UBC Faculty Research and Publications

Studies of one and two component aerosols using IR/VUV single particle mass spectrometry: Insights into.. Hanna, Sarah J.; Campuzano-Jost, Pedro; Simpson, Emily A.; Hepburn, John W.; Kanan, Khalid M. M.; Bertram, Allan K.; Blades, Michael W. 2010

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
[if-you-see-this-DO-NOT-CLICK]
Bertram_2010_CPPC_c0cp00462f.pdf [ 1.08MB ]
[if-you-see-this-DO-NOT-CLICK]
Metadata
JSON: 1.0041860.json
JSON-LD: 1.0041860+ld.json
RDF/XML (Pretty): 1.0041860.xml
RDF/JSON: 1.0041860+rdf.json
Turtle: 1.0041860+rdf-turtle.txt
N-Triples: 1.0041860+rdf-ntriples.txt
Original Record: 1.0041860 +original-record.json
Full Text
1.0041860.txt
Citation
1.0041860.ris

Full Text

Studies of one and two component aerosols using IR/VUV single particlemass spectrometry: Insights into the vaporization process andquantitative limitationsEmily A. Simpson,aPedro Campuzano-Jost,*aSarah J. Hanna,aKhalid M. M. Kanan,bJohn W. Hepburn,aMichael W. BladesaandAllan K. Bertram*aReceived 5th May 2010, Accepted 7th July 2010DOI: 10.1039/c0cp00462fThis paper presents the studies of one and two component particles using a CO2laser forvaporization and VUV ionization in an ion trap mass spectrometer. The degree of fragmentationfor a one component system was demonstrated to be a function of CO2laser energy. In a twocomponent system, the degree of fragmentation was shown to be a function of the particlecomposition. This observation indicates that the analysis of mixed particles may be far morecomplicated than anticipated for a two step process with soft vaporization. In addition toshowing that fragmentation is a function of CO2laser energy and particle composition, we alsoshow that a key parameter that determines the extent of fragmentation is the energy absorbedby the particle during desorption. The ionization delay profile in a one component system is alsoshown to be strongly dependent on the vaporization energy. In a two component system, thedelay profile is shown to strongly depend on the composition of the particle. The combined datasuggest that the key parameter that governs the delay profile is the energy absorbed by theparticle during desorption. This finding has implications for potential field measurements.Finally, for a two component system where the absorption crosssections are different, the changein the degree of fragmentation with particle composition resulted in a non-linear dependence ofion signal on composition. This makes any attempt at quantification difficult.IntroductionAerosols are an important topic of research given theirubiquitous presence in the atmosphere and their significantrole in climate, chemistry of the atmosphere, and humanhealth.1–4Atmospheric aerosols contain numerous species,including sulfates,nitrates, andorganics. Oneof thechallengesin real-time organic aerosol analysis is the fragility of themolecules under study. One step laser desorption/ionizationrequires high laser powers which can cause extensive fragmentationof organic species through successive absorption of severalphotons and charge-transfer matrix effects.5A solution to thisis to separate the vaporization and ionization steps6and use a‘‘soft’’ ionization source to reduce fragmentation of manyorganics.7–9Recently we have developed a single particle massspectrometer that incorporates a CO2laser for vaporization, atunable laser-based VUV source for ionization, and an iontrap mass spectrometer for mass analysis. The instrumentwas characterized with single component particles byHanna et al.8,10By using a ‘‘soft’’ ionization source, ions are generated fromthe vaporized aerosol neutrals with a minimum of excessenergy. This reduces the extent of fragmentation andthus simplifies the chemical characterization of the aerosol.To date, several research groups have implemented softionization sources for aerosol mass spectrometry. Someexamples include metal attachment,7chemical ionization,11–14attachment of low-energy photoelectrons (PERCI),15–17resonance enhanced multiphoton ionization (REMPI),6,18–24and single photon ionization.9,25–31In the following, we carry out detailed studies of pure oleicacid, pure oleyl alcohol (see Fig. 1 for chemical structures),and mixed oleic acid : oleyl alcohol particles using our singleparticle ion trap mass spectrometer. The three specificquestions we are trying to address are detailed below.Question 1: how does the fragmentation vary with CO2energy and particle composition for one and two componentsystems?Thisiscritical forinterpretingmassspectraforsimilarinstruments and also for determining if similar instruments willyield a linear responsewith particle composition. Forexample,if the fragmentation pattern of one species in a mixed particlechanges as a function of the particle composition, evaluatingtheionsignalfromasingularmolecular orfragmentionwouldnot give a linear response for that species.Question 2: how does the ionization delay profile vary withCO2energy and composition for one and two componentsystems? The ionization delay profile is a scan of total ionsignal as a function of the delay time between the firing of thedesorption laser and ionization laser. Understanding theionization delay profiles is important for achieving optimalinstrument performance, particularly in a field instrumentwhere a single delay time will be used.aDepartment of Chemistry, University of British Columbia,Vancouver, BC V6T 1Z1, Canada.E-mail: pcampuzano@chem.ubc.ca, bertram@chem.ubc.cabDepartment of Chemistry, Al-Quds University, JerusalemThis journal is C13c the Owner Societies 2010 Phys.Chem.Chem.Phys.,2010, 12, 11565–11575 | 11565PAPER www.rsc.org/pccp | Physical Chemistry Chemical PhysicsDownloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView OnlineQuestion 3: for a two component system in which theabsorption crosssections are different, does the mass spectrumshow a linear response with composition? The results from thisstudy will help determine the