EPA-600/4-78-011
January 1978
Environmental Monitoring Series
> ' v ' <*' ''''%>''""*' y^.-''#/V <-', ., ''''^"^ " *sr **"*^v* :/'A
^^ '' ^---'- AA-'-J-^^J.. ^L.r .f i^^rf'^i ^^ t '.jTS^^^-g^."-t'jgk " '%
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/4-78-011
January 1978
PREPARATION OF WATER SAMPLES
FOR ASBESTOS FIBER COUNTING
BY ELECTRON MICROSCOPY
Eric J. Chatfield
Roger W. Glass
M. Jane Dillon
Ontario Research Foundation
Mississauga, Ontario L5K IBS, Canada
Contract Number 68-03-2389
Project Officer
Charles H. Anderson
Analytical Chemistry Branch
Environmental Research Laboratory
Athens, Georgia 30605
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GEORGIA 30605
-------
DISCLAIMER
This report has been reviewed by the Environmental Research Laboratory,
U.S. Environmental Protection Agency, Athens, GA, and approved for publi-
cation. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute endorsement
or recommendation for use.
11
-------
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.
The widespread use of asbestos-containing materials gives rise to
concern about exposure of the general population to low level concentrations
in air, water supplies, and food. Although hazards associated with the
inhalation of asbestos at high concentrations are recognized, the health
significance of ingested particles is not fully understood. An important
first step is the development of a method for quantitative determination
of trace concentrations of asbestos minerals in water. This report,
which investigates several critical factors involved in the analytical
procedure, is a contribution toward the development of a standard asbestos
analytical method.
David W. Duttweiler
Director
Environmental Research Laboratory
Athens, Georgia
m
-------
ABSTRACT
The analytical procedures used by a number of laboratories for the
analysis of asbestos fibers in water samples have been investigated, with the
overall objective of recommending an optimum technique.
Standardized aqueous dispersions of chrysotile, crocidolite and taconite
tailings were prepared, samples of which were filtered so as to generate
approximately 150 identically loaded membrane filters for each type of fiber.
Filtration problems when using polycarbonate filters (Nuclepore) were solved,
and an optimum filtration technique was developed which permits microscopically
uniform deposits to be obtained. Such uniform deposits are essential, since
only a small area of the filter is eventually examined and assumed to be
representative. A large number of replicate analyses were made using the
filters, to establish fiber losses and reproducibility of the various tech-
niques. The preparation techniques investigated were:
(a) Carbon-coated Nuclepore with Jaffe washer dissolution;
(b) Uncoated Nuclepore with Jaffe washer dissolution;
(c) Both 0.22 vim and 0.45 ym pore size Millipore, with
Jaffe washer dissolution;
(d) 0.45 ym pore size Millipore with condensation washer
dissolution;
(e) Ontario Research Foundation (ORF) ashing technique.
In this method a Millipore filter is first ashed and the
ash redispersed in water. The resulting dispersion is
filtered through a 0.1 ym Nuclepore which is then processed
by technique (a).
The carbon-coated Nuclepore technique was found to have essentially zero
fiber losses. Electron micrographs obtained both before and after the Jaffe
washing operation show that all particles visible on the filter surface prior
to washing can be located on the final replica. The Ontario Research
Foundation ashing technique was also found to have essentially zero fiber
losses for chrysotile and taconite, but a statistically significant 41% mean
loss was found for crocidolite. The uncoated Nuclepore technique was found to
give mean losses of up to 33%. There was some evidence that in the case of
taconite, both the ORF ashing method and the Nuclepore-Jaffe method yielded
final preparations in which the fibers were less aggregated.
iv
-------
In contrast, all of the Millipore techniques, using either the Jaffe
washer or the condensation washer, produced mean fiber losses of up to 80%.
For the 0.45 ym pore size Millipore preparations, no significant differences
were found in fiber losses between the use of the Jaffe washer and the conden-
sation washer. The results from the replicate Millipore samples indicate that
these techniques are not reproducible in that coefficients of variation for 10
measurements as high as 0.8 were observed. Both of the Nuclepore techniques
and the ORF ashing technique consistently achieved coefficients of variation
better than 0.4.
The uniformity of the deposits on all the final electron microscope
preparations was expressed quantitatively, and the most satisfactory perform-
ance was given by all of the Nuclepore techniques, both with and without the
ashing step. The uniformity of the deposits using the Millipore techniques
was generally very poor, indicating significant particle migration during the
washing operations.
The effects of ultrasonic treatment at 20 kHz on fiber dispersions of
both chrysotile and crocidolite were also evaluated. The fiber concentration
of suspensions of chrysotile was shown to increase by about 0.2% per second
at an ultrasonic power density of 0.5 watt/ml, and by about 0.1% per second
at 0.05 watt/ml. No effect could be detected on crocidolite, even after 300
seconds.
This report was submitted in fulfillment of Contract No. 68-03-2389 by
the Ontario Research Foundation under the sponsorship of the U.S. Environ-
mental Protection Agency. This report covers the period March 15th 1976 to
March 14th 1977, and work was completed as of March 14th 1977.
-------
-------
CONTENTS
Foreword i i i
Abstract iv
Fi gures vi i i
Tab! es xi
Acknowledgement xi i i
1. Introduction 1
2. Conclusions 5
3. Review of Techniques Investigated
3.1 Carbon Coated Nuclepore Technique 7
3.2 Direct Dissolution of Nuclepore Filters onto Carbon-
Coated Grids 7
3.3 Jaffe Washer Dissolution of 0.45 ym and 0.22 ym Pore
Size Millipore Filters 11
3.4 Condensation Washer Technique 11
3.5 Ontario Research Foundation Ashing Technique 16
4. Experimental Approach 21
5. Preparation of Standard Dispersions 23
6. Filtration 25
7. Sample Preparation and Analysis 32
7.1 Carbon-Coated Nuclepore Preparation 32
7.2 Direct Dissolution of Uncoated Nuclepore onto Carbon-
Coated Gri ds 33
7.3 Jaffe Washer Dissolution of Mi Hi pore Filters 33
7.4 Condensation Washer Dissolution of 0.45 ym Pore Size
Mi 11i pore Fi1ters 33
7.5 Ontario Research Foundation Ashing Technique 34
7.6 Fiber Counting 35
8. Statistical Treatment of Results 35
8.1 Uniformity of Deposit on the Electron Microscope Grids 36
8.2 Best Estimate and Confidence Interval of the Fiber
Concentration 36
8.3 Statistical Significance of Loss Measurements 37
8.4 Variability Within Techniques 38
9. Results 39
9.1 Contamination of 0.45 ym Millipore Filters by
Chrysotile 53
9.2 Identification of Amphiboles by SAED 53
vii
-------
9.3 Discussion of Fiber Losses and Reproducibility 53
9.3.1 Chrysotile 57
9.3.2 Crocidolite 57
9.3.3 Taconite 57
9.4 Summary of Technique Characteristics 60
9.5 A Demonstration of Negligible Fiber Losses by the
Coated Nuclepore Technique 62
10. Effects of Ultrasonification 71
11. References 88
Appendix A: Individual Sample Results 90
viii
-------
FIGURES
Number Page
1 SEM Micrograph of 0.1 ym Pore Size Nuclepore Filter Surface 8
2 Jaffe Washer 9
3 Alternative Design of Jaffe Washer 10
4 SEM Micrograph of Millipore Filter Surface 12
5 Condensation Washer, Shown Inside Protective Draft Shield 13
6 Detail of Condensation Washer Cold Finger and Brass Specimen
Gri d Hoi der 14
7 Condensation Washer Cold Finger in Position 15
8 Low Temperature Plasma Asher (Two Chamber Type) 17
9 Ashed Samples in Beakers Being Removed from Single Chamber Asher 18
10 Ultrasonic Probe for Redispersal of Ashed Residue 19
11 Positive Pressure Clean Room Required for Contamination Control
During Specimen Preparation 20
12 Filtration Manifold in use for Simultaneous Filtration of Water
Sampl es 26
13 Optical Micrograph of Nuclepore Filter, Showing Uneven Deposit
of Particulate 27
14 SEM Micrograph of Taconite deposit on Nuclepore Filter, Showing
Uneven Deposit of Particulate 27
15 SEM Micrograph of Glass Frit Filter Support, Showing Flat Areas
Which Restrict Filtration Rate Locally 28
16 Optical Micrograph of Nuclepore Filter with Deposit of Taconite,
Showing Difference Between Backed and Unbacked Areas 28
17 Optical Micrograph of Nuclepore Filter with Deposit of Taconite,
Showing Comparison of Untreated and Plasma Treated Filters 30
ix
-------
18 Optical Micrograph of Nuclepore Filter with Deposit of
Particulate, Showing Hydrophobic Areas of Filter which
Permitted no Liquid to pass through 30
19 Fiber Length Distribution, Chrysotile 40
20 Fiber Length Distribution, Confirmed Crocidolite 41
21 Fiber Length Distribution, Probable Crocidolite 42
22 Fiber Length Distribution, Total Crocidolite Fibers 43
23 Fiber Length Distribution, Confirmed Taconite 44
24 Fiber Length Distribution, Probable Taconite 45
25 Fiber Length Distribution, Total Taconite Fibers 46
26 Effect of Fiber Length on SAED Identification 52
27 SEM Micrograph of Carbon-Coated Nuclepore Filter, Showing
Taconite Deposit, Prior to Jaffe Washing 64
28 TEM Micrograph of Carbon Replica After Jaffe Washing, Showing the
Same Area of Deposit (as in Figure 27). Note the Correct Relative
Positions of the Particles, Indicating No Particle Loss or Move-
ment 65
29 SEM Micrograph of Carbon-Coated Nuclepore Filter with Taconite
Deposit, Prior to Jaffe Washing 66
30 TEM Micrograph of Carbon Replica after Jaffe Washing, Showing the
Same Area of Deposit (as in Figure 29X Note that Even the Very
Fine Fibers Visible in the TEM Image can be Traced on the SEM
Image 67
31 SEM Micrograph of Carbon-Coated Nuclepore Filter with Chrysotile
Deposit, Prior to Jaffe Washing. Note Very Low Contrast of
Chrysotile Fibrils 68
32 TEM Micrograph of Carbon Replica After Jaffe Washing, Showing the
Same Area of Deposit (as in Figure 30) Note that all Detectable
Fibrils and Other Particles have been Retained in Position,
Indicating No Particle Loss 69
33 Effects of Ultrasonification, Chrysotile 73
34 Effects of Ultrasonification, Crocidolite 74
35 Effects of Ultrasonification, Chrysotile Size Distributions 75
36 Effects of Ultrasonification, Chrysotile Size Distributions 76
-------
37 Effects of UHrasonification, Chrysotile Size Distributions 77
38 Effects of UHrasonification, Chrysotile Size Distributions 78
39 Effects of UHrasonification, Crocidolite Size Distributions 79
40 Effects of UHrasonification, Crocidolite Size Distributions 80
41 Effects of UHrasonification, Crocidolite Size Distributions 81
42 Effects of UHrasonification, Crocidolite Size Distributions 82
43 Effects of UHrasonification, Percentage Increase as a Function
of Exposure Time - Chrysotile 86
44 Effects of UHrasonification, Percentage Increase as a Function
of Exposure Time - Crocidolite 37
-------
TABLES
Number Page
1 Summary of Experimental Results: Chrysotile 47
2 Summary of Experimental Results: Crocicolite 48
3 Summary of Experimental Results: Crocidolite 49
4 Summary of Experimental Results: Taconite 50
5 Summary of Experimental Results: Taconite 51
6 Fiber Loss and Variability Analysis: Chrysotile 54
7 Fiber Loss and Variability Analysis: Crocidolite 55
8 Fiber Loss and Variability Analysis: Taconite 56
9 Fi ber Loss Di stri buti on Summary 58
10 Summary of Principal Results 59
11 Summary of Grid Uniformity and Fiber Loss Measurements 61
12 Effects of Ultrasonification, Aqueous Chrysotile Dispersion 72
13 Effects of Ultrasonification, Aqueous Crocidolite Dispersion .... 72
14 Effects of Ultrasonification, Chrysotile: Comparison of Mean
Gri d Square Fi ber Counts 84
15 Effects of Ultrasonification, Crocidolite: Comparison of Mean
Grid Square Fiber Counts 85
TABLES IN APPENDIX
A-l Carbon-Coated Nuclepore Technique (Initial 10 Samples) Chrysotile 91
A-2 Carbon-Coated Nuclepore Technique (Second 10 Samples) Chrysotile 92
A-3 Jaffe Washer Technique, 0.1 ym Nuclepore, Chrysotile 93
XT
-------
A-4 Jaffe Washer Technique, 0.45 ym Millipore, Chrysotile 94
A-5 Condensation Washer Technique, 0.45 ym Millipore, Chrysotile 95
A-6 Jaffe Washer Technique, 0.22 ym Millipore, Chrysotile 96
A-7 ORF Ashing Technique, Chrysotile 97
A-8 Carbon-Coated Nuclepore Technique (Initial 10 Samples) Crocidolite 93
A-9 Carbon-Coated Nuclepore Technique (Second 10 Samples) Crocidolite 99
A-10 Jaffe Washer Technique, 0.1 ym Nuclepore, Crocidolite 100
A-ll Jaffe Washer Technique, 0.45 ym Millipore, Crocidolite 101
A-12 Condensation Washer Technique, 0.45 ym Millipore, Crocidolite ... 102
A-13 Jaffe Washer Technique, 0.22 ym Millipore, Crocidolite 103
A-14 ORF Ashing Technique, Crocidolite 104
A-15 Carbon-Coated Nuclepore Technique (Initial 10 Samples) Taconite . 105
A-16 Carbon-Coated Nuclepore Technique (Second 10 Samples) Taconite .. 106
A-17 Jaffe Washer Technique, 0.1 ym Nuclepore, Taconite 107
A-18 Jaffe Washer Technique, 0.45 ym Millipore, Taconite 103
A-19 Condensation Washer Technique, 0.45 ym Millipore, Taconite 109
A-20 Jaffe Washer Technique, 0.22 ym Millipore, Taconite 110
A-21 ORF Ashing Technique, Taconite Ill
A-22 ORF Ashing Technique, Chrysotile 112
A-23 Chrysotile Contamination in Mi Hi pore Filters 113
A-24 Effects of Ultrasonification, Aqueous ChrysotiIB Dispersion 114
A-25 Effects of Ultrasonification, Aqueous Chrysotile Dispersion 115
A-26 Effects of Ultrasonification, Aqueous Crocidolite Dispersion 115
A-27 Effects of Ultrasonification, Aqueous Crocidolite Dispersion 117
xm
-------
ACKNOWLEDGEMENTS
The authors wish to acknowledge Dr. Philip M. Cook, of the National
Water Quality Laboratory, Duluth, for his assistance in obtaining the
taconite sample. They also wish to express their appreciation to
Dr. 6.S. Watson, Department of Statistics, Princeton University for suggesting
the grid uniformity test, and to Dr. C.H. Anderson for his continued interest.
xiv
-------
SECTION 1
INTRODUCTION
Asbestos is a term used to describe a variety of hydrated silicate
minerals which have one common attribute: the ability to be separated into
soft, silky fibers of colloidal dimensions. There is no single mineral known
as asbestos; rather there are two major groups, containing a total of six
varieties of minerals (1). The two groups are classified on the basis of
their crystal structures: serpentine and amphiboles. The sole member of the
serpentine class is chrysotile, which accounts for some 90% of the asbestos
fiber produced today. There are five recognized types of amphibole asbestos:
crocidolite, amosite, anthophyllite, tremolite and actinolite.
Chemically, chrysotile comprises hydrated magnesium silicate, Mg3.Si205.
(OH)^ with minor inclusions of iron and aluminum oxides. The amphibole
asbestoses are much more complex, being based on the general composition
(X)7.Si8 022-(OH)2, in which X may be Na, Fe2+, Fe3+, Mg or Ca in various
combinations. It is a common misconception that the asbestos minerals are
"indestructible". In reality all are susceptible to both chemical and thermal
degradation. Chrysotile is susceptible to acid attack, the degree and rate
of dissolution being dependent on the temperature, concentration and nature
of the acid, and also to some extent on the source of the chrysotile (2-3).
In contrast to chrysotile, the amphiboles are more acid resistant, although
there are significant differences between varieties. The resistance of
asbestos minerals to attack by reagents other than acids is generally consid-
ered excellent at temperatures up to 100°C, but deteriorates rapidly at higher
temperatures. Despite their relatively high fusion temperatures, asbestos
minerals are all completely decomposed at temperatures of 1000°C or lower,
depending on the variety. Independently of the surrounding atmosphere,
chrysotile decomposes to form forsterite and silica at temperatures in excess
of 500°C. However, finely sub-divided samples have been shown to dehydroxylate
at much lower temperatures (4). The decomposition of amphiboles is extremely
complex, and depends greatly on the surrounding atmosphere. Pyroxenes,
Cristobalite and iron oxide are frequent products of amphibole pyrolysis. Both
chemical and thermal decomposition of asbestoses, even when complete, may leave
decomposition products which retain the fibrous morphology of the original
sample. This may lead to complications when an analysis relying on morpho-
logical characteristics alone is used without precaution.
The widespread use of asbestos containing materials gives rise to concern
about exposure of the general population to low level concentrations of these
materials in air, water supplies and food. Although hazards associated with
-------
the inhalation of asbestos at high concentrations are well established, (5,6,7)
the significance of asbestos particles when ingested is not fully understood
(8). However, there is some evidence of increased incidence of tumors of the
gastro-intestinal tract where individuals have been exposed to the material
over a long period (5,9). Methods are therefore required by which trace
concentrations of asbestos minerals in water can be monitored.
Asbestos contamination can be reported in terms of either mass, or number
of fibers, per unit volume of liquid. If the data are reported as mass con-
centrations, using for example, X-ray diffraction or infrared spectroscopy,
no information is available about the dimensions of individual particles. On
the other hand, microscopy techniques, which report number concentrations and
number distributions, give detailed information about the size and shape
characteristics of individual fibers. Epidemiological studies indicate that
there appears to be a significant relationship between the degree of pene-
tration of the fibers into tissue and the physical dimensions or aerodynamic
diameters of the fibers (6). Similar size dependent phenomena have been
suggested for ingestion of asbestos fibers (6,10). It is for this reason that
fiber number concentrations, together with size measurements, should be
determined, rather than simple measurements of the mass of asbestos fibers.
The requirement that the number concentration be reported limits the analytical
techniques available to either light or electron microscopy. Optical micro-
scopy has for some years been used for industrial hygiene measurements of
airborne particles in asbestos plants and mining operations. In principle,
a membrane filter on which particles have been collected is examined by phase
contrast microscopy at a magnification of approximately 450X, and all fibers
longer than 5 micrometers are counted, regardless of their composition. Whilst
this method is adequate as a guide for routine monitoring of air in asbestos
processing plants, the environmental water sample presents an entirely
different problem. Firstly, in most water samples the fibers observed are
extremely small, and 5 micrometers would be a commonly encountered upper limit
of length. Secondly, water usually contains a large amount of diatomaceous
material, fragments of which can be mistaken for asbestos fibers. Thus for
the environmental water sample, methods of identification of asbestos fibers
are required.
The transmission electron microscope (TEM) has a resolution limit of the
order of 0.2 - 0.5 nm, which is more than adequate for the observation of even
the smallest fibers. The technique of selected area electron diffraction (SAED)
is also available on this instrument, permitting investigations of the crystal
structure of an individual fiber. More recently, it has become possible to
incorporate the energy dispersive X-ray spectrometer (EDS) in the instrument,
thus permitting information to be obtained about the chemical composition of
the individual fiber. The use of the scanning electron microscope (SEM) for
this purpose has also been suggested. Like the TEM, this instrument can also
be fitted with an energy dispersive X-ray spectrometer. A typical resolution
for a modern instrument would be about 10 nm. However, this resolution limit
generally applies to an ideal, electrically conducting specimen. For the
examination of asbestos fibers, non-conducting particles, with diameters down
to 20 nm, are collected on a polymer filter surface. The filter is then
vacuum coated with carbon or gold to prevent charging of the particles, and the
overall contrast between the smallest fibers and the background is poor.
-------
Additionally, the thickness of the gold or carbon coating can be of the order
of 10 - 20 nm, expanding the apparent size of the fibers. Under these con-
ditions the instrumental resolution limit is rarely achieved and the morphology
of the narrowest fibers is difficult to observe. These sample-related
resolution problems are largely associated with the analysis of chrysotile,
rather than the amphibole fiber, since amphibole fibers are usually present in
much larger sizes. A discussion of the applicability of the two instruments
has been given by Ruud (11), who concluded that fewer errors are made in the
identification procedure using the TEM-SAED combination than are made using the
SEM-EDS combination.
