BRULAND. KENNETH W. Complexation by pengxiuhui


									 LIMNOLOGY                                                                                                      March 1989
            AND                                                                                                  Volume     34
                                                                                                                  Number    2
Limnol. Oceanogr., 34(2), 1989, 269-285
0 1989, by the American   Society of Limnology   and Oceanography,   Inc.

Complexation of zinc by natural organic ligands in the
central North Pacific
Kenneth W. Bruland
Institute   of Marine      Sciences, University         of California,       Santa Cruz 95064

            The complexation     of zinc by natural organic ligands in the upper 600 m of the central North
         Pacific was determined with differential pulse anodic stripping voltammetry at a thin mercury film,
         rotating glassy-carbon disk electrode. Of the dissolved zinc in surface waters >98% was bound in
         strong complexes (log Klcond,=“z+ = 11 .O) by relatively zinc-specific organic ligands existing at low
         concentrations (1.2 nM). Although the vertical distribution of this ligand class is relatively uniform,
         total dissolved zinc varies from concentrations close to 0.1 nM in surface waters of the central
         North Pacific to 3 nM at 500 m. Complexation         of dissolved zinc by this class of organic ligands
         causes the concentration of inorganic zinc to vary from 2 pM in surface waters to 2 nM at 500 m,
         a 1,OOO-fold variation. The free zinc ion activity, expressed as pZn, is calculated to be of the order
         of 12.7 in the surface waters of the central North Pacific. This low free zinc ion activity may
         influence distributions   of oceanic and neritic phytoplankton    species.

   The nutrient-type       distribution of dis-                                   and manganese) availability is an important
solved zinc in seawater indicates that it is                                      selective force acting on phytoplankton        and
removed from surface waters and trans-                                            ultimately affecting the structure of phyto-
ported to depth as a trace constituent of                                         plankton communities. Because the avail-
biogenic particles (Bruland et al. 1978; Bru-                                     ability of zinc to phytoplankton        is thought
land 1980). The low concentration of zinc                                         to be dependent on the concentration (or
in the surface photic zone (N 0.1 nmol kg-‘)                                      activity) of free zinc ions, critical to the above
led Anderson et al. (1978) to speculate that                                      arguments is the possibility of complexation
zinc is a biolimiting nutrient to phytoplank-                                     of dissolved zinc by organic ligands. Organ-
ton in the open ocean. Based on the rela-                                         ic ligands, capable of forming strong com-
tionship between zinc concentrations in sur-                                      plexes with zinc, could reduce free zinc ion
face waters and zinc requirements of marine                                       activities to extremely low levels. In addi-
phytoplankton       and their geographic distri-                                  tion, models of the interactions of dissolved
butions in the surface ocean, Brand et al.                                        zinc with suspended particles require infor-
(1983) suggested that zinc (as well as iron                                       mation on the extent of competition for free
                                                                                  zinc between dissolved ligands and binding
Acknowledgments                                                                   sites on particle surfaces (Whitfield          and
    This work was supported by ONR contract NO00 14-                              Turner 1987). Estimates of the abundance
83-K-0683. I am grateful to Kenneth Coale, Rob Franks,
and Geoff Smith for technical assistance, and to John
                                                                                  of zinc-complexing organic ligands and their
Donat, Gary Gill, Tony Michaels, and Bill Sunda for                               conditional stability constants are of partic-
critical reviews of this manuscript.                                              ular importance.
270                                          B&and

    Previous studies of zinc complexation          ing the extent and strength of zinc-organic
with anodic stripping voltammetry         to de-   ligand complexation       by DPASV involves
termine solution speciation, although li m-        titrating the ligand(s) in a sample with added
ited, have suggested that natural dissolved        zinc and measuring the oxidation current of
organic ligands can play an important role         inorganic zinc deposited in the TMF-RGCD
in the complexation of zinc in oceanic waters      electrode. This approach is based on the
(Fisher and Fabris 1982; Piotrowicz et al.         principle that an electroinactive     complex,
 1983). A study of zinc-organic complexa-          ZnL, is formed when a dissolved ligand (or
tion with a technique involving ligand com-        class of ligands), L, complexes with (elec-
petition followed by cathodic stripping vol-       troactive) zlinc:
tammetric detection suggested that natural
                                                                       Zn’ + L’ = ZnL.
zinc-organic complexes are the dominant
dissolved zinc species in a sample from the        The equilibrium   expression              for this com-
Irish Sea (van den Berg 1985). However,            plexation reaction is:
detailed information is not available on the
                                                            K’ cond,    Zn’   = [ZnL]/[Zn’][L’]
extent of zinc complexation with natural or-
ganic ligands in the open ocean.                   where [Zn’] is the total concentration of dis-
    Using differential pulse anodic stripping      solved inorganic zinc species present, [L’]
voltammetry      (DPASV) at a thin mercury         the concentration         of organic ligands that
film, rotating glassy-carbon disk (TMF-            have the potential to strongly bind zinc, and
RGCD) electrode, I here present the results        [ZnL] the concentration of the zinc-organic
of laboratory studies and held observations        ligand complex. Klcond,Zn’ is the conditional
on the complexation       of zinc with organic     stability constant (with respect to inorganic
ligands in seawater. Laboratory studies with       zinc) of the zinc-organic         ligand complex
the chelating ligand ethylenediaminetetraa-        measured under specific conditions (e.g. pH,
cetic acid (EDTA) verified the accuracy of         ionic strength, and concentrations of com-
the DPASV approach in determining the              peting cations).
extent of organically complexed zinc in sea-           [Zn’] is related to the concentration of the
water. Field studies at a station in the central   free solvated zinc ion, [Zn2+], by an inor-
North Pacific determined the vertical dis-         ganic side reaction coefficient, aZn, ([Zn’] =
tributions of the concentrations of natural        [Zn2+]cuz,,: Ringbom and Still 1972; Turner
organic ligands forming strong complexes           et al. 198 1). Theoretical estimates of aZn, in
with zinc, the conditional stability constants     surface seawater average - 2.1 (Turner et al.
of the complexes, the inorganic zinc con-           198 1; van den Berg 1985), indicating that
centrations, and the free zinc ion activities.     about 47% of the dissolved inorganic zinc
Additional information was obtained on the         exists as the free solvated Zn2+ ion and 53%
specificity of the natural organic ligands for     as complexes, principally ZnCl+. The zinc
zinc and the lability of the complexes.            species constituting Zn’ are either directly
                                                   elcctroactive (such as solvated Zn”) or the
7’heory                                            dissociation rates of the complexes are so
   The application of anodic stripping vol-        rapid, relative to their residence times in
tammetry to studies of metal speciation in         the TMF-RGCD           electrode diffusion layer,
natural waters has been promoted by scv-           that they are kinetically labile and detected
era1 workers, most notably the late H. W.          as elcctroactive (e.g. ZnCl+). [L’] includes
Nurnberg and coworkers (Nurnberg 1982;             the concentrations of free dissolved ligand,
Nurnberg and Valenta 1983), M. Branica             L”-, protonated forms (e.g. HL’- “), and li-
and coworkers (Raspor et al. 1980; Plavsic         gand that may be complexed with other cat-
et al. 1982), and I. Ruzic (Ruzic 1982, 1984).     ions (e.g. CaL). Because DPASV, as used
Voltammetric     stripping techniques such as      here, measures total inorganic zinc, the con-
DPASV at a TMF-RGCD           electrode are ex-    ditional stability constants for natural zinc-
tremely sensitive and technically relatively       organic complexes determined by DPASV
straightforward.                                   are reported with respect to Zn’ (hence de-
   The most direct approach for determin-          noted K’ cond, Zn’)* K’cond, Zn’ can be converted
                                            Oceanic zinc speciation                                                                                                    271

to a conditional stability constant with re-                 4.0         I   I   I   ,    I    I   I   I   I    I    I   I   I   I    I    I      I      I         , ,
spect to the concentration of free solvated                         -        [Ll = l.OnM                                                             /            Y'
Zn2+ (Kcond, Zn2 by multiplying Xcond, Znl by                                                                                                  / // /                  -

