Anal. Chem. 2003, 75, 2105-2111
Exploration of Microwave Plasma Source Cavity Ring-Down Spectroscopy for Elemental Measurements Yixiang Duan*
C-ACS, MS K484, Los Alamos National Laboratory, Los Alamos, New Mexico 87545 Chuji Wang
Diagnostic Instrumentation & Analysis Laboratory, Mississippi State University, Starkville, Mississippi 39759 Christopher B. Winstead
Department of Physics and Astronomy, University of Southern Mississippi, USM Box 5046, Hattiesburg, Mississippi 39406
We are exploring sensitive techniques for elemental measurements using cavity ring-down spectroscopy (CRDS) combined with a compact microwave plasma source as an atomic absorption cell. The research work marries the high sensitivity of CRDS with a low-power microwave plasma source to develop a new instrument that yields high sensitivity and capability for elemental measurements. CRDS can provide orders of magnitude improvement in sensitivity over conventional absorption techniques. Additional benefit is gained from a compact microwave plasma source that possesses the advantages of low power and low-plasma gas flow rate, which are of benefit for atomic absorption measurements. A laboratory CRDS system consisting of a tunable dye laser is used in this work for developing a scientific base and demonstrating the feasibility of the technique. A laboratory-designed and -built sampling system for solution sample introduction is used for testing. The ring-down signals are monitored using a photomultiplier tube and recorded using a digital oscilloscope interfaced to a computer. Lead is chosen as a typical element for the system optimization and characterization. The effects of baseline noise from the plasma source are reported. A detection limit of 0.8 ppb (10-10) is obtained with such a device. Inductively coupled plasma (ICP) sources have been widely used in atomic emission spectrometry and mass spectrometry.1,2 Most work in atomic absorption spectrometry (AAS) utilizes flames and electrothermal devices as atomizers for AAS measurement,3,4 although the use of plasma sources as atomization cells * Corresponding author. E-mail:
[email protected]. (1) Montaser, A.; Golightly, D. W. Inductively Coupled Plasma in Analytical Atomic Spectrometry; VCH Publishers: New York, 1992. (2) Houk R. S.,Anal. Chem. 1986, 58, 97A-105A. (3) Ingle, J. D., Jr.; Crouch, S. R. Spectrochemical Analysis; Prentice Hall Inc.: Upper Saddle River, NJ, 1988. (4) Price, W. J. Spectrochemical Analysis by Atomic Absorption; John Wiley & Sons: New York, 1985. 10.1021/ac0207832 CCC: $25.00 Published on Web 04/02/2003
© 2003 American Chemical Society
for AAS measurement was proposed in the early development stage of plasma spectrometry. In the first report, a multiple beam system for ICP-AAS measurement was described that used a hollow cathode lamp as a radiation source and an ICP as an atomization cell.5 A later work reported detection limits for ICPAAS for determination of Ag, Al, Ca, Cu, Mo, Ta, and V at about ppm levels using a short plasma torch as an absorption cell.6 On the basis of theoretical considerations and calculations, Magyar and Aeschbach suggested that the ICP might not be an ideal source for AAS measurement.7 The following reasons were identified: (1) the high plasma gas flow rate required for maintaining the ICP dilutes the concentration of analyte atoms, resulting in a short residence time of analyte in the plasma; (2) the absorption path length in an ICP is relatively short and thus is not beneficial for AAS measurement; and (3) the high temperatures in the ICP source favor the production of excited and ionized species while AAS needs ground-level populations. Therefore, subsequent research on ICP-AAS has mainly focused on fundamental studies and plasma diagnostics.8-10 Although cavity ring-down spectroscopy (CRDS) has rapidly gained popularity in the molecular spectroscopy community, there are few reports exploring atomic absorption with CRDS.11-14 So far, the only published papers used an inductively coupled plasma as an atomization cell for CRDS measurement.11,12 This exploratory research showed very promising results to adapt CRDS for atomic (5) Wendt, R. H.; Fassel, V. A. Anal. Chem. 1966, 38, 337-338. (6) Vellion, C.; Margoshes, M. Spectrochim. Acta 1968, 23B, 503-512. (7) Magyer, B.; Aeschbach, F. Spectrochim. Acta 1980, 35B, 839-848. (8) Hart, L. P.; Smith, B. W.; Omenetto, N. Spectrochim. Acta 1986, 41B, 13671380. (9) Gillson, G.; Horlick, G. Spectrochim. Acta 1986, 41B, 431-438. (10) Mignardi, M. A.; Smith, B. W.; Winefordner, J. D. Anal. Chem. 1990, 62, 586-592. (11) Miller, G. P.; Winstead, C. B. J. Anal. At. Spectrosc. 1997, 12, 907-912. (12) Wang, C.; Mazzotti, F. J.; Miller, G. P.; Winstead, C. B. Appl. Spectrosc. 2002, 56, 386-397. (13) Winstead, C. B.; Mazzotti, F. J.; Mierzwa, J.; Miller, G. P. Anal. Commun. 1999, 36, 277-279. (14) Spuler, S.; Linne, M.; Sappey, A.; Snyder, S. Appl. Opt. 2000, 39, 2480.
