oxides. Sc and Mn are both inferred to be present in augite, and are significantly enhanced in the presence of organic acids. In granite, the large percentage of phosphate lost documents the loss of significant apatite. The large percentage of Th (in sphene) documents the large percentage of sphene lost; La and Ce are both present in apatite and sphene as well as hornblende. V, Zn, Pb and Ni are all present in biotite and hornblende. Therefore, the trace elements released are consistent with the enhanced loss of apatite, sphene, hornblende and biotite. Laboratory Scale: Elemental Release With and Without Citrate Basalt batch experiments.—Elemental release from our column experiments can be compared to release from previous batch experiments with basalt (Neaman and 651 from dissolving basalt and granite with and without organic ligands 10000 A) Basalt Ratio of release citrate:noligand 1000 Al 100 10 Th 1 0.1 10000 0 2 4 6 8 10 12 2 4 6 8 10 12 B) Granite 1000 100 10 1 0.1 0 log K1 for metal-citrate Fig. 5. The ratio of the percentage released with citrate (Pcitrate) to without citrate (Pnoligand) versus the stability constant for the element with citrate (K1 ⫽ [ML]/[M][L] where [ML], [M] and [L] are concentrations of the element-ligand complex, free ion, and free ligand respectively). The stability constant for Fe3⫹ was used as the columns were exposed to the atmosphere. Stability constants were obtained from the NIST data base. Elements not included in the NIST database include: Zr, Mo, W, P, S, Sc, Ti, V, Cr, and Si. The pK3 value of citric acid ([HL]/[L][H]) is at pH⫽5.75 (NIST Database 46). Therefore, in all of the columns, citrate should be fully deprotonated, and present as citrate ligand. others, 2005a; Neaman and others, 2005b) and granite (Neaman and others, 2006). These previous batch experiments were also designed to test the effect of citrate in creating biosignatures. Therefore, a comparison as to whether the same biosignatures were inferred in the oxic batch experiments as in the present study is an interesting study in scaling. The percentage of each element released with and without citrate is relevant to whether that element may serve as a useful biosignature for citrate. However, since the previous batch experiments and present column experiments were performed for different lengths of time, the percentages of each element released cannot be simply compared. Therefore, we normalize each element released from the basalt dissolution experiments to the percentage of silicon released. When we compare the normalized elemental release from the oxic basalt batch experiments and the oxic basalt dissolution columns, we find that the normalized percentage is within a factor of two for the following elements: Al, Fe, Mg, P, Ti, V, Y, Zr, and Ba. In contrast, Cr, Cu and Rb are much more strongly enhanced relative to Si in the batch experiments than in the column experiments. The basalt dissolution rates measured by Neaman and others (2005b) are significantly higher than the rates calculated in these column experiments (table 6), which is not surprising given the different durations of the experiments, the different waterrock ratios, and the fact that the batch experiments include initial rates. Release rates 652 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates with citrate from Eick and others (1996a, 1996b) were not compared to our results because only parabolic release rates were given. Granite batch experiments.—Similarly, we here compare normalized total element release for the granite dissolution experiments. Silicon was not reported for the granite batch dissolution experiments, therefore we normalized to Mg. The normalized percentage is within a factor of two for the columns and batch experiments for: Fe, P, Ti, V, Cu and Rb. Much less Al, Cr, and Y, and significantly more Ba was released from the batch experiments than the column experiments normalized to Mg. Laboratory Scale: Dissolution Rates With and Without Citrate The logarithm of the basalt rock dissolution rate (mol m⫺2 s⫺1) calculated from the Si release rate (equation 1) is significantly higher (⫺12.51 log dissolution rate in mol m⫺2 s⫺1) than the log basalt dissolution rate without citrate (⫺13.22 log dissolution rate in mol m⫺2 s⫺1) (fig. 3A). This trend was also observed in previous studies comparing basalt dissolution rates with and without citrate in batch experiments (Neaman and others, 2005b). However, dissolution rates from this study are much slower than other basalt dissolution studies (fig. 3A) (Gislason and Eugster, 1987; Eick and others, 1996a; Eick and others, 1996b; Oelkers and Gislason, 2001; Gislason and Oelkers, 2003; Wolff-Boenisch and others, 2004). Such a discrepancy is likely due to the long duration of our experiments and also to the column method itself. In dissolution columns, solutions can become relatively concentrated compared to batch