QJZn’-                                                       3.0 -
    Simulated DPASV titration curves for
samples containing an identical concentra-
tion of organic ligand (1 nM), but with dif-
fering Kond, Znf7can be used to demonstrate
how the extent of zinc complexation          by a
particular ligand is affected by the product
ofK’ cond, Zn’ x [L’] which equals [ZnL]/[Zn’]
(Fig. 1). Ligands with K’co,,d,Zn’ X [L’] < 10e3
do not interact with zinc appreciably, and
the titration curve obtained for a sample
containing such ligands would be indistin-                         0.0                   1.0                   2.0                   3.0                               4.0

guishable from that for a blank solution                                                               [ZnTl (nM)
containing no ligand. In contrast, the titra-               Fig. 1. Simulated titrations of a ligand concentra-
tion curves obtained for samples containing              tion of 1 nM, demonstrating the effect of varying the
ligands forming stronger complexes, with                 conditional stability constant, Klcond,7,,,.
K’ cm-ad, Zn’ x [L’] > 103, would show no de-
tectable Zn’ until L is almost completely                the stability constants that could be deter-
titrated, after which the increase in the Zn’            mined to between 10’ and 10’ I.
signal as a function of zinc added would be                  For the purpose of estimating the extent
identical to that in a blank. Samples con-               of zinc complexation with natural organic
taining ligands forming complexes of mod-                ligands, the linearization        technique de-
erate strength would exhibit titration curves            scribed above can be effectively applied t,o
intermediate between these two cases.                    the analysis of samples containing one or
    The data obtained from a zinc titration              two ligands of varying concentrations and
of a sample containing a moderate or strong              zinc complexes of varying conditional sta-
zinc-complexing           ligand can be linearly         bility constants. In a solution with a com-
transformed to allow calculation of both the             plex mixture of zinc-complexing ligands, the
total concentration of the ligand [L] and the            complexation parameters derived from this
conditional      stability constant of its zinc          treatment of the titration data may not be
complex (Ruzic 1982). For a single zinc-                 thermodynamically      correct with respect to
complexing ligand, the linearization        equa-        the actual organic ligands and natural zinc-
tion is                                                  organic complexes existing in the sample.
                                                        Nevertheless, if the data fit such a simple
     [Zn’1 _ [Zn’1 +               1                     model, determination      of these parameters
    [ZnLl Ll           cK’cond,   Zn’   X    ITI) *     allows us to describe (or conceptualize) the
                                                        zinc-complexing      characteristics of the nat-
The ratio [Zn’] : [ZnL] plotted against [Zn’]           ural ligands and their apparent influence on
for each titration point yields a straight line         zinc speciation.
with a slope of [L]-I, and intercept of                      Historically, much of the DPASV specia-
W cond, Zn’ X [L])‘. For a ligand concentra-            tion research has been performed with
tion of 1 nM (e.g. Fig. l), a window of sta-            hanging mercury drop (HMD) electrodes
bility constant values between about 1O8and             (Raspor et al. 1980). Results of metal spe-
 lOI can be determined by the DPASV ap-                 ciation research with HMD electrodes have
proach. For stability constants > 1012, only            prompted a number of criticisms of the
the ligand concentration and a lower limit              DPASV approach to studying metal-organ-
to the stability constant (dependent on the             ic ligand interactions in natural waters (Tu-
sensitivity) is obtainable. An increase of li-          schall and Brezonik 1983; Bhat et al. 198 1).
gand concentration by an order of magni-                These criticisms center on three main con-
tude (10 nM) would decrease the values of               cerns: possible underestimation         of the or-
272                                           Bruland

ganically complexed metal fraction due to            bile with respect to this DPASV approach.
a “kinetic contribution”        to the stripping     In contrast, zinc forms stronger complexes
current if the dissociation rate of the metal-       with organic ligands such as EDTA which,
organic complex is rapid with respect to the         under seawater conditions, have dissocia-
residence time of the metal-organic           com-   tion rate constants of - 1Op4s-l and are elec-
plex within the electrode diffusion layer;           trochemically inactive or inert with respect
possible underestimation       of the organically    to the TMF-RGCD           electrodes. Thus, the
complexed metal fraction due to direct elec-         diffusion layer thickness can be described
trochemical     reduction     of metal-organic       as defining the time scale of the deposition
complexes during the deposition step, which          step, and the DPASV-available          fraction of
would produce an apparent “inorganic                 complexed metal is that which can disso-
metal” signal upon stripping; and possible           ciate to the electroactive form during this
blocking of the electrode surface by adsorp-         interval.
tion of organics, thereby decreasing sensi-              Possibility of direct reduction -Problems
tivity or invalidating      the measurements.        associated with the direct electrochemical
Another concern (pertinent to any specia-            reduction of metal-organic complexes at the
tion technique) is the time required for             electrode surface can be circumvented            by
equilibration between the titrated metal and         performing the clectrodeposition        step at po-
the organic ligand(s). These concerns are            tentials negative enough to reduce only the
discussed below along with the steps taken           labile (i.e. inorganic) forms of the metal.
to eliminate, minimize, or address them.             Stronger metal-organic        ligand complexes
    Kinetic currents-The        theory for esti-     can tolerate more negative deposition po-
mating kinetic currents from the dissocia-           tentials before electroreduction of the metal
tion of labile complexes within the diffusion        in the complex occurs. An appropriate de-
layer of TMF-RGCD         electrodes is well es-     position potential for the sample of interest
tablished (Davison        1978; van Leeuwen          can be determined experimentally           by con-
 1987). A rotating disk electrode is well-de-        ducting the electrodeposition       at a series of
fined hydrodynamically;       consequently, the      potentials and plotting the measured strip-
diffusion layer thickness can be calculated          ping current as a function of the applied
directly from rotation rate. The time avail-         deposition potential to yield a “pseudo-
able for a complex to dissociate in the dif-         voltammogram”          (Figura and McDuffie
fusion layer of the TMF-RGCD             electrode    1979). There is generally a range of depo-
surface can be controlled by varying the rate        sition potentials over which reduction of the
of rotation.                                         metal strongly complexed by organic li-
    At a rotation speed of 5,000 rpm, the            gands does not occur.
TMF-RGCD         electrodes used in this study          Possible interference by adsorbed organ-
have a diffusion layer thickness of -6 x             its-The    performance of the TMF-RGCD
 low4 cm. Under these conditions a metal-            electrodes in open ocean waters does not
ligand complex with a dissociation rate con-         appear to be affected by adsorption of or-
stant, kd, ~0.1 s-l is “inert” with respect to       ganics. This problem has generally been ob-
its residence time in the diffusion layer.           served in analyzing samples from humic-
Complexes with a kd > lo5 S-I are almost             rich lakes, rivers, and estuaries with HMD
completely “labile,” and the zinc dissociat-         electrodes (Brezonik et al. 1976), but has
ing from these complexes within the diffu-           not been observed in analyzing oceanic
sion layer will provide an additional con-           samples with much larger surface area TMF-
tribution to the stripping current (corrections      RGCD electrodes (Plavsic et al. 1982).
due to differences in diffusion coefficients,           Equilibration of metal spikes with organic
cf. van Leeuwen 19 87, arc ignored here).            ligands-An assessment of the importance
 Most zinc complexes with inorganic ligands          of this concern involves determining       the
such as Cl- or weak complexes with organic           formation    rate of metal-organic    ligand
ligands such as acetate have rapid dissocia-         complexes (Ruzic and Nikolic 1982). The
tion kinetics (Pankow and Morgan 198 1)              time required for the titrated zinc to reach
and are thus electrochemically       active or la-   equilibrium with the natural organic ligands
                                  Oceanic zinc speciation                                    273

