Photoreductive Dissolution of IronOxides Trapped in Ice and ItsEnvironmental ImplicationsK I T A E K I M , † W O N Y O N G C H O I , * , †

M I C H A E L R . H O F F M A N N , ‡ H O - I L Y O O N , §

A N D B Y O N G - K W O N P A R K §

School of Environmental Science and Engineering, PohangUniversity of Science and Technology (POSTECH),Pohang 790-784, South Korea; W. M. Keck Laboratories,California Institute of Technology, Pasadena, California,United States of America; and Korea Polar Research Institute,Incheon, South Korea

Received December 14, 2009. Revised manuscript receivedApril 15, 2010. Accepted April 20, 2010.

The availability of iron has been thought to be a mainlimiting factor for the productivity of phytoplankton and relatedwith the uptake of atmospheric CO2 and algal blooms infresh and sea waters. In this work, the formation of bioavailableiron (Fe(II)aq) from the dissolution of iron oxide particles wasinvestigated in the ice phase under both UV and visible lightirradiation.Thephotoreductivedissolutionof ironoxidesproceededslowly in aqueous solution (pH 3.5) but was significantlyaccelerated in polycrystalline ice, subsequently releasing morebioavailable ferrous iron upon thawing. The enhancedphotogeneration of Fe(II)aq in ice was confirmed regardless ofthe type of iron oxides [hematite, maghemite (γ-Fe2O3),goethite (R-FeOOH)] and the kind of electron donors. The ice-enhanced dissolution of iron oxides was also observedunder visible light irradiation, although the dissolution ratewas much slower compared with the case of UV radiation.The iron oxide particles and organic electron donors (if any) inice are concentrated and aggregated in the liquid-like grainboundary region (freeze concentration effect) where protonsare also highly concentrated (lower pH). The enhancedphotodissolution of iron oxides should occur in this confinedboundary region. We hypothesized that electron hopping throughthe interconnected grain boundaries of iron oxide particlesfacilitates the separation of photoinduced charge pairs. Theoutdoor experiments carried out under ambient solar radiationof Ny-Ålesund (Svalbard, 78°55′N) also showed that thegeneration of dissolved Fe(II)aq via photoreductive dissolutionis enhanced when iron oxides are trapped in ice. Our resultsimply that the ice(snow)-covered surfaces and ice-cloud particlescontaining iron-rich mineral dusts in the polar and coldenvironments provide a source of bioavailable iron when theythaw.

IntroductionThe role of bioavailable iron in aquatic and marine environ-ments is critical in biota (1-4). Martin’s (5, 6) iron-fertilizationhypothesis has been tested in mesoscale field experimentsthat yield enhanced primary productivity (7), although theextent of the CO2 uptake (8, 9) triggered by iron addition isthe subject of much debate (10). The dominant input of ironto the open ocean is supplied by aeolian dusts that aretransported mainly from the desert area (11-13). Thebioavailable iron largely comes from the light-induceddissolution occurring on iron (ferric) (oxyhydr)oxide (acomponent of mineral dust and ambient aerosol) (1, 14-16).For example, Finden et al. (16) reported that particulate ironoxides provide bioavailable iron via photoreductive dissolu-tion for phytoplankton growth.

Previous reports on the photoreductive dissolution of iron(oxyhydr)oxides in the aqueous phase indicate that the ratesof reaction depend on the incident light intensity, wavelengthrange, concentration and specific nature of the electrondonors in solution, pH, ionic strength, and the crystallinephase of the iron (oxyhydr)oxides (e.g., goethite, hematite)(17-19). In this work, we focus our attention on thephotochemical reaction of iron (oxyhydr)oxides trapped inice and its possible contribution to the redox cycle of ironunder sunlight. The iron-containing dust particles trappedin ice, snow, and frozen cloud aerosols may undergophotoreductive dissolution under solar light and release morebioavailable iron upon thawing. The heterogeneous pho-tochemical reactions taking place in polycrystalline ice maydiffer dramatically from those in the aqueous phase. Mostchemical reactions in aqueous solution are generally slowedat lower temperature. However, many bimolecular chemicalreactions are reported to be accelerated by freezing abovethe eutectic point, in which the solutes are concentrated asthey are excluded from crystalline ice into a liquid-like grainboundary region within the polycrystalline ice matrix (freezeconcentration effect) (20-24). For example, Takenaka et al.(20, 21) found that the oxidation of nitrite ion (NO2

-) withdissolved oxygen to form nitrate (NO3

-) was accelerated bya factor of 105 during freezing of aqueous nitrite solutions.It has also been reported that the photochemical processestaking place in ice can be very different from those in theaqueous phase and might play an important role in thechemical transformation in cold environments (25, 26). Thisstudy investigated the photoreductive dissolution of variousiron (oxyhydr)oxides in ice (typically around pH 3.5 thatrepresents cloudwater condition) under UV or visible lightirradiation and demonstrated that the heterogeneous icephotochemical process is significantly different from theaqueous counterpart. The experimental parameters affectingthe ice-enhanced photochemical process, the mechanism,and the environmental implications are discussed.

