Characteristics of a volcano field ion quadrupole mass spectrometer

Characteristics of a volcano field ion quadrupole mass spectrometer

Intemationai Jouma! of Muss Spectrometry cmd Ion Phyaiy Q Elsevier Scientific Publishing _ Company, , Am&sdam _ - - 25 (1977) Printed -- ...

2MB Sizes 0 Downloads 125 Views

Intemationai Jouma! of Muss Spectrometry cmd Ion Phyaiy Q Elsevier

Scientific

Publishing

_

Company,

,

Am&sdam

_

-

-

25 (1977)

Printed

--

_

in Tbe

c

183-198 ~I~%@& Netb&i&&&~~~~

_ _

__

_

Linus Pouting Institute of Science a& Medicine. Menlo Park. California 94025

_ =ssg$_ _ =C&s

(U.S.A.)

CHilRLESA.SPINDT Physicul (U.S.A.) (First

Ekctrcmics

received

Croup.

1 February

Stunford Research 1977:

in final form

Institute, 28 March

Menlo

Pmk.

California 94025

1977)

A33sTRAcr

A field ionization source ahaped in the form of a mi croecopic vokano is coupled with a quadrupole mass spectrometer. and the perfo rmance characteristic9 of the 6yxtem are measured. The vokano source structure, consisting of a ZO-JUII diameter rim where fidd ionization and a lO+rn diameter throat for tbe passage of sample material. permit3 the use of high pressure sample gases and yields b&b effAency ionization with excellent ion beam optics. Both water and toluene are used as sample gases at source presures ranging up to the room temperature vapor presures of these materials (20 torr and 30 to?r for water and toluene v ). and 1-butene bss been used up to pwsaurss of 140 torr. The characteristic l&e background signal produced at low m.atxeswiwntbe field ion source is positioned at the entrance of the quadrupole is shown to -be due to hi& energy electrons produced by positive ions colliding with the field ion source Structure. This background is effectively elimina~~witbout off3etting tbe field ion source from the quadrupoIe cente!rline.by t&e use of magnetic and ion optical techniques. Energy analyses are performed on H30*, C& (T-butene) and C&H-s (tuluene) ion species by tbe retarding potential me&xi. The energy fulLwidth at baWheight of these specieaismeacured at about 1.9.1.8 and 1.9 eV respectively.

INTRODUCTION

The use of a quadrupole mass spectrometer (QMS) coupled with a field desorption (FD) ion source [1,2] and field ionization (FI) source [3] have recently been reported. The QMS, with its rapid scann.in~~ capab=tys -licity of operation and resolution c+rol fl&bilitJr is an at&active ent to couple with m-type sources. Since the QMS mquires low energy ions for proper analysis (typically 10-30 ,eV) and F’I produces _ w-energe_tic ions (@pica&r lSUO-10 000 eV), good ion optical conbeam losses d-wing the deceleration ditions are reqwred -to_ prevent exceske __ [4] have linear apdyaons. of an aqay o,f_posts apact 26

Fig-l_ AeannSngdectrw

nGmgr&xofavolcano

field

ionization

souse_

The

diameter

ofthewAcanohokis2Opm.

~aparto~eranareaofaboutlmm~,alsosuffersfromIargecurrentlosses during deceleration of the sampleionsintothequadrupole.The FIsource with the bes&ion-opticaI properties is a single point;however.its limited signal for many practical ionizingregion cannotproduceanadequate applications_ The geometrp of a volcano-type FI source [S,9] offers a useful compromiseforQMSappIications.Thesource(seeFig.1)bastheshapeofa microscopicvoIcano~a20_Crmdiame~rimwfiereFIcrccurs,Thesmnll sizeoftfieio~~gmesthesourceexcellention-optical~~ permitting good focusingin bothx- andy-directions.Since the sample materialisintroducedclosetotheio~grimthrougfithecentralvolcano boIe, efficientionization of the sample resuIts_Also, the low gasconducoperationpossibIe. t2nceofthevoIcano~~icemakes~samplepressure VOLCANO

