Complex chemical composition of colored surface films formed from reactions of propanal in ...

15 Pages·2015·4.33 MB·English

Atmos.Chem.Phys.,15,4225–4239,2015
www.atmos-chem-phys.net/15/4225/2015/
doi:10.5194/acp-15-4225-2015
©Author(s)2015.CCAttribution3.0License.
Complex chemical composition of colored surface films formed from
reactions of propanal in sulfuric acid at upper troposphere/lower
stratosphere aerosol acidities
A.L.VanWyngarden1,S.Pérez-Montaño1,J.V.H.Bui1,E.S.W.Li1,T.E.Nelson1,K.T.Ha1,L.Leong1,and
L.T.Iraci2
1DepartmentofChemistry,SanJoséStateUniversity,SanJosé,CA95192,USA
2AtmosphericScienceBranch,NASAAmesResearchCenter,MoffettField,CA94035,USA
Correspondenceto:A.L.VanWyngarden([email protected])
Received:27October2014–PublishedinAtmos.Chem.Phys.Discuss.:19November2014
Revised:28March2015–Accepted:1April2015–Published:24April2015
Abstract.Particlesintheuppertroposphereandlowerstrato- propanalwithglyoxaland/ormethylglyoxalalsoshowedev-
sphere(UT/LS)consistmostlyofconcentratedsulfuricacid idence of products of cross-reactions. Since cross-reactions
(40–80wt%) in water. However, airborne measurements would be more likely than self-reactions under atmospheric
haveshownthattheseparticlesalsocontainasignificantfrac- conditions,similarreactionsofaldehydeslikepropanalwith
tion of organic compounds of unknown chemical compo- commonaerosolorganicspecieslikeglyoxalandmethylgly-
sition. Acid-catalyzed reactions of carbonyl species are be- oxalhavethepotentialtoproducesignificantorganicaerosol
lievedtoberesponsibleforsignificanttransferofgasphase mass and therefore could potentially impact chemical, opti-
organicspeciesintotroposphericaerosolsandarepotentially caland/orcloud-formingpropertiesofaerosols,especiallyif
moreimportantatthehighaciditiescharacteristicofUT/LS theproductspartitiontotheaerosolsurface.
particles.Inthisstudy,experimentscombiningsulfuricacid
(H SO ) with propanal and with mixtures of propanal with
2 4
glyoxal and/or methylglyoxal at acidities typical of UT/LS
1 Introduction
aerosols produced highly colored surface films (and solu-
tions) that may have implications for aerosol properties. In
Aerosols in the upper troposphere and lower stratosphere
order to identify the chemical processes responsible for the
(UT/LS) are composed primarily of sulfuric acid (40–
formation of the surface films, attenuated total reflectance–
80wt%)(Cleggetal.,1998;Finlayson-PittsandPitts,2000;
Fouriertransforminfrared(ATR-FTIR)and1Hnuclearmag-
Tabazadehetal.,1997)andwater,buttheyalsocontainsig-
netic resonance (NMR) spectroscopies were used to ana-
nificantfractionsoforganiccompounds(Froydetal.,2009;
lyze the chemical composition of the films. Films formed
Murphyetal.,2007,2014,1998).InthecaseofUTaerosols,
frompropanalwereacomplexmixtureofaldolcondensation
theamountoforganicmaterialcanevenexceedtheamount
products, acetals and propanal itself. The major aldol con-
of sulfate present (Murphy et al., 1998). The potential im-
densationproductswerethedimer(2-methyl-2-pentenal)and
pactsofthisorganicmaterialonchemical,opticalandcloud-
1,3,5-trimethylbenzenethatwasformedbycyclizationofthe
forming properties of UT/LS aerosols are highly uncertain
linearaldolcondensationtrimer.Additionally,thestrongvis-
since relatively little is known about the chemical composi-
ibleabsorptionofthefilmsindicatesthathigher-orderaldol
tionoftheorganicfractionbecauseavailablesamplingtech-
condensationproductsmustalsobepresentasminorspecies.
niques and frequencies are limited by the high-altitude air-
The major acetal species were 2,4,6-triethyl-1,3,5-trioxane
bornemissionsrequired.
and longer-chain linear polyacetals which are likely to sep-
arate from the aqueous phase. Films formed on mixtures of
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.
4226 A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO
2 4
In contrast to UT/LS aerosols, tropospheric aerosols are 2007; McNeill et al., 2006; Park et al., 2007; Riemer et al.,
better sampled so it is well established that they contain 2009; Thornton and Abbatt, 2005). Similarly, organic coat-
major fractions of organics (up to 90%) (e.g., Calvo et al., ingsonsulfateaerosolswouldalteropticalproperties,espe-
2013; Hallquist et al., 2009; Jacobson et al., 2000; Jimenez ciallyiftheorganicsarehighlyabsorbingintheUV–visible.
et al., 2009; Kanakidou et al., 2005; Murphy et al., 2006; In order to assess whether species that form surface films
Zhangetal.,2007),andtherehavebeenmanystudiesaimed onpropanal/H SO mixturesinthelaboratorycouldbeim-
2 4
atchemicalcharacterizationoftroposphericorganicaerosol portantinUT/LSaerosols,thereactionsresponsibleforfilm
(OA) particles and at determining the physical/chemical formation must be identified, which is, therefore, the focus
pathwaysfortheformationofOA.Inparticular,reactionsof ofthepresentwork.
carbonyl-containingorganicspeciesincludingaldolconden- Recentworkwithvariousotheraldehydes(Lietal.,2011;
sation,hemiacetal/acetalformation,organosulfateformation Sareen et al., 2010; Schwier et al., 2010) demonstrated that
andvariouspolymerizationreactionshaveallbeenidentified products of reactions of formaldehyde, acetaldehyde, gly-
aspotentialsourcesoflow-volatilityorganicproductsintro- oxal, methylglyoxal and their mixtures are surface-active
posphericorganicaerosols(BarsantiandPankow,2004;Er- even in water and ammonium sulfate/water solutions char-
vens and Volkamer, 2010; Gao et al., 2004; Garland et al., acteristic of less acidic lower tropospheric aerosols. Their
2006; Holmes and Petrucci, 2007; Jang et al., 2002, 2004; chemicalcharacterizationofthereactionproductsidentified
Kalberer et al., 2004; Liggio and Li, 2006, 2008; Liggio et hemiacetal oligomers and aldol condensation products, but
al., 2007; Lim et al., 2010; Michelsen et al., 2004; Nozière thesurface-activespecieswerenotspecificallyidentified.
