LIMNOLOGY

and
OCEANOGRAPHY: METHODS Limnol. Oceanogr.: Methods 11, 2013, 466–474
© 2013, by the American Society of Limnology and Oceanography, Inc.

Direct chromatographic separation and quantification of calcium

and magnesium in seawater and sediment porewaters
Melissa Meléndez1*, Ekaterina P. Nesterenko2,3, Pavel N. Nesterenko2, and Jorge E. Corredor1
1
Department of Marine Sciences, University of Puerto Rico, Mayagüez Campus, Puerto Rico
2
Australian Centre for Research on Separation Science (ACROSS), School of Chemistry, University of Tasmania
3
Irish Separation Science Cluster (ISSC), National Centre for Sensor Research, Dublin City University, Glasnevin, Dublin, Ireland

Abstract
Direct analysis of Ca2+ and Mg2+ is required for accurate determination of metastable carbonate mineral phase
saturation states (ΩCaCO3; ΩMgCO3) in seawater, sediment porewaters, and other high ionic strength brines. To
this end, we have implemented a method using High Performance Chelation Ion Chromatography (HPCIC) in
which metal ion complexation at the stationary phase renders separation efficiency insensitive to high ionic
strength matrix effects common to other ion chromatography (IC) methods. This method, using direct auto-
mated on-column injection, vastly increases sample throughput capacity in comparison to current titration
methods. Calcium and magnesium ions in IAPSO standard seawater were selectively separated using a mono-
lithic silica column (100 × 4.6 mm ID) activated with a covalently bonded iminodiacetic acid (IDA) chelator.
The colored ion complexes resulting from post-column reaction (PCR) of the ions with a metallochromic indi-
cator, in this case 4-(2-pyridylazo)-resorcinol (PAR), were detected spectrophotometricaly at 510 nm.
Optimization of flow rate, eluent concentration, pH, and sample injection volume allowed baseline separation
of Mg2+(0.05474 mol kg–1) and Ca2+ (0.01065 mol kg–1) in less than 8 min using 2 μL seawater sample injections.
At a flow rate of 1 mL min–1, peak elutions occurred respectively at 4 and 5 min, using an eluent containing 0.1
M potassium chloride and 1 mM nitric acid adjusted to pH 2.5. Retention time variability below 0.5% for both
metals following more than 200 injections indicates long-term stability of the derivatized monolithic silica col-
umn. Method application to marine sediment porewaters is discussed.

Accurate determination of Mg2+and Ca2+ ion concentrations processes including biogenic and abiogenic precipitation, as
and their relative proportions in seawater, marine sediment well as carbonate sediment dissolution, can result in devia-
porewaters and other environmental high ionic strength tions from this norm (Gledhill 2005; Ribou et al. 2007). Dur-
brines is troublesome despite their high concentrations due to ing carbonate precipitation from seawater, Mg2+ for example,
the complexities of the matrix and chemical similarity which can co-precipitate with Ca2+ yielding high-magnesium calcite
results in their co-precipitation (Traganza and Szabo 1967; thereby altering the Ca2+/Mg2+ ratio and the ion activity prod-
Carpenter and Manella 1973; Kanamori and Ikegami 1980). uct relative to Mg-calcite mineral phases. These Mg-calcite
Although Mg2+and Ca2+ ion concentrations in open ocean mineral phases (low and high Mg-calcites) are not well under-
waters are largely conservative with respect to salinity, coastal stood due to problems involving the precision of mea-
surements and uncertainty regarding the basic thermody-
*Corresponding author: E-mail: melissa.melendezoyola@upr.edu namic solubility and kinetic properties of these phases (Morse
et al. 2006). Accurate determination of Mg2+ and Ca2+ ion con-
Acknowledgments
Authors would like to give thanks to Drs. Dwight Gledhill and centrations can help elucidate such difficulties.
Andreas Andersson for their assistance and recommendations through- Changes in seawater saturation state (Ω) with respect to
out this research. Special thanks are due to the research group at the metastable carbonate phases are not only affected by changes
Australian Centre for Research on Separation Science (ACROSS) and to in the seawater CO32– concentration. Variations in the Mg2+
Dr. Brett Paull at University of Tasmania, Australia. Support for this study
and Ca2+ ion seawater concentrations and their ratio imply
was provided, in part, by the Caribbean Coastal Ocean Observing
System (CariCOOS), NOAA Coral Reef Conservation Program, Puerto changes in Mg-calcite composition and solubility, Mg content
Rico Sea Grant College Program, ExxonMobil, Department of Marine in marine calcifying organisms skeletons, as well as changes in
Science UPR-M and ACROSS. seawater Ω (Andersson et al. 2008). The delicate equilibrium
DOI 10.4319/lom.2013.11.466 between these two alkaline earth metal cations is triggered by