usefulness of similar instrumentsfor future quantitative and kinetic studies. Our results shouldapply to similar instruments that use a two step vaporization/ionization process with a CO2laser for desorption. This pointis discussed in more detail in the Summary and conclusionssection. Furthermore, this study will help identify some of thepotential weaknesses of a two laser system for quantitativeanalysis of multi-component particles.Oleic acid is present in the atmosphere at concentrationsofB1ngmC03in the particle phase and is known to originatefrom a variety of sources, including meat cooking operations.32Oleyl alcohol is structurally comparable to oleic acid with thecarboxylicacidgroupbeingreplacedbyanalcohol(seeFig.1).The two compounds are liquids and miscible in each other.Additionally, astheabsorptioncrosssectionsofthe 2speciesat1056 cmC01(the CO2wavelength utilized) differed by a factorof B4 (based on the IR spectra measured; see Table 1), theinstrument’s ability to carry out quantitative measurements ofmixtures of this kind is investigated.There have not been many studies related to the questionsoutlined above. Studies of the effects of CO2power onfragmentation have been previously conducted for singleparticles of oleic acid,31,33ethylene glycol,28,31and aniline5with CO2/VUV in a TOF-MS. Hanna et al. investigated thesame effect for oleic acid, 2,4-dihydroxybenzoic acid, andcaffeine in an ion trap using a CO2/VUV combination.8,10There have been no previous studies of the change infragmentation as a function of composition for mixed particles.Studies of the ionization delay profile as a function of CO2energy and/or composition have been done for single componentparticles of ethylene glycol,28aniline,5oleic acid,8and 2,4-dihydroxybenzoic acid.8In addition, Woods et al.30discussedionization delay profiles of mixed glycerol–oleic acid particles(immiscible mixtures) using a CO2/VUV system. To ourknowledge, no one has explored the effect of changing particlecomposition on ionization delay profiles (or noted such aneffect) in a miscible mixed particle system.Lastly, there has only been one study investigating the linearresponse with particle composition for mixed particles usingCO2/VUV. Baer and coworkers explored the quantitative useof CO2/VUV for four varying particle compositions of a threecomponent polycyclic aromatic hydrocarbon (PAH) mixture.29Thesearomaticspeciesarereasonably stable under fragmentationand primarily produce molecular ions which makes the overallanalysis more straightforward. Conversely oleic acid and oleylalcohol are not as robust as PAHs and tend to undergo morefragmentation,27,33particularly in the ion trap system wherestorage time is longer8(see below for further discussion).The two component system was studied at two CO2desorption energies of 10 mJ per pulse and 7 mJ per pulse(1056 cmC01). At 10 mJ per pulse, both species fragmentedextensively and it was impossible to separate the mass spectralpeaks of the individual components; hence this paper focuseson results from the 7 mJ per pulse case.Materials and methodInstrument descriptionThe instrument has been described in detail in previouspublications8,10,34and therefore only a brief description isgiven here. The main components of the system are an aerosolinlet, a sizing region, and a particle analysis region whereaerosols are vaporized by a pulsed CO2laser, the gas phasemolecules ionized by pulsed vacuum ultraviolet (VUV), andthe ions mass analyzed by an ion trap mass spectrometer(Fig. 2).The mid-IR laser for desorption is a single mode, tunablepulsed CO2laser (9.2–10.8mm) with a maximum output of 50 mJon the strongest lines (MTL-3G, Edinburgh Instruments Ltd).In these studies, IR pulses with energies between 4–30 mJ wereused to vaporize the particles with a laser spot size of B1mmdiameter inside the trap. For all the studies discussed in thispaper a wavelength of 1056 cmC01was used.Particles are optically thin with respect to the IR energy inthese experiments and are therefore expected to be uniformlyheated by the CO2laser pulse. A particle can be defined asoptically thin if the product of the radius, r, the absorptioncrosssection, s, and the concentration, C,is{1.35The IRabsorption crosssections of oleic acid and oleyl alcohol(solutions prepared in dichloromethane) were measured usinga Bruker Equinox 55 FTIR. At 1056 cmC01, the absorptioncrosssections were measured to be 2.85 C2 10C020cm2molecC01Fig. 1 Chemical structures of oleic acid (left) and oleyl alcohol (right).Table 1 Relevant properties of oleic acid and oleyl alcoholCompoundMW/gmolC01Density/gcmC03sCO2/cm2molecC01IE/eVOleic acid 282.4614 0.895 2.85 C2 10C0208.68278.65 C6 0.058Oleylalcohol268.4778 0.850 1.03 C2 10C019o9.14a8.56 C6 0.05baBased on energies given for 2-propen-1-ol and 2-butene-1-ol from NISTChemistry Webbook and trends shown in Adam and Zimmermann37forionization energy as a function of increasing hydrocarbon chain length.bMeasured in a separate study in our group.11566 | Phys. Chem. Chem. Phys., 2010, 12, 11565–11575 This journal is C13c the Owner Societies 2010Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView Onlineand 1.03 C2 10C019cm2molecC01, respectively, for oleic acid andoleyl alcohol. Based on these numbers, rsC { 1 for theexperimental conditions. Additionally, the particle size is smallin comparison to the CO2wavelength making it unlikely thatinternal focusing of IR light will give rise to a temperaturegradient in the particle.36The tunable vacuum UV light is produced by resonanceenhanced four wave difference mixing in xenon gas and isdescribed in detail elsewhere.8,10The source is continuouslytunable from 10.2 eV (122 nm) to 7.4 eV (168 nm) andproduces between 1010and 1013photons per pulse dependingon the wavelength. The generated VUV is separated from thepump wavelengths by a custom monochromator and focusedto a slightly vertically elongated spot with an area ofB1mm2.For all the current studies an ionization energy of 8.75 eV wasused. For comparison, the measured ionization energy of oleicacid is 8.68 eV.27See Table 1 for some select properties of oleicacid and oleyl alcohol.The ion trap is operated in mass selective instability modewith an accessible mass range of 10 to 340 Da. Mass scanningrates of 4000 Da sC01were used for all the experiments whichgivesamassresolution ofB500m/Dmatm/z=264.Themassaxis is calibrated daily by recording 70 eV EI spectra ofperfluorotributylamine.It should be noted that the paths of the vaporization andionization lasers and the path of the particle beam do notintersect in the center of the trap. Instead the particles firstpass through the IR and then the VUV beam as they traverseFig. 