and Esteve, 2007; Nozière and Riemer, 2003; Sareen et al., In order to identify the chemical species present in films
2010; Shapiro et al., 2009; Surratt et al., 2007, 2006; Tan formedbypropanalandsulfuricacid,weconsidertheprod-
etal.,2010;Tolockaetal.,2004;Zhaoetal.,2005;Ziemann uctsofthefollowingpotentialreactions(identifiedbyalet-
andAtkinson,2012).Sincethesereactionsarealleitheracid- ter in Fig. 1): (a) aldol condensation, (b) trimethylbenzene
catalyzed orrequire sulfate, theyare likely to beeven more formation via cyclization of the linear trimer produced by
favorable at the high sulfuric acid concentrations typical of aldol condensation, (c) hemiacetal, acetal and/or polyacetal
UT/LSaerosols. formation,(d)trioxaneformationviacyclotrimerizationand
Preliminary experiments for the current work, in which (e) organosulfate formation. Each of these processes result
various carbonyl species (propanal, glyoxal and/or methyl- in higher molecular weight products, which could result in
glyoxal) were combined with highly concentrated sulfuric partitioningtothesolidphaseasasurfacefilm.
acid to simulate UT/LS aerosol acidities, produced highly Aldol condensation products are expected since it can be
colored solutions; solutions containing propanal also pro- seen from Fig. 1 that they are the only potential products
duced reaction products that partitioned to the liquid sur- containingsufficientconjugationtoabsorbvisiblelight,but
face as macroscopic semi-solid surface films that were also they are not necessarily the major component of the films
highly colored. The possibility that similar organic prod- since only tiny amounts of such chromophores are neces-
ucts could partition to thin layers or films on the surface sary for color (McLaren, 1983). Products of propanal aldol
of UT/LS aerosols is of particular interest because organic condensation reactions have been observed in aqueous me-
compounds that coat aerosol particles would have the most dia containing various catalysts, including anion exchange
dramatic effects on aerosol chemical, optical and/or cloud- resin(Pyoetal.,2011),ammoniumandcarbonatesalts(Noz-
formingproperties(seeDonaldsonandVaida,2006,andMc- ière et al., 2010), mixed metal oxides (Tichit et al., 2002)
Neilletal.,2013,forreviewsofaerosolsurfacecoatingsand and zeolites (Hoang et al., 2010). In the case of zeolite cat-
their impacts on aerosol properties). For example, organic alysts, 1,3,5-trimethylbenzene was also observed and pro-
coatingsonaqueousdropletsandsulfuricacidaerosolshave posedtoformfromthelineartrimerproducedbyaldolcon-
beenobservedtoimpedewateruptakeand/orevaporationin densation reactions (Fig. 1b). Aldol condensation reactions
laboratory experiments (e.g., Davies et al., 2013; Otani and of propanal have also been studied in concentrated sulfuric
Wang, 1984; Rubel and Gentry, 1984; Seaver et al., 1992; acid (60–96wt%) solutions by Nozière and Esteve (2007)
Xiong et al., 1998), so organic coatings on UT/LS aerosols andCasaleetal.(2007).NozièreandEsteve(2007)reported
and/ordroplets couldpotentiallyinhibit watercondensation theUV–visiblespectraofaldolcondensationproductsofsix
and therefore cloud formation and/or growth. Organic coat- carbonylcompoundsincludingpropanalandconcludedthat
ingsmayalsoimpactheterogeneousreactionsataerosolsur- theirabsorptionindexcouldbecomesignificantovertheap-
faces;forexample,reactiveuptakeofN O hasbeenshown proximately2-yearresidencetimeofstratosphericaerosols.
2 5
to be impeded by various organic coatings which could re- Casale et al. (2007) measured bulk reaction rates for a se-
duce the rate of hydrolysis of N O to HNO on sulfuric riesofaliphaticaldehydes(C –C ),showingthatbutanaland
2 5 3 2 8
acidaerosols,affectingNO andOHbudgets(Anttilaetal., propanal had the highest reaction rates but concluding that
x
2006;Badgeretal.,2006;CosmanandBertram,2008;Cos- therateswerenotfastenoughtoberesponsiblefortransfer
man et al., 2008; Escorcia et al., 2010; Evans and Jacob, of significant organic mass into tropospheric aerosols. Both
2005;Folkersetal.,2003;Gastonetal.,2014;Knopfetal., studiesfocusedonaldolcondensationreactionsduetotheir
Atmos.Chem.Phys.,15,4225–4239,2015 www.atmos-chem-phys.net/15/4225/2015/
A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO 4227
2 4
a. Aldolcondensation
aldol aldolcondensation products
B
+ …
‐ A
enol 2‐methyl‐2‐pentenal 2,4‐dimethyl‐2,4‐heptadienal 
b. Trimethylbenzeneformation
‐
2,4‐dimethyl‐2,4‐heptadienal  1,3,5‐trimethylbenzene
c. Hemiacetal/acetalformation and further polymerization
n
diol(hydrate) hemiacetal acetal polyacetal
d. Cyclotrimerization
2,4,6‐triethyl‐1,3,5‐trioxane 
e. Organosulfateformation
Figure1.Potentialreactionsofpropanalinthepresenceofsulfuricacid.SelectedhydrogenpositionsarelabeledAandBinbluetofacilitate
discussioninthetext.
potential to form light absorbing compounds and therefore Allen, 2004, 2008; Vinnik et al., 1986), so alcohol species
used UV–visible spectroscopy for product detection, which including the diol formed by hydration of propanal and/or
is not sensitive to products of the other potential reactions (hemi-)acetals (Surratt et al., 2008) formed from propanal
(Fig.1c–e)consideredhere. (Fig. 1c) could react directly with sulfuric acid to form
Propanal may also undergo acid-catalyzed reactions with organosulfatessimilartothoseformedbyreactionofglyoxal
its hydrated form (diol) to form hemiacetals, acetals and/or onsulfuricacidaerosols(Liggioetal.,2005).Anexampleis
linear polyacetals as shown in Fig. 1c. In addition to these shownforreactionofthepropanalhydrateinFig.1e.
linearspecies,propanalmayalsoundergoacid-catalyzedcy- In the present study we first employ a combination of at-
clotrimerizationtoformacyclicpolyacetal(atrioxane)(Fig. tenuated total reflectance–Fourier transform infrared (ATR-
1d). These reactions have not been reported specifically for FTIR)and1Hnuclearmagneticresonance(1HNMR)spec-
propanal in sulfuric acid, but Garland et al. (2006) have troscopies to identify the major species in the films formed
shown that sulfuric acid aerosols exposed to hexanal vapor by propanal on sulfuric acidsolutions. In order to approach
containedhemiacetals,whileLietal.(2008)identifiedatri- more atmospherically realistic mixtures of organics and to
oxaneinbulkreactionsofoctanalwithsulfuricacid(butnot address the possibility of cross-reactions between different
in sulfuric acid aerosols exposed to octanal vapor). In both carbonyl species, we also examined films formed on mix-
studies,aldolcondensationproductswerealsoobserved.Fur- turesofpropanalwithglyoxaland/ormethylglyoxal.Finally,
thermore, propanal has been shown to form a trioxane in wealsousedUV–visiblespectroscopyoftheliquidsolutions
aqueous solution (Corrochano et al., 2010) and to form a togainchemicalinsightintotheidentityofthechromophores
mixtureofaldolcondensationproducts,hemiacetalsandac- andtoillustratetheirpotentialimportanceforUT/LSaerosol
etalsinthepresenceofananion-exchangeresincatalyst(Pyo opticalproperties.
etal.,2011).
