Report
SMOKE CURVE CALIBRATION
March 1969
New York University
> .
SchooJ of Engineering and Science
University Heights, New York, N.Y. 16453
Research ^Division
-------
Report
Smoke Curve Calibration
PHS Contract PH-86-68-66
Prepared by
William T. Ingram
Project Director
Research Division
New York University
School of Engineering and Science
March 1969
-------
Table of Contents
List of Tables. . .
. . . .
. . .
.......
. . .
List of Figures
List of Exhibits
. . .
. . . . . .
. . . .
. . . . . .
. . .
. " . .
. . . .
. . .
. . . . .
Objectives. . .
Summary
. . . . .
. . . .
. . .
. . . .
. . .
. . .
. . . . . .
. . . .
. . .
.......
Particulate Measurement Systems
Experimental Procedure. .
. . . . .
. . . . . .
.' . .
. . . . .
, . .
. . .
Results
.......
. . .
..........
. . .
Development of European vs. United States Reporting
Unit Relationship. . . . . . . . . . . . . . . . . .
Relationship Between Particulate Density and Particle
C oun t . . . . . . . . . . . . . . . . . . . . . . . .
Summary Discussion.
Acknowledgments
. . .
. . . . . .
. . . .
. . . .
. . .
. . . . .
........
. . .
References. . . . .
. . . . .
............
Appendix A - Apparatus Used to Determine the Form of
the Smoke Calibration Curve. . . . . .
Appendix B - Standard Operating Procedure for
Photovo1t Ref1ectometer Model 610
. . .
Appendix C - Apparatus Used to Determine Scale to be
Attached to the Calibration Curve. . .
Appendix D - Instruments and Procedures for Comparison
of Light Transmission and Reflection. .
iie
Page
iii
iv
v
I
I
4
7
25
37
52
67
68
70
A-I
B-1
C-1
D-1
-------
Table No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
List of Tables
Title
Pump Operation During 15 Minute Cycle
Normalization of Arbitrary Units. . . . .
NYU (Bronx) Station 1966 and 1968 Values
in Arbitrary Units. . . . . . . . . . .
Christ odor a (Manhattan) Station 1968
Values in Arbitrary Units. . . . . . . .
Comparison of Reflection Readings - Eel
and Photovolt Reflectometers ......
Comparison International Smoke Calibration
Curve and Proposed New York City Smoke
Calibration Curve in ~g/cm2 . . . . . . .
Comparison International-Great Britain
Smoke Calibration Curve and Proposed New
York3City Smoke Calibration Curve in
~g/m . . . . . . . . . . . . . . . . . .
Percent Transmission vs. Percent
Reflection Values. . . . . . . . . . . .
Frequency Distribution Coh/1000' Average
of Two 1 Hour Readings. . . . . . . . .
Frequency Distribution Coh/1000' One 2
Hour Reading. . . . . . . . . . . . . .
Conversion Table Coh/1000' vs. ~g/m3
Related Spot Evaluation and Particle Count
Linear Regression Analysis - Coh/1000 Ft.
vs. Particle Count. . . . . . . . . . .
Calculated Values - Coh/1000' and RUD
iii.
Page
14
23
26
28
31
36
39
42
48
49
53
57
58
61
-------
Figure No'.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Lis t of Figures
Title
Surface Concentration
Reflection. . . . .
Surface Concentration
Transmission. . . .
IJ.g/m3 vs. %
. . . . .
. . . . .
Coh/1000' vs. %
. . . . .
. . . . .
Location of Sampling Sites. . . . . . .
Apparatus for Test Program. . . . . . .
Apparatus to Determine Scale of Smoke
Curve. . . . . . . . . . . . . . . . .
NYU Bronx Station Smoke Curve in
Arbitrary Units. . . . . . . . . . . .
NYU Christodora Station Smoke Curve in
Arbitrary Units. . . . . . . . .
Mean - International - Bronx - Manhattan
Smoke Calibration Curves in Arbitrary
Units. . . . . . . . . . . . . . . . .
Proposed ~ew York City Calibration Curve
in ~g / em ...............
Propo~ed Standard Calibration Curve in
~g / m .................
Comparison % Transmission vs. %
Reflection. . . . . . . . . . . . . . .
Relationship 2 Hour Readings vs. 1 Hour
Readings - Coh/1000' . . . . . . . . . .
Conversion Curve - IJ.g/m3 to Coh/1000
Linear Feet. . . . . . . . . . . . . .
Regression Lines for Coh/1000 Feet and
Particle Count - Particle Size Range:
0.3-1.0 IJ., 1.0-2.0 IJ. . . . . . . . . . .
Regression Lines for Coh/1000 Feet and
Particle Count - Particle Size Range:
2.0-6.0 IJ., 6.0-30.0 IJ. . . . . . . . . .
Comparison of Values of Coh/1000 Feet as
a Function of Evaluation Technique. . .
iv.
Page
9
10
13
16
20
27
29
32
35
38
43
50
51
59
60
62
-------
Exhibit No.
1
2
3
List of Exhibits
Title
General View of Test Apparatus ....,
Illustration of Sequence of Flow Shown
in Figure 4 .. . . . . . . . . . . . .
Hi Volume Sampling System with Meter
and Pump. . . . . . . . . . . . . . . .
v.
Page
17
18
21
-------
REPORT
SMOKE CURVE CALIBRATION
Obiective
The original goals of this project were to establish
the shape of a smoke curve for fine suspended particulates in
New York City comparable to the International Smoke Curve used
in British and other European practice, and to relate two com-
monly used reporting units of surface concentration, ~g/m3
(European practice) and Coh/1000 lineal feet (United States
practice), to each other. An additional goal was developed
during the project which was to examine the possible relation-
ship between size of particle and particulate density as mea-
sured in United States practice. While not directly a goal of
the project, it was essential in the course of work to study
the relationship between particulate density readings taken
for one hour and two hour sampling periods. Progress toward
achievement of these goals is presented in the following pages.
Summary
This report reviews briefly the background and develop-
ment of particulate density measurement systems in the United
States and Great Britain and discusses the difference in re-
porting units which has made it impossible, from routine
reporting, to make comparisons of particulate density concen-
trations in the atmosphere between locations in the United
States, using either gravimetric measurement or light reflec-
tion or transmission techniques, and locations in Great Britain
or other European countries using the International Smoke Curve
for estimating concentration. Procedures for the establishment
of a relationship between United States reporting units (Coh/
1000 ft.) and British equivalent units (~g/m3), each expressing
-------
2.
a unit measure of concentration of particulates in atmosphere
but neither expressing a true unit weight concentration of sus-
pended particles, have been developed and described in detail.
A smoke curve for New York City has been developed at
two locations, one representing a heavily populated area with
industrial, commercial and vehicular influence on particle
emissions, the other representing urban residential living with
less obvious industrial, vehicular and commercial contributions.
The relationship between curves for these sites and between
curves derived for New York and the International Smoke Curve
is shown to be of basically the same form between 0 and 45
darkness index. It is further shown that the New York City
curve can be normalized to 30 ~g/cm2 at a darkness index of 25.
This positioning compares with l7~g/cm2 (England) and 20 ~g/
cm2 (International Smoke Curve). In effect then the surface
concentration of particulates required to produce the same
darkness index is higher in New York than in either England,
France or Holland, but in other regards the New York curve is
one of the family of curves common to all. The final form of
the New York City Standard Smoke Curve has been taken as:
~g/cm2 = 1.00 (Darkness Index) + .0079 (DI)2
(13)
That expression has been converted into a concentration equa-
tion standardized at an air volume of 2 M3 passing through the
filter as follows:
~g/m3 = 5.06 xv~g/cm2
(15)
Data are presented that show that a one hour Coh/lOOO'
reading alone cannot be representative of a two hour Coh/lOOO'
reading. The calculated line of best fit is:
Y = - 0.14 + 1.31 X
(23)
where
Y = average of two one hour values, Coh/lOOO'
X = one two hour value, Coh/lOOO'
-------
3.
Data are presented to show that the
tance value as measured by RAC spot
tance value measured on a Photovolt
relationship of transmit-
evaluation, and the reflec-
Reflectometer was
y = 2.71 + 1.008 X
(21)
where y = % Transmittance
X = % Reflectance
A graphical presentation is shown that will permit the
conversion of either United States or European data by entering
with a measured darkness index and reading the equivalent con-
centration in ~g/m3. A standard curve is also provided for
New York City whereby entry can be made in Coh/IOOO fto and
read as ~g/m3, thus providing an easy and direct reading that
can be used in comparison of particulate density concentrations
measured in the European system.
An exploratory study was made for the purpose of inves-
tigating possible relationships between particle count differ-
entiated by size range and particulate density in terms of
light transmittance and/or light reflectance. A series of
curves is presented to show that relationship. The concentra-
tion is greatest and most significant when the Coh/IOOO' value
is related to total particle count in the size range 1.0-2.0 ~.
The relationship appears to be most stable for ambient air
relative humidities between 31% and 47%, however further work
on this relationship is needed to be definitive in either
humidity effects or controls.
Obviously more work on calibration curves for places
elsewhere than New York City is necessary. However, it is
believed that with relatively small error, the New York City
curve could be used for making relative comparisons of small
size suspended particulates in the atmosphere based on Coh/IOOO'
readings as they are routinely taken. Likewise, with the pro-
cedures offered a calibration curve index for other locations
could be established quickly for curve positioning.
-------
4.
It would not be necessary for practical and routine
measurement to undertake the more rigorous and exacting work
required to revalidate the form of curve for each locality,
since both in New York, Great Britain, France and Holland the
studies have uniformly confirmed that curve form, within the
normal measurement limits, is consistent and needs only to be
positioned at the standardized index point.
Particulate Measurement Systems
United States Practice
Although there are several methods for measuring par-
ticulates in air, it has been somewhat routine practice in the
United States to measure suspended particulates in atmosphere
by two basic systems. One is a gravimetric process the other
is related to measurement of optical density of filtered
deposit.
