Journal of the Air & Waste Management Association, 48, 77-81 (1998).

Real Time, Ultrasensitive Monitoring of Air Toxics by Laser Photoionization Time-of-Flight Mass Spectrometry

Marco J. Castaldi and Selim M. Senkan^*, Department of Chemical Engineering, University of California, Los Angeles, CA 90095 (Submitted: October 15, 1996, Accepted: February 12, 1997)

^* Corresponding author, (310) 206-4106, senkan@seas.ucla.edu

Abstract

Gas phase concentrations of individual polycyclic aromatic hydrocarbons (PAH) were measured in real time in combustion products from a co-flow diffusion flame using laser photoionization (LP) time of flight mass spectrometry (TOF/MS). In particular, a naphthalene detection sensitivity of 4 parts per billion (ppb) was demonstrated. The use of calibration mixtures with argon indicated the feasibility of naphthalene detection at about 45 parts per trillion (ppt) at a signal to noise (S/N) ratio of 20. This suggests the possibility of low ppt level detection at a S/N of 1. The novelty of the system is the use of a heated sampling probe and a continuously purged, heated pulse valve that was positioned close to the ionization zone, thereby allowing the generation of photoions in the high-density region of the sample jet where the concentrations of PAH are high. Because the system developed allows for the real time detection of select species, it represents a useful tool in Continuous Emissions Monitoring (CEM) for environmental compliance as well as direct process control.

Introduction

Polycyclic aromatic hydrocarbons (PAH) are the largest single class of chemical carcinogens known today¹, and are generated primarily by combustion processes². The latter include stationary sources such as power generation plants, boilers, and heaters as well as mobile ones that include gasoline and diesel engines. In addition, PAH are often associated with soot production, which itself is an environmental hazard and its formation represents a loss in combustion efficiency.

At present, virtually all PAH determinations are made via the "pre-concentration" method^3-5. In this approach, gases bearing trace PAHs are withdrawn from combustion processes using probes and passed over a variety of organic adsorbents, most common ones being XAD (styrene-divinylbenzene polymer) resins and Tenax (2,6-diphenyl-p-phenylene oxide polymer), for extended periods of time (anywhere from 10 min to several hours, depending on conditions), thereby accumulating enough PAH so that they can be detected by available analytical instruments. The PAHs captured on the adsorbates are then recovered by Soxhlet extraction using solvents such as methylene chloride, cyclohexane, benzene, toluene and ether, with each solvent providing for the improved extraction of select PAHs⁶. The extraction process typically takes anywhere from 6-48 hrs to ensure the recovery of most of the PAHs. The extracts are then concentrated, internal standards added, and analyzed using Gas Chromatography/ Mass Spectrometry (GC/MS) or Liquid Chromatography/ Fluorescence (LC/F).

As a consequence of the many steps involved, the "pre-concentration" techniques are far too slow, expensive and possess considerable uncertainties, and as such are not suitable for the continuous monitoring of these pollutants in combustion effluents. Thus, there is a need for the development of sensitive, fast, reliable and cost effective techniques.

Laser photoionization (LP) coupled to time of flight mass spectrometry (LP-TOF/MS) provides an effective solution to these problems and provides near-instantaneous and sensitive detection of individual PAHs and has been used in the literature^7-10. However, these methods were not broadly applicable and the reported sensitivities of detection were on the order of hundreds parts per billion (ppb) to parts per million (ppm), thus were not particularly suitable for continuous monitoring of PAH emissions from practical combustion systems.

It should be noted that several optical techniques, such as laser induced fluorescence (LIF) and photoionization detection (PID), have also been developed over the years to monitor PAH concentrations in the gas phase^11-14. However, these approaches provide information only on the total amount of PAHs present in the system, without speciation. This is a significant drawback because only a select number of PAHs, such as dibenz(a,h)-anthracene, benzo(a)pyrene and indeno(123,cd)pyrene induce the greatest health risks¹. Consequently, total PAH measurements are an insensitive marker to assess the risks induced by combustion devices.

In recognizing these problems, we developed a highly sensitive, real-time detector to monitor PAH in hot combustion products by exploiting Laser Photoionization (LP) and Time of Flight Mass Spectrometry (TOF/MS).

