Fragmentation reactions of alkylphenylammonium ions

JOURNAL OF MASS SPECTROMETRYJ. Mass Spectrom. 34, 1253–1273 (1999)

Fragmentation Reactions ofAlkylphenylammonium Ions

Alex G. Harrison*Department of Chemistry, University of Toronto, Toronto, Ontario M5S 3H6, Canada

The fragmentation reactions of a variety of alkylphenylammonium ions, C6H5NH3−nRnY (n ≥ 1, R = CH3, C2H5,

i -C3H7, n-C4H9) were studied by energy-resolved mass spectrometry. Ionization was by fast atom bombardment(FAB) or electrospray ionization. Energy-resolved fragmentation data were obtained by low-energy collision-induced dissociation (CID) in the quadrupole cell of a hybrid sector/quadrupole instrument following FABionization and by cone-voltage CID in the interface region of the electrospray/quadrupole instrument. Acomparison of the two methods of obtaining energy-resolved data showed that very similar results are obtainedby the two methods. The fragmentation reactions of the alkylphenylammonium ions are rationalized in termsof competitive formation of an [RY—NC6H5H3−nRn−1] complex or a [C6H5H3−nRn−1NYž—žR] complex. Theformer complex fragments by internal proton transfer to yield C6H5H3−nRn−1NHY and [R− H] whereas thelatter complex fragments to form C6H5H3−nRn−1NY and an alkyl radical. Alkane elimination, which is veryprominent for tetraalkylammonium ions, most likely involves sequential elimination of an alkyl radical andeither an H atom or an alkyl radical for the phenyl-substituted ammonium ions. Copyright 1999 John Wiley& Sons, Ltd.

KEYWORDS: alkylphenylammonium ions; fragmentation mechanisms; energy-resolved mass spectrometry; cone-voltagecollision-induced dissociation; electrospray

INTRODUCTION

The gas-phase ion chemistry of alkylammonium ions,RnH4�nNC, has seen extensive study1–13 and presentsan interesting case in which the mode of fragmenta-tion depends on the extent of alkyl group substitu-tion. When R is propyl or larger andn � 3, the majorlow-energy fragmentation reactions involve either elim-ination of an alkene (R� H) to form an ammoniumion of lower carbon content or formation of the alkylion RC,7,8,10,12 although in high-energy collision-induceddissociation (CID) significant alkane elimination mayoccur.3,4,10 By contrast, tetraalkylammonium ions.n D4/ show very abundant elimination of alkanes both inlow-energy metastable ion and CID fragmentation andin high-energy CID.1,2,4–6,9,11,12 The alkanes eliminatedfrom R4NC correspond to RH and to the alkane derivednominally by combination of R with the radical (often[R� CH2]) formed by fission of a C–C bond to thenitrogen. Thus, the tetramethylammonium ion shows elim-ination of CH4,11 the tetraethylammonium ion shows elim-ination of C2H6 and C3H8,11 the tetra-n-propylammoniumion shows elimination of C3H8 and C5H12

5,12 and thetetra-n-butylammonium ion shows elimination of C4H10and C7H16.5,6,9,12

* Correspondence to: A. G. Harrison, Department of Chemistry,University of Toronto, Toronto, Ontario M5S 3H6, Canada.E-mail: [email protected]

Contract/grant sponsor: Natural Sciences and Engineering Re-search Council (Canada).

The fragmentation of alkyl- and dialkylammonium ionshas been considered8,10,13 to proceed by formation ofan [RC—NH2R0] ion–dipole complex; this complex mayfragment either by HC transfer from RC to the amine toform R0NH3

C or by simple cleavage to form RC. In con-trast, the fragmentation of tetraalkylammonium ions hasbeen considered2,5,6,9,11 to proceed by homolytic cleav-age of the C–N bond to form the complex [Rž—žCNR3],which may fragment to form R3NCž or, within the com-plex, H-atom or alkyl group abstraction by Rž leads toalkane elimination and formation of immonium ions. Asa logical merging of these two separate mechanisms,Beranova and Wesdemiotis11 and Bosma and Harrison12

proposed that these two fragmentation routes are in com-petition, as outlined in the generalized Scheme 1, andthat the relative energetics determine which fragmenta-tion route predominates. Bosma and Harrison12 showedthat the fragmentation ofn-propyl-.n D 2–4/ and n-butylammonium.n D 2–4/ ions could be rationalized onthis basis. Beranova and Wesdemiotis11 suggested that thecompetition between the two routes was determined by therelative ionization energies of Rn�1H4�nN and Rž. Indeed,it has been shown12 that for the two competing reactions

RnH4�nNC ��! RC C Rn�1H4�nN .1/

��! Rž C Rn�1H4�nNCž .2/

the difference in reaction enthalpies is given by

H°1�H°

2 D IE.Rž/� IE.Rn�1H4�nN/ .3/

In this paper, we report a study of the fragmenta-tion reactions of phenylalkylammonium ions containing

