Boron isotope fractionation during brucite deposition from artificial seawater

Boron isotope fractionation during brucite deposition from artificial seawater J. Xiao, Y. K. Xiao, C. Q. Liu, and Z. D. Jin State Key Laboratory of Loess and Quaternary Geology, Institute of Earth Environment, Chinese Academy of Sciences, Xi’an, 710075, China Qinghai Institute of Salk Lakes, Chinese Academy of Sciences, Xining, Qinghai, 810008, China State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang, 550002, China


Introduction
Mg is a common trace element in corals and the Mg/Ca ratio has been used to investigate the paleothermometry of seawater (Watanabe et al., 2001;Mitsuguchi et al., 2008;Wei et al., 2009;Allison et al., 2010).The incorporation of Mg into coral skeleton is controlled by varying factors.Mg can exist in the form of brucite in corals (Smith and Delong, 1978;Nothdurft et al., 2005) and Mg content in scleractinian corals varies between different genera and localities (Fallon et al., 1999).Nothdurft et al. (2005) studied brucite in living scleractinian corals colonies (Acropora, Pocillopora, Porites) from subtidal and intertidal settings in the Great Barrier Reef, Australia, and subtidal Montastraea from the Florida Keys, United States.Their results suggested that high containing activity combined with high pH and low pCO 2 led to localized brucite precipitation in organic biofilms of coral skeletons and that brucite in living coral skeletons is an indicator of extreme microenvironments in shallow-marine settings.The occurrence of brucite in corals can bring a higher Mg/Ca ratio and be responsible for reported anomalies in Mg/Ca vs. sea-surface temperature (SST) plots in corals (Nothdurft et al., 2005).Defining the source of such anomalies is critical for using Mg/Ca in corals to reconstruct the paleothermometry of seawater (Mg/Ca-SST proxy).
In recent years, topics such as the reconstruction of ancient seawater pH using the isotopic composition of boron in corals, the calculation of the past pCO 2 , and the influence of these two factors on changes in the ancient climate, have become important issues for the international boron isotope geochemistry community (δ 11 B-pH proxy) (Hemming and Hanson, 1992;Gaillardet and Allègre, 1995;Pelejero et al., 2005;Liu et al., 2009;Wei et al., 2009;Douville et al., 2010).
Published by Copernicus Publications on behalf of the European Geosciences Union.Whether the δ 11 B of corals is equal to that of B(OH) − 4 has emerged as one of the main questions involved in these issues.A series of inorganic calcite precipitation experiments have shown that B(OH) − 4 is the only or dominant species incorporated into calcite (Hemming et al., 1995;Sanyal et al., 1996Sanyal et al., , 2000)).However, Klochko et al. (2009) recently found that both trigonal and tetrahedral coordinated boron existed in biogenic and hydrothermal carbonates.Rollion- Bard et al. (2011) also found both boron coordination species, but in different proportions depending on the coral microstructure, i.e. centres of calcification versus fibres.They suggested that careful sampling is necessary before performing boron isotopic measurements in deep-sea corals (Rollion-Bard et al., 2011).Inorganic calcite precipitation experiment has been carried out by Xiao et al. (2006a), indicating that the δ 11 B of inorganic calcium carbonate was not parallel with the calculated curve of B(OH) − 4 , but deviated increasingly from the parallel trend as pH increased.When the pH was increased to a certain value, the isotopic fractionation factor of boron between precipitation and solution was greater than 1.Xiao et al. (2006a) reasoned that the presence of Mg 2+ or other microelements was the main reason for this observation and concluded that B(OH) 3 incorporated preferentially into brucite.If this is true, the isotopic compositions of boron in corals can be affected by the existence of brucite in corals and the preferential incorporation of B(OH) 3 into brucite.If the brucite-bearing corals were used in the δ 11 B-pH proxy, the measured δ 11 B of corals and the calculated pH will be higher than the normal value, which influences the δ 11 B-pH proxy negatively.However, little attention has been paid to the mechanism of boron incorporated into brucite and its influence on the boron isotopic composition of corals.
In this study, experiments on the incorporation of boron during the deposition of brucite from magnesium-free artificial seawater at various pH values were carried out.The incorporation species of boron into brucite and the the boron isotope fractionation during deposition of brucite were determined.The influence of brucite in corals on Mg/Ca and the δ 11 B, judgment method for existence of brucite in corals and its potential application in the δ 11 B-pH proxy and the Mg/Ca-SST proxy were discussed.This result will shed some light on the application of the Mg/Ca-SST proxy and the δ 11 B-pH proxy.

Reagents and equipment
1.0 M HCl, 0.5 M NaOH, and 0.5 M MgCl 2 were prepared using boron-free water with twice distilled HCl (GR), NaOH (GR), and MgCl 2 (GR), respectively.The concentration of the Cs 2 CO 3 (99.994% purity) solution used in our study was 12.3 g Cs l −1 .A suspended graphite solution was prepared using spectrum pure (SP) graphite mixed with 80 % ethanol and 20 % water.For a mannitol concentration of 1.82 g l −1 , GR mannitol mixed with boron-free water was used.The boron-free water was obtained by using 18.2 M milliQ of water exchanged with an Amberlite IRA 743 boron specific resin.A GV IsoProbe T-single magnetic sector thermal ionization mass spectrometer was used for boron isotope analysis.A 2100 type and UV-2100 type spectrophotometer from Shanghai Unico Ltd was used for boron concentration analysis.A D/Max-2200 type X-ray diffraction equipment from Japan neo-confucianism company was used for XRD analysis.A TP310 type pH meter from Beijing Time Power Co. LTD was used for pH analysis.
Throughout all the experiments, vessels of Teflon, polyethylene or chert were used for avoiding boron contamination.

