- Email: [email protected]
In situ measurements of soot formation in simple flames using small angle X-ray scattering
Nuclear Instruments and Methods in Physics Research B 238 (2005) 334–339 www.elsevier.com/locate/nimb
In situ measurements of soot formation in simple flames using small angle X-ray scattering C. Gardner a, G.N. Greaves a,*, G.K. Hargrave b, S. Jarvis b, P. Wildman b, F. Meneau a,c, W. Bras c, G. Thomas a b
a Institute of Mathematical and Physical Sciences, University of Wales, Aberystwyth SY23 3BZ, United Kingdom Wolfson School of Mechanical and Manufacturing Engineering, Loughborough University, Loughborough LE11 3TU, United Kingdom c Netherlands Organisation for Scientific Research (NWO), DUBBLE CRG/ESRF, P.O. Box 220, F38043 Grenoble Cedex, France
Available online 1 September 2005
Abstract Direct SAXS measurements of soot formation from ethylene have been made using laminar pre-mixed flames for the first time. The slot burner was configured to maximise the signal from particulates. The geometry also enabled the thermal background from the surrounding hot gasses to be accurately removed. With cold flame speeds of 40 cm s1 we have been able to identify particle sizes and densities from moderately sooty to rich flame conditions. By adjusting the height of the burner in the beam, the development of particles as a function of position above the flame tip and therefore as a function of time from ignition have been obtained. These reveal evidence for bimodal particle nucleation and growth at different stages in the continuous combustion of ethylene. Ó 2005 Elsevier B.V. All rights reserved. PACS: 60.10.Eq; 61.46.+w; 81.05.Uw Keywords: Soot; Flame; Particle size; Nucleation; Aggregation; SAXS
1. Introduction Understanding soot generation in simple hydrocarbon fuels is fundamental to establishing practical strategies for increasing combustion efficiency, *
Corresponding author. Tel.: +44 1970 622802; fax: +44 1970 622826. E-mail address: [email protected] (G.N. Greaves).
at the same time as limiting particulate emissions. Quantitative knowledge of the processes from soot inception to the generation of visible particles has, until recently, relied on in situ static and dynamic optical techniques [1,2] and ex situ transmission electron microscopy (TEM) [3–5]. Optical techniques, however, are insensitive to particle sizes ˚ and therefore to the primary soot below 100 A particles that nucleate as well as the smaller
0168-583X/$ - see front matter Ó 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2005.06.072
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339
polycyclic aromatic hydrocarbons (PAHs) [6] in the flame. Ex situ methods like TEM and scanning mobility particle sizing (SMPS) [7,8], are able to ˚ , but the sampling measure sizes below 100 A probes generally result in significant particle coagulation. Despite extremely low volume fractions (108) and aerosol densities (1011 cm3) [7], in situ small angle X-ray scattering (SAXS) offers attractive possibilities for probing particulates in flames [5,9–11], but the main empirical challenge is to separate the scattering from soot from the scatter due to the high temperature gaseous background. In small angle neutron scattering (SANS) the background is dominated by incoherent scatter from hydrogen and water vapour [12] which masks hydrocarbon structure in the ‘‘dark region’’ between the soot particles. This, however, is accessible through SAXS because scattering from hydrogen is much smaller. Previously sooting flames have been ignited within sootless flames, enabling rapid in situ measurements to be made [10,11], but by incurring large temperature gradients that make the sootless background difficult to remove. Ultra
335
SAXS is necessarily ex situ, avoids the hot background problem, but requires more concentrated samples than the volume fraction available in living flames. As a result the PAHs and primary particles are heavily masked by surface scattering from the recovered large soot aggregates which ˚ in diameter [5]. Soot particles are typically 250 A ˚ in diameter are also analysed from ex situ of 460 A SAXS measurements [13] using form factor analysis [11].