Lastly, alcohols may react with sulfuric acid to form sul-
fate esters (Deno and Newman, 1950; Iraci et al., 2002;
Michelsenetal.,2006;Minerathetal.,2008;VanLoonand
www.atmos-chem-phys.net/15/4225/2015/ Atmos.Chem.Phys.,15,4225–4239,2015
4228 A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO
2 4
2 Experimentalmethods analysis by FTIR spectroscopy. ATR-FTIR spectra of the
films and standards were collected on a Nicolet 6700 spec-
Surface films were first detected on solutions of propanal trophotometer from 4000 to 700cm−1 at 1cm−1 resolution
anditsmixtureswithglyoxaland/ormethylglyoxalinsulfu- usingamercurycadmiumtelluride(MCT)detectoranda10-
ricacid,whichwereallowedtoreactforseveralweeks(see bounce AMTIR ATR crystal with 45◦ mirrors from PIKE.
Fig.S1intheSupplementforphotosoftypicalsurfacefilms). ATR-FTIRwaschosenforchemicalanalysissincethesemi-
Subsequently, controlled survey studies were performed to solidfilmscouldbedirectlyanalyzedonacrystalcompatible
examine the conditions required for formation of surface with concentrated sulfuric acid and without any need to al-
films. In these experiments, samples of propanal, glyoxal ter the chemical environment by dissolving the sample in a
and/ormethylglyoxalinallpossiblecombinationsof1,2or solvent.Inordertoprovidemorechemicalspecificity,films
all3species(0.030Mineachorganicpresent)wereprepared were also analyzed by 1H NMR spectroscopy using a Var-
in stock solutions of 19, 37, 48 and 76wt% sulfuric acid ianINOVA400MHzspectrometer.NMRsampleswerepre-
(H SO ).Sinceinitialexperimentsindicatedthatsolutionsof pared by dissolving film samples in deuterated chloroform
2 4
glyoxaland/ormethylglyoxaldidnotformfilmswithoutthe (CDCl )inquartzNMRtubes(5mmouterdiameter).ATR-
3
presence of propanal (ultimately confirmed by these survey FTIRand/orNMRspectrawerealsorecordedforthefollow-
experiments),concentrationsofmixedorganicswerechosen ing commercially available standards: 2-methyl-2-pentenal
tokeepthepropanalconcentrationconstant,sothatanydif- (97wt%Sigma-Aldrich),1,3,5-trimethylbenzeneand2,4,6-
ferencesinfilmformationratesinthemixturescomparedto triethyl-1,3,5-trioxane(AKosGmbH,customsynthesis).
propanalalonecouldnotsimplybeduetoadifferentconcen- Finally, the UV–visible absorption spectra of solutions
trationofpropanalandthereforewouldindicatethatglyoxal (0.030Mineachorganic)wereobtainedusingaVarianCary
and/or methylglyoxal could impact the ability of propanal 50 Bio UV–visible spectrometer with a diode array detec-
to form films. This results in samples that have a total or- tor and quartz cuvettes of various path lengths from 0.01 to
ganicconcentrationthatincreaseswiththenumberoforgan- 10mm for different regions of the spectrum. Prior to anal-
icspresentupto0.09Mforsolutionsthatcontainallthreeor- ysis, solutions were filtered through 2.5µm Teflon filters to
ganics.AlthoughUT/LSaerosolconcentrationsoftheseor- removeanysuspendedsolidparticulates.
ganiccompoundsareunknown,0.03Mislikelymuchlarger
than UT/LS concentrations of any one carbonyl species but
ismorereasonableifconsideredasrepresentativeofthetotal 3 Results
aldehydeorcarbonylconcentration.Sulfuricacidstocksolu-
tionswerepreparedbydilutionofconcentratedsulfuricacid 3.1 Formationoforganicsurfacefilms
(96–98wt%, Sigma-Aldrich, ACS grade) with Milli-Q wa-
ter,andconcentrationswereconfirmedbytitrationwithstan- Carbonyl-containing organics (propanal, glyoxal and/or
dardizedsodiumhydroxide(0.5N,Sigma-Aldrich).Thefol- methylglyoxal) mixed with sulfuric acid (19–76wt%) to
lowing Sigma-Aldrich organics were used: 97wt% reagent simulateUT/LSaerosolaciditiesproducedcoloredsolutions,
gradepropanal,40wt%glyoxaland40wt%methylglyoxal precipitatesandsurfacefilms.Atthehighestacidities,allin-
in water. 4.0mL aliquots of each mixture were transferred dividual organics and organic mixtures examined (0.030M
to multiple 8mL glass vials and stored under each of the ineachorganic)producedvisiblycoloredsolutionsthatdark-
following temperature and lighting conditions: room tem- ened with time. Mixtures containing propanal produced the
perature(21–24◦C)/constantfluorescentlight,roomtemper- most deeply colored solutions, progressing from yellow to
ature/dark, 0◦C/dark, −19◦C/dark. Samples were visually orangetoredtobrownovertimescalesrangingfromminutes
monitored daily for color changes and formation of surface to months. This color darkening progressed faster at higher
films in order to survey which mixtures formed films and acidities, consistent with an acid-catalyzed reaction. Many
toassesstheimpactofacidity,organicmixture,temperature propanal-containing mixtures also eventually produced col-
andfluorescentlightonfilmformationrates. ored precipitates and/or surface films (mixtures containing
Chemicalanalysisofthefilmsrequiredproductionoffilms only glyoxal and/or methylglyoxal did not produce surface
in sufficient quantity to allow physical removal of a portion films). These solids or semi-solids were observed either as
withoutdisturbingtheunderlyingsulfuricacidsolutionsand particlessuspendedintheliquid(usuallycollectingnearthe
therebyavoidingspectroscopicinterferencesfromwaterand surface) and/or as semi-rigid macroscopic films on the sur-
sulfuric acid. Therefore, samples used for chemical analy- face. In principle, the films could potentially be formed ei-
sis were prepared as above, except at the higher concentra- therbyheterogeneousreactionsattheair/liquidinterfaceor
tion of 0.30M in each organic and were stored in volumet- byliquid-phasereactionsresultinginproductsthatpartition
ricflasks(roomtemperature/fluorescentlight)whichcaused to the surface. The latter process, however, is supported by
the film to concentrate on the small liquid surface area in theobservationthatwhensolutionswerestoredinvolumet-
theneckoftheflask.Filmsampleswereremovedandtrans- ricflaskssolid,dark-coloredmaterialsometimescollectedon
ferred with a glass rod to the surface of an ATR crystal for theupperslantedwallsinthebodyoftheflaskbeforemigrat-
Atmos.Chem.Phys.,15,4225–4239,2015 www.atmos-chem-phys.net/15/4225/2015/
A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO 4229
2 4
ingtothesurface;presumablythematerialroseduetoitslow FTIRand1HNMRspectroscopieswereusedtoanalyzethe
densityrelativetothesolutionbutwastemporarilyimpeded chemical composition of the surface films. The combined
fromreachingthesurfacebytheflaskwalls.Furthermore,the results of these two techniques provide evidence that the
quantityoffilmmaterialobservedcannotbeeasilyexplained films are a mixture of aldol condensation products (mainly
byheterogeneoussurfacereactionsalone. 2-methyl-2-pentenaland1,3,5-trimethylbenzene)andacetals
There was variability in film formation rates for repli- (mainly2,4,6-triethyl-1,3,5-trioxaneandlonger-chainlinear
cates of the survey experiments most likely due to differ- polyacetals) as detailed in Sect 3.2.1 through 3.2.3 below.