2 Instrument schematic of the single particle ion trap mass spectrometer (SPITMS).This journal is C13c the Owner Societies 2010 Phys.Chem.Chem.Phys.,2010, 12, 11565–11575 | 11567Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView Onlinethe ion trap (inset of Fig. 2). The distance that the aerosolstravel between the two intersection points is on the order of2 mm. As a result of this distance, there is no ion signal if theionization pulse is fired immediately after the CO2laser(0 ms delay time). However, if the delay between the two laserpulses is varied, a profile of the expanding plume from thevaporized aerosol can be obtained.Experiments consisted of collecting mass spectra as afunction of the delay time between the two laser pulses.Typically we scanned from 0–30 ms delay time. From thesedata we extracted the ionization delay profiles (scan of totalion signal as a function of time between desorption andionization). See Fig. 5 as an example. The mass spectrapresented in this paper were obtained by integrating over theentire delay profile. In other words, we summed all the massspectra at each different delay time to obtain a representativemass spectrum for the entire vaporization process.For the exploration of effects on fragmentation ratios, theintegrated values over the entire delay profile were also used,again to give a fragmentation ratio consistent with the entirevaporization process.ChemicalsOleic acid (Fluka, Z99%) and oleyl alcohol (Aldrich, 99%)were used as purchased without further purification. For pureparticles, solutionswere prepared in 2-propanol (Aldrich, 99.9%)in the concentration range of 10C05gmLC01and particleswere generated using a vibrating orifice aerosol generator(TSI Model 3450).Mixed particles of oleic acid and oleyl alcohol wereprepared by aerosolizing solutions with mass ratios of90 : 10, 75 : 25, 50 : 50, and 25 : 75 oleic acid : oleyl alcoholin isopropanol with the total mass concentration of thesolution remaining unchanged (2 C2 10C05gmLC01). Since themolar masses of oleic acid and oleyl alcohol are very similar,the mole ratios vary from the mass ratios by at most 1%.All particles were passed through a85Kr charge neutralizerwith B25 LPM dilution air flow before entering the singleparticle ion trap mass spectrometer (SPITMS). This acted as adrying tube and prevented additional drying and size changesin the aerosol lens.ResultsPure oleic acid particlesParticle mass spectra were acquired for pure one componentparticles of oleic acid using a range of CO2energies forvaporization. Fig. 3 shows an example spectrum recorded at7 mJ per pulse CO2energy. No ions from the CO2laser alonewere observed over the entire range of energies. The massspectra showed a shift towards lower m/z fragments withincreasing vaporization energy as expected.31Several peaks from the oleic acid mass spectrum were usedto assess the degree of fragmentation as a function of CO2energy. The peaks chosen for this analysis were selected tocover a range of fragment ions from high m/z to low m/zacross the spectrum. These peaks are labeled in Fig. 3. Eachfragment’s relative intensity is determined with respect to thetotal ion signal. As can be seen in Fig. 4, the general trend is adecrease in high mass peaks and an increase in lower massfragments with increasing desorption energy. This changingdegree in fragmentation reflects the increasing internal energyof the vaporized molecules.28Fig. 5 shows the ionization delay profiles obtained for oleicacid over a range of desorption energies. The exact shape ofthe profile is dependent on the distance between the two laserbeams (which is constant in these experiments) and thetranslational energy of the vaporized molecules which expandoutward from the particle, filling, and then passing beyond theionization volume. As can be seen, the expansion speed of thedesorbed aerosol plume increases with increasing CO2laserenergy, with the maximum total ion signal shifting to shorterdelay times. These profiles were normalized for ease ofcomparing the optimal delay time where maximum ion signalFig. 3 Oleic acid normalized mass spectra (averaged over the entiredelay profile) acquired at 7 mJ 1056 cmC01CO2, 142 nm VUV.Fig. 4 Relative intensities of fragment peaks compared to the totalion signal as a function of desorption energy for pure oleic acidparticles.11568 | Phys. Chem. Chem. Phys., 2010, 12, 11565–11575 This journal is C13c the Owner Societies 2010Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView Onlineoccurs. It should be noted that total ion intensities do changewith increasing desorption power, reaching a plateau in totalion signal at approximately 15 mJ per pulse. Also, note thatthe ionization delay profile is only a qualitative indicator oftranslational energy, since energy distributions from thevaporization process can be complex and are not wellunderstood.Pure oleyl alcohol particlesParticle mass spectra were also acquired for pure one componentparticles of oleyl alcohol using a range of CO2energies. Fig. 6shows an example spectrum recorded at 7 mJ per pulse CO2energy. Increasing the desorption energy resulted in increasinglevels of fragmentation in concurrence with previousstudies.5,8,10,28,31,33No molecular ion was observed at any ofthe desorption energies used