Analytical Chemistry, Vol. 75, No. 9, May 1, 2003 2105
absorption measurement in plasma sources. However, as indicated in previous research,7 using a conventional ICP as an atomization cell for AAS measurement has some limitations. The prevalent theme in the literature has been that the ICP can be a poor source for ground-state neutral species because most of the analyte atoms are either excited or ionized in conventional ICPs. With such knowledge in mind, the most recent work on ICP-CRDS has lowered the ICP power in combination with a modification to the ICP torch design so that ground-state populations of analytes can be significantly enhanced, allowing improved detection limits to be achieved using cavity ring-down spectroscopy.12 These results suggest that continued optimization of the plasma source may significantly improve the sensitivity for atomic absorption measurement with CRDS. Microwave-induced plasma (MIP) is a powerful alternative source for elemental determination and has been extensively used in analytical atomic spectrometry.15 Compared with other types of plasma sources, MIP offers some attractive characteristics, such as its unique features of high excitation efficiency for metal and nonmetal elements, capability of working with various gases, simplicity, and low cost for instrumentation and maintenance. In addition, microwave plasmas can be sustained at fairly low power and low gas flow rate, making them a desirable source for absorption measurement.16 So far, there have been a fair number of publications reporting the use of microwave plasma sources as atomization cells for conventional AAS measurements with hollow cathode lamps.17-20 Additional efforts to improve sensitivity include designing high-efficiency desolvation devices to remove water vapor loading,19 designing various plasma discharges for better absorption measurement,16 adapting different sampling devices,18 and regulating the plasma gas flow system. With conventional lamps as a light source, MIP-AAS yields about 2-3 orders of magnitude lower detection limits than similar ICP-AAS experiments.16 These results encourage us to pursue a new combination of microwave plasma atomizer with cavity ring-down measurement. In this work, we first combine a microwave plasma source with cavity ring-down measurement. The research takes advantage of the extreme absorption sensitivity of CRDS combined with the microwave plasma source advantages of low power, low gas flow rate, and ease of operation. Here we report initial results to be followed by further extensive characterization of MIP-CRDS. In this initial effort, we have tested the feasibility of combining the two technologies and examined the potential of using a microwave plasma source with cavity ring-down measurement. EXPERIMENTAL SECTION Instrument Assembly and Setup. A schematic diagram of the experimental setup is shown in Figure 1. The system consists of five primary parts: laser source, plasma source, ring-down cavity, sampling device, and detection electronics. The optical configuration in this work is similar to the assembly reported in the literature.12 Briefly, a tunable ultraviolet laser beam is gener(15) Jin, Q.; Duan, Y.; Olivares, J. A. Spectrochim. Acta 1997, 52B, 131-161. (16) Duan, Y.; Huo, M.; Du, Z.; Jin, Q. Appl. Spectrosc. 1993, 47, 1871-1879. (17) Ng, K. C.; Garner, T. G. Appl. Spectrosc. 1993, 47, 241-243. (18) Duan, Y.; Li, X.; Jin, Q. J. Anal. At. Spectrom. 1993, 8, 1091-1096. (19) Duan, Y.; Zhang, H.; Huo, M.; Jin, Q. Spectrochim. Acta 1994, 49B, 583592. (20) Duan, Y.; Huo, M.; Liu, J.; Jin, Q. F. J. Anal. Chem. 1994, 349, 277-282.
