Energy-resolved mass spectrometry: a comparison of quadrupole cell and cone-voltage collision-induced dissociation

Energy-resolved Mass Spectrometry: aComparison of Quadrupole Cell andCone-voltage Collision-induced Dissociation

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

Collision-induced dissociation (CID) can be effected in the interface region between atmospheric pressureionization sources and single quadrupole mass analyzers. By varying the electric field in these cone-voltageCID experiments energy-resolved mass spectra can be obtained leading to breakdown graphs exploring theenergy evolution of the fragmentation pathways. The breakdown graphs obtained from these cone-voltageCID studies are very comparable to those obtained by varying the collision energy in the quadrupolecollision cell of a BEqQ mass spectrometer. The comparison has been made for the protonated peptidesH-Leu-Gly-Gly-OH, H-Gly-Leu-Gly-OH, H-Gly-Gly-Leu-OH and Leu-enkephalin Copyright # 1999 JohnWiley & Sons, Ltd.

Received 11 May 1999; Revised 21 June 1999; Accepted 23 June 1999

The study of the fragmentation of gaseous ions has longbeen a vital source of information concerning the structuresand chemistry of such species.1,2 This information is moreuseful if the internal energy of the fragmenting ions can bevaried in a controlled fashion; such energy-resolved studiespermit the establishment of breakdown graphs expressingthe dependence of fragment ion yields on the internal energyof the decomposing ions. For odd-electron molecularcations energy-resolved data can be obtained with highprecision by photoelectron/photoion coincidence experi-ments3–5 and, with reduced precision, by charge exchangeexperiments.6–8However, for even-electron species, such asprotonated or deprotonated molecules, the approaches toenergy-resolved data are more limited. While energy-resolved data can be obtained from single-photon photo-dissociation studies by varying the wavelength of theincident photons,9,10the majority of energy-resolved studiesof even-electron ions have employed low, variable energycollision-induced dissociation (CID) experiments, the so-called energy-resolved mass spectrometric (ERMS) techni-que.11–13In this approach the CID mass spectra of the ionsof interest are recorded as a function of collision energyover the range from ca. 2 to 50–100 eV. There is ampleevidence11–18 that, at these low collision energies, thekinetic energy transformed into internal energy increases asthe collision energy is increased, although the relationshipbetween collision energy and mean internal energy islargely unknown and appears to vary from system to system.As a result, the fractional ion yields as a function of collisionenergy provide only qualitative, but nevertheless important,information as to the energy evolution of the fragmentationpathways of gaseous even-electron ions. Such low, variable

energy collisional experiments have been carried out for themost part in radio frequency only quadrupole collision cellsas part of multiple quadrupole instruments or hybrid sector/quadrupole instruments although similar experiments can becarried out using Fourier transform ion cyclotron resonancemass spectrometers or ion trap instruments.19

A major development over the past decade has been theintroduction of electrospray ionization20,21 using atmo-spheric pressure ion sources frequently coupled to a singlequadrupole mass analyzer. It is well known that, in suchinstruments, fragmentation can be induced by collisionalactivation in the higher pressure regions as the ions passfrom the source into the mass analyzer.22–25 This CIDprocess has been variously called in-source CID, nozzle-skimmer fragmentation, cone-voltage fragmentation or highorifice potential fragmentation.26 A number of studies27–31

have shown that, as the electric field in this sampling regionis increased, the average energy imparted to the decompos-ing ions increases. This raises the possibility that one mightbe able to obtain energy-resolved CID data by cone-voltagevariation similar to that obtained in quadrupole cells byvarying the collision energy. The present paper reports acomparison of energy-resolved CID data obtained byvarying the electric field in the sampling region of anatmospheric pressure electrospray ionization quadrupoleinstrument with energy-resolved CID data obtained byvarying the collision energy in the quadrupole cell of ahybrid BEqQ mass spectrometer.

EXPERIMENTAL

The quadrupole cell CID experiments were carried out usinga VG Analytical (Manchester, UK) ZAB-2FQ hybrid BEqQmass spectrometer which has been described in detailpreviously.32 Briefly, the instrument is a reversed-geometry(BE) double-focusing mass spectrometer that is followed bya third stage consisting of a deceleration lens system, a radiofrequency (rf)-only quadrupole collision cell q and aquadrupole mass analyzer Q. In the CID experiments the

*Correspondence to: A. G. Harrison, Department of Chemistry,University of Toronto, 80 St. George St., Toronto, Ontario M5S3H6, Canada.E-mail: [email protected]/grant sponsor: Natural Sciences and Engineering ResearchCouncil, Canada.

