Influence of Instrumental Parameters on the Kinetic Energyof Ions and Plasma Temperature for a Hexapole Collision/Reaction-Cell-Based Inductively Coupled Plasma QuadrupoleMass Spectrometer

GEORGES FAVRE,* RENE BRENNETOT,* FREDERIC CHARTIER, andJEANINE TORTAJADACommissariat a l’Energie Atomique, Saclay, DEN/DPC/SECR/LANIE, 91191 Gif-sur-Yvette Cedex, France (G.F., R.B., F.C.); Laboratoire

Analyse et Modelisation pour la Biologie et l’Environnement, Universite d’Evry-Val d’Essonne, CNRS UMR 8587, Evry, France (J.T.)

Inductively coupled plasma mass spectrometry (ICP-MS) is widely used in

inorganic analytical chemistry for element and/or isotope ratio measure-

ments. The presence of interferences, which is one of the main limitations

of this method, has been addressed in recent years with the introduction of

collision/reaction cell devices on ICP-MS apparatus. The study of ion–

molecule reactions in the gas phase then became of great importance for

the development of new analytical strategies. Knowing the kinetic energy

and the electronic states of the ions prior to their entrance into the cell,

i.e., just before they react, thereby constitutes crucial information for the

interpretation of the observed reactivities. Such studies on an ICP-MS

commonly used for routine analyses require the determination of the

influence of different instrumental parameters on the energy of the ions

and on the plasma temperature from where ions are sampled. The kinetic

energy of ions prior to their entrance into the cell has been connected to

the voltage applied to the hexapole according to a linear relationship

determined from measurements of ion energy losses due to collisions with

neutral gas molecules. The effects of the plasma forward power, sampling

depth, and the addition of a torch shield to the ICP source were then

examined. A decrease of the plasma potential due to the torch shielding,

already mentioned in the literature, has been quantified in this study at

about 3 V.

Index Headings: Inductively coupled plasma mass spectrometry; ICP-MS;

Plasma temperature; Potential plasma; Kinetic energy; Collision cell;

Reaction cell; Torch shield; Instrumental parameters.

INTRODUCTION

Inductively coupled plasma mass spectrometry (ICP-MS) isone of the most powerful inorganic mass spectrometrictechniques for applications in environmental, industrial, andnuclear fields. However, the presence of mass interferences canhamper the measurement of an isotope at a given m/z ratio. Achromatographic step before analysis is classically used toseparate interfering elements and to lead to mono-elementalpurified fractions. However, this method presents drawbacks inthe nuclear field, where isotopes are quantified in irradiatedsamples, with the increase of (1) the risk of naturalcontamination, (2) handling time on radioactive solutions, and(3) production of radioactive wastes. Alternative analyticalapproaches can be found to overcome this interference problemwhile avoiding such supplementary steps and thus lowering theirradiation of operators. The cold plasma technique1–4 isencountered in the case of argide-based interfering molecularions and high resolution devices5,6 for separating ions differingin mass by a fraction of a mass unit. However, the required mass

resolution for solving isobaric interference in the nuclear field(m/Dm . 30 000) stays beyond the actual reach of the double-focusing sector field inductively coupled plasma mass spec-trometers (SF-ICP-MS) and the use of cold plasma leads to anenhancement of matrix-derived polyatomic ions, in addition to adecrease in sensitivity, which is problematic when very lowconcentrations of radionuclides (’ ng�mL�1) are encountered.

Thus, the introduction of collision/reaction cells7–13 oncommercial ICP-MS instruments ten years ago provided aninteresting way to solve various types of interferences bykeeping a good sensitivity. Resolution of some of them may beachieved by means of well-chosen gases and definedinstrumental parameters. Dissociation of polyatomic speciesand selective reaction of the gas with the interfering element orthe analyte are the main mechanisms leading to the resolutionof interference.7,8,12,13 However, in both cases the influence ofthe ion kinetic energy on the chemical reactivity in the cell,discussed previously elsewhere,14–16 is of prime interest. Theenergy distribution of ions is furthermore narrowed in somecases to improve ion transmission17 and abundance sensitiv-ity.18 Thus, even if the accessible energy range is limited oncommercial ICP-MS instruments, knowing the kinetic energyof ions on such devices is useful for understanding theirreactivity and for optimizing instrumental parameters.