2106 Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
Figure 1. Schematic diagram of the experimental setup: L1, L2, lenses; M, broadband turning mirror; M1, M2, cavity mirrors; PMT, photomultiplier tube.
ated using a narrow line width, dual grating dye laser (Radiant NarrowScan) followed by frequency doubling (Inrad Autotracker III). A 20-Hz repetition rate Nd:YAG laser (Continuum Powerlite 8020) is used to pump the dye laser. The minimum scanning step of the dye laser is 0.0003 nm, with a line width of ∼0.08 cm-1 at 283.3 nm. A spatial filter system consisting of two focusing lenses and a pinhole is used to approximately mode match the laser beam to the ring-down cavity. Two plano-concave mirrors (R ∼99.72%) with a 6-m radius of curvature (Los Gatos Research) are used to form a 770-mm cavity. This mirror separation results in a stable optical cavity while allowing ample room for inserting various plasma, flame, or other atomization sources. Broadband light emission from the plasma is rejected by a 10-nm band-pass filter prior to the detection of the ring-down signal using a photomultiplier tube (PMT; Hamamatsu R928). The signals are digitized using a digital oscilloscope (Tektronix, TDS 410A) interfaced to a computer. Microwave Plasma Source. For these initial experimental tests, a robust microwave plasma source known as a microwave plasma torch was used as an atomization cell for ring-down measurement. The plasma takes a toroidal shape and can be operated with a power ranging from tens of watts to several hundred watts. The plasma source used here is the same as we used in our previous research for atomic emission measurement.21 The flamelike plasma formed by the torch has been demonstrated to be an excellent stable source for atomic spectrometry.22,23 The plasma torch, consisting of three coaxial tubes, offers some additional advantages over conventional microwave plasma sources and has been widely applied in emission, mass spectrometry, and fluorescence measurements.15 The torch is connected to a 2450MHz microwave power supply through 1-m coaxial cable. The plasma source is mounted inside the optical cavity on a X-Y-Z three-dimensional adjustable stage for precise alignment of optical beams for maximum absorption. Typical operational parameters for the instrument are summarized in Table 1. Sampling Device. Samples are delivered through a commercial peristaltic pump into an ultrasonic nebulizer (U-5000 AT+, CETAC), where the liquid samples are generated into fine, wet (21) Duan, Y.; Su, Y.; Jin, Z.; Abeln, S. Rev. Sci. Instrum. 2000, 71, 1557-1563. (22) Jin, Q.; Zhu, C.; Borer, M. W.; Hieftje, G. M. Spectrochim. Acta 1991, 46B, 417-430. (23) Duan, Y.; Su, Y.; Jin, Z.; Abeln, S. P. Anal. Chem. 2000, 72, 1672-1679.