466
 15415856, 2013, 9, Downloaded from https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lom.2013.11.466 by Algeria Hinari NPL, Wiley Online Library on [11/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Meléndez et al. Calcium and Magnesium in Seawater

slight changes in alkalinity and carbon dioxide tension that substrates serve as the stationary phase. Metal ion analytes
may cause their precipitation or dissolution (Laurence 1926; form very stable complexes with these ligands and hence effi-
Brewer et al. 1975). In sediment porewaters, changes in Ca2+, cient separation is achieved (Nesterenko et al. 2011;
Mg2+, and CO32– concentrations arise from the precipitation or Nesterenko et al. 2013).The use of chelating ion-exchangers to
dissolution of calcium carbonate minerals (Koczy 1956; Tra- form kinetically labile surface complexes and retain metal ions
ganza and Szabo 1967; Kanamori and Ikegami 1980; Kleypas according to the stability of corresponding complexes is one
et al. 2006; Ribou et al. 2007). The direct quantification of the of the multiple advantages in high performance liquid chela-
Mg2+ and Ca2+ seawater ion concentrations can help in under- tion ion chromatography (HPLCIC) (Nesterenko and Jones
standing the chemical behavior of seawater metastable car- 2007). Modification of monolithic silica columns with cova-
bonate mineral phases and in more accurate determination of lently bonded chelating iminodiacetic acid (IDA) groups has
their corresponding seawater Ω (Kleypas et al. 2006; Ribou et proven to allow excellent cation separation and increased
al. 2007). peak efficiencies compared with other columns.
Current methods for the determination of Mg2+ and Ca2+ We here describe implementation of an HPLCIC method
ion concentrations in seawater use gravimetric procedures and for separation and quantification of Mg2+ and Ca2+ ions in sea-
ion-exchange separation combined with titration methods. water in less than 8 min. We use a monolithic silica column
Due to the precision difficulties, seawater Mg2+ ion concentra- derivatized with a covalently bonded IDA chelator for separa-
tion is usually determined as the difference between total alka- tion, and post-column derivatization of the ions with 4-(2-
line earth metals and Ca2+ plus strontium (Kanamori and pyridylazo)-resorcinol (PAR) for optical detection and quan-
Ikegami 1980). Meanwhile Ca2+ is selectively titrated with Zin- tification. The metallochromic reagent PAR forms
con (Zn- ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′- water-soluble complexes with Mg2+ and Ca2+ ions of moderate
tetraacetic acid [EGTA]) (Culkin and Cox 1976; Kanamori and molar absorptivities (~ 104 at about 500 nm), therefore
Ikegami 1980), ethylenediamine-N,N,N′,N′-tetraacetic acid exhibiting robust sensitivity for spectrophotometric detection
(EDTA) (Riley and Tongudai 1967), or glyoxal-bis(2-hydrox- (Jezorek and Freiser 1979). Monolithic HPLC columns,
yanil) (GBHA) (Tsunogai et al. 1968). Other methods incorpo- employing a continuous silica matrix etched with porous
rate ion selective electrodes as end-point indicators (Whitfield channels, surpass traditional packed bead column perform-
et al. 1969; Růžička et al. 1973; Lebel and Poisson 1976; ance with higher separation efficiency, reduced retention
Kanamori and Ikegami 1980). These methods are laborious times, and low column backpressure.
and time-consuming, and as a result, sample throughput is Benefits of this method include elimination of the need for
limited in the best cases to a few tens of sample analyses per sample pretreatment or manipulation, high sample through-
day. Additionally, most of these techniques do not have suffi- put achievable with automated sample injection, low sample
cient resolution to detect the small changes due to calcifica- volume required, reduced number of solutions necessary, lack
tion processes. A direct in situ method using a custom-made of interference from other ionic compounds, method simplic-
ion-selective electrode has been described (Wenzhöfer et al. ity and reliability, and reduced sensitivity to the ionic strength
2001), but resolution is poor and the electrode is not com- of the sample matrix. Whereas we have yet to achieve the
mercially available. Other instrumental methods include canonical precision of 0.1% quoted for seawater applications
inductively coupled plasma spectrometry (ICP-MS), atomic (Carpenter and Manella 1973; Kanamori and Ikegami 1980;
absorption spectrophotometry (AAS), and flame atomic Olson and Chen 1982), the method can currently be applied
absorption spectrophotometry (FAA). However sample pre- to sediment porewaters and further method refinement is
treatment is needed, interferences from other major ionic expected to achieve this requirement.
components in the matrix are expected, and analysis costs can To explore anticipated improvement in method repro-
be high. ducibility with increased injection volume but given limita-
Recently, chromatographic techniques have been devel- tion to seawater Ca2+ and Mg2+ analysis imposed by column-
oped that can provide higher energy interactions between the loading capacity and detector saturation, we performed a
ionic analytes of interest and selected adsorbents or stationary series of experiments using increasing injection volumes of
phases increasing significantly the degree of separation selec- Mn2+ ion proxy at low concentration. Manganese ion was cho-
tivity. Chelation ion chromatography (CIC), first described by sen because it exhibits greater molar absorptivity with the PAR
Moyers and Fritz in 1977, is a retention mechanism that reagent allowing use of more dilute and less acidic solutions.
allows specific interactions between a dissolved metal ion ana-
lyte and a chelating stationary phase. Paull et al. (1996) Materials and procedures
demonstrated the potential application of CIC to the problem Monolithic silica IDA modified column
of Mg2+ and Ca2+ ion separation using a dynamic chelating ion A monolithic bare silica column (Phenomenex 100 × 4.6
exchange mechanism whereby a chelator dissolved in the car- mm) was modified with IDA chelator through the activation
rier coats a porous graphitic carbon column. In recent devel- of silanol groups at the surface of the silica monolith column
opments, selected organic ligands covalently bonded to inert with distilled water at 60°C followed by recycling of mixture

