Professional Documents
Culture Documents
Estimates of the sea-to-air flux of dimethylsulfide (DMS) are based on sea surface concentration
measurementsand gas exchange calculations. Such calculations are dependent on the diffusivity of
DMS (DDMs), which has never been experimentallydetermined.In this study the diffusivity of DMS
in pure water was measuredover a temperaturerange of 5ø-30øC.The measurementswere made using
a dynamic diffusioncell in which the diffusinggasflows over one side of an agar gel membrane and the
inert gas flows over the other side. The diffusion coefficient can be estimated from either time
dependentor steady state analysisof the data, with an estimated uncertainty of less than 8% (1 rr) in
eachmeasurement.A bestfit to alltheexperimental
results
yieldstheequationDDMs (incm2sec
= 0.020exp(-18.1/RT), whereR = 8.314 x 10-3 kJmole-1 K -• andT is temperature
in kelvin.
The values of DDMS obtained in this study were 7-28% larger than estimatesfrom the empirical
formula of Hayduk and Laudie (1974) which has previously been used for DMS in gas exchange
calculations. Applying these values to seawater results in an increase of less than 5% in the global
oceanic flux of DMS.
INTRODUCTION tude of the gas exchange coefficient has been deduced from
a variety of artificial and natural tracer experiments utilizing
The sea-to-air exchange of dimethylsulfide (DMS) is a •4C,Rn, SF6, and3He [Penget al., 1979;Smethieet al.,
major component of the global atmospheric sulfur budget. 1985; Wanninkhof et al., 1985; Watson et al., 1991], and the
Current estimates of this flux are based on the measurement results are extrapolated to DMS using the above relation-
of sea surface DMS concentrationsand various parameter- ships. The diffusivity of DMS has not previously been
izations of gas exchange across the sea surface [Andreae, measured. Instead, it has been estimated from the diffusiv-
1990; Bates et al., 1987]. These parameterizations take the ities of other nonelectrolytes using various empirical formu-
general form of lae relating diffusivity to molar volume, viscosity, and tem-
perature [Wilke and Chang, 1955; Hayduk and Laudie,
Flux= Kl(Cl - Ca/a) = Ka(Cla - Ca) (1) 1974].
In this study we experimentally determine the diffusion
where K is the gas exchange coefficient or piston velocity
coefficient for DMS, compare the measured and calculated
(expressedon a liquid or gas phase basis), C is the concen-
values, and discuss the implications for the global flux of
tration in the liquid or gas phase, and a is the dimensionless
DMS. We also determined the diffusion coefficient of meth-
solubility of the gas in seawater [Liss and Slater, 1974]. The
ane in order to evaluate systematic errors introduced by the
gas exchangecoefficient contains both liquid and gas phase
experimental apparatus and to compare our results to diffu-
components
(kI andka) asfollows: sivities reported in the literature for other gases.
1 1 1
-+ (2) EXPERIMENTAL METHOD
KI kl aka
The experimental approach used in this study is a varia-
Experimentalstudieshave shownthat ka >> kI under tion on the method of Barter [1941], and the apparatus was
natural conditions [Liss, 1973]. Therefore gas exchange is modified from the design of Jiihne et al. [1987a]. The
dominated by resistance in the liquid phase for slightly diffusion cell consists of a stainless steel housing with two
soluble gases. This is also the case for DMS, which has chambers on either side of an aqueous gel membrane (Figure
intermediate solubility in water. It has also been demon- l a). At the onset of the experiment, a flow of DMS in
stratedexperimentallythat k I is dependenton the diffusivity nitrogen is introduced on the ::high-concentration" side of
of the diffusinggas D with a functional form of the membrane (referred to as chamber 1), while pure nitro-
gen flows on the "low-concentration" side of the membrane
kI =f(Sc -n) =f(D n) (3) (referred to as chamber 2). The experiment consists of
measuring the ratio of the DMS concentrations in the gas
where n may vary from 1/2-2/3 dependingon the sea state
flows from each chamber as a function of time from the start
[Liss and Merlivat, 1986; Ji•hne et al., 1987b]. The magni-
of the experiment or after equilibration of the membrane. In
Copyright 1993 by the American Geophysical Union. this section we briefly discuss the use of diffusion theory to
Paper number 93JC01858. model the experimental results and describe the experimen-
0148-0227/93/93 JC-01858505.00 tal procedure.
