EPA (Environmental Protection Agency)
Method Study 17, Method 607 (Nitrosamines)
Southwest Research Inst., San Antonio, TX
• #
Prepared for
Environmental Monitoring and Support Lab.
Cincinnati, OH	,
Jun 84

-------


PB84-207646
'¦C",\
jt.;:;r:
EPA-600/4-84-051
June 1984
t—-
.•;• r:UNS
?A Gi\
L AST UNt JTZ*—
Cjf T.r/7
-j^j-	
___ 11
3/8'
EPA METHOD STUDY 17
METHOD 607 (NITROSAMNES)
by
i
John D. Millar
Richard E. Thomas
Herbert J. Schattenberg
Southwest Research Institute
San Antonio, Texas 78284
... -l
0-1 '"8"
1 'r-
Contract 68-03-2606
Robert L. Graves and Edwarc
i
L. Berg, Project Officers
Quality Assurance Branch
Environmental Monitoring and Support Laboratory
I U.S. Environmental.Protection Agency
'	Cincinnati, ;0hio 45268
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY '
| OFFICE OF RESEARCH AND DEVELOPMENT
, U.S. ENVIRONMENTAL:PROTECTION AGENCY
I	CINCINNATI, OHIO 45268
L
—
E?A Form 2350-4 (4-CO)
IPnCVIOUSLV CIS. L'PA FCNM 2S;
PAC1" NI.'.YGEm
HHODKCO B»
NATIONAL TECHNICAL
ilNFCRMATION SERVICE
i - Hi. OEPAQlkURT Of CGHUERCE
. sriincfiao.««. Z2i6i
BOTTOM OF .
IMAGE AnfA;
-e- Oi/TSIfJE
^¦DhViHNSIOfJ
FOII TABLES
.1 AMD ILLUS-
TRATIONS
>S GOvfuuMrfii	Ot:r»C£
i t l.-.	d'C Ir*-^	IrLtoii¦
TYPING GUIDE SHEET , . p
. .•	V	W'tk
—	'	^ - \	^	r	t'jb * £ ^	J

-------
TECHNICAL REPORT DATA
If lease read Inunctions on the reverse before completing)
1. REPORT NO. 2.
EPA-600/4-S4-051
3. RECIPIENT'S ACCESSIOf*NO.
PB8 k 207 a lift
4. TITLE ANO SUBTITLE
EPA Method Study 17, Method 607 (Nitrosamines)
5. REPORT DATE
June 19G4
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
John D. Millar, Richard E. Thomas, Herbert J.
Schattenberg
a. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORG "VNIZATION NAME ANO ADDRESS
Southwest Research Institute
6220 Culebra Road, P.O. Drawer 28510
San Antonio, Texas 73284
10. PROGRAM E'.EMENT NO.
CBS DIA, BEBIC
11. CONTRACT/GRANT NO.
68-03-2606
12. SPONSORING AGENCY NAME ANO AOORESS
Quality Assurance Branch
Environmental Horn toring and Support Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final 9/78-12/81
14. SPONSORING AGENCY CODE
EPA-ORD/ 06
15; SUPPLEMENTARY NOTES
16. ABSTRACT
"this report describes the results obtained and data analyses from an interlaboratory
evaluation of ZPA Method 607 (Nitrosamines). The method is designed to analyze for
three nitrosamines, N-nitrosodimethylamine, N-nitrosodi-n-propylamine, and N-nitroso-
diphenylimine, in water and wastewater.
The study design ranuired the analyst to dose six waters with each of six mixtures of
the three nitrosamines. The six dosing levels represented three Youden pairs, one
each at a low, an intermediate, and a high level. A total of 17 laboratories
participated in the study.
The method was studied to estimate the accuracy and precision that can be expected,
Including effects on the accuracy and precision of analysis of different matrices.
In addition, results of method detection limit and analytical curve studies and
qualitative assessments of the method based upon comments by the participating
laboratories are included.
17. KEY WORDS ANO DOCUMENT ANALYSIS
a. DESCRIPTORS
b.lOENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group



18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLMSS (Tf,(s Report)
Unclassified
21. NO OF PAGES
74
20. SECURITY CLASS (Thispage/
Unclassified
22. PRICE
SPA Form 2220*1 (9-73)

-------
	i n i ¦	¦ I	I HM	"I ^'fl «w ¦*! II ¦ I				——»—1¦¦mlWfgiW
c: f-'l r.-l 'J*
n.v or
"\ • •-*

DISCLAIMER
~T
¦ rg	
iifr.iN'
lasr L.iNr-»^
OF "I EXT
The information in this document has been funded wholly or in part by
the United States Environmental Protection Agency under Contract 68-03-2606
to Southwest Research Institute. It has been subject to the agency's peer
and administrative review, and it has been approved for publication as an
EPA document. :
6-1/2'
y-i -'8'
I
I
3/8"
L

EPA Form 235(M (4-BOI
(PKEVIOUSLV ON KPA TOHM 28?)
. ,;'V\' ¦. .v.;-y •
Fi	11" ^
PAGi: ,\IU.VI8L,'1
eOTTOM OF
IMAGE AHEA
0'JISIDE
rC-JMhNSION
.'OR TABLES i
I AND ILLUS-
TRATIONS
K U j te-.CHMlfNT PH.UTlNO OFF.qC l
-------
tryfAW* WJftT--'.-'*"-1

.z
(}£rjW
L/oT LINES
Of 7 = X T
FOREWORD	-;
•	!	t
Environmental measurements are required to determine the quality of
ambient waters and the character of waste effluents. The Environmental
Monitoring and Support Laboratory (EMSL)-Cincinnati, conducts research to:
•	develop and evaluate techniques to measure the presence and
concentration of physical, chemical, and radiological pollutants in
water, wastewater, bottom sediments, and solid waste,
:	i
•	investigate methods for the concentration, recovery, and
I'..;-	 loentification of viruses, bacteria and other microorganisms 1n ~
water,i	j	j
•	conduct studies to determine the responses of aquatic organisms to \
water quality,	!	j
•	conduct an agency-wide quality assurance program tc assure
standardization and quality control of systems for monitoring water *
and wastewater.	|	\
This publication reports the results of EPA's Interlaboratory study 17
for the following priority pollutants, which are analyzed using EPA Method
607 (Nltrosamines):
i
N-nitrosod1methylamine
N-n1trosodi-n-propylamine
N-nltrosodi phenyl amine
Federal agencies, states, municipalities, universities, private
laboratories, and industry should find this evaluative study of assistance
1n monitoring and controlling pollution 1n the environment.
R. L. Booth
Director, EMSL-C1nc1nnat1
J 3/8"
i:
L.
EPA Form ?3[>0-4 (4-80)
iPnrviousi-Y cin. lpa torm ?q7)

PAGE fMUMCC!)
13CTTO.M OF
IMAGE
	OUi'JIJiE
DiMENSlON '
FOR TABLES
	I AImO ILLUS-
TRATIONS

-------

ABSTRACT
This report describes the results obtained and data analyses from an
Interlaboratory evaluation of EPA Method 607 (Nitrosamines). The method is
designed to analyze for three nitrosaminfis, N-nitrosodimethylamine, N-
nitrosodi-n-propyl amine, and N-nitrosodiphenylamine, in water and
wastewater. As tested here, the method utilized three GO-mL extractions
with dichloromethane, cleanup/separation on an alumina column, and
injection into a gas chromatograph equipped with a nitro9en-phosphorus
detector.	i
U«oT . I.'
\jr < \
The study design required the analyst to dose six waters with each of
six mixtures of the three nitrosamines. The six dosing levels represented
three Youden pairs, one each at a low, an intermediate, and a high level.
The six waters used were a laboratory pure water, a finished drinking
water, and a surface water, all collected by the participant, and three
low-background industrial effluents furnished by the prime contractor. A
total of 17 laboratories participated !in the study.
'	I
The method was studied to estimate the accuracy and precision that can
be expected, including effects on the-accuracy and precision of analysis of
different matrices. In addition, results of method detection limit and
analytical curve studies and qualitative assessments of the method based
upon comments by the participating laboratories are included.
Th*s report covers work accomplished over the period from September
1978 to December 1981 under EPA Contract 68-03-2606.
: jfr-
I.
3 '3"
EPA For.n 2350-4 (4-SfJj
[PUCVIC USLY CIN. I.PA FORM ?.B7|
PACE NUM8EH
BO rTO.Yi OF
IMAGE AHEA;
DU1MDL:
•DIMENSION
FOR TABLES
AND ILLUS-
TRATIONS '
£!J<; CC Vt-nNMlNT kfl MTiNG	JSdJ-Cch-e.-)*
TYPING GUIDE SHEET

-------

' H o1
ifcCW
.Asr Lir:&rsK
>r ro.r
COflTEMTS
1 j	Page
i
i
Foreword 				111
Abstract 				1v
Tables 		v1
Acknowledgmcnt	|		vi1
1.	Introduction 	;		1
2.	Summary.	:		4
3.	Description of Study 				6
;r	 Test design 				-				6 -
Preparation of samples, wastewater selection and
procurement 		7
Analysis and reporting 		9
Distribution of samples ...:		10
Stability of samples 		10
4.	Treatment of Data 	1		12
Preprocessing	1		12
Rejection of outliers	;		13
Youden's laboratory ranking procedure 		13
Tests for individual outliers 		14
Statistical summaries				14
Statement of method accuracy 		16
Statement of method precision 		18
Comparison of accuracy and precision across
water types	;		19
5.	Discussion and Conclusions . ..f		25
Accuracy of the method	i		25
Precision of the method ...!r		25
Comparison across water types 		30
Method evaluation		 30
References
Appendices
A* Raw data
B. EPA Method 607 (Nitrosaminesj
—	(	
i_ r!l _ i	
EPA l:orm 23r.0-4 («BO>
{PREVIOUSLY ON. ~PA TOPM 2B7)
38
39
48
™%
:v.V '
vv.'.Myj
PAGE NUVCCR
BvjTTGiVJ Or
l\1*GE Af*E A,
OUTSIDE
DIMENSION
FOR TACIES
I AND ILLUS-
TRATIONS
TV U S. GOVERNMENT PfllNTlKG OFFICE 1960^860-639
TYPING GUIDE SHEET



.J
'X
' vi
M

-------
I,i,jm niynf—u^yi,iiw	"'4 >J':- ¦«"W'»W"""T!JV-1" «';y1"» *:i •¦ ¦¦ yi	, i iW.jii. ' ""i ¦ *?'[¦ " '¦¦¦¦;» wm
TABLES
i
Number	Page
1.	Participating Laboratories 	 3
2.	Accuracy and Precision Equations 	 5
3.	True Concentrations in Study Samples 	 8
4.	Statistical Summary for N-N1trosodimethylamine Analyses
by Water Type	 26
- 5. Statistical Summary for N-N1trosod1-n-propylam1ne Analyses
by Mater Type	 27
6.	Statistical Summary for N-Nitrosodiphenyl amine Analyses
by Water Type	 28
7.	Estimated Percent Recovery for Various Water Types at Midrange
Concentration			 29
i
8.	Estimated Ovorall Percent Relative Standard Deviation for
Various Water Types at Midrange Concentration 	 31
9.	Estimated Single-Analyst Percent Relative Standard Deviation
for Various Water Types at Midrange Concentration 	 31
10.	Effect of Water Type on N-N1trosod1methylamine Analysis 	 32
l
11.	Effect of W2ter Type on N-N1trosod1-n-propylamine Analysis 	 33
i	!
12.	Effect on Water Type on N-Nitrosod1phenyl amine Analysis	 34
i	I
13.	Method Detection Limits and Lowest Concentrations Used
In Study	i	 35
i	'
A-l. Raw Data for N-N1trosomethylamine Analysis by Water Type 	 39
I	!
A-2. Raw Data for N-N1trosod1-n-propylamine Analysis by Water Type .. 42
A-3. Raw Data for N-Nitrosodlphenyl amine Analysis by Water Type	 45
WQM
LAi>f l.{ML
0«' K/.T
I-
3/8'
EPA Potm 2350-4 M-80)
t
(PHCVIOUSLY CIN. HPA rOHM 237)
TAGL' WUMGi-'f!
BOVTC.iV Ot
IVAGc AHf I\,
OUTSiDE
l)iME\-SlC.N
TOR lASLCS'
I AND ILL'JS-
TRA~:ONr.
¦U U.S. GOVKHNVfF.NT PRINTING Ol'FICE 1bP(V£e»-6X>
TYPING GUIDE SHEET -
l n i i *<1 *
...	„ 1

-------
©
J	'	c	'	-	'	'
.,. • - —-	.			¦	mfvurr**"""'**-"'^

 U i 00/E=i!.\UUT pniNllhCi OTIC!-' WcSiUCMSi

-------
SECTION 1
INTRODUCTION
j EPA first promulgated guidelines [1] establishing test procedures for j
i the analysis of pollutants 1n 1973, following the passage of the Federal :
Water Pollution Control Act in 1972 by Congress. Pursuant to the amendment j
and publication of these guidelines, EPA entered Into a Settlement j
Agreement—the Consent Decree—which required the study and, 1f necessary, :
regulation ot: 65 "priority" pollutants and classes of pollutants of known j
i or suspected toxicity to the biota, j Subsequently,.. Co/igress passed thej
[clean Water Act of 1977 [2], mandating the control cf toxic pollutants
j discharged into ambient waters by Industry.
In order to facilitate the Implementation of the Clean Water Act, EPA
selected 129'specific toxic pollutants, 113 organic and 16 Inorganic, for
Initial study.! The organic pollutants were divided into 12 categories
based on their chemical structure.) Analytical methods were developed by
EPA for these 12 categories through 1n-house and contracted research.
These analytical methods may eventually be required for the monitoring of
the 113 toxic pollutants in Industrial wastewater effluents, as specified
by the Clean Water Act of 1S77.
As a logical subsequence to the work that produced proposed EPA Method
607 (Nltrosamlnes) [3], an 1nterlaboratory study was conducted to test the
validity of the proposed method. Thl^ report describes the work performed,
presents the data acquired, and gives the conclusions drawn from the
collaborative effort.
i
The three compounds undergolngianalyses 1n the 1nterlaboratory study
were N-n1trosod1methylamine (MDMA), N-n1trosod1-n-propylamine (NDPrA), and
N-nitrosodlphenyl amine (NDPhA).
The laboratories participating 1n this study were the 17 lowest
bidders from the 11st of qualifying {laboratories that responded to the
request for bids. Qualifications of the laboratories were established by
review of information submitted on past experience and available equipment.
Previous experience with the laboratories was also included in the

I
53/8'
1

W:
CPA Form 23WM W-30)
IfRCVIOUSLV CIN. t'PA FORM 207)
PAGE NUMUEH
	^
ivys tiOYlrtl.'ll.KT 'Pl'i IiHf»	i:-U' CC '"o-
OOTTOM OF
IMAGE AP.EA;
OUTSIDE
DIUCNSION
FOR TAHLtS
AMD ILLUS-
TRATIONS
TYPING GUIDE SHEET

-------



~evaluation where applicable. The participants were selected to be typical .;
: of laboratories that would be using the methodology in its intended
application. The niHnber of participating laboratories was cut from 20 to
; 17 for budgetary i-easons, and no volunteer laboratories could be acquired.
! The laboratories are identified by number 1n this report and no correlation
--between the identifying number and the order of laboratories in the list of,'
! participating laboratories should be presumed.	i
v- ,2
%
* *
:%
-"5S
•cr u:

' ^3/a-
L_ 1_1
t
-t
M/ii"
I
.J	
fcPA Form 23(4-80)
(PMCVIOUSLY CtN. CPA FORM ?D>)
	SEE-
PAGE NUMBIT.
BOTTOM OP
•	IMAOE AREA;
OUICjDE
*	DIMENSION
FOH TABl.tU
	I AND III US-
i Nations
ft \j£	r-i\ir.!i'i»oof.cJCe isso-feGew

TYPING GUIDE SHEET
.!'¦&>* >iQ)£B-.,Vn£y



-------

.TZ'i Of-
TABLE 1.. PARTICIPATING LABORATORIES
• Analytical Development Corporation
I 1875 Mil low -Park Way
j Monument, Colorado 80132
! Analytical Research Laboratories
; 160 Taylor Street
j Monrovia, California 91016
i	:
j Battel!e	:
! Columbus Laboratories. • ....	
j 505 King Avenue
Colunbus, Ohio ; 43201
i
1
Blospherics, Inc.
4928 Wyaconda Road
Rockville, Maryland 20852
t
Camp, Dresser and McKee, Inc.
Environmental Sciences Division
6132 West Fond du Lac Avenue
Milwaukee, Wisconsin 53218
!
Environmental Research Group
117 North First
Ann Arbor, Michigan 48104
I
Environmental Science and Engineering,
Inc.
P.O. Box ESE 32602
Newberry Road (5 ml west of 175)
Gainesville, Florida 32604
i
Radian Corporation
8500 Shoal Creek Boulevard
Austin, Texas 78766
1
Raltech Scientific Services, Inc.
3301 Kinsman Boulevard
P.O. Box 7545
Madison, Wisconsin 53707
I ! I !.M:
7x r
; r--.: i	
i_ r _ 1	
EPA Form 23'JC 4 (4-&I)
(PHEV'IOUSLY CIN. El'4 FOHM 1*1)
Recra Research, Inc.
Ill Wales Avenue
P.O. Box 448
Tonawarida, New York 14150
Southern Research Institute
2000 Ninth Avenue South
Birmingham, Alabama 35205
„SRI International _ . 	
333 Ravenswood Avenue
Fenlo Park, California 94025
Technical Services, Inc.
103-7 Stockton Street
Jacksonville, Florida 32201
Texas Instruments, Inc.
13500 North Central Expressway
P.O. Box 225621
Dallas, Texas 75265
Versar, Inc.
6621 Electronic Drive
Springfield, Virginia 22151
West Coast Technical Service,
Inc.
17605 Fabrlca Way, Suite D
Cerrltos, California 90701
Wilson Laboratories
528 North Ninth
Sallna, Kansas 67401
PAGE NUMBf-r)

BOTTOM OF
I MAG!: Ar-iCA;
OUTSIDE
DIMENSION
POH TASLEa
I and ILL'. >-
¦ TKATIONS
£ 'jS 00MOillKG Of !'«U-
. TYPING GUIDE SHEF.T ;
I
i
L. .


-------
SECTION 2
SUMMARY
As a result of the collaborative study conducted and the data
analysis, the following conclusions Cun be drawn concerning EPA Method 607
(NUrosamines).	J
•	The accuracy of the methcd could be expressed as a linear function
of the true concentration. The regression equations for accuracy
are shown in Table 2.	|
x			^ '	i		 ^
•	The precision of the metliod could be expressed as a linear function
of the mean recovery, both as single-analyst and overall standard
deviations. The regression equations for precision are also shown
in Table 2.
IJEGIN!
i.AST I
Of-
.7

"T
The percent recovery of the method was similar to that obtained i
during|the developmental phase. The method has an extreme negative ;
bias for NDMA, nearly quantitative recovery for NDPrA, and a !
moderate negative bias for NDPhA.
i
Percent recoveries at the midrange concentration were from 36 to
43* foif NDMA, 84 to 102% for NDPrA, and 58 to 67% for NDPhA.
The precision of the method, was about as expected for NDMA and
NDPrA (and higher than expected for NDPhA. The additional
variability likely results from the column elution procedure.
Six water types were used 1n this study: laboratory pure, finished
drinking, surface, and three relatively interference-free
Industrial effluents. No difference 1n method performance was
attributable to the water t!ype from which the analysis was
performed.
Verifying the activity of the alumina and separating NDPhA from
d1phenyl amine proved to be a difficult step 1n the analytical
procedure and several laboratories were unable to achieve
satisfactory separation. }

EfA Form 2350-4 ($-80)
(PRCVIOUSLY c:s. CPA FORM 287)
PAGE NLH.iUER
BOTTOM Or
tMAGc. A3EA:'
OUTSIDE
DIMENSION
FOR TABLES
At.'D ILLUS-
TRATIONS
POW-1Wt.Nl "VtiNNNo OriiCf 1930 660-639
TYPING GUIDE SHEET,
. I .

