Anal. Chem. 2008, 80, 1546-1551
Photophoretic Velocimetry for the Characterization of Aerosols Christoph Haisch,* Carsten Kykal, and Reinhard Niessner
Chair for Analytical Chemistry, Institute of Hydrochemistry, Technische Universita¨t Mu¨nchen, Marchioninistrasse 17, D-81377 Munich, Germany
Aerosols are particles in a size range from some nanometers to some micrometers suspended in air or other gases. Their relevance varies as wide as their origin and composition. In the earth’s atmosphere they influence the global radiation balance and human health. Artificially produced aerosols are applied, e.g., for drug administration, as paint and print pigments, or in rubber tire production. In all these fields, an exact characterization of single particles as well as of the particle ensemble is essential. Beyond characterization, continuous separation is often required. State-of-the-art separation techniques are based on electrical, thermal, or flow fields. In this work we present an approach to apply light in the form of photophoretic (PP) forces for characterization and separation of aerosol particles according to their optical properties. Such separation technique would allow, e.g., the separation of organic from inorganic particles of the same aerodynamic size. We present a system which automatically records velocities induced by PP forces and does a statistical evaluation in order to characterize the particle ensemble properties. The experimental system essentially consists of a flow cell with rectangular cross section (1 cm2, length 7 cm), where the aerosol stream is pumped through in the vertical direction at ambient pressure. In the cell, a laser beam is directed orthogonally to the particle flow direction, which results in a lateral displacement of the particles. In an alternative configuration, the beam is directed in the opposite direction to the aerosol flow; hence, the particles are slowed down by the PP force. In any case, the photophoretically induced variations of speed and position are visualized by a second laser illumination and a camera system, feeding a mathematical particle tracking algorithm. The light source inducing the PP force is a diode laser (λ ) 806 nm, P ) 0.5 W).
The term aerosol addresses any particle suspended in gas and moving with a gas flow. This name does not reveal information about the chemical composition nor the physical properties. Aerosol particles can be solid or liquid or even both; they can be * Corresponding author. E-mail:
[email protected].
1546 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
generated by condensation from the gas phase as well be dispersion of coarse particles. Aerosols can be geogenic, e.g., volcanic ash, mineral dust suspended by wind, forest fires, or sea spray, as well as anthropogenic, e.g., incineration processes like diesel engines, dust generated by open-pit mining, or abrasion of rubber tires. Given their properties, significant influence on the earth’s climate is observed either directly by light absorption and scattering or indirectly when acting as condensation nuclei, hence influencing cloud formation. Various health effects are attributed to the influence of aerosol particles due to their physical structures as well as to their chemical composition and surface reactivity. The variety of aerosol manifestation and effects explains the vast variety of aerosol characterization and separation techniques. The particle number concentration can be measured by a condensation nucleus counter.1 In combination with different separation techniques, particle size distributions can be assessed. Typical examples of techniques which separate according to sizerelated physical quantities, are, e.g., filtration, cascade impaction,2 the electrical low-pressure cascade impactor,3 or the electrostatic classifier (differential mobility particle sizer, DMPS).4 Beyond electrostatic forces, thermophoresis is a physical effect which is widely applied for aerosol sampling and separation.5,6 Particles passing a temperature gradient are forced to the low-temperature side and can be deposited by this effect on a cooled surface.7 To our knowledge there is no established method for continuous sorting of particles according to their optical properties. Besides filter-based techniques like the Aethalometer8 or the PSAP,9 photoacoustics is the most common method to measure the optical (1) Holl, W.; Mu ¨ hleisen, R. Naturwissenschaften 1954, 41, 300-301. (2) Keskinen, J.; Marjamaki, M.; Virtanen, A.; Makela, T.; Hillamo, R. J. Aerosol Sci. 1998, 30, 111-116. (3) Marjamaki, M.; Keskinen, J.; Chen, D.