CCC 1076–5174/99/121253–21 $17.50 Received 29 May 1999Copyright 1999 John Wiley & Sons, Ltd. Accepted 30 August 1999

1254 A. G. HARRISON

Scheme 1

different alkyl groups and different numbers of alkylgroups in an attempt to obtain additional evidence as tothe competition between the two pathways in Scheme 1.Beranova and Wesdemiotis11 studied the metastable ionfragmentation and high-energy CID fragmentation of thephenyltrimethylammonium ion and noted that CH4 elim-ination, which is observed for metastable fragmenta-tion of the tetramethylammonium ion, is not observedin metastable ion fragmentation but is observed underhigh-energy CID conditions for the phenyl-substitutedammonium ion. It will be shown in the following thatalkane elimination, which is commonly observed fortetraalkylammonium ions, does not occur for phenylalky-lammonium ions; rather, the nominal alkane eliminationreaction involves sequential elimination of two alkyl rad-icals (viz. for C6H5N.C2H5/3C sequential elimination ofC2Hž5 and CHž3).

The ammonium ions were prepared by fast atom bom-bardment (FAB) ionization and/or electrospray ionization(ESI). In the FAB studies the fragmentation reactionsof the ammonium ions were examined by metastableion studies and by variable low-energy CID studies,the so-called energy-resolved mass spectrometric (ERMS)technique.14–16 The ESI studies were carried out using anelectrospray source coupled to a single-quadrupole massanalyzer. It is well known17–20 that, in such instruments,fragmentation can be induced by collisional activation inthe higher pressure regions as the ions pass from the sourceinto the quadrupole mass analyzer. This CID processhas been variously called in-source CID, nozzle–skimmerfragmentation, cone-voltage fragmentation or high-orificepotential fragmentation.21 A number of studies22–26 haveshown that, as the electric field in this sampling region isincreased, the average energy imparted to the decomposingions increases. This raises the possibility that one might beable to obtain energy-resolved CID data by ‘cone-voltage’CID similar to that obtained by varying the collision energyin quadrupole collision cells. A significant aspect of thepresent study is a comparison of energy-resolved CID dataobtained by varying the electric field, cone-voltage CID, inthe interface region of the ESI/quadrupole instrument with

energy-resolved CID data obtained by quadrupole cell CIDusing a hybrid BEqQ mass spectrometer. A similar com-parison involving the fragmentation of protonated peptidesis presented elsewhere.27

EXPERIMENTAL

The experimental work using FAB ionization was carriedout using a VG Analytical (Wythenshawe, Manchester,UK) ZAB-2FQ hybrid BEqQ mass spectrometer whichhas been described in detail previously.28 Briefly, theinstrument is a reversed-geometry (BE) double-focusingmass spectrometer that is followed by a third stage con-sisting of a deceleration lens system, a radiofrequency(r.f.)-only quadrupole collision cell q and a quadrupolemass analyzer Q. In the unimolecular metastable ion frag-mentation studies the appropriate ion beam was mass-selected by the BE double-focusing mass spectrometer at6 keV ion energy, decelerated to¾20 eV kinetic energyand introduced into the r.f.-only quadrupole cell in theabsence of collision gas. The products of unimolecularfragmentation in the cell were measured using the finalmass-analyzing quadrupole. A similar procedure was fol-lowed in the low-energy CID experiments but with theaddition to the quadrupole cell of N2 collision gas at apressure of 1ð 10�7–2ð 10�7 Torr (1 TorrD 133.3 Pa)as indicated by the ionization gauge attached to the pump-ing line for the region. In the CID studies the collisionenergy typically was varied from 5 to 45 or 50 eV. Theresults are presented in the following as breakdown graphsexpressing the percentage fragment ion abundance as afunction of the collision energy (laboratory scale). Inboth the unimolecular and CID studies, 20–50 2 s scans

Figure 1. Schematic diagram of electrospray/quadrupole massspectrometer.

Copyright 1999JohnWiley & Sons,Ltd. J. MassSpectrom. 34, 1253–1273(1999)

FRAGMENTATION REACTIONS OF ALKYLPHENYLAMMONIUM IONS 1255

were accumulated on a multi-channel analyzer to improvethe signal-to-noise ratio. The phenylalkylammonium ionswere prepared by FAB ionization using an Ar atombeam of 7–8 keV kinetic energy with the samples dis-solved in either a glycerol or a 1 : 1 thioglycerol–2,20-dithiodiethanol (saturated with oxalic acid) matrix. Thereis substantial evidence that anilines are protonated on thenitrogen in FAB ionization, although gas-phase Brønstedacid chemical ionization does lead to substantial ringprotonation.29–31

The electrospray ionization experiments were carriedout using a VG Platform quadrupole mass spectrom-eter (Micromass, Manchester, UK) equipped with anatmospheric pressure electrospray source. Samples at

micromolar concentrations in either a 1 : 1 CH3CN–H2Oor 1 : 1 CH3OH–H2O solvent were introduced at a flow-rate of 20µl min�1. Nitrogen was used as nebulizing gas(10 l h�1) and drying gas (200–250 l h�1). A potential of¾2.5 kV was applied to the capillary. A schematic dia-gram of the electrospray source is presented in Fig. 1.Collisional activation can be carried out in the regionbetween the sample cone and the skimmer by varying thepotential applied to the sample cone (the skimmer beingheld at ground potential). Since the pressure in this regionis ¾1 mbar, CID clearly is being carried out under mul-tiple collision conditions compared with the near-singlecollision conditions in the quadrupole cell of the BEqQinstrument. Typically 16 2 s scans were accumulated by

Figure 2. Comparison of breakdown graphs obtained by quadrupole cell (A) and cone-voltage (B) CID of trimethylphenylammo-nium ion.


