The Atmospheric Chemistry of Hydrogen Cyanide (HCN)

UC Irvine UC Irvine Previously Published Works

Title The atmospheric chemistry of hydrogen cyanide (HCN)

Permalink https://escholarship.org/uc/item/98d9x8wp

Journal Journal of Geophysical Research, 88(C15)

ISSN 0148-0227

Authors Cicerone, RJ Zellner, R

Publication Date 1983

DOI 10.1029/JC088iC15p10689

License https://creativecommons.org/licenses/by/4.0/ 4.0

Peer reviewed

eScholarship.org Powered by the California Digital Library University of California JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 88, NO. C15, PAGES 10,689-10,696, DECEMBER 20, 1983

The AtmosphericChemistry of Hydrogen Cyanide (HCN)

R. J. CICERONE

National Centerfor AtmosphericResearch

R. ZELLNER

UniversitdtGSttin•Ten, lnstitut fur PhysikalischeChemie

Since 1981,three groupshave reportedspectroscopic detections and measurementsof hydrogency- anidein the atmosphere.HCN concentrations(volume mixing ratios) of (1.5-1.7)x 10-•ø appearto characterizethe stratosphereand the northernhemisphere's nonurban troposphere. In this paper we explorethe atmosphericbehavior of HCN by examiningits chemicaland photochemicalproperties. Its principalsinks are reactionswith atmosphericOH and O(•D); precipitationappears to be a negligible sink. In the stratosphere,vacuum UV photonsalso attack HCN. Atmosphericmodel calculationsshow that HCN shouldbe relativelywell mixedin the troposphereand that its concentrationdecreases slowly with altitudein the stratosphere.Its atmosphericresidence time appearsto be about 2.5 years,although 1-5 yearsis a possiblerange. To maintain the observedatmospheric burden of HCN, an annual source of about 2 x 10TM g nitrogenas HCN is required;we speculateas to the identityof thesesources. Oxidation of HCN by OH, while the major sink for atmosphericHCN, is not simpleor direct.Instead, oxidation proceedsfrom the HCN-OH adduct formed in HCN + OH reactions.These pathways and their uncertainties are outlined here.

1. INTRODUCTION be explored.It is alsoof interestto knowwhether the avail- HCN was first seenin the earth's atmosphereby Coffey et able HCN measurementsshed any light on the origin and al. [1981]. They employed infrared absorption instrumenta- behaviorof stratosphericCH3CN detectedby Arnoldet al. tion aboard an aircraft that flew between 5øN and 50øN lati- [1978],Arijs et al. [1980],and B6hrinqer and Arnold [1981]. tude. Coffey et al. measuredthe total amount of HCN above In thispaper we explore the atmospheric chemistry of HCN 12 km altitude; their vertical column was 6.5 x 10•½ cm-2. by evaluatingits propertiesand availablekinetic and thermo- Assuming that HCN is distributed uniformly with altitude dynamicdata; this discussion comprises section 2. Section3 above 12 km, they deduced a stratosphericmixing ratio of presentsour atmosphericmodel calculations and estimates of 1.55x 10-•ø by volume.These corrected HCN amountsare HCN verticalprofiles, source strengths, and atmosphericresi- slightly lower (M. T. Coffey, private communication, 1982) dence times. than those reported by Coffey et al. [1981]. Similarly, Carli et 2. ATMOSPHERICCHEMISTRY OF HCN al. [1982] discoveredHCN by examiningstratospheric micro- wave emissionspectra. They obtained data between20 and 40 HCN is a stronglybound molecule. Its bond dissociation km altitude. They state that the HCN mixing ratio was (1.3- energyD hasbeen determined by Davisand Okabe[1968]; 2.6) x 10-•ø. More recently,Rinsland et al. [1982] detected D(H-CN) = 5.20 eV or 120 kcal/mol.This D value corre- HCN in the troposphere. With IR absorption measurements spondsto a long-wavelengthcutoff of 238.4nm for photo- from Kitt Peak (altitude = 2.095 kin), they found vertical dissociation.In reality, the dissociationof HCN at wave- columnamounts of 2.73x 10•5 cm-2, yieldingan average lengthsnear 238 nm has not been observed.Instead, Hilgen- troposphericmixing ratio of 1.66x 10-•0 between2.1 and 12 dorff [1935] observeda regular band seriesabsorption be- km. They were also able to estimatestratospheric H CN con- tween 178 and 200 nm, with an onset to continuous absorp- centrations. tion below 178 nm. Herzberg and Innes [1957] observedpre- Becausethe earth's atmosphereis oxidizing,reasons for the dissociation for /l < 178.5 nm (6.93 eV), although they presenceof HCN are not entirely obvious.Questions as to the admitted that weak predissociationabsorption might have sourcesof this compound and the rate of oxidation arise im- escapedtheir detection.In their research,Herzberg and Innes mediately. Further, it is important to know if the oxidation of found that HCN representsa significant in situ source of gaseousni- HCN + hv• H(2S)+ CN(2•) trogen oxides or of NO3- and CN- in precipitation and whether atmospheric HCN can interfere with attempts to the 2n stateof CN is 1.15eV abovethe X2E groundstate. In measureNOz or peroxyacetylnitrate, for example,by conver- Davis and Okabe's [1968] work, they actually excited HCN sion to NO followed by chemiluminescentdetection of NO. At with much shorter wavelengthsand observedfluorescence of conc•ntrauons Ol I tu ppt, za•'• concentrationscan exceed CN(B2y•),an excitedstate ot c;• mat is .•.zue V aboveground. those of NO and NO2 in the background troposphere[see There are no available data on photodissociationof HCN for Kley et al., 1981], so the possibilitiesmentioned above should 175 nm