Table 1. Operational Conditions of the Instrument Plasma microwave power plasma supporting gas flow rate plasma central gas flow rate
120 W 0.35 L/min 0.45 L/min
Sampling sample uptake rate heating temperature of the ultrasonic nebulizer chamber cooling temperature of the ultrasonic nebulizer desolvator heating temperature of the membrane device N2 gas flow rate in the drier sample concentration
80 °C 0.5 L/min 100 ng/mL
Optical Setup laser repetition rate scanning step reflectivity of cavity mirrors filter band-pass
20 Hz 0.0003 nm 99.8% 10 nm
Data Acquisition number of laser pulses for average
20
0.75 mL/min 140 °C
emerged, CRDS typically is based upon the injection of a laser beam through one end mirror of a stable optical cavity. Due to the finite reflectivity of the cavity mirrors, optical absorption, and optical scattering, the intensity of the light trapped in the cavity decays exponentially with time. The time constant of the exponential decay or ring-down lifetime is given by
τ ) d/c(1 - R + σ(ν)nls)
(1)
-5 °C
aerosols through contact with an ultrasonic transducer. The heating temperature inside the ultrasonic nebulizer is ∼140 °C, with a cooling chiller operating around -5 °C at the aerosol outlet. The aerosol generation efficiency by the system is ∼7-10% with a sample uptake rate of 0.75 mL/min. An additional desolvation device is employed to further control the solvent loading in the low-power plasma source and to enhance the system performance. The desolvation device consists of a membrane desolvator and a nitrogen gas stream. A flow rate of ∼0.5 L/min nitrogen is introduced into the space between the desolvation chamber and the membrane tube to remove the water vapor produced during the desolvation process. With this arrangement, a good desolvation efficiency is obtained and the plasma source can operate in a stable manner with heavy aerosol loading. Chemicals and Reagents. High-purity argon (99.999%) was used as the working gas for supporting the plasma source. Sample solutions were prepared by diluting standard solutions (1000 µg/ mL, Absolute Standard Inc, Hamden, CT) and introduced into the plasma source through the sampling system. RESULTS AND DISCUSSION Measurement Principle of Plasma Source CRDS. A number of CRDS reviews24-27 have been published in the years since O’Keefe and Deacon’s first report.28 A CRDS absorption measurement is based upon determining the decay time for light trapped in an optical cavity rather than measuring transmitted light intensity. The technique is relatively unaffected by laser power fluctuation noise, allowing for absolute absorption measurements even using pulsed lasers. Although numerous variations have (24) Scherer, J. J.; Paul, J. B.; O′Keefe, A.; Saykally, R. J. Chem. Rev. 1997, 97, 25. (25) Wheeler, M. D.; Newman, S. M.; Orr-Ewing, A. J.; Ashfold, M. N. R. J. Chem. Soc., Faraday Trans. 1998, 94, 337-351. (26) Busch, K. W., Busch, M. A., Eds. Cavity Ringdown Spectroscopy: An Ultratrace Absorption Measurement Technique; ACS Symposium Series 720; Oxford University Press: New York, 1999. (27) Miller, G. P.; Winstead, C. B. Cavity Ringdown Laser Absorption Spectroscopy. In Encyclopedia of Analytical Chemistry; Meyers, R. A., Ed.; John Wiley and Sons Ltd.: Chichester, 2000. (28) O’Keefe, A.; Deacon, D. A. G. Rev. Sci. Instrum. 1988, 59, 2544-2551.
where c is the speed of light, d is the cavity length, R is the reflectivity of the cavity mirrors, n is the sample density, σ(ν) is absorption cross section at laser frequency ν, and ls is the path length of the light through the sample medium. For a gas uniformly filling the cavity, ls ) d. The term σ(ν)nls is the wellknown exponent from the Beer-Lambert law. By determining the ring-down time with and without a sample in the cavity, absolute absorbance measurements at the level of a few partsper-million fractional absorbance can be readily obtained. In the present configuration, the plasma source does not fill the entire cavity length. Optical losses due to scattering from the plasma region and the air in the open cavity are accounted for by substituting a lower effective reflectivity Reff for R in eq 1. This effective reflectivity is determined by measuring the ring-down time without an analyte present (blank). The absorbance can be rewritten as
absorbance )
(
)
d 1 1 c τ τ0
(2)
where τ and τ0 are the measured ring-down times with an analyte and blank solution, respectively. Optimization of the Experimental Parameters. In this study, the optimization work has been performed based on varying only one parameter at a time when all other parameters were fixed. To find relative optimum conditions for each parameter, a preexamination was performed to determine the values that are proper as the basis for further examination. All major parameters, including microwave power influence, carrier gas flow rate, plasma gas flow rate, and observation height, were optimized before any further exploration. For comparison purposes with the ICP plasma source in our previous work, lead was chosen for this optimization work and performance evaluation. (1) Plasma Power Influence. The plasma source can be operated from as low as 50 to ∼300 W, which is the highest power we can use with this particular power supply. In this wide range, as shown in Figure 2, we found that the lead signal increases with microwave power from 50 to ∼100 W, and then the signal tends to level off from about 100 to 180 W. Increased power may contribute to expansion of the plasma volume, but presumably the increased path length for absorption measurement is compensated for by the decreased analyte number density. Further increase of the power decreases the lead signal significantly. This hints that high power may benefit excitation and ionization states and may reduce atom populations in the ground state. In this point of view, a relatively low power is preferred for the atomic absorption measurement. From this initial characterization, we chose 120 W as a baseline operating power since the absorbance Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
2107
Figure 2. Microwave power influence on the signal intensity: Lead concentration, 100 ppb; plasma observation height, 6 mm above the top of the torch; plasma gas flow rate, 0.35 L/min; carrier gas flow rate, 0.45 L/min. Other experimental conditions are the same as in Table 1. Figure 4. Plasma support gas flow rate influence on signal intensity. Experimental conditions are the same as in Table 1.