1256 A. G. HARRISON

the data system. The results of the skimmer cone CIDexperiments are presented in the following in the formof breakdown graphs showing either the percentage frag-ment ion abundance as a function of cone voltage or thepercentage total ion abundance as a function of cone volt-age, the latter plots including the precursor or parent ionabundance.

All samples studied were obtained from AldrichChemical (Milwaukee, WI, USA), with the exceptionof triethylphenylammonium hydroxide (10% aqueoussolution), which was obtained from Pfaltz and Bauer

(Waterbury, CT, USA). The samples showed no impuritiesin their mass spectra and were used as received.

RESULTS AND DISCUSSION

Comparison of quadrupole cell and cone-voltage CID

Figures 2–5 compare the breakdown graphs (percent-age fragment ion abundance versus collision energy

Figure 3. Comparison of breakdown graphs obtained by quadrupole cell (A) and cone-voltage (B) CID of triethylphenylammonium ion.



Figure 4. Comparison of breakdown graphs obtained by quadrupole cell (A) and cone-voltage (B) CID of diethylphenylammonium ion.

or cone voltage) obtained by quadrupole cell CIDand by cone-voltage CID for the trimethylphenyl-, triethylphenyl-, diethylphenyl- and ethylphenylammo-nium ions, respectively.The relevantions werepreparedby FAB ionization for the quadrupolecell CID experi-mentsandby ESI for the skimmerconeCID studies.Forthe trimethylphenylammoniumion the primary fragmen-tation reaction involves loss of CH3 with further frag-mentation by loss of H from the [M � CH3]C ion athigher collision energy. The major primary fragmenta-tion reactionsfor the ethylphenylammoniumions involve

elimination of ethyleneor elimination of an ethyl rad-ical. For the triethylphenylammoniumion a secondaryfragmentationreaction involves loss of a methyl radi-cal from [M � C2H5]C (m/z 149) to give m/z 134, whilefor the diethylphenylammoniumion a secondaryreactioninvolves loss of CH3 from the [M � C2H5]C ion to givem/z 106.A detaileddiscussionof the fragmentationreac-tions of the variousammoniumions is presentedbelow.

The important observationat the moment is that thevariationof relativefragmention abundanceswith increas-ing conevoltageclosely mimics the variation of relative


1258 A. G. HARRISON

Figure 5. Comparison of breakdown graphs obtained by quadrupole cell (A) and cone-voltage (B) CID of ethylphenylammonium ion.

fragmention abundanceswith increasingcollision energyin the quadrupolecell. It is well established16,32–36 that,in low-energy CID in quadrupolecells, thekinetic energytransformedinto internal energy increaseswith increas-ing collision energy. As a result,breakdowngraphssuchas those displayedin Figs 2–5 reflect, in a qualitativesense,the evolution of the fragmentationof the selectedion with increasinginternalenergy. The closecorrespon-denceof theresultsobtainedin theESI/cone-voltageCIDstudieswith thoseobtainedin the quadrupolecell CIDstudieslendsfurther supportto earlier observations22–27

that the averageinternal energy of the fragmentingionsincreaseswith increasingcone voltage in cone-voltageor in-sourceCID. As a result of thesestudieswe con-cludethatbreakdowngraphsexpressing,qualitatively,theenergy dependenceof the fragmentationof gaseousionscan be obtainedby cone-voltageCID with the ions pre-paredby electrosprayionization or atmosphericpressurechemicalionization. The disadvantageof this approach,comparedwith quadrupolecell CID, is that one cannotemploymassselectionof theion to bestudied.As aresult,the methodcanbe usedonly whereone dominantion is



formed in the ionization process. This was the case in thepresent studies.

Fragmentation of methylphenylammonium ions

The fragmentation reactions of the trimethylphenyl- anddimethylphenylammonium ions were investigated. In uni-molecular metastable ion fragmentation in the quadrupolecell both ammonium ions showed elimination of CH3only. This is in agreement with the observations of Bera-nova and Wesdemiotis11 for the trimethylphenylammo-nium ion. By contrast, the tetramethylammonium ionshows metastable ion fragmentation by loss of both CH3and CH4.11

The breakdown graphs obtained by cone-voltage CIDare shown in Figs 6 and 7, that for the trimethylpheny-lammonium ion (Fig. 6) differing from the plot in Fig. 2by extending the data to higher cone voltages and plot-ting the results as percentage of total ion abundance ratherthan as percentage of fragment ion abundance. It appearsfrom Figs 2, 6 and 7 that the fragmentation reaction lead-ing, nominally, to methane elimination involves sequentialelimination of CH3 C H. This is in agreement with theresults of Glishet al.,37 who showed, from MS/MS/MSexperiments in a quadrupole ion trap, that them/z 121ion derived from the trimethylphenylammonium ion frag-mented by loss of H to givem/z 120. Beranova andWesdemiotis11 observed significant fragmentation of the

Figure 6. Breakdown graph for trimethylphenylammonium ion.


