Soil surface acidity plays a determining role in theatmospheric-terrestrial exchange of nitrous acidMelissa A. Donaldsona, David L. Bishb, and Jonathan D. Raffa,c,1

aSchool of Public and Environmental Affairs, Indiana University, Bloomington, IN 47405-2204; bDepartment of Geological Sciences, Indiana University,Bloomington, IN 47405-1405; and cDepartment of Chemistry, Indiana University, Bloomington, IN 47405-7102

Edited* by Barbara J. Finlayson-Pitts, University of California, Irvine, CA, and approved November 21, 2014 (received for review September 25, 2014)

Nitrous acid (HONO) is an important hydroxyl (OH) radical sourcethat is formed on both ground and aerosol surfaces in the well-mixed boundary layer. Recent studies report the release of HONOfrom nonacidic soils, although it is unclear how soil that is morebasic than the pKa of HONO (∼3) is capable of protonating soilnitrite to serve as an atmospheric HONO source. Here, we useda coated-wall flow tube and chemical ionization mass spectrome-try (CIMS) to study the pH dependence of HONO uptake ontoagricultural soil and model substrates under atmospherically rele-vant conditions (1 atm and 30% relative humidity). Experimentsmeasuring the evolution of HONO from pH-adjusted surfaces treatedwith nitrite and potentiometric titrations of the substrates show,to our knowledge for the first time, that surface acidity ratherthan bulk aqueous pH determines HONO uptake and desorptionefficiency on soil, in a process controlled by amphoteric aluminumand iron (hydr)oxides present. The results have important implica-tions for predicting when soil nitrite, whether microbially derivedor atmospherically deposited, will act as a net source or sink ofatmospheric HONO. This process represents an unrecognizedmechanism of HONO release from soil that will contribute toHONO emissions throughout the day.

nitrous acid | nitrite | surface acidity | air pollution | nitrogen cycle

Nitrous acid (HONO) plays a significant role in regulating theoxidative capacity of the atmosphere as it is readily photo-

lyzed to produce nitric oxide (NO) and hydroxyl radical (OH)(1). The latter product, OH, is an important atmospheric oxidantthat leads to ozone and secondary aerosol formation (2, 3).Despite extensive research over three decades (4–12), HONOformation and loss processes are not completely understood, andcurrent atmospheric models do not accurately predict nighttimeand daytime levels (13–18). Models and field measurements ofvertical HONO gradients indicate that reactions on groundsurfaces may explain the inconsistencies (12, 13, 15, 19–21) andthat the hydrolysis of NO2 on these surfaces is the most impor-tant HONO source at night (22–25) that likely continues throughthe daytime (19). Recent investigations demonstrate that nitritederived from microbial activity is a source of atmospheric HONOfrom nonacidic soil (i.e., soil pH >5) (26, 27). However, themechanism responsible for HONO release from soil nitrite re-mains unclear because the pKa of HONO in bulk water is ∼3(28–30); this suggests that our understanding of bulk aqueousequilibria may not be applicable to understanding the speciationof soil-adsorbed N(III) species [N(III) = NO2ˉ, HONO,H2ONO+]. Here we report that the surface acidity of boundarylayer minerals plays a critical role in the ability of a surface torelease HONO or sequester it as nitrite (NO2ˉ) or the nitro-acidium ion (H2ONO+) on the surface. We also demonstrate thedisproportionate importance of aluminum and iron (hydr)oxidesin controlling HONO uptake and release on soil surfaces.In our previous study of HONO uptake onto soil surfaces at

varied relative humidity (RH), we observed that as the RH in-creased above 30%, the mineral surfaces were covered by theequivalent of a monolayer of adsorbed H2O and that masstransfer from the gas phase into the liquid phase and subsequent

acid–base equilibria control the net uptake of HONO onto soilunder atmospherically relevant conditions (31–33). It thereforeseems reasonable that soil pH is an important variable in de-termining the uptake of HONO that deserves further attention. Soilis a complex mixture of amphoteric minerals, including aluminumand iron (hydr)oxides whose surface reactivity is pH dependentand can differ greatly from that measured in bulk water (34).Surface metal oxides are coordinatively unsaturated, and chemi-sorption of H2O leads to the formation of neutral M–OH andcharged M–Oˉ or M–OH2

+ groups (where M = Al3+ and Fe3+),depending on the pH of the environment (Fig. 1B) (35–37).Surface M–OH2

+ groups in particular are acidic due to the abilityof the metal cation to withdraw electrons from the bound H2Oand weaken the O–H bond. By analogy, it is well known thataqua complexes of metal cations such as [Fe(OH2)6]

3+ and[Al(OH2)6]

3+ readily donate protons, an effect that is attenuatedby outer-sphere hydration (hydrogen bonding) (38).The existence of chemisorbed H2O in the soil system means

that HONO and NO2ˉ not only interact with species such asH3O

+ in bulk water as shown in Fig. 1A, but they may also un-dergo proton transfer reactions with mineral substrates (Fig. 1B);if this is true, then it is possible that the apparent pKa of theNO2ˉ/HONO system could be influenced by the availability ofsurface protons, whose surface density may not necessarily bereflected by the bulk pH. In one example of how surface acidityis different from bulk acid–base equilibria, the pKa of ammoniawas shown to be lowered from 9.3 in bulk water to 8.4 on a hy-drated hematite interface (39).We propose that the degree of protonation of amphoteric soil

mineral surfaces yielding M–Oˉ, M–OH, and M–OH2+ deter-

mines the pH dependence of HONO uptake onto soil mineralsurfaces rather than its pKa in bulk water. To test this hypothesis,