experiments, slowing the rate of dissolution. In basalt, the mineral dissolution rates calculated by PHREEQC using equation 5 for augite, glass, magnetite and apatite were enhanced by more than a factor of two in the presence of citrate in comparison to the absence of citrate (table 11). Similarly, for the granite, the mineral dissolution rates calculated using PHREEQC for plagioclase, apatite, magnetite and hornblende were enhanced by more than a factor of two in the presence of citrate compared to the absence of citrate (table 11). Dissolution rates of these minerals are shown in figure 3, compared to previous literature values. Data are shown either as rate laws derived from multiple experiments (Bandstra and Brantley, 2008), or from individual experiments (Siegel and Pfannkuch, 1984; Schott and Berner, 1985; Holdren and Speyer, 1987; Sverdrup, 1990; Casey and others, 1991; White and others, 1994; Stillings and Brantley, 1995; Stillings and others, 1996; Hoch and others, 1996; van Hees and others, 2002). Dissolution rates are also compared to rates from the literature for experiments containing citrate for basalt (Neaman and others, 2005b), oligoclase (Lundstrom and Ohman, 1990; Stillings and others, 1996), apatite (Welch and others, 2002), and hornblende (Zhang, ms, 1990; Zhang and others, 1996; Zhang and Bloom, 1999). More than one data point for a condition indicates a range in values (table 11). Mineral dissolution rates enhanced in the presence of citrate calculated from major element chemistry using PHREEQC are: augite, apatite, glass, and magnetite (in basalt) and apatite, hornblende, magnetite and plagioclase (in granite). These minerals can be compared to the trace element hosts significantly enhanced in the presence of citrate (basalt: augite, apatite, Cu-Fe sulfides, and Fe-Ti oxides and granite: apatite, hornblende, sphene, and biotite). The overlap in the minerals showing enhanced dissolution when calculated from major element chemistry by PHREEQC or inferred from the trace element hosts builds confidence in the observation that organic acids enhance mineral dissolution rates either by indirect mechanisms such as changing the complexation of the ions in solution (Oelkers and Schott, 1998; Kubicki and others, 1999), or by directly affecting the mineral surface (Furrer and Stumm, 1983). from dissolving basalt and granite with and without organic ligands 653 Field Scale: Rivers We also compare our results from the column experiments to observations from organic-rich and -poor rivers. Multiple comparisons of element concentrations between organic-rich rivers (or “black” rivers), and organic-poor rivers indicate increased concentrations of elements such as Al, Fe, Th, Zr, Y, and rare earth elements (REE) in the organic-rich waters as compared to the organic-poor waters, but low concentrations of Ca, Na, Sr, K, Ba, Rb, and U (Dai and Martin, 1995; Dupré and others, 1996; Viers and others, 1997; Braun and others, 1998; Oliva and others, 1999; Viers and others, 2000; Millot and others, 2003; Tosiani and others, 2004; Braun and others, 2005; Pokrovsky and others, 2006). The enhanced dissolution and mobility of these elements in the organic-rich rivers supports the observations of these experiments that this element list is more mobile in the presence of citrate, and may therefore be useful as biosignatures (table 7). Viers and others (1997) observe that in the comparison between organic-poor and organic-rich waters in a tropical watershed in Cameroon, the organic-poor waters contain low concentrations of major and trace elements [as well as Dissolved Organic Carbon (DOC)], and that the elements are in a true dissolved form except for Al and REE. In contrast, in the organic-rich waters, concentrations of elements usually considered immobile (Al, Ga, Fe, Ti, Th, Zr, Y and the REE) are increased. In addition, when these organic-rich waters were successively filtered, thereby removing organic carbon colloids, the relationship between the elements Al, Ga, Fe, Ti, Th, Zr, Y and the REE, and DOC form a straight line, suggesting that the concentrations are controlled by the organic colloidal fraction. When we compare the elements enhanced with DOC in the rivers to the elements enhanced in the presence of the granite columns containing citrate (Y, La, Ce, Pb, Th, Al, P, Ti, V, Fe, Ni and Zn), we notice that the enhanced elements are very similar. This builds confidence that these elements may be useful as biosignatures. Similar results are observed in a more spatially detailed study of the Nyong basin rivers in Cameroon (Viers and others, 2000) in which Al, Fe, Y, Zr, Th and the REE were found to exhibit a strong correlation with dissolved organic carbon. Similarly, Dupré and others (1996) observe in the Congo Basin that both organic-rich and organic-poor rivers contain very low concentrations of very soluble elements such as Ca, Na, Sr, K, Ba, Rb, and U, and very