Assuming that the TEM-SAED technique is to be used, the analytical problem
is reduced to that of filtration of the water sample and transfer of a
representative fraction of the retained solid material to a carbon-coated
electron microscope grid. Various techniques by which this can be achieved
have been reported. The techniques can be classified according to whether it
is a direct preparation, or a preparation which incorporates an intermediate
ashing step to remove organic materials. A direct method using an ultra-
centrifuge was reported by Durham and Pang (12), in which an ultra-centrifuge
is used at an acceleration of about 30,000g to concentrate the solids from a
one liter sample of water. The solids are then redispersed in a small volume
of distilled water and an aliquot of this is transferred to a carbon-coated
electron microscope grid using an microliter pipette. This technique has few
advantages over filtration and has the disadvantage that the equipment is
expensive. Extensive cleaning of the centrifuge containers is also required to
ensure that there is no cross-contamination between samples. In a technique
by Kramer and Mudroch (13), the water is first filtered through a Millipore or
Nuclepore membrane filter, which is then given ultrasonic treatment whilst
immersed in a small volume of fiber-free water. The filter is then removed
and a drop of the remaining suspension applied using a microliter pipette
to a carbon or collodion coated electron microscope grid. Using this technique,
the efficiency of removal of fibers from the filter by the ultrasonic treat-
ment can be questioned and the uniformity of the deposit on the final electron
microscope preparation is also in doubt. A number of techniques are based on
the work of Jaffe (14) and Kalmus (15), in which the particulate material on
the surface of a membrane filter is directly transferred to a carbon-coated
electron microscope grid by solvent dissolution of the membrane filter. The
solvent dissolution can be achieved using either a Jaffe washer or a conden-
sation washer. In a technique reported by Cook (16), a Nuclepore filter is
used for the original water filtration, after which it is carbon-coated by
vacuum evaporation. The filter is then dissolved by the Jaffe washer technique,
leaving a carbon extraction replica of the original filter surface, in which
the particulate material is embedded. Techniques incorporating an ashing step
to remove organic materials have been reported by Biles and Emerson (17),
Cunningham and Pontefract (18), Chatfield (19) and Chatfield and Glass (20).
The requirement for the ashing step arises when there is significant content
of organic material in the water. Attempts to produce samples by any of the
direct filtration techniques may result in a large deposit of slime on the
surface of the filter which contains the inorganic particles. Alternatively,
the requirement for all particles to be separate and in a single layer on the
final microscope preparation may involve a dilution such that the detection
level for asbestos in the presence of this other material is totally unaccept-
-------
able. In these cases, oxidation of the organic content has to be achieved
prior to the preparation of the electron microscope sample. Biles and Emerson
(17) and Cunningham and Pontefract (18) used a muffle furnace, in which a
filter or a centrifugate was oxidized at high temperatures. More recently,
low temperature plasma ashing has been applied to this problem. The technique
of Chatfield and Glass involves initial filtration of the water sample through
a 0.1 pm Millipore membrane filter, which is then ashed in a low temperature
asher. The ash is redispersed ultrasonically in a small volume of double
distilled water, and this suspension is filtered through a 0.1 ym pore size
Nuclepore filter. This filter is then prepared for the electron microscope
by the carbon-coating replication process.
The consensus of opinions in the U.S.A. and the majority in Canada now
favor the filtration methods and direct transfer techniques for preparation of
electron microscope grids. In this report the results of an investigation of
the quantitative performance of these methods are presented. For those
samples containing large amounts of organic materials an alternate technique
incorporating an ashing step is required. The most promising technique from
those available appears to be that of Chatfield and Glass; this method was
investigated along with the direct transfer techniques. All of the analytical
methods referred to occasionally require the use of ultrasonic treatment. In
some cases, if a water sample has been allowed to stand for any length of time,
the particulate material, including the asbestos fibers if present, may have
coagulated and possibly also been deposited on the interior surface of the
container. In such cases ultrasonic treatment must be used to ensure repre-
sentative sampling from the container. In addition, all techniques which
incorporate an ashing step require the use of ultrasonic treatment in order to
redisperse the ash. There is little published work on the effect of ultra-
sonic treatment on dispersions of asbestos fibers, but work by Chatfield and
Glass (21) indicates that at low power densities the effects are minimal on
dispersions of chrysotile asbestos. Since ultrasonic treatment is an
essential step in some preparation techniques, the effects of such treatment
on the fiber number concentrations and their size distributions are investi-
gated in a quantitative manner.
-------
SECTION 2
CONCLUSIONS
The carbon-coated Nuclepore technique has essentially zero fiber losses,
and particles remain fixed in position during the Jaffe washing operation.
This technique gives a narrower spectrum of results for replicate analyses
than any of the Mi Hi pore direct dissolution methods.
The ORF ashing technique yields a spectrum of results and grid uniformity
similar to those of the carbon coated Nuclepore method. Any losses which
occur do so in the ashing step, and the subsequent preparation procedure has
essentially zero losses. Fiber losses were negligible for chrysotile and
taconite, but a mean loss of 41% occurred in the crocidolite preparations.
The reason for the crocidolite loss is not clear, but may be associated with
the surface properties of such small amphibole fibers. There is some evidence
in the crocidolite fiber length data to suggest a preferential loss of the
smaller fibers. In the case of the taconite dispersion, the fibers were
generally somewhat longer, and any such effects were not noticed. Both of the
Nuclepore techniques and the ashing technique are superior to the Millipore
preparations in that they have generally lower or even zero fiber losses, and
give a significantly more uniform fiber deposition on the final electron
microscope grids.
All three Millipore dissolution techniques introduce large preparation
losses, regardless of whether the Jaffe washer or the condensation washer is
used. Significant particle movement occurs during dissolution of the filter:
this particle movement leads to non-uniform deposition of fibers on the final
electron microscope grids. The losses when using the Jaffe washer were not
significantly different from those using the condensation washer for the 0.45
ym pore size Millipore filter.
There is less fiber loss when using the 0.22 ym pore size Millipore
filter than when the 0.45 ym filters are used in the Jaffe washer dissolution
process. This is possibly due to a shallower penetration of the particulate
into the 0.22 ym filter structure, and thus more efficient deposition of the
material on the carbon-coated grids. This conclusion is supported by the
observation of chrysotile contamination in the Millipore filters when analyzed
by the ashing technique, while such contamination was not detected by the
direct transfer Millipore techniques.
The fiber loss using the uncoated Nuclepore preparation technique was
significantly less than those when using the direct transfer Millipore
techniques.
-------
At the 5% significance level, an increase in the fiber concentration was
detectable when a chrysotile asbestos dispersion was exposed to ultrasonic
treatment. If an approximately linear effect with time is assumed, at 0.05
watt/ml the numerical increase in fiber concentration was about 0.1% per
second. At 0.5 watt/ml the corresponding increase was about 0.2% per second.
Even after treatment at 0.5 watt/ml, undispersed bundles of chrysotile were
still evident.
At the 5% significance level, no numerical increase in fiber concen-
tration was detected in the case of crocidolite. In neither case could a
systematic change in size distribution be detected.
-------
SECTION 3
REVIEW OF TECHNIQUES INVESTIGATED
3.1 CARBON COATED NUCLEPORE TECHNIQUE
The Nuclepore filter consists of a polycarbonate material which is
soluble in chloroform. This type of filter is unique, in that it consists of
a continuous, featureless plastic film, perforated by cylindrical holes of a
narrowly defined size range. Figure 1 shows a scanning electron micrograph
of a 0.1 ym pore size Nuclepore filter. It can be seen that the surface
structure of this filter would present no obstacles in the identification of
particles on its surface.
Using this technique, a small volume of the water in question is first
filtered, after which the filter is dried and carbon coated using a vacuum
evaporator. A small square of the coated filter is placed on a 200 mesh
copper electron microscope grid, and the filter dissolved away using chloro-
form. For the dissolution process, two designs of Jaffe washer are available.
Figure 2 shows the design of washer used by Kalmus (15), which has proved
satisfactory in many fields of application since 1954. It consists of a
supporting bridge, made from a rectangular strip of stainless steel wire mesh
bent sharply to form an inverted "U". The upper flat surface is covered
lengthwise with a paper strip cut from a Whatman filter of slightly smaller
width than that of the bridge. The end of the Whatman paper strip is bent
downwards so as to touch the floor of the petri dish in which it is placed.
Grids are placed in the position illustrated, on the top of which are placed
portions of the carbon-coated Nuclepore filter. The lid of the petri dish is
then placed in position and the assembly allowed to stand for periods of up to
two days, after which the plastic filter medium is completely dissolved,
leaving a thin carbon film containing the embedded particulate. The alter-
native design is shown in Figure 3, in which a pile of 4 or 5 glass microscope
slides is placed in the bottom of a petri dish. These are covered by two or
three Whatman filters in the position shown, which in turn are covered by a
flat strip of stainless steel mesh. The grids are placed on top of the
stainless steel mesh and the assembly is used generally in the same manner as
the other design.
3.2 DIRECT DISSOLUTION OF NUCLEPORE FILTERS ONTO CARBON-COATED GRIDS
In this technique, particulate from the surface of an uncoated Nuclepore
filter is deposited by direct transfer to a vacuum evaporated thin carbon
film, which is in turn supported by a 200 mesh copper electron microscope grid.
A number of techniques are available for preparation of the carbon-coated grids
-------
Fig. 1 Scanning Electron Micrograph
of 0.1 ym Pore Size Nuclepore
Filter Surface
(Cambridge Stereoscan S4, 20 kV,
gold coated)
-------
'6j.j
-------
Fig. 3 Alternative Design of Jaffe Washer
10
-------
required for this technique. A suitable and reliable method is to cast a
single drop of a 1 or 2% solution of parlodion in n-pentyl acetate onto a clean
water surface contained in a crystallizing dish. After the solvent has
evaporated, a thin film of nitrocellulose is left floating on the water
surface. A number of copper electron microscope grids are placed on the
surface of this floating plastic layer, and a Whatman filter paper is then
dropped onto the water surface. The Whatman filter paper may then be lifted
from the surface of the water, when it brings with it the plastic film and the
grids. This assembly is allowed to dry and carbon coated in the vacuum
evaporator. The copper grids, supporting a carbon coated nitrocellulose film,
can be picked from the surface of the paper. These grids are then placed in
a Jaffe washer, using n-pentyl acetate as the solvent. The nitrocellulose
layer is thus dissolved away from the grids, leaving only the carbon film.
This technique produces nearly perfect carbon-coated grids, having very few
areas in which the carbon film is broken. A large number of coated grids can
be prepared at one time, and they do not appear to deteriorate in storage.
3.3 JAFFE WASHER DISSOLUTION OF 0.45 ym & 0.22 ym PORE SIZE MILLIPORE FILTERS
Figure 4 shows an SEN! micrograph of a Millipore filter surface. In
contrast to the Nuclepore filter, there is much surface detail, and if repli-
cated by carbon evaporation, particles would be difficult to locate within it.
Using this dissolution technique, a small square of the Millipore filter is
placed, deposit side down, onto a carbon coated copper grid in a Jaffe washer,
using acetone as the solvent. Beaman and File considered that fiber losses
associated with Jaffe washer preparations were small (22) and in general less
than 10%. However, they used tetrahydrofuran as the solvent, rather than
acetone which is more commonly used with this technique.
3.4 CONDENSATION WASHER TECHNIQUE
This technique has been used by Beaman and File (22), McCrone and Stewart
(23) and McFarren and Millette (24). Figure 5 shows a condensation washer.
Essentially it consists of a heated flask containing solvent, vapor from which
is prevented from escaping by a vertical condenser. A side arm permits entry
of a water-cooled cold finger, on which the membrane filter and carbon-coated
grid assemblies are placed. Dissolution of the filter takes place over a
period of some hours by a reflux washing action. The original purpose of this
device was for the dissolution of plastic from carbon-coated plastic replicas
where loose particle movement is of no consequence. It has since been applied
to the preparation of particulate samples with varying degrees of success.
Opinions as to the proper operating conditions for the condensation washer
vary, but there is agreement that fiber losses will at least be minimised if
the washer is operated with the condensation level approximately at the
position of the grids. Rapid operation will result in violent condensation
action in the vicinity of the grids, and will inevitably result in fiber
losses. A brass holder device, designed by Millette, permits easier handling
of the carbon-coated grids and filter portions when assembling the washer.
It has also been suggested that the condensation washer with chloroform as
the solvent could be used to dissolve carbon coated Nuclepore filters. The
11
-------
Fig. 4 Scanning Electron Micrograph of
Mi Hi pore Filter Surface
(Cambridge Stereoscan S4, 20 kV,
gold coated)
12
-------
Fig. 5 Condensation Washer, Shown Inside
Protective Draft Shield
13
-------
Fig. 6 Detail of Condensation Washer Cold
Finger and Brass Specimen Grid Holder
14
-------
Fig. 7 Condensation Washer Cold Finger in Position
15
-------
feasibility of the latter approach was investigated in this program.
3.5 ONTARIO RESEARCH FOUNDATION ASHING TECHNIQUE
This technique was reported by Chatfield and Glass (20,21) as an optimum
technique for the analysis of water supplies having high concentrations of
organic material. A typical plasma asher used for this procedure is shown in
Figure 8. A beaker containing a 0.1 ym pore size Millipore filter, on which
the particulate material is deposited, is placed inside the chamber. The
chamber is first evacuated, and then a supply of oxygen is admitted through a
needle valve so that the chamber pressure rises to approximately 1 Torr. An
electrical discharge is maintained by the absorption of radio frequency energy
supplied from a coil which surrounds the chamber. At an oxygen pressure of
approximately 1 Torr and an indicated power of approximately 20 W, Millipore
filters can be completely ashed within 24 hours. Figure 9 shows beakers
being removed from a single chamber plasma asher. To obtain adequate re-
dispersal of the ash, it has been found necessary to use an ultrasonic probe,
rather than the bath which was reported previously. A suitable probe is shown
in Figure 10; this unit is capable of generating 40 or 50 W of ultrasonic
energy. After ashing, the residual material is redispersed in a small
volume of double-distilled water. This suspension is then filtered through
a 0.1 ym pore size Nuclepore filter, which is prepared for the electron micro-
scope using the carbon-coating procedure discussed earlier. The validity of
this preparation technique is, of course, contingent on a successful demon-
stration that the ultrasonic treatment used does not significantly change the
fiber dimensions or concentrations reported.
16
-------
Fig. 8 Low Temperature Plasma Asher (Two Chamber Type)
17
-------
Fig. 9 Ashed Samples in Beakers Being Removed from Single
Chamber Asher
18
-------
Fig. 10 Ultrasonic Probe for Redispersal of Ashed Residues
19
-------
Fig. 11 Positive Pressure Clean Room Required for Contamination Control During Specimen Preparation
-------
SECTION 4
EXPERIMENTAL APPROACH
The principal objectives are to establish the reproducibility and fiber
losses when using the various methods of electron microscope sample prepar-
ation. In order to achieve this, a large number of membrane filters, both
Millipore and Nuclepore, were loaded with identical concentrations of
asbestos fibers. For the Millipore filter, 0.22 ym and 0.45 ym pore sizes
were selected, whilst for the Nuclepore filter the 0.1 ym pore size was used.
Three types of asbestos fibers were investigated: UICC chrysotile, UICC
crocidolite, and a suspension of taconite tailings obtained from the Duluth
area of Minnesota. In order to eliminate as many experimental errors as
possible, all of the filtrations were carried out by one person, using
standardized techniques. Since the stability of the standard fiber suspen-
sions was questionable, the filtrations for each type of asbestos fiber were
all performed as rapidly as possible, shaking the stock suspension between
samples. In fact, with proper organization, it was possible to prepare each
series within a period of one day. All of the sample preparation was carried
out by one person. So far as scheduling would permit, fiber counting was
generally performed by the allocation of one material and one instrument to
one person; for example, all of the taconite samples were counted by a single
person using one instrument only. In this way, when attempting to compare
different preparation techniques for the same type of fiber, differences
between individuals in fiber counting philosophy were eliminated.
In order to control contamination, all sample preparation was performed
in an asbestos-free clean room, maintained at positive pressure by air
supplied through filters and electrostatic precipitators. Figure 11 shows the
clean room, in which can be seen a low temperature plasma asher, a multiple
filtering manifold, an ultrasonic probe and a glass water still which is used
to redistill the primary supply of distilled water. All glassware and equip-
ment used is washed using double distilled water. For the actual fiber
counting, chrysotile fibers were classified on the basis of morphology alone,
since this was the only material present. Crocidolite and taconite fibers
were categorized as confirmed amphibole, probable amphibole, and the sum of
these. Fibers placed into the "confirmed" category yielded selected area
electron diffraction patterns recognizable and similar to those from reference
amphibole fibers. Fibers giving either no pattern, or sufficient diffraction
spots for reliable identification, were placed into the "probable" category.
Initially, approximately 10 samples were processed for each type of asbestos
fiber, and each preparation technique. On the basis of the results, an
optimum technique was selected; an additional 10 filters for each fiber type
were then analyzed using this technique, in order to establish better statis-
tical validity. The technique which incorporates the ashing step required an
21
-------
investigation into the effects of ultrasonic treatment on the fibers. This
investigation was carried out for both chrysotile and crocidolite dispersions.
Effects of the ultrasonic treatment were anticipated to be most serious in
the case of chrysotile, and for this part of the work a natural water sample
was used, originating from a contaminated lake in Northern Ontario. This
sample was ideal for the purpose, since it contained many fiber bundles. In
the absence of any suitable natural specimen, the crocidolite sample was
prepared from UICC material, avoiding the use of ultrasonics in this
particular case.
22
-------
SECTION 5
PREPARATION OF STANDARD DISPERSIONS
Experience has shown that it is almost impossible to estimate the
final numerical concentration of a fiber dispersion prepared artifically from
the solid material, particularly if only the shorter fibers are required.
Furthermore, such dispersions are usually unstable and cannot be relied upon
for more than a day or so. Previous work (20,21) indicated that natural lake
water samples, spiked with artificial dispersions of chrysotile, crocidolite
and taconite tailings, remain stable for periods of some months. However,
later work in methodology investigation has shown that dispersions of these
materials in distilled water are not particularly stable, and that floccu-
lation occurs, giving rise to difficulty in obtaining a representative sample.
In order to prepare a large number of identically loaded filters, therefore,
it is necessary to prepare the dispersion and then to perform the filtrations
in a very short time.
For preparation of the chrysotile and crocidolite dispersions, approx-
imately 5 mg of the UICC fiber was placed into a clean agate pestle and
mortar; the material was then hand ground in about 1 ml of double distilled
water. This suspension was transferred to 1 liter of double-distilled water
in a 1.5 liter beaker, and stirred gently with a glass rod to break up the
aggregates. Caution was exercised in stirring, since grinding of a glass rod
on the inside of a beaker can generate a surprising amount of glass powder.
The dispersion was then exposed to ultrasonic treatment, using a probe, for
approximately 5 minutes at a power density of 0.05 Watt/ml, after which the
suspension was allowed to stand overnight. Using a large pipette, the top
500 to 700 ml was transferred to a 1 liter polyethylene bottle. The remainder,
including the settled material, was discarded. This dispersion formed the
concentrated stock suspension. A 10 ml sample of this stock was dispersed in
3 liters of double-distilled water, giving the final fiber dispersion. A
small volume of this was filtered onto a Nuclepore filter and examined by the
carbon-coating technique in order to establish the approximate fiber
concentration. A calculation was then made to determine how much of the
dispersion should be filtered so as to produce the required filter loading.
The target concentration was that equivalent to approximately 100 fibers in
ten grid openings of the final electron microscope sample. For chrysotlle,
it was found necessary to filter 20 ml and for crocidolite 15 ml. For the
work on Cummingtonite, a suspension of taconite tailings was obtained from
Dr. P.M. Cook of the National Water Quality Laboratory, Duluth, Minnesota.
This stock suspension was diluted 1:10, and 6 ml of this was dispersed in 3
liters of double-distilled water. The required filter loading was obtained by
the filtration of 10 ml of the resulting suspension. It is necessary to
emphasize that synthetic fiber dispersions of chrysotile and crocidolite, and
23
-------
probably taconite, are usually quite unstable and that the only reliable
method of obtaining a reproducible standard is by fixation of a known volume
of such a suspension on filters. On the other hand the work of Chatfield and
Glass (20,21) demonstrated the stability of some natural samples when spiked
with artificial fiber dispersions. In effect, a fiber dispersion in water may
be stable, but its stability appears to be a complex function of many variables
and should not be relied upon.
