Multiwavelength Raman Microspectroscopy for Rapid Prediction of

Jan 24, 2011 - Alexander Rinkenburger , Takaaki Toriyama , Kazuhiro Yasuda , Reinhard ... Chethan K. Gaddam , Randy L. Vander Wal , Xu Chen , Aleksey ...
0 downloads 0 Views 2MB Size
Download PDF
LETTER pubs.acs.org/ac

Multiwavelength Raman Microspectroscopy for Rapid Prediction of Soot Oxidation Reactivity Johannes Schmid, Benedikt Grob, Reinhard Niessner, and Natalia P. Ivleva* Chair for Analytical Chemistry, Institute of Hydrochemistry, Technische Universit€at M€unchen, Marchioninistrasse 17, D-81377 Munich, Germany

bS Supporting Information ABSTRACT: Multiwavelength Raman microspectroscopy (MWRM) analysis for characterization of soot structure and reactivity was developed. This new method is based on the dispersive character of carbon D mode in Raman spectra (i.e., red shift and increase in intensity at higher excitation wavelength, λ0). The approach was proven by investigating various diesel soot samples and related carbonaceous materials at different λ0 (785, 633, 532, and 514 nm). In order to compare the behavior of the D mode for various samples and to derive a single parameter characterizing the soot structure, the difference of integrals for pairs of spectra collected at different λ0 was calculated. MWRM analysis revealed substantial differences in the structural ordering which decreases from graphite, over Printex XE2 and various diesel soot samples, to spark discharge soot. To obtain the relation between structure and reactivity of soot, MWRM analysis was combined with temperature-programmed oxidation (TPO). TPO allowed us to characterize the oxidation behavior of soot in terms of the maximum emission (CO þ CO2) temperature and reactivity index. The latter was calculated by introducing the reactivity limits: spark discharge soot containing a large amount of disorder represents the upper limit, whereas the lower limit is given by graphite powder with high structural order. The comparison of MWRM (viz., the observed Raman difference integrals) and TPO data revealed a linear correlation between soot structure and oxidation reactivity. Thus, we demonstrated for the first time the potential of MWRM for a robust and rapid prediction of diesel soot reactivity based on the structure-reactivity correlation.

T

he reduction of anthropogenic aerosol sources is essential for the overall improvement of air quality.1 Soot nanoparticles emitted by diesel engines account for a major fraction of air pollutants in urban areas. To meet the present and future emission limits, soot particles must be removed from the engine exhaust.2 Therefore, a wide range of particle trapping systems and exhaust aftertreatment technologies are currently under investigation. Diesel particulate filters, which have been applied for this purpose, need to be regenerated by gasification of the deposited soot.3-7 The efficiency of this regeneration step is strongly affected by the oxidation reactivity of the deposited soot particles.6,8-10 In particular, the formation of highly reactive soot would make it possible to reduce energy consumption at the regeneration step (due to shorter combustion times and lower temperatures).10 Thermo-analytical methods are usually applied for analysis of soot oxidation reactivity. In particular, thermo-gravimetric analysis (TGA), measuring the mass decrease, and temperatureprogrammed oxidation (TPO), measuring the gasification products by mass spectrometry8 or infrared spectroscopy,5 are used. Additionally, the information on functional groups and desorbing r 2011 American Chemical Society

fractions can be deduced from diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) and temperature-programmed desorption mass spectroscopy (TPD-MS) analysis.11,12 Recent studies showed that the soot reactivity is clearly connected to the structural order of soot.6,8,9,13-16 Thus, knowledge about the soot structure can help in predicting soot reactivity. Usually high-resolution transmission electron microscopy (HRTEM) is applied for investigation of soot structure,6,8,9,16,17 but this method is too demanding for routine analysis. The X-ray diffraction (XRD) technique, which allows one to overcome the limitation of HRTEM imaging associated with probing a nanoscale area, can also be used for structural analysis. However, for some diesel soot samples, no clear relation has been found between stacking order of the graphene layers (XRD) and oxidation rate (TGA).14,15 Therefore, it is highly desirable to establish a rapid and robust analytical tool for determining the soot structure and reactivity. Received: November 9, 2010 Accepted: January 14, 2011 Published: January 24, 2011 1173