-R.; Pui, D. Y. H. J. Aerosol Sci. 1999, 31, 249-261. (4) Seifert, M.; Tiede, R.; Schnaiter, M.; Linke, C.; Mo ¨hler, O.; Schurath, U.; Strom, J. J. Aerosol Sci. 2004, 35, 981-993. (5) Wang, C.-C. Int. Commun. Heat Mass Transfer 2005, 32, 1337-1349. (6) Messerer, A.; Niessner, R.; Po¨schl, U. J. Aerosol Sci. 2003, 34, 10091021. (7) Zheng, F. Adv. Colloid Interface Sci. 2002, 97, 253-276. (8) Fialho, P.; Hansen, A. A. D.; Honrath, E. R. J. Aerosol Sci. 2005, 36, 267282. (9) Sheridan, P.; Arnott, W.; Ogren, J.; Andrews, E.; Atkinson, D.; Covert, D.; Moosmu ¨ ller, H.; Petzold, A.; Schmid, B.; Strawa, A.; Varma, R.; Virkkula, A. Aerosol Sci. Technol. 2005, 39, 1-16. 10.1021/ac7021019 CCC: $40.75
© 2008 American Chemical Society Published on Web 02/07/2008
absorption of aerosol,10,11 while light scattering is routinely determined by a nephelometer.12 Photophoresis (PP), first described by Ehrenhaft,13 denotes the phenomenon that small particles suspended in gases or liquids start migrating when illuminated by an intense beam of light. Two causes of the particle’s migration can be distinguished: Direct PP is a consequence of the momentum transfer of photons to the particle when the light is reflected, diffracted, or absorbed by the particle. The dominating effect for aerosol particles is generally assumed to be the indirect PP. This is also a consequence of light absorption by the particle, leading to a rising temperature on the irradiated side of the particle and the surrounding medium. The statistically higher collision rate of the medium’s molecules with the surface of the particle results in a net force acting on the particle. According to the literature,14 it was Fresnel who first observed these indirect forces while investigating the effect of radiation pressure in 1825 and Ehrenhaft who devised the term photophoresis for this force in 1917. In recent literature, the indirect force is usually addressed as thermophotophoresis, photothermophoresis, or radiometric levitation. PP forces were identified and intensively discussed as the driving mechanism for the transport of atmospheric aerosol in stratospheric or even mesospheric regions.15-19 Very little is published on the technical and especially analytical application of PP forces acting on aerosol particles. To our knowledge the socalled photophoretic spectroscopy,20-23 which characterizes the optical properties of a single aerosol particle based on an equilibrium between gravitational and PP forces induced on different wavelengths, is the only approach heading for this goal. The number of publications dealing with the characterization of particles suspended in liquids is much higher (e.g., refs 14, 24, and 25 and literature therein). A theoretical discussion of such approach is discussed by Greene et al.26 (10) Moosmu ¨ ller, H.; Arnott, P. W.; Rogers, F. C. J. Air Waste Manage. Assoc. 1997, 47, 157-166. (11) Beck, A. H.; Niessner, R.; Haisch, C. Anal. Bioanal. Chem. 2003, 375, 11361143. (12) Pettersson, A.; Lovejoy, R. E.; Brock, A. C.; Brown, S. S.; Ravishankara, R. A. J. Aerosol Sci. 2004, 35, 995-1011. (13) Ehrenhaft, F. Annalen der Physik 1918, 56, 81-132. (14) Zhao, S. B.; Koo, Y.-M.; Chung, S. D. Anal. Chim. Acta 2006, 556, 97103. (15) Rohatschek, H. J. Aerosol Sci. 1996, 27, 467-475. (16) Cheremisin, A. A.; Vassilyev, V. Y.; Horvath, H. J. Aerosol Sci. 2005, 36, 1277-1299. (17) Beresnev, S. A.; Kochneva, L. B.; Suetin, P. E.; Zakharov, V. I.; Gribanov, K. G. Atmos. Oceanic Opt. 2003, 16, 431-438. (18) Beresnev, S. A.; Kochneva, L. B.; Suetin, P. E. Atmos. Oceanic Opt. 2002, 15, 472-479. (19) Pueschel, R. F.; Verma, S.; Rohatschek, H.; Ferry, G. V.; Boiadjieva, N.; Howard, S. D.; Strawa, A. W. J. Geophys. Res., [Atmos.] 2000, 105, 37273736. (20) Arnold, S.; Amani, Y. Opt. Lett. 1980, 5, 242-244. (21) Arnold, S.; Amani, Y.; Orenstein, A. Rev. Sci. Instrum. 1980, 51, 12021204. (22) Lin, B. H. Opt. Lett. 1985, 10, 68-70. (23) Pope, M.; Arnold, S.; Rozenshtein, L. Chem. Phys. Lett. 1979, 62, 589591. (24) Helmbrecht, C.; Niessner, R.; Haisch, C. Anal. Chem. 2007, 79, 70977103. (25) MacDonald, P. M.; Spalding, C. G.; Dholakia, K. Nat. Biotechnol. 2003, 426, 421-424. (26) Greene, M. W.; Spjut, E. R.; Bar-Ziv, E.; Longwell, P. J.; Sarofim, F. A. Langmuir 1985, 1, 361-365.