Significance

Nitrous acid (HONO) emitted from soil plays an important rolein regulating the oxidizing capacity of the atmosphere. Un-expectedly high emissions of HONO are observed from soilswith close to neutral pH, where aqueous acid–base and Henry’slaw equilibria would predict that nitrite is the dominant ni-trogen species. We show that surface acidity of soil mineralsrather than aqueous acid–base equilibria controls nitrite spe-ciation in soil, implying that up to 70% of global soils are ca-pable of emitting HONO due to their acidic or close to neutralpH; this suggests that soil nitrite, whether microbially derivedor otherwise deposited, likely plays a more widespread role interrestrial–atmospheric cycling of nitrogen affecting air pollu-tion and climate than previously thought.

Author contributions: M.A.D. and J.D.R. designed research; M.A.D. performed research;M.A.D., D.L.B., and J.D.R. analyzed data; and M.A.D., D.L.B., and J.D.R. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.1To whom correspondence should be addressed. Email: jdraff@indiana.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1418545112/-/DCSupplemental.

18472–18477 | PNAS | December 30, 2014 | vol. 111 | no. 52 www.pnas.org/cgi/doi/10.1073/pnas.1418545112

Dow

nloa

ded

by g

uest

on

May

20,

202

0

we measured the uptake coefficients for HONO adsorption asa function of soil pH and monitored the evolution of HONOfrom nitrite additions to pH-adjusted soil and other model sub-strates under atmospherically relevant conditions of pressure(1 atm), HONOmixing ratios (10 ppb), and relative humidity (30%RH). In addition, we used potentiometric titrations to determinethe surface charge of each substrate to better understand the re-lationship between surface acidity and the speciation of N(III) inthe soil system. This study provides a fundamental understandingof how atmospherically significant emissions of HONO arise fromsoils with close to neutral pH where traditional aqueous phaseequilibria would predict that nitrite is the dominant nitrogenspecies. To our knowledge, the surface acidity of soil minerals hasnever before been linked to the pH-dependent exchange ofHONO between the terrestrial and atmospheric environments.The results have important implications for predicting when soilNO2ˉ, whether microbially derived or otherwise deposited, willact as a net source or sink of atmospheric HONO. Including thismechanism within parameterizations of HONO sources will im-prove the accuracy of atmospheric models used for air pollutionprediction and control.

Results and DiscussionDependence of HONO Uptake on pH. A coated-wall flow reactorcoupled to a chemical ionization mass spectrometer (CIMS)was used to study the adsorption of HONO on pH-adjustedagricultural soil and other soil components. Uptake of HONOonto the substrate-coated surface was initiated by retractingthe moveable injector to a position upstream of the coatedtube, thereby exposing the substrate layer to a well-mixed and

laminar flow of HONO and humidified air. The uptake ofHONO onto a surface is described by the uptake coefficient,γnet, a unitless measure of the fraction of collisions with thewall surface that results in the removal of the species from thegas phase (40–42). This ratio reflects the balance of HONOfluxes due to adsorption (Jads) and desorption (Jdes) divided bythe total flux of HONO molecules hitting the surface (Jcoll), asillustrated by Eq. 1.

γnet =Jads − Jdes

Jcoll[1]

When considering the pH dependence of γnet for HONO, weexpect a parabolic trend reflective of the pH-dependent specia-tion of HONO: Adsorption is most likely at high pH whereHONO can remain on the surface as nitrite (NO2ˉ), or at verylow pH where nitroacidium ion (H2ONO+) is the predominantform of N(III) (Fig. 1). At high pH, HONO is expected to reactwith basic M–Oˉ groups, leaving behind solvated nitrite and co-valently bonded nitrate, nitrito (–ONOˉ), and nitro (–NO2ˉ)groups, the existence of which is supported by infrared spectros-copy studies on related metal oxide systems (43–46). At theacidic and basic extremes, the γnet is high because adsorptionoutcompetes desorption. Considerations of bulk aqueous equi-libria for HONO (Fig. 1A) suggest that the lowest γnet valueswould be expected in the range of pH 2–3, because this is wheredesorption of volatile HONO becomes possible. Wetted-wallflow tube experiments show that values of γ for the uptake ofHONO into bulk water are indeed lowest at between pH 2 and 3(47). Experiments on soil surfaces, however, show the occur-rence of the minimum at pH 6.2 (Fig. 2A), nearly 3 pH unitsabove the pKa of HONO, which suggests that the interaction ofHONO with solid surfaces is different from its interaction withwater and that the surface charge of the amphoteric minerals insoil plays an important role.To further explore the interaction of HONO with mineral sur-

faces, we chose kaolinite as a model system because it is bothpresent in our soil sample (SI Appendix) and its surface propertieshave been well characterized in previous studies (48–52). Kaoliniteis a phyllosilicate mineral of alternating tetrahedral silicon andoctahedral aluminum sheets that has been shown to have a neutralsurface charge at pH ∼3 (48, 53); i.e., at a pH of 3, the surfacedensity of M–OH2

+ and M–Oˉ groups is equal. We also modifiedthe kaolinite surface by adding 2% wt/wt AlCl3, because this hasbeen shown to coat the kaolinite with amorphous Al3+-based hy-drolysis products, e.g., Al(OH)3. The coating shifts its neutralsurface charge to several pH units above that of pure kaolinitewhile maintaining the integrity of the clay (49). These two sub-strates represent surfaces where Al–OH2

+ groups are favored un-der acid (for kaolinite) or close to neutral pH values (for kaolinite +AlCl3), a difference that we would expect to dramatically impactthe pH of the minimum in HONO uptake. As shown in Fig. 2A, thetrends in HONO uptake on both surfaces show minima at pH 3.0and 6.9 for the kaolinite and the AlCl3 + kaolinite mixture, re-spectively, that track the surface charge of the substrate rather thanthe pKa of HONO in bulk water.

