A Low-Power, Atmospheric Pressure, Pulsed Plasma Source for

Nov 30, 2000 - The possibility of constructing a compact and low-power plasma detector has been pointed out;7,8 however, to our best knowledge, there ...
1 downloads 9 Views 82KB Size
Download PDF
Anal. Chem. 2001, 73, 360-365

A Low-Power, Atmospheric Pressure, Pulsed Plasma Source for Molecular Emission Spectrometry Zhe Jin, Yongxuan Su, and Yixiang Duan*

Chemical Science and Technology Division, CST-9, MS K484, Los Alamos National Laboratory, Los Alamos, New Mexico 87545

A low-power, plasma source-based, portable molecular emission detector is described in this paper. The detector employs a pulsed-plasma source operated at atmospheric pressure for molecular fragmentation and excitation. The plasma was generated with a home-built high-voltage pulsed power supply. The average operational power of the detector was less than 0.2 W. The effects of operational parameters such as plasma gas, voltage, and plasma gas flow rate were investigated. Molecular emission spectra of a variety of organic compounds were studied. The features of the emission spectra obtained with the pulsed plasma source were significantly different from those obtained with direct current (dc) discharge at a power higher than 10 W. The spectra obtained in this work showed strong CH emission at 431.2 nm; however, the typical CN emission observed with a conventional dc plasma source at 383-388 nm was very weak in most cases. The strong CN emission was only obtained for compounds containing nitrogen, such as aniline. Dimethyl sulfoxide can be detected at a limit of 200 ppb using helium plasma by observing the emission band of the CH radical. The detector was very stable and did not experience electrode fouling even with the introduction of organic vapors. Such a detector is very promising for organic vapor detection. Innovative, field-portable monitoring technologies are required in the chemical industry, national defense, and environmental protection in order to obtain real-time data on chemical emissions in air, identify the sources of chemicals, and reduce or eliminate the emissions of toxic chemicals. Large platform analytical equipment such as gas chromatography (GC), high-performance liquid chromatography (HPLC), mass spectrometer (MS), inductively coupled plasma atomic emission spectrometer (ICP-AES), ICPMS, and GC/MS have been used to analyze chemical emissions in the environment. However, most of the equipment currently commercially available is large, expensive, and not suitable for real-time field use. Therefore, there is a strong need to develop miniature, field-portable analytical instruments. * Corresponding author: (fax) 505-665-5982; (e-mail) [email protected].

360 Analytical Chemistry, Vol. 73, No. 2, January 15, 2001

By employing a separation technique such as capillary gas chromatography, it is possible to develop a compact separation instrument. A field-portable gas chromatography/mass spectrometer was developed by Sinha and Guinikov.1 A short microbore column 3 m in length and 50 µm in inner diameter was used for separation. Such a design lowers the flow rate of carrier gas to 0.05 atm cm3 min-1 and significantly reduces the weight and power needs of the mass spectrometer, thus making the system portable. However, the cost of the system, which is much higher than GC equipped with other detectors such as flame ionization detectors or electron capture detectors, significantly limits its applications. Plasma-based emission spectrometry was first used in a GC detector by McCormack et al.2 A microwave-induced plasma (MIP) was used as an energy source for molecular fragmentation and excitation. Detection limits as low as 10-16 g/s were obtained for hexane. Because these emission detectors are very sensitive and able to simultaneously detect a wide variety of analytes, their use has been studied recently.3,4 Eijkel et al.3 prepared a direct current (dc) microplasma detector on a glass chip. Plasma was generated in a chamber of 50-nL volume. Methane could be detected to 600 ppm. Because of the sputtering of cathode material at reduced pressure (130 Torr), the lifetime of the detector was limited to ∼2 h. The emission intensity also showed considerable quantitative variation; therefore, the microplasma detector was not suitable for practical use in GC. A low-power microwave plasma detector was recently developed by Engel et al.4 A plasma with a longitudinal extension of 2-3 cm was generated with a forward power of 10-40 W. Because the power consumption is low, it is possible to operate the detector with a semiconductor microwave source powered by a car battery. Because of the low operational power, the rotational temperature of the plasma was reported to be ∼650 K; thus, the plasma has very low tolerance for water-loaded aerosols. Additionally, gas pressure and sample vapor clouds had a strong effect on plasma characteristics. Therefore, the performance of the detector would strongly depend on the sample types. (1) Sinha, M. P.; Gutnikov, G. Anal. Chem. 1991, 63, 2012-2016. (2) McCormack, A. J.; Tong, S. C.; Cooke, W. D. Anal. Chem. 1965, 37, 14701476. (3) Eijkel, J. C. T.; Stoeri, H.; Manz, A. Anal. Chem. 1999, 71, 2600-2606. (4) Engel, U.; Bilgic, A. M.; Haase, O.; Voges, E.; Broekaert, A. C. Anal. Chem. 2000, 72, 193-197. 10.1021/ac000678x CCC: $20.00

© 2001 American Chemical Society Published on Web 11/30/2000

Figure 1. Schematic diagram of the experimental setup.