The kinetic energy of ions in the cell may be controlled byvarying the hexapole voltage, which acts as an accelerationpotential. In this study, the measurements of maximal energylosses of ions due to elastic collisions with neutral gasmolecules have been carried out for different values of theacceleration potential. An explicit link between the voltageapplied to the collision/reaction cell and the initial kineticenergy of ions, Ei, has been determined for an X7 Series I ICP-MS Q (Thermo-Fisher Scientific) spectrometer and its evolu-tion, the X7 Series II spectrometer. A characteristic instrumentalfactor, revealing the influence on the energy of ions of the ‘‘Pi’’lens, added on the Series II just after the skimming cone, ishighlighted. The effect of adding a grounded shield platebetween the ICP torch and the load coil, known for loweringand homogenizing the energy of ions through the diminishmentof the plasma potential,19,20 has also been quantified and itsinfluence on the ICP temperature will be discussed.

EXPERIMENTAL

Instrumentation and Optimization. Two quadrupoleThermo Elemental X7 inductively coupled plasma massspectrometers (Thermo-Fisher Scientific, Winsford, UK),

Received 5 March 2008; accepted 28 November 2008.* Authors to whom correspondence should be sent. E-mail: georges.favre@cea.fr, rene.brennetot@cea.fr.

Volume 63, Number 2, 2009 APPLIED SPECTROSCOPY 2070003-7028/09/6302-0207$2.00/0

� 2009 Society for Applied Spectroscopy

equipped with collision cell technology (CCT), were used forthis study. While the first one, an X7 Series I, has beenmodified in order to work with radioactive materials, thesecond one, an X7 Series II, was kept in its classicalconfiguration.

Concerning the nuclearized ICP-MS apparatus, a glove box(JACOMEX, Dagneux, France), in which are located thesample introduction system, the plasma ion source, and theinterface, has been installed. An autosampler CETAX ASX260, also situated in the glove box, minimizes humanmanipulation of radioactive samples. Air-tightness is ensuredby a modified interface, and the pumping system was alsoadapted: a new pump was added (Edwards 65 m3�h�1) to thestandard configuration to ensure a better vacuum in theexpansion chamber and to compensate the head losses due tothe addition of filters. Two independent cooling systems wereused, the first one for the spray chamber (maintained at 3 8C byPeltier effect), and the second one for the interface and the RF(radio frequency) coil.

For both instruments the Ar plasma, in which the nebulizedsample is introduced, is provided by the application of a 1400 Wradio frequency potential (RF, 27.12 MHz). The ion lenssettings, nebulizer gas-flow rate, and torch position of theinstrument were tuned daily for the apparatus in standard mode(i.e., without any gas in the cell) in order to maximize sensitivityat m/z 115 (In) and 238 (U) for a multi-element tuning solution(SPEX) during short-term stability tests. Typical sensitivities, instandard mode, were better than 350 000 and 450 000counts�s�1�ng�1�mL, respectively, for In and U in the case ofthe nuclearized X7 Series I and better than 180 000 and 200 000counts�s�1�ng�1�mL in the case of the X7 Series II. Thedifference is due to the supplementary pump installed at theinterface on the nuclearized ICP-MS (S-option), which allows abetter vacuum to be obtained. The ICP-MS operating conditionsused for this study are given in Table I. However, it can beoutlined from these data that the most outstanding differencesbetween the parameters of the two instruments appear in the

range of VH and VQ voltage accessible values that are appliedrespectively to the cell and to the quadrupole used for theseparation according to the m/z ratio (Fig. 1).

Materials and Reagents. Experiments have been performedwith a solution of 10 ng�mL�1 Zr, In, Sm, and U in 0.2 mol�L�1

nitric acid. This solution was obtained from 1 lg�mL�1 stocksolutions prepared from 1000 lg�mL�1 mono-elementalstandard solutions (SPEX) diluted in 0.2 mol�L�1 nitric acid.The 0.2 mol�L�1 nitric acid solution was prepared by diluting a65% Normatom Prolabo solution with milliQ (Millipore) water(18.2 MX�cm resistivity). High purity neon (99.9999% purity)was obtained from Messer France and was used as the collisiongas for the experiments in the pressurized mode.