The gravimetric system requires the collection of
particulates taken from a large volume of air on fairly large
size filter paper (usually 4" diameter or 8" x lO"). The
method has been known as the Hi-Vol method. The filter paper
is weighed before and after the sample, and the air volume pro-
ducing the quantity of particulate is measured by rate of flow
and time of run. Results are reported as ~g/cubic meter of
airo
A commonly used method in the United States employing
the optical density relationship is described fully in
ASTM-DI704-6l (Standard Method of Test for Particulate Matter
in the Atmosphere). The system is applicable to fine particles
(usually less than 40 ~ in diameter). Air is passed through a
definite circular area of filter paper at a measured sampling
rate for a one or two hour period and is repeated on a new area
in sequence so that in the selected increments of time it is
possible to obtain a series of spots or stains that can be com-
pared by measuring either reflectance or transmission of light
-------
5.
as compared with clean paper. The optical density relationship
measured by transmission thus obtained is then converted into
a relationship with the volume of air producing the stain and
has commonly been reported as Coh/1000 linear feet. A variant
in reporting is introduced by using the reflection value con-
verted to a relationship known as the rud, or a deposit which
produces an optical reflectance of 0.01 due to 10,000 lineal
feet of air. The rud then equals 0.1 Coh/1000 ft.
European Practice
Practice in Europe stems from early work of Harris, and
Owens and Clark. In 1936 Hilll used transmitted light to
assess the stain produced on a filter paper after a known vol-
ume of air had been passed through it. A relation was estab-
lished between the optical density and the mass concentration
of smoke. Calibration curves were established to show rela-
tionships between % reflectance and weight in ~g/cm2 filter
paper surface.
Waller2 has reviewed the historical development of
British Smoke Curves that are based on optical density rela-
tionships and not on mass concentrationo
The importance of differentiating between weight and
darkness was examined by the Working Party on Methods of Mea-
suring Air Pollution and Survey Techniques, Organization for
Economic Cooperation and Development and reported in 1964.3
That group pointed out that darkness of the stain cannot be
proportional to the total weight of suspended matter in the
volume samples and should be considered simply as an indication
of the dark material in the air. However, experimental work
of the group has shown that the shape of the calibration curve
does not vary beyond reasonable limits, so that it is possible
to have a workable index of dark smoke concentration related
to standard smoke from a reflectometer reading. And, in fact,
the form of curve between darkness index of 10 and 60 was
sufficiently reliable that it could be referred to as the
-------
6.
Generalized International Standard Calibration Curve. By
normalizing the curve at an equal point (in this case 20 ~g/cm2
at darkness index of 25) the curve could then be used to define
the proposed equivalent international standard smoke scale. It
has been reasoned therefore that the adaptation of an estab-
lished universal form of smoke calibration curve would allow
results of air pollution surveys made allover the world to be
compared directly.
The British form of that curve has been taken by the
British Standards Institution as a basis for the measurement
of smoke and is now used in the National Survey. The Warren
Spring Laboratory carried out an experimental program that
stemmed from a basic assumption that the deduced (from curve)
surface concentration on the filter is linearly proportional
to the volume of the sample passed through it.4 It was estab-
lished that properties of the proposed curve held good for a
variety of smokes in England, France and Holland as presented
in the DECD report3 and since 1964 the curve has been used in
England for the estimate of smoke concentrations in equivalent
standard smoke units. The surface concentration determined by
standard curve was transferred into smoke concentrations by
calculations according to equations derived for the purpose
and this calculation permitted the expression of a darkness
index measurement in equivalent units of ~g/m3 for reporting
purposes, valid .for a darkness index ranging from 0 to 60.
Reporting Units of Both Practices
Thus British data on particulate density are now reported
as ~g/m3, and it should be clear that though this unit is re-
lated to a gravimetric reading of surface concentration in its
original derivation, it is not and cannot be a reporting of
actual weight of material trapped on filter paper per unit
volume of air passed through the paper.
It is also quite apparent that if particulate density
as measured nationally in the United States by the ASTM Standard
-------
7 .
Method (DI704-61) is to be compared to particulate density as
measured in England or the European countries such as France
and Holland, there must be a relationship established between
Coh/IOOO' and ~g/m3 the reporting units of the two light index
measurement systems.
This project has been directed to the major task of
establishing that relationship for at least one city of the
United States, New York Cityo If, as has been found in the
European studies, the curve form is the same between acceptable
limits, and an index point placement of the curve is possible,
then procedures for the transfer of United States reporting
units into English units and vice versa become a reality. The
experimental procedures and the results obtained are presented
in the following pages.
Correlation of Particle Size and Particulate Density
In the latter stages of project work it became apparent
that size of particle trapped on the filter might influence the
index. Exploratory work related to the establishment of those
relationships that might exist has been performed and is also
reported here.
Experimental Procedure
It was the intent of the procedure to duplicate as
closely as possible the calibration procedures used by the
Warren Spring Laboratory. Both before the start of the project
and during the project, personal visits and correspondence were
employed to check out the procedures and findings. Several
items of equipment and, for portions of the work, the same fil-
ter paper used in England were borrowed so that any differences
attributable to the equipment or paper could be reduced or
eliminated.
-------
8.
Relationship .Between Units of Surface Concentration
The equations used in Great Britain for determining
smoke concentrations in the air are calculated by the use of
the British Standard Smoke Calibration Curve. For reflectom-
eter reading of 99 to 40, equation (1) is used:
C = ~ (91,679.22 - 3,332.0460 R + 49.618884 R2
- 0.3532977S R3 + 0.0009863435 R4)
( 1)
where C = concentration, ~g/m3
V = volume of sample in cubic feet
F = constant equal to 1000 for 1 inch filter
For reflectometer readings between 40 and 20, equation (2) is
used:
C = ~ (214,245.1 - 15,130.512 R + 508.181 R2
- 8.831144 R3 + 000628057 R4)
(2)
where the key is as above.
In the United States smoke concentration is expressed
as soiling index where the Coh unit is used. The Coh unit is
defined by equation (3):
O.D. X area of spot x 105
Coh/1000 linear feet = volume of air sample
(3)
where O.D. = optical density determined by light transmission
The solution of equation (1) for sample volumes of
30 ft3* and 70.62 ft3** (2 M3) is plotted in Figure 1.
The solution of equation (3), plotted in Figure 2, was
solved for a 1" diameter spot and a sample of 30 ft3.
It is noted that light reflection is used to determine
~g/m3 whereas light transmission is used to determine Coh
*Represents volume for two hour sampling period on AISlo
**Represents volume for typical sampling period in Great
Britain.
-------
250
M 200
s
..........
b.()
::J.
~
0
.r-!
.j...J 150
ru
H
.j...J
~
C.J
()
~
0
u
C.J 100
()
ru
LI-i
H
;:j
rJ)
50
300
9.
Surface Concentration
~g/m3 VB. % Reflection
o
100
90
80
70
60
50
Percent Reflection
Figure 1
-------
5.0
.u
ClJ
ClJ
~
I-< 4.0
cu
-------
11.
values. Due to the nature of the instruments, there is almost
a 1:1 relationship between these two quantities.
Rationale of Comparison
The ultimate goal required a method of comparison
between ~g/m3 and Coh/lOOO LF. If, for example, 100 ~g/m3 is
reported, what would be the comparable Coh value? From Figure
1 we note that a surface concentration of 100 ~g/m3 is obtained
when the resultant spot produces a reflectance of 60%, when
sampling at the normal flow rate in Great Britain and sampling
a volume of 70.62 cubic feet. If we now assume that an AISI
instrument was sampling the same atmosphere, it would require
a percent transmission (or reflection) reading, after a sample
of 30 cubic feet, of 75%0 This 75% value would be equivalent
to a Coh reading of 2.3.
The British curve is based on a
darkness index of 25 equals 17 ~g/cm2.
form, a darkness index of 25 equals 20
The sub-goals then became:
1. the comparison of curve form to determine that a
curve for New York did or did not fit into the
European family of curves, and
2. the determination of the value for New York data
that would normalize at a darkness index of 25.
standard
For the
~g/cm2.
curve where a
international
Description of Sampling Sites
The New York University Bronx Station is located in the
northwest section of the Bronx, on the NYU Campus, at 1911
Osborne Place. The surrounding land usage includes a .steam
power plant, several thousand feet to the north, operated by
Consolidated Edison Company of New York; five-story apartment
buildings to the south and east; and the Major Deegan Highway
and the Harlem River to the west. Two sampling probes were
placed on the outside of the building facing toward the north-
west. These probes were three feet from the building at a
height of about seven feet above the ground. This site was
-------
12.
used for runs covering the periods March 1966 to October 1966,
and January 1968 to February 1968.
The station designated as Christodora Station (Manhat-
tan) is located at 601 East 9th Street on the lower east side
of Manhattan. Immediate surroundings include a park, five-
story apartment buildings, and eighteen-story buildings to
the north and east about 2,000 feet on the peripherYn A steam
power plant, operated by Consolidated Edison Company of New
York is northeast of the area. The two sampling probes ex-
tended three feet beyond the building and about 189 feet above
the ground, facing in a southwesterly direction. Sampling
was conducted at this station from February 1968 to August
1968.
The stations are located as shown in Figure 3.
approximate distance separating the New York University
Station and Christodora Station is ten miles.
The
Bronx
Determination of the Form of the Calibration Curve
In this first phase of the investigation, proportional
sampling of the air provided a series of smoke stains with
increasing values of darkness index (darkness index equals
100 - i~) for corresponding increasing volumes of air sampled.
This was accomplished by using several fractional timers,
arranged so that the flow to each individual filter was oper-
ating either 11.1%, 16.6%, 33.3%, 50%, 66.6% or 100% (contin-
uous flow) of the time. The actual on time percentage was
slightly less, as shown in Table 1.
Basically, air was drawn through an inverted funnel
four inches in diameter and fed into a wide mouthed cylindri-
cal mixing bottle by means of a one-half inch I.Do glass tubing.
The tubing entered the stopper centrally and dropped to one
inch above the base. Seven symetrica11y placed one-quarter
inch I.Do glass tubes were placed one-half inch into the bottle
and each tube led to a one inch diameter filter clamp contain-
ing Whatman No.1 filter paper and then to a dry gas meter
-------
1:-1
o
(')
I1J
rt
~.
o
~
I':r:j 0
~. I-t1
()Q
~ CJ)
Ii I1J
CD a
'1::1
W t-'
~.
~
()Q
CJ)
~.
rt
CD
en
~
I-
( =
TH£ 6RONX
q
v
~
Sample Site
Locations
IV
s
E
uPPER ~y ~
I y-'~~\
4 / 8 \, J'.F.K. t-'
~ $ A.RP~ w
o 0
8 R
~ A-/ /
-------
14.