Experimental SectionMaterials. Optically transparentR-Fe2O3 (hematite) colloidswere prepared from the hydrolysis of FeCl3 following themethod described by Faust et al. (27). The as-preparedaqueous colloid (800 mg ·Fe2O3/L) was diluted to 8-16 mg/Lin most experiments. The diluted hematite colloid with pH3.5 and 600 µM of chlorides was stable without coagulationfor 10 days. The synthesized hematite phase was confirmedby X-ray diffraction analysis (PANalytical X’Pert diffracto-meter with an X’Celerator detector) using Cu KR radiation(see Supporting Information, Figure S1(a)). The electronicbandgap transition of the synthesized hematite was observed

* Corresponding author phone: +82-54-279-2283; fax: +82-54-279-8299; e-mail: wchoi@postech.edu.

† School of Environmental Science and Engineering, PohangUniversity of Science and Technology (POSTECH).

‡ W. M. Keck Laboratories, California Institute of Technology.§ Korea Polar Research Institute.

Environ. Sci. Technol. 2010, 44, 4142–4148

4142 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010 10.1021/es9037808 2010 American Chemical SocietyPublished on Web 05/06/2010

at the onset of 550 nm in the UV-visible absorption spectrum(see Supporting Information, Figure S1(b)). The primarycolloidal particle sizes were in the range of 10-40 nm asobserved by high resolution field emission transmissionelectron microscopy (HR-FETEM). Other commercially avail-able iron oxides of hematite (Aldrich, BET surface area 8m2/g), maghemite (Aldrich, BET surface area 36 m2/g), andgoethite (Aldrich, BET surface area 185 m2/g) were also usedand compared with the colloidal hematite for their photo-dissolution reactions. Commercial iron oxides were groundinto fine powder before preparing the stock aqueoussuspension to increase the dispersion. Suwannee River fulvicand humic acids were purchased from the InternationalHumic Substances Society (http://www.ihss.gatech.edu).

Photolysis. Aqueous suspension/colloid of iron oxides(10 mL) with a desired concentration (8-16 mg/L) was placedin a quartz tube (12 × 125-mm), sealed with septa, andsolidified in an ethanol bath cooled at -20 °C. The tem-perature of the freezing bath was gradually lowered from -5to-20 °C within 30 min to prevent the breakage of the tubes.After freezing, the colloidal hematite was uniformly dispersedalong the ice column without aggregation while the com-mercial iron oxide powder was slightly concentrated in thelower part of the ice column because of the gravitationalsettling during freezing. The whole ice column was irradiatedby the lamp after the freezing process. Sample tubes(maximum 16 tubes) were located in a merry-go-roundphotolysis reactor that was rotated at a constant speed(0.8-1.0 rpm) around a 100-W mercury lamp (Ace Glass Inc.)for uniform irradiation. Light was filtered by a pyrex jacket(transmitting λ > 300 nm; see Supporting Information FigureS1(b)) surrounding the mercury lamp that was immersed inthe ethanol bath. Visible light irradiation experiments useda 100-W halogen lamp (Phillips). A 300-W Xe arc lamp withvertical beam turning assemblies was also used for naturalsunlight simulation. In this case, samples were placed inPetri dishes covered with quartz tops and solidified using aPeltier device (Ace-Tec, Korea). All irradiated ice sampleswere thawed for sampling and the subsequent analysis forFe2+. Aqueous photochemical experiments of iron oxidesdissolution were also carried out as a control at 25 °C usingthe same experimental setup.

Analysis. Fe2+ concentrations were measured spectro-photometrically by the 1,10-phenanthroline method (28).One-milliliter aliquots of the sample solution were withdrawnfrom the quartz reaction tube after thawing the ice sample.The liquid aliquot was filtered with a 0.45-µm filter to removeiron oxides particles. Because of their aggregated state, noparticles passed through the 0.45-µm filter. The filtrate wasthen added to an amber vial containing 2 mL of 0.2% 1,10-phenanthroline, 1.5 mL of an buffer, which was diluted withwater to the total volume of 10 mL. The vial was mixedvigorously and kept for 1 h in the dark before the analysis.The absorbance measurements at 510 nm (εm ) 2.5 × 104 Lmol-1 cm-1) were done using a UV-visible spectrophotom-eter (UV-2401PC Shimadzu). The absorbance calibration forthe determination of [Fe2+] was carried out using a freshlyprepared ferrous solution (Fe(ClO4)2 ·6H2O). The total dis-solved iron (Fe2+ and Fe3+) concentration was determinedby atomic absorption spectroscopy (AAS, SpectrAA-800).