As a first step toward PP separation of an aerosol according to its optical properties, we present PP velocimetry as a tool for the quantification and statistical evaluation of PP forces acting on aerosol particles. THEORY The direct PP effect is induced by a transfer of photon momentum to a particle by refraction and reflection.27 The resulting net force acting on a particle in a Gaussian laser beam can be split into an axial and a radial component. The gradient forces acts in the radial direction to the center of the laser beam while the scattering forces act in the axial direction. Depending on the optical properties of the particle, the gradient force moves the particles parallel or in the opposite direction of the laser beam. When particles absorb the incident light only on the irradiated side, a temperature gradient within the particles is generated. The surrounding gas layer reaches a temperature equilibrium with the surface of the particle. The molecules with higher kinetic energy in the region of higher gas temperature impinge on the particle with greater momenta than molecules in the cold region, thereby leading to the migration of the particle in the direction opposite to the surface temperature gradient. The dependency of the indirect PP force FP on the physical properties of the particle and the surrounding as can be described in the following way:28
κη2 If FT kp
FP ) -2πr
( )( 1
)
1 kg λ 1 + 3Cm 1 + 2 + 2C λ t r k r p
(1)
The relevant quantities can be grouped in those which depend on the gas medium, i.e., the mean free path of the gas molecules λ, and viscosity η, density F, temperature T, and thermal conductivity kg of the gas, the particle’s properties, i.e., radius r and thermal conductivity kp, and those which are specific to the experimental setup, i.e., the absolute light intensity I. The factor f quantifies the fraction of incoming light which is absorbed by the particle and therefore summarizes the optical properties of the particle. The two dimensionless constants Cm and Ct with typical values near 1.25 and 2.0, respectively, describe the properties of the particle/gas interface while κ is a dimensionless constant specific for the surrounding gas. It is possible to deduce the properties of the aerosol particles if they are suspended in a well-defined gas and the experimental conditions are precisely controlled. However, it is a priori not possible to calculate one of these properties, i.e., r, kp, or f, separately but rather a cumulative value including all three factors. For monodisperse particles, so-called optothermal properties can be deduced from the PP force acting on the particle. For all these calculations, it is essential to know precisely the light power irradiated on the particle which includes the total light power as well as its distribution. Spherical aerosol particles which are irradiated and accelerated by a PP force Fp experience a counteracting force by the friction (27) Ashkin, A. IEEE J. Sel. Top. Quantum Electron. 2000, 6, 841-856. (28) Reed, D. L. J. Aerosol Sci. 1977, 8, 123-131.