10,689 10,690 CICERONEAND ZELLNER' ATMOSPHERICCHEMISTRY OF HYDROGENCYANIDE

TABLE 1. Enthalpyof ReactionAH• for PossibleReactions of electronshell reactants exhibit relativelylarge activationener- HCN With 14 AtmosphericGases gies[see, e.g., Kneba and 'Wolfrum, 1980]. Reactants • Products AHs, kcal/mol In other casesfrom Table 2, for example,HCN + O3-• HOCN + 02, while one can imagine reasonablepathways H CN + O2--• H OCN + O +14 like • HNCO + O +2O HCN + OH--• HOCN + H >+2 HCN + 03--} HOCN + 02 HCN + O3--• OH + CN + 02 +43 HCN + HO: • H 2 + 02 + CN +66 -• H202 + CN +35 primaryattacks on a closedelectron shell molecule (HCN) are HCN + H:O: • H 2 + HO 2 + CN + 102 required.For example,the reactionof CO (isoelectronicwith HCN + NO 2--• HNO 2 + CN +43 HCN) and 03 is unmeasureablyslow (k < 10-23 cm3/ +94 HCN + HNO: • H: + NO 2 + CN molecules). In severalof the possiblereactions listed in Table HCN + CH:O--, H: + CN + CHO + 104 HCN + SO 2• NCO + HSO +101 2, extensiverearrangements are alsorequired. Finally, thereis HCN + O• OH + CN +19 no available evidencethat any of the possiblereactions in H CN + C1--• H C1 + CN +18 Table 2 actually occur exceptfor one of the pathwayswith HCN + H • H 2 + CN +17 OH. HCN + HC1---•H 2 + CICN +23 HCN + CO• CHO + CN + 105 There are two likely atmosphericreactants of significancein the atmosphericchemistry of HCN: O(•D) and OH. Indeed, Heats of formation for these calculations are taken from Okabe the reaction of OH and HCN has been studied in the labora- [1978] and apply for 0øK except for HO 2 for which we took tory [Phillips, 1978, 1979; Fritz et al., 1983; R. Zellner and B. Howard's [1980] value of 2.5 _ 1 kcal/mol and for NCO we adopted 48 kcal/mol from Sullivan et al. [1983]. Note that all these potential Fritz, unpublishedmanuscript, 1983]. From theselaboratory reactionsare endothermic(AHs > 0); rates of these reactionsin the experiments,it appearsthat an adduct is formed; the thermo- atmospherewould be negligiblyslow. chemistryof the possiblereaction pathways is shownin Table 2. The central valuesof AH298ø for HCN, OH, CN, and H20 lead to the conclusion that HCN + OH--}CN + H20 is Contrary to extrapolations from information in several slightly endothermic (A/-/a = + 1.5 kcal/mol). Further, for a popular handbooks,HCN is not very solublein water when hydrogen abstraction reaction like this, BEBO calculations low partial pressures(for example, < 1 torr) of HCN are in indicate that there should be an activationenergy of at least 9 question.Henry's constantKn = P/X, where P is the partial kcal/mol (R. Zellner and B. Fritz, unpublishedmanuscript, pressure(torr) of HCN and X is the mole fraction of HCN in 1983). The experimentsof Fritz et al. [1983] imply that this the liquid, is Kn = 4000 at 18øCand 0.01 atm HCN [Interna- reaction path is not followed at atmospheric temperatures. tional Critical Tables of Numerical Data, 1928]. At the same Further, the exchange reaction path (OH + HCN temperaturethe Ostwald solubility coefficient(see, e.g., Wil- HOCN + H) which has been suggestedto be one of the domi- helm et al., 1977] is about 252 [Linke, 1958]. Further confirm- nant HCN oxidation routes in flames [Haynes, 1977] is en- ing data are found in the works by Lewisand Randall [1923] dothermicby 2.4 kcal/mol, and it is probably preventedby an and Landolt-B6rnstein[1962]. On the basis of this solubility additional dissociation barrier between HCN.OH and and the fact that HCN is a very weak acid HOCN + H. Formation of HNCO + H is less likely than HCNaq•- H + + CN- (K = 1.3x 10-9) HOCN due to an additional isomerizationof the heavy atom moiety. one deducesthat HCN has a very long atmosphericresidence time against rainout. If 1 m of rain falls per unit area each year and it were saturatedwith HCN (with respectto, say,200 TABLE 2. Possible Exothermic Reactions of HCN With ppt of gaseousHCN) it would take 34 yearsof averagerainfall AtmosphericGases to remove the 200 ppt of HCN. This calculation is not very Reactants--} Products AH•, kcal/mol sensitiveto rainfall pH; we took pH = 4 here. Let us proceednow to examinepossible gas phase reactions HCN + 0 3--} HO 2 + NCO --16 that can affect the atmosphericbehavior of HCN. We will --} HNCO + 0 2 -81 concludethat the principal sinks for HCN are reactionswith --} HOCN + 0 2 -75 HCN + NO--} N 2 + CHO -45 OH and O(•D) and in the upperstratosphere, photolysis. First HCN + NO 2 --} N20 + CHO -11 we will examineand dismissother potential reactions.Table 1 --} HNCO + NO -33 lists enthalpiesof reaction, AHR, for potential reactions be- --} HOCN + NO -27 tween HCN and 14 atmosphericgases. The significantlyposi- HCN + O2(tA)--} OH + NCO -14 HCN + C10--} HC1 + NCO -30 tive values of AHR shown in Table 1 indicate that each of HCN + H202--• HNCO + H20 -71 these reactionswould proceed very slowly if at all at atmo- --} HOCN + H20 -66 spherictemperatures. Once again, the impressionarises that HCN + OH--} H20 + CN -1.5 < AH• < +4.5 HCN can be a relatively stable gas in the atmosphere,as was -• HNCO + H -4 suggestedby Crutzenet al. ['1979]. --} HCNOH - 35(+ 4) It is possibleto proposeexothermic gas phaseatmospheric See Table 1 for definitions and sourcesof thermodynamicdata; a reactions of HCN. Table 2 lists seven such reactants that heat of formation,AH s, of 9 kcal/molhas beentaken for CHO branch into 14 setsof reaction products.Close inspectionof [Benson,1976]. We havetaken AHs > -14 and > -8 kcal/molfor molecular structures leads one to the conclusion that most of HNCO and HOCN, respectively[Sullivan et al., 1983; Benson,1978]. Although exothermic, most of these reactions are concluded to be thesereactions are unlikely to proceedin the atmosphere.For highly unlikely as discussedin the text. The only likely reactions example, HCN +C10-•NCO+ HC1 would be a four- affectingatmospheric HCN is with OH in which an adduct is formed centered reaction; four-centered reactions involving closed- andreaction with O(•D). CICERONEAND ZELLNER'ATMOSPHERIC CHEMISTRY OF HYDROGENCYANIDE 10,691