Table 1
Pump Operation During 15 Minute Cycle
Theoretical Actual Actual
On Time On Time On Time
Pump Noo (Minutes) (Minutes) (%)
1 1500 15 100
2 10.0 10.03 66091
3 7.5 7.22 48.1
4 5,,0 5,,0 33.3
5 2.5 2.57 17.1
6 1.7 1.64 10.9
-------
15.
capable of rea4ing 0.1 ft3. One filter was used as a control
and the remaining were used as test filters. Each meter was
connected to a 1.3 cfm pump which had been orificed to produce
a flow of approximately .073 ft3/min.
The fractional timers were arranged to operate the
pumps for different periods of time over a 15 minute cycle.
The number of cycles that the test was run was set so that
there was an approximate darkness index of less than 60 on a
second continuously operated control filter. This filter was
used only to check the progress of the test and, therefore,
was removed from the filter clamp periodically to check the
darkness index.
Table 1 gives the actual time each pump was running
during a 15 minute cyc1eo The apparatus for the test program
is diagrammed in Figure 4 and listed in Appendix A.
A general view of the test apparatus is shown in
Exhibit 1. Exhibit 2 illustrates pictorially the sequence of
flow diagrammed in Figure 4.
At the end of the test period, usually 24 hours, the
volume of air passed through each filter was read on dry gas
meters and the darkness index of each stain was read with a
Photovo1t Ref1ectometer using a green tristimu1us filter (see
Appendix B).
This procedure was repeated several times and each
individual test was graphed, volume of air vs. darkness index.
As each curve was plotted, the values at darkness indexes of
5, 10, 15, 20, 25 through 50 were summed. Each individual
value of "volume of air" was then divided by the summation and
multiplied by 100. This new unit was now an "arbitrary unit"
which brought the curves into coincidence over a normal work-
ing range.
Determination of Weight of Total
Suspended Matter in the Air
For the determination of smoke concentrations in
-------
Vacuum
Pump
% time
on
Orifice
~
.. Dry Gas
~ Meter
tot
(I)
~
Filter Clamp
and Holder
Sample Line
Mixing Bottle
Air
Intake
.....
0\
Apparatus for Test Program
-------
General View of Test Apparatus
t--'
-....J
-------
~-
~nb_~
Exhibit 2
Illustration of Sequence of Flow '
Shown in Figure 4
~
JO
------
-------
19.
absolute gravimetric units, the procedure illustrated in
Figure 5 was used.
The air sample was initially drawn through an inverted
60°, 8.5 inch diameter glass funnel into a wide mouthed mixing
bottle. The intake line was within one inch of the bottom of
the bottle and the sample lines leaving it were placed one-
half inch into the bottle. Since the volume flow was different
in the two lines, the diameter of the lines were varied to give
a similar velocity in each line. The smaller diameter line led
to a one inch filter clamp, housing Whatman No.1 filter paper,
then to a dry gas meter, and finally to a 1.3 cfm vacuum pump
which had been orificed to give a flow of approximately 0.073
ft3/min. The larger diameter line was connected to a four inch
glass fiber filter housed in line with a dry gas meter and
attached to a high vacuum pump orificed to deliver a flow of
1.5 ft3/min. (see Exhibit 3).
A secondary four inch glass fiber filter was required
as the "control" filter and was used in the test to determine
extraneous changes in weight arising from handling procedures.
Prior to testing, the glass fiber filters were placed
in a desiccator containing magnesium perchlorate for 24 hours
and weighed on a Mettler electronic balance. One glass filter
was placed in a four inch diameter filter unit in the high
volume sampling line, and the other four inch filter was placed
in the control unit. Apparatus used in this procedure is
listed in Appendix Co
The one inch Whatman No.1 filter paper was placed in
its filter clamp and the testing was startedo
A test period generally continued 24 to 36 hours depend-
ing on the time required to produce a stain of suitable dark-
ness, approximately a darkness index of 50 on the one inch
Whatman filter. The glass fiber filters were removed and
allowed to dry in a desiccator for 24 hours before being
weighed. The increase in weight on the test filter was a
measure of the dry deposited material or smoke and any change
-------
tzj
....
(JQ
c::
11
(\)
VI
Air
Intake
.
.
.
.
- ~--
Mixing
Bottle
1" diam. Filter
o
Dry Gas
Meter
Vacuum
PUIl1p
.073 cfm
4" diam. Filter
o
Dry Gas
Meter
Apparatus to Determine Scale of Smoke Curve
Vacuum
Pump
1.5 cfm
N
o
.
-------
21.
Exhibit 3
Hi Volume Sampling System
With Meter and Pump
-------
220
in weight arising from the handling procedure. A correction
factor for the error due to handling was made by either adding
or subtracting the difference between the initial and final
weights of the control filterG
If:
W
VM
= weight of dry deposited matter, ~g
= volume of air passed by the high volume
sampler, M3
the mean concentration of suspended matter, C, is equal
then
to:
C
=
W
VM
or
,
(4)
As the same concentration of suspended matter is sampled
in the normal sampling line, the surface concentration of sus-
pended matter present on the one inch diameter filter can be
calculated from the corresponding volume of air passed in this
line and the area of the filter.
If:
v = volume of air sampled by the normal sampler, M3
D = diameter of the filter, cm
then the surface concentration of suspended matter on the
filter, S, is given by:
S
=
(4W) (-y.)
DT VM
or
(3-)
cm
(5)
Surface concentrations computed in this manner are seen
in Table 2.
of Relationshi Between
an 0 Transm1ss10n
The standard instrument used to sample particulate
matter in CohjlOOO' is an AISI Model No. F2, manufactured by
the Research Appliance Company of Pennsylvania. This instru-
ment had an adjustable timer so that samples could be taken
for any time period between one minute and three hours. The
-------
23.
Table 2
Normalization of Arbitrary Units
Norma1ized*
Normalized Surface
Average Surface Concentration
Darkness Arbitrary Concentration ~g/cm2
Index Unit-NYU 'fJ-g/cm2. NYU International
5 0.7 109
10 1.7 4.8 507
15 3.1 807 905
20 4.9 1308 1405
25 701 2000 2000
30 908 27.6 2609
35 12.8 3600 34.8
40 1602 45.6 45.2
45 19.8 55.8 5707
50 24.8 69.8 73.1
*Data from private communication from DSIR, Warren Spring
Laboratory.
-------
24.
instrument is also equipped with an adjustable flowmeter vary-
ing from 2 ft3/hr to 20 ft3/hro Calibration of the instrument
was done by using a Precision Dry Gas Meter with varying flow
rates on the f10wratero The standard filter paper used was
Whatman #4.
Percent transmission was measured by a Research
Appliance Company Spot Evaluator, Model No. 363. This instru-
ment is capable of measuring optical density between 0-100
percent transmission. Standardization is made by adjusting
the built-in rheostat to 100% T on a clean section of Whatman
#4 filter paper before and after samples are recorded.
Percent reflection was measured by a PhotovQ1t Reflec-
tion Meter, Model No. 610, capable of measuring percent
reflectance between 0-100. One clean sheet of Whatman No. 1
filter paper placed on a white enamel plaque was used in the
standardization procedure. The galvanometer was then adjusted
to read 100% R.
One Hour vs. Two Hour Coh/1000' Values
In order to determine whether one hour Coh/1000' values
would be the same as two hour Coh/1000' values (i.eo, the
average of two one-hour Coh values equaling a two-hour Coh
value), the following procedure was used.
Two Research Appliance Company AISI Model F2 Samplers
were attached to a common mixing jug. One AISI sampled for
one hour periods while the other sampled for two hour periods.
The f10wraters were set at 15 cfh or .25 cfm and were kept at
constant flow.
Each sample was measured on a RAC Model 363 Spot Evalu-
ator in % T and then converted to Coh/1000'. The one hour Coh
values were then averaged into two hour sampling periods cor-
responding to the two hour Coh values. The results of these
comparisons are discussed in a later section of this report.
-------
25.
Results
Shape or Form of Smoke Calibration Curve
Two series of data (1966, 1968) were used in determining
the shape of the NYU Bronx Station curve in arbitrary units.
These arbitrary units were computed by plotting the volume of
air sampled in cubic feet vs. the darkness index (DI = 100 -
% R) for each run. After each curve was plotted, the values
at darkness indexes 5, 10, 15 through 50 were summed. Each
volume of air was then divided by the sum and multiplied by
100. This new unit was the arbitrary unit for the different
darkness indexes
ft3
arbitrary unit = x 100
~ft3
(6)
The values in arbitrary units for each series of runs
were tabulated in Table 30 These values were averaged for each
darkness index and the NYU Bronx Station curve in arbitrary
units is shown in Figure 6D
Manhattan Curve
In order to determine whether the shape of the smoke
curve would vary according to geographic location, the smoke
curve apparatus was moved to Christodora Station in Manhattan.
The same procedure for calculating arbitrary units in the Bronx
curve was used in the Manhattan curve. Data obtained from the
Christodora runs are given in Table 4. Each run was then aver-
aged for the varying darkness indexes and the Manhattan curve
constructed from the data is given in Figure 7.
Comparison of International-Bronx-Manhattan
Curves.in Arbitrary Units
The proposed generalized International Standard Calibra-
tion Curve as reported in the DECD report3 is the average of
natural curves found in France, Holland and Great Britain. An
-------
26.