in a sample was determined experimentally        ~1 of a 5,000 ppm solution of Hg2+. The
by varying the equilibration time from sev-      mercury films were formed on the RGCD
eral minutes to more than a day.                 electrode by degassing the mercury-film for-
                                                 mation blanks with oxygen-free N, for 10
Materials and methods                            min and then depositing for 15 min at - 1.2
 Sample collection and handling -The             V with a working electrode rotation rate of
procedures used for sample collection and         5,000 rpm. After the mercury-film         forma-
filtration    have been reported previously      tion step, the TMF-RGCD          electrode rota-
(Bruland et al. 1979, 1985; Bruland 1980).       tion was stopped, and after a 30-s quiescent
Seawater samples were collected in Teflon-       period the mercury film was stripped by
coated, 30-liter Go-F10 bottles (General         scanning the potential in the positive direc-
Oceanics). The samples were filtered by          tion in the differential pulse mode (scan rate,
transferring the seawater from the Go-F10        20 mV s-l; pulse modulation amplitude, 50
bottle through Teflon tubing to Teflon filter    mV; pulse frequency, 10 s-l). If the voltam-
sandwiches (Millipore) containing 142-mm         mogram obtained from the mercury-film
diameter, 0.3~pm pore-size polycarbonate         formation blanks showed low (~0.1 nM)
filters (Nuclepore) and finally into 2-liter     levels of zinc, sample analysis would then
FEP Teflon bottles. Samples were triple          proceed. The zinc-stripping       peak occurs at
bagged, placed in buckets, and stored at         about - 1.1 V (vs. an Ag/AgCl reference
room temperature until analysis. Samples         electrode).
were normally analyzed within 18 h of col-           Sample titrations -After       mercury-film
lection, but a few were analyzed beginning       formation blank deposition, a purged sam-
 1 h after collection and up to 2 d later to     ple was used to rinse the cell electrodes. This
study storage effects. All sample manipu-        sample rinse solution was then discarded
lations and analyses, both at sea and at UC-     and a freshly purged sample introduced. The
Santa Cruz, were conducted in a positive-        sample was electrodeposited for 15 min at
pressure work area equipped with a Class-         - 1.2 V with a TMF-RGCD           electrode at a
 100 laminar flow, clean-air bench.              rotation rate of 5,000 rpm. The TMF-
     Instrumentation -The DPASV appara-          RGCD electrode was then stripped by scan-
tus has been described elsewhere (Bruland        ning its potential in the differential pulse
et al. 1985). Briefly, the system uses a mod-    mode as described earlier. The scan was
ified Princeton Applied Research (PAR) 174       stopped and held at about -0.2 V with the
A voltammetric       analyzer connected to an    TMF-RGCD         electrode rotating for at least
electrochemical cell housed in a Plexiglas       2 min between depositions to completely
cell stand. The electrochemical cell consists    strip the mercury film of residual metals.
of a 60-ml Teflon sample cup, a rotating         The 15-min deposition was repeated to ob-
glassy-carbon disk (RGCD) working elcc-          tain a second “zero addition”         voltammo-
trode onto which a thin Hg film (TMF) is         gram, ensuring that the working electrode
deposited, a platinum wire counter elec-         was conditioned for seawater analysis and
trode, and an Ag/AgCl reference electrode.       that the initial scan was reproducible.
The Plexiglas stand also accommodates ad-           Titrations were conducted by successive
ditional cell-mounting positions for sample      standard additions in the range of 0.2-0.3
and blank purging. The output from the vol-      nM zinc. After each zinc addition, the sam-
tammetric analyzer is displayed on a strip       ple was allowed to equilibrate for a mini-
chart recorder.                                  mum of 10 min before the next deposition-
    Mercury-film formation -Before a sam-        stripping cycle was repeated. As the titration
ple titration, the RGCD electrode was pol-       proceeded, zinc additions were doubled.
ished with 0.05-pm A1203 at a slow rotation      Most titrations were carried out to roughly
rate (100-500 rpm) and rinsed with dilute,       5 nM (about a fourfold excess over the cal-
quartz-distilled    HCI. Mercury-film  forma-    culated ligand concentration), although in a
tion blanks were prepared with 60 ml of          few titrations up to 30 nM of zinc was added.
Ultra-pure Milli-Q HZ0 (Millipore),     200 ~1   The sample volume was measured after
of saturated KC1 (Merck Supra-pur) and 100       completing each titration for concentration
274                                                                                                                    Bruland

                                                                                                                            within a few weeks of production.      This
          aooo                                                                                    ’       ’                 treated seawater is hereafter called UVSW.
                  t           A
          6000                                                                                                              Results
 P                t                                                                                                           Laboratory studies: Evaluation of the
                                                                          .                                                 DPAS V approach to zinc speciation -The
                                                                                                                            accuracy of the DPASV approach used here
                                                                                                                            was first evaluated by titrating the well-
                                                                                                                            studied model ligand EDTA; 9.5 nM EDTA
                                                                                                                            and various concentrations        of zinc were
                                                                                                                            added to IJVSW aliquots and allowed to
                                                                                                                            equilibrate for 24 h. The data (Fig. 2) fit a
                                                                                                                            single metal-single      ligand model (Ruzic
                  0                           IO                                                 30           40             1982, 1984), and the experimentally      deter-
                                                                                                                            mined values for the EDTA concentration
                                                                                                                            (10.0 nM) and conditional stability constant
            3.2                                                                                     //’            -        for the ZnEDTA2- complex (log K’cond,zn, =
                               B                                                                 //                         7.9) agree closely with the actual concen-
            2.8                                                                           . /a                              tration of EDTA added (9.5 nM) and the
                                                                              l       /

            2.4                                               .
                                                                      /                                                     predicted conditional stability constant (log
                                                          /                                                                 K’ cond. Zn’ = 7.6). The predicted conditional
                                                                                                                            stability constant for ZnEDTA2- in seawa-
     c!                                        /
     p      1.6
                                          /                                                                                 ter was calculated from the thermodynamic
     c!                        /=                                                                                           stability constant (Martell and Smith 1974)
            1.2          4                                                                                                  after correction to an ionic strength of 0.7



            0.0   I- / IA
                                  (ni$    520

   Fig. 2. A. Zinc titration of 9.5 nM EDTA in UVSW
                                                   / 7
                                                                                                                            with free ion activity coefficients obtained
                                                                                                                            with the Davies equation. A side-reaction
                                                                                                                            coefficient for EDTA (aEDTA) of 108.0 and
                                                                                                                            an inorganic side-reaction coefficient for
                                                                                                                            zinc (cuzn) of 2.1 were used to convert to
                                                                                                                            K’ cond.   Zn’-
                                                                                                                                A solution of 0.1 M KC1 containing 9.5
(estimated log Klcond,Zn, = 7.6). B. Linear transforma-                                                                     nM EDTA was then titrated with zinc to
tion of the EDTA titration data yielding ligand con-
centration of 10.0 nM and log Klcond,Zn~ 7.9.
                                           =                                                                                evaluate the accuracy of the DPASV ap-
                                                                                                                            proach in analyzing a sample containing a
                                                                                                                            ligand with a much higher conditional sta-
determinations.                                     Throughout the course of                                                                             =
                                                                                                                            bility constant (K’cond,Zn, 1013.8).The con-
these titrations,                                  pH stayed constant (-to. 15                                              ditional stability constant of the ZnEDTA2’
pH units).                                                                                                                  complex is much greater in 0.1 M KC1 be-
     Total dissolved zinc determinations-To-                                                                                cause the absence of Ca2+ and Mg2+ greatly
tal dissolved zinc [Zn-,.] was determined with                                                                              decreases the side-reaction coefficient for
graphite furnace atomic absorption spec-                                                                                    EDTA. In this case, the product of the con-
trometry (GFAAS) (Bruland et al. 1979,                                                                                      ditional stability constant and the ligand
 1985; Bruland 1980).                                                                                                       concentration is 6 x 105. Therefore, no Zn’
    Preparation of U VS W- Seawater con-                                                                                    should be detected until the EDTA is almost
taining nondetectable levels of trace metals                                                                                completely titrated. Experimental titration
and metal-complexing         dissolved organic                                                                              data verified this prediction (Fig. 3A). ‘The
matter was prepared for use in model ligand                                                                                 linear transformation      of the titration data
titrations and other experiments by an ul-                                                                                  (Fig. 3B) yielded an EDTA concentration of
traviolet photo-oxidation     and resin column                                                                              9.9 nM and a lower       limit for K’cond, Znt of
system described in detail by Donat and                                                                                      10”. That no Zn’ (co.01 nM) was detected
Bruland ( 1988). The seawater was collected                                                                                 for additions of zinc less than the ligand
in 2-liter FEP Teflon bottles and was used                                                                                  concentration     is consistent with the Zn-
                                    Oceanic zinc speciation                                                 275