HOCN + HOa We cannot dismissthis possibility,but we considerit unlikely that this bimolecularabstraction reaction can competewith termolecularO2 addition, especiallyin the troposphere.To H• 02,M H competesuccessfully, ka must exceedthe effectivesecond-order rate constant, k., for the O2 addition shown at point A in HO• C • N• O0 Figure i. At ambient pressuresk. should be near i0 -"2 (D/.o cm3/molecules. Severallines of reasoningindicate that ka cannot be this fast. First, new experimentsin G6ttingen have testedthe influenceof 02 by adding 02 and NO to the reac- H• tion mixture. The apparent rate constant for OH + HCN C:N•) should decreaseif HOCN and HO 2 are produced(due to the HO ' fast regenerationof OH through HO 2 + NO--} OH + NO2). This effectwas not observed;the experimentsplaced an upper limit of 5 x 10-23 cm3/molecules on ka.If the reactiondoes H•. proceed at this upper limit rate, it could representan impor- C --N =O tant pathway in the oxidation of the HCNOH adduct, if the HO/ adduct's structure is as shown in Figure 1. A few other reactions in which 02 abstractsan H atom from a free radical have been studied. Table 3 lists six such reactions and their thermochemistry.Unfortunately, reaction rates are not avail- C •N:O able for all these reactionsespecially at atmospherictemper- o • atures. The reactions of Table 3 are not fast; some are very slow. The reaction of C2H4OH most closely resemblesour candidate reaction both in complexity and exothermicity (for HCNOH + O2-} HOCN + HO2, AHR = - 12.5 kcal/mol). NO + CHO The reactionrate listedin Table 3, 1.3 x 10-23 cm3/molecule Fig. 1. Atmosphericoxidation pathway of the HCN. OH adduct s [Peeters et al., 1982] applies for vibrationally excited assumingthat the OH is associatedwith the carbon atom. Processes C2H4OH. Peeters et al. also found that deactivation of labeled A-G are discussedin the text, for example, F is the decompo- C2H4OH* by 0 2 was faster than reactionswith 02. Hence in sition of nitrosoformaldehyde.This overall oxidation pathway is by analogy with that of hydrocarbonsand is essentiallythat of Fritz et air the dominant pathway would be deactivation by N 2 and al. [1983]. 02, and the vibrationally excitedstates would be unimportant.

The HCNOH adduct could appear either as shown at the /OH top of Figure 1 or at the top of Figure 2. Experimentally,Fritz H--(• :N et al. [1983] and R. Zellner and B. Fritz (unpublishedmanuscript, 1983) saw no dissociationof the HCNOH adduct on a few millisecond time scale, leading them to deduce that the adduct is at least -15 kcal/mol stable with respect to the Q•02'M /OH reactants that form it. The thermochemistryof addition reac- H--C--N' tions can be used to estimate the heat of formation of the adduct. For example,for HO• OH /c = N H•C:N H we estimateAHy to be 7 + 4 kcal/mol.With thisvalue of AHy the adduct would be 35 (+4) kcal/mol stable with respectto the reactants.This structureof the adduct might be preferred OH becauseC-O bonds generally tend to be more stable than H• N-O bonds. Nonetheless,following Fritz et al., we consider o•C--•/ the possibleoxidation chains of both isomersof HCNOH; these are sketchedin Figures 1 and 2. These oxidation pathways, beginningwith attack either by 02 or 03 in Figure 1 ©J,o H and by 02 in Figure 2, are suggestedby analogywith hydro- •C• N---O carbon oxidation pathways.At point G in Figure 1 we indi- o • cate the possibilitythat a bimolecularreaction with 0 2 could abstract an H atom from the HCNOH adduct if the adduct is of the structureshown in Figure 1, that is,

NO + ½HO HO• Fig. 2. Atmosphericoxidation pathway of the HCN.OH adduct assumingthat theOH is associatedwith the nitrogen atom. Processes H/C=Iq+02 • HOCN+HO 2 labeled0•,/•, ..., • arediscussed in thetext. See also Figure 1. 10,692 CICERONEAND ZELLNER.'ATMOSPHERIC CHEMISTRY OF HYDROGENCYANIDE

TABLE 3. Thermochemistryand Rate Coefficientsfor SeveralRadical ReactionsWith O 2 That Form HO,•

Reaction Rate,cm3/molecule s AHR, kcal/mol

C2H 5 + 0 2--} C2H,, + HO 2 ki= 2.1 x 10-•3 -11.5 C2H 3 +O 2-•C2H 2 + HO 2 kii(1000K)= 1.5 x 10-•2 -12.3 CH30 + 0 2-• CH20 + HO 2 kill(298K)= 6 x 10-•6 -23 kiii= 5 x 10-•3 exp(-2000/T) CHO + 0 2-• CO + HO 2 kiv• 5 x 10-•2 -32 C2H4OH q-O2• C2H3OH q- HO 2 kv • 1.3x 10-•3 - 13.5 (for vibrationallyexcited C2H,•OH ) kv/kdcactivationby02 • 0.6 + 0 2--} + HO 2 -26 k,, < 10-•5(T = 350--400 K)

HO ß Forcomparison, thereaction H/•C=N q- 0 2-•}HOCN + 0 2would beexothermic by12.5 kcal/mol.