Nitrite Protonation as a Function of Soil pH.A series of experimentswas conducted to measure the amount of HONO evolved as, ineach case, 1 nmol of aqueous NO2ˉ was added onto pH-adjustedsubstrates to directly probe the influence of substrate pH on theJdes term in Eq. 1. Fig. 2B shows titration curves for NO2ˉadditions to soil, kaolinite, and kaolinite + AlCl3 mixtures. Mostnotably, the inflection points in the titration curves in Fig. 2Bcorrespond to the minima in HONO uptake for the respectivesubstrates shown in Fig. 2A. At pH values significantly above theHONO uptake minima, where basic M–O− species dominate thesurface of minerals, HONO was not produced due to the low

Fig. 1. An illustration of the speciation of HONO in natural systems. (A)Partitioning between bulk aqueous and gas phases is determined by Henry’slaw. Once in the aqueous phase, HONO dissociates into NO2ˉ or may beprotonated to form the nitroacidium cation (H2NO2

+), depending on the pH.(B) Speciation of HONO on soil surfaces depends on the surface acidity,which is controlled in part by amphoteric metal oxides present.

Donaldson et al. PNAS | December 30, 2014 | vol. 111 | no. 52 | 18473

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S

Dow

nloa

ded

by g

uest

on

May

20,

202

0

density of exchangeable surface protons. Because the surface isprotonated at lower pH, the proportion of M–OH2

+ groupsincreases and the surface is capable of protonating NO2ˉ to formHONO. Consistent with these results is the observation that thepresence of surface-adsorbed H2O is required to observe anydependency of γnet on pH (SI Appendix, Fig. S1). Water is clearlyrequired to hydrate the oxide surfaces and promote protontransfer from acidic M–OH2

+ groups to NO2ˉ (54). Fig. 2B showsthat soil and the 2% AlCl3 in kaolinite both demonstrated theability to protonate nitrite at pH values that are much greaterthan the pKa of HONO in the bulk aqueous phase. Once again,these experiments confirm the importance of the substrate sur-face chemistry—namely, the degree of protonation of the metaloxides, rather than typical bulk aqueous behavior.

Effect of Soil Surface Charge on HONO Uptake and Release. To cor-relate the pH dependence of γnet to the particle surface charge,potentiometric titration was used to determine the point of zerosalt effect (PZSE), the common intersection point where the H+

and OHˉ activities are unaffected by changes in electrolyte

concentrations, indicating a neutral surface charge (37, 55, 56).Potentiometric titration is a common method for finding thesuspension pH at which the net surface charge is zero, i.e., thepoint of zero charge (PZC) based on the amount of H+ adsorbedto exchangeable surface sites (37, 55, 57). At the PZSE of asubstrate, the surface is partially hydroxylated and the surfacedensity of M–OH2

+ and M–Oˉ groups is equal. As shown in Fig.2C, the PZSE values for soil, kaolinite, and the AlCl3 mixtureoccur at pH values of 6.1, 3.4, and 7.2, respectively. The PZSEvalues correspond to the minimum in HONO uptake efficiency(Fig. 2A) and the beginning of HONO desorption observed forthe nitrite titrations (Fig. 2B), because the surface is neithersufficiently basic to retain the HONO as nitrite nor acidicenough to retain the HONO as the nitroacidium cation. Thus,the Jdes term in Eq. 1 is greatest at the PZSE. The excellentagreement among the occurrence of the substrate PZSE values,the pH of minimum HONO uptake, and the onset of HONOdesorption strongly suggests that it is the presence of ampho-teric metal oxides that determines whether HONO is released orremains adsorbed to soil.

Fig. 2. A comparison of HONO speciation and surface acidity. This matrix shows the agreement between measurements of (A) HONO uptake; (B) HONO productionfrom nitrite; and (C) surface charge for the three substrates: agricultural soil, kaolinite, and 2% wt/wt AlCl3 in kaolinite. Lines on the HONO uptake and the HONOproduction plots are connections of the data to guide the eye, and the potentiometric titration has a polynomial fit used to derive the point of intersection, the PZSE.The error bars represent the 95% confidence interval of the mean of replicate measurements. The pH values highlighted by yellow bars represent the intersection ofminimal HONO uptake, the beginning of HONO production from soil nitrite, and the PZSE, all of which are indicative of neutral surface charge.

18474 | www.pnas.org/cgi/doi/10.1073/pnas.1418545112 Donaldson et al.