dx.doi.org/10.1021/ac102939w | Anal. Chem. 2011, 83, 1173–1179

Analytical Chemistry Raman spectroscopy is a standard nondestructive tool for the structural characterization of carbonaceous materials based on vibrational fingerprint spectra.18 When Raman spectroscopy is combined with optical microscopy, Raman microspectroscopy (RM), it is possible to obtain full spectral information with spatial resolution in the micrometer range. A link between Raman spectroscopic parameters and the structure of carbonaceous materials, in particular soot, has already been discussed in the literature.13,15,19-32 RM has been used for the analysis of changes in the chemical structure of soot oxidized in air13 as well as at conditions relevant to common diesel aftertreatment systems (i.e., oxidation with nitrogen dioxide23 or 5% oxygen33 in nitrogen). Recently, we combined RM, HRTEM, electron energy loss spectroscopy (EELS), and TPO to obtain comprehensive information on soot structure and reactivity. The better insight into the complex heterogenic structure of soot provided by HRTEM and EELS allowed us to validate RM as an efficient tool for analyzing structural ordering of soot.26 The development of modern diesel engines can result in more reactive soot particles. Moreover, NOx reducing methods (which use the trade-off between soot and NOx) additionally increase soot emissions.13,34,35 Improved knowledge about the structure and reactivity can facilitate the optimization process of diesel exhaust aftertreatment systems. It is important to note that, in order to correlate structure and reactivity, both characteristics have to be determined by rapid, robust, and well-defined approaches. However, up to now, there is also a need for a clear definition of soot reactivity that can help to describe the oxidation behavior of soot with one single parameter. On the other hand, structural analysis should provide a reliable ordering parameter. In this study, we propose multiwavelength Raman microspectroscopy (MWRM) analysis of soot, which allows for rapid characterization of diesel soot structure by implementing the dispersive character of the D mode in Raman spectra. Moreover, we introduce a robust index, which describes the oxidation behavior of soot in terms of the maximum temperature of CO and CO2 emission. Analysis of the relation between structural (MWRM) and reactivity (TPO) data allowed us to derive for the first time a direct correlation between soot structure and reactivity of complex diesel soot samples and related carbon materials and to validate MWRM as an effective tool for predicting soot reactivity.

’ EXPERIMENTAL SECTION Materials. Different types of soot and carbonaceous materials have been analyzed. Spark discharge soot generated by a spark discharge aerosol generator (GfG 1000, Palas, Karlsruhe, Germany) was used as the upper limit for reactivity studies, and a commercially available graphite powder (C g 99.0%, particle size of 1 to 2 μm (Fluka/Sigma-Aldrich, St. Louis, USA)) was used as the lower limit. As the reference for ideal graphitic lattices, a highly oriented pyrolytic graphite (HOPG) from Mateck (Juelich, Germany) was used. Printex XE-2 soot was purchased from Degussa AG (Frankfurt, Germany). The diesel soot samples (DS1-DS12) were collected from different engines at varying driving cycles. Sampling of Soot Samples. The sampling of aerosol particles, especially in diesel exhaust systems, needs to be done carefully. Usually soot samples are generated by filtering a welldefined and diluted partial exhaust stream at stable conditions.

