Limnol. Oceapogr.

, 25(5), 1980, 807-822

@ 1980, by the American Society of Limnology and Oceanography, Inc.

Microelectrophoretic properties of natural surface-active

organic matter in coastal seawater
K. A. Hunter1g2
School of Environmental Sciences, University of East Anglia,
Norwich NR4 7TJ, United Kingdom

Abstract
Measurements by the technique of particle electrophoresis rcvcal that different well char-
actcrizcd solid surfaces freshly cxposcd to seawater become covered by a tenacious film of
natural surface-active organic matter which dominates their subsequent surface chemistry.
The dependence of elcctrophorctic mobility (uB) on pH and metal ion concentration (pM)
indicates that carboxylic acid (-COOH) and phenolic (-011) groups are the major ionizable
functional groups iti the organic fihn exposed to the solution. Detailed treatment of the data
on uK vs. pH and uE vs. pM allows calculation of intrinsic equilibrium constants for the
dissociation of surface carboxylic acid groups and the complexation of metal ions by the
organic film. The results obtained are in substantial agreement with data derived from more
classical rlotentiometric titrations of marine and freshwater organic matter isolated by ad-
sorption Or ultrafiltration methods.

Many processes of geochemical signif- of adsorbed organic films (Loeb and

icance taking place in the ocean occur at Neihof 1975, 1977; Neihof and Loeb
interfacial boundaries and may be pro- 1974), which apparently also promote the
foundly influenced by the action of dis- accumulation of particulate radionu-
solved and colloidal surface-active organ- elides on the surfaces of marine organ-
ic compounds. Among these are several isms by adhesion (Hodge et al. 1979).
interfacial processes of importance to the Despite the obvious importance of
transport of particulate material and as- these effects, studies of the chemical na-
sociated trace elements in the sea and ture of organic matter in seawater with an
their ultimate deposition in marine sed- emphasis on the major surface-active
iments. These include flotation and ag- species are relatively few. Much of the
gregation of particulate materials by bub- colloidal organic matter in seawater is
bles (Wallace et al. 1977; Wallace and largely uncharacterized, with a complex
Duce 1975, 1978; Hunter 1980), floccu- and probably variable structure (Ogura
lation and transport of clay minerals and 1977).
other colloids in estuarine and shelf In the present work, the surface chem-
waters (Gibbs 1977; Hunter and Liss ical properties of the organic films
1979), inhibition of mineral-seawater in- formed by natural surface-active com-
teractions (Suess 1970, 1973), solubility pounds in coastal seawater on the sur-
of particulate trace elements recently de- faces of solid particulates were examined
posited on the sea (Wallace et al. 1977), directly by the technique of microelec-
and adsorption of dissolved trace ele- trophoresis. The ionization behavior of
ments by organic films on particles (Jaf- functional groups in the adsorbed organic
fritsic-Renault et al. 1980). The ecology films was studied by use of the charac-
of periphytic microorganisms and macro- teristic dependence of electrophoretic
fouling growths on immersed surfaces mobility on pH and metal ion concentra-
and the marine corrosion of metallic sur- tion in a manner analogous to conven-
faces are also influenced by the formation tioanl potentiometric titrations. The
methods involved resemble those used to
Financial assistance from the Rutherford Com- study bacteria cell surfaces (Abramson et
mittee of the Royal Society, London, and the N.Z.
University Grants Committee is acknowledged.
al. 1942; Bungenberg de Jong 1949;
2 Present address: Chemistry Department, Uni- Douglas and Shaw 1958; Shaw 1969).
versity of Otago, Box 56, Dunedin, New Zealand. I thank P. S. Liss, P. Brimblecombe,
807
808 Hunter

and R. Chesselet for advice and helpful used after UV irradiation. NaCl solutions
discussions. free of surfal:e-active contaminants were
prepared by dissolving NaCl preignited
Methods at 500C for 13h in double-distilled water.
Electrophoretic measurements were All synthetic and irradiated electrolytes
made at 25 + 0.2C with a Rank Mk II were rcplacc:d frequently and stored at
particle microelectrophoresis apparatus 6C when not in use to retard microbial
and a lo- x l-mm rectangular quartz cell. growth.
Palladium electrodes were used (Neihof
and Schuldiner 1960). The electropho- Results and Discussion
retic mobility UV, i.e. the velocity per Some indic:ation of the ultimate state of
unit field u,/E, was calculated from the surfaces immersed in seawater can be
mean velocities of at least 10 particles in gained by ellamining natural particulate
each direction at each of the stationary matter. The mean mobility of 20 particles
levels. The field strength E was calculat- in a seawater sample (33s0 salinity) was
ed as E = V/Z, where V is the applied -0.65 x 10-j rn* s-l *V-l, with 23% C.V.
voltage and Z the effective interelectrode A variation of 6-8% for an electrophoret-
distance, obtained by measuring the re- ically uniforrn dispersion is typical of the
sistance of the cell + electrodes filled microscope :nethod (Shaw 1969). I saw
with a standard KC1 solution by means of no positive particles, in contrast to the
a 3-kHz transformer bridge. predictions elf Pravdic (1970) and Martin
Four types of solid particles intended et al. (1971) With the lo- x l-mm cell
to represent a broad range of surface these low r-nobilities were difficult to
types were used: an acidic oxide (SiO,, measure in a concentrated electrolyte
Cab-o-sil), a basic oxide (A&O,, BDII), an such as sea-water owing to convective
organic polymer with cationic functional heating and gassing at the electrodes
groups (Amberlite IRA-402 anion ex- with the high currents (>20 mA) in-
change resin), and a hydrophobic organic volved. Measurements were more reli-
polymer nominally devoid of functional able after dilution of the electrolyte with
groups (Amberlite XAD-2). The raw sol- double-distilled water, when the nega-
ids were ground as a wet slurry in a mor- tive mobility of natural particles was
tar, suspended in double-distilled water, found to increase regularly in magnitude
and particles were selected for size in the with dilutior., in accordance with expan-
approximate range of 0.55-pm radius by sion of the electrically diffuse layer
differential sedimentation. Adsorbed or- (Overbeek and Lykl ema 1959). At a salin-
ganic material was removed by repeated ity of 30/00, Inobility was increased to
centrifugation and redispersion in dou- -1.19 x lOhE rn.s-l.V1, with 20% C.V.
ble-distilled water followed by pro- Other results obtained after improve-
longed dialysis. ments in electrode and cell design which
Seawater samples were collected in allow precise measurements in seawater
the North Sea 6-10 km SE of Gt. Yar- (Hunter and Liss 1979) confirm the re-
mouth, Norfolk, at a depth of 2040 cm sults describ zd here-natural particulate
and stored in chromic acid-cleaned 2.5- matter from the coastal North Sea has a
liter glass bottles at 6C in the dark after narrow range of mobilities, with all par-
removal of natural particulate matter by ticles being negatively charged. Micro-
centrifugation. Organic-free seawater of scopic and X. ray examination showed the
the same ionic composition was prepared dominant co nponents of the particulate
from filtered aliquots by UV photo-oxi- matter to be microorganisms, detritus,
dation (Armstrong et al. 1966) for periods quartz, and clay minerals.
of 3040 h (Neihof pers. comm.). Artifi- The mobil ties of the four types of car-
cial seawater prepared according to Kes- rier particles were identical in irradiated
ter et al. (1967) to match the salinity of seawater and artificial seawater solutions
the North Sea water (-34x0) was also of the same salinity (-33x0) and varied
 Surface-active organic matter 809