Table 3
NYU (Bronx) Station
1966 and 1968 Values in Arbitrary Units
Darkness Index
Date 5 10 15 20 25 30 35 40 45 50
3/29/66 0.6 1.,5 3.1 5.1 7,,4 9.9 12,,8 16.0 19.7 23.6
3/30/66 0.,5 1.6 3.,1 5.,0 7.,2 9.8 12.,0 16.0 19.8 23.,9
3/31/66 0.6 1,,7 3.2 5.,2 7,,4 10.,0 12.8 16.1 1905 23.4
4/05/66 0.8 1.9 3.3 5.,2 7,,4 9.,8 12.,6 15.7 19.7 23.6
4/07/66 0.,5 1.,5 2.,7 4.,7 6.,6 9.3 12.5 16.2 2005 25.3
4/18/66 0.5 1.4 2.6 4.3 6.2 8.7 12.2 16.2 21.2 2607
6/14/66 006 107 301 5.0 701 907 12.5 16.0 19.9 24.3
6/20/66 0.,9 2.2 306 5.,3 7.4 9.7 1204 15.8 19.3 23.7
6/27/66 008 2.0 306 5.3 7.3 9.8 12.3 15.6 18.9 22.4
9/28/66 0.,7 108 3.2 5.0 7.2 9.9 1300 16.2 19.6 23.3
10/03/66 0.7 1.1 2.5 4.3 6.5 9.4 12.7 16.7 20.3 2507
1/03/68 008 1.5 3.3 5.1 7.3 907 12.5 16.0 1906 24.2
1/18/68 0.5 1.6 3.0 4.9 6.8 9.5 12.2 15.8 20.1 25.5
1/22/68 0.6 1.9 3.3 5.0 7.1 905 12.4 15.8 19.5 24.9
1/25/68 0.7 1.8 3.2 4.7 6.7 9.4 12.4 15.9 20.1 25.0
1/30/68 0.,8 1.7 3.0 4.5 6.7 9.4 12.2 16.0 20.2 25.5
2/05/68 0.8 1.8 3.2 5.0 7.3 9.8 12.8 15.9 19.6 23.8
2/12/68 0.6 1.6 3.0 4.8 6.8 9.2 12.2 16.0 20.4 25.1
Average 0.7 1.7 3.1 4.9 7.0 9.6 12.5 16.0 19.9 24.4
-------
30
25
(JJ
.j..J
.r-!
C
::J
:>-.
~
co 20
,
h
.j..J
.r-!
,.0
~
-------
28.
Table 4
Christodora (Manhattan) Station
1968 Values in Arbitrary Units
Darkness Index
Date 5 10 15 20 25 30 35 40 45 50
2/27/68 1.0 2.0 3.3 5.0 7.0 8.7 1200 15.3 20.0 25.6
3/05/68 1.0 1.9 3.4 4.9 6.9 8..9 11.8 15.5 20..1 25.5
3/12/68 0..6 107 3.0 4.7 607 9.2 12.3 16 00 \ 2000 2508
3/26/68 004 009 202 4.0 5.8 8.9 12.1 16.5 21.5 2703
4/02/68 0.7 1..7 302 500 7..0 9.2 11.5 1403 19.1 28.2
4/23/68 0.5 1..5 2.9 4.9 7.2 906 12.4 1503 1905 2601
4/30/68 007 108 3.3 5.0 7.0 901 12.0 1500 2000 26.5
5/09/68 007 107 3.1 4..8 609 809 11.7 1506 2002 2604
5/14/68 0..3 1..2 2.6 403 606 9.3 1202 1600 20.7 26.7
6/10/68 005 102 205 4.3 605 901 1006 1507 2101 2802
6/17/68 0.7 108 3.1 407 605 807 1106 15.5 20.6 26.7
6/24/68 005 104 2..8 407 6..8 9..5 12.2 15.3 19.7 27..1
7/01/68 0..7 1.9 3..3 5.1 7.3 9.8 12..7 16..0 1906 23..5
Average 0.6 1..6 2.9 407 6.8 9.2 11.9 15.5 20..2 26.4
-------
30
25
rJJ
.\J
'''';
r::
:;:J
>..
H
Ct! 20
H
.\J
''''';
,.0
H
«
r::
'r-1
r:: 15
o
.,..,
.\J
Ct!
H
.\J
r::
~ 10
r::
o
u
QJ
tJ
Ct!
4-1
~ 5
U)
29.
NYU Christodora Station Smoke Curve
in Arbitrary Units
10
20
30
40
50
Darkness Index (100-%R)
Filure 7
-------
30.
Eel Reflectometer was used in determining the form of the
international calibration curve, but data obtained from private
correspondence with the DSIR in England* indicated that read-
ings obtained with a Photovolt Reflectometer are almost iden-
tical with those taken with an Eel Reflectometer (see Table 5).
A comparison of the NYU Bronx Station, Christodora
(Manhattan) Station and the International Curves in arbitrary
units (Figure 8) shows that these curves are almost identical.
This relationship established that the form of the smoke curve
for New York City stations is relatively the same and in turn,
the New York City curve is relatively the same as the Inter-
national curve within the limits of darkness index normally
measured.
Determination of the Scale to be Attached
to the Smoke Calibration Curve
An absolute, constant scale cannot be attached to a
calibration curve as the weight of material deposited depends
upon the nature of the material forming the smoke and this
varies from place to place and with time at anyone location.
The International Standard Calibration Curve now in use, there-
fore, is in terms of surface concentration of a hypothetical
equivalent standard smoke established by arbitrary definition,
but based on extensive experimental data.3
To determine the absolute scale of surface concentra-
tions to be assigned to the ordinate of a calibration curve at
any particular place and time, it is necessary to determine the
absolute weight of total suspended matter or smoke in the air.
The procedure and apparatus used for this determination have
been discussed previouslYD
It is recommended by the DECD that for Whatman No. 1
filter paper and the Eel Reflectometer, the Generalized
*These data obtained by sending papers read with Photovolt
Reflectometer to Warren Spring Laboratory for independent
reading with Eel Reflectometer.
-------
31
Table 5
Comparison of Reflection Readings
Eel and Photovolt Ref1ectometers
Using 1" Whatman Noo 1 Filter Paper
Eel Photovo1t
80.0 79.5
79.0 7805
73.0 75.5
65.5 65.0
64.5 64.5
57.5 55.5
54.5 53.5
-------
32.
Mean -
International - Bronx - Manhattan
Smoke Calibration Curves
in Arbitrary Units
30
tI) 25 i('~O
.j.J
'..-1
!::
~
:>,
~
CtI
~
.j.J 20
-..-I
,.!) ,'(' NYU (Bronx)
~
-< A Interna.tiona1
!:: 0 Christodora (Man.)
...-1
!:: 15
0
...-1
.j.J
CtI
~
.j.J
!::
(l)
u
!:: 10
0
u *£
(l) /
u
CtI
4-1
~
::1
C/)
5 *£0
/
*&0
~
10 20 30 40 50
Darkness Index (100-%R)
Figure 8
-------
33.
International Calibration Curve, the form of which has been
defined above, should be normalized so that a darkness index
of 25 (DI = 100 - % R) corresponds to a value of 20 ~g/cm2 for
the surface concentration of an equivalent international stan-
dard smokeo By using 20 ~g/cm2 at a darkness index of 25, the
normalization formula becomes:
20
= Normalized ~g/cm2 at X
(7)
NYU Bronx and Christodora (Manhattan)
Station Results
In order to determine the surface concentration scale
to be assigned to the ordinate of the New York City Smoke
Calibration Curve, three series of runs, during two different
years (1966 and 1968) at two different locations (NYU Bronx
and Christodora Manhattan) were made. For each set of data a
two variable non-linear correlation was used
where a = 0, X = DI, Y = ~g/cm2
Y = b ~ X + c ~ X2
XY = b ~ X2 + c 2:.. X3
and was substituted in
( 8)
(9)
Y = a + b X + c X2
(10)
for the equation of the line.
calculated:
The following equations were
NYU Bronx 1966 Series
Y = .4871 X + .02375 X2
(11)
Christodora (Manhattan) 1968 Series
Y = 089 X + .011 X2
(12)
By combining these two similar series of data with additional
data taken at our NYU Station in 1968, the proposed New York
City curve equation becomes:
-------
34.
y = 1.00 X + .0079 X2
(13)
where
x = darkness index
y = ~g/cm2
Proposed New York City Standard Calibration Curve
From these results it was calculated that the standard
value for Christodora (Manhattan) Station is 29 ~g/cm2 at a
darkness index of 25 and the NYU Bronx Station yields a result
of 28 ~g/cm2 at a darkness index of 25. The combined data of
these two curves (equations 11 and 12) plus additional data
from our NYU Station in 1968, produced the proposed New York
City Standard Calibration Curve in ~g/cm2.
The value of 30 ~g/cm2 at a darkness index of 25 was
selected as the representative positioning value and the
equation for normalizing the data was:
30
= Normalized ~g/cm2 at X
(14)
These data would shift the curve i& from the International
curve and ~ from the British curve, and would introduce a
1~6% correction in the conversion using the British curve.
The results of normalization process calculations then
yielded the proposed New York City Standard Calibration Curve
presented in Figure 9.
A tabulated comparison of the relation between darkness
index, the Internation Smoke Curve, and the proposed New York
City curve is given in Table 6.
Conversion of that curve into the reporting unit ~g/m3
is then accomplished by calculating values using the relation-
ship of equations (4) and (5) expressed as
~g/m3 = 5.06 xv~g/cm2
(15)
The curve of equivalent standard smoke units with
V = 2 M3 (70062 ft3) derived for New York City is shown in
-------
N
i=
CJ
-
Ol
;:j
!::
o
.0-1
W
C\j
~
w
!::
Q)
CJ
!::
o
U
Q)
CJ
C\j
4-j
~
;:j
Cf.)
35.
Proposed New York City Calibration Curve
in Ilg/cm2
Photovolt Reflectometer-Whatman No.1 Filter
Normalized to DI = 25 Ilg/cm2 = 30
Paper
120
100
80
N.Y.C. Curve
60
40
20
International Curve
10
20
30
40
50
Darkness Index (100-% R)
Figure 9
-------
36.
Table 6
Comparison
International Smoke Calibration Curve
and
Proposed New York City Smoke Calibration Curve
in IJ-g/cm2
Darkness International New York City
Index Curve Curve
10 5.7 7.1
15 9.5 13.2
20 14.5 21.2
25 20.0 30.0
30 26.9 40.8
35 34.8 53.3
40 45.2 68.7
45 57.7 87.1
50 73.1 110.1
-------
37.
Figure 10 and comparison data are tabulated in Table 7.
Figure 10 is now in a form that can be used to trans-
late the equivalent unit values for comparison that might be
made between United States cities and those of Great Britain
and other European countries. However, there remains one fur-
ther step applied to United States practice where the reporting
unit is Coh/1000' rather than ~g/m3.
Development of European vs. United States
Reporting Unit Relationship
The development of Figure 10 was accomplished according
to the Warren Spring and DECD procedure and utilized darkness
index as measured by photovo1t meter.
As soon as it became necessary to convert
based on light transmittance into ~g/m3 based on
tance, questions arose concerning
1. the validity of using % T or % R, and
2. the comparability of readings taken at one hour
or two hour intervals on particulate spots
developed by passing air through the filter
paper of the AISI sampler.
Accordingly, these two issues were investigated.
