PB83-262915
Rapid Screening Technique for
Detection of Asbestos
Fibers in Water Samples
Ontario Research Foundation, Mississauga
Prepared for
Environmental Research Lab., Athens, GA
Sep 83
U.S. DEPARTMENT OF COMMERCE
National Technical Information Service
«
-------
EPA-600/4-83-041
. ,_September 1983
PB83-262915
' RAPID SCREENING TECHNIQUE ;
FOR*DETECTION OF ASBESTOS FIBERS
IN WATER SAMPLES
by
E.J.. Chatfield and P. Riis
. Electron Optical Laboratory
Department of Applied Physics
Ontario Research Foundation
Sheridan Park Research Community
Mississauga, Ontario, Canada L5K IBS
Contract 68-03-2717
Project Officer
J. MacArthur Long
Analytical Chemistry Branch
Environmental Research Laboratory
Athens, Georgia
30613
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENA, GEORGIAJQ613
" ^ TOlONAL TECHNICAL
_ .. _ INFORMATION SERVICE
f US. DEPARTMENT OF COMMERCE
I SPRINGFItLO. V*. 22161
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-60Q/4-83-Q41
4, TITLE AND SUBTITLE
Rapid Screening Technique for Detection of Asbestos
Fibers in Water Samples
3..RECIPIENT-S ACCESSION-NO.
PB83-262915
BT REPORT DATt
1QR-3 .
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
E.J. Chatfield and P. Riis
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department, of Applied Physics.
Ontario Research Foundation
Sheridan Park Research Community
Mississauga, Ontario, Canada L5K 1B3
10. PROGRAM ELEMENT NO.
CBNC1A
11. CONTRACT/GRANT NO;
68-03-2717
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Research LaboratoryAthens GA
Office of Research and Development
U.S. Environmental Protection Agency
Athens, Georgia 30613
13. TYPE OF REPORT AND PERIOD COVERED
Final, 10/78-9/81
14. SPONSORING AGENCY CODE
EPA/600/01
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A rapid screening method is presented that will allow samples containing less
than a pre-defined concentration of asbestos fibers to be rejected from further"
analysis, allowing more detailed transmission electron microscopy characterization
to be confined to those samples that have high fiber concentrations. Alignment of
asbestos fibers in magnetic fields, combined with measurements of the scattered
light from the. aligned dispersions, was investigated. A fixed-fiber alignment
method and a dynamic method of fiber measurement were studied. The dynamic fiber
method proved to be the more sensitive method. Detection limits of 0.5 million
fibers per liter (MFL) and 5 MFL were achieved for crocidolite and chrysotile,
respectively. These detection limits were achieved directly from the water sample
without any preconcentration steps. The scattered light measurement techniques
were applied to the determination of the fiber concentrations in drinking water
samples from three sources, and the results were .consistent with those obtained
independently by transmission electron microscopy.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
21. NO. OF PAGES
86 ~
22. PRICE
EPA Form 2220-1 (9-73)
-------
NOTICE
THIS DOCUMENT: HAS'. B'.E.EN . R.E PR ODUCE.D
FROM THE BEST, COPY FURNISHED .US BY
THE SPONSORING AGENCY. ALTHOUGH IT
IS RECOGNIZED THAT CERTAIN PORTIONS
ARE ILLEGIBLE, IT IS BEING RELEASED'
IN THE INTEREST OF MAKING A VAI LA B L E
AS MUCH INFORMATION "AS" POSSIBLE:".' "~
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the
United States Environmental Protection Agency under Contract No. 68-03-
2717 to Ontario Research Foundation, lit has been subject to the Agency's
peer and administrative review, and it has been approved for publication
as an EPA .document. Mention of trade inames or commercial products does
not constitute endorsement or recommendation for use.
TT
-------
FOREWORD
Nearly every phase of. environmental protection depends on a capability to
identify and measure specific pollutants in the environment. As part of
this Laboratory's research on the occurrence, movement, transformation,
impact, and control of environmental contaminants, the Analytical Chemistry
Branch develops and assesses new techniques for identifying and measuring
chemical constituents of water and soil.
A 3-year study was conducted to develop improvements in the analytical
method for determination of asbestos fiber concentrations in water samples.
The research produced an improved sample preparation and analysis method-
ology, a rapid screening technique to reduce analysis cost, and a new
reference analytical methodology for asbestos in water. The analytical
method for determining asbestos fibers in water is perceived as representing
the current state-of-the-art.
. William T. Donaldson
Acting Director
Environmental Research Laboratory
Athens, Georgia
iii
-------
PREFACE
Analyses of water samples, for the presence of asbestos fibers are made by
transmission electron microscopy (TEM), using selected area electron diffrac-
tion and energy dispersive X-ray analysis, to identify each individual fiber.
The fiber identification and counting procedure is labor-intensive, and the
resulting high analytical costs have limited the extent to which water
supplies can be monitored routinely for asbestos fibers. A rapid method is
required which will allow samples containing lower than a pre-defined concen-
tration of asbestos fibers to be rejected from further analysis, allowing more
detailed TEM characterization to be confined to those samples which have high
fiber concentrations. , A rapid method is also required for routine monitoring
of fiber concentrations in water sources where asbestos fibers are known to be
present at concentrations of concern. Measurement of the light scattered by
magnetically aligned fibers was investigated as a rapid screening method.
iv
-------
ABSTRACT
When placed in a strong magnetic-field, asbestos fibers in liquid suspension
adopt preferred alignment directions relative to the field direction. Fibers
may align parallel to, normal to, or at a constant angle to the magnetic field,
depending on the mineralogical variety of fiber. Light is scattered more
strongly in directions normal to the length of the fiber, and thus observation
of the scattered light pattern from an aligned fiber dispersion can yield a
measurement of fiber concentration. The application of this method as a means
for rapid measurement of asbestos fiber concentration in water samples has
been investigated. . ,
--'.(' ' -
A fixed fiber alignment method"has been studied "in "which an'aqueous fiber dis-
persion is filtered through a membrane filter located in a strong magnetic
field. This'results in a filter on which the asbestos fibers are permanently
aligned in preferred directions. When :the filter is exposed to solvent vapor,
the structure collapses and the filter becomes transparent. Rotation of the ,
filter in a collimated beam of light yields maxima in the intensity of the
scattered light;, the positions of these maxima are related to the alignment
direction of the fibers.'. ~. : . ;;:
A dynamic fiber method of measurement has also been investigated in which the
behavior of aqueous asbestos fiber dispersions in a rotating magnetic field
is observed. A spectrophotometer cell which contains the fiber dispersion is
placed between the poles of a rotating magnet,, and is illuminated by a colli-
mated beam of light. The fibers rotate in synchronism with the magnetic field,
and maxima in the intensity of the forward scattered light are observed. Since
light is scattered more strongly in directions normal to the lengths of fibers,
a maximum in intensity of the scattered light occurs for every 180" of fiber
rotation. When the scattered light is monitored and the intensity displayed
as a function of magnet position,, the areas under the peaks are proportional
to the fiber concentration. . .
It was found, using the fixed fiber alignment method, that the lowest detection
level was limited by the residual structure in the collapsed membrane filter.
For UICC crocidolite and UICC amosite, the detection level was about 0.1 ng/mm^,
and for chrysotile about 1.ng/mm^. If 25 mm diameter filters were used, these
detection levels correspond to filter loadings of about 20 ng and 200 ng res-
pectively. The dynamic fiber method achieved much lower detection limits of
180 ng/L and 30 ng/L for crocidolite and chrysotile respectively. These.
detection limits apply to. the 5 ml volume of aqueous fiber dispersion in the
spectrophotometer cell, and correspond to detection of 0.9 ng of crocidolite
and 0.15 ng of chrysotile. It.was found that the required detection limits of
0.2 million fibers per liter (MFL) or 1 ng/L can be achieved with the incor-
-------
poration of a selective fiber concentration technique. A limited study was
made of the high gradient magnetic separation technique for amphibole fibers.
A new method was also devised for separation of chrysotile fibers by selective
adhesion to organic materials.
The developed scattered light measurement techniques were applied to determina-
tion of the fiber concentrations in drinking water samples from three sources,
and. the'results were consistent with "those obtained independently by trans-
mission electron microscopy.
Mineral species other than the asbestos varieties were examined with the
dynamic fiber method in order to determine .possible interferences. The results
.indicated that non-fibrous material which rotates with the magnetic field
yields broad scattered light maxima at about 45° and 225° to the magnetic
field direction. Qualitative measurements showed that many other fibrous
mineral species yielded alignment modes similar to those obtained with the
asbestos varieties.
The analytical time required for a single measurement using the current instru-
mentation is -less than 10 minutes; labor requirements for sample preparation
are variable depending on the fiber concentration steps incorporated, but
these need not exceed 1 man-hour per sample.
vv
-------
CONTENTS
Foreword i i i
Preface i v
Abstract v
Fi gures. i x
Tables xii
Acknowledgment xiii
.1. INTRODUCTION - 1
1.1 Applications of a Rapid Screening-Technique 1
1.2. Required Features of a Rapid Screening Technique ., 1
1.3 Possible Technical Approaches 2
1.3.1 Chemical Analysis Methods 2
1.3.2 X-ray Diffraction Methods ' 3
1.3 3 Infrared Spectroscopy 3
1.3.4 Two Phase Liquid Separation 3
1.3.5 Single-Particle Scattered Light Measurement
Technique ' 4
1.3.6 Measurement of Scattered Light from
" Magnetically Aligned. Fibers. . ...4..
1.4 Summary of Available Techniques 4
2. CONCLUSIONS AND RECOMMENDATIONS 6
3. OBSERVATION OF MAGNETICALLY-ALIGNED FIBER SAMPLES 8
3.1 Types of Fiber Alignment 9
3.2 Observation of Alignment Effects by Light
Scattering Techniques 10
4. FIXED FIBER ALIGNMENT TECHNIQUE 13
4.1 Equipment Design 13
4.2 Sample Preparation 15
4.2.1 Filtration of Aqueous Fiber Dispersions 15
4.2.2 Clearing of Membrane Filters 17
4.3. Measurements and Results 21
4.3.1 Measurement of Detection Levels for Asbestos
Fibers Dispersed in Double-Distilled Water 21
4.3.2 Effects of Non-Fibrous Particulate .^ 26
4.4 Magnetic Field Requirements for Fiber Alignment 27
4.5 Signal Enhancement Techniques for.Improvement of
Detection Levels 28
4.5.1 Complete Dissolution of Filter Medium 28
vi i
-------
4.5.2 Electronic Noise Reduction , 28
4.5.3 .Use of Reflective Scattering Techniques 28
4.5.4 Radiofrequency Plasma Etching of Filters 29
'4.5.5 Shadowing of Particulate by Vacuum Deposition
; of Opaque Films- 30
4.6 Evaluation of the Fixed Fiber Alignment Technique 31
5.DYNAMIC -FIBER- TECHNIQUE^-.."; r.:rrrr77Trr.r:7::r:::.... /. /...... '33
5.1 Equipment Design 33
5.2 Sample Preparation .-..- 37
5.3 Measurements and Results 37
5.3.1 Measurement of Blank Samples 38
5.3.2 Measurement of Detection Levels for Asbestos
. V ^Fibers Dispersed in Double-Distilled Water 40
5.3.3 Effects of Non-Fibrous Particulate 43
5.4 Evaluation of the: Dynamic Fiber Technique 44
6. ALIGNMENT MODES OF SELECTED MINERAL SPECIES IN MAGNETIC FIELDS . 45
.7. METHODS'FOR"CONCENTRATION"bF~"FIBERS::/.T.'.'."..""."..'.".'.".'.". 59
7.1 Non-Specific Fiber Concentration 59
7.2 Removal of Organic Particles ... 59
7.3. Specific Concentration of Chrysotile 59
7.4 Specific Concentration of Amphiboles .. 62
' " . ' -''-'' '-":'« '. . '-' ' ' 1 .''-,...