LETTER

As the diesel particulate filters, where the soot reactivity is most relevant, are located within the exhaust system, a filter sampling within the exhaust pipe is a direct approach to gain easy accessible soot surfaces. The filter material needs to withstand high back pressure, temperature, and high mass flow. Moreover, the temperature at high loads, which could be high enough to cause partial oxidation, needs to be monitored. In this study, all diesel soot samples where collected by a pretreated metal fiber fleece filter (Bekipor; Kortrijk Bekaert, Belgium)26 at full flow or side stream conditions. To prevent systematic errors due to volatile compounds, all samples where preheated in filtered air at 423 K for 30 min. Raman Microspectroscopy, RM. The Raman spectra have been recorded by two different Raman microscope systems. A Renishaw 2000 Raman microscope system (Renishaw, Wottonunder-Edge, U.K.), using laser wavelengths of λ0 = 514 nm (Arþ laser), 25 mW, and λ0 = 633 nm (He-Ne laser), 25 mW was used. Calibration was performed with a silicon wafer by utilizing the first-order phonon band of Si at 520 cm-1. Spectra on the soot surface were taken in the range of 800-2000 cm-1 (Stokes shift) with a 50 magnification objective (NA = 0.75, Leica, Wetzlar, Germany). The integration time with 10-60 s and the number of accumulated spectra from 5 to 20 were adjusted to obtain an optimal signal-to-noise ratio. A defocused laser beam (diameter ≈ 40 μm) and 25% of the source power were applied to avoid laser-induced decomposition of the soot samples. On every sample, 5-7 randomized spots were measured. The spectra were multipoint baseline corrected via WiRE 1.2 software (Renishaw, Wotton-under-Edge, U.K.) running under GRAMS/ 32 (Thermo Galactic, Waltham, USA). A Horiba LabRAM HR Raman microscope system (Horiba Jobin Yvon, Kyoto, Japan), using laser wavelengths of λ0;1 = 532 nm (frequency doubled Nd:YAG-solid phase laser), 100 mW, λ0;2 = 633 nm (He-Ne laser), 35 mW, and λ0;3 = 785 nm (diode laser), 300 mW, was used. Calibration was performed with a silicon wafer and additionally by zero-order-correction of the used grating. Soot spectra were obtained in the range of 800-2000 cm-1 (Stokes shift) with a 50 magnification objective (NA = 0.75, Olympus, Tokyo, Germany). The laser beam intensity was reduced to 1-0.1% and distributed over a 15 μm  15 μm surface using DuoScan. For 120 s integration time and 15 accumulated spectra, a very good signal-to-noise ratio was obtained. Each sample was measured at 3-5 randomly distributed spots. The spectra were multipoint baseline corrected via LabSpec 5.58.25 software (Horiba Jobin Yvon, Kyoto, Japan). For each sample, the spectra were normalized with Matlab 7.0.4 (The Mathworks, Natick, USA) at the G peak around 1600 cm-1. One mean spectrum per sample and λ0 was obtained. The spectrum collected at the lower λ0 was subtracted from that at higher λ0, resulting in the difference spectrum that could be integrated afterward. It should be noted that the signal-to-noise ratio, the G mode position, and the baseline correction are critical parameters upon collecting and processing spectra. Temperature-Programmed Oxidation, TPO. A diesel exhaust aftertreatment model system4 has been modified and partially rebuilt to perform the oxidation experiments within a temperature range from 373 K up to 973 K (or 1053 K), with a total gas flow of 3 L/min of nitrogen including 5% of oxygen. The heating rate during the experiments was set to 5 K/min. The temperature of the gas stream was adjusted with a temperature controller, using a type K thermocouple that was placed in the immediate vicinity of the filter substrate surface. Analysis of 1174

dx.doi.org/10.1021/ac102939w |Anal. Chem. 2011, 83, 1173–1179

Analytical Chemistry

LETTER

Figure 1. Raman spectra of (a) HOPG, (b) graphite powder, (c) DS 12, and (d) spark discharge soot (sorted by increasing structural disorder) measured at different excitation wavelengths (λ0;1 = 532 nm, λ0;2 = 633 nm, λ0;3 = 785 nm) using the Horiba system. The grayish areas are the result of the subtraction of the λ0;1 spectra from the λ0;3 spectra for dark gray and the λ0;2 spectra for light gray, respectively.

oxidation products was accomplished by IR spectroscopy using a IFS 66/s FTIR spectrometer (Bruker, Billerica, Germany), equipped with a 2 L gas flow cell with an optical path length of 6.4 m.4

’ RESULTS AND DISCUSSION Structure Analysis by Multiwavelength Raman Microspectroscopy, MWRM. Many different attempts have been

made to characterize soot structure by RM. Raman spectra (λ0 = 514 nm) of soot consist of two strong overlapping bands around 1600 cm-1 (G or “graphite”) and 1350 cm-1 (D or “defect”). Therefore, in order to derive structural information, these spectra are usually fitted with band combinations ranging from two36 (G, D) to five bands23,24,26,28 (G, D1-D4 around 1580, 1350, 1620, 1500, and 1200 cm-1, respectively). The G band is assigned to ideal graphitic layers, D1 and D2 to disordered graphitic structures, D3 to amorphous or molecular carbon, and D4 to disordered graphitic lattices and inorganic and organic impurities.28 We successfully used the five-band fitting28 for analysis of structure of a large variety of different diesel soot samples and related carbonaceous materials.23,24,26,28,33 However, for some soot samples, we observed unusual signals in the D4 area. The form of these signals varied from broad shoulder(s) to small narrow peak(s). The presence of such signal(s) did not allow us to perform the five-band fitting properly without introducing an additional band (see Supporting Information, Figure SI 1). Similar limitations of the fitting procedure have been reported by Liu et al.37 and Larouche et al.38 To be able to investigate all soot samples (also in the presence of unexpected signals), we searched for a robust method to

