Kim Y.J. (Ed.) Advanced Environmental Monitoring

Подождите немного. Документ загружается.

172 W.W. Szymanski et al.

Probably the most interesting feature of this new instrumental approach is the

detail that for spherical particles with different refractive indices, the DWOPS does

not require any recalibration vs. some particle standards, as might be needed with

currently existing single optical particle spectrometers. Common optical spectrom-

eters are typically calibrated by means of polystyrene latex (PSL) particles with a

refractive index m = 1.59−0i. Accurate size measurements of particles with refrac-

tive indices different from the PSL refractive index would require a new calibration

of the instrument with apt particles (Liu et al. 1983; Hering and McMurry 1991).

In the case of DWOPS, each single particle measurement results in information

about the particle size and its complex refractive index.

13.4 Modeling of the Performance of the New Spectrometer

The strength of this method is its principal simplicity, as it is shown in Fig. 13.3.

The measuring system of the DWOPS contains two colinear, monochromatic,

unpolarized, focused laser light sources with different wavelengths (λ1=685 nm and

λ2=532 nm) illuminating aerosol particles from opposite sides which are passing

through the sensing volume. Light scattered from each single particle is then col-

lected over angular ranges (10°–30° and 150°–170°) forward and backward with

respect to the incoming beam of each light source. This yields a quadruple set of

independent scattered light pulses from each detected single particle, which is in

contrast to commonly operating optical particle spectrometers which provide a sin-

gle pulse from each measured particle.

The information contained in these four scattered light pulses allows us, using

our evaluation method described below, to determine for each measured particle

the size and complex refractive index. For modeling of the DWOPS performance,

a Monte–Carlo method was used. Each randomly generated particle is described by

its size and complex refractive index m = Re(m) + Im(m). Data related to such a

particle is converted into four digitized responses (RF, RB, GF, GB). This set of

four numbers can be superimposed with noise simulating an actual measuring

situation. This quadruple of values is then compared with a precomputed table

calculated using the Mie theory of light scattering (Brown 1996; Bohren and

Hufman 1983). The table consists of four columns, in which each row was

calculated for a particle with given size and a real and an imaginary part of the

refractive index. The values in the precalculated table correspond to DWOPS

responses based on the scattering geometry of each sensor. Fig. 13.4 shows the

main steps of the simulation procedure.

Tables with various sizes were examined in order to optimize the evaluation

performance from the point of view of system accuracy and computing time.

Obviously, with the increasing number of rows (i.e. decreasing step size) in the

table the numerical resolution of size and refractive index would increase. However,

due to the high complexity of the morphology of light scattering curves vs. particle

size and vs. complex refractive index, the numerical increase of resolution might

13 An Innovative Approach to Optical Measurement 173

yield single-valued results for a given quadruple of measured data (Nagy et al.

2005). In this contribution, the numerically investigated range of parameters covers

the size of particles from 0.1 µm to 10 µm. The real part of examined refractive

indices is between 1.1 and 2.0 and the imaginary part between 0 and 1, which

corresponds to properties of most environmental aerosols. Randomly generated

particles with a certain, known size and refractive index are used to obtain light

scattering response values based on the DWOPS geometry. These values are then

employed to recover the original particle data based on best fitting them to

precalculated ideal responses.

The Monte–Carlo procedure was used to build up frequency diagrams of

relative errors between the “measured” and ideal value for each particle. In the

above specified size and complex refractive index space random numbers with

uniform distribution were generated, yielding a particle ensemble with certain

sizes and refractive indices. In this way, particles were numerically obtained and

used for the exploration of capabilities of the DWOPS. The dimension of the

evaluation table has a considerable impact on the accuracy and the speed of

the data recovery. An evaluation table with 50 equidistant logarithmic steps in

the above given size range from 0.1 to 10 µm was used in this study. The number

of spacing intervals in the real part of the refractive index range was 19 with

linear spacing (0.05 step size), and the number of sections in the imaginary part

of the refractive index range was 21 with linear spacing (0.05 step size). Results

of modelling the DWOPS performance using the Monte–Carlo procedure are

shown in Fig. 13.5.