ences in the gentle movement of the samples that was re- Thedetailedchemicalanalysisinthesesectionsispresented
quiredtodetectfilmsduringdailyvisualobservations;how- forsurfacefilmsformedon0.30Mpropanal/48wt%H SO
2 4
ever,thefollowinggeneraltrendsemerged(seeFigs.S5and as a starting point, since surface films were only formed
S6 in the Supplement for trends and variability). First, the onsolutionscontainingpropanalandpropanalformedfilms
precise dependence of film-formation rate on acidity was fastest at 48wt% H SO . These films were stored at room
2 4
complex, but, in general, the films formed faster at higher temperature under constant fluorescent light and were sam-
acidity,consistentwithacid-catalyzedprocesses.Infact,the pled and analyzed 7 days after mixing the solutions. Sec-
mostacidic(76wt%H SO )propanal/glyoxalmixturepro- tion3.3–3.5subsequentlyaddresstheimpactofvaryingthe
2 4
duced a surface film immediately upon combining the re- temperature, illumination, organic concentration, film age,
actants, although the other organic mixtures formed films acidityandorganicmixturefromthisbasecase.
moreslowlyat76wt%thanat48wt%H SO .Specifically,
2 4
films were first observed on propanal-only samples after 3.2.1 Aldolcondensationproducts
4 days in 48wt% H SO vs. 5–10 days in 76wt% H SO ,
2 4 2 4
and visible film formation on propanal/methylglyoxal and Figure2presentsatypicalATR-FTIRspectrumofasurface
propanal/glyoxal/methylglyoxalsamplesrequired5–22days filmformedona7-day-old0.30Mpropanal/48wt%H SO
2 4
in48wt%H SO ,whilesamplesin76wt%H SO stilldid mixture (in green) along with spectra of four standards for
2 4 2 4
nothavevisiblefilmsafter180days.Second,film-formation comparison. The strong absorption band in the film spec-
ratesalsovariedasafunctionoforganicmixture.Ingeneral, trum at 1689cm−1 and the band at 1643cm−1 are consis-
mostmixturescontainingglyoxalformedfilmsmorerapidly tent with the characteristic C=O and C=C stretching vi-
thanthosewithout,whilemixturescontainingmethylglyoxal brations,respectively,ofanα,β-unsaturatedaldehydewhich
consistently formed films more slowly whenever there was is produced by aldol condensation (Fig. 1a). The spectrum
a detectable difference in rates (see Fig. S6 in the Supple- of neat2-methyl-2-pentenalshown inFig. 2(blue) displays
ment).Third,filmsformedbothinthedarkandunderfluores- these bands at 1687 and 1643cm−1 and is scaled to illus-
centlightwithnoconsistenttrendinformationrate.Finally, trate the maximum amount of the film spectrum that could
filmsformeddaystomonthsmoreslowlyatcoldertempera- beexplainedbyitspresence(limitedbythesizeoftheC=C
turesbut,importantlyforapplicationtothecoldUT/LS,were bandat1643cm−1).AnadditionalC=Opeakat1722cm−1
eventually observed (after approximately 100 days) even at occursinthesaturatedaldehydestretchingregionandisas-
thelowesttemperature(−19◦C)examined. signedtounreactedpropanal.InFig.2,thespectrumofneat
propanal(red)isalsoscaledtoillustrateitspotentialcontri-
3.2 Chemicalcompositionofsurfacefilms butiontothespectrumofthefilm.
The 1H NMR spectrum for this film presented in Fig. 3
Thehighlycolorednatureofthesurfacefilms(onlyformed indicates that 2-methyl-2-pentenal is the dominant species
onsolutionscontainingpropanal)isstrongevidenceforaldol sinceitcontainsstrongpeaks(assignedinFig.3)correspond-
condensationproducts,asaldolcondensationistheonlypo- ing to all five types of hydrogens in 2-methyl-2-pentenal in
tentialreaction(Fig.1)ofpropanalinsulfuricacidthatcan thecorrectmultiplicityandwithin0.03ppmofourstandard.
resultinproductswiththeconjugationrequiredtocauseab- Although some of the peaks are too small or too close to
sorptionofvisiblelight.Infact,multiplealdolcondensation interferingpeakstointegratereliably,therelativepeakinten-
steps are required to produce sufficient conjugation, since sities are also roughly consistent with the standard. Resid-
thefirstaldolcondensationproductofpropanal(2-methyl-2- ual propanal is similarly positively identified by compari-
pentenal,seeFig.1a)iscolorlesswithλ fortheπ →π* sontothestandardasshownbypeakassignmentsinFig.3.