2. Experiment Direct SAXS measurements of soot formation that overcome many of these difficulties have recently been made at the ESRF using BM26B. The experimental arrangement is shown in Fig. 1. Ten kilo-electron-volt X-rays and with two sample–detector lengths provided a wave vec˚ 1, encomtor range of 0.016 > 4p sin h/k > 0.3 A passing the scattering for the range of particle sizes expected within living flames. A stabilised water-cooled slot burner was employed,
Fig. 1. Schematic arrangement of laminar flow slot burner and experimental configuration used for SAXS.
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339
sufficiently long to maximise the signal from the soot and dark region and to minimise air scatter from outside the flame and the effects of temperature gradients at the edges. Precautions were also taken to preserve the integrity of the flame, such as the chimney and gauze guards. The burner, the chimney and the heat exchanger were mounted on a vertical traverse so that the region above the flame tip could be probed with the 0.3 mm beam up to 50 mm above the burner surface. A cold flame speed of 40 cm s1 was used, enabling measurements to be made for combustion times of up to 5/40 = 125 ± 1 ms. Measurements were made in both vertical directions in order to determine the size of systematic errors from the repeatability of the analysed SAXS parameters. Three ethylene/air flames were examined: a rich flame with an equivalence ratio, U = 2, a moderately sooty flame (U = 1.85) and a lean sootless flame (U = 1.3). The leaner flame provided background profiles at the same points above the burner where the data for the sooting flames were recorded. SAXS from both flames were obtained in pairs over 40 min exposures at each height above the burner. From separate thermocouple measurements for this burner sooting flames are known to be 50 °C hotter than leaner flames at the same height, resulting in small but measurable differences in scatter from the residual thermal density fluctuations. These were corrected for together with detector variations. SAXS profiles from the particles in the rich and moderately sooty flames were obtained by subtracting the leaner flame data from the sooting flame data at the same height. The resulting wave vector ranges were generally a decade smaller than those previously reported by other groups for SAXS [10,11,13] or for SANS [12] experiments. This enabled reliable SAXS measurements to be made of soot particles in their native environment without resorting to form factor analysis and covering the important ˚ down to 5 A ˚. size range from 100 A
3. Results Preliminary results are shown in Figs. 2–4. Fig. 2 shows a typical example of the SAXS data.
10
1
A
I(q) Relative Intensity
336
B
10
0
0.000
0.005 2
0.010
—2
q (Α )
Fig. 2. Example of background subtracted log I versus q2 from particles within the moderately sooty flame 22 mm above the burner obtained with 2.5 m flame-to-detector distance. Two regimes, marked A and B, relate the two different particle size groups identified at all heights above the burner.
Plotted as log I(q) versus q2, this exhibits two regimes, A and B. Both regimes were evident for all the flames and the different sample-to-detector distances employed enabled accurate data to be obtained for each region. For separated particle systems the SAXS 2 2 intensity at low q values, Rg q IðqÞ ¼ Ið0Þ exp [14], where Rg is the Gui3 nier radius and measures the particle size. Our results therefore point to the existence of two groups of particles, each with different characteristic sizes. Fig. 3 shows the Rg values obtained for the rich flame as a function of height above the burner and therefore of ignition time. For one group, average Rg values rise slowly from around ˚ to around 65 A ˚ , whereas for the other they 50 A ˚ to 40 A ˚ . The rise more sharply from around 25 A dotted line in Fig. 3 corresponds to Rg values of ˚ and was obtained from the leaner around 8 A flame used for background subtraction. These smaller particles vary little in size with height above the burner and relate to nanostructure in the residual gases within the flame.
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339
337
TIME (ms) 50
75
100
125
80
GUINIER RADIUS (ANGSTROMS)
70
A 60
50
40
B 30
20
10
0 15
20
25
30
35
40
45
50
HEIGHT (mm)
Fig. 3. Guinier radii for rich flame (U = 2) measured for sample-to-detector distances of 2.5 m (j) and 1.7 m (m). A and B relate to the two particle regimes identified in Fig. 2. The arrow marks the onset of SAXS signal taken from Fig. 4. The dotted line indicates the Guinier radii of the background sootless flame (U = 1.3). Curves are included to guide the eye. The combustion time, t, is related to the height above the burner, H, by t = H/m, where m is the flame velocity of 40 cm s1.
