EPA-670/2-75-050e
June 1975
Environmental Protection Technology Series


-------
                                                     EPA-670/2-75-050e
                                                     June 1975
                   DIRECT FILTRATION OF LAKE SUPERIOR

                   WATER FOR ASBESTIFORM FIBER REMOVAL


                               Appendix E

    Ontario Research Foundation Electron Microscope Analysis Results

                               Appendix F

EPA National Water Quality Laboratory X-Ray Diffraction Analysis Results

                               Appendix G

 University of Minnesota at Duluth Electron Microscope Analysis Results
                                   By
                  Black $ Veatch,  Consulting Engineers
                      Kansas City, Missouri  64114
                       Program Element No.  1CB047
                     Contract No.  DACW 37-74-C-0079
                  Interagency Agreement EPA-IAG-D4-0388
            USEPA,  Region V and Corps of Engineers, St. Paul
                             Project Officer

                             Gary S. Logsdon
                    Water Supply Research Laboratory
                 National Environmental Research Center
                         Cincinnati, Ohio  45268
                 NATIONAL ENVIRONMENTAL RESEARCH CENTER
                   OFFICE OF RESEARCH AND DEVELOPMENT
                  U.S.  ENVIRONMENTAL PROTECTION AGENCY
                         CINCINNATI, OHIO  45268

-------
                          REVIEW NOTICE
         The National Environmental Research Center -- Cincinnati
has reviewed this report and approved its publication.  Approval
does not signifiy that the contents necessarily reflect the views
and policies of the U.S. Environmental Protection Agency,  nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
                                11

-------
                           FOREWORD
     Man and his environment must be protected from the adverse
effects of pesticides, radiation, noise,  and other forms of
pollution, and the unwise management of solid waste.   Efforts
to protect the environment require a focus that recognizes the
interplay between the components of our physical environment --
air, water and land.   The National Environmental Research Centers
provide this multidisciplinary focus through programs engaged in

     •    studies on the effects of environmental contaminants
          on man and the biosphere, and

     •    a search for ways to prevent contamination and to
          recycle valuable resources.

     This report and its appendices present the results of pilot
plant filtration research for the removal of asbestiform fibers
from drinking water.   The several appendices present detailed
information on water quality, pilot plant equipment and operation,
individual filter run data, asbestiform fiber and amphibole mass
concentrations in raw and filtered water, and diatomite filter
optimization.  Appendix E contains electron microscope analytical
results from the Ontario Research Foundation.  Appendix F contains
x-ray diffraction analytical results from the EPA National Water
Quality Laboratory in Duluth.  Appendix G contains electron
microscope analysis results from the University of Minnesota at
Duluth.
                               A,  W.  Breidenbach,  Ph.D.
                               Director
                               National Environmental
                               Research Center,  Cincinnati
                               111

-------
                           ABSTRACT
     Pilot plant research conducted in 1974 at Duluth,  Minnesota,
demonstrated that asbestiform fiber counts in Lake Superior water
could be effectively reduced by municipal filtration plants.
During the study, engineering data were also obtained for making
cost estimates for construction and operation of both granular
and diatomaceous earth (DE) filtration plants ranging in size from
0.03 to 30 mgd.

     Data provided to the contractor by the Ontario Research
Foundation are presented in Appendix E.  ORF performed asbesti-
form, fiber analysis of water samples by the transmission electron
microscope method in this project.  In order to place the data in
better perspective, a description of the analytical method used by
ORF is reproduced in Appendix E.

     In Appendix F, the amphibole mass data obtained by the
National Water Quality Laboratory in Duluth are presented.  This
appendix also includes information on the analytical methods used
at NWQL.  The x-ray diffraction analysis for amphibole mass
provided confirmation of electron microscope amphibole fiber
results.

     Fiber count data obtained at the University of Minnesota at
Duluth are tabulated in Appendix G.  A statement describing the
electron microscope analytical method is also included.
                                IV

-------
                           CONTENTS
Appendix
          Ontario Research Foundation Electron Microscope
            Analysis Results of Pilot Water Treatment
            Units - Raw Water from Duluth Lakewood Intake     1

          Quantitative Analysis of Asbestos Minerals in
            Air and Water (E. J. Chatfield)                   9

          Measuring Asbestos in the Environment (E. J.
            Chatfield and H. Pullan)                         11
          EPA National Water Quality Laboratory X-Ray
            Diffraction Analysis Results of Pilot Water
            Treatment Units - Raw Water from Duluth
            Lakewood Intake                                  16

          Semi-quantitative Determination of Asbestiform
            Amphibole Mineral Concentrations in Western
            Lake Superior Water Samples (P.  M.  Cook)          24

          Asbestiform Amphibole Minerals:   Detection  and
            Measurement of High Concentrations  in Muni-
            cipal Water Supplies (P.  M. Cook, G.  E.
            Glass,  and J. H.  Tucker)                          35
          University of Minnesota at Duluth Electron
            Microscope Analysis Results of Pilot Water
            Treatment Units  -  Raw Water from Duluth
            Lakewood Intake                                   39

-------
APPENDIX E ONTARIO RESEARCH FOUNDATION ELECTRON MICROSCOPE ANALYSIS
           RESULTS OF PILOT WATER TREATMENT UNITS - RAW WATER FROM
1 '
Date of
sample
4/19

4/25

5/7

5/9

5/16

5/16

5/16

5/16

5/16

5/30

5/30

5/30

5/30

5/30

6/4

6/4

6/4

6/4

6/4

6/6

6/6

Filter
utilized
MM-2

MM- 2

BIF

BIF



MM-1

MM-2

BIF

ERD-2



MM-1

MM-2

BIF

ERD-2



MM-1

MM-2

BIF

ERD-2



MM-1

Run
No.
1

7

10-T

12-T



7

37

8

10



26

50

17

14



29

53

19

15



31

f/1
Raw
Sample0
A=0.304
C=0.217
A=0.348
C=0.174
A=0.522
C=0.130
A=0.870
C=1.78
A=2.61
C=1.35








A=1.43
C=1.83








A=1.74
C=2.91








A=1.52
C=2.35


x 106

Finished Per Cent
sample0 Removal
A=<0.0435
C=0.0435
A=<0.0435
C=0.348
A=<0.0435
C=1.43
A=<0.0435
C-<0.0435


A=<0.0435
C=0.130
A=<0.0435
C=0.174
A=0.565
C=0.652
A=0.261
C=1.04


A=<0.0435
C=0.261
A=<0.0435
C=0.913
A=0.739
C=1.83
A=<0.0435
CO.0435


A=<0..0435
C=0.478
A=<0.0435
C=1.87
A=0.217
C=0.0870
A=0.391
C=l . 00


A=<0.0435
C=0.348
85
80
87
—






98
90
98
87


90
23


97
86
96
50


97
98


97
83
97
36


77
66


97
85

-------
APPENDIX E  (CONTINUED) .
Date of
sample
6/6

6/6

6/6

6/11

6/11

6/11

6/11

6/11

6/13

6/13

6/13

6/13

6/13

6/17

6/17

6/17

6/17

6/17

6/24

6/24

6/24

6/24

Filter
utilized
MM-2

BIF

ERD-2



MM-1

MM-2

BIF

ERD-1



MM-l

MM-2

BIF

ERD-1



MM-1

MM-2

BIF

ERD-1



MM-1

MM-2

BIF

Run
No.
54

20

15



34

55

24

1A



36

57

28

1A



39

61

31

4A



44

71

37

f/1 x 106

Raw Finished Per Cent
sample0 sample0 Removal
A=0.0870
C=0.0870
A=0.0870
C=0.739
A=0.130
C=0.435
A=1.04
C=1.65
A=0.174
C=0.174
A=<0.087
C=2.26
A=0.174
C=0.261
A=0.957
C=9.31
A=0.48
C=0.41
A=<0.0217
C=0.57
A=0.15
C=2.02
A=0.80
C=0.46
A=0.09
C=0.71
A=0.61
C=2.15
A=<0.0217
C=0.44
A=0.04
C=0.37
A=0 . 09
C=0.39
A=<0.0217
C=1.37
A=1.04
C-3.91
A=.<0.0217
C=0.130
A=0.0217
C=0.978
A=<0.0217
C=0.239
40
65


91
82


83
89
91
—


8
—


96
—
69
—






97
79
93
83


97
36


80
97
98
75



-------
Date of
sample
6/24

6/28

6/28

6/28

6/28

6/28

7/3

7/3

7/3

7/3

7/3

7/19

7/19

7/19

7/19

7/19

7/23

7/23

7/23

7/23

7/23

Filter
utilized
ERD-2



MM-1

MM- 2

BIF

ERD-2



MM-1

MM-2

BIF

ERD-2



MM-1

MM-2

BIF

ERD-2



M^l

MM-2

BIF

ERD-2

Run
No. .
22



48

76

47

28



52

78

51

32



67

96

70

40



70

99

72

41

f/1 x 106

Raw Finished Per Cent
sampled sampleP Removal
A=0.0217
C=0.239
A-l.ll A=0.91b
C=8.12 C=3.35b
A=< 0,0217
C=1.37
A=< 0.0217
C=0.52b
A=0.15
C=6.03
A=0.11
C=5.3
A=0.565
C=3.57
A=< 0.0217
C=0.283
A=0.0217
C=0.544
A=0.196
C=1.67
A=0.130
C=0.804
A=0.52
C=0.35
A=< 0.0217
C=0.15
A=0.02
C=0.04
A=<0.0435
C=<0.0435
A=0.11
C=0.22
A=0.54
C=0.09
A=<0.0217
C=0.07
A=<0.0217
C=0.20
A=.<0.0217
C=0.76
A=0.09
C=0.09




98
83
97
84


90
35


96
92
97
86


77
77


96
57
96
91


79
40


96
22
96
—


83


-------
APPENDIX E fCONTINURm.
Date of
sample
7/25

7/25

7/25

7/25

7/25

7/30

7/30

7/30

7/30

7/30

7/31

7/31

8/1

8/1

8/1

8/1

8/1

8/6

8/6

8/6

8/6

8/6

Filter
utilized


MM-1

MM-2

BIF

ERD-2



MM-1

MM-2

BIF

ERD-2



MM-2



MM-1

MM-2

BIF

ERD-2



MM-1

MM-2

BIF

ERD-2

Run
No.


72

101

73

43



74

105

77

45



106



76

107

79

46



78

109

82

48

f/1 x 106

Raw Finished Per Cent
sample0 Sample0 Removal
A=0.11
C=1.43
A=<0.0217
C=0.37
A=<0.0217
C=0.15
A=0.04
C=0.35
A=<0.0217
C=0.44
A=0.11
C=0.11
A=<0.0217
C=<0.0217
A=<0.0217
C=0.04
A=<0.0217
C=0.07
A=<0.0217
C=0.07
A=0.33
C=0.22b
A=<0.0217
C=<0.0217
A=0.22
C=0.15
A=<0.0217
C=0.22
A=0.02
C=<0.0217
A=<0.0217
C=0.04
A=<0.0217
C=0.13
A=0.6
C=0.3
A=<0.0217
C=0.06
A=<0.0217
C=<0.0217
A=<0.0217
C=0.06
A=0.02
C=<0.0217


82
74
82
89


82
69


82
82
82
64


82
36


94
91


91
—
91
87


91
13


67
80
97
93


97
93

-------
APPENDIX E (CONTINUED).
Date of
sample
8/8

8/8

8/8

8/8

8/8

8/13

8/13

8/13

8/13

8/13

8/15

8/15

8/15

8/15

8/15

8/20

8/20

8/20

8/22

8/21

8/22

8/22

8/23

8/23

Filter
utilized


MM-1

MM-2

BIF

ERD-2



MM-1

MM-2

BIF

ERD-2



MM-1

MM-2

BIF

EKD-2



MM-2

EKD-2



MM-1

MM-1

MM-1



MM-2

Run
No.