0.2 04 0.6 0.8 1.0 a

XAD-2
-l.O-

I
a
&/L---l-++r-

a
II. i I I 1
60 160 240 Vol/cm3

5.0
Fig. 1. Electrophoretic mobility of carrier par-
ticles in diluted seawater (3%0 salinity) after cxpo-
sure to successive 40-cm aliquots of clarified sea- 4.6
water (33%0 salinity).
(cl IRA-402

4,2 e
I
over a wide range from positive (IRA-
402) to negative (SiO,) according to the
ionization of surface groups and specific 0.2 0.4 0.6 0 8 1 0 a

ion adsorption for each surface type in Fig. 2. Apparent acidity constant pII - log[a/
organic-free seawater (Table 1, Fig. 1). (1 - a)] of seawater-treated carrier particles as a
Suspensions of the four solids (107-lo8 function of ionization degree CL
particles. cm-) were exposed repeatedly
to particle-free seawater by centrifuga- for each type of solid surface as a result
tion and redispersion (Neihof and Loeb of the progressive adsorption of acidic
1972). Table 1 shows the resultant mo- organic material. The effect is made
bilitics after 24 h of exposure to one ali- clearer in Fig. 1, which shows the mo-
quot of seawater and after repeated ex- bilities of carrier particles in a separate
posures. Also shown are results from a experiment after dilution to a salinity of
seawater sample stored in the dark at 6C 3%0. In each case (Table 1, Fig. 1) the
for 6 months. Th ese results clearly show mobilities of the carrier particles con-
a shift toward more negative mobilities verge after exposure to seawater onto a

Table 1. Electrophoretic mobilities (u,, 10 -R m.S-lv-l) o f carrier particles in UV-irradiated

seawater and in freshly collected or stored seawater (6 months) after a single exposure and repeated
exposures.

Fresh Slorcd
Carrier
particle 1 exposure Rcpcated 1 exposure Repented

IRA-402 1-1.16 k 0.07 +0.68 k 0.04 -0.65 2 0.09 +0.72 k 0.04 -0.30 5 0.05
XAD-2 +O.lO +- 0.07 -0 -0.30 IL 0.07 -0 -0.10 k 0.07
AI,O,, -10.20 iI 0.07 -0.10 2 0.0 -0.40 2 0.08 -
SiO, -0.41 2 0.07 -0.52 k 0.05 -0.51 Ik 0.05 - - 0.39: 0.07
810 Hunter

-2.0 1 08" I rn2s4v -' the surface chemistry of natural particles

E in seawater.
The experimental results described so
t
far confirm in all essential respects the
earlier finlings of Neihof and Loeb
(1972, 1974). Quite different types of sol-
id surface readily adsorb dissolved and
colloidal surface-active material from
seawater, leading to film-covered sur-
faces that are electrochemically similar in
behavior. lifter prolonged exposure to
T I 1 1 I I
seawater, the adsorbed organic films
2 4 6 8 IO 12 PH were not readily removed by washing
with irradiated seawater or more dilute
electrolytes. This stability of the ad-
sorbed organic films allowed transfer of
the treated particles to a simpler electro-
lyte (NaCI) for study of the mobility vari-
ations with pH and metal ion concentra-
[bl
tions,
-l.O-
Electrophoretic behavior with chang-
ing pH-Suspensions of A1,03, XAD-2,
and IRA-4C2 particles previously treated
with excess seawater were centrifuged
and the p,rrticles redispcrsed several
times in 10 mol. rnp3 NaCl electrolyte
2 4 6 8 IO 12 pH (the SiO, suspension had a strong ten-
dency to coagulate and further work with
it was abandoned). The suspensions
were then divided into two parts for mo-
bility measurements in acidic and alka-
line solutions to avoid irreversible effects
at the pH I:xtremes, and pH was varied
by additions of NaOH or HCl.
ICI The uE v.;. pH curves of the three types
of seawatela-treated solids are similar in
form and quite different from the curves
characteris tic of the solids themselves
(Fig. 3). A 20:, , as a basic oxide with a
PZC in the region 8-9 (Ottewill and Hol-
loway 1972 ), is positively charged below
pH 8-9, i.e most of the range shown. The
anion exch;mge resin IRA-402 is positive-
PH ly charged except at high pH, owing to
the presen:e of amino-N cationic groups
on the surface. XAD-2 showed fairly uni-
Fig. 3. Mobility vs. pH behavior in 10 rnol *rn+ form positive mobilities up to high pH
NaCl solution of seawater-treated carrier particles. whose overall magnitudes depended on
length of storage (in distilled water). Pro-
mobility range that encompasses the nat- longed storage gave rise to substantial
ural particulate material. This suggests positive m obilities (2-3 x lOA rn. s-l.
that adsorption of surface-active organic V-l), appa.-cntly through depolymeriza-
matter is a major mechanism controlling tion reactions.
 Surface-active organic matter 811