Measurement Procedure. It should be emphasized that thedevices used in the present work are designed for elemental orisotope ratio analyses and not for fundamental studies. Ionsenter the cell with a given kinetic energy, Ei, determined by thevoltage, VH, applied to the hexapole. They will lose energywhen a gas (neon in our case) is injected into the cell becauseof collisions with the gas molecules. By assuming an elasticcollisional process of ions of mass Mi and kinetic energy E

ðkÞi

with stagnant (ENe¼ 0) neon atoms of mass MNe, the energy ofions after the collision k is given by:8

Eðkþ1Þi ¼ E

ðkÞi

M2i þM2

Ne

ðMi þMNeÞ2

" #ð1Þ

The more the gas flow rate in the cell is raised, the more ionsundergo collisions and also, the higher the energy losses are.Neon was chosen because its atomic mass is sufficiently highto efficiently induce temporal homogenization without causinglarge losses from collisional scattering. Some multiple collectordevices (MC ICP-MS) mainly use the collision cell forcollisional focusing of ions as a replacement for theelectrostatic sector analyzer classically encountered in dou-ble-focusing apparatus. In this study the cell is instead used as areactive area, and knowing the kinetic energy of ions isinteresting for looking at their reactivity with gases.

Determinations of such ion energy losses were performed bytaking advantage of the kinetic energy discrimination (KED)effect. Its principle, described elsewhere in greater detail,7,14,15

consists of using a ‘‘potential barrier’’ between the hexapolecollision cell and the quadrupole mass analyzer (Fig. 1) toexclude the slowest ions. This potential barrier corresponds tothe difference, VQ � VH, applied between the cell and theanalyzer and can be modified for a given VH value by changingthe voltage VQ applied to the quadrupole. Thus, when thepotential barrier VQ � VH is increased, more ions are stopped

TABLE I. Typical ICP-MS operating conditions.

Solution uptake rate 1 mL�min�1

Nebulizer Quartz concentricSpray chamber Quartz chamber with impact beadSpray chamber temperature 3 8C

ICP parameters

Plasma forward power RF 1400 WReflected power ,2 WNebulizer gas flow rate 0.85 L�min�1

Auxiliary gas flow rate 0.8 L�min�1

Cooling gas flow rate 14 L�min�1

x, y position Optimized daily to maximize signal forInþ and Uþ

z position 8 mma

13 mmb

Cell geometry HexapoleCollision gas (Ne) flow rate 0 to 3 mL�min�1

VExtraction �600 V

VH (hexapole bias) [�10; þ10 V]a

[�20; þ20 V]b

VQ (quadrupole bias) [�10; þ10 V]a

[�20; þ12 V]b

a X7 Series I.b X7 Series II.

FIG. 1. A schematic diagram of the ICP-MS X7 (Thermo-Fisher Scientific)with three characteristic parameters related to ion energy (VP: plasma potential,VH: hexapole bias voltage and VQ: quadrupole bias voltage).

208 Volume 63, Number 2, 2009

since only the most energetic can get through it. As a result, themeasured intensity of the ion signal decreases. With apressurized cell, a lower potential barrier corresponding to alower VQ value will suffice to stop a maximum of ions sincethey would have lost energy through collisions with the gasmolecules. Stopping-curve measurements (Fig. 2) thus ob-tained give access to the ion energy loss associated with agiven gas flow rate and VH value. The energy of ions canindeed be defined as the quadrupole potential VQ required toreduce the ion intensity by one order of magnitude.14 Finallythe difference between the VQ values obtained for the two casesof a vented cell and a neon pressurized cell corresponds to theamount of kinetic energy lost by the ions through collisionswith neon molecules.

The intensities of 90Zr, 115In, 147Sm, and 238U have beenmeasured to cover as broad a mass range as possible. Elementswhose molecular weights are below 90 u.a. have not been takeninto account in the experiments because of a lack of sensibilitywhen applying the KED method. A data extrapolation to thelow molecular masses could offset that experimental limitation.Data were collected for a wide range of VH and VQ settings atNe flow rates of 0, 0.5, 2.5, and 3 mL�min�1. The sequence ofVH values was taken in random order to minimize any drift, aswere the VQ values at each VH setting. Ion optic parameterswere optimized at each VH change to maximize 115Insensitivity. The extraction voltage applied on the skimmercone, VExtraction, was held constant at�600 V in order to avoidinterference with the energy measurements, while the forwardpower and the nebulizer gas-flow rate were initially set at 1400W and 0.85 L�min�1, respectively, i.e., the values classicallyused for analyses, since previous papers have already shownthe influence of these two parameters on ion energy.4,21–23

RESULTS AND DISCUSSION

Collision Conditions. The gas density, n (cm�3), isestimated from the volume flow rate of gas, F (cm3�s�1),delivered by the mass flow controller since no pressure gaugewas installed on the collision cell. The mass flow controller iscalibrated for a delivery pressure, P0, of 1 bar. Thus, the gasflow rate into the cell, Gin (molecules/s), is roughly estimatedaccording to the following relationship:16

