A PRELIMINARY REPORV
OK ASBJIRTOS
IN TI1L DULUTH, NJNi;bSOTA,ARKA.
OFE1CE OF TECHNICAL ANALYSIS
OFFICE OF ENFORCEMENT AND GENKK/J CO
ENVJUONME^TAL PROTECTION AGL^CY
January, 1974
-------
r
L
AUENCt
-------
II. Air
Water
A, /'•'::; 1 v£c/j_ of _ \".-j-et^.££?lr_?.-'^_ar''> about: 4o% complete. Duplicate
'r1 contractors shows:
1. A definite preserve of asbestiform fibers in
19 saiapiit.! (44%).
2. A possible presence in 7 samples (16%).
3. A definite absence in 17 samples (40%).
B. W? t:or supply__ gy_steian _a£f££t^ed_:
1. Dafinitely: Daluth, Silver Bay, Beaver Bayy and
IVo Harbors, Minnesota, and Superior, Wisconsin.
2. Strong possibility: Grand Marais, Aurora, and
Cloquet) Minnesota, and White Pirre, Michigan.
3. These water supply PJ^Sterns serve not more than
153,232 people.
C• Asbentiform fiber identity and size:
1. Fibers found by both independent EPA contractors
iu all buc tnree samples were identified as being
the asiphibole "grunerite-cummingtonite" similar
to that found on the Mesabi Range.
2, Fibers were found to be predominantly shorter than
5 micrometers.
A. Analyses of air sagpj.es exclusive of stack samplesand
court-ordered samp1es :
1. First sampling effort:
a. Analyses by Battelle 95% completed (106 samples),
but they will not stand behind the sample preparation
technique used ro obtain those results.
b. Reanalysed 33 of those samples which results they
will support.
-------
11
c. Onl-y the remits or L:h2 ro?-.nalysis have
bec-r, cc.-i'-.iuerc--l in this interpretation
of the data, sJ. though there is no
significant Difference betvaon the results
obtcrLr.eo oithc-r tdr.ie.
2. .Second c..u'"p3 ir.g eiiort: Analyses by two other EPA
contractors are Hi complete.
&* Concentrations of Fiberr-, ir.ti;e_^Air_aj^ Silver Bay
The samples usc-d for comparative purposer, have been taken
under different scr.i.piin? conditions and for different programs.
Generally the data ware generated during research programs where
sampling and analytical methods were being tested for validity.
Keeping the limitations of the analysis in mind, the
following conclusions have been determined:
1) Amph.i bole concentrations measured down-
wind of the Reserve Mining processing facilities
are not uniquely high. Chrysotile concentrations
equal to or greater than the Silver Bay amphibole
concentrations have been measured in Ohio and
California near asbestos plants.
2) The average concentration of two back-
ground (or clean site) samples measured, in the
Silver Bay area are ten times greater than the
urban mean chrysotile concentration. The average
amphibole concentration for the background site
in Silver Bay is greater than 98.8% of the 314
chrysotile concentrations measured in 55 U. S.
cities during the 1969-1972 period.
-------
TABLE OF CONTENTS
l^age
Introduction -L
Chapter I
Prelirainary Interpretation Of Analytical Data Reported To
Date On Levels Of Asbentiform Fibers In The Water Supplies
Of The Duluth Area And Western Lake Superior Basin
A. SUMMARY OF FINDINGS AND TENTATIVE CONCLUSIONS
1. Summary of Findings 1
2. Summary of Tentative Conclusions 5
B. INTRODUCTION 10
C. WATER SAMPLING EFFORT - WESTERN LAKE SUPERIOR BASIN 10
D. ANALYTICAL EFFORT 12
E. RESULTS OF ANALYTICAL EFFORT
1. Findings '• 14
2. Tentative Conclusions 23
Chapter II
Preliminary Analysis Of Data On Asbes, i.iforui Fibers In
Air From Silver Bay, Minnesota
A. INTRODUCTION 27
B. SOURCES OF DATA 28
C. DATA FROM THE EPA MEMBRANE FILTER NETWORK 32
D. DATA FROM THE EPA SAMPLING EFFORT IN MINNESOTA 32
E. COMPARISON OF SILVER BAY DOWNWIND DATA TO OTHER
DOWNWIND-OF-SOURCE LOCATION DATA 37
F. COMPARISON OF SILVER BAY BACKGROUND DATA TO URBAN
CONCENTRATIONS 40
G. COMPARISON OF SILVER BAY BACKGROUND DATA TO NON-
URBAN AND REMOTE DATA 42
H. CONCLUSION ON THE SILVER BAY AMPHIBOLE CONCENTRA-
TIONS 42
-------
iv
Chapter III
Preliminary Estimate Of The Occurrence Of Asbestiform And
Other Potentially Hazardous Fibers In The Taconite Industry
A. INTRODUCTION 46
B. THE GENERAL GEOl^OCY AM) MINERALOGY OF THE
BIWABIK IRON FORMATION 47
C. DISTRIBUTION OF TACONITE MIKES ALONG THE
MESABI RANGE 48
D. ASBESTIFORM MINERALS OTHER THAN GRTJNERITE--
CUKMINGTONITE ALONG THE MESABI RANGE 49
E. ASBESTIFORM MINERALS EXPECTED IN EMISSIONS
FROM LOADING PELLETS 50
F. TACONITE MINING IN STATES OTHER THAN
MINNESOTA 51
G. POTENTIAL POLLUTION ASSOCIATED WITH ACT-
IVITIES OTHER THAN THE MINING AND MILLING OF
IRON ORE 51
H. CONCLUSIONS 5rt
APPENDIX I ANALYTICAL METHODOLOGY USED BY EPA
CONTRACTORS 57
*APPENDIX II ANALYTICAL RESULTS REPORTED TO DATE
FOR WATER AND AIR SAMPLES 81
APPENDIX III STATISTICAL ANALYSIS OF THE ANALYTICAL
RESULTS 82
*Due to its bulk, Appendix II can be seen on file in the
Office of Technical Analysis
-------
INTRODUCTION
The data upon which this analysis is based are the analytical
results obtained to date from the contractors and cooperators. The
analysis were about 50% completed as of November, 1973. Table I
shows the EPA contractors as of that date. On that date about 366
samples had been completed, reported, and submitted for clearance
and about 163 remained to be analyzed. This accounting does not include
the stack samples or process samples to be analyzed. Table II shows
the number of samples handled from each locale. Since the results are
only about half completed, conclusions included herein must be
viewed as preliminary, pending completion of all samples and final
detailed analysis and interpretations of the results.
Definition of Asbestos
"Asbestos" is defined by the n S. Geological Survey as a term
applied in commerce to naturally fibrous silicates that are amenable
to mechanical separation into fine filaments of considerable tensile
strength and flexibility.A These fibers have, by virtue of unique
combinations of physical and chemical properties, a great variety of
industrial uses.
/!_ United States Mineral Resdurces, D. A. Brobst and W. P. Pratt,
eds.; U. S. Geological Survey Professional Paper 820, p. 64, 1973.
-------
vi
Classification of Asbest os_
The silicate minerals called asbestos vary greatly in their
physical and chemical properties. The minerals are
classified as follows:^
CHRYSOTILE AMPHIBOLES
l
ACTINOLITE AMOSITE ANTHOPHYLLITE CROCIDOLITE TREMOL1TE
GRUNERITE CUMMINGT
CUMMINGTONITE
7~ Evaluation of Carcinogenic Risk of Chemicals to Man;
Monograph, International Agency for Research on Cancer,
vol. 2, 1973.
-------
ro
W
ra
s
W
>
o
O
O
W
1-3
pa
to
10
co CJ
t-j' M
•^ s
-
l-i T} o
0> }-J ,n '•:
•o o ro t>
i-' i u:
o w a;
4J 3 4J r*
M O S "D
y CJ CU CO <<-(
O JS *- O M
0 • O a f.j -~.
^d M i-i i 3 '••> -D i->
o -cucjcJOr-ji'c;
05 !-f 4-J 4J ;J O p 03
iJ -H fl (3 CT fll _'.» J-:
CO ^ t^ 13- *& P-s ^ *-•> •* '
o c-a o v
O u •.
l-i (" .-
« i ,1
W C.) •* .->
CJ H '0
0 p. , , n
o r; ,-; J
P, I!H a . r-!
x«-
*"O
CJ
co a>
X i
rt r^
Q CO
Di
m o
d
i/^
~v»
o
CN
•» «
W P O
Cn • cy te o
o- ^ ,0 co
o w, S Go
'ft! OS !
~ -x i-j r-.'
P-i O ^ ?
t ,
l-.
":
z -.. r
'• - '
• • , ;
• '
<."• ' , ' - -
r c ^ • '-
CT1
t/'i
0
r" ' s"-'
P. 'O
c/:> cj p
0.' U)
O CS !
t/') ^~
r-i O C,
1 f- p.
O i-i O
CO •-*'
CO
i^r
rH
f>
VO
O 1
rH 0
O
M
r>
'"^j
M
0
i
J.J
b
CJ M
a)
j^ 13
iH ^
<-C-
" 0
tr, rH
CSJ
CQ
cu jr1
rH U
a, rj
fi CU
it!
co o
CM
O rH
r~- -co-
co
r--
r^
rH
V.
crT
O
O
W
o
CO
c
o
•H
c
o
W
^ o
§Sx
00
a »-i o
0 3 O
*-• S t~}
M-l
4J O"
•H O LT,
rH C^
-------
1
i
;.;!
^?
W
- ,J
^ -i
O »
'
I.LJ
0 c--
CO! Sz
1 ''
'Jj t-
i-i <
vj1- !*
5!
^
6
0
coi
p.!
CO
«
col
o
c3
b
eft!
HI
B
s
o
*M
|
t/o
1
M
M
W
H
j-
!
t
O
Q
O
H
-rj G)
QJ CJ
•u ^ d
4J O tfl
•H VM i-4
0 CO
,£1 CU
3 rH
CO CJ
3 T3
3 W
2 4J
4 l-l
3 O
3 P.
5 «
"S
4J
CU
1 1
§4
O
U
O
0)
Prf
*
CO
VJ CD
CO rH
*§ "o i*
,3 Co
iz; co
**"*' ' ---.--.-.. ._. - .. _...•_
/*^*
t-H
CM
=«=
*
X
O
)-l
a.
P. ^
* ^^ * 5 *
003 <* 0 O 0 OOOO. rHOOOOf-OC
j^, »-*.r->^-»»-i mcT\>
os -H eg nj to
pq •< *J }J
O rHOt^OOO^Cy
(0>rl cflCdrHtH-H. iH >
ni u a> W t-4 *•< >-i w *^ rH
rHQ) COKrH0CU CO 0) C^7o
§*§* rH-4-lH3 3 3 ^^^
r^^ F^liSa S0CO O ,^p.aj u'oo
o 4-> >-5 vi • K • iJ 3 iJ iJ ,592 ^J ,, F, r; X ( i
?S! »« mi «iH r* tH O^ Cd =S= PO<5n) *O MCI) C fi CJ c
SJ •?£ S < C > M n C CCM w 3 o-CdrJOSv
r-P H CO •H0oU«C0n J-< tfi 4-) 4-IC_)Oi~lM 0-<
5 H£j rH1 tl'^'WUS-H4J(§2">*">30 .Htscn O- a>j
< 3« « 3 WCOCOlM0rHrHCd rH "? ?- ^ S £" fe ",?, "
OCD f* O»* PJtciC.)CUG)3'H34-1 Cy ^CTOCOOp-iCU
te 0 t< -^ u rH eg :-: M rs P4 ^: P-^ cn o cn -H w «aj K cn o O O P
4°J 1-1 g_
^
H
cn
cn
o
r-l
Sfr
-t
3
rH
-ct
cn
TOTALS
~" l
1
•
« «
10 C !
& O
rH N
fi O
a ,0
05 r* *
/-t 0
•H 03 .
O
oo je
rj
rCJ M
C 3
1
w o l
CU JJ
0 0, 4-1
CO (3 C! 1
r-Vl rJ 0*
03 « co
>
-1 5-1 CU
u o ,n
1 | 0 I
CO vo «
.^ O
CTJ M CO
0 >-.
CO LW rH
0) OJ
CO W 0
>> 5) t\l
HI £>-, vo
id "ca to
0) 3 -H
it «.
4-J M
C3 CJ iH
O 4-» ?
tH 0) O*
rH rH (J
n) f.x)
c/j f^i c^"1
-X *'^
-K -;< -i:
-------
CliAPTER I
PRELIMINARY INTERPRETATION OF
ANALYTICAL DATA REPORTED TO DATE
ON LEVELS OF ASRESTIFORM FIBERS IN THE
WATER SUPPLIES OF THE DULUTH
AREA AND WESTERN LAKE SUPERIOR BASIN
-------
CiiAPTFR I
TASLE OF CONTENTS
A. S\J:;/:A:;Y OF IT-DIMCS AKI; TENTATIVE CONCLUSIONS
»
1. Summary of Fi/idings 1
2.. Summary cf Tentative Conclusions 5
B. INTRODUCTION 10
C. WATER SAMPLING EFFORT - WESTERN LAKE SUPERIOR BASIN 10
D. ANALYTICAL EFFORT 12
E. RESULTS OF ANALYTICAL EFFORT
1. Findings 14
2. Tentative Conclusions 23
-------
A. stii-giAKY OF FILINGS A'.n
-1' S1M1ARY C~ riNDIKC'i
The air iytical re::-.: Its-1 preo'intly in h«md address the extent of
at.bcstifor,:, fiber cone,: •:?>' :i scion of wrier samples collected by EPA in Into
June and early July of Li;7^, as determined by electron microscopy par forwci
by two ET'A-contractor ].-,bo,'.ator ies , the University of California School i.-f
V'ublic Heal-'.h (UCSPII) ;t l^rV'jKry, California, and McCrone Associates,
(KCCA) of Chicago, 111 l^ois. Only those results for which analyses were
coiapieted by both laboratories are reported on here, and any conclusions
drawn at this Lime must reuiain subject to change until all samples
collected at various points and localities are completed.
While the overall June/July EPA water sampling program produced
samples at various points in the water source-supply-distribution chsin
from 30 major locations in the Western Lake. Superior Basin — covering
12 cities and communities in Minnesota, 2 in Wisconsin, A in Michigan,
as well as additional Minnesota locations, including well water and
surface supplies from rivers and lakes — analytical results reported by
both laboratories to date yield only the partial coverage indicated in
Table B, and as summarized below. _!_/
Total # water samples:
(Western Lake Superior Area)
UC School of Public Health ..
McCrone Associates *......».,
(US Army ERDLATOR Samples)
UC School of Public Health .,
McCrone Associates. ,
Total Received/Completed...
Received (
. ... 88
79
.... 11
9
187
Completed to date
45
54
11
9
119
Total * of analyses completed for duplicate
samples taken at the same location and
comprising the basis' for the tentative
conclusions reached in this report:
86 for 43 water
The 86 data points above represent analytical findings reported by
UCSFH and MCCA to date of asbestifomi fiber levels found in 43 water samples.
including 8 ERDL.ATOR treatment unit samples. Neither laboratory had
knowledge of the type or origin of any sample, each sample being marked
solely with an identification number which was decoded by EPA on receipt
of the analytical results.
Excluding results for the experimental runs of the US Array ERDLATOR
Treatment Unit (Diatomaceous Earth Filtration) , findings respecting
asbestiform fiber levels for duplicate samples from identical locations
are now in hand for 35 separate locations in the Western Lake Superior
See also Data Appendix
-------
Table B — Listing of Number and Type of Water Samples Received
and Status of Completion of Electron Microscope Analyses
by Both EPA-Contract or Labs, by Major Sampling Location. 2/
Over-all Sampling Program1
Number £• Type of"Sample Received/Completed
Samples Received and R^S
Reported by Both Lahp *,•
Duplicate Samples Tsk<.\
the Same Locations - to
a,
v.
Lii.
Is),
-0.
21.
24.
28.
29.
30.
» 6 Type of Sample
/•^ cation* Received
(See Legend)
Silver Bay 3-DS/2-EFF/3-IN
jl-iaver Bay 3-DS/2-EFF/2-IN
Duluth 11-DS/10-EFF/3-IN/2-HYD/
2-WW
3-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN/2-HYD
2-DS/5-JN
4-DS/2-EFF/2-IN/2-HYD
2-DS/2-EFF/2-IN/2-Hyd
2-DS/2-EFF/2-IN/2-HYD
2-DS/2-EFF/2-IN/2-HYD
2-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN/2-HYD
2-DS/2-EFF/2-IN/2-HYD
2-WW
Two Harbors
Ashland,Wis.
Cloquet
Superior,Wis
VMts Fine, Mich
Eagte Harbor, "
Grand Marais
Virginia
Eveieth
Aurora
Hoyt Lakes
Minneapolis
Ramsay,Mich.
U&n. P&L Well
Fond du La c Wen 2-WW
Lslosvocd School, Mch, 2-DS
NE Exp.Station 2-DS
Gooseberry
State Park 2-DS
Edibeck Carpgrf. 2-DS
Elde's Cojnar^rrg 2-DS
Beaver River 2-RW
Stewart River 2-RW
Baptism River 2-RW
St.Louis River 2-RW
Wild Rice Lake 2-LW
Pike Lake 2-LW
Birch Lake 2-LW
Subtotal - 30 major locations:
56-DS/38-EFF/37-IN/8-RW
6-LW/16-HYD/6-WW
Add ERDLATOR
Samples: 20
Analyses
Completed
2-DS/1-EFF/2-IN
2-DS/l-EFF/l-IN
8-DS/7-EFF/1-IN
2-WW
2-DS/l-EFF/l-IN
2-DS/l-EFF
Total Total
Received Completed
All completed
2-DS/l-EFF/l-IN
1-DS/l-EFF/l-IN
2-DS/1-EFF/2-IN/1
1-DS/l-EFF
1-DS/l-EFF/l-IN
All completed
All completed
1-DS/l-EFF/l-IN/l
1-DS
1-DS
2-RW
2-RW
2-RW
2-RW
2-LW
1-LW
1-LW
8
7
28
7
8
1
10
8
8
-HYD 8
6
6
6
6
-HYD 8
8
2
2
2
2
2
2
2
2
2
2
2
2
2
2
36-DS/23-EFF/22-IN/8-
RW/4-LW/4-HYB/2-BWivV 167
20
20
Total for all samples 187
4
2
16
2
2
6
10
2
0
4
0
0
6
6
0
0
0
0
0
0
0
0
0
2
2
2
2
2
0
0
86
(Pei
( 50% '
(Ji'-v •
(.IT';
(29-; '
(2;?:.
( 100',' ,
(100% ';
( \ y *' ^
(50%. •
( 0*0
ClOO'o i
(100% j
( 0% )
( 0% i
( 0% }
( 0% '
( 0% )
( 0%; }
( 0% )
( 0% )
(100% )
(100% )
(100? )
(100% )
(100% )
(.0% )
< 0% )
70 ( 42%)
16 (100% )
( 46% )
* = Minnesota locations unless otherwise shown
Legend: DS = Distribution System/EFF = Effluent/ IN = Influent/ RW
LW = local lake water/ HYD = Fire hydrant/ WW = well water
2/ See also Data Appendix
= River water
-------
The tabulation of fiber counts and size distributions summarized below
is restricted to locations from which duplicate samples were received and reporieo
on by both EPA-contractor laboratories to date. Results of samples for which .inaJx
of the duplicate remains outstanding will be reported after they are received ar.u
evaluated*,
Table C --
Asbestiform Fibers and Fiber Size Distributions Found in
Duplicate Samples via Electron Microscopy Analysis performed
by University of California School of Public Health and
McCrone Associates, Inc., of Chicago, 111. o/
I feation*
I,Silver Bay
"-' ^eever Bay
'..Duiuth
:'.Two Harbors
- Silver Bay
' .Superior,Wis.
J.Jeno's *1 (Duluth)
] . " £2
:1. " #3
^.Superior,Wis.
M.Cloquet
1 ri "
± -j«
• it "
.T.Superior, Wis.
'B.White Pine, Mich.
"'.Grand Marais
.jO. Aurora
M.Hallett We 11, Duhith
i.2,Ashland,Wis.
'3.Superior, Wis.
'>4 " "
U' -* •*
l:.r>.Grand Marais
26.Aurora
27. "
'.''.Hoyt Lakes
JV. "
•/) " "
;i .Beaver River
•v Stewart River
• ''.Baptism River
31.St. Louis River
Vi.wild Rice Lake
Type
Sample
DS
DS
DS1
DS2
DS3
DS4
DS
IN
DS
EFF
EFF
EFF
EFF
IN
DS
IN
DS
DS
DS
IN
ww
DS
IN
HYD
IN
EFF
DS
IN
EFF
DS
RW
RW
RW
RW
Number of Fibers
UCSPH MCCA
(Millions of Fibers
per liter water)
2.0
3.0
5.0
2.0
1.0
1.0
2.0'
.5
.3
.7
1.0
.1
.7
BLACL**
•II
II.
It
11
11
II
11
4.37
5.29
2.76
6.08
2.26
1.47
2.53
1.47
1.10
3.87
.6V
.71
.37
.39***
.93***
.98***
.39
.18
.25
.26
BLACL**
it
it
Fiber Length - Percent Distribiitj
UCSPH MCCA ^
<3 3-5 >5 <3 3-5 >5
[Micrometer size classes)
«
3.2 -
2.6 -
1.5
10.2 6.1
9.1 3.6
7.1 -
4.8 -
6.7 6.7
9.1
LW
Continued next page with US Army ERDL.ATOR results)
-------
(Tabulation continued)
JLS ARMY ERDLATOR RESULTS
1.
ri
5!
4.
j.
t).
•7
*.'' .
Run z$
Run *9
Run #9
Run *7
Run »7
Run =<>
Run »8
Run #8
IN
EFF
IN
EFF
EFF
EFF
EFF
EFF
Number of Fibers
UCSPH MCCA
(Millions
.3
,2
.0
.3
.1
.3
BLACL**
BLACL**
per Liter)
2.04
6.05
8.24
.39
.99
.46
BLACL**
BLACL**
Fiber Length
UCSPH
3
50.0
100.0
54.8
83.3
100.0
100.0
-
-
3-5
25.0
-
2.3
15.0
-
-
-
-
- Percent Distribur.i
MCCA
5 3 3-5
25.0 96.2 3.8
100.0
2.3 95.3 2.0 1
1.7 100.0
- 100.0
- 100.0
_
. -
* Minnesota location unless otherwise indicated
** BLACL means "Below lowest acceptable counting limit"; this is a variable ericounte-x
from sample to sample, is dependent on other suspended particulates content of £;>' ,
fie samples, and may comprise anywhere from "more than zero" to "less than 50,001)'
fibers.
***• Chrysotile asbestos fibers were found in these samples; no amphibole fibers.
3/ See also Data Appendix
A detailed discussion of findings in hand for individual sampling
locations is presented'in E. RESULTS OF ANALYTICAL EFFORT - Item 1. Findings,
The tentative conclusions arising from the findings are summarized below, and
are addressed in additional detail in E.2. Tenttrtr^et€3i?a£sr$35fias.
A map showing the approximate locations of the water sampling points
for which both labs reported analytical results to date appears on the fol-
lowing page, with a legend categorizing the findings thus far reported.
-------
<->
o
3C
>— <
t/>
O
g
rri
V — '
E'-%.°
>"c * ...
*"*" \jy P *
LO o ^ v y
= St-srVjL
rX.. 2 E. 2. £ ^ V ~
Ia>*-";» -s c-f ^ C- --. , —
Ifcasr _j. rr o o co ^
^ o • ./. =° =o- ft ~-
-i -',^-N.- • M- T-t-'O v 1*1
. •££% s 5 ss f \ * §
. 4t> \ g^ VY> 5
\ /^ \ 2 $3=* v
•\ / V . ^Jsp r.; co
•'\ Jj S- 33 < _,.
A V - V . >-'•« o — «
- \ x. « S <-
• V XfrocD on
, V T " " ^
* \ \ fc-V-<< co m
rf O
SJL
X
o"
v<
" ^
oT o~
Or-T
*
to \-
^ . 7s / -s . \- -
•CV- . \AV < V-
• ;*-*•>. - . •!«>* (i> v
•v ^«s ... \ \j
•\ *— "X:.*. . -J» -^ \
*
.