24
-------
SECTION 6
FILTRATION
Although no particular problem has been demonstrated when using the
Millipore type of membrane filter, filtration is undoubtedly the most critical
step in the Nuclepore preparation procedures. Filtration is performed using
commercially available 1 inch diameter assemblies, consisting of filter
funnels with a sintered glass frit support, and liquid reservoirs with vertical
sides. Using vertical sided reservoirs, the geometry is optimized so as to
minimize preferential deposition of particulate as a function of position on
the filter. Figure 13 shows an optical micrograph of a Nuclepore filter which
has been used to filter a concentrated dispersion of taconite tailings. It
can be seen that the deposit is extremely non-uniform, with some areas having
little or no deposit at all. It has been found experimentally that in the
absence of any precautions this is the type of deposit which may be expected
if a Nuclepore filter is used in one of these filtration assemblies. Figure
14 shows a scanning electron micrograph of such a taconite deposit on a
Nuclepore filter. It can be seen that significant areas, comparable with that
of a 3 mm diameter electron microscope grid, could be extracted from such a
filter by direct transfer techniques so as to give a totally unrepresentative
idea of the filter loading. A heavy deposit of this type permits the non-
uniformities to be recognized. However, at the filter loadings normally used
for electron microscope sample preparation, such irregularities would go
unnoticed and could easily lead to some of the unsatisfactory inter-laboratory
comparisons which have so far been conducted (25). This observation is
extremely worrying, since the direct transfer techniques of specimen prepar-
ation rely on absolute uniformity of particulate material deposit over the
active area of the filter. It was necessary to investigate this effect before
the standard series of filters could be prepared. Chatfield and Glass (20)
observed this type of non-uniform deposition when filtering redispersed ash
on to Nuclepore filters. Assuming the effect to be a property of the
suspension, a detergent was added and this technique appeared to yield a
uniform deposit. Later ethyl alcohol was also used satisfactorily as the
redispersal medium, together with an alcohol resistant Millipore as a backing
filter for the Nuclepore. However, the addition of detergent was considered
inadvisable when the Nuclepore was to be subsequently carbon-replicated, since
there is an increased possibility of detachment of particles from the carbon
replica. A technique for satisfactory filtration of water samples was there-
fore still required.
The principal origin of the non-uniformity can be envisioned by
inspection of Figure 15, in which is shown a scanning electron micrograph of
the sintered glass frit used to support the Nuclepore filter during the
filtration process. Areas of filter closely contacting the flat ground areas
25
-------
Fig. 12 Filtration Manifold in use for
Simultaneous Filtration of
Water Samples
26
-------
100
Fig. 13 Optical Micrograph of Nuclepore Filter Showing
Uneven Deposit of Particulate
Fig. 14 SEM Micrograph of Taconite Deposit on Nuclepore
Filter, Showing Uneven Deposit of Participate
(Cambridge Stereoscan S4, 20 kV, gold coated)
27
-------
100 pm
Fig. 15 SEM Micrograph of Glass Frit Filter Support, Showing
Flat Areas Which Restrict Filtration Rate Locally
(Cambridge Stereoscan S4, 20 kV, gold coated)
250
Unbacked
Backed
Fig. 16 Optical Micrograph of Nuclepore Filter with Deposit of
Taconite, Showing Difference Between Backed and Unbacked
Areas
28
-------
of this frit will permit very little filtration to occur, whereas the open
areas will permit efficient filtration. The solution to this problem lies in
the use of a backing filter. Figure 16 shows an optical micrograph of a
Nuclepore filter with, as before, a heavy deposit of taconite tailings, in
which only half of the Nuclepore filter was provided with a backing filter.
The comparison of the unbacked and backed areas is striking. However, even
when a backing filter is used, non-uniformities in the deposit are still
found and these appear to be related to the fact that the Nuclepore filter is
basically a hydrophobic material. The manufacturer applies a detergent to the
surface of the filter, in order to render it hydrophilic; this process, how-
ever, does not appear to be entirely satisfactory in some batches. The plasma
asher can be used to render the surfaces of many materials hydrophilic. Some
success has been obtained in achieving a more uniform particulate deposit by
a pretreatment of the Nuclepore filter in the plasma asher. The treatment
used was approximately 10 seconds in the asher at a very low power of about
10 watts. A significant improvement in the uniformity of deposit was achieved,
although some areas appeared to be resistant to this treatment. Figure 17
shows a comparison of treated and untreated filters, in their response to a
heavy deposit of taconite. It can be seen that the islands of sparse deposit
are not present on the treated filter.
It can now be seen that water filtration is not the simple topic it at
first appears, if a uniform deposit of material on the filter is required.
The problems can be largely overcome by bulk ordering of filters, specified
with separators of polypropylene, rather than the usual paper variety to
which release agents are sometimes applied. If a backing filter is also used,
problems of non-uniformity can be minimized. However, problems still
occasionally appear; some possibly caused by the filter clamping arrangements
on the commercially available equipment, and some by localized hydrophobic
areas on the filters themselves. Figure 18 shows an optical micrograph of a
highly colored particulate deposit on a Nuclepore filter suffering from this
particular problem. The effect is characterized by elongated oval shaped
areas, having no particulate deposit at all. The filters concerned appear to
be hydrophobic in these areas, since no filtration has occurred within them.
From the above discussion, it can be seen that filtration in the case of
the Nuclepore is an extremely critical step, and the following instructions
must be followed precisely if a uniform deposit is to be obtained. Initially,
all equipment must be dry; partial wetting of the backing filters leads to a
clogged situation which varies the filtration rate over the area of the
filter. A conventional water aspirator may be used as the vacuum source,
yielding a pressure of about 2kPa. Filtration rates are usually slow. The
backing filter to be used may be any medium or large pore size Mi Hi pore
filter; the 0.45 pm to 5 ym pore sizes have been used satisfactorily. The
backing filter is placed onto the glass frit support with the vacuum turned
on. The Nuclepore filter, shiny side up, is then placed on top of the
backing filter. The suction permits the filters to settle firmly onto their
support. If any folds appear in the Nuclepore filter it should be rejected
and replaced. The liquid reservoir should then be clamped in position,
keeping the vacuum on. The water sample should then be poured directly on to
the Nuclepore and allowed to filter. If the reservoir is not large enough to
contain the required volume, the additional liquid should be carefully added
29
-------
Treated Untreated
Fig. 17 Optical Micrograph of Nuclepore Filter with Deposit
of Taconite, Showing Comparison of Untreated and
Plasma Treated Filters
Fig. 1s Optical Micrograph of Nuclepore Filter with Deposit
of Particulate, Showing Hydrophobic Areas of Filter
which Permitted no Liquid to pass through
30
-------
well before completion of filtration. After filtration, the sides of the
funnel should not be rinsed. The Nuclepore filter should then be removed and
dried. Chatfield and Glass (20) reported use of a Polyvic backing filter, in
their ashing technique. This particular filter was selected because of its
resistance to ethyl alcohol, the redispersal medium for the ash. This later
work has shown that the non-uniformity problems experienced in the development
of their method were, in fact, a consequence of the filter properties, rather
than of the dispersion itself. A hydrophilic backing filter, such as a normal
cellulose ester membrane, is now preferred, together with water as the
redispersal medium.
31
-------
SECTION 7
SAMPLE PREPARATION AND ANALYSIS
For all of this work the design of Jaffe washer shown in Figure 3 was
used, although other work has shown that the washer described by Kalmus is
simpler and easier to use, and also appears not to affect the results obtained.
7.1 CARBON COATED NUCLEPORE PREPARATION
For this preparation technique a portion of the Nuclepore filter is
carbon-coated by vacuum evaporation. The distance between the filter and the
carbon arc should be about 10 cm. During the evaporation the sample is
rotated and tilted at various angles to ensure efficient coverage by carbon
of all features on the filter. It is most important that the Nuclepore filter
not be overheated; if the surface of the filter is exposed to excessive heat
during the evaporation process the filter plastic exhibits cross-linking and
becomes insoluble in chloroform. Such a filter gives an unsatisfactory
electron microscope sample, since there is always undissolved plastic remain-
ing after solvent extraction. The thickness of carbon to be used is found by
experience; if it is too thin the replica will be fragile and breakage will
occur, if it is too thick there will be lack of image contrast and inter-
ference with the diffraction process. Initially, problems were experienced
in handling the Nuclepore filters. Attempts were made to dissolve the
Nuclepore filters in the Jaffe washer, the carbon-coated side being placed
downwards in contact with the copper electron microscope grid. Exposure to
the chloroform vapor invariably caused the filter segments to curl up into a
scroll, leading to an unsatisfactory preparation. This effect could be
eliminated by placing a drop of chloroform from a microliter pipette directly
onto the surface of the filter before filling up the Jaffe washer with
chloroform. It was not possible to produce satisfactory preparations if the
carbon-coated side was placed downwards, unless the drop of chloroform was
added. However, a more satisfactory technique was found, in which the carbon-
coated Nuclepore was placed with the carbon face upwards on the copper grid.
Using this technique, the Jaffe washer is first filled with solvent. The
portion of filter is cut from the Nuclepore, using a scalpel blade, and placed
onto a copper electron microscope grid with the carbon facing upwards. The
entire assembly is picked up using a pair of fine tweezers and placed boldly
onto the Jaffe washer. Using this method, no curling occurs and almost a
100% success rate is obtained in the sample preparation. The minimum time for
filter dissolution is approximately seven hours, but more satisfactory
preparations are obtained if the assembly is allowed to stand for up to 48
hours.
32
-------
7.2 DIRECT DISSOLUTION OF UNCOATED NUCLEPORE ON CARBON-COATED GRIDS
In this preparation technique the uncoated Nuclepore filter is placed
with its active side down on to a carbon-coated electron microscope grid.
This assembly is placed in the Jaffe washer, prior to addition of the
chloroform. A small drop of chloroform was applied to the filter from a
microliter pipette, as in the case of the coated Nuclepore method, without
which it was not possible to produce grids with undamaged carbon films. The
preparation technique using the initial drop of chloroform appeared to be
satisfactory, and samples were produced which contained no confusing repli-
cation detail. Attempts were made to use the condensation washer for this
dissolution process, using chloroform as the solvent. After many experiments
it was found that the condensation washer was extremely unstable when operated
using chloroform as the solvent. In order to obtain condensation in the
vicinity of the grids, it was necessary to operate the washer fairly close to
the boiling point of chloroform. Unfortunately, explosive generation of
vapor would occasionally occur, leading to total loss of the samples. The
condensation washer technique using chloroform was therefore abandoned in
favor of the more reliable and simple Jaffe washer.
7.3 JAFFE WASHER DISSOLUTION OF MILLIPORE FILTERS
Before adding acetone to the Jaffe washer, carbon-coated grids were
placed on the wick, and portions of the Millipore to be processed were placed
on them, the active filter surface facing downwards in contact with the carbon
film. A 5 microliter drop of acetone was initially applied to each filter,
after which the washer was charged with acetone. The washer was best left
for a minimum period of 24 hours. Even after periods of 24 hours undissolved
filter medium remained on the samples.
7.4 CONDENSATION WASHER DISSOLUTION OF 0.45 ym PORE SIZE MILLIPORE FILTERS
A substantial amount of effort was expended in investigating this
technique. The condensation washer was operated under conditions as close to
those of the published interim method (26) as possible. The brass holder
designed by Millette (24) improved the ease of handling samples in this
washer. Carbon-coated grids were placed in the brass holder, and portions of
the Millipore filters to be processed were placed on them with the active
surfaces facing downwards. A 5-10 microliter droplet of acetone was applied
to each filter from a microliter pipette. The cold finger with the brass
holder on it was then inserted into the washer, which was then switched on.
Heating of the solvent reservoir was adjusted such that the condensation
level of solvent on the interior of the equipment appeared to be approximately
the level of the grids. Under these conditions washing action was very gentle
and even after a period of 6 to 12 hours undissolved plastic was still
observed. The degree of washing and the residual plastic found was comparable,
therefore, with the treatment given in the Jaffe washer. In order to obtain
stable operation of the condensation washer it was found necessary to control
both the temperature and the flow of the cooling water in the cold finger. In
addition, the washer required almost constant observation, since minor changes
33
-------
of water flow and temperature in the cold finger and also changes in supply
voltage could cause variation in its operating characteristics. A large
number of initial experiments were made to investigate the operation of the
washer before any samples were processed. Thermocouples, placed in the
vicinity of the samples, indicated that movement of the "condensation level"
was a very sensitive function of the heat input to the solvent flask.
However, the height of the condensation level on the interior of the glass-
ware can only be an indication of the solvent action on the grids themselves.
This will be controlled also by the cooling water temperature in the cold
finger, and there appears to be no simple way to monitor the condensation
rate on the grids, apart from counting drops as they fall back into the
reservoir flask.
7.5 ONTARIO RESEARCH FOUNDATION ASHING TECHNIQUE
Although the original method was developed using a 0.1 ym pore size
Millipore filter, in this particular program the 0.45 ym filters were used.
It has been found in previous work that if the Millipore is ashed directly,
some samples yield ash which is difficult to redisperse. Satisfactory
ashing behavior is usually obtained if the filter is first dissolved in
filtered acetone in a 50 ml beaker, one drop of distilled water added, and
the acetone then evaporated to dryness. A white deposit remains in the bottom
of the beaker. The deposit was ashed for a period of 24 hours under standard
conditions, i.e. at a pressure of 1 Torr of pure oxygen and an indicated
power of approximately 20 watts. About half way through the ashing cycle the
power was increased to 40 watts so that complete oxidation could be assured.
After ashing, 20 ml of water was added to the beaker, and the solid material
was dispersed by ultrasonic treatment at a power of 0.5 watt/ml for 30
seconds. This dispersion was then filtered through a 0.1 ym pore size
Nuclepore filter. The Nuclepore filter was then processed by the carbon-
coated Nuclepore technique. The initial group of filters processed indicated
that there was far more insoluble ash residues produced than had been observed
previously. The function of the ashing procedure is to remove solid materials
which interfere with the analysis; in fact this particular batch of Millipore
filters yielded more solid material than the ashing procedure was intended to
remove. The composition of the ash residue was investigated by energy
dispersive X-ray spectrometry, and found to consist principally of calcium,
sulphur and phosphorus. Discussions with Millipore Corporation revealed that
detergents are added to the surface of the Millipore filters to assist
filtration properties, and that the origin of the sulphur and phosphorus was
from these detergents. However, the origin of the calcium is a present
uncertain. What is certain is that the addition of water to this mixture of
chemical radicals will yield insoluble calcium phosphate and calcium sulphate.
To overcome this problem, the redispersal medium was changed from distilled
water to a 0.01% solution of the disodium salt of ethylenediamine tetraacetic
acid. This material complexes all free calcium, and inhibits the formation
of insoluble precipitates of calcium sulphate and calcium phosphate. The
procedure was found entirely satisfactory, and very little residual ash was
found on the Nuclepore filters when this redispersal technique was in use.
There was also no evidence of any adverse affect on the asbestos fibers con-
tained. However, as in the case of all other reagents used in the preparation
34
-------
of the samples for asbestos counting, the EDTA solution was filtered through a
0.1 ym Nuclepore filter prior to use. Examination of the Nuclepore filter
used for this reagent filtration indicated that there was significant con-
tamination of the crystalline EDTA salt by chrysotile asbestos.
7.6 FIBER COUNTING
For each sample several specimen grids were prepared. A total of 10 grid
openings from the available specimen grids were examined for the presence of
fibers. Grids were counted at a magnification of about 20,000 - 25,000, and
the actual grid opening dimensions were measured at a magnification of
approximately 2,000. Since the dimensions of the grid openings have been
observed to vary significantly within an individual grid, it is the practise
to count and measure all fibers within the grid opening, and then at a reduced
magnification to measure the dimensions of the grid opening so that the error
due to this effect can be eliminated. Wherever possible, fibers touching the
grid bars were counted on two sides of the opening only. In these cases their
lengths cannot be accurately stated. However, the grid loadings were
generally low, and such instances were rare. Fibers were placed into the
three categories previously discussed, i.e. confirmed amphibole, probable
amphibole, and chrysotile. Where fibers, particularly amphibole fibers, were
too large to identify by electron diffraction, a search was made around the
edge of the particle in order to find suitable areas which would give SAED
patterns. In most cases this could be achieved. The fiber counts obtained
were processed by a computer program which performed the relevant statistical
analyses and also computed the fiber number and mass concentrations and their
size distributions.
35
-------
SECTION 8
STATISTICAL TREATMENT OF RESULTS
8.1 UNIFORMITY OF DEPOSIT ON THE ELECTRON MICROSCOPE GRIDS
A check was made using the chi-squared test, to determine whether the
number of fibers found on individual grid squares were randomly and uniformly
distributed among the grid squares. If the total number of fibers found in
k grid squares is n, and the areas of the k individual grid squares are
designated A. to A. , then the total area examined
The fraction of the total area represented by the individual grid square area,
Pi , is given by Aj/A. If the fibers are randomly and uniformly dispersed over
the k grid squares counted, the expected number of fibers falling in the
region of one grid square with area A. = np.. If the observed number found in
that grid square is n.. , then
"j -
1=1 np1
This value is compared with the significance point of the x2 distribution,
having (k - 1) degrees of freedom. We may express our reluctance to discard
the idea that the deposit is uniform by establishing a very low value of a,
the significance level, and in this work a significance level a of 0.1% has
been used. The use of such a loose criterion for uniformity permits an
economic and realistic proportion of the samples to be considered acceptable.
8.2 THE BEST ESTIMATE AND CONFIDENCE INTERVAL OF THE FIBER CONCENTRATION
In the fiber analysis we wish to sample about 10 grid openings from the
population of grid openings and determine the mean grid opening fiber count
for the population on the basis of our sample. We also wish to determine the
interval about the sample mean, which, with a stated degree of confidence,
will contain the population mean. This is achieved by calculating the simple
arithmetic mean, followed by computation of a confidence interval using the
36
-------
Student "t" distribution. For the two-sided "t" test, n values of grid
square fiber count are used. The sample estimate of variance s2 is first
calculated, where
n (n - 1)
If the desired confidence is 100 (1 - a)%, for the two-sided interval the
value of t = t, /2 is obtained for (n - 1) degrees of freedom. For example
if the desired a/ confidence level is 95%, for the two-sided interval the
value of tQ Q7t- is obtained for (n - 1) degrees of_freedom. If the mean value
of fiber u-y/D concentration is calculated to be X, the upper and lower values
of the confidence interval are given by
v _ "\T i tS
and XL =X -
This confidence interval is the range of values within which, with a stated
degree of confidence, the mean value of all grid squares may be expected to
lie. It is important to recognize that the chi-squared test and the calcu-
lation of the confidence interval are not the same procedure. The chi-squared
test is the appropriate test to demonstrate that the fibers are randomly and
uniformly distributed on the grid squares selected. A very lossy preparation,
for example, which has lost all the fibers from the specimen grid except one,
all other grid squares containing no fibers, will give a very low value of
chi-squared. This is a statement that those fibers present are very uniformly
dispersed, i.e. a nearly constant zero. However, the 95% confidence interval
of such a preparation would be very large, indicating an imprecise result.
8.3 STATISTICAL SIGNIFICANCE OF LOSS MEASUREMENTS
Using this procedure a one sided test is used to determine if the mean
value for measurements by one technique significantly exceeds the mean value
for measurements using another technique.
Initially, the means_ and sample estimates of variance are calculated for
the two techniques. If X. and XB are the means for the two techniques and s2»
and s2B are the sample estimates of variance_ for n. and ng
measurements respectively, the estimated variances of X. and XR are
VA = s2A/nA and VB = s2B/nB. A
The effective number of degrees of freedom is f, where
37
-------
f - 2
"
(nA + 1) + (nB
If the significance level of the test is a, then the value of t/, _ \ is
obtained for f degrees of freedom, where f is the nearest * ~ a'
integer to f.
The value of u = tn v vfy. + VR is obtained^ and_this is compared _
with the difference in u " a) _ b the means (X. - Xj. If u >(X. - XB)
there is no reason to believe that X. exceeds Xg at the stated level of
significance.
8.4 VARIABILITY WITHIN TECHNIQUES
Within a particular technique, the spread of results obtained may be
different from the spread of results within another technique. The vari-
ability of a particular technique can be expressed using the relative standard
deviation, of coefficient of variation, i.e. the standard deviation of the
results divided by their mean value.
38
-------
SECTION 9
RESULTS
Tables A-l to A-23 in the Appendix summarize results obtained for all
samples analyzed. For each sample, the sequence number in the filtration
series is stated, followed by the mean fiber concentration, the 95% confidence
interval, and the mass concentration calculated by summation of the volume of
fibers. The grid distribution is expressed as a value of chi-squared,
together with the significance level at which grid uniformity could be
accepted. Table A-22 shows the chrysotile results as measured by the ashing
technique. Examination of the crocidolite and taconite results processed by
the ashing technique showed that the 0.45 ym Millipore filters were contamin-
ated by chrysotile. The chrysotile fiber counts obtained on the ashed
crocidolite samples are reported in Table A-23. The significance of this
filter contamination on the results is discussed later.