Gin ¼ NaP0F=RT ¼ 2:34 3 1019F ð2Þ

where Na is Avogadro’s number (6.022 3 1023 mol�1), R is theideal gas constant (8.314 J�mol�1�K�1), and T is the absolutetemperature of the collision cell (310 K). The gas leaves thecell through entrance and exit apertures at a flow rate of Gout.Both apertures can be considered as circles of diameter D ’0.2 cm. By assuming that the gas flows out effusively throughthese apertures,24 Gout can be expressed as:

Gout ¼ 0:25nuA ð3Þ

where u is the mean gas velocity [u¼ (8kBT/pm)1/2, where kB¼1.38 3 10�23 J�K�1 is the Boltzmann’s constant and m is themolecular weight] and A ¼ 2 (pD2/4) is the total area of bothgas exits. The gas density in the cell is at a steady state, so Gin

¼ Gout and the mean free path k is given by:

k ¼ 1=ðffiffiffi2p

rnÞ ð4Þ

In Eq. 4, r is the collision cross-section for the species underconsideration. We approximate it to the kinetic cross-sectionof neon calculated in the hard sphere model rNe¼ pd2

Ne ¼ 2.53 10�15 cm2 (where dNe ¼ 2.8 A).25 The partial pressure ofNe is also calculated from n by using the ideal gas law. Forthe neon flow rates used in this study, FNe ¼ 0.5 cm3�min�1,nNe is calculated to be 2.2 3 1014 cm�3, leading to anapproximately 7 mTorr neon partial pressure (PNe) and amean free path, kNe, around 13 mm; while for FNe ¼ 2.5cm3�min�1 the following values are found: nNe ¼ 1.1 3 1015

cm�3; PNe ’ 35 mTorr; and kNe ’ 2.5 mm.The cell is about 10 cm long. Ions will therefore undergo

approximately 10/0.25 ’ 40 collisions with neon moleculeswhen they pass through the cell. However, this estimateminimizes the true number of collisions, as the effective path ofions is longer than the geometric length of the cell and the gaskinetic cross-section for Ne–Ne collisions, used for thecalculation of the mean free path of the ions, lower than thereal r values associated with elements of interest in this work(Zr, In, Sm, and U).

By applying Eq. 1 to these four ions, for which a 25 eVinitial kinetic energy E

ð0Þi is assumed, it appears that a number

of collisions of 40 leads to an efficient damping of the kinetic

FIG. 2. Stopping curves for ions (&, 90Zr; m, 115In; �, 147Sm; and ¤, 238U)measured at Ne flow rates of 0 (solid black symbols), 0.5 (dark gray symbols),2.5 (open symbols), and 3 (light gray symbols) mL�min�1. The hexapolepotential VH was set at�8 V, plasma power RF¼1400 W, and nebulizer flow¼0.85 L�min�1.

FIG. 3. Kinetic energy, Ei, of four ions (90Zrþ, 115Inþ, 147Smþ, and 238Uþ)according to the number of collisions, Ncoll, undergone with the neon atoms. A25 eV initial kinetic energy, E

ð0Þi , is assumed. It appears that the energy is

efficiently damped after 30 collisions, leading to the thermalization of the ionbeam for the highest neon flow rates used in this study (3 mL�min�1).

APPLIED SPECTROSCOPY 209

energy of the ions. Figure 3 shows that after 40 collisions thekinetic energy of the heaviest ion 238Uþ is reduced at 0.05 eV atthe exit of the cell, demonstrating the efficient thermalization ofthe ion beam for the highest neon flow rates used in this study.

Maximal Energy Loss of Ions Due to Collisions. Theacceleration potential, VH, applied to the cell, controls thekinetic energy of ions, Ei, which was evaluated through themeasurement of the maximal energy released by ions duringtheir transit in the pressurized cell. It corresponds to the energyof ions at the cell entrance since they are thermalized at its exit(3/2kBT ¼ 0.04 eV). This maximal loss is observed whenstopping curves linked to different gas-flow rates overlap asreported in Fig. 2 for 2.5 and 3 mL�min�1 Ne flow rates and aVH value set at �8 V. This overlap of two stopping curvesshows that for both gas-flow rates the energy loss remainsidentical. Stopping curves corresponding to higher flow rateswould be the same as the ones reported for 2.5 and 3mL�min�1. Similar determinations of maximal energy losseswere carried out for different VH values in order to exhibit anexplicit link between the VH acceleration voltage and thekinetic energy of ions. In the case of the nuclearized instrument�8 V, �3 V, þ2 V, and þ7 V VH settings were taken intoaccount, while a wider VH range was used for the X7 Series IIICP-MS in its classical configuration (�16 V,�8 V,þ2V, andþ10 V). Numerical results obtained with the nuclearizedapparatus through such an approach are reported in Table II.