... _ . -
L ?. •;• '•
^;>.^Mvl -«\i>!,-.v
.- / "H- "••%;? •'
/ * ol ?'. • \=?
s
-
*
N.\ *
/ \ ' A
r • A
.**!-" * • I
"V"
- rX,- • . .-••••• • ^ •
„*. ' » *. * * -• . • ."
o^s'-g/ - 1
•^^ 3 M»
fD r+ V
" -3- ^ -V "
±? ~~ xT^V 0 . O St CQ
*-* -^ x ~^>. t-< °i o. >-< n
. cr> • f ."'v • c ". i-'- oj t— •
-c" ? 4 corfco3CD
^z- • ». A M. S o o
iQ A o c M- c+
O *"^x^ 3 3 co co H
rs ~i "*- en co co o
'/. . ' • r+ M- 0 3
-J C O 3
i-t 3 i— ' s:
. • CD *< M.
• " )-*' O
O 3 i^v »-$
Ms O O
f* ^5 CO
. . *" r+ CD O
v^r^r g"Sg5
-^{v ^ 3 «<
. .\ M> Q} "O
• ^ 3 O. CD 3
CL t~{ «rf QJ
3 co v:
tQ CO f+ r*
CO M- O M-
3 O
CO tQ CL CO
3 O) »-•
D. <-*<-+
3- CD
CD -•
V.
^ V "
V ;• . • " •
\ /
\ ../.*.
O
* A ^V >E
\7 ^U^ "
IF \£8T
TF
11 II U
> O "T3
t-5o»CO3 1]CD "fl "^ "•]
M- r* CO 0> M- h+) l-J- O I-1'
g i-.. 3 H- cr M- cr tr cr
^ O *X3 **Z CD 3 CD O) CD
3 i— • «n- Hjl-"- >-(Cr H
CD M- en e-f co i— ' co
33COO -CD -CD
CD o: cr cr cr
gt_5l_i OJJ> OJIT3 M
03 0) co O* CO >-j co
M«?T33 CDco CDCD CD
3CDCD CLCD CXco C.
3 to 3 CD
O C O O O 3 O
C 0) H" 3CD 3O 3
CO <•+ r* CD
,r+ CO H( O M M
03 r+ CD >-»> CD O CD
3 !3" Hf, CO CO M> CO
D. CD O C > C C
M- >-J 1— " CO 1— ' i» H-1
3 co <-•• cr <-*• to r+
tQ CD 3> COCD COD* CO
3CL CO CD
03CDCL r+<-* r*CO r+
r-f M. O *->• O r* O
pi r+ M> M.
rtQI-" QO D.W> D.
3-OO O>H( 03O 03
(-••|3 r-f 3 r+ H< r+
CO 03 CD CD 3 CD
0
CJ
•^
CD C t-H R
Q ^0 CO Z
CD KI ra >-€
3 33 33 3=.
Q. H CO >~3
O l~i
12 Pj
?^V S S *^
fe/ CO > H
V M H 2
s: ra c
il 33 t-H
O 33 CO Ci
CD m G co
M- - ~c 33
3 r1 m
M- 03 l-l CO
r+ K; m ^3
CD co m
o o
13 no1-)
H O "T M
CD O 2!
co 33 S C73
CD > PJ
3 T3 CO 3>
O - cc H CO
CD H-l m CO
O 33 R
O > Z CO
H> p H
r~ •— <
3> i" ^ 1* T]
co o P^: o
cr o m 33
CD 3> SI
to H
c-+ t-H
I-" O
0
H(
3
-------
Area, comprising 70 analyses completed as compared to 167 samples
received, a completion percentage 01 about 42 percent. If ERDLATOR
results are included, the completion rstio is 46 percent.
The individual water source/supply areas sampled for which results
have been reported by both laboratories are identified in the tabulation
below, along with an accounting of results still outstanding as well as
indicated completion ratios.
2. SUMMARY OF TENTATIVE CONCLUSIONS
In view of the fact that analytical results respecting the presence
or absence of asbestiform fibers, as presently in hand, comprise only
about 46 percent of the total data base, all conclusions drawn below
must be viewed as tentative. Aside from considerations of statistical
validity, it is entirely possible that for given locations for which the
partial results presently available show an absence of fibers may, on
examination of samples for which results remain outstanding, show their
presence. Notwithstanding these obvious shortcomings, it is patently
desirable — and reasonable — to advance tentative conclusions respecting
levels of water contamination by asbestiform fibers in the Western Lake
Superior Basin area.
1) Results reported for 43 duplicate water samples analyzed by
electron microscopy by two independent EPA contractor labs
show a definite presence of asbestiform fibers in 19 samples
(44%), a probablp p^qence in 7 samples (1G%), and a definite
absence in the remaining 17 samples (40%). 4/
a) The definitional framework relied upon in framing these
tentative conclusions is as follows:
(1) All conclusions relating to the presence or absence
of asbestiform fibers refer to water sampled in late
June - early July of 1973, and thus reflect levels
then prevailing.
(2) A "definite" presence of asbestiform fibers is assumed
where both EPA contractor labs reported finding asbesti-
form fibers greater than the "Lowest Acceptable Counting
Limit" for any duplicate samples which they both examined.
(See (3) below)
(3) A "definite" absence of asbestiform fibers is assumed
where both EPA contractor labs reported finding no
asbestiform fibers in any duplicate samples which they
both examined. This is identified as "BLACL", meaning
"Below Lowest Acceptable Counting Limit" — a variable
encountered from sample-to-sample which is dependent
on other suspended particulates content of specific
samples, and may comprise anywhere from "more than zero"
to "less than 50,000 fibers".
kj Derived from Data Appendix Tables III - V.
-------
A "probable" presence of asbestif ora fibers is assumed
where one T,:'A contractor lab reported finding no asbesti-
form fibers (BLA.CL.) , whereas the other reported fiber
levels above "ELAGL". This divergence is encountered in
a total of 7 out of the A3 duplicate samples examined
by both labi:, and its significance is discussed later.
The asbestiform fibers found by both EPA contractor labs for all
but three cluplicat i__s rtagles we re identified as amphibole ashes tojs
fibers i n_ t lie "grin i e r ite_ r_£"'-Ti;m A^L? tonite" series of fibrous mine ral
_ _ _
of the type found in various Geposlts in the Iron Range. In three
samples, all from the Cloquet area, one lab found not amphibole but
chrysotile asbestos fibers, a serpentine asbestos such as is used
in most asbestos-containing manufactured products; the other lab
saw no asbestif orm fibers in these three samples.
3) In the absence of _ medical studies showing that asbestif orm fj.ber
ingestion is beneficial to human health, and in the presence of
a variety of scientific studies affirming that asbestif orm fiber
inhalation is hazar doua to human health, the conclusion dictated
by prudence is that the absence of such fibers from drinking water
supplies is clear ly preferable to their presence.
4) This conclusion forms the basis for arranging the analytical
results respecting gsbestif orm fiber content for varjLous^localities
for which duplicate samples have been reported on by both EPA
contractor labs in a fashion ranging from "Contaminated" to
"Not Contaminated^. Table D shows these sampling pointy without
intermediate ranking reflecting degrees of severity of contamination.
5) Pending development of a consensus by medical authorities respecting
"harmful" versus "acceptable" levels of asbestiform fibers in water
destined for ingestion, it makes little sense to attempt to rank
analytical results for the several sampling points by "intermediate"
degrees of severity of contamination by contrasting, say, 3 million
fibers per liter versus 1 million fibers per liter. Whatever
degree of precision would obtain from such an attempt could be at
least partially obliterated by the lack of precision in presently
available electron microscopy techniques for determining numbers
and sizes of asbestiform fibers in water.
6) The only certainty which obtains at this point in time is the
certainty of the electron microscopist who sees an asbestiform
fiber under his microscope and identifies it as such beyond the
shadow of doubt, using his equipment to its full capability,
and drawing on his knowledge and experience with full confidence.
Beyond that certainty, we find ourselves at the "frontier" of the
state-of-the-art, and attempts to impart precision beyond that
which is deserved are not warranted at this point in time.
-------
7) Thus, while for 7 of the 43 samples reported on here, no asbesti-
form fibers were i ouiid by one lab., V7hereas the other did see and
identify ssbestlforni fibers., cannot be dismissed or taken lightly.
Although the corroboracive validity present for the other 36
samples is here lack!mr, the presumption should be that asbesti-
fonn f 1 hers were J-na.vci present in the water at the time th
-------
Table D — List of Water Sampling Voints for which Electron Microscopy
Analytical Results Presently in Hand for Duplicate Samples
Examined by Two EPA Contractor Labs Show the Presence or
Absence of Asbestiform Fibers 5_/
a) Contaminated — Definite Presence of Asbestos Fibers
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Duluth
ii
ii
it
ii
ii
ii
it
ti
ii
ti
ii
tt
Silver
ii
Beaver
Area —
ii __
ii __
ii __
ii
ii
it
ii __
ii __
ii
ti __
ii __
ii __
Bay
ii
Bay
Two Harbors —
Superior ,Wis . —
Arlington Water Pumping Station; tap off west pump
Fire Station at 3501 Woodlawn Ave.; tap in fire
truck bay
Duluth Airport; custodian closet tap, main floor
Duluth Pump. Station, 4424 Grand Ave.; north pump tap
Jeno's Diatomaceous Filter #1
ii it n j»2
It II It JiO
Duluth Raw Lake Water before ERDLATOR Run #8
11 " " before " Run #9
11 " " after " Run #9
11 " " after " Run //7
" " " after " Run #7
" " " after " Run //9
Standard Gas Station, Corner of Highway 6; and
Outer Drive; outside tap eastside of building
Pump House influent; tap off incoming line
Jim's Shell Station, Hwy. 61; car bay tap
Fire Station sink tap, 7th St. and 2d Ave.
Underbill Service Center, 2915 Hill Ave., tap in
meter shop, Superior Power & Light Co.
19. " " — Water Plant effluent; tap in laboratory sink
b) Probably Contaminated — Probable Presence of Asbestiform Fibers
1. Superior, Wis. — Lake Superior water tap outside water treatment
plant Superior Power & Light Co. (from Cloquet Pipeline)
2. White Pine, — Town site heating plant; boiler room tap.
Mich.
3. Grand Marais, — Municipal building and fire station tap in
Minn. fire truck bay
4. Aurora, Minn. — Water Plant intake
*5. Cloquet.Mirin. — Raw well water; intake tap inside building in
pumphouse
*6. " " — Influent - Northwest Paper Co. Lake Superior
influent tap in plant off pipeline
*7. " " — Well water - fire station Cloquet Ave. and 5th St.
(sometimes uses Lake Superior water from pipeline)
* Fibers found in these samples were chrysotile, not amphibole,
raising the possibility that contamination occurs via an
asbestos-containing manufactured product, such as asbestos-
cement pipe, or from some other source.
-------
Table D — Continued
c) Not Contaminated — Definite Absence, of Asbestiform Fibers
1.
2.
3.
4.
Duluth Area
Ashland,Wis. —
5. Superior,Wis.—
611 it
. —
7. Grand Marais —
8. Aurora —
9. "
10. Hoyt Lakes
11. "
12.
13.
14.
Beaver River —
Stewart River—
15. Baptism River—
16. St.Louis Rvr.—
17. Wild Rice Lake-
Hallet-.i. Well (Jeno's) 59th Ave.W. Tap off pump
DuJuth Raw Lake Water, after ERDLATOR Run #8
" " after ERDLATOR Run //8
Water utility office; tap in basement at 320 E.2d St.
Superior Power & Light Co. well; intake for water
plant; tap in laboratory sink.
Superior Power & Light Co. (well); fire hydrant,
underbill service center near building 2900 Kill Ave.
Water Plant Intake; tap off spigot on intake pump
Water Plant effluent; outside tap
City Hall; utility room tap
Water Plant; intake tap in main building
Water Plant; effluent tap after punip, main building
basement
Police Department; garage tap
Stream sample off Highway 61 near Beaver Bay
Kwy. 2, N. of Two Harbors, downstream side of bridge
Eckbeck State Park near Finland, Minn.
Hwy. 2 bridge, 200 yards downstream of bridge
Dam Rd. off Minn. Route #4; sampled at the dam
J>/ Derived from Data Appendix Tables I, III, IV, and V.
-------
10
B. INTRODUCTION
On June 15, 1973, a news release issued by the Environmental
Protection Agency in Washington, D.C. announced that high concentrations
of asbestos fibers had been discovered in the drinking water supply of
Duluth and several communities or. the Minnesota shore of Lake Superior. 6_/
The source of these fibers was believed to be the discharge of taconite
tailings by the Reserve Mining Company at Silver Bay, Minnesota. The
Reserve Mining Company, an equally-owned subsidiary of ARMCO and REPUBLIC
Steel Companies, mines low-grade iron ore (taconite) at Babbitt, Minnesota,
and converts it into high grade pellets at Silver Bay, Minnesota. In
operation since 1956, the Company dumps waste ore tailings, exceeding
60,000 tons per day at full production, into Lake Superior at Silver Bay.
The abatement of this pollutiong discharge is the subject of a civil action
filed by the U.S. Department of Justice in February 1972, at the request
of the Environmental Protection Agency, in the U.S. District Court for the
District of Minnesota, naming Reserve Mining Company as the defendant.
Trial commenced August 1, 1973, and is still in progress at this time.
C. WATER SAMPLING EFFORT — WESTERN LAKE SUPERIOR BASIN
Immediately after issuing the news release, the Environmental
Protection Agency began efforts to secure samples of water thought to be
contaminated with asbestos fibers. A field operations center was established
in Duluth, and personnel of EPA Region V proceeded to collect water samples
at more than 30 sites in the Western Lake Superior Basin Area, covering
various representative points in the water source-supply~distri'.»utioa
setting of 12 cities and communities in Minnesota, 2 in Wisconsin, 4 in
Michigan, as well as 14 additional Minnesota locations comprising well
water and surface supplies from rivers and lakes. Tj
Concurrently, the Agency commenced a search across the United States
to locate analytical talent and facilities capable of processing large
numbers of water samples to yield an early determination of the nature
and extent of the asbestiform fiber contamination problem. The search
soon resulted in the painful realization that meaningful answers to the
problem would be long in the coming. Few laboratories had ever tried to
determine asbestos content of water; even fewer investigators had ever
looked for asbestiform fibers with an electron microscope. Some who had
done the latter had familiarity solely with techniques designed to deter-
mine asbestos dust levels in occupationally-exposed segments of the population.
By mid-July of 1973, it was apparent that the Agency found itself at the
frontier of knowledge and that no quick answers were possible or even
within sight.
fj "Asbestos Fibers Found in Duluth Water Supply5', U.S. Environmental
Protection Agency News Release, dated June 15, 1973
Tj See Data Appendix Table I
-------
11
By early July, more than 700 water samples had been collected and
were awaiting analysis. The exact locations are identified in Table I
of the Appendix, together with collection dates, sample sizes, container
types, preservatives used, and laboratories receiving the samples for
analysis. For convenient overview, the sampling locations are summarized
in alphabetical order below, together with an indication of the type of
sample taken, coded as follows: DS = Distribution system sample; IN =
Influent water; EFF = Effluent water (such as from a water treatment plant)
HYD = Fire Hydrant Sample; RW = River water; LW = Lake water from local
lakes; WW = Well water.
Table A
Alphabetical Summary of Water Sampling
Locations and Sample Type Categories ^/
Ashland, Wis IN/EFF/DS/HYD
Aurora, Minn IN/EFF/DS/HYD
Baptism River, Minn RW
Beaver Bay, Minn IN/EFF/DS/HYD
Beaver River, Minn RW
Birch Lake Reservoir, Minn LW
Cloquet, Minn IN/DS/HYD
Duluth, Minn IN/EFF/DS/HYD
Eagle Harbor, Mich IN/EFF/DS/HYD
Eckbeck State Park, Minn DS
Elde's Corner Spring, Minn DS
Eveleth, Minn IN/EFF/DS/HYD
Fond du Lac Well, Minn. DS (WW)
Gooseberry State Park, Minn. ... DS
Grand Marais, Minn IN/EFF/DS/HYD
Hallett Well, Minn. DS (WW)
Hoyt Lakes, Minn IN/EFF/DS/HYD
Jeno's Filter (Duluth), Minn. .. IN/EFF
Lakewood School, Minn DS (WW)
Minneapolis, Minn IN/EFF/DS/HYD
Minn. Power & Light, Minn DS
NE Exp. Station, Minn DS
Pike Lake, Minn LW
Ramsay, Mich IN/EFF/DS/HYD
St. Louis River, Minn RW
Stewart River, Minn RW
Silver Bay, Minn IN/EFF/DS/HYD
Superior, Wis IN/EFF/DS/HYD
Two Harbors, Minn IN/EFF/DS/HYD
U.S. Army ERDLATOR Unit IN/EFF
Virginia, Minn. IN/EFF/DS/HYD
White Pine, Mich IN/EFF/DS/HYD
Wild Rice Lake, Minn LW
JB/ Derived from Data Appendix Table I
-------
12
D. ANALYTICAL EFFORT
By raid-July 1973, EPA had arranged to secure.the services of three
laboratories thought capable of snaking determinations respecting the
nature and level of asbestifonn fiber contamination in the water supplies
of the Western Lake Superior Basin area. Two of these laboratories,
WALTER C. MCCRONE ASSOCIATES, INC. of Chicago, Illinois, and the SCHOOL
OF PUBLIC HEALTH OF THE UNIVERSITY OF CALIFORNIA at Berkeley had had
prior experience in identifying asbestos in water. The third lab,
BATTELLE COLUMBUS MEMORIAL INSTITUTE at Columbus, Ohio, had had prior
experience, and a continuing contract with EPA, respecting the determination
of asbestos levels in air samples collected by the Agency under a continuing
national air sampling network program effort. All three laboratories were
deemed to have the analytical equipment necessary to make valid determinations
in the problem at hand.
Of the 769 water and filter samples collected by early July, some
268 were filter samples; the remaining 501 were water samples suitably pre-
served. 9/ About 300 of these actual water samples, comprising duplicates
of the original samples, were sent to the three EPA contractor labs named
above. Another 93 water and filter samples were sent to the EPA Water
Supply Research Laboratory at Cincinnati, Ohio, for analysis by optical
microscope, and 33 samples were sent to the EPA National Water Quality
Laboratory at Duluth, Minnesota, for analysis by that facility's
investigators. About 168 water samples remain in the custody of EPA
Region V at its Minnesota-Wisconsin District Office. Additional details
of sample disposition are contained in Appendix Tables I and II. Appendix
Table II is an alphabetical listing of samples remaining in the custody
of Region V.
From the beginning of the search for analytical capability, the Agency
realized that it was crucial to pinpoint the identity, dimensions and levels
of asbestiform fibers as these were present in the water actually being
ingested by the people in the affected communities, with the greatest
possible, precision. This approach alone would secure the factual basis
essential to informed decision-making by medical authorities respecting
the significance of such contaminants to human health, when ingested.
There were severe doubts that optical microscopy had either sufficient
power of resolution and definition or the essential versatility of electron
microscopy, particularly in the sub-micrometer size ranges of particles.
These doubts were proven justified via a comparison of results obtained
by optical microscopy versus those obtained by electron microscopy; there
was a statistically significant lack of relationship between the two sets
of results, obtained by examination of duplicates of the same water samples.
With electron microscopy, too, there were dangers that certain methods
of sample preparation could yield inaccurate information on numbers of
asbestiform fibers and their size distributions as actually present in the
water ingested, because the preparation techniques had the potential of
breaking fibers into two or more pieces and of separating fiber bundles
into individual fibrils, thus artificially multiplying the number of fibers
actually present.
9/ From Data Appendix Table 1
-------
13
Time — rather than accuracy — thus became the tradeoff. One of the
three EPA contractor labs, RATTELLE MEMORIAL INSTITUTE of Columbus, Ohio,
and largely completed its vmter analyses by the end of August 1973;
however, its sample preparation procedures essentially were those followed
to date for determining asbestos dust levels in air, using scnication
techniques (shaking by sound waves) to prepare samples for examination
under the electron microscope. The number of fibers subsequently observed
under the electron microscope way or may not have resembled the number
actually present in the water ingested; most likely, they did not.
BATTELLE reported no fiber size distributions, i.e. reported no data
on the length and width of the fibers found, but instead constructed asbesti-
fonn fiber mass concentrations, expressed-in nanograms per liter of water,
and reported these together with total numbers of fibers calculated to be
present in each sample.
When subsequent analyses were reported on by MCCRONE ASSOCIATES and
the UNIVERSITY OF CALIFORNIA SCHOOL OF PUBLIC HEALTH, using essentially
identical techniques which differed from BATTELLE'S, and it became apparent
that the fibers found in the water samples analyzed were predominantly in
the shorter length category (i.e. 0 to 5 micrometers), mass concentration as
a means of expressing levels of asbestos contamination was largely abandoned
in favor of total fiber counts and individual size distributions. The
rationale was that mass concentration was a potentially misleading indicator
of asbestiform fiber contamination, since a single large fiber present in the
water destined for ingestion could, if broken up into multiple shorter fibers
by sample preparation techniques, be too easily misconstrued as constituting
the presence of a great number of smaller fibers. Therefore, the work done.
by BATTELLE has not been continued, and the results as reported have not been
used in defining the extent 'of asbestiform fiber contamination of water
supplies in the Western Lake Superior Basin area.
*
The discussion of analytical results which follows below relies
exclusively on the analytical findings reported to date by MCCRONE
ASSOCIATES, INC. and the UNIVERSITY OF CALIFORNIA SCHOOL OF PUBLIC HEALTH.
Neither lab had knowledge of the type or origin of the samples examined,
each sample having been marked solely with an identification number which
was decoded by EPA on receipt of the contractors' analytical findings.
The results discussed below represent approximately 46 percent of the total
analytical program contracted for, and the principal investigator of each
of the two labs has testified on the results of his findings as a Federal
witness in the litigation ongoing in Federal District Court.
Owing to the fact that each of the two EPA contractor lab s above is
currently engaged, at the direction of the Federal Court conducting the
Reserve Mining trial, in asbestiform fiber analyses of Court-collected
air and water samples, the timing for completion of analyses for the
original EPA-collected Western Lake Superior Basin area water samples is
uncertain and may well extend into the early part of 1974. Until these
outstanding results are secured, the coverage is incomplete and will
remain so.
-------
14
E. RESULTS OF ANALYTICAL EFFORT
1. FINDINGS
The analytical results presently in hand address the extent of
asbestifonn fiber contamination of water samples collected in late June
and early July of 1973. Appendix Table 1 sets forth the exact dates and
locations for the individual water samples. Analyses were performed by
electron microscopy by two EPA contractor laboratories, the UNIVERSITY OF
CALIFORNIA SCHOOL OF PUBLIC HEALTH at Berkeley (UCSPH) and WALTER C.
MCCRONE ASSOCIATES, INC. of Chicago, Illinois (MCCA). Only those water
samples for which analyses were fully completed by both labs are reported
on here — a completion percentage of roughly 46 percent, as explained
below.
The overall June/July 1973 EPA water sampling program, as pointed
out previously, included water samples from more than 30 distinct locations
in the Western Lake Superior Basin, drawn from various locations in the
water source-supply-distribution chain. The analytical results thus far
fully reported by both contractor labs yield only the partial coverage
summarized below, and shown in detail in Table B. 10/
EPA Contractor Total Number of Water Samples
and Water Sample Category Received Completed to Date
Western Lake Superior,excl. US
Army ERDLATOR Treatment Unit:
UC School of Public Health 88 45
McCrone Associates...; 79 54
US Army ERDLATOR Treatment Unit
Runs (Diatomaceous Earth) :
UC School of Public Health 11 11
McCrone Associates 9 9
Total Received/Completed 187 119
Total number of analyses completed for duplicate
samples taken at the same location and comprising
the data base of the findings and tentative
conclusions presented in this report 86 for 43 water
samples
107 See Data Appendix Tables I, II, IV, and V
-------
15
The 86 data points above represent the sum total of analytical
results reported to date by UCSTli and MCCA of asbestitorm fiber levels
and size distributions found in -'i'? water samples, including 8 U.S. Army
ERDLATOR Treatment Unit samples designed to t^st that unit's fiber removal
efficiency when operating on rrv Lake Superior water sampled near EPA's
National Water Quality Laboratory at Duluth.