16,481
16,482 SALTZMAN
ET AL.' DIFFUSIONCOEFFICIENT
OFDIMETHYLSULFIDE
OC (C2- C•) 2D
cI) = -D -- = -D
Ox I I
0el
nx 71'
TFE/Polgethglene
Z (C2cosnrr- C1)cos l exp- (Dn2rc2t/l
2)
n=l
Support
(5)
Diffusion Cell
At x - I this expression reduces to
OC (C2- C•) 2D
• = -g
Ox I
ß Z (C2cosnrr- C1)cosnrrexp-(Dn2rr2t/l
2)
n=l
Temperature Controller/
Circulating Bath
4DCo
I Z cos
(2m
+1)rr
exp-[D(2m
+1)2rc2t/l
2]
JThermometer
J
m=l
(6)
J' I
Assumingthat C2 and Co are both equal to zero gives
I Aqueous I ----- C2,f2 OC C1 2DC1
N
:2
Supply I Bubblers
I
Cell I
I
I '--'" CI,fl
--=
Ox
+D--+•
I I
'
IDMS
I Reservoir
I
•_
I
I
I
I
ß Z cosnrrexp- (Dn2rc2t/l
n=l
2) (7)
Fig. 1. (a) Schematiccross sectionof the diffusioncell. The In our experimentsthe concentrationratio of the diffusing
high- and low-concentrationchambersare labeled as 1 and 2, gas in contact with either side of the diffusionmembrane
respectively,as in (4). (b) Schematicof the experimentalapparatus.
The gas flows from the cell are labeled as in (10). (C2g/Clg)
canbedetermined
moreaccurately thanabsolute
concentrations. Thus we do not determine the absolute flux
through the membranebut, rather, the ratio of the flux
Theoretical Analysis
through
themembrane
atx = I to C•g. C• canbeconverted
Crank [1975] gives the following expressionfor the con- to C• if the solubilityis known.We cansolvethe theoretical
centration of a diffusingsubstance(C) in an infinite sheet or expressionfor the flux for this ratio:
membrane'
where I is the membrane thickness, C• and C2 are the As describedabove, the experimental quantity of interest
concentrationsat x - 0 and x = 1, respectively, and Co is is the ratio of the gas flux throughthe "low-concentration"
the concentration in the sheet at t - 0, which is assumed to surface of the membrane to the gas concentration at the
be uniform throughout the sheet. "high-concentration"sideof the membrane.This is givenby
SALTZMANET AL.: DIFFUSIONCOEFFICIENT
OF DIMETHYLSULFIDE 16,483
el) c2f2 The gel thickness used in each experiment was calculated
(lO) from the gel weight and the known diameter of the cell. The
•'•= Claa
A gel density was determined experimentally to be 0.992 g
C2aandCla arethe cm-3 (ltr = 0.05%)at 25øC.The uncertainty
wherecI)isthefluxoutofthemembrane, involvedin
gasphaseconcentrationson either sideof the membrane,f2 determining the gel thickness has two components. The first
is the gas flow on the low-concentration side of the mem- is the uncertainty in the physical measurement of the thick-
brane, a is the dimensionless Ostwald coefficient for DMS in ness, which is largely due to the uncertainty in the determi-
water (ratio of aqueous phase to gas phase concentration), nation of the density of the gel. The second results from the
and A is the cross-sectional area of the membrane.
loss of some of the gel to evaporation during the course of
The experimental apparatus is shown in Figure 1b. The the experiment, which was greatest at the higher tempera-
cell is immersed in a stirred, thermostated water bath which tures. The combined uncertainty in the measured thickness
was varied at 5 ø intervals from 5 ø to 30øC over the course of
is less than 5%. For steady state calculations the latter
the experiments. The nitrogen gas supplies were passed uncertainty was removed through the measurement of the
through glass-fritted bubblers immersed in the water bath to gel thickness at the conclusion of the experiment.
saturatethem with water vapor prior to entering the cell. The The concentrations of DMS on either side of the mem-
temperature of the cell was monitored using a thermocouple brane are determined by withdrawing samples from their
sensor placed in a well drilled near the center of the cell.