Direct current,5-6 alternating current,7 and high-voltage pulsed8 plasma detectors have been developed. These detectors are simple to build and consume less power than other plasma sources. The possibility of constructing a compact and low-power plasma detector has been pointed out;7,8 however, to our best knowledge, there has been no report on a compact plasma source-based emission detector that can be operated with an energy source such as alkaline batteries. This work describes a new low-power pulsed plasma detector for molecular emission spectrometry. This detector is simple, low cost, and sensitive to organic vapors. The detector was designed to have very low power consumption. A home-built, high-voltage pulsed power supply that can be operated with alkaline batteries was used to generate the plasma for molecular fragmentation and excitation. The influence of plasma gases, flow rates, and discharge voltages on the performance of the detector was studied. The reproducibility and sensitivity of the detector to organic vapors was studied using dimethyl sulfoxide (DMSO). EXPERIMENTAL SECTION Detector and High-Voltage Pulsed Power Supply. A schematic diagram of the detector is shown in Figure 1. Two platinum plate electrodes (0.04 mm in thickness and 3 mm in width) were placed face to face inside Teflon tubing (40 mm long and 7 mm o.d.). The distance between these two discharge electrodes was ∼1.5 mm. A high-voltage pulse was applied to the electrodes for atmospheric pressure plasma generation. Epoxy was used to seal the outside wall of the Teflon tubing to prevent gas leakage. To reduce the weight of the power supply, a small power transformer (T1, primary voltage 230 V; secondary voltage 6.3 V) was used to generate high voltage. The transformer and driver circuitry for generating the high-voltage pulse was housed in a shielded 21 cm × 15 cm × 5.5 cm aluminum box. The total weight of the pulsed power supply is ∼3 lbs. High voltage is generated with a transformer in a manner similar to that of an ignition coil8 and a neon transformer.9 To generate high voltage, the secondary wires of the transformer were used as primary wires, and the primary wires were used as secondary wires. The experimental (5) Braman, R. S.; Dynako, A. Anal. Chem. 1968, 40, 95-106. (6) Decker, R. J. Spectrochim. Acta 1980, 35B, 19-31. (7) Costanzo, R. B.; Barry, E. F. Anal. Chem. 1988, 60, 826-829. (8) Wentworth, W. E.; Vasnin, S. V.; Stearns, S. D.; Meyer, C. J. Chromatographia 1992, 34, 219-225.

parameters such as charging time and discharging time were modulated with a home-built pulse generator based on a 555 timer, which is a device for generating accurate time delays or oscillation. A low voltage from a 15 V dc power supply or alkaline batteries was applied to the transformer (see Figure 1), and a high-voltage output was obtained during charging and discharging cycles, which were controlled by the pulse generator and a power transistor (Q1). The adjustable parameters that are important for discharging performance include charging time (Tc), discharging time (Td), and voltage applied to the primary wires (Epri). The operating frequency ranged from 350 to 1000 Hz. A stable pulsed plasma source could be generated with this power supply. Emission Spectrum Measurement. Plasma was generated using plasma gas such as helium, argon, or nitrogen when a highvoltage pulse was applied to the discharge electrodes. The emission spectrum from the plasma was monitored with an Ocean Optics PC2000 spectrometer system (Dunedin, FL) as shown in Figure 1. The light collection system included a collimating lens, which converts divergent beams of light into a parallel beam and focuses the beam into an optical fiber. The spectrometer system was optimized for the wavelength range of 200-480 nm. A Sony ILX511 liner CCD-array detector with 2048-element pixels was used for light detection. A 200-µm-diameter fiber and 2400 grooves/nm gratings were used for the instrument. A Windowsbased OOIChem operating software was used for data acquisition and signal processing. A notebook computer with a 100-kHz sampling frequency DAQ-700 card (National Instruments Inc.) was used to show real-time spectrum and store data. An interface cable was used to make a connection between the spectrometer and the computer. Discharge Voltage and Current Measurements. A HP 54520A oscilloscope (Palo Alto, CA) was used to monitor discharge parameters. The discharge voltage was measured with a 100:1 voltage divider, and the current was determined by measuring the voltage drop across a 100 Ω resistor with the ground return. Organic Vapor Introduction. Two different methods were used for organic vapor introduction. Since DMSO has a vapor pressure of 0.4 Torr at 20 °C, helium gas saturated with DMSO vapor could be introduced directly into the plasma without affecting the plasma stability. However, when helium gas saturated with other highly volatile organic vapors such as methanol or acetone was continuously fed into the plasma, the plasma became unstable. When helium saturated with ethanol was used, plasma could not be generated at a discharge pulse of 4000 V, which required that these volatile compounds be introduced into the plasma by a different method. First, a piece of Tygon tubing was rinsed with a volatile organic solvent. The Tygon tubing with the organic solvent adsorbed into it was then connected to the plasma gas line. Helium gas passed through the gas line, carrying the organic compounds adsorbed on the surface of the Tygon tubing into the plasma. Using this method, volatile organic compounds can be introduced and the plasma can be maintained. Caution: aniline is highly toxic and care should be taken in the handling, analysis, and disposal of aniline and its oxidation products. Chemicals and Reagents. High-purity argon, helium, and nitrogen (99.999%, Trigas Industrial Gases) were used as plasma gases. Methanol, ethanol, acetone, dichloromethane, aniline, and DMSO (Aldrich, Milwaukee, WI) were used as received. Analytical Chemistry, Vol. 73, No. 2, January 15, 2001

