JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1992 VOL. 7 7 Evaluation of a Low-powered Argon Microwave Plasma Discharge as an Atomizer for the Determination of Mercury by Atomic Fluorescence Spectrometry Yixiang Duan Xiangxing Kong Hanqi Zhang Jun Liu and Qinhan Jin* Department of Chemistry Jilin University Changchun 130023 China A low-powered argon microwave plasma torch discharge is introduced into atomic fluorescence spectrometry (AFS) for the first time. The configuration and the characteristics of the plasma for AFS are described in this paper. Some factors influencing the determination of mercury such as the flow rates of the carrier gas and the plasma gas are discussed in detail. The detection limit for mercury by this method is shown to be 3 ppb and the relative standard deviation for a solution concentration of 1 pg ml-i is 2.5% (n=11). The proposed method is relatively free from background interference and gives a large dynamic range.Keywords Microwave plasma torch; atomic fluorescence spectrometry; mercury determination Work on the use of an inductively coupled plasma (ICP) as an atomization cell for atomic fluorescence spectrometry (AFS) has developed rapidly since Montaser and Fassel' suggested that the ICP might be a promising atomization cell for AFS in 1976. In 1981 a commercial ICP-AFS instrument manufactured by the Baird Corporation became available and its performance was described by Demem2 The microwave-induced plasma (MIP) has received con- siderable attention as an atomization cell for atomic emission spectrometry (AES) and atomic absorption spec- trometry (AAS)3*4 during the past decade since the introduc- tion of the Beenakker c a ~ i t y ~ and great success has been achieved using both techniques.Since atomic fluorescence is a hybrid phenomenon of atomic emission and atomic absorption it possesses the characteristics of both AES and AAS but is not limited by the characteristics of either of the techniques. The MIP may also be used as an atomization cell in AFS. More recently Perkins and Long6 reported the first use of a low-wattage MIP as an atomization cell for AFS measurements in which a TMolo cavity was used to produce the plasma and a hollow cathode lamp (HCL) as well as a continuum xenon arc lamp were used for excitation. The preliminary results for some metals were not very satisfactory.The detection limits were from the sub-ppm to ppm level. However these results were greatly improved by their later work7 with a high-efficiency helium MIP as the atomization source in which the detection limits for 14 elements were from the ppb to sub-ppm level. These results are comparable to and even a little better than those obtained with MIP-AES under the same conditions indicating that the MIP has great potential in AFS. In this paper the microwave plasma torch (MPT)* was introduced for the first time into AFS. The configuration of this torch is similar to that of the ICP torch and it has already been applied in AES.8 In the present work preliminary studies of the characteristics of the plasma and its performance as an atomizer for AFS as well as its application in the determination of mercury were carried out. Some other elements such as zinc and cadmium have been also studied with a similar system and detection limits at the ppb and sub-ppb level were achieved respec- tively.Detailed results for these two elements will be reported elsewhere. ~ ~ * To whom correspondence should be addressed. r r r I 1 I Cleaning reagent 4 I - ' Reducing 10 11 - 1 reagent - dr Fig. I Schematic diagram of MPT-AFS system 1 high-voltage supply; 2 preamplifier; 3 detection circuit; 4 computer; 5 PMT; 6 HCLs; 7 lens; 8 MPT; 9 microwave power supply; 10 concentrated H,SO,; 1 1 cold vapour generator; 12 recorder; and 13 power supply Experimental Reagents All chemicals used were of analytical-reagent grade.Water was distilled and de-ionized. The stock solution of mercury (1000 ppm) was prepared from HgC12. This solution was diluted as required for use. The SnC12 solution (about 5%) was prepared by adding 5 g of SnC12.H20 to 5 ml of concentrated hydrochloric acid heating slightly for dissolu- tion and then diluted to 100 ml. Instrumentation and Procedure The block diagram of the MPT-AFS system used is shown in Fig. 1. The apparatus used is listed in Table 1. A specially designed HCL was used in this experiment working in a pulsed mode. The radiation from the mercury HCL was focused with a suitable lens onto the atomization cell for excitation. The resulting fluorescence (at 253.7 nm) was collected at an angle of 45" with respect to the excitation beam.After amplification by a preamplifier and a8 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1992 VOL. 7 Table 1 Apparatus used for MPT-AFS Component MPT Photomultiplier tube High voltage power supply Computer Dual-channel spectrometer Pulsed mercury HCL Microwave power generator ModeVsize Manufacturer - Laboratory built R106 - Beijing Instrumental Factory of Geology