Asbestos real-time monitorin an atmospheric environment

Norihisa Hiromoto, Kosei Hashiguchi, Shigeo Ito, and Toshikazu Itabe

The concentration of asbestos fiber aerosols can be monitored by measuring the polarization of laser lightscattered by asbestos fibers. The principle of discriminating asbestos fibers is based on the theoreticallyexpected difference in polarization at a scattering angle of 170 deg between cylindrical and sphericalairborne particles; polarization at this scattering angle should be positive for cylindrical particles such asasbestos fibers but should be negative or close to zero for spherical mineral particles. We constructed anexperimental asbestos real-time monitor that uses a strong electric field to align the airborne particles,that uses lasers having linear polarization with an equal amplitude in parallel and perpendicularcomponents to the aligned long axis of particles, and that simultaneously detects the two components ofthe linear polarization of light scattered at 170 deg, i.e., close to the backscatter. Experiments that wereperformed to detect the light scattered from airborne standard asbestos fibers showed that the measuredpolarization fits theoretical prediction. The concentrations of airborne asbestos fibers obtained by theasbestos real-time monitor were consistent with those estimated by the standard phase contrast micro-scope method. © 1997 Optical Society of America

Key words: Asbestos fibers, light scattering, polarization, cylindrical particles, aerosols.

1. Introduction

Asbestos is a group of fibrous minerals that can causeasbestosis as well as lung and pleural tumors and istherefore no longer used on exposed surfaces in build-ings as fireproofing and thermal-isolation material orin brake linings and clutch pads as friction material~see Ref. 1 for a review of asbestos and problemsassociated with it!. It is, however, still mixed intocement construction materials, and large quantitiesof asbestos are still used in friction materials such asthe mold-type brake linings of cars. It is thereforeimportant to monitor the concentration of airborneasbestos fibers in the atmosphere so that we canavoid the air pollution they cause. Researchers atthe Environmental Agency measure the concentra-tion of asbestos fibers in the general atmosphericenvironment in Japan and reported that this concen-

N. Hiromoto and T. Itabe are with the Communications Re-search Laboratory, 4-2-1 Nukui-kita, Koganei, Tokyo 184, Japan.K. Hashiguchi is with ESCOM, Incorporated, 5-11-3 Higashi-oi,Shinagawa 140, Japan. S. Ito is with the Faculty of Engineering,Toyo University, 2100 Kujirai, Kawagoe, Saitama 350, Japan.

Received 2 January 1997; revised manuscript received 30 May1997.

0003-6935y97y369475-06$10.00y0© 1997 Optical Society of America

tration has been less than 1 fiberyliter over the pastfew years.2

A standard method for measuring the asbestos con-centration is for one to collect asbestos fibers on amembrane filter by using a vacuum sampler and thencount them under a phase contrast microscope~PCM!. For real-time measurement, a fiber aerosolmonitor ~MIE, Inc., U.S.A.! can be used to detect laserlight scattered around a scattering angle of approxi-mately 90 deg from fibers swung by an oscillatingelectric field ~see Ref. 3 for details on the principle ofthe fiber aerosol monitor!. Other attempts to dis-criminate fibers from spherical particles have beenmade by using intensity and polarization4,5 or bywide angle spatial intensity distribution6 in the mea-surements of forward-scattered light. We propose anew real-time method for optical detection of airborneasbestos fibers in a general atmospheric environmentby observing two components of the polarization ofscattered light at a scattering angle of 170 deg, de-scribe an experimental asbestos real-time monitorthat incorporates this method, and compare the re-sults obtained with this monitor and the PCMmethod.

2. Theoretical Considerations

Because asbestos fibers are more than a few mi-crometers long and have diameters of the order of 10nm ~20–60 nm for chrysotile fibers and slightly

20 December 1997 y Vol. 36, No. 36 y APPLIED OPTICS 9475

larger values for amosite and crocidolite fibers!, wecan reasonably assume them to be nearly cylindrical.The refractive index n of asbestos is 1.55 for chryso-tile fibers and 1.68 and 1.70 for amosite and crocido-lite fibers, respectively.7 When we used a computerprogram8 and calculated scattering matrix elementsfor a normally illuminated infinite cylinder, we foundit noteworthy that the polarization P given by P 5~uT1u2 2 uT2u2!y~uT1u2 1 uT2u2!, where T1 and T2 are,respectively, the diagonal components of the scatter-ing matrix corresponding to components of the fieldparallel and perpendicular to the cylinder axis, ispositive at a fs scattering angle of 170 deg when theradius of the cylinder is smaller than l, the wave-length of incident light.9 The definitions of paralleland perpendicular components of incident and scat-tered fields, the scattering plane, the scattering an-gles fs and us are illustrated in Fig. 1. Thescattering angle fs, in the scattering plane orthogo-nal to the cylinder axis, is defined such that 0 degdenotes forward scatter and 180 deg denotes back-scatter.