1260 A. G. HARRISON

Figure 7. Breakdown graph for dimethylphenylammonium ion.

trimethylphenylammoniumion by elimination of C6H5and C6H6 in high-energy CID; theseproducts,at m/z 59andm/z 58, respectively,wereobservedin only low yieldat the highest collision energies in the quadrupolecellCID studies(Fig. 2) and werenot observedin the cone-voltage CID experiments.Glish et al.37 did not observetheseproductsin their ion trapCID studies.It appearsthateliminationsof C6H5 andC6H6 havehigh critical reactionenergieswhich aresurpassedto a significantextentonlyin high-energy collisional activation.

Glish et al.,37 from MS/MS/MS/MSexperimentsin theion trap, reportedthat the m/z 120 ion derivedfrom thetrimethylphenylammoniumion fragmentedto give ionsof m/z 118, 103, 91, 79 and 77; the samefragmentation

reactionswere observed38,39 in ion trap CID of the m/z120 ion .[M � H]C/ derived by dissociativeionizationof dimethylaniline. As Fig. 6 shows, we observeonlyminor formation of m/z 118 with somewhatgreaterfor-mation of m/z 105 and 104 on further fragmentationofm/z 120, the major fragmentationchannelgiving m/z 77;in particular, we did not observeany productsof m/z103, 91 or 79. We are unable to explain this differ-ence.The presentresultssuggestrelatively simple frag-mentationsequencesfor both the trimethylphenyl- anddimethylphenylammoniumions asoutlined in Scheme2.Significant rearrangementwould needto be invoked torationalize the formation of productswith m/z 103, 91and79.



Scheme 2

It should be noted that the multiple collision condi-tionsof CID in the interfaceregionof theESI/quadrupoleinstrumentallow one to follow the fragmentationof therelevantion further along the degradationpathwaythanis possiblein quadrupolecell CID, at leastundersingle-collision conditionsin thelatter.This is clearfrom a com-parisonof Fig. 2(A) and Fig. 6. In addition, quadrupolecell CID of the dimethylphenylammoniumion showedamuch lower extent of fragmentationto m/z 106 and 77evenat a 55 eV collision energy31 than occursat higherconevoltagesin skimmerconeCID (Fig. 7).

Fragmentation of ethylphenylammonium ions

The fragmentation reactions of the triethylphenyl-,diethylphenyl- and ethylphenylammoniumions werestudiedin detail. For comparisonpurposesthe tetraethy-lammoniumion wasincludedin thestudy.Themetastableion spectra of the four ammonium ions, as recordedfor unimolecularfragmentationreactionsoccurringin thequadrupolecell, are given in Table1. The breakdowngraphobtainedby quadrupolecell CID for the tetraethy-lammoniumion is presentedin Fig. 8, while the break-down graphs for the three ethylphenylammoniumionsobtainedby cone-voltageCID arepresentedin Figs 9–11.The latter extendto higherconevoltagesthanthe resultsshownin Figs 3–5 andareplottedasa percentageof thetotal ion abundancerather than as a percentageof thefragmention abundance.

The tetraethylammoniumion fragmentsin metastableion reactionsby elimination of C2H5, C2H6 and C3H8;the last elimination dominatesthe breakdowngraphpar-ticularly at higher collision energies. Theseresults arein essentialagreementwith the resultsof Beranova and