The individual water source-supply-distribution areas and points
sampled in the overall June/July 1973 progran are shown in Table B,
together with an accounting of the status of completion of analyses for
each of 30 major locatiorvs. While 119 analytical results have been reported
to date, out of a total of 187 samples received, only 86 of the former
relate to sampling points from which duplicate samples were taken and for
which the analytical results have been fully reported by both labs to
date — a net completion status of approximately 46 percent, ll/
Table C gives a detailed accounting of the analytical findings,
reported by each of 35 geographical locations, plus 8 U.S. Army ERDLATOR
Treatment Unit samples, for which duplicate samples were examined by
each of the two labs. The dimensions addressed are numbers of asbestiform
fibers per liter of water and percent distributions of fiber lengths for
three classes; 1) less than 3 micrometers: 2) 3 to 5 micrometers: and
3) longer than 5 micrometers. In the absence of agreement by medical
authorities respecting the human health significance of asbestiform fibers
of various lengths when ingested, the grouping of these data into the
three length classes indicated must be viewed as arbitrary, presented
primarily for convenience. Appendix Tables III and IV comprise, the bflsi<-
data reported by the two labs from which the above groupings were con-
structed, and from which different size groupings may be computed as the
needs arise.
Where Table C indicates the presence of asbestiform fibers, these have
been identified by the labs as asbestiform fibers of the amphibole family
in the grunerite-cummingtonite series of fibrous minerals of the type found
in various deposits of the Iron Range. Three samples, all from the Cloquet
area, constitute exceptions; one lab found chryeoti-le asbestos fibers, which
is,a serpentine variety of asbestos used in most asbestos-containing manu-
factured products, such as auto brake linings, asbestos cement water pipes,
etc., but found no amphibole fibers. The other lab found no asbestos
fibers of any type in the duplicates of these 3 samples.
Both labs used identical criteria for defining what constitutes a
fiber; this working definition identified as a fiber any particle with a
length to width ratio (aspect ratio) of at least 3 to 1. Morphology,
electron diffraction patterns, and other techniques were then employed to
provide identification of such fibers as asbestiform fibers of a given
asbestos class.
ll/ Derived from Data Appendix Tables I, III, IV, and V
-------
Table B — Listing of Aumb.-r and Typ;; of Water Samples Received
and Status of Completion of Electron Microscopy Analyses
by Both KPA-Oni.ractor bibs, by Major Sampling Location.
Over-all Sampling Program
Samples Received nnd Rosi'i.
Reported by Both i.nhs f'o
Duplicate Siimplos Taken fi<
the Samp Locntions - to
1
• I
_>.
4.
*~
6!
• .
• > .
j *
H).
f t
1.2.
; ;-5 .
VI.
5.
!6.
' 7.
o!
. 9.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Nuaibftr 5- Type of Sample R<
tt & Type of Sample
Location* Received
(See Legend)
Silver Bay 3-DS/2-EFF/3-IN
Beaver Bay 3-DS/2-EFF/2-IN
Duluth 11-DS/10-EFF/3-IN/2-HYD/
2-WW
Two Harbors 3-DS/2-EFF/2-IN
Ashland , Wis . 2-DS/2-EFF/2-IN/2-HYD
Cloquet 2-DS/5-IN
Superi or , Wi s . 4-DS/2-EFF/2--IN/2-HYD
Write Pine, Mich. 2-DS/2-EFF/2-IN/2-Hyd -
Eagfe Harbor, " 2-DS/2-EFF/2-IN/2-HYD
Grand Harais 2-DS/2-EFF/2-IN/2-HYD
Virginia 2-DS/2-EFF/2-IN
Eveleth 2-DS/2-EFF/2-IN
Aurora 2-DS/2-EFF/2-IN
Hoyt Lakes 2-DS/2-EFF/2-IN
Minneapolis 2-DS/2-EFF/2-IN/2-HYD
Ramsay , Mi ch . 2-DS/2-EFF/2-IN/2-HYD
Hfim. P&L Well 2-WW
Fond du Lac Well 2-WW
F pfra^rwH ^V^irrvl Mw^t O T"\ C
NE Exp. Station 2-DS
Gooseberry
State Park 2-DS
Eckbetk Carpgid . 2-DS
Elde's ComsrSpcig 2-DS
Beaver River 2-RW
Stewart River 2-RW
Baptism River 2-RW
St. Louis River 2-RW
Wild Rice Lake 2-LW
Pike Lake 2-LW
Birch Lake 2-LW
Subtotal - 30 major locations*
56-DS/38-EFF/37-IN/8-RW
6-LW/16-HYD/6-WW
Add ERDLATOR »
Samples: 20
Total for all samples
* = Minnesota locations unless otherwise
Leqend: DS ~ Distribution System/EFF =
"l/*G11f«.*t/J / f * f\ '•'•"I Y\ "1 £5 t" O rl *"*" "
.- Ut; .* Vt/U / *-•(* it;t j.G L ^y
Analyses Total
Completed Received
2-DS/J -EFF/2-IN
2-DS/l-FFF/l-IN
8-DS/7-EFF/1-IN -
8
7
2-'Arl'.' ^8
2-DS/l-EFF/l-IN
2-DS/l-EFF
Al % CG1Fi;^r]t(5"tGCl
All completed
2-DS/l-EFF/l-IN
1-DS/l-EFF/l-IN
2-DS/1-EFF/2-IN/1-HYD
1-DS/l-EFF
1-DS/l-EFF/l-IN
All completed
All completed
1-DS/l-EFF/l-IN/l-HYD
—
—
—
—
—
1-DS
1-DS
—
2-RW
2-RW
2-RW
2-RW
2-LW
1-LW
1-LW
36-DS/23-EFF/22-IN/8-
7
8
7
10
8
8
8
6
6
6
6
8
8
2
2
2
2
2
2
2
2
2
2
2
2
2
2
RW/4-LW/4-HYD/2-WW 167
20
shown
20
07
O 1
Effluent/ IN - Influent/ RW
Total
Completed
4
2
16
2
2
6
10
2
0
4
0
0
6
6
0
0
0
0
0
0
0
0
0
2
2
2
2
2
0
0
— ,
70
16
PA
oo
( Per '.•••'.;
Comp 1 •-; i. '
( 5( )''•_!
( 2C %
( 57^0 ;
( 29?;
( 2%'. •
( IOC1 -',
( 100;;
( 2?"',
( o-...
( 50,;
( 0, j
( 0,'e
( 100'v
( 1 (:> )•;.'.
( o;1,'
( o;:
(' 0°0
( o;'
( 07,
( O/
( O/o
( o;o
( 0%
( loovo
(100%
( lOO/o
( 100%
( 100%
( o;.
( 0;;i
( 42%:
(100%
f .,0,
'iO,.i
= River water
LW = local lake water/ HYD = Fire hydrant/ WW = well water
JL2/ Derived from Data Appendix Tables I - V.
-------
Table B —
Li si Ing
of ^ui'ibcf Lnd Typo of Water Samples Received
and Status of tk/npletion of Electron Microscope Analyses
by Both EPA-Ci.ri tract or Labs, by Major Sampling Location.^/
Over-all Sampling Progrem
Number & Type of Samp 1 e Re c:ej ved / Co nip1 e t ed
Satrrples Received nnd Ro?s I
Reported hy Hoih LiUis fo\
Duplicate Samplos Taken <\
the Same Lor.ntions - to *!,•
Location*
. Silver Bay
. Beaver Bay
, Duluth
.
)2.
i 3.
!'i.
5.
16.
"7.
0.
,9.
20.
21.
22.
23.
24.
25.
26
27.
28.
29
30.
_
» & Type of Sample
Received
(See Legend!'
3-DS/2-EFF/3-1N
3-DS/2-EFF/2-IN
11-DS/10-EFF/3-IN/2-HYD/
2-WW
3-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN/2-HYD
2-DS/5-IN
4-DS/2-EFF/2-IN/2-HYD
2-DS/2-EFF/2-IN/2-Hyd -
2-DS/2-EFF/2-IN/2-HYD
2-DS/2-EFF/2-IN/2-HYD
2-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN
2-DS/2-EFF/2-IN/2-HYD
2-DS/2-EFF/2-IN/2-HYD
2-ww
Analyses
Two Harbors
Ashland.Wis.
Cloquet
Superior.Wis
Wite Pine, Mich
Eagfe Harbor, "
Grand Marais
Virginia
Eve let h
Aurora
Hoyt Lakes
Minneapolis
Ramsay, Mich.
ftfim. P&L Well
Fond duLac WeH2-WV
Laksvocd Sdiod, !«*, 2-DS
NE Exp. Station 2-DS
Gooseberry
State Park 2-DS
Ecfeedc Carpgid . 2-DS
Elde's CoOTrSprig 2-DS
Beaver River 2-RW
Stewart River 2-RW
Baptism River 2-RW
St. Louis River 2-RW
Wild Rice Lake 2-LW
Pike Lake 2-LW
Birch Lake 2-LW
Subtotal - 30 major locations: . . .
56-DS/38-EFF/37-IN/8-RW
6-LW/16-HYD/6-WW
Add ERDLATOR t
Samples: 20
Total
Received
Total
Completed
2-DS/1-EFF/2-IN
2-DS/l-EFF/l-IN
0-DS/7-EFF/1-IN -
2-IW
2-DS/l-EFF/l-IN
2-DS/l-EFF
All completed
All completed
2-DS/l-EFF/l-IN
1-DS/l-EFF/l-IN
2-DS/1-EFF/2-IN/1-
1-DS/l-EFF
1-DS/l-EFF/l-IN
All completed
All completed
1-DS/l-EFF/l-IN/l-
1-DS
1-DS
2-RW
2-RW
2-RW
2-RW
2-LW
1-LW
1-LW
8
7
28
7
8
7
10
8
8
HYD 8
6
6
6
6
HYD 8
8
2
2
2
2
2
2
2
2
2
2
2
2
2
2
4
2
16
2
2
£
10
2
0
4
0
0
6
6
0
0
0
0
0
0
0
0
0
2
2
2
2
2
0
0
36-DS/23-EFF/22-IN/8-
RW/4-LW/4-HYD/2-WW 167
20
20
Total for all samples 187
70
16
86
( 29%
( 25"<
( 10r;;<
( 100;.'-
( ov,
( soy:
( o;:
(• o%
( o'-
( o:-.'
(100%
(100?o
(100%
(100%
( o?,
( 0"o
* = Minnesota locations unless otherwise shown
Legend: DS ~ Distribution System/EFF ~ Effluent/ IN ~ Influent/ RW = River water
LW = local lake water/ HYD = Fire hydrant/ WW = well water
121 Derived from Data Appendix Tables I - V.
-------
17
As Table C reveals, both labs were in agreement in 36 out of the
43 water samples respecting the presence or absence of asbestiform fibers;
that is, both found a definite presence of such fibers in 19 of the samples
of which they had examined duplicates, and a definite absence of such fibers
in 17 samples. Three of the 7 samples for which one lab found a presence of
asbestiform fibers, whereas the other lab found none, involved the Cloquet
area samples which contained chrysotile asbestos fibers; the remaining 4
samples were found by one lab to contain amphibole asbestos fibers. __
Table C —r Asbestiform Fibers and Fiber Si7,e Distributions Found in
Duplicate Samples via Electron Microscopy Analysis performed
by University of California School of Public Health and
McCrone Associates, Inc., of Chicago, 111. 137
Number of Fibers
Type
Location* Sample
1. Silver Bay
2. Beaver Bay
3.Duluth
4. "
5. "
6. "
7. Two Harbors
* 5. Silver Bay
9.SuDerior , Wi s.
lO.Jeiio's *1 (Duluth)
11. " #2
12. " »3
13. Superior, Wis.
14. Cloquet
15. "
16. "
17. Superior, Wis.
IP,. White Pine, Mich.
19. Grand Mara is
20. Aurora
21.Hallett Well.Dulrth
22. Ashland, Wis.
23. Superior, Wis.
24. "
25. Grand Mara is
26. Aurora
27. "
20.Hoyt Lakes
29. "
••>/-* M ii
•jl) .
31 .Beaver River
32. Stewart River
33. Baptism River
34. St. Louis River
35. Wild Rice Lake
DS
DS
DS1
DS2
DS3
DS4
DS
IN
DS
EFF
EFF
EFF
EFF
IN
DS
IN
DS
DS
DS
IN
WW
DS
IN
HYD
IN
EFF
DS
IN
EFF
DS
RW
RW
RW
RW
LW •
MCCA
UCSPH
(Millions of Fibers
per 1 iter water)
2.0
3.0
5.0
2.0
1.0
1.0
2.0
.5
.3
.7
1.0
.1
.7
BLACL**
II
It
tt
ft
It
tt
Ii
4.37
5.29
2.76
6.03
2.26
1.47
2.53
1.47
1.10
3.87
.67
.71
.37
.39***
9Q#$$
.98***
.39
.18
.25
.26
BLACL**
II
II
Fiber Length -
Percent
UCSPH
<3
o/
/o
50.
35.
64.
44.
45.
3-5 >5
(Micrometer
7
8
9
1
8
60.0
32.
76.
71.
68.
56.
56.
34.
-
-
-
-
-
-
-
5
0
4
4
9
1
2
o/
/O
43.
38.
29.
39.
39.
30.
36.
16.
14.
28.
29.
41.
52.
-
-
-
-
-
-
-
7
9
8
0
6
0
4
0
3
9
4
5
6
o/
/O
5.
22.
5.
16.
14.
10.
31.
8.
14.
2.
13.
2.
13.
-
-
-
-
-
-
-
Dis
tri bul i
MCCA
<3 3-5
size classes
6
5
3
9
6
0
1
0
3
7
7
4
2
«/
/O
96.
97.
100.
98.
83.
100.
87.
92.
95.
100.
100.
100.
100.
86.
100.
100.
100.
100.
90.
100.
8
4
C
5
7
0
3
9
2
0
0
0
0
6
0
0
0
0
9
0
3.
o
fca »
—
1.
10.
_
9.
7.
4.
_
_
—
-
6.
—
_
_
_
9.
_
>5
o/
2
6
_
5
2 6.]
_
1 3.(
i
8
_
_
_
_
7 6.'
_
_
_
_
1
_
Continued next page with US Army ERDLATOR results)
-------
Table C — Continued
18
Number of Fibers Fiber Length - Percent Distribution
US ARMY ERDLATOR RESULTS
UCSPH
MCCA
UCSPH
MCCA
1.
li.
3.
4.
v •
6.
7.
B.
Run »8
Run #9
Run #9
Run »7
Run #7
Run #9
Run #3
Run ^8
IN
EFF
IN
EFF
EFF
EFF
EFF
EFF
( Mi 1 1 i on s
.3
.2
.8
.3
.1
.3
BLACL**
BLACL**
per Liter)
2.04
6.05
8.24
.39
.99
.46
BLACL**
BLACL**
3
50.
100.
54.
83.
100.
100.
-
-
3-5
0
0
8
3
0
0
25.
-
2.
15.
-
-
-
-
0
3
0
5
25.0
-
2.3
1.7
-
. -
-
-
3
96.2
100.0
95.3
100.0
100.0
100.0
—
-
3-5 :
3.8 -
-
2.8 1.9
_
_ , _
— _
_
* Minnesota location unless otherwise indicated
** BLACL means "Below lowest acceptable counting limit"; this is a variable encounter,' i
from sample to sample, is dependent on other suspended particulates content of speci
fie samples, and may comprise anywhere from "more than zero" to "less than 50,000"
fibers.
***" Chrysotile asbestos fibers were found in these samples; no amphibole fibers.
13/ Derived from Data Appendix Tables I, III, IV, and V
-------
19
Communities in whose water supplies asbestiform fibers were
fcnirid to be present by both labs include Silver Bay, Beaver Bay,
Duluth, Two Harbors, -- all in Minnesota, cs well as Superior,
Wisconsin. Asbestiform fibar counts ranged from a high of 8.24
million fibers per liter for raw Lake Superior water, sampled near
the'EPA National Water Quality Laboratory at Duluth and used in one
of the U.S. Army ERDLATOR Treatment Unit tests, to a low of .1
million fibers per liter for the effluent from Jeno's Diatomaceous
Filber #3, also at Duluth,
Communities where one lab found asbestiform fibers to be present,
whereas the other lab found none, include White Pine, Michigan,
Grand Marais and Aurora, Minnesota, and Cloquet. Chrysotile, rather
than aiaphibole asbestos fibers were found in the Cloquet samples, as
previously discussed.
Sampling points for which both labs determined an absence of
asbestiform fibers include the Hallett Well at Duluth; Ashland,
Wisconsin; Superior, Wisconsin; Grand Marais, Aurora, Hoyt Lakes,
Beaver River, Stewart River, Baptism River, St. Louis River, and
Wild Rice Lake in Minnesota.
It should be noted that several communities are listed in more
than one of the categories identified above, and caution should be
exercised in translating these largsly incomplete findings into a
basis for remedial action. A quick reference to Table B will estab-
lish that complete analytical results, as reported by both labs for
duplicate samples which they both examined, are presently in hand
for only 9 of the 30 major locations sampled. These include Cloquet,
Hoyt Lakes, Aurora, the 4 Minnesota river locations, Wild Rice Lake,
and Superior, Wisconsin. For 6 of these 9 locations, no asbestiform
fibers were found by either lab. For Aurora, Superior and Cloquet,
however, the situation is different.
The Aurora water plant intake was found to contain asbestiform
fibers by one lab; the other lab saw none. The Aurora water plant
effluent and distribution system, however, as sampled at the water
plant outside tap and the City Hall utility room tap, respectively,
were found to contain no asbestiforra fibers by both labs. It would
appear desirable to subject the Aurora supply to re-sampling and
re-analysis at an early date.
For the City of Superior, Wisconsin, the presence of asbestiforra
fibers was reported by both labs for the water plant effluent (tap
in laboratory sink) and for the Superior Power and Light Company's
Underbill Service Center at 2915 Hill Avenue (tap in the meter shop).
One lab found asbestiform fibers in the water tap outside the Superior
Power and Light Company water treatment plant (Lake Superior water
from the Cloquet Pipeline); the other lab found no fibers in that
sample. Neither lab found asbestiform fibers in the intake water
(well water) for the Superior Power and Light Company water plant
-------
20
(tap in laboratory sink) or the, fire hydrant (well water) near the
Company's Underbill Service Center (hydrant near building at 2900 Hill
Avenue).
For 3 samples taken at Cloqutst, the lab found chrysotile asbastoa
fibers; the other lab found no asbestiform fibers. The locations
examined included a raw well water intake tap inside the pumphouse
building, influent water from Lake Superior off the Cloquet Pipeline
in the Northwest Paper Company Plant, and a tap at the fire station
at Cloquet Avenue and 5th Street.
The City of Cloquet recently converted to use of water from a
30 million gallons per day pipeline carrying water from Lake Superior 14/
but apparently has retained a flexibility to use well water previously
constituting the sole supply in conjunction with Cloquet Pipeline
water. The shource of the chrysotile asbestos contamination is
presently unknown. A review of construction contract specifications
for the Cloquet Water Line 15/ shows that the piping used was steel,
rather than asbestos-cement pipe of the type likely to contain
chrysotile asbestos. Detailed contract specifications did call for
coating all subaqueous steel pip© "..with coal tar enamel, fibrous
glass mat, bonded asbestos felt wrap coating per AWWA C-203,
Section AL.4, 'Steel Pipe Exterior Protection for Submarine Lines'...."
Further, the interior and exterior surface of the inlet structure
"....shall be cleaned, priir.ed and coated with coal tar enamel and the
exterior shall be wrapped with bonded double asbestos felt wrappers...."
Pending a thorough on-scene search for the origin of the chrysotile
contamination, it would appear desirable to resample Cloquet Pipeline
and Cioquet well supplies at an early date.
The probable presence of asbestiform fibers at Grand Marais and at
White Pine, Michigan, is indicated by the findings of one lab; the
other lab saw no fibers in the samples it examined. Although neither
lab found asbastiform fibers in the intake water of the Grand Marais
Water Plant (tap off spigot on intake pump), one lab found such fibers
in the Grand Marais Municipal Building Fire Station (tap in the fire
truck bay). Only 50 percent of the Grand Marais results are presently
in hand.
For White Pine, Michigan, only 25 percent of the analytical results
are presently available. One lab found asbestiform fibers in water
sampled at the boiler room tap of the Townsite central heating plant;
the other lab saw no asbestiform fibers in the duplicate sample it
examined.
14/ June 28, 1973, "Report on the Duluth Water* Situation" to
Gov. Anderson, by the Minnesota Department of Natural Resources
IS/ "Specifications for Cloquet Water Line, Cloquet, Minnesota"
Ruble Miller Associates, .Inc., Duluth, Minnesota; "Conformed
Contract" (undated) Set #32
-------
21
Caution is again advised in viewing theso findings an definitive
for locations where results, as reported by both laba, ars not in hand
and remain to be evaluated.
Respecting the essentially arbitrary grouping of asbcsntif orrn fibers
into the length classes discussed previously, the following findings
obtain for samples of which duplicates were examined by both EPA
contractor labs. As evident from Table C, one lab found asbestiform
fibers longer than 5 micrometers in 16 of the 19 samples found to
contain fibers, whereas the other lab found fibers of that length
class in only 4 of the 26 samples which it determined to contain fibsrs
of any length. For the 3 to 5 micrometer length group, the former
lab found fibers of that length in 16 of the 19 samples, contrasted
with 11 out of 26 samples for the latter lab. Asbestiforra fibers
shorter than 3 micrometers were present in all samples found to
contain fibers of any length by either lab.
No evidence is now in hand which could serve to explain or
reconcile the apparent divergence of findings respecting the lengths of
fibers found by the two labs. Pending the receipt of additional results
which would form a basis sufficient for thorough evaluation, it appears
unwarranted to attribute the divergence to inter-lab differences in
sample preparation and transfer techniques since both labs employ
procedures which aim to prevent the breaking up of fibers present
in the samples to the maximum possible extent. On the basis of the
findings presented in Table C, it is possible to generalize that the
asbestiform fibers found to be present in the samples examined to date
by both labs are predominantly of a length shorter than 5 raicronetars.
Pending the development of precise location-by-location data, the
overall number of people exposed to drinking water supplies determined
thus far to contain asbestiform fibers remains unknown. An upper
limit can be established by positing a "worst case" assumption to the
end that the entire population of the communities whose supplies
were found to contain asbestiform fibers to date was being served at
the time of sampling by those same supplies. Local water supply
authorities thoroughly familiar with the layouts and peculiarities
of their individual systems will alone be able to determine how many
people were served with the supplies present in the systems at the
locations and dates of sampling. A tabulation giving the 1970 Census
populations for the communities in whose supplies asbestiform fibers
were found to be present to date follows below; it would constitute
the "worst case" upper limit of exposure addressed above.
-------
22
1970 Population
Community (UP.Iessr othery;ise shQnr,)_ _16/
Silver Bay, Minn. 3,504
Beaver Bay, Minn. 905
Duluth, Minn. 98,400 *
Two Harbors, Minn. 4,437
Cloquet, Minn. 8,699
Grand Marais, Minn. 1,301
Aurora, Minn. 2,531
Total Minnesota 119,777
White Pine, Michigan 1,218
Superior, Wisconsin 32,237
Total above communities 153,232
* 1970 population was 100,578, but Minn. Dept. of Natural
Resources estimates that 98,400 persons are served through
the Lakewood Pumping Station IT/
16/ US Census of Population, 1970; General Population Characteristics -
Minnesota, US Dept. of Commerce, Bureau of the Census
JT7/ June 28, 1973 "Report on the Duluth Water Situation" to Gov. Anderson,
by the Minnesota Department of Natural Resources
-------
23
2. TENTATIVE CONCLUSIONS
Although analytical results presently in hand represent approxi-
mately 46 percent of the data bose at this point in time, it is
reasonable, nevertheless,, to advance some tentative conclusions
respecting levels of water contamination by asbestiform fibers in
the Western Lake Superior Basin area.
1) Results reported for 43 duplicate water samples analysed via
electron nicroscopy by two indapendent EPA-contractor lab-
oratories show a definite presence of asbestiform fibers
in 19 samples (44%), a probable presence in 7 samples (16%),
and a definite absence in the remaining 17 samples (40%). IB/
2) The definitional framework relied upon in framing these
tentative conclusions is as follows:
a) All conclusions related to the presence or absence of
asbestiform fibers refer to water sampled in late June -
early July of 1973, and therefore reflect levels of
contamination then prevailing.
b) A "definite" presence of asbestiform fibers is assumed
where both EPA-contractor labs reported finding asbesti-
form fibers greater than the "Lowest Acceptable Counting
Limit" for any duplicate sajsplfea which they both examined.
See c) below.
c) A "definite" absence of asbestiform fibers is assumed
where both EPA-contractor labs reported finding no
asbestiform fibers for any duplicate samples which they
both examined; this is identified as "BLACL", meaning
"Below Lowest Acceptable Counting Limit" - a variable
encountered from sample to sample which is dependent on
other suspended particulates content of specific samples,
and may comprise anywhere from "more than zero" to "less
than 50,000 fibers".
d) A "probable" presence of asbestiform fibers is assumed
where one EPA-contractor lab reported finding no asbesti-
form fiber (BLACL), whereas the other reported levels
above "BLACL". This divergence is encountered in a total
of 7 samples out of the 43 duplicate samples examined by
both labs; its significance is discussed later.