Vibrationallyexcited HCNOH would probablybe quenched [1983] of the decayrate of OH andthe generalbehavior of similarly.Also, for C2H5 it has beenshown that the addition recombinationreactions, we have adoptedthe following ex- of 02 proceedsabout 20 timesfaster than the formationof pressionfor the effectivealtitude-dependent rate constant, C2H4 + HO2, at leastat troposphericpressures [Plumb and k•(z),for the reactionof HCN and OH [Zellner,1978; Troe, Ryan,1981]. Finally, note that our candidatereaction requires 1979]: that an H atom be abstracted from an s-carbon atom that is double bonded to the N atom with a free-radicalposition, a situation unlike those for (CH3)2N + 0 2--• CH2 - N-CH3 kl(Z)-'I -•- ko[ko-•'•/k•' M] ]J X 0'8{1 +(log(•o[M•/•©)),} -1 (1) + HO 2 [Lindleyeta!., 1979] or C2H5 + 0 2--} C2H4 + HO 2. where[M] is total atmosphericdensity (cm-3) as a function We will proceednow to discussthe other, possibledominant of altitude, z, and ko= 1.5 x 10-31 exp (-875/T) pathsof HCNOH oxidation;further discussion of the conse- cmO/molecule2 s, and k• = 1.16x 10-13 exp (--400/T) quencesof a rapid bimolecularreaction with 02 is presented cm3/molecules, are the experimentallydetermined rate coef- below. ficientsin the limit of low and highpressure, respectively. At point B in Figure 1 a rate constantof 10-• cm3/molecules and 10 ppt of NO would lead to a lifetimeof 3. ATMOSPHERICMODEL CALCULATIONSAND DISCUSSION 1000 s for By performingcalculations with an atmosphericphoto- chemical/transportmodel, we can learnseveral things about HOM atmosphericHCN. For example,by assumingsteady states for /C = N-O0 its atmosphericvertical profile and for its ground level H sources,we can deduceits atmosphericresidence time. Fur- One can also imaginethat OH or CO could competewith ther, we can estimatethe total sizeof its sourcesby summing NO at this point. At point C in Figure 1, the proposedattack the loss rates for HCN at all altitudes. Individual HCN by 03 is by analogywith NH 2 + O3--} products,a reaction sources(discussed below) should, of course,sum to the total that is foundto be relativelyfast [Hack et al, 1982] and whose source estimate. productsare suggestedto be NH20 + 02. Also,at point F in The numerical model we used for thesecalculations is very Figure 1 the decay of nitrosoformaldehydeto NO + CHO similarto that describedby Cicerone[1979]. For HCN, we would occurin about 0.1 s [Napier and Norfish, 1967]. adopteda fixed-densitylower boundarycondition at (z = 0 If the structure of the HCNOH adduct is as shown at the km)equivalent to a volumemixing ratio of 1.7x 10-•ø.At top of Figure2, onemight imagine that the reactionat point/J themodel upper boundary (80 km)we employed the condition (reaction with NO by analogy to hydrocarbon oxidation) that the upwardflux, •p(80)of HCN equalsthe steadystate might be with HO 2 instead,to yield loss above that altitude. This flux was calculated from HO0X c)=fMH(J + k,[OH] + k2EO('D)]) (2) /C= N-OH+ 02 wherethe termin parenthesesis the sumof the first-orderloss H processesfor HCN (in s-•) dueto photolysis(J), reaction with This might occur in clean troposphericair where H O2 mixing OH (k•[OH])and with O(•D)(k2[O•D]).H andf are the ratios might be comparable to those of NO [Thompsonand verticalscale height and the volumemixing ratio of HCN at Cicerone, 1982]. Note that nitrosoformaldehyde would be 80 km, respectively.The threeloss processes for H CN were formed in the schemeof Figure 2 as in Figure 1. Thus, what- included in the model at all altitudes,z, from zero to 80 km. ever the structure of HCNOH, theseoxidation pathways sug- Valuesfor the altitude-dependentsteady state concentrations gestthat NO and CHO are formed within minutes.Of course, of [OH(z)] and [O(•DXz)] are shownin Table 4; theseare the CHO reacts quickly with 02 to yield CO + HO 2. Prob- from Cicerone et al. [1983] and Thompsonand Cicerone ably the slowestreactions in Figures 1 and 2 are those that [1982]and represent 24-hour averages. The verticalprofile of require NO. eddydiffusion coefficient was taken from thesesame refer- On the basis of the laboratory experimentsof Fritz et al. ences.The rate coefficientk•(z) is calculatedfrom (1); k2 has CICERONEAND ZELLNER' ATMOSPHERICCHEMISTRY OF HYDROGENCYANIDE 10,693

TABLE 4. Input Data for Atmospheric Model Calculations' OH mixing ratio between2.1 and 12 km, assuminga nearly con- and O(•D) DensitiesFrom Ciceroneet al. [1983] and Thompsonand stant mixing ratio with altitude. Our calculationshows that Cicerone[1982] this is a good assumption,that is, the mixing ratio of HCN Altitude, OH, O(•D), J k•(z), should decreaseonly about 5% between0 and 12 km. With km cm- 3 cm- 3 s- • cm3/molecule s OH densities twice those of Table 4 we calculate that the