Magnesium-free artificial seawater
Magnesium-free artificial seawater was prepared by Mocledon prescription.The chemical composition of the magnesium-free artificial seawater is given in Table 1.In order to get sufficient quantities of boron for precise isotopic analysis in the precipitated Mg(OH) 2 , boron concentration in the artificial seawater was adjusted to 45.9 ppm, approximately 10 times that of normal seawater (∼4.5 ppm).According to the result reported by Ingri et al. (1957), there was no polyborate species present when the boron concentration of aqueous solution was 45.9 ppm.The pH value and boron isotopic composition of the initial artificial seawater was 3.64 and −7.00 ± 0.07 ‰, respectively.Since magnesium-free artificial seawater was used in our experiments, no brucite was deposited while the pH was being adjusted.Brucite was deposited only when MgCl 2 was added after the pH adjustments were completed.

Experimental deposition of brucite
Deposition experiments were carried out in a pH-controlled system at a temperature of 22 ± 0.5 • C in a super-clean laboratory (100 Class).Firstly, about 200 ml of artificial seawater was placed inside a plastic beaker.Solutions of NaOH were slowly added to each beaker to achieve the following range of pH values: 9.5, 10.0, 10.5, 11.0, 11.5, 12.0, 12.5, and 13. 0. 8 ml of MgCl 2 solution was then added dropwise into the artificial seawater with different pH values.During this process, NaOH was added to keep the pH at the predetermined pH values as the brucite was deposited.The pH fluctuations were measured by a pH meter with a precision of 0.02.A magnetic stirrer was used during the experimental process to ensure the reactions were completed.The experiment was terminated after 8 ml of MgCl 2 was added.After mixing the deposit and solution well, the mixture was divided into two parts equally.One was filtrated by pumping immediately to isolate deposit from the solution (defined as a repose time T r = 0 h) and the other was filtrated by pumping after reposing for 20 h (defined as a repose time T r = 20 h).The separated solution was placed in air-tight plastic bottles, and the brucite deposit was washed with boron-free water until Cl − was no longer detected.Then the deposited material was dried at 60 • C and placed in air-tight plastic bottles.

Separation of boron from the samples
In preparation for isotopic measurement, boron was extracted from the artificial seawater and the brucite using a twostep chromatographic technique (Wang et al., 2002).About 0.5 ml of Amberlite IRA 743 boron-specific resin with 80-100 mesh was placed in a polyethylene column with a diameter of 0.2 cm.The length of the resin bed was 1.5 cm.About 200 mg of brucite was first dissolved using a slightly excessive amount of boron-free HCl solution, passed through a column filled with Amberlite IRA 743 resin, and then eluted using approximately 5 ml of 0.1 M HCl at 75 • C.An amount of mannitol approximately equimolar to the boron content was added to the eluate, which was then desiccated to near dryness by partial evaporation in a super clean oven with laminar flow at 60 • C (Xiao et al., 2003).The solution was then loaded again into the column filled with a mixed resin composed of 0.5 ml cation-exchange resin (H + form) and 0.5 ml anion-exchange resin (ion-exchange II, HCO − 3 form).The boron was eluted by 5 ml of boron-free water.The eluate was evaporated again at 60 • C to produce a solution with a boron concentration of about 1 µg B µl −1 for isotopic measurement.The procedures for extracting boron in the liquid samples were the same as those described above.The total boron blank in chemical process and filament loading was determined by isotope dilution mass spectrometry (IDMS) to contain 46 ng of boron.

Measurements of boron concentration and isotopic composition
The boron concentrations were determined by the azomethine-H spectrophotometric method.One ml of sample solution, 2 ml of buffer solution, and 2 ml of azomethine-H solution were added in that order.After thorough mixing, each solution was allowed to stand for 120 min.The absorption of the boron-azomethine-H complex was measured at 420 nm using a spectrophotometer.The external precision of the azomethine-H method used here was 2 %.
The isotopic compositions of boron in all samples were measured by a GV IsoProbe T single magnetic sector thermal ionization mass spectrometer and the P-TIMS method using Cs 2 BO + 2 ions with a graphite loading (Xiao et al., 1988).Two µl of graphite slurry was first loaded onto a degassed tantalum filament, and then approximately 1 µl of sample solution containing about 1-4 µg of boron and an equimolar amount of mannitol were also loaded.This was followed by 1-2 µl of a Cs 2 CO 3 solution containing an equimolar amount of cesium.The loaded material was dried by heating the filament at 1.2 A for 5 min.
The data were collected by switching the magnetic field between the masses 308 ( 133 Cs 10 2 B 16 O + 2 ) and 309 ( 133 Cs 11 2 B 16 O + 2 ), and the intensity ratios of the ion beams at masses 308 and 309 (R 309/308 ) were calculated.Calibrated with oxygen isotopes, the 11 B/ 10 B ratio was calculated as R 309/308 − 0.00078.The isotopic composition of boron was expressed as an δ 11 B value according to the following formula: Here the standard material is NIST SRM 951, the recommended value of which is 4.04362 ± 0.00137 (Catanzaro et al., 1970).The measured average δ 11 B/ 10 B ratio of NIST SRM 951 was 4.0525 ± 0.0027 (2 σ m , n = 5).