WEIGHTED INTEGRATED INTENSITY (arbitrary units)
TIME (ms) 50
75
100
125
4
3
2
1
0 15
20
25
30
35
40
45
50
HEIGHT (mm)
Fig. 4. Weighted integrated SAXS intensities for rich flame (U = 2) (j) compared to a moderately sooty flame (U = 1.85) (h), measured with a sample-to-detector distance of 2.5 m. Curves are included to guide the eye. The combustion time and height above the burner are related to the flame velocity as in Fig. 3.
338
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339
Fig. 4 shows the weighted integrated SAXS R qmax intensity, Q ¼ qmin IðqÞq2 dq, for the two flames with the same 2 m camera length. A step function can be clearly identified for both flames, the step sizes reflecting the different levels of soot being generated and the onsets the different delays for combustion. The error bars in Figs. 3 and 4 indicate the reproducibility of the measurements at the same height for repeated traverses of the burner at times separated by several hours. By comparison the individual statistical errors are far smaller, being contained within the sizes of the symbols. 4. Discussion The abrupt rise in Q for both flames in Fig. 4 illustrates the rapid nucleation of soot in laminar pre-mixed flames. This occurs for ethylene within 10 mm of the flame tip or 25 ms from ignition and compares well with the rise time in number density of 24 ms obtained from SMPS experiments for equivalent flames (U = 2.07) [7]. A similar figure emerges from SANS measurements of rich ethylene flames [12]. Turning to the particle sizes presented in Fig. 3, these clearly fall into three groups. For leaner flames the molecular particles ˚ ) whilst for rich flames the largare small (Rg 8 A ˚ both values est soot particles have Rgs of 60 A are relatively constant with height above the burner surface. The sizes of the third group are intermediate between these limits, growing progressively from sizes close to the molecular structure at the flame tip and eventually attaining ˚ further up the flame. values with Rgs of 40 A In order to relate these particle sizes with those measured by different techniques, we recall that TEM images of primary soot particles sampled from combusting fuels like ethylene reveal ‘‘onion-like’’ shapesp [4,5]. For spherical particles the radius equals (5/3)Rg [14]. Scaled accordingly, the largest particles we measure have average ˚ , in good agreement with diameters of 140–160 A the mean diameters for the largest particles reported from SANS data on rich ethylene flames from form factor analysis [12]. Our in situ SAXS ˚ diameter results are also similar to the 120–140 A values for particulates sampled from diesel engines
measured by USAXS [5], in this case deconvoluted from surface scattering from much larger soot aggregates. These authors also report evidence for much smaller ‘‘sub unit’’ particles with diameters ˚ , close to the 20 A ˚ particle sizes we measure of 17 A directly from in situ SAXS from lean ethylene flames (dotted line in Fig. 3). Turning finally to the particles of intermediate size which grow during soot formation, these appear to nucleate from the ˚ ) and aggregate to create the larger priPAHs (20 A ˚ ) – a bimodal process that mary particles (160 A starts almost immediately at the flame tip. Bimodal particle size distribution functions have recently been reported for rich ethylene flames (U 2), using ex situ sampling methods in conjunction with SMPS [7,8]. These develop with height above the burner surface. Both groups report bimodality, with an exponential component stemming from smaller particles extending to ˚ in diameter, together with a very around 100 A broad lognormal component centred around parti˚ but with r values as large as cle diameters of 200 A 1.5 [7]. Our in situ experiments reveal smaller diameters for both groups of particles (Fig. 3). Moreover the linearity of the two Guinier regions A and B (Fig. 2) suggest far less polydispersity in soot particles probed in situ than sampled ex situ. Finally our SAXS data obtained from laminar living flames offer precise height/time information on the way the different particle groups develop along the flame.
