JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1987, VOL. 2 13 Langmuir Probe Potential Measurements in the Plasma and their Correlation with Mass Spectral Characteristics in Inductively Coupled Plasma Mass Spectrometry Alan L. Gray Department of Chemistry, University of Surrey, Guildford, Surrey GU2 5XH, UK R. S. Houk Ames Laboratory-US Department of Energy and Department of Chemistry, lowa State University, Ames, /A 5001 7, USA John G. Williams Department of Chemistry, University of Surrey, Guildford, Surrey GU2 5XH, UK A floating Langmuir probe is used to measure the apparent d.c. potential in an inductively coupled plasma (ICP) while the latter is used as an ion source for mass spectrometry (MS). The probe is swung through the plasma to provide potential measurements with some spatial resolution and to obviate cooling of the probe.The d.c. potential in the centre of the plasma is influenced by the presence of the metal sampling cone in the plasma and also by the gas flow through the orifice. In general, the potential correlates with the characteristics of the mass spectra; these parameters depend in a sensitive fashion upon the manner by which the load coil is grounded and shielded. For the load coil geometries investigated in this study, the potential and ion energies generally increase with aerosol gas flow-rate and decrease as power increases. As potential increases the abundance of doubly charged ions generally increases, ArO+/Co+ and Ar2+/Co+ decrease and CeO+ decreases slightly. The measured potential in front of the sampling orifice is generally a few volts below the mean ion energy, which indicates that both measurements are reasonable approximations to the actual d.c. plasma potential.Keywords: Inductively coupled plasma mass spectrometry; ion sampling processes The analytical performance of ICP-MS is critically dependent on the ion extraction process. A crucial step in the develop- ment of ICP-MS into a useful analytical technique was the empirical study of the plasma - sampling cone interaction172 and the establishment of practical means of controlling r.f. potentials in the plasma, any resulting discharges to the grounded sampling cone and r.f. interference in the counting and control circuits. The load coil normally used to sustain an ICP for emission spectrometry has the downstream (top) end at high potential and the upstream (bottom) end grounded.A basic approach to reduce the r.f. potentials between the plasma and the sampling cone is to modify the load coil and its grounding arrangements. Initially this simply involved grounding the downstream end of the coil and this has been used by Gray,l Gray and Date,2 Olivares and Ho~k39~ and in the PlasmaQuad device marketed by VG Isotopes, An effective load coil configuration that reduces the r.f. potentials in the plasma to very low levels has been described by Douglas and French5 and incorporated into the Sciex Elan ICP-MS device. High voltage from the impedence matching network is applied to both ends of the coil and a ground strap is connected to a point equidistant from either end.Douglas and French report that this configuration also improves various other performance figures such as orifice lifetime and mass spectrometric peak shapes. More recently Grays has reported the use of other coil configurations which, although still grounded at one end, greatly reduce the r.f. potentials in the plasma by the geometrical disposition of the turns of the coil and by the use of a grounded screen between the coil and the plasma. Measurements of ion kinetic energy and other mass spectrometric parameters indicate that the plasma potentials obtained with the simple reversed load coil are higher than those from either the modified but asymmetrically grounded coils or from the centre tapped coil and that with all the asymmetrically grounded coils these potentials are more dependent on plasma operating parameters than with the centre tapped load ~0i1.4~6-8 Nevertheless, proper adjustment of operating conditions yields good analytical performance from all of these systems.This paper describes a way of experimentally estimating the apparent plasma potential directly and shows that the results of such measurements correlate with the mass spectral characteristics for the simple reversed load coil geometry. Experimental Potential Measurements The apparatus used to estimate plasma potential is shown in Fig. 1. It consisted of a W rod (1 mm diameter) which was allowed to swing through the plasma on a PTFE pivot. The tip of the rod protruded from a quartz sheath so that a short section 1 mm long was exposed to the plasma.A single swing of the probe took ca. 0.2 s; thus external cooling of the probe was not necessary because it was not exposed to the plasma long enough to cause melting or formation of an appreciable oxide coating. The probe path was positioned to traverse the centre of the plasma between the torch end and sampling orifice as shown in Fig. 1. The probe was connected via coaxial cables and a voltage divider (X 9) to the vertical input (impedance 1 MO) of a storage oscilloscope. A simple switch triggered one sweep of the oscilloscope when the probe was released. The resulting trace was a record of potential sensed by the probe as a function of position along the probe path through the plasma. This profile was then traced by hand to yield the potential profiles shown below.Langmuir probes are generally used to measure the electron temperature and density in a plasma, which requires determi- nation of probe current as a function of externally applied voltage.9 In the experiments described herein no external bias voltage was applied to the probe; the probe therefore sensed only the potential induced upon it by its contact with the plasma. Previous work by Douglas and French reported theJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1987, VOL. 2 F I .