£EPA

                   Labo'i-
                   Athens GA

Development of a
Rapid Analytical
Method for
Determining
Asbestos in Water

-------
                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-6ooA-78-o66
                                           December 1978
DEVELOPMENT OF A RAPID ANALYTICAL METHOD
    FOR DETERMINING ASBESTOS IN WATER
                     by
  Carl W. Melton, Sandra J. Anderson, Carolyn F. Dye,
     W. Eugene Chase, and Richard E. Heffelfinger
                  BATTELLE
              Columbus Laboratories
               505 King Avenue
             Columbus, Ohio  43201
            Contract No. 68-03-2199
                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. Environ-
mental Protection Agency, Athens, Georgia, and approved for publication.  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 en-
dorsement 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 mea-
suring 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. Although electron microscopic
methods are required for the quantitative determination of asbestos levels  in water,  such
methods are expensive, time-consuming and require sophisticated equipment not usually avail-
able in a conventional chemical laboratory.  There is a need, therefore, for a simple, rapid ana-
lytical method that can be used for the preliminary examination of a large number  of water
samples.  This report describes the  progress that has been made toward the development of a
rapid analytical procedure for chrysotile asbestos.
                                             David W. Duttweiler
                                             Director
                                             Environmental Research Laboratory
                                             Athens, Georgia
                                            ui

-------
                                       ABSTRACT
     The development of a rapid analytical method for determining chrysotile asbestos in water
that requires substantially less time per analysis than electron microscopy methods is described.
Based on the proposition that separation of chrysotile from other waterborne particulate would
greatly simplify the task of detection,  the research effort was directed toward establishing sepa-
ration and  concentration techniques.  This investigation led to the development of a separation
procedure whereby chrysotile is extracted from a water sample into an immiscible organic
liquid phase. The procedure is called two-phase liquid separation (TPLS).

     TPLS has been combined with a light microscopic intercept  counting technique, and also
with a colorimetric spot test  detection technique to result in two complete rapid analytical
methods.  The TPLS — light  microscopic (LM) method requires more expensive equipment than
the TPLS - spot test method; however, the limit of detection of TPLS-LM method  is  1.0 ng,
approximately 40,000 fibers, at the 99 percent confidence level, whereas the limit of detection
of the TPLS - spot test method is approximately  100 ng or 4.0 x 106 fibers. The TPLS-LM
method, therefore, is recommended as a first choice, and the TPLS — spot test method is
recommended to  be used under conditions that require no greater detection sensitivity than
100 ng per sample.

     TPLS extracts chrysotile from water into isooctane after the chrysotile surface  has been
rendered hydrophobic by reaction with an anionic surfactant (dioctyl sodium sulfosuccinate).
Extraction of the  chrysotile from the water phase  into the isooctane phase occurs as the two
liquids are  shaken in a separatory funnel.  Agitation creates an  emulsion  that is broken by
adding sodium chloride solution. The isooctane is then filtered to deposit the chrysotile on a
filter where its concentration is  analyzed by light microscopy or spot test procedures.

     For light microscopic  detection, the chrysotile is deposited on a Nuclepore filter, which
has a relatively smooth surface.  The filtered deposit is carbon  coated to make it reflective and
also to enhance the apparent fiber diameters to make more fibers detectable  by light microscopy.
Intercept counts are made of fiber segment lengths equal to the spacing between concentric
circles in a calibrated  eyepiece reticule. The measurement of total fiber  length as the diameter
of the filter is traversed is referred to a standard curve to obtain mass of chrysotile per filter
deposit.

     For spot test detection, the extracted chrysotile is deposited in a small (3-mm diameter)
area on a Millipore filter. The  chrysotile is reacted with manganous ion  and  washed, and the
reacted manganous ion is detected by a spot test reaction between potassium periodate and
tetrabase, which is catalyzed  by manganese.   Varying shades of blue serve to  indicate different
concentrations of chrysotile.
                                            IV

-------
     Experimental results have shown that 75 percent of the chrysotile is extracted by TPLS
procedures.  This is taken into account when evaluating  analytical results.

     This report was submitted in fulfillment of Contract No. 68-03-2199 by Battelle's
Columbus Laboratories, under the sponsorship of the U.S. Environmental Protection Agency.
This report covers the period from June 1, 1975, to June 30, 1978, and  work was completed.
as of June 1978.

-------
                                       CONTENTS

Foreword	    iii
Abstract	    iv
Figures	    viii
Tables	    x
Acknowledgments	    xi

     1.  Introduction	      1
     2.  Conclusions	      3
     3.  Recommendations	      5
     4.  Experimental Work	      6
           Phase I:   Separation Experiments	      6
                Initial Separation Experiments	      6
                Dielectric Separation	      9
                Two-Phase  Liquid Separation	    13
                Evolution of TPLS for Chrysotile	    13
                Factors Affecting TPLS  Optimization	    21
                Effects of Critical TPLS Factors on Selectivity and
                  Percent Recovery	    22
                TPLS Recovery
                                  	    22
                TPLS Selectivity	    25
                Handling Large Water Sample Volumes by TPLS	    27
                TPLS Recovery as a Function of Particle Size and
                  Aspect Ratio	    29
            Phase II:  Investigation of Detection Methods	    33
                Change of Flow Rate as a Function of Chrysotile
                  Concentration	    33
                Spot-Test Detection	    34
                X-Ray Diffraction	    35
                Light Microscopy	    36
            Phase III:  Application of Analytical Procedures	    39
                TPLS — Light Microscopic Analysis of Real Water Samples	    39
                TPLS - Spot Test Analysis of Real Water Samples	    45
     5.  Intel-laboratory Analysis of Four Water Samples	    49

References	    52
Appendixes

     A.  Instructions for Carrying Out Two-Phase Liquid Separation Coupled
        With Light Microscopic Detection of Chrysotile in Water Samples	    53
     B.  Instructions for Carrying Out  Two-Phase Liquid Separation (TPLS) Coupled
        With Spot Test Detection of Chrysotile in Water Samples	    63

                                           vii

-------
                                       FIGURES

Number                                                                               Page

    1     CPE instrument	      8

    2     Electron micrographs of effluents collected in test tubes 22 through 29
           in a CPE run made at pH 8.6 showing how pure chrysotile is confined
           to a band position in the curtain cell	     10

    3     Electron micrographs of effluents collected in CPE test tubes 20 through 28
           showing separation of diatoms from chrysotile in a run at pH 10.7  ....     11

    4     An electron micrograph of the original mixture of diatoms and chrysotile
           before CPE separation at pH 10.7	     12

    5     Chrysotile  collected on the anode in the dielectric cell containing
           isopropyl alcohol	     14

    6     Chrysotile  separated from diatomaceous earth in the first TPLS
           experiment	     16

    7     Particulate remaining in the water phase in the first TPLS experiment....     16

    8     Distribution of chrysotile  fiber lengths in sample 96-1 before (•) and
           after (A) TPLS	     31

    9     Distribution of chrysotile  fiber aspect ratios in sample 96-1 before (•)
           and after (A) TPLS	     31

  10     Distribution of chrysotile  fiber lengths in sample 96-2 before (•) and
           after (A) TPLS	     31

  11     Distribution of chrysotile  fiber aspect ratios in sample 96-2 before (•)
           and after (A) TPLS	     31

  12     Distribution of chrysotile  fiber lengths in sample 96-3 before (•) and
           after (A) TPLS	     32

  13     Distribution of chrysotile  fiber aspect ratios in sample 96-3 before (•)
           and after (A)  TPLS	     32
                                           Vlll

-------
 14     Distribution of chrysotile fiber lengths in sample 96-4 before (•)
          and after (A) TPLS	    32

 15     Distribution of chrysotile fiber aspect ratios in sample 96-4 before (•)
          and after (A) TPLS	    32

 16     Plot of pressure drop and flow rate vs chrysotile concentration	    34

 17     The eyepiece reticule used in fiber counting	    37

 18     Standard curve that relates light microscope counts to mass (/ug)  of
          chrysotile per filter	    41

 19     Light micrograph showing a  heavy mat of TPLS-extracted pure
          chrysotile from sample M3 0307 1500 D2 WGO	    47

 20     Comparison of spot test results on real water samples to spot test
          results on a series of chrysotile standards	    48

A-l     The mechanical shaker	    57

A-2     Separatory  funnels attached  to mechanical shaker with rubber bands  ....    57

A-3     Before (right) and after (left) breaking the emulsion by "salting out"
          with NaCl	    58

A-4     Draining off the water phase	    59

A-5     Filtering the isooctane	    59

A-6     A TPLS specimen prepared for light microscopic analysis.  The filtered
          TPLS fraction  was taped to a microscope slide and coated  with vapor-
          deposited carbon in a vacuum evaporator	    60

A-7     Concentric  circles in eyepiece reticle  superimposed on image of TPLS
          fraction from ORF water sample	    61

A-8     Standard curve which relates light microscope measurement of total
          fiber length to mass (jug) of chrysotile per filter	    62

B-l     Special reservoir with 3-mm-diameter hole	    66

B-2     Spraying device with vial containing spot test working solution	    66

B-3     A typical spot test result obtained on an 0.8-/jg chrysotile standard	    67
                                           IX

-------
                                       TABLES

Number                                                                            Page

    1     Effect of Agitation by Stirring	    24

    2     Percent Recovery With 2.5 x ID'3  M MT-70	    24

    3     Percent Recovery With 1.0 x 1(T3  M MT-70	    26

    4     Percent TPLS Recovery of Particles Detectable by Light
         Microscopy	    38

    5     Detection Limit Determinations	    39

    6     Light Microscopic Detection of Chrysotile at Twenty
         Concentrations	    40

    7     Results of TPLS  Applied to Real Water Samples	    42

    8     LM Analytical Results on Real  Water Samples	    46

    9     Spot Test Results on Real Water Samples	    48

   10     Comparison  of Analytical Results Obtained at Battelle and
         ORF on Four Water Samples	    51

-------
                                 ACKNOWLEDGMENTS
     The authors wish to acknowledge Mr. George Reimschussel of the Johns Manville Research
Laboratory, Denver, for supplying samples of pure chrysotile having different electrophoretic
mobilities.  They also wish to express appreciation to Dr. Eric Chatfield and Ms. M. Jane Dillon
of the Ontario Research Foundation, Toronto, Canada, for their aid in the evaluation of two-
phase liquid separation coupled with light microscopic detection and to Dr. C. H. Anderson for
his helpful  suggestions and interest in the development of analytical methodology for chrysotile
in water.
                                            XI

-------
                                        SECTION 1

                                     INTRODUCTION
     The presence of asbestos in water supplies in the United States and Canada, as well as
other parts of the world, has led to the need to rapidly analyze large numbers of natural- and
industrial-water samples for asbestos content.  Streams, runoff, lakes and city water systems
are the major receptacles of asbestos contamination. Natural weathering of asbestos deposits,
eroding of asbestos-cement city water pipes, and dumping of tailings from mining operations
are some of the mechanisms by which water supplies become contaminated with asbestos.  In
the past, transmission electron microscopy (TEM) has been the primary method for deter-
mining asbestos concentrations, but the time and cost required to analyze a sample becomes
prohibitive when  dealing with a large number of sampling sites.  Clearly, there is a need for a
more rapid and less expensive analytical method.

     The objective of the research program reported here was to develop a rapid analytical
method for asbestos in water requiring substantially less time and cost and less sophisticated
equipment than a method based on TEM.  Research was performed to develop a rapid monitor-
ing method that would eliminate  from further examination those samples containing asbestos
concentrations below significant levels.  After screening, only those samples identified as having
significant quantities of asbestos would be submitted for more  definitive TEM analysis.

     Asbestos analytical methodology employs  TEM because the fiber  sizes range well below  the
resolution capabilities of the light  microscope, and health effects investigators require fiber-size-
frequency data.   This does not imply that the rapid method must obtain fiber sizes. It need
only identify those samples that need to be fully characterized. Therefore, the detection
method employed in the rapid procedure  is based on bulk analysis.

     The logic behind the approach taken in the development of a rapid method was guided  by
an awareness of problems arising from gross amounts of interfering particulate prevalent in most
water samples. The sample could  be rid of interfering organic matter by low-temperature
ashing, but the inorganic particulate could not  easily be eliminated  to  leave only the asbestos.
Analysis  either had to be carried out in the presence of both organic and inorganic particulate
or in the presence of inorganic particulate or the asbestos had to be separated from the inter-
fering particulate. Analysis could  be performed for separated  and concentrated asbestos with
much less difficulty than for asbestos in the presence of interfering particulate.  Accordingly,
our research effort emphasized the development of a separation method for asbestos in order
to facilitate its subsequent detection. The separation of chrysotile was investigated first because
chrysotile is the most prevalent form of asbestos found in most sampling locations.  Also, its
surface chemistry is sufficiently different  than most other waterborne  particulate to make its
selective  separation possible.

-------
     The goal has been to develop a separation method that will not only eliminate interfering
extraneous participate, but will also reduce the time required for specimen preparation.  Speci-
men preparation methods employed in conjunction with electron microscopy detection often
require lengthy ashing periods to rid the sample of organic matter, but fail to eliminate masking
inorganic particles from the ash residue.  The developed separation method does not include an
ashing step, but does selectively and rapidly separate and concentrate the asbestos directly from
the original water sample.

     Three approaches to chrysotile separation were explored.  All exploited differences in
electrical properties between chrysotile and other prevalent  waterborne particles. The methods
investigated were continuous particle electrophoresis (CPE), dielectric separation, and extraction
of asbestos from aqueous suspension into an immiscible oil  phase (two-phase liquid separation).
Although CPE proved promising, there was greater  success with the two-phase liquid extraction
technique.   Consequently, most of the research effort was directed toward the study of two-
phase liquid separation (TPLS).

     Attempts to apply TPLS to amphibole separation were not successful; however, this was a
minor research effort. Early in the program it was decided to concentrate on chrysotile separation
rather than concurrently try to devise a procedure for the separation of amphiboles.  It is believed
that the separation of amphiboles by modified TPLS procedures remains a possibility.

     The complete analytical method for chrysotile includes TPLS  coupled  with light micro-
scopic analysis  of the chrysotile in the separated fraction.  Alternatively, the separated and
concentrated chrysotile is analyzed by a catalytic spot test procedure.

     Two-phase liquid separation promotes the selective transfer of chrysotile from  water to an
immiscible organic liquid after the  chrysotile surface has been rendered hydrophobic by reaction
with an anionic surfactant. The organic liquid is then filtered to deposit the separated chrysotile
on a Nuclepore filter.  The filtered chrysotile is carbon coated  and viewed with a light micro-
scope employing bright field incident illumination at 500X, and counts are  made of fiber
segments as the diameter of the effective filter area is traversed.  The counts thus obtained  are
translated into  chrysotile mass (jug) per filter by means of a standard curve  that was generated
using standard filtered chrysotile suspensions.

     The spot-test analytical procedure is applied to the TPLS  fraction that has been deposited
on an «3 mm diameter area on a Millipore filter.  First, the deposited chrysotile is  reacted  with
manganous ion and washed, and then the reacted manganese is detected by a chromogenic spot
test reaction between potassium periodate and tetrabase, which is catalyzed by manganese.

     The limit of detection  of the TPLS - light microscopic method is 1 nanogram at the
99 percent confidence level.  The smallest quantity of chrysotile detectable by the  spot test is
100 nanograms. Although the spot test  is much less sensitive,  it can be used effectively in
certain monitoring situations. The relationship between mass and total number of  chrysotile
fibers is variable and strongly dependent upon fiber dimensions.  All water samples  show a
broad range of fiber size distribution; for simplification in this  work it has been assumed that
the average fiber is 1  jum long and 0.1 jum wide.  When this assumption is made 1 ng is
equivalent to approximately 40,000 fibers.

-------
                                       SECTION 2

                                     CONCLUSIONS
 (1)   Continuous particle electrophoresis (CPE) will separate chrysotile, but CPE will not
       handle sufficiently large volumes of particulate to be used in an analytical method.

 (2)   Dielectric separation as investigated is ineffective for chrysotile.

 (3)   Two-phase liquid separation of chrysotile from water samples coupled with light
       microscopic analysis of the separated chrysotile by intercept counting will detect
       chrysotile quantities as low as 1.0 ng at a 99  percent confidence level.  Assuming an
       average chrysotile fiber size of 1  /xm length and 0.1  jum width, 1 ng is the equivalent
       to approximately 40,000  fibers.

 (4)   TPLS extracts 75 percent of those chrysotile  particle sizes detectable by light
       microscopy.