Ferrioxalate actinometry was used for comparing the UVlight intensity absorbed by aqueous and ice samples (29).The ferrioxalate solution was solidified prior to irradiation,irradiated under the same conditions as the iron oxidephotolysis, and then thawed in the dark for the analysis ofFe2+.

The optical images of ice were obtained with a ZeissJENALAB-pol polarizing microscope equipped with a LinkamLTS 350 thermal stage (temperature range of-196 to 350 °C)and a Linkam LNP94 liquid nitrogen pump. One or two

droplets of sample solutions were dropped on the cover glassand put on the stage. The desired temperature (-20 °C) ofthe airtight stage containing sample solution was controlledby the liquid nitrogen pump. It took 2 min to cool the stageat -20 °C.

Outdoor Experiments. The photodissolution experimentswere also carried out under ambient solar radiation. Sincethe freezing temperature is not attained and the solar fluxis highly varying in our geographical location and season,the outdoor experiments were conducted in the arctic region,Ny-Ålesund (78°55′N, 11°56′E, sea level) from May 14 to 28,2009 where sunlight is available 24 h a day. Quartz tubescontaining the various iron oxides with electron donors werefrozen in a refrigerator before being exposed to sunlight. Thesolidified ice samples were placed horizontally on the surfaceof ambient snow for exposure to incident solar radiation.The irradiated samples remained solid during the entireexposure to sunlight. Control photolyses of aqueous samplescontaining iron oxides were carried out simultaneously underthe same irradiation conditions. In order to prevent thefreezing of the aqueous samples under ambient exposure,samples were placed on an electrically heated mat on thesnow. The concentrations of photogenerated Fe2+ wereimmediately determined after solar irradiation in the KoreaPolar Research Institute (KOPRI) DASAN Station (Ny-Åle-sund). The ambient temperature ranged between -11 and-1 °C. The integrated solar irradiance as measured at theKoldewey Station in Ny-Ålesund varied from 1.8 to 34.1 W/m2

for UV band of 300 < λ < 370 nm, depending on the angularposition of the sun and the weather condition, with a dailyaverage of 11.0 W/m2 (corresponding to about 31 µEinsteinm-2sec-1 assuming 335 nm photons).

Results and DiscussionFormation of Fe(II)aq via Photoreductive Dissolution of IronOxides. We conducted a series of heterogeneous photo-chemical experiments with hematite (R-Fe2O3: nanosizedcolloidal sol synthesized in the laboratory) to quantify solubleFe2+ generated in ice under conditions characteristic ofambient solar radiation. We found that the photodissolutionof hematite under UV irradiation (mercury lamp) of λ > 300nm, which occurs very slowly in aerated aqueous solutionsof pH 3.5, is significantly accelerated in the polycrystallineice phase both in the absence and presence of an addedelectron donor, as shown in Figure 1. The addition of variouselectron donors markedly enhanced the dissolution ofhematite in ice, while their presence little influenced thephotodissolution in water except for oxalic acid (Figure 1b).In all cases, the dissolution of hematite was enhanced in theice phase compared to the aqueous counterpart reactionsthat were performed under the same irradiation conditions.In the absence of light (i.e., dark control), there was nodetectable generation of Fe2+ in both ice and aqueous phases.The photodissolution of hematite in water and ice was carriedout under Xe-arc lamp irradiation (solar simulating condition)as well and the same trend was observed. The concentrationsof Fe2+ generated from hematite after 12 h irradiation (Xelamp) with formic acid, citric acid, and oxalic acid (600 µMeach) were 5, 23, and 8 µM, in aqueous solution; 24, 52, and41 µM, in ice, respectively.

To test whether there is the ice-induced dissolution ofiron oxides to generate Fe3+ ions in the dark, which can befollowed by their photoreduction to Fe2+, the followingcontrol experiment was done. First, the hematite solutionwas frozen then thawed in the dark, which was immediatelyfollowed by UV irradiation. The production of Fe2+ was aslittle as that of the aqueous photodissolution. Furthermore,we compared the concentration of Fe2+ (determined by the1,10-phenanthroline method) with the total dissolved iron(Fe2+ plus Fe3+ determined by AAS) and found little difference

VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4143

between them. This excludes the possibility of the ice-induced(photo)dissolution of Fe3+. Therefore, the mechanism of theice-induced dissolution (in the dark) of iron oxides and thesubsequent photoreduction of Fe3+ to Fe2+ can be dismissed.