interpret Raman spectra. In a recent study,26 we have shown that the dispersive character of D mode, i.e., red shift and increase in intensity at higher λ0, can be used to describe structural properties of carbon materials. The dispersive character of the D mode is well-known for different kinds of carbonaceous materials, such as crystalline graphite, glassy carbon, annealed hydrogenated amorphous carbon, multicomponent carbon films, and carbon nanotubes.39-49 The origin and dispersion of the D mode was explained by Matthews et al. in terms of the resonant Raman coupling between electrons and phonons containing the same wave vector near the K point of the Brillouin zone.40 Reich and Thomsen have attributed the dispersive character of the D mode to the double-resonant Raman process, which (for a given laser energy and phonon branch) selectively enhances a particular phonon wave vector and phonon frequency.44 Sood et al. have proposed double resonance and disorder-induced two-phonon scattering combining an optic photon at the K point in the Brillouin zone and acoustic phonon (whose momentum is determined uniquely by the double resonance condition) to explain the D peak dispersion.45 By utilizing the dispersive character of the D mode, we developed multiwavelength Raman microspectroscopy (i.e., MWRM) analysis of soot, which allowed us to obtain information on the structural order of soot and to predict the reactivity of real diesel soot samples based on a structure-reactivity correlation. The efficiency of MWRM was demonstrated by investigating carbonaceous materials with different structural order at three λ0 (532, 633, and 785 nm) using the Horiba Raman system (see Figure 1). Theoretically, different excitation wavelengths λ0 should have no effect on the position of Raman bands. Indeed, the peak for 1175

dx.doi.org/10.1021/ac102939w |Anal. Chem. 2011, 83, 1173–1179

Analytical Chemistry

LETTER

Figure 2. Oxidation behavior of soot with increasing temperature described by (a) emission of oxidation products and (b) mass conversion for three diesel soot samples and reactivity limits (spark discharge soot and graphite).

ideal graphitic lattice vibration at 1580 cm-1 in spectra of HOPG is independent of λ0 (Figure 1a). This invariance of the G mode has been well described in the literature for different carbonaceous materials.18,28,40-42,47,48 However, the D mode of disordered graphitic lattice in spectra of graphite powder (Figure 1b) collected at λ0 = 532 nm (2.33 eV), 633 nm (1.96 eV), and 785 nm (1.58 eV) demonstrates an increase in intensity and the red shift of the peak, from 1348 cm-1 to 1335 cm-1 and 1314 cm-1. That is in good agreement with the published data for the D peak shift (40-50 cm-1/eV).46 Additionally, a small wavelength dependent shoulder was observed at 1620 cm-1 that can be assigned to the vibrations of the surface graphene layers of disordered graphitic lattices.28 We chose different diesel soot (DS) samples as representatives of higher structural disorder. For these materials, the dispersive character of D mode is noticeably more pronounced (Figure 1c). All soot spectra consist of a strong D peak which is partially overlapping with the G peak. For DS samples (Figure 1c) in this region, a minor intensity increase at higher λ0 was observed. The D peak is now shifted from 1345 cm-1 to 1335 cm-1 and 1312 cm-1. Spark discharge soot, which is highly disordered and, therefore, highly reactive (selected as the upper reactivity limit) shows even more significant overlap of peaks (Figure 1d) and exhibits the highest observed dispersive character. The D peak position is shifted from 1338 cm-1 to 1320 cm-1 and 1302 cm-1. For better structural characterization, we obtained difference spectra for various carbon samples. The difference spectrum was achieved by subtracting the spectrum collected at the lower λ0 from the spectrum at higher λ0. The difference spectra (area beneath the spectra) are plotted in Figure 1. With increasing structural disorder (from Figure 1a-d), the differences between the spectra become more pronounced. Furthermore, a difference integral (DI) can be obtained, which transfers the structural information performed by MWRM into a single parameter that can be used for structure-reactivity correlations discussed below. Reactivity of Soot. In order to validate the structural data obtained by MWRM, the reactivity of soot was investigated in detail. As soot is an inhomogeneous carbonaceous aerosol with different size distributions, element impurities, and condensed substances on the surface, the analysis of the overall oxidation behavior is not trivial. Using a temperature at which the overall mass is reduced to, e.g., 5% as the measure of reactivity bears the risk of different artifacts (e.g., of not taking into consideration high amounts of ash content and diffusion protected soot