The similarity of the UV vs. pH curves ments. The interfacial region of a charged
after organic adso.rption illustrates the organic surface in solution is convention-
high degree of uniformity of surface type ally divided into three regions of surface
attained by surfaces exposed to seawater charge density which maintain overall
much more dramatically than uE mea- charge neutrality: the surface itself (us),
surements at a single pH value (Fig. 1). the Stern layer of specifically adsorbed
The zero mobilities attained at pH 2 and ions (al), and the diffuse layer of electro-
below indicate that positively charged statically adsorbed ions in the solution
functional groups, e.g. amino-N, are rel- (4:
atively unimportant. Although mobility
measurement below pH 2 was difficult as
a result of electrode gassing in acid so- Protonation of ionized acidic groups such
lution, there was no evidence for large as carboxylate anions at the surface can
positive mobilities in the acid region. Be- be considered as specific adsorption of
tween pH 2 and 6 there is a rapid in- H+ ions into the Stern layer with the in-
crease in negative mobility to 50-80% of trinsic charge density of the surface re-
the maximum values, with an apparent maining constant at a limit set by the
inflection point near pH 4.5. The single- number of ionizable groups per unit area
step ionization with an apparent pK of exposed to the solution. Changes in
~4.5 indicates the presence of carboxylic charge density with pH are then given by
acid groups. Phosphate groups also ion-
AU, = -Acr2, AU~ = 0.
ize in this pH region but display a two-
step ionization. Sulfate groups have pK I propose that the ionization behavior of
values typically <3 and would remain a charged organic surface can be de-
negatively charged at low pH (Shaw scribed by an equation analogous to that
1969). From pH 6 to 11 the negative mo- used to describe the acid-base equilibria
bility increases more gradually. The po- of charged macromolecules such as pro-
sition of this less steep increase in ioni- teins (Tanford 1961):
zation coupled with the apparent absence
of positive mobilities indicates weak PKinl = pH - log[[al( 1 - a)]
+ FV/2,303RT
acidic groups with pK values in the re-
gion 9-10, e.g. phenolic groups. At high where cy is the degree of ionization, pKint
pH, in the vicinity of pH 11, uF: changes is an intrinsic dissociation constant ap-
abruptly to values typical of the solid sur- plicable to the same functional group on
face itself as a result of the desorption of an uncharged surface, and F, R, and T
the adsorbed organic acids. take their usual meanings. The quantity
The electrophoretic mobility vs. pH V is an arbitrary potential which in-
curves of organic surfaces resemble the creases in magnitude with (x, vanishing
titration curves of equivalent solution at cy = 0, and corrects the basic mass-ac-
forms (Shaw 1969) and can be regarded tion equation for nonspecific effects aris-
as an electrophoretic titration of ioniza- ing from the surface charge, interactions
ble surface functional groups. Electroki- between functional groups that affect
netic methods sense only part of the their independent ionization, and small
change in surface charge density with differences in pKinl of functional groups
ionization, since the electrokinetic slip- of the same chemical type.
ping plane is generally located some dis- The degree of ionization cy is related to
tance from the surface. Thus, careful con- the change in (TPor op with pH. If cio and
trol of experimental conditions is oil are the appropriate charge densities
necessary so that there is a simple rela- (i = 1, 2) at a = 0 and (x = 1, then
tionship between ionization parameters
(Ti - Ui 0 Agi
and the electrokinetic results if thermo- a= l- o= 1 .
oi - cTi
dynamic dissociation constants are to be Oi ui

derived from electrophoresis measure- The relationshin of the charge densities

812 Hunter

oI or crZto electrokinetic parameters such stant pKinl, the quantity a was calculated
as mobility, uE , is quantitatively complex as a functior of pH after assuming values
and incomplete, since there is no explicit for extrema IY = 0 and 1. Zero mobilities
theory linking them (Ottewill and Hol- were assumc:d for a! = 0 at low pH, while
loway 1975). In sufficiently dilute solu- values of -1.0 (Al,O,), -1.33 (XAD-2),
tions, 5 can be taken as equal to the po- and -1.78 :< 10m8 rn2*s-l.V- (IRA-402)
tential at the Stern plane &, so that the were estimated from the curves, making
electrophoretic charge density oR is due allowar ce for the mobility changes
equal in magnitude to the charge density above pH 6 where the carboxylic acid
in the diffuse layer cr2(aE = -crZ = or + groups are completely ionized. From
a.$). If the effective particle radius of cur- these calculations, the quantity pH -
vature a is large in relation to the Debye- log[cx/( 1 - a,~] was obtained and is plotted
Hiickel thickness of the double layer (I/ in Fig. 2 for the three types of carrier par-
K) the electrophoretic charge density oE ticles.
and 6 are related by the expression For both 41,0, and XAD-2, this quan-
tity increases linearly with a over the ac-
(TE = (8RTa&,cb)~sinh[XF~/2RT] cessible experimental range and allows
(Ka > 100) extrapolatio 1 to pKint values of 3.92 for
Al203 and 2.68 for XAD-2. The plot for
which, at low values of 5 (x6 < 40 mV)
the IRA-402, carrier particles is linear at
simplifies to
low values of cx (~0.4) with more compli-
CTE = BEoK (Ka > 100, zt < 40 mv). cated behavior at higher values. The pKint
value is 4.02 in this case. The intrinsic
Under the same conditions, the Smolu-
pK values found here are similar to those
chowskii equation can be used to relate
found by potentiometric titration of (x-
uE and ,$ (Shaw 1969):
COOH groitps on proteins, and slightly
UE = &E,)tJI7). more acidic than protein side-chain car-
boxylic acid groups (Tanford 1961).
For larger potentials and smaller values The slopes of the three plots in Fig. 2
Of Ku, the relationships between uE, 6, GE, are similar and give values of !P = 55.
and the double-layer parameters become This result is significant since it indicates
very complex and demand in particular that the clectrokinetic potential 5 is un-
a knowledge of the effective radius of able by itself to account for the nonspe-
curvature. This is obviously unsatisfac- cific effect of surface charge on the acid-
tory where polydisperse suspensions arc base equilibrium, contrary to what has
used, as in the present work. According- been frequl:ntly assulned (Ottewill and
ly, I chose the particle size range and the Shaw 1967; Shaw 1969). Although other
electrolyte concentrations so as to remain effects suck as variable pKint values and
within the limitations outlined. For 10 functional group interference cannot be
molsm- NaCl solution, K has a value of discounted, the simple linear depen-
3.3 x 10H.mB3, implying that particles of dence of q on (Y and 4 indicated in Fig.
~O.5-~m radius as used in the uls vs. pH 2a and b is strong evidence that the be-
studies satisfy the requirements of large havior illus rated arises purely from elec-
Ka. The largest mobilities seen in Fig. 3 trostatic elects. The factor 5 can arise
in the region of carboxylic acid ioniza- from penetration by adsorbing II ions
tion, z-2 X lo+ m2.s-.V1, are equiv- into the Stc:rn layer. The potential drop
alent to a maximum 5 of 26 mV. Thus 5, in this interface region is at least expo-
crfiJ,and uE are linearly related under the nential with distance (Overbeek and
experimental conditions used and the Lyklema 1!359), so that the potential !P
mobility vs. pH behavior will have the will increase by a factor of 5 above (;
same form as conventional titration at the slipljing plane within a distance
curves (Fig. 3a-c). (In 5)/K = 5 nm. This distance is an order
For the derivation of the acidity con- of magnitude smaller than the thick-
 Surface-active organic matter 813