80

111

84

49



82

113

88

51



84

114

89

55



118

59



86

86

86



119

f/1 x 106

Raw Finished Per Cent
sample sample Removal
A=0.06
C=0.09
A=0.04
C=0 . 20
A=<0.0217
C=0.09
A=<0.0217
C=0.4
A=<0.0217
C=0.04
A=0.13
C=0.20
A=<0.0217
C=0.03
A=<0.0217
C=1.4
A=<0.0217
C=0.1
A=0.04
C=0.50
A=0.09
C=0.17
A=<0.0217
C=0.52
A=0.02
C=0.22
A=0.15
C=0.44
A=0.04
C=0.04
A=0.30
C=0.72
A=0.02
C=2.1
A=<0.0217
0=0.17
A=0.13
C=0.37
A=<0.0217
C=0.13
A=<0.0217
C=0.22
A=<0.0217
C=0.80
A=0.72
C=0.44
A= 0.0217
C=0.28


33
—
67
—


67
55


85
85
85
—


69
—


78
—
78
—


55
76


93
—
93
76


85
65
85
40
85
—


97
36

-------
APPENDIX E  (CONTINUED).
Date of
sample
8/23

8/28

8/28

8/28

8/30

8/30

8/30

8/30

9/4

9/4

9/4

9/4

9/6

9/6

9/6

9/6

9/9

9/9

9/9

9/10

9/10

9/11

9/11

Filter
utilized
ERD-2



MM-2

BIF



MM-2

BIF

ERD-2



MM-2

BIF

ERD-2



MM-2

BIF

ERD-2



BIF

ERD-2



MM-2



MM-2

Run
No.
63



122

103



124

105

68



126

108

70



128

111

72



113

73



131



133

f/1
Kaw
sample0


A=0.39
C=0.48




A=0.78
C=0.33






A=1.61
C=0.33






A=0.72
C=0.39






A=1.0
C=0.5




A=0.60
C=0.50


A=0.30
C=0.30


x 106
finished
	 sample0
A=0.02
C=0.59


A=<0.0217
C=0.04
A=0.15
C=0.24


A=0.02
C=0.30
A=0.04
C=0.17
A=<0.0217
C=0.54


A=<0.0217
C=<0.0217
A=0.17
01.15
A=<0.0217
C=0.54


A=<0.0217
C=0.37
A=0.02
C=0.59
A=<0.0217
C=1.67


A=0.04
C=1.0
A=< 0.0217
C=1.0


A=< 0.0217
C=0.5


A=0.02
C-0 . 30
Per Cent
Removal
97
__


95
92




97
— _


97
__


99
94


99
	


97
50


97
__




98
—


97
—


93


-------
APPENDIX E (CONTINUED).
Date of
sample
9/11

9/13

9/13

9/13

9/13

9/16

9/16

9/16

9/16

9/17

9/17

9/19

9/19

9/19

9/19

9/20

9/20

Filter
utilized
BIF



MM-2

BIF

ERD-2



MM-2

BIF

ERD-2



MM-2



MM-1

MM-2

BIF



ERD-2

Run
No.
115



137

117

78



138

118

79



139



87

140

120



85

f/1 x 106

Raw Finished Per Cent
sample0 sample0 Removal
A=< 0.0217
C=0.10
A=0.02
C=0.06
A=0.02
C=0.10
A=0.02
C=0.40
A=< 0.0217
C=0.06
A=0.90
C=0.10
A=0.02
C=0.30
A=<0.0217
C=1.30
A=0.02
C=0.06
A=0.20
C=0.10
A-<0.0217
C=0.10
A=0.60
C=0.09
A=1.00
C=0.70
A=< 0.0217
C=0.30
A=0.02
C=1.80
A=1.3
C=0.2
A=0.02
C=0.32
93
67


—
__






98
__


98
40


90
__


—
__
97
—




98
~*~
   BDL -  Below detectable  limits  of  analysis  equipment




   Cloquet  Pipeline water




   A  = amphibole,   C =  chrysotile

-------
                10
00
              <0  8
              O
              cr
              UJ
              UJ
              m
              OL
              O
              —  3
              I-
              co
              UJ
              00  «
                                AMPHIBOLE
                         	-^  CHYRSOTILE
                    5     15    25
                        APRIL
 15   25
MAY
 15
JUNE
                                                               25
  15    25    5     15    25
JULY           AUGUST
5    15    25
 SEPTEMBER
                     FIGURE   I.     ONTARIO  RESEARCH  FOUNDATION
                                  ASBESTIFORM  FIBER COUNTS
                                  RAW WATER AT DULUTH  LAKEWOOD INTAKE- 1974

-------
 QUANTITATIVE ANALYSIS OF ASBESTOS MINERALS IN AIR AND WATER

 E. J. Chatfield

 Dept. of Applied Physics, Ontario Research Foundation, Sheridan Park,
 Mississauga, Ontario, Canada

 The effects of inhalation o.f asbestos particles are well known. (*)   However,
 although the significance of asbestos particles when ingested is not fully
 understood, there is  some evidence of increased incidence 6f gastrointestinal
 carcinoma where individuals  have been exposed to the material over  a long
 period. (s)  Methods are  therefore required by which trace concentrations of
 asbestos minerals in  both air and ingestibles can be monitored.

 Asbestifonn minerals  fall into two groups, serpentine and amphibole.  Chryso-
 tile is the only asbestifonn member of the serpentine type, consisting of
 magnesium silicate Mg3Si205(OH).  The asbestifonn amphiboles have a range of
 composition X7Si8022(OK)2 where X may be Na, Fe2*, Fea+, Mg or Ca in various
 combinations.   The principal  types of amphibole commonly encountered include
 crocidolite, anthophyllite,  tremolite, and amosite, although amosite itself
 has a number of sub-species  of variable compositions.  Although it  is  common-
 ly thought that asbestos minerals are indestructible, in practice they are
 decomposed by heat or acids  to a variable degree depending on the particular
 variety.  This property  restricts the treatment possible in any analytical
 procedure.  This paper describes an analytical technique which is suitable
 for detection and measurement of low concentrations of asbestifonn  minerals
 in air and water samples.

 The first step in the procedure is to collect some of the solid material on a
 O.lpjn pore size membrane filter.  In the case of air, about 5m3 is  filtered;
 about 200ml in the case  of a  measurement on water.  The membrane filter is
 then ashed in a clean glass vial using a plasma microincineration technique.
 The ashing takes place at low temperatures (typically 70°C), thus whilst no
 decomposition of the  mineral  fibers can occur, organic materials and the
 filter itself  are oxidised to C02.  The residue is gently redispersed  ultra-
 sonically in filtered distilled water, and the suspension centrifuged  on to a
 1cm diameter glass cover disc.  The disc is dried and a thin carbon coating
 is applied by  evaporation.  The carbon film   is scored and floated off on to
 water, carrying the particles with it; pieces of this are then picked  up on
 200 mesh copper grids.   A maximum of 10 grid squares is searched for asbestos
 particles at a magnification  of about 25000.   Particles are identified at the
 instrument by  electron diffraction, and measured in both length and width by
 comparison with a series  of geometrically spaced calibration circles scored
 on the fluorescent screen of  the microscope.   Typical crocidolite particles
 are shown in Figure 1, and their single fiber diffraction pattern is shown in
 Figure 2.  Figures 3  and  U show typical amosite particles and their single
 fiber diffraction pattern.   It can be seen that the diffraction patterns in
 this case are  quite different, although identification within the amphibole
 series is not  generally  so simple.  In contrast, chrysotile can be  easily
 distinguished  by both morphology and its diffraction pattern.

 The particle counts are processed by a computer program,  which calculates the
 particle number and mass  concentrations;  it also plots the number and  mass
 sice distributions.   Using the sample volumes specified,  air concentrations
 can be measured down  to about 0.01 fibers/cc, and water concentrations to 10*
 fibers/liter in typical cases.  The actual sensitivity depends on the  degree
 of concentration which can be achieved,  whilst retaining  a suitable micro-
 scope specimen.   This  in  turn is determined by the concentration of extra-
 neous material  present in the original sample.

Quantitative Analysis of Asbestos  Minerals  in  Air  and Water by E.  J.
Chatfield  is reprinted from 32nd Ann.  Proc.  Electron Microscopy  Soc.
Amer.,  St. Louis,  Missouri, 1974,  C. J. Arceneaux  (ed.),  with  permis-
sion of Claitors Publishing Division,  3165 South Arcadin, Baton
Rouge,  La.

-------
                Reprinted from:
                32nd Ann. Proc. Electron Microscopy  Soc. Amer."
                St. Louis, Missouri, 1974. C. J. Arceneaux  (ed.).

 An extensive study has been made  of water samples in Ontario,(3) following a
 small pilot study by a different  technique.(4)   The water in Ontario general-
 ly contains an average of about 2 million fibers of chrysotile per liter, and
 usually no detectable amphibole types.

 1.  I.  J.  Selikoff et al, Arch. Environ. Health, 25, p. 1-13,  (1972).
 2. . T.  F.  Mancuso and E. J.  Coulter, Arch. Environ.  Health,  6,  p. 210, (1963).
 3.  G.  Kay, Water and Pollution Control, Sept.  1973, p. 33-35.
 4.  J.  M.  Cunningham and R.  Pontefract, Nature, 232, p. 332, (1971).
Figure  1.  Crocidolite Fibers
                                         Figure  2.  Crocidolite  Single  Fiber
                                                    Diffraction  Pattern
Figure 3.  Amosite Fibers
                                         Figure 4.  Amosite Single Fiber
                                                    Diffraction Pattern
                                     10

-------
Measuring Asbestos  in  the  Environment by E.  J.  Chatfield  and H.  Pullan is  reprinted
from Canadian Research & Development, Nov-Dec,  1974,  with permission  of Maclean-
Hunter Ltd.,  481  Univ.  Ave., Toronto, Ontario M5W-1A7.    (Pages  23-27.)

   Measuring asbestos  in the environment
   THE  INCIDENCE  of asbestos-related
   disease reported by medical author-
   ities and its classification as a hazard-
   ous material " "' has created a need
   for techniques  by  which  it can be
   measured at low concentrations in
   air and water samples.
     To appreciate the criteria involved
   in  the design of such low level ana-
   lytical methods, it is useful to review
   some  of the recent  history of  as-
   bestos-related  disease,  the  estab-
   lished maximum occupational expo-
   sure levels, and the properties of the
   materials themselves. Until recently,
   inhalation was considered to be the
   only hazard associated with this ma-
   terial, and fairly reliable data are
   available relating the incidence of
   the progressive  disease  asbestosis
   with  the  individual's exposure. In-
   deed,  occupational exposure levels
   have been defined for some years in
   both the USA and Britain. However,
   in the last few years a previously rare
   malignancy  condition,  mesothe-
   lioma, has been linked with exposure
   to asbestos, this exposure not always
   being  an occupational one.  "  Fur-
   thermore,  statistical  studies  have
   shown  that  persons  exposed  to
   asbestos minerals show a greater in-
   cidence of various types of gastroin-
   testinal carcinoma. -' • '• As  a  con-
   sequence, public health authorities
   in most civilized countries are re-as-
   sessing  the significance of the  pres-
   ence of asbestos minerals in the envi-
   ronment.  The  discovery that many
   Canadian water supplies contain up-
   wards of 1 million asbestos fibres per
   litre '- " has also given cause for sur-
   veys of municipal water sources to
   be made, whilst in the United States,
   disposal of mine tailings in Lake Su-
   perior in the vicinity of Duluth, Min-
   nesota has led to an extended  Fed-
   eral court action 1J  against  the
   company involved. These  recent
   events indicate the  seriousness with
   which the authorities  now view the
   natural presence of, or the discharge
   of,  this material into the environ-
   ment, and it appears to be  only a
   matter of time before some guide-
   lines are established concerning ac-
   ceptable levels in both air and water
by E. J. Chatfield and H. Pullan
Department of Applied  Physics
 Ontario Research  Foundation
     Sheridan  Park, Ontario

    for the general population. '"• On the
    other hand, the importance  of as-
    bestos in most sectors of the econ-
    omy cannot be denied

      As  previously mentioned,  occu-
    pational levels for workers in the as-
    bestos industry have existed for some
    time, however, there appears to be a
    considerable divergence  of opinion
    on just how these levels should be es-
    tablished. The recommended  North
    American occupational MFC  (Max-
    imum  Permissible  Concentration)
    for air has been  set at 5 fibres/ml;  '"
    a further condition is that only those
    fibres greater than 5 jum in length are
    included in this figure. At the time of
    writing there is  no MFC for  inges-
    tibles. The occupational MFC in the
    UK has been set at 2 fibres/ml, '"i:
    with the same fibre length limitation.
    In this case, some discrimination be-
    tween the various asbestos mineral
    types is exercised, and mass concen-
    tration values are also equated  to
    these  for convenience of measure-
    ment. It  must  be emphasized that
    these are all OCCUPATIONAL lev-
    els which are  established using sta-
    tistical data and by defining some
    small incidence of asbestos-related
    disease in the industry as acceptable.
    If the  normal  philosophy used  in
    radiation protection were to  be ap-
    plied, acceptable levels for exposure
    of the general population would log-
    ically  be set at  substantially  lower
    values, perhaps one order of magni-
    tude or so  lower. The design of  a
    suitable measurement technique for
    asbestos  minerals  must  therefore
    take account of the low levels to be
    measured  For air, a suitable  lower
    detection limit appears to be  about
    0.01 fibres/ml, whilst for water the
    corresponding value would be  about
    1O fibres/litre,  this latter figure be-
    ing about 100 times less than the av-
    erage  concentrations  normally
    found.