particles).22 Interrupting the oxidation process (e.g., at 773 K) and focusing on the partial mass conversion at this point can help to understand structural changes but gives only limited access to reactivity since the oxidation process is not completely monitored. During a full thermo-analytical method, the emission of carbon oxides as function of temperature is easily accessible. This dependence has a very narrow peak which determines the temperature necessary for diesel particulate filter regeneration. In this study, we applied TPO to get detailed information on the oxidation behavior of diesel soot samples and related carbonaceous materials. Figure 2a shows the CO and CO2 emission of a selection of samples, while Figure 2b illustrates the oxidation process in mass conversion versus temperature. Graphite powder exhibits almost no emission up to 773 K, at higher temperatures a small disordered fraction starts reacting. This results in an exponential increase of oxidation, which eventually peaks at 1033 K. However, technical limitations of our test bench prevented oxidation to full mass conversion. The DS 6, DS 7, and DS 9 samples exhibit a representative oxidation behavior for diesel soot. Oxidation typically starts at 523 K with a small emission shoulder or peak at 623 K. This can be assigned to amorphous or highly disordered structures which dominate the oxidation process at low temperatures. Beyond 698 K, emission increases rapidly, peaking in the range between 800 and 900 K (depending on the structural properties of the soot). After a relatively sharp maximum, the emission ceases within a few measuring points. Since spark discharge soot contains a large amount of amorphous domains and disordered graphitic structures, the early oxidation behavior is very pronounced.26 The main oxidation peak appearing at 788 K is broad in comparison with other soot samples and shows a slow decline after the maximum. It should be mentioned that full oxidation for spark discharge soot cannot be achieved up to 973 K, suggesting that spark discharge soot either contains small amounts of highly ordered graphite (possibly from the graphite electrode) or/and is graphitized during heating up. This is in good agreement with different gravimetric oxidation experiments showing a small percentage of residual mass. Structure-Reactivity Correlation of Real Diesel Soot Samples. When a well-defined reactivity (TPO data) is combined with structural information (MWRM data), a structure-reactivity correlation can be obtained. In order to interpret the TPO data more easily, a reactivity index RIsoot is introduced. Spark discharge soot represents the high reactivity limit and, therefore, 1176

dx.doi.org/10.1021/ac102939w |Anal. Chem. 2011, 83, 1173–1179

Analytical Chemistry

LETTER

Figure 3. Correlations of reactivity with the structure of soot. Correlation of the maximum emission (CO2 þ CO) temperature and the reactivity index, respectively, versus the difference integral for various soot samples and carbonaceous materials. DI633-514nm (middle, blue) were measured at the Renishaw system while DI785-532nm (upper, red) and DI633-532nm (lower, black) were measured at the Horiba system. For the linear equation (DI = s  T þ t) of the black data points, s = -0.2554 and t = 245.4 (black filled line, R2 = 0.998, n = 3, m = 1). For the blue data points, s = -0.2369 and t = 281.1 (blue dashed line, R2 = 0.994, n = 10, m = 1), and for the red data points, s = -0.1910 and t = 206.6 (red pointed line, R2 = 0.990, n = 4, m = 1).