high concentrations for trace elements and REE, particularly in the organic-rich rivers. These authors again attribute the high concentrations of these trace and REE to transport in a colloidal phase. In a study of groundwater in the tropical watershed Nsimi-Zoetele in Cameroon, Oliva and others (1999) have found that organic acids enhance dissolution of minerals and transport of the generally insoluble elements Al, Fe, Ti, Zr and REE. In further studies of the same watershed, Braun and others (2005) have similarly found that degradation of swamp organic matter produces organic colloids, which mobilize Fe, Al, Zr, Ti, and Th from the top meter of the swamp. When the organic colloid content is low, Th and Zr are close to detection, controlled by the solubility of the secondary mineral thorianite (ThO2) and the primary mineral zircon (ZrSiO4). Tosiani and others (2004) also observe that enhanced transport of Al and Fe correlates with DOC, and attribute the mobility of these elements (as well as La) to transport by organic colloids in the Cuyuni basin in Southern Venezuela. In a very different climate regime in the Mackenzie River Basin in Canada, Millot and others (2003) have observed that the presence of organic matter in the northern watersheds increases weathering by a factor of approximately 3 to 4 between the low and high DOC rivers. Trace elements were not measured, so it is not possible to compare the effect of DOC on trace elements in these environments. In two arctic rivers, Dai and Martin (1995) have also concluded that the low metal concentrations 654 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates Fig. 6. Basalt column discharge concentrations and rivers draining a similar lithology (average weeks 40-45) normalized to BCR-1. Column discharges are indicated by heavy black lines (solid ⫽ citratecontaining, dotted⫽ligand-free), and the rivers by the gray solid lines. River concentrations are from Tosiani and others (2004), and Pokrovsky and others (2006). The higher the concentration of organic carbon in the rivers, the thicker the gray line. In most cases, the normalized citrate-containing and ligand-free column discharges bracket the normalized concentrations of the rivers. Where elements are not plotted, concentrations were unavailable for the rivers (for example, P), or below detection (ligand-free column). Elements are organized after the periodic table, with first s block elements, from smallest to largest atomic radius, then d block elements, then p block and f block elements. This allows the comparison of elements with similar properties. found in these rivers are largely controlled by the colloidal fraction, although they have not determined whether the colloids are organic, inorganic, or both. These results span a wide range of climatic conditions. The similarity of the elements that appear to be enhanced by the presence of organics in natural environments (Al, Fe, Th, Zr, Y and REE) to those enhanced in the column experiments (Y, La, Ce, Th, Al, P, Ti, Fe, Pb and Ni in both granite and basalt-containing columns, Zr, Sc, and Mn in basalt-containing columns, and V and Zn in granite-containing columns) is additional evidence that these metals may function as important biosignatures. We have compared the average elemental concentrations for weeks 40 to 45 of the basalt and granite-containing columns with and without citrate to river concentrations draining basalt and granite lithologies (figs. 6 and 7). To make the comparison we normalized the river concentrations either to the basalt BCR-1 or the granite from the Tuolomne River Series. In most cases, the citrate-containing column and the ligandfree column bracket the normalized river values (figs. 6 and 7). The heavier lines represent higher concentrations of dissolved organic carbon, and in some cases it from dissolving basalt and granite with and without organic ligands 655 Fig. 7. Granite column discharge (average of weeks 40-45) and rivers draining a similar lithology concentrations normalized to Tuolomne river granite. Column discharges are indicated by heavy black lines (solid ⫽ citrate-containing, dotted⫽ligand-free), and the rivers by the gray solid lines. River concentrations are from Tosiani and others (2004), Viers and others (1997), and Viers and others (2000). The higher the concentration of organic carbon in the rivers, the thicker the gray line. In most cases, the normalized citrate-containing and ligand-free column discharges bracket the normalized concentrations of the rivers. Where elements are not plotted, concentrations were unavailable for the rivers (for example, P), or below detection (ligand-free column). Elements are organized after the periodic table, with first s block elements, from smallest to largest atomic radius, then d block elements, then p block and f block elements. This allows the comparison of elements with similar properties. appears that