EDTA2- complex being kinetically inert and
electroinactive with respect to the DPASV
approach used.
   Field studies-Most of the samples for
this study were collected in the central North
Pacific at the VERTEX-IV      station (-28”N,
 158”W) during a 2-week occupation in July
 1983. A few additional samples were col-
lccted on the VERTEX-V           cruise during
June-July 1984 from a station at the edge
of the central North Pacific gyre (33”N,
 139”W). Similar physical and chemical
oceanographic conditions existed at both
stations. The depth of the mixed layer was                      0       10                         30       40

25 m and the base of the euphotic zone and                                        [Zn,;4nM)

nutrient-depleted    layer extended to about
 150 m. This shallow mixed-layer depth is
typical of summer stratification,      whereas
winter mixing extends the layer to 150 m
(Bathen 1972).
  Vertical prqjile of total dissolved zinc con-
centrations-The     vertical profile of total
dissolved zinc (Fig. 4A) is consistent with
previous data from this oceanographic re-
gime (Bruland 1980), and the correspon-
dence between the vertical profiles of dis-
solved zinc and silicic acid (Fig. 4B) supports
the oceanographic consistency of the total
dissolved zinc results (Bruland 1980, 1983).              0.0   k
                                                                    I         I           I    I        I

                                                                0            IO
   DPASV zinc titrations- DPASV zinc ti-                                                      20            30
                                                                                [Zn’1 hM)
trations were conducted on samples from
                                                      Fig. 3. A. Zinc titration of 9.5 nM EDTA in 0.1 M
selected depths in the upper 600 m of the          KC1 (estimated log Kcond Znz= 13.8). B. Linear trans-
water column at VERTEX-IV.              The dis-   formation of the EDTA iitration data yielding ligand
solved zinc initially present in a typical sam-    concentration of 9.9 nM and log Krcond,Znf > 11.
ple from 90 m was 0.35 nM (Fig. 5A, Table
 1). For total zinc concentrations ~0.5 nM,
no electroactive zinc was detected (i.e. [Zn’]     tions (Ruzic and Nikolic 1982). The trans-
 < 0.01 nM). Thus, at least 98% of the zinc        formed data give a zinc-complexing            li-
(for [Zn,.] < 0.5 nM) was bound by organic         gand concentration of 0.94 nM and a log
ligands in complexes that were electrochem-        K’ cond, Zn’ of 10.6.
ically inert. For zinc additions yielding total       With this data set it is impossible to assess
concentrations of 0.5-l .5 nM, the Zn’ cur-        whether there may have been another class
rent responses initially form a curve that         of ligands which form weaker zinc com-
approaches a straight line at total zinc con-      plexes that might slightly influence zinc
centrations > 1.5 nM. The slope of the lin-        complexation in the sample (e.g. 100 nM of
ear Zn’ response at the higher additions was       a ligand with a Ktcond, Zn, of 106). Such a
used to calculate [Zn’].                           ligand class could not account, however, for
   The plot of [Zn’]/[ZnL]     vs. [Zn’] for the   more than 10 or 20% of additional organic
90-m sample was linear, indicating that the        zinc complexation, since the linear Zn’ re-
data can be simply modeled by assuming             sponse after titrating a sample with 4 nM
formation of a one-metal, one-ligand com-          zinc was 80-90% of the response of the same
plex (Fig. 5B) and that an equilibration time      electrode film to Zn’ in UVSW.
of 10 min was adequate for the zinc addi-             All of the sample titrations performed in
276                                                                                    B&and

                                                      [Zn,] (nM)                                                                                         lIH4SiOJ (PM)
             0.0                       1.0           2.0           3.0       4.0        5.0                                        10           20       30           40       50        60      70               80
             o-                I         I    I        I       I     I   I     I   1                      o"               I        I       I    I   I    I     I      I   I    I    I    II      I       I
                           .                                                                                   .
                  -. .                                                             A     -                      .
                                                                                                                   .                                                                                  B           -
       100-**                                                                                       100 -0
                           .                                                                               .
                  -.                                                                                    -.
                       .                                                                                               .
       200        -    l                                                                            200-                       .

   :300-                           l                                                     -     2    300        -                        l

   F                                                                                           F

   cl 400-                                                                               -
                                                                                               Q 400-                                                    .

       500 -                                                                                        500        -

       600        -                                                                    . -          600        -                                                                                              c

       700                     "              "            '        '    '    "                     7OO.-L-.L-LI-L-L-L-LI-I-I--LI                                                                 I       1

       Fig. 4.                     Vertical       profiles of (A) total dissolved            zinc and (B) silicic acid at the VERTEX-IV                                                        station.

this study fit a simple one-metal, one-ligand                                                      of 4 pM, which is only 2% of [ZnT]. The
model. Within the surface 150 m of the                                                             average concentration      of free zinc ion,
VERTEX-IV           station, the average concen-                                                   [Zn’+], is then estimated to be 1.8 pM. If
tration     of ligands forming        strong zinc                                                  we assume an activity coefficient for Zn2+
complexes was 1.2k0.2 nM (Fig. 6A) and                                                             of ~0.2, the zinc ion activity ({Zn’+}) in the
the average value for the log K’cond,ZnP       was                                                 upper 200 m averages 0.4 pM, which is
 10.7kO.3 (Fig. 6B). Little or no vertical                                                         equivalent to a pZn value of 12.4. At VER-
structure was apparent for either quantity                                                         TEX-IV,    [Zn’] and {Zn2+} exhibit a dra-
(Table 2).                                                                                         matic increase of three orders of magnitude
     In the titrations of samples from 400 to                                                      from the sutiace to intermediate depths (600
600 m, electroactive Zn’ was detected be-                                                          m), varying between 4 and 4,000 pM and
fore any zinc was added. The 400-m sample                                                          0.4 and 400 pM, respectively.
was the deepest for which the ligand con-
centration and conditional          stability con-                                                    Table 1. Data from the 90-m sample (VERTEX-
stant could be calculated with confidence.                                                         IV). A sensitivity of 184 nA nM-I was used lo calculate
                                                                                                   [Zn’]. The Ruzic’s linearization    yielded a slope of IW
At depths >400 m, the increasing concen-                                                           = 1.06 and an intercept of h = 0.026, thus resulting in
trations of dissolved zinc had completely                                                          model-derived                                        Zna
                                                                                                                     values of [L] = 0.94 and log JCcond, =
titrated the ligands present and conditional                                                        10.6.
stability constants could not be estimated.                                                                                                             --
The concentration of organically complexed                                                             P&l                              current                P-4                  [ZnL1
zinc in these deeper samples appeared to be                                                            WV                                @A)                   (nM)                  b-W               rZnLl

 l-l.5 nM.                                                                                             0.35                                   0               10.01                 0.35              ~0.029
     Since no detectable Zn’ was observed at                                                           0.62                                   4                 0.02                0.60                0.033
depths shallower than 200 m (Fig. 7), the                                                              0.88                                  24                 0.13                0.75                0.17
                                                                                                       1.15                                  56                 0.30                0.85                0.35
model-derived values for [L’] and Ktcond,Znj                                                           1.42                                  88                 0.48                0.94                0.5 1
and the total dissolved zinc concentration                                                             1.69                                 140                 0.76                0.93                0.82
([Zn,]) originally present in the sample were                                                          1.95                                 192                 1.04                0.91                1.14
used to calculate [Zn’] (e.g. [Zn,] = 0.2 nM,                                                          2.49                                 292                 1.59                0.90                1.77
                                                                                                       3.02                                 384                 2.09                0.93                2.25
 [L] = 1.2 nM, [L’] = 1.0 nM, K’cond, =       Znl                                                      3.56                                 480                 2.6 1               0.95                2.75
 1010.7). These average values yield a [Zn’]
                                                      Oceanic zinc speciation                                                                         277

                                                                                                             Ll (nMI
                                                                                       0,                                              I
                            [Zn,] (nM)
                                                                                                   I    I          I
                                                                                                                       . I                        I
                                                                                     100                             .*
                                                           P                                                       .
                                                                                        t                            .
                                                                                                                       .                         1
             i                                                                 E

                                                                                            1                      .