and the fragmentation was quiteextensive even at the lowest vaporization energy used.The peaks chosen for this analysis were selected to cover arange of fragment ions from high m/z to low m/z across thespectrum. The intensities of selected fragment peaks in themass spectrum (see Fig. 6) were again compared to the totalion signal as a function of desorption energy to give anindication of the degree of fragmentation for oleyl alcohol.Fig. 7 shows the same trend observed for oleic acid: a decreasein high mass fragments and increase in low mass fragments asdesorption energy increases.Fig. 8 shows the normalized ionization delay profilesobtained for oleyl alcohol at a range of desorption energies.These profiles were normalized for ease of comparing theoptimal delay time where maximum ion signal occurs. Itshould be noted that total ion intensities do change withincreasing desorption power, reaching a plateau in total ionsignal at approximately 10 mJ per pulse for oleyl alcohol.An increase in the expansion speed of the desorbed aerosolplume can be observed with increasing CO2laser energysimilar to what was observed for pure oleic acid particlesand in other studies. However, it is observed that at acomparable desorption energy, for instance 7 mJ, there is asignificant difference in the delay profiles observed for oleicacid and oleyl alcohol. For pure oleyl alcohol, the maximumion signal occurs at 6 ms compared to 14 ms for pure oleic acid.The delay profile reaches a maximum ion signal at a shorterdelay time and decays faster for oleyl alcohol, suggesting amuch faster translational energy of desorbed molecules ascompared to oleic acid. The differences in delay profiles arelikely due to the differences in IR absorption crosssections ofoleyl alcohol and oleic acid for an equivalent CO2laser power.As mentioned previously, the absorption crosssection of oleylalcohol was approximately a factor of 4 higher in magnitudethan that of oleic acid based on the IR spectra of the purecomponents. This concept is explored in more detail in theDiscussion section.Fig. 5 Total ion signal as a function of ionization delay time for oleicacid. Each trace is for a different CO2pulse energy (at 1056 cmC01)asindicated in the legend (VUV at 142 nm). All traces have beennormalized to make comparison easier.Fig. 6 Oleyl alcohol normalized mass spectra (averaged over theentire delay profile) acquired at 7 mJ 1056 cmC01CO2, 142 nm VUV.Fig. 7 Relative intensities of fragment peaks compared to the totalion signal as a function of desorption energy for pure oleyl alcoholparticles.This journal is C13c the Owner Societies 2010 Phys.Chem.Chem.Phys.,2010, 12, 11565–11575 | 11569Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView OnlineTwo component particles of oleic acid and oleyl alcoholAll results discussed for the two component particles wereobtained with 7 mJ per pulse desorption energy, where theoleic acid mass spectrum has distinctive high mass peaks(m/z = 264, 246, 148, 134, 127) that do not overlap with oleylalcohol fragments. Due to the fragmentation of metastableions observed in the ion trap previously,8we don’t seesignificant molecular ion peaks for the analytes chosen in thisstudy (Fig. 3 and 6). This adds to the complexity of analysis.Normalized mass spectra are shown as a function of particlecomposition in Fig. 9. The pure particle spectra for oleic acidand oleyl alcohol are distinctly different at 7 mJ 1056 cmC01CO2.Oleic acid experiences far less fragmentation and producesfragments primarily above m/z 125, while oleyl alcohol under-goes much more fragmentation and gives fragments primarilybelow m/z 125. The mass spectra of the mixed particles showsome interesting features. The 90 : 10 wt% spectrumappears to most closely resemble a combination of the peakintensities found in each of the 7 mJ pure component spectra.By 75 : 25 wt%, the dominant peaks for oleic acid (at m/z= 264for example) are extremely diminished, more so than would beexpected based on composition alone. This could be due to ashift in the fragmentation pattern of oleic acid, perhaps due toexperiencing more heating in the vaporization step or itcould be due to a preferential vaporization of oleyl alcohol.However, as the two components are both liquids andmiscible, it is not expected that oleyl alcohol partitions tothe surface and undergoes preferential desorption at low CO2energies; therefore a change in the oleic acid fragmentationpattern seems most likely.Thefragmentation of the mixed particleswas thenexaminedas a function of composition. Since oleyl alcohol is moreheavily fragmented at 7 mJ, selected high mass peaks in oleicacid were used to monitor the change in fragmentation. Peaksat m/z 127, 134, 235, 246, and 264 were ratioed to m/z 148 todetermine the degree of fragmentation of oleic acid. All ofthese peaks are assigned exclusively to oleic acid. Oleyl alcoholspectra were examined in detail at a range of different CO2energies to ensure that none of the peaks selected to evaluateoleic acid were observed. Therefore we can use these peakswith confidence to look at the change in fragmentation foroleic acid by comparing to the peak at m/z 148, which alsobelongs exclusively to oleic acid and is present at a range ofenergies in the pure oleic acid spectra. All other peaks that weconsidered using for comparison had extensive overlap witholeyl alcohol fragments. For example, the oleic acid fragmentat 56 m/z overlaps with an oleyl alcohol peak within the clusterof fragments around 57 m/z that can be seen in Fig. 6.Fig. 10 shows the degree of fragmentation as a function ofcomposition for oleic acid in mixed particles. The fragmentationpattern for oleic acid clearly shows that as the percentagecomposition of oleic acid decreases (oleyl alcohol increases),oleic acid fragments more extensively.The ionization delay profiles (total ion signal versus ionizationdelay time) also reflect the effect of composition on vaporizationof the particle. In Fig. 11, it can be