Acknowledgements Experimental work was supported by the UK Engineering and Physical Sciences Research Council (EPSRC), the Netherlands Organisation for Scientific Research (NWO) and the European Synchrotron Radiation Source (ESRF).
References [1] A. DÕAlessio, in: D.C. Siegla, G.W. Smith (Eds.), Particulate carbon – Formation during combustion, Plenum, New York, 1982, p. 207. [2] G. Kroner, H. Fuchs, R. Tatschl, O. Glatter, Part. Part. Syst. Char. 20 (2003) 111.
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339 [3] H. Bockhorn, F. Fetting, A. Heddrich, U. Meyer, G. Wannemarcher, J. Aerosol Sci. 19 (1988) 591, and references therein. [4] S. di Stasio, Carbon 39 (2001) 109. [5] A. Braun, F.E. Huggins, S. Seifert, J. Ilavsky, N. Shah, K.E. Kelly, A. Sarofim, G.P. Huffman, Combust. Flame 137 (2004) 63. [6] H. Wang, M. Frenklach, Combust. Flame 96 (1994) 163. [7] B. Zhao, Z. Yang, M.V. Johnston, H. Wang, A.S. Wexler, M. Balthasar, M. Kraft, Combust. Flame 133 (2003) 173. [8] M.M. Maricq, Combust. Flame 137 (2004) 340.
339
[9] W.A. England, Combust. Sci. Technol. 46 (1986) 83. [10] J.P. Hessler, S. Seifert, R.E. Winans, T.H. Fletcher, Faraday Discuss. 119 (2002) 395. [11] J.P. Hessler, S. Seifert, R.E. Winans, Proc. Combust. Inst. 29 (2002) 2743. [12] H. Wang, B. Zhao, B. Wyslouzil, K. Streletzky, Proc. Combust. Inst. 29 (2002) 2749. [13] C.M. Sorensen, C. Oh, P.W. Schmidt, T.P. Rieker, Phys. Rev. E 58 (1998) 4666. [14] A. Guinier, G. Fournet, Small Angle X-ray Scattering, Academic Press, New York, 1982.
In situ measurements of soot formation in simple flames using small angle X-ray scattering C. Gardner a, G.N. Greaves a,*, G.K. Hargrave b, S. Jarvis b, P. Wildman b, F. Meneau a,c, W. Bras c, G. Thomas a b
a Institute of Mathematical and Physical Sciences, University of Wales, Aberystwyth SY23 3BZ, United Kingdom Wolfson School of Mechanical and Manufacturing Engineering, Loughborough University, Loughborough LE11 3TU, United Kingdom c Netherlands Organisation for Scientific Research (NWO), DUBBLE CRG/ESRF, P.O. Box 220, F38043 Grenoble Cedex, France
Available online 1 September 2005
Abstract Direct SAXS measurements of soot formation from ethylene have been made using laminar pre-mixed flames for the first time. The slot burner was configured to maximise the signal from particulates. The geometry also enabled the thermal background from the surrounding hot gasses to be accurately removed. With cold flame speeds of 40 cm s1 we have been able to identify particle sizes and densities from moderately sooty to rich flame conditions. By adjusting the height of the burner in the beam, the development of particles as a function of position above the flame tip and therefore as a function of time from ignition have been obtained. These reveal evidence for bimodal particle nucleation and growth at different stages in the continuous combustion of ethylene. Ó 2005 Elsevier B.V. All rights reserved. PACS: 60.10.Eq; 61.46.+w; 81.05.Uw Keywords: Soot; Flame; Particle size; Nucleation; Aggregation; SAXS
1. Introduction Understanding soot generation in simple hydrocarbon fuels is fundamental to establishing practical strategies for increasing combustion efficiency, *
Corresponding author. Tel.: +44 1970 622802; fax: +44 1970 622826. E-mail address: [email protected] (G.N. Greaves).