nesses of films of surface-active organic (a) carboxyl

00
1
M?
l
matter adsorbed from seawater onto plat-
inum plates as determined by ellipso- U$Pb
11 l
Cd NI
111
metry (Loeb and Neihof 1975). It is rea- Cu Zn

sonable that carboxylic groups should be TP A. cc,cl

I 1 11
found at least this far inside the slipping
plane and that H+ ions from solution can IJ I
-3 -2 -1 0 I 2 3
reach them.
Above pH 6, the increase in negative log(RCClmol mm31
mobility is attributed to phenolic groups.
Monophenols have pK values in the
range 9-10 (Sillen and Martell 1964),
while these functional groups on proteins Mg Ba
have intrinsic pK values of 9.5 < pKint < (b) phosphate n 1

10.8 (Tanford 1961). This range of pK val-

UO, Pp<n NI
ues is appropriate to the behavior in Fig. n IIM 1

2, but there is some uncertainty intro- &C

Th Al Ce La
duced by the rapid dcsorption of the or- m 111

ganic film above pH 11. Phenolic and car- I I 1 1 I I

boxy1 acid groups are the major oxygen- -3 -2 -1 0 1 2 3

containing ionizable groups that occur in

log(RCC/mol mm31
terrestrial humic substances (Stevenson
and Butler 1969). An approximate mea-
sure of the relative numbers of carboxylic
acid and phenolic groups can be obtained
from the uI$ vs. pH curves; values of r =
OH/COOH of 0.28 (IRA-402), 0.26 (XAD-
Zn Ni Cd
9, and 0.24 (Al,O,) so obtained appear to (c) IRA-LO2 11 n

be comparable within experimental lim- Th Fe Al La Ce

itations. Higher values of r are found by ma n l rn

titration or chemical analysis for tcrres-

trial humus (Stevenson and Butler 1969),
and Wilson and Kinney (1977) obtained
IogiRCClmol mm31
a value of 0.71 for freshwater humus by
potentiometric titration. There is some Fig. 4. Reversal of charge concentration spectra
for metal cations of (a) carboxyl colloid-sodium
evidence for preferential orientation of arabinate, (1,) phosphate colloid-egg lecithin, (c)
phenolic groups toward the surface when seawater-trcatcd IRA-402.
adsorption takes place onto solids con-
taining hydroxyl groups (e.g. oxides)
through hydrogen bonding (Giles 1959).
When proteins adsorb onto solid sur- tration (RCC) spectrum characteristic of
faces, many of the functional groups are these functional groups does not depend
not exposed to the solution (Bull 1958). very much on the nature of the underly-
Electrophoretic behavior with chang- ing organic structure (Douglas and Shaw
ing metal ion concentrcitions-Bungen- 1958; IIunter and Liss 1976). Figure 4
berg dc Jong (1949) has outlined the prin- shows characteristic RCC spectra for car-
ciples by which carboxylate, phosphate, boxy1 and phosphate colloids reproduced
and sulfate functional groups on organic from data given by Bungenberg de Jong
colloids can be distinguished by their dif- (1949). The general pattern of results, i.e.
ferent spectra of metal ion concentrations the relative positions of different metal
at which the electrophoretic charge on ions, is quite similar for all colloids hav-
the colloid is reversed from negative to ing the same type of dominant functional
positive. The reversal of charge conccn- group. An alternative representation of
814 Hunter

log(RCC/ mal m-3)

egg lec~tl~in

Hexol
uo 8
2 .Th

l AI
0 Ce
l Ce l La
La 0
0 uo*
.Pb

0 LUteO 0 Lute0

(a) carboxyl Q . Pb

(b) phosphate
% l M
colloids 2n colloids
) l ca

Ni
Be

I 1 ,
2 1 0 I 1
lw(FtCC/mol m-3 ) 2 1 0
log(RCC/mol mw3)
arabinate
soya bean phosphatlde II

Fig. 5. Comparison of log RCC for metal cations of (a) two carl~oxyl colloids and (b) two phosphate
colloids.

Fig. 6. Electrophoretic mobility of seawater-treated IRA-402 pulicles as a function of the lognrithm of

metal ion concentration for 13 metnl cations.
 Surface-active organic mutter 815

this concept not previously used in the 1 IRA-LO2

-3 -
literature is suggested here: in Fig. 5 are
shown plots of log RCC data for a suite l Th
-2 -
of metal ions of two carboxyl colloids and A-
two phosphate colloids. These compari- k

sons show a simple, almost linear rela-

tionship between the log RCC data for Pb
..uo,
two colloids having the same type of Cu
l LhOCe
group. Similar plots for different colloids,
e.g. carboxyl against phosphate, are scat-
tered. Zn
RCC concentrations of 13 different me- l m
0 Ni (al COOH colloid
tallic cations were obtained for seawater- Mg Cd
treated IRA-402 particles by measuring I
2 1
I I
0
the mobility as a function of metal ion log(RCC/mol mm31
concentration in double-distilled water. a Th
The film-coated IRA-402 particles were -2 -
washed with double-distilled water by i7
l AI

centrifugation and redispcrsion, and k -I-

known quantities of standard solutions of z

E Pb
UO*
metal salts were added successively. 2; o-
a
La
l
0
Metal salts used were chlorides (Mg, Ca, E Cu
l a
Ce
Ba, Ni, Cd), nitrates (Pb, Ce, La, UOz, F - a Ba
Th), sulfates (Fe, Cu), and acetate (Zn).
The pH of the resulting dispersions was 2
Ni .Zn