Depending on the investigated particle size range, an optical system of a

spectrometer may sometimes provide multi-valued responses. For that reason the

Fig. 13.4 Schematic presentation of the simulation procedure. Scattered light pulses are defined

here as follows: RF = red laser, forward scattering; RB = red laser, backward scattering; GF =

green laser, forward scattering; GB = green laser, backward scattering. The “*” sign indicates the

retrieved values, usually different from the input values

174 W.W. Szymanski et al.

investigated particle size range interval was divided into subintervals covering

following size ranges: 0.1–0.4 µm, 0.4–0.8 µm, 0.8–1.2 µm, 1.2–3 µm, 3–5 µm, 5–7

µm and 7–10 µm. It is evident that the instrument’s performance is similar in all

sizing intervals. Nearly bell-shaped distribution of relative error vs. counts proves

the sizing feasibility of the optical system. Certain secondary structures depending

on the size range can be observed; however, a pronounced maximum is unmistaka-

ble, yielding for the examined particles a unique sizing result with the relative siz-

ing error of about ±5% (FWHM). With the increasing signal noise levels the width

of these distributions increase, however, as long as the noise levels do not exceed

about 15%, a unique signal evaluation can be obtained.

Similarly to the modeling of sizing errors we have investigated relative errors for the

determination of particle refractive indices. The relative error for the assessment of the

real part of the refractive index is within about ±10% of the actual value (FWHM). The

accuracy of the determination of the imaginary part of the refractive index is somewhat

lower. Here the relative error spreads within about ±15% (FWHM). These findings lead

to the conclusion that the evaluation table with the above specified dimensions is suita-

ble for accurate measurements within the targeted size range.

13.5 Measuring Performance of the New Spectrometer

The decisive proof of the feasibility of any prototype of a new measuring system is

the experimental verification of the performance. For this study, well-defined

laboratory aerosols with known sizes and complex refractive indices were used and

are summarized in Table 13.1. The experimental arrangement is schematically

shown in Fig. 13.6.

Fig. 13.5 Sizing accuracy for different particle size ranges (legend shows size ranges in µm). The

noise level superimposed on the signal amounts to 5%

13 An Innovative Approach to Optical Measurement 175

All particles were generated from a solution or suspension using pneumatic

atomization. PSL particles were obtained from diluted aqueous suspensions of cer-

tified particle size standards (Duke Scientific Corp.). DEHS particles were gener-

ated from an isopropyl alcohol-DEHS 0.5% solution. The absorbing, carbon-like

particles were generated from a black ink 50% diluted in ultrapure distilled water.

The atomized (collison atomizer) and dried (diffusion dryer) particles were

brought to charge equilibrium in a bipolar ion atmosphere obtained with a weak

radioactive source (

210

Po charger). Following the charging process, particles enter

the differential mobility analyzer (DMA), where they move on trajectories deter-

mined by the clean airflow and the electric field resulting from the applied voltage

(Liu and Pui 1975). The relationship between the voltage V and monodisperse

particle size (D

DMA

) extracted from the DMA is given by: D

DMA

= (constant)

neC(D

DMA

)V.

C(D

DMA

) is the particle slip correction, e is the elementary charge, n is the

number of charges carried by a particle. An optimization of operating conditions of

such a system yields test aerosols where the number of particles carrying n >1

elementary charges is negligible.

Table 13.1 Refractive indices of particle materials used in this study

Material Complex refractive index

Polystyrene latex (PSL) 1.59 + 0.0i

Di-ethyl-hexyl-sebaccate (DEHS) 1.45 + 0.0i

Black ink (Staedtler Mars Ink No. 745) 1.70 + 0.32i (Hogan 1985)

Fig. 13.6 Experimental arrangement for experimental performance verification of the DWOPS

176 W.W. Szymanski et al.

Thus the obtained DEHS and black ink particles have exactly known sizes

DMA

) with the width of the distribution of the order of 5% or better.

The atomization process of PSL suspensions results in test aerosols containing not

only the required calibration particles with defined size but also those residual precur-

sor particles. They are much smaller in size, typically below 100 nm, but due to their

high numbers they may cause measurable pulses influencing the determination of the

instrument’s performance. For that reason these PSL particles were also size classi-

fied by means of the DMA. This process removes the residual particles yielding

nearly perfectly monodispersed challenge aerosol of exactly known particle size.