Dissolved boron species and boron isotope in seawater
The dominant aqueous species of boron in seawater are B(OH) 3 and B(OH) − 4 .The relative proportions of these species are a function of pH (Fig. 1a) and given by the following relation: Figure 1 shows the theoretical distribution of boron species as a function of pH for seawater.At low pH (pH < 7), virtually all of the boron in seawater is in the B(OH) 3 species; conversely, at high pH (pH > 10), virtually all of the boron  (Dickson, 1990) and stable boron isotope fractionation (b) between them as a function of pH, using α 4-3 =0.974 (Klochko et al., 2006).At low pH (pH<7), virtually all of the boron in seawater is in the B(OH) 3 species; conversely, at high pH (pH>10), virtually all of the boron is in the B(OH) 4 -species (Fig. 1a).Only the B(OH) 4 − incorporated into the marine bio-carbonates is one of the essential hypotheses for using boron isotopic composition in marine bio-carbonate to reconstruct the paleo-environment (Hemming and Hanson, 1992).
19 Fig. 1.Fraction (a) of the major dissolved boron species B(OH) 3 (green curve) and B(OH) − 4 (red curve) in seawater, using pK b = 8.597 (Dickson, 1990) and stable boron isotope fractionation (b) between them as a function of pH, using α 4−3 = 0.974 (Klochko et al., 2006).At low pH (pH < 7), virtually all of the boron in seawater is in the B(OH) 3 species; conversely, at high pH (pH > 10), virtually all of the boron is in the B(OH) − 4 species (Fig. 1a).Only the B(OH) − 4 incorporated into the marine bio-carbonates is one of the essential hypotheses for using boron isotopic composition in marine bio-carbonate to reconstruct the paleo-environment (Hemming and Hanson, 1992).
is in the B(OH) − 4 species (Fig. 1a).The isotope exchange reaction between these species is given by: 10 B(OH The distribution of species and the isotope exchange reaction are solved simultaneously to give the isotopic composition of the individual boron species vs. pH for seawater (Fig. 1b).The abundances of two stable isotopes 11 B and 10 B at pH 8 make up approximately 80 % and 20 % of the total boron, respectively.Modern seawater has a boron concentration of about 4.5 ppm, and a consistent, worldwide δ 11 B isotopic composition of +39.61 ± 0.04‰ relative to NIST SRM 951 (Foster et al., 2010).The isotopic composition of B(OH) − 4 increases with pH and so does the δ 11 B of carbonates, provided that B(OH) − 4 is preferentially incorporated in the carbonates.As a result, variations of seawater pH in the past should be traceable by variations of δ 11 B in fossil carbonates.This is the basis of the boron paleo-pH proxy (Hemming and Hanson, 1992).

The partition coefficient of boron between Brucite and final seawater
The boron concentrations and the partition coefficients (K d ) between brucite and final seawater are listed in Table 2, and shown in Fig. 2. The results are shown as follows: 1.The variation trend of the results for T r = 0 h and T r = 20 h are basically the same (Fig. 2a), indicating the incorporation of boron into brucite deposit is instantaneous, and the reaction time is not a main limiting factor.This finding is consistent with the previous study (Liu et al., 2004).
2. When the pH is lower than 10, [B] fsw decreases as the pH increases, and reaches the lowest value at pH 10 (Fig. 2a).This indicates that the amount of boron incorporated into deposited brucite increases as the pH increases, and reaches the highest value at pH 10. [B] fsw increases quickly at higher pH values, and no smooth trend is observed, indicating that the amount of incorporated boron decreases quickly.
3. When pH is lower than 10, K d increases as the pH increases (Fig. 2b).The K d reaches its highest values of 487.5 and 494.2 for T r = 0 h and T r = 20 h at pH 10, respectively.When the pH is higher than 10, K d decreases and shows a mild trend after pH 12.
The K d between brucite deposit and final seawater can be shown as: ] fluid , K d3 and K d4 are the partition coefficient of B(OH) 3 and B(OH) − 4 , respectively.f 3 is the ratio of B(OH) 3 to B(OH) − 4 in solution.This equation indicates that K d is not only related with K d3 and K d4 , but also with the fraction of B(OH) 3 in solution.The K d between clay minerals and seawater is 3.60 at T = 5 • C and pH 8.45 (Palmer et al., 1987), but the highest K d in our experiments is 494 at pH 10, which is much higher than 3.6.It indicates  that the K d3 is much greater than K d4 , i.e. the incorporation capacity of B(OH) 3 into brucite is much higher than that of B(OH) − 4 and the K d3 contribution is dominating.The quantity of brucite reaches its highest at pH 10, indicating the reaction equivalence point of brucite deposition occurs approximately at pH 10.The K d also reaches its highest at pH 10.When pH is higher than 10, the f and K d3 decreases while K d4 increases as the increasing pH.Although the increasing concentration of OH − in solution can influence the adsorption of B(OH) − 4 (Keren et al., 1981), the adsorption of B(OH) − 4 continues, and the K d4 still increases perhaps.Because of K d3 is greater than K d4 , the decreasing of K d3 is dominating, so that the K d decreases with the increasing pH.In the study of Xu and Ye (1997), the point of zero charge (PZC) of brucite is close to 11.9, which approximates to the inflexion point (pH 12) in our experiment.When the pH is higher than PZC, anions will not be adsorbed by goethite and gibbsite (Kingston et al., 1972), so that the adsorption ability of B(OH) − 4 by brucite decreases greatly when the pH is higher than the PZC of brucite.Therefore, K d should monotonically decrease with increasing pH as well.
Boron removal experiments by magnesium hydroxide at different pH values have been carried out by Liu et al. (2004) (Sanyal et al., 1996).The variation trend of δ 11 B for T r = 0 h and T r = 20 h are basically same and all the δ 11 B d values are enriched in the heavy isotope relative to the δ 11 B fsw (Fig. 3a).All the α d-fsw are greater than 1 and shows a waved pattern (Fig. 3b), which is entirely different from that of the carbonates deposits (Fig. 3c).
21 boron balance time of the deposited Mg(OH) 2 , the value for T r = 20 h may be more reliable.
The evolution pattern of boron incorporated into the Mg(OH) 2 , and the variations in partition coefficients with the pH, are basically the same as those of oxides and clay minerals.However, the maximum amounts of incorporated boron in brucite or other hydroxides are much higher than those of the oxides and clay minerals (Keren et al., 1981;Keren and Gast, 1983;Goldberg andGlaubig, 1985, 1986;Lemarchand et al., 2005Lemarchand et al., , 2007)), indicating that the incorporation capacity of boron on hydroxides is stronger than that of both oxides and clay minerals.The boron concentration in corals varies from 39 to 117.5 ppm (Vengosh et al., 1991;Gaillardet and Allègre, 1995;Xiao et al., 2006b;Liu et al., 2009).If 4.5 ppm of the seawater boron concentration is used, the K d between coral and seawater varies from 8.7 to 26.1.In our experiments, the boron concentration in the initial artificial seawater is 45.9 ppm, finally the level of boron concentration in brucite depositions can be divided by 10.In this case, the expected boron content would range between 22 and 95 ppm (excluding the aberrant value of 185.9 ppm obtained at pH 9.5), so the K d between brucite and final seawater varies from 4.8 to 19.2.Such a range of boron concentrations and K d is of the same order of magnitude as the common boron contents measured in corals.