SOURCE

The ~oIcan0 FI sohrce is constructed using microfabIicationtechnoIogy [lO,lll ~SGCI on p&ysicaI vapor deposition and photoIithography- The sourceusedinthese ~entsismadeofcopperandiseppicaily75lun ~witbaBdCrm-diamet;errimaad120_Crm~base.The~~of ~of~e~edgeisof~eorderof0.l~Inaddition,~~rim surfacehasarough Euzb&ructuretbattendstoeaturnce thtifieldionization process_The vokano &ucture ismountedon-aM&inSeterpedesW containingal-mm~diamekrhakforth~~ bfs@MG_e(__j+*:?: and 3)- Aholeofa1001Lines-;cm"-~~gdd~~~~

Fig- 2_ A sa-~ electron micrograph of a volcano field ionization soarce mounted on its support pedestal by means of a press-fitted cap ring. The diameter of the cap ring is 3mLm

%pxn above thevolcano rim(see Fig_ 4), Anegatiegridpotentialof between 1000 and 4000 V relative to the volcano willthen field-ionize samplegasesintroducedthro~~evolcanohole_ ,-FOCUS

I COU?dTER-EI_EmHEATER _ CARTRIDGE

I

,

‘\

ELECTROOE

F-i-i-4_ A miaummpe photogmpb cm-l coun~nickel~pasitioned ringincenterof gridhole).

of arolcanoGeldioaizationsourawitha1CKUine &out 25pmsbcweti1erolcsnorim

(bright

Thedesigncon~tbehiPdthevolcanoshapeistointroducethesample matenalascloseas~~totbefieldio~~onofthesouroe,IftEre samplemoleculespcrsssufficientlyclosetothe~fieldrimoftbevolcano, ffieywillbedeflectedintotheximbythe fo'rceresulting fromtheinteraction oftbefZeld-induceddipolemomentandtheelecMcEeldgradient 1121. ionizationefficiencyandbeam Theanticiptedeffecbwouldbetoincrease inkmsity.Inaddition,therestrictedprrssaseway ofthevolcanothroat,wfrirh narrows downtoabout lOjrminsidediameterat4Opmbelonthevolcnmo rirll,resul~inreduceiigztsloadingofthesystemandpermitshi&k8ource -- _ pressureoperation, Gfetime studksofthevolcano~~havenoiyetbeen~~. However,tk!ae8oul.ce8 havebeenusedforover6Oh,underawkk~ ofopemtingcoLuiitions,wzthnoobaerPahle to~ev~?riowhL~so~ presuresdonotappearr' _- J atrwtumevwswheaahi& pOteatial--~~COrmteielectnjdedpoacimO~when

asomceo~occras, hoRow anode~(~m~),~m-jn&&

ashiftis- pIlm!miinnopentian-~-sLr‘to -. _-

-_ ., -_

-_-

.__~ y._:=_ -._--_I ST.

regionnoto~~~~volcanoorifice,however,canbed~~e totbevolcano(asisthe~wiehother~ofFIso~). FigureSshowsatypicalplotoftotalionbeam~t~potentirnldifference betweentbecounter#ectrodeandvolcano.Thebeamcuuentwas obtainedfromal-butenesampleatO.ltorrpressure andwasmeasuredona 3cmdiametercollectorplatepositionedabout4cmdownstreamfromtbe _- _ Volvo_ The sem.ilogarM~Gc plot shows the m c_initialsteep straight alope followed by a more gxwlual slope as the‘electricfield is betweenl.Gand 2.5 in~.Thebreakinthetwoslopesnormallyoccurs kVdependingupontheseparationbe~~~evolcaaoand~ountereIectrode_aswellasthesharpnessandconditionofthevo~orim.Duringall the me asurem&ts, the source is heated ti 250°C by means of aheater locatedinthecountereIectrodesupport(seeFQ.3). Activation~[4] of the rim edge has been accomplished by introducing toluene at about 1-torrpressure l&roughthevolcanoholewhilemaintahG~g a volcano-counterelecte potential difference of GL 3500 V_ For the measurementspresen tedhere,however,noa~ptswereIlnadetoactivate tfiesourcealtbougfisomeactivationmayhaveoccurredduringtbeexperimen&