Figure 2 shows the dependence of the polarizationof light ~l 5 0.488 mm! scattered by an infinite cyl-inder ~n 5 1.55! as a function of the radius of thecylinder. The wavelength of 0.488 mm was adoptedbecause it is the strongest line of the Ar-ion laser thatwas adopted in the experimental asbestos real-timemonitor. The polarization is positive for the cylinderwith a radius less than approximately this wave-length of light as shown in Fig. 2. The polarizationis also positive for a cylinder with a very small radiuscompared with l, because Rayleigh scattering by theinfinite cylinder produces P 5 ~un 1 1u2 2 4 cos2

fs!y~un 1 1u2 1 4 cos2 fs! 5 0.25 for refractive indexn 5 1.55 at fs 5 170 deg.

Scattering by a cylinder of finite length was alsoexamined using the scattering matrix derived underthe assumption that the induced electric field insidethe particle is equal to that inside the infinite cylin-der.10 The calculated polarization at fs 5 170 degas a function of the cylinder radius is almost the same

Fig. 1. Definition of parallel and perpendicular components ofincident and scattered fields, the scattering plane, and the scat-tering angles of fs and us for scattering by a normally illuminatedcylinder.

9476 APPLIED OPTICS y Vol. 36, No. 36 y 20 December 1997

as in Fig. 2, if us, the scattering angle measured fromthe plane orthogonal to the cylinder axis ~see Fig. 1!,is within 90 6 5 deg. Figure 3 shows a comparisonof differential scattering cross sections at fs 5 170deg between infinite and finite cylinders with a ra-dius of 0.05 mm for l 5 0.488 mm. The scatteringcross section of a finite cylinder integrated for us 50–180 deg ~i.e., for the whole range of us! is exactly thesame as that of an infinite cylinder. Furthermore,the scattering cross section of a finite cylinder longerthan 2 mm integrated for us 5 90 6 5 deg is within afactor of 2 of that of an infinite cylinder. These re-sults mean that those derived in the previous para-graph for an infinite cylinder are also valid for a finite

Fig. 2. Polarization dependence of 0.488-mm scattered light, ascattering angle of fs 5 170 deg, by a normally illuminated infinitecylinder with a refractive index of n 5 1.55 on the radius of thecylinder.

Fig. 3. Differential scattering cross sections at fs 5 170 degversus cylinder length for infinite ~open triangles! and finite ~opencircles and squares! cylinders of n 5 1.55 with a radius of 0.05 mm.The scattering cross sections of a finite cylinder are integrated forus 5 0–180 deg and for us 5 85–95 deg.

cylinder when scattering at us 5 90 6 5 deg is ob-served.

As shown in Fig. 4, however, for spherical particleswith n 5 1.55 the polarization of 0.488-mm light scat-tered at fs 5 170 deg is close to zero or negative whenit is measured in the same way as for the cylinder andthe radius of the particle is less than twice the wave-length of light. This is also true at a Rayleigh limit@i.e., P 5 ~1 2 cos2 fs!y~1 1 cos2 fs! 5 0.015 for fs 5170 deg#.

Because the radii at the maximum number densityin the aerosol number size distribution in the tropo-sphere are below 0.1 mm and most airborne particleshave radii less than 1 mm,11,12 negative polarizationof spherical particles should be common under gen-eral atmospheric conditions. Larger particles canalso be removed at the inlet of the instrument bymethods such as an inertial impactor or a coarsefilter. We therefore inferred that a cylindrical par-ticle such as an asbestos fiber can be discriminatedfrom a near-spherical mineral particle by the polar-ization of light scattered at 170 deg.