Isotopic fractionation of boron between deposited brucite and solution
Isotopic compositions of boron in deposited brucite (δ 11 B d ) and final seawater (δ 11 B fsw ) are listed in Table 3 and shown in Fig. 3.The crucial points illustrated in Fig. 3 are summarized as follows: 1.The variation trend of the results for T r = 0 h and T r = 20 h are basically same (Fig. 3a).This indicates that the isotopic balance between brucite deposit and solution is instantaneous, and that the results of the experiment are exact.A minor difference between T r = 0 h and T r = 20 h appears as the pH increases from 9.5 to 10.0.When T r = 0 h, δ 11 B d and α d−fsw both increase as the pH increases from 9.5 to 10.0.However, when T r = 20 h, δ 11 B d and α d−fsw both decrease as the pH increases from 9.5 to 10.0.This is perhaps related to the fact that the formation of brucite is slower at low pH than at high pH.
2. All the δ 11 B d values are enriched in the heavy isotope relative to the δ 11 B fsw (Fig. 3a).All the α d−fsw are greater than 1 (Fig. 3b), which is entirely different from those of the carbonates deposits (Fig. 3c).As B(OH) − 4 is incorporated preferentially into carbonate deposits, all the δ 11 B values of carbonates are lower than that of the parent seawater.Meanwhile, all the isotopic fractionation factors α carb−sw between carbonates and seawater are lower than 1 and increase with pH.In this study, the α d−fsw between brucite deposit and solution shows a waved pattern (Fig. 3b), suggesting a complex isotopic fractionation process.
The δ 11 B d , δ 11 B isw3 and δ 11 B isw4 of B(OH) 3 and B(OH) − 4 in initial seawater, calculated using the different α 4−3 versus pH values of the parent solutions, are shown in Fig. 4. When T r = 0 h, the δ 11 B d values are equal to or lower than the δ 11 B isw3 calculated using α 4−3 = 0.9600.When T r = 20 h, the δ 11 B d values are equal to or lower than the δ 11 B isw3 calculated using α 4−3 = 0.9397, but all the δ 11 B d values are higher than δ 11 B isw (Fig. 4).The α 4−3 of 0.9600 and 0.9397 is calculated according to the highest δ 11 B of deposited brucite for T r = 0 h and T r = 20 h.If δ 11 B d is equal to δ 11 B isw3 , it implies that only B(OH) 3 is   incorporated into Mg(OH) 2 , which is impossible under normal circumstances.Moreover a situation in which δ 11 B d > δ 11 B isw3 would be even more improbable.Except for pH 9.5, δ 11 B d values are lower than δ 11 B isw3 calculated using α 4−3 = 0.9600 (Fig. 4), so that the true α 4−3 would have to be lower than 0.9600, which is a little lower than the currently expected value between 0.983 (Sanchez-Valle et al., 2005) and 0.952 (Zeebe, 2005).The pH values used in previous carbonate precipitation experiments are lower than 9.0, while that of our experiments are between 9.5 and 13.0.The differences of the experiments conditions may result diverse α 4−3 , but which value is much more reliable or the true α 4−3 is still uncertain.Our results also indicate that δ 11 B d is not parallel with δ 11 B isw3 , but first decreases and then increases with an inflexion at pH 12.All of the data indicate that both B(OH) 3 and B(OH) − 4 are incorporated into the deposited brucite.
The fractions of B(OH) 3 (F d3 ) in the brucite deposit calculated using α 4−3 = 0.9397 and α 4−3 = 0.9600 are shown in Fig. 5.It seems that F d3 tends to decrease with decreasing α 4−3 .Because the adopted α 4−3 is not true, the calculated F d3 is not exact.The true F d3 could be lower than the value indicated in Fig. 5, but the variation trend of F d3 with pH should be certain and much higher than that of the initial artificial seawater (F isw3 ).The δ 11 B d is the concurrent result of δ 11 B d3 and δ 11 B d4 .Here δ 11 B d can be shown as: where f d3 is the fraction of B(OH) 3 in solid.During pH 10-12, both δ 11 B d3 and δ 11 B d4 increases with increasing pH (Fig. 1b).Therefore, the δ 11 B d should be increased with the increasing pH, but the increasing trend is inconspicuous (Fig. 3a).There are two reasons accounting for this.On the one hand, the increasing δ 11 B d3 and δ 11 B d4 can result in the increasing of δ 11 B d .On the other hand, with the increasing pH, the dominant decreasing incorporation fraction of B(OH) 3 and increasing incorporation fraction of B(OH) − 4 will result the deceasing of δ 11 B d .With the common influence of these two aspects, the variation of δ 11 B d with pH is small.When the pH reaches 12 (PZC of brucite), the adsorption quantity of B(OH) − 4 decreases suddenly, while that of B(OH) 3 increases (Fig. 5).So the δ 11 B d increases.