max
transitionof∼266and∼233nmin75wt%H SO (Casale There are no additional detectable NMR peaks consistent
2 4
et al., 2007) and water (our standard), respectively. Further with linear compounds with additional units of conjugation
conjugationfromadditionalaldolcondensationreactionsof duetomultiplealdolcondensationsteps,indicatingthatthey
2-methyl-2-pentenal with propanal or with itself is required mustbesignificantlylessabundantthan2-methyl-2-pentenal
to shift absorption into the visible. Although products from and therefore will not contribute substantially to the FTIR
multiple aldol condensation steps are almost certainly re- spectrum either. For example, the protons labeled A and B
sponsibleforthefilmcolor,thesechromophoresarenotnec- in 2,4-dimethyl-2,4-heptadienal (Fig. 1) would be expected
essarilythemajorchemicalcomponentsofthefilms,soATR- to appear as singlets with chemical shifts near those for
www.atmos-chem-phys.net/15/4225/2015/ Atmos.Chem.Phys.,15,4225–4239,2015
4230 A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO
2 4
9 4 5 - P o ly m e r
I
F ilm fo rm e d o n 0 .3 M p ro p a n a l in 4 8 w t.% H S O
2 4
TT
S ta n d a rd s (n e a t):
T
p ro p a n a l (P )
2 -m e th y l-2 -p e n te n a l (2 M 2 P ) I
1 ,3 ,5 -trim e th y lb e n z e n e (T M B )
0 .6
2 ,4 ,6 -trie th y l-1 ,3 ,5 -trio x a n e (T )
a ,b -u n s a tu ra te d T T
a ld e h y d e (m o s tly 2 M 2 P ) I
e 1 6 8 9
c
n C = O
rba0 .4 I T
o
bs I T
A
I
T 8 3 4
0 .2 p r o p1a7n2a2l 1 6 4 3 I T T M B
C = C I
I I 1 6 0 9
T M B
I
0 .0
3 6 0 0 3 4 0 0 3 2 0 0 3 0 0 0 2 8 0 0 2 6 0 0 1 8 0 0 1 6 0 0 1 4 0 0 1 2 0 0 1 0 0 0 8 0 0
W a v e n u m b e r (c m -1)
Figure2.TypicalATR-FTIRspectrumofasurfacefilmformedon0.30Mpropanalin48wt%H2SO4 (7daysaftermixing)comparedto
neatstandards.Spectraofstandardsforpropanal,2-methyl-2-pentenal,1,3,5-trimethylbenzeneand2,4,6-triethyl-1,3,5-trioxanearescaledto
indicatetheirmaximumpossiblecontributiontothefilmspectrum.Thepositionsofthemajortrioxanepeaksareindicatedwiththeabbrevi-
ationTtoillustratetheirpresenceinthespectrumofthefilm.Otherimportantpeaksarelabeledwithwavenumberandtheirassignmentsas
discussedinthetext.Notethattheregionfrom2500to1800cm−1 lackspeaksandisomittedforclarity.Fordetailsofthelower-intensity
tracesseeFig.S2intheSupplement,whichprovidesaversionofthisfigurecoveringthesmallerabsorbancerangeof0–0.15.
2,4-hexadienal (Spectral Database for Organic Compounds, and, therefore, cannot be explained by any other potential
SDBS,2014)at9.5and7.1ppm,respectively. products.
Although there is no NMR evidence for linear aldol Although 2-methyl-2-pentenal and 1,3,5-trimethyl-
condensation products beyond 2-methyl-2-pentenal, NMR benzene are shown here to be the major products resulting
peaks at 2.26 and 6.79ppm confirm the presence of 1,3,5- from aldol reactions, both are colorless, so the more highly
trimethylbenzene (mesitylene) (SDBS), which has previ- conjugated compounds formed by further aldol conden-
ously been observed to form in reactions of propanal over sation steps that are presumably responsible for the film
acidic zeolite catalysts (Hoang et al., 2010). Hoang et color must be minor constituents. Therefore, there must
al.(2010)proposedthat1,3,5-trimethylbenzenewasformed also be additional compounds present in the film to explain
by acid-catalyzed cyclization and subsequent dehydration the strength of the peaks that appear in the FTIR spectrum
of the trimer formed by aldol condensation (2,4-dimethyl- between1500–800and3000–2800cm−1.
2,4-heptadienal) of propanal as shown in Fig. 1b, which
is a reasonable mechanism for sulfuric acid solutions as 3.2.2 Ethers:acetals/hemiacetalsandlinear/cyclic
well. Furthermore, the trimer formed by aldol condensation polyacetals
of acetone has also been shown to cyclize to form 1,3,5-
In addition to aldol condensation products, the FTIR and
trimethylbenzeneinsulfuricacid(Duncanetal.,1998;Kane
NMRspectrabothalsodisplayevidenceforethergroups(C-
et al., 1999; Klassen et al., 1999). In Fig. 2, the ATR-FTIR
O-C) due to strong peaks in the 1200–1000cm−1 region of
spectrumofneat1,3,5-trimethylbenzene(black)isscaledto
theFTIRspectrum(Fig.2)andpeaksinthe4.5–5.1ppmre-
thefilmspectrumtoindicateitsmaximumpotentialcontribu-
gionoftheNMRspectrum.Speciesthatcouldberesponsible
tion.Thecomparisonshowsthatthefilmspectrumisconsis-
fortheseethersignaturesincludehemiacetals,acetalsand/or
tentwiththepresenceofsome1,3,5-trimethylbenzeneinthe
higher-order polyacetal polymers which can form from the
filmsinceithasbandscorrespondingtothetwomostintense
1,3,5-trimethylbenzenebandsat834and1609cm−1,thelat- reactionofpropanalwithoneormoreofitshydrates(diols)
(Fig.1c)orfromcyclotrimerizationofpropanaltoformthe
terofwhichliesintheregionforaromaticskeletalvibrations
cyclicacetal,2,4,6-triethyl-1,3,5-trioxane(Fig.1d).Ofthese
Atmos.Chem.Phys.,15,4225–4239,2015 www.atmos-chem-phys.net/15/4225/2015/
A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO 4231
2 4
P
2P 2P, o
M M P
2 2 &
4 0 T
M2P − 1.7 1.11, 1.1 0.94
2 P −
− 9.79 P − 9.39 − 7.27 CHCl3 − 6.79 TMB − 6.48 2M2P − 4.98 Po ? − 4.78 T − 4.50 Po 2.47 P − 2.36 2M2− 2.26 TMB 1.69 T
PPM 9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0
Chemical Shift (ppm)
Figure3.1HNMRspectrumofasurfacefilmformedon0.30Mpropanalin48wt%H2SO4(7daysaftermixing).Thefilmwasdissolved
inCDCl3.TheATR-FTIRspectrumforthissamefilmisshowninFig.2.Allmajorpeakshavebeenassignedtothefollowingfivedominant
species:P=propanal,2M2P=2-methyl-2-pentenal,TMB=1,3,5-trimethylbenzene,Po=polymer,T=2,4,6-triethyl-1,3,5-trioxane.
potential products, the cyclotrimer is most easily confirmed and Pinchas, 1952). Furthermore, both the hemiacetal and
since it is readily identified by comparison of the FTIR and acetal are also unlikely to be major ether constituents since
NMRspectraofthefilmtothespectraof2,4,6-triethyl-1,3,5- onlyaweakpeakexistsintheOHstretchingregion(3500–
trioxane as indicated by the peaks assigned to the trioxane 3400cm−1)whereastrongerpeak(withrespecttothepeaks
(T) in Figs. 2 and 3. Specifically, the 1H NMR spectrum intheetherregion)wouldbeexpectedduetotheOHgroups.