Dow

nloa

ded

by g

uest

on

May

20,

202

0

Soil Components Affecting HONO Adsorption and Desorption. Tofurther investigate the relative importance of the differentcomponents of the soil matrix on HONO uptake, we performeda series of experiments on treated soil. The substrates includedsoil oxidized with hydrogen peroxide and a reconstituted fractioncomposed of the oxidized soil and 3% wt/wt of an isolated soilhumic acid to examine the effect of soil organic matter on the pHdependence of HONO uptake to soil surfaces. Fig. 3 shows thatthe pH of minimum HONO uptake was lower by ∼2 pH unitswhen soil is oxidized, relative to unmodified soil. To further in-vestigate the role of organic matter, the oxidized soil wasreconstituted with 3% wt/wt Elliott soil humic acid to simulatethe organic matter content of the native soil. Elliott soil humicacid is isolated from soil collected in Joliet, IL, and is typical ofthe soil organic matter common to the US states of Indiana,Illinois, and Iowa (58). As shown in Fig. 3, the minimum HONOuptake at pH 3.9 for the reconstituted soil was similar to that ofthe oxidized soil, suggesting that soil organic matter has little tono effect on the pH dependence of HONO uptake onto soil.Although hydrogen peroxide removes soil organic matter, it

can also induce severe structural and chemical changes in the soilthat result in the mobilization of iron and aluminum ions frommineral surfaces and loss of structural OH groups (59–62). Theloss of iron and aluminum is confirmed by energy-dispersiveX-ray spectroscopy (EDX) measurements, which show that the ironand aluminum content in our soil decreases by 1–2% upon oxi-dation (Fig. 3). Hints into the nature of the iron- and aluminum-containing minerals affected by soil oxidation are obtainedby comparing EDX results with those obtained using X-raypowder diffraction (XRD; SI Appendix), which reveals that themain crystalline phases present in the soil are quartz, feldspar,

amphibole, mica, smectite, and kaolinite. None of these mineralsare expected to possess a PZC above 4, and the notable absenceof crystalline Fe3+ and Al3+ (hydr)oxides in the XRD pattern of theunmodified soil suggests that small amounts of amorphous ironoxides and aluminum hydroxides, not detectable by XRD, arelikely present (34). It is important to note that the determinedsoil PZSE represents an ensemble of many solid phases whoseindividual points of zero charge range from 1 to 8 (Fig. 4). Itwould appear that the PZSE of our soil sample is heavily influ-enced by amorphous Fe3+ and Al3+ (hydr)oxides present in thesoil, both of which have PZC values in the range of 7–9 (Fig. 4)and are known to cover mineral grains with a large surface areadespite their low relative abundance (53, 63). Similar to ourkaolinite + AlCl3 experiments, Sakurai et al. (50) demonstratedthat even small percentages of iron or aluminum hydroxides (2–10%) added to soils can significantly shift the PZC to highervalues, in some cases by as much as 4 pH units. Elliott andSparks (63) also demonstrated that when soil was treated toremove amorphous iron oxide coatings, the soils shifted fromhaving a PZC of ∼7.5 to behavior more like silica with a PZC of∼2. In conclusion, the data suggest that it is the loss of amor-phous iron and aluminum (hydr)oxide in the soil that bestexplains the significant decrease in the pH of minimum uptakefrom the 6.2 of native soil to the 4.0 observed for oxidized soil.

Environmental Implications. This study presents direct evidencethat reactive uptake and release of HONO on soil surfaces arepH dependent, consistent with the ability of amphoteric mineralsurfaces to influence the speciation of N(III) in the soil envi-ronment, independent of the acid–base equilibria expected inbulk water. The data support a mechanism whereby efficientuptake of gas-phase HONO at high soil pH is driven by the re-action of HONO with basic M–Oˉ groups to generate aqueousnitrite and covalently bonded nitrate, nitrito, and nitro groups,whereas at low pH values, higher uptake efficiency is driven bythe formation of surface-adsorbed H2ONO+ cations. At in-termediate pH values, any nitrite contributed by HONO to thesurface will be protonated by M–OH2

+ groups and a reversibleequilibrium will be established. These findings suggest that fu-ture parameterization of HONO uptake onto soil surfaces can-not only consider Henry’s law partitioning to and from anadsorbed or pore water phase as previously assumed (31) but

Fig. 3. The dependence of HONO uptake on mineral content and organicmatter. (A) The pH dependence of HONO uptake on treated soil samplesexposed to 10 ppb of HONO in air at 30% relative humidity: reconstitutedsoil, oxidized soil, and whole soil. The dashed lines are guides for the eye.The error bars represent the 95% confidence interval of the mean of repli-cate measurements. (B) Absolute change in wt% of elements in soil uponoxidation with H2O2, as determined by EDX.

Fig. 4. The range of measured PZC for soil, common minerals, and soilconstituents (53, 63, 67, 69). Note that bentonite consists largely of the claymineral smectite.