467
 15415856, 2013, 9, Downloaded from https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lom.2013.11.466 by Algeria Hinari NPL, Wiley Online Library on [11/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Meléndez et al. Calcium and Magnesium in Seawater

IDA and 3-glycidoxypropyltriethoxysilane through the col- ment and then replacing it with the well sampler. After inser-
umn at 70°C (for method details see Sugrue et al. 2003; tion into the sediment, the sampler was left on site allowing
Nesterenko and Jones 2007; Nesterenko et al. 2013). Surface repetitive sampling at identical locations and depth intervals.
treatment and functionalization of the continuous unitary Porewater samples were collected in situ by withdrawing pore-
porous structure and structure of the bonded layer within water using two 60 cc syringes and storing in 125 mL plastic
such columns have been described by Sugrue et al. (2004) and sample bottles. Each sample was filtered through 0.45 μm
Nesterenko et al. (2013). membrane filters and poisoned with 60 μL of a saturated HgCl2
Reagents and solutions solution to prevent biological alteration of the sample. Pore-
For photometric detection, we used PAR reagent (CAS# water salinity was determined using the Guildline Autosal
1141-59-9, acid form – Fluka, 99% purity) as a post-column 8400B salinometer with a precision of ± 0.003. Conductivity,
reagent. We prepared stock solutions of 1 mM PAR and 2 M Temperature, and Depth (CTD) casts of the overlying water col-
ammonium hydroxide (analytical reagent grade). The high pH umn were routinely performed. Surface and bottom samples of
of the stock solution prevents adsorption onto plastic surfaces. the overlying seawater were collected for analyses as well using
The standard post column reagent was prepared by dilution to a Van Dorn bottle.
0.05 mM PAR. To adjust the pH to ~ 10.4, we used 2 M nitric
acid (analytical reagent grade). The post-column reagent thus Assessments
prepared is stable for weeks if not months, and will not need Optimization of the method
filtering, degassing, or an overpressure of inert gas. We tested eluent concentration over ranges of 0.1 to 0.5 M
The mobile phase was prepared using 0.1 M potassium chlo- KCl and 1 to 4 mM HNO3 with pH between 2.5 and 3.0. Base-
ride (KCl) and 1 mM HNO3, pH of ~ 2.5. Standard seawater line cation separation and peak shape were found to be opti-
(International Association for the Physical Sciences of the Ocean mal at 0.1 M of KCl and 1 mM HNO3. Systematic reduction of
– IAPSO, batch 149; 10 May 2007) with salinity 34.994 was pur- HNO3 and KCl concentrations improved the response and
chased from OSIL (Havant, UK). Stoichiometric reference com- produced sharper and narrower peak shapes. These optimized
position of IAPSO standard seawater provides the best current eluent concentrations allowed return to baseline between
estimation of Mg2+ (0.05474 mol kg–1) and Ca2+ (0.01065 mol peaks for up to 0.5 s. Variations in eluent pH from 2 to 3 were
kg–1) concentrations in seawater (Millero et al. 2008). tested, but no significant changes in retention or peak shapes
We used Nalgene bottles for storage of all stock and work- were observed.
ing solutions due to their low metal contamination. Glassware The effect on reaction completion of varying PCR reagent
and plasticware were acid washed before use with 10 mL of 1 flow rate was tested. Increasing PCR reagent flow rate from 0.7
M nitric acid followed by a rinse with deionized water pro- to 1 mL min–1 resulted in increased photometric response for
vided form a Milli-Q system (Millipore, Bedford, USA). both analytes in standard seawater (Fig. 1). Maximum
Chromatographic instrumentation absorbance response was obtained at a flow rate of 1 mL min–1.
A Waters 2695 HPLC Separations Module (Waters, Milford, To minimize the ambiguities introduced in addressing both
MA, USA) chromatography system was used. The autosampler the reagent and eluent delivery proportions, absorbance
built in to the Separation Module allowed runs of 178 samples
in a single analytical sequence. Column oven was set to 30°C
for all separations. A post column reaction (PCR) flow system
was used to allow cation detection. The 1/16” polypropylene
mixing coil used in the PCR was about 2.5 m long using a
high-pressure pump (Model 350 Scientific System Inc.). The
colored PAR-derivatized cations were detected spectrophoto-
metricaly using a model 2487 UV/VIS spectrophotometric
detector operated at 510 nm. Data were processed using the
Waters Empower 3 Software.
Porewater samples
Stainless steel well samplers (3/4-inch ID) were developed,
which allow porewater sampling down to 20 cm sediment
depth. Each sampler is placed 1 m apart on a 10 m transect
along the reef. Each sampling port consists of thirty 1/16-inch
holes drilled around the sample in a 1 cm span. Samples were
taken at 2 cm resolution through the upper 20 cm of the sedi-
ment column. Following the technique described by Falter and Fig. 1. Effect of PCR reagent flow rates on Ca2+ and Mg2+ responses using
Sansone (2000), we installed the samplers by first hammering standard seawater as a probe. Maximum absorbance response is observed
a stainless steel tube (54 cm long, 5/8-inch ID) into the sedi- at 1 mL min–1.