Coh/1000'
light ref1ec-
Validity of Using % T or % R
If a light beam is placed on a piece of paper the light
will either be reflected, R; transmitted, T; absorbed by the
dirt, A; or absorbed by the paper and scattered, E. This can
be represented by:
T + R + A + E = 100
(16)
Therefore the equations for transmittance and reflectance are:
T = 100 - R - E - A
(17)
R = 100 - T - E - A
(18)
-------
320
280
240
200
M
~ 160
tIC
:i
120
80
40
38.
Proposed Standard Calibration Curve
in IJ.g/m3
DI = 25, IJ.g/cm2 m 30, V = 2 M3
o
New York City
International
Great Britain
10
20
30
40
(100 - ;oR)
10
60
50
Darkness Index
Figure
-------
39.
Table 7
Comparison
International-Great Britain Smoke Calibration Curve
and
Proposed New York City Smoke Calibration Curve
in
~g/m3 at V = 2 M3
Darkness
Index International Great Britain New York City
10 14.4 12.4 18.0
15 24.0 2100 33.4
20 36.7 3009 53.6
25 50.6 43.1 75.9
30 6801 5707 103.2
35 8800 75.4 134 . 8
40 114.5 9702 173.8
45 146.0 12400 22003
50 18500 15701 27805
-------
40.
When the light transmission of a sample is read, clean paper
is inserted and the light transmittance is adjusted to 100%.
The sample spot is then inserted and a reading is obtained.
Initially we have:
For
A = 0
T = 100%
which says
R + E = 0
and the equation for transmittance becomes
T = 100 - A
(19)
When reflectometer readings are made the reflectance is set at
100% on a clean sheet of filter paper. Therefore initially:
For
A = 0
R = 100
which says
T + E = 0
and the equation for reflectance is
R = 100 - A
(20)
Since the equations for transmittance and reflectance
are the same, a given sample spot should give identical read-
ings for both methods of evaluation.
Procedure
A clean roll of Whatman No.4, I-inch filter paper was
placed in a spot evaluator (reading percent transmission) with
the initial reading set to 90%. The roll was continuously
passed through the light beam and % transmission was noted.
-------
41.
The results are as follows:
% Transmission
Interval
No. of values
In Interval
77-79
80-82
83-85
86 -88
89-91
92-94
95-97
o
9
44
247
130
20
o
Initial Setting
It is noted that while these values can be reset to an
initial reading on a clean portion of the paper proceeding the
dirt spot, one is never sure that the portion where the spot
appears would not have the same initial setting. It is obvious
that the paper is not uniform. The most nearly acceptable
procedure for determining initial setting appears to be the
average of a reading Lmmediate1y preceding and one Lmmediate1y
following the actual spoto There is no assurance that that is
the correct reading, however the errors so introduced may be
no greater than errors due to humidity, and other factors
influencing either light transmission or reflectance values.
Table 8 indicates the data used in the comparison between
transmittance and reflectance, and Figure 11 compares the data
with the theoretical 1:1 relationship. Procedures and instru-
ments for making this comparison are described in Appendix D.
The 1:1 relationship is therefore not absolutely true.
The difference introduced at 50% reflection is in the relation-
ship of the curve as shown by the equation:
y = 2.71 + 1.008 X
(21)
where y = % Transmission
X = % Reflection
The relationship is apparently linear and therefore in the
normal range of measurement (say 90 to 50% transmittance) can
easily be converted by using Figure 11 (i.e., 50% R = 53.11% T;
-------
42.
Table 8
Percent Transmission vs. Percent Reflection Values
1966 and 1968 Data
%T %R %T %R %T %R %T %R
71 70 90 93.5 85 85 46 40
72 7305 62 58 93 88.5 48 50
75 7805 74 66 91 9305 64 57
66 6405 76 61 98 94 70 63
70 6905 84 7705 54 52 80 75
64 6105 87 86 66 61 81 82
64 59 95 9005 67 67 52 47.5
69 65 76 74 72 74.5 57 58.5
76 7305 86 81 79 84 66 66
80 79 82 8405 86 88 78 72.5
94 8505 89 89 43 42 84 8005
97 89.5 94 9205 51 50.5 88 87
62 63.5 98 95 53 5705 74 6805
70 70 50 53 63 66 54 52.5
74 75 63 58.5 72 77 63 66.5
81 82 62 56 82 83.5 38 39.5
88 89 62 57 45 39.5 64 64
97 92.5 72 63.5 58 48 56 53.5
66 60 78 72 59 54 52 52.5
67 67.5 83 81 64 63.5 50 50
74 74 84 88 79 74 62 61
82 79.5 42 41 84 81.5 71 70
93 86 50 49.5 38 28.5 80 75.5
90 91 57 55.5 46 38 42 36.5
65 64.5 71 63.5 60 46 54 51.5
74 71 78 73.5 62 54 64 60
78 78.5 86 78.5 76 74 77 78.5
84 8205 67 7305 72 68 78 79.5
92 89 80 80 80 75 70 75.5
55 5305 55 5505 67 65 63 58
82 7805 65 64.5 68 6805 76 72
60 61
-------
c
o
~
en
en
~
e
~ 70
CIS
~
~
~
60
80
90 -
/
/
/
/
>/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
Comparison
% Transmission vs. % Reflection
40
50
1:1 Relationship
90
80
70
7. Reflection
Figure 11
43.
Y = 2.71 + 1.008X
60
50
40
-------
440
90% R = 93.43% T). If % T were used directly in calculating
darkness index (100% - % R), the difference in equivalent smoke
units would be 47 ~g/m3 less at 50% R than at 53.11% T. This
is about as large an error as would normally be introduced by
that practice. On the other hand, the minimum error in using
the 1:1 relationship at 90% R would be 8 ~g/m3. The routine
field observer might then at some point wish to correct for
this by calculating darkness index using % R calculated for
% T according to equation (21), whereas research work of
greater refinement would probably require the correction on
all % T readings.
When the ambient air is passed through filter paper,
the dirt in that air will be retained on the papero The re-
sultant "spot" can be reported in terms of soiling index or
Coh/lOOO LFo This term combines into a quantitative expres-
sion, the Coh, which is an arbitrary unit of measurement
describing light scattering potential and the volume of air
from which the particulates producing that potential have been
trappedo
The Coh unit is defined as that value of light scatter-
ing particulates that produces an optical density of 0.01 when
I
measured by light transmission. This is the term Log l in
the Beers Law equation
Io
Log T
=
KQ
(22)
where K = constant of proportionality
Q = quantity of sample in air
Plots of that optical density are proportional to the air
volume and the plot of one against the other results in a
straight line when light transmission values are above 50%
(i.e., low optical densities). There is also in practicality
-------
45.
an upper lLmit (about 90%) at which ability to read the light
transmission value accurately is Lmpaired and error is
introduced.
A 100% transmission value is taken on a piece of clean
filter paper and then the transmission value of the spot
created on a given area of paper passing air at a given flow
rate through that area is read as a percent of the original
optical transmission. DLmensional analysis shows that the
quantitation of
ft3
ft7
=
ffi
=
linear feet
Thus if the number of Cohs of particulate potential is in pro-
portion to the air flow, the expression Coh/1000 linear feet
expresses the light scattering potential in relation to the
air sample volume and the area of entrapment.
The measurement can be illustrated by the following
example:
Given Q = 0.250 ft3/min
TLme of sampling = 120 minutes
Sample volume = 30000 ft3
Given area of filter paper 1" diameter =
5045 x 10-3 ft2
30000 ft3 = 5504.5 ft
Air Sample = 5045 x 10-3 ft2
= 5.5045 1000 ft
If the resultant spot gives a light transmission value of 87%
log 100 = 0.0603 = 6.03 Coh
87
5~5g~5 = 1.10 Coh/lOOO ft
A question that arises concerns the influence of a
change in sample volume on the resultant transmission reading.
-------
46.
That is, if, as in the above example, the sample volume is
changed to 15 ft3 (60 min @ .25 ft3/min), will the light trans-
mission of the spot vary to the degree that Coh/lOOO LF will
indeed be 1.17 The problem is further complicated by the fact
that if the concentration of particles in the air changes
during the sampling period, the result would not be expected
to be the same. However, each one hour sampling period in a
given two hour density determination would be correct. This
assumption can be demonstrated readily by averaging two 1 hour
Coh readings taken simultaneously with one 2 hour determination.
The following example illustrates this point:
60 min cycle 120 min cycle
% transmission after 1 hour 87 not read
% transmission after 2 hours 100* 87
Coh end of first hour 202 not read
Coh for end of second hour 000 1.1
Average for 2 hour period 101 1.1
It is seen by this example that time and therefore
sample volume should not affect the final answer, and that
each segment, i.e., one hour particulate density value, is
correct unto itself.
Two main additional items also require consideration.
The quality of the filter paper fluctuates. In a roll of
Whatman No.4, for example, as used on the AISI instrument,
the percent transmission can vary as much as 9% (see page 41).
In the low ranges this is quite significant and a lar~er sam-
pling volume would be needed to provide a darker spot, and
thereby minimize the initial error. At high Coh/1000' values,
the filtering action may be more efficient than at low values,
80 that lower sampling volumes might be required.
*Assume clean air during second one hour sampling period.
-------
47.
Results of 1 Hour and 2 Hour Coh/1000' Testing
Two Research Appliance Corporation AISI Instruments were
placed into operation sampling the same atmosphere. One instru-
ment was set to cycle every 60 minutes and the other every 120
minutes. The flow rates were identical at 15 cfh.
For each two hour sampling period, two 1 hour readings
and one 2 hour reading were obtainedo An average of the two
1 hour readings was made and these average values were compared
to the one 2 hour value (see Tables 9 and 10). A line of best
fit was calculated and a correlation between the two values was
obtainedo The equation relating the two 1 hour Coh/lOOO' aver-
ages and the one 2 hour Coh/1000' is
y
=
- 0.14 + 1.31 X
(23)
where Y = average of two 1 hour values, Coh/lOOO'
X = one 2 hour value, Coh/lOOO'
The correlation coefficient was 0.9434.
Figure 12 represents 2 hour Coh/lOOO' readings vs. 1
hour Coh/1000' readings and shows that a one hour reading
alone cannot be directly representative of a two hour reading.
Relationship Between Equivalent Units
The final steps in completing a relationship between
United States re.porting units and European reporting units
required that all normalized data be averaged thereby obtain-
ing final values at varying darkness index values.