The photodissolution reactions were also investigatedusing commercially available iron oxide powders havingdifferent particle size and crystallinity [hematite, maghemite(γ-Fe2O3), goethite (R-FeOOH)]. The dark dissolution wasnegligible for all samples. The enhanced photoactivities inice were also confirmed regardless of the type of iron oxidestrapped in ice (Figure 1c). The photodissolution of iron oxidesin water and ice was also carried out under visible light

irradiation (λ > 400 nm) because the majority of sunlight isin the visible region. Although the dissolution rates weremuch slower than under UV irradiation as previously reported(30), the trend remained the same. Figure 2 shows that thevisible light-induced formation of Fe2+ from the dissolutionof colloidal hematite was clearly enhanced in ice.

The light intensities absorbed by the aqueous and icesamples were measured using the ferrioxalate actinometry.The intensities of the UV light absorbed by the aqueous andice samples were 3.2 × 10-4 and 1.5 × 10-4 einstein min-1

L-1 (λ > 300 nm), respectively, although the incident lightflux should be the same (see Supporting Information, FigureS2). The light intensity absorbed by the ice sample shouldbe taken only as a rough estimate because the actinometryin ice phase has some uncertainty. The light intensityabsorbed in the ice sample was about half of that in aqueouswater because the turbid (white) ice sample scatters out theincident light. Therefore, the photodissolution rate of ironoxides trapped in ice is even higher than that in water despitemuch reduced light intensity absorbed by the ice sample.

The enhanced photodissolution of iron oxides in ice wasalso confirmed from the TEM images. The hematite colloidparticles were compared before and after the UV irradiation(Figure 3). The sizes of the primary particles in ice weresignificantly reduced after the UV exposure as a result of thephotodissolution, while those in the aqueous phase werelittle changed under the same irradiation condition. Weestimate that the dissolved amount of Fe2+ after 48 h of UVirradiation in ice corresponds to the photodissolution of 30to 40% of the mass of the hematite particles.

The enhanced photodissolution of iron oxides trapped inice may be related with the singular chemical process (e.g.,freeze concentration effect (20-24)) occurring in the ice grain

FIGURE 1. Production of Fe(II)aq via photoreductive dissolutionof iron oxides under UV light (λ > 300 nm). (a) Hematite colloidin water at 25 °C and hematite trapped in ice at -20 °C. Theformic acid was added as an electron donor (ED).[r-Fe2O3(colloid)] ) 8 mg/L. (b) Comparison of Fe2+ productionafter 48 h of UV irradiation in the presence of various organicacids as ED in aqueous solution at 25 °C (black bar) and ice at-20 °C (gray bar). Experimental conditions were as follows:[r-Fe2O3(colloid)] ) 16 mg/L, [ED] ) 600 µM, [humic or fulvicacid] ) 1 ppm, pHi ) 3.2-3.5. (c) Formation of Fe(II)aq viaphotoreductive dissolution of various types of iron oxides(commercially available powder) under 48 h UV irradiation inaqueous solution (black bar) and ice phase (gray bar), [Ironoxide] ) 0.2 g/L, [formic acid] ) 6000 µM, pHi ) 3.5.

FIGURE 2. Photoreductive dissolution of colloidal hematiteunder visible light (λ > 400 nm, 100-W halogen lamp). (a)[r-Fe2O3] ) 8 mg/L, [formic acid] ) 600 µM, pHi ) 3.2-3.5. (b)Comparison of Fe2+ production after 10 days of visible lightirradiation in the presence of various organic acids as ED inaqueous solution at 25 °C (black bar) and ice at -20 °C (graybar). [r-Fe2O3] ) 16 mg/L, [ED] ) 600 µM, pHi ) 3.2-3.5.

4144 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010

boundary region. In this respect, the change in the oxygenavailability in ice should be considered. Takenaka et al. (21)suggested that oxygen is accumulated in the grain boundaryregion during the freezing process. The presence of oxygengenerally inhibits the photoreductive dissolution of ironoxides through either reoxidizing the surface-bound Fe(II)or scavenging conduction band electrons on the photoexcitediron oxide particle (17, 31). This actually implies a retardedphotodissolution of iron oxides in ice, contrary to the

observations of this study. To investigate the role of oxygenin the photochemical process, we carried out the photo-dissolution experiments in both aerated and deaeratedconditions, and the result is shown in Figure 4. The dissolutionof iron oxides was retarded in the presence of oxygen asexpected but the enhanced production of Fe2+ was observedin both aerated and deaerated ice samples. This indicatesthat the enhanced photodissolution of iron oxides in icecannot be related with the role of oxygen.