seen that the ionizationdelay profiles change noticeably with composition until acertain percentage of oleyl alcohol is reached in the particle(B25% or greater). There is also a significant change in totalion signal with an increasing amount of particle materialdesorbed with increasing oleyl alcohol composition which isnot reflected by the normalized delay profiles.Fig. 8 Total ion signal as a function of ionization delay time for oleylalcohol at an ionization wavelength of 142 nm. Each trace is for adifferent CO2pulse energy (at 1056 cmC01) as indicated in the legend.All traces have been normalized to make comparison easier.Fig. 9 Normalized mass spectra (averaged over the entire delayprofile) as a function of particle composition (7 mJ 1056 cmC01CO2,142 nm VUV).11570 | Phys. Chem. Chem. Phys., 2010, 12, 11565–11575 This journal is C13c the Owner Societies 2010Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView OnlineDiscussionDependence of degree of fragmentation on CO2energy andcompositionThe degree of fragmentation for a species has been shownto be strongly related to the internal energy of themolecule.8,9,28,31,33,35,38,39If the internal energy of the vaporizedmolecules is high, then increased fragmentation isexpected.35,38–40For the pure particles in this study, fragmentationincreased with increasing desorption energy, which isconsistent with results from other groups that show the degreeof fragmentation to be a strong function of particle heating,regardless of whether a heater or laser was used for thevaporization step.8,9,28,31,33,35,38,39Additionally, in this studythe extent of fragmentation was seen to depend on particlecomposition. At an equal desorption energy, pure oleylalcohol showed more fragmentation than pure oleic acid.The partitioning of energy into the different degrees offreedomiscomplexandnotwellunderstoodinlaservaporizationof particles. Here we only carry out a simple qualitativeanalysis to show that the fragmentation behavior observedfor pure particles as well as the two component particles comesfrom the different amounts of absorbed CO2energy. In otherwords, this simple analysis shows that the presence of oleylalcohol in the mixed particle enhances the absorption ofCO2laser radiation and leads to an increase in internal energy(and hence fragmentation) of the vaporized molecules.The respective absorption crosssections determined in theIR measurements were used to calculate the amount of energyabsorbed by the particle (Eabs) for both the pure oleic acidparticles and the mixed particles according to Beer’s law andsome simple approximations as described by eqn (1) and (2).The derivation of eqn (1) (for a one component system) isgiven in the Appendix.Eabs¼ ECO2ApACO21C0exp C0sNp4rp3C18C19C20C21C26C27ð1ÞECO2is the measured CO2energy, Apis the area of the particle,ACO2is the area of the CO2beam where it intersects theparticle, s is the absorption crosssection, Npis the numberdensity of molecules in the particle, and 4rp/3 is an approximationof the path length the light travels through the particle.Similarly for a two component system, the equation for energyabsorbed by the particle during desorption is given below:Eabs¼ ECO2ApACO21C0exp C0ðs1N1þs2N2Þ4rp3C20C21C26C27ð2Þwhere s1and s2are the respective crosssections of eachcomponent and N1and N2are the respective number densitiesof each component in the mixed particle.Fig. 12 shows the ratios of the peak intensities at m/z = 127,134 and 264 relative to the peak at m/z 148 as a function of thecalculated CO2energy absorbed per particle for oleic acid inboth the one component and two component particles. Theclosed data points represent the pure oleic acid particles andthe open data points representthe mixed particles. The ratio ofm/z = 264 to m/z = 148 decreases with increasing energyabsorbed whereas the ratio of peaks at m/z 127 and 134 tom/z 148 increases with increasing energy absorbed. At higherenergies for the pure oleic acid, the error bars are large due tothe low signal to noise ratio which occurs because themass spectra are dominated by very low mass fragments(om/z = 50) at these energies. The data in Fig. 12 demonstratethat the degree of fragmentation for oleic acid in both onecomponent and two component particles is due to the energyabsorbed per particle in the desorption step. As expected,more energy absorbed results in more fragmentation.Dependence of ionization delay profiles on CO2laser energy andparticle compositionThe expansion of the aerosol plume can be followed byvarying the delay between the CO2and VUV pulses. Theshapeoftheionization delayprofileasillustratedbytheprofileFWHM or the delay time at maximum ion signal is aqualitative indicator of the translational energy of theFig. 10 Fragmentation ratio of oleic acid (selected peaks relative tom/z 148) as a function of weight percent oleic acid.Fig. 11 Normalized ion signal as a function of ionization delay timefor mixed oleic acid : oleyl alcohol particles.This journal is C13c the Owner Societies 2010 Phys.Chem.Chem.Phys.,2010, 12, 11565–11575 | 11571Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView Onlinevaporized molecules.28The pure component data suggest thatthe relative translational energy of the desorbed moleculesincreased with increasing CO2energy leading to narrowerdelay profiles and earlier optimal delay times. In the casewhere a higher translational energy is imparted to themolecules, they will enter and leave the ionization region morequickly, whereas at lower energies the molecules will takelonger to fill and then expand beyond the ionization volume.In the mixed particle study, all particles are vaporized at thesame CO2laser energy of 7 mJ, yet the relative translationalenergy of the molecules appears to be changing with compositionas shown in Fig. 10.Similar to the previous section, we perform a simplequalitative analysis to show that the trends in the ionizationdelay profiles for both the one component and two componentaerosols come