     Asbestos is a generic term used for
    two families of minerals which have
    a fibruous texture and which can be
    split into individual, sometimes very
    flexible fibres. The first group are the
    serpentines, of which  chrysotile  is
the only fibrous member. This min-
eral comprises most of the world's
production, and nearly all of the Ca-
nadian contribution to it  It is a hy-
drated magnesium  silicate, having
the composition Mg.Si.O, (OH),,
and is not the "indestructible" mate-
rial it is commonly thought to be  It
is attacked  by even the weak acids,
such as acetic acid, and decomposes
on  heating at 450C to  Forstente,
which is not fibrous, although it may
still retain the fibrous morphology of
the parent material Individual fibres
of chrysotile have been observed to
decompose  at temperatures as low as
250C'" The other  group  of fibrous
minerals are the amphiboles, which
have  the  general composition
X;Si_O,,(OH), where X may be Na#,
Fejd, Fe'd,  MgJ#, or CaJ$ in various
combinations.1" These minerals pos-
sess greater resistance to acids and
heat than chrysotile, and a number
of specific compositions are  recog-
nized and named as individual min-
erals. The principal types thus recog-
nized are amosite,  grunente,
anthophylhte, crocidolite,  actinohte
and tremolite. The amosite-grunente
series  has   a  variable  composition
which has  led to the naming of a
number of other sub-species such as
Cummingtonite, the mineral in-
volved in the western arm of Lake
Superior.

  From the  compositions  given
above, it can be seen that many other
mineral  species may have similar
compositions, yet not be asbestiform
types. The only identification tech-
nique open to  us,  therefore,  is one
sensitive to  the crystal structure, i.e
either electron  or X-ray diffraction
Chemical analysis by itself  is not
useful. In some cases, only a combi-
nation of the two is adequate  Since
most of the  health criteria are stated
in terms of fibre number concentra-
tions,  all  mass measurement  tech-
niques, including X-ray  diffraction,
are excluded. In any case, X-ray dif-
fraction  is  of  marginal  sensitivity,
particularly   when  other  common
minerals such as kaolmite are also
present.-"1 Along with most other m-
                                                    11

-------
  Figure 6-Elcctron diffraction pattern of amos/fe
Figure 5—Election micrograph of amoiite  Arf.CCC1
 Figure 4—Electron ijiltraction pattern^of crocidolite
                                                             Figure 3—Electron micrograph ot ciucicjolne AJJ
Figure 2—Electron diffraction pattern ot chrysolite asbestos
                                                              Figuie  1 —Electron  micrograph  of chrysotili'-  asbestos,
                                                              X84.000
                                                          12

-------
vestigators. we  elected for the only
obvious approach, particle counting
by  electron microscopy, combined
with electron diffraction  for identi-
fication.1J  -'--'' Although scanning
electron  microscopy has also  been
suggested as a suitable alternative, it
is evident that identification can only
be  based on  morphology, with per-
haps some  chemical  data  from
energy   dispersive  X-ray analysis.
Optical  microscopy, using  tech-
niques such as that currently used for
health control in the industry, is also
inapplicable to  trace measurement in
environmental  samples,  since this
technique assumes all fibrous  mate-
rial to be asbestos and lacks any easy
identification  technique  for  small
particles. It is also found that  many
fibres have widths much  lower than
those capable of being resolved by
the optical microscope, even though
their lengths may exceed the  5  /xm
length limit sometimes specified.

  The whole  topic of trace asbestos
measurement is surrounded by con-
troversy, and the area subject to most
dispute is undoubtedly that of speci-
men preparation. The requirement
of  the preparation  technique  is  to
quantitatively deposit the solid con-^
tent of an air or water sample onto
an  electron  microscope  specimen
support film. It would be desirable to
use  the  technique of Kalmus,J?  in
which direct dissolution of the mem-
brane  filter is  achieved by  reflux
washing  in acetone vapour, thus de-
positing the paniculate material
quantitatively  on a  carbon  coated
electron  microscope grid. This  tech-
nique suffers  from the disadvantage
that some smaller  particles are
washed away; a fact that can easily
be  checked by processing a radio-
active particulate  sample. More sig-
nificantly, raw  water  samples  often
contain an overwhelming  proportion
of organic material, which necessi-
tates dilution  of the sample so that a
reasonably loaded  electron  micro-
scope sample can be obtained. This
dilution has the undesirable effect -of
separating  the asbestos fibres  more
widely, thus requiring more counting
time. Also the large amount of extra-
neous  organic material  obscures
many of the  fibres. It is therefore
preferable  to contrate the asbestos
fibres at  the expense of the organic
materials present.

  The technique developed for  anal-
ysis of water  samples at  ORF is a
modification of those described  by
Cunningham  and Pontefract,1-' and
Biles and Emerson.-'' At this time it
represents the only  published  tech-
nique which has been tested for  mass
balance using standard asbestos dis-
persions. Even  this  has  only  been
tested in the case of chrysotile  The
same method  is  also  used  for air
samples collected on membrane fil-
ters.
  The first step in the procedure is to
collect some of the solid material on
a filler. In the case of air, the collec-
tion properties  of membrane filters
allow pore sizes of 0.4 jum or even 0.8
,um  to be  used,  and a  volume "of
about 5m is filtered. For water sam-
ples, the particle  sizes  collected are
strictly a function of pore size, and
the  smallest pore  size compatible
with a  reasonable flow-rate  is se-
lected. In practice 0.1 fus the smallest
convenient pore size, and a volume
of about 200ml  of the water is fil-
tered.  The  remaining  steps in  the
alaytical  procedure for  both types of
sample  are identical.  The filter is
transferred  to  a  clean  glass   vial,
which is then placed in a plasma mi-
cro-incinerator (low  temperature
asher). In this device the filter is oxi-
dized, along with all other organic
materials present, with very little dis-
turbance at a  temperature  of less
                        than 80CV- The oxidation takes
                        place  in  oxygen at a  pressure of
                        about  1 Torr, which is excited by a
                        radio-frequency discharge. After
                        some hours the vial containing the
                        residue is removed and double dis-
                        tilled water added. The ashed resi-
                        dues are  gently dispersed ultrason-
                        ically and an aliquot  of  the
                        dispersion is centnfuged on to a 1cm
                        diameter  cover  glass at an  acceler-
                        ation of 8,000g for about 20 minutes
                        In practice,  a drop of very dilute de-
                        tergent is also  added to the centri-
                        fuge tube for a reason  which is re-
                        ferred  to later.  The disc is removed,
                        dried and a thin carbon coating ap-
                        plied by  vacuum evaporation.  The
                        carbon film  is scored  by a scalpel
                        blade,  and is then floated  on to wa-
                        ter,  carrying the deposit of particles
                        with it.  The  detergent which  was
                        added  to  the centrifuge tube assists
                        in the removal  of  the  carbon  film
                        from the glass disc. Pieces of the car-
                        bon are then picked up on 200 mesh
                        electron microscope support  grids.
                          About  10 grid squares,  selected
                        from several grids, are  searched for
                        asbestos  particles using a trans-
             ASBESTOS   FIBRE   COUNT   ANALYSIS


   SIZE     PARTICLE   NUMBER    CUM    CUM  NO
CATEGORY   SIZE,  UM   COUNTED  NUMBER  PERCENT
     I
     2
     3
     A
     5
     6
     7
     8
     9
    10
    11
    12
    13
    14
    15
    16
    17
 0.040
 0.057
 0.080
 0. 1 13
 0.J60
 0.226
 0.320
 0.453
 0.640
 0.905
 1.280
   810
 I
 2.560
 3.620
 5.1 20
 7.241
10.240
 0
 0
 1
20
16
16
16
 8
 7
 2
 I
 1
 0
 0
 1
 •0
 0.0
 0.0
 0.0
 0.5
11.0
29.0
45.0
61.0
73.0
80.5
85.0
86.5
87.5
88.0
88.0
88.5
89.0
  0.00
  0.00
  0.00
  0.56
 12.36
 32.58
 50.56
 68.54
 82.02
 90.45
 95.51
 97.19
 98.31
 98.88
 98.88
 99.44
100.00
DATE:  15/12/73

 CUM MASS
 PERCENT

   0.00
   0.00
   0.00
   0.09
   2.86
   8.94
  16.88
  27.47
  38.87
  49.50
  58.00
  64.23
  71.53
  75.81
  75.81
  87.90
 100.00
       TOTAL  NUMBER  OF  PARTICLES COUNTED =    89
       SIZE OF  GRID  SQUARE USED           =    85 MICROMETRES
       NUMBER OF GRID SQUARES  COUNTED    =    10
       DENSITY  USED  IN  MASS  CALCULATION  :     2.40  G/CC

       TYPE OF  ASBESTIFORM COUNTED   - CHRYSOTILE

                                                 6
 CONCENTRATION OF ASBESTOS      = 1.93 X 10    FIBRES/LITRE
                                               -3
 ESTIMATED MASS CONCENTRATION  : 3.93 X 10    MICROGRAMS/LITRE
                                                It
 I.E.  ONE PART OF ASBESTOS  IN    2.55 X 10     PARTS OF LIQUID

           LOWEST  DETECTABLE LEVELS  UNDER THE
           CONDITIONS USED IN THESE  MEASUREMENTS
                                   4
           NUMBER   -   2.17 X 10    FIBRES/LITRE
                                  -6
           MASS     -   7.42 X 10    MICROGRAMS/LITRE

Figure 7—Computer printout for a chrysotile-in-water sample
                                                 13

-------
               ASBESTOS FIBRE LENGTH DISTRIBUTION
                                                         DATE:  IV12/7J
FIBRt LtNGTH,
MICROMETRES
  10+
   8+

   6+

   4+
   2+



   1 +
 0.8+

 0.6+


 0.4+


 0.2+
 0.1 +
 .08+
TYPE OF ASBESTIFORM
NUMBER OF FIBRES SIZED
CHRYSOTILE
89
 .06+
    0.5 1
                       10   20  50  40  50  60   70  80   90   95   98 99 .5
                         CUMULATIVE  PERCENTAGE NUMBER
              LESS THAN STATED LENGTH   (LOT,.  PROBABILITY SCALE)

 Figure 8—Fibre length distribution of a typical sample
 Figure 9—Removing a set of samples from the asher
mission  electron  microscope  at  a
magnification    of   approximately
25,000.   Asbestiform  particles  are
identified individual!}  by  electron
diffraction,  and  their  lengths  and
widths are measured
  The particle  counting is normally
terminated after 10 grid squares have
been  examined, or when about 100
particles have  been  counted  Both
the detection level and  the accuracy
can  be  improved  by  additional
counting, but  for  economic  reasons
these arbitrary  limits have  usually
been  applied  The minimum  level,
i.e. one  particle detected,  corre-
sponds  to about  10' particles/litre
for water samples, or 001  particles/
ml  for air samples, whilst detection
of about 100 particles yields an accu-
racy of \0r/f
  The data  are  then  processed by a
computer program, which  calculates
both  number and mass  concentra-
tions, and also plots the  size distribu-
tions.
   It is the practice of some workers
to identify only a minor proportion
of the asbestos particles  by diffrac-
tion  Although chrysotile  may  be
identified primarily by  its  character-
istic morphology in the transmission
electron image, the amphiboles pos-
sess no  such  characteristic  appear-
ance  In water samples,  particularly,
diatomaceous  and mica  fragments
can often be mistaken for amphibolc
fibres It proves little  to identify only
 }5''i of the reported fibres by diffrac-
tion,  and the ORh alaytical team re-
port  only  those  amphibole  fibres
which have been  so confirmed.
   1 he technique  has been criticized
for Us use of ultrasonics to redisperse
the sample  residues after the ashing
procedure  The concern is  that the
ultrasonic treatment  may break up
the fibres, giving rise  to an artificially
 high measurement In fact if can he-
shown  that power densities  of some
 watts/ml are required before signifi-
cant  breakage of suspended  particles
 occurs,  whereas the  power densities
 used in  Ihis technique are only a few
 milliwatts/ml   However,  although
 we do  not regard this as a cause lor
 concern, there are many unresolved
 questions K>r  example, water sam-
 ples  have to be collected in the field.
 and  these are  usually stored in bot-
 tles during transit to the laboratory
 Nothing is known  about  the  scav-
 enging action  of  the bottle's interior
 surfaces on the suspended  particles
 during  storage, the effect of pH, or
 whether  plastic  or glass  bottles
 should be used on this account. This
 effect may also be dependent on par-
 ticle concentration The only reliable
 procedure may be that of immediate
                                                              14

-------
filtration. The mass balance in many
of the  analytical methods in use  is
also suspect, and may indicate that
results being obtained are lower than
the real values.

  Contamination  is  a very  serious
problem: almost all  reagents, glass-
ware, water, etc.,  are contaminated
in some  degree with chrysotile as-
bestos,  and the most extreme  pre-
cautions must be taken  to  eliminate
this.  Even  some  membrane filters
have been found to contain amounts
of chrysotile which can  disrupt the
measurement  At  ORF  we perform
critical  phases of sample  preparation
in a positive  pressure clean room,
from which all known sources of as-
bestos  have been  excluded  The air
supply  is filtered and passed  through
an   electrostatic  precipitator.  The
floor is of pure vinyl, rather  than vi-
nyl-asbestos, and a suspended ceiling
was fitted to prevent possible fallout
from insulated  air ducts and pipes.
To minimize the dust problem,  steel
furniture was  installed  rather  than
the wood variety. Any visible dust  is
treated with suspicion and  cleaned
up using wet tissues to prevent its
dispersal.   Disposable  laboratory
coats and overshoes  are  used by all
personnel entering this  area  Glass-
ware is cleaned before  use in chro-
mic acid, and then rinsed  in double
distilled  water. Only by  observing
the   strictest  hygiene,  comparable
with that necessary during handling
of radioactive isotopes, is it  possible
to  maintain  the   low  background
measurements which we  routinely
achieve Even  use  of some types of
talcum powder or cosmetics by the
individuals performing sample prep-
aration can  cause a perplexing series
of  contaminated  samples to  arise
where  blank measurements were ex-
pected.