The microenvironment in the liquid-like grain boundaryregion is characterized by a low pH, which also needs to beconsidered. Even though the present experiments werecarried out by freezing a solution initially at pH 3.5, the pHin the liquid-like grain boundary region is predicted to beeven lower due to the freeze concentration effects of theice-excluded protons. It was previously estimated that thelocal concentration of acids in the ice grain boundary regionis increased by 2-3 orders of magnitude in contrast to theaqueous solution (24, 32). The photoreductive dissolution ofhematite should be enhanced at lower pH (14, 18, 19), asconfirmed in this study: [Fe2+] after 48 h UV irradiation ofhematite in aqueous solution was 5 µM at pH 3.50 and 31µM at pH 1.15 (under the condition of Figure 1b with formicacid). Even at pH 1.15, the dark reductive dissolution ofhematite was insignificant. Therefore, the lower pH in theice grain boundary should be an important factor responsiblefor the enhanced photodissolution in ice. The ice-enhancedphotodissolution effect rapidly decreased with increasingpH. The concentration of Fe2+ generated after 48 h UVirradiation (with formic acid under the condition of Figure1b) was about 83 µM at pHi 3.5, 16 µM at pHi 4.0, and negligibleabove pH 5. The fact that the reoxidation of the photoge-nerated Fe(II) to Fe(III) exponentially increases with raisingpH from 3 to 7 (33) should also limit the photogeneration offerrous ions to acidic conditions.

Outdoor Experiments under Solar Radiation. The pho-todissolution of iron oxides was studied in the arctic region(Ny-Ålesund, Svalbard, 78°55′N) in order to confirm thelaboratory-observed phenomenon under natural environ-mental conditions. The outdoor experimental results aresummarized in Table 1. The production of Fe2+ from thephotodissolution of iron oxides in ice was found to beconsistently higher than that in the corresponding aqueousphase, which confirms the laboratory results. Incidentally,the polar environment (where most surface is covered withice and snow) in which the outdoor experiments were carriedout provides an ideal natural condition where the studiedprocess can actually occur.

Photochemical Mechanisms. The observed photodisso-lution of iron oxides should be a result of the reductive processthat occurs through the photoinduced electron transfer upon

FIGURE 3. TEM images of hematite colloid particles before andafter UV light irradiation with [r-Fe2O3] ) 8 mg/L, formic acid(600 µM), pHi ) 3.5. (a) initial sample, (b) after 48 h irradiationin the aqueous phase (25 °C), (c) after 48 h irradiation in the icephase (-20 °C).

FIGURE 4. Effect of oxygen on the photoreductive dissolution ofhematite under UV irradiation. [r-Fe2O3] ) 16 mg/L, [formicacid] ) 600 µM, pHi ) 3.5, 48 h irradiation.

VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4145

absorbing photons. The photoreductive transformation ofFe(III) to Fe(II) in the oxide lattice is initiated by either theligand-to-metal charge transfer (LMCT) from the adsorbedorganic reductant to Fe(III) species on the surface (eq 1) or thebandgap excitation of iron oxide semiconductor (Eg ) 2.2 eVor 565 nm) followed by the trapping of conduction band electronin the lattice Fe(III) sites (eqs 2,3). Photochemically formedFe(II) at the surface of iron oxides (Fe(II)surf) then desorbs intothe solution (eq 4). The adsorption of Fe(II) on iron oxides shouldbe negligible under the acidic condition where Fe2+ ions areelectrostatically repelled from the positively charged surface(pHzpc ) 7.8-8.3 for hematite).

The present results imply that the above photoinducedelectron transfer processes are accelerated in ice. Freezeconcentration effects increase the concentration of ironoxide particles and organic ligands in the grain boundaryregion and may enhance the surface complexation (eq 1)and the subsequent dissolution of iron oxides. Oxalic acidsthat form strong surface complexes on iron oxides (31, 34)exhibited the highest photodissolution rates in not onlythe aqueous but also the ice phase (Figure 1b). However,whether the LMCT mechanism is working in the very acidicice boundary layer (where the ligand complexation is notvery favored) is uncertain. However, the enhanced pho-todissolution in ice was also observed in the absence ofadded electron donors, which implies that the enhancedLMCT mechanism cannot be the sole reason.