-------

W)f.iw
illj,1; V-T*1*" ~'UI<\iv"rmj.i";*';'

TABLE 2. ACCURACY AND PRECISION EQUATIONS

Water type
N-NI trosod 1 rimthy 1 ami ne N-N1 trosod 1 -n-propy laml ne
N-N1trosod1pheny1 amine
Ranqe, uq/L
O.M-24.2 1.22-26.7
8.22-54.8
Laboratory Pure
fikUN
LAO": UiMCJlJ-
01- Tc XT
* C - true concentration
X - nean concentration
Accuracy
X
=
0.3 TC«
+ 0.06
X
» 0.96C
-
0.07
X
= 0.64C
~
0.52
* Precision





-


-
	



Overa11
S
a
0.25X
+
0.11
S
¦» 0.21X
•f
0.15
S
» 0.46X
-
0.47
Single analyst
SR
a
0.25X
-
0.04
SR
= 0.15X
+
0.13
SR
= 0.36X
-
1.55
i
Finished Drinking












|
Accuracy
X
a
0.37C
+
0.23
X
» 0.84C
-
0.02
X
» 0.60C
-
0.03
Precision |












1
Overa 11
S
•
0.2JX
+
0.J3
s
= 0.28X
+
0.05
S
- 0.37X
+
0.67
Single analyst 1
1
SR
¦
0.16X
~
0.15
SR
" 0.24X


SR
» 0.23X
4
0.81
Surface i













Accuracy j
X
"
0.42C
+
0.14
X
» 0.92C
~
0.05
X
- 0.S2C
-
0.56
Precision |













Overa11 ,
S
B
0.34X
+
0.17
s
« 0.26X
~
0.24
S
« 0.32X
V
1.03
Single analyst r
SR
a
0.29X
+
0.15
" SR
" 0.16X
+
0.24
"" SK
* 0.23X
4
0.24"
i
Industrial Effluent 1













Accuracy
X
a
0.38C
+
0.17
X
- 1.00C
~
0.21
X
- 0.6X
•
0.44
Precision j













Overa 11 <. i
S
B
0.33X
+
0.09
s
X
«o
•
o
¦
+
0.39
S
X
•
o
0
4
0.14
Single analyst ;
SR
n
0.13X
+
0.21
SR
- 0.18X
+
0.27
SR
» 0.34X
-
0.63
Industrial Effluent 2













Accuracy i
X
B
0.35C
~
0.13
X
» 0.86C
~
0.21
X
¦ 0.58C
¥
0.15
Precision













Overa11 j
S
B
0.33X
•f
0.09
s
¦ 0.33X
+
C.18
S
- 0.42X
4
0.66
Single analyst
1
SR

0.25X
+
0.03
SR
» 0.26X
-
0.04
SR
» 0.22X
4
0.65
Industrial Effluent 3













Accuracy |
X
¦
0.36C
+
0.30
X
° 0.94C
+
0.14
X
- 0.62C
4
0.54
Precision .













Overa11 1
s
W
0.27X
+
0.21
S
- 0.37T
+
0.25
S
¦ 0.37X
4
0.50
Single analyst
SR
a
0.28X
~
0.07
SR
- 0.22X
~
0.44
SR
« 0.21X
4
0.21
d
!_ il — _
El-A Torm 2350-4 (4-80>
(PREVIOUSLY CIN. LPA FORM ZSi)
PAGE N'JMBEH
BOTTOM <.'r"'
IMAGE AREA
OUTSiD'.-. 1
SJ-DIMcNblON J
1 FOR TABLES
.1 AW iL'_US- j
TFUTIONS 3
VrUS uOVr'
-------


WWHI4 ii||UNV«l£|[|i|

"1
SECTION 3
DESCRIPTION OF STUDY
sir o i fj
I A^T L t\-
Ui i[K!
The study design was based on Youden's original plan [4] for
collaborative evaluation of precision and accuracy for analytical methods.
According to Youden's design, samples are analyzed in pairs where each
sample of a pair has a slightly different concentration of the constituent.
The analyst is directed to do a single analysis and report one value for
each sample, as if for a normal, routine sample.
—	1	¦ '	'
~ In this study, samples were prepared as concentrates"in sealed glass
ampules and shipped to the analyst along with portions of final effluents
from manufacturing plants from three relevant Industries. Each
participating laboratory was responsible for supplying laboratory pure
water, finished water, and a surface water, thus giving a total of six
water matrices involved 1n the study, j The analyst was required to add an
aliquot of each concentrate to a volume of water from each of the six
waters and submit the spiked water to1 analysis. Three pairs of samples
were used. One pair contained the substances at what was considered to be
equivalent to a low level for the Industrial effluents; a second pair
contained the,substances at an intermediate level; and the third pair
contained the substances at a high level.
TEST DESIGN |
A summary of the test design using Youden's nonreplicate technique for
X and Y samples 1s given below:
i
1.	Threje Youden pairs were used for each parameter with the
f^ ^tion from the mean of each pair being at least 5% but not
more Ithan 20%.
2.	The three Youden pairs were spread over a usable and realistic
range with the lowest level 'estimated to be near the detection
Hmitj in the Industrial effluents with the highest background.
3.	Analyses were performed!1n six waters. Therefore, each
	 laboratory was to generate1 36 data points for each compounds.

i:
3,V



EPA Form 2350-4 (4-801
IPBCVIOUSUV CIN. EPA FCRM 267)
PAGC NUMBC-Ft
BOTTOM OF
I WAGS
Ou'TSlOE
Dlf.'.cNSiON
FOR TAUliS
Ai-.'D ILLUS-
TRATIONS
u S GOVERW >1T cpiNTiNCi 01' iCC 1880-660



TYPING GUIDE SHEET

-------
rnywy' 'i-v	¦¦-¦ww.'n1	. ,Wl,	inyiw ¦¦"	"-'»y1"
~
L AST ! ' .i !¦.'
Of P.,-.;
6.
7.
8.
9.
Three of the waters were selected from relevant industries as
determined from the information contained in a memorandum of
December 29, 1978 from M. Dean Neptune, Analytical Programs,
Effluent Guidelines Division to R. B. Schaffer, Director,
Effluent Guidelines Division, through W. A. Telliard, Chief,
Enerqy and Mining Branch.
Thirty-six concentrates in sealed glass ampules were shipped with
approximate!y 12 liters (3 gallons) of each of the three
industrial effluents to the 17 laboratories. The concentrations
of substances in the ampules were unknown to the participants.
Each participant was supplied with a copy of Method 607 and
supplementary instructions relative to spiking procedure, cleanup
column to be used, GC column and detector to be used, and GC
injection technique.	j
To commence an analysis, th-» analyst was instructed to open "an
ampule, add 1 ml of concentrate to 1 L of water, then analyze as
per instructions.	j
I	|
Each sample was to be analyzed once.
i	j
Before the formal study began, each participant was sent a pair
of ampules (not one of the pairs used in the study) for a trial
analysis by Method 607. After submitting da*.a from these
analyses to SwRI, participants met in Cincinnati to discuss and
resolve problems encountered during the trial run.
Fifty ampules of each concentrate prepared were supplied to the
project officer.	{

i

1
£CA l-'orm 2350-4 (4-80!
|f RUVIOUSLY CIN. CPA FORM 28/)
PREPARATION OF 'SAMPLES, WASTEWATER SELECTION AND PROCUREMENT
The N-nitrosamines used to prepare sample concentrates were obtained
from two commercial sources: Aldrich Chemical Company and Eastman Organic
Chemicals. These compounds were compared to high purity compounds obtained
from Ultra Scientific, Inc. (formerly :RFR Corporation) and were found to be
suitable for use as received.
I
The detailed protocol for concentrate and ampule production was
reviewed and approved by the project officer before ampule production
commenced. Salient points of the protocol are described in the following
paragraph. '

of
'MAGE AREA;
O'JTSlSfc
DIMENSION
rOH TABLES
AND ILLUS-
TRATIONS
us c.ov« fNvfcNr f'Piu'JNo orrirr

TYPING GUIDE SHEET

-------
^ »\ f n	vjm jy jmv-'ir*	rmi»«wc
w#wu'^y«y-"j|j*P'J|>'T»M.n ¦	g?f*i '•
¦ j 4-u#uyyi i > inra»p>ri^w;w»i*.n*<'q»'t"
Sample concentrates were prepared by dissolving precisely-weighed
(analytical balance) amounts in acetone in Class A volumetric flasks.
Where dilutions were required, Class A volumetric glass pipettes were used
to transfer the required volumes. All glassware had been fired overnight
at approximately 400°C before use. The volumes transferred were never less
than 4 mL. Solutions were put into brown, borosilicate glass ampules,
chilled, heat-sealed, and wrapped in aluminum foil. The sealed ampules
were stored in paperboard boxes .*t 4°C until shipped to participants. All
weighing, pipetting, and filling operations were performed under subdued
light in as short a time span as possible.
Before each concentrate was used to fill ampules, the concentrations
of substances were compared with a star.dard that had been prepared from
separate weighings of the substances. | Two ampules taken at random from the
ampules produced were checked against the same external standard mentioned
immediately above. These verification checks served to prevent gross
eVrors from being committed; the true values were "assumed to be those
established by the weighings of substances for concentrate preparation. It
was rare to find a verification analysis which deviated more than 5% from
the true value. .The -true values for all test substances are given in Table
3.	V	I
1	I
i	I
TABLE 3. TRUE CONCENTRATIONS IN STUDY SAMPLES
| (1 mL concentrate in 1 L water; ug/L)
Compound
Lowest
pair
Medium
pair
Highest pair
Ampule lot number
3
5
1
2
4
6
N-nitrosodimethylamine
N-nitrosodi-Ji-propyl amine
N-ni trosodi pheny1 ami ne
0.837
1.484
8.216
lJo08
1.217
10.956
6.696
10.388
16.432
8.064
8.519
21.912
20.088
26.71*
41.080
2^.192
17.038*
54.780
Industrial effluents selected
li'JOIN
last urjtej—
Oi Tc/.T
* Slightly 1n;excess of 20% deviation from the mean due to pipetting
error. This' exception approved by project officer.
for the Interlaboratory study were
obtained from the following Industrlalj categories:
Industrial effluent 1—rubber plant
Industrial effluent 2—textile plant
Industrial effluent 3—organics and plastics plant
Industrial, effluents 1 and 2 presented no problems for the analyst.
Industrial effluer.t 3 contained N-nitrosodimethylamine, or an artifact, at
3 '8''

1
L- 11
EPA form 23S0-4 W-80)
IrBCVIOUSLY CIN. ffH rORM 2»7|
^ j** -
PAGE NUMt'EH
507TOM Or
IMAGE ARIIA;
UTSiDS
ENSIGN
TOP) TABLES
i A\a illus-
trations
rtr Ui, GOVrfi .MfcNI f'HINTihC^r'lCE ISBO-M-Wi
TYPING GUIDE SHEtT
:

-------


a concentration in the low ppb range. Also in the third effluent was a
: peak eluting in the near vicinity cf the N-nitrosodi-.i-propylamine peak
I that could present some problem in peak measurements for this compound.
i Each wastewater was collected in two 55-gallon untreated iron drums,
'treated for residual chlorine as required, shipped to SwRI, pooled, and
, reshipped to participants in l-gaMon glass containers obtained from
i Burdick and Jackson, Inc.	;
i	j
| ANALYSIS AND REPORTING
j In addition to admonitions to follow Method 607 procedure (Appendix
| B), supplementary instructions were sent to participants at the time the
trial run ampules were shipped. Participants were advised how to dose the
water and how to make injections of extract into the GC using the solvent-
flush technique described by Burke.[5]. Only N-P detectors were to be
used. Only alumina cleanup columns were permitted; the substitution of
~Baker~aluminum oxide, basic (1-0539) or Fisher Scientific alumina, basic
(A-941) for the Woelm product (Superl, basic, 04571) specified in the
method was permissible. Also, the stipulation was made that all pertinent
GC recorder charts were to be sent with the data submitted for the trial
run and the formal tests to follow.
At the Cincinnati conference, trial run data were presented and a
step-by-step discussion of the method was given. The most common errors
(calculation, standards preparation, and Improper expression of results)
were pointed out. At the meeting, the following supplementary instructions
were given: '
1.	Column 1, as given in Method 607, and N-P detectors will be used
by all participants (earlier instructions as to the detector to
be used were unclear to some).
:	i
2.	Tap water and surface water ,samples should be tested for residual
chlor.ine. If present, 1t Should be removed by adding sodium
thlos'ulfate as described in ithe method.
I	• I
3.	The N-nitrosamine ampules should be protected from light and kept
at 4°C until used. Subcontractors were requested to use the
cmpules within a 30-day period, 1f possible, and to include the
date of use of each ampule in their final report.
After the Cincinnati conference, ja follow-up letter was sent to each
participant requesting that the example data calculation sheets and data
AFGIN
U\3t
OF TEXT
\
M _ JL _ __


\
EM'i Form J&WM \4-EO)
(1'HCVIOUSL.y CIn\ EPA FORM S07)
PAGlT NUMOER
BOITOMOF
IMAGE AREA;'
OUTSIDE
DIMENSION
1 f-OH tables
I AND ILLUS- '
THATIONS
X
\
V
OOVrHMMFNT t'MNTitto OPPiCE 1J8*6Gr>-&K*
*
TYPING GUIDE SHEET
¦i-k-n	. r..	¦ .tC	...p. /¦
hU&aZJuluktti

-------

i fjy	Tyj)arffw > m

"«JII»|*||»|^H|	/JUJ1 J

1
. -'¦ or-
•:*- 1 'A- '-
I
iit r;iN
LAST LINE Er-
>)f- TLXT
.sumniary sheets that were enclosed be used du' ing the study. The same
letter requested that the final report include information on the
following:	i
i
1.	method of quantitating; !
2.	sources of standards;	;
~	-3. source of surface water and the nature of any possible
contaminants;
4. suggestions as to how Method 607 could be improved.
DISTRIBUTION OF SAMPLES	|
r
A single shipment of 36 ampules was made by overnight air express to
all laboratories. No instance of ampule breakage during shipment was
reported. However, some participants reported shortages or received empty,
unbroken ampules and others accidentally broke ampules during the study.
Replacement ampules were provided in these cases.
. _	i	l
,jr The water: matrices were also shipped by overnight air express, each'
participant receiving approximately 12 liters (3 gallons) of each of the
three industrial effluents. The time required to collect and distribute
the effluents was about' three weeks.; No breakage in transit occurred.
Several gallons were broken or otherwise made unusable through laboratory
mishaps. Replacement shipments were made in these instances.
STABILITY OF SAMPLES
Since N-n1t!-osamine solutions, especially ones containing N-
nltrosodiphenylamine, had exhibited measurable deterioration after storage
of 30 to 45 days at 24°C, every effort was made to keep the period of time
between solution preparation and date of use to a minimum. Extra
precautions were taken by wrapping the ampules in aluminum foil immediately
after preparation, advising participants to keep the ampules 1n the
refrigerator until used, and asking tljiat they be used within 30 days after
delivery. Solutions and ampules were prepared and shipped within a period
of one week. All laboratories, except one, reported that their ampules had
been used within nine weeks from the time the solutions had been prepared
at Southwest Research Institute. Theltime period required for this last
laboratory to: use the ampules was 'about 11 weeks. Samples from excess
ampules held in storage at Southwest Research Institute in the dark at 4°C
were analyzed!at the end of four weeks and again at the end of nine weeks.
No evidence of deterioration was found. Method 607 was followed 1n the
analyses at the four- and nine-week points. At the r.ine-week point, an
HPLC method was also used. The latter procedure can separate N-
i
3/8"

EPA Form 2350-4 (4-SO)
(PREVIOUSLY CIN. EPA FORM Zf. >|
PAGE NIJM9E«
OCT TOM ClC
IMACGARCA;
-rJ OLT5IDC
T^DIMEN&ION
•' ron tables
	I AND ILLUS-
TRATIONS
^ruf. oovmM.iCN!	Ofic-fc itej-iGC-M?

TYPING GUIDE SHEET


-------

T^r> -jr
jnitrosodiphenyl amine and di phenyl amine (degradatiri product) on the
janalytical column, whereas Method 607 depends upon separation of these two
icompounds by liquid-solid chromatography before injection into the 6C-NPD
¦ instrument. The absence of serious deterioration of N-nitrosodiphenyl amine
j 1 s considered to be unequivocal. Therefore, deterioration of the N-
• ^riitrpsamines before use in tha Method 607 study should not be a factor,
[assuming that all subcontractors stored the ampules at 4°C as instructed.
r*s-
iCGiN
LAS'I LINE64—
OF TEXT
„j u-1.7'
P-1/8"

• A
$ 3/8"
J	
w
.V	'
EPA Fn.-m j!35&-4 (4-80)
(PRCVIUUSLV (.IN. CPA FORM 287)
j.-.;;:;:
PAGE NUMCEH
JUS CiOVEHNMtN" fll.NHNC. OFFiCfc :330-6w> 639
TYPING GUIDE SHEET
BOTTOM OF
IMAGE ARf A:
OUTSIDE
DIMENSION
FOR TABLES
I and illus-
tration:: ¦
- v... vct^T.	' -v-	y

-------
\ ^uuuwp
i-^.j	v,v^v' "V^1--


::l v;T" :i O
r *
'fv-
the precision, statistics and mean
water type on accuracy and precision.
ueg w
i -\sr l11\c
Oi TEXT
SECTION 4
TREATMENT OF DATA
The objective of this interlaboratory study was to obtain Information
about the accuracy and precision associated with measurements generated by
Method 607. This objective was not through the use of statistical analysis
techniques designed to extract and summarize the relevant information about
accuracy and precision from the data reported by the participating
1 aboratories. The statistical techniques employed in the data reduction
process ..are similar to the.techniques suggestedinthe ASTM Standard
Practice D2777r77. ~	I ' "		
i
The algorithms required to perform the statistical analyses have been
integrated into a system of computer programs referred to as IMVS
(Inter!aboratory Method Validation Study). The analyses performed by IMVS
[6] include several tests for the rejection of outliers (laboratories and
individual data points), summary statistics by concentration level for mean
recovery (accuracy), overall and Single-analyst standard deviation
(precision), determination of the linear relationship between mean recovery
and concentration level, determination of the linear relationship between
recovery, and a test for the effect of
A detailed description of each of the statistical analysis procedures
Is presented below.
PREPROCESSING '
I
An initial review of the data was performed to determine 1f a
systematic error was evident 1n the data that could be Identified and
legitimately corrected prior to| data analysis. Chromatograms and
supporting data were investigated t ..J

-------
•inconsistency in order not to prejudice the results. If an error was
[found, the corrected values were used in the data analyses. If the analyst
ireported no error could be found, the data were allowed to stand as
'reported.	;
i	:
1_ . All analyses reported as less than a detection limit and results that
;the analysts noted as influenced by spillage or loss of sample were removed
jfrom the data set prior tc insertion into the computer program. The data
set thus prepared was utilized in the statistical analysis supplied by the
sponsor.
Spurious data points are always a part of any set of data collected
during an interlaboratory test program. It is important to identify and
remove these data points because they can lead to values of summary
statistics which are not representative of the general behavior of the
method. However, some erratic behavior in the data may be directly related
to some facet of the method under the study. Therefore, spurious data
points should not be removed indiscriminantly, and any points that are
removed should be clearly identified since further investigation of the
analytical conditions related to the outliers might be of value. Data
rejected as outliers for this study as a result of any of the following
LAi>T U,¦>:!:' ft:
OF TC I
tests for outliers have been identif
tables.	'
i
YOUDEN'S LABORATORY RANKING PROCEDURE
ed by the symbol in the raw data
1
!_ L L
1
*
EPA form 23D0-4 M-80i
(PRLVIOUSLV CIN. iIPA FORM 2H7|

PAGE NUMtlEII
In some cases the analytical values reported by a specific laboratory
are so consistently high or low that a large systematic error may be
attributed to that laboratory. These data are not representative of the
method and should be rejected. Youden's [4] ranking test for outlying
laboratories was applied separately to!data from each of the waters used 1n
this study. Since six water types were used in this study, the laboratory
ranking procedure was applied to these1 six different subsets of the data.
Each laboratory,ranking test was performed at the 5% level of significance.
;	I
The Youden laboratory ranking procedure requires a complete set of
data from every:laboratory within a given water type. Missing data from
labortory i for water type j were replaced by the following procedure.
Letting denote the reported measurement from laboratory 1 for water
type j and concentration level C^, it 1s assumed that
I
~	'	*1 jk s Bj'Ck^J'Li'Cijk