In Tables A-l to A-21, it can be seen immediately that the variability
of the results within a particular method is highest for the direct transfer
Millipore techniques. Figures 19-25 show the fiber length distributions for
the various categories of reported fibers. Examination of the size distri-
butions for total fibers indicates that there is very little distortion of
the size distributions by the individual methods of sample preparation. The
size distribution in each case is obtained by summing the individual distri-
butions for all samples within any one technique. Each size distribution
plotted, therefore, represents a minimum of 10 individual measurements.
Tables 1-5 show summaries of the experimental results. Table 1, for example,
shows the mean value of the fiber concentration for 26 separate preparations
of chrysotile filters by the carbon-coated Nuclepore technique. The mean
values of the technique results are given together with their standard
deviations. The data for the ORF ashing technique have been corrected for the
observed chrysotile contamination of the Millipore filters, using the mean
chrysotile count observed on the crocidolite filters as a means of establishing
the contamination level. As the best available expedient, the mean chrysotile
count obtained on these filters was subtracted from each individual chrysotile
result.
It should also be mentioned that the computed mass concentrations quoted
are presented as an indication of the reproducibility of this measurement.
Although the values obtained are not sufficiently accurate for their use in
loss calculations, it does appear that a useful estimate of mass concentration
can be made.
39
-------
Figure 19 Fiber Length Distribution, Chrysotile
100
tt ts 10
80 70 60 SO 40 3O
Fiber Length,
Micrometers
© 0.1 ym Nuclepore, Coated
A 0.45 ym Millipore, Jaffe
Xo.45 urn Millipore, Cond. Washer
00.22 pm Millipore, Jaffe
i 0.1 ym Nuclepore, Uncoated
Vo.45 ym Millipore, ORF Ashing
0.1
90 9!
Cumulative Percentage Number Less than Stated Length
-------
Figure 20 Fiber Length Distribution, Confirmed Crocldollte
70 6O 50 40 30 ZO
1 OS 0.2 0.1 0.05
i oo ^M
n
7
6
5
4
3
2
10 *-
9
g
5
3
z
3
8
T
6
5
4
3
2
- - -
. i . .. _
: --
P
»
-;-
--:
-'-
ibe
tier
ex
r Length,
ometers
--
3-
- ~
: rrt
;-:r:
^
.iT
\7::
^_Lli
i .::
\J?I
4f
:--- -
.4.; :.:
1
<^
i
i :
1 i.
1 i
1
... 1 i
i
'
-. ..
.
I
1 1 - /
i -
i
V^f l\
V "B
1
IS
^LlJ
"«M_i/I_J
j
. .j... ..
*?z^
f
fol
T$
r t
1 ' 1 !
-4- 1~ -
1 "~ ! j_
r^
i
l_
rv
X
^x
... .
I T I
: ! i
;
i
i
^ i
i
' " ;]
-
^
ra-
-
0 0
A o
Xo
Ho
-Ho
i
I.
k)
1 -
ri
LJ
-©
|\^>
G
.; : :
: -
.1 um Nuclepore, Coated
.45 wm Millipore, Jaffe
.45 ym Millipore, Cond. W
.22 ym Millipore, Jaffe
.1 ym Nuclepore, Uncoated
45 ym Millipore, ORF Ash
i
O.05 01 0 2 0.5
20 30
Cumulative Percentage Nunber Leva than Stated Length
-------
Figure 21. Fiber Length Distribution, Probable Crocidolite
ro
M.
100 ;-
a
7
6
5
4
3
2
1 ft ^
10 ,
8
7
6
5
4
3
2
U'
9
9
7
6
5
4
3
2
- r- .; -H f -- 4.- i <-i H4f -i-t H4 I - i-h^-
Fiber Length,
Micrometers
-- :. _r-:r-
L f
; ;f
..---.
i
t:::
:
;;;
::::
- ^
^
7
- .t i-
-
J :
.^f|
~"~
'-'-':':
:::
^
-
:
~
Jr- i
*- .;
- -~
_i
- ;
^i
^- -
yj^j
.;: :
-|-r- f'
T~
*
l.i:-
= .-.
: FT1
-f :..4:
"---: 5:11
t-fri
"T i.-a
^n;i
- fc
r
:r"-
-
i_ :
- r ;.
^-.-: ';
E- ' -
E-^iii
:.:Z-
-
- 1
: : = :
; I i
: : ::
*
^L-
-t
-
^
"-
H^*
*
.
£
::::
--f-
,
_,.
__
r r 1 ^ /
"v^nr£
I!:'
1
T-
jiX
/v Cj""7x>
ti l/*^
;
E|o
~r'
'
1 1 V
e-
:
. .
.
1
-
;
V'
T
*
)/
\
A4-
V-^e
.-_ . I
r':
J
4:
r S
x^
trf"^ 1
I
T
rrr
)
:
-'---
j
(.
(
\
?
r
....
i
OO.l vim Nuclepore, Coated
A 0.45 vim Millipore, Jaffe
Xo.45 Vim Millipore, Cond. W<
H0.22 vim Millipore, Jaffe
"TO.l vim Nuclepore, Uneoated
^0.45 ym Millipore, ORF Ashi
01
1
9
7
5
4
3
Z
4
3
I
1
ft
isher
ng
i
99
Cumulative Percentage Number Less than Stated Length
-------
Figure 22. Fiber Length Distribution, Total Crocidolite Fibers
100
.£»
CO
10 '-
1.0
0.1
- .T-4-1 --t :<---* !-r ff -t- 1- f-t-'-r-
Fiber Length,
Micrometers
: '
: :~
".::
_4_
^
:-
r ~T~
r:
7 777}
: i.r:
r -! ! !
"rr-H
vi:-
:-:;:
:.:::
1 : J L
^i
iVT
f : : :
1 . - .
" ~" 1
- : r
; _ ; -
' :- if^
-* -
^
-zfet^
V:
i i i
-
-4
- .
-^
_
--
- -
: *"
N;
^ **
""isy*
i
. .
i }..::
j i
^
i
-
\- ' .r
! ^
HjLr
5N
^X
*S£|
^
y
S
p
3
' 1
t \ m^L I
1 ^w^-i
i i 1
| ' 1
t
I
0
A
X
m
1
\
ff"
w*~
B
LJ
_
t?
n
&
t
«
5
4
3
2
1
9
«
7
«
S
4
3
2
0.1 ym Nuclepore, Coated
0.45 ym Millipore, Jaffe
0.45 vm Millipore, Cond. Washer
0.22 ym Millipore, Jaffe
0.1 ym Nuclepore, Uncoated
0.45 ym Millipore, ORF Ashing
0.05 0 1 0.2 0.5
Cuaulatlve Percentage Number Less than Stated Length
-------
100'
Figure 23 Fiber Length Distribution, Confirmed Taconite
9 96 9S 90 80 70 60 50 4O 30 20 1O
0 5 0.2 0.1 0.05
7
6
5
4
3
2
9
6
3
Z
.-
9
e
3
2
Fiber Length,
Micrometers
:
--- T-
- - - -
;;:-:
... ;
--
;
~:
rt^jr
: : ; ;
3i^!
-
sy
/\
A^
~ -;- -
,1^
^ '\e,
-..-
r I
1 j
W-:
^f
\
- /
:
\ 1 I
^ ^1
1
j
:
j
1
t
1
i
y
>^^\r:
*i ^
i
X
zzs
---
7?
4
1 1
PT
7-
_._
j . i
. i
....
- j
__^
_
!
i
-------
Figure 24 Fiber Length Distribution, Probable Taconite
01
100
10 -
1.0
0.1
80 70 60 50 4O 30
O 5 O.Z 0.1 0.05
Fiber Length,
Micrometers
-~f Hr
i":!
I'm
: ::: : :::
;;l;j : ::;
! ! i;Li 1 ;i
i!nFV ^
^Trifl
;;;; iil!
; i;;:;;
:.:!;:::
1 TT - ,
- . - ~- .
^^
I3|
."-1 :
: T
; :
M 1 i-
:
_4 o_ r :
I' 7-
*^ *
: T r ;
_
-;4
rilt
^
: :.~
* -*-f -
'. '~^~
:.r :
: : T ;
S?i
\
i
=trJ^
t+-
:.t:--
r++~
01 0.05 O.I 0.2 0.5 1 2 5 10
-ll-U-
Tii^r^:
: : ::
i! \j-
1; tt'
E3H"
am
;.; r;
r= r -:
Hll
E^ : :
m
l_/^
V
-M-
\
A
tttr
:4r
£$.(
nfvar-
. - -
.,.: Li.
T; : r
/
^ *
Tfre
T: : :
ltH~
\ -\
( sJ
fi
-rx
B-
-
*
zz
^3
n ^
O o.
Ao.
Xo.
Elo.
-ho.
Vo.
,
^
i
-&-
SH\B*/
?"t'""
i
-
rs
J
_
i
' " i
1 ym Nuclepore, Co
45 ym Millipore, J
45 ym Millipore, C
22 ym Millipore, J
1 ym Nuclepore, Un
45 ym Millipore, 0
T
ated
affe
ond. W<
affe
coated
RF Ashd
7
«
5
4
3
2
9
a
e
5
4
2
isher
Lng
»»
Cumulative Percentage Number Less than Stated Length
-------
Figure 25 Fiber Length Distribution, Total Taconite Fibers
100
70 <0 90 40 30
1 05 O.Z 0.1 0.05
10
1.0
0.1
9 I 1 ! ' H!h + H--' 1 i i ill ! Mil. ..' ._ =bi
7 :
Fiber Length,
Micrometers
4 .
3
2
7
6
5
4
3
2
9
B
S
4
3
2
_t : -
- - -
." .JiT. ; :
: :
i ':
: - : .
'-.'-.'I
-.'-::.
"~
- -
Ulj1
/Hi
f; ;:
....
;::r
: :i|
.....
.....
- r -r r-T-
....
::-
: :;:
""IT
- -
4iH
J v .: r
«!l! ':~
'."~Z~,
-Iri-
- ,-!:_
-..^--i..
:- r-
. : : ^
rr-'r:
*«r
- -
^rH-
^k: 1
- T : -
; : -I.
: :v
" r T' : ; : ;
i^^t
em;
alii 5=
^i=
. -j~ J--J- -^-
TTTT*-
: .. r .
4«i j
: r T i
^^ T~^-
tt te
i
*
(-
^~jr:
F-r:---
E^r
-.: ,rT-rt-
: :i : I :
;f - : : : -.
i»
-rrrr
-rrrr
:::v
TTT
*
*t=
i
rt-
-
r^
T^
sffi
^--
...
ytor7
r:.: :
T?TT
"*"T"*"T
«fr
n: :
£&
t- ^
r*
-----
§«
-f4-^--
^*!._J_
v/
*
j
-i
:-
-r :
X-
....
4TVi
r
!iA-
*L±
H-
rr
r
^~
.. . j...
i
r.ll
9
t
7
e
5
4
J
2
1
*
a
.1
s
4
3
2
O 0.1 vim Nuclepore, Coated
A 0.45 vim Millipore, Jaffe
Xo.45 vim Millipore, Cond. Washer
HO. 22 vim Millipore, Jaffe
r 0.1 vim Nuclepore, Uncoated
V 0.45 ym Millipore, ORF Ashing
0.03 0.1 0.2
Cumulative Percentage Nuaber Less than Stated Length
-------
TABLE 1
Summary of Experimental Results; Chrysotile
Numerical Concentrations
(Values In 106 Fibers/Liter)
Preparation
Technique
Carbon Coated
Nuclepore
Uncoated Nuclepore
Jaffe Wick
0.45lim Milllpore
Jaffe Wick
0.45vm Milllpore
Condensation Washer
0.22Wm Milllpore
Jaffe Wick
ORF Ashing
Technique
Number
of
Samples
Analysed
26
12
14
12
11
10
Concentration of Chrysotile
x 106 Fibers/liter
Mean
Value
23.42
18.52
A. 71
9.62
10.02
25.28
Standard
Deviation
8.58
7.34
3.66
5.79
3.32
9.67
Mass Concentration of
Chrysotile, ug/liter
Mean
Value
0.320
0.246
0.063
0.113
0.167
0.234
Standard
Deviation
0.227
0.144
0.057
0.069
0.137
0.072
-------
TABLE 2
Summary of Experimental Results: Crocidollte
Numerical Concentrations
(Values in 106 Fibers/Liter)
p.
00
Preparation
Technique
Carbon Coated
Nuclepore
Uncoated Nuclepore,
Jaffe Wick
0.45pm Millipore
Jaffe Wick
0.45pm Millipore
Condensation Washer
0.22Pm Millipore
Jaffe Wick
ORF Ashing
Technique
Number
of
Samples
Analysed
20
10
10
14
13
10
Confirmed Amphibole
Mean
Count
8.11
5.64
1.50
1.85
3.31
4.39
Standard
Deviation
2.88
2.42
0.96
1.60
2.49
0.95
Probable Amphibole
Mean
Count
5.81
3.74
1.37
2.13
3.45
3.85
Standard
Deviation
1.86
1.28
0.78
1.74
2.35
0.66
Total Fibers
Mean
Count
13.92
9.38
2.88
3.98
6.75
8.24
Standard
Deviation
3.24
3.48
1.59
3.17
4.70
1.36
Proportion
Yielding
Identified SAED
Patterns, %
58.3
60.1
52.1
46.5
49.0
53.3
-------
TABLE 3
Summary of Experimental Results: Crocldolite
Mass Concentrations
(Values in micrograms/llter)
<£>
Preparation
Technique
Carbon Coated
Nuclepore
Uncoated Nuclepore
Jaffe Wick
0.45Pm Millipore
Jaffe Wick
0.45pm Millipore
Condensation Washer
0.22pm Millipore
Jaffe Wick
ORF Ashing
Technique
Number
of
Samples
Analysed
20
10
10
U
13
10
Confirmed Amphibole
Mean
Mass
Cone.
1.294
0.626
0.192
0.410
0.331
0.761
Standard
Deviation
0.694
0.318
0.230
0.438
0.206
0.324
Probable Amphibole
Mean
Mass
Cone.
0.155
0.093
0.032
0.068
0.075
0.144
Standard
Deviation
0.121
0.085
0.023
0.104
0.065
0.055
Total Amphibole Mass
Mean
Mass
Cone.
1.449
0.719
0.224
0.478
0.406
0.904
Standard
Deviation
0.690
0.331
0.234
0.487
0.232
0.338
Proportion of
Mass which
Yielded SAED
Patterns, %
89.3
87.1
85.7
85.8
81.5
84.2
-------
TABLE 4
cn
O
Summary of Experimental Results; Taconite
Numerical Concentrations
(Values in 106 Fibers/Liter)
Preparation
Technique
Carbon Coated
Nuclepore
Uncoated Nuclepore
Jaffe Wick
0.45pm Millipore
Jaffe Wick
0.4 Sum Millipore
Condensation Washer
0.22Mm Millipore
Jaffe Wick
ORF Ashing
Technique
Number
of
Samples
Analysed
20
10
10
10
10
10
Confirmed Amphlbole
Mean
Count
16.26
25.36
5.24
5.45
7.65
22.34
Standard
Deviation
3.86
5.26
1.42
2.84
2.68
2.46
Probable Amphibole
Mean
Count
8.71
15.9
2.66
3.48
4.91
10.15
Standard
Deviation
3.11
1.52
1.11
1.42
1.94
2.70
Total Fibers
Mean
Count
24.96
40.75
7.90
8.92
12.56
32.49
Standard
Deviation
5.69
5.22
2.28
4.08
4.09
4.26
Proportion
Yielding
Identified SAED
Patterns, %
65.1
61.0
66.3
61.1
60.9
68.6
-------
TABLE 5
Summary of Experimental Results: Taconite
Mass Concentrations
(Values In mlcrograma/llter)
Preparation
Technique
Carbon Coated
Nuclepore
Uncoated Nuclepore
Jaffe Wick
0.45um Millipore
Jaffe Wick
0.45um Millipore
Condensation Washer
0.22pm Millipore
Jaffe Wick
ORF Ashing
Technique
Number
of
Samples
Analysed
20
10
10
10
10
10
Confirmed Amphibole
Mean
Mass
Cone.
10.29
10.22
3.82
5.21
8.26
14.35
Standard
Deviation
6.78
8.45
2.62
4.64
6.21
6.02
Probable Amphibole
Mean
Mass
Cone.
0.194
0.148
0.048
0.034
0.031
0.169
Standard
Deviation
0.162
0.048
0.039
0.013
0.015
0.064
Total Amphibole Mass
Mean
Mass
Cotic .
10.49
10.36
3.87
5.24
8.29
14.52
Standard
Deviation
6.83
8.48
2.63
4.63
6.21
6.02
Proportion of
Mass which
Yielded SAED
Patterns, %
98.1
98.1
98.7
99.4
99.6
98.8
-------
Figure 26 Effect of Fiber Length on SAED Identification
cn
ro
100
Percentage of Fibers
Displaying Satisfactory
SAED Pattern
100
Fiber Length, Micrometers
-------
9.1 CONTAMINATION OF 0.45 ym MILLIPORE FILTERS BY CHRYSOTILE
At this point some remarks about the significance of the chrysotile
contamination should be made. The normal practise when using the ORF ashing
technique is to remove silicate contamination, fibrous or otherwise, by
washing the Millipore filters in concentrated hydrofluoric acid prior to use.
On this occasion, this was not done because of possible deleterious effects
of the treatment on the solubility of the filters in acetone and consequent
interference with the direct dissolution methods. Previously reported (27)
chrysotile contamination of filters was sufficiently low that it would be of
little concern in this study. However, since the mean contamination level was
equivalent to a result of 12.72 x 106 fibers/1, it is interesting to speculate
on the reasons why it was not observed at these levels on the Mi Hi pore
dissolution methods. The reason would appear to be that the efficiency of
particle transfer to the grid during the dissolution methods is a function of
the depth of the fibers within the open filter structure. This conclusion is
supported by the fact that in Tables 1, 2 and 4 it can be seen that the 0.22
ym pore size Millipore gave a higher result on every occasion than the 0.45
ym Mi Hi pore, in which the intentional fibers would be more deeply buried in
its structure. Thus during the filter dissolution process, most of the
uniformly distributed contamination fibers in the depth of the filter were
probably lost, whereas they were efficiently collected by the ORF ashing
procedure to be deposited on the final replica.
9.2 IDENTIFICATION OF AMPHIBOLES BY SAED
In Table 2 it can be seen that the proportion of particles yielding
identified SAED patterns is remarkably constant between individual preparation
techniques, the actual proportion varying from 49% to 60%. On the other hand,
in Table 3 it can be seen that the majority of the mass is in the identified
category. The same conclusion can be drawn for taconite, in which the
identified numerical proportion is between 51% and 69% of the total. Figure
26 shows the success of identification by SAED as a function of fiber length,
for both crocidolite and taconite. It can be seen that the chance of positive
identification falls markedly for fibers below 1 ym in length, whereas above
about 4 ym to the maximum size studied, i.e. 32 ym, nearly all the fibers
fall within the identified category.
9.3 DISCUSSION OF FIBER LOSSES AND REPRODUCIBILITY
The principal results of the study are shown in Tables 6-8, in which
fiber losses and variability analyses are tabulated for the three asbestos
materials and all of the preparation methods. In these tables the carbon-
coated Nuclepore preparation is assumed to have negligible losses, and the
recoveries of the other methods are referred to it. For each preparation
method, the question of whether the declared loss is detectable at 5%
significance is asked. This test is necessary in view of the large standard
deviations of the results obtained by some methods. Durther evidence regard-
ing the negligible losses of the coated Nuclepore preparation is presented
later in this section.
53
-------
TABLE 6
Fiber Loss and Variability Analysis: Chrysotile
cn
Preparation
Technique
Carbon Coated
Muclepore
Dncoated Nuclepore
Jafff Hick
0.45um Mlllipore
Ja£f£ Hick
0,45um Mllllpore
Condensation Hasher
0.22po Mllllpore
Jaff £ Hick
ORF Ashing
Technique
Mean
Value,
x 106
Fibers/1
23.42
18.52
4.71
9.62
10.02
25.28
952 Confidence
Interval,
x 106
Fibers/liter
19.95 - 26.89
13.86 - 23.18
2.59 - 6.82
5.94 - 13.30
7.79 - 12.25
18.36 - 32.20
Number
of
Filters
Analyzed
26
12
14
12
11
10
Std.Dev.
of Mean,
x 106
Fibers/1
8.58
7.34
3.66
5.79
3.32
9.67
Loss * In
Preparation.
%
-
21
80
59
57
0
Is the
Numerical
loss
detectable at
5%
Significance?
-
Yes
Yes
Yes
Yes
No
Significance
Level
Required
to Reverse
Decision,
X
-
2.5
<0.5
<0.5
<0.5
30
Coeff. of
Variation
(Std. Dev.)