at the same time as limiting particulate emissions. Quantitative knowledge of the processes from soot inception to the generation of visible particles has, until recently, relied on in situ static and dynamic optical techniques [1,2] and ex situ transmission electron microscopy (TEM) [3–5]. Optical techniques, however, are insensitive to particle sizes ˚ and therefore to the primary soot below 100 A particles that nucleate as well as the smaller
0168-583X/$ - see front matter Ó 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2005.06.072
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339
polycyclic aromatic hydrocarbons (PAHs) [6] in the flame. Ex situ methods like TEM and scanning mobility particle sizing (SMPS) [7,8], are able to ˚ , but the sampling measure sizes below 100 A probes generally result in significant particle coagulation. Despite extremely low volume fractions (108) and aerosol densities (1011 cm3) [7], in situ small angle X-ray scattering (SAXS) offers attractive possibilities for probing particulates in flames [5,9–11], but the main empirical challenge is to separate the scattering from soot from the scatter due to the high temperature gaseous background. In small angle neutron scattering (SANS) the background is dominated by incoherent scatter from hydrogen and water vapour [12] which masks hydrocarbon structure in the ‘‘dark region’’ between the soot particles. This, however, is accessible through SAXS because scattering from hydrogen is much smaller. Previously sooting flames have been ignited within sootless flames, enabling rapid in situ measurements to be made [10,11], but by incurring large temperature gradients that make the sootless background difficult to remove. Ultra
335
SAXS is necessarily ex situ, avoids the hot background problem, but requires more concentrated samples than the volume fraction available in living flames. As a result the PAHs and primary particles are heavily masked by surface scattering from the recovered large soot aggregates which ˚ in diameter [5]. Soot particles are typically 250 A ˚ in diameter are also analysed from ex situ of 460 A SAXS measurements [13] using form factor analysis [11].
2. Experiment Direct SAXS measurements of soot formation that overcome many of these difficulties have recently been made at the ESRF using BM26B. The experimental arrangement is shown in Fig. 1. Ten kilo-electron-volt X-rays and with two sample–detector lengths provided a wave vec˚ 1, encomtor range of 0.016 > 4p sin h/k > 0.3 A passing the scattering for the range of particle sizes expected within living flames. A stabilised water-cooled slot burner was employed,
Fig. 1. Schematic arrangement of laminar flow slot burner and experimental configuration used for SAXS.
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339
sufficiently long to maximise the signal from the soot and dark region and to minimise air scatter from outside the flame and the effects of temperature gradients at the edges. Precautions were also taken to preserve the integrity of the flame, such as the chimney and gauze guards. The burner, the chimney and the heat exchanger were mounted on a vertical traverse so that the region above the flame tip could be probed with the 0.3 mm beam up to 50 mm above the burner surface. A cold flame speed of 40 cm s1 was used, enabling measurements to be made for combustion times of up to 5/40 = 125 ± 1 ms. Measurements were made in both vertical directions in order to determine the size of systematic errors from the repeatability of the analysed SAXS parameters. Three ethylene/air flames were examined: a rich flame with an equivalence ratio, U = 2, a moderately sooty flame (U = 1.85) and a lean sootless flame (U = 1.3). The leaner flame provided background profiles at the same points above the burner where the data for the sooting flames were recorded. SAXS from both flames were obtained in pairs over 40 min exposures at each height above the burner. From separate thermocouple measurements for this burner sooting flames are known to be 50 °C hotter than leaner flames at the same height, resulting in small but measurable differences in scatter from the residual thermal density fluctuations. These were corrected for together with detector variations. SAXS profiles from the particles in the rich and moderately sooty flames were obtained by subtracting the leaner flame data from the sooting flame data at the same height. The resulting wave vector ranges were generally a decade smaller than those previously reported by other groups for SAXS [10,11,13] or for SANS [12] experiments. This enabled reliable SAXS measurements to be made of soot particles in their native environment without resorting to form factor analysis and covering the important ˚ down to 5 A ˚. size range from 100 A
3. Results Preliminary results are shown in Figs. 2–4. Fig. 2 shows a typical example of the SAXS data.
10
1
A
I(q) Relative Intensity
336
B
10
0
0.000
0.005 2
0.010
—2
q (Α )
Fig. 2. Example of background subtracted log I versus q2 from particles within the moderately sooty flame 22 mm above the burner obtained with 2.5 m flame-to-detector distance. Two regimes, marked A and B, relate the two different particle size groups identified at all heights above the burner.