the increase in elemental concentration corresponds to an increase in dissolved organic carbon (for example, elements such as Zr) and in other cases it appears that other factors such as climate, vegetation, or relief might be playing a stronger role. If the organic content of the river is controlling the elemental concentrations in the river water, and the stability complex with citrate were similar to the stability complex of the natural organic matter, we would expect increasing concentrations with increasing strength of the stability constant as was observed for the different elements in the basalt and granite-containing columns (fig. 8). To normalize the river concentrations for effects that might change all elemental concentrations (for example, water-rock ratios), we divided each normalized elemental concentration by the concentration of another element (Ca). The concentrations from the citratecontaining columns increase with increasing stability constant (fig. 8), while the ligand-free columns display no apparent trend. However, when we plot the river data in the same manner, we observe that the river concentrations, instead of increasing with increasing stability constant as is observed in the citrate-containing columns, instead display a decreasing trend more typical of the ligand-free columns. 656 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates Fig. 8. Elemental concentrations normalized by Ca for (A) basalt citrate and ligand free columns (B) granite citrate and ligand-free columns, (C) an example organic-rich river draining basaltic lithology and (D) an example organic-rich river draining granitic lithology. Normalized elemental concentrations increase in the citrate-containing columns with increasing NIST stability constant, whereas normalized concentrations in the citrate-free columns and rivers show no trend or decrease with increasing NIST stability constant. Concentrations were first normalized to elemental concentrations in BCR-1 (if basalt were present), or granite Tuolumne series (if granite were present). This decreasing trend could be explained since the organic matter in the rivers is much more dilute than the organic acids used in our experiments. Alternately, it is possible that the soils may already have been depleted by organic-rich weathering solutions. If the soils drained by these rivers have been leached by organic-rich solutions, the depletions in the soils and sediments might prove biosignatures. This has previously been tested in paleosols (Neaman and others, 2005a), and it would be interesting to test it in modern soils as well. Field Scale: Implications for Element Normalization Elements often used as an assumed immobile element for mass balance calculations include Zr, Ti, Fe, and Al. In the absence of citrate, Ti is the least mobile element from both basalt and granite. In the presence of citrate, Ti is still relatively immobile (⬍0.14% removed from basalt and ⬍0.35% removed from granite), and is still a good choice as an immobile element for basalt weathered in the presence of citrate. However, Zr (0.065% removed) and Al (0.28%) are both less mobile from granite in the presence of citrate than Ti. This suggests that Zr and Al might both be better choices as an immobile element for mass balance calculations in weathering granite. from dissolving basalt and granite with and without organic ligands 657 conclusions Citrate clearly enhances the dissolution of both basalt and granite, and the chemical signatures of this enhancement may prove a useful biosignature of biological activity. Elemental releases indicate a strong correlation with the stability constant of the element with citrate, and this suggests that values of association constants may be used to infer the presence of other organic acids. Elements which showed a strong enrichment in the presence of citrate, and might therefore be useful biosignatures include: Zr, Sc and Mn (basalt), V and Zn (granite), and Y, La, Ce, Th, Ti, Al, P, Pb, Ni and Fe (both basalt and granite). Release of these elements is consistent with enhanced dissolution of apatite ⫹ Fe sulfides ⫹ Fe oxides ⫹ augite in basalt and apatite ⫹ sphene ⫹ hornblende in granite suggesting mineralogical as well as elemental biosignatures. Compilation of elemental concentrations in river waters indicates that enhanced concentrations of Al, Fe, Ti, Th, Zr, Y and the REE correlate with higher DOC in natural waters: these elements were also enhanced by the presence of citrate in our column experiments. Differences in mineral reactivity between ligand-free and ligand-containing solutions suggest that the presence of organic acids in paleosols from early Earth may indeed be documented in soil minerals and element compositions. acknowledgments We thank B. Alexander, J. Kittleson, H. Gong, M. Angelone, J. Catalina, T. Rusnak, D. Voigt, L. Liermann, A. Zimmerman, R. Conrey, Maya Bhatt, and D. Eggler