8. EVALUATION OF THE RAPID SCREENING'TECHNIQUE: '
APPLICATION TO MUNICIPAL DRINKING WATER SAMPLES 66
. -REFERENCES ..... .V:.',...:....,....,.';.. :.;.':. 69
viii
-------
FIGURES
Number . - -. Page
1 Alignment modes of asbestos fibers'in magnetic fields 8
2 Phase contrast optical micrographs 9
3 Scattered, light distributions ... 11.
4 Schematic1 of system used for the analysis of
... light scattered-from aligned asbestos fibers . ... 13
.5 System for scattered light measurements of fixed
alignment samples 14
6 Sample turntable and photomultiplier ..... ;... 15
7 Filtration assembly located between poles of an electromagnet . 16
8 Equipment used..to. collapse membrane .filters-.. 17
9 UICC Crocidolite: Scattered light profile obtained from a.
filter loading of 15 ng/mm2 22
10 UICC Amosite: Scattered light profile obtained from a
filter loading of 16 ng/mm2 22
11 Union Carbide Chrysotile: Scattered light profile obtained
from a filter loading,of 12 ng/mnr. . 22
12. UICC Crocidolite: Scattered light profile obtained from a
filter loading of 0.4 ng/mm2 23
13 UICC Amosite: Scattered light profile obtained from a
filter loading of 0.2 ng/mm2 23
14 Union Carbide Chrysotile: Scattered light profile obtained
from a filter loading of 0.6 ng/mm2 23
15 UICC Crocidolite: Area of P-fiber peak as a function of
mass and fiber concentration on filter 24 .
ix
-------
16 UICC Amosite: Area of P-fiber peak as a function
of mass and fiber concentration on filter 24
17 UICC Amosite: Area of N-fiber peak as a function
of mass and fiber concentration on filter 25
18 Union Carbide Chrysotile: Area of peak as a function
of mass and'fiber concentration on'filter ...... .v. 25
19 Scattered light profile showing peaks from amosite in
muni ci pal dri nki ng .water 26
20 Union Carbide Chrysotile: Variation of peak area
with magnetic field strength 27
21 Schematic of equipment used for analysis of light scattered '
from magnetically-aligned, fibers in liquid suspension 34
22. Equipment for measurement of scattered light intensities
from fibers rotating in liquid suspensions 35
23 Rotating magnet and detection optics of dynamic fiber system .. 35
24 Position encoder located on magnet periphery used to
communicate magnet position to the computer 36
25 Sample loading mechanism 37
26 UICC Crocidolite: Dynamic scattered light profile 39
27 UICC Amosite: Dynamic, scattered light profile 39
28 Union Carbide Chrysotile: Dynamic, scattered Tight profile 39
29 UICC Crocidolite: Variation of scattered light profile
with fiber concentration ...... 41
30 Union Carbide Chrysotile: Variation of scattered light
profile with fiber concentration 41
31 UICC Crocidolite: Area of P-fiber peak as a function
of mass and fiber concentration 42
32 Union Carbide Chrysotile: Area of peak as a function
of mass, and fiber concentration' 42
33 Scattered light profile of borosilicate glass .
particle suspension ... -. 43
34 - Scattered light profiles 46-
81 (selected minerals) 57
x
-------
82 High gradient magnetic separator 63
83 The effect of magnetic field strength on
retention of amosite fibers 64
84 Magnetic separator retention efficiency for UICC amosite 64
85 Scattered light profile of water sample from
Beaver Bay, Minnesota, before and after subtraction of
general particle peak 67
86 Scattered light profile of water sample from
Sherbrooke, Quebec, before and after subtraction
of general particle peak 67
87 Scattered light profile of water sample from
Mississauga, Ontario, before and after subtraction
of general parti cle peak 68
-------
..JABLES
Number ... Page
1 Measurement of Scattered Light from Sample
Preparation Media .......'.......... 20
. 2 Measurement of Scattered Light from Samples
Prepared, by the Dichloroethane-AIA Technique 20
3 The Effect of R.F. Plasma Etching on Aligned
- Crocido!ite-Fiber-Samples 30
4 Improvements in Peak/Background Ratio Produced
by Gold Coating 31
5 Summary of Fiber Removal from Water Samples 61
xii
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the invaluable assistance of Mr. F. Bottone,
Mr. J. Hackett and .Mr. T. Scott, who were responsible for much, of the instru-
mental and software development. They also wish to express their apprecia-
tion to Mr. L. Doehler and Mrs. A. Liebert for their patience during many
hours of sample preparation and TEM evaluation. The authors also' wish to
thank Dr. G. Plant of the Geological Survey of Canada, and Drs. F.J. Wicks
and R.I. Gait of the Royal Ontario Museum for helpful discussions and for
supply of the many mineral samples. The authors especially want to thank
Mrs. M. Cochrane,. Mrs. M.J. Dillon and Mrs. S. Newman for their outstanding
effort in preparing this manuscript...
xiii
-------
- -SECTION! - .
-INTRODUCTION-
1.1 APPLICATIONS OF A RAPID SCREENING TECHNIQUE
Current methods for measurement of the concentration of asbestos fibers
in water samples1 require a minimum of approximately five hours of labor
for each sample. This procedure is time-consuming and expensive in both
labor and equipment, requiring about three hours of fiber identification
and counting by a skilled operator using an analytical electron micro-
scope. A rapid analytical method is .required for.the following two
applications:
(a) to select for detailed electron microscope analysis only those
samples which have fiber concentrations exceeding some
specified value, and;
(b) to allow frequent and economically viable Monitoring of water
sources in which the fiber content has already been adequately
characterized.
1.2 REQUIRED FEATURES OF A RAPID SCREENING TECHNIQUE
If a rapid screening method is to be useful, the economic advantages
over conventional methods based on electron microscopy must be worth-
while. In particular, the equipment required should be significantly
less expensive than the analytical electron microscope, and the instru-
ment time required for each analysis should be short. Elimination of
the requirement for highly qualified and skilled personnel to conduct
the analysis would also be very desirable. In assessing the suitability
of the available methods, sample preparation labor and the time required
for the actual measurement must be considered carefully. The features
of a. rapid screening technique should include:
(a) substantially less labor expenditure than that required for
conventional electron microscopy techniques;
(b) as a consequence of (a), the technique should not rely on
manual fiber counting techniques, either by electron microscopy
or light microscopy;
(c) routine analysis should be possible without the requirement for
a high degree of skill;
-------
(d) it should not require expensive equipment;
(e) it must be sufficiently sensitive for detection of 0.2 million
fibers per liter (MFL) or 1 nanogram per liter (ng/L);
(f) it should be able to distinguish between chrysotile and amphibole
asbestos-fibers.- ......;.-...., ._.. _ __
1.3 POSSIBLE TECHNICAL APPROACHES '
If electron microscopy is excluded,, the possibilities for a rapid
screening technique fall into three categories: those which consist of
a simple chemical measurement, and other more sophisticated approaches
which selectively detect either the mineral species or the actual
mineral ^faeA4 present. Certainly, a rapid screening technique cannot
. incorporate manual fiber counting.;in any form.
1.3.1 Chemical Analysis Methods "' ": ' '.
Chemical methods are based on the detection of one or more of the
constituent elements of asbestos. The element or elements are
usually selected for analysis on the basis of instrumental sensi-
tivity and the ease of analytical technique.. Unfortunately, the
constituent-elements of the asbestos minerals are not specific.
All of the elements concerned are present in many other minerals
and compounds generally present in air, 'water, and soil at con- ;
centrations which .are orders of magnitude higher than the level
' -.of asbestos.;. Methods.-based-on..analysis-of~speci.fic-elements can
therefore be used only when itv is-known-that, the samples contain
.. considerable amounts of asbestos relative to other species which
might interfere. This may .be-the case for some samples
associated with asbestos-processing industries, but in the
general case little is known about the particular sample.
Atomic absorption, flame spectroscopy, emission spectroscopy and
X-ray fluorescence have been Considered for analysis of environ-
mental samples for asbestos.* For measurement of magnesium, the
lowest detection level of 100 ng is achieved by atomic absorption.
This would still require concentration of asbestos from about
400 liters of water in order to achieve a detection limit of
1 ng/liter for chrysotile in water.
Moreover, in a recent study of the chrysotile concentrations in
, some wine and water samples,3 it was found that the insoluble
.magnesium concentrations were of the order of micrograms/liter,
compared with the nanogram/liter levels of asbestos reported in
similar samples by electron microscope fiber counting. This
indicates that the chrysotile fibers may contain only 0.1% of
the total insoluble magnesium in a typical water sample, an
observation which invalidates the use of simple chemical methods
for this type of sample.
: " '' " -1 ' . .. .:;'. 2 "'
-------
In summary, the only justifiable application of the simple
chemical measurement is the case where the particular variety of
asbestos is known to be the only species- present which contains
the element being analyzed.. Furthermore, compositional variations
within the same species of asbestos demand that a calibration be
performed using asbestos from the tame, source.
T.3'.2 X-ray Diffraction Methods """"" :" .'"""
X-ray diffraction (XRD) is capable of specific detection of
individual mineral species. Quantitative determination by X-ray
diffraction requires the use of a. series of known standards; the .
intensities of a particular characteristic reflection are deter-
mined for the unknown and for the standards.. From the ratio of
the.:Intensities, the unknown quantity of asbestos in the samples .'
can be obtained.1* The best minimum detectable weight is reported
to be approximately 1 microgram (yg), and the asbestos must com-
prise, more than about 11 of the total sample.2 For reliable
determinations, 15 to 20% asbestos may be required.5*6 Grinding
... .... Q.p..a sampie during preparation" has been shown "to" affect signifi-
cantly the sensitivity of the measurements by X-ray diffraction.7*8
. A feasibility study has been reported9 in which the fibers were
aligned electrostatically before the X-ray measurements were made.
In this way the detection level of the technique for chrysotile
asbestos was improved to about 0.2 yg, but it was estimated that
this would-be degraded to about 0.4 yg if the asbestos were
present, in a mixture of other.materials.
Thus the detection level of X-ray diffraction methods is inade-
quate, for application as a screening technique for unknown
'samples. . ; ' . .
1.3.3 Infrared Spectroscopy -. ; :
. . .- i '... . . .
The detection level of the infrared spectroscopy technique is
only about 20 yg, and the method is subject to many inter-
ferences.2*10 Accordingly, it is not appropriate as a screening
method for asbestos in unknown environmental water samples.
1.3.4 Two Phase Liquid Separation . .
A technique based on two phase liquid separation (TPLS) has
recently been described11 in which selective extraction of
chrysotile from water into an organic solvent phase is promoted
by an anionic surfactant. Over most of the pH range chrysotile
has a positive zeta potential, in contrast to the negative zeta
potentials of most other waterborne particuTate species. The
surfactant consequently reacts selectively with the chrysotile,
and the fibers become hydrophobic. The aqueous- phase is then
shaken with iso-octane until an emulsion is formed. The emulsion
-------
is salted out with sodium chloride solution, and the chrysotile
is found in the iso-octane.
Several successive extractions are necessary to achieve a recovery
of about 75%. The iso-octane is then filtered, using a Nuclepore*
polycarbonate filter. The filter is carbon coated and fibers on
"~ its surface are counted in reflected light. The sensitivity for
chrysotile was reported as.about 1 ng/L. Unfortunately, the
surface properties of the amphiboles are similar to those of many
other particulate. species;.consequently it is doubtful whether
any useful separation- of amphibole fibers could be achieved using
this technique.
1.3.5 Single-Particle Scattered Light Measurement Technique
A method developed by Diehl et a!12 is based on light scattering
using a. focused laser beam, and observation of the li-ght pulse
from a single particle. Use of several detectors which view the
light pulse from different directions permits some degree of
>.-- -particle shape discrimination. However, the technique requires
a complete initial characterization of the water source by
electron microscopy for calibration of the output. Any fluctua-
tion in the relative proportions of the different types of
particulate may lead to an erroneous result. The possibility of
error, and the requirement for prior electron microscopic charac-
terization of the sample, make this method unsuitable as a rapid
screening technique., :
1.3.6 Measurement of Scattered Light from Magnetically Aligned Fibers
.., Preliminary experiments described by Timbrel!13 have shown that
. . asbestos fibers adopt preferred alignment directions when sus-
pended in a magnetic field. Using relatively simple equipment,
the combination of this effect with scattered light measurements
was:reported to be capable of detecting a few nanograms of fibers.
Moreover, since the technique is based on light scattering, it is
not limited to measurement of only those fibers which can be
resolved in the optical microscope. Fibers which do not align
in the magnetic field, and: randomly shaped particles, do not
contribute to the measurement. Some degree of discrimination
between the fiber types is:possible because their alignment modes
are different.
1.4 SUMMARY OF AVAILABLE TECHNIQUES .
Techniques such as chemical analysis, X-ray diffraction, and infrared
spectroscopy are unsuitable as the basis of a rapid screening technique
either on the grounds of inadequate detection level or lack of speci-
ficity. Although detection levels could possibly be improved by some
pre-concentration step, low specificity would lead to an unacceptable
- number of false positive results.
-------
A screening technique should preferably be capable of detection of
asbestos ^-tfacM, and this is why the previous work in this area has
concentrated either on some unique property of asbestos, as in the
TPLS method, or on specific methods for detecting elongated particles.
Application of the TPLS method appears to be limited to chrysotile, and,
in order to achieve the detection levels required, the sample analysis
is based on manual fiber counting using an optical microscope. The
single-particle scattered light technique would be~capable of significant
further development, but it is not capable of species discrimination and
all elongated particles are counted regardless of composition.
The magnetic alignment approach has many of the features required for a -
rapid screening technique. It appears to be more sensitive than any
other technique considered, apart from the single-particle scattered
light method. It has also been demonstrated that it is sensitive only
to fibers, with a further restriction that they must adopt some pre-
ferred alignment in a magnetic field. The alignment mode is also
species-dependent to some limited extent, leading to the possibility of
discrimination between different asbestos varieties. For these reasons
the magnetic'alignment technique was selected'Tor further development.
-------
.-SECTION 2- .
CONCLUSIONS AND. RECOMMENDATIONS
The measurement of scattered light from magnetically aligned asbestos fibers
:has been demonstrated as a suitable method for detection of asbestos fibers
in water samples. Two techniques were investigated.
The technique based on filtration of fiber dispersions in a magnetic field
was found to have an inadequate detection level, primarily limited by the
residual structure of the membrane filter. The detection limit for fiber
size distributions..similar ..to. those...found, .in water samples was about
102 fibers/mm2 for UICC crocidolite and about 105 fibers/mm2 for Union
Carbide chrysotile. In order to detect a concentration of 0.2 MFL of
chrysotile, filtration of about.100L of water through an active filter area
of 200 mm2 would be required. Signal enhancement techniques such as RF
plasma etching of the filter and shadowing of the particulate by evaporated
.metal films failed to improve the detection limits significantly.
The liquid suspension technique was shown to have detection limits of 0.5 MFL
.for-UICC. crocidolite and.5 MFL.for.Union Carbide-chrysotile. These detection
limits can be achieved directly from the water sample without any pre-
concentration steps. . . . . .