468
 15415856, 2013, 9, Downloaded from https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lom.2013.11.466 by Algeria Hinari NPL, Wiley Online Library on [11/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Meléndez et al. Calcium and Magnesium in Seawater

response was investigated through standard addition (Sugrue from 0.99 to 0.95, this does not significantly compromise ana-
et al. 2003). Reagent flow rates tested ranged from 0.5 to 1 mL lyte determination. Fig. 2 shows the dependence of response
min–1. Table 1 shows the changes in linear regression coeffi- against concentration of both metals using different standard
cients with variation of PCR flow rate of Mg2+ and Ca2+ ions. addition concentrations of Mg2+ and Ca2+, at 1 mL min–1 elu-
Peak absorbance exhibits a linear relationship with Ca2+ ion ent flow rate and 0.9 mL min–1 PCR reagent flow rate.
concentration (R2 = 0.99, n = 8) at reagent flow rate of 0.5 mL Effect of sample injection volume was tested for 2, 5, and 10
min–1, but Mg2+ is incompletely derivatized at this low reagent μL by assessing replicate reproducibility of three standard sea-
flow rate. The correlation coefficient for magnesium increased water injections (n = 9). Increasing sample volume resulted in
significantly with increased reagent flow rate. Highest linear lower reproducibility as shown by the increased standard devi-
correlation between reagent flow rate and absorbance ation of peak areas (Fig. 3). Response reproducibility was best
response for Mg2+ was achieved at 0.9 mL min–1 (R2 = 0.98, n = with smallest injected volume of the sample. Injection volume
10). Despite a modest concurrent Ca2+ coefficient decrease of 2 μL was consequently selected for routine operation.
Increased detector response with increased PRC reagent
flow rate, but the absence of plateaus for either Ca2+ or Mg2+
Table 1. Coefficients of determination of linear regressions (R2)
of Mg2+ and Ca2+ standard concentrations versus absorbance at (Fig. 1), indicates that post-column reaction completion was
different reagent flow rates using standard additions. not reached even at the highest flow rate possible (1 mL
min–1). These circumstances result in limited method repro-
Reagent flow rate ducibility despite intentional minimization of injection vol-
(mL min–1) R2 Mg2+ R2 Ca2+ ume to the smallest injection loop possible (1 mL).
0.5 0.05 0.99 To explore the effect of varying sample injection volume on
0.6 0.23 0.97 reproducibility using the Mn2+ ion proxy (given the column
0.7 0.51 0.96 loading limitation for Ca2+ and Mg2+ ions) at a concentration
0.8 0.73 0.82 of 45.5 nM prepared from standard 1000 ppm in 0.5 M nitric
0.9 0.98 0.95 acid, we performed 10 consecutive runs each for 2, 4, 10, and
20 μL injection volumes keeping column temperature and
sample temperature constant at 25°C. The eluent used in this
test was slightly different (3 mM HNO3; 0.1 M KCl) from that
used for seawater Mg2+ and Ca2+ analysis to minimize run time.
The post column reagent was as previously described delivered
at 0.9 mL min–1. Increased injection volume using the Mn2+
ion proxy resulted in reduction of the relative standard devia-
tion (RSD) from 4.8% at 2 μL injection volume to 0.7% at 20
μL injection volume (Fig. 4).
Method performance following optimization
A chromatogram performed under optimized parameters
using the IDA-modified silica monolithic column shows com-

Fig. 2. Linearity of Ca2+ (R2 = 0.99, n = 8) and Mg2+ (R2 = 0.98, n = 10)
standard addition at reagent flow rate of 0.9 mL min–1. Standards were Fig. 3. Effect of sample volume injection on relative instrument
measured in duplicate. response.

469
 15415856, 2013, 9, Downloaded from https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lom.2013.11.466 by Algeria Hinari NPL, Wiley Online Library on [11/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Meléndez et al. Calcium and Magnesium in Seawater

Fig. 6. Column stability as indicated by retention time consistency for

Mg2+ and Ca2+ in standard triplicate seawater samples analyzed following
every ~ 30 porewater sample injections.

