Anal. Chem. 1998, 70, 2745-2749
Mass Spectrometry of Particles Formed in a Deuterated Ethene Diffusion Flame R. A. Fletcher,*,† R. A. Dobbins,‡ and H.-C. Chang§
Surface and Microanalysis Science Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, Division of Engineering, Brown University, Providence, Rhode Island 02912, and Saint Gobain/Norton, Worcester, Massachusetts 01615
Nanometer-sized spherule soot precursor particles have been collected by thermophoretic sampling from the interior of a laminar diffusion flame and mass analyzed by laser microprobe mass spectrometry. Mass spectra of the precursor particles formed in an ethene diffusion flame have indicated the presence of polycyclic aromatic hydrocarbons (PAHs) in the m/z range of 202-300 and higher mass peaks extending out to m/z 472. The mass resolution of the time-of-flight mass spectrometer used did not provide conclusive identification of PAHs because of ambiguities in assignment for the relative amounts of carbon and hydrogen (CxHy) for each PAH peak and the possibilities of spectral interferences. To determine the chemical formula that can be assigned to each molecular ion peak, an isotopically pure deuterated ethene (C2D4) fuel was burned in place of normal ethene (C2H4) in the diffusion flame. For the normal ethene fuel, mass peaks tentatively identified as C16H10 to C38H16 were obtained. Accordingly, deuterated PAH peaks ranging from C16D10 to C38D16 were found when C2D4 was burned. These m/z values correspond to molecular ion, M•+, peaks for an array of PAH compounds. The deuterated PAH mass peaks (CxDy) were entirely consistent with a mass shift of y mass units with respect to the normal PAH mass peaks. The carbonaceous particle aggregates collected from the upper flame region have mass peaks characteristic of Cx+ and CxH+, while the deuterated soot has Cx+ and CxD+. The deuterated ethene experiment has verified the identities of x and y in the PAH (CxHy) compounds present in the precursor particle samples. No prior experiment using pure deuterium-based fuel as a combustion diagnostic to form aerosol-containing deuterated PAH compounds has been reported. Fundamental studies of soot particle formation mechanisms are motivated by wide interdisciplinary goals spanning fuel combustion efficiency, health and environmental concerns, and forensic applications. Earlier work reported the existence of small (5-10 nm diameter) precursor particles formed in the lower regions of a diffusion flame and the significant role that these precursor particles play in early particle formation processes in †
NIST. Brown University. § Saint Gobain/Norton. ‡
S0003-2700(97)01293-6 CCC: $15.00 Published on Web 06/02/1998
© 1998 American Chemical Society
flames.1,2 Precursor particles are small, solitary spherules with liquid-like characteristics and a high electron beam transmission when compared to that of the fully developed aggregated structures found higher in the same flame. Collections of precursor and carbonaceous soot aggregated particles were subjected to analysis by transmission electron microscopy (TEM) and laser microprobe mass spectrometry (LMMS).3 Figure 1 illustrates the striking particle morphology differences. The results of the laser microprobe analysis on the precursor particles indicated the presence of polycyclic aromatic hydrocarbon (PAH) mass peaks. Pure PAH compounds and a standard reference material (SRM 1491) composed of 24 PAHs were analyzed in the past, and the molecular ion peaks were found to coincide with some of the peaks found in the precursor particle samples.3 However, due to the mass resolution of the time-of-flight mass spectrometer used, the PAH-like peaks could not be positively identified as aromatic hydrocarbons especially, in light of the possible carbon cluster (Cx+) peak interferences. To address this concern, pure deuterated ethene (C2D4) was substituted as fuel in place of normal ethene (C2H4), and the precursor particles formed in this flame, as well as carbonaceous aggregates present higher in the flame, were collected for analysis. Peaks in the timeof-flight mass spectra of the LMMS that correspond to PAHs from precursor particles undoubtedly result from a combination of various isomers (CxHy). Examples of isomeric PAHs for C20H12 previously identified as intermediates in the combustion of hydrocarbon fuels are benzo[e]pyrene, perylene, and benzo[b]fluoranthene. Although exact compound identification is not possible by this analysis, positive identification of the PAH formula (CxHy) for species found in soot precursor particles is a goal of the present study. This paper also demonstrates the possibility of utilizing labeled fuels to better understand particle formation mechanisms. A computer search of the abstracted literature has revealed no prior use of a pure deuterium-carbon compound as the fuel in combustion diagnostic experiments. (1) Dobbins, R. A.; Subramaniasivan, H. In Soot Formation in Combustion; Bockhorn, H., Ed.; Springer-Verlag: Berlin, 1994, pp 81-103. (2) Dobbins, R. A. The Early Soot Particle Formation in Hydrocarbon Flames. Invited Lecture at the Glassman 70 Symposium, Princeton University, October 28, 1993. In Physical and Chemical Aspects of Combustion: A Tribute to Irvin Glassman; Dryer, F. L., Sawyer, R. F., Eds.; Gordon and Breach: Newark, NJ, 1997, pp 107-133. (3) Dobbins, R. A.; Fletcher, R. A.; Lu, W. Combustion Flame 1995, 100, 301309.