Mean
0.366
0.396
0.777
0.602
0.331
0.383
* Losses are referred to Coated Nuclepore Technique Values
-------
TABLE 7
Fiber Loss and Variability Analysis: Crocidolite
01
01
Preparation
Technique
Carbon Coated
Nuclepore
Uncoated Nuclepore
Jafffi Wick
0.45pm Milllpore
JaffS Wick
0.45um Hillipore
Condensation Washer
0.22um Milllpore
Jaffe" Wick
ORF Ashing
Technique
Mean
Value,
x 106
Fibers/1
13.92
9.38
2.88
3.98
6.75
8.24
95Z Confidence
Interval,
x 106
Fibers/liter
12.40 - 15.44
6.89 - 11.87
1.74 - 4.02
2.15 - 5.81
3.91 - 9.59
7.27 - 9.21
Number
of
Filters
Analyzed
20
10
10
14
13
10
Std.Dev.
of Mean,
x 106
Fibers/1
3.24
3.48
1.59
3.17
4.70
1.36
Loss * in
Preparation
%
1
33
79
71
52
41
Is the
Numerical
loss
detectable at
52
Significance?
-
Yes
Yes
Yes
Yes
Yes
Significance
Level
Required
to Reverse
Decision,
Z
-
<0.5
<0.5
<0.5
<0.5
<0.5
Coeff. of
Variation
(Std. Dev.)
Mean
0.233
0.371
0.552
0.796
0.696
0.165
Losses are referred to Coated Nuclepore Technique Values
-------
TABLE 8
Fiber Loss and Variability Analysis: Taconite
cn
cn
Preparation
Technique
Carbon Coated
Nuclepore
Uncoated Nuclepore
Jaffg Wick
0.45um Millipore
Jafffi Wick
0.45pm Millipore
Condensation Washer
0.22pm Millipore
Jaffe1 Wick
ORF Ashing
Technique
Mean
Value,
x 106
Fibers/1
24.96
40.75
7.90
8.92
12.56
32.49
95% Confidence
Interval,
x 106
Fibers/liter
22.30 - 27.62
36.80 - 44.68
6.27 - 9.53
6.00 - 11.84
9.63 - 15.49
29.44 - 35.54
Number
of
Filters
Analyzed
20
10
10
10
10
10
Std.Dev.
of Mean,
x 106
Fibers/1
5.69
5.22
2.28
4.08
4.09
4.26
Loss * in
Preparation
%
-
0+
68
64
50
0
Is the
Numerical
loss
detectable at
5%
Significance?
-
No
Yes
Yes
Yes
No
Significance
Level
Required
to Reverse
Decision,
Z
-
+
<0.5
<0.5
<0.5
+
Coeff. of
Variation
(Std. Dev.)
Mean
0.228
0.128
0.289
0.457
0.326
0.131
* Losses are referred to Coated Nuclepore Technique Values
+ Result Higher than that for Coated Nuclepore Technique
-------
9.3.1 Chrysotile
In Table 6 it can be seen that for chrysotile, all of the Millipore
dissolution techniques give large and statistically significant losses rela-
tive to the coated Nuclepore technique, although the losses of the uncoated
Nuclepore method were less significant and indeed become insignificant at
the 2.5% level. This is further evidence for the conclusion that the
efficiency of particle transfer is related to its initial distance in the
filter structure from the carbon film substrate. In the case of the
Nuclepore, all of the particulate is at its surface and efficient direct
transfer appears to take place. In the case of chrysotile, no statistically
significant fiber loss was observed when the samples were prepared by the ORF
ashing technique. The variability can be considered in terms of particle
movement during dissolution. The high values of this parameter, indicating
a wide spectrum of results, occur principally in connection with the Millipore
dissolution methods, although the 0.22 ym Millipore is an exception to this
rule. If it is recalled that the filters were all of identical, or at least
very similar fiber loadings, the variability should be constant if there is no
contribution to it from the preparation technique. It is interesting that
the variability of the coated Nuclepore technique using ashing (ORF ashing
technique), is very nearly equal to that for the direct coating technique,
indicating that the ashing step does not increase the range of results
obtained.
9.3.2 Crocidolite
Table 7 summarizes the results for crocidolite. As in the case of
chrysotile, the largest losses and the highest variabilities were displayed
by the Millipore direct transfer technique, notably those using the 0.45 ym
Millipore. The uncoated Nuclepore technique yielded a 33% mean loss, with
an increased variability. Although the ashing technique gave a 41% mean loss,
the variability of the results was smaller than that for any other technique,
reflecting the more complete dispersion of particles obtained when ultra-
sonic treatment is used.
9.3.3 Taconite
In Table 8 the results for taconite show generally the same pattern, in
that the high losses were observed for the Millipore direct dissolution
preparations. The variabilities were also, as before, highest for these
techniques. However, both the ORF ashing technique and the direct dissolution
Nuclepore technique yielded higher results on this occasion than the direct
carbon-coated preparation. The reason perhaps lies in more efficient
dispersal of fibers by these methods, either by the action of ultrasonic
energy in the one case, or chloroform during the dissolution step in the
other. Fibers which would normally remain aggregated thus become available
for inspection as individual fibers, increasing the numerical result. This is
confirmed by observations on specimens of taconite prepared by the carbon-
coated Nuclepore technique, which indicate that a significant proportion of
the material is overlooked during counting because of aggregation of fibers.
57
-------
TABLE 9
tri
00
Fiber Length Distribution Summary
Number Medium Lengths In Micrometres
Preparation
Technique
Carbon Coated
Nuclepore
Uncoated Nuclepore
Jaffe1 Wick
0.45ym Millipore
Jaffa" Wick
0.45pm Millipore
Condensation Hasher
0.22 urn Millipore
Jaffa" Wick
ORF Ashing
Technique
Chrysotile
0.72
0.75
0.68
0.76
0.66
0.80
Crocidolite
Confirmed by
SAED
0.87
0.88
0.90
0.96
0.79
0.95
Probable
0.43
0.70
0.63
0.51
0.48
0.63
Total
Fibers
0.66
0.70
0.63
0.66
0.58
0.79
Taconite
Confirmd by
SAED
1.2
1.2
1.6
1.7
1.6
1.2
Probable
0.92
1.0
0.90
0.80
0.72
0.75
Total
Fibers
1.1
1.1
1.3
1.3
1.1
1.1
-------
TABLE 10
Summary of Principal Results
cn
IO
Item
Is the Coated Nuclepore Preparation Mean Value
significantly higher than any of the Millipore
methods?
Are the losses using the 0.45pm Millipore and
condensation washer significantly less than
those using the 0.45pm Millipore and the
Jaffe wick?
Are the losses using the 0.22pm Millipore and
the Jaff£ wick significantly less than those
using the 0.45pm Millipore and Jaffe wick?
Are the losses using the O.lpm Nuclepore and
the Jaff6 wick significantly higher than
those using the Coated Nuclepore preparation?
Are the losses using the ORF ashing technique
significantly higher than those using the
direct Coated Nuclepore preparation?
Result at the 5%
Significance Level
Yes (For all 3 materials)
Chrysotile - Yes
Crocidolite - No
Taconite - No
Chrysotile - Yes
Crocidolite - Yes
Taconite - Yes
Chrysotile - Yes
Crocidolite - Yes
Taconite - No
Chrysotile - No
Crocidolite - Yes
Taconite - Ho
Significance Level Required
to Reverse Decision
<0.5
I
20
30
<0.5
0.5
<0.5
2.5
<0.5
*
3.0
<0.5
*
* Result obtained significantly higher than that of Coated Nuclepore Technique
-------
In many cases a group of fibers is so closely packed that none can be seen
as fibers and are not included in the count. Evidently when using the ORF
ashing technique, or the direct dissolution of Nuclepore filters, fibers are
spread more uniformly on the grids and a greater percentage of them are
available for inspection and checking by SAED. In a smaller way, this is an
illustration of an effect which has been observed, particularly with
chrysotile, where the ashing technique has reported a much higher value than
did the carbon-coated Nuclepore technique. Particularly if large amounts of
organic materials are present, fibers are observed to associate with the
organic material and few of them can be identified either visually or by SAED.
When the organic material is destroyed by ashing, the fibers become available
for dispersion and inspection in the final preparation. Experience has
shown that with organically dirty water containing slimes the direct carbon-
coated Nuclepore technique can be somewhat misleading.
9.4 SUMMARY OF TECHNIQUE CHARACTERISTICS
Table 9 shows a summary of the fiber median lengths for the various
preparation techniques. It can be seen that there is no significant shift
in the median length of total fibers as a function of preparation technique.
Table 10 summarizes the principal results of this study. The signifi-
cance of the reported loss is tested on each occasion at the 5% significance
level, and a significance level is also reported which would be required to
cause a reversal of the decision.
The carbon-coated Nuclepore technique gave significantly higher mean
values than any of the Millipore techniques. No difference between the
losses using the condensation washer, as opposed to the Jaffe washer, could
be seen, except for a marginal one in the case of chrysotile. The ORF ashing
technique did not introduce losses into preparations for chrysotile and
taconite, although a 41% loss was demonstrated for crocidolite. The reason
for this loss is not understood. A curious, but quite definite conclusion,
is that the losses involved in the Jaffe washer preparation for all three
fiber types are significantly less if a 0.22 ym pore size Millipore is used
than if the 0.45 ym Millipore is used.
The Millipore direct dissolution methods were characterized by large
fiber losses and large variabilities. Examination of the individual sample
data shown in Appendix A shows that particularly in the case of chrysotile,
use of the 0.45 ym Millipore with either the condensation washer or the
Jaffe washer is capable of giving practically any value from zero up to those
values yielded by the coated Nuclepore or ashing techniques. Therein lies,
perhaps, some explanation for the controversy surrounding the use of the
Millipore techniques. Even under these closely controlled operating
conditions using clean fiber dispersions, the inevitable conclusion is that
for chrysotile and crocidolite, which consisted largely or high aspect ratio
fibers, the Mi Hi pore techniques are unreproducible. They are rather more
consistent for the taconite. However, the mean losses for all three fiber
types are high, between 60-80%, relative to the coated Nuclepore technique.
A further indication of the extent of the losses when using these techniques
60
-------
TABLE 11
Summary of Grid Uniformity and Fiber Loss Measurements
Technique
Carbon Coated
O.lpm Nuclepore
O.lpm Nuclepore
Jaffe Wick
0.45pm Millipore
Jaffe Wick
0.45pm Millipore
Condensation Washer
0.22pm Millipore
Jaffe Wick
ORF/Ashing
Chrysotile
Percentage
of Samples
with uniform
Fiber
distribution
+
85
90
50
25
10
100
Fiber
Loss
*
-
21
80
59
57
0
Crocidolite
Percentage
of Samples
with uniform
Fiber
distribution
+
100
80
80
70
60
80
Fiber
Loss
*
-
33
79
71
52
41
Taconite
Percentage
of Samples
with uniform
Fiber
distribution
+
95
90
90
70
90
100
Fiber
Loss
*
-
0
68
64
50
0
* Fiber Loss calculated assuming Coated Nuclepore Value as 100%
-------
is given by the behavior of the chrysotile contamination in the Millipore
filters. It is significant that this contamination was not detectable by any
of the direct transfer preparation techniques; it only became apparent after
an ashing procedure. Blanks processed simultaneously were satisfactory. A
conclusion can be drawn that the direct transfer Millipore preparation
techniques do not efficiently transfer material which is buried deeply within
the filter structure. All of the 0.45 ym pore size Mi Hi pore filters used
were contaminated to the same degree, and yet this was not noticeable using
the direct transfer preparations. As discussed previously, it is logical to
suggest that when using a 0.45 ym pore size Millipore filter, the deposited
material penetrates more deeply into the structure than it does in the case of
the 0.22 ym filter. The efficiency of deposition by the direct transfer
technique appears to be related to its distance from the carbon film during
the dissolution process. This conclusion is supported both by the fact that
the fiber losses in the case of the 0.22 ym Millipore were less severe, and
also by the behavior of the 0.45 ym Millipore chrysotile contamination. It
is relevant to ask whether the direct transfer techniques would work even
better using the 0.1 ym Mi Hi pore filter.
Table 11 shows a summary of the grid uniformity and fiber loss measure-
ments for the six preparation techniques and the three types of asbestos
fibers. It can be seen that chrysotile is the most difficult material to
arrange in a uniform manner on the filter; in particular the 0.22 ym
Mi Hi pore filter displayed only 10% of the samples which were uniform at a
significance level of 0.1%, whereas 85% of the carbon-coated Nuclepore
samples were uniform at that significance level. In the case of crocidolite
and taconite the uniformity of all methods appears to be satisfactory. For
example, using crocidolite with the 0.45 ym Millipore Jaffe washer prepar-
ation, 80% of the samples had uniform distributions on the grids. However,
this is not to say that the preparation technique was satisfactory, since
nearly 80% of the fibers had been lost. We are merely saying that those
which remained were in fact relatively uniform in distribution. Thus in
Table 11 uniformity data and the fiber loss data must be read together. The
ORF ashing technique also has excellent uniformity of fiber distribution.
This is to be expected, since after the ashing step the technique is identical
with the carbon-coated Nuclepore method. The uncoated Nuclepore technique
also yielded satisfactory grid uniformity, indicating minimal particle move-
ment during preparation.
9.5 A DEMONSTRATION OF NEGLIGIBLE FIBER LOSS BY THE COATED NUCLEPORE
TECHNIQUE
In the preceding analysis is has been assumed that the losses associated
with the carbon-coated Nuclepore technique are effectively zero. Figures 27
to 32 demonstrate that this is indeed the case. Figures 27 and 28 are SEM
and TEM micrographs respectively of a Nuclepore filter which has been used to
filter a taconite tailings dispersion, shown at a magnification of 14,000.
Figure 27, the SEM micrograph, was obtained after carbon coating the filter
but before carrying out the washing operation in the Jaffe washer. Figure 28
shows a TEM micrograph of the same area of the filter replica after it had
been washed in the Jaffe washer. It can be seen that all of the particles
62
-------
The following illustrations are pairs of electron
micrographs presented facing each other for
comparison purposes.
63
-------
SEM Micrograph of Carbon Coated Nuclepore Filter, Showing
Taconite Deposit, Prior to Jaffe Washing
(Cambridge Stereoscan S4, 20 kV, carbon coating)
64
-------
1.0
Fig. 28 TEM Micrograph of Carbon Replica After Jaffe Washing, Showing
the Same Area of Deposit as in Fig. 27. Note the Correct
Relative Positions of the Particles, Indicating no Particle
Loss or Movement
(Philips EM301, 80 kV)
65
-------
Fig. 29 SEM Micrograph of Carbon Coated Nuclepore Filter with
Taconite Deposit, Prior to Jaffe Washing
(Cambridge Stereoscan S4, 20 kV, carbon coating)
66
-------
1.0
o
Fig. 30 TEM Micrograph of Carbon Replica After Jaffe Washing, Showing
the Same Area of Deposit as in Fig. 29. Note that Even the
Very Fine Fibers Visible in the TEM Image can be Traced on
the SEM Image
(Philips EM301, 80 kV)
67
-------
Fig. 31 SEM Micrograph of Carbon Coated Nuclepore Filter with
Chrysotile Deposit, Prior to Jaffe Washing. Note Very
Low Contrast of Chrysotile Fibrils
(Cambridge Stereoscan S4, 20 kV, carbon coating)
68
-------
1.0 ym
Fig. 32 TEM Micrograph of Carbon Replica After Jaffe Washing, Showing
the Same Area of Deposit as in Fig. 31. Note that all
Detectable Fibrils and Other Particles have been Retained in
Position Indicating No Particle Loss
(Philips EM301, 80 kV)
69
-------
present on the SEM image prior to the filter dissolution are also present in
the TEM image of the replica. There will however, be some small distortions
of the two images, since the SEM specimen was viewed at a tilt angle of 20
degrees; secondly the carbon replica is not particularly strong, and some
sagging of the replica occurs during the washing operation. Thus the pro-
jected image of this is somewhat distorted. However, it is an incontrovert-
ible fact that each particle discernible on the SEM image is also found on the
TEM image in the same relative position, indicating no particle movement or
loss. The same conclusion is drawn regarding Figures 29 and 30 at a
magnification of 10,000. Figures 31 and 32 are a similar pair of micrographs
using a chrysotile asbestos sample. Thus, as anticipated, the carbon coating
embeds the fibers and holds them in position during the dissolution operation,
whereas other techniques allow the particles to move around during the
washing operation. The non-uniformities of grid deposition observable in the
Millipore preparations are also indicative of particle motion during the
washing operation, since there is no reason to suspect the uniformity of
deposition on the Millipore filter. The use of the Millipore filter as a
backing filter to achieve better uniformity of deposition on the Nuclepore
filter indicates that the pressure distribution at the surface of the
Millipore during the filtration operation is relatively uniform, hence the
distribution of particles on the surface of the Millipore will also be
uniform. Any differences in the uniformity of fiber distribution in the
actual counting results would therefore indicate that one or other of the
techniques concerned had incurred particle motion during preparation.
It should be mentioned that for the SEM-TEM comparison, extreme care had
to be exercised in limiting electron dosage and heating deposited in the
surface of the Nuclepore filter during the SEM examination. The use of
normal beam currents in the SEM leads to plastic cross linking, and con-
sequently to insolubility of the upper layers in chloroform. The technique
was to use the minimum beam current possible, so that the image was noise-
limited. Even under these conditions, insoluble plastic was observed after
Jaffe washing, to an extent that examination in the TEM was impossible.
Eventually it was found possible to obtain SEM micrographs in which dosage
to the area concerned was limited to that given by one pass of the electron
beam. This, however, required that the micrographed area was never viewed
on the SEM screen.
70
-------
SECTION 10
EFFECTS OF ULTRASONIFICATION
Before conducting these experiments it was necessary to calibrate the
ultrasonic energy dissipation of the probe in use. The technique to
accomplish this is relatively simple. The ultrasonic probe is operated at a
known power setting with its tip immersed in a beaker containing a known
volume of water, and the rate of temperature rise is measured. This rate of
temperature rise is then compared with that produced by dissipation of a known
power from an electrical heater immerse'1 in the same volume of water. It is
thus possible to calibrate the ultrasonic energy deposition in the known
volume of water as a function of setting on the ultrasonic generator.
The effect of ultrasonic treatment at a frequency of 20 kHz was tested at
two values of power dissipation, 0.05 watt/ml and 0.5 watt/ml, using disper-
sions of chrysotile and crocidolite. The chrysotile dispersion was a natural
sample from a contaminated lake in Northern Ontario, and the crocidolite
dispersion was artificially prepared using UICC material. Approximately 100
ml of the dispersion was taken in a 250 ml beaker and ultrasonic treatment
was administered at a known power for a measured time interval. Two filters
were then prepared by the carbon-coated Nuclepore technique from the treated
dispersion, and the grids prepared from these two filters were counted by
two individual operators, A and B. The results obtained are shown in Tables
A-24 - A-27. In Table A-27 for the 15 second exposure time, operator A
reported that the grids obtained were very badly dispersed. A second pair of
grids were prepared from the original filter, but no significant improvement
was obtained. Thus this appears to be an example of a Nuclepore filtration
problem of some kind.
Tables 12 and 13 show the results computed in terms of fibers/g. These
values are obtained by dividing the reported mean fiber concentration by the
mass concentration. This normalizes the mean fiber concentration to a
constant mass, and compensates to some extent for the inevitable sampling
errors. It can be seen in Tables 12 and 13 that the only apparent increase in
fiber numbers per gram occurs in the case of chrysotile at the high power and
at 300 seconds exposure, although even here there is a discrepancy between
the results obtained by operator A and operator B. No consistent effect can
be seen in the case of crocidolite. Figures 33 and 34 show the changes in
fiber number concentrations, plotted graphically, for operators A and B,
together with the 95% confidence intervals. Although a consistent increase
can be seen for chrysotile, no such increase can be seen in the case of
crocidolite. Figures 35-42 show size distributions for the dispersions as
functions of power and exposure time to the ultrasonic treatment, in which no
consistent shift in the size distributions can be seen as a function of
71
-------
TABLE 12
Effects of Ultrasonification
Aqueous Chrysotile Dispersion, Fibers/g x 10
13
Exposure
Time,
Seconds
0
15
60
300
Power 0.05 watt/ml
Operator A
2.15
1.38
1.43
1.62
Operator B
2.29
1.94
1.87
2.09
Power »0.5 watt/ml
Operator A
2.22
1.42
1.22
2.17
Operator B
2.08
1.39
1.68
2.92
TABLE 13
Effects of Ultrasonification
Aqueous Crocidolite Dispersion, Fibers/g x 10
13
Exposure
Time,
Seconds
0
15
60
300
Power " 0.05 watt/ml
Operator A
0.76
0.71
0.65
0.59
Operator B
0.69
0.62
0.80
0.79
Power - 0.5 watt/ml
Operator A
0.70
* 0.98
1.04
0.62
Operator B
0.81
0.83
0.83
0.60
* Poor grid distribution reported by operator
72
-------
Figure 33 Effects of Ultrasoniflcation, Chrysotile
100
50-
Power - 0.5 watts/ml
0 Operator A 13 Operator B
Fiber Concentration
x 10 Fibers/liter
Power - 0.05 watts/ml
© Operator A Q Operator B
100
Exposure Time, Seconds (Square Root Scale)
73
-------
Figure 34 Effects of Ultrasonification, Crocldolite
150-r
iooa
300
Exposure Tlae, Seconds (Square loot Scale)
74
-------
Figure 35 Effects of Ultrasonification, Chrysotile Size Distributions
ui
100
» 98 »5 90 80 70 60 SO 40 30 20
2 1 05 0.2 0.1 O.OS 0.01
10
1.0
, -..--. |-4 -
r
Fiber Length,
Micrometers
; :- :-
- r
i :
i . . .