3) The asbestiform fibers found by both EPA-contractor labs for all
but three duplicate samples were identified as amphibole asbestos
fibers in the 'grunerite-cummingtonite" series of fibrous
_18/ Derived from Data Appendix Tables III - V
-------
minerals of the type found in various deposits in the "Iron
Range". (In three samples, all frova the Cloquet area, one
EPA-contractor lab found chryaotile asbestos fibers, a
serpentine variety of asbestos used in most asbestos-
containing manufactured products, such as auto brake linings,
asbestos cement water pipes, etc.; the other lab found no
asbestos fibers (BLACL) in tho same three samples.)
4) Pending agreement by medical authorities respecting the
significance to human health of ingesting asbestos fibers and
respecting the significance and plaxisibility of a quantitative
determination of an "acceptable" level of asbestiform fiber
content of water supplies, prudence dictates that an absence •
of such fibers is clearly preferable to their presence.
5) Using this prudent approach makes sense, particularly in the
absence of scientific studies showing that asbestiform fiber
ingestion is beneficial to human health, and thus forms the
basis for arranging the analytical results for various
localities in a fashion ranging from "Contaminated" to
"Not Contaminated". Such a ranking follows below, without
intermediate ranking reflecting degrees of severity, for all
samples for which fiber counts were reported by both labs to
date. 19/
Contaminated — Definite Presence of Asbestos Fibers
1. Duluth Area
211 *t
« i
3
511 ii
• i
6. " , "
7. " , "
8. " , "
g ti ii
10 *. " i "
11. " , "
12. " , "
13. " , "
14. Silver Bay
15.
16. Beaver Bay
17. Two Harbors
18. Super!or,Wis,
19. " , "
— Arlington Water Pumping Station; tap off west pump
— Fire Station at 3501 Woodlawn Ave.; tap in fire
trvrck bay
— Duluth Airport; custodian closet tap, main floor
— Duluth Pump. Station, 4424 Grand Ave.;north pump tap
— Jeno's Diatomaceous Filter #1
__ " " " #2
— " " " «3
— Duluth Raw Lake Water,before ERDLATOR Run #8
— " " " " , before " Run *9
- , after " Run *9
— " " " " , after " Run #7
- " , after " Run #7
— " " " " , after " Run »9
— Standard Gas Station, corner of Highway 6; and
Outer Drive; outside tap eastside of building
— Pump House influent; tap off incoming line
— Jim's Shell Station,Hwy.61; car bay tap
— Fire Station sink tap, 7th St. and 2d Ave.
— Underbill Service Center, 2915 Hill Ave., tap in
meter shop,Superior Power & Light Co.
— Water Plant effluent; tap in laboratory sink
JL9/ Derived from Data Appendix Tables I, III, IV and V
-------
25
b) Probably Con tarn j natcd --Probable pre s e ?: c e of A she s t i rorm_FU)e r
2. White Pine, MJdu
3. Grand Marais,
Minn.
4. Aurora,Minn.
*5. Cloquet,Minn.
1. Superior,Wis. -- Lake Superior water tsp outside water treatment, plant
Superior Power & Light Co. (from Cloquet Pipeline)
Town site heating plant; boiler room tap
Municipal ouilding or.ci fire station tap in
fire truck bay
Water Plant intake
Raw well water; intake tap inside building in pumphoup
*6. " , " — Influent - Northwest Paper Co. Lake Superior influent
. tap in plant off pipeline
*7. " , " -- Well water - fire station Cloquet Ave. and 5th St.
(sometimes uses Lake Superior water from pipeline)
* Fibers found in these samples were chrysotile, not amphibole,
raising the possibility that contamination occurs via an asbestos-
containing manufactured product, such as asbestos-cement pipe, or
from some other source.
c) Not Contaminated — Definite Absence of Asbestiform Fibers
1. Duluth Area
2.
3TI tl
.
4. Ashland,Wis. -
5. Superior,Wis. -
6 « ii
.
7. Grand Marais
8. Aurora ' -
9. "
10. Hoyt Lakes
11. "
12.
13. Beaver River -
14. Stewart River -
15. Baptism River -
16. St. Louis River •
17. Wild Rice Lake
Hallett Well (Jeno's) 59th Ave.W. Tap off pump.
Duluth Raw Lake Water, after ERDLATOR Run #8
" , after ERDLATOR Run «8
Water utility office; tap in basement at 320 E.2d St.
Superior Power & Light Co. well;intake for water
plant; tap in laboratory sink.
Superior Power & Light Co. (well); fire hydrant,under-
hill service center near building 2900 Hill Ave.
Water Plant Intake; tap off spigot on intake pump
Water Plant effluent; outside tap
City Hall; utility room tap
Water Plant; intake tap in main building
Water Plant;effluent tap after pump,main building basement
Police Department; garage tap
Stream sample off Highway 61 near Beaver Bay
Hwy, 2, N. of Two Harbors,downstream side of bridge
Eckbeck State Park near Finland,Minn.
Hwy. 2 bridge, 200 yards downstream of bridge
Dam Rd. off Minn. Route #4; sampled at the dam
-------
26
6) Until medical authorities develop a consensus respecting
"harmful" versus "acceptable" .lovels of asbestiform fibers
in water, it makes little sense to attempt to rank analytical
results for the various localities by "intermediate" degrees
of severity of contamination by contrasting, say, 3 million
fibers per liter versus 1 million fibers per liter. Whatever
degree of precision would obtain from such an attempt would
be partially obliterated by the lack of precision inherent
in electron microscopy techniques employed to determine
numbers and sizes of asbestiform fibers in water.
7) Because we find ourselves at the "frontier" of the state-
of-the-art in this respect, differences in water sample
preparation techniques, differences in analytical procedures,
and, indeed, differences in operator experience and expertise
— all speak against attempting to impart precision beyond
that which is deserved. The only certainty which obtains
at this point in time is the certainty of the electron micro-
scopist who sees an asbestiform fiber under his microscope
and identifies it as such beyond the shadow of a doubt, using
his equipment to its full capability, and drawing on his
knowledge and prior experience with full confidence.
8) Thus, while for 7 of the 43 samples reported on here, no fibers
were found by one EPA contractor, the fact that the other lab
did see and identify asbestiform fibers cannot be dismissed or
taken lightly. Although these 7 sample results thus lack the
corroborative validity present in the case of the other 35
samples, the presumption should be that asbestiform fibers
were indeed present in the water at the time the samples were
taken. The communities involved should have early priority
in any future re-sampling effort, particularly also since
chrysotile asbestos fibers were found at three locations,
suggesting that contamination may come from asbestos-cement
water conveyance pipe or some other yet unknown source.
9) No medical consensus presently exists respecting critical
lengths or widths of asbestiform fibers for production of
disease; an Occupational Health and Safety Administration
emergency standard for occupational exposure of asbestos
industry workers addresses inhalation limits only, and
prescribes maximum allowable exposures couched in terms of
asbestiform fibers longer than 5 micrometers. As now on
th© books -- aside from specified peak short-period exposure
limits — the OSHA emergency standard for asbestos fiber
inhalation limits exposure to no more than 5 fibers per cubic
centimeter of air inhaled over 8 hours in a 5-day week,
effective July 7, 1973, and no more than 2 fibers per cubic
centimeter of air, effective July 1, 1976.
-------
27
10) In the absence of medical studies establishing that the
5 micrometer fibor length, is critical in terms of inoestod
asbestiforra fibers, there is no baais for assuming that any
particular fiber length >
-------
CHAPTER II
PRELIMINARY ANALYSIS OF DATA ON
ASBESTIFORM FIBERS IN THE AIR FROM
SILVER BAY, MINNESOTA
-------
CHAPTER II
TABLE OF CONTENTS
Page
A. Introduction 28
B. Sources of Data 2«
C. Data from the EPA Membrane Filter Network 32
D. Data from the EPA Scrap ling Effort in Minnesota ,2
E. Comparison of Silver Bay Downwind Data to Other Downwind-of- 3-
Source Location Data
F. Comparison of Silver Bay Background Data to Urban Concentrations 40
G. Comparison of Silver Bay Background Data to Non-Urban and
Remote Data 4 2
H. Conclusion on the Silver Bay Amphibole Concentrations 42
-------
CHAPTER II
ANALYSIS OF ASBESTOS DATA
A. Introduction
Chrysotile asbestos concentrations in air have been measured at a
number of locations in the United States during the past four years.
In order to place the Duluth/Silver Bay amphibole asbestos data in
perspective, the following section consists of a comparative analysis
of the ambient concentrations of asbestos in the Silver Bay
area with chrysotile asbestos concentrations measured in other places
in the United States. Whereas the amphibole asbestos concentrations
in Silver Bay are likely to be the residuals from the minerals processed
in the area, it should be noted that the chrysotile asbestos concentrations
measured in other U. S. locations may be emitted either by point
sources (asbestos mines, fabricating facilities, processing facilities
and the like) or by other, more diverse, human activities such as
auto braking, asbestos spraying, and building demolitions.
The conclusions presented here should be used with great
caution. A comparative analysis such as this one should compare
similar pollutants for different locations. This is not possible in
this case because of the limited data on both amphiboles and chrysotile
which are available. Electron microscope analytical techniques have not
been used previously to measure asbestiform amphibole concentrations in
the ambient air. As a consequence, data on ambient air concentrations of
the asbestiform amphiboles have not been collected.
The analysis which follows assumes that the two forms of
asbestos — chrysotile and amphibole — are comparable.
The significance of this assumption is not known with regard to
the objectives of the analysis which is to compare the mass
concentrations of amphibole asbestos in the atmosphere of Silver
Bay to chrysotile concentrations in other parts of the country.
This analysis does not attempt to relate potential health
risks to mass concentrations in air. Current research findings do
not provide answers to the question: Are there different health
responses when humans are exposed to giv-2n mass concentrations of
chrysotile and amphibole asbestos? The assumption that the two
asbestos groups are similar permits the undertaking of such a
comparative analysis. However, it is important to recognize that
there are differences in fiber shape, density, chemistry, size, and
aerodynamic properties between the two forms.
B. Sources of Data
The data used in the comparative analysis have been taken from
-------
29
published and unpublished studies that have been conducted over the
last four years by:
Battelle Memorial Institute,
Mount Sinai School of Medicine,
U. S. Environmental Protection Agency, and
State of California Air anci Industrial Hygiene Laboratory
(References are provided at the end of this section.)
Most chrysotile measurements to date have been collected by the EPA
National Membrane Filter Network. The Network is located in 55 cities
across the country. A map shown in Figure 1 indicates the cities in
which the samplers are operated. The Network, which has been operating
since 1969, has been used to monitor chrysotile asbestos concentrations
in urban areas. The sampling sites are located in downtown areas where
the traffic and population density is substantial. The sites are generally
located on roofs of post offices, fire houses, or other public
facilities. Composite samples /i are analyzed quarterly. The membrane
filters were analyzed by the Mt. Sinai School of Medicine from the second
quarter of 1969 through the fourth quarter of 1970. Since that time,
Battelle Memorial Institute has analyzed the samples.
Other data that are used in the comparative analysis have been taken
from various studies. Air sampling for asbestos was conducted by
Bathe!le near sources of asbestos emissions and in urban and rural areas
at the time the electron microscopy analytical procedure was developed.
The work performed by Mt. Sinai generally examined alleged asbestos
problem areas in New York City such as at construction sites where asbestos
is sprayed, street corners where fibers are released by auto braking, and
the like. The EPA data for Ambler, Pennsylvania, were developed during
a field investigation of a potentially hazardous source of asbestos.
Finally, the State of California has conducted air sampling near asbestos
mills and in urban locations while developing analytical methods for
asbestos.
/!_ A composite sample for an urban sampling station is prepared in
the following manner: (1) a 24-hour sample is taken every twelfth
day during a calendar quarter; (2) some scheduled sampling days are
missed because of other priorities, hence there are not always 7
or 8 sampling periods per calendar quarter; (3) the 5 to 8 membrane
filters collected in a quarter are cut into pie-shaped wedges of
1/5 or 1/8 and pieced together to form.a full-sized filter; (4) the
composite filter is then analyzed by the electron microscope
analytical procedure.
-------
/ »i
1 *l •
\ !-'/
y / i.
' : l
/ r
f
•
1
1
H
i
/2
;l
*.i
I
1
»
a *
i
; }
; '
1
x'
^x-^^
TL.n.
« ti
I* !
i i
3 J
I i
/ /-
i ;
—-If- j
1! l!
si Jl
, °! «!
i !
i !
i s.l
I J* i
1 • !
i r—
;rf !
j|i j . !
j |i
11 !
j f
/ J i
/ ii i
/ ? "i i
1 > \ '
1 I" 18- !
r ! i i
• i i I ! ^-^-
!» ! " i-tT
i 3 ! aXij
s'
iS^^s -
"%'^ L
' 50-^^:j
«,- *^- F'! i
^\, f*** «,^, 5
cn ° r-1
-------
h-'.!.- '
It;"., •
r.;-i i =
^thi"
h,.!:1-!
u,'—'.-.
'. ' *.-'. I >
$f-'- \"-\
i .f
t •
H L}
?. O
MO
1^,.!;,-
>5^
-------
30
A summary of asbestos concentrations for selected locations is
presented in Figure 2/i, to show r.he range of values-that have baen
observed. Notice that material in Figurs 2 ia plotted on seven cycle semi-
log paper. In aorno instances, raw data points h&vo been indicated. In
other instances, where only a ra?ig« of concentrations was presented in
the literature, this range has b?en indicated. Further, in some cases
where the number of analyzed samples is large, a range has been indicated
for the sake of clarity.
The data in Figure 2 have been grouped into three categories which
are based on sampling site location. The categories are:
1. Downwind-of-source locations,
2. Urban locations, and
3. Non-urban/remote locations.
This classification permits the grouping of data collected under similar
conditions and, hopefully, minimises the likelihood of comparing data that
are dissimilar.
Caution should be taken before making comparisons between the
"downwind-of-source" data and the data found in urban and non-urban
locations. There are two significant reasons why the downwind-of-source
data arc unlike the other data.
First, short-term samples at sites similar to site #3 detailed in Silver
Bay (see Section II 2 for a description of the sampling site and results
for Silver Bay) which is 700 feet from an emission source yield concentrations
which are far from representative of the long-term average concentrations at
such location because wind directional data indicate that the location is
an upwind site a majority of the time. The data from site #3 represent
measurements which were taken to indicate the absence or presence of amphibole
asbestos near the Reserve Mining Facility. Accordingly, the samples were
taken under atmospheric conditions that would yield high amphibole
concentrations. Long-term measurements at the same location would yield
average results significantly lower than the short-term peak concentration.
This relationship is true for all of the downwind-of-source data. The
EPA National Membrane Filter Network, on the other hand, measures
chrysotile concentrations at equal intervals over extended periods of time.
Atmospheric conditions vary from one sampling period to another. As a
consequence, the "urban mean chrysotile concentration" (see Figure 2),
which is the mean of a number of long-term (i.e, quarterly) ambient
measurements for 55 cities, is not directly comparable to the short-term
Silver Bay results.
The data plotted in Figure 2 have been taken from the references
that are listed at the 'end of this chapter.
-------
31
Secondly, samples taken at sites immediately downwind of emission
point sources measure the plurae before significant atmospheric diffusion
and dispersion dilute the pollutant. A sampling site further downwind of
an emission source will generally measure concentrations that are lower
than those at a nearby site. Concentrations are generally plotted as a
function of distance from a point source as isopleths which are shown in
Figure 3.
As a result, the concentrations measured at site #3 at Silver Bay
are not representative of the concentration expected in the residential
area of Silver Bay. On the other hand, the measurements obtained from
the EPA National Membrane Filter Network were taken at locations in
populated urban areas. Thus, the data are not comparable directly.
Contrary to the downwind-of-source relationship between short-term
and long-term measurements, short-term concentrations for most urban and
non-urban/remote locations are generally representative of the long-term
concentrations at a given site. This is primarily because of the lack of
a major sole source of the pollutant. Thus, for such areas the expected
difference between short-term measurements and the long-term mean
diminishes since the statistical variation is smaller.
The laboratory techniques used in the past studies have not been
verified completely by EPA for reproducibility or accuracy. Past data
were used on the assumption that the accuracy and reproducibility of the
techniques were similar to that of the techniques used to measure
amnhibole concentrations. Certain evaluations of laboratories were
conducted. For example, a detailed review of the State of California
Air and Industrial Hygiene Laboratory techniques and work revealed
problems that would place past data in question.
Further, recent analyses of identical arnphibole air samples by
Battelle and Mt. Sinai yielded the results in Table 1 A.
TABLE 1 A: COMPARISON OF BATTELLE AND MT. SINAI REUSLTS FOR IDENTICAL FILTERS
Filter Amphibole Mass Concentrations (ng/nP) (Mt. Sinai Result)
Number Battelle Mt. Sinai (Battelle Result)
A-84 100 *
A-85 62 20 0.3
A-97 570 3700 6.5
A-98 950 600 0.6
* No amphibole fibers detected.
-------
32
Such variations have not been fully explained. The comparative
analysis which follows relies only on analytical results from Battelle
for samples taken in the four Minnesota communities since Mt. Sinai did
not analyze samples other than the four shown above.
^• Data from the EPA Membrane Filter Network
An analysis of the results from the EPA National Membrane Filter
Network is presented because, although the data are not for amphibole asbestos,
they allow at least some comparison with levels of the more common
variety of asbestos and give some perspective as to relative concentrations
of asbestos, in so far as present data permit. The Network has yielded
a substantial quantity of data for a number of the populated areas of
the United States.
The distribution of chrysotile concentrations for the period starting
in 1969 and ending with the first quarter of 1972 is presented in
Table 1, Frequency distribution plots are shown in Figures 3 through
5. Note on Figures 3, 4, and 5 that the data are skewed significantly in
the direction of the lower values. Although the simple arithmetic mean
concentrations for all cities during the 1969-1972 time period is 3.2
nanograms/m3 (io~9 g/m Or ng/m3), the median value is 1.6 ng/m3. This relation
ship indicates that the mean is significantly affected by some high values.
However, only 14, or four percent, of the 314 samples are greater than 10
The mean values for 187 samples analyzed by Mt. Sinai and the 127
samples analyzed by Battcllc are very similar:
Arithmetic Mean (Mt. Sinai) = 3.3 ng/m3,
, Arithmetic Mean (Battelle) =3.2 ng/m3'
a difference of approximately 3 percent. The small difference in these
results from Battelle and Mt. Sinai for chrysotile asbestos concentrations
can be contracted to recent analyses of identical amphibole ambient air
samples by the two laboratories which have yielded mass concentrations
by nearly an order of magnitude.
High chrysotile concentrations (i.e., values greater than 10 ng/m3)
obtained by the Network were evaluated to determine if one or two urban
areas consistently experienced high concentrations. It was concluded
that no one urban area had high readings consistently. As Table 2
indicates, only 2 urban areas had more than one reading which exceeded
10 ng/m3 for the 1969 - 1972 period. Tables 3 and 4 show a ranking of
the 15 urban areas with the highest average concentration as measured
by Battelle and Mt. Sinai.
D. Data from the EPA Sampling Effort in Minnesota
EPA conducted an ambient air sampling program for asbestiform
amphibole concentrations in Babbitt, Hoyt Lake, Duluth, and Silver Bay,
Minnesota during June 1973. Membrane samplers, identical to those used
-------
33
TABLE 1
Distribution of Chrysor.ile Concentration in the Ambient Air of United States
Urban Areas
Chrysotile
Concentrations
(ng/m3)
0 - .99
1.0-1.99
2.0-2.99
3.0-3.99
4.0-4.99
5.0-5.99
6.0-6.99
7.0-7.99
8.0-8.99
9.0-9.99
10.0-10.99
11. 0-11. 99
12. 0-12. 99
L3. 0-13. 99
L4. 0-14. 99
16.0-16.99
18.0-18.99
34.0-34.99
38.0-38.99
40.0-40.99
95.0-95.99
TOTALS
Samples
Analyzed by
Mt. Sinai
Number of
Samples
61
58
23
13
9
4
2
1
4
1
1
2
-
2
1
1
1
-
1
1
1
187
Percent of
Total
32.6
31.0
12.3
7.0
4.8
2.1
1.1
.5
2.1
.5
.5
1.0
-
1.0
.5
.5
.5
-
.5
.5
.5
- 100%
Samples
Analyzed by
Battelle
Number of
Samples
27
33
22
11
9
9
4
3
3
3
-
1
-
—
—
-
1
1
-
-
-
127
Percent
of Total
21.3
26.0
17.3
8.7
7.1
7.1 '
3.2
2.4
2.4
2.4
-
.8
-
-
—
-
.8
.8
-
-
-
^100%
Samples Analyzed by
Mt. Sinai & Battelle
Number of
Samples
88
91
45
24
18
13
6
4
7
*4
1
3
-
2
1
1
2
1
1
1
1
314
Percent of
Total
28,0
29.0
14.3
7.6
5.7
4.1
1.9
1.3
2.3
1.3
.3
1.0
.6
.3
.3
.6
.3
.3
,3
.3
.100%
-------
PERCENTREE
IBLEI
Z0J3
30.0
Has
fr* n .,
j
-
o~"
M
*-
—
s-
»-
)
\
\
-I
••Hit
?
3
3
o j> »
i - .
U)
• in
H W Ifl
»t^ ^ »y
\X3 /
ua 3 T]
\ to JJ
3 —
3
1
H
3
3 i
01
HI
i
i
c n
3 m
m 1 >
! z
j] i
z
J
| i
PI
i 1
in i
c
in
A
D
Z
A
PI
Z
H
J
I
^
D
Z
in
z
i
PI
in
H
n
c
H
D
Z
D
I
A
I
•MM*
in
D
H
r
PI
s
-------
PI
El
y
j
h
D
in
n:
i
v
L
n
z
n
P
3
10
E
h
in
6
i
i
y
i
h
z
U1
z
n
i
D:
z
y
V
z
D
in
^
tt tt
i I
J ^ « . C
i (N CM
h « i
E § a
ID i §
D
C
••••M*^^
-5
(••^
-8
-
BU»
1 !
1
1
1
1 i 1 1 I I
1 1 1 I 1
0*02 0'0i 01
-------
• y
j
h
Q
in
M*
y.
I
v
L
D
Z
D
h
3
to
K
h
in
in
a
i
y
i
i-
z
in
z
D
P
Q:
1-
Z
y
z
D
V
f [
^ /N-
«! OB
3 11 \
b. in - u c
° 5 if!1 !