 (5)   The sensitivity  of the light microscopic detection procedure is high over the concentra-
       tion range above 1.0 ng; an average of nine counts is registered at 1.0 ng and 15,000
       counts at 1000 ng.

 (6)   Few interferences occur in TPLS; it is surprisingly selective.  Organic filaments and platy
       mineral particles, however, are occasionally extracted along with chrysotile.  The fila-
       ments can be distinguished from  chrysotile  with a little practice.  The presence of the
       platy mineral can result in some  of the chrysotile being obscured from view.

 (7)   The TPLS — spot test procedure requires relatively inexpensive equipment and can be
       run with a minimum  of training  by an individual with no specific background.

 (8)   The detection limit of the spot test procedure is approximately 100 ng.

 (9)   The spot test is selective for chrysotile; no interferences were encountered while running
       real water samples.

(10)   TPLS can separate as large as 0.4-mg quantities of chrysotile from 100-ml water aliquots.

(11)   Degree of agitation, pH, concentration of surfactant, the ratio of the volumes of oil/
       water, and the number of extractions from a  single  water aliquot are the most critical
       TPLS factors.

-------
(12)   Percent chrysotile recovery by TPLS is optimized:
       • at pH 3.5 to 5.0
       • when the oil/water volume ratio is 1:2
       • when shaking is the mode of agitation
       • when 2.5 x 10~3  M anionic surfactant concentration is used
       • when five extractions are made from  the same water aliquot.

(13)   TPLS favors the extraction of longer fibers with larger aspect ratios.

(14)   Filtration  and resuspension is a practical means of concentrating particulate from large
       water volumes for TPLS.

(15)   It is not as feasible  to scale up TPLS to handle large water volumes as it is to filter and
       resuspend  particulate  in smaller, more manageable water volumes for  TPLS.

(16)   The results of the interlaboratory analyses at Battelle and ORF were  in agreement
       within a factor of 4, except on Sample No. 4 in which Battelle found only 1/10 of the
       chrysotile  reported  by ORF.  Battelle's TPLS — light microscopic result in this case
       agreed with the electron microscopic result obtained at ORF.

(17)   Approximately eight samples  can be prepared and analyzed by the TPLS — light
       microscopic method by two people in one day.

-------
                                       SECTION 3

                                  RECOMMENDATIONS
     As with any new analytical technique, one must proceed with caution and become
thoroughly familiar with the application of two-phase liquid separation, light microscopic
detection and spot test application before relying on the analytical results.  It is recommended,
therefore, that the results of these procedures be  compared initially to results obtained on the
same samples by electron microscopy.  Afterwards, when there is sufficient confidence in the
method's effectiveness, one can progress toward the application of TPLS coupled with either of
the detection procedures as requirements dictate without comparison to electron microscopic
analytical results.

     Although water samples were analyzed using both TPLS coupled with light microscopic
counting and the TPLS  — spot test procedure, it was not possible to investigate all varieties of
water samples.  Thus, unanticipated interferences may occur to give high or low results under
sample conditions not investigated.

     It is recommended that the developed methodology be applied first to monitoring varia-
tions in chrysotile concentrations from well characterized sources, and then, as  experience
develops, be applied to  the analysis of chrysotile  from different sources.

-------
                                       SECTION 4

                                EXPERIMENTAL WORK
     This section of the report discusses the experimental work performed to develop separation
and detection methods for chrysotile.  Separation experiments explored dielectric separation,
continuous particle electrophoresis, and two-phase liquid separation.  When promising results
were obtained using TPLS to separate chrysotile, emphasis was placed on the optimization of
this process.  No further search was made for other separation methods for chrysotile.  Also,
when initial experiments to separate amphiboles by various modifications of TPLS were not
completely successful, it was decided to concentrate the entire effort on improvement of the
separation of chrysotile.

     Several detection methods were investigated with the  objective of developing a method
requiring less time and less expensive equipment than electron microscopy.   Spot-test reactions,
changes in flow  rate through filtered chrysotile  deposits, X-ray diffraction, and light microscopy
were evaluated.  Light microscopy of the separated and filtered chrysotile was the most success-
ful detection method found, and it has been written into the detailed instructions for the final
method. A  spot-test procedure was also  developed for chrysotile detection and is recommended
as an alternative.
PHASE I:  SEPARATION EXPERIMENTS

Initial Separation Experiments

Continuous Particle Electrophoresis —
     The Beckman Continuous Particle Electrophoresis (CPE) System was used in experiments
designed to separate and concentrate chrysotile from synthetic mixtures of chrysotile and
diatomaceous earth.  The operating principle of CPE and related experiments are described in
the following sections.

Operating Principle  of CPE —
     Simply stated,  the  operating principle of Beckman CPE is as follows.  A suspension of
sample material is introduced as  a fine, steady stream into a vertical, flowing "curtain" of
electrolyte.  A horizontal dc electrical field is applied to the curtain. Each particle has two
components of motion:   a vertical component (Vv) that is the same as the  curtain velocity, and
a horizontal component (V^) that is proportional to the electrophoretic mobility or zeta
potential of the particle. Each different particulate component is therefore deviated a different
angle, a, from the vertical, given by

                                      tan a = ~
                                              vv

-------
The net result at the lower end of the cell is a particle band position that is proportional to the
zeta potential of the particles.  Where a mixture of particles having different electrophoretic
mobilities is analyzed,  each type of particle will assume a different band position.

     The possibility of fractionating different types of particles by CPE assumes a certain mini-
mum  difference in  their electrophoretic mobilities. The relative mobilities in a  mixture can
frequently be manipulated.  One useful procedure is to change the pH. Others  include the use
of polyvalent cations,  and use of surfactants and various agents that alter the surface char-
acteristics of the particles.

     The CPE instrument is shown in Figure 1. This system can be separated into five sub-
sections:  (1) curtain flow system, (2) sample flow system, (3) electrode rinse systems,
(4) cooling system, and (5) power supply. The curtain flow system provides a filtered source
of the electrolyte at a  constant flow  rate to the cell.   The sample flow system provides a means
of continuously metering a sample into the flowing curtain medium.  The electrode rinse system
is designed to remove  electrolysis products from the electrodes and thereby maintain a high
degree of pH stability  within the curtain. The power  supply is designed to provide a constant
voltage gradient within the cell regardless of changes in internal impedance or electrolyte
conductivity.

     The electrodes, located in two chambers positioned at the sides of the curtain cell,  are
isolated from the curtain by semipermeable membranes.  This separation prevents mixing of  the
electrode  rinse with the curtain buffer. The sample is introduced at the top of the  curtain and
is thereby diluted by the curtain buffer solution as it flows downward  between  the electrodes.
The sample particles are influenced by the voltage gradient to migrate laterally,  and finally to be
discharged through drainage tubes positioned at uniform (1 mm) intervals  across the bottom of
the curtain cell.  The effluent is thus divided into 48 fractions and collected in  test  tubes.

     According to the  instructions in the use of the CPE system, most particles possess a higher
zeta potential at higher pH conditions. Consequently, they recommend the utilization of buffer
media at relatively  high pH values. The instructions also suggest the use of a 30 V/cm voltage
gradient for most applications.  Therefore, most of the initial CPE experiments  were carried  out
at pH values greater than 7.0 — employing a voltage gradient of 30 V/cm.

CPE Experiments —
      CPE experiments were conducted at pH 8.6, 3.6,  and 10.7.  Runs were made at  pH 8.6
and 10.7  because the  CPE instructions stated that better separation usually was effected at
higher pH conditions.  The run at pH 3.6 was made because literature^) indicated that the
greatest positive surface charge  was developed  on chrysotile at this pH.

     The first attempt to separate chrysotile from diatomaceous earth was made in a veronal
buffer at pH 8.6. An  initial run was made with pure chrysotile dispersed in filtered, deionized
water containing 1.9% aerosol OT to determine into which tubes the chrysotile would be dis-
charged.   It was found that the bulk  of the chrysotile was collected in the test  tubes numbered
20 through 29.  This was determined by filtering 5-ml aliquots from each tube  onto 25-mm
HAWP Millipore  filters and examining prepared specimens by electron microscopy.  The speci-
mens were prepared by cutting a piece from the center of the filter and dissolving the filter
material away in an acetone bath while the deposit side was in contact with a carbon support

-------
                                               Cell F il ling Control
    Magnetic Sample Holder
         Sample Inlet Tube
       Electrophoresis Cell

        Curtain Flow Meter

Fractionation Viewing  Area


          Collecting Tubes
48-Tube Fraction Collector
                                                               Cell Current Monitor
  Cell Voltage Monitor




  Voltage Gradient Monitor



  Vcltage Gradient Adjust

  Voltage Gradient Runye Switch

  Curtain Pump Speed Control

  Curtain Pump Switch
— Rinse  Pump Switch
  Cell Voltage Switch


  Power Swi tch
                                        Figure 1.  CPE instrument.

-------
film held by a mesh electron microscope grid.  Electron micrographs were taken of preparations
containing asbestos, and those micrographs showing the bulk of the asbestos were contact
printed.  The resulting contact print is included in  Figure 2.

     A second run at pH 8.6 was made on pure diatomaceous earth also dispersed in filtered
deionized water containing 1.9% aerosol OT.  The diatoms were discharged in the same tubes
as was the chrysotile; consequently, no further experiments were performed at pH 8.6.

     Subsequently, separations were attempted at pH 3.6 (phosphate buffer) at which agglomer-
ation of particulate resulted in plugging of the tubes through which the effluent from the
curtain cell was discharged. The addition of the dispersant Aerosol OT did  not prevent
agglomeration at pH 3.6.

     Next, runs were made at pH 10.7 (in 0.001 M Na2CO3> first separately on pure chrysotile
and pure diatomaceous earth dispersed in 1.9% Aerosol OT in deionized water.  These experi-
ments indicated that separation  might be obtained  at this pH; the chrysotile was concentrated
primarily in tubes 20 through 24 and diatoms in tubes  25 through 30.  Then, a run was made
in which separation was attempted on a 10:1  by weight (7.0 mg and 0.7  mg in 100 ml H2O)
mixture of diatomaceous earth and chrysotile, respectively.  This mixture also was dispersed in
1.0% Aerosol OT in filtered deionized water.  As predicted by the runs on the pure materials,
a separation was effected.  Figure 3 shows electron microscope preparations from test tubes
numbered 20 through 27.  Although some diatoms were found in tubes 22, 23, and 24, the
degree of separation represents a marked improvement over the  original mixture shown in
Figure 4.

     Possibly, deviation from  the 30-V/cm voltage  gradient would yield even better separations.
Also, instrument instructions state  that lower curtain flow rates result  in better separations.
All experiments were conducted at a curtain flow rate of 20 ml/minute with a sample introduc-
tion rate of 50 /xl/minute.  Of course, as the curtain flow and sample introduction rates  are
decreased, longer times are required for CPE separations. Compromises must be made in
achieving the desired balance between degrees of CPE separation and run duration.

     At this point, concurrent experiments with two-phase liquid separation proved  more
successful, and CPE studies were discontinued.

Dielectric Separation

     Dielectric separation^) is based on the relationship between the dielectric constants of
suspended particulate material and that of the suspending liquid.  When the suspending liquid
has a lower dielectric constant than the  particle, the particle is attracted to immersed electrodes
when either an ac or dc voltage is imposed.  Theoretically, when the difference is the least
between the dielectric constants of the liquid and the particle type to be separated, the  most
effective separations are obtained.

     First experiments  were performed on pure chrysotile suspended in liquids having relatively
low dielectric constants compared to chrysotile (dielectric constant = 33.7)  in order to study the
attraction of chrysotile to  the electrodes at various voltages under both ac and dc conditions.
Most of these trials were carried out in Freon MF  having a dielectric constant of 2.4.  It was

-------
,-



>
     x
     -'


                                     ; -v   -5
                                    "%/
                                      /;'
                                                                                 ,
                                                                                         p .. »ffl&
                                                                                          4d

           22
           23
24
25
           26
           27




29
          Figure 2. Electron micrographs of effluents collected in test tubes 22 through 29 in a CPE run made
                   at pH 8.6 showing how pure chrysotile is confined to a band position in  the curtain cell.

                                     5000X

-------
 20

                               22

                               23
  25
26
27
28      5000X
Figure 3.  Electron micrographs of effluents collected in CPE test tubes 20 through 28 showing separation
          of diatoms from chrysotile in a run at pH 10.7.

-------
found that chrysotile was best attracted and held at 200 Vdc to electrodes separated by a
distance of approximately  1 cm.  Below 200 V the chrysotile did not collect as rapidly on the
electrodes, and above 200  V a turbulence developed that dislodged much of the deposit.
Consequently, 200 Vdc was employed in subsequent dielectric separation experiments.
                                                      5000X

           Figure 4.  An electron micrograph of the original mixture of diatoms and
                     chrysotile before CPE separation at pH  10.7.
     Although chrysotile was deposited on the electrodes, it was not held well enough to lift
the electrodes from the  liquid and thereby remove the deposit.  In order to be able to retain
the chrysotile deposit on the electrodes while undeposited particulate was removed, a flow-
through cell was designed.  The cell design permitted gentle flushing of the liquid containing
undeposited suspended particulate from the cell  without disturbing the chrysotile deposit. A
reservoir that  contained particle-free liquid and that had a drain on the side delivered this
liquid to the dielectric cell positioned below. The dielectric cell, which held 40 ml, also had a
drain at the bottom which could be adjusted to  a flow which matched the rate of introduction
of the particle-free liquid. After the deposit had formed on the electrodes, the unattached
particulate was flushed from the cell with 400 ml of liquid.  Then the deposited  participate
was resuspended in the remaining liquid and drained from the cell into a separate container.

     Investigations of dielectric separation were carried out primarily in methyl, ethyl, and
isopropyl alcohols because their dielectric constants are below that of  chrysotile  but relatively
high compared to most other liquids. Also their dielectric constants can be altered upward with
the addition of water to create solutions  that are even nearer to the dielectric constant of

-------
chrysotile.  For example, water has a dielectric constant of 80 and ethyl alcohol a constant of
24.  The change in dielectric constant is a straight line function, and thus 90% ethyl alcohol-
10% water solution has a dielectric constant of 30.  Separation trials were made, therefore,
first with pure ethyl, methyl, and isopropyl alcohols and later with ethyl alcohol to which 10%
water was added  to elevate the dielectric constant.

     First,  separate suspensions of pure chrysotile and pure diatomaceous earth  were run in
various candidate liquids.  In cases where it was observed that  chrysotile was deposited and
diatomaceous earth was not, 10:1 mixtures by weight of diatomaceous earth and chrysotile
were run.  Figure 5 illustrates how chrysotile was deposited on the anode from  an  isopropyl
alcohol suspension of chrysotile.  Despite the fact that this was originally a suspension of
chrysotile fibrils in the colloidal size range, a visible deposit was formed on the  electrode.

     Always  before making a run, the particulate was dispersed ultrasonically in the dielectric
liquid without the aid of additives. On the basis of previously described experiments, pure
methyl alcohol, pure isopropyl alcohol, and 90% ethyl alcohol-10% water solution  were  used in
dielectric separation experiments. Although in each case some enrichment of asbestos was
realized,  the  results were inferior to continuous particle electrophoresis and two-phase liquid
separation.  It appears that interaction may occur between chrysotile and diatoms to result in
the attraction of both species to  the electrode.  Also, inevitably, as asbestos particles agglom-
erated, a mass was formed that trapped diatoms as they circulated in the dielectric liquid.

     Experimentation with dielectric separation was curtailed in favor of the pursuit of the
more promising two-phase liquid separation.

Two-Phase Liquid Separation

     Two-phase liquid separation was  conceived as a possible means of chrysotile separation.
First experiments were successful, but success occurred somewhat accidentally.  The following
section discusses  the process of TPLS  discovery.

Evolution of TPLS for Chrysotile

     The first TPLS experiment was conducted in an attempt to separate chrysotile from a
10:1 by  weight synthetic mixture of diatomaceous earth  and chrysotile that had been used in
CPE experiments. The original mixture contained 7.0 mg of diatomaceous earth and 0.7 mg of
chrysotile dispersed ultrasonically in 100  ml of 0.001 N Na2CC»3 containing 0.1% Aerosol OT.
The mixture had been treated ultrasonically to create a dispersion that would not clog the CPE
discharge tubes.