Apple I1 Nanjing Electric Instrument Factory XDY-I1 Beijing Instrumental Factory of Geology VEKY-AF The 12th Institute of Mechanics and Electronics Ministry of China DW-2 Beijing Instrumental Factory of Geology Table 2 Operating parameters for MPT-AFS Forward power Reflected power Observation height Current of HCL Photomultiplier tube power supply Flow rate of carrier gas (F,) Flow rate of plasma gas (F,,) Sample volume Reduction agent volume Time for measurement and integration 10-50 W ow 0-36 mm 15-90 mA 300 V 400- 1000 ml min-' 400-1000 ml min-' 2 ml 2 ml 5-15 s detection circuit the fluorescence signal was fed to an analogue-to-digital ( N D ) converter and then printed and stored using a computer.The data presented in this paper are given as background-corrected values. In order to obtain AFS profiles the background signal was subtracted by the computer from the analyte signals. The working curve and inter-element effect plots were treated similarly. Operating Conditions The operating conditions for MPT-AFS are shown in Table 2. The optimum microwave forward power is 50 W with 0 W reflected. The practical forward power was usually much lower than 80 W so it was not necessary to cool the torch with water during operation.The plasma is very easy to ignite by touching a metal rod which was insulated from the operator by a rubber tube to the top of the central tube. Sample introduction was performed by the use of a cold vapour generator (aerator). A 5% SnC1 solution was used as a reducing agent to produce mercury vapour. In order to prevent a large amount of air from entering the plasma and affecting the performance of the plasma the aerator was equipped with a septum. The sample solution was injected into the reaction vessel (aerator) through the septum. The aerator was cleaned with de-ionized water after each measurement. The areas of the fluorescence signals were recorded. The sample turnaround time is about 30 s.Results and Discussion Characteristics of the Plasma The plasma was produced using an MIP similar to that used by Jin et at.* The torch consists of three concentric copper tubes. The dimensions of the outer tube are 22 mm i.d. x 25 mm 0.d. The intermediate tube is of 5.3 mm i.d. x 5.8 mm 0.d. The inner tube is of 1.8 mm i.d. x 2.7 mm 0.d. For 3 2 4 Fig. 2 Microwave plasma torch I first bright zone; 2 second bright zone; 3 plasma tail plume; 4 crossing point; 5 central tube; 6 intermediate tube; and 7 outer tube all experiments the argon (plasma gas) flow is introduced continuously into the intermediate tube to maintain the plasma and the sample gas was introduced into the inner tube along with the carrier gas. The plasma was formed at t:he top of the torch (see Fig.2). The plasma can be optically divided into three zones first bright zone second bright zone and tail plume. The heights of different zones are about 6 5 and 13 mm respectively. The plume is a good observation zone for atomic fluores- cence because it has a very pale colour and the background emission is weak. When forward power is low (1 0 W) the plasma plume is short and there is a little tremble in the plasma. When the forward power is higher than 30 W the plasma becomes stable. Microwave leakage from the MPT discharge was in- spected with a radiation hazard meter and was found to be less than 5 mW cm-2 at a distance of 5 cm from the top of the unshielded torch. Atomic Fluorescence Profiles and Working Curves In order to determine the position of maximum analyte signal in the plasma tail plume atomic fluorescence profiles olf different analytes were studied.Fig. 3 represents the atomic fluorescence profile of a 1 ppm mercury solution at a plasma forward power of 70 W. In this profile the relative fluorescence intensity (Ir) of mercury versus the observation h,eight (H) above the top of the torch is plotted. The maximum intensity of the fluorescence was observed to occur at a height of 16 mm above the top of the torch. AfterJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1992 VOL. 7 9 180 I I 1 ’-I 40 A Him m Fig. 3 Atomic fluorescence profile of mercury versus the observa- tion height above the top of the torch obtained with MPT A mercury atomic fluorescence (at 253.76 nm); and B background under the same conditions as those for fluorescence I I t - A B C D E I 1 I I VS 0 5 10 15 Fig.4 Influence of flow rate of carrier gas on the intensity of mercury fluorescence A 160; B 400; C 600; D 800; and E 1000 ml min-I. F Background this point intensity diminished rapidly in the plasma plume with the increase of height. It was also noted that at the first and second bright zones there exists a strong plasma flicker noise and spectral background. Therefore it is difficult to observe the mercury fluorescence signal in these areas so the plasma tail plume was usually chosen to measure the fluorescence signals. The linearity of the working curve for mercury extends to about three order of magnitude. As can be seen in the range shown (Fig. 3) no obvious