not controlled by buffering since most .Mg (b) PO4 colloid

buffers for the range pH 6-8 are either
organic or nutrients and contain signifi-
cant quantities of surface-active contam-
inants, especially after storage. The uE
vs. log M curves (where M = metal ion
concentration) so obtained are given in
Fig. 6. Reversal of charge occurred with
all of the cations used, with the lowest
RCC found for the polyvalent Th(IV) ion,
5 X 1O-3 mo1.m 3.
The RCC spectrum derived from these
results is shown in Fig. 4c for comparison
with the two model spectra for carboxyl
(Fig. 4a) and phosphate (Fig. 4b) colloids.
The distinctive features of Fig. 4c which
are in common with those of carboxyl col-
loids are several: lower RCC of Ba than
of Mg; general order of RCC values for
the transition metals, Pb < Cu < Zn <
Ni (the unusually high RCC of Cd prob-
ably results from competitive formation
of chloro-complexe,s); RCC of UOz2+ is
not unusually low as it is with phosphate
colloids; RCC values of Ce and La are
not unusually low compared with diva-
lent ions. Apart from the order La < Ce,
the IRA-402 spectrum displays all ofthe
t
log(RCC/ mol mm31
Fig. 7. Comparison of log RCC of seawater-
treated IRA-402 with (a) carboxyl colloid-sodium
arabinate and (11)phosphate colloid-egg lecithin.
features normally taken as characteristic
of carboxyl colloids (Bungenberg de Jong
1949). In Fig. 7a and b, the log RCC data
arc compared with those for the two
types of colloids in the same manner as
Fig. 5a and b. A clear pattern is seen in
the comparison with the carboxyl colloid
(Fig. 7a), although there are some small
variations for the transition metals. The
comparison plot with the phosphate col-
loid (Fig. 7b) is quite scattered.
On the basis of this evidence I con-
clude that the major ionizable surface
functional group of the organic material
adsorbed from seawater onto IRA-402 is
carboxylic acid. This result is in agree-
ment with the form of the uR vs. pH curve
for the same surface as already discussed.
The phenolic groups do not seem to enter
 +

816 Hunter

0 uo2
l La
0 Ce
@ Th

3-
l- \ @-------6 I

\ P -,
2-
2-

u /I cP%l*,v-
E

log(M,/mol I$~)
-2 -1 0 1

Fig. 8. Electrophoretic mobility of seawater-treated AI,@ particlt:s in 10 mol. m 3 NaCl solution as a

function of logarithm of metal ion concentrations. Left-strongly adrarl)ed ions; right-divalent cations.
Lowest curve for Cu+ is drawn correctly against uE axis. All successive curves are displaced upward by
-0.40 X 10-R m2. s-l *V- 1 for greater clarity.

into consideration, There has been no tion. It shculd be apparent, therefore,

explicit investigation of the effect of com- that thermoclynamic data for surface com-
binations of functional groups such as plexation of metal ions should be obtain-
phenolic-OH on the form of RCC spectra. able from electrophoretic measurements
My work suggests that the presence of in a manner analogous to pKinl acidity
minor phenolic groups does not perturb constants. 1~ successful method for ob-
the qualitative nature of the functional taining surace metal-complex associa-
group-metal ion equilibria to any signif- tion constants is presented here. The
icant extent. IRA-402 results in Fig. 6 are not suitable
The characteristic RCC spectrum of an for this purpose because the total electro-
organic surface, which is a summary of its lyte concenration varies widely over the
ufiJ vs. pM behavior in a simple form, is, us vs. pM curves and does not fulfill the
like the uE vs. pH behavior, a result of requirements of large KU which allow cal-
chemical interactions between functional culation of the surface charge densities
groups on the surfLIce and ions in solu- from the eh:ctrokinetic results. In a sec-
 Surface-active organic matter 817

ond series of experiments, the uE vs. pM

behavior of seawater-treated A&O, parti-
cles was therefore measured in 10 1.5
mol -mm3 NaCl solution as with the uK vs.
pH behavior, thus maintaining the ionic
strength (and hence K) so that KU is al-
ways large, During these measurements '3
0
pH was held constant within the range - 1.0
E
8.0 + 0.2 by adding NaOH or HCl to re-
verse pH changes caused by the metal
ions used.
The mobility vs. pM curves obtained
for A&O, after seawater treatment are giv-
en in Fig. 8. It is apparent that most of
the curves display the same general form
as the ug vs. pH curves, an extended S
shape typical of a mass-action equilibri-
um. A notable difference from the form 0.2 0.4 0.6 0.8 1.0
of results in Fig. 6 is that there is no re-
versal of charge for the divalent cations. 0
Extrapolation of the curves in Fig. 8 Fig. 9. Apparent metal ion association constant
(right) to zero mobility indicates RCC log (6/l - 0) - log M. Lower squares: after inclusion
values generally higher than the maxi- of an electrostatic correction factor 2F(/2.303RT.
mum concentrations used. The most im-
portant reason for this would seem to be
the competitive formation of surface com- intrinsic association constant for the for-
plexes of the metal cations with the uni- mation of the surface complex. The lower
valent ion Na in the background electro- limit of uE: at e = 1 is poorly defined in
lyte. the curves of Fig. 8, so that e was ob-
Intrinsic association constants for the tained from the data by a procedure
formation of surface metal complexes are which reduces the effect of uncertainty
obtained from a detailed analysis of the in uE1. The equilibrium equation can be
u& vs. pM curves which is analogous to rearranged to
that already carried out for the uIc vs. pH l/AcrE = llx,eN + l/k,c,x,eN
curves. The charge density UV must be - exp( -x,FWRT),
calculated from 4 by the expression for a
mixed electrolyte (NaCl + metal salt) allowing estimation of N from the inter-
given by Oldham (1975): cept of a plot of l/As, vs. l/c,. The quan-
tity u,~() was obtained by mobility mea-
CT&J
= (2RTEEJf surements in an ultraclean 10 mol *m-3
* (I& ci[exp( - xiF{IRT) - l]}f NaCl solution, as it is sensitive to small
where the summation extends over all quantities of trace metal cations present
ions i of charge zi in solution. The com- as contaminants in the electrolyte. The
plex equilibrium can be represented in degree of association e at each point on
the form the curve was calculated from

log k, = pM + log@/1 - e) e/I - 8 = Aa&,eN.