CCC 0951–4198/99/161663–08 $17.50 Copyright# 1999 John Wiley & Sons, Ltd.

RAPID COMMUNICATIONS IN MASS SPECTROMETRYRapid Commun. Mass Spectrom.13, 1663–1670 (1999)

Figure 1. Schematicof the electrospray/quadrupolemassspectro-meter.

Figure 2. Comparisonof breakdowngraphsobtained by quadrupolecell andcone-voltageCID of protonatedH-Leu-Gly-Gly-OH.

Figure 3. Comparisonof breakdowngraphsobtainedby quadrupolecell andcone-voltageCID of protonatedH-Gly-Leu-Gly-OH.

Figure 4. Comparisonof breakdowngraphsobtainedby quadrupolecell andcone-voltageCID of protonatedH-Gly-Gly-Leu-OH.

Rapid Commun.MassSpectrom.13, 1663–1670(1999) Copyright# 1999JohnWiley & Sons,Ltd.

1664 QUADRUPOLE CELL VS. CONE-VOLTAGE CID

Figure 5. % of total ion abundanceasa function of conevoltageforprotonatedH-Leu-Gly-Gly-OH.Electrosprayionization.

Figure 6. % of total ion abundanceasa function of conevoltageforprotonatedH-Gly-Leu-Gly-OH.Electrosprayionization.

Figure 7. % of total ion abundanceasa function of conevoltageforprotonatedH-Gly-Gly-Leu-OH.Electrosprayionization.

Figure 8. % of total ion abundanceasa function of conevoltageforprotonated H-Leu-Gly-Gly-OH. Atmospheric pressure chemicalionization.

Copyright# 1999JohnWiley & Sons,Ltd. Rapid Commun.MassSpectrom.13, 1663–1670(1999)

QUADRUPOLE CELL VS. CONE-VOLTAGE CID 1665

appropriate ion beamwas mass-selectedby the BE massspectrometer at 6 keV ion energy, decelerated to kineticenergiesin the5 to 50 eV rangeandintroducedinto the rf-only collisioncell containingN2 collision gasat a pressureof 1–2� 10ÿ7 Torr (1 Torr = 133.3Pa)asindicated by theionization gauge attached to the pumping line for thequadrupole region.The ionic CID fragmentation productswere measured using the final mass-analyzing quadrupoleQ. Typically 20–502 s scanswereaccumulatedon a multi-channel analyzerto improvesignal-to-noise.The resultsarepresentedin the following asbreakdowngraphs expressingthe percent fragment abundance as a function of thecollision energy (laboratory scale). The ions of interestwerepreparedby fastatombombardment (FAB) ionizationusing anAr atombeamof 7–8keV energywith thesampledissolved in either glycerol or a 1:1 thioglycerol/2,2'-dithiodiethanol (saturatedwith oxalic acid) matrix.

The cone-voltage CID experiments were carried outusing a single quadrupole VG Platform (Micromass,Manchester, UK) mass spectrometer controlled by aMassLynx Version 2–10 data system. The appropriateprotonatedspecieswere preparedat atmospheric pressureeither by electrosprayionization (ESI) or by atmosphericpressure chemical ionization (APCI). In the electrosprayexperiments the sample, at micromolar concentration ineither a 1:1 CH3CN/H2O or CH3OH/H2O solution, wasintroduced into the electrospraysourceat a flow rate of20mL minÿ1. The electrospraycapillary potential washeldat 2.5–3.0kV. N2 wasusedasnebulizing gasat a flow rateof 10 L hrÿ1 andasdrying gasat a flow rateof 250L hrÿ1.In the APCI experimentsthe sample in a 1:1 CH3CN/H2Osolution was introduced into the source at a flow rate of

50mL minÿ1 with the APCI probe held at 370°C. Thecoronadischargeneedle wassetat ca.2.9 kV. N2 wasusedasnebulizing gasat a flow rateof 100L hrÿ1 andasdryinggasat a flow rateof 250L hrÿ1.