     Three ml of the original synthetic mixture was diluted by the addition of 10 ml of 0.001 N
sodium carbonate, combined with 50 ml of benzene containing 0.5 ml of oleic acid,  and agitated
ultrasonically in a separatory funnel.   The oleic acid addition was made because it  was stated'*)
that long-chain aliphatic acids are chemisorbed by chrysotile.  It was reasoned that chemisorp-
tion of oleic acid would create a hydrophobic surface on the chrysotile fiber and thus promote
its transfer to the benzene phase.
                                             13

-------
Figure 5.  Chrysotile collected on the anode in the dielectric
          cell containing isopropyl alcohol.

-------
     An emulsion formed upon agitation separated into a benzene layer, a cream layer, and a
water layer after standing overnight.  One ml of the benzene layer was pipetted off and filtered
through a 25-mm Millipore filter having a 0.45  /xm pore size. Ten drops of the water layer were
also filtered.  Both filtered residues were prepared for electron microscopy.  When the prepara-
tions were examined in the electron microscope it was found that the particulate in the benzene
phase was comprised almost entirely of chrysotile (Figure 6) and that the particulate in the
water phase was primarily  diatomaceous earth (Figure 7).

     These experimental results set the stage for further experiments designed to determine the
critical factors involved in  TPLS.

     Results of the first experiment indicated that the oleic acid addition might be effecting a
separation as predicted.  In additional experiments employing comparable synthetic mixtures
of 10:1 by weight of diatoms and chrysotile suspended in water containing no sodium
carbonate or Aerosol OT, however, the addition of oleic acid did not promote chrysotile
separation.  Then, 0.05% of Aerosol OT was added to the water suspension, and another trial
was made in  which 50 ml of the water suspension was shaken in a separatory funnel with 50 ml
of benzene containing 0.5  ml of  oleic acid; again chrysotile was found to be selectively trans-
ferred to the benzene phase.  Subsequent experiments in which the  oleic acid was eliminated,
but in which Aerosol OT was added, effected chrysotile separation.  Other experiments without
Aerosol OT or any other additions resulted in no chrysotile separation.  It was concluded,
therefore, that Aerosol OT was reacting with the surface of the chrysotile to promote its
transfer to the benzene phase.

     This discovery led to  the investigation of TPLS mechanisms and conditions that affect
percent recovery and selectivity.  The early experiments were carried out with benzene; how-
ever, because of the possible toxicity of benzene, isooctane was ultimately used as the oil phase.
The following section discusses the mechanisms identified through pertinent literature and
experimentation.

The Mechanisms of TPLS -
     Essentially,  two-phase liquid separation selectively extracts chrysotile from a water sample
into a  water-immiscible oil phase, such as benzene, mineral spirits, or isooctane, after the sur-
face of chrysotile fibers is made  hydrophobic through a reaction with an anionic surfactant.
TPLS selectivity  is based on the  difference  between the zeta potential of chrysotile and the
zeta potentials of most other  waterborne particulate.  The surface charge of chrysotile is posi-
tive over a pH range from  2 to 11.4; whereas most other waterborne particulate is negative
over this pH  range.  The chrysotile, therefore, selectively reacts with the negatively  charged
anionic surfactant, becomes hydrophobic, and is transferred from the water phase to the oil
phase.

     Speil and Leinweber^^ state that chrysotile reacts with certain anionic wetting agents,
such as Aerosol OT, and that  this reaction is accompanied by a strong chemisorption of the
agents  with permanent modification of the  surface.  Because MgOH  groups of the brucite layer
are presented for reaction on  the surface of chrysotile(3), the Aerosol OT (dioctyl sodium
sulfosuccinate) most likely reacts with the brucite layer to create a  tightly bound organic coat-
ing covering the fiber surface.
                                             15

-------
                                             t    _        I
                                                              7500X
Figure 6.  Chrysotile separated from diatomaceous earth in the first TPLS experiment.
J22542
          J22545                                             7500X

  Figure 7.  Particulate remaining in the water phase in the first TPLS experiment.


                                      16

-------
     Reinders(4) has defined interfacial tension conditions governing transfer of dispersed solids
from the aqueous phase to the oil phase. If 7WO is the interfacial tension of the  oil-water
interface, 7SW is the interfacial tension of the solid-water interface, and 7SO is the interfacial
tension of the solid-oil interface, the following conditions may exist.

     (1) If 7SO > 7WO + 7SW, the solid remains suspended in the water phase.
     (2) If 7SW > 7WO + 7SO, the solid is transferred to the  oil  phase.

     (3) If 7WO > 7SO + 7SW, or if none of the interfacial tensions is greater than
         the sum of the  other two, the solid will concentrate at the oil-water
         interface.

     Condition (1) exists  when the solid  surface is hydrophilic and condition (2) or (3) may
exist when the solid surface is hydrophobic. Experimentally, it  can be determined which condi-
tion prevails by contact angle measurement.

     No contact angle measurements have been made during the  investigation of the chrysotile
separation.  They are rather difficult to  make on a fibrous mineral because the surface cannot
be polished smooth.

     The sign and magnitude  of the zeta potential of the mineral surface is affected by the pH.
In turn, the zeta potential determines whether and to what degree a reagent will react with the
mineral surface.

     In the case of stoichiometric chrysotile, the zeta potential is positive in the pH range below
the isoelectric point (zero point charge), which occurs at pH  11.4, attaining a maximum positive
value in the range pH 3 to pH 4.  Above pH 11.4 the zeta potential of chrysotile is negative.

     Having a positive zeta potential at pH  7.0 and below is relatively unique among water-
borne  particles. This phenomenon provides the  basis for separation of chrysotile  from other
particles. By making its surface hydrophobic through reaction with anionic reagents at
 7WO + 7SO is created to promote selective transfer of chrysotile
to the oil phase. Aerosol OT, MT-70, and  MO-70 surfactants, which are anionic analogous
alkyl sodium sulfosuccinates,  have been  employed successfully in the separation of chrysotile.

     The idealized  formula for stoichiometric chrysotile^) is Mg3Si2C>5 (OH)4- The fibrils have
a rolled sheet structure composed of an outer  layer of magnesia  bonded to a layer of silica.
Hydrated chrysotile presents MgOH groups  for reaction with  reagents.  Chowdhury  and
Kitchener state that the higher the magnesia content the greater is the positive zeta potential.
Also, there is evidence that a greater degree of hydration of the  magnesia surface increases
electrophoretic mobility.  When all magnesia is leached out under acid conditions or by
weathering, a silica gel pseudomorph that is left  has a strongly negative zeta potential. This
structure, not meeting the compositional criteria, is not chrysotile.

     According to Lai and Fuerstenau(6), anionic reagent (collector) adsorption by  or reaction
with the mineral surface is primarily determined by the reagent  concentration and the pH as it
affects zeta potential.  When reagent concentration and mineral  zeta potential are optimum, the
adsorption density of the reagent, as expressed in moles/cm^, on the mineral surface is at a
                                             17

-------
maximum. As a result, the hydrophobicity also becomes the highest possible with that par-
ticular reagent.

     The degree of hydrophobicity is also a function of reagent type.  Those reagents having
larger hydrophobic alkyl groups render the mineral surface more hydrophobic.  Lai and
Fuerstenau also found that the percent recovery of alumina from water suspension into
isooctane, using sodium akyl  sulfonate reagents with C8, CIO, C12, and C14 alkyl groups,
increased markedly with larger carbon chain length.

     Heeding this finding, a search was made for a sodium  sulfosuccinate having larger alkyl
groups than the octyl groups of Aerosol OT.  A reagent was found, Monawet 70 (MT 70),
which is ditridecyl sodium sulfosuccinate. Experiments with this compound established that
it also promoted transfer of chrysotile to the oil phase. Subsequent experiments (to be
discussed later) indicated improved recovery but also reduced selectivity when compared to
dioctyl sodium sulfosuccinate (MO-70 and Aerosol OT).

     Most of the literature discusses  adsorption of reagents. The reaction of the sulfo-
succinates with the chrysotile surface may be more than adsorption, however:  Speil and
Leinweben')  state that the highly reactive chrysotile surfaces causes many interesting surface
reactions to take place that are intermediate between simple adsorption and  true chemical re-
action.  They  also  report that the reaction of chrysotile with certain anionic wetting agents,
such as Aerosol OT,  is accompanied  by a strong chemisorption of the reagents with permanent
modification of the surface.

     In TPLS, an organic liquid that is immiscible with water is agitated with water, and the
transfer of suspended chrysotile from the water phase to the oil phase  is promoted by the
anionic surface active agent.  The degree  of interpenetration of phases as it affects the total
interfacial contact  is of critical importance in recovery  of particles.  Interfacial contact is
determined by the intensity and mode of agitation, the oil/water volume ratio, and  the
surfactant concentration that controls emulsion formation.

     Continual agitation is more effective than intermittent agitation because the droplets are
not allowed time to  form a stable emulsion.  Consequently, the total interfacial contact becomes
a function of agitation time and intensity.

     The anionic reagent concentration must be high enough to lower the 7WO to the point that
contact between the phases is optimized.  Lai and Fuerstenau found that a 2.5  x 10~4 M con-
centration of C14  sulfonate promoted maximum alumina recoveries.

     The water/oil volume ratio of 25% contains the minimum oil volume to attain  close pack-
ing of ideal uniform  spheres.   Lai and Fuerstenau state that, under these conditions, the pos-
sible phase inversion  from an oil-in-water to a water-in-oil  emulsion is  diminished.

     As one might anticipate, the more wetting agent added, the more stable and finely
divided was the emulsion formed during agitation. When effective concentrations of surfactant
were added, the time required for the emulsion to break and separate completely into two
layers became  a problem. The dispersed  oil droplets first coagulated, rose to the top of the
water layer to form a cream layer; then the cream layer very slowly coalesced contributing


                                            18

-------
to the formation of a benzene layer above the cream layer.  Complete coalescence of the cream
layer  sometimes required as long as overnight.

     Emulsification and Demulsification - There are conflicting requirements in the TPLS sys-
tem.  Ideally, a finely divided oil in water emulsion must be formed to extract the major
quantity of asbestos present, and then this emulsion must break rapidly and completely into
a water layer and an oil layer.  The more finely divided the emulsion, the more stable it is.
The time required for complete separation becomes too long, without some aid to accelerate
the process.

     Emulsification — As previously described,  the first dispersion method employed to form
the emulsion was ultrasonic treatment.  Then, merely hand shaking was used. In either case,
when «0.04% of Aerosol OT was added to the  water phase or benzene phase, an emulsion was
formed that required a standing period of approximately 12 hours to break completely.

     Literature^) indicates that intermittent shaking was more effective in emulsion formation
than  uninterrupted shaking.  Sherman suggests that it requires a small amount of time for the
surface-active reagent to be adsorbed  on the surfaces of newly formed droplets and to stabilize
them fully. Experimentally he found that five  intermittent shakings (2 shakes each) with a rest
period of 20 to 30 seconds between shakings would completely emulsify 60% by volume of
benzene in 1% aqueous sodium oleate.  It required  3000 uninterrupted shakes in a machine,
ksting about 7 minutes, to  completely emulsify the same mixture. First, intermittent shaking
was incorporated into the TPLS procedure, and appeared to be reasonably effective.  However,
Sherman also states that hand shaking results in globules ranging from 50 to 100 /xm and that
to get smaller droplets one must apply more vigorous agitations to the liquids. It was speculated
that more  complete extraction of chrysotile would  be promoted if the surface contact were in-
creased between the water phase and  benzene phase by the creation of smaller oil globules, but
ultimately it was found that more complete chrysotile separation was promoted by continual
agitation than by intermittent agitation.  Continual agitation presumably increases the  contact
between the water and oil phases by not allowing time for dispersed oil droplets to acquire the
stabilizing  surfactant layer; the droplets form, coalesce, and reform, and thereby extract more
chrysotile from the water phase than  when the  first droplets formed are allowed time  to
stabilize.

      Demulsification — In the initial stages of development of TPLS, it was obvious that some
procedure  must be devised to promote rapid and complete demulsification.  Consequently, a
search was made for a simple, reproducible method of demulsification that would not interfere
with  the selective extraction of chrysotile into the benzene phase.

     In consulting the literature^), it was found that there are several ways to promote
demulsification. Some of the approaches to breaking an emulsion are:  addition of excess
solvent, addition of a solvent miscible with each of the two phases, destruction of the emulsify-
ing agent, salting out, filtration, heating, freezing, electrolyzing, centrifuging, or changing the pH.

     The approaches to demulsification considered most appropriate for application to the TPLS
system from the standpoint of simplicity and possible lack of interference were heating,  salting
out, and changing the pH.  Changing the pH was tried first, and it was found that the emulsion
                                            19

-------
broke at pH 3. In view of the rather high solubility of chrysotile in acid, this approach was
abandoned.

     Heating the emulsion was tried next, and it was found necessary to heat almost to the
boiling point of water to break the emulsion.  Heating to such a high temperature caused the
benzene phase to boil off explosively evolving large quantities of noxious benzene vapor. How-
ever, demulsification by heating has the attraction of not changing the chemical environment;
thus successive extractions of chrysotile under similar conditions from the same water sample
are possible.

     Salting out was the most successful method.  G. M. Sutheim(^) states that if a strong
electrolyte salt is added, the emulsion coagulates as a result of interference with the adsorbed
emulsifying agent at the  benzene-water interface.  He found the coagulating  effect of anions
to decrease as follows:  citrate, sulfate, acetate, chloride; for cations:  sodium, potassium,
ammonium, magnesium,  calcium,  barium. According to this listing, sodium citrate should have
the greatest coagulating effect.  Therefore, sodium citrate was the first reagent to  be tried in a
salting out experiment.

     Ten ml of 25% sodium citrate was added to an emulsion created by shaking  together 50 ml
of water containing a suspension of 0.6 mg of chrysotile and 9.2 mg of diatoms and 50 ml of
benzene to which 2.0 ml of 1.0% Aerosol OT dissolved in benzene was added. After the addi-
tion of sodium citrate the mixture was shaken once and allowed to stand. Within 5  minutes the
emulsion broke completely into two layers.  The benzene layer was pipetted off and filtered,
and a preparation was made for electron microscopy. Both diatoms and chrysotile were found
to be present in the benzene fraction.  The interference with selective separation was attributed
to the organic citrate ion.  Consequently, sodium chloride was selected for subsequent trials.
Ten ml of 25% NaCl was added to an emulsion formed by intermittently shaking  together:
     (a)  50 ml of water containing  a suspension of 0.2 mg of  chrysotile and 5.0 mg
          of diatoms
     (b)  50 ml of benzene to which was added 2.0 ml of 1.0% Aerosol OT in benzene.

After adding the  NaCl, the emulsion was shaken again and allowed to stand until  it separated
completely into benzene and water layers.  This required approximately 5 minutes. The
benzene layer  was then pipetted off and filtered, and a preparation was made for  electron
microscopy. Only asbestos was found in this preparation indicating that salting out with NaCl
did not interfere  with TPLS selectivity.

     When these developments indicated that TPLS could be speeded up, effort was directed
toward TPLS optimization.  Attention was focused on the investigation of the identified
factors affecting degree of TPLS selectivity and recovery of chrysotile from the water.

     The following  listed factors were taken into consideration  during the experimental studies
to establish the best TPLS procedure.  Those factors subject to control were varied to search
for optimum conditions  for TPLS selectivity and recovery.
                                            20

-------
Factors Affecting TPLS Optimization

      (1)  Relative interfacial tensions in the solid-liquid-liquid system
      (2)  Hydrogen ion activity (pH)
      (3)  Concentration of the surfactant
      (4)  Oil/water volume ratio
      (5)  Alkyl chain length on the surfactant
      (6)  Zeta potential of the particle type
      (7)  Adsorption density (moles/cm^) of surfactant surface
      (8)  Chemistry of the chrysotile surface
      (9)  Effects of added ions
     (10)  Agitation conditions
     (11)  Emulsification.

Experimental Procedures in TPLS Optimization

     The criteria used to judge TPLS optimization were degree of recovery and selectivity.
Experimental results were evaluated by transmission electron microscopy.  In recovery studies,
aliquots from a standard chrysotile suspension were subjected to TPLS and recovery was
evaluated by comparing the quantity of extracted chrysotile to that in comparable aliquots not
subjected to TPLS.   By using heavy concentrations of chrysotile (^4.0 mg/1), it was possible to
make rapid comparisons of chrysotile concentrations without resorting to fiber counting.
Although this analytical procedure was quite effective,  no  numbers expressing recoveries were
derived.  However, when conditions were found that favored maximum recovery, atomic
absorption spectrometric analysis (AAS) for magnesium was run  to obtain data expressing
percentage recovery.  Again large quantities of chrysotile were employed to obtain the most
sensitive AAS measurement of recovery.