Donaldson et al. PNAS | December 30, 2014 | vol. 111 | no. 52 | 18475

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S

Dow

nloa

ded

by g

uest

on

May

20,

202

0

must also take into account the pH-dependent reactivity ofmineral surfaces that controls N(III) speciation in this system.This work has important implications for understanding how

soil pH determines whether soil NO2ˉ will be a net sink or sourceof HONO to the atmosphere and provides the theoretical frame-work to support recent reports of atmospherically significantHONO emissions from soils with pH values of up to 8.8, far abovethe pKa of HONO (26, 27). Whether soil nitrite is accumulated viaatmospheric deposition of HONO or by microbial nitrification, itcan act as a source of HONO to the atmosphere in the presenceof ubiquitous aluminum and iron (hydr)oxides, in addition tobeing influenced by deposition of photochemically derived acids assuggested by VandenBoer et al. (19). It is apparent that out-gassing of nocturnally accumulated soil nitrite over the courseof the next day could represent an important source of daytimeHONO and OH radical. Previously, it was shown that atmo-spheric–terrestrial exchange of HONO could be approximatedbased on the aqueous pKa of HONO, aqueous N(III) concen-trations, Henry’s law, and pore-water pH (27). However, our re-sults show that not accounting for the surface acidity of soilminerals may lead to an underestimation of the amount of HONOoutgassed from soil.The importance of soil as a widespread source of atmospheric

HONO is apparent from land-cover analyses that classify theEarth’s surface as 15% bare soil, with an additional 13% con-sidered cropland that often remains as bare soil surfaces throughfallow seasons (64). The knowledge that surface acidity is thedominant factor in nitrite protonation and release as HONO,rather than bulk pH, changes the proportion of soil types pos-sibly involved in HONO production from the small subset withpH <3 to the vast majority of acidic and nonacidic soils—astaggering 70% of global soils (65). As Fig. 4 demonstrates, soilsare extremely heterogeneous and their reactivity in terms ofbeing able to protonate soil nitrite will be heavily influenced bytheir constituents, which include both the mineral grains andcoatings of amorphous iron and aluminum (hydr)oxides (50).HONO release and uptake are clearly coupled to soil mineral-ogy, and an understanding of the soil characteristics is needed tofully understand the atmospheric–terrestrial exchange of N(III)in the global nitrogen cycle. This additional source of HONOacts as a precursor to OH and thus plays a role in determiningozone and aerosol formation and in limiting the lifetime ofgreenhouse gases in the atmosphere.

MethodsSoil Characterization. Soil samples were collected from an agricultural fieldlocated in Bartholomew County, IN (39.17, −85.89) at a depth of 0–5 cm in May2013. The soil was autoclaved for 1 h at 394 K for three successive days andsieved with a 120-mesh sieve. The soil is classified as Genesee, a fine loamyalluvium comprised of quartz, aluminosilicates, and 1% organic matter (66).The unadjusted soil pH was measured as recommended by Hendershot et al.(67) [1:2 soil/water (vol/vol); Thermo Scientific, OrionStar A211 pH meter] andwas determined to be 6.48. SEM (FEI Quanta 600 FEG instrument) was used tocharacterize the soil–glass interface and soil surface; the coating covered theentire inner surface of the tube and was uniform to the eye.

Coated-Wall Flow Tube Studies. Five substrates were used in this study:Genesee soil, kaolinite (Fluka Analytical), a 2% wt/wt AlCl3 (Sigma Aldrich) inkaolinite mixture, oxidized soil, and reconstituted soil. The substrate slurrypH was adjusted using H2SO4 (Fluka Analytical) and NaOH (EMD Chemicals)solutions, and was measured using the Thermo Scientific, OrionStar A211 pHmeter; due to the heterogeneity of the substrates and the variations in slurrycomposition, an error of ±0.2 is assumed. The slurry composition (2:1 soil/

water and 1:1 kaolinite/water, wt/wt) was chosen based on the suitabilityfor coating. The slurry of substrate and deionized water (18.2 MΩ·cm; Milli-QIntegral) was dripped onto the inner walls of the cylindrical glass tubeinserts and dried overnight in an oven at 393 K, yielding dry weights thatwere determined to be well within the mass independent region to bestrepresent an infinitely thick layer of substrate (soil: 1.5 ± 0.2 g; kaolinite:0.7 ± 0.1 g) (31).

Uptake of HONO onto the substrate-coated surface was initiated byretracting the moveable injector to a position upstream of the coatedsection, exposing the substrate layer to a well-mixed and laminar flow ofHONO and humidified air (SI Appendix). Uptake coefficients reported hereare derived from the steady-state portion of the signal vs. time and arebased on the geometric surface area; this best represents real boundarylayer soil surfaces and provide results that are directly comparable to fieldmeasurements and can be implemented into models.

Soil Oxidation and Reconstitution. Based on a modification of the methodof Favilli et al. (68), a slurry of soil (120 g), water (150 mL), and hydrogenperoxide (Macron Fine Chemicals; 30% wt/wt, 500 mL) was stirred in avented flask for 30 d. The soil was filtered and rinsed with water (1 L).More H2O2 (400 mL) was added portion-wise over 48 h to ensure completereaction (i.e., to the point where no more gas was evolved), and the soilwas filtered and rinsed again thoroughly. The reconstituted soil wascomprised of a mixture of the oxidized soil and 3% wt/wt of Elliott soilhumic acid (International Humic Substance Society).

Nitrite Titration Studies. Three substrates (soil, kaolinite, and a 2%wt/wt AlCl3in kaolinite mixture) were used in this study, and the pH-adjusted slurrieswere prepared and adjusted using the method described above. However, inthis case the substrate was then dried and the powder (0.1 ± 0.02 g) waspacked into a small Teflon trough (13 mm inner diameter, 2 mm depth). Thistrough was placed beneath a port inside the flow tube (flow: 2,180 cm3·min–1)where a rubber stopper supported a 10-μL Hamilton syringe. After the troughwas inserted and had been allowed to equilibrate with air at 30% RH, a 5-μLdrop of 200 μM NaNO2 was added and the resulting outgassed HONO wasmonitored and averaged over a 10-min interval. This procedure was carriedout with two fresh surfaces for each substrate and pH.