BOTTOM OF
IMAGE AHcA
(.UTGIDE. -
DIMENSION
ron TABLES
I Wv'D ILLUS-
TRATIONS
ius covTM' 
L.
TYPING GUIDE SHEET
—

-------

III I ¦¦ I»I'.)II.WW«'<*''.1, "» 	 			
¦¦pijuiftwimi¦
where 8j and Yj are fixed parameters which determine the effect of water
type j, Lj is the systematic error due to laboratory i and is the
random within laboratory error. Taking natural logarithms, it follows that
' i
an
^ijk	+ "Yj *n + *n + *n

Br.Gi\
L-\i>1 Li::fc*a»
cr r~
which is a linear regression model with dependent variable in and
independent variable in C|<. (Details and justification for this model are
discussed in the section "Comparison of Accuracy and Precision Across Water
Types.") ,	|
The natural logarithms of the individual laboratory's data were
regressed against the natural logarithms of the true concentration levels
for the six ampules 1n each water type. The predicted values irTX-jj^ were
obtained from^the regression equation j and the missing values for X-f were
^estimated by X-jj^ = expUn	where exp (c) denotes the constant e
raised to the c power.
1
If the ranking test rejected a'laboratory for a specific water type,
then all of the laboratory data for that water type were rejected as
outliers. The rejected values were excluded from all the remaining
analyses. In addition, after completion of the laboratory ranking
procedure, the predicted values created to fill in for the missing data
were rejected and excluded from further analyses for all laboratories.
TESTS FOR INDIVIDUAL OUTLIERS
1
The data remaining after the laboratory ranking procedure were grouped
by water typew For each water type, the data were broken down into six
subsets defined by the six concentration levels (ampules) used in the
study. For eich subset of the data, all missing, zero, "less than" and
"nondetect" data were rejected. Nextl, the test for individual outliers
constructed by Thompson ["/] and suggested in the ASTM Standard Practice
D2777-77 was applied to the data using a 5% significance level. If an
Individual data point was rejected based on this test, it was removed from
the subset, and the test was repeated using the remaining data in the
subset. This.process was continued until no additional data could be
rejected. J
STATISTICAL SUMMARIES
Several summary statistics were calculated using the data remaining
for each concentration level after the outlier rejection tests were
Or Tfc/7
T
!__ 1.
3 e"
L
EPA Form : 350-4 (4-80)
(PHEVIOU! LY CI N.	F-ORV ?.97)
31:
PAGE NUMBER

UOTTOM or
If.iA'iE AREA;
C
DIMHr.'SION
FOR TASL^S
I AND ILLUS-
TRATIONS >
iui GOVLKKi.'iM purjtiiia or-ncc taw-fico-fey;
TYPING GUIDE SHEET



-------



*J*P
y.*t
•cm
VST Li:>
: TE/7
or
¦t',\ ;*,r
'-^performed. These summary statistics include: the number of retained data
• points, the mean recovery, accuracy as a percent relative error, the
| absolute overall standard deviation, the percent relative overall standard
| deviation, the absolute single-analyst standard deviation, and the percent
! relative single-analyst standard deviation. The basic formulas used to
^calculate these statistics are presented below where Xj, X2, ..., Xn denpte
the values of the n retained data points for a specific concentration
1evel.	'
Mean Recovery (X):
I	!
x = 17 * xi
i	»» j=i 1
The conventional notation for mean Recovery is X; however the symbol X is
used in this report to be consistent'with the output from the computer
TOr ¦:)-
t,A. iw.cr

• I
- 0
program.
I
Accuracy as a % Relative Error:
9-1/8"
I

one _ X - True Value „
*RE ' True Value x 100
Overall Standard Deviation:
i
1 5
1
and	1

Percent Relative Overall Standard Deviation:
X
%RSD = (f) x 100
The overall standard deviation
with measurenents generated by a group
S indicates the precision associated
of laboratories. This represents
the broad variation in the data collected in an interlaboratory study.
However, a measure of how well an individual analyst can expect to perform
in his own laboratory 1s another important measure of precision. This
single-analyst precision, denoted by S
by
_ 1_
A 3/8"
R, was estimated for each Youden pair
S-:M:

EPA Form 2350-4 (4-80)
(PREVIOUSLY CIN. EPA FORM 20?}
PAGE NUMbfc"n
BOTTOM OF
IMAGE AREA;
OUTSfOc
OlMENBtON
FOR TAULES
.1 AND ILLUS-
• TF1ATIONS
Cr US C-O'/CHNMt'NT PfilrtllUC. OCFiCE 1930-WWW

j i r.£
/•' &
TYPING GUIDE SHEE1
., *	s" ¦ ¦

.. >r .

-------
^T-.'.u.l-1-lH.-AJUi Ijuiyi-	l-yi WMUJISTWWWJWLK) ¦"? "	" 'IK J"'	"-""-V,'-t >.,.'J-a	H-WH'1.	l.4VJ.L1".«- '"J 7.- rvy-r?- v- ;¦>	*^5!
.... - 5,
SR ¦	*
D)<
where m - the number of complete sets of Youden pair observations
1	remaining after outliers have been removed,
i— D-j - the difference between the observations in the i^h
—_	Youden pair,		- -		
D - average of the D-j values.
by
The percent relative single-analyst standard deviation was calculated
XRSD-SR = x 100
where X* is the average of the two mean recovery statistics corresponding
to the two concentration levels defining the particular Youden pair.
These summary statistics provide detailed Information on the accuracy
J and precision of the data obtained for each concentration level. One
| objective of the statistical analysis of the data 1s to summarize the
Information about accuracy and precision which is contained in the
statistics.	i
A systematic relationship often exists between the mean recovery (X)
and the true concentration level (C)iof the analyte 1n the sample. In
addition, there are often systematic relationships between the precision
statistics (S and SR) and the mean recovery (X). Usually these systematic
relationships can be adequately approximated by a linear relationship
(i.e., by a straight line). Once these straight lines are established,
they can be useci to conveniently summarize the behavior of the method
i within a water type, and they can aid 1n comparing the behavior of the
method across water types. In addition they can be used to obtain
estimates of the accuracy and precision at any concentration level within
the applicable range studied. They can also be used to predict the
behavior of the, method when used under! similar conditions. These Important
relationships are discussed below.
STATEMENT OF METHOD ACCURACY
The accuracy of the method 1s characterized by the relationship of the
mean recovery (X) to the true concentration (C) of the analyte 1n the water
sample. In order to obtain a mathematical expression for this
relationship, a regression line of the form
C,! !
.J.i l
' Tl.M

!_ fcZ. _ I	
EPA Forrr 23S0-4 (4-CO)
(PRCVIOU'ilY CIN. EPA FORM 2«7j
I S.
COT TOM ut-
IMAGC- AIltA.
OUTSIDfc
FOK TAKLEo
Ar' J ILLUS-
TRATIONS
A U'i r.OviUNULNT XFllr.!.UC. Cl'r.Cf lBtft. 050-639
TYPING GUIDE SHEET
¦22
T-X

; n-f. 'ur rO-'


-------
P"

»v?*y u^»*'ji"'"viair-v|5UTT
<¦	'V
«• V"*TV
-?^r"v-.'J "T-^m
' Vs.l
H'.i'i
<• ,\
X = a + b • C
!was fitted to the data by regression techniques
(lM
The true concentration values often vary over a wide range. In such
cases, the mean recovery statistics associated with the larger
concentration values tend to dominate the fitted regression line producing
relatively larger errors in the estimates of mean recovery at the lower
concentration values. In order to eliminate this problem, a weighted least
squares technique was used to fit the mean recovery data to the true
concentration values. The weighted least square technique was performed by
dividing both sides of Equation (1) by C resulting in Equation (2)

I* b
(2)
	
!jJ U\
. !" A "
The X/C values were regressed against the 1/C values using ordlnatory
least squares to obtain estimates for the values of a and b. (This is
equivalent to performing a weighted! least squares with weights w = l/C^;
see Reference' 8;15 ;>age 108 for details.) Equation (2) can easily be
converted to the desired relationship given by Equation (1). The Intercept
(b) from Equation (2) becomes the slope (b) for Equation (1) and the slope
(a) from Equation (2) becomes the Intercept (a) for Equation (1). Equation
(1) can be used:to calculate the percent recovery over the applicable range
of concentrations used in the study.
The percent recovery is given by
Percent Recovery

£]x 100 = + b] x 100	(3)
If the absolute value of the ratio (a/C) 1s small relative to the slope (d)
for concentration 1n the low end of the range of concentration levels used
1n the study, then the percent recovery can be approximated by b x 100.
For example, suppose the true concentration values range from 25 pg/L to
515 tig/L, the fitted line 1s given !by X = 0.20 + 0.85 • C. The percent
recovery would be approximated by (0.85) x 100 * 85% over the specified
range of 25 ug/l to 515 pg/L.	j
If the ratio (a/C) 1s not small relative to the slope (b), then the
percent recovery depend; upon the true concentration (C), and it must be
evaluated at each concentration value within the specified range.
"L.P.
rL „l
*5- VV
&
m
a
Ji
LiorroM ot:
IMAGE AKUA:
-v - 0 jToIDE
^•D'^CNSION
' PCR TA2L!:3
i
f
4-.¦.
EPA Form 235'>4 (4-83i
|pncviousi.v an. er»A ror.M 2»7|

maibcr
-LA\0 1LLUG-
T3,
rSATUMS
m

':^4
fe*
>-T
, ' A
"V*.

%n
vva novi.wiv.rNr I'minhnO DMir;r itfSaaoo-W9
TYPING GUIDE SHEET-


-------
P"
V

K

Sc ;5
H flr
i STATEMENT OF METHOD PRECISION
l	1
! The precision of the method is characterized by tl.e relationship.;
! between precision statistics (S and SR) and mean recovery (X). In order to
J obtain a mathematical expression for these relationships, regression lines
• -of the form
S • d + e
(4)
and
SR - f + g • X*
(5)
were fitted to the data,
i
As discussed previously with respect to accuracy, the values of X and
ofteri vary over a wide rarige." In such cases the standard deviation
statistics associated with the larger mean recovery values will dominate
the regression lines. This will produce relatively larger errors in the
estimates of S and SR at the lower mean recovery values. Therefore, a
weighted least squares technique was also used to establish the values of
the parameters d, e, f and g in Equations (4) and (5). The weighted least
squares technique was performed by dividing both sides of Equation (4) by
X* resulting in Equation (6)
eeG!N
LAST IKEJS^
OF TEXT
S. d Ji- +
x 0 x*
(6)
and by dividing both sides of Equation (5) by X* resulting 1n Equation (7)
SR
X*
1_
x*
S 3/8"
1	
(7)
The {S/X} values were regressed against the {1/X> values and the
{5R/X*} values were regressed against the {1/X*} values using ordinary least
squares to obtain estimates for the parameters d, e, f and g.
Equations:(4) and (5) were obtained from Equations (6) and (7) 1n a
manner simil.;r to that discussed for mean recovery. The slope (d) for
Equation (6) is the intercept (d) for Equation (4), and the intercept (e)
for Equation (6) is the slope (e) for Equation (4). Similarly, the slope
EPA Form 2250-4 (4-«0>
(PREVIOUSLY C*,H. EPA FORM 287)
mm
PAGE NUMBER
>
BOTTOM OF
IMAHE ARfcAJ
ounios A
DIMUMSION?
FOR TAULESJ
I AMD ILLUS- !
THATIONS
ft Lis COVCRNMfeNT PniNT.Kf"5 OfFiC^I 1930-6C0-F3*
TYPING GUIDE SHEET
rJiiri.* nlV.im'TUii iftfcWw-*

-'A	1lW»> nMViV. »f?V* 1 rrt'i'Itit'i '**» ¦ ¦ ¦

-------
.(f) for Equation (7) is the intercept (f) for Equation (5), and the
intercept (g) for Equation (7) is the slope (g) for Equation (5).
Given Equations (4) and (5), the percent relative overall standard
deviation and the percent relative single-analyst standard deviation are
%RSD
¦GH
x 100
(8)
I and
%RSD-SA
"[I**']
x 100
(9)
• respectively. If the absolute value of the ratio (d/X) is small relative
; to the slope (e), then the percent relative overall standard deviation can
i be approximated by (e x 100) over the applicable range of mean recovery
f values. Similarly if the ratio (f/X+) is small relative to the slope (g),
j then the percent relative single-analyst standard deviation cun be
! approximated by (g x 100) over the applicable range of mean recovery
! values.	i
If the ratios (d/X and f/X*j are not spall relative to the slopes (e)
and (f), then the percent relative stands'd deviations depend upon the
values of the meai recovery statistics X and X*, and they should be
evaluated separately for each value of X and X*.
COMPARISON OF ACCURACY AND PRECISION ACROSS WATER TYPES
t
t
It 1s possible that the accuracy and precision of Method 607 depend
upon the type of water being analyzed. The summary statistics X, S and SR
are calculated separately for each j concentration level within each water
type. They can be compared across water types in order to obtain
information about the effects of water type on accuracy and precision.
However, the use of these summary statistics 1n this manner has several
disadvantages. First, 1t 1s cumbersome since there are 36 mean recovery
statistics (X).(slx concentrations x six waters), 36 percent statistics (S)
and 18 precision statistics (SR) calculated for each compound. Comparison
of these statistics across concentration levels and across water types
becomes unwieldly. Second, the statistical properties of this type of
comparisons procedure are difficult to determine. Finally, due to
variation associated with X, S and SR, comparisons based on these
statistics can lead to Inconsistent conclusions about the effect of water
type. For example, distilled water may produce a significantly lower value-

I —
3/3"
Ei'A Form 2350-4 M-80)-
(I'REVIOUSLV CIN. CPA FORM 267)
PAGE NUMbCK
£r US. GCV/EKNk'tNY ®ft'NTlNG OFFICE 19MKC0-639
TYPING GUIDE SHEET

-------
than surface water for the precision statistic S at a high concentration,,
but a significantly higher value for S at a low concentration.
An alternative approach, described in detail in Reference [9], has
been developed to test for the effects of water type. This alternative
approach is based on the concept of summarizing the average effect of water
type across concentration levels rather than studying the local effects at
each concentration level. If significant differences are established by
this alternative technique, then the summary statistics can be used for
further local analysis.	,
The test for the effect of water type is based on the following
statistical model. If X-jj^ denotes the measurement reported by laboratory
1 for water type j and ampule k, then
uEGIN
LAST
OF TEXT
m jk ,s

Ei
i jk
1 = 1,2,...,n
"j =" 1,2, •,6
k = 1,2	6
(10)
The model components Bj and Yj are fixed parameters which determine the
effect of water type j on the behavior the observed measurements (X-jjfc}.
The parameter is the true concentration level associated with ampule k.
The model component L-j 1s a random factor which accounts for the systematic
error associated with laboratory i. The model component is the random
factor which accounts for the within laboratory error.
The model 1s designed to approximate the global behavior of the data.
The multiplicative structure was chosen because of two Important
properties. First, It allows for a; possible curvilinear relationship
between the data {X^} and the true concentration level through the use
of the exponent Yj on C^. This makes|the model more flexible 1n comparison
to straight line models. Second, 'as will be seen below, there 1s an
Inherent increasing relationship between the variability in the data and
the concentration level C|< 1n this;model. This property 1s Important
because 1t 1s topical of interlaboratory data collected under conditions
where the true xoncentration levels vary widely.
Accuracy 1s related directly to the mean recovery or expected value of
the measurements tX-j}. The expected value for the data modeled by
Equation (10) 1s
£(xijk) = Pj * Ck'J • E(L-j • r-i jk)
77,a-
i	
—s.
EPA Form 2350-4 (4-80)
(PREVIOUSLY CiN. I'HA FORM 287J
_ PAGE NUMCEH
BOTTOM OF -
IMAGF AKf.»: 3
CUTSIDE
DIMENSION
FOFi TABLES.
I AND ILLUS-
TRATIONS
O U.S GOVERNMENT HUNTING OFFICE 1880-660-639
.TV TYPING GUIDE SHEET

'/ * ' ' 3 ' ' ' '

-------
r
1

Precision is related to the variability in the measurements (X^).
The variance of the data modeled by Equation (10) is
Var(Xijk) = [bj CkYJ]2 Vardi • eijk)
(12)
which is an increasing function of Ck.
The effect of water type on the accuracy and precision of Method 607
is determined by the values of the parameters (8j } and {-yj) in Equations
(11) and (12). If the {0j} and {yj} vary with j (i.e., vary across water
type), then the accuracy and precision of the method also vary across water
type.	j
In order to determine if these parameters do vary across water type
and to compare their values, they must be estimated from the laboratory
data using regression techniques. Equation (10) represents the basic
model. However, taking natural logarithms of both sides of Equation (10),
the following straight line regression model is obtained,
'/An Xijk
An f$j + Yj
An Ck + in L-j + An e-jjk
(13)
which can be analyzed using standard
inear model analysis techniques. The
parameter An pj is the intercept and Yj the slope of the regression line
associated with water type j. It is assumed that An 1s normally
distributed with mean 0
and that the An Li
and variance of and that *n e1jk normally
distributed with mean 0 and variance o<
^ and that the An L
-------

versus

Ha: An Pj - *n Pi i 0 and/or Yj - Yi ^ 0 for some j = 2,3,4,5,6
iiCGiN
LAST LINE
C." TEXT
The test of null hypothesis Hg against the alternative hypothesis
is based on an F-statistic derived from standard linear model theory. The
probability of obtaining a value of an F-statistic as large as the value
which was actually observed (F OBS), denoted by P(F > F 08S), 1s calculated
under the assumption that Hg is true. The null hypothesis Hg 1s rejected
1n favor of if P(F > F OBS) 1s less than 0.05.	1
If Hg 1s rejected, then some linear combination of the differences
An 0j - *n Pi and Yj - Yi is statistically different from zero. However,
this does not guarantee there will be a statistically signiflcant^direct
effect attributable to any specific water type since the overall F test can
be overly sensitive to minor systematic effects ccirmon to several \water
types. The effect du? to water type is judged to be statistically
significant only if one of the differences in Pj - An Pi and/or Yj - ti is
statistically different from zero. This is determined by checkingYthe
simultaneous 95% confidence intervals which are constructed for each, of
these differences. • Each tru® difference can be stated to He within ^ts
respective confidence Interval with 95% confidence. If zero 1s contained
within the confidence interval, then there is no evidence that th£
corresponding difference is significantly different from zero.	\
If at least one of the confidence intervals for the differences)
in pj - in Pi or yj - yj fails to include zero, then the statistical
significance of the effect due to water type has been established^
However, establishment of a statistically significant effect due to water
type does not necessarily mean that the effect 1s of practical importance.
Practical importance is related to, the size and interpretation of the
difference, j	j
The interpretation of the differences involves comparing the mear
recovery and standard deviation of the	data for each water type to
the mean recovery and standard deviation obtained for distilled water.
These comparisons are made on a relative basis. The mean recovery for
water type j 'is given by Equation (11). The mean recovery for water type j
fs compared to that for distilled water (j=l) on a relative basis by
EOjiik) _ B|1 CtXi eg, • Jjn) _ a
EtX-jijJ a. r. Yi rh , .	fti
Pl Cku E(Li • ejlk) Pi
(14)


3/8"

if A Form 2350-4 «-80)
(J'REVHMJSLY CIN. CPA l-ORM 207)
22.
PAGE NUMeErt
\t
¦¦N
KOTTOM OF ;
llv.AGK AREA,
outside j
'DIMENSION
' FOR TASLES
.1 AND ILLUS-
TRATIONS
>
O US GOVfcRNMENT POINTING OFFICE l98£>-«0-63P
TYPING GUIDE SHEET
s

-------
PVT

ry-^rrrr.v

. ' i\!