When that had been done, the values could be used in
equation (15) with the air volume V = 0.8496 M3 (the AISI sam-
pling flow for two hours at 0025 cfm), thus:
~g/m3
=
5.06 x ~g/cm2
0084
(24)
The curve so calculated is given in Figure 13 and can be used
for the AISI sampling unit set to pass 0025 cfm through the
-------
48.
Table 9
Frequency Distribution Coh/1000'
Average of Two 1 Hour Readings
February 15, 1968 to March 13, 1968
Interval No. of Cases in
(Coh/1000') Given Interval % Total
0.0-0.4 91 28.9 J
0.5-0.9 92 29.2 58.1
1.0-1.4 55 17.51
1.5-1.9 33 10.5 28.0
2.0-2.4 12 3.8 J
2.5-2.9 8 2.5 6.3
3.0-3.4 4 1.3
3.5-3.9 2 .6
4.0-4.4 7 2.2
4.5-4.9 5 1.6
5.0;'5.4 2 .6
5.5-5.9 1 .3
6.0-6.4
605-6.9 3 1.0
315 100.0%
-------
Table 10
Frequency Distribution Coh/1000'
One 2 Hour Reading
February 15, 1968 to March 13, 1968
Interval
(Coh/ 1000' t
0.0-004
0.5-0.9
1.0-1.4
1.5-109
2.0-2.4
2.5-2.9
300-3.4
3.5-3.9
4.0-4.4
4.5-4~9
5.0-5.4
5.5-5.9
6.0-6.4
6.5-6.9
No. of Cases in
Given Interval
76
122
62
22
9
7
9
4
4
315
% Total
24.11
3807)
1907 }
700
2.9 ]
2.1
2.9
1.3
1.3
100.0%
49.
62.8
2607
500
-------
50.
Relationship
2 Hour Readings vs. 1 Hour
Coh/1000'
Readings
6.
5.0
. .
.
.
. .
.
. . .
. .
. .
. . . .
. . . .
4.0
o
o
o
r-I
.......
~
8 3.0
J.I
:3
o
=:
.
.
.
. . .
.
.
.
. .
.
1:1 Relationship
r-I
. .
. . . .
2.0
.
. ~.
..
. .
. .
. .
. . .
.
1.0
. .
. . .
.. ...
......
.
1.0
2.0
3.0 4.0
2 Hour Coh/1000'
5.0
6.0
Figure
12
-------
.(\'\
a
.......
~
300
51.
Conversion Curve
~g/m3 to Coh/1000 Linear Feet
700
600
500
400
Sampling Period: 2 hours
Filter Diameter: 2.54 em,
Filter Area: 5.06 ~m'
Volume of Air: .a496 M
200
100
1.0
2.0
3.0
4.0
5.0
7.0
6.0
Coh/lOOO Linear Feet
figure 13
-------
52.
filter paper tape for two hours. It should not be used for
any other set of conditions without running check tests. Under
the conditions stated it is now possible to take United States
reporting units and convert them directly to European units.
That is, a Coh/1000' value of 3.0 in New York City represents
260 ~g/m3 in equivalent standard smoke units. A tabulation
of Coh/1000' values by 0.1 intervals is given in Table 11 with
its equivalent in ~g/m3.
Relationship Between Particulate Density
and Particle Count
Late in the study it became possible to obtain on loan
a particle counting instrument that could be used to explore
briefly the number of particles counted in size ranges related
to the Coh/1000' value obtained by standard particulate den-
sity measurementso
It was then the purpose of this short study to inves-
tigate the relationship between air borne particulate density
defined in terms of light transmittance (and/or light reflec-
tance) and the airstream particle count differentiated by size
ranges.
The work was made feasible by Particle Technology, Inc.
who supplied, on loan, their Model 140 Airborne Particle Sensor
equipped with two Model 3116 counter racks.
Instrumentation
The filter tape samples were collected by a Research
Appliance Corporation Model F2SE AISI Tape Sampler. This unit
was modified so that its pump could be shut down independent
of the remainder of its electrical system. The sampler is
equipped with a R.A.C. Spot Evaluator meter (% T or O.D.) to
provide a continuous readout of the transmittance. In addition
to the meter readout, transmittance values were read on a
Research Appliance Company Spot Evaluator, Model Noo 363.
-------
53.
Table 11
Conversion Table Coh/1000' vs. IJ.g/m3
Coh/1000' w.g/m3 Coh/1000' w.g/m3 Coh/1000' w.g/m3
.1 5 2.6 212 5.1 585
.2 7 2.7 223 5.2 600
03 13 2.8 243 5.3 620
04 20 2.9 250 5.4 638
05 25 3.0 260 505 650
06 30 3.1 275 5.6 675
07 37 3.2 288 5.7 690
.8 42 303 300 508 710
.9 50 304 317 509 728
100 57 3.5 330 6.0 745
101 65 306 345 6.1 755
102 73 3.7 358 6.2 775
1.3 79 3.8 370 6.3 790
1.4 90 3.9 385 6.4 810
1.5 100 4.0 400 6.5 835
1.6 108 4.1 418 6.6 850
1.7 117 4.2 433 6.7 865
1.8 126 4.3 450 6.8 890
1.9 135 4.4 465 6.9 905
2.0 145 4.5 480 7.0 925
2.1 155 4.6 500 7.1 945
2.2 168 4.7 519 7.2 965
2.3 179 4.8 530 7.3 985
2.4 190 4.9 550 7.4 1005
2.5 200 5.0 568 7.5 1025
-------
54.
(Reflectance values were read on a Photovo1t Ref1ectometer
Model No. 610.)
The particle counts were made using a Particle Technology,
Inc. Model 140 Airborne Particle Sensor/Photometer equipped
with two Model 3116 counter racks. Each counter rack was
capable of providing the particle counts in two adjacent size
ranges, thus the particles were counted in four different size
ranges simultaneously. The P.T.I. Model 140 Airborne Particle
Sensor/Photometer is a near forward scatter (30°-71° ha1f-
angle) electro-optical systemo This optical system minimizes
the effect of particle coloration influencing the size deter-
mination of each particle and maximizes the intensity of light
scattered by each particle.
Definitions
The following definitions have been used (in accordance
with ASTM Standards, Industrial Water: Atmospheric Analysis,
D1704, Part 23, 1965):
10
1. Optical Density (O.D.) = 10g10 ~
where 10 = the intensity of transmitted light
through the clean paper,
and I = the intensity of transmitted light
through the sample.
1 Coh Unit = that quantity of particulate matter
which produces an optical density of
0.01 on filter paper.
Smoke concentration = Coh Units per 1000 linear
feet of air drawn through
filte~ paper.
Coh/1000 ft = (O.D.t(105)
2.
3.
40
where L is the total linear feet of air. drawn
through the filter paper.
Ro
Optical Reflectance (O.Ro) = loglO 1r
where Ro = the intensity of the light reflected
from the clean paper, and
R = the intensity of the light reflected
from the samp1eo
-------
550
5.
1 RUD Unit
= that quantity of particulate matter
which produces an optical reflectance
of 0.01 due to 10,000 linear feet of
air.
6.
RUD = 0.1 Coh/1000 ft.
Program and Procedure
The particle counter racks were set to discriminate and
count all particles in the following four ranges:
003 to 1.0 ~
1.0 to 2.0 ~
2.0 to 6.0 ~
600 to 3000 ~
Because of time limitations, all particle size ranges on the
digital counters were set in accordance with the "Signal in
Volts vSo Particle Size" calibration curve supplied by Particle
Technology, Inco No attempt was made to size calibrate with
an aerosol generator according to the manufacturer's prescribed
procedures 0 The precise technique involving the use of a
pulse generator and oscilloscope was not employed to set the
discriminator voltages at the low end of the scale (003-1.0 ~)o
Outside air was drawn to the instruments through a
common mixing vessel using the pumps of the respective instru-
ments. The P.ToI. Sensor sampled at the rate of 1.0 cfm and
the R.A.Co, AISI at the rate of 0.20 cfm. The total sample
period was two hours (1 hour sample periods were tried but they
did not produce filter stains dark enough to be in the range
of reliable measurement).
Operational Procedure
Both instruments were turned on and allowed to warm up.
The pump of the AISI was turned off and filter tape advanced
to a fresh spoto The "initial" % T (after a 15 minute aging
period--see discussion) was noted on the readout meter. The
P.T.I. counter panels were set at zero and then the AISI pump
and the P.T.I. counters were turned on simultaneously. After
an elapsed time of 10 minutes, the AISI pump and the P.T.I.
-------
560
counters were turned off simultaneouslyo The % T and the
indicated particle counts were then transcribed. The counters
were reset to zero and the entire process repeated until two
hours of running time had been achieved. At the end of the
two hour period, the AISI tape was removed and read according
to standard procedure on the R.A.C. Spot Evaluator for % Trans-
mission, and on the Photovolt Reflectometer for % Reflectance.
Results
The summary of observed
in Table 12. Runs 1 and 2 are
different time base (one hour)
as to be considered unreliable
and measured results is presented
omitted since they were for a
and their spots were so light
particularly in view of the
overall sparsity of data.
The results of the regression analysis are presented in
Table 13 with selected regressions lines shown in Figures 14
and 15. It is noted that the total counts have been reduced
by a factor of 5 to make the volume sampled equivalent to that
which produced the Coh/1000'. The correlation coefficients
are all significantly non-zero with 99% confidence except for
those obtained for Coh/1000 feet vs. 1/5 (Particle count:
6-30 ~).
Discussion of Results
Variations in % T Readings
For the purpose of analysis of results, the values of
% T used to determine Coh/1000' are restricted to those mea-
sured on the R.A.C. Model 363 Spot Evaluator. This became
necessary when, in three cases (Runs 8, 9 and 12), the initial
% T was such that the final transmission value on the attached
Spot Evaluator meter fell below the 50% level thereby intro-
ducing inaccuracies. The calculated values of Coh/1000' on
the basis of % T measurement technique are given in Table 14
and shown in the scatter diagram of Figure 16.