for their assistance. This work was supported by funding from the NASA grant NAG5-12330, National Science Foundation Integrative Graduate Education and Research Traineeship grant DGE-9972759, the Penn State Biogeochemical Research Initiative for Education and the NASA Astrobiology Institute, Grant # NNA04CC06A. E.M.H. is grateful for support from the National Science Foundation Graduate Research Fellowship Program. S.L.B. acknowledges support from the NSF-funded Center for Environmental Kinetics Analysis NSF-CHE-0431328. We appreciate the thoughtful reviews by K. Maher and M. Velbel. References Anderson, S. P., Dietrich, W. E., and Brimhall, G. H., Jr., 2002, Weathering profiles, mass balance analysis, and rates of solute loss: Linkages between weathering and erosion in a small, steep catchment: Geological Society of America Bulletin, v. 114, n. 9, p. 1143–1158, doi: 10.1130/00167606(2002)114具1143:WPMBAA典2.0.CO;2. Bandstra, J. Z., and Brantley, S. L., 2008, Data Fitting Techniques with Applications to mineral dissolution kinetics, in Brantley, S. L., Kubicki, J. D., and White, A. F., editors, Kinetics of Water-Rock Interaction: New York, Springer, p. 211–257, doi:10.1007/978-0-387-73563-4_6. Banfield, J. F., Jones, B. F., and Veblen, D. R., 1991, An AEM-TEM study of weathering and diagenesis, Abert Lake, Oregon: I, Weathering reactions in the volcanics: Geochimica et Cosmochimica Acta, v. 55, n. 10, p. 2781–2793, doi:10.1016/0016-7037(91)90444-A. Banfield, J. F., Moreau, J. W., Chan, C. S., Welch, S. A., and Little, B., 2001, Mineralogical biosignatures and the search for life on Mars: Astrobiology, v. 1, n. 4, p. 447– 465, doi:10.1089/153110701753593856. Barrett, E. P., Joyner, L. G., and Halenda, P. P., 1951, The determination of pore volume and area distributions in porous substances. I. Computations from nitrogen isotherms: Journal of the American Chemical Society, v. 73, p. 373–380, doi:10.1021/ja01145a126. Bateman, P. C., and Chappell, B. W., 1979, Crystallization, fractionation, and solidification of the Tuolumne Intrusive Series, Yosemite National Park, California: Geological Society of America Bulletin, v. 90, p. 465– 482, doi:10.1130/0016-7606(1979)90具465:CFASOT典2.0.CO;2. Baziramakenga, R., Simard, R. R., and Leroux, G. D., 1995, Determination of organic acids in soil extracts by ion chromatography: Soil Biology and Biochemistry, v. 27, p. 349 –356, doi:10.1016/00380717(94)00178-4. Bennett, P. C., 1991, Quartz dissolution in organic-rich aqueous systems: Geochimica et Cosmochimica Acta, v. 55, p. 1781–1797, doi:10.1016/0016-7037(91)90023-X. Bennett, P. C., Melcer, M. E., Siegel, D. I., and Hassett, J. P., 1988, The dissolution of quartz in dilute aqueous solutions of organic acids at 25°C: Geochimica et Cosmochimica Acta, v. 52, p. 1521–1530, doi:10.1016/ 0016-7037(88)90222-0. Blake, R. E., and Walter, L. M., 1999, Kinetics of feldspar and quartz dissolution at 70 – 80°C and near-neutral 658 E.M. Hausrath, A. Neaman, and S.L. Brantley—Elemental release rates pH: effects of organic acids and NaCl: Geochimica et Cosmochimica Acta, v. 63, p. 2043–2059, doi:10.1016/S0016-7037(99)00072-1. Brantley, S. L., and Conrad, C. F., 2008, Analysis of Rates of Geochemical Reactions, in Brantley, S. L., Kubicki, J. D., and White, A. F., editors, Kinetics of Water-Rock Interaction: New York, Springer, p. 1–37, doi:10.1007/978-0-387-73563-4_1. Braun, J.-J., Viers, J., Dupre, B., Polve, M., Ndam, J., and Muller, J.-P., 1998, Solid/liquid REE fractionation in the lateritic system of Goyoum, East Cameroon: the implication for the present dynamics of the soil covers of the humid tropical regions: Geochimica et Cosmochimica Acta, v. 62, p. 273–300, doi:10.1016/ S0016-7037(97)00344-X. Braun, J.-J., Ngoupayou, J. R. N., Viers, J., Dupre, B., Bedimo Bedimo, J.-P., Boeglin, J.-L., Robain, H., Nyeck, B., Freydier, R., Nkamdjou, L. S., Rouiller, J., and Muller, J.-P., 2005, Present weathering rates in a humid tropical watershed: Nsimi, South Cameroon: Geochimica et Cosmochimica Acta, v. 69, p. 357– 387, doi:10.1016/j.gca.2004.06.022. Brunauer, S., Emmett, P. H., and Teller, E., 1938, Adsorption of gases in multimolecular layers: Journal of the American Chemical Society, v. 60, n. 2, p. 309 –319, doi:10.1021/ja01269a023. Buss, H. L., ms, 2006, Biogeochemical Weathering of Iron-Silicate Minerals: University Park, Pennsylvania, Penn State University, Ph. D., thesis. Buss, H. L., Brantley, S. L., and Liermann, L. J., 2003, Nondestructive methods for removal of bacteria from silicate surfaces: Geomicrobiology Journal, v. 20, n. 1, p. 25– 42, doi:10.1080/01490450303888. Casey, W. H., and Westrich, H. R., 1992, Control of dissolution rates of orthosilicate minerals by divalent metal-oxygen bonds: Nature, v. 355, p. 157–159, doi:10.1038/355157a0. Casey, W. H., Westrich, H. R., and Holdren, G. R., 1991, Dissolution rates of plagioclase at pH ⫽ 2 and 3: American Mineralogist, v. 76, p. 211–217. Dai, M.