     Specimens for  TEM recovery analysis were prepared by depositing the chrysotile from an
aliquot of the standard suspension on a  0.45 //m, 25 mm,  Millipore filter, and depositing the
chrysotile separated by TPLS from a similar aliquot on another Millipore filter. A piece was cut
from each filter and placed deposit side down on an electron microscope specimen grid bearing
a carbon support  film. The Millipore filter material was dissolved by the Jaffe Wick Method
to prepare a specimen suited to viewing in the TEM where recovery was evaluated  by comparing
asbestos concentrations at 5000X magnification.

     A great many experiments were run before achieving  recovery approaching 100 percent.
Consequently, the recovery results of most experiments could be quickly judged by electron
microscopy. Specimen preparations were saved from the various recovery experiments in order
to be able to make  comparisons among results as  TPLS development progressed.

     After TPLS recovery was optimized and combined with light microscopic detection, re-
covery of those fibers detectable by light microscopy was  assessed by comparing light


                                            21

-------
microscopic counts on as-filtered aliquots to counts made on TPLS fractions from similar
aliquots.  The results are discussed in the report section dealing with determination of limits
of detection.

     Selectivity was assessed from the results of TPLS applied to synthetic mixtures of
diatomaceous earth and chrysotile as well as from results obtained when TPLS was run on
real water samples.  Interferences giving both high and  low results  were assessed.  TPLS
selectivity was not found to be as great a problem as recovery. Selectivity of the final TPLS
procedure for chrysotile in  the presence of diatomaceous earth is excellent; only an occasional
diatom was found among matted pure chrysotile fibers in the electron microscope preparations.

     TPLS selectivity when applied to real water samples is discussed  in a later section entitled
"TPLS Spot Test  Analysis of Real Water Samples".  Although interferences giving both high and
low results were identified in these studies, they did not prevent chrysotile detection in these
samples.

Effects of Critical TPLS Factors on
Selectivity and Percent Recovery

     The  most critical factors in TPLS  proved to be (1) mode of agitation, (2) type of
surfactant, (3) concentration of surfactant, (4) number of extractions from a single water
aliquot, (5) ratio of oil to water phase, (6) pH, and (7) the promotion of the reaction  between
the chrysotile surface and the surfactant.  Although the effects of  some factors overshadowed
others, all were found to be important in the optimization of recovery and selectivity.

     The  problems of recovery and selectivity were dealt with separately during experimenta-
tion primarily because there was no satisfactory procedure to evaluate recovery when there
were large amounts of interfering extraneous particulate in the original water suspension.  Con-
sequently, all recovery experiments were run on standard suspensions  of pure chrysotile.

     The  general TPLS procedure in its stage of development at the beginning of these optimi-
zation studies consisted of adding anionic surfactant to the water  phase, agitating the water
phase and oil phase together, salting out the emulsion that formed upon agitation, and after
phase separation, filtering the oil phase to concentrate  the extracted chrysotile on a filter.

     Progress toward optimization of TPLS recovery and selectivity is summarized in the fol-
lowing  discussions of effects of the critical factors.

TPLS Recovery

Effects of Agitation —
     The  mode of agitation is perhaps the most critical of all the investigated factors.  When all
other factors are  optimized, if agitation is inadequate, low recoveries are inevitable.

     First success was realized using ultrasonic  treatment.  Later, hand shaking in a separatory
funnel produced good results. After consulting the literature(6), it was decided to attempt  to
obtain better control of agitation by substituting mechanical stirring for shaking by hand.  Lai
and Fuerstenau reported «95 percent recovery of alumina from water suspension into  isooctane


                                             22

-------
using mechanical stirring and they were able to use it while varying other parameters to work
toward optimum recovery. The effects of stirring appeared to be quite reproducible.

     After switching to stirring, varying recoveries were obtained; some were quite high but
others were  low. This variability was attributed to other, possibly unidentified, parameters
when the manipulation of identified parameters did not consistently improve results.

     Mechanical stirring was performed at speeds of 600, 1200 and 2200 RPM and recovery was
found to increase with stirring  speed (Table 1).  However, stirring speeds of 2200 RPM and
above were impractical.  We settled on  1600 RPM, ran 18 experiments under identical condi-
tions, and analyzed recovery  by atomic absorption analysis for magnesium. A  mean recovery of
28% ± 14 was obtained.  This degree of recovery was considered to be too low and to have un-
acceptable reproducibility. Consequently, a superior means of agitation was sought.

     At this point, shaking by hand was tried again and better recoveries were  found.  A
mechanical shaker was used and consistently higher recoveries were obtained than by stirring.
Then differences were observed as factors other than agitation were varied. Ultimately, a
recovery of  74% ± 9.6 was obtained (Table 2) using shaking at 150 cycles per  minute for
5 minutes.  This was considered adequate for the purposes of the rapid analytical method.

Effect of pH -
     Hydrogen ion activity is not extremely critical in  the range from pH 3 to  7, and there is
no problem in maintaining the  pH in this range during TPLS.  However, to attain the highest
positive charge on the chrysotile surface to promote maximum reaction with the anionic
surfactant, the pH of the water phase was adjusted to  3.5 where the zeta potential has the
highest positive  value. This is accomplished usually by addition of 0.1 N HC1.   In all cases,
the pH of water samples has been found to be  above 3.5.

Effects of Surfactant Compound and Surfactant Concentration —
     Only one type  of surfactant has been investigated to any extent, this is the anionic alkyl
sodium sulfosuccinate.  However, two analogous compounds of this type have  been investi-
gated.  Because  of initial  success with these compounds, dioctyl sodium sulfosuccinate (Aerosol
OT and MO-70) and ditridecyl sodium  sulfosuccinate  (MT-70), no others were sought.

     There was  some evidence that MT-70 promoted slightly higher TPLS  yields than Aerosol
OT, but  later experiments indicated comparable recoveries with MT-70 and MO-70, the dioctyl
sodium sulfosuccinate supplied by Mona Industries Inc., Paterson, New Jersey  07524.  Greater
TPLS selectivity was observed while using MO-70.  Therefore, MO-70 has  been recommended
in the finalized  TPLS procedure.

     Surfactant  concentration is critical in TPLS, not  only from the standpoint of recovery but
also in that  too  high concentration produces emulsion  stability and phase inversion. Therefore,
concentration must be carefully adjusted to obtain acceptable TPLS recovery and to avoid
emulsification problems.

     While employing MT-70, optimized recoveries were observed to occur at concentrations
ranging from 1  x 10-3 M to 2.5 x 10~3 M. No phase inversion was detected at 1  x 10-3 M, but
it was occasionally seen at the  2.5 x 10~3 M concentration.  Recoveries were determined by


                                             23

-------
   TABLE 1. EFFECT OF AGITATION BY STIRRING
Run
3 1370-93 A
31370-93B
31370-85
32608-1
Stirring Speed, rpm
600
600
1600
2200
Percent Recovery Based
on Mg Analysis
7
7
25
39
TABLE 2. PERCENT RECOVERY WITH 2.5 X lO'3 M MT-70
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
Mean = 74%
% Recovery After Five
Consecutive Extractions
From Same Sample
77
59
81
64
67
54
77
72
86
81
72
69
99
Recovery
pH After Five
Consecutive
Extractions
4.5
4.7
4.7
4.3
4.4
4.5
4.5
4.3
4.4
4.5
4.6
4.3
4.4

Standard Deviation = ±9.6
                      24

-------
AAS analysis for magnesium at these two surfactant concentrations to determine whether there
was a significant difference in recovery.  Twenty-nine determinations were made at 1 x 10"3 M
MT-70 concentration and the  recovery was measured to be 63% ±11.  At 2.5 x 10~3 M con-
centration, it was measured to be 74% ± 9.6 (Tables 2 and 3).

     Because no further problems with phase inversion were encountered, the higher  surfactant
concentration is recommended in the description of the rapid analytical method.

Number of TPLS Extractions  From a Single Water Aliquot -
     In initial TPLS studies, only one extraction was made from the water aliquot.  Upon mak-
ing several extractions and separately filtering each extracted fraction, it was discovered that the
bulk of the chrysotile was not necessarily extracted in the first pass. The TPLS procedure was
modified to include three extractions.  As further studies were made of TPLS yields, it was
again modified to include five passes.

     Five extractions from the same water aliquot are recommended in the final TPLS pro-
cedure.  Under these conditions, 75  percent of the  chrysotile is extracted.

Ratio of Oil Phase  to Water Phase —
     The ratio of the volume  of oil phase to the volume of water phase is important with
respect to  phase inversion. The oil volume was varied from 25 percent of the total liquid
volume to 50 percent  (1:1 water-oil).  We finally settled on 100 ml of water phase and 50 ml
of isooctane.  Under these conditions satisfactory interfacial contact occurs and phase inversion
is, in most cases, avoided.

Promotion of the Reaction Between Chrysotile and Surfactant —
     It was found that TPLS  recovery was improved by heating the water phase after the addi-
tion of the surfactant.  Presumably, heating promotes the reaction between the chrysotile sur-
face and surfactant.  Perhaps  one of the mechanisms by which this occurs is the desorption of
interfering species from the chrysotile surface.

     Heating at 60°C is recommended in the routine TPLS method.

TPLS Selectivity

     Preliminary experiments to investigate  TPLS selectivity were run on synthetic mixtures of
chrysotile  and diatomaceous earth.  Diatomaceous earth was chosen because of the prevalence
of diatoms in water samples.  The recommended method is quite selective for chrysotile in  the
presence of diatoms.

     It is impossible to predict the interferences that might be  encountered when applying
TPLS to water samples; however, TPLS  has been run on water samples  from 26 different
sources. Some interferences were observed, but they were not extremely serious. All were re-
lated to lack of TPLS selectivity.  Interferences that could give both high and low results were
observed.

     High  results were obtained on some samples from which  organic filaments were extracted
by TPLS.  Nevertheless, these filaments  can be identified as not asbestos with a little experience.


                                             25

-------
TABLE 3. PERCENT RECOVERY WITH 1.0 X KT3 M MT-70
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Mean = 63%
Standard Deviation
% Recovery After Five
Consecutive Extractions
From Same Sample
71
68
61
50
75
75
82
89
82
46
55
82
65
68
71
48
48
63
55
55
55
63
68
51
40
55
55
63
63

= ±11
pH After Five
Consecutive
Extractions
4.8
4.7
4.8
4.2
4.3
4.6
4.7
4.8
4.8
4.4
4.6
4.7
4.8
4.5
4.6
4.6
4.7
4.5
4.6
4.6
4.6
4.7
4.7
4.7
4.8
4.8
4.7
4.8
4.8


                     26

-------
These filamentous structures are approximately 0.1 urn in diameter and have varying lengths.
One of their identifying characteristics is that along their lengths they have swellings 0.2 i*m in
diameter and approximately 0.5 fim long.  These swellings can be seen in the light microscope
with vertical illumination at 500X magnification after the filtered TPLS deposit has been carbon
coated. No other interferences giving high results were discovered.

     Two types of interferences produce low results.  One is an extracted platy mineral that
could obscure chrysotile fibers. These plates yield a diffraction pattern the same  as that of
talc.  In no case was a sufficient quantity extracted to prevent the detection of chrysotile.

     The other type of interference that will give  low results is related to the nature of the
chrysotile surface.  If magnesium has been leached from the fiber surface, the zeta potential
becomes negative, hence, it will not react with the surfactant.  Also,  coatings may form on the
fiber and thus prevent access of the surfactant. Observed in the  TEM, these coatings have low
electron densities indicating that they are probably organic.

     In the cases in which interference from organic material is a problem, heating after
surfactant addition improves recoveries of coated  fibers, but removal  of organic fractions by a
chemical treatment possibly would result in  further improvement.  Because of the time re-
quired, low-temperature ashing was not  considered during the development of the rapid analyti-
cal method.  Neither were "wet ashing"  techniques explored.  It is believed that any future work
should include the investigation of "wet ashing".

Handling Large Water Sample Volumes by TPLS

     It is desirable to apply TPLS to the particulate in relatively large volumes of water in order
to realize the most sensitive detection possible. One of  the approaches taken was to increase the
volume of water sample handled by TPLS.   This resulted in a corresponding increase in the
volume of isooctane and  an increase in the time required to filter the separated chrysotile.
Thus,  it appears that scaling up TPLS is not a  satisfactory solution.

     A scaled-up procedure, was used to  run 500-ml water aliquots.  The filtration time required
to filter the 500 ml of  isooctane from each  of 5 extractions through a 25-mm,  0.2-jum
Nuclepore filter using the standard Millipore reservoir was approximately  2-1/2 hours.  Two
anionic surfactants (MO-70 and MT-70)  were used in these experiments.  Scaling  up to handle
the 500-ml aliquots involved the use of  a 2000-ml separatory funnel equipped with a Teflon
stopcock to avoid stopcock grease.  Pure chrysotile was  suspended in 500 ml of filtered
deionized water; 250 ml of 0.01 M  surfactant  was added and the suspension was  sonicated
for 15 minutes. The suspension was heated  to 150°F while stirring and subsequently cooled
to room temperature.   Then, 250  ml of  water  and 500 ml of isooctane were added.  TPLS
was performed and the separated fractions were deposited on Nuclepore filters. Those separa-
tions using MO-70 (dioctyl sodium sulfosuccinate) were relatively free of NaCl crystals and
exhibited good recovery.   MT-70 (ditridecyl sodium sulfosuccinate) separations contained a
great many NaCl crystals; however,  chrysotile recovery was also quite good.

     Even if smaller oil-to-water ratios were used, the filtration time would still pose a problem.
Filtration and resuspension was found to be a  better way of handling large water volumes.
                                             27

-------
Filtration and Resuspension —
     Work has been done to determine the feasibility of filtering relatively large volumes of
water, removing the filtered deposit, and resuspending the particulate in smaller water volumes
to facilitate  the application of two-phase liquid separation.  Six different types of filters were
used, and removal of the deposits was tried using ultrasonic treatment and shaking.

     Standard chrysotile suspensions (200- to 500-ml aliquots) were filtered using MF-Millipore,
Duralon, Polyvic, Celotate, Nuclepore, Gelman Metricel  (DM 45), and Mitex filters.  Fluoropore
filters were also tried, but water would not pass through them without the addition of alcohol
or other liquid to promote wetting.  Ultrasonic treatment disintegrated the MF-Millipore,
Duralon, Polyvic, and Celotate filters; however, the best recoveries were obtained by ultrasonic
treatment of Nuclepore, Mitex, and Metricel DM  450 filter deposits.  Recoveries from Metricel
DM  450 and Nuclepore filters were best of all and were judged by microscopic examination to
approach 100 percent. Some Teflon fibers were dislodged from Mitex filters by ultrasonic
treatment, but, despite this problem, the redeposited chrysotile was quite visible.

     Shaking, although not as effective in particle removal, did  not break any of the filter types.
In all cases,  some of the chrysotile was removed, and was found to  contain no extraneous
debris when redeposited.  A means of agitation employing less energy than the particular ultra-
sonic treatment used and greater energy than shaking would likely make it feasible  to remove
the major quantity of deposited particulate from any of the filter types.

     TPLS did not separate the chrysotile from filter fragments; both were transferred to the
isooctane.  Chrysotile removed ultrasonically  from Mitex, Metricel DM 450, and Nuclepore
filters was separated  from the water phase by TPLS.  Although Nuclepore is one of the two
best filters from the  standpoint of particle removal, it is fragile and difficult to handle in a field
operation.  The procedure found most effective for the  removal of the filtered deposit in
preparation  for TPLS was to add 50 ml of 0.005 M MT-70 to the filter,  deposit side down,
in a 250-ml flask and to  sonicate for 15 minutes; 25 ml of deionized water was then used to
rinse the filter as it was removed from the beaker, bringing the volume to 75 ml.  At this
point, the resulting suspension was heated to 150°F while stirring to promote the reaction
between the  surfactant and the chrysotile surface. Next, the volume was increased to 100 ml
by adding deionized  water to make the MT-70 concentration 2.5 x  10-3  M; then 50 ml of
isooctane was added  and the TPLS procedure was carried out.  Five extractions were made and
the separated fractions were combined on one 0.2-Mm-pore-size Nuclepore filter. Evaluations
of recovery  were made on carbon-coated deposits with the light microscope, or on  Jaffe
preparations  in the electron microscope.

     Twenty-four experiments were made in which filter deposits were removed ultrasonically
and  subjected to TPLS. Good recoveries were obtained.