In the investigation of light scattering from liquid suspensions, it was
found that particles of random shape which rotate with the magnetic field
.produce a broad maximum of scattered light intensity corresponding to align-
ment at an angle of 45° to the magnetic field direction. This effect was
observed, for example, with borosilicate glass fragments, and is in contrast
;with the simple increase in constant scattering obtained from particulate
which is unaffected by the magnetic field. In general,, however, the presence
of other particulate degrades the detection limit, and therefore specific
fiber separation techniques were investigated. HGMS (high gradient magnetic
separation) was successful in extracting UICC crocidolite and amosite9 having
a 95% numericar collection efficiency for dispersions of amosite. Because
they are not strongly magnetic, fibers of chrysotile were not retained by
the magnetic separator. Therefore HGMS is a useful technique for separating
chrysotile asbestos from amphibole asbestoses which contain high concentra-
tions of iron.
A new separation technique which is based on scavenging of fibers by organic
materials was successful for specific separation of chrysotile. The same
technique also appears to allow concentration of crocidolite and amosite
fibers, but it is not yet established if the separation is specific. The
-------
recovery of separated chrysotile was between 87% and 100% for standard dis-
persions, falling to about 45% in the case of drinking water samples.
Three municipal water supplies were analyzed directly by the rotating fiber
method. The particulate of random shapes yielded prominent, broad peaks at
45° and 225°, and it was necessary to perform profile subtractions in order
to extract the signal originating from the fibers present. The residual
peaks after this procedure agreed with the known asbestos fiber levels. For
a water sample from Beaver Bay, Minnesota, the 45° component was subtracted
and this resulted in residual peaks at 0° and 180°, and at 90° and 270°.
This agrees with the known asbestos content of the water: cumnringtonite is
known to align parallel and grunerite normal to the magnetic field direction.
It was possible to measure directly the chrysotile fiber concentration in a
municipal water which has a concentration of 40 MFL. Application of the
fiber separation technique to the same sample yielded a concentrated sus-
pension for analysis which contained only chrysotile.
With pre-concentration of the fibers from water samples, the rotating magnet
method is capable of detecting concentrations of 0.2 MFL or 1 ng/L of
asbestos fibers. Development of computer profile subtraction techniques will
permit the separation of the components corresponding to mineral fibers from
the total scattered light profile. This will reduce the amount of sample
preparation required for separation and pre-concentration. Variation of the
rotation rate and strength of the magnetic field may provide additional
information by which particle species may be differentiated. In routine use,
it is estimated that water samples could be analyzed directly in five to ten
minutes, while samples requiring separation or pre-concentration would require
less than one man-hour for preparation^and analysis.
The alignment modes of a number of fibrous mineral species, in a magnetic
field were investigated qualitatively. Some yielded broad scattered light
profiles similar to those from chrysotile, while others displayed sharper
peaks from fibers aligned in directions parallel or normal to the magnetic
field. If the primary purpose is the detection of "asbestos" then there is
some potential for interference by fibrous species other than those normally
considered to be asbestos. Assuming that the purpose of the technique is to
determine if any fibrous mineral species are present, then it is highly
successful, extremely sensitive, and allows for some discrimination between
.mineralogical species. ='...
Assuming some pre-concentration of the; sample, the magnetic alignment tech-
nique has the required detection level and sensitivity for measurement of
fiber concentrations in water. It is capable of significant further develop-
ment, particularly for the determination of fiber dimensions. More extensive
fiber characterization could also be achieved on the basis of iron content
and alignment mode. Further research is also required to optimize the
specific fiber separation techniques.
7
-------
...^SECTION .3. - .,
OBSERVATION OF MAGNETICALLY-ALIGNED FIBER SAMPLES
,3,1 TYPES OF FIBER ALIGNMENT . i
When asbestos fibers are suspended in a liquid and placed in a strong
magnetic field, about 1.0 tesla (T), they become aligned in one of the
three possible modes illustrated: in Figure 1. Depending on the type of
fibers, and in some cases their origin, they may align parallel to the
field direction (P-type), normal to the field (N-type), or transversely
_tp. the field at a constant angle (T-type). Optical microscope slides
(""~cahb"e"prepared..on which asbestos"fibers are permanently fixed in their
preferred alignments relative to a magnetic field. The suspension of
fibers is initially prepared either in an aqueous agar solution or in a
dilute solution of nitrocellulose1in n-pentyl acetate. A drop of the
suspension is placed on a microscope slide, which is then placed between
the poles of a magnet and in a horizontal .position such that the field
direction .is in the plane of the slide. As the suspension medium
evaporates.and solidifies, the fibers-contact the surface of the micro-
scope slide, thus placing an additional constraint on their orientations.
Once the suspension medium has dried, the microscope slide may be viewed
in opticalphase contrast, illumination .to determine the mode of alignment
for the particular species of fiber.
(a)
P-FIBERS
I b)
N-FIBERS
10
T-FIBERS
Figure 1. Alignment modes of asbestos fibers in
: '-.-.. magnetic fields. (After Timbrel!, 1975)
Some examples of samples prepared by this procedure are shown in
Figure 2. Figure 2a shows a phase contrast optical micrograph of an
unaligned dispersion of UICC (Union Internationale Centre le Cancer)
crocidolite; this can be compared with Figure 2b which shows a similar
:! 8.
-------
! lOjum
1 I0>jm
Figure 2.
Phase contrast optical micrographs of:
a) unaligned UICC crocidolite;
b) aligned UICC crocidolite;
c) aligned UICC amosite;
d) aligned UICC Canadian chrysotile;
e)' aligned New Zealand cummingtonite,
(The magnetic field direction is
indicated on ttie figures).
--.. *
-------
sample prepared 1n a magnetic field of 1.0 T. In the case of
crocidolite, the majority of the fibers are aligned parallel to the
magnetic field and a smaller number are aligned in directions perpen-
dicular to the.field. Figure 2c shows a phase contrast optical micro-
graph of an aligned UICC amosite dispersion, and illustrates that for
this mineral there are large numbers of both N-type fibers and P-type
fibers. Figure 2d shows the alignment effect observed with UICC
Canadian chrysotile. Although'the fibers "are" P-type, their curvatures' ~
do not permit the precise alignment which is obtained with crocidolite
and amosite. Figure 2e shows an aligned sample of a variety of
cummingtonite from New Zealand, which contains T-type fibers with their
alignment directions symmetrically disposed at a constant angle about
the field direction........ ',
It is reported that the alignment.of asbestos fibers in a magnetic field
'< occurs because they are either paramagnetic or weakly ferromagnetic, and
that.chrysotile behaves in this way because of the presence of particles
of magnetite in the fibers.13 The P-fiber alignment mode is a conse-
quence of the direction of maximum magnetic susceptibility being parallel
% -to-the length-of the fiber;-for N-fibers it'is normal to the length of
the fibers, and for.T-fibers it is at some angle to the fiber length.
Unpublished work by Cressey and Whittaker11* on amphiboles indicates that
the existence of the alignment effect is probably associated with the
fiber structure itself, rather than being a consequence of any inclusions
of strongly magnetic materials. In a study of,the crystallographic
orientations of aligned amphibole fibers by selected area electron
diffraction,.they found that N-type UICC amosite fibers were aligned with
the crystal- y-axis'orientechwithin ±20° of the field direction. This is
explained if the axis of greatest!magnetic susceptibility is parallel to
the t/-axis, and further supports the idea that the alignment effect is
due to the'crystallographic structure of the fiber. Angular restric-
tions about the z-axis of P-type fibers were also found, but were of a
much broader range.; Nor precise explanation for the different alignment
behaviors of P-type-and N-type fibers has so far been reported. In the
case of chrysotile, although no systematic study has been made, the
alignment effect can still be demonstrated for varieties of chrysotile
which have very low iron concentrations.
3.2. OBSERVATION OF ALIGNMENT EFFECTS.BY LIGHT SCATTERING TECHNIQUES
When a beam of light is used to illuminate a fiber, light is scattered
preferentially in directions perpendicular to the length of the fiber.
For a distribution of randomly-oriented fibers this results in random
light scattering in all directions. To observe the effect of fiber
alignment on the light scattering behavior, a microscope slide prepared .
as described in 3.1 is placed, in the path of a collimated beam of light
projecting on to a white opaque screen. The image-obtained is the
scattered light distribution of the sample. Examples are shown in
Figure 3. These figures were obtained by photographing the distribution
projected on the screen.. .
''" " ' - ;'-10::-":
-------
Figure 3.
Scattered light distributions from:
a) unaligned UlCC crocidolite;
aligned UICC crocidolite;
aligned UICC amosite;
d) aligned UICC Canadian chrysotile;
e) aligned New Zealand cummingtonite.
(The.magnetic field direction is
indicated on the figures).
11
-------
An unaligned sample of UICC crocidolite yields the circularly symmetrical
pattern in Figure 3a. When fibers have been aligned in a magnetic field,
the scattered light distributions from individual fibers are also
aligned, resulting in a maximum in the scattered light intensity in
directions perpendicular to the lengths of the fibers. This is illus-
trated in Figures 3b to 3e for magnetically-aligned distributions of
crocidolite, amosite, chrysotile and New Zealand cummingtonite respec-
tively. Most of the crocidolite fibers are aligned parallel to the
magnetic field, resulting in sharp scattered light maxima perpendicular
to the field direction, but small N-fiber maxima are also visible.
Amosite contains large quantities of both P-type and N-type fibers,
resulting in scattered light maxima both perpendicular and parallel to
the magnetic field. Chrysotile fibers are aligned parallel to the field
direction, but because the fibers*are often curved, there is incomplete
alignment and the maximum in the scattered intensity is much broader.
New Zealand cummingtonite contains fibers which align at a constant
angle to the field (T-type), giving rise to the "X" pattern of Figure 3e.
It should be understood that the distributions of Figure 3 are obtained
by illumination of a large area of the sample, and they are insensitive
to sample-translation. However, if the sample is rotated, the scattered-
light distribution also rotates. This is the basis of the measurement
technique.
To obtain quantitative information from these scattered light patterns,
a detector is used to measure the intensity of the scattered light at a
fixed angle from the optical axis. The difference between the intensi-
ties at a maximum position and a minimum position is a measure of the
amount of aligned fibers in the sample. A convenient method of examining
the alignment behavior is to rotate the scattered light pattern so that
its maxima and minima sweep over the detector area. This can be achieved
by rotating the slide on which the orientations of aligned fibers have
been permanently fixed. Another approach is to rotate a magnetic field
slowly around a cell containing a free suspension of fibers in a liquid.
The first technique is instrumentally simpler, and so this approach was
.investigated initially. The method was found to have some limitations,
which prompted a fuller investigation of the rotating magnet system.
For each system, the initial task:was to determine the detection level
for asbestos fibers. : .
12
-------
.......SECTION 4
FIXED FIBER ALIGNMENT TECHNIQUE
4.1 EQUIPMENT DESIGN :
The principle of the scattered light analysis technique is illustrated
schematically in Figure 4. The sample is a transparent plastic film
containing fibers which have been aligned magnetically. A beam of light
from a laser is used to illuminate the aligned fiber sample, which is
rotated about an axis coincident with the center of the beam. A photo-
multiplier detector is mounted so'that its axis intersects'the incident
"beam at the center of the sample. The angle of intersection, <(>, can be
varied. The detector output is a measure of the scattered light inten-
sity for a particular value of the detector angle , and can be
expressed as a function of 6, the angular rotation of the sample from
' the original magnetic field direction used during preparation.
LASER
CLEARED MEMBRANE FILTER
CONTAINING ALIGNED FIBERS
e
DETECTOR
OPTICAL AXIS
Figure 4. Schematic of system used for the analysis of
1 ight scattered from aligned asbestos f ibers.
13 ;
-------
The actual instrumentation used is shown in Figure 5. The light beam
of 514.5 nm wavelength from an argon ion laser was expanded to a diameter
of approximately 1.5 cm. The diameter of the beam illuminating the sample
could be varied using an iris. This arrangement allowed a large area of
the sample to be illuminated with uniform intensity. The sample was
mounted at the center of a turntable, as shown in Figure 6. The
scattered light was detected by a photomultiplier assembly mounted on an
arm which could be rotated about the center of the sample. The signal
from the photomultiplier was fed to an oscilloscope for initial evalua-
tion of the samples, and to an x-y recorder which gave permanent records
of the scattered light distributions. The sample turntable could be
rotated at different speeds to accommodate this dual display technique.
Figure 5. System for scattered light measurements of fixed alignment
samples. The laser is on the lower level of the optical
bench, and the beam is reflected by mirrors to the upper
level. The beam is expanded before illuminating the sample,
and forward scattered light is detected by the photomultiplier.
14
-------
mm. a i
Figure 6. Sample turntable and photomultiplier.
4.2 SAMPLE PREPARATION
4.2.1 Filtration of Aqueous Fiber Dispersions
The sample preparation technique was designed to produce a
membrane filter on which asbestos fibers were aligned by a
magnetic field applied during a filtration procedure (referred
to as magnetic filtration). In this technique a 25 mm diameter
glass filtration assembly (Millipore Corporation, Cat. No.
XX10 025 00) was located between the poles of an electromagnet,
and a non-magnetic clamp was used to attach the filter reservoir.
Suction for the filtration was provided by a water jet pump
(aspirator). The filtration apparatus is shown in Figure 7.