from the different amounts of absorbed CO2energy. Fig. 13 contains the pure and mixed particle datafor oleic acid plotted as ionization delay profile FWHM(full width at half maximum) versus absorbed energy in thetop panel and optimum delay time versus absorbed energy inthe bottom panel. Closed data points represent the pureparticle and open data points represent the two componentparticle. Both panels demonstrate a fairly consistent trend: theFWHM and delay time both decrease with an increase in CO2energy absorbed. Overall, Fig. 13 illustrates that moreabsorbed CO2energy results in a higher relative translationalenergy, as expected.Non-linear response of ion signal with particle compositionIn a previous study, Woods et al. showed that a two stepprocess using VUV ionization gave quantitative detection ofaromatics (PAHs) in a mixed particle.29In other words, thesignal from a single component in a mixed particle was linearwith the amount (mole fraction) of that component in themixed particle. Shown in Fig. 14 is a plot of three peaks(assigned exclusively to oleic acid) normalized to the total ionsignal as a function of concentration. By normalizing to thetotal ion signal any effect of partial or incomplete vaporizationon the quantitative response should be removed. The spectrawere collected from 30–300 m/z and it is not believed that anysignificant amount of fragments occurs below 30 m/z based onpreviously collected data. Therefore, the relative amount ofoleic acid fragments should give a linear response in relation tocomposition even if the amount of material evaporatingFig. 12 Fragmentationratios foroleicacidas afunctionofIR energyabsorbed during vaporization. Closed symbols show pure oleic acidparticle data, open symbols show mixed particle data.Fig. 13 (Top) Ionization delay profile FWHM for both pure oleicacid particles and mixed particles as a function of CO2energyabsorbed/particle and (bottom) delay time at maximum total ionsignal as a function of CO2energy absorbed/particle.Fig. 14 Relativeintensityofoleicacidfragmentionsignaltototalionsignal for all peaks (oleic and oleyl) as a function of the oleic acid moleratio.11572 | Phys. Chem. Chem. Phys., 2010, 12, 11565–11575 This journal is C13c the Owner Societies 2010Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView Onlinechanges as a function of composition. However, the trendshown in Fig. 14 is clearly non-linear over the range ofcompositions. This trend most likely occurs because therelative amount of signal at m/z 127, 134 and 264 is changingdue to: (1) a change in composition and (2) a change infragmentation with composition. If every peak in the massspectrum from oleic acid could be included in the analysis, alinear trend would be expected. Conversely, this would be verydifficult due to the issue with overlapping peaks betweensimilar aliphatic species. To illustrate the problem with over-lapping peaks, in our experiments only peaks at m/z 282, 264,246, 148, 134, and 127 can beexclusively assigned to oleic acid.All other peaks overlap with oleyl alcohol peaks. For a 50 : 50mixture of oleic acid and oleyl alcohol, these peaks assignedexclusively to oleic acid contribute o5% to the total ionsignal.Fig. 14 shows that a quantitative analysis is difficult if theenergy absorbed per particle during desorption changessignificantly with composition. Conditions where quantitativeanalysis should be successful are (1) where the fragmentationdoes not change with composition or (2) where the IRabsorption crosssections do not change with composition(assuming heat capacities are similar). An example of the firstcase would be analytes that do not fragment or producerelatively few fragments, such as the mixtures of PAHs studiedby Woods et al.29For the second case, species with similarabsorption crosssections and heat capacities are needed or thechange in composition restricted to a range where thedependence is not as strong. These conditions also assumeliquid phaseandmiscibility of speciessuch thattheparticle isahomogeneous mixture.Summary and conclusionsThis paper presents the first results obtained from a twocomponent aerosol study of miscible organic species usingCO2laser vaporization and VUV ionization in an ion trapmass spectrometer. The analysis of mixed particles, evenlaboratory generated two component aerosols, is a complicatedprocess. Below we address the three questions posed in theIntroduction.Question 1: how does the fragmentation vary with CO2laserenergy and particle composition for one and two componentsystems? The degree of fragmentation observed in singleparticlemassspectraofaonecomponentsystemwasdemonstratedto be a function of CO2laser energy. In a two componentsystem, the degree of fragmentation was shown to be afunction of the particle composition. This observationindicates that the analysis of mixed particles may be far morecomplicated than anticipated for a two step process with softvaporization if the degree of fragmentation can change as afunction of composition. The mass spectra of the purecomponent cannot necessarily be used as a fingerprint for thatsame component in a mixture. The implications of this effectare also important for laboratory studies of particle reactionsand subsequent product analysis, depending on the extent ofparticle composition change. In addition to showing thatfragmentation is a function of CO2energy and composition,we also showed that the key parameter that determines theextent of fragmentation, which is a proxy for the internalenergy of the vaporized molecules, is the energy absorbed bythe particle during desorption.Question 2: how does the ionization delay profile vary withCO2energy and composition for one and two componentsystems? The ionization delay profile in a one componentsystem is shown to be strongly dependent on the vaporizationenergy. In a two component system, the delay profile is shownto strongly depend on the composition of the particle. Thecombined data suggest that the key parameter