80 4.81 (6) 17 1.2 (-6) 6.9 (- 19) Figure 3 the Coffeyet al. [1981] data are shownas a uniform 78 4.70 (6) 13 9.2 (-7) 1.0 (- 18) 76 4.66 (6) 10 6.6 (-7) 1.5 (- 18) mixingratio (1.55• 0.19)x 10-•ø between12 and 40 km. 74 4.72 (6) 11 4.5 (- 7) 2.3 (- 18) Actually, Coffey et al. reported a total column amount at all 72 4.80 (6) 14 2.9 (-7) 3.4 (- 18) altitudes above 12 km; we have indicated that their deduced 70 4.78 (6) 20 1.9 (- 7) 5.0 (- 18) mixing ratio applies to altitudes below 40 km becauseonly 68 4.74 (6) 27 1.2 (-7) 7.4 (- 18) about 1% of the measuredabsorption could arise from over 66 4.74 (6) 37 8.4 (-8) 1.1 (- 17) 64 4.86 (6) 49 6.0 (-8) 1.6 (- 17) 40 km even if HCN were distributed uniformly. In other 62 5.09 (6) 63 4.4 (-8) 2.2 (- 17) words, their data contain virtually no information on the 60 5.48 (6) 81 3.3 (-8) 3.0 (- 17) HCN above 40 km, so comparisonwith our calculatedprofile 58 6.04 (6) 102 2.4 (-8) 4.1 (- 17) is meaninglessfor altitudesabove, say, 40 km. Our model does 56 6.73 (6) 124 1.8 (-8) 5.5 (- 17) 54 7.65 (6) 148 1.3 (- 8) 7.2 (- 17) predicta slow but significantdecrease of HCN above 20 km. 52 8.77 (6) 172 9.6 (- 9) 9.4 (- 17) The Carli et al. [1982] data were reported as HCN mixing 50 1.01 (7) 195 7.0 (-9) 1.2 (- 16) ratios of 1.3 to 2.6 x 10-•o at all altitudes between 20 and 40 48 1.16 (7) 216 5.2 (-9) 1.5 (- 16) km. With OH densities50% those of Table 4, HCN decreases 46 1.30 (7) 229 3.8 (-9) 1.9 (- 16) more slowlywith altitude; at 40 km it is 1.13 timesthat shown 44 1.41 (7) 226 2.8 ( - 9) 2.4 ( - 16) 42 1.42(7) 197 2.0 (-9) 2.9 (- 16) in Figure 3. 40 1.29 (7) 152 1.4 (- 9) 3.5 (- 16) Attack by OH (with subsequent rapid oxidation to 38 1.07 (7) 108 9.6 (- 10) 4.3 (- 16) CO + NO, as shown in Figures 1 and 2) is the dominant 36 8.28 (6) 74 6.1 (- 10) 5.3 (- 16) HCN lossprocess throughout the troposphere.With the OH 34 5.96(6) 47 3.5 (- 10) 6.6 (- 16) 32 3.92 (6) 28 1.8 (- 10) 8.3 (- 16) and O(;D) densitiesof Table 4, the rate of attack by O(•D) 30 2.35 (6) 15 8.0 (- 11) 1.0 (- 15) dominates in the stratosphereabove 34 km, and photolysis 28 1.33 (6) 8 3.0 (- 11) 1.3 (- 15) matchesthe O(;D) rate at 54 km. Only about2% of the HCN 26 7.52 (5) 4 9.4 (- 12) 1.7 (- 15) (from presumed ground level sources)reaches the strato- 24 4.42 (5) 2 2.5 (- 12) 2.1 (- 15) sphere;the remaining98% is oxidizedin the troposphereby 22 2.78 (5) 1 5.3 (- 13) 2.6 (- 15) 20 1.95 (5) 0.6 9.0 ( - 14) 3.3 ( - 15) OH addition followed by the reactions shown in Figures 1 18 1.55 (5) 0.3 1.1 (- 14) 4.0 (- 15) and 2. This tropospheric oxidation of HCN represents a 16 1.47(5) 0.14 0 5.0 (- 15) source of NOx for the troposphere; it amounts to 4.5 14 3.08 (5) 0.03 0 6.0 (- 15) x 107/cm2 s whencolumn-integrated. If HCN is distributed 12 3.20 (5) 0.008 0 7.3 ( - 15) uniformlyover the globe(see below) this is 1.7 x 10TM g N/yr 10 2.13 (5) 0.004 0 8.8 (-15) 8 7.30 (5) 0 0 1.1 (- 14) or 3.6 x 10TM g NO/yr. This NOx sourceis only about one 6 8.60 (5) 0 0 1.4 (- 14) quarter as large as the productionrate of NOx (due to N20 4 8.10 (5) 0 0 1.7 (- 14) + O(•D)• 2 NO) in the stratosphere[Liu et al., 1980],most 2 7.30 (5) 0 0 1.9 (- 14) of which flows downward into the troposphere.About half of 0 3.20 (6) 0 0 2.2 (- 14) this NO would be produced between 0 and 2 km altitude. When OH densities twice those of Table 4 are used in the The reaction rate k•(z), for HCN + OH is from equation (1). Photodissociationfrequencies, J(s-•) werecalculated as describedin calculations,the column-integratedloss of HCN (production the text. All quantities represent24-hour averages.Further calculations were performed with OH densitiesequal to 50% and 200% those listed here. 6O i I i I %[II I] i i 50- Calculated been assumedto be 10-•ø cm3/molecules, independentof i iI I i it _1 pressureand temperatureand hencealtitude. By evaluating all terms in (2) at 80 km and iterating numerically, a suitable upper boundarycondition obtains.In additional calculations, c3 30- Carli et al. - OH densities were varied from those shown in Table 4. Be- Coffey e , causewe are not aware of any data on the photoabsorption • 2:0- crosssection of HCN as a function of wavelength(except for I0- 2 < 155 nm [Lee, 1980]), we have had to estimate this quan- tity. Specifically,we have assumedthat HCN photoabsorption 0 I RinslandI I I etI al.Ii1•• i •I i and photolysisshould be similar to that for HC1. A better i0-• i0-•o analogy might be with C2H2 (isoelectronicwith HCN), but we HCN MOLE FRACTION could find no suitableC2H2 data for wavelengthslonger than Fig. 3. Verticalprofile of HCN calculatedwith the atmospheric 200 nm. The J values shown in Table 4 for HCN are those for model described in the text and the data of Table 4. Also shown are HC1; they are taken from Ciceroneet al. [1983]. the troposphericobservations of Rinslandet al. [1982] and the strato- sphericobservations of Coffeyet al. [1981] and Carli et al. [1982]. Figure 3 showsour calculatedvertical profile of HCN along Lower troposphericconcentrations are likely in the southernhemi- with all availableHCN measurements.In the tropospherethe sphere'stroposphere (see text) and no appreciablediurnal variations Rinsland et al. [1982] data were reported as the average are expectedbelow about 40 km. 10,694 CICERONEAND ZELLNER:ATMOSPHERIC CHEMISTRY OF HYDROGENCYANIDE