  The  techniques  in  use  at ORF for
trace  asbestos  measurement   have
been developed over a  period  of
three years. During that time many
of the  difficulties  have been identi-
fied and  solutions  found  However,
there is a pressing need for further
development  work,  particularly  in
the field of quantitative sample prep-
aration A  development  program  is
therefore being initiated at ORF to
investigate  the various  aspects  of
sample preparation and storage, with
the  eventual aim of defining a rec-
ommended  analytical  procedure
which  has  been thoroughly tested.
Until this program is completed, the
current method, which gives reason-
ably reliable data at an economically
acceptable  cost, will  remain  in  use
References
 1 Selikofl I  J  et al Arch Environ  Health V ol 25
   1972 page 1
 2 Mancuso  1  F and Coulter F J  Arch Environ
   Health Vol  6  1963 page 210
 3 Howihane D O  Ann  NY Acad Sci  132  165
   page 647
 4 Wagner J C' Ann NY Acid Sci  132  1965, page
   575
 5 Knox j  I  et al Bnl  J  Indus!  Mcd  25 1968
   page 293
 6 Newhouse M L and V. agner J C  Brit J  Indus!
   Med 26. 1969 page 302
 7 Real F  f  Fancet n  1960 page  1211
 8 Hmson K F V.  Bit  I Dis Chest 59  |965  page
   121
 9 Federal Register {L S A 1  Vol 32 No  110 1972
   page 11318
 10 Statutory  Instruments (L K ) 1969 No 691) Facto-
   ries  I he  Asbestos Regulations
 1) Nevcbouse M  1  and  I hompson H  Bnl I  ln-
   dustr Med  22  1965. page 261
 12 C unnmgham H  M and  I'ontetract, R  Nature
   232  1971 page 332
 13 Kav  (j  H  lournal A W V> A  Sept  1974  page
   513  and  Watei and Pollution Control Sept  1973
15 Bnckman  L and Rubmo RA "National Behind a
  Proposed  Asbestos  A;r Qualitv  Standard"  67th
  Meeting ol Air Pollution C onlrol Association  Den-
  ver  Col  1974
16 Aver II 1  et al Ann N\ Acad Sciences 132 Arl
  I 1965 page 274
17 Brit  Oec Health Soc  Report ot ( ommiltee on Hv -
  giene Standards  Hvgiene Standards lor C hr\sotile
  Asbestos Dust  fXc  1967
18 Berrv F  Private C ommumcation
19 Industrial  Minerals and Rods Fd .1  I  (nlson
  AIM!   page 48  (1968|
JO Rickards  A  I  and Badami  D  \ Nature 234
   1971 page 93
Jl Smith Cr  R  et al Proceedings Fleetron Micro-
  scopv Societv ol America 1973 page 310
22 \\alkcr C  V,  ct al Proceedings Fleetron Micro-
  scopv Socielv ol \merica 1974 page s24

23 Mudroch  <) and Kramer I  R Ibid page 526

24 C hatlielcl  E J  Ibid page 52k
js ( hatlicld  I I  Proceedings  Microscopical Societv
  ol Canada 1974 page 24

26 Biles B  and  i merson  I  R Nature 219  1968
   page 93

27  Kalmus 1 H  I Appl  Plus 2s (19^41

2K  Thomas. R S  and C orletl M I Proceedings Flee-
   tron Microscopv Noeietv ol America  1973 page
   334                               fj
 Figure 10—Identifying an  asbestos fibre in the transmission electron microscope

-------
APPENDIX F EPA NATIONAL WATER QUALITY LABORATORY X-RAY DIFFRACTION ANALYSIS
           RESULTS OF PILOT WATER TREATMENT UNITS - RAW WATER FROM DULUTH
           LAKEWOOD INTAKE.
Date of
sample
4/19
4/25
5/7
5/9
5/16
5/16
5/16
5/16
5/16
5/22
5/22
5/22
5/22
5/22
5/30
5/30
5/30
5/30
5/30
6/4
6/4
6/4
6/4
6/4
6/6
6/6
6/6
6/6
6/6
6/11
6/11
6/11
6/11
6/11
6/17
6/17
6/17
6/17
6/17
6/28
6/28
Filter
utilized
MM- 2
MM-2
BIF
BIF

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-1

MM-1
MM-2
BIF
ERD-1


Run
No.
1
7
10-T
12-T

7
37
8
10

16
43
11
13

26
50
17
14

29
53
19
15

31
54
20
15

34
55
24
1A

39
61
31
4A


Amphibole mass
SS (mg/1) concentration (mg/1) Per Cent
Raw Finished
0.90 0.43
0.86 0.39
0.69 0.09
0.61 0.31
0.64
0.13
0.05
0.12
0.05
0.74
0.33
0.13
2.63a
0.04
0.68
0.10
0.05
1.02a
0.10
0.57
0.09
0.02
0.11
0.10
0.77
0.12
0.04
0.17
0.36a
1.30
0.24
0.07
0.59a
1.47
0.86
0.12
0.08
3.22a
0.06
0.74
2.09b
Raw Finished
0.10 <0.02
0.12 <0.02
0.06 <0.01
0.22 0.003
0.14
0.005
0.003
0.006
0.003
0.20
<0.01
<0.003
<0.03
0.002
0.19
<0.01
0.005
<0.02
0.004
0.26
<0.006
<0.006
0.004
0.003
0.18
<0.005
<0.005
0.004
<0.004
0.18
0.006
<0.004
0.008
<0.02
0.17
<0.006
<0.01
<0.03
<0.005
0.16.
f\
0.11°
Removal
80
83
83
98

96
98
96
98

95
98

99

95
97

98

98
98

99

97
97

98

97
98

89

96
94

97


                                          16

-------
APPENDIX F  (CONTINUED)
Date of
sample
6/28
6/28
6/28
6/28
7/3
7/3
7/3
7/3
7/3
7/19
7/19
7/19
7/19
7/19
7/23
7/23
7/23
7/23
7/23
7/25
7/25
7/25
7/25
7/25
7/30
7/30
7/30
7/30
7/30
7/30
7/30
7/31
7/31
8/1
8/1
8/1
8/1
8/1
8/6
8/6
8/6
8/6
8/6
8/8
8/8
8/8
Filter
utilized
MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-2

MM-1
MM-2
BIF
ERD-2

MM-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
Run
No.
48
76
47
28

52
78
51
32

67
96
70
40

70
99
72
41

72
101
73
43

104

74
105
77
45

106

76
107
79
46

78
109
82
48

80
111
Amphibole mass
SS (mg/1) concentration (mg/l)Per Cent
Raw Finished
0.76
0.04b
0.38
0.68
0.75
0.11
0.08
0.65a
0.12
0.88
0.10
0.18
0.60a
0.04
0.59
0.03
0.08
0.70a
0.06
0.68
0.08
0.10
0.45a
0.03
0.68
0.78°
0.64
0.03
0.14
0.13
0.04
2. 58 b
0.09b
0.90
0.06
0.06
0.24a
0.06
0.70
0.03
0.06
0.71a
0.03
0.64
0.06
0.10
Raw Finished
0.007
<0.005b
<0.008
<0.01
0.07
<0.004
<0.003
<0.004
0.003
0.13
<0.003
<0.005
<0.003
<0.003
0.09
<0.003
<0.004
0.007
<0.003
0.04
<0.005
<0.005
<0.004
<0.003
0.10
0.02
0.06
<0.003
<0.005
<0.003
<0.003
0.10b
<0.003b
0.08
<0.004
<0.003
0.01
<0.002
0.09
<0.003
<0.003
<0.007
<0.002
0.05
<0.003
<0.003
Removal
97
95

91

95
96

96

98
96

98

97
95

97

87
87

92

80

95
92

95

97

95
96

97

97
97

98

94
94
                                         17

-------
APPENDIX F  (CONTINUED),
Date of
sample
8/8
8/8
8/13
8/13
8/13
8/13
8/13
8/15
8/15
8/15
8/15
8/15
8/20
8/20
8/20
8/21
8/22
8/22
8/22
8/23
8/23
8/23
8/28
8/28
8/28
8/30
8/30
8/30
8/30
9/4
9/4
9/4
9/4
9/6
9/6
9/6
9/6
9/9
9/9
9/9
9/10
9/10
Filter
utilized
BIF
ERD-2

MM-1
MM- 2
BIF
ERD-2

MM-1
MM- 2
BIF
ERD-2

MM- 2
ERD-2
MM-1

MM-1
MM-1

MM- 2
ERD-2

MM-2
BIF

MM-2
BIF
ERD-2

MM-2
BIF
ERD-2

MM-2
BIF
ERD-2

BIF
ERD-2

MM-2
Run
No.
84
49

82
113
88
51

84
114
89
55

118
59
86

86
86

119
63

122
103

124
105
68

126
108
70

128
111
72

113
73

131
Amphibole mass
SS (mg/1) concentration (mg/l)Per Cent
Raw Finished
0.59a
0.04
0.67
0.03
0.04
0.47a
0.09
0.81
0.08
0.05
0.25
0.16
0.81
0.09
0.04
0.05
0.70
0.06
0.06
0.63
0.02
0.05
0.54
0.02
0.14
0.41
0.03
0.27a
0.008
0.72
0.03
0.55a
0.03
0.61
0.04
0.38a
0.01
0.60
0.10
0.03
0.62
0.08
Raw Finished
<0.006
<0.002
0.06
<0.003
<0.003
<0.005
<0.003
0.06
<0.003
<0.003
<0.003
<0.003
0.02
<0.003
<0.003
<0.003
0.03
<0.003
<0.003
0.04
<0.003
<0.003
0.09
<0.003
<0.003
0.07
<0.003
<0.003
<0.003
0.10
<0.003
<0.006
<0.003
0.07
<0.003
<0.003
<0.003
0.04
<0.003
<0.003
0.05
<0.003
Removal

96

95
95

95

95
95

95

85
85
85

90
90

92
92

97


96

97

97

97

96

96


92

94
                                         18

-------
APPENDIX F  (CONTINUED)
Date of
sample
9/11
9/11
9/11
9/13
9/13
9/13
9/13
9/16
9/16
9/16
9/16
9/17
9/17
9/19
9/19
9/19
9/19
9/20
9/20
Filter
utilized

MM-2
BIF

MM-2
BIF
ERD-2

MM-2
BIF
ERD-2

MM-2

MM-1
MM-2
BIF

ERD-2
Run
No.

133
115

137
117
78

138
118
79

139

87
140
120

85
SS (mg/1)
Raw Finished
0.63
0.08
0.52a
0.51
0.03
0.25
0.02
0.31
0.04
0.39a
0.04
0.58
0.07
0.50
0.06
0.03
0.40a
0.40
0.003
Amphibble mass
concentration (mg/l)Per Cent
Raw
0.06


0.09



0.08



0.06

0.10



0.07

Finished

<0.003
<0.005

<0.003
<0.003
<0.003

<0.003
<0.004
<0.003

<0.003

<0.003
0.003
<0.004

<0.003
Removal

95


97

97

96

96

95

97
97


96

 a Mostly DE
 b Cloquet Pipeline water
 c Mostly alum and DE, sample purposely collected after  turbidity break-
   through
                                       19

-------
     At the concentrations of fibers encountered in Lake Superior water,  there
appeared to be no correlation between turbidity and fiber counts for either
amphibole  or chrysotile fibers.  However, there was a relationship between
raw water amphibole mass and amphibole fiber counts, as shown in Figure  1.
Using linear regression analysis, the following was obtained:

    F/l = (0.15 + 6.2 Mass) x 106

    F/l = amphibole fiber count, 10  f/1

    Mass = amphibole mass, mg/1

The correlation coefficient, r, was 0.48.
               4.0
          cr.
          o
O  £
00 —
         >• h-
         I- <
               3.0
            LJ
            o
               2.0
         a:
         UJ
            o
         z x
         o a-
               1.0
                             1.0          2.0          3.0

                        AMPHIBOLE   FIBERS  /LITER X |Q6
                        ONTARIO  RESEARCH  FOUNDATION
                                                        4.0
      FIGURE   |.    CORRELATION  BETWEEN  NWQL AMPHIBOLE MASS CONCENTRATION
                  AND ORF AMPHIBOLE FIBER COUNTS-RAW WATER AT DULUTH
                  LAKEWOOD INTAKE  -1974.
                                      20

-------
    o
    (- 0.9
      0-8
    O
    o
      0.7
    O
    m 0.6
^  Q.
   0.5
 O"  Q
 e  z
    **• 0.4
    CO
    Q

    O 0.3
    CO
Q
z
UJ
Q.