;;:;
i!::
-R
-77
~
-.
---
-
--:-
.....
*&
-----
:-::
--7-
-----
bx
' ~~
--r
- -
--
. " i -
i
-::
-
- -
;.i.
-
.
!
'
'
' j-'Vyvrv
"TA^jv
,
r
r~
3+X
A
*x
...-
-H--
>.
>J
i
9 4f*2i}
«->
--- --- --:---
i
i
... .. ._.,
_
r "
H:r
t
i-ff-
im
T-
t
#
-H-4-
u - - A
A fi
hX
_i^_-i
l~ : i ! j
"I
. . . . i
i
j
i
i
...i _i"
:
:
|
i ;
....
~~
....
r
^^
^T"
*
^
-T-J--
Hit
i
-
1 ;
....
..; ..: ..
: i
...
i '
_
Power = 0.05 watt/ml
Operator A
r" Zero time
A 15 seconds
-f" 60 seconds
X 300 seconds
i
:...... j j...
" 05 0 i 0 2 0.5 t 2
10 20 30 40 50 60
80 9O 95 96 99
Cumulative Percentage Number Less than Stated Length
-------
Figure 36 Effects of UltrasonifIcation. Chrysotile Size Distributions
2 1 05 O.Z O.t 0.05 0.01
±UU9"
8
6
5
4
3
Z
10 ,
8
7
6
5
4
3
Z
1.0' -
9
a
6
5
4
3
2
0.1' '
0
- -4 1 I- \ | p|- -t-{ - f -t - -t-J- - - i T f - - -i
f^«:t:,
Fiber Lengt
Micrometers
: "r.."-!
: : r:
.:. : r
- -^--4
:LL-3
:IE
-4
: 7
: .:
-
i.-^.-ff -
:h,
5
4
fci':
r-4-
4^
;-;::-
i
A
«"
^ . - .
- -
A
t',i_v
-^ 1
m. :
[
^y
, ^
I i j
, - - -
<
to
--I
I .
'
i
a&+-x
5*
mjfn -
: i
i > i
\
-
-
.....
i .
t t"
T ;
' : !
31 0.05 0.1 0.2 0.5 1 2 5 10 20 3
.. .-
i 1
i i
A/
2K
v^l
- ^
i^(:
^
r '
». '
|
' \
I
i
L ! !
r r
Power =0.5 watt/ml
Operator A
0 Zero time
A 15 seconds
"T~ 60 seconds
X 300 seconds
i i
i
j
Cumulative Percentage Number Less than Stated Length
-------
Figure 37 Effects of UltrasonifIcation, Chrysotile Size Distributions
vl
-J
100'
99 98 95 90
60 70 60 50 40 30 20
2 t 05 0.2 O.t 0.05 0.01
10
1.0'
0.1
T - - -+-
-14,4 '.)..... .,.:ii:u
- 1
Fiber Length,
Micrometers
~ ;-T-
~: -
H~L
-----
' i'-t-
!::
::::
~ -
~.
iHi
---
---
__..:
--
-f 4-
y
: :H
~r:
-rr1^
: H:
- i-
--
^
^₯^
------
-:-
-
-
---
-
~t ""
. ^
'.
'.
}-- -
j
1 A
c3
4
.Li
tJ2l^if
*§:
i -
.. i
T
i
...
' -
'
;
""
_...
..... _.-z
T
- .....U4(
A^
4T
f *
48
{_ .
h
-
- --
.....
!
i
i
5
pr
-
-
T
"^rA I
'"*
&
.._
^X
i^r
---
-I4-
--
----- - -
!
!
1 |
1 !
1
..j....
1
tr.
~
....
1-
Power = 0.05 watt/ml
Operator B
0 Zero time
A 15 seconds
~t~ 60 seconds
X 300 seconds
i
O.Ot 0.05 0.1 0.2 0.5 1 Z
90 95 98 99
Cumulative Percentage Number Less than Stated Length
-------
Figure 38 Effects of Ultraaonification. Chrysotile Size Distributions
00
100'
II < 9! 90
I 0 5 o.2 O.I O.O5
10 ;
1.0.
0* 1 001 0.05 0.1 02 0.5 1
90 95
- T.rr^li;j_'l^ -Tp:7t|r;-;| ; ; | r -
Fiber Length,
Micrometers
r.:-^T-T:
ni:
i fi
: : :
. *.-~
"Tm
j~
:E
::::
i :
~
f. ;
± t^:t
^H
;!?;
: ::.:
X ~^
-- :-- -
^
; fi-
: : : ;
...::_.
---
. .- 1- --
1
. J
1*
:-:
X'
'
-----
t
1 '
1
Lu
1
aL
1
H
'
'
ea
i
~t
4-3$
&
1
1
k
-:--
A
Aid
*ti y
QUA
ft/
fv
^3
^
?' '
^-
....
I
Power =0.5 watt/ml
Operator B
0 Zero time
^ 15 seconds
i 60 seconds
X 300 seconds
Cumulative Percentage Number Less than Stated Length
-------
Figure 39 Effects of Ultrasonification. Crocidolite Size Distributions
100-
99 98 95 90
0 5 0.2 0.1 0.05 0.0!
' '
Fiber Length,
Micrometers
;
;
-
w
'--
;
i
..
i
[ .. .
.
-,
-"
-
i !
.....
(:-,
; : ; : : i
:
,
- :
^ X ©
-
- - -
1
T "
*
i
L.._.. .
1
~~T
t
^ .. j
... -
--
..
I
"x
X|£
'A
- -4--
- - - -
'---r - "
... .., .
;_.4..._
I i
i
...
--
..
[
H-
t
-je A
--V
A
w
V*k
zr
;;
i.
^L
r
1
tirf
i
.
i
i
L
; i i
Power = 0.05 watt/ml
Operator A
0 Zero time
A 15 seconds
i 60 seconds
X 300 seconds
j i
i
10
1.0, -
0.1,
n 01 n.05 o 1 n ? o.s i 2
20 30 40 50 60 70 BO 90 95 98 9» 99.8 99.9
Cunulatlve Percentage Number Less than Stated Length
-------
Figure 40 Effects of Ultr-asonlfication, Croc-ldollte Size Distributions
00
O
100
80 70 6O 50 40 30 20
2 1 05 0.2 0.1 O.05 0.01
10
1.0
0.1
001 O.O5 0.1 0.2 0,5 I 2
fi
5
4
3
2
9
a
7
6
5
4
3
2
9
8
6
5
4
3
2
p--,-
- v* jr"
i : - :
- --
-f -
- ~ :-
: "!l ' 1 ' 1 : ' '\ ' ! '| ;
Fiber Length,
Micrometers
i ~
z '
_ ...
- --
_.:..
-
~-
h":
'-'
'--
-:
rrt.-
: :rr
- r
_r:4|
___
;,:
_
-- : --
- i" "
(
i- ;
,«*
_.
-
.
1 i
i
i;:T
i j
I j I ";
! i 1 '
i !
' ;
fa
...
-
r
i
-
v
-
;
"1
M
i
i
;
i
.
.
i
&&
-
;
f
\SA\
XP
4£
i
=
a i
i
x
7\
"
... .:.
I
-
i
.:..
Power =0.5 watt/ml
Operator A
G Zero time
A 15 seconds
"T" 60 seconds
X 300 seconds
1 i
Cumulative Percentage Number Less than Stated Length
-------
Figur* 41 Effects of Ultrasonification, Crocidolite Size Distributions
00
100!
1.0,
0.1
I- - -t--| | |t4- - Y--I-- j - j- 4 ~H-J ^ l |~H- *- +~ i~"p^~^^
-* I * j-H-»-4- ' T-* - |--*-' '-fr i~t~l~"t4- - - ' 4 '
h
: n
-'. - i
. - .
- '
'-. t
: \\
Fiber Length,
Micrometers
HH :^j[-
: ;^;
. .~
:/|"
-^-
~^
f
':.:=
'-: ': :
' "^
: £
~r':
-^-ii
-r- "
r
:-;;;
: : : ;
: : K
r-t : :
V
rf?
r i"1" .
T:
-r-
: : i ^
. : : : :
I : :.7
4-
;^;
$i
;
.-T4 :
^ft
T ;
: : : ; *
'K: i :
t"il
^
-
7--;-:
1
^
b&K
I
'1
:
Mff
Bfi
1
^1
^3f
&
A
i
H-
W
>
-------
Figure 42 Effects of Ultrasonificatlon. Crocidolite Size Distributions
«.;
7
6
3
«
3
2
10 .
9
8
6
4
00
ro 3
z
i n,
3
a
6
5
4
3
2
0.1,
0
h -.->-"
Fiber
Micron
: : :
- - --
--'---
-'-.
:"
--
__-
. . Y , ^..t . . . .j . .. ( . . .
Length,
Deters
1 :
-
-~
i I;!
- : Y
i :H
-~
.-_!
\-^W
--
\~ ^
-.,:.
..
- '--.
"-
17
.:-.;
:--^
' :
,
-~-
...
t
. ~
1
i i
i ,. , i- . ~
|-
1
--I--
ill . i : 1 j
] ;
i | |
." -,-.- . -,
! i
1
.
__
!
":-jg
\-. i . . t..-i.-.i_.-j_
..L . i . . 1 L._ ..4_
. _ ' L.. ^W _ J i .
1 X (>!% HK
i _L ! i^ j ^^ Ti
Jfl
lA
£^i
:-
J
/»J
IXx ; '
oK.
.
I
i
1
^. k^4 L. ._
*-
i ; i
.-!_ i
:::;
Operator B
j
tts/ml
-~-
~
. .
'
:
O Zero time
^ 15 seconds
"T~ 60 secc
i /( 300 secc
' i i i j
nds
nds
Cumulative Percentage Number Less than Stated Length
-------
exposure time or power. The summarized results of the effects of ultra-
sonification on the two dispersions are shown in Tables 14 and 15, in which a
statistical analysis is carried out at 5% significance. In Table 14 for
chrysotile, it can be seen that no significant effect was demonstrated at a
power of 0.05 watt/ml up to times of 60 seconds. At the higher power, all
periods of exposure demonstrated a statistically significant effect. However,
even at the high power of 0.5 watt/ml for an exposure time of up to 60
seconds, the increase in fiber numbers was only 35%, which is at the limits
of normal measurement accuracy for this kind of analysis. In Table 15 no
statistically significant effect was demonstrated for fiber concentration
change at 5% significance at any power or exposure time. In Tables 14 and 15
the observations of operators A and B are combined. Figures 43 and 44 show
these results plotted graphically, and expressed as percentages of the initial
value of fiber concentration. In Figure 43, it can be seen that if an
approximately linear effect were to be assumed, the increase in fiber
concentration as a function of time was approximately 0.1% per second at 0.05
watt/ml, whereas at 0.5 watt/ml the rate of increase was approximately 0.2%
per second. Consideration of Figure 44 indicates graphically that no
consistent effect on the fiber concentration was demonstrated as a function
of either power or exposure time for crocidolite.
83
-------
TABLE 14
00
Effects of Ultrasoniflcatlon
Chryaotlle: Comparison of Mean Grid Square Fiber Counts
(Combined Observations of Operators A and B)
Power
Density
watt/nl
0.05
0.50
Exposure
Time
Seconds
0
15
60
300
0
15
60
300
Mean Grid
Square Fiber ')
Count
17.61
15.18
19.83
24.29
14.68
19.87
19.69
24.90
Sample
Standard
Deviation^
7.29
5.94
9.58
12,72
7.07
8.12
7.96
6.58
Increase over
initial value
Z
-
-13.8
12.6
37.9
-
35.4
34.1
69.6
Detectable
Increase at
51 Significance
-
No
No
Yes
-
Yes
Yes
Yes
Approximate Significance
Level Required for
Change of Decision, %
-
20
30
2.5
-
1
1
<0.5
* Normalized to mean grid opening area
+ Total number of grid openings counted 20
-------
00
tJI
TABLE IS
Effects of Ultrasonificatlon
Crocidolite; Comparison of Mean Grid Square Fiber Counts
(Combined Observations of Operators A and B)
Power
Density
watt/ml
0.05
0.5
0.5+
Exposure
Time
Seconds
0
15
60
300
0
15
60
300
0
15
Mean Grid
Square Fiber
Count
12.28
13.95
14.89
14.35
16.63
17.31
20.25
18.00
14.70
17.70
Sample
Standard
Deviation
5.75
5.39
5.91
4.22
7.15
16.11(a)
7.88
7.24
6.88
7.89
Increase over
initial value
Z
-
13.6
21.3
16.9
-
4.1
21.8
8.2
20.4
Detectable
Increase at
5% Significance
-
No
No
No
-
No
No
No
No
Approximate Significance
Level Required for
Change of Decision, %
-
20
10
20
-
>40
10
30
40
* Normalized to mean grid opening area
+ Total number of grid openings counted - 20
+ Results for Operator A only, 10 grid openings
(a) Very poor grid distribution, additional result for
Operator A only is appended.
-------
Figure 43 Effects of Ultraaonification: Percentage Increase as a Function of Exposure Time - Chrysotlle
00
leor
150
100
Percentage Increase
over Initial Value
0.05 watt/ml
E) 0.5 watt/ml
60
Exposure Time, Seconds (Square Root Scale)
300
-------
Figure 44 Mfects of Ultrasonification; Percentage Increase as a Function of Exposure Time - Crocidolite
00
180
150
100
50
15
60
Exposure Time Seconds (Square Root Scale)
300
1
- rej
Inc
ove
In!
Val
1
: ~t~
! T-
' !
: i
! j
1
4-
1
i
i
1
i
..1
centagt
rease
r
tial
ue
'
i r .
_..
1
"j~
1
j
; t
i ,
. i
i
i
r
i
ft
--
1
I
4
i
! J
i
_,_
~^--
.
;
t
j
. t
1
i
~^-
i
.
f~
i
i
i
4-
l
i
_.
i-
:
I
i
j
1
1
1
1
i
j
i
!
!
i
_!
]
i
i
i
l
i .
i
i
i
i
\
\
,
V-
\
i
1
a
i
i
j
... .1 . .
_.
T
1
j
1
I
1
j
\
\
T
i
i
i
J
*
1
\
m
I
. ..{....
i
t
!
i
j
1
. l
i
I
i
i
j
O 0.05 watt/ml
Q 0.5 watt /ml
; i i
"" \"~r~
j
1
! '
j
-------
REFERENCES
1. Industrial Minerals and Rocks, Ed. J.L. Gilson, A.I.M.E., p48, 1960.
2. Speil, S., and Leineweber, J.O. Environmental Research, 2_, (3), 166-208,
1969.
3. Monkman, L.J. 2nd International Conference Phys. and Chem. of Asbestos
Minerals, Paper 3:2, Louvain (Belgium) 1971.
4. Berry, E.E. Paper 2:7, Ibid.
5. Selikoff, I.J., Nicholson, W.J., and Langer, A.M. Arch. Environ. Health,
25, 1, 1972.
6. Bogovski, P., Gilson, J.L., Timbrell, V., and Wagner, J.C. Proceedings
of the Conference on Biological Effects of Asbestos, Int. ASSA Res. Con.
Scientific Pub. No. 8, Lyon, 1972, WHO 1973.
7. Wagner, J.C., Sleggs, C.A., and Marchand, P. Brit. J. Ind. Med. Y7_> 260»
1960.
8. Stanton, M.F. J. Nat. Cancer Inst. Vol. 52, 633, 1974.
9. Mancuso, T.F., and Coulter, E.J. Arch. Environ. Health, 6., 210, 1963.
10. Pott, F., Huth, F., and Fredericks, K.H. Environ. Health Perspect., 9_,
313, 1974.
11. Ruud, C.O., Barrett, C.D., Russell, P.A., and Clark, R.L. Selected Area
Electron Diffraction and Energy Dispersive X-ray Analysis for the
Identification of Asbestos Fibers, a Comparison. Micron, ]_, 115, 1976.
12. Durham, R.W., and Pang, T.W.S. Asbestos Fibers in Lake Superior. Water
Quality Parameters, ASTM STP 573, 5, 1975.
13. Kramer, J.R., and Mudroch, 0. Asbestos Research at McMaster University,
Canadian Research and Development, Nov-Dec. 1975.
14. Jaffe, M.A. Proceedings, Electron Microscopy Society of America,
Toronto, Sept. 1948.
15. Kalmus, E.H. Preparation of Aerosols for Electron Microscopy, J. Appl.
Phys. 25, 87, 1954.
-------
16. Cook, P.M., Rubin, I.B., Maggiore, C.J., and Nicholson, W.J. X-ray
Diffraction and Electron Beam Analysis of Asbestiform Minerals in Lake
Superior Waters. Proc. In. Conf. Environ. Sensing and Assessment, IEEE,
New York, 2, 34-1-1, 1975.
17. Biles, B. and Emerson, T.R. Nature, 219, 93, 1968.
18. Cunningham, J.M., and Pontefract, R. Nature, 232. 332, 1971.
19. Chatfield, E.J. Quantitative Analysis of Asbestos Minerals in Air and
Water, Proceedings 32nd Ann. Proc. EMSA, St. Louis, Missouri, 528, 1974.
20. Chatfield, E.J., and Glass, R.W. Improved Methodology for Determination
of Asbestos as a Water Pollutant, Ontario Res. Found. Report, 1976.
21. Chatfield, E.J., and Glass, R.W. Analysis of Water Samples for Asbestos:
Sample Storage and Technique Development Studies, F.D.A. Conference,
Electron Microscopy of Microfibers, (In Press).
22. Beaman, D.R., and File, D.M. Quantitative Determination of Asbestos
Fiber Concentrations, Analyt. Chem. 48, 101 1976.
23. McCrone, W.C., and Stewart, I.M. American Laboratory, 13 1974.
24. Millette, J.R., and McFarren, E.F. EDS of Waterborne Asbestos Fibers
in TEM, SEM and STEM, Scanning Electron Microscopy 1976, (Part III),
IITRI, Chicago, 451, 1976.
35. An Inter-laboratory Study of Asbestiform Mineral Fiber Levels in the
Water Supply of Thunder Bay, Ontario. Ontario Ministry of the Environ-
ment Report, Sept. 1975.
26. Anderson, C.H., and Long, J.M. Preliminary Interim Procedure for
Fibrous Asbestos, U.S, Environmental Protection Agency, July 31, 1976.
27. Chatfield, E.J. Asbestos Background Levels in Three Filter Media Used
for Environmental Monitoring, 33rd Ann. Proc. EMSA, Las Vegas, Nevada,
1975.
89
-------
APPENDIX A
INDIVIDUAL SAMPLE RESULTS
90
-------
TABLE A-l
Carbon Coated Nuclepore Technique (Initial 10 Samples)
Chrysotile
Filter
No.
5
10
15
20
25
30
35
40
45
50
5 * +
10 * +
15 *
30 *
35 *
40 *
Mean Fiber
Concentration
x 106
36.9
46.4
32.1
33.2
33.3
20.5
18.0
17.0
14.5
16.5
30.9
16.2
25.7
21.6
10.6
10.7
95% Confidence
Interval
x 106
28.6 - 45.2
33.1 - 59.7
26.0 - 38.3
24.5 - 41.9
24.2 - 42.4
14.5 - 26.5
9.59 - 26.4
10.7 - 23.3
8.16 - 20.8
11.0 - 22.0
20.1 - 41.7
8.90 - 23.5
14.7 - 36.7
12.5 - 30.7
6.00 - 15.2
7.40 - 14.0
Mass
Concentration
Pg/1
0.284
0.214
0.185
0.201
0.667
0.127
0.110
0.136
0.144
0.140
0.360
0.953
0.899
0.402
0.475
0.418
Grid Distribution
X2
19.47
43.09
14.84
27.03
28.25
20.28
192.14
26.38
34.21
20.49
24.22
20.24
11.23
25.58
26.59
9.22
Significance
Level of
Uniformity %
1
<0.1
5
0.1
<0.1
1
<0.1
0.1
<0.1
1
0.1
0.5
2.5
0.1
0.1
10
* Repeat
Preparations
+ 9 grid square
count
+ 6 grid square
count
-------
TABLE A-2
Carbon Coated Nuclepore Technique (Second 10 Samples)
Chrysotile
Filter
No.