RIsoot is set to 100%. In contrast, graphite powder consists of stacks of graphene layers with predominant well-defined longrange ordering26 and, therefore, represents the lower reactivity limit (0%), with T being the maximum emission (CO2 þ CO) temperature. ! TSoot - TGraphite ð1Þ RIsoot ¼ 100 TGf G - TGraphite This index can be used to compare results from different thermoanalytical methods, e.g., TGA vs TPO, more easily and enables for correlation of structure and reactivity. Additionally, on the basis of T and RIsoot (see Figure 3), samples can be classified as very reactive (T e 823 K or RIsoot g 85%), reactive (823 K < T e 873 K or 85% > RIsoot g 65%), and less reactive soot (T > 873 K or RIsoot < 65%). From this classification, DS 1, DS 2, and DS 11 are found to be very reactive toward oxidation, and their generation conditions should be investigated further. DS 3-DS 6 and DS 12 can be classified as reactive soot samples and should enable readily regeneration of DPF structures. However, DS 7-DS 10 appear to be more resistant toward oxidation in comparison to other diesel soot samples. Figure 3 shows three series of DI determined by MWRM reflecting soot structure plotted against reactivity data, T and RIsoot obtained by TPO. A clear link between MWRM and TPO data can be observed that enables a structure-reactivity correlation. For the DI633-514 nm (middle data set, blue), a very good correlation can be found for the measured reactivity limits and diesel soot samples, while two points slightly deviate from the linear trend (DS 8 and DS 9).

To further prove the multiwavelength approach, two additional excitation wavelength differences were used to obtain the DI785-532nm (upper, red) and the DI633-532nm (lower, black) data set. The investigated samples (graphite, Printex XE-2, DS 11 and DS 12, and spark discharge soot) also show the same trend of DI with increasing T (or decreasing RIsoot). Only graphite strongly deviates from the trend in the lower data set and DS 11 in the upper and lower sets. These samples show the limitations of this approach. For the deviating graphite data point, the wavelength difference (101 nm) seems to be too small for a linear trend below 923 K, since this would result in a negative integral. However, the use of a larger wavelength difference (129 and 253 nm in the middle and upper data sets, respectively) allowed us to overcome this problem. DS 11 has shown a significantly higher reactivity than expected from the DI values, which can be explained by impurities detected in this sample. We found a residual white ash on the filter surface after oxidation. Raman analysis revealed the presence of sulfate and different iron oxides as residual material (see Supporting Information, Figure SI 2). These impurities could act as an oxygen carrier or could be catalytically active. In other ongoing studies, we observed a similar increase in reactivity when iron oxides and/or oxide hydroxide were present as an internal mixture within the studied particles. Moreover, Song et al.14,15 have found that not only the initial diesel soot structure but also the initial oxygen groups have a strong influence on the oxidation rate. Additionally, since our MWRM and the TPO experiments of one diesel soot type were carried out on two different filter samples that were collected next to each other or consecutively with the same engine conditions, a deviation between samples cannot be excluded completely. 1177

dx.doi.org/10.1021/ac102939w |Anal. Chem. 2011, 83, 1173–1179

Analytical Chemistry Generally, the correlations can be observed at various wavelength differences with similar linear trends. These trends appear to be shifted almost parallel depending on the wavelength difference and, therefore, show similar slopes. The calculated average of the slopes is -0.2278 with a standard deviation of 0.0332. This clearly shows that the MWRM can be applied for various wavelength differences and for a large number of different soot samples.

’ CONCLUSIONS In this study, we developed the novel multiwavelength Raman microspectroscopy analysis of soot, to determine the structure of carbonaceous materials by utilizing the dispersive character of the D mode in Raman spectra. The dispersion can be explained by double phonon resonance46 which is dependent on the abundance of defects in the graphitic layers. This multiwavelength approach was used to generate a single value (DI) describing structural properties. To validate the structural information by reactivity data, we applied TPO. We discussed different approaches to determine the reactivity and defined a robust value that describes the oxidation behavior using the maximum emission (CO þ CO2) temperature and reactivity index. The latter was calculated by introducing the reactivity limits: spark discharge soot containing a large amount of disorder represents the upper limit whereas the lower limit is given by graphite powder with high structural order. Altogether, it allowed us to validate the MWRM structural analysis and demonstrated for the first time a clear correlation between structure and reactivity. It has clearly been shown that the MWRM is a rapid analytical tool to predict the oxidation behavior of diesel soot samples and other carbon materials. This opens up a novel approach for the prediction of soot reactivity that could be used to promote the optimization of diesel engine parameters and aftertreatment systems for a fast and energy efficient diesel particulate filter regeneration. Moreover, knowledge about reactivity of soot nanoparticles can help in understanding their influence on public health and environmental chemistry. ’ ASSOCIATED CONTENT

bS

Supporting Information. Additional data on Raman analysis of soot samples. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Tel: þ49 89 2180 78238. Fax: þ49 89 2180 78255.