Fig. 4. Effect of increasing sample injection volumes on method repro-

ducibility using Mn2+ metal as an ion proxy.
deliver separation in a period of less than 4 min, selectivity,
peak symmetry, and resolution of the massive Ca2+ and Mg2+
peaks in our samples are favored at the lower flow rate with
the higher retention time.
Retention time variability was used to assess long-term col-
umn stability. During operational runs, a sample of standard
seawater was analyzed every ~ 30 injections. Twenty-one sam-
ples of standard seawater were analyzed through the sequence
of 60 sediment porewater samples. The average retention
times were 4.15 ± 0.01 min for Mg2+ and 5.44 ± 0.02 min for
Ca2+. Retention time variability over the 656 min (10.9 h) of
consecutive injections was 0.4% and 0.5%, respectively, for
Mg2+ and Ca2+ (Fig. 6).
Throughout a sequence of 243 consecutive sample injec-
tions, the maximum variability from the stoichiometric refer-
ence composition of standard seawater defined by Millero et
al. (2008) (Mg2+ 0.05474 mol kg–1 and Ca2+ 0.01065 mol kg–1)
was 1% for Mg2+ and 2% for Ca2+ (Fig. 7 and Table 2).
Determination of Mg and Ca in sediment porewaters
Fig. 5. Chromatogram of Mg2+ and Ca2+ in standard seawater obtained Example of method application
using eluent of 0.1 M of KCl and 1 mM of HNO3 at pH ~ 2.5, flow-rate,
The optimized HPCIC method enabled analysis of high
1.0 mL min–1; sample injection volume, 2 μL; PCR 0.05 mM PAR at pH ~
10.4, flow rate, 0.9 mL min–1; photometric detection at 510 nm. ionic strength seawater and porewaters samples. Vertical dis-
tribution and temporal variability of Mg2+ and Ca2+ in sedi-
ment porewaters collected at a mid-shelf reef off La Parguera,
plete baseline separation of Mg2+ and Ca2+ ions in standard sea- Puerto Rico, from June to September 2011 was examined. We
water achieved in less than 8 min at a flow rate of 0.9 mL analyzed a series of 60 samples in triplicate (180 injections)
min–1 (Fig. 5). Column efficiencies calculated from chromato- under optimized conditions as described above. Average RSD
graphic peaks are 13,720 and 24,800 theoretical plates per for triplicate samples was 1% and 2%, respectively, for Mg2+
meter for Mg2+ and Ca2+, correspondingly. These numbers are and Ca2+ ions. The maximum and the minimum RSD regis-
in a good agreement with values ranging 18,000 to 37,560 tered for one sample in triplicate was 2% and 7% and 0.1%
reported for such columns (Sugrue et al. 2003). The difference and 0.2%, respectively, for Mg2+ and Ca2+ ions.
in efficiency calculated for the massive Mg2+ peak is connected In general, Mg2+ and Ca2+ sediment porewater concentra-
with partial overloading of the column, which caused some tions at Enrique Reef increased with depth in the sediment col-
peak broadening. Although increased eluent flow rate can umn. Whereas temporal changes are apparent, no definite tem-

470
 15415856, 2013, 9, Downloaded from https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lom.2013.11.466 by Algeria Hinari NPL, Wiley Online Library on [11/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Meléndez et al. Calcium and Magnesium in Seawater

Table 2. Variability of apparent Mg2+ and Ca2+ ion concentra-

tions and ionic ratio in standard seawater (SSW) analyzed
throughout the porewater sample analysis run. Percent difference
from Mg2+ and Ca2+ ion concentrations in standard seawater as
established by Millero et al. (2008).