CO
      0.2
      o.i
           .O SUSPENDED  SOLIDS
            • AMPHIBOLE
5    15    25
    APRIL
                                15
                               MAY
                                     25
                                       5    IS    25
                                           JUNE
 IS    25
JULY
5    IS    25
   AUGUST
5    15    25
 SEPTEMBER
            FIGURE 2.    ENVIRONMENTAL PROTECTION AGENCY
                         NATIONAL  WATER QUALITY  LABORATORY
                         SUSPENDED SOLIDS AND  AMPHIBOLE  MASS CONCENTRATION
                         RAW WATER AT DULUTH LAKEWOOD INTAKE - 1974

-------
      ADVANCES  IN
    X-RAY ANALYSIS
                Volume 18
                Edited by

             William L. Pickles
           Lawrence Livermore Ijiboratory
              Liverniore, California

                  and

Charles S. Barrett, John B. Newkirk, and Clayton O. Ruud
             Denver Research Institute
             The University of Denver
               Denver, Colorado
                Sponsored by
              University of Denver
            Denver Research Institute
        Metallurgy and Materials Science Division
  PLENUM PRESS • NEW YORK AND LONDON
                 22

-------
 The Library of Congress cataloged the first volume of this title as follows:


       Conference on Application of X-ray Analysis.

            Proceedings.  6>.h-                  1957-
          iDenveri
               T.  tllus.  24-28 cm  annual
            No proceedings published for the tint ~t i-onferenoes.
            Vols.  for VJfift-    called al*'> Advances in X-tuy analyst*, T 2-
            Proceedlne* for  19~»7 f"'jed by rip Ci.nferercs un^fr an et\-.ier
          name: Conference  on Industrial Applications of  X-ray Analyst*.
          Other slight rnri.itlons In name o? conference
            Vol. for 19T>7 ptjbl'
-------
SEMI-QUANTITATIVE DETERMINATION OF ASBESTIPORM AMPHIBOLE MINERAL

CONCENTRATIONS IN WESTERN LAKE SUPERIOR WATER SAMPLES


            P. H. Cook

            National Water Quality Laboratory, U.  S. EPA

            Duluth, Minnesota  55804


                             ABSTRACT

     The amphibole mineral, cunmingtonite-grunerite, has been used
as a tracer for taconite tailings discharged into  Western Lake
Superior.  The discovery of many asbestiforra amphibole fibers in
the tailings and Western Lake Superior water lead  to concern over
fiber concentrations in municipal water supplies using this water.
This concern was based on the association between  human asbestos
exposure and increased rates of cancer of the gastrointestinal
tract and peritoneum.  An x-ray diffraction external standard tech-
nique has been developed for rapid, inexpensive, semi-quantitative
determinations of amphibole mass concentration in  water.  The av-
erage amphibole mass concentrations for different  Western Lake
Superior water intakes compare very well with the  average electron
microscope fiber counts for the same samples.  Daily  amphibole
analysis of the Duluth water supply indicates an average amphibole
concentration of 0.19 milligrams per liter.


                           INTRODUCTION

     For several years x-ray diffractometry has been  the key an-
alytical technique for National Water Quality Laboratory studies of
the distribution and fate of taconite tailings which have been
discharged into Western Lake Superior at Silver Bay, Minnesota
since 1956.  A major component of this 67,000 ton  per  day discharge,
the amphibole mineral cummingtonite-grunerite, provides an ideal
tracer for the tailings.  The cummingtonite-grunerite  (310) peak
at 29.1° 20 for copper K= radiation (d = 3.07 A) is not found in
x-ray diffraction patterns for natural lake sediments  or suspended

                                557
                               24

-------
558
                               P. M. Cook
          373 m«/P
                                     LATITUDE N47M341
                                     LONGITUDE W9-I5T
                                     LAKE DEPTH 27> METEKS
          9EBWCNT <
          so-raw*
         7S-100 im
         Ming
 FIGURE 1	X-RAY  DIFFRACTION PATTERNS  (COPPER RADIATION)  FROM SEDI-
 MENT SAMPLES TAKEN AT SUCCESSIVE 25  MM INTERVALS IN AN  AREA OF
 TACONITE TAILINGS DEPOSITION.  CUMMINGTOKITE-GRUNERITE,
 (Mg,Fe)7Si8022(OH)2,  PEAKS ARE SHADED.   THE (110) PEAK  AT APPROXI-
 MATELY 10.6° 29 IS COMMON TO MOST AMPHIBOLES
                                 25

-------
                             P. M. Cook
                                                                559
solids.   X—ray diffraction patterns  (Figure  1)  of  lake water sus-
pended solids which contain taconite  tailings  and  sediment trora
successive 25 ram sections  of the  lake bottom in an area of tailings
deposition show a clear gradation from large amounts of
cununingtonite-grunerite (shaded peaks)  in very  recent surficial
sediments to no cummingtonite-grunerite and  little amphibole in the
older, underlying sediments (75-100 mm).   X-ray diffraction study
of hundreds of river suspended sediment samples also indicates no
detectable cummingtonite-grunerite (<1%)  and only  1-2% amphibole in
the natural sediments entering Western Lake  Superior.  Much or all
of the trace amphibole is  the common, non-asbestiforra mineral horn-
blende.

     Further indication of the recent addition  of  cummingtonite-
grunerite to Western Lake.  Superior water is  provided b\ x-ray dif-
fraction patterns of many  suspended sediment samples saved from the
years 1940, 1950, and 1964 (Figure 2).   All  samples from 1940 and
1950 did not contain detectable amounts of cumr.ingtonite-grunerite
and little If any other ampljifaole minerals as indicated by a (110)
peak at 10.6° 29 (d = 8.34 A), All of the 1964 samples, however,
contained large concentrations of cummingtonite-grunerite as shown
by the appearance of large (110)  and  (310) peaks.   The (110;/(310)
peak ratios for these samples are typical of these found for tac-
onite tailings samples.
                                             0.1940
               
-------
560                          p. M. Cook

     In 1973, study of the morphology of amphibole particles in fine
taconite tailings by transmission electron microscopy revealed the
presence of many asbestiform fibers (Figure 3).   The realization
that many of these cummingtonite-grunerite fibers are indistin-
guishable from amosite asbestos fibers lead to concern over the use
of Western Lake Superior water for municipal drinking water
supplies.  This concern was based on the association between human
asbestos exposure and increased rates of cancer of the gastro-
intestinal tract and peritoneum (1) and daily x-ray diffraction
analyses of Duluth, Minnesota drinking water samples which indicated
the constant presence of high concentrations of taconite tailings.
Transmission electron microscope analysis of Duluth water samples
confirmed the presence of many amphibole fibers.

     Since the discovery of asbestiform amphibole fibers in the
water supplies of Silver Bay, Beaver Bay, Two Harbors, Duluth, and
Cloquet, Minnesota, extensive sampling programs by the Environmental
Protection Agency and other groups have been undertaken for electron
microscope fiber counts.  These analyses while in agreement with
the x-ray diffraction results, are very expensive, time-consuming,
and Imprecise.  At this time fiber counts done by different lab-
oratories are not comparable and intralaboratory replicate results
usually vary by ± 50% of the mean.   The amphibole fiber concen-
trations generally correlate with the. asphitule mass concentrations
determined by x-ray diffraction.  Thus x-ray diffraction nonitoring
of water samples combined with occasional electron microscope fiber
counts offers a faster, less expensive, and probably more accurate
measure of amphibole fiber contamination.  This technique has been
particularly useful for evaluating various filtration media's abil-
ities to remove amphibole fibers from drinking water.
                AMPHIBOLE ANALYSIS OF WATER SAMPLES

     Water samples from Western Lake Superior public water supplies,
normally ten liters in volume, are pressure filtered through 0.45p
•membrane fibers.  When the turbidity of the sample is known, the
volume filtered is adjusted to give a 4-8 mg sediment sample.  The
total suspended solids are determined by difference and a. weighing
correction applied to compensate for a small filter weight loss
due to leaching (2).  Distilled water blanks are run periodically
to check for contamination.  The dry membrane filter with sample is
fastened to a glass slide with a thin filia of lacquer, the filter
edges trimmed, and the slide directly examined with a Norelco ver-
tical diffractometer (copper Id radiation) with a graphite crystal
focusing monochromator.

     The amphibole fibers and cleavage fragments assume a preferred
orientation such that the c-axis, which corresponds to the long
dimension of the fiber,  is parallel to the filter surface.  This
                               27

-------
                            P. M. Cook
                                                               561
            r^ -^v*-?'''  rf^'*
                                                           »,,
                                                     *'   XV
                                *
                                s.
                                           i
                                        VV_»
        3
        3q-..
                                      •>-t*.;'
                                    '^*      V
FIGURE 3	ELHCTRON MICKOCRAPH OF <2u TACONITE TAILINGS.   a)  LOW
MAGNIFICATION  C.SC^X).  fa) HIC.ilER MAGSIFICAriOX U2.500X)  VIEW OF
AX AMPHIBOLE FIBER BUNDLE
causes Che  (110)  reflection and, to a lesser extent,  the (310)
reflection  intensities  to 3e enhanced, permitting the detection  of
trace amounts  of  .npliibcle.  As little as 0.05 ng of  <2\.
curamingtonite-grunerite produces measurable (110) and (310)  peaks.

     A semi-quantitative measurement of the amphibole concentration
is made by  an  external  standard technique.  This technique has been
                                28

-------
562                          p. M. Cook

used to estimate trace amounts of chrysotile asbestos and amphibole
asbestos in dust samples on membrane filters (3,4) and fulfills the
need for rapid, standardized estimates of amphiboie concentration
in samples which are not amenable to the use of an internal stan-
dard.  Three potentially large sources of systematic error had to
be considered before accepting the external standard model; varia-
bility of particle size, sample mass absorption coefficient, and
amphibole preferred orientation.

     The external standard chosen for the preparation of standard
curves of x-ray peak intensity versus mass of amphibole was the
amphibole mixture found in the <2u taconite tailings.  This choice
was made since the predominant amphibole in Western Lake. Superior
water is cumaingtonite-grunerita from taconite tailings and natural
amphihole concentrations in Lake Superior water are normally not
detectable by x-ray diffraction.  The <2y taconite tailings were
determined by the x-ray diffraction of cumraingtonite-grunerite/
quartz mixtures to contain approximately 80% amphiboie and 20%
quartz.  Most of the amphiaole is cummingtonite-grunerite with some
actinolite-tremolite.  Larger size fractions of the tailings contain
less amphibole and more quartz with a small percentage of magnetite.

     Reference samples were prepared by adding known amounts of the
<2u amphibole standard fo ten liter samples of Lake Superior water
having no detectable amphibole minerals.  This water, obtained from
Grand Marais, Minnesota, (.oucainecl 0.4 mg/1 «uspeTd2d soliJs. which
consisted primarily of organic debris, diatoms, quartz, and clay
minerals.  These standard samples were then filtered and analyzed by
x-ray diffraction in the same manner as unknown samples.  The re-
sulting x-ray diffraction patterns are identical in appearance to
those for Duluth water samples.

     The <2y amphibole particle size  (by gravity settling) for the
external standard ^'as shown to be appropriate by a centrifugation si^e-
separation of Duluth water suspended solids from samples taken on
fifteen different days.  Ninety-five percent of the suspended solids
were in the <2u fraction witn only a small amount of amphibole in
the 5% which was >2p.  Thus variability in diffracted x-ray inten-
sity due to mineral particle size >2y is insignificant.

     The filtration of ten liters of Duluth water normally results
in 4-8 mg of suspended solids retained on the 0.45;. membrane filter.
When the suspended solids exceed 0.3 mg/1, smaller volumes are
filtered.  A sample weight of 8 rag and an average density of 2 g/cm~
results in a hypothetical sample thickness of 3u on the filter.
This thin sample thickness should preclude variability due to dif-
ferences in sample absorption coefficients.  Direct evidence for
this is provided by the linearity of a plot of percent amphiboie
versus x-ray intensity for samples in this weight range; the uniform
                                29

-------
                             P. M. Cook
                                                                563
intensity of filter background in the x-ray diffraction pattern with
increasing sample weight to 10 mg; and the linearity of a plot of
quartz peak (d = 3.33 A) intensity versus weight of quartz, regard-
less of total sample weight in the range 0—12 mg.
    AMPHIBOLE
    (110) PEAK
      X-RAY
    INTENSITY
   counts/second
                           MILLIGRAMS  AMPHIBOLE L2. >j

FIGURE 4	EXTERNAL STANDARD CURVE FOR AMPHIBOLE SEMI-QUANTITATIVE
ANALYSIS
     The non-linearity of the external standard curve (Figure H) is
due to a decreasing degree of preferred orientation as the ar.ount
of amphibole increases.  This is indicated by decreasing araphibole
(110)/C31Q) and amphibole (.110) /quartz peak ratios with increasing
weight of the standard amphibole—quartz mixture on the filter.  The
utility of the external standard curve depends on how well the curve
models amphibole preferred orientation in environmental samples.
Similar curves based on samples prepared with increased amounts of
natural sediment agreed well with the standard curve used.  K'ith
large amounts of natural sediment, the amphibole peak intensity is
weakened which would cause an underestimation of amphibole concen-
trations.  Other standard curves were employed to estimate the
amphibole concentration in the few samples with a very high
concentration of non-amphibole minerals.
                               30

-------
 564                          P. M. Cook

     External standard curves, such as Figure 4  were plotted from
 the nor.-lincar least squares refinement of araphibole mass versus
 amphibole  (130) peak intensity data points.  The data fit an equa-
 tion of the form: Ic = IQ + !«(!- exp-kC), where TC = intensity at
 concentration C (mg amphibole); Io = intensity at C = 0; !„ =
 intensity at C = °°; and k is a constant.  This equation is consist-
 ent with a model in which the degree of preferred orientation
 decreases as more amphibole particles are placed on the membrane
 filter.  Standard curves utilizing amphibole (110) peak height above
 background are identical to curves plotted from the (110) peak
 areas.  Both measurements are used and give the same amphibole
 concentrations for environmental samples.  Use of an amphibole (310)
 peak curve gives the same results but with less precision due to
 lower peak intensity.