Experimental Approach

A sketch of the LP-TOF/MS system developed is shown in Figure 1 together with the co-flow burner used in the flame experiments. The burner was made from three concentric stainless steel tubing, with the fuel flow through the inner tubing that was 1 cm. in diameter.

Figure 1. Sketch of the Laser Photoionization Time of Flight Mass Spectrometer System Developed for the Real Time, Ultrasensitive Monitoring of Process Emissions.

The fuel flow was surrounded by the air flow in the annular region and argon shield gas through the outermost tube. Flame sampling was accomplished by withdrawing gases from within the flame along the central axis of the burner using a heated 6 mm OD quartz microprobe having an orifice diameter of about 0.150 mm at its tapered tip. A heated quartz wool filter was used to remove soot and particulate matter from the gas stream. Because of the possibility of PAH to adsorb onto soot surfaces and thus remain on the quartz filter, we kept the filter at a minimum temperature of 200^oC and sub-atmospheric pressure. This approach will allow more than 95% of phenanthrene to desorb off the filter and remain in the gas phase². The samples were transferred to the LP-TOF/MS system through silica lined steel tubing and were introduced into the vacuum chamber via a pulse valve through a 0.3 mm diameter orifice. The pulse valve was a commercially available solenoid unit (General Valve Series 9, NJ) that was modified to provide continuous purging of the valve orifice area by welding a pair of 1/16" nickel tubing to the valve body. The sampling system, which includes the probe, transfer lines and the pulse valve was maintained at temperature between 150^oC-250^oC and at slightly sub-ambient pressures to minimize the condensation and/or adsorption of PAHs on surfaces. The feasibility of this sampling technique for PAH determination has recently been demonstrated¹⁵.

The TOF system consists of two vacuum chambers: One for the generation of the pulsed jet and photoions and the other for ion drift. The first chamber was pumped with a 1000 lt/s turbomolecular pump (Varian TMP1000) and the drift tube by 300 lt/s turbomolecular pump (Varian HT300). The pressure in the first and second chambers were maintained at about 2x10^-5 and 1x10^-6 Torr, respectively during the experiments. The two chambers were separated by a 1 cm diameter orifice through which ions were transported. This orifice was covered with a high transmission wire mesh. The TOF/MS system is based on the Wiley-McLaren design¹⁶ which was equipped with a reflectron (R.M. Jordan Co, CA).

Sample gases passing through the pulse valve were then jetted between the repeller and extractor plates of the TOF/MS in 300 µs pulses at 2-10 Hz repetition rate using a controller (Iota One, General Valve, NJ). One of the unique features of our system has been the close proximity of the pulse valve to the repeller and extractor plates. This called for a novel repeller and extractor plate design as shown in Figure 1. Briefly, the repeller and extractor plates, that were made from 4 cm square stainless steel sheets, were spaced about 1.0 cm from one another using ceramic posts. The plate assembly was then mounted on to a base plate as shown in Figure 1. The top 0.50 cm of the repeller and extractor plates were bent about 45^o towards each other to minimize the electrical field penetration by the pulse valve into the ionization region. The exact geometry and the spacing of the plates were determined using the SIMION ion trajectory program¹⁷. As seen in Figure 1, the resulting plate design also provided a skimming action to the jet emerging from the pulse valve, thereby improving both the resolution and signal-to-noise ratio of the measurements. The minimum plate spacing at the top was about 4 mm. With this plate design, it was possible to position the pulse valve within 1.5-2.0 cm of the ionization point with minimal field penetration effects. The close proximity of the pulse valve allows the generation of photoions in the high-density region of the expanding free jet where PAH concentrations are also high. The gases eventually reach molecular flow regime further downstream, a condition where most of the prior work has been conducted^7,9. It should be noted that, although photoionization in the high-density region of an expanding jet has previously been proposed²¹, the highly complicated ion optics design renders this approach difficult to implement.

The core of the skimmed jet was then intercepted by a KrF excimer laser light ( 2.5 ns pulse, 100 Hz, MPB PSX-100, Canada). The laser light was synchronized with the pulse valve using a timed delay mechanism. The photoions generated were then rapidly removed from the gas jet using high extraction voltages (e.g. 3100 V and 2100 V at the repeller and extractor plates, respectively) to minimize ion molecule collisions. Another unique feature of the experimental system developed was the use of a ceramic shroud ring between the extractor and the base plate. This ring protects the photoions as they move into the drift tube from collisions with neutrals.

