JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 199 1 VOL. 6 493 INTER-LABORATORY NOTE Inserted Injector Tubes for Inductively Coupled Plasma Spectrometry Lyne S. Gervais and Eric D. Salin" Department of Chemistry McGill University 80 1 Sherbrooke Street West Montreal Quebec H3A 2K6 Canada Experiments are described in which an inductively coupled plasma is operated with an injector tube that is inserted into the body of the plasma. The system is highly tolerant to pressure fluctuations from the sample introduction system. Both wide (6 mm o.d. 3 mm i.d.) and narrow (2 mm o.d. 1 mm i.d.) injector tubes were studied. Interestingly the intensities observed with the two different injector tubes were fairly similar. Alumina was found to be a superior material since graphite appeared to reduce the power of the plasma.Keywords Inductively coupled plasma; torch; design; injector; sample introduction Work has been carried out on the development of sample introduction systems for inductively coupled plasma (ICP) spectrometry. 1-5 Recently studies have been undertaken on a system that utilizes a Beckman burner with a hydrogen- oxygen flame as a sample n e b ~ l i z e r . ~ * ~ With a conventional ICP torch this sytem produced severe fluctuations in the aerosol flow rate. By the use of an 'inserted injector' torch design this sample introduction system could be stablized. The inserted injector system is extremely tolerant to pressure fluctuations from a sample introduction system which might otherwise extinguish the plasma. The design may be of interest to researchers involved with novel sample introduction systems.All work was carried out on a Jarrell-Ash Model 61 direct reading spectrometer system with a 27 MHz generator operated at 1.0 kW unless otherwise indicated. An outer (plasma) gas flow of 16 1 min-l was used for all experiments. The auxiliary gas flow was varied as described in the given experiment but never exceeded 1.0 1 min-'. A Ligere V-type nebulizer was used with a 0.9 1 min-l argon flow. A heated spray sample introduction system was used with modifications to a conventional torch as de- scribed in the various sections. All solutions were prepared from 1000 ppm standards obtained from Fisher Scientific. Multi-element solutions of Cd Cu Ni Pb and Zn at the 20 ppm level were used for the experiments. Conventional torches use a configuration which places the injector tip several millimetres below the plasma.Since the injector tubes are commonly made of quartz this is 10 mm - 5 mm H essential to their survival. Other materials of course are possible including boron nitride and alumina. Alumina is particularly attractive owing to its ability to tolerate higher temperatures. An undesirable feature is the more demand- ing requirements for shaping this material. Alumina cannot be conveniently shaped in machine or glass shops but must be fabricated by the manufacturer in a given form. In attempts to develop a torch system which was relatively immune to aerosol flow fluctuations initially small extensions were placed on a conventional injector tube.The design is illustrated in Fig. 1. The materials used were boron nitride and graphite. The second set of experiments used the tubular system illustrated in Fig. 2. The tube was commercially available alumina in two dimensions 6 mm o.d. 3 mm i.d. (6/3) and 2 mm o.d. 1 mm i.d. (2/1). The height reference system used for insertion heights is illustrated in Fig. 3. The bottom of the load coil was selected as a logical 0 (zero) reference point. The inner torch tube was located 2 mm below the 0 reference point (- 2 mm in this reference system). Viewing heights were measured conventionally from the top of the load coil. All numeric data (as compared with visual observations) presented were collected from the tubular system. Background corrected signals were calculated for Standard injector tube If )I- Fig.1 Cap insertion configuration * To whom correspondence should be addressed.494 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 199 1 VOL. 6 Coil 0 O / - - _ _ - - Reference point for viewing height measurements I 0 “I 12mm Reference point for insertion mm height measurements Auxiliary gas Plasma