       2.0   -                                /                                %     300                       .
  T                                     /
  L!         -                      /
  A          -                  /                                                    400                       .
  5          -
       1.0 -          //’                                      7                     500 l--I
                                                                                         9             IO                    11                  12
                                                                                                       log    K’cond,      Zn’
                                                                       Fig. 6. A. Vertical distribution     of the strong zinc-
                                                                    complexing ligand, [L]. The line represents the con-
                                                                    centration of total dissolved zinc, [Zn.,]. B. Vertical
                                                                    distribution of the conditional     stability constant, log
                             b-Cl (nM)                              K' cond,   Zn”
    Fig. 5. A. Zinc titration of a typical sample from
90 m (VERTEX-IV).        The slope of the line is the sen-
sitivity of the electrode. B. Linear transformation     of
the 90-m titration data indicating a ligand concentra-
tion of 0.94 nM and log Klcond,ZnS 10.6.
                                                                       Table 2. Results from the VERTEX-IV         station.
                                                                    Values of [ZN’] at depths of 300 m and deeper were
    Methodology tests-Analyses      of filtered                     measured voltammetrically;  shallower values are cal-
and unfiltered aliquots of the same surface                         culated with model-derived values of [L] and Klcond,Zne.
sample indicate that the bulk of zinc-com-                            Sample
plexing ligands occurs in the dissolved frac-                          dcplh                L&l         Ll                                      [Zn’1
                                                                        (ml                 WV         WV              log I’d    Zn            (nMI
tion (~0.3 pm; Table 3). The zinc naturally
present in the surface seawater of the central                          22                  0.30       0.98                10.7                (0.008)
                                                                        50                  0.15       1.34                10.8                (0.002)
North Pacific is typically 80-90% dissolved,                            50                  0.30       1.33                10.9                (0.004)
with lo-20% in the particulate fraction                                 90                  0.35       0.94                10.6                (0.014)
(Bruland unpubl. data).                                                100                  0.17       1.37                10.7                (0.003)
    No significant differences were observed                           125                  0.20       1.12                10.5                (0.007)
in ligand concentrations or conditional sta-                           150                  0.15       0.92                10.6                (0.005)
                                                                       175                  0.20       1.31                10.6                (0.004)
bility constants between samples titrated                              200                  0.23       1.07                10.5                (0.008)
immediately after collection and tempera-                              300                  0.65       1.05                10.4                 0.033
ture equilibration  (1 h) and those whose ti-                          400                  1.62       1.3                 10.4                 0.44
tration began 2 1, 23, or 35 h later (Table                            500                  3.30        -                                       1.9
                                                                       600                  4.77        -                                       3.5
3). Thus, the zinc-complexing    character of
278                                                                                         B&and
                                                     Zinc       (nM)                                                        log Zinc   (nM)
               0             I         I
                                     1.0   I         I
                                                   2.0      I         ,
                                                                    3.0       ,   4;o   ,   5io    013        .   -12        -11        -10         -9             -6
                                                                                                          I         I   I      I   I      ,
                                                                                                         . .
                                                                                                           . .

      2300'                      +

      a 400-             l

         500    -                              .

         600 -                                                            .

   Fig. 7. A. Vertical distribution of total dissolved zinc (+) and inorganic                                               zinc (0). B. Vertical        distribution   of
log [Zn,] (+) and log{Zn2+} (0) ({Zn2+} is the free zinc ion activity).

the ligands appears to be stable for days. In                                                     exists between -- 1.2 and - I .3    V in which
addition, no significant change in ligand                                                         little or no electroreduction       of the Zn-
concentrations or conditional stability con-                                                      EDTA2- complex was apparent.         Most of the
stants was observed when zinc additions                                                           titrations were therefore carried   out at - 1.2
were allowed to equilibrate for 10 min or                                                         V-the most positive potential       possible.
24 h. This observation provides a lower lim-                                                          Interferences by other metals- Prelimi -
it to the zinc-ligand association kinetics.                                                       nary studies on possible interferences in zinc
   Samples from 50-m depth were collected                                                         titrations from the presence of copper or
at various times during the day and night                                                         cadmium were performed. Doubling the
and titrated with zinc to assess the influence                                                    concentration of total copper initially pres-
of diurnal variability on zinc complexation                                                       ent in a surface sample (0.6 nM) yielded a
(Table 3). No significant differences in li-                                                      copper concentration       close to that of the
gand concentrations or conditional stability                                                      zinc-binding ligands (1.2 nM), but this level
constants were observed, supporting the rel-                                                      of copper did not significantly affect the zinc
ative stability of the zinc-complexing     char-                                                  titrations. Coale and Bruland ( 1988) have
acter of the ligands on time scales of days.                                                      shown that two classes of organic ligands
   Pscudovoltammograms       were obtained on                                                     strongly complex >99.7% of the total dis-
both UVSW and seawater samples from                                                               solved copper in surface waters of the cast-
 100-m depth. Both solutions were adjusted                                                        ern North Pacific. The stronger ligand class
to contain a total zinc concentration of - 0.5                                                    has a concentration of - 1.8 nM and dom-
nM. In the UVSW, all of this was Zn’, while                                                       inates copper complexation           in surface
in the natural sample 98% of the zinc ap-                                                         waters. If sufficient copper was added (>2
peared complexed with organic ligands. At                                                         nM) to completely complex the strong cop-
deposition potentials between - 1.2 and                                                           per-binding ligands, a copper-stripping peak
 - 1.3 V, the Zn’-stripping   peak currents in                                                    was apparent, and Zn’ sensitivity was se-
both types of solutions were constant, but                                                        verely depressed. This depression of Zn’
only just detectable in the natural sample                                                        sensitivity was due to formation of a Cu-Zn
with magnitudes about 2% of those ob-                                                             intermetallic compound in the mercury film
served in UVSW. Additional           studies of                                                   (Shuman and Woodward 1976).
UVSW containing ZnEDTA2 - also indicat-                                                               During the stripping step following each
ed that a window of deposition potentials                                                         zinc addition in a titration, the potential was
                                           Oceanic zinc speciation                                           279

  Table 3. Results from diurnal, storage, and filtration studies. The variability in any of these experiments   is
not considered to bc significant with respect to the experimental precision of the method.

                         Nom                                     L&l (nM)          tL1W)           1%Rd. Zn’
      Diurnal study, 50 m (VERTEX-IV)
        Time of collection
           2230 1 Aug                                              0.15             1.43              10.6
           0400 2 Aug                                              0.30             1.34              10.8
           1100 2Aug                                               0.30             1.33              10.9
           1715 3 Aug                                              0.15             1.48              10.9
      Storage, 50 m (VERTEX-IV)
        Time between collection and start of analysis (h)
            1                                                      0.30             1.33              10.9
           21                                                      0.30             1.33              10.6
           35                                                      0.30             1.23              10.7
      Filtration, 50 m (VERTEX-IV)
         Filtered                                                  0.15              1.28             11.0
         Unfiltered                                                0.15              1.48             10.9
         Unfiltered                                                0.15              1.20             11.2
      Filtration and storage, 130 m (VERTEX-V)
         Collected 0930 30 Jan 84
            Filtered (14 kPa)
               Analyzed 1040 30 Jan 84                             0.34              1.33             11.0
               Analyzed 0900    1 Jul 84                           0.34              1.41             10.7
          Filtered (55 kPa)
             Analyzed 1500     30 Jun 84                           0.34              1.45             10.8
             Analyzed   1040   30 Jan 84                           0.34              1.31             10.6
             Analyzed   1500   30 Jan 84                           0.34              1.41             10.8
             Analyzed   0900    1 Jul 84                           0.34              1.43             11.0

scanned through the region where any elec-                  amount of inadvertent contamination could
troactive copper deposited in the TMF-                      lead to a significant underestimate of the
RGCD electrode would be oxidized, pro-                      concentration of ligands and the strength of
ducing a stripping current. No electroactive                their interaction with zinc. It is imperative
copper was apparent in samples from depths                  that concentrations of total dissolved zinc
 ~500 m at this station. Thus, there seemed                 be determined in the samples used to de-
to be no interference from copper via in-                   termine metal complexation and compared
tcrmetallic compound formation with zinc                    with published data sets exhibiting ocean-
at surface and intermediate depths. Inter-                  ographic consistency (e.g. Bruland 1980) in
ference from levels of copper of the order                  order to establish that the samples were not
of 1 nM was not expected because of the                     contaminated      at the outset. The vertical
strong specificity observed by Coale and                    profile of total dissolved zinc concentra-
Bruland (1988) of the copper-binding        li-             tions at this VERTEX-IV         station (Fig. 4)
gands for copper. Finally, additions of cad-                agrees well with that of a previous report
mium at concentrations of 1 nM did not                      (Bruland 1980) and is consistent with the
affect the zinc titrations either.                          Zn-Si relationship previously reported for
                                                            this area.
Discussion                                                      The model-derived      values for the con-
  The ability to collect, process, and store                ditional stability constants of natural zinc-
seawater samples without contamination     is               organic complexes and concentrations of the
an absolute prerequisite for determining                    zinc-complexing      ligand show a fairly uni-
metal complexation.      Even a very small                  form distribution in the upper 400 m at this
280                                               B&and