Hi — rr VD
I 1 ffl S : : ;
ILL • £j
, ••
-------
TABLE 2
Chrysotile Concentrations Greater Than iOng/m" As Measured by the
Environmental Protection Agency l-fembrane Filter Network
Chrysotile
Urban Area
Dayton, Ohio
Long Beach, California
Memphis, Tennessee
Memphis, Tennessee
San Diego, California
Seattle, Washington
Dallas, Texas
Ft. Worth, Texas
Detroit, Michigan
El Paso, Texas
Los Angeles, California
New York, New York
Los Angeles, California
San Antonio, Texas
:entrat3 on
££/ui3)
95.0
40.0
38.0
34.3
18.8
18.0
16.5
14.4
13.4
13.3
11.8
11.0
11.0
10.2 ,
Date of
Measurement
- 4Q 1969
4Q 1969
4Q 1969
1Q 1972
1Q 1972
3Q 1969
4Q 1969
3Q 1969
3Q 1969
4Q 1969
3Q 1969
2Q 1969
4Q 1971
2Q 1969
-------
35
TABLE 3
Cities
Chrysotile Asbestos Coucentrat i
Laboratory; Ht. Sinai
Dates ; 2nd Quarter 1969 through 4th Quarter 1970
Sample Size; 4 Composite Samples
Chrysotile Concentration (ng/m*')
Rank City
1 Dayton, Ohio
2 Long Beach, Calif.
3 Memphis, Tenn.
4 Seattle, Wash.
5 El Paso, Texas
6 Kansas City, Kans.
7 Detroit, Mich.
8 Dallas, Texas
9 Ft. Worth, Tex.
10 San Antonio, Tex.
11 Phoenix, Arix.
12 Rochester, N. Y.
13 Los Angeles, Calif.
14 San Francisco, Calif.
15 New York, N. Y.
S*£fc
95.0
40.0
38.0
18.0
13.3
8.0
13.4
16.5
14.4
10,2
8.9
8.6
11.8
5.0
11.0
Low
0.4
1.3
0.4
0.6
0.3
3.9
0.5
0.3
0.9
0.9
1.3
1.4
0.7
1.8
0.5
Average
24.3
11.3
10.2
5.8
5.8
5.7
5.1
4.9
4.6
4.3
4.3
4.3
3.8
3.7
3.3
-------
36
TABLE 4
Chrysotile Asbestos Concentrations for Selectgj...IL...S ^jC
Laboratory: Battelle
Dates ; 1st and 4th Quarters 1971
1st Quarter 1972
Sample Size: 3 Composite Samples unless otherwise shown
f\
Chrysotile Concentration (ng/irr)
Rank City High Low Average
1 Memphis, Tenn. (1 sample) 34.3 - 34.3
2 San Diego, Calif 18.8 3.0 10.1
3 Los Angeles, Calif 11.0 6.8 9.1
4 New York, N. Y. 9.4 1.7 6.8
5 El Paso, Tex. 7.6 5.5 6.2
6 Kansas City, Kans. 7.4 4.0 5.8
7 New Orleans, La. ' 5.9 4.5 5.2
8 Phoenix, Ariz. 8.8 1.8 . 5.0
9 Denver, Colo. (2 samples) 6.5 3.0 4.8
•
10 Atlanta, Ga. (2 samples) 8.8 0.8 4.8
11 San Francisco, Calif. 5.8 , 3.2 4.7
12 Long Beach, Calif. 5.3 4.0 4.5
13 Chicago, 111. (2 samples) 7.2 1.2 4.2
14 St. Louis, MO.(2 samples) 6.1 2.1 4.1
15 Jersey City, N. J. 5.8 2.3 3.5
-------
37
in the EPA National Membrane Filter Network, were installed at selected
locations in the four Minnesota communities. Maps showing the sampling
sites for Silver Bay, Hoyt Lake, Babbitt, and Duluth are presented in
Figures 6, 7, 8, and 9, respectively. One or more meteorological stations
were operated in the vicinity of the samplers to collect wind speed and
wind direction data for the area.
Millipore filters (0.45 u) were used to collect the samples, which
were analyzed by Battelle Memorial Institute. The analytical method used
in the program is described in Appendix I.
The results for thirty-one (31) samples are reported in Table 5.
The analyses are classified as "re-runs" since they are the product of
a second attempt by Battelle to analyze the amphibole mass concentrations
gathered during the June sampling program. Problems with the first-run
analyses left those results in question in as much as Battelle refused to
stand behind them.
The analytical results of samples taken in Silver Bay have been
used in the comparative analysis. A detailed map of Silver Bay, which
shows sampling sites #1, #3, and #4 is presented in Figure 10. Site #1
is located 1.2 miles from the source and is referred to as a "background
site" (the sampling station was located adjacent to a road in the woods
and was distant from a significant amount of human activity). Site #3 is
referred to as a "process site" since it is immediately downwind of the
Reserve Mining facility when oil-lake brcescc prevail. Site #4 is so
located that it was downwind of the process part of the time and at other
times it served as a "background site."
The data for Duluth, Babbitt, and Hoyt Lake have not been used
directly in the comparative analysis. This is so primarily because most
attention has been directed recently to the potential health problem in
Silver Bay. In addition, the samples collected in Duluth and Hoyt Lake
in the June 1973 field sampling program were less than adequate for a
comparative analysis for those areas because of such problems as the siting
of samplers. Consequently, a second field sampling program was undertaken
in August 1973 in Silver Bay, Hoyt Lake, and Duluth. Hoyt Lake, and not
Babbitt, was indued in the second sampling program because of a desire
to assess the potential emissions from on-land disposal operations. The
samples from the second field program have not been analyzed thus far
because of analytical problems with the nylon-backed Millipore filters.
E. Comparison of Silver Bay Downwind Data to Other Downwind-of-Source
Location Data
Asbestos concentrations measured downwind of known sources are presented
in Table 6. No precise comparison can be made between the data for
Silver Bay and the other data because, as discussed in Section B, short-
term concentrations measured downwind of a given source can vary extensively
depending upon:
-------
•p
•H
O
0) O
-------
w
o
-------
00
o
C\4
©
Vl O
•»*
O -H
O FJ
O O
XAS
II
-------
tP K
as
o
H
1
H
2
W
§
u
CO
VO
o
O O r-l CO
«# e-H V0
I I I I
P» in
CN (N
I I
O O
CN r-
I
f*>
m r- t-
in
O O
CN in o
^3* &> i—i
i i i
ro O CN
CN m in
ro
(1)
-P
t7>
C
Gi
oo
in
CN
CO
(1)
4J
•H
CO
pj
S T)
O
bn tl*
o >;
sa o
(0
n
rocess Sites
PH
•o
n
•H
^
c
o
Q
5
•H
W
0)
•H
tn
c
-H
H
•H
^
C
•H
^
C
O
Q
H
4-* M t
o
o o
O t-«
r~ in
4J
•H
to
O QJ
4J 4J
•H -^
•O 10 CO
3 "O Tf
O C C
U ? ?
ifl O O
PQ Q Q
JS
•d
4)
3
W
0)
o
-p
(0
•H
•o
0)
-------
0)
fO
NO
3
fl)
cu o
^ %
H b
co
to
C
T-t
CO 4-1
O cd
CO -H4J
CU M C
rn rA 1
•
_i
cu
O CU
0 CJ
TO 0 H
•POP
CO M O
1-1 Fu CO
Q
/~s
o oo,£
pj \— •*
.0 1-1 t3
4-» rH O
00 D.-H
0 B H
CU CQ CU
»J co fn
rH ?•>
OS k
O O
•H 4J
•U C}
r^t r-4
tH O
C d3
< i-5
^
CO
t~-
rH
1
co
m
CN
CU
rH
rH
CU
4-1
4-1
rt
M
O
C
O
li i i
^•^ ^n
0 4J *-N
^vTO
Q) 1-1 0
60 4J --x
(3 (3 tvC
H >
^4
6
rH
•
O
ft
a
o
•H
4J
CO
U
^
•H
4J
c
•H
O
•H
U
•
M
»Z5
*ri (T1 W)
(x, 6 c
4J Tl
rH V! 4 '
CO CU CO
CQ ^Q O
O W -r-l -U
W ,0 I'J
•rH 4J tfj i — 1
Q Ct) li| fX(
^.
CM
CU
TH
rH
CU
4J
4->
TO
ta
o
0
o .
f*^t
fM
CM
Q)
rH
•H
4-1
o
CO
?"*
^1
6
»
•
«d
fi<
A
M
CU
rH
^Q
g
^
•r^ ' i' '
<~i >_;
pi ;_i C-.
H TO C'
•IJ
<-o (-0 iu
o: C>
Cj .H rO
^-i >H ( ".;
O CM -£j
•
«w
m
o
0
•
0)
}*4
JS
m
•
CO
Cs)
£
M
4-)
n)
o
ON
o
m
-*
Q)
rH
0
J3
•H
J3
Q-
f
•
U
2
««
t^
rt
ta
M
u
>
rH
tH
•H
CO
S5
01 -H
4-J CO
•H CO
a >
&
n
rH
_\j
(J
O
tJ
«4-l
O
CO
pi
s
o
H
0
4-1
CO
M
O
rQ
to
»1
-.
w
1
rH
<0
•H
M
4J
CO
•0
0
rH
Md
8
n
^
*
rt
•H
e
o
(M
Tl
rH
Ct)
O
Vi-4
O
CU
4J
rt
4-1
CO
CM|
-------
! Vi • '• •"'. -r^TL-^i—-*•"• TV-***^ .~~\T
', [ l\A.^'\^-,*"V^~£/*&^^'-r'^~^- ^rfW
' ) J^'.IS-'' \^-<<:i^--"-^iLv A.Jfi'-.-fjrW-V'^^C 'J/
•
• //If % <-}•/_ "P^.'. "f~'.^
~ !-C\ \L.; -° '-"-'^
iim^^i«
-------
wind speed
wind direction
wind varibility
- atmospheric stability
the distance between sampling site and the source, and
other factors such as topography and precipitation.
Neither: (1) the atmospheric variables; nor (2) the distances between
samplers and source were fixed during the data-gathering efforts that
developed the information in Table 6.
Relevant observations that can be made from Table 6 are:
1) The amphibole concentrations reported for Silver Bay are
not uniquely high when compared to chrysotile data
that were taken immediately downwind of known asbestos sources.
2) High asbestos concentrations may be observed immediately
downwind of asbestos mills and fabricating facilities the
emission being wind-blown fibers from one or more
tailing piles or from process exhaust stacks.
Some of the measurements reported in Table 6 were taken in residential
areas. However, others, such as those in Silver Bay, Minnesota, and
King City, California, were not.
F. Comparison of Silver Bay Background Data to Urban Concentrations
Long-term sampling for amphibole asbestos in the town of Silver
Bay has not been conducted. As a result, data on annual ambient concentrations
of amphibole asbestos for the residential area of Silver Bay do not exist.
Such data are necessary to make a comparison with chrysotile concentrations
for urban areas in the United States. /!_
However, inasmuch as the source is located within 0.5 miles of the
shopping center in Silver Bay one can reasonably assume that the long-term
concentrations there will be equal to or greater than the background
concentration detected approximately two miles from Silver Bay. An
average background concentration of 33 ng/m3 was determined from two samples.
/I "In the Reserve Mining case a court-ordered air sampling program has
been undertaken that is designed to yield amphibole concentrations
which are representative of the conditions in residential Silver Bay
on an annual average basis.
-------
41
Chrysotile asbestos concentrations for a number of urban areas are
shown in Table 7. The average background sample of amphibole concentration
near Silver Bay is greater than 98.8% of the measurements reported by EPA
for 55 United States urban areas. The av?.rage background concentration for
Silver Bay is ten times the urban mean chzysotile concentration for 55 U.S.
cities.
TABLE ? : Reported Asbestos Concentrations in Urban Areas
Range of
Location
Berkeley, Calif.
Auto Braking
Sites, NYC
EPA Urban
Network
(55 Cities)
World Trade
Center, NYC
Philadelphia
N.Y.C.
Dayton
Houston
San Francisco
Pittsburgh
Wash., D.C.
Asbestos
Form
Chrysotile
Chrysotile
Chrysotile
Chrysotile •
Chrysotile
Chrysotile
Chrysotile
Chrysotile
Chrysotile
Chrysotile
Chrysotile
Concentr
Low
22
32
0.1
3.5
45
1.7
0.4
4.0
1.5
2.0
1.6
atioa(ng/o3)
Hi£h
72
165
95
365
100
65
11
6.0
40
8.0
40
Number of Mean Median
Samples (ng/m^) (ng/m3)
2 — — —
6 76
314 3.2 -C2.0
22 40 21
7 __
17 ~ — —
7 — — __
4
3
3 __
4
-------
G. Comparison of__S_il._ver_priy Background Data to Non-Urban and Remote'^Datci
Measurements of asbestos concentrations that were taken neither
in urban locations nor downwind of known asbestos sources have been
summarized and are presented in Tczble 8. The data are separated into
three categories, these being:
1) upwind sites in vicinity of asbestos point source,
2) suburban locations, arid
3) remote locations.
The background measurements for Silver Bay are not different from
measurements taken upwind of other known point sources of asbestos. In
addition, it is worth noting that the average background amphibole concentra-
tion for Silver Bay (33 ng/rn^, which is the average of two 8-hour samples --
see Table 8) is ten times the urban mean chrysotile concentration. The
other background concentrations measured in the vicinity of known asbestos
point sources are also substantially higher than the urban mean chrysotile
concentration. These observations may support the generally accepted
assumption that asbestos fibers in the ecosystem surrounding a source
will tend to remain in the area by adhesion to vegetation and other
ground level objects and not be totally removed with clearing weather
fronts. The suspension of fibers from water bodies containing fibers
under high wind velocity conditions has also been hypothesized.
No conclusive statement can be made regarding the reported suburban
concentrations because of the small sample size. HOwever, the data
appear to be more representative of urban locution data thandata reported
for upwind sites near asbestos point sources.
The remote concentrations reported for Frankfort, Kentucky were
measured near a fish hatchery that was at the time seven miles from
a paved road. The measurements were taken at a remote site where
chrysotile concentrations were believed to be extremely low. The average
background amphibole concentration at Silver Bay (i.e., 33 ng/m^) is
412 times greater than the average of three chrysotile measurements
taken at the fish hatchery. Additional data were not found that would
indicate if amphibole mass concentrations in the range of 0.01 to 0.1 ng/m-3
are representative of concentrations in other remote locations in the
United States.
H. Conclusions on the Silver Bay Amphibole Concentrations
A comparative analysis of available data on amphibole and chrysotile
concentrations has been conducted. The results of the analysis should
be used with caution. Many reasons support this conservative viewpoint:
a. The analytical techniques used to measure amphible asbestos
are not yet standardized. Different techniques have been used at
each laboratory that has analyzed samples. In addition, the
analytical results for the laboratories have differed. As an example,
four identical filters that were analyzed by the Mt. Sinai School of
Medicine and the Battelle Memorial Institute have yielded results
that differed by an order of magnitude.
-------
b. The laboratory techniques that yielded past data used in the
comparative analysis have not been verified for reproducibility
and accuracy. Past data was used on the assumption that the accuracy
and reproducibility of tho techniques were similar to that of the
techniques used to measure amphibole concentrations,
c. The samples used in the analysis hove been taken under different
sampling conditions and fcr c3ifferent purposes (i.e., the EPA
National Membrane Filter Network vs. the Minnesota studies).
Generally, the data frora 1970-1972 wore generated during research
programs where sampling and analytical methods were being tested
for validity.
Keeping the limitations of the analysis in mind, the following
conclusions have been determined:
1). Araphibole concentrations measured downwind of the Reserve
Mining processing facilities are not uniquely high. Chrysotile
concentrations equal to or greater than the Silver Bay amphible
concentrations have been measured in Ohio and California near
asbestos plants.
2). The average concentration of two background (or clean site)
samples measured in the Silver Bay area are ten times greater
than the urban mean chrysotile concentration. The average
amphible concentration for the background site in Silver Bay
is greater than 98.8% of the 314 chrysotile concentrations
ir.sasursd in 55 U. S. cities during the 1970-1972 period.
-------
44
TABLE 8: Asbestos Concentration for Background (Upwind) Sites
Type
Background
Site
Location
Range of
Asbestos Concentration(ng/m^) Number of
Form Low High Samples
Upwind site in vicinity
of Asbestos Point Source
Upwind site in vicinity
of Asbestos Point Source
Upwind site in vicinity
of Asbestos Point Source
Upwind site in vicinity
of Asbestos Point Source
Upwind site in vicinity
of Asbestos Point Source
Port
Alleghany,Pa. Chrysotile 10
Country
Club,
Ambler,Pa. Chrysotile —
King City,
Calif. Chrysotile
Remote Location
Hoyt Lake,
Minn. Amphibole 20
Silver Ray;
Minn. Amphibole 23
Frankfort,
Kentucky Chrysotile
0.03
30
— 57 —
27
42
Suburban Location
Suburban Location
Ridgewood,
New Jersey
NWQL Roof
Duluth, Minn.
Chrysotile — 20 —
Amphibole 8 11
1
2
0.12
-------
REFERENCES
Heffelfinger, R. F., C. W. Melton, and D. L. Kiefer, "Development
of A Rapid Survey of Sampling and Analysis for Asbestos in Ambient
Air (Final Report)," Battelle Memorial Institute, Columbus, Ohio,
Columbus Laboratories, Environmental Protection Agency Contract
CPA 22-68-110, February 29, 1972.
Data not formally published that was developed in conjunction
with the Reserve Mining case.
Draft Report, "Asbestos Pollution Problem at Ambler, Pennsylvania,"
April 11, 1972 (Prepared by Region III of the Environmental
Protection Agency).
Nicholson, William J., A. N. Rohl, "Asbestos Air Pollution in
New York City (Final Report)," Environmental Sciences Laboratory,
The Mount Sinai School of Medicine of the City University of
New York, New York, N. Y., December 2, 1971.
Personal communications, Dr. S. David Shearer, Environmental
Protection Agency, Office of Research and Development,
September 1973.
Nicholson, W. J., Ruul, A. N. turn Ferranu, E. F., "Asbestos Air
Pollution in New York City," In Proceedings of the Second Clean
Air Congress. (Edited by England, H. M. and W. T. Barry), Academic
Press, New York (1971).
Stanley, R. L., et al, "Status of Analytical Methods for
Determination of Asbestos in the Environment (AIHL Report No, 151),"
Air and Industrial Hygiene Laboratory, California State Department
of Health, Berkeley, California, September 1973.
-------
CHAPTER III
PRELIMINARY ESTIMATE OF THE OCCURRENCE
OF ASBESTIFORM AND OTHER POTENTIALLY HAZARDOUS
FIBERS' IN THE TACONITE INDUSTRY
-------
CHAPTER III
TABLE OF CONTENTS
Page
A. Introduction 46
B. The General Geology and Mineralogy of the Biwabik
Iron Formation • " 47
C. Distribution of Taconite Mines Along the Mesabi Range 48
D. Asbestiform Minerals Other Than Grunerite-Cummingtonite
Along the Mesabi Range 49
E. Asbestiform Minerals Expected in Emissions from Loading
Pellets 50
F. Taconite Mining in States Other Than Minnesota 51
G. Potential Pollution Associated with Activities Other
Than the Mining and Milling of Iron Ore 51
H. Conclusions 54
-------
CHAPTFK III
PRELIMINARY ESTIMATE OF TlIF, OCCURRENCE
OF ASBESTIFORM AND OTHER PO1TW a ALLY HAZARDOUS
FIBERS IN TIIK TACONITL INDUSTRY
A. Introducti on
The geology of the Biwabik iron formation from which iron ore
is mined along the Mesabi Range in northern Minnesota is well known
because of its economic importance. Long-standing work by the U. S.
Geological Survey, studies by the Minnesota Geological Survey,
investigations by expert witnesses, as well as observations and
analyses conducted by the staff of the Office of Technical Analysis,
all indicate that the following questions can be answered:
(1) Are the five other companies currently mining taconite
along the Mesabi Range, and the company which has announced
it will begin operations in 1975, likely to be found to have
the same minerals in the same concentrations in their emissions
to the air as does Reserve Mining Company at Silver Bay?
(2) Are any asbestiform minerals not present in Reserve's
emissions expected to HP .in the emissions of the other
companies? If so, what are the expected concentrations of
such asbestiform minerals?
(3) What asbestiform minerals are to be expected in the
emissions resulting from the loading of pellets of beneficiated
iron ore into interlake ore carriers?
(4) Is the taconite mined by Reserve mineralogically similar
to that mined by other companies outside of Minnesota? Would
the same problems of air pollution by asbestiform minerals be
expected there?
(5) What other sorts of mining, manufacturing, or agricultural
activities would be expected to have high concentrations.of
asbestiform fibers?
-------
47
B* The General Geology and Mineralogy of the Biwabik Iron Formation
The Mesabi Range consists of a low-lying ridge which trends
northeasterly across almost 130 miles of central and northeastern
Minnesota (Figure 1).
A prominent type of rock in the Mesabi Range is the Biwabik
Iron Formation, made up of a series of primary sedimentary iron
oxides, hydroxides, silicates arid carbonates deposited in a
sequence of layers which ranges from 400 to over 600 feet thick
and dips underground toward the southeast at about 10°. The
mineralogical characteristics vary considerably along the outcrop
of the formation as well as down its southeasterly dip. This provides
a basis for dividing the formation vertically into 5 major members
and as many as 22 submembers (Figure 2).
A massive igneous intrusion, the Duluth gabbro, occurs at the
northeastern-most end of the Biwabik Iron Formation. During the
emplacement of this igneous body, the eastern Mesabi Range between
Mesaba, Minnesota and Babbitt, Minnesota underwent high-grade contact
metamorphism at temperatures which may have reached 700-750 degrees
centigrade (Gundersen and Schwartz, 1962). Silicate minerals,
including amosite, as well as other minerals were formed from pre-
existing ones within about 3.5 miles of the Duluth Gabbro where this
contact metamorphism was effective. The rest of the formation
remained relatively unchanged by this event. The mineralogical
patterns jjxouuced by contact metamorphism of the Biwabik Iron
Formation provides the basis for dividing the formation laterally
into zones, and for predicting the mineralogy at any point
along the formation (Figures 3 & 4).
Portions of the Biwabik Iron Formation exposed at the surface
of the earth underwent weathering for hundreds of millions of
years. During weathering, the iron minerals were oxidized so
that a rich iron ore (hematite, Fe203) remained as the major residue
where many other iron-bearing minerals had previously existed.
These pockets of hematite were mined as a "high-grade" primary
ore and shipped to the blast furnaces without further concentrations.
As these deposits were depleted (oxidation extended only to the
uppermost exposed layers of the formation and in limited areas
along the formation), the "low-grade" iron ores containing magnetite
(Fe30^) admixed with silicate minerals were examined for their economic
potential. Low-grade siliceous iron ores are called taconites.
They may consist of up to 30% magnetite in a matrix of quartz and
-------
U
z
u.
O
u
g
-in
O
M
O
•H
M
0)
CO \D
O>
0) iH
§w
O W
00 M
l-i
C§
O \-t
t~t v^
I
•H
(U
60
-------
ii?
*~ >*
V
5
_j
i/>
a
UJ
a
CL
3
>-
^-
rc
UJ
C>
K
u;
>
>-
<
/{
T
(1
V."
J
SV
kl CS
1
o
jlSl&l '>P-J»?Oje.itxix)
1 (KKXlft UJ«!t03)
| «>.QUi»rt
•§
<
J3
3
<
J?
<
i1
<
4'
•5
*f
-I*®
«- ? - C71
5 3 5 '.
fc 7 £ s
>/> s ili s
irf -S 0
CO "
& _.pr;
8 06)
C (12)
0 t?)
E <6>
F (20)
G (25)
H 00)
t (5)
J (i 6)
K (35)
L (30)
M (20}
N (4)
0 (17}
P (60)
Q (ZG)
R (Ml
S (1)
T fol
U CO)
k-v~T'~
CENEffCLl^ED COLUMNAR BCOTtCN Or THE bf^-aDx '^OvJ-irO.SMaT(QN
IN THE EASTERN Mt,S>,lB! DiSTRlCT, f.iNNfcSOU
!cyered {<.i'Qp$ide} chert toco^itt 'oroliy wiih ho'nbip-nrff , hedfnbprgifc and
lcminaf€d {feifohyper^the^e-mjLfi^tife) Quartz focon,>e with hedenbergite
oid fa>o!tte (eo;.!) and fan snytcd (c L'fr.fnif g fan- fe-f^ogr^ fife) chert
toconife (•nest)
fwavy lan-.mcted factinfilite-rrrcgnefifr) chert tacowtt locdly wtfh cummfng-
/ Joniie end 'ninor hedfinfagf gi ,e <
•^ quortz taconite with abundant granule structures end Icco'Iy w-ifh quortz-
Ss^ filled sep'ona structures, minci mognelite, cuiTiminctonits ond actinolifs
Shaly tzdded (cunvntngtarete-mug-^dte) quartz tacwts wrin m.nor orxirod:lc
ond hedenbergtte (cost), locally abundant cumrningtonire (wtsf)
quartz focomtg (east) ond moft/fd (andratiite) qucrtz tcconite (*esJ) w»lh
obundonl mcgnsliie- bear ing granules Ihraughou!
S wavy layered foe tmohte -magnet it?) quartz tacantte wetita-rich granules ond
>i pebbles, congiomefal'C fobnc through ou 1 ; minor hjmoli'e
granule (magnetite) quartz tacontte wdn abundant mjQnet'tB-nch pebbles near
\. lop ond thickly layered (magnetite) qvartz toconite neoc boltom
wavy layered (silicate-magnetite) Quartz f aconite wiih obu^don) mogneiite-
rich granules o"d pebblei; S'licoies o^e actmotils ond f^rronypsrjih»ne (eott)
ond cumnrnojontlB (w«st]
wavy layered (stiicafe'Tflogretitf} sibcote- quartz (aconite w,th abundant mog-
ond ho'oblendft (eosl) and cumfrinqtoniis ond ochno!ti« (west), silicates with
layered (magnetite) fayabte- quartz tccomle with l«rrohypersth«ne (eosfl ond
layered (magnetite) cummtngtontfc -quartz toocnife (wsst)
S fayolits -quartz tocomte *ith (errohypersthene (eosl) end cummingtanrte '-quartz
bedded granw'e (magnetite) quartz- fcyctite tcconite wiih some (errohypersthene
\. n^|if*-beonng granules (w«st)
i \
shaty quartz -fcyolite toconite to fayalitf taccnite (eos)) ond shafy qvortz-
cummingtorxfe facoftt/e to cummngtoryfe taconite (west); minor mogne'ile
/ m.nor foyclite, b.otite, aimandits ond pyrrhoMte (cost) ond traces of pynte, pyr-
/ fholite ond cumnTn^ionife («(»sl)
/ 1
/ I layered (magnetite) faya'>f? -quartz tocomie wilh minor cummmglonite
i 1 io^cred (mogne'if?) cjuGrtz tacanife with minor cunmingtoniie throughout, and
// ^ede foyolife (cost)
Figure 2~ After Gundersen and Schwartz, 1962.