3
7
13
17
23
27
33
37
43
47
Mean Fiber
Concentration
x 106
30.0
24.1
25.6
24.1
27.4
21.1
22.8
15.6
17.3
16.9
95% Confidence
Interval
x 106
23.0 - 37.0
19.1 - 29.1
19.0 - 32.2
18.3 - 29.9
19.5 - 35.3
16.4 - 25.8
17.3 - 28.3
11.7 - 19.5
14.2 - 20.4
12.6 - 21.2
Mass
Concentration
Pg/1
0.466
0.298
0.286
0.314
0.418
0.166
0.187
0.080
0.164
0.225
Grid Distribution
X2
12.34
11.04
18.33
18.86
22.54
15.19
10.61
7.85
3.67
12.16
Significance
Level of
Uniformity %
10
10
2.5
2.5
0.5
5
10
10
10
10
-------
ID
co
TABLE A-3
Jaffe Wick Technique, O.lyim Nuclepore
Chrysotile
Filter
No.
5
10
15
20
25
30
35
40
45
50
30 *
50 *
Mean Fiber
Concentration
x 106
34.6
23.7
25.2
21.0
19.7
20.3
13.9
17.1
9.70
8.16
12.1
16.8
95* Confidence
Interval
x 106
28.1 - 41.1
17.1 - 30.3
14.5 - 35.9
16.3 - 25.7
15.4 - 24.0
14.1 - 26.5
8.70 - 19.1
12.1 - 22.1
7.05 - 12.4
4.10 - 12.2
8.90 - 15.3
11.5 - 22.1
Mass
Concentration
Pg/1
0.443
0.247
0.232
0.199
0.239
0.151
0.216
0.133
0.135
0.620
0.198
0.144
Grid Distribution
X2
21.23
16.39
44.82
12.43
12.33
19.40
23.67
9.74
8.06
21.31
6.21
17.77
Significance
Level of
Uniformity %
1
5
<0.1
10
10
1
0.1
10
10
1
10
2.5
* Repeat Preparations
-------
TABLE A-4
Jaffe Wick Technique, 0.45iJtn Millipore
Chrysotile
Filter
No.
5
10
15
20
25
30
35
40
45
50
30 *
35 *
40 *
50 *
Mean Fiber
Concentration
x 106
6.18
5.29
10.2
7.80
13.3
2.38
0.79
2.33
5.12
2.64
2.41
0.35
3.29
3.79
95% Confidence
Interval
x 106
0.88 - 11.5
0 - 13.8
7.60 - 12.8
0 - 18.1
3.60 - 23.0
0 - 5.65
0 - 1.88
0 - 6.07
1.86 - 8.38
0 - 5.40
0.55 - 4.27
0 - 0.85
0.48 - 6.10
0.87 - 6.71
Mass
Concentration
Vg/1
0.165
0.047
0.133
0.153
0.122
0.072
0.006
0.007
0.040
0.032
0.022
0.002
0.035
0.048
Grid Distribution
X2
52.03
188.13
12.20
137.37
77.26
64.15
26.47
60.02
21.73
37.53
17.58
8.56
22.82
15.96
Significance
Level of
Uniformity %
<0.1
<0.1
10
<0.1
<0.1
<0.1
0.1
<0.1
1
<0.1
2.5
10
0.5
5
* Repeat Preparations
-------
UD
tn
TABLE A-5
Condensation Washer Technique, 0.45tim Millipore
Chrysotile
Filter
No.
5
10
15
20
25
30
35
40
45
50
30 *
50 *
Mean Fiber
Concentration
x 106
8.05
10.6
21.6
8.80
17.9
0.50
10.0
12.7
5.71
4.85
5.01
9.69
95% Confidence
Interval
x 106
2.29 - 13.8
0 - 22.1
11.1 - 32.1
3.53 - 14.1
10.7 - 25.1
0 - 1.29
4.50 - 15.5
5.10 - 20.3
0.41 - 11.0
0.22 - 9.48
2.23 - 7.79
0 - 19.6
Mass
Concentration
Pg/1
0.084
0.089
0.273
0.099
0.159
0.009
0.140
0.105
0.183
0..041
0.080
0.099
Grid Distribution
x2
49.65
132.41
53.74
26.79
32.61
14.10
37.57
35.45
45.69
35.00
14.32
72.26
Significance
Level of
Uniformity %
<0.1
<0.1
<0.1
0.1
<0.1
10
<0.1
<0.1
<0.1
<0.1
10
<0.1
* Repeat Preparations
-------
01
TABLE A-6
Jaffe Wick Technique, 0.22pm Millipore
Chrysotile
Filter
No.
2
4
6
8
10
12
14
16
18
20
4 *
Mean Fiber
Concentration
x 106
12.5
5.26
16.1
8.87
9.02
10.1
4.99
13.9
8.88
9.93
10.7
95% Confidence
Interval
x 106
3.30 - 21.7
0.38 - 10.1
7.00 - 25.2
0.42 - 17.3
2.72 - 25.3
3.10 - 17.1
1.47 - 8.51
9.30 - 18.5
3.41 - 14.4
3.23 - 16.6
2.80 - 18.6
Mass
Concentration
Ug/1
0.081
0.125
0.131
0.123
0.074
0.065
0.102
0.124
0.475
0.399
0.142
Grid Distribution
X2
73.21
41.68
65.55
115.63
48.07
65.16
29.35
18.23
45.97
60.44
39.57
Significance
Level of
Uniformity %
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
2.5
<0.1
<0.1
<0.1
* Repeat Preparation
-------
TABLE A-7
ORF Ashing Technique: Chrysotlle *
ID
Filter
No.
3
7
13
17
23
27
33
37
A3
47
Mean Fiber
Concentration
x 106
11.28
23.68
16.48
38.18
23.18
17.78
33.28
28.98
29.48
30.48
95% Confidence
Interval
x 106
4.41 - 18.15
16.58 - 30.78
9.07 - 23.89
31.17 - 45.19
15.69 - 30.67
10.74 - 24.82
26.46 - 40.10
22.33 - 35.63
20.68 - 38.28
23.09 - 37.87
Mass
Concentration
yg/1
0.145
0.306
0.274
0.370
0.213
0.130
0.236
0.245
0.218
0.200
Grid Distribution
X2
16.04
10.20
13.35
5.98
12.97
9.94
6.73
5.60
11.06
10.93
Significance
Level of
Uniformity %
5
10
10
10
10
10
10
10
10
10
* Results corrected for contamination detected in Millipore polymer
-------
TABLE A-8
Carbon Coated Huclepore Technique (Initial 10 samples)
Crocidolite
CO
Filter
No.
5
10
15
20
25
30
35
40
45
50
Confirmed Amphibole Fibers
Mean
Cone.
x 106
14.4
7.38
12.7
6.84
8.72
8.72
8.23
11.0
11.0
10.9
95% Confidence
Interval
x 106
9.8 - 19.0
4.71 - 10.1
8.10 - 17.3
3.80 - 9.88
3.32 - 14.1
6.96 - 10.5
5.79 - 10.7
7.40 - 14.6
8.00 - 14.0-
6.90 - 14.9
Mass
Cone.
Ug/1
1.88
2.13
1.35
1.75
2.53
2.07
1.11
2.. 31
1.88
1.74
- Probable Amphlbole Fibers
Mean
Cone .
'x 106,*
6.43
7.38
6.37
7.26
7.06
4.46
3.82
4.33
3.68
5.24
95Z Confidence
Interval
x 106
3.68 - 9.18
3.66 - 11.1
3.73 - 9.01
3.99 - 10.5
3.90 - 10.2
2.27 - 6.65
2.22 - 5.42
2.89 - 5.77
1.41 - 5.95
2.80 - 7.68
Mass
Cone.
Wg/1
0.08
0.21
0.10
0.23
0.19
0.07
0.03
0.05
0.05
0.09
Total Cone.
Mean
Cone.
x 106
20.8
14.8
19.1
14.1
15.8
13.2
12.1
15.3
14.7
16.1
Mass
Cone.
yg/l
1.96
2.34
1.45
1.98
2.72
2.14
1.14
2.36
1.93
1.83
Grid Distribution
X2 (Total)
26.73
15.32
6.08
18.93
23.04
5.50
7.46
8.59
8.95
14.39
Significance
Level of
Uniformity Z
0.1
5
10
2.5
0.5
10
10
10
10
10
-------
TABLE A-9
Carbon Coated Nuclepore Technique (Second 10 Samples)
Croctdolite
UD
Filter
No.
3
7
13
17
23
27
33
37
43
47
Confirmed Amphibole Fibers
Mean
Cone.
x 106
4.98
4.60
5.34
6.30
6.96
2.55
6.49
9.03
8.78
7.32
95% Confidence
Interval
x 106
2.30 - 7.66
1.14 - 8.06
2.31 - 8.37
3.06 - 9.54
4.96 - 8.96
1.24 - 3.86
3.52 - 9.46
5.24 - 12.8
5.86 - 11.7
4.92 - 9.72
Mass
Cone.
yg/l
0.76
0.47
0.61
1.01
1.07
0.23
0.39
0.61
1.00
0.98
Probable Amphibole Fibers
Mean
Cone.
x 106
10.4
6.48
7.26
6.50
8.02
2.55
3.82
5.42
5.31
4.43
95% Confidence
Interval
x 106
6.90 - 13.9
2.86 - 10.1
3.98 - 10.5
2.89 - 10.1
3.55 - 12.5
0.82 - 4.28
1.96 - 5.68
3.01 - 7.83
2.76 - 7.86
1.96 - 6.90
Mass
Cone.
yg/l
0.32
0.24
0.15
0.14
0.36
0.04
0.06
0.10
0.48
0.10
Total Cone.
Mean
Cone.
x 106
15.4
11.1
12.6
12.8
15.0
5.1
10.3
14.5
14.1
11.8
Mass
Cone.
yg/l
1.08
0.71
0.76
1.15
1.43
0.27
0.45
0.71
1.48
1.08
Grid Distribution
X2 (Total)
12.86
17.61
25.45
17.71
8.44
16.03
12.09
13.71
9.39
12.32
Significance
Level of
Uniformity %
10
2.5
0.1
2.5
10
5
10
10
10
10
-------
TABLE A-10
Jaffc Washer Technique, O.lMm Nuclepore
Crocldollte
O
O
Filter
Ho.
5
10
15
20
25
30
35
40
45
50
Confirmed Amphibole Fibers
Mean
Cone.
x 106
9.45
7.53
4.75
1.86
6.36
2.40
7.23
4.10
7.50
5.19
95* Confidence
Interval
x 106
6.60 - 12.3
3.79 - 11.3
1.40 - 8.10
0.04 - 3.68
4.06 - 8.66
0 - 4.82
4.30 - 10.2
2.18 - 6.02
3.41 - 11.6
3.41 - 6.97
Mass
Cone.
Ug/1
0.969
0.615
0.255
0.319
0.517
0.449
1.190
0.301
0.727
0.915
Probable Amphibole Fibers
Mean
Cone.
x 106
4.82
5.73
4.18
1.66
5.05
2.40
3.62
3.92
3.48
2.51
95% Confidence
Interval
x 106
3.09 - 6.55
3.48 - 7.98
1.82 - 6.54
0 - 3.48
3.40 - 6.70
0 - 5.04
1.69 - 5.55
2.25 - 5.59
1.56 - 5.40
1.04 - 3.98
Mass
Cone.
WS/1
0.100
0.312
0.127
0.022
0.076
0.029
0.068
0.054
0.106
0.033
Total Cone.
Mean
Cone.
x 106
14.3
13.3
8.93
3.52
11.4
4.80
10.9
8.02
11.0
7.70
Mass
Cone.
wg/1
1.069
0.927
0.382
0.341
0.593
0.478
1.258
0.355
0.833
0.948
Grid Distribution
X2 (Total)
8.35
14.81
13.17
30.66
9.22
53.41
9.62
12.08
15.55
6.72
Significance
Level of
Uniformity Z
10
5
10
<0.1
10
<0.1
10
10
5
10
-------
TABLE A-ll
Jaffe Washer Technique, 0.45um Milllpore
Crocidolite
Filter
No.
5
10
15
20
25
30
35
40
45
50
Confirmed Amphibole Fibers
Mean
Cone.
x 106
2.82
2.94
1.38
0.22
1.46
0.44
1.79
0.61
2.32
1.05
95% Confidence
Interval
x 106
0 - 6.48
0.86 - 5.02
0.21 - 2.55
0 - 0.72
0 - 3.54
0 - 1.35
0 - 3.60
0 - 1.33
0 - 4.71
0 - 2.48
Mass
Cone.
yg/i
0.479
0.709
0.080
0.007
0.100
0.023
0.260
0.021
0.112
0.126
Probable Amphibole Fibers
Mean
Cone.
x 106
1.30
2.73
2.08
<0.22
1.46
0.65
0.90
1.02
1.68
1.90
95% Confidence
Interval
x 106
0.27 - 2.33
0.51 - 4.95
0 - 4.44
-
0 - 3.41
0 - 2.03
0 - 2.38
0 - 2.05
0 - 3.56
0.56 - 3.24
Mass
Cone.
W8/1
0.028
0.033
0.047
-
0.083
0.012
0.045
0.024
0.028
0.017
Total Cone.
Mean
Cone.
x 106
4.12
5.67
3.46
0.22
2.92
1.09
2.69
1.63
4.00
2.95
Mass
Cone.
Ug/1
0.507
0.742
0.127
0.007
0.183
0.035
0.305
0.045
0.140
0.143
Grid Distribution
X2 (Total)
35.76
23.90
18.08
8.29
38.02
22.71
21.81
14.43
15.19
14.68
Significance
Level of
Uniformity %
<0.1
0.1
2.5
10
<0.1
0.5
0.5
10
5
5
-------
TABLE A-12
Condensation Washer Technique, 0.45um Milllpore
Crocidolite
O
l\3
Filter
No.
5
10
15
20
25
30
35
40
45
50
20 *
25 *
30 *
45 *
Confirmed Amphlbole Fibers
Mean
Cone.
x 106
4.50
2.86
4.18
<0.19
0.58
0.40
1.53
1.37
0.41
3.87
3.53
1.47
0.55
0.58
95% Confidence
Interval
x 106
1.75 - 7.25
0.21 - 5.51
1.55 - 6.81
-
0 - 1.24
0 - 1.28
0 - 3.51
0.43 - 2.31
0 - 1.02
1.68 - 6.06
0 "- 7.08
0 - 3.05
0 - 1.20
0 - 1.43
Mass
Cone.
Mg/1
0.941
0.195
0.815
-
0.136
0.016
0.319
0.113
0.644
1.440
0.733
0.317
0.067
0.010
Probable Amphlbole Fibers
Mean
Cone.
x 106
4.30
4.40
5.23
0.19
0.97
0.20
3.83
2.94
0.41
2.37
2.23
0.92
1.28
0.58
95% Confidence
Interval
x 106
1.67 - 6.93
1.48 - 7.32
0.85 - 9.61
0 - 0.64
0 - 2.15
0 - 0.64
0.15 - 7.51
0.95 - 4.93
0 - 1.34
0.07 - 4.67
0 - 5.00
0.09 - 1.83
0 - 2.97
0 - 1.50
Mass
Cone.
ug/l
0.399
0.108
0.129
0.002
0.017
0.002
0.079
0.041
0.007
0.029
0.036
0.016
0.083
0.006
Total Cone.
Mean
Cone.
x 106
8.80
7.26
9.41
0.19
1.55
0.60
5.36
4.31
0.82
6.24
5.76
2.39
1.83
1.16
Mass
Cone.
Ug/1
1.340
0.303
0.944
0.002
0.153
0.018
0.398
0.154
0.651
1.469
0.769
0.333
0.150
0.016
Grid Distribution
X2 (Total)
14.74
24.17
40.51
51.14
12.91
14.64
63.90
12.92
11.10
19.93
51.14
12.91
14.64
19.41
Significance
Level of
Uniformity %
5
0.1
<0.1
<0.1
10
10
<0.1
10
10
1
<0.1
10
10
1
* Repeat Preparations
-------
TABLE A-13
Jaff6 Washer Technique. 0.22iJm Mlllipore
Crocidolite
O
CO
Filter
No.
2
it
6
8
10
12
14
16
18
20
6 *
16 *
18 *
Confirmed Amphlbole Fibers
Mean
Cone.
x 106
1.96
A. 48
0.68
1.13
4.18
6.95
6.40
1.58
0.66
7.92
1.93
3.45
1.67
95% Confidence
Interval
x 106
0.18 - 3.74
0 - ,9.02
0 - 1.75
0.21 - 2.05
1.38 - 6.98
3.53 - 10.4
3.55 - 9.25
0 - 3.55
0.15 - 1.16
0.46 - 15.4
0.64 - 3.22
0.31 - 6.59
0.51 - 2.83
Mass
Cone.
yg/i
0.766
0.204
0.237
0.096
0.397
0.364
0.441
0.188
0.101
0.660
0.172
0.426
0.251
Probable Amphibole Fibers
Mean
Cone.
x 106
2.40
2.77
1.14
2.82
5.28
8.18
6.19
1.58
2.14
6.38
0.86
3.85
1.25
95% Confidence
Interval
x 106
0.43 - 4.37
0 - 5.78
0 - 2.49
0.24 - 5.40
2.45 - 8.11
4.17 - 12.2
2.54 - 9.84
0 - 3.38
0 - 5.57
1.84 - 10.9
0.07 - 1.65
0 - 7.76
0 - 3.01
Mass
Cone.
yg/l
0.018
0.080
0.030
0.044
0.231
0.162
0.113
0.020
0.021
0.099
0.003
0.076
0.081
Total Cone.
Mean
Cone.
x 106
4.36
7.25
1.82
3.95
9.46
15.1
12.6
3.16
2.80
14.3
2.79
7.30
2.92
Mass
Cone.
Wg/1
0.784
0.284
0.267
0.140
0.628
0.526
0.554
0.208
0.122
0.759
0.175
0.502
0.332
Grid Distribution
X2 (Total)
42.65
40.33
7.29
15.78
9.13
15.06
25.74
32.93
23.46
64.24
7.29
32.93
23.46
Significance
Level of
Uniformity %
<0.1
<0.1
10
5
10
5
0.1
<0.1
0.5
<0.1
10
<0.1
0.5
* Repeat Preparations
-------
TABLE
ORF Ashine Technique
Crocidoljte
Filter
No,
3
7
13
17
23
27
33
37
43
47
Confirmed Amphibole Fibers
Mean
Cone.
x 106
5.16
3.81
2,77
5.61
5.87
4.09
3,67
4.60
3.90
4.44
95% Confidence
Interval
x 106
2.71 - 7.61
2.00 - 5.62
0.58 - 4.96
3.77 - 7.45
4.01 - 7.73
1.47 - 6,71
1.56 - 5.78
2.60-6,60
1,18 - 6.62
1.32 - 7,56
Mass
Cone,
Wg/1
0.654
1.36
0.263
0,777
0,818
1.12
0.395
0.560
0.901
0.761
Probable Amphlbole Fibers
Mean
Cone.
x 106
4.17
2.80
3,96
3,67
5,09
4.48
4,06
3,60
3.08
3-55
95% Confidence
Interval
x 106
1.54 - 6.80
0.95 - 4.65
0.37 - 7.55
1.91 - 5.43
2.45 - 7.73
3,19 - 5,77
1.98 - 6.14
1.29 - 5.91
0.37 - 5.79
1.55 - 5.55
Mass
Cone.
Ug/1
0.155
0,072
0.130
0.122
0.214
0.251
0.073
0.143
0.139
0.136
Total Cone.
Mean
Cone.
x 106
9,33
6.61
6.73
9.28
10.96
8.57
7.73
8.20
6.98
7.99
Mass
Cone.
Mg/1
0.809
1.43
0.393
0.899
1.03
1.37
0.468
0.703
1.04
0.897
Grid Distribution
X2 (Total)
11.58
14.08
29.02
10.43
9.97
7.05
10.59
11.45
30.96
11,45
Significance
Level of
Uniformity %
10
10
<0.1
10
10
10
10
10
<0.1
10
-------
TABLE A-15
O
en
Carbon Coated Nuclepore Technique (Initial 10 samples)
Taconite
Filter
No.