A schematic of the instrument in the electrosprayconfiguration is shown in Fig. 1. Collision-induceddissociation is effectedin the region betweenthe sampleconeandtheskimmer by increasingthepotentialappliedtothe sample cone, the skimmer being held at groundpotential. Sincethe pressurein this region is of the orderof 1 Torr, one is clearly operating in the multi-collisionregime. Energy-resolved CID spectrawere obtained byincreasingthe conevoltagein steps.At eachconevoltage,15 2 s scanswere accumulated by the data system. Theresultsare presented in the following either asplots of %fragmention abundanceasa functionof theconevoltageoras plots of the % total ion abundance (including theprecursor ion) asa function of conevoltage. Sincethereisno massselectionof the precursor ion, isotopecorrectionsweremade where necessary.

The peptides studied were obtained from BACHEMBiosciences(King of Prussia, PA, USA) andwereusedasreceived.

RESULTS AND DISCUSSION

To compareenergy-resolved CID data obtained by cone-voltagevariation with dataobtained by varyingthecollisionenergyin thequadrupolecell, threetripeptides,H-Leu-Gly-Gly-OH, H-Gly-Leu-Gly-OH andH-Gly-Gly-Leu-OH,andthepentapeptideLeu-enkephalin (H-Tyr-Gly-Gly-Phe-Leu-OH) werestudied.

Figures2–4 comparethe breakdown graphs obtained bythetwo approaches for thethreeprotonatedtripeptides.Thebreakdown graphs for protonatedH-Leu-Gly-Gly-OH andH-Gly-Leu-Gly-OH are redrawn from an earlier publica-tion.33 For protonated H-Leu-Gly-Gly-OH (Fig. 2) andprotonatedH-Gly-Leu-Gly-OH (Fig. 3) thevariationsof ionabundanceswith conevoltageclosely mirror thevariationsof ion abundanceswith collision energy in the quadrupolecell CID experiments.For protonatedH-Gly-Gly-Leu-OH(Fig. 4) thevariationsof ion abundanceswith conevoltageand with collision energy are in qualitative agreementalthoughthere are somequantitativedifferences,particu-larly a considerably moreintensesignal at m/z189(y2@) inthe electrospray/cone-voltageCID experiments. Neverthe-less,the agreement between the threesetsof datasupportsthe contention that energy-resolved massspectral datacanbe obtained by cone-voltage CID studies which providesessentially the same information as energy-resolvedCIDdataobtained by varyingthecollision energyin quadrupolecell CID studies.

Thebreakdown graphsobtained from thequadrupolecellCID experiments are necessarily plotted as a percent offragment ion abundance since, under the near-singlecollision conditionsemployed, only a small fraction of theparent ions undergo fragmentation. However, under themultiple-collision conditions of the cone-voltage CIDexperiments, the parent ion signal is largely attenuated athigher cone voltages.An alternative, and equally useful,methodof presentingtheCID datais shown in Figs5–7forthethreeprotonatedtripeptides;in theseplots,thepercentoftotal ion abundance(includingtheparention) is plottedasafunction of the cone voltage and the fragment ions arelabelledaccordingto theusualnomenclature34,35for peptide

Figure 9. % of total ion abundanceasa function of conevoltageforprotonated H-Gly-Leu-Gly-OH. Atmospheric pressure chemicalionization.



fragment ions where possible. It is obvious that thebreakdown graphsplotted in this fashionprovide asmuchinformation concerning the energy evolution of thefragmentation of the parention asdo the plots of Figs 2–4. For example,it is clear from eitherFig. 3 or Fig. 6 thatprotonated H-Gly-Leu-Gly-OH undergoesthe sequentialfragmentation

MH� ! b2! a2! m=z 86 �1�

where m/z 86 is the internal immonium ion derived fromleucine. Similarly, the results of Figs 2 or 5 providesubstantial evidence that, on the present experimentaltimescale, protonatedH-Leu-Gly-Gly-OH fragments di-rectly to a1, theleucineimmoniumion; in addition theb2 ionfragments directly to the samea1 ion. Both pathways toimmonium ions have been elucidated previously.36 The

fragmentation reactions of protonated H-Gly-Gly-Leu-OH(Fig. 7) aremorecomplexwith severalinitial products, thetwo majorbeingthey1@ andy2@ ionsincorporatingthemorebasic C-terminal leucine. At higher collision energies theincreasein the abundance of y1@ as the y2@ abundancedecreasesstronglysuggeststhatthey2@ ion is fragmentingtoform y1@.