\ \
\ \
\ '¦
\
\
*
\
\ .
i-
.[The ratio of the\standard deviations would be equivalent to Equation (14)
j and therefore the^ interpretation of the effect on precision 5s the same as,
; that for the effect) on mein recovery."]
\	H
i	\	\	¦ \
j The ratio in\EC|Juation U4) is a measure of the relative difference in
mean recovery between water\type j. anci\distil led water. It is composed of
jtwo parts (a) gj/Bi, which is^independert of the true concentration level
ji'I.e., the constant bias) ahd (b)	^ which depends upon the true
iconcentration level (i).e., the concentration dependent bias). If Yj - Yj
{is zero, ^then the relaVive difference in znean recovery is just Cj/Pi which
lis indep^ndent of concentration T^vel C^. ^t can then be stated that the
linear recovery of water type j - i'« (Bj/Bj) Vc 100% of the mean recovery for
j distilled wa*er, If Yj - Yi is nottzert, thin the meat) recovery of water
itype\j is Ci.Bj/Bi) 'C^ •'j-'M) y \oOX of t\jat for distilled water and
J there'Fore depervis upon the\true concentration '^evel C^.
nniiv1
cIllustrate\Vhese points cor
signif1cant\\F-val'je f\?.s b
onsiajir the following example.
been Obtained *nd the confidence
!- : — In order to
!Suppose that a signif1can^\^-*aii^-..
! interval's for all \the differences contain *ero 'txcept for water type 5.
Tor water type 5,' the point estimate ixtr in 30\- xn &i 1s -0.38 and the
: rnnfiripnop inf.prurfl for tn ft5\- JM \fi (-0.&J, \-0.0/) . The point
tj 1s 0.0A >nd the confidence1).!nterval for Y5 - Yi 1s
\estimate for y;; .	v
iV-0.04, 0.1ft). Jn this case a\stat1stfealty significant effect due to
jwatef typeihas been established which Involves aj'.ly water type 5. The
( practical significance of this effect 1s judged by considering Equation
i(140. \The ra^io o^ mean recoveries for water type 5 and distilled water 1s
| given by;	* 1	x
\ \
\
V
V.
E(x15k)
^qiiJUi
c S?
(15)

and the r^atlo of the standard deviations is given by
BEGIN
LAO t LINEBf*-
Or TEXT
\
VarUijjJ
YaFOqtiJ
.{1 cJ5-^l
1*1
(16)
3,8"
Since the confidence internal for Y5 -Jyi contains zero this difference 1s
assumed to be Insignificant and 1s set to zero. Therefore, Equations (15)
and (16) reduce to 85/^1 • The point estimate for &n $5 - in pi was -0.38.
Therefore, the point estimate for Bg/pi 1s 0.68, and the mean recovery for
water type 5 1s:est1mated to be 68% of the mean recovery for distilled
water. Similarly the standard deviation for t!ie data for water type 5 1s
EPA Form 2350-4 (4-80)
(PREVIOUSLY CIN. CPA FORI! 267)

PAGE NUMBER
	^
,1
L' 5 GOVERNMENT POINTING OM'tCE 19ffl>8GO-«3
TYPING GUIDE SHEET
BOTTOM OF
IMAGr" ARC A;
OUTSIDE
u'VE^r.iow
FCn 7ADLES
A MO ILLUS-
TRATIONS



-------

'J'	*-*wr*wr"
Miy v-v^»«v^
estimated to be 68% of the standard deviation for distilled water. Since
the 95% confidence interval for tn 35 - an gj was (-0.69, -0.07), any value
1n the interval (0.50, 0.93) is a reasonable estimate for &5/B1, and the
mean recovery (standard deviation) for water type 5 can be claimed to be
from 50% to 93% of the mean recovery (standard deviation) for distilled
water, The practical significance of the effect due to water type 5 would
depend upon the importance of a mean recovery (standard deviation) which 1s
between 50% and 93% of the mean recovery (standard deviation) observed for
distilled water.
The comparison of accuracy and precision across water types just
discussed is based on the assumption that Equation (10) approximately
models the data. It is clear that In practical monitoring programs of this
type such models cannot model the data completely 1n every case. This
analysis, therefore, is viewed as a screening procedure which identifies
those cases where differences 1n water types are likely to be present. A
'more detailed, local analysis can then be pursued using the basic summary
statistics for precision and accuracy.
1 1
1
EFGIN
LAST
or- tr.XT u
I
I.
Tliz- _
tPA form 2350-4 (4-80)
(PRTVIOUSLY CIN. EPA FORM ZH7
*
I'AUE NUMF3EU
BOTTOM OF
IMAC.F AflGA*
OJT3IDE
^DIW.FfJSION
FOR TABLES
I AMD ILLUS-
TRATIONS
ftUS CiQVtRNMCNT PRINTING OT-ICE \M0-6to-529
TYPING GUIDE SHEET

i'-li'l "'i*'tt 1 m 1' rili , i~11	¦¦ 1 ¦- ^	-.it'	»		 ¦-

" 'in

-------
SECTION 5
DISCUSSION AND CONCLUSIONS
DTfilN
i.i\
TCXT
!£S>~
The rejection rate for these data were similar to expectations for an
interlaboratory method study. A total of 356 values were rejected,
Including missing data, data reported as less than some value, and
statistical outliers rejected according to the criteria stated earlier.
These represent 19.4% of all data requested.
The summary statistics obtained from the collaborative study data are
presented in Tables 4 through 6 for the three nitrosamines studied.
Discussion of the accuracy, precision and consistency across water types is
presented separately in the following sections.
ACCURACY OF THE HETHOO	!
t
The accuracy of the method is estimated by the linear regression of
mean recovery versus true concentration presented earlier. The equations
are valid only over the range of true values studied and should not be
extrapolated beyond this range. J
As an Illustration of the accuracy that can be expected for the three
nitrosamines under study, the mldrange concentrations were Inserted into
the regression equations to estimate percent recoveries. These predicted
recoveries at the mldrange, shown 1n Table 7, provide a basis for comparing
the recoveries across water types and among nitrosamines. Recoveries at
other concentrations would vary due to the relative Impact of the slope and
Intercept of the regression line upon the calculated result.
1	' I
There 1s good consistency across water types for the nitrosamines,
with recoveries ranging from 36 to 43% for NOMA, 84 to 102% for NDPrA, and
58 to 67% for NDPhA.
I
PRECISION OF THE METHOD
i
The precision of the method 1s estimated by calculating regression
equations for the precision components versus the mean recovery. In each
case, the regression model is assumed and both the overall standard

EPA Foriv 2300-4 (4-80)
(PREVIOUSLY C»»/. CFA FORM 207)
PAGE: fJuMBTK
£ US GOv^R.JMUJTPflt»*T'NQ OFFICE 1980-664-639
, TYPING GUIDE SHEET
BOTTOM 01'
IMAGE AHLA;
OUTSl^C
DIMENSION
FOK TADl.tS
A^D ILLUS-
TRATIONS

-------
W~'~
X C
-< -7
pbrr:
I tJ
I
I
h*-
S1"''1
Pp.
W-;'¦
TABLE 4.
EHVtRTMEKTAL «3iIT0R1HG AND SU»?0»T IARQBATQR?
OFFICE OF RESEARCH AW OEVELORfENT
EWIR3H»E*TAL PROTECTION AGENCY
•**£»« *ETM30 VALIDATION STUDY - SRI NITROSAKINES*
STATISTICAL SUMARY FOR N-NITROSIOINETHYLAMNE INALYSES SY WATER TY»E
WATER 1	WATER 2	WATER 3	WATER *	WATER
- t »
I * 1 ?i
PAGE 1*
WATER 6
L-^J
law TOUOEN PAIR
3 9
1 9
3
9
3
9
3 9
3
*
NUK4ER OP DATA POINTS
14 1«
It 16
12
13
14
19
14 14
13
13
TRUE VALUf UG/l
0.84 I.01
0.8* 1.01
0.94
1.01
0.84
1.01
0.94 1.01
9.94
1*01
he an recovery
.1* .*T
.44 .76
.99
.46
.40
.67
.44 .49
.92
.79
ACCURACY AS t REl ERROR
-99.93 -92.92
-47.09 -26.10
-33.74
-94.23
-92.76
-33.0)
-47.06 -99.33
-31.96
-22*60
OVERALL STO DC* (SI
.13 .26
.74 .68
.39
.32
.17
.49
.24 .24
.32
*32
OVERALL REL STO OEVt *
91.90 59.61
94.20 91.49
69 .M
68.66
41.86
72.09
94.33 94.29
61.07
66*64
SINGLE STO OEVt (SRI
.07
.24

.29

29
.14

• 25
ANALYST REl OEVt t
i6.tr
40.30
97.7?
93.
00
31.62
39
• 01
IIEDIU* TOUOEN PAIR

1 2
1
2
1
2
1 2
1
2
NUMBER OF DATA POINTS
i« ti
14 13
12
13
19
13
13 13
13
15
TRUE VALUE UG/l
6.TO S.06
6.70 8.06
6.70
8.06
6.70
8.06
6.70 8.06
6.70
e*06
MEAN RECOVERY
2.74 1.09
2.60 3.11
2.88
3.93
2.73
2.99
2.42 2.80
2.29
3*39
ACCURACY AS * REl ERROR
*99.03 -61.73
-61.14 -61.40
-97.01
-91.23
-99.26
-63.46
-43.89 -69.27
-64.39
-97*93
OVERALL STO OEV (SI
.93 .TJ
.68 .70
• 62
2.08
1.01
1.01
.84 1.00
.36
1.00
OVERALL PEL STO OEVt t
19.A3 23.78
26.19 22.99
21.69
92.89
36.99
94.34
34.84 39.86
16.13
31*77
SINGLE STO OEVt (SRI
• A3
.69

1.23
•
94
.41

• 76
ANALYST REL OEVt t
14.87
21.09
36.19
18.
99
19.79
27
*01
HIGH YOUDEN PAIR
A 6
4 6
4
6
4
6
4 6
\
6
NUMBER OF DATA POINTS
13 19
19 19
14
14
14
14
14 13
14
16
TRUE VALUE UG/L
10.99 IA.19
20.09 24.19
29.09
24.19
20.09
24.19
20.09 24.19
23*09
24*19
NEAN RECOVERY
6.80 '.*3
*.80 9.14
8.49
10.14
7.69
9.42
7.97 8.07
7.71
8.66
ACCURACY AS * REL ERROR
-64.17 -64.12
-60.20 -62.21
-97.94
-98.09
-61.92
-61.06
-60.34 -66.64
•61*54
-43*40
OVERALL STO DEV (SI
1.69 3.70
1.49 2.49
3.09
3.79
1.99
3.07
2.39 3.33
2.31
3.9J
OVERALL REL STO UEVt »
21.91 42.89
19.44 27.24
36.60
37.38
29.49
32.94
29.90 41.24
?S»11
*4.41
SINGLE STO OEV# (SRI
Z.ST
1.4t

2.94
1.
90
2.72
2
>60
ANALYST REl OEVt S
13.26
17.29
27.28
17.
92
33.86
31
.3*
WATER LEGEND
OISTILLEO WATER
TAP WATER
SURFACE WATER
WASTE WATER 1
WASTE WATER 2
WASTE WATER 3

-!>
§1
H _
CP
ZC
*/¦>
¦no?; -
O - C c-
13 -» > -I
C — n, > s.




-------
O r w
—. > T*

ft- .. ~ -
8
S
i.
r-fj
i
hri
2	8
>	^
*
0	.
X
2

f .-
? ¦ -
^..
t'.-.
H
¦<
- tj
"Z
: o
¦ a
c
5 •
. m
co
X
m
m
H

i
TABLE 5. 	~ 4
EMViRnwviTAi instioRiNG and su»»ort laboratory
OFFICE OF RESEARCH ANO DEVELOPMENT
E-WIR1NNENTAL "RlTECtrai A6ENCY
•~•EPA METHOD VALIDATION STUDY - SRI NITROSANINES*
STATISTICAL SUHARY FOR N-NITROSODt-i-PRlPYlANI^E ANALYSES BY HATER TYPE
MATER I	MATE* 2	HATES 1	WATER 4	WATER
I * * SI »»SE 57
WATER

LOW TOUOEN PAIR
3
*
1
1
1
9
1
9
1
5
1
&
NUMBER OF DATA POINTS
11
11
12
11
11
19
1*
11
14
15
11
12
TRUE VALUE UG/L
1.48
i.>;
1.48
1.2*
1.49
1.22
1.48
1.22
1.48
1.2?
1.48
1.22
MEAN RECOVERY
1.2]
1.18
1.12
1.09
1.19
1.19
2.09
1.19
1.44
1.27
1.76
1.11
ACCURACY AS T REL ERROR
-17.15
-2.40
-24.18
-11.*9
-6.93
-1.21
17.34
-2.11
-2.84
4.05
11.29
-7.15
OVERALL STD OEV IS)
.*>
.11
.41
.28
.11
.61
1.28
.59
.66
.62
1.17
.56
OVERALL REL STO DEV. (
16.27
29.OA
16.79
26.09
16.99
92.11
62.9)
99.92
41,9 5
48.74
66.>0
49.37
WATER LEGEND
OISTILLEO WATER
TAR HATER
SURFACE WATER
WASTE WATER 1
WASTE WATER 2
WASTE WATER 1
SINGLE STD 0EV» «SR>

.10

.26

.43
.16
.12

.75
ANALYST REL OEV. *
14.72
21.10
11.85
34.68
21.96

51 .68
N'OIU* YOUDEN PAIR
1
2
1
2
1
2
1 2
1 2

I 2
NUMBER OF DATA POINTS
11
11
14
14
16
16
11 14
15 16

15 15
TRUE VALUE US/L
10.19
9.52
10.19
8.52
10.19
9.52
10.IV 8.5*
10.19 8.52
10
.39 8.62
RE AN RECOVERY
9.93
8.39
9.10
6.57
9.16
8.18
11.27 8.41
9.05 8.11
9
.59 9.24 J
ACCURACY AS t REL ERROR
— 5.42
-5.99
-10.46
-22.94
-11.81
-1.94
8.47 -1.05
-12.91 -4.52
-1
.69 8.42
OVERALL STD DEV (SI
2.09
>.48
2.54
2.18
2.37
2.79
2*-58 2.49
2.54 2.75
4
.20 3.65 l
OVERALL REL STO OEV. «
21.31
10.71
27.29
11.16
29.91
14.09
22.87 29."9
28.11 11.75
41
.76 39.56 )
SINGLF STO OEVi (SRI

l.ai

2.66

1.89
1.89
1.80

3.11 ;
ANALYST Rtl DtV, 1
20.25
11.51
21.ac
19.04
20.90

13.04 ;
1
HIGH YOUDEN PAIR
9
6
4
6
4
6
4 6
4 6

4 6
NUMBER OF DATA POINTS
11
14
14
11
11
16
14 11
16 16

14 14 i
TRUE VALUE US/L
26.71
17.04
26.71
17.04
26.71
IT. 04
26.71 17.04
24.71 17.04
26
71 17.04 |
MEAN RECOVERY
29.63
16.91
21.09
11.61
24.17
16.01
25.55 16.89
22.01 14.57
24
13 14.99
ACCURACY AS t REL ERROR
-1.8T
-2.19
-10.5T
-20.00
-9.51
-6.06
-4.11 -.81
-17.52 -14.47
-9
65 -12.03 1
OVERALL STD OEV (S)
2.12
4.41
4.49
4.52
4.6J
5.59
7.59 4.40
9.40 4.98
r
91 5.16 i
OVERALL REL STO DEV. (
9.02
26.51
19.90
11.18
19.32
14.95
29.72 26.01
42.69 14.'}
32
77 34.41 I
SINGLE STO 0EV» ISRI

t.«l

t.91

2.77
4.16
5.66

3.44 [
ANALYST REL 0€V, «
11.40
15.90
11.77
20.54
10.91

17.60 t
o —
L_JL

-t > C „
-G ?! O — "*¦ O
V rn 2 m ^
-A


-------
- 3- n
. ^<0
~
X •
-* -
H.
<
O


V-h
J I.
o
c
o
m
(A
X
m
m
H
a
<
o
c
. «•
r*
•<
r>
2
n
>
Tl
0
1
u«sr --
'nz
_I I I	_ —	i	—_	
			TABLE 6.
EltftRniREITAi N*mTQRING AND SUPPORT LABORATORY
OFFICE OF ftfSFARC*4 AND DFVfin?»fNT
ENVIR3NNEKTAL PROTECTION AGENCY
~~•EPA 1ET4QD VALIOATION STUDY - SRI *IT*OSANINES»
STATISTICAL SUN«ARY FOR N-NITROSOOtPHENYLININE ANALYSES BY WATER TYPE
~'V!
I " V St
WATER t
WATER 2
WATER 3
WATEft *
WAT«R 9
WATER 6
>
O
f\3
"GOT
| LOW Y04I0EN PAIR
I KUPBER OF OATA POINTS
I TRUE VALUE U6/L
: MEAN RECOVERY
I ACCURACY AS t ftEL ERROR
j OVERALL STO OFV fSI
OVERALL RFL STO 0EW» t
3
9
1
9
3
9
3
9
1
9
«
•
t «
1
1
5
14
11
11
13
11
13
12
12
13
19
u
1*
9*22
19*96
9*22
10*96
9*»2
19.96
8*22
10*96
8*22
10.96
t.n
10.94
9*86
7.04
4*9J
6*04
6*29
4*38
9*20
7*07
4*99
6.62
5.56
6.97
•29*88
-39.79
-60*02
-61.86
•68*29
-41.00
-36*76
-19.91
-39*23
-39.93
-42.10
-16.41
2.16
2*66
2*76
1*89
2*19
.3*06
2*21
3*06
3*12
3*49
>.6t
J.07
16*89
37*46
99.96
6**16
96*21
67*99
62*R9
43*31
62*49
9 2*74
St.>7
44.14
SINGLE
ANALYST
STO OEVt (SRI
REL OEV* t
• 66
10*26
<•11
39*02
1*33
29*09
1*23
20*01
2.17
37*61
1*61
26.0*
HIGH YQUOEN PAIR
NUHRER OF OATA POINTS
TRUE VALUE UG/L
HE AN RECOVERY
ACCURACY AS t REL ERROR
OVERALL STO OEV tS)
OVERALL REL STO OEV* *
6
16
61*09
26.66
-19*19
11*19
61*99
6
19
96*711
96*37
-37*26
16*92
61*42
6
11
*1*09
29*67
-37*91
19*30
63*11
6
16
96*79
39*3*
-66*62
11*43
67*97
6
19
61*09
21*4*
-67*91
6*16
37*99
6
11
96*78
34* 29
-17*61
12*66
36*29
6
12
61*09
29*27
-38.49
8*46
33*49
6
13
56*79
39*2*
-28*12
17*10
63*96
6
19
61*08
29*17
-38*72
14.07
99*89
6
19
94*79
31.R9
-41*79
16.38
91.38
4
11
61*98
29.»3
-31*99
10*83
62*93
6
19
94 « 78
J2*?9
-41*12
12.14
37*69
SINGLE
ANALYST
STO OEV» ISRI
REL OEV# X
8*62
27.60
7*93
26*30
9.0*
18*10
9*78
30*30
6*68
30.42
7*18
26*94
WATER LEGEND
1	- OISTULEO WATER
2	- TAP WATER
3	- SURFACE WATER
4	- WASTE WATER 1
9 - 4ASTE WATER 2
6 - WASTE WATER 3
NEOIIM VOUOEN PAIR
1 2
1
2
I
2
1
2
1 2
1
2

N'JM<1ER OF 0&TA POINTS
11 11
1*
15
13
12
11
12
13 14
14
14
*
TRUE VALUE UC/l
16.41 21.91
16.41
21.91
16.41
21.91
16.41
21.91
14.43 21.91
14.4 3
21.91

HEAN RECOVERY
11.3* 14.44
11.01
12.81
9.9}
14.00
10.41
11.11
9.14 12.6)
11.34
15.47

ACCURACY IS X REL ERROR
-30.9* -32.24
-«.91
-41.44
—19.43
-16.11
-34.4?
-30.13
-43.1) -42.27
-31.01
-29.39

0*ER411 STO OEV (SI
5.52 (.17
!.)2
1.06
4.))
4. TO
1.09
7.31
4.09 3.19
3.97
6.60
' c
OVERALL REL STO OEV. *
48.66 42.91
14.43
17.41
41.76
13.60
29.61
49.07
43.7) 2).17
15.04
42.64

SINGLE STO OEV* (SRI
1.80
1.
02
4.
11
1.
81
1.96

2.47

ANALYST REL OEV# I
29.02
2).
30
14.
15
29.
92
17.83
1
8.40


2*
¦If
§5
tl
o ? w
5>3
i/i —i
s>i
SO

J, '--'1

-------
V
s
3
E3
6
h!
h5
>
c

Iv
t--'-
3
2
z ¦
o
C5
c
o
m
rV:S2
m
m
o
<
r;
TABLE 7. ESTIMATED PERCENT RECOVERY FOR VARIOUS WATER 'TYPES AT MIDRANGE CONCENTRATION
f i
Compound
	~ Percent recovery 1n given water types	1
Laboratory Finished	Industrial Industrial Industrial.
Mldrange pure drinking Surface effluent 1 effluent 2 effluent 31
N-n1trosod1methylam1ne	12.52	37
N-n1trosod1propylam1ne	 13.96	95
N-nitrosod1pheriylMi1he	31.51" " 66"
39
84
60
43
92
"60
39
102
67
36
88
58
.yl
h>^oo;o
_ . m (st H
5r>"csi
2 L C n; > ^
?r> Z '£ O
38
95
64
'i