Measured intensities of scattered light, Imeas, for acylindrical particle are expressed by

Imeas 5 IiLy~pk!~uT1u2 1 uT2u2!Dfs, (1)

where k ~52pyl! is the wave number in inverse cen-timeters, Ii is the power density of the illuminatinglaser beam in watts per inverse square centimeters, Lis the length of the cylinder in centimeters, Dfs is themeasured angular width in the scattering plane inradians, and T1 and T2 have the same values asmentioned above. Measured intensities of scatteredlight for a spherical particle are expressed by

Imeas 5 Iiy~2k2!~uS1u2 1 uS2u2!DV, (2)

where DV is a solid angle in which the scattering lightis measured, and S1 and S2 are, respectively, thediagonal components of the scattering matrix for di-rections corresponding to T1 and T2. Because the

Fig. 4. Polarization dependence of 0.488-mm light scattered at fs

5 170 deg by a sphere with n 5 1.55 on the radius of the sphere.

light scattered by the infinite cylinder is confinedwithin a plane perpendicular to the cylinder axis, anangular width in the scattering plane Dfs wasadopted in Eq. ~1!. The light scattered by thesphere, however, is distributed over the whole solidangle, and hence solid angle DV, which should beconsidered to be approximately p~Dfsy2!2, is intro-duced in Eq. ~2!.

Table 1 lists the calculated intensities of scatteredlight for a 30-mW Ar-ion laser with a 0.61-mm-diameter beam and for a 10-mW He–Ne laser with a0.68-mm-diameter beam. These intensities are cal-culated, at fs 5 170 deg and with n 5 1.55 and 1.68,for a cylinder and a sphere with almost the samegeometric cross section. Because Dfs is 0.1 rad,solid angle DV is 7.85 3 1023 sr. These values weredetermined from the parameters for the experimen-tal asbestos real-time monitor described in Section 3.Table 1 shows that the intensity of light scatteredfrom a cylindrical particle is much higher than that oflight scattered from a spherical particle. Therefore,we can conclude that scattered light with both posi-tive polarization and considerable intensity can beattributed to an asbestos fiber rather than a near-spherical mineral particle.

3. Experimental Instrument

The experimental asbestos real-time monitor consistsof a fiber-alignment apparatus with high voltage toalign the long axis of airborne particles that flow intoa plastic tube with the aligned electric field across thetube, two illuminating lasers having linear polariza-tion inclined 45 deg from the aligned electric field,scattered light detectors that simultaneously mea-sure two orthogonal polarizations at a scattering an-gle of 170 deg, and laser power monitors to observethe variation of the output of the lasers ~Fig. 5!.

Airborne particles are drawn by a small air pumpinto the plastic tube, a part of which has two semi-cylindrical electrodes of gold-coated aluminum, of afiber-alignment apparatus at a flow rate of 2000 cm3ymin, which is within the recommended condition inthe National Institute of Occupational Safety andHealth Method 7400.13 The particles are aligned by

Table 1. Comparison of the Intensities of Light Scattered at fs 5 170deg by Cylindrical and Spherical Particles

Wavelength~mm!

Refractive Indexof Particles

Intensity of ScatteredLighta ~1029 W!

Cylinderb Spherec

0.488 1.55 1.63 0.0241.68 4.19 0.043

0.633 1.55 0.509 0.1571.68 0.556 0.086

aIlluminating lasers were a 30-mW Ar-ion laser with a 0.61-mm-diameter beam and a 10-mW He–Ne laser with a 0.68-mm-diameter beam. The angle subtended by the scattered lightdetector was 0.1 rad as described in the text.

bRadius of 0.2 mm and length of 5 mm.cRadius of 0.8 mm.

20 December 1997 y Vol. 36, No. 36 y APPLIED OPTICS 9477

the moment force acting on them through the dielec-tric polarization raised by the strong electric field,such that their long axis becomes parallel to the di-rection of the electric field applied between the twoelectrodes. When an asbestos fiber with a radius of0.05 mm and a length of 10 mm is subjected to anelectric field of 61 kVycm, alignment is typicallyachieved in approximately 1023 s. A particle flows adistance of only approximately 0.4 mm within thistime because the flow velocity in a 1-cm-diametertube is 42.4 cmys. The fiber alignment apparatushas two antireflection windows through which laserbeams pass and scattered light can be observed.

The two illuminating lasers, an Ar-ion ~l 5 0.488mm! and a He–Ne ~l 5 0.633 mm!, each have linearpolarization, the direction of which can be adjusted bya half-wave plate to 45 deg from the direction of thealigned electric field such that the laser light has anequal amplitude in parallel and perpendicular com-ponents to the aligned long axis of the particle. Twolasers with different wavelengths were used in theexperimental instrument to prove that the methodwe describe is applicable to a wide range of wave-lengths. The beams of the two lasers are mixed by adichroic beam splitter and are normally incident tothe long axis of the particles. Because the lasershave beam diameters of 0.61 and 0.68 mm, a smallparticle that passes through the beam produces scat-tered light with a pulse width of approximately 1.5ms. The output power of the Ar-ion laser is approx-imately 30 mW and that of the He–Ne laser is ap-proximately 10 mW.