Cone-voltageCID experimentswerealso carriedoutwithpreparation of the protonated peptides by atmosphericpressure chemical ionization (APCI). The breakdowngraphs for protonatedH-Leu-Gly-Gly-OH andH-Gly-Leu-Gly-OH (Figs 8 and 9) can be comparedwith the similarplots (Figs 5 and 6) obtained following electrosprayionization. In general, the agreement of the data issatisfactory although the b2 ion is formed in greaterabundance in the fragmentation of protonatedH-Leu-Gly-Gly-OH producedby APCI thanfor thesame ion produced

Figure 10. Breakdowngraphfor protonatedLeu-enkephalin.Quadrupolecell CID. Datafrom Ref. 37



by ESI. The reasonsfor this differenceare not clear.Themost striking observationis thattheonsetfor fragmentationoccurs at distinctly lower cone voltages in the APCIexperiments than in the ESI experiments. Although thisdifferencemay arise, in part, from slight differences infocusing voltages,29,30 it is most likely that other effectsplayamoreimportantrole.Onemightexpect theMH� ionsproduced by APCI to havea higher initial internal energysince not only wasthe APCI probeoperated at 370°C butalso gas-phaseprotonation may impart greater internalenergy to the MH� ions produced, althoughit is not clearhow muchof this excess energy would be retainedby theions by the time they have reached the cone/skimmerregion. In ESI it is likely that the initially produced MH�

ionsaremorehighly solvatedthanthose producedby APCIand may require one or more low-energy collisions toproduce the free MH� ions, although no specificallysolvated MH� ions were detected at low cone voltages.Both theseeffectswould be expectedto result in a loweronsetfor fragmentation in APCI experiments than in ESIexperiments. It also should be noted that the MH� ionsproduced by FAB ionization in the quadrupole cell CIDexperiments are likely to have higher initial internalenergies than thoseproducedby ESI or APCI. The maineffect of this initial internal energy will be to lower thecollision energy at which fragmentation is first observedratherthanto altertheshapeof thebreakdowngraphs.Thus,

it is considered that the comparisonof Figs 2–4 remainsvalid despite the different methods of ionization.

As the fragmentationpathways become morenumerousandcomplex, the informationwhich canbe obtained fromsimple breakdown graphs becomesmore limited. This isillustrated to some extent by the results obtained forprotonatedH-Gly-Gly-Leu-OH (Fig. 7) andis shown moreclearly by the results obtained for protonated Leu-enkephalin;Fig. 10 presentsthebreakdown graphobtainedby quadrupole cell CID while Fig. 11 presentsthe resultsobtainedby cone-voltageCID. Onenotes the considerablyhigher onsetvoltagefor fragmentation (ca. 36V, Fig. 11)for protonated Leu-enkephalin compared with the onsetvoltages of ca. 20V for the tripeptides. Although thebreakdown graphsof Figs 10 and 11 are in reasonableagreement, theamountof detailedinformationobtainableasto reactionpathwaysis ratherlimit ed.The plotsdo indicatea major reactionpathway asbeing

MH� ! b4! a4! b3 �2�

but the detailed pathways to more minor ions remainsunclear.MoresophisticatedMSn experiments,suchasthosecarriedout by AlexanderandBoyd38 andby Chenget al.39

on protonated Leu-enkephalin, are needed to map out thepathways in detail. For suchcomplex systems the energyevolution of the fragmentation reactions can be revealedqualitatively by examining the CID spectraobtained atdifferent cone voltages. This is illustrated by the CIDspectraof protonatedLeu-enkephalin atconevoltagesof 46,55 and67V, asreproducedin Fig. 12.

CONCLUSIONS

The presentstudy has shown that energy-resolvedmassspectraobtained by cone-voltageCID using anatmosphericpressuresource coupled to a single quadrupole massspectrometer can be used to establish breakdown graphssimilar to those obtainedby variable low-energyCID in aquadrupole collision cell. Thesebreakdown graphs estab-lish, at leastin a qualitative sense,the energy evolution ofthe fragmentation of the speciesof interest. Cone-voltageCID studieshavebeenusedby Paradisi andco-workers40 toestablishbreakdown graphs similar to thosereported here;however, no comparisonwith quadrupolecell CID studieswasmade.