Plotted as log I(q) versus q2, this exhibits two regimes, A and B. Both regimes were evident for all the flames and the different sample-to-detector distances employed enabled accurate data to be obtained for each region. For separated particle systems the SAXS 2 2 intensity at low q values, Rg q IðqÞ ¼ Ið0Þ exp [14], where Rg is the Gui3 nier radius and measures the particle size. Our results therefore point to the existence of two groups of particles, each with different characteristic sizes. Fig. 3 shows the Rg values obtained for the rich flame as a function of height above the burner and therefore of ignition time. For one group, average Rg values rise slowly from around ˚ to around 65 A ˚ , whereas for the other they 50 A ˚ to 40 A ˚ . The rise more sharply from around 25 A dotted line in Fig. 3 corresponds to Rg values of ˚ and was obtained from the leaner around 8 A flame used for background subtraction. These smaller particles vary little in size with height above the burner and relate to nanostructure in the residual gases within the flame.
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339
337
TIME (ms) 50
75
100
125
80
GUINIER RADIUS (ANGSTROMS)
70
A 60
50
40
B 30
20
10
0 15
20
25
30
35
40
45
50
HEIGHT (mm)
Fig. 3. Guinier radii for rich flame (U = 2) measured for sample-to-detector distances of 2.5 m (j) and 1.7 m (m). A and B relate to the two particle regimes identified in Fig. 2. The arrow marks the onset of SAXS signal taken from Fig. 4. The dotted line indicates the Guinier radii of the background sootless flame (U = 1.3). Curves are included to guide the eye. The combustion time, t, is related to the height above the burner, H, by t = H/m, where m is the flame velocity of 40 cm s1.
WEIGHTED INTEGRATED INTENSITY (arbitrary units)
TIME (ms) 50
75
100
125
4
3
2
1
0 15
20
25
30
35
40
45
50
HEIGHT (mm)
Fig. 4. Weighted integrated SAXS intensities for rich flame (U = 2) (j) compared to a moderately sooty flame (U = 1.85) (h), measured with a sample-to-detector distance of 2.5 m. Curves are included to guide the eye. The combustion time and height above the burner are related to the flame velocity as in Fig. 3.
338
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339
Fig. 4 shows the weighted integrated SAXS R qmax intensity, Q ¼ qmin IðqÞq2 dq, for the two flames with the same 2 m camera length. A step function can be clearly identified for both flames, the step sizes reflecting the different levels of soot being generated and the onsets the different delays for combustion. The error bars in Figs. 3 and 4 indicate the reproducibility of the measurements at the same height for repeated traverses of the burner at times separated by several hours. By comparison the individual statistical errors are far smaller, being contained within the sizes of the symbols. 4. Discussion The abrupt rise in Q for both flames in Fig. 4 illustrates the rapid nucleation of soot in laminar pre-mixed flames. This occurs for ethylene within 10 mm of the flame tip or 25 ms from ignition and compares well with the rise time in number density of 24 ms obtained from SMPS experiments for equivalent flames (U = 2.07) [7]. A similar figure emerges from SANS measurements of rich ethylene flames [12]. Turning to the particle sizes presented in Fig. 3, these clearly fall into three groups. For leaner flames the molecular particles ˚ ) whilst for rich flames the largare small (Rg 8 A ˚ both values est soot particles have Rgs of 60 A are relatively constant with height above the burner surface. The sizes of the third group are intermediate between these limits, growing progressively from sizes close to the molecular structure at the flame tip and eventually attaining ˚ further up the flame. values with Rgs of 40 A In order to relate these particle sizes with those measured by different techniques, we recall that TEM images of primary soot particles sampled from combusting fuels like ethylene reveal ‘‘onion-like’’ shapesp [4,5]. For spherical particles the radius equals (5/3)Rg [14]. Scaled accordingly, the largest particles we measure have average ˚ , in good agreement with diameters of 140–160 A the mean diameters for the largest particles reported from SANS data on rich ethylene flames from form factor analysis [12]. Our in situ SAXS ˚ diameter results are also similar to the 120–140 A values for particulates sampled from diesel engines
measured by USAXS [5], in this case deconvoluted from surface scattering from much larger soot aggregates. These authors also report evidence for much smaller ‘‘sub unit’’ particles with diameters ˚ , close to the 20 A ˚ particle sizes we measure of 17 A directly from in situ SAXS from lean ethylene flames (dotted line in Fig. 3). Turning finally to the particles of intermediate size which grow during soot formation, these appear to nucleate from the ˚ ) and aggregate to create the larger priPAHs (20 A ˚ ) – a bimodal process that mary particles (160 A starts almost immediately at the flame tip. Bimodal particle size distribution functions have recently been reported for rich ethylene flames (U 2), using ex situ sampling methods in conjunction with SMPS [7,8]. These develop with height above the burner surface. Both groups report bimodality, with an exponential component stemming from smaller particles extending to ˚ in diameter, together with a very around 100 A broad lognormal component centred around parti˚ but with r values as large as cle diameters of 200 A 1.5 [7]. Our in situ experiments reveal smaller diameters for both groups of particles (Fig. 3). Moreover the linearity of the two Guinier regions A and B (Fig. 2) suggest far less polydispersity in soot particles probed in situ than sampled ex situ. Finally our SAXS data obtained from laminar living flames offer precise height/time information on the way the different particle groups develop along the flame.
Acknowledgements Experimental work was supported by the UK Engineering and Physical Sciences Research Council (EPSRC), the Netherlands Organisation for Scientific Research (NWO) and the European Synchrotron Radiation Source (ESRF).
References [1] A. DÕAlessio, in: D.C. Siegla, G.W. Smith (Eds.), Particulate carbon – Formation during combustion, Plenum, New York, 1982, p. 207. [2] G. Kroner, H. Fuchs, R. Tatschl, O. Glatter, Part. Part. Syst. Char. 20 (2003) 111.
C. Gardner et al. / Nucl. Instr. and Meth. in Phys. Res. B 238 (2005) 334–339 [3] H. Bockhorn, F. Fetting, A. Heddrich, U. Meyer, G. Wannemarcher, J. Aerosol Sci. 19 (1988) 591, and references therein. [4] S. di Stasio, Carbon 39 (2001) 109. [5] A. Braun, F.E. Huggins, S. Seifert, J. Ilavsky, N. Shah, K.E. Kelly, A. Sarofim, G.P. Huffman, Combust. Flame 137 (2004) 63. [6] H. Wang, M. Frenklach, Combust. Flame 96 (1994) 163. [7] B. Zhao, Z. Yang, M.V. Johnston, H. Wang, A.S. Wexler, M. Balthasar, M. Kraft, Combust. Flame 133 (2003) 173. [8] M.M. Maricq, Combust. Flame 137 (2004) 340.
339
[9] W.A. England, Combust. Sci. Technol. 46 (1986) 83. [10] J.P. Hessler, S. Seifert, R.E. Winans, T.H. Fletcher, Faraday Discuss. 119 (2002) 395. [11] J.P. Hessler, S. Seifert, R.E. Winans, Proc. Combust. Inst. 29 (2002) 2743. [12] H. Wang, B. Zhao, B. Wyslouzil, K. Streletzky, Proc. Combust. Inst. 29 (2002) 2749. [13] C.M. Sorensen, C. Oh, P.W. Schmidt, T.P. Rieker, Phys. Rev. E 58 (1998) 4666. [14] A. Guinier, G. Fournet, Small Angle X-ray Scattering, Academic Press, New York, 1982.