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1547
force FF according to Stokes’ law, resulting in an equilibrium velocity. With dependence on the migration direction of the particles, the gravitational force FG has to be taken into account as well. The equilibrium state is described by the following equation:
B FP ) B FF ( B FG
(2)
b B FF ) -6πηrv
(3)
with
being the friction force exerted on a particle with the velocity v and
4 FG ) πr3FPg 3
(4)
the gravitational force (rP is the particle density, g the acceleration of gravity). Substituting eq 1, 3, and 4 into eq 2 reveals the equilibrium velocity vpp:
( )(
1 1 If vPP ) κη 3 FTkP
)
2 1 1 ( r2Fg kg λ 9 η λ 1 + 3Cm 1 + + 2C t r k r 1
p
(5)
With dependence on the orientation of the particle migration with respect to gravitation, the second additive part of the right side of eq 5 has to be added or subtracted or can be omitted completely. This equilibrium velocity is referred to in this paper as photophoretic velocity. For a practical application, it has to be noted that the PP velocity in first approximation is inversely proportional to the thermal conductivity of the particle which may allow for a characterization of the particle regarding this physical quantity. Neglecting the influence of gravity, there is a low dependency on the particle geometrical dimension, while vpp is directly proportional to the applied light intensity as long as there are no physical or chemical modifications of the particle generated by the radiation. These modifications can be evaporation of particle material or thermal decomposition of the particle’s structure, for instance. The influence of the optical properties of the particle is manifested in the formula by the factor f, which essentially represents its optical properties. Again there is linear dependency of the PP velocity on this quantity. As all experiments described below were carried out under ambient pressure, the influence of pressure changes is not discussed here, i.e., λ is constant. For particles suspended in air, it is generally assumed that the indirect thermo-PP effect dominates the direct effect.22,29 On the basis of eq 1 and the theoretical considerations given in ref 24, a comparison of the forces exerted by indirect and direct PP on a particle is possible. According to our experimental conditions, we assumed a laser power of 500 mW at a laser wavelength of 806 nm, focused to a spot of a 200 µm diameter. Acting on a polystyrene latex particle with a radius of rP ) 1 µm and a (29) Keh, J. H.; Chen, S. Y. Colloids Surf., A 2002, 196, 153-162.
1548 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
Figure 1. Experimental setup for photophoretic velocimetry.
refractive index of 1.58, this radiation results in a indirect PP force of FPindirect) 2.4 × 10-10 N while the direct PP force is FPdirect ) 2.6 × 10-11 N. The other optical and thermal properties of the particles are summarized in the results section (subsection Size Dependency of the PP Velocities). Even for latex particles, having a low optical absorption, the indirect PP effect is 1 order of magnitude greater than the direct PP. For strongly absorbing particles, the difference is even more significant and can reach several orders of magnitude. Hence, we neglect the influence of direct PP in the following considerations. EXPERIMENTAL SECTION Flow Cell. The experiments were carried out in an in-house made flow cell (Figure 1) with a rectangular cross section of 10 mm × 10 mm, subsequently referred to the xy-plane. The total length is 90 mm; inlet and outlet are positioned at both ends under an angle of 45° relative to the long (z-) axis of the cell. The aerosol flow was directed along the z-axis of the cell. Fused silica windows are located in the center of the x- and the y-side. To avoid turbulences in the gas flow, the windows are flush with the flow cell walls. For all experiments, the aerosol flow was directed vertically upward, the particles moving against gravitation. With the choice of a maximum flow rate of 0.6 L min-1, a laminar flow in the interaction volume was assured. The corresponding flow velocities ranged from 10 to 100 mm s-1. Optical System. The PP force was generated by means of a fiber-coupled diode laser system (in-house made) with an emission wavelength of 806 nm. The maximum power at the distal end of the fiber was 500 mW, measured by a Field Max II power meter (Coherent). Two geometrical configurations of the laser beam vs the flow direction of the aerosol were tested: The so-called in-flow configuration means a parallel arrangement of the light beam and the aerosol flow with either identical or opposing directions while