Fibers become aligned as the liquid passes ..through the magnetic
field, and retain their orientation when collected on the filter
surface. This technique has the advantage of allowing concentra-
tion of the fibers from a known volume of liquid onto the active
area of the filter. The complete procedure is specified by
steps (a) to (e).
15
-------
Figure 7. Filtration assembly located between
the poles of an electromagnet.
16
-------
(a) With a 0.22 ym pore size type GS MiniporeR filter mounted
in the filtration assembly, the aspirator is turned on and
the magnetic field is adjusted to"the desired value (about
1.0 T).
(b) The desired volume of liquid is-filtered through the assembly.
The filtration rate should be restricted to ensure that the
fibers have adequate time to become aligned before contacting
the filter surface. The filtration rate can be adjusted by
changing the applied vacuum. A filtration rate of about
10 mL/minute has been found to be satisfactory.
(c) The magnetic field is then turned off.
(d) A mark is made on the edge of the filter nearest to one pole
of the magnet. This provides a record on the filter of the
field direction during filtration.
(e) The filter is removed and dried for approximately 15 minutes
- '- ' at 70°C.
4.2.2 Clearing of Membrane Filters
The filter on which the aligned fibers have been deposited must
be rendered transparent before it can be examined using the light
scattering equipment. Since it. was discovered that the detection
level for the presence of fibers was dependent on the degree of
filter transparency which could be obtained, several different
preparation techniques were investigated. All of the filter
clearing techniques involve collapse of the membrane filter
sponge structure by exposure to acetone vapor. During this pro-
cedure the filter must not'become distorted and the fibers must
retain their orientations. To prepare a satisfactory sample the
filter must be held in position on a flat substrate.
In the original technique, a clean glass microscope slide was
used as a substrate for the filter. The slide was first dipped
in a 20% solution of collodion in ethanol and allowed to stand
for about 30 seconds, after which period the surface film became
viscous. The filter was then placed onto this surface film,
using a rolling action to inhibit formation of air bubbles between
the glass slide and the filter. The filter was then collapsed by
. exposing its surface to acetone vapor. This was done by inverting
the slide over a beaker containing acetone. When the filter had
.become transparent it was removed from the acetone vapor and
allowed to stand for a few minutes, during which the filter
plastic solidified. _ .
For scattered -light analysis, the filter was carefully removed
from the glass slide after cutting one edge, to release it. The
clear plastic membrane obtained contained aligned fibers, and it
: 17 .
-------
could be supported 1n a holder for scattered light analysis.
Although the sample could be removed from the glass slide
immediately after clearing, it generally peeled from the surface
more easily if it was stored for about 24 hours after it had been
removed from the acetone vapor.
Cleared filter samples: prepared from unused filters were
examined using' the equipment described in 4.1. An index' of the
transparency of such a sample is.given by measurement of the
total forward scattered light intensity at a fixed angle from the
optical axis., The limit of detection for asbestos fibers is in
.fact determined by the variation in the scattered light intensity
from a blank, sample as it is rotated. It is therefore desirable
.that this, intensity and its variation with sample rotation be
minimized, so that the scattered light intensity from filtered
... particulate material, will form as large a proportion of the total
signal as possible-. Accordingly, measurements were made on a
series of blank samples prepared by various techniques, and these
were compared.with values obtained from cleaned microscope slides,
. cover-slips and .thin collodion membranes.
Initial measurements indicated that the collodion used to attach
. the membrane filter1 to the microscope slide was itself a source
of .scattered light, and alternative means of attachment were
investigated. It was found that the buckling and distortion of a
membrane filter during the acetone clearing procedure could be
prevented if. the pores of the filter were filled with a compatible,
but:acetone-miscible, solvent. The solvent used was 1,2-dichloro-
ethane, and"an optimum procedure was developed for clearing of
0.22 ytn pore size: type GS Millipore filters without disturbing
the fiber alignment. In this procedure, a small amount of
dichloroethane was first placed on a 5 cm x 7.5 cm microscope
slide. The filter was placed on the dichloroethane and immediately
transferred to a cleaned microscope slide. The edges of the
filter were then rapidly attached to the slide using an acetone-
based adhesive (nail polish). This step must be performed
quickly,, since the filter lifts from the slide as the dich-
loroethane evaporates. The method was rapid and effective in
yielding filters which lay in contact with the slide surfaces.
The filter was immediately cleared by exposure to acetone vapor.
In a study of acetone vapor clearing methods, it was found that
the technique published by the Asbestos International
Association15 (AIA) yielded clearer filters than those produced
by various modifications of the acetone vapor chamber method of
Ortiz and Isom.16 In the AIA method, the equipment shown in
Figure 8 is used to direct a stream of acetone vapor onto the
filter surface.. Although the AIA clearing process was somewhat
variable and the samples produced were found to yield a wide range
of scattered light intensities, the actual values were lower than
those given by any other technique tested.
.:.. ;.':: is . :
-------
Simple Condensing Column
Acetone
Heating device with
temperature regulator
Figure 8.. Equipment used to collapse membrane
filters. (After AIA, 1979).
:A summary of the measurements made.is given in Table 1. It can
be seen that scattered light from the glass microscope slide did
not contribute significantly to the background intensity when
""compared with the values"obtained for'cTeared membrane filters.
Moreover, the collodion previously used to attach the filter to
the glass slide scattered light strongly, and its use should
therefore be avoided in the sample preparation.
The sample preparation procedure which produced the lowest back-
ground scattered light was to mount the filter by the dichloro-
ethane technique and then to clear the filter using the AIA
acetone vapor method. The. variability of the scattered light
signal obtained from a group of ten such preparations was
V measured; these results are shown in Table 2. The significant
feature of the results is that, with one exception, all of the
filters yielded two peaks in the scattered light profile of
similar shape to those obtained for aligned.fibers of chrysotile
asbestos. When these peaks occurred in the same positions as
thos/> expected for chrysotile, they imposed a lower limit for the
detection level. It was not possible to obtain a flat response
from the majority of preparations.
:"-: 19- :"
-------
TABLE 1. MEASUREMENT OF SCATTERED LIGHT
FROM SAMPLE PREPARATION MEDIA
MEDIUM
Cleaned Glass
Microscope Slide
Cleaned Glass
Cover-slip
50 ran Thick Collodion
Membrane . -
0.22 um pore size type GS
MUfipore Filter
(AIA cleared)
0.22 um pore size type GS
Mlllipore Filter
(Vapor- chamber cleared)
0.45 vim pore size PVC
Copolymer Gelman Filter
(cleared with'a dioxane
cycl ohexanone solution)
BACKGROUND
SIGNAL, mV
90 - 250
300 - 900
400 - 1500
". 3000 -' 6000
9000 - 20000
8000. - 25000
NOISE AMPLITUDE
(peak-to-peak), mV
15 - '100
75 - 150
120 - 1000
150 - 1000
500 - 1000
1000 - 2000
TABLE 2. MEASUREMENT OF SCATTERED LIGHT FROM
TECHNIQUE
SAMPLE
NUMBER
1
2
3
4 '
5
6' .
7"
8
9
10
BACKGROUND-
'S IGNAL, mV
1990
1760,'
1940
2170
; : 3960
2560
3280
2160
1720
1820
NOISE AMPLITUDE
(peak-to-peak), mV
190
120
180
-...-.. 30°
1000
'..''. 330
750
190
310
190
NOISE
PATTERN
H-
++
H-
++
+
H-
++.
-H-
H-
H-
An approximate sine curve with one peak per revolution.
Scattered light peaks 180° apart which were very similar to
and indistinguishable from those produced by chrysotile when
the orientation is appropriate.
: 20
-------
4.3 MEASUREMENTS AND RESULTS
The objectives of the measurements are as follows:
(a) to determine the sensitivity of. the system using suspensions of
the three asbestos varieties UICC crocidolite, UICC amosite and .
...purified Union .Carbide. Caljdria.chrysotile, each ..dispersed...! n
double-distilled water;
(b) to determine the effects of additions of non-fibrous particulate;
and
(c) to investigate possible methods of improving the detection level,
if required.
4.3.1 Measurement of Detection Levels for Asbestos
Fibers Dispersed in Double-Pi stilled Water
Samples were prepared by filtration of a range of volumes of a
stock dispersion of UICC crocidolite in double-distilled water.
The procedure was repeated for both UICC amosite and Union
Carbide chrysotile. Scattered light profiles were obtained for
all of the samples. Figures 9, 10 and 11 show examples of the
profiles obtained for filter loadings of about 15 ng/mm2 of the
three materials. It can be seen that for this filter loading
there was a strong signal from the aligned fibers, which was
easily detectable. When the filter loading was further reduced
by a factor of 20 - 100, the profiles shown in Figures 12, 13
and 14 were obtained. The peaks were still easily detectable at
these reduced concentrations, and would still be measurable if
the fiber concentration were reduced by a further factor of 5.
This indicates that in the-absence of major proportions of other
particulate.material, a fiber concentration of about 0.1 ng/mm2
would be detectable, for any one of the three varieties.
' . . -
The;peak areas for all of the scattered light profiles obtained
from each material were measured using a planimeter, and the
areas expressed in arbitrary units were plotted as functions of
mass concentrations and numerical fiber concentrations. The
calibration curves obtained are shown in Figures 15 - 18.
In the results for crocidolite shown in Figure 15, it can be seen
that for mass concentrations below 0.1 ng/mm2 replicate measure-
ments range over a factor of 8. At this concentration, the peaks
from the fibers are sometimes difficult to separate from the noise
and therefore this concentration can be taken as the approximate
minimum detection level. Detailed detection level studies were
not performed for amosite, but the patterns'shown in Figures 16
and 17 are similar to that obtained from crocidolite. The data
for-chrysotile, shown in Figure 18, indicate that at low
chrysotile concentrations, peaks are present which are unrelated
'' 21
-------
360U
Figure 9. UICC Crocidolite: Scattered light profile2
obtained from a filter loading of 15 rig/mm .
1.6 '
J^ C
to en
0) -r-
Q. in
M-
-------
-* c
co 01
If* QJ
o c
O i^
r- a;
+>
0.23
0 -
180°
360L
Figure 12. UICC Crocidolite: Scattered light profile 2
obtained from a filter loading of 0.4 ng/mm
(Mass loading corresponds to 850 fibers/mm2
of median length 0.6 vm).
-------
too
MASS CONCENTRATION , nq / mmz
0-1 1-0 10
10s 10*
FIBER CONCENTRATION, FIBERS/mm2
I05
Figure 15." UICC Crocidolite: Area of P-fiber peak as a function
.. '' of mass and fiber concentration on filter.
10
TOTAL- MASS- CONCENTRATION ,- ng /mm"1
1-0 10
' Z
3
5
s '<>-
111
-------
10
UJ
H
z
K
5 i-o
-------
to the presence of chrysotile. Accordingly, peaks similar to p f>\
those shown in Figure 14 should be interpreted with considerable
caution. : -
For fiber concentrations greater than about 10^ fibers/mm^ or
1.0 ng/mm2 there appears to be a direct relationship with peak
area. . At concentrations below:these values, the peak areas main-
~tained-an- approximate constant value similar to that observed at
1Q5 fibers/mm2, regardless of fiber concentration. . It is thought
that these are artifact peaks originating from residual structure
in the.collapsed filter, and that this residual structure defines
a minimum detectable limit in these measurements. These results
indicate that the method of filtering in a magnetic field,
. . /followed by collapsing of the Mill i pore filter, gives a detection
" limit for. chrysotile of the order of 10^ fibers/mm' or 1.0 ng/mm^.
Ten filter samples prepared with a 105 fibers/mm2 loading of
chrysotile further supported this value for the detection limit; '
..only three yielded easily recognizable fiber peaks.
-4.3.2- -Effects of Non-Fibrous Particulate ' - --- --
To investigate the effects of non-fibrous particulate, various
volumes of municipal drinking water containing 0.3 ng/mL of total
insoluble solids were.mixed with 5 ml aliquots of a 0.2 ng/mL
dispersion of amosite, and the resulting dispersions were prepared
for scattered light measurements. It was found that the effect of
... " . the,'non-fibrous particulate was to increase the value of the back-
: ground intensity of the scattered Tight.- The. sizes of the peaks
^remained approximately constant. Figure 19 shows the scattered
:. light profile for which.the" filter loading was approximately
' 50 ng/mm^ of total insoluble solids and 1 ng/mm^ of amosite. The
. . experiments indicated that approximately 0.1% by weight of amosite
as a proportion of total insoluble solids was still detectable.
i 1 0.06
', ci '.-
t. ' *r-
" ta '. '
O 0)
O <3
Figure 19. Scattered Aight profile showing peaks from
amosite in municipal drinking water. Filter
loading was 1 ng/mm2 amosite and 50 ng/mrn
total insoluble solids.
26 ; :-
-------
~- This conclusion is reasonable in view of the fact that most of the
intensity of the.scattered light originates from the filter itself.
However, the results do indicate that at approximately the
0.2 ng/mm2 level of an amphibole, the other components in this
.municipal water sample did not contribute any measurable inter-
ference to the determination of the fiber concentration.
4;4' MAGNETIC' FIELD-REQUIREMENTS' FOR FIBER-ALIGNMENT': ~
Initial work has shown that crocidolite and amosite align very, precisely
in magnetic fields. It was found that even the residual magnetic field
present when the magnet was not energized was adequate to produce some
alignment effect on these fibers.; In contrast, alignment of.chrysotile
filters requires much stronger fields. Because the higher value of the
detection level which was obtained for chrysotile may have been a conse-
quence of inadequate field strength, .it was decided to investigate the
minimum magnetic field .requirement for efficient alignment of chrysotile
fibers. Since the cost of the magnet escalates very rapidly with increase
of field strength, it was also required to determine whether an increase
of-magnetic field strength above the minimum value required would yield
a useful increase in the scattered light signal.