’ ACKNOWLEDGMENT Financial support by the Hanns-Seidel-Stiftung (BMBF funding) is gratefully acknowledged. We thank D. Rothe (MAN Nutzfahrzeuge AG, N€urnberg; Chairman of the FVV, Forschungsvereinigung Verbrennungskraftmaschinen e.V., research project “Filter regeneration by reactive soot”) for helpful discussions and diesel soot samples DS 11 and DS 12. Additionally, we thank M. Knauer for fruitful discussions. ’ REFERENCES (1) Bernstein, J. A.; Alexis, N.; Barnes, C.; Bernstein, I. L.; Bernstein, J. A.; Nel, A.; Peden, D.; Diaz-Sanchez, D.; Tarlo Susan, M.; Williams, P. B. J. Allergy Clin. Immunol. 2004, 114, 1116–1123.

LETTER

(2) Sydbom, A.; Blomberg, A.; Parnia, S.; Stenfors, N.; Sandstrom, T.; Dahlen, S. E. Eur. Respir. J. 2001, 17, 733–746. (3) Higgins, K. J.; Jung, H.; Kittelson, D. B.; Roberts, J. T.; Zachariah, M. R. J. Phys. Chem. A 2002, 106, 96–103. (4) Messerer, A.; Niessner, R.; P€ oschl, U. Carbon 2006, 44, 307–324. (5) Messerer, A.; Rothe, D.; P€ oschl, U.; Niessner, R. Top. Catal. 2004, 30/31, 247–250. (6) M€uller, J. O.; Su, D. S.; Jentoft, R. E.; Wild, U.; Schl€ogl, R. Environ. Sci. Technol. 2006, 40, 1231–1236. (7) Saracco, G.; Badini, C.; Specchia, V. Chem. Eng. Sci. 1999, 54, 3035–3041. (8) M€uller, J. O.; Su, D. S.; Jentoft, R. E.; Kr€ohnert, J.; Jentoft, F. C.; Schl€ ogl, R. Catal. Today 2005, 102-103, 259–265. (9) Su, D. S.; Jentoft, R. E.; M€uller, J. O.; Rothe, D.; Jacob, E.; Simpson, C. D.; Tomovic, Z.; M€ullen, K.; Messerer, A.; P€oschl, U.; Niessner, R.; Schl€ogl, R. Catal. Today 2004, 90, 127–132. (10) Su, D. S.; M€uller, J. O.; Jentoft, R. E.; Rothe, D.; Jacob, E.; Schl€ ogl, R. Top. Catal. 2004, 30/31, 241–245. (11) Muckenhuber, H.; Grothe, H. Carbon 2006, 44, 546–559. (12) Muckenhuber, H.; Grothe, H. Carbon 2007, 45, 321–329. (13) Al-Qurashi, K.; Boehman, A. L. Combust. Flame 2008, 155, 675–695. (14) Song, J.; Alam, M.; Boehman, A. L. Combust. Sci. Technol. 2007, 179, 1991–2037. (15) Song, J.; Alam, M.; Boehman, A. L.; Kim, U. Combust. Flame 2006, 146, 589–604. (16) Vander Wal, R. L.; Tomasek, A. J. Combust. Flame 2003, 134, 1–9. (17) Vander Wal, R. L.; Tomasek, A. J. Combust. Flame 2004, 136, 129–140. (18) Ferrari, A. C.; Robertson, J. Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61, 14095–14107. (19) Beyssac, O.; Goffe, B.; Petitet, J.-P.; Froigneux, E.; Moreau, M.; Rouzaud, J.-N. Spectrochim. Acta, Part A: Mol. Biomol. Spectrosc. 2003, 59A, 2267–2276. (20) Charbonneau, P.; Wallace, J. S. Soc. Automot. Eng., [Spec. Publ.] SP 2009, SP-2254, 637–651 (Diesel Exhaust Emission Control). (21) Dippel, B.; Jander, H.; Heintzenberg, J. Phys. Chem. Chem. Phys. 1999, 1, 4707–4712. (22) Fang, H. L.; Lance, M. J. Soc. Automot. Eng., SP 2004, SP-1898, 89–97. (23) Ivleva, N. P.; McKeon, U.; Niessner, R.; P€oschl, U. Aerosol Sci. Technol. 2007, 41, 655–671. (24) Ivleva, N. P.; Messerer, A.; Yang, X.; Niessner, R.; P€ oschl, U. Environ. Sci. Technol. 2007, 41, 3702–3707. (25) Jawhari, T.; Roid, A.; Casado, J. Carbon 1995, 33, 1561–1565. (26) Knauer, M.; Schuster, M. E.; Su, D. S.; Schl€ogl, R.; Niessner, R.; Ivleva, N. P. J. Phys. Chem. A 2009, 113, 13871–13880. (27) Rosen, N.; Novakov, T. Nature 1977, 266, 708–710. (28) Sadezky, A.; Muckenhuber, H.; Grothe, H.; Niessner, R.; P€oschl, U. Carbon 2005, 43, 1731–1742. (29) Sheng, C. Fuel 2007, 86, 2316–2324. (30) Sze, S.-K.; Siddique, N.; Sloan, J. J.; Escribano, R. Atmos. Environ. 2001, 35, 561–568. (31) Vander Wal, R. L.; Tomasek, A. J.; Street, K.; Hull, D. R.; Thompson, W. K. Appl. Spectrosc. 2004, 58, 230–237. (32) Zhu, X.-L.; Sheng, C.-D. Fuel Process. Technol. 2010, 91, 837–842. (33) Knauer, M.; Carrara, M.; Rothe, D.; Niessner, R.; Ivleva, N. P. Aerosol Sci. Technol. 2009, 43, 1–8. (34) Lloyd, A. C.; Cackette, T. A. J. Air Waste Manage. Assoc. 2001, 51, 809–847. (35) Schmid, M.; Leipertz, A. Int. J. Vehicle Design 2006, 41, 188– 205. (36) Tuinstra, F.; Koenig, J. L. J. Chem. Phys. 1970, 53, 1126–1130. (37) Liu, Y.; Liu, C.; Ma, J.; Ma, Q.; He, H. Phys. Chem. Chem. Phys. 2010, 12, 10896–10903. (38) Larouche, N.; Stansfield, B. L. Carbon 2010, 48, 620–629. 1178