Table 1. Metastable ion fragmentation of ammonium ions

Neutral loss (%)Ion �C2H4 �C2H5 �C2H6 �C3H8

.C2H5/4NC 4.0 38.1 31.0 26.9C6H5.C2H5/3NC 38.7 56.2 3.6 1.5C6H5.C2H5/2NHC 63.6 36.4C6H5.C2H5/NH2

C 90.0 10.0

Scheme 3

Wesdemiotis.11 Replacementof an ethyl by a phenylgroup,asin the triethylphenylammoniumion, resultsin amajorchangein thefragmentationreactionsobserved.Thedominantmetastableion fragmentationreactionsbecomeelimination of C2H4 and C2H5 with only minor elimina-tion of C2H6 and C3H8. Indeed,the breakdowngraphs(Figs 3 and9) suggeststrongly that the major pathwayto the m/z 134 ion (nominal C3H8 elimination) involvessequentialelimination of C2H5 and CH3 .m/z 178!m/z 149! m/z 134/. Similarly, the breakdowngraphs(Figs 4 and10) for the diethylphenylammoniumion sug-gestthat them/z 106 ion (nominalC3H8 elimination)alsois formed by sequentialelimination of C2H5 and CH3.m/z 150! m/z 121! m/z 106/. The ethyleneelimi-nation reaction forming m/z 150 for the triethylpheny-lammoniumion andm/z 122 for the diethylphenylammo-nium ion decreasein importanceas the collision energyis increased.For the ethylphenylammoniumion, ethyleneelimination to form m/z 94 is the dominant metastableion reaction;however,elimination of C2H5 to form m/z93 increasesrapidly in importanceas the internalenergyof the fragmentingions increases(Figs 5 and11). Themajor fragmentationsequencesfor thetriethylphenyl-anddiethylphenylammoniumions originateby initial loss ofC2H5 andareoutlined in Scheme3. It alsoappearsthat,in bothcases,them/z 106ion fragmentsto a minor extentto form m/z 79 and 78 in addition to the major channelforming m/z 77 .C6H5

C/. The m/z 94 and 93 fragments


1262 A. G. HARRISON

Figure 8. Breakdown graph obtained by quadrupole cell CID for tetraethylammonium ion. FAB ionization.

observed(Fig. 10) in low abundancein the fragmenta-tion of the diethylphenylammoniumion presumablyorig-inate from further fragmentationof them/z 122 ion (seeFig. 11).

The fragmentationreactionsof the mixed methylethyl-phenylammoniumion werestudiedby cone-voltageCIDwith the resultsshownin Fig. 12. The major fragmenta-tion pathwayoriginatesby elimination of an ethyl radi-cal to give theN-methylanilinemolecularion (m/z 107),which fragmentsfurther by sequentialelimination of Hand CH2 NH. As a result, the breakdowngraphhasastrong resemblanceto that of the dimethylphenylammo-nium ion. Therealsois minor eliminationof ethyleneandeliminationof a methyl radical.

Fragmentation of isopropylphenylammonium ions

Isopropylphenyl-and diisopropylphenylammoniumionswere studied using electrosprayionization and cone-voltageCID. The breakdowngraphsobtainedareshownin Figs 13 and14.For theisopropylphenylammoniumionthe major fragmentationreactioninvolves elimination ofC3H6 to giveprotonatedaniline(m/z 94),which fragmentsat highercollision energiesby eliminationof ammoniatogivem/z 77. The diisopropylphenylammoniumion showssequentialelimination of two moleculesof propenetogive eventuallyprotonatedaniline. In both systemsthereis minor formation of C3H7

C (m/z 43). In particularonenotesthe absenceof alkyl radical .C3Hž7/ elimination, a



Figure 9. Breakdown graph for triethylphenylammonium ion.

reactionof major importancefor the ethylphenylammo-nium ions.

The mixed isopropylethylphenylammoniumion alsowasstudiedby cone-voltageCID with the resultsshownin Fig. 15. Although the major primary fragmentation

reaction involves elimination of propene to give theethylphenylammoniumion (m/z 122), there is minorelimination of a propyl radical to give m/z 121 andpropaneto give m/z 120. The ethylphenylammoniumion(m/z 122) fragmentsfurther to give protonatedaniline


1264 A. G. HARRISON

Figure 10. Breakdown graph for diethylphenylammonium ion.

(m/z 94) and ionized aniline (m/z 93) as observedear-lier (see Fig. 11). No primary fragmentationreactionswere observedinvolving elimination of C2H4 or elimi-nationof C2H5.

Fragmentation of n-butylphenylammonium ions

The fragmentation of n-butylphenyl- and di-n-butyl-phenylammoniumions were studied by cone-voltage



Figure 11. Breakdown graph for ethylphenylammonium ion.

CID with the results shown in Figs 16 and17. Thebutylphenylammoniumionsshowsprimaryfragmentationeitherby eliminationof buteneto give protonatedaniline(m/z 94) or by elimination of aniline to give thebutyl ion C4H9

C (m/z 57) with the latter fragmentingfurther to give C3H5

C and C2H5C. The fragmentation

reactionsof the dibutylphenylammoniumion are morecomplex.In addition to the formation of C4H9

C and itsfurther fragmentationproducts,one observessequentialelimination of two molecules of butene .m/z 206!m/z 150! m/z 94/. Therealsois significanteliminationof abutyl radicalto formm/z 149,whichfragmentsfurtherby elimination of C3Hž7 to give the product with m/z

106; the latter fragmentsto m/z 77 .C6H5C/, asobserved

previously.