[Mg2+] [Ca2+] % Delta Mg2+ % Delta Ca2+

n (mol kg–1) (mol kg–1) to SSW to SSW
1 0.0550 0.0107 0 0
2 0.0548 0.0107 0 0
3 0.0544 0.0106 –1 –1
4 0.0547 0.0111 0 4
5 0.0545 0.0105 0 –1
6 0.0550 0.0104 0 –3
7 0.0546 0.0106 0 –1
Fig. 7. Column stability as indicated by the concentration of Mg2+ and 8 0.0550 0.0106 0 0
Ca2+ in standard seawater analyzed following every ~ 30 porewater sam- 9 0.0547 0.0108 0 1
ple injections. 10 0.0548 0.0107 0 0
11 0.0547 0.0107 0 0
poral trend was evident (Fig. 8).The maximum increases over 12 0.0547 0.0106 0 –1
sediment-water interface surface values to 16 cm depth within 13 0.0545 0.0108 0 1
the sediment were 0.0063 and 0.0038 mol kg–1 for Mg2+ and 14 0.0563 0.0110 3 3
Ca2+, respectively. These changes with depth are large relative 15 0.0534 0.0102 –2 –4
to the measurement error. Porewater Mg2+/Ca2+ ratios decreased 16 0.0548 0.0106 0 0
with depth presumably as a result of sediment dissolution of 17 0.0547 0.0107 0 0
metastable carbonate phases. Maximum and minimum 18 0.0549 0.0106 0 0
Mg2+/Ca2+ ratios were 5.37 and 4.22, respectively (Fig. 9). 19 0.0546 0.0106 0 0
20 0.0547 0.0107 0 1
Discussion Avg 0.0547 0.0107 0 0
The method here described optimizes chelation-based sep- SD 0.0005 0.0002 1 2
aration of the alkaline earth metal ions Mg2+ and Ca2+ at high
concentrations on the monolithic IDA column using a high bonded chelating reagents in the stationary phase of the mono-
ionic strength/low pH eluent. The method makes possible lithic column reduced the necessity for dilution, sample pre-
rapid automated analysis of Mg2+ and Ca2+ in high ionic treatment, or the use of multi-column separation techniques.
strength matrices such as marine sediment porewaters. The high retention of Mg2+ and Ca2+ on the surface monolithic
Completion of the post-column complexation reaction phase is evident. The column can be used to analyze other alka-
with the colored reagent posed an analytical challenge due to line earth metals, such as Sr2+ and Ba2+ in samples containing
the high concentration of Mg2+ and Ca2+, third and fourth excess Mg2+ and Ca2+ (Sugrue et al. 2003; Nesterenko et al. 2013).
most abundant ions in seawater. Magnesium was a particular The method using the IDA-modified silica monolithic col-
challenge because of its high concentration in seawater and umn described in this study offers new possibilities to gain
mainly because of its short residence time in the chromato- meaningful insight into the biogeochemical processes occur-
graphic column. The latter factor means that the Mg2+ band ring in permeable sediments. Organic matter remineralization
migrates through a significant part of the chromatographic processes and the concomitant metabolic CO2 production
column together with the massive band of alkali metal cations force carbonate dissolution in aerobic surface layers of the cal-
from seawater resulting in column overloading and peak careous marine sediments resulting in increased Mg2+ and Ca2+
broadening. Kinetically, complex formation with PAR was concentrations at depth within the sediment (Burdige and
“fast.” Increasing the PCR reagent delivery allowed the com- Zimmerman 2002; Andersson et al. 2006). Direct porewater
plexes to form through the post-column reaction. This process Mg2+ and Ca2+ ion measurements provide additional evidence
is apparent in the resulting chromatogram (Fig. 5), which for the occurrence of sediment carbonate dissolution and can
because of the fast kinetics of chelate formation and dissocia- be used to address the question of preferential dissolution of
tion shows relatively narrow peak shapes. metastable carbonate phases (Mackenzie et al. 1983; Morse et
Column efficiency was not compromised throughout the al. 1985, 2006; Burdige and Zimmerman 2002).
analytical run of 243 samples, and no significant variability of System limitations for handling the massive alkaline earth
retention times was observed (see Fig. 6). The use of covalently cation concentrations of seawater remain. Addressing the

471
 15415856, 2013, 9, Downloaded from https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lom.2013.11.466 by Algeria Hinari NPL, Wiley Online Library on [11/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Meléndez et al. Calcium and Magnesium in Seawater

Fig. 8. Vertical porewater profiles for Mg2+ and Ca2+ ion concentration at Enrique Reef during June (left), July (center), and September 2011 (right).

Fig. 9. Vertical porewater profiles for Mg2+/Ca2+ ion ratios at Enrique Reef during June (left), July (center), and September 2011 (right).

“chemical” problems of column overload, detector saturation, injected volumes (up to 20 μL) (see Fig. 4) confirms the poor
and reaction completion by sample volume reduction resulted performance of low volume sample injection and points the
in the “mechanical” problem of poor injection reproducibil- way toward method optimization.
ity. To confirm the dependence of reproducibility on injection
sample volume, we used a very dilute Mn2+ solution so as to Comments and recommendations
assure operation within the linear range of the calibration Although the method as here presented is applicable to
plot. Dramatic improvement of reproducibility with larger study large variations of Mg2+ and Ca2+ in marine sediment

472
 15415856, 2013, 9, Downloaded from https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lom.2013.11.466 by Algeria Hinari NPL, Wiley Online Library on [11/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Meléndez et al. Calcium and Magnesium in Seawater