The manner in which α d−fsw changes with pH for
T r = 0 h and 20 h is shown in Fig. 3b.Except for pH 9.5, both patterns are basically the same for each T r , and match the change in F d3 shown in Fig. 5.The α d−fsw values decrease as the pH increases from 9.5 to 12, and reaches the lowest point at pH 12. Then α d−fsw values increase rapidly.The α d−fsw reflects the relative ratio of B(OH) 3 to B(OH) − 4 in brucite.The explanation of δ 11 B d with pH can be also used for the explanation of the boron fractionation factor between brucite and solution.When the pH ranges from 10 to 12, the incorporation fraction of B(OH) 3 decreases continuously (Fig. 5), which is contrary to that of B(OH) − 4 .This lead to the deceases ratio of B(OH) 3 to B(OH) − 4 , so the α d−sw decreases.When pH reaches 12 (PZC of Mg(OH) 2 ), the adsorption quantity of B(OH) − 4 decreases suddenly, and the relative ratio of B(OH) 3 to B(OH) − 4 increases (Fig. 5), so the α d−fsw also increases.Our results indicate that B(OH) 3 is preferentially incorporated into the solid phase during the deposition of brucite.In a precipitated experiment of calcium carbonates from artificial seawater, Xiao et al. (2006a) obtained the result that the boron isotopic fractionation factors between calcium carbonates and seawaters were higher than 1.They reasoned that the presence of Mg 2+ or other microelements was the main reason for this observation and concluded that the heavier B(OH) 3 species was incorporated preferentially into Mg(OH) 2 .The results of our study have further confirmed the conclusion of Xiao et al. (2006a).

Model 1: boron adsorption by minerals
The boron adsorption by minerals can be explained by phenomenological equation (Keren et al., 1981).It is supposed that when B(OH) 3 , B(OH) − 4 and OH − coexist in solution, they can be adsorbed by Mg(OH) 2 , metal oxides and clay minerals, and the affinity coefficients K B3 , K B4 and K OH are 4 and OH − , respectively (Keren et al., 1981).They compete for the same adsorption site on solid.When the pH is less than 7, B(OH) 3 is dominant.Because K B3 to clay mineral is very low, the boron adsorption is relatively minor.Under this pH value, the concentration of B(OH) − 4 and OH − is very low.Although they have a relatively stronger affinity to clay, their contribution to the total boron adsorption is still very small.When the pH increases to 9, B(OH) − 4 concentration increases rapidly.Compared with B(OH) − 4 , OH − concentration is still lower.So the adsorbed boron increases rapidly.When the pH further increases, OH − concentration will increase evidently, and the boron contribution decreases because of the competition for adsorption site with B(OH) − 4 .In this study, when the pH is lower than 10, K d increases as the pH increases (Fig. 2b).The K d reaches its highest values at pH 10.When the pH is higher than 10, K d decreases and shows a mild trend after pH 12.The variation characteristics of K d with the pH can also be explained partly by this model.But isotopic fractionation characteristics of boron, the maximum adsorption value and K d value of brucite deposit can not be explained by this totally.
The incorporation of B(OH) − 4 into brucite may be related to adsorption.Previous studies (Liu et al., 2004;Garcia-Soto and Camacho, 2006;Yuan et al., 2006) have indicated that the incorporation of boron into brucite or MgO was an adsorption process.When Mg 2+ exists in solution, it can be adsorbed by brucite crystallite, and the resulting positive charge can adsorb the negatively charged B(OH) − 4 .The equation is shown as follows: If model 1 is applied, the δ 11 B d should be lower than δ 11 B isw and α d−sw should be lower than 1.In this study, however, the results show that all δ 11 B d are higher than Clim.Past, 7, 693-706, 2011 Fig. 6.The fraction (a) and δ 11 B (b) of boron species calculated using different pK a versus pH values.The influence of pK a on fraction and δ 11 B of boron species is big for solution with pH<10.When pH>10, the influence of pK a is insignificant.For example, the δ 11 BB isw3 calculated using pK a 7.0, 8.0 and 9.0 (α 4/3 =0.975) are 16.08‰, 5.57‰ and -4.74‰, respectively for pH 8, and 18.44‰, 18.21‰ and 15.89‰, respectively for pH 10.When pH >11, there are basically no differences among them (Fig. 6b).
24 Fig. 6.The fraction (a) and δ 11 B (b) of boron species calculated using different pK a versus pH values.The influence of pK a on fraction and δ 11 B of boron species is big for a solution with the pH < 10.When the pH > 10, the influence of pK a is insignificant.For example, the δ 11 B isw3 calculated using pK a 7.0, 8.0 and 9.0 (α 4/3 = 0.975) are 16.08 ‰, 5.57 ‰ and −4.74 ‰, respectively for pH 8 and 18.44 ‰, 18.21 ‰ and 15.89 ‰, respectively, for pH 10.When the pH > 11, there are basically no differences among them (Fig. 6b).
δ 11 B isw and α d−sw values are higher than 1.This suggests that model 1 is not suitable to explain all the data in our study.
The apparent pH dependence of the fractionation between carbonates and clays can be described in terms of the α 3−4 and the pK a of the boric acid-borate equilibrium (Sanchez-Valle et al., 2005).This model explains the B isotopic fractionation data for inorganically and biologically precipitated carbonates and clays, and suggests that in minerals like clay and other layered hydroxides, surface charge properties play an important role by changing the apparent acid-base equilibrium constant.
In our study, pK a of the boric acid-borate equilibrium may be affected due to the presence of a positive charge on the Mg(OH) 2 surface.The fraction and δ 11 B of boron species calculated using different pK a versus pH values is shown in Fig. 6, which suggests that the influence of pK a on fraction and δ 11 B of boron species is big for a solution with the pH < 10.When the pH > 10, the influence of pK a is insignificant.For example, the δ 11 B isw3 calculated using pK a 7.0, 8.0 and 9.0 (α 4/3 = 0.975) are 16.08 ‰, 5.57 ‰ and −4.74 ‰, respectively for pH 8, and 18.44 ‰, 18.21 ‰ and 15.89 ‰, respectively for pH 10.When the pH > 11, there are basically no differences among them.The pH values of our experiments are higher 10, so the isotopic fractionation characteristics of boron are not attributed to the change of pK a .