361

Figure 2. Typical background of emission spectrum of the pulsed helium plasma: He flow rate, 3.0 L/min; discharge frequency, 830 Hz; Tc, 1100 µs; Td, 100 µs; Epri, 4.55 V; integration time, 200 ms; spectrum average number, 10.

All the experiments were performed at room temperature and atmospheric pressure. RESULTS AND DISCUSSION Plasma Gas and Plasma Generation. The no-load output voltage of the home-built, high-voltage pulsed power supply ranges from 2000 to 4000 V. The pulse parameters of 1100 µs charging time (Tc) and 100 µs discharging time (Td) were commonly used. With an input voltage (Epri) of 2.5 V, a pulse height of ∼2000 V was obtained. At a pulse height of 2000 V, bright plasma was instantly observed when argon or helium was used as the plasma gas at a flow rate of 0.5 L/min. However, no discharge was observed when nitrogen was used as the plasma gas under the same conditions. A stable nitrogen plasma could only be generated when the Epri was further increased to 7.92 V, corresponding to a pulse height of ∼3000 V. The operational power increases linearly with Epri up to 5.53 V and then levels off with a further increase of Epri. The average operational power ranges from 0.04 W at an Epri of 2.5 V to 0.16 W at an Epri of 7.91 V. To reduce the power consumption of the detector, either argon or helium was needed as the plasma gas. Because helium has a higher metastable-state energy for excitation and ionization of organic compounds, we selected it as the plasma gas for our studies. Two Duracell size D batteries (1.5 V) were used to power the transformer circuitry. A stable helium plasma was maintained for more than 10 h in our preliminary test with a He flow rate of 0.1 L/min, a pulse frequency of 670 Hz, and Epri of 1.5 V. Since a single size D battery has a capacity of 15 000 MAH and the maximum average discharge current in the experiment is ∼50 mA, the operational time was estimated to be ∼600 h with two size D alkaline batteries. However, because a voltage regulator and additional circuitry were used to adjust the output voltage, the practical operational time should be shorter than the estimated value due to the additional power consumption by these added electric components. Plasma Background. Figure 2 shows the emission spectrum of the atmospheric pressure helium plasma for the wavelength 362 Analytical Chemistry, Vol. 73, No. 2, January 15, 2001