-------
Table 12
Related Spot Evaluation and Particle Count
(2 Hour Sample Period)
Spot Particle Counts (Np) USWB @ Central Park
Evaluation 1 cfm Relative
Run @ 0020 cfm 0.3- 1.0- 2.0- 6.0- Wind Humidit)
No. Date Time %T* 1m ~ ~ ~ ~ Velocity (Average
3 3/10/69 1405-1605 94 9505 14,141 71 , 3 86 21,388 2,736 NW@ 18 MPH 40%
4 3/12/69 0925-1125 92 9305 18,200 75,887 20,030 2,306 NW@ 18 MPH 47%
5 3/14/69 0925-1125 95 95.0 12,436 45,526 9,439 960 NW@ 12 MPH 47%
6 3/17/69 0915-1115 78 8300 58,576 185,942 24,977 2,137 SW @ 9 MPH 31%
7 3/17/69 1335-1535 88 9005 30,540 146,458 32,550 2,336 SW @ 11 MPH 23%
8 3/18/69 0600-0800 55 6700 95,576 353,027 44,167 2,033 SW @ 6 MPH 45%
9 3/18/69 0815-1015 56 6905 96,791 367,747 49,928 2,377 SW @ 5 MPH 32%
10 3/19/69 1206-1406 84 86.5 35,724 150,986 29,470 1,745 NE @ 13 MPH 40%
11 3/20/69 1135-1335 82 84.5 69,047 154,071 23,970 1,442 SSE @ 5 MPH 68%
12 3/20/69 1900-2100 71 77.5 24,365 98,949 21,532 1,313 SE @ 11 MPH 86%
13 3/21/69 1720-1920 95 95.5 12,579 65,832 22,309 1,031 W@ 12 MPH 42%
14 3/22/69 1520-1720 94 94.5 13,856 73,154 26,045 1,389 NW@ 14 MPH 43%
*Va1ues from R.A.C. Spot Evaluator Model 363.
V1
"'-J
.
-------
Table 13
Linear Regression Analysis
Coh/1000 Ft. VSo Particle Count (Np/5)
Standard Corre-
Particle No. 1ation
Size of Average Deviation C oe f - Scat-
Range Cases Coh Np/5 G""Coh UNp/5 ficent ter*
003-1.0 \-l 12 2.13 8,030 1.86 6,109 0.887 0.86 Coh =
10** 2.08 8,538 1.98 6,569 0.950 0.61 Coh =
1.0-2.0 \-l 12 2.13 29,816 1.86 20,866 0.929 0.69 Coh =
10** 2.08 30,871 1.98 22,390 0.995 0.20 Coh =
2.0-6.0 \-l 12 2.13 5,430 1.86 2,091 0.905 0.79 Coh =
10** 2.08 5,435 1.98 2,236 0.910 0.82 Coh =
6.0-30 \-l 12 2.13 363 1.86 112 0.259 1. 79 Coh =
10** 2.08 363 1.98 118 0.368 1.84 Coh =
*Scatter = ~Coh J 1 - r2.
**Runs 7 and 12 omitted.
Regression Equation
y=mx+b
0.270 x 10-33 Np
00286 x 10- Np
0.834 x 10-~ Np
0.879 x 10- Np
0.804 x 10-3 Np
0.804 x 10-3 Np
0.430 x 10-2 Np +
0.617 x 10-2 Np -
- 0043
.362
.362
.634
2.24
2.29
.563
.160
\..JI
00
.
-------
59.
Figure 14
Regression Lines for Coh/1000 Feet and Particle Count (Np/5)
For Selected Particle Size Ranges
a. Particle Size Range: 0.3-1.0 ~
~
Coh = 0.27 x 10-3 ~ - 0.04
Coh = 0.29
x 10-3 ¥
- - - - 12
10
- 0.36
R.H. = 86%
+
cases
cases
+
R.H. = 68%
2
4
16
18
20
6 8 10 12 14
Particle Count (Np/5 x 103)
b. Particle Size Range: 1.0-2.0 ~
6
Coh = 0.83
x 10-4 ~ - 0.36
0.88 x 10-4 ~ - 0.63
R.H. .. 86%
+
---- 12 cases
10 c~ses
= 68%
1
1
2
3 456 7
Particle Count (Np/5 x 104)
8
9
10
-------
600
Figure 15
Regression Lines for Coh/1000 Feet and Particle Count (Np/5)
For Selected Particle Size Ranges
a. Particle Size Range: 2.0-6.0 ~
6
o
Coh = 0.80 x 10-3 ~ - 2.24
R.H. = 86%
+
Coh = 0.80
x 10-3 ¥ - 2.29
- - -- 12 cases
10 cases
R.H. =
X R.H. = 23%
3 456 7
Particle Count (Np/5 x 103)
b. Particle Size Range: 600-3000 ~
1
2
8
9
10
6
5
o 0
Coh = 0062 x 10-2 ~ - 0.16
....'"
.....
.....
......
""
""
",,""
",,"toh = 0.43 x 10-2 ¥
"" "" + 0.56 2
"" - - -- 1 cases
......
~.
10 cases
~ 4
Q)
rz..
g 3
o
~
........
.g 2
u
R.Ho+ 86%
= 23%
1
1
2
3 4 5 6 7
Particle Count (Np/5 x 102)
8
9
10
-------
Table 14
Calculated Values
Coh/1000' and RUD
Run No. Coh/1000'* RUD Coh/1000'**
3 0.6 4.5 0.6
4 0.8 6.5 0.7
5 0.5 5.0 0.9
6 2.5 18.0 2.3
7 1.3 9.6 1.1
8 5.9 39.0 4.9
9 5.7 36.0 4.0
10 1.7 14.0 1.5
11 2.0 16.0 2.0
12 3.4 25.0 2.8
13 0.5 4.5 0.6
14 0.6 5.5 0.8
*Va1ues from R.A.C. Spot Evaluator, Model 363.
**Va1ues from attached readout meter.
61
-------
62.
Figure 16
Comparison of Values of Coh/1000 Feet
As a Function of Evaluation Technique
6
o
o
~5
\0
C")
.-I
Q)
"0
o
~
Coh363= 1.~5Co~oO. - 0.37
u 4
~
~
o
.j.J
co
~
.-I
~ 3
~
.j.J
o
0..
rJ)
.......
1:1 Relationship
.j.J 2
Q)
Q)
~
o
o
o
.-I
:c 1
o
u
1 2 3 4
Coh/1000 Feet (Continuous Read-out Meter)
5
-------
63.
It was noted earlier that the AISI tapes were aged
under the light source incorporated in the AISI Sampler for
15 minutes prior to the start of the air flow pumpo This
practice was begun when it was observed that there was invar-
iably an initial drop of up to 4-5% T that bore no apparent
relationship to the particulate loading of the incoming air
stream. Were this initial drop in % T not discounted, the
calculated values of the Coh/1000 feet would have been dis-
portionate1y high. It is not at all clear at this stage as
to the cause of this phenomena. Separate tests conducted
with a tape alternately humidified and desiccated suggest that
the "initial drop" may be, at least in part, due to realign-
ment of filter fibers due to the heat generated by light
source as a result of desiccation and/or fiber roasting. It
is of further interest to note that once the filter spot was
obtained by two hours of sampling, there were only rsndom,
insignificant variations in the value of the final % T when
evaluated on the Model 363 Spot Evaluator immediately after
the test completion, one hour later, 24 hours later and after
24 hours of desiccation (all sections of record filter tapes
were stored in plastic "Baggies" to minimize external
contamination) 0
Apparent Effect of Humidity
It was obvious, even as the tests were being conducted,
from preliminary scatter diagrams maintained for the values of
Coh/1000 feet versus particle counts that excellent relation-
ships were developing. Except for Runs 7, 11 and 12, there
was a strong linear relationship particularly evident in the
size ranges 0.3 to 1.0 ~ and 1.0 to 2.0 ~ (see Figure 14). In
all instances (disregarding size range 6-30 ~) Run 11 showed a
very high Coh/1000 feet value for a markedly low particle
count. Except in the 0.3 to 1.0 ~ size range, Run 7 showed a
high particle count for a comparatively low Coh/1000 feet
value. This latter feature was also true of Run 11 in the
-------
64.
003 to 100 ~ size range. The only Lmmediately common denom-
inator between these anomalous points seems to be the value
of the relative humidity. From Table 12 it is found that Run
7 had an extremely low relative humidity average, 23%, while
Runs 11 and 12 had high relative humidity averages, 68% and
86% respectively. The relative humidity for all other runs
averaged between 32% and 47%.
Junge5 notes that "as the relative humidity increases,
the aerosol particles gradually change both in size and
physical properties, and finally beeome cloud and fog drop-
lets 0.. They pick up other material from the surrounding air
by a variety of processes and thus their composition and
physical structure changes continually." He further notes
that below 70% relative humidity a noticeable fraction of
particles behaved as though they were dry, but above 70% they
soon assumed the properties of droplets. Similar observations
of effects of relative humidity on light scattering properties
of particles have been recently reported by Lundgren, et al.6
It can be surmised in the case of Run 11, where the
relative humidity is 86%, that "large" droplets have been
formed which have swept up and coalesced and/or agglomerated
smaller particles thereby producing a relatively low total
particle count in all size ranges on the P.ToI. instrument.
However, when these aggregates are trapped on the AISI tape
their moisture is evaporated and the smaller individual par-
ticles return to reduce the light transmission. Similar rea-
soning can also be applied to Run 11. The relative humidity
has begun to increase the size of the submicron particles
substantially increasing the count in the 0.3 to 1.0 ~ size
range (but not as yet in the 1.0-2.0 ~ range); however when
impinged on the filter paper and dried out, they return to
their former smaller size, apparently ineffective in reducing
light transmission.
The case of Run 7, where the relative humidity was
obviously low, cannot be explained by any of the above
-------
65.
reasoning. If one assumes, however, that trapping of particles
on filter paper is a process of polymerization which is related
to cohesive properties of the particle surface, then one can
speculate that the low relative humidity results in a low
cohesive force thereby retarding the polymerization process
and effectively producing a low value of Coh/1000 feet for a
relatively high particle count. As has been observed, this
phenomena would be most active in the larger particle size
ranges where the surface area is greatest and least effective
in the smallest size range.
Particle Size Range Responsible for Coh/1000 Feet
From Table 13 and Figure 14 the highest correlation and
least scatter of data points is observed, without question,
for particles in the size range from 1.0-2.0~. In a separate
sequence of tests 25 readings were taken at one minute inter-
vals with the P.T.I. counter. Ten readings were taken with
the counter racks set to read in the following ranges:
0.3 -
0.8 -
2
6
0.8 ~
2.0 ~
6 ~
- 30 ~
These were followed Lmmediately by twelve readings in the stan-
dard ranges:
003 -
1.0 -
2
6
100 ~
2.0 ~
6 ~
- 30 ~
After the above, three additional readings were taken in the
ranges:
0.3 - 0.8 ~, etc.