The scattered light detector has channels for twowavelengths, 0.488 and 0.633 mm, and each channelcontains two photomultipliers so that it can be usedto measure the parallel and perpendicular compo-nents of polarization simultaneously. The anglesubtended by the scattered light detector is 5 degaround a scattering angle of 170 deg, which is close tothe backscattering direction.

Fig. 5. Configuration of the experimental ARM.

9478 APPLIED OPTICS y Vol. 36, No. 36 y 20 December 1997

Two silicon photodiodes used as the laser powermonitor measure the light of approximately 0.5% ofthe respective laser output that passes through thedichroic beam splitter.

4. Experiments and Results

The experimental asbestos real-time monitor wasused to detect light scattered from airborne asbestosfibers that were drawn into the instrument from thesampling chamber of the fiber aerosol generator de-veloped by Kitasato Health Science Center and Shi-bata Scientific Technology, Ltd. Standard amositefibers ~JAWE 221, produced in South Africa! wereused in the experiment. Airborne fibers that passthrough the instrument were collected by a 25-mm-diameter white membrane filter placed in front of theair pump. Four channels of signals of scattered lightpulses were sampled by a digital oscilloscope andanalyzed on a personal computer. Examples oftime-dependent signals observed for parallel and per-pendicular components of the scattered light areshown in Fig. 6, where 100 mV corresponds to anintensity of approximately 1 3 10210 W for the Ar-ionlaser and approximately 4 3 10211 W for the He–Nelaser.

Figure 6 also shows that the polarization of lightpulses scattered from asbestos fibers is positive be-cause the intensity of the parallel component is stron-ger than that of the perpendicular component. Italso shows that the pulse width of the scattered lightis a few milliseconds. These observations are con-sistent with the theoretical predictions discussed inSection 2. In the experiment with standard amositefibers, approximately 80% of the observed pulses dis-played degrees of polarization greater than 0.2.

An experiment with near-spherical particles ~air-cleaner test dust! whose radii are smaller than 0.5mm showed that almost all the particles were distin-guished by the small magnitude of scattered light incomparison with the case of asbestos fibers. Thisresult is consistent with the estimation of scatteredlight intensities discussed in Section 2.

Particles collected on the membrane filter duringreal-time monitoring were sampled for evaluation bythe PCM method. The membrane filters were madetransparent with acetone and the number of asbestosfibers in 100 visual 300-mm-diameter fields wascounted under a PCM. This counting was done ac-cording to the asbestos monitoring manual of theEnvironmental Agency of Japan ~1987! as well as theAsbestos International Association, Health andSafety Publication Recommended Technical MethodNo. 1, both of which specify that fibers longer than 5mm are counted.

The concentration of asbestos fibers, CMF in fibersycm3, determined by the PCM method can be calcu-lated as

CMF 5ANMF

anQ, (3)

Fig. 6. Examples of time-dependent signals observed for parallel and perpendicular components of light pulses scattered from standardamosite fibers. The top and second rows indicate the light pulses of parallel and perpendicular components of polarization at 633 nm,respectively, and the third and bottom rows indicate the same, respectively, at 488 nm. Each column represents the detection of anasbestos fiber.

where A is the effective collection area of the filter,NMF is the total number of counted fibers, a is thearea of one visual field, n is the number of measuredvisual fields, and Q ~cm3! is the air volume sampled.The concentration of asbestos fibers measured by theasbestos real-time monitor ~ARM!, CARM in fibersycm3, can be calculated as

CARM 5aNARM

Q, (4)

where a is the calibration coefficient determined bycomparison with CMF, and NARM is the number ofasbestos fibers detected by the ARM, that is, the

Fig. 7. Comparison of asbestos fiber concentrations measured bythe ARM and determined by the PCM method. Filled and opencircles are, respectively, for measurements with the Ar-ion andHe–Ne lasers.

number of signal pulses with a degree of polarizationgreater than 0.2.

The asbestos fiber concentrations measured at thetwo wavelengths by the ARM are compared in Fig. 7with the concentration determined by the PCMmethod. The calibration coefficient a was found tobe 26.10 6 0.36 under a correlation coefficient of0.981 in Fig. 7. This value is reasonable as the ratiofor the section of tube to the area of the field of viewof the scattered light detector falling on the incidentlaser beam. The durations of these measurementswere from 20 to 200 s. The standard deviation er-rors were approximately 16% for the ARM and ap-proximately 20% for the PCM method, which werebasically determined from the inverse square root ofthe total number of counted fibers. Figure 7 clearlyshows good correlation between the concentrationsmeasured by the two methods, which means that theARM is suitable for detecting airborne asbestos fi-bers. The ARM has a minimum detectable concen-tration of 1.3 3 1022 fibersycm3 for a 1-minmeasurement if we substitute a 5 26.1, NARM 5 1fiber, and Q 5 2000 cm23 in Eq. ~4!.