region between 240 and 500 nm. The spectrum shows typical helium emission lines and significant molecular bands of OH, NH, N2, and N2+ species10-13 due to the presence of impurities such as water, N2, O2, and hydrocarbons. The C(I) atomic line at 247.86 nm,10,11,13 which is commonly observed in other plasmas such as microwave-induced plasma, is absent. This result may be due to several factors, such as high purity of helium, low discharge power, and short residence time (6.9 ms at a plasma gas flow rate of 1 L/min). To simplify the design of the detector, no glass or quartz optical window was placed in the gas outlet of the detector; therefore, back-diffusion of air into the detector at low flow rates is expected. As shown in Figure 3, at a flow rate of 0.5 L/min, the spectrum is dominated by the molecular emission bands of N2 between 340 and 360 nm. When the flow rate is increased, the intensity of N2 bands decreases gradually. However, even at a flow rate of 5.0 L/min, weak N2 bands still exist. Since it is very unlikely that the N2 bands at such a high flow rate are contributed by the backdiffusion of air into the detector, the nitrogen bands are probably due to the trace impurities of N2 in helium.14 The influence of the back-diffusion of air can be reduced by using a higher plasma gas flow rate, reducing the inner diameter of the detector, or shielding the detector gas outlet with an optical window. Current/Voltage Characteristics. The impurities introduced into the detector through back-diffusion can affect the discharge characteristics. To minimize the effect of impurities on the discharge voltage and current measurements, it is best to reduce the amount of impurities in the plasma. We can reduce the influence of impurities introduced by back-diffusion of air by increasing the flow rate of helium to above 3 L/min. Figure 4 shows the current/voltage curve at a flow rate of 4.0 L/min. The drop of plasma voltage with increasing plasma current was reported for a dc microplasma due to heating of the plasma gas.3 In this work, the discharge voltage increases with the discharge current (see Figure 4). The characteristics of current and voltage observed for the detector indicate that the plasma gas temperature is low. The gas temperature can be estimated from the rotational temperature, which is measured from the intensity distribution of the rotational lines in the OH bands.15 The rotational temperature of the current plasma source was determined to be 320 ( 20 K. The low gas temperature can be associated with the low power used for the discharge and relatively high gas flow rates. Because of the low gas temperature, the surface of the detector body did not become warm to the touch even after continuously running the detector for several hours. Plasma Stability. A stable helium plasma can be generated with an Epri in the range of 2.5-9 V independent of pulse width and pulse frequencies. The intensity of the molecular emission of nitrogen at 357.69 at 0.5 L/min is strongly dependent on the input voltage. At Epri ) 2.5 V, a stable helium plasma was obtained; however, the emission intensity was low. The emission intensity (9) Yu, T.; Winefordner, J. D. Spectrosc. Lett. 1988, 21, 465-476. (10) Estes, S. A.; Uden, P. C.; Barnes, R. M. Anal. Chem. 1981, 53, 1829-1837. (11) Patel, B. M.; Heithmar, E.; Winefordner, J. D. Anal. Chem. 1987, 59, 23742377. (12) Rice, G. W.; D′Silva, A. P.; Fassel, V. A. Spectrochim. Acta 1985, 40B, 15731584. (13) Quimby, B. D.; Sulliva, J. J. Anal. Chem. 1990, 62, 1027-1034. (14) Ogino, H.; Seki, T. Anal. Chem. 1997, 69, 3636-3640. (15) Ishii, I.; Montaser, A. Spectrochim. Acta 1991, 46B, 1197-1206.

Figure 3. Effect of helium gas flow rate on the background spectra: discharge frequency, 830 Hz; Tc, 1100 µs; Td, 100 µs; Epri, 4.55 V; integration time, 200 ms; spectrum average number, 10.

Figure 4. Current/voltage curve of the pulsed helium plasma detector at atmospheric pressure: helium flow rate, 4.0 L/min; discharge frequency, 830 Hz; Tc, 1100 µs; Td, 100 µs.

of N2 at 357.69 nm reached its maximum at Epri ) 5 V. There was no further increase in emission intensity when Epri was increased above 5 V. The stability of the detector over a period of 76 min was examined by measuring the background emission intensity at 431 nm, He(I) at 388.86 nm, and N2 at 357.69 nm. Figure 5 shows the change of emission intensity versus operating time. There is only a slight drift in background, N2, and He(I) signals. The relative standard deviation (RSD) for the emission intensity was 1.6% for background, 2.7% for N2, and 3.3% for He(I).

Figure 5. Stability of the pulsed helium plasma source: He flow rate, 0.5 L/min; discharge frequency, 830 Hz; Tc, 1100 µs; Td, 100 µs; Epri, 5.06 V.

Organic Vapor Detection. Emission Characteristics. The emission spectra of organic compounds in dc discharge5 and MIP2 showed prominent bands of CN and CH in the region between 350 and 450 nm without deliberate addition of nitrogen into the plasma. Trace impurities in the plasma gas can produce an intense CN band at 388.3 nm.2 The typical emission spectra of organic vapors obtained with the pulsed helium plasma are shown in Figure 6. All compounds show the typical CH emission band at 431.3 nm and its rotational fine structure. However, the CN band system16 from 383 to 388 nm is very weak for most compounds except aniline. Previous work5 indicated that the CN band at 388.3 nm observed for n-hexane samples in dc discharge was signifiAnalytical Chemistry, Vol. 73, No. 2, January 15, 2001