     Metricel DM 450 filters were the best found. They are easily handled in the field,
particles are readily removed from the filter surface by ultrasonification,  and they do not
break up in  the ultrasonic bath.  The DM 450 filter has a 0.45-Mm pore size and it is composed
of a PVC copolymer.  The diameter of the DM 450 filters used  was 47 mm. Larger diameters,
of course, could be used  to filter larger quantities of water or water containing heavy concen-
trations of suspended particulate.
                                            28

-------
     If large water samples that are heavily loaded with particulate must be handled, several
47-mm-diameter filters can be used.  The collected particulate can be removed ultrasonically in
a 100-ml volume of water containing the anionic surfactant.

     Filtration and resuspension is recommended as an alternate procedure in the TPLS - light
microscopic analytical method.  When greater sensitivity is desired than is possible by extracting
a 75-ml aliquot, then it is suggested that filtration and resuspension be employed.

     Another  advantage of filtration and resuspension is that samples can be collected and
filtered soon afterwards. This would minimize sample storage problems.

TPLS Recovery as a Function of Particle
Size and Aspect Ratio

     Experiments were conducted to determine whether TPLS favors the extraction of certain
chrysotile fiber sizes and aspect ratios.  Several groups of experiments have been run investi-
gating TPLS particle size recovery; however, the set of experiments that  used the finalized
 TPLS conditions are the most pertinent.  This last set of  experiments also was designed to
determine the effects on particle size recovery as a function of zeta potential.

     Four samples of chrysotile having different zeta potentials were supplied by Mr. George
 Reimschussel  of Johns Mansville  Research and Engineering Center.  Mr. Reimschussel measured
their electrophoretic mobilities as listed below:

                    Sample          Electrophoretic Mobility at pH 10
                     96-1                        +3.9 MV
                     96-2                        +0.8 MV
                     96-3                        -1.2 MV
                     96-4                        -2.6 MV

 No electrophoretic mobilities were measured at lower pH conditions.  Nevertheless, they must
 all have positive zeta potentials at pH 3.5-4. Good TPLS recoveries were obtained in this
 pH range.  The existence of  zeta potential variations among chrysotile types makes it important
 to control the pH during TPLS.

      Standard chrysotile suspensions were prepared from  samples 96-1, 96-2,  96-3 and 96-4,
 and equal (50 ml) aliquots were  taken from the same suspension.  One aliquot was merely
 filtered onto a 25-mm, 0.2-Mm-pore-size Nuclepore filter; three other aliquots from the same
 suspension were subjected to TPLS and  the extracted chrysotile was also deposited on the
 same type Nuclepore filters.  Specimens for TEM were prepared from  the filtered deposits by
 carbon coating followed by dissolution of the filter material by the Jaffe Wick Method.

     The following TPLS procedure was used.

       1.  To  a 75-ml water aliquot adjusted to pH 3.5 with 0.1 N HC1, add  25 ml of
          0.01 M MO-70, heat to 140°F while stirring and cool to room temperature.
                                             29

-------
      2.  Combine resulting solution with 50 ml of isooctane in a 250-ml separatory
         funnel.
      3.  Shake the isooctane and water phase together in the separatory funnel at
         150 cycles per minute for 5 minutes.
      4.  Add  10 ml of  10 percent NaCl and shake gently to break the emulsion.
      5.  When the phases have separated drain off the water phase completely allow-
         ing a small amount of isooctane to drain into the water.
      6.  Pour the isooctane into a beaker from the separatory funnel avoiding water
         droplets that may remain in the bottom or on the sides of the separatory
         funnel.
      7.  Filter the isooctane onto  a 25-mm, 0.2-/um-pore-size Nuclepore filter using
         the standard filter reservoir.
      8.  Repeat the extraction procedure four times filtering the isooctane each time
         through the same  filter.
      9.  Dry filter and coat the deposit side with carbon by vacuum evaporating  1  cm
         of carbon rod  1 mm in diameter at a distance of 15 cm.
     10.  Prepare TEM specimens by the Jaffe  Wick Method.

     Specimens were prepared for electron microscopy by carbon coating the filtered chrysotile
deposits and dissolving away the Nuclepore filter by the Jaffe technique.  TEM fiber counts and
measurements of fiber lengths and widths were made of at least 100 fibers in each preparation.
A computer was used to assign lengths,  widths, and aspect ratios to size classes and to plot
each versus cumulative number percent  and cumulative volume percent.

     Upon examination of the plots, it was evident that the cumulative volume percent plots
were too greatly affected by occasional  fibers having unusually large diameters.  Also, the
widths of fibers extracted by TPLS did  not differ significantly from those not subjected to
TPLS. Consequently, it was decided to include in this report only two sets of plots for each
sample, i.e., length versus cumulative number percent and aspect ratio versus cumulative num-
ber percent.  Each plot presents the results of four determinations; one of the as-deposited
chrysotile, and three of TPLS extractions from comparable aliquots.  The plots are designated
as Figures 8 through 15.

     The results show that TPLS favors  the extraction  of the longer fibers with larger aspect
ratios. In this respect, these latest results are consistent with those previously obtained under
different  TPLS conditions. This is acceptable in the sense that it is more desirable that the
more  visible larger fibers are selectively extracted when light microscopic detection is used.
In addition,  investigators of  health effects have reported that the longer fibers are more
carcinogenic than the  shorter ones.
                                            30

-------
33.33
99.9
99
«- 98
£ 95
o
fc 90
o.
80
fc" 70
-£ 60
| 50
Z 40
„> 30
(1)
-> 20
1 '°
0 2
1
0.5
°0'.f
0.05
0.0.Q
—
	
—
—

• AA

• A A
• At
• AA
- *A AA
•AA
- A A

•A
•A
—
—
-
	
-
!i i i i l l 1 1 i i l i i i 1 1 1
2 05 1.0 2.0 5.0 10 2C
                  Length,
 Figure 8. Distribution of chrysotile fiber
 lengths in sample 96-1 before (•) and
 after (A) TPLS.
33.33
99.9
99
-H- 98
£ 95
5 90
o.
80
of 70
-§ 60
§ 50
Z 40
a, 30
~ 20
0
1 '°
3 5
° 2
0.5
0.2
0§>5
o.ob
—
—
—
—
• AA A
• AA
• AA
• AA
— • At
• A A
• AA
•A AA
• AA
—
1 1 1 1 1 1 ll 1 1 | 1 1 1 1 1 1
2 0.5 1.0 2.0 5.0 10 2C
                  Length, fj.m

Figure 10.  Distribution of chrysotile fiber
lengths in sample 96-2 before (•) and
after (A) TPLS.
                                             c
                                             0>
                                             o
                                             Iw
                                             03
                                             Q.
                                             -0
                                             E
                                             3
                                             3

                                             3
33.33
99.9
99
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5
0.2

0.05
0.01

—
	
—
—
• m.
• AAA
• AAA
• A A
— «AA
•AAA
•tA
•At
•A*
—
-
—
-
_
^~
-
1 1 1 1 1 1 ll 1 1 1 I 1 M ll
          5.0   10   20     50
               Length/Width
                                                                                100   200
 Figure 9.  Distribution of chrysotile fiber
 aspect ratios in sample 96-1 before (•)
 and after (A) TPLS.




"E

a.
_
<5
JO
E
3
Z
0)
~
0
3
E
3
O




33.33
99.9
99
98

95
90

80
70
60
50
40
30
20

10

5

2
0.5
0.0 12

—
	
	

—
• A

• AA
• A
— A A^A
	 A\ AA
• AAA
• AA
• A

• AA

—

—
-
l 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ll
.0 5.0 10 20 50 100 20
Length/ Width
Figure 11.  Distribution of chrysotile fiber
aspect ratios in sample 96-2 before (•)
and after (A) TPLS.
                                          31

-------
99.9
99
98

S 95
fc 90
0.
80
a3~ 70
-£ 60
| 50
Z 40
 20
1 10
E 5
13 J
0 2
1
0.5
0.2
O.I
0.05
001
—
	
—

• —
• A A

• A AA
• AM
• AM
— • AAA
• AAA
• A A A
• AAA
•MA
	

-
—
-
-
—
-
l i i l 1 l 1 1 l 1 1 1 1 l 1 i 1



c
O)
o
k 8.

1__
co
.a
E


co
'•g
I
3
O







0.2 05 1.0 2.0 5.0 10 20
Length, ^.m
Figure 1 2. Distribution of chrysotile fiber
lengths in sample 96-3 before (•) and
after (A) TPLS.
99.99
99.9
99
*- 98
c
v 95
o
S3 90
Q.
80
of 70
-g 60
i 50
Z 40
a, 30
~ 20
"5 10
1 5
o
2
0.5
0.2
0.05


—
—
—

—

• A A

• MA
• M
• M
— • M
• M
- tMft.
• M\
•Wk
—

~
-
L
i i i i 1 1 1 1 i i i i i i 1 1 1




"c

•^
-5
E
(_j




b.2 0.5 1.0 2.0 5.0 10 20
                 Length, fj.m
Figure 14.  Distribution of chrysotile fiber
lengths  in sample 96-4 before (•) and
after (A) TPLS.
33.33
99.9
99
98
95
90
80
70
60
50
40
30
20
10
5
2
1
0.5
0.2
0.05
0.0 12

—
—
—
—
• AA,
• A M
• AA A
• AA A
- • AA A
• M A
• A A
• A AA
• 4*
—
-
—
-
-
-
i i i i 1 1 1 1 i i i i i i 1 1 !
0 5.0 10 20 50 100 20
                                                              Length/Width
                                               Figure 13.  Distribution of chrysotile fiber
                                               aspect ratios in sample 96-3 before (•)
                                               and after (A) TPLS.
33.33
99.9
99
98
95
90
80
70
60
50
40
30
20
10
5
2
0.5
a?
0.05
nni
—
	
—
—
•4A
• MA
•tA
A JM
^K ^^^V
A 4k
•i
• M
—
-
_
1 1 1 1 1 1 1 1 1 1 1 1 l 1 1 1 1
                                                                                    200
               Length/Width
Figure 15.  Distribution of chrysotile fiber
aspect ratios in sample 96-4 before (•)
and after (A) TPLS.
                                         32

-------
PHASE II:  INVESTIGATION OF DETECTION METHODS

Change of Flow Rate as a-Function  of Chrysotile Concentration

     A survey of detection methods was made in an effort to provide rapid semiquantitative
analysis of chrysotile separated and concentrated by TPLS.  The methods investigated were
change in flow rate of a liquid through a filter as a function of the deposited mass, spot-test
procedures, X-ray diffraction, and light microscopy.  The objective was to develop a simple
routine detection procedure with sufficient sensitivity to detect asbestos concentrations below
the significant level in order to eliminate insignificant samples from electron microscopic
analysis.

     No attempt was made to confine detection to fiber counting.  Rather, a search was made
for a bulk detector that could be used without the need for a high degree of skill to screen large
numbers of water samples rapidly.  This was the reason for  the investigation of a method so
simple as that which attempted to correlate changes in flow rate through a filtered chrysotile
deposit with the mass of chrysotile extracted by TPLS.

     The flow rate of a liquid through a filtered deposit becomes slower as the mass of
deposited particulate becomes greater.  It was believed that  this phenomenon could provide
the basis for a rapid detection method.

     Experiments were run to determine the flow rates through measured quantities of
chrysotile deposited on 25-mm-diameter Nuclepore (0.1  jum pore size) filters and dried.
Initial experiments were run using the standard Millipore 25-mm filtering apparatus.

     The measurements were made by introducing  15 ml of isooctane into the reservoir,
applying the vacuum, and recording the time required for the isooctane to pass through the
filter.  The pressure drop was noted for each point.

     Quantities of chrysotile ranging from 5.0 to 40.0 jug were used.  Both pressure drop across
the filter and flow rate were measured.  The plot of pressure drop and flow rate  versus
chrysotile concentration is shown in Figure 16. Pressure drop was not a very sensitive indi-
cation of chrysotile concentration.   Flow rate was  a more sensitive measurement, but not
sufficiently so under these experimental conditions.

     Further experimentation to obtain more sensitive measurements by flow rate resulted in
detection of quantities as low as 0.5 /xg-  This was  accomplished by making a special filter
reservoir with a 3-mm hole in order to confine the filtered  chrysotile to a smaller filter area,
thus increasing the change in flow rate for a given  amount of chrysotile.

     Attempts to extend detection below  0.5 /ug were not successful; furthermore, when analysis
of the TPLS fraction from a real water sample was tried, a  relatively  small amount of inter-
fering platy mineral particles gave too high results. The 0.5 Mg detection limit was not con-
sidered to be adequate. No further effort was made  to use this approach.
                                             33

-------
   no

   100

   90

tn  80

§  70
o>
tn
 ~ 60
4, and (4) detection of reacted manganese by the periodate-tetrabase spot test.
                                            34

-------
     The detailed spot-test procedure is as follows:

     1.  Chrysotile as extracted by TPLS is deposited on a 25-mm, 0.45-jum-pore-size
        Millipore filter in an ^S-mm-diameter area and dried.  This is done by using
        a special reservoir with a 3-mm-diameter opening.

     2.  The special reservoir is then removed and the filter bearing the deposited
        chrysotile is covered with a  Millipore filter, and the standard 25-mm reservoir
        is used for Step  3.

     3.  Fifteen ml of 25% MnSC^-P^O at room temperature is poured into the filter
        reservoir and aspirated through the deposit over a period of 1 minute.
     4.  Next, the  filtered deposit is washed with 15-ml quantities of room temperature
        deionized water  5 times aspirating each washing completely through the filter.
     5.  The filter reservoir is removed while applying suction to the filter to hold it
        to the frit and then  the covering Millipore filter is removed with tweezers.
     6.  The filter bearing the chrysotile deposit is removed and dried under a heat
        lamp  at 145°F (63°C).
     7.  The filter is attached by the edge  to a piece of white blotter paper deposit
        side up and sprayed  with the spot testing working solution and placed under
        the heat lamp at 145°F until development is maximized («*2 minutes).
     8.  The depth of blue color  that develops is compared to photographs of a  series
        of standards  made using  known quantities of chrysotile.

     The working  spot test solution  is made  up by combining equal volumes of the Solutions A
and  B:
     Solution  A:  1.0% solution of tetrabase in 95% ethyl alcohol.
     Solution  B:  A saturated solution of potassium periodate in 2N acetic acid.

     Comparison of the  depth of color of  the spot test result is possible over a chrysotile con-
centration range from 0.1 ^g to «0.8 ng.   At 0.8 /ug and above the color  is so intense that little
change with increase  in concentration can  be detected.  Nevertheless, when used primarily to
select samples for  TEM analysis the  lack of good quantitative indications above 0.8 fig poses no
problem.

     The spot test procedure has the advantage that  it can be carried out  quickly with a mini-
mum of expensive apparatus and training;  but the preparation, unlike that used  for light
microscopic detection, cannot be readily used for electron microscopy when confirmation of
results is desired.

X-Ray Diffraction

     The techniques of powder X-ray diffraction were explored in an effort to establish a
detection  procedure.  The primary difficulties were quantification and sensitivity of detection.

     The X-ray powder camera was  used in most of the experiments; diffractometer methods
were not considered to have  sufficient sensitivity.  The major problem was with specimen

                                             35

-------
preparation when dealing with lower quantities of chrysotile (<30 ng). No way was found to
prepare a specimen so that it would be entirely irradiated by the X-ray beam.

     Light microscopy was found to be far superior to X-ray diffraction.   Fibers detectable by
light microscopy have always been found in water samples containing  chrysotile even when the
mean fiber length is below 1.0 /um.

Light Microscopy

     It was found  that light microscopy employing vertical illumination as obtained with metal-
lographic optics is an  extremely sensitive means of detecting fiber shapes  in a carbon-coated
TPLS deposit on a Nuclepore filter.  When interferences from particulate  other than chrysotile
are virtually eliminated by TPLS, the separated chrysotile fibers are quite visible.  The carbon
coating increases the detectability of the chrysotile by two mechanisms; it enhances image
contrast by increasing the reflectivity of the Nuclepore substrate and also increases the apparent
fiber diameter to  the point that some of the fibers with  diameters too small to be seen by light
microscopy become detectable if their lengths are greater than  «1.0 Mm-  The fibers are seen as
dark structures against a bright background  because the microscope illumination is along the
axis of the objective lens.  The Nuclepore filter surface is perpendicular to the objective axis
reflecting light  directly back into the  objective and the fiber surfaces are not perpendicular
reflecting light  away from the objective.