respective outlet tubing with gastight syringesusing perflu-
DMS was introducedinto chamber 1 by passageover a small
oroalkoxy needles. Thus samples contact only glass and
glass bulb containing liquid DMS (purity >99%, Aldrich,
Teflon during sampling. Samples from the high-concentra-
Milwaukee, Wisconsin). The concentration on this side of
tion side are diluted by a factor of roughly 10 (500 for
the cell is therefore slightly undersaturatedwith respect to
methane) in the syringe by adding pure nitrogen. The sam-
pure DMS. For methane runs, the glass bulb was removed,
ples are then loaded into the injection loops of a 10-port gas
and the pure gas (purity 99.0%, Liquid Carbonic, Chicago,
injection valve and analyzed by gas chromatography with
Illinois) was introduced into the bubbler in place of the
nitrogen.Gasflowratesof 10cm3min- • and20 cm3 min- • flame ionization detection. Two different volume loops (30
were used on the high- and low-concentration sides of the and 300 taL) are used for the high (Cla) and low-
membrane, respectively, during the experiments. concentration
(C2a)sidesof themembrane,
respectively.
As
The membrane is an agar gel (0.5%) which is 3.8 cm in a result, the amount of DMS (or methane) injected on
column from both sides of the cell is similar once the cell
diameter and approximately0.6 cm thick. Agarosegelshave
been used in many previous studiesof diffusionthrough a reaches steady state. Linearity in detector response was
liquid membrane [Schantz and Lauffer, 1962; Spalding, tested (using DMS) by successive dilutions of the effluent
1969; Langdon and Thomas, 1971]. Jiihne et al. [1987a] from the low-concentrationsideof the cell to cover the range
compared gas diffusion through agarose gels to wetted frit of concentrationsobserved during time-resolved measure-
diaphragms.Their results demonstratedagreementbetween ments. A linear relationship between peak area and dilution
the two methods, with more reproducible results from the factor was obtained. A least squareslinear regressionto the
gels. This is presumably because convective or other turbu- data gave a slope of 1.01 _+0.05 (1 tr); thus no corrections for
lent motion is inhibited in the gels. Two small corrections detector linearity are needed. The absolute concentrations of
must be accounted for when calculating aqueous diffusion the gas streams are not determined; instead, the ratio of the
coefficients from measurements made in gel membranes. peak areas is taken as the ratio of their concentrations. The
The first is the reductionof solubilityof the diffusinggasdue analyseswere done on an HP5890 gas chromatographwith a
to the lowering of the activity of water. Second, there is a 2-m Chromasil 330 column, an oven temperature of 50øC,
hindrance to the diffusion path due to the formation of a and nitrogencarriergas at a flow rate of 30 cm3 min-•.
three-dimensionalnetwork not found in pure water. Lang- Chromatographicdata was acquired and integrated using a
don and Thomas [1971] estimated that both effects reduce PC-based data acquisition system.
the diffusion coefficient for a substancethrough a gel in The gas flow rate on the low-concentration side of the
comparisonto pure water by 1.36% for a 0.5% gel. membrane (f2) was determined using a soap bubble flow
The gel in our apparatusrestsdirectly on a sheetof porous meter corrected for temperature and water vapor variations.
polytetrafluoroethylene filter membrane 0.13 mm thick with The gas flow on the high-concentration side of the membrane
mean pore size 10-20/am (Zitex, Norton Company, Wayne, (fl) was also monitored to insure constantflow during the
New Jersey). The sheet is supportedby a porous polyethyl- course of each experiment, but this flow rate is not used in
the calculation of the diffusion coefficient.
ene sheet 1.59 mm thick with mean pore size 15-45 tam
The Ostwald coefficient of DMS in water was obtained
(X-4900, Porex Corporation, Fairburn, Georgia). The poros-
ity of both sheets is sufficiently large that they make a from Dacey et al. [1984]. A linear regressionwas fit to their
negligible contribution to the resistance of the membrane to data and yielded the equation
gas diffusion. The gel is cylindrical in shape,but for a small In a = -10.1794 + 3761.33(1/T) (11)
portion of its total length the gel diameter is slightly in-
creased by the presence of a small (0.79 mm x 0.40 mm) where T is temperature (in kelvin). The estimated uncer-
groove machined in the cell wall. This groove provides tainty in this fit is 2.52% (lit). The uncertainty in the
friction for the gel to prevent it from slidingupward in the coefficient,however, was due mainly to the variations in the
event of a slight pressure gradient between the two sides temperature of the cell. The accuracy of the thermocouple
during setup. The presence of a groove of these dimensions used for temperature measurement in this experiment is
has a negligibleeffect on the diffusion of gasesthrough the -+0.4øC.This results in an uncertainty of less than 2% (1tr) in
cell, as demonstrated by Barrer et al. [1962]. the Ostwald coefficient.