Analytical Chemistry, Vol. 70, No. 13, July 1, 1998 2745
Figure 1. Micrographs of particles captured in an ethene C2H4 diffusion flame. (a) Precursor particles captured at Z ) 20 mm above the burner mouth on the axis of the flame. (b) Carbonaceous aggregates found at Z ) 50 mm.
EXPERIMENTAL SECTION The burner configuration consists of a central tube of 11.1 mm i.d. with a surrounding air stream of 102 mm diameter. In the normal configuration that has been used to study the nonsmoking ethene flame, the fuel (C2H4) flow rate is 3.9 cm3/s, while the air flow rate is 715 cm3/s. These conditions result in a stable laminar flame whose visible flame height is 88 mm (measured from the base of the burner mouth) when the burner has achieved thermal equilibrium after about 30 min of operation. Five liters of C2D4 (98%) was provided for these experiments by the Cambridge Isotope Laboratories under its Research Grant Program.4 The high value of this gas necessitated special procedures for its conservation. The flame height of the C 2D4 flame was adjusted by a needle valve setting to be equal to the 88 mm value that results in the case of the C2H4 flame under the flow conditions and the burner configuration stated above. This measure circumvented the need to consume any of the limited C2D4 supply in a flow calibration procedure. A three-way valve that permitted a rapid selection of either C2H4 or C2D4 was installed adjacent to the burner entrance port. With this feature, it was possible to use C2H4 to bring the burner to thermal equilibrium and then rapidly shift to C2D4 to conduct sampling tests. The height of the flame was measured after the full test sampling sequence was completed and was found to be 88.9 mm. The use of the C2D4 fuel caused a reduction of the visual flame luminosity that is interpreted as signifying a reduction in the (4) The 2nd Research Grant Program, Cambridge Isotope Laboratories, 50 Frontage Rd., Andover, MA 01810; 1995.