     A quick examination with the light microscope is all that  is required to determine whether
a filtered chrysotile deposit is greater than 100 nanograms.  Analysis of intermediate concentra-
tions is performed with an eyepiece reticule having 10 concentric circles to minimize fiber
counting time as fiber concentration varies (Figure 17).  When  concentrations are relatively
high, counts are made within circles near the center, and lower concentration counts are made
within circles of greater diameter — sometimes across the diameter of  the filter deposit.

     Light microscopic detection of chrysotile was placed on a quantitative basis by generating
a standard curve relating chrysotile mass per filter to counts made at 500X magnification while
viewing a carbon-coated Nuclepore filter with vertical illumination.  Counts are made while
scanning a 250-jum-wide path across the diameter of the effective filter area.  The eyepiece
reticule having  10 concentric circles spaced  12.5 p.m apart is employed (Figure 17), and one
count is registered for a fiber segment equal to the spacing between concentric circles.  Compen-
sation  for variations in fiber width are made by estimating the  number of minimum detectable
fiber widths in  larger  fibers or bundles.  Because  the outermost circle  delineates a 250-/um-
diameter field at 500X magnification, it is used to determine the path width over which counts
were made.  Counts are made only on fibers falling within that field as the path is traversed.

     Standard chrysotile suspensions were filtered onto Nuclepore filters using the standard
25-mm Millipore filtering  apparatus.  Concentrations ranging from  1.0 nanogram to 10,000
nanograms were deposited on Nuclepore filters, which were subsequently  taped flat on micro-
scope slides and coated with  carbon.  At least six filtered deposits were prepared at each of
seven concentrations from two  sources  of pure chrysotile.

     Two persons have made counts on several of these preparations and  the reproducibility
between the two sets  of results appears to be quite acceptable  as demonstrated by  the  agreement


                                            36

-------
Figure 17.  The eyepiece reticule used in fiber counting.

-------
 of data. Counting a preparation requires approximately 15 minutes when dealing with
 100-nanogram concentrations where the total number of counts is «1500. When higher
 concentrations are encountered, various sized fields may be selected from the set of concentric
 circles in the eyepiece reticule to  count smaller areas and the data may be normalized by
 multiplying the results by the appropriate area factor.

     The standard curve that was thus generated is presented in Figure  18. Counts versus
 chrysotile mass per filter are shown on a  log-log plot. The sensitivity of light microscopic
 detection appears quite adequate  for the  purposes of the rapid analytical method.  Also, the
 time per analysis is sufficiently short, especially if the method is to indicate only if a sample
 should  be analyzed more precisely by electron microscopy. Then an upper count limit may be
 set according to allowable chrysotile concentration to minimize the required counting time.

 Determination of Percent TPLS Recovery of Particle
 Sizes Detectable by Light Microscopy —
     The same preparations used to determine the recovery of chrysotile by TPLS by TEM
 methods were also used to  determine  the recovery of particle sizes detectable by light
 microscopy. The counting procedure described in the preceding section was applied to com-
 parable aliquots before and after TPLS.

     Results are presented in Table 4.  The percent recovery as determined from three separate
 runs from each of the four samples are quite comparable to the percent recovery as determined
 by atomic absorption speclrometry.  When all results were combined, the mean and standard
 deviation were calculated to be 75% ± 15.  Consequently, when applying light microscopic
 detection to TPLS fractions, it  is  suggested that a factor of 2 be applied to the counts made
 with the light microscope to compensate  for fiber loss during TPLS.

         TABLE 4.  PERCENT TPLS RECOVERY OF PARTICLES DETECTABLE
                    BY LIGHT MICROSCOPY
       Sample         Run 1         Run 2         Run 3        Mean ± Std. Dev.
96-1
96-2
96-3
96-4
92
76
97
74
83
92
57
64
65
72
49
81
80 ± 13
80 ± 10
67 ± 25
73 ± 9
Limits of Light Microscopic Detection —
     Limits of detection have been determined at 80, 90, 95, 98 and 99 percent confidence
levels.  The data used in the calculations were obtained by  the light microscopic counting
technique.  Counts were made on 22 different concentrations ranging from 0.8  to 1000 nano-
grams. An average of approximately six preparations at each concentration were counted.
These data  were analyzed by a computer program to determine the limits of detection of
chrysotile per Nuclepore filter deposit.  Table 5 presents the detection limits at 100% TPLS re-
covery and also at 75% ± 15 TPLS recovery. The latter recovery was determined by a com-
parison of light microscope counts made on as-deposited aliquots versus similar aliquots from

                                           38

-------
the same standard suspensions subjected to TPLS.  At 60% recovery (75% -  15) the limit of
detection at the 99% confidence level was computed to be  1 nanogram per filter deposit.

                   TABLE 5.  DETECTION LIMIT DETERMINATIONS


     Percent            Detection Limit (/ug)               Detection Limit (/xg) at
Confidence Level	at 100% Recovery               75% ± IS TPLS Recovery
75%
80
90
95
98
99
2
2
3
5
5
.2 x
.7 x
.4 x
.3 x
.2 x
ID'4
JO'4
ID'4
ID'4
ID'4
2
3
4
5
6
.9 x
.6 x
.5 x
.7 x
.9 x
ID'4
lO-4
ID'4
lO-4
10-4
2.4
3.0
3.8
4.8
5.8
+15
x
x
X
X
X
lO-4
ID'4
lO-4
ID'4
10-4
-1
3.7 x
4.5 x
7.5 x
2.6 x
8.7 x
5
JO'4
ID'4
ID'4
ID'4
ID"4
     The data points used in the calculation of detection limits are tabulated in Table 6.  These
data were also used to generate the standard curve.

PHASE III:  APPLICATION OF ANALYTICAL PROCEDURES

TPLS — Light Microscopic Analysis of Real Water Samples

     Seventeen water  samples were obtained from Dr. Ian Stewart of W. C. McCrone Associates
and from  Dr. Eric Chatfield of Ontario  Research Foundation. All were subjected to TPLS and
the extracted fractions were analyzed by the light microscopic counting technique (see Fig-
ure 18), and some of  them were analyzed by the spot test procedure.  Specimens from these
samples were also examined by  TEM before and after TPLS to determine whether chrysotile
was present and the nature of particulate extracted by TPLS along with chrysotile.

     In the as-received condition, many of the samples contained large quantities of extraneous
particulate from which it was necessary to  separate the chrysotile.  First, TEM comparisons
were made between two-phase liquid separations from 50-ml aliquots and filtered deposits
prepared from 10-ml aliquots of the as-received samples to determine whether there was agree-
ment between the presence and concentration of chrysotile in the as-received sample and the
corresponding TPLS fraction.  The same preparations were also examined by light microscopy
to evaluate the feasibility of using the light microscope for rapid  screening in a detection
procedure.

    Most of the samples had a  pH of 7 or higher, which is too high to obtain optimum TPLS
results.  Consequently, the pH of the 75-ml aliquot was adjusted to pH 3.5  with 0.1  N HC1
where the positive zeta potential of chrysotile is maximized.  Then, 25 ml of 0.01 M MO-70
was added, the sample heated to 140°C, cooled to room temperature, and 50 ml of isooctane
added.  TPLS was then carried out and  the extracted chrysotile was deposited over a  16-mm-
diameter filter area.

                                          39

-------
TABLE 6. LIGHT MICROSCOPIC DETECTION OF CHRYSOTILE AT TWENTY CONCENTRATIONS
Concentration
(Mg) Avg
0.00079
0.001
0.0018
0.005
0.0079
0.009
0.01
0.018
0.02
0.025
0.036
0.045
0.079
0.1
0.18
0.2
0.36
0.4
0.72
0.79
0.8
1.0
5.67
8.8
31.7
81.8
93.3
157.3
119.2
260.3
272.4
235.5
350.5
527.8
959.7
1356.3
1679.8
2188
6908.3
5953.2
12910.4
12426.7


6
10
40
78
62
137
139
228
338
225
410
583
972
1,435
1,708
1,951
7,298
6,232
13,202
16,705
10,496
12,300
6
6
29
72
91
128
139
272
294
244
368
486
1,076
1,064
1,924
2,425
8,528
5,904
12,218
10,707
7,954
13,776
5
7
6
93
118
186
95
386
286
264
310
522
831
1,433
1,852

5,166
5,904
11,726
9,868
13,366
12,710

8
29
88

164
77
210
251
231
296
500

1,316
1,780

4,592
5,986
13,858

11,726
16,400

5
38
86

137
116
238
225
259
392
590

1,600
1,146

6,658
5,740
1 1 ,070

12,628
11,480
Counts

8 13 11 9 11
48
86 84 70 69 102 79 75

192
149
228
254 259
190
327
486

1,290
1,669

7,298 5,986 7,462 6,806 8,118 8,364 6,624

15,744 12,628 12,464 13,284

10,414 9,184 13,694 9,184 13,038 11,168.4
12,054 13,448 16,646 15,990 17,056 15,006 14,260.5

-------
                1,000,000 p
                  100,000 —
                   10,000  —
                §  1,000
                o
0.0001   QOOI    0.01      O.I
             Pure Chrysotile,
                                                       1.0
10
              Figure 18.  Standard curve that relates light microscope
                          counts to mass (f/g) of chrysotile per filter.

Results are presented in Table 7.  In general, the results obtained were as follows.

1.  In no case was chrysotile found  in the as-received preparation and not found in
   the TPLS fraction by TEM and light microscopy.
2.  In five cases, chrysotile was found  in the TPLS fraction and not in the as-
   received preparation; this was attributed to masking by extraneous particulate.
3.  Chrysotile was detected in only  two of the as-received preparations by light
   microscopy.  This was also the result of masking by other particulate.
4.  Filamentous structures, which appear to be organic when seen in the TEM, are
   present in four of the TPLS fractions. To the inexperienced observer using
   light microscopy, these filaments might be identified as chrysotile to give false
   positive results.
5.  In two cases, some interference  was encountered from other particulate being
   transferred along with chrysotile to the isooctane; however, chrysotile was
   detected more readily by  TEM and LM in TPLS preparations than in the filter
   deposits of the corresponding  as-received  samples.
6.  Detection of chrysotile in TPLS fractions from real water samples was found
   feasible using the light microscope.
7.  Light scattering, as seen with the unaided eye, from carbon-coated deposits of
   the TPLS  fractions were observed  to correlate fairly well with the presence of
   chrysotile. False positive  results are more likely than false  negative results when
   relying on light scattering.
                                         41

-------
                                          TABLE 7.  RESULTS OF TPLS APPLIED TO REAL WATER SAMPLES
            Water Sample Number
      Observation in TEM of
         TPLS Fraction
TEM Observations of
 As-Received Sample
Debris    Chrysotile
             LM Observations of
               TPLS Fraction
                             Remarks
to
       C3 0514 1530 U3WGO 11/17 I
        (filter sample WG3)
      C3 0514 1530 D3 WGO 11/17 II
       (filter sample WG3)
      C9 0305 1215 U4 WGO 11/17 III
       (filter sample WG3)
      C9 0305 1350 U5 WGO 11/22 I
        (filter sample WG1)
       C9 0305 1340 D5 WGO 11/17 IV
        (filter sample WG3)
       P4 0212 0920 D5 WGO 11 /15 II
        (filter sample WG2)
Long filamentous organic structures   Heavy
present; very little chrysotile seen
Many bacteria-like structures plus
long organic filaments. Medium
amount of chrysotile

No chrysotile, a few organic fila-
ments,  and very little extraneous
particulate

Many clumps of chrysotile, a fairly
clean separation containing a few
bacteria
A clean separation containing many
short chrysotile fibers
Very few chrysotile fibers; some
extraneous organic particulate
          Little
Medium   None
Little
None
Medium   None
Little
None
Very
heavy
None
             Long filaments seen,
             but no chrysotile
Medium number of
fiber-shared struc-
tures seen

Very few fiber
shapes
             Many individual
             fibers and several
             clumps of fibers
             detected
Many short fibers
seen
Very few fibers
seen
With some experience, these
filaments would not be con-
strued as chrysotile when
seen with the light
microscope

Both chrysotile fibers and
organic filaments are visible
in the light microscope

Not a significant population
of fiber shapes
Relatively high fiber popula-
tion detected by light
microscopy indicating pos-
sibly significant quantity of
chrysotile

The medium to high con-
centration of short fibers
was unique to this particular
sample; despite this char-
acteristic, many fibers were
detected by light microscopy

No significant number of
fibers shapes were seen by
light microscopy

-------
                                                           TABLE?. (Continued)
     Water Sample Number
      Observation in TEM of
         TPLS Fraction
TEM Observations of
As-Received Sample
Debris    Chrysotile
LM Observations of
  TPLS Fraction
         Remarks
M3 0307 0730 Dl WIO 11/151
 (filter sample WIe)
M30307 1500D2WGO11/15IV
 (filter sample WG2)
PI 0130 0555 D2 WIO 11/15 III
 (filter sample WI2)
 R2 0305 0615 Dl WIO 11/10 III
  (filter sample WI3)
 SI 0326 0700 D4 WIO 11/10 IV
A very clean heavy mat of chrysotile  Heavy     Heavy
fibers too thick to penetrate with
electron beam except at edge of
filtered deposit
A heavy mat of clean chrysotile       Medium   Very
fibers too thick to penetrate with     to heavy   heavy
electron beam except at edge of
deposit
A little extraneous organic par-
ticulate; very few chrysotile fibers
Medium    None
A clean separation containing a       Very
few long chrysotile fibers; very        heavy
little extraneous particulate
 A heavy mat of chrysotile coated     Very
 with a structureless organic film      heavy
          Very few
          (other fiber
          types
          present)
           None
           visible
                        Mat of fibers
                        Heavy mat of fibers
                        seen
Almost no fibers
seen
A few long fibers
Heavy mat of fibers
plus heavy extraneous
particles
The TPLS fraction when de-
posited on the Nuclepore
filter contained so many
fibers that individual fibers
could be discerned only at
the edge  of the deposit

A very dense chrysotile mat
from the TPLS fraction in
which individual fibers were
evident only at the edge
of the filtered  deposit

No  significant numbers of
fiber shapes seen in the
light microscope

Although the as-received
sample contained other fiber
types than chrysotile, only
chrysotile was  detected in
the  TPLS fraction

This sample is one of two
in the sample set that con-
tained a large quantity of
interfering material in the
TPLS fraction

-------
                                                           TABLE 7.  (Continued)
     Water Sample Number
      Observation in TEM of
         TPLS Fraction
 TEM Observations of
 As-Received Sample
 Debris     Chrysotile
              LM Observations of
                TPLS Fraction
                              Remarks
Gl 0505 0450 D4WIO 11/8 II
 11/22 II (filter sample WI3)
N2 0302 1805 D3 WGO 11 /10 II
 (filter sample WG1)

N8 0403 1730 D6 WGO 11/101
 (filter sample WG2)
Nl 0606 1335 D2 WGO 11/29 I
 11/8 I (filter sample WG2)
Chatfield No. 1
 (US sample)
Chatfield No. 2 & 3
 (Canadian sample)
No chrysotile and very few
extraneous particles
No chrysotile found, but organic
filaments were present

No chrysotile found, but organic
filaments were present
A medium to heavy amount of
chrysotile, also some platy
mineral particles
Heavy
None
Medium    None
Heavy
None
Heavy
None
Gross numbers of chrysotile fibers
and bundles of fibers; a clean
separation

Many chrysotile fibers plus several
diatoms and other unidentified
structures transferred along with
chrysotile
Heavy
Very
heavy
Medium
Heavy
A few fiber shapes
seen
A few fiber shapes
seen

A few fiber shapes
seen, but not con-
strued as chrysotile
Many fibers detected
Dense population of
chrysotile
Heavy fiber con-
centration plus
diatoms and other
particles
The few fiber shapes de-
tected in the light micro-
scope gave a weak false posi-
tive result

A weak false positive result
Although fiber shapes were
detected in the light micro-
scope, they did not have the
appearance of chrysotile

A case in which the debris
was so heavy in the as-
received preparation that no
chrysotile was detected
despite a relative high
chrysotile concentration

A very good separation  of
gross amounts of chrysotile
This is the other sample of
two in the sample set that
contained a fairly heavy
concentration of interfering
particulate in the TPLS
fraction

-------
     8.  TPLS procedures will extract wide ranges of chrysotile concentrations from
        real water samples.  Some extractions resulted in mats of fibers too thick
        to see through with either the TEM or LM, whereas, others contained only
        a few or no fibers.