Mechanistic implications

The experimentalresults presentedabove will be dis-cussedin termsof the competingreactionsin Scheme1involving eitherheterolyticcleavageor homolytic cleav-age of the R—N bond. As discussedpreviously andillustrated in Scheme4, for the competingreactions(1)and (2) the difference in reaction enthalpiesis givenby IE.Rž/� IE.Rn�1H4�nN/. Accordingly, the resultswill be discussedin terms of the relative ionization


1266 A. G. HARRISON

Figure 12. Breakdown graph for methylethylphenylammonium ion.

energies, although it is recognizedthat theseare onlyapproximatemeasuresof the relative energies neces-sary to reachthe two complexes[RC—NRn�1H4�n] and[Rn�1H4�nNCž—žR]. Table2 lists the availableioniza-tion energies40 for the speciesof interest in the presentstudy.The ionizationenergiesreportedfor N-ethylanilineand N-isopropylaniline are possibly anomalouslyhighwhen comparedwith the reportedionization energy forN-methylaniline.

The dominant loss of CH3 from the methylpheny-lammonium ions is consistentwith the formation ofa ionized aniline–methyl radical complex which frag-ments by methyl radical loss. As the data in Table2

Table 2. Ionization energiesof anilines and alkyl radicals

Aniline IE (eV) Alkyl radical IE (eV)

C6H5NH2 7.72 CH3 9.84C6H5NHCH3 7.33 C2H5 8.13C6H5N.CH3/2 7.12 i-C3H7 7.36C6H5NHC2H5 7.56 n-C4H9 8.02C6H5N.C2H5/2 6.98 s-C4H9 7.25C6H5NHC3H7 7.5

show, the ionization energy of CHž3 is much greaterthan that of either aniline or N-methylaniline. The



Figure 13. Breakdown graph for isopropylphenylammonium ion.

results for the isopropylphenylammoniumions are alsoeasily interpreted.Since IE(isopropyl) is considerablylower than IE(aniline), fragmentation occurs exclu-sively by heterolytic bond cleavageto form the [i�C3H7

C—NH2C6H5] complexwithin which proton trans-fer occurs to give protonatedaniline since the protonaffinity of aniline (210 kcal mol�1 (1 kcalD 4.184 kJ))41

is greaterthan the proton affinity of propene(180 kcalmol�1).41 As expected,thereis someformationof C3H7

C(and fragmentation products thereof) at higher colli-sion energies. The dominant elimination of C3H6 from

the isopropylethylphenyl-and diisopropylphenylammo-nium ions also is readily understandablein terms ofthe relative ionization energies if the ionization ener-gies of N-ethylanilineandN-isopropylanilineare indeed¾7.5 eV. If they are significantly lower one might haveexpected greater homolytic bond cleavage and elim-ination of a propyl radical, but this is observed toonly a minor extent. The relative ionization energiesalso rationalize the absenceof ethyleneor ethyl radi-cal loss from the isopropylethylphenylammonium ionsincetheformationof eitherthe[C2H5

C—NH.i-Pr)C6H5]


1268 A. G. HARRISON

Figure 14. Breakdown graph for diisopropylphenylammonium ion.

or the [C6H5.i-Pr)NHCž—žC2H5] complex is moreenergy demandingthan the formation of the [C3H7

C—NH(Et)C6H5] complex.

The dominantelimination of butenefrom the n-butyl-phenyl- and di-n-butylphenylammoniumions also canbe rationalized in terms of Scheme1 if one assumesa hydrogen shift during heterolytic bond cleavagetoform a .sec-butyl/C–aniline complex rather than an.n-butyl/C–aniline complex. A similar rearrangementappearsto occur during the fragmentationof butylam-monium ions.12 We cannot rule out the possibility thatthere is skeletal rearrangementof the butyl group to

give an incipient tert-butyl ion. There is considerablygreaterformationof theC4H9

C ion (andfragmentsderivedtherefrom) than there is formation of the propyl ionfrom the isopropylphenylammoniumions. This could bean indication that the tert-butyl ion is formed sincethe proton affinity of isobutene(191 kcal mol�1)41 isconsiderablyhigher than that of propene or 2-butene(179 kcal mol�1).41 For the dibutylphenylammoniumionthereis somefragmentationby loss of a butyl radical togive m/z 149, which subsequentlyfragmentsfurther bypropyl radical loss to give m/z 106. This suggeststhatthe [C6H5.C4H9/NHCž—žC4H9] complex is not much



Figure 15. Breakdown graph for isopropylethylphenylammonium ion.

higher in energy than the [C4H9C—NH(C4H9/C6H5]

complex.The high ionization energy of the ethyl radical

(8.13eV) suggeststhat fragmentation of the variousethylphenylammoniumions shouldoccur by way of the

[Rž—CžNRn�1H4�n] complexand,indeed,eliminationofan ethyl radical is a prominent fragmentationreactionfor the three ethylphenylammoniumions and for theethylmethylphenylammoniumion, particularly at higherinternalenergies.However,at lower internalenergiesand,


1270 A. G. HARRISON

Figure 16. Breakdown graph for n-butylphenylammonium ion.

particularly for metastableion fragmentation,eliminationof ethylene is a significant fragmentation reaction.The relative ionization energies indicate that ethyleneelimination by way of a C2H5

C–aniline complexshouldnotbecompetitivewith fragmentationby ethyl radicallossby way of an ionizedaniline–ethyl radicalcomplex.Thisleadsto thesuggestionthatfor theethylphenylammoniumions ethylene loss occurs by a direct mechanism,asoutlined in Scheme5, leading to a ring-protonatedaniline. Sucha direct fragmentationreactionshouldhavedemandingsteric requirementsand thus an unfavourableentropyof activation.This is in line with the decreaseinimportanceof this reactionchannelasthe internalenergyof the fragmentingions increases.