thatgoverns thedelay profile is also the energy absorbed by the particle duringdesorption. This finding has implications for potential fieldmeasurements. In the field, where each particle is not necessarilyidentical, it will be necessary to use a fixed ionization delaytime (fixed time between the vaporization laser and ionizationlaser). This however could be problematic. For instance, if theinstrument parameters are optimized for the maximum totalion signal with one species, these parameters may be far fromideal for a different species and will result in a sensitivity bias.Hanna et al. have shown delay profiles for 2,4-dihydroxy-benzoic acid (DHB) and oleic acid at 15 mJ vaporizationenergy that varied widely; collecting data at the optimal delaytime for oleic acid would result in almost no signal for DHB asits vaporization profile evolved much more quickly.8Both onecomponent and two component delay profiles in this studyindicate this to be a valid concern. This effect may be diminishedat high vaporization energies where there appears to be less ofa composition dependence for pure particles, but this willcome at the cost of increased fragmentation.Question 3: for a two component system where theabsorption crosssections are different, does the mass spectrumshow a linear response with composition? No. The change inthe degree of fragmentation with particle composition resultedin a non-linear dependence of ion signal on composition. Thismakes any attempt at quantification difficult. One possibleway around this problem is to monitor all ion peaks associatedwith a species. But this also seems difficult to achieve in mostcases, since organic species often have significant fragmentationeven with VUV ionization, which can result in overlappingmass spectral peaks. For species that do not fragment aseasily, or have similar absorption crosssections, quantitativeanalysis should be possible. Another possible way around thisproblem is to change the desorption method: impinging theparticles on a heated surface would ensure that the desorptiontemperature is independent of the particle composition.For atmospheric aerosols, the variation in the IR cross-section is not known. Many particles consist of inorganicspecies and a multitude of organic species. Studies on the IRcrosssection of atmospheric aerosol particles (specifically forIR wavelengths relevant for a CO2laser) would be interesting.Our studies have focused on a case where the crosssectionchanged by a factor of 4. Whether or not this represents a‘‘worst-case scenario’’ for field measurements remains to bedetermined.Our system included a two step laser desorption/ionizationsystem followed by an ion trap for mass analysis. Other twostep laser desorption/ionization systems with a CO2laser alluse a TOF mass analyzer, which will result in less fragmentationthan what is observed with an ion trap.8One of the mainThis journal is C13c the Owner Societies 2010 Phys.Chem.Chem.Phys.,2010, 12, 11565–11575 | 11573Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView Onlinedifferences between the TOF mass analyzer and an ion trap isthe storage and measurement time for the ions. For a TOFmass analyzer the measurement/extraction time is on the orderof microseconds. For an ion trap, the storage and measurementtime is on the order of tens of milliseconds. As a result morefragmentation can occur in the ion trap experiments due to thelonger time for unimolecular dissociation. Nevertheless theeffectsobservedaboveshouldstillberelevantforTOFinstrumentsas long assome level offragmentation isobserved, which is thecase for many types of analytes, particularly long chainhydrocarbons.9,25,30,31,33AppendixCalculation of energy absorbed per particle during desorptionBeer’s law, which describes the absorption of light by a sampleis given below, where I is the intensity of light measured afterthe sample, Iois the initial intensity of light, s is the absorptioncrosssection, c is the concentration of the species, and l is thepath length the light travels:C0lnIIoC18C19¼ scl ðA1ÞIn this case we are interested in the CO2energy absorbed bythe particle (Eabs). The energy measured after the particle isequal to the difference of the initial CO2energy (Ei) and theenergy absorbed by the particle (Eabs), thus the equation canbe rewritten:C0lnEiC0EabsEiC18C19¼ scl ðA2ÞSolving for the absorbed CO2energy gives:Eabs=Ei(1 C0 exp(C0scl)) (A3)The initial CO2energy the particle experiences is defined as themeasured CO2energy times the fraction of the CO2beamintercepted by the particle:Ei¼ ECO2ApACO2ðA4Þwhere ECO2is the measured CO2energy, Apis the area of theparticle and ACO2is the area of the CO2beam where itintersects the particle. The fraction of Apover ACO2is includedto take into account the geometric fraction of a flat beamprofile intersecting the particle. We know the CO2profile isGaussian, so we could be off by as much as a factor of 2 inthe absolute energy absorbed per particle, but the relativecomparison between the pure and mixed particles will not beaffected. Eqn (A4) can be rewritten as follows:Eabs¼ ECO2ApACO2ð1C0expðC0sclÞÞ ðA5ÞThe absorption crosssection (s) was measured with an FTIRspectrometer for pure components of oleyl alcohol or oleicacid and is given in units of cm2molecC01. The concentration cis given as the number density (Np) of molecules in the particle(molecule per cm3). The path length l(units of cm) is approximatedby calculating the height of a cylinder with the same volumeand radius (rp) as the particle studied which yields a pathlength of 4rp/3. (For the purposes of this paper, the exactcorrection factor for path length is irrelevant since we are onlycomparing relative ratios).Thus Eabscan be determined as follows:Eabs¼ ECO2ApACO21C0exp C0sNp4rp3C18C19C20C21C26C27ðA6ÞFor the two component particle, eqn (A6) can be rewritten asfollows where the contribution from each component is takeninto account:Eabs¼ ECO2ApACO21C0exp C0ðs1N1þs2N2Þ4rp3C20C21C26C27ðA7Þs1and s2are the respective