Figure 3. Sample carrier gas flow rate influence on lead signal intensity: sample concentration, 100 ppb; microwave power, 120 W. Other experimental conditions are the same as in Figure 1.
is highest at this power, though not significantly. This power is used for subsequent measurements unless otherwise stated. (2) Carrier Gas Flow Rate. Another important parameter for system performance is carrier gas flow rate, which serves to bring samples into the plasma. Lower carrier gas flow rate is good for a longer analyte residence time but carries less sample into the plasma, which is not favorable for absorption measurement. From this point of view, a higher carrier flow rate should be good for introducing more sample; however, if the flow rate is too high, it will significantly reduce the analyte residence time and, therefore, decrease signal intensity. This phenomenon is evident in Figure 3, which gives an optimum value at ∼0.45 L/min. From Figure 3, we can also see that carrier gas flow rate is critical as the curve shape is sharp before and after the maximum peak. Thus, we need to carefully control the flow rate in order to obtain the best sensitivity. (3) Plasma Support Gas Flow Rate. The plasma support gas is critical for plasma operation. A stable plasma usually needs a particular flow rate to operate. A low flow rate may cause plasma instabilities and influence the system performance. Figure 4 depicts plasma support gas flow rate influence on the lead signal intensity. As shown in the figure, a signal maximum occurs at a flow rate of ∼0.35 L/min. Further increase of plasma gas flow rate obviously decreases the signal. The asymmetric shape of the curve in Figure 4 demonstrates that higher flow rate sharply lowers signal intensity when the flow rate is over 0.35 L/min. This phenomenon can be explained by the possibility that higher plasma support gas flow rate may influence the analyte concentra2108 Analytical Chemistry, Vol. 75, No. 9, May 1, 2003
Figure 5. Lead line shape at 283.3 nm. Scan was performed at an observation height of 9 mm through the center of the plasma source with a step width of 0.0003 nm: microwave power used, 120 W, lead concentration, 250 ppb. All other experimental conditions are the same as in Table 1.
tion in the plasma through gas dilution of the analytes. For this reason, we chose 0.35 L/min as a baseline value for plasma support gas flow rate in our further experiments. Analyte Line Shape. Lead line shapes were obtained at different observation heights by scanning the laser over a narrow wavelength range. Figure 5 shows a typical line shape recorded through the center of the plasma at an observation height of 9 mm. Background absorption from a blank solution was used as a correction. The scanning range for the lead line (283.3 nm) is ∼0.03 nm, and the scanning step is 0.0003 nm. At each data point, 100 laser shots were averaged. A 3-5 point-smooth was applied to the data before plotting. Previous CRDS measurements demonstrated that an ICP plasma is more stable and the baseline noise is relatively smaller at lower observation positions than that at higher positions.12 At each position in the microwave plasma, the baseline noise is significantly lower than that identified in the ICP plasma source. The absorption line width is determined by both instrument and physical broadening.29,30 The instrument broadening results from the laser line width (