363

species, the emission of CN is closely related to experimental conditions such as operational power. In helium plasma, helium ions (21.2 eV) and metastable helium species (19 eV) have sufficient energy to break molecular bonds; therefore, CN can be directly produced from nitrogen-containing organic compounds such as aniline. The high emission intensity of CN for aniline is tentatively attributed to the direct formation of CN under the impact of highly energetic helium ions and metastable helium species. The formation of CN from an organic compound in helium plasma was represented by the following reactions:5

CH + N2 f CN + NH

(1)

NH + CH f CN + H2 (or 2H)

(2)

2C + N2 f 2CN

(3)

2C + 2NH f 2CN + H2 (or 2H)

(4)

cantly affected by discharge power, and the CN emission intensity was small at relatively low power levels. Because the average power used for our plasma detector was less than 0.2 W, which is much lower than that employed in dc discharge,5 only very weak CN emission peaks were observed at 383-388 nm in most cases. For an organic compound that does not contain nitrogen, the formation of CN species can be attributed to chemical reactions following fragmentation of the compound into C and CH species. Because multistep reactions are involved in the formation of CN

reactions 1 and 2 were believed to be the major source of CN, and reactions 3 and 4 were relatively less important.5 However, while we observed relatively intense CH, N2, and NH emissions with the low-power helium plasma, we did not observe significant CN bands. It seems that there is no direct correlation between CH and CN; therefore, we believe that CN formation is mainly through reactions 3 and 4 rather than reactions 1 and 2. Carbon Deposition. Reagents gases such as oxygen, hydrogen, nitrogen, or air are usually added to the helium plasma in order to prevent or reduce carbon deposition inside a plasmabased detector.17,18 In this work, we studied whether there was significant carbon deposition inside the detector following organic vapor introduction. Helium gas saturated with DMSO vapor was continuously delivered to the detector for more than 30 min. We observed very little carbon deposition inside the detector. Several factors can account for low carbon deposition in the newly developed plasma detector. The plasma gas temperature is only ∼320 K. At such a low temperature, it is difficult to produce carbon from organic vapors. Additionally, there is no optical window enclosing the gas exit of the detector, causing a back-diffusion of air into the detector, which reduces or eliminates carbon deposition. Finally, the short residence time of the analyte in the detector may also lessen carbon deposition. Analytical Performance. The back-diffusion of air into the detector generates no observable signal at 431.2 nm; therefore, the emission lines originating from the most common compounds in air do not interfere with organic vapor detection. A molecular emission detector based on pulsed discharge helium plasma source is well suited for trace organic vapor detection in air. The reproducibility for organic vapor detection was studied. The CH emission was monitored at 431.2 nm following DMSO vapor injections into the helium plasma (Figure 7). The RSD for the emission intensity of CH was calculated to be 4.5% (N ) 21). The detection limit for DMSO in ppm (v/v) was calculated from the emission intensity of 61.6 ppm DMSO as 2N/S using helium plasma,3 where N is the noise level of the plasma

(16) Sutton, D. G.; Westberg, K. R.; Melzer, J. E. Anal. Chem. 1979, 51, 13991401.

(17) McLean, W. R.; Stanton, D. L. Penketh, G. E. Analyst 1973, 98, 432-442. (18) Ebdon, L.; Hill, S.; Ward, R. W. Analyst 1986, 111, 1113-1138.

Figure 6. Typical emission spectra of various organic compounds obtained with the pulsed helium plasma detector: He flow rate, 1.0 L/min; sischarge frequency, 830 Hz; Tc, 1100 µs; Td, 100 µs; Epri, 6.92 V; integration time, 500 ms; spectrum average number, 5.

Figure 7. Investigation of the stability of the pulsed helium plasma detector for a 60 ppm DMSO: He flow rate, 0.5 L/min; integration time, 100 ms; discharge frequency, 830 Hz; Tc, 1100 µs; Td, 100 µs; Epri, 5.06 V.

364

Analytical Chemistry, Vol. 73, No. 2, January 15, 2001

background at 431.2 nm and S is expressed as the emission intensity at 431.2 nm per ppm of DMSO. A detection limit of 0.20 ppm for DMSO could be achieved. This result is much better than that reported for methane (600 ppm) in a dc microplasma based on the same CH band.3 CONCLUSIONS A molecular emission detector has been developed based on a pulsed helium plasma source. The detector has an average operating power less than 0.2 W. Because of the low power requirement, the detector can be operated with two 1.5-V alkaline batteries for more than 10 h. This low-power pulsed plasma detector was evaluated for organic vapor detection. Organic vapors can be detected by monitoring the CH emission at 431.2 nm. Typical emission bands of C and CN originating from organic compounds are very weak for organic compounds that do not contain nitrogen. The advantages of the pulsed plasma detector

include (1) low power consumption (