Aliquots of 10 ml volume from a 250 yg/L chrysotile dispersion were
filtered at differing field strengths from 0.2 T up to 1.0 T in incre-
ments of:0.2 T.- The resulting filter'samples, were prepared and the
scattered light distributions measured. The areas of the scattered light
- peaks were measured and' plotted against the magnetic field strength,
giving the results shown in Figure 20. Maximum alignment was achieved
for field-strengths-greater than 0.4-T. -These-results demonstrate good
repeatability and show that 0,4 T is an adequate field strength for the
magnetic filtration technique. ; . .
w
t 40
cc
S
on
cc
<
. 20
Ul
10
0 0-2 0-4 0-6 0-8 I-O
MAGNETIC: FIELD STRENGTH , m
Figure 20. Union Carbide Chrysotile: Variation of
peak area_wi th magnetic__field strength.
27
-------
4.5 SIGNAL ENHANCEMENT TECHNIQUES FOR IMPROVEMENT OF DETECTION LEVELS
The measurements made using crocidolite and chrysotile fibers have indi-
cated that the minimum concentrations which can be detected on filters
were about 0.1 ng/mm2 and 1.0 ng/mm2 respectively. For detection of
chrysotile at a level of about 1 ng/L in water, concentration of the
chrysotile from 200 liters of water onto a filter of 200 mm2 active area
'would be required". The'presence of large proportions of other particu-
late in this volume of water would probably degrade the detection limit
still further. A significant improvement in the detection level is
therefore required._
Since it appears that the scattered light.from a blank filter cannot be
reduced below some minimum level, an alternative approach would be to
enhance the scattered light signal from the aligned fibers.
4.5.1 Complete Dissolution of Filter Medium
The scattered light from the filter background could be eliminated
if the fiber orientations could be retained while the filter
itself was dissolved or oxidized. A number of different techniques
were investigated, all of which were based on extraction replica-
tion of the filter surface. A thick layer of silicon monoxide was
deposited by vacuum evaporation in order to fix the fiber orien-
tations. The filter was then placed on a glass slide and proce-
dures such as solvent extraction or low temperature ashing were
used to remove the filter medium. All of these techniques failed
- to yield optically flat samples. Residual cracks and wrinkles in
the thin film yielded large "peaks in the scattered light unrelated
to the presence of fibrous material. Accordingly, this approach
: was abandoned. ;.
4.5.2 . Electronic Noise Reduction
. . Electronic noise suppression techniques such as use of a lock-in
- amplifier were considered as a means of extraction of periodic
- information from the light scattering profiles. Unfortunately,
their application to this problem is inappropriate, since the
. background signal variability from the sample is identical for
each rotation. These techniques are only useful for extraction of
periodic signals from a background of random noise, and in this
', case all of the variability in the sample has been permanently
. fixed in position so that it repeats itself periodically once per
revolution.'/. i . .
4.5.3 Use of Reflective Scattering Techniques
One way to avoid the limitation of detection level imposed by the
residual internal structure of the cleared membrane filter would
be-to arrange, the equipment so that reflective scattering from the
filter surface is measured. In the early work on the magnetic
. alignment technique_a Nuclepore capillary pore_filter was used to
. . ::' za
-------
filter the fiber dispersion, after which a coating of gold was
applied to the filter surface either by vacuum evaporation or by
sputtering. The reflected scattered light intensity at an angle
of about 30° from the optical axis was measured. Although strong
signals from aligned fibers were obtained, reflective scattering
was found to be very sensitive to imperfections on the filter sur-
face. In particular, scratches on the filter surface were found
to yield "very large'spurious peaks" in"the scattered light profiles.
The surface imperfection problem was reduced by using a Millipore
filter and collapsing the structure in the same way as for the
forward scattering technique. However, spurious peaks from many
of the.samples were still observed and at that time the technique
was abandoned. At the improved detection levels already achieved
by the forward scattering technique, the spurious peaks associated
with reflective scattering would present an even more serious
obstacle to reliable measurements.
4-.5.4 Radiofrequency Plasma Etching of Filters
During collapse of the membrane filters by exposure to acetone
vapor, many of the fibers are known to become totally embedded in
the filter plastic. This reduces their contribution to the
scattered light signal. Fibers embedded in the filter can be
brought to the surface by etching the sample in an R.F. plasma
asher. . In order to evaluate the potential of this procedure for
signal enhancement, a set of filter samples was prepared which had
a 3500 fibers/mm2 loading of aligned crocidolite fibers. The
scattered light profile of each sample was measured. The samples
were then exposed to different etching times ranging from one to
eight minutes at an R.F.. power level of 50 watts, after which the
scattered light profile'of each was recorded again. The results
are summarized in Table 3. For a two-minute etching time the peak
height was found to increase by a factor of 2.2, which was the
maximum improvement obtained. Further increase of etching time up
.to about 6 minutes resulted in a gradual reduction of peak height,
after which point there was a. sudden drop to about 0.14 of the
initial aligned fiber peak. The background scattered light inten-
sity was also found to increase with etching time. These results
suggest that:
(a) the optimum etching time was about 2 minutes;
(b) . the increase in background intensity was due to
development of structure on the filter surface;
(c) the.sudden loss of peak height may be due to detach-
ment, loss or disturbance of fibers when the sample
was excessively etched.
A similar experiment was performed with filters which had loadings
.of_105_fibers/mm2_of,aligned chrysotile-.fibers. .For an etching
29 .:.-
-------
TABLE 3. THE EFFECT OF R.F. PLASMA ETCHING ON ALIGNED CROCIDOLITE FIBER SAMPLES
SAMPLE
NUMBER
1
2
3
4
5
6
7
8
9
BEFORE ETCHING
Background
Signal, Volts
3.9
4.4
4.4
5.1
5.1
4.4
4.Y
5.0-
4.6
P- Fiber Peak
Height, Volts
. 1.33
2.89
3.68 .
.5.36. -
5-37
.- 4:76
5.14-
4.53.
4.39
ETCHING
TIME
1 min.
2 m1n.
. 3 m1n.
4 m1n.
5 min.
6 m1n.
7 m1n.
8 min.
-
AFTER ETCHING
Background
Signal, Volts
4.3
8.8
12.5
16.1.
. 22.4
31.3
32.3
32.4
-
P-Fiber Peak
Height, Volts
2.05
6.40
6.80
8.90
7.25
0.66
0.22
0.15
-
PEAK HEIGHTS
RATIO
After/Before
1.54
2.21
1.85
1.66
' 1.35
0.14
0.04
0.03
-
time of \h minutes the peak height was found to increase by a
factor of 3.5, while the background intensity increased by a
. ... factor of 2... For longer etching times the peaks at 0° and 180°
became unequal in height. ...
'It is clear .that although some improvement in the scattered light
signal can be Obtained using the. plasma etching technique, it does
not offer significant potential for the major improvement of
detection level which is required.
4.5.5 .Shadowing of Parti oil ate by Vacuum Deposition of Opaque Films .
Angular asymmetry of the scattered light intensity from the cleared
membrane filter is responsible for evaluation of the minimum detec-
tion, limit. The effect of this can be eliminated by application of
an opaque layer on the-filter except in those areas occupied by
particles. Such a layer would eliminate the contribution of the
substrate to the scattered light signal. An opaque layer of either
gold or aluminum can be applied by vacuum deposition, either normal
to the filter surface or at some other angle. If the deposition is
applied at normal incidence to the filter surface, the layer will
be continuous except at the edges of the particles; if it is
applied at some other angle, uncoated areas will occur on the
filter surface where it is shielded from the_ evaporation source
by the particles. It was thought that large" improvements in the
detection limit might be achieved by correlating the direction of
deposition of the opaque film with the alignment direction expected
for the fibers.
:.. 30
-------
Experiments were conducted with opaque evaporated gold coatings._
Some filter samples were coated at normal incidence, and others
were coated at 45° incidence. The gold evaporation source was
arranged so that the P-fibers were oriented with their lengths
perpendicular to the evaporation direction. This geometry caused
the P-fibers to produce uncoated slits on the filter surface, the
dimensions of which were roughly the same as those of the fibers,
while the rest'of"the filter was"rendered opaque. " " ~
The results obtained from experiments conducted using aligned
fiber, samples of crocidolite and chrysotile are shown in Table 4.
Significant improvements in peak to background ratios were
obtained, but it was found that signals from non-fibrous particu-
late were generated which were similar to those from fibrous .
material. It is possible that further investigation of the tech-
nique,, using two evaporation directions and signal processing,
could eliminate, the signals from the non-fibrous components.
However, the technique was abandoned in favor of the liquid sus-
pension technique described in Section 5.
TABLE 4. IMPROVEMENTS IN PEAK/BACKGROUND
RATIO PRODUCED BY GOLD COATING.
SAMPLE TREATMENT
Crocidolite,
Gol d Coated at
Normal Incidence
Crocidolite,
Gold Coated at
45° Incidence
Chrysotile,
Gold Coated at
Normal Incidence
: . PEAK/ BACKGROUND RATIO -
Before Treatment,
0.32' ";'
' i .
0.54 ."-:.
;
0.22
:
.After Treatment
2.14
4.00
1.76
IMPROVEMENT
RATIO
6.7
7.4
8.0
4.6 EVALUATION OF THE FIXED FIBER ALIGNMENT TECHNIQUE .
The best sensitivity achieved using this, technique was about 0.08 ng/mm2
: of crocidolite and 1 ng/mm2 of chrysotile. If a filter of 200 mm2 active
area is used, the minimum filter loadings detectable are 16 ng and 200 ng
";: __ respectively.._..For_chryspti_le_at an Jnitial _.concentration.of_ 1 ng/L,.
>; .. . ;-. 31:.:."
-------
concentration of the fibers from 200 liters of water would be needed to
achieve the minimum detection limit. Even assuming a further order of
magnitude improvement by the gold coating technique to be possible, the
target detection level of 0.2 MFL or 1 ng/L could not be achieved. In a
typical waterborne fiber size distribution, a crocidolite fiber concen-
tration of 0.2 MFL corresponds approximately to 70 ng/L. To obtain a
filter loading of 16 ng would require filtration of about 230 mL of
water'.."which is marginally possible for relatively clean water. 'However,
the target mass concentration of 1 ng/L is still not possible, since it
would require filtration of 16 liters of water to achieve this detection
limit.
-------
SECTION 5
DYNAMIC FIBER TECHNIQUE
In'this experimental arrangement the asbestos fibers are suspended in a liquid
during the measurement.. A magnetic field rotates around the liquid suspen-
sion, and the fibers also rotate, maintaining their orientation with the field
direction. The forward scattered light intensity is measured at a fixed angle
to the optical axis. . In this system the problems associated with variability
of the light scattered by. the rotating filter, in the fixed fiber measurement.
system, are eliminated since only the fibers themselves rotate. Furthermore,'
the scattering is random for individual particles which are not affected by
'the magnetic field but move under Brownian motion only.
5.1 EQUIPMENT DESIGN
! A schematic diagram of the equipment is shown in Figure 21. The fiber
suspension is contained in a cylindrical quartz spectrophotometer cell
of 5 ml volume and 2 cm-path length. The magnet has a field strength of
0.9 T across a gap of 2.5 cm; since no suitable permanent magnets are
commercially available it was necessary to construct this. The magnet
rotation system was also constructed, and operates at speeds of 60 rpm
down to 1.0 rpm. In this system, the signal from the photomultiplier
detector is fed to a small computer and the scattered light profile is
displayed on the computer video screen. This use of a computer allows
much more flexibility when measuring the scattered light and analyzing
; the patterns produced. In particular, scattered light intensity data can
be accumulated over a number of revolutions^ of the magnet, thus permitting
averaging of the effects of random scattering from particulate other than
fibers. This "dynamic fiber" scattered light measurement system is shown
in Figure 22. Figure 23 shows the detail of the rotating magnet and
detection optics. The instantaneous position of the magnet is detected
by two optical encoders shown in Figure 24, one to count revolutions and
another to count the 212 divisions within a single revolution. The sample
loading mechanism consists of the cannon-shaped device shown in Figure 25,
which accommodates the sample cell on the end. The cannon device is
moved horizontally so that the sample is between the poles of the magnet.
The light beam passes down the axis of the hollow cannon tube.
The operating procedure for the equipment, is very simple. The liquid
sample is first loaded in the system and positioned" between the poles of
the magnet. The magnet rotation is initiated, and the computer derives
the angular velocity by timing one revolution. The intensity of the
scattered light is measured at each of the 212 positions
' '.-:. 33'v:-
-------
LENS
COLL I MATED
LIGHT BEAM
r^- ~i
I ROTATING
I MAGNET >»
-CYLINDRICAL
SAMPLE CELL
PHOTOMULTIPLIER
12 BIT D/A.
CONVERTER
12 BIT A/D
CONVERTER
LT
DIFFER-
ENCE
AMPLIFIER
GAINX
(A-B)
Figure 21. Schematic of equipment used for analysis
: of light scattered from magnetically-
. aligned fibers in liquid suspension.
34
-------
Figure 22. Equipment for measurement of scattered light intensities
from fibers rotating in liquid suspensions.
Figure 23. Rotating magnet and detection optics of dynamic fiber system.