     Replicate (five) analyses of Duluth water samples indicate a
 standard deviation of + 3% for determining amphibole concentrations
 in typical samples with 0.1-0.3 mg/1 amphibole.  For samples having
 lower amphibole concentrations (<0.1 mg/1) and high suspended solids
 (>1.0 mg/1), this precision is reduced to + 25%.  Overall suspended
 solids determinations have a standard deviation of + 67, of th? mean.
Detection limits for determining amphibole concentration depend on
 the volume of water filtered and can be as low as 0.5 ug/1.
                  WATER SUPPLY AMPHIBOLE ANALYSIS

     Daily analyses of Duluth water samples for amphibole and sus-
pended solids concentrations began in March 1973 and continues to
date.  Results through January of 1974 are shown in Figure 5 with
climatological data and intake water temperatures.  X-ray diffrac-
tion analysis provides a picture of daily and seasonal fluctuations
in amphibole and suspended solids concentrations.  For example,
periods of heavy rainfall are followed by abrupt increases in
suspended solids due to river run-off and shore erosion.  These
increases In suspended solids do not coincide with increases in
amphibole, indicating a different source for atnphibole sediment.

     Maximum amphibole concentrations (up to 0.8 mg/1) occur in
the late fall and spring.  Minimum amphibole concentrations  (0.04
mg/1) occur during the late summer and early fall when a thermo-
cline is present in Western Lake Superior.  The average amphibole
concentration measured was 0.19 milligrams per liter with 0.83
milligrams per liter total suspended solids.

     During the period August 22-November 28, 1973, personnel from
Region V of the Environmental Protection Agency obtained weekly
water samples from municipal water supplies using Lake Superior
water from Grand Marais, Minnesota to Marquette, Michigan.  These
                                31

-------
CO
1X3
                                £~   '5
                                lio  10
                                Sf   *
                                32    o
                                    E
                                    NE
                                    N
                                    NW
                                •D
                                e   W
                                $   SE
                                    S
                                    sw
   I.   .1  t lUftaJ t  I I*. 
-------
566
                             P. M. Cook
samples were analyzed for amphibole mass concentration at the
National Water Quality Laboratory and amphibole  fiber  concentration
by transmission  electron microscopy at  the  Ontario Research
Foundation in Sheridan Park, Ontario and McCrone Associates in
Chicago, Illinois.   Figure 6 depicts the average x-ray diffraction
and electron microscope measurements for each station.  The agree-
ment between these  two measurements is obviously very  good.  The
pattern of maximum  concentrations at Beaver Bay  and  decreasing
concentrations in a counterclockwise direction around  Western Lake
Superior is consistent with large quantities of  amphibole fiber
discharged at a  point between the Silver Bay and Beaver Bay,
Minnesota water  supply intakes and then transported  towards Duluth
(southwest) by the  predominantly counterclockwise  currents of
Western Lake Superior (5).
    Comparison  of  NWOL  X-Ray Diffraction  Amphibols Analyses to  EPA Region V
       Electron Microscope  Fiber Counts  for  Public  Water  Supply Samples
     Average Concentrations  for  Weekly Samples Taken Aug 22-Nov 28, 1973
oia
• O44
* 040
* 036
If 03*
S 5 02«
1 or O20
^ 0.6
r°"
g OO8
J OO4
0

o





s
LL




S
o _
IB ! ~
B B o o o o 5
• B • 2 £ JL °
"3
1 » I 1 *
1 * ! 1 . « i M I »
1 1 1 1 1 1 1 1 1 1 1
13
"
II
H'
w 8
11,
CS
I'"'
a







'Ui






8
liiiliii
'S
j >. & I S 1 • -
! J i* 1 1! ! if 1
FIGURE 6	COMPARISON OF AMPHIBOLE MASS CONCENTRATION DETERMINED
BY X-RAY DIFFRACTION TO TRANSMISSION ELECTRON MICROSCOPE AMPHIBOLE
FIBER COUNTS FOR  LAKE SUPERIOR WATER INTAKES
                                 33

-------
                            P. M. Cook                           567

                        ACKNOWLEDGEMENTS

     The author wishes to gratefully acknowledge the assistance of
Mr. James Tucker of the National Water Quality Laboratory for
electron microscope examinations of water and tailings samples;
Mr. Robert Fulton and Mr. David Marklund for their excellent work
in preparing many of the samples examined by x-ray diffraction;
and Dr. Billy Fairless of the Environmental Protection Agency,
Region V, Central Regional Laboratory, for providing water intake
fiber counts.


                           REFERENCES

1.   I. J. Selikoff, E. C. Hammond and J. Churg, "Carcinogenicity
     of Amosite Asbestos," Arch. Environ. Health Z5_, 183-186 (1972).

2.   J. G. Eaton and G. E. Likens, "Use of Membrane Filters in
     Gravimetric Analyses of Particulate Matter in Natural Waters,"
     Water Resources Res. _5, 1151-1156 (1969).

3.   A. L. Rickards, "Estimation of Trace Amounts of Chrysotile
     Asbestos by X-Ray Diffraction," Anal. Chem. 44, 1872-1373
     (1972).

4.   J. V. Crable, "Quantitative Determination of Chrysotile,
     Amosite, and Crocidolite by X-Ray Diffraction," Am. Ind. Hyg.
     Assoc. J. rt_, 293-298 (1966).

5.   C. E. Adams, "Summer Circulation in Western Lake Superior,"
     Proc. 13th Conf. on Great Lakes Res., 862-879 (1970).
                                34

-------
^sprinted from
b September 1974, Volume 185, pp. 853-855
   Asbestiform Amphibole Minerals: Detection and Measurement
         of High Concentrations in Municipal Water Supplies
                       Philip M. Cook, Gary E. Glass and James H. Tucker
       Asbestiform Amphibol Minerals:  Detection and Measurement of High Concentrations
       in Municipal Water Supplies by Philip M. Cook, et. al.,  is reprinted from
       Science,  6 September 1974, with permission of the American Association for
       the Advancement of Science, 1515 Mass. Ave., N.W., Washington, D.C.  20005.
               Copyright© 1974 by the American Association for the Advancement of Science

                                         35

-------
Asbestiform Amphibole Minerals: Detection and Measurement
of High Concentrations in Municipal Water Supplies
   Abstract. Asbestiform  amphibole  minerals, which  have been demonstrated to
 he associated  with  human health problems,  have been  detected in substantial
 quantities in municipal water supplies taken from  western  Luke  Sitpeiior water.
 The total concentration of amphibole  minerals in the Dulutlt. Minnesota,  water
 supply, as measured by x-ray diffraction  for  daily samples  of suspended solids,
 averages 0.19 milligram per liter with large fluctuations due to seasonal and dinia-
 tological  effects on  lake  circulation. Election nucioscopic  examination oj these
 water samples  confirms the presence  of asbestiform amplnhole fibers  A tonseiva-
 tive estimate of the fiber count for 1973 Duluth water supply samples is (! to 30}
 X 10'' amphibole  fibers identifiable bv  election difjiaction pel liter oj watei with
 a mass concentration  ol  I to 30 micrognims pel  litci
   The inhalation  of  asbestos fibers has
 long  been recognized as a  serious oc-
 cupational  and  environmental  health
 problem.  Moreover,  excessive rates of
 gastrointestinal  and  peritoneal  cancer
 are  associated  with occupational  ex-
 posure  to asbestos (/).  Recently it has
 been  suggested that the ingestion of as-
 bestiform minerals causes an increased
 incidence  of  gastrointestinal  cancers
 (2).  The presence of asbestiform par-
 ticles in parenteral drugs  (J), beverages
 (4, 5), food (6), and  drinking  water
 (5, 7)  has  been reported, and the mi-
 gration of these fibers  through the rat
 bowel wall has been demonstraled by
several  workers (8). The rapid trans-
port of  large intact starch granules and
other particles throughout  the  human
body  after ingestion  has  also been  re-
ported (9).
  Although natural  sources  of  asbesti-
form  minerals are known !o contribute
to  fiber  concentrations   in  drinking
water, particularly  in areas  of  serpen-
tine rock, industrial discharge and min-
ing operations can  also  produce high
concentrations  of  asbestiform  minerals
in drinking water supplies (7). The con-
tribution to water supplies from asbes-
tos-cement pipe is now being studied by
the Environmental  Protection  Agency.
                                                       ,0-5 P",      *«f
 Fig. 1.  Electron micrographs of amphibole fibers filtered  from Diiluth drinking water:
 (a) fiber approximately 2.2 /^m long and 0.04 ^m wide; (b) fiber approximately 2.9 /^m
 long which  is a bundle of many  individual fibrils.  Amphibole fibers are  present with
 other minerals,  diatoms, and organic detriuis; thus it is difficult  to  identify all the
 amphibo'e fibers
Such contamination is invariably due to
chrysotile asbestos, since approximately
95 percent of the  asbestos fiber used in
North America is  chrysotile (10}. Other
asbestos  minerals, all of which  arc in
the amphibole group of hydrated  sili-
cates, include amosite, crocidolite,  an-
thophyllite,  trcmohte, and actinolite.
   We report  here the discovery of as-
bestiform  amphibole  fibers  in  public
water supplies taken from western Lake
Superior  water. We  have  studied  the
variations  in  the  concentration  of as-
bestiform minerals in  this water  over
the past  year by  x-ray  diffraction  and
electron   microscopic techniques.  The
predominant amphibole  present is cum-
mingtomte-grunerite,   which  is  repre-
sented by the formula (Mg>Fe)7SiNO.j;>-
(OH)j. The asbestiform cummingtonite-
grunerite of commercial importance  is
amosite.   In  addition, smaller amounts
of  tremolite-actinolite  and hornblende
are found in the amphibole fraction of
suspended solids  filtered from  western
Lake Superior water.  The concentration
of  amphibole  (//).  particularly cum-
mingtonite-grunerite,  was  found  to be
below detection  limits (< 0.02 ing/liter)
at Thunder  Bay,  Ontario, and  Grand
Marais,  Minnesota; detectable at Silver
Bay.  Minnesota, high O  0 1  mg/liter)
at  Beaver  Bay.  Two  Harbors,   and
Duluth,  Minnesota,  and  detectable in
Cloquct.  Minnesota   water,  which  is
also used bv Superior. Wisconsin
   Examination of samples of  suspended
solid« from the Duluth water  supply by
transmission  electron  microscope   re-
veals the presence of  diatom fragments,
organic   debris,  quartz particles, some
clay  minerals, and amphibole particles
ranging from blocky cleavage  fragments
to ashestiform  fibers (Fig   1). High-
magnification electron micrographs (Fig.
Ib)  show that  many fibers  consist of
••mailer  fibers, or  fibrils, held together
in  bundles.  The   bundle  nature,  the
lineation observed  owing to the presence
of  fibrils  within   the  fiber,   and  the
ragged ends of the fibers have all been
listed as criteria for  the morphological
identification  of   asbestos   fibers   by
transmission electron  microscopy  (12).
   Although amphibole fibers  as long as
20 iim  have  neen observed  in Duluth
                                                          36