tional k

scale

c&s&&&

dlagrani &king the anaIm construction details. l%e hnal~ker qf s&$jd&& steel.‘p_nrh analyzerlens element is positioned and

~~~'by~a-setoffoar4mmd;nmetersapphireballs(uOt shownlinPig,)lolcatedoneach~~e~~-Theanalgzeroperates byfocnsingtbeiolus~m~vol~osourceontbel-mm diameternDerture -2and3andaset of four Iocabediilekckxi~3bymeans~fdectrbd~~, defkctorslocatedin &ectrode_2_The~divergingion~thatemergesfrorp the~qe&ur&isfocusedintoauaxiaUyparalielbeambythe 1ensingefGct ofekctrodes3 axid4-Aset-of~dsfollowingektrode-4can t;henb4iusedto+.dfo~_retardingpotedial energyanalysisontheionbeamThe grids ked in the-analyzer are copper with 30 lines cm-' and9056 is not gold-plated or maintained at an tEUqsuq, s+e-theanalyzer surfaazpotentialoftberetarding elevatedtempemture,v8riationsinco&xt gridscausedbyabsorbedgasesareiarge and noabsoluteenergymeasurementswereattempted. Aschematicvievroftbeenergyanalpzerisshownin Fig.?alongwith sometgpicaIo~~gvoltages_Figures8,9andlOshowenergyanalysis curvy for H,O*,C,H,CIT,(toluene) and Cfln (l-butene).Curve Ain these ~wasproducedbypIottingtherebnLingpotentialgridvoltageasa function of beam txurent(swit& positions a,in Fig. 7),CurveBisthe dSferentiaIofCurve A andisproducedbysuperimposinga0.2-Vsquare waveontherekudingpotential@idvoltageandamplifyinganddetecting

Fig_6. SC& ontheqndnapokmru

Fig. 7_ Schematic diagram of v0lcan0 source mass Spectrometer operated for ion eaergp ~~.Atypicalretardingpotentialcurve~obtainedwithsPlri~esinP06iti0ila(see Curves A of Pigs- 8. 9 and lO).whiIe the differe~t.ial of this curve isobtainfdwitb switchesinposition b(seeCurvesB,Figs_8,9andlO).

-10

0

-5

RETRRDlNG

10

5

GRID FalENnAL

bob)

Fig. 9_ Energy Andy of RsO+ ions obtained by manual Ming of the rehrding tid po~ntislis400oV.~B~thcdiEferentialofCweA voltage. The coonterelectmde of2OOcpssuperim_pc6edonthe retwding obtainedusing aOT2-V amplitudesquarewave

only the ac signal component by means of a Iqckin amp_Lifief (switch p$$$ tionbinFig~?).DuringalIenergyanalysismeasurem en&, the QMS is tuned to &eparticuIarion species beinganaIyz& and set at a+ut unit znass resolutiOIL

The narrow energy spread of 2-3, 2.3 and 2.2 eV m eSu& for-C&H&H;, HsO* and Gas demonstrates the exceQention-opti&lcapabilities of the vokano source. These energy spreads should be reduced by 0.4 eV to compensate for the resolution loss caused by the 0.2-V source square-wave analyz&g voltage. The more correct values sbouId therefore be l-9. i-9 and l-8 eV_ respectively_ These spreads appear comparable to those obtained &om single point emithzs [13,14]. QUADRUPOLE

MASS

When

sources

FI

ANALYSIS

are dkctlyconnectedto

aQMSalargebackground the lower mass portion of

the spectrum is typically observed Cl.31 (see Fig. 11(a)). We found that the background signal could be reduced by an order of magnitude without dxzturbing the mass peak amplitudes by placing a small magnet near the source (Fig. 11(b)). This result indicates that the background signal is probably due to electrons generated in the source region. The larger magnetic effect on the electrons relative to the toluene ions is anticipated since the radius of curvature of a i&V ion (energy determined by the 15-V

signal that gradually increases toward

TOLUENE b)