-------
MB^DCEk
7
Lf\C*
I
deviation (S) and single-analyst standard deviation (SR) are expressed in
this manner. As was the case with the accuracy equation, these equations
may only be said to be valid over the range of concentrations studied.
As was done with the accuracy equation, the precision of the method is
Illustrated by the calculation of percent relative standard deviations at
the mldrange concentration. Although the variability at any concentration
over the range studied will depend upon the relative impact of the slope
and intercept of the precision equations, these relative standard
deviations, shown in Tables 8 and 9, allow a general evaluation of method
performance.
COMPARISON ACROSS HATER TYPES
The summaries on the effects of water type cr» the results obtained are
presented in Tables 10, 11, and 12. In none of the instances was the
statistic significantly large to warrant rejection of the null hypothesis
of equality. The conclusion Is that the method performed 1n a comparaM ?¦
manner over the six waters used in this study. It should be remembeie-,
however, that the Industrial effluents selected for this study were choser.
because of their relatively low background of interfering substances. Some
differences can be expected to occur when analyzing a variety of effluents
and the user should evaluate method performance with each matrix on which
1t 1s used.	j
i
METHOD EVALUATION	!
J
The accuracy of the method as determined from the collaborative study
data compared favorably to the results obtained by SwRI 1n the
developmental phase of the method [10]. In developmental studies 1n
Interference-free water, recoveries of 40, 98 and 80% were obtained for
NDMA, NDPrA, and NDPhA, respectively, and In samples of five effluents, the
average recoveries ranged from 25 to 38 for NDMA, 54 to 103 for NDPrA, and
46 to 89 for NDPhA. By comparison,!the percent accuracy at mldranges
spanned 36 to 43 for NDMA, 84 to 102 for NDPrA, and 58 to 67 for NDPhA 1n
the waters used in this study.
Method detection limit (MDL) was determined for each of the three
compounds 1n Interference-free water and 1n two Industrial effluents, using
a procedure specified by the Environmental Monitoring and Support
Laboratory. In this procedure, at least seven sample replicates containing
the compounds at concentrations near the estimated detection limit of each j
were analyzed by Method 607. The standard deviation of the replicate !
measurements was calculated and multiplied by the Student's t value J
	;	J		—
X'S"
L


EPA Form 2350-4 («-80)
(PflCVIUUSH Cll.. CP« l-O.IM 267}
PAGE rJUMSF-.rt
f-'JTTOM OF
IMAGE ARC-A:
OUTS'DC
UlUGNSiO.M
FOR TADLCS
a--;d illus
TRATIOND
Ol5. GOVEIl'iMFWT PfllMTING OFHCE 1
hb.
TYPING GUIDE SHEET


-------
K.
> . ¦
r "'V'mV-
t %¦
-' \'J

y'i \-. j
fc'V\. :
r.v
.\'o -
¦U',
& •
V.
T 2 h-t»
TABLE 8.
ESTIMATED OVERALL PERCENT RELATIVE STANDARD DEVIATION FOR
VARIOUS WATER TYPES AT MIDRANGE CONCENTRATION	
i i
l7'.
ur--
I"-:;

£ foi>
2 t"

•H,
-<
. "0
z
o,
o'
c
D
m
gv
m
m
' tr
§
3
5
¦o
O¦
U -\'.r 1
:vv>^ ;
£
§ <

L -4
Compound
Mldranqe
Laboratory
pure
Overall RSD for given water type	~
Industrial Industrial
Finished
drinking
Surface effluent 1 effluent 2
Industrial
eff1uent 3
N-n1trosodimethylam1ne	12.52 - 26
N-n1trosodipropy!am1ne 13.96 22
N-nitrosod1phenylamine	31.51	45_
25
28
39
35
28
35
34
29
39
34
34
44
29
39
39
TABLE 9. ESTIMATED SINGLE-ANALYST PERCENT RELATIVE STANDARD DEVIATION FOR
VARIOUS WATER TYPES AT MIDRANGE CONCENTRATION



Single-analyst RSD for given water type
Compound
Mldranqe
Laboratory Finished Industrial Industrial Industrial
sure drinking Surface effluent 1 effluent 2 effluent 3
N-n1trosodimethylamine
N-n1trosod1propyl amine
N-nitrosod1phenyl amine
12.52
13.96
31.51
25
16
31
17 30 15 25 29
24 18 20 26 25
26 24 31 24 22
i k
2 4- O V C C" o
5C>£5™||
.

-------
i miimil

rr~.
TABLE 10.
ENVIRONMENTAL HON I TORINO AND SUPPORT LABORATORY
3F'IC; OF PESEARCH «NO DEVELOPMENT
ENVIRONMENTAL PROTECT ION AUENCY
~••|P» METHOD VALIOATION STUOT - SRI NITSOSAKINES*
EFVFCT V MATE* TYPE ON N-NITROSODINETHYLAMNE ANaLTSIS
< V "
i ' I
-Tin
I
i » V 51 page ja

»» POINT ESTIMATES
DHTIUID KATE* SLOPE IGAHMAII) ¦ .9963*
VAT®* INT	.19901	.016*	I
.1669	(	-.1907 •	.*9091	-.0933	I
.0*12	I	-.2993	.	.36771	-.0*28	I
.3387	(	.0073	>	.67001	-.1193	<
-.19*9 . .10721
-.1389 > .17211
-.2092 .	.0989)
-.1961 •	.1X0*1
-.2726 f	.0j;0>
NOTE I
IF ZERO IS CONTAINED iflTHIN A 5IVEN CONFIDENCE INTERVAL THEN T^ERf IS NO TTATISTICAL SIGNIFICANCE BITUEEN
RISTILLEO HATER ANO THE CORRESPONDING WASTE HATER FOR TH£ ASSOCIATED PARANETERtINTERCEPT'SIOPEI.
THE SLOPE AND INTERCEPT E STINATES eROM TMIS ANALYSIS ARE NIT THE SANE AS TllCIE 34TAINE0 FR01 THE "RECISION
AND ACCURACY RE6RESSI0NS PERFORMED EA'LIFR.
FOR COMPLETE DETAILS ON INTERPRETING THIS R:PQ«T> SEE APPEIDIt A IN THE PR06«AMNER(S1 DOCUMENTATION.
S£•.	-T ,
	
_ —fyi—-i-
h>'soO;e
IZO^C-'O
-t	, C3 l •> C, -I
~ r v 2 r- I11 O
o u. > o o .
2 L. a? — ni > ^
.. -n m -r.
w. >T'



-------
TARLE 11.
ENVIRONIENTAL MONITORING »HO SUPPORT LA83RAT0RY
OFFICE Of RESEARCH AND DEVELOPMENT
envirinnental PRirrcriON agfncy
•••EPA METHJD VALIDATION STUDY - SRI NITP3SAMNES*
t*ercr IP WATER TYPE 0* N-NITR3S00I-N-P»0PYLANI«IE ANALYSIS
• » POINT ESTIMATES **
DISTILLED WITER SLOPEIGAHMA(II • 1.04130
MATER INTERCSPTIWATER-DISTILIEO). SLOPE IWATER-wISTlLLEOl
t N V S» P»Gf 76 !
-.0394
•	0321
.119*
.1296
•	1490
••0290
-.0278
-.0343
-.1068
-.0T74
nu«te
•• ANALYSIS OF VARIANCE ••
OF SUN OF SQUARES «EAN SQUARE
•¦C(01STTLL*0I	1
RESIWATER/0TSTILLE9) 10
ERROR	472
Tim	483
646.87932
1.94096
12.99233
731.80860
646.67932
.19410
•17383
F PR09
1.10 .3972
»• TABLE OF 991 CONFIDENCE INTERVALS FOR THE DIFFERENCES 9ETWEEN INTERCEPTS AND THE DIFFERENCES 8ETMEEN SLOPES ~•
HATER
INTERCEPT!HATER-0ISTILLEOI
ESTIIATe	INTERVAL
SLO>E(WATER-DISTILLFO>
ESTIIATE	INTERVAL
-.0394	I -.4069 •	.32801	-.0298	(
.0321	( -.3179 >	.3817)	-.0278	I
.1194	I -.2160 .	.4748)	-.036)	(
.1296	( -.2183 •	.47791	-.106a	(
.1490	(	-.2122 .	.9102)	-.0774	I
-.1992 t	.1396*
-.1873 »	.1318)
-.1976 t	.1290)
-.2639 •	.0903)
-.2403 >	.0894)
IF IEBO IS CONTAINED WITHIN A GIVEN CONFIDENCE INTERVAL THEN T-tEBE IS NO STATISTICAL StGN£F'CANCE 9FTWEEN
DISTILLED WATER ANO THE CORRESPONOfNS WASTE WATER FOR THE ASSOCIATED PARAMETER!INTERCEPT/SLOPE)•
THE SLOPE AND INTERCEPT ESTIMATES FROM THIS ANALYSIS ARE NOT THE SAME AS THOSE OBTAINED F»!JN THE PRECISION
AND ACCURACY REGRESSIONS PERF1>"ED FARLIFR.
FOR COMPLETE DFTAILS ON INTERACTING THTS REPORT. SEE APPENOII A IN THE PR0GRANN5RIS) DOCU'ENTATION.
8k!->. •
W>-noO;iv
IjO;c?C
5|:>g5^s
•ztc -
yt >11

-------

'hz
B
S
—1
<
2
2
C>
ci
c
o
m
(A
x
m
m
' H ¦
H*--
>
o
ni »:
3 [:
TABLE 12.
ENVTRONIENTAL N0iIT0*lNG ANO SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL PROTECTION AGENCY
•»~«» OCTHIO VALI OATtON STIJOY - SPI NITROSANINES«
FFFECT OF WATER TYPE ON N-NITROSOD1PHENYLAMNE ANALYSIS
»• POINT ESTIMATES ••
DlSTILLc1) HATER SLJPtiSAMHA(11 • .94289
WATER INTERCEPTtWATER-DISVILLE}) SLOPECWATER-DISTILLEO)
1 M >1
PAGE 11*
-.mi
-.6*13
-.1*02
-.2)31
-.1117
• CI"*
.1901
.0905
.0*17
.0232
MURCE
•• ANALYSIS OF VARIANCE
IF SUN 0« SQUARES MEAN SQUARE
tEGfDISTILLED)	1
RECIVATER/OISTIILEO) 10
ERROR	447
TOTAL	458
215.77399
2.45695
92.104**
310.33939
215.77398
.24)70
.20605
F PROB
1.19 .2937
*• TABLE OF 95* CONFIDENCE INTERVALS FOR THE DIFFERENCES BETWEEN INTERCEPTS AND THE DIFFERENCES 9ETWECN SLOPES •«
WAT'R
INTERCEPT!WATER-OISTILLED)
FSTINATE	INTERVAL
SLOPF(WATER-DISTILLED)
ESTIMATE	INTFPVAL
-.2551	I	-1.1794	,	.6453)	.035ft	I	-.2409 ,	.3321)
-.6813	I	-1.6432	>	.2807)	.1901	(	-.1230 »	.5031)
-.3402	I	-1.2X41	,	.6038)	.0995	(	-.2057 ,	.4027)
-.2833	(	-1.1929	,	.6263)	.0417	(	-.2513 ,	.3348)
-.1117	(	-1.0275	•	.8040)	.0232	I	-.2723 .	.31P7)
NOTE I
IF 2ER0 IS CONTAINED WITHIN A 6TV?N CONMOENCE INTERVAL THEN THESE IS NO STATISTICAL SIGNIFICANCE BETWEEN
DISTHLEO WATER AND THE CORRESPONDING WASTE WATER FOR THE ASSOCIATED PARAPE T£R« INTERCEPT/SLOPE).
THE SL DOCUMENTATION.
-<> -n a o = ra
>o=>£h>h
•H _ j m w CI -H
cf:>'5o^g

-------

- appropriate for a 99% confidence level with n-1 degrees of freedom (3.143
for seven replicates) to give the MDL value. The MDL values obtained are
j given in Table 13. Also shown in Table 13 for easy comparisons with the
! MDLs are the lowest concentrations used in the interlaboratory study.
TAB'.E 13. METHOD DETECTION LIMITS AND LOWEST
CONCENTRATIONS USED IN STUDY

NDMA
Concentration,
NDPrA
H9/1-
NDPhA
Interference-free water, MDL
• 0.149
0.462
0.807
Industrial effluent 1, MDL
0.099
0.359
0.757
Industrial effluent 3, MDL
1 0.121
0.741
1.57
Lowest conc. in study
0.837
1.217
8.216
.•i'UlM
t AST
'.'¦F TEXT
In conjunction with the MDL determinations, analytical curves for the
compounds were determined on duplicate samples at five concentration levels
chosen in conference with the project officer. The analytical curve study
showed that an essentially linear response was observed for N-
nltrosodlmethyl amine up to concentrations of 148 ug/L and for N-n1trosodi-
n-propylamine up to 466 ug/L; these concentrations are far above the
highest concentrations used in this study. For N-n1trosodiphenylamine, a
significant deviation from linearity was observed with an approximate one-
third drop 1n response factor when the concentration went from 8 ug/L to
800 wg/L. However, the concentration span from the lowest Youden pair to
the highest Youden pair in the interlaboratory study was only from 8 tn 55
wg/L, a segment of the analytical curve where essentially linear response
was observed.	I
The lowest spiking levels used 1n the study ranged from five to eight
times the MDL values for NDMA and NDPhA and from two to three times the MDL
values for NDPrA. The smaller difference between MDL and spiking level for
NDPrA had no adverse effect on the assay of this substance, and study
results for 1t and NDMA are about as expected. The results for NDPhA,
although acceptable, are not as good as anticipated. The critical part of
the procedure for NDPhA Is Its separation from the potentially Interfering
compound, dlphenylamine (DPA) and NDMA, on the alumina cleanup column. To
achieve the separation, the alumina activity must be within a narrow range
that 1s presumably attained by adding 2 mL of water to 100 g of Woelm,
Super 1 activity, basic alumina as received. Although this procedure was
n-
i	•_
EPA Fo'm 2350-4 (4-80)
lPMCV»OUSLY CIN. LPA f'ORM 2B7J
f'AGF NUMuLR
	^
it U S OOVfcHNMKNT PHINTIN3 OFFICE 1930-660 <39
TYPING GUIDE SHEET
ft'l 1 OM OF
I VIAGl AFJFA.
MirS'Dh
DIMENSION
POR TABUS
AND ILLUS-
TRATIONS
*Tt 'Alk hliT?





-------
3
L'ECIN
I.-'-ST IJM: ]
OF TE.
EPA Fo ir 21&0-4 «-80)
(PRr:VI 3USLV CIN. LPA FORM 287)
PAGE NUVDFR
BC1 TOM Oc
IMAGE AHEA-
OUTSiDE
imitr.'SiON
POR TAB_IS
I ILLUS-
TRATIONS
dUS GOVERNMENT PRINTING OFFICE 1980-660-639
N.
I: \
TYPJNG GUIDE SHEET



¦ "VV

-------
Another major trouble spot that appeared during the study was the
great variation in response of the N-P detector. This trouble is greatly
reduced by venting the injected solvent, and all participants reporting
difficulties with response were advised to install a vent in their GC
system and to inject standards several times per day.
----- Several laboratories reported that they had great difficulty Injecting
the first fraction of eluate (30* ethyl ether/70% pentane) because of its
low boiling point and suggested that it, too, should be solvent exchanged
with methyl alcohol before analysis by GC.	!
Several participants complained that the temperatures given in the
method for concentrating the fractions in a Kuderna-Dan1 sh apparatus were :
too low and therefore the times required to perform this step were too ¦
long.	j
—
BEGIN
LAST
Or TEXT
£ 3"'8"
!_ __
r

EPA Form 2350A M-801
« (PREVIOUSLY C> ki. CPA .'OHM 2»7)
PAGE NUMBEH
BOTTOM OF
fW^GT; AREA.
OUTSIOE
oifvEwnioN
FOH TABLFS
i MCj ILLUS-
TRATIONS
ft U S GOVERNMENT POINTING OFFICE 1MHWC0-C19
; /TYPING GUIDE SHEET	^ if *
'A" >• ^ I" -5'1..-
•1 -	• s' c -va, ¦ • .,jT; %*>/:

-------
- '.ft- <-**Ty;VTyfrT
r"fl	-

REFERENCES
1.
2.
3.
4.
;5.
6.
7.
8.
9.
10.
Code of Federal Regulations, 40:Part 136, October 16, 1973.
U.S. Code Congressional and Administrative News, 3, 1977, 95th
Congress, First Session 1977.
Federal Register 44, 233, 69496-69499, December 3, 1979.
Youden, W. J. Statistical Techniques for Collabwrative Tests.
Association of Official Analytical Chemists, Washington, D.C., 1969.
Burke, J. A., J. Amer. Chem. Soc., 48:1037, 1965.
— ^**-1
BLGIN
Outler, E. C. and McCreery, J. H. Interlaboratory Method Validation
Study: Program Documentation, Battelle Columbus Laboratories, 1982.	,
Thompson, W. R. Annals of Mathematical Statistics, 6:214, 1935.	j
Draper, N. R. and Smith, H. Applied Regression Analysis. Wiley, New	j
York, 1966.	!	-	j
!	|
Bishop, T. A., Brydon, F. E.; and Outler, E. Development of	'
Appropriate Statistical Techniques to Compare Analytical Methods
Across Wastewaters. Battelle Report to Environmental Protection	i
Agency.	1	j
1	I	i
Rhoades, J. W., Thomas, R. E.^ and Johnson, D. E. Determination of	;
Nitrosamines 1n Industrial and Municipal Wastewaters. EPA Report
EM3L-141,,October 1981.
. L/-27 sp-
ur- iuxt
'¥i

EPA Form 2350-4 (4-80)
(PREVIOUSLY CIN. LPA FORM 287)
PACir. KU.MCf.K
BOTTOM OP
IMAGE ARKA:
OUTolOC
V" DIMENSION
FOR TABLES
	I AND U.LUG-
THATIOfcS
¦0 U.S. GOVERNMENT PRINTING OFHCt 1 SSO-tSO-tM
TYPING GUIDE SHEET


-------
m
*o
>
i' !
...
hr
INS-
TABLE A-l.
r *
CO
to:
3-
2 •
*2- "•
C)
0
c
6
m
c/?
1	'
m
m .
o
. <
o
-S L
-f
u«
c>l*TRnmtENTAl (10XIT0JIHG ANO SUPPORT LABORATORY
OFFICE Oc RSSEARC* *N0 0EVF10»"E*T
ENVIRONMENTAL PROTECTION A3ENCY
1STH00 WiLIOATIDM STUOY - SRI NITROSAHINES*
data mo N-NiT«o?joineTHri4miE analysis by water type
t 1 V St PASS ?0
HATE*
KITES 2
WATER 3
WATE»
WATER
4ATER
ion TOUOEH PAIR
3
3
*
3
3
3
3
3
3
3
3
3

TRUE VALUE UGH
0*6%
1.01
0.94
1.01
0* 64
l.Cl
0.64
1.01
0.84
i.01
0.84
1.01
f
LABORATORY NUH8ER












<
701
I.03*
1.03
1.02
1.97
1.00
1.13
2.80*
1.73
1.00
1.30*
0.31
1.10

T02
0,1«
0,090
0*38
0.21
1.44
0.41
C. 24
0.21
0.29
0.21
0.6?
0.41

709
0*69
0.61
0* 39
0.43
0.29
a.ci
0.56
1.20
0.44
0.48
0.17
0.44
1
704
0*43
0.42
0.31
0.31
0.21
0.21
0.40 '
0.47
0.49
0.34
o.eo
0.97

703
0.70
0.91
0.72
0.92
0.34
0.99
0.33
0.93
0.39
O.Q6
0. 71
1.02
TO ^9 ' .Y
706
0* 3b
0*46
0*43
0.39
0.43
9.68
0.43
0.93
0.37
0.49
0.49
C . 4 3
> "O -
707
17.03*
6.9*0
•
2.14
33.00*
6.40*
26.60*
29.80*
9.16*
2.40*
4 ? « 90*
7.00*
C m —t
708
' 0*29
0.37
0.*4
0.36
0.30
0.36
0.27
0.36
0.28
0.44
•
0.39
•o o •
700
0.11
O.C*
0*13
0.64
0.«l
0.30
0.31
0.49
0.13
0.29
1.09
1.66

710
0*21
0.33
0* 31
0.36
0.23
0.38
0,23
0*27
0.33
0.26
0.40
~
—1 X ^
711
0,21
0.30
0* 34
3.43
0.28
9.33
0.34
0.30
0.29
0.38
0.49
0.98
3>
712
0.30
0*39 «
0.30*
0.30*<
0.30*<
0.3d*<
0. 30*<
0.30*<
0. 30*
0.31
0. 30
0.46

713
0*30
0.49
0*30
0.39
0.30
0.30
0.40
0.30
C. 30
0.40
0.30
•

714
0.20
0.30
0. 40
0.60
*
•
0.20
0.30
0.20*
0 >20*
0.10
0.20

713
2.40*
2.20*
2* 10*
2.10
2. 30*
1.60*
0.70
0.80
0.90
1.0C
2.60*
1.90

716
0.27
0*29
0*24
0.34
•
0.28
C. 22
1.61
0.22
0.36
9.11*
10.0J*

717
0*29
0*43
0.81
0.29
0.83
0.20
0.67
0.41
0.*6
0.22
1.10
0.36

• • REJECTED












t
tfATER LEGEND













DISTILLED WATER
TAP WATER
SURFACE WATER
WASTE WATER 1
rfASTE WATER 2
WASTE WATER 3
I !