of the film contains all three of the peaks in the reference Instead, the strong peak at 945cm−1 most likely re-
spectrum (SDBS): a triplet at 4.78ppm, a complex multi- sults from longer-chain polymers of propanal (polyacetal
pletat1.67ppmandatripletat0.94ppm(althoughthebroad in Fig. 1c) due to C-O-C-O-C stretching bands that are
peak group at 0.94ppm can only be partially due to the tri- shifted to lower frequencies with the addition of additional
oxane due to its strong intensity relative to the other triox- ethergroups.Spectraofpolymersofvarioussmallaldehydes
anepeaks).Similarly,asshownbyassignmentsinFig.2,at (formaldehyde, acetaldehyde and propanal) which contain
least 13 peaks in the FTIR spectrum of the film correspond thesamepolymethoxy(-C-O-) backbonedisplayonlyvery
n
topeaksinthespectrumofneat2,4,6-triethyl-1,3,5-trioxane weak OH stretching bands but multiple very strong, broad,
(including all six of the strongest peaks between 1500 and overlapping absorption bands between 925 and 975cm−1
900cm−1). Furthermore, previous studies of 2,4,6-triethyl- (Novak and Whalley, 1959a, b, 1962; Vogl, 1964a, b). Al-
1,3,5-trioxane report that it phase separates upon formation thoughthepeakat945cm−1 doesnotexactlymatchanyof
from propanal/catalyst solutions (Sato et al., 1993, 1991), the three strongest peaks (975, 960 and 925cm−1) in this
consistentwithoursurfacefilmformation. region in the Novak and Whalley (1959a) spectrum of the
Uponassignmentofthecyclotrimerpeaks,onlyonemajor polymethoxypolymerformedbypressurizationofpropanal,
peak in the FTIR spectrum of the film remains unexplained thereissuchbroadabsorptionintheentire980–920cm−1re-
by species identified thus far (2,4,6-triethyl-1,3,5-trioxane, gionofthepolymerspectrumthatapeaknear945cm−1may
2-methyl-2-pentenal, 1,3,5-trimethylbenzene and propanal). not be distinguishable. Furthermore, the polymer present in
Thispeakat945cm−1 is,however,thestrongestpeakinthe oursurfacefilmislikelytodisplaydifferentrelativeintensi-
spectrum of the film and therefore must be a major peak in ties of the peaks in this region due to differences in degree
the spectrum of the absorbing species. The hemiacetal and ofpolymerizationand/ordifferencesinrelativequantitiesof
singleacetalformedbypropanal(Fig.1c)areunlikelytobe rotational isomers (Novak and Whalley, 1959a). Addition-
responsibleforthepeakat945cm−1sincetheywouldbeex- ally,bandsmayalsobeshiftedinfrequencyduetodifferent
pectedtoproducetheirstrongestbandsathigherfrequencies. interactions between polymer chains (Novak and Whalley,
Specifically,thehemiacetalwouldproduceastrongFTIRab- 1959a)inthecomplexsurfacefilmmatrix.Finally,theNMR
sorptionbandinthe1150–1085cm−1regionfromtheasym- spectrum of the film is also consistent with the presence of
metricstretchofitssingleethergroup,whiletheacetalcon- propanalpolymersincethe4.5–5.1ppmregioncontainsmul-
tainstheC-O-C-O-Cmoietywhichwouldproducefivechar- tiple unassigned peaks consistent with ethers and similar to
acteristic bands between 1200 and 1020cm−1 (Bergmann the broad group of unresolved peaks from 4.5 to 5.0ppm
www.atmos-chem-phys.net/15/4225/2015/ Atmos.Chem.Phys.,15,4225–4239,2015
4232 A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO
2 4
that characterizes the NMR spectrum of the polymethoxy 1.0
A c id u s e d :
polymer formed by acetaldehyde (Vogl, 1964a), while CH
2 H C l
andCH3protonsfromtheethylchainsarelikelyresponsible H S O
0.8 2 4
forpeaksinthe1.0–1.7ppmregionandforaportionofthe
tripletat0.94ppm,respectively.
Afterthisidentificationofpolymersofpropanal,wenote e*0.6
c
thatallofthemajorbandsintheinfraredspectrumthatcould an
rb
notbeexplainedbyaldolcondensationproductseithercorre- bso0.4
spondto2,4,6-triethyl-1,3,5-trioxaneorcouldreasonablybe A
assignedtolonger-chainlinearpropanalpolymers.
0.2
3.2.3 Otherpotentialfilmcomponents:organosulfates
0.0
andminorspecies 3600 3400 3200 3000 2800 2600 1800 1600 1400 1200 1000 800
W a v e n u m b e r (c m -1)
In order to test for the presence of organosulfates, reac- Figure4.ATR-FTIRspectraoffilmsformedon0.30Mpropanalat
tion mixtures were prepared with hydrochloric acid at the pH=−0.85inH2SO4(48wt%)andinHCl(7daysaftermixing).
samepHasthesulfuricacidmixtures.Theformationofsur- The region from 2500 to 1800cm−1 is omitted for clarity. *Ab-
face films on these mixtures demonstrates that organosul- sorbancespectraarescaledtotheC=Opeakat1690cm−1 from
fates are not necessary for film formation, and the simi- aldolcondensationproducts(predominantly2-methyl-2-pentenal).
larity of the ATR-FTIR spectra for films formed on sulfu-
ric acid and hydrochloric acid solutions (example shown in
Fig. 4 for 0.30M propanal/48wt% H SO ) demonstrates
2 4 ferences in relative peak areas between chemical species
that organosulfates are not present in significant quantities.
whencomparedtothebasecase.
Wenote,however,thatorganosulfatescouldstillbeproduced
There were two detectable trends in NMR peak area
inthesulfuricacidsolutions,wheretheywouldbeexpected
ratios with film age. First, the trioxane (4.78ppm) peak
toremainduetotheirionizability.
area decreased with age relative to all other species that
Although all of the major peaks (and many of the minor
produced peaks separated well enough for integration (2-
peaks) in both the FTIR and NMR spectra of the film can
methyl-2-pentenal (9.39ppm), trimethylbenzene (6.79ppm)
beassignedtothechemicalspeciesdiscussedthusfar,some
and propanal (9.79ppm)); furthermore, the oldest samples
smallunassignedpeaks(e.g.,NMRpeaksat3.2and3.9ppm)
(68 and 134 days) lacked any detectable trioxane. Trioxane
indicatethepresenceofotherminorspecies.Thesecouldin-
peaks also decreased relative to the peaks in the ATR-FTIR
clude products of multiple aldol condensation steps, aldols
spectraassignedtolong-chainpolymers.Therefore,sincetri-
that have not lost water through the condensation process
oxane decreases with time relative to all other major film
(see Fig. 1), the hemiacetal and acetal formed by propanal,
species and the films grow thicker with time, it is possible
other acetals that could also potentially be formed by reac-
that trioxane is initially formed rapidly, followed by slower
tions of aldol condensation products with propanal and/or
formation of all other film species. Second, the trimethyl-
productsfromoxidationoffilmsbylight/air.