Tanner et al.8 reported in their review on ICP-MS collision/reaction cells the relationship (Eq. 5) below, which exhibits alinear dependence of the ion energy, Ei, with the cell offsetpotential VH:

Ei ¼ Esource þ VP � VH ð5Þ

where VP is the contribution from the plasma potential at theion kinetic energy and Esource is the energy gained by ions atmass Mi through the supersonic expansion in the interfaceregion. Authors agree that the plasma offset voltage may lie

between 0 and about 20 V in practical operating systemsaccording to the load coil geometry and its groundingarrangements.26–29 By assuming that ions and neutral Aratoms acquire a common supersonic velocity through theexpansion, the Esource term can be simply formulated:

Esource ¼Mi

MAr

EAr ¼Mi

MAr

5

2kBT0

� �ð6Þ

where MAr is the mass of the bulk plasma component and EAr

is the kinetic energy gained by the Ar neutral species throughthe expansion, which is related to the gas kinetic temperatureT0 outside the sampler in the ICP.4,23

The following relationship (Eq. 7), similar to the one above,was extracted from our experimental measurements of energylosses for the four ions under consideration. The parameters aand bi, characteristic of the operating conditions and of the ionmass, are summarized in Table III.

Ei ¼ bi � aVH ð7Þ

Thus, increasing the VH potential slows down ions that enterinto the cell as shown in Fig. 4. However, a significantdifference appears with the theoretical Eq. 5 in the case of thenuclearized ICP-MS. The VH multiplying factor a was indeedfound at a aI ¼ 0.72 6 0.01 value, lower than the oneforecasted by the theory, ath ¼ 1, while an identicalexperimental protocol on the X7 Series II led to aII ¼ 0.93 60.01, in much better agreement.

Two reasons have been initially advanced to explain thisdiscrepancy of aI with the theoretical value, which can beinterpreted as a characteristic instrumental factor of theThermo-Fisher Scientific ICP-MS X7 Series I instrument. Incomparison with the instruments used in previously reportedstudies, only differences in the ionic lens system located beforethe cell or consecutive to the nuclearization of the device canindeed be propounded as possible explanations. This secondpossibility was immediately rejected by a similar complemen-tary study carried out on an ICP-MS X7 Series I in a classicalconfiguration (i.e., without any modification due to thenuclearization). The a coefficient derived from Eq. 7 wasfound at 0.71 6 0.01, consistent with previous aI values. Thisobservation confirmed the effect of the optics located betweenthe interface and the cell on the energy of the ions. It revealedmore precisely the influence of the Protective Ion extractionlens, known as the p lens, which was added on the X7 Series IIafter the skimming cone (Fig. 5) in order to modify theextraction field and ensure that the material deposited on thecone is not released and focused in the ion beam.

It appears through the comparison of Eqs. 5, 6, and 7 that the

TABLE II. Maximal energy losses.a

VH (V)

DEmax (eV)

90Zr 115In 147Sm 238U

�8 16.3 17 17.6 20.3�3 12.1 13.2 14 16.1þ2 8.7 9.8 10.7 13.1þ7 5.2 6.1 7 9.6

a Determined from stopping-curve measurements on the nuclearized X7 SeriesI for a 1400 W plasma forward power without the torch shield.

TABLE III. Calculated energy parameters (b in eV, VP in V, and T0 in K).

aa

b

VP T090Zr 115In 147Sm 238U

X7 Series I, nuclearized configurationb RF ¼ 1400 W 0.72 10.21 11.16 11.96 14.42 7.7 6 0.6 5340 6 190RF ¼ 1200 W 0.73 10.16 10.95 11.85 14.13 7.7 6 0.3 5150 6 120

X7 Series IIc Without the torch shield 0.93 9.99 10.61 11.36 12.89 8.2 6 0.3 3730 6 115With the torch shield 0.83 6.90 7.41 7.92 9.27 5.5 6 0.25 2970 6 80

a Reported values were estimated at 6 0.01.b Without a torch shield, nebulization flow rate¼ 0.85 L�min�1.c RF ¼ 1400 W, nebulization flow rate ¼ 0.85 L�min�1.