H > -n o O
JZOjt
-•
oP>wU
2 cl a 7\
u> h r C
Is
Cl -i
mo
33 o
rn T1

''I


-------
V. '
S\
K'.
rv
f.u

K' ¦
-< '
' j2
2
O
a
c
O '
,m
w
x
. m
m,
H
X
m
3
N
>
<
T1
O
0
c
1
V)
r

<
•8
o
j>
2
s

g
>



0

i

e>

fr
fv
¦
l^ii
>
o
c t-o
rn
3
O
o
i i
TABLE A-l (CONT'D)
I 1 Y S« P436 II
MHIRimEMTAL NONtTORINS »N1 SUPPORT LARORATORY
JFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL PROtECTION AGENCY
•*»cpi METHOD VALIDATION STUDY - SRI NITROSAMINES*
RAH OATA FOR N-NITR0S30IMETHYLAMINE ANALYSIS BY HATER TYPE

HATE*
1
WATER
2
¥476*
3
bATE*
4
tfATE*
9
4ATE*
6
NEOtU* VOUDE* MI»
I
2
t
2
1
2
1
2
1
2
1
Z
TRUE VALUE (It/L
6.70
••04
6*70
9,06
6.70
9.06
6.70
8*06
6.70
8.06
6.70
8.06
IA80W0RT NUKSER












701
1.16 «
0.20*
2.34
3.49
0.31 +
0.88
0.38
1.42
0.97
1.02
2.00
1.98
*02
2.3)
U91
1.90
1.44
2.13
>.61
2.91
2.64
2,09
2 * *1
2.64
3.16
TC3
2*69
3.29
7.79
3.49
2.92
4.34
3.27
3*91
•
2.79
2.44
2.39
70*
2.21
2*3?
2.63
2.66
2.02
2.23
3.02
2*90
2. 32
3.90
2.97
4.02
7w>
3.97
%.32
3.36
4.32
3.68
4.09
3.30
4.73
3.33
4.34
3.6«>+
*.76
T06
3.22
4.41
2.T2
2.74
3.07
1.18
2.93
3.67
2.89
3.13
2.23
2.40
707
30*00+
38*49+
29.19+
20.00+
46.20+
69.10+
91.00+
47.70+
37.70+
18.70+
33*10+
69.90+
70S
2*79
3*11
3.19
3*66
3.84
4.90
2.BC
2*49
2.90
3.99
2.37
2.61
709
3*01
3*61
1.23
3.00
3.91
7.42
1.92
1*93
1.64
1.98
2.9)
4,tfl
710
8*19 +
10.70+
11.60+
19.99+
9.31 +
19*60+
4.07
11.60+
6.40+
19.40+
10.93+
19.90+
711
2*19
2,92
7*63
3.99
3.19
3*94
2,06
2.86
2.89
3.79
1.97
2.99
712
3.72
3*10
1.97+
1.47+
1.19 +
1.11 +
0.69+
2.96+
2.31
2.71
1*72
1.96
713
2.30
2.90
2*39
2.90
2.»0
3*60
3.80
3*90
3.10
3.90
4. 60+
3.10
71%
2*99
2*10
3.90
».90
~
*
2.90
1.90
1.40+
1.90*
2.39
3.70
719
*•90*
3.80+
3.C0
J.40
2.80
8.60
1.90
3.00
1.4C
1.90
1*60
4,90
716
2.2%
2*81
2.6 9
9.46+
1.88
2.34
3.96
~
3.63
7.90+
2.33
3.68
717
2.20
j«eo
1.79
3.40
2*60
3.40
2.90
3.40
2.80
1.80
2.20
9.'«C
I • • REJECTED
' MATER LfSKO
1	- DISTILLED MATER
2	- TAP HATER
: i - surface water
' 4 - HASTE HATER i
' 5 - HASTE HATER I
| 6 - HASTE HATER }
f; 1
8 ¦ I
¦"	.
H > f> C O ^ M
X ^ O r, c =- O
—i	1 i?
o jT. > cs -5 !T 5
Z JL EC — n ; J>
c/> 5 1- 9 33 o
V5 fV «c fTl TI
v> >
._LK.



-------
=132
7 /*, .
if .

S:-:
i
> v.'
K:\.'
¦<
2
'
o
o
c
c
rn
in
X
.rn •
j-etJni
r-
A
• i
i *
TABLE A-l (CONT'D)
> (:¦¦¦:•:¦
II
SK
I'-:-
Enrtaii^EHTti houtorug »ho support laboratory
OFFICE Of RESEAiCH 4N1 OcVELOPMENT
EHtflRflVXFNTAL PROTECTION .*GESCY
•«««PA "ETHOO VALIOATIO* STUB* - S»I N1TR0J*"INES»
RAW 0»TA f}» N-NITHQSlDINETHTLttlMF ANALYSIS BY WATER TYPE
I fl V Si base 22

HATES
1
H4TER 2
HATES
3
WATER 4
WATER
9
tfATER
6
t
HIGH VOUOE* PAIR
4
6
4
6
4
6
4
6
4
6
4
6

TRUE VALUE U6/1
20*09
2%.19
?0«09
24*19
20.09
24.19
2C.09
24.19
20.09
24.19
20.09
24.19

LABORATORY NUMBER












1
5
701
~
7.1?
~ .19
8.29
2.09
4.14
7.43
9.13
69,706
6.17
9.2?
9.14
1
702
6*99
2.21
7.04
6*63
7.49
3.79
6.27
9.92
9.93
1.79
6.22
3.67
1
793
6*79
9.79
7.M
9.27
6.26
6.97
9.29
9.21
13.90
ft. 99
9.92
7.9J

704
6.*2
6.0^
6.91
8.90
7.31
10.20
10.10
12*00
8.91 <
8.84*
9.24
13.10

70S
9*50
19*90
9. B0
19.00
6.89
14.90
7. 69
13.80
7.39
14.40
9. 3*
12.30

706
6*73
9.91
9.46
13.10
9.21
10.67
9.72
10.30
6.93
P.27
8.29
10.40

707
49.40*
46.60*
66*096
161.906
93.106
37.40*
134*706
210.90*
l4e«t>o»
63.706
216.70*
94.90*
—!
7011
6*69
1*41
9.4»
9.99
9.64
9.43
9.29
6.39
7.9*
10.20
10.90
2.99
i
»
739
*.7*
7.99
4.44
6.74
7.44
9.74
4.03
4.66
4.93
t. 9*
5. 96
12.10

710
6*17
12*90
7.19
8.02
o.OI
13.40
7.89
9.10
7.48
6.66
r«04
6.97

711
7. 36
9.00
9.9®
10.63
8.91
13.70
8.19
9.39
9.10
10.10
9» 19
8.93

712
9,49
10.61
7. 616
.1*036
3.826
7.P96
9.606
8.196
4.?»
3.69
2. 99*
6.06

713
7.80
10.23
10.10
10.40
8.90
13.19
8.60
11.40
6.70
10.90
8. 50
13.90
!
71%
9*00
10.40
A. 30
9.90
~
4
3.90
•
3.406
9.10*
6.03
4.60

715
10.*0*
19.00*
7.00
13.90
14.90
17.40
0.10
9.10
6.10
10.90
11.80
12.60

716
17.90*
9.84
9.64
10.20
14.00
7,41
16.60*
19.70
11.10
20.90*
17.10*
14.90

717
9*10
7*10
e.oo
3*80
8.10
4.99
7.10
6.60
7.70
7.20
6.40
9.70
i
t
IJl
• • REJECTED
WATER LESEHO
1	- 01STIUED HATER
2	- TAP WATER '
1 - SURFACE W3TER
* - WASTE WATER 1
5	- WASTE WATER 2
6	- WASTE W4TFR ?
'S:;0^c JO
52 = 5^3
£ U ® — m £ —
""Si 3?
Cfl >
*^>	1



-------
f> r- n
t > rr»
21)
n
*T3
>
c
3
K
8

hr

H
¦•¦ <
¦TJ
—	O
O	j
CD	f:
C	S
O	5
it.	;
v-	I
5'	c
El- ' ?
rn	c,
' J
L>
~ • REJfCTEO
HATES IC6EN0
TABLE A-2
I 1 V Sl *A3E 99
SHVIRIIIEHTAL N01TTQRING ANO SU»°0RT LA43RATQRY
OFFICE OF RESEARCH ANO OtVELOP^ENT
ENVIRONMENTAL PROTeCMOW AGENCY
•«•?»« ^ETHOO VALIDATION STUOY - SRI NITROSAMINES*
RAW DATA F09 N-NtTRQSODI-N-PRODYLAMNE ANALYSIS BY HATE# TYPE


WATER
1
WATER
2
WATER
3
tfATER
4
WATER
9
WATE*
6
LOW YOUDEN
PAIR
3
9
3
9
3
9
3
9
3
9
3
9
ir.Vl VALUE
U6/L
i.*e
1*2?
1.40
1.22
1.48
1.22
1.49
1.22
1.48
1.22
1.49
1.22
LABORATORY
H'JP»BE«









1.90*
1.69
1.09
701

0.89
0*96
0.9T
9.699
C.96
0.29
2.93
1.33
9.93*
702

1*70
1*21
1. 84
1.23
2.47
1.23
1.92
0.94
I. 79
1.13
3.71
1.60
70S

2.1R*
1.6*9
1.09*
1.69*
1.909
2.17*
1.90
2.23
2.02
1.39
2.00*
2.46*
70%

1.29
1.26
0.99
1*00
1.40
1.20
4 • 83*
2.3»*
3.09
2.03
4.23*
4.07*
703

1.68
1.71
1.739
i.999
1.92
1.99
1.69
1.63
1.71
1.64
1.67
1.90
706

1.03
1.09
1.39
1.20
1.79
2.27
1.30
1.14
1.17
1«11
1.92
0.9*
707

*
9
9
*
9
#
*.79
1.99
1.98
•
3.69
1.36
708

1.54
1*29
0.77
•
1.96
1.62
1.63
0.98
e.tv
1.17
4
0.92
709

0.7*9
0.64*
0.42
0.69
0.81
0.86
0.76*
0.669
C. 67
0.79
0.86
0.73
710

0*39
0*63
0.79
0.62
0.77
0.69
0.89
0*97
0.68
0.46
0.93
9
711

o.re
1.19
t.O*
l.oa
1. 19
0.69
1.99
1.99
1.03
1.71
0.72
e.97
712

1.29
1.19
0. 70*
1.67
1.18
0.80
1.14
1.00
0.80*
0.62
0.83
0.90
713

1.40
1*30
1.40
1.10
0.90
1.40
1.80
1.40
1.60
1.40
1.00
0.90
71%

1*00
1*00
1.10
1.90
1.20
1.20
1. 90
0.70
0.90
0*i0
0.90
2.34
719

3.00*
2.39*
4.499
3.109
9.909
2.10
9.10
4.109
9.60*
2. 80
3.70
9.10*
716

1.93
1.26
0.99
1.11
•
1.14
0.90ft
0.96*
1.94
1.37
4
4
717

0.79
1.70
1.70
0,99
1.80
0.34
1.30
0.02
1.30
0.62
1.70
0.45
OISTILLEO WATER
TAP WATER
SURFACE WATER
WASTE WATER 1
WASTE WATER 2
WASTE WATFR 3

~i> Tl O O =: 03
X2D *-C C ^ O
>	O 2 ^ -1 ^ "j
5p>££"9
>	f_ u — m > S.

—-~Jy.
' c
. c -

-------
Orr,
-n >
m
vI
"l rr
~rn
X..: ¦
L- -
' ?
s i
o
- jj
- z
£.=
H
3-,;. o
ijT
o
,m
co •
X
m
m
.H
*-;
)¦¦ A
&
r^v„,
¦'
!&•
f: ...t4.V;'
j-<3C -
•. s
«/

TABLE A-2 (CONT'D)
Elft'imEITAL KONITO'm ANO SU»»0»T L«q09AT?IRV
OFFICE OF RESEARCH AH") DEVELOPMENT
ENVIRONMENTAL PR3TECTJ9N AGENCY
»EH30 VALlOATlcm STUOY - SRI MITRQSAMHES*
RAW OATA FT N-NlTR1S0Ot-N-«0»Y'.A9INE ANALYSIS BY WATER TYPE
I 1 Ml PAGE 54
• ¦ OEJECTEO
WA'EB l«Sf»a
1	- OISTILtfO WATER
2	- tap water
3	- SURFACE WATER
4	- WASTE WATER 1
9 - WASTE WATER 2
6 - WASTE WATER 1
H>Jco;n
l?0;c^0
¦J> U 33 5: > "1
H _ m to O -}
z V- T. — ITl > -=»
w £ r 5 3 o
.. ** rn .2 n> in
cn > ^
CS>:
1
1

WATER I
WATER
2
WAT'S 3
WATER 4
WATES
5
WATCfl
6
1

' NEOtim YOUOfl PAIR
1
2
1
2
1
2
1
2
1
2
1
2

1
1
TRUE VALUE US/l
10.39
8.5»
10.19
8.52
10.39
9.52
10.39
8.92
10.39
8.92
10*39
8*«2


1 LABORATORY NU"«ER










15*75



701
12.19 <
0.40*
12.63
9.31
6.79
3.82
13.10
9.00
13.90*<
10.60*
9.8*

1
| 702
6.57
4.SI
7.19
5.35
7.H
4.24
6.79
4.83
7.09
3.94
10.69
4*00


¦ | 70S
11.00°
11.60*
11.20*
9.46*
15.10*
14.30*
• 13.AO
10.90
•
8.61
11 * 50*
13.70*
;

i 704
10.50
9.99
9.41
i.l»
10.90
8.C8
13.10*
8.46*
11.30
11.40
15*20*
12.10*
I

70S
12.40
12.50
12.50*
11.69*
12.80
13.50
11.50
13.50
11.10
12.90
12*20
13.60
i

! 706
e.63
8.74
5.99
7.83
11.30
8.97
10.20
7.74
9.64
7.57
A* 92
7.27
; j <

J 797
29.10*
9.37
4.46*
0.98*
6.73
11.70
11.90
6.92
5.19
3.26
c.ei
12*70
1	1 -V!
W';
1 70S
10.20
8.26
0.02
6.45
11.30
8.89
9.48
7.34
5.S3
6*66
9. 3*
7.91
1 » ' '
|rY:

709
5.47*
5.21*
5.71
4.43
5.43
5.01
6.00*
5.86*
7.17
5 30
6*e*
3.67

; 710
10.00
19.10
9.95
7.59
7.09
l».49
10.90
11.10
10.10
9,86
5*77
10*60
: 1 -

i 714
7.44
t.6«
9.82
8.40
10.50
8.10
8.82
6.88
1C.10
10.70
6*67
5*92
i !

712
12.09
6.0!
6.30
4.10
6.47
5.14
3.70*
7.40
B, 77
9*00
6* 3&
4.39
|

713
9.30
9.00
9.70
7.70
10.90
a.90
19.10
10.50
12.60
10*09
11.13
11.00


71*
B.10
5.60
9.10
6.70
11.10
8.10
10.70
4.20
*.80
6.90
16.50
16.00


713
16.10*
10.00*
9.40
8.70
9.60
8.70
8.80
9.30
7.60
10*09
9.60
9.60
t
1"
716
12.80
9.0«
11.10
3.60
6.45
7.68
3.58*
»
9.72
e.6*
10*90
10*20


717
7.40
7.90
19.10
1.60
11.60
7.50
15.60
8.80
13.30
9*70
10.30
11*60
' i
¦i—^
t> -r
¦x?o
J


-------

prr-r
p;..

Ji'A.

- ¦ **
 r
o {¦:
11
m y.-
J i '-.. -
HIGH TOUOE" *«It
TRUE VALUE UC/l
LABORATORY NU«»ER
701
702
7CJ
70*
70S
7C6
- 707
70B
709
710
711
71?
713
71*
7H
, 716
717
! » • rejecteo
i WATER LECEN0
A
TABLE A-2 (CONT'D)
CNVTVKOMTtL "iHITORIHC *10 SUPPORT LAB0R AT0RY
OFFICE OF RESEARCH AN! 0*¥fLOPKENT
ENHIRONHEMTAL "RITfCTIn* AGENCY
•••sot "ETN00 VALIDATION STU0V - SRI lITPflSAMNES*
RAH OATA FIR M-1ITR1S00I-N-PRQPVLA1 the ANALYSIS BY HATER Ti PE
I 1 H <« PAGE »0
£
o
&
f
s •
DISTILLED HATE*
TAP HATER
SURFACE WATER
HASTE HATER I
HASTE HATfR 2
HASTE HtTER 3
Sgc'ISp
> u "¦» 2; -I ?
¦H	I 3 in -• H
rr f C? ^ T> s j
wSrQ ~ o
m -i rr »p
Cfl >
HATE#
1
tflTER
2
WATER 1
HATE*
4
WATER
9
4ATE*
6
i
4
6
4
6
4
6
4
4
4
6
4
4
j
26.71
17*04
*6.71
17.04
26.71
17.04
26.71
17.04
26*71
17.04
26.71
17.04
\
j
1
#
17.09
29.90
20.41
20. 10
12*50
40*10
21.10
\10.00*
29.00*
IT. 30
17.30
!
22*10
11*20
19*00
11*90
20.00
14.00
20*90
14*90
21.90
11.00
26.60
16.90

31.40*
19.90*
11*20*
22*10*
14.90*
26.90*
11*00
19.10
17.90
14.30
16.40*
24.10*

24*19
16*70
t»*74
19*00
11.00*
21.10
19.20*
24.80*
2^ * 40
15.20
2T.40»
26.10*
' i
j
27*40
21*90
11*10*
21» 70*
11,00
23.60
11*60
21.90
20.20
22.70
34* 9 J
21.20
|
23.20
22.69
19.91
12.20
27.00
16.00
27.00
16.30
29.20
14.00
30.30
21.40
i
1
28*60
22*10
l.«l*
1*99*
2.79*
4*10
15*90
22*90
7*39
2.71
T.0*
2.37
29*70
16*60
«*70
10*20
27.00
17*60
29. 00
14*10
26.00
IP.30
19.20
14.^0
14.30*
11.50*
19.20
10.90
19.40
1>*4C»
17.30*
4*01*
14.60
10.10
17.40
13.30
!
10*10*
7.81
19*49
4.69
16.60
6*00
11* *0
6.49
14.40
9.96
IT,60
9.S4

21*90
11.90
27.00
16.00
22.10
10.CO
21*10
14.00
29.90
16 . 30
21.49
19.00
t
25*41
13.21
?fl.2l
4.00*
11.01*
19*20
21*24
19.00
14*09
10.10
13.90*
10.23

27*20
16.73
29.70
17.90
29.50
26*10
12*10
21*10
32.00
10.30
25.60
21.40
J
21*70
16.10
>1*20
16.40
21.20
17*70
16*10
«
2.00
19.90
16.70
14. 50
•
29*00*
17.20*
22*70
19.90
27*30
19*90
26*10
17*70
10.10
17.10
20.90
16.20