-------
-------
Distonco from Contact (D) (miles)
00 SO • ir
1 1 1 "1 1 1 '1 I
ZONE 1
*
< Grcenolile
Chomosilo ^
* 000
Minnesolcjjte a. a A
*•-, *
We
e
X Incipient decomposition
D Melomorphic fobrlc
In quorlz groin*
Quortt
8^-§ — -$r%— ^~u~
Siderite
ftnUrilc
Hematite
o & Pi
Mognetita
Pyrila
1 II 1 1 1 1 1
OO 5O 1O
.0 S.O
TFTT 1 1 i
ZOK'£ 2
r- -i n
M — §-® —
O O ©6
•D SfflQ
6.
1 » ' >
II III 1
0 5.O
ZC£L
0J?
xy
=u£
r
oQ
i
1.0 O.', Q
1 1 1 I II I
ZONE 4
*
1, ja -— _ An
Hedonborgi «
Kerrohyperslhene
FoyoI'Me
— r~! n TT
Colcile
Pyrholilt
IIMI it i
1.0 0.5 0.
.Distance from Contact (D) (miles)
Figure 4 - Diagram showing the mineralogy of the Biwabik Iron Formation
in different metamorphic zones as a function of distance from
the contact with the Duluth Gabbro. Small circles indicate
locations of specimens; dashed lines indicate uncertainty of
occurrence, (after French 1968)
* Fibrous minerals discussed herein.
-------
43
other minerals. The Biwabik icon formation also contains a number
of other silicate minerals associated with the magnetite and quartz.
When suitable techniques were developed vhJch made thece rocks
economically exploitable, th«ce taconite oros became the primary
mining objective. Recovery of the magnetite necessitated processing
the associated rocks as well, fibrous amphiboles among them.
C. Distribution of Taconite Mines along the MeF-;.bi Range
Reserve Mining Company's Peter Mitch-.-11 pit lies within Zone 4.
(Figure 1) The asbestiform variety of the amphiboly grunerite-
cummingtonite (called amosite) comprises 10% to 60% of the Biwabik
Iron Formation in Zone 4 (see Figure 4). Burrell (1973) reported
that amosite composed about 20% of the ore mined by Reserve. Erie
Mining Company operates a small pit (the Dunka pit) east of Reserve's
in Zone 4. The mineralogy there is magnetite, quartz, hedenbergite,
ferrohypersthene, garnet, with but little grunerite-cummingtonite.
Erie also operates three larger pits in Zone 3 east of Mesaba
(Figure 1). In Zone 3 the Biwabik Iron Formation has been reported
to contain from 20% to 50% grunerite-cummingtonite. All the other
taconite mines on the Mesabi Range lie in Zone 1.
Erie Mining Company, the only taconite operation in Zone 3,
has completed primary mining of their pits which lie within that
zone. By the end of the year, the cleanup operations currently
underway in these pits will ba completed and no more ore will be
taken from them. Ore subsequently will be mined mainly from pits
in Zone 2, with some small fraction coming from the Dunka pit.
At that time, the emissions of amosite from Erie's plant should
decline to a low level maintained by the small volume
of rich ore mined from the Dunka pit.* Erie Mining Company is
aware of the asbestos problem and plans to step up its existing
program of revegetating the tailings dumps. This should further
stabilize the tailings and control this source of amphibole and
other silicate particulates to the air and water.
The process water used to carry tailings to the dumps is recirculated;
little runoff enters the surface drainage. The company also has
a program underway to further raduce emissions of particulates to
the air from stacks on the pelletizers.
Thus, amosite is not expected to present a significant threat
as a pollutant at any other taconite mine along the Mesabi
Range other than at Reserve's.
* See the section on asbestos in air samples for levels around Erie's plant.
-------
D. Asbestiform Minerals Other Than Crc'inerite-Cummingtonite
Along the Mesabl Ramie" " " ~~~~"~
Crocidolite and actinclite are other asbestiform amphibole
minerals which occur along the Messbi Range. Crocidolite, the
asbestiform variety of the f,:r;phibolc rxebeckite, has been reported
present in Zone 4 in "minute." amounts (Gundersen & Schwartz, 1962;
White, 1954). The amphiboio actinolice has been identified in
amounts of up to 5% in "cno <1 (Gundd-rsen & Schwartz, 1962) . The
presence of these minerals recognised as asbestos contributes to
the overall level of asbestos pollution of air and water and
increases the analytical problems of determining particle identity
and quantity.
Additional minerals not classified as asbestos may behave as
such physiologically because of their size and shape. Greenalite,
minnesotaite, talc, and hornblende are such minerals which occur in
the Biwabik Iron Formation. The mineral greenalite occurs in the
relatively unmetainorphosed part of the formation, and thus occurring
through Zone 1 and 2 and into Zone 3 (see Figure 4). It composes up
to 20% of the formation in Zone 1. Greenalite has not been included
among the asbestiform minerals which have been considered to be
hazardous in the past, in spite of the fact that its structure is
that of a serpentine, which is the mineral family containing the
principal commercial asbestos mineral chrysotile. In light of the
medical testimony that the health effect is primarily related to the
sice olid shape of a. particle, there is reason to believe that high
concentrations of greenalite may present a significant health problem
along the western-most 100 miles of the Mesabi Range.
The same is true of the mineral minnesotaite, which may make
up as much as 40% of the taconite ore in Zones 1 and 2. Minnesotaite
commonly occurs in small, fibrous shoeves about 50 microns long from
which smaller individual fibrils can be seen to radiate. Minnesotaite
has a structure and crystal form similar to that of talc, and, there-
fore, can be expected to pose a potential air pollution problem.
Talc itself occurs in the formation in Zone 1 in amounts less than 5%.
The effects of fugitive dust in talc mills on the respiratory system
of exposed workers is well known.
The amphibole hornblende is present in Zone 4 in amounts up to
5% according to French (19&8) . Dr. Irving Selikoff has described an
incomplete study done at Mt. Sinai Hospital which indicates preliminarily
that the lungs of workers engaged in driving a tunnel in New York City
through rock containing hornblende are heavily loaded. They have also found
surprisingly high levels of hornblende in the lungs of residents of
New York City. The residents were most likely exposed by the rock the
city is built upon, for hornblende is not used as commercial asbestos.
-------
-------
50
An estimate of the long tons of asbestiform minerals mined and
milled by each operator shipping pellets in 1972 allows an assessment
and comparison of the attendant potential hazards involved. Table
II summarizes the tons of pellets each operator shipped from the
Mesabi Range during 1972. Such an estimate can be made if it is
assumed that:
(1) The ratio of ore mined to pellets shipped is 3:1, most
operators must mine and process 3 tons of ore to produce 1 ton
of pellets.
(2) The asbestiform minerals identified in the Biwabik Iron
Formation are present in the ore mined in amounts equal to the
median percentage abundance reported in the literature for those
minerals.
The comparison on Table II shows that almost as much asbestiform and
potentially hazardous minerals are produced as pellets. Until medical
research identifies some threshold dosage of asbestos below which no
health hazard is likely to result, the potential for the various health
problems caused by amphiboles, and other silicates not now recognized
as hazardous, but which are emitted by taconite operations from all
zones of the Biwabik Iron Formation, remains problematical.
E* Asbestiform Minerals Expected in Emissions from Loading Pellets
Green taconite pellets are heated to greater than 1100°F in the
pelletizing furnaccc for periods of up to 6 hours. This firing melts
silicate minerals, and when tho pellets are removed from the
furnaces, the silicates are quenched to a glass. Glass has no
crystal structure, and consequently no crystalline fibers would be
expected to be emitted during the handling of pellets. Small sub-
sequant pieces of silicate glass may be emitted during pellet-
handling, however, as something less than onehalf of one percent
of the pellets break during handling. The pellets themselves then
are not expected to be a source of asbestos emissions. The potential
for silicosis among workers occupationally exposed to emissions
during pellet-handling must be considered, however.
An additional potential hazard may exist due to the possible
emissions of asbestiform minerals from the handling of so-called
"natural" ore. This ore is ground only to 0.75 inches, and is
handled wet, so it is not expected to be a potential source of air
pollution. Twelve million long tons were shipped by rail from the
Mesabi Iron Range in 1972, or about 29% of the volume of pellets
shipped from the Mesabi Range during the same period.
-------
-------
1
1-4
m
<
w
w
S
w
g
o
M
o\
§r*
CQ -
MM
*g
gg
0
8
<
N
d
«
?*
H
ESTIPORM AND POTEt;
to
a
tj
o
1
ft
H
&
M
M
TABLE
t
n
C ja r
O « r-
•H <
i-i IM .
S ° «
n i
U-t C
00.
81 T;
W tO [
JO «
^,
WO S
C 0 "e
<0 K <(
i-i VI
Jj M «
*" -v
~(3 t> r-
v u
*•» c
i-s •?
•H C C
*J 3 a
£5£
5
X
•~< c
n v
4J t) CO
e n MS
u d, ^
o a
pk 1-1 •
« .e
"e SJ c
« c w
•H l-i
B i: (u
O o? g
•w 3 O
iH O P
y 7 Si
a I* *•- '
.88 .
SSS
CM
a.
in
•0 **
in
a C H
| 1 M
* 3 ^
*« *
V
H
t~l
4)
Ck
Location
fMetanorphlc Zone)
K
t
S
1
(
b
1
•D
ftt
d
•H
H
H 0)
3 M
^ (3 JN
i r; —i
i X '
•i "O
} t r-(
: o -M
E
_P3 C
..
S
a
- «
>«
r
u
tn
Vj
u ^
•0 «N
C kO
3 -( o t>
>- M U
o u <;
^
o
-a-
g
g
i
(4
1
M
8
to
5
Reserve M:
•M
u
^ g
V V)
t)
v> "
o !n
8 &.
5S
*
•X
"T S »» (N ^ ^
22^2 ^2i S5!
-*
5s,
»A *O
•-I
«*>
A'jCMCS
« C
U U u
I JJ « ^ 41 i-( 0!
4) 1-1 U U 4J « £
WrH H *J -H S M
•HO iH O i-l O t-H
IJ C Q CO d W q)
a) O C U C 01 c
§3 Si gS3 gS^i
o^as a£i5 s^^
n ~ o
o " n
•^
3 s
«o "^
s" ~ s
i d a-
!• i i
5 5
3 •£ «
- 1 S
1 1 I
«*N
3 ,-
d S
" S
*S 8
?s a
5 . H
^1 s s
^^ ia
iifc S S
m
1
a 5 'J
s s s
£ 2
« 2 S -32
«J -H i-l 4J -rt
O -H r-1 o i-*
we c 5 c
cojo oo c£o
COIiH Gr-( Ctlr-J
Sofi es ss.3
*
CO CO Cj<»
01 CJ ^l1*
c
o
1
S3 p
§ § a S
da s '
* * S g.
13 s !
18^5
£ S S3
O .
*o
u
«
B
•¥-1
6C
V
o g
O -H
5 *« u
E
^ S
i *; c
i 1
^
i C M
3u B
: •*
&H
4J
n
M
3
<
c
•H
4)
' 3
O *-< o
<-> • 0
* 1 "
1 8 1
2 *
S S
C *J
£ 2
ca
i-(
!
!-H
•u
C
tf
.C
4J
•H
I
X
60
3
1 O
.e
ca
•H
O
S
K
•H
M
iH
Ct
§
N
1
;|
B
•3
ag
q
^
c
"
c
c
c
E
o
V
g
Q
u
§
Cr
iH
d
00
•H
Q
S
O
•o
».'
u
T<
B
O
1
s
V
fi
u
o
c
o
3
r-t
O
tx
-------
F • Taconite I- i r\ing_j.n 01l^r Statcs
The mineralogy of the Biwabik Tron Formation is not unique.
Similar iron fonnationn are found on the Marquette Range, Menominee
Range and the Iron River District, all in Michigan, and on the
Penoche-Gogeb.ic Range which lies primarily in Wisconsin (Figure 1) .
French (1968) has noted tb-t except for a lesser abundance of hematite
and the presence of fayallte in the highly metamorphosed taconite .
of Zone 4, the Eiwabik lfvon Formation is virtually identical
mineralogically to those iron forir.etions in other states which have
been described by James (1955) and interpreted by him to have formed
simply by progressive regional metamorphism. The amphibole minerals,
as well as the silicates greenalite, minnesotaite, and talc,
occur on these iron ranges, and are expected to have grain sizes
comparable to those of their counterparts on the Mesabi Range.
Table II lists the amount of pellets produced from taconite
mined from these iron formations.
G. Pollution Potential Associated with Activities other^ than the
Mining and Milling of Iron Ore
The occurrence of airphiboles in geologic settings other than in
iron formations is too widespread to cite by individual industry or
location. The fact that arnphiboles occur in igneous and metamorphic
rocks in substantial volumes as well as in the soils produced by the
weathering of those rocks, makes it probable that amphiboles can be
found in large quantities over a significant fraction of the country.
It is particulajc.ly important to note tnat amphiboles are common in
both igneous and metamorphic rocks associated with ore deposits. Thus,
problems similar to those attributed to amphiboles on the Mesabi
Range arising from the mining and milling may be expected wherever
those activities are conducted. Independent efforts by the U. S.
Geological Survey and Battelle Memorial Laboratory for OAWP are
underway to identify all mines which are expected to be sources
of asbestos.
Chrysotile is not nearly as abundant as the amphiboles. The U.S.
Geological Survey has cataloged more than 50 chrysotile mines and
prospects in the 50 states. Nine mines were listed as significant
producers in 1966 (Table III). The emissions from, and environmental
concentrations around, these sites have never been determined. However,
monitoring of air and water quality is currently being conducted at
the chrysotile mine near Lowell, Vermont, for Region I by the Quality
Assurance and Environmental Monitoring Laboratory, NERC, Research
Triangle Park, North Carolina, A similar investigation should be
conducted at the sites of several of the largest mines and processing
plants.
-------
CO
a
0
•rl Cfl
. _J f->
r~i 1-4
iH O
gH
OD
« C3
>, 0
4J h4
O >+-(
rt o
81 '
u
* rl
CM O
!•>. 4-1
CT\ C6
H K
01
ex
CO O
•
fc>
2;
M
CO
fe
4J
CO
CO
0)
4-1
rt
4J
CO
T3
01
4J
•rl
C
J3
*
0
a
•H
s
«!
g fl
co o
PL) )-l
t'' 1 1
S 5
rt
4J
C
3
S
r^
•
0
rH
•
8
60
C
•rl
C
S
E
01
co
«s
o
•H
i-H
fi O
^ 0
o. e
o> C
rt •,
cd
«
rl
0)
>
rH
•H
CO
cn
•
o
rH
k
8
00
. C
•H
C
s
0)
•H
rl
W
ctt
•O
(0
U
rH KH
Ol O
0)
4-» rH •
CO 0) O
O 0
0 4-1
01 CO 01
4= ' rl
0) TJ O
rH C
X rt G
4J H O
0> C rl
(Q M H
«
a
p!
S
•k
CO
0>
•3
K4
4-1
t^
O
«
co
*
CM
t
8
4J
01
r-l
r-l
flj
fri
rH
i
0!
•U
CO
rH
cd
§
•H
4J
rt
is
•
8
00
•S
g
cd
d
§
W
•
C
0
H
S
ti
•rl
4J
cd
S
0)
0)
«
co
•
CN
01
4-1
•H
C
0
O
rt
H
rf
o>
rH
4J
3
«
•
8
60
fl
•H
a
g
rt
C
C
at
K
•
g
•H
S
*
3
CO
S
43
CO
S
O
«
01
^
o
U
cu
4J
vH
C
O
o
(3
H
43
4J
0)
i-H
01
£
•
O
o
C
o
4J
M
f»
2
t^
R}
,C
OJ
.-1
«>
O
•
C
a
d
s
»t
^
j_>
o>
i-H
0)
^
in
•
fl
*
o
u
00
C
•H
C
g
C
0
M
M
a>
rl
iH
P.
1
•
8
o
0
M
M
CO
14H
tH
•H
rH
CJ
tj
§
rH
01
>
01
rH
0
§5
O A
M O
e-H
S
S -
rl
0)
a
rH
cd
PK
m
»
CN
«
8
CO
C
•H
0
s
§
rl
M
O>
4J
4.)
W
1
S
•
8
C
o
rl
M
m
in
U-l
•H
rH
U
4>
§
rH
0)
>
0)
rH
U
*
.a
o
•rl
S
*»
o
•rl
rH
*Q
3
(X
«S
m
-*
*
H
4-1
r!
i-H
PM
4J
01
rH
i-H
0) '
PM
rl
0)
01
C
o
•H
PH
•
O
0
C!
0
rl
H
co
<4-l
0)
rH
0
.
•
43
c;
d
S
«t
co
rH
rH
S
0)
rH
60
CO
M
m
CO rH
• •
O N
•
8
00
•S
c
JH
S
-o
4J g
•o S
rH rH
O C
0) C
rH q)
U «
•
43
U
•H
S
•
•S a
rl -3
S «
•P
J §
"° 5
rH JS
O
ffi
0
rl
K H
m
t-»
•
o
•
o
o
rH
<0
u;
4J
CO
"O
rH
M
•
CO
•H
&
A
CO
rH
>3 H
rH Cd
CO P4
2J
S rl
c_? oi
co >
S S
•rf
15
rt
rH
m
-------
en
in
rH
^
to
Q
W
EH
H
W
S3
CO
M
2
(O
O
B
CO
ta
m
CO
i
H
M
H
ro
IW i— I
O OH
!>1 C7*
•P c
•H -H
O W
tn
M CJ
O O
O
J>1 >H
•P PH
O
U
1
d CD
o d
•H -H
•P 2
o
s
J^
Q)
o
£l
-O
0
M
CM
MH
O
CU
e
fd
a
d
o
•H
-PX
o d
3 fd
T3 K
o
CM
H
rfl
^
0)
d
•H
2
d
•rH
rH
fd
O
o
tN
Q)
rH
•rH
-P
O
tn
J>i
^.4
,s^
U
d)
d
•H
2
j_i
fd
Q)
Z
tn
rd
i-i J>i
> d
fd 3
rH O
rd O
U
•
CM
^i
O
U
tn
o
4J
W
o
43
tn
itjj
o
•H
M-l
O
fd
CM
CO
0,
rH
•H
-P
O
to
>-)
J-(
45
o
4J
d
73
o ~
O >i
4J
J»i"H
•P d
d -H
o «
2 v-'
o
•p
•H
C ^i
Q) 4J
« d
d o
rd U
to
•
CM
O
U
QJ
t3
•rH
43
M
fd
U
d
0
•H
d -
D
^j«
0
rH
•H
•P
O
tn
>i
M
r]
O
rd
CM
§
Q) -rH
> Q
in
Q)
•H
•rH
•p
O
in
t>i
M
r]
O
•p
c
o
e
M
0)
J^
0) MH
> 0
•rH >, ^
« CD 43 (U
*~H 4-^ jQ
•P H M O
rH TO O rH
fd > 2 O
•
CM
V_i
O
U
tn
d
•rH
d
•H
2
tn
»>i
rd
2
u1
O
fd
10
vc
•
rH
•rH
4J
O
tn
!>i
IH
r]
U
*d
c
o
N
•H
a
o
o
t
tr>
MH
2
tn
O
tn
(U
43
tn
rij
d
o
in
0)
5
r-
0)
iH
•rl
4J
O
in
^i
"^
r*
U
(0
N
•H
<
IH
O
CM
rH O)
o 1
CO
cu
iH
•rH
O
in
>t
43
o
4J
d
3
o
u
Jx,
(J)
0
d
s2
fd
CM
(_i
o
CJ
tn
O O
-P U
tn
to
d
JH
CQ
%«"
(^5
d
•H
rH
O
rH
rd
O
tn
iH
ITS
(!)
>i
t'j
rJ
o
•H
>
0)
n
CM
tr>
d
•M
M
3
•a
§
•H
4J
(C
M
0)
CM
O
-P
O
-------
70
small; sovae method of concentration is therefore necessary. Much
of the solid content of the water is organic, e.g./ plant debris,
bacteria, etc., and unfortunately this is also concentrated along
with the asbestos, an ashing stage was therefore introduced into
the process to destry this organic material, which is of no interest
in the analysis. Two methods of processing such specimens have
already been described, (7,s) both of which have some undesirable
features. The method selected for these samples combines the best
features of the published techniques. The water sample is thoroughly
shaken to redisperse any solid material which may have settled out
during storage. A 200 ml sample of this is filtered through a 1000 A
pore size membrane filter. This pore size is adequately small, since
fibers even shorter than this are unlikely to follow the tortuous
path through such a inembrance. It is necessary to assist this
filtration by applying a vaccum to the clean side of the filtration
equipment. Filtration tiire is variable, depending on the cleanliness
of the water. Times varying between 15 and 90 minutes are observed;
in some cases the filtered volume is reduced, since complete blackage
occurs. Pressure filtration methods are used for some very dirty
samples. The membrane filter is then placed in a silica crucible,
covered with a silica lid and this is heated to about 450°C for
a period of more than three hours. The filter residue after ashing
is dispersed in 4 ml of freshly distilled water, and agitated gently
in an ultrasonic vibrator for 10 minutes. One ml of this suspension
is added to a specially designed centrifuge tube which contains a
carbon coated electron microscope grid. The suspension is deposited
uniformly on the electron microscope grid by operating the centrifuge
at 8300 g for a period of 30 minutes. The grid is removed from the
centrifuge tube and, after drying, is ready for inspection in the
electron microscope.
Fiber Counting
The grid is examined at a direct screen magnification of 25,000
times at an accelerating voltage of 80 kv. Using a series of
calibration circles scribed on the microscope fluorescent screen,
asbestos fibers are placed into size categories according to their
length and diameters. A minimum of 10 grid squares (200 mesh size)
is examined, and all asbestos fibers in this area are counted. One
problem in fiber counting is caused by the presence of large amounts
of other material, consisting mostly of fragmented diatoms. As a
magnification of 25,000 the area of sample examined amounts to some
40 square metres, on which fibers perhaps as small as 3 mm are being
sought among other material. This illustrates the main difficulty
of the technique — that of manual counting of very widely spaced
fibers; a method for which there is no substitute on samples such as
these. However, counting times of about 2 hours are adequate to
size the 10 grid squares required.
-------
69
Where: the No. of fibers counted = 5 (occasionally 10)
Grid area factor = 1.84 for grids with thick
bars, or
1.34 for gr:ds with thin bars
The mine tailing sample (slurry) was analy/.ed as follows:
1. The sample was shaken well to disperse the larger
solids which had settled to the bottom.
2. One ml was removed quickly with a pipette.
3. A 1:100 ml dilution was made in a volumetric flask
using glass-distilled water.
4. Ten ml of the dilution were added to approximately 100 ml
glass-distilled water previously put into the funnel of a
glass Millipore filter holder with an 0.8u pore size, 47 mm
diameter, Nuclepore filter in place.
5. The filter assembly was gently rocked to disperse the
sample.
6. The diluted sample is filtered as described in Appendix A.
The filtrate was filtered through an 0.2u pore size
Nuclepore filter. Both filters were prepared for and examined
by electron microscopy. The method of calculation is as
follows:
No. of fibers on filter x 452 X 129.4 X 1.34 XI
5 .0001
fibers/liter.
A solid sample of tailings from the Reserve Mining
Company operations on Lake Superior was requested by me.
This sample was labeled "Tailings - West Launder Delta,"
6/27/73. A portion of this sample was ground in a mortar
and the powder analyzed by X-ray diffraction by Mr. Robert
Jack in the Department of Geology and Geophysics, University
of California, Berkeley.
E. Ontario Research Foundation, Sheridan Park, Ontario, Canada:
Analysis of Water, Sediment and Air Samples for Asbestos.
Water Samples
The amounts of asbestos in water samples are usually very
-------
68
The procedure for analysis of asbestos fibers in water
differs from the analysis of asbestos fibers in air in the
following aspects.