benzene(6.79ppm)to2-methyl-2-pentenal(9.39ppm)peak
arearatioincreasedwithage(byafactorof2to3goingfrom
3.3 Effectsoflightexposure,temperature,propanal 1–7-day-oldsamplesto68-and134-day-oldsamples).Since
concentrationandfilmage 2-methyl-2-pentenalisaprecursorfortrimethylbenzenefor-
mation,thisresultsuggeststhattrimethylbenzeneformation
The preceding detailed chemical analyses were presented continuesbeyond1week.
for the base case of a film formed on a 7-day-old solu- Although solutions stored at the lowest temperature of
tion of 0.30M propanal in 48wt% sulfuric acid, stored at −19◦Cdidnotproducesufficientquantitiesoffilmforanal-
room temperature under fluorescent room light. Very simi- ysis,theNMRspectrumofa73-day-oldfilmformedat0◦C
lar NMR spectra were obtained from films formed on solu- showed higher relative levels of trioxane and lower relative
tionsthatwerestoredunderdifferentconditions(darkand/or levels of trimethylbenzene than those formed at room tem-
0◦C), that were younger (1 and 4 days) and older (68 and perature.Thisresultisconsistentwithreactionsthatproceed
134 days) or that were formed at lower propanal concen- more slowly at lower temperature, according to the previ-
tration (0.030M); these spectra confirm the presence of the ouslynotedtrendswithage.
samemajorchemicalspecies.Spectraoffilmsformedinthe Finally, solutions with lower propanal concentration
dark are not detectably different than those formed in the (0.030 vs. 0.30M) did not produce a sufficient quantity of
light, but films formed at different ages, at 0◦C or at lower film for reliable removal and spectral analysis without con-
propanalconcentrationdisplaythefollowingsignificantdif- taminationbytheunderlyingsulfuricacidsolution.However,
Atmos.Chem.Phys.,15,4225–4239,2015 www.atmos-chem-phys.net/15/4225/2015/
A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO 4233
2 4
one weak NMR spectrum of a 16-day-old sample was ob- 1.5
H S O c o n c e n tra tio n :
tainedthatallowspositivedetectionoftrimethylbenzeneand 2 4
4 8 w t.%
2-methyl-2-pentenal and indicates likely presence of long-
3 7 w t.%
chainpolymersduetomultipleoverlappingpeakssimilarto
those previously assigned to protons on the polymer ethyl 1.0
chains(1.0–1.7and∼0.94ppm).Trioxanecouldnotbede- e*
c
n
tected above the noise; however, we note that low trioxane a
rb
o
contentcouldbeduetotheolderfilmagesincethetrimethyl- s
b
A
benzene to 2-methyl-2-pentenal ratio is high and therefore 0.5
also consistent with older films formed on 0.30M propanal
solutions.
3.4 Effectofacidity 0.0
3600 3400 3200 3000 2800 2600 1800 1600 1400 1200 1000 800
W a v e n u m b e r (c m -1)
AsdiscussedinSect.3.1,acidityhasacomplexeffectonthe
Figure5.EffectofacidityontheATR-FTIRspectraofsurfacefilms
formationratesofthesurfacefilmsthatvariesdependingon
formedon0.30MpropanalinH2SO4 (7daysaftermixing).Trip-
the organic mixture. In general, films tended to form faster
licatesareshownforboth48and37wt%H2SO4.Theregionfrom
as the acidity increased from 19 to 37 to 48wt% H2SO4, 2500to1800cm−1 isomittedforclarity.*Absorbancespectraare
butfilmsformedmoreslowlyornotatallatthehighestacid- scaled to the C=O peak at 1690cm−1 from aldol condensation
ity(76wt%H2SO4)inallmixturesexceptpropanal/glyoxal. products(predominantly2-methyl-2-pentenal)inordertoillustrate
The FTIR spectra of films formed on mixtures of 0.30M differencesbetweenrelativepeakintensities.
propanalin48and37wt%H SO solutionsinFig.5(shown
2 4
intriplicate)demonstratethattherearealsochemicaldiffer-
propanaland0.03Mglyoxalformedfilmsfasterthan0.03M
ences in films formed at different acidities. The spectra are
scaledtotheC=Opeakat1690cm−1fromaldolcondensa- propanal alone suggesting that products of cross-reactions
betweenglyoxalandpropanalparticipateinfilmformation,
tion products (predominantly 2-methyl-2-pentenal) in order
resultinginfasterfilmformationduetohighertotalconcen-
toillustratedifferencesinrelativepeakintensities.Although
trations of reactants available for film-forming reactions. In
there is considerable variability in relative peak intensities
contrast,mixturesof0.03Mpropanaland0.03Mmethylgly-
among the spectra of replicates (most likely due to film in-
homogeneity), the peaks in the 1200–900cm−1 region are oxal formed films more slowly than 0.03M propanal alone.
generallylargerrelativetotheC=Opeakatthehigheracid- A comparison of FTIR spectra of films formed on various
organicmixturesin48wt%H SO allpreparedonthesame
ity (red), indicating a larger relative contribution to the film 2 4
day are shown in Fig. 6. Because significant variability in
fromthe2,4,6-triethyl-1,3,5-trioxaneandlonger-chainpoly-
relative peak intensities exists in replicate FTIR spectra of
acetal polymers that absorb in this region. In addition, the
thefilmsmostlikelyduetoinhomogeneityinthesolidmix-
films formed at the higher acidity also have smaller peaks
at 1608cm−1, indicating smaller concentrations of 1,3,5- turesofmultiplechemicalspecies,acompletesetofreplicate
spectraareprovidedintheSupplement(Figs.S3andS4)to
trimethylbenzene relative to aldol condensation products.
demonstrate that the differences between organic mixtures
Both of these trends are confirmed by NMR spectroscopy
discussed here are in fact due to differing chemical path-
(datanotshown).Thepresenceofmore2,4,6-triethyl-1,3,5-
ways and are not simply sampling artifacts. The spectra in
trioxane and polymers at higher acidities is consistent with
Fig.6areagainscaledtotheC=Opeakat1690cm−1from
faster film formation at higher acidities in the 19–48wt%
propanal aldol condensation products in order to illustrate
H SO range since these species are most likely responsi-
2 4
differences between relative peak intensities. The spectra
bleforthephaseseparationintoasurfacefilm.Additionally,
of the films from propanal and propanal/glyoxal are nearly
slowerfilmformationatthehighestacidity(76wt%)ispo-
identicalinthe1800–1600cm−1 region,indicatingthatboth
tentiallyduetolowwatercontentthatreducestheformation
filmsinclude2-methyl-2-pentenalandunreactedpropanalin
ofthediolsrequiredtobeginthepolymerizationprocess(see
similar ratios. Conversely, the spectral pattern in the 1200–
Fig.1c).