2T.ro
16*40
24* *0
19.00
26*20
19*20
10*60*
7*91*
26*10
21.90
20.70
#
\
16.40*
11*40
29*20
7.60
27.60
11*70
27*40
11.90
27.90
12*40
29.90
lfr.10
i
	l_
i	.	i	-	- - 		 Mm in i	• • '••	~—'""¦ 1 ¦¦¦¦— 			
jgSfcs«agS8affig^aiMa«gSBaaBg^^^es«attaB&as8^Mateg«aBgaBBafceBiaKaBaa

-------

TABLE A-3
MvnimfNTAi Nn4iTn«i«e and wort laboratory
OFFICE OF RESEARCH ANJ DEVELOPMENT
ENVIRONMENTAL PA3I6CTION AGENC*
~»*EP« NETM1D VALIDATION ST'JOT - SKI NITPOSAMNE J»
01 r* FOR M-NITROSaDIPIENVlAMIE ANALYSIS BT WATER TYPE
1 « SI »AGE lb
LOW TOUDEN PUR
TRUE VALUE US/I
LtlGSATORV NUH«E«
TO I

1

2
WATE9
9
HATER
4
WATER
9
HATER
6
9
9

9
9
. 9
9
9
3
9
3
9
e.«
10*94
Q.I!
19*99
8*22
10.96
8*22
10.96
8.22
10.96
8.22
10.96
9*04
66*80*<
9.00*<
9.09*<
9*00*
13*60 «
9.0C*
12.60 <
9.00*
4.90
17.90*<
4.00*
4*69
9.21
9*11
9*96
9* 31
9.21
4*36
9.79
3*31
4.38
9.27
3.13
o.eo
7.69
9.94
9.41
10*90*
12.90*
3.11
6.39
8.39
14.60
6.72
q.4i
6*17
0.4?
9.7t
6.00
9*92
7.17
4*14
7.94
9.97
b.34
6.n
8.0)
10*60*
I2»ee*
9.19*
12.10*
9*50*
19*00*
1C.70*
12.00*
9.95*
12.80*
11.00*
12.30*
0* 10
9*10
T.47
9.9J
6*96
9.«#9
9.54
6*90
4.97
6.44
>.9J
9.06
#
•
*
t
0
•
0
0
•
•
•
•
7*94
11*60
9.69
0
11.10*
14.40*
9.60*
14.00*
9.67
13*30
•
11.10
9*91
6.91
4.94
9*91
9*49
7*29
6.09
9.A6
6.7?
8.74
10.30
8.44
i*e3
2.29
1*06
I.*7
2,91
2.40
2.39
2.66
7*32
3.43
0.64
0.94
9.29
4*0*
9.7*
4*99
4.21
4.2*
9.98
3.90
4.29
6.41
3.99
9.29
6*09
8.09 «
I.00*
7.90
9*39
6*92
6* 90
fi4 33 «
1.00*
2.67
7.93
9.96
7.10
9.20
6.29
6*90
6*90
9*90
2.^0
3.00
2.90
4.8C
2.99
9.10
6*70
7*69
2*40
9*49
0* 90
7.00
7.40
10.30
0.60
4.60
4.30
10.93
19.60*
17*60*
7.10
9*90
9.40
7.26
9.00
9.60
9.30
6.90
6.93
0.30
2.91
9*66
1.97
2*0%
*
2.P4
1.18*
2.76*
2. 38
3.04
2.44
3.96
6*90
8*20
ll.no
9.60
0*70
3.10
6*30
0.00*
11.00
7.90
8.60
10.4g
REJECTED
MATER LEGEND
01*TILLED MATER
UP MATER
SURFACE MATER
WASTE WATER 1
WASTE WATER 2
WASTE WATER 1
!!:::
i1

i
~i— ^


-------
¦ ¦¦I |UBI IT IT»W"	I
C- r- co
-n > rr,
.Jf
3
r.
.<
O
c

•<
"T7
: 5
o
a.
a
o
m
oo
X
m
'IT!
H

*
&
* -«>-
I to
j-KU _
l

" V"!
TABLE A-3 (CONT'D)
o>.:

, o — •
EHmimCNTM. t "n c O 5. do
IZ0;c?0
> C3 S; -H 5 "1
h		 rn t/> O -H
PISSfV-,
2C£^m>-
«5>C9 '2iCJ


-------
~i2
3
n "
5 ?
c 3
p u
r $
z s
4l.S
<.
2
2
a
a
c
5
rr.
co
X
rn
rn
hr
h*'

TABLE A-3 (CONT'D)
- i
*WT*"*flcNTAL **n>ITTQRING 410 SUPPORT 14H0»AT0RY
1FFIC6 fl* RESEARCH AN*) OEmO**EHT
ENVIRONMENTAL PROTECTION AGENCY
MKO* 1ETH10 VALIDATION STUDY - SRI NITROSANINES*
RAH OftTA FOB N-^ITROSJOPMENYLAMNE ANALYSIS *Y WATER TYPE
I * V SI page 96

WEB
E 1
tflTER 2
tfATEfl
9
«ATE* 4
WATER
5
4ATER
6
HIGH T0U9CN PAIB
4
6
4
6
4
6
4
6
4
6
4
6
TRUE VAUJE U6ft
~ 1.03
5* .7*
41*08
54*79
41*04
94*79
41.06
94.78
41.08
54.7ft
41.06
54.78
U80RAT0R7 NUNQcR
_











701
10*10 <
3*00*
22*60
3*47 <
4«00*<
4.CO*
61.20*
74.20
91.10
44.60 <
4.GO*
25.20
7C2
20.90
19.70
13*90
23*70
19*70
28.20
19.70
25.40
11.30
23.00
13.30
36.30
703
96*00
94.20
96.69
49*70
33.60*
79.40*
3b. 20
47.70
91.80
48.40
40.60
52.80
704
32.00
49*70
13*59
44*00
0
4J.PG
i:.oo
47.20
22.00
35.10
29.90
28,-0
709
94.00*
70.50*
54.40*
76*50*
59.00*
79.80*
99.20*
73.10*
96.10*
70.90*
95.60*
73.40*
796
36*40
25.20
36*89
31*90
26.70
24.10
18.30
42.50
20.50
38.90
30.30
*3.00
707
~
~
*
$
0
•
~
~
•
•
*
•
708
42*40
>8.19
16.00
38*60
46.90*
60.90*
90.10*
93.00*
42.80
67.90
33.20
14. 80
709
31*00
25*39
19*50
42*70
39.20
42.70
39.90
13.10
31.30
il.10
39.40
4?.50
710
7*69
14.19
9*93
13*90
7.59
16.00
10.70
19.60
10.20
17.30
7.35
16.30
711
25*40
25.49
26*50
26*90
24.70
44.10
16.00
27.00
29.60
31.10
23.10
27.73
712
34*60
44.60
13*64
12.99*
18*46*
36.94
26.33
31.33
12.90
3.33
19.31*
30.62
713
33*90
47.30
36*10
49*90
90*20
51.20
22.30
32.60
29.10
35.00
14.70
30.90
71*
29*70
45*40
18*30
36*70
26*00
40.90
24.90
46.90
12.20
?9.10
20.60
18.90
719
49*90*
54.890
30*90
39.00
15* 90
24.80
35.90
62.10
21.60
37.70
37.00
46.70
716
11*20
10*00
7*88
8*03
12*10
u.eo
6.33*
9.96*
10.40
19.80
13.80
10.40
717
19*90
47*40
26*30
22*60
18*90
9.00*
27.00
42.00
29,70
7.90
28.80
43.70
-v.- 8
L^l
• ¦ REJECTEO
WATER LE4ENO
0ISTILLE0 HATER
TAP MATER
SURFACE WATER
HASTE rfiTER 1
HASTE MATER 2
MASTE MATFR 1
.1


«
32C;cgO
> a » i ¦ - (> -i
-H _ _j r'^ i/> O H
^ r- w ^ W o
" O P > in O v j»
crthrO o
V* r.i rn
Cn

.? " -o
' ¦

-------
I. • c
APPENDIX B
NITROSAMINES
METHOD 607
"> 1. Scope and Application
1.1 This method covers the determination of certain nltrosamines. The
following parameters may be determined by this method:
Parameter	STORET No.
BEGIN
LAST LINf:
or itxr

N-n1trosodlmethylamine
N-n1tro sod1pheny1 ami ne
N-n1 trosodl-n-propy1amine
34438
34433
34428
1.2	This method Is applicable to the determination of these compounds
In municipal and Industrial discharges. It 1s designed to be used
to meet the monitoring requirements of the National Pollutant
Discharge Elimination System (NPDES). As suchr it presupposes a
high expectation of finding the specific compounds of interest. If
the user is attempting to screen samples for any or all of the
compounds above, he must develop independent protocols for the
verification of identity.
1.3	The sensitivity of this method 1s usually dependent upon the level
of Interferences rather than Instrumental limitations. The limits
of detection listed 1n Table I represent sensitivities that can be
achieved in wastewaters 1n the absence of interferences.
1.4	This method 1s recomnended for use only by experienced residue
analysts or under the close supervision of such qualified persons.
1.5	The analyst must understand that nltrosamines are known carcino-
gens. Utmost care must be exercised in the handling of materials
13/!L _ i	
EPA Form 2350-4 '4-SO)
(PttEVIOUStV CIN. CPA rORM 287)
i;ag;:48-
page number
nOTTOM or
I IMAGE AREA;
—outside
DIMENSION
1 FOR 7A3LES
	I AND II.LU3-
~ RATIONS
: L' S GO'hI'HHVEn; PWITin'S r:il ICL 1*130-600-639
typYng guide sheet


-------
DGGlN	i
USTUMCi
CJ-" "i Ex I
which are known or believed to contain nitrosamlnes.
2.	Sunroary of Method
2.1	A 1-liter sample of wastewater is extracted with methylene chloride
using separatory funnel techniques. The extract is dried and
concentrated to a volume of 10 ml or less. Depending i:,:on the
nitrosamlnes being measured, a column cleanup procedure may be
required. Chromatographic conditions are described which allow for
the accurate measurement of the compounds in the extract.
2.2	If Interferences are encountered, the method provides selected
general purpose cleanup procedures to aid the analyst 1n their
elimination.
3.	Interferences
3.1	Solvents, reagents, glassware, and other sample processing hardwara
may yield discrete artifacts and/or elevated baselines causing
misinterpretation of gas chromatograms. All of these materials
must be demonstrated to be free from interferences under the
conditions of the analysis by running method blanks. Specific
selection of reagents and purification of solvents by distillation
In all-glass systems may be required.
3.2	Interferences coextracted from the samples will vary considerably
from source to source, depending upon the diversity of the indus-
trial complex or municipality being sampled. While general
clean-up techniques are provided as part of this method, unique
samples may require additional cleanup approaches to achieve the
.	i
sensitivities stated in Table 1.
.3.3 It 1s necessary to remove dlphenylamine from the sample extract sorrow of
,			. IMAGf AREA;
	,-J OUTSIDt
DIMENSION
1 Fori TABLES
TRATIONS

!_ i _ _ I	
EPA Form 23504 W-ROi
(PREVIOUSLY cm.	FORM 2A7J
('AGE NUMBER
iv u -i bu^sw :n "HiNr omof tgso-f-oo-bja
TYPING GUIDE SHEET
JJ

-------
1
-'t GIN-
LAST Llfib 2S
oFTcXr
I
prior to gas chromatography because 1t will interfere with the
determination of N-nitrosodiphenylamine. Removal is achieved if
the sample is processed conpletely through one of the clean-up
procedures detailed 1n the method.
Apparatus and Materials	[
4.1	Sampling equipment, for discrete or composite sampling.
4.1.1	Grab sample bottle - amber glass, 1-I1ter or 1-quart
volume. French or Boston Round design 1s recommended.
The container must be washed and solvent rinsed before
use to minimize Interferences.
4.1.2	Bottle caps - Threaded to screw on to the sample
bottles. Caps must be lined with Teflon. Foil may be
substituted 1f saople 1s not corrosive.
4.1.3	Compositing equipment - Automatic or manual compositing
systen. Must Incorporate glass sample containers for the
»1 lection of a minimum:of 2S0 ml. Sample containers
must be kept refrigerated during sampling. No tygon or
rubber tubing may be used 1n the system.
4.2	Separatory funnels - 2000 ml and 2S0 ml, with Teflon stopcock.
4.3	Drying column - 20 mn ID pyrex chromatographic column with coarse
frit.
4.4	Kuderna-Danlsh (K-D) Apparatus
4.4.1 Concentrator tube - 10 ml, graduated (Kontes
K-570050-1025 or equivalent). Calibration must be
checked. Ground glass stopper (size 19/22 Joint) is used
to prevent evaporation of extracts.
^3/8"
-7	~
1	

EPA f orm 2350-i (4-80)
(PRE'«'IOU***LY ClN. CPA FORM 28?)
P/.'oE NUMBUH
COTTOM OF
IMAGE AREA,
J cursiOE •
DIMENSION
> FOR TAF.LES
I AND ILLUS-
TRATIONS
V»Or. OOVLHNV.SNT rH!Nll\'G Oi-"eir5 660-Gin
TYPING GUSDE SHEET

-------

i"L\ Ti i"
¦ ;v- -j-
.'•» 'A
3cGlf-.'
I.-ST l.lii
0>" T'.XT
?.»
4.4.2	Evaporative flask - 500 ml (Kontes K-57001-0500 or equiva-
lent). Attach to concentrator tube with springs. (Kontes
K-662750-0012).
4.4.3	Snyder colu»nn - three-ball macro (Kontes K503000-0121 or
equivalent).
4.4.4* Snyder column - two-ball micro (Kontes K-569001-0219 or
equivalent).
4.4.5 Boiling chips - solvent extracted, approximately 10/40 mssh.
4.5	Water bath - Heated, with concentric ring cover, capable of temper-
ature control (+ 2°C). The bath should be used 1n a hood.
4.6	Gas chromatogrcph - Analytical systan complete with gas ehromato-
graph suitable for on-column Injection and all required accessories
Including nitrogen-phosphorus or reductive Hall detector, column
supplies, recorder, gases, syringes. A data system for measuring
peak areas 1s recommended.
4.7	Chromatographic column - Pyrex (approximately 300 mm long x 10 mm
ID) with coarse fritted disc at bottom and Teflon stopcock (Kontes
K-420540-0213 or equivalent).
4.8	Chromatographic column - Pyrex (approximately 400 mm long x 22 mm
10) with coarse fritted disc at bottom and Teflon stopcock (Kontes
*
K-420540-0234 or equivalent).
5. Reagents
5.1 Preservatives:
5.1.1	Sodium hydroxide - (ACS) 10 N 1n distilled water.
5.1.2	Sulfuric acid - (ACS) Mix equal volumes of conc. HgSO^
with distilled water.
! ivs" i
i	j			j	
EPA Form 2350-4 (4-001
(PnCVIUUSLY CIN. CPA FORM 28V)
PAGE NUMBEH
BOTTOM OF
I MAG!: iREA:
4CUTSIDU
DIMENSION
FOR TABLES
_l AND ILLUS-
THATI VMS
6 us ijOvf>.NVErirrniMiinr; 07'C^ v.iec-cw-639
TYPING GUIDE SHE£T

' \

-------
•; .ip •
'A*
^ji-gin j
LAo r uirjL
or TEXT
5.1.3 Sodium thlosulfate - (ACS) Granular.
5.2	Methylene chloride - Pesticide quality or equivalent.
5.3	Sodium Sulfate - (ACS) Granular, anhydrous (purified by heating at
400°C for 4 hrs. 1n a shallow tray).
5.4	Stock standards - Prepare stock standard solutions at a concentra-
tion of 1.00 ug/ul by dissolving 0.100 grams of assayed reference
material 1n pesticide quality Isooctane or other appropriate
solvent and diluting to volume in a 100 ml ground glass stoppered
volumetric flask. The stock solution 1s transferred to ground
glass stoppered reagent bottles, stored irf a refrigerator, and
checked frequently for signs of degradation or evaporation,
especially just prior to preparing working standards from them.
5.5	Methyl alcohol, pentane, acetone - Pesticide quality or equivalent.
5.6	Diethyl Ether -Nanograde, redistilled 1n glass 1f necessary.
5.6.1	Must be free of peroxides as Indicated by EM Quant test
strips. (Test strips, are available from EM Laboratories,
Inc., 500 Executive Blvd., Elmsford, N.Y. 10523.)
5.6.2	Procedures recommended for removal of peroxides are provided
with the test strips. After cleanup, 20 ml ethyl alcohol
preservative must be added to each liter of ether.
5.7	Florisll - PR grade (60/100 mesh); purchase activated at 1250°F
and store 1n dark 1n glass containers with glass stoppers or
foil-lined screw caps. Before use, activate each batch at least 16
"i i
hours at 130°C 1n a foil covered glass container.
5.8	Alumina - Activity Super I, Basic, W200 series (ICN Life Sciences
Group., No. 404571).
	!	
llZL _ 1	
SPA rorm 2350-J (4-BOI
(.quVIOUSCV CIN. EPA I "OHM ZJV)
:::::-::::::-:-52 V:-;
I'AGE NUMSEFt
101 TOM Of-
I IMAGE1 Aft CM.
	irj OUTSJOE
^ diffusion
' FOR TABLCS
— I AMD ILLUS-
THATIONS
?IK» GU'.iH.WtM PfciJIttiti ,C£ «960-Si:o WV
TYPING GUIDE SHEET
_ tlfi.
'J

-------
e-'G'N
LAS'?
Of- TEXT j
I
i 5.9 Hydrochloric acl.d, 105C-(ACS) Add one volume of conc. HC1 to nine
I
i	volumes distilled water.
6.	Calibration
6.1	Prepare calibration standards that contain the compounds of
Interest, either singly or mixed together. The standards should be
t
prepared at concentrations covering two or more orders of magnitude
that will completely bracket the. working range of the chromato-
graphic system. If the sensitivity of the detection system can be
calculated from Table I as 100 ug/1 in the final extract, for
example, prepare standards at 10 ug/1, 50 ug/1, 100 ug/1, 500 ug/1,
etc. so that injections of 1-5 ul of each calibration standard will
define the linearity of the detector 1n the working range.
6.2	Assemble the necessary gas chromatographic apparatus and establish
operating parameters equivalent to those indicated in Table I. By j
Injecting calibration standards, establish the sensitivity limit of
the detector and the linear range of the analytical system for each
compound.
i
6.3	Before using any cleanup procedure, the analyst must process a	j
I
series of calibration standards through the procedure to validate	I
elutlon patterns and the absence of Interferences from the reagents.
7.	Quality Control	:
7.1 Before processing any samples, the analyst should demonstrate , !
through the analysis of a distilled water method blank, that all
glassware and reagents are Interference-free. Each time a set of
sanples 1s extracted or there 1s a change In reagents, a method
"	" 4 dui f'OrVI OF
IMAGP AHEA;
	:	ourstoc
k
3'8
.A	
S:;: 53;;'
"SS uUibTL-'l:
dimension
' f-GR TADLEii
EPA F.irm 21504 (4-W)
|PREV 'OUSL.Y CIN. F.PA rOHM 28>)
PAGE NUMBER
¦JSION
TADLES
J AND ILLUS-
• TI1ATIONS
6 US OUVIKNVCHT KiiNnuC. OfvliT l9SO-6i)0-fJ9
TYPING GUIDE SHEET
t IT,.