Potentiometric Titrations. Suspensions of soil (100 g·L−1) in 0.01 and 0.1 MNaCl were adjusted to range from pH 2.3 to 10.2 using HCl and NaOH. Thesuspensions were kept at room temperature in sealed 20-mL vials andshaken twice a day for 3 d. The pH of each was then measured again on thethird day. The initial amount of acid or base minus the amount necessaryto adjust a blank NaCl solution to the final equilibrium pH was used to es-timate the amount of H+ and OHˉ adsorbed by the soil or clay samples (56).Aqueous substrate dispersions were titrated at two different ionic strengthsto allow for identification of the inflection point of the titration curve (thecommon intersection point that occurs at the PZSE) that is not sensitive tochanges in ionic strength. The amount of H+ adsorbed and released dependson ionic strength above and below this point, which affects the thickness ofthe diffuse double layer and hence the degree to which the surface charge isshielded (37).

Energy-Dispersive X-ray Spectroscopy. EDX with an Oxford Inca detector wasused to determine the elemental composition of the soil. The results for thenative soil (in weight %) are as follows: O, 63; Si, 19; C, 10; Al, 5; Fe, 2; Ca, 0.4;and K, 1. The results for the oxidized soil (in weight%) are as follows: O, 56; Si,33; C, 3; Al, 4; Fe, 1; Ca, 0.7; and K, 2 (SI Appendix).

ACKNOWLEDGMENTS. We thank Mulu Kebede for assistance with SEM-EDXmeasurements; Patrick Cavanaugh for assistance with XRD analysis; theIndiana University Nanoscale Characterization Facility for access to thescanning electron microscope; Jeffrey White and Allen Siedle for helpfuldiscussions; and Jörg Meyer and Donald Garvin for glassblowing. This workwas funded by National Science Foundation CAREER Award AGS-1352375 (toJ.D.R.), Indiana University, and National Science Foundation Graduate Re-search Fellowship Program DGE-1342962 (to M.A.D.).

1. Finlayson-Pitts BJ, Pitts JN, Jr (2000) Chemistry of the Upper and Lower Atmosphere:

Theory, Experiments, and Applications (Academic, San Diego).2. Harrison RM, Peak JD, Collins GM (1996) Tropospheric cycle of nitrous acid. J Geophys

Res Atmos 101(D9):14429–14439.3. Lammel G, Cape JN (1996) Nitrous acid and nitrite in the atmosphere. Chem Soc Rev

25(5):361–370.

4. Perner D, Platt U (1979) Detection of nitrous acid in the atmosphere by differential

optical absorption. Geophys Res Lett 6(12):917–920.5. Platt U, Perner D, Harris GW, Winer AM, Pitts JN, Jr (1980) Observations of nitrous acid

in an urban atmosphere by differential optical absorption. Nature 285(5763):312–314.6. Harris GW, et al. (1982) Observations of nitrous acid in the Los Angeles atmosphere and im-

plications for predictions of ozone-precursor relationships. Environ Sci Technol 16(7):414–419.

18476 | www.pnas.org/cgi/doi/10.1073/pnas.1418545112 Donaldson et al.

Dow

nloa

ded

by g

uest

on

May

20,

202

0

7. Alicke B, Platt U, Stutz J (2002) Impact of nitrous acid photolysis on the total hydroxylradical budget during the LOOP/PIPAO study in Milan. J Geophys Res Atmos 107(D22):LOP 9-1–LOP 9-17.

8. Alicke B, et al. (2003) OH formation by HONO photolysis during the BERLIOZ exper-iment. J Geophys Res Atmos 108(D4):PHO 3-1–PHO 3-17.

9. Winer AM, Biermann HW (1994) Long pathlength differential optical absorptionspectroscopy (DOAS) measurements of gaseous HONO, NO2 and HCHO in the Cal-ifornia south coast air basin. Res Chem Intermed 20(3):423–445.

10. Kleffmann J, Wiesen P (2005) Heterogeneous conversion of NO2 and NO on HNO3

treated soot surfaces: Atmospheric implications. Atmos Chem Phys 5(1):77–83.11. Monge ME, et al. (2010) Light changes the atmospheric reactivity of soot. Proc Natl

Acad Sci USA 107(15):6605–6609.12. Kleffmann J (2007) Daytime sources of nitrous acid (HONO) in the atmospheric

boundary layer. ChemPhysChem 8(8):1137–1144.13. Stutz J, Alicke B, Neftel A (2002) Nitrous acid formation in the urban atmosphere:

Gradient measurements of NO2 and HONO over grass in Milan, Italy. J Geophys ResAtmos 107(D22):LOP 5-1–LOP 5-15.

14. Elshorbany YF, Steil B, Brühl C, Lelieveld J (2012) Impact of HONO on global atmo-spheric chemistry calculated with an empirical parameterization in the EMAC model.Atmos Chem Phys 12(20):9977–10000.

15. Wong KW, Tsai C, Lefer B, Grossberg N, Stutz J (2013) Modeling of daytime HONOvertical gradients during SHARP 2009. Atmos Chem Phys 13(7):3587–3601.