- c 0.89 E 2 0.82 F 2 0.75 $ 0.67 2 0.60 2 0.53 - - 0 - + u, - $? 14 ( a ) - - - - - - l-z;d** ICP To \ Probe scope To path PTFE ground arm 0 10 20rnrn u Shield? f':te STpler Induction region Load CAA Tokh 4 0 10 20 30mm Fig. 1. Scale diagrams of apparatus: (a) view down barre- of torc..showin probe, insulating sleeve and pivot, and probe path through ICP; (by side view of torch, probe path and sampling cone. The orifice diameters through and the spacing between the sampler and skimmer have been enlarged for clarity; actual values for these dimensions are listed in Table 1 peak to peak r.f. voltage in the plasma, whereas the present work reports d.c. potentials because the r.f. component was filtered out by the long time constant of the input circuit of the oscilloscope used. Measurements with a larger (2 mm diameter) hand-held probe at various spatial positions in the plasma yielded potentials of the same polarity and similar magnitude to those measured with the swung probe. This experiment was carried out to verify that the potentials measured were not critically dependent on probe size or the rate of travel of the probe through the plasma. The hand-held probe was also used to check the d.c.potential with r.f. power applied but the plasma unlit. No potential greater than 0.1 V was seen, confirming that the potentials measured were not due to r.f. pick-up. ICP-MS Instrumentation The essential components and performance of this device have been described previously.2 Standard operating conditions are listed in Table 1; these conditions are close to those preferred for multi-element analysis with this equipment when used with a coil similar to coil X (Fig. 2). This table shows standard operating parameters when values for particular operating conditions are not specified for the data presented below.Three slightly different configurations for the load coil and shielding were used; these are depicted in Fig. 2. Coil X (referred to as the strapped and shielded configuration) was used unless otherwise stated. These coil configurations were chosen because they correspond closely to those used in earlier publications describing ion kinetic energies and mass spectral characteristics with two different ICP-MS instru- ments,4?7 and it was considered desirable to determine whether the plasma potential measurements were consistent with this previous work. The present paper is not meant to be a comprehensive evaluation of the best possible load coil geometry for ICP-MS.5.6 The sample solution contained 1 mg 1-1 of Co, Ba and of Ce in 1% HN03 in distilled de-ionised water.Cobalt was chosen because it does not form appreciable doubly charged or oxide ions and is close in mlz to Ba2+ and Ce2+. Barium and cerium I 0 29 :Omm c b I 0 0 4 HV uooh - X Y 2 Fig. 2. Scale diagrams of load coils, shielding and grounding configurations used in uresent work. A, Plasma torch; B, load coil; C, quartz bonnet; D, high voltage (HV) end of load coil; E grounded end of coil; and F, front screen I 1 1 1 1 1 1 1 1 1 1 1 ~ 1 ~ ~ -8 -7 -6 -5 -4 -3 -2 - 1 0 1 2 3 4 5 6 Position of IRZ/mm from coil edge Fig. 3. Position of tip of initial radiation zone relative to downstream turn of load coil for various powers and aerosol gas flow-rates A, 1 .O; B, 1.2; C , 1.4; and D, 1.8 kW were chosen because barium has the lowest second ionisation energy and CeO+ is one of the more refractory oxide species observed in ICP-MS.Mass spectra were acquired for cu. 1 min by scanning repetitively from rnlz 50 to 180; the resulting spectra were summed and transferred from the multi-channel scaler to cassette tape. Regions of interest were later selected corresponding to various ionic species of interest (e.g., background, singly charged analyte, doubly charged analyte and analyte oxide ions) and the resulting net peak integrals determined. Ion kinetic energies were determined for Co+ with the mass analyser transmitting only mlz 59 by applying a retarding positive potential to the entrance lens to the rods; otherwise, this lens was normally grounded. The potential required to stop the Co+ beam (denoted V,,, below) was taken as the applied voltage necessary to attenuate the Co+ signal to 0.1% of its original value for the particular set of parameters selected. This measurement method yielded similar trends to those reported by Olivares and Houk on their ICP-MS device except that the V,,, values