-------
•vater samples, most are less than 5  ^m
long with  many  less than  1  /j.m long.
There  has  been  considerable  debate
(13)  over  the  carcmogenicity  of  in-
haled  asbestos fibers  smaller  than 5
/>m, although occupational and environ-
mental exposure  to  asbestos  which re-
sults in cancer invariably involves more
fibers  smaller than  5  /im than  fibers
larger  than 5 /im. Less  is known of the
significance of fiber length  when  the
fibers  are  ingested.
  Amphibole  fiber  counts  by electron
microscopy (14)  showed  millions  of
amphibole fibers  per liter in samples of
Duluth  water.   The   amphibole-like
fibers  may  be positively  identified by
their  selected-area  electron diffraction
patterns  (SAED). For  reasons  of size,
orientation, or particulate  interference
many  amphibole fibers do  not provide
diagnostic diffraction patterns, and thus
not all (he fibers  present were counted.
The  presence  ot  some  chrysotile fibers
was  also noted.
  A comparison  of the analysis of the
water  samples by x-ray diffraction  and
electron  microscopy permits the estima-
tion  of  fiber  counts for other  Duluth
water  samples  The comparison  rests
on  the  assumption that  the  mass  of
total amphibole present is related to the
number  of amphibole  fibers.  This re-
                                quires a constant particle size  distribu-
                                tion for the samples compared, as was
                                observed for the Duluth water  samples.
                                We estimate a range of (1 to 30) X 106
                                SAED  identified  amphibole fibers per
                                liter  of  water with  a  mass  concentra-
                                tion of 1 to 30 /ig/liter. The concentra-
                                tion  of  fibers  in  the  drinking  water
                                vanes with lake conditions and tends  to
                                decrease with the increasing residence
                                time of  the  water   in  the  distribution
                                system   Occasional peak concentrations
                                (up to   10° fibers  per liter)  can result
                                from the  resuspension  of settled sedi-
                                ment in the water lines.  These amphi-
                                bole asbestiform  fiber counts and  par-
                                ticularly mass concentrations are much
                                higher   than  the values  reported for
                                chrysotile   fiber  contamination  in  22
                                municipal water supplies in the Province
                                of Ontario (7). At Thunder Bay, which,
                                like Duluth, uses unfiltered  Lake Supe-
                                rior  water, 0.8 X 10" chrysotile fibers
                                per liter with a mass  concentration  of
                                0.0002 /ig/liter were found.
                                  The  daily  variations in  the  amphi-
                                bole concentrations  of  Duluth  water
                                supply  samples,  as calculated from the
                                amphibole x-ray diffraction  peaks (11),
                                are depicted  in Fig. 2.  During 1973,
                                the  amphibole  concentrations  varied
                                from 0 03 to  0 80 mg/liter with a mean
                                concentration  of  0.19  mg/liter.  The
S    W
     SE
     S
c    swt
o
S ~  30
                            ,*. .  .  i .  .1*1 li it, tt  I,   iJl   ..*   J,,]
                                                    •   .     .   .     1
   I  . I i  .Ik   i.    i.i                .     .  .            i I      . . ]
bill,        .  i  u  W      til'	i    It,     II I 11.. Ill I .,  l  I i. ]
I !•,.  it .«  n       II    I   k<  I «t I. <  i .   Hull i I  ti   ii   I n hM ill
       I    I         .1.1    I*   i.lllillll  I  h I        lllli     I.
>   I    I   I    I  ill         . I  J II   k   ill    I  U J .  .     ,     .1
  \t II I  I I     I.I      III.   I.I.I    iJ,    J  .1   II III     «
     E  r    i.   ..  § id*! i i it.
     NE      i,      .    i.l.  . I
     N
     NW
- —  3 °t                                        I         I
a E  2 of                         i            ..J         I
§-  I0li,	_,L.-.I, L.  l.a..L....,ull,.  i  .,J  I
          Jan   Feb  March April   May  June   July   Aug   Sept  Oct   Nov   Dec    Jan
                                       1973

 Fig. 2 Results  of analyses  (lanuary  1973  through January  1974)  of  10-liter  Duluth,
 Minnesota, drinking water samples  for amphibole  and  suspended solid concentrations.
 Dailv sampling began on  19  March  1973. The  dashed plot for prior dates indicates
 the penod of less frequent sampling Measinements of the concentrations of suspended
 solids were not made on  samples taken before d Maich  Resultant  wind  direction and
 speed (vMiid scale, 0 to 30 km hour) anci precipitation  data are  the  values  as reported
 by  the U S Department ot Commerce. National  Oceanic and Atmospheric Administra-
 tion,  National Weather Service, for Dnlulh  Inte; national  Airport.  Daily mean  water
 temperatures  aie calculated from houily Duluth water intake tempeiature  data provided
 by  Ihe Duluth Water and Gas Department l.akeuood pumping  station. Climatological
 events which  alleet Ihe amount and mineralogical  nature of  the suspended solids noi-
 inullv pierede the observed change  in \valer cniahtv r»  ) to  2 days
mean percent  (by  weight)  of the sus-
pended  solids  identified  as amphibole
was  23  percent.
  The effect  of Climatological  condi-
tions  on  the amount and mineralogical
nature of the  suspended solids  in  the
Duluth  water  supply is  most evident
when heavy rainfalls are followed  by
an increase  in the amount of suspended
solids resulting from river runoff and
shore erosion. On 24 May, for example.
4.!  cm of  rainfall  was  recorded; this
was  followed  by a  brief  period, begin-
ning on 26  May, characterized by high
concentrations  of  suspended  solids  in
the Duluth  water supply. The lag time
represents the time  needed for the river
runoff to move downshore to the water
intake. These  storm-caused high  con-
centrations of suspended solids have low
percentages  of amphibole;  this  finding
was  expected,  since our  study  shows
that  suspended river sediments  enter-
ing Lake Superior  contain  only  0 to 3
percent amphibole,  mainly  hornblende.
The   prevailing  water  circulation  in
western Lake Superior is known to  be
counterclockwise (/5), consistent  with
the pattern  of progressively  increasing
amphibole concentration which we find
in  lake  water  to  the  northeast  of
Duluth.  The  Duluth water  intake,  lo-
cated at  a depth of 20 m, may  receive
water with increased  amphibole concen-
tration when the surface  water circula-
tion   from   the  northeast  is  promoted
by  extended  periods of  easterly and
northeasterly winds,  as during the pe-
riods of  29  March  to 9 April, 29 April
to 1  May. and  I  to 7 May. These same-
winds may  also cause the resuspension
of recently  settled  amphibole-nch sedi-
ment by  wave action in  the shallow
water  area  around  the  water  intake.
A period characterized  by  very  high
concentrations of suspended solids (ap-
proximately 20 percent amphibole) oc-
curred in December 1973 when strong
easterly winds  resuspended surface sedi-
ments and the river sediment input was
low.  Ice cover, which normally  begins
in  January,  prevents  such  wind-gen-
erated resuspension  of lake sediment.
   Amphibole concentrations  in  Duluth
water diminish  during  the  period  of
increasing   summer  stratification  of
western  Lake  Superior water until fall
overturn  (the  time  period in  Fig.  2
when  -vater temperatures were  greater
than  4°C),  probably because  of  the  de-
creased circulation  of deeper lake water
from  the  northeast.  During  times  of
isothermal conditions without ice cover
this  circulation is more pronounced, and
thus the  peak amphibole  concentrations
                                                           37

-------
occur in spring and late fall.  Changes
of water temperature at the intake dur-
ing the  months of summer  stratification
are  often  wind-related.  Offstore  winds
(westerly or northwesterly)  can  cause
upwelling which brings colder  water  to
the intake such as on 6 and 11 Septem-
ber. Easterly or northeasterly winds dur-
ing  the  months of  stratification push
warm  surface  water into  the Duluth
water intake area, causing higher water
temperatures such as on 24 July.
   A  historical record of  the  types  of
amphibole   minerals   previously   sus-
pended  in  Lake Superior water may  be
derived  from a study  of  the bottom
sediments.  Dell  (16)  reported  horn-
blende  as  the  predominant  amphibole
in the sand fraction  of Lake  Superior
postglacial  sediments with  a  trace  of
tremolite-actinolite also present in some
cases. Our  study (17)  of  the surficial
sediments  of western   Lake  Superior
shows a clear pattern of a  recently de-
posited  sediment  layer rich  in  cum-
mingtonite-grunerite  on  top  of  older
sediment  which  does  not  contain  de-
tectable amounts «  1  percent) of cum-
mingtonite-grunerite. This layer rich  in
cummingtonite-grunerite is  thickest (90
m or  more)  and coarsest in the vicinity
of a large  taconite tailings  discharge  at
Silver Bay,  Minnesota  (18).  It spreads
throughout much of western Lake Su-
perior, becoming thin  and  diluted with
other  sediment  at  Duluth,  which is  at
the western  tip  of the  lake.  Indication
of recent changes in the mineralogy  of
suspended solids in  western Lake  Su-
perior water is  provided by our x-ray
diffraction analysis  of  suspended  sedi-
ment  samples collected  for several pe-
riods  in the  past by the  City of Duluth
water utility. Samples from 1939-1940
and   1949-1950  contain   only  trace
amounts of  amphibole  with  no detect-
able   cummingtonite-grunerite,   but   all
samples  studied  for the period 1964—
1965  contained large  amounts of  am-
phibole  (average,  31   percent  of  the
total inorganic solids), most  of which
was cummingtonite-grunerite.  The geo-
logical and  limnological data  indicate
that the source of this large increase  in
amphibole material is  the taconite tail-
ings  (18) that, since  1956, have  been
discharged into western Lake  Superior
at Silver Bay.
                         PHILIP  M. COOK
                         GARY E. GLASS
                       JAMES H. TUCKER
U.S. Environmental Protection Agency,
National  Water Quality Laboratory,
6201  Congdon Boulevard,
Duluth, Minnesota 55804


            Refmacei ud No«c<

 1. E. E. Keal, Lancet  19fO.II, 1214 (1960); I. I.
   Selikoft, J.  Churg, E. C. Hammond, /. Am.
   Med. Assoe. US, 142 (1964); P. E. Enterline
   and M.  A.  Kendrick, Arch. Environ.  Health
   IS,  181  (1967); I. I.  Selikoff, E. C. Hammond,
   J. Churg, /. Am. Med. Assoc. 204. 106 (1968);
   M.  L.  Newhouse and J. C. Wagner, Br.  J.
   Ind. Med.  M,  302  (1969).
 2. Joint  National  Institute  of  Environmental
   Health   Sciences-Environmental   Protection
   Agency  Conference  on  Biological  Effects  of
   Ingested Asbestos, Durham, North  Carolina,
   18-20 November  1973.
 3. W.  I. Nicholson, C.  J. Maggiore, I. J. Selikoff,
   Science 177, 171  (1972).
 4. C. L. Berry, Nature (Land.) 21», 93 (1968).
 5. H.  M.  Cunningham and  R.  D. Pontefract,
   Ibid. 232, 332 (1971).
 6. R. R. Merliss, Science 173, 1141 (1971);  H.  E.
   Blejer and  R. J. Arlon, /. Occup.  Med. IS,
   92  (1973).
 7. O. Kay, Water Pollut. Control 111, 33 (1973).
 8. G.  E. Westlake,  H. J.  Spjut, M. N. Smith,
   Lab. Invest. 14, 2029 (1965); R. D. Pontefract
   and H. M.  Cunningham, Nature (Land.) 243,
   352 (1973);  H. M.  Cunningham  and R.  D.
   Pontefract, I. Assoc. Off. Anal. Chem. 56, 976
   (1973).
 9. C. Volkheimer and F. H. Schulz, Qual.  Plant.
   Mater. Vtg. 18, 17 (1968); G. Schreiber, Arch.
   Environ. Health W, 39 (1974).
10. S. Speil and J. P. Leineweber, Environ. Res.
   2, 166 (1969).
11. By  x-ray diffraction  analysis of  suspended
   solids filtered from 10-liter water samples, we
   made semiquantitative estimates of the amphi-
   bole concentration with the use of an external
   standard. This technique  has been  used  to
   estimate trace amounts of chrysotile asbestos
   and  amphibole asbestos in dust  samples  on
   membrane filters [A. L. Rickards, Anal. Chem.
   44, 1872 (1972); J. V. Crable, Am. Ind. Hyg.
   Assoc. 1. 27, 293 (1966)] and fulfills the need
   for rapid, standardized estimates of amphibole
   concentration in samples  that are not amenable
   to the  use  of an internal standard.
12. P. Gross, R. T. P.  deTreville, M. N. Halter,
   Arch. Environ. Health 20, 571 (1970).
13. Doubts  about  the  carcinogenicity  of smaCl-
   fibers  stem primarily from  the  absence  of
   tumors in animals subjected to short  fibers.
   A recent  study  [F.  Pott, F.  Huth,  K.  H.
   Friedricks, Zentralbl.  Bakterlol. Parasitenkd.
   Infektlonskr. Hyg. Abt. I Orlg. 15S (5/6), 463
   (1972)],  however,  reported  that  rats  intra-
   peritoneally injected with chrysotile  fibers  in-
   curred about a 40 percent incidence of tumors
   for two different tests with  small fibers  (95
   percent less than 5 tan and 99 percent  less
   than  3  /an).
14. The  fiber counts  were  carried out  by   the
   Ontario Research Foundation (ORF), Sheridan
   Park,  Ontario. A comparison with  literature
   values is  normally not possible since  different
   preparation and counting  methods are often
   used.  Because ORF results have been  reported
   for other water samples (7),  these values can
   be compared  with  those  results. The ORF
   fiber-counting  technique consists  of  filtering
   the water sample with  a 0.1-nm  membrane
   filter,  ashing the filter by maintaining it at
   450°C for 3  hours,  dispersing the  ashed
   sample in 4 ml of distilled  water,  and cen-
   trifuging a 1-ml aliquot onto a carbon-coated
   electron  microscope grid which is examined
   at X  25,000 magnification  on a transmission
   electron microscope (Jeolco model JEM 100U)
   at 80  kv. The  Environmental  Protection Agen-
   cy is  currently developing  a  standard method
   for the counting of asbestos fibers in  environ-
   mental samples.
15. C. E. Adams,  Proc.  13th  Con/. Great Lake*
   Res.  (1970), p. 862.
16. C. I.  Dell, "Late Quaternary Sedimentation
   in Lake Superior," thesis, University  of Michi-
   gan (1971).
17. P. M. Cook  G. E.  Glass, R. W. Anu.cw, in
   preparation; R. W. Andrew and G.  E.  Glass.
   in Proceedings of a Conference on the Matter
   of Pollution  of Lake  Superior (U S. Depart-
   ment  of  Interior-Federal Water Quality  Ad-
   ministration hearings, 2nd session, April 1970)
   (O-401-869, Government Printing Office,  Wash-
   ington, D.C.,  1970), vol.  1,  PP. 226-250.
18. Cummingtonite-grunerite  is found almost  ex-
   clusively  in  metamorphic rocks, usually in
   metamorphosed  iron  formations. The eastern
   Biwabik iron  formation  in northeastern Min-
   nesota has been contact-metamorphosed  by
   the Duluth  gabbro.  B.  M.  French  [M/nn.
   Geol.  Sun. Butt. 45,  1  (1968)1 has described
   in  detail  the  formation  of cummingtonite-
   grunerite  in  the  metamorphosed iron forma-
   tion near  the Duluth  gabbro. This cumming-
   tonite-grunerite in many cases is acicular to
   asbestiform in  habit and varies from  iron-rich
   grunerite  to  magnesium-rich cummingtomte.
   (The  infrared spectrum of a sample  of cum-
   mingtonite-grunerite from  the taconite  iron
   ore body  is identical to  that of amosite  as-
   bestos. The  infrared  interpretation  technique
   described  by  R. G.  Burns  and  R.  G. J.
   Strens [Science 153, 890 (1966)] for  cumming-
   tonite-grunerite  indicates  that both  samples
   have  an  FeAFe + Mg)  atom ratio of 0.76.)
   The taconite  iron ore body  has been  mined,
   and,  after the  extraction of magnetite,  the
   amphibole-rich tailings have been discharged
   since  1956 into western  Lake Superior at
   Silver Bay, Minnesota.
 25 January 1974; revised 31 May 1974
                                                                38