16. Vogel B, Vogel H, Kleffmann J, Kurtenbach R (2003) Measured and simulated verticalprofiles of nitrous acid-part II. Model simulations and indications for a photolyticsource. Atmos Environ 37(21):2957–2966.

17. Sarwar G, et al. (2008) A comparison of CMAQ HONO predictions with observationsfrom the Northeast Oxidant and Particle Study. Atmos Environ 42(23):5760–5770.

18. Gonçalves M, Dabdub D, Chang WL, Jorba O, Baldasano JM (2012) Impact of HONOsources on the performance of mesoscale air quality models. Atmos Environ54:168–176.

19. VandenBoer TC, et al. (2013) Understanding the role of the ground surface in HONOvertical structure: High resolution vertical profiles during NACHTT-11. J. Geophys ResAtmos 118(17):10155–10171.

20. Wong KW, et al. (2012) Daytime HONO vertical gradients during SHARP 2009 inHouston, TX. Atmos Chem Phys 12:635–652.

21. Wong KW, Oh H-J, Lefer BL, Rappenglück B, Stutz J (2011) Vertical profiles of nitrousacid formation in the nocturnal urban atmosphere of Houston, TX. Atmos Chem Phys11(8):3595–3609.

22. Underwood GM, Song CH, Phadnis M, Carmichael GR, Grassian VH (2001) Hetero-geneous reactions of NO2 and HNO3 on oxides and mineral dust: A combined labo-ratory and modeling study. J Geophys Res Atmos 106(D16):18055–18066.

23. Finlayson-Pitts BJ, Wingen LM, Sumner AL, Syomin D, Ramazan KA (2003) The het-erogeneous hydrolysis of NO2 in laboratory systems and in outdoor and indoor at-mospheres: An integrated mechanism. Phys Chem Chem Phys 5(2):223–242.

24. Ramazan KA, Syomin D, Finlayson-Pitts BJ (2004) The photochemical production ofHONO during the heterogeneous hydrolysis of NO2. Phys Chem Chem Phys6:3836–3843.

25. Wang S, et al. (2003) Atmospheric observations of enhanced NO2-HONO conversionon mineral dust particles. Geophys Res Lett 30(11):1595.

26. Oswald R, et al. (2013) HONO emissions from soil bacteria as a major source of at-mospheric reactive nitrogen. Science 341(6151):1233–1235.

27. Su H, et al. (2011) Soil nitrite as a source of atmospheric HONO and OH radicals.Science 333(6049):1616–1618.

28. Hellebust S, Roddis T, Sodeau JR (2007) Potential role of the nitroacidium ion onHONO emissions from the snowpack. J Phys Chem A 111(7):1167–1171.

29. Riordan E, Minogue N, Healy D, O’Driscoll P, Sodeau JR (2005) Spectroscopic andoptimization modeling study of nitrous acid in aqueous solution. J Phys Chem A109(5):779–786.

30. Anastasio C, Chu L (2009) Photochemistry of nitrous acid (HONO) and nitrous acidiumion (H2ONO) in aqueous solution and ice. Environ Sci Technol 43(4):1108–1114.

31. Donaldson MA, Berke AE, Raff JD (2014) Uptake of gas phase nitrous acid ontoboundary layer soil surfaces. Environ Sci Technol 48(1):375–383.

32. Ewing GE (2004) Thin film water. J Phys Chem B 108(41):15953–15961.33. Ewing GE (2006) Ambient thin film water on insulator surfaces. Chem Rev 106(4):

1511–1526.34. Doyle CS, et al. (1999) XANES studies at the Al K-edge of aluminium-rich surface

phases in the soil environment. J Synchrotron Radiat 6(Pt 3):621–623.35. Parks GA, deBruyn PL (1962) The zero point of charge of oxides. J Phys Chem 66(6):

967–973.36. Stumm W, Morgan JJ (1996) Aquatic Chemistry (Wiley, New York), 3rd Ed.37. Parker JC, Zelazny LW, Sampath S, Harris WG (1979) A critical evaluation of the ex-

tension of zero point of charge (zpc) theory to soil systems. Soil Sci Soc Am J 43(4):668–674.

38. Duffy JA, Ingram MD (1978) Acidic nature of metal aquo complexes: Proton-transferequilibriums in concentrated aqueous media. Inorg Chem 17(10):2798–2802.

39. Shimizu K, Shchukarev A, Boily J-F (2011) X-ray photoelectron spectroscopy of fast-frozen hematite colloids in aqueous solutions. 3. Stabilization of ammonium speciesby surface (hydr)oxo groups. J Phys Chem C 115(14):6796–6801.

40. Hanson DR (1997) Surface-specific reactions on liquids. J Phys Chem B 101(25):4998–5001.

41. Ammann M, Pöschl U (2007) Kinetic model framework for aerosol and cloud surfacechemistry and gas-particle interactions; Part 2: Exemplary practical applications andnumerical simulations. Atmos Chem Phys 7(23):6025–6045.

42. Pöschl U, Rudich Y, AmmannM (2007) Kinetic model framework for aerosol and cloudsurface chemistry and gas-particle interactions; Part 1: General equations, parame-ters, and terminology. Atmos Chem Phys 7(23):5989–6023.

43. Hadjiivanov KI (2000) Identification of neutral and charged NxOy surface species by IRspectroscopy. Catal Rev Sci Eng 42(1-2):71–144.