Condensational nuclei counter (CNC, TSI Inc.) and laser aerosol spectrometer

(LASX, Particle Measuring Systems) were used to control the number, concentration

and size of test aerosols generated with the atomizer-dryer-differential mobility ana-

lyzer system and to provide reproducible experimental conditions.

Analogously to the numerical experiments described above, actual particles with

known properties were generated and used to study the capability of the DWOPS

to optically determine the size of the particles in question and their complex refrac-

tive index. Results of these experiments are summarized in Table 13.2 and show

clearly the exceptional sizing capability of the DWOPS along with the possibility

to distinguish between absorbing and non-absorbing particles. The agreement

between the nominal complex refractive indices given in Table 13.1 and refractive

indices obtained from measurements with DWOPS is very credible.

Such measurements are not possible with optical particle spectrometers currently

available. It is also evident that there is no need to recalibrate the instrument when

measuring particles with various refractive indices, as is usual with currently existing

single optical particle spectrometers. Providing that the particles in question are spheri-

cal, the described instrument directly yields information about particle size and its

complex refractive index based on the signal evaluation procedure present. This experi-

mental data, although limited by the current optical arrangement toward smaller parti-

cles by the laser power and toward larger particles by the DMA capability, verify the

postulated performance of the DWOPS system (Szymanski et al. 2002). Measurements

performed earlier with conventional spectrometers using aerosol particles with refrac-

tive indices corresponding to those used in this study showed that erroneous particle

sizing up to 100% due to changes in complex refractive index may occur (Garvey and

Pinnick 1983; Liu et al. 1983; Szymanski and Liu 1986).

Table 13.2 Values for particle size and its complex refractive index m

obtained with the DWOPS

Material D

DMA

[µm] D

DWOPS

[µm] Re (m) Im (m)

DEHS 0.589 0.588 ± 0.01 1.55 ± 0.20 0.04 ± 0.04

DEHS 0.633 0.643 ± 0.04 1.45 ± 0.09 0.02 ± 0.03

DEHS 0.730 0.731 ± 0.09 1.46 ± 0.02 0 ± 0.04

DEHS 0.846 0.834 ± 0.03 1.46 ± 0.02 0 ± 0.04

PSL 1.000 0.949 ± 0.01 1.60 ± 0.02 0 ± 0.04

Black ink 0.750 0.728 ± 0.09 1.62 ± 0.08 0.34 ± 0.1

Black ink 0.845 0.811 ± 0.08 1.61 ± 0.20 0.32 ± 0.2

13 An Innovative Approach to Optical Measurement 177

13.6 Summary

Single optical particle spectrometry, although very suitable for measurement of envi-

ronmental particles in the size range from about 100 nm up to tens of micrometers,

appears to have been stagnating in recent years. Modifications with regard to the

source of illumination, decreasing size of the instrument, or comfort of handling were

commenced; however, a factual innovation in terms of novel design yielding infor-

mation beyond particle size distribution seems to be missing. The DWOPS system

described here is a new approach to intensity-based optical particle spectrometry. The

instrument’s performance was numerically investigated in the size range from 0.1 to

10 µm and for a range of complex refractive indices occurring in environmental aero-

sols. First experimental results, although within a limited size range, verify the poten-

tial of this technique. It shows clear advantages over the traditional single optical

particle measuring method. The problem of the variability of calibration curves with

changing refractive indices and the issue of a non- monotonic response of the instru-

ment with regard to the particle size—the well-known limitation of conventional

systems—does not arise for the DWOPS, because the combined response of all four

sensors and the described signal evaluation method determine directly the size, along

with the complex refractive index of each detected particle. Numerical modeling and

experimental data are very promising; however, further experimental studies with an

optimized optical system, allowing measurements over the entire size range, also with

nonhomogeneous spherical, as well as nonspherical particles, are planned and will be

reported later. Making reasonable assumptions regarding the range of particle sizes

and the range of the real and imaginary parts of refractive indices, it was numerically

explored and experimentally verified that a unique measurement of particle size along

with an unambiguous determination of the complex refractive indices of measured

particles in real time is possible with the DWOPS system. This approach may pave

the way for new applications for single optical particle measurement in atmospheric

science and environmental monitoring.

Acknowledgements The support of this work was provided by Austrian Science Foundation

(FWF), Proj. P15619 to W.W.S., by National Research and Development Program, Proj.