The iron oxide particles confined/concentrated inliquid-like boundary domains in ice might be extensivelyaggregated. The presence of boundary domains at the edgeof ice crystals and the extensive aggregation of hematiteparticles in the boundary edge are clearly observed in theoptical images of Figure 5. Iron oxide particles should beexcluded from the ice lattice into the ice crystal boundariesupon freezing, which concurs with the drastic change ofthe environmental properties around the particles such aspH, ionic strength, and the concentration of solutes andelectron donors. Such conditions force the iron oxideparticles to aggregate. Within the agglomerates of semi-

TABLE 1. Production of Fe2+ under Arctic Sunlight (OutdoorExperiments in Ny-Ålesund)

aqueous(µM)

ice(µM)

various ED(after 72 h irradiation)a

no ED 15 38formic acid 13 73acetic acid 13 51citric acid 33 111oxalic acid 18 100

various iron oxides(after 48 h irradiation)b

synthesized colloidal hematite(R-Fe2O3)

13 82

hematite (R-Fe2O3) 6 19maghemite (γ-Fe2O3) 17 389goethite (R-FeOOH) 35 517a [colloidal R-Fe2O3] ) 16 mg/L, [ED] ) 600 µM. b [Iron

oxide] ) 0.2 g/L, [formic acid] ) 6 mM.

FIGURE 5. Optical images of polycrystalline ice showing the highly aggregated iron oxide particles trapped in ice veins (grainboundary region) at -20 °C. (a) pure ice, (b) ice with maghemite (γ-Fe2O3) (1 g/L), (c) ice with colloidal hematite (0.8 g/L), and (d)schematic illustration of concentrated hematite particles (orange circles) and formic acids (ball-and-stick models) in ice grainboundary region.

≡Fe(III)-L + hv f ≡ Fe(II)-L·+(photoinduced LMCT)(1)

FeIII2O3 + hv f ecb

- + hvb+ (bandgap excitation) (2)

Fe(III)(at lattice or surface site) + ecb- f Fe(II)surf (3)

Fe(II)surf(at lattice or surface site) f Fe(II)aq (4)

4146 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010

conductor nanoparticles, the charge-pair separation canbe facilitated by the electron hopping through the inter-connected grain boundaries and the following interfacialelectron transfer reactions can be enhanced comparedwith those occurring on an isolated particle (35, 36).

Environmental Implications. The present finding thatthe photodissolution of iron oxides can be significantlyenhanced in the ice phase has many implications in thecold environmental condition where ice and snow con-taining mineral dusts are exposed to sunlight. Since theproposed phenomenon is particularly prominent in theacidic condition, acidic ice and snow should be of interest.One example is related to the fate of cloud dropletscontaining iron dusts. The pH of cloudwater is typically3.5-5.5 or lower (12). Atmospheric deposition of mineraldusts supplies much of the nutrient iron to the ocean. TheAeolian mineral dust particles containing iron form cloudcondensation nuclei and undergo freeze-thaw cycles inthe upper troposphere, bringing a more soluble ironfraction upon deposition. For example, one study reportedthat the soluble fraction of iron in aerosols was ∼1% nearsources, but often 10-40% farther away, producing asignificant increase in soluble Fe deposition on remoteocean region (12, 13). The ice-enhanced photodissolutionof iron oxides may contribute to the improved solubilityof iron in acidic aerosols during the upper atmospherictransport. Another example is acidic snow/ice containingiron dusts. A snow/ice layer on sea ice under solar radiationmay contribute to the iron supply into the polar oceanthrough the proposed ice-enhanced photochemical pro-cess. This may be partly responsible for ice-edge algalblooms that are observed during polar springtime (37, 38).Since the photogenerated ferrous ions would be rapidlyreoxidized in circumneutral natural waters, the implica-tions of the ice-induced photodissolution on aquatic biotawould depend on how fast microorganisms can assimilatethe dissolved ferrous ions. In general, the proposedphenomenon can be operating wherever there is acidicsnow/ice and the iron supply through the photochemicalpath is important. However, the actual dissolution of ironoxides should sensitively depend on various experimentalconditions such as pH, crystallinity and surface area ofthe minerals, and electrolyte/organics concentrations.Therefore, extrapolating any laboratory results to naturalenvironmental conditions requires careful confirmation,which can be a subject of further studies.

AcknowledgmentsFunding for this work was provided by KOSEF NRL program(No. R0A-2008-000-20068-0), KOSEF EPB center (No. R11-2008-052-02002), KCAP (Sogang Univ.) funded by NRF (2009-C1AAA001-2009-0093879), and Korea Polar Research Institute(KOPRI). K.K. thanks J. Klanova and P. Klan for their kindsupport of his visit and training at Masaryk University, CzechRepublic. We gratefully acknowledge the contributions of D.Hrazdira and Y. Ahn.

Supporting Information AvailableX-ray diffractogram/UV-visible absorption spectrum ofsynthesized hematite, and the ferrioxalate actinometry forice and aqueous samples. This information is available freeof charge via the Internet at http://pubs.acs.org/.

Literature Cited(1) Turner, D. R.; Hunter, K. A. The Biogeochemistry of Iron in

Seawater. 7th ed.; John Wiley & Sons, Ltd: Chichester,2001.