The KrF (248 nm, 5.0 eV photon energy) excimer laser light was used in the experiments since many aromatic hydrocarbons have high UV absorption cross sections in the 200-300 nm range¹⁸, and the production of their photoions by 2-photon ionization (R2PI) process is an efficient one⁹. Consequently, the detection of a large spectrum of PAH are possible using this one-color light source. The photoionization of background gases, that include H₂O, CO₂, CO, N₂ and O₂ are also largely avoided because of their higher ionization potentials and the low probability of 3-photon ionization process¹⁹.

Once the photoions were formed and extracted, they entered a 100 cm field-free drift region. The time of arrival of the ions were then recorded by dual microchannel plate (MCP) detectors either in the normal mode or the reflectron mode. In the normal mode, the ions are directly detected by a line-of-sight MCP unit. In the reflectron mode of operation, the ions enter a reflecting field that reverses their direction and improves resolution, and are detected by another MCP detector. The latter method was used in all the experiments reported here. Signals from the MCPs were then amplified by a fast amplifier (EG&G, VT120A) recorded either by a digital oscilloscope (LeCroy, 9350A) and a multichannel scaler (EG&G, Turbo MCS).

Results and Discussion

Experiments were conducted first to evaluate the sensitivity, the resolution and the time of flight response of the system developed. To attain these goals, a calibration gas mixture was prepared by introducing 1 µL of liquid standard solution containing PAH into a 2 lt stainless steel tank that has been previously heated and evacuated. For this, a commercially available PAH calibration mixture in methylene chloride solution was used (Ultrasystems Inc., RI). The tank temperature was maintained at about 250^oC to minimize the adsorption of PAH on surfaces. The transfer line from the tank to the pulse valve was kept at 200^oC and the pulse valve was a 180^oC. As will be evident from the results presented below, even higher temperatures would have been desirable to further reduce adsorption/desorption effects.

Following the introduction of the liquid sample, the tank was slowly pressurized to 2 atm using argon gas. The pressure in the tank was monitored by a high precision capacitance manometer gauge (MKS, Burlington, MA), and maintained at the same pressure for a period of 30 minutes to ensure the complete mixing of PAH and argon bath gas. The resulting gas phase concentrations of PAH were then determined from the liquid standard solution introduced, the pressure and volume of the tank and the ideal gas law equation.

In Figure 2, the TOF spectrum of a 5 parts per million (ppm) PAH mixture is shown. As evident from this figure, excellent sensitivities with high signal-to-noise ratios and resolutions were attained. The time of arrival of the various PAH ions to the detector were also determined to be correlated by the relationship t_TOF = a + bM, where M is the molecular weight and, a and b are instrument specific empirical constants. Moreover, it should be noted that the spectrum generated in Figure 2 represents a single laser shot measurement and not an average of many shots over a period of time.

Figure 2. Single Laser Shot LP-TOF/MS Spectrum of a 5 ppm Mixture of PAH.

In addition, experiments were made with naphthalene/ argon mixture in order to better assess the ultimate detection sensitivity of the instrument. This was achieved by systematically reducing the naphthalene concentration in the calibration tank and by monitoring the ion signal intensity. Naphthalene concentrations were reduced by venting the tank to a suitable lower pressure and then re-pressurizing it to 2 atm with pure argon. The resulting gas mixture was then passed through the transfer lines and the pulse valve system for a period of 10-15 minutes in order to establish the new adsorption/desorption equilibria at the new naphthalene concentration.

In Figure 3, the relative naphthalene ion intensities measured in all of the experiments are plotted as a function of concentration, in parts per billion (ppb) by volume units. As can be seen from this figure, a linear relationship was observed between naphthalene levels and signal intensity over a broad concentration range. It should be noted, however, that different instrument settings at different starting naphthalene concentrations were used to generate the composite results presented in Figure 3, yet the detector gain was kept at a constant value. In addition, the ion signals reported in Figure 3 correspond to single laser pulse measurements, without averaging or scaling. With multichannel scaling, improvements in sensitivity in the order 100 can be attained²⁰.