The lens is located at an angle of 12° from the optical axis
and detects scattered light within 1CT2 steradians. The lens
is arranged so that the sample is imaged in the plane of the
photomultiplier cathode.
35
-i. *
-------
Figure 24. Position encoder located on magnet periphery used to
communicate magnet position to the computer.
-------
Figure 25. Sample loading mechanism.
within one revolution. The average value is transmitted to a difference
amplifier, the gain of which is then adjusted so that peaks are amplified
to about 1 - 5 volts. The output of the difference amplifier at each of
the 212 positions for a single rotation is then recorded. Data from a
specified number of revolutions of the magnet are accumulated and
averaged. The scattered light profile can then be displayed on a video
terminal or plotted.
5.2 SAMPLE PREPARATION
The fiber dispersion in water is placed in the spectrophotometer cell and
measured directly. Where fiber concentrations are low, either non-
selective or selective pre-concentration may be necessary. These tech-
niques are considered in Section 6.
5.3 MEASUREMENTS AND RESULTS
It was found that there are three types of particulate which are relevant
to the measurement technique:
37
-------
(a) fibers which align with and rotate with the magnetic field;
(b) particles of equant or random shapes which rotate with the magnetic
field; and .
(c) fibers or particles of equant or random shapes which do not respond
to.the magnetic field.
The fibers in category (a) produce scattered light profiles which have
peaks. At. high magnet rotation speeds fibers may lag behind the orien-
tation they would assume in a stationary field. Category (b) particles
always produce broad scattered light peaks corresponding to alignment at
45°. Surprisingly, glass particles fall into this category. Category
(c) particles simply contribute .to a uniform background scattered light.
signal. Some particles of (b) and (c) are usually present in all samples,
and certainly they will be present .in environmental water samples.
The scattered light profiles obtained for three varieties of asbestos
were examined to confirm that the dynamic fiber technique would produce
data comparable with that from the filtration method. Figures 26, 27 and
28 show the profiles for.UICC crocido!ite, UICC amosite and Union Carbide
chrysotile.respectively. Although the curves for the two amphibole
varieties were similar to those obtained from the filtration method, the
chrysotile peaks were observed to lag behind the magnetic field by about
109 under the conditions, of this, measurement.
5.3.1 Measurement of Blank Samples
One'of the most challenging problems in the use of this equipment
was to obtain a satisfactory blank profile. The system is
extremely sensitive to suspended particulate in the sample, and
it was found that two effects initially prevented achievement of
a flat profile which did not contain scattered light peaks.
The first effect was the presence of a sinusoidal signal dis-
playing one cycle per revolution of the magnet. To investigate
this effect,, a source of scattered light which did not vary with
magnet rotation was required. This was conveniently arranged by
positioning a thin sheet of polyethylene in place of the spectro-
photometer cell. It was found that variation in stray magnetic
field during rotation of the magnet was altering the gain of the
photomultiplier. The effect could not be eliminated without a
significant revision of the experimental equipment, but it was
..minimized by a combination of magnetic shielding and re-positioning
of the photomultiplier. At high values of amplifier gain it was
found that the baseline of the signal still displayed a slight
sinusoidal variation, which could then be removed by profile
subtraction to leave only the relevant signal.
The second effect was a consequence of the inability to produce
absolutely particle-free water. The only water suitable for pre-
paration..^ Jow background samples was found to be water double-
38
-------
90 180 WIT
MAGNETIC FIELD DIRECTION
MAGNETIC FIELD DIRECTION
Figure 26. UlCC.Crocidolite: Dynamic, Figure 27. UICC Amosite. Dynamic
scattered light profile. scattered light profile.
;g
1 **
if
1
ii
l . MAGNETIC FIELD DIRECTION
Figure 28. Union Carbide Chrysotile:
Dynamic scattered light profile.
: 39 :
-------
distilled in glass, with precautions being taken not to disturb .
ground glass joints during the distillation. Filtered water was
in no way satisfactory, since in the context of these measurements
filtration merely replaces one kind of particulate with another
which originates from the downstream surfaces of both the filter
. and the-filtration equipment. For the low background work, double-
distilled water from a_Corning MEGA-PURE still has been found
~ satisfactory,".however,~it~must~be collected.directly in boro-
silicate glass bottles having plastic screw caps with Teflon"
liners. ;
. To obtain a blank scattered light profile, the "particle-free"
double-distilled water must be handled only in a laminar flow
hood. As many as ten successive washings of the cell may be
necessary before a sufficiently low level of particulate conta-
mination is achieved.. Ultrasonic treatment should be avoided,
since this tends to generate particulate.
5.3.2 Measurement, of Detection Levels for Asbestos
Fibers-Dispersed rn Double-Distilled Hater - -
Using dispersions of UICC crocidolite and Union Carbide chrysotile,
the system was. calibrated to obtain values for the minimum detec-
tion limits. The calibrations were performed by preparing a series
of fiber dispersions from serum bottles of standard fiber suspen-
sions., The serum bottles had concentrations of 140 MFL for
.crocidolite and 200 MFL for chrysotile, and the test dispersions
were prepared by dilution of these with "particle-free" water.
The concentrations were spaced "logarithmically at intervals of a
factor of 3 from the serum bottle concentration down to about
0.2. MFL. The spectrophotometer cell was loaded with each disper-
sion in turn, starting with the lowest concentration and working
upwards to avoid cross-contamination. The scattered light profile
of each dispersion was measured for 10 to 20 revolutions at a
magnet speed of 10 rpm for1crocidolite, and for 5 revolutions at
1 rpm for chrysotile. Figures 29 and 30 illustrate the variation
in scattered light signal with fiber concentration.
Because of the effect of the magnetic field on the photomultiplier,
and the response of the non-fibrous contamination to the rotating
field, it 1s necessary to employ profile subtraction techniques to
extract the relevant peaks for very low fiber concentrations. For
these calibration experiments the subtraction was performed
manually; a computer program has ^ince been written to do sub-
tractions. The calibrations obtained are shown in Figures 31 and
32, 'in which peak areas are shown as functions of both numerical
and mass concentrations. The detection level for crocidolite
was found to be about 0.5 MFL, and for chrysotile about 5 MFL.
These detection levels apply to the fiber dispersion in the spec-
trophotometer cell, and correspond to mass concentration detection
levels: of about 180 ng/L and 30 ng/L for crocidolite and chrysotile
...respectively. , . .
-------
fIBGNETIC FULO DIRECTION
Figure 29. UICC Crocidolite: Variation of scattered
light profile with fiber concentration.
T to IM»o Sifl
BMNET1C FIELD OIltCCTION
Figure 30. Union Carbide Chrysotile: Variation of
_ _ _s^.attere^ light profile with fiber concentration.
' ' .41:' ;
-------
MASS CONCENTRATION, ng /L
10 11 ' ' "
10
I0
oc
UJ
0.:
0-1
0-01
0-1.
l-0> :..-, 10
'.' FIBER'. CONCENTRATION . MFL
100
Figure;:31. UICC Crocidolite: Area of P-fiber peak as
-V ' ;-'.a- function of mass and fiber concentration.
10
MASS CONCENTRATION, nq/L
: 10 -." " I02 " I03
1-0
CD
(E
-------
In general, to achieve the target detection level of 0.2 MFL or...
1 ng/L, some selective or non-selective fiber concentration will
be necessary. Although the dynamic fiber analysis technique is
significantly more sensitive than the filtration method, the
ability to detect asbestos fibers can be degraded by the presence
of some other types of particulate.
5V3V3"' Effects 'ofNon-Fibrbus' Parti cuTate ." ""
Figure 33 shows the profile obtained from borosilicate glass par-
ticulate in water. . This material rotates with the magnetic field
and gives rise ;to broad peaks at 45° and 225°. This effect was
not noticed when, using the,fixed fiber alignment technique, pro-
bably because, it was below the detection level of the method and
perhaps also because any preferred orientations of particles
having random shapes would be disturbed when they contacted the ;
filter. . ;.-..'. | ; .
.-.' .:"'.- ')'
The origin of the peaks has not been fully investigated, but can
- -be-explained 'if- the particles- adopt orientations such that their
longest dimensions are parallel" or normal to the magnetic field
direction. The facets of the particles would scatter light pre-
ferentially throughout a broad angle, centered on ± 45° to the
magnetic field direction. This would still occur if the particles
had complete rotational freedom about their axes parallel to the
magnetic field. . . .'..'
Non-fibrous particulate which does not.respond to the magnetic
"field" contributes only to the-constant componenfof the scattered
light intensity. ' ' .
a io Teo ZTO
numeric FIELD DIRECTION
Figure 33. Scattered light profile of borosilicate
glass particle suspension.
:.'..; 43'-":':;.
-------
5.4 EVALUATION OF THE DYNAMIC FIBER TECHNIQUE. .
Of the various approaches examined, the rotating magnet system is capable
of achieving the lowest, detection limits and has the most potential for
future development. At the lowest fiber concentrations of interest,
other particle species present in even very clean water contribute sig-
nificantly to signals which overlap the peaks from some types of asbestos
fiber. General particulate which rotates'with the magnetic field yields
peaks corresponding to an alignment direction of 45° to the field.
However, profile subtraction techniques can be used to extract the peaks
originating from asbestos fibers, which always occur either at 0° and
180°, or at 90° and 2708,. The system is capable of detecting a mass of
1 ng for crocidolite or 0.2 ng for chrysotile in a volume of 5 ml. With
some pre-concentration of the water sample,, it will be possible to detect
0.2 MFL or 1 ng/L of either material. Pre-concentration techniques are
discussed in Section 7. .
44
-------
........... SECTION. 6 .
ALIGNMENT MODES OF SELECTED MINERAL SPECIES IN MAGNETIC FIELDS
The alignment modes of UICC crocidolite, UICC amosite and UICC chrysotile have
already been described in Section 3.1. The studies reported in Sections 4
and 5 have been conducted using UICC crocidolite, UICC amosite, and refined
Union Carbide Calidria chrysotile. For calibration purposes, it was con-
sidered that the purified chrysotile would be more suitable than the UICC
chrysotile. : .. .
The alignment modes of a selected group of minerals were determined in order
that: " - -
(a) potential interferences with the measurement of an
asbestos fiber dispersion can be specified; and,
(b) the different alignment modes can be used as means
of identification where possible.
Aqueous dispersions of a range of fibrous minerals were prepared for deter-
mination of their alignment modes using the dynamic fiber method. Other
fibrous and non-fibrous minerals related to the primary varieties were also
examined in the same way. The scattered light profiles are shown in Figures
34-81. . . :
It is important to recognize that these profiles are qualitative, and indicate
only the possibility of interference. 'Some of the materials, such as
halloysite, did not appear to display strong alignment and on an equivalent
:mass basis may not constitute an interference with the measurement of other
fibrous material. Some of the .scattered light profiles contain contributions
from other particulate which has rotated with the magnetic field, similar to
the effect displayed by borosilicate glass particles. These contributions
could be removed by a profile subtraction technique.
The following general conclusions can be drawn from the scattered Tight pro-
files of the selected minerals. :
(a) All varieties of chrysotile behave in the same way; they
display broad P-peaks which lag behind the magnetic field.
(b) Antigorite can be discriminated from chrysotile; the
antigorite displays relatively sharp N-peaks..
45
-------
MflGNETIC FIELD DIRECTION
Figure 34. Chrysotile .
(UICC Canadian)
! 5
: i
0 _ 90 _UO 2
MRGNET1C FIELD DIRECTION
Figure 35. Chrysotile
(UICC Rhodesian)
370 3<0
MflCNETIC FIELD DIRECTION | ---?
Figure 36. Chrysotile .
(Thetford, Quebec)
Figure 37. Chrysotile
-(Union Carbide)
SCATTEREDLIGHT PROFILES
-------
II
270580
ITOGNETIC FIELD DIRECTION.
Figure 38. Lizardite
.("Owens.Pit", Ontario)
s
a
Iflti 270
URGNETIC FIELD DIRECTION
Figure 39. Picrolite
.. (Broughton, Quebec)
0 90 ISO WO
. nmnric ritui OIHICTION
HRGNeTIC FIELD DIKECTION
Figure 40. Antigorite i
(East Broughton, Quebec)
Figure 41. Talc
{Broughton, Quebec)
47
-------
Figure 42.
(Asbestos, Quebec)
Figure 43. Brucite
(Asbestos, Quebec)
II
3
a
S
MWNETIC FIELD DIRECTION
180 270 380 - . I !
0 10 IK WO
muwrric rieui OIHECTIOH
Figure 44. Tremolite
. (Elzivir, Ontario)
Figure 45. Tremolite
{.Clarendon, Ontario)
SCATTERED LIGHT PROFILES
-------
Figure 46. Tremolite
(Arctic)
Figure 47. Tremolite
(Inyo County, Cal.)