2746 Analytical Chemistry, Vol. 70, No. 13, July 1, 1998
formation rate and concentration of soot particles. This reduction may be caused by the smaller diffusion coefficient of deuterium that replaces hydrogen as a key component in the early soot formation reactions. The flame-borne particles were captured at well-defined flame locations by the thermophoretic sampling procedure.5 This method employs a pneumatically controlled system to insert an electron microscope grid to a precisely controlled position in a flame for time interval of about 50 ms. Particles in the flame are driven to the cool grid from the hot gas stream by the thermal gradient-driven thermophoretic force. Two locations of interest in the present tests are on the axis of the flame at heights Z above the burner of 20 and 50 mm. The spatial positioning of the sampling grid is estimated to be within 0.5 mm in both the vertical and the radial positions. Precleaned plasma-ashed TEM grids with no supporting thin film were used as particle collection surfaces. Multiple exposures of 50 ms are required in order to capture a sample of adequate mass for LMMS analysis. Eight exposures were used at Z ) 20 mm, and six exposures were adequate at Z ) 50 mm. The exposed TEM grids were conveyed to the NIST in polyethylene vials for analysis by laser microprobe mass spectrometry on the LAMMA-500 (Leybold-Heraeus, Cologne, Germany), which has been described extensively.6,7 Total sample mass of collected particulate material was estimated to be less than 100 pg and required a microprobe analytical technique with high sensitivity to carbon compounds. LMMS permits comparison of both the precursor particles and the fully mature carbonaceous soot aggregates formed and captured in the diffusion flame. The use of time-of-flight mass spectrometry to analyze the composition of the precursor particles formed in the flames burning, alternately, C2H4 and C2D4 allows the juxtaposition of the mass spectrasa presentation that displays the mass shift across the entire spectrum. The LAMMA-500 has transmission geometry with respect to the laser beam and the time-of-flight mass spectrometer so that the beam is incident on the front surface of the TEM grid containing particles. The laser beam was defocused to produce spot sizes on the grid surface of approximately 50-100 µm diameter, irradiating an estimated 1-20 pg sample material. Irradiance levels were approximately 1-20 MW/cm2 of 266 nm radiation. Positive ions were mass-analyzed by a time-of-flight mass spectrometer with mass resolving power, M/∆M, of approximately 500 at m/z of 208. The mass spectra were analyzed using software developed by the University of Antwerp. RESULTS AND DISCUSSION Most important was the comparison made between the mass peaks of normal PAH and deutero-PAH species produced from combustion of normal ethene and deuterated ethene. The positive ion mass spectra shown in Figure 2 illustrate the mass shift in the molecular ions due to replacement of hydrogen by deuterium. Figure 2a is a positive ion mass spectrum derived from averaging (5) Dobbins, R. A.; Megaridis, C. M. Langmuir 1987, 3, 254-259. (6) Denoyer, E.; Van Grieken, R.; Adams, F.; Natusch, D. F. S. Anal. Chem. 1982, 54, 26A-41A. (7) Van Vaeck, L.; Van Roy, W.; Gijbels, R.; Adams, F. Lasers in Mass Spectrometry: Organic and Inorganic Instrumentation. In Laser Ionization Mass Analysis; Vertes, A., Gijbels, R., Adams, F., Eds.; John Wiley & Sons: New York, 1993; p 41-51.
Table 1. Number of Carbon and Hydrogen Atoms Per Formula, m/z Values Found, and Relative Peak Intensities for Normal and Deuterated PAH Pairs Normalized to Total Peak Area, Expressed in Percent
Figure 2. Mass spectra of particles sampled from Z ) 20 mm height in a diffusion flame burning (a) normal ethene (C2H4) and (b) deuterated ethene (C2D4). The asterisk denotes significant minor peaks found in the literature to be common in this type of mass spectra related to PAHs.