-------
APPENDIX G UNIVERSITY OF MINNESOTA AT DULUTH ELECTRON MICROSCOPE
           ANALYSIS RESULTS OF PILOT WATER TREATMENT UNITS -
           RAW WATER FROM DULUTH LAKEWOOD INTAKE.
f/1 x 10°
Date of
sample
6/13
6/13
6/13
6/13
6/17
6/17
6/17
6/17
6/17
6/24
6/24
6/24
6/24
6/24
6/28
6/28
6/28
6/28
6/28
7/3
7/3
7/3
7/3
7/3
7/19
7/19
7/19
7/19
7/19
7/23
7/23
7/23
7/23
7/23
7/25
7/25
7/25
7/25
7/25
7/30
7/30
7/30
Filter
utilized

MM-1
MM-2
ERD-1

MM-1
MM-2
BIF
ERD-1

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
Run Raw
No. sample
33.25
36
57
1
10.64
39
61
31
4
6.03
44
71
37
22
5.54
48
76
47
28
2.99
52
78
51
32
266.0
67
96
70
40
26.0
70
99
72
41
60.5
72
101
73
43
26.6
74
105
Finished
sample

0.53
0.07
6.93

0.23
0.41
7.59a
7.07a

0.92
0.825
8.17a
1.85a

0.51
0.53b
0.83
0.73

0.91
0.16
1.96
1.22

0.66
0.95
2.00
4.16

0.35
0.23
4.41
2.04

0.46
0.43
0.921
0.964

0.18
0.10
Per Cent
Removal

98
99
79

98
96
—
66

85
86
—
69

91
90
85
87

69
95
34
59

—
—
—
—

99
99
83
92

99
99
99
98

99
99
                                      39

-------
APPENDIX G  (CONTINUED).
Date of
sample
7/30
7/30
7/31
7/31
8/1
8/1
8/1
8/1
8/1
8/6
8/6
8/6
8/6
8/6
8/8
8/8
8/8
8/8
8/8
8/13
8/13
8/13
8/13
8/13
8/15
8/15
8/15
8/15
8/15
8/20
8/20
8/20
8/22
8/21
8/22
8/22
8/23
8/23
8/23
8/28
8/28
8/28
8/30
8/30
8/30
8/30
Filter
utilized
BIF
ERD-2

MM-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-1
MM-2
BIF
ERD-2

MM-2
ERD-2

MM-1
MM-1
MM-1

MM-2
ERD-2

MM-2
BIF

MM-2
BIF
ERD-2
Run
No.
77
45

106

76
107
79
46

78
109
82
48

80
111
84
49

82
113
88
51

84
114
89
55

118
59

86
86
86

119
63

122
103

124
105
68
f/1 x 106
Raw Finished
sample sample
1.01
0.87
106. 6b
0.64b
30.0
0.72
0.14
1.70
0.79
10.2
0.20
0.19
0.42
2.00
15.1
0.45
0.16
1.52
0.34
20.0
0.38
0.37
0.59
0.51
8.98
0.55
0.43
1.27
1.53
25.3
0.15
0.81
13.6
0.31
0.49
0.37
17.0
0.26
1.71
10.4
0.11
1.05
17.8
0.18
0.65
0.29
Per Cent
Removal
96
99

99

98
99
94
97

98
98
96
80

97
99
90
98

98
98
97
97

94
95
86
83

99
97

98
96
97

98
90

99
90

99
96
98
                                     40

-------
APPENDIX G (CONTINUED).

Date of
sample
9/4
9/4
9/4
9/4
9/6
9/6
9/6
9/6
9/9
9/9
9/9
9/10
9/10
9/11
9/11
9/11
9/13
9/13
9/13
9/13
9/16
9/16
9/16
9/16
9/17
9/17
9/19
9/19
9/19
9/19
9/20
9/20

Filter
utilized

MM-2
BIF
ERD-2

MM-2
BIF
ERD-2

BIF
ERD-2

MM-2

MM-2
BIF

MM-2
BIF
ERD-2

MM-2
BIF
ERD-2

MM-2

MM-1
MM-2
BIF

ERD-2
a Value may be in error due
Cloquet
Pipeline water

Run
No.

126
108
70

128
111
72

113
73

131

133
115

137
117
78

138
118
79

139

87
140
120

85
to DE present

f/1
Raw
sample
30.3



13.6



15.4


—

13.0


30.0



20.3



12.8

13.8



19.1

in sample

x 106
Finished
aample

0.33
0.96
0.70

0.17
0.36
0.26

0.33
0.64

0.27

0.30
0.28

0.24
0.47
0.51

0.51
4.89
0.81

0.64

1.46
0.33
0.76

0.45



Per Cent
Removal

99
97
98

99
97
98

98
96

—

98
98

99
98
98

97
76
96

95

89
98
94

98


                                       41

-------
                                                                           266
K3
             1C
             O
                100
                 90
                 80
                 70
             X
             
-------
           UNIVERSITY OF MINNESOTA   School of Medicine
           DULUTH                      ; ?205 East 5th Street
                                       • Duluth, Minnesota 55812
                                                             September 19, 1974
Mr. 0. John Schmidt
Black & Veatch
Consulting Engineers
P. 0. Box No. 8405
Kansas City, Missouri 64114

Dear-Mr. Schmidt:

I am pleased to forward the following description of the methods by which we have
been counting fibers here at the School of Medicine in Duluth.

Water samples were obtained from Black & Veatch Engineers at the Lakewood Pumping
Station for the City of Duluth.  The code information on each bottle was recorded
and the bottles were then assigned random numbers in order to "blind" personnel
involved in the subsequent steps of the analysis.  The code was not broken until
counting of all samples from a given week or weeks was completed.

Two-hundred millileters of each sample were filtered through Nuclcoporc* filters
with pore size of 0.2 microns.  Particulate matter within the sample was collected
on the dull surface side of the membrane filter, and a sediment faintly visible to
the naked eye was observed with the majority of samples filtered.  The filter was
air dried and placed in a sealed plastic container until electron microscope grids
were prepared.

At the time of grid preparation, each filter was cut in half with a fine scissors.
One-half was returned to the plastic container for possible future use.  The other
half was inverted on two randomly placed copper electron microscope grids which
had been placed on six layers of filter paper in a clean glass petri dish.  The
EM grids had been covered with a Formvar plastic film followed by a carbon coating.
A single drop of chloroform was placed on the filter membrane over each grid
position, fixing that portion of the membrane firmly to the surface of the coated
grid.   The filter paper in the petri dish was then saturated with chloroform to
dissolve the remainder of the plastic filter, leaving the filtered particulate
matter adhering to the coated surface of the grid.

Fibers were enumerated using a Philips Series 200 Electron Microscope at 10,000 X
magnification.  Any particle with nearly parallel sides, square ends, and an aspect
ratio of 3 to 1 or greater was counted as a fiber.  Electron diffraction was per-
formed on a sufficient number of particles in each sample to ascertain that fibers
of a given morphologic appearance had a crystalline diffraction pattern typical of
chrysotile or amphibolc asbestos.

                                                               \
                                       43

-------
Mr. 0. John Schmidt
September 19, 1974
Page Two
Both of the grids prepared from a sample were studied.  Particle enumeration continu
until at least 25 fibers had been seen or until at least 20 grid squares along an
equatorial plane had been studied.  The total number of fibers seen in each grid
square studied were recorded and averaged for the two grids from each filter.
Appropriate mathematical factors were applied to convert particles seen into fibers
per liter based upon the volume of water filtered and the cross-section area of the
grid examined.

Agreement between the replicate samples averaged plus or minus 15 per cent and the
analysis of the variance indicated that the correlation between the replicate
samples was significantly greater than would have been expected by chance alone.

                                        Sincerely yours,
                                        Robert E. Carter, M.D.
                                        Dean
REC:mk
                                         44

-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
DEPORT NO. 2.
EPA-670/2-75-050e
TITLE AND SUBTITLE
DIRECT FILTRATION OF LAKE SUPERIOR
WATER FOR ASBESTIFORM FIBER REMOVAL
Appendixes E, F, and G
\UTHOR(S)
Black § Veatch, Consulting Engineers
'ERFORMING ORGANIZATION NAME AND ADDRESS
Black £ Veatch, Consulting Engineers
1500 Meadow Lake Parkway
Kansas City, Missouri 64114
SPONSORING AGENCY NAME AND ADDRESS
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
SUPPLEMENTARY NOTES
This work conducted through interagency agreement betwe
of Engineers, St. Paul District. See also EPA-670/2-75
3. RECIPIENT'S ACCESSIOWNO.
5. REPORT DATE
June 1975; Issuing Date
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1CB047; ROAP 21AQB; Task 024
1 1 . CONTR ACT/SKAN* NO.
DACW 37-74-C-0079
IAG #EPA-IAG-D4-0388
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
en EPA Region V and the Corps
-050a, b, c, d, f, and g.

Pilot plant research conducted  in  1974  at  Duluth,  Minnesota, demonstrated that
asbestiform fiber counts in Lake Superior  water could be effectively reduced by
municipal filtration plants.  During  the study engineering data were also obtained
for making cost estimates for construction and operation of both granular and
diatomaceous earth  (DE) filtration plants  ranging  in size from 0.03 to 30 mgd.
Data provided to the contractor by the  Ontario Research Foundation are presented
in Appendix E.  ORF performed asbestiform  fiber analysis of water samples by the
transmission electron microscope method in this project.  In order to place the
data in better perspective, a description  of the analytical method used by ORF is
reproduced in Appendix E.  In Appendix  F,  the amphibole mass data obtained by the
National Water Quality Laboratory  in  Duluth are presented.  This appendix also
includes information on the analytical  method used at NWQL.  The x-ray diffraction
analysis for amphibole mass provided  confirmation  of electron microscope amphibole
fiber results.  Fiber count data obtained  at the University of Minnesota at Duluth
are tabulated in Appendix G.  A statement  describing the electron microscope
analytical method is also included.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Asbestos
Amphiboles
Serpentine
Water supply
Filtration
Water treatment
Pilot plants
DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b. IDENTIFIERS/OPEN ENDED TERMS
Mixed media filtration
Diatomaceous earth fil-
tration
Asbestiform
Chrysotile
Fiber removal
Duluth (Minnesota)
Lake Superior
19. SECURITY CLASS (This Report)
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
c. COSATI Field/Group
13B
21. NO. OF PAGES
51
22. PRICE
 Form 2220-1 (9-73)
45
                                                 U. S. GOVERNMENT PRINTING OFFICE: 1975-657-59VS39B Region No. 5-II

-------