0 10
muaimc rinjj omtcrioit
W«S«TIC riELD OIRECTIOM
Figure 48. Tremolite
.(Transvaal, RSA)
Figure 49. Tremolite
(Yakutya, USSR)
SCATTERED LIGHT PROFILES
49'
-------
numeric new OIUCCTIOR
Figure 50. Actinolite
: (Marbridge,, Quebec)
; nBGNETIC FIELD DIRECTION
Figure 51. Amosite (UICC)
Figure 52.. Amosite
(Lyndberg,. RSA)
Figure 53. Cummingtonite
"(Soutpansberg, RSA)
SCATTERED LIGHT PROFILES
-------
180 270 3GO
iwoirric riEU DIRECTION
0 SO IBO 270
IWQtlETIC FICLO OlltECTION
Figure 54. Cuimiingtonite Figure 55. Cummingtonite
(Mikanui, New Zealand) (Lead, South Dakota)
Figure 56. Grunerite
(Health Lake, Ontario)
Figure 57. Grunerite
(Humbolt, Michigan)
_SCATTEREp LIGHT PROFILES
.:-'.:;- 51:''
-------
Figure 58. Minnesotaite
.(Mesabi, Minnesota)
~53o f"
nRGNETlC FIELD DIRECT I ON
Figure 59. Stilpnomelane
(French Ridge, N.Z.)
d i'o" ido
HBCNETIC FIELD DIRECTION
Figure 60. Hornblende
(Ross, Ontario)
^Figure 61. Suspected Omphacite
(Frontenac Co. Ont.)
SCAERE°
ROFILES
-------
V
HBCNETIC FIELD DIRECTION
Figure 62. Crocidolite
(UICC)
mwiieric ricui oirecrion
Figure 63. Crocidolite
(Prieska, RSA)
Figure 64. Riebeckite Figure 65. Anthophyllite
(St. Peter's Dome,. Col.) (UICC)
SCATTERED LIGHT PROFILES
. . 53
-------
io i»o
FIELD D I SECT lot!
Figure 66. Anthophyllite
(Montauban,. Quebec)
o so 180 270
HRGNETK FIELD DIRECTION
Figure 67. Anthophyllite
(Salt Mountain,
Georgia)
nooime FIELD oimeritm
a so no
KflGNETIC FIELD DIRECTION
Figure 68. Gedrite
(Telemark,, Norway)
Figure 69. Howieite
(Laytonville, CaT.)
SCATTERED LIGHT PROFILES
__ _ &L 54 ::;:':.
-------
~~aSB rio wo
KBCNETIC FIELD DIRECTION
Figure 70. Wollastonite
(Asbestos, Quebec)
0 90 180 270 360
. HBGNETIC FIELD DIRECTION
Figure 71. Wollastonite
(Meldon, Quebec)
;i
180 270 SCO
"'" "' KBCHtTIC FIELD DIRECTION
Figure 72.. Halloysite
(Eureka, Utah)
S 180 270
HflGNETIC FIELD DIRECTION
Figure 73. Halloysite
(Delta, Utah)
SCATTERED LIGHT PROFILES
::; 55.
-------
HflGNETIC FIELD DIRECTION
T ao Tso
MRGNETIC FIELD DIRECTION
Figure 74. Palygorskite" , Figure 75. Palygorskite
.(Metaline Falls, Wash.) (Pormona, Cal.)
4 ;
".',. - MBGNETIC FIEUO DIRECTION ' f,;*-
270360
Figure 76. Xonotlite
(Asbestos, Quebec)
' '! KB6NETIC FIELD DIRECTION
Figure 77. Xonotlite
(Wakefield, Quebec)
SCATTERED LIGHT PROFILES
-------
270 380
MAGNETIC FIELD DIRECTION
Figure 78. Phlogopite
(Phalaborwa, RSA)
90 190 Z70
WWMETIC FIELD OIRKTtQH
Figure 79. Pectolite
(Thetford, Quebec)
0 90 180 WO
ItPGNETIC riCUl DIRECTION
I
SO 180 270 560
HflONETIC FIELD DIRECTION
Figure 80. Pectolite
.(Japan)
Figure 81. Diatomaceous Earth
tSeitz Supra EKS Filter)
SCATTERED^LIGHT PROFILES
57
-------
(c) Brucite. and, lizardite can/be mistaken for chrysotile,
since they both display broad P-peaks.
(d) All tremolite samples and the actinolite sample display
sharp N-peaks only. : :
(e) Cummingtonite generally displays sharp P-peaks with the
exception of"the Mikanuf sample"which aligns at about
15° in the T-mode. .
(f) The.grunerite sample displays sharp N-peaks.
(g) The crocidolite samples display sharp P-peaks.
(h) Anthophyllite .displays broad P-peaks similar to those
of chrysotile. .....-)
(i) Wollastonite displays broad P-peaks similar to those
of chrysotile.
(j) 'The amosite samples display-sharp P-peaks and N-peaks.
It is evident that if the primary purpose is. the detection of "asbestos" there
is some potential for interference by fibrous species other than those nor-
mally considered as asbestos. Broad peaks also occur from platy minerals such
as phlogopite; these are a consequence of the presence of cleavage fragments
which have high aspect ratios. Assuming- that the purpose of the technique is
to determine if,ant/ jj-tbtoui mote/ui£ .4pe.ct.eA' are present, then it is highly
successful, extremely sensitive, and allows'some discrimination between
species. -.. * ; ..''/'..- -,. . :,- ..,..',%..;.-.:'. ':,/;.'.. ... .
It was at first thought that general non-fibrous particulate would contribute
only to the base-line level, and would not yield any peaks. As discussed in
Section 5.3.3,.this is not the case. .The peaks at 45" and 225° from general
particulate ^seriously overlap those at;0° and 180° from chrysotile, and may
also embed the peaks from low concentrations of amphibole. Where there are
.large general particulate peaks, measurements of low fiber concentrations
will be possible with the use of profile subtraction techniques. Where the
general particulate level is too high for an effective measurement to be made,
methods for specific concentration of fibers must be employed.
58
-------
SECTION 7
METHODS FOR CONCENTRATION OF FIBERS
7.1 NON-SPECIFIC FIBER CONCENTRATION
Samples which on direct examination fail to yield any scattered light
peaks at 0° and 180°, or at 90° and 270°,may have fiber concentrations
below the detection limit of the equipment. In these cases non-specific
fiber concentration may be applied to increase the actual fiber concen-
tration to a value within the sensitive range. This can be achieved by
filtration of a large volume of water and redispersal of the collected
particulate in a smaller volume of double-distilled particle-free water.
A Nuclepore capillary pore polycarbonate or polyester filter is used for
this filtration, and the collected .particulate is redispersed in a small
plastic beaker of water by placing it in an ultrasonic bath for a few
minutes. The filter is then removed and discarded, after which the con-
centrated dispersion can be examined using the light scattering equipment.
Along with concentration of the fibers, this simple procedure also con-
centrates the other particulate. The non-fibrous component of the par-
ticipate, particularly that which-is affected by the magnetic field, may
be of too high a concentration for a reliable measurement of the fiber
content to be made. In this case specific, fiber concentration methods
must be employed. . ;
.7.2 REMOVAL OF ORGANIC PARTICLES .
The ozone-ultraviolet oxidation procedure described elsewhere17*18.19
is an effective means for removal of organic particles from drinking
water supplies. Although some refractory organics are not affected,
improvement in filtration rate always occurs after the treatment, which
indicates that significant removal of organic particulate has been
achieved. Essentially, 1% ozone gas is bubbled through the sample while
it is irradiated by short wavelength ultraviolet light from a submerged
lamp. Treatment for about 3 hours has been found adequate for most water
samples; a noticeable reduction in turbidity is usually observed.
7.3 SPECIFIC CONCENTRATION OF CHRYSOTILE .
The two-phase:liquid separation (TPLS) technique11-was originally thought
to be directly applicable to specific concentration of chrysotile.
However, although it was highly effective in separation of chrysotile
from artificially-prepared dispersions, it produced no extraction of
chrysotile from drinking water collected in.Sherbrooke,-Quebec.. This
:-. - 59"''
-------
drinking water is known to contain 40 - 80 MFL of chrysotile, measured
by electron microscopy. This technique, therefore, was not investigated
further.
As part of the overall research program on analytical methods,17 a dis-
covery was made that chrysotile asbestos can be collected on the inside
surfaces of unsterile plastic containers. Although the precise mechanism
is not yet fully understood,' it .is known that the' surface electrical
charge of'chrysotile causes some complex organic materials of high mole-
cular weight to attach to its surface. These complex organic materials,
probably polysaccharides, are secreted by some single cell biological
organisms which seem always to be present in drinking water. When the
plastic container.is shaken in a reciprocal laboratory shaker, the
organisms, along with scavenged chrysotile fibers, contact the internal
surface of the container more frequently and become attached to it.
Complete attachment of all chrysotile is obtained using this procedure
on artificial dispersions of Union Carbide or UICC Canadian chrysotile
in distilled, but unsterile, water. The most abundant organism in this
water was a species of pseudomonas, at a concentration of about 10'
viable organisms per liter. The technique is also effective in removal
of chrysotile from Sherbrooke drinking water.
The attachment of chrysotile to the container walls is achieved to a
maximum extent after the sample has been shaken at a frequency of about
2.5 cycles/second for a minimum of 48 hours. The liquid sample itself,
which then contains only a. small proportion of .the original chrysotile
and all of the other particulate,...is discarded. The bottle is refilled
using clean double-distil.led water, and .treated using the ozone-UV tech-
nique, which oxidizes the: organic-:materials~and-releases the fibers- from-
the container walls. The bottle is afterwards treated in an ultrasonic
bath for about 15 minutes to ensure complete dispersal of the fibers. A
suspension of suitable concentration of chrysotile for measurement in the
light scattering system can then.be prepared by the filtration technique
described in 7.1. -.- ' ; . -','; ,, .
The observation of this biological scavenging effect has serious impli-
cations for the sample preparation techniques normally used for measure-
ments of fiber concentration by electron microscopy; a complete account
of the studies has been described.17*20
The biplogical scavenging of asbestos fibers did not appear to be com-
pletely specific for chrysotile asbestos; crocidolite and amosite also
displayed some separation, although studies of the separation of these
amphibole types were not completed in view of the greater promise shown
by magnetic separation techniques. Table 5 shows examples of the results
obtained.
The recoveries for artificially-prepared dispersions in distilled water
were usually close to 100%, and with very few fibers remaining in sus-
pension after the shaking operation. For a naturally-occurring chryso-
tile dispersion from Sherbrooke, Quebec the recovery was about 45%,
indicating either that some fundamental, property of the fibers was
60 :
-------
TABLE 5. SUMMARY OF FIBER REMOVAL FROM WATER SAMPLES
Sample
UICC Chrysotile
in unsterile
distilled
water.
Union Carbide
Chrysotile in
unsterile
distilled
water.
Sherbrooke
(Quebec)
Municipal
water
UICC
Crocidolite in
unsterile
distilled
water
UICC
Crocidolite in
unsterile
distilled
water
UICC Amosite
in unsterile
distilled
water
Mississauga
(Ontario)
Drinking
Water
Spiked with
Crocidolite
Concentration
of Initial
Suspension,
MFL
23.6
78.4
42
32.3
140
Not
Measured
12.9
Concentration
of Fibers
Remaining in
Suspension,
MFL
5.0
< 0.5
24
0.7
47
2.0.
6.0
Concentration
of Fibers
Separated
from the
Initial
Suspension,
MFL
32.5
64.5
18
24.5
118
12.1
5.6
Recovery
Efficiency,*
(Percent)
87
99
45
97
70
85
45
*The recovery efficiency was determined as the ratio of the
concentration of separated fibers to the sum of the separated
fibers and those remaining in suspension.
Concentration expressed as a value referring to the volume
of the initial suspension.
61
-------
different or that the bacteria were in some way less efficient in their..,
scavenging behavior. The same consideration applies to crocidolite added
to a municipal drinking water,..which was separated with a reduced
efficiency of 45%.. .,.;"._:
Because of the simplicity, of the technique, it is directly applicable to
the rapid screening method. Further investigation of the effect is
: needed to develop" it'lnto"a 'controlTed"procedure. " " "
7.4 SPECIFIC CONCENTRATION OF AMPHIBOLES. . .. .
Magnetic separation21*22 of UICC amosite from aqueous suspensions was
described by Timbrel!.23 The published data, however, are based on
: retention-efficiencies-measured bythe'magnetic alignment method, and do
not indicate the numerical fiber retention efficiency, nor whether the
separation is strongly, size selective. , As part of the investigation of
concentration methods,,, the published design of magnetic separator was
investigated, using an aqueous suspension of UICC amosite. The apparatus
used is shown in .Figure 82; 0.5 grams of magnetic stainless steel wool with
...... - a.wire;diameter of;80;ym;.occupied;:1.^5 cm -immediately above the outlet valve
at .the.bottom of. a;.50 mL..buret.,, The-.buret was arranged so that-the steel
wool was between the poles of,a magnet. Each 10 ml aliquot of the -
v amosite suspension was successively passed through the separator at a
flow rate of about 2,:mL/minute without attempting to remove the fibers :
retained from previous, samples. The magnetic field was increased to a
higher value for each;successive aliquot, to the maximum field of 1.5 T.
The filtrates were prepared for fiber counting'in the TEM. Filtrations
. for the TEM:samples were carried out in a.magnetic field, so that
"separate measurements, could be made for P-type and N-type fibers. The
.- results are. shown'in Figures 83 and 84, in which the numerical retention
. efficiencies are shown;as functions of magnetic field strength. The
retention efficiencies for both fiber types are similar. The maximum
retention efficiency,, for both.?- and N-fibers, occurred at a field
strength above 1.0 T.;;v>t. ;.,.'; .."-./ ; _ -...... _
' .- The'observation that'the retention efficiency showed a similar correla-
tion with magnetic field strength.for both P- and N-fibers shows that
the retention is not an effect of fiber orientation. P-fibers are
aligned in a horizontal plane by the magnetic field and pass through the
steel wool .in a direction normal to their lengths. Since N-fibers are
aligned with their lengths normal to the field, their orientations rela-
tive, to the direction of movement are random. If retention depended only
on fiber orientation relative to direction of motion, then the retention
efficiency for N-fibers would be substantially less than for P-fibers.