210 Volume 63, Number 2, 2009

bi parameter has a physical meaning. This experimentalparameter is actually linked to the plasma potential VP and tothe neutral plasma temperature T0 through Eq. 8:

bi ¼ Esource þ VP ¼Mi

MArEAr þ VP ¼

Mi

MAr

5

2kBT0

� �þ VP

ð8Þ

Simultaneous equations are obtained by applying theserelationships to our experimental bi data. The mean values ofVP and T0 have been determined for different cases and arereported in Table III with their associated standard deviations.It can be noticed that these parameters are in agreement withpreviously published values,4,22,23,26,28,29 giving confidence inthese results.

Effect of the Torch z-Position: Sampling Zone. Acomment can be made concerning the results obtained on thetwo devices without using the torch shield system. Theparameter T0 linked to the energy was indeed evaluated as1600 K lower in the case of the experiments on the X7 Series IIICP-MS, although the RF power was set in both cases at 1400W (Table III). The sampling zone of ions in the ICP was inreality different for the two series of experiments. Themeasurements on the nuclearized instrument have been donewith the torch in end-stop position, while those on the X7

Series II device were carried out for a backed-off z-position at13 mm (see operating conditions in Table I). Thus, the plasmaarea from which ions were extracted had a lower temperatureT0. The plasma potentials VP found in both cases are, on thecontrary, the same when the uncertainties are taken intoaccount.

A complementary study has been carried out on the X7Series I ICP-MS for five z-positions of the torch (8, 9.5, 10.5,12, and 13 mm) and a 1400 W plasma forward power. Thesevalues correspond to the distance from the top of the load coilto the sampling plate, as represented in Fig. 6. The resultsprovided additional information about the characteristics of thisinstrument and confirmed the conclusions drawn from theresults obtained on the X7 Series II ICP-MS for a torch z-position set at 13 mm. For the nebulizer gas-flow rate used inthis study, a lowering of the temperature according to a linearmodel is observed when the sampling depth is increased (Fig.7). It appears that the temperature decreases about 250 K forevery supplementary mm in the sampling depth for the [8–13mm] range used in the experiments. As mentioned above, theplasma potential VP remained stable for the five z-positions ofthe torch around the mean value VP,mean ¼ 8.7 6 0.4 V,consistent with the values reported in Table III.

Effect of Plasma Forward Power. The results shown abovewere obtained for a RF power setting at 1400 W, our usualworking condition. The link between the forward power andion energy was examined by using the same experimentalprotocol on the nuclearized ICP-MS with the RF set at 1200 W.Unfortunately, energy losses could not be evaluated for lowerplasma forward powers because the stopping curves no longerhave the same shape as the one shown in Fig. 2 when theplasma conditions are no longer robust. It appeared that theplasma RF power indirectly influences the energy of ions byacting slightly on the plasma temperature T0. As previouslyreported,4,21 T0 follows the forward power variations: a 5340 Ktemperature was associated with our usual working RF setting,while a lower T0 value was calculated (5150 K) for the 1200 Wforward power. These values furthermore appear to be in goodagreement with temperatures expected in the ICP.4,22,23,26,28

On the contrary this small modification of the RF plasma

FIG. 4. Dependence of the ion kinetic energy with the VH acceleration voltageapplied to the collision/reaction cell of the nuclearized ICP-MS Q X7 Series Ispectrometer. RF power was set at 1400 W and nebulizer flow at 0.85 L�min�1.No torch shield was used.

FIG. 5. Modification of the optics located before the cell between the ICP-MSX7 Series I and its evolution, the X7 Series II. The Protective Ion extractionlens (p lens) plays on the extraction fields that do not penetrate into the cavity atthe back of the skimmer cone; the material deposited on this cone thus remainsdeposited and does not get sputtered off and re-ionized.

FIG. 6. Schematic representation of the extraction region. The distance fromthe top of the load coil to the sampling cone is shown.

APPLIED SPECTROSCOPY 211

power seems to influence neither the slope a nor the plasmapotential VP, which maintains a constant value at 7.7 V, also inagreement with literature data27,28 (Table III).

Playing on the RF power parameter while at the same timekeeping robust plasma conditions does not enable obtaining amarked change in the ion kinetic energy. Indeed, it appears thatonly a maximal energy variation of 0.5 eV is possible bymodifying the RF plasma power by 200 W.

Effect of the Plasma Torch Shield. Yamada et al.14 added atorch shield to the ICP source in their experiments to minimizethe plasma offset voltage in order to obtain low initial energyand to narrow the energy distribution. For that matter, certainMC ICP-MS instruments, equipped with a collision cell as areplacement for the electrostatic sector analyzer classicallyencountered in double-focusing apparatus, use this principle formore easily thermalizing the ions and significantly improvingthe instrument performances with, in particular, an increase insensitivity. However, to our knowledge papers referring to thisinstrumental improvement never give concrete values for thedecrease of VP resulting from the addition of the torch shieldsystem.

Our study on the ICP-MS X7 Series II confirms thisinfluence of the torch shield on the ion energy and furthermorebrings data to quantify this effect, even if a wider gap wasexpected between the two cases. While the plasma potentialwas evaluated at 8.2 6 0.3 V in the classic configuration(without using the capacitive shield), its value decreased to 5.56 0.25 V when the Pt guard electrode is inserted between theICP torch and the load coil, leading to ions with a lower initialenergy. Other contributions to the energy are also impacted.The plasma temperature T0 actually decreased by 800 K,passing from 3730 K to 2970 K, while the coefficient a,initially near the theoretical value (a¼ 0.93), fell to 0.83. Theaddition of a torch shield therefore plays a major role in theconditions of extraction of ions from the ICP since thetemperature and the potential of the plasma are significantlydropped. The main modification of the kinetic energy of ionsnevertheless comes from the decrease of the VP parameter. Thiseffect can be used to prevent or at least limit the formation ofunwanted species in the cell through endothermic processes.

Figure 8 illustrates this possibility via the formation of EuOþ

from the endothermic reaction of 153Euþ ions with O2. When noPt guard is inserted between the torch and the load coil and theVH potential is set toþ5 V, almost 20% of the 153EuOþ ions areobserved from around 1 mL�min�1 O2 in the cell despite thepositive enthalpy of the reaction (DrH ¼þ108.7 kJ�mol�1).30

The kinetic energy of ions contributes to exceeding thethermodynamic barrier that should normally inhibit thereactivity. By using the results obtained for 147Sm, whose massis close to that of 153Eu, the kinetic energy of europium cationsis evaluated around 8.5 eV in the laboratory frame, i.e., 140kJ�mol�1 in the Center of Mass frame, which is higher than theendothermic barrier. By shielding the ICP, the plasma potentialVP loses almost 3 V, leading to an ion energy expressed in thelaboratory frame close to 5.5 eV, i.e., only 90 kJ�mol�1 in theCenter of Mass frame. Measurements of Euþ and EuOþ signalintensities according to the O2 flow rate show this time thatproportions of the EuOþ product are far less significant. Thisobservation is consistent with the endothermic feature of thecorresponding reaction and underlines the relevance of knowingthe kinetic energy of the ions at their entrance into the collision/reaction cell in order to be able to interpret the reactivity.

CONCLUSION

In this study, the influence of different instrumental settingsof the ICP-MS Q X7 spectrometer (Thermo-Fisher Scientific)on the physical parameters linked to the energy of ions hasbeen examined. This device, which is commonly used forroutine analyses, has also been involved in our laboratory instudies relative to the understanding of the chemistry of ion–molecule reactions in the gas phase. Assessing the energy of

FIG. 8. Signal intensities of the 153Euþ cation and the 153EuOþ product ionaccording to the O2 flow rate for a europium solution at 5 ppb. The hexapolepotential VH was set at�5 V, plasma power RF¼ 1400 W, and nebulizer flowrate ¼ 0.85 L�min�1. Solid black symbols correspond to the case where nocapacitive shield is used on the ICP torch, while the open symbols represent theresults obtained for a shielded torch.

FIG. 7. Influence of the sampling depth on (left) the plasma potential VP and(right) the plasma temperature T0 on the ICP-MS X7 Series I. Error barsrepresent standard deviations. On the left diagram the gray line corresponds tothe plasma potential mean value (VP,mean ¼ 8.7 V) and the dashed gray linescorrespond to the values VP,mean 6 r (r¼ 0.4 V).

212 Volume 63, Number 2, 2009

ions that enter into the collision/reaction cell is therefore usefulfor interpreting certain reactivities. A linear relationshipbetween the voltage applied to the cell and the energy of ionshas been established through the measurement of energy lossesdue to elastic collisions of ions with the gas molecules. Itshowed the kinetic energy range of the ions accessible with theapparatus and the way according to which this one can bemodified by playing on the multipole voltage, which behavesas an acceleration potential.

Experiments carried out on both the ICP-MS X7 Series I andits evolution, the Series II, highlighted that the presence of thep lens after the extraction contributes to slow down thediminishment of energy of the ions when the voltage VH isincreased. The plasma forward power appeared to slightlyinfluence the temperature of the plasma, T0, while the additionof a torch shield led to a decided lowering of T0 and to adecrease of the plasma potential VP, quantified in this study tobe on the order of 3 V.

1. K. Kehm and E. H. Hauri, C. M. O’D. Alexander, and R. W. Carlson,Geochim. Cosmochim. Acta 67, 2879 (2003).

2. M. Segura, Y. Madrid, and C. Camara, J. Anal. At. Spectrom. 18, 1103(2003).

3. S. D. Tanner, J. Anal. At. Spectrom. 10, 905 (1995).4. H. Niu and R. S. Houk, Spectrochim. Acta, Part B 51, 779 (1996).5. A. P. Vonderheide, M. V. Zoriy, A. V. Izmer, C. Pickhardt, J. A. Caruso,

P. Ostapczuk, R. Hille, and J. S. Becker, J. Anal. At. Spectrom. 19, 675(2004).

6. F. Vanhaecke, Anal. Bioanal. Chem. 372, 20 (2002).7. S. D. Tanner, V. I. Baranov, and D. R. Bandura, Spectrochim. Acta, Part B

57, 1361 (2002).8. D. R. Bandura, V. I. Baranov, and S. D. Tanner, Fresenius J. Anal. Chem.

370, 454 (2001).

9. G. C. Eiden, J. Barinaga, and D. W. Koppenaal, Rapid Commun. MassSpectrom. 11, 37 (1997).

10. G. K. Koyanagi, D. Caraiman, V. Blagojevic, and D. K. Bohme, J. Phys.Chem. A 106, 4581 (2002).

11. D. W. Koppenaal, G. C. Eiden, and C. J. Barinaga, J. Anal. At. Spectrom.19, 561 (2004).

12. H. Isnard, M. Aubert, P. Blanchet, R. Brennetot, F. Chartier, V. Geersten,and F. Manuguerra, Spectrochim. Acta, Part B 61, 150 (2006).

13. G. Favre, R. Brennetot, F. Chartier, and P. Vitorge, Int. J. Mass Spectrom.265, 15 (2007).

14. N. Yamada, J. Takahashi, and K. Sakata, J. Anal. At. Spectrom. 17, 1213(2002).

15. M. A. Dexter, H. J. Reid, and B. L. Sharp, J. Anal. At. Spectrom. 17, 676(2002).

16. Z. Du and R. S. Houk, J. Anal. At. Spectrom. 15, 383 (2000).17. S. Becker and H. J. Dietze, Fresenius J. Anal. Chem. 368, 23 (2000).18. S. F. Boulyga and S. Becker, J. Anal. At. Spectrom. 17, 1202 (2002).19. K. Sakata and K. Kawabata, Spectrochim. Acta, Part B 49, 1027 (1994).20. N. S. Nonose, N. Matsuda, N. Fudagawa, and M. Kubota, Spectrochim.

Acta, Part B 49, 955 (1994).21. J. B. Olsen, J. H. Macedone, and P. B. Farnsworth, J. Anal. At. Spectrom.

21, 856 (2006).22. S. D. Tanner, J. Anal. At. Spectrom. 8, 891 (1993).23. G. Holland and S. D. Tanner, ‘‘Experimental Studies of Ion Kinetic

Energies in ICP-MS’’, in Applications of Plasma Source Mass Spectrom-etry II, A. N. Eaton, Ed. (Springer Verlag, London, 1993), p. 222.

24. W. Chen, T. Olney, and D. J. Douglas, J. Anal. At. Spectrom. 14, 9 (1999).25. W. Umrath, ‘‘Tables, Formulas, Nomograms and Symbols’’, in Funda-

mental of Vacuum Technology (Leybold Vacuum, Products and ReferencesBook, Cologne, 2001/2002), Chap. 9, p. 142, Table 3.

26. A. L. Gray and J. G. Williams, J. Anal. At. Spectrom. 2, 599 (1987).27. R. S. Houk, J. K. Schoer, and J. S. Crain, J. Anal. At. Spectrom. 2, 283

(1987).28. J. E. Fulford and D. J. Douglas, Appl. Spectrosc. 40, 971 (1986).29. A. L. Gray, R. S. Houk, and J. G. Williams, J. Anal. At. Spectrom. 2, 13

(1987).30. G. K. Koyanagi and D. K. Bohme, J. Phys. Chem. A 105, 8964 (2001).

APPLIED SPECTROSCOPY 213