+ x,FW2,303RT
where the adsorbing ion M has charge xhl
and concentration pM = -log Cu. e is the
degree of coverage of complexing sites
by the ion M, and the constant k, is an
The quantity log (e/l - e) - log M was
then calculated and is shown as a func-
tion of 8 for the ion Cu+ in Fig. 9. This
is reasonably linear with e over most of
the range and extrapolates at e = 1 (i.e.
near zero charge on the surface) to
818 Hunter

Table 2. Experimental parameters for uE varia-

1 08 m%- v-1
tion of seawater-treated A&O, carrier particles in 10 -3.0 Ef
molmm- NaCl solution with addition of CuSO, so-
lution.

-2.76 + 0.05 10-8 m2.s-l.V-l

UEO
5 -35.4 k 0.6 mV
-8.8 If: 0.2 lop3 C*m-2 -2.0
UK0
N 2.16 x 10 rnp2
n,eN = uE1 - uEo 6.9 + 0.2 10m3 C*rnp2
k, 1.6 m3*mol-

-1.0

log k, = 0.210, or k, = 1.62 m3. mall =

1.62 x 10 dm3*mol-1 in more usual
units. The experimental values of the var- t I I 1.,
-12 -1 0 1 2 log(Cu/mol K3)
ious parameters involved for Cu2+ are
given in Table 2. Other metal ions ex- Fig. 10. M ability of seawater-treated Al,O, par-
amined gave similar behavior to that in ticles in 10 mol. rnp3 NaCl solution as a function of
CU-~ concentr Ition. Points are experimental, curve
Fig. 9 (Mg, Ca, Mn, Zn, UO,). The log k, calculated from theory.
values obtained by this procedure for
several of these ions are given in Table
3.
From the slope of the log@/1 - 0) - log of the specal nature of H+ ions in solu-
M VS. 6,plot for CU 2+, the correction factor tion with regard to size and mobility.
!P is seen to be very close to the experi- A complete test of the validity of the
mental 5; !P = 1.065. This was confirmed equilibriun- approach to the uE vs. pM
by plotting log@/1 - 0) - log CU + 2FiJ behavior developed here can be made by
2.303RT, as in the lower part of Fig. 9. calculating the uE vs. pM behavior ex-
An electrostatic correction for the charge pected front the theory. Values of N, zM,
on the surface, using 5, removes all vari- uEo from Table 2 were taken together
ation in the apparent association constant with CNa(;l= 10 mol.mP3 and the assump-
with 8. By the same procedure, ?P was tion that V = 5 to calculate u,~ as a func-
found to be similarly close to 5 for the tion of C&O, concentration. The equa-
other cations, V/c having the values 1.60, tions involved follow logically from the
1.55, 1.40, and 1.86 for the ions Mg2+, treatment of the experimental data al-
Ca2+ Zn2+ and U022+. This result is sub- ready outlined and, because of the ap-
staniially different from that found for the pearance elf TP = 5 in the exponential
acid-base behavior and suggests that un- terms, are nonlinear in 8. They were
like H+, the hydrated metal ions are un- solved by computer with a geometrically
able to penetrate very far into the organic converging iterative method. The curve
film when they bond to the surface func- in Fig. 10 resulting from this calculation
tional groups. This is reasonable in view shows that the theory completely ac-

Table 3. Comparison of electrophoretic log k, values with metal ion association constants of natural
organic colloids and small molecule models.

Compound Mn2 Zn2 Cu2+ Mg21 Ca2k uo,* +

Al,O, films 2.2 3.0 3.2 2.1 2.6 4.5 This work
Smith Lake organics 2.50 3.37 3.68 - - Wilson and Kinney 1977
Phthalic acid 2.74 2.89 3.46 - GI3 4.38 Sillen and Martell 1964
Salicylic acid 5.90 6.85 10.60 ==3 =3 =:I2 Sillen and Martell 1964
Seawater IR2 4.51 5.31 9.71 3.98 4.12 - Mantoura et al. 1978
 Surface-active organic matter 819

counts for the uE variation with Cu con- -3 o lo8";m2s-' v-1

centration in the range studied. Similar
results applied to the other ions when the
appropriate ratio Yr:[ was taken. In fact, > %
h
the electrostatic distortion of the uE curve
is quite small when !P is so near to 5, as -2,o- %%
shown by the curves in Fig. 11, which \
were calculated from the appropriate %
\
equilibrium parameters for Mg2+, Ca2+,
and Zn2+ but with the assumption that 5 -10 \
can be used to correct for the electrostatic \ 0
interaction as with Cu+. The agreement
still falls largely within observed exper-
imental errors. log ( Mglmol rrf3)
1 i I
Comparison with other work-There -3 -2 -1 0 1 2
have been relatively few studies of the
ionizable functional group chemistry of
-3.e 10 m2s-V-
dissolved or colloidal organic matter in El
seawater. Wilson and Kinney (1977) car-
ried out potentiometric titrations in 0.15 \a 0
mol. dm-3 KC1 solution of high molecular
weight material harvested from seawater
in the Gulf of Alaska and from freshwater
in Smith Lake, Alaska, by ultrafiltration.
Intrinsic acid dissociation constants for
carboxylic acid and phenolic groups were
-1.1
-2.0 \
found to be 3.9 (-COOH) and 9.3 (phe-
nolic -OH) for the seawater extract, and 0

4.23 (-COOH) and 8.71 (phenolic -OH)

for the freshwater extract. The seawater
log(Ca/mo I me31
results are in excellent agreement with 1 1 , I
my work with intrinsic pK values of 4.02 -3 -2 -1 0 1 2
(IRA-402), 3.68 (XAD-2), and 3.92 (Al,O,) G
for carboxylic acid groups and pK in the -30 108u E! m2s-V-
range 9-10 for phenolic groups. Wilson
and Kinney also determined metal-ion
\ 0
association constants for the freshwater
extract for several divalent cations. Their
results, given in Table 3, show a reason- -2.0 $
able similarity with the intrinsic log k,
values determined in the present work,
despite the fact that freshwater and ma- 0
rine organic colloids are being compared.
The similarity in ionization behavior al-
ready noted with regard to acid-base -1.0 \
properties (Wilson and Kinney 1977; Hu-
izenga and Kestcr 1979) therefore ex- log(Zn/mol mm31
I
tends also to complex formation with me- -3 -2 -1 0 1 2
tallic cations. Fig. 11. Mobility of seawater-treated Al,O,, par-
Table 3 also shows stability constant ticles in 10 mol*m-3 NaCI solution as a function of
data from the literature for two simple Mgz+, Ca2+, and Znzk concentration. Points arc cx-
monomer model ligands with carboxylic perimental, curves calculated frorn theory.
acid and phenolic functional groups-
820 Hunter