-H., and Martin, J.-M., 1995, First data on trace metal level and behaviour in two major Arctic river-estuarine systems (Ob and Yenisey) and in the adjacent Kara Sea, Russia: Earth and Planetary Science Letters, v. 131, p. 127–141, doi:10.1016/0012-821X(95)00021-4. Dodge, F. C. W., Millard, H. T., and Elsheimer, H. N., 1982, Compositional variations and abundances of selected elements in granitoid rocks and constituent minerals, central Sierra Nevada Batholith, California: U. S. Geological Survey Professional Paper, v. 1248, 24 p. Drever, J. I., and Stillings, L. L., 1997, The role of organic acids in mineral weathering: Colloids and surfaces A: Physicochemical and Engineering Aspects, v. 120, p. 167–181, doi:10.1016/S0927-7757(96)03720-X. Dupré, B., Gaillardet, J., Rousseau, D., and Allegre, C. J., 1996, Major and trace elements of river-borne material: The Congo Basin: Geochimica et Cosmochimica Acta, v. 60, p. 1301–1321, doi:10.1016/00167037(96)00043-9. Eick, M. J., Grossl, P. R., Golden, D. C., Sparks, D. L., and Ming, D. W., 1996a, Dissolution kinetics of a lunar glass simulant at 25 °C: The effect of pH and organic acids: Geochimica et Cosmochimica Acta, v. 60, p. 157–170, doi:10.1016/0016-7037(95)00377-0. –––––– 1996b, Dissolution of a lunar basalt simulant as affected by pH and organic anions: Geoderma, v. 74, p. 139 –160, doi:10.1016/S0016-7061(96)00055-9. Fisk, M. R., Popa, R., Mason, O. U., Storrie-Lombardi, M. C., and Vicenzi, E. P., 2006, Iron-Magnesium silicate bioweathering on Earth (and Mars?): Astrobiology, v. 6, p. 48 – 68, doi:10.1089/ast.2006.6.48. Flanagan, F. J., 1967, U.S. Geological Survey silicate rock standards: Geochimica et Cosmochimica Acta, v. 31, p. 289 –308, doi:10.1016/0016-7037(67)90043-9. Furrer, G., and Stumm, W., 1983, The role of surface coordination in the dissolution of ␦–Al2O3 in dilute acids: Chimie, v. 37, p. 338 –341. Gislason, S. R., and Eugster, H. P., 1987, Meteoric water-basalt interactions. I: A laboratory study: Geochimica et Cosmochimica Acta, v. 51, p. 2827–2840, doi:10.1016/0016-7037(87)90161-X. Gislason, S. R., and Oelkers, E. H., 2003, Mechanism, rates and consequences of basaltic glass dissolution: II. An experimental study of the dissolution rates of basaltic glass as a function of pH and temperature: Geochimica et Cosmochimica Acta, v. 67, p. 3817–3832, doi:10.1016/S0016-7037(03)00176-5. Govindaraju, K., 1994, 1994 compilation of working values and sample description for 383 geostandards: Geostandards Newsletter, v. 18, p. 1–158 p., doi: 10.1111/j.1751-908X.1994.tb00526.x. Goyne, K. W., Brantley, S. L., and Chorover, J., 2006, Effects of organic acids and dissolved oxygen on apatite and chalcopyrite dissolution: Implications for using elements as organomarkers and oxymarkers: Chemical Geology, v. 234, p. 28 – 45, doi:10.1016/j.chemgeo.2006.04.003. Halsey, G., 1948, Physical adsorption on non-uniform surfaces: The Journal of Chemical Physics, v. 16, p. 931, doi:10.1063/1.1746689. Hart, S. R., and Dunn, T., 1993, Experimental cpx/melt partitioning of 24 trace elements: Contributions to Mineralogy and Petrology, v. 113, p. 8, doi:10.1007/BF00320827. Hausrath, E. M., Liermann, L. J., House, C. H., Ferry, J. G., and Brantley, S. L., 2007, The effect of methanogen growth on mineral substrates: will Ni markers of methanogen-based communities be detectable in the rock record?: Geobiology, v. 5, p. 49 – 61, doi:10.1111/j.1472-4669.2007.00095.x. Hausrath, E. M., Treiman, A. H., Vicenzi, E., Bish, D. L., Blake, D., Sarrazin, P., Hoehler, T., Midtkandl, I., Steele, A., and Brantley, S. L., 2008, Short- and long-term olivine weathering in Svalbard: Implications for Mars: Astrobiology, v. 8, n. 6, p. 1079 –1092, doi:10.1089/ast.2007.0195. Ho, A. M., and Cashman, K. V., 1997, Temperature constraints on the Ginkgo flow of the Columbia River Basalt Group: Geology, v. 25, p. 403– 406, doi:10.1130/0091-7613(1997)025具0403:TCOTGF典2.3.CO;2. Hoch, A. R., Reddy, M. M., and Drever, J. I., 1996, The effect of iron content and dissolved O2 on dissolution rates of clinopyroxene at pH 5.8 and 25°C: preliminary results: Chemical Geology, v. 132, p. 151–156, doi:10.1016/S0009-2541(96)00050-2. Holdren, G. R., Jr., and Speyer, P. M., 1987, Reaction rate-surface area relationships during the early stages of from dissolving basalt and granite with and without organic ligands 659 weathering. II. Data on eight additional feldspars: Geochimica Cosmochimica Acta, v. 51, p. 2311–2318, doi:10.1016/0016-7037(87)90284-5. Huang, W. H., and Kiang, W. C., 1972, Laboratory dissolution of plagioclase feldspars in water and organic acids at room temperature: American Mineralogist, v. 57, p. 1849 –1859. Jones, D. L., and Edwards, A. C., 1998, Influence of sorption on the biological utilization of two simple carbon substrates: Soil Biology and Biochemistry, v. 30, p. 1895–1902, doi:10.1016/S00380717(98)00060-1. Kalinowski, B. E., Liermann, L. J., Brantley, S. L., Barnes, A., and