10 individual spectra of precursor particles collected at Z ) 20 mm and formed from burning normal C2H4 fuel. Figure 2b is the analogue of the Figure 2a spectrum for precursor particles formed by burning C2D4. It is apparent that the normal PAH peaks formed in the C2H4 burning flame at m/z 202, 252, 276, 300, 326, 350, and 374 shown in Figure 2a are shifted to deutero-PAH mass peaks at m/z 212, 264, 288, 312, 340, 364, and 388 in Figure 2b for the deuterated ethene (C2D4) flame. The relative mass shifts of 10, 12, and 14 are the appropriate amounts corresponding to the hydrogen/deuterium content of the molecular species. This is readily seen in Table 1, where the relative peak intensities that have been calculated from the sum of the individual peaks for the respective number of spectra analyzed are shown. If the peaks at m/z 252, 264, 276, 300, etc. were a result of only carbon clusters (void of H or D), no shift in m/z would have been observed. The results confirm the presence of PAH compounds on the TEM grids containing the soot precursor particles. Since only pure deuterated ethene was used in the experiments, we expect the PAH peaks to contain ion signals from deutero-PAH. The PAH and deutero-PAH peaks are directly collated, and, given the position of the normal PAH peaks, the deutero-PAH peak position can be unambiguously predicted. Deutero-PAHs, CxDy analogues of the PAHs, CxHy, should have equivalent values for x and y, and for equal values of y, the m/z shift is y. For example, in the deuterated ethene flame, the peak corresponding to C20D12 (m/z 264) appears in lieu of the peak resulting from C20H12 (m/z 252) that is prominent in the normal ethene burns. The m/z difference is y ) 12. A selected set of the
x
y
CxHy (m/z)
relative peak intensitya,c
CxDy (m/z)
relative peak intensityb,c
16 18 18 19 20 20 21 21 22 22 24 24 26 26 28 30 32 38
10 10 12 11 10 12 11 12 10 12 12 14 12 14 14 14 14 16
202 226 228 239 250 252 263 264 274 276 300 302 324 326 350 374 398 472
8.31 2.62 5.39 5.69 4.27 26.2 4.78 4.94 1.34 19.2 5.86 3.90 0.74 2.76 1.95 1.36 0.50 0.22
212 236 240 250 260 264 274 276 284 288 312 316 336 340 364 388 412 488
13.8 0.57 2.76 4.49 0.75 12.8 5.17 3.55 1.43 23.1 10.8 0.50 1.28 3.48 6.19 6.68 2.14 0.43
a Based on 44 spectra. b Based on 18 spectra. c The expanded uncertainty in the relative peak intensity is approximately 20%. Expanded uncertainty in the relative peak intensity consists of twice the square root of the sum of the squares of the Poisson counting uncertainty associated with the area of the mass peaks and the estimated uncertainty due to peak integration.
mass peaks given in Table 1 are presented in Figure 3 in histogram representation, where the relative peak intensities are plotted vs m/z. The logarithmic scale is chosen for the Figure 3a ordinate to display the smaller mass fractions more prominently. Both the normal and the deutero-PAHs are shown in Figure 3a, representing a combined mass spectrum of selected species found in the precursor particles at Z ) 20 mm of both flames. The 20 mass peaks in Figure 3a are collapsed into 10 peaks that are shown in the linear stacked bar plot, Figure 3b, when the masses of the deutero-PAH are reduced by the expected amount of hydrogen (y amounts) for the respective isomeric group. The perfect coincidence of the PAHs and deutero-PAHs shown in Figure 3b, occurring over a m/z range of 202-488 and for numbers of hydrogen atoms ranging over values of y ) 1016, strongly supports the choice of the x and y assignments for both PAH and deutero-PAH related peaks. In the region of higher mass, the deutero-PAHs display higher count fractions than the normal PAHs. Mass peaks that correspond to PAHs with relative molecular masses below 200 amu are rarely seen in the precursor particles, presumably because of the higher vapor pressure of these compounds that would cause them to vaporize in the instrument’s vacuum chamber before they could be analyzed. Carbonaceous soot particles collected at Z ) 50 mm height in the flame are mature, agglomerated aggregates that differ from the precursors by morphology and hydrogen content. In Figure 4, we compare single positive ion mass spectral results for carbonaceous aggregates created from C2H4 and C2D4. For the Z ) 50 mm particles, the large m/z shift, as previously observed for the precursor particles, is absent for particles collected at this height from either fuel. All the carbon-associated peaks appear to result from positive ion carbon clusters, Cx+, and CxH+ or CxD+ in the case of the deuterated ethene flame; no PAH compounds Analytical Chemistry, Vol. 70, No. 13, July 1, 1998
2747
Figure 4. Mass spectra of particles sampled from Z ) 50 mm height in a diffusion flame burning (a) normal ethene (C2H4) and (b) deuterated ethene (C2D4).