we address the orthogonal arrangement of laser beam and flow as cross-flow configuration. Although both variations have distinct advantages, we will concentrate on the latter. In this case, the exit of the fiber was imaged into the middle of the flow cell, resulting in a 280 µm spot (fwhm). By means of a beam profiling system (Laser Cam-HR, Coherent), both the beam diameter and the intensity distribution were measured, revealing a nearly circular profile (circularity > 0.9) with a quasi-Gaussian power distribution and a peak power of 476 W cm-2. In order to get a well-resolved map of the focal spot, it was magnified by a factor of 1:4.3 onto the beam profiler. The applied laser power can be adjusted by a set of neutral density optical filters (Linos GmbH, Germany) down to 10% of the maximum value. Care has been taken to block the laser light behind the cell completely as the high laser power may be harmful to human skin and eyes and cause burning of plastics and paper when focused. To monitor the particle movement by a camera, the particles were illuminated by a cw Nd:YAG-laser (λ ) 532 nm, P ) 15 mW, OEM) perpendicular to the excitation laser. Its profile was modified to a vertical (xz-plane) line with a width of 5 mm and a thickness of approximately 0.5 mm. Evaluation System. The migration of the particles was recorded by means of a CCD camera (Guppy, F 080B, Allied Vision Technology GmbH, Germany) adapted to a macrolens AF Micro-Nikkor 50 mm 1:2,8DG (Sigma GmbH, Germany). The field of view has a size of 5 mm in the y direction and 3.75 mm in the z direction, resolved on 1024 × 768 pixels. From a series of images, the PP velocimetry of the single particles is obtained by an adapted Matlab code (The MathWorks Inc.), based on an IDL tracking tool by Crocker and Weeks.30 The code links the particle positions found in each frame together and forms time scaled trajectories for every particle. With a frame rate of 30 frames per second, the maximum velocity which can be resolved is as high as 10 mm s-1. The actual PP velocity is calculated by subtracting the bulk flow velocity, which is also extracted by the particle tracking tool, from the measured particle velocity. Aerosol Generation. Aerosol was generated by nebulization of white polystyrene latex suspensions (PSL, Bang Laboratories) with particle diameters of 0.18, 0.17, 0.46, 0.99, 1.99, 2.88, and 4.13 µm, and fluorescent particles of the same size (1.0 µm) but different colors (white, yellow, red) (FluoSpheres, Molecular Probes). Although the absorption properties of these particles are not specified, significant differences can be expected at the excitation wavelength of 806 nm. Because of the strong scattering of these particles, optical absorption measurements in solution were not possible. The fluorescence of these particles is in the visible range. Fluorescence excitation at 806 nm is not to be expected, and radiative emission as a competitive process to local warming, which leads to photothermophoresis, can be neglected. A cross-flow nebulizer was applied to extract the particles. After nebulization of the particle suspension, the generated aerosol was sucked through a diffusion dryer in order to obtain a dry aerosol. All volume flows were controlled by means of rotameters and needle valves.
RESULTS AND DISCUSSION Power-Dependency of the PP Velocities. As predicted by eq 5, the PP velocity is directly proportional to the applied laser power. This correlation was verified on the example of the polystyrene latex particles with a diameter of 1.9 µm. As summarized in Figure 2, the expected linear correlation between PP velocity and laser power density was found. All data shown in this paper are calculated from a number n of at least 100 separately tracked particles. With dependence on the particle concentration, this usually takes between 10 and 20 min. These times can be significantly reduced by applying a more powerful laser source, illuminating a larger focal area. The PP velocities given in Figure 2 as well as in all further graphs correspond to the maximum velocity each particle reaches in the center of the laser beam, i.e., the position of maximum power density. To avoid artifacts by single faulty image frames, the velocity profiles for each particle were smoothed by a SavitzkyGolay moving average filter.31 With the equation of motion for the particle under the influence of light and friction force solved, it can be shown that the acceleration of a particle in the light beam is quasi-instantaneous.32 This theoretical prediction, whose mathematical derivation is beyond the scope of this publication, can be verified by comparing the PP velocity of a particle passing through the laser beam with the intensity distribution as shown in Figure 3. Because of the linear dependency of the PP velocity on the laser power density, the velocities of the particles directly follow the power density distribution of the laser. It can be noticed that no significant gradient force, as described in the literature for PP forces acting on colloidal particles suspended in liquids,24 is observed. The reason for this observation is the fact that the laser beam diameter is significantly larger than the particle dimensions. Size Dependency of the PP Velocities. The intention of this project is to define a system for the characterization of unknown particles, first by measuring different PP velocities and statistical evaluation, later in the form of a continuous separation system. The intrinsic particle properties influencing the PP velocity are the particle’s size and its optical and thermal properties. In a next step we investigated the influence of the particle size on the velocity while keeping constant its further properties and the applied laser power density. These experiments were again carried
(30) Crocker; C. J.; Weeks, R. E. http://physics.georgetown.edu/matlab/ code.html, 2007.