5. Conclusions

We have developed a real-time monitor to detect air-borne asbestos fibers. It consists of a high-voltagefiber-alignment apparatus, two normally illuminat-ing lasers ~an Ar-ion and a He–Ne! with linear polar-ization inclined 45 deg from the long axis of fibers,scattered light detectors that simultaneously mea-sure two orthogonal polarizations at a scattering an-gle of 170 deg, and laser power monitors. Asbestos

20 December 1997 y Vol. 36, No. 36 y APPLIED OPTICS 9479

fibers are distinguished, according to theoretical scat-tering from a normally illuminated cylindrical parti-cle, by a combination of positive polarization andconsiderable intensity of the scattered light. The ob-served pulses of light scattered in an experiment withstandard amosite asbestos fibers were consistentwith theoretical predictions in terms of both polar-ization and pulse width. The concentration of air-borne asbestos fibers measured with the asbestosreal-time monitor had good correlation with thatmeasured by use of the standard phase contrast mi-croscope method, demonstrating that the asbestosreal-time monitor can be efficiently used to detectairborne asbestos fibers.

The authors acknowledge Yoshihito Konishi andhis colleagues at the Kitasato Health Science Centerfor their detailed guidance in asbestos measurementwith the PCM method and also in the generation ofairborne asbestos fibers. This research was sup-ported by the Environmental Agency of Japan as oneof the Research Projects on Environmental Researchand Studies by National Institutions.

References1. B. T. Mossman, J. Bignon, M. Corn, A. Seaton, and J. B. L. Gee,

“Asbestos: scientific developments and implications for pub-lic policy,” Science 247, 294–301 ~1990!.

2. White Paper of the Environmental Agency of Japan ~Environ-mental Agency of Japan, Tokyo, 1985!, pp. 7–8, in Japanese.

3. P. Lilienfeld, “Light scattering from oscillating fibers at normalincidence,” J. Aerosol Sci. 18, 389–400 ~1987!.

9480 APPLIED OPTICS y Vol. 36, No. 36 y 20 December 1997

4. A. R. Jones and H. Savaloni, “A light scattering instrument todiscriminate and size fibers. Part I: A theoretical descrip-tion using the RGD approximation,” Part. Part. Syst. Charact.6, 110–118 ~1989!.

5. A. R. Jones and H. Savaloni, “A light scattering instrument todiscriminate and size fibers. Part II: Experimental system,”Part. Part. Syst. Charact. 6, 144–150 ~1989!.

6. E. Hirst and P. H. Kaye, “Experimental and theoretical lightscattering profiles from spherical and nonspherical particles,”J. Geophys. Res. 101, 19,231–19,235 ~1996!.

7. N. Kamiyama, “Mineralogy of asbestos,” Labor Health 29,32–37 ~1988!, in Japanese.

8. C. F. Bohren and D. R. Huffman, Absorption and Scattering ofLight by Small Particles ~Wiley, New York, 1983!, AppendicesA and C.

9. N. Hiromoto, T. Itabe, K. Araki, and S. Ito, “A new method formonitoring asbestos fiber aerosols by measuring polarizationof scattered light: principle,” in Abstracts of Papers, 17th In-ternational Laser Radar Conference ~Laser Radar Society ofJapan, Sendai, Japan, 1994!, No. 25 D 2, pp. 52–53.

10. M. A. Karam, A. K. Fung, and Y. M. M. Antar, “Electromag-netic wave scattering from some vegetation samples,” IEEETrans. Geosci. Remote Sensing 26, 799–808 ~1988!.

11. F. J. Dentener, G. R. Carmichael, Y. Zhang, J. Lelieveld, andP. J. Crutzen, “Role of mineral aerosol as a reactive surface inthe global troposphere,” J. Geophys. Res. 101, 22,869–22,889~1996!.

12. R. D. Cadle, Particles in the Atmosphere and Space ~Reinhold,New York, 1966!, Chap. 2, pp. 37–38.

13. NIOSH Manual of Analytical Methods, Method 7400 ~NationalInstitute of Occupational Safety and Health, U.S. Dept. ofHealth, Education, and Welfare, Public Health Service, Wash-ington, D.C., 1989!, pp. 7400-1–7400-14.