A number of limitations to the cone-voltage CIDapproachmustbe noted. Theserelate primarily to the factthatthereis nomassselectionof theparention asthereis inquadrupole cell CID experiments.Thus, the cone-voltageCID studiesareessentiallylimit edto situationswhere thereis onedominant ion formedin the ionization process;suchwasthe casein the present study. In addition, therecanbeinterferences from backgroundions; in our experience thisis more seriousin APCI studiesthanin ESI studies.In thelatter thenumber of ions formedfrom thematrix is limitedandtheir identity remainsconstantwith the resultthat theyusually can be ignored when manipulating the data.However, backgroundionscould beaproblemif acompleteunknown were to be studied and if the samplesize (orconcentration) is limit ed. In the present study we havesetthedatasystemthresholdto ignoreion signalslessthan2 %of the basepeak.As shown by the spectrain Fig. 12, mostbackgroundpeakswereeliminated by thisapproach(signalsat m/z149and88 appear to bebackgroundpeaks). Clearly,

Figure 11. Breakdowngraph for protonatedLeu-enkephalin.Cone-voltageCID andelectrosprayionization.



with anunknownonerunstherisk of omittingsomerelevantpeaksby this approach.

On the positive side cone-voltage CID doesprovide arelatively simple approachto obtainingbreakdown graphswhich are particularly useful in unravelling fragmentationpathwaysfor smaller species.Further examplesof this usewill be presentedin a future publication on the fragmenta-tion of alkylphenyl ammonium ions (A.G. Harrison, inpreparation). In our experience higher effective internalenergies can be attained in the multiple-collision cone-voltage CID experiments than in quadrupole cell CID

experiments,thus enablingthe fragmentation pathways tobe followed further along the degradation chain. Inprinciple, similar resultsmight be attained in quadrupolecell CID studies by operating at higher collision gaspressuresand higher collision energies; however, at leaston our BEqQinstrument,ion transmissiondecreases ratherdrastically under suchconditions.

Acknowledgements

The financial support of the Natural Sciencesand Engineering

Figure 12. CID massspectraof protonatedLeu-enkephalinat threeconevoltages.Electrosprayionization.



ResearchCouncil (Canada)is acknowledged.The gift of the VGPlatform to the Departmentof Chemistryby MicromassCanadaisgreatly appreciated,as is the assistanceof Dr Alex B. Young inkeepingthe instrumentoperational.

REFERENCES

1. K. Levsen,FundamentalAspectsof MassSpectrometry, VerlagChemie,New York (1978).

2. F. W. McLafferty andF. Turecek, Interpretationof MassSpectra,4th Edn,University ScienceBooks,Mill Valley, CA (1993).

3. J. Donnacher,Org. MassSpectrom.19, 253 (1984).4. T. Baer,Adv.Chem.Phys.64, 111 (1986).5. T. Baer, J. Booze and K. M. Weitzel, in VacuumUltraviolet

Photoionizationof MoleculesandClusters, C. Y. Ng (Ed.),WorldScientific,Singapore,p. 259 (1991).

6. E. Lindholm, in Ion-Molecule Reactionsin the Gas Phase, P.Ausloos (Ed.), American Chemical Society, Washington,p. 1(1966).