of NO) essentiallydoubles. OH densitieshalf those of Table4 can occur, but reaction rates for each of these reactions is lead to about half as much HCN loss (NO production).The lower than k3 [Bullock and Cooper, 1972b], so nearly all residencetime, •:a,for HCN can be obtainedfrom CN--} NCO. Due to its strongabsorption bands between 360 and 450 nm [Holland et al., 1958], NCO must decompose TR-! '-- L(z)[HCN(z)]dz [HeN(z)] dz rapidly (minutes)in daylight to N + CO [I-lerzberg,1966]. An © © apparent consequenceof thesefacts is that decompositionof atmosphericC2N2 (if any) would not yield HeN. where L(z) is the total lossrate for HeN due to attack by OH, by O(!D) and photolysisat eachaltitude, z, and [HCN] is the The high rate of the reactionbetween CN and 02 can also density(cm -3) of HeN. For OH densitiesequal to thoseof serve to partially decouple the atmosphericchemistry of CH3CN and HeN. CN radicalsare formedin one of the two Table 4, twice those of Table 4, and half those of Table 4 we photolysispathways for CH3CN at 184.9 nm [McElcheran et find ,• = 2.5, 1.3,and 5.0 years,respectively. al., 1958]. Can this representa stratosphericsource of HeN? From this estimate of atmosphericresidence time, we may Becausek3102]/k,•[CH•] is between105 and 106 throughout proceed to predict southern hemispheric(SH) tropospheric the atmosphere,only 10-5 or 10-6 of the CN radicalsfrom mixing ratios of HeN which are as yet unmeasured.By as- CH3CN (or any other source)can become HeN. Fui'ther, if sumingthat the sourcesof HeN are primarily in the NH (see below) and applying the analysisthat Khalil and Rasmussen thereis onlyabout 10-!2 by volumeof CH3CNin themiddle stratosphere[I-lenschen and Arnold, 1981], the magnitude of [1983] used for CH4 (whose principal sink is also OH) we this sourceof CN radicalsis very small. Recent reports of estimate that the ratio of northern hemisphericHeN mixing much higher concentrationsof CH3CN near ground level ratio to that in the SH is 1.21 if ,• = 2.5 years.Accordingly, if thereis a NH mixingratio of 1.66x 10-!o (in the planetary [Beckerand Ionescu, 1982] requirefurther consideration. boundarylayer), then there would be 1.37x 10-!o in the SH. What is the origin of atmosphericHeN? Accordingto our calculations,an annualsource of (1-3) x 10TM g(N) as HeN is '_For,• = 1.3 and 5.0 years,the N/S ratio wo_uldbe 1.36 and 1.11,I respectively.All of thesevalues are approximate,of requiredto maintain the measuredatmospheric HeN burden. course. This annual sourceof HeN to the atmosphereis lessthan 1% Based on our calculations and the measurements summa- of the estimatedglobal annual amount of N in NO,, [Logan, 1983] from all sources.Significant indirect sourcessuch as the rized in Figure 3, HCN might be as abundant as any other odd-nitrogencompound in the middleand upper troposphere; atmosphericdecomposition of R-CN species(e.g., CH3CN ) are very unlikelyas discussedabove. Also, in situ injectionsby see Kley et al. [1981] for NO,, data and Sin•lh and Salas jet aircraft are apparentlyinsignificant [Robertson et al., 1979] [1983] for peroxyacetylnitrate data. Owing to its relatively low water solubility and weak acidity and to its relative becausewith adequateO2, turbine enginesproduce very little HeN. Some catalytic convertersthat reduceautomotive emis- photochemicalinertness, there shouldbe no significantdiurnal sionsof NO,,, when in ill repair, are known to produce high variation in HCN concentrations,and there are no obvious concentrations(500 ppm) of HeN. and major roles for HCN in troposphericchemistry. Further, The possibilitythat HeN is producedby lightninghas been due to overlapswith other major infrared absorbersthere is consideredby Chameidesand Walker [1981] and J. T. Kasting no significantrole for HCN in climate(V. Ramanathanand J. (private communication, 1981). With present epoch oxygen Kiehl, private communication,1982). levels,this HeN sourceappears negligible, although it should In the stratospherethere are two potentially interesting be noted that emissions from CN radicals are observed oc- paths for HCN oxidation that produceCN radicals' casionallyfrom lightning-disturbedair [Wallace, 1960; Sala- HCN+ hv-} H + CN nave et al., 1962]. Similarly, thermodynamicequilibrium cal- culationsindicate that volcanoescan emit HeN, especiallyin O(!D) +HCN--} OH + CN high-temperature,low-oxygen zones (R. Prinn, private com- It is alsopossible that theexothermic reaction O(3p) q- HCN munication, 1982), but this sourceis difficult to quantify, and + M--} HOCN + M, or the bimolecular abstraction of H it is probably relatively small. from the HCNOH adduct by O2 (discussedabove) followed Direct evident does exist for the release of HeN from sev- by HOCN + hv--} HO + CN could be important in the eral biological sourcesand from severalindustrial practices stratosphere.The fate of the CN radical needsexplored; it is suchas the productionof coke. The productionand releaseof conceivable that insights into the origin of stratospheric HeN from higher plants and from bacteriaand fungi is well CH3CN could be gained.First, it must be noted that documented [Conn, 1980; Bach, 