As indicated in Figure 3, the lowest naphthalene concentration that was reliably detected was 45 parts per trillion (ppt) by volume. Experiments at lower naphthalene concentrations were also conducted but are not reported here because of desorption rate limitations. These were apparent even in the 45 ppt experiments, as evidenced by the departure of the experimental data from the 45^o line in Figure 3. It should be possible to reduce desorption effects by conducting the experiments at higher temperatures.

Figure 3.Results of Sensitivity Study for Naphthalene Detection, Relative Ion Signal vs. Concentration of Naphthalene.

The use of longer equilibration times would not be feasible because of significant reductions in the sample tank pressure, which, in return, reduces the pulse valve delivery pressure and thus the amount of sample introduced into the TOF/MS system. In addition, excessive delays in measurements render the technique unsuitable for continuous emissions monitoring.

The measurements at 45 ppt naphthalene concentration also resulted in ion signals with S/N ratio of about 20. This suggests the feasibility of detection of nearly 4 ppt naphthalene at S/N ratio of 1 in each laser shot. With signal averaging or scaling, even sub ppt detection limits appear feasible. The 4 ppt naphthalene sensitivity is substantially below the 500-1000 ppt limit reported in the literature⁷, and establishes a new standard for continuous emissions monitoring. The dashed line shown in Figure 3 indicates the expected low ppt detection limit.

Finally experiments were conducted using the co-flow burner to measure PAH levels in the diffusion flames of methane. In these experiments a heated quartz micro-probe was used to withdraw samples from within the flame followed by LP-TOF/MS analysis in a manner as described above. The probe and sampling line temperatures were also maintained at 200^oC. Two separate flames were studied in this series of experiments. In Figure 4, the concentration profiles for benzene (78 amu), naphthalene (128 amu) and phenanthrene/anthracene (178 amu) along the central axis of the flame are presented for the following pre-combustion flow rates: CH₄ 2.17 cc/s fuel, O₂ 15.15 cc/s and Ar 59.44 cc/s in the annular region, and Ar shield gas flow of 59.9 cc/s. Under these conditions, the flame exhibited a yellow tip indicative of soot and carbon formation. Although higher molecular weight PAH were formed, they were not observed in these experiments because of their adsorption on surfaces in the probe and/or transfer lines at the low temperatures employed (200^oC). The profiles presented in Figure 4 also represent the average of 3 set of experiments, with each data point determined from the average of 50 laser shots at 5 Hz.

Figure 4.Concentration Profiles for Species With Molecular Weights 78 (Benzene), 128 (Naphthalene), 178 (Phenanthrene) as a Function of Distance Above Burner Surface for a Co-Flowing Methane Diffusion Flame.

As evident from Figure 4, the PAH profiles exhibited trends as expected along this flame. That is, the PAH concentrations first increased as the flame front is approached, and then decreased in the post flame zone due to oxidation. In addition, the peak locations shifted towards longer reaction times with increasing molecular weight. From Figure 4, it can be seen that the PAH levels along this flame was substantial reaching levels as high as 70 ppm. As a result of these high PAH levels, it was not possible to probe this flame for low PAH concentrations. Consequently, another flame with slightly lower CH₄ and higher O₂ flow rates were also studied. This flame, which did not exhibit yellow tipping, had the following pre-flame flow rates: CH₄ 1.91 cc/s, O₂ 18.18 cc/s and Ar 100 cc/s in the annular region, and 119 cc/s of Ar shield gas. In Figure 5, the concentration profile for naphthalene along this flame is presented. These results also correspond to the average of 3 sets of experiments and 50 laser shots at 5 Hz for each measurement. As evident from this figure, naphthalene concentrations varied over several orders of magnitude along the flame, and the LP-TOF/MS system was able to provide real time detection over this entire range. It is particularly significant to note the feasibility of detection of naphthalene at levels as low as 4 parts per billion (ppb) in combustion products in real time. At these levels the experimental S/N ratios were about 1-5. Although this detection level is gratifying, it is significantly higher than the 45 ppt naphthalene sensitivity determined by the calibration mixture experiments.

Figure 5. Concentration Profile for Species With Molecular Weight 128 (Naphthalene) as a Function of Distance Above Burner Surface for a Non-Sooting, Non-Luminous Co-Flowing Methane Diffusion Flame.