Pantano, C. G., 2000, X-ray photoelectron evidence for bacteria-enhanced dissolution of hornblende: Geochimica et Cosmochimica Acta, v. 64, p. 1331–1343, doi:10.1016/S0016-7037(99)00371-3. Kpomblekou-a, K., and Tabatabai, M. A., 1994, Effect of organic acids on release of phosphorus from phosphate rocks: Soil Science, v. 158, p. 442– 453, doi:10.1097/00010694-199415860-00006. Krzyszowska, A. J., Blaylock, M. J., Vance, G. F., and David, M. B., 1996, Ion chromatographic analyses of low molecular weight organic acids in Spodosol forest floor solutions: Soil Science Society of America Journal, v. 60, p. 1565–1571. Kubicki, J. D., Schroeter, L. M., Itoh, M. J., Nguyen, B. N., and Apitz, S. E., 1999, Attenuated total reflectance Fourier-transform infrared spectroscopy of carboxylic acids adsorbed onto mineral surfaces: Geochimica Cosmochimica Acta, v. 63, p. 2709 –2725, doi:10.1016/S0016-7037(99)00194-5. Kuroda, P. K., and Sandell, E. B., 1954, Geochemistry of molybdenum: Geochimica et Cosmochimica Acta, v. 6, p. 35– 63, doi:10.1016/0016-7037(54)90028-9. Lundstrom, U., and Ohman, L.-O., 1990, Dissolution of feldspars in the presence of natural organic solutes: Journal of Soil Science, v. 41, p. 359 –369. Manley, E. P., and Evans, L. J., 1986, Dissolution of feldspars by low-molecular weight aliphatic and aromatic acids: Soil Science, v. 141, p. 106 –112, doi:10.1097/00010694-198602000-00002. Martell, A. E., and Smith, R. M., 2001, NIST Critically Selected Stability Constants of Metal Complexes Database: NIST Standard Reference Database 46, Version 6.0: Gaithersburg, Maryland, U.S. Department of Commerce, NIST. Millot, R., Gaillardet, J., Dupre, B., and Allegre, C. J., 2003, Northern latitude chemical weathering rates: Clues from the Mackenzie River Basin, Canada: Geochimica Cosmochimica Acta, v. 67, p. 1305–1329, doi:10.1016/S0016-7037(02)01207-3. Neaman, A., Chorover, J., and Brantley, S. L., 2004, The effect of organic ligands on basalt and granite weathering, in Wanty, R., and Seal, R., editors, Water-Rock Interaction 11, Saratoga Springs, New York, Proceedings of the 11th International symposium on Water-Rock Interaction: London, Taylor and Francis Group, v. 1, p. 1347–1350. –––––– 2005a, Element mobility patterns record organic ligands in soils on early Earth: Geology, v. 33, p. 117–120, doi:10.1130/G20687.1. –––––– 2005b, Implications of the evolution of organic acid moieties for basalt weathering over geological time: American Journal of Science, v. 305, p. 147–185, doi:10.2475/ajs.305.2.147. –––––– 2006, Effects of organic ligands on granite dissolution in batch experiments at pH 6: American Journal of Science, v. 306, p. 451– 473, doi:10.2475/06.2006.03. Oelkers, E. H., and Gislason, S. R., 2001, The mechanism, rates and consequences of basaltic glass dissolution: I. An experimental study of the dissolution rates of basaltic glass as a function of aqueous Al, Si, and oxalic acid concentration at 25°C and pH ⫽ 3 and 11: Geochimica et Cosmochimica Acta, v. 65, p. 3671–3681, doi:10.1016/S0016-7037(01)00664-0. Oelkers, E. H., and Schott, J., 1998, Does organic acid adsorption affect alkali-feldspar dissolution rates?: Chemical Geology, v. 151, p. 235–245, doi:10.1016/S0009-2541(98)00082-5. Oliva, P., Viers, J., Dupre, B., Fortune, J. P., Martin, F., Braun, J. J., Nahon, D., and Robain, H., 1999, The effect of organic matter on chemical weathering: study of a small tropical watershed: nsimi-zoetele site, Cameroon: Geochimica et Cosmochimica Acta, v. 63, p. 4013– 4035, doi:10.1016/S0016-7037(99)00306-3. Parkhurst, D. L., and Appelo, C. A. J., 1999, User’s guide to PHREEQC (version 2)—A computer program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations.: U.S. Geological Survey Water-Resources Investigations Report 99-4259, 312 p. Pokrovsky, O. S., Schott, J., and Dupre, B., 2006, Trace element fractionation and transport in boreal rivers and soil porewaters of permafrost-dominated basaltic terrain in Central Siberia: Geochimica et Cosmochimica Acta, v. 70, p. 3239 –3260, doi:10.1016/j.gca.2006.04.008. Prowatke, S., and Klemme, S., 2006, Trace element partitioning between apatite and silicate melts: Geochimica et Cosmochimica Acta, v. 70, p. 4513– 4527, doi:10.1016/j.gca.2006.06.162. Schott, J., and Berner, R. A., 1985, Dissolution mechanisms of pyroxenes and olivines during weathering, in Drever, J. I., editor, The Chemistry of Weathering: NATO ASI series. Series C. Mathematical and Physical Sciences: Boston, D. Reidel, p. 35–53. Shen, Y., Ström, L., Jönsson, J.