5
10
15
20
25
30
35
40
45
50
Confirmed Amphibole Fibers
Mean
Cone.
x 106
10.3
12.5
12.5
12.3
14.3
14.0
24.1
21.4
15.6
17.9
95* Confidence
Interval
x 106
5.70 - 14.9
8.10 - 16.9
8.60 - 16.4
7.60 - 17.0
6.60 - 22.0
10.1 - 17.9
11.3 - 36.9
15.3 - 27.5
11.9 - 19.3
11.8 - 24.0
Mass
Cone.
Pg/1
17.5
26.8
18.0
13.1
22.6
4.49
7.66
11.1
11.2
10.7
Probable Amphibole Fibers
Mean
Cone.
x 106
4.49
4.94
3.72
5.21
7.13
9.31
9.00
8.43
5.55
4.72
95% Confidence
Interval
x 106
1.67 - 7.31
1.55 - 8.33
0 - 7.48
3.02 - 7.40
5.32 - 8.94
5.97 - 12.7
2.80 - 15.2
5.33 - 11.5
2.21 - 8.89
2.57 - 6.87
Mass
Cone.
Pg/1
0.30
0.20
0.33
0.76
0.28
0.35
0.21
0.20
0.07
0.04
Total Cone.
Mean
Cone.
x 106
14.8
17.4
16.2
17.5
21.4
23.3
33.1
29.8
21.2
22.6
Mass
Cone.
Vg/1
17.8
27.0
18.3
13.9
22.9
4.84
7.87
11.3
11.3
10.7
Grid Distribution
X2 (Total)
6.25
14.24
12.69
7.99
9.35
5.83
31.86
14.42
5.55
7.14
Significance
Level of
Uniformity %
10
10
10
10
10
10
<0.1
10
10
10
-------
TABLE A-16
O
CTi
Carbon Coated Nuclepore Technique (Second 10 Samples)
Taconite
Filter
No.
3
7
13
17
23
27
33
37
43
47
Confirmed Amphibole Fibers
Mean
Cone.
x 106
15.1
17.4
23.7
17.8
16.7
11.9
16.7
13.2
18.0
19.7
95* Confidence
Interval
x 106
8.50 - 21.7
10.3 - 24.5
19.9 - 27.5
14.2 - 21.4
12.7 - 20.7
8.10 - 15.7
12.8 - 20.6
7.20 - 19.2
12.3 "- 23.7
11.8 - 27.6
Mass
Cone.
Pg/1
9.81
3.75
3.75
2.88
14.6
6.43
5.19
8.82
3.85
3.64
Probable Amphibole Fibers
Mean
Cone.
x 106
8.43
.12.4
10.2
9.52
10.8
12.5
12.7
12.1
13.4
9.55
95* Confidence
Interval
x 106
5.36 - 11.5
6.90 - 17.9
7.20 - 13.2
5.99 - 13.1
6.00 - 15.6
7.30 - 17.7
8.10 - 17.3
8.60 - 15.6
9.40 - 17.4
5.24 - 13,9
Mass
Cone.
US/1
0.07
0.16
0.08
0.08
0.10
0.18
0.12
0.12
0.15
0.08
Total Cone.
Mean
Cone.
x 106
23.5
29.8
33.9
27.3
27.5
24.4
29.4
25.3
31.4
29.3
Mass
Cone.
yg/l
9.88
3.91
3.83
2.96
14.7
6.61
5.31
8.94
4.00
3.72
Grid Distribution
X2 (Total)
11.18
11.38
11.34
13.35
16.14
8.23
4.18
15.42
3.76
27.64
Significance
Level of
Uniformity %
10
10
10
10
5
10
10
5
10
0.1
-------
TABLE A-17
Jaffe Washer Technique, 0.1 pro Nuclepore
Taconite
Filter
No.
5
10
15
20
25
30
35
40
45
50
Confirmed Amphibole Fibers
Mean
Cone.
x 106
14.7
27.9
30.3
31.0
26.3
19.4
29.3
22.1
28.3
24.3
95% Confidence
Interval
x 106
11.0 - 18.4
22.7 - 33.1
20.6 - 40.0
22.8 - 39.2
19.7 - 32.9
16.4 - 22.4
20.2 - 38.4
11.6 - 32.6
24.0 - 32.6
18.1 - 30.5
Mass
Cone.
yg/1
6.75
4.23
8.13
21.7
3.69
29.5
6.66
7.97
5.44
8.14
Probable Amphibole Fibers
Mean
Cone.
x 106
17.0
14.0
14.8
16.5
13.4
15.7
14.6
13.7
17.9
16.3
95% Confidence
Interval
x 106
10.2 - 23.8
7.90 - 20.1
10.1 - 19.5
10.1 - 22.9
8.40 - 18.4
9.50" - 21.9
10.2 - 19.0
7.80 - 19.6
13.4 - 22.4
7.90 - 24.7
Mass
Cone.
Ug/1
0.220
0.144
0.105
0.177
0.111
0.240
0.127
0.116
0.110
0.127
Total Cone.
Mean
Cone.
x 106
31.7
41.9
45.1
47.5
39.7
35.1
43.9
35.8
46.2
40.6
Mass
Cone.
ug/l
6.97
4.37
8.24
21.88
3.80
29.7
6.79
8.09
5.55
8.27
Grid Distribution
X2 (Total)
11.85
6.49
13.05
10.02
10.96
6.93
5.70
28.39
6.37
7.92
Significance
Level of
Uniformity %
10
10
10
10
10
10
10
<0.1
10
10
-------
TABLE A-18
Jaffe Washer Technique, 0.45um Mllllpore
Taconlte
O
00
Filter
No.
5
10
15
20
25
30
35
40
45
50
Confirmed Amphlbole Fibers
Mean
Cone.
x 106
7.74
5.57
5.49
5.14
3.22
6.74
6.16
3.58
3.96
4.77
95X Confidence
Interval
x 106
0.58 - 14.9
0 - 12.1
1.50 - 9.48
1.05 - 9.23
0.56 - 5.88
1.36 - 12.1
2.46 - 9.86
0.74 - 6.42
2.14 - 5.78
2.42 - 7.12
Mass
Cone.
Mg/1
8.86
2.45
2.96
1.29
1.42
7.55
3.59
3.23
5.32
1.55
Probable Amphibole Fibers
Mean
Cone.
x 106
4.15
4.11
1.16
3.02
1.47
2.64
2.93
1.93
1.52
3.71
95% Confidence
Interval
x 106
0.14 - 8.16
0.86 - 7.36
0 - 2.64
1.36 - 4.68
0.13 - 2.81
0.73 - 4.55
1.17 - 4.69
0.11 - 3.75
0 - 3.10
0.49 - 6.93
Mass
Cone.
Mg/1
0.04
0.06
0.01
0.07
0.01
0.06
0.02
0.02
0.14
0.05
Total Cone.
Mean
Cone.
x 106
11.9
9.68
6.65
8.16
4.69
9.38
9.09
5.51
5.48
8.48
Mass
Cone.
yg/i
8.90
2.51
2.97
1.36
1.43
7.61
3.61
3.25
5.46
1.60
Grid Distribution
X2 (Total)
19.90
32.17
12.46
7,89
15.14
22.50
15.50
15.11
7.20
15,66
Significance
Level of
Uniformity %
1
<0.1
10
10
5
0.5
5
5
10
5
-------
TABLE A-19
Condensation Washer Technique, 0.45wm Mlllipore
Taconlte
Filter
No.
5
10
15
20
25
30
35
40
45
50
Confirmed Amphibole Fibers
Mean
Cone.
x 106
8.39
9.68
1.87
3.91
2.98
7.04
3.46
2.87
8.87
5.39
95Z Confidence
Interval
x 106
3.49 - 13.3
5.59 - 13.8
0 - 4.05
0 - 8.41
0.46 - 5.50
0 - 14.4
0.31 - 6.61
0.58 - 5.16
5.91 - 11.8
1.35 - 9.43
Mass
Cone.
Pg/1
12.9
4.15
2.41
3.18
0.28
8.53
0.72
0.89
11.8
7.23
Probable Amphibole Fibers
Mean
Cone.
x 106
4.84
6.08
2.80
3.91
1.99
3.67
2.13
1.43
3.67
4.25
952 Confidence
Interval
x 106
2.48 - 7.20
3.58 - 8.58
0.95 - 4.65
1.33 - 6.49
0 - 4.01
0.35 - 6.99
0.44 - 3.82
0.03 - 2.83
1.01 - 6.33
1.02 - 7.48
Mass
Cone.
Pg/1
0.03
0.05
0.03
0.03
0.02
0.03
0.04
0.02
0.03
0.06
Total Cone.
Mean
Cone.
x 106
13.2
15.8
4.67
7.82
4.97
10.7
5.59
4.30
12.5
9.64
Mass
Cone.
Pg/1
12.9
4.20
2.44
3.21
0.30
8.56
0.76
0.91
11.8
7.29
Grid Distribution
X2 (Total)
10.11
6.94
17". 98
35.29
15.20
46.96
29.68
9.50
7.23
15.19
Significance
Level of
Uniformity %
10
10
2.5
<0.1
5
<0.1
<0.1
10
10
5
-------
TABLE A-20
Jaffe Washer Technique. 0.22pm Millipore
Taconite
Filter
Mo.
2
4
6
8
10
12
14
16
18
20
Confirmed Amphibole Fibers
Mean
Cone.
x 106
4.73
11.6
6.37
11.2
10.3
4.36
5.62
9.00
7.15
6.18
95% Confidence
Interval
x 106
1.41 - 8.05
7.00 - 16.2
3.60 - 9.14
5.10 - 17.3
5.10 - 15.5
1.63 - 7.09
2.60 - 8.64
5.73 - 12.3
2.28 - 12.0
1.73 - 10.6
Mass
Cone.
yg/l
20.0
5.83
6.29
14.1
5.49
5.79
2.05
16.0
5.61
1.45
Probable Amphibole Fibers
Mean
Cone.
x 106
3.68
4.49
3.76
5.32
9.51
2.18
4.02
4.91
5.30
5.90
95% Confidence
Interval
x 106
0.98 - 6.38
1.69 - 7.29
1.61 - 5.91
2.95 - 7.69
5.07 - 14.0
0.11 - 4.25
2.18 - 5.86
2.12 - 7.70
2.39 - 8.21
0 - 12.7
Mass
Cone.
U8/1
0.02
0.05
0.02
0.02
0.05
0.01
0.02
0.05
0.04
0.03
Total Cone.
Mean
Cone.
x 106
8.41
16.1
10.1
16.5
19.8
6.54
9.64
13.9
12.5
12.1
Mass
Cone.
Hg/1
20.0
5.88
6.31
14.1
5.54
5.80
2.07
16.1
5.65
1.48
Grid Distribution
X2 (Total)
22.88
15.57
7.01
17.85
13.59
24.86
9.53
7.93
12.02
42.01
Significance
Level of
Uniformity %
0.5
5
10
2.5
10
0.1
10
10
10
<0.1
-------
TABLE A-21
ORF Ashing Technique
Taconite
Filter
No.
2
8
12
18
22
28
32
38
42
48
Confirmed Amphibole Fibers
Mean
Cone.
x 106
22.6
21.3
26.3
23.7
22.6
20.2
25.0
23.6
18.5
19.6
95% Confidence
Interval
x 10s
15.8 - 29.4
11.2 - 31.4
20.3 - 32.3
18.8 - 28.6
17.9 - 27.3
14.2 - 26.2
13.5 - 36.5
14.0 - 33.2
13.1 - 23.9
14.0 - 25.2
Mass
Cone.
Ug/1
9.66
8.76
26.6
10.6
18.3
13.6
17.8
6.00
14.6
17.6
Probable Amphibole Fibers
Mean
Cone.
x 106
12.1
4.11
12.6
13.1
10.2
11.8
10.5
8.55
10.5
7.70
95% Confidence
Interval
x 106
6.20 - 18.0
0.79 - 7.43
7.90 - 17.3
9.30 - 16.9
7.00 - 13.4
8.20 - 15.4
6.20 - 14.8
4.78 - 12.9
5.30 - 15.7
4.59 - 10.8
Mass
Cone.
Ug/1
0.23
0.06
0.15
0.23
0.13
0.27
0.17
0.19
0.16
0.10
Total Cone.
Mean
Cone.
x 106
34.7
25.4
38.9
36.8
32.8
32.0
35.5
32.5
29.0
27.3
Mass
Cone.
Ug/1
9.89
8.82
26.8
10.8
18.4
13.9
18.0
6.19
14.8
17.7
Grid Distribution
X2 (Total)
15.00
12.71
5.33
7.13
2.77
4.05
16.17
12.11
13.05
8.71
Significance
Level of
Uniformity %
5
10
10
10
95
10
5
10
10
10
-------
TABLE A-22
ORF Ashing Technique: Chrysotlle *
Filter
No.
3
7
13
17
23
27
33
37
A3
47
Mean Fiber
Concentration
x 10s
24.0
36.4
29.2
50.9
35.9
30.5
46.0
41.7
42.2
43.2
95% Confidence
Interval
x 106
18.2 - 29.8
30.1 - 42.7
22.7 - 35.7
44.8 - 57.0
29.3 - 42.5
24.4 - 36.6
40.3 - 51.7
36.2 - 47.2
33.9 - 50.5
36.7 - 49.7
Mass
Concentration
yg/l
0.236
0.397
0.365
0.461
0.304
0.221
0.327
0.336
0.309
0.291
Grid Distribution
X2
16.04
10.20
13.35
5.98
12.97
9.94
6.73
5.60
11.06
10.93
Significance
Level of
Uniformity %
5
10
10
10
10
10
10
10
10
10
* Concentrations include contamination contribution
-------
TABLE A-23
Chrysotile Contamination in Millipore Filters *
Filter
No.
Croc. 3
Croc. 7
Croc. 13
Croc. 17
Croc. 23
Croc. 27
Croc. 33
Croc . 37
Croc. 43
Croc. 4 7
Mean Fiber
Concentration
x 106
13.0
8.70
5.64
10.4
19.8
9.68
20.0
15.9
8.78
15.3
95% Confidence
Interval
x 106
9.15 - 16.8
4.43 - 13.0
4.37 - 6.91
7.58 - 13.3
14.6 - 25.1
5.18 - 14.2
13.4 - 26.6
10.7 - 21.2
5.25 - 12.3
10.4 - 20.3
Mass
Concentration
Wg/1
0.121
0.044
0.044
0.055
0.267
0.067
0.095
0.098
0.042
0.080
Grid Distribution
x2
12.01
17.10
3.96
9.15
15.47
31.24
17.45
26.54
15.33
19.71
Significance
Level of
Uniformity %
10
2.5
10
10
5
< 0.1
2.5
0.1
5
1
* Results obtained by counting chrysotile on filters loaded with crocidolite
-------
TABLE A-24
Effects of Ultrasonlficatlon
Aqueous Chrysotlle Dispersion, 0.05 watt/ml
Operator
A
B
Exposure
Time
Seconds
0
15
60
300
0
15
60
300
Mean
Crocldollte
Concentration
x 106 fibers/1
46.3
36.4
52.9
53.8
73.3
67.4
82.6
111.0
952 Confidence
Interval
x 106 fiber s/1
32.8 -- 59.8
28.3 - 44.5
39.2 - 66.6
40.8 - 66.8
56.3 - 90.3
56.3 - 78.5
57.2 - 108.0
80.0 - 142.0
Mass
Concentration
V8/1
2.15
2.64
3.71
3.32
3.20
3.48
4.41
5.32
Grid Distribution
X2
16.59
7.36
17.27
12.80
19.13
9.48
41.11
36.13
Significance
Level of
Uniformity Z
5
10
2.5
10
1
10
<0.1
<0.1
-------
TABLE A-25
Effects of Ultrasonlfication
Aqueous Chrysotile Dispersion, 0.5 watt/ml
Operator
A
B
Exposure
Time
Seconds
0
15
60
300
0
15
60
300
Mean
Crocidolite
Concentration
x 106 fibers/1
42.2
56.9
47.8
79.3
59.8
79.9
89.7
92.6
95% Confidence
Interval
x 106 fibers/I
29.1- 55.3
44.6 - 69.2
32.1 - 63.5
62.9 - 95.7
41.4 - 78.2
55.3 - 104.5
77.2 - 102.2
77.9 - 107.3
Mass
Concentration
ug/1
1.90
4.02
3.91
3.66
2.87
5.76
5.35
3.17
Grid Distribution
X2
21.02
18.16
18.03
14.53
35.17
23.93
39.88
9.04
Significance
Level of
Uniformity Z
1
2.5
2.5
10
<0.1
0.1
<0.1
10
-------
TABLE A-26
Effects of Ultrasonification
Aqueous Crocidollte Dispersion, 0.05 watt/ml
CTi
Operator
A
B
Exposure
Time
Seconds
0
15
60
300
0
15
60
300
Mean
Crocidolite
Concentration
x 106 fibers/1
50.3
58.7
64.8
60.3
53.1
57.3
59.2
59.6
95% Confidence
Interval
x 106 fibers/1
36.3 - 64.3
43.1 - 74.3
48.8 - 80.8
47.6 - 73.0
34.5 - 71.7
40.6 - 74.0
37.9 - 80.5
45.2 - 74.0
Mass
Concentration
Mg/1
6.61
8.31
10.0
10.2
7.72
9.29
7.37
7.54
Grid Distribution
X2
20.29
12.36
16.29
6.19
34.77
19.82
23.20
18.28
Significance
Level of
Uniformity %
1
10
5
10
<0.1
1
0.5
2.5
-------
TABLE A-27
Effects of Ultrasoniflcation
Aqueous Crocidolite Dispersion, 0.5 watt/ml
Operator
A
B
Exposure
Time
Seconds
0
15 *
60
300
0
15
60
300
Mean
Crocidolite
Concentration
x 106 fibers/1
74.7
76.0
93.9
104.0
73.6
77.9
86.9
58.0
95Z Confidence
Interval
x 106 flbers/1
51.2 - 98.2
5.3 - 146.7
67.3 - 120.5
92.0 - 116.0
49.0 - 98.2
53.1 - 102.7
60.7 - 113.1
37.8 - 78.2
Mass
Concentration
Kg/1
10.6
7.72
9.03
16.7
9.07
9.38
10.5
9.70
Grid Distribution
X2
22.02
201.98
16.01
10.21
27.13
48.30
20.11
21.05
Significance
Level of
Uniformity Z
0.5
<0.1
5
10
0.1
<0.1
1
1
* Filter reported non-uniform by operator
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/4-78-OH
2.
3. RECIPIENT'S ACCESSI Ol> NO.
4. TITLE AND SUBTITLE
Preparation of Water Samples for Asbestos
Fiber Counting by Electron Microscopy
5. REPORT DATE
January 1978 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
E.J. Chatfield, R.W. Glass, and M.J. Dillon
8. PERFORMING ORGANIZATION REPORT NO.
ADDRESS
Sheridan Park Research Community
Mississauga, Ontario L5K IBS, Canada
10. PROGRAM ELEMENT NO.
1CC614
11. CONTRACT/GRANT NO.
68-03-2389
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Research Laboratory - Athens, GA
Office of Research and Development
U.S. Environmental Protection Agency
Athens, GA 30605
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/01
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Procedures used by a number of laboratories for the analysis of asbestos
fibers in water samples were investigated using standardized aqueous dispersions
of chrysotile, crocidolite, and taconite tailings. Filtration problems when
using polycarbonate filters were solved, and an optimum filtration technique
was developed that permits microscopically uniform deposits to be obtained.
Replicate analyses established fiber losses and reproducibility of five filter
preparation techniques. The uniformity of the deposits on all the final electron
microscope preparations was expressed quantitatively, and the most satisfactory
performance was given by all of the Nuclepore techniques. Ultrasonic treatment
at 20 kHz increased fiber concentrations of chrysotile suspensions but had no
effect on crocidolite suspensions.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Asbestos
Analyzing
Electron microscopy
Water pollution
Ultrasonic treatment
48A
68D
99A
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
132
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
118
* U.S. GOVERNMENT POINTING OFFICE: 1978 757-140/6675
-------
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Environmental Research Information Center
Cincinnati, Ohio 45268
OFFICIAL BUSINESS
PENALTY FOR PRIVATE USE, S3OO
AN EQUAL OPPORTUNITY EMPLOYER
POSTAGE AND FEES PAID
U.S. ENVIRONMENTAL PROTECTION AGENCY
EPA-335
Special Fourth-C lass R ate
Book
I
5
o
iu
(3
Mr. Gary McGee
Rm 564
Environmental Monitoring
and Support Laboratory
U.S. Environmental Protection
Agency
Cincinnati OH 45268
If your address is incorrect, please change on the above label
tear off; and return to the above address.
If you do not desire to continue receiving these technical
reports. CHECK HERE O; tear off label, and return it to the
above address.
EPA-600/4-78-011
------- |