     It was concluded from this study of the application of TPLS to actual water samples, that
TPLS is an effective approach to more sensitive detection of chrysotile by TEM or LM in water
samples than in  transmission electron microscopy applied to unashed filtered deposits of as-
received  samples.  This is true even under the conditions of incomplete TPLS chrysotile
recovery. Also,  light microscopy detected extracted chrysotile from samples in which both
LM and TEM examination of as-received filtered deposits revealed none.

     Light microscopic counts were made on the carbon-coated TPLS deposits on the 0.2-Mm-
pore-size Nuclepore filters.  Although possible interferences existed in a few samples in the form
of a platy mineral and filamentous organic structures that might be identified as chrysotile by
an inexperienced observer, they did not obscure the presence of significant quantities of
chrysotile or contribute to erroneously high counts. The results are presented in Table 8.

     Four samples contained gross amounts of chrysotile making it impossible to perform
counting. An example of the concentration of fibers is shown in Figure  19. This is a photo-
micrograph taken at the edge of the filtered deposit extracted by TPLS.   TPLS can handle these
large quantities  without difficulty.

TPLS-Spot Test Analysis of Real Water Samples

     The periodate-tetrabase catalytic spot test procedure,  as previously described, was applied
to TPLS fractions filtered onto  Millipore  filters.  The deposit was confined to a 4-mm-diameter
area by using the filter reservoir with a 3-mm-diameter hole.  All deposits were  dried before
reacting the chrysotile with manganous ion.

     Figure 20 shows a comparison between spot test results on standard chrysotile concentra-
tions and on chrysotile concentrations extracts from 50-ml sample aliquots by TPLS of spot
test analyses of  four real water  samples.

     In cases where no  chrysotile was indicated by the  spot test, some extraneous particulate
was observed. The color prints shown are not the  same hue, and it is difficult to obtain prints
which are comparable.  The positive transparencies from which the prints were made, however,
have the same hue and  show differences in depth of color  better.  In view of this, it is recom-
mended that color transparencies of standards be viewed on a light box when comparisons of
analytical results from real water samples are made to standards in the filter deposit indicating
that the spot  test  procedure was selective for chrysotile in these cases.

     Table 9 presents results of the estimated chrysotile concentrations in nine water samples.
                                            45

-------
       TABLE 8. LM ANALYTICAL RESULTS ON REAL WATER SAMPLES
Sample
C9305 1215U4WGO
N2 03021 805 D3WGO
Nl 0606 1335D2WGO
M30307 1500D2 WGO
SI 0326 0700 D4 WIO
M3 037 0730 Dl WIO
PI 01 30 0555 D2 WIO
P402120920D5 WGO
R203050615D1 WIO
C90305 1340D5 WGO
C90305 1350U5 WGO
Gl 0505 0450 D4 WIO
N80403 1730D6WGO
C3 05141 530 D3 WGO
C3 05141 530 U3 WGO
Chatfield Sample 3
(American Source)
Chatfield Sample 1
(Canadian Source)
Light Microscope Counts
0
0
439
Too many fibers to count
Too many fibers to count
Too many fibers to count
134
78
85
Too many to count
705
0
0
382
14
7,926*
594,500*
Nanograms
per Filter
0
0
32
>500
>500
>500
10
6
6.8
>500
50
0
0
21
1.2
400
28,000
Nanograms
per Liter
0
0
640
> 10,000
>1 0,000
>1 0,000
200
120
136
>1 0,000
1,000
0
0
420
24
5,300
373,000
*For Sample 3, three fields within #10 ring in the concentric circle eyepiece reticule were counted; for
 Sample #1, one field delineated by ring #2 was counted.
                                      46

-------
Figure 19. Light micrograph showing a heavy mat of TPLS-extracted
          pure chrysotile from sample M3 0307 1500 D2 WGO.

-------
            Series of standard chrysotile concentrations
              Spot test results on real water samples
Figure 20. Comparison of spot test results on real water samples to spot
          test results on a series of chrysotile standards.

            TABLE 9. SPOT TEST RESULTS ON REAL
                      WATER SAMPLES
         Sample
    PI 0130 0555 D2
    M3 0307 1500 D2
    R2 0305 0615 Dl
    C9 0305 1215  U4
    M3 0307 0730 Dl
    C9 0305 1350 U5
    C9 0305 1340 D5
    C9 0305 1215 U4
Estimated Concentration
        Per Filter
                              48

-------
                                       SECTION 5

             INTERLABORATORY ANALYSIS OF FOUR WATER SAMPLES
     Four water samples supplied by Dr. Eric Chatfield of the Ontario Research Foundation
(ORF) were analyzed at both Battelle and ORF by the TPLS-light microscopic analytical
procedure.

     Instructions describing the procedure were supplied ORF by Battelle.  The first attempt at
ORF to apply the technique as described revealed that the instructions required more  clarifica-
tion.  In particular, the isooctane remained cloudy after the separation procedure, which led the
ORF group to believe  that they were not following the correct procedure.  Cloudy isooctane is
normal, however.  Another problem concerned the precise conditions for the optical microscope
count. As initially planned, Dr.  E. J. Chatfield and Ms. M. J. Dillon visited Battelle's Columbus
Laboratories to obtain firsthand experience with the technique before attempting analysis of
the interlaboratory samples.

     The samples supplied by ORF were from Thetford Mines, Quebec; Toronto, Ontario;
Asbestos, Quebec; and Lloyd  Lake, Ontario.  Both heavily loaded and lightly loaded samples,
which  also contained some extraneous debris, were provided.  By electron microscopy, ORF
estimated these samples to range from about 100 Mg/liter to 10 ng/liter of chrysotile asbestos.
Battelle did not analyze these samples by electron microscopy.

     At ORF, some deviation from conditions in the instructions was necessary because of
equipment and supply  limitations.  In particular, Nuclepore  filters were of 0.1-/urn pore size,
and backing Millipore filters were of 0.45-jum pore size. All shaking was by hand, and double-
distilled water was used instead of deionized  water.

     ORF detected some artifacts on the surface of some blank Nuclepore filters that  might be
mistaken for chrysotile fibers. They state that in many cases the artifact can be recognized for
what it is:  a glancing perforation.  In  others there is some difficulty in rejecting it as  a genuine
fiber.  No comparable  artifacts were observed at Battelle.

     The ORF researchers concluded that the two-phase liquid separation technique is a "most
effective chrysotile extraction method".  They  report that the technique allowed processing of
much larger volumes of dirty water samples than was possible by direct TEM methods. Also,
as Battelle  reports, extraction into the organic phase is not totally specific to chrysotile, which
sometimes gives rise to difficulties in interpretation of  the optical fiber  counting.  Dr.  Chatfield
and Ms. Dillon conclude, however, that an experienced microscopist could in many cases  dis-
tinguish the chrysotile  from the  other debris.
                                            49

-------
     Specific suggestions were made by ORF to assist in clarification of TPLS-light microscopic
analytical instructions.  These have been heeded and the instructions modified accordingly.

     A general suggestion was made concerning the requirements for optical microscopy. ORF
was unable to duplicate the exact  light microscope magnification conditions employed at
Battelle.  The distance between adjacent concentric circles of the eyepiece reticule was 10.0 ,um
with their microscope rather than  12.5 jum as with the Battelle microscope.  Consequently,
Dr. Chatfield suggested  that the standard curve relate total fiber length to chrysotile mass in-
stead of numbers of 12.5-pm lengths to chrysotile mass.  Total fiber length is independent of
the features of a particular microscope, so long as an adequate magnification is in use.  The
eyepiece  reticule in any microscope then need only be calibrated and measurements expressed
as total fiber length.

     Table 10 presents the  comparison of the ORF and Battelle results of analyses on four
water samples.  Agreement  is approximately within a factor of 4 in all cases except for Sample
4, in which the Battelle result is ^0.1  that of ORF. In this case, the Battelle result agrees with
the electron microscopy result obtained by ORF.  According to results from both laboratories,
all of these samples would  be destined for more accurate analysis by electron microscopy,  with
the possible exception of Sample 4.

     The results of this  interlaboratory comparison emphasizes the need for more extensive
round robin testing of the methodology to  lead to further improvement of the reliability of
results.
                                           50

-------
TABLE 10. COMPARISON OF ANALYTICAL RESULTS OBTAINED AT BATTELLE
         AND ORF ON FOUR WATER SAMPLES
Sample Details

Sample
Number
1

3

4

5



Source
Lloyd Lake
Ontario
Thetford Mines
Quebec
Toronto
Ontario
Asbestos
Quebec

Water
Type
Raw

Treated

Raw

Treated



Description
Sand, day

Municipal water
General debris
Lake Ontario

Municipal water
Fine sand
Estimated Mass
by TEM
(jug/liter)
50

1.0

1.3 x 10~2

8.0 x 10~2

Optical
Number of
10.0 Aim
Segments
233

830

92

234

Counts
Area
Examined
(mm2)
0.0785

3.20

3.20

3.20

Equivalent
Number of
12.5 /on
Segments
186

664

74

187

Number of
12.5 /An Segments
Equivalent to
Diameter Scan
9478 ORF
4113 Battelle
830 ORF
5480 Battelle
92 ORF
9 Battelle
234 ORF
668 Battelle

Mass of
Chrysotile
(jug/filter)
0.60
0.28
0.065
0.35
0.0085
0.0013
0.020
0.050

Aliquot
Volume
(ml)
50
75
75
75
75
75
50
75
Calculated
Mass of
Chrysotile
(jug/liter)
16
3.7
1.16
4.7
0.15
0.0177
0.53
0.66

-------
                                    REFERENCES
1.   Speil, S. and Leinweber, J. P., "Asbestos Minerals in Modern Technology", Environmental
    Research,  March,  1969.

2.   Rosenhaltz, V. L., Smith, F. T., "The Dielectric Constants of Mineral Powders", American
    Mineralogist, 21, 115-120, 1936.

3.   Mumpton, F. A., "Characterization of Chrysotile Asbestos and Other Members of the Ser-
    pentine Group of Minerals", Siemens Review, 75-84, 1974.

4.   von Reinders, W., Kolloid Zeitschrift, Vol.  13, p 235,  1913.

5.   Chowdhury, S. and Kitchener, J. A., "The Zeta Potentials of Natural and Synthetic
    Chrysotiles", Int.  J. of Min. Proc., 2, 277-285, 1975.

6.   Lai, R.W.M. and Fuerstenau, D. W., "Liquid-Liquid Extraction of Ultrafine Particles",
    Trans. AIME, Vol. 241,  549-566, 1968.

7.   Sherman,  P., "Emulsion  Science", Academic Press, p 6, 1968.

8.   Bennett, H., Bishop, J. L., Wulfinghoff, M.  F., "Practical  Emulsions", Chemical Publishing
    Company, p 33, 1968.

9.   Sutheim, G. M., "Introduction  to Emulsions", Chemical Publishing Co., p  161,  1946.
                                          52

-------
                                      APPENDIX A

              INSTRUCTIONS FOR CARRYING OUT TWO-PHASE LIQUID
                  SEPARATION COUPLED WITH LIGHT MICROSCOPIC
                   DETECTION OF CHRYSOTILE IN WATER SAMPLES
     The rapid analytical method developed at Battelle-Columbus for determining chrysotile
asbestos in water involves first the separation  of chrysotile from other waterborne particulate
followed by quantitative analysis of the separated fraction by light microscopy.  Separation is
accomplished by a procedure in  which the selective transfer of chrysotile from water to iso-
octane is promoted by a reaction with an anionic surfactant.  Subsequently, the separated
chrysotile is deposited on a Nuclepore filter and its concentration is determined  by  a light
microscopic counting technique.

     The separation procedure is called two-phase liquid separation (TPLS). The selective trans-
fer of chrysotile occurs  when the isooctane and water are shaken together in a separatory funnel
after the chrysotile surface is rendered hydrophobic through a surface reaction with an anionic
surfactant MO-70, dioctyl sodium sulfosuccinate.  The presence of the surfactant causes an
emulsion to form,  which is broken by "salting out" with  sodium chloride.  After the phases
have separated, the isooctane is filtered to deposit the chrysotile on a Nuclepore filter. Then,
the filtered deposit is coated with vacuum-evaporated carbon.  The carbon coating serves to
make the preparation highly reflective and also increases the apparent fiber diameters to bring
more fibers into the detectable range of light  microscopy.

     Light microscopic analysis is carried out  at 500X magnification using incident bright-field
(vertical) illumination to count fiber segments equal in length to the distance (12.5 Mm) be-
tween concentric circles in an  eyepiece reticle. The counts expressed as total fiber length are
then translated into micrograms  of chrysotile  per filter deposit by reference to a standard curve
that relates total fiber length to  mass of chrysotile per filter.  Finally, the chrysotile mass Gzg)
per liter of water sample is calculated  from the  aliquot taken, assuming 75 percent recovery.

     The following detailed procedure presents step-by-step  the instructions for carrying out
TPLS coupled with light microscopic analysis.
Step 1:   Shake  the bottle of water sample with a mechanical shaker (Figure A-l) for approxi-
         mately 5 minutes at «180 cycles per minute to create a uniform particulate
         suspension.
Step 2:   Take an  aliquot either directly from the water sample, or, alternatively, filter an ali-
         quot onto a Metricel DM450 filter.
         Alternative  2A: Take a 75-ml aliquot directly from the uniform water sample. Adjust
         the pH to 3.5 ±0.1  with 0.1  N HC1, add 25 ml of 0.01M MO-70, heat to 140°F


                                           53

-------
          while stirring, and allow to cool to room temperature.  If acid additions overshoot the
          correct pH value, correct by  adding 0.1 N NaOH.  Heat at least to 140°F; heating as
          high as the boiling point has no adverse effect except to extend the time required for
          cooling. Cooling may be accomplished by refrigeration or in an ice bath.  Also, other
          than 75-ml aliquots may be  taken; then the amount of surfactant added must be ad-
          justed to maintain the 2.5 x  10~3 M concentration, and the amount of isooctane must
          be 1/3  of the total liquid volume.
          Alternative 2B:  Filter a relatively large aliquot (1 liter or more) of water sample to
          deposit the particulate on 47-mm Metricel DM450 (0.45Mm pore  size) filter(s)
          (change filters if necessary as they  become loaded to the degree that the water flow
          stops).  Before the  filter deposits have dried, ultrasonically remove the deposited par-
          ticulate from the filter(s) in a beaker containing 75 ml of water to which 25 ml of
          0.01 M MO-70 was  added to make a total volume of 100 ml.  Adjust the pH to
          3.5 ±0.1 with 0.1 N HC1, heat to 140°F, and allow to cool to room temperature.
 Step 3:   Combine the water phase from Step 2A or 2B with 50 ml  of isooctane  in a 250-ml
          separatory funnel equipped with a Teflon stopcock.
 Step 4:   Attach the separatory funnel to a mechanical shaker and shake at  ^150 ±25 cycles
          per minute for 5 minutes (Figure A-2).
 Step 5:   Add 10 ml of 10% NaCl to  break the emulsion (Figure A-3).  Shake lightly to dis-
          tribute the NaCl in the emulsion.  Occasionally the isooctane layer will form under
          the aqueous  layer.  Further gentle shaking will eliminate this condition.
 Step 6:   Remove the  separatory funnel stopper during separation of the two liquid layers.
          Drain off the water  phase through the stopcock (Figure A-4).  Allow time for the
          water droplets to settle and drain them off before pouring the isooctane from the  top
          of the separatory funnel into the 25-mm Millipore filtering apparatus (Figure A-5).
          It is extremely important to  take care that no water is filtered along with the isooc-
          tane. The isooctane is filtered through a 0.2-ptm-pore-size, 25-mm Nuclepore filter,
          shiny side up, backed with an 8.0-jLim-pore-size, 25-mm Millipore filter to obtain  a
          uniform chrysotile distribution.
 Step 7:   Deposit the separated chrysotile onto the Nuclepore filter by aspirating  the isooctane
          through the filter.
 Step 8:   Combine the same water phase again with 50 ml of isooctane  in the separatory funnel
          and repeat the extraction procedure without adding additional  sodium chloride, and
          deposit the extracted chrysotile on the same Nuclepore filter.   Repeat the extraction
          procedure three more times to make a total of five extractions from the same water
          aliquot deposited on the same Nuclepore filter.
 Step 9:   Remove the Nuclepore filter, allow it to  air dry and tape it by the edges, deposit
          side up, so that it lies flat on a 3 x 1-inch labeled microscope slide (Figure A-6).
          Double sided adhesive tape may be used  to hold the filter flat for  microscopic
          examination.