1. The participates in water can be collected directly on
a Nuclepore filter; that is, the transfer technique from a
Millipore filter (sample collection filter) to a Nuclepore
filter (filter used for electron microscopy) as described in
Appendix A is not necessary.
2. The sample size used for analysis of fibers in water is
in the range of one liter while approximately 5000 liters of air
are sampled for analysis of fibers in air.
3. In the analysis of fibers in water, correction for a
segment removed for analysis by other methods is not necessary;
that is, the normalizing factor described in Appendix A is not
necessary.
A sequential filtration using 0.8u pore size and 0.2u pore
size Nuclepore filters was used for collection of particulates
from the water samples when it was found that many fibers were
very small and some could pass through the 0.8u pore size filter.
The sequential filtration will be used on all water samples until
sufficient data on the percent breakthrough of small fibers through
the 0.8u Nuclepore filter is obtained. The number of fibers/liter
calculated for both filters for all samples are reported as well
as their totals.
The formula for calculation of the fibers/liter in water is
from Appendix A and refers to 47 mm diameter filters:
No. of fibers counted X Total fields in grid X
No. of fields counted
Effective area of filter X Grid Area Factor X
Area of disc
1 = f/1
Vol. sampled
For a typical sample:
No. of fibers counted X 452 x 129.4 x 1
5 1 liter
fibers/liter
-------
to line on their flat side rather than on edge. There is,
however, a finite possibility that some laths will be on
edge and, due to the very small size of many of the fibers
of interest, the approximation to a square fiber will not
give more than a slightly high bias to the mass readings.
Also presented on the computer printout is the calculated
number of fiber per liter of water, the calculated mass of
fiber per liter of water, the size distribution of the fibers
based on length and width, ana the distribution of fibers
by aspect ratio. The program automatically assigns the
longest dimension of the fiber to the length and excludes all
fibers with an aspect ration below three.
Electron micrographs are being taken of the samples as
they are examined to show the typical appearance of the
amphibole fibers and the electron diffraction patterns by
which they may be identified.
D. Mr. Jack Murchio, School of Public Health, University
of California, Berkeley, California: Analysis by Electron
Microscopy of Asbestos Fibers in Water Samples
The method of analysis of the water samples is one adapted
from the method developed in this laboratory for analysis of
asbestos fibers collected from air on membrane filters. This
method is appended to this report as Appendix A (and is Appendix
A from a report entitled "Asbestos Fibers in t-he Ambient Air
of California," to the California Air Resources Board).
Modifications necessary to adapt the method to the analysis of
asbestos fibers in water will be reported in the Method Section.
the method of analysis of the mine tailings is also described
in the Method Section.
METHOD
One liter of a water sample is filtered through an 0.8 u
pore size, 47 mm diameter, Nuclepore filter. The filtration
method is described in Appendix A. A 500 ml aliquot of the
filtrate is filtered through an 0.2 u pore size, 47 mm diameter,
Nuclepore filter. The particulates on both filters are prepared
for and examined by electron microscopy as described in Appendix A.
-------
66
portions of the sample is minimal. Techniques involving transfer
of a liquid suspension directly to the electron microscope grid
are more subject to error since in the drying process there is
frequently a size separation as the meniscus of the drying drop
recedes.
The samples, after preparation as described above, were
examined on the JEM 200 Electron Microscope using an accelerating
voltage of 150 kV at a magnification of 17,OOOX on the M3 range
of the instrument. This magnification was chosen since it is
at this magnification that the intermediate lens aperture is in
focus in the specimen plane. It is thus possible, by inserting
the aperture and switching to the diffraction position, to
obtain a selected area electron diffraction (SAED) pattern of the
fiber being examined with no other adjustments to the microscope.
In this way it is possible to spot check individual fibers very
rapidly while measurements are being made on them.
A considerable amount of time was spent with a sample of
the Reserve Mining Company tailings and with Samples Numbered
AM0081A and AM0090A to familiarize the operators with the
morphological appearance and also the diffraction patterns of
the amphibole asbestiforms present in these samples. Following
this initial familiarization, the remaining samples were examined
and measured, the length and width of each fiber being recorded
while the fiber was viewed directly on the screen. Interpolation
from a screen scribed at half-centimeter intervals allows an
accuracy of measurement on the screen of approximately + 0.05 cm.
There is a calibration-factor between the magnification range on
the photomicrographic plate (which was 17,OOOX) of 0.86: That
is to say, the image as observed on the fluorescent screen is
0.86 times 17,OOOX. Measurements of the individual fibers were
read into a voice recorder to permit the operator to devote his
entire attention to the screen and to prevent any possibility
either of duplicating the count on one fiber, or of missing
fibers and of the operator losing his dark accomodation by having
to stop to write data.
The recorded data were transcribed onto punchtape and were
computer processed to give listings of the length and width of
the fibers, together with a computed mass of each fiber computed
on the basis of density times length times width squared
(D x L x W^) with the value of 3.40 taken as the mean density of
the amphibole fibers. Because many of the araphiboles are lath
shaped rather than square in cross section, this figure may
well be slightly high, since the laths will, in general, tend
-------
65
The following detailed procedure was employed to generate
the standard curve:
(1) A weighed amount of amosite was placed in a 600 ml
beaker and sufficient water containing 0.1 percent Aerosol
OT was added to make a stock suspension which would contain
4 ug/ml after the amosite was uniformly dispersed by
ultrasonic treatment.
(2) After ultrasonic treatment aliquots were taken from
the original stock suspension and further diluted with water
containing Aerosol OT to create appropriate concentrations
from which aliquots could be taken and filtered to deposit
the desired quantities on 25-mm diameter HAWP Millipore filters.
(3) The filters bearing the various concentrations of amosite
were then carbon coated and prepared for examination in the
electron microscope as described in Steps 4 through 9 in the
outlined procedure for handling liquid water samples.
C. Mr. Ian Stewart, McCrone Associates, Chicago, Illinois:
Methodology for Determination of Amphibole Asbestos in Duluth
Water Samples
The samples were received in two forms: (1) filters
which had been taken on the site; and (2) one-liter polyethylene
bottles of the water,
Our initial attempts were to use the filters taken at
the site since this would involve less suspicion of possible
loss of sample due to adherence to the walls of the container
or contamination of the sample during furhter handling. Un-
fortunately, the filters as we received them were too heavily
loaded to be usable for electron microscopy without further
preparation. We, therefore, prepared our own filter samples
by filtering small volumes (typically in the range of 30-200 ml)
of the water onto 13 mm diameter Millipore filters. A section
was then cut from the Millipore filter approximately 2-3 mm.
square and was placed facedown on a previously carbon-coated
electron microscope grid and the membrane filter dissolved
using acetone in a Soxhlet extractor. Previous work with this
technique has shown us that there is very little risk of
contamination in transfering the filter on the electron
microscope grid to the Soxhlet extractor; and, furthermore,
by dissolving the filter in situ on the grid, the risk of losing
-------
2. Filtered Water Samples
(1) The as-received sample is a membrane filter through
which three (3) liters of saupled water were filtered.
i2^ A 30° radial section is cut from the filter and thus
represents 1/12 of the residao filtered from 3 liters of
water or that residue from 250 ml of original sample.
(3) The radial section is placed in a 15 ml test tube and
the test tube in turn is placed in the low-temperature asher
for 2 days.
(4) After ashing has been completed, 10 ml of water containing
0.1 percent Aerosol OT is added to the ash residue and treated
ultrasonically to disperse the suspended ash.
(5) The ultrasonically treated suspension is poured into a
600 ml beaker, the test tube is rinsed with water containing
0.1 percent Aerosol OT into the 600 ml beaker, and the sample
volume is diulted to 250 ml with water containing 0.1 percent
Aerosol OT.
(6) The 250 ml is then subjected to ultrasonic treatment.
(7) Next aliquots (25, 50, 75, and 100 ml) are filtered through
25-nun (0.45 u) !LV.\T Millipore filters.
The sample is then processed in the same manner as is the liquid
air sample (Steps 4 through 9).
3. Generation of Standard Curve for Water Samples
A sample of UICC amosite was obtained from Dr. Leineweber of
Johns-Manville. When the sample was examined in the electron microscope,
the asbestos particles were found to be quite similar with respect to
morphology and particle size to those seen in the Duluth samples.
Consequently, this amosite sample was employed as a standard material
in the generation of the standard curve which relates counts made
in the electron microscope to micrograms of amphible type asbestos on
25-mm HAWP Millipore filter.
-------
63
B. Procedure Employed in the Analysis of Water Samples
1. Unfiltered Water Samples
(1) The as-received 1 liter water saff.ple is agitated by
rolling one hour on mixing rollers.
(2) A 250 ml aliquot is taT:en from the liter sample, 2.5 ml
of 10 percent Aerosol OT is added, and it is ultrasonically
treated in a 600 ml beaker to disperse the suspended particulate
uniformly.
(3) Aliquots (25, 50, 75, and 100 ml) are filtered through 25-mm
HAWP Millipore filters (o.45 u).
(4) The filter bearing the residue is allowed to dry and then
200 A of carbon film is vacuum deposited over the residue on the
filter.
(5) Then 3 x 3-mm squares are cut from the filter and the
filter material is dissolved away in an acetone, bath.
(6) The remaining carbon film bearing the filtered particles
is picked up on a 200 mesh electron microscope support grid.
(7) Counts are made in the electron microscope at 20,OOOX
by scanning 5 to 10 grid openings systematically. WVien a
fiber is encountered in the path of a scan one count is
registered. If a single fiber traverses the path of more than
one scan it is counted as many times as it is encountered.
This "intercept counting" approach minimizes uncertainty
arising from large variations in fiber lengths when performing
mass determinations.
(8) The average counts per grid opening obtained from the
sample are translated by means of a standard curve to micrograms
of asbestos on the 25-mm filter.
(9) Taking into consideration the aliquot which was filtered
on the 25-mm filter calculations are then made to express final
results in terms of micrograms asbestos per liter of original
sampled water.
-------
62
The modified procedure is as follows :
(1) Two 45° radial sections are cut from the nylon web
supported membrane filter.
(2) The collected particulate is removed from each 45°
radial section of the filter by ultrasonification in 100 ml
of water containing 0.1 percent Aerosol OT.
(3) The combined solution from Step 2 was duluted to 900 ml
in a clear plastic bottle.
(4) The bottle containing the 900 ml of diluted solution was
placed on mixing rollers for one (1) hour.
(5) A 250 ml aliquot was removed from the bottle and
ultrasoned for two minutes.
(6) A 10 ml aliquot was filtered onto a 25-mm HAWP
Millipore filter.
(7) After thoroughly rinsing with water the particulate
deposited on the filter was leached with 10 ml of 0.1 N
HCL and then again rinsed.
prepared for electron microscopy by coating with carbon.
(9) A 3 x 3-mm square was cut from the filter and the
filter material dissolved away to leave the carbon film
bearing the deposited particulate.
(10) The carbon film is picked up on 200 mesh electon
microscope grid and counts are made at 20,OOOX magnifi-
cation in the electron microscope.
(11) The ug/m^ of amphibole is calculated using data
obtained from standard curve and the total volume of
air sampled.
-------
61
(8) The supernatant liquid containing small suspended
particles not sedimonted by centrifugation is poured
into a 100 ml graduate.
(9) Steps 6, 7, and 8 are repeated twice to extract as
much asbestos as possible from extraneous particulate.
(10) The combined supernatant extractions in the 100 ml
graduate are then diluted to 100 ml and mixed by ultra-
sonification to create a uniform suspension.
(11) JQiquots (usually 70 and 30 ml) are filtered through
25-ml Millipore filters.
(12) The filters are cut in half and coated with a carbon
film by vacuum evaporation.
(13) A 2 x 2-mm square is cut from the carbon coated filter
bearing the deposited asbestos and the filter material is
dissolved away in an acetone bath to leave the carbon film in
which the extracted particles are supported.
(14) The carbon film is picked up on a 200 mesh electron
microscope support grid and examined in the electron
microscope at 20,OOOX where counts are made as described
in Steps 4 throngh 9 in the procedure for analyzing
liquid air samples.
2. Revised Procedure for Air Samples
When it was discovered that appreciable amounts of amphibole
filters were being lost during centrifugation an alternate method
for removing extraneous material from the sample was sought. It
was found that 0.1 normal hydrochloric acid removed sufficient
quantities of the masking material to expose the fibers to view in
the electron microscope. Consequently, a modified procedure was
developed which employed hydrochloric acid leaching as a substitute
for centrifugation. Ashing was also eliminated from the procedure.
-------
60
III. Compilation of Actual Procedures Followed by EPA Contractors
A. Procedures Employed in the Analysis of Air Samples
by Battelle Memorial Institute, Columbus, Ohio
1. General Procedures
The collection of airborne particulate were made by EPA on membrane
filters containing an insoluble web support. The web cannot be shed
in the low-temperature asher, consequently, it was necessary to remove
the collected particulate from these filters by ultrasonic treatment,
redeposit the particulate on a 0.45u white plain Millipore filter (HAWP)
and subject the redeposited material to low-temperature ashing. Then
the procedure which was developed for chrysotile was followed. The
detailed procedure which was first used in the analysis of air samples is
as follows:
(1) A 10° radial section is cut from the as-received membrane
filter bearing the collected air particulate.
(2) The collected air particulate is removed from the as-
received filter sample in 10 ml of water containing 0.1 percent
Aerosol OT by ultrasonic treatment in a 100 ml beaker.
(3) The dislodged particulate is then deposited on a 25-mm HAWP
Millipore filter.
(4) Steps 2 and 3 are repeated three times to extract as much
particulate as possible from the surface of the filter.
(5) The redeposited particulate on the Millipore filter is low-
temperature ashed for 48 hours in a 15 ml test tube.
(6) The 10 ml of water containing 0.1 percent Aerosol OT is
added to the ash in the test tube and ultrasonic treatment is
employed ot disperse the ash residue uniformly.
(7) The resulting suspension is then centrifuged at 800 G for
20 minutes.
-------
59
determined grid openings on each grid prepared are
scanned (approximately 8 openings) until at least
50 fibers ate counted and sized.
2. An upper limit of mass concentration is cal-
culated the same way as does McCrone Associates.
The same assumptions are made.
L x W^ x 3.3g/cmJ x Area - Volume =
Upper limit of Mass/Vol.
3. A lower limit of mass concentration is cal-
culated by using a third dimension factor of
0.15u. This factor is chosen because it is the
most probable value for width for these fibers,
and it is also the limit of transparency of
fibers as seen under the microscope.
L x W O.lbu x 3.3g/crr.3 x Area - Volume factor =
lower limit of Mass/Vol,
4. The upper and lower limits differ by up to a
factor of 2. These two numbers are averaged and
this average value is reported as the mass con-
concentration .
C. Battelle Memorial Institute
«**.
1. The prepared grids are placed in the electron
microscope.
2. Counts are made in the electron microscope at
20,OOOX by scanning 5 to 10 grid openings systemati-
cally. When a fiber is encountered in the path of
a scan, one count is registered. If a single fiber
intersects the path of more than one scan, it is
counted as many times as it is encountered. This
"intercept counting" approach minimizes uncertainty
arising from large variations in fiber lengths when
performing mass determinations.
3. The mass concentration is read from a standard
curve generated by weighing out different amounts
of a standard amosite sample, preparing grids and
counting in the same manner as above.
4. This mass value is then multiplied by a volume -
area factor to arrive at mass per unit volume.
-------
II. Summary of Methodology Used by EPA Contractors for the
Determination of Mass Concentration
A. McCrone Associates - calculation
1. A small amount of the water sample (30-200 ml)
is filtered through a 13nun, 0.45u millipore filter.
2. A section, 2 to 3mm square, is cut from the
filter. This section of filter is placed face
down on a carbon-coated grid and the mcmbrance
filter dissolved using acetone in a Soxhlet ex- •
tractor.
3. The samples on grids were examined on a JEM
200 Electron Microscope using an accelerating
voltage of 150KV at a magnification of 17,OOOX.
4. Approximately 50 fibrs are counted and sized.
5. Mass is calculated by the formula L x W^ x
3.4g/cm^ x 4.5 x 10^ approximately = Mass/Vol.
L = Length
W - Width (thickness assumed to be equal to
the width).
3.4 g/cm^ = Density factor
4.5 x 10^ = Area factor for grid and filter size
and for volume of water filtered.
The assumptions implicit in this are:
1. The fibers are deposited in uniform abundance on
the grid such that the number of fibers counted in
a few grid openings are representative of the number of
fibers on all grid openings.
2. The particle size distribution of those counted is
representative of the particle size distribution of the
whole sample.
B. Mt. Sinai Cancer Institute
1. Rub--out procedure followed. Fibers are trans-
ferred to the electron microscope grids and pre-
-------
F. X I2IH/L. EERKELEYxE.M.
M
O
HE3SS.
n
13
31
r
2 ?
tn
n
in
A
D
13
n
r
33
n
-------
57
I. A Statistical Analysis of the Validity and Accuracy of
Using Optical Microscopy to Estimate the Number of Asbestso-
Like Fibers in Water as Determined by Electron Microscopy
To determine the value of optical microscopy (OM) in estimating
results obtained from electron microscopy (EM) a limited series of
paried OM-EM observations were obtained. Jack Huchio, School of
Public Health, Berkeley, used SM methodology (17,000x magnification)
and DeLeon, also of the School of Public Health, employed the NIOSH
optical technique at 450x. A third set of data, an optical determina-
tion at lOOOx will be forthcoming shortly. Since each OM-EM reading
in fibers per liter was for a single sample, these data (N = 28 pairs
of observations) were subjected to least-square regression analysis.
The regression coefficient, r, was emp'.jyed as a measure of predictive
value. Perfect predictability between these estimates would be
reflected in a high r value (r Ranges from 1. to 0.0) . Good agreement
(r = .5) would imply that OM could be used as a rapid and relatively
inexpensive estimator of true fiber quantity.
Time estimates for OM vs. EM analysis time are approximately
1/2 hour vs. 1 day and cost estimates are about $15 versxis $200
respectively. Thus, if OM predicts EM with even imperfect accuracy,
the former could be used in an Agency program to screen water supplies
outside of the Duluth area.
Figure A-l shows graphically a plot of the two methods for paired
samples. The fitted line appears anomalous because many data points
fall on the x axis and are obscured by it. The correlation coefficient
for these data was .16 .and r is not statistically significant at the 1
or 5% level of confidence. It can be concluded that OM cannot be
employed to estimate EM with respect to fiber counts. OM counts, for
these conditions, bear little more than a random relationship to EM
counts.
The poor relationship between OM and EM counts is due to the size
distribution of the asbestos fibers (generally smaller than 1 micron)
and the inability of the optical microscope to resolve the vastly more
numerous smaller particles.
-------
APPENDIX I
TABLE OF CONTENTS
PAGE
I. A Statistical Analysis of the Validity and Accuracy
of Using Optical Microscopy to Estimate the Number
of Asbestos-Like Fibers in Water as Determined by
Electron Microscopy .. ......... 57
II. Summary of Methodology Used by EPA Contractors for
the Determination of Mass Concentration ....... 58
A. McCrone Associates - Calculation ......... 58
B. Mt. Sinai Cancer Institute ............ 58
C. Battelle Memorial Institute 59
III. Compilation of Actual Procedures Followed by EPA
Contractors ......................60
A. Procedures Employed in the Analysis of Air Samples
by Battelle Memorial Institute, Columbus, Ohio . . 60
B. Procedures Employed in the Analysis of Water
Samples by Battelle Memorial Institute, Columbus,
Ohio 63
C» M£ . loll SttiWdtL, I'icCX'Cuie AafeOCj-, CjiiCQCjO ,
Illinois: Methodology for Determination of Amphibole
Asbestos in Duluth Water Samples .... 65
D. Mr. Jack Murchio, School of Public Health, University
of California, Berkeley, California: Analysis by
Electron Microscopy of Asbestos Fibers in Water
Samples 67
E. Ontario Research Foundation, Sheridan Park, Ontario,
Canada: Analysis of Water, Sediment and Air Samples
for Asbestos 69
F. Mt. Sinai School of Medicine, New York, New York:
Asbestos Sample Preparation and Analysis
Methodology 73
G. DeLeon - NIOSH Method 76
H. Procedures from EMventions, Inc., Rockville, MD.
(Not Submitted) 80
-------
Appendix I
ANALYTICAL METHODOLOGY USED
BY EPA CONTRACTORS
-------
APPENDICES
-------
56
REFERENCES CITED
Burrell, S. (1973) Mineralogy of Reserve Mining Company Ores. Unpub. report.
French, Sevan M., (1968) Progressive Contact Hetamorphism of the Biwabik
Iron Formation, Mesabi Range, Minn., Minn. Geological Survey
Bulletin 45, U. of Minn. Press, Minneapolis.
Gundersen, James N. and George M. Schwartz,(1962) The Geology of the
Metamorphosed Biwabik Iron Formation, Eastern Mesabi District,
Minn., Minnesota Geological Survey Bulletin 43, Univ. of Minn.
Press, Minneapolis.
James, H. B. (1955) Zones of Regional Met amor phi. sm in the Precsmbrian
of Northern Michigan, Geological Society of America Bulletin, v. 66,
p. 1455-1488.
Marsden, R. W., Emanuelson, M. W., Owens, J. S., Walker, N.E., and R. F.
Werner (1968) The Mesabi Iron Range, Minnesota, in Ore Deposits
of the United States, 1933-1966, Bridge, J. D., ed; p. 518-537.
White, D. A. (1954) The Stratigraphy and Structure of the Mesabi Range,
Minnesota; Minnesota Geological Survey Bulletin, pp. 38, 92.
-------
35
3. No significant emissions of asbestiform minerals or
other silicates are expected from the handling of pellets
of beneficiated ore. Hovrever, the potential for silicosis
developing among workers occupationally exposed to emissions
during handling of pellets must be considered.
4. The taconite mine by Reserve Mining Company, and occurring
in Minnesota on the Mesabi Range, is mineralogically similar
to that occurring in Wisconsin, Michigan, and Wyoming. There
are at least five plants in Michigan, one in Wisconsin, and
one in Wyoming about which air and water pollution are expected.
None of the plants have a capacity greater than 35% of that
of Reserve Mining Company's plant, but a health hazard may exist,
and these sites should have high priority in any future sampling
programs.
5. Other activities expected to emit high concentrations of
asbestiform fibers are numerous. Many, many mines operate in
rocks containing amphibole minerals which are produced as a by-
product. Sampling at selected sites should be done in the future.
Soils formed from rocks containing asbestiform minerals are
widespread and are expected to contain those minerals themselves.
The occupational and environmental exposure levels in all of the
above cases are unknown. An effort of a magnitxide similar to
that dedicated to determining the ambient levels of asbestos
in air in urban areas must be made if we are to accurately determine
the levels of exposure in each of the above cases.
-------
54
Soils containing asbestiform minerals are a largely
uninvestigated source of potentially hazardous air pollution in
the United States. The medical literature reports that, in
Bulgaria, soil near a serpentine mine contained chrysotile and
amphiboles. Farmers working the land in this area have been
exposed to fibers found in the soil and a significant number were
found to have heavily loaded lung tissue as a result.
That soils are locally significant sources of asbestiform
minerals must be considered in selecting control populations for use
in epidemiological studies. In northeastern Minnesota, for
example, igneous and metamorphic rocks containing amphiboles and
volcanic rocks containing some serpentine are exposed at the surface.
Background concentrations of asbestiform minerals there are expected
to be significantly higher than in those parts of the state where
tills of sedimentary virgin or sedimentary rocks themselves are
exposed at the surface.
H. Conclusions
1. No other company currently mining taconite along the
Mesabi Range handles as much amosite annually as does Reserve
Mining Company. Reserve handles 2.5 times as much amosite
as does the next largest producer.
2. Amosite is not the only asbestiform mineral present
in emissions from Reserve Mining Company; crocidolite and
actinolite are axpect-.ed to occur in small quantities as
well. Additional minerals not now classed as asbestiform/
and whose potential as a health hazard are unknown,
are: hornblende, minnesotaite, greenalite, and talc.
Respiratory diseases are known to be prevalent among workers
in talc mills with fugitive dust. When the contributions
of these minerals to the emissions are considered, two plants
are expecting to emit a larger tonnage of potentially hazardous
fibers annually than does Reserve Mining Company. Those two
companies are: Erie Mining Company and U. S. Steel's Minntac
plant. The other three companies on the Mesabi Range all emit
about half the tonnage of potentially harmful fibers that
Reserve Mining Company does. Even these emissions must be
viewed with alarm, however, if there is no threshold
concentration associated with the incidence of mesothelioma,
or other serious diseases resulting from exposure to asbesti-
form minerals. Air emissions from all these plants should
have priority in any future sampling program on the Iron Range.