(2) Falkowski, P. G.; Barber, R. T.; Smetacek, V. Biogeochemicalcontrols and feedbacks on ocean primary production. Science1998, 281, 200–206.

(3) Morel, F. M. M.; Price, N. M. The biogeochemical cycles of tracemetals in the oceans. Science 2003, 300, 944–947.

(4) Boyd, P. W.; Watson, A. J.; Law, C. S.; Abraham, E. R.; Trull, T.;Murdoch, R.; Bakker, D. C. E.; Bowie, A. R.; Buesseler, K. O.;Chang, H.; et al. A mesoscale phytoplankton bloom in the polarSouthern Ocean stimulated by iron fertilization. Nature 2000,407, 695–702.

(5) Martin, J. H.; Fitzwater, S. E. Iron deficiency limits phytoplanktongrowth in the north-east Pacific subarctic. Nature 1988, 331,341–343.

(6) Martin, J. H. Glacial-interglacial CO2 change: The iron hy-pothesis. Paleoceanography 1990, 5, 1–13.

(7) Boyd, P. W; Jickells, T.; Law, C. S.; Blain, S.; Boyle, E. A.; Buesseler,K. O.; Coale, K. H.; Cullen, J. J.; Baar, H. J. W. d.; Follows, M.;Harvey, M.; Lancelot, C.; Levasseur, M.; Owens, N. P. J.; Pollard,R.; Rivkin, R. B.; Sarmiento, J.; Schoemann, V.; Smetacek, V.;Takeda, S.; Tsuda, A.; Turner, S.; Watson, A. J. Mesoscale ironenrichment experiments 1993-2005: Synthesis and futuredirection. Science 2007, 315, 612–617.

(8) Baar, H. J. W. d.; Jong, J. T. M. d.; Bakker, D. C. E.; Loscher, B. M.;Veth, C.; Bathmann, U.; Smetacek, V. Importance of iron forplankton blooms and carbon dioxide drawdown in the SouthernOcean. Nature 1995, 373, 412–415.

(9) Chisholm, S. W. Stirring times in the Southern Ocean. Nature2000, 407, 685–686.

(10) Blain, S.; Queguiner, B.; Armand, L.; Belviso, S.; Bombled, B.;Bopp, L. Effect of natural iron fertilization on carbonsequestration in the Southern Ocean. Nature 2007, 446, 1070–1075.

(11) Jichells, T. D.; An, Z. S.; Andersen, K. K.; Baker, A. R.; Bergametti,G.; Brooks, N.; Cao, J. J. Global iron connection between desertdust, ocean biogeochemistry, and climate. Science 2005, 308,66–71.

(12) Zhuang, G.; Yi, Z.; Duce, R. A.; Brown, P. R. Link between ironand sulphur cycles suggested by detection of ferrous in remotemarine aerosols. Nature 1992, 355, 537–539.

(13) Fan, S.-M.; Moxim, W. J.; II, H. L. Aeolian input of bioavailableiron to the ocean. Geophys. Res. Lett. 2006, 33, L07602.

(14) Cornell, R. M.; Schwertmann, U. The Iron Oxides: Structure,Properties, Reactions, Occurrence and Uses. VCH Publishers: NewYork, NY (USA): 1996.

(15) Johnson, K. S.; Coale, K. H.; Elrod, V. A.; Neil, W, T. Ironphotochemistry in seawater from the equatorial Pacific. Mar.Chem. 1994, 46, 319–334.

(16) Finden, D. A. S.; Tipping, E.; Jaworski, G. H. M.; Reynolds,C. S. Light-induced reduction of natural iron(III) oxide andits relevance to phytoplankton. Nature 1984, 309, 783-784.

(17) Sulzberger, B.; Laubscher, H. Reactivity of various types ofiron(hydr)oxides toward light-induced dissolution. Mar. Chem.1995, 50, 103–115.

(18) Pehkonen, S. O.; Slefert, R.; Erel, Y.; Webb, S.; Hoffmann, M. R.Photoreduction of iron oxyhydroxides in the presence ofimportant atmospheric organic compounds. Environ. Sci.Technol. 1993, 27, 2056–2062.

(19) Waite, T. D.; Morel, F. M. M. Photoreductive dissolution ofcolloidal iron oxides in natural waters. Environ. Sci. Technol.1984, 18, 860–868.

(20) Takenaka, N.; Ueda, A.; Maeda, Y. Acceleration of the rate ofnitrite oxidation by freezing in aqueous solution. Nature 1992,358, 736–738.

(21) Takenaka, N.; Ueda, A.; Daimon, T.; Bandow, H.; Dohmaru, T.;Maeda, Y. Acceleration mechanism of chemical reaction byfreezing: the reaction of nitrous acid with dissolved oxygen. J.Phys. Chem. 1996, 100, 13874–13884.