16,484 SALTZMANET AL.' DIFFUSION COEFFICIENTOF DIMETHYLSULFIDE
i i i
2.8
Temperature(degrees K)
2.8 I-1 -
øooooo DO d• 3.0
278
i
28,.3
i
288
i
293
i
298
i
303
i
o 2.4 _
2.4
D=O.O31exp(-18.3/RT)
2.5 •7
E 2.0
_
2.0 '"'
•
1.2
0.8
1=0.61
+0.004
(1
a) 0.8
(o
w
v
I
1.0 A
O this study
iT 0.4 O Cl - 0.4 0.5
V Johne
FI Wilke-Chong
et ol.
O Flux/C 1
A Hoyduk-Loudie
0.0 ' ' ' ' 0.0 0.0 I I I I I
10 15 20 25 30 ,.5 10 1,5 20 25 30 35
Absolute Relative
Parameter Range Uncertainty Uncertainty %
TABLE 2. Schmidt Numbers for Dimethylsulfide in Seawater of Erickson, D. J., III, S. J. Ghan, and J. E. Penner, Global ocean-to-
35 Parts-per-Thousand Salinity atmosphere dimethylsulfide flux, J. Geophys. Res., 95, 7543-
7552, 1990.
Temperature, Schmidt Number, Eyring, H., Plasticity and diffusion as examples of absolute reaction
degrees Celsius Sc (v/D) rates, J. Chem. Phys., 4, 283-291, 1936.
Havilcek, L. L., and R. D. Crain, Practical Statistics for the
5 2027 Physical Sciences, 489 pp., American Chemical Society, Wash-
10 1537 ington, D.C., 1988.
15 1173 Hayduk, W., and H. Laudie, Prediction of diffusion coefficientsfor
20 920 nonelectrolytes in dilute aqueous solutions, AIChE J., 20, 611-
25 720 615, 1974.
30 577 Jfihne, B., G. Heinz, and W. Dietrich, Measurement of the diffusion
coefficientsof sparingly soluble gasesin water, J. Geophys. Res.,
92, 10,767-10,776, 1987a.
Laudie correlation. Despite the large differences between Jfihne, B., O. Munnich, R. Bosinger, A. Dutzi, W. Huber, and P.
those estimates and the experimentally determined diffusiv- Libner, On the parameters influencing air-water gas exchange, J.
Geophys. Res., 92, 1937-1949, 1987b.
ities (7-28%), the global flux of DMS is not greatly affected.
Korson, L., W. Drost-Hansen, and F. J. Millero, Viscosity of water
Since the square root of the diffusivity is used in the flux at various temperatures, J. Phys. Chem., 73, 34-39, 1969.
calculation, the global flux should increase on the order of Langdon, A. C., and H. C. Thomas, Self-diffusion studies of gel
4-5%. A correction to the global flux estimate of these hydration and the obstruction effect, J. Phys. Chem., 75, 1821-
studies cannot be made with only a single calculation, 1826, 1971.
Liss, P.S., Processesof gas exchange acrossan air-water interface,
because the difference in diffusivity is temperature depen- Deep Sea Res., 20, 221-238, 1973.
dent. The correction must be made for the flux at individual Liss, P.S., and L. Merlivat, Air-sea gas exchange rates: Introduc-
locations and then factored into the global flux estimate. tion and synthesis, in The Role of Air-Sea Exchange in Geochem-
ical Cycling, edited by P. Buat-Menard, pp. 113-127, D. Reidel,
Norwell, Mass., 1986.
SUMMARY Liss, P.S., and P. G. Slater, Flux of gases across the air-sea
interface, Nature, 247, 181-184, 1974.
In this study the diffusion coefficient of DMS in pure water
Marquardt, D. W., An algorithm for least-squares estimation of
was experimentally determined in order to provide a basis nonlinear parameters, J. Soc. Ind. Appl. Math., Set. A, 11(2),
for sea-to-air gas exchange calculations. The measured dif- 431-441, 1963.
fusivities agree reasonably well with empirical estimates, Millero, F. J., Seawater as a multicomponentelectrolyte solution, in
with the closestagreementprovided by the Wilke and Chang The Sea, vol. 5, edited by E. D. Goldberg, pp. 3-80, John Wiley,
New York, 1974.