station. In contrast, Coale and Bruland                   surface waters of the central North Pacific
(1988) observed that the concentration of                 as well; as stated earlier, however, they
the ligand primarily responsible for strongly             should not substantially influence zinc com-
complexing copper in surface waters of the                plexation. Instead, the data in this report
eastern North Pacific showed a substantial                demonstrate that organic ligands existing in
vertical gradient and essentially was below               low nanomolar concentrations and forming
detection at depths >200 m. The maximal                   strong zinc complexes effectively control the
concentration       of the strong coppcr-com-             speciation of zinc in surface waters of the
plexing ligand occurred within the euphotic               central North Pacific.
zone at depths between 60 and 80 m.                           Copper is the only other metal for which
    The only other open-ocean data set of                 there is convincing evidence of the domi-
zinc complexation having oceanographical-                 nant role of organic complexation        in spe-
ly consistent concentrations            of total dis-     ciation in oceanic surface waters. Although
solved zinc with which to compare these                   comprehensive comparisons have not yet
present results is that obtained by Donat                 been performed,        several different     ap-
and Bruland (1989) in a comparison of                     proaches suggest that within surface waters
DPASV and competitive ligand equilibra-                    >99% of the Cu2+ exists complexed with
tion-differential      pulse cathodic stripping           organic ligands (DPASV: Kramer 1986;
voltammetry (CLE-DPCSV). At a station in                  Coale and Bruland 1988; CLE-DPCSV: van
the Northeast Pacific (VERTEX-VII,                 sta-   den Berg 1984; MnO, adsorption: van den
tion T-5, 39”36’N, 140”46’W), they ob-                    Berg 1982; ligand competition-solvent        ex-
served that > 95% of the total dissolved zinc             traction: Moffet and Zika 1987; ligand com-
at 60- and 150-m depths was bound in                      petition-solid   phase extraction: Sunda and
strong organic complexes [log rcond, Zn2+ =               Hanson 1987; bioassay: Sunda and Fergu-
 11.2kO.2 (DPASV) and 10.3kO.2 (CLE-                      son 1983).
DPCSV)] by a low concentration (-2 nM)                        The findings reported here are markedly
of zinc-complexing         ligands. Thus, the zinc        different from observations on the zinc-
speciation results reported here are in good              complexing characteristics of humic-type
agreement with those obtained by Donat                    substances generally referenced in the lit-
and Bruland (1989), although the ligand                   erature (e.g. Mantoura et al. 1978) (Fig. 8).
concentration at the VERTEX-VII                station    The concentration of humic substances es-
in the Northeast Pacific (IV 1,000 nm north-              timated by Mantoura et al. (1978) is three
east ofthe VERTEX-IV            station studied here)     orders of magnitude larger than the conccn-
is slightly higher.                                       tration of the zinc-complexing      ligand ob-
    Van den Berg (1985) has reported the ex-              served in this study, but the conditional sta-
tent of zinc complexation for a surface water             bility constant of the zinc-humic      complex
sample from the Irish Sea. His data best fit              was almost six orders of magnitude less.
an equilibrium        model for complexation of           Mantoura et al. (1978) argued that humics,
zinc by two organic ligands having concen-                and dissolved organics in general, were not
trations of 26 and 62 nM and forming zinc                 expected to substantially influence the spe-
complexes having values for log Ktcond,Zn~+               ciation of zinc in seawater.
of 8.4 and 7.5. The total dissolved zinc con-                 Inputs of zinc from both the atmosphere
centration he found in his sample was 28                  and vertical mixing from intermediate
nM. Van den Berg’s observations from the                  depths have been estimated to supply sig-
Irish Sea stand in marked contrast to the                 nificant zinc to surface waters (Bruland
average ligand concentrations (1.2 nM), av-                1980). For example, elevated dissolved zinc
erage conditional         stability constants (log        concentrations in near-surface waters in the
K’ cond, Zn2 I- = 11 .O), and initial zinc concen-        central Pacific have been measured after a
trations (0.2 nM) estimated in the present                rain event (Bruland 1980). Such inputs are
 study for the upper 200 m of the VERTEX-                  sporadic and would tend to supply zinc in
 IV station.                                              pulses, potentially resulting in variable free
    There may be classes of organic ligands               zinc ion activities within the surface layers.
 able to form weak zinc complexes in the                  Within the upper 200 m of the central North
                                      Oceanic zinc speciation                                           281

Pacific, however, there is an average of 1             10'2

nM of strong zinc-complexing        ligand in ex-
cess of the dissolved zinc concentration. The
 1 nM of excess concentration of ligand could
readily complex zinc introduced to the sur-
face layer from such inputs, thereby strongly          10'0

buffering free zinc ion activity at low levels.
   At depths >300 m, Zn’ concentrations
become measurable because dissolved zinc
concentrations increase with depth to levels
exceeding the zinc-complexing          capacity of
the organic ligands (Fig. 5A). In all of the
deep-water samples analyzed, dissolved zinc
existed predominantly      as electroactive, in-
organic Zn’. If the zinc-complexing         ligand
exists in deep waters at concentrations com-            IO6

parable to those found in surface waters, 80-
90% of the zinc would occur as inorganic
species, due to the large excess of dissolved
zinc over the complexing organic ligand.
    Zinc-organic complex lability- Previous             IO4
                                                          10-10          1O-8             10-6           10-4
studies of zinc complexation in the Gulf of
                                                                                D-1 (Ml
Mexico (Piotrowicz et al. 1983) have indi-
                                                        Fig. 8. A log-log plot showing the percent com-
cated that zinc-organic complexes were ex-           plexation for a given concentration   of zinc-complcx-
tremely labile on time scales of hours to            ing ligand, [L], and its conditional stability constant,
days. Because analysis normally begins be-           K’cond.Zn'.
tween 1 and 18 h after collection and a sam-
ple titration in the present study took up to         valent cations in seawater-do     not signifi-
8 h to complete, complex lability and the             cantly interact with the natural zinc-com-
potential of ligand degradation were of great         plexing ligands, and therefore have little
concern. In marked contrast to the results            effect on zinc complexation kinetics.
of Piotrowicz et al. (1983), however, the re-            In contrast, the metal complexation      ki-
sults of the time series experiments reported         netics of the model ligand EDTA are strong-
here indicate that both the concentration of          ly affected by Ca2+ and Mg2-t. Consequently,
zinc-complexing     ligands and the condition-        in seawater EDTA has a low formation rate
al stability constants of their zinc complexes       constant with zinc, of the order of 102-lo3
did not change significantly over a period           M-l s-l (Raspor et al. 1980). The apparent
of days. Sporadic zinc contamination         could   lack of interaction between Ca2+ and Mg2+
have caused the variation in results reported        and the natural zinc-complexing     ligands is
by Piotrowicz et al. (1983), prompting their          suggestive of the presence of donor atoms
conclusion regarding the apparent lability           such as sulfur and nitrogen in the natural
of the zinc-organic complexes.                       ligands rather than oxygen- the donor atom
  Zinc-organic ligand equilibration time             preferred by Ca2+ and Mg2+. A formation
and specijcity-The  1O-min period within             rate constant of the order of lo7 M-l s-l,
which zinc additions seemed to equilibrate           combined with a conditional stability con-
completely with the natural organic ligands          stant of lOI’, would imply an upper limit
provides a lower limit to the kinetics of            to the dissociation rate constant for the ZnL
complex formation. Since there is - 1 nM             complex of - 1O-4 s-l. This dissociation rate
of excess ligand, its formation rate constant        constant is consistent with the assumed
with zinc would have to be at least 7 x lo6          electrochemical inertness of the zinc-organ-
M-’ s-l in order to equilibrate fully in 10          ic ligand complex with respect to the dif-
min. This high formation        rate constant        fusion layer of our TMF-RGCD         electrode
suggests that Ca2+ and Mg2 ‘-the major di-           during the deposition step.
 282                                           Bruland