(22) Heger, D.; Jirkovsky, J.; Klan, P. Aggregation of methylene bluein frozen aqueous solutions studied by absorption spectroscopy.J. Phys. Chem. A 2005, 109, 6702–6709.

(23) Betterton, E. A.; Anderson, D. J. Autoxidation of N(III), S(IV),and other species in frozen solutionsA possible pathway forenhanced chemical transformation in freezing systems. J. Atmos.Chem. 2001, 40, 171–189.

(24) Grannas, A. M.; Jones, A. E.; Dibb, J.; Ammann, M.; Anastasio,C.; Beine, H. J.; Bergin, M.; Bottenheim, J.; Boxe, C. S.; Carver,G.; Chen, G.; Crawford, J. H.; Domin’e, F.; Frey, M. M.; Guzm’an,M. I.; Heard, D. E.; Helmig, D.; Hoffmann, M. R.; Honrath, R. E.;Huey, L. G.; Hutterli, M.; Jacobi, H. W.; Kl’an, P.; Lefer, B.;McConnell, J.; Plane, J.; Sander, R.; Savarino, J.; Shepson, P. B.;Simpson, W. R.; Sodeau, J. R.; Glasow, R. v.; Weller, R.; Wolff,E. W.; Zhu, T. An overview of snow photochemistry: evidence,mechanisms and impacts. Atmos. Chem. Phys. 2007, 7, 4329–4373.

VOL. 44, NO. 11, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4147

(25) Klanova, J.; Klan, P.; Nosek, J.; Holoubek, I. Environmental icephotochemistry: Monochlorophenols. Environ. Sci. Technol.2003, 37, 1568–1574.

(26) Blaha, L.; Klanova, J.; Klan, P.; Janosek, J.; MichalSkarek; Ruzicka,R. Toxicity increases in ice containing monochlorophenols uponphotolysis: environmental consequences. Environ. Sci. Technol.2004, 38, 2873–2878.

(27) Faust, B. C.; Hoffmann, M. R.; Bahnemann, D. W. Photocatalyticoxidation of sulfur dioxide in aqueous suspensions of iron oxide(R-Fe2O3). J. Phys. Chem. 1989, 93, 6371–6381.

(28) Fortune, W. B.; Mellon, M. G. Determination of iron witho-phenanthroline. Ind. Eng. Chem., Anal. Ed. 1938, 10, 60–64.

(29) Hatchard, C. G.; Parker, C. A. A new sensitive chemicalactinometer. II. Potassium ferrioxalate as a standard chemicalactinometer. Proc. R. Soc. London, A 1956, 235, 518–536.

(30) David, F.; David, P. G. Photoredox chemistry of Iron(III) chlorideand Iron(III) perchlorate in aqueous media. a comparative study.J. Phys. Chem. 1976, 80, 579–583.

(31) Siffert, C.; Sulzberger, B. Light-induced dissolution of hematitein the presence of oxalate: A case study. Langmuir 1991, 7,1627–1634.

(32) Heger, D.; Klanova, J.; Klan, P. Enhanced protonation of cresolred in acidic aqueous solutions caused by freezing. J. Phys. Chem.B 2006, 110, 1277–1287.

(33) Nordstrom, D. K.; Alpers, C. N. Negative pH, efflorescentmineralogy, and consequences for environmental restorationat the Iron Mountain Superfund site, California. Proc. Natl. Acad.Sci. 1999, 96, 3455–3462.

(34) Banwart, S.; Davids, S.; Stumm, W. The role of oxalate inaccelerating the reductive dissolution of Hematite by ascorbate.Colloids Surf. 1989, 39, 303–309.

(35) Wang, C.-Y.; Bottcher, C.; Bahnemann, D. W.; Dohrmann, J. K.A comparative study of nanometer sized Fe(III)-doped TiO2

photocatalysts: Synthesis, characterization and activity. J. Mater.Chem. 2003, 13, 2322–2329.

(36) Lakshminarasimhan, N.; Kim, W.; Choi, W. Effect of theagglomerated state on the photocatalytic hydrogen productionwith In situ agglomeration of colloidal TiO2 nanoparticles. J.Phys. Chem. C 2008, 112, 20451–20457.

(37) Aguilar-Islas, A. M.; Rember, R. D.; Mordy, C. W.; Wu, J. Seaice-derived dissolved iron and its potential influence on thespring algal bloom in the Bering Sea. Geophys. Res. Lett. 2008,35, L24601.

(38) Sedwick, P. N.; DiTullio, G. R. Regulation of algal blooms inAntarctic shelf water by the release of iron from melting sea icewaters. Geophys. Res. Lett. 1997, 24, 2515–2518.

ES9037808

4148 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 11, 2010