phthalic acid, C,jH,(COOH), , and salicyl- is further sltpported by the elevated crit-
ic acid, C,H,(OH)(COOH). The pattern ical surf:nce tensions of film surfactants
of log k, values found for organic films (Baier et al, 1974). Among the functional
on AI203 conforms closely to that of the groups meiitioned, carboxylic acid and
dicarboxylate, phthalic acid, suggesting phenolic groups are the major ones ca-
that the stability constants of this material pable of io.rization in aqueous solution,
could be used in a preliminary speciation The electrophoretic studies provide no
model for natural waters containing col- evidence for functional groups capable of
loidal organic ligands. Salicylic acid, positive ch,lrge such as -NH, guanidyl,
with single carboxylate and hydroxyl and imidazole groups, all of which would
groups, gives a quite different speciation act as positive indicators of protein.
pattern. Mantoura et al. (1978) have mea- There arc considerable limitations to the
sured metal ion complex stability con- ability of electrokinetic methods to de-
stants for low molecular weight (mol wt tect positively charged groups in small
~700) fulvic acids extracted from fresh- quantity, compared to the anionic func-
water and seawater by adsorption onto tions such as -COO-, since the important
Amberlite XAD-2. Some of their results diagnostic region is at low pH where uE
are also given in Table 3; comparison values become small and difficult to mea-
with the present work suggests that these sure. Marine fulvic acids have a C:N ratio
low molecular weight species are not of of ~7.7 with most of the nitrogen present
major importance in consolidated organic in nonhydrolyzable forms such as poly-
films on particles such as Al,O,. Their sta- functional amines and heterocyclics
bility constant data indicate a speciation (Gagosian ;md Stuermer 1977). With a
pattern much closer to that of salicylic molar ratio of -COOH:-NII of >lO, pos-
acid than of phthalic acid. itive mobil. ties at pH < 2 would almost
Huizenga and Kester (1979) reported certainly bca undetected by present tech-
potentiometric titration results for sev- niques.
eral samples of organic matter isolated
from seawater by adsorption onto activat- ConclusionSs
ed charcoal. They found little variation in Electrophoretic measurements reveal
the titration behavior between pH 2 and that widely different surfaces freshly ex-
8 for samples from different marine and posed to seawater become covered by a
freshwater environments, with a mean tenacious film of natural surface-active
intrinsic pK for nine samples of 3.6 * 0.1 organic material which dominates their
for the groups they describe as type 1 subsequent surface chemistry. The mo-
(presumably -COOH). Titration data bility variation with pH and metal ion
above pH 8 are not given, but many of concentration point to carboxylic acid
their curves show a steady increase in the and phenolic groups as the major ioniz-
number of H+ ions titrated above pH 7 able functions, and the films formed by
which may be evidence for the presence adsorption Into three different solid sur-
of weaker groups such as aromatic -OH. faces are strikingly uniform in this re-
Spectroscopic studies (IR,13C, and [HI spect. Detailed treatment of the ZC~vs.
NMR) of marine fulvic acids isolated by pII and u,~ vs. pM data allow calculation .
adsorption onto Amberlite XAD-2 indi- of intrinsic equilibrium constants which
cate a functionality dominated by carbox- compare well with similar data obtained
ylic acids, esters, and amides; aliphatic by more ch.ssical methods with different
and aromatic OH and heterocyclic-N fractions of marine dissolved organic
(Stuermer and Payne 1976; Stuermer and matter. This indicates that reliable thcr-
Harvey 1978). The IR spectra of natural modynamic data can be obtained by elec-
sea-surface films obtained by the prism- trophoretic measurements on suspended
dip technique (Baier et al. 1974) are in particles, provided experimental condi-
broad agreement with this assignment. tions are carefully controlled. The tech-
This high degree of surface oxygenation nique thus has advantages over classical
 Surface-active organic matter 821

methods in that uncertain extraction lonium in the biogeochemistries of the coastal