NRDP3:3/005/2001, and by the ÖAD-WTZ, Proj. Nr. A13/2003.

References

Barnard J.C. and Harrison L.C. (1988), Monotonic responses from monochromatic optical particle

counters, Appl. Opt., 27,584–592.

Bohren C.F. and Hufman D.R. (1983), Absorption and Scattering of Light by Small Particles.,

Wiley Interscience Publication, New York.

Chang H., Okuyama K., and Szymanski W.W. (2003), Experimental evaluation of optical proper-

ties of porous silica/carbon composite particles, Aerosol Sci. Technol., 37, 735–751.

Cooke D.D. and Kerker M. (1975), Response calculations for light scattering aerosol particle

counters, Appl. Opt., 14, 734 –745.

178 W.W. Szymanski et al.

Czitrovsky A. and Jani P. (1993), New design for a light scattering airborne particle counter and

its applications, Opt. Eng., 32, 2557–2582.

Dick W.D., Sachweh B.A., and McMurry P.H. (1996), Distinction of coal dust particles from liquid

droplets by variations in azimuthal light scattering, Appl. Occup. Environ. Hyg., 11(7), 637–646.

Dick W.D., McMurry P.H., and Bottiger J.R. (1994), Size and composition dependent response

of the DAWN-A multi-angle single particle detector, Aerosol Sci. Technol., 20, 345–355.

Garvey D.M. and Pinnick R.G. (1983), Response characteristics of the particle measuring systems

active scattering aerosol spectrometer probe, Aerosol Sci. Technol., 2, 477–488.

Hanusch T. and Jaenicke R. (1993), Simulation of optical particle counter FSSP-100. Consequences

for size distribution measurements, J. Aerosol Sci., 23, 112–120.

Hering S. and McMurry P.H. (1991), Response of a PMS LAS-X laser optical particle counter to

monodisperse atmospheric aerosols, Atmos. Environ., 25A, 463–461.

Hinds W.C. and Kraske G. (1986), Performance of PMS model LAS-X optical particle counter,

J. Aerosol Sci., 17, 67–72.

Hogan A., Ahmed N., Black J., and Barnard S. (1985), Some physical properties of black aerosol,

J. Aerosol Sci., 16, 391–396.

Kerker M. (1997), Light scattering instrumentation for aerosol studies and historical overview,

Aerosol Sci. Technol., 27, 522–535.

Liu B.Y.H. and Pui D.Y.H. (1975), On the performance of the electrical aerosol analyzer,

J. Aerosol Sci., 6, 249–264.

Liu B.Y.H., Szymanski W.W., and Pui D.Y.H. (1983), Response of laser optical particle counter

to transparent and light absorbing particles, ASHRAE Trans., 92, 518–527.

Liu Y. and Daum P.H. (2000), The effect of refractive index on size distributions and light

scattering coefficients derived from optical particle counters, J. Aerosol Sci., 31, 945–956.

Liu B.Y.H., Szymanski W.W., and Ahn K.H. (1985), Aerosol size distribution measurement by

laser and white light optical particle counters, J. Environ. Sci., 28, 19–25.

McMurry P.H., Zhang X. and Lee C.-T. (1996), Issues in aerosol measurement for optics

assessment, J. Geophys. Res., 101, 19189–19195.

Nagy A., Szymanski W.W., Golczewski A., Gál P., and Czitrovszky A. (2005), Effects of the

evaluation table dimensions on the DWOPS sizing accuracy, In: Proceedings of the European

Aerosol Conference EAC2005 Ghent, Belgium.

Pinnick R.G., Pendleton J.D., and Videen G. (2000), Characteristics of the particle measuring

systems active scattering aerosol spectrometer probes, Aerosol Sci. Technol., 33, 334–352.

Renliang Xu. (2000), Particle characterization: Light scattering methods, 13, 432 p, Kluwer,

Dordrecht.

Szymanski W.W. and Liu B.Y.H. (1986), On the sizing accuracy of laser optical particle counters,

Part Charact., 3, 1–8.

Szymanski W.W., Nagy A., Czitrovszky A., and Jani P. (2002), A new method for the simultane-

ous measurement of aerosol particle size, complex refractive index and particle density, Meas.

Sci. Tech., 13, 303–307.