gas Conventional torch cross-section Fig. 3 Measurement reference system I 1 I I 4 6 8 10 12 14 16 18 Observation height/mm Fig. 4 Comparison of profiles for ‘hot’ versus ‘normal’ systems using a 2/1 injector tube 2 mm below the load coil (see Fig. 3 for observation height reference point). A Cd ‘normal’; B Cd ‘hot’ C Cu ‘normal’ D. Cu ‘hot’; E Zn ‘normal’; and F Zn ‘hot’ each element using a two point background correction method.Signals were obtained from the average of 30 1 s integrations. A small gap between the alumina injector tube and the graphite plug allowed an auxiliary flow to be introduced. Previous experience with direct sample insertion (DSI)l0 suggested that the graphite a conducting material might affect the plasma. Experiments with graphite extensions (Fig. 1) confirmed that larger masses of graphite affected the power required to start the plasma. In a similar vein an insertion of a 6/3 graphite tube 3 mm into the plasma required 1.5 kW for plasma stability versus 1.0 kW for a 1 mm extension of the same material. Contrary to our DSI experience the graphite extension does not heat within seconds. It reaches an apparent ‘red heat’ after about 1 min depending on the height and mass of the system.Interestin- gly the graphite does not become sufficiently hot that it affects the quartz injector tube on which it is resting. Boron nitride on the other hand becomes sufficiently hot after 15 min that a visible orange Si emission is present in the plasma from the injector tube. The length of the extension material into the plasma will also affect the amount of auxiliary gas required to place the plasma in its conventional position directly above the auxiliary tube. With an extension of 6 mm into the load coil region no auxiliary gas is necessary. Further extensions cause the plasma to move up. The adverse effects of longer graphite extensions led to the development of the tubular system. The tubular system was made of alumina a non- B ,c 10 - 1 oh -I- - - -*- 4 6 8 10 12 14 16 18 30t 20 “4 6 8 10 12 14 16 18 Zn E (4 60 - “ U 4 6 8 10 12 14 16 18 ,Y :‘ H ‘I 4 6 8 10 12 14 16 18 Observation heighvmm Fig.5 Profiles of relative intensity versus observation height. (a)-(c) 6/3 Injector tubes at insertion heights of A 3; B 8; and C 12 mm. (4-m 2/1 Injector tubes at insertion heights of D -2; E 1; F 3; G 5 ; H 8; and I 12 mm. See Fig. 3 for reference pointsJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY SEPTEMBER 1991 VOL. 6 495 8 - 7 - x. c 1 - 0 4 8 12 16 18 1 L Observation heighthm Fig. 6 Reproducibility of the signal for the 2/1 injector tube system using ‘hot’ and ‘normal’ sample introduction systems at various insertion heights A ‘normal’ -2 mm; B ‘hot’ -2 mm; C ‘hot’ + 3 mm; D.‘hot’ + 8 mm; and E ‘hot’ + 12 mm. See Fig. 3 for reference points conducting ceramic material and consequently allowed a 1.0 kW operation; however with insertion heights greater than 8 mm one must apply more power to light the plasma. The amount of auxiliary gas required to position the plasma depends on both the tube size and the insertion height. With the 6/3 tube a gas flow of 0.3 1 min-’ was required at both 6 and 12 mm. In contrast the smaller 2/1 tube required 1.0 1 min-I at 6 mm and 0.3 1 min-’ at 12 mm. Apparently the tubes exert an upward force related to their diameters and insertion heights. The auxiliary flow pushes the plasma upwards in a manner similar to a normal auxiliary flow. There is no apparent sheathing effect on the injector tube. When the injector tube is low in the plasma the appearance of the plasma is as anticipated from normal operation.When the injector tube is high in the plasma the central channel wedges outward slightly and is much darker than the usual central channel. The inserted injector is in part a byproduct of research performed with a heated sample introduction system called the ‘hot’ ~ y s t e m . ~ . ~ The system uses a heated spray chamber and a condenser. Data from what is called the ‘normal’ system have been included for comparison purposes. This system differs from a conventional system only in one respect it has a condenser between the spray chamber and the torch. Fig. 4 contains both ‘normal’ and ‘hot’ system data. For clarity of presentation the ‘normal’ system data have been scaled so that the largest value in each ‘normal’ set matches the largest value in each ‘hot’ set.In both instances the injector is at the conventional position 2 mm below (-2 mm) the load coil using the reference system illustrated in Fig. 3. As previous work demon~trated,~ the differences in performance between this and a conventional system are minimal with the exception that the ‘hot’ system does provide higher signal levels (approximately a factor of 4) due to an increased mass transfer rate.’19 Since all data were obtained under the same conditions (power data acquisition etc.) the primary differences seem to be simply an increase in signal level and a degradation in precision by a factor of approximately 2 as discussed below. For a given element the profiles are fairly similar.This leads to speculation that the results may be comparable to those that would be obtained with a conventional nebulizer-spray chamber system. Height profiles for a range of elements are presented for the 613 tube [Fig. 5(a)-(c)] and the 2/1 tube [Fig. 5(d)-(f)]. It is not surprising that the intensitiey would fall off with deep insertions however it is surprising that the 6/3 system with its large central channel produces intensities equivalent to those of the much narrower 211 system at lower insertion depths. This indicates that some type of pinching effect may be channelling the analyte into the viewing zone. Unfortu- Table 1 Differential pressures Insertion height/ Differential pressure mm above atmospheric/atm -2 + 1 +3 +5 +8 + 12 0.0 17 0.023 0.028 0.029 0.029 0.036 nately no profiles are available to verify this.Certainly this is not what one would expect from simple geometric and optical considerations. Fig. 6 shows a noise profile of the 2/1 system using both ‘hot’ and ‘normal’ sample introduction systems. While the ‘normal’ system is a factor of 2 lower in noise its behaviour is otherwise similar to the ‘hot’ system with -2 and + 3 mm insertion heights. At higher insertion levels 8 and 12 mm the noise distribution is fairly flat indicating an independence between viewing height and noise. This is interesting because it suggests that some noise reduction process probably drying of the droplets is taking place in the tube. Furthermore it indicates that the process takes place more efficiently in the tube than in the plasma.Considering the variety of mechanisms available for energy transfer this does not seem unreasonable. Another noteworthy facet of the inserted injector system is the pressure induced in the sample introduction system. Whenever a gas is forced through an orifice one would expect pressure differentials. When using a 613 system the differential pressure in the spray chamber was very low approximately 0.0025 atm (1 atm= 9.86923 x lo5 Pa). This was used for the flame sample introduction system. Using a 2/1 system the pressures increased as indicated in Table 1. The inserted injector system allows the ICP to be used with sample introduction systems which have large pressure fluctuations. For the flame sample introduction system described previou~ly,~*~ the inserted injector system allowed a system which could not be maintained for more than 30 s to be run continu- ously with approximately 1 O/o relative standard deviation. The delay in introducing the analyte stream into the plasma body may allow other types of experiments which could not be carried out with a conventional system. References 1 Blain L. Salin E. D. and Boomer D. W. J. Anal. A?. Spectrom 1989 4 721. 2 Sing R. L. A. and Salin E. D. Anal. Chem. 1989 61 163. 3 Monasterios C. J. and Salin. E. D. Anal. Chem. 1986 58 780. 4 Habib M. M. and Salin E. D. Anal. Chem. 1985 57 2055. 5 Gervais L. S. and Salin E. D. J. Anal. Ai. Spectrom. 1991,6 41. 6 Usypchuck L. Undergraduate Honours Thesis McGill University 1989. 7 Usypchuck L. Moss P. Karanassios V. and Salin E. D. in preparation. 8 LCgere G. and Burgener P. ICP Inf Newsl. 1985 11 447. 9 Gervais L. S. M.Sc. Thesis McGill University 1989. 10 Salin E. D. and Horlick G. Anal. Chem. 1979 51 2284. Paper 0/04983B Received November 5th 1990 Accepted April 29th 1991