An importantobservationis that theonsetsfor nominaleliminationof CH4 from themethylphenylammoniumionsandfor nominaleliminationof C3H8 from theethylpheny-lammoniumions areconsiderablyhigher than the onsetsfor elimination of an alkyl radical.This is in contrasttothesituationfor the tetramethyl-andtetraethylammoniumions, where the onsetsfor elimination of an alkyl radi-cal and for elimination of an alkaneare very similar. Inthe discussionabovewe havetakenthis to meanthat thenominal alkaneelimination reactionfor the alkylpheny-lammoniumions involves sequentialelimination of tworadicals, CHž3C Hž for the methylphenylammoniumionsand C2Hž5 C CHž3 for the ethylphenylammoniumions.Alkane elimination is believedto occur by attackof the



Figure 17. Breakdown graph for di-n-butylphenylammonium ion.

alkyl radicalRž in the [alkylaniline ion–Rž] complexona C—H or C—C bondof the alkylaniline. In, for exam-ple, the [.C2H5/3NCž—žC2H5] complex the charge andradicalsiteslocalizedon thenitrogenserveto weakentheadjacentC—C bond,makingattackof theethyl radicalon

the C—C bondenergetically feasible.By contrast,in the[C6H5.C2H5/2NCž—žC2H5] complexthechargeandrad-ical sitescanbe delocalizedinto the phenylring with theresultthattheC—C bondis notweakened.This, in effect,raises the energy barrier for alkane elimination in the


1272 A. G. HARRISON

Scheme 4

Scheme 5

alkylphenylammoniumions comparedwith the tetraalky-lammoniumions.Two possibilitiesthenexist.Oneis thatalkane elimination does occur but with a higher onsetenergy thanalkyl radicalloss.Theotheris that thebarrierfor attackon theC—C or C—H bondis sufficiently largethat,asmoreenergy is pumpedinto theprecursorion, thealkylanilineion–radicalcomplexseparatesandtheexcitedfree alkylaniline molecularion subsequentlylosesa radi-cal, i.e. the nominalalkaneelimination reactioninvolvesthesequentiallossof two radicals.Thepresentstudydoesnotpermitadefinitivedistinctionbetweenthesetwo possi-bilities. However,oneshouldnotethatGlish et al.37 haveshown, by MS/MS/MS experiments,that the sequential

lossof CHž3C Hž doesoccur for the trimethylphenylam-monium ion. In addition, loss of CHž3 is the dominantfragmentationreaction of the N-ethylaniline molecularions formed by ethyl radical loss from the ethylpheny-lammoniumions. The questionreally is whether,as theinternalenergy of the alkylphenylammoniumion exceedstheenergy barrierfor alkaneelimination,therateof attackof the alkyl radical in the complexon a C—H or C—Cbondis sufficiently larger thanthe rateof dissociationofthe complexto maketrue alkaneeliminationa significantfragmentationreaction.We suggestthat this is not thecaseandthat the major route to nominalalkaneelimina-tion is the sequentialelimination of two radicalsfor thealkylphenylammoniumions.

CONCLUSIONS

An important conclusionfrom the presentwork is thatbreakdowngraphs,expressingthe energy dependenceoffragmentationsequences,canbeobtainedby cone-voltageCID in atmosphericpressureionization/quadrupoleinstru-ments;thesebreakdowngraphsarevery similar to thoseobtainedby low variable-energy CID in quadrupolecolli-sioncells.This observationthatenergy-resolvedCID datacanbeobtainedby cone-voltageCID is in agreementwithearlierstudies.24–27

The major fragmentationchannelsof alkylphenylam-monium ions have been shown to involve competitionbetween heterolytic C—N bond cleavage leading toformation of an [RC—NC6H5H3�nRn�1] complex andhomolytic C—N bondcleavageleadingto formationof a[C6H5H3�nRn�1NCž—žR] complex.Theformercomplexleadseither to the formation of RC or to [R� H] elimi-nationandthe formationof a protonatedaniline whereasthe latter complex leads to the formation of an anilinemolecularion. Thecompetitionbetweenthe two channelsis determinedby the relative ionization energies of thealkyl radicalandthe substitutedaniline.

Acknowledgements

The financial support of the Natural Sciencesand EngineeringResearchCouncil (Canada)is acknowledged.Thegift of theVG Plat-form to theDepartmentof Chemistryby MicromassCanadais greatlyappreciated,as is the assistanceof Dr Alex B. Young in keepingitoperational.