   The zinc-complexing     ligands showed at            nM;    log K’cond, Zn’ = 8.1 and 7.2, respective-
least some selectivity for zinc. Amounts of             ly) is indicative of neritic waters, then free
copper and cadmium roughly equivalent to               zinc ion activities in neritic waters could be
the concentration of the zinc-complexing    li-         one to three orders of magnitude greater
gands could be added to a sample before the            than in the central North Pacific. If, how-
zinc titration without significantly changing          ever, the ligand concentrations          in neritic
the results.                                           waters are similar to those observed in this
   Influence of zinc speciation on biota -The          study for oceanic waters, then the amount
  response of aquatic organisms to zinc con-           of total dissolved zinc present in neritic
  centrations in their external environment is         waters (on the order of a few nM) would
 generally thought to reflect changes in free          exceed the ligand concentration,          and free
 Zn2+ ion activity rather than those of the            zinc activities would be even greater than
 total dissolved zinc concentration (Ander-            in the former case.
 son et al. 1978). Vertical profiles of total               At these higher free zinc ion activities,
 dissolved zinc concentrations in the central          the neritic phytoplankton species studied by
 North Pacific generally exhibit concentra-            Brand et al. (1983) reproduced at maximal
 tions varying from -0.1 nM in surface                rates and exhibited no evidence of zinc lim-
 waters to - 3 nM at 500 m-a 30-fold vari-            itation (or zinc toxicity). Thus, the zinc spe-
 ation. Taking account of organic complex-            ciation data reported here are consistent with
 ation, the inorganic      zinc concentration         the hypothesis of Brand et al. (1983) that
 ([Zn’]) would vary from 2 pM to 2 nM -a              zinc, as a biolimiting        nutrient, may influ-
  1,OOO-fold variation. Fret zinc ion activity,       ence distributions of phytoplankton          species
 expressed aspZn, would be - 12.7 in surface          in the oceans. If the extent of organic com-
 waters and 9.7 at 500 m. This dramatic dif-          plexation of zinc observed here is represcn-
 ference of three orders of magnitude in free         tative of the open sea, then low free zinc ion
 zinc ion activity between surface waters and         activities may not limit primary production
 500 m is similar to that observed for free           in oceanic waters. Free zinc ion activity may
 cupric ion (Coale and Bruland 1988) in the          play an important role, however, in influ-
Northeast Pacific.                                   encing the species composition and ecology
    Brand et al. (1983) using single phyto-           of phytoplankton         populations.
plankton species cultured in media buffered                The results of this study raise important
with EDTA over a range of pZn from 10.5              questions. For example, what is the nature
to 13.2, showed that reproductive rates of           of these strong zinc-complcxing          organic li-
eucaryotic algae and cyanobacteria arc dif-          gands that seem to show a remarkable de-
ferentially affected by free zinc ion activities     gree of specificity? How stable are the li-
in the range reported here for the upper 200         gands? What are the sources and sinks of
m of the central North Pacific. Phytoplank-          these ligands? How are they spatially and
ton species with primarily neritic distribu-         temporally distributed? A number of inter-
tions generally exhibited substantially        re-   esting scenarios or hypotheses can be for-
duced growth rates at pZn values of 12.7-            mulated. The scavenging of metal cations
the value representative of central North            by particles from surface waters has gen-
Pacific surface waters. Species with primar-         erally been modeled as a competition              for
ily oceanic distributions     did not show evi-      the metal cation by surface adsorption sites
dence of zinc limitation at pZn values < 13,         on particles. Perhaps the binding of zinc by
however, and thus would not be zinc limited          dissolved organic by
in the central North Pacific.
    Neritic surface waters have substantially
higher concentrations of total dissolved zinc
than oceanic surface waters (Bruland and
Franks 1983). If the decreased degree of or-
ganic complexation of zinc (relative to the
present study) observed by van den Berg
(1985) for the Irish Sea ([L] = 26 and 62
                                     Oceanic zinc speciation                                            283