porewaters, further improvement of method precision will be pling in reef sediments. Coral Reefs 19(1):93-97 [doi:10.
necessary for the determination of small changes in seawater. 1007/s003380050233].
Observations of alkalinity changes indicate that calcification Gledhill, D. K. 2005.Calcite dissolution kinetics and solubility
in coral reef environments can be expected to change Ca2+ in Na-Ca- Mg-Cl brines of geologically relevant composi-
concentration by less than ~ 0.050 mM over a daily cycle pos- tion at 0.1 to 1 bar pCO2 and 25 to 80°C. Ph.D. thesis, Texas
ing a considerable analytical challenge. A&M Univ.
Accuracy of our analyses was compromised by the inability Laurence, I. 1926. The precipitation of calcium and magne-
to reduce sample injection below 1 μL since we injected very sium from sea water. J. Mar. Biol. Assoc. U.K. 14(02):441-
concentrated samples into a column of internal diameter 4 446 [doi:10.1017/S002531540000792X].
mm. Our experiments with the Mn2+ ion proxy, however, Jezorek, J. R., and H. Freiser. 1979. 4-(Pyridylazo)resorcinol-
demonstrate significant improvement of method repro- based continuous detection system for trace levels of metal
ducibility using greater injection volume (Fig.4). We conse- ions. Anal. Chem. 51(3):373-376 [doi:10.1021/ac50039
quently recommend the use of higher capacity columns a012].
(internal diameter 8 mm for example) to allow increasing sam- Kleypas, J. A., R. A. Feely, V. J. Fabry, C. Langdon, C. L. Sabine,
ple volume injection to more reproducible volumes in the and L. L. Robbins. 2006. Impacts of ocean acidification on
range of 10-20 μL. Auto sampler injection accuracy, in partic- coral reefs and other marine calcifiers: A guide for future
ular, increases substantially in this range. research, report of a workshop held 18-20 April 2005, St.
Fully automated chromatographic analysis of Mg2+ and Ca2+ Petersburg, FL, sponsored by NSF, NOAA, and the U.S. Geo-
in seawater exhibits significant advantages over titration logical Survey, p. 1-88.
methods including rapid sample throughput, low sample vol- Kanamori, S., and H. Ikegami. 1980. Computer-processed
ume, and decreased operator labor. A single chromatographic potentiometric titration for the determination of calcium
determination is 5 to10 times faster than the corresponding and magnesium in seawater. J. Oceanogr. Soc. Japan
determination by titration. Although autotitrators can be 36(4):177-184 [doi:10.1007/BF02070330].
equipped with autosamplers, sample volumes used are orders Koczy, F. F. 1956.The specific alkalinity. Deep Sea Res. 3(4):279-
of magnitude higher than for HPCIC resulting in lower sam- 288 [doi:10.1016/0146-6313(56)90018-1].
ple loading capacity. Because the cost of the basic equipment, Lebel, J., and A. Poisson. 1976. Potentiometric determination
reagents, and materials is comparable for both methods, chro- of calcium and magnesium in seawater. Mar. Chem.
matographic determination is more cost-effective, especially 4(4):321-332 [doi:10.1016/0304-4203(76)90018-9].
for large sample numbers. Mackenzie, F.T.,W. D. Bischoff, F.C. Bishop, M. Loijens, J. Schoon-
maker, and R. Wollast. 1983. Mg-calcites: low temperature
References occurrence, solubility and solid-solution behavior, p. 97-143.
Andersson, A. J., F. T. Mackenzie, and N. R. Bates. 2008. Life on In R. J. Reeder [ed.], Reviews in mineralogy, carbonates: Min-
the margin: implications of ocean acidification on Mg-cal- eralogy and chemistry. Mineralogical Society of America.
cite, high latitude and cold-water marine calcifiers. Mar. Millero, F., R. Feistel, D.Wright, and T. Mcdougall. 2008. The
Ecol. Prog. Ser. 373:265-273 [doi:10.3354/meps07639]. composition of standard seawater and the definition of the
———, F. T. Mackenzie, and A. Lerman. 2006. Coastal ocean reference-composition salinity scale. Deep Sea Res. I
CO2–carbonic acid–carbonate sediment system of the 55(1):50-72 [doi:10.1016/j.dsr.2007.10.001].
Anthropocene. Glob. Biogeochem. Cycles 20(1):1-13 Morse, J., A. J. Andersson, and F. Mackenzie. 2006. Initial
[doi:10.1029/2005GB002506]. responses of carbonate-rich shelf sediments to rising atmos-
Brewer, P. G., G. T. F. Wong, M. P. Bacon, and D.W. Spencer. pheric pCO2 and “ocean acidification:” Role of high Mg-
1975. An oceanic calcium problem? Earth Planet. Sci. Lett. calcites. Geochim. Cosmochim. Acta 70(23):5814-5830
26(1):81-87 [doi:10.1016/0012-821X(75)90179-X]. [doi:10.1016/j.gca.2006.08.017].
Burdige, D. J., and R. C. Zimmerman. 2002. Impact of sea grass Morse, J. W., J. J. Zullig, L. D. Bernstein, F. J Millero, P. J. Milne,
density on carbonate dissolution in Bahamian sediments. A. Mucci, and G. R. Choppin. 1985. Chemistry of calcium
Limnol. Oceanogr. 47(6):1751-1763 [doi:10.4319/lo.2002. carbonate-rich shallow water sediments in the Bahamas.
47.6.1751]. Am. J. Sci. 285(2):147-185 [doi:10.2475/ajs.285.2.147].
Carpenter, J. H., and M. E. Manella. 1973. Magnesium to chlo- Moyers, E. M., and J. S. Fritz. 1977. Preparation and analytical
rinity ratios in seawater. J. Geophys. Res. 78(18):3621-3626 applications of a propylenediaminetetraacetic acid resin.
[doi:10.1029/JC078i018p03621]. Anal. Chem. 49(3):418-423 [doi:10.1021/ac50011a023].
Culkin, F., and R. A. Cox. 1976. Sodium, potassium, magne- Nesterenko, E. P., P. N. Nesterenko, B. Paull, M. Meléndez, and
sium, calcium and strontium in sea water. Deep Sea Res. J. E. Corredor. 2013. Fast direct determination of strontium
Oceanogr. Abstr. 13(5):789-804 [doi:10.1016/0011-7471 in seawater using high-performance chelation ion chro-
(76)90905-0]. matography. Microchem. J. 111:8-15 [doi:10.1016/j.microc.
Falter, J. L., and F. J. Sansone. 2000. Shallow porewater sam- 2012.09.003].