7. E. Lindholm, in Ion-Molecule Reactions,Vol. 2, J. L. Franklin(Ed.), PlenumPress,N.Y., p. 457 (1972).

8. A. G. Harrison, in Ionic Processesin the Gas Phase, M. A.AlmosterFerreira(Ed.), Reidel,Dordrecht,p. 23 (1984).

9. F.M. HarrisandJ.H. Beynon,in GasPhaseIon Chemistry,Vol. 3,M. T. Bowers(Ed.), AcademicPress,N.Y., p. 100 (1984).

10. R. C. Dunbar,in GasPhaseIon Chemistry,Vol. 3, M. T. Bowers(Ed.), AcademicPress,N.Y., p. 130 (1984).

11. S. A. McLuckey, G. L. Glish and R. G. Cooks, Int. J. MassSpectrom.Ion Phys.39, 219 (1981).

12. D. D. FetterolfandR.A. Yost,Int. J.MassSpectrom.Ion Phys.44,37 (1982).

13. S.A. McLuckeyandR. G. Cooks,in TandemMassSpectrometry,F. W. McLafferty (Ed.), Wiley, New York, p. 303 (1983).

14. D. J. Douglas,J. Phys.Chem.86, 185 (1982).15. P.H. DawsonandW.-F. Sun,Int. J. MassSpectrom.Ion Phys.44,

51 (1982).16. P. H. Dawson,Int. J. MassSpectrom.Ion Phys.50, 287 (1983).17. S. Nacson and A. G. Harrison, Int. J. Mass Spectrom.Ion

Processes63, 325 (1985).18. J. J. Zwinselman,S. Nacsonand A. G. Harrison, Int. J. Mass

Spectrom.Ion Processes67, 93 (1985).19. K. L. Busch,G. L. GlishandS.A. McLuckey,MassSpectrometry/

MassSpectrometry.TechniquesandApplicationsof TandemMassSpectrometry, VCH Publishers,New York (1988).

20. C. M. Whitehouse,R. N. Dreyer,M. Yamashitaand J. B. Fenn,Anal. Chem.57, 675 (1985).

21. R. B. Cole(Ed.),ElectrosprayMassSpectrometry:Fundamentals,Instrumentationand Applications, Wiley and Sons,New York(1997).

22. J. A. Loo, H. R. UdsethandR. D. Smith, Rapid Commun.MassSpectrom.2, 207 (1988).

23. R. D. Smith,J. A. Loo, C. J. Barinaga,C. G. EdmondsandH. R.Udseth,J. Am.Soc.MassSpectrom.1, 53 (1990).

24. V. Katta,S. K. ChowdhuryandB. T. Chait,Anal. Chem.63, 174(1991).

25. M. J.Huddleston,M. F. BeanandS.A. Carr,Anal.Chem.65, 877(1993).

26. A. P. Bruins, in ElectrosprayMassSpectrometry:Fundamentals,Instrumentationand Applications, R. B. Cole (Ed.), Wiley andSons,New York, Chap.3 (1997).

27. R. D. VoyksnerandT. Pack,Rapid Commun.MassSpectrom.5,263 (1991).

28. C. L. Hunter,A. G. Marek and D. J. Douglas,Biochemistry36,1018(1997).

29. C. ColletteandE. DePauw,Rapid Commun.MassSpectrom.12,165 (1998).

30. C. Collette,L. Drahos,E. DePauwandK. Vekey,RapidCommun.MassSpectrom.12, 1673(1998).

31. P. J. R. Sjobergand K. E. Markides,J. MassSpectrom.33, 872(1998).

32. A. G. Harrison,R. S. Mercer,E. J. Reiner,A. B. Young, R. K.Boyd, R. E. March and C. J. Porter, Int. J. MassSpectrom.IonProcesses74, 13 (1986).

33. T. Yalcin, C. Khouw, I. G. Csizmadia,M. R. PetersonandA. G.Harrison,J. Am.Soc.MassSpectrom.6, 1165(1995).

34. P. Roepstorffand J. Fohlman,Biomed.MassSpectrom.11, 601(1984).

35. K. Biemann,Biomed.Environ.MassSpectrom.16, 99 (1988).36. K. Ambihapathy,T. Yalcin, H.-W. LeungandA. G. Harrison,J.

MassSpectrom.32, 209 (1997).37. T. Yalcin, I. G. Csizmadia,M. R. PetersonandA. G. Harrison,J.

Am.Soc.MassSpectrom.7, 233 (1996).38. A. J. Alexander and R. K. Boyd, Int. J. Mass Spectrom.Ion

Processes90, 211 (1989).39. X. Cheng,Z. Wu, C. Fenselau,M. IshiharaandB. D. Musselman,

J. Am.Soc.MassSpectrom.6, 175 (1995).40. A. Domo, C. Paradisiand G. Scorrano,Rapid Commun.Mass

Spectrom.11, 1687(1997).



Documents

Energy-resolved mass spectrometry: a comparison of quadrupole cell and cone-voltage collision-induced dissociation