-------
				it'»«iM«tmujgap«558iBK2i3C5rSC!3$lS^2SatC!2SSflt.
\ I LH 'J>

."EG IN
L/.ST LI'-jCW
or
blank should be processed as a safeguard against chronic laboratory
contamination.
7.2 Standard quality assurance practices should be used with this
method. Field replicates should be collected to validate the
precision of the sampling technique. Laboratory replicates should
be analyzed to validate the precision of the analysis. Fortified
samples should be analyzed to validate the accuracy of the
analysis. Where doubt exists over the Identification of a peak on
the chromatogran, confirmatory techniques such as mass spectroscopy
should be used.
8. Sample Collection. Preservation, and Handling
8.1	Grab samples must be collected in glass containers. Conventional
sanpling practices should be followed, except that the bottle must
not be prewashed with sample before collection. Composite samples
should be collected 1n refrigerated glass containers 1n accordance
with the requirements of the program. Automatic sanpling equipment
must be free of tygon and other potential sources of contamination.
8.2	The sanples must be Iced or refrigerated from the time of collec-
tion until extraction. Chemical preservatives should not be used
in the field unless more than 24 hours will elapse before delivery
to the laboratory. If the samples will not be extracted within 48
hours of collection, they must be preserved as follows:
8.2.1	Add 35 mg of sodium thlosulfate per part per million of free
chlorine per liter of sample.
8.2.2	Adjust the pH of the water sample to pH 7 to 10 using sodium
hydroxide or sulfuric acid. Record the volume of add or,
base added.
\ 	
. L _¦
; • j 3/8-
i	j	—
EPA Form 2300-4 (4-8M .-
•(PSCVIOUSLV C1N. H.PA FOttM 287)
. PACSNUf/GEH
' O'-
er. TOM OF
flA-^E /- R E A;
mf'.F.NSlON
Hon TABLHS
I AfvO ILLUS-
TRATIONS -
A us crvf:i',M£'.,T miN/iNC; rncf.
1 TYPING GUIDE SHEET
pi
m


i/l'.ix ThVi1'. ~ rr f* n -r - v ¦T7 •'"

-------
]
C.GIN |
,iyr
F UX i
j 8.3 All samples must be extracted within 7 days and completely analyzed
within 30 days of collection.
\ 9. Sample Extraction
9.1	Mark the water meniscus on the side of the sample bottle for later
determination of sample volume. Pour the entire sample Into a
two-liter separatory funnel. Check the pH of the sample with
wide-range pH paper and adjust to within the range of 7 to 10 with
sodium hydroxide or sulfuric acid.
9.2	Add 60 ml methylene chloride to the sample bottle, saal, and shake
30 seconds to rinse the inner walls. Transfer the solvent into the
separatory funnel, and extract the sample by shaking the funnel for
two minutes with periodic venting to release vapor pressure. Allow
the organic l^yer to separate from the water phase for a minimum of
ten minutes. If the emulsion interface between layers 1s more than
one-third the size of the solvent layer, the analyst must employ
mechanical techniques to complete the phase separation. The
optimum technique depends upon the sample, but may include
Stirring, filtration of the anulslon through glass wool, or
centrlfugatlon. Collect the methylene chloride extract in a 250-ml
separatory funnel.
9.3	Add a second 60-ml volume of methylene chloride* to the sample
bottle and complete the extraction procedure a second time,
combining the extracts In the 250-ml separatory funnel.
9.4	Perform a third extraction 1n the same manner. Add 10 ml of 10%
HC1 solution to the combined extracts and shake for 2 minutes.
Allow the layers to separate. Drain the methylene chloride layer
£PA Fo.m 23'30-4 (4-80)
(PREVIOUSLY CIN. CPA POHM 287)

PAGE NUMBER
ii u '> Govt-fii.vi«r PhiNriNGrff.tf; iono i.jO cos
ITTO U OF
AREA.
Ol!T5!D£
TOItetiviSlGM
FOrt TAtiLtS
TAtiLtS
I AND 'IlUS-
TRATIONS
. TYPING GUIDE SHEET

-------
BBK0S9OT 0

through a drying column containing 3-4 Inches of anhydrous sodium
sulfate, and collect 1t in a 500-ml Kuderna-Danish (K-D) flask
equipped with a 10 nl concentrator tube. Rinse the column with
20-30 ml methylene chloride to complete the quantitative transfer.
9.5	Add 1-2 clean boiling chips to the flask and attach a three-ball
Snyder column. Prewet the Snyder column by adding about 1 ml
methylene chloride to the top. Place the K-D apparatus on a hot
water bath (60-65°C) so that the concentrator tube 1s partially
Irrmersed 1n the hot water, and the entire lower rounded surface or
the flask 1s bathed in vapor. Adjust the vertical position of the
apparatus and the water ten^erature as required to complete the
concentration In 15-20 minutes. At the proper rate of distillation
the balls of the column will actively chatter but the chambers will
not flood. Because of the volatility of N-n1trosod1methylam1ne,
K-0 concentration must be carefully carried out. When the apparent
volume of liquid reaches 1 ml, remove the K-0 apparatus and allow
it to drain for at least 10 minutes while cooling. Remove the
Snyder column and rinse the flask and Its lower joint Into the
concentrator tube with 1-2 ml of methylene chloride. A 5-ml
syringe Is recoosended for this operation. Unless the entire
extract will be subjected to a cleanup operation (Section 10),
adjust the extract volume to 10.0 ml with methylene chloride, add
stopper, and refrigerate.
9.6	If the sample 1s being analyzed for N-n1trosod1phenylam1ne, the
analyst, must 1nraed1ately proceed with one of the cleanup methods in
fl C:! '"'1	I	
AST i.lf.CEf-
E. vr T t	-
F- TCaT
U-
r"L !.
EPA Form L3WW 14-801
(FT^VIOUSLY CIN. LP A FORM 2fi7)
PACK NUMBER
-a Ol».OO^ri\M:N7 '-rt.'MiNG OhCG 1it6OW0-O3^
TYPING GUIDE SHEET .
r? -
f-i -
ti ••
3 -,*¦
i
	i lici tom or-
It."AG-. AREA:
YDIWCNSION
1 FCrt TABLtS
	I AND ILLUS-
TRATIONS

-------

C-.r. III! OP
iylamine Is of no interest, the analyst must
choose between proceeding directly to Section II, or submitting the
extract to a cleanup procedure before gas chromatography. A
solvent exchange from methylene chloride to methyl alcohol Is
required for direct gas chromatography. Once the entire extract is
in methyl alcohol 1t cannot be treated to either of the cleanup
procedures In Section 10. Therefore, In the absence of previous
experience with the sample matrix, the.analyst should remove a 2.0
ml aliquot of the extract for gas chromatography and retain the
remainder for cleanup if required later.
9.3 Determine the original sample volume by refilling the sample bottle
to the mark and transferring the liquid to a 1000 ml graduated
cylinder. Record the sample volume to the nearest S ml.
10. Cleanup and Separation
10.1 If the entire extract 1s to be cleaned up by one of the following
procedures, 1t must be concentrated to 2.0 ml. To the concentrator
tube 1n 9.S, Add a clean boiling chip and attach a two-ball
olcro-Snyder column Prewet the coluim by adding about 0.5 ml
methylene chloride to the top. Place the K-0 apparatus on a
steaming hot (60-65°C) water bath so that the concentrator tube
1s partially imnersed in the hot water. Adjust the vertical
				.. I.. —UU
J Ffil
3/8"
_... I_
EPA Form 2350-4 (4 801
(PPCVIOUCUY CIN. EPA FOBW 2«7>

PAGE NUMBER
us OOVCItWENf OO'NTI.nG OfflCE l9M-
-------
Kl
position of the apparatus and the water temperature as required to
complete the concentration 1n 5-10 minutes. At the proper rate of
distillation the balls of the column will actively chatter but the
chambers will not flood. When the apparent volume of liquid
reaches about 0.5 ml, remove the K-D and allow it to drain for at
least 10 minutes while cooling. Remove the micro-Snyder column and
rinse its lower joint Into the concentrator tube with 0.2 ml of
methylene chloride. Adjust the final volume to 2.0 ml and proceed
with one of the following cleanup procedures.
10.2 Flor1s11 Column Cleanup for N1trosam1nes
10.2.1	Place 22g of activated Florlsil In a 22 mm ID chromato-
graphic column. After settling-the Florlsil by tapping the
column, add about a 5 mm layer of anhydrous granular sodium
sulfate to the top.
10.2.2	Prtelute the column, after cooling, with 40 ml of 1555 -ethyl
ether/853, pentane. Discard the eluate and just prior to
exposure of the sodium sulfate layer to air, quantitatively
transfer 2.0 ml of sanple extract Into the column b> decan-
tatlon using an additional 2 ml of pentane to complete the
transfer.
10.2.3	Perform the first elutlon with 90 ml of 15SS ethyl ether/8555
pentane (V/V) and discard the eluate. This fraction will
contain any d1phenylamine.
10.2.4	Perform the second elutlon with 100 ml of 5% acetone/955
ethyl ether (V/V) and collect the eluate 1n a 500-ml K-D
. flask equipped with a ".0-ml concentrator tube. This
tet-LI:	\
QPA Form 23f>0-4 (4-80)
(PREVIOUSLY CIN. EPA FORM 287)
mx. 58:;
PAGE NUMBER
liOTTOM OF
- IMAGE AREA;
.0 JT^IDE;
^¦Llir/.LNSION
1 FOR TABLES
	I AND !LLUS-
TFIATIONC ¦
TVUS GCVUtNVi.Nf fPiKTiNG Off (C.h 1060 JxJO-WO
TYPING GUIDE SHEET
:-'4>

j.
¦ A. 'a a ll Adi rtfftu	:&V.

-------
' -1) L*H 0 s
'N
r li\:i
c-XT
i
?"r _i	
EPA Fonn 235C-4 (&Q0)
(fnEV»OU3LV CIN. CPA POHM 267)
PAGE NUMBER
TO-' >
A ....	I
fraction will contain all of the nltrosamines.
10.2.5	Add 15 ml of methanol to the collected ekate and concen-
trate as 1n 9.5 at 70-75°C, substituting pentane for
methylene chloride.
10.2.6	Analyze by gas chromatography.
10.3 Alunlna Column Cleanup for Nltrosamines
10.3.1	Place lOOg of alumina, as 1t comes from the manufacturer.
Into a 500 ml reagent bottle and add 2 ml of distilled
water, which 1s free of nltrosanrlnes and Interferences. Mix
the alunlna preparation thoroughly by shaking or rolling for
10 minutes and let It stand for at least 2 hours. The
preparation should be homogeneous before use. Keep the
bottle sealed tightly to ensure proper activity.
10.3.2	Place 12 grams of the alumina preparation Into a 10 nan 10
chromatographic column and tap the column to settle the
alumina. Add 1-2 cm of anhydrous sodium sulfate to the top
of the alunlna.
10.3.3	Preelute the column with 10 ml of 30£ ethyl ether/70%
pentane (V/V). Discard the eluate (about 2 ml) and, just
prior to exposure of the iicdium sulfate layer the air,
transfer 2.0 ml of sample jxtract onto the column by decan-
tatlon using an addition? i 2 ml of pentane to complete the
transfer.
10.3.4	Just prior to exposure of the sodium sulfate layer to the
air, add 70 ml of 30JS ethyl ether/70% pentane. Discard the
first 10 ml of eluate but collect the rest of the eluate 1n
&
[20TTOM OF
i'.'AGE AI1EA.
•V OUTSIDE
'.'""i'/.f N'JION
fOH TABLES
.1 AND ILLUS-
TRATIONS
iU5 r.OvCIlr.MfcHT Pflimnjc OlfiCt lS80-«0-f.3N
TYPING GUIDE SHEET
MM
'• >v'
U--t
M
SB?
•pi
m

iSf


-------
K,r";r
KCGIN
LAST I IMEB
or- 1 EXT •,
	—			I
I
a 500-mT <-0 flask equipped with a 10 ml concentrator tube.
This fraction contains N-n1trosodiphenylamine and probably a
small amount of N-n1trosod1-n-propylam1ne.
10.3.5	Next elute the column with 60 ml of .50% ethyl ether/50%
pentane, collecting the eluate. In > second K-0 flask
equipped with a 10 ml concentrator tube. Add 15 ml methyl
»
alcohol to the K-0. This fraction will contain
N-n1trosodircethylam1ne, most of the
N-n1trosod1-n-propy1am1ne and any dlphenylamlne.
10.3.6	Concentrate both fractions as In 9.5 substituting pentane
for methylene chloride.
10.3./ Analyze by gas chromatography.
11. 6as Chromatography
11.1	N-n1trosodiphenylamine conpletely reacts to form dlphenylamlne at
normal operating temperatures of the GC Injection port. There-
fore, N-n1trosod1pheny1am1ne Is actually chromatographed and
detected as dlphenylsnlne. The determination of either of the
confounds In the original sample would be uncertain without the
us0 of one of the previous cleanup procedures which separate the
two confounds.
11.2	Table I stznmarizes soma reconsnended gas chromatographic column
materials and operating conditions for the Instrument. Included
1n this table are estimated retention times and sensitivities that
should be achieved by this method. Examples of the separations
achieved by the primary column are shown 1n Figures 1 and 2.
Calibrate the systen dally with a minimum of three Injections' of
i
3'r;930-6-30-Co0
TYPING GUIDE SHEET
¦ ¦ .-.I-		.-A. ¦ J..... 1. —	"is-m
¦ isiit.r mrrahi'iWim
253

-------
i : :,r.
k ¦-./ r
dCGIN
LAST LINE .
Or «EXT
•_ >1 OP
iv '
I
TO? ;;.k
.it/, ¦: ^
calibration standards.
11.3	If the extract has not been submitted to one of the cleanup proce-
dures 1n Section 10, 1t 1s necessary to exchange the solvent from
methylene chloride to methyl alcohol before the thermionic detec-
tor can be used. To a 1-10 ml volume of methylene chloride
extract 1n a concentrrator tube, add 2 ml methyl alcohol, and a
clean boiling chip. Attach a two-ball mlcro-Snyder column.
Preset the column by adding about 0.5 ml methylene chloride
through the top. Place the K-0 apparatus on a boiling water bath
so that the concentrator tube 1s partially Immersed In the hot
water. Adjust the vertical position and Insulate the apparatus as
necessary to complete the concentration 1n 5-10 minutes. At the
proper rate of distillation the balls of the column will actively
chatter but the chambers will not flood. When the apparent voliane
of liquid reaches about 0.5 ml, remove the K-D and allow It to
drain for at least 10 minutes while cooling. Remove the
mlcro-Snyder column and rinse Its lower joint into the concen-
trator tube with 0.2 ml of methyl alcohol. Adjust the final
voliaae to 2.0 ml.
11.4	Inject 2-5 ul of the sample extract using the solvent-flush
f
technique. Ssnaller (1.0 ul) volumes can be Injected If automatic
devices are deployed. Record the volisne Injected to the nearest
0.05 ul, and the resulting peak size, 1n area units.
11.5	If the peak area exceeds the linear range of the system, dilute
the extract and reanalyze.
11.6	If the peak area measurement Is prevented by the presence of
¦ i
3/8'
I
i —
EPA Form 2350-4 (4-90)
(PRLVIOUSLV cm, CPA FORM 287)
	1	
mm*
PAGE NUMBER
|TrOM OF
AGE AKEA;
OUTSIDE
'DIMENSION
FOR TAHl.ES
AWO ILLUS-
TRATIONS

iVu3 gcvlw.'uemi '•pintingofr"ii't. isao-1
TYPING GUIDE SHEET

-------
l-'r
T
Cl 'X ' L r '.A'
""f	
Interferences, further cleanuo 1s required.
12.	Calculations
12.1	Determine the concentration of Individual compounds according to
the formula:
Concentration, ug/1 » (A) (B) (Vt)
(Vi) (Vs)
where A » Calibration factor for chromatographic systen, 1n
nanograms material per area unit.
B ¦ Peak size 1n Injection of sample extract, 1n area units
Vj ¦ volume of extract Injected (ul)
Vt ¦ Volume of total extract (ul)
Vs ¦ Volume of water extracted (ml)
12.2	Report results In micrograms per liter without correction for
recovery data. When duplicate and spiked samples are analyzed,
all data obtained should be reported.
13.	Accuracy and Precision
The U.S. EPA Environmental Monitoring and Support Laboratory 1n
Cincinnati 1s 1n the process of conducting an Interlaboratroy method
study to determine the accuracy and precision of this test procedure.
BIBLIOGRAPHY
"Development and Application of Test Procedures for Specific Organic Toxic
Substances 1n Wastewaters. Category 5 - Nltrosamlnes," Report for EPA
Contract 68-03-2606 (In preparation).
BEGIN
LAST U\"l
Or TCXT
EPA Form 2350-4 (4-80)
(PREVIOUSLY CIN/CPA FORM 2«7)
PACE NUM13ER
1

1 j
f"'s i lid
sm . ^
.J
BOTTOM OF
IMAGE AREA;
OUTSIDE
¦"DIMENSION
FOR TA3LES
AND ILLUS-
TBATJONS
£ US GOVFHN.VENT PRiMl t'JG C^MCE
TYPING GUIDE SHEET


-------
KL-Mrj
t-nr.:i
L!\f or ,
•u tit
CL:M Or
TO"1 nf
.6 IV* .u
L
TABLE I
GAS CHROMATOGRAPHY OF NITROSAMINES
Nltrosamine
Retention Time (m1n.)
Column 1 Column 2
N-n1trosod1dimethylamine	4.1
N-n1trosod1-n-d1propy1am1ne 12.1
N-n1trosod1d1phenylam1ne	12.8*
0.88
4.2
6.4**
Detection
Um1t (ug/1)
0.3
0.5
1.0
Column 1 conditions: Chromosorb WAW 80/100 mesh coated with 10% Carbowax
2CM/22 KOH packed 1n a 180 cm long x 4 mm ID glass column with helium
carrier gas at 40 ml/m1n flow rate. Isothermal column temperature 1s
110°C except where * indicates 220°C.
Column 2 conditions: Supelcoport 100/120 mesh coated with 102 SP-2250
* packed 1n a 180 cm long x 4 mm ID glass column with helium carrier gas
at 40 ml/mln flow rate. Isothermal column temperature 1s 120°C except
where ** Indicates 210°C.
Detection limit Is calculated from the mlnlmun detectable GC response being
equal to five times the GC background noise, assuming a 10 ml final
volume of the 1 liter sample extract, and assuming a GC Injection of 5
microliters. A nitrogen-phosphorus detector was used to collect this
data, but a Thermal Energy Analyzer exhibited equivalent sensitivity.
B'lGIW
L/.ST LINE
Cf T L X T
I
I.
'tl1-
EPA Form 2250 4 (4-80)
(PREVIOUSLY CIN. EPA FORM 267)
mmm
PAGE NUMBER
BOTTOM OF
IMAGE AREA;
OUTSIDE
DIMENSION
FOR TABLES
I AND ILLUS-
TRATIONS
.£U.S uOvE-RNMCNr PRJNTlNG OFFiCr 19aO-COO-63i>
TYPING GUIDE SHEET

-------
y S- GIW
' iss r
1.1.oi . j—
r.-A'r ^r-
"i ^r-o j
i C'.tn-
;£C'. ;ONS
,:t
"seSr —
\ t-
S6C-IN
LAST LINECt*-
OF TEXT
I
I.
II'"
T
Ci.-wri-'R 01-
C r
TOP OF
Ja.J 0:G
COLUMN: 10% CARBOWAX 20IU + 2% KOH ON CHR0M0S0R8 W-AW
TEMPERATURE: 110*
DETECTOR: PHOSPHORUS/NITROGEN
Ol
Z
ui
S
'*3
8
o
ee
2
UJ
2
3
<
s
a.
f
O
VI
2 4 I 8 10 12 14
RETENTION TIME-MINUTES
Figure 1. Gas chromatograir of nitrosamines

L,
EPA Form 2350-4 (4-80) '
¦;P»EVIOL!Sl-Y C1N. EPA rORM 287)
	64
PAGE NUMBER
— Ee»
BOTTOM OF
IMAGE AREA;
OUTSIDt
'DIMENSION
FOR TABLES
__l AND ILL'JS- ¦,
TRATIONS

" V: U.GOVEINM'.-Nf PRINT MG OFPiCf- '9S0 660-630
TYPING guide sheet
IMC



-------
i3
•itCIN
r I R.ST
H'.fiL
r:_AU,
^r-ou'oni
CI:.\Tr» OF
i Ac.e

GEC-IN
'_A5T UNC>4>-
OF TeXT
"1"
A

COLUMN: 10% CARBOWAX 20M + 2% KOH ON CHROMOSORB V»-AW
TEMPERATURE: 220°C.
DETECTOR: PHOSPHORUS/NITROGEN
TO? Or
jl.	/
— E2-

0 2 4 6 8 10 12 14 16 18
RETENTION TIME-MINUTES
Figure 2. Gas chromatogram of N-nitrosodiphenylamine
as diphenylamine
* 3/8"
F.PA Form 2350-4 (4-80)
(Pft l£ V iOU^UY CIN. fc'PA r-OflM
iiSIS
PAGE NUMBER
.. BOTTOM OF '
| IMAGE AREA:
¦J OUTSIDE
DIMENSION
FOR TABLES
.I. AND ILLUS-
, TRATIONS
n t.\s Govr.RNMtNi tTiujiiiNa Office lseo-teo^M
TYPING GUIDE SHEET


-------