(31) Savitzky, A.; Golay, E. M. J. Anal. Chem. 1964, 36, 1627-1639. (32) Arnold, S.; Lewittes, M. J. Appl. Phys. 1982, 53, 5314-5319.
Figure 2. Correlation between the laser power density and the PP velocity for polystyrene latex particles (1.99 µm).
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1549
Table 1. Photophoretic Velocities of Differently Colored and Sized PSL Particles (n > 100, Laser Power Density 476 W cm-2)
Figure 3. Tracked migration path of a polystyrene latex particle (PSL, 1.99 µm) under the influence of the laser beam (upper) and comparison of PP velocity with the laser power density distribution (lower).
Figure 4. Dependency of the PP velocity on the particle’s diameter (PLS) in comparison with the theoretical prediction, fitted to the experimental data with dependence on the optical absorption f (laser power density 389 W cm-1).
out with the white PSL particles, having well-defined diameters and identical optical properties. The correlation between particle diameter and resulting PP velocities is presented in Figure 4 (n > 100). The black curve representing the theoretical data is generated by fitting eq 5 without the gravitation term to the measured results. The only fitting parameter is f, the fraction of incoming light which is absorbed by the particle. A thermal conductivity of kp ) 0.08 W(K m)-1 was assumed for the polystyrene latex particles, and kg ) 0.025 W (K m)-1, λ ) 59.3 nm, T ) 293 K, F ) 1.292 kg m-3, µ ) 18.27 × 10-5 µPa s, and λ ) 59.8 nm were taken as constants describing the gas. All physical properties of gas and particles are cited from ref 33. The resulting absorbed fraction of incoming light was found to be f ) 0.00065. This very low value can be considered realistic since the latex particles are mainly scattering, are nearly transparent, and have only a very low optical absorption. As it can be seen in Figure 4, the theoretical data show a lower increase with particle size than our experimentally found results. This again may be due to the very low optical absorption. It may be suspected that the theoretical considerations summarized above are not perfectly suitable for these specific optical properties, (33) Weast, C. R. CRC Handbook of Chemistry and Physics, 61st ed.; CRC Press, Inc.: Boca Raton, FL, 1980.