Further research is underway to improve the sensitivity of detection of naphthalene and other PAH in combustion products via Resonance Enhanced Multiphoton Ionization (REMPI) process. In this approach, a laser beam with a specific photon energy (obtained from a tunable laser) is used for the REMPI of a desired target molecule. Because REMPI efficiencies are generally very large, it will be possible to significantly increase the number of ions generated, thus detection sensitivity under this mode of operation.

It is important to note that the LP-TOF/MS developed holds the promise to be a universal real time detector. The universality comes from the fact that any compound can be detected by mass spectrometry, and that the selective photoionization of species can be accomplished by the use of suitable lasers. As a result of its universality, sensitivity and real time capability, the LP-TOF/MS represents an attractive Continuous Emissions Monitor (CEM) for environmental compliance purposes. In addition, the LP-TOF/MS also opens the possibility to implement real time process control to minimize process emissions and to maintain product quality.

Acknowledgements

This work was supported, in part, by funds from the Toxic Combustion By-Products Consortium of Companies, the National Science Foundation, the Environmental Protection Agency, the University of California Energy Institute and the UCLA Center for Clean Technology.

References

1. Klaasssen, C.D., Amdur, A.O., Doull, J., Eds. Casarett and Doull's TOXICOLOGY, Macmillan, NY (1986).

2. Longwell, J.P., 19th Symposium (Int'l) on Combustion, The Combustion Institute, p.1339 (1982).

3. Bjorseth, A., editor, "Handbook of Polycyclic Aromatic Hydrocarbons", Marcel Dekker Inc. NY (1983).

4. Otsen, R.; Leach, J.M.; Chung, L.T.K., Anal. Chem., 59, 1701-1705, (1987).

5. CARB, California Air Resources Board, "Determination of Polycyclic Aromatic Hydrocarbon Emissions from Stationary Sources", Method 429, Sept. 12, (1989).

6. Griest, W.H., and Caton, J.E., in "Handbook of Polycyclic Aromatic Hydrocarbons", Bjorseth, A., Ed., Marcel Dekker, New York (1983).

7. Tanada, T.N., J., Velasquez, N. Hemmi, and T.A. Cool, Combust. Sci. Tech., 101, 333 (1994).

8. Siegmann, K., H. Hepp, K. Sattler, H. Burscher, and H.C. Siegmann, "Fast Growth of PAH in Flames: A New Approach", poster paper presented at the 25th Symposium (Int'l) on Combustion, Irvine CA (1994).

9. Lubman, D.M., and L. Li, in "Lasers and Mass Spectrometry", D.M. Lubman, Ed.,p.353, Oxford University Press, New York, (1990).

10. Weeks, S., A.P. D'Silva, R.L.M. Dobson, in "Lasers and Mass Spectrometry", D.M. Lubman, Ed.,p.510, Oxford University Press, New York, (1990).

11. Thijssen, J.H., M.A. Toqan, J.M. Beer, and A.F. Sarofim, Combust. Sci. Tech., 90, 101 (1992).

12. Coe, D.S., and J.I. Steinfeld, Chem. Phys. Lett. 76, 485 (1980).

13. DiLorenzo, A., A. D'Alessio, V. Cincotti, S. Masi, P. Menna, C. Venitozzi, 18th Symposium (Int'l) on Combustion, 485 (1981).

14. Prado, G., Garo, A., Ko, A., and Sarofim, A., Twentieth Symposium (International) on Combustion, The Combustion Institute, 989, (1984).

15. Castaldi, M.J., Vincitore, A.M., and Senkan, S.M., Combust. Sci. Tech., 107, 1, (1995).

16. Wiley, W.C., and I.H. McLaren, Rev. Sci. Inst., 26, 1150 (1955).

17. Dahl and Delmore, Simion Program Ver. 6.0, INEL, U.S.D.O.E., (1994).

18. Friedel, R.A., and M. Orchin, The Ultraviolet Spectra of Aromatic Compounds, John-Wiley, New York (1951).

19. Lias, S.G., J.E. Bartmess, J.F. Liebman, J.L. Holmes, R.D., Levin, W.G. Mallard, J. Phys. Chem. Ref. Data, 17, Supplement No. 1, (1988).

20. Cotter, R.J., in "Time of flight Spectrometry", ACS Symposium Series No:549, p.16 (1994).

21. Oser,H., Thanner,R., Grotheer,H.H., Combust. Sci. Tech., 116, 567 (1996).