dx.doi.org/10.1021/ac102939w |Anal. Chem. 2011, 83, 1173–1179

Analytical Chemistry

LETTER

(39) Kastner, J.; Pichler, T.; Kuzmany, H.; Curran, S.; Blau, W.; Weldon, D. N.; Delamesiere, M.; Draper, S.; Zandbergen, H. Chem. Phys. Lett. 1994, 221, 53–58. (40) Matthews, M. J.; Pimenta, M. A.; Dresselhaus, G.; Dresselhaus, M. S.; Endo, M. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, R6585–R6588. (41) Pocsik, I.; Hundhausen, M.; Koos, M.; Ley, L. J. Non-Cryst. Solids 1998, 227-230, 1083–1086. (42) Ramsteiner, M.; Wagner, J. Appl. Phys. Lett. 1987, 51, 1355– 1357. (43) Rao, A. M.; Richter, E.; Bandow, S.; Chase, B.; Eklund, P. C.; Williams, K. A.; Fang, S.; Subbaswamy, K. R.; Menon, M.; Thess, A.; Smalley, R. E.; Dresselhaus, G.; Dresselhaus, M. S. Science 1997, 275, 187–191. (44) Reich, S.; Thomsen, C. Philos. Trans. R. Soc. Lond. A 2004, 362, 2271–2288. (45) Sood, A. K.; Gupta, R.; Asher, S. A. J. Appl. Phys. 2001, 90, 4494–4497. (46) Thomsen, C.; Reich, S. Phys. Rev. Lett. 2000, 85, 5214–5217. (47) Vidano, R. P.; Fischbach, D. B.; Willis, L. J.; Loehr, T. M. Solid State Commun. 1981, 39, 341–344. (48) Wang, Y.; Alsmeyer, D. C.; McCreery, R. L. Chem. Mater. 1990, 2, 557–563. (49) Castiglioni, C.; Negri, F.; Rigolio, M.; Zerbi, G. J. Chem. Phys. 2001, 115, 3769–3778.

1179

dx.doi.org/10.1021/ac102939w |Anal. Chem. 2011, 83, 1173–1179