-A., and Tyler, G., 1996, Low molecular organic acids in the rhizosphere soil solution of beech forest (Fagus sylvatica L.) Cambisols determined by ion chromatography using supported liquid membrane enrichment technique: Soil Biology and Biochemistry, v. 28, p. 1163–1169, doi:10.1016/0038-0717(96)00119-8. Siegel, D. I., and Pfannkuch, H. O., 1984, Silicate mineral dissolution at pH 4 and near standard temperature and pressure: Geochimica et Cosmochimica Acta, v. 48, p. 197–201, doi:10.1016/0016-7037(84)90362-4. Sigg, L., and Stumm, W., 1981, The interaction of anions and weak acids with the hydrous goethite (␣-FeOOH) surface: Colloids Surfaces, v. 2, p. 101–117, doi:10.1016/0166-6622(81)80001-7. Stevenson, F. J., 1991, Organic matter-micronutrient reactions in soils, in Mortvedt, J. J., editor, Micronutrients in Agriculture: Soil Science Society of America Book Series, N. 4, p. 145–186. Stillings, L. L., and Brantley, S. L., 1995, Feldspar dissolution at 25°C and pH 3: Reaction stoichiometry and 660 E.M. Hausrath, A. Neaman, and S.L. Brantley the effect of cations: Geochimica et Cosmochimica Acta, v. 59, p. 1483–1496, doi:10.1016/00167037(95)00057-7. Stillings, L. L., Drever, J. I., Brantley, S. L., Sun, Y., and Oxburgh, R., 1996, Rates of feldspar dissolution at pH 3–7 with 0 – 8 M oxalic acid: Chemical Geology, v. 132, p. 79 – 89, doi:10.1016/S0009-2541(96)00043-5. Sverdrup, H. U., 1990, The Kinetics of Base Cation Release Due to Chemical Weathering: Lund, Sweden, Lund University Press, 245 p. Tosiani, T., Loubet, M., Viers, J., Valladon, M., Tapia, J., Marrero, S., Yanes, C., Ramirez, A., and Dupre, B., 2004, Major and trace elements in river-borne materials from the Cuyuni basin (southern Venezuela): evidence for organo-colloidal control on the dissolved load and element redistribution between the suspended and dissolved load: Chemical Geology, v. 211, p. 305–334, doi:10.1016/j.chemgeo.2004.07.001. van Hees, P. A. W., Lundstrom, U. S., and Morth, C.-M., 2002, Dissolution of microcline and labradorite in a forest O horizon extract: the effect of naturally occurring organic acids: Chemical Geology, v. 189, p. 199 –211, doi:10.1016/S0009-2541(02)00141-9. Viers, J., Dupré, B., Polve, M., Schott, J., Dandurand, J.-L., and Braun, J.-J., 1997, Chemical weathering in the drainage basin of a tropical watershed (Nsimi-Zoetele site, Cameroon): comparison between organicpoor and organic-rich waters: Chemical Geology, v. 140, p. 181–206, doi:10.1016/S00092541(97)00048-X. Viers, J., Dupré, B., Braun, J.-J., Deberdt, S., Angeletti, B., Ngoupayou, J. N., and Michard, A., 2000, Major and trace element abundances, and strontium isotopes in the Nyong basin rivers (Cameroon): constraints on chemical weathering processes and elements transport mechanisms in humid tropical environments: Chemical Geology, v. 169, p. 211–241, doi:10.1016/S0009-2541(00)00298-9. Wang, X., Li, Q., Hu, H., Zhang, T., and Zhou, Y., 2005, Dissolution of kaolinite induced by citric, oxalic, and malic acids: Journal of Colloid and Interface Science, v. 290, p. 481– 488, doi:10.1016/j.jcis.2005.04.066. Welch, S. A., and Ullman, W. J., 1993, The effect of organic acids on plagioclase dissolution rates and stoichiometry: Geochimica et Cosmochimica Acta, v. 57, p. 2725–2736, doi:10.1016/00167037(93)90386-B. Welch, S. A., Taunton, A. E., and Banfield, J. F., 2002, Effect of Microorganisms and Microbial Metabolites on Apatite Dissolution: Geomicrobiology Journal, v. 19, p. 343–367, doi:10.1080/01490450290098414. White, A. F., and Brantley, S. L., 1995, Chemical Weathering Rates of Silicate Minerals: Washington, D. C., Mineralogical Society of American, Reviews in Mineralogy, v. 31, 583 p. –––––– 2003, The effect of time on the weathering of silicate minerals: why do weathering rates differ in the laboratory and field?: Chemical Geology, v. 202, p. 479 –506, doi:10.1016/j.chemgeo.2003.03.001. White, A. F., Peterson, M. L., and Hochella, M. F., Jr., 1994, Electrochemistry and dissolution kinetics of magnetite and ilmenite: Geochimica et Cosmochimica Acta, v. 58, p. 1859 –1875, doi:10.1016/00167037(94)90420-0. White, A. F., Bullen, T. D., Vivit, D. V., Schulz, M. S., and Clow, D. W., 1999, The role of disseminated calcite in the chemical weathering of granitoid rocks: Geochimica et Cosmochimica Acta, v. 63, p. 1939 –1953, doi:10.1016/S0016-7037(99)00082-4. Wolff-Boenisch, D., Gislason, S. R., Oelkers, E. H., and Putnis, C. V., 2004, The dissolution rates of natural glasses as a function of their composition at pH 4 and 10.6, and temperatures from 25 to 74°C: Geochimica et Cosmochimica Acta, v. 68, p. 4843– 4858, doi:10.1016/j.gca.2004.05.027. Zhang, H., ms, 1990, Factors determining the rate and stoichiometry of hornblende dissolution: University of Minnesota, Ph. D. thesis. Zhang, H., and Bloom, P. R., 1999, Dissolution Kinetics of Hornblende in Organic Acid Solutions: Soil Science Society of America Journal, v. 63, p. 815– 822. Zhang, H., Bloom, P. R., Nater, E. A., and Erich, M. S., 1996, Rates and stoichiometry of hornblende dissolution over 115 days of laboratory weathering at pH 3.6 – 4.0 and 25°C in 0.01 M lithium acetate: Geochimica et Cosmochimica Acta, v. 60, p. 941–950, doi:10.1016/0016-7037(95)00447-5.
© Copyright 2026 Paperzz