473
 15415856, 2013, 9, Downloaded from https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lom.2013.11.466 by Algeria Hinari NPL, Wiley Online Library on [11/04/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Meléndez et al. Calcium and Magnesium in Seawater

Nesterenko, P. N., and P. Jones. 2007. Recent developments in acid functionalised monolithic silica chelating ion
the high-performance chelation ion chromatography of exchanger for rapid determination of alkaline earth metal
trace metals. J. Separat. Sci. 30(11):1773-1793 [doi:10.1002/ ions in high ionic strength samples. Analyst 128(5):417-
jssc.200700126]. 420 [doi:10.1039/b303467d].
———, P. Jones, B. Paull, and Royal Society of Chemistry ———, P. Nesterenko, and B. Paull. 2004. Ion exchange prop-
(Great Britain). 2011. High performance chelation ion erties of monolithic and particle type iminodiacetic acid
chromatography. In R. M. Smith [ed.], RSC Chromato- modified silica. J. Separat. Sci. 27(10-11):921-930 [doi:10.
graphic Series, RSC Chemistry, Cambridge, 2011, p. 1002/jssc.200401794].
284–303 [doi:10.1039/9781849732314-00284]. Traganza, E. D., and B. J. Szabo. 1967. Calculation of calcium
Olson, E. J., and C.T. A. Chen. 1982. Interference in the deter- anomalies on the Great Bahama Bank from alkalinity and
mination calcium in seawater. Limnol. Oceanogr. chlorinity data. Limnol. Oceanogr. 12(2):281-286 [doi:10.
27(2):375-380 [doi:10.4319/lo.1982.27.2.0375]. 4319/lo.1967.12.2.0281].
Paull, B., P. A. Fagan, and P. R Haddad. 1996. Determination of Tsunogai, S., M. Nishimura, and S. Nakaya. 1968. Calcium and
calcium and magnesium in sea-water using a dynamically magnesium in sea water and the ratio of calcium to chlo-
coated porous graphitic carbon column with a selective rinity as a tracer of water-masses. J. Oceanogr. Soc. Japan
metallochromic ligand as a component of the mobile 24(4):153-159.
phase. Anal. Commun. 33(6):193-196 [doi:10.1039/ac9963 Wenzhöfer, F., M. Adler, O. Kohls, C. Hensen, B. Strotmann, S.
300193]. Boehme, and H. Schulz. 2001. Calcite dissolution driven by
Ribou, A. C., J. M. Salmon, J. Vigo, and C. Goyet. 2007. Mea- benthic mineralization in the deep-sea: in situ mea-
surements of calcium with a fluorescent probe Rhod-5N: surements of Ca2+, pH, pCO2 and O2. Geochim. Cos-
Influence of high ionic strength and pH. Talanta 71(1):437- mochim. Acta 65(16):2677-2690 [doi:10.1016/S0016-7037
442 [doi:10.1016/j.talanta.2006.04.011]. (01)00620-2].
Riley, J. P., and M. Tongudai. 1967. The major cation/chlorin- Whitfield, M., J. V. Leyendekkers, and J. D. Kerr. 1969. Liquid
ity ratios in seawater. Chem. Geol. 2:263-269 [doi:10.1016/ ion-exchange electrodes as end-point detectors in com-
0009-2541(67)90026-5]. pleximetric titrations: Part II. Determination of calcium
Růžička, J., E. H. Hansen, and J.C. Tjell. 1973. Selectrode—the and magnesium in the presence of sodium. Anal. Chim.
universal ion-selective electrode: Part VI. The calcium(II) Acta 45(3):399-410 [doi:10.1016/S0003-2670(01)95569-2].
selectrode employing a new ion exchanger in a nonporous Submitted 24 January 2013
membrane and a solid-state reference system. Anal. Chim. Revised 28 June 2013
Acta 67(1):155-178 [doi:10.1016/S0003-2670(01)80243-9]. Accepted 20 August 2013
Sugrue, E., P. Nesterenko, and B. Paull. 2003. Iminodiacetic

474