44. Underwood GM, Miller TM, Grassian VH (1999) Transmission FT-IR and Knudsen cellstudy of the heterogeneous reactivity of gaseous nitrogen dioxide on mineral oxideparticles. J Phys Chem A 103(31):6184–6190.

45. Ozensoy E, Peden CHF, Szanyi J (2005) NO2 adsorption on ultrathin theta-Al2O3 films:Formation of nitrite and nitrate species. J Phys Chem B 109(33):15977–15984.

46. Szanyi J, Kwak JH, Chimentao RJ, Peden CHF (2007) Effect of H2O on the adsorptionof NO2 on γ-Al2O3: an in situ FTIR/MS study. J Phys Chem C 111(6):2661–2669.

47. Hirokawa J, Kato T, Mafuné F (2008) Uptake of gas-phase nitrous acid by pH-con-trolled aqueous solution studied by a wetted wall flow tube. J Phys Chem A 112(47):12143–12150.

48. Sposito G (1984) The Surface Chemistry of Soils (Oxford University Press, New York), p 234.49. Johnson SB, Dixon DR, Scales PJ (1999) The electrokinetic and shear yield stress

properties of kaolinite in the presence of aluminium ions. Colloids Surf A PhysicochemEng Asp 146(1–3):281–291.

50. Sakurai K, Teshima A, Kyuma K (1990) Changes in zero point of charge (ZPC), specificsurface area (SSA), and cation exchange capacity (CEC) of kaolinite and montmoril-lonite, and strongly weathered soils caused by Fe and Al coatings. Soil Sci Plant Nutr36(1):73–81.

51. Wieland E, Stumm W (1992) Dissolution kinetics of kaolinite in acidic aqueous sol-utions at 25 C. Geochim Cosmochim Acta 56(9):3339–3355.

52. Schroth BK, Sposito G (1997) Surface charge properties of kaolinite. Clays Clay Miner45(1):85–91.

53. Kosmulski M (2009) pH-dependent surface charging and points of zero charge. IV.Update and new approach. J Colloid Interface Sci 337(2):439–448.

54. Rubasinghege G, Grassian VH (2013) Role(s) of adsorbed water in the surface chem-istry of environmental interfaces. Chem Commun 49(30):3071–3094.

55. Appel C, Ma LQ, Dean Rhue R, Kennelley E (2003) Point of zero charge determinationin soils and minerals via traditional methods and detection of electroacoustic mo-bility. Geoderma 113(1–2):77–93.

56. Marcano-Martinez E, McBride MB (1989) Comparison of the titration and ion ad-sorption methods for surface charge measurement in oxisols. Soil Sci Soc Am J 53(4):1040–1045.

57. Sparks D (2003) Environmental Soil Chemistry (Academic, San Diego), 2nd Ed, pp 352.58. Swift RS (1996) Organic matter characterization. Methods of Soil Analysis. Part 3.

Chemical Methods, Soil Sci Soc Am Book Series: 5, ed Sparks DL et al. (Soil Sci Soc Am,Madison, WI), pp 1018–1020.

59. Mikutta R, Kleber M, Kaiser K, Jahn R (2005) Review: Organic matter removal fromsoils using hydrogen peroxide, sodium hypochlorite, and disodium peroxodisulfate.Soil Sci Soc Am J 69(1):120–135.

60. Fialips C-I, Huo D, Yan L, Wu J, Stucki JW (2002) Infrared study of reduced andreduced-reoxidized ferruginous smectite. Clays Clay Miner 50(4):455–469.

61. Komadel P, Madejova J, Stucki JW (1995) Reduction and reoxidation of nontronite:Questions of reversibility. Clays Clay Miner 43(1):105–110.

62. Manceau A, et al. (2000) Oxidation-reduction mechanism of iron in dioctahedralsmectites: II. Crystal chemistry of reduced Garfield nontronite. Am Mineral85:153–172.

63. Elliott HA, Sparks DL (1981) Electrokinetic behavior of a paleudult profile in relationto mineralogical composition. Soil Sci 132(6):402–409.

64. Latham J, Cumani R, Rosati I, Bloise M (2014) FAO Global Land Cover (GLC-SHARE)Beta-Release 1.0 Database (Food and Agriculture Organization of the United Nations,Rome).

65. Merry RH (2009) Acidity and alkalinity of soils. Environmental and Ecological Chem-istry, ed Sabljic A (Encyclopedia of Life Support Systems, Paris), Vol 2.

66. US Department of Agriculture (USDA) Staff (2013)Web Soil Survey. Natural ResourcesConservation Service, USDA. Available at websoilsurvey.nrcs.usda.gov/. AccessedNovember 7, 2014.

67. Hendershot WH (1978) Measurement technique effects of the value of zero point ofcharge and its displacement from zero point of titration. Can J Soil Sci 58(3):439–442.

68. Favilli F, et al. (2008) Comparison of different methods of obtaining a resilient organicmatter fraction in Alpine soils. Geoderma 145(3–4):355–369.

69. Sakurai K, Teshima A, Kyuma K (1989) Factors affecting zero points of charge (ZPC) ofvariable charge soils. Soil Sci Plant Nutr 35:21–31.

Donaldson et al. PNAS | December 30, 2014 | vol. 111 | no. 52 | 18477

EART

H,A

TMOSP

HER

IC,

ANDPL

ANET

ARY

SCIENCE

S

Dow

nloa

ded

by g

uest

on

May

20,

202

0