1550 Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
size/µm
particle color
photophoretic velocities/mm s-1
1 1 0.99 1.9 2.88
red yellow white white white
0.23 ( 0.031 0.33 ( 0.039 0.22 ( 0.028 0.39 ( 0.038 0.52 ( 0.036
as they are optimized for rather high optical absorption values, where the optical energy is deposed close to the surface.28 A detailed discussion of the discrepancy of the observed results and the theoretical prediction is beyond the scope of this paper and will be considered in a separate communication. Nonetheless it can be stated that this kind of experiment is a feasible approach to measure the optical absorption properties of a large number of individual particles in a very short time. To our knowledge, there is no other technique to fulfill this task. For completely unknown particles the combination with a size-fractioning system like a DMPS allows for the determination of a cumulative factor combining optical absorption and thermal conductivity of the particles. Characterization of Optical Properties. The intention of the presented system is to characterize the optical properties of different particles. For this purpose we employed the colored particles. The PP velocities of these particles were measured in the way described above, applying a laser power density of 476 W cm-2. The results are summarized in Table 1. The size dependency of the PP velocity described above is confirmed. Furthermore a significant difference between differently colored particles was found. Apparently the optical absorption differences are small between the visually different colors at the applied invisible laser wavelength of 806 nm. Lateral Displacement. As stated above, a parallel as well as an orthogonal alignment of the laser beam relative to the aerosol flow can be realized. Although all results shown here were obtained with the orthogonal configuration, similar results were obtained with a parallel setup. Because of the geometry of the flow cell, the last beam-forming optics has to be placed much further apart from the observed interaction volume for the parallel arrangement which leads to significantly lower power densities acting on the particles. Beyond that we see a better applicability for the orthogonal configuration with respect to a continuous separation system. A first step into this direction is made with the following experiment. Aerosol particles of different sizes and colors were moved by the orthogonal excitation laser beam, resulting in a lateral displacement (see, e.g., Figure 3, upper section). The displacement depends on the bulk flow velocity since the interaction time of the particle with the laser beam is lower at higher bulk flow velocities. In consequence the displacement decreases, while for a fixed bulk flow the displacement depends on the achieved PP velocity. These correlations are summarized in Figure 5. For a given species and diameter of particles, the displacement depends linearly on the inverse bulk flow velocity as expected. Particles with larger diameters or higher optical absorptions achieve larger displacements since they are driven
Figure 5. Correlation between bulk flow velocity and lateral PP displacement in the laser beam (n > 100, laser power density 389 W cm-2).
further during the given interaction time with the laser light, which is directly proportional to the bulk flow velocity. For a given bulk flow velocity corresponding to a vertical cross-section in Figure 5, each species and size is displaced by a significantly different distance. In accordance to the measured PP velocities (Table 1), the displacements of red particles are only slightly higher than the one found for white particles of nearly the same size while the displacements of the yellow particles are much higher and are between the white ones of the same size and the one found for larger particles. By placement of a suitable mechanical separation system in the flow cell, or performing a spatially resolved counting downstream of the laser beam, a continuous separation or, in the latter case, statistical evaluation can be realized. From the results given in Figure 4 and the error bars therein, a size resolution of roughly 100 nm for particles in the 1 µm range can be expected with our current setup. Generally a higher resolution can be expected for smaller particles. Our results indicate that by a modified flow system and with the application of a more powerful laser source, this value can be improved by at least 1 order of magnitude. Our results from experiments in the liquid phase underline these expectations.24 SUMMARY AND CONCLUSION We present and characterize an experimental setup for the measurement of photophoretic velocities. It allows a continuous
monitoring of PP velocities of a large number of particles in a short time, where each particle is separately measured. With the measured velocities, deductions on the optical and thermal conductivity properties of the particles are possible. In combination with other techniques, especially size separation, the optothermal absorption of the particles can be calculated. A next step in this direction is the integration of a parallel measurement of thermophoretic velocities within the same instrument in order to distinguish optical and thermal properties. To our knowledge there is no other experimental method available which reveals optical properties of large numbers of particles rather than ensemble properties as, e.g., photoacoustic spectroscopy does. Beyond characterization and statistical evaluation, we present an approach to physically separate particles of different optothermal properties, i.e., different susceptibilities to photophoresis. Further work will be dedicated to the application of different wavelengths and the systematic assessment of different particle systems and properties (e.g., internally mixed particles and surface-coated particles) concerning PP susceptibility. Possible applications are on the one hand the atmospheric aerosol, its migration due to photophoresis, and its influence on the global radiation balance, on the other hand the characterization of industrial aerosol, e.g., for dyestuff pigments, paint, or in cosmetics. A large field of application may be the synthesis of particulate chemical catalysts. An optical characterization system based on PP may allow one to directly distinguish externally mixed aerosol from internally mixed particle systems.
ACKNOWLEDGMENT This work is financially supported by the DFG Grant NI 261/ 16-1. The authors would also like to acknowledge S. Wiesemann, head of the mechanical workshop, for his support during development and construction of the experimental setup.
Received for review December 20, 2007.
October
11,
2007.
Accepted
AC7021019
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
1551