[1955] correlation using the original solvent association Millero, F. J., and A. Poisson, International one-atmosphere equa-
factor. Diffusivities calculated using the Hayduk and Laudie tion of state of seawater, Deep Sea Res., Part A, 28, 625-629,
[1974] expression, which are commonly used in the gas 1981.
exchange literature, were lower than the experimental re- Peng, T. H., W. S. Broecker, G. G. Mathieu, and Y.-H. Li, Radon
evasion rates in the Atlantic and Pacific oceans as determined
sults by 7-28%, depending on the temperature. Using the
during the GEOSECS program, J. Geophys. Res., 84, 2471-2486,
measured diffusivities results in an increase in the global 1979.
DMS sea-to-air flux estimate of approximately 5%. The Rackett, H. G., Equation of state for saturated liquids, J. Chem.
diffusivities are used to derive a set of Schmidt numbers for Eng. Data, 15, 514-517, 1970.
DMS in seawater which are recommended for use in gas Schantz, E. J., and M. A. Lauffer, Diffusion measurements in agar
gel, Biochemistry, 1,658-663, 1962.
exchange calculations. Smethie, W. M., Jr., T. Takahashi, D. W. Chipman, and J. R.
Ledwell, Gas exchange and CO2 flux in the tropical Atlantic
Oceandetermined
from222Rn
andpCO2 measurements,
J. Geo-
Acknowledgments. The authors would like to thank John Plane, phys. Res., 90, 7005-7022, 1985.
Frank Millero, Rik Wanninkhof, Rana Fine, Tim Bates, and Bill Spalding, G. E., A sensitivemethod for measuringdiffusion coeffi-
Asher for helpful scientific discussion during the course of this cients in agarosegels of electrolyte solutions,J. Phys. Chem., 73,
work. This research was supported by the National Science Foun- 3380-3383, 1969.
dation (Grant ATM 9120498) and the Swedish Natural Science Spencer,C. F., and R. P. Danner, Improved equationfor prediction
Research Council (Contract E-EG/GU 9906-303). We also would of liquid density, J. Chem. Eng. Data, 17, 236-241, 1972.
like to thank Henning Rodhe for his support of this collaborative Wanninkhof, R., J. R. Ledwell, and W. S. Broecker, Gas exchange-
effort and the International Meteorological Institute in Stockholm, wind speed relation measured with sulfur hexafluoride on a lake,
Sweden, for providing travel funds. Science, 227, 1224-1226, 1985.
Watson, A. J., R. C. Upstill-Goddard, and P.S. Liss, Air-sea gas
REFERENCES exchange and stormy seas measured by a dual-tracer technique,
Nature, 349, 145-147, 1991.
Andreae, M. O., Ocean-atmosphere interactions in the global bio- Wiesenburg,D. A., and N. L. Guinasso,Jr., Equilibrium solubilities
geochemical sulfur cycle, Mar. Chem., 30, 1-29, 1990. of CH4, CO, and H 2 in water and sea water, J. Chem. and Eng.
Barrer, R. M., Diffusion in and Through Solids, 464 pp., Cambridge Data, 24, 356-360, 1979.
University Press, New York, 1941. Wilke, C. R., and P. Chang, Correlation of diffusion coefficientsin
Barrer,R. M., J. A. Barrie,andM. G. Rogers,Permeation
through dilute solutions, AIChE J., 1,264-270, 1955.
a membrane with mixed boundary conditions, Trans. Faraday
Soc., 58, 2473-2483, 1962. K. Holmen and C. Leck, Department of Meteorology, Arrhenius-
Bates, T. S., J. D. Cline, R. H. Gammon, and S. R. Kelly-Hansen, laboratory, Stockholm University, Stockholm S-106 91, Sweden.
Regional and seasonalvariations in the flux of oceanic dimethyl- D. B. King and E. S. Saltzman, RosenstielSchool of Marine and
sulfide to the atmosphere, J. Geophys. Res., 92, 2930-2938, 1987. Atmospheric Science, University of Miami, Miami, FL 33149.
Crank, J., The Mathematics of Diffusion, 414 pp., Oxford Univer-
sity Press, New York, 1975.
Dacey, J. W. H., S. G. Wakeham, and B. L. Howes, Henry's law (Received January 14, 1993;
constants for dimethylsulfide in freshwater and seawater, Geo- revised July 7, 1993;
phys. Res. Lett., 11,991-994, 1984. accepted July 8, 1993.)