determined in the present work tended to be somewhat higher.4 The position of the sampling orifice with respect to the load coil was left constant for most of the experiments reported herein.However, aerosol gas flow-rate and forward power to the plasma were varied over wide ranges so that the effective sampling position relative to the initial radiation zone (IRZ) varied between experiments.11 Therefore, the position of the downstream tip of the IRZ was determined visually during nebulisation of 1000 mg 1-1 of Y for various plasma operating parameters. The results are depicted in Fig.3 so that the changes in the spatial structure of the plasma during the experiments described below can be followed. The sampling orifice was downstream from the IRZ for all the powers and aerosol gas flow-rates used. Ion lens parameters were not fully re-optimised as would be done for analytical use after eachJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1987, VOL. 2 15 Table 1. Standard operating conditions PowerIkW . . . . . . . . 1.2 Torch . , . . . . . . Standard length Fassel typelo Outer Ar gas flow-rate/ Auxiliary Ar gas flow-;aie/ Nebuliser . , . . . . . . Jarrel- Ash cross-flow Aerosol gas flow-ratell min-1 0.53-0.60 Sampling orifice: lmin-1 .. . . 12 1 min-’ . . . . . . . . 0 single-pass spray chamber lo Position . . . . . . . . 12 mm from end of load coil on centre Holediametedmm . . . . 1.0 Conematerial . . . . Ni Position . . . . . . . . 6 mm behind sampler on centre Holediametedmm . . . . 1.0 Skimmer: parameter change, so that absolute values of the Co+ signal did not represent optimised performance at each plasma parameter setting. However, the analytical significance of polyatomic ions such as ArO+ and Ar2+ lies in their relative size in relation to analyte reponse so these values are quoted as a ratio to the Co+ response. Results and Discussion Interpretation of Probe Potential Measurements Typical potential profiles obtained with load coil configura- tion X (Fig. 2) are illustrated in Fig.4. The approximate outer limits of the luminous white section of the plasma are also illustrated on the figure. In general the potential profiles showed three maxima across the plasma. The outermost potential maxima were just outside the plasma boundary. The farthest right maximum was slightly closer to the centre than the farthest left maximum because the probe velocity naturally increased during its travel down through the plasma. The structure on the farthest left maxima was not reproducible and thus is probably just noise. Most of the potential profiles had a sharp central maximum such as that depicted in Fig. 4. The actual potential at this position is denoted V, (where c denotes the centre) and is measured relative to ground because the sampling orifice is grounded.The observation that V, was generally of positive polarity and in the range 5-20 V agrees qualitatively with ion kinetic energies measured previously with the reversed load coil ge0metry.4~7.8 The probe revealed a plasma “warm-up” effect in that it took ca. 30 min for the potential profiles to stabilise after the plasma had been started and positioned in contact with the sampler. No such time lag was observed for a simple change of plasma parameters after this initial warm-up period, however. The trends of V, and profile shape with plasma parameters were quite reproducible although the magnitude of the V, values measured was reproducible only to within k 30% from day to day. It should also be noted that an isolated probe immersed in a plasma will assume a potential that is less than the actual plasma potential.The magnitude of this potential difference (often called the floating potential) depends upon the extent of electron cooling and ion - electron recombination in the vicinity of the probe. These effects are difficult to estimate accurately for probe measurements with an ICP. However, the floating potential could be of the order of several kT, (k = Boltzmann constant, T, = electron temperature in the unperturbed plasma near the probe), i.e., a few eV.12-14 The potential measurements reported below underestimate the actual plasma potential by up to this amount. Despite this offset in the absolute potentials, the measured potentials still illustrate the trends in actual plasma potential with plasma parameters, coil geometry, etc.Table 2. V, values (volts) measured for various plasma configurations, power 1.2 kW, coil to aperture spacing 12 mm Gas No gas extraction extraction Aerosol 1 min-1 Dry* vapourt aerosol$ gas flow-rate/ H20 H2O H2O aerosol - - - 0 34 0.53 33 34 27 8 0.60 32 32 25 10 0.67 33 31 23 12 0.75 34 29 22 16 0.82 34 27 21 18 0.89 33 25 20 18 * No H20 introduced. t HzO vapour introduced from residual H20 in spray chamber. 5 HzO aerosol introduced. 20 10 0 20 - ? : l o m .