potentiaIof‘the‘v;dlcymoso~)Esinbout27~~greatertbanIt3500eV electson(~ergp'de~~-bythe;3500_Vpotentialoftbecountereleotrodegrid), Akxond'inni;#tion ofehxtronoriginofthe backgrcitidsignal arisesfiomtpelargeeffectthat electxode-l(see Fig_ 6)hason thebackV for optimal groundlevel. Electro&l~~voltagG is normally about-270 focking oSftaizpo&ive~idfls_& else-lwol~e~~emorepasitive. thebackgroundl&r&iiicrease& andtbepositiveionsigualdecresesanddis toover appearsatabou&-100 V.l%e backgro~levelcontinuestoincrease 100 timesnormaIIevel atanelwtrode-l.voltageof150 V(seeFig. 12).This strongIenseff~on~~ba;c~~dlevelcanbeexplainedbyanincregsed renectionofpositmeionsbylens-lbacktowardthecounterelec~e~das theelectrode-lvol~becomesmore~~.These~ectedionswill~d

-_g_ ewith 35l5-eVener&-on_thesidefacin~&i&~ strike the cotm~ trod*l_and eject e&ctsoni_ihm the cou.ntewhct+&.toward_ le-e.& aperCam of ebxtrode-1.The energeticelectrons-produced atthemcounter-

to

electrode have a transit time of less than one himdiiedthbf the period of the quadxupole high frequency field and can -the quadrupole duringthe low field portion of the RFcycle [I]_ They then produce a mukiplier signal by tdkionally ejecting positive iqns or photons when they strike the surface .opposite the quadrupole exit apertum. Also, since the coIlisions required to produce the background signal are energy dependent, one would anticipate t&at making the counkrelectrode voltage less negative would ckxease thebackground+gnal.T&iseffectis indeed observed here and elsewhere [3]. The use of the retarding potential energy analyzer shown in Fig. 6 offers a remedy for the electron-induced background. The ions or&inating from the volcano source are focused on the aperkve in electrode3 with little loss-in beam iniznsity. The energetic electrons, however, will not benefit f.rom the focusing conditions and will be attenuated by the reduced acceptance solid angle of the aperture. If, in addition, a weak deflecting magnetic field is introduced along the beam path length of the analyzer, a further reductlon in background is produced. Figu~ 13 shows the results of using the Harding potential analyzer and a weak magnetic field on the mass analysis of a

TOLUENE

canbeanimportant Theabilitpofionso~tooperateathigbpresures attribute,Most biologicalsamples, for exampIe, contain rehtiveiy large amolmtsof~_knmanycasesthiswatermustfirstbesubs&ntially

vc=a!5uovo

0 c) 8

EACKGROUMD 0 0

ooo%

=o

00%

& OV,--=v

removedbeforeamass~canbeeffectmelyperfonnedTheremo~~f, thiswatercan~difficultandmayaffect-tfienaRveoftheother_pI~ componentstobemeasured The geometry of the volcano FI source inakes it well suited tohigh pressure operation.Tbegeome~~pumpingspeedofthecentrslorificeis about lo-= cm's_'_Thus,asourcepressure of1OOtorrwiUhxreasepressure in thepresentvacuumsystem,which hasapumpingsp&dof1OOO1s-'. fromabasevalueof2 X 10'7torrto10-6torr. Figure 14 shows the effect on signaI and background ssasource pressure due to l-buteneisvariedbetwecn10-2 and140torzTwosetsof curvesareshown,utilizingcounterelectrodevoltagesof-35OOand-2500V, mqectively. The background signal (signaIlevel at mass 52.5)ofeach set tendstoillcrease gladuauywit.hpressum_ Theionsignaifor V,= -3500 V increasesindirectproportion~thel-butenesourcepressure up toca.30 torrwhere itlevelsoff.The45" slopeIineshowuinPig_14represen tsthe conditionofsignalbeing~yproportional~sourcepressure, Thesignal for V,= -2500Vtendstodropbelowthe4Ssiopeatpressums abovelo-' toxr andthesignalwasgenerally morenoisythanforV,=-35OOV. Atthehigh pressure endofthecurves, amoreiapidincrease in backg.roundsignaI occurs coupledwithadecreaseinl-butenesignal.Thepositive