-------
Identifi cation
Identification of chrysotile asbestos fibers gives risa to
little difficulty. The fiber itself has a characteristic tubular
appearance (Figures 1, 2), and in moct cases (depending on previous
history) such fibers give an equally characteristic electron
diffraction pattern (Figure 3) . Thr.r, pattern can be compared with
that obtained from specimens of known chrysotile. The appearance
and single-fiber diffraction patterns of the other varieties are
subtly different. A standard of Amosite is shewn in Figure 4, and
its single-fiber diffraction pattern in Figure 5. The tubular
morphology of chrysotile is not present, and the diffraction
pattern is more regular than that of chrysotile. Crocidolite is
shown in Figure 6. The morphology is similar in this size range to
that of chrysotile, i.e., very long thin fibers, though it does no
have the tubular appearance. The single-fiber electron diffraction
pattern (Figure 7) shows fibrous morphology, and is quite different
in appearance from that of chrysotile. Tremolite is shown in Figure
8, and its single-fiber diffraction pattern in Figure 9.
Whilst it is not economically to identify the asbestos type
of every fiber by electron diffraction in all samples; the technique
can be used whenever doubt exists through observation of the fibrous
morphology alone.
Standard Samples
Plank samples of distilled water have Leea processed in order
to demonstrate that asbestos is not being introduced as a result of
the analytical procedures. It is very easy to think of possible
sources, such as airbore material, or contamination form furnace
insulation, etc. In these blank samples 100 ml of distilled water
were filtered and the samples processed in the same way as for the
other specimens. In one of these samples, no fibers were detected,
and in the other one only 3 fibers.
In further standardization experiments, a known suspension of
fibers in the size range of interest was added to membrane filters
prior to ashing. These samples indicated that the method was
satisfactory, giving confidence in the results obtained on the
actual samples.
Particle Size Considerations
The techniques used have been developed on the premise that
the fiber lengths are in the size range of the transmission electron
microscope, i.e., .Olum up to about 20 urn. Samples so far examined
have confirmed this, though in the case of sediment samples and some
air samples an optical examination is also necessary.
-------
72
Levels of Detection (Water Sampl es_)_
The level of detection for fiber concentrations in this type of
measurement corresponds to the detection of one fiber in the area
counted. This is equivalent to about lo'* fibers p?r liter. At the
fiber Lengths observed, this figure corresponds to about 2 X 10
micrograms per liter, or 1 part in 5.10-1-' bv weight.
The lowest level of detection using this method depends to a
large extent on the time which is spent on counting particles. This
time can be reduced if the samples were initially concentrated to a
greater extent, e.g, by filtering more than 200 ml of the original
specimen. Unfortunately, the amount of other material in the water
dictates a limit to the degree of concentration possible. Above
this, the final electron microscope sample grid becomes so loaded
with other residues that most of the area is obscured and anything
approaching an accurate fiber count is not possible. This limit-
varies widely between water samples from different locations, and in
several samples it has even been necessary to allow heavy material
to settle before final preparation of the microscope grid from the
ashed residues. In other samples it has been only possible to force
100 ml of water through the membrane filter before complete blockage
took place. In some cases all of the ashed residues were of small
particle size, and here the only recourse was to prepare an electron
microscope specimen with a reduced proportion of ashed residues, and
to spend more time counting the fibers.
It is most important to realize that many diatomaceous fragments
appear at low magnifications to be linear and fibrous. Diffraction
patterns and the morphology under higher magnifications are the only
means of rejecting such artifacts from the fiber count.
Calculation and Presentation of Results
The fiber counts and measurements obtained on each sample are
processed by a specially written computer program, with results
being expressed in appropriate units for air, water or sediment
samples. Typical computer outputs for the three cases are attached.
The mass concentrations shown are determined by assuming the X-ray
density for the principal type of asbestos present.
A typical number - fiber length distribution is also shown.
On this cumulative curve it can be seen that above a particular fiber
length, no fibers were counted. Fibers above this upper limit of
length correspond to only a few percent of the total number counted,
-------
73
so, the error in the total number reported is very small. However,
those few fibers above the upper limit of size contribute progressively
more to the total mass present. For example, because of their low
numbers, very large fibers at the 104 per liter concentration level
would not be found in these measurements, but they could contribute
significantly to the mass concentration. The measured number-size
distribution would allow cloner estimates to be made of the mass
concentrations, but this d r> c, procedure which requires a considerable
amount of computation.
F. Mt. Sinai School of Medicine, New York, New York: .Asbestos
Sample Preparation and Analysis Methodology
Ashing
Samples collected on rnembrance filters are prepared for ashing
by cutting a I cm square of the filter and placing it, dust side down,
on a clean microscope side. Two or rhree drops of acetone are added
to the square to partially dissolve the filter and to fix the material
to the glass slide. The sample is ashed in a low temperature activated
oxygen asher for 10 minutes to one hour, depending on sample composition.
(The cleaner samples will require the longer ashing period.) It is
best to stop the ashing before complet combustion of the filter to
minimize any loss of sample mineral material.
Rubout Procedure
Following ashing, the samples are dispersed in a nitrocellulose
film and mounted on formvar coated electron microscope gruids. The
dispersal is accomplished by a "rubout" technique in which asbestos
fibers are broken into individual fibrils. This procedure allows
positive identification of chrysotile asbestos to be made on the
basis of morphology alone. Moreover, large agglomerates of mineral
particles, which could obscure the presence of asbestos fibers of
fibrils, and to which asbestos fibrils could be attached, are broken
into particles sufficiently small to allow all asbestos fibrils to
be seen.
The rubout is accomplished by placing a large drop of 1% solution
of nitrocellulose in amyl acetate on the ashed residue. The flat
surface near the edge of a clean watch glass is used to grind the
residue into submicron sized particles. This grinding usually takes
from 5 to 10 minues, depending on the amount of residue present.
Typically, the amyl acetate will evaporate in less than 5 minutes
and additional drops of pur amyl acetate must be added to complete
the rubout. (Adding the 1% nitrocellulose solution will result in
an undesirably thick film.) During the final stages of the rubout,
-------
the solution and residue arc dispersed over several cm of the microscope
slide. After dispersing the ashed residue, some of which still remains
on the watch glass, another urop of the nitrocellulose solution is
placed on a second clean slide. Duriny the dispersal of this drop
over several cm of the slide, the major portion of the residue remaining
on the watch glass is removed. By this procedure, less than 10% of
the residue will remain on the watch glass.
The two slides are placed in contact and the ground residue and
nitrocellulose solution further dispersed. The residue typically
is spread over the width of the slide for a distance of from 5 to
7 cm. The two slides are thn pulled apart. No pressure need be
exerted during this procedure as surface tension will hold them in
contact.
With practive , a uniform film is thus produced which can be tested
by viewing against a light. (We earlier did particle counts to verify
uniformity, but visual appearance proved to be as reliable an indicator
of uniformity.) If the film is not uniform, amyl acetate can be added
and the above procedure repeated.
Mounting
The edges of the slide with the attached film are scrap ed with
a scalpel blade. By carefully dipping the slide in water at an angle
of from 30° - 50°, the nitrocellulose film can be floated onto the water.
Several procedures can b
-------
75
Electron Microscope Scanning and_ Coyr.i.i nc;
Typically, 8 grids are prepared from each sample ashed using
material from widely separated portions of the prepared films
from both slides. A single square of each grid is scanned at
40,000 X magnification to determine thn quantity of chrysotile
present. The identification of chrysotile is on the basis of
morphology, either that of the classic tubular structure or that
of altered chrysotile (as a result of either beam damage or physical
damage prior to collection). Here the structure exhibits a dense,
irregular, inner region, sometimes with a thin capillary, and an
electron transparent irregular outer region. To gain experience in
recognizing various forms of chrysotile, one can select an unaltered
fiber and watch it deform in the electron beam. The use of carbon
coated films can reduce this damage, but in practice this procedure
has not been found necessary.
The length and diameter of each fiber are estimated with the
aid of fiducial marks on the viewing screen and the mass of chrysotile
per grid square determined. The total area of film prepared is used
to calculate a dilution factor from the rubout. (Typically, 1 cm of
membrance filter sample is dispersed in 25 cm^ of film.) Knowing the
air volume passed through a given area of filter paper, the concentration
of chrysotile asbestos in the sampled air can be obtained.
In practice, a reasonable statistical variation exists between
the number of fibrils found on different grid squares. The variation
exceeds that expected on pure statistical grounds as occasionally
clumps of fibrils, resulting from the incomplete dispersal of a
fiber bundle, are seen. However, greater variation occurs in the
volumes seen on different grid squares as one large fibril can contribute
100 times the volume of a small one, hence, the need for counting
several grid squares. Eight lOOu X 100 U squares have proved to
be sufficient. The inaccuracies inherent in the sample preparation
do not warrant additional effort in scanning.
Calibration
The above procedures are calibrated by processing filters prepared
with known amount of chrysotile. Here both clean membrane filters
and filters previously used to collect ambient air samples are used.
Triple air jet milled chrysotile is dispersed in f^O to which Aerosol
OT has been added. The solution is subject to ultrasonic energy and
diluted so as to produce a concentration of about 1 nanogram of
chrysotile per cc. This solution is filtered through the membrance
filter until 10 to 50 ngms of chrysotile is collected per cm^ of filter.
The above procedure is then followed as in the case of collected air
samples. Recoveries range, typically, from 30% to 50% of the added
chrysotile and this factor is applied to the data from the processed samples,
-------
76
Clean Room Procedures
It has been found that strict adherence to clean room procedures
must be followed. All sample processing should take place in a
filtered air environment. The water used should be filtered thorough
0.22u membrane filters. Blank controls must be run routinely to
assure absence of contamination.
G. Del eon - NIOSH Method
Introduction
In the early epidemiologic studies of the asbestos industry by
Fulton (1935) and Dreesen (1938) , the impinger was settled upon as
the standard method for appraising dustiness. Since few fibers were
seen it was necessary to count both grains and fibers to obtain a
statistically useful count depsite the assumption that asbestosis
was caused by fibers. This inconsistency notwithstanding, dust control
based on impinger results has greatly reduced the incidence of
asbestosis where it has been conscientiously applied in the asbestos
textile industry. However, a number of investigators (Lynch and
Smith, 1955; Doll, 1955) have suggested other health hazards associated
with asbestos, and a more relevant and sensititve method is required
for estimation of the hazard, especially in industries using only
part asbestos in their product.
A further weakness in the impinger method is that it is not
amenable to long personal samples. It is believed that personal
samples best approximate the worker's actual exposure, and Roach (1966)
has recommended a sample duration of ten work shifts for mineral dusts.
New methods were therefore needed, and among those studied were
fiber counts at 430X phase contrast on membrane filters that had
been made transparent. This paper deals with the development and
standardization of this method.
The principal variable which needed to be considered was the
mounting procedure in relation to maximum transparency and the
effect of time on the dimensional stability of the mount.
Methods using immersion oil, airplace glue or acetone were found
to be either too cumbersome or resulted in an imperfectly cleared
or mottled filter, particularly as viewed with phase contrast optics.
In methods which involved the application of liquid over the dust
deposit the movement of the meniscus across the deposit sometimes
caused redistribution of the dust. The use of a 50-50 mixture of
diethyl oxalate-dimethyl phtalate was attempted and found to result
-------
77
in satisfactory clearance. However, since this medixim dissolved
the filter, the migration of particles resulted in changing counts
with time and needed to be examined to evaluate the method.
Sampling Procedure
(R)
Samples for count are collected on Killipore type AA filters
by personal samplers operated by battery powered pumps worm by the
employees, and as simultaneous samples collected in conjunction with
impingers. The filters are contained in plastic filter holders and
are supported on thick pads which also aid in controlling the distribu-
tion of air through the filter. Several methods of admitting air to
the filter have been attempted. With the face cap removed and the
filter completely exposed, the air may enter the filter evenly, but
large high velocity particles and foreign objects may also hit the
filter and possibly cause damage. As an alternative, the face cap
may be left in place and the small plug removed. The passage of air
through this 4 mm opening at a flow rate of 2 liters per mirute produces
a velocity of 265 cm/sec, which yields an impaction parameter of
0.081 for 1 micron (u) unit density spheres. Although this impaction
parameter would predict a very low collection efficiency of 1 u
particles by impaction, some larger particles would tend to be impacted
on the center of the filter. This was found not to affect count
distribution, which is dominated by small particles; however, it does
result in obscuring a small portion of the filter, particularly if
many large particles are present, as in friction product plants. The
third alternative was to drill six additional 4 mm holes in some face
caps and to fit these perforated caps to the filter holder during
sampling. This reduced the velocity to 38 on/sec thus eliminating
the impaction problem' while still protecting the filter from damage.
Mounting the Sample
The mounting medium used in this method is prepared by dissovling
0.1 gm of membrane filter material per ml of a 1:1 solution of
dimethyl phthalate and diethyl oxalate. The exact proportions of
the 3 components are not critical, but the medium must have as high
a viscosity as possible without being difficult to handle. The index
of refraction of the medium thus prepared is ND = 1.47.
To prepare a sample for microscopic examination, a drop
of the mounting medium is placed on a freshly cleaned standard
(25 mm x 75 mm) microscope slide, using a dropper or applicator.
The volume of the drop is approximately 0.05 ml. A wedge-shaped
piece about 1 cm by 2 cm is excised from the filter using
a scalpal and forceps, and placed duet side up on the
-------
78
drop of mounting solution (Figure 1). A #1-1/2 cover-slip
carefully cleaned with lens tissue ic placed over the filter
wedge. Slight pressure on the covers!ip achieves contact between
it and the mounting mediarn.
Clearing of the filter with this method is slow, requiring
about 15 minutes. The r ample may be examined as soon as the
mount is transparent.
The transition of the dust-supporting substrate .from
translucent, solid membrane filter cellulose ester material to
a transparent, optically homogeneous gel is very gentle, and
at no time does a meniscus sweep across the dust deposit.
Rather, the substrate slowly softens beneath the dust, and there
is no opportunity for washing away or redistribtuion of the dust.
The optical homogeneity of the resulting mount is nearly
perfect, with only a slight background granularity under phase
contrast, which disappears within 1 or 2 days.
Stability
Observers noted a reduction with time in apparent con-
centration for samples mounted with only dimethyl phthalate
and diethyl oxalate. By observation, this loss was due to the
migration of the dust particles outward from the center through
the gelatinous ruo.tGjTj.al of tli£ fiA^u mount, with & ciecreasy in
aread concentration. .To determine the magnitude of this loss,
a group of samples was mounted using 1:1 dimethyl phthalate
and diethyl oxalate only, and counted at one day to three-day
intervals for periods ranging up to 76 days. Counts were displayed
as fractions in percent of the original, or first day counts
(Figure 2). At the end of one week, samples mounted by this
method yielded only 75% of their original values. By 30 days
after mounting, the counts using this method had declined to
40% of thier first day value, and continued at this level with no
further decline until counting was terminated 76 days after
mounting.
A group of samples mounted in the method described in the
section "Mounting the Sample" using high-viscosity mounting
medium were counted at one-day intervals for eleven days. No
loss of areal concentration was apparent over this short period
and a second group was similarly mounted and counted for a
period of fourty-six days. The composite of these two series
in shown in Figure 2. No loss of concentration was detectable
during the first 30 days, but the count declined to 85% of its
original value by the 46th day, when the series was terminated.
-------
79
A 6 mm square of membrane filter with a vaccurn-deposited
film of carbon was mounted in the manner described to provde
a macroscopic observation of the distortion of the filter areas.
After a few days, the square began to disintegrate at the edges,
but the major portion shoved no discontinuity during the initial
30 day period. By the end of 62 days, most of the square had
disintegrated and the fragments migrated over an area 16 mm by 17
mm.
Evaluation
The filter samples mounted in the manner previously described
are evaluated in terms of grain and fiber concentration, and fiber
size. Binocular research microscopes equipped v/ith fixed mechanical
stages have been used. Phase contrast optics fitted to these
instruments include an Abbe condenser with roatating phase turret
and a 4 mm "high-dry" achromatic objective used for all membrane
filter determinations. In all cases 10X Huygens eyepieces, one
of which contained a Forton reticle at the level of the field-limiting
diaphragm, were used. The left half of the Porton reticle field
served to define the counting area or field.
A ribbon filament illuminator was cirtically adjusted to
provide Kohler illumination through a clear blue filter.
Twenty fields located at random on the sample were counted
and total grains, total fibcrc, fibers greater than 5 u, and
fibers greater than 10 u were recorded. Any particle having an
aspect ratio of 3 or greater was considered a fiber.
Recording of Data
Since it was recognized early in the study of the asbestos
products industry that the large number of samples to be collected
would present a burdensome data processing problem, the data were
collected on forms designed for transcription to punch cards (Figure
3). Computer programs were then written to calculate dust
concentrations and to produce statistical summaries consisting of
mean, standard deviation, median, etc., for any group of samples
by plant, operation, job, type, etc. These programs, written in
Automath (Fortran II) are applicable to any problem using the sample
and count forms as source documents. Additional analyses possible
are correlation and regression of simultaneous samples and Poisson
index of dispersion of field counts for various counters.
-------
80
Summary
A method for sampling and counting asbestos dust using
membrane filters has been developed and evaluated. The mounting
medium developed clears the filter gently so as not to
redistribute the dust deposit, and produces a mount with adequate
dimensional stability for a period of thirty days. This method
has been used in developing environmental data in connection
with the Public Health Service study of the asbestos products
industry, and is recommended for research studies in other dusty
trades although not for application to Threshold Limit Values.
A system of data handling involving electronic computer processing,
which is also applicable to other studies, was discussed.
References
Doll, R. Mortality from Lung Cancer in Asbestos Workers.
British Journal of Industrial Medicine, 12; 81-85 (1955) .
Dreesen, W. C., J. M. DallaValle, T. I. Edwards, J. M.
Miller, and R. R. Sayers. A Study of Asbestos in the Asbestos
Textile Industry. Public Health Bulletin No. 241, u. S.
Government Printing Office, Washington, D. C. (1938).
Fulton, W. B., A. Dooley, J. L. Matthews, and R. L. Houtz.
Asbestosis. Pennsylvania Department of Labor and Industry,
Special Bulletin 42 (1935).
Lynch, S. A. A More Rational Basis for Air Sampling
Programmes. Am. Ind. Hyg. Assn. J., 27, 1 (Jan.-Feb. 1966).
Note
(R)
Millipore filters are named in this paper since the
method was developed using this brand of filters. However,
this does not imply endorsement by the United Stated Department
of Health, Education, and Welfare or the Public Health Service.
Similar techniques may be developed for other filter brands.
H. Procedures from EmVentions, Inc., Rockville, Maryland
(Not submitted)
-------
81
APPENDIX II
ANALYTICAL RESULTS REPORTED THROUGH SEPTEMBER,
1973, FOR WATER AND AIR SAMPLES*
*Due to its bulk, APPENDIX II can
be seen on file in the Office of
Technical Analysis , EPA.
-------
APPENDIX III
STATISTICAL ANALYSIS OF THE
ANALYTICAL RESULTS
-------
APPENDIX III
TABLE OF CONTENTS
Page
A. Replication Reliability and the Validity of Ordering
of Numeric Assignments for Selected Battelle Air Data ^2
B. An Examination of Inter-operator (contractor) variability
in Electron Microscope Fiber Counts on Replicates of the
Same Sample 84
-------
82
A. Replication Re3.lability end the VcJ 'ditv_ci Ordering of Numeric
Assignments for Selected^crtrirell ^ /.i.r_JDpty.
Subsets of paired air samples analysed by Battelle for fiber
concentrations were statistically examined to obtain answers to the
following general questions:
(1) What kind of relative variability exists between
analytical results obtained for suVsamples taken from a single
synoptic sample replicate (i.e., subdivisions of two filter pads
from a single location.}
(2) What kind of absolute variability exists between
analytical results obtained for subsanples taken from a single
synoptic sample replicate (i.e., subdivisions of two filter
pads from a single location.)
These questions bear on the reliability and replication-
validity of a sample value. Both parametric and non-parametric statis-
tical tests were performed to obtain estimates cf this variability.
A rank correlation analysis was performed (on split samples)
to determine if the ordering of quantitative values for a given
analytical run was approximately constant for two replicates of a
single sample. This test was run for two data sets and the results
are given below in Tables 1 and 2. The correlation coefficients, Rho,
for these two runs were .86 and .66, respectively, and are indicative
of good relative agreement between the order of assigned values on two
separate replicates. Thus,, the analytical assignments are the some
regardless of when the analyses are performed.
With respect to the absolute numerical values assigned to sub-
samples, data for selected samples were subjected to a paired "t" test.
This test asked the question, "Are numeric values assigned to a. sample
constant when analyses are performed on two separate subsamples?" Results
of this analysis are given below in Talbe 1 and Table 2:
-------
Table 1:
83
Sample
1
2
3
4
5
7
8
9
10
First RAT- i Battellc Air Samples
Subsample 1 Subsample 2
.09
.03
.06
.06
.01
.07
.02
.02
.02
(1.0)
(5.0)
(3,5)
(3.5)
(9.0)
(2.0)
(7.0)
(7.0)
(7.0)
(Rhc)
.19
.03
.07
.01
.02
.07
.02
.01
.02
Results of Rank Correlation
r^ .663
»*« good serial agreement
Results of Paired "t" Test
t - .508 with Sdf
• %. no significant difference between
in Subsample 1 and Subsample 2.
Rank
(1.0)
(4.0)
(2.5)
(8.5)
(6.0)
(2.5)
(6.0)
(8.5)
(6.0)
values
Table 2:
Rerun: Battelle Air Samples
Sample //
1
2
3
4
5
6
7
8
9
10
Subsample 1
-------
84
From these results it is possible to conclude that both the
relative and absolute analytical results obtained for air subsamples at
a single point are in good acjreement. Thus, the data can be used to
estimate relative, and possibly, absolute levels, of fiber mass within a
given area. The results of these analyses do not speak to the accuracy
of the measurement or to the size, type, or chemistry of the fiber mass
present. These analyses do not reflect the errors stated by Battelle
to invalidate the first set of air data analyzed (10 samples) nor to
validate the second analyses (of 31 samples) said by Battelle to be valid.
B. An examination of inter-operator (contractor) variability in
Electron Microscopic fiber counts on_ rejplicates of the
saiae sample
For the Duluth asbestos study, the various Government contractors
(McCrone, Inc.; Jack Murchio, The School of Public Health, University of
California, Berkeley) were supplied split samples of water from a given
location and McCrone received filter pads of material filtered in the
field as well. These were analyzed using electron microscopes (EM)
and with the data being both the number of fibers/liter and the length
and width distribution of those fibers. Of the samples analyzed to
date, only 13 samples have been completed by both contractors for all
parameters to be estimated.
The data on fibers/liter from these 13 samples have been
subjected to two-way analysis of varianca. This test permits the
estimation of the statistical significance due to two factors:
contractor effect and sample effect. The questions asked of the data
were, "are there significant differences in counts produced by the
various contractors,, and are there significant differences da« to samples?:
Significant differences between contractors could possibly mean that they
are counting asbestos in different ways but comparisons would be invalid.
Significant differences between samples are to be expected a_ priori since
the number of fibers per unit volumes should vary from sample to sample.
It is cautioned that these are preliminary and partial data so that the
following results must be accepted as tentative, pending completion of
all samples.
Raw data are given in Table 1. Results of this analysis are
given in Table 2. The F value for columns (= contractor) is not
significant while the row variation (= between individual samples) is
significant at the 5% level. These results are encouraging since they
suggest that two independent analyses of the same sample produce similar
results. The means of the data from the two operators, 2.3 X 10^ for
McCrone and 1.4 X 10^ F/l for Murchio, are somewhat different but the
difference is not statistically significant.
-------
85
Table 1
EM-Water Sample Results
SOURCE
S ample
1
2
3
4
5
6
7
8
9
10
11
12
13
Me Crone
4.37*
- 5.29
2.76
6.08
2.26
1.47
2.53
0.25
3.87
0.67
0.71
. 0.39
0.46
Murchio
2.0
3.0
5.0
2.0
1.0
1.0
2.0
0
0.7
1.0
0.1
0.3
0.3
* 6
Fibers/liter X 10
-------
-------
86
Table 2
Analysis of Varionce Table
Source df SS MS
Total
Columns
Rows
Error
25
1
12
12
77.092
6.213
54.472
16.407
3.084
' 6.213
4.539
1.367
4.545 NS
*
3.320
NS = Not Significant
* = Significant at 5% Level
------- | |