Szymanski W.W., Ciach T., Podgorski A., and Gradon L. (2000), Optimized response character-

istics of an optical particle spectrometer for size measurement of aerosols, J. Quant. Spectrosc.

Radiat. Transf., 64, 75–85.

Wyatt P.J. (1998), Submicrometer particle sizing by multi-angle light scattering following

fractionation, J. Colloid Interface Sci., 197, 9–20.

Chapter 14

Remote Sensing of Aerosols by Sunphotometer

and Lidar Techniques

Anna M. Tafuro, F. De Tomasi, and Maria R. Perrone

Abstract Active and passive remote sensing devices such as lidars and

sunphotometers, respectively, are peculiar tools to follow the spatial and temporal

evolution of aerosol loads and get complementary data to properly characterize

aerosol optical and microphysical properties. A XeF-based Raman lidar is routinely

used at the physics department of Lecce’s University (40° 20′ N, 18° 6′ E), to

monitor aerosol vertical distributions and characterize aerosol optical properties

by the vertical profiles of the backscatter and extinction coefficient, lidar

ratio, and depolarization ratio. In addition, a sun/sky radiometer operating within

AERONET is used to supplement lidar measurements and better infer aerosol

types and properties by columnar values of the particle size distribution, the real

and imaginary refractive index, the single scattering albedo and the Angstrom

exponent (Å). The main objective of this paper is to provide some results on the

spatial and temporal evolution of the aerosol properties over south-east Italy, in the

central-east Mediterranean basin, by using lidar and sunphotometer measurements.

Speciﬁ cally, results on the characterization of the aerosol load from July 18 to July

21, 2005 are reported and particular attention is devoted to the Sahara dust outbreak

that has occurred over south-east Italy on July 18 and 19, 2005.

Keywords: Aerosols, lidar, remote sensing, sunphotometer

14.1 Introduction

Several studies have shown that both natural and anthropogenic aerosols have

important effects on the climate of the earth-atmosphere system (Haywood and

Shine 1997). Aerosol particles affect the climate directly by scattering and absorbing

solar radiation and indirectly by modifying cloud microphysical properties. But, as

a consequence of their high spatial and temporal variability, these effects can

CNISM, Dipartimento di Fisica, Università di Lecce, via per Arnesano, Lecce, Italy

179

Y.J. Kim and U. Platt (eds.), Advanced Environmental Monitoring,

179–189.

180 A.M. Tafuro et al.

be strongly regional (Nakajima et al. 2003) and the current uncertainty in the quantita-

tive assessment of the Earth’s radiative balance, is a direct consequence of the

variable nature of aerosols on regional and seasonal scales. A multiple-measurement

approach is currently used to assess aerosol impacts on global climate. Specifically,

long-term continuous observations from satellites, networks of ground-based

instruments and dedicated field experiments in clean and polluted environments are

used to feed global aerosol and climate models (Kaufman et al. 2002). The aerosol

robotic network (AERONET), which is an international network coordinated by the

NASA Goddard Space Flight Center, has been established to assess aerosol optical

properties from ground-based sun/sky radiometer measurements (Holben et al.

1998). The need of assessing a vertical resolved aerosol climatology has led to the

establishment of the European aerosol research lidar network (EARLINET) that at

present consists of a network of 22 lidar stations spread over whole Europe

(Bösenberg et al. 2003).

In this paper, an elastic-Raman lidar operating within EARLINET and a sun/sky

photometer operating within AERONET are used to characterize the aerosol load

from July 18 to July 21, 2005. Particular attention is devoted to the Saharan dust event

occurred in the south-east of Italy on July 18–19. A main paper’s objective is to show

that complementary measurements are required to infer optical and microphysical

properties of aerosols of different type. Measurements have been performed at the

physics department of Lecce’s University that is located on a flat peninsula of

south-east Italy, nearly 50 km away from large industrial areas. The site is well

suited to contribute to the aerosol characterization of the east Mediterranean basin,

which represents a unique area in terms of suspended particulate matter. Aerosols

from different sources converge to the Mediterranean basin: urban/industrial

aerosols and seasonal biomass burning from Europe (Zerefos et al. 2000; Lelieveld

et al. 2002), maritime and long-range transported polluted air masses from the

Atlantic Ocean, mineral dust from North Africa (Gobbi et al. 2000; Tafuro et al.