Figure 3. Histogram of relative peak intensities of selected peaks ranging from m/z 202 to 488 detected in the precursor particles at Z ) 20 mm height in both the C2H4 and C2D4 fueled flames. (a) The combined set of peaks associated with PAH and deutero-PAH with values of x and y common to both sets. (b) The same combined set of peaks with deutero-PAH m/z values reduced by y, the number of hydrogen atoms in the PAH, illustrating the match of the chemical formula for normal PAH and deutero-PAH reduced by y amu for all the peaks shown.
are detected. Although the C2D4 flame did not yield appreciable amounts of carbonaceous aggregate at these heights, the figure shows the expected negligible effect of the deuterium substitution for hydrogen on the carbon cluster series, Cx+ (the carbon cluster series that has m/z values corresponding to 12x). Peaks are observed for the particles formed from the C2D4 fuel which are the analogues of (CxH)+ [m/z 12x + 1] for particles collected at this height in the normal ethene. Examples of such CxD+ [12x + 2] peaks seen in Figure 4b are m/z 134, 146, 158, 170, and 182. The stability of PAHs CxHy, with x and y even, under conditions approximating these flame temperatures, pressures, and major gas composition, was determined from thermodynamic calculation by Stein and Fahr.8 The results of this study were presented as a grid of x vs y that displays the most stable isomer for each PAH class. These classes are termed stabilomer masses, and the graph represents a valley along which growth to larger PAH is favored to occur. It has been pointed out that the PAH classes identified in soot precursor particles sampled from the C2H4 flame correspond to the stabilomer masses.9 In Figure 5, the masses (a) CxHy and (b) CxDy that correspond to those of the stabilomers in (8) Stein, S. E.; Fahr, A. J. Phys. Chem. 1985, 89, 3714.
2748 Analytical Chemistry, Vol. 70, No. 13, July 1, 1998
the x-y grid of Stein and Fahr are plotted for the samples from Z ) 20 mm in both natural ethene (C2H4) and deuterated ethene (C2D4) flames. The shaded area represents the possible x and y values thermodynamically allowed per Stein and Fahr. The number, fn, given in each cell represents the peak area for each mass divided by the total peak area for all stabilomers masses. Minor constituent masses with odd values of x and y, such as C19H11, are formed by methylene bridges, as discussed elsewhere.9 Figure 5 shows the major species found in the precursor particles of both C2H4 and C2D4 diffusion flames mapped onto the thermodynamically allowed space of the stabilomer grid. This result indicates that growth to larger masses occurs along the stabilomer grid for both the normal PAH and deutero-PAH species. Both electron impact and laser ionization mass spectra exhibit (M - 2H)+ peaks for certain PAH compounds.10-12 Mass spectra for precursor particles derived from normal ethene indicate minor peaks at m/z 250 and 274 (as noted by the asterisk in Figure 2), with major peaks at 252 and 276. This (M - 2H)+ occurrence lends further support that PAH compounds are present in the precursor particles. In conclusion, we present the first experiments that use deuterated ethene (C2D4) as a substitute fuel for combustion diagnostic studies. The results of the experiments provide positive identification of the chemical formulas, CxHy, of the normal PAH (9) Dobbins R. A.; Fletcher, R. A.; Chang, H.-C. Combustion Flame, in press. (10) EPA/NIH Mass Spectral Data Base, Vol. 3, U.S. Department of Commerce, 1978. (11) Van Vaeck, L.; Claereboudt, J.; De Waele, J.; Esmans, E.; Gijbels, R. Anal. Chem. 1985, 57, 2944. (12) Balasanmugam, K.; Viswanadham, S. K.; Hercules, D. M. Anal. Chem. 1986, 58, 1103.
Figure 5. Count fraction, fn, of (a) normal PAH and (b) deutero-PAH masses represented in a stabilomer grid. The shaded area represents the boundaries of the most stable PAH classes as reported by Stein and Fahr.8
isomers found to be present in the precursor particles of a C2H4 diffusion flame and rule out the possible interference of Cx+ clusters. ACKNOWLEDGMENT The authors thank the Cambridge Isotope Laboratories for the contribution of the C2D4 that was used in this investigation. This work was sponsored by the U.S. Army Research Office under Grant No. DAAH04-95-1-0429.
NOTE Commercial equipment, instruments, materials, and software are identified in this report to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the NIST, nor does it imply that they are the best available for the purpose. Received for review November 26, 1997. Accepted April 9, 1998. AC971293Q
Analytical Chemistry, Vol. 70, No. 13, July 1, 1998
2749