1948; Robbinset al., 1950; Marshall and Hutchinson,1970], although no estimatesof CN + O2-• NCO + O (3) global releaserates into the atmosphereare available. These is a very fast reaction.At room temperature,its reaction rate, biogenicHeN sourcesclearly need further investigation.Re- k3,is 10- TMcm3/molecule s [see, e.g., Boden and Thrush, 1968; turning to nonbiogenic sources,large amounts of HeN Bullock and Cooper, 1972a; Albers et al., 1975]. No required appearin coke oven gaswhen coke is produced[Grosick and activation energy (or perhaps a slight negative activation Kovacic, 1981] and in other coal carbonizationprocesses. energy)was observedby Bullock and Cooper [1972a]. Clearly, Generally,high-temperature, low-oxygen processes are likely it would be difficult for any hydrogendonor, R-H, to compete to produce HeN; an examplemight be the processof steel with O2 for any available CN radicals. Formation of HeN degassingwhen Argon is usedin placeof oxygenas a purge through gas [Cottrell, 1967]. Many other hydrocarbon-richflames are throught to produce HeN, but little if any HeN escapes CN + CH½--• HeN + CH3 (4) oxidationto NO in general[Glassman, 1977]. Smoldering bio- CN + C2H 6 --• HeN + C2H 5 (5) massmight produce HeN and CH3CN; this should be explored.Thus, in principle,several likely sourcesexist for atmo- CN + Call 8• HeN + C3H8 (6) sphericHeN. Biologicalsources ivould be of greatinterest to CICERONEAND ZELLNER' ATMOSPHERICCHEMISTRY OF HYDROGEN CYANIDE 10,695 quantify to determineif atmosphericHCN is mostly natural Herzberg, G., ElectronicSpectra and ElectronicStructure of Polyato- or anthropogenic. rnic Molecules, D. Van Nostrand, New York, 1966. Herzberg, G., and K. K. Innes, Ultraviolet absorptionspectra of HCN and DCN, Can. J. Phys.,35, 842-879, 1957. Acknowledgments.R. Zellner thanks the Deutsche For- Hilgendorff,H. J., Die Absorptionsspektrenvon Blaus/iure,Hydrazin. schungsgenmeinschaftfor a Heisenbergstipendium.R. Cicerone ac- Athylen und Ammoniak im Schumann-Gebietund vond Hydrazin knowledgesexpert computing assistancefrom S. Walters. The Na- im Ouarzultraviolett, Z. Phys., 95, 781-788. 1935. tional Center for AtmosphericResearch is sponsoredby the National Holland, R., D. W. G. Style, R. N. Dixon, and D. A. Ramsay,Emis- Science Foundation. sion and absorptionspectra of NCO and NCS, Nature, 182, 336- 337, 1958. REFERENCES Howard, C. J., Kinetic study of the equilibrium HOe + NO • OH Albers, E. A., K. Hoyermann, H. Schacke,K. J. Schmatjko, H. G. + NO e and the thermochemistryof HO e, J. Am. Chern.Soc., 102, 6937-6941, 1980. Wagner, and J. Wolfrum, Absolute rate coefficientsfor the reac- InternationalCritical Tablesof NumericalData, vol. 3, Physics,Chem- tions of H atoms with NeO and reactionsof CN radicals,in 15th istry and Technology,McGraw-Hill, New York, 1928. InternationalSymposium on Combustion,pp. 765-772, Combustion Khalil, M. A. K., and R. A. Rasmussen,Sources, sinks and seasonal Institute, Pittsburgh, 1975. cyclesof atmosphericmethane, J. Geophys.Res., 88, 5131-5144, Arijs, E., D. Nevejans, and J. Ingels, Unambiguous mass determi- 1983. nation of major stratosphericpositive ions, Nature, 288, 684-686, 1980. Kley, D., J. W. Drummond, M. McFarland, and S.C. Liu, Tropo- Arnold, F., H. B/Jhringer,and G. Henschen,Composition measure- sphericprofiles of NO,,, J. Geophys.Res., 86, 3153-3161, 1981. Kneba, M., and J. Wolfrum, Bimolecular reactionsof vibrationally ments of stratosphericpositive ions, Geophys.Res. Lett., 5, 653- excitedmolecules, Annu. Rev. Phys.Chern., 31, 47-79, 1980. 656, 1978. Landolt-B/Srnstein,H. H., Zahlenwerte und Functionen aus Physik, Bach, E., On hydrocyanic acid formation in mushrooms,Physiol. Chernie, Astronornie,Geophysik und Technik, teil 2, pp. 1-29, Plant., 1,387-388, 1948. Springer-Verlag,New York, 1962. Becker, K. H., and A. Ionescu, Acetonitrile in the lower troposphere, Lee, L. C., CN(Ae•Xe•) and CN(B2• Xe•) yieldsfrom HCN Geophys.Res. Lett, 9, 1349-1351, 1982. photodissociation,J. Chern.Phys., 72, 6414-6421, 1980. Benson, S. W., ThermochemicalKinetics, 2nd ed., John Wiley, New Lewis, G. N., and M. Randall, Thermodynamicsand the Free Energy York, 1976. of ChemicalSubstances, 1st ed., 590 pp., McGraw-Hill, New York, Benson, S. W., Thermochemistry and kinetics of sulfur-containing 1923. molecules and radicals, Chern.Rev., 78, 23-35, 1978. Lindley, C. R. C., J. G. Calvert, and J. H. Shaw, Rate studiesof the Boden, J. C., and B. A. Thrush, Kinetics of reactions involving CN emission,IV, Study of the reactionsof CN by electronicabsorption reactionsof the (CH3)2N radical with 0 2, NO and NOe, Chem. Phys.Lett., 67, 57-62, 1979. spectroscopy,Proc. R. Soc.London Ser. A, 305, 107-123, 1968. Linke, W. F., Solubilitiesof Inorganic and Metal-Organic Compounds, B/Shringer,H., and F. Arnold, Acetonitrile in the stratosphere-- vol. 1, 4th ed., 1106pp., D. Van Nostrand, New York, 1958. Implicationsfrom laboratory studies,Nature, 290, 321-322, 1981. Liu, S.C., D. Kley, M. McFarland, J. D. Mahlman, and H. Levy III, Bullock, G. E., and R. Cooper, Reactions of cyanogen radicals, 2, On the origin of troposphericozone, J. Geophys.Res., 85, 7546- Reactionswith (CN)e and 02, J. Chern.Soc. Faraday Trans. 