crosssections of each of the purecomponents and N1and N2are the molecular densities of eachspecies in the mixed particle. There is some uncertainty in theabsorption crosssection terms as absorption crosssections canvary with solvents and vary with temperature during theheating process. Also, the absorption crosssections determinedfor bulk solutions do not necessarily apply directly to aerosolparticles.AcknowledgementsThis work was performed at the UBC Laboratory forAdvanced Spectroscopy and Imaging Research (LASIR).The authors would like to thank the National Sciences andEngineering Research Council of Canada (NSERC), theCanadian Foundation for Climate and Atmospheric Sciences(CFCAS), the Canadian Foundation for Innovation (CFI),the British Columbia Knowledge Development Fund(BCKDF), the NRC-NSERC-BDC Nanotechnology Initiative,and the Canada Research Chair Program for financialsupport.References1U.Po¨ schel, Angew. Chem., Int. Ed., 2005, 44, 7520–7540.2 B. J. Finlayson-Pitts and J. N. Pitts Jr., Chemistry of the Upper andLower Atmosphere: Theory, Experiments, and Applications,Academic Press, San Diego, 2000.3 M. Kanakidou, J. H. Seinfeld, S. N. Pandis, I. Barnes,F. J. Dentener, M. C. Facchini, R. Van Dingenen, B. Ervens,A. Nenes, C. J. Nielsen, E. Swietlicki, J. P. Putaud, Y. Balkanski,S. Fuzzi, J. Horth, G. K. Moortgat, R. Winterhalter, C. E. L.Myhre, K. Tsigaridis, E. Vignati, E. G. Stephanou and J. Wilson,Atmos. Chem. Phys., 2005, 5, 1053–1123.4 Y. Rudich, Chem. Rev., 2003, 103, 5097–5124.5 J. Cabalo, A. Zelenyuk, T. Baer and R. E. Miller, Aerosol Sci.Technol., 2000, 33, 3–19.6 A. Lazar, P. T. A. Reilly, W. B. Whitten and J. M. Ramsey,Environ. Sci. Technol., 1999, 33, 3993–4001.7 M. R. Canagaratna, J. T. Jayne, J. L. Jimenez, J. D. Allan,M. R. Alfarra, Q. Zhang, T. B. Onasch, F. Drewnick, H. Coe,A. Middlebrook, A. Delia, L. R. Williams, A. M. Trimborn,M. J. Northway, P. F. DeCarlo, C. E. Kolb, P. Davidovits andD. R. Worsnop, Mass Spectrom. Rev., 2007, 26, 185–222.8 S. J. Hanna, P. Campuzano-Jost, E. A. Simpson, I. Burak,M. W. Blades, J. H. Hepburn and A. K. Bertram, Phys. Chem.Chem. Phys., 2009, 11, 7963–7975.9 M. J. Northway, J. T. Jayne, D. W. Toohey, M. R. Canagaratna,A. Trimborn, K. I. Akiyama, A. Shimono, J. L. Jimenez,P. F. DeCarlo, K. R. Wilson and D. R. Worsnop, Aerosol Sci.Technol., 2007, 41, 828–839.11574 | Phys. Chem. Chem. Phys., 2010, 12, 11565–11575 This journal is C13c the Owner Societies 2010Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView Online10 S. J. Hanna, P. Campuzano-Jost, E. A. Simpson, D. B. Robb,I. Burak, M. W. Blades, J. W. Hepburn and A. K. Bertram, Int. J.Mass Spectrom., 2009, 279, 134–146.11 J. D. Hearn and G. D. Smith, J. Phys. Chem. A, 2004, 108,10019–10029.12 J. D. Hearn, A. J. Lovett and G. D. Smith, Phys. Chem. Chem.Phys., 2005, 7, 501–511.13 J. D. Hearn and G. D. Smith, Int. J. Mass Spectrom., 2006, 258,95–103.14 D. Voisin, J. N. Smith, H. Sakurai, P. H. McMurry andF. L. Eisele, Aerosol Sci. Technol., 2003, 37, 471–475.15 B. W. LaFranchi, J. Zahardis and G. A. Petrucci, Rapid Commun.Mass Spectrom., 2004, 18, 2517–2521.16 B. W. LaFranchi and G. A. Petrucci, Int. J. Mass Spectrom., 2006,258, 120–133.17 J. Zahardis, B. W. LaFranchi and G. A. Petrucci, Atmos. Environ.,2006, 40, 1661–1670.18 Y. X. Su, M. F. Sipin, H. Furutani and K. A. Prather, Anal.Chem., 2004, 76, 712–719.19 A. Zelenyuk, J. Cabalo, T. Baer and R. E. Miller, Anal. Chem.,1999, 71, 1802–1808.20 A. Zelenyuk and D. Imre, Aerosol Sci. Technol., 2005, 39, 554–568.21 M. Bente, T. Adam, T. Ferge, S. Gallavardin, M. Sklorz, T. Streibeland R. Zimmermann, Int. J. Mass Spectrom., 2006, 258, 86–94.22 M. Bente, M. Sklorz, T. Streibel and R. Zimmermann, Anal.Chem., 2009, 81, 2525–2536.23 B. D. Morrical, D. P. Fergenson and K. A. Prather, J. Am. Soc.Mass Spectrom., 1998, 9, 1068–1073.24 B. D. Morrical and R. Zenobi, Atmos. Environ., 2002, 36, 801–811.25 J. N. Shu, S. K. Gao and Y. Li, Aerosol Sci. Technol., 2008, 42,110–113.26 J. N. Shu, K. R. Wilson, M. Ahmed and S. R. Leone, Rev. Sci.Instrum., 2006, 77, 043106.27 E. R. Mysak, K. R. Wilson, M. Jimenez-Cruz, M. Ahmed andT. Baer, Anal. Chem., 2005, 77, 5953–5960.28 E. Woods, R. E. Miller and T. Baer, J. Phys. Chem. A, 2003, 107,2119–2125.29 E. Woods, G. D. Smith, Y. Dessiaterik, T. Baer and R. E. Miller,Anal. Chem., 2001, 73, 2317–2322.30 E. Woods, G. D. Smith, R. E. Miller and T. Baer, Anal. Chem.,2002, 74, 1642–1649.31 D. C. Sykes, E. Woods, G. D. Smith, T. Baer and R. E. Miller,Anal. Chem., 2002, 74, 2048–2052.32 W. F. Rogge, L. M. Hildemann, M. A. Mazurek, G. R. Cass andB. R. T. Simonelt, Environ. Sci. Technol., 1991, 25, 1112–1125.33 D. G. Nash, X. F. Liu, E. R. Mysak and T. Baer, Int. J. MassSpectrom., 2005, 241, 89–97.34 E. A. Simpson, P. Campuzano-Jost, S. J. Hanna, D. B. Robb,J. W. Hepburn, M. W. Blades and A. K. Bertram, Int. J. MassSpectrom., 2009, 281, 140–149.35 T. Ferge, F. Mu¨ hlberger and R. Zimmermann, Anal. Chem., 2005,77, 4528–4538.36 W. A. Harris, P. T. A. Reilly and W. B. Whitten, Int. J. MassSpectrom., 2006, 258, 113–119.37 T. Adam and R. Zimmermann, Anal. Bioanal. Chem., 2007, 389,1941–1951.38 B. Oktem, M. P. Tolocka and M. V. Johnston, Anal. Chem., 2004,76, 253–261.39 E. Gloaguen, E. R. Mysak, S. R. Leone, M. Ahmed andK. R. Wilson, Int. J. Mass Spectrom., 2006, 258, 74–85.40 K. R. Wilson, M. Jimenez-Cruz, C. Nicolas, L. Belau, S. R. Leonetand M. Ahmed, J. Phys. Chem. A, 2006, 110, 2106–2113.This journal is C13c the Owner Societies 2010 Phys.Chem.Chem.Phys.,2010, 12, 11565–11575 | 11575Downloaded by The University of British Columbia Library on 18 April 2011Published on 16 August 2010 on http://pubs.rsc.org | doi:10.1039/C0CP00462FView Online

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 10 8
Japan 3 0
Canada 2 0
China 2 0
City Views Downloads
Mountain View 6 0
Ashburn 4 0
Tokyo 3 0
Vancouver 2 0
Beijing 1 0
Shenzhen 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.32536.1-0041860/manifest

Comment

Related Items