The average counts/cubic foot/minute are tabulated below:
-------
66u
Size Range Np/min Size Range Np/min
0.3 - Ou8 ~ 18 0.3 - 1.0 ~ 204
008 - 2g0 ~ 1114 1.0 - 2.0 ~ 980
2 6 ~ 150 2 6 ~ 166
6 - 30 ~ 8 6 - 30 ~ 11
It is observed that the particle count in the range from 0.3 -
1.0 ~ is over 100 times that of the count of 0.3 - 0.8 ~.
Furthermore it is noted that when the difference between these
two counts is subtracted from the count in the 0.8 - 2.0 ~
range, the new count is consistent with the actual count mea-
sured in the 1.0 - 2.0 ~ range. This may imply that the strong
correlation between Coh/1000 feet and particle count earlier
observed in the 0.3 - 1.0 ~ size range may in fact be generated
by those particles in the 0.8 - 1.0 ~ range. Unfortunately a
similar study was not conducted between the 1.0 - 2.0 ~ and
2uO - 6.0 ~ range bands. It is quite possible that one would
find the dominate correlation results from those particles
close to 2.0 ~ in size.
Comment
On the basis of limited data, there is a very strong
and significant correlation between the actual count of par-
ticles in various size ranges and the commonly used parameter
Coh/1000 feet.
The correlation is greatest and most significant when
the Coh/1000 feet value is related to total particle count in
the size range from 100 - 200~. There is strong evidence to
believe that the relationship can be still more firmly estab-
lished if the lower limit of the size range is reduced to 0.8 ~
and the upper limit increased to some value between 2 - 6 ~.
It appears that this relationship between Coh/1000 feet
and particle count is most stable for ambient air relative
humidities between 31% and 47%. Aberrations occur at higher
and lower values of humidity, but it is suggested that this
relationship too, could be firmly established with sufficient
-------
670
controlled experiments. It is quite possible that the problem
may be alleviated simply by establishing a standard humidity
control value on the incoming air stream.
Summary Discussion
This project has reported a series of operations that
have worked sequentially from a basic concept of reporting
equivalent smoke concentration based on light reflectance
techniques adapted by GECD and Great Britain to the ultimate
conversion of United States reporting units of particulate
density into the same equivalent reporting units. In brief,
using techniques, methodologies, curves and equations presented
in this report it is now possible for New York City (and prob-
able with only minor error for any United States reporting
unit) to take routine readings of particulate density in Coh/
1000 feet and convert these to ~g/m3 as used in European
practice. The reverse is also trueo That is, those in Great
Britain or other European countries who wish to make compari-
sons of particulate concentration in the respective atmospheres
of United States and European locations may use the curve pre-
sented here for converting levels of particulate concentration
and comparing them in epidemiological studies of the effects
of particulate concentration. It has been established that
the New York City curve is, in form, practically identical to
the International Smoke Calibration Curve and to the British
Standard Smoke Calibration Curve. Positioning of the curve
is accomplished according to procedures proposed by GECD in
1964 with the index point established at 25% darkness index.
The postioning value for the International curve is 20 ~g/cm2,
for Great Britain 17 ~g/cm2, and for New York City 30 ~g/cm2.
For New York City major points of conversion include:
1.0 Coh/1000 feet = 57 ~g/m3
2.0 = 145
3.0 = 260
-------
68.
4.0 Coh/1000 feet
500
6.0
7.0
~g/m3
= 400
= 568
= 745
= 925
Curve divergence does not permit carrying the comparison to
higher Coh/1000 feet levels, or greater particulate concentra-
tion in the atmosphere, without special calculations. It has
also been established in this study that there is a slight
divergence from a 1:1 relationship between % light transmission
and % light reflection.
Likewise it has been shown that one hour particulate
density readings are not directly representative of two hour
readings and can not be extrapolated in the comparison of
readings taken with a substantially different total flow pass-
ing through a spot on a tape sampler.
It has been established on the basis of limited data
that there is a strong and significant correlation between the
actual particle count of particles in various size ranges and
the parameter of concentration, Coh/1000'.
It is believed that for preliminary estimating purposes,
at least, the New York Calibration Curve may be applied in the
interpretation of AISI data from other United States cities.
This belief is based on the established similarity of smoke
curve data from various cities in Great Britain. For more
preci~e comparison it is necessary only to establish with rela-
tively few measurements the proper index point for a calibra-
tion curve for any city, and the procedure for establishing
that index point has been presented in this report.
Acknowledgments
The staff of the Research Division, New York University,
School of Engineering and Science has devoted an unusual amount
of time (not always compensated) in pursuing the work of this
project. The staff included, in addition to the Project
-------
69.
Director, Jack Golden, Research Scientist, Edward J. Kaplin,
Research Scientist, and Raymond Broglie, Research Technician.
They are all to be commended for the interest they have shown.
The Bureau of Disease Prevention and Environmental
Control, National Center for Air Pollution Control, Public
Health Service, has contributed both funds and equipment to
this projecto These are gratefully acknowledged.
The support of the several project monitors including
Dr. Roy McCaldin, John O'Connor and Ferris B. Benson has been
most helpful.
We wish to extend our thanks particularly to the
Ministry of Technology, Warren Spring Laboratory (formerly
DSIR), for the continued advice, assistance and checking pro-
vided throughout this prqjecto Through the cooperation of the
staff including Dr. Marjorie Clifton, Dr. David Gall, and Mr.
Desmond Bailey it was possible for us to carry out the calibra-
tion procedures duplicating British practice as closely as
possible.
The particle size studies could not have been under-
taken without the cooperation of Particle Technology, Inc.
The loan of essential equipment by that company is greatly
appreciated.
Needless to say, the efforts of the project secretary,
Miss Patricia Twomey, have been invaluable. Her services are
greatly appreciated.
-------
70.
References
1.
Hill, A. S. G., "Measurement of the Optical Densities of
Smoke Stains on Filter Papers," Trans. Farad. Soc.,
32, 1125 (1936).
2.
Waller, R. E., "Experiments on the Calibration of Smoke
Filters," APCA Journal, 14, 8, 323 (August 1964).
, Methods of Measurin Air Pollution, Chapter
Two, "Smoke," Pu ~cat~on No. , Organ~sat~on for
Economic Co-operation and Development, Paris (1964).
3.
4.
, "Estimation of Smoke in the Atmosphere
by the Smoke Filter Method," Internal Report No.
RR/AP/84, Warren Spring Laboratory, Stevenage, Great
Britain (March 1965).
5.
Junge, C. E., Air Chemistry and Radioactivity, International
Geophysical Series, Volume 4, Academic Press (1963).
Lundgren, D. A. and Cooper, D. W., "Effect of Humidity on
Light-Scattering Methods of Measuring Particle Concen-
tration," APCA Journal, 19,4, 243 (April 1969).
6.
-------
A-l
APPENDIX A
Apparatus Used to Determine
the
Form of the Smoke Calibration Curve
In order to determine the shape or form of the smoke
calibration curve, the following equipment was used:
a) Seven 1" diameter brass filter holders.
b) Seven filter holder clamps.
c) Seven "Superior" dry gas meters adjusted to read
to 0.1 ft3.
d) Seven 1.3 cfm vacuum pumps.
e) Fractional timer unit with sampling time ratios
(over a 15 minute cycle) of 1/9, 1/6, 1/3, 1/2
and 1 (i.e., continuous operation).
f) Yale Regular Point Hypodermic Needles (20 gauge).
g) Wide-mouth glass mixing jug--capacity 5 liters.
h) Glass funnel--4" diameter inlet and a 1/2" bore
glass outlet.
i) Whatman No.1 filter papero
-------
5.
6.
7.
8.
9a.
10.
B-1
APPENDIX B
Standard Operating Procedure
for
Photovolt Reflectometer Model 610
1.
20
3.
4.
Insert plug of search unit into instrument panelo
Insert plug of instrument into 110 volt AC outlet.
Insert green TRISTIMULUS filter into search unit.
Place 1 clean sheet of Whatman No.1 filter paper on
enamel plaque.
Place search unit on top of Whatman Noo 1 filter paper.
Switch "On" knob on the instrument panel.
Allow 15 minutes for instrument to warm up.
Set reading to 100%. R by adjusting course and fine dials.
After each sample (for small numbers of samples) check
100% R by inserting original clean Whatman No.1 filter
paper. Adjust if necessary with fine adjustment dial.
For numerous samples, check arbitrarily and adjust, if
necessary, with fine adjustment dial.
After completing all samples, disassemble unit in reverse
of steps 1 through 6 and clean components.
b.
-------
C-1
APPENDIX C
Apparatus Used to Determine Scale
To be Attached to the Calibration Curve
a) One 1.3 cfm vacuum pump orificed to produce approximately
.073 ft3/min.
b) High volume suction pump, of normal rating 50-200 ft3/hr.,
capab1e30f drawing the air through the apparatus at 1100-
4400 ft /day.
c) Two "Superior" dry gas meters adjusted to read to 0.1 ft3.
d) Two filter clamps 4" IoD.
e) Hard glass funnel 8.5" IoD. (to produce inlet velocity of
approximately 2 em/see).
f) Glass tubing of 1/2" I.Do with all bends having an 8"
radius.
g) One 1" brass filter c1ampo
h) Desiccator containing magnesium perchlorate for drying
filter medium.
i) Electronic balance capable of measuring 001 mg.
j) Filter Medium - Glass Fiber Filter (Type E--Ge1man Instru-
ment Company).
-------
D-l
APPENDIX D
Instruments and Procedures for Comparison
of
Light Transmission and Reflection
Reflectance values were read on a Photovolt Reflec-
tometer, Model Number 610. The instrument was checked against
a standard after each reading. The standard was one clean
sheet of Whatman No.1 paper placed on a white enamel plaque
supplied by the manufacturer. In all cases a green tristimu-
Ius filter was used in the search unit. The initial reading
produced by the standard was 100%.
Transmittance vlaues were read on a Research Appliance
Company Spot Evaluator, Model Number 363. The instrument is
reset, if necessary, to 100% before each sample is read. This
is done on a clean piece of Whatman No.4 papero
------- |