The close similarity in retention,efficiency for P- and N-fibers suggests
that fiber entrapment is; a result.:of magnetic attraction between fibers
and, the steel wool. :
Figure 84 shows the cumulative fiber number retention efficiency of the
separator as a function of both magnetic field and fiber length. It can
. be seen that the separator is most efficient at the highest magnetic
_ field strengths, butJthat.fields..considerably higher than 1.5,T would be
. ' - '.-' \. - ":: 62. ..- '
-------
Figure 82. High gradient magnetic separator.
63
-------
o) N-FIBERS
b) P-FIBERS
100
a
UJ
z
UJ
IT
V)
ir
80
60
O
uj 40
o
UJ
O
K
Ul.
0.
20
0 0-5 1-0 1-5
MAGNETIC FIELD STRENGTH CM
0 0-5 1-0 1-5
MAGNETIC FIELD STRENGTH(T)
Figure 83. The effect of magnetic field strength on retention
. of amosite fibers; a) N-fibers and b) P-fibers.
100
u
o
UJ
Q.
UJ
Ul
K
CC
Ul
m
u.
UJ
S
U
30
OI.5T.
*I.IT
*0.7T
0.4 T
.2T
i i i i i I
0-3
10
FIBER LENGTH,
100
Figure 84. Magnetic separator retention efficiency for UICC amosite.
':.'. ' :*:,';: 64 /:.'-!.: . '
-------
required to improve the performance significantly for fibers longer than
about 1 urn. The efficiency for collection of fibers shorter than 1 ym
falls very rapidly with fiber length. Nevertheless, in this configura-
tion at a field of 1.5 T.the; cumulative efficiency still exceeds 50%
for fibers shorter than 0.5 ym. At 1.5 T the efficiency of the separator
in terms of mass is over 99.8%. This is a consequence of the fact that
the larger fibers which_comprise most of thejnass are retained more
efficiently".""" " - -;--
Measurements using crocidolite dispersions indicate that the collection
efficiency for crocidolite was very high in a magnetic field of 1.5 T.
In these experiments, aliquots of a crocidolite dispersion were passed
through the separator in a magnetic field of 1.5 T. The separator was
then removed from the field, and 5 ml of distilled water were added. The
collected.fibers were redispersed by vigorous shaking, after, which this
washing procedure was repeated. The redispersed fiber suspensions, the
filtrates,, and samples of the original dispersion were then filtered in
a magnetic field in order to prepare aligned fiber samples. The scat-
tered light profiles for these samples were then obtained, and the
results-indicated .that although the collection efficiency was high, the
redispersal step'yielded a recovery of only about.50%. It is possible
that more complete removal couTd.be obtained by the use of ultrasonic
treatment. . . .
Some studies have been performed using 400 mesh size electrolytically-
etched nickel mesh as the magnetic separator element instead of the stain-
less steel wool. This shows promise as an element from which more com-
plete removal of the collected fibers can be achieved without contamina-
tion from fragments of stainless steel. Collection efficiencies have not
been measured systematically, but the initial results show that values
comparable with those o.f the steel wool can be achieved. There is also
considerable potential for further improvements in the design of magnetic
separator elements for this application. It is: important to recognize,
however, that, the separation is dependent on the iron content, and that
the efficiency of collection of low-Iron amphiboles such as tremolite may
not be possible with simple systems such as that discussed.
65
-------
..SECTION,8 _.,_ ..
' EVALUATION OF'THE RAPID SCREENING TECHNIQUE:
APPLICATION TO MUNICIPAL:DRINKING WATER SAMPLES
The dynamic-fiber screening technique has, been applied to the examination of
three well-characterized municipal water samples. Beaver Bay, Minnesota
drinking water is unfiltered Lake Superior water taken from a location about
10 miles .south-west of Silver Bay. A sample- of this water was ozone-UV treated
to remove organics,, and then-agitated in the ultrasonic bath for about 15
minutes. The scattered'light profile obtained directly and without concentra-
tion is shown as the,solid curve.in Figure 85. .It can be seen that the peaks
are asymmetric, > This profile, contains a component from the fibrous particulate
superimposed on large general particulate peaks centered on angles of 45° and
225° to the magnetic field direction. A general particulate profile obtained
from ground glass particles has been subtracted from the total profile, leaving
an extracted profile which contains residual peaks at 0° and 180°, and at 90°
and 270°. The extracted profile is plotted as.the. broken line. The two sets
of peaks correspond to the known presence, of cummingtonite-grunerite fibers at
a concentration of about 17-MFL.;-:7 ". : -:'_
A sample from Sherbrooke, Quebec was examined, following the same method of
specimen preparation. The scattered light profile obtained without further
concentration is shown in Figure 86. Peaks occur in the profile, from which
the general particulate contribution at 45° and 225° can be subtracted, leaving
the chrysotile-type peaks at 0° and 180° in the extracted profile. This water
contains about 40 MFL of chrysotile,. as determined by TEM methods.
Drinking water in Mississauga, Ontario was last-measured by TEM methods in
1977 and was found to contain less than about 2 MFL of chrysotile. The pro-
file obtained for this water without any oxidation treatment is shown in
Figure 87. The general particulate profile was subtracted, leaving an
extracted profile containing; only random instrumental noise. The profile is
consistent wi th the; known low fiber concentration, which is below the detec-
tion level of the equipment. , :
These three examples show the scattered light profiles obtained from measure-
ment of municipal waters containing 17 MFL amphibole, 40 MFL chrysotile, and
less than 2 MFL chrysotile .respectively. It has been shown in Section 5.3.2
that the detection level for direct measurement of chrysotile is 5 MFL, and
for amphibole is 0.5 MFL.. In order to achieve the target detection level of
0.2 MFL for chrysotile, a pre-concentration by a. factor of 25 is required;
this can be achieved by either selective or non-selective pre-concentration.
Amphibole fibers require pre-concentration by only a^ factor of 2.5 to achieve
-------
0 95 lia z5o
KPGI4CTIC flELO DIRECTION
Figure 85. Scattered light profile of water sample from
Beaver Bay, Minnesota, before .and after
subtraction of general particle peak.
4»
SO 1*0270
HSBKSTIC FIELD DIRECTION
Figure 86. Scattered light profile of water sample from
Sherbrooke, Quebec, before and after
__ subtraction of general particle peak.
67
-------
MEASURED
PROFILE
HBGKETIC PItUJ DIRECTION
Figure 87. Scattered light profile of water sample from
Mississauga, Ontario,' before and after
subtraction of general particle peak.
the target detection level and, depending on the iron content, this can be
achieved by simple filtration and re-dispersal, or by the use of magnetic
separation. .,.
68
-------
REFERENCES.
1. Anderson, C.H. and Long, J.M. (1980) Interim Method for Determining
Asbestos in Water. U.S. Environmental Protection Agency, Athens,
Georgia. EPA-600/4-80-005, National Technical Information Service,
Springfield, Virginia 22161. ; - :-rc
2. Zielhuis,. R.L.:(editor)'(1977) Public Health Risks of Exposure to
Asbestos. Commission of the European Communities, Directorate-General
for Social Affairs, Health,and, Safety Directorate,. Pergamon Press Inc.,
: Maxwell House, Fairview Park, Elmsford, New. York 10523, U.S.A.
3. Barbeau,~C.~'(1979) Evaluation of Chrysotile by Chemical Methods. In:
Mineralogical Techniques of Asbestos Determination (R.L. Ledoux, ed.).
Mineralogical Association of Canada, Department of Mineralogy, Royal
Ontario Museum, 100 Queen's Park, Toronto, Ontario, Canada, M5S 2C6,
p. 197-211.
\ ,- " " . '. -
4. Crable, J.V.. (1966) Quantitative. Determination of Chrysotile, Amosite
and Crocidolite by X-ray Diffraction. Amer. Indust. Hygiene Assoc.
J. 27., p. .293-298. . .
5. Crable, J.V. and Knott, M.S. (1966a) Application, of X-ray Diffraction
to the Determination of Chrysotile in Bulk or Settled Dust Samples.
Amer- Indust. Hygiene Assoc. J.;2_7_, p. 383-387.
6. Crable, J.V.. and Knott;. M.S. (1966b) Quantitative X-ray Diffraction
Analysis of Crocidolite and Amosite in Bulk or Settled Dust Samples.
Amer.. Indust. Hygiene Assoc. J; 27, p< 449-453.
7. Ocella, E. and Maddalon,-G. (1963) X-ray Diffraction Characteristics of
Some Types of Asbestos in Relation to Different Techniques of Comminution.
Medna. Lav., 54_, p. 628-636./, ; : .. .
' . '''' ' '"' '. - ' ' '' -
8. Heidermanns,: G.. (1973) Asbestos'Content Determination by Optical,
Chemical, Radiographic and IRSpectographic Analysis Procedures.
Staub (Engl..)-33, p. 67-72. : , /: \.;^ ::;.'.
9. Birks, L.S.,; Fatemi.M., Gilfrich, J.V. and Johnson, E.T. (1975)
Quantitative Analysis of Airborne Asbestos by X-ray.Diffraction:
: Final Report on. Feasibility Study,. EPA-650/2-75-004. U.S. Environ-
mental Protection Agency,- National Environmental Research Center,
Research Triangle. Park,, North Carolina:,, 27711.
..69
-------
.10. Middleton, A.P. (1979) The Identification of Asbestos in Solid Materials.
In: Asbestos Properties.Applications and Hazards, ed. Michaels, L. and
Chissick, S.S. Published by John Wiley & Sons Ltd., New York, p. 260-277.
11. Melton, C.W., Anderson,, S.J., Dye; C.F., Chase, W.E." and Heffelfinger,
R.E. (1978) Development of a Rapid Analytical Method for Determining
Asbestos in Water, EPA-600/4-78-066. U.S. Environmental Protection
Agency, Environmental'Research Laboratory, Athens, Georgia 30613. -
12. Diehl, S.R.,. Smith, D.T.. and Sydor, M. (1979) Optical Detection of
Fiber Particles in Water, EPA-600/2-79-127. U.S. Environmental
Protection .Agency, Municipal Environmental Research Laboratory,
i Cincinnati, Ohio, 45268. : - ;. , .;
13. Timbrel!', V. (1975) Alignment of Respirable Asbestos.Fibres by Magnetic
Fields. Annals of Occupational. Hygiene, JL8_, p. 299-311.
" ' * .-.' .' '. ' ; - ' .-.':'..'..£. ' ' '
14. . Cressey, B.A. and Whittaker, E.J.W. (1980) Magnetic Orientation of
Amphibole Fibres. Personal communication from B.A. Cressey.
15. .Asbestos International Association (1979) Reference Method for the
Determination of Airborne Asbestos- Fibre Concentrations at Workplaces
: by Light Microscopy (Membrane Filter Method). Asbestos International
Association, ,68 Gloucester Place, .London, W1H 3HL, England.
.16. Ortiz, LAV and Isom, B.L. (1974) Transfer Technique for Electron
Microscopy of Membrane Filter Samples. American Indust. Hygiene Assoc. .
: Journal 35 ,(No.-7) v p. 423^425.. :;; - ::. :
17. ChatfieTd, E.J..,. Stott, W.R. and bilTon,.M.J. (1982) Development of
: Improved Analytical Techniques for Determination of Asbestos in Water
' Samples (Final,Report). Contract No. 68-03-2717, S.E. Environmental
Research Laboratory, U.S. Environmental Protection Agency, College
i Station Road, Athens, .Georgia,-30613.VU.S.A./
18. Chatfield, E.J. (1979) Preparation and^Analysis of Particulate Samples
by Electron Microscopy, with Special Reference to Asbestos. Scanning
Electron Microscopy/1979/1,; SEM Inc.,.AMF: 0'Hare, IL 60666, p. 563-578,
486. . '. ...;...-. v; -.^/;- '. '. :'. .''..';.',: .
19. Chatfield, E.J. (1978) A New Technique for Preparation of Beverage
Samples for Asbestos Measurement by Electron Microscopy. Ninth
International Congress on Electron Microscopy, Toronto, Vol. II,
; ' p. 102-103.;. ; ........ ../:..',-.; .-". }.;.. />. .",.-.
20. Chatfield, E.J. (1980) Analytical:'Procedures and Standardization for
Asbestos Fiber Counting in Air, Water and Solid Samples. N.B.S.
Special Publication 619, Proceedings, of the NBS/EPA Asbestos Standards
Workshop: Materials and Analytical Methods, National Bureau of Standards,
Gaithersburg, Maryland, (Issued 1982), U.S. Government Printing Office,
'.. Washington, D.C.,, 20402, p. 91-107.
-------
I ^ ,. ^a. -. ,\ . ;' j , - ,. , ~ -. ;-. * - -. _,.-.»... -I-.. ,. ..,.._-1.-..,.r.i .,^-mi«r-,-- -,w*i. , u> - . .,.-..-.. __ -. , f. , . ;.
21. Oberteuffer, J.A._ (1973) High Gradient Magnetic Separation, IEEE
Transactions on Magnetics, 9., P- 303-306.'
22. Oberteuffer, J.A. (1974) Magnetic Separation: A Review of Principles,
Devices and Applications, IEEE Transactions on Magnetics, 10_, p. 223-238.
23. Timbrel], V. (1977) Magnetic Separation of Respirable Asbestos Fibres,
Filtration and Separation, 1£, Number 3," pi 241-242." "'"' "
. 71:
------- |