self-absorption was observed.Influence of Flow Rate of Carrier Gas It was shown that the fluorescence signals were greatly affected by the flow rate of the carrier gas (Fig. 4). When the flow rate is too low the mercury generated can only be’ carried out of the reaction vessel slowly and incompletely. When the flow rate is too high the concentration of the mercury generated is diluted by the carrier gas and the fluorescence signal is also diminished. The optimum flow rate was found to be 600 ml min-l. The flow rate of the plasma gas is another important factor that affects the intensity of fluorescence and the shape of the plasma. With an increase in the flow rate of the plasma gas the plasma plume became elongated and the optimum observation height was altered. The optimum flow rate of plasma gas was found to be about 600 ml min-I.Influence of Microwave Power The influence of microwave power on the height of the plasma is obvious in the range of lower forward power (Fig. 5). The plasma height is increased with an increase in 700 8oo t \ 3 600 3 500 w .- c - 200 100 d 1 I 1 I I I 1 Microwave powerw 0 10 20 30 40 50 Fig. 5 Influence of microwave power on the height of plasma and the intensity of fluorescence of mercury A height of plasma; B mercury fluorescence intensity; and C background 0 15 30 45 60 75 90 105 120 ifmA Fig. 6 Influence of HCL current on the intensity of fluorescence of mercury microwave power and then levels off when the power is greater than 30 W. The influence of microwave power on the intensity of fluorescence is relatively complex.The intensity of fluorescence decreases with an increase in the microwave power. Influence of HCL Current The influence of the HCL current (i) on the intensity of the fluorescence signal is shown in Fig. 6. The intensity of fluorescence increases sharply with increasing HCL current at first and then levels off. It is obvious that the mercury HCL reaches a maximum emission intensity when the current is increased. Other Influences It has been noted that the reducing agent SnCl is easily absorbed on the wall of the cold vapour generator. It is very difficult to clean up even using a large amount of water. An oxidizing agent (1% KM,04) was used during the course of cleaning to eliminate the influence of SnCl on the system. Recovery An artificial solution containing 60 ppb of mercury and various metal ions including 1 ppm of Zn2+ Ca2+ A13+ and Cu2+ 10 ppm of Mg2+ Fe3+ Na+ K+ Cd2+ and Co2+ and 100 ppb of Pb2+ was used to test the recovery of mercury.The recovery was shown t o be more than 90% for ten measurements. No obvious interfering effect was observed.10 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY FEBRUARY 1992 VOL. 7 Table 3 Comparison of detection limits of several atomic fluores- cence methods for the determination of mercury Excitation Atomization Detection source cell limits (ppb) Ref. Microwave excited electrodeless discharge lamp ICP 10 1 Pulsed HCL ICP 25 9 Low pressure mercury vapour lamp ICP 0.2 10 Pulsed HCL MPT 3 This work Detection Limits The detection limit was measured under the optimum conditions and calculated according to the guidelines of the International Union of Pure and Applied Chemistry.For the calculation 1 1 blank readings were taken. The analytical sensitivities were calculated from the working curve of the element under study. The detection limit for mercury was shown to be 3 ppb. The relative standard deviation for a 1 ,ug ml-l solution was 2.5% (n= 11). A comparison of the detection limits for mercury obtained with different AFS methods is shown in Table 3. Conclusions This work shows that the MPT discharge is promising as an atomization cell for AFS. It exhibits a large dynamic range over a concentration range of several orders of magnitude and is relatively free from background interferences. Fur- thermore its detection limit for mercury is comparable to or even better than that of the ICP with the same excitation source. There is no doubt that MPT-AFS can be used in simultaneous multi-element analysis. A study on the use of HCL with MPT-AFS for a number of other elements and their simultaneous determination is underway. This work was supported by the National Natural Science Foundation of China. References 1 Montaser A. and Fassel V. A. Anal. Chem. 1976,48 1490. 2 Demers D. R. paper presented at the Pittsburgh Conference Atlantic City USA 198 1. 3 Jin Q. Zhang H. Yu S. Spectrosc. Spectral Anal. (Beijing) 1989 9(4) 32. 4 Lin X. Zhang H. Bing G. and Jin Q. ZCP Znf Newsl. 1988 42 1285. 5 Beenakker C. I. M. Spectrochim. Acta Part B 1976,31,483. 6 Perkins L. D. and Long G. L. Appl. Spectrosc. 1988 42 1285. 7 Perkins L. D. and Long G. L. Appl. Spectrosc. 1989,43,499. 8 Jin Q. Zhu C. Borer M. W. and Hieftje G. M. Spectro- chim. Acta Part B 1991 46 417. 9 Demers D. R. and Allemand C. D. Anal. Chem. 1981 53 1915. 10 Lancione R. L. and Drew D. M. Spectrochim. Acta Part B 1985 40 107. Paper I /02398E Received May 22 I991 Accepted September 19 I991