ion-decreases atthehighpresmues primarilybecauseofareductionin field strength at the Volvo rim causedby the formationofaconductive plasmainthehighpressure region of the vohno orifice. The nature of the conductive plasma is characterized in a form analogous to the paschen gsseousbreakdownc~e~15].Thisisdemo~~byFig.l5whichshows aplotofthe pressure atw~chthe~tobackgroundratioislverrmsthe volcano-count e potential difference_Thevaluesalongtbiscurve reflecttheconditionsfortheonset ofgaseous breakdown.Thecurve shapeis dmilnrtotbeleft-handportionofatypicalPaschenc~.Itindicatesthatif one operates with avolcano-counterelectrode potentialdifferencevalue

TOLUENE

I. 0

10

-20

30

r

1

.

50

40

*

60

70

t 90

80

nlfe Pig-16_Massscauchiaimdwi~tdaene oftohlene). WFP====

at 30 torr (room

source p-

temperature

tx?.lowthe curve_minimum, no voltage breakdown will occuzatthevolcano IegaRilessofsoalrcY=pressure sincetfieenergygathered bytheionbetween collisionsisinsufficienfto cause ionization of the molecule. However, cBargeexchangeandscattering~causedbythehigfibac~undpres_ sUre%$llWZIltWIlycauSe theionsignaltodeteriorate. E'igure16shows a~scanob~edwiththevolcanosourcepTessurein equilibrium with liquid toluene at room tempemture(m 30 torr)-The voicano+.xmterelecte potential &fference used for thisscan was 2315 V, I

I

WATER

I

I

FL-’

Figurel?isaspec&um obtained~mwatervaporattheroomtem~. equilibriumpressure ofca.20torr.Thevolcan
Thevo~oFIsourcehasdemonstrateditscapabilityofo~~effiappear feasible. cientiyatpressures upto100torr,andevenhigherpmssures capabilitpshouklincrease theeffectivenessofFIMSin ThkhighpressUre numerous applications. In particular, the inmrfacmg of -quid chromatog7aphytoFIMSshould_bemoreeffective_Also,tbedirectanalysis of vapors from volatile liquids can be performed without any sample pretreatment. me-excellent ion-optical characteristicsobtained from a voIcano~sourceenahIetheperforman ceofhighresolutionenergyanaIysisof thefinmnleions.Thesenumerous~~~~shouldrnalrethevolcanosourcse ausefuI_additiontoanaIyticaImassspectrometry.

ACKNOWLEDGEZKENTS

‘I&e autho~g w&-to express their gratitude to Arthur B. Rob&on for general experimental discusions, to Ivor Brodie for helpful discuszions on elec&ical breakdown, &o Frank Allen for superb technical asistance and to Hazel Pakka for the fahication of the volcano structures, We also wish to thank the National In&itute of HeaItb Division of General Medical Sciences for support of this work under Grant No. 21197-03.

1 2 3 4 H.D_l3e&ey.F'ieldIo&aiiorrMass ~etrg,P~~~NewYor~.1971_ 5 W. Abertb,CaI. Spindt and KaT. Rogers, 11th Symposium onEIectrons, Ion, and LaserEkamTe&_.SaaFnn&coPreu,Inc,197l,p.63lm W_Aberth,RBSpenya~dC_kSpiPdt,20thAnnaalCo~~noeonMassS~trome~andAlliedTopicr.~TeJEy197~p,269. W, Aberth.CA SpindtandRJLSperry.2~AMualConferenceon~Spec~ metrgand~Topics,SanFrandsco.1973,p.463. C-k Spindt, Z&d International Fiid-Emission Symposium, Pennsylvania State Unhersitp,A~l976. W3L Abertb,J_ WiZder.k Bauibgame and CA Spindt,?tb internationalB&u= 1976.tobepuMishedinAdvan. Spectrome~Conr~a.FIo~ce,I~.Septcmber MasSpfftmm,?. 10 C~Spindt.~ApplPbgk.39(1~)3604. 11 CA Spindt, paper in preparation 12 R Gomer,FieId Emimion andFieldIonization,HarvardUniversityPres~,Cambridge. 196X_ 13 ~_~~F_W,Rijn&enandH-D_BecLeg,Z_Na~~A.29(1974)773. 14 LV. Goldenfehi, 1.2. Korcmtyshemkyand B.G. B%chan&uk,fnt.J.MassSpectrom. 10nPhys.,13(1974)2!97. 15 ~.'Dm~andIP~Penning,Re~.MLdocLI~,12(1940)87.