Model 2: chemical reaction of B(OH) 3 with
Mg(OH) 2 Rodionov et al. (1991) reported that when limewater was added to seawater, CaO • B 2 O 3 • nH 2 O was deposited.This reaction is shown as follows: During deposition of brucite, a similar reaction may occur.MgO • B 2 O 3 • nH 2 O may also be deposited.This hypothesis was confirmed by the XRD analyses of brucite.Compared with Fig. 7b, the brucite peak in Fig. 8 is clear, with apparent peaks of pinnoite (Mg(BO 2 ) 2 • 3H 2 O) and szaibelyite (MgBO 2 (OH)).Judging from the quantity and intensity of these peaks, pinnoite content should be higher than that of szaibelyite.As the pH increases, the borate peak gradually weakens (Fig. 7c-h).When the pH is 12.5 (Fig. 7h), its peak is very close to that of Mg(OH) 2 , though the characteristic pinnoite peak is still observed.The pinnoite is the result of the following reaction between brucite and H 3 BO 3 : The above reaction cannot occur for B(OH) − 4 .Thus, F d3 can exceed F fsw3 due to the preferential incorporation of  all the α d−fsw are greater than 1 (Fig. 3b), which are consistent with model 2. But the α d−fsw shows a waved pattern (Fig. 3b), which indicates a complex isotopic fractionation process and cannot be explained completely by model 2.
The fraction of boric acid in solution is negligible at high pH values (Fig. 1a), so that the amount of B(OH) 3 incorporated into brucite should be little.But the F d3 is much higher than the F isw3 (Fig. 5).This may be related to the transformation of borate to boric acid.After B(OH) 3 is incorporated into brucite, F isw3 decreases.In order to keep the balance of the B(OH) 3 /B(OH)  .Judged from quantity and intensity of these peaks, pinnoite content should be higher than that of szaibelyite.As pH increases, the borate peak gradually weakens (Fig. 7c-h).When pH is 12.5 (Fig. 7h), its peak is very close to that of Mg(OH) 2 standard (Fig. 7a), though the characteristic pinnoite peak is still observed.  .Judging from the quantity and intensity of these peaks, pinnoite content should be higher than that of szaibelyite.As the pH increases, the borate peak gradually weakens (Fig. 7c-h).When the pH is 12.5 (Fig. 7h), its peak is very close to that of Mg(OH) 2 standard (Fig. 7a), although the characteristic pinnoite peak is still observed.
quantity in solution decreases compared with initial solution.
When a new equilibrium is reached, B(OH) 3 continues to incorporate into brucite, and B(OH) − 4 in solution will transform to B(OH) 3 to keep the equilibrium.This process will repeat, and finally the F d3 is much higher than the F isw3 .However, the total boron quantity in solution is much less than that in the initial solution.