2006), and sea spray from the Mediterranean sea itself (Perrone et al. 2005). A brief

description of the XeF lidar and of the sunphotometer is given in Section 14.2. The

advection pattern characterization over south-east Italy from July 18 to July 21, 2005

is provided in Section 14.3 by using 7-day analytical backtrajectories and satellite

MODIS images. Lidar and sunphotometer measurements are analyzed in Sections

14.4 and 14.5, respectively. Summary and conclusion are reported in Section 14.6.

14.2 Experimental Facilities

An elastic-Raman lidar employing a XeF excimer laser (Lambda Physik LPX 210 I)

as radiation source has been used in this study for the aerosol vertical monitoring.

The lidar system is located at the physics department of Lecce’s University (40° 20′ N,

18° 6′ E), on a flat area of the Salento peninsula in Italy. The XeF laser emits light

pulses of 30 ns duration at 351 nm, with a maximum energy and repetition rate of

250 mJ and 100 Hz, respectively. Collection of the backscatter radiation is

14 Remote Sensing of Aerosols by Sunphotometer and Lidar Techniques 181

obtained by a Newtonian telescope, whose primary mirror has 30 cm-diameter

and 120 cm-focal length. The lidar system has been designed to measure as a function

of the altitude, the aerosol backscatter coefficient, the extinction coefficient, the

extinction-to-backscatter ratio, known as lidar ratio, the depolarization ratio, and

the water vapor mixing ratio. More details on the experimental apparatus are

reported in De Tomasi and Perrone (2003).

A sun/sky radiometer made by CIMEL (France) and operating within AERONET

has been used to derive columnar values of the aerosol size distribution, the real and

imaginary parts of the complex refractive index, the single scattering albedo, and the

aerosol optical thickness. The CIMEL sunphotometer (CE318-1) is an automatic

sun/sky radiometer, with a 1.2° field of view and two detectors, to measure direct Sun

and sky radiances at eight spectral channels: 340, 380, 440, 500, 670, 870, 940, and

1020 nm. Holben et al. (1998, 2001) have provided a detailed description of the

instrument, the data acquisition system and the data inversion procedure. The instru-

ment is located on the roof of the physics department of Lecce’s University.

14.3 Characterization of Advection Patterns

by Backtrajectories and Satellite Images

The 7-day analytical backtrajectories and MODIS satellite images have been used

to infer the source regions of the air masses advected over south-east Italy from July

18 to July 21, 2005. The 7-days backtrajectories, for seven distinct arrival height

levels (950, 850, 700, 500, 300, 250, and 200 hPa) and for two arrival times (00:00

and 12:00 UTC), are provided by NASA for each AERONET site (http://croc.gsfc.

nasa.gov/aeronet/index.html). Figure 14.1 shows the 7-day analytical backtrajectories

of the air masses reaching the monitoring site on (a) July 18 at 12:00 UTC, (c) July

19 at 00:00 UTC, and (e) July 21 at 00:00 UTC, respectively. Figure 14.1a indicates

that on July 18, at 12:00 UTC, the 950 and 850 hPa air masses have been advected

to Lecce from Europe, while the 700 hPa air mass has been advected from the

Atlantic Ocean. Figure 14.1b shows as function of time, the change of height of

the 950, 850, and 700 hPa air masses for the 12:00 UTC arrival time.

Figure 14.1c, which displays the July 19 backtrajectories, indicates that north-west

Africa was the source region of the 700 hPa air mass advected to Lecce on July 19,

while Fig. 14.1d shows as function of time, the change of height of the 950, 850,

and 700 hPa air masses, for the 00:00 UTC arrival time, and we observe that the air

mass reaching Lecce on July 19 at about 700 hPa, was at ∼900 hPa over north-west

Africa. Hence, this air mass is expected to be affected by Sahara dust particles. Last

conclusion is supported by satellite images.

Figures 14.2a and 14.2b show two MODIS images of the central Mediterranean

provided by the Aqua satellite (http://rapidfire.sci.gsfc.nasa.gov/) for July 18 at

12:00 UTC and for July 19 at 12:40 UTC, respectively. It is worth observing from

Fig. 14.2a that a dust plume, moving from Sahara, was present over the south-west

Mediterranean on July 18, at midday. Then, according to the backtrajectories of