methods are not necessary. Furthermore, zone. Nature 277: 206-209.
HUIZENGA, D. L., AND D. R. KESTER. 1979. Pro-
natural suspensions with 10 particles. tonation equilibria of marine dissolved organic
ml-* or more can be directly examined matter. Limnol. Oceanogr. 24: 145-150.
(Hunter and Liss 1979). HUNTER, K. A. 1977. Chemistry of the sea surface
Additional studies by this method in- microlayer. Ph.D. thesis, Univ. East Anglia.
363 p.
dicate that macromolecular species and -. 1980. Processes affecting particulate trace
hydrophobic humic material isolable us- metals in the sea surface microlayer. Mar.
ing Amberlite XAD-2 resin are important Chcm. 9: 49-70.
contributors to the ambient organic films -, AND P. S. LTSS. 1976. Mcasuremcnt of the
(Hunter 1977). Although the present solubility product of various metal ion carbox-
ylates. J. Electroanal. Chcm. 73: 347-358.
work applies specifically to coastal waters, -, AND -. 1979. The surface charge of
its combination with other studies sug- suspended particles in estuarine and coastal
gests that an important fraction of seawa- waters. Nature 282: 823-825.
ter DOM, at least 10% (Stuermcr and JAFFR~SIC-RENAULT, N., K. A. IIUNTER, R. CIKES-
SET,ET, AND C. E. LAMBERT. 1980. Etude ex-
Harvey 1977) and possibly as much as perimental de ladsorption du Zn et du Se
70% (Huizenga and Kester 1979), dis- sur des particulcs dalumine dam leau de mer.
plays a systematic uniformity of ioniza- Oceanis 5: 597-601.
tion behavior. Because of its inherent KESTER, D. R., I. W. DUEDEI,I., D. N. CONNOR, AND
surface activity, this may be the fraction R. M. PYTKOWICZ. 1967. Preparation of arti-
ficial seawater. Limnol. Oceanogr. 12: 176-
of greatest geochemical significance. 179.
LOEB, G. I., AND 1~. A. NEIIIOF. 1975. Marine con-
ditioning films. Adv. Chem. Ser. 145: 319-335.
References -, AND -. 1977. Adsorption of an organic
film at the platinum-seawater in&f&e. J. Mar.
ABMMSON, H. A., L. S. MOYER, AND M. I-I. Gon- Res. 35: 283-291.
IN. 1942. Elcctrophorcsis of proteins and the MANTOUI~~, R. F., A. DICKSON, AND J. P. RILEY.
chemistry of cell surfaces. Reinhold. 1978. The complexation of metals with humic
ARMSTRONG, F. A., P. M. WILLIAMS, AND J. D. materials in natllral waters. Estuarine Coastal
STRICKLAND. 1966. Photo-oxidation of organic Mar. Sci. 6: 387408.
matter in seawater by ultraviolet irradiation; MARTIN, J. M., J. JEDNACEK, AND V. PRAVDIC.
analytical and other applications. Nature 2 11: 1971. The physico-chemical aspects of trace
481483. element behaviour in estuarine environments.
BAIER, R. E., D. W. GOUPIL, S. PERLMUTTIXR, AND Thalassia Jugosl. 7: 619-637.
R. KING. 1974. Dominant chemical composi- NEIIIOF, R. A., AND G. I. LOEU. 1972. Surface
tion of sea surface films, natural slicks and charge characterization of particulate matter in
foams. J, Rech. Atmos. 8: 571-600. seawater. Limnol. Oceanogr. 17: 7-16.
BULL, II. 1% 1958. The effective electrophoretic -, AND -, 1974. Dissolved organic rnatter
radii of adsorbed protein molecules. J. Am. in seawater and the electric charge of immersed
Chem. Sot. 80: 1901-1904. surfaces. J. Mar. Rcs. 32: 5-12.
RUNGENBERG DE JONG, II. G. 1949. Reversal of -7 AND S. SCIIULDINER. 1960. Simple non-
charge phenomena, equivalent weight and spc- gassing electrodes for use in electrophoresis.
cific properties of the ionized groups, p. 259- Nature 185: 526-527.
334. Zn H. R. Kruyt [ed.], Colloid scicncc, v. 2. OWJM, N. 1977. High molecular weight organic
Elsevicr. matter in seawater. Mar. Chem. 5: 535-549.
DOUGLAS, II. W., AND D. J. SIIAW. 1958. Electro- OI,DIIAM, K. R. 1975. Composition of the diffuse
phoretic studies on model particles. Trans. Far- double layer in seawater of other media con-
aday Sot. 54: 1748-1753. taining ionic species of +2, + 1, -1 and -2
GA<:OSIAN, R. R., ANI) 11. EI. STUERMER. 1977. The charge types. J. Electroanal. Chcm. 63: 139-
cycling of biogenic compounds and their di- 156.
agcnetically transformed products in seawater. OTTXWILI,, R. II., ANI> L. R. HOLI,OWAY. 1975.
Mar. Chcm. 5: 605-632. Electrokinctic properties of natllral particles, p.
GIBBS, R. J. - 1977. Clay mineral segregation in the 599-621. Zn E. 1). Goldberg [ed.], The nature
marine environment. J. Sediment. Petrol. 47: of seawater. Dahlcm Conf.
237-243. -> AND J. N. SHAW. 1967. Studies on the prep-
GILES, C. II. 1959. Hydrogen bonding in adsorp- aration and charactcrisation of rnonodisperse
tion processes, p. 449460. Zn D. Hadzi red.], polystyrene lattices. Kholloid. Zh. 218: 3440.
IIydrogen bonding. Pergnmon. OVERIIEEK, J. T., AND J. L,YKI,EMA. 1959. Electric
IIODGE, V. F., M. KOIDE, AND E. I>. GOI,DBERG. potentials in colloidal systems, p. l-33. Zn M.
1979. Particulate uranium, plutonium and po- Rier red.], Electrophoresis. Academic.
822 Hunter

PRAVDIC, V. 1970. Surface charge characterization -. 1973. Interaction of organic compounds

of sea sediments. Limnol. Oceanogr. 15: 230- with calcilun carbonate. 2. Organo-carbonate
233. association in recent sediments. Geochim. Cos-
SIIAW, D. J. 1969. Electrophoresis. Academic. mochim. Acta 37: 2435-2447.
SILL~N,L.G.,AND A. E. MARTELL. 1964. Stability TANFORD, C. 1961. The physical chemistry of mac-
constants of metal ion complexes. Chem. Sot. romolccul~s. Wiley.
Spcc. Publ. 17. WALLACE, G. 'I?., AND R. A. DUCK 1975. Concen-
STEVENSON, F.J., AND J.H. BUTLER. 1969. Chem- tration of 1)articulate trace metals and particu-
istry of humic acids and related compounds, p. late organi : carbon in marine surface waters by
534-557. Zn G. Eglinton and J. H. Murphy a bubble f otation mechanism. Mar. Chem. 3:
[eds.], Organic geochemistry. Longmans. 157-181.
STUERMEH, D. H., AND G. R. HARVEY. 1977. The -,AND---. 1978. Open-ocean transport of
isolation of humic substances and alcohol-sol- particulate trace metals by bubbles. Deep-Sea
uble organic matter from seawater. Deep-Sea Res. 25: 8?7-835.
ResL. 24. . 303-309. -, G. I,. I-IOFFMAN, AND R. A. DUCE. 1977.
-, AND -. 1978. Structural studies on The influcllcc of organic matter and atmospher-
marine humus: A new reduction sequence for ic deposition on the particulate trace metal con-
carbon skeleton determination. Mar. Chem. 6: centration Df Northwest Atlantic surface water.
55-70. Mar. Chenl. 5: 143-170.
- AND J. R. PAYNE. 1976. Investigation of WILSON, D. E, AND P. KINNEY. 1977. Effect of
seawater and terrestrial humic substances with polymeric charge variations on the proton and
13-C and 1-H nuclear magnetic resonance. metal ion quilibria of humic materials. Lim-
Geochim. Cosmochim. Acta 40: 1109-l 114. nol. Oceanogr. 22: 281-289.
SUESS, E. 1970. Interaction of organic compounds
with calcium carbonate. 1. Association phe- Submitted: 2 July 1979
nomena and geochemical implications. Gco- Accepted: 4 April 1980
chim. Cosmochim. Acta 34: 157-168.