 lating zinc at extremely low levels to have        havior with respect to particle interactions,
 a competitive advantage over neritic species       particularly within the surface layers of the
 in oceanic regimes. A shortcoming of this          ocean where complexation        of zinc by or-
 hypothesis is that it appears to require group     ganic ligands may reduce scavenging of zinc
 selection because the individual        that se-   by particles.
 cretes the ligand is unlikely to benefit from          Because the results reported here on the
 its own secretions.                                organic complexation      of zinc, along with
     Another possibility requiring investiga-       similar findings for copper, have been ob-
 tion is that oceanic species of phytoplankton      tained only recently, a need exists to verify
 might be able to assimilate and utilize or-        these results with independent methods. One
ganically bound zinc. The current dogma             such independent method showing promise
 that biological response to metals such as         for zinc speciation studies is cathodic strip-
zinc reflects changes in free metal ion activ-      ping voltammetry with competitive ligand
 ities is based primarily on studies that used      equilibration   (CLE-DPCSV) (van den Berg
 the synthetic chelators EDTA and NTA to             1985). A comparison of the DPASV ap-
 set free metal ion activities specified by the     proach prcsentcd here with the CLE-DPCSV
 experimenters.    Perhaps, as has been ob-         approach for zinc has recently been under-
 served with iron-siderophores,       there are     taken on a set of samples from the North
 oceanic species of phytoplankton       that can    Pacific, and good agreement has been dcm-
 assimilate organically complexed zinc, thus        onstrated for both ligand concentrations and
gaining a competitive advantage over species        conditional stability constants (Donat and
that cannot. If organisms actually can as-          Bruland 1989). We also have initiated stud-
 similate the zinc-organic     ligand complex,      ies on the spatial and temporal variability
 such biological utilization could bc the ma-       of zinc complexation within the North Pa-
jor process by which the ligand is consumed.        cific to further elucidate the biogeochem-
Perhaps a dynamic balance exists for this           istry of zinc.
class of ligands between its production and
consumption, resulting in the relatively uni-
form and ubiquitous presence of the zinc-           References
complexing ligand in the central North and           ANDERSON, M. A., F. M. M. MOREL, AND R. R. L.
Northeast Paci fit.                                      GUILLARD. 1978. Growth limitation of a coastal
                                                         diatom by low zinc ion activity. Nature 276: 70-
Conclusions                                              71.
                                                     BATHEN, K. H. 1972. On the seasonal changes in the
   DPASV methods with a TMF-RGCD                         depth ofthe mixed layer in the North Pacific Ocean.
electrode show great promise for studies of              J. Geophys. Res. 77: 7 138-7 150.
metal speciation in natural waters. The spe-         BHAT, G. A., R. A. SAAR, R. B. SMART, AND J. H.
ciation results obtained for zinc in this study          WEBER. 198 1. Titration of soil-derived fulvic acid
                                                         by coppcr(II) and measurement of free copper(H)
demonstrate the overwhelming importance                  by anodic stripping voltammetry      and copper(B)
of organic complexation      to the speciation           selective elcctrodc. Anal. Chem. 53: 2275-2280.
of zinc. This observation is in marked con-          BRAND, L. E., W.G. SUNDA,AND R. R. L. GUILLARD.
trast with most previous speculations which              1983. Limitation     of marinc phytoplankton     rc-
suggested that organic complexation would                productive rates by zinc, manganese, and iron.
                                                         Limnol. Oceanogr. 28: 1182-l 195.
be of only minor importance to the specia-          BREZONIK, P.L.,P.A. BRAUNER,AND W. STUMM. 1976.
tion of zinc in seawater. In intermediate and           Tract metal analysis by anodic stripping voltam-
deep waters, however, organic complexa-                 metry: Effect of sorption by natural and model
tion does not play a dominant role, and zinc            compounds. Water Res. 10: 605-6 12.
                                                    BRULAND, K. W. 1980. Oceanographic distributions
exists primarily as the free solvated ion and           of cadmium, zinc, nickel and copper in the North
in inorganic complexes (e.g. ZnCl’).                    Pacific. Earth Planet. Sci. Lett. 47: 176-198.
   Zinc-complexing    organic ligands strongly      -          1983. Trace elcmcnts in seawater, p. 157-
buffer free zinc ion activity in surface sea-           22b. In J. P. Riley and R. Chester [eds.], Chemical
water with respect to perturbations in total            oceanography. V. 8. Academic.
                                                    -        K. H. COALE, AND L. MART. 1985. Analysis
dissolved zinc. The observed speciation of              of’seawater for dissolved cadmium, copper and
zinc should influence its geochemical be-               lead: An inter-comparison     of voltammetric   and
284                                                      B&and
       atomic absorption methods. Mar. Chem. 17: 285-                  teractions with marine fulvic and humic materials
       300.                                                            in seawater using anodic stripping voltammetry,
 --,        AND R. P. FRANKS. 1983. Mn, Ni, Cu, Zn,                    p. 699-7 17. In Trace metals in seawater. NATO
       and Cd in the western North Atlantic, p. 395-4 14.              Conf. Ser. 4. Mar. Sci. V. 9. Plenum.
       Zn Trace metals in seawater. NATO Conf. Ser. 4.            PLAVSIC, M., D. KRZNARIC, AND M. BRANICA. 1982.
       Mar. Sci. V. 9. Plenum.                                         Determination    of the apparent copper complexing
--,         -,      G. A. KNAUER, AND J. H. MARTIN.                    capacity of seawater by anodic stripping voltam-
       1979. Sampling and analytical methods for the                   metry. Mar. Chem. 11: 17-3 1.
      determination ofcopper, cadmium, zinc, and nickel           RASPOR, B., H. W. NURNBERG, P. VALENTA, AND M.
      at the nanogram per liter level in seawater. Anal.               BRANICA. 1980. Kinetics and mechanisms of trace
      Chim. Acta 105: 223-245.                                         metal chelation in sea water. J. Elcctroanal. Chem.
mp,        G. A. KNAUER, ANDJ. H. MARTIN. 1978. Zinc                   115: 293-308.
      in northeast Pacific waters. Nature 271: 741-743.           RINGBOM, A., AND E. STILL. 1972. The calculation
COALE, K. H., AND K. W. BRULAND. 1988. Copper                         and use ofalpha coefficients. Anal. Chim. Acta 59:
      complexation     in the Northcast Pacific. Limnol.               143-146.
      Oceanogr. 33: 1084-l 101.                                  Ruzrc, I. 1982. Theoretical aspects of the direct ti-
 DAVISON, W. 1978. Defining the electroanalytically                   tration of natural waters and its information        yield
      measured species in a natural water sample. J.                   for trace metal speciation. Anal. Chim. Acta 140:
      Electroanal. Chem. 87: 395-404.                                 99-l 13.
 DONAT, J. R., AND K. W. BRULAND. 1988. Direct                   -.          1984. Kinetics of complexation          and deter-
      determination     of dissolved cobalt and nickel in              mination of complexation        parameters in natural
      seawater by differential pulse cathodic stripping               waters, p. 13 l-147. In C. J. M. Kramer and J. C.
      voltammetry preceded by adsorptive collection of                 Duinker [eds.], Complexation         of trace metals in
      cyclohexane- 1,2-clione dioxime complexes. Anal.                natural waters. Nijhoff/Junk.
      Chem. 60: 240-244.                                         -,         AND S. NIKOLIC. 1982. The influence of ki-
---,       AND -.         1989. A comparison of two vol-              netics on the direct titration curves of natural water
      tammetric techniques for determining zinc spe-                  systems-theoretical       considerations. Anal. Chim.
      ciation in Northeast Pacific ocean waters. Mar.                 Acta 140: 131-147.
      Chem. In press.                                            SHUMAN, M. S., AND G. P. WOODWARD, JR. 1976.
FIGURA, P., AND B. MCDUFFIE. 1979. Use of chelex                      Intermetallic compound formation between cop-
      resin for determination     of labile trace metal frac-         per and zinc in mercury and its effects on anodic
      tions in aqueous ligand media and comparison of                 stripping voltammetry.        Anal. Chem. 48: 1979-
      the method with anodic stripping voltammetry.                    1983.
      Anal. Chem. 51: 120-125.                                   SUNDA, W. G., AND R. L. FERGUSON. 1983. Sensi-
FISHER, N. S., AND J. G. FABRIS. 1982. Complexation                   tivity of natural bacterial communities to addi-
      of Cu, Zn and Cd by metabolites excreted from                   tions of copper and to cupric ion activity: A bioas-
      marine diatoms. Mar. Chem. 11: 245-255.                         say of copper complexation         in seawater, p. 87 l-
KRAMER, C. J. M. 1986. Apparent copper complcx-                       890. In Trace metals in seawater. NATO Conf.
      ation capacity and conditional stability constants              Ser. 4. Mar. Sci. V. 9. Plenum.
      in North Atlantic waters. Mar. Chem. 18: 335-              -,         AND A. K. HANSON. 1987. Measurement of
      349.                                                            free cupric ion concentration in seawater by a li-
 MANTOURA, R. F. C., A. DICKSON, AND J. P. RILEY.                     gand competition        technique involving        copper
       1978. The complexation         of metals with humic            sorption onto C,, SEP-PAK cartridges. Limnol.
      materials in natural waters. Estuarine Coastal Mar.             Oceanogr. 32: 537-55 1.
      Sci. 6: 387-408.                                           TURNER, D. R., M. WHITFIELD, AND A. G. DICKSON.
MARTELL, A. E., AND R. M. SMITH. 1974. Critical                        198 1. The equilibrium        speciation of dissolved
      stability constants. V. 1. Plenum.                              components in freshwater and seawater at 25°C
MOFFET, J. W., AND R. G. ZIKA. 1987. Solvent ex-                      and 1 atm. pressure. Geochim. Cosmochim. Acta
      traction of copper acetylacetonate        in studies 01’        45: 855-882.
      coppcr(I1) speciation in seawater. Mar. Chem. 21:          TUSCHALL, J. R., AND P. L. BREZONIK. 1983. Com-
      301-313.                                                        plexation of heavy metals by aquatic humus: A
NURNBERG, H. W. 1982. Voltammetric              trace analysis        comparative study of five analytical methods, p.
      in ecological chemistry of toxic metals. Pure Appl.             275-294. Zn R. F. Christman and E. T. Gjessing
      Chcm. 54: 853-878.                                              [eds.], Aquatic and terrcstrial humic materials.
----,       AND P. VALENTA. 1983. Potentialities           and        Butterworth.
      applications of voltammetry        in chemical specia-     VAN DEN BERG, C. M. G. 1982. Determination                    of
      tion oftrace metals in the sea, p. 67 l-7 17. In Tract          copper complexation with natural organic ligands
      metals in seawater. NATO Conf. Ser. 4. Mar. Sci.                in seawater by equilibration          with MnO,. Mar.
      V. 9. Plenum.                                                   Chem. 11: 307-342.
PANKOW, J. F., AND J. J. MORGAN. 198 1. Kinetics                 -.          1984. Determination       of the complexing ca-
      for the aquatic environment.       Environ. Sci. Tech-          pacity and conditional stability constants of cop-
      nol. 15: 1155-l 164.                                            per(I1) with natural organic ligands in seawater by
PIOTROWICZ, S. R., G. R. HARVEY, M. SPRING-                           cathodic stripping voltammetry            of copper-cate-
      ER-YOUNG, R. A. COURANT, AND D. A. BORAN.                       chol complex ions. Mar. Chem. 15: 1-18.
       1983. Studies of cadmium, copper, and zinc in-            -.          1985. Determination      of the zinc complexing
                                           Oceanic zinc speciation                                       285

    capacity in seawater by cathodic stripping voltam-      of particles in regulating the composition of sea-
    metry of zinc-APDC complex ions. Mar. Chem.             water, p. 457-493. In W. Stumm [ed.], Aquatic
    16: 121-130.                                            surface chemistry. Wiley-Interscience.
VAN LEEUWEN. H. P. 1987. Voltammetric          titrations
    involving’mctal    complexes: Effect of kinetics and
    diffusion coefficients. Sci. Total Environ. 60: 45-
                                                                          Submitted: 1.5August I988
    55.                                                                    Accepted: 10 October 1988
WHITFIELD, M., AND D. R. TURNER. 1987. The role                           Revised: 30 November 1988

To top