1, 68, 7552, 1980. 2175-2184, 1972a. Logan,J. A., Nitrogen oxidesin the troposphere:Global and regional Bullock, G. E., and R. Cooper, Reactions of cyanogen radicals, 3, budgets, J. Geophys. Res., this issue, 1983. Arrhenius parametersfor reactions with alkanes, J. Chern. Soc. Marshall, A.M., and S. A. Hutchinson, Biological activity of volatile Faraday Trans. 1, 68, 2185-2190, 1972b. metabolitesfrom cultures of fornes scuttelatus,Trans. Br. Mycol. Carli, B., F. Mencaraglia, and A. Bonetti, New assignmentsin the Soc., 55, 239-251, 1970. submillimeteremission spectrum of the stratosphere,lnt. J. Infra- McElcheran, D. E., M. H. J. Wijnen, and E. W. R. Steacie, The red Millimeter Waves, 3, 385-394, 1982. photolysisof methylcyanide at 1849,/•, Can. J. Chern.,36, 321-329, Chameides,W. L., and J. C. G. Walker, Rates of fixation by lightning 1958. of carbon and nitrogen in possibleprimitive atmospheres,Origins Napier, I. M., and R. G. W. Norrish, The photosynthesisof nitro- Life, 11, 291-302, 1981. soformaldehydeand the identificationof its spectrum,Proc. R. Soc. Cicerone,R. J., Atmosphericcarbon tetrafluoride:A nearly inert gas, London Ser. A, 299, 337-347, 1967. Science,206, 59-61, 1979. Okabe, H., Photochemistryof Small Molecules, John Wiley, New Cicerone, R. J., S. Walters, and S.C. Liu, Nonlinear responseof York, 1978. stratosphericozone column to chlorine injections,J. Geophys.Res, Peeters,J., V. Fonderie,G. Huybrechts,and D. Maes, paper presented 88, 3647-3661, 1983. at 7th InternationalSymposium on Gas Kinetics,G/Sttingen, 1982. Coffey, M. T., W. G. Mankin, and R. J. Cicerone,Spectroscopic detec- Phillips,L. F., Pressuredependence of the rate of reactionof OH with tion of stratospherichydrogen cyanide, Science, 214, 333-335, 1981. HCN, Chern.Phys. Lett., 57, 538-539, 1978. Conn, E. E., Cyanogeniccompounds, Annu. Rev. Plant. Physiol.,31, Phillips,L. F., Rate of reactionof OH with HCN between298 and 433-451, 1980. 563 K, Aust. J. Chern.,32, 2571-2577, 1979. Cottrell, A., An Introduction to Metallurgy, St. Martins, New York, 1967. Plumb, I. C., and K. R. Ryan, Kinetic studiesof the reactionof C2H Crutzen, P. J., L. E. Heidt, J.P. Krasnec, W. H. Pollock, and W. with O 2 at 295 K, lnt. J. Chern.Kinet., 13, 1011-1028,1981. Rinsland, C. P., M. A. H. Smith, P. L. Rinsland, A. Goldman, J. W. Seiler,Biomass burning as a sourceof atmosphericgases CO, Brault, and G. M. Stokes, Ground-based spectroscopicmeasure- NeO, NO, CH3Cl and COS, Nature, 282, 253-256, 1979. ments of atmospherichydrogen cyanide, $. Geophys.Res., 87, Davis, D. D., and H. Okabe, Determination of bond dissociation 11119-11125, 1982. energiesin hydrogen cyanide, cyanogenand cyanogenhalides by Robbins, W. J., A. Rolnick, and F. Kavanaugh, Production of hy- the photodissociationmethod, J. Chern.Phys., 49, 5526-5531, 1968. drocyanicacid by culturesof a basidiomycete,Mycologia, 52, 161- Fritz, B., K. Lorenz, W. Steinert, and R. Zellner, Rate of oxidation of 165, 1950. HCN by OH radicals at lower temperatures, Oxid. Cornrnun.,in press,1983. Robertson, D. J., R. H. Groth, and A. G. Glastris, HCN content of Glassman, I., Combustion,Academic, New York, 1977. turbine engine exhaust, J. Air Pollut. Control Assoc.,29, 50-51, 1979. Grosick, H. A., and J. E. Kovacic, Coke-oven gas and effluent treat- ment, in Chemistryof Coal Utilization, edited by M. A. Elliot, pp. Salanave,L. E., R. E. Orville, and C. N. Richards, Slitlessspectra of 1085-1152, Interscience,New York, 1981. lightning in the region from 3850 to 6900 angstroms,J. Geophys. Hack, W., O. Horie, and H. G. Wagner, Determination of the rate of Res, 67, 1877-1884, 1962. the reactionof NH e with 0 2,J. Phys.Chern., 86, 765-771, 1982. Singh, H. B., and L. J. Salas, Peroxyacetylnitratein the free tropo- Haynes, B. S., The oxidation of HCN in fuel-rich flames, Cornbust. sphere,Nature, 302, 326-328, 1983. Flame, 28, 113-121, 1977. Sullivan,B. J., G. P. Smith,and D. R. Crosley,Lifetimes in the Henschen,G., and F. Arnold, New positive ion speciesin the strato- state and the heat of formation of NCO, Chern. Phys. Lett., 96, sphere,Nature, 291, 211-213, 1981. 307-310, 1983. 10,696 CICERONEAND ZELLNER:ATMOSPHERIC CHEMISTRY OF HYDROGENCYANIDE

Thompson, A.M., and R. J. Cicerone, Clouds and wet removal as R. J. Cicerone,National Center for AtmosphericResearch, P.O. causesof variability in the trace-gas composition of the marine Box 3000, Boulder, CO 80307. troposphere,J. Geophys.Res., 87, 8811-8826, 1982. R. Zellner, Universit/it G/Sttingen,Institut fur Physikalische Troe, J., Predictive possibilitiesof unimolecularrate theory, J. Phys. Chemie,G/Sttingen, Federal Republic of Germany. Chern.,83, 114-126, 1978. Wallace, L., Note on the spectrumof lightning in the region 3670 to 4280•, J. Geophys.Res, 65, 1211-1214,1960. Wilhelm, E., R. Battino, and R. L. Wilcock, Low-pressuresolubility of gasesin liquid water, Chern.Ret•., 77, 219-262, 1977. (ReceivedJune 7, 1983; Zellner, R., Recombination reactionsin atmosphericchemistry, Ber. revisedSeptember 23, 1983; Bunsenvies.Phys. Chern., 82, 1172-1179, 1978. acceptedSeptember 23, 1983.)