 Step 10:  Evaporate carbon onto the surface of filter bearing the deposit.  (Evaporate 10.0
          millimeters by 1.0 mm graphite rod at a distance of 20 cm at normal incidence.)
Step 11:  Use vertical (incident bright-field) illumination  to view the carbon-coated filter at
                 magnification with a  light microscope equipped with an eyepiece reticle having

                                           54

-------
          10 concentric circles (Figure A-7).  In the eyepiece reticle, the radius of the center
          circle is equal to the distance between adjacent concentric circles.
Step 12:  Position the filter beneath the objective lens so that a traverse may be made across
          the full diameter of the particulate  deposit.

Step 13:  Determine the smallest detectable fiber diameter.  One count by  the light microscopic
          method is registered  for a fiber with the  minimum detectable diameter and a segment
          length equal to the distance between concentric circles.  At  500X, this distance is
          12.5 Mm in the Leitz Orthoplan microscope used at Battelle. For other microscope
          setups the field delineated by the concentric circles must be calibrated using a stage
          micrometer.

Step 14:  Perform fiber segment  counts on fibers that fall within the outside concentric circle
          as  the major diameter of the filtered deposit is  traversed. As counts are made,
          roughly estimate the number of minimum fiber diameters encountered in fiber
          bundles and adjust the fiber segment count accordingly.
Step 15:  Translate fiber segment counts into total fiber length by multiplying the number  of
          counts times the calibrated distance between adjacent concentric  circles.  Refer the
          total fiber length to  the standard curve (Figure A-8), which relates total fiber length
          to chrysotile mass  (jug) per filter.
Step 16:  Calculate the chrysotile concentration  (jug) per liter from the mass per filter obtained
          from the standard  curve as follows:

          pig/filter x 100 x 1000  ml             = /ig/liter
                     75    ml  of water aliquot

          This calculation is  based on 75% recovery.  When the outside circle of eyepiece reticle
          delineates a field 250 ,um in diameter,  and  the major diameter of the deposit is
          16 mm, one 250-Mm-diameter field area is  1/82 of the area of the path 250 /zm wide
          and 16 mm long across the filter.  Because the standard curve was derived using this
          path area (a»4.0 mm^), areas measured with other microscope conditions must be de-
          termined and compensations made  for differences by dividing the total path area by
          the area over which  measurements were made  and multiplying this ratio times the
          number of fiber counts times the calibrated distance between concentric circles.

     When the radius of the  smallest circle in a set of 10 circles is equal to the spacing between
adjacent circles, the following  area percentages may be used in calculations to normalize fiber
length data obtained by  counting less area than that within the outermost circle.

          Circle No.  10          100%
          Circle No.  9           81%
          Circle No.  8           64%
          Circle No.  7           49%
          Circle No.  6           36%
          Circle No.  5           25%
          Circle No.  4           16%
          Circle No.  3           9%
          Circle No.  2           4%
          Circle No.   1            1%

                                             55

-------
Making Up Solutions

     It is imperative that all solutions be filtered before use.  It is suggested that solutions be
filtered through 0.2-/xm-pore-size, 47-mm-diameter Millipore filters.

MO-70 (Dioctyl Sodium Sulfosuccinate)

     MO-70 is obtainable from Mona Industries, Inc., Paterson, New Jersey 07524, in 70 per-
cent aqueous solution.  To make up the 0.01 M stock solution, weigh out 6.34 ±0.02 g.  Heat
the solution to  140-150°F to obtain a uniform solution while stirring.  Dilute to  1 liter with
deionized  water, and filter.   (Shelf life — 1 month).

0.1 N HC1 and 0.1 N NaOH

     Purchase vial of Dilut-lt (G. T.  Baker Chemical Co.) for making up 0.1 N HC1 or 0.1  N
NaOH.  Dilute the contents  of the plastic ampoule to 1  liter with deionized water, and filter
through 0.2-pm-pore-size, 47-mm Millepore filter.

10% NaCl

     Add  90 ml of deionized water to lOg of reagent grade NaCl and filter.
                                           56

-------
        Figure A-l.  The mechanical shaker.
Figure A-2. Separatory funnels attached to mechanical
            shaker with rubber bands.
                        57

-------
Figure A-3.  Before (right) and after (left) breaking the
             emulsion by "salting out" with NaCl.

                          58

-------
Figure A-4.  Draining off the water phase.
                                                      :
   Figure A-5.  Filtering the isooctane.
                     59

-------
Figure A-6. A TPLS specimen prepared for light microscopic analysis.  The
            filtered TPLS fraction was taped to a microscope slide and
            coated with vapor-deposited carbon in a vacuum evaporator.
                                   60

-------
Figure A-7. Concentric circles in eyepiece reticle superimposed on image
            of TPLS fraction from ORF water sample.
                                   61

-------
             125,000,000 p
              12,500,000 —
                          I  I I II   I  I I II   I  I 11 I   I  I  Ill  I  III
                     0.0001   0.001    0.01     O.I       1.0       10
                                  Pure Chrysotile, fj.g


Figure A-8.  Standard curve which relates light microscope measurement of
             total fiber length to mass (jug) of chrysotile per filter.
                                    62

-------
                                       APPENDIX B

                INSTRUCTIONS FOR CARRYING OUT TWO PHASE LIQUID
                   SEPARATION (TPLS) COUPLED WITH SPOT TEST
                  DETECTION OF CHRYSOT1LE IN WATER SAMPLES
     The rapid analytical method alternative to two-phase liquid separation - light microscopic
detection procedure is the TPLS - spot test procedure.  Although spot test detection is less
sensitive than light microscopic detection (0.1 Mg versus  0.001 /xg), spot testing is more rapid,
and its sensitivity is adequate for some monitoring situations. Spot testing also has the advan-
tage of being routine, rapid, requiring no equipment as expensive as a light microscope, and
very little training.  After separation of the chrysotile is  accomplished by TPLS, spot test de-
tection can be run in 10 minutes.

     The only equipment needed are a 25-mm Millipore filtering apparatus, a special reservoir
with a 3-mm hole for the filter apparatus, a heat lamp, a wash bottle, spot test chemicals,  and
a spraying device such as is used in thin layer chromatography.  The materials required are
25-mm-diameter, 0.45-jum-pore-size Millipore filters, spot-test  chemicals (potassium periodate
and tetrabase), and white blotters.

     A technician who has no formal scientific training can be taught to run the spot test de-
tection procedure in less than 1 hour.  Semiquantitative  results can readily be evaluated with a
little practice to determine whether there are changes in chrysotile concentration within the
range of detection of the method.

     The chrysotile extraction procedure is the same as that  recommended prior to light micro-
scopic detection; however, a Millipore filter is used in the spot-test technique rather than a
Nuclepore filter as is used for the light microscopic detection.

     The selective transfer of chrysotile in the TPLS procedure occurs when the isooctane and
water are shaken together in a separatory funnel after the chryostile surface is rendered hydro-
phobic through a surface reaction with the anionic surfactant MO-70, dioctyl sodium sulfo-
succinate.  The presence of the surfactant causes an emulsion to form, which is broken by
"salting out" with sodium chloride. After the phases have separated, the isooctane is filtered to
deposit the chrysotile on a Millipore filter.  The  deposit  is confined to a 4-mm-diameter filter
area to enhance spot test detection.

     Spot test detection involves first the reaction of manganous ion with chrysotile extracted
from a water sample followed by the detection of the reacted manganese by a  spot test reaction
which is catalyzed by manganese. The manganese catalyzes the reaction between potassium
periodate and tetrabase [4, 4' Methylenebis (N, N-dimethylaniline)] to yield a  bright blue


                                            63

-------
reaction product.  The depth of color developed is indicative of the quantity of manganese
present in the chrysotile structure.

     The following detailed procedure presents step-by-step instructions for carrying out TPLS
coupled with chrysotile detection by the catalytic periodate-tetrabase spot test.
Step 1:   Shake the bottle of water sample with a mechanical  shaker (Figure A-l) for approxi-
         mately 5 minutes at ^ISO cycles per minute to create a uniform particulate
         suspension.
Step 2:   Take an aliquot either directly from the water sample, or, alternatively, filter an ali-
         quot onto a Metricel DM450 filter.
         Alternative 2A:  Take a 75-ml aliquot directly from the uniform  water sample.  Ad-
         just the pH to 3.5 ±0.1 with 0.1  N HC1, add 25  ml of  0.01M MO-70, heat to  140°F
         while stirring and allow to cool to room temperature. If acid additions overshoot the
         correct pH value, adjust by adding 0.1 N NaOH.  Heat at least to  140°F; heating as
         high as the boiling point has no adverse effect except to extend the time required for
         cooling.  Cooling may be  accomplished by  refrigeration  or an ice bath.  Also, other
         than 75-ml aliquots may be taken; then the amount of  surfactant added must be ad-
         justed to maintain the 2.5 x  10~3 M concentration, and the amount of isooctane must
         be 1/3 of the total liquid volume.
         Alternative 2B:  Filter a relatively large aliquot (1 liter  or more) of water sample to
         deposit the particulate on 47-mm Metricel DM450 (0.45-jum-pore-size) filter(s)
         (change filters if necessary as they become loaded to the degree that the water  flow
         stops).  Before the  filter deposits have dried, ultrasonically remove the deposited par-
         ticulate from the filter(s)  in a beaker containing 75 ml of water to which 25 ml of
         0.01  M MO-70 was added to make a total  volume of 100 ml.  Adjust the pH to
         3.5 ±0.1 with 0.1 n HC1  or 0.1 n NaOH, heat to  140°F, and  allow to cool to room
         temperature.
Step 3:   Combine the water phase from Step 2A or 2B with 50  ml of isooctane  in a 250-ml
         separatory funnel equipped with a Teflon stopcock.

Step 4:   Attach the separatory funnel to a mechanical shaker and shake at  «* 150 ±25 cycles
         per minute for 5 minutes (Figure A-2).
Step 5:   Add  10 ml of 10% NaCl  to break the emulsion (Figure A-3). Shake lightly to dis-
         tribute the NaCl  in the emulsion. Occasionally the isooctane layer will form under
         the aqueous layer.  Further gentle shaking  will eliminate this condition.
Step 6:   Remove the separatory funnel stopper during separation of the two liquid layers.
         Drain off the water phase through the stopcock (Figure A-4). Allow time for the
         water droplets to settle and drain them off before pouring the isooctane from the top
         of the separatory funnel into the 25-mm Millipore filtering apparatus (Figure A-5).
         It is extremely important to take care  that no water is filtered along with the
         isooctane.

         A special filter reservoir is used to concentrate the chrysotile into  a small area
         (Figure B-l).  [The special reservoir, used as a substitute for the standard reservoir, is
         made with a 3-mm hole to concentrate the asbestos in  a 4-mm-diameter filter area
         (Figure  B-l).  A glass blower  can readily make this special part.]
                                            64

-------
Step 7:   Aspirate the isooctane through a 25-mm, 0.45-/«n, Millipore filter to deposit the ex-
          tracted chrysotile.

Step 8:   Combine the same water phase again with 50 ml of isooctane in the separately funnel
          and repeat the extraction procedure without adding additional sodium chloride, and
          deposit the extracted chrysotile on the same Nuclepore filter. Repeat the extraction
          procedure three more times to make a total of five extractions from the same water
          aliquot deposited on the same Millipore filter.

Step 9:   Remove the filter, air dry and store in a labeled plastic petri dish.
Step 10:  Place the filter, deposit side up on the frit of a 25-mm Millipore filtering apparatus
          and place a  blank Millipore filter on top of the filter bearing the chrysotile deposit.
          Clamp the standard Millipore reservoir (16-mm-hole) in position.
Step 11:  Pour water into the reservoir and adjust the aspiration rate to 15 ml per minute.
Step 12:  Aspirate  15  ml of 25% manganous sulfate (25 g MnSO4 • H^O/lOO ml solution)
          through the filter.
Step 13:  Wash five times aspirating each washing completely through the filter before  adding
          the next washing.
Step 14:  Increase the aspiration rate to hold the filters to  the frit and remove the filter
          reservoir.
Step 15:  Separate the filters and discard the top filter.
Step 16:  Place the bottom filter deposit side up on a white blotter and dry under the  heat
          lamp at 145°F for 1 minute.
Step 17:  Attach the filter deposit side up by its edges with cellophane tape to the white
          blotter.
Step 18:  Use a spraying device (Figure  B-2) to spray the filter two times lightly with  the spot
          test working solution which is made up by mixing equal (5  ml)  quantities of the fol-
          lowing Solutions A and B:
          Solution A:  Saturated solution of potassium periodate in 2N acetic acid (11.5 ml
          glacial acetic diluted to 100 ml with water).
          Solution B:   1.0 g of tetrabase dissolved in  100 ml of 95  percent ethyl alcohol.
Step 19:  Place the filter under the heat lamp at 145°F until the color of the spot does not
          change,  or if no spot becomes visible, until  the filter background turns a light blue
          color.
Step 20:  Compare the color of the spot to  colors developed in  similar preparations made from
          known chrysotile concentrations.  The latter may be recorded on 35-mm color trans-
          parencies that can be viewed to make comparisons to  unknowns.  Color prints were
          not found satisfactory for  this purpose.
          A color Xerox copy of a spot test result  obtained on  an 0.8 /ig chrysotile standard is
          shown in Figure B-3.
          Note:  Filter all solutions through 0.45-Mm-pore-size  Millipore filters before use.
                                             65

-------

          Figure B-l.  Special reservoir with 3-mm-diameter hole.
Figure B-2.  Spraying device with vial containing spot test working solution.



                                    66

-------

                   •:! I'-•*- •
                   ^,-«*-.• ,.
Figure B-3.  A typical spot test result obtained on an 0.8-/ig chrysotile standard.

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1. DEPORT \o.
  EPA-600/4-78-066
                                                           3. RECIPIENT'S ACCESSION NO.
  TITLE A\'D SUBTITLE
  Development of a Rapid Analytical Method for Deter-
  mining Asbestos in Water
                       5. REPORT DATE-
                        December ]978 issuing date
                       6. PERFORMING ORGANIZATION CODE
 7 AUTHOR(S)
  C.W.  Melton, S.J. Anderson,  C.F.
  R.E.  Heffelfinger
                                                           8. PERFORMING ORGANIZATION REPORT NO
Dye, W.E. Chase,  and
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
  BATTELLE  Columbus Laboratories
  505 King Avenue
  Columbus, OH  43201
                        10. PROGRAM ELEMENT NO.

                          1BD713	
                        11. CONTRACT/GRANT NO.
                          68-03-2199
 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   6/75 -  6/78
                        14. SPONSORING AGENCY CODE

                          EPA/600/01
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
       The development of  a  rapid analytical method for  determining chrysotile asbestos
  in water that requires substantially less time per analysis  than electron microscopy
  methods is described.  Based on the proposition that separation of chrysotile from
  other waterborne particulate would greatly simplify the  task of detection, the researc
  effort was directed toward establishing separation and concentration techniques.  This
  investigation led to the development of a separation procedure whereby chrysotile is
  extracted from a water sample into an immiscible organic liquid phase.  The procedure
  is called two-phase liquid separation (TPLS).

       TPLS has been combined with a light microscopic intercept counting technique and
  with acolorimetric spot  test detection technique to result in two complete rapid anal-
  ytical methods.  The TPLS-light microscopic (LM) method  requires more expensive equip-
  ment than the TPLS-spot  test method; however, the limit  of detection of TPLS-LM
  method is 1.0 ng at the  99 percent confidence level, whereas the limit of detection of
  the TPLS-spot test method  is approximately 100 ng.  The  TPLS-LM method, therefore, is
  recommended as a first choice,  and the TPLS-spot test  method is recommended for use
  under conditions that require no greater detection sensitivity than 100 ng per sample.

       Experiments have shown  that 75 percent of the chrysotile is extracted by TPLS.
 17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
          b.IDENTIFIERS/OPEN ENDED TERMS  C.  COSATI Field/Group
       Asbestos
       Chemical  analysis
       Electron  microscopy
       Water pollution
                                           07B
                                           13B
 8. DISTRIBUTION STATEMENT

 RELEASE TO PUBLIC
          19. SECURITY CLASS (ThisReport)
             UNCLASSIFIED
                                                                         21. NO. OF PAGES
80
                                              20. SECURITY CLASS (This page)
                                                 UNCLASSIFIED
                                                                         22. PRICE
EPA Form 2220-1 (9-73)
                                            68
                                                                     -.: U.S. GOVERNMENT PRINTING OfFICEr 1979 -657-060 /1558

-------