Model 3: both chemical reaction and adsorption of boron incorporated into brucite
Many studies have indicated that only B(OH) − 4 is incorporated preferentially into corals marine carbonates (Vengosh et al., 1991;Hemming and Hanson, 1992;Hemming et al., 1995;Spivack et al., 1993;Sanyal et al., 1996Sanyal et al., , 2000)).Under this circumstance, all the δ 11 B values of carbonates are considered to be paralleled with the theoretical δ 11 B sw3 and are lower than that of the parent seawater; α carb−sw between carbonates and seawater are lower than 1 and increase with pH.When boron is adsorbed by oxides or clay minerals, the adsorption ability of B(OH) − 4 is much stronger than that of B(OH) 3 .The B(OH) − 4 is also incorporated preferentially into oxides or clay minerals, causing the enrichment of 10 B in oxides or clay minerals.Our results show that δ 11 B d is higher than δ 11 B isw and parallels neither with δ 11 B isw4 nor with δ 11 B isw3 , but changes from decrease to increase with an inflexion at pH 12 (Fig. 3b).This indicates that both B(OH) − 4 and B(OH) 3 are incorporated into brucite, and the latter is incorporated into brucite preferentially.Boron incorporation into brucite may be controlled by two processes: chemical reaction of B(OH) 3 with brucite and B(OH) − 4 adsorption onto brucite.The simultaneous occurrence of these two processes decides the boron concentration and isotopic fractionation of brucite.The mechanisms of boron incorporated into brucite are distinct from those of carbonates and oxides or clay minerals.
5 Judgment method for the existence of brucite in corals and its potentialapplication in δ 11 B-pH proxy and Mg/Ca-SST proxy The incorporation of Mg into coral skeleton was controlled by varying factors.Fallon et al. (1999) suggested that variations in Mg data in Porites could be a result, not of temperature changes but of possible micro-scale heterogeneities.Mitsuguchi et al. (2003) (Xiao et al., 2006a), and (c) between Mg/Ca and δ 11 B in corals in Sanya Bay (Xiao et al., 2006b).The Mg/Ca was positive with δ 11 B in inorganic calcite and increased with the seawater pH.
that overall Mg incorporation was controlled by temperature.But the physiology within the coral colony may account for differing amounts of Mg among skeletal structures.A recent study (Nothdurft et al., 2005) showed that brucite exists in a wide range of common reef-building coral in Great Barrier Reef and Florida.Elevated Mg concentrations in modern scleractinians may promote the formation of high-Mg calcite cements, as observed in Holocene corals from Heron Reef (Nothdurft et al., 2005), and then cause the deviation from the normal SST-Mg/Ca curves.Samples of brucite-bearing corals could be responsible for anomalies Mg/Ca vs. SST plots in corals (Nothdurft et al., 2005).Thus, corals should be cautiously used, especially brucite-bearing corals, to reconstruct SST.If the variation of Mg/Ca can not be defined exactly, the Mg/Ca-SST proxy would become more complex.
In addition, the influence of brucite in corals on the δ 11 B-pH proxy is more severe.Because one of the essential hypotheses for the δ 11 B-pH proxy is that only the B(OH) − 4 incorporated into corals, that is 10 B incorporated into corals preferentially (Vengosh et al., 1991;Hemming and Hanson, 1992;Sanyal et al., 1996).Under this circumstance, the boron isotope fractionation factor between coral and seawater is less than one.When brucite deposited from seawater, 11 B enriched into brucite, the fractionation factor between brucite and seawater is greater than one, which is entirely different from those into corals (Fig. 3).The occurrence of brucite in coral can change the isotopic composition of boron in coral due to the heavier B(OH) 3 species incorporated preferentially into brucite and the deviant high boron isotopic composition of corals may associate with the occurrence of brucite.If the brucite-bearing corals were used in δ 11 B-pH proxy, the measured δ 11 B of coral and the calculated pH will be higher than the normal value.Rollion-Bard et al. (2011) suggested that careful sampling is necessary before performing boron isotopic measurements in deep-sea corals.So the existence of brucite in corals can negatively influence the Mg/Ca-SST proxy and the δ 11 B-pH proxy.How to judge the variation of δ 11 B and Mg/Ca in corals caused by brucite is important for the δ 11 B-pH proxy and the Mg/Ca-SST proxy.Abnormal boron isotope fractionation (α carb/sw > 1) was observed in an inorganic calcite precipitation experiment carried out by Xiao et al. (2006a).They reasoned that the simultaneous deposition of brucite with inorganic calcite was the main reason for this observation.Under this circumstance, the δ 11 B and Mg/Ca of inorganic calcite have a good positive relationship with correlation coefficients of 0.98 (Fig. 9a).In addition, Mg/Ca was independent of SST, but increased with seawater pH (Fig. 9b), indicating the high Mg/Ca ratio and δ 11 B were due to the increasing brucite deposition as seawater pH increases.These observations provide a new method for differentiating the existence of brucite in corals.The weak negative relationship between the δ 11 B and Mg/Ca in corals in Sanya Bay (Fig. 9b) indicated there is no brucite existing in corals in this area.Thus, the relationship between δ 11 B and Mg/Ca in corals can be used to judge the existence of brucite in corals, which should provide a reliable method for better use of δ 11 B and Mg/Ca in corals to reconstruct paleo-marine environment.