900cm−1 regionforthepropanal/glyoxalfilmdoesnotcor-
respondtothespectrumof2,4,6-triethyl-1,3,5-trioxaneasit
3.5 Cross-reactionswithglyoxalandmethylglyoxal
does for the propanal-only film. Since glyoxal did not form
films by itself, this infrared signature is most likely due to
Toexaminethepotentialeffectofadditionalorganicspecies
productsofcross-reactionsbetweenpropanalandglyoxal.
withcarbonylgroupsontheformationoffilms,mixturesof
propanal with glyoxal and/or methylglyoxal were also ex-
amined. Although glyoxal and methylglyoxal did not form
films in the absence of propanal, the mixtures of 0.03M
www.atmos-chem-phys.net/15/4225/2015/ Atmos.Chem.Phys.,15,4225–4239,2015
4234 A.L.VanWyngardenetal.:Compositionofcoloredsurfacefilmsformedonpropanal/H SO
2 4
der to give some insight into the potential formation of
P ro p a n a l
highly absorbing species over the long residence time of
P ro p a n a l, G ly o x a l
3 P ro p a n a l, M e th y lg ly o x a l lower stratospheric aerosols (∼2 years). Figure 7a (green)
P ro p a n a l, G ly o x a l, M e th y lg ly o x a l showsthatasolutionof0.030Mpropanalin48wt%sulfuric
acid allowed to age for 274 days had two strong absorption
* peaksaround200and245nm,mostlikelycorrespondingto
ce2
an species also observed in the films: 1,3,5-trimethylbenzene
orb and2-methyl-2-pentenal,whichabsorbinwaterat∼200and
s
b
A 234nm, respectively. More importantly, the absorbance ex-
1 tends significantly into the visible. There are no other dis-
tinguishable peaks, but the absorbance is most likely due
tooverlappingpeaksfromvariouslongeroligomersformed
by additional aldol condensation reactions of 2-methyl-2-
0
3600 3400 3200 3000 2800 2600 1800 1600 1400 1200 1000 800 pentenal with propanal and/or with itself. Each sequential
W a v e n u m b e r (c m -1)
aldol condensation step would add another unit(s) of con-
Figure6.ATR-FTIRspectraofsurfacefilmsformedonmixtures jugation and thereby shift the absorption peak to longer
ofpropanalwithglyoxaland/ormethylglyoxalin48wt%H2SO4 wavelengths. This interpretation is supported by the obser-
(7daysaftermixing).Solutionsare0.30Mineachorganic.There- vationthatwhentheaciditywasincreasedto76wt%sulfu-
gionfrom2500to1800cm−1 isomittedforclarity.*Absorbance
ricacid(Fig.7b),theintensityofthepeakcorrespondingto
spectraarescaledtotheC=Opeakat1690cm−1fromaldolcon-
2-methyl-2-pentenalwasreduced(orevenabsent)andaddi-
densationproducts(predominantly2-methyl-2-pentenal)inorderto
tional peaks became distinguishable at longer wavelengths
illustrate differences between relative peak intensities. Spectra of
(270, 365, 388 (shoulder) and 458nm). Nozière and Es-
replicatesareprovidedintheSupplement(Figs.S3andS4).
teve(2007)observedasimilarspectrumforreactionproducts
of propanal in 96wt% sulfuric acid and also ascribe these
TheFTIRspectrumofthefilmfrompropanalandmethyl- longwavelengthpeakstooligomersfromaldolcondensation
glyoxal deviates even farther from that of only propanal. reactions.Althoughtheysuggestthatthepeakintheirspec-
Not only does it lack the signature of 2,4,6-triethyl-1,3,5- trumnear270nmmaybepropanalitself,thiscannotbethe
trioxane, indicating products of cross-reactions, but, addi- caseforoursamplessincethemolarabsorptivityofpropanal
tionally, absorbance in the entire 1500–900cm−1 region is istoosmallat∼9cm−1M−1(Xuetal.,1993).
much stronger relative to the aldol condensation peak at The absorption spectra of mixtures of propanal with gly-
1690cm−1, indicating a stronger relative contribution from oxaland/ormethylglyoxalarealsopresentedinFig.7.“Ef-
acetalspecies.Finally,itisintriguingthat(1)thespectrumof fective”molarabsorptivitiesarecalculatedbasedonlyonthe
thefilmformedonthepropanal/glyoxal/methylglyoxalmix- concentrationofthepropanalreactant(0.030M)sothatany
ture is quite similar to that for the propanal/methylglyoxal changesinabsorbance(comparedtothepropanal-onlyspec-
mixture, differing only in relative peak ratios (a result that trum) must be due to the presence of the additional organic
remained true even when the methylglyoxal concentration species.Atbothacidities,absorbanceinmostofthespectrum
was reduced by up to a factor of 10 but not by a factor of is increased, with methylglyoxal having a larger effect than
100); and (2) that the rate of film formation was decreased glyoxal,suggestingthattheaddedorganicspeciesareunder-
fromtherateforthepropanal/glyoxalmixture.Thiscouldin- going aldol condensation either via reactions with propanal
dicate that glyoxal is somehow inhibited from participating orself-reactions.Althoughsomeoftheadditionalabsorption
infilm-formingreactionsbythepresenceofmethylglyoxal. maybeduetoglyoxalandmethylglyoxalthemselves,molar
The mechanism for such inhibition is unclear, but plausible absorptivitiesofthesespeciesaretoosmall(Horowitzetal.,
explanationsincludecross-reactionsofglyoxalwithmethyl- 2001;MalikandJoens,2000;Plumetal.,1983)tocontribute
glyoxalthatarefasterthanthosewithpropanalbutthatdonot significantlyatleastbelow350nm.
result in products that partition to the film and/or dimeriza-
tionreactionsofpropanalwithmethylglyoxalthatarefaster
than those with glyoxal but that subsequently require more 4 Discussionandatmosphericimplications
timetoformpolymersthatarelargeenoughtopartitioninto
thefilm. The major species present in surface films formed on bulk
solutions of propanal in sulfuric acid were identified as
3.6 UV–visiblespectraofsolutions aldol condensation products (mainly 2-methyl-2-pentenal
and 1,3,5-trimethylbenzene) and polyacetals (mainly 2,4,6-
Although the focus of this work is on characterization of triethyl-1,3,5-trioxane and longer-chain linear polyacetals).
thesurfacefilms,theUV–visibleabsorptionspectraofaged Oftheseproducts,thepolyacetalspecies(bothcyclicandlin-
organic/sulfuric acid solutions were also examined in or- ear) are most likely to be primarily responsible for the sep-
Atmos.Chem.Phys.,15,4225–4239,2015 www.atmos-chem-phys.net/15/4225/2015/