REFERENCES

1. Gierlach HH, Rollgen FW, Borchers F, Levsen K. Org. MassSpectrom. 1977; 12: 367.

2. Fischer M, Veith H-J. Helv. Chim. Acta 1978; 61: 3038.3. Sigsby ML, Day RJ, Cooks RG. Org. Mass Spectrom. 1979;

14: 556.4. Wagner W, Heimbach H, Levsen K. Int. J. Mass Spectrom.

Ion Phys. 1980; 36: 125.5. Ohashi M, Barrow RP, Benson WR. J. Am. Chem. Soc. 1981;

103: 3943.6. Veith H-J. Org. Mass Spectrom. 1983; 18: 154.7. Reiner EJ, Poirier RA, Peterson MR, Csizmadia IG, Harri-

son AG. Can. J. Chem. 1986; 64: 1652.8. Bowen RD, Harrison AG, Reiner EJ. J. Chem. Soc., Perkin

Trans. 2 1988; 1009.

9. Mestdagh H, Morin N, Rolando C. Org. Mass Spectrom.1988; 23: 246.

10. Reiner EJ, Harrison AG, Bowen RD. Can J. Chem. 1989; 67:2081.

11. Beranova S, Wesdemiotis C. Int. J. Mass Spectrom. IonProcesses 1994; 134: 83.

12. Bosma NL, Harrison AG. Can. J. Chem. 1994; 72: 2205.13. Audier HE, Morton TH. Org. Mass Spectrom. 1993; 28: 1218.14. McLuckey SA, Glish GL, Cooks RG. Int. J. Mass Spectrom.

Ion Phys. 1981; 39: 219.15. Fetterolf DD, Yost RA. Int. J. Mass Spectrom. Ion Phys. 1982;

44: 37.16. McLuckey SA, Cooks RG. In Tandem Mass Spectrometry,

McLafferty FW (ed). Wiley: New York, 1983; Chapt. 15.



17. Loo JA, Udseth HR, Smith RD. Rapid Commun. Mass Spec-trom. 1988; 2: 207.

18. Smith RD, Loo JA, Barinaga CJ, Edmonds CG, Udseth HR.J. Am. Soc. Mass Spectrom. 1990; 1: 53.

19. Katta V, Chowdhury SK, Chait BT. Anal. Chem. 1991; 63: 174.20. Huddleston MJ, Bean MF, Carr SA. Anal. Chem. 1993; 65:

877.21. Bruins AP. In Electrospray Mass Spectrometry: Fundamen-

tals, Instrumentation and Applications, Cole RB (ed). Wiley:New York, 1997; Chapt. 3.

22. Voyksner RD, Pack T. Rapid Commun. Mass Spectrom. 1991;5: 263.

23. Hunter CL, Mauk AG, Douglas DJ. Biochemistry 1997; 36:1018.

24. Collette C, De Pauw E. Rapid Commun. Mass Spectrom.1998; 12: 165.

25. Collette C, Drahos L, De Pauw E, Vekey K. Rapid Commun.Mass Spectrom. 1998; 12: 1673.

26. Sjoberg PJR, Markides KE. J. Mass Spectrom. 1998; 33: 872.27. Harrison AG. Rapid Commun. Mass Spectrom. 1999; 13:

1663.28. Harrison AG, Mercer RS, Reiner EJ, Young AB, Boyd RK,

March RE, Porter CJ. Int. J. Mass Spectrom. Ion Phys. 1986;74: 13.

29. Dookeran NN, Harrison AG. J. Am. Soc. Mass Spectrom.1995; 6: 19.

30. Nold MJ, Wesdemiotis C. J. Mass Spectrom. 1996; 31: 1169.31. Harrison AG, Tu Y-P. Int. J. Mass Spectrom. in press.32. Douglas DJ. J. Phys. Chem. 1982; 86: 185.33. Dawson PH, Sun W-F. Int. J. Mass Spectrom. Ion Phys. 1982;

44: 51.34. Dawson PH. Int. J. Mass Spectrom. Ion Phys. 1983; 50: 287.35. Nacson S, Harrison AG. Int. J. Mass Spectrom. Ion Phys.

1985; 63: 325.36. Zwinselman JJ, Nacson S, Harrison AG. Int. J. Mass Spec-

trom. Ion Phys. 1985; 67: 93.37. Glish GL, Goeringer DE, Asano KG, McLuckey SA. Int.

J. Mass Spectrom. Ion Phys. 1989; 94: 15.38. Glish GL, McLuckey SA, Asano KG. J. Am. Soc. Mass Spec-

trom. 1990; 1: 166.39. McLuckey SA, Glish GL, Van Berkel GJ. Int. J. Mass Spec-

trom. Ion Processes 1991; 106: 213.40. Lias SG, Bartmess JE, Liebman JF, Holmes JL, Levin RD,

Mallard WG. J. Phys. Chem. Ref. Data 1988; 17: Suppl. 1.41. Hunter EPL, Lias SG. J. Phys. Chem. Ref. Data 1998; 27: 413.

Copyright 1999 John Wiley & Sons, Ltd. J. Mass Spectrom. 34, 1253–1273 (1999)

Documents

Fragmentation reactions of alkylphenylammonium ions