Conclusions
Based on the above experimental results, the following conclusions can be drawn: 1. Boron concentration of deposited brucite and partition coefficient K d between deposited brucite and resultant seawater is controlled by the pH of solution.The incorporation capacity of boron into brucite is much stronger than that into oxides and clay minerals.
2. All the δ 11 B d values are higher than δ 11 B fsw and all the fractionation factors α d−fsw are higher than 1, indicating the preferential incorporation of H 3 BO 3 into Mg(OH) 2 .The mechanism of boron incorporated into brucite is distinct from that of carbonates deposition.
3. During deposition of brucite, both H 3 BO 3 and B(OH) − 4 incorporated into brucite.The simultaneous occurrence of boron adsorption onto brucite and the precipitation reaction of H 3 BO 3 with brucite decide the boron concentration and isotopic fractionation of the resulting brucite.
4. The existence of brucite in corals can affected the δ 11 B and Mg/Ca in corals and influences the Mg/Ca-SST proxy and δ 11 B-pH proxy negatively.The relationship between δ 11 B and Mg/Ca in corals can be used to judge the existence of brucite in corals, which should provide a reliable method for better using of δ 11 11B and Mg/Ca in corals for the reconstruction of the paleo-marine environment.

Fig. 1 .
Fig. 1.Fraction (a) of the major dissolved boron species B(OH) 3 (green curve) and B(OH) 4-(red curve) in seawater, using pK b =8.597(Dickson, 1990) and stable boron isotope fractionation (b) between them as a function of pH, using α 4-3 =0.974(Klochko et al., 2006).At low pH (pH<7), virtually all of the boron in seawater is in the B(OH) 3 species; conversely, at high pH (pH>10), virtually all of the boron is in the B(OH) 4 -species (Fig.1a).Only the B(OH) 4 − incorporated into the marine bio-carbonates is one of the

Fig. 2 .Fig. 2 .
Fig. 2. Boron concentrations (a) and partition coefficient K d (b) between brucite deposit and final seawaterfor T r =0 h and 20 h.The variation trend of the results for T r = 0 h and T r = 20 h are basically the same (Fig.2a), indicating the incorporation of boron into brucite deposit is instantaneous.[B]fsw decreases firstly and then increases quickly with the lowest value at pH 10 (Fig.2a), indicating the amount of boron incorporated into deposited brucite increases firstly and decrease as pH increases, and reaches the highest value at pH 10.The variation trend of K d is contrary to that of [B] fsw (Fig.2b).

Fig. 3 .
Fig. 3. Isotopic compositions (a) and fractionation factor (α d−fsw ) (b) of boron between brucite deposit and solution for T r = 0 h and 20 h and (c) between inorganic carbonate and seawater(Sanyal et al., 1996).The variation trend of δ 11 B for T r = 0 h and T r = 20 h are basically same and all the δ 11 B d values are enriched in the heavy isotope relative to the δ 11 B fsw (a).All the α d−fsw are greater than 1 and shows a waved pattern (b), which is entirely different from that of the carbonates deposits (c).

Fig. 5 .
Fig.5.Fraction F d3 of B(OH) 3 in brucite deposit and final seawater calculated according to δ 11 B d with different α 4−3 .The variation trend of F d3 for a different repose time is basically the same.F d3 tends to decrease with decreasing α 4−3 and is much higher than that of the initial artificial seawater (blue curve).

Fig. 7 .
Fig. 7. X-Ray diffractogram of brucite deposited from artificial seawater at different pH values.(a) XRD of brucite standard, (b) XRD of brucite from boron-free artificial seawater at pH 10.0,(c-h) XRD of brucite in our experiments.The brucite peak in our samples is clear compared with brucite standard (Fig.7a).

Fig. 7 .
Fig. 7. X-Ray diffractogram of brucite deposited from artificial seawater at different pH values.(a) XRD of brucite standard, (b) XRD of brucite from boron-free artificial seawater at pH 10.0, (c)-(h) XRD of brucite in our experiments.The brucite peak in our samples is clear compared with the brucite standard (a).

Fig. 8 .
Fig. 8. X-Ray diffractogram of brucite deposited from B-containing artificial seawater at pH 9.5.It is a magnified figure of Fig. 7c.Compared with Fig.7b, the brucite peak in Fig.8is clear, with apparent peaks of pinnoite (Mg(BO 2 ) 2 • 3H 2 O) and szaibelyite (MgBO 2 (OH)).Judging from the quantity and intensity of these peaks, pinnoite content should be higher than that of szaibelyite.As the pH increases, the borate peak gradually weakens (Fig.7c-h).When the pH is 12.5 (Fig.7h), its peak is very close to that of Mg(OH) 2 standard (Fig.7a), although the characteristic pinnoite peak is still observed.

Table 2 .
Boron concentration (ppm) and partition coefficient K d between brucite deposit and final seawater.
[B]fsw and [B] d are the boron concentrations in final seawater and deposit, respectively.

Table 3 .
Isotopic fractionation of boron between deposited brucite and solution.α d−fsw is the isotopic fractionation factor of boron between deposit and final seawater, calculated as α d−fsw = (1000 + δ 11 B d )/(1000 + δ 11 B fsw ). b (Xiao et al., 2006b)een Mg/Ca and δ 11 B in inorganic calcite, (b) between Mg/Ca and seawater pH(Xiao et al., 2006a), and (c)between Mg/Ca and δ 11 B in corals in Sanya Bay(Xiao et al., 2006b).The Mg/Ca was positive with δ 11 B in inorganic calcite and increased with seawater pH.Relationship (a) between Mg/Ca and δ 11 B in inorganic calcite, (b) between Mg/Ca and seawater pH Weinbauer et al. (2000)t and believed to be a result of a biological/metabolic effect.Weinbauer et al. (2000)studied the potential use of Mg as an environmental indicator in the coral Corallium rubrum, founding www.clim-past.net/7/693/2011/Clim.Past, 7, 693-706, 2011