- +- CI 0 a 0 20 i a 0 (13) (12) 0 Probe position - Fig. 4. Measured potential profiles for coil X at 1.2 kW forward power and aerosol gas-flow rates of: (a) 0.75; (b) 0.60; and (c) 0.53 1 min-1. The central potential (V,) and the approximate outer boundaries of the ICP are indicated.Position 0 indicates the centre of the ICP Effects of Ion Sampling Conditions on Potential Measurements Potential profiles were obtained for a variety of plasma conditions that deviate from normal analytical application. Nevertheless, some interesting observations were made that pertain to the general objectives of this study. High negative potentials (ca. -50 V) were observed with the plasma on but separated from the sampler by ca. 0.5 m. In this situation the plasma was not in contact with a grounded object, the probe being isolated by 10 MQ. Formation of the space charge sheath around the floating probe caused the probe to assume a net negative potential.12-14 The polarity and magnitude of the potentials changed when the plasma was positioned in contact with the sampling orifice, for reasons to be described below.The V, values measured at the normal operating power level of 1.2 kW are shown in Table 2. The values obtained with the expansion stage pump shut off, so that no gas was extracted, are shown in the three columns under “No gas extraction.” When the plasma was unpunched, i.e., with no aerosol gas flow, the measured V, was 34 V. Punching the central channel through the plasma with a flow of dry argon at flow-rates from 0.53 to 0.89 1 min-1 made no significant difference to V,. The16 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1987, VOL. 2 introduction of water vapour obtained by passing the aerosol gas through the nebuliser spray chamber without water uptake (column “HzO vapour”) made no difference to the potential at low gas flow-rates but as the flow-rate increased and more vapour was taken up from the water surface in the spray chamber the potential steadily decreased.The introduction of still more water by operating the nebuliser uptake pump, so that aerosol droplets were introduced as well as vapour, resulted in still lower potentials, down to 20 V at 0.89 1 min-1. Under these “extractionless” conditions the boundary layer over the surface of the sampling cone remained unbroken. As soon as the expansion stage pump was started, however, the boundary layer over the aperture was broken, and the potential became very much lower. Although not shown in Table 2 the potential observed when the plasma was unpunched during ion extraction was ca.+2 V. When the axial channel was punched with dry argon the potential rose to ca. 4 V. When water was added as nebulised aerosol the potential increased with flow as shown in the final column, i. e., the opposite trend to that seen under “extractionless” conditions. This marked difference in behaviour when ions were extracted was thought to be due to the much closer contact between the plasma and the cone once the boundary layer over the aperture had been penetrated. The complete boundary layer (ie., with the pump off) formed a relatively high impedence between the plasma and grounded cone, which was affected by the amount of water introduced, possibly due to the higher electron density from ionised H and 0.15 When the plasma was drawn through the aperture it was only separated from the cone by a much thinner boundary layer along the inner surfaces of the orifice superimposed on the very thin plasma sheath.16 Therefore, a much lower potential developed across the thinner boundary layer because the impedance to the flow of charged particles was lower.It is the presence of this sheath that produces the positive d.c. offset potential (relative to the grounded cone) from the r.f. potentials in the plasma,l2 the differing mobilities of ions and electrons causing the sheath to behave partly as a rectifying layer. Behaviour similar to this has already been reported from the ion energy measurements.8 Conditions in the sheath region may be expected to depend critically on plasma parameters such as temperature and electron density, which are themselves determined by aerosol gas flow-rate and composition, plasma power, etc., and geometrical considera- tions such as coil configuration, plasma coil - aperture spacing and aperture diameter.The above hypothesis is only intended to provide a qualitative picture of the ion extraction process. A precise description of these phenomena involving time varying electric fields, fluid effects, spatial gradients of temperature and compositions, etc. is a very challenging objective. A useful discussion of the behaviour of plasma sheaths in the presence of high frequency r.f. potentials has been given by Chapman. 12 Potential and Mass Spectrometric Measurements Under Analytical Operating Conditions Numerous potential profiles (Figs. 4-7) and mass spectra (Table 3) were obtained for a range of plasma operating conditions spanning those of use for actual analysis with this particular device (Table 1) except that the coil - aperture spacing was 15 mm.Mass spectra (but not potential profiles) were obtained with this load coil at a spacing of 12 mm; the trends of Ba2+/Co+, etc. with power and aerosol gas flow-rate were similar to those shown in Table 3. For all the tables in this paper the ratio Cez+/Co+ was also determined; this ratio was always smaller than Ba2+/Co+ and followed the same trends with power and flow-rate. Inspection of the data in Table 3 yielded the following general trends with some exceptions as V C "pel t 4 l o t (6) y o y 0 Probe position - Fig. 5. Potential profiles at low aerosol gas flow-rate (0.53 1 min-I), coil X: (a) 1.4; (b) 1.6; and (c) 1.8 kW.Compare with Table 3 20 0 - Probe position - Fig. 6. Potential profiles for coil x at 0.75 1 min-1 aerosol gas flow-rate and powers: of (a) 1.0; (b) 1.2; (c) 1.4; and (d) 1.6 kW. Compare with Table 3 Probe position - Fig. 7. Potential rofiles for coil X at 0.89 1 min-l aerosol gas flow-rate: (a) 1.2; (g) 1.4; and (c) 1.6 kW. Compare with Table 3JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1987, VOL. 2 17 Table 3. Measured potential values compared with mass spectral characteristics for various plasma operating parameters. The sampling orifice was 15 mm from the load coil (X) for this data Aerosol/ 1 min-1 Power/kW 0.53 1 .o 1.2 1.4 1.6 1.8 0.60 1 .o 1.2 1.4 1.6 0.75 1.0 1.2 1.4 1.6 0.89 1 .o 1.2 1.4 1.6 V,N* 13 11 10 8 6 17 13 10 8 16 18 10 8 13 13 14 9 vmax/v-t 25.8 25.6 19.6 19.1 18.1 29.3 25.8 23.8 19.9 38.2 30.7 28.1 23.7 38.8 38.5 32.8 28.2 B aZ+/Co + , ratio, YO 1.40 0.50 0.41 0.50 0.84 5.40 0.64 0.34 0.57 50.3 5.6 0.66 0.46 82.0 48.0 9.3 1.9 * Uncertainties typically hl V.t Uncertainties typically k0.5 V. $ ArO+ monitored at rnlz 56 and Ar2+ at rnlz 80 for this and subsequent tables. CeO+/Ce+, ratio, % 0.69 0.99 1.30 1.70 2.10 0.55 0.85 1 S O 1.60 0.91 0.78 0.85 1.20 1 .oo 0.95 0.94 1 .oo ArO+/Co+, ratio, %$ 22 35 68 96 135 13 19 53 68 4.2 7 1.2 21 4.5 3.5 5.2 11 Ar,+/Co+, ratio, O/O$ 479 505 221 189 139 173 334 461 194 24 97 91 290 16 14 47 97 noted. For the same operating conditions the V, values listed on the figures do not always agree exactly with those in Table 3 because the latter were averaged over several swings of the probe, whereas the figures were traced after only one swing.Increasing the aerosol gas flow-rate at constant power induced either an increase or no appreciable change in V, (e.g., Fig. 4) except at low power and high aerosol gas flow-rate. Increasing the power at constant aerosol gas flow-rate generally induced a decrease in V, except at low power and high aerosol gas flow-rate (e.g., 1.0 kW and 0.75 1 min-1, Table 4, Figs. 6 and 7). In the last instance the valleys in the potential profiles (marked by the arrows in Figs. 6 and 7) were elevated considerably further above ground than under other operating conditions. In contrast, use of high power and low aerosol gas flow-rate suppressed these valley potentials (see arrows, Figs.5 and 6) to small positive or even negative values, which tended to pull V, down also. Increasing aerosol gas flow-rate andor decreasing power caused V,,, and Ba2+/Co+ to increase. The ratio CeO+/Ce+ increased with the aerosol gas flow-rate at low power (Table 3), decreased as aerosol gas flow-rate increased at high power, and generally increased with power at constant aerosol gas flow-rate. In other words, changing a plasma parameter that caused an increase in V, generally induced a corresponding increase in ion energies and doubly charged ion abundances and a small decrease in metal oxide abundances. The behaviour described above is consistent with the following interpretation. The majority of the gas species in the centre channel of the plasma that enter the aperture are neutral (ca.99.9%) and acquire an energy from the supersonic expansion of about 1.5 eV.17 A relatively small proportion (ca. 0.1 %) of the total species are charged with equal numbers of positive ions and electrons present. The potential plots obtained with the probe show the potential to have a maximum at the centre of the plasma, the size of which depends on the plasma operating parameters. Relative to the grounded sampling orifice, the ions and electrons generated in the plasma appear to originate from a region at this potential and thus ions enter the aperture with a kinetic energy greater than that of the neutrals. At the higher plasma potentials, the d.c. and r.f. fields in the region of the plasma sheath at the aperture contribute to the electron energy so that excited states are populated and decay to produce a visible glow inside the expansion chamber when water is introduced into the central channel.At higher energies still, for V, values of ca. 13 V and above, additional double ionisation is produced in species such as Ba with low second ionisation energies. The extent of this does not appear to correlate with V, values above 13 V but does increase with Vmax above a threshold at ca. 28 V. At the flow-rates (0.53 and 0.6 1 min-1) and power (1.2 kW) normally used for analysis with this coil, values for V,, V,,,, Ba2+/Co+ and CeO+/Ce+ are close to the lowest obtained. As reported previously, the trends of Baz+/Co+ and CeO+/Ce+ with power and flow-rate were opposite to those expected from the ICP al0ne.18~19 This behaviour indicates that the potential difference between the plasma and the orifice could be responsible for helping to keep oxide ions dissociated during the extraction process.However, Table 3 shows that ( a ) there is only a factor of 2-3 to be gained in reducing CeO+/Ce+ by deliberately inducing a high plasma potential, (b) at low power, increasing the aerosol gas flow-rate actually makes CeO+/Ce+ less favourable and ( c ) deliberate elevation of the plasma potential to overly high values would make doubly charged ions more of a problem. The interpretation of the effect of power and aerosol gas flow-rate on the polyatomic (cluster) ion levels is superficially more straightforward. Both ArO+/Co+ and Ar2+/Co+ dec- rease consistently with increasing aerosol flow, suggesting that the increased energy available at higher plasma potentials does produce dissociation.These species are probably formed behind the aperture and are relatively weakly bound. They may be expected to show the highest cross-sections for their formation reactions at low ion energies, so the response may merely reflect the probability of their formation in the supersonic expansion. This behaviour may also simply reflect changes in the abundance of the atomic precursor ion (e.g., O+ for ArO+, Ar+ for Arz+) in the plasma with operating conditions, i.e., reduction of the O+ concentration in the plasma means there is less O+ available to react with neutral Ar during the expansion, so less ArO+ is detected.Apart from a few instances at the lowest values of V, obtained with this coil, and which are insufficiently general to carry much weight, the trends reported are generally consis- tent with those reported previously for load coils that are asymmetrically grounded,6?7 and opposite to those reported for the centre tapped load coil with which the behaviour of doubly charged, oxide and cluster ions tends to follow what would be expected from the ICP itself.lg.19 With the centre tapped coil, ion energy is almost independent of plasma18 10 5 - 0 -5 ? .- c(1 - 1 0 - - c JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1987, VOL. 2 - ( a ) - (C) (8) (5) V t - t I ( 4 0 2l Probe position - Fig. 8. Potential profiles for coil X, 1.0 kW: (a) no aerosol gas flow; and dry Ar only in aerosol gas flow at (b) 0.53 and (c) 0.60 1 min-1 operating parameters.In spite of this apparent advantage the analytical performance of ICP-MS systems with either type of load coil has been shown to be similarly dependent on plasma operating parameters, especially aerosol carrier gas flow- rate.3.7.18 The real significance of this dependence however lies in the variations experienced by a user in operating the system when plasma parameters are nominally stabilised and this appears to depend more on other aspects of system engineering than on the specific load coil geometry involved. Effects of Water Content on Probe Potential Measurements Additional potential profiles were obtained while gas was extracted through the orifice from ICPs with no axial channel and with dry Ar rather than Ar plus nebulised H20 injected into the axial channel.Typical results are depicted in Fig. 8 and V, values are shown in Table 2. In the absence of aerosol gas flow the measured potential in the centre of the ICP was very low; punching the axial channel with dry Ar induced an increase in V,. These two observations are in agreement with those reported previously for one of the Ames instruments4 and for the Surrey instrument used here.8 However, compari- son of the potential profiles and V, data in Fig. 8 with those for injection of wet aerosol (e.g., Fig. 5 and the right-hand column of Table 2) indicate that introduction of water while gas was being extracted caused V, to increase sharply compared with that obtained when only dry Ar was injected into the axial channel.This behaviour has also been reported previously.8 As noted above, the opposite trend was observed when there was no gas flow through the orifice. It is also interesting that the injection of water into the axial channel raised the voltage signal outside this region of the plasma. In other words, the potentials measured in the valleys of the profiles (marked by arrows in Fig. 8) were affected by the presence or absence of H20 in the axial channel. It would seem that the injected H20 plays an active role in determining the over-all chemical and even electrical characteristics of the ICP when the latter is used as an ion source. Effects of Shielding Experiments such as those described above were also perfor- med using the same load coil but a larger aperture in the shielding box (Y, Fig.2). Typical results are depicted in Figs. 9-11 and Table 4. For these data the sampling orifice was slightly closer to the load coil (12 mm) than for those shown in Table 3. Because the mass spectral characteristics were also influenced by the ion sampling position, the absolute values in Table 3 are not strictly comparable to those in Table 4. The discussion below thus emphasises the trends of V, with V,,, 10 5 0 -5 -10 ? 10 .- 2 5 - Q, a 5 0 -5 -10 10 5 0 -5 -10 t Probe position - Fig. 9. Potential profiles for coil Y at 1.2 kW and aerosol gas flow-rates of: (a) 0.53; (b) 0.60; and (c) 0.67 1 min-1. Compare with Table 4. Arrows mark region of apparent negative potential h $21 10 5 0 -5 c v c t I 5 t I Probe position - Fig.10. Potential profiles for coil Y, 1.6 kW for aerosol gas flow-rates of: (a) 0.6; (b) 0.67; (c) 0.75; (d) 0.82; and (e) 0.89 1 min-1. Compare with Table 4 -5 ? - ru .- .I-- != '0 n .I- 0 5 -10' I Probe position - Fig. 11. Potential rofiles for coil Y at 0.89 1 min-1, dry Ar only in aerosol gas flow: (nf1.2; and (b) 1.6 kWJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1987, VOL. 2 19 ~~ ~ Table 4. Measured potential values and mass spectral characteristics for less shielded load coil (Y), sampling position 12 mm Aerosol gas flow-rate/ Power/ Ba2+ /Co + CeO +lCe + ArO + /Co + Ar2+ /Co + 1 min-1 k W vfl V,,,N ratio, Yo ratio, Yo ratio, % ratio, YO 0.53 1.0 1.2 1.6 1.6 0.67 1 .o 1.2 1.4 1.6 8 f l 6 6 t 67 11 10 7 7 19.2 k 0.5 16.7 21.8 18.4 26.0 22.3 18.6 17.2 0.1* 0.1* 0.91 0.87 1.1 0.3 0.03* 0.09* 1.5 1.7 2.4 2.1 0.9 1.4 2.0 2.3 8.4 8.3 60 69 2.4 3.2 6.7 7.7 226 46 54 105 66 132 138 27 0.75 1.2 10 25.4 1.1 1.1 2.2 55 1.6 8 20.8 O* 1.5 3.5 52 1.4 8 24.1 0.2* 1.4 3.8 68 0.89 1.4 12 28.0 1.9 1.1 1.1 14 1.6 12 25.2 1.1 1.3 1.8 23 1.8 12 25.0 0.8 1.5 3.2 50 1.8 127 24.3 0.4 1.4 2.7 31 1.6 121.25.2 0.8 1.2 1.4 22 * Ba2+ levels under these conditions were comparable to or smaller than the background, so the Ba2+/Co+ values listed are only t Results of dudicate trials when indicated power and flow-rate were reset after ca. 30 min interval. approximations. and how these reflect changes in mass spectral characteristics with power and flow-rate. As shown in Figs.9-11, the potential profiles were less symmetrical for the less shielded coil (geometry Y). This was particularly so for the initial valley (marked by arrows in Figs. 9 and lo), which even assumed a negative polarity under conditions of low aerosol gas flow-rate and high power. Increasing aerosol gas flow-rate at constant power displaced this valley to more positive potentials with corresponding increases in V,, V,, and Ba2+/Co+. In Table 4 the general trends in ArO+/Co+ and Ar2+/Co+ with changing power and aerosol gas flow-rate are also similar to those in Table 3. Thus, changing a plasma parameter that altered Vc also changed the characteristics of the mass spectra in ways similar to those seen with coil X. Although the precise cause of the negative valley is not certain, its effect was to reduce the potential difference between the centre of the plasma and the sampling orifice and apparently attenuate the influence of this potential difference on the mass spectra. Several sets of data obtained separately under the same operating conditions are also listed in Table 4; these illustrate the typical reproducibility of these parameters during a single day’s operation.It is very interesting that less extensive shielding of the load coil (geometry Y) should yield unsymmetrical potential profiles and lower plasma potentials and ion energies than a closely screened coil (geometry X). The following two observations with load coil Y also merit discussion. The probe dropped vertically close to the down- stream turn of the load coil, which was connected to the ground at its bottom end (Fig. 1).For the profiles shown in Figs. 9-11 the top side of the coil [marked in Fig. l(a)] floated at higher potential than the bottom side, the latter being closer to the ground strap. The negative valleys in Figs. 9-11 appeared when the probe had just entered the ICP. In other words, the plasma region where negative potentials were apparently developed was adjacent to the load coil section at higher applied r.f. potentials. To test this explanation a load coil of similar dimensions but with the direction of the windings reversed was then substituted so that the higher voltage side of the coil was below centre. The negative valley in the potential profiles now occurred below the centre, i. e., it followed the high voltage side of the coil.Finally, some potential profiles obtained at high aerosol gas flow-rate with only dry Ar injected are shown in Fig. 11. These profiles showed pronounced negative valleys and very small values of V, (2-3 V). Thus the exclusion of H20 yielded lower plasma potentials with either the tightly shielded or loosely shielded load coil. This observation concurs with the general experience in Surrey and Ames that introduction of sample without concomitant H20 yields lower ion energies.20 Effects of Grounding The data described up to this point were obtained with a short copper braid connected from the downstream turn of the load coil to the grounded shielding box. Numerous spectra and potential profiles were also obtained with a load coil grounded only at the coupling box (configuration Z , Fig.2); the results are summarised in Table 5. Comparison of these data with Table 4 indicates that higher levels of V,, Vmax and Ba2+ were generated with the ground strap removed and the diameter of the shielding aperture held constant. This agrees with other observations that grounding the load coil with an additional strap helps to diminish the plasma potential.5 The potential profiles still had negative valleys at high power and low aerosol gas flow-rate, although these valleys were less pronounced than for load coil Y. The trends of V, and mass spectral characteristics with changing aerosol gas flow-rate and power were similar to those seen with the other load coils, except that the ratio CeO+/Ce+ was largely independent of plasma operating conditions with the ground strap absent. It is also interesting that load coil Z at 0.89 1 min-1 yielded levels of Ba2+ comparable to or even slightly less than those obtained with load coil X even though the former lacked a ground strap.Thus, the plasma potential and the characteristics of the mass spectra depend in a sensitive fashion upon the grounding and shielding of the load coil. Nevertheless, careful adjustment and control of operating parameters can yield good analytical performance with various versions of the reversed load coil. Comparison of Plasma Potential with Ion Kinetic Energy It is interesting to note that V, was generally much less than V,, for Co+. However, if V, represented the actual plasma potential, it should have been comparable to the mean ion energy E , which was approximately 4ym,. Inspection of Tables 3-5 shows that in most instances E was higher than Vc by +1 to +4 V; in a few instances ( E - V,) was slightly negative or as high as +10 V.?ere are at least two phenomena that could have caused E to be a few volts above V,. First, as pointed out by Fulford and Douglas, the ions20 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1987, VOL. 2 Table 5. Potential values and mass spectral data for unstrapped, less shielded load coil (Z). Sampling position 12 mm above load coil Aerosol gas flow-rate/ 1 min-1 0.53 0.60 0.67 0.75 0.89 Power/ kW 1 .o 1.2 1.4 1.6 1.8 1 .o 1.2 1.4 1.6 1.8 1.0 1.2 1.4 1.6 1.8 1.0 1.2 1.4 1.6 1.8 1.0 1.2 1.4 1.6 1.8 v f l 1 3 f 1 11 9 9 9 15 12 9 10 9 16 14 11 11 10 15 16 13 12 11 13 16 17 16 14 VmaxN 35.0 f 0.5 31.0 29.3 25.4 23.5 39.4 34.5 30.6 29.6 26.8 43.7 37.4 35.6 31.5 30.2 45.8 40.8 37.5 34.9 32.3 46.9 47.9 45.9 41.2 37.4 Ba2+/Co+ ratio, YO 5.8 0.2 O* O* O* 2.6 O* O* O* 9.3 2.7 O* O* 20 32 31 20 8.7 1.8 0.2* 30 38 29 15 8.6 CeO +/Ce+ ratio, Yo 0.7 0.7 0.7 0.8 1.2 0.5 0.8 0.8 0.8 0.9 0.3 0.6 0.8 1 .o 0.9 0.6 0.7 0.8 1.1 1.1 0.4* 0.6 0.8 0.8 0.9 ArO +/Co+ ratio, YO 5.8 9.5 6.0 4.8 4.7 2.6 5.9 6.0 5.5 7.3 1.9 3.8 3.4 5.2 7.8 1.8 1.7 2.1 3.6 7.0 1.4 0.9 1 .o 1.5 5.9 Ar2+/Co+ ratio, YO 200 150 135 87 54 65 169 160 150 144 7 100 140 161 207 4 18 92 100 107 2.0 2.0 2.6 17 73 * Ba2+ or CeO+ levels under these conditions were comparable to or less than the background, so the ratios listed are only approximations. acquire kinetic energy by being entrained in the Ar flow in the supersonic jet.For Co+ this contribution amounts to ca. 1.5 eV.17 The second source of voltage offset is the floating potential effect described above. The magnitude of this effect tends to increase with the flow velocity of gas in the plasma,13 which could partly explain why V, tended to level off or even decrease at high aerosol gas flow-rate while V,, continued to increase. Incorporation of the floating potential and the energy acquired by the ions during the sup_ersonic expansion brings V, into approximate agreement with E, i.e., the average ion energy is comparable to the plasma potential. The work at Surrey was supported jointly by the British Geological Survey (NERC) and the Directorate General for Science, Research and Development (DG XIUG-2) of the European Community. J.G. W. acknowledges support from the Procurement Executive, Ministry of Defence, UK. R. S. H. was supported by the Ames Laboratory-US Department of Energy, Contract No. W-7405-Eng-82 via the Director for Energy Research, Office of Basic Energy Sciences. R. S. H. also acknowledges travel support provided by VG Isotopes. References 1. Gray, A. L., Ph.D. Thesis, University of Surrey, 1982. 2. Gray, A. L., and Date, A. R., Analyst, 1983, 108, 1033. 3. Olivares, J. A., and Houk, R. S., Anal. Chem., 1985,57,2674. 4. Olivares, J. A., and Houk, R. S., Appl. Spectrosc., 1985,39, 1070. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Douglas, D. J., and French, J. B., Spectrochim. Acta, Part B, 1986,41, 197. Gray, A. L., J. Anal. At. Spectrom., 1986, 1, 247. Gray, A. L., Spectrochim. Acta, Part B, 1986,41, 151. Gray, A. L., Fresenius 2. Anal. Chem., 1986,324,561. Smy, P. R., Adv. Phys., 1976, 5,517. Scott, R. H., Fassel, V. A., Kniseley, R. N., andNixon, D. E., Anal. Chem., 1974,46,75. Koirtyohann, S . R., Jones, J. S., and Yates, D. A., Anal. Chem., 1980, 52, 1965. Chapman, B., “Glow Discharge Processes,” Wiley, New York, 1980, Chapters 3 and 5. Clements, R. M., and Smy, P. R., J. Phys. D, Appl. Phys., 1974, 7 , 551. Swift, J. D., and Schwar, M. J. R., ‘‘Electrical Probes for Plasma Diagnostics,” Iliffe Books, London, 1969, Chapters 1,7 and 12. Alder, J. F., Bornbelka, R. M., and Kirkbright, G. F., Spectrochim. Acta, Part B, 1980,35, 163. Houk, R. S., Fassel, V. A., and Svec, H. J., Dynamic Mass Spectrom., 1981,6, 234. Fulford, J. E., and Douglas, D. J., Appl. Spectrosc., 1986,40, 971. Horlick, G., Tan, S. H., Vaughan, M. A., and Rose, C. A., Spectrochim. Acta, Part B, 1985,40, 1555. Vaughan, M. A., and Horlick, G., Appl. Spectrosc., 1986,40, 434. Jiang, S.-J., and Houk, R. S